Index: projects/building-blocks/Makefile.inc1 =================================================================== --- projects/building-blocks/Makefile.inc1 (revision 275663) +++ projects/building-blocks/Makefile.inc1 (revision 275664) @@ -1,2141 +1,2146 @@ # # $FreeBSD$ # # Make command line options: # -DNO_CLEANDIR run ${MAKE} clean, instead of ${MAKE} cleandir # -DNO_CLEAN do not clean at all # -DDB_FROM_SRC use the user/group databases in src/etc instead of # the system database when installing. # -DNO_SHARE do not go into share subdir # -DKERNFAST define NO_KERNEL{CONFIG,CLEAN,DEPEND,OBJ} # -DNO_KERNELCONFIG do not run config in ${MAKE} buildkernel # -DNO_KERNELCLEAN do not run ${MAKE} clean in ${MAKE} buildkernel # -DNO_KERNELDEPEND do not run ${MAKE} depend in ${MAKE} buildkernel # -DNO_KERNELOBJ do not run ${MAKE} obj in ${MAKE} buildkernel # -DNO_PORTSUPDATE do not update ports in ${MAKE} update # -DNO_ROOT install without using root privilege # -DNO_DOCUPDATE do not update doc in ${MAKE} update # -DWITHOUT_CTF do not run the DTrace CTF conversion tools on built objects # LOCAL_DIRS="list of dirs" to add additional dirs to the SUBDIR list # LOCAL_ITOOLS="list of tools" to add additional tools to the ITOOLS list # LOCAL_LIB_DIRS="list of dirs" to add additional dirs to libraries target # LOCAL_MTREE="list of mtree files" to process to allow local directories # to be created before files are installed # LOCAL_TOOL_DIRS="list of dirs" to add additional dirs to the build-tools # list # METALOG="path to metadata log" to write permission and ownership # when NO_ROOT is set. (default: ${DESTDIR}/METALOG) # TARGET="machine" to crossbuild world for a different machine type # TARGET_ARCH= may be required when a TARGET supports multiple endians # BUILDENV_SHELL= shell to launch for the buildenv target (def:/bin/sh) # WORLD_FLAGS= additional flags to pass to make(1) during buildworld # KERNEL_FLAGS= additional flags to pass to make(1) during buildkernel # # The intended user-driven targets are: # buildworld - rebuild *everything*, including glue to help do upgrades # installworld- install everything built by "buildworld" # doxygen - build API documentation of the kernel # update - convenient way to update your source tree (eg: svn/svnup) # # Standard targets (not defined here) are documented in the makefiles in # /usr/share/mk. These include: # obj depend all install clean cleandepend cleanobj .if !defined(TARGET) || !defined(TARGET_ARCH) .error "Both TARGET and TARGET_ARCH must be defined." .endif .include "share/mk/src.opts.mk" .include .include # We must do share/info early so that installation of info `dir' # entries works correctly. Do it first since it is less likely to # grow dependencies on include and lib than vice versa. # # We must do lib/ and libexec/ before bin/, because if installworld # installs a new /bin/sh, the 'make' command will *immediately* # use that new version. And the new (dynamically-linked) /bin/sh # will expect to find appropriate libraries in /lib and /libexec. # SRCDIR?= ${.CURDIR} .if defined(SUBDIR_OVERRIDE) SUBDIR= ${SUBDIR_OVERRIDE} .else SUBDIR= share/info lib libexec SUBDIR+=bin .if ${MK_GAMES} != "no" SUBDIR+=games .endif .if ${MK_CDDL} != "no" SUBDIR+=cddl .endif SUBDIR+=gnu include .if ${MK_KERBEROS} != "no" SUBDIR+=kerberos5 .endif .if ${MK_RESCUE} != "no" SUBDIR+=rescue .endif SUBDIR+=sbin .if ${MK_CRYPT} != "no" SUBDIR+=secure .endif .if !defined(NO_SHARE) SUBDIR+=share .endif SUBDIR+=sys usr.bin usr.sbin .if ${MK_TESTS} != "no" SUBDIR+= tests .endif .if ${MK_OFED} != "no" SUBDIR+=contrib/ofed .endif # # We must do etc/ last for install/distribute to work. # SUBDIR+=etc # Local directories are last, since it is nice to at least get the base # system rebuilt before you do them. .for _DIR in ${LOCAL_DIRS} .if exists(${.CURDIR}/${_DIR}/Makefile) SUBDIR+= ${_DIR} .endif .endfor # Add LOCAL_LIB_DIRS, but only if they will not be picked up as a SUBDIR # of a LOCAL_DIRS directory. This allows LOCAL_DIRS=foo and # LOCAL_LIB_DIRS=foo/lib to behave as expected. .for _DIR in ${LOCAL_DIRS:M*/} ${LOCAL_DIRS:N*/:S|$|/|} _REDUNDENT_LIB_DIRS+= ${LOCAL_LIB_DIRS:M${_DIR}*} .endfor .for _DIR in ${LOCAL_LIB_DIRS} .if empty(_REDUNDENT_LIB_DIRS:M${_DIR}) && exists(${.CURDIR}/${_DIR}/Makefile) SUBDIR+= ${_DIR} .endif .endfor .endif .if defined(NOCLEAN) NO_CLEAN= ${NOCLEAN} .endif .if defined(NO_CLEANDIR) CLEANDIR= clean cleandepend .else CLEANDIR= cleandir .endif LOCAL_TOOL_DIRS?= BUILDENV_SHELL?=/bin/sh SVN?= /usr/local/bin/svn SVNFLAGS?= -r HEAD MAKEOBJDIRPREFIX?= /usr/obj .if !defined(OSRELDATE) .if exists(/usr/include/osreldate.h) OSRELDATE!= awk '/^\#define[[:space:]]*__FreeBSD_version/ { print $$3 }' \ /usr/include/osreldate.h .else OSRELDATE= 0 .endif .endif .if !defined(VERSION) REVISION!= ${MAKE} -C ${SRCDIR}/release -V REVISION BRANCH!= ${MAKE} -C ${SRCDIR}/release -V BRANCH SRCRELDATE!= awk '/^\#define[[:space:]]*__FreeBSD_version/ { print $$3 }' \ ${SRCDIR}/sys/sys/param.h VERSION= FreeBSD ${REVISION}-${BRANCH:C/-p[0-9]+$//} ${TARGET_ARCH} ${SRCRELDATE} .endif KNOWN_ARCHES?= amd64 arm armeb/arm armv6/arm armv6hf/arm i386 i386/pc98 mips mipsel/mips mips64el/mips mips64/mips mipsn32el/mips mipsn32/mips powerpc powerpc64/powerpc sparc64 .if ${TARGET} == ${TARGET_ARCH} _t= ${TARGET} .else _t= ${TARGET_ARCH}/${TARGET} .endif .for _t in ${_t} .if empty(KNOWN_ARCHES:M${_t}) .error Unknown target ${TARGET_ARCH}:${TARGET}. .endif .endfor .if ${TARGET} == ${MACHINE} TARGET_CPUTYPE?=${CPUTYPE} .else TARGET_CPUTYPE?= .endif .if !empty(TARGET_CPUTYPE) _TARGET_CPUTYPE=${TARGET_CPUTYPE} .else _TARGET_CPUTYPE=dummy .endif _CPUTYPE!= MAKEFLAGS= CPUTYPE=${_TARGET_CPUTYPE} ${MAKE} \ -f /dev/null -m ${.CURDIR}/share/mk -V CPUTYPE .if ${_CPUTYPE} != ${_TARGET_CPUTYPE} .error CPUTYPE global should be set with ?=. .endif .if make(buildworld) BUILD_ARCH!= uname -p .if ${MACHINE_ARCH} != ${BUILD_ARCH} .error To cross-build, set TARGET_ARCH. .endif .endif .if ${MACHINE} == ${TARGET} && ${MACHINE_ARCH} == ${TARGET_ARCH} && !defined(CROSS_BUILD_TESTING) OBJTREE= ${MAKEOBJDIRPREFIX} .else OBJTREE= ${MAKEOBJDIRPREFIX}/${TARGET}.${TARGET_ARCH} .endif WORLDTMP= ${OBJTREE}${.CURDIR}/tmp # /usr/games added for fortune which depend on strfile BPATH= ${WORLDTMP}/legacy/usr/sbin:${WORLDTMP}/legacy/usr/bin:${WORLDTMP}/legacy/usr/games:${WORLDTMP}/legacy/bin XPATH= ${WORLDTMP}/usr/sbin:${WORLDTMP}/usr/bin:${WORLDTMP}/usr/games STRICTTMPPATH= ${BPATH}:${XPATH} TMPPATH= ${STRICTTMPPATH}:${PATH} # # Avoid running mktemp(1) unless actually needed. # It may not be functional, e.g., due to new ABI # when in the middle of installing over this system. # .if make(distributeworld) || make(installworld) INSTALLTMP!= /usr/bin/mktemp -d -u -t install .endif # # Building a world goes through the following stages # # 1. legacy stage [BMAKE] # This stage is responsible for creating compatibility # shims that are needed by the bootstrap-tools, # build-tools and cross-tools stages. # 1. bootstrap-tools stage [BMAKE] # This stage is responsible for creating programs that # are needed for backward compatibility reasons. They # are not built as cross-tools. # 2. build-tools stage [TMAKE] # This stage is responsible for creating the object # tree and building any tools that are needed during # the build process. # 3. cross-tools stage [XMAKE] # This stage is responsible for creating any tools that # are needed for cross-builds. A cross-compiler is one # of them. # 4. world stage [WMAKE] # This stage actually builds the world. # 5. install stage (optional) [IMAKE] # This stage installs a previously built world. # BOOTSTRAPPING?= 0 # Common environment for world related stages CROSSENV= MAKEOBJDIRPREFIX=${OBJTREE} \ MACHINE_ARCH=${TARGET_ARCH} \ MACHINE=${TARGET} \ CPUTYPE=${TARGET_CPUTYPE} .if ${MK_GROFF} != "no" CROSSENV+= GROFF_BIN_PATH=${WORLDTMP}/legacy/usr/bin \ GROFF_FONT_PATH=${WORLDTMP}/legacy/usr/share/groff_font \ GROFF_TMAC_PATH=${WORLDTMP}/legacy/usr/share/tmac .endif .if defined(TARGET_CFLAGS) CROSSENV+= ${TARGET_CFLAGS} .endif # bootstrap-tools stage BMAKEENV= INSTALL="sh ${.CURDIR}/tools/install.sh" \ PATH=${BPATH}:${PATH} \ WORLDTMP=${WORLDTMP} \ VERSION="${VERSION}" \ MAKEFLAGS="-m ${.CURDIR}/tools/build/mk ${.MAKEFLAGS}" BMAKE= MAKEOBJDIRPREFIX=${WORLDTMP} \ ${BMAKEENV} ${MAKE} ${WORLD_FLAGS} -f Makefile.inc1 \ DESTDIR= \ BOOTSTRAPPING=${OSRELDATE} \ SSP_CFLAGS= \ MK_HTML=no MK_INFO=no NO_LINT=yes MK_MAN=no \ -DNO_PIC MK_PROFILE=no -DNO_SHARED \ -DNO_CPU_CFLAGS MK_WARNS=no MK_CTF=no \ MK_CLANG_FULL=no MK_LLDB=no MK_TESTS=no # build-tools stage TMAKE= MAKEOBJDIRPREFIX=${OBJTREE} \ ${BMAKEENV} ${MAKE} ${WORLD_FLAGS} -f Makefile.inc1 \ TARGET=${TARGET} TARGET_ARCH=${TARGET_ARCH} \ DESTDIR= \ BOOTSTRAPPING=${OSRELDATE} \ SSP_CFLAGS= \ -DNO_LINT \ -DNO_CPU_CFLAGS \ MK_WARNS=no MK_CTF=no MK_CLANG_FULL=no MK_LLDB=no MK_MAN=no \ MK_SHAREDOCS=no MK_TESTS=no # cross-tools stage XMAKE= TOOLS_PREFIX=${WORLDTMP} ${BMAKE} \ TARGET=${TARGET} TARGET_ARCH=${TARGET_ARCH} \ MK_GDB=no MK_TESTS=no # kernel-tools stage KTMAKEENV= INSTALL="sh ${.CURDIR}/tools/install.sh" \ PATH=${BPATH}:${PATH} \ WORLDTMP=${WORLDTMP} \ VERSION="${VERSION}" KTMAKE= TOOLS_PREFIX=${WORLDTMP} MAKEOBJDIRPREFIX=${WORLDTMP} \ ${KTMAKEENV} ${MAKE} ${WORLD_FLAGS} -f Makefile.inc1 \ DESTDIR= \ BOOTSTRAPPING=${OSRELDATE} \ SSP_CFLAGS= \ MK_HTML=no MK_INFO=no -DNO_LINT MK_MAN=no \ -DNO_PIC MK_PROFILE=no -DNO_SHARED \ -DNO_CPU_CFLAGS MK_WARNS=no MK_CTF=no # world stage WMAKEENV= ${CROSSENV} \ _SHLIBDIRPREFIX=${WORLDTMP} \ _LDSCRIPTROOT= \ VERSION="${VERSION}" \ INSTALL="sh ${.CURDIR}/tools/install.sh" \ PATH=${TMPPATH} # make hierarchy HMAKE= PATH=${TMPPATH} ${MAKE} LOCAL_MTREE=${LOCAL_MTREE:Q} .if defined(NO_ROOT) HMAKE+= PATH=${TMPPATH} METALOG=${METALOG} -DNO_ROOT .endif .if ${MK_CDDL} == "no" WMAKEENV+= MK_CTF=no .endif .if defined(CROSS_TOOLCHAIN) LOCALBASE?= /usr/local .include "${LOCALBASE}/share/toolchains/${CROSS_TOOLCHAIN}.mk" .endif .if defined(CROSS_TOOLCHAIN_PREFIX) CROSS_COMPILER_PREFIX?=${CROSS_TOOLCHAIN_PREFIX} CROSS_BINUTILS_PREFIX?=${CROSS_TOOLCHAIN_PREFIX} .endif XCOMPILERS= CC CXX CPP .for COMPILER in ${XCOMPILERS} .if defined(CROSS_COMPILER_PREFIX) X${COMPILER}?= ${CROSS_COMPILER_PREFIX}${${COMPILER}} .else X${COMPILER}?= ${${COMPILER}} .endif .endfor XBINUTILS= AS AR LD NM OBJCOPY OBJDUMP RANLIB SIZE STRINGS .for BINUTIL in ${XBINUTILS} .if defined(CROSS_BINUTILS_PREFIX) X${BINUTIL}?= ${CROSS_BINUTILS_PREFIX}${${BINUTIL}} .else X${BINUTIL}?= ${${BINUTIL}} .endif .endfor WMAKEENV+= CC="${XCC} ${XCFLAGS}" CXX="${XCXX} ${XCFLAGS} ${XCXXFLAGS}" \ DEPFLAGS="${DEPFLAGS}" \ CPP="${XCPP} ${XCFLAGS}" \ AS="${XAS}" AR="${XAR}" LD="${XLD}" NM=${XNM} \ OBJDUMP=${XOBJDUMP} OBJCOPY="${XOBJCOPY}" \ RANLIB=${XRANLIB} STRINGS=${XSTRINGS} \ SIZE="${XSIZE}" .if ${XCC:M/*} XFLAGS= --sysroot=${WORLDTMP} .if defined(CROSS_BINUTILS_PREFIX) # In the case of xdev-build tools, CROSS_BINUTILS_PREFIX won't be a # directory, but the compiler will look in the right place for it's # tools so we don't need to tell it where to look. .if exists(${CROSS_BINUTILS_PREFIX}) XFLAGS+= -B${CROSS_BINUTILS_PREFIX} .endif .else XFLAGS+= -B${WORLDTMP}/usr/bin .endif .if ${TARGET} == "arm" .if ${TARGET_ARCH:M*hf*} != "" TARGET_ABI= gnueabihf .else TARGET_ABI= gnueabi .endif .endif .if defined(X_COMPILER_TYPE) && ${X_COMPILER_TYPE} == gcc XCFLAGS+= -isystem ${WORLDTMP}/usr/include -L${WORLDTMP}/usr/lib XCXXFLAGS+= -I${WORLDTMP}/usr/include/c++/v1 -std=gnu++11 -L${WORLDTMP}/../lib/libc++ DEPFLAGS+= -I${WORLDTMP}/usr/include/c++/v1 .else TARGET_ABI?= unknown TARGET_TRIPLE?= ${TARGET_ARCH:C/amd64/x86_64/}-${TARGET_ABI}-freebsd11.0 XCFLAGS+= -target ${TARGET_TRIPLE} .endif .endif WMAKE= ${WMAKEENV} ${MAKE} ${WORLD_FLAGS} -f Makefile.inc1 DESTDIR=${WORLDTMP} .if ${TARGET_ARCH} == "amd64" || ${TARGET_ARCH} == "powerpc64" # 32 bit world LIB32_OBJTREE= ${OBJTREE}${.CURDIR}/world32 LIB32TMP= ${OBJTREE}${.CURDIR}/lib32 .if ${TARGET_ARCH} == "amd64" .if empty(TARGET_CPUTYPE) LIB32CPUFLAGS= -march=i686 -mmmx -msse -msse2 .else LIB32CPUFLAGS= -march=${TARGET_CPUTYPE} .endif LIB32WMAKEENV= MACHINE=i386 MACHINE_ARCH=i386 \ MACHINE_CPU="i686 mmx sse sse2" LIB32WMAKEFLAGS= \ AS="${AS} --32" \ LD="${LD} -m elf_i386_fbsd -Y P,${LIB32TMP}/usr/lib32" .elif ${TARGET_ARCH} == "powerpc64" .if empty(TARGET_CPUTYPE) LIB32CPUFLAGS= -mcpu=powerpc .else LIB32CPUFLAGS= -mcpu=${TARGET_CPUTYPE} .endif LIB32WMAKEENV= MACHINE=powerpc MACHINE_ARCH=powerpc LIB32WMAKEFLAGS= \ LD="${LD} -m elf32ppc_fbsd" .endif LIB32FLAGS= -m32 ${LIB32CPUFLAGS} -DCOMPAT_32BIT \ -isystem ${LIB32TMP}/usr/include/ \ -L${LIB32TMP}/usr/lib32 \ -B${LIB32TMP}/usr/lib32 .if ${XCC:M/*} LIB32FLAGS+= --sysroot=${WORLDTMP} .endif # Yes, the flags are redundant. LIB32WMAKEENV+= MAKEOBJDIRPREFIX=${LIB32_OBJTREE} \ _SHLIBDIRPREFIX=${LIB32TMP} \ _LDSCRIPTROOT=${LIB32TMP} \ VERSION="${VERSION}" \ INSTALL="sh ${.CURDIR}/tools/install.sh" \ PATH=${TMPPATH} \ LIBDIR=/usr/lib32 \ SHLIBDIR=/usr/lib32 \ LIBPRIVATEDIR=/usr/lib32/private \ DTRACE="${DTRACE} -32" LIB32WMAKEFLAGS+= CC="${XCC} ${LIB32FLAGS}" \ CXX="${XCXX} ${LIB32FLAGS}" \ DESTDIR=${LIB32TMP} \ -DCOMPAT_32BIT \ -DLIBRARIES_ONLY \ -DNO_CPU_CFLAGS \ MK_CTF=no \ -DNO_LINT \ MK_TESTS=no LIB32WMAKE= ${LIB32WMAKEENV} ${MAKE} ${LIB32WMAKEFLAGS} \ MK_MAN=no MK_INFO=no MK_HTML=no MK_HYPERV=no MK_BHYVE=no LIB32IMAKE= ${LIB32WMAKE:NINSTALL=*:NDESTDIR=*:N_LDSCRIPTROOT=*} \ MK_TOOLCHAIN=no ${IMAKE_INSTALL} .endif IMAKEENV= ${CROSSENV:N_LDSCRIPTROOT=*} IMAKE= ${IMAKEENV} ${MAKE} -f Makefile.inc1 \ ${IMAKE_INSTALL} ${IMAKE_MTREE} .if empty(.MAKEFLAGS:M-n) IMAKEENV+= PATH=${STRICTTMPPATH}:${INSTALLTMP} \ LD_LIBRARY_PATH=${INSTALLTMP} \ PATH_LOCALE=${INSTALLTMP}/locale IMAKE+= __MAKE_SHELL=${INSTALLTMP}/sh .else IMAKEENV+= PATH=${TMPPATH}:${INSTALLTMP} .endif .if defined(DB_FROM_SRC) INSTALLFLAGS+= -N ${.CURDIR}/etc MTREEFLAGS+= -N ${.CURDIR}/etc .endif _INSTALL_DDIR= ${DESTDIR}/${DISTDIR} INSTALL_DDIR= ${_INSTALL_DDIR:S://:/:g:C:/$::} .if defined(NO_ROOT) METALOG?= ${DESTDIR}/${DISTDIR}/METALOG IMAKE+= -DNO_ROOT METALOG=${METALOG} INSTALLFLAGS+= -U -M ${METALOG} -D ${INSTALL_DDIR} MTREEFLAGS+= -W .endif .if defined(DB_FROM_SRC) || defined(NO_ROOT) IMAKE_INSTALL= INSTALL="install ${INSTALLFLAGS}" IMAKE_MTREE= MTREE_CMD="mtree ${MTREEFLAGS}" .endif # kernel stage KMAKEENV= ${WMAKEENV} KMAKE= ${KMAKEENV} ${MAKE} ${.MAKEFLAGS} ${KERNEL_FLAGS} KERNEL=${INSTKERNNAME} # # buildworld # # Attempt to rebuild the entire system, with reasonable chance of # success, regardless of how old your existing system is. # _worldtmp: .if ${.CURDIR:C/[^,]//g} != "" # The m4 build of sendmail files doesn't like it if ',' is used # anywhere in the path of it's files. @echo @echo "*** Error: path to source tree contains a comma ','" @echo false .endif @echo @echo "--------------------------------------------------------------" @echo ">>> Rebuilding the temporary build tree" @echo "--------------------------------------------------------------" .if !defined(NO_CLEAN) rm -rf ${WORLDTMP} .if defined(LIB32TMP) rm -rf ${LIB32TMP} .endif .else rm -rf ${WORLDTMP}/legacy/usr/include # XXX - These three can depend on any header file. rm -f ${OBJTREE}${.CURDIR}/usr.bin/kdump/ioctl.c rm -f ${OBJTREE}${.CURDIR}/usr.bin/kdump/kdump_subr.c rm -f ${OBJTREE}${.CURDIR}/usr.bin/truss/ioctl.c .endif .for _dir in \ lib usr legacy/bin legacy/usr mkdir -p ${WORLDTMP}/${_dir} .endfor mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${WORLDTMP}/legacy/usr >/dev/null .if ${MK_GROFF} != "no" mtree -deU -f ${.CURDIR}/etc/mtree/BSD.groff.dist \ -p ${WORLDTMP}/legacy/usr >/dev/null .endif mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${WORLDTMP}/usr >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.include.dist \ -p ${WORLDTMP}/usr/include >/dev/null ln -sf ${.CURDIR}/sys ${WORLDTMP} .if ${MK_DEBUG_FILES} != "no" # We could instead disable debug files for these build stages mtree -deU -f ${.CURDIR}/etc/mtree/BSD.debug.dist \ -p ${WORLDTMP}/legacy/usr/lib >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.debug.dist \ -p ${WORLDTMP}/usr/lib >/dev/null .endif .if ${MK_TESTS} != "no" mtree -deU -f ${.CURDIR}/etc/mtree/BSD.tests.dist \ -p ${WORLDTMP}/usr >/dev/null .endif .for _mtree in ${LOCAL_MTREE} mtree -deU -f ${.CURDIR}/${_mtree} -p ${WORLDTMP} > /dev/null .endfor _legacy: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 1.1: legacy release compatibility shims" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${BMAKE} legacy _bootstrap-tools: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 1.2: bootstrap tools" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${BMAKE} bootstrap-tools _cleanobj: .if !defined(NO_CLEAN) @echo @echo "--------------------------------------------------------------" @echo ">>> stage 2.1: cleaning up the object tree" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${WMAKE} ${CLEANDIR:S/^/par-/} .if defined(LIB32TMP) ${_+_}cd ${.CURDIR}; ${LIB32WMAKE} -f Makefile.inc1 ${CLEANDIR:S/^/par-/} .endif .endif _obj: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 2.2: rebuilding the object tree" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${WMAKE} par-obj _build-tools: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 2.3: build tools" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${TMAKE} build-tools _cross-tools: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 3: cross tools" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${XMAKE} cross-tools ${_+_}cd ${.CURDIR}; ${XMAKE} kernel-tools _includes: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 4.1: building includes" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${WMAKE} SHARED=symlinks par-includes _libraries: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 4.2: building libraries" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; \ ${WMAKE} -DNO_FSCHG MK_HTML=no MK_INFO=no -DNO_LINT MK_MAN=no \ MK_PROFILE=no MK_TESTS=no MK_TESTS_SUPPORT=${MK_TESTS} libraries _depend: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 4.3: make dependencies" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${WMAKE} par-depend everything: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 4.4: building everything" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${WMAKE} par-all .if defined(LIB32TMP) build32: @echo @echo "--------------------------------------------------------------" @echo ">>> stage 5.1: building 32 bit shim libraries" @echo "--------------------------------------------------------------" mkdir -p ${LIB32TMP}/usr/include mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${LIB32TMP}/usr >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.include.dist \ -p ${LIB32TMP}/usr/include >/dev/null .if ${MK_DEBUG_FILES} != "no" mtree -deU -f ${.CURDIR}/etc/mtree/BSD.debug.dist \ -p ${LIB32TMP}/usr/lib >/dev/null .endif mkdir -p ${WORLDTMP} ln -sf ${.CURDIR}/sys ${WORLDTMP} .for _t in obj includes cd ${.CURDIR}/include; ${LIB32WMAKE} DIRPRFX=include/ ${_t} cd ${.CURDIR}/lib; ${LIB32WMAKE} DIRPRFX=lib/ ${_t} .if ${MK_CDDL} != "no" cd ${.CURDIR}/cddl/lib; ${LIB32WMAKE} DIRPRFX=cddl/lib/ ${_t} .endif cd ${.CURDIR}/gnu/lib; ${LIB32WMAKE} DIRPRFX=gnu/lib/ ${_t} .if ${MK_CRYPT} != "no" cd ${.CURDIR}/secure/lib; ${LIB32WMAKE} DIRPRFX=secure/lib/ ${_t} .endif .if ${MK_KERBEROS} != "no" cd ${.CURDIR}/kerberos5/lib; ${LIB32WMAKE} DIRPRFX=kerberos5/lib ${_t} .endif .endfor .for _dir in usr.bin/lex/lib cd ${.CURDIR}/${_dir}; ${LIB32WMAKE} DIRPRFX=${_dir}/ obj .endfor .for _dir in lib/ncurses/ncurses lib/ncurses/ncursesw lib/libmagic cd ${.CURDIR}/${_dir}; \ WORLDTMP=${WORLDTMP} \ MAKEFLAGS="-m ${.CURDIR}/tools/build/mk ${.MAKEFLAGS}" \ MAKEOBJDIRPREFIX=${LIB32_OBJTREE} ${MAKE} SSP_CFLAGS= DESTDIR= \ DIRPRFX=${_dir}/ -DNO_LINT -DNO_CPU_CFLAGS MK_WARNS=no MK_CTF=no \ build-tools .endfor cd ${.CURDIR}; \ ${LIB32WMAKE} -f Makefile.inc1 libraries .for _t in obj depend all cd ${.CURDIR}/libexec/rtld-elf; PROG=ld-elf32.so.1 ${LIB32WMAKE} \ DIRPRFX=libexec/rtld-elf/ ${_t} cd ${.CURDIR}/usr.bin/ldd; PROG=ldd32 ${LIB32WMAKE} \ DIRPRFX=usr.bin/ldd ${_t} .endfor distribute32 install32: cd ${.CURDIR}/lib; ${LIB32IMAKE} ${.TARGET:S/32$//} .if ${MK_CDDL} != "no" cd ${.CURDIR}/cddl/lib; ${LIB32IMAKE} ${.TARGET:S/32$//} .endif cd ${.CURDIR}/gnu/lib; ${LIB32IMAKE} ${.TARGET:S/32$//} .if ${MK_CRYPT} != "no" cd ${.CURDIR}/secure/lib; ${LIB32IMAKE} ${.TARGET:S/32$//} .endif .if ${MK_KERBEROS} != "no" cd ${.CURDIR}/kerberos5/lib; ${LIB32IMAKE} ${.TARGET:S/32$//} .endif cd ${.CURDIR}/libexec/rtld-elf; \ PROG=ld-elf32.so.1 ${LIB32IMAKE} ${.TARGET:S/32$//} cd ${.CURDIR}/usr.bin/ldd; PROG=ldd32 ${LIB32IMAKE} ${.TARGET:S/32$//} .endif WMAKE_TGTS= .if !defined(SUBDIR_OVERRIDE) WMAKE_TGTS+= _worldtmp _legacy _bootstrap-tools .endif WMAKE_TGTS+= _cleanobj _obj _build-tools .if !defined(SUBDIR_OVERRIDE) WMAKE_TGTS+= _cross-tools .endif WMAKE_TGTS+= _includes _libraries _depend everything .if defined(LIB32TMP) && ${MK_LIB32} != "no" WMAKE_TGTS+= build32 .endif buildworld: buildworld_prologue ${WMAKE_TGTS} buildworld_epilogue .ORDER: buildworld_prologue ${WMAKE_TGTS} buildworld_epilogue buildworld_prologue: @echo "--------------------------------------------------------------" @echo ">>> World build started on `LC_ALL=C date`" @echo "--------------------------------------------------------------" buildworld_epilogue: @echo @echo "--------------------------------------------------------------" @echo ">>> World build completed on `LC_ALL=C date`" @echo "--------------------------------------------------------------" # # We need to have this as a target because the indirection between Makefile # and Makefile.inc1 causes the correct PATH to be used, rather than a # modification of the current environment's PATH. In addition, we need # to quote multiword values. # buildenvvars: @echo ${WMAKEENV:Q} .if ${.TARGETS:Mbuildenv} .if ${.MAKEFLAGS:M-j} .error The buildenv target is incompatible with -j .endif .endif buildenv: @echo Entering world for ${TARGET_ARCH}:${TARGET} @cd ${.CURDIR} && env ${WMAKEENV} ${BUILDENV_SHELL} || true TOOLCHAIN_TGTS= ${WMAKE_TGTS:N_depend:Neverything:Nbuild32} toolchain: ${TOOLCHAIN_TGTS} kernel-toolchain: ${TOOLCHAIN_TGTS:N_includes:N_libraries} # # installcheck # # Checks to be sure system is ready for installworld/installkernel. # installcheck: _installcheck_world _installcheck_kernel _installcheck_world: _installcheck_kernel: # # Require DESTDIR to be set if installing for a different architecture or # using the user/group database in the source tree. # .if ${TARGET_ARCH} != ${MACHINE_ARCH} || ${TARGET} != ${MACHINE} || \ defined(DB_FROM_SRC) .if !make(distributeworld) _installcheck_world: __installcheck_DESTDIR _installcheck_kernel: __installcheck_DESTDIR __installcheck_DESTDIR: .if !defined(DESTDIR) || empty(DESTDIR) @echo "ERROR: Please set DESTDIR!"; \ false .endif .endif .endif .if !defined(DB_FROM_SRC) # # Check for missing UIDs/GIDs. # CHECK_UIDS= auditdistd CHECK_GIDS= audit .if ${MK_SENDMAIL} != "no" CHECK_UIDS+= smmsp CHECK_GIDS+= smmsp .endif .if ${MK_PF} != "no" CHECK_UIDS+= proxy CHECK_GIDS+= proxy authpf .endif .if ${MK_UNBOUND} != "no" CHECK_UIDS+= unbound CHECK_GIDS+= unbound .endif _installcheck_world: __installcheck_UGID __installcheck_UGID: .for uid in ${CHECK_UIDS} @if ! `id -u ${uid} >/dev/null 2>&1`; then \ echo "ERROR: Required ${uid} user is missing, see /usr/src/UPDATING."; \ false; \ fi .endfor .for gid in ${CHECK_GIDS} @if ! `find / -prune -group ${gid} >/dev/null 2>&1`; then \ echo "ERROR: Required ${gid} group is missing, see /usr/src/UPDATING."; \ false; \ fi .endfor .endif # # Required install tools to be saved in a scratch dir for safety. # .if ${MK_INFO} != "no" _install-info= install-info .endif .if ${MK_ZONEINFO} != "no" _zoneinfo= zic tzsetup .endif ITOOLS= [ awk cap_mkdb cat chflags chmod chown \ date echo egrep find grep id install ${_install-info} \ ln lockf make mkdir mtree mv pwd_mkdb \ rm sed services_mkdb sh sysctl test true uname wc ${_zoneinfo} \ ${LOCAL_ITOOLS} +# Needed for share/man +.if ${MK_MAN} != "no" +ITOOLS+=makewhatis +.endif + # # distributeworld # # Distributes everything compiled by a `buildworld'. # # installworld # # Installs everything compiled by a 'buildworld'. # # Non-base distributions produced by the base system EXTRA_DISTRIBUTIONS= doc .if ${MK_GAMES} != "no" EXTRA_DISTRIBUTIONS+= games .endif .if defined(LIB32TMP) && ${MK_LIB32} != "no" EXTRA_DISTRIBUTIONS+= lib32 .endif .if ${MK_TESTS} != "no" EXTRA_DISTRIBUTIONS+= tests .endif DEBUG_DISTRIBUTIONS= .if ${MK_DEBUG_FILES} != "no" DEBUG_DISTRIBUTIONS+= base ${EXTRA_DISTRIBUTIONS:S,doc,,} .endif MTREE_MAGIC?= mtree 2.0 distributeworld installworld: _installcheck_world mkdir -p ${INSTALLTMP} progs=$$(for prog in ${ITOOLS}; do \ if progpath=`which $$prog`; then \ echo $$progpath; \ else \ echo "Required tool $$prog not found in PATH." >&2; \ exit 1; \ fi; \ done); \ libs=$$(ldd -f "%o %p\n" -f "%o %p\n" $$progs 2>/dev/null | sort -u | \ while read line; do \ set -- $$line; \ if [ "$$2 $$3" != "not found" ]; then \ echo $$2; \ else \ echo "Required library $$1 not found." >&2; \ exit 1; \ fi; \ done); \ cp $$libs $$progs ${INSTALLTMP} cp -R $${PATH_LOCALE:-"/usr/share/locale"} ${INSTALLTMP}/locale .if defined(NO_ROOT) echo "#${MTREE_MAGIC}" > ${METALOG} .endif .if make(distributeworld) .for dist in ${EXTRA_DISTRIBUTIONS} -mkdir ${DESTDIR}/${DISTDIR}/${dist} mtree -deU -f ${.CURDIR}/etc/mtree/BSD.root.dist \ -p ${DESTDIR}/${DISTDIR}/${dist} >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${DESTDIR}/${DISTDIR}/${dist}/usr >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.include.dist \ -p ${DESTDIR}/${DISTDIR}/${dist}/usr/include >/dev/null .if ${MK_DEBUG_FILES} != "no" mtree -deU -f ${.CURDIR}/etc/mtree/BSD.debug.dist \ -p ${DESTDIR}/${DISTDIR}/${dist}/usr/lib >/dev/null .endif .if ${MK_TESTS} != "no" && ${dist} == "tests" mtree -deU -f ${.CURDIR}/etc/mtree/BSD.tests.dist \ -p ${DESTDIR}/${DISTDIR}/${dist}/usr >/dev/null .endif .if defined(NO_ROOT) ${IMAKEENV} mtree -C -f ${.CURDIR}/etc/mtree/BSD.root.dist | \ sed -e 's#^\./#./${dist}/#' >> ${METALOG} ${IMAKEENV} mtree -C -f ${.CURDIR}/etc/mtree/BSD.usr.dist | \ sed -e 's#^\./#./${dist}/usr/#' >> ${METALOG} ${IMAKEENV} mtree -C -f ${.CURDIR}/etc/mtree/BSD.include.dist | \ sed -e 's#^\./#./${dist}/usr/include/#' >> ${METALOG} .endif .endfor -mkdir ${DESTDIR}/${DISTDIR}/base cd ${.CURDIR}/etc; ${CROSSENV} PATH=${TMPPATH} ${MAKE} \ METALOG=${METALOG} ${IMAKE_INSTALL} ${IMAKE_MTREE} \ DISTBASE=/base DESTDIR=${DESTDIR}/${DISTDIR}/base \ LOCAL_MTREE=${LOCAL_MTREE:Q} distrib-dirs .endif ${_+_}cd ${.CURDIR}; ${IMAKE} re${.TARGET:S/world$//}; \ ${IMAKEENV} rm -rf ${INSTALLTMP} .if make(distributeworld) .for dist in ${EXTRA_DISTRIBUTIONS} find ${DESTDIR}/${DISTDIR}/${dist} -mindepth 1 -empty -delete .endfor .if defined(NO_ROOT) .for dist in base ${EXTRA_DISTRIBUTIONS} @# For each file that exists in this dist, print the corresponding @# line from the METALOG. This relies on the fact that @# a line containing only the filename will sort immediatly before @# the relevant mtree line. cd ${DESTDIR}/${DISTDIR}; \ find ./${dist} | sort -u ${METALOG} - | \ awk 'BEGIN { print "#${MTREE_MAGIC}" } !/ type=/ { file = $$1 } / type=/ { if ($$1 == file) { sub(/^\.\/${dist}\//, "./"); print } }' > \ ${DESTDIR}/${DISTDIR}/${dist}.meta .endfor .for dist in ${DEBUG_DISTRIBUTIONS} @# For each file that exists in this dist, print the corresponding @# line from the METALOG. This relies on the fact that @# a line containing only the filename will sort immediatly before @# the relevant mtree line. cd ${DESTDIR}/${DISTDIR}; \ find ./${dist}/usr/lib/debug | sort -u ${METALOG} - | \ awk 'BEGIN { print "#${MTREE_MAGIC}" } !/ type=/ { file = $$1 } / type=/ { if ($$1 == file) { sub(/^\.\/${dist}\//, "./"); print } }' > \ ${DESTDIR}/${DISTDIR}/${dist}.debug.meta .endfor .endif .endif packageworld: .for dist in base ${EXTRA_DISTRIBUTIONS} .if defined(NO_ROOT) ${_+_}cd ${DESTDIR}/${DISTDIR}/${dist}; \ tar cvJf ${DESTDIR}/${DISTDIR}/${dist}.txz \ --exclude usr/lib/debug \ @${DESTDIR}/${DISTDIR}/${dist}.meta .else ${_+_}cd ${DESTDIR}/${DISTDIR}/${dist}; \ tar cvJf ${DESTDIR}/${DISTDIR}/${dist}.txz \ --exclude usr/lib/debug . .endif .endfor .for dist in ${DEBUG_DISTRIBUTIONS} . if defined(NO_ROOT) ${_+_}cd ${DESTDIR}/${DISTDIR}/${dist}; \ tar cvJf ${DESTDIR}/${DISTDIR}/${dist}-dbg.txz \ @${DESTDIR}/${DISTDIR}/${dist}.debug.meta . else ${_+_}cd ${DESTDIR}/${DISTDIR}/${dist}; \ tar cvJfL ${DESTDIR}/${DISTDIR}/${dist}-dbg.txz \ usr/lib/debug . endif .endfor # # reinstall # # If you have a build server, you can NFS mount the source and obj directories # and do a 'make reinstall' on the *client* to install new binaries from the # most recent server build. # reinstall: .MAKE @echo "--------------------------------------------------------------" @echo ">>> Making hierarchy" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${MAKE} -f Makefile.inc1 \ LOCAL_MTREE=${LOCAL_MTREE:Q} hierarchy @echo @echo "--------------------------------------------------------------" @echo ">>> Installing everything" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${MAKE} -f Makefile.inc1 install .if defined(LIB32TMP) && ${MK_LIB32} != "no" ${_+_}cd ${.CURDIR}; ${MAKE} -f Makefile.inc1 install32 .endif redistribute: .MAKE @echo "--------------------------------------------------------------" @echo ">>> Distributing everything" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${MAKE} -f Makefile.inc1 distribute .if defined(LIB32TMP) && ${MK_LIB32} != "no" ${_+_}cd ${.CURDIR}; ${MAKE} -f Makefile.inc1 distribute32 \ DISTRIBUTION=lib32 .endif distrib-dirs distribution: .MAKE cd ${.CURDIR}/etc; ${CROSSENV} PATH=${TMPPATH} ${MAKE} \ ${IMAKE_INSTALL} ${IMAKE_MTREE} METALOG=${METALOG} ${.TARGET} # # buildkernel and installkernel # # Which kernels to build and/or install is specified by setting # KERNCONF. If not defined a GENERIC kernel is built/installed. # Only the existing (depending TARGET) config files are used # for building kernels and only the first of these is designated # as the one being installed. # # Note that we have to use TARGET instead of TARGET_ARCH when # we're in kernel-land. Since only TARGET_ARCH is (expected) to # be set to cross-build, we have to make sure TARGET is set # properly. .if defined(KERNFAST) NO_KERNELCLEAN= t NO_KERNELCONFIG= t NO_KERNELDEPEND= t NO_KERNELOBJ= t # Shortcut for KERNCONF=Blah -DKERNFAST is now KERNFAST=Blah .if !defined(KERNCONF) && ${KERNFAST} != "1" KERNCONF=${KERNFAST} .endif .endif .if ${TARGET_ARCH} == "powerpc64" KERNCONF?= GENERIC64 .else KERNCONF?= GENERIC .endif INSTKERNNAME?= kernel KERNSRCDIR?= ${.CURDIR}/sys KRNLCONFDIR= ${KERNSRCDIR}/${TARGET}/conf KRNLOBJDIR= ${OBJTREE}${KERNSRCDIR} KERNCONFDIR?= ${KRNLCONFDIR} BUILDKERNELS= INSTALLKERNEL= .for _kernel in ${KERNCONF} .if exists(${KERNCONFDIR}/${_kernel}) BUILDKERNELS+= ${_kernel} .if empty(INSTALLKERNEL) INSTALLKERNEL= ${_kernel} .endif .endif .endfor buildkernel ${WMAKE_TGTS} ${.ALLTARGETS:M_*}: .MAKE # # buildkernel # # Builds all kernels defined by BUILDKERNELS. # buildkernel: .if empty(BUILDKERNELS) @echo "ERROR: Missing kernel configuration file(s) (${KERNCONF})."; \ false .endif @echo .for _kernel in ${BUILDKERNELS} @echo "--------------------------------------------------------------" @echo ">>> Kernel build for ${_kernel} started on `LC_ALL=C date`" @echo "--------------------------------------------------------------" @echo "===> ${_kernel}" mkdir -p ${KRNLOBJDIR} .if !defined(NO_KERNELCONFIG) @echo @echo "--------------------------------------------------------------" @echo ">>> stage 1: configuring the kernel" @echo "--------------------------------------------------------------" cd ${KRNLCONFDIR}; \ PATH=${TMPPATH} \ config ${CONFIGARGS} -d ${KRNLOBJDIR}/${_kernel} \ -I '${KERNCONFDIR}' '${KERNCONFDIR}/${_kernel}' .endif .if !defined(NO_CLEAN) && !defined(NO_KERNELCLEAN) @echo @echo "--------------------------------------------------------------" @echo ">>> stage 2.1: cleaning up the object tree" @echo "--------------------------------------------------------------" cd ${KRNLOBJDIR}/${_kernel}; ${KMAKE} ${CLEANDIR} .endif .if !defined(NO_KERNELOBJ) @echo @echo "--------------------------------------------------------------" @echo ">>> stage 2.2: rebuilding the object tree" @echo "--------------------------------------------------------------" cd ${KRNLOBJDIR}/${_kernel}; ${KMAKE} obj .endif @echo @echo "--------------------------------------------------------------" @echo ">>> stage 2.3: build tools" @echo "--------------------------------------------------------------" ${_+_}cd ${.CURDIR}; ${KTMAKE} kernel-tools .if !defined(NO_KERNELDEPEND) @echo @echo "--------------------------------------------------------------" @echo ">>> stage 3.1: making dependencies" @echo "--------------------------------------------------------------" cd ${KRNLOBJDIR}/${_kernel}; ${KMAKE} depend -DNO_MODULES_OBJ .endif @echo @echo "--------------------------------------------------------------" @echo ">>> stage 3.2: building everything" @echo "--------------------------------------------------------------" cd ${KRNLOBJDIR}/${_kernel}; ${KMAKE} all -DNO_MODULES_OBJ @echo "--------------------------------------------------------------" @echo ">>> Kernel build for ${_kernel} completed on `LC_ALL=C date`" @echo "--------------------------------------------------------------" .endfor # # installkernel, etc. # # Install the kernel defined by INSTALLKERNEL # installkernel installkernel.debug \ reinstallkernel reinstallkernel.debug: _installcheck_kernel .if empty(INSTALLKERNEL) @echo "ERROR: No kernel \"${KERNCONF}\" to install."; \ false .endif @echo "--------------------------------------------------------------" @echo ">>> Installing kernel ${INSTALLKERNEL}" @echo "--------------------------------------------------------------" cd ${KRNLOBJDIR}/${INSTALLKERNEL}; \ ${CROSSENV} PATH=${TMPPATH} \ ${MAKE} ${IMAKE_INSTALL} KERNEL=${INSTKERNNAME} ${.TARGET:S/kernel//} distributekernel distributekernel.debug: .if empty(INSTALLKERNEL) @echo "ERROR: No kernel \"${KERNCONF}\" to install."; \ false .endif mkdir -p ${DESTDIR}/${DISTDIR} .if defined(NO_ROOT) echo "#${MTREE_MAGIC}" > ${DESTDIR}/${DISTDIR}/kernel.premeta .endif cd ${KRNLOBJDIR}/${INSTALLKERNEL}; \ ${IMAKEENV} ${IMAKE_INSTALL:S/METALOG/kernel.premeta/} \ ${IMAKE_MTREE} PATH=${TMPPATH} ${MAKE} KERNEL=${INSTKERNNAME} \ DESTDIR=${INSTALL_DDIR}/kernel \ ${.TARGET:S/distributekernel/install/} .if defined(NO_ROOT) sed -e 's|^./kernel|.|' ${DESTDIR}/${DISTDIR}/kernel.premeta > \ ${DESTDIR}/${DISTDIR}/kernel.meta .endif .for _kernel in ${BUILDKERNELS:S/${INSTALLKERNEL}//} .if defined(NO_ROOT) echo "#${MTREE_MAGIC}" > ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.premeta .endif cd ${KRNLOBJDIR}/${_kernel}; \ ${IMAKEENV} ${IMAKE_INSTALL:S/METALOG/kernel.${_kernel}.premeta/} \ ${IMAKE_MTREE} PATH=${TMPPATH} ${MAKE} \ KERNEL=${INSTKERNNAME}.${_kernel} \ DESTDIR=${INSTALL_DDIR}/kernel.${_kernel} \ ${.TARGET:S/distributekernel/install/} .if defined(NO_ROOT) sed -e 's|^./kernel|.|' \ ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.premeta > \ ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.meta .endif .endfor packagekernel: .if defined(NO_ROOT) cd ${DESTDIR}/${DISTDIR}/kernel; \ tar cvJf ${DESTDIR}/${DISTDIR}/kernel.txz \ @${DESTDIR}/${DISTDIR}/kernel.meta .for _kernel in ${BUILDKERNELS:S/${INSTALLKERNEL}//} cd ${DESTDIR}/${DISTDIR}/kernel.${_kernel}; \ tar cvJf ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.txz \ @${DESTDIR}/${DISTDIR}/kernel.${_kernel}.meta .endfor .else cd ${DESTDIR}/${DISTDIR}/kernel; \ tar cvJf ${DESTDIR}/${DISTDIR}/kernel.txz . .for _kernel in ${BUILDKERNELS:S/${INSTALLKERNEL}//} cd ${DESTDIR}/${DISTDIR}/kernel.${_kernel}; \ tar cvJf ${DESTDIR}/${DISTDIR}/kernel.${_kernel}.txz . .endfor .endif # # doxygen # # Build the API documentation with doxygen # doxygen: @if [ ! -x `/usr/bin/which doxygen` ]; then \ echo "You need doxygen (devel/doxygen) to generate the API documentation of the kernel." | /usr/bin/fmt; \ exit 1; \ fi cd ${.CURDIR}/tools/kerneldoc/subsys && ${MAKE} obj all # # update # # Update the source tree(s), by running svn/svnup to update to the # latest copy. # update: .if (defined(CVS_UPDATE) || defined(SUP_UPDATE)) && !defined(SVN_UPDATE) @echo "--------------------------------------------------------------" @echo "CVS_UPDATE and SUP_UPDATE are no longer supported." @echo "Please see: https://wiki.freebsd.org/CvsIsDeprecated" @echo "--------------------------------------------------------------" @exit 1 .endif .if defined(SVN_UPDATE) @echo "--------------------------------------------------------------" @echo ">>> Updating ${.CURDIR} using Subversion" @echo "--------------------------------------------------------------" @(cd ${.CURDIR} && ${SVN} update ${SVNFLAGS}) .endif # # ------------------------------------------------------------------------ # # From here onwards are utility targets used by the 'make world' and # related targets. If your 'world' breaks, you may like to try to fix # the problem and manually run the following targets to attempt to # complete the build. Beware, this is *not* guaranteed to work, you # need to have a pretty good grip on the current state of the system # to attempt to manually finish it. If in doubt, 'make world' again. # # # legacy: Build compatibility shims for the next three targets # legacy: .if ${BOOTSTRAPPING} < 800107 && ${BOOTSTRAPPING} != 0 @echo "ERROR: Source upgrades from versions prior to 8.0 not supported."; \ false .endif .for _tool in tools/build ${_+_}@${ECHODIR} "===> ${_tool} (obj,includes,depend,all,install)"; \ cd ${.CURDIR}/${_tool} && \ ${MAKE} DIRPRFX=${_tool}/ obj && \ ${MAKE} DIRPRFX=${_tool}/ DESTDIR=${MAKEOBJDIRPREFIX}/legacy includes && \ ${MAKE} DIRPRFX=${_tool}/ depend && \ ${MAKE} DIRPRFX=${_tool}/ all && \ ${MAKE} DIRPRFX=${_tool}/ DESTDIR=${MAKEOBJDIRPREFIX}/legacy install .endfor # # bootstrap-tools: Build tools needed for compatibility # .if ${MK_GAMES} != "no" _strfile= games/fortune/strfile .endif .if ${MK_CXX} != "no" _gperf= gnu/usr.bin/gperf .endif .if ${MK_GROFF} != "no" _groff= gnu/usr.bin/groff .endif .if ${MK_VT} != "no" _vtfontcvt= usr.bin/vtfontcvt .endif .if ${BOOTSTRAPPING} < 900002 _sed= usr.bin/sed .endif .if ${BOOTSTRAPPING} < 1000002 _m4= lib/libohash \ usr.bin/m4 .endif .if ${BOOTSTRAPPING} < 1000013 _yacc= lib/liby \ usr.bin/yacc .endif .if ${BOOTSTRAPPING} < 1000014 _crunch= usr.sbin/crunch .endif .if ${BOOTSTRAPPING} < 1000026 _nmtree= lib/libnetbsd \ usr.sbin/nmtree .endif .if ${BOOTSTRAPPING} < 1000027 _cat= bin/cat .endif .if ${BOOTSTRAPPING} < 1000033 _lex= usr.bin/lex .endif .if ${BOOTSTRAPPING} >= 900040 && ${BOOTSTRAPPING} < 900041 _awk= usr.bin/awk .endif .if ${MK_BSNMP} != "no" _gensnmptree= usr.sbin/bsnmpd/gensnmptree .endif # We need to build tblgen when we're building clang either as # the bootstrap compiler, or as the part of the normal build. .if ${MK_CLANG_BOOTSTRAP} != "no" || ${MK_CLANG} != "no" _clang_tblgen= \ lib/clang/libllvmsupport \ lib/clang/libllvmtablegen \ usr.bin/clang/tblgen \ usr.bin/clang/clang-tblgen .endif # dtrace tools are required for older bootstrap env and cross-build # pre libdwarf .if ${MK_CDDL} != "no" && (${BOOTSTRAPPING} < 1100006 \ || (${MACHINE} != ${TARGET} || ${MACHINE_ARCH} != ${TARGET_ARCH})) _dtrace_tools= cddl/usr.bin/sgsmsg cddl/lib/libctf lib/libelf \ lib/libdwarf cddl/usr.bin/ctfconvert cddl/usr.bin/ctfmerge .endif # Default to building the GPL DTC, but build the BSDL one if users explicitly # request it. _dtc= usr.bin/dtc .if ${MK_GPL_DTC} != "no" _dtc= gnu/usr.bin/dtc .endif .if ${MK_KERBEROS} != "no" _kerberos5_bootstrap_tools= \ kerberos5/tools/make-roken \ kerberos5/lib/libroken \ kerberos5/lib/libvers \ kerberos5/tools/asn1_compile \ kerberos5/tools/slc \ usr.bin/compile_et .endif # Please document (add comment) why something is in 'bootstrap-tools'. # Try to bound the building of the bootstrap-tool to just the # FreeBSD versions that need the tool built at this stage of the build. bootstrap-tools: .MAKE .for _tool in \ ${_clang_tblgen} \ ${_kerberos5_bootstrap_tools} \ ${_dtrace_tools} \ ${_strfile} \ ${_gperf} \ ${_groff} \ ${_dtc} \ ${_awk} \ ${_cat} \ usr.bin/lorder \ usr.bin/makewhatis \ usr.bin/rpcgen \ ${_sed} \ ${_yacc} \ ${_m4} \ ${_lex} \ lib/libmd \ usr.bin/xinstall \ ${_gensnmptree} \ usr.sbin/config \ ${_crunch} \ ${_nmtree} \ ${_vtfontcvt} ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_tool} && \ ${MAKE} DIRPRFX=${_tool}/ obj && \ ${MAKE} DIRPRFX=${_tool}/ depend && \ ${MAKE} DIRPRFX=${_tool}/ all && \ ${MAKE} DIRPRFX=${_tool}/ DESTDIR=${MAKEOBJDIRPREFIX}/legacy install .endfor # # build-tools: Build special purpose build tools # .if !defined(NO_SHARE) _share= share/syscons/scrnmaps .endif .if ${MK_GCC} != "no" _gcc_tools= gnu/usr.bin/cc/cc_tools .endif .if ${MK_RESCUE} != "no" _rescue= rescue/rescue .endif build-tools: .MAKE .for _tool in \ bin/csh \ bin/sh \ ${_rescue} \ ${LOCAL_TOOL_DIRS} \ lib/ncurses/ncurses \ lib/ncurses/ncursesw \ ${_share} \ usr.bin/awk \ lib/libmagic \ usr.bin/mkesdb_static \ usr.bin/mkcsmapper_static \ usr.bin/vi/catalog ${_+_}@${ECHODIR} "===> ${_tool} (obj,build-tools)"; \ cd ${.CURDIR}/${_tool} && \ ${MAKE} DIRPRFX=${_tool}/ obj && \ ${MAKE} DIRPRFX=${_tool}/ build-tools .endfor .for _tool in \ ${_gcc_tools} ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all)"; \ cd ${.CURDIR}/${_tool} && \ ${MAKE} DIRPRFX=${_tool}/ obj && \ ${MAKE} DIRPRFX=${_tool}/ depend && \ ${MAKE} DIRPRFX=${_tool}/ all .endfor # usr.bin/vi is required to process files in share/termcap .if !defined(NO_SHARE) .for _tool in \ usr.bin/vi ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_tool} && \ ${MAKE} DIRPRFX=${_tool}/ obj && \ ${MAKE} DIRPRFX=${_tool}/ depend MK_MAN=no SUBDIR= && \ ${MAKE} DIRPRFX=${_tool}/ all MK_MAN=no SUBDIR= && \ ${MAKE} DIRPRFX=${_tool}/ install MK_MAN=no SUBDIR= DESTDIR=${WORLDTMP}/ .endfor .endif # # kernel-tools: Build kernel-building tools # kernel-tools: .MAKE mkdir -p ${MAKEOBJDIRPREFIX}/usr mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${MAKEOBJDIRPREFIX}/usr >/dev/null # # cross-tools: Build cross-building tools # .if ${TARGET_ARCH} != ${MACHINE_ARCH} .if ${TARGET_ARCH} == "amd64" || ${TARGET_ARCH} == "i386" _btxld= usr.sbin/btxld .endif .endif .if ${TARGET_ARCH} != ${MACHINE_ARCH} .if ${MK_RESCUE} != "no" || defined(RELEASEDIR) _crunchide= usr.sbin/crunch/crunchide .endif .if ${TARGET_ARCH} == "i386" && defined(RELEASEDIR) _kgzip= usr.sbin/kgzip .endif .endif # If we're given an XAS, don't build binutils. .if ${XAS:M/*} == "" && ${MK_BINUTILS_BOOTSTRAP} != "no" _binutils= gnu/usr.bin/binutils .if ${MK_ELFTOOLCHAIN_TOOLS} != "no" _elftctools= lib/libelftc \ usr.bin/addr2line \ usr.bin/elfcopy \ usr.bin/nm \ usr.bin/size \ usr.bin/strings .endif .endif # If an full path to an external cross compiler is given, don't build # a cross compiler. .if ${XCC:M/*} == "" && ${MK_CROSS_COMPILER} != "no" .if ${MK_CLANG_BOOTSTRAP} != "no" _clang= usr.bin/clang _clang_libs= lib/clang .endif .if ${MK_GCC_BOOTSTRAP} != "no" _cc= gnu/usr.bin/cc .endif .endif cross-tools: .MAKE .for _tool in \ ${_clang_libs} \ ${_clang} \ ${_binutils} \ ${_elftctools} \ ${_cc} \ usr.bin/xlint/lint1 usr.bin/xlint/lint2 usr.bin/xlint/xlint \ ${_btxld} \ ${_crunchide} \ ${_kgzip} \ sys/boot/usb/tools ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_tool} && \ ${MAKE} DIRPRFX=${_tool}/ obj && \ ${MAKE} DIRPRFX=${_tool}/ depend && \ ${MAKE} DIRPRFX=${_tool}/ all && \ ${MAKE} DIRPRFX=${_tool}/ DESTDIR=${MAKEOBJDIRPREFIX} install .endfor NXBENV= MAKEOBJDIRPREFIX=${OBJTREE}/nxb \ INSTALL="sh ${.CURDIR}/tools/install.sh" \ VERSION="${VERSION}" NXBMAKE= ${NXBENV} ${MAKE} \ TBLGEN=${OBJTREE}/nxb-bin/usr/bin/tblgen \ CLANG_TBLGEN=${OBJTREE}/nxb-bin/usr/bin/clang-tblgen \ MACHINE=${TARGET} MACHINE_ARCH=${TARGET_ARCH} \ MK_GDB=no MK_TESTS=no \ SSP_CFLAGS= \ MK_HTML=no MK_INFO=no NO_LINT=yes MK_MAN=no \ -DNO_PIC MK_PROFILE=no -DNO_SHARED \ -DNO_CPU_CFLAGS MK_WARNS=no MK_CTF=no \ MK_CLANG_FULL=no MK_LLDB=no native-xtools: .MAKE mkdir -p ${OBJTREE}/nxb-bin/bin mkdir -p ${OBJTREE}/nxb-bin/sbin mkdir -p ${OBJTREE}/nxb-bin/usr mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${OBJTREE}/nxb-bin/usr >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.include.dist \ -p ${OBJTREE}/nxb-bin/usr/include >/dev/null .for _tool in \ bin/cat \ bin/chmod \ bin/cp \ bin/csh \ bin/echo \ bin/expr \ bin/hostname \ bin/ln \ bin/ls \ bin/mkdir \ bin/mv \ bin/ps \ bin/realpath \ bin/rm \ bin/rmdir \ bin/sh \ bin/sleep \ ${_clang_tblgen} \ usr.bin/ar \ ${_binutils} \ ${_elftctools} \ ${_cc} \ ${_gcc_tools} \ ${_clang_libs} \ ${_clang} \ sbin/md5 \ sbin/sysctl \ gnu/usr.bin/diff \ usr.bin/awk \ usr.bin/basename \ usr.bin/bmake \ usr.bin/bzip2 \ usr.bin/cmp \ usr.bin/dirname \ usr.bin/env \ usr.bin/fetch \ usr.bin/find \ usr.bin/grep \ usr.bin/gzip \ usr.bin/id \ usr.bin/lex \ usr.bin/lorder \ usr.bin/mktemp \ usr.bin/mt \ usr.bin/patch \ usr.bin/sed \ usr.bin/sort \ usr.bin/tar \ usr.bin/touch \ usr.bin/tr \ usr.bin/true \ usr.bin/uniq \ usr.bin/unzip \ usr.bin/xargs \ usr.bin/xinstall \ usr.bin/xz \ usr.bin/yacc \ usr.sbin/chown ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_tool} && \ ${NXBMAKE} DIRPRFX=${_tool}/ obj && \ ${NXBMAKE} DIRPRFX=${_tool}/ depend && \ ${NXBMAKE} DIRPRFX=${_tool}/ all && \ ${NXBMAKE} DIRPRFX=${_tool}/ DESTDIR=${OBJTREE}/nxb-bin install .endfor # # hierarchy - ensure that all the needed directories are present # hierarchy hier: .MAKE cd ${.CURDIR}/etc && ${HMAKE} distrib-dirs # # libraries - build all libraries, and install them under ${DESTDIR}. # # The list of libraries with dependents (${_prebuild_libs}) and their # interdependencies (__L) are built automatically by the # ${.CURDIR}/tools/make_libdeps.sh script. # libraries: .MAKE cd ${.CURDIR} && \ ${MAKE} -f Makefile.inc1 _prereq_libs && \ ${MAKE} -f Makefile.inc1 _startup_libs && \ ${MAKE} -f Makefile.inc1 _prebuild_libs && \ ${MAKE} -f Makefile.inc1 _generic_libs # # static libgcc.a prerequisite for shared libc # _prereq_libs= gnu/lib/libssp/libssp_nonshared gnu/lib/libgcc lib/libcompiler_rt # These dependencies are not automatically generated: # # gnu/lib/csu, gnu/lib/libgcc, lib/csu and lib/libc must be built before # all shared libraries for ELF. # _startup_libs= gnu/lib/csu .if exists(${.CURDIR}/lib/csu/${MACHINE_ARCH}-elf) _startup_libs+= lib/csu/${MACHINE_ARCH}-elf .elif exists(${.CURDIR}/lib/csu/${MACHINE_ARCH}) _startup_libs+= lib/csu/${MACHINE_ARCH} .else _startup_libs+= lib/csu/${MACHINE_CPUARCH} .endif _startup_libs+= gnu/lib/libgcc _startup_libs+= lib/libcompiler_rt _startup_libs+= lib/libc _startup_libs+= lib/libc_nonshared .if ${MK_LIBCPLUSPLUS} != "no" _startup_libs+= lib/libcxxrt .endif gnu/lib/libgcc__L: lib/libc__L gnu/lib/libgcc__L: lib/libc_nonshared__L .if ${MK_LIBCPLUSPLUS} != "no" lib/libcxxrt__L: gnu/lib/libgcc__L .endif _prebuild_libs= ${_kerberos5_lib_libasn1} \ ${_kerberos5_lib_libhdb} \ ${_kerberos5_lib_libheimbase} \ ${_kerberos5_lib_libheimntlm} \ ${_kerberos5_lib_libheimsqlite} \ ${_kerberos5_lib_libheimipcc} \ ${_kerberos5_lib_libhx509} ${_kerberos5_lib_libkrb5} \ ${_kerberos5_lib_libroken} \ ${_kerberos5_lib_libwind} \ lib/libbz2 ${_libcom_err} lib/libcrypt \ lib/libelf lib/libexpat \ lib/libfigpar \ ${_lib_libgssapi} \ lib/libkiconv lib/libkvm lib/liblzma lib/libmd lib/libnv \ ${_lib_libcapsicum} \ lib/ncurses/ncurses lib/ncurses/ncursesw \ lib/libopie lib/libpam ${_lib_libthr} \ ${_lib_libradius} lib/libsbuf lib/libtacplus \ lib/libgeom \ ${_cddl_lib_libumem} ${_cddl_lib_libnvpair} \ ${_cddl_lib_libuutil} \ ${_cddl_lib_libavl} \ ${_cddl_lib_libzfs_core} \ ${_cddl_lib_libctf} \ lib/libutil lib/libpjdlog ${_lib_libypclnt} lib/libz lib/msun \ ${_secure_lib_libcrypto} ${_lib_libldns} \ ${_secure_lib_libssh} ${_secure_lib_libssl} \ gnu/lib/libdialog .if ${MK_GNUCXX} != "no" _prebuild_libs+= gnu/lib/libstdc++ gnu/lib/libsupc++ gnu/lib/libstdc++__L: lib/msun__L gnu/lib/libsupc++__L: gnu/lib/libstdc++__L .endif lib/libgeom__L: lib/libexpat__L .if ${MK_LIBTHR} != "no" _lib_libthr= lib/libthr .endif .if ${MK_RADIUS_SUPPORT} != "no" _lib_libradius= lib/libradius .endif .if ${MK_OFED} != "no" _ofed_lib= contrib/ofed/usr.lib/ .endif .if ${MK_CASPER} != "no" _lib_libcapsicum=lib/libcapsicum .endif lib/libcapsicum__L: lib/libnv__L lib/libpjdlog__L: lib/libutil__L _generic_libs= ${_cddl_lib} gnu/lib ${_kerberos5_lib} lib ${_secure_lib} usr.bin/lex/lib ${_ofed_lib} .for _DIR in ${LOCAL_LIB_DIRS} .if exists(${.CURDIR}/${_DIR}/Makefile) _generic_libs+= ${_DIR} .endif .endfor lib/libopie__L lib/libtacplus__L: lib/libmd__L .if ${MK_CDDL} != "no" _cddl_lib_libumem= cddl/lib/libumem _cddl_lib_libnvpair= cddl/lib/libnvpair _cddl_lib_libavl= cddl/lib/libavl _cddl_lib_libuutil= cddl/lib/libuutil _cddl_lib_libzfs_core= cddl/lib/libzfs_core _cddl_lib_libctf= cddl/lib/libctf _cddl_lib= cddl/lib cddl/lib/libzfs_core__L: cddl/lib/libnvpair__L cddl/lib/libzfs__L: lib/libgeom__L cddl/lib/libctf__L: lib/libz__L .endif .if ${MK_CRYPT} != "no" .if ${MK_OPENSSL} != "no" _secure_lib_libcrypto= secure/lib/libcrypto _secure_lib_libssl= secure/lib/libssl lib/libradius__L secure/lib/libssl__L: secure/lib/libcrypto__L .if ${MK_LDNS} != "no" _lib_libldns= lib/libldns lib/libldns__L: secure/lib/libcrypto__L .endif .if ${MK_OPENSSH} != "no" _secure_lib_libssh= secure/lib/libssh secure/lib/libssh__L: lib/libz__L secure/lib/libcrypto__L lib/libcrypt__L .if ${MK_LDNS} != "no" secure/lib/libssh__L: lib/libldns__L .endif .if ${MK_KERBEROS_SUPPORT} != "no" secure/lib/libssh__L: lib/libgssapi__L kerberos5/lib/libkrb5__L \ kerberos5/lib/libhx509__L kerberos5/lib/libasn1__L lib/libcom_err__L \ lib/libmd__L kerberos5/lib/libroken__L .endif .endif .endif _secure_lib= secure/lib .endif .if ${MK_KERBEROS} != "no" kerberos5/lib/libasn1__L: lib/libcom_err__L kerberos5/lib/libroken__L kerberos5/lib/libhdb__L: kerberos5/lib/libasn1__L lib/libcom_err__L \ kerberos5/lib/libkrb5__L kerberos5/lib/libroken__L \ kerberos5/lib/libwind__L kerberos5/lib/libheimsqlite__L kerberos5/lib/libheimntlm__L: secure/lib/libcrypto__L kerberos5/lib/libkrb5__L \ kerberos5/lib/libroken__L lib/libcom_err__L kerberos5/lib/libhx509__L: kerberos5/lib/libasn1__L lib/libcom_err__L \ secure/lib/libcrypto__L kerberos5/lib/libroken__L kerberos5/lib/libwind__L kerberos5/lib/libkrb5__L: kerberos5/lib/libasn1__L lib/libcom_err__L \ lib/libcrypt__L secure/lib/libcrypto__L kerberos5/lib/libhx509__L \ kerberos5/lib/libroken__L kerberos5/lib/libwind__L \ kerberos5/lib/libheimbase__L kerberos5/lib/libheimipcc__L kerberos5/lib/libroken__L: lib/libcrypt__L kerberos5/lib/libwind__L: kerberos5/lib/libroken__L lib/libcom_err__L kerberos5/lib/libheimbase__L: lib/libthr__L kerberos5/lib/libheimipcc__L: kerberos5/lib/libroken__L kerberos5/lib/libheimbase__L lib/libthr__L kerberos5/lib/libheimsqlite__L: lib/libthr__L .endif .if ${MK_GSSAPI} != "no" _lib_libgssapi= lib/libgssapi .endif .if ${MK_KERBEROS} != "no" _kerberos5_lib= kerberos5/lib _kerberos5_lib_libasn1= kerberos5/lib/libasn1 _kerberos5_lib_libhdb= kerberos5/lib/libhdb _kerberos5_lib_libheimbase= kerberos5/lib/libheimbase _kerberos5_lib_libkrb5= kerberos5/lib/libkrb5 _kerberos5_lib_libhx509= kerberos5/lib/libhx509 _kerberos5_lib_libroken= kerberos5/lib/libroken _kerberos5_lib_libheimntlm= kerberos5/lib/libheimntlm _kerberos5_lib_libheimsqlite= kerberos5/lib/libheimsqlite _kerberos5_lib_libheimipcc= kerberos5/lib/libheimipcc _kerberos5_lib_libwind= kerberos5/lib/libwind _libcom_err= lib/libcom_err .endif .if ${MK_NIS} != "no" _lib_libypclnt= lib/libypclnt .endif .if ${MK_OPENSSL} == "no" lib/libradius__L: lib/libmd__L .endif gnu/lib/libdialog__L: lib/msun__L lib/ncurses/ncursesw__L .for _lib in ${_prereq_libs} ${_lib}__PL: .PHONY .MAKE .if exists(${.CURDIR}/${_lib}) ${_+_}@${ECHODIR} "===> ${_lib} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_lib} && \ ${MAKE} MK_TESTS=no DIRPRFX=${_lib}/ obj && \ ${MAKE} MK_TESTS=no DIRPRFX=${_lib}/ depend && \ ${MAKE} MK_TESTS=no MK_PROFILE=no -DNO_PIC \ DIRPRFX=${_lib}/ all && \ ${MAKE} MK_TESTS=no MK_PROFILE=no -DNO_PIC \ DIRPRFX=${_lib}/ install .endif .endfor .for _lib in ${_startup_libs} ${_prebuild_libs:Nlib/libpam} ${_generic_libs} ${_lib}__L: .PHONY .MAKE .if exists(${.CURDIR}/${_lib}) ${_+_}@${ECHODIR} "===> ${_lib} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_lib} && \ ${MAKE} MK_TESTS=no DIRPRFX=${_lib}/ obj && \ ${MAKE} MK_TESTS=no DIRPRFX=${_lib}/ depend && \ ${MAKE} MK_TESTS=no DIRPRFX=${_lib}/ all && \ ${MAKE} MK_TESTS=no DIRPRFX=${_lib}/ install .endif .endfor # libpam is special: we need to build static PAM modules before # static PAM library, and dynamic PAM library before dynamic PAM # modules. lib/libpam__L: .PHONY .MAKE ${_+_}@${ECHODIR} "===> lib/libpam (obj,depend,all,install)"; \ cd ${.CURDIR}/lib/libpam && \ ${MAKE} MK_TESTS=no DIRPRFX=lib/libpam/ obj && \ ${MAKE} MK_TESTS=no DIRPRFX=lib/libpam/ depend && \ ${MAKE} MK_TESTS=no DIRPRFX=lib/libpam/ \ -D_NO_LIBPAM_SO_YET all && \ ${MAKE} MK_TESTS=no DIRPRFX=lib/libpam/ \ -D_NO_LIBPAM_SO_YET install _prereq_libs: ${_prereq_libs:S/$/__PL/} _startup_libs: ${_startup_libs:S/$/__L/} _prebuild_libs: ${_prebuild_libs:S/$/__L/} _generic_libs: ${_generic_libs:S/$/__L/} .for __target in all clean cleandepend cleandir depend includes obj .for entry in ${SUBDIR} ${entry}.${__target}__D: .PHONY .MAKE ${_+_}@set -e; if test -d ${.CURDIR}/${entry}.${MACHINE_ARCH}; then \ ${ECHODIR} "===> ${DIRPRFX}${entry}.${MACHINE_ARCH} (${__target})"; \ edir=${entry}.${MACHINE_ARCH}; \ cd ${.CURDIR}/$${edir}; \ else \ ${ECHODIR} "===> ${DIRPRFX}${entry} (${__target})"; \ edir=${entry}; \ cd ${.CURDIR}/$${edir}; \ fi; \ ${MAKE} ${__target} DIRPRFX=${DIRPRFX}$${edir}/ .endfor par-${__target}: ${SUBDIR:S/$/.${__target}__D/} .endfor .include .if make(check-old) || make(check-old-dirs) || \ make(check-old-files) || make(check-old-libs) || \ make(delete-old) || make(delete-old-dirs) || \ make(delete-old-files) || make(delete-old-libs) # # check for / delete old files section # .include "ObsoleteFiles.inc" OLD_LIBS_MESSAGE="Please be sure no application still uses those libraries, \ else you can not start such an application. Consult UPDATING for more \ information regarding how to cope with the removal/revision bump of a \ specific library." .if !defined(BATCH_DELETE_OLD_FILES) RM_I=-i .else RM_I=-v .endif delete-old-files: @echo ">>> Removing old files (only deletes safe to delete libs)" # Ask for every old file if the user really wants to remove it. # It's annoying, but better safe than sorry. # NB: We cannot pass the list of OLD_FILES as a parameter because the # argument list will get too long. Using .for/.endfor make "loops" will make # the Makefile parser segfault. @exec 3<&0; \ cd ${.CURDIR}; \ ${MAKE} -f ${.CURDIR}/Makefile.inc1 ${.MAKEFLAGS} ${.TARGET} \ -V OLD_FILES -V "OLD_FILES:Musr/share/*.gz:R" | xargs -n1 | \ while read file; do \ if [ -f "${DESTDIR}/$${file}" -o -L "${DESTDIR}/$${file}" ]; then \ chflags noschg "${DESTDIR}/$${file}" 2>/dev/null || true; \ rm ${RM_I} "${DESTDIR}/$${file}" <&3; \ fi; \ done # Remove catpages without corresponding manpages. @exec 3<&0; \ find ${DESTDIR}/usr/share/man/cat* ! -type d | \ sed -ep -e's:${DESTDIR}/usr/share/man/cat:${DESTDIR}/usr/share/man/man:' | \ while read catpage; do \ read manpage; \ if [ ! -e "$${manpage}" ]; then \ rm ${RM_I} $${catpage} <&3; \ fi; \ done @echo ">>> Old files removed" check-old-files: @echo ">>> Checking for old files" @cd ${.CURDIR}; \ ${MAKE} -f ${.CURDIR}/Makefile.inc1 ${.MAKEFLAGS} ${.TARGET} \ -V OLD_FILES -V "OLD_FILES:Musr/share/*.gz:R" | xargs -n1 | \ while read file; do \ if [ -f "${DESTDIR}/$${file}" -o -L "${DESTDIR}/$${file}" ]; then \ echo "${DESTDIR}/$${file}"; \ fi; \ done # Check for catpages without corresponding manpages. @find ${DESTDIR}/usr/share/man/cat* ! -type d | \ sed -ep -e's:${DESTDIR}/usr/share/man/cat:${DESTDIR}/usr/share/man/man:' | \ while read catpage; do \ read manpage; \ if [ ! -e "$${manpage}" ]; then \ echo $${catpage}; \ fi; \ done delete-old-libs: @echo ">>> Removing old libraries" @echo "${OLD_LIBS_MESSAGE}" | fmt @exec 3<&0; \ cd ${.CURDIR}; \ ${MAKE} -f ${.CURDIR}/Makefile.inc1 ${.MAKEFLAGS} ${.TARGET} \ -V OLD_LIBS | xargs -n1 | \ while read file; do \ if [ -f "${DESTDIR}/$${file}" -o -L "${DESTDIR}/$${file}" ]; then \ chflags noschg "${DESTDIR}/$${file}" 2>/dev/null || true; \ rm ${RM_I} "${DESTDIR}/$${file}" <&3; \ fi; \ for ext in debug symbols; do \ if ! [ -e "${DESTDIR}/$${file}" ] && [ -f \ "${DESTDIR}${DEBUGDIR}/$${file}.$${ext}" ]; then \ rm ${RM_I} "${DESTDIR}${DEBUGDIR}/$${file}.$${ext}" \ <&3; \ fi; \ done; \ done @echo ">>> Old libraries removed" check-old-libs: @echo ">>> Checking for old libraries" @cd ${.CURDIR}; \ ${MAKE} -f ${.CURDIR}/Makefile.inc1 ${.MAKEFLAGS} ${.TARGET} \ -V OLD_LIBS | xargs -n1 | \ while read file; do \ if [ -f "${DESTDIR}/$${file}" -o -L "${DESTDIR}/$${file}" ]; then \ echo "${DESTDIR}/$${file}"; \ fi; \ for ext in debug symbols; do \ if [ -f "${DESTDIR}${DEBUGDIR}/$${file}.$${ext}" ]; then \ echo "${DESTDIR}${DEBUGDIR}/$${file}.$${ext}"; \ fi; \ done; \ done delete-old-dirs: @echo ">>> Removing old directories" @cd ${.CURDIR}; \ ${MAKE} -f ${.CURDIR}/Makefile.inc1 ${.MAKEFLAGS} ${.TARGET} \ -V OLD_DIRS | xargs -n1 | sort -r | \ while read dir; do \ if [ -d "${DESTDIR}/$${dir}" ]; then \ rmdir -v "${DESTDIR}/$${dir}" || true; \ elif [ -L "${DESTDIR}/$${dir}" ]; then \ echo "${DESTDIR}/$${dir} is a link, please remove everything manually."; \ fi; \ done @echo ">>> Old directories removed" check-old-dirs: @echo ">>> Checking for old directories" @cd ${.CURDIR}; \ ${MAKE} -f ${.CURDIR}/Makefile.inc1 ${.MAKEFLAGS} ${.TARGET} \ -V OLD_DIRS | xargs -n1 | \ while read dir; do \ if [ -d "${DESTDIR}/$${dir}" ]; then \ echo "${DESTDIR}/$${dir}"; \ elif [ -L "${DESTDIR}/$${dir}" ]; then \ echo "${DESTDIR}/$${dir} is a link, please remove everything manually."; \ fi; \ done delete-old: delete-old-files delete-old-dirs @echo "To remove old libraries run '${MAKE} delete-old-libs'." check-old: check-old-files check-old-libs check-old-dirs @echo "To remove old files and directories run '${MAKE} delete-old'." @echo "To remove old libraries run '${MAKE} delete-old-libs'." .endif # # showconfig - show build configuration. # showconfig: @(${MAKE} -n -f ${.CURDIR}/sys/conf/kern.opts.mk -V dummy -dg1; \ ${MAKE} -n -f ${.CURDIR}/share/mk/src.opts.mk -V dummy -dg1) 2>&1 | grep ^MK_ | sort -u .if !empty(KRNLOBJDIR) && !empty(KERNCONF) DTBOUTPUTPATH= ${KRNLOBJDIR}/${KERNCONF}/ .if !defined(FDT_DTS_FILE) || empty(FDT_DTS_FILE) .if exists(${KERNCONFDIR}/${KERNCONF}) FDT_DTS_FILE!= awk 'BEGIN {FS="="} /^makeoptions[[:space:]]+FDT_DTS_FILE/ {print $$2}' \ '${KERNCONFDIR}/${KERNCONF}' ; echo .endif .endif .endif .if !defined(DTBOUTPUTPATH) || !exists(${DTBOUTPUTPATH}) DTBOUTPUTPATH= ${.CURDIR} .endif # # Build 'standalone' Device Tree Blob # builddtb: @PATH=${TMPPATH} MACHINE=${TARGET} \ ${.CURDIR}/sys/tools/fdt/make_dtb.sh ${.CURDIR}/sys \ "${FDT_DTS_FILE}" ${DTBOUTPUTPATH} ############### .if defined(TARGET) && defined(TARGET_ARCH) .if ${TARGET} == ${MACHINE} && ${TARGET_ARCH} == ${MACHINE_ARCH} XDEV_CPUTYPE?=${CPUTYPE} .else XDEV_CPUTYPE?=${TARGET_CPUTYPE} .endif NOFUN=-DNO_FSCHG MK_HTML=no MK_INFO=no -DNO_LINT \ MK_MAN=no MK_NLS=no MK_PROFILE=no \ MK_KERBEROS=no MK_RESCUE=no MK_TESTS=no MK_WARNS=no \ MK_VI=no \ TARGET=${TARGET} TARGET_ARCH=${TARGET_ARCH} \ CPUTYPE=${XDEV_CPUTYPE} -DNO_SHARE XDDIR=${TARGET_ARCH}-freebsd XDTP?=/usr/${XDDIR} .if ${XDTP:N/*} .error XDTP variable should be an absolute path .endif CDBENV=MAKEOBJDIRPREFIX=${MAKEOBJDIRPREFIX}/${XDDIR} \ INSTALL="sh ${.CURDIR}/tools/install.sh" CDENV= ${CDBENV} \ _SHLIBDIRPREFIX=${XDDESTDIR} \ TOOLS_PREFIX=${XDTP} CD2CFLAGS=-isystem ${XDDESTDIR}/usr/include -L${XDDESTDIR}/usr/lib \ --sysroot=${XDDESTDIR}/ -B${XDDESTDIR}/usr/libexec \ -B${XDDESTDIR}/usr/bin -B${XDDESTDIR}/usr/lib CD2ENV=${CDENV} CC="${CC} ${CD2CFLAGS}" CXX="${CXX} ${CD2CFLAGS}" \ CPP="${CPP} ${CD2CFLAGS}" \ MACHINE=${TARGET} MACHINE_ARCH=${TARGET_ARCH} CDTMP= ${MAKEOBJDIRPREFIX}/${XDDIR}/${.CURDIR}/tmp CDMAKE=${CDENV} PATH=${CDTMP}/usr/bin:${PATH} ${MAKE} ${NOFUN} CD2MAKE=${CD2ENV} PATH=${CDTMP}/usr/bin:${XDDESTDIR}/usr/bin:${PATH} ${MAKE} ${NOFUN} XDDESTDIR=${DESTDIR}/${XDTP} .if !defined(OSREL) OSREL!= uname -r | sed -e 's/[-(].*//' .endif .ORDER: xdev-build xdev-install xdev-links xdev: xdev-build xdev-install .ORDER: _xb-worldtmp _xb-bootstrap-tools _xb-build-tools _xb-cross-tools xdev-build: _xb-worldtmp _xb-bootstrap-tools _xb-build-tools _xb-cross-tools _xb-worldtmp: mkdir -p ${CDTMP}/usr mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${CDTMP}/usr >/dev/null _xb-bootstrap-tools: .for _tool in \ ${_clang_tblgen} \ ${_gperf} ${_+_}@${ECHODIR} "===> ${_tool} (obj,depend,all,install)"; \ cd ${.CURDIR}/${_tool} && \ ${CDMAKE} DIRPRFX=${_tool}/ obj && \ ${CDMAKE} DIRPRFX=${_tool}/ depend && \ ${CDMAKE} DIRPRFX=${_tool}/ all && \ ${CDMAKE} DIRPRFX=${_tool}/ DESTDIR=${CDTMP} install .endfor _xb-build-tools: ${_+_}@cd ${.CURDIR}; \ ${CDBENV} ${MAKE} -f Makefile.inc1 ${NOFUN} build-tools _xb-cross-tools: .for _tool in \ ${_binutils} \ ${_elftctools} \ usr.bin/ar \ ${_clang_libs} \ ${_clang} \ ${_cc} ${_+_}@${ECHODIR} "===> xdev ${_tool} (obj,depend,all)"; \ cd ${.CURDIR}/${_tool} && \ ${CDMAKE} DIRPRFX=${_tool}/ obj && \ ${CDMAKE} DIRPRFX=${_tool}/ depend && \ ${CDMAKE} DIRPRFX=${_tool}/ all .endfor _xi-mtree: ${_+_}@${ECHODIR} "mtree populating ${XDDESTDIR}" mkdir -p ${XDDESTDIR} mtree -deU -f ${.CURDIR}/etc/mtree/BSD.root.dist \ -p ${XDDESTDIR} >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.usr.dist \ -p ${XDDESTDIR}/usr >/dev/null mtree -deU -f ${.CURDIR}/etc/mtree/BSD.include.dist \ -p ${XDDESTDIR}/usr/include >/dev/null .if ${MK_TESTS} != "no" mtree -deU -f ${.CURDIR}/etc/mtree/BSD.tests.dist \ -p ${XDDESTDIR}/usr >/dev/null .endif .ORDER: xdev-build _xi-mtree _xi-cross-tools _xi-includes _xi-libraries xdev-install: xdev-build _xi-mtree _xi-cross-tools _xi-includes _xi-libraries _xi-cross-tools: @echo "_xi-cross-tools" .for _tool in \ ${_binutils} \ ${_elftctools} \ usr.bin/ar \ ${_clang_libs} \ ${_clang} \ ${_cc} ${_+_}@${ECHODIR} "===> xdev ${_tool} (install)"; \ cd ${.CURDIR}/${_tool}; \ ${CDMAKE} DIRPRFX=${_tool}/ install DESTDIR=${XDDESTDIR} .endfor _xi-includes: ${_+_}cd ${.CURDIR}; ${CD2MAKE} -f Makefile.inc1 par-includes \ DESTDIR=${XDDESTDIR} _xi-libraries: ${_+_}cd ${.CURDIR}; ${CD2MAKE} -f Makefile.inc1 libraries \ DESTDIR=${XDDESTDIR} xdev-links: ${_+_}cd ${XDDESTDIR}/usr/bin; \ mkdir -p ../../../../usr/bin; \ for i in *; do \ ln -sf ../../${XDTP}/usr/bin/$$i \ ../../../../usr/bin/${XDDIR}-$$i; \ ln -sf ../../${XDTP}/usr/bin/$$i \ ../../../../usr/bin/${XDDIR}${OSREL}-$$i; \ done .else xdev xdev-build xdev-install xdev-links: @echo "*** Error: Both TARGET and TARGET_ARCH must be defined for \"${.TARGET}\" target" .endif Index: projects/building-blocks/bin/freebsd-version/Makefile =================================================================== --- projects/building-blocks/bin/freebsd-version/Makefile (revision 275663) +++ projects/building-blocks/bin/freebsd-version/Makefile (revision 275664) @@ -1,20 +1,19 @@ # $FreeBSD$ SCRIPTS = freebsd-version MAN = freebsd-version.1 CLEANFILES = freebsd-version.sh NEWVERS = ${.CURDIR}/../../sys/conf/newvers.sh -freebsd-version.sh.in: ${NEWVERS} -freebsd-version.sh: ${.CURDIR}/freebsd-version.sh.in +freebsd-version.sh: ${.CURDIR}/freebsd-version.sh.in ${NEWVERS} eval $$(egrep '^(TYPE|REVISION|BRANCH)=' ${NEWVERS}) ; \ if ! sed -e "\ s/@@TYPE@@/$${TYPE}/g; \ s/@@REVISION@@/$${REVISION}/g; \ s/@@BRANCH@@/$${BRANCH}/g; \ " ${.ALLSRC} >${.TARGET} ; then \ rm -f ${.TARGET} ; \ exit 1 ; \ fi .include Index: projects/building-blocks/cddl/compat/opensolaris/misc/thread_pool.c =================================================================== --- projects/building-blocks/cddl/compat/opensolaris/misc/thread_pool.c (revision 275663) +++ projects/building-blocks/cddl/compat/opensolaris/misc/thread_pool.c (revision 275664) @@ -1,430 +1,428 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #include __FBSDID("$FreeBSD$"); #pragma ident "%Z%%M% %I% %E% SMI" #include #include #include #include "thread_pool_impl.h" typedef void (*_Voidfp)(void*); /* pointer to extern "C" function */ static void delete_pool(tpool_t *tpool) { tpool_job_t *job; /* * There should be no pending jobs, but just in case... */ for (job = tpool->tp_head; job != NULL; job = tpool->tp_head) { tpool->tp_head = job->tpj_next; free(job); } (void) pthread_attr_destroy(&tpool->tp_attr); free(tpool); } /* * Worker thread is terminating. */ static void worker_cleanup(void *arg) { tpool_t *tpool = arg; if (--tpool->tp_current == 0 && (tpool->tp_flags & (TP_DESTROY | TP_ABANDON))) { if (tpool->tp_flags & TP_ABANDON) { pthread_mutex_unlock(&tpool->tp_mutex); delete_pool(tpool); return; } if (tpool->tp_flags & TP_DESTROY) (void) pthread_cond_broadcast(&tpool->tp_busycv); } pthread_mutex_unlock(&tpool->tp_mutex); } static void notify_waiters(tpool_t *tpool) { if (tpool->tp_head == NULL && tpool->tp_active == NULL) { tpool->tp_flags &= ~TP_WAIT; (void) pthread_cond_broadcast(&tpool->tp_waitcv); } } /* * Called by a worker thread on return from a tpool_dispatch()d job. */ static void job_cleanup(void *arg) { tpool_t *tpool = arg; pthread_t my_tid = pthread_self(); tpool_active_t *activep; tpool_active_t **activepp; pthread_mutex_lock(&tpool->tp_mutex); /* CSTYLED */ for (activepp = &tpool->tp_active;; activepp = &activep->tpa_next) { activep = *activepp; if (activep->tpa_tid == my_tid) { *activepp = activep->tpa_next; break; } } if (tpool->tp_flags & TP_WAIT) notify_waiters(tpool); } static void * tpool_worker(void *arg) { tpool_t *tpool = (tpool_t *)arg; int elapsed; tpool_job_t *job; void (*func)(void *); tpool_active_t active; sigset_t maskset; pthread_mutex_lock(&tpool->tp_mutex); pthread_cleanup_push(worker_cleanup, tpool); /* * This is the worker's main loop. * It will only be left if a timeout or an error has occured. */ active.tpa_tid = pthread_self(); for (;;) { elapsed = 0; tpool->tp_idle++; if (tpool->tp_flags & TP_WAIT) notify_waiters(tpool); while ((tpool->tp_head == NULL || (tpool->tp_flags & TP_SUSPEND)) && !(tpool->tp_flags & (TP_DESTROY | TP_ABANDON))) { if (tpool->tp_current <= tpool->tp_minimum || tpool->tp_linger == 0) { (void) pthread_cond_wait(&tpool->tp_workcv, &tpool->tp_mutex); } else { struct timespec timeout; clock_gettime(CLOCK_MONOTONIC, &timeout); timeout.tv_sec += tpool->tp_linger; if (pthread_cond_timedwait(&tpool->tp_workcv, &tpool->tp_mutex, &timeout) != 0) { elapsed = 1; break; } } } tpool->tp_idle--; if (tpool->tp_flags & TP_DESTROY) break; if (tpool->tp_flags & TP_ABANDON) { /* can't abandon a suspended pool */ if (tpool->tp_flags & TP_SUSPEND) { tpool->tp_flags &= ~TP_SUSPEND; (void) pthread_cond_broadcast(&tpool->tp_workcv); } if (tpool->tp_head == NULL) break; } if ((job = tpool->tp_head) != NULL && !(tpool->tp_flags & TP_SUSPEND)) { elapsed = 0; func = job->tpj_func; arg = job->tpj_arg; tpool->tp_head = job->tpj_next; if (job == tpool->tp_tail) tpool->tp_tail = NULL; tpool->tp_njobs--; active.tpa_next = tpool->tp_active; tpool->tp_active = &active; pthread_mutex_unlock(&tpool->tp_mutex); pthread_cleanup_push(job_cleanup, tpool); free(job); /* * Call the specified function. */ func(arg); /* * We don't know what this thread has been doing, * so we reset its signal mask and cancellation * state back to the initial values. */ sigfillset(&maskset); (void) pthread_sigmask(SIG_SETMASK, &maskset, NULL); (void) pthread_setcanceltype(PTHREAD_CANCEL_DEFERRED, NULL); (void) pthread_setcancelstate(PTHREAD_CANCEL_ENABLE, NULL); pthread_cleanup_pop(1); } if (elapsed && tpool->tp_current > tpool->tp_minimum) { /* * We timed out and there is no work to be done * and the number of workers exceeds the minimum. * Exit now to reduce the size of the pool. */ break; } } pthread_cleanup_pop(1); return (arg); } /* * Create a worker thread, with all signals blocked. */ static int create_worker(tpool_t *tpool) { sigset_t maskset, oset; pthread_t thread; int error; sigfillset(&maskset); (void) pthread_sigmask(SIG_SETMASK, &maskset, &oset); error = pthread_create(&thread, &tpool->tp_attr, tpool_worker, tpool); (void) pthread_sigmask(SIG_SETMASK, &oset, NULL); return (error); } tpool_t * tpool_create(uint_t min_threads, uint_t max_threads, uint_t linger, pthread_attr_t *attr) { tpool_t *tpool; int error; if (min_threads > max_threads || max_threads < 1) { errno = EINVAL; return (NULL); } - tpool = malloc(sizeof (*tpool)); + tpool = calloc(1, sizeof (*tpool)); if (tpool == NULL) { errno = ENOMEM; return (NULL); } - bzero(tpool, sizeof(*tpool)); (void) pthread_mutex_init(&tpool->tp_mutex, NULL); (void) pthread_cond_init(&tpool->tp_busycv, NULL); (void) pthread_cond_init(&tpool->tp_workcv, NULL); (void) pthread_cond_init(&tpool->tp_waitcv, NULL); tpool->tp_minimum = min_threads; tpool->tp_maximum = max_threads; tpool->tp_linger = linger; /* make all pool threads be detached daemon threads */ (void) pthread_attr_init(&tpool->tp_attr); (void) pthread_attr_setdetachstate(&tpool->tp_attr, PTHREAD_CREATE_DETACHED); return (tpool); } /* * Dispatch a work request to the thread pool. * If there are idle workers, awaken one. * Else, if the maximum number of workers has * not been reached, spawn a new worker thread. * Else just return with the job added to the queue. */ int tpool_dispatch(tpool_t *tpool, void (*func)(void *), void *arg) { tpool_job_t *job; - if ((job = malloc(sizeof (*job))) == NULL) + if ((job = calloc(1, sizeof (*job))) == NULL) return (-1); - bzero(job, sizeof(*job)); job->tpj_next = NULL; job->tpj_func = func; job->tpj_arg = arg; pthread_mutex_lock(&tpool->tp_mutex); if (tpool->tp_head == NULL) tpool->tp_head = job; else tpool->tp_tail->tpj_next = job; tpool->tp_tail = job; tpool->tp_njobs++; if (!(tpool->tp_flags & TP_SUSPEND)) { if (tpool->tp_idle > 0) (void) pthread_cond_signal(&tpool->tp_workcv); else if (tpool->tp_current < tpool->tp_maximum && create_worker(tpool) == 0) tpool->tp_current++; } pthread_mutex_unlock(&tpool->tp_mutex); return (0); } /* * Assumes: by the time tpool_destroy() is called no one will use this * thread pool in any way and no one will try to dispatch entries to it. * Calling tpool_destroy() from a job in the pool will cause deadlock. */ void tpool_destroy(tpool_t *tpool) { tpool_active_t *activep; pthread_mutex_lock(&tpool->tp_mutex); pthread_cleanup_push((_Voidfp)pthread_mutex_unlock, &tpool->tp_mutex); /* mark the pool as being destroyed; wakeup idle workers */ tpool->tp_flags |= TP_DESTROY; tpool->tp_flags &= ~TP_SUSPEND; (void) pthread_cond_broadcast(&tpool->tp_workcv); /* cancel all active workers */ for (activep = tpool->tp_active; activep; activep = activep->tpa_next) (void) pthread_cancel(activep->tpa_tid); /* wait for all active workers to finish */ while (tpool->tp_active != NULL) { tpool->tp_flags |= TP_WAIT; (void) pthread_cond_wait(&tpool->tp_waitcv, &tpool->tp_mutex); } /* the last worker to terminate will wake us up */ while (tpool->tp_current != 0) (void) pthread_cond_wait(&tpool->tp_busycv, &tpool->tp_mutex); pthread_cleanup_pop(1); /* pthread_mutex_unlock(&tpool->tp_mutex); */ delete_pool(tpool); } /* * Like tpool_destroy(), but don't cancel workers or wait for them to finish. * The last worker to terminate will delete the pool. */ void tpool_abandon(tpool_t *tpool) { pthread_mutex_lock(&tpool->tp_mutex); if (tpool->tp_current == 0) { /* no workers, just delete the pool */ pthread_mutex_unlock(&tpool->tp_mutex); delete_pool(tpool); } else { /* wake up all workers, last one will delete the pool */ tpool->tp_flags |= TP_ABANDON; tpool->tp_flags &= ~TP_SUSPEND; (void) pthread_cond_broadcast(&tpool->tp_workcv); pthread_mutex_unlock(&tpool->tp_mutex); } } /* * Wait for all jobs to complete. * Calling tpool_wait() from a job in the pool will cause deadlock. */ void tpool_wait(tpool_t *tpool) { pthread_mutex_lock(&tpool->tp_mutex); pthread_cleanup_push((_Voidfp)pthread_mutex_unlock, &tpool->tp_mutex); while (tpool->tp_head != NULL || tpool->tp_active != NULL) { tpool->tp_flags |= TP_WAIT; (void) pthread_cond_wait(&tpool->tp_waitcv, &tpool->tp_mutex); } pthread_cleanup_pop(1); /* pthread_mutex_unlock(&tpool->tp_mutex); */ } void tpool_suspend(tpool_t *tpool) { pthread_mutex_lock(&tpool->tp_mutex); tpool->tp_flags |= TP_SUSPEND; pthread_mutex_unlock(&tpool->tp_mutex); } int tpool_suspended(tpool_t *tpool) { int suspended; pthread_mutex_lock(&tpool->tp_mutex); suspended = (tpool->tp_flags & TP_SUSPEND) != 0; pthread_mutex_unlock(&tpool->tp_mutex); return (suspended); } void tpool_resume(tpool_t *tpool) { int excess; pthread_mutex_lock(&tpool->tp_mutex); if (!(tpool->tp_flags & TP_SUSPEND)) { pthread_mutex_unlock(&tpool->tp_mutex); return; } tpool->tp_flags &= ~TP_SUSPEND; (void) pthread_cond_broadcast(&tpool->tp_workcv); excess = tpool->tp_njobs - tpool->tp_idle; while (excess-- > 0 && tpool->tp_current < tpool->tp_maximum) { if (create_worker(tpool) != 0) break; /* pthread_create() failed */ tpool->tp_current++; } pthread_mutex_unlock(&tpool->tp_mutex); } int tpool_member(tpool_t *tpool) { pthread_t my_tid = pthread_self(); tpool_active_t *activep; pthread_mutex_lock(&tpool->tp_mutex); for (activep = tpool->tp_active; activep; activep = activep->tpa_next) { if (activep->tpa_tid == my_tid) { pthread_mutex_unlock(&tpool->tp_mutex); return (1); } } pthread_mutex_unlock(&tpool->tp_mutex); return (0); } Index: projects/building-blocks/cddl/contrib/opensolaris/lib/libctf/common/ctf.5 =================================================================== --- projects/building-blocks/cddl/contrib/opensolaris/lib/libctf/common/ctf.5 (nonexistent) +++ projects/building-blocks/cddl/contrib/opensolaris/lib/libctf/common/ctf.5 (revision 275664) @@ -0,0 +1,1140 @@ +.\" +.\" This file and its contents are supplied under the terms of the +.\" Common Development and Distribution License ("CDDL"), version 1.0. +.\" You may only use this file in accordance with the terms of version +.\" 1.0 of the CDDL. +.\" +.\" A full copy of the text of the CDDL should have accompanied this +.\" source. A copy of the CDDL is also available via the Internet at +.\" http://www.illumos.org/license/CDDL. +.\" +.\" +.\" Copyright (c) 2014 Joyent, Inc. +.\" +.Dd Sep 26, 2014 +.Dt CTF 5 +.Os +.Sh NAME +.Nm ctf +.Nd Compact C Type Format +.Sh SYNOPSIS +.In sys/ctf.h +.Sh DESCRIPTION +.Nm +is designed to be a compact representation of the C programming +language's type information focused on serving the needs of dynamic +tracing, debuggers, and other in-situ and post-mortem introspection +tools. +.Nm +data is generally included in +.Sy ELF +objects and is tagged as +.Sy SHT_PROGBITS +to ensure that the data is accessible in a running process and in subsequent +core dumps, if generated. +.Lp +The +.Nm +data contained in each file has information about the layout and +sizes of C types, including intrinsic types, enumerations, structures, +typedefs, and unions, that are used by the corresponding +.Sy ELF +object. The +.Nm +data may also include information about the types of global objects and +the return type and arguments of functions in the symbol table. +.Lp +Because a +.Nm +file is often embedded inside a file, rather than being a standalone +file itself, it may also be referred to as a +.Nm +.Sy container . +.Lp +On illumos systems, +.Nm +data is consumed by multiple programs. It can be used by the modular +debugger, +.Xr mdb 1 , +as well as by +.Xr dtrace 1M . +Programmatic access to +.Nm +data can be obtained through +.Xr libctf 3LIB . +.Lp +The +.Nm +file format is broken down into seven different sections. The first +section is the +.Sy preamble +and +.Sy header , +which describes the version of the +.Nm +file, links it has to other +.Nm +files, and the sizes of the other sections. The next section is the +.Sy label +section, +which provides a way of identifying similar groups of +.Nm +data across multiple files. This is followed by the +.Sy object +information section, which describes the type of global +symbols. The subsequent section is the +.Sy function +information section, which describes the return +types and arguments of functions. The next section is the +.Sy type +information section, which describes +the format and layout of the C types themselves, and finally the last +section is the +.Sy string +section, which contains the names of types, enumerations, members, and +labels. +.Lp +While strictly speaking, only the +.Sy preamble +and +.Sy header +are required, to be actually useful, both the type and string +sections are necessary. +.Lp +A +.Nm +file may contain all of the type information that it requires, or it +may optionally refer to another +.Nm +file which holds the remaining types. When a +.Nm +file refers to another file, it is called the +.Sy child +and the file it refers to is called the +.Sy parent . +A given file may only refer to one parent. This process is called +.Em uniquification +because it ensures each child only has type information that is +unique to it. A common example of this is that most kernel modules in +illumos are uniquified against the kernel module +.Sy genunix +and the type information that comes from the +.Sy IP +module. This means that a module only has types that are unique to +itself and the most common types in the kernel are not duplicated. +.Sh FILE FORMAT +This documents version +.Em two +of the +.Nm +file format. All applications and tools currently produce and operate on +this version. +.Lp +The file format can be summarized with the following image, the +following sections will cover this in more detail. +.Bd -literal + + +-------------+ 0t0 ++--------| Preamble | +| +-------------+ 0t4 +|+-------| Header | +|| +-------------+ 0t36 + cth_lbloff +||+------| Labels | +||| +-------------+ 0t36 + cth_objtoff +|||+-----| Objects | +|||| +-------------+ 0t36 + cth_funcoff +||||+----| Functions | +||||| +-------------+ 0t36 + cth_typeoff +|||||+---| Types | +|||||| +-------------+ 0t36 + cth_stroff +||||||+--| Strings | +||||||| +-------------+ 0t36 + cth_stroff + cth_strlen +||||||| +||||||| +||||||| +||||||| +-- magic - vers flags +||||||| | | | | +||||||| +------+------+------+------+ ++---------| 0xcf | 0xf1 | 0x02 | 0x00 | + |||||| +------+------+------+------+ + |||||| 0 1 2 3 4 + |||||| + |||||| + parent label + objects + |||||| | + parent name | + functions + strings + |||||| | | + label | | + types | + strlen + |||||| | | | | | | | | + |||||| +------+------+------+------+------+-------+-------+-------+ + +--------| 0x00 | 0x00 | 0x00 | 0x08 | 0x36 | 0x110 | 0x5f4 | 0x611 | + ||||| +------+------+------+------+------+-------+-------+-------+ + ||||| 0x04 0x08 0x0c 0x10 0x14 0x18 0x1c 0x20 0x24 + ||||| + ||||| + Label name + ||||| | + Label type + ||||| | | + Next label + ||||| | | | + ||||| +-------+------+-----+ + +-----------| 0x01 | 0x42 | ... | + |||| +-------+------+-----+ + |||| cth_lbloff +0x4 +0x8 cth_objtoff + |||| + |||| + |||| Symidx 0t15 0t43 0t44 + |||| +------+------+------+-----+ + +----------| 0x00 | 0x42 | 0x36 | ... | + ||| +------+------+------+-----+ + ||| cth_objtoff +0x2 +0x4 +0x6 cth_funcoff + ||| + ||| + CTF_TYPE_INFO + CTF_TYPE_INFO + ||| | + Return type | + ||| | | + arg0 | + ||| +--------+------+------+-----+ + +---------| 0x2c10 | 0x08 | 0x0c | ... | + || +--------+------+------+-----+ + || cth_funcff +0x2 +0x4 +0x6 cth_typeoff + || + || + ctf_stype_t for type 1 + || | integer + integer encoding + || | | + ctf_stype_t for type 2 + || | | | + || +--------------------+-----------+-----+ + +--------| 0x19 * 0xc01 * 0x0 | 0x1000000 | ... | + | +--------------------+-----------+-----+ + | cth_typeoff +0x08 +0x0c cth_stroff + | + | +--- str 0 + | | +--- str 1 + str 2 + | | | | + | v v v + | +----+---+---+---+----+---+---+---+---+---+----+ + +---| \\0 | i | n | t | \\0 | f | o | o | _ | t | \\0 | + +----+---+---+---+----+---+---+---+---+---+----+ + 0 1 2 3 4 5 6 7 8 9 10 11 +.Ed +.Lp +Every +.Nm +file begins with a +.Sy preamble , +followed by a +.Sy header . +The +.Sy preamble +is defined as follows: +.Bd -literal +typedef struct ctf_preamble { + ushort_t ctp_magic; /* magic number (CTF_MAGIC) */ + uchar_t ctp_version; /* data format version number (CTF_VERSION) */ + uchar_t ctp_flags; /* flags (see below) */ +} ctf_preamble_t; +.Ed +.Pp +The +.Sy preamble +is four bytes long and must be four byte aligned. +This +.Sy preamble +defines the version of the +.Nm +file which defines the format of the rest of the header. While the +header may change in subsequent versions, the preamble will not change +across versions, though the interpretation of its flags may change from +version to version. The +.Em ctp_magic +member defines the magic number for the +.Nm +file format. This must always be +.Li 0xcff1 . +If another value is encountered, then the file should not be treated as +a +.Nm +file. The +.Em ctp_version +member defines the version of the +.Nm +file. The current version is +.Li 2 . +It is possible to encounter an unsupported version. In that case, +software should not try to parse the format, as it may have changed. +Finally, the +.Em ctp_flags +member describes aspects of the file which modify its interpretation. +The following flags are currently defined: +.Bd -literal +#define CTF_F_COMPRESS 0x01 +.Ed +.Pp +The flag +.Sy CTF_F_COMPRESS +indicates that the body of the +.Nm +file, all the data following the +.Sy header , +has been compressed through the +.Sy zlib +library and its +.Sy deflate +algorithm. If this flag is not present, then the body has not been +compressed and no special action is needed to interpret it. All offsets +into the data as described by +.Sy header , +always refer to the +.Sy uncompressed +data. +.Lp +In version two of the +.Nm +file format, the +.Sy header +denotes whether whether or not this +.Nm +file is the child of another +.Nm +file and also indicates the size of the remaining sections. The +structure for the +.Sy header , +logically contains a copy of the +.Sy preamble +and the two have a combined size of 36 bytes. +.Bd -literal +typedef struct ctf_header { + ctf_preamble_t cth_preamble; + uint_t cth_parlabel; /* ref to name of parent lbl uniq'd against */ + uint_t cth_parname; /* ref to basename of parent */ + uint_t cth_lbloff; /* offset of label section */ + uint_t cth_objtoff; /* offset of object section */ + uint_t cth_funcoff; /* offset of function section */ + uint_t cth_typeoff; /* offset of type section */ + uint_t cth_stroff; /* offset of string section */ + uint_t cth_strlen; /* length of string section in bytes */ +} ctf_header_t; +.Ed +.Pp +After the +.Sy preamble , +the next two members +.Em cth_parlablel +and +.Em cth_parname , +are used to identify the parent. The value of both members are offsets +into the +.Sy string +section which point to the start of a null-terminated string. For more +information on the encoding of strings, see the subsection on +.Sx String Identifiers . +If the value of either is zero, then there is no entry for that +member. If the member +.Em cth_parlabel +is set, then the +.Em ctf_parname +member must be set, otherwise it will not be possible to find the +parent. If +.Em ctf_parname +is set, it is not necessary to define +.Em cth_parlabel , +as the parent may not have a label. For more information on labels +and their interpretation, see +.Sx The Label Section . +.Lp +The remaining members (excepting +.Em cth_strlen ) +describe the beginning of the corresponding sections. These offsets are +relative to the end of the +.Sy header . +Therefore, something with an offset of 0 is at an offset of thirty-six +bytes relative to the start of the +.Nm +file. The difference between members +indicates the size of the section itself. Different offsets have +different alignment requirements. The start of the +.Em cth_objotoff +and +.Em cth_funcoff +must be two byte aligned, while the sections +.Em cth_lbloff +and +.Em cth_typeoff +must be four-byte aligned. The section +.Em cth_stroff +has no alignment requirements. To calculate the size of a given section, +excepting the +.Sy string +section, one should subtract the offset of the section from the following one. For +example, the size of the +.Sy types +section can be calculated by subtracting +.Em cth_stroff +from +.Em cth_typeoff . +.Lp +Finally, the member +.Em cth_strlen +describes the length of the string section itself. From it, you can also +calculate the size of the entire +.Nm +file by adding together the size of the +.Sy ctf_header_t , +the offset of the string section in +.Em cth_stroff , +and the size of the string section in +.Em cth_srlen . +.Ss Type Identifiers +Through the +.Nm ctf +data, types are referred to by identifiers. A given +.Nm +file supports up to 32767 (0x7fff) types. The first valid type identifier is 0x1. +When a given +.Nm +file is a child, indicated by a non-zero entry for the +.Sy header Ns 's +.Em cth_parname , +then the first valid type identifier is 0x8000 and the last is 0xffff. +In this case, type identifiers 0x1 through 0x7fff are references to the +parent. +.Lp +The type identifier zero is a sentinel value used to indicate that there +is no type information available or it is an unknown type. +.Lp +Throughout the file format, the identifier is stored in different sized +values; however, the minimum size to represent a given identifier is a +.Sy uint16_t . +Other consumers of +.Nm +information may use larger or opaque identifiers. +.Ss String Identifiers +String identifiers are always encoded as four byte unsigned integers +which are an offset into a string table. The +.Nm +format supports two different string tables which have an identifier of +zero or one. This identifier is stored in the high-order bit of the +unsigned four byte offset. Therefore, the maximum supported offset into +one of these tables is 0x7ffffffff. +.Lp +Table identifier zero, always refers to the +.Sy string +section in the CTF file itself. String table identifier one refers to an +external string table which is the ELF string table for the ELF symbol +table associated with the +.Nm +container. +.Ss Type Encoding +Every +.Nm +type begins with metadata encoded into a +.Sy uint16_t . +This encoded information tells us three different pieces of information: +.Bl -bullet -offset indent -compact +.It +The kind of the type +.It +Whether this type is a root type or not +.It +The length of the variable data +.El +.Lp +The 16 bits that make up the encoding are broken down such that you have +five bits for the kind, one bit for indicating whether or not it is a +root type, and 10 bits for the variable length. This is laid out as +follows: +.Bd -literal -offset indent ++--------------------+ +| kind | root | vlen | ++--------------------+ +15 11 10 9 0 +.Ed +.Lp +The current version of the file format defines 14 different kinds. The +interpretation of these different kinds will be discussed in the section +.Sx The Type Section . +If a kind is encountered that is not listed below, then it is not a valid +.Nm +file. The kinds are defined as follows: +.Bd -literal -offset indent +#define CTF_K_UNKNOWN 0 +#define CTF_K_INTEGER 1 +#define CTF_K_FLOAT 2 +#define CTF_K_POINTER 3 +#define CTF_K_ARRAY 4 +#define CTF_K_FUNCTION 5 +#define CTF_K_STRUCT 6 +#define CTF_K_UNION 7 +#define CTF_K_ENUM 8 +#define CTF_K_FORWARD 9 +#define CTF_K_TYPEDEF 10 +#define CTF_K_VOLATILE 11 +#define CTF_K_CONST 12 +#define CTF_K_RESTRICT 13 +.Ed +.Lp +Programs directly reference many types; however, other types are referenced +indirectly because they are part of some other structure. These types that are +referenced directly and used are called +.Sy root +types. Other types may be used indirectly, for example, a program may reference +a structure directly, but not one of its members which has a type. That type is +not considered a +.Sy root +type. If a type is a +.Sy root +type, then it will have bit 10 set. +.Lp +The variable length section is specific to each kind and is discussed in the +section +.Sx The Type Section . +.Lp +The following macros are useful for constructing and deconstructing the encoded +type information: +.Bd -literal -offset indent + +#define CTF_MAX_VLEN 0x3ff +#define CTF_INFO_KIND(info) (((info) & 0xf800) >> 11) +#define CTF_INFO_ISROOT(info) (((info) & 0x0400) >> 10) +#define CTF_INFO_VLEN(info) (((info) & CTF_MAX_VLEN)) + +#define CTF_TYPE_INFO(kind, isroot, vlen) \\ + (((kind) << 11) | (((isroot) ? 1 : 0) << 10) | ((vlen) & CTF_MAX_VLEN)) +.Ed +.Ss The Label Section +When consuming +.Nm +data, it is often useful to know whether two different +.Nm +containers come from the same source base and version. For example, when +building illumos, there are many kernel modules that are built against a +single collection of source code. A label is encoded into the +.Nm +files that corresponds with the particular build. This ensures that if +files on the system were to become mixed up from multiple releases, that +they are not used together by tools, particularly when a child needs to +refer to a type in the parent. Because they are linked used the type +identifiers, if the wrong parent is used then the wrong type will be +encountered. +.Lp +Each label is encoded in the file format using the following eight byte +structure: +.Bd -literal +typedef struct ctf_lblent { + uint_t ctl_label; /* ref to name of label */ + uint_t ctl_typeidx; /* last type associated with this label */ +} ctf_lblent_t; +.Ed +.Lp +Each label has two different components, a name and a type identifier. +The name is encoded in the +.Em ctl_label +member which is in the format defined in the section +.Sx String Identifiers . +Generally, the names of all labels are found in the internal string +section. +.Lp +The type identifier encoded in the member +.Em ctl_typeidx +refers to the last type identifier that a label refers to in the current +file. Labels only refer to types in the current file, if the +.Nm +file is a child, then it will have the same label as its parent; +however, its label will only refer to its types, not its parents. +.Lp +It is also possible, though rather uncommon, for a +.Nm +file to have multiple labels. Labels are placed one after another, every +eight bytes. When multiple labels are present, types may only belong to +a single label. +.Ss The Object Section +The object section provides a mapping from ELF symbols of type +.Sy STT_OBJECT +in the symbol table to a type identifier. Every entry in this section is +a +.Sy uint16_t +which contains a type identifier as described in the section +.Sx Type Identifiers . +If there is no information for an object, then the type identifier 0x0 +is stored for that entry. +.Lp +To walk the object section, you need to have a corresponding +.Sy symbol table +in the ELF object that contains the +.Nm +data. Not every object is included in this section. Specifically, when +walking the symbol table. An entry is skipped if it matches any of the +following conditions: +.Lp +.Bl -bullet -offset indent -compact +.It +The symbol type is not +.Sy STT_OBJECT +.It +The symbol's section index is +.Sy SHN_UNDEF +.It +The symbol's name offset is zero +.It +The symbol's section index is +.Sy SHN_ABS +and the value of the symbol is zero. +.It +The symbol's name is +.Li _START_ +or +.Li _END_ . +These are skipped because they are used for scoping local symbols in +ELF. +.El +.Lp +The following sample code shows an example of iterating the object +section and skipping the correct symbols: +.Bd -literal +#include +#include + +/* + * Given the start of the object section in the CTF file, the number of symbols, + * and the ELF Data sections for the symbol table and the string table, this + * prints the type identifiers that correspond to objects. Note, a more robust + * implementation should ensure that they don't walk beyond the end of the CTF + * object section. + */ +static int +walk_symbols(uint16_t *objtoff, Elf_Data *symdata, Elf_Data *strdata, + long nsyms) +{ + long i; + uintptr_t strbase = strdata->d_buf; + + for (i = 1; i < nsyms; i++, objftoff++) { + const char *name; + GElf_Sym sym; + + if (gelf_getsym(symdata, i, &sym) == NULL) + return (1); + + if (GELF_ST_TYPE(sym.st_info) != STT_OBJECT) + continue; + if (sym.st_shndx == SHN_UNDEF || sym.st_name == 0) + continue; + if (sym.st_shndx == SHN_ABS && sym.st_value == 0) + continue; + name = (const char *)(strbase + sym.st_name); + if (strcmp(name, "_START_") == 0 || strcmp(name, "_END_") == 0) + continue; + + (void) printf("Symbol %d has type %d\n", i, *objtoff); + } + + return (0); +} +.Ed +.Ss The Function Section +The function section of the +.Nm +file encodes the types of both the function's arguments and the function's +return type. Similar to +.Sx The Object Section , +the function section encodes information for all symbols of type +.Sy STT_FUNCTION , +excepting those that fit specific criteria. Unlike with objects, because +functions have a variable number of arguments, they start with a type encoding +as defined in +.Sx Type Encoding , +which is the size of a +.Sy uint16_t . +For functions which have no type information available, they are encoded as +.Li CTF_TYPE_INFO(CTF_K_UNKNOWN, 0, 0) . +Functions with arguments are encoded differently. Here, the variable length is +turned into the number of arguments in the function. If a function is a +.Sy varargs +type function, then the number of arguments is increased by one. Functions with +type information are encoded as: +.Li CTF_TYPE_INFO(CTF_K_FUNCTION, 0, nargs) . +.Lp +For functions that have no type information, nothing else is encoded, and the +next function is encoded. For functions with type information, the next +.Sy uint16_t +is encoded with the type identifier of the return type of the function. It is +followed by each of the type identifiers of the arguments, if any exist, in the +order that they appear in the function. Therefore, argument 0 is the first type +identifier and so on. When a function has a final varargs argument, that is +encoded with the type identifier of zero. +.Lp +Like +.Sx The Object Section , +the function section is encoded in the order of the symbol table. It has +similar, but slightly different considerations from objects. While iterating the +symbol table, if any of the following conditions are true, then the entry is +skipped and no corresponding entry is written: +.Lp +.Bl -bullet -offset indent -compact +.It +The symbol type is not +.Sy STT_FUNCTION +.It +The symbol's section index is +.Sy SHN_UNDEF +.It +The symbol's name offset is zero +.It +The symbol's name is +.Li _START_ +or +.Li _END_ . +These are skipped because they are used for scoping local symbols in +ELF. +.El +.Ss The Type Section +The type section is the heart of the +.Nm +data. It encodes all of the information about the types themselves. The base of +the type information comes in two forms, a short form and a long form, each of +which may be followed by a variable number of arguments. The following +definitions describe the short and long forms: +.Bd -literal +#define CTF_MAX_SIZE 0xfffe /* max size of a type in bytes */ +#define CTF_LSIZE_SENT 0xffff /* sentinel for ctt_size */ +#define CTF_MAX_LSIZE UINT64_MAX + +typedef struct ctf_stype { + uint_t ctt_name; /* reference to name in string table */ + ushort_t ctt_info; /* encoded kind, variant length */ + union { + ushort_t _size; /* size of entire type in bytes */ + ushort_t _type; /* reference to another type */ + } _u; +} ctf_stype_t; + +typedef struct ctf_type { + uint_t ctt_name; /* reference to name in string table */ + ushort_t ctt_info; /* encoded kind, variant length */ + union { + ushort_t _size; /* always CTF_LSIZE_SENT */ + ushort_t _type; /* do not use */ + } _u; + uint_t ctt_lsizehi; /* high 32 bits of type size in bytes */ + uint_t ctt_lsizelo; /* low 32 bits of type size in bytes */ +} ctf_type_t; + +#define ctt_size _u._size /* for fundamental types that have a size */ +#define ctt_type _u._type /* for types that reference another type */ +.Ed +.Pp +Type sizes are stored in +.Sy bytes . +The basic small form uses a +.Sy ushort_t +to store the number of bytes. If the number of bytes in a structure would exceed +0xfffe, then the alternate form, the +.Sy ctf_type_t , +is used instead. To indicate that the larger form is being used, the member +.Em ctt_size +is set to value of +.Sy CTF_LSIZE_SENT +(0xffff). In general, when going through the type section, consumers use the +.Sy ctf_type_t +structure, but pay attention to the value of the member +.Em ctt_size +to determine whether they should increment their scan by the size of the +.Sy ctf_stype_t +or +.Sy ctf_type_t . +Not all kinds of types use +.Sy ctt_size . +Those which do not, will always use the +.Sy ctf_stype_t +structure. The individual sections for each kind have more information. +.Lp +Types are written out in order. Therefore the first entry encountered has a type +id of 0x1, or 0x8000 if a child. The member +.Em ctt_name +is encoded as described in the section +.Sx String Identifiers . +The string that it points to is the name of the type. If the identifier points +to an empty string (one that consists solely of a null terminator) then the type +does not have a name, this is common with anonymous structures and unions that +only have a typedef to name them, as well as, pointers and qualifiers. +.Lp +The next member, the +.Em ctt_info , +is encoded as described in the section +.Sx Type Encoding . +The types kind tells us how to interpret the remaining data in the +.Sy ctf_type_t +and any variable length data that may exist. The rest of this section will be +broken down into the interpretation of the various kinds. +.Ss Encoding of Integers +Integers, which are of type +.Sy CTF_K_INTEGER , +have no variable length arguments. Instead, they are followed by a four byte +.Sy uint_t +which describes their encoding. All integers must be encoded with a variable +length of zero. The +.Em ctt_size +member describes the length of the integer in bytes. In general, integer sizes +will be rounded up to the closest power of two. +.Lp +The integer encoding contains three different pieces of information: +.Bl -bullet -offset indent -compact +.It +The encoding of the integer +.It +The offset in +.Sy bits +of the type +.It +The size in +.Sy bits +of the type +.El +.Pp +This encoding can be expressed through the following macros: +.Bd -literal -offset indent +#define CTF_INT_ENCODING(data) (((data) & 0xff000000) >> 24) +#define CTF_INT_OFFSET(data) (((data) & 0x00ff0000) >> 16) +#define CTF_INT_BITS(data) (((data) & 0x0000ffff)) + +#define CTF_INT_DATA(encoding, offset, bits) \\ + (((encoding) << 24) | ((offset) << 16) | (bits)) +.Ed +.Pp +The following flags are defined for the encoding at this time: +.Bd -literal -offset indent +#define CTF_INT_SIGNED 0x01 +#define CTF_INT_CHAR 0x02 +#define CTF_INT_BOOL 0x04 +#define CTF_INT_VARARGS 0x08 +.Ed +.Lp +By default, an integer is considered to be unsigned, unless it has the +.Sy CTF_INT_SIGNED +flag set. If the flag +.Sy CTF_INT_CHAR +is set, that indicates that the integer is of a type that stores character +data, for example the intrinsic C type +.Sy char +would have the +.Sy CTF_INT_CHAR +flag set. If the flag +.Sy CTF_INT_BOOL +is set, that indicates that the integer represents a boolean type. For example, +the intrinsic C type +.Sy _Bool +would have the +.Sy CTF_INT_BOOL +flag set. Finally, the flag +.Sy CTF_INT_VARARGS +indicates that the integer is used as part of a variable number of arguments. +This encoding is rather uncommon. +.Ss Encoding of Floats +Floats, which are of type +.Sy CTF_K_FLOAT , +are similar to their integer counterparts. They have no variable length +arguments and are followed by a four byte encoding which describes the kind of +float that exists. The +.Em ctt_size +member is the size, in bytes, of the float. The float encoding has three +different pieces of information inside of it: +.Lp +.Bl -bullet -offset indent -compact +.It +The specific kind of float that exists +.It +The offset in +.Sy bits +of the float +.It +The size in +.Sy bits +of the float +.El +.Lp +This encoding can be expressed through the following macros: +.Bd -literal -offset indent +#define CTF_FP_ENCODING(data) (((data) & 0xff000000) >> 24) +#define CTF_FP_OFFSET(data) (((data) & 0x00ff0000) >> 16) +#define CTF_FP_BITS(data) (((data) & 0x0000ffff)) + +#define CTF_FP_DATA(encoding, offset, bits) \\ + (((encoding) << 24) | ((offset) << 16) | (bits)) +.Ed +.Lp +Where as the encoding for integers was a series of flags, the encoding for +floats maps to a specific kind of float. It is not a flag-based value. The kinds of floats +correspond to both their size, and the encoding. This covers all of the basic C +intrinsic floating point types. The following are the different kinds of floats +represented in the encoding: +.Bd -literal -offset indent +#define CTF_FP_SINGLE 1 /* IEEE 32-bit float encoding */ +#define CTF_FP_DOUBLE 2 /* IEEE 64-bit float encoding */ +#define CTF_FP_CPLX 3 /* Complex encoding */ +#define CTF_FP_DCPLX 4 /* Double complex encoding */ +#define CTF_FP_LDCPLX 5 /* Long double complex encoding */ +#define CTF_FP_LDOUBLE 6 /* Long double encoding */ +#define CTF_FP_INTRVL 7 /* Interval (2x32-bit) encoding */ +#define CTF_FP_DINTRVL 8 /* Double interval (2x64-bit) encoding */ +#define CTF_FP_LDINTRVL 9 /* Long double interval (2x128-bit) encoding */ +#define CTF_FP_IMAGRY 10 /* Imaginary (32-bit) encoding */ +#define CTF_FP_DIMAGRY 11 /* Long imaginary (64-bit) encoding */ +#define CTF_FP_LDIMAGRY 12 /* Long double imaginary (128-bit) encoding */ +.Ed +.Ss Encoding of Arrays +Arrays, which are of type +.Sy CTF_K_ARRAY , +have no variable length arguments. They are followed by a structure which +describes the number of elements in the array, the type identifier of the +elements in the array, and the type identifier of the index of the array. With +arrays, the +.Em ctt_size +member is set to zero. The structure that follows an array is defined as: +.Bd -literal +typedef struct ctf_array { + ushort_t cta_contents; /* reference to type of array contents */ + ushort_t cta_index; /* reference to type of array index */ + uint_t cta_nelems; /* number of elements */ +} ctf_array_t; +.Ed +.Lp +The +.Em cta_contents +and +.Em cta_index +members of the +.Sy ctf_array_t +are type identifiers which are encoded as per the section +.Sx Type Identifiers . +The member +.Em cta_nelems +is a simple four byte unsigned count of the number of elements. This count may +be zero when encountering C99's flexible array members. +.Ss Encoding of Functions +Function types, which are of type +.Sy CTF_K_FUNCTION , +use the variable length list to be the number of arguments in the function. When +the function has a final member which is a varargs, then the argument count is +incremented by one to account for the variable argument. Here, the +.Em ctt_type +member is encoded with the type identifier of the return type of the function. +Note that the +.Em ctt_size +member is not used here. +.Lp +The variable argument list contains the type identifiers for the arguments of +the function, if any. Each one is represented by a +.Sy uint16_t +and encoded according to the +.Sx Type Identifiers +section. If the function's last argument is of type varargs, then it is also +written out, but the type identifier is zero. This is included in the count of +the function's arguments. +.Ss Encoding of Structures and Unions +Structures and Unions, which are encoded with +.Sy CTF_K_STRUCT +and +.Sy CTF_K_UNION +respectively, are very similar constructs in C. The main difference +between them is the fact that every member of a structure follows one another, +where as in a union, all members share the same memory. They are also very +similar in terms of their encoding in +.Nm . +The variable length argument for structures and unions represents the number of +members that they have. The value of the member +.Em ctt_size +is the size of the structure and union. There are two different structures which +are used to encode members in the variable list. When the size of a structure or +union is greater than or equal to the large member threshold, 8192, then a +different structure is used to encode the member, all members are encoded using +the same structure. The structure for members is as follows: +.Bd -literal +typedef struct ctf_member { + uint_t ctm_name; /* reference to name in string table */ + ushort_t ctm_type; /* reference to type of member */ + ushort_t ctm_offset; /* offset of this member in bits */ +} ctf_member_t; + +typedef struct ctf_lmember { + uint_t ctlm_name; /* reference to name in string table */ + ushort_t ctlm_type; /* reference to type of member */ + ushort_t ctlm_pad; /* padding */ + uint_t ctlm_offsethi; /* high 32 bits of member offset in bits */ + uint_t ctlm_offsetlo; /* low 32 bits of member offset in bits */ +} ctf_lmember_t; +.Ed +.Lp +Both the +.Em ctm_name +and +.Em ctlm_name +refer to the name of the member. The name is encoded as an offset into the +string table as described by the section +.Sx String Identifiers . +The members +.Sy ctm_type +and +.Sy ctlm_type +both refer to the type of the member. They are encoded as per the section +.Sx Type Identifiers . +.Lp +The last piece of information that is present is the offset which describes the +offset in memory that the member begins at. For unions, this value will always +be zero because the start of unions in memory is always zero. For structures, +this is the offset in +.Sy bits +that the member begins at. Note that a compiler may lay out a type with padding. +This means that the difference in offset between two consecutive members may be +larger than the size of the member. When the size of the overall structure is +strictly less than 8192 bytes, the normal structure, +.Sy ctf_member_t , +is used and the offset in bits is stored in the member +.Em ctm_offset . +However, when the size of the structure is greater than or equal to 8192 bytes, +then the number of bits is split into two 32-bit quantities. One member, +.Em ctlm_offsethi , +represents the upper 32 bits of the offset, while the other member, +.Em ctlm_offsetlo , +represents the lower 32 bits of the offset. These can be joined together to get +a 64-bit sized offset in bits by shifting the member +.Em ctlm_offsethi +to the left by thirty two and then doing a binary or of +.Em ctlm_offsetlo . +.Ss Encoding of Enumerations +Enumerations, noted by the type +.Sy CTF_K_ENUM , +are similar to structures. Enumerations use the variable list to note the number +of values that the enumeration contains, which we'll term enumerators. In C, an +enumeration is always equivalent to the intrinsic type +.Sy int , +thus the value of the member +.Em ctt_size +is always the size of an integer which is determined based on the current model. +For illumos systems, this will always be 4, as an integer is always defined to +be 4 bytes large in both +.Sy ILP32 +and +.Sy LP64 , +regardless of the architecture. +.Lp +The enumerators encoded in an enumeration have the following structure in the +variable list: +.Bd -literal +typedef struct ctf_enum { + uint_t cte_name; /* reference to name in string table */ + int cte_value; /* value associated with this name */ +} ctf_enum_t; +.Ed +.Pp +The member +.Em cte_name +refers to the name of the enumerator's value, it is encoded according to the +rules in the section +.Sx String Identifiers . +The member +.Em cte_value +contains the integer value of this enumerator. +.Ss Encoding of Forward References +Forward references, types of kind +.Sy CTF_K_FORWARD , +in a +.Nm +file refer to types which may not have a definition at all, only a name. If +the +.Nm +file is a child, then it may be that the forward is resolved to an +actual type in the parent, otherwise the definition may be in another +.Nm +container or may not be known at all. The only member of the +.Sy ctf_type_t +that matters for a forward declaration is the +.Em ctt_name +which points to the name of the forward reference in the string table as +described earlier. There is no other information recorded for forward +references. +.Ss Encoding of Pointers, Typedefs, Volatile, Const, and Restrict +Pointers, typedefs, volatile, const, and restrict are all similar in +.Nm . +They all refer to another type. In the case of typedefs, they provide an +alternate name, while volatile, const, and restrict change how the type is +interpreted in the C programming language. This covers the +.Nm +kinds +.Sy CTF_K_POINTER , +.Sy CTF_K_TYPEDEF , +.Sy CTF_K_VOLATILE , +.Sy CTF_K_RESTRICT , +and +.Sy CTF_K_CONST . +.Lp +These types have no variable list entries and use the member +.Em ctt_type +to refer to the base type that they modify. +.Ss Encoding of Unknown Types +Types with the kind +.Sy CTF_K_UNKNOWN +are used to indicate gaps in the type identifier space. These entries consume an +identifier, but do not define anything. Nothing should refer to these gap +identifiers. +.Ss Dependencies Between Types +C types can be imagined as a directed, cyclic, graph. Structures and unions may +refer to each other in a way that creates a cyclic dependency. In cases such as +these, the entire type section must be read in and processed. Consumers must +not assume that every type can be laid out in dependency order; they +cannot. +.Ss The String Section +The last section of the +.Nm +file is the +.Sy string +section. This section encodes all of the strings that appear throughout +the other sections. It is laid out as a series of characters followed by +a null terminator. Generally, all names are written out in ASCII, as +most C compilers do not allow and characters to appear in identifiers +outside of a subset of ASCII. However, any extended characters sets +should be written out as a series of UTF-8 bytes. +.Lp +The first entry in the section, at offset zero, is a single null +terminator to reference the empty string. Following that, each C string +should be written out, including the null terminator. Offsets that refer +to something in this section should refer to the first byte which begins +a string. Beyond the first byte in the section being the null +terminator, the order of strings is unimportant. +.Ss Data Encoding and ELF Considerations +.Nm +data is generally included in ELF objects which specify information to +identify the architecture and endianness of the file. A +.Nm +container inside such an object must match the endianness of the ELF +object. Aside from the question of the endian encoding of data, there +should be no other differences between architectures. While many of the +types in this document refer to non-fixed size C integral types, they +are equivalent in the models +.Sy ILP32 +and +.Sy LP64 . +If any other model is being used with +.Nm +data that has different sizes, then it must not use the model's sizes for +those integral types and instead use the fixed size equivalents based on an +.Sy ILP32 +environment. +.Lp +When placing a +.Nm +container inside of an ELF object, there are certain conventions that are +expected for the purposes of tooling being able to find the +.Nm +data. In particular, a given ELF object should only contain a single +.Nm +section. Multiple containers should be merged together into a single +one. +.Lp +The +.Nm +file should be included in its own ELF section. The section's name +must be +.Ql .SUNW_ctf . +The type of the section must be +.Sy SHT_PROGBITS . +The section should have a link set to the symbol table and its address +alignment must be 4. +.Sh SEE ALSO +.Xr dtrace 1 , +.Xr elf 3 , +.Xr gelf 3 , +.Xr a.out 5 , +.Xr elf 5 Property changes on: projects/building-blocks/cddl/contrib/opensolaris/lib/libctf/common/ctf.5 ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/cddl/contrib/opensolaris =================================================================== --- projects/building-blocks/cddl/contrib/opensolaris (revision 275663) +++ projects/building-blocks/cddl/contrib/opensolaris (revision 275664) Property changes on: projects/building-blocks/cddl/contrib/opensolaris ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,2 ## Merged /vendor/illumos/dist:r272851 Merged /head/cddl/contrib/opensolaris:r275590-275663 Index: projects/building-blocks/cddl/lib/libctf/Makefile =================================================================== --- projects/building-blocks/cddl/lib/libctf/Makefile (revision 275663) +++ projects/building-blocks/cddl/lib/libctf/Makefile (revision 275664) @@ -1,34 +1,35 @@ # $FreeBSD$ .PATH: ${.CURDIR}/../../../cddl/contrib/opensolaris/common/ctf .PATH: ${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libctf/common .PATH: ${.CURDIR}/../../../sys/cddl/contrib/opensolaris/common/ctf LIB= ctf SRCS= ctf_create.c \ ctf_decl.c \ ctf_error.c \ ctf_hash.c \ ctf_labels.c \ ctf_lib.c \ ctf_lookup.c \ ctf_open.c \ ctf_subr.c \ ctf_types.c \ ctf_util.c +MAN= ctf.5 WARNS?= 0 CFLAGS+= -DCTF_OLD_VERSIONS CFLAGS+= -I${.CURDIR}/../../../sys/cddl/compat/opensolaris \ -I${.CURDIR}/../../../cddl/compat/opensolaris/include \ -I${OPENSOLARIS_USR_DISTDIR}/head \ -I${OPENSOLARIS_USR_DISTDIR}/common/ctf \ -I${OPENSOLARIS_USR_DISTDIR}/lib/libctf/common \ -I${OPENSOLARIS_SYS_DISTDIR}/uts/common DPADD= ${LIBZ} LDADD= -lz .include Index: projects/building-blocks/cddl =================================================================== --- projects/building-blocks/cddl (revision 275663) +++ projects/building-blocks/cddl (revision 275664) Property changes on: projects/building-blocks/cddl ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/cddl:r275590-275663 Index: projects/building-blocks/contrib/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp =================================================================== --- projects/building-blocks/contrib/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp (revision 275663) +++ projects/building-blocks/contrib/llvm/lib/Transforms/Vectorize/LoopVectorize.cpp (revision 275664) @@ -1,5166 +1,5174 @@ //===- LoopVectorize.cpp - A Loop Vectorizer ------------------------------===// // // The LLVM Compiler Infrastructure // // This file is distributed under the University of Illinois Open Source // License. See LICENSE.TXT for details. // //===----------------------------------------------------------------------===// // // This is the LLVM loop vectorizer. This pass modifies 'vectorizable' loops // and generates target-independent LLVM-IR. // The vectorizer uses the TargetTransformInfo analysis to estimate the costs // of instructions in order to estimate the profitability of vectorization. // // The loop vectorizer combines consecutive loop iterations into a single // 'wide' iteration. After this transformation the index is incremented // by the SIMD vector width, and not by one. // // This pass has three parts: // 1. The main loop pass that drives the different parts. // 2. LoopVectorizationLegality - A unit that checks for the legality // of the vectorization. // 3. InnerLoopVectorizer - A unit that performs the actual // widening of instructions. // 4. LoopVectorizationCostModel - A unit that checks for the profitability // of vectorization. It decides on the optimal vector width, which // can be one, if vectorization is not profitable. // //===----------------------------------------------------------------------===// // // The reduction-variable vectorization is based on the paper: // D. Nuzman and R. Henderson. Multi-platform Auto-vectorization. // // Variable uniformity checks are inspired by: // Karrenberg, R. and Hack, S. Whole Function Vectorization. // // Other ideas/concepts are from: // A. Zaks and D. Nuzman. Autovectorization in GCC-two years later. // // S. Maleki, Y. Gao, M. Garzaran, T. Wong and D. Padua. An Evaluation of // Vectorizing Compilers. // //===----------------------------------------------------------------------===// #define LV_NAME "loop-vectorize" #define DEBUG_TYPE LV_NAME #include "llvm/Transforms/Vectorize.h" #include "llvm/ADT/DenseMap.h" #include "llvm/ADT/EquivalenceClasses.h" #include "llvm/ADT/Hashing.h" #include "llvm/ADT/MapVector.h" #include "llvm/ADT/SetVector.h" #include "llvm/ADT/SmallPtrSet.h" #include "llvm/ADT/SmallSet.h" #include "llvm/ADT/SmallVector.h" #include "llvm/ADT/StringExtras.h" #include "llvm/Analysis/AliasAnalysis.h" #include "llvm/Analysis/Dominators.h" #include "llvm/Analysis/LoopInfo.h" #include "llvm/Analysis/LoopIterator.h" #include "llvm/Analysis/LoopPass.h" #include "llvm/Analysis/ScalarEvolution.h" #include "llvm/Analysis/ScalarEvolutionExpander.h" #include "llvm/Analysis/ScalarEvolutionExpressions.h" #include "llvm/Analysis/TargetTransformInfo.h" #include "llvm/Analysis/ValueTracking.h" #include "llvm/Analysis/Verifier.h" #include "llvm/IR/Constants.h" #include "llvm/IR/DataLayout.h" #include "llvm/IR/DerivedTypes.h" #include "llvm/IR/Function.h" #include "llvm/IR/IRBuilder.h" #include "llvm/IR/Instructions.h" #include "llvm/IR/IntrinsicInst.h" #include "llvm/IR/LLVMContext.h" #include "llvm/IR/Module.h" #include "llvm/IR/Type.h" #include "llvm/IR/Value.h" #include "llvm/Pass.h" #include "llvm/Support/CommandLine.h" #include "llvm/Support/Debug.h" #include "llvm/Support/PatternMatch.h" #include "llvm/Support/raw_ostream.h" #include "llvm/Support/ValueHandle.h" #include "llvm/Target/TargetLibraryInfo.h" #include "llvm/Transforms/Scalar.h" #include "llvm/Transforms/Utils/BasicBlockUtils.h" #include "llvm/Transforms/Utils/Local.h" #include #include using namespace llvm; using namespace llvm::PatternMatch; static cl::opt VectorizationFactor("force-vector-width", cl::init(0), cl::Hidden, cl::desc("Sets the SIMD width. Zero is autoselect.")); static cl::opt VectorizationUnroll("force-vector-unroll", cl::init(0), cl::Hidden, cl::desc("Sets the vectorization unroll count. " "Zero is autoselect.")); static cl::opt EnableIfConversion("enable-if-conversion", cl::init(true), cl::Hidden, cl::desc("Enable if-conversion during vectorization.")); /// We don't vectorize loops with a known constant trip count below this number. static cl::opt TinyTripCountVectorThreshold("vectorizer-min-trip-count", cl::init(16), cl::Hidden, cl::desc("Don't vectorize loops with a constant " "trip count that is smaller than this " "value.")); /// We don't unroll loops with a known constant trip count below this number. static const unsigned TinyTripCountUnrollThreshold = 128; /// When performing memory disambiguation checks at runtime do not make more /// than this number of comparisons. static const unsigned RuntimeMemoryCheckThreshold = 8; /// Maximum simd width. static const unsigned MaxVectorWidth = 64; /// Maximum vectorization unroll count. static const unsigned MaxUnrollFactor = 16; /// The cost of a loop that is considered 'small' by the unroller. static const unsigned SmallLoopCost = 20; namespace { // Forward declarations. class LoopVectorizationLegality; class LoopVectorizationCostModel; /// InnerLoopVectorizer vectorizes loops which contain only one basic /// block to a specified vectorization factor (VF). /// This class performs the widening of scalars into vectors, or multiple /// scalars. This class also implements the following features: /// * It inserts an epilogue loop for handling loops that don't have iteration /// counts that are known to be a multiple of the vectorization factor. /// * It handles the code generation for reduction variables. /// * Scalarization (implementation using scalars) of un-vectorizable /// instructions. /// InnerLoopVectorizer does not perform any vectorization-legality /// checks, and relies on the caller to check for the different legality /// aspects. The InnerLoopVectorizer relies on the /// LoopVectorizationLegality class to provide information about the induction /// and reduction variables that were found to a given vectorization factor. class InnerLoopVectorizer { public: InnerLoopVectorizer(Loop *OrigLoop, ScalarEvolution *SE, LoopInfo *LI, DominatorTree *DT, DataLayout *DL, const TargetLibraryInfo *TLI, unsigned VecWidth, unsigned UnrollFactor) : OrigLoop(OrigLoop), SE(SE), LI(LI), DT(DT), DL(DL), TLI(TLI), VF(VecWidth), UF(UnrollFactor), Builder(SE->getContext()), Induction(0), OldInduction(0), WidenMap(UnrollFactor) {} // Perform the actual loop widening (vectorization). void vectorize(LoopVectorizationLegality *Legal) { // Create a new empty loop. Unlink the old loop and connect the new one. createEmptyLoop(Legal); // Widen each instruction in the old loop to a new one in the new loop. // Use the Legality module to find the induction and reduction variables. vectorizeLoop(Legal); // Register the new loop and update the analysis passes. updateAnalysis(); } virtual ~InnerLoopVectorizer() {} protected: /// A small list of PHINodes. typedef SmallVector PhiVector; /// When we unroll loops we have multiple vector values for each scalar. /// This data structure holds the unrolled and vectorized values that /// originated from one scalar instruction. typedef SmallVector VectorParts; // When we if-convert we need create edge masks. We have to cache values so // that we don't end up with exponential recursion/IR. typedef DenseMap, VectorParts> EdgeMaskCache; /// Add code that checks at runtime if the accessed arrays overlap. /// Returns the comparator value or NULL if no check is needed. Instruction *addRuntimeCheck(LoopVectorizationLegality *Legal, Instruction *Loc); /// Create an empty loop, based on the loop ranges of the old loop. void createEmptyLoop(LoopVectorizationLegality *Legal); /// Copy and widen the instructions from the old loop. virtual void vectorizeLoop(LoopVectorizationLegality *Legal); /// \brief The Loop exit block may have single value PHI nodes where the /// incoming value is 'Undef'. While vectorizing we only handled real values /// that were defined inside the loop. Here we fix the 'undef case'. /// See PR14725. void fixLCSSAPHIs(); /// A helper function that computes the predicate of the block BB, assuming /// that the header block of the loop is set to True. It returns the *entry* /// mask for the block BB. VectorParts createBlockInMask(BasicBlock *BB); /// A helper function that computes the predicate of the edge between SRC /// and DST. VectorParts createEdgeMask(BasicBlock *Src, BasicBlock *Dst); /// A helper function to vectorize a single BB within the innermost loop. void vectorizeBlockInLoop(LoopVectorizationLegality *Legal, BasicBlock *BB, PhiVector *PV); /// Vectorize a single PHINode in a block. This method handles the induction /// variable canonicalization. It supports both VF = 1 for unrolled loops and /// arbitrary length vectors. void widenPHIInstruction(Instruction *PN, VectorParts &Entry, LoopVectorizationLegality *Legal, unsigned UF, unsigned VF, PhiVector *PV); /// Insert the new loop to the loop hierarchy and pass manager /// and update the analysis passes. void updateAnalysis(); /// This instruction is un-vectorizable. Implement it as a sequence /// of scalars. virtual void scalarizeInstruction(Instruction *Instr); /// Vectorize Load and Store instructions, virtual void vectorizeMemoryInstruction(Instruction *Instr, LoopVectorizationLegality *Legal); /// Create a broadcast instruction. This method generates a broadcast /// instruction (shuffle) for loop invariant values and for the induction /// value. If this is the induction variable then we extend it to N, N+1, ... /// this is needed because each iteration in the loop corresponds to a SIMD /// element. virtual Value *getBroadcastInstrs(Value *V); /// This function adds 0, 1, 2 ... to each vector element, starting at zero. /// If Negate is set then negative numbers are added e.g. (0, -1, -2, ...). /// The sequence starts at StartIndex. virtual Value *getConsecutiveVector(Value* Val, int StartIdx, bool Negate); /// When we go over instructions in the basic block we rely on previous /// values within the current basic block or on loop invariant values. /// When we widen (vectorize) values we place them in the map. If the values /// are not within the map, they have to be loop invariant, so we simply /// broadcast them into a vector. VectorParts &getVectorValue(Value *V); /// Generate a shuffle sequence that will reverse the vector Vec. virtual Value *reverseVector(Value *Vec); /// This is a helper class that holds the vectorizer state. It maps scalar /// instructions to vector instructions. When the code is 'unrolled' then /// then a single scalar value is mapped to multiple vector parts. The parts /// are stored in the VectorPart type. struct ValueMap { /// C'tor. UnrollFactor controls the number of vectors ('parts') that /// are mapped. ValueMap(unsigned UnrollFactor) : UF(UnrollFactor) {} /// \return True if 'Key' is saved in the Value Map. bool has(Value *Key) const { return MapStorage.count(Key); } /// Initializes a new entry in the map. Sets all of the vector parts to the /// save value in 'Val'. /// \return A reference to a vector with splat values. VectorParts &splat(Value *Key, Value *Val) { VectorParts &Entry = MapStorage[Key]; Entry.assign(UF, Val); return Entry; } ///\return A reference to the value that is stored at 'Key'. VectorParts &get(Value *Key) { VectorParts &Entry = MapStorage[Key]; if (Entry.empty()) Entry.resize(UF); assert(Entry.size() == UF); return Entry; } private: /// The unroll factor. Each entry in the map stores this number of vector /// elements. unsigned UF; /// Map storage. We use std::map and not DenseMap because insertions to a /// dense map invalidates its iterators. std::map MapStorage; }; /// The original loop. Loop *OrigLoop; /// Scev analysis to use. ScalarEvolution *SE; /// Loop Info. LoopInfo *LI; /// Dominator Tree. DominatorTree *DT; /// Data Layout. DataLayout *DL; /// Target Library Info. const TargetLibraryInfo *TLI; /// The vectorization SIMD factor to use. Each vector will have this many /// vector elements. unsigned VF; protected: /// The vectorization unroll factor to use. Each scalar is vectorized to this /// many different vector instructions. unsigned UF; /// The builder that we use IRBuilder<> Builder; // --- Vectorization state --- /// The vector-loop preheader. BasicBlock *LoopVectorPreHeader; /// The scalar-loop preheader. BasicBlock *LoopScalarPreHeader; /// Middle Block between the vector and the scalar. BasicBlock *LoopMiddleBlock; ///The ExitBlock of the scalar loop. BasicBlock *LoopExitBlock; ///The vector loop body. BasicBlock *LoopVectorBody; ///The scalar loop body. BasicBlock *LoopScalarBody; /// A list of all bypass blocks. The first block is the entry of the loop. SmallVector LoopBypassBlocks; /// The new Induction variable which was added to the new block. PHINode *Induction; /// The induction variable of the old basic block. PHINode *OldInduction; /// Holds the extended (to the widest induction type) start index. Value *ExtendedIdx; /// Maps scalars to widened vectors. ValueMap WidenMap; EdgeMaskCache MaskCache; }; class InnerLoopUnroller : public InnerLoopVectorizer { public: InnerLoopUnroller(Loop *OrigLoop, ScalarEvolution *SE, LoopInfo *LI, DominatorTree *DT, DataLayout *DL, const TargetLibraryInfo *TLI, unsigned UnrollFactor) : InnerLoopVectorizer(OrigLoop, SE, LI, DT, DL, TLI, 1, UnrollFactor) { } private: virtual void scalarizeInstruction(Instruction *Instr); virtual void vectorizeMemoryInstruction(Instruction *Instr, LoopVectorizationLegality *Legal); virtual Value *getBroadcastInstrs(Value *V); virtual Value *getConsecutiveVector(Value* Val, int StartIdx, bool Negate); virtual Value *reverseVector(Value *Vec); }; /// \brief Look for a meaningful debug location on the instruction or it's /// operands. static Instruction *getDebugLocFromInstOrOperands(Instruction *I) { if (!I) return I; DebugLoc Empty; if (I->getDebugLoc() != Empty) return I; for (User::op_iterator OI = I->op_begin(), OE = I->op_end(); OI != OE; ++OI) { if (Instruction *OpInst = dyn_cast(*OI)) if (OpInst->getDebugLoc() != Empty) return OpInst; } return I; } /// \brief Set the debug location in the builder using the debug location in the /// instruction. static void setDebugLocFromInst(IRBuilder<> &B, const Value *Ptr) { if (const Instruction *Inst = dyn_cast_or_null(Ptr)) B.SetCurrentDebugLocation(Inst->getDebugLoc()); else B.SetCurrentDebugLocation(DebugLoc()); } /// LoopVectorizationLegality checks if it is legal to vectorize a loop, and /// to what vectorization factor. /// This class does not look at the profitability of vectorization, only the /// legality. This class has two main kinds of checks: /// * Memory checks - The code in canVectorizeMemory checks if vectorization /// will change the order of memory accesses in a way that will change the /// correctness of the program. /// * Scalars checks - The code in canVectorizeInstrs and canVectorizeMemory /// checks for a number of different conditions, such as the availability of a /// single induction variable, that all types are supported and vectorize-able, /// etc. This code reflects the capabilities of InnerLoopVectorizer. /// This class is also used by InnerLoopVectorizer for identifying /// induction variable and the different reduction variables. class LoopVectorizationLegality { public: LoopVectorizationLegality(Loop *L, ScalarEvolution *SE, DataLayout *DL, DominatorTree *DT, TargetLibraryInfo *TLI) : TheLoop(L), SE(SE), DL(DL), DT(DT), TLI(TLI), Induction(0), WidestIndTy(0), HasFunNoNaNAttr(false), MaxSafeDepDistBytes(-1U) {} /// This enum represents the kinds of reductions that we support. enum ReductionKind { RK_NoReduction, ///< Not a reduction. RK_IntegerAdd, ///< Sum of integers. RK_IntegerMult, ///< Product of integers. RK_IntegerOr, ///< Bitwise or logical OR of numbers. RK_IntegerAnd, ///< Bitwise or logical AND of numbers. RK_IntegerXor, ///< Bitwise or logical XOR of numbers. RK_IntegerMinMax, ///< Min/max implemented in terms of select(cmp()). RK_FloatAdd, ///< Sum of floats. RK_FloatMult, ///< Product of floats. RK_FloatMinMax ///< Min/max implemented in terms of select(cmp()). }; /// This enum represents the kinds of inductions that we support. enum InductionKind { IK_NoInduction, ///< Not an induction variable. IK_IntInduction, ///< Integer induction variable. Step = 1. IK_ReverseIntInduction, ///< Reverse int induction variable. Step = -1. IK_PtrInduction, ///< Pointer induction var. Step = sizeof(elem). IK_ReversePtrInduction ///< Reverse ptr indvar. Step = - sizeof(elem). }; // This enum represents the kind of minmax reduction. enum MinMaxReductionKind { MRK_Invalid, MRK_UIntMin, MRK_UIntMax, MRK_SIntMin, MRK_SIntMax, MRK_FloatMin, MRK_FloatMax }; /// This struct holds information about reduction variables. struct ReductionDescriptor { ReductionDescriptor() : StartValue(0), LoopExitInstr(0), Kind(RK_NoReduction), MinMaxKind(MRK_Invalid) {} ReductionDescriptor(Value *Start, Instruction *Exit, ReductionKind K, MinMaxReductionKind MK) : StartValue(Start), LoopExitInstr(Exit), Kind(K), MinMaxKind(MK) {} // The starting value of the reduction. // It does not have to be zero! TrackingVH StartValue; // The instruction who's value is used outside the loop. Instruction *LoopExitInstr; // The kind of the reduction. ReductionKind Kind; // If this a min/max reduction the kind of reduction. MinMaxReductionKind MinMaxKind; }; /// This POD struct holds information about a potential reduction operation. struct ReductionInstDesc { ReductionInstDesc(bool IsRedux, Instruction *I) : IsReduction(IsRedux), PatternLastInst(I), MinMaxKind(MRK_Invalid) {} ReductionInstDesc(Instruction *I, MinMaxReductionKind K) : IsReduction(true), PatternLastInst(I), MinMaxKind(K) {} // Is this instruction a reduction candidate. bool IsReduction; // The last instruction in a min/max pattern (select of the select(icmp()) // pattern), or the current reduction instruction otherwise. Instruction *PatternLastInst; // If this is a min/max pattern the comparison predicate. MinMaxReductionKind MinMaxKind; }; /// This struct holds information about the memory runtime legality /// check that a group of pointers do not overlap. struct RuntimePointerCheck { RuntimePointerCheck() : Need(false) {} /// Reset the state of the pointer runtime information. void reset() { Need = false; Pointers.clear(); Starts.clear(); Ends.clear(); IsWritePtr.clear(); DependencySetId.clear(); } /// Insert a pointer and calculate the start and end SCEVs. void insert(ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr, unsigned DepSetId); /// This flag indicates if we need to add the runtime check. bool Need; /// Holds the pointers that we need to check. SmallVector, 2> Pointers; /// Holds the pointer value at the beginning of the loop. SmallVector Starts; /// Holds the pointer value at the end of the loop. SmallVector Ends; /// Holds the information if this pointer is used for writing to memory. SmallVector IsWritePtr; /// Holds the id of the set of pointers that could be dependent because of a /// shared underlying object. SmallVector DependencySetId; }; /// A struct for saving information about induction variables. struct InductionInfo { InductionInfo(Value *Start, InductionKind K) : StartValue(Start), IK(K) {} InductionInfo() : StartValue(0), IK(IK_NoInduction) {} /// Start value. TrackingVH StartValue; /// Induction kind. InductionKind IK; }; /// ReductionList contains the reduction descriptors for all /// of the reductions that were found in the loop. typedef DenseMap ReductionList; /// InductionList saves induction variables and maps them to the /// induction descriptor. typedef MapVector InductionList; /// Returns true if it is legal to vectorize this loop. /// This does not mean that it is profitable to vectorize this /// loop, only that it is legal to do so. bool canVectorize(); /// Returns the Induction variable. PHINode *getInduction() { return Induction; } /// Returns the reduction variables found in the loop. ReductionList *getReductionVars() { return &Reductions; } /// Returns the induction variables found in the loop. InductionList *getInductionVars() { return &Inductions; } /// Returns the widest induction type. Type *getWidestInductionType() { return WidestIndTy; } /// Returns True if V is an induction variable in this loop. bool isInductionVariable(const Value *V); /// Return true if the block BB needs to be predicated in order for the loop /// to be vectorized. bool blockNeedsPredication(BasicBlock *BB); /// Check if this pointer is consecutive when vectorizing. This happens /// when the last index of the GEP is the induction variable, or that the /// pointer itself is an induction variable. /// This check allows us to vectorize A[idx] into a wide load/store. /// Returns: /// 0 - Stride is unknown or non consecutive. /// 1 - Address is consecutive. /// -1 - Address is consecutive, and decreasing. int isConsecutivePtr(Value *Ptr); /// Returns true if the value V is uniform within the loop. bool isUniform(Value *V); /// Returns true if this instruction will remain scalar after vectorization. bool isUniformAfterVectorization(Instruction* I) { return Uniforms.count(I); } /// Returns the information that we collected about runtime memory check. RuntimePointerCheck *getRuntimePointerCheck() { return &PtrRtCheck; } /// This function returns the identity element (or neutral element) for /// the operation K. static Constant *getReductionIdentity(ReductionKind K, Type *Tp); unsigned getMaxSafeDepDistBytes() { return MaxSafeDepDistBytes; } private: /// Check if a single basic block loop is vectorizable. /// At this point we know that this is a loop with a constant trip count /// and we only need to check individual instructions. bool canVectorizeInstrs(); /// When we vectorize loops we may change the order in which /// we read and write from memory. This method checks if it is /// legal to vectorize the code, considering only memory constrains. /// Returns true if the loop is vectorizable bool canVectorizeMemory(); /// Return true if we can vectorize this loop using the IF-conversion /// transformation. bool canVectorizeWithIfConvert(); /// Collect the variables that need to stay uniform after vectorization. void collectLoopUniforms(); /// Return true if all of the instructions in the block can be speculatively /// executed. \p SafePtrs is a list of addresses that are known to be legal /// and we know that we can read from them without segfault. bool blockCanBePredicated(BasicBlock *BB, SmallPtrSet& SafePtrs); /// Returns True, if 'Phi' is the kind of reduction variable for type /// 'Kind'. If this is a reduction variable, it adds it to ReductionList. bool AddReductionVar(PHINode *Phi, ReductionKind Kind); /// Returns a struct describing if the instruction 'I' can be a reduction /// variable of type 'Kind'. If the reduction is a min/max pattern of /// select(icmp()) this function advances the instruction pointer 'I' from the /// compare instruction to the select instruction and stores this pointer in /// 'PatternLastInst' member of the returned struct. ReductionInstDesc isReductionInstr(Instruction *I, ReductionKind Kind, ReductionInstDesc &Desc); /// Returns true if the instruction is a Select(ICmp(X, Y), X, Y) instruction /// pattern corresponding to a min(X, Y) or max(X, Y). static ReductionInstDesc isMinMaxSelectCmpPattern(Instruction *I, ReductionInstDesc &Prev); /// Returns the induction kind of Phi. This function may return NoInduction /// if the PHI is not an induction variable. InductionKind isInductionVariable(PHINode *Phi); /// The loop that we evaluate. Loop *TheLoop; /// Scev analysis. ScalarEvolution *SE; /// DataLayout analysis. DataLayout *DL; /// Dominators. DominatorTree *DT; /// Target Library Info. TargetLibraryInfo *TLI; // --- vectorization state --- // /// Holds the integer induction variable. This is the counter of the /// loop. PHINode *Induction; /// Holds the reduction variables. ReductionList Reductions; /// Holds all of the induction variables that we found in the loop. /// Notice that inductions don't need to start at zero and that induction /// variables can be pointers. InductionList Inductions; /// Holds the widest induction type encountered. Type *WidestIndTy; /// Allowed outside users. This holds the reduction /// vars which can be accessed from outside the loop. SmallPtrSet AllowedExit; /// This set holds the variables which are known to be uniform after /// vectorization. SmallPtrSet Uniforms; /// We need to check that all of the pointers in this list are disjoint /// at runtime. RuntimePointerCheck PtrRtCheck; /// Can we assume the absence of NaNs. bool HasFunNoNaNAttr; unsigned MaxSafeDepDistBytes; }; /// LoopVectorizationCostModel - estimates the expected speedups due to /// vectorization. /// In many cases vectorization is not profitable. This can happen because of /// a number of reasons. In this class we mainly attempt to predict the /// expected speedup/slowdowns due to the supported instruction set. We use the /// TargetTransformInfo to query the different backends for the cost of /// different operations. class LoopVectorizationCostModel { public: LoopVectorizationCostModel(Loop *L, ScalarEvolution *SE, LoopInfo *LI, LoopVectorizationLegality *Legal, const TargetTransformInfo &TTI, DataLayout *DL, const TargetLibraryInfo *TLI) : TheLoop(L), SE(SE), LI(LI), Legal(Legal), TTI(TTI), DL(DL), TLI(TLI) {} /// Information about vectorization costs struct VectorizationFactor { unsigned Width; // Vector width with best cost unsigned Cost; // Cost of the loop with that width }; /// \return The most profitable vectorization factor and the cost of that VF. /// This method checks every power of two up to VF. If UserVF is not ZERO /// then this vectorization factor will be selected if vectorization is /// possible. VectorizationFactor selectVectorizationFactor(bool OptForSize, unsigned UserVF); /// \return The size (in bits) of the widest type in the code that /// needs to be vectorized. We ignore values that remain scalar such as /// 64 bit loop indices. unsigned getWidestType(); /// \return The most profitable unroll factor. /// If UserUF is non-zero then this method finds the best unroll-factor /// based on register pressure and other parameters. /// VF and LoopCost are the selected vectorization factor and the cost of the /// selected VF. unsigned selectUnrollFactor(bool OptForSize, unsigned UserUF, unsigned VF, unsigned LoopCost); /// \brief A struct that represents some properties of the register usage /// of a loop. struct RegisterUsage { /// Holds the number of loop invariant values that are used in the loop. unsigned LoopInvariantRegs; /// Holds the maximum number of concurrent live intervals in the loop. unsigned MaxLocalUsers; /// Holds the number of instructions in the loop. unsigned NumInstructions; }; /// \return information about the register usage of the loop. RegisterUsage calculateRegisterUsage(); private: /// Returns the expected execution cost. The unit of the cost does /// not matter because we use the 'cost' units to compare different /// vector widths. The cost that is returned is *not* normalized by /// the factor width. unsigned expectedCost(unsigned VF); /// Returns the execution time cost of an instruction for a given vector /// width. Vector width of one means scalar. unsigned getInstructionCost(Instruction *I, unsigned VF); /// A helper function for converting Scalar types to vector types. /// If the incoming type is void, we return void. If the VF is 1, we return /// the scalar type. static Type* ToVectorTy(Type *Scalar, unsigned VF); /// Returns whether the instruction is a load or store and will be a emitted /// as a vector operation. bool isConsecutiveLoadOrStore(Instruction *I); /// The loop that we evaluate. Loop *TheLoop; /// Scev analysis. ScalarEvolution *SE; /// Loop Info analysis. LoopInfo *LI; /// Vectorization legality. LoopVectorizationLegality *Legal; /// Vector target information. const TargetTransformInfo &TTI; /// Target data layout information. DataLayout *DL; /// Target Library Info. const TargetLibraryInfo *TLI; }; /// Utility class for getting and setting loop vectorizer hints in the form /// of loop metadata. struct LoopVectorizeHints { /// Vectorization width. unsigned Width; /// Vectorization unroll factor. unsigned Unroll; LoopVectorizeHints(const Loop *L, bool DisableUnrolling) : Width(VectorizationFactor) , Unroll(DisableUnrolling ? 1 : VectorizationUnroll) , LoopID(L->getLoopID()) { getHints(L); // The command line options override any loop metadata except for when // width == 1 which is used to indicate the loop is already vectorized. if (VectorizationFactor.getNumOccurrences() > 0 && Width != 1) Width = VectorizationFactor; if (VectorizationUnroll.getNumOccurrences() > 0) Unroll = VectorizationUnroll; DEBUG(if (DisableUnrolling && Unroll == 1) dbgs() << "LV: Unrolling disabled by the pass manager\n"); } /// Return the loop vectorizer metadata prefix. static StringRef Prefix() { return "llvm.vectorizer."; } MDNode *createHint(LLVMContext &Context, StringRef Name, unsigned V) { SmallVector Vals; Vals.push_back(MDString::get(Context, Name)); Vals.push_back(ConstantInt::get(Type::getInt32Ty(Context), V)); return MDNode::get(Context, Vals); } /// Mark the loop L as already vectorized by setting the width to 1. void setAlreadyVectorized(Loop *L) { LLVMContext &Context = L->getHeader()->getContext(); Width = 1; // Create a new loop id with one more operand for the already_vectorized // hint. If the loop already has a loop id then copy the existing operands. SmallVector Vals(1); if (LoopID) for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) Vals.push_back(LoopID->getOperand(i)); Vals.push_back(createHint(Context, Twine(Prefix(), "width").str(), Width)); Vals.push_back(createHint(Context, Twine(Prefix(), "unroll").str(), 1)); MDNode *NewLoopID = MDNode::get(Context, Vals); // Set operand 0 to refer to the loop id itself. NewLoopID->replaceOperandWith(0, NewLoopID); L->setLoopID(NewLoopID); if (LoopID) LoopID->replaceAllUsesWith(NewLoopID); LoopID = NewLoopID; } private: MDNode *LoopID; /// Find hints specified in the loop metadata. void getHints(const Loop *L) { if (!LoopID) return; // First operand should refer to the loop id itself. assert(LoopID->getNumOperands() > 0 && "requires at least one operand"); assert(LoopID->getOperand(0) == LoopID && "invalid loop id"); for (unsigned i = 1, ie = LoopID->getNumOperands(); i < ie; ++i) { const MDString *S = 0; SmallVector Args; // The expected hint is either a MDString or a MDNode with the first // operand a MDString. if (const MDNode *MD = dyn_cast(LoopID->getOperand(i))) { if (!MD || MD->getNumOperands() == 0) continue; S = dyn_cast(MD->getOperand(0)); for (unsigned i = 1, ie = MD->getNumOperands(); i < ie; ++i) Args.push_back(MD->getOperand(i)); } else { S = dyn_cast(LoopID->getOperand(i)); assert(Args.size() == 0 && "too many arguments for MDString"); } if (!S) continue; // Check if the hint starts with the vectorizer prefix. StringRef Hint = S->getString(); if (!Hint.startswith(Prefix())) continue; // Remove the prefix. Hint = Hint.substr(Prefix().size(), StringRef::npos); if (Args.size() == 1) getHint(Hint, Args[0]); } } // Check string hint with one operand. void getHint(StringRef Hint, Value *Arg) { const ConstantInt *C = dyn_cast(Arg); if (!C) return; unsigned Val = C->getZExtValue(); if (Hint == "width") { if (isPowerOf2_32(Val) && Val <= MaxVectorWidth) Width = Val; else DEBUG(dbgs() << "LV: ignoring invalid width hint metadata\n"); } else if (Hint == "unroll") { if (isPowerOf2_32(Val) && Val <= MaxUnrollFactor) Unroll = Val; else DEBUG(dbgs() << "LV: ignoring invalid unroll hint metadata\n"); } else { DEBUG(dbgs() << "LV: ignoring unknown hint " << Hint << '\n'); } } }; /// The LoopVectorize Pass. struct LoopVectorize : public LoopPass { /// Pass identification, replacement for typeid static char ID; explicit LoopVectorize(bool NoUnrolling = false) : LoopPass(ID), DisableUnrolling(NoUnrolling) { initializeLoopVectorizePass(*PassRegistry::getPassRegistry()); } ScalarEvolution *SE; DataLayout *DL; LoopInfo *LI; TargetTransformInfo *TTI; DominatorTree *DT; TargetLibraryInfo *TLI; bool DisableUnrolling; virtual bool runOnLoop(Loop *L, LPPassManager &LPM) { // We only vectorize innermost loops. if (!L->empty()) return false; SE = &getAnalysis(); DL = getAnalysisIfAvailable(); LI = &getAnalysis(); TTI = &getAnalysis(); DT = &getAnalysis(); TLI = getAnalysisIfAvailable(); // If the target claims to have no vector registers don't attempt // vectorization. if (!TTI->getNumberOfRegisters(true)) return false; if (DL == NULL) { DEBUG(dbgs() << "LV: Not vectorizing because of missing data layout\n"); return false; } DEBUG(dbgs() << "LV: Checking a loop in \"" << L->getHeader()->getParent()->getName() << "\"\n"); LoopVectorizeHints Hints(L, DisableUnrolling); if (Hints.Width == 1 && Hints.Unroll == 1) { DEBUG(dbgs() << "LV: Not vectorizing.\n"); return false; } // Check if it is legal to vectorize the loop. LoopVectorizationLegality LVL(L, SE, DL, DT, TLI); if (!LVL.canVectorize()) { DEBUG(dbgs() << "LV: Not vectorizing.\n"); return false; } // Use the cost model. LoopVectorizationCostModel CM(L, SE, LI, &LVL, *TTI, DL, TLI); // Check the function attributes to find out if this function should be // optimized for size. Function *F = L->getHeader()->getParent(); Attribute::AttrKind SzAttr = Attribute::OptimizeForSize; Attribute::AttrKind FlAttr = Attribute::NoImplicitFloat; unsigned FnIndex = AttributeSet::FunctionIndex; bool OptForSize = F->getAttributes().hasAttribute(FnIndex, SzAttr); bool NoFloat = F->getAttributes().hasAttribute(FnIndex, FlAttr); if (NoFloat) { DEBUG(dbgs() << "LV: Can't vectorize when the NoImplicitFloat" "attribute is used.\n"); return false; } // Select the optimal vectorization factor. LoopVectorizationCostModel::VectorizationFactor VF; VF = CM.selectVectorizationFactor(OptForSize, Hints.Width); // Select the unroll factor. unsigned UF = CM.selectUnrollFactor(OptForSize, Hints.Unroll, VF.Width, VF.Cost); DEBUG(dbgs() << "LV: Found a vectorizable loop ("<< VF.Width << ") in "<< F->getParent()->getModuleIdentifier() << '\n'); DEBUG(dbgs() << "LV: Unroll Factor is " << UF << '\n'); if (VF.Width == 1) { DEBUG(dbgs() << "LV: Vectorization is possible but not beneficial.\n"); if (UF == 1) return false; // We decided not to vectorize, but we may want to unroll. InnerLoopUnroller Unroller(L, SE, LI, DT, DL, TLI, UF); Unroller.vectorize(&LVL); } else { // If we decided that it is *legal* to vectorize the loop then do it. InnerLoopVectorizer LB(L, SE, LI, DT, DL, TLI, VF.Width, UF); LB.vectorize(&LVL); } // Mark the loop as already vectorized to avoid vectorizing again. Hints.setAlreadyVectorized(L); DEBUG(verifyFunction(*L->getHeader()->getParent())); return true; } virtual void getAnalysisUsage(AnalysisUsage &AU) const { LoopPass::getAnalysisUsage(AU); AU.addRequiredID(LoopSimplifyID); AU.addRequiredID(LCSSAID); AU.addRequired(); AU.addRequired(); AU.addRequired(); AU.addRequired(); AU.addPreserved(); AU.addPreserved(); } }; } // end anonymous namespace //===----------------------------------------------------------------------===// // Implementation of LoopVectorizationLegality, InnerLoopVectorizer and // LoopVectorizationCostModel. //===----------------------------------------------------------------------===// void LoopVectorizationLegality::RuntimePointerCheck::insert(ScalarEvolution *SE, Loop *Lp, Value *Ptr, bool WritePtr, unsigned DepSetId) { const SCEV *Sc = SE->getSCEV(Ptr); const SCEVAddRecExpr *AR = dyn_cast(Sc); assert(AR && "Invalid addrec expression"); const SCEV *Ex = SE->getBackedgeTakenCount(Lp); const SCEV *ScEnd = AR->evaluateAtIteration(Ex, *SE); Pointers.push_back(Ptr); Starts.push_back(AR->getStart()); Ends.push_back(ScEnd); IsWritePtr.push_back(WritePtr); DependencySetId.push_back(DepSetId); } Value *InnerLoopVectorizer::getBroadcastInstrs(Value *V) { // We need to place the broadcast of invariant variables outside the loop. Instruction *Instr = dyn_cast(V); bool NewInstr = (Instr && Instr->getParent() == LoopVectorBody); bool Invariant = OrigLoop->isLoopInvariant(V) && !NewInstr; // Place the code for broadcasting invariant variables in the new preheader. IRBuilder<>::InsertPointGuard Guard(Builder); if (Invariant) Builder.SetInsertPoint(LoopVectorPreHeader->getTerminator()); // Broadcast the scalar into all locations in the vector. Value *Shuf = Builder.CreateVectorSplat(VF, V, "broadcast"); return Shuf; } Value *InnerLoopVectorizer::getConsecutiveVector(Value* Val, int StartIdx, bool Negate) { assert(Val->getType()->isVectorTy() && "Must be a vector"); assert(Val->getType()->getScalarType()->isIntegerTy() && "Elem must be an integer"); // Create the types. Type *ITy = Val->getType()->getScalarType(); VectorType *Ty = cast(Val->getType()); int VLen = Ty->getNumElements(); SmallVector Indices; // Create a vector of consecutive numbers from zero to VF. for (int i = 0; i < VLen; ++i) { int64_t Idx = Negate ? (-i) : i; Indices.push_back(ConstantInt::get(ITy, StartIdx + Idx, Negate)); } // Add the consecutive indices to the vector value. Constant *Cv = ConstantVector::get(Indices); assert(Cv->getType() == Val->getType() && "Invalid consecutive vec"); return Builder.CreateAdd(Val, Cv, "induction"); } /// \brief Find the operand of the GEP that should be checked for consecutive /// stores. This ignores trailing indices that have no effect on the final /// pointer. static unsigned getGEPInductionOperand(DataLayout *DL, const GetElementPtrInst *Gep) { unsigned LastOperand = Gep->getNumOperands() - 1; unsigned GEPAllocSize = DL->getTypeAllocSize( cast(Gep->getType()->getScalarType())->getElementType()); // Walk backwards and try to peel off zeros. while (LastOperand > 1 && match(Gep->getOperand(LastOperand), m_Zero())) { // Find the type we're currently indexing into. gep_type_iterator GEPTI = gep_type_begin(Gep); std::advance(GEPTI, LastOperand - 1); // If it's a type with the same allocation size as the result of the GEP we // can peel off the zero index. if (DL->getTypeAllocSize(*GEPTI) != GEPAllocSize) break; --LastOperand; } return LastOperand; } int LoopVectorizationLegality::isConsecutivePtr(Value *Ptr) { assert(Ptr->getType()->isPointerTy() && "Unexpected non ptr"); // Make sure that the pointer does not point to structs. if (Ptr->getType()->getPointerElementType()->isAggregateType()) return 0; // If this value is a pointer induction variable we know it is consecutive. PHINode *Phi = dyn_cast_or_null(Ptr); if (Phi && Inductions.count(Phi)) { InductionInfo II = Inductions[Phi]; if (IK_PtrInduction == II.IK) return 1; else if (IK_ReversePtrInduction == II.IK) return -1; } GetElementPtrInst *Gep = dyn_cast_or_null(Ptr); if (!Gep) return 0; unsigned NumOperands = Gep->getNumOperands(); Value *GpPtr = Gep->getPointerOperand(); // If this GEP value is a consecutive pointer induction variable and all of // the indices are constant then we know it is consecutive. We can Phi = dyn_cast(GpPtr); if (Phi && Inductions.count(Phi)) { // Make sure that the pointer does not point to structs. PointerType *GepPtrType = cast(GpPtr->getType()); if (GepPtrType->getElementType()->isAggregateType()) return 0; // Make sure that all of the index operands are loop invariant. for (unsigned i = 1; i < NumOperands; ++i) if (!SE->isLoopInvariant(SE->getSCEV(Gep->getOperand(i)), TheLoop)) return 0; InductionInfo II = Inductions[Phi]; if (IK_PtrInduction == II.IK) return 1; else if (IK_ReversePtrInduction == II.IK) return -1; } unsigned InductionOperand = getGEPInductionOperand(DL, Gep); // Check that all of the gep indices are uniform except for our induction // operand. for (unsigned i = 0; i != NumOperands; ++i) if (i != InductionOperand && !SE->isLoopInvariant(SE->getSCEV(Gep->getOperand(i)), TheLoop)) return 0; // We can emit wide load/stores only if the last non-zero index is the // induction variable. const SCEV *Last = SE->getSCEV(Gep->getOperand(InductionOperand)); if (const SCEVAddRecExpr *AR = dyn_cast(Last)) { const SCEV *Step = AR->getStepRecurrence(*SE); // The memory is consecutive because the last index is consecutive // and all other indices are loop invariant. if (Step->isOne()) return 1; if (Step->isAllOnesValue()) return -1; } return 0; } bool LoopVectorizationLegality::isUniform(Value *V) { return (SE->isLoopInvariant(SE->getSCEV(V), TheLoop)); } InnerLoopVectorizer::VectorParts& InnerLoopVectorizer::getVectorValue(Value *V) { assert(V != Induction && "The new induction variable should not be used."); assert(!V->getType()->isVectorTy() && "Can't widen a vector"); // If we have this scalar in the map, return it. if (WidenMap.has(V)) return WidenMap.get(V); // If this scalar is unknown, assume that it is a constant or that it is // loop invariant. Broadcast V and save the value for future uses. Value *B = getBroadcastInstrs(V); return WidenMap.splat(V, B); } Value *InnerLoopVectorizer::reverseVector(Value *Vec) { assert(Vec->getType()->isVectorTy() && "Invalid type"); SmallVector ShuffleMask; for (unsigned i = 0; i < VF; ++i) ShuffleMask.push_back(Builder.getInt32(VF - i - 1)); return Builder.CreateShuffleVector(Vec, UndefValue::get(Vec->getType()), ConstantVector::get(ShuffleMask), "reverse"); } void InnerLoopVectorizer::vectorizeMemoryInstruction(Instruction *Instr, LoopVectorizationLegality *Legal) { // Attempt to issue a wide load. LoadInst *LI = dyn_cast(Instr); StoreInst *SI = dyn_cast(Instr); assert((LI || SI) && "Invalid Load/Store instruction"); Type *ScalarDataTy = LI ? LI->getType() : SI->getValueOperand()->getType(); Type *DataTy = VectorType::get(ScalarDataTy, VF); Value *Ptr = LI ? LI->getPointerOperand() : SI->getPointerOperand(); unsigned Alignment = LI ? LI->getAlignment() : SI->getAlignment(); // An alignment of 0 means target abi alignment. We need to use the scalar's // target abi alignment in such a case. if (!Alignment) Alignment = DL->getABITypeAlignment(ScalarDataTy); unsigned AddressSpace = Ptr->getType()->getPointerAddressSpace(); unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ScalarDataTy); unsigned VectorElementSize = DL->getTypeStoreSize(DataTy)/VF; if (ScalarAllocatedSize != VectorElementSize) return scalarizeInstruction(Instr); // If the pointer is loop invariant or if it is non consecutive, // scalarize the load. int ConsecutiveStride = Legal->isConsecutivePtr(Ptr); bool Reverse = ConsecutiveStride < 0; bool UniformLoad = LI && Legal->isUniform(Ptr); if (!ConsecutiveStride || UniformLoad) return scalarizeInstruction(Instr); Constant *Zero = Builder.getInt32(0); VectorParts &Entry = WidenMap.get(Instr); // Handle consecutive loads/stores. GetElementPtrInst *Gep = dyn_cast(Ptr); if (Gep && Legal->isInductionVariable(Gep->getPointerOperand())) { setDebugLocFromInst(Builder, Gep); Value *PtrOperand = Gep->getPointerOperand(); Value *FirstBasePtr = getVectorValue(PtrOperand)[0]; FirstBasePtr = Builder.CreateExtractElement(FirstBasePtr, Zero); // Create the new GEP with the new induction variable. GetElementPtrInst *Gep2 = cast(Gep->clone()); Gep2->setOperand(0, FirstBasePtr); Gep2->setName("gep.indvar.base"); Ptr = Builder.Insert(Gep2); } else if (Gep) { setDebugLocFromInst(Builder, Gep); assert(SE->isLoopInvariant(SE->getSCEV(Gep->getPointerOperand()), OrigLoop) && "Base ptr must be invariant"); // The last index does not have to be the induction. It can be // consecutive and be a function of the index. For example A[I+1]; unsigned NumOperands = Gep->getNumOperands(); unsigned InductionOperand = getGEPInductionOperand(DL, Gep); // Create the new GEP with the new induction variable. GetElementPtrInst *Gep2 = cast(Gep->clone()); for (unsigned i = 0; i < NumOperands; ++i) { Value *GepOperand = Gep->getOperand(i); Instruction *GepOperandInst = dyn_cast(GepOperand); // Update last index or loop invariant instruction anchored in loop. if (i == InductionOperand || (GepOperandInst && OrigLoop->contains(GepOperandInst))) { assert((i == InductionOperand || SE->isLoopInvariant(SE->getSCEV(GepOperandInst), OrigLoop)) && "Must be last index or loop invariant"); VectorParts &GEPParts = getVectorValue(GepOperand); Value *Index = GEPParts[0]; Index = Builder.CreateExtractElement(Index, Zero); Gep2->setOperand(i, Index); Gep2->setName("gep.indvar.idx"); } } Ptr = Builder.Insert(Gep2); } else { // Use the induction element ptr. assert(isa(Ptr) && "Invalid induction ptr"); setDebugLocFromInst(Builder, Ptr); VectorParts &PtrVal = getVectorValue(Ptr); Ptr = Builder.CreateExtractElement(PtrVal[0], Zero); } // Handle Stores: if (SI) { assert(!Legal->isUniform(SI->getPointerOperand()) && "We do not allow storing to uniform addresses"); setDebugLocFromInst(Builder, SI); // We don't want to update the value in the map as it might be used in // another expression. So don't use a reference type for "StoredVal". VectorParts StoredVal = getVectorValue(SI->getValueOperand()); for (unsigned Part = 0; Part < UF; ++Part) { // Calculate the pointer for the specific unroll-part. Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF)); if (Reverse) { // If we store to reverse consecutive memory locations then we need // to reverse the order of elements in the stored value. StoredVal[Part] = reverseVector(StoredVal[Part]); // If the address is consecutive but reversed, then the // wide store needs to start at the last vector element. PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(-Part * VF)); PartPtr = Builder.CreateGEP(PartPtr, Builder.getInt32(1 - VF)); } Value *VecPtr = Builder.CreateBitCast(PartPtr, DataTy->getPointerTo(AddressSpace)); Builder.CreateStore(StoredVal[Part], VecPtr)->setAlignment(Alignment); } return; } // Handle loads. assert(LI && "Must have a load instruction"); setDebugLocFromInst(Builder, LI); for (unsigned Part = 0; Part < UF; ++Part) { // Calculate the pointer for the specific unroll-part. Value *PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(Part * VF)); if (Reverse) { // If the address is consecutive but reversed, then the // wide store needs to start at the last vector element. PartPtr = Builder.CreateGEP(Ptr, Builder.getInt32(-Part * VF)); PartPtr = Builder.CreateGEP(PartPtr, Builder.getInt32(1 - VF)); } Value *VecPtr = Builder.CreateBitCast(PartPtr, DataTy->getPointerTo(AddressSpace)); Value *LI = Builder.CreateLoad(VecPtr, "wide.load"); cast(LI)->setAlignment(Alignment); Entry[Part] = Reverse ? reverseVector(LI) : LI; } } void InnerLoopVectorizer::scalarizeInstruction(Instruction *Instr) { assert(!Instr->getType()->isAggregateType() && "Can't handle vectors"); // Holds vector parameters or scalars, in case of uniform vals. SmallVector Params; setDebugLocFromInst(Builder, Instr); // Find all of the vectorized parameters. for (unsigned op = 0, e = Instr->getNumOperands(); op != e; ++op) { Value *SrcOp = Instr->getOperand(op); // If we are accessing the old induction variable, use the new one. if (SrcOp == OldInduction) { Params.push_back(getVectorValue(SrcOp)); continue; } // Try using previously calculated values. Instruction *SrcInst = dyn_cast(SrcOp); // If the src is an instruction that appeared earlier in the basic block // then it should already be vectorized. if (SrcInst && OrigLoop->contains(SrcInst)) { assert(WidenMap.has(SrcInst) && "Source operand is unavailable"); // The parameter is a vector value from earlier. Params.push_back(WidenMap.get(SrcInst)); } else { // The parameter is a scalar from outside the loop. Maybe even a constant. VectorParts Scalars; Scalars.append(UF, SrcOp); Params.push_back(Scalars); } } assert(Params.size() == Instr->getNumOperands() && "Invalid number of operands"); // Does this instruction return a value ? bool IsVoidRetTy = Instr->getType()->isVoidTy(); Value *UndefVec = IsVoidRetTy ? 0 : UndefValue::get(VectorType::get(Instr->getType(), VF)); // Create a new entry in the WidenMap and initialize it to Undef or Null. VectorParts &VecResults = WidenMap.splat(Instr, UndefVec); // For each vector unroll 'part': for (unsigned Part = 0; Part < UF; ++Part) { // For each scalar that we create: for (unsigned Width = 0; Width < VF; ++Width) { Instruction *Cloned = Instr->clone(); if (!IsVoidRetTy) Cloned->setName(Instr->getName() + ".cloned"); // Replace the operands of the cloned instructions with extracted scalars. for (unsigned op = 0, e = Instr->getNumOperands(); op != e; ++op) { Value *Op = Params[op][Part]; // Param is a vector. Need to extract the right lane. if (Op->getType()->isVectorTy()) Op = Builder.CreateExtractElement(Op, Builder.getInt32(Width)); Cloned->setOperand(op, Op); } // Place the cloned scalar in the new loop. Builder.Insert(Cloned); // If the original scalar returns a value we need to place it in a vector // so that future users will be able to use it. if (!IsVoidRetTy) VecResults[Part] = Builder.CreateInsertElement(VecResults[Part], Cloned, Builder.getInt32(Width)); } } } Instruction * InnerLoopVectorizer::addRuntimeCheck(LoopVectorizationLegality *Legal, Instruction *Loc) { LoopVectorizationLegality::RuntimePointerCheck *PtrRtCheck = Legal->getRuntimePointerCheck(); if (!PtrRtCheck->Need) return NULL; unsigned NumPointers = PtrRtCheck->Pointers.size(); SmallVector , 2> Starts; SmallVector , 2> Ends; LLVMContext &Ctx = Loc->getContext(); SCEVExpander Exp(*SE, "induction"); for (unsigned i = 0; i < NumPointers; ++i) { Value *Ptr = PtrRtCheck->Pointers[i]; const SCEV *Sc = SE->getSCEV(Ptr); if (SE->isLoopInvariant(Sc, OrigLoop)) { DEBUG(dbgs() << "LV: Adding RT check for a loop invariant ptr:" << *Ptr <<"\n"); Starts.push_back(Ptr); Ends.push_back(Ptr); } else { DEBUG(dbgs() << "LV: Adding RT check for range:" << *Ptr << '\n'); unsigned AS = Ptr->getType()->getPointerAddressSpace(); // Use this type for pointer arithmetic. Type *PtrArithTy = Type::getInt8PtrTy(Ctx, AS); Value *Start = Exp.expandCodeFor(PtrRtCheck->Starts[i], PtrArithTy, Loc); Value *End = Exp.expandCodeFor(PtrRtCheck->Ends[i], PtrArithTy, Loc); Starts.push_back(Start); Ends.push_back(End); } } IRBuilder<> ChkBuilder(Loc); // Our instructions might fold to a constant. Value *MemoryRuntimeCheck = 0; for (unsigned i = 0; i < NumPointers; ++i) { for (unsigned j = i+1; j < NumPointers; ++j) { // No need to check if two readonly pointers intersect. if (!PtrRtCheck->IsWritePtr[i] && !PtrRtCheck->IsWritePtr[j]) continue; // Only need to check pointers between two different dependency sets. if (PtrRtCheck->DependencySetId[i] == PtrRtCheck->DependencySetId[j]) continue; unsigned AS0 = Starts[i]->getType()->getPointerAddressSpace(); unsigned AS1 = Starts[j]->getType()->getPointerAddressSpace(); assert((AS0 == Ends[j]->getType()->getPointerAddressSpace()) && (AS1 == Ends[i]->getType()->getPointerAddressSpace()) && "Trying to bounds check pointers with different address spaces"); Type *PtrArithTy0 = Type::getInt8PtrTy(Ctx, AS0); Type *PtrArithTy1 = Type::getInt8PtrTy(Ctx, AS1); Value *Start0 = ChkBuilder.CreateBitCast(Starts[i], PtrArithTy0, "bc"); Value *Start1 = ChkBuilder.CreateBitCast(Starts[j], PtrArithTy1, "bc"); Value *End0 = ChkBuilder.CreateBitCast(Ends[i], PtrArithTy1, "bc"); Value *End1 = ChkBuilder.CreateBitCast(Ends[j], PtrArithTy0, "bc"); Value *Cmp0 = ChkBuilder.CreateICmpULE(Start0, End1, "bound0"); Value *Cmp1 = ChkBuilder.CreateICmpULE(Start1, End0, "bound1"); Value *IsConflict = ChkBuilder.CreateAnd(Cmp0, Cmp1, "found.conflict"); if (MemoryRuntimeCheck) IsConflict = ChkBuilder.CreateOr(MemoryRuntimeCheck, IsConflict, "conflict.rdx"); MemoryRuntimeCheck = IsConflict; } } // We have to do this trickery because the IRBuilder might fold the check to a // constant expression in which case there is no Instruction anchored in a // the block. Instruction *Check = BinaryOperator::CreateAnd(MemoryRuntimeCheck, ConstantInt::getTrue(Ctx)); ChkBuilder.Insert(Check, "memcheck.conflict"); return Check; } void InnerLoopVectorizer::createEmptyLoop(LoopVectorizationLegality *Legal) { /* In this function we generate a new loop. The new loop will contain the vectorized instructions while the old loop will continue to run the scalar remainder. [ ] <-- vector loop bypass (may consist of multiple blocks). / | / v | [ ] <-- vector pre header. | | | v | [ ] \ | [ ]_| <-- vector loop. | | \ v >[ ] <--- middle-block. / | / v | [ ] <--- new preheader. | | | v | [ ] \ | [ ]_| <-- old scalar loop to handle remainder. \ | \ v >[ ] <-- exit block. ... */ BasicBlock *OldBasicBlock = OrigLoop->getHeader(); BasicBlock *BypassBlock = OrigLoop->getLoopPreheader(); BasicBlock *ExitBlock = OrigLoop->getExitBlock(); assert(ExitBlock && "Must have an exit block"); // Some loops have a single integer induction variable, while other loops // don't. One example is c++ iterators that often have multiple pointer // induction variables. In the code below we also support a case where we // don't have a single induction variable. OldInduction = Legal->getInduction(); Type *IdxTy = Legal->getWidestInductionType(); // Find the loop boundaries. const SCEV *ExitCount = SE->getBackedgeTakenCount(OrigLoop); assert(ExitCount != SE->getCouldNotCompute() && "Invalid loop count"); // The exit count might have the type of i64 while the phi is i32. This can // happen if we have an induction variable that is sign extended before the // compare. The only way that we get a backedge taken count is that the // induction variable was signed and as such will not overflow. In such a case // truncation is legal. if (ExitCount->getType()->getPrimitiveSizeInBits() > IdxTy->getPrimitiveSizeInBits()) ExitCount = SE->getTruncateOrNoop(ExitCount, IdxTy); ExitCount = SE->getNoopOrZeroExtend(ExitCount, IdxTy); // Get the total trip count from the count by adding 1. ExitCount = SE->getAddExpr(ExitCount, SE->getConstant(ExitCount->getType(), 1)); // Expand the trip count and place the new instructions in the preheader. // Notice that the pre-header does not change, only the loop body. SCEVExpander Exp(*SE, "induction"); // Count holds the overall loop count (N). Value *Count = Exp.expandCodeFor(ExitCount, ExitCount->getType(), BypassBlock->getTerminator()); // The loop index does not have to start at Zero. Find the original start // value from the induction PHI node. If we don't have an induction variable // then we know that it starts at zero. Builder.SetInsertPoint(BypassBlock->getTerminator()); Value *StartIdx = ExtendedIdx = OldInduction ? Builder.CreateZExt(OldInduction->getIncomingValueForBlock(BypassBlock), IdxTy): ConstantInt::get(IdxTy, 0); assert(BypassBlock && "Invalid loop structure"); LoopBypassBlocks.push_back(BypassBlock); // Split the single block loop into the two loop structure described above. BasicBlock *VectorPH = BypassBlock->splitBasicBlock(BypassBlock->getTerminator(), "vector.ph"); BasicBlock *VecBody = VectorPH->splitBasicBlock(VectorPH->getTerminator(), "vector.body"); BasicBlock *MiddleBlock = VecBody->splitBasicBlock(VecBody->getTerminator(), "middle.block"); BasicBlock *ScalarPH = MiddleBlock->splitBasicBlock(MiddleBlock->getTerminator(), "scalar.ph"); // Create and register the new vector loop. Loop* Lp = new Loop(); Loop *ParentLoop = OrigLoop->getParentLoop(); // Insert the new loop into the loop nest and register the new basic blocks // before calling any utilities such as SCEV that require valid LoopInfo. if (ParentLoop) { ParentLoop->addChildLoop(Lp); ParentLoop->addBasicBlockToLoop(ScalarPH, LI->getBase()); ParentLoop->addBasicBlockToLoop(VectorPH, LI->getBase()); ParentLoop->addBasicBlockToLoop(MiddleBlock, LI->getBase()); } else { LI->addTopLevelLoop(Lp); } Lp->addBasicBlockToLoop(VecBody, LI->getBase()); // Use this IR builder to create the loop instructions (Phi, Br, Cmp) // inside the loop. Builder.SetInsertPoint(VecBody->getFirstNonPHI()); // Generate the induction variable. setDebugLocFromInst(Builder, getDebugLocFromInstOrOperands(OldInduction)); Induction = Builder.CreatePHI(IdxTy, 2, "index"); // The loop step is equal to the vectorization factor (num of SIMD elements) // times the unroll factor (num of SIMD instructions). Constant *Step = ConstantInt::get(IdxTy, VF * UF); // This is the IR builder that we use to add all of the logic for bypassing // the new vector loop. IRBuilder<> BypassBuilder(BypassBlock->getTerminator()); setDebugLocFromInst(BypassBuilder, getDebugLocFromInstOrOperands(OldInduction)); // We may need to extend the index in case there is a type mismatch. // We know that the count starts at zero and does not overflow. if (Count->getType() != IdxTy) { // The exit count can be of pointer type. Convert it to the correct // integer type. if (ExitCount->getType()->isPointerTy()) Count = BypassBuilder.CreatePointerCast(Count, IdxTy, "ptrcnt.to.int"); else Count = BypassBuilder.CreateZExtOrTrunc(Count, IdxTy, "cnt.cast"); } // Add the start index to the loop count to get the new end index. Value *IdxEnd = BypassBuilder.CreateAdd(Count, StartIdx, "end.idx"); // Now we need to generate the expression for N - (N % VF), which is // the part that the vectorized body will execute. Value *R = BypassBuilder.CreateURem(Count, Step, "n.mod.vf"); Value *CountRoundDown = BypassBuilder.CreateSub(Count, R, "n.vec"); Value *IdxEndRoundDown = BypassBuilder.CreateAdd(CountRoundDown, StartIdx, "end.idx.rnd.down"); // Now, compare the new count to zero. If it is zero skip the vector loop and // jump to the scalar loop. Value *Cmp = BypassBuilder.CreateICmpEQ(IdxEndRoundDown, StartIdx, "cmp.zero"); BasicBlock *LastBypassBlock = BypassBlock; // Generate the code that checks in runtime if arrays overlap. We put the // checks into a separate block to make the more common case of few elements // faster. Instruction *MemRuntimeCheck = addRuntimeCheck(Legal, BypassBlock->getTerminator()); if (MemRuntimeCheck) { // Create a new block containing the memory check. BasicBlock *CheckBlock = BypassBlock->splitBasicBlock(MemRuntimeCheck, "vector.memcheck"); if (ParentLoop) ParentLoop->addBasicBlockToLoop(CheckBlock, LI->getBase()); LoopBypassBlocks.push_back(CheckBlock); // Replace the branch into the memory check block with a conditional branch // for the "few elements case". Instruction *OldTerm = BypassBlock->getTerminator(); BranchInst::Create(MiddleBlock, CheckBlock, Cmp, OldTerm); OldTerm->eraseFromParent(); Cmp = MemRuntimeCheck; LastBypassBlock = CheckBlock; } LastBypassBlock->getTerminator()->eraseFromParent(); BranchInst::Create(MiddleBlock, VectorPH, Cmp, LastBypassBlock); // We are going to resume the execution of the scalar loop. // Go over all of the induction variables that we found and fix the // PHIs that are left in the scalar version of the loop. // The starting values of PHI nodes depend on the counter of the last // iteration in the vectorized loop. // If we come from a bypass edge then we need to start from the original // start value. // This variable saves the new starting index for the scalar loop. PHINode *ResumeIndex = 0; LoopVectorizationLegality::InductionList::iterator I, E; LoopVectorizationLegality::InductionList *List = Legal->getInductionVars(); // Set builder to point to last bypass block. BypassBuilder.SetInsertPoint(LoopBypassBlocks.back()->getTerminator()); for (I = List->begin(), E = List->end(); I != E; ++I) { PHINode *OrigPhi = I->first; LoopVectorizationLegality::InductionInfo II = I->second; Type *ResumeValTy = (OrigPhi == OldInduction) ? IdxTy : OrigPhi->getType(); PHINode *ResumeVal = PHINode::Create(ResumeValTy, 2, "resume.val", MiddleBlock->getTerminator()); // We might have extended the type of the induction variable but we need a // truncated version for the scalar loop. PHINode *TruncResumeVal = (OrigPhi == OldInduction) ? PHINode::Create(OrigPhi->getType(), 2, "trunc.resume.val", MiddleBlock->getTerminator()) : 0; Value *EndValue = 0; switch (II.IK) { case LoopVectorizationLegality::IK_NoInduction: llvm_unreachable("Unknown induction"); case LoopVectorizationLegality::IK_IntInduction: { // Handle the integer induction counter. assert(OrigPhi->getType()->isIntegerTy() && "Invalid type"); // We have the canonical induction variable. if (OrigPhi == OldInduction) { // Create a truncated version of the resume value for the scalar loop, // we might have promoted the type to a larger width. EndValue = BypassBuilder.CreateTrunc(IdxEndRoundDown, OrigPhi->getType()); // The new PHI merges the original incoming value, in case of a bypass, // or the value at the end of the vectorized loop. for (unsigned I = 0, E = LoopBypassBlocks.size(); I != E; ++I) TruncResumeVal->addIncoming(II.StartValue, LoopBypassBlocks[I]); TruncResumeVal->addIncoming(EndValue, VecBody); // We know what the end value is. EndValue = IdxEndRoundDown; // We also know which PHI node holds it. ResumeIndex = ResumeVal; break; } // Not the canonical induction variable - add the vector loop count to the // start value. Value *CRD = BypassBuilder.CreateSExtOrTrunc(CountRoundDown, II.StartValue->getType(), "cast.crd"); EndValue = BypassBuilder.CreateAdd(CRD, II.StartValue , "ind.end"); break; } case LoopVectorizationLegality::IK_ReverseIntInduction: { // Convert the CountRoundDown variable to the PHI size. Value *CRD = BypassBuilder.CreateSExtOrTrunc(CountRoundDown, II.StartValue->getType(), "cast.crd"); // Handle reverse integer induction counter. EndValue = BypassBuilder.CreateSub(II.StartValue, CRD, "rev.ind.end"); break; } case LoopVectorizationLegality::IK_PtrInduction: { // For pointer induction variables, calculate the offset using // the end index. EndValue = BypassBuilder.CreateGEP(II.StartValue, CountRoundDown, "ptr.ind.end"); break; } case LoopVectorizationLegality::IK_ReversePtrInduction: { // The value at the end of the loop for the reverse pointer is calculated // by creating a GEP with a negative index starting from the start value. Value *Zero = ConstantInt::get(CountRoundDown->getType(), 0); Value *NegIdx = BypassBuilder.CreateSub(Zero, CountRoundDown, "rev.ind.end"); EndValue = BypassBuilder.CreateGEP(II.StartValue, NegIdx, "rev.ptr.ind.end"); break; } }// end of case // The new PHI merges the original incoming value, in case of a bypass, // or the value at the end of the vectorized loop. for (unsigned I = 0, E = LoopBypassBlocks.size(); I != E; ++I) { if (OrigPhi == OldInduction) ResumeVal->addIncoming(StartIdx, LoopBypassBlocks[I]); else ResumeVal->addIncoming(II.StartValue, LoopBypassBlocks[I]); } ResumeVal->addIncoming(EndValue, VecBody); // Fix the scalar body counter (PHI node). unsigned BlockIdx = OrigPhi->getBasicBlockIndex(ScalarPH); // The old inductions phi node in the scalar body needs the truncated value. if (OrigPhi == OldInduction) OrigPhi->setIncomingValue(BlockIdx, TruncResumeVal); else OrigPhi->setIncomingValue(BlockIdx, ResumeVal); } // If we are generating a new induction variable then we also need to // generate the code that calculates the exit value. This value is not // simply the end of the counter because we may skip the vectorized body // in case of a runtime check. if (!OldInduction){ assert(!ResumeIndex && "Unexpected resume value found"); ResumeIndex = PHINode::Create(IdxTy, 2, "new.indc.resume.val", MiddleBlock->getTerminator()); for (unsigned I = 0, E = LoopBypassBlocks.size(); I != E; ++I) ResumeIndex->addIncoming(StartIdx, LoopBypassBlocks[I]); ResumeIndex->addIncoming(IdxEndRoundDown, VecBody); } // Make sure that we found the index where scalar loop needs to continue. assert(ResumeIndex && ResumeIndex->getType()->isIntegerTy() && "Invalid resume Index"); // Add a check in the middle block to see if we have completed // all of the iterations in the first vector loop. // If (N - N%VF) == N, then we *don't* need to run the remainder. Value *CmpN = CmpInst::Create(Instruction::ICmp, CmpInst::ICMP_EQ, IdxEnd, ResumeIndex, "cmp.n", MiddleBlock->getTerminator()); BranchInst::Create(ExitBlock, ScalarPH, CmpN, MiddleBlock->getTerminator()); // Remove the old terminator. MiddleBlock->getTerminator()->eraseFromParent(); // Create i+1 and fill the PHINode. Value *NextIdx = Builder.CreateAdd(Induction, Step, "index.next"); Induction->addIncoming(StartIdx, VectorPH); Induction->addIncoming(NextIdx, VecBody); // Create the compare. Value *ICmp = Builder.CreateICmpEQ(NextIdx, IdxEndRoundDown); Builder.CreateCondBr(ICmp, MiddleBlock, VecBody); // Now we have two terminators. Remove the old one from the block. VecBody->getTerminator()->eraseFromParent(); // Get ready to start creating new instructions into the vectorized body. Builder.SetInsertPoint(VecBody->getFirstInsertionPt()); // Save the state. LoopVectorPreHeader = VectorPH; LoopScalarPreHeader = ScalarPH; LoopMiddleBlock = MiddleBlock; LoopExitBlock = ExitBlock; LoopVectorBody = VecBody; LoopScalarBody = OldBasicBlock; LoopVectorizeHints Hints(Lp, true); Hints.setAlreadyVectorized(Lp); } /// This function returns the identity element (or neutral element) for /// the operation K. Constant* LoopVectorizationLegality::getReductionIdentity(ReductionKind K, Type *Tp) { switch (K) { case RK_IntegerXor: case RK_IntegerAdd: case RK_IntegerOr: // Adding, Xoring, Oring zero to a number does not change it. return ConstantInt::get(Tp, 0); case RK_IntegerMult: // Multiplying a number by 1 does not change it. return ConstantInt::get(Tp, 1); case RK_IntegerAnd: // AND-ing a number with an all-1 value does not change it. return ConstantInt::get(Tp, -1, true); case RK_FloatMult: // Multiplying a number by 1 does not change it. return ConstantFP::get(Tp, 1.0L); case RK_FloatAdd: // Adding zero to a number does not change it. return ConstantFP::get(Tp, 0.0L); default: llvm_unreachable("Unknown reduction kind"); } } static Intrinsic::ID checkUnaryFloatSignature(const CallInst &I, Intrinsic::ID ValidIntrinsicID) { if (I.getNumArgOperands() != 1 || !I.getArgOperand(0)->getType()->isFloatingPointTy() || I.getType() != I.getArgOperand(0)->getType() || !I.onlyReadsMemory()) return Intrinsic::not_intrinsic; return ValidIntrinsicID; } static Intrinsic::ID checkBinaryFloatSignature(const CallInst &I, Intrinsic::ID ValidIntrinsicID) { if (I.getNumArgOperands() != 2 || !I.getArgOperand(0)->getType()->isFloatingPointTy() || !I.getArgOperand(1)->getType()->isFloatingPointTy() || I.getType() != I.getArgOperand(0)->getType() || I.getType() != I.getArgOperand(1)->getType() || !I.onlyReadsMemory()) return Intrinsic::not_intrinsic; return ValidIntrinsicID; } static Intrinsic::ID getIntrinsicIDForCall(CallInst *CI, const TargetLibraryInfo *TLI) { // If we have an intrinsic call, check if it is trivially vectorizable. if (IntrinsicInst *II = dyn_cast(CI)) { switch (II->getIntrinsicID()) { case Intrinsic::sqrt: case Intrinsic::sin: case Intrinsic::cos: case Intrinsic::exp: case Intrinsic::exp2: case Intrinsic::log: case Intrinsic::log10: case Intrinsic::log2: case Intrinsic::fabs: case Intrinsic::copysign: case Intrinsic::floor: case Intrinsic::ceil: case Intrinsic::trunc: case Intrinsic::rint: case Intrinsic::nearbyint: case Intrinsic::round: case Intrinsic::pow: case Intrinsic::fma: case Intrinsic::fmuladd: case Intrinsic::lifetime_start: case Intrinsic::lifetime_end: return II->getIntrinsicID(); default: return Intrinsic::not_intrinsic; } } if (!TLI) return Intrinsic::not_intrinsic; LibFunc::Func Func; Function *F = CI->getCalledFunction(); // We're going to make assumptions on the semantics of the functions, check // that the target knows that it's available in this environment and it does // not have local linkage. if (!F || F->hasLocalLinkage() || !TLI->getLibFunc(F->getName(), Func)) return Intrinsic::not_intrinsic; // Otherwise check if we have a call to a function that can be turned into a // vector intrinsic. switch (Func) { default: break; case LibFunc::sin: case LibFunc::sinf: case LibFunc::sinl: return checkUnaryFloatSignature(*CI, Intrinsic::sin); case LibFunc::cos: case LibFunc::cosf: case LibFunc::cosl: return checkUnaryFloatSignature(*CI, Intrinsic::cos); case LibFunc::exp: case LibFunc::expf: case LibFunc::expl: return checkUnaryFloatSignature(*CI, Intrinsic::exp); case LibFunc::exp2: case LibFunc::exp2f: case LibFunc::exp2l: return checkUnaryFloatSignature(*CI, Intrinsic::exp2); case LibFunc::log: case LibFunc::logf: case LibFunc::logl: return checkUnaryFloatSignature(*CI, Intrinsic::log); case LibFunc::log10: case LibFunc::log10f: case LibFunc::log10l: return checkUnaryFloatSignature(*CI, Intrinsic::log10); case LibFunc::log2: case LibFunc::log2f: case LibFunc::log2l: return checkUnaryFloatSignature(*CI, Intrinsic::log2); case LibFunc::fabs: case LibFunc::fabsf: case LibFunc::fabsl: return checkUnaryFloatSignature(*CI, Intrinsic::fabs); case LibFunc::copysign: case LibFunc::copysignf: case LibFunc::copysignl: return checkBinaryFloatSignature(*CI, Intrinsic::copysign); case LibFunc::floor: case LibFunc::floorf: case LibFunc::floorl: return checkUnaryFloatSignature(*CI, Intrinsic::floor); case LibFunc::ceil: case LibFunc::ceilf: case LibFunc::ceill: return checkUnaryFloatSignature(*CI, Intrinsic::ceil); case LibFunc::trunc: case LibFunc::truncf: case LibFunc::truncl: return checkUnaryFloatSignature(*CI, Intrinsic::trunc); case LibFunc::rint: case LibFunc::rintf: case LibFunc::rintl: return checkUnaryFloatSignature(*CI, Intrinsic::rint); case LibFunc::nearbyint: case LibFunc::nearbyintf: case LibFunc::nearbyintl: return checkUnaryFloatSignature(*CI, Intrinsic::nearbyint); case LibFunc::round: case LibFunc::roundf: case LibFunc::roundl: return checkUnaryFloatSignature(*CI, Intrinsic::round); case LibFunc::pow: case LibFunc::powf: case LibFunc::powl: return checkBinaryFloatSignature(*CI, Intrinsic::pow); } return Intrinsic::not_intrinsic; } /// This function translates the reduction kind to an LLVM binary operator. static unsigned getReductionBinOp(LoopVectorizationLegality::ReductionKind Kind) { switch (Kind) { case LoopVectorizationLegality::RK_IntegerAdd: return Instruction::Add; case LoopVectorizationLegality::RK_IntegerMult: return Instruction::Mul; case LoopVectorizationLegality::RK_IntegerOr: return Instruction::Or; case LoopVectorizationLegality::RK_IntegerAnd: return Instruction::And; case LoopVectorizationLegality::RK_IntegerXor: return Instruction::Xor; case LoopVectorizationLegality::RK_FloatMult: return Instruction::FMul; case LoopVectorizationLegality::RK_FloatAdd: return Instruction::FAdd; case LoopVectorizationLegality::RK_IntegerMinMax: return Instruction::ICmp; case LoopVectorizationLegality::RK_FloatMinMax: return Instruction::FCmp; default: llvm_unreachable("Unknown reduction operation"); } } Value *createMinMaxOp(IRBuilder<> &Builder, LoopVectorizationLegality::MinMaxReductionKind RK, Value *Left, Value *Right) { CmpInst::Predicate P = CmpInst::ICMP_NE; switch (RK) { default: llvm_unreachable("Unknown min/max reduction kind"); case LoopVectorizationLegality::MRK_UIntMin: P = CmpInst::ICMP_ULT; break; case LoopVectorizationLegality::MRK_UIntMax: P = CmpInst::ICMP_UGT; break; case LoopVectorizationLegality::MRK_SIntMin: P = CmpInst::ICMP_SLT; break; case LoopVectorizationLegality::MRK_SIntMax: P = CmpInst::ICMP_SGT; break; case LoopVectorizationLegality::MRK_FloatMin: P = CmpInst::FCMP_OLT; break; case LoopVectorizationLegality::MRK_FloatMax: P = CmpInst::FCMP_OGT; break; } Value *Cmp; if (RK == LoopVectorizationLegality::MRK_FloatMin || RK == LoopVectorizationLegality::MRK_FloatMax) Cmp = Builder.CreateFCmp(P, Left, Right, "rdx.minmax.cmp"); else Cmp = Builder.CreateICmp(P, Left, Right, "rdx.minmax.cmp"); Value *Select = Builder.CreateSelect(Cmp, Left, Right, "rdx.minmax.select"); return Select; } namespace { struct CSEDenseMapInfo { static bool canHandle(Instruction *I) { return isa(I) || isa(I) || isa(I) || isa(I); } static inline Instruction *getEmptyKey() { return DenseMapInfo::getEmptyKey(); } static inline Instruction *getTombstoneKey() { return DenseMapInfo::getTombstoneKey(); } static unsigned getHashValue(Instruction *I) { assert(canHandle(I) && "Unknown instruction!"); return hash_combine(I->getOpcode(), hash_combine_range(I->value_op_begin(), I->value_op_end())); } static bool isEqual(Instruction *LHS, Instruction *RHS) { if (LHS == getEmptyKey() || RHS == getEmptyKey() || LHS == getTombstoneKey() || RHS == getTombstoneKey()) return LHS == RHS; return LHS->isIdenticalTo(RHS); } }; } ///\brief Perform cse of induction variable instructions. static void cse(BasicBlock *BB) { // Perform simple cse. SmallDenseMap CSEMap; for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E;) { Instruction *In = I++; if (!CSEDenseMapInfo::canHandle(In)) continue; // Check if we can replace this instruction with any of the // visited instructions. if (Instruction *V = CSEMap.lookup(In)) { In->replaceAllUsesWith(V); In->eraseFromParent(); continue; } CSEMap[In] = In; } } void InnerLoopVectorizer::vectorizeLoop(LoopVectorizationLegality *Legal) { //===------------------------------------------------===// // // Notice: any optimization or new instruction that go // into the code below should be also be implemented in // the cost-model. // //===------------------------------------------------===// Constant *Zero = Builder.getInt32(0); // In order to support reduction variables we need to be able to vectorize // Phi nodes. Phi nodes have cycles, so we need to vectorize them in two // stages. First, we create a new vector PHI node with no incoming edges. // We use this value when we vectorize all of the instructions that use the // PHI. Next, after all of the instructions in the block are complete we // add the new incoming edges to the PHI. At this point all of the // instructions in the basic block are vectorized, so we can use them to // construct the PHI. PhiVector RdxPHIsToFix; // Scan the loop in a topological order to ensure that defs are vectorized // before users. LoopBlocksDFS DFS(OrigLoop); DFS.perform(LI); // Vectorize all of the blocks in the original loop. for (LoopBlocksDFS::RPOIterator bb = DFS.beginRPO(), be = DFS.endRPO(); bb != be; ++bb) vectorizeBlockInLoop(Legal, *bb, &RdxPHIsToFix); // At this point every instruction in the original loop is widened to // a vector form. We are almost done. Now, we need to fix the PHI nodes // that we vectorized. The PHI nodes are currently empty because we did // not want to introduce cycles. Notice that the remaining PHI nodes // that we need to fix are reduction variables. // Create the 'reduced' values for each of the induction vars. // The reduced values are the vector values that we scalarize and combine // after the loop is finished. for (PhiVector::iterator it = RdxPHIsToFix.begin(), e = RdxPHIsToFix.end(); it != e; ++it) { PHINode *RdxPhi = *it; assert(RdxPhi && "Unable to recover vectorized PHI"); // Find the reduction variable descriptor. assert(Legal->getReductionVars()->count(RdxPhi) && "Unable to find the reduction variable"); LoopVectorizationLegality::ReductionDescriptor RdxDesc = (*Legal->getReductionVars())[RdxPhi]; setDebugLocFromInst(Builder, RdxDesc.StartValue); // We need to generate a reduction vector from the incoming scalar. // To do so, we need to generate the 'identity' vector and overide // one of the elements with the incoming scalar reduction. We need // to do it in the vector-loop preheader. Builder.SetInsertPoint(LoopBypassBlocks.front()->getTerminator()); // This is the vector-clone of the value that leaves the loop. VectorParts &VectorExit = getVectorValue(RdxDesc.LoopExitInstr); Type *VecTy = VectorExit[0]->getType(); // Find the reduction identity variable. Zero for addition, or, xor, // one for multiplication, -1 for And. Value *Identity; Value *VectorStart; if (RdxDesc.Kind == LoopVectorizationLegality::RK_IntegerMinMax || RdxDesc.Kind == LoopVectorizationLegality::RK_FloatMinMax) { // MinMax reduction have the start value as their identify. if (VF == 1) { VectorStart = Identity = RdxDesc.StartValue; } else { VectorStart = Identity = Builder.CreateVectorSplat(VF, RdxDesc.StartValue, "minmax.ident"); } } else { // Handle other reduction kinds: Constant *Iden = LoopVectorizationLegality::getReductionIdentity(RdxDesc.Kind, VecTy->getScalarType()); if (VF == 1) { Identity = Iden; // This vector is the Identity vector where the first element is the // incoming scalar reduction. VectorStart = RdxDesc.StartValue; } else { Identity = ConstantVector::getSplat(VF, Iden); // This vector is the Identity vector where the first element is the // incoming scalar reduction. VectorStart = Builder.CreateInsertElement(Identity, RdxDesc.StartValue, Zero); } } // Fix the vector-loop phi. // We created the induction variable so we know that the // preheader is the first entry. BasicBlock *VecPreheader = Induction->getIncomingBlock(0); // Reductions do not have to start at zero. They can start with // any loop invariant values. VectorParts &VecRdxPhi = WidenMap.get(RdxPhi); BasicBlock *Latch = OrigLoop->getLoopLatch(); Value *LoopVal = RdxPhi->getIncomingValueForBlock(Latch); VectorParts &Val = getVectorValue(LoopVal); for (unsigned part = 0; part < UF; ++part) { // Make sure to add the reduction stat value only to the // first unroll part. Value *StartVal = (part == 0) ? VectorStart : Identity; cast(VecRdxPhi[part])->addIncoming(StartVal, VecPreheader); cast(VecRdxPhi[part])->addIncoming(Val[part], LoopVectorBody); } // Before each round, move the insertion point right between // the PHIs and the values we are going to write. // This allows us to write both PHINodes and the extractelement // instructions. Builder.SetInsertPoint(LoopMiddleBlock->getFirstInsertionPt()); VectorParts RdxParts; setDebugLocFromInst(Builder, RdxDesc.LoopExitInstr); for (unsigned part = 0; part < UF; ++part) { // This PHINode contains the vectorized reduction variable, or // the initial value vector, if we bypass the vector loop. VectorParts &RdxExitVal = getVectorValue(RdxDesc.LoopExitInstr); PHINode *NewPhi = Builder.CreatePHI(VecTy, 2, "rdx.vec.exit.phi"); Value *StartVal = (part == 0) ? VectorStart : Identity; for (unsigned I = 0, E = LoopBypassBlocks.size(); I != E; ++I) NewPhi->addIncoming(StartVal, LoopBypassBlocks[I]); NewPhi->addIncoming(RdxExitVal[part], LoopVectorBody); RdxParts.push_back(NewPhi); } // Reduce all of the unrolled parts into a single vector. Value *ReducedPartRdx = RdxParts[0]; unsigned Op = getReductionBinOp(RdxDesc.Kind); setDebugLocFromInst(Builder, ReducedPartRdx); for (unsigned part = 1; part < UF; ++part) { if (Op != Instruction::ICmp && Op != Instruction::FCmp) ReducedPartRdx = Builder.CreateBinOp((Instruction::BinaryOps)Op, RdxParts[part], ReducedPartRdx, "bin.rdx"); else ReducedPartRdx = createMinMaxOp(Builder, RdxDesc.MinMaxKind, ReducedPartRdx, RdxParts[part]); } if (VF > 1) { // VF is a power of 2 so we can emit the reduction using log2(VF) shuffles // and vector ops, reducing the set of values being computed by half each // round. assert(isPowerOf2_32(VF) && "Reduction emission only supported for pow2 vectors!"); Value *TmpVec = ReducedPartRdx; SmallVector ShuffleMask(VF, 0); for (unsigned i = VF; i != 1; i >>= 1) { // Move the upper half of the vector to the lower half. for (unsigned j = 0; j != i/2; ++j) ShuffleMask[j] = Builder.getInt32(i/2 + j); // Fill the rest of the mask with undef. std::fill(&ShuffleMask[i/2], ShuffleMask.end(), UndefValue::get(Builder.getInt32Ty())); Value *Shuf = Builder.CreateShuffleVector(TmpVec, UndefValue::get(TmpVec->getType()), ConstantVector::get(ShuffleMask), "rdx.shuf"); if (Op != Instruction::ICmp && Op != Instruction::FCmp) TmpVec = Builder.CreateBinOp((Instruction::BinaryOps)Op, TmpVec, Shuf, "bin.rdx"); else TmpVec = createMinMaxOp(Builder, RdxDesc.MinMaxKind, TmpVec, Shuf); } // The result is in the first element of the vector. ReducedPartRdx = Builder.CreateExtractElement(TmpVec, Builder.getInt32(0)); } // Now, we need to fix the users of the reduction variable // inside and outside of the scalar remainder loop. // We know that the loop is in LCSSA form. We need to update the // PHI nodes in the exit blocks. for (BasicBlock::iterator LEI = LoopExitBlock->begin(), LEE = LoopExitBlock->end(); LEI != LEE; ++LEI) { PHINode *LCSSAPhi = dyn_cast(LEI); if (!LCSSAPhi) break; // All PHINodes need to have a single entry edge, or two if // we already fixed them. assert(LCSSAPhi->getNumIncomingValues() < 3 && "Invalid LCSSA PHI"); // We found our reduction value exit-PHI. Update it with the // incoming bypass edge. if (LCSSAPhi->getIncomingValue(0) == RdxDesc.LoopExitInstr) { // Add an edge coming from the bypass. LCSSAPhi->addIncoming(ReducedPartRdx, LoopMiddleBlock); break; } }// end of the LCSSA phi scan. // Fix the scalar loop reduction variable with the incoming reduction sum // from the vector body and from the backedge value. int IncomingEdgeBlockIdx = (RdxPhi)->getBasicBlockIndex(OrigLoop->getLoopLatch()); assert(IncomingEdgeBlockIdx >= 0 && "Invalid block index"); // Pick the other block. int SelfEdgeBlockIdx = (IncomingEdgeBlockIdx ? 0 : 1); (RdxPhi)->setIncomingValue(SelfEdgeBlockIdx, ReducedPartRdx); (RdxPhi)->setIncomingValue(IncomingEdgeBlockIdx, RdxDesc.LoopExitInstr); }// end of for each redux variable. fixLCSSAPHIs(); // Remove redundant induction instructions. cse(LoopVectorBody); } void InnerLoopVectorizer::fixLCSSAPHIs() { for (BasicBlock::iterator LEI = LoopExitBlock->begin(), LEE = LoopExitBlock->end(); LEI != LEE; ++LEI) { PHINode *LCSSAPhi = dyn_cast(LEI); if (!LCSSAPhi) break; if (LCSSAPhi->getNumIncomingValues() == 1) LCSSAPhi->addIncoming(UndefValue::get(LCSSAPhi->getType()), LoopMiddleBlock); } } InnerLoopVectorizer::VectorParts InnerLoopVectorizer::createEdgeMask(BasicBlock *Src, BasicBlock *Dst) { assert(std::find(pred_begin(Dst), pred_end(Dst), Src) != pred_end(Dst) && "Invalid edge"); // Look for cached value. std::pair Edge(Src, Dst); EdgeMaskCache::iterator ECEntryIt = MaskCache.find(Edge); if (ECEntryIt != MaskCache.end()) return ECEntryIt->second; VectorParts SrcMask = createBlockInMask(Src); // The terminator has to be a branch inst! BranchInst *BI = dyn_cast(Src->getTerminator()); assert(BI && "Unexpected terminator found"); if (BI->isConditional()) { VectorParts EdgeMask = getVectorValue(BI->getCondition()); if (BI->getSuccessor(0) != Dst) for (unsigned part = 0; part < UF; ++part) EdgeMask[part] = Builder.CreateNot(EdgeMask[part]); for (unsigned part = 0; part < UF; ++part) EdgeMask[part] = Builder.CreateAnd(EdgeMask[part], SrcMask[part]); MaskCache[Edge] = EdgeMask; return EdgeMask; } MaskCache[Edge] = SrcMask; return SrcMask; } InnerLoopVectorizer::VectorParts InnerLoopVectorizer::createBlockInMask(BasicBlock *BB) { assert(OrigLoop->contains(BB) && "Block is not a part of a loop"); // Loop incoming mask is all-one. if (OrigLoop->getHeader() == BB) { Value *C = ConstantInt::get(IntegerType::getInt1Ty(BB->getContext()), 1); return getVectorValue(C); } // This is the block mask. We OR all incoming edges, and with zero. Value *Zero = ConstantInt::get(IntegerType::getInt1Ty(BB->getContext()), 0); VectorParts BlockMask = getVectorValue(Zero); // For each pred: for (pred_iterator it = pred_begin(BB), e = pred_end(BB); it != e; ++it) { VectorParts EM = createEdgeMask(*it, BB); for (unsigned part = 0; part < UF; ++part) BlockMask[part] = Builder.CreateOr(BlockMask[part], EM[part]); } return BlockMask; } void InnerLoopVectorizer::widenPHIInstruction(Instruction *PN, InnerLoopVectorizer::VectorParts &Entry, LoopVectorizationLegality *Legal, unsigned UF, unsigned VF, PhiVector *PV) { PHINode* P = cast(PN); // Handle reduction variables: if (Legal->getReductionVars()->count(P)) { for (unsigned part = 0; part < UF; ++part) { // This is phase one of vectorizing PHIs. Type *VecTy = (VF == 1) ? PN->getType() : VectorType::get(PN->getType(), VF); Entry[part] = PHINode::Create(VecTy, 2, "vec.phi", LoopVectorBody-> getFirstInsertionPt()); } PV->push_back(P); return; } setDebugLocFromInst(Builder, P); // Check for PHI nodes that are lowered to vector selects. if (P->getParent() != OrigLoop->getHeader()) { // We know that all PHIs in non header blocks are converted into // selects, so we don't have to worry about the insertion order and we // can just use the builder. // At this point we generate the predication tree. There may be // duplications since this is a simple recursive scan, but future // optimizations will clean it up. unsigned NumIncoming = P->getNumIncomingValues(); // Generate a sequence of selects of the form: // SELECT(Mask3, In3, // SELECT(Mask2, In2, // ( ...))) for (unsigned In = 0; In < NumIncoming; In++) { VectorParts Cond = createEdgeMask(P->getIncomingBlock(In), P->getParent()); VectorParts &In0 = getVectorValue(P->getIncomingValue(In)); for (unsigned part = 0; part < UF; ++part) { // We might have single edge PHIs (blocks) - use an identity // 'select' for the first PHI operand. if (In == 0) Entry[part] = Builder.CreateSelect(Cond[part], In0[part], In0[part]); else // Select between the current value and the previous incoming edge // based on the incoming mask. Entry[part] = Builder.CreateSelect(Cond[part], In0[part], Entry[part], "predphi"); } } return; } // This PHINode must be an induction variable. // Make sure that we know about it. assert(Legal->getInductionVars()->count(P) && "Not an induction variable"); LoopVectorizationLegality::InductionInfo II = Legal->getInductionVars()->lookup(P); switch (II.IK) { case LoopVectorizationLegality::IK_NoInduction: llvm_unreachable("Unknown induction"); case LoopVectorizationLegality::IK_IntInduction: { assert(P->getType() == II.StartValue->getType() && "Types must match"); Type *PhiTy = P->getType(); Value *Broadcasted; if (P == OldInduction) { // Handle the canonical induction variable. We might have had to // extend the type. Broadcasted = Builder.CreateTrunc(Induction, PhiTy); } else { // Handle other induction variables that are now based on the // canonical one. Value *NormalizedIdx = Builder.CreateSub(Induction, ExtendedIdx, "normalized.idx"); NormalizedIdx = Builder.CreateSExtOrTrunc(NormalizedIdx, PhiTy); Broadcasted = Builder.CreateAdd(II.StartValue, NormalizedIdx, "offset.idx"); } Broadcasted = getBroadcastInstrs(Broadcasted); // After broadcasting the induction variable we need to make the vector // consecutive by adding 0, 1, 2, etc. for (unsigned part = 0; part < UF; ++part) Entry[part] = getConsecutiveVector(Broadcasted, VF * part, false); return; } case LoopVectorizationLegality::IK_ReverseIntInduction: case LoopVectorizationLegality::IK_PtrInduction: case LoopVectorizationLegality::IK_ReversePtrInduction: // Handle reverse integer and pointer inductions. Value *StartIdx = ExtendedIdx; // This is the normalized GEP that starts counting at zero. Value *NormalizedIdx = Builder.CreateSub(Induction, StartIdx, "normalized.idx"); // Handle the reverse integer induction variable case. if (LoopVectorizationLegality::IK_ReverseIntInduction == II.IK) { IntegerType *DstTy = cast(II.StartValue->getType()); Value *CNI = Builder.CreateSExtOrTrunc(NormalizedIdx, DstTy, "resize.norm.idx"); Value *ReverseInd = Builder.CreateSub(II.StartValue, CNI, "reverse.idx"); // This is a new value so do not hoist it out. Value *Broadcasted = getBroadcastInstrs(ReverseInd); // After broadcasting the induction variable we need to make the // vector consecutive by adding ... -3, -2, -1, 0. for (unsigned part = 0; part < UF; ++part) Entry[part] = getConsecutiveVector(Broadcasted, -(int)VF * part, true); return; } // Handle the pointer induction variable case. assert(P->getType()->isPointerTy() && "Unexpected type."); // Is this a reverse induction ptr or a consecutive induction ptr. bool Reverse = (LoopVectorizationLegality::IK_ReversePtrInduction == II.IK); // This is the vector of results. Notice that we don't generate // vector geps because scalar geps result in better code. for (unsigned part = 0; part < UF; ++part) { if (VF == 1) { int EltIndex = (part) * (Reverse ? -1 : 1); Constant *Idx = ConstantInt::get(Induction->getType(), EltIndex); Value *GlobalIdx; if (Reverse) GlobalIdx = Builder.CreateSub(Idx, NormalizedIdx, "gep.ridx"); else GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx, "gep.idx"); Value *SclrGep = Builder.CreateGEP(II.StartValue, GlobalIdx, "next.gep"); Entry[part] = SclrGep; continue; } Value *VecVal = UndefValue::get(VectorType::get(P->getType(), VF)); for (unsigned int i = 0; i < VF; ++i) { int EltIndex = (i + part * VF) * (Reverse ? -1 : 1); Constant *Idx = ConstantInt::get(Induction->getType(), EltIndex); Value *GlobalIdx; if (!Reverse) GlobalIdx = Builder.CreateAdd(NormalizedIdx, Idx, "gep.idx"); else GlobalIdx = Builder.CreateSub(Idx, NormalizedIdx, "gep.ridx"); Value *SclrGep = Builder.CreateGEP(II.StartValue, GlobalIdx, "next.gep"); VecVal = Builder.CreateInsertElement(VecVal, SclrGep, Builder.getInt32(i), "insert.gep"); } Entry[part] = VecVal; } return; } } void InnerLoopVectorizer::vectorizeBlockInLoop(LoopVectorizationLegality *Legal, BasicBlock *BB, PhiVector *PV) { // For each instruction in the old loop. for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { VectorParts &Entry = WidenMap.get(it); switch (it->getOpcode()) { case Instruction::Br: // Nothing to do for PHIs and BR, since we already took care of the // loop control flow instructions. continue; case Instruction::PHI:{ // Vectorize PHINodes. widenPHIInstruction(it, Entry, Legal, UF, VF, PV); continue; }// End of PHI. case Instruction::Add: case Instruction::FAdd: case Instruction::Sub: case Instruction::FSub: case Instruction::Mul: case Instruction::FMul: case Instruction::UDiv: case Instruction::SDiv: case Instruction::FDiv: case Instruction::URem: case Instruction::SRem: case Instruction::FRem: case Instruction::Shl: case Instruction::LShr: case Instruction::AShr: case Instruction::And: case Instruction::Or: case Instruction::Xor: { // Just widen binops. BinaryOperator *BinOp = dyn_cast(it); setDebugLocFromInst(Builder, BinOp); VectorParts &A = getVectorValue(it->getOperand(0)); VectorParts &B = getVectorValue(it->getOperand(1)); // Use this vector value for all users of the original instruction. for (unsigned Part = 0; Part < UF; ++Part) { Value *V = Builder.CreateBinOp(BinOp->getOpcode(), A[Part], B[Part]); // Update the NSW, NUW and Exact flags. Notice: V can be an Undef. BinaryOperator *VecOp = dyn_cast(V); if (VecOp && isa(BinOp)) { VecOp->setHasNoSignedWrap(BinOp->hasNoSignedWrap()); VecOp->setHasNoUnsignedWrap(BinOp->hasNoUnsignedWrap()); } if (VecOp && isa(VecOp)) VecOp->setIsExact(BinOp->isExact()); Entry[Part] = V; } break; } case Instruction::Select: { // Widen selects. // If the selector is loop invariant we can create a select // instruction with a scalar condition. Otherwise, use vector-select. bool InvariantCond = SE->isLoopInvariant(SE->getSCEV(it->getOperand(0)), OrigLoop); setDebugLocFromInst(Builder, it); // The condition can be loop invariant but still defined inside the // loop. This means that we can't just use the original 'cond' value. // We have to take the 'vectorized' value and pick the first lane. // Instcombine will make this a no-op. VectorParts &Cond = getVectorValue(it->getOperand(0)); VectorParts &Op0 = getVectorValue(it->getOperand(1)); VectorParts &Op1 = getVectorValue(it->getOperand(2)); Value *ScalarCond = (VF == 1) ? Cond[0] : Builder.CreateExtractElement(Cond[0], Builder.getInt32(0)); for (unsigned Part = 0; Part < UF; ++Part) { Entry[Part] = Builder.CreateSelect( InvariantCond ? ScalarCond : Cond[Part], Op0[Part], Op1[Part]); } break; } case Instruction::ICmp: case Instruction::FCmp: { // Widen compares. Generate vector compares. bool FCmp = (it->getOpcode() == Instruction::FCmp); CmpInst *Cmp = dyn_cast(it); setDebugLocFromInst(Builder, it); VectorParts &A = getVectorValue(it->getOperand(0)); VectorParts &B = getVectorValue(it->getOperand(1)); for (unsigned Part = 0; Part < UF; ++Part) { Value *C = 0; if (FCmp) C = Builder.CreateFCmp(Cmp->getPredicate(), A[Part], B[Part]); else C = Builder.CreateICmp(Cmp->getPredicate(), A[Part], B[Part]); Entry[Part] = C; } break; } case Instruction::Store: case Instruction::Load: vectorizeMemoryInstruction(it, Legal); break; case Instruction::ZExt: case Instruction::SExt: case Instruction::FPToUI: case Instruction::FPToSI: case Instruction::FPExt: case Instruction::PtrToInt: case Instruction::IntToPtr: case Instruction::SIToFP: case Instruction::UIToFP: case Instruction::Trunc: case Instruction::FPTrunc: case Instruction::BitCast: { CastInst *CI = dyn_cast(it); setDebugLocFromInst(Builder, it); /// Optimize the special case where the source is the induction /// variable. Notice that we can only optimize the 'trunc' case /// because: a. FP conversions lose precision, b. sext/zext may wrap, /// c. other casts depend on pointer size. if (CI->getOperand(0) == OldInduction && it->getOpcode() == Instruction::Trunc) { Value *ScalarCast = Builder.CreateCast(CI->getOpcode(), Induction, CI->getType()); Value *Broadcasted = getBroadcastInstrs(ScalarCast); for (unsigned Part = 0; Part < UF; ++Part) Entry[Part] = getConsecutiveVector(Broadcasted, VF * Part, false); break; } /// Vectorize casts. Type *DestTy = (VF == 1) ? CI->getType() : VectorType::get(CI->getType(), VF); VectorParts &A = getVectorValue(it->getOperand(0)); for (unsigned Part = 0; Part < UF; ++Part) Entry[Part] = Builder.CreateCast(CI->getOpcode(), A[Part], DestTy); break; } case Instruction::Call: { // Ignore dbg intrinsics. if (isa(it)) break; setDebugLocFromInst(Builder, it); Module *M = BB->getParent()->getParent(); CallInst *CI = cast(it); Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI); assert(ID && "Not an intrinsic call!"); switch (ID) { case Intrinsic::lifetime_end: case Intrinsic::lifetime_start: scalarizeInstruction(it); break; default: for (unsigned Part = 0; Part < UF; ++Part) { SmallVector Args; for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i) { VectorParts &Arg = getVectorValue(CI->getArgOperand(i)); Args.push_back(Arg[Part]); } Type *Tys[] = {CI->getType()}; if (VF > 1) Tys[0] = VectorType::get(CI->getType()->getScalarType(), VF); Function *F = Intrinsic::getDeclaration(M, ID, Tys); Entry[Part] = Builder.CreateCall(F, Args); } break; } break; } default: // All other instructions are unsupported. Scalarize them. scalarizeInstruction(it); break; }// end of switch. }// end of for_each instr. } void InnerLoopVectorizer::updateAnalysis() { // Forget the original basic block. SE->forgetLoop(OrigLoop); // Update the dominator tree information. assert(DT->properlyDominates(LoopBypassBlocks.front(), LoopExitBlock) && "Entry does not dominate exit."); for (unsigned I = 1, E = LoopBypassBlocks.size(); I != E; ++I) DT->addNewBlock(LoopBypassBlocks[I], LoopBypassBlocks[I-1]); DT->addNewBlock(LoopVectorPreHeader, LoopBypassBlocks.back()); DT->addNewBlock(LoopVectorBody, LoopVectorPreHeader); DT->addNewBlock(LoopMiddleBlock, LoopBypassBlocks.front()); DT->addNewBlock(LoopScalarPreHeader, LoopMiddleBlock); DT->changeImmediateDominator(LoopScalarBody, LoopScalarPreHeader); DT->changeImmediateDominator(LoopExitBlock, LoopMiddleBlock); DEBUG(DT->verifyAnalysis()); } /// \brief Check whether it is safe to if-convert this phi node. /// /// Phi nodes with constant expressions that can trap are not safe to if /// convert. static bool canIfConvertPHINodes(BasicBlock *BB) { for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E; ++I) { PHINode *Phi = dyn_cast(I); if (!Phi) return true; for (unsigned p = 0, e = Phi->getNumIncomingValues(); p != e; ++p) if (Constant *C = dyn_cast(Phi->getIncomingValue(p))) if (C->canTrap()) return false; } return true; } bool LoopVectorizationLegality::canVectorizeWithIfConvert() { if (!EnableIfConversion) return false; assert(TheLoop->getNumBlocks() > 1 && "Single block loops are vectorizable"); // A list of pointers that we can safely read and write to. SmallPtrSet SafePointes; // Collect safe addresses. for (Loop::block_iterator BI = TheLoop->block_begin(), BE = TheLoop->block_end(); BI != BE; ++BI) { BasicBlock *BB = *BI; if (blockNeedsPredication(BB)) continue; for (BasicBlock::iterator I = BB->begin(), E = BB->end(); I != E; ++I) { if (LoadInst *LI = dyn_cast(I)) SafePointes.insert(LI->getPointerOperand()); else if (StoreInst *SI = dyn_cast(I)) SafePointes.insert(SI->getPointerOperand()); } } // Collect the blocks that need predication. BasicBlock *Header = TheLoop->getHeader(); for (Loop::block_iterator BI = TheLoop->block_begin(), BE = TheLoop->block_end(); BI != BE; ++BI) { BasicBlock *BB = *BI; // We don't support switch statements inside loops. if (!isa(BB->getTerminator())) return false; // We must be able to predicate all blocks that need to be predicated. if (blockNeedsPredication(BB)) { if (!blockCanBePredicated(BB, SafePointes)) return false; } else if (BB != Header && !canIfConvertPHINodes(BB)) return false; } // We can if-convert this loop. return true; } bool LoopVectorizationLegality::canVectorize() { // We must have a loop in canonical form. Loops with indirectbr in them cannot // be canonicalized. if (!TheLoop->getLoopPreheader()) return false; // We can only vectorize innermost loops. if (TheLoop->getSubLoopsVector().size()) return false; // We must have a single backedge. if (TheLoop->getNumBackEdges() != 1) return false; // We must have a single exiting block. if (!TheLoop->getExitingBlock()) return false; + // We only handle bottom-tested loops, i.e. loop in which the condition is + // checked at the end of each iteration. With that we can assume that all + // instructions in the loop are executed the same number of times. + if (TheLoop->getExitingBlock() != TheLoop->getLoopLatch()) { + DEBUG(dbgs() << "LV: loop control flow is not understood by vectorizer\n"); + return false; + } + // We need to have a loop header. DEBUG(dbgs() << "LV: Found a loop: " << TheLoop->getHeader()->getName() << '\n'); // Check if we can if-convert non single-bb loops. unsigned NumBlocks = TheLoop->getNumBlocks(); if (NumBlocks != 1 && !canVectorizeWithIfConvert()) { DEBUG(dbgs() << "LV: Can't if-convert the loop.\n"); return false; } // ScalarEvolution needs to be able to find the exit count. const SCEV *ExitCount = SE->getBackedgeTakenCount(TheLoop); if (ExitCount == SE->getCouldNotCompute()) { DEBUG(dbgs() << "LV: SCEV could not compute the loop exit count.\n"); return false; } // Do not loop-vectorize loops with a tiny trip count. BasicBlock *Latch = TheLoop->getLoopLatch(); unsigned TC = SE->getSmallConstantTripCount(TheLoop, Latch); if (TC > 0u && TC < TinyTripCountVectorThreshold) { DEBUG(dbgs() << "LV: Found a loop with a very small trip count. " << "This loop is not worth vectorizing.\n"); return false; } // Check if we can vectorize the instructions and CFG in this loop. if (!canVectorizeInstrs()) { DEBUG(dbgs() << "LV: Can't vectorize the instructions or CFG\n"); return false; } // Go over each instruction and look at memory deps. if (!canVectorizeMemory()) { DEBUG(dbgs() << "LV: Can't vectorize due to memory conflicts\n"); return false; } // Collect all of the variables that remain uniform after vectorization. collectLoopUniforms(); DEBUG(dbgs() << "LV: We can vectorize this loop" << (PtrRtCheck.Need ? " (with a runtime bound check)" : "") <<"!\n"); // Okay! We can vectorize. At this point we don't have any other mem analysis // which may limit our maximum vectorization factor, so just return true with // no restrictions. return true; } static Type *convertPointerToIntegerType(DataLayout &DL, Type *Ty) { if (Ty->isPointerTy()) return DL.getIntPtrType(Ty); // It is possible that char's or short's overflow when we ask for the loop's // trip count, work around this by changing the type size. if (Ty->getScalarSizeInBits() < 32) return Type::getInt32Ty(Ty->getContext()); return Ty; } static Type* getWiderType(DataLayout &DL, Type *Ty0, Type *Ty1) { Ty0 = convertPointerToIntegerType(DL, Ty0); Ty1 = convertPointerToIntegerType(DL, Ty1); if (Ty0->getScalarSizeInBits() > Ty1->getScalarSizeInBits()) return Ty0; return Ty1; } /// \brief Check that the instruction has outside loop users and is not an /// identified reduction variable. static bool hasOutsideLoopUser(const Loop *TheLoop, Instruction *Inst, SmallPtrSet &Reductions) { // Reduction instructions are allowed to have exit users. All other // instructions must not have external users. if (!Reductions.count(Inst)) //Check that all of the users of the loop are inside the BB. for (Value::use_iterator I = Inst->use_begin(), E = Inst->use_end(); I != E; ++I) { Instruction *U = cast(*I); // This user may be a reduction exit value. if (!TheLoop->contains(U)) { DEBUG(dbgs() << "LV: Found an outside user for : " << *U << '\n'); return true; } } return false; } bool LoopVectorizationLegality::canVectorizeInstrs() { BasicBlock *PreHeader = TheLoop->getLoopPreheader(); BasicBlock *Header = TheLoop->getHeader(); // Look for the attribute signaling the absence of NaNs. Function &F = *Header->getParent(); if (F.hasFnAttribute("no-nans-fp-math")) HasFunNoNaNAttr = F.getAttributes().getAttribute( AttributeSet::FunctionIndex, "no-nans-fp-math").getValueAsString() == "true"; // For each block in the loop. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { // Scan the instructions in the block and look for hazards. for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e; ++it) { if (PHINode *Phi = dyn_cast(it)) { Type *PhiTy = Phi->getType(); // Check that this PHI type is allowed. if (!PhiTy->isIntegerTy() && !PhiTy->isFloatingPointTy() && !PhiTy->isPointerTy()) { DEBUG(dbgs() << "LV: Found an non-int non-pointer PHI.\n"); return false; } // If this PHINode is not in the header block, then we know that we // can convert it to select during if-conversion. No need to check if // the PHIs in this block are induction or reduction variables. if (*bb != Header) { // Check that this instruction has no outside users or is an // identified reduction value with an outside user. if(!hasOutsideLoopUser(TheLoop, it, AllowedExit)) continue; return false; } // We only allow if-converted PHIs with more than two incoming values. if (Phi->getNumIncomingValues() != 2) { DEBUG(dbgs() << "LV: Found an invalid PHI.\n"); return false; } // This is the value coming from the preheader. Value *StartValue = Phi->getIncomingValueForBlock(PreHeader); // Check if this is an induction variable. InductionKind IK = isInductionVariable(Phi); if (IK_NoInduction != IK) { // Get the widest type. if (!WidestIndTy) WidestIndTy = convertPointerToIntegerType(*DL, PhiTy); else WidestIndTy = getWiderType(*DL, PhiTy, WidestIndTy); // Int inductions are special because we only allow one IV. if (IK == IK_IntInduction) { // Use the phi node with the widest type as induction. Use the last // one if there are multiple (no good reason for doing this other // than it is expedient). if (!Induction || PhiTy == WidestIndTy) Induction = Phi; } DEBUG(dbgs() << "LV: Found an induction variable.\n"); Inductions[Phi] = InductionInfo(StartValue, IK); // Until we explicitly handle the case of an induction variable with // an outside loop user we have to give up vectorizing this loop. if (hasOutsideLoopUser(TheLoop, it, AllowedExit)) return false; continue; } if (AddReductionVar(Phi, RK_IntegerAdd)) { DEBUG(dbgs() << "LV: Found an ADD reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerMult)) { DEBUG(dbgs() << "LV: Found a MUL reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerOr)) { DEBUG(dbgs() << "LV: Found an OR reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerAnd)) { DEBUG(dbgs() << "LV: Found an AND reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerXor)) { DEBUG(dbgs() << "LV: Found a XOR reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_IntegerMinMax)) { DEBUG(dbgs() << "LV: Found a MINMAX reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_FloatMult)) { DEBUG(dbgs() << "LV: Found an FMult reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_FloatAdd)) { DEBUG(dbgs() << "LV: Found an FAdd reduction PHI."<< *Phi <<"\n"); continue; } if (AddReductionVar(Phi, RK_FloatMinMax)) { DEBUG(dbgs() << "LV: Found an float MINMAX reduction PHI."<< *Phi << "\n"); continue; } DEBUG(dbgs() << "LV: Found an unidentified PHI."<< *Phi <<"\n"); return false; }// end of PHI handling // We still don't handle functions. However, we can ignore dbg intrinsic // calls and we do handle certain intrinsic and libm functions. CallInst *CI = dyn_cast(it); if (CI && !getIntrinsicIDForCall(CI, TLI) && !isa(CI)) { DEBUG(dbgs() << "LV: Found a call site.\n"); return false; } // Check that the instruction return type is vectorizable. // Also, we can't vectorize extractelement instructions. if ((!VectorType::isValidElementType(it->getType()) && !it->getType()->isVoidTy()) || isa(it)) { DEBUG(dbgs() << "LV: Found unvectorizable type.\n"); return false; } // Check that the stored type is vectorizable. if (StoreInst *ST = dyn_cast(it)) { Type *T = ST->getValueOperand()->getType(); if (!VectorType::isValidElementType(T)) return false; } // Reduction instructions are allowed to have exit users. // All other instructions must not have external users. if (hasOutsideLoopUser(TheLoop, it, AllowedExit)) return false; } // next instr. } if (!Induction) { DEBUG(dbgs() << "LV: Did not find one integer induction var.\n"); if (Inductions.empty()) return false; } return true; } void LoopVectorizationLegality::collectLoopUniforms() { // We now know that the loop is vectorizable! // Collect variables that will remain uniform after vectorization. std::vector Worklist; BasicBlock *Latch = TheLoop->getLoopLatch(); // Start with the conditional branch and walk up the block. Worklist.push_back(Latch->getTerminator()->getOperand(0)); while (Worklist.size()) { Instruction *I = dyn_cast(Worklist.back()); Worklist.pop_back(); // Look at instructions inside this loop. // Stop when reaching PHI nodes. // TODO: we need to follow values all over the loop, not only in this block. if (!I || !TheLoop->contains(I) || isa(I)) continue; // This is a known uniform. Uniforms.insert(I); // Insert all operands. Worklist.insert(Worklist.end(), I->op_begin(), I->op_end()); } } namespace { /// \brief Analyses memory accesses in a loop. /// /// Checks whether run time pointer checks are needed and builds sets for data /// dependence checking. class AccessAnalysis { public: /// \brief Read or write access location. typedef PointerIntPair MemAccessInfo; typedef SmallPtrSet MemAccessInfoSet; /// \brief Set of potential dependent memory accesses. typedef EquivalenceClasses DepCandidates; AccessAnalysis(DataLayout *Dl, DepCandidates &DA) : DL(Dl), DepCands(DA), AreAllWritesIdentified(true), AreAllReadsIdentified(true), IsRTCheckNeeded(false) {} /// \brief Register a load and whether it is only read from. void addLoad(Value *Ptr, bool IsReadOnly) { Accesses.insert(MemAccessInfo(Ptr, false)); if (IsReadOnly) ReadOnlyPtr.insert(Ptr); } /// \brief Register a store. void addStore(Value *Ptr) { Accesses.insert(MemAccessInfo(Ptr, true)); } /// \brief Check whether we can check the pointers at runtime for /// non-intersection. bool canCheckPtrAtRT(LoopVectorizationLegality::RuntimePointerCheck &RtCheck, unsigned &NumComparisons, ScalarEvolution *SE, Loop *TheLoop, bool ShouldCheckStride = false); /// \brief Goes over all memory accesses, checks whether a RT check is needed /// and builds sets of dependent accesses. void buildDependenceSets() { // Process read-write pointers first. processMemAccesses(false); // Next, process read pointers. processMemAccesses(true); } bool isRTCheckNeeded() { return IsRTCheckNeeded; } bool isDependencyCheckNeeded() { return !CheckDeps.empty(); } void resetDepChecks() { CheckDeps.clear(); } MemAccessInfoSet &getDependenciesToCheck() { return CheckDeps; } private: typedef SetVector PtrAccessSet; typedef DenseMap UnderlyingObjToAccessMap; /// \brief Go over all memory access or only the deferred ones if /// \p UseDeferred is true and check whether runtime pointer checks are needed /// and build sets of dependency check candidates. void processMemAccesses(bool UseDeferred); /// Set of all accesses. PtrAccessSet Accesses; /// Set of access to check after all writes have been processed. PtrAccessSet DeferredAccesses; /// Map of pointers to last access encountered. UnderlyingObjToAccessMap ObjToLastAccess; /// Set of accesses that need a further dependence check. MemAccessInfoSet CheckDeps; /// Set of pointers that are read only. SmallPtrSet ReadOnlyPtr; /// Set of underlying objects already written to. SmallPtrSet WriteObjects; DataLayout *DL; /// Sets of potentially dependent accesses - members of one set share an /// underlying pointer. The set "CheckDeps" identfies which sets really need a /// dependence check. DepCandidates &DepCands; bool AreAllWritesIdentified; bool AreAllReadsIdentified; bool IsRTCheckNeeded; }; } // end anonymous namespace /// \brief Check whether a pointer can participate in a runtime bounds check. static bool hasComputableBounds(ScalarEvolution *SE, Value *Ptr) { const SCEV *PtrScev = SE->getSCEV(Ptr); const SCEVAddRecExpr *AR = dyn_cast(PtrScev); if (!AR) return false; return AR->isAffine(); } /// \brief Check the stride of the pointer and ensure that it does not wrap in /// the address space. static int isStridedPtr(ScalarEvolution *SE, DataLayout *DL, Value *Ptr, const Loop *Lp); bool AccessAnalysis::canCheckPtrAtRT( LoopVectorizationLegality::RuntimePointerCheck &RtCheck, unsigned &NumComparisons, ScalarEvolution *SE, Loop *TheLoop, bool ShouldCheckStride) { // Find pointers with computable bounds. We are going to use this information // to place a runtime bound check. unsigned NumReadPtrChecks = 0; unsigned NumWritePtrChecks = 0; bool CanDoRT = true; bool IsDepCheckNeeded = isDependencyCheckNeeded(); // We assign consecutive id to access from different dependence sets. // Accesses within the same set don't need a runtime check. unsigned RunningDepId = 1; DenseMap DepSetId; for (PtrAccessSet::iterator AI = Accesses.begin(), AE = Accesses.end(); AI != AE; ++AI) { const MemAccessInfo &Access = *AI; Value *Ptr = Access.getPointer(); bool IsWrite = Access.getInt(); // Just add write checks if we have both. if (!IsWrite && Accesses.count(MemAccessInfo(Ptr, true))) continue; if (IsWrite) ++NumWritePtrChecks; else ++NumReadPtrChecks; if (hasComputableBounds(SE, Ptr) && // When we run after a failing dependency check we have to make sure we // don't have wrapping pointers. (!ShouldCheckStride || isStridedPtr(SE, DL, Ptr, TheLoop) == 1)) { // The id of the dependence set. unsigned DepId; if (IsDepCheckNeeded) { Value *Leader = DepCands.getLeaderValue(Access).getPointer(); unsigned &LeaderId = DepSetId[Leader]; if (!LeaderId) LeaderId = RunningDepId++; DepId = LeaderId; } else // Each access has its own dependence set. DepId = RunningDepId++; RtCheck.insert(SE, TheLoop, Ptr, IsWrite, DepId); DEBUG(dbgs() << "LV: Found a runtime check ptr:" << *Ptr << '\n'); } else { CanDoRT = false; } } if (IsDepCheckNeeded && CanDoRT && RunningDepId == 2) NumComparisons = 0; // Only one dependence set. else { NumComparisons = (NumWritePtrChecks * (NumReadPtrChecks + NumWritePtrChecks - 1)); } // If the pointers that we would use for the bounds comparison have different // address spaces, assume the values aren't directly comparable, so we can't // use them for the runtime check. We also have to assume they could // overlap. In the future there should be metadata for whether address spaces // are disjoint. unsigned NumPointers = RtCheck.Pointers.size(); for (unsigned i = 0; i < NumPointers; ++i) { for (unsigned j = i + 1; j < NumPointers; ++j) { // Only need to check pointers between two different dependency sets. if (RtCheck.DependencySetId[i] == RtCheck.DependencySetId[j]) continue; Value *PtrI = RtCheck.Pointers[i]; Value *PtrJ = RtCheck.Pointers[j]; unsigned ASi = PtrI->getType()->getPointerAddressSpace(); unsigned ASj = PtrJ->getType()->getPointerAddressSpace(); if (ASi != ASj) { DEBUG(dbgs() << "LV: Runtime check would require comparison between" " different address spaces\n"); return false; } } } return CanDoRT; } static bool isFunctionScopeIdentifiedObject(Value *Ptr) { return isNoAliasArgument(Ptr) || isNoAliasCall(Ptr) || isa(Ptr); } void AccessAnalysis::processMemAccesses(bool UseDeferred) { // We process the set twice: first we process read-write pointers, last we // process read-only pointers. This allows us to skip dependence tests for // read-only pointers. PtrAccessSet &S = UseDeferred ? DeferredAccesses : Accesses; for (PtrAccessSet::iterator AI = S.begin(), AE = S.end(); AI != AE; ++AI) { const MemAccessInfo &Access = *AI; Value *Ptr = Access.getPointer(); bool IsWrite = Access.getInt(); DepCands.insert(Access); // Memorize read-only pointers for later processing and skip them in the // first round (they need to be checked after we have seen all write // pointers). Note: we also mark pointer that are not consecutive as // "read-only" pointers (so that we check "a[b[i]] +="). Hence, we need the // second check for "!IsWrite". bool IsReadOnlyPtr = ReadOnlyPtr.count(Ptr) && !IsWrite; if (!UseDeferred && IsReadOnlyPtr) { DeferredAccesses.insert(Access); continue; } bool NeedDepCheck = false; // Check whether there is the possiblity of dependency because of underlying // objects being the same. typedef SmallVector ValueVector; ValueVector TempObjects; GetUnderlyingObjects(Ptr, TempObjects, DL); for (ValueVector::iterator UI = TempObjects.begin(), UE = TempObjects.end(); UI != UE; ++UI) { Value *UnderlyingObj = *UI; // If this is a write then it needs to be an identified object. If this a // read and all writes (so far) are identified function scope objects we // don't need an identified underlying object but only an Argument (the // next write is going to invalidate this assumption if it is // unidentified). // This is a micro-optimization for the case where all writes are // identified and we have one argument pointer. // Otherwise, we do need a runtime check. if ((IsWrite && !isFunctionScopeIdentifiedObject(UnderlyingObj)) || (!IsWrite && (!AreAllWritesIdentified || !isa(UnderlyingObj)) && !isIdentifiedObject(UnderlyingObj))) { DEBUG(dbgs() << "LV: Found an unidentified " << (IsWrite ? "write" : "read" ) << " ptr: " << *UnderlyingObj << "\n"); IsRTCheckNeeded = (IsRTCheckNeeded || !isIdentifiedObject(UnderlyingObj) || !AreAllReadsIdentified); if (IsWrite) AreAllWritesIdentified = false; if (!IsWrite) AreAllReadsIdentified = false; } // If this is a write - check other reads and writes for conflicts. If // this is a read only check other writes for conflicts (but only if there // is no other write to the ptr - this is an optimization to catch "a[i] = // a[i] + " without having to do a dependence check). if ((IsWrite || IsReadOnlyPtr) && WriteObjects.count(UnderlyingObj)) NeedDepCheck = true; if (IsWrite) WriteObjects.insert(UnderlyingObj); // Create sets of pointers connected by shared underlying objects. UnderlyingObjToAccessMap::iterator Prev = ObjToLastAccess.find(UnderlyingObj); if (Prev != ObjToLastAccess.end()) DepCands.unionSets(Access, Prev->second); ObjToLastAccess[UnderlyingObj] = Access; } if (NeedDepCheck) CheckDeps.insert(Access); } } namespace { /// \brief Checks memory dependences among accesses to the same underlying /// object to determine whether there vectorization is legal or not (and at /// which vectorization factor). /// /// This class works under the assumption that we already checked that memory /// locations with different underlying pointers are "must-not alias". /// We use the ScalarEvolution framework to symbolically evalutate access /// functions pairs. Since we currently don't restructure the loop we can rely /// on the program order of memory accesses to determine their safety. /// At the moment we will only deem accesses as safe for: /// * A negative constant distance assuming program order. /// /// Safe: tmp = a[i + 1]; OR a[i + 1] = x; /// a[i] = tmp; y = a[i]; /// /// The latter case is safe because later checks guarantuee that there can't /// be a cycle through a phi node (that is, we check that "x" and "y" is not /// the same variable: a header phi can only be an induction or a reduction, a /// reduction can't have a memory sink, an induction can't have a memory /// source). This is important and must not be violated (or we have to /// resort to checking for cycles through memory). /// /// * A positive constant distance assuming program order that is bigger /// than the biggest memory access. /// /// tmp = a[i] OR b[i] = x /// a[i+2] = tmp y = b[i+2]; /// /// Safe distance: 2 x sizeof(a[0]), and 2 x sizeof(b[0]), respectively. /// /// * Zero distances and all accesses have the same size. /// class MemoryDepChecker { public: typedef PointerIntPair MemAccessInfo; typedef SmallPtrSet MemAccessInfoSet; MemoryDepChecker(ScalarEvolution *Se, DataLayout *Dl, const Loop *L) : SE(Se), DL(Dl), InnermostLoop(L), AccessIdx(0), ShouldRetryWithRuntimeCheck(false) {} /// \brief Register the location (instructions are given increasing numbers) /// of a write access. void addAccess(StoreInst *SI) { Value *Ptr = SI->getPointerOperand(); Accesses[MemAccessInfo(Ptr, true)].push_back(AccessIdx); InstMap.push_back(SI); ++AccessIdx; } /// \brief Register the location (instructions are given increasing numbers) /// of a write access. void addAccess(LoadInst *LI) { Value *Ptr = LI->getPointerOperand(); Accesses[MemAccessInfo(Ptr, false)].push_back(AccessIdx); InstMap.push_back(LI); ++AccessIdx; } /// \brief Check whether the dependencies between the accesses are safe. /// /// Only checks sets with elements in \p CheckDeps. bool areDepsSafe(AccessAnalysis::DepCandidates &AccessSets, MemAccessInfoSet &CheckDeps); /// \brief The maximum number of bytes of a vector register we can vectorize /// the accesses safely with. unsigned getMaxSafeDepDistBytes() { return MaxSafeDepDistBytes; } /// \brief In same cases when the dependency check fails we can still /// vectorize the loop with a dynamic array access check. bool shouldRetryWithRuntimeCheck() { return ShouldRetryWithRuntimeCheck; } private: ScalarEvolution *SE; DataLayout *DL; const Loop *InnermostLoop; /// \brief Maps access locations (ptr, read/write) to program order. DenseMap > Accesses; /// \brief Memory access instructions in program order. SmallVector InstMap; /// \brief The program order index to be used for the next instruction. unsigned AccessIdx; // We can access this many bytes in parallel safely. unsigned MaxSafeDepDistBytes; /// \brief If we see a non constant dependence distance we can still try to /// vectorize this loop with runtime checks. bool ShouldRetryWithRuntimeCheck; /// \brief Check whether there is a plausible dependence between the two /// accesses. /// /// Access \p A must happen before \p B in program order. The two indices /// identify the index into the program order map. /// /// This function checks whether there is a plausible dependence (or the /// absence of such can't be proved) between the two accesses. If there is a /// plausible dependence but the dependence distance is bigger than one /// element access it records this distance in \p MaxSafeDepDistBytes (if this /// distance is smaller than any other distance encountered so far). /// Otherwise, this function returns true signaling a possible dependence. bool isDependent(const MemAccessInfo &A, unsigned AIdx, const MemAccessInfo &B, unsigned BIdx); /// \brief Check whether the data dependence could prevent store-load /// forwarding. bool couldPreventStoreLoadForward(unsigned Distance, unsigned TypeByteSize); }; } // end anonymous namespace static bool isInBoundsGep(Value *Ptr) { if (GetElementPtrInst *GEP = dyn_cast(Ptr)) return GEP->isInBounds(); return false; } /// \brief Check whether the access through \p Ptr has a constant stride. static int isStridedPtr(ScalarEvolution *SE, DataLayout *DL, Value *Ptr, const Loop *Lp) { const Type *Ty = Ptr->getType(); assert(Ty->isPointerTy() && "Unexpected non ptr"); // Make sure that the pointer does not point to aggregate types. const PointerType *PtrTy = cast(Ty); if (PtrTy->getElementType()->isAggregateType()) { DEBUG(dbgs() << "LV: Bad stride - Not a pointer to a scalar type" << *Ptr << "\n"); return 0; } const SCEV *PtrScev = SE->getSCEV(Ptr); const SCEVAddRecExpr *AR = dyn_cast(PtrScev); if (!AR) { DEBUG(dbgs() << "LV: Bad stride - Not an AddRecExpr pointer " << *Ptr << " SCEV: " << *PtrScev << "\n"); return 0; } // The accesss function must stride over the innermost loop. if (Lp != AR->getLoop()) { DEBUG(dbgs() << "LV: Bad stride - Not striding over innermost loop " << *Ptr << " SCEV: " << *PtrScev << "\n"); } // The address calculation must not wrap. Otherwise, a dependence could be // inverted. // An inbounds getelementptr that is a AddRec with a unit stride // cannot wrap per definition. The unit stride requirement is checked later. // An getelementptr without an inbounds attribute and unit stride would have // to access the pointer value "0" which is undefined behavior in address // space 0, therefore we can also vectorize this case. bool IsInBoundsGEP = isInBoundsGep(Ptr); bool IsNoWrapAddRec = AR->getNoWrapFlags(SCEV::NoWrapMask); bool IsInAddressSpaceZero = PtrTy->getAddressSpace() == 0; if (!IsNoWrapAddRec && !IsInBoundsGEP && !IsInAddressSpaceZero) { DEBUG(dbgs() << "LV: Bad stride - Pointer may wrap in the address space " << *Ptr << " SCEV: " << *PtrScev << "\n"); return 0; } // Check the step is constant. const SCEV *Step = AR->getStepRecurrence(*SE); // Calculate the pointer stride and check if it is consecutive. const SCEVConstant *C = dyn_cast(Step); if (!C) { DEBUG(dbgs() << "LV: Bad stride - Not a constant strided " << *Ptr << " SCEV: " << *PtrScev << "\n"); return 0; } int64_t Size = DL->getTypeAllocSize(PtrTy->getElementType()); const APInt &APStepVal = C->getValue()->getValue(); // Huge step value - give up. if (APStepVal.getBitWidth() > 64) return 0; int64_t StepVal = APStepVal.getSExtValue(); // Strided access. int64_t Stride = StepVal / Size; int64_t Rem = StepVal % Size; if (Rem) return 0; // If the SCEV could wrap but we have an inbounds gep with a unit stride we // know we can't "wrap around the address space". In case of address space // zero we know that this won't happen without triggering undefined behavior. if (!IsNoWrapAddRec && (IsInBoundsGEP || IsInAddressSpaceZero) && Stride != 1 && Stride != -1) return 0; return Stride; } bool MemoryDepChecker::couldPreventStoreLoadForward(unsigned Distance, unsigned TypeByteSize) { // If loads occur at a distance that is not a multiple of a feasible vector // factor store-load forwarding does not take place. // Positive dependences might cause troubles because vectorizing them might // prevent store-load forwarding making vectorized code run a lot slower. // a[i] = a[i-3] ^ a[i-8]; // The stores to a[i:i+1] don't align with the stores to a[i-3:i-2] and // hence on your typical architecture store-load forwarding does not take // place. Vectorizing in such cases does not make sense. // Store-load forwarding distance. const unsigned NumCyclesForStoreLoadThroughMemory = 8*TypeByteSize; // Maximum vector factor. unsigned MaxVFWithoutSLForwardIssues = MaxVectorWidth*TypeByteSize; if(MaxSafeDepDistBytes < MaxVFWithoutSLForwardIssues) MaxVFWithoutSLForwardIssues = MaxSafeDepDistBytes; for (unsigned vf = 2*TypeByteSize; vf <= MaxVFWithoutSLForwardIssues; vf *= 2) { if (Distance % vf && Distance / vf < NumCyclesForStoreLoadThroughMemory) { MaxVFWithoutSLForwardIssues = (vf >>=1); break; } } if (MaxVFWithoutSLForwardIssues< 2*TypeByteSize) { DEBUG(dbgs() << "LV: Distance " << Distance << " that could cause a store-load forwarding conflict\n"); return true; } if (MaxVFWithoutSLForwardIssues < MaxSafeDepDistBytes && MaxVFWithoutSLForwardIssues != MaxVectorWidth*TypeByteSize) MaxSafeDepDistBytes = MaxVFWithoutSLForwardIssues; return false; } bool MemoryDepChecker::isDependent(const MemAccessInfo &A, unsigned AIdx, const MemAccessInfo &B, unsigned BIdx) { assert (AIdx < BIdx && "Must pass arguments in program order"); Value *APtr = A.getPointer(); Value *BPtr = B.getPointer(); bool AIsWrite = A.getInt(); bool BIsWrite = B.getInt(); // Two reads are independent. if (!AIsWrite && !BIsWrite) return false; const SCEV *AScev = SE->getSCEV(APtr); const SCEV *BScev = SE->getSCEV(BPtr); int StrideAPtr = isStridedPtr(SE, DL, APtr, InnermostLoop); int StrideBPtr = isStridedPtr(SE, DL, BPtr, InnermostLoop); const SCEV *Src = AScev; const SCEV *Sink = BScev; // If the induction step is negative we have to invert source and sink of the // dependence. if (StrideAPtr < 0) { //Src = BScev; //Sink = AScev; std::swap(APtr, BPtr); std::swap(Src, Sink); std::swap(AIsWrite, BIsWrite); std::swap(AIdx, BIdx); std::swap(StrideAPtr, StrideBPtr); } const SCEV *Dist = SE->getMinusSCEV(Sink, Src); DEBUG(dbgs() << "LV: Src Scev: " << *Src << "Sink Scev: " << *Sink << "(Induction step: " << StrideAPtr << ")\n"); DEBUG(dbgs() << "LV: Distance for " << *InstMap[AIdx] << " to " << *InstMap[BIdx] << ": " << *Dist << "\n"); // Need consecutive accesses. We don't want to vectorize // "A[B[i]] += ..." and similar code or pointer arithmetic that could wrap in // the address space. if (!StrideAPtr || !StrideBPtr || StrideAPtr != StrideBPtr){ DEBUG(dbgs() << "Non-consecutive pointer access\n"); return true; } const SCEVConstant *C = dyn_cast(Dist); if (!C) { DEBUG(dbgs() << "LV: Dependence because of non constant distance\n"); ShouldRetryWithRuntimeCheck = true; return true; } Type *ATy = APtr->getType()->getPointerElementType(); Type *BTy = BPtr->getType()->getPointerElementType(); unsigned TypeByteSize = DL->getTypeAllocSize(ATy); // Negative distances are not plausible dependencies. const APInt &Val = C->getValue()->getValue(); if (Val.isNegative()) { bool IsTrueDataDependence = (AIsWrite && !BIsWrite); if (IsTrueDataDependence && (couldPreventStoreLoadForward(Val.abs().getZExtValue(), TypeByteSize) || ATy != BTy)) return true; DEBUG(dbgs() << "LV: Dependence is negative: NoDep\n"); return false; } // Write to the same location with the same size. // Could be improved to assert type sizes are the same (i32 == float, etc). if (Val == 0) { if (ATy == BTy) return false; DEBUG(dbgs() << "LV: Zero dependence difference but different types\n"); return true; } assert(Val.isStrictlyPositive() && "Expect a positive value"); // Positive distance bigger than max vectorization factor. if (ATy != BTy) { DEBUG(dbgs() << "LV: ReadWrite-Write positive dependency with different types\n"); return false; } unsigned Distance = (unsigned) Val.getZExtValue(); // Bail out early if passed-in parameters make vectorization not feasible. unsigned ForcedFactor = VectorizationFactor ? VectorizationFactor : 1; unsigned ForcedUnroll = VectorizationUnroll ? VectorizationUnroll : 1; // The distance must be bigger than the size needed for a vectorized version // of the operation and the size of the vectorized operation must not be // bigger than the currrent maximum size. if (Distance < 2*TypeByteSize || 2*TypeByteSize > MaxSafeDepDistBytes || Distance < TypeByteSize * ForcedUnroll * ForcedFactor) { DEBUG(dbgs() << "LV: Failure because of Positive distance " << Val.getSExtValue() << '\n'); return true; } MaxSafeDepDistBytes = Distance < MaxSafeDepDistBytes ? Distance : MaxSafeDepDistBytes; bool IsTrueDataDependence = (!AIsWrite && BIsWrite); if (IsTrueDataDependence && couldPreventStoreLoadForward(Distance, TypeByteSize)) return true; DEBUG(dbgs() << "LV: Positive distance " << Val.getSExtValue() << " with max VF = " << MaxSafeDepDistBytes / TypeByteSize << '\n'); return false; } bool MemoryDepChecker::areDepsSafe(AccessAnalysis::DepCandidates &AccessSets, MemAccessInfoSet &CheckDeps) { MaxSafeDepDistBytes = -1U; while (!CheckDeps.empty()) { MemAccessInfo CurAccess = *CheckDeps.begin(); // Get the relevant memory access set. EquivalenceClasses::iterator I = AccessSets.findValue(AccessSets.getLeaderValue(CurAccess)); // Check accesses within this set. EquivalenceClasses::member_iterator AI, AE; AI = AccessSets.member_begin(I), AE = AccessSets.member_end(); // Check every access pair. while (AI != AE) { CheckDeps.erase(*AI); EquivalenceClasses::member_iterator OI = llvm::next(AI); while (OI != AE) { // Check every accessing instruction pair in program order. for (std::vector::iterator I1 = Accesses[*AI].begin(), I1E = Accesses[*AI].end(); I1 != I1E; ++I1) for (std::vector::iterator I2 = Accesses[*OI].begin(), I2E = Accesses[*OI].end(); I2 != I2E; ++I2) { if (*I1 < *I2 && isDependent(*AI, *I1, *OI, *I2)) return false; if (*I2 < *I1 && isDependent(*OI, *I2, *AI, *I1)) return false; } ++OI; } AI++; } } return true; } bool LoopVectorizationLegality::canVectorizeMemory() { typedef SmallVector ValueVector; typedef SmallPtrSet ValueSet; // Holds the Load and Store *instructions*. ValueVector Loads; ValueVector Stores; // Holds all the different accesses in the loop. unsigned NumReads = 0; unsigned NumReadWrites = 0; PtrRtCheck.Pointers.clear(); PtrRtCheck.Need = false; const bool IsAnnotatedParallel = TheLoop->isAnnotatedParallel(); MemoryDepChecker DepChecker(SE, DL, TheLoop); // For each block. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { // Scan the BB and collect legal loads and stores. for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e; ++it) { // If this is a load, save it. If this instruction can read from memory // but is not a load, then we quit. Notice that we don't handle function // calls that read or write. if (it->mayReadFromMemory()) { // Many math library functions read the rounding mode. We will only // vectorize a loop if it contains known function calls that don't set // the flag. Therefore, it is safe to ignore this read from memory. CallInst *Call = dyn_cast(it); if (Call && getIntrinsicIDForCall(Call, TLI)) continue; LoadInst *Ld = dyn_cast(it); if (!Ld) return false; if (!Ld->isSimple() && !IsAnnotatedParallel) { DEBUG(dbgs() << "LV: Found a non-simple load.\n"); return false; } Loads.push_back(Ld); DepChecker.addAccess(Ld); continue; } // Save 'store' instructions. Abort if other instructions write to memory. if (it->mayWriteToMemory()) { StoreInst *St = dyn_cast(it); if (!St) return false; if (!St->isSimple() && !IsAnnotatedParallel) { DEBUG(dbgs() << "LV: Found a non-simple store.\n"); return false; } Stores.push_back(St); DepChecker.addAccess(St); } } // Next instr. } // Next block. // Now we have two lists that hold the loads and the stores. // Next, we find the pointers that they use. // Check if we see any stores. If there are no stores, then we don't // care if the pointers are *restrict*. if (!Stores.size()) { DEBUG(dbgs() << "LV: Found a read-only loop!\n"); return true; } AccessAnalysis::DepCandidates DependentAccesses; AccessAnalysis Accesses(DL, DependentAccesses); // Holds the analyzed pointers. We don't want to call GetUnderlyingObjects // multiple times on the same object. If the ptr is accessed twice, once // for read and once for write, it will only appear once (on the write // list). This is okay, since we are going to check for conflicts between // writes and between reads and writes, but not between reads and reads. ValueSet Seen; ValueVector::iterator I, IE; for (I = Stores.begin(), IE = Stores.end(); I != IE; ++I) { StoreInst *ST = cast(*I); Value* Ptr = ST->getPointerOperand(); if (isUniform(Ptr)) { DEBUG(dbgs() << "LV: We don't allow storing to uniform addresses\n"); return false; } // If we did *not* see this pointer before, insert it to the read-write // list. At this phase it is only a 'write' list. if (Seen.insert(Ptr)) { ++NumReadWrites; Accesses.addStore(Ptr); } } if (IsAnnotatedParallel) { DEBUG(dbgs() << "LV: A loop annotated parallel, ignore memory dependency " << "checks.\n"); return true; } for (I = Loads.begin(), IE = Loads.end(); I != IE; ++I) { LoadInst *LD = cast(*I); Value* Ptr = LD->getPointerOperand(); // If we did *not* see this pointer before, insert it to the // read list. If we *did* see it before, then it is already in // the read-write list. This allows us to vectorize expressions // such as A[i] += x; Because the address of A[i] is a read-write // pointer. This only works if the index of A[i] is consecutive. // If the address of i is unknown (for example A[B[i]]) then we may // read a few words, modify, and write a few words, and some of the // words may be written to the same address. bool IsReadOnlyPtr = false; if (Seen.insert(Ptr) || !isStridedPtr(SE, DL, Ptr, TheLoop)) { ++NumReads; IsReadOnlyPtr = true; } Accesses.addLoad(Ptr, IsReadOnlyPtr); } // If we write (or read-write) to a single destination and there are no // other reads in this loop then is it safe to vectorize. if (NumReadWrites == 1 && NumReads == 0) { DEBUG(dbgs() << "LV: Found a write-only loop!\n"); return true; } // Build dependence sets and check whether we need a runtime pointer bounds // check. Accesses.buildDependenceSets(); bool NeedRTCheck = Accesses.isRTCheckNeeded(); // Find pointers with computable bounds. We are going to use this information // to place a runtime bound check. unsigned NumComparisons = 0; bool CanDoRT = false; if (NeedRTCheck) CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE, TheLoop); DEBUG(dbgs() << "LV: We need to do " << NumComparisons << " pointer comparisons.\n"); // If we only have one set of dependences to check pointers among we don't // need a runtime check. if (NumComparisons == 0 && NeedRTCheck) NeedRTCheck = false; // Check that we did not collect too many pointers or found an unsizeable // pointer. if (!CanDoRT || NumComparisons > RuntimeMemoryCheckThreshold) { PtrRtCheck.reset(); CanDoRT = false; } if (CanDoRT) { DEBUG(dbgs() << "LV: We can perform a memory runtime check if needed.\n"); } if (NeedRTCheck && !CanDoRT) { DEBUG(dbgs() << "LV: We can't vectorize because we can't find " << "the array bounds.\n"); PtrRtCheck.reset(); return false; } PtrRtCheck.Need = NeedRTCheck; bool CanVecMem = true; if (Accesses.isDependencyCheckNeeded()) { DEBUG(dbgs() << "LV: Checking memory dependencies\n"); CanVecMem = DepChecker.areDepsSafe(DependentAccesses, Accesses.getDependenciesToCheck()); MaxSafeDepDistBytes = DepChecker.getMaxSafeDepDistBytes(); if (!CanVecMem && DepChecker.shouldRetryWithRuntimeCheck()) { DEBUG(dbgs() << "LV: Retrying with memory checks\n"); NeedRTCheck = true; // Clear the dependency checks. We assume they are not needed. Accesses.resetDepChecks(); PtrRtCheck.reset(); PtrRtCheck.Need = true; CanDoRT = Accesses.canCheckPtrAtRT(PtrRtCheck, NumComparisons, SE, TheLoop, true); // Check that we did not collect too many pointers or found an unsizeable // pointer. if (!CanDoRT || NumComparisons > RuntimeMemoryCheckThreshold) { DEBUG(dbgs() << "LV: Can't vectorize with memory checks\n"); PtrRtCheck.reset(); return false; } CanVecMem = true; } } DEBUG(dbgs() << "LV: We" << (NeedRTCheck ? "" : " don't") << " need a runtime memory check.\n"); return CanVecMem; } static bool hasMultipleUsesOf(Instruction *I, SmallPtrSet &Insts) { unsigned NumUses = 0; for(User::op_iterator Use = I->op_begin(), E = I->op_end(); Use != E; ++Use) { if (Insts.count(dyn_cast(*Use))) ++NumUses; if (NumUses > 1) return true; } return false; } static bool areAllUsesIn(Instruction *I, SmallPtrSet &Set) { for(User::op_iterator Use = I->op_begin(), E = I->op_end(); Use != E; ++Use) if (!Set.count(dyn_cast(*Use))) return false; return true; } bool LoopVectorizationLegality::AddReductionVar(PHINode *Phi, ReductionKind Kind) { if (Phi->getNumIncomingValues() != 2) return false; // Reduction variables are only found in the loop header block. if (Phi->getParent() != TheLoop->getHeader()) return false; // Obtain the reduction start value from the value that comes from the loop // preheader. Value *RdxStart = Phi->getIncomingValueForBlock(TheLoop->getLoopPreheader()); // ExitInstruction is the single value which is used outside the loop. // We only allow for a single reduction value to be used outside the loop. // This includes users of the reduction, variables (which form a cycle // which ends in the phi node). Instruction *ExitInstruction = 0; // Indicates that we found a reduction operation in our scan. bool FoundReduxOp = false; // We start with the PHI node and scan for all of the users of this // instruction. All users must be instructions that can be used as reduction // variables (such as ADD). We must have a single out-of-block user. The cycle // must include the original PHI. bool FoundStartPHI = false; // To recognize min/max patterns formed by a icmp select sequence, we store // the number of instruction we saw from the recognized min/max pattern, // to make sure we only see exactly the two instructions. unsigned NumCmpSelectPatternInst = 0; ReductionInstDesc ReduxDesc(false, 0); SmallPtrSet VisitedInsts; SmallVector Worklist; Worklist.push_back(Phi); VisitedInsts.insert(Phi); // A value in the reduction can be used: // - By the reduction: // - Reduction operation: // - One use of reduction value (safe). // - Multiple use of reduction value (not safe). // - PHI: // - All uses of the PHI must be the reduction (safe). // - Otherwise, not safe. // - By one instruction outside of the loop (safe). // - By further instructions outside of the loop (not safe). // - By an instruction that is not part of the reduction (not safe). // This is either: // * An instruction type other than PHI or the reduction operation. // * A PHI in the header other than the initial PHI. while (!Worklist.empty()) { Instruction *Cur = Worklist.back(); Worklist.pop_back(); // No Users. // If the instruction has no users then this is a broken chain and can't be // a reduction variable. if (Cur->use_empty()) return false; bool IsAPhi = isa(Cur); // A header PHI use other than the original PHI. if (Cur != Phi && IsAPhi && Cur->getParent() == Phi->getParent()) return false; // Reductions of instructions such as Div, and Sub is only possible if the // LHS is the reduction variable. if (!Cur->isCommutative() && !IsAPhi && !isa(Cur) && !isa(Cur) && !isa(Cur) && !VisitedInsts.count(dyn_cast(Cur->getOperand(0)))) return false; // Any reduction instruction must be of one of the allowed kinds. ReduxDesc = isReductionInstr(Cur, Kind, ReduxDesc); if (!ReduxDesc.IsReduction) return false; // A reduction operation must only have one use of the reduction value. if (!IsAPhi && Kind != RK_IntegerMinMax && Kind != RK_FloatMinMax && hasMultipleUsesOf(Cur, VisitedInsts)) return false; // All inputs to a PHI node must be a reduction value. if(IsAPhi && Cur != Phi && !areAllUsesIn(Cur, VisitedInsts)) return false; if (Kind == RK_IntegerMinMax && (isa(Cur) || isa(Cur))) ++NumCmpSelectPatternInst; if (Kind == RK_FloatMinMax && (isa(Cur) || isa(Cur))) ++NumCmpSelectPatternInst; // Check whether we found a reduction operator. FoundReduxOp |= !IsAPhi; // Process users of current instruction. Push non PHI nodes after PHI nodes // onto the stack. This way we are going to have seen all inputs to PHI // nodes once we get to them. SmallVector NonPHIs; SmallVector PHIs; for (Value::use_iterator UI = Cur->use_begin(), E = Cur->use_end(); UI != E; ++UI) { Instruction *Usr = cast(*UI); // Check if we found the exit user. BasicBlock *Parent = Usr->getParent(); if (!TheLoop->contains(Parent)) { // Exit if you find multiple outside users or if the header phi node is // being used. In this case the user uses the value of the previous // iteration, in which case we would loose "VF-1" iterations of the // reduction operation if we vectorize. if (ExitInstruction != 0 || Cur == Phi) return false; // The instruction used by an outside user must be the last instruction // before we feed back to the reduction phi. Otherwise, we loose VF-1 // operations on the value. if (std::find(Phi->op_begin(), Phi->op_end(), Cur) == Phi->op_end()) return false; ExitInstruction = Cur; continue; } // Process instructions only once (termination). Each reduction cycle // value must only be used once, except by phi nodes and min/max // reductions which are represented as a cmp followed by a select. ReductionInstDesc IgnoredVal(false, 0); if (VisitedInsts.insert(Usr)) { if (isa(Usr)) PHIs.push_back(Usr); else NonPHIs.push_back(Usr); } else if (!isa(Usr) && ((!isa(Usr) && !isa(Usr) && !isa(Usr)) || !isMinMaxSelectCmpPattern(Usr, IgnoredVal).IsReduction)) return false; // Remember that we completed the cycle. if (Usr == Phi) FoundStartPHI = true; } Worklist.append(PHIs.begin(), PHIs.end()); Worklist.append(NonPHIs.begin(), NonPHIs.end()); } // This means we have seen one but not the other instruction of the // pattern or more than just a select and cmp. if ((Kind == RK_IntegerMinMax || Kind == RK_FloatMinMax) && NumCmpSelectPatternInst != 2) return false; if (!FoundStartPHI || !FoundReduxOp || !ExitInstruction) return false; // We found a reduction var if we have reached the original phi node and we // only have a single instruction with out-of-loop users. // This instruction is allowed to have out-of-loop users. AllowedExit.insert(ExitInstruction); // Save the description of this reduction variable. ReductionDescriptor RD(RdxStart, ExitInstruction, Kind, ReduxDesc.MinMaxKind); Reductions[Phi] = RD; // We've ended the cycle. This is a reduction variable if we have an // outside user and it has a binary op. return true; } /// Returns true if the instruction is a Select(ICmp(X, Y), X, Y) instruction /// pattern corresponding to a min(X, Y) or max(X, Y). LoopVectorizationLegality::ReductionInstDesc LoopVectorizationLegality::isMinMaxSelectCmpPattern(Instruction *I, ReductionInstDesc &Prev) { assert((isa(I) || isa(I) || isa(I)) && "Expect a select instruction"); Instruction *Cmp = 0; SelectInst *Select = 0; // We must handle the select(cmp()) as a single instruction. Advance to the // select. if ((Cmp = dyn_cast(I)) || (Cmp = dyn_cast(I))) { if (!Cmp->hasOneUse() || !(Select = dyn_cast(*I->use_begin()))) return ReductionInstDesc(false, I); return ReductionInstDesc(Select, Prev.MinMaxKind); } // Only handle single use cases for now. if (!(Select = dyn_cast(I))) return ReductionInstDesc(false, I); if (!(Cmp = dyn_cast(I->getOperand(0))) && !(Cmp = dyn_cast(I->getOperand(0)))) return ReductionInstDesc(false, I); if (!Cmp->hasOneUse()) return ReductionInstDesc(false, I); Value *CmpLeft; Value *CmpRight; // Look for a min/max pattern. if (m_UMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_UIntMin); else if (m_UMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_UIntMax); else if (m_SMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_SIntMax); else if (m_SMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_SIntMin); else if (m_OrdFMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMin); else if (m_OrdFMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMax); else if (m_UnordFMin(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMin); else if (m_UnordFMax(m_Value(CmpLeft), m_Value(CmpRight)).match(Select)) return ReductionInstDesc(Select, MRK_FloatMax); return ReductionInstDesc(false, I); } LoopVectorizationLegality::ReductionInstDesc LoopVectorizationLegality::isReductionInstr(Instruction *I, ReductionKind Kind, ReductionInstDesc &Prev) { bool FP = I->getType()->isFloatingPointTy(); bool FastMath = (FP && I->isCommutative() && I->isAssociative()); switch (I->getOpcode()) { default: return ReductionInstDesc(false, I); case Instruction::PHI: if (FP && (Kind != RK_FloatMult && Kind != RK_FloatAdd && Kind != RK_FloatMinMax)) return ReductionInstDesc(false, I); return ReductionInstDesc(I, Prev.MinMaxKind); case Instruction::Sub: case Instruction::Add: return ReductionInstDesc(Kind == RK_IntegerAdd, I); case Instruction::Mul: return ReductionInstDesc(Kind == RK_IntegerMult, I); case Instruction::And: return ReductionInstDesc(Kind == RK_IntegerAnd, I); case Instruction::Or: return ReductionInstDesc(Kind == RK_IntegerOr, I); case Instruction::Xor: return ReductionInstDesc(Kind == RK_IntegerXor, I); case Instruction::FMul: return ReductionInstDesc(Kind == RK_FloatMult && FastMath, I); case Instruction::FAdd: return ReductionInstDesc(Kind == RK_FloatAdd && FastMath, I); case Instruction::FCmp: case Instruction::ICmp: case Instruction::Select: if (Kind != RK_IntegerMinMax && (!HasFunNoNaNAttr || Kind != RK_FloatMinMax)) return ReductionInstDesc(false, I); return isMinMaxSelectCmpPattern(I, Prev); } } LoopVectorizationLegality::InductionKind LoopVectorizationLegality::isInductionVariable(PHINode *Phi) { Type *PhiTy = Phi->getType(); // We only handle integer and pointer inductions variables. if (!PhiTy->isIntegerTy() && !PhiTy->isPointerTy()) return IK_NoInduction; // Check that the PHI is consecutive. const SCEV *PhiScev = SE->getSCEV(Phi); const SCEVAddRecExpr *AR = dyn_cast(PhiScev); if (!AR) { DEBUG(dbgs() << "LV: PHI is not a poly recurrence.\n"); return IK_NoInduction; } const SCEV *Step = AR->getStepRecurrence(*SE); // Integer inductions need to have a stride of one. if (PhiTy->isIntegerTy()) { if (Step->isOne()) return IK_IntInduction; if (Step->isAllOnesValue()) return IK_ReverseIntInduction; return IK_NoInduction; } // Calculate the pointer stride and check if it is consecutive. const SCEVConstant *C = dyn_cast(Step); if (!C) return IK_NoInduction; assert(PhiTy->isPointerTy() && "The PHI must be a pointer"); uint64_t Size = DL->getTypeAllocSize(PhiTy->getPointerElementType()); if (C->getValue()->equalsInt(Size)) return IK_PtrInduction; else if (C->getValue()->equalsInt(0 - Size)) return IK_ReversePtrInduction; return IK_NoInduction; } bool LoopVectorizationLegality::isInductionVariable(const Value *V) { Value *In0 = const_cast(V); PHINode *PN = dyn_cast_or_null(In0); if (!PN) return false; return Inductions.count(PN); } bool LoopVectorizationLegality::blockNeedsPredication(BasicBlock *BB) { assert(TheLoop->contains(BB) && "Unknown block used"); // Blocks that do not dominate the latch need predication. BasicBlock* Latch = TheLoop->getLoopLatch(); return !DT->dominates(BB, Latch); } bool LoopVectorizationLegality::blockCanBePredicated(BasicBlock *BB, SmallPtrSet& SafePtrs) { for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { // We might be able to hoist the load. if (it->mayReadFromMemory()) { LoadInst *LI = dyn_cast(it); if (!LI || !SafePtrs.count(LI->getPointerOperand())) return false; } // We don't predicate stores at the moment. if (it->mayWriteToMemory() || it->mayThrow()) return false; // Check that we don't have a constant expression that can trap as operand. for (Instruction::op_iterator OI = it->op_begin(), OE = it->op_end(); OI != OE; ++OI) { if (Constant *C = dyn_cast(*OI)) if (C->canTrap()) return false; } // The instructions below can trap. switch (it->getOpcode()) { default: continue; case Instruction::UDiv: case Instruction::SDiv: case Instruction::URem: case Instruction::SRem: return false; } } return true; } LoopVectorizationCostModel::VectorizationFactor LoopVectorizationCostModel::selectVectorizationFactor(bool OptForSize, unsigned UserVF) { // Width 1 means no vectorize VectorizationFactor Factor = { 1U, 0U }; if (OptForSize && Legal->getRuntimePointerCheck()->Need) { DEBUG(dbgs() << "LV: Aborting. Runtime ptr check is required in Os.\n"); return Factor; } // Find the trip count. unsigned TC = SE->getSmallConstantTripCount(TheLoop, TheLoop->getLoopLatch()); DEBUG(dbgs() << "LV: Found trip count: " << TC << '\n'); unsigned WidestType = getWidestType(); unsigned WidestRegister = TTI.getRegisterBitWidth(true); unsigned MaxSafeDepDist = -1U; if (Legal->getMaxSafeDepDistBytes() != -1U) MaxSafeDepDist = Legal->getMaxSafeDepDistBytes() * 8; WidestRegister = ((WidestRegister < MaxSafeDepDist) ? WidestRegister : MaxSafeDepDist); unsigned MaxVectorSize = WidestRegister / WidestType; DEBUG(dbgs() << "LV: The Widest type: " << WidestType << " bits.\n"); DEBUG(dbgs() << "LV: The Widest register is: " << WidestRegister << " bits.\n"); if (MaxVectorSize == 0) { DEBUG(dbgs() << "LV: The target has no vector registers.\n"); MaxVectorSize = 1; } assert(MaxVectorSize <= 32 && "Did not expect to pack so many elements" " into one vector!"); unsigned VF = MaxVectorSize; // If we optimize the program for size, avoid creating the tail loop. if (OptForSize) { // If we are unable to calculate the trip count then don't try to vectorize. if (TC < 2) { DEBUG(dbgs() << "LV: Aborting. A tail loop is required in Os.\n"); return Factor; } // Find the maximum SIMD width that can fit within the trip count. VF = TC % MaxVectorSize; if (VF == 0) VF = MaxVectorSize; // If the trip count that we found modulo the vectorization factor is not // zero then we require a tail. if (VF < 2) { DEBUG(dbgs() << "LV: Aborting. A tail loop is required in Os.\n"); return Factor; } } if (UserVF != 0) { assert(isPowerOf2_32(UserVF) && "VF needs to be a power of two"); DEBUG(dbgs() << "LV: Using user VF " << UserVF << ".\n"); Factor.Width = UserVF; return Factor; } float Cost = expectedCost(1); unsigned Width = 1; DEBUG(dbgs() << "LV: Scalar loop costs: " << (int)Cost << ".\n"); for (unsigned i=2; i <= VF; i*=2) { // Notice that the vector loop needs to be executed less times, so // we need to divide the cost of the vector loops by the width of // the vector elements. float VectorCost = expectedCost(i) / (float)i; DEBUG(dbgs() << "LV: Vector loop of width " << i << " costs: " << (int)VectorCost << ".\n"); if (VectorCost < Cost) { Cost = VectorCost; Width = i; } } DEBUG(dbgs() << "LV: Selecting VF = : "<< Width << ".\n"); Factor.Width = Width; Factor.Cost = Width * Cost; return Factor; } unsigned LoopVectorizationCostModel::getWidestType() { unsigned MaxWidth = 8; // For each block. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { BasicBlock *BB = *bb; // For each instruction in the loop. for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { Type *T = it->getType(); // Only examine Loads, Stores and PHINodes. if (!isa(it) && !isa(it) && !isa(it)) continue; // Examine PHI nodes that are reduction variables. if (PHINode *PN = dyn_cast(it)) if (!Legal->getReductionVars()->count(PN)) continue; // Examine the stored values. if (StoreInst *ST = dyn_cast(it)) T = ST->getValueOperand()->getType(); // Ignore loaded pointer types and stored pointer types that are not // consecutive. However, we do want to take consecutive stores/loads of // pointer vectors into account. if (T->isPointerTy() && !isConsecutiveLoadOrStore(it)) continue; MaxWidth = std::max(MaxWidth, (unsigned)DL->getTypeSizeInBits(T->getScalarType())); } } return MaxWidth; } unsigned LoopVectorizationCostModel::selectUnrollFactor(bool OptForSize, unsigned UserUF, unsigned VF, unsigned LoopCost) { // -- The unroll heuristics -- // We unroll the loop in order to expose ILP and reduce the loop overhead. // There are many micro-architectural considerations that we can't predict // at this level. For example frontend pressure (on decode or fetch) due to // code size, or the number and capabilities of the execution ports. // // We use the following heuristics to select the unroll factor: // 1. If the code has reductions the we unroll in order to break the cross // iteration dependency. // 2. If the loop is really small then we unroll in order to reduce the loop // overhead. // 3. We don't unroll if we think that we will spill registers to memory due // to the increased register pressure. // Use the user preference, unless 'auto' is selected. if (UserUF != 0) return UserUF; // When we optimize for size we don't unroll. if (OptForSize) return 1; // We used the distance for the unroll factor. if (Legal->getMaxSafeDepDistBytes() != -1U) return 1; // Do not unroll loops with a relatively small trip count. unsigned TC = SE->getSmallConstantTripCount(TheLoop, TheLoop->getLoopLatch()); if (TC > 1 && TC < TinyTripCountUnrollThreshold) return 1; unsigned TargetVectorRegisters = TTI.getNumberOfRegisters(true); DEBUG(dbgs() << "LV: The target has " << TargetVectorRegisters << " vector registers\n"); LoopVectorizationCostModel::RegisterUsage R = calculateRegisterUsage(); // We divide by these constants so assume that we have at least one // instruction that uses at least one register. R.MaxLocalUsers = std::max(R.MaxLocalUsers, 1U); R.NumInstructions = std::max(R.NumInstructions, 1U); // We calculate the unroll factor using the following formula. // Subtract the number of loop invariants from the number of available // registers. These registers are used by all of the unrolled instances. // Next, divide the remaining registers by the number of registers that is // required by the loop, in order to estimate how many parallel instances // fit without causing spills. unsigned UF = (TargetVectorRegisters - R.LoopInvariantRegs) / R.MaxLocalUsers; // Clamp the unroll factor ranges to reasonable factors. unsigned MaxUnrollSize = TTI.getMaximumUnrollFactor(); // If we did not calculate the cost for VF (because the user selected the VF) // then we calculate the cost of VF here. if (LoopCost == 0) LoopCost = expectedCost(VF); // Clamp the calculated UF to be between the 1 and the max unroll factor // that the target allows. if (UF > MaxUnrollSize) UF = MaxUnrollSize; else if (UF < 1) UF = 1; bool HasReductions = Legal->getReductionVars()->size(); // Decide if we want to unroll if we decided that it is legal to vectorize // but not profitable. if (VF == 1) { if (TheLoop->getNumBlocks() > 1 || !HasReductions || LoopCost > SmallLoopCost) return 1; return UF; } if (HasReductions) { DEBUG(dbgs() << "LV: Unrolling because of reductions.\n"); return UF; } // We want to unroll tiny loops in order to reduce the loop overhead. // We assume that the cost overhead is 1 and we use the cost model // to estimate the cost of the loop and unroll until the cost of the // loop overhead is about 5% of the cost of the loop. DEBUG(dbgs() << "LV: Loop cost is " << LoopCost << '\n'); if (LoopCost < SmallLoopCost) { DEBUG(dbgs() << "LV: Unrolling to reduce branch cost.\n"); unsigned NewUF = SmallLoopCost / (LoopCost + 1); return std::min(NewUF, UF); } DEBUG(dbgs() << "LV: Not Unrolling.\n"); return 1; } LoopVectorizationCostModel::RegisterUsage LoopVectorizationCostModel::calculateRegisterUsage() { // This function calculates the register usage by measuring the highest number // of values that are alive at a single location. Obviously, this is a very // rough estimation. We scan the loop in a topological order in order and // assign a number to each instruction. We use RPO to ensure that defs are // met before their users. We assume that each instruction that has in-loop // users starts an interval. We record every time that an in-loop value is // used, so we have a list of the first and last occurrences of each // instruction. Next, we transpose this data structure into a multi map that // holds the list of intervals that *end* at a specific location. This multi // map allows us to perform a linear search. We scan the instructions linearly // and record each time that a new interval starts, by placing it in a set. // If we find this value in the multi-map then we remove it from the set. // The max register usage is the maximum size of the set. // We also search for instructions that are defined outside the loop, but are // used inside the loop. We need this number separately from the max-interval // usage number because when we unroll, loop-invariant values do not take // more register. LoopBlocksDFS DFS(TheLoop); DFS.perform(LI); RegisterUsage R; R.NumInstructions = 0; // Each 'key' in the map opens a new interval. The values // of the map are the index of the 'last seen' usage of the // instruction that is the key. typedef DenseMap IntervalMap; // Maps instruction to its index. DenseMap IdxToInstr; // Marks the end of each interval. IntervalMap EndPoint; // Saves the list of instruction indices that are used in the loop. SmallSet Ends; // Saves the list of values that are used in the loop but are // defined outside the loop, such as arguments and constants. SmallPtrSet LoopInvariants; unsigned Index = 0; for (LoopBlocksDFS::RPOIterator bb = DFS.beginRPO(), be = DFS.endRPO(); bb != be; ++bb) { R.NumInstructions += (*bb)->size(); for (BasicBlock::iterator it = (*bb)->begin(), e = (*bb)->end(); it != e; ++it) { Instruction *I = it; IdxToInstr[Index++] = I; // Save the end location of each USE. for (unsigned i = 0; i < I->getNumOperands(); ++i) { Value *U = I->getOperand(i); Instruction *Instr = dyn_cast(U); // Ignore non-instruction values such as arguments, constants, etc. if (!Instr) continue; // If this instruction is outside the loop then record it and continue. if (!TheLoop->contains(Instr)) { LoopInvariants.insert(Instr); continue; } // Overwrite previous end points. EndPoint[Instr] = Index; Ends.insert(Instr); } } } // Saves the list of intervals that end with the index in 'key'. typedef SmallVector InstrList; DenseMap TransposeEnds; // Transpose the EndPoints to a list of values that end at each index. for (IntervalMap::iterator it = EndPoint.begin(), e = EndPoint.end(); it != e; ++it) TransposeEnds[it->second].push_back(it->first); SmallSet OpenIntervals; unsigned MaxUsage = 0; DEBUG(dbgs() << "LV(REG): Calculating max register usage:\n"); for (unsigned int i = 0; i < Index; ++i) { Instruction *I = IdxToInstr[i]; // Ignore instructions that are never used within the loop. if (!Ends.count(I)) continue; // Remove all of the instructions that end at this location. InstrList &List = TransposeEnds[i]; for (unsigned int j=0, e = List.size(); j < e; ++j) OpenIntervals.erase(List[j]); // Count the number of live interals. MaxUsage = std::max(MaxUsage, OpenIntervals.size()); DEBUG(dbgs() << "LV(REG): At #" << i << " Interval # " << OpenIntervals.size() << '\n'); // Add the current instruction to the list of open intervals. OpenIntervals.insert(I); } unsigned Invariant = LoopInvariants.size(); DEBUG(dbgs() << "LV(REG): Found max usage: " << MaxUsage << '\n'); DEBUG(dbgs() << "LV(REG): Found invariant usage: " << Invariant << '\n'); DEBUG(dbgs() << "LV(REG): LoopSize: " << R.NumInstructions << '\n'); R.LoopInvariantRegs = Invariant; R.MaxLocalUsers = MaxUsage; return R; } unsigned LoopVectorizationCostModel::expectedCost(unsigned VF) { unsigned Cost = 0; // For each block. for (Loop::block_iterator bb = TheLoop->block_begin(), be = TheLoop->block_end(); bb != be; ++bb) { unsigned BlockCost = 0; BasicBlock *BB = *bb; // For each instruction in the old loop. for (BasicBlock::iterator it = BB->begin(), e = BB->end(); it != e; ++it) { // Skip dbg intrinsics. if (isa(it)) continue; unsigned C = getInstructionCost(it, VF); BlockCost += C; DEBUG(dbgs() << "LV: Found an estimated cost of " << C << " for VF " << VF << " For instruction: " << *it << '\n'); } // We assume that if-converted blocks have a 50% chance of being executed. // When the code is scalar then some of the blocks are avoided due to CF. // When the code is vectorized we execute all code paths. if (VF == 1 && Legal->blockNeedsPredication(*bb)) BlockCost /= 2; Cost += BlockCost; } return Cost; } /// \brief Check whether the address computation for a non-consecutive memory /// access looks like an unlikely candidate for being merged into the indexing /// mode. /// /// We look for a GEP which has one index that is an induction variable and all /// other indices are loop invariant. If the stride of this access is also /// within a small bound we decide that this address computation can likely be /// merged into the addressing mode. /// In all other cases, we identify the address computation as complex. static bool isLikelyComplexAddressComputation(Value *Ptr, LoopVectorizationLegality *Legal, ScalarEvolution *SE, const Loop *TheLoop) { GetElementPtrInst *Gep = dyn_cast(Ptr); if (!Gep) return true; // We are looking for a gep with all loop invariant indices except for one // which should be an induction variable. unsigned NumOperands = Gep->getNumOperands(); for (unsigned i = 1; i < NumOperands; ++i) { Value *Opd = Gep->getOperand(i); if (!SE->isLoopInvariant(SE->getSCEV(Opd), TheLoop) && !Legal->isInductionVariable(Opd)) return true; } // Now we know we have a GEP ptr, %inv, %ind, %inv. Make sure that the step // can likely be merged into the address computation. unsigned MaxMergeDistance = 64; const SCEVAddRecExpr *AddRec = dyn_cast(SE->getSCEV(Ptr)); if (!AddRec) return true; // Check the step is constant. const SCEV *Step = AddRec->getStepRecurrence(*SE); // Calculate the pointer stride and check if it is consecutive. const SCEVConstant *C = dyn_cast(Step); if (!C) return true; const APInt &APStepVal = C->getValue()->getValue(); // Huge step value - give up. if (APStepVal.getBitWidth() > 64) return true; int64_t StepVal = APStepVal.getSExtValue(); return StepVal > MaxMergeDistance; } unsigned LoopVectorizationCostModel::getInstructionCost(Instruction *I, unsigned VF) { // If we know that this instruction will remain uniform, check the cost of // the scalar version. if (Legal->isUniformAfterVectorization(I)) VF = 1; Type *RetTy = I->getType(); Type *VectorTy = ToVectorTy(RetTy, VF); // TODO: We need to estimate the cost of intrinsic calls. switch (I->getOpcode()) { case Instruction::GetElementPtr: // We mark this instruction as zero-cost because the cost of GEPs in // vectorized code depends on whether the corresponding memory instruction // is scalarized or not. Therefore, we handle GEPs with the memory // instruction cost. return 0; case Instruction::Br: { return TTI.getCFInstrCost(I->getOpcode()); } case Instruction::PHI: //TODO: IF-converted IFs become selects. return 0; case Instruction::Add: case Instruction::FAdd: case Instruction::Sub: case Instruction::FSub: case Instruction::Mul: case Instruction::FMul: case Instruction::UDiv: case Instruction::SDiv: case Instruction::FDiv: case Instruction::URem: case Instruction::SRem: case Instruction::FRem: case Instruction::Shl: case Instruction::LShr: case Instruction::AShr: case Instruction::And: case Instruction::Or: case Instruction::Xor: { // Certain instructions can be cheaper to vectorize if they have a constant // second vector operand. One example of this are shifts on x86. TargetTransformInfo::OperandValueKind Op1VK = TargetTransformInfo::OK_AnyValue; TargetTransformInfo::OperandValueKind Op2VK = TargetTransformInfo::OK_AnyValue; if (isa(I->getOperand(1))) Op2VK = TargetTransformInfo::OK_UniformConstantValue; return TTI.getArithmeticInstrCost(I->getOpcode(), VectorTy, Op1VK, Op2VK); } case Instruction::Select: { SelectInst *SI = cast(I); const SCEV *CondSCEV = SE->getSCEV(SI->getCondition()); bool ScalarCond = (SE->isLoopInvariant(CondSCEV, TheLoop)); Type *CondTy = SI->getCondition()->getType(); if (!ScalarCond) CondTy = VectorType::get(CondTy, VF); return TTI.getCmpSelInstrCost(I->getOpcode(), VectorTy, CondTy); } case Instruction::ICmp: case Instruction::FCmp: { Type *ValTy = I->getOperand(0)->getType(); VectorTy = ToVectorTy(ValTy, VF); return TTI.getCmpSelInstrCost(I->getOpcode(), VectorTy); } case Instruction::Store: case Instruction::Load: { StoreInst *SI = dyn_cast(I); LoadInst *LI = dyn_cast(I); Type *ValTy = (SI ? SI->getValueOperand()->getType() : LI->getType()); VectorTy = ToVectorTy(ValTy, VF); unsigned Alignment = SI ? SI->getAlignment() : LI->getAlignment(); unsigned AS = SI ? SI->getPointerAddressSpace() : LI->getPointerAddressSpace(); Value *Ptr = SI ? SI->getPointerOperand() : LI->getPointerOperand(); // We add the cost of address computation here instead of with the gep // instruction because only here we know whether the operation is // scalarized. if (VF == 1) return TTI.getAddressComputationCost(VectorTy) + TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS); // Scalarized loads/stores. int ConsecutiveStride = Legal->isConsecutivePtr(Ptr); bool Reverse = ConsecutiveStride < 0; unsigned ScalarAllocatedSize = DL->getTypeAllocSize(ValTy); unsigned VectorElementSize = DL->getTypeStoreSize(VectorTy)/VF; if (!ConsecutiveStride || ScalarAllocatedSize != VectorElementSize) { bool IsComplexComputation = isLikelyComplexAddressComputation(Ptr, Legal, SE, TheLoop); unsigned Cost = 0; // The cost of extracting from the value vector and pointer vector. Type *PtrTy = ToVectorTy(Ptr->getType(), VF); for (unsigned i = 0; i < VF; ++i) { // The cost of extracting the pointer operand. Cost += TTI.getVectorInstrCost(Instruction::ExtractElement, PtrTy, i); // In case of STORE, the cost of ExtractElement from the vector. // In case of LOAD, the cost of InsertElement into the returned // vector. Cost += TTI.getVectorInstrCost(SI ? Instruction::ExtractElement : Instruction::InsertElement, VectorTy, i); } // The cost of the scalar loads/stores. Cost += VF * TTI.getAddressComputationCost(PtrTy, IsComplexComputation); Cost += VF * TTI.getMemoryOpCost(I->getOpcode(), ValTy->getScalarType(), Alignment, AS); return Cost; } // Wide load/stores. unsigned Cost = TTI.getAddressComputationCost(VectorTy); Cost += TTI.getMemoryOpCost(I->getOpcode(), VectorTy, Alignment, AS); if (Reverse) Cost += TTI.getShuffleCost(TargetTransformInfo::SK_Reverse, VectorTy, 0); return Cost; } case Instruction::ZExt: case Instruction::SExt: case Instruction::FPToUI: case Instruction::FPToSI: case Instruction::FPExt: case Instruction::PtrToInt: case Instruction::IntToPtr: case Instruction::SIToFP: case Instruction::UIToFP: case Instruction::Trunc: case Instruction::FPTrunc: case Instruction::BitCast: { // We optimize the truncation of induction variable. // The cost of these is the same as the scalar operation. if (I->getOpcode() == Instruction::Trunc && Legal->isInductionVariable(I->getOperand(0))) return TTI.getCastInstrCost(I->getOpcode(), I->getType(), I->getOperand(0)->getType()); Type *SrcVecTy = ToVectorTy(I->getOperand(0)->getType(), VF); return TTI.getCastInstrCost(I->getOpcode(), VectorTy, SrcVecTy); } case Instruction::Call: { CallInst *CI = cast(I); Intrinsic::ID ID = getIntrinsicIDForCall(CI, TLI); assert(ID && "Not an intrinsic call!"); Type *RetTy = ToVectorTy(CI->getType(), VF); SmallVector Tys; for (unsigned i = 0, ie = CI->getNumArgOperands(); i != ie; ++i) Tys.push_back(ToVectorTy(CI->getArgOperand(i)->getType(), VF)); return TTI.getIntrinsicInstrCost(ID, RetTy, Tys); } default: { // We are scalarizing the instruction. Return the cost of the scalar // instruction, plus the cost of insert and extract into vector // elements, times the vector width. unsigned Cost = 0; if (!RetTy->isVoidTy() && VF != 1) { unsigned InsCost = TTI.getVectorInstrCost(Instruction::InsertElement, VectorTy); unsigned ExtCost = TTI.getVectorInstrCost(Instruction::ExtractElement, VectorTy); // The cost of inserting the results plus extracting each one of the // operands. Cost += VF * (InsCost + ExtCost * I->getNumOperands()); } // The cost of executing VF copies of the scalar instruction. This opcode // is unknown. Assume that it is the same as 'mul'. Cost += VF * TTI.getArithmeticInstrCost(Instruction::Mul, VectorTy); return Cost; } }// end of switch. } Type* LoopVectorizationCostModel::ToVectorTy(Type *Scalar, unsigned VF) { if (Scalar->isVoidTy() || VF == 1) return Scalar; return VectorType::get(Scalar, VF); } char LoopVectorize::ID = 0; static const char lv_name[] = "Loop Vectorization"; INITIALIZE_PASS_BEGIN(LoopVectorize, LV_NAME, lv_name, false, false) INITIALIZE_AG_DEPENDENCY(TargetTransformInfo) INITIALIZE_PASS_DEPENDENCY(DominatorTree) INITIALIZE_PASS_DEPENDENCY(ScalarEvolution) INITIALIZE_PASS_DEPENDENCY(LCSSA) INITIALIZE_PASS_DEPENDENCY(LoopInfo) INITIALIZE_PASS_DEPENDENCY(LoopSimplify) INITIALIZE_PASS_END(LoopVectorize, LV_NAME, lv_name, false, false) namespace llvm { Pass *createLoopVectorizePass(bool NoUnrolling) { return new LoopVectorize(NoUnrolling); } } bool LoopVectorizationCostModel::isConsecutiveLoadOrStore(Instruction *Inst) { // Check for a store. if (StoreInst *ST = dyn_cast(Inst)) return Legal->isConsecutivePtr(ST->getPointerOperand()) != 0; // Check for a load. if (LoadInst *LI = dyn_cast(Inst)) return Legal->isConsecutivePtr(LI->getPointerOperand()) != 0; return false; } void InnerLoopUnroller::scalarizeInstruction(Instruction *Instr) { assert(!Instr->getType()->isAggregateType() && "Can't handle vectors"); // Holds vector parameters or scalars, in case of uniform vals. SmallVector Params; setDebugLocFromInst(Builder, Instr); // Find all of the vectorized parameters. for (unsigned op = 0, e = Instr->getNumOperands(); op != e; ++op) { Value *SrcOp = Instr->getOperand(op); // If we are accessing the old induction variable, use the new one. if (SrcOp == OldInduction) { Params.push_back(getVectorValue(SrcOp)); continue; } // Try using previously calculated values. Instruction *SrcInst = dyn_cast(SrcOp); // If the src is an instruction that appeared earlier in the basic block // then it should already be vectorized. if (SrcInst && OrigLoop->contains(SrcInst)) { assert(WidenMap.has(SrcInst) && "Source operand is unavailable"); // The parameter is a vector value from earlier. Params.push_back(WidenMap.get(SrcInst)); } else { // The parameter is a scalar from outside the loop. Maybe even a constant. VectorParts Scalars; Scalars.append(UF, SrcOp); Params.push_back(Scalars); } } assert(Params.size() == Instr->getNumOperands() && "Invalid number of operands"); // Does this instruction return a value ? bool IsVoidRetTy = Instr->getType()->isVoidTy(); Value *UndefVec = IsVoidRetTy ? 0 : UndefValue::get(Instr->getType()); // Create a new entry in the WidenMap and initialize it to Undef or Null. VectorParts &VecResults = WidenMap.splat(Instr, UndefVec); // For each vector unroll 'part': for (unsigned Part = 0; Part < UF; ++Part) { // For each scalar that we create: Instruction *Cloned = Instr->clone(); if (!IsVoidRetTy) Cloned->setName(Instr->getName() + ".cloned"); // Replace the operands of the cloned instructions with extracted scalars. for (unsigned op = 0, e = Instr->getNumOperands(); op != e; ++op) { Value *Op = Params[op][Part]; Cloned->setOperand(op, Op); } // Place the cloned scalar in the new loop. Builder.Insert(Cloned); // If the original scalar returns a value we need to place it in a vector // so that future users will be able to use it. if (!IsVoidRetTy) VecResults[Part] = Cloned; } } void InnerLoopUnroller::vectorizeMemoryInstruction(Instruction *Instr, LoopVectorizationLegality*) { return scalarizeInstruction(Instr); } Value *InnerLoopUnroller::reverseVector(Value *Vec) { return Vec; } Value *InnerLoopUnroller::getBroadcastInstrs(Value *V) { return V; } Value *InnerLoopUnroller::getConsecutiveVector(Value* Val, int StartIdx, bool Negate) { // When unrolling and the VF is 1, we only need to add a simple scalar. Type *ITy = Val->getType(); assert(!ITy->isVectorTy() && "Val must be a scalar"); Constant *C = ConstantInt::get(ITy, StartIdx, Negate); return Builder.CreateAdd(Val, C, "induction"); } Index: projects/building-blocks/contrib/llvm/patches/patch-r275633-llvm-r223171-fix-vectorizer.diff =================================================================== --- projects/building-blocks/contrib/llvm/patches/patch-r275633-llvm-r223171-fix-vectorizer.diff (nonexistent) +++ projects/building-blocks/contrib/llvm/patches/patch-r275633-llvm-r223171-fix-vectorizer.diff (revision 275664) @@ -0,0 +1,71 @@ +Pull in r223171 from upstream llvm trunk (by Michael Zolotukhin): + + PR21302. Vectorize only bottom-tested loops. + + rdar://problem/18886083 + +This fixes a bug in the llvm vectorizer, which could sometimes cause +vectorized loops to perform an additional iteration, leading to possible +buffer overruns. Symptoms of this, which are usually segfaults, were +first noticed when building gcc ports, here: + +https://lists.freebsd.org/pipermail/freebsd-ports/2014-September/095466.html +https://lists.freebsd.org/pipermail/freebsd-toolchain/2014-September/001211.html + +Introduced here: http://svnweb.freebsd.org/changeset/base/275633 + +Index: lib/Transforms/Vectorize/LoopVectorize.cpp +=================================================================== +--- lib/Transforms/Vectorize/LoopVectorize.cpp (revision 21) ++++ lib/Transforms/Vectorize/LoopVectorize.cpp (revision 22) +@@ -2864,6 +2864,14 @@ bool LoopVectorizationLegality::canVectorize() { + if (!TheLoop->getExitingBlock()) + return false; + ++ // We only handle bottom-tested loops, i.e. loop in which the condition is ++ // checked at the end of each iteration. With that we can assume that all ++ // instructions in the loop are executed the same number of times. ++ if (TheLoop->getExitingBlock() != TheLoop->getLoopLatch()) { ++ DEBUG(dbgs() << "LV: loop control flow is not understood by vectorizer\n"); ++ return false; ++ } ++ + // We need to have a loop header. + DEBUG(dbgs() << "LV: Found a loop: " << + TheLoop->getHeader()->getName() << '\n'); +Index: test/Transforms/LoopVectorize/loop-form.ll +=================================================================== +--- test/Transforms/LoopVectorize/loop-form.ll (revision 0) ++++ test/Transforms/LoopVectorize/loop-form.ll (revision 22) +@@ -0,0 +1,31 @@ ++; RUN: opt -S -loop-vectorize < %s | FileCheck %s ++target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" ++ ++; Check that we vectorize only bottom-tested loops. ++; This is a reduced testcase from PR21302. ++; ++; rdar://problem/18886083 ++ ++%struct.X = type { i32, i16 } ++; CHECK-LABEL: @foo( ++; CHECK-NOT: vector.body ++ ++define void @foo(i32 %n) { ++entry: ++ br label %for.cond ++ ++for.cond: ++ %i = phi i32 [ 0, %entry ], [ %inc, %for.body ] ++ %cmp = icmp slt i32 %i, %n ++ br i1 %cmp, label %for.body, label %if.end ++ ++for.body: ++ %iprom = sext i32 %i to i64 ++ %b = getelementptr inbounds %struct.X* undef, i64 %iprom, i32 1 ++ store i16 0, i16* %b, align 4 ++ %inc = add nsw i32 %i, 1 ++ br label %for.cond ++ ++if.end: ++ ret void ++} Index: projects/building-blocks/contrib/llvm =================================================================== --- projects/building-blocks/contrib/llvm (revision 275663) +++ projects/building-blocks/contrib/llvm (revision 275664) Property changes on: projects/building-blocks/contrib/llvm ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/contrib/llvm:r274984-275663 Index: projects/building-blocks/gnu/usr.bin/gdb/kgdb/kthr.c =================================================================== --- projects/building-blocks/gnu/usr.bin/gdb/kgdb/kthr.c (revision 275663) +++ projects/building-blocks/gnu/usr.bin/gdb/kgdb/kthr.c (revision 275664) @@ -1,252 +1,252 @@ /* * Copyright (c) 2004 Marcel Moolenaar * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "kgdb.h" static CORE_ADDR dumppcb; static int dumptid; static cpuset_t stopped_cpus; static struct kthr *first; struct kthr *curkthr; CORE_ADDR kgdb_lookup(const char *sym) { CORE_ADDR addr; char *name; asprintf(&name, "&%s", sym); addr = kgdb_parse(name); free(name); return (addr); } struct kthr * kgdb_thr_first(void) { return (first); } static void kgdb_thr_add_procs(uintptr_t paddr) { struct proc p; struct thread td; struct kthr *kt; CORE_ADDR addr; while (paddr != 0) { if (kvm_read(kvm, paddr, &p, sizeof(p)) != sizeof(p)) { warnx("kvm_read: %s", kvm_geterr(kvm)); break; } addr = (uintptr_t)TAILQ_FIRST(&p.p_threads); while (addr != 0) { if (kvm_read(kvm, addr, &td, sizeof(td)) != sizeof(td)) { warnx("kvm_read: %s", kvm_geterr(kvm)); break; } kt = malloc(sizeof(*kt)); kt->next = first; kt->kaddr = addr; if (td.td_tid == dumptid) kt->pcb = dumppcb; - else if (td.td_state == TDS_RUNNING && + else if (td.td_oncpu != NOCPU && CPU_ISSET(td.td_oncpu, &stopped_cpus)) kt->pcb = kgdb_trgt_core_pcb(td.td_oncpu); else kt->pcb = (uintptr_t)td.td_pcb; kt->kstack = td.td_kstack; kt->tid = td.td_tid; kt->pid = p.p_pid; kt->paddr = paddr; kt->cpu = td.td_oncpu; first = kt; addr = (uintptr_t)TAILQ_NEXT(&td, td_plist); } paddr = (uintptr_t)LIST_NEXT(&p, p_list); } } struct kthr * kgdb_thr_init(void) { long cpusetsize; struct kthr *kt; CORE_ADDR addr; uintptr_t paddr; while (first != NULL) { kt = first; first = kt->next; free(kt); } addr = kgdb_lookup("allproc"); if (addr == 0) return (NULL); kvm_read(kvm, addr, &paddr, sizeof(paddr)); dumppcb = kgdb_lookup("dumppcb"); if (dumppcb == 0) return (NULL); addr = kgdb_lookup("dumptid"); if (addr != 0) kvm_read(kvm, addr, &dumptid, sizeof(dumptid)); else dumptid = -1; addr = kgdb_lookup("stopped_cpus"); CPU_ZERO(&stopped_cpus); cpusetsize = sysconf(_SC_CPUSET_SIZE); if (cpusetsize != -1 && (u_long)cpusetsize <= sizeof(cpuset_t) && addr != 0) kvm_read(kvm, addr, &stopped_cpus, cpusetsize); kgdb_thr_add_procs(paddr); addr = kgdb_lookup("zombproc"); if (addr != 0) { kvm_read(kvm, addr, &paddr, sizeof(paddr)); kgdb_thr_add_procs(paddr); } curkthr = kgdb_thr_lookup_tid(dumptid); if (curkthr == NULL) curkthr = first; return (first); } struct kthr * kgdb_thr_lookup_tid(int tid) { struct kthr *kt; kt = first; while (kt != NULL && kt->tid != tid) kt = kt->next; return (kt); } struct kthr * kgdb_thr_lookup_taddr(uintptr_t taddr) { struct kthr *kt; kt = first; while (kt != NULL && kt->kaddr != taddr) kt = kt->next; return (kt); } struct kthr * kgdb_thr_lookup_pid(int pid) { struct kthr *kt; kt = first; while (kt != NULL && kt->pid != pid) kt = kt->next; return (kt); } struct kthr * kgdb_thr_lookup_paddr(uintptr_t paddr) { struct kthr *kt; kt = first; while (kt != NULL && kt->paddr != paddr) kt = kt->next; return (kt); } struct kthr * kgdb_thr_next(struct kthr *kt) { return (kt->next); } struct kthr * kgdb_thr_select(struct kthr *kt) { struct kthr *pcur; pcur = curkthr; curkthr = kt; return (pcur); } char * kgdb_thr_extra_thread_info(int tid) { char comm[MAXCOMLEN + 1]; char td_name[MAXCOMLEN + 1]; struct kthr *kt; struct proc *p; struct thread *t; static char buf[64]; kt = kgdb_thr_lookup_tid(tid); if (kt == NULL) return (NULL); snprintf(buf, sizeof(buf), "PID=%d", kt->pid); p = (struct proc *)kt->paddr; if (kvm_read(kvm, (uintptr_t)&p->p_comm[0], &comm, sizeof(comm)) != sizeof(comm)) return (buf); strlcat(buf, ": ", sizeof(buf)); strlcat(buf, comm, sizeof(buf)); t = (struct thread *)kt->kaddr; if (kvm_read(kvm, (uintptr_t)&t->td_name[0], &td_name, sizeof(td_name)) == sizeof(td_name) && strcmp(comm, td_name) != 0) { strlcat(buf, "/", sizeof(buf)); strlcat(buf, td_name, sizeof(buf)); } return (buf); } Index: projects/building-blocks/gnu/usr.bin/gdb =================================================================== --- projects/building-blocks/gnu/usr.bin/gdb (revision 275663) +++ projects/building-blocks/gnu/usr.bin/gdb (revision 275664) Property changes on: projects/building-blocks/gnu/usr.bin/gdb ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/gnu/usr.bin/gdb:r275233-275663 Index: projects/building-blocks/lib/msun/src/e_j0f.c =================================================================== --- projects/building-blocks/lib/msun/src/e_j0f.c (revision 275663) +++ projects/building-blocks/lib/msun/src/e_j0f.c (revision 275664) @@ -1,337 +1,337 @@ /* e_j0f.c -- float version of e_j0.c. * Conversion to float by Ian Lance Taylor, Cygnus Support, ian@cygnus.com. */ /* * ==================================================== * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved. * * Developed at SunPro, a Sun Microsystems, Inc. business. * Permission to use, copy, modify, and distribute this * software is freely granted, provided that this notice * is preserved. * ==================================================== */ #include __FBSDID("$FreeBSD$"); #include "math.h" #include "math_private.h" static float pzerof(float), qzerof(float); static const float huge = 1e30, one = 1.0, invsqrtpi= 5.6418961287e-01, /* 0x3f106ebb */ tpi = 6.3661974669e-01, /* 0x3f22f983 */ /* R0/S0 on [0, 2.00] */ R02 = 1.5625000000e-02, /* 0x3c800000 */ R03 = -1.8997929874e-04, /* 0xb947352e */ R04 = 1.8295404516e-06, /* 0x35f58e88 */ R05 = -4.6183270541e-09, /* 0xb19eaf3c */ S01 = 1.5619102865e-02, /* 0x3c7fe744 */ S02 = 1.1692678527e-04, /* 0x38f53697 */ S03 = 5.1354652442e-07, /* 0x3509daa6 */ S04 = 1.1661400734e-09; /* 0x30a045e8 */ static const float zero = 0.0; float __ieee754_j0f(float x) { float z, s,c,ss,cc,r,u,v; int32_t hx,ix; GET_FLOAT_WORD(hx,x); ix = hx&0x7fffffff; if(ix>=0x7f800000) return one/(x*x); x = fabsf(x); if(ix >= 0x40000000) { /* |x| >= 2.0 */ s = sinf(x); c = cosf(x); ss = s-c; cc = s+c; if(ix<0x7f000000) { /* make sure x+x not overflow */ z = -cosf(x+x); if ((s*c)0x54000000) z = (invsqrtpi*cc)/sqrtf(x); else { u = pzerof(x); v = qzerof(x); z = invsqrtpi*(u*cc-v*ss)/sqrtf(x); } return z; } - if(ix<0x3c000000) { /* |x| < 2**-7 */ + if(ix<0x3b000000) { /* |x| < 2**-9 */ if(huge+x>one) { /* raise inexact if x != 0 */ if(ix<0x39800000) return one; /* |x|<2**-12 */ else return one - x*x/4; } } z = x*x; r = z*(R02+z*(R03+z*(R04+z*R05))); s = one+z*(S01+z*(S02+z*(S03+z*S04))); if(ix < 0x3F800000) { /* |x| < 1.00 */ return one + z*((float)-0.25+(r/s)); } else { u = (float)0.5*x; return((one+u)*(one-u)+z*(r/s)); } } static const float u00 = -7.3804296553e-02, /* 0xbd9726b5 */ u01 = 1.7666645348e-01, /* 0x3e34e80d */ u02 = -1.3818567619e-02, /* 0xbc626746 */ u03 = 3.4745343146e-04, /* 0x39b62a69 */ u04 = -3.8140706238e-06, /* 0xb67ff53c */ u05 = 1.9559013964e-08, /* 0x32a802ba */ u06 = -3.9820518410e-11, /* 0xae2f21eb */ v01 = 1.2730483897e-02, /* 0x3c509385 */ v02 = 7.6006865129e-05, /* 0x389f65e0 */ v03 = 2.5915085189e-07, /* 0x348b216c */ v04 = 4.4111031494e-10; /* 0x2ff280c2 */ float __ieee754_y0f(float x) { float z, s,c,ss,cc,u,v; int32_t hx,ix; GET_FLOAT_WORD(hx,x); ix = 0x7fffffff&hx; /* Y0(NaN) is NaN, y0(-inf) is Nan, y0(inf) is 0 */ if(ix>=0x7f800000) return one/(x+x*x); if(ix==0) return -one/zero; if(hx<0) return zero/zero; if(ix >= 0x40000000) { /* |x| >= 2.0 */ /* y0(x) = sqrt(2/(pi*x))*(p0(x)*sin(x0)+q0(x)*cos(x0)) * where x0 = x-pi/4 * Better formula: * cos(x0) = cos(x)cos(pi/4)+sin(x)sin(pi/4) * = 1/sqrt(2) * (sin(x) + cos(x)) * sin(x0) = sin(x)cos(3pi/4)-cos(x)sin(3pi/4) * = 1/sqrt(2) * (sin(x) - cos(x)) * To avoid cancellation, use * sin(x) +- cos(x) = -cos(2x)/(sin(x) -+ cos(x)) * to compute the worse one. */ s = sinf(x); c = cosf(x); ss = s-c; cc = s+c; /* * j0(x) = 1/sqrt(pi) * (P(0,x)*cc - Q(0,x)*ss) / sqrt(x) * y0(x) = 1/sqrt(pi) * (P(0,x)*ss + Q(0,x)*cc) / sqrt(x) */ if(ix<0x7f000000) { /* make sure x+x not overflow */ z = -cosf(x+x); if ((s*c)0x54800000) z = (invsqrtpi*ss)/sqrtf(x); else { u = pzerof(x); v = qzerof(x); z = invsqrtpi*(u*ss+v*cc)/sqrtf(x); } return z; } if(ix<=0x32000000) { /* x < 2**-27 */ return(u00 + tpi*__ieee754_logf(x)); } z = x*x; u = u00+z*(u01+z*(u02+z*(u03+z*(u04+z*(u05+z*u06))))); v = one+z*(v01+z*(v02+z*(v03+z*v04))); return(u/v + tpi*(__ieee754_j0f(x)*__ieee754_logf(x))); } /* The asymptotic expansions of pzero is * 1 - 9/128 s^2 + 11025/98304 s^4 - ..., where s = 1/x. * For x >= 2, We approximate pzero by * pzero(x) = 1 + (R/S) * where R = pR0 + pR1*s^2 + pR2*s^4 + ... + pR5*s^10 * S = 1 + pS0*s^2 + ... + pS4*s^10 * and * | pzero(x)-1-R/S | <= 2 ** ( -60.26) */ static const float pR8[6] = { /* for x in [inf, 8]=1/[0,0.125] */ 0.0000000000e+00, /* 0x00000000 */ -7.0312500000e-02, /* 0xbd900000 */ -8.0816707611e+00, /* 0xc1014e86 */ -2.5706311035e+02, /* 0xc3808814 */ -2.4852163086e+03, /* 0xc51b5376 */ -5.2530439453e+03, /* 0xc5a4285a */ }; static const float pS8[5] = { 1.1653436279e+02, /* 0x42e91198 */ 3.8337448730e+03, /* 0x456f9beb */ 4.0597855469e+04, /* 0x471e95db */ 1.1675296875e+05, /* 0x47e4087c */ 4.7627726562e+04, /* 0x473a0bba */ }; static const float pR5[6] = { /* for x in [8,4.5454]=1/[0.125,0.22001] */ -1.1412546255e-11, /* 0xad48c58a */ -7.0312492549e-02, /* 0xbd8fffff */ -4.1596107483e+00, /* 0xc0851b88 */ -6.7674766541e+01, /* 0xc287597b */ -3.3123129272e+02, /* 0xc3a59d9b */ -3.4643338013e+02, /* 0xc3ad3779 */ }; static const float pS5[5] = { 6.0753936768e+01, /* 0x42730408 */ 1.0512523193e+03, /* 0x44836813 */ 5.9789707031e+03, /* 0x45bad7c4 */ 9.6254453125e+03, /* 0x461665c8 */ 2.4060581055e+03, /* 0x451660ee */ }; static const float pR3[6] = {/* for x in [4.547,2.8571]=1/[0.2199,0.35001] */ -2.5470459075e-09, /* 0xb12f081b */ -7.0311963558e-02, /* 0xbd8fffb8 */ -2.4090321064e+00, /* 0xc01a2d95 */ -2.1965976715e+01, /* 0xc1afba52 */ -5.8079170227e+01, /* 0xc2685112 */ -3.1447946548e+01, /* 0xc1fb9565 */ }; static const float pS3[5] = { 3.5856033325e+01, /* 0x420f6c94 */ 3.6151397705e+02, /* 0x43b4c1ca */ 1.1936077881e+03, /* 0x44953373 */ 1.1279968262e+03, /* 0x448cffe6 */ 1.7358093262e+02, /* 0x432d94b8 */ }; static const float pR2[6] = {/* for x in [2.8570,2]=1/[0.3499,0.5] */ -8.8753431271e-08, /* 0xb3be98b7 */ -7.0303097367e-02, /* 0xbd8ffb12 */ -1.4507384300e+00, /* 0xbfb9b1cc */ -7.6356959343e+00, /* 0xc0f4579f */ -1.1193166733e+01, /* 0xc1331736 */ -3.2336456776e+00, /* 0xc04ef40d */ }; static const float pS2[5] = { 2.2220300674e+01, /* 0x41b1c32d */ 1.3620678711e+02, /* 0x430834f0 */ 2.7047027588e+02, /* 0x43873c32 */ 1.5387539673e+02, /* 0x4319e01a */ 1.4657617569e+01, /* 0x416a859a */ }; static float pzerof(float x) { const float *p,*q; float z,r,s; int32_t ix; GET_FLOAT_WORD(ix,x); ix &= 0x7fffffff; if(ix>=0x41000000) {p = pR8; q= pS8;} else if(ix>=0x40f71c58){p = pR5; q= pS5;} else if(ix>=0x4036db68){p = pR3; q= pS3;} else if(ix>=0x40000000){p = pR2; q= pS2;} z = one/(x*x); r = p[0]+z*(p[1]+z*(p[2]+z*(p[3]+z*(p[4]+z*p[5])))); s = one+z*(q[0]+z*(q[1]+z*(q[2]+z*(q[3]+z*q[4])))); return one+ r/s; } /* For x >= 8, the asymptotic expansions of qzero is * -1/8 s + 75/1024 s^3 - ..., where s = 1/x. * We approximate pzero by * qzero(x) = s*(-1.25 + (R/S)) * where R = qR0 + qR1*s^2 + qR2*s^4 + ... + qR5*s^10 * S = 1 + qS0*s^2 + ... + qS5*s^12 * and * | qzero(x)/s +1.25-R/S | <= 2 ** ( -61.22) */ static const float qR8[6] = { /* for x in [inf, 8]=1/[0,0.125] */ 0.0000000000e+00, /* 0x00000000 */ 7.3242187500e-02, /* 0x3d960000 */ 1.1768206596e+01, /* 0x413c4a93 */ 5.5767340088e+02, /* 0x440b6b19 */ 8.8591972656e+03, /* 0x460a6cca */ 3.7014625000e+04, /* 0x471096a0 */ }; static const float qS8[6] = { 1.6377603149e+02, /* 0x4323c6aa */ 8.0983447266e+03, /* 0x45fd12c2 */ 1.4253829688e+05, /* 0x480b3293 */ 8.0330925000e+05, /* 0x49441ed4 */ 8.4050156250e+05, /* 0x494d3359 */ -3.4389928125e+05, /* 0xc8a7eb69 */ }; static const float qR5[6] = { /* for x in [8,4.5454]=1/[0.125,0.22001] */ 1.8408595828e-11, /* 0x2da1ec79 */ 7.3242180049e-02, /* 0x3d95ffff */ 5.8356351852e+00, /* 0x40babd86 */ 1.3511157227e+02, /* 0x43071c90 */ 1.0272437744e+03, /* 0x448067cd */ 1.9899779053e+03, /* 0x44f8bf4b */ }; static const float qS5[6] = { 8.2776611328e+01, /* 0x42a58da0 */ 2.0778142090e+03, /* 0x4501dd07 */ 1.8847289062e+04, /* 0x46933e94 */ 5.6751113281e+04, /* 0x475daf1d */ 3.5976753906e+04, /* 0x470c88c1 */ -5.3543427734e+03, /* 0xc5a752be */ }; static const float qR3[6] = {/* for x in [4.547,2.8571]=1/[0.2199,0.35001] */ 4.3774099900e-09, /* 0x3196681b */ 7.3241114616e-02, /* 0x3d95ff70 */ 3.3442313671e+00, /* 0x405607e3 */ 4.2621845245e+01, /* 0x422a7cc5 */ 1.7080809021e+02, /* 0x432acedf */ 1.6673394775e+02, /* 0x4326bbe4 */ }; static const float qS3[6] = { 4.8758872986e+01, /* 0x42430916 */ 7.0968920898e+02, /* 0x44316c1c */ 3.7041481934e+03, /* 0x4567825f */ 6.4604252930e+03, /* 0x45c9e367 */ 2.5163337402e+03, /* 0x451d4557 */ -1.4924745178e+02, /* 0xc3153f59 */ }; static const float qR2[6] = {/* for x in [2.8570,2]=1/[0.3499,0.5] */ 1.5044444979e-07, /* 0x342189db */ 7.3223426938e-02, /* 0x3d95f62a */ 1.9981917143e+00, /* 0x3fffc4bf */ 1.4495602608e+01, /* 0x4167edfd */ 3.1666231155e+01, /* 0x41fd5471 */ 1.6252708435e+01, /* 0x4182058c */ }; static const float qS2[6] = { 3.0365585327e+01, /* 0x41f2ecb8 */ 2.6934811401e+02, /* 0x4386ac8f */ 8.4478375244e+02, /* 0x44533229 */ 8.8293585205e+02, /* 0x445cbbe5 */ 2.1266638184e+02, /* 0x4354aa98 */ -5.3109550476e+00, /* 0xc0a9f358 */ }; static float qzerof(float x) { const float *p,*q; float s,r,z; int32_t ix; GET_FLOAT_WORD(ix,x); ix &= 0x7fffffff; if(ix>=0x41000000) {p = qR8; q= qS8;} else if(ix>=0x40f71c58){p = qR5; q= qS5;} else if(ix>=0x4036db68){p = qR3; q= qS3;} else if(ix>=0x40000000){p = qR2; q= qS2;} z = one/(x*x); r = p[0]+z*(p[1]+z*(p[2]+z*(p[3]+z*(p[4]+z*p[5])))); s = one+z*(q[0]+z*(q[1]+z*(q[2]+z*(q[3]+z*(q[4]+z*q[5]))))); return (-(float).125 + r/s)/x; } Index: projects/building-blocks/release/scripts/relnotes-search.sh =================================================================== --- projects/building-blocks/release/scripts/relnotes-search.sh (revision 275663) +++ projects/building-blocks/release/scripts/relnotes-search.sh (revision 275664) @@ -1,133 +1,133 @@ #!/bin/sh #- # Copyright (c) 2014 The FreeBSD Foundation # All rights reserved. # # This software were developed by Glen Barber # under sponsorship from the FreeBSD Foundation. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # 1. Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # 2. Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # # THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE # ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS # OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) # HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT # LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY # OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF # SUCH DAMAGE. # # $FreeBSD$ # set -C PATH="/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin" export PATH usage() { echo "Usage:" echo -n "$(basename ${0}) [-rNNNNNN]" echo " [-l /path/for/output] /path/to/branch" echo " -r: The oldest commit to include in the search" echo "" exit 1 } main() { while getopts "l:r:" arg ; do case ${arg} in l) # Disallow '-rNNNNNN' argument for oldest # revision # from becoming the log file # accidentally. where="${OPTARG##-r*}" [ -z "${where}" ] && usage if [ -e "${where}" ]; then echo "Log file already exists:" echo " (${where})" return 2 fi ;; r) rev="${OPTARG##-r}" c=$(echo -n ${rev} | tr -d '0-9' | wc -c) if [ ${c} -ne 0 ]; then echo "Revision number must be numeric." return 2 fi # Since the last specified revision is # specified, mangle the variable to # make svn syntax happy. rev="-r${rev}:rHEAD" ;; *) usage ;; esac done shift $(( ${OPTIND} - 1 )) # This assumes a local working copy, which svn search # allows exactly one repository path (although the root # can still be the path). [ "$#" -ne 1 ] && usage # If no log file, write to stdout. [ -z "${where}" ] && where=/dev/stdout svn= # Where is svn? for s in /usr/bin /usr/local/bin; do if [ -x ${s}/svn ]; then svn=${s}/svn break fi if [ -x ${s}/svnlite ]; then svn=${s}/svnlite break fi done # Did we find svn? if [ -z "${svn}" ]; then echo "svn(1) binary not found." return 2 fi # Is more than one path specified? (This should never # be triggered, because the argument count is checked # above, but better safe than sorry.) if [ $# -gt 1 ]; then echo "Cannot specify more than one working path." return 2 fi # Does the directory exist? if [ ! -d "${1}" ]; then echo "Specified path (${1}) is not a directory." return 2 fi # Is it a subversion repository checkout? ${svn} info ${1} >/dev/null 2>&1 if [ "$?" -ne 0 ]; then echo "Cannot determine svn repository information for ${1}" return 2 fi # All tests passed. Let's see what can possibly go wrong # from here. The search string specified should match this # in PCRE speak: ':[\t ]*' - ${svn} log ${rev} --search 'Relnotes:?[^ ]*' ${1} > ${where} + ${svn} log ${rev} --search 'Relnotes:*[A-Za-z0-9]*' ${1} > ${where} return $? } main "${@}" exit $? Index: projects/building-blocks/share/man/man4/upgt.4 =================================================================== --- projects/building-blocks/share/man/man4/upgt.4 (revision 275663) +++ projects/building-blocks/share/man/man4/upgt.4 (revision 275664) @@ -1,222 +1,224 @@ .\" $OpenBSD: upgt.4,v 1.6 2008/04/17 14:01:22 jmc Exp $ .\" $FreeBSD$ .\" .\" Copyright (c) 2007 Marcus Glocker .\" Copyright (c) 2005-2007 .\" Damien Bergamini .\" .\" Permission to use, copy, modify, and distribute this software for any .\" purpose with or without fee is hereby granted, provided that the above .\" copyright notice and this permission notice appear in all copies. .\" .\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES .\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF .\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR .\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES .\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN .\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF .\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. .\" .\" .\" .\" Copyright (c) 2006 Theo de Raadt. .\" Copyright (c) 2006 The DragonFly Project. All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" .\" 1. Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" 2. Redistributions in binary form must reproduce the above copyright .\" notice, this list of conditions and the following disclaimer in .\" the documentation and/or other materials provided with the .\" distribution. .\" 3. Neither the name of The DragonFly Project nor the names of its .\" contributors may be used to endorse or promote products derived .\" from this software without specific, prior written permission. .\" .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS .\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT .\" LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS .\" FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE .\" COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, .\" INCIDENTAL, SPECIAL, EXEMPLARY OR CONSEQUENTIAL DAMAGES (INCLUDING, .\" BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; .\" LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED .\" AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, .\" OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT .\" OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF .\" SUCH DAMAGE. .\" .Dd April 17, 2008 .Dt UPGT 4 .Os .Sh NAME .Nm upgt .Nd Conexant/Intersil PrismGT SoftMAC USB IEEE 802.11b/g wireless network device .Sh SYNOPSIS To compile this driver into the kernel, place the following lines in your kernel configuration file: .Bd -ragged -offset indent .Cd "device ehci" .Cd "device uhci" .Cd "device ohci" .Cd "device usb" .Cd "device upgt" .Cd "device wlan" .Ed .Pp Alternatively, to load the driver as a module at boot time, place the following line in .Xr loader.conf 5 : .Bd -literal -offset indent if_upgt_load="YES" .Ed .Sh DESCRIPTION The .Nm driver supports the USB 2.0 Conexant/Intersil PrismGT series wireless adapters based on the GW3887 chipset. .Pp These are the modes the .Nm driver can operate in: .Bl -tag -width "IBSS-masterXX" .It BSS mode Also known as .Em infrastructure mode, this is used when associating with an access point, through which all traffic passes. This mode is the default. .\" .It IBSS mode .\" Also known as .\" .Em IEEE ad-hoc .\" mode or .\" .Em peer-to-peer .\" mode. .\" This is the standardized method of operating without an access point. .\" Stations associate with a service set. .\" However, actual connections between stations are peer-to-peer. .\" .It Host AP .\" In this mode the driver acts as an access point (base station) .\" for other cards. .It monitor mode In this mode the driver is able to receive packets without associating with an access point. This disables the internal receive filter and enables the card to capture packets from networks which it wouldn't normally have access to, or to scan for access points. .El .Pp .Nm supports software WEP. Wired Equivalent Privacy (WEP) is the de facto encryption standard for wireless networks. It can be typically configured in one of three modes: no encryption; 40-bit encryption; or 104-bit encryption. Unfortunately, due to serious weaknesses in WEP protocol it is strongly recommended that it not be used as the sole mechanism to secure wireless communication. WEP is not enabled by default. .\".Pp .\"The transmit speed is user-selectable or can be adapted automatically by the .\"driver depending on the received signal strength and on the number of hardware .\"transmission retries. .Pp The .Nm driver can be configured at runtime with .Xr ifconfig 8 . .Sh FILES .\".Pp .\"These firmware files are not free because Conexant/Intersil refuses .\"to grant distribution rights. .\"As a result, even though .\".Ox .\"includes the driver, the firmware files cannot be included and .\"users have to download these files on their own. This driver requires the .Nm upgtfw firmware to be installed before it will work. The firmware files are not publicly available. A package of the firmware which can be installed via .Xr pkg_add 1 is available: .Bd -literal -offset indent http://weongyo.org/project/upgt/upgt-firmware-2.13.1.0.tar.gz .Ed .Sh HARDWARE The .Nm driver supports USB 2.0 Conexant/Intersil PrismGT series wireless adapters based on the GW3887 chipset, among them: .Pp .Bl -bullet -compact .It Belkin F5D7050 (version 1000) .It Cohiba Proto Board .It D-Link DWL-G120 Cohiba .It FSC Connect2Air E-5400 USB D1700 .It Gigaset USB Adapter 54 .It Inventel UR045G .It +Netgear WG111v1 (rev2) +.It SMC EZ ConnectG SMC2862W-G .It Sagem XG703A .It Spinnaker DUT .It Spinnaker Proto Board .El .Sh EXAMPLES Join an existing BSS network (i.e., connect to an access point): .Bd -literal -offset indent ifconfig wlan create wlandev upgt0 inet 192.168.0.20 \e netmask 0xffffff00 .Ed .Pp Join a specific BSS network with network name .Dq Li my_net : .Pp .Dl "ifconfig wlan create wlandev upgt0 ssid my_net up" .Pp Join a specific BSS network with 64-bit WEP encryption: .Bd -literal -offset indent ifconfig wlan create wlandev upgt0 ssid my_net \e wepmode on wepkey 0x1234567890 weptxkey 1 up .Ed .Sh SEE ALSO .Xr arp 4 , .Xr netintro 4 , .Xr usb 4 , .Xr wlan 4 , .Xr ifconfig 8 .Sh HISTORY The .Nm driver first appeared in .Ox 4.3 . .Sh AUTHORS .An -nosplit The .Nm driver was written by .An Marcus Glocker Aq Mt mglocker@openbsd.org . .Pp The hardware specification was reverse engineered by the people at .Pa http://www.prism54.org . .Sh CAVEATS The .Nm driver just supports the USB 2.0 devices (GW3887 chipset) but not the USB 1.0 devices containing the NET2280, ISL3880, and ISL3886 chipsets. Some further efforts would be necessary to add USB 1.0 support to the driver. Index: projects/building-blocks/share/man/man4 =================================================================== --- projects/building-blocks/share/man/man4 (revision 275663) +++ projects/building-blocks/share/man/man4 (revision 275664) Property changes on: projects/building-blocks/share/man/man4 ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/share/man/man4:r275497-275663 Index: projects/building-blocks/share =================================================================== --- projects/building-blocks/share (revision 275663) +++ projects/building-blocks/share (revision 275664) Property changes on: projects/building-blocks/share ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/share:r275590-275663 Index: projects/building-blocks/sys/arm/altera/socfpga/files.socfpga =================================================================== --- projects/building-blocks/sys/arm/altera/socfpga/files.socfpga (revision 275663) +++ projects/building-blocks/sys/arm/altera/socfpga/files.socfpga (revision 275664) @@ -1,30 +1,31 @@ # $FreeBSD$ kern/kern_clocksource.c standard arm/arm/bus_space_generic.c standard arm/arm/bus_space_asm_generic.S standard arm/arm/cpufunc_asm_armv5.S standard arm/arm/cpufunc_asm_arm10.S standard arm/arm/cpufunc_asm_arm11.S standard arm/arm/cpufunc_asm_armv7.S standard arm/arm/bus_space-v6.c standard arm/arm/gic.c standard arm/arm/mpcore_timer.c standard arm/altera/socfpga/socfpga_common.c standard arm/altera/socfpga/socfpga_machdep.c standard arm/altera/socfpga/socfpga_manager.c standard arm/altera/socfpga/socfpga_rstmgr.c standard arm/altera/socfpga/socfpga_mp.c optional smp dev/dwc/if_dwc.c optional dwc dev/mii/micphy.c optional micphy dev/mmc/host/dwmmc.c optional dwmmc # BERI specific dev/beri/beri_ring.c optional beri_ring dev/beri/beri_mem.c optional beri_mem -dev/beri/virtio/virtio.c optional beri_vtblk +dev/beri/virtio/virtio.c optional beri_vtblk | vtbe dev/beri/virtio/virtio_block.c optional beri_vtblk +dev/beri/virtio/network/if_vtbe.c optional vtbe Index: projects/building-blocks/sys/arm/arm/syscall.c =================================================================== --- projects/building-blocks/sys/arm/arm/syscall.c (revision 275663) +++ projects/building-blocks/sys/arm/arm/syscall.c (revision 275664) @@ -1,196 +1,197 @@ /* $NetBSD: fault.c,v 1.45 2003/11/20 14:44:36 scw Exp $ */ /*- * Copyright 2004 Olivier Houchard * Copyright 2003 Wasabi Systems, Inc. * All rights reserved. * * Written by Steve C. Woodford for Wasabi Systems, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed for the NetBSD Project by * Wasabi Systems, Inc. * 4. The name of Wasabi Systems, Inc. may not be used to endorse * or promote products derived from this software without specific prior * written permission. * * THIS SOFTWARE IS PROVIDED BY WASABI SYSTEMS, INC. ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL WASABI SYSTEMS, INC * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. */ /*- * Copyright (c) 1994-1997 Mark Brinicombe. * Copyright (c) 1994 Brini. * All rights reserved. * * This code is derived from software written for Brini by Mark Brinicombe * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by Brini. * 4. The name of the company nor the name of the author may be used to * endorse or promote products derived from this software without specific * prior written permission. * * THIS SOFTWARE IS PROVIDED BY BRINI ``AS IS'' AND ANY EXPRESS OR IMPLIED * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL BRINI OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * RiscBSD kernel project * * fault.c * * Fault handlers * * Created : 28/11/94 */ #include __FBSDID("$FreeBSD$"); #include #include +#include #include #include #include #include #include #include #include #include #include void swi_handler(struct trapframe *); static __inline void call_trapsignal(struct thread *td, int sig, u_long code) { ksiginfo_t ksi; ksiginfo_init_trap(&ksi); ksi.ksi_signo = sig; ksi.ksi_code = (int)code; trapsignal(td, &ksi); } int cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa) { struct proc *p; register_t *ap; int error; sa->code = td->td_frame->tf_r7; ap = &td->td_frame->tf_r0; if (sa->code == SYS_syscall) { sa->code = *ap++; sa->nap--; } else if (sa->code == SYS___syscall) { sa->code = ap[_QUAD_LOWWORD]; sa->nap -= 2; ap += 2; } p = td->td_proc; if (p->p_sysent->sv_mask) sa->code &= p->p_sysent->sv_mask; if (sa->code >= p->p_sysent->sv_size) sa->callp = &p->p_sysent->sv_table[0]; else sa->callp = &p->p_sysent->sv_table[sa->code]; sa->narg = sa->callp->sy_narg; error = 0; memcpy(sa->args, ap, sa->nap * sizeof(register_t)); if (sa->narg > sa->nap) { error = copyin((void *)td->td_frame->tf_usr_sp, sa->args + sa->nap, (sa->narg - sa->nap) * sizeof(register_t)); } if (error == 0) { td->td_retval[0] = 0; td->td_retval[1] = 0; } return (error); } #include "../../kern/subr_syscall.c" static void syscall(struct thread *td, struct trapframe *frame) { struct syscall_args sa; int error; sa.nap = 4; error = syscallenter(td, &sa); KASSERT(error != 0 || td->td_ar == NULL, ("returning from syscall with td_ar set!")); syscallret(td, error, &sa); } void swi_handler(struct trapframe *frame) { struct thread *td = curthread; td->td_frame = frame; td->td_pticks = 0; /* * Make sure the program counter is correctly aligned so we * don't take an alignment fault trying to read the opcode. * XXX: Fix for Thumb mode */ if (__predict_false(((frame->tf_pc - INSN_SIZE) & 3) != 0)) { call_trapsignal(td, SIGILL, 0); userret(td, frame); return; } /* * Enable interrupts if they were enabled before the exception. * Since all syscalls *should* come from user mode it will always * be safe to enable them, but check anyway. */ if (td->td_md.md_spinlock_count == 0) { if (__predict_true(frame->tf_spsr & PSR_I) == 0) enable_interrupts(PSR_I); if (__predict_true(frame->tf_spsr & PSR_F) == 0) enable_interrupts(PSR_F); } syscall(td, frame); } Index: projects/building-blocks/sys/arm/broadcom/bcm2835/bcm2835_wdog.c =================================================================== --- projects/building-blocks/sys/arm/broadcom/bcm2835/bcm2835_wdog.c (revision 275663) +++ projects/building-blocks/sys/arm/broadcom/bcm2835/bcm2835_wdog.c (revision 275664) @@ -1,209 +1,209 @@ /*- * Copyright (c) 2012 Alexander Rybalko * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define BCM2835_PASWORD 0x5a #define BCM2835_WDOG_RESET 0 #define BCM2835_PASSWORD_MASK 0xff000000 #define BCM2835_PASSWORD_SHIFT 24 #define BCM2835_WDOG_TIME_MASK 0x000fffff #define BCM2835_WDOG_TIME_SHIFT 0 #define READ(_sc, _r) bus_space_read_4((_sc)->bst, (_sc)->bsh, (_r)) #define WRITE(_sc, _r, _v) bus_space_write_4((_sc)->bst, (_sc)->bsh, (_r), (_v)) #define BCM2835_RSTC_WRCFG_CLR 0xffffffcf #define BCM2835_RSTC_WRCFG_SET 0x00000030 #define BCM2835_RSTC_WRCFG_FULL_RESET 0x00000020 #define BCM2835_RSTC_RESET 0x00000102 #define BCM2835_RSTC_REG 0x00 #define BCM2835_RSTS_REG 0x04 #define BCM2835_WDOG_REG 0x08 static struct bcmwd_softc *bcmwd_lsc = NULL; struct bcmwd_softc { device_t dev; struct resource * res; bus_space_tag_t bst; bus_space_handle_t bsh; int wdog_armed; int wdog_period; char wdog_passwd; struct mtx mtx; }; static void bcmwd_watchdog_fn(void *private, u_int cmd, int *error); static int bcmwd_probe(device_t dev) { if (!ofw_bus_status_okay(dev)) return (ENXIO); if (ofw_bus_is_compatible(dev, "broadcom,bcm2835-wdt")) { device_set_desc(dev, "BCM2708/2835 Watchdog"); return (BUS_PROBE_DEFAULT); } return (ENXIO); } static int bcmwd_attach(device_t dev) { struct bcmwd_softc *sc; int rid; if (bcmwd_lsc != NULL) return (ENXIO); sc = device_get_softc(dev); sc->wdog_period = 7; sc->wdog_passwd = BCM2835_PASWORD; sc->wdog_armed = 0; sc->dev = dev; rid = 0; sc->res = bus_alloc_resource_any(dev, SYS_RES_MEMORY, &rid, RF_ACTIVE); if (sc->res == NULL) { device_printf(dev, "could not allocate memory resource\n"); return (ENXIO); } sc->bst = rman_get_bustag(sc->res); sc->bsh = rman_get_bushandle(sc->res); bcmwd_lsc = sc; mtx_init(&sc->mtx, "BCM2835 Watchdog", "bcmwd", MTX_DEF); EVENTHANDLER_REGISTER(watchdog_list, bcmwd_watchdog_fn, sc, 0); return (0); } static void bcmwd_watchdog_fn(void *private, u_int cmd, int *error) { struct bcmwd_softc *sc; uint64_t sec; uint32_t ticks, reg; sc = private; mtx_lock(&sc->mtx); cmd &= WD_INTERVAL; if (cmd > 0) { sec = ((uint64_t)1 << (cmd & WD_INTERVAL)) / 1000000000; - ticks = (sec << 16) & BCM2835_WDOG_TIME_MASK; - if (ticks == 0) { + if (sec == 0 || sec > 15) { /* * Can't arm * disable watchdog as watchdog(9) requires */ device_printf(sc->dev, - "Can't arm, timeout is less than 1 second\n"); + "Can't arm, timeout must be between 1-15 seconds\n"); WRITE(sc, BCM2835_RSTC_REG, (BCM2835_PASWORD << BCM2835_PASSWORD_SHIFT) | BCM2835_RSTC_RESET); mtx_unlock(&sc->mtx); return; } + ticks = (sec << 16) & BCM2835_WDOG_TIME_MASK; reg = (BCM2835_PASWORD << BCM2835_PASSWORD_SHIFT) | ticks; WRITE(sc, BCM2835_WDOG_REG, reg); reg = READ(sc, BCM2835_RSTC_REG); reg &= BCM2835_RSTC_WRCFG_CLR; reg |= BCM2835_RSTC_WRCFG_FULL_RESET; reg |= (BCM2835_PASWORD << BCM2835_PASSWORD_SHIFT); WRITE(sc, BCM2835_RSTC_REG, reg); *error = 0; } else WRITE(sc, BCM2835_RSTC_REG, (BCM2835_PASWORD << BCM2835_PASSWORD_SHIFT) | BCM2835_RSTC_RESET); mtx_unlock(&sc->mtx); } void bcmwd_watchdog_reset() { if (bcmwd_lsc == NULL) return; WRITE(bcmwd_lsc, BCM2835_WDOG_REG, (BCM2835_PASWORD << BCM2835_PASSWORD_SHIFT) | 10); WRITE(bcmwd_lsc, BCM2835_RSTC_REG, (READ(bcmwd_lsc, BCM2835_RSTC_REG) & BCM2835_RSTC_WRCFG_CLR) | (BCM2835_PASWORD << BCM2835_PASSWORD_SHIFT) | BCM2835_RSTC_WRCFG_FULL_RESET); } static device_method_t bcmwd_methods[] = { DEVMETHOD(device_probe, bcmwd_probe), DEVMETHOD(device_attach, bcmwd_attach), DEVMETHOD_END }; static driver_t bcmwd_driver = { "bcmwd", bcmwd_methods, sizeof(struct bcmwd_softc), }; static devclass_t bcmwd_devclass; DRIVER_MODULE(bcmwd, simplebus, bcmwd_driver, bcmwd_devclass, 0, 0); Index: projects/building-blocks/sys/arm/conf/SOCKIT-BERI =================================================================== --- projects/building-blocks/sys/arm/conf/SOCKIT-BERI (revision 275663) +++ projects/building-blocks/sys/arm/conf/SOCKIT-BERI (revision 275664) @@ -1,144 +1,145 @@ # Kernel configuration for Terasic SoCKit (Altera Cyclone V SoC). # # For more information on this file, please read the config(5) manual page, # and/or the handbook section on Kernel Configuration Files: # # http://www.FreeBSD.org/doc/en_US.ISO8859-1/books/handbook/kernelconfig-config.html # # The handbook is also available locally in /usr/share/doc/handbook # if you've installed the doc distribution, otherwise always see the # FreeBSD World Wide Web server (http://www.FreeBSD.org/) for the # latest information. # # An exhaustive list of options and more detailed explanations of the # device lines is also present in the ../../conf/NOTES and NOTES files. # If you are in doubt as to the purpose or necessity of a line, check first # in NOTES. # # $FreeBSD$ ident SOCKIT-BERI include "../altera/socfpga/std.socfpga" makeoptions MODULES_OVERRIDE="" makeoptions DEBUG=-g # Build kernel with gdb(1) debug symbols makeoptions WERROR="-Werror" options HZ=100 options SCHED_4BSD # 4BSD scheduler options INET # InterNETworking options INET6 # IPv6 communications protocols options GEOM_PART_BSD # BSD partition scheme options GEOM_PART_MBR # MBR partition scheme options GEOM_PART_GPT # GUID partition tables options TMPFS # Efficient memory filesystem options FFS # Berkeley Fast Filesystem options SOFTUPDATES options UFS_ACL # Support for access control lists options UFS_DIRHASH # Improve performance on big directories options MSDOSFS # MSDOS Filesystem options CD9660 # ISO 9660 Filesystem options PROCFS # Process filesystem (requires PSEUDOFS) options PSEUDOFS # Pseudo-filesystem framework options COMPAT_43 # Compatible with BSD 4.3 [KEEP THIS!] options SCSI_DELAY=5000 # Delay (in ms) before probing SCSI options KTRACE options SYSVSHM # SYSV-style shared memory options SYSVMSG # SYSV-style message queues options SYSVSEM # SYSV-style semaphores options _KPOSIX_PRIORITY_SCHEDULING # Posix P1003_1B real-time extensions options KBD_INSTALL_CDEV options PREEMPTION options FREEBSD_BOOT_LOADER options VFP # vfp/neon options SMP # Debugging makeoptions DEBUG=-g # Build kernel with gdb(1) debug symbols options BREAK_TO_DEBUGGER #options VERBOSE_SYSINIT # Enable verbose sysinit messages options KDB options DDB # Enable the kernel debugger options INVARIANTS # Enable calls of extra sanity checking options INVARIANT_SUPPORT # Extra sanity checks of internal structures, required by INVARIANTS #options WITNESS # Enable checks to detect deadlocks and cycles #options WITNESS_SKIPSPIN # Don't run witness on spinlocks for speed #options DIAGNOSTIC # NFS support options NFSCL # Network Filesystem Client options NFSLOCKD # Network Lock Manager options NFS_ROOT # NFS usable as /, requires NFSCLIENT # Uncomment this for NFS root #options NFS_ROOT # NFS usable as /, requires NFSCL #options BOOTP_NFSROOT #options BOOTP_COMPAT #options BOOTP #options BOOTP_NFSV3 #options BOOTP_WIRED_TO=ue0 device mmc # mmc/sd bus device mmcsd # mmc/sd flash cards device dwmmc options ROOTDEVNAME=\"ufs:/dev/mmcsd0s4\" # Pseudo devices device loop device random device pty device md device gpio # USB support options USB_HOST_ALIGN=64 # Align usb buffers to cache line size. device usb options USB_DEBUG #options USB_REQ_DEBUG #options USB_VERBOSE #device musb device dwcotg device umass device scbus # SCSI bus (required for ATA/SCSI) device da # Direct Access (disks) device pass # Serial ports device uart device uart_ns8250 # I2C (TWSI) device iic device iicbus # SPI device spibus # BERI specific device beri_ring device beri_mem device beri_vtblk +device vtbe device altera_pio # Ethernet device ether device mii device smsc device smscphy device dwc device micphy # USB ethernet support, requires miibus device miibus device axe # ASIX Electronics USB Ethernet device bpf # Berkeley packet filter #FDT options FDT options FDT_DTB_STATIC makeoptions FDT_DTS_FILE=socfpga-sockit-beri.dts Index: projects/building-blocks/sys/boot/fdt/dts/arm/socfpga-sockit-beri.dts =================================================================== --- projects/building-blocks/sys/boot/fdt/dts/arm/socfpga-sockit-beri.dts (revision 275663) +++ projects/building-blocks/sys/boot/fdt/dts/arm/socfpga-sockit-beri.dts (revision 275664) @@ -1,151 +1,161 @@ /*- * Copyright (c) 2014 Ruslan Bukin * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ /dts-v1/; /include/ "socfpga.dtsi" / { model = "Terasic SoCKit"; compatible = "altr,socfpga-cyclone5", "altr,socfpga"; memreserve = < 0x00000000 0x1000 >, /* SMP trampoline */ - < 0x00001000 0x1000 >; /* virtio block */ + < 0x00001000 0x1000 >, /* virtio block */ + < 0x00002000 0x1000 >; /* virtio net */ memory { device_type = "memory"; reg = < 0x00000000 0x40000000 >; /* 1G RAM */ }; SOC: socfpga { serial0: serial@ffc02000 { status = "okay"; }; usb1: usb@ffb40000 { status = "okay"; }; gmac1: ethernet@ff702000 { status = "okay"; rxd0-skew-ps = <0>; rxd1-skew-ps = <0>; rxd2-skew-ps = <0>; rxd3-skew-ps = <0>; txen-skew-ps = <0>; txc-skew-ps = <2600>; rxdv-skew-ps = <0>; rxc-skew-ps = <2000>; }; mmc: dwmmc@ff704000 { #address-cells = <1>; #size-cells = <0>; status = "okay"; num-slots = <1>; supports-highspeed; broken-cd; bus-frequency = <25000000>; slot@0 { reg = <0>; bus-width = <4>; }; }; beri_mem0: mem@d0000000 { compatible = "sri-cambridge,beri-mem"; reg = <0xd0000000 0x10000000>; /* 256mb */ status = "okay"; }; pio0: pio@c0020000 { compatible = "altr,pio"; reg = <0xc0020000 0x1000>; /* recv */ interrupts = < 76 >; interrupt-parent = <&GIC>; status = "okay"; }; pio1: pio@c0021000 { compatible = "altr,pio"; reg = <0xc0021000 0x1000>; /* send */ interrupts = < 82 >; /* not in use on arm side */ interrupt-parent = <&GIC>; status = "okay"; }; beri_vtblk: vtblk@00001000 { compatible = "sri-cambridge,beri-vtblk"; reg = <0x00001000 0x1000>; + pio-recv = <&pio0>; + pio-send = <&pio1>; + beri-mem = <&beri_mem0>; + status = "disabled"; + }; + + beri_vtnet: vtnet@00002000 { + compatible = "sri-cambridge,beri-vtnet"; + reg = <0x00002000 0x1000>; pio-recv = <&pio0>; pio-send = <&pio1>; beri-mem = <&beri_mem0>; status = "okay"; }; beri_debug: ring@c0000000 { compatible = "sri-cambridge,beri-ring"; reg = <0xc0000000 0x3000>; interrupts = < 72 73 >; interrupt-parent = <&GIC>; device_name = "beri_debug"; data_size = <0x1000>; data_read = <0x0>; data_write = <0x1000>; control_read = <0x2000>; control_write = <0x2010>; status = "okay"; }; beri_console: ring@c0004000 { compatible = "sri-cambridge,beri-ring"; reg = <0xc0004000 0x3000>; interrupts = < 74 75 >; interrupt-parent = <&GIC>; device_name = "beri_console"; data_size = <0x1000>; data_read = <0x0>; data_write = <0x1000>; control_read = <0x2000>; control_write = <0x2010>; status = "okay"; }; }; chosen { bootargs = "-v"; stdin = "serial0"; stdout = "serial0"; }; }; Index: projects/building-blocks/sys/boot =================================================================== --- projects/building-blocks/sys/boot (revision 275663) +++ projects/building-blocks/sys/boot (revision 275664) Property changes on: projects/building-blocks/sys/boot ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/sys/boot:r275590-275663 Index: projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/metaslab.c =================================================================== --- projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/metaslab.c (revision 275663) +++ projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/metaslab.c (revision 275664) @@ -1,2665 +1,2675 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2014 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. */ #include #include #include #include #include #include #include #include #include SYSCTL_DECL(_vfs_zfs); SYSCTL_NODE(_vfs_zfs, OID_AUTO, metaslab, CTLFLAG_RW, 0, "ZFS metaslab"); /* * Allow allocations to switch to gang blocks quickly. We do this to * avoid having to load lots of space_maps in a given txg. There are, * however, some cases where we want to avoid "fast" ganging and instead * we want to do an exhaustive search of all metaslabs on this device. * Currently we don't allow any gang, slog, or dump device related allocations * to "fast" gang. */ #define CAN_FASTGANG(flags) \ (!((flags) & (METASLAB_GANG_CHILD | METASLAB_GANG_HEADER | \ METASLAB_GANG_AVOID))) #define METASLAB_WEIGHT_PRIMARY (1ULL << 63) #define METASLAB_WEIGHT_SECONDARY (1ULL << 62) #define METASLAB_ACTIVE_MASK \ (METASLAB_WEIGHT_PRIMARY | METASLAB_WEIGHT_SECONDARY) uint64_t metaslab_aliquot = 512ULL << 10; uint64_t metaslab_gang_bang = SPA_MAXBLOCKSIZE + 1; /* force gang blocks */ SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, gang_bang, CTLFLAG_RWTUN, &metaslab_gang_bang, 0, "Force gang block allocation for blocks larger than or equal to this value"); /* * The in-core space map representation is more compact than its on-disk form. * The zfs_condense_pct determines how much more compact the in-core * space_map representation must be before we compact it on-disk. * Values should be greater than or equal to 100. */ int zfs_condense_pct = 200; SYSCTL_INT(_vfs_zfs, OID_AUTO, condense_pct, CTLFLAG_RWTUN, &zfs_condense_pct, 0, "Condense on-disk spacemap when it is more than this many percents" " of in-memory counterpart"); /* * Condensing a metaslab is not guaranteed to actually reduce the amount of * space used on disk. In particular, a space map uses data in increments of * MAX(1 << ashift, space_map_blksize), so a metaslab might use the * same number of blocks after condensing. Since the goal of condensing is to * reduce the number of IOPs required to read the space map, we only want to * condense when we can be sure we will reduce the number of blocks used by the * space map. Unfortunately, we cannot precisely compute whether or not this is * the case in metaslab_should_condense since we are holding ms_lock. Instead, * we apply the following heuristic: do not condense a spacemap unless the * uncondensed size consumes greater than zfs_metaslab_condense_block_threshold * blocks. */ int zfs_metaslab_condense_block_threshold = 4; /* * The zfs_mg_noalloc_threshold defines which metaslab groups should * be eligible for allocation. The value is defined as a percentage of * free space. Metaslab groups that have more free space than * zfs_mg_noalloc_threshold are always eligible for allocations. Once * a metaslab group's free space is less than or equal to the * zfs_mg_noalloc_threshold the allocator will avoid allocating to that * group unless all groups in the pool have reached zfs_mg_noalloc_threshold. * Once all groups in the pool reach zfs_mg_noalloc_threshold then all * groups are allowed to accept allocations. Gang blocks are always * eligible to allocate on any metaslab group. The default value of 0 means * no metaslab group will be excluded based on this criterion. */ int zfs_mg_noalloc_threshold = 0; SYSCTL_INT(_vfs_zfs, OID_AUTO, mg_noalloc_threshold, CTLFLAG_RWTUN, &zfs_mg_noalloc_threshold, 0, "Percentage of metaslab group size that should be free" " to make it eligible for allocation"); /* * Metaslab groups are considered eligible for allocations if their * fragmenation metric (measured as a percentage) is less than or equal to * zfs_mg_fragmentation_threshold. If a metaslab group exceeds this threshold * then it will be skipped unless all metaslab groups within the metaslab * class have also crossed this threshold. */ int zfs_mg_fragmentation_threshold = 85; SYSCTL_INT(_vfs_zfs, OID_AUTO, mg_fragmentation_threshold, CTLFLAG_RWTUN, &zfs_mg_fragmentation_threshold, 0, "Percentage of metaslab group size that should be considered " "eligible for allocations unless all metaslab groups within the metaslab class " "have also crossed this threshold"); /* * Allow metaslabs to keep their active state as long as their fragmentation * percentage is less than or equal to zfs_metaslab_fragmentation_threshold. An * active metaslab that exceeds this threshold will no longer keep its active * status allowing better metaslabs to be selected. */ int zfs_metaslab_fragmentation_threshold = 70; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, fragmentation_threshold, CTLFLAG_RWTUN, &zfs_metaslab_fragmentation_threshold, 0, "Maximum percentage of metaslab fragmentation level to keep their active state"); /* * When set will load all metaslabs when pool is first opened. */ int metaslab_debug_load = 0; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, debug_load, CTLFLAG_RWTUN, &metaslab_debug_load, 0, "Load all metaslabs when pool is first opened"); /* * When set will prevent metaslabs from being unloaded. */ int metaslab_debug_unload = 0; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, debug_unload, CTLFLAG_RWTUN, &metaslab_debug_unload, 0, "Prevent metaslabs from being unloaded"); /* * Minimum size which forces the dynamic allocator to change * it's allocation strategy. Once the space map cannot satisfy * an allocation of this size then it switches to using more * aggressive strategy (i.e search by size rather than offset). */ uint64_t metaslab_df_alloc_threshold = SPA_OLD_MAXBLOCKSIZE; SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, df_alloc_threshold, CTLFLAG_RWTUN, &metaslab_df_alloc_threshold, 0, "Minimum size which forces the dynamic allocator to change it's allocation strategy"); /* * The minimum free space, in percent, which must be available * in a space map to continue allocations in a first-fit fashion. * Once the space_map's free space drops below this level we dynamically * switch to using best-fit allocations. */ int metaslab_df_free_pct = 4; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, df_free_pct, CTLFLAG_RWTUN, &metaslab_df_free_pct, 0, "The minimum free space, in percent, which must be available in a " "space map to continue allocations in a first-fit fashion"); /* * A metaslab is considered "free" if it contains a contiguous * segment which is greater than metaslab_min_alloc_size. */ uint64_t metaslab_min_alloc_size = DMU_MAX_ACCESS; SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, min_alloc_size, CTLFLAG_RWTUN, &metaslab_min_alloc_size, 0, "A metaslab is considered \"free\" if it contains a contiguous " "segment which is greater than vfs.zfs.metaslab.min_alloc_size"); /* * Percentage of all cpus that can be used by the metaslab taskq. */ int metaslab_load_pct = 50; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, load_pct, CTLFLAG_RWTUN, &metaslab_load_pct, 0, "Percentage of cpus that can be used by the metaslab taskq"); /* * Determines how many txgs a metaslab may remain loaded without having any * allocations from it. As long as a metaslab continues to be used we will * keep it loaded. */ int metaslab_unload_delay = TXG_SIZE * 2; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, unload_delay, CTLFLAG_RWTUN, &metaslab_unload_delay, 0, "Number of TXGs that an unused metaslab can be kept in memory"); /* * Max number of metaslabs per group to preload. */ int metaslab_preload_limit = SPA_DVAS_PER_BP; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, preload_limit, CTLFLAG_RWTUN, &metaslab_preload_limit, 0, "Max number of metaslabs per group to preload"); /* * Enable/disable preloading of metaslab. */ boolean_t metaslab_preload_enabled = B_TRUE; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, preload_enabled, CTLFLAG_RWTUN, &metaslab_preload_enabled, 0, "Max number of metaslabs per group to preload"); /* * Enable/disable fragmentation weighting on metaslabs. */ boolean_t metaslab_fragmentation_factor_enabled = B_TRUE; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, fragmentation_factor_enabled, CTLFLAG_RWTUN, &metaslab_fragmentation_factor_enabled, 0, "Enable fragmentation weighting on metaslabs"); /* * Enable/disable lba weighting (i.e. outer tracks are given preference). */ boolean_t metaslab_lba_weighting_enabled = B_TRUE; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, lba_weighting_enabled, CTLFLAG_RWTUN, &metaslab_lba_weighting_enabled, 0, "Enable LBA weighting (i.e. outer tracks are given preference)"); /* * Enable/disable metaslab group biasing. */ boolean_t metaslab_bias_enabled = B_TRUE; SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, bias_enabled, CTLFLAG_RWTUN, &metaslab_bias_enabled, 0, "Enable metaslab group biasing"); static uint64_t metaslab_fragmentation(metaslab_t *); /* * ========================================================================== * Metaslab classes * ========================================================================== */ metaslab_class_t * metaslab_class_create(spa_t *spa, metaslab_ops_t *ops) { metaslab_class_t *mc; mc = kmem_zalloc(sizeof (metaslab_class_t), KM_SLEEP); mc->mc_spa = spa; mc->mc_rotor = NULL; mc->mc_ops = ops; return (mc); } void metaslab_class_destroy(metaslab_class_t *mc) { ASSERT(mc->mc_rotor == NULL); ASSERT(mc->mc_alloc == 0); ASSERT(mc->mc_deferred == 0); ASSERT(mc->mc_space == 0); ASSERT(mc->mc_dspace == 0); kmem_free(mc, sizeof (metaslab_class_t)); } int metaslab_class_validate(metaslab_class_t *mc) { metaslab_group_t *mg; vdev_t *vd; /* * Must hold one of the spa_config locks. */ ASSERT(spa_config_held(mc->mc_spa, SCL_ALL, RW_READER) || spa_config_held(mc->mc_spa, SCL_ALL, RW_WRITER)); if ((mg = mc->mc_rotor) == NULL) return (0); do { vd = mg->mg_vd; ASSERT(vd->vdev_mg != NULL); ASSERT3P(vd->vdev_top, ==, vd); ASSERT3P(mg->mg_class, ==, mc); ASSERT3P(vd->vdev_ops, !=, &vdev_hole_ops); } while ((mg = mg->mg_next) != mc->mc_rotor); return (0); } void metaslab_class_space_update(metaslab_class_t *mc, int64_t alloc_delta, int64_t defer_delta, int64_t space_delta, int64_t dspace_delta) { atomic_add_64(&mc->mc_alloc, alloc_delta); atomic_add_64(&mc->mc_deferred, defer_delta); atomic_add_64(&mc->mc_space, space_delta); atomic_add_64(&mc->mc_dspace, dspace_delta); } void metaslab_class_minblocksize_update(metaslab_class_t *mc) { metaslab_group_t *mg; vdev_t *vd; uint64_t minashift = UINT64_MAX; if ((mg = mc->mc_rotor) == NULL) { mc->mc_minblocksize = SPA_MINBLOCKSIZE; return; } do { vd = mg->mg_vd; if (vd->vdev_ashift < minashift) minashift = vd->vdev_ashift; } while ((mg = mg->mg_next) != mc->mc_rotor); mc->mc_minblocksize = 1ULL << minashift; } uint64_t metaslab_class_get_alloc(metaslab_class_t *mc) { return (mc->mc_alloc); } uint64_t metaslab_class_get_deferred(metaslab_class_t *mc) { return (mc->mc_deferred); } uint64_t metaslab_class_get_space(metaslab_class_t *mc) { return (mc->mc_space); } uint64_t metaslab_class_get_dspace(metaslab_class_t *mc) { return (spa_deflate(mc->mc_spa) ? mc->mc_dspace : mc->mc_space); } uint64_t metaslab_class_get_minblocksize(metaslab_class_t *mc) { return (mc->mc_minblocksize); } void metaslab_class_histogram_verify(metaslab_class_t *mc) { vdev_t *rvd = mc->mc_spa->spa_root_vdev; uint64_t *mc_hist; int i; if ((zfs_flags & ZFS_DEBUG_HISTOGRAM_VERIFY) == 0) return; mc_hist = kmem_zalloc(sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE, KM_SLEEP); for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; metaslab_group_t *mg = tvd->vdev_mg; /* * Skip any holes, uninitialized top-levels, or * vdevs that are not in this metalab class. */ if (tvd->vdev_ishole || tvd->vdev_ms_shift == 0 || mg->mg_class != mc) { continue; } for (i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) mc_hist[i] += mg->mg_histogram[i]; } for (i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) VERIFY3U(mc_hist[i], ==, mc->mc_histogram[i]); kmem_free(mc_hist, sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE); } /* * Calculate the metaslab class's fragmentation metric. The metric * is weighted based on the space contribution of each metaslab group. * The return value will be a number between 0 and 100 (inclusive), or * ZFS_FRAG_INVALID if the metric has not been set. See comment above the * zfs_frag_table for more information about the metric. */ uint64_t metaslab_class_fragmentation(metaslab_class_t *mc) { vdev_t *rvd = mc->mc_spa->spa_root_vdev; uint64_t fragmentation = 0; spa_config_enter(mc->mc_spa, SCL_VDEV, FTAG, RW_READER); for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; metaslab_group_t *mg = tvd->vdev_mg; /* * Skip any holes, uninitialized top-levels, or * vdevs that are not in this metalab class. */ if (tvd->vdev_ishole || tvd->vdev_ms_shift == 0 || mg->mg_class != mc) { continue; } /* * If a metaslab group does not contain a fragmentation * metric then just bail out. */ if (mg->mg_fragmentation == ZFS_FRAG_INVALID) { spa_config_exit(mc->mc_spa, SCL_VDEV, FTAG); return (ZFS_FRAG_INVALID); } /* * Determine how much this metaslab_group is contributing * to the overall pool fragmentation metric. */ fragmentation += mg->mg_fragmentation * metaslab_group_get_space(mg); } fragmentation /= metaslab_class_get_space(mc); ASSERT3U(fragmentation, <=, 100); spa_config_exit(mc->mc_spa, SCL_VDEV, FTAG); return (fragmentation); } /* * Calculate the amount of expandable space that is available in * this metaslab class. If a device is expanded then its expandable * space will be the amount of allocatable space that is currently not * part of this metaslab class. */ uint64_t metaslab_class_expandable_space(metaslab_class_t *mc) { vdev_t *rvd = mc->mc_spa->spa_root_vdev; uint64_t space = 0; spa_config_enter(mc->mc_spa, SCL_VDEV, FTAG, RW_READER); for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; metaslab_group_t *mg = tvd->vdev_mg; if (tvd->vdev_ishole || tvd->vdev_ms_shift == 0 || mg->mg_class != mc) { continue; } space += tvd->vdev_max_asize - tvd->vdev_asize; } spa_config_exit(mc->mc_spa, SCL_VDEV, FTAG); return (space); } /* * ========================================================================== * Metaslab groups * ========================================================================== */ static int metaslab_compare(const void *x1, const void *x2) { const metaslab_t *m1 = x1; const metaslab_t *m2 = x2; if (m1->ms_weight < m2->ms_weight) return (1); if (m1->ms_weight > m2->ms_weight) return (-1); /* * If the weights are identical, use the offset to force uniqueness. */ if (m1->ms_start < m2->ms_start) return (-1); if (m1->ms_start > m2->ms_start) return (1); ASSERT3P(m1, ==, m2); return (0); } /* * Update the allocatable flag and the metaslab group's capacity. * The allocatable flag is set to true if the capacity is below * the zfs_mg_noalloc_threshold. If a metaslab group transitions * from allocatable to non-allocatable or vice versa then the metaslab * group's class is updated to reflect the transition. */ static void metaslab_group_alloc_update(metaslab_group_t *mg) { vdev_t *vd = mg->mg_vd; metaslab_class_t *mc = mg->mg_class; vdev_stat_t *vs = &vd->vdev_stat; boolean_t was_allocatable; ASSERT(vd == vd->vdev_top); mutex_enter(&mg->mg_lock); was_allocatable = mg->mg_allocatable; mg->mg_free_capacity = ((vs->vs_space - vs->vs_alloc) * 100) / (vs->vs_space + 1); /* * A metaslab group is considered allocatable if it has plenty * of free space or is not heavily fragmented. We only take * fragmentation into account if the metaslab group has a valid * fragmentation metric (i.e. a value between 0 and 100). */ mg->mg_allocatable = (mg->mg_free_capacity > zfs_mg_noalloc_threshold && (mg->mg_fragmentation == ZFS_FRAG_INVALID || mg->mg_fragmentation <= zfs_mg_fragmentation_threshold)); /* * The mc_alloc_groups maintains a count of the number of * groups in this metaslab class that are still above the * zfs_mg_noalloc_threshold. This is used by the allocating * threads to determine if they should avoid allocations to * a given group. The allocator will avoid allocations to a group * if that group has reached or is below the zfs_mg_noalloc_threshold * and there are still other groups that are above the threshold. * When a group transitions from allocatable to non-allocatable or * vice versa we update the metaslab class to reflect that change. * When the mc_alloc_groups value drops to 0 that means that all * groups have reached the zfs_mg_noalloc_threshold making all groups * eligible for allocations. This effectively means that all devices * are balanced again. */ if (was_allocatable && !mg->mg_allocatable) mc->mc_alloc_groups--; else if (!was_allocatable && mg->mg_allocatable) mc->mc_alloc_groups++; mutex_exit(&mg->mg_lock); } metaslab_group_t * metaslab_group_create(metaslab_class_t *mc, vdev_t *vd) { metaslab_group_t *mg; mg = kmem_zalloc(sizeof (metaslab_group_t), KM_SLEEP); mutex_init(&mg->mg_lock, NULL, MUTEX_DEFAULT, NULL); avl_create(&mg->mg_metaslab_tree, metaslab_compare, sizeof (metaslab_t), offsetof(struct metaslab, ms_group_node)); mg->mg_vd = vd; mg->mg_class = mc; mg->mg_activation_count = 0; mg->mg_taskq = taskq_create("metaslab_group_taskq", metaslab_load_pct, minclsyspri, 10, INT_MAX, TASKQ_THREADS_CPU_PCT); return (mg); } void metaslab_group_destroy(metaslab_group_t *mg) { ASSERT(mg->mg_prev == NULL); ASSERT(mg->mg_next == NULL); /* * We may have gone below zero with the activation count * either because we never activated in the first place or * because we're done, and possibly removing the vdev. */ ASSERT(mg->mg_activation_count <= 0); taskq_destroy(mg->mg_taskq); avl_destroy(&mg->mg_metaslab_tree); mutex_destroy(&mg->mg_lock); kmem_free(mg, sizeof (metaslab_group_t)); } void metaslab_group_activate(metaslab_group_t *mg) { metaslab_class_t *mc = mg->mg_class; metaslab_group_t *mgprev, *mgnext; ASSERT(spa_config_held(mc->mc_spa, SCL_ALLOC, RW_WRITER)); ASSERT(mc->mc_rotor != mg); ASSERT(mg->mg_prev == NULL); ASSERT(mg->mg_next == NULL); ASSERT(mg->mg_activation_count <= 0); if (++mg->mg_activation_count <= 0) return; mg->mg_aliquot = metaslab_aliquot * MAX(1, mg->mg_vd->vdev_children); metaslab_group_alloc_update(mg); if ((mgprev = mc->mc_rotor) == NULL) { mg->mg_prev = mg; mg->mg_next = mg; } else { mgnext = mgprev->mg_next; mg->mg_prev = mgprev; mg->mg_next = mgnext; mgprev->mg_next = mg; mgnext->mg_prev = mg; } mc->mc_rotor = mg; metaslab_class_minblocksize_update(mc); } void metaslab_group_passivate(metaslab_group_t *mg) { metaslab_class_t *mc = mg->mg_class; metaslab_group_t *mgprev, *mgnext; ASSERT(spa_config_held(mc->mc_spa, SCL_ALLOC, RW_WRITER)); if (--mg->mg_activation_count != 0) { ASSERT(mc->mc_rotor != mg); ASSERT(mg->mg_prev == NULL); ASSERT(mg->mg_next == NULL); ASSERT(mg->mg_activation_count < 0); return; } taskq_wait(mg->mg_taskq); metaslab_group_alloc_update(mg); mgprev = mg->mg_prev; mgnext = mg->mg_next; if (mg == mgnext) { mc->mc_rotor = NULL; } else { mc->mc_rotor = mgnext; mgprev->mg_next = mgnext; mgnext->mg_prev = mgprev; } mg->mg_prev = NULL; mg->mg_next = NULL; metaslab_class_minblocksize_update(mc); } uint64_t metaslab_group_get_space(metaslab_group_t *mg) { return ((1ULL << mg->mg_vd->vdev_ms_shift) * mg->mg_vd->vdev_ms_count); } void metaslab_group_histogram_verify(metaslab_group_t *mg) { uint64_t *mg_hist; vdev_t *vd = mg->mg_vd; uint64_t ashift = vd->vdev_ashift; int i; if ((zfs_flags & ZFS_DEBUG_HISTOGRAM_VERIFY) == 0) return; mg_hist = kmem_zalloc(sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE, KM_SLEEP); ASSERT3U(RANGE_TREE_HISTOGRAM_SIZE, >=, SPACE_MAP_HISTOGRAM_SIZE + ashift); for (int m = 0; m < vd->vdev_ms_count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp->ms_sm == NULL) continue; for (i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) mg_hist[i + ashift] += msp->ms_sm->sm_phys->smp_histogram[i]; } for (i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i ++) VERIFY3U(mg_hist[i], ==, mg->mg_histogram[i]); kmem_free(mg_hist, sizeof (uint64_t) * RANGE_TREE_HISTOGRAM_SIZE); } static void metaslab_group_histogram_add(metaslab_group_t *mg, metaslab_t *msp) { metaslab_class_t *mc = mg->mg_class; uint64_t ashift = mg->mg_vd->vdev_ashift; ASSERT(MUTEX_HELD(&msp->ms_lock)); if (msp->ms_sm == NULL) return; mutex_enter(&mg->mg_lock); for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) { mg->mg_histogram[i + ashift] += msp->ms_sm->sm_phys->smp_histogram[i]; mc->mc_histogram[i + ashift] += msp->ms_sm->sm_phys->smp_histogram[i]; } mutex_exit(&mg->mg_lock); } void metaslab_group_histogram_remove(metaslab_group_t *mg, metaslab_t *msp) { metaslab_class_t *mc = mg->mg_class; uint64_t ashift = mg->mg_vd->vdev_ashift; ASSERT(MUTEX_HELD(&msp->ms_lock)); if (msp->ms_sm == NULL) return; mutex_enter(&mg->mg_lock); for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) { ASSERT3U(mg->mg_histogram[i + ashift], >=, msp->ms_sm->sm_phys->smp_histogram[i]); ASSERT3U(mc->mc_histogram[i + ashift], >=, msp->ms_sm->sm_phys->smp_histogram[i]); mg->mg_histogram[i + ashift] -= msp->ms_sm->sm_phys->smp_histogram[i]; mc->mc_histogram[i + ashift] -= msp->ms_sm->sm_phys->smp_histogram[i]; } mutex_exit(&mg->mg_lock); } static void metaslab_group_add(metaslab_group_t *mg, metaslab_t *msp) { ASSERT(msp->ms_group == NULL); mutex_enter(&mg->mg_lock); msp->ms_group = mg; msp->ms_weight = 0; avl_add(&mg->mg_metaslab_tree, msp); mutex_exit(&mg->mg_lock); mutex_enter(&msp->ms_lock); metaslab_group_histogram_add(mg, msp); mutex_exit(&msp->ms_lock); } static void metaslab_group_remove(metaslab_group_t *mg, metaslab_t *msp) { mutex_enter(&msp->ms_lock); metaslab_group_histogram_remove(mg, msp); mutex_exit(&msp->ms_lock); mutex_enter(&mg->mg_lock); ASSERT(msp->ms_group == mg); avl_remove(&mg->mg_metaslab_tree, msp); msp->ms_group = NULL; mutex_exit(&mg->mg_lock); } static void metaslab_group_sort(metaslab_group_t *mg, metaslab_t *msp, uint64_t weight) { /* * Although in principle the weight can be any value, in * practice we do not use values in the range [1, 511]. */ ASSERT(weight >= SPA_MINBLOCKSIZE || weight == 0); ASSERT(MUTEX_HELD(&msp->ms_lock)); mutex_enter(&mg->mg_lock); ASSERT(msp->ms_group == mg); avl_remove(&mg->mg_metaslab_tree, msp); msp->ms_weight = weight; avl_add(&mg->mg_metaslab_tree, msp); mutex_exit(&mg->mg_lock); } /* * Calculate the fragmentation for a given metaslab group. We can use * a simple average here since all metaslabs within the group must have * the same size. The return value will be a value between 0 and 100 * (inclusive), or ZFS_FRAG_INVALID if less than half of the metaslab in this * group have a fragmentation metric. */ uint64_t metaslab_group_fragmentation(metaslab_group_t *mg) { vdev_t *vd = mg->mg_vd; uint64_t fragmentation = 0; uint64_t valid_ms = 0; for (int m = 0; m < vd->vdev_ms_count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp->ms_fragmentation == ZFS_FRAG_INVALID) continue; valid_ms++; fragmentation += msp->ms_fragmentation; } if (valid_ms <= vd->vdev_ms_count / 2) return (ZFS_FRAG_INVALID); fragmentation /= valid_ms; ASSERT3U(fragmentation, <=, 100); return (fragmentation); } /* * Determine if a given metaslab group should skip allocations. A metaslab * group should avoid allocations if its free capacity is less than the * zfs_mg_noalloc_threshold or its fragmentation metric is greater than * zfs_mg_fragmentation_threshold and there is at least one metaslab group * that can still handle allocations. */ static boolean_t metaslab_group_allocatable(metaslab_group_t *mg) { vdev_t *vd = mg->mg_vd; spa_t *spa = vd->vdev_spa; metaslab_class_t *mc = mg->mg_class; /* * We use two key metrics to determine if a metaslab group is * considered allocatable -- free space and fragmentation. If * the free space is greater than the free space threshold and * the fragmentation is less than the fragmentation threshold then * consider the group allocatable. There are two case when we will * not consider these key metrics. The first is if the group is * associated with a slog device and the second is if all groups * in this metaslab class have already been consider ineligible * for allocations. */ return ((mg->mg_free_capacity > zfs_mg_noalloc_threshold && (mg->mg_fragmentation == ZFS_FRAG_INVALID || mg->mg_fragmentation <= zfs_mg_fragmentation_threshold)) || mc != spa_normal_class(spa) || mc->mc_alloc_groups == 0); } /* * ========================================================================== * Range tree callbacks * ========================================================================== */ /* * Comparison function for the private size-ordered tree. Tree is sorted * by size, larger sizes at the end of the tree. */ static int metaslab_rangesize_compare(const void *x1, const void *x2) { const range_seg_t *r1 = x1; const range_seg_t *r2 = x2; uint64_t rs_size1 = r1->rs_end - r1->rs_start; uint64_t rs_size2 = r2->rs_end - r2->rs_start; if (rs_size1 < rs_size2) return (-1); if (rs_size1 > rs_size2) return (1); if (r1->rs_start < r2->rs_start) return (-1); if (r1->rs_start > r2->rs_start) return (1); return (0); } /* * Create any block allocator specific components. The current allocators * rely on using both a size-ordered range_tree_t and an array of uint64_t's. */ static void metaslab_rt_create(range_tree_t *rt, void *arg) { metaslab_t *msp = arg; ASSERT3P(rt->rt_arg, ==, msp); ASSERT(msp->ms_tree == NULL); avl_create(&msp->ms_size_tree, metaslab_rangesize_compare, sizeof (range_seg_t), offsetof(range_seg_t, rs_pp_node)); } /* * Destroy the block allocator specific components. */ static void metaslab_rt_destroy(range_tree_t *rt, void *arg) { metaslab_t *msp = arg; ASSERT3P(rt->rt_arg, ==, msp); ASSERT3P(msp->ms_tree, ==, rt); ASSERT0(avl_numnodes(&msp->ms_size_tree)); avl_destroy(&msp->ms_size_tree); } static void metaslab_rt_add(range_tree_t *rt, range_seg_t *rs, void *arg) { metaslab_t *msp = arg; ASSERT3P(rt->rt_arg, ==, msp); ASSERT3P(msp->ms_tree, ==, rt); VERIFY(!msp->ms_condensing); avl_add(&msp->ms_size_tree, rs); } static void metaslab_rt_remove(range_tree_t *rt, range_seg_t *rs, void *arg) { metaslab_t *msp = arg; ASSERT3P(rt->rt_arg, ==, msp); ASSERT3P(msp->ms_tree, ==, rt); VERIFY(!msp->ms_condensing); avl_remove(&msp->ms_size_tree, rs); } static void metaslab_rt_vacate(range_tree_t *rt, void *arg) { metaslab_t *msp = arg; ASSERT3P(rt->rt_arg, ==, msp); ASSERT3P(msp->ms_tree, ==, rt); /* * Normally one would walk the tree freeing nodes along the way. * Since the nodes are shared with the range trees we can avoid * walking all nodes and just reinitialize the avl tree. The nodes * will be freed by the range tree, so we don't want to free them here. */ avl_create(&msp->ms_size_tree, metaslab_rangesize_compare, sizeof (range_seg_t), offsetof(range_seg_t, rs_pp_node)); } static range_tree_ops_t metaslab_rt_ops = { metaslab_rt_create, metaslab_rt_destroy, metaslab_rt_add, metaslab_rt_remove, metaslab_rt_vacate }; /* * ========================================================================== * Metaslab block operations * ========================================================================== */ /* * Return the maximum contiguous segment within the metaslab. */ uint64_t metaslab_block_maxsize(metaslab_t *msp) { avl_tree_t *t = &msp->ms_size_tree; range_seg_t *rs; if (t == NULL || (rs = avl_last(t)) == NULL) return (0ULL); return (rs->rs_end - rs->rs_start); } uint64_t metaslab_block_alloc(metaslab_t *msp, uint64_t size) { uint64_t start; range_tree_t *rt = msp->ms_tree; VERIFY(!msp->ms_condensing); start = msp->ms_ops->msop_alloc(msp, size); if (start != -1ULL) { vdev_t *vd = msp->ms_group->mg_vd; VERIFY0(P2PHASE(start, 1ULL << vd->vdev_ashift)); VERIFY0(P2PHASE(size, 1ULL << vd->vdev_ashift)); VERIFY3U(range_tree_space(rt) - size, <=, msp->ms_size); range_tree_remove(rt, start, size); } return (start); } /* * ========================================================================== * Common allocator routines * ========================================================================== */ /* * This is a helper function that can be used by the allocator to find * a suitable block to allocate. This will search the specified AVL * tree looking for a block that matches the specified criteria. */ static uint64_t metaslab_block_picker(avl_tree_t *t, uint64_t *cursor, uint64_t size, uint64_t align) { range_seg_t *rs, rsearch; avl_index_t where; rsearch.rs_start = *cursor; rsearch.rs_end = *cursor + size; rs = avl_find(t, &rsearch, &where); if (rs == NULL) rs = avl_nearest(t, where, AVL_AFTER); while (rs != NULL) { uint64_t offset = P2ROUNDUP(rs->rs_start, align); if (offset + size <= rs->rs_end) { *cursor = offset + size; return (offset); } rs = AVL_NEXT(t, rs); } /* * If we know we've searched the whole map (*cursor == 0), give up. * Otherwise, reset the cursor to the beginning and try again. */ if (*cursor == 0) return (-1ULL); *cursor = 0; return (metaslab_block_picker(t, cursor, size, align)); } /* * ========================================================================== * The first-fit block allocator * ========================================================================== */ static uint64_t metaslab_ff_alloc(metaslab_t *msp, uint64_t size) { /* * Find the largest power of 2 block size that evenly divides the * requested size. This is used to try to allocate blocks with similar * alignment from the same area of the metaslab (i.e. same cursor * bucket) but it does not guarantee that other allocations sizes * may exist in the same region. */ uint64_t align = size & -size; uint64_t *cursor = &msp->ms_lbas[highbit64(align) - 1]; avl_tree_t *t = &msp->ms_tree->rt_root; return (metaslab_block_picker(t, cursor, size, align)); } static metaslab_ops_t metaslab_ff_ops = { metaslab_ff_alloc }; /* * ========================================================================== * Dynamic block allocator - * Uses the first fit allocation scheme until space get low and then * adjusts to a best fit allocation method. Uses metaslab_df_alloc_threshold * and metaslab_df_free_pct to determine when to switch the allocation scheme. * ========================================================================== */ static uint64_t metaslab_df_alloc(metaslab_t *msp, uint64_t size) { /* * Find the largest power of 2 block size that evenly divides the * requested size. This is used to try to allocate blocks with similar * alignment from the same area of the metaslab (i.e. same cursor * bucket) but it does not guarantee that other allocations sizes * may exist in the same region. */ uint64_t align = size & -size; uint64_t *cursor = &msp->ms_lbas[highbit64(align) - 1]; range_tree_t *rt = msp->ms_tree; avl_tree_t *t = &rt->rt_root; uint64_t max_size = metaslab_block_maxsize(msp); int free_pct = range_tree_space(rt) * 100 / msp->ms_size; ASSERT(MUTEX_HELD(&msp->ms_lock)); ASSERT3U(avl_numnodes(t), ==, avl_numnodes(&msp->ms_size_tree)); if (max_size < size) return (-1ULL); /* * If we're running low on space switch to using the size * sorted AVL tree (best-fit). */ if (max_size < metaslab_df_alloc_threshold || free_pct < metaslab_df_free_pct) { t = &msp->ms_size_tree; *cursor = 0; } return (metaslab_block_picker(t, cursor, size, 1ULL)); } static metaslab_ops_t metaslab_df_ops = { metaslab_df_alloc }; /* * ========================================================================== * Cursor fit block allocator - * Select the largest region in the metaslab, set the cursor to the beginning * of the range and the cursor_end to the end of the range. As allocations * are made advance the cursor. Continue allocating from the cursor until * the range is exhausted and then find a new range. * ========================================================================== */ static uint64_t metaslab_cf_alloc(metaslab_t *msp, uint64_t size) { range_tree_t *rt = msp->ms_tree; avl_tree_t *t = &msp->ms_size_tree; uint64_t *cursor = &msp->ms_lbas[0]; uint64_t *cursor_end = &msp->ms_lbas[1]; uint64_t offset = 0; ASSERT(MUTEX_HELD(&msp->ms_lock)); ASSERT3U(avl_numnodes(t), ==, avl_numnodes(&rt->rt_root)); ASSERT3U(*cursor_end, >=, *cursor); if ((*cursor + size) > *cursor_end) { range_seg_t *rs; rs = avl_last(&msp->ms_size_tree); if (rs == NULL || (rs->rs_end - rs->rs_start) < size) return (-1ULL); *cursor = rs->rs_start; *cursor_end = rs->rs_end; } offset = *cursor; *cursor += size; return (offset); } static metaslab_ops_t metaslab_cf_ops = { metaslab_cf_alloc }; /* * ========================================================================== * New dynamic fit allocator - * Select a region that is large enough to allocate 2^metaslab_ndf_clump_shift * contiguous blocks. If no region is found then just use the largest segment * that remains. * ========================================================================== */ /* * Determines desired number of contiguous blocks (2^metaslab_ndf_clump_shift) * to request from the allocator. */ uint64_t metaslab_ndf_clump_shift = 4; static uint64_t metaslab_ndf_alloc(metaslab_t *msp, uint64_t size) { avl_tree_t *t = &msp->ms_tree->rt_root; avl_index_t where; range_seg_t *rs, rsearch; uint64_t hbit = highbit64(size); uint64_t *cursor = &msp->ms_lbas[hbit - 1]; uint64_t max_size = metaslab_block_maxsize(msp); ASSERT(MUTEX_HELD(&msp->ms_lock)); ASSERT3U(avl_numnodes(t), ==, avl_numnodes(&msp->ms_size_tree)); if (max_size < size) return (-1ULL); rsearch.rs_start = *cursor; rsearch.rs_end = *cursor + size; rs = avl_find(t, &rsearch, &where); if (rs == NULL || (rs->rs_end - rs->rs_start) < size) { t = &msp->ms_size_tree; rsearch.rs_start = 0; rsearch.rs_end = MIN(max_size, 1ULL << (hbit + metaslab_ndf_clump_shift)); rs = avl_find(t, &rsearch, &where); if (rs == NULL) rs = avl_nearest(t, where, AVL_AFTER); ASSERT(rs != NULL); } if ((rs->rs_end - rs->rs_start) >= size) { *cursor = rs->rs_start + size; return (rs->rs_start); } return (-1ULL); } static metaslab_ops_t metaslab_ndf_ops = { metaslab_ndf_alloc }; metaslab_ops_t *zfs_metaslab_ops = &metaslab_df_ops; /* * ========================================================================== * Metaslabs * ========================================================================== */ /* * Wait for any in-progress metaslab loads to complete. */ void metaslab_load_wait(metaslab_t *msp) { ASSERT(MUTEX_HELD(&msp->ms_lock)); while (msp->ms_loading) { ASSERT(!msp->ms_loaded); cv_wait(&msp->ms_load_cv, &msp->ms_lock); } } int metaslab_load(metaslab_t *msp) { int error = 0; ASSERT(MUTEX_HELD(&msp->ms_lock)); ASSERT(!msp->ms_loaded); ASSERT(!msp->ms_loading); msp->ms_loading = B_TRUE; /* * If the space map has not been allocated yet, then treat * all the space in the metaslab as free and add it to the * ms_tree. */ if (msp->ms_sm != NULL) error = space_map_load(msp->ms_sm, msp->ms_tree, SM_FREE); else range_tree_add(msp->ms_tree, msp->ms_start, msp->ms_size); msp->ms_loaded = (error == 0); msp->ms_loading = B_FALSE; if (msp->ms_loaded) { for (int t = 0; t < TXG_DEFER_SIZE; t++) { range_tree_walk(msp->ms_defertree[t], range_tree_remove, msp->ms_tree); } } cv_broadcast(&msp->ms_load_cv); return (error); } void metaslab_unload(metaslab_t *msp) { ASSERT(MUTEX_HELD(&msp->ms_lock)); range_tree_vacate(msp->ms_tree, NULL, NULL); msp->ms_loaded = B_FALSE; msp->ms_weight &= ~METASLAB_ACTIVE_MASK; } -metaslab_t * -metaslab_init(metaslab_group_t *mg, uint64_t id, uint64_t object, uint64_t txg) +int +metaslab_init(metaslab_group_t *mg, uint64_t id, uint64_t object, uint64_t txg, + metaslab_t **msp) { vdev_t *vd = mg->mg_vd; objset_t *mos = vd->vdev_spa->spa_meta_objset; - metaslab_t *msp; + metaslab_t *ms; + int error; - msp = kmem_zalloc(sizeof (metaslab_t), KM_SLEEP); - mutex_init(&msp->ms_lock, NULL, MUTEX_DEFAULT, NULL); - cv_init(&msp->ms_load_cv, NULL, CV_DEFAULT, NULL); - msp->ms_id = id; - msp->ms_start = id << vd->vdev_ms_shift; - msp->ms_size = 1ULL << vd->vdev_ms_shift; + ms = kmem_zalloc(sizeof (metaslab_t), KM_SLEEP); + mutex_init(&ms->ms_lock, NULL, MUTEX_DEFAULT, NULL); + cv_init(&ms->ms_load_cv, NULL, CV_DEFAULT, NULL); + ms->ms_id = id; + ms->ms_start = id << vd->vdev_ms_shift; + ms->ms_size = 1ULL << vd->vdev_ms_shift; /* * We only open space map objects that already exist. All others * will be opened when we finally allocate an object for it. */ if (object != 0) { - VERIFY0(space_map_open(&msp->ms_sm, mos, object, msp->ms_start, - msp->ms_size, vd->vdev_ashift, &msp->ms_lock)); - ASSERT(msp->ms_sm != NULL); + error = space_map_open(&ms->ms_sm, mos, object, ms->ms_start, + ms->ms_size, vd->vdev_ashift, &ms->ms_lock); + + if (error != 0) { + kmem_free(ms, sizeof (metaslab_t)); + return (error); + } + + ASSERT(ms->ms_sm != NULL); } /* * We create the main range tree here, but we don't create the * alloctree and freetree until metaslab_sync_done(). This serves * two purposes: it allows metaslab_sync_done() to detect the * addition of new space; and for debugging, it ensures that we'd * data fault on any attempt to use this metaslab before it's ready. */ - msp->ms_tree = range_tree_create(&metaslab_rt_ops, msp, &msp->ms_lock); - metaslab_group_add(mg, msp); + ms->ms_tree = range_tree_create(&metaslab_rt_ops, ms, &ms->ms_lock); + metaslab_group_add(mg, ms); - msp->ms_fragmentation = metaslab_fragmentation(msp); - msp->ms_ops = mg->mg_class->mc_ops; + ms->ms_fragmentation = metaslab_fragmentation(ms); + ms->ms_ops = mg->mg_class->mc_ops; /* * If we're opening an existing pool (txg == 0) or creating * a new one (txg == TXG_INITIAL), all space is available now. * If we're adding space to an existing pool, the new space * does not become available until after this txg has synced. */ if (txg <= TXG_INITIAL) - metaslab_sync_done(msp, 0); + metaslab_sync_done(ms, 0); /* * If metaslab_debug_load is set and we're initializing a metaslab * that has an allocated space_map object then load the its space * map so that can verify frees. */ - if (metaslab_debug_load && msp->ms_sm != NULL) { - mutex_enter(&msp->ms_lock); - VERIFY0(metaslab_load(msp)); - mutex_exit(&msp->ms_lock); + if (metaslab_debug_load && ms->ms_sm != NULL) { + mutex_enter(&ms->ms_lock); + VERIFY0(metaslab_load(ms)); + mutex_exit(&ms->ms_lock); } if (txg != 0) { vdev_dirty(vd, 0, NULL, txg); - vdev_dirty(vd, VDD_METASLAB, msp, txg); + vdev_dirty(vd, VDD_METASLAB, ms, txg); } - return (msp); + *msp = ms; + + return (0); } void metaslab_fini(metaslab_t *msp) { metaslab_group_t *mg = msp->ms_group; metaslab_group_remove(mg, msp); mutex_enter(&msp->ms_lock); VERIFY(msp->ms_group == NULL); vdev_space_update(mg->mg_vd, -space_map_allocated(msp->ms_sm), 0, -msp->ms_size); space_map_close(msp->ms_sm); metaslab_unload(msp); range_tree_destroy(msp->ms_tree); for (int t = 0; t < TXG_SIZE; t++) { range_tree_destroy(msp->ms_alloctree[t]); range_tree_destroy(msp->ms_freetree[t]); } for (int t = 0; t < TXG_DEFER_SIZE; t++) { range_tree_destroy(msp->ms_defertree[t]); } ASSERT0(msp->ms_deferspace); mutex_exit(&msp->ms_lock); cv_destroy(&msp->ms_load_cv); mutex_destroy(&msp->ms_lock); kmem_free(msp, sizeof (metaslab_t)); } #define FRAGMENTATION_TABLE_SIZE 17 /* * This table defines a segment size based fragmentation metric that will * allow each metaslab to derive its own fragmentation value. This is done * by calculating the space in each bucket of the spacemap histogram and * multiplying that by the fragmetation metric in this table. Doing * this for all buckets and dividing it by the total amount of free * space in this metaslab (i.e. the total free space in all buckets) gives * us the fragmentation metric. This means that a high fragmentation metric * equates to most of the free space being comprised of small segments. * Conversely, if the metric is low, then most of the free space is in * large segments. A 10% change in fragmentation equates to approximately * double the number of segments. * * This table defines 0% fragmented space using 16MB segments. Testing has * shown that segments that are greater than or equal to 16MB do not suffer * from drastic performance problems. Using this value, we derive the rest * of the table. Since the fragmentation value is never stored on disk, it * is possible to change these calculations in the future. */ int zfs_frag_table[FRAGMENTATION_TABLE_SIZE] = { 100, /* 512B */ 100, /* 1K */ 98, /* 2K */ 95, /* 4K */ 90, /* 8K */ 80, /* 16K */ 70, /* 32K */ 60, /* 64K */ 50, /* 128K */ 40, /* 256K */ 30, /* 512K */ 20, /* 1M */ 15, /* 2M */ 10, /* 4M */ 5, /* 8M */ 0 /* 16M */ }; /* * Calclate the metaslab's fragmentation metric. A return value * of ZFS_FRAG_INVALID means that the metaslab has not been upgraded and does * not support this metric. Otherwise, the return value should be in the * range [0, 100]. */ static uint64_t metaslab_fragmentation(metaslab_t *msp) { spa_t *spa = msp->ms_group->mg_vd->vdev_spa; uint64_t fragmentation = 0; uint64_t total = 0; boolean_t feature_enabled = spa_feature_is_enabled(spa, SPA_FEATURE_SPACEMAP_HISTOGRAM); if (!feature_enabled) return (ZFS_FRAG_INVALID); /* * A null space map means that the entire metaslab is free * and thus is not fragmented. */ if (msp->ms_sm == NULL) return (0); /* * If this metaslab's space_map has not been upgraded, flag it * so that we upgrade next time we encounter it. */ if (msp->ms_sm->sm_dbuf->db_size != sizeof (space_map_phys_t)) { uint64_t txg = spa_syncing_txg(spa); vdev_t *vd = msp->ms_group->mg_vd; if (spa_writeable(spa)) { msp->ms_condense_wanted = B_TRUE; vdev_dirty(vd, VDD_METASLAB, msp, txg + 1); spa_dbgmsg(spa, "txg %llu, requesting force condense: " "msp %p, vd %p", txg, msp, vd); } return (ZFS_FRAG_INVALID); } for (int i = 0; i < SPACE_MAP_HISTOGRAM_SIZE; i++) { uint64_t space = 0; uint8_t shift = msp->ms_sm->sm_shift; int idx = MIN(shift - SPA_MINBLOCKSHIFT + i, FRAGMENTATION_TABLE_SIZE - 1); if (msp->ms_sm->sm_phys->smp_histogram[i] == 0) continue; space = msp->ms_sm->sm_phys->smp_histogram[i] << (i + shift); total += space; ASSERT3U(idx, <, FRAGMENTATION_TABLE_SIZE); fragmentation += space * zfs_frag_table[idx]; } if (total > 0) fragmentation /= total; ASSERT3U(fragmentation, <=, 100); return (fragmentation); } /* * Compute a weight -- a selection preference value -- for the given metaslab. * This is based on the amount of free space, the level of fragmentation, * the LBA range, and whether the metaslab is loaded. */ static uint64_t metaslab_weight(metaslab_t *msp) { metaslab_group_t *mg = msp->ms_group; vdev_t *vd = mg->mg_vd; uint64_t weight, space; ASSERT(MUTEX_HELD(&msp->ms_lock)); /* * This vdev is in the process of being removed so there is nothing * for us to do here. */ if (vd->vdev_removing) { ASSERT0(space_map_allocated(msp->ms_sm)); ASSERT0(vd->vdev_ms_shift); return (0); } /* * The baseline weight is the metaslab's free space. */ space = msp->ms_size - space_map_allocated(msp->ms_sm); msp->ms_fragmentation = metaslab_fragmentation(msp); if (metaslab_fragmentation_factor_enabled && msp->ms_fragmentation != ZFS_FRAG_INVALID) { /* * Use the fragmentation information to inversely scale * down the baseline weight. We need to ensure that we * don't exclude this metaslab completely when it's 100% * fragmented. To avoid this we reduce the fragmented value * by 1. */ space = (space * (100 - (msp->ms_fragmentation - 1))) / 100; /* * If space < SPA_MINBLOCKSIZE, then we will not allocate from * this metaslab again. The fragmentation metric may have * decreased the space to something smaller than * SPA_MINBLOCKSIZE, so reset the space to SPA_MINBLOCKSIZE * so that we can consume any remaining space. */ if (space > 0 && space < SPA_MINBLOCKSIZE) space = SPA_MINBLOCKSIZE; } weight = space; /* * Modern disks have uniform bit density and constant angular velocity. * Therefore, the outer recording zones are faster (higher bandwidth) * than the inner zones by the ratio of outer to inner track diameter, * which is typically around 2:1. We account for this by assigning * higher weight to lower metaslabs (multiplier ranging from 2x to 1x). * In effect, this means that we'll select the metaslab with the most * free bandwidth rather than simply the one with the most free space. */ if (metaslab_lba_weighting_enabled) { weight = 2 * weight - (msp->ms_id * weight) / vd->vdev_ms_count; ASSERT(weight >= space && weight <= 2 * space); } /* * If this metaslab is one we're actively using, adjust its * weight to make it preferable to any inactive metaslab so * we'll polish it off. If the fragmentation on this metaslab * has exceed our threshold, then don't mark it active. */ if (msp->ms_loaded && msp->ms_fragmentation != ZFS_FRAG_INVALID && msp->ms_fragmentation <= zfs_metaslab_fragmentation_threshold) { weight |= (msp->ms_weight & METASLAB_ACTIVE_MASK); } return (weight); } static int metaslab_activate(metaslab_t *msp, uint64_t activation_weight) { ASSERT(MUTEX_HELD(&msp->ms_lock)); if ((msp->ms_weight & METASLAB_ACTIVE_MASK) == 0) { metaslab_load_wait(msp); if (!msp->ms_loaded) { int error = metaslab_load(msp); if (error) { metaslab_group_sort(msp->ms_group, msp, 0); return (error); } } metaslab_group_sort(msp->ms_group, msp, msp->ms_weight | activation_weight); } ASSERT(msp->ms_loaded); ASSERT(msp->ms_weight & METASLAB_ACTIVE_MASK); return (0); } static void metaslab_passivate(metaslab_t *msp, uint64_t size) { /* * If size < SPA_MINBLOCKSIZE, then we will not allocate from * this metaslab again. In that case, it had better be empty, * or we would be leaving space on the table. */ ASSERT(size >= SPA_MINBLOCKSIZE || range_tree_space(msp->ms_tree) == 0); metaslab_group_sort(msp->ms_group, msp, MIN(msp->ms_weight, size)); ASSERT((msp->ms_weight & METASLAB_ACTIVE_MASK) == 0); } static void metaslab_preload(void *arg) { metaslab_t *msp = arg; spa_t *spa = msp->ms_group->mg_vd->vdev_spa; ASSERT(!MUTEX_HELD(&msp->ms_group->mg_lock)); mutex_enter(&msp->ms_lock); metaslab_load_wait(msp); if (!msp->ms_loaded) (void) metaslab_load(msp); /* * Set the ms_access_txg value so that we don't unload it right away. */ msp->ms_access_txg = spa_syncing_txg(spa) + metaslab_unload_delay + 1; mutex_exit(&msp->ms_lock); } static void metaslab_group_preload(metaslab_group_t *mg) { spa_t *spa = mg->mg_vd->vdev_spa; metaslab_t *msp; avl_tree_t *t = &mg->mg_metaslab_tree; int m = 0; if (spa_shutting_down(spa) || !metaslab_preload_enabled) { taskq_wait(mg->mg_taskq); return; } mutex_enter(&mg->mg_lock); /* * Load the next potential metaslabs */ msp = avl_first(t); while (msp != NULL) { metaslab_t *msp_next = AVL_NEXT(t, msp); /* * We preload only the maximum number of metaslabs specified * by metaslab_preload_limit. If a metaslab is being forced * to condense then we preload it too. This will ensure * that force condensing happens in the next txg. */ if (++m > metaslab_preload_limit && !msp->ms_condense_wanted) { msp = msp_next; continue; } /* * We must drop the metaslab group lock here to preserve * lock ordering with the ms_lock (when grabbing both * the mg_lock and the ms_lock, the ms_lock must be taken * first). As a result, it is possible that the ordering * of the metaslabs within the avl tree may change before * we reacquire the lock. The metaslab cannot be removed from * the tree while we're in syncing context so it is safe to * drop the mg_lock here. If the metaslabs are reordered * nothing will break -- we just may end up loading a * less than optimal one. */ mutex_exit(&mg->mg_lock); VERIFY(taskq_dispatch(mg->mg_taskq, metaslab_preload, msp, TQ_SLEEP) != 0); mutex_enter(&mg->mg_lock); msp = msp_next; } mutex_exit(&mg->mg_lock); } /* * Determine if the space map's on-disk footprint is past our tolerance * for inefficiency. We would like to use the following criteria to make * our decision: * * 1. The size of the space map object should not dramatically increase as a * result of writing out the free space range tree. * * 2. The minimal on-disk space map representation is zfs_condense_pct/100 * times the size than the free space range tree representation * (i.e. zfs_condense_pct = 110 and in-core = 1MB, minimal = 1.1.MB). * * 3. The on-disk size of the space map should actually decrease. * * Checking the first condition is tricky since we don't want to walk * the entire AVL tree calculating the estimated on-disk size. Instead we * use the size-ordered range tree in the metaslab and calculate the * size required to write out the largest segment in our free tree. If the * size required to represent that segment on disk is larger than the space * map object then we avoid condensing this map. * * To determine the second criterion we use a best-case estimate and assume * each segment can be represented on-disk as a single 64-bit entry. We refer * to this best-case estimate as the space map's minimal form. * * Unfortunately, we cannot compute the on-disk size of the space map in this * context because we cannot accurately compute the effects of compression, etc. * Instead, we apply the heuristic described in the block comment for * zfs_metaslab_condense_block_threshold - we only condense if the space used * is greater than a threshold number of blocks. */ static boolean_t metaslab_should_condense(metaslab_t *msp) { space_map_t *sm = msp->ms_sm; range_seg_t *rs; uint64_t size, entries, segsz, object_size, optimal_size, record_size; dmu_object_info_t doi; uint64_t vdev_blocksize = 1 << msp->ms_group->mg_vd->vdev_ashift; ASSERT(MUTEX_HELD(&msp->ms_lock)); ASSERT(msp->ms_loaded); /* * Use the ms_size_tree range tree, which is ordered by size, to * obtain the largest segment in the free tree. We always condense * metaslabs that are empty and metaslabs for which a condense * request has been made. */ rs = avl_last(&msp->ms_size_tree); if (rs == NULL || msp->ms_condense_wanted) return (B_TRUE); /* * Calculate the number of 64-bit entries this segment would * require when written to disk. If this single segment would be * larger on-disk than the entire current on-disk structure, then * clearly condensing will increase the on-disk structure size. */ size = (rs->rs_end - rs->rs_start) >> sm->sm_shift; entries = size / (MIN(size, SM_RUN_MAX)); segsz = entries * sizeof (uint64_t); optimal_size = sizeof (uint64_t) * avl_numnodes(&msp->ms_tree->rt_root); object_size = space_map_length(msp->ms_sm); dmu_object_info_from_db(sm->sm_dbuf, &doi); record_size = MAX(doi.doi_data_block_size, vdev_blocksize); return (segsz <= object_size && object_size >= (optimal_size * zfs_condense_pct / 100) && object_size > zfs_metaslab_condense_block_threshold * record_size); } /* * Condense the on-disk space map representation to its minimized form. * The minimized form consists of a small number of allocations followed by * the entries of the free range tree. */ static void metaslab_condense(metaslab_t *msp, uint64_t txg, dmu_tx_t *tx) { spa_t *spa = msp->ms_group->mg_vd->vdev_spa; range_tree_t *freetree = msp->ms_freetree[txg & TXG_MASK]; range_tree_t *condense_tree; space_map_t *sm = msp->ms_sm; ASSERT(MUTEX_HELD(&msp->ms_lock)); ASSERT3U(spa_sync_pass(spa), ==, 1); ASSERT(msp->ms_loaded); spa_dbgmsg(spa, "condensing: txg %llu, msp[%llu] %p, " "smp size %llu, segments %lu, forcing condense=%s", txg, msp->ms_id, msp, space_map_length(msp->ms_sm), avl_numnodes(&msp->ms_tree->rt_root), msp->ms_condense_wanted ? "TRUE" : "FALSE"); msp->ms_condense_wanted = B_FALSE; /* * Create an range tree that is 100% allocated. We remove segments * that have been freed in this txg, any deferred frees that exist, * and any allocation in the future. Removing segments should be * a relatively inexpensive operation since we expect these trees to * have a small number of nodes. */ condense_tree = range_tree_create(NULL, NULL, &msp->ms_lock); range_tree_add(condense_tree, msp->ms_start, msp->ms_size); /* * Remove what's been freed in this txg from the condense_tree. * Since we're in sync_pass 1, we know that all the frees from * this txg are in the freetree. */ range_tree_walk(freetree, range_tree_remove, condense_tree); for (int t = 0; t < TXG_DEFER_SIZE; t++) { range_tree_walk(msp->ms_defertree[t], range_tree_remove, condense_tree); } for (int t = 1; t < TXG_CONCURRENT_STATES; t++) { range_tree_walk(msp->ms_alloctree[(txg + t) & TXG_MASK], range_tree_remove, condense_tree); } /* * We're about to drop the metaslab's lock thus allowing * other consumers to change it's content. Set the * metaslab's ms_condensing flag to ensure that * allocations on this metaslab do not occur while we're * in the middle of committing it to disk. This is only critical * for the ms_tree as all other range trees use per txg * views of their content. */ msp->ms_condensing = B_TRUE; mutex_exit(&msp->ms_lock); space_map_truncate(sm, tx); mutex_enter(&msp->ms_lock); /* * While we would ideally like to create a space_map representation * that consists only of allocation records, doing so can be * prohibitively expensive because the in-core free tree can be * large, and therefore computationally expensive to subtract * from the condense_tree. Instead we sync out two trees, a cheap * allocation only tree followed by the in-core free tree. While not * optimal, this is typically close to optimal, and much cheaper to * compute. */ space_map_write(sm, condense_tree, SM_ALLOC, tx); range_tree_vacate(condense_tree, NULL, NULL); range_tree_destroy(condense_tree); space_map_write(sm, msp->ms_tree, SM_FREE, tx); msp->ms_condensing = B_FALSE; } /* * Write a metaslab to disk in the context of the specified transaction group. */ void metaslab_sync(metaslab_t *msp, uint64_t txg) { metaslab_group_t *mg = msp->ms_group; vdev_t *vd = mg->mg_vd; spa_t *spa = vd->vdev_spa; objset_t *mos = spa_meta_objset(spa); range_tree_t *alloctree = msp->ms_alloctree[txg & TXG_MASK]; range_tree_t **freetree = &msp->ms_freetree[txg & TXG_MASK]; range_tree_t **freed_tree = &msp->ms_freetree[TXG_CLEAN(txg) & TXG_MASK]; dmu_tx_t *tx; uint64_t object = space_map_object(msp->ms_sm); ASSERT(!vd->vdev_ishole); /* * This metaslab has just been added so there's no work to do now. */ if (*freetree == NULL) { ASSERT3P(alloctree, ==, NULL); return; } ASSERT3P(alloctree, !=, NULL); ASSERT3P(*freetree, !=, NULL); ASSERT3P(*freed_tree, !=, NULL); /* * Normally, we don't want to process a metaslab if there * are no allocations or frees to perform. However, if the metaslab * is being forced to condense we need to let it through. */ if (range_tree_space(alloctree) == 0 && range_tree_space(*freetree) == 0 && !msp->ms_condense_wanted) return; /* * The only state that can actually be changing concurrently with * metaslab_sync() is the metaslab's ms_tree. No other thread can * be modifying this txg's alloctree, freetree, freed_tree, or * space_map_phys_t. Therefore, we only hold ms_lock to satify * space_map ASSERTs. We drop it whenever we call into the DMU, * because the DMU can call down to us (e.g. via zio_free()) at * any time. */ tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg); if (msp->ms_sm == NULL) { uint64_t new_object; new_object = space_map_alloc(mos, tx); VERIFY3U(new_object, !=, 0); VERIFY0(space_map_open(&msp->ms_sm, mos, new_object, msp->ms_start, msp->ms_size, vd->vdev_ashift, &msp->ms_lock)); ASSERT(msp->ms_sm != NULL); } mutex_enter(&msp->ms_lock); /* * Note: metaslab_condense() clears the space_map's histogram. * Therefore we must verify and remove this histogram before * condensing. */ metaslab_group_histogram_verify(mg); metaslab_class_histogram_verify(mg->mg_class); metaslab_group_histogram_remove(mg, msp); if (msp->ms_loaded && spa_sync_pass(spa) == 1 && metaslab_should_condense(msp)) { metaslab_condense(msp, txg, tx); } else { space_map_write(msp->ms_sm, alloctree, SM_ALLOC, tx); space_map_write(msp->ms_sm, *freetree, SM_FREE, tx); } if (msp->ms_loaded) { /* * When the space map is loaded, we have an accruate * histogram in the range tree. This gives us an opportunity * to bring the space map's histogram up-to-date so we clear * it first before updating it. */ space_map_histogram_clear(msp->ms_sm); space_map_histogram_add(msp->ms_sm, msp->ms_tree, tx); } else { /* * Since the space map is not loaded we simply update the * exisiting histogram with what was freed in this txg. This * means that the on-disk histogram may not have an accurate * view of the free space but it's close enough to allow * us to make allocation decisions. */ space_map_histogram_add(msp->ms_sm, *freetree, tx); } metaslab_group_histogram_add(mg, msp); metaslab_group_histogram_verify(mg); metaslab_class_histogram_verify(mg->mg_class); /* * For sync pass 1, we avoid traversing this txg's free range tree * and instead will just swap the pointers for freetree and * freed_tree. We can safely do this since the freed_tree is * guaranteed to be empty on the initial pass. */ if (spa_sync_pass(spa) == 1) { range_tree_swap(freetree, freed_tree); } else { range_tree_vacate(*freetree, range_tree_add, *freed_tree); } range_tree_vacate(alloctree, NULL, NULL); ASSERT0(range_tree_space(msp->ms_alloctree[txg & TXG_MASK])); ASSERT0(range_tree_space(msp->ms_freetree[txg & TXG_MASK])); mutex_exit(&msp->ms_lock); if (object != space_map_object(msp->ms_sm)) { object = space_map_object(msp->ms_sm); dmu_write(mos, vd->vdev_ms_array, sizeof (uint64_t) * msp->ms_id, sizeof (uint64_t), &object, tx); } dmu_tx_commit(tx); } /* * Called after a transaction group has completely synced to mark * all of the metaslab's free space as usable. */ void metaslab_sync_done(metaslab_t *msp, uint64_t txg) { metaslab_group_t *mg = msp->ms_group; vdev_t *vd = mg->mg_vd; range_tree_t **freed_tree; range_tree_t **defer_tree; int64_t alloc_delta, defer_delta; ASSERT(!vd->vdev_ishole); mutex_enter(&msp->ms_lock); /* * If this metaslab is just becoming available, initialize its * alloctrees, freetrees, and defertree and add its capacity to * the vdev. */ if (msp->ms_freetree[TXG_CLEAN(txg) & TXG_MASK] == NULL) { for (int t = 0; t < TXG_SIZE; t++) { ASSERT(msp->ms_alloctree[t] == NULL); ASSERT(msp->ms_freetree[t] == NULL); msp->ms_alloctree[t] = range_tree_create(NULL, msp, &msp->ms_lock); msp->ms_freetree[t] = range_tree_create(NULL, msp, &msp->ms_lock); } for (int t = 0; t < TXG_DEFER_SIZE; t++) { ASSERT(msp->ms_defertree[t] == NULL); msp->ms_defertree[t] = range_tree_create(NULL, msp, &msp->ms_lock); } vdev_space_update(vd, 0, 0, msp->ms_size); } freed_tree = &msp->ms_freetree[TXG_CLEAN(txg) & TXG_MASK]; defer_tree = &msp->ms_defertree[txg % TXG_DEFER_SIZE]; alloc_delta = space_map_alloc_delta(msp->ms_sm); defer_delta = range_tree_space(*freed_tree) - range_tree_space(*defer_tree); vdev_space_update(vd, alloc_delta + defer_delta, defer_delta, 0); ASSERT0(range_tree_space(msp->ms_alloctree[txg & TXG_MASK])); ASSERT0(range_tree_space(msp->ms_freetree[txg & TXG_MASK])); /* * If there's a metaslab_load() in progress, wait for it to complete * so that we have a consistent view of the in-core space map. */ metaslab_load_wait(msp); /* * Move the frees from the defer_tree back to the free * range tree (if it's loaded). Swap the freed_tree and the * defer_tree -- this is safe to do because we've just emptied out * the defer_tree. */ range_tree_vacate(*defer_tree, msp->ms_loaded ? range_tree_add : NULL, msp->ms_tree); range_tree_swap(freed_tree, defer_tree); space_map_update(msp->ms_sm); msp->ms_deferspace += defer_delta; ASSERT3S(msp->ms_deferspace, >=, 0); ASSERT3S(msp->ms_deferspace, <=, msp->ms_size); if (msp->ms_deferspace != 0) { /* * Keep syncing this metaslab until all deferred frees * are back in circulation. */ vdev_dirty(vd, VDD_METASLAB, msp, txg + 1); } if (msp->ms_loaded && msp->ms_access_txg < txg) { for (int t = 1; t < TXG_CONCURRENT_STATES; t++) { VERIFY0(range_tree_space( msp->ms_alloctree[(txg + t) & TXG_MASK])); } if (!metaslab_debug_unload) metaslab_unload(msp); } metaslab_group_sort(mg, msp, metaslab_weight(msp)); mutex_exit(&msp->ms_lock); } void metaslab_sync_reassess(metaslab_group_t *mg) { metaslab_group_alloc_update(mg); mg->mg_fragmentation = metaslab_group_fragmentation(mg); /* * Preload the next potential metaslabs */ metaslab_group_preload(mg); } static uint64_t metaslab_distance(metaslab_t *msp, dva_t *dva) { uint64_t ms_shift = msp->ms_group->mg_vd->vdev_ms_shift; uint64_t offset = DVA_GET_OFFSET(dva) >> ms_shift; uint64_t start = msp->ms_id; if (msp->ms_group->mg_vd->vdev_id != DVA_GET_VDEV(dva)) return (1ULL << 63); if (offset < start) return ((start - offset) << ms_shift); if (offset > start) return ((offset - start) << ms_shift); return (0); } static uint64_t metaslab_group_alloc(metaslab_group_t *mg, uint64_t psize, uint64_t asize, uint64_t txg, uint64_t min_distance, dva_t *dva, int d) { spa_t *spa = mg->mg_vd->vdev_spa; metaslab_t *msp = NULL; uint64_t offset = -1ULL; avl_tree_t *t = &mg->mg_metaslab_tree; uint64_t activation_weight; uint64_t target_distance; int i; activation_weight = METASLAB_WEIGHT_PRIMARY; for (i = 0; i < d; i++) { if (DVA_GET_VDEV(&dva[i]) == mg->mg_vd->vdev_id) { activation_weight = METASLAB_WEIGHT_SECONDARY; break; } } for (;;) { boolean_t was_active; mutex_enter(&mg->mg_lock); for (msp = avl_first(t); msp; msp = AVL_NEXT(t, msp)) { if (msp->ms_weight < asize) { spa_dbgmsg(spa, "%s: failed to meet weight " "requirement: vdev %llu, txg %llu, mg %p, " "msp %p, psize %llu, asize %llu, " "weight %llu", spa_name(spa), mg->mg_vd->vdev_id, txg, mg, msp, psize, asize, msp->ms_weight); mutex_exit(&mg->mg_lock); return (-1ULL); } /* * If the selected metaslab is condensing, skip it. */ if (msp->ms_condensing) continue; was_active = msp->ms_weight & METASLAB_ACTIVE_MASK; if (activation_weight == METASLAB_WEIGHT_PRIMARY) break; target_distance = min_distance + (space_map_allocated(msp->ms_sm) != 0 ? 0 : min_distance >> 1); for (i = 0; i < d; i++) if (metaslab_distance(msp, &dva[i]) < target_distance) break; if (i == d) break; } mutex_exit(&mg->mg_lock); if (msp == NULL) return (-1ULL); mutex_enter(&msp->ms_lock); /* * Ensure that the metaslab we have selected is still * capable of handling our request. It's possible that * another thread may have changed the weight while we * were blocked on the metaslab lock. */ if (msp->ms_weight < asize || (was_active && !(msp->ms_weight & METASLAB_ACTIVE_MASK) && activation_weight == METASLAB_WEIGHT_PRIMARY)) { mutex_exit(&msp->ms_lock); continue; } if ((msp->ms_weight & METASLAB_WEIGHT_SECONDARY) && activation_weight == METASLAB_WEIGHT_PRIMARY) { metaslab_passivate(msp, msp->ms_weight & ~METASLAB_ACTIVE_MASK); mutex_exit(&msp->ms_lock); continue; } if (metaslab_activate(msp, activation_weight) != 0) { mutex_exit(&msp->ms_lock); continue; } /* * If this metaslab is currently condensing then pick again as * we can't manipulate this metaslab until it's committed * to disk. */ if (msp->ms_condensing) { mutex_exit(&msp->ms_lock); continue; } if ((offset = metaslab_block_alloc(msp, asize)) != -1ULL) break; metaslab_passivate(msp, metaslab_block_maxsize(msp)); mutex_exit(&msp->ms_lock); } if (range_tree_space(msp->ms_alloctree[txg & TXG_MASK]) == 0) vdev_dirty(mg->mg_vd, VDD_METASLAB, msp, txg); range_tree_add(msp->ms_alloctree[txg & TXG_MASK], offset, asize); msp->ms_access_txg = txg + metaslab_unload_delay; mutex_exit(&msp->ms_lock); return (offset); } /* * Allocate a block for the specified i/o. */ static int metaslab_alloc_dva(spa_t *spa, metaslab_class_t *mc, uint64_t psize, dva_t *dva, int d, dva_t *hintdva, uint64_t txg, int flags) { metaslab_group_t *mg, *rotor; vdev_t *vd; int dshift = 3; int all_zero; int zio_lock = B_FALSE; boolean_t allocatable; uint64_t offset = -1ULL; uint64_t asize; uint64_t distance; ASSERT(!DVA_IS_VALID(&dva[d])); /* * For testing, make some blocks above a certain size be gang blocks. */ if (psize >= metaslab_gang_bang && (ddi_get_lbolt() & 3) == 0) return (SET_ERROR(ENOSPC)); /* * Start at the rotor and loop through all mgs until we find something. * Note that there's no locking on mc_rotor or mc_aliquot because * nothing actually breaks if we miss a few updates -- we just won't * allocate quite as evenly. It all balances out over time. * * If we are doing ditto or log blocks, try to spread them across * consecutive vdevs. If we're forced to reuse a vdev before we've * allocated all of our ditto blocks, then try and spread them out on * that vdev as much as possible. If it turns out to not be possible, * gradually lower our standards until anything becomes acceptable. * Also, allocating on consecutive vdevs (as opposed to random vdevs) * gives us hope of containing our fault domains to something we're * able to reason about. Otherwise, any two top-level vdev failures * will guarantee the loss of data. With consecutive allocation, * only two adjacent top-level vdev failures will result in data loss. * * If we are doing gang blocks (hintdva is non-NULL), try to keep * ourselves on the same vdev as our gang block header. That * way, we can hope for locality in vdev_cache, plus it makes our * fault domains something tractable. */ if (hintdva) { vd = vdev_lookup_top(spa, DVA_GET_VDEV(&hintdva[d])); /* * It's possible the vdev we're using as the hint no * longer exists (i.e. removed). Consult the rotor when * all else fails. */ if (vd != NULL) { mg = vd->vdev_mg; if (flags & METASLAB_HINTBP_AVOID && mg->mg_next != NULL) mg = mg->mg_next; } else { mg = mc->mc_rotor; } } else if (d != 0) { vd = vdev_lookup_top(spa, DVA_GET_VDEV(&dva[d - 1])); mg = vd->vdev_mg->mg_next; } else { mg = mc->mc_rotor; } /* * If the hint put us into the wrong metaslab class, or into a * metaslab group that has been passivated, just follow the rotor. */ if (mg->mg_class != mc || mg->mg_activation_count <= 0) mg = mc->mc_rotor; rotor = mg; top: all_zero = B_TRUE; do { ASSERT(mg->mg_activation_count == 1); vd = mg->mg_vd; /* * Don't allocate from faulted devices. */ if (zio_lock) { spa_config_enter(spa, SCL_ZIO, FTAG, RW_READER); allocatable = vdev_allocatable(vd); spa_config_exit(spa, SCL_ZIO, FTAG); } else { allocatable = vdev_allocatable(vd); } /* * Determine if the selected metaslab group is eligible * for allocations. If we're ganging or have requested * an allocation for the smallest gang block size * then we don't want to avoid allocating to the this * metaslab group. If we're in this condition we should * try to allocate from any device possible so that we * don't inadvertently return ENOSPC and suspend the pool * even though space is still available. */ if (allocatable && CAN_FASTGANG(flags) && psize > SPA_GANGBLOCKSIZE) allocatable = metaslab_group_allocatable(mg); if (!allocatable) goto next; /* * Avoid writing single-copy data to a failing vdev * unless the user instructs us that it is okay. */ if ((vd->vdev_stat.vs_write_errors > 0 || vd->vdev_state < VDEV_STATE_HEALTHY) && d == 0 && dshift == 3 && vd->vdev_children == 0) { all_zero = B_FALSE; goto next; } ASSERT(mg->mg_class == mc); distance = vd->vdev_asize >> dshift; if (distance <= (1ULL << vd->vdev_ms_shift)) distance = 0; else all_zero = B_FALSE; asize = vdev_psize_to_asize(vd, psize); ASSERT(P2PHASE(asize, 1ULL << vd->vdev_ashift) == 0); offset = metaslab_group_alloc(mg, psize, asize, txg, distance, dva, d); if (offset != -1ULL) { /* * If we've just selected this metaslab group, * figure out whether the corresponding vdev is * over- or under-used relative to the pool, * and set an allocation bias to even it out. */ if (mc->mc_aliquot == 0 && metaslab_bias_enabled) { vdev_stat_t *vs = &vd->vdev_stat; int64_t vu, cu; vu = (vs->vs_alloc * 100) / (vs->vs_space + 1); cu = (mc->mc_alloc * 100) / (mc->mc_space + 1); /* * Calculate how much more or less we should * try to allocate from this device during * this iteration around the rotor. * For example, if a device is 80% full * and the pool is 20% full then we should * reduce allocations by 60% on this device. * * mg_bias = (20 - 80) * 512K / 100 = -307K * * This reduces allocations by 307K for this * iteration. */ mg->mg_bias = ((cu - vu) * (int64_t)mg->mg_aliquot) / 100; } else if (!metaslab_bias_enabled) { mg->mg_bias = 0; } if (atomic_add_64_nv(&mc->mc_aliquot, asize) >= mg->mg_aliquot + mg->mg_bias) { mc->mc_rotor = mg->mg_next; mc->mc_aliquot = 0; } DVA_SET_VDEV(&dva[d], vd->vdev_id); DVA_SET_OFFSET(&dva[d], offset); DVA_SET_GANG(&dva[d], !!(flags & METASLAB_GANG_HEADER)); DVA_SET_ASIZE(&dva[d], asize); return (0); } next: mc->mc_rotor = mg->mg_next; mc->mc_aliquot = 0; } while ((mg = mg->mg_next) != rotor); if (!all_zero) { dshift++; ASSERT(dshift < 64); goto top; } if (!allocatable && !zio_lock) { dshift = 3; zio_lock = B_TRUE; goto top; } bzero(&dva[d], sizeof (dva_t)); return (SET_ERROR(ENOSPC)); } /* * Free the block represented by DVA in the context of the specified * transaction group. */ static void metaslab_free_dva(spa_t *spa, const dva_t *dva, uint64_t txg, boolean_t now) { uint64_t vdev = DVA_GET_VDEV(dva); uint64_t offset = DVA_GET_OFFSET(dva); uint64_t size = DVA_GET_ASIZE(dva); vdev_t *vd; metaslab_t *msp; ASSERT(DVA_IS_VALID(dva)); if (txg > spa_freeze_txg(spa)) return; if ((vd = vdev_lookup_top(spa, vdev)) == NULL || (offset >> vd->vdev_ms_shift) >= vd->vdev_ms_count) { cmn_err(CE_WARN, "metaslab_free_dva(): bad DVA %llu:%llu", (u_longlong_t)vdev, (u_longlong_t)offset); ASSERT(0); return; } msp = vd->vdev_ms[offset >> vd->vdev_ms_shift]; if (DVA_GET_GANG(dva)) size = vdev_psize_to_asize(vd, SPA_GANGBLOCKSIZE); mutex_enter(&msp->ms_lock); if (now) { range_tree_remove(msp->ms_alloctree[txg & TXG_MASK], offset, size); VERIFY(!msp->ms_condensing); VERIFY3U(offset, >=, msp->ms_start); VERIFY3U(offset + size, <=, msp->ms_start + msp->ms_size); VERIFY3U(range_tree_space(msp->ms_tree) + size, <=, msp->ms_size); VERIFY0(P2PHASE(offset, 1ULL << vd->vdev_ashift)); VERIFY0(P2PHASE(size, 1ULL << vd->vdev_ashift)); range_tree_add(msp->ms_tree, offset, size); } else { if (range_tree_space(msp->ms_freetree[txg & TXG_MASK]) == 0) vdev_dirty(vd, VDD_METASLAB, msp, txg); range_tree_add(msp->ms_freetree[txg & TXG_MASK], offset, size); } mutex_exit(&msp->ms_lock); } /* * Intent log support: upon opening the pool after a crash, notify the SPA * of blocks that the intent log has allocated for immediate write, but * which are still considered free by the SPA because the last transaction * group didn't commit yet. */ static int metaslab_claim_dva(spa_t *spa, const dva_t *dva, uint64_t txg) { uint64_t vdev = DVA_GET_VDEV(dva); uint64_t offset = DVA_GET_OFFSET(dva); uint64_t size = DVA_GET_ASIZE(dva); vdev_t *vd; metaslab_t *msp; int error = 0; ASSERT(DVA_IS_VALID(dva)); if ((vd = vdev_lookup_top(spa, vdev)) == NULL || (offset >> vd->vdev_ms_shift) >= vd->vdev_ms_count) return (SET_ERROR(ENXIO)); msp = vd->vdev_ms[offset >> vd->vdev_ms_shift]; if (DVA_GET_GANG(dva)) size = vdev_psize_to_asize(vd, SPA_GANGBLOCKSIZE); mutex_enter(&msp->ms_lock); if ((txg != 0 && spa_writeable(spa)) || !msp->ms_loaded) error = metaslab_activate(msp, METASLAB_WEIGHT_SECONDARY); if (error == 0 && !range_tree_contains(msp->ms_tree, offset, size)) error = SET_ERROR(ENOENT); if (error || txg == 0) { /* txg == 0 indicates dry run */ mutex_exit(&msp->ms_lock); return (error); } VERIFY(!msp->ms_condensing); VERIFY0(P2PHASE(offset, 1ULL << vd->vdev_ashift)); VERIFY0(P2PHASE(size, 1ULL << vd->vdev_ashift)); VERIFY3U(range_tree_space(msp->ms_tree) - size, <=, msp->ms_size); range_tree_remove(msp->ms_tree, offset, size); if (spa_writeable(spa)) { /* don't dirty if we're zdb(1M) */ if (range_tree_space(msp->ms_alloctree[txg & TXG_MASK]) == 0) vdev_dirty(vd, VDD_METASLAB, msp, txg); range_tree_add(msp->ms_alloctree[txg & TXG_MASK], offset, size); } mutex_exit(&msp->ms_lock); return (0); } int metaslab_alloc(spa_t *spa, metaslab_class_t *mc, uint64_t psize, blkptr_t *bp, int ndvas, uint64_t txg, blkptr_t *hintbp, int flags) { dva_t *dva = bp->blk_dva; dva_t *hintdva = hintbp->blk_dva; int error = 0; ASSERT(bp->blk_birth == 0); ASSERT(BP_PHYSICAL_BIRTH(bp) == 0); spa_config_enter(spa, SCL_ALLOC, FTAG, RW_READER); if (mc->mc_rotor == NULL) { /* no vdevs in this class */ spa_config_exit(spa, SCL_ALLOC, FTAG); return (SET_ERROR(ENOSPC)); } ASSERT(ndvas > 0 && ndvas <= spa_max_replication(spa)); ASSERT(BP_GET_NDVAS(bp) == 0); ASSERT(hintbp == NULL || ndvas <= BP_GET_NDVAS(hintbp)); for (int d = 0; d < ndvas; d++) { error = metaslab_alloc_dva(spa, mc, psize, dva, d, hintdva, txg, flags); if (error != 0) { for (d--; d >= 0; d--) { metaslab_free_dva(spa, &dva[d], txg, B_TRUE); bzero(&dva[d], sizeof (dva_t)); } spa_config_exit(spa, SCL_ALLOC, FTAG); return (error); } } ASSERT(error == 0); ASSERT(BP_GET_NDVAS(bp) == ndvas); spa_config_exit(spa, SCL_ALLOC, FTAG); BP_SET_BIRTH(bp, txg, txg); return (0); } void metaslab_free(spa_t *spa, const blkptr_t *bp, uint64_t txg, boolean_t now) { const dva_t *dva = bp->blk_dva; int ndvas = BP_GET_NDVAS(bp); ASSERT(!BP_IS_HOLE(bp)); ASSERT(!now || bp->blk_birth >= spa_syncing_txg(spa)); spa_config_enter(spa, SCL_FREE, FTAG, RW_READER); for (int d = 0; d < ndvas; d++) metaslab_free_dva(spa, &dva[d], txg, now); spa_config_exit(spa, SCL_FREE, FTAG); } int metaslab_claim(spa_t *spa, const blkptr_t *bp, uint64_t txg) { const dva_t *dva = bp->blk_dva; int ndvas = BP_GET_NDVAS(bp); int error = 0; ASSERT(!BP_IS_HOLE(bp)); if (txg != 0) { /* * First do a dry run to make sure all DVAs are claimable, * so we don't have to unwind from partial failures below. */ if ((error = metaslab_claim(spa, bp, 0)) != 0) return (error); } spa_config_enter(spa, SCL_ALLOC, FTAG, RW_READER); for (int d = 0; d < ndvas; d++) if ((error = metaslab_claim_dva(spa, &dva[d], txg)) != 0) break; spa_config_exit(spa, SCL_ALLOC, FTAG); ASSERT(error == 0 || txg == 0); return (error); } void metaslab_check_free(spa_t *spa, const blkptr_t *bp) { if ((zfs_flags & ZFS_DEBUG_ZIO_FREE) == 0) return; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); for (int i = 0; i < BP_GET_NDVAS(bp); i++) { uint64_t vdev = DVA_GET_VDEV(&bp->blk_dva[i]); vdev_t *vd = vdev_lookup_top(spa, vdev); uint64_t offset = DVA_GET_OFFSET(&bp->blk_dva[i]); uint64_t size = DVA_GET_ASIZE(&bp->blk_dva[i]); metaslab_t *msp = vd->vdev_ms[offset >> vd->vdev_ms_shift]; if (msp->ms_loaded) range_tree_verify(msp->ms_tree, offset, size); for (int j = 0; j < TXG_SIZE; j++) range_tree_verify(msp->ms_freetree[j], offset, size); for (int j = 0; j < TXG_DEFER_SIZE; j++) range_tree_verify(msp->ms_defertree[j], offset, size); } spa_config_exit(spa, SCL_VDEV, FTAG); } Index: projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/metaslab.h =================================================================== --- projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/metaslab.h (revision 275663) +++ projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/metaslab.h (revision 275664) @@ -1,98 +1,98 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2014 by Delphix. All rights reserved. */ #ifndef _SYS_METASLAB_H #define _SYS_METASLAB_H #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif typedef struct metaslab_ops { uint64_t (*msop_alloc)(metaslab_t *msp, uint64_t size); } metaslab_ops_t; extern metaslab_ops_t *zfs_metaslab_ops; -metaslab_t *metaslab_init(metaslab_group_t *, uint64_t, - uint64_t, uint64_t); +int metaslab_init(metaslab_group_t *, uint64_t, uint64_t, uint64_t, + metaslab_t **); void metaslab_fini(metaslab_t *); void metaslab_load_wait(metaslab_t *); int metaslab_load(metaslab_t *); void metaslab_unload(metaslab_t *); void metaslab_sync(metaslab_t *, uint64_t); void metaslab_sync_done(metaslab_t *, uint64_t); void metaslab_sync_reassess(metaslab_group_t *); uint64_t metaslab_block_maxsize(metaslab_t *); #define METASLAB_HINTBP_FAVOR 0x0 #define METASLAB_HINTBP_AVOID 0x1 #define METASLAB_GANG_HEADER 0x2 #define METASLAB_GANG_CHILD 0x4 #define METASLAB_GANG_AVOID 0x8 int metaslab_alloc(spa_t *, metaslab_class_t *, uint64_t, blkptr_t *, int, uint64_t, blkptr_t *, int); void metaslab_free(spa_t *, const blkptr_t *, uint64_t, boolean_t); int metaslab_claim(spa_t *, const blkptr_t *, uint64_t); void metaslab_check_free(spa_t *, const blkptr_t *); metaslab_class_t *metaslab_class_create(spa_t *, metaslab_ops_t *); void metaslab_class_destroy(metaslab_class_t *); int metaslab_class_validate(metaslab_class_t *); void metaslab_class_histogram_verify(metaslab_class_t *); uint64_t metaslab_class_fragmentation(metaslab_class_t *); uint64_t metaslab_class_expandable_space(metaslab_class_t *); void metaslab_class_space_update(metaslab_class_t *, int64_t, int64_t, int64_t, int64_t); uint64_t metaslab_class_get_alloc(metaslab_class_t *); uint64_t metaslab_class_get_space(metaslab_class_t *); uint64_t metaslab_class_get_dspace(metaslab_class_t *); uint64_t metaslab_class_get_deferred(metaslab_class_t *); uint64_t metaslab_class_get_minblocksize(metaslab_class_t *mc); metaslab_group_t *metaslab_group_create(metaslab_class_t *, vdev_t *); void metaslab_group_destroy(metaslab_group_t *); void metaslab_group_activate(metaslab_group_t *); void metaslab_group_passivate(metaslab_group_t *); uint64_t metaslab_group_get_space(metaslab_group_t *); void metaslab_group_histogram_verify(metaslab_group_t *); uint64_t metaslab_group_fragmentation(metaslab_group_t *); void metaslab_group_histogram_remove(metaslab_group_t *, metaslab_t *); #ifdef __cplusplus } #endif #endif /* _SYS_METASLAB_H */ Index: projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c =================================================================== --- projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c (revision 275663) +++ projects/building-blocks/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev.c (revision 275664) @@ -1,3502 +1,3506 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2011, 2014 by Delphix. All rights reserved. * Copyright 2013 Martin Matuska . All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include SYSCTL_DECL(_vfs_zfs); SYSCTL_NODE(_vfs_zfs, OID_AUTO, vdev, CTLFLAG_RW, 0, "ZFS VDEV"); /* * Virtual device management. */ /* * The limit for ZFS to automatically increase a top-level vdev's ashift * from logical ashift to physical ashift. * * Example: one or more 512B emulation child vdevs * child->vdev_ashift = 9 (512 bytes) * child->vdev_physical_ashift = 12 (4096 bytes) * zfs_max_auto_ashift = 11 (2048 bytes) * zfs_min_auto_ashift = 9 (512 bytes) * * On pool creation or the addition of a new top-level vdev, ZFS will * increase the ashift of the top-level vdev to 2048 as limited by * zfs_max_auto_ashift. * * Example: one or more 512B emulation child vdevs * child->vdev_ashift = 9 (512 bytes) * child->vdev_physical_ashift = 12 (4096 bytes) * zfs_max_auto_ashift = 13 (8192 bytes) * zfs_min_auto_ashift = 9 (512 bytes) * * On pool creation or the addition of a new top-level vdev, ZFS will * increase the ashift of the top-level vdev to 4096 to match the * max vdev_physical_ashift. * * Example: one or more 512B emulation child vdevs * child->vdev_ashift = 9 (512 bytes) * child->vdev_physical_ashift = 9 (512 bytes) * zfs_max_auto_ashift = 13 (8192 bytes) * zfs_min_auto_ashift = 12 (4096 bytes) * * On pool creation or the addition of a new top-level vdev, ZFS will * increase the ashift of the top-level vdev to 4096 to match the * zfs_min_auto_ashift. */ static uint64_t zfs_max_auto_ashift = SPA_MAXASHIFT; static uint64_t zfs_min_auto_ashift = SPA_MINASHIFT; static int sysctl_vfs_zfs_max_auto_ashift(SYSCTL_HANDLER_ARGS) { uint64_t val; int err; val = zfs_max_auto_ashift; err = sysctl_handle_64(oidp, &val, 0, req); if (err != 0 || req->newptr == NULL) return (err); if (val > SPA_MAXASHIFT || val < zfs_min_auto_ashift) return (EINVAL); zfs_max_auto_ashift = val; return (0); } SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift, CTLTYPE_U64 | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof(uint64_t), sysctl_vfs_zfs_max_auto_ashift, "QU", "Max ashift used when optimising for logical -> physical sectors size on " "new top-level vdevs."); static int sysctl_vfs_zfs_min_auto_ashift(SYSCTL_HANDLER_ARGS) { uint64_t val; int err; val = zfs_min_auto_ashift; err = sysctl_handle_64(oidp, &val, 0, req); if (err != 0 || req->newptr == NULL) return (err); if (val < SPA_MINASHIFT || val > zfs_max_auto_ashift) return (EINVAL); zfs_min_auto_ashift = val; return (0); } SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift, CTLTYPE_U64 | CTLFLAG_MPSAFE | CTLFLAG_RW, 0, sizeof(uint64_t), sysctl_vfs_zfs_min_auto_ashift, "QU", "Min ashift used when creating new top-level vdevs."); static vdev_ops_t *vdev_ops_table[] = { &vdev_root_ops, &vdev_raidz_ops, &vdev_mirror_ops, &vdev_replacing_ops, &vdev_spare_ops, #ifdef _KERNEL &vdev_geom_ops, #else &vdev_disk_ops, #endif &vdev_file_ops, &vdev_missing_ops, &vdev_hole_ops, NULL }; /* * When a vdev is added, it will be divided into approximately (but no * more than) this number of metaslabs. */ int metaslabs_per_vdev = 200; SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, metaslabs_per_vdev, CTLFLAG_RDTUN, &metaslabs_per_vdev, 0, "When a vdev is added, how many metaslabs the vdev should be divided into"); /* * Given a vdev type, return the appropriate ops vector. */ static vdev_ops_t * vdev_getops(const char *type) { vdev_ops_t *ops, **opspp; for (opspp = vdev_ops_table; (ops = *opspp) != NULL; opspp++) if (strcmp(ops->vdev_op_type, type) == 0) break; return (ops); } /* * Default asize function: return the MAX of psize with the asize of * all children. This is what's used by anything other than RAID-Z. */ uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize) { uint64_t asize = P2ROUNDUP(psize, 1ULL << vd->vdev_top->vdev_ashift); uint64_t csize; for (int c = 0; c < vd->vdev_children; c++) { csize = vdev_psize_to_asize(vd->vdev_child[c], psize); asize = MAX(asize, csize); } return (asize); } /* * Get the minimum allocatable size. We define the allocatable size as * the vdev's asize rounded to the nearest metaslab. This allows us to * replace or attach devices which don't have the same physical size but * can still satisfy the same number of allocations. */ uint64_t vdev_get_min_asize(vdev_t *vd) { vdev_t *pvd = vd->vdev_parent; /* * If our parent is NULL (inactive spare or cache) or is the root, * just return our own asize. */ if (pvd == NULL) return (vd->vdev_asize); /* * The top-level vdev just returns the allocatable size rounded * to the nearest metaslab. */ if (vd == vd->vdev_top) return (P2ALIGN(vd->vdev_asize, 1ULL << vd->vdev_ms_shift)); /* * The allocatable space for a raidz vdev is N * sizeof(smallest child), * so each child must provide at least 1/Nth of its asize. */ if (pvd->vdev_ops == &vdev_raidz_ops) return (pvd->vdev_min_asize / pvd->vdev_children); return (pvd->vdev_min_asize); } void vdev_set_min_asize(vdev_t *vd) { vd->vdev_min_asize = vdev_get_min_asize(vd); for (int c = 0; c < vd->vdev_children; c++) vdev_set_min_asize(vd->vdev_child[c]); } vdev_t * vdev_lookup_top(spa_t *spa, uint64_t vdev) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); if (vdev < rvd->vdev_children) { ASSERT(rvd->vdev_child[vdev] != NULL); return (rvd->vdev_child[vdev]); } return (NULL); } vdev_t * vdev_lookup_by_guid(vdev_t *vd, uint64_t guid) { vdev_t *mvd; if (vd->vdev_guid == guid) return (vd); for (int c = 0; c < vd->vdev_children; c++) if ((mvd = vdev_lookup_by_guid(vd->vdev_child[c], guid)) != NULL) return (mvd); return (NULL); } void vdev_add_child(vdev_t *pvd, vdev_t *cvd) { size_t oldsize, newsize; uint64_t id = cvd->vdev_id; vdev_t **newchild; ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(cvd->vdev_parent == NULL); cvd->vdev_parent = pvd; if (pvd == NULL) return; ASSERT(id >= pvd->vdev_children || pvd->vdev_child[id] == NULL); oldsize = pvd->vdev_children * sizeof (vdev_t *); pvd->vdev_children = MAX(pvd->vdev_children, id + 1); newsize = pvd->vdev_children * sizeof (vdev_t *); newchild = kmem_zalloc(newsize, KM_SLEEP); if (pvd->vdev_child != NULL) { bcopy(pvd->vdev_child, newchild, oldsize); kmem_free(pvd->vdev_child, oldsize); } pvd->vdev_child = newchild; pvd->vdev_child[id] = cvd; cvd->vdev_top = (pvd->vdev_top ? pvd->vdev_top: cvd); ASSERT(cvd->vdev_top->vdev_parent->vdev_parent == NULL); /* * Walk up all ancestors to update guid sum. */ for (; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum += cvd->vdev_guid_sum; } void vdev_remove_child(vdev_t *pvd, vdev_t *cvd) { int c; uint_t id = cvd->vdev_id; ASSERT(cvd->vdev_parent == pvd); if (pvd == NULL) return; ASSERT(id < pvd->vdev_children); ASSERT(pvd->vdev_child[id] == cvd); pvd->vdev_child[id] = NULL; cvd->vdev_parent = NULL; for (c = 0; c < pvd->vdev_children; c++) if (pvd->vdev_child[c]) break; if (c == pvd->vdev_children) { kmem_free(pvd->vdev_child, c * sizeof (vdev_t *)); pvd->vdev_child = NULL; pvd->vdev_children = 0; } /* * Walk up all ancestors to update guid sum. */ for (; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum -= cvd->vdev_guid_sum; } /* * Remove any holes in the child array. */ void vdev_compact_children(vdev_t *pvd) { vdev_t **newchild, *cvd; int oldc = pvd->vdev_children; int newc; ASSERT(spa_config_held(pvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); for (int c = newc = 0; c < oldc; c++) if (pvd->vdev_child[c]) newc++; newchild = kmem_alloc(newc * sizeof (vdev_t *), KM_SLEEP); for (int c = newc = 0; c < oldc; c++) { if ((cvd = pvd->vdev_child[c]) != NULL) { newchild[newc] = cvd; cvd->vdev_id = newc++; } } kmem_free(pvd->vdev_child, oldc * sizeof (vdev_t *)); pvd->vdev_child = newchild; pvd->vdev_children = newc; } /* * Allocate and minimally initialize a vdev_t. */ vdev_t * vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops) { vdev_t *vd; vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP); if (spa->spa_root_vdev == NULL) { ASSERT(ops == &vdev_root_ops); spa->spa_root_vdev = vd; spa->spa_load_guid = spa_generate_guid(NULL); } if (guid == 0 && ops != &vdev_hole_ops) { if (spa->spa_root_vdev == vd) { /* * The root vdev's guid will also be the pool guid, * which must be unique among all pools. */ guid = spa_generate_guid(NULL); } else { /* * Any other vdev's guid must be unique within the pool. */ guid = spa_generate_guid(spa); } ASSERT(!spa_guid_exists(spa_guid(spa), guid)); } vd->vdev_spa = spa; vd->vdev_id = id; vd->vdev_guid = guid; vd->vdev_guid_sum = guid; vd->vdev_ops = ops; vd->vdev_state = VDEV_STATE_CLOSED; vd->vdev_ishole = (ops == &vdev_hole_ops); mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_stat_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_probe_lock, NULL, MUTEX_DEFAULT, NULL); for (int t = 0; t < DTL_TYPES; t++) { vd->vdev_dtl[t] = range_tree_create(NULL, NULL, &vd->vdev_dtl_lock); } txg_list_create(&vd->vdev_ms_list, offsetof(struct metaslab, ms_txg_node)); txg_list_create(&vd->vdev_dtl_list, offsetof(struct vdev, vdev_dtl_node)); vd->vdev_stat.vs_timestamp = gethrtime(); vdev_queue_init(vd); vdev_cache_init(vd); return (vd); } /* * Allocate a new vdev. The 'alloctype' is used to control whether we are * creating a new vdev or loading an existing one - the behavior is slightly * different for each case. */ int vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id, int alloctype) { vdev_ops_t *ops; char *type; uint64_t guid = 0, islog, nparity; vdev_t *vd; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0) return (SET_ERROR(EINVAL)); if ((ops = vdev_getops(type)) == NULL) return (SET_ERROR(EINVAL)); /* * If this is a load, get the vdev guid from the nvlist. * Otherwise, vdev_alloc_common() will generate one for us. */ if (alloctype == VDEV_ALLOC_LOAD) { uint64_t label_id; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &label_id) || label_id != id) return (SET_ERROR(EINVAL)); if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_SPARE) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_L2CACHE) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_ROOTPOOL) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } /* * The first allocated vdev must be of type 'root'. */ if (ops != &vdev_root_ops && spa->spa_root_vdev == NULL) return (SET_ERROR(EINVAL)); /* * Determine whether we're a log vdev. */ islog = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &islog); if (islog && spa_version(spa) < SPA_VERSION_SLOGS) return (SET_ERROR(ENOTSUP)); if (ops == &vdev_hole_ops && spa_version(spa) < SPA_VERSION_HOLES) return (SET_ERROR(ENOTSUP)); /* * Set the nparity property for RAID-Z vdevs. */ nparity = -1ULL; if (ops == &vdev_raidz_ops) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY, &nparity) == 0) { if (nparity == 0 || nparity > VDEV_RAIDZ_MAXPARITY) return (SET_ERROR(EINVAL)); /* * Previous versions could only support 1 or 2 parity * device. */ if (nparity > 1 && spa_version(spa) < SPA_VERSION_RAIDZ2) return (SET_ERROR(ENOTSUP)); if (nparity > 2 && spa_version(spa) < SPA_VERSION_RAIDZ3) return (SET_ERROR(ENOTSUP)); } else { /* * We require the parity to be specified for SPAs that * support multiple parity levels. */ if (spa_version(spa) >= SPA_VERSION_RAIDZ2) return (SET_ERROR(EINVAL)); /* * Otherwise, we default to 1 parity device for RAID-Z. */ nparity = 1; } } else { nparity = 0; } ASSERT(nparity != -1ULL); vd = vdev_alloc_common(spa, id, guid, ops); vd->vdev_islog = islog; vd->vdev_nparity = nparity; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &vd->vdev_path) == 0) vd->vdev_path = spa_strdup(vd->vdev_path); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &vd->vdev_devid) == 0) vd->vdev_devid = spa_strdup(vd->vdev_devid); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PHYS_PATH, &vd->vdev_physpath) == 0) vd->vdev_physpath = spa_strdup(vd->vdev_physpath); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_FRU, &vd->vdev_fru) == 0) vd->vdev_fru = spa_strdup(vd->vdev_fru); /* * Set the whole_disk property. If it's not specified, leave the value * as -1. */ if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &vd->vdev_wholedisk) != 0) vd->vdev_wholedisk = -1ULL; /* * Look for the 'not present' flag. This will only be set if the device * was not present at the time of import. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &vd->vdev_not_present); /* * Get the alignment requirement. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASHIFT, &vd->vdev_ashift); /* * Retrieve the vdev creation time. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, &vd->vdev_crtxg); /* * If we're a top-level vdev, try to load the allocation parameters. */ if (parent && !parent->vdev_parent && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY, &vd->vdev_ms_array); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT, &vd->vdev_ms_shift); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASIZE, &vd->vdev_asize); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVING, &vd->vdev_removing); } if (parent && !parent->vdev_parent && alloctype != VDEV_ALLOC_ATTACH) { ASSERT(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_ADD || alloctype == VDEV_ALLOC_SPLIT || alloctype == VDEV_ALLOC_ROOTPOOL); vd->vdev_mg = metaslab_group_create(islog ? spa_log_class(spa) : spa_normal_class(spa), vd); } /* * If we're a leaf vdev, try to load the DTL object and other state. */ if (vd->vdev_ops->vdev_op_leaf && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_L2CACHE || alloctype == VDEV_ALLOC_ROOTPOOL)) { if (alloctype == VDEV_ALLOC_LOAD) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DTL, &vd->vdev_dtl_object); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_UNSPARE, &vd->vdev_unspare); } if (alloctype == VDEV_ALLOC_ROOTPOOL) { uint64_t spare = 0; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_SPARE, &spare) == 0 && spare) spa_spare_add(vd); } (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE, &vd->vdev_offline); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG, &vd->vdev_resilver_txg); /* * When importing a pool, we want to ignore the persistent fault * state, as the diagnosis made on another system may not be * valid in the current context. Local vdevs will * remain in the faulted state. */ if (spa_load_state(spa) == SPA_LOAD_OPEN) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED, &vd->vdev_faulted); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DEGRADED, &vd->vdev_degraded); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED, &vd->vdev_removed); if (vd->vdev_faulted || vd->vdev_degraded) { char *aux; vd->vdev_label_aux = VDEV_AUX_ERR_EXCEEDED; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_AUX_STATE, &aux) == 0 && strcmp(aux, "external") == 0) vd->vdev_label_aux = VDEV_AUX_EXTERNAL; } } } /* * Add ourselves to the parent's list of children. */ vdev_add_child(parent, vd); *vdp = vd; return (0); } void vdev_free(vdev_t *vd) { spa_t *spa = vd->vdev_spa; /* * vdev_free() implies closing the vdev first. This is simpler than * trying to ensure complicated semantics for all callers. */ vdev_close(vd); ASSERT(!list_link_active(&vd->vdev_config_dirty_node)); ASSERT(!list_link_active(&vd->vdev_state_dirty_node)); /* * Free all children. */ for (int c = 0; c < vd->vdev_children; c++) vdev_free(vd->vdev_child[c]); ASSERT(vd->vdev_child == NULL); ASSERT(vd->vdev_guid_sum == vd->vdev_guid); /* * Discard allocation state. */ if (vd->vdev_mg != NULL) { vdev_metaslab_fini(vd); metaslab_group_destroy(vd->vdev_mg); } ASSERT0(vd->vdev_stat.vs_space); ASSERT0(vd->vdev_stat.vs_dspace); ASSERT0(vd->vdev_stat.vs_alloc); /* * Remove this vdev from its parent's child list. */ vdev_remove_child(vd->vdev_parent, vd); ASSERT(vd->vdev_parent == NULL); /* * Clean up vdev structure. */ vdev_queue_fini(vd); vdev_cache_fini(vd); if (vd->vdev_path) spa_strfree(vd->vdev_path); if (vd->vdev_devid) spa_strfree(vd->vdev_devid); if (vd->vdev_physpath) spa_strfree(vd->vdev_physpath); if (vd->vdev_fru) spa_strfree(vd->vdev_fru); if (vd->vdev_isspare) spa_spare_remove(vd); if (vd->vdev_isl2cache) spa_l2cache_remove(vd); txg_list_destroy(&vd->vdev_ms_list); txg_list_destroy(&vd->vdev_dtl_list); mutex_enter(&vd->vdev_dtl_lock); space_map_close(vd->vdev_dtl_sm); for (int t = 0; t < DTL_TYPES; t++) { range_tree_vacate(vd->vdev_dtl[t], NULL, NULL); range_tree_destroy(vd->vdev_dtl[t]); } mutex_exit(&vd->vdev_dtl_lock); mutex_destroy(&vd->vdev_dtl_lock); mutex_destroy(&vd->vdev_stat_lock); mutex_destroy(&vd->vdev_probe_lock); if (vd == spa->spa_root_vdev) spa->spa_root_vdev = NULL; kmem_free(vd, sizeof (vdev_t)); } /* * Transfer top-level vdev state from svd to tvd. */ static void vdev_top_transfer(vdev_t *svd, vdev_t *tvd) { spa_t *spa = svd->vdev_spa; metaslab_t *msp; vdev_t *vd; int t; ASSERT(tvd == tvd->vdev_top); tvd->vdev_ms_array = svd->vdev_ms_array; tvd->vdev_ms_shift = svd->vdev_ms_shift; tvd->vdev_ms_count = svd->vdev_ms_count; svd->vdev_ms_array = 0; svd->vdev_ms_shift = 0; svd->vdev_ms_count = 0; if (tvd->vdev_mg) ASSERT3P(tvd->vdev_mg, ==, svd->vdev_mg); tvd->vdev_mg = svd->vdev_mg; tvd->vdev_ms = svd->vdev_ms; svd->vdev_mg = NULL; svd->vdev_ms = NULL; if (tvd->vdev_mg != NULL) tvd->vdev_mg->mg_vd = tvd; tvd->vdev_stat.vs_alloc = svd->vdev_stat.vs_alloc; tvd->vdev_stat.vs_space = svd->vdev_stat.vs_space; tvd->vdev_stat.vs_dspace = svd->vdev_stat.vs_dspace; svd->vdev_stat.vs_alloc = 0; svd->vdev_stat.vs_space = 0; svd->vdev_stat.vs_dspace = 0; for (t = 0; t < TXG_SIZE; t++) { while ((msp = txg_list_remove(&svd->vdev_ms_list, t)) != NULL) (void) txg_list_add(&tvd->vdev_ms_list, msp, t); while ((vd = txg_list_remove(&svd->vdev_dtl_list, t)) != NULL) (void) txg_list_add(&tvd->vdev_dtl_list, vd, t); if (txg_list_remove_this(&spa->spa_vdev_txg_list, svd, t)) (void) txg_list_add(&spa->spa_vdev_txg_list, tvd, t); } if (list_link_active(&svd->vdev_config_dirty_node)) { vdev_config_clean(svd); vdev_config_dirty(tvd); } if (list_link_active(&svd->vdev_state_dirty_node)) { vdev_state_clean(svd); vdev_state_dirty(tvd); } tvd->vdev_deflate_ratio = svd->vdev_deflate_ratio; svd->vdev_deflate_ratio = 0; tvd->vdev_islog = svd->vdev_islog; svd->vdev_islog = 0; } static void vdev_top_update(vdev_t *tvd, vdev_t *vd) { if (vd == NULL) return; vd->vdev_top = tvd; for (int c = 0; c < vd->vdev_children; c++) vdev_top_update(tvd, vd->vdev_child[c]); } /* * Add a mirror/replacing vdev above an existing vdev. */ vdev_t * vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops) { spa_t *spa = cvd->vdev_spa; vdev_t *pvd = cvd->vdev_parent; vdev_t *mvd; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); mvd = vdev_alloc_common(spa, cvd->vdev_id, 0, ops); mvd->vdev_asize = cvd->vdev_asize; mvd->vdev_min_asize = cvd->vdev_min_asize; mvd->vdev_max_asize = cvd->vdev_max_asize; mvd->vdev_ashift = cvd->vdev_ashift; mvd->vdev_logical_ashift = cvd->vdev_logical_ashift; mvd->vdev_physical_ashift = cvd->vdev_physical_ashift; mvd->vdev_state = cvd->vdev_state; mvd->vdev_crtxg = cvd->vdev_crtxg; vdev_remove_child(pvd, cvd); vdev_add_child(pvd, mvd); cvd->vdev_id = mvd->vdev_children; vdev_add_child(mvd, cvd); vdev_top_update(cvd->vdev_top, cvd->vdev_top); if (mvd == mvd->vdev_top) vdev_top_transfer(cvd, mvd); return (mvd); } /* * Remove a 1-way mirror/replacing vdev from the tree. */ void vdev_remove_parent(vdev_t *cvd) { vdev_t *mvd = cvd->vdev_parent; vdev_t *pvd = mvd->vdev_parent; ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(mvd->vdev_children == 1); ASSERT(mvd->vdev_ops == &vdev_mirror_ops || mvd->vdev_ops == &vdev_replacing_ops || mvd->vdev_ops == &vdev_spare_ops); cvd->vdev_ashift = mvd->vdev_ashift; cvd->vdev_logical_ashift = mvd->vdev_logical_ashift; cvd->vdev_physical_ashift = mvd->vdev_physical_ashift; vdev_remove_child(mvd, cvd); vdev_remove_child(pvd, mvd); /* * If cvd will replace mvd as a top-level vdev, preserve mvd's guid. * Otherwise, we could have detached an offline device, and when we * go to import the pool we'll think we have two top-level vdevs, * instead of a different version of the same top-level vdev. */ if (mvd->vdev_top == mvd) { uint64_t guid_delta = mvd->vdev_guid - cvd->vdev_guid; cvd->vdev_orig_guid = cvd->vdev_guid; cvd->vdev_guid += guid_delta; cvd->vdev_guid_sum += guid_delta; } cvd->vdev_id = mvd->vdev_id; vdev_add_child(pvd, cvd); vdev_top_update(cvd->vdev_top, cvd->vdev_top); if (cvd == cvd->vdev_top) vdev_top_transfer(mvd, cvd); ASSERT(mvd->vdev_children == 0); vdev_free(mvd); } int vdev_metaslab_init(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; uint64_t m; uint64_t oldc = vd->vdev_ms_count; uint64_t newc = vd->vdev_asize >> vd->vdev_ms_shift; metaslab_t **mspp; int error; ASSERT(txg == 0 || spa_config_held(spa, SCL_ALLOC, RW_WRITER)); /* * This vdev is not being allocated from yet or is a hole. */ if (vd->vdev_ms_shift == 0) return (0); ASSERT(!vd->vdev_ishole); /* * Compute the raidz-deflation ratio. Note, we hard-code * in 128k (1 << 17) because it is the "typical" blocksize. * Even though SPA_MAXBLOCKSIZE changed, this algorithm can not change, * otherwise it would inconsistently account for existing bp's. */ vd->vdev_deflate_ratio = (1 << 17) / (vdev_psize_to_asize(vd, 1 << 17) >> SPA_MINBLOCKSHIFT); ASSERT(oldc <= newc); mspp = kmem_zalloc(newc * sizeof (*mspp), KM_SLEEP); if (oldc != 0) { bcopy(vd->vdev_ms, mspp, oldc * sizeof (*mspp)); kmem_free(vd->vdev_ms, oldc * sizeof (*mspp)); } vd->vdev_ms = mspp; vd->vdev_ms_count = newc; for (m = oldc; m < newc; m++) { uint64_t object = 0; if (txg == 0) { error = dmu_read(mos, vd->vdev_ms_array, m * sizeof (uint64_t), sizeof (uint64_t), &object, DMU_READ_PREFETCH); if (error) return (error); } - vd->vdev_ms[m] = metaslab_init(vd->vdev_mg, m, object, txg); + + error = metaslab_init(vd->vdev_mg, m, object, txg, + &(vd->vdev_ms[m])); + if (error) + return (error); } if (txg == 0) spa_config_enter(spa, SCL_ALLOC, FTAG, RW_WRITER); /* * If the vdev is being removed we don't activate * the metaslabs since we want to ensure that no new * allocations are performed on this device. */ if (oldc == 0 && !vd->vdev_removing) metaslab_group_activate(vd->vdev_mg); if (txg == 0) spa_config_exit(spa, SCL_ALLOC, FTAG); return (0); } void vdev_metaslab_fini(vdev_t *vd) { uint64_t m; uint64_t count = vd->vdev_ms_count; if (vd->vdev_ms != NULL) { metaslab_group_passivate(vd->vdev_mg); for (m = 0; m < count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp != NULL) metaslab_fini(msp); } kmem_free(vd->vdev_ms, count * sizeof (metaslab_t *)); vd->vdev_ms = NULL; } } typedef struct vdev_probe_stats { boolean_t vps_readable; boolean_t vps_writeable; int vps_flags; } vdev_probe_stats_t; static void vdev_probe_done(zio_t *zio) { spa_t *spa = zio->io_spa; vdev_t *vd = zio->io_vd; vdev_probe_stats_t *vps = zio->io_private; ASSERT(vd->vdev_probe_zio != NULL); if (zio->io_type == ZIO_TYPE_READ) { if (zio->io_error == 0) vps->vps_readable = 1; if (zio->io_error == 0 && spa_writeable(spa)) { zio_nowait(zio_write_phys(vd->vdev_probe_zio, vd, zio->io_offset, zio->io_size, zio->io_data, ZIO_CHECKSUM_OFF, vdev_probe_done, vps, ZIO_PRIORITY_SYNC_WRITE, vps->vps_flags, B_TRUE)); } else { zio_buf_free(zio->io_data, zio->io_size); } } else if (zio->io_type == ZIO_TYPE_WRITE) { if (zio->io_error == 0) vps->vps_writeable = 1; zio_buf_free(zio->io_data, zio->io_size); } else if (zio->io_type == ZIO_TYPE_NULL) { zio_t *pio; vd->vdev_cant_read |= !vps->vps_readable; vd->vdev_cant_write |= !vps->vps_writeable; if (vdev_readable(vd) && (vdev_writeable(vd) || !spa_writeable(spa))) { zio->io_error = 0; } else { ASSERT(zio->io_error != 0); zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE, spa, vd, NULL, 0, 0); zio->io_error = SET_ERROR(ENXIO); } mutex_enter(&vd->vdev_probe_lock); ASSERT(vd->vdev_probe_zio == zio); vd->vdev_probe_zio = NULL; mutex_exit(&vd->vdev_probe_lock); while ((pio = zio_walk_parents(zio)) != NULL) if (!vdev_accessible(vd, pio)) pio->io_error = SET_ERROR(ENXIO); kmem_free(vps, sizeof (*vps)); } } /* * Determine whether this device is accessible. * * Read and write to several known locations: the pad regions of each * vdev label but the first, which we leave alone in case it contains * a VTOC. */ zio_t * vdev_probe(vdev_t *vd, zio_t *zio) { spa_t *spa = vd->vdev_spa; vdev_probe_stats_t *vps = NULL; zio_t *pio; ASSERT(vd->vdev_ops->vdev_op_leaf); /* * Don't probe the probe. */ if (zio && (zio->io_flags & ZIO_FLAG_PROBE)) return (NULL); /* * To prevent 'probe storms' when a device fails, we create * just one probe i/o at a time. All zios that want to probe * this vdev will become parents of the probe io. */ mutex_enter(&vd->vdev_probe_lock); if ((pio = vd->vdev_probe_zio) == NULL) { vps = kmem_zalloc(sizeof (*vps), KM_SLEEP); vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE | ZIO_FLAG_TRYHARD; if (spa_config_held(spa, SCL_ZIO, RW_WRITER)) { /* * vdev_cant_read and vdev_cant_write can only * transition from TRUE to FALSE when we have the * SCL_ZIO lock as writer; otherwise they can only * transition from FALSE to TRUE. This ensures that * any zio looking at these values can assume that * failures persist for the life of the I/O. That's * important because when a device has intermittent * connectivity problems, we want to ensure that * they're ascribed to the device (ENXIO) and not * the zio (EIO). * * Since we hold SCL_ZIO as writer here, clear both * values so the probe can reevaluate from first * principles. */ vps->vps_flags |= ZIO_FLAG_CONFIG_WRITER; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; } vd->vdev_probe_zio = pio = zio_null(NULL, spa, vd, vdev_probe_done, vps, vps->vps_flags | ZIO_FLAG_DONT_PROPAGATE); /* * We can't change the vdev state in this context, so we * kick off an async task to do it on our behalf. */ if (zio != NULL) { vd->vdev_probe_wanted = B_TRUE; spa_async_request(spa, SPA_ASYNC_PROBE); } } if (zio != NULL) zio_add_child(zio, pio); mutex_exit(&vd->vdev_probe_lock); if (vps == NULL) { ASSERT(zio != NULL); return (NULL); } for (int l = 1; l < VDEV_LABELS; l++) { zio_nowait(zio_read_phys(pio, vd, vdev_label_offset(vd->vdev_psize, l, offsetof(vdev_label_t, vl_pad2)), VDEV_PAD_SIZE, zio_buf_alloc(VDEV_PAD_SIZE), ZIO_CHECKSUM_OFF, vdev_probe_done, vps, ZIO_PRIORITY_SYNC_READ, vps->vps_flags, B_TRUE)); } if (zio == NULL) return (pio); zio_nowait(pio); return (NULL); } static void vdev_open_child(void *arg) { vdev_t *vd = arg; vd->vdev_open_thread = curthread; vd->vdev_open_error = vdev_open(vd); vd->vdev_open_thread = NULL; } boolean_t vdev_uses_zvols(vdev_t *vd) { if (vd->vdev_path && strncmp(vd->vdev_path, ZVOL_DIR, strlen(ZVOL_DIR)) == 0) return (B_TRUE); for (int c = 0; c < vd->vdev_children; c++) if (vdev_uses_zvols(vd->vdev_child[c])) return (B_TRUE); return (B_FALSE); } void vdev_open_children(vdev_t *vd) { taskq_t *tq; int children = vd->vdev_children; /* * in order to handle pools on top of zvols, do the opens * in a single thread so that the same thread holds the * spa_namespace_lock */ if (B_TRUE || vdev_uses_zvols(vd)) { for (int c = 0; c < children; c++) vd->vdev_child[c]->vdev_open_error = vdev_open(vd->vdev_child[c]); return; } tq = taskq_create("vdev_open", children, minclsyspri, children, children, TASKQ_PREPOPULATE); for (int c = 0; c < children; c++) VERIFY(taskq_dispatch(tq, vdev_open_child, vd->vdev_child[c], TQ_SLEEP) != 0); taskq_destroy(tq); } /* * Prepare a virtual device for access. */ int vdev_open(vdev_t *vd) { spa_t *spa = vd->vdev_spa; int error; uint64_t osize = 0; uint64_t max_osize = 0; uint64_t asize, max_asize, psize; uint64_t logical_ashift = 0; uint64_t physical_ashift = 0; ASSERT(vd->vdev_open_thread == curthread || spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); ASSERT(vd->vdev_state == VDEV_STATE_CLOSED || vd->vdev_state == VDEV_STATE_CANT_OPEN || vd->vdev_state == VDEV_STATE_OFFLINE); vd->vdev_stat.vs_aux = VDEV_AUX_NONE; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; vd->vdev_notrim = B_FALSE; vd->vdev_min_asize = vdev_get_min_asize(vd); /* * If this vdev is not removed, check its fault status. If it's * faulted, bail out of the open. */ if (!vd->vdev_removed && vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED || vd->vdev_label_aux == VDEV_AUX_EXTERNAL); vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, vd->vdev_label_aux); return (SET_ERROR(ENXIO)); } else if (vd->vdev_offline) { ASSERT(vd->vdev_children == 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, VDEV_AUX_NONE); return (SET_ERROR(ENXIO)); } error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize, &logical_ashift, &physical_ashift); /* * Reset the vdev_reopening flag so that we actually close * the vdev on error. */ vd->vdev_reopening = B_FALSE; if (zio_injection_enabled && error == 0) error = zio_handle_device_injection(vd, NULL, ENXIO); if (error) { if (vd->vdev_removed && vd->vdev_stat.vs_aux != VDEV_AUX_OPEN_FAILED) vd->vdev_removed = B_FALSE; vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, vd->vdev_stat.vs_aux); return (error); } vd->vdev_removed = B_FALSE; /* * Recheck the faulted flag now that we have confirmed that * the vdev is accessible. If we're faulted, bail. */ if (vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED || vd->vdev_label_aux == VDEV_AUX_EXTERNAL); vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, vd->vdev_label_aux); return (SET_ERROR(ENXIO)); } if (vd->vdev_degraded) { ASSERT(vd->vdev_children == 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED, VDEV_AUX_ERR_EXCEEDED); } else { vdev_set_state(vd, B_TRUE, VDEV_STATE_HEALTHY, 0); } /* * For hole or missing vdevs we just return success. */ if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops) return (0); if (zfs_trim_enabled && !vd->vdev_notrim && vd->vdev_ops->vdev_op_leaf) trim_map_create(vd); for (int c = 0; c < vd->vdev_children; c++) { if (vd->vdev_child[c]->vdev_state != VDEV_STATE_HEALTHY) { vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED, VDEV_AUX_NONE); break; } } osize = P2ALIGN(osize, (uint64_t)sizeof (vdev_label_t)); max_osize = P2ALIGN(max_osize, (uint64_t)sizeof (vdev_label_t)); if (vd->vdev_children == 0) { if (osize < SPA_MINDEVSIZE) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_TOO_SMALL); return (SET_ERROR(EOVERFLOW)); } psize = osize; asize = osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE); max_asize = max_osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE); } else { if (vd->vdev_parent != NULL && osize < SPA_MINDEVSIZE - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE)) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_TOO_SMALL); return (SET_ERROR(EOVERFLOW)); } psize = 0; asize = osize; max_asize = max_osize; } vd->vdev_psize = psize; /* * Make sure the allocatable size hasn't shrunk. */ if (asize < vd->vdev_min_asize) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (SET_ERROR(EINVAL)); } vd->vdev_physical_ashift = MAX(physical_ashift, vd->vdev_physical_ashift); vd->vdev_logical_ashift = MAX(logical_ashift, vd->vdev_logical_ashift); vd->vdev_ashift = MAX(vd->vdev_logical_ashift, vd->vdev_ashift); if (vd->vdev_logical_ashift > SPA_MAXASHIFT) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_ASHIFT_TOO_BIG); return (EINVAL); } if (vd->vdev_asize == 0) { /* * This is the first-ever open, so use the computed values. * For testing purposes, a higher ashift can be requested. */ vd->vdev_asize = asize; vd->vdev_max_asize = max_asize; } else { /* * Make sure the alignment requirement hasn't increased. */ if (vd->vdev_ashift > vd->vdev_top->vdev_ashift && vd->vdev_ops->vdev_op_leaf) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (EINVAL); } vd->vdev_max_asize = max_asize; } /* * If all children are healthy and the asize has increased, * then we've experienced dynamic LUN growth. If automatic * expansion is enabled then use the additional space. */ if (vd->vdev_state == VDEV_STATE_HEALTHY && asize > vd->vdev_asize && (vd->vdev_expanding || spa->spa_autoexpand)) vd->vdev_asize = asize; vdev_set_min_asize(vd); /* * Ensure we can issue some IO before declaring the * vdev open for business. */ if (vd->vdev_ops->vdev_op_leaf && (error = zio_wait(vdev_probe(vd, NULL))) != 0) { vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, VDEV_AUX_ERR_EXCEEDED); return (error); } /* * If a leaf vdev has a DTL, and seems healthy, then kick off a * resilver. But don't do this if we are doing a reopen for a scrub, * since this would just restart the scrub we are already doing. */ if (vd->vdev_ops->vdev_op_leaf && !spa->spa_scrub_reopen && vdev_resilver_needed(vd, NULL, NULL)) spa_async_request(spa, SPA_ASYNC_RESILVER); return (0); } /* * Called once the vdevs are all opened, this routine validates the label * contents. This needs to be done before vdev_load() so that we don't * inadvertently do repair I/Os to the wrong device. * * If 'strict' is false ignore the spa guid check. This is necessary because * if the machine crashed during a re-guid the new guid might have been written * to all of the vdev labels, but not the cached config. The strict check * will be performed when the pool is opened again using the mos config. * * This function will only return failure if one of the vdevs indicates that it * has since been destroyed or exported. This is only possible if * /etc/zfs/zpool.cache was readonly at the time. Otherwise, the vdev state * will be updated but the function will return 0. */ int vdev_validate(vdev_t *vd, boolean_t strict) { spa_t *spa = vd->vdev_spa; nvlist_t *label; uint64_t guid = 0, top_guid; uint64_t state; for (int c = 0; c < vd->vdev_children; c++) if (vdev_validate(vd->vdev_child[c], strict) != 0) return (SET_ERROR(EBADF)); /* * If the device has already failed, or was marked offline, don't do * any further validation. Otherwise, label I/O will fail and we will * overwrite the previous state. */ if (vd->vdev_ops->vdev_op_leaf && vdev_readable(vd)) { uint64_t aux_guid = 0; nvlist_t *nvl; uint64_t txg = spa_last_synced_txg(spa) != 0 ? spa_last_synced_txg(spa) : -1ULL; if ((label = vdev_label_read_config(vd, txg)) == NULL) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (0); } /* * Determine if this vdev has been split off into another * pool. If so, then refuse to open it. */ if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_SPLIT_GUID, &aux_guid) == 0 && aux_guid == spa_guid(spa)) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_SPLIT_POOL); nvlist_free(label); return (0); } if (strict && (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &guid) != 0 || guid != spa_guid(spa))) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (0); } if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvl) != 0 || nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_ORIG_GUID, &aux_guid) != 0) aux_guid = 0; /* * If this vdev just became a top-level vdev because its * sibling was detached, it will have adopted the parent's * vdev guid -- but the label may or may not be on disk yet. * Fortunately, either version of the label will have the * same top guid, so if we're a top-level vdev, we can * safely compare to that instead. * * If we split this vdev off instead, then we also check the * original pool's guid. We don't want to consider the vdev * corrupt if it is partway through a split operation. */ if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 || nvlist_lookup_uint64(label, ZPOOL_CONFIG_TOP_GUID, &top_guid) != 0 || ((vd->vdev_guid != guid && vd->vdev_guid != aux_guid) && (vd->vdev_guid != top_guid || vd != vd->vdev_top))) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (0); } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (0); } nvlist_free(label); /* * If this is a verbatim import, no need to check the * state of the pool. */ if (!(spa->spa_import_flags & ZFS_IMPORT_VERBATIM) && spa_load_state(spa) == SPA_LOAD_OPEN && state != POOL_STATE_ACTIVE) return (SET_ERROR(EBADF)); /* * If we were able to open and validate a vdev that was * previously marked permanently unavailable, clear that state * now. */ if (vd->vdev_not_present) vd->vdev_not_present = 0; } return (0); } /* * Close a virtual device. */ void vdev_close(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *pvd = vd->vdev_parent; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); /* * If our parent is reopening, then we are as well, unless we are * going offline. */ if (pvd != NULL && pvd->vdev_reopening) vd->vdev_reopening = (pvd->vdev_reopening && !vd->vdev_offline); vd->vdev_ops->vdev_op_close(vd); vdev_cache_purge(vd); if (vd->vdev_ops->vdev_op_leaf) trim_map_destroy(vd); /* * We record the previous state before we close it, so that if we are * doing a reopen(), we don't generate FMA ereports if we notice that * it's still faulted. */ vd->vdev_prevstate = vd->vdev_state; if (vd->vdev_offline) vd->vdev_state = VDEV_STATE_OFFLINE; else vd->vdev_state = VDEV_STATE_CLOSED; vd->vdev_stat.vs_aux = VDEV_AUX_NONE; } void vdev_hold(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_is_root(spa)); if (spa->spa_state == POOL_STATE_UNINITIALIZED) return; for (int c = 0; c < vd->vdev_children; c++) vdev_hold(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_hold(vd); } void vdev_rele(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_is_root(spa)); for (int c = 0; c < vd->vdev_children; c++) vdev_rele(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_rele(vd); } /* * Reopen all interior vdevs and any unopened leaves. We don't actually * reopen leaf vdevs which had previously been opened as they might deadlock * on the spa_config_lock. Instead we only obtain the leaf's physical size. * If the leaf has never been opened then open it, as usual. */ void vdev_reopen(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); /* set the reopening flag unless we're taking the vdev offline */ vd->vdev_reopening = !vd->vdev_offline; vdev_close(vd); (void) vdev_open(vd); /* * Call vdev_validate() here to make sure we have the same device. * Otherwise, a device with an invalid label could be successfully * opened in response to vdev_reopen(). */ if (vd->vdev_aux) { (void) vdev_validate_aux(vd); if (vdev_readable(vd) && vdev_writeable(vd) && vd->vdev_aux == &spa->spa_l2cache && !l2arc_vdev_present(vd)) l2arc_add_vdev(spa, vd); } else { (void) vdev_validate(vd, B_TRUE); } /* * Reassess parent vdev's health. */ vdev_propagate_state(vd); } int vdev_create(vdev_t *vd, uint64_t txg, boolean_t isreplacing) { int error; /* * Normally, partial opens (e.g. of a mirror) are allowed. * For a create, however, we want to fail the request if * there are any components we can't open. */ error = vdev_open(vd); if (error || vd->vdev_state != VDEV_STATE_HEALTHY) { vdev_close(vd); return (error ? error : ENXIO); } /* * Recursively load DTLs and initialize all labels. */ if ((error = vdev_dtl_load(vd)) != 0 || (error = vdev_label_init(vd, txg, isreplacing ? VDEV_LABEL_REPLACE : VDEV_LABEL_CREATE)) != 0) { vdev_close(vd); return (error); } return (0); } void vdev_metaslab_set_size(vdev_t *vd) { /* * Aim for roughly metaslabs_per_vdev (default 200) metaslabs per vdev. */ vd->vdev_ms_shift = highbit64(vd->vdev_asize / metaslabs_per_vdev); vd->vdev_ms_shift = MAX(vd->vdev_ms_shift, SPA_MAXBLOCKSHIFT); } /* * Maximize performance by inflating the configured ashift for top level * vdevs to be as close to the physical ashift as possible while maintaining * administrator defined limits and ensuring it doesn't go below the * logical ashift. */ void vdev_ashift_optimize(vdev_t *vd) { if (vd == vd->vdev_top) { if (vd->vdev_ashift < vd->vdev_physical_ashift) { vd->vdev_ashift = MIN( MAX(zfs_max_auto_ashift, vd->vdev_ashift), MAX(zfs_min_auto_ashift, vd->vdev_physical_ashift)); } else { /* * Unusual case where logical ashift > physical ashift * so we can't cap the calculated ashift based on max * ashift as that would cause failures. * We still check if we need to increase it to match * the min ashift. */ vd->vdev_ashift = MAX(zfs_min_auto_ashift, vd->vdev_ashift); } } } void vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg) { ASSERT(vd == vd->vdev_top); ASSERT(!vd->vdev_ishole); ASSERT(ISP2(flags)); ASSERT(spa_writeable(vd->vdev_spa)); if (flags & VDD_METASLAB) (void) txg_list_add(&vd->vdev_ms_list, arg, txg); if (flags & VDD_DTL) (void) txg_list_add(&vd->vdev_dtl_list, arg, txg); (void) txg_list_add(&vd->vdev_spa->spa_vdev_txg_list, vd, txg); } void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg) { for (int c = 0; c < vd->vdev_children; c++) vdev_dirty_leaves(vd->vdev_child[c], flags, txg); if (vd->vdev_ops->vdev_op_leaf) vdev_dirty(vd->vdev_top, flags, vd, txg); } /* * DTLs. * * A vdev's DTL (dirty time log) is the set of transaction groups for which * the vdev has less than perfect replication. There are four kinds of DTL: * * DTL_MISSING: txgs for which the vdev has no valid copies of the data * * DTL_PARTIAL: txgs for which data is available, but not fully replicated * * DTL_SCRUB: the txgs that could not be repaired by the last scrub; upon * scrub completion, DTL_SCRUB replaces DTL_MISSING in the range of * txgs that was scrubbed. * * DTL_OUTAGE: txgs which cannot currently be read, whether due to * persistent errors or just some device being offline. * Unlike the other three, the DTL_OUTAGE map is not generally * maintained; it's only computed when needed, typically to * determine whether a device can be detached. * * For leaf vdevs, DTL_MISSING and DTL_PARTIAL are identical: the device * either has the data or it doesn't. * * For interior vdevs such as mirror and RAID-Z the picture is more complex. * A vdev's DTL_PARTIAL is the union of its children's DTL_PARTIALs, because * if any child is less than fully replicated, then so is its parent. * A vdev's DTL_MISSING is a modified union of its children's DTL_MISSINGs, * comprising only those txgs which appear in 'maxfaults' or more children; * those are the txgs we don't have enough replication to read. For example, * double-parity RAID-Z can tolerate up to two missing devices (maxfaults == 2); * thus, its DTL_MISSING consists of the set of txgs that appear in more than * two child DTL_MISSING maps. * * It should be clear from the above that to compute the DTLs and outage maps * for all vdevs, it suffices to know just the leaf vdevs' DTL_MISSING maps. * Therefore, that is all we keep on disk. When loading the pool, or after * a configuration change, we generate all other DTLs from first principles. */ void vdev_dtl_dirty(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size) { range_tree_t *rt = vd->vdev_dtl[t]; ASSERT(t < DTL_TYPES); ASSERT(vd != vd->vdev_spa->spa_root_vdev); ASSERT(spa_writeable(vd->vdev_spa)); mutex_enter(rt->rt_lock); if (!range_tree_contains(rt, txg, size)) range_tree_add(rt, txg, size); mutex_exit(rt->rt_lock); } boolean_t vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size) { range_tree_t *rt = vd->vdev_dtl[t]; boolean_t dirty = B_FALSE; ASSERT(t < DTL_TYPES); ASSERT(vd != vd->vdev_spa->spa_root_vdev); mutex_enter(rt->rt_lock); if (range_tree_space(rt) != 0) dirty = range_tree_contains(rt, txg, size); mutex_exit(rt->rt_lock); return (dirty); } boolean_t vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t t) { range_tree_t *rt = vd->vdev_dtl[t]; boolean_t empty; mutex_enter(rt->rt_lock); empty = (range_tree_space(rt) == 0); mutex_exit(rt->rt_lock); return (empty); } /* * Returns the lowest txg in the DTL range. */ static uint64_t vdev_dtl_min(vdev_t *vd) { range_seg_t *rs; ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); rs = avl_first(&vd->vdev_dtl[DTL_MISSING]->rt_root); return (rs->rs_start - 1); } /* * Returns the highest txg in the DTL. */ static uint64_t vdev_dtl_max(vdev_t *vd) { range_seg_t *rs; ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); rs = avl_last(&vd->vdev_dtl[DTL_MISSING]->rt_root); return (rs->rs_end); } /* * Determine if a resilvering vdev should remove any DTL entries from * its range. If the vdev was resilvering for the entire duration of the * scan then it should excise that range from its DTLs. Otherwise, this * vdev is considered partially resilvered and should leave its DTL * entries intact. The comment in vdev_dtl_reassess() describes how we * excise the DTLs. */ static boolean_t vdev_dtl_should_excise(vdev_t *vd) { spa_t *spa = vd->vdev_spa; dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; ASSERT0(scn->scn_phys.scn_errors); ASSERT0(vd->vdev_children); if (vd->vdev_resilver_txg == 0 || range_tree_space(vd->vdev_dtl[DTL_MISSING]) == 0) return (B_TRUE); /* * When a resilver is initiated the scan will assign the scn_max_txg * value to the highest txg value that exists in all DTLs. If this * device's max DTL is not part of this scan (i.e. it is not in * the range (scn_min_txg, scn_max_txg] then it is not eligible * for excision. */ if (vdev_dtl_max(vd) <= scn->scn_phys.scn_max_txg) { ASSERT3U(scn->scn_phys.scn_min_txg, <=, vdev_dtl_min(vd)); ASSERT3U(scn->scn_phys.scn_min_txg, <, vd->vdev_resilver_txg); ASSERT3U(vd->vdev_resilver_txg, <=, scn->scn_phys.scn_max_txg); return (B_TRUE); } return (B_FALSE); } /* * Reassess DTLs after a config change or scrub completion. */ void vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg, int scrub_done) { spa_t *spa = vd->vdev_spa; avl_tree_t reftree; int minref; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); for (int c = 0; c < vd->vdev_children; c++) vdev_dtl_reassess(vd->vdev_child[c], txg, scrub_txg, scrub_done); if (vd == spa->spa_root_vdev || vd->vdev_ishole || vd->vdev_aux) return; if (vd->vdev_ops->vdev_op_leaf) { dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; mutex_enter(&vd->vdev_dtl_lock); /* * If we've completed a scan cleanly then determine * if this vdev should remove any DTLs. We only want to * excise regions on vdevs that were available during * the entire duration of this scan. */ if (scrub_txg != 0 && (spa->spa_scrub_started || (scn != NULL && scn->scn_phys.scn_errors == 0)) && vdev_dtl_should_excise(vd)) { /* * We completed a scrub up to scrub_txg. If we * did it without rebooting, then the scrub dtl * will be valid, so excise the old region and * fold in the scrub dtl. Otherwise, leave the * dtl as-is if there was an error. * * There's little trick here: to excise the beginning * of the DTL_MISSING map, we put it into a reference * tree and then add a segment with refcnt -1 that * covers the range [0, scrub_txg). This means * that each txg in that range has refcnt -1 or 0. * We then add DTL_SCRUB with a refcnt of 2, so that * entries in the range [0, scrub_txg) will have a * positive refcnt -- either 1 or 2. We then convert * the reference tree into the new DTL_MISSING map. */ space_reftree_create(&reftree); space_reftree_add_map(&reftree, vd->vdev_dtl[DTL_MISSING], 1); space_reftree_add_seg(&reftree, 0, scrub_txg, -1); space_reftree_add_map(&reftree, vd->vdev_dtl[DTL_SCRUB], 2); space_reftree_generate_map(&reftree, vd->vdev_dtl[DTL_MISSING], 1); space_reftree_destroy(&reftree); } range_tree_vacate(vd->vdev_dtl[DTL_PARTIAL], NULL, NULL); range_tree_walk(vd->vdev_dtl[DTL_MISSING], range_tree_add, vd->vdev_dtl[DTL_PARTIAL]); if (scrub_done) range_tree_vacate(vd->vdev_dtl[DTL_SCRUB], NULL, NULL); range_tree_vacate(vd->vdev_dtl[DTL_OUTAGE], NULL, NULL); if (!vdev_readable(vd)) range_tree_add(vd->vdev_dtl[DTL_OUTAGE], 0, -1ULL); else range_tree_walk(vd->vdev_dtl[DTL_MISSING], range_tree_add, vd->vdev_dtl[DTL_OUTAGE]); /* * If the vdev was resilvering and no longer has any * DTLs then reset its resilvering flag and dirty * the top level so that we persist the change. */ if (vd->vdev_resilver_txg != 0 && range_tree_space(vd->vdev_dtl[DTL_MISSING]) == 0 && range_tree_space(vd->vdev_dtl[DTL_OUTAGE]) == 0) { vd->vdev_resilver_txg = 0; vdev_config_dirty(vd->vdev_top); } mutex_exit(&vd->vdev_dtl_lock); if (txg != 0) vdev_dirty(vd->vdev_top, VDD_DTL, vd, txg); return; } mutex_enter(&vd->vdev_dtl_lock); for (int t = 0; t < DTL_TYPES; t++) { /* account for child's outage in parent's missing map */ int s = (t == DTL_MISSING) ? DTL_OUTAGE: t; if (t == DTL_SCRUB) continue; /* leaf vdevs only */ if (t == DTL_PARTIAL) minref = 1; /* i.e. non-zero */ else if (vd->vdev_nparity != 0) minref = vd->vdev_nparity + 1; /* RAID-Z */ else minref = vd->vdev_children; /* any kind of mirror */ space_reftree_create(&reftree); for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; mutex_enter(&cvd->vdev_dtl_lock); space_reftree_add_map(&reftree, cvd->vdev_dtl[s], 1); mutex_exit(&cvd->vdev_dtl_lock); } space_reftree_generate_map(&reftree, vd->vdev_dtl[t], minref); space_reftree_destroy(&reftree); } mutex_exit(&vd->vdev_dtl_lock); } int vdev_dtl_load(vdev_t *vd) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; int error = 0; if (vd->vdev_ops->vdev_op_leaf && vd->vdev_dtl_object != 0) { ASSERT(!vd->vdev_ishole); error = space_map_open(&vd->vdev_dtl_sm, mos, vd->vdev_dtl_object, 0, -1ULL, 0, &vd->vdev_dtl_lock); if (error) return (error); ASSERT(vd->vdev_dtl_sm != NULL); mutex_enter(&vd->vdev_dtl_lock); /* * Now that we've opened the space_map we need to update * the in-core DTL. */ space_map_update(vd->vdev_dtl_sm); error = space_map_load(vd->vdev_dtl_sm, vd->vdev_dtl[DTL_MISSING], SM_ALLOC); mutex_exit(&vd->vdev_dtl_lock); return (error); } for (int c = 0; c < vd->vdev_children; c++) { error = vdev_dtl_load(vd->vdev_child[c]); if (error != 0) break; } return (error); } void vdev_dtl_sync(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; range_tree_t *rt = vd->vdev_dtl[DTL_MISSING]; objset_t *mos = spa->spa_meta_objset; range_tree_t *rtsync; kmutex_t rtlock; dmu_tx_t *tx; uint64_t object = space_map_object(vd->vdev_dtl_sm); ASSERT(!vd->vdev_ishole); ASSERT(vd->vdev_ops->vdev_op_leaf); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); if (vd->vdev_detached || vd->vdev_top->vdev_removing) { mutex_enter(&vd->vdev_dtl_lock); space_map_free(vd->vdev_dtl_sm, tx); space_map_close(vd->vdev_dtl_sm); vd->vdev_dtl_sm = NULL; mutex_exit(&vd->vdev_dtl_lock); dmu_tx_commit(tx); return; } if (vd->vdev_dtl_sm == NULL) { uint64_t new_object; new_object = space_map_alloc(mos, tx); VERIFY3U(new_object, !=, 0); VERIFY0(space_map_open(&vd->vdev_dtl_sm, mos, new_object, 0, -1ULL, 0, &vd->vdev_dtl_lock)); ASSERT(vd->vdev_dtl_sm != NULL); } bzero(&rtlock, sizeof(rtlock)); mutex_init(&rtlock, NULL, MUTEX_DEFAULT, NULL); rtsync = range_tree_create(NULL, NULL, &rtlock); mutex_enter(&rtlock); mutex_enter(&vd->vdev_dtl_lock); range_tree_walk(rt, range_tree_add, rtsync); mutex_exit(&vd->vdev_dtl_lock); space_map_truncate(vd->vdev_dtl_sm, tx); space_map_write(vd->vdev_dtl_sm, rtsync, SM_ALLOC, tx); range_tree_vacate(rtsync, NULL, NULL); range_tree_destroy(rtsync); mutex_exit(&rtlock); mutex_destroy(&rtlock); /* * If the object for the space map has changed then dirty * the top level so that we update the config. */ if (object != space_map_object(vd->vdev_dtl_sm)) { zfs_dbgmsg("txg %llu, spa %s, DTL old object %llu, " "new object %llu", txg, spa_name(spa), object, space_map_object(vd->vdev_dtl_sm)); vdev_config_dirty(vd->vdev_top); } dmu_tx_commit(tx); mutex_enter(&vd->vdev_dtl_lock); space_map_update(vd->vdev_dtl_sm); mutex_exit(&vd->vdev_dtl_lock); } /* * Determine whether the specified vdev can be offlined/detached/removed * without losing data. */ boolean_t vdev_dtl_required(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *tvd = vd->vdev_top; uint8_t cant_read = vd->vdev_cant_read; boolean_t required; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); if (vd == spa->spa_root_vdev || vd == tvd) return (B_TRUE); /* * Temporarily mark the device as unreadable, and then determine * whether this results in any DTL outages in the top-level vdev. * If not, we can safely offline/detach/remove the device. */ vd->vdev_cant_read = B_TRUE; vdev_dtl_reassess(tvd, 0, 0, B_FALSE); required = !vdev_dtl_empty(tvd, DTL_OUTAGE); vd->vdev_cant_read = cant_read; vdev_dtl_reassess(tvd, 0, 0, B_FALSE); if (!required && zio_injection_enabled) required = !!zio_handle_device_injection(vd, NULL, ECHILD); return (required); } /* * Determine if resilver is needed, and if so the txg range. */ boolean_t vdev_resilver_needed(vdev_t *vd, uint64_t *minp, uint64_t *maxp) { boolean_t needed = B_FALSE; uint64_t thismin = UINT64_MAX; uint64_t thismax = 0; if (vd->vdev_children == 0) { mutex_enter(&vd->vdev_dtl_lock); if (range_tree_space(vd->vdev_dtl[DTL_MISSING]) != 0 && vdev_writeable(vd)) { thismin = vdev_dtl_min(vd); thismax = vdev_dtl_max(vd); needed = B_TRUE; } mutex_exit(&vd->vdev_dtl_lock); } else { for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; uint64_t cmin, cmax; if (vdev_resilver_needed(cvd, &cmin, &cmax)) { thismin = MIN(thismin, cmin); thismax = MAX(thismax, cmax); needed = B_TRUE; } } } if (needed && minp) { *minp = thismin; *maxp = thismax; } return (needed); } void vdev_load(vdev_t *vd) { /* * Recursively load all children. */ for (int c = 0; c < vd->vdev_children; c++) vdev_load(vd->vdev_child[c]); /* * If this is a top-level vdev, initialize its metaslabs. */ if (vd == vd->vdev_top && !vd->vdev_ishole && (vd->vdev_ashift == 0 || vd->vdev_asize == 0 || vdev_metaslab_init(vd, 0) != 0)) vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); /* * If this is a leaf vdev, load its DTL. */ if (vd->vdev_ops->vdev_op_leaf && vdev_dtl_load(vd) != 0) vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); } /* * The special vdev case is used for hot spares and l2cache devices. Its * sole purpose it to set the vdev state for the associated vdev. To do this, * we make sure that we can open the underlying device, then try to read the * label, and make sure that the label is sane and that it hasn't been * repurposed to another pool. */ int vdev_validate_aux(vdev_t *vd) { nvlist_t *label; uint64_t guid, version; uint64_t state; if (!vdev_readable(vd)) return (0); if ((label = vdev_label_read_config(vd, -1ULL)) == NULL) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); return (-1); } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_VERSION, &version) != 0 || !SPA_VERSION_IS_SUPPORTED(version) || nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 || guid != vd->vdev_guid || nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (-1); } /* * We don't actually check the pool state here. If it's in fact in * use by another pool, we update this fact on the fly when requested. */ nvlist_free(label); return (0); } void vdev_remove(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; dmu_tx_t *tx; tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg); if (vd->vdev_ms != NULL) { metaslab_group_t *mg = vd->vdev_mg; metaslab_group_histogram_verify(mg); metaslab_class_histogram_verify(mg->mg_class); for (int m = 0; m < vd->vdev_ms_count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp == NULL || msp->ms_sm == NULL) continue; mutex_enter(&msp->ms_lock); /* * If the metaslab was not loaded when the vdev * was removed then the histogram accounting may * not be accurate. Update the histogram information * here so that we ensure that the metaslab group * and metaslab class are up-to-date. */ metaslab_group_histogram_remove(mg, msp); VERIFY0(space_map_allocated(msp->ms_sm)); space_map_free(msp->ms_sm, tx); space_map_close(msp->ms_sm); msp->ms_sm = NULL; mutex_exit(&msp->ms_lock); } metaslab_group_histogram_verify(mg); metaslab_class_histogram_verify(mg->mg_class); for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) ASSERT0(mg->mg_histogram[i]); } if (vd->vdev_ms_array) { (void) dmu_object_free(mos, vd->vdev_ms_array, tx); vd->vdev_ms_array = 0; } dmu_tx_commit(tx); } void vdev_sync_done(vdev_t *vd, uint64_t txg) { metaslab_t *msp; boolean_t reassess = !txg_list_empty(&vd->vdev_ms_list, TXG_CLEAN(txg)); ASSERT(!vd->vdev_ishole); while (msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg))) metaslab_sync_done(msp, txg); if (reassess) metaslab_sync_reassess(vd->vdev_mg); } void vdev_sync(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; vdev_t *lvd; metaslab_t *msp; dmu_tx_t *tx; ASSERT(!vd->vdev_ishole); if (vd->vdev_ms_array == 0 && vd->vdev_ms_shift != 0) { ASSERT(vd == vd->vdev_top); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); vd->vdev_ms_array = dmu_object_alloc(spa->spa_meta_objset, DMU_OT_OBJECT_ARRAY, 0, DMU_OT_NONE, 0, tx); ASSERT(vd->vdev_ms_array != 0); vdev_config_dirty(vd); dmu_tx_commit(tx); } /* * Remove the metadata associated with this vdev once it's empty. */ if (vd->vdev_stat.vs_alloc == 0 && vd->vdev_removing) vdev_remove(vd, txg); while ((msp = txg_list_remove(&vd->vdev_ms_list, txg)) != NULL) { metaslab_sync(msp, txg); (void) txg_list_add(&vd->vdev_ms_list, msp, TXG_CLEAN(txg)); } while ((lvd = txg_list_remove(&vd->vdev_dtl_list, txg)) != NULL) vdev_dtl_sync(lvd, txg); (void) txg_list_add(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg)); } uint64_t vdev_psize_to_asize(vdev_t *vd, uint64_t psize) { return (vd->vdev_ops->vdev_op_asize(vd, psize)); } /* * Mark the given vdev faulted. A faulted vdev behaves as if the device could * not be opened, and no I/O is attempted. */ int vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux) { vdev_t *vd, *tvd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; /* * We don't directly use the aux state here, but if we do a * vdev_reopen(), we need this value to be present to remember why we * were faulted. */ vd->vdev_label_aux = aux; /* * Faulted state takes precedence over degraded. */ vd->vdev_delayed_close = B_FALSE; vd->vdev_faulted = 1ULL; vd->vdev_degraded = 0ULL; vdev_set_state(vd, B_FALSE, VDEV_STATE_FAULTED, aux); /* * If this device has the only valid copy of the data, then * back off and simply mark the vdev as degraded instead. */ if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) { vd->vdev_degraded = 1ULL; vd->vdev_faulted = 0ULL; /* * If we reopen the device and it's not dead, only then do we * mark it degraded. */ vdev_reopen(tvd); if (vdev_readable(vd)) vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux); } return (spa_vdev_state_exit(spa, vd, 0)); } /* * Mark the given vdev degraded. A degraded vdev is purely an indication to the * user that something is wrong. The vdev continues to operate as normal as far * as I/O is concerned. */ int vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux) { vdev_t *vd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); /* * If the vdev is already faulted, then don't do anything. */ if (vd->vdev_faulted || vd->vdev_degraded) return (spa_vdev_state_exit(spa, NULL, 0)); vd->vdev_degraded = 1ULL; if (!vdev_is_dead(vd)) vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux); return (spa_vdev_state_exit(spa, vd, 0)); } /* * Online the given vdev. * * If 'ZFS_ONLINE_UNSPARE' is set, it implies two things. First, any attached * spare device should be detached when the device finishes resilvering. * Second, the online should be treated like a 'test' online case, so no FMA * events are generated if the device fails to open. */ int vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *newstate) { vdev_t *vd, *tvd, *pvd, *rvd = spa->spa_root_vdev; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; vd->vdev_offline = B_FALSE; vd->vdev_tmpoffline = B_FALSE; vd->vdev_checkremove = !!(flags & ZFS_ONLINE_CHECKREMOVE); vd->vdev_forcefault = !!(flags & ZFS_ONLINE_FORCEFAULT); /* XXX - L2ARC 1.0 does not support expansion */ if (!vd->vdev_aux) { for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) pvd->vdev_expanding = !!(flags & ZFS_ONLINE_EXPAND); } vdev_reopen(tvd); vd->vdev_checkremove = vd->vdev_forcefault = B_FALSE; if (!vd->vdev_aux) { for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) pvd->vdev_expanding = B_FALSE; } if (newstate) *newstate = vd->vdev_state; if ((flags & ZFS_ONLINE_UNSPARE) && !vdev_is_dead(vd) && vd->vdev_parent && vd->vdev_parent->vdev_ops == &vdev_spare_ops && vd->vdev_parent->vdev_child[0] == vd) vd->vdev_unspare = B_TRUE; if ((flags & ZFS_ONLINE_EXPAND) || spa->spa_autoexpand) { /* XXX - L2ARC 1.0 does not support expansion */ if (vd->vdev_aux) return (spa_vdev_state_exit(spa, vd, ENOTSUP)); spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } return (spa_vdev_state_exit(spa, vd, 0)); } static int vdev_offline_locked(spa_t *spa, uint64_t guid, uint64_t flags) { vdev_t *vd, *tvd; int error = 0; uint64_t generation; metaslab_group_t *mg; top: spa_vdev_state_enter(spa, SCL_ALLOC); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; mg = tvd->vdev_mg; generation = spa->spa_config_generation + 1; /* * If the device isn't already offline, try to offline it. */ if (!vd->vdev_offline) { /* * If this device has the only valid copy of some data, * don't allow it to be offlined. Log devices are always * expendable. */ if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) return (spa_vdev_state_exit(spa, NULL, EBUSY)); /* * If the top-level is a slog and it has had allocations * then proceed. We check that the vdev's metaslab group * is not NULL since it's possible that we may have just * added this vdev but not yet initialized its metaslabs. */ if (tvd->vdev_islog && mg != NULL) { /* * Prevent any future allocations. */ metaslab_group_passivate(mg); (void) spa_vdev_state_exit(spa, vd, 0); error = spa_offline_log(spa); spa_vdev_state_enter(spa, SCL_ALLOC); /* * Check to see if the config has changed. */ if (error || generation != spa->spa_config_generation) { metaslab_group_activate(mg); if (error) return (spa_vdev_state_exit(spa, vd, error)); (void) spa_vdev_state_exit(spa, vd, 0); goto top; } ASSERT0(tvd->vdev_stat.vs_alloc); } /* * Offline this device and reopen its top-level vdev. * If the top-level vdev is a log device then just offline * it. Otherwise, if this action results in the top-level * vdev becoming unusable, undo it and fail the request. */ vd->vdev_offline = B_TRUE; vdev_reopen(tvd); if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_is_dead(tvd)) { vd->vdev_offline = B_FALSE; vdev_reopen(tvd); return (spa_vdev_state_exit(spa, NULL, EBUSY)); } /* * Add the device back into the metaslab rotor so that * once we online the device it's open for business. */ if (tvd->vdev_islog && mg != NULL) metaslab_group_activate(mg); } vd->vdev_tmpoffline = !!(flags & ZFS_OFFLINE_TEMPORARY); return (spa_vdev_state_exit(spa, vd, 0)); } int vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags) { int error; mutex_enter(&spa->spa_vdev_top_lock); error = vdev_offline_locked(spa, guid, flags); mutex_exit(&spa->spa_vdev_top_lock); return (error); } /* * Clear the error counts associated with this vdev. Unlike vdev_online() and * vdev_offline(), we assume the spa config is locked. We also clear all * children. If 'vd' is NULL, then the user wants to clear all vdevs. */ void vdev_clear(spa_t *spa, vdev_t *vd) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); if (vd == NULL) vd = rvd; vd->vdev_stat.vs_read_errors = 0; vd->vdev_stat.vs_write_errors = 0; vd->vdev_stat.vs_checksum_errors = 0; for (int c = 0; c < vd->vdev_children; c++) vdev_clear(spa, vd->vdev_child[c]); if (vd == rvd) { for (int c = 0; c < spa->spa_l2cache.sav_count; c++) vdev_clear(spa, spa->spa_l2cache.sav_vdevs[c]); for (int c = 0; c < spa->spa_spares.sav_count; c++) vdev_clear(spa, spa->spa_spares.sav_vdevs[c]); } /* * If we're in the FAULTED state or have experienced failed I/O, then * clear the persistent state and attempt to reopen the device. We * also mark the vdev config dirty, so that the new faulted state is * written out to disk. */ if (vd->vdev_faulted || vd->vdev_degraded || !vdev_readable(vd) || !vdev_writeable(vd)) { /* * When reopening in reponse to a clear event, it may be due to * a fmadm repair request. In this case, if the device is * still broken, we want to still post the ereport again. */ vd->vdev_forcefault = B_TRUE; vd->vdev_faulted = vd->vdev_degraded = 0ULL; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; vdev_reopen(vd == rvd ? rvd : vd->vdev_top); vd->vdev_forcefault = B_FALSE; if (vd != rvd && vdev_writeable(vd->vdev_top)) vdev_state_dirty(vd->vdev_top); if (vd->vdev_aux == NULL && !vdev_is_dead(vd)) spa_async_request(spa, SPA_ASYNC_RESILVER); spa_event_notify(spa, vd, ESC_ZFS_VDEV_CLEAR); } /* * When clearing a FMA-diagnosed fault, we always want to * unspare the device, as we assume that the original spare was * done in response to the FMA fault. */ if (!vdev_is_dead(vd) && vd->vdev_parent != NULL && vd->vdev_parent->vdev_ops == &vdev_spare_ops && vd->vdev_parent->vdev_child[0] == vd) vd->vdev_unspare = B_TRUE; } boolean_t vdev_is_dead(vdev_t *vd) { /* * Holes and missing devices are always considered "dead". * This simplifies the code since we don't have to check for * these types of devices in the various code paths. * Instead we rely on the fact that we skip over dead devices * before issuing I/O to them. */ return (vd->vdev_state < VDEV_STATE_DEGRADED || vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops); } boolean_t vdev_readable(vdev_t *vd) { return (!vdev_is_dead(vd) && !vd->vdev_cant_read); } boolean_t vdev_writeable(vdev_t *vd) { return (!vdev_is_dead(vd) && !vd->vdev_cant_write); } boolean_t vdev_allocatable(vdev_t *vd) { uint64_t state = vd->vdev_state; /* * We currently allow allocations from vdevs which may be in the * process of reopening (i.e. VDEV_STATE_CLOSED). If the device * fails to reopen then we'll catch it later when we're holding * the proper locks. Note that we have to get the vdev state * in a local variable because although it changes atomically, * we're asking two separate questions about it. */ return (!(state < VDEV_STATE_DEGRADED && state != VDEV_STATE_CLOSED) && !vd->vdev_cant_write && !vd->vdev_ishole); } boolean_t vdev_accessible(vdev_t *vd, zio_t *zio) { ASSERT(zio->io_vd == vd); if (vdev_is_dead(vd) || vd->vdev_remove_wanted) return (B_FALSE); if (zio->io_type == ZIO_TYPE_READ) return (!vd->vdev_cant_read); if (zio->io_type == ZIO_TYPE_WRITE) return (!vd->vdev_cant_write); return (B_TRUE); } /* * Get statistics for the given vdev. */ void vdev_get_stats(vdev_t *vd, vdev_stat_t *vs) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); mutex_enter(&vd->vdev_stat_lock); bcopy(&vd->vdev_stat, vs, sizeof (*vs)); vs->vs_timestamp = gethrtime() - vs->vs_timestamp; vs->vs_state = vd->vdev_state; vs->vs_rsize = vdev_get_min_asize(vd); if (vd->vdev_ops->vdev_op_leaf) vs->vs_rsize += VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE; vs->vs_esize = vd->vdev_max_asize - vd->vdev_asize; vs->vs_configured_ashift = vd->vdev_top != NULL ? vd->vdev_top->vdev_ashift : vd->vdev_ashift; vs->vs_logical_ashift = vd->vdev_logical_ashift; vs->vs_physical_ashift = vd->vdev_physical_ashift; if (vd->vdev_aux == NULL && vd == vd->vdev_top && !vd->vdev_ishole) { vs->vs_fragmentation = vd->vdev_mg->mg_fragmentation; } /* * If we're getting stats on the root vdev, aggregate the I/O counts * over all top-level vdevs (i.e. the direct children of the root). */ if (vd == rvd) { for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *cvd = rvd->vdev_child[c]; vdev_stat_t *cvs = &cvd->vdev_stat; for (int t = 0; t < ZIO_TYPES; t++) { vs->vs_ops[t] += cvs->vs_ops[t]; vs->vs_bytes[t] += cvs->vs_bytes[t]; } cvs->vs_scan_removing = cvd->vdev_removing; } } mutex_exit(&vd->vdev_stat_lock); } void vdev_clear_stats(vdev_t *vd) { mutex_enter(&vd->vdev_stat_lock); vd->vdev_stat.vs_space = 0; vd->vdev_stat.vs_dspace = 0; vd->vdev_stat.vs_alloc = 0; mutex_exit(&vd->vdev_stat_lock); } void vdev_scan_stat_init(vdev_t *vd) { vdev_stat_t *vs = &vd->vdev_stat; for (int c = 0; c < vd->vdev_children; c++) vdev_scan_stat_init(vd->vdev_child[c]); mutex_enter(&vd->vdev_stat_lock); vs->vs_scan_processed = 0; mutex_exit(&vd->vdev_stat_lock); } void vdev_stat_update(zio_t *zio, uint64_t psize) { spa_t *spa = zio->io_spa; vdev_t *rvd = spa->spa_root_vdev; vdev_t *vd = zio->io_vd ? zio->io_vd : rvd; vdev_t *pvd; uint64_t txg = zio->io_txg; vdev_stat_t *vs = &vd->vdev_stat; zio_type_t type = zio->io_type; int flags = zio->io_flags; /* * If this i/o is a gang leader, it didn't do any actual work. */ if (zio->io_gang_tree) return; if (zio->io_error == 0) { /* * If this is a root i/o, don't count it -- we've already * counted the top-level vdevs, and vdev_get_stats() will * aggregate them when asked. This reduces contention on * the root vdev_stat_lock and implicitly handles blocks * that compress away to holes, for which there is no i/o. * (Holes never create vdev children, so all the counters * remain zero, which is what we want.) * * Note: this only applies to successful i/o (io_error == 0) * because unlike i/o counts, errors are not additive. * When reading a ditto block, for example, failure of * one top-level vdev does not imply a root-level error. */ if (vd == rvd) return; ASSERT(vd == zio->io_vd); if (flags & ZIO_FLAG_IO_BYPASS) return; mutex_enter(&vd->vdev_stat_lock); if (flags & ZIO_FLAG_IO_REPAIR) { if (flags & ZIO_FLAG_SCAN_THREAD) { dsl_scan_phys_t *scn_phys = &spa->spa_dsl_pool->dp_scan->scn_phys; uint64_t *processed = &scn_phys->scn_processed; /* XXX cleanup? */ if (vd->vdev_ops->vdev_op_leaf) atomic_add_64(processed, psize); vs->vs_scan_processed += psize; } if (flags & ZIO_FLAG_SELF_HEAL) vs->vs_self_healed += psize; } vs->vs_ops[type]++; vs->vs_bytes[type] += psize; mutex_exit(&vd->vdev_stat_lock); return; } if (flags & ZIO_FLAG_SPECULATIVE) return; /* * If this is an I/O error that is going to be retried, then ignore the * error. Otherwise, the user may interpret B_FAILFAST I/O errors as * hard errors, when in reality they can happen for any number of * innocuous reasons (bus resets, MPxIO link failure, etc). */ if (zio->io_error == EIO && !(zio->io_flags & ZIO_FLAG_IO_RETRY)) return; /* * Intent logs writes won't propagate their error to the root * I/O so don't mark these types of failures as pool-level * errors. */ if (zio->io_vd == NULL && (zio->io_flags & ZIO_FLAG_DONT_PROPAGATE)) return; mutex_enter(&vd->vdev_stat_lock); if (type == ZIO_TYPE_READ && !vdev_is_dead(vd)) { if (zio->io_error == ECKSUM) vs->vs_checksum_errors++; else vs->vs_read_errors++; } if (type == ZIO_TYPE_WRITE && !vdev_is_dead(vd)) vs->vs_write_errors++; mutex_exit(&vd->vdev_stat_lock); if (type == ZIO_TYPE_WRITE && txg != 0 && (!(flags & ZIO_FLAG_IO_REPAIR) || (flags & ZIO_FLAG_SCAN_THREAD) || spa->spa_claiming)) { /* * This is either a normal write (not a repair), or it's * a repair induced by the scrub thread, or it's a repair * made by zil_claim() during spa_load() in the first txg. * In the normal case, we commit the DTL change in the same * txg as the block was born. In the scrub-induced repair * case, we know that scrubs run in first-pass syncing context, * so we commit the DTL change in spa_syncing_txg(spa). * In the zil_claim() case, we commit in spa_first_txg(spa). * * We currently do not make DTL entries for failed spontaneous * self-healing writes triggered by normal (non-scrubbing) * reads, because we have no transactional context in which to * do so -- and it's not clear that it'd be desirable anyway. */ if (vd->vdev_ops->vdev_op_leaf) { uint64_t commit_txg = txg; if (flags & ZIO_FLAG_SCAN_THREAD) { ASSERT(flags & ZIO_FLAG_IO_REPAIR); ASSERT(spa_sync_pass(spa) == 1); vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1); commit_txg = spa_syncing_txg(spa); } else if (spa->spa_claiming) { ASSERT(flags & ZIO_FLAG_IO_REPAIR); commit_txg = spa_first_txg(spa); } ASSERT(commit_txg >= spa_syncing_txg(spa)); if (vdev_dtl_contains(vd, DTL_MISSING, txg, 1)) return; for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) vdev_dtl_dirty(pvd, DTL_PARTIAL, txg, 1); vdev_dirty(vd->vdev_top, VDD_DTL, vd, commit_txg); } if (vd != rvd) vdev_dtl_dirty(vd, DTL_MISSING, txg, 1); } } /* * Update the in-core space usage stats for this vdev, its metaslab class, * and the root vdev. */ void vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta, int64_t space_delta) { int64_t dspace_delta = space_delta; spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; metaslab_group_t *mg = vd->vdev_mg; metaslab_class_t *mc = mg ? mg->mg_class : NULL; ASSERT(vd == vd->vdev_top); /* * Apply the inverse of the psize-to-asize (ie. RAID-Z) space-expansion * factor. We must calculate this here and not at the root vdev * because the root vdev's psize-to-asize is simply the max of its * childrens', thus not accurate enough for us. */ ASSERT((dspace_delta & (SPA_MINBLOCKSIZE-1)) == 0); ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache); dspace_delta = (dspace_delta >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio; mutex_enter(&vd->vdev_stat_lock); vd->vdev_stat.vs_alloc += alloc_delta; vd->vdev_stat.vs_space += space_delta; vd->vdev_stat.vs_dspace += dspace_delta; mutex_exit(&vd->vdev_stat_lock); if (mc == spa_normal_class(spa)) { mutex_enter(&rvd->vdev_stat_lock); rvd->vdev_stat.vs_alloc += alloc_delta; rvd->vdev_stat.vs_space += space_delta; rvd->vdev_stat.vs_dspace += dspace_delta; mutex_exit(&rvd->vdev_stat_lock); } if (mc != NULL) { ASSERT(rvd == vd->vdev_parent); ASSERT(vd->vdev_ms_count != 0); metaslab_class_space_update(mc, alloc_delta, defer_delta, space_delta, dspace_delta); } } /* * Mark a top-level vdev's config as dirty, placing it on the dirty list * so that it will be written out next time the vdev configuration is synced. * If the root vdev is specified (vdev_top == NULL), dirty all top-level vdevs. */ void vdev_config_dirty(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; int c; ASSERT(spa_writeable(spa)); /* * If this is an aux vdev (as with l2cache and spare devices), then we * update the vdev config manually and set the sync flag. */ if (vd->vdev_aux != NULL) { spa_aux_vdev_t *sav = vd->vdev_aux; nvlist_t **aux; uint_t naux; for (c = 0; c < sav->sav_count; c++) { if (sav->sav_vdevs[c] == vd) break; } if (c == sav->sav_count) { /* * We're being removed. There's nothing more to do. */ ASSERT(sav->sav_sync == B_TRUE); return; } sav->sav_sync = B_TRUE; if (nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, &aux, &naux) != 0) { VERIFY(nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_SPARES, &aux, &naux) == 0); } ASSERT(c < naux); /* * Setting the nvlist in the middle if the array is a little * sketchy, but it will work. */ nvlist_free(aux[c]); aux[c] = vdev_config_generate(spa, vd, B_TRUE, 0); return; } /* * The dirty list is protected by the SCL_CONFIG lock. The caller * must either hold SCL_CONFIG as writer, or must be the sync thread * (which holds SCL_CONFIG as reader). There's only one sync thread, * so this is sufficient to ensure mutual exclusion. */ ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_CONFIG, RW_READER))); if (vd == rvd) { for (c = 0; c < rvd->vdev_children; c++) vdev_config_dirty(rvd->vdev_child[c]); } else { ASSERT(vd == vd->vdev_top); if (!list_link_active(&vd->vdev_config_dirty_node) && !vd->vdev_ishole) list_insert_head(&spa->spa_config_dirty_list, vd); } } void vdev_config_clean(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_CONFIG, RW_READER))); ASSERT(list_link_active(&vd->vdev_config_dirty_node)); list_remove(&spa->spa_config_dirty_list, vd); } /* * Mark a top-level vdev's state as dirty, so that the next pass of * spa_sync() can convert this into vdev_config_dirty(). We distinguish * the state changes from larger config changes because they require * much less locking, and are often needed for administrative actions. */ void vdev_state_dirty(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_writeable(spa)); ASSERT(vd == vd->vdev_top); /* * The state list is protected by the SCL_STATE lock. The caller * must either hold SCL_STATE as writer, or must be the sync thread * (which holds SCL_STATE as reader). There's only one sync thread, * so this is sufficient to ensure mutual exclusion. */ ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_STATE, RW_READER))); if (!list_link_active(&vd->vdev_state_dirty_node) && !vd->vdev_ishole) list_insert_head(&spa->spa_state_dirty_list, vd); } void vdev_state_clean(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_STATE, RW_READER))); ASSERT(list_link_active(&vd->vdev_state_dirty_node)); list_remove(&spa->spa_state_dirty_list, vd); } /* * Propagate vdev state up from children to parent. */ void vdev_propagate_state(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; int degraded = 0, faulted = 0; int corrupted = 0; vdev_t *child; if (vd->vdev_children > 0) { for (int c = 0; c < vd->vdev_children; c++) { child = vd->vdev_child[c]; /* * Don't factor holes into the decision. */ if (child->vdev_ishole) continue; if (!vdev_readable(child) || (!vdev_writeable(child) && spa_writeable(spa))) { /* * Root special: if there is a top-level log * device, treat the root vdev as if it were * degraded. */ if (child->vdev_islog && vd == rvd) degraded++; else faulted++; } else if (child->vdev_state <= VDEV_STATE_DEGRADED) { degraded++; } if (child->vdev_stat.vs_aux == VDEV_AUX_CORRUPT_DATA) corrupted++; } vd->vdev_ops->vdev_op_state_change(vd, faulted, degraded); /* * Root special: if there is a top-level vdev that cannot be * opened due to corrupted metadata, then propagate the root * vdev's aux state as 'corrupt' rather than 'insufficient * replicas'. */ if (corrupted && vd == rvd && rvd->vdev_state == VDEV_STATE_CANT_OPEN) vdev_set_state(rvd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); } if (vd->vdev_parent) vdev_propagate_state(vd->vdev_parent); } /* * Set a vdev's state. If this is during an open, we don't update the parent * state, because we're in the process of opening children depth-first. * Otherwise, we propagate the change to the parent. * * If this routine places a device in a faulted state, an appropriate ereport is * generated. */ void vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux) { uint64_t save_state; spa_t *spa = vd->vdev_spa; if (state == vd->vdev_state) { vd->vdev_stat.vs_aux = aux; return; } save_state = vd->vdev_state; vd->vdev_state = state; vd->vdev_stat.vs_aux = aux; /* * If we are setting the vdev state to anything but an open state, then * always close the underlying device unless the device has requested * a delayed close (i.e. we're about to remove or fault the device). * Otherwise, we keep accessible but invalid devices open forever. * We don't call vdev_close() itself, because that implies some extra * checks (offline, etc) that we don't want here. This is limited to * leaf devices, because otherwise closing the device will affect other * children. */ if (!vd->vdev_delayed_close && vdev_is_dead(vd) && vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_close(vd); /* * If we have brought this vdev back into service, we need * to notify fmd so that it can gracefully repair any outstanding * cases due to a missing device. We do this in all cases, even those * that probably don't correlate to a repaired fault. This is sure to * catch all cases, and we let the zfs-retire agent sort it out. If * this is a transient state it's OK, as the retire agent will * double-check the state of the vdev before repairing it. */ if (state == VDEV_STATE_HEALTHY && vd->vdev_ops->vdev_op_leaf && vd->vdev_prevstate != state) zfs_post_state_change(spa, vd); if (vd->vdev_removed && state == VDEV_STATE_CANT_OPEN && (aux == VDEV_AUX_OPEN_FAILED || vd->vdev_checkremove)) { /* * If the previous state is set to VDEV_STATE_REMOVED, then this * device was previously marked removed and someone attempted to * reopen it. If this failed due to a nonexistent device, then * keep the device in the REMOVED state. We also let this be if * it is one of our special test online cases, which is only * attempting to online the device and shouldn't generate an FMA * fault. */ vd->vdev_state = VDEV_STATE_REMOVED; vd->vdev_stat.vs_aux = VDEV_AUX_NONE; } else if (state == VDEV_STATE_REMOVED) { vd->vdev_removed = B_TRUE; } else if (state == VDEV_STATE_CANT_OPEN) { /* * If we fail to open a vdev during an import or recovery, we * mark it as "not available", which signifies that it was * never there to begin with. Failure to open such a device * is not considered an error. */ if ((spa_load_state(spa) == SPA_LOAD_IMPORT || spa_load_state(spa) == SPA_LOAD_RECOVER) && vd->vdev_ops->vdev_op_leaf) vd->vdev_not_present = 1; /* * Post the appropriate ereport. If the 'prevstate' field is * set to something other than VDEV_STATE_UNKNOWN, it indicates * that this is part of a vdev_reopen(). In this case, we don't * want to post the ereport if the device was already in the * CANT_OPEN state beforehand. * * If the 'checkremove' flag is set, then this is an attempt to * online the device in response to an insertion event. If we * hit this case, then we have detected an insertion event for a * faulted or offline device that wasn't in the removed state. * In this scenario, we don't post an ereport because we are * about to replace the device, or attempt an online with * vdev_forcefault, which will generate the fault for us. */ if ((vd->vdev_prevstate != state || vd->vdev_forcefault) && !vd->vdev_not_present && !vd->vdev_checkremove && vd != spa->spa_root_vdev) { const char *class; switch (aux) { case VDEV_AUX_OPEN_FAILED: class = FM_EREPORT_ZFS_DEVICE_OPEN_FAILED; break; case VDEV_AUX_CORRUPT_DATA: class = FM_EREPORT_ZFS_DEVICE_CORRUPT_DATA; break; case VDEV_AUX_NO_REPLICAS: class = FM_EREPORT_ZFS_DEVICE_NO_REPLICAS; break; case VDEV_AUX_BAD_GUID_SUM: class = FM_EREPORT_ZFS_DEVICE_BAD_GUID_SUM; break; case VDEV_AUX_TOO_SMALL: class = FM_EREPORT_ZFS_DEVICE_TOO_SMALL; break; case VDEV_AUX_BAD_LABEL: class = FM_EREPORT_ZFS_DEVICE_BAD_LABEL; break; default: class = FM_EREPORT_ZFS_DEVICE_UNKNOWN; } zfs_ereport_post(class, spa, vd, NULL, save_state, 0); } /* Erase any notion of persistent removed state */ vd->vdev_removed = B_FALSE; } else { vd->vdev_removed = B_FALSE; } if (!isopen && vd->vdev_parent) vdev_propagate_state(vd->vdev_parent); } /* * Check the vdev configuration to ensure that it's capable of supporting * a root pool. * * On Solaris, we do not support RAID-Z or partial configuration. In * addition, only a single top-level vdev is allowed and none of the * leaves can be wholedisks. * * For FreeBSD, we can boot from any configuration. There is a * limitation that the boot filesystem must be either uncompressed or * compresses with lzjb compression but I'm not sure how to enforce * that here. */ boolean_t vdev_is_bootable(vdev_t *vd) { #ifdef sun if (!vd->vdev_ops->vdev_op_leaf) { char *vdev_type = vd->vdev_ops->vdev_op_type; if (strcmp(vdev_type, VDEV_TYPE_ROOT) == 0 && vd->vdev_children > 1) { return (B_FALSE); } else if (strcmp(vdev_type, VDEV_TYPE_RAIDZ) == 0 || strcmp(vdev_type, VDEV_TYPE_MISSING) == 0) { return (B_FALSE); } } else if (vd->vdev_wholedisk == 1) { return (B_FALSE); } for (int c = 0; c < vd->vdev_children; c++) { if (!vdev_is_bootable(vd->vdev_child[c])) return (B_FALSE); } #endif /* sun */ return (B_TRUE); } /* * Load the state from the original vdev tree (ovd) which * we've retrieved from the MOS config object. If the original * vdev was offline or faulted then we transfer that state to the * device in the current vdev tree (nvd). */ void vdev_load_log_state(vdev_t *nvd, vdev_t *ovd) { spa_t *spa = nvd->vdev_spa; ASSERT(nvd->vdev_top->vdev_islog); ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); ASSERT3U(nvd->vdev_guid, ==, ovd->vdev_guid); for (int c = 0; c < nvd->vdev_children; c++) vdev_load_log_state(nvd->vdev_child[c], ovd->vdev_child[c]); if (nvd->vdev_ops->vdev_op_leaf) { /* * Restore the persistent vdev state */ nvd->vdev_offline = ovd->vdev_offline; nvd->vdev_faulted = ovd->vdev_faulted; nvd->vdev_degraded = ovd->vdev_degraded; nvd->vdev_removed = ovd->vdev_removed; } } /* * Determine if a log device has valid content. If the vdev was * removed or faulted in the MOS config then we know that * the content on the log device has already been written to the pool. */ boolean_t vdev_log_state_valid(vdev_t *vd) { if (vd->vdev_ops->vdev_op_leaf && !vd->vdev_faulted && !vd->vdev_removed) return (B_TRUE); for (int c = 0; c < vd->vdev_children; c++) if (vdev_log_state_valid(vd->vdev_child[c])) return (B_TRUE); return (B_FALSE); } /* * Expand a vdev if possible. */ void vdev_expand(vdev_t *vd, uint64_t txg) { ASSERT(vd->vdev_top == vd); ASSERT(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); if ((vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count) { VERIFY(vdev_metaslab_init(vd, txg) == 0); vdev_config_dirty(vd); } } /* * Split a vdev. */ void vdev_split(vdev_t *vd) { vdev_t *cvd, *pvd = vd->vdev_parent; vdev_remove_child(pvd, vd); vdev_compact_children(pvd); cvd = pvd->vdev_child[0]; if (pvd->vdev_children == 1) { vdev_remove_parent(cvd); cvd->vdev_splitting = B_TRUE; } vdev_propagate_state(cvd); } void vdev_deadman(vdev_t *vd) { for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; vdev_deadman(cvd); } if (vd->vdev_ops->vdev_op_leaf) { vdev_queue_t *vq = &vd->vdev_queue; mutex_enter(&vq->vq_lock); if (avl_numnodes(&vq->vq_active_tree) > 0) { spa_t *spa = vd->vdev_spa; zio_t *fio; uint64_t delta; /* * Look at the head of all the pending queues, * if any I/O has been outstanding for longer than * the spa_deadman_synctime we panic the system. */ fio = avl_first(&vq->vq_active_tree); delta = gethrtime() - fio->io_timestamp; if (delta > spa_deadman_synctime(spa)) { zfs_dbgmsg("SLOW IO: zio timestamp %lluns, " "delta %lluns, last io %lluns", fio->io_timestamp, delta, vq->vq_io_complete_ts); fm_panic("I/O to pool '%s' appears to be " "hung on vdev guid %llu at '%s'.", spa_name(spa), (long long unsigned int) vd->vdev_guid, vd->vdev_path); } } mutex_exit(&vq->vq_lock); } } Index: projects/building-blocks/sys/cddl/contrib/opensolaris =================================================================== --- projects/building-blocks/sys/cddl/contrib/opensolaris (revision 275663) +++ projects/building-blocks/sys/cddl/contrib/opensolaris (revision 275664) Property changes on: projects/building-blocks/sys/cddl/contrib/opensolaris ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,2 ## Merged /head/sys/cddl/contrib/opensolaris:r275590-275663 Merged /vendor-sys/illumos/dist:r275540 Index: projects/building-blocks/sys/conf/files.amd64 =================================================================== --- projects/building-blocks/sys/conf/files.amd64 (revision 275663) +++ projects/building-blocks/sys/conf/files.amd64 (revision 275664) @@ -1,575 +1,577 @@ # This file tells config what files go into building a kernel, # files marked standard are always included. # # $FreeBSD$ # # The long compile-with and dependency lines are required because of # limitations in config: backslash-newline doesn't work in strings, and # dependency lines other than the first are silently ignored. # # linux32_genassym.o optional compat_linux32 \ dependency "$S/amd64/linux32/linux32_genassym.c" \ compile-with "${CC} ${CFLAGS:N-fno-common} -c ${.IMPSRC}" \ no-obj no-implicit-rule \ clean "linux32_genassym.o" # linux32_assym.h optional compat_linux32 \ dependency "$S/kern/genassym.sh linux32_genassym.o" \ compile-with "sh $S/kern/genassym.sh linux32_genassym.o > ${.TARGET}" \ no-obj no-implicit-rule before-depend \ clean "linux32_assym.h" # ia32_genassym.o standard \ dependency "$S/compat/ia32/ia32_genassym.c" \ compile-with "${CC} ${CFLAGS:N-fno-common} -c ${.IMPSRC}" \ no-obj no-implicit-rule \ clean "ia32_genassym.o" # ia32_assym.h standard \ dependency "$S/kern/genassym.sh ia32_genassym.o" \ compile-with "env NM='${NM}' sh $S/kern/genassym.sh ia32_genassym.o > ${.TARGET}" \ no-obj no-implicit-rule before-depend \ clean "ia32_assym.h" # font.h optional sc_dflt_font \ compile-with "uudecode < /usr/share/syscons/fonts/${SC_DFLT_FONT}-8x16.fnt && file2c 'static u_char dflt_font_16[16*256] = {' '};' < ${SC_DFLT_FONT}-8x16 > font.h && uudecode < /usr/share/syscons/fonts/${SC_DFLT_FONT}-8x14.fnt && file2c 'static u_char dflt_font_14[14*256] = {' '};' < ${SC_DFLT_FONT}-8x14 >> font.h && uudecode < /usr/share/syscons/fonts/${SC_DFLT_FONT}-8x8.fnt && file2c 'static u_char dflt_font_8[8*256] = {' '};' < ${SC_DFLT_FONT}-8x8 >> font.h" \ no-obj no-implicit-rule before-depend \ clean "font.h ${SC_DFLT_FONT}-8x14 ${SC_DFLT_FONT}-8x16 ${SC_DFLT_FONT}-8x8" # atkbdmap.h optional atkbd_dflt_keymap \ compile-with "/usr/sbin/kbdcontrol -L ${ATKBD_DFLT_KEYMAP} | sed -e 's/^static keymap_t.* = /static keymap_t key_map = /' -e 's/^static accentmap_t.* = /static accentmap_t accent_map = /' > atkbdmap.h" \ no-obj no-implicit-rule before-depend \ clean "atkbdmap.h" # ukbdmap.h optional ukbd_dflt_keymap \ compile-with "/usr/sbin/kbdcontrol -L ${UKBD_DFLT_KEYMAP} | sed -e 's/^static keymap_t.* = /static keymap_t key_map = /' -e 's/^static accentmap_t.* = /static accentmap_t accent_map = /' > ukbdmap.h" \ no-obj no-implicit-rule before-depend \ clean "ukbdmap.h" # hpt27xx_lib.o optional hpt27xx \ dependency "$S/dev/hpt27xx/amd64-elf.hpt27xx_lib.o.uu" \ compile-with "uudecode < $S/dev/hpt27xx/amd64-elf.hpt27xx_lib.o.uu" \ no-implicit-rule # hptmvraid.o optional hptmv \ dependency "$S/dev/hptmv/amd64-elf.raid.o.uu" \ compile-with "uudecode < $S/dev/hptmv/amd64-elf.raid.o.uu" \ no-implicit-rule # hptnr_lib.o optional hptnr \ dependency "$S/dev/hptnr/amd64-elf.hptnr_lib.o.uu" \ compile-with "uudecode < $S/dev/hptnr/amd64-elf.hptnr_lib.o.uu" \ no-implicit-rule # hptrr_lib.o optional hptrr \ dependency "$S/dev/hptrr/amd64-elf.hptrr_lib.o.uu" \ compile-with "uudecode < $S/dev/hptrr/amd64-elf.hptrr_lib.o.uu" \ no-implicit-rule # amd64/acpica/acpi_machdep.c optional acpi acpi_wakecode.o optional acpi \ dependency "$S/amd64/acpica/acpi_wakecode.S assym.s" \ compile-with "${NORMAL_S}" \ no-obj no-implicit-rule before-depend \ clean "acpi_wakecode.o" acpi_wakecode.bin optional acpi \ dependency "acpi_wakecode.o" \ compile-with "${OBJCOPY} -S -O binary acpi_wakecode.o ${.TARGET}" \ no-obj no-implicit-rule before-depend \ clean "acpi_wakecode.bin" acpi_wakecode.h optional acpi \ dependency "acpi_wakecode.bin" \ compile-with "file2c -sx 'static char wakecode[] = {' '};' < acpi_wakecode.bin > ${.TARGET}" \ no-obj no-implicit-rule before-depend \ clean "acpi_wakecode.h" acpi_wakedata.h optional acpi \ dependency "acpi_wakecode.o" \ compile-with '${NM} -n --defined-only acpi_wakecode.o | while read offset dummy what; do echo "#define $${what} 0x$${offset}"; done > ${.TARGET}' \ no-obj no-implicit-rule before-depend \ clean "acpi_wakedata.h" # amd64/amd64/amd64_mem.c optional mem #amd64/amd64/apic_vector.S standard amd64/amd64/atomic.c standard amd64/amd64/autoconf.c standard amd64/amd64/bios.c standard amd64/amd64/bpf_jit_machdep.c optional bpf_jitter amd64/amd64/cpu_switch.S standard amd64/amd64/db_disasm.c optional ddb amd64/amd64/db_interface.c optional ddb amd64/amd64/db_trace.c optional ddb amd64/amd64/elf_machdep.c standard amd64/amd64/exception.S standard amd64/amd64/fpu.c standard amd64/amd64/gdb_machdep.c optional gdb amd64/amd64/in_cksum.c optional inet | inet6 amd64/amd64/initcpu.c standard amd64/amd64/io.c optional io amd64/amd64/locore.S standard no-obj amd64/amd64/xen-locore.S optional xenhvm amd64/amd64/machdep.c standard amd64/amd64/mem.c optional mem amd64/amd64/minidump_machdep.c standard amd64/amd64/mp_machdep.c optional smp amd64/amd64/mp_watchdog.c optional mp_watchdog smp amd64/amd64/mpboot.S optional smp amd64/amd64/pmap.c standard amd64/amd64/prof_machdep.c optional profiling-routine amd64/amd64/ptrace_machdep.c standard amd64/amd64/sigtramp.S standard amd64/amd64/stack_machdep.c optional ddb | stack amd64/amd64/support.S standard amd64/amd64/sys_machdep.c standard amd64/amd64/trap.c standard amd64/amd64/uio_machdep.c standard amd64/amd64/uma_machdep.c standard amd64/amd64/vm_machdep.c standard amd64/pci/pci_cfgreg.c optional pci cddl/contrib/opensolaris/common/atomic/amd64/opensolaris_atomic.S optional zfs compile-with "${ZFS_S}" crypto/aesni/aeskeys_amd64.S optional aesni crypto/aesni/aesni.c optional aesni aesni_wrap.o optional aesni \ dependency "$S/crypto/aesni/aesni_wrap.c" \ compile-with "${CC} -c ${CFLAGS:C/^-O2$/-O3/:N-nostdinc} ${WERROR} ${PROF} -mmmx -msse -maes ${.IMPSRC}" \ no-implicit-rule \ clean "aesni_wrap.o" crypto/blowfish/bf_enc.c optional crypto | ipsec crypto/des/des_enc.c optional crypto | ipsec | netsmb crypto/via/padlock.c optional padlock crypto/via/padlock_cipher.c optional padlock crypto/via/padlock_hash.c optional padlock dev/acpica/acpi_if.m standard dev/acpi_support/acpi_wmi_if.m standard dev/agp/agp_amd64.c optional agp dev/agp/agp_i810.c optional agp dev/agp/agp_via.c optional agp dev/amdsbwd/amdsbwd.c optional amdsbwd dev/amdtemp/amdtemp.c optional amdtemp dev/arcmsr/arcmsr.c optional arcmsr pci dev/asmc/asmc.c optional asmc isa dev/atkbdc/atkbd.c optional atkbd atkbdc dev/atkbdc/atkbd_atkbdc.c optional atkbd atkbdc dev/atkbdc/atkbdc.c optional atkbdc dev/atkbdc/atkbdc_isa.c optional atkbdc isa dev/atkbdc/atkbdc_subr.c optional atkbdc dev/atkbdc/psm.c optional psm atkbdc dev/bxe/bxe.c optional bxe pci dev/bxe/bxe_stats.c optional bxe pci dev/bxe/bxe_debug.c optional bxe pci dev/bxe/ecore_sp.c optional bxe pci dev/bxe/bxe_elink.c optional bxe pci dev/bxe/57710_init_values.c optional bxe pci dev/bxe/57711_init_values.c optional bxe pci dev/bxe/57712_init_values.c optional bxe pci dev/coretemp/coretemp.c optional coretemp dev/cpuctl/cpuctl.c optional cpuctl dev/dpms/dpms.c optional dpms # There are no systems with isa slots, so all ed isa entries should go.. dev/ed/if_ed_3c503.c optional ed isa ed_3c503 dev/ed/if_ed_isa.c optional ed isa dev/ed/if_ed_wd80x3.c optional ed isa dev/ed/if_ed_hpp.c optional ed isa ed_hpp dev/ed/if_ed_sic.c optional ed isa ed_sic dev/fb/fb.c optional fb | vga dev/fb/s3_pci.c optional s3pci dev/fb/vesa.c optional vga vesa dev/fb/vga.c optional vga dev/ichwd/ichwd.c optional ichwd dev/if_ndis/if_ndis.c optional ndis dev/if_ndis/if_ndis_pccard.c optional ndis pccard dev/if_ndis/if_ndis_pci.c optional ndis cardbus | ndis pci dev/if_ndis/if_ndis_usb.c optional ndis usb dev/io/iodev.c optional io dev/ipmi/ipmi.c optional ipmi dev/ipmi/ipmi_acpi.c optional ipmi acpi dev/ipmi/ipmi_isa.c optional ipmi isa dev/ipmi/ipmi_kcs.c optional ipmi dev/ipmi/ipmi_smic.c optional ipmi dev/ipmi/ipmi_smbus.c optional ipmi smbus dev/ipmi/ipmi_smbios.c optional ipmi dev/ipmi/ipmi_ssif.c optional ipmi smbus dev/ipmi/ipmi_pci.c optional ipmi pci dev/ipmi/ipmi_linux.c optional ipmi compat_linux32 dev/ixl/if_ixl.c optional ixl pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/if_ixlv.c optional ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/ixlvc.c optional ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/ixl_txrx.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/i40e_osdep.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/i40e_lan_hmc.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/i40e_hmc.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/i40e_common.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/i40e_nvm.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/ixl/i40e_adminq.c optional ixl pci | ixlv pci \ compile-with "${NORMAL_C} -I$S/dev/ixl" dev/fdc/fdc.c optional fdc dev/fdc/fdc_acpi.c optional fdc dev/fdc/fdc_isa.c optional fdc isa dev/fdc/fdc_pccard.c optional fdc pccard dev/fdt/fdt_x86.c optional fdt dev/hpt27xx/hpt27xx_os_bsd.c optional hpt27xx dev/hpt27xx/hpt27xx_osm_bsd.c optional hpt27xx dev/hpt27xx/hpt27xx_config.c optional hpt27xx dev/hptmv/entry.c optional hptmv dev/hptmv/mv.c optional hptmv dev/hptmv/gui_lib.c optional hptmv dev/hptmv/hptproc.c optional hptmv dev/hptmv/ioctl.c optional hptmv dev/hptnr/hptnr_os_bsd.c optional hptnr dev/hptnr/hptnr_osm_bsd.c optional hptnr dev/hptnr/hptnr_config.c optional hptnr dev/hptrr/hptrr_os_bsd.c optional hptrr dev/hptrr/hptrr_osm_bsd.c optional hptrr dev/hptrr/hptrr_config.c optional hptrr dev/hwpmc/hwpmc_amd.c optional hwpmc dev/hwpmc/hwpmc_intel.c optional hwpmc dev/hwpmc/hwpmc_core.c optional hwpmc dev/hwpmc/hwpmc_uncore.c optional hwpmc dev/hwpmc/hwpmc_piv.c optional hwpmc dev/hwpmc/hwpmc_tsc.c optional hwpmc dev/hwpmc/hwpmc_x86.c optional hwpmc dev/hyperv/netvsc/hv_net_vsc.c optional hyperv dev/hyperv/netvsc/hv_netvsc_drv_freebsd.c optional hyperv dev/hyperv/netvsc/hv_rndis_filter.c optional hyperv dev/hyperv/stordisengage/hv_ata_pci_disengage.c optional hyperv dev/hyperv/storvsc/hv_storvsc_drv_freebsd.c optional hyperv dev/hyperv/utilities/hv_kvp.c optional hyperv dev/hyperv/utilities/hv_util.c optional hyperv dev/hyperv/vmbus/hv_channel.c optional hyperv dev/hyperv/vmbus/hv_channel_mgmt.c optional hyperv dev/hyperv/vmbus/hv_connection.c optional hyperv dev/hyperv/vmbus/hv_hv.c optional hyperv dev/hyperv/vmbus/hv_ring_buffer.c optional hyperv dev/hyperv/vmbus/hv_vmbus_drv_freebsd.c optional hyperv dev/kbd/kbd.c optional atkbd | sc | ukbd | vt dev/nfe/if_nfe.c optional nfe pci dev/ntb/if_ntb/if_ntb.c optional if_ntb dev/ntb/ntb_hw/ntb_hw.c optional if_ntb ntb_hw dev/nvd/nvd.c optional nvd nvme dev/nvme/nvme.c optional nvme dev/nvme/nvme_ctrlr.c optional nvme dev/nvme/nvme_ctrlr_cmd.c optional nvme dev/nvme/nvme_ns.c optional nvme dev/nvme/nvme_ns_cmd.c optional nvme dev/nvme/nvme_qpair.c optional nvme dev/nvme/nvme_sysctl.c optional nvme dev/nvme/nvme_test.c optional nvme dev/nvme/nvme_util.c optional nvme dev/nvram/nvram.c optional nvram isa dev/random/ivy.c optional rdrand_rng dev/random/nehemiah.c optional padlock_rng dev/qlxge/qls_dbg.c optional qlxge pci dev/qlxge/qls_dump.c optional qlxge pci dev/qlxge/qls_hw.c optional qlxge pci dev/qlxge/qls_ioctl.c optional qlxge pci dev/qlxge/qls_isr.c optional qlxge pci dev/qlxge/qls_os.c optional qlxge pci dev/qlxgb/qla_dbg.c optional qlxgb pci dev/qlxgb/qla_hw.c optional qlxgb pci dev/qlxgb/qla_ioctl.c optional qlxgb pci dev/qlxgb/qla_isr.c optional qlxgb pci dev/qlxgb/qla_misc.c optional qlxgb pci dev/qlxgb/qla_os.c optional qlxgb pci dev/qlxgbe/ql_dbg.c optional qlxgbe pci dev/qlxgbe/ql_hw.c optional qlxgbe pci dev/qlxgbe/ql_ioctl.c optional qlxgbe pci dev/qlxgbe/ql_isr.c optional qlxgbe pci dev/qlxgbe/ql_misc.c optional qlxgbe pci dev/qlxgbe/ql_os.c optional qlxgbe pci dev/qlxgbe/ql_reset.c optional qlxgbe pci dev/sfxge/common/efx_bootcfg.c optional sfxge inet pci dev/sfxge/common/efx_ev.c optional sfxge inet pci dev/sfxge/common/efx_filter.c optional sfxge inet pci dev/sfxge/common/efx_intr.c optional sfxge inet pci dev/sfxge/common/efx_mac.c optional sfxge inet pci dev/sfxge/common/efx_mcdi.c optional sfxge inet pci dev/sfxge/common/efx_mon.c optional sfxge inet pci dev/sfxge/common/efx_nic.c optional sfxge inet pci dev/sfxge/common/efx_nvram.c optional sfxge inet pci dev/sfxge/common/efx_phy.c optional sfxge inet pci dev/sfxge/common/efx_port.c optional sfxge inet pci dev/sfxge/common/efx_rx.c optional sfxge inet pci dev/sfxge/common/efx_sram.c optional sfxge inet pci dev/sfxge/common/efx_tx.c optional sfxge inet pci dev/sfxge/common/efx_vpd.c optional sfxge inet pci dev/sfxge/common/efx_wol.c optional sfxge inet pci dev/sfxge/common/siena_mac.c optional sfxge inet pci dev/sfxge/common/siena_mon.c optional sfxge inet pci dev/sfxge/common/siena_nic.c optional sfxge inet pci dev/sfxge/common/siena_nvram.c optional sfxge inet pci dev/sfxge/common/siena_phy.c optional sfxge inet pci dev/sfxge/common/siena_sram.c optional sfxge inet pci dev/sfxge/common/siena_vpd.c optional sfxge inet pci dev/sfxge/sfxge.c optional sfxge inet pci dev/sfxge/sfxge_dma.c optional sfxge inet pci dev/sfxge/sfxge_ev.c optional sfxge inet pci dev/sfxge/sfxge_intr.c optional sfxge inet pci dev/sfxge/sfxge_mcdi.c optional sfxge inet pci dev/sfxge/sfxge_port.c optional sfxge inet pci dev/sfxge/sfxge_rx.c optional sfxge inet pci dev/sfxge/sfxge_tx.c optional sfxge inet pci dev/sio/sio.c optional sio dev/sio/sio_isa.c optional sio isa dev/sio/sio_pccard.c optional sio pccard dev/sio/sio_pci.c optional sio pci dev/sio/sio_puc.c optional sio puc dev/speaker/spkr.c optional speaker dev/syscons/apm/apm_saver.c optional apm_saver apm dev/syscons/scterm-teken.c optional sc dev/syscons/scvesactl.c optional sc vga vesa dev/syscons/scvgarndr.c optional sc vga dev/syscons/scvtb.c optional sc dev/tpm/tpm.c optional tpm dev/tpm/tpm_acpi.c optional tpm acpi dev/tpm/tpm_isa.c optional tpm isa dev/uart/uart_cpu_x86.c optional uart dev/viawd/viawd.c optional viawd dev/vmware/vmxnet3/if_vmx.c optional vmx dev/wbwd/wbwd.c optional wbwd dev/wpi/if_wpi.c optional wpi +dev/xen/pci/xen_acpi_pci.c optional xenhvm +dev/xen/pci/xen_pci.c optional xenhvm dev/isci/isci.c optional isci dev/isci/isci_controller.c optional isci dev/isci/isci_domain.c optional isci dev/isci/isci_interrupt.c optional isci dev/isci/isci_io_request.c optional isci dev/isci/isci_logger.c optional isci dev/isci/isci_oem_parameters.c optional isci dev/isci/isci_remote_device.c optional isci dev/isci/isci_sysctl.c optional isci dev/isci/isci_task_request.c optional isci dev/isci/isci_timer.c optional isci dev/isci/scil/sati.c optional isci dev/isci/scil/sati_abort_task_set.c optional isci dev/isci/scil/sati_atapi.c optional isci dev/isci/scil/sati_device.c optional isci dev/isci/scil/sati_inquiry.c optional isci dev/isci/scil/sati_log_sense.c optional isci dev/isci/scil/sati_lun_reset.c optional isci dev/isci/scil/sati_mode_pages.c optional isci dev/isci/scil/sati_mode_select.c optional isci dev/isci/scil/sati_mode_sense.c optional isci dev/isci/scil/sati_mode_sense_10.c optional isci dev/isci/scil/sati_mode_sense_6.c optional isci dev/isci/scil/sati_move.c optional isci dev/isci/scil/sati_passthrough.c optional isci dev/isci/scil/sati_read.c optional isci dev/isci/scil/sati_read_buffer.c optional isci dev/isci/scil/sati_read_capacity.c optional isci dev/isci/scil/sati_reassign_blocks.c optional isci dev/isci/scil/sati_report_luns.c optional isci dev/isci/scil/sati_request_sense.c optional isci dev/isci/scil/sati_start_stop_unit.c optional isci dev/isci/scil/sati_synchronize_cache.c optional isci dev/isci/scil/sati_test_unit_ready.c optional isci dev/isci/scil/sati_unmap.c optional isci dev/isci/scil/sati_util.c optional isci dev/isci/scil/sati_verify.c optional isci dev/isci/scil/sati_write.c optional isci dev/isci/scil/sati_write_and_verify.c optional isci dev/isci/scil/sati_write_buffer.c optional isci dev/isci/scil/sati_write_long.c optional isci dev/isci/scil/sci_abstract_list.c optional isci dev/isci/scil/sci_base_controller.c optional isci dev/isci/scil/sci_base_domain.c optional isci dev/isci/scil/sci_base_iterator.c optional isci dev/isci/scil/sci_base_library.c optional isci dev/isci/scil/sci_base_logger.c optional isci dev/isci/scil/sci_base_memory_descriptor_list.c optional isci dev/isci/scil/sci_base_memory_descriptor_list_decorator.c optional isci dev/isci/scil/sci_base_object.c optional isci dev/isci/scil/sci_base_observer.c optional isci dev/isci/scil/sci_base_phy.c optional isci dev/isci/scil/sci_base_port.c optional isci dev/isci/scil/sci_base_remote_device.c optional isci dev/isci/scil/sci_base_request.c optional isci dev/isci/scil/sci_base_state_machine.c optional isci dev/isci/scil/sci_base_state_machine_logger.c optional isci dev/isci/scil/sci_base_state_machine_observer.c optional isci dev/isci/scil/sci_base_subject.c optional isci dev/isci/scil/sci_util.c optional isci dev/isci/scil/scic_sds_controller.c optional isci dev/isci/scil/scic_sds_library.c optional isci dev/isci/scil/scic_sds_pci.c optional isci dev/isci/scil/scic_sds_phy.c optional isci dev/isci/scil/scic_sds_port.c optional isci dev/isci/scil/scic_sds_port_configuration_agent.c optional isci dev/isci/scil/scic_sds_remote_device.c optional isci dev/isci/scil/scic_sds_remote_node_context.c optional isci dev/isci/scil/scic_sds_remote_node_table.c optional isci dev/isci/scil/scic_sds_request.c optional isci dev/isci/scil/scic_sds_sgpio.c optional isci dev/isci/scil/scic_sds_smp_remote_device.c optional isci dev/isci/scil/scic_sds_smp_request.c optional isci dev/isci/scil/scic_sds_ssp_request.c optional isci dev/isci/scil/scic_sds_stp_packet_request.c optional isci dev/isci/scil/scic_sds_stp_remote_device.c optional isci dev/isci/scil/scic_sds_stp_request.c optional isci dev/isci/scil/scic_sds_unsolicited_frame_control.c optional isci dev/isci/scil/scif_sas_controller.c optional isci dev/isci/scil/scif_sas_controller_state_handlers.c optional isci dev/isci/scil/scif_sas_controller_states.c optional isci dev/isci/scil/scif_sas_domain.c optional isci dev/isci/scil/scif_sas_domain_state_handlers.c optional isci dev/isci/scil/scif_sas_domain_states.c optional isci dev/isci/scil/scif_sas_high_priority_request_queue.c optional isci dev/isci/scil/scif_sas_internal_io_request.c optional isci dev/isci/scil/scif_sas_io_request.c optional isci dev/isci/scil/scif_sas_io_request_state_handlers.c optional isci dev/isci/scil/scif_sas_io_request_states.c optional isci dev/isci/scil/scif_sas_library.c optional isci dev/isci/scil/scif_sas_remote_device.c optional isci dev/isci/scil/scif_sas_remote_device_ready_substate_handlers.c optional isci dev/isci/scil/scif_sas_remote_device_ready_substates.c optional isci dev/isci/scil/scif_sas_remote_device_starting_substate_handlers.c optional isci dev/isci/scil/scif_sas_remote_device_starting_substates.c optional isci dev/isci/scil/scif_sas_remote_device_state_handlers.c optional isci dev/isci/scil/scif_sas_remote_device_states.c optional isci dev/isci/scil/scif_sas_request.c optional isci dev/isci/scil/scif_sas_smp_activity_clear_affiliation.c optional isci dev/isci/scil/scif_sas_smp_io_request.c optional isci dev/isci/scil/scif_sas_smp_phy.c optional isci dev/isci/scil/scif_sas_smp_remote_device.c optional isci dev/isci/scil/scif_sas_stp_io_request.c optional isci dev/isci/scil/scif_sas_stp_remote_device.c optional isci dev/isci/scil/scif_sas_stp_task_request.c optional isci dev/isci/scil/scif_sas_task_request.c optional isci dev/isci/scil/scif_sas_task_request_state_handlers.c optional isci dev/isci/scil/scif_sas_task_request_states.c optional isci dev/isci/scil/scif_sas_timer.c optional isci isa/syscons_isa.c optional sc isa/vga_isa.c optional vga kern/kern_clocksource.c standard kern/link_elf_obj.c standard # # IA32 binary support # #amd64/ia32/ia32_exception.S optional compat_freebsd32 amd64/ia32/ia32_reg.c optional compat_freebsd32 amd64/ia32/ia32_signal.c optional compat_freebsd32 amd64/ia32/ia32_sigtramp.S optional compat_freebsd32 amd64/ia32/ia32_syscall.c optional compat_freebsd32 amd64/ia32/ia32_misc.c optional compat_freebsd32 compat/ia32/ia32_sysvec.c optional compat_freebsd32 compat/linprocfs/linprocfs.c optional linprocfs compat/linsysfs/linsysfs.c optional linsysfs # # Linux/i386 binary support # amd64/linux32/linux32_dummy.c optional compat_linux32 amd64/linux32/linux32_locore.s optional compat_linux32 \ dependency "linux32_assym.h" amd64/linux32/linux32_machdep.c optional compat_linux32 amd64/linux32/linux32_support.s optional compat_linux32 \ dependency "linux32_assym.h" amd64/linux32/linux32_sysent.c optional compat_linux32 amd64/linux32/linux32_sysvec.c optional compat_linux32 compat/linux/linux_emul.c optional compat_linux32 compat/linux/linux_file.c optional compat_linux32 compat/linux/linux_fork.c optional compat_linux32 compat/linux/linux_futex.c optional compat_linux32 compat/linux/linux_getcwd.c optional compat_linux32 compat/linux/linux_ioctl.c optional compat_linux32 compat/linux/linux_ipc.c optional compat_linux32 compat/linux/linux_mib.c optional compat_linux32 compat/linux/linux_misc.c optional compat_linux32 compat/linux/linux_signal.c optional compat_linux32 compat/linux/linux_socket.c optional compat_linux32 compat/linux/linux_stats.c optional compat_linux32 compat/linux/linux_sysctl.c optional compat_linux32 compat/linux/linux_time.c optional compat_linux32 compat/linux/linux_timer.c optional compat_linux32 compat/linux/linux_uid16.c optional compat_linux32 compat/linux/linux_util.c optional compat_linux32 dev/amr/amr_linux.c optional compat_linux32 amr dev/mfi/mfi_linux.c optional compat_linux32 mfi # # Windows NDIS driver support # compat/ndis/kern_ndis.c optional ndisapi pci compat/ndis/kern_windrv.c optional ndisapi pci compat/ndis/subr_hal.c optional ndisapi pci compat/ndis/subr_ndis.c optional ndisapi pci compat/ndis/subr_ntoskrnl.c optional ndisapi pci compat/ndis/subr_pe.c optional ndisapi pci compat/ndis/subr_usbd.c optional ndisapi pci compat/ndis/winx64_wrap.S optional ndisapi pci # libkern/memmove.c standard libkern/memset.c standard # # x86 real mode BIOS emulator, required by atkbdc/dpms/vesa # compat/x86bios/x86bios.c optional x86bios | atkbd | dpms | vesa contrib/x86emu/x86emu.c optional x86bios | atkbd | dpms | vesa # # bvm console # dev/bvm/bvm_console.c optional bvmconsole dev/bvm/bvm_dbg.c optional bvmdebug # # x86 shared code between IA32, AMD64 and PC98 architectures # x86/acpica/OsdEnvironment.c optional acpi x86/acpica/acpi_apm.c optional acpi x86/acpica/acpi_wakeup.c optional acpi x86/acpica/madt.c optional acpi x86/acpica/srat.c optional acpi x86/bios/smbios.c optional smbios x86/bios/vpd.c optional vpd x86/cpufreq/powernow.c optional cpufreq x86/cpufreq/est.c optional cpufreq x86/cpufreq/hwpstate.c optional cpufreq x86/cpufreq/p4tcc.c optional cpufreq x86/iommu/busdma_dmar.c optional acpi acpi_dmar pci x86/iommu/intel_ctx.c optional acpi acpi_dmar pci x86/iommu/intel_drv.c optional acpi acpi_dmar pci x86/iommu/intel_fault.c optional acpi acpi_dmar pci x86/iommu/intel_gas.c optional acpi acpi_dmar pci x86/iommu/intel_idpgtbl.c optional acpi acpi_dmar pci x86/iommu/intel_qi.c optional acpi acpi_dmar pci x86/iommu/intel_quirks.c optional acpi acpi_dmar pci x86/iommu/intel_utils.c optional acpi acpi_dmar pci x86/isa/atpic.c optional atpic isa x86/isa/atrtc.c standard x86/isa/clock.c standard x86/isa/elcr.c optional atpic isa | mptable x86/isa/isa.c standard x86/isa/isa_dma.c standard x86/isa/nmi.c standard x86/isa/orm.c optional isa x86/pci/pci_bus.c optional pci x86/pci/qpi.c optional pci x86/x86/busdma_bounce.c standard x86/x86/busdma_machdep.c standard x86/x86/dump_machdep.c standard x86/x86/fdt_machdep.c optional fdt x86/x86/identcpu.c standard x86/x86/intr_machdep.c standard x86/x86/io_apic.c standard x86/x86/legacy.c standard x86/x86/local_apic.c standard x86/x86/mca.c standard x86/x86/mptable.c optional mptable x86/x86/mptable_pci.c optional mptable pci x86/x86/msi.c optional pci x86/x86/nexus.c standard x86/x86/tsc.c standard x86/x86/delay.c standard x86/xen/hvm.c optional xenhvm x86/xen/xen_intr.c optional xen | xenhvm x86/xen/pv.c optional xenhvm x86/xen/pvcpu_enum.c optional xenhvm x86/xen/xen_apic.c optional xenhvm x86/xen/xenpv.c optional xenhvm x86/xen/xen_nexus.c optional xenhvm x86/xen/xen_msi.c optional xenhvm -x86/xen/xen_pci.c optional xenhvm +x86/xen/xen_pci_bus.c optional xenhvm Index: projects/building-blocks/sys/conf =================================================================== --- projects/building-blocks/sys/conf (revision 275663) +++ projects/building-blocks/sys/conf (revision 275664) Property changes on: projects/building-blocks/sys/conf ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/sys/conf:r275303-275663 Index: projects/building-blocks/sys/dev/beri/virtio/network/if_vtbe.c =================================================================== --- projects/building-blocks/sys/dev/beri/virtio/network/if_vtbe.c (nonexistent) +++ projects/building-blocks/sys/dev/beri/virtio/network/if_vtbe.c (revision 275664) @@ -0,0 +1,655 @@ +/*- + * Copyright (c) 2014 Ruslan Bukin + * All rights reserved. + * + * This software was developed by SRI International and the University of + * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) + * ("CTSRD"), as part of the DARPA CRASH research programme. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +/* + * BERI Virtio Networking Frontend + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#include +#include + +#include + +#include +#include +#include +#include + +#include "pio_if.h" + +#define DPRINTF(fmt, args...) printf(fmt, ##args) + +#define READ4(_sc, _reg) \ + bus_read_4((_sc)->res[0], _reg) +#define WRITE4(_sc, _reg, _val) \ + bus_write_4((_sc)->res[0], _reg, _val) + +#define VTBE_LOCK(sc) mtx_lock(&(sc)->mtx) +#define VTBE_UNLOCK(sc) mtx_unlock(&(sc)->mtx) +#define VTBE_ASSERT_LOCKED(sc) mtx_assert(&(sc)->mtx, MA_OWNED); +#define VTBE_ASSERT_UNLOCKED(sc) mtx_assert(&(sc)->mtx, MA_NOTOWNED); + +/* + * Driver data and defines. + */ +#define DESC_COUNT 256 + +struct vtbe_softc { + struct resource *res[2]; + bus_space_tag_t bst; + bus_space_handle_t bsh; + device_t dev; + struct ifnet *ifp; + int if_flags; + struct mtx mtx; + boolean_t is_attached; + + int beri_mem_offset; + device_t pio_send; + device_t pio_recv; + int opened; + + struct vqueue_info vs_queues[2]; + int vs_curq; + int hdrsize; +}; + +static struct resource_spec vtbe_spec[] = { + { SYS_RES_MEMORY, 0, RF_ACTIVE }, + { -1, 0 } +}; + +static void vtbe_txfinish_locked(struct vtbe_softc *sc); +static void vtbe_rxfinish_locked(struct vtbe_softc *sc); +static void vtbe_stop_locked(struct vtbe_softc *sc); +static int pio_enable_irq(struct vtbe_softc *sc, int enable); + +static void +vtbe_txstart_locked(struct vtbe_softc *sc) +{ + struct virtio_net_hdr_mrg_rxbuf *vnh; + struct iovec iov[DESC_COUNT]; + struct vqueue_info *vq; + struct iovec *riov; + struct ifnet *ifp; + struct mbuf *m; + struct uio uio; + int enqueued; + int iolen; + int error; + int *addr; + int reg; + int len; + int n; + + VTBE_ASSERT_LOCKED(sc); + + /* RX queue */ + vq = &sc->vs_queues[0]; + if (!vq_has_descs(vq)) { + return; + } + + ifp = sc->ifp; + if (ifp->if_drv_flags & IFF_DRV_OACTIVE) { + return; + } + + enqueued = 0; + + if (!vq_ring_ready(vq)) + return; + + vq->vq_save_used = be16toh(vq->vq_used->idx); + + for (;;) { + if (!vq_has_descs(vq)) { + ifp->if_drv_flags |= IFF_DRV_OACTIVE; + break; + } + + IFQ_DRV_DEQUEUE(&ifp->if_snd, m); + if (m == NULL) { + break; + } + + n = vq_getchain(sc->beri_mem_offset, vq, iov, + DESC_COUNT, NULL); + + KASSERT(n >= 1 && n <= DESC_COUNT, + ("wrong descriptors num %d", n)); + + addr = iov[0].iov_base; + len = iov[0].iov_len; + + vnh = iov[0].iov_base; + memset(vnh, 0, sc->hdrsize); + vnh->num_buffers = htobe16(1); + + iov[0].iov_len -= sc->hdrsize; + iov[0].iov_base = (void *)((uintptr_t)iov[0].iov_base + + sc->hdrsize); + riov = &iov[0]; + + uio.uio_resid = iov[0].iov_len; + uio.uio_iov = riov; + uio.uio_segflg = UIO_SYSSPACE; + uio.uio_iovcnt = 1; + uio.uio_offset = 0; + uio.uio_rw = UIO_READ; + + error = m_mbuftouio(&uio, m, 0); + if (error) + panic("m_mbuftouio failed\n"); + + iolen = (len - iov[0].iov_len - sc->hdrsize); + vq_relchain(vq, iov, 0, iolen + sc->hdrsize); + paddr_unmap((void *)addr, len); + + if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); + + BPF_MTAP(ifp, m); + m_freem(m); + + ++enqueued; + } + + if (enqueued != 0) { + reg = htobe32(VIRTIO_MMIO_INT_VRING); + WRITE4(sc, VIRTIO_MMIO_INTERRUPT_STATUS, reg); + + PIO_SET(sc->pio_send, Q_INTR, 1); + } +} + +static void +vtbe_txstart(struct ifnet *ifp) +{ + struct vtbe_softc *sc = ifp->if_softc; + + VTBE_LOCK(sc); + vtbe_txstart_locked(sc); + VTBE_UNLOCK(sc); +} + +static void +vtbe_stop_locked(struct vtbe_softc *sc) +{ + struct ifnet *ifp; + + VTBE_ASSERT_LOCKED(sc); + + ifp = sc->ifp; + ifp->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE); +} + +static void +vtbe_init_locked(struct vtbe_softc *sc) +{ + struct ifnet *ifp = sc->ifp; + + VTBE_ASSERT_LOCKED(sc); + + if (ifp->if_drv_flags & IFF_DRV_RUNNING) + return; + + ifp->if_drv_flags |= IFF_DRV_RUNNING; +} + +static void +vtbe_init(void *if_softc) +{ + struct vtbe_softc *sc = if_softc; + + VTBE_LOCK(sc); + vtbe_init_locked(sc); + VTBE_UNLOCK(sc); +} + +static int +vtbe_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data) +{ + struct ifmediareq *ifmr; + struct vtbe_softc *sc; + struct ifreq *ifr; + int mask, error; + + sc = ifp->if_softc; + ifr = (struct ifreq *)data; + + error = 0; + switch (cmd) { + case SIOCSIFFLAGS: + VTBE_LOCK(sc); + if (ifp->if_flags & IFF_UP) { + pio_enable_irq(sc, 1); + + if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) { + vtbe_init_locked(sc); + } + } else { + pio_enable_irq(sc, 0); + + if (ifp->if_drv_flags & IFF_DRV_RUNNING) { + vtbe_stop_locked(sc); + } + } + sc->if_flags = ifp->if_flags; + VTBE_UNLOCK(sc); + break; + case SIOCADDMULTI: + case SIOCDELMULTI: + break; + case SIOCSIFMEDIA: + case SIOCGIFMEDIA: + ifmr = (struct ifmediareq *)data; + ifmr->ifm_count = 1; + ifmr->ifm_status = (IFM_AVALID | IFM_ACTIVE); + ifmr->ifm_active = (IFM_ETHER | IFM_10G_T | IFM_FDX); + ifmr->ifm_current = ifmr->ifm_active; + break; + case SIOCSIFCAP: + mask = ifp->if_capenable ^ ifr->ifr_reqcap; + if (mask & IFCAP_VLAN_MTU) { + ifp->if_capenable ^= IFCAP_VLAN_MTU; + } + break; + + case SIOCSIFADDR: + pio_enable_irq(sc, 1); + default: + error = ether_ioctl(ifp, cmd, data); + break; + } + + return (error); +} + +static void +vtbe_txfinish_locked(struct vtbe_softc *sc) +{ + struct ifnet *ifp; + + VTBE_ASSERT_LOCKED(sc); + + ifp = sc->ifp; +} + +static int +vq_init(struct vtbe_softc *sc) +{ + struct vqueue_info *vq; + uint8_t *base; + int size; + int reg; + int pfn; + + vq = &sc->vs_queues[sc->vs_curq]; + vq->vq_qsize = DESC_COUNT; + + reg = READ4(sc, VIRTIO_MMIO_QUEUE_PFN); + pfn = be32toh(reg); + vq->vq_pfn = pfn; + + size = vring_size(vq->vq_qsize, VRING_ALIGN); + base = paddr_map(sc->beri_mem_offset, + (pfn << PAGE_SHIFT), size); + + /* First pages are descriptors */ + vq->vq_desc = (struct vring_desc *)base; + base += vq->vq_qsize * sizeof(struct vring_desc); + + /* Then avail ring */ + vq->vq_avail = (struct vring_avail *)base; + base += (2 + vq->vq_qsize + 1) * sizeof(uint16_t); + + /* Then it's rounded up to the next page */ + base = (uint8_t *)roundup2((uintptr_t)base, VRING_ALIGN); + + /* And the last pages are the used ring */ + vq->vq_used = (struct vring_used *)base; + + /* Mark queue as allocated, and start at 0 when we use it. */ + vq->vq_flags = VQ_ALLOC; + vq->vq_last_avail = 0; + + return (0); +} + +static void +vtbe_proc_rx(struct vtbe_softc *sc, struct vqueue_info *vq) +{ + struct iovec iov[DESC_COUNT]; + struct ifnet *ifp; + struct uio uio; + struct mbuf *m; + int iolen; + int i; + int n; + + ifp = sc->ifp; + + n = vq_getchain(sc->beri_mem_offset, vq, iov, + DESC_COUNT, NULL); + + KASSERT(n >= 1 && n <= DESC_COUNT, + ("wrong n %d", n)); + + iolen = 0; + for (i = 1; i < n; i++) { + iolen += iov[i].iov_len; + } + + uio.uio_resid = iolen; + uio.uio_iov = &iov[1]; + uio.uio_segflg = UIO_SYSSPACE; + uio.uio_iovcnt = (n - 1); + uio.uio_rw = UIO_WRITE; + + if ((m = m_uiotombuf(&uio, M_NOWAIT, 0, ETHER_ALIGN, + M_PKTHDR)) == NULL) { + if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); + goto done; + } + + m->m_pkthdr.rcvif = ifp; + + if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); + + CURVNET_SET(ifp->if_vnet); + VTBE_UNLOCK(sc); + (*ifp->if_input)(ifp, m); + VTBE_LOCK(sc); + CURVNET_RESTORE(); + +done: + vq_relchain(vq, iov, n, iolen + sc->hdrsize); +} + +static void +vtbe_rxfinish_locked(struct vtbe_softc *sc) +{ + struct vqueue_info *vq; + int reg; + + /* TX queue */ + vq = &sc->vs_queues[1]; + if (!vq_ring_ready(vq)) + return; + + /* Process new descriptors */ + vq->vq_save_used = be16toh(vq->vq_used->idx); + + while (vq_has_descs(vq)) { + vtbe_proc_rx(sc, vq); + } + + /* Interrupt the other side */ + reg = htobe32(VIRTIO_MMIO_INT_VRING); + WRITE4(sc, VIRTIO_MMIO_INTERRUPT_STATUS, reg); + + PIO_SET(sc->pio_send, Q_INTR, 1); +} + +static void +vtbe_intr(void *arg) +{ + struct vtbe_softc *sc; + int pending; + uint32_t reg; + + sc = arg; + + VTBE_LOCK(sc); + + reg = PIO_READ(sc->pio_recv); + + /* Ack */ + PIO_SET(sc->pio_recv, reg, 0); + + pending = htobe32(reg); + if (pending & Q_SEL) { + reg = READ4(sc, VIRTIO_MMIO_QUEUE_SEL); + sc->vs_curq = be32toh(reg); + } + + if (pending & Q_PFN) { + vq_init(sc); + } + + if (pending & Q_NOTIFY) { + /* beri rx / arm tx notify */ + vtbe_txfinish_locked(sc); + } + + if (pending & Q_NOTIFY1) { + vtbe_rxfinish_locked(sc); + } + + VTBE_UNLOCK(sc); +} + +static int +vtbe_get_hwaddr(struct vtbe_softc *sc, uint8_t *hwaddr) +{ + int rnd; + + /* + * Generate MAC address, use 'bsd' + random 24 low-order bits. + */ + + rnd = arc4random() & 0x00ffffff; + + hwaddr[0] = 'b'; + hwaddr[1] = 's'; + hwaddr[2] = 'd'; + hwaddr[3] = rnd >> 16; + hwaddr[4] = rnd >> 8; + hwaddr[5] = rnd >> 0; + + return (0); +} + +static int +pio_enable_irq(struct vtbe_softc *sc, int enable) +{ + + /* + * IRQ lines should be disabled while reprogram FPGA core. + */ + + if (enable) { + if (sc->opened == 0) { + sc->opened = 1; + PIO_SETUP_IRQ(sc->pio_recv, vtbe_intr, sc); + } + } else { + if (sc->opened == 1) { + PIO_TEARDOWN_IRQ(sc->pio_recv); + sc->opened = 0; + } + } + + return (0); +} + +static int +vtbe_probe(device_t dev) +{ + + if (!ofw_bus_status_okay(dev)) + return (ENXIO); + + if (!ofw_bus_is_compatible(dev, "sri-cambridge,beri-vtnet")) + return (ENXIO); + + device_set_desc(dev, "Virtio BERI Ethernet Controller"); + return (BUS_PROBE_DEFAULT); +} + +static int +vtbe_attach(device_t dev) +{ + uint8_t macaddr[ETHER_ADDR_LEN]; + struct vtbe_softc *sc; + struct ifnet *ifp; + int reg; + + sc = device_get_softc(dev); + sc->dev = dev; + + sc->hdrsize = sizeof(struct virtio_net_hdr_mrg_rxbuf); + + if (bus_alloc_resources(dev, vtbe_spec, sc->res)) { + device_printf(dev, "could not allocate resources\n"); + return (ENXIO); + } + + /* Memory interface */ + sc->bst = rman_get_bustag(sc->res[0]); + sc->bsh = rman_get_bushandle(sc->res[0]); + + mtx_init(&sc->mtx, device_get_nameunit(sc->dev), + MTX_NETWORK_LOCK, MTX_DEF); + + if (setup_offset(dev, &sc->beri_mem_offset) != 0) + return (ENXIO); + if (setup_pio(dev, "pio-send", &sc->pio_send) != 0) + return (ENXIO); + if (setup_pio(dev, "pio-recv", &sc->pio_recv) != 0) + return (ENXIO); + + /* Setup MMIO */ + + /* Specify that we provide network device */ + reg = htobe32(VIRTIO_ID_NETWORK); + WRITE4(sc, VIRTIO_MMIO_DEVICE_ID, reg); + + /* The number of desc we support */ + reg = htobe32(DESC_COUNT); + WRITE4(sc, VIRTIO_MMIO_QUEUE_NUM_MAX, reg); + + /* Our features */ + reg = htobe32(VIRTIO_NET_F_MAC | + VIRTIO_NET_F_MRG_RXBUF | + VIRTIO_F_NOTIFY_ON_EMPTY); + WRITE4(sc, VIRTIO_MMIO_HOST_FEATURES, reg); + + /* Get MAC */ + if (vtbe_get_hwaddr(sc, macaddr)) { + device_printf(sc->dev, "can't get mac\n"); + return (ENXIO); + } + + /* Set up the ethernet interface. */ + sc->ifp = ifp = if_alloc(IFT_ETHER); + ifp->if_baudrate = IF_Gbps(10); + ifp->if_softc = sc; + if_initname(ifp, device_get_name(dev), device_get_unit(dev)); + ifp->if_flags = (IFF_BROADCAST | IFF_SIMPLEX | + IFF_MULTICAST | IFF_PROMISC); + ifp->if_capabilities = IFCAP_VLAN_MTU; + ifp->if_capenable = ifp->if_capabilities; + ifp->if_start = vtbe_txstart; + ifp->if_ioctl = vtbe_ioctl; + ifp->if_init = vtbe_init; + IFQ_SET_MAXLEN(&ifp->if_snd, DESC_COUNT - 1); + ifp->if_snd.ifq_drv_maxlen = DESC_COUNT - 1; + IFQ_SET_READY(&ifp->if_snd); + ifp->if_hdrlen = sizeof(struct ether_vlan_header); + + /* All ready to run, attach the ethernet interface. */ + ether_ifattach(ifp, macaddr); + + sc->is_attached = true; + + return (0); +} + +static device_method_t vtbe_methods[] = { + DEVMETHOD(device_probe, vtbe_probe), + DEVMETHOD(device_attach, vtbe_attach), + + { 0, 0 } +}; + +static driver_t vtbe_driver = { + "vtbe", + vtbe_methods, + sizeof(struct vtbe_softc), +}; + +static devclass_t vtbe_devclass; + +DRIVER_MODULE(vtbe, simplebus, vtbe_driver, vtbe_devclass, 0, 0); +MODULE_DEPEND(vtbe, ether, 1, 1, 1); Property changes on: projects/building-blocks/sys/dev/beri/virtio/network/if_vtbe.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/sys/dev/beri/virtio/virtio.c =================================================================== --- projects/building-blocks/sys/dev/beri/virtio/virtio.c (revision 275663) +++ projects/building-blocks/sys/dev/beri/virtio/virtio.c (revision 275664) @@ -1,183 +1,248 @@ /*- * Copyright (c) 2014 Ruslan Bukin * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * BERI virtio mmio backend common methods */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include +#include +#include +#include + #include #include #include #include +#include +#include "pio_if.h" + int vq_ring_ready(struct vqueue_info *vq) { return (vq->vq_flags & VQ_ALLOC); } int vq_has_descs(struct vqueue_info *vq) { return (vq_ring_ready(vq) && vq->vq_last_avail != be16toh(vq->vq_avail->idx)); } void * paddr_map(uint32_t offset, uint32_t phys, uint32_t size) { bus_space_handle_t bsh; if (bus_space_map(fdtbus_bs_tag, (phys + offset), size, 0, &bsh) != 0) { panic("Couldn't map 0x%08x\n", (phys + offset)); } return (void *)(bsh); } void paddr_unmap(void *phys, uint32_t size) { bus_space_unmap(fdtbus_bs_tag, (bus_space_handle_t)phys, size); } static inline void _vq_record(uint32_t offs, int i, volatile struct vring_desc *vd, struct iovec *iov, int n_iov, uint16_t *flags) { if (i >= n_iov) return; iov[i].iov_base = paddr_map(offs, be64toh(vd->addr), be32toh(vd->len)); iov[i].iov_len = be32toh(vd->len); if (flags != NULL) flags[i] = be16toh(vd->flags); } int vq_getchain(uint32_t offs, struct vqueue_info *vq, struct iovec *iov, int n_iov, uint16_t *flags) { volatile struct vring_desc *vdir, *vindir, *vp; int idx, ndesc, n_indir; int head, next; int i; idx = vq->vq_last_avail; ndesc = (be16toh(vq->vq_avail->idx) - idx); if (ndesc == 0) return (0); head = be16toh(vq->vq_avail->ring[idx & (vq->vq_qsize - 1)]); next = head; for (i = 0; i < VQ_MAX_DESCRIPTORS; next = be16toh(vdir->next)) { vdir = &vq->vq_desc[next]; if ((be16toh(vdir->flags) & VRING_DESC_F_INDIRECT) == 0) { _vq_record(offs, i, vdir, iov, n_iov, flags); i++; } else { n_indir = be32toh(vdir->len) / 16; vindir = paddr_map(offs, be64toh(vdir->addr), be32toh(vdir->len)); next = 0; for (;;) { vp = &vindir[next]; _vq_record(offs, i, vp, iov, n_iov, flags); i+=1; if ((be16toh(vp->flags) & \ VRING_DESC_F_NEXT) == 0) break; next = be16toh(vp->next); } paddr_unmap((void *)vindir, be32toh(vdir->len)); } if ((be16toh(vdir->flags) & VRING_DESC_F_NEXT) == 0) return (i); } return (i); } void vq_relchain(struct vqueue_info *vq, struct iovec *iov, int n, uint32_t iolen) { volatile struct vring_used_elem *vue; volatile struct vring_used *vu; uint16_t head, uidx, mask; int i; mask = vq->vq_qsize - 1; + vu = vq->vq_used; head = be16toh(vq->vq_avail->ring[vq->vq_last_avail++ & mask]); - vu = vq->vq_used; uidx = be16toh(vu->idx); vue = &vu->ring[uidx++ & mask]; - vue->id = htobe16(head); + vue->id = htobe32(head); + vue->len = htobe32(iolen); vu->idx = htobe16(uidx); /* Clean up */ for (i = 1; i < (n-1); i++) { paddr_unmap((void *)iov[i].iov_base, iov[i].iov_len); } } + +int +setup_pio(device_t dev, char *name, device_t *pio_dev) +{ + phandle_t pio_node; + struct fdt_ic *ic; + phandle_t xref; + phandle_t node; + + if ((node = ofw_bus_get_node(dev)) == -1) + return (ENXIO); + + if (OF_searchencprop(node, name, &xref, + sizeof(xref)) == -1) { + return (ENXIO); + } + + pio_node = OF_node_from_xref(xref); + SLIST_FOREACH(ic, &fdt_ic_list_head, fdt_ics) { + if (ic->iph == pio_node) { + *pio_dev = ic->dev; + PIO_CONFIGURE(*pio_dev, PIO_OUT_ALL, + PIO_UNMASK_ALL); + return (0); + } + } + + return (ENXIO); +} + +int +setup_offset(device_t dev, uint32_t *offset) +{ + pcell_t dts_value[2]; + phandle_t mem_node; + phandle_t xref; + phandle_t node; + int len; + + if ((node = ofw_bus_get_node(dev)) == -1) + return (ENXIO); + + if (OF_searchencprop(node, "beri-mem", &xref, + sizeof(xref)) == -1) { + return (ENXIO); + } + + mem_node = OF_node_from_xref(xref); + if ((len = OF_getproplen(mem_node, "reg")) <= 0) + return (ENXIO); + OF_getencprop(mem_node, "reg", dts_value, len); + *offset = dts_value[0]; + + return (0); +} + Index: projects/building-blocks/sys/dev/beri/virtio/virtio.h =================================================================== --- projects/building-blocks/sys/dev/beri/virtio/virtio.h (revision 275663) +++ projects/building-blocks/sys/dev/beri/virtio/virtio.h (revision 275664) @@ -1,72 +1,70 @@ /*- * Copyright (c) 2014 Ruslan Bukin * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ #define READ2(_sc, _reg) \ bus_read_2((_sc)->res[0], _reg) #define READ4(_sc, _reg) \ bus_read_4((_sc)->res[0], _reg) #define WRITE2(_sc, _reg, _val) \ bus_write_2((_sc)->res[0], _reg, _val) #define WRITE4(_sc, _reg, _val) \ bus_write_4((_sc)->res[0], _reg, _val) #define PAGE_SHIFT 12 #define VRING_ALIGN 4096 -#define NUM_QUEUES 1 #define VQ_ALLOC 0x01 /* set once we have a pfn */ #define VQ_MAX_DESCRIPTORS 512 -#define VTBLK_BLK_ID_BYTES 20 -#define VTBLK_MAXSEGS 256 - struct vqueue_info { uint16_t vq_qsize; /* size of this queue (a power of 2) */ uint16_t vq_num; uint16_t vq_flags; uint16_t vq_last_avail; /* a recent value of vq_avail->va_idx */ uint16_t vq_save_used; /* saved vq_used->vu_idx; see vq_endchains */ uint32_t vq_pfn; /* PFN of virt queue (not shifted!) */ volatile struct vring_desc *vq_desc; /* descriptor array */ volatile struct vring_avail *vq_avail; /* the "avail" ring */ volatile struct vring_used *vq_used; /* the "used" ring */ }; int vq_ring_ready(struct vqueue_info *vq); int vq_has_descs(struct vqueue_info *vq); void * paddr_map(uint32_t offset, uint32_t phys, uint32_t size); void paddr_unmap(void *phys, uint32_t size); int vq_getchain(uint32_t beri_mem_offset, struct vqueue_info *vq, struct iovec *iov, int n_iov, uint16_t *flags); void vq_relchain(struct vqueue_info *vq, struct iovec *iov, int n, uint32_t iolen); +int setup_pio(device_t dev, char *name, device_t *pio_dev); +int setup_offset(device_t dev, uint32_t *offset); Index: projects/building-blocks/sys/dev/beri/virtio/virtio_block.c =================================================================== --- projects/building-blocks/sys/dev/beri/virtio/virtio_block.c (revision 275663) +++ projects/building-blocks/sys/dev/beri/virtio/virtio_block.c (revision 275664) @@ -1,602 +1,558 @@ /*- * Copyright (c) 2014 Ruslan Bukin * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * BERI virtio block backend driver */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "pio_if.h" #define DPRINTF(fmt, ...) +/* We use indirect descriptors */ +#define NUM_DESCS 1 +#define NUM_QUEUES 1 + +#define VTBLK_BLK_ID_BYTES 20 +#define VTBLK_MAXSEGS 256 + struct beri_vtblk_softc { struct resource *res[1]; bus_space_tag_t bst; bus_space_handle_t bsh; struct cdev *cdev; device_t dev; int opened; device_t pio_recv; device_t pio_send; struct vqueue_info vs_queues[NUM_QUEUES]; char ident[VTBLK_BLK_ID_BYTES]; struct ucred *cred; struct vnode *vnode; struct thread *vtblk_ktd; struct sx sc_mtx; int beri_mem_offset; struct md_ioctl *mdio; struct virtio_blk_config *cfg; }; static struct resource_spec beri_spec[] = { { SYS_RES_MEMORY, 0, RF_ACTIVE }, { -1, 0 } }; static int vtblk_rdwr(struct beri_vtblk_softc *sc, struct iovec *iov, int cnt, int offset, int operation, int iolen) { struct vnode *vp; struct mount *mp; struct uio auio; int error; bzero(&auio, sizeof(auio)); vp = sc->vnode; KASSERT(vp != NULL, ("file not opened")); auio.uio_iov = iov; auio.uio_iovcnt = cnt; auio.uio_offset = offset; auio.uio_segflg = UIO_SYSSPACE; auio.uio_rw = operation; auio.uio_resid = iolen; auio.uio_td = curthread; if (operation == 0) { vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); error = VOP_READ(vp, &auio, IO_DIRECT, sc->cred); VOP_UNLOCK(vp, 0); } else { (void) vn_start_write(vp, &mp, V_WAIT); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); error = VOP_WRITE(vp, &auio, IO_SYNC, sc->cred); VOP_UNLOCK(vp, 0); vn_finished_write(mp); } return (error); } static void vtblk_proc(struct beri_vtblk_softc *sc, struct vqueue_info *vq) { struct iovec iov[VTBLK_MAXSEGS + 2]; uint16_t flags[VTBLK_MAXSEGS + 2]; struct virtio_blk_outhdr *vbh; uint8_t *status; off_t offset; int iolen; int type; int i, n; int err; n = vq_getchain(sc->beri_mem_offset, vq, iov, VTBLK_MAXSEGS + 2, flags); KASSERT(n >= 2 && n <= VTBLK_MAXSEGS + 2, ("wrong n value %d", n)); vbh = iov[0].iov_base; status = iov[n-1].iov_base; KASSERT(iov[n-1].iov_len == 1, ("iov_len == %d", iov[n-1].iov_len)); type = be32toh(vbh->type) & ~VIRTIO_BLK_T_BARRIER; offset = be64toh(vbh->sector) * DEV_BSIZE; iolen = 0; for (i = 1; i < (n-1); i++) { iolen += iov[i].iov_len; } switch (type) { case VIRTIO_BLK_T_OUT: case VIRTIO_BLK_T_IN: err = vtblk_rdwr(sc, iov + 1, i - 1, offset, type, iolen); break; case VIRTIO_BLK_T_GET_ID: /* Assume a single buffer */ strlcpy(iov[1].iov_base, sc->ident, MIN(iov[1].iov_len, sizeof(sc->ident))); err = 0; break; case VIRTIO_BLK_T_FLUSH: /* Possible? */ default: err = -ENOSYS; break; } if (err < 0) { if (err == -ENOSYS) { *status = VIRTIO_BLK_S_UNSUPP; } else *status = VIRTIO_BLK_S_IOERR; } else *status = VIRTIO_BLK_S_OK; vq_relchain(vq, iov, n, 1); } static int close_file(struct beri_vtblk_softc *sc, struct thread *td) { int error; if (sc->vnode != NULL) { vn_lock(sc->vnode, LK_EXCLUSIVE | LK_RETRY); sc->vnode->v_vflag &= ~VV_MD; VOP_UNLOCK(sc->vnode, 0); error = vn_close(sc->vnode, (FREAD|FWRITE), sc->cred, td); if (error != 0) return (error); sc->vnode = NULL; } if (sc->cred != NULL) crfree(sc->cred); return (0); } static int open_file(struct beri_vtblk_softc *sc, struct thread *td) { struct nameidata nd; struct vattr vattr; int error; int flags; flags = (FREAD | FWRITE); NDINIT(&nd, LOOKUP, FOLLOW, UIO_SYSSPACE, sc->mdio->md_file, td); error = vn_open(&nd, &flags, 0, NULL); if (error != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); if (nd.ni_vp->v_type != VREG) { return (EINVAL); } error = VOP_GETATTR(nd.ni_vp, &vattr, td->td_ucred); if (error != 0) return (error); if (VOP_ISLOCKED(nd.ni_vp) != LK_EXCLUSIVE) { vn_lock(nd.ni_vp, LK_UPGRADE | LK_RETRY); if (nd.ni_vp->v_iflag & VI_DOOMED) { return (1); } } nd.ni_vp->v_vflag |= VV_MD; VOP_UNLOCK(nd.ni_vp, 0); sc->vnode = nd.ni_vp; sc->cred = crhold(td->td_ucred); return (0); } static int vtblk_notify(struct beri_vtblk_softc *sc) { struct vqueue_info *vq; int queue; int reg; vq = &sc->vs_queues[0]; if (!vq_ring_ready(vq)) return (0); if (!sc->opened) return (0); reg = READ2(sc, VIRTIO_MMIO_QUEUE_NOTIFY); queue = be16toh(reg); KASSERT(queue == 0, ("we support single queue only")); /* Process new descriptors */ vq = &sc->vs_queues[queue]; vq->vq_save_used = be16toh(vq->vq_used->idx); while (vq_has_descs(vq)) vtblk_proc(sc, vq); - /* Interrupt other side */ - PIO_SET(sc->pio_send, Q_INTR, 1); + /* Interrupt the other side */ + if ((be16toh(vq->vq_avail->flags) & VRING_AVAIL_F_NO_INTERRUPT) == 0) { + reg = htobe32(VIRTIO_MMIO_INT_VRING); + WRITE4(sc, VIRTIO_MMIO_INTERRUPT_STATUS, reg); + PIO_SET(sc->pio_send, Q_INTR, 1); + } return (0); } static int vq_init(struct beri_vtblk_softc *sc) { struct vqueue_info *vq; uint8_t *base; int size; int reg; int pfn; vq = &sc->vs_queues[0]; - vq->vq_qsize = NUM_QUEUES; + vq->vq_qsize = NUM_DESCS; reg = READ4(sc, VIRTIO_MMIO_QUEUE_PFN); pfn = be32toh(reg); vq->vq_pfn = pfn; size = vring_size(vq->vq_qsize, VRING_ALIGN); base = paddr_map(sc->beri_mem_offset, (pfn << PAGE_SHIFT), size); /* First pages are descriptors */ vq->vq_desc = (struct vring_desc *)base; base += vq->vq_qsize * sizeof(struct vring_desc); /* Then avail ring */ vq->vq_avail = (struct vring_avail *)base; base += (2 + vq->vq_qsize + 1) * sizeof(uint16_t); /* Then it's rounded up to the next page */ base = (uint8_t *)roundup2((uintptr_t)base, VRING_ALIGN); /* And the last pages are the used ring */ vq->vq_used = (struct vring_used *)base; /* Mark queue as allocated, and start at 0 when we use it. */ vq->vq_flags = VQ_ALLOC; vq->vq_last_avail = 0; return (0); } static void vtblk_thread(void *arg) { struct beri_vtblk_softc *sc; int err; sc = arg; sx_xlock(&sc->sc_mtx); for (;;) { err = msleep(sc, &sc->sc_mtx, PCATCH | PZERO, "prd", hz); vtblk_notify(sc); } sx_xunlock(&sc->sc_mtx); kthread_exit(); } static int -setup_pio(struct beri_vtblk_softc *sc, char *name, device_t *dev) -{ - phandle_t pio_node; - struct fdt_ic *ic; - phandle_t xref; - phandle_t node; - - if ((node = ofw_bus_get_node(sc->dev)) == -1) - return (ENXIO); - - if (OF_searchencprop(node, name, &xref, - sizeof(xref)) == -1) { - return (ENXIO); - } - - pio_node = OF_node_from_xref(xref); - SLIST_FOREACH(ic, &fdt_ic_list_head, fdt_ics) { - if (ic->iph == pio_node) { - *dev = ic->dev; - PIO_CONFIGURE(*dev, PIO_OUT_ALL, - PIO_UNMASK_ALL); - return (0); - } - } - - return (ENXIO); -} - -static int -setup_offset(struct beri_vtblk_softc *sc) -{ - pcell_t dts_value[2]; - phandle_t mem_node; - phandle_t xref; - phandle_t node; - int len; - - if ((node = ofw_bus_get_node(sc->dev)) == -1) - return (ENXIO); - - if (OF_searchencprop(node, "beri-mem", &xref, - sizeof(xref)) == -1) { - return (ENXIO); - } - - mem_node = OF_node_from_xref(xref); - if ((len = OF_getproplen(mem_node, "reg")) <= 0) - return (ENXIO); - OF_getencprop(mem_node, "reg", dts_value, len); - sc->beri_mem_offset = dts_value[0]; - - return (0); -} - -static int backend_info(struct beri_vtblk_softc *sc) { struct virtio_blk_config *cfg; uint32_t *s; int reg; int i; /* Specify that we provide block device */ reg = htobe32(VIRTIO_ID_BLOCK); WRITE4(sc, VIRTIO_MMIO_DEVICE_ID, reg); - /* The number of queues we support */ - reg = htobe16(NUM_QUEUES); - WRITE2(sc, VIRTIO_MMIO_QUEUE_NUM, reg); + /* Queue size */ + reg = htobe32(NUM_DESCS); + WRITE4(sc, VIRTIO_MMIO_QUEUE_NUM_MAX, reg); /* Our features */ reg = htobe32(VIRTIO_RING_F_INDIRECT_DESC | VIRTIO_BLK_F_BLK_SIZE | VIRTIO_BLK_F_SEG_MAX); WRITE4(sc, VIRTIO_MMIO_HOST_FEATURES, reg); cfg = sc->cfg; cfg->capacity = htobe64(sc->mdio->md_mediasize / DEV_BSIZE); cfg->size_max = 0; /* not negotiated */ cfg->seg_max = htobe32(VTBLK_MAXSEGS); cfg->blk_size = htobe32(DEV_BSIZE); s = (uint32_t *)cfg; for (i = 0; i < sizeof(struct virtio_blk_config); i+=4) { WRITE4(sc, VIRTIO_MMIO_CONFIG + i, *s); s+=1; } sprintf(sc->ident, "Virtio block backend"); return (0); } static void vtblk_intr(void *arg) { struct beri_vtblk_softc *sc; int pending; int reg; sc = arg; reg = PIO_READ(sc->pio_recv); /* Ack */ PIO_SET(sc->pio_recv, reg, 0); pending = htobe32(reg); if (pending & Q_PFN) { vq_init(sc); } if (pending & Q_NOTIFY) { wakeup(sc); } } static int beri_ioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags, struct thread *td) { struct beri_vtblk_softc *sc; int err; sc = dev->si_drv1; switch (cmd) { case MDIOCATTACH: /* take file as argument */ if (sc->vnode != NULL) { /* Already opened */ return (1); } sc->mdio = (struct md_ioctl *)addr; backend_info(sc); DPRINTF("opening file, td 0x%08x\n", (int)td); err = open_file(sc, td); if (err) return (err); PIO_SETUP_IRQ(sc->pio_recv, vtblk_intr, sc); sc->opened = 1; break; case MDIOCDETACH: if (sc->vnode == 0) { /* File not opened */ return (1); } sc->opened = 0; DPRINTF("closing file, td 0x%08x\n", (int)td); err = close_file(sc, td); if (err) return (err); PIO_TEARDOWN_IRQ(sc->pio_recv); break; default: break; } return (0); } static struct cdevsw beri_cdevsw = { .d_version = D_VERSION, .d_ioctl = beri_ioctl, .d_name = "virtio block backend", }; static int beri_vtblk_probe(device_t dev) { if (!ofw_bus_status_okay(dev)) return (ENXIO); if (!ofw_bus_is_compatible(dev, "sri-cambridge,beri-vtblk")) return (ENXIO); device_set_desc(dev, "SRI-Cambridge BERI block"); return (BUS_PROBE_DEFAULT); } static int beri_vtblk_attach(device_t dev) { struct beri_vtblk_softc *sc; int error; sc = device_get_softc(dev); sc->dev = dev; if (bus_alloc_resources(dev, beri_spec, sc->res)) { device_printf(dev, "could not allocate resources\n"); return (ENXIO); } /* Memory interface */ sc->bst = rman_get_bustag(sc->res[0]); sc->bsh = rman_get_bushandle(sc->res[0]); sc->cfg = malloc(sizeof(struct virtio_blk_config), M_DEVBUF, M_NOWAIT|M_ZERO); sx_init(&sc->sc_mtx, device_get_nameunit(sc->dev)); error = kthread_add(vtblk_thread, sc, NULL, &sc->vtblk_ktd, 0, 0, "beri_virtio_block"); if (error) { device_printf(dev, "cannot create kthread\n"); return (ENXIO); } - if (setup_offset(sc) != 0) + if (setup_offset(dev, &sc->beri_mem_offset) != 0) return (ENXIO); - if (setup_pio(sc, "pio-send", &sc->pio_send) != 0) + if (setup_pio(dev, "pio-send", &sc->pio_send) != 0) return (ENXIO); - if (setup_pio(sc, "pio-recv", &sc->pio_recv) != 0) + if (setup_pio(dev, "pio-recv", &sc->pio_recv) != 0) return (ENXIO); sc->cdev = make_dev(&beri_cdevsw, 0, UID_ROOT, GID_WHEEL, S_IRWXU, "beri_vtblk"); if (sc->cdev == NULL) { device_printf(dev, "Failed to create character device.\n"); return (ENXIO); } sc->cdev->si_drv1 = sc; return (0); } static device_method_t beri_vtblk_methods[] = { DEVMETHOD(device_probe, beri_vtblk_probe), DEVMETHOD(device_attach, beri_vtblk_attach), { 0, 0 } }; static driver_t beri_vtblk_driver = { "beri_vtblk", beri_vtblk_methods, sizeof(struct beri_vtblk_softc), }; static devclass_t beri_vtblk_devclass; DRIVER_MODULE(beri_vtblk, simplebus, beri_vtblk_driver, beri_vtblk_devclass, 0, 0); Index: projects/building-blocks/sys/dev/beri/virtio/virtio_mmio_platform.c =================================================================== --- projects/building-blocks/sys/dev/beri/virtio/virtio_mmio_platform.c (revision 275663) +++ projects/building-blocks/sys/dev/beri/virtio/virtio_mmio_platform.c (revision 275664) @@ -1,227 +1,256 @@ /*- * Copyright (c) 2014 Ruslan Bukin * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * BERI interface for Virtio MMIO bus. * * This driver provides interrupt-engine for software-implemented * Virtio MMIO backend. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "virtio_mmio_if.h" #include "pio_if.h" static void platform_intr(void *arg); struct virtio_mmio_platform_softc { struct resource *res[1]; bus_space_tag_t bst; bus_space_handle_t bsh; device_t dev; void (*intr_handler)(void *); void *ih_user; device_t pio_recv; device_t pio_send; }; static int setup_pio(struct virtio_mmio_platform_softc *sc, char *name, device_t *dev) { phandle_t pio_node; struct fdt_ic *ic; phandle_t xref; phandle_t node; if ((node = ofw_bus_get_node(sc->dev)) == -1) return (ENXIO); if (OF_searchencprop(node, name, &xref, sizeof(xref)) == -1) { return (ENXIO); } pio_node = OF_node_from_xref(xref); SLIST_FOREACH(ic, &fdt_ic_list_head, fdt_ics) { if (ic->iph == pio_node) { *dev = ic->dev; PIO_CONFIGURE(*dev, PIO_OUT_ALL, PIO_UNMASK_ALL); return (0); } } return (ENXIO); } static int virtio_mmio_platform_probe(device_t dev) { if (!ofw_bus_status_okay(dev)) return (ENXIO); if (!ofw_bus_is_compatible(dev, "beri,virtio_mmio_platform")) return (ENXIO); device_set_desc(dev, "Virtio MMIO platform"); return (BUS_PROBE_DEFAULT); } static int virtio_mmio_platform_attach(device_t dev) { struct virtio_mmio_platform_softc *sc; struct fdt_ic *fic; phandle_t node; sc = device_get_softc(dev); sc->dev = dev; if (setup_pio(sc, "pio-send", &sc->pio_send) != 0) return (ENXIO); if (setup_pio(sc, "pio-recv", &sc->pio_recv) != 0) return (ENXIO); if ((node = ofw_bus_get_node(sc->dev)) == -1) return (ENXIO); fic = malloc(sizeof(*fic), M_DEVBUF, M_WAITOK|M_ZERO); fic->iph = node; fic->dev = dev; SLIST_INSERT_HEAD(&fdt_ic_list_head, fic, fdt_ics); return (0); } static int -platform_note(device_t dev, size_t offset) +platform_note(device_t dev, size_t offset, int val) { struct virtio_mmio_platform_softc *sc; + int note; + int i; sc = device_get_softc(dev); - if (offset == VIRTIO_MMIO_QUEUE_NOTIFY) { - mips_dcache_wbinv_all(); - PIO_SET(sc->pio_send, Q_NOTIFY, 1); + switch (offset) { + case (VIRTIO_MMIO_QUEUE_NOTIFY): + if (val == 0) + note = Q_NOTIFY; + else if (val == 1) + note = Q_NOTIFY1; + break; + case (VIRTIO_MMIO_QUEUE_PFN): + note = Q_PFN; + break; + case (VIRTIO_MMIO_QUEUE_SEL): + note = Q_SEL; + break; + default: + note = 0; } - if (offset == VIRTIO_MMIO_QUEUE_PFN) { + if (note) { mips_dcache_wbinv_all(); - PIO_SET(sc->pio_send, Q_PFN, 1); + + PIO_SET(sc->pio_send, note, 1); + + /* + * Wait until host ack the request. + * Usually done within few cycles. + * TODO: bad + */ + + for (i = 100; i > 0; i--) { + if (PIO_READ(sc->pio_send) == 0) + break; + } + + if (i == 0) + device_printf(sc->dev, "Warning: host busy\n"); } return (0); } static void platform_intr(void *arg) { struct virtio_mmio_platform_softc *sc; int reg; sc = arg; /* Read pending */ reg = PIO_READ(sc->pio_recv); /* Ack */ PIO_SET(sc->pio_recv, reg, 0); /* Writeback, invalidate cache */ mips_dcache_wbinv_all(); if (sc->intr_handler != NULL) sc->intr_handler(sc->ih_user); } static int platform_setup_intr(device_t dev, device_t mmio_dev, void *intr_handler, void *ih_user) { struct virtio_mmio_platform_softc *sc; sc = device_get_softc(dev); sc->intr_handler = intr_handler; sc->ih_user = ih_user; PIO_SETUP_IRQ(sc->pio_recv, platform_intr, sc); return (0); } static device_method_t virtio_mmio_platform_methods[] = { DEVMETHOD(device_probe, virtio_mmio_platform_probe), DEVMETHOD(device_attach, virtio_mmio_platform_attach), /* virtio_mmio_if.h */ DEVMETHOD(virtio_mmio_note, platform_note), DEVMETHOD(virtio_mmio_setup_intr, platform_setup_intr), DEVMETHOD_END }; static driver_t virtio_mmio_platform_driver = { "virtio_mmio_platform", virtio_mmio_platform_methods, sizeof(struct virtio_mmio_platform_softc), }; static devclass_t virtio_mmio_platform_devclass; DRIVER_MODULE(virtio_mmio_platform, simplebus, virtio_mmio_platform_driver, virtio_mmio_platform_devclass, 0, 0); Index: projects/building-blocks/sys/dev/beri/virtio/virtio_mmio_platform.h =================================================================== --- projects/building-blocks/sys/dev/beri/virtio/virtio_mmio_platform.h (revision 275663) +++ projects/building-blocks/sys/dev/beri/virtio/virtio_mmio_platform.h (revision 275664) @@ -1,35 +1,37 @@ /*- * Copyright (c) 2014 Ruslan Bukin * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $FreeBSD$ */ -#define Q_NOTIFY 0x1 -#define Q_PFN 0x2 -#define Q_INTR 0x4 +#define Q_NOTIFY 0x01 +#define Q_PFN 0x02 +#define Q_INTR 0x04 +#define Q_SEL 0x08 +#define Q_NOTIFY1 0x10 Index: projects/building-blocks/sys/dev/usb/controller/xhci_pci.c =================================================================== --- projects/building-blocks/sys/dev/usb/controller/xhci_pci.c (revision 275663) +++ projects/building-blocks/sys/dev/usb/controller/xhci_pci.c (revision 275664) @@ -1,371 +1,372 @@ /*- * Copyright (c) 2010 Hans Petter Selasky. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "usb_if.h" static device_probe_t xhci_pci_probe; static device_attach_t xhci_pci_attach; static device_detach_t xhci_pci_detach; static usb_take_controller_t xhci_pci_take_controller; static device_method_t xhci_device_methods[] = { /* device interface */ DEVMETHOD(device_probe, xhci_pci_probe), DEVMETHOD(device_attach, xhci_pci_attach), DEVMETHOD(device_detach, xhci_pci_detach), DEVMETHOD(device_suspend, bus_generic_suspend), DEVMETHOD(device_resume, bus_generic_resume), DEVMETHOD(device_shutdown, bus_generic_shutdown), DEVMETHOD(usb_take_controller, xhci_pci_take_controller), DEVMETHOD_END }; static driver_t xhci_driver = { .name = "xhci", .methods = xhci_device_methods, .size = sizeof(struct xhci_softc), }; static devclass_t xhci_devclass; DRIVER_MODULE(xhci, pci, xhci_driver, xhci_devclass, 0, 0); MODULE_DEPEND(xhci, usb, 1, 1, 1); static const char * xhci_pci_match(device_t self) { uint32_t device_id = pci_get_devid(self); switch (device_id) { case 0x01941033: return ("NEC uPD720200 USB 3.0 controller"); case 0x10421b21: return ("ASMedia ASM1042 USB 3.0 controller"); case 0x0f358086: return ("Intel Intel BayTrail USB 3.0 controller"); case 0x9c318086: case 0x1e318086: return ("Intel Panther Point USB 3.0 controller"); case 0x8c318086: return ("Intel Lynx Point USB 3.0 controller"); case 0x8cb18086: return ("Intel Wildcat Point USB 3.0 controller"); default: break; } if ((pci_get_class(self) == PCIC_SERIALBUS) && (pci_get_subclass(self) == PCIS_SERIALBUS_USB) && (pci_get_progif(self) == PCIP_SERIALBUS_USB_XHCI)) { return ("XHCI (generic) USB 3.0 controller"); } return (NULL); /* dunno */ } static int xhci_pci_probe(device_t self) { const char *desc = xhci_pci_match(self); if (desc) { device_set_desc(self, desc); return (0); } else { return (ENXIO); } } static int xhci_use_msi = 1; TUNABLE_INT("hw.usb.xhci.msi", &xhci_use_msi); static void xhci_interrupt_poll(void *_sc) { struct xhci_softc *sc = _sc; USB_BUS_UNLOCK(&sc->sc_bus); xhci_interrupt(sc); USB_BUS_LOCK(&sc->sc_bus); usb_callout_reset(&sc->sc_callout, 1, (void *)&xhci_interrupt_poll, sc); } static int xhci_pci_port_route(device_t self, uint32_t set, uint32_t clear) { uint32_t temp; uint32_t usb3_mask; uint32_t usb2_mask; temp = pci_read_config(self, PCI_XHCI_INTEL_USB3_PSSEN, 4) | pci_read_config(self, PCI_XHCI_INTEL_XUSB2PR, 4); temp |= set; temp &= ~clear; /* Don't set bits which the hardware doesn't support */ usb3_mask = pci_read_config(self, PCI_XHCI_INTEL_USB3PRM, 4); usb2_mask = pci_read_config(self, PCI_XHCI_INTEL_USB2PRM, 4); pci_write_config(self, PCI_XHCI_INTEL_USB3_PSSEN, temp & usb3_mask, 4); pci_write_config(self, PCI_XHCI_INTEL_XUSB2PR, temp & usb2_mask, 4); device_printf(self, "Port routing mask set to 0x%08x\n", temp); return (0); } static int xhci_pci_attach(device_t self) { struct xhci_softc *sc = device_get_softc(self); int count, err, rid; /* XXX check for 64-bit capability */ if (xhci_init(sc, self)) { device_printf(self, "Could not initialize softc\n"); goto error; } pci_enable_busmaster(self); rid = PCI_XHCI_CBMEM; sc->sc_io_res = bus_alloc_resource_any(self, SYS_RES_MEMORY, &rid, RF_ACTIVE); if (!sc->sc_io_res) { device_printf(self, "Could not map memory\n"); goto error; } sc->sc_io_tag = rman_get_bustag(sc->sc_io_res); sc->sc_io_hdl = rman_get_bushandle(sc->sc_io_res); sc->sc_io_size = rman_get_size(sc->sc_io_res); usb_callout_init_mtx(&sc->sc_callout, &sc->sc_bus.bus_mtx, 0); sc->sc_irq_rid = 0; if (xhci_use_msi) { count = pci_msi_count(self); if (count >= 1) { count = 1; if (pci_alloc_msi(self, &count) == 0) { if (bootverbose) device_printf(self, "MSI enabled\n"); sc->sc_irq_rid = 1; } } } sc->sc_irq_res = bus_alloc_resource_any(self, SYS_RES_IRQ, &sc->sc_irq_rid, RF_SHAREABLE | RF_ACTIVE); if (sc->sc_irq_res == NULL) { device_printf(self, "Could not allocate IRQ\n"); /* goto error; FALLTHROUGH - use polling */ } sc->sc_bus.bdev = device_add_child(self, "usbus", -1); if (sc->sc_bus.bdev == NULL) { device_printf(self, "Could not add USB device\n"); goto error; } device_set_ivars(sc->sc_bus.bdev, &sc->sc_bus); sprintf(sc->sc_vendor, "0x%04x", pci_get_vendor(self)); if (sc->sc_irq_res != NULL) { err = bus_setup_intr(self, sc->sc_irq_res, INTR_TYPE_BIO | INTR_MPSAFE, NULL, (driver_intr_t *)xhci_interrupt, sc, &sc->sc_intr_hdl); if (err != 0) { device_printf(self, "Could not setup IRQ, err=%d\n", err); sc->sc_intr_hdl = NULL; } } if (sc->sc_irq_res == NULL || sc->sc_intr_hdl == NULL || xhci_use_polling() != 0) { device_printf(self, "Interrupt polling at %dHz\n", hz); USB_BUS_LOCK(&sc->sc_bus); xhci_interrupt_poll(sc); USB_BUS_UNLOCK(&sc->sc_bus); } /* On Intel chipsets reroute ports from EHCI to XHCI controller. */ switch (pci_get_devid(self)) { case 0x0f358086: /* BayTrail */ case 0x9c318086: /* Panther Point */ case 0x1e318086: /* Panther Point */ case 0x8c318086: /* Lynx Point */ + case 0x8cb18086: /* Wildcat Point */ sc->sc_port_route = &xhci_pci_port_route; sc->sc_imod_default = XHCI_IMOD_DEFAULT_LP; break; default: break; } xhci_pci_take_controller(self); err = xhci_halt_controller(sc); if (err == 0) err = xhci_start_controller(sc); if (err == 0) err = device_probe_and_attach(sc->sc_bus.bdev); if (err) { device_printf(self, "XHCI halt/start/probe failed err=%d\n", err); goto error; } return (0); error: xhci_pci_detach(self); return (ENXIO); } static int xhci_pci_detach(device_t self) { struct xhci_softc *sc = device_get_softc(self); device_t bdev; if (sc->sc_bus.bdev != NULL) { bdev = sc->sc_bus.bdev; device_detach(bdev); device_delete_child(self, bdev); } /* during module unload there are lots of children leftover */ device_delete_children(self); if (sc->sc_io_res) { usb_callout_drain(&sc->sc_callout); xhci_halt_controller(sc); } pci_disable_busmaster(self); if (sc->sc_irq_res && sc->sc_intr_hdl) { bus_teardown_intr(self, sc->sc_irq_res, sc->sc_intr_hdl); sc->sc_intr_hdl = NULL; } if (sc->sc_irq_res) { if (sc->sc_irq_rid == 1) pci_release_msi(self); bus_release_resource(self, SYS_RES_IRQ, sc->sc_irq_rid, sc->sc_irq_res); sc->sc_irq_res = NULL; } if (sc->sc_io_res) { bus_release_resource(self, SYS_RES_MEMORY, PCI_XHCI_CBMEM, sc->sc_io_res); sc->sc_io_res = NULL; } xhci_uninit(sc); return (0); } static int xhci_pci_take_controller(device_t self) { struct xhci_softc *sc = device_get_softc(self); uint32_t cparams; uint32_t eecp; uint32_t eec; uint16_t to; uint8_t bios_sem; cparams = XREAD4(sc, capa, XHCI_HCSPARAMS0); eec = -1; /* Synchronise with the BIOS if it owns the controller. */ for (eecp = XHCI_HCS0_XECP(cparams) << 2; eecp != 0 && XHCI_XECP_NEXT(eec); eecp += XHCI_XECP_NEXT(eec) << 2) { eec = XREAD4(sc, capa, eecp); if (XHCI_XECP_ID(eec) != XHCI_ID_USB_LEGACY) continue; bios_sem = XREAD1(sc, capa, eecp + XHCI_XECP_BIOS_SEM); if (bios_sem == 0) continue; device_printf(sc->sc_bus.bdev, "waiting for BIOS " "to give up control\n"); XWRITE1(sc, capa, eecp + XHCI_XECP_OS_SEM, 1); to = 500; while (1) { bios_sem = XREAD1(sc, capa, eecp + XHCI_XECP_BIOS_SEM); if (bios_sem == 0) break; if (--to == 0) { device_printf(sc->sc_bus.bdev, "timed out waiting for BIOS\n"); break; } usb_pause_mtx(NULL, hz / 100); /* wait 10ms */ } } return (0); } Index: projects/building-blocks/sys/dev/usb/quirk/usb_quirk.c =================================================================== --- projects/building-blocks/sys/dev/usb/quirk/usb_quirk.c (revision 275663) +++ projects/building-blocks/sys/dev/usb/quirk/usb_quirk.c (revision 275664) @@ -1,851 +1,870 @@ /* $FreeBSD$ */ /*- * Copyright (c) 1998 The NetBSD Foundation, Inc. All rights reserved. * Copyright (c) 1998 Lennart Augustsson. All rights reserved. * Copyright (c) 2008 Hans Petter Selasky. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "usbdevs.h" #define USB_DEBUG_VAR usb_debug #include #include #include MODULE_DEPEND(usb_quirk, usb, 1, 1, 1); MODULE_VERSION(usb_quirk, 1); -#define USB_DEV_QUIRKS_MAX 320 +#define USB_DEV_QUIRKS_MAX 384 #define USB_SUB_QUIRKS_MAX 8 struct usb_quirk_entry { uint16_t vid; uint16_t pid; uint16_t lo_rev; uint16_t hi_rev; uint16_t quirks[USB_SUB_QUIRKS_MAX]; }; static struct mtx usb_quirk_mtx; #define USB_QUIRK_VP(v,p,l,h,...) \ { .vid = (v), .pid = (p), .lo_rev = (l), .hi_rev = (h), \ .quirks = { __VA_ARGS__ } } #define USB_QUIRK(v,p,l,h,...) \ USB_QUIRK_VP(USB_VENDOR_##v, USB_PRODUCT_##v##_##p, l, h, __VA_ARGS__) static struct usb_quirk_entry usb_quirks[USB_DEV_QUIRKS_MAX] = { USB_QUIRK(ASUS, LCM, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(INSIDEOUT, EDGEPORT4, 0x094, 0x094, UQ_SWAP_UNICODE), USB_QUIRK(DALLAS, J6502, 0x0a2, 0x0a2, UQ_BAD_ADC), USB_QUIRK(DALLAS, J6502, 0x0a2, 0x0a2, UQ_AU_NO_XU), USB_QUIRK(ALTEC, ADA70, 0x103, 0x103, UQ_BAD_ADC), USB_QUIRK(ALTEC, ASC495, 0x000, 0x000, UQ_BAD_AUDIO), USB_QUIRK(QTRONIX, 980N, 0x110, 0x110, UQ_SPUR_BUT_UP), USB_QUIRK(ALCOR2, KBD_HUB, 0x001, 0x001, UQ_SPUR_BUT_UP), USB_QUIRK(MCT, HUB0100, 0x102, 0x102, UQ_BUS_POWERED), USB_QUIRK(MCT, USB232, 0x102, 0x102, UQ_BUS_POWERED), USB_QUIRK(TI, UTUSB41, 0x110, 0x110, UQ_POWER_CLAIM), USB_QUIRK(TELEX, MIC1, 0x009, 0x009, UQ_AU_NO_FRAC), USB_QUIRK(SILICONPORTALS, YAPPHONE, 0x100, 0x100, UQ_AU_INP_ASYNC), USB_QUIRK(LOGITECH, UN53B, 0x0000, 0xffff, UQ_NO_STRINGS), USB_QUIRK(REALTEK, RTL8196EU, 0x0000, 0xffff, UQ_CFG_INDEX_1), USB_QUIRK(ELSA, MODEM1, 0x0000, 0xffff, UQ_CFG_INDEX_1), USB_QUIRK(PLANEX2, MZKUE150N, 0x0000, 0xffff, UQ_CFG_INDEX_1), /* Quirks for printer devices */ USB_QUIRK(HP, 895C, 0x0000, 0xffff, UQ_BROKEN_BIDIR), USB_QUIRK(HP, 880C, 0x0000, 0xffff, UQ_BROKEN_BIDIR), USB_QUIRK(HP, 815C, 0x0000, 0xffff, UQ_BROKEN_BIDIR), USB_QUIRK(HP, 810C, 0x0000, 0xffff, UQ_BROKEN_BIDIR), USB_QUIRK(HP, 830C, 0x0000, 0xffff, UQ_BROKEN_BIDIR), USB_QUIRK(HP, 1220C, 0x0000, 0xffff, UQ_BROKEN_BIDIR), USB_QUIRK(XEROX, WCM15, 0x0000, 0xffff, UQ_BROKEN_BIDIR), /* Devices which should be ignored by uhid */ USB_QUIRK(APC, UPS, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(BELKIN, F6C550AVR, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(CYBERPOWER, 1500CAVRLCD, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(CYPRESS, SILVERSHIELD, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(DELORME, EARTHMATE, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(DREAMLINK, DL100B, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(ITUNERNET, USBLCD2X20, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(ITUNERNET, USBLCD4X20, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(LIEBERT, POWERSURE_PXT, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(LIEBERT2, PSI1000, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(MGE, UPS1, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(MGE, UPS2, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(APPLE, IPHONE, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(APPLE, IPHONE_3G, 0x0000, 0xffff, UQ_HID_IGNORE), USB_QUIRK(MEGATEC, UPS, 0x0000, 0xffff, UQ_HID_IGNORE), /* Devices which should be ignored by both ukbd and uhid */ USB_QUIRK(CYPRESS, WISPY1A, 0x0000, 0xffff, UQ_KBD_IGNORE, UQ_HID_IGNORE), USB_QUIRK(METAGEEK, WISPY1B, 0x0000, 0xffff, UQ_KBD_IGNORE, UQ_HID_IGNORE), USB_QUIRK(METAGEEK, WISPY24X, 0x0000, 0xffff, UQ_KBD_IGNORE, UQ_HID_IGNORE), USB_QUIRK(METAGEEK2, WISPYDBX, 0x0000, 0xffff, UQ_KBD_IGNORE, UQ_HID_IGNORE), USB_QUIRK(TENX, UAUDIO0, 0x0101, 0x0101, UQ_AUDIO_SWAP_LR), /* MS keyboards do weird things */ USB_QUIRK(MICROSOFT, NATURAL4000, 0x0000, 0xFFFF, UQ_KBD_BOOTPROTO), USB_QUIRK(MICROSOFT, WLINTELLIMOUSE, 0x0000, 0xffff, UQ_MS_LEADING_BYTE), /* Quirk for Corsair Vengeance K60 keyboard */ USB_QUIRK(CORSAIR, K60, 0x0000, 0xffff, UQ_KBD_BOOTPROTO), /* Quirk for Corsair Vengeance K70 keyboard */ USB_QUIRK(CORSAIR, K70, 0x0000, 0xffff, UQ_KBD_BOOTPROTO), /* umodem(4) device quirks */ USB_QUIRK(METRICOM, RICOCHET_GS, 0x100, 0x100, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(SANYO, SCP4900, 0x000, 0x000, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(MOTOROLA2, T720C, 0x001, 0x001, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(EICON, DIVA852, 0x100, 0x100, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(SIEMENS2, ES75, 0x000, 0x000, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(QUALCOMM, CDMA_MSM, 0x0000, 0xffff, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(QUALCOMM2, CDMA_MSM, 0x0000, 0xffff, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(CURITEL, UM150, 0x0000, 0xffff, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(CURITEL, UM175, 0x0000, 0xffff, UQ_ASSUME_CM_OVER_DATA), USB_QUIRK(VERTEX, VW110L, 0x0000, 0xffff, UQ_ASSUME_CM_OVER_DATA), /* USB Mass Storage Class Quirks */ USB_QUIRK_VP(USB_VENDOR_ASAHIOPTICAL, 0, UQ_MSC_NO_RS_CLEAR_UA, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(ADDON, ATTACHE, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(ADDON, A256MB, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(ADDON, DISKPRO512, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(ADDONICS2, CABLE_205, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(AIPTEK, POCKETCAM3M, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ALCOR, UMCR_9361, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(ALCOR, TRANSCEND, 0x0000, 0xffff, UQ_MSC_NO_GETMAXLUN, UQ_MSC_NO_SYNC_CACHE, UQ_MSC_NO_TEST_UNIT_READY), USB_QUIRK(APACER, HT202, 0x0000, 0xffff, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(ASAHIOPTICAL, OPTIO230, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(ASAHIOPTICAL, OPTIO330, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(ATP, EUSB, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(BELKIN, USB2SCSI, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(CASIO, QV_DIGICAM, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(CCYU, ED1064, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(CENTURY, EX35QUAT, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(CYPRESS, XX6830XX, 0x0000, 0xffff, UQ_MSC_NO_GETMAXLUN, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(DESKNOTE, UCR_61S2B, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(DMI, CFSM_RW, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(EMTEC, RUF2PS, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(EPSON, STYLUS_875DC, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(EPSON, STYLUS_895, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(FEIYA, 5IN1, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(FEIYA, ELANGO, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(FREECOM, DVD, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(FUJIPHOTO, MASS0100, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_RS_CLEAR_UA, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(GENESYS, GL641USB2IDE, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(GENESYS, GL641USB2IDE_2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(GENESYS, GL641USB, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(GENESYS, GL641USB_2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_WRONG_CSWSIG), USB_QUIRK(HAGIWARA, FG, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(HAGIWARA, FGSM, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(HITACHI, DVDCAM_DZ_MV100A, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(HITACHI, DVDCAM_USB, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_INQUIRY), USB_QUIRK(HP, CDW4E, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(HP, CDW8200, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_START_STOP), USB_QUIRK(IMAGINATION, DBX1, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_WRONG_CSWSIG), USB_QUIRK(INSYSTEM, USBCABLE, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_START_STOP, UQ_MSC_ALT_IFACE_1), USB_QUIRK(INSYSTEM, ATAPI, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(INSYSTEM, STORAGE_V2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(IODATA, IU_CD2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(IODATA, DVR_UEH8, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(IOMEGA, ZIP100, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_TEST_UNIT_READY), /* XXX ZIP drives can also use ATAPI */ USB_QUIRK(JMICRON, JM20337, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(KINGSTON, HYPERX3_0, 0x0000, 0xffff, UQ_MSC_NO_INQUIRY), USB_QUIRK(KYOCERA, FINECAM_L3, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(KYOCERA, FINECAM_S3X, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_INQUIRY), USB_QUIRK(KYOCERA, FINECAM_S4, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_INQUIRY), USB_QUIRK(KYOCERA, FINECAM_S5, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(LACIE, HD, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(LEXAR, CF_READER, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(LEXAR, JUMPSHOT, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(LEXAR, JUMPDRIVE, 0x0000, 0xffff, UQ_MSC_NO_INQUIRY), USB_QUIRK(LOGITEC, LDR_H443SU2, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(LOGITEC, LDR_H443U2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI,), USB_QUIRK(MELCO, DUBPXXG, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(MICROTECH, DPCM, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_START_STOP), USB_QUIRK(MICRON, REALSSD, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(MICROTECH, SCSIDB25, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(MICROTECH, SCSIHD50, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(MINOLTA, E223, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(MINOLTA, F300, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(MITSUMI, CDRRW, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI | UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(MOTOROLA2, E398, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_INQUIRY_EVPD, UQ_MSC_NO_GETMAXLUN), USB_QUIRK_VP(USB_VENDOR_MPMAN, 0, UQ_MSC_NO_SYNC_CACHE, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(MSYSTEMS, DISKONKEY, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE, UQ_MSC_NO_GETMAXLUN, UQ_MSC_NO_RS_CLEAR_UA), USB_QUIRK(MSYSTEMS, DISKONKEY2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(MYSON, HEDEN, 0x0000, 0xffff, UQ_MSC_IGNORE_RESIDUE, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(NEODIO, ND3260, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ), USB_QUIRK(NETAC, CF_CARD, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(NETAC, ONLYDISK, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(NETCHIP, CLIK_40, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_INQUIRY), USB_QUIRK(NETCHIP, POCKETBOOK, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(NIKON, D300, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(OLYMPUS, C1, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_WRONG_CSWSIG), USB_QUIRK(OLYMPUS, C700, 0x0000, 0xffff, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(ONSPEC, SDS_HOTFIND_D, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(ONSPEC, CFMS_RW, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ONSPEC, CFSM_COMBO, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ONSPEC, CFSM_READER, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ONSPEC, CFSM_READER2, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ONSPEC, MDCFE_B_CF_READER, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ONSPEC, MDSM_B_READER, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(ONSPEC, READER, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(ONSPEC, UCF100, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_INQUIRY, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(ONSPEC2, IMAGEMATE_SDDR55, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(PANASONIC, KXL840AN, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(PANASONIC, KXLCB20AN, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(PANASONIC, KXLCB35AN, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(PANASONIC, LS120CAM, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_UFI), USB_QUIRK(PLEXTOR, 40_12_40U, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_TEST_UNIT_READY), USB_QUIRK(PNY, ATTACHE2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE, UQ_MSC_NO_START_STOP), USB_QUIRK(PROLIFIC, PL2506, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK_VP(USB_VENDOR_SAMSUNG_TECHWIN, USB_PRODUCT_SAMSUNG_TECHWIN_DIGIMAX_410, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(SANDISK, SDDR05A, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_READ_CAP_OFFBY1, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SANDISK, SDDR09, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_READ_CAP_OFFBY1, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SANDISK, SDDR12, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_READ_CAP_OFFBY1, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SANDISK, SDCZ2_128, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(SANDISK, SDCZ2_256, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(SANDISK, SDCZ4_128, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(SANDISK, SDCZ4_256, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(SANDISK, SDDR31, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_READ_CAP_OFFBY1), USB_QUIRK(SANDISK, IMAGEMATE_SDDR289, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SCANLOGIC, SL11R, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_INQUIRY), USB_QUIRK(SHUTTLE, EUSB, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_START_STOP, UQ_MSC_SHUTTLE_INIT), USB_QUIRK(SHUTTLE, CDRW, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(SHUTTLE, CF, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(SHUTTLE, EUSBATAPI, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(SHUTTLE, EUSBCFSM, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(SHUTTLE, EUSCSI, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(SHUTTLE, HIFD, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SHUTTLE, SDDR09, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SHUTTLE, ZIOMMC, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SIGMATEL, I_BEAD100, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_SHUTTLE_INIT), USB_QUIRK(SIIG, WINTERREADER, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(SKANHEX, MD_7425, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(SKANHEX, SX_520Z, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(SONY, HANDYCAM, 0x0500, 0x0500, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC, UQ_MSC_RBC_PAD_TO_12), USB_QUIRK(SONY, CLIE_40_MS, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(SONY, DSC, 0x0500, 0x0500, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC, UQ_MSC_RBC_PAD_TO_12), USB_QUIRK(SONY, DSC, 0x0600, 0x0600, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC, UQ_MSC_RBC_PAD_TO_12), USB_QUIRK(SONY, DSC, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(SONY, HANDYCAM, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(SONY, MSC, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(SONY, MS_MSC_U03, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_UFI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SONY, MS_NW_MS7, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SONY, MS_PEG_N760C, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(SONY, MSACUS1, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(SONY, PORTABLE_HDD_V2, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(STMICRO, ST72682, 0x0000, 0xffff, UQ_MSC_NO_PREVENT_ALLOW), USB_QUIRK(SUPERTOP, IDE, 0x0000, 0xffff, UQ_MSC_IGNORE_RESIDUE, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(SUPERTOP, FLASHDRIVE, 0x0000, 0xffff, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(TAUGA, CAMERAMATE, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(TEAC, FD05PUB, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_UFI), USB_QUIRK(TECLAST, TLC300, 0x0000, 0xffff, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(TREK, MEMKEY, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(TREK, THUMBDRIVE_8MB, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(TRUMPION, C3310, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_UFI), USB_QUIRK(TRUMPION, MP3, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_RBC), USB_QUIRK(TRUMPION, T33520, 0x0000, 0xffff, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(TWINMOS, MDIV, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI), USB_QUIRK(VIA, USB2IDEBRIDGE, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(VIVITAR, 35XX, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(WESTERN, COMBO, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(WESTERN, EXTHDD, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(WESTERN, MYBOOK, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY_EVPD, UQ_MSC_NO_SYNC_CACHE), - USB_QUIRK(WESTERN, MYPASSWORD, 0x0000, 0xffff, UQ_MSC_FORCE_SHORT_INQ), - USB_QUIRK(WESTERN, MYPASSPORT, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_00, 0x0000, 0xffff, UQ_MSC_FORCE_SHORT_INQ), + USB_QUIRK(WESTERN, MYPASSPORT_01, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_02, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_03, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_04, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_05, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_06, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_07, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_08, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_09, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORT_10, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_00, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_01, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_02, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_03, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_04, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_05, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_06, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_07, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_08, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), + USB_QUIRK(WESTERN, MYPASSPORTES_09, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(WINMAXGROUP, FLASH64MC, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY), USB_QUIRK(YANO, FW800HD, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_FORCE_SHORT_INQ, UQ_MSC_NO_START_STOP, UQ_MSC_IGNORE_RESIDUE), USB_QUIRK(YANO, U640MO, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_ATAPI, UQ_MSC_FORCE_SHORT_INQ), USB_QUIRK(YEDATA, FLASHBUSTERU, 0x0000, 0x007F, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_UFI, UQ_MSC_NO_RS_CLEAR_UA, UQ_MSC_FLOPPY_SPEED, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(YEDATA, FLASHBUSTERU, 0x0080, 0x0080, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_UFI, UQ_MSC_NO_RS_CLEAR_UA, UQ_MSC_FLOPPY_SPEED, UQ_MSC_NO_TEST_UNIT_READY, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(YEDATA, FLASHBUSTERU, 0x0081, 0xFFFF, UQ_MSC_FORCE_WIRE_CBI_I, UQ_MSC_FORCE_PROTO_UFI, UQ_MSC_NO_RS_CLEAR_UA, UQ_MSC_FLOPPY_SPEED, UQ_MSC_NO_GETMAXLUN), USB_QUIRK(ZORAN, EX20DSC, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_CBI, UQ_MSC_FORCE_PROTO_ATAPI), USB_QUIRK(MEIZU, M6_SL, 0x0000, 0xffff, UQ_MSC_FORCE_WIRE_BBB, UQ_MSC_FORCE_PROTO_SCSI, UQ_MSC_NO_INQUIRY, UQ_MSC_NO_SYNC_CACHE), USB_QUIRK(TOSHIBA, TRANSMEMORY, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MSC_NO_PREVENT_ALLOW), USB_QUIRK(VIALABS, USB30SATABRIDGE, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE), /* Non-standard USB MIDI devices */ USB_QUIRK(ROLAND, UM1, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SC8850, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SD90, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, UM880N, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, UA100, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, UM4, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, U8, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, UM2, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SC8820, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, PC300, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SK500, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SCD70, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, UM550, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SD20, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, SD80, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(ROLAND, UA700, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(EGO, M4U, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI), USB_QUIRK(LOGILINK, U2M, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI), USB_QUIRK(MEDELI, DD305, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(REDOCTANE, GHMIDI, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI), USB_QUIRK(TEXTECH, U2M_1, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI), USB_QUIRK(TEXTECH, U2M_2, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI), USB_QUIRK(WCH2, U2M, 0x0000, 0xffff, UQ_SINGLE_CMD_MIDI), /* Non-standard USB AUDIO devices */ USB_QUIRK(MAUDIO, FASTTRACKULTRA, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), USB_QUIRK(MAUDIO, FASTTRACKULTRA8R, 0x0000, 0xffff, UQ_AU_VENDOR_CLASS), /* * Quirks for manufacturers which USB devices does not respond * after issuing non-supported commands: */ USB_QUIRK(ALCOR, DUMMY, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MSC_NO_TEST_UNIT_READY, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(APPLE, DUMMY, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(FEIYA, DUMMY, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(REALTEK, DUMMY, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MATCH_VENDOR_ONLY), USB_QUIRK(INITIO, DUMMY, 0x0000, 0xffff, UQ_MSC_NO_SYNC_CACHE, UQ_MATCH_VENDOR_ONLY), }; #undef USB_QUIRK_VP #undef USB_QUIRK static const char *usb_quirk_str[USB_QUIRK_MAX] = { [UQ_NONE] = "UQ_NONE", [UQ_MATCH_VENDOR_ONLY] = "UQ_MATCH_VENDOR_ONLY", [UQ_AUDIO_SWAP_LR] = "UQ_AUDIO_SWAP_LR", [UQ_AU_INP_ASYNC] = "UQ_AU_INP_ASYNC", [UQ_AU_NO_FRAC] = "UQ_AU_NO_FRAC", [UQ_AU_NO_XU] = "UQ_AU_NO_XU", [UQ_BAD_ADC] = "UQ_BAD_ADC", [UQ_BAD_AUDIO] = "UQ_BAD_AUDIO", [UQ_BROKEN_BIDIR] = "UQ_BROKEN_BIDIR", [UQ_BUS_POWERED] = "UQ_BUS_POWERED", [UQ_HID_IGNORE] = "UQ_HID_IGNORE", [UQ_KBD_IGNORE] = "UQ_KBD_IGNORE", [UQ_KBD_BOOTPROTO] = "UQ_KBD_BOOTPROTO", [UQ_UMS_IGNORE] = "UQ_UMS_IGNORE", [UQ_MS_BAD_CLASS] = "UQ_MS_BAD_CLASS", [UQ_MS_LEADING_BYTE] = "UQ_MS_LEADING_BYTE", [UQ_MS_REVZ] = "UQ_MS_REVZ", [UQ_NO_STRINGS] = "UQ_NO_STRINGS", [UQ_POWER_CLAIM] = "UQ_POWER_CLAIM", [UQ_SPUR_BUT_UP] = "UQ_SPUR_BUT_UP", [UQ_SWAP_UNICODE] = "UQ_SWAP_UNICODE", [UQ_CFG_INDEX_1] = "UQ_CFG_INDEX_1", [UQ_CFG_INDEX_2] = "UQ_CFG_INDEX_2", [UQ_CFG_INDEX_3] = "UQ_CFG_INDEX_3", [UQ_CFG_INDEX_4] = "UQ_CFG_INDEX_4", [UQ_CFG_INDEX_0] = "UQ_CFG_INDEX_0", [UQ_ASSUME_CM_OVER_DATA] = "UQ_ASSUME_CM_OVER_DATA", [UQ_MSC_NO_TEST_UNIT_READY] = "UQ_MSC_NO_TEST_UNIT_READY", [UQ_MSC_NO_RS_CLEAR_UA] = "UQ_MSC_NO_RS_CLEAR_UA", [UQ_MSC_NO_START_STOP] = "UQ_MSC_NO_START_STOP", [UQ_MSC_NO_GETMAXLUN] = "UQ_MSC_NO_GETMAXLUN", [UQ_MSC_NO_INQUIRY] = "UQ_MSC_NO_INQUIRY", [UQ_MSC_NO_INQUIRY_EVPD] = "UQ_MSC_NO_INQUIRY_EVPD", [UQ_MSC_NO_PREVENT_ALLOW] = "UQ_MSC_NO_PREVENT_ALLOW", [UQ_MSC_NO_SYNC_CACHE] = "UQ_MSC_NO_SYNC_CACHE", [UQ_MSC_SHUTTLE_INIT] = "UQ_MSC_SHUTTLE_INIT", [UQ_MSC_ALT_IFACE_1] = "UQ_MSC_ALT_IFACE_1", [UQ_MSC_FLOPPY_SPEED] = "UQ_MSC_FLOPPY_SPEED", [UQ_MSC_IGNORE_RESIDUE] = "UQ_MSC_IGNORE_RESIDUE", [UQ_MSC_WRONG_CSWSIG] = "UQ_MSC_WRONG_CSWSIG", [UQ_MSC_RBC_PAD_TO_12] = "UQ_MSC_RBC_PAD_TO_12", [UQ_MSC_READ_CAP_OFFBY1] = "UQ_MSC_READ_CAP_OFFBY1", [UQ_MSC_FORCE_SHORT_INQ] = "UQ_MSC_FORCE_SHORT_INQ", [UQ_MSC_FORCE_WIRE_BBB] = "UQ_MSC_FORCE_WIRE_BBB", [UQ_MSC_FORCE_WIRE_CBI] = "UQ_MSC_FORCE_WIRE_CBI", [UQ_MSC_FORCE_WIRE_CBI_I] = "UQ_MSC_FORCE_WIRE_CBI_I", [UQ_MSC_FORCE_PROTO_SCSI] = "UQ_MSC_FORCE_PROTO_SCSI", [UQ_MSC_FORCE_PROTO_ATAPI] = "UQ_MSC_FORCE_PROTO_ATAPI", [UQ_MSC_FORCE_PROTO_UFI] = "UQ_MSC_FORCE_PROTO_UFI", [UQ_MSC_FORCE_PROTO_RBC] = "UQ_MSC_FORCE_PROTO_RBC", [UQ_MSC_EJECT_HUAWEI] = "UQ_MSC_EJECT_HUAWEI", [UQ_MSC_EJECT_SIERRA] = "UQ_MSC_EJECT_SIERRA", [UQ_MSC_EJECT_SCSIEJECT] = "UQ_MSC_EJECT_SCSIEJECT", [UQ_MSC_EJECT_REZERO] = "UQ_MSC_EJECT_REZERO", [UQ_MSC_EJECT_ZTESTOR] = "UQ_MSC_EJECT_ZTESTOR", [UQ_MSC_EJECT_CMOTECH] = "UQ_MSC_EJECT_CMOTECH", [UQ_MSC_EJECT_WAIT] = "UQ_MSC_EJECT_WAIT", [UQ_MSC_EJECT_SAEL_M460] = "UQ_MSC_EJECT_SAEL_M460", [UQ_MSC_EJECT_HUAWEISCSI] = "UQ_MSC_EJECT_HUAWEISCSI", [UQ_MSC_EJECT_HUAWEISCSI2] = "UQ_MSC_EJECT_HUAWEISCSI2", [UQ_MSC_EJECT_TCT] = "UQ_MSC_EJECT_TCT", [UQ_BAD_MIDI] = "UQ_BAD_MIDI", [UQ_AU_VENDOR_CLASS] = "UQ_AU_VENDOR_CLASS", [UQ_SINGLE_CMD_MIDI] = "UQ_SINGLE_CMD_MIDI", }; /*------------------------------------------------------------------------* * usb_quirkstr * * This function converts an USB quirk code into a string. *------------------------------------------------------------------------*/ static const char * usb_quirkstr(uint16_t quirk) { return ((quirk < USB_QUIRK_MAX) ? usb_quirk_str[quirk] : "USB_QUIRK_UNKNOWN"); } /*------------------------------------------------------------------------* * usb_test_quirk_by_info * * Returns: * 0: Quirk not found * Else: Quirk found *------------------------------------------------------------------------*/ static uint8_t usb_test_quirk_by_info(const struct usbd_lookup_info *info, uint16_t quirk) { uint16_t x; uint16_t y; if (quirk == UQ_NONE) goto done; mtx_lock(&usb_quirk_mtx); for (x = 0; x != USB_DEV_QUIRKS_MAX; x++) { /* see if quirk information does not match */ if ((usb_quirks[x].vid != info->idVendor) || (usb_quirks[x].lo_rev > info->bcdDevice) || (usb_quirks[x].hi_rev < info->bcdDevice)) { continue; } /* see if quirk only should match vendor ID */ if (usb_quirks[x].pid != info->idProduct) { if (usb_quirks[x].pid != 0) continue; for (y = 0; y != USB_SUB_QUIRKS_MAX; y++) { if (usb_quirks[x].quirks[y] == UQ_MATCH_VENDOR_ONLY) break; } if (y == USB_SUB_QUIRKS_MAX) continue; } /* lookup quirk */ for (y = 0; y != USB_SUB_QUIRKS_MAX; y++) { if (usb_quirks[x].quirks[y] == quirk) { mtx_unlock(&usb_quirk_mtx); DPRINTF("Found quirk '%s'.\n", usb_quirkstr(quirk)); return (1); } } /* no quirk found */ break; } mtx_unlock(&usb_quirk_mtx); done: return (0); /* no quirk match */ } static struct usb_quirk_entry * usb_quirk_get_entry(uint16_t vid, uint16_t pid, uint16_t lo_rev, uint16_t hi_rev, uint8_t do_alloc) { uint16_t x; mtx_assert(&usb_quirk_mtx, MA_OWNED); if ((vid | pid | lo_rev | hi_rev) == 0) { /* all zero - special case */ return (usb_quirks + USB_DEV_QUIRKS_MAX - 1); } /* search for an existing entry */ for (x = 0; x != USB_DEV_QUIRKS_MAX; x++) { /* see if quirk information does not match */ if ((usb_quirks[x].vid != vid) || (usb_quirks[x].pid != pid) || (usb_quirks[x].lo_rev != lo_rev) || (usb_quirks[x].hi_rev != hi_rev)) { continue; } return (usb_quirks + x); } if (do_alloc == 0) { /* no match */ return (NULL); } /* search for a free entry */ for (x = 0; x != USB_DEV_QUIRKS_MAX; x++) { /* see if quirk information does not match */ if ((usb_quirks[x].vid | usb_quirks[x].pid | usb_quirks[x].lo_rev | usb_quirks[x].hi_rev) != 0) { continue; } usb_quirks[x].vid = vid; usb_quirks[x].pid = pid; usb_quirks[x].lo_rev = lo_rev; usb_quirks[x].hi_rev = hi_rev; return (usb_quirks + x); } /* no entry found */ return (NULL); } /*------------------------------------------------------------------------* * usb_quirk_ioctl - handle quirk IOCTLs * * Returns: * 0: Success * Else: Failure *------------------------------------------------------------------------*/ static int usb_quirk_ioctl(unsigned long cmd, caddr_t data, int fflag, struct thread *td) { struct usb_gen_quirk *pgq; struct usb_quirk_entry *pqe; uint32_t x; uint32_t y; int err; switch (cmd) { case USB_DEV_QUIRK_GET: pgq = (void *)data; x = pgq->index % USB_SUB_QUIRKS_MAX; y = pgq->index / USB_SUB_QUIRKS_MAX; if (y >= USB_DEV_QUIRKS_MAX) { return (EINVAL); } mtx_lock(&usb_quirk_mtx); /* copy out data */ pgq->vid = usb_quirks[y].vid; pgq->pid = usb_quirks[y].pid; pgq->bcdDeviceLow = usb_quirks[y].lo_rev; pgq->bcdDeviceHigh = usb_quirks[y].hi_rev; strlcpy(pgq->quirkname, usb_quirkstr(usb_quirks[y].quirks[x]), sizeof(pgq->quirkname)); mtx_unlock(&usb_quirk_mtx); return (0); /* success */ case USB_QUIRK_NAME_GET: pgq = (void *)data; x = pgq->index; if (x >= USB_QUIRK_MAX) { return (EINVAL); } strlcpy(pgq->quirkname, usb_quirkstr(x), sizeof(pgq->quirkname)); return (0); /* success */ case USB_DEV_QUIRK_ADD: pgq = (void *)data; /* check privileges */ err = priv_check(curthread, PRIV_DRIVER); if (err) { return (err); } /* convert quirk string into numerical */ for (y = 0; y != USB_DEV_QUIRKS_MAX; y++) { if (strcmp(pgq->quirkname, usb_quirkstr(y)) == 0) { break; } } if (y == USB_DEV_QUIRKS_MAX) { return (EINVAL); } if (y == UQ_NONE) { return (EINVAL); } mtx_lock(&usb_quirk_mtx); pqe = usb_quirk_get_entry(pgq->vid, pgq->pid, pgq->bcdDeviceLow, pgq->bcdDeviceHigh, 1); if (pqe == NULL) { mtx_unlock(&usb_quirk_mtx); return (EINVAL); } for (x = 0; x != USB_SUB_QUIRKS_MAX; x++) { if (pqe->quirks[x] == UQ_NONE) { pqe->quirks[x] = y; break; } } mtx_unlock(&usb_quirk_mtx); if (x == USB_SUB_QUIRKS_MAX) { return (ENOMEM); } return (0); /* success */ case USB_DEV_QUIRK_REMOVE: pgq = (void *)data; /* check privileges */ err = priv_check(curthread, PRIV_DRIVER); if (err) { return (err); } /* convert quirk string into numerical */ for (y = 0; y != USB_DEV_QUIRKS_MAX; y++) { if (strcmp(pgq->quirkname, usb_quirkstr(y)) == 0) { break; } } if (y == USB_DEV_QUIRKS_MAX) { return (EINVAL); } if (y == UQ_NONE) { return (EINVAL); } mtx_lock(&usb_quirk_mtx); pqe = usb_quirk_get_entry(pgq->vid, pgq->pid, pgq->bcdDeviceLow, pgq->bcdDeviceHigh, 0); if (pqe == NULL) { mtx_unlock(&usb_quirk_mtx); return (EINVAL); } for (x = 0; x != USB_SUB_QUIRKS_MAX; x++) { if (pqe->quirks[x] == y) { pqe->quirks[x] = UQ_NONE; break; } } if (x == USB_SUB_QUIRKS_MAX) { mtx_unlock(&usb_quirk_mtx); return (ENOMEM); } for (x = 0; x != USB_SUB_QUIRKS_MAX; x++) { if (pqe->quirks[x] != UQ_NONE) { break; } } if (x == USB_SUB_QUIRKS_MAX) { /* all quirk entries are unused - release */ memset(pqe, 0, sizeof(*pqe)); } mtx_unlock(&usb_quirk_mtx); return (0); /* success */ default: break; } return (ENOIOCTL); } static void usb_quirk_init(void *arg) { /* initialize mutex */ mtx_init(&usb_quirk_mtx, "USB quirk", NULL, MTX_DEF); /* register our function */ usb_test_quirk_p = &usb_test_quirk_by_info; usb_quirk_ioctl_p = &usb_quirk_ioctl; } static void usb_quirk_uninit(void *arg) { usb_quirk_unload(arg); /* destroy mutex */ mtx_destroy(&usb_quirk_mtx); } SYSINIT(usb_quirk_init, SI_SUB_LOCK, SI_ORDER_FIRST, usb_quirk_init, NULL); SYSUNINIT(usb_quirk_uninit, SI_SUB_LOCK, SI_ORDER_ANY, usb_quirk_uninit, NULL); Index: projects/building-blocks/sys/dev/usb/serial/u3g.c =================================================================== --- projects/building-blocks/sys/dev/usb/serial/u3g.c (revision 275663) +++ projects/building-blocks/sys/dev/usb/serial/u3g.c (revision 275664) @@ -1,1213 +1,1217 @@ /* * Copyright (c) 2008 AnyWi Technologies * Author: Andrea Guzzo * * based on uark.c 1.1 2006/08/14 08:30:22 jsg * * * parts from ubsa.c 183348 2008-09-25 12:00:56Z phk * * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. * * $FreeBSD$ */ /* * NOTE: * * - The detour through the tty layer is ridiculously expensive wrt * buffering due to the high speeds. * * We should consider adding a simple r/w device which allows * attaching of PPP in a more efficient way. * */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "usbdevs.h" #define USB_DEBUG_VAR u3g_debug #include #include #include #include #include #ifdef USB_DEBUG static int u3g_debug = 0; static SYSCTL_NODE(_hw_usb, OID_AUTO, u3g, CTLFLAG_RW, 0, "USB 3g"); SYSCTL_INT(_hw_usb_u3g, OID_AUTO, debug, CTLFLAG_RW, &u3g_debug, 0, "Debug level"); #endif #define U3G_MAXPORTS 12 #define U3G_CONFIG_INDEX 0 #define U3G_BSIZE 2048 #define U3G_TXSIZE (U3G_BSIZE / U3G_TXFRAMES) #define U3G_TXFRAMES 4 /* Eject methods; See also usb_quirks.h:UQ_MSC_EJECT_* */ #define U3GINIT_HUAWEI 1 /* Requires Huawei init command */ #define U3GINIT_SIERRA 2 /* Requires Sierra init command */ #define U3GINIT_SCSIEJECT 3 /* Requires SCSI eject command */ #define U3GINIT_REZERO 4 /* Requires SCSI rezero command */ #define U3GINIT_ZTESTOR 5 /* Requires ZTE SCSI command */ #define U3GINIT_CMOTECH 6 /* Requires CMOTECH SCSI command */ #define U3GINIT_WAIT 7 /* Device reappears after a delay */ #define U3GINIT_SAEL_M460 8 /* Requires vendor init */ #define U3GINIT_HUAWEISCSI 9 /* Requires Huawei SCSI init command */ #define U3GINIT_HUAWEISCSI2 10 /* Requires Huawei SCSI init command (2) */ #define U3GINIT_TCT 11 /* Requires TCT Mobile init command */ enum { U3G_BULK_WR, U3G_BULK_RD, U3G_INTR, U3G_N_TRANSFER, }; struct u3g_softc { struct ucom_super_softc sc_super_ucom; struct ucom_softc sc_ucom[U3G_MAXPORTS]; struct usb_xfer *sc_xfer[U3G_MAXPORTS][U3G_N_TRANSFER]; uint8_t sc_iface[U3G_MAXPORTS]; /* local status register */ uint8_t sc_lsr[U3G_MAXPORTS]; /* local status register */ uint8_t sc_msr[U3G_MAXPORTS]; /* u3g status register */ uint16_t sc_line[U3G_MAXPORTS]; /* line status */ struct usb_device *sc_udev; struct mtx sc_mtx; uint8_t sc_numports; }; static device_probe_t u3g_probe; static device_attach_t u3g_attach; static device_detach_t u3g_detach; static void u3g_free_softc(struct u3g_softc *); static usb_callback_t u3g_write_callback; static usb_callback_t u3g_read_callback; static usb_callback_t u3g_intr_callback; static void u3g_cfg_get_status(struct ucom_softc *, uint8_t *, uint8_t *); static void u3g_cfg_set_dtr(struct ucom_softc *, uint8_t); static void u3g_cfg_set_rts(struct ucom_softc *, uint8_t); static void u3g_start_read(struct ucom_softc *ucom); static void u3g_stop_read(struct ucom_softc *ucom); static void u3g_start_write(struct ucom_softc *ucom); static void u3g_stop_write(struct ucom_softc *ucom); static void u3g_poll(struct ucom_softc *ucom); static void u3g_free(struct ucom_softc *ucom); static void u3g_test_autoinst(void *, struct usb_device *, struct usb_attach_arg *); static int u3g_driver_loaded(struct module *mod, int what, void *arg); static eventhandler_tag u3g_etag; static const struct usb_config u3g_config[U3G_N_TRANSFER] = { [U3G_BULK_WR] = { .type = UE_BULK, .endpoint = UE_ADDR_ANY, .direction = UE_DIR_OUT, .bufsize = U3G_BSIZE,/* bytes */ .frames = U3G_TXFRAMES, .flags = {.pipe_bof = 1,.force_short_xfer = 1,}, .callback = &u3g_write_callback, }, [U3G_BULK_RD] = { .type = UE_BULK, .endpoint = UE_ADDR_ANY, .direction = UE_DIR_IN, .bufsize = U3G_BSIZE,/* bytes */ .flags = {.pipe_bof = 1,.short_xfer_ok = 1,}, .callback = &u3g_read_callback, }, [U3G_INTR] = { .type = UE_INTERRUPT, .endpoint = UE_ADDR_ANY, .direction = UE_DIR_IN, .flags = {.pipe_bof = 1,.short_xfer_ok = 1,.no_pipe_ok = 1,}, .bufsize = 0, /* use wMaxPacketSize */ .callback = &u3g_intr_callback, }, }; static const struct ucom_callback u3g_callback = { .ucom_cfg_get_status = &u3g_cfg_get_status, .ucom_cfg_set_dtr = &u3g_cfg_set_dtr, .ucom_cfg_set_rts = &u3g_cfg_set_rts, .ucom_start_read = &u3g_start_read, .ucom_stop_read = &u3g_stop_read, .ucom_start_write = &u3g_start_write, .ucom_stop_write = &u3g_stop_write, .ucom_poll = &u3g_poll, .ucom_free = &u3g_free, }; static device_method_t u3g_methods[] = { DEVMETHOD(device_probe, u3g_probe), DEVMETHOD(device_attach, u3g_attach), DEVMETHOD(device_detach, u3g_detach), DEVMETHOD_END }; static devclass_t u3g_devclass; static driver_t u3g_driver = { .name = "u3g", .methods = u3g_methods, .size = sizeof(struct u3g_softc), }; DRIVER_MODULE(u3g, uhub, u3g_driver, u3g_devclass, u3g_driver_loaded, 0); MODULE_DEPEND(u3g, ucom, 1, 1, 1); MODULE_DEPEND(u3g, usb, 1, 1, 1); MODULE_VERSION(u3g, 1); static const STRUCT_USB_HOST_ID u3g_devs[] = { #define U3G_DEV(v,p,i) { USB_VPI(USB_VENDOR_##v, USB_PRODUCT_##v##_##p, i) } U3G_DEV(ACERP, H10, 0), U3G_DEV(AIRPLUS, MCD650, 0), U3G_DEV(AIRPRIME, PC5220, 0), U3G_DEV(AIRPRIME, AC313U, 0), U3G_DEV(ALINK, 3G, 0), U3G_DEV(ALINK, 3GU, 0), U3G_DEV(ALINK, DWM652U5, 0), U3G_DEV(AMOI, H01, 0), U3G_DEV(AMOI, H01A, 0), U3G_DEV(AMOI, H02, 0), U3G_DEV(ANYDATA, ADU_500A, 0), U3G_DEV(ANYDATA, ADU_620UW, 0), U3G_DEV(ANYDATA, ADU_E100X, 0), U3G_DEV(AXESSTEL, DATAMODEM, 0), U3G_DEV(CMOTECH, CDMA_MODEM1, 0), U3G_DEV(CMOTECH, CGU628, U3GINIT_CMOTECH), U3G_DEV(DELL, U5500, 0), U3G_DEV(DELL, U5505, 0), U3G_DEV(DELL, U5510, 0), U3G_DEV(DELL, U5520, 0), U3G_DEV(DELL, U5520_2, 0), U3G_DEV(DELL, U5520_3, 0), U3G_DEV(DELL, U5700, 0), U3G_DEV(DELL, U5700_2, 0), U3G_DEV(DELL, U5700_3, 0), U3G_DEV(DELL, U5700_4, 0), U3G_DEV(DELL, U5720, 0), U3G_DEV(DELL, U5720_2, 0), U3G_DEV(DELL, U5730, 0), U3G_DEV(DELL, U5730_2, 0), U3G_DEV(DELL, U5730_3, 0), U3G_DEV(DELL, U740, 0), U3G_DEV(DLINK, DWR510_CD, U3GINIT_SCSIEJECT), U3G_DEV(DLINK, DWR510, 0), U3G_DEV(DLINK3, DWM652, 0), U3G_DEV(HP, EV2200, 0), U3G_DEV(HP, HS2300, 0), U3G_DEV(HP, UN2420_QDL, 0), U3G_DEV(HP, UN2420, 0), U3G_DEV(HUAWEI, E1401, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1402, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1403, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1404, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1405, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1406, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1407, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1408, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1409, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E140A, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E140B, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E140D, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E140E, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E140F, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1410, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1411, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1412, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1413, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1414, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1415, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1416, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1417, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1418, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1419, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E141A, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E141B, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E141C, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E141D, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E141E, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E141F, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1420, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1421, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1422, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1423, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1424, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1425, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1426, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1427, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1428, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1429, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E142A, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E142B, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E142C, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E142D, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E142E, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E142F, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1430, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1431, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1432, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1433, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1434, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1435, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1436, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1437, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1438, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1439, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E143A, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E143B, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E143C, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E143D, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E143E, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E143F, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E173, 0), U3G_DEV(HUAWEI, E173_INIT, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, E3131, 0), U3G_DEV(HUAWEI, E3131_INIT, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, E180V, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E220, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E220BIS, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E392, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, MOBILE, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, E1752, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, E1820, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, K3772, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, K3772_INIT, U3GINIT_HUAWEISCSI2), U3G_DEV(HUAWEI, K3765, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, K3765_INIT, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, K3770, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, K3770_INIT, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, K4505, U3GINIT_HUAWEI), U3G_DEV(HUAWEI, K4505_INIT, U3GINIT_HUAWEISCSI), U3G_DEV(HUAWEI, ETS2055, U3GINIT_HUAWEI), U3G_DEV(KYOCERA2, CDMA_MSM_K, 0), U3G_DEV(KYOCERA2, KPC680, 0), U3G_DEV(LONGCHEER, WM66, U3GINIT_HUAWEI), U3G_DEV(LONGCHEER, DISK, U3GINIT_TCT), U3G_DEV(LONGCHEER, W14, 0), U3G_DEV(LONGCHEER, XSSTICK, 0), U3G_DEV(MERLIN, V620, 0), U3G_DEV(NEOTEL, PRIME, 0), U3G_DEV(NOVATEL, E725, 0), U3G_DEV(NOVATEL, ES620, 0), U3G_DEV(NOVATEL, ES620_2, 0), U3G_DEV(NOVATEL, EU730, 0), U3G_DEV(NOVATEL, EU740, 0), U3G_DEV(NOVATEL, EU870D, 0), U3G_DEV(NOVATEL, MC760, 0), U3G_DEV(NOVATEL, MC547, 0), U3G_DEV(NOVATEL, MC679, 0), U3G_DEV(NOVATEL, MC950D, 0), U3G_DEV(NOVATEL, MC990D, 0), U3G_DEV(NOVATEL, MIFI2200, U3GINIT_SCSIEJECT), U3G_DEV(NOVATEL, MIFI2200V, U3GINIT_SCSIEJECT), U3G_DEV(NOVATEL, U720, 0), U3G_DEV(NOVATEL, U727, 0), U3G_DEV(NOVATEL, U727_2, 0), U3G_DEV(NOVATEL, U740, 0), U3G_DEV(NOVATEL, U740_2, 0), U3G_DEV(NOVATEL, U760, U3GINIT_SCSIEJECT), U3G_DEV(NOVATEL, U870, 0), U3G_DEV(NOVATEL, V620, 0), U3G_DEV(NOVATEL, V640, 0), U3G_DEV(NOVATEL, V720, 0), U3G_DEV(NOVATEL, V740, 0), U3G_DEV(NOVATEL, X950D, 0), U3G_DEV(NOVATEL, XU870, 0), U3G_DEV(MOTOROLA2, MB886, U3GINIT_SCSIEJECT), U3G_DEV(OPTION, E6500, 0), U3G_DEV(OPTION, E6501, 0), U3G_DEV(OPTION, E6601, 0), U3G_DEV(OPTION, E6721, 0), U3G_DEV(OPTION, E6741, 0), U3G_DEV(OPTION, E6761, 0), U3G_DEV(OPTION, E6800, 0), U3G_DEV(OPTION, E7021, 0), U3G_DEV(OPTION, E7041, 0), U3G_DEV(OPTION, E7061, 0), U3G_DEV(OPTION, E7100, 0), U3G_DEV(OPTION, GE40X, 0), U3G_DEV(OPTION, GT3G, 0), U3G_DEV(OPTION, GT3GPLUS, 0), U3G_DEV(OPTION, GT3GQUAD, 0), U3G_DEV(OPTION, GT3G_1, 0), U3G_DEV(OPTION, GT3G_2, 0), U3G_DEV(OPTION, GT3G_3, 0), U3G_DEV(OPTION, GT3G_4, 0), U3G_DEV(OPTION, GT3G_5, 0), U3G_DEV(OPTION, GT3G_6, 0), U3G_DEV(OPTION, GTHSDPA, 0), U3G_DEV(OPTION, GTM380, 0), U3G_DEV(OPTION, GTMAX36, 0), U3G_DEV(OPTION, GTMAX380HSUPAE, 0), U3G_DEV(OPTION, GTMAXHSUPA, 0), U3G_DEV(OPTION, GTMAXHSUPAE, 0), U3G_DEV(OPTION, VODAFONEMC3G, 0), U3G_DEV(QISDA, H20_1, 0), U3G_DEV(QISDA, H20_2, 0), U3G_DEV(QISDA, H21_1, 0), U3G_DEV(QISDA, H21_2, 0), U3G_DEV(QUALCOMM, NTT_L02C_MODEM, U3GINIT_SCSIEJECT), U3G_DEV(QUALCOMM2, AC8700, 0), U3G_DEV(QUALCOMM2, MF330, 0), U3G_DEV(QUALCOMM2, SIM5218, 0), U3G_DEV(QUALCOMM2, VW110L, U3GINIT_SCSIEJECT), U3G_DEV(QUALCOMM2, GOBI2000_QDL, 0), U3G_DEV(QUALCOMM2, GOBI2000, 0), + U3G_DEV(QUALCOMM2, VT80N, 0), + U3G_DEV(QUALCOMM3, VFAST2, 0), U3G_DEV(QUALCOMMINC, AC2726, 0), + U3G_DEV(QUALCOMMINC, AC682_INIT, U3GINIT_SCSIEJECT), + U3G_DEV(QUALCOMMINC, AC682, 0), U3G_DEV(QUALCOMMINC, AC8700, 0), U3G_DEV(QUALCOMMINC, AC8710, 0), U3G_DEV(QUALCOMMINC, CDMA_MSM, U3GINIT_SCSIEJECT), U3G_DEV(QUALCOMMINC, E0002, 0), U3G_DEV(QUALCOMMINC, E0003, 0), U3G_DEV(QUALCOMMINC, E0004, 0), U3G_DEV(QUALCOMMINC, E0005, 0), U3G_DEV(QUALCOMMINC, E0006, 0), U3G_DEV(QUALCOMMINC, E0007, 0), U3G_DEV(QUALCOMMINC, E0008, 0), U3G_DEV(QUALCOMMINC, E0009, 0), U3G_DEV(QUALCOMMINC, E000A, 0), U3G_DEV(QUALCOMMINC, E000B, 0), U3G_DEV(QUALCOMMINC, E000C, 0), U3G_DEV(QUALCOMMINC, E000D, 0), U3G_DEV(QUALCOMMINC, E000E, 0), U3G_DEV(QUALCOMMINC, E000F, 0), U3G_DEV(QUALCOMMINC, E0010, 0), U3G_DEV(QUALCOMMINC, E0011, 0), U3G_DEV(QUALCOMMINC, E0012, 0), U3G_DEV(QUALCOMMINC, E0013, 0), U3G_DEV(QUALCOMMINC, E0014, 0), U3G_DEV(QUALCOMMINC, E0017, 0), U3G_DEV(QUALCOMMINC, E0018, 0), U3G_DEV(QUALCOMMINC, E0019, 0), U3G_DEV(QUALCOMMINC, E0020, 0), U3G_DEV(QUALCOMMINC, E0021, 0), U3G_DEV(QUALCOMMINC, E0022, 0), U3G_DEV(QUALCOMMINC, E0023, 0), U3G_DEV(QUALCOMMINC, E0024, 0), U3G_DEV(QUALCOMMINC, E0025, 0), U3G_DEV(QUALCOMMINC, E0026, 0), U3G_DEV(QUALCOMMINC, E0027, 0), U3G_DEV(QUALCOMMINC, E0028, 0), U3G_DEV(QUALCOMMINC, E0029, 0), U3G_DEV(QUALCOMMINC, E0030, 0), U3G_DEV(QUALCOMMINC, E0032, 0), U3G_DEV(QUALCOMMINC, E0033, 0), U3G_DEV(QUALCOMMINC, E0037, 0), U3G_DEV(QUALCOMMINC, E0039, 0), U3G_DEV(QUALCOMMINC, E0042, 0), U3G_DEV(QUALCOMMINC, E0043, 0), U3G_DEV(QUALCOMMINC, E0048, 0), U3G_DEV(QUALCOMMINC, E0049, 0), U3G_DEV(QUALCOMMINC, E0051, 0), U3G_DEV(QUALCOMMINC, E0052, 0), U3G_DEV(QUALCOMMINC, E0054, 0), U3G_DEV(QUALCOMMINC, E0055, 0), U3G_DEV(QUALCOMMINC, E0057, 0), U3G_DEV(QUALCOMMINC, E0058, 0), U3G_DEV(QUALCOMMINC, E0059, 0), U3G_DEV(QUALCOMMINC, E0060, 0), U3G_DEV(QUALCOMMINC, E0061, 0), U3G_DEV(QUALCOMMINC, E0062, 0), U3G_DEV(QUALCOMMINC, E0063, 0), U3G_DEV(QUALCOMMINC, E0064, 0), U3G_DEV(QUALCOMMINC, E0066, 0), U3G_DEV(QUALCOMMINC, E0069, 0), U3G_DEV(QUALCOMMINC, E0070, 0), U3G_DEV(QUALCOMMINC, E0073, 0), U3G_DEV(QUALCOMMINC, E0076, 0), U3G_DEV(QUALCOMMINC, E0078, 0), U3G_DEV(QUALCOMMINC, E0082, 0), U3G_DEV(QUALCOMMINC, E0086, 0), U3G_DEV(QUALCOMMINC, SURFSTICK, 0), U3G_DEV(QUALCOMMINC, E2002, 0), U3G_DEV(QUALCOMMINC, E2003, 0), U3G_DEV(QUALCOMMINC, K3772_Z, 0), U3G_DEV(QUALCOMMINC, K3772_Z_INIT, U3GINIT_SCSIEJECT), U3G_DEV(QUALCOMMINC, MF195E, 0), U3G_DEV(QUALCOMMINC, MF195E_INIT, U3GINIT_SCSIEJECT), U3G_DEV(QUALCOMMINC, MF626, 0), U3G_DEV(QUALCOMMINC, MF628, 0), U3G_DEV(QUALCOMMINC, MF633R, 0), U3G_DEV(QUANTA, GKE, 0), U3G_DEV(QUANTA, GLE, 0), U3G_DEV(QUANTA, GLX, 0), U3G_DEV(QUANTA, Q101, 0), U3G_DEV(QUANTA, Q111, 0), U3G_DEV(SIERRA, AC402, 0), U3G_DEV(SIERRA, AC595U, 0), U3G_DEV(SIERRA, AC313U, 0), U3G_DEV(SIERRA, AC597E, 0), U3G_DEV(SIERRA, AC875E, 0), U3G_DEV(SIERRA, AC875U, 0), U3G_DEV(SIERRA, AC875U_2, 0), U3G_DEV(SIERRA, AC880, 0), U3G_DEV(SIERRA, AC880E, 0), U3G_DEV(SIERRA, AC880U, 0), U3G_DEV(SIERRA, AC881, 0), U3G_DEV(SIERRA, AC881E, 0), U3G_DEV(SIERRA, AC881U, 0), U3G_DEV(SIERRA, AC885E, 0), U3G_DEV(SIERRA, AC885E_2, 0), U3G_DEV(SIERRA, AC885U, 0), U3G_DEV(SIERRA, AIRCARD580, 0), U3G_DEV(SIERRA, AIRCARD595, 0), U3G_DEV(SIERRA, AIRCARD875, 0), U3G_DEV(SIERRA, C22, 0), U3G_DEV(SIERRA, C597, 0), U3G_DEV(SIERRA, C888, 0), U3G_DEV(SIERRA, E0029, 0), U3G_DEV(SIERRA, E6892, 0), U3G_DEV(SIERRA, E6893, 0), U3G_DEV(SIERRA, EM5625, 0), U3G_DEV(SIERRA, EM5725, 0), U3G_DEV(SIERRA, MC5720, 0), U3G_DEV(SIERRA, MC5720_2, 0), U3G_DEV(SIERRA, MC5725, 0), U3G_DEV(SIERRA, MC5727, 0), U3G_DEV(SIERRA, MC5727_2, 0), U3G_DEV(SIERRA, MC5728, 0), U3G_DEV(SIERRA, MC8700, 0), U3G_DEV(SIERRA, MC8755, 0), U3G_DEV(SIERRA, MC8755_2, 0), U3G_DEV(SIERRA, MC8755_3, 0), U3G_DEV(SIERRA, MC8755_4, 0), U3G_DEV(SIERRA, MC8765, 0), U3G_DEV(SIERRA, MC8765_2, 0), U3G_DEV(SIERRA, MC8765_3, 0), U3G_DEV(SIERRA, MC8775, 0), U3G_DEV(SIERRA, MC8775_2, 0), U3G_DEV(SIERRA, MC8780, 0), U3G_DEV(SIERRA, MC8780_2, 0), U3G_DEV(SIERRA, MC8780_3, 0), U3G_DEV(SIERRA, MC8781, 0), U3G_DEV(SIERRA, MC8781_2, 0), U3G_DEV(SIERRA, MC8781_3, 0), U3G_DEV(SIERRA, MC8785, 0), U3G_DEV(SIERRA, MC8785_2, 0), U3G_DEV(SIERRA, MC8790, 0), U3G_DEV(SIERRA, MC8791, 0), U3G_DEV(SIERRA, MC8792, 0), U3G_DEV(SIERRA, MINI5725, 0), U3G_DEV(SIERRA, T11, 0), U3G_DEV(SIERRA, T598, 0), U3G_DEV(SILABS, SAEL, U3GINIT_SAEL_M460), U3G_DEV(STELERA, C105, 0), U3G_DEV(STELERA, E1003, 0), U3G_DEV(STELERA, E1004, 0), U3G_DEV(STELERA, E1005, 0), U3G_DEV(STELERA, E1006, 0), U3G_DEV(STELERA, E1007, 0), U3G_DEV(STELERA, E1008, 0), U3G_DEV(STELERA, E1009, 0), U3G_DEV(STELERA, E100A, 0), U3G_DEV(STELERA, E100B, 0), U3G_DEV(STELERA, E100C, 0), U3G_DEV(STELERA, E100D, 0), U3G_DEV(STELERA, E100E, 0), U3G_DEV(STELERA, E100F, 0), U3G_DEV(STELERA, E1010, 0), U3G_DEV(STELERA, E1011, 0), U3G_DEV(STELERA, E1012, 0), U3G_DEV(TCTMOBILE, X060S, 0), U3G_DEV(TCTMOBILE, X080S, U3GINIT_TCT), U3G_DEV(TELIT, UC864E, 0), U3G_DEV(TELIT, UC864G, 0), U3G_DEV(TLAYTECH, TEU800, 0), U3G_DEV(TOSHIBA, G450, 0), U3G_DEV(TOSHIBA, HSDPA, 0), U3G_DEV(YISO, C893, 0), U3G_DEV(WETELECOM, WM_D200, 0), /* Autoinstallers */ U3G_DEV(NOVATEL, ZEROCD, U3GINIT_SCSIEJECT), U3G_DEV(OPTION, GTICON322, U3GINIT_REZERO), U3G_DEV(QUALCOMMINC, ZTE_STOR, U3GINIT_ZTESTOR), U3G_DEV(QUALCOMMINC, ZTE_STOR2, U3GINIT_SCSIEJECT), U3G_DEV(QUANTA, Q101_STOR, U3GINIT_SCSIEJECT), U3G_DEV(SIERRA, TRUINSTALL, U3GINIT_SIERRA), #undef U3G_DEV }; static int u3g_sierra_init(struct usb_device *udev) { struct usb_device_request req; req.bmRequestType = UT_VENDOR; req.bRequest = UR_SET_INTERFACE; USETW(req.wValue, UF_DEVICE_REMOTE_WAKEUP); USETW(req.wIndex, UHF_PORT_CONNECTION); USETW(req.wLength, 0); if (usbd_do_request_flags(udev, NULL, &req, NULL, 0, NULL, USB_MS_HZ)) { /* ignore any errors */ } return (0); } static int u3g_huawei_init(struct usb_device *udev) { struct usb_device_request req; req.bmRequestType = UT_WRITE_DEVICE; req.bRequest = UR_SET_FEATURE; USETW(req.wValue, UF_DEVICE_REMOTE_WAKEUP); USETW(req.wIndex, UHF_PORT_SUSPEND); USETW(req.wLength, 0); if (usbd_do_request_flags(udev, NULL, &req, NULL, 0, NULL, USB_MS_HZ)) { /* ignore any errors */ } return (0); } static void u3g_sael_m460_init(struct usb_device *udev) { static const uint8_t setup[][24] = { { 0x41, 0x11, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x13, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x01, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }, { 0xc1, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x40, 0x02 }, { 0xc1, 0x08, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00 }, { 0x41, 0x07, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00 }, { 0xc1, 0x0f, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02 }, { 0x41, 0x01, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x07, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x03, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x19, 0x00, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x00, 0x00, 0x11, 0x13 }, { 0x41, 0x13, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x09, 0x00, 0x00, 0x00, 0x80, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00 }, { 0x41, 0x12, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x01, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x07, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x03, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00 }, { 0x41, 0x19, 0x00, 0x00, 0x00, 0x00, 0x06, 0x00, 0x00, 0x00, 0x00, 0x00, 0x11, 0x13 }, { 0x41, 0x13, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x09, 0x00, 0x00, 0x00, 0x80, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00, 0x0a, 0x00, 0x00, 0x00 }, { 0x41, 0x07, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00 }, }; struct usb_device_request req; usb_error_t err; uint16_t len; uint8_t buf[0x300]; uint8_t n; DPRINTFN(1, "\n"); if (usbd_req_set_alt_interface_no(udev, NULL, 0, 0)) { DPRINTFN(0, "Alt setting 0 failed\n"); return; } for (n = 0; n != (sizeof(setup)/sizeof(setup[0])); n++) { memcpy(&req, setup[n], sizeof(req)); len = UGETW(req.wLength); if (req.bmRequestType & UE_DIR_IN) { if (len > sizeof(buf)) { DPRINTFN(0, "too small buffer\n"); continue; } err = usbd_do_request(udev, NULL, &req, buf); } else { if (len > (sizeof(setup[0]) - 8)) { DPRINTFN(0, "too small buffer\n"); continue; } err = usbd_do_request(udev, NULL, &req, __DECONST(uint8_t *, &setup[n][8])); } if (err) { DPRINTFN(1, "request %u failed\n", (unsigned int)n); /* * Some of the requests will fail. Stop doing * requests when we are getting timeouts so * that we don't block the explore/attach * thread forever. */ if (err == USB_ERR_TIMEOUT) break; } } } /* * The following function handles 3G modem devices (E220, Mobile, * etc.) with auto-install flash disks for Windows/MacOSX on the first * interface. After some command or some delay they change appearance * to a modem. */ static void u3g_test_autoinst(void *arg, struct usb_device *udev, struct usb_attach_arg *uaa) { struct usb_interface *iface; struct usb_interface_descriptor *id; int error; unsigned long method; if (uaa->dev_state != UAA_DEV_READY) return; iface = usbd_get_iface(udev, 0); if (iface == NULL) return; id = iface->idesc; if (id == NULL || id->bInterfaceClass != UICLASS_MASS) return; if (usb_test_quirk(uaa, UQ_MSC_EJECT_HUAWEI)) method = U3GINIT_HUAWEI; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_SIERRA)) method = U3GINIT_SIERRA; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_SCSIEJECT)) method = U3GINIT_SCSIEJECT; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_REZERO)) method = U3GINIT_REZERO; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_ZTESTOR)) method = U3GINIT_ZTESTOR; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_CMOTECH)) method = U3GINIT_CMOTECH; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_WAIT)) method = U3GINIT_WAIT; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_HUAWEISCSI)) method = U3GINIT_HUAWEISCSI; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_HUAWEISCSI2)) method = U3GINIT_HUAWEISCSI2; else if (usb_test_quirk(uaa, UQ_MSC_EJECT_TCT)) method = U3GINIT_TCT; else if (usbd_lookup_id_by_uaa(u3g_devs, sizeof(u3g_devs), uaa) == 0) method = USB_GET_DRIVER_INFO(uaa); else return; /* no device match */ if (bootverbose) { printf("Ejecting %s %s using method %ld\n", usb_get_manufacturer(udev), usb_get_product(udev), method); } switch (method) { case U3GINIT_HUAWEI: error = u3g_huawei_init(udev); break; case U3GINIT_HUAWEISCSI: error = usb_msc_eject(udev, 0, MSC_EJECT_HUAWEI); break; case U3GINIT_HUAWEISCSI2: error = usb_msc_eject(udev, 0, MSC_EJECT_HUAWEI2); break; case U3GINIT_SCSIEJECT: error = usb_msc_eject(udev, 0, MSC_EJECT_STOPUNIT); break; case U3GINIT_REZERO: error = usb_msc_eject(udev, 0, MSC_EJECT_REZERO); break; case U3GINIT_ZTESTOR: error = usb_msc_eject(udev, 0, MSC_EJECT_STOPUNIT); if (error == 0) error = usb_msc_eject(udev, 0, MSC_EJECT_ZTESTOR); break; case U3GINIT_CMOTECH: error = usb_msc_eject(udev, 0, MSC_EJECT_CMOTECH); break; case U3GINIT_TCT: error = usb_msc_eject(udev, 0, MSC_EJECT_TCT); break; case U3GINIT_SIERRA: error = u3g_sierra_init(udev); break; case U3GINIT_WAIT: /* Just pretend we ejected, the card will timeout */ error = 0; break; default: /* no 3G eject quirks */ error = EOPNOTSUPP; break; } if (error == 0) { /* success, mark the udev as disappearing */ uaa->dev_state = UAA_DEV_EJECTING; } } static int u3g_driver_loaded(struct module *mod, int what, void *arg) { switch (what) { case MOD_LOAD: /* register our autoinstall handler */ u3g_etag = EVENTHANDLER_REGISTER(usb_dev_configured, u3g_test_autoinst, NULL, EVENTHANDLER_PRI_ANY); break; case MOD_UNLOAD: EVENTHANDLER_DEREGISTER(usb_dev_configured, u3g_etag); break; default: return (EOPNOTSUPP); } return (0); } static int u3g_probe(device_t self) { struct usb_attach_arg *uaa = device_get_ivars(self); if (uaa->usb_mode != USB_MODE_HOST) { return (ENXIO); } if (uaa->info.bConfigIndex != U3G_CONFIG_INDEX) { return (ENXIO); } if (uaa->info.bInterfaceClass != UICLASS_VENDOR) { return (ENXIO); } return (usbd_lookup_id_by_uaa(u3g_devs, sizeof(u3g_devs), uaa)); } static int u3g_attach(device_t dev) { struct usb_config u3g_config_tmp[U3G_N_TRANSFER]; struct usb_attach_arg *uaa = device_get_ivars(dev); struct u3g_softc *sc = device_get_softc(dev); struct usb_interface *iface; struct usb_interface_descriptor *id; uint32_t iface_valid; int error, type, nports; int ep, n; uint8_t i; DPRINTF("sc=%p\n", sc); type = USB_GET_DRIVER_INFO(uaa); if (type == U3GINIT_SAEL_M460 || usb_test_quirk(uaa, UQ_MSC_EJECT_SAEL_M460)) { u3g_sael_m460_init(uaa->device); } /* copy in USB config */ for (n = 0; n != U3G_N_TRANSFER; n++) u3g_config_tmp[n] = u3g_config[n]; device_set_usb_desc(dev); mtx_init(&sc->sc_mtx, "u3g", NULL, MTX_DEF); ucom_ref(&sc->sc_super_ucom); sc->sc_udev = uaa->device; /* Claim all interfaces on the device */ iface_valid = 0; for (i = uaa->info.bIfaceIndex; i < USB_IFACE_MAX; i++) { iface = usbd_get_iface(uaa->device, i); if (iface == NULL) break; id = usbd_get_interface_descriptor(iface); if (id == NULL || id->bInterfaceClass != UICLASS_VENDOR) continue; usbd_set_parent_iface(uaa->device, i, uaa->info.bIfaceIndex); iface_valid |= (1<device, &i, sc->sc_xfer[nports], u3g_config_tmp, U3G_N_TRANSFER, &sc->sc_ucom[nports], &sc->sc_mtx); if (error) { /* next interface */ i++; ep = 0; continue; } iface = usbd_get_iface(uaa->device, i); id = usbd_get_interface_descriptor(iface); sc->sc_iface[nports] = id->bInterfaceNumber; if (bootverbose && sc->sc_xfer[nports][U3G_INTR]) { device_printf(dev, "port %d supports modem control\n", nports); } /* set stall by default */ mtx_lock(&sc->sc_mtx); usbd_xfer_set_stall(sc->sc_xfer[nports][U3G_BULK_WR]); usbd_xfer_set_stall(sc->sc_xfer[nports][U3G_BULK_RD]); mtx_unlock(&sc->sc_mtx); nports++; /* found one port */ ep++; if (nports == U3G_MAXPORTS) break; } if (nports == 0) { device_printf(dev, "no ports found\n"); goto detach; } sc->sc_numports = nports; error = ucom_attach(&sc->sc_super_ucom, sc->sc_ucom, sc->sc_numports, sc, &u3g_callback, &sc->sc_mtx); if (error) { DPRINTF("ucom_attach failed\n"); goto detach; } ucom_set_pnpinfo_usb(&sc->sc_super_ucom, dev); device_printf(dev, "Found %u port%s.\n", sc->sc_numports, sc->sc_numports > 1 ? "s":""); return (0); detach: u3g_detach(dev); return (ENXIO); } static int u3g_detach(device_t dev) { struct u3g_softc *sc = device_get_softc(dev); uint8_t subunit; DPRINTF("sc=%p\n", sc); /* NOTE: It is not dangerous to detach more ports than attached! */ ucom_detach(&sc->sc_super_ucom, sc->sc_ucom); for (subunit = 0; subunit != U3G_MAXPORTS; subunit++) usbd_transfer_unsetup(sc->sc_xfer[subunit], U3G_N_TRANSFER); device_claim_softc(dev); u3g_free_softc(sc); return (0); } UCOM_UNLOAD_DRAIN(u3g); static void u3g_free_softc(struct u3g_softc *sc) { if (ucom_unref(&sc->sc_super_ucom)) { mtx_destroy(&sc->sc_mtx); device_free_softc(sc); } } static void u3g_free(struct ucom_softc *ucom) { u3g_free_softc(ucom->sc_parent); } static void u3g_start_read(struct ucom_softc *ucom) { struct u3g_softc *sc = ucom->sc_parent; /* start interrupt endpoint (if configured) */ usbd_transfer_start(sc->sc_xfer[ucom->sc_subunit][U3G_INTR]); /* start read endpoint */ usbd_transfer_start(sc->sc_xfer[ucom->sc_subunit][U3G_BULK_RD]); } static void u3g_stop_read(struct ucom_softc *ucom) { struct u3g_softc *sc = ucom->sc_parent; /* stop interrupt endpoint (if configured) */ usbd_transfer_stop(sc->sc_xfer[ucom->sc_subunit][U3G_INTR]); /* stop read endpoint */ usbd_transfer_stop(sc->sc_xfer[ucom->sc_subunit][U3G_BULK_RD]); } static void u3g_start_write(struct ucom_softc *ucom) { struct u3g_softc *sc = ucom->sc_parent; usbd_transfer_start(sc->sc_xfer[ucom->sc_subunit][U3G_BULK_WR]); } static void u3g_stop_write(struct ucom_softc *ucom) { struct u3g_softc *sc = ucom->sc_parent; usbd_transfer_stop(sc->sc_xfer[ucom->sc_subunit][U3G_BULK_WR]); } static void u3g_write_callback(struct usb_xfer *xfer, usb_error_t error) { struct ucom_softc *ucom = usbd_xfer_softc(xfer); struct usb_page_cache *pc; uint32_t actlen; uint32_t frame; switch (USB_GET_STATE(xfer)) { case USB_ST_TRANSFERRED: case USB_ST_SETUP: tr_setup: for (frame = 0; frame != U3G_TXFRAMES; frame++) { usbd_xfer_set_frame_offset(xfer, frame * U3G_TXSIZE, frame); pc = usbd_xfer_get_frame(xfer, frame); if (ucom_get_data(ucom, pc, 0, U3G_TXSIZE, &actlen) == 0) break; usbd_xfer_set_frame_len(xfer, frame, actlen); } if (frame != 0) { usbd_xfer_set_frames(xfer, frame); usbd_transfer_submit(xfer); } break; default: /* Error */ if (error != USB_ERR_CANCELLED) { /* do a builtin clear-stall */ usbd_xfer_set_stall(xfer); goto tr_setup; } break; } } static void u3g_read_callback(struct usb_xfer *xfer, usb_error_t error) { struct ucom_softc *ucom = usbd_xfer_softc(xfer); struct usb_page_cache *pc; int actlen; usbd_xfer_status(xfer, &actlen, NULL, NULL, NULL); switch (USB_GET_STATE(xfer)) { case USB_ST_TRANSFERRED: pc = usbd_xfer_get_frame(xfer, 0); ucom_put_data(ucom, pc, 0, actlen); case USB_ST_SETUP: tr_setup: usbd_xfer_set_frame_len(xfer, 0, usbd_xfer_max_len(xfer)); usbd_transfer_submit(xfer); break; default: /* Error */ if (error != USB_ERR_CANCELLED) { /* do a builtin clear-stall */ usbd_xfer_set_stall(xfer); goto tr_setup; } break; } } static void u3g_cfg_get_status(struct ucom_softc *ucom, uint8_t *lsr, uint8_t *msr) { struct u3g_softc *sc = ucom->sc_parent; *lsr = sc->sc_lsr[ucom->sc_subunit]; *msr = sc->sc_msr[ucom->sc_subunit]; } static void u3g_cfg_set_line(struct ucom_softc *ucom) { struct u3g_softc *sc = ucom->sc_parent; struct usb_device_request req; req.bmRequestType = UT_WRITE_CLASS_INTERFACE; req.bRequest = UCDC_SET_CONTROL_LINE_STATE; USETW(req.wValue, sc->sc_line[ucom->sc_subunit]); req.wIndex[0] = sc->sc_iface[ucom->sc_subunit]; req.wIndex[1] = 0; USETW(req.wLength, 0); ucom_cfg_do_request(sc->sc_udev, ucom, &req, NULL, 0, 1000); } static void u3g_cfg_set_dtr(struct ucom_softc *ucom, uint8_t onoff) { struct u3g_softc *sc = ucom->sc_parent; DPRINTF("onoff = %d\n", onoff); if (onoff) sc->sc_line[ucom->sc_subunit] |= UCDC_LINE_DTR; else sc->sc_line[ucom->sc_subunit] &= ~UCDC_LINE_DTR; u3g_cfg_set_line(ucom); } static void u3g_cfg_set_rts(struct ucom_softc *ucom, uint8_t onoff) { struct u3g_softc *sc = ucom->sc_parent; DPRINTF("onoff = %d\n", onoff); if (onoff) sc->sc_line[ucom->sc_subunit] |= UCDC_LINE_RTS; else sc->sc_line[ucom->sc_subunit] &= ~UCDC_LINE_RTS; u3g_cfg_set_line(ucom); } static void u3g_intr_callback(struct usb_xfer *xfer, usb_error_t error) { struct ucom_softc *ucom = usbd_xfer_softc(xfer); struct u3g_softc *sc = ucom->sc_parent; struct usb_page_cache *pc; struct usb_cdc_notification pkt; int actlen; uint16_t wLen; uint8_t mstatus; usbd_xfer_status(xfer, &actlen, NULL, NULL, NULL); switch (USB_GET_STATE(xfer)) { case USB_ST_TRANSFERRED: if (actlen < 8) { /* usb_cdc_notification with 2 data bytes */ DPRINTF("message too short (expected 8, received %d)\n", actlen); goto tr_setup; } pc = usbd_xfer_get_frame(xfer, 0); usbd_copy_out(pc, 0, &pkt, actlen); wLen = UGETW(pkt.wLength); if (wLen < 2) { DPRINTF("message too short (expected 2 data bytes, received %d)\n", wLen); goto tr_setup; } if (pkt.bmRequestType == UCDC_NOTIFICATION && pkt.bNotification == UCDC_N_SERIAL_STATE) { /* * Set the serial state in ucom driver based on * the bits from the notify message */ DPRINTF("notify bytes = 0x%02x, 0x%02x\n", pkt.data[0], pkt.data[1]); /* currently, lsr is always zero. */ sc->sc_lsr[ucom->sc_subunit] = 0; sc->sc_msr[ucom->sc_subunit] = 0; mstatus = pkt.data[0]; if (mstatus & UCDC_N_SERIAL_RI) sc->sc_msr[ucom->sc_subunit] |= SER_RI; if (mstatus & UCDC_N_SERIAL_DSR) sc->sc_msr[ucom->sc_subunit] |= SER_DSR; if (mstatus & UCDC_N_SERIAL_DCD) sc->sc_msr[ucom->sc_subunit] |= SER_DCD; ucom_status_change(ucom); } case USB_ST_SETUP: tr_setup: usbd_xfer_set_frame_len(xfer, 0, usbd_xfer_max_len(xfer)); usbd_transfer_submit(xfer); return; default: /* Error */ if (error != USB_ERR_CANCELLED) { /* try to clear stall first */ usbd_xfer_set_stall(xfer); goto tr_setup; } return; } } static void u3g_poll(struct ucom_softc *ucom) { struct u3g_softc *sc = ucom->sc_parent; usbd_transfer_poll(sc->sc_xfer[ucom->sc_subunit], U3G_N_TRANSFER); } Index: projects/building-blocks/sys/dev/usb/usbdevs =================================================================== --- projects/building-blocks/sys/dev/usb/usbdevs (revision 275663) +++ projects/building-blocks/sys/dev/usb/usbdevs (revision 275664) @@ -1,4591 +1,4615 @@ $FreeBSD$ /* $NetBSD: usbdevs,v 1.392 2004/12/29 08:38:44 imp Exp $ */ /*- * Copyright (c) 1998-2004 The NetBSD Foundation, Inc. * All rights reserved. * * This code is derived from software contributed to The NetBSD Foundation * by Lennart Augustsson (lennart@augustsson.net) at * Carlstedt Research & Technology. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. */ /* * List of known USB vendors * * USB.org publishes a VID list of USB-IF member companies at * http://www.usb.org/developers/tools * Note that it does not show companies that have obtained a Vendor ID * without becoming full members. * * Please note that these IDs do not do anything. Adding an ID here and * regenerating the usbdevs.h and usbdevs_data.h only makes a symbolic name * available to the source code and does not change any functionality, nor * does it make your device available to a specific driver. * It will however make the descriptive string available if a device does not * provide the string itself. * * After adding a vendor ID VNDR and a product ID PRDCT you will have the * following extra defines: * #define USB_VENDOR_VNDR 0x???? * #define USB_PRODUCT_VNDR_PRDCT 0x???? * * You may have to add these defines to the respective probe routines to * make the device recognised by the appropriate device driver. */ vendor UNKNOWN1 0x0053 Unknown vendor vendor UNKNOWN2 0x0105 Unknown vendor vendor EGALAX2 0x0123 eGalax, Inc. vendor CHIPSBANK 0x0204 Chipsbank Microelectronics Co. vendor HUMAX 0x02ad HUMAX vendor LTS 0x0386 LTS vendor BWCT 0x03da Bernd Walter Computer Technology vendor AOX 0x03e8 AOX vendor THESYS 0x03e9 Thesys vendor DATABROADCAST 0x03ea Data Broadcasting vendor ATMEL 0x03eb Atmel vendor IWATSU 0x03ec Iwatsu America vendor MITSUMI 0x03ee Mitsumi vendor HP 0x03f0 Hewlett Packard vendor GENOA 0x03f1 Genoa vendor OAK 0x03f2 Oak vendor ADAPTEC 0x03f3 Adaptec vendor DIEBOLD 0x03f4 Diebold vendor SIEMENSELECTRO 0x03f5 Siemens Electromechanical vendor EPSONIMAGING 0x03f8 Epson Imaging vendor KEYTRONIC 0x03f9 KeyTronic vendor OPTI 0x03fb OPTi vendor ELITEGROUP 0x03fc Elitegroup vendor XILINX 0x03fd Xilinx vendor FARALLON 0x03fe Farallon Communications vendor NATIONAL 0x0400 National Semiconductor vendor NATIONALREG 0x0401 National Registry vendor ACERLABS 0x0402 Acer Labs vendor FTDI 0x0403 Future Technology Devices vendor NCR 0x0404 NCR vendor SYNOPSYS2 0x0405 Synopsys vendor FUJITSUICL 0x0406 Fujitsu-ICL vendor FUJITSU2 0x0407 Fujitsu Personal Systems vendor QUANTA 0x0408 Quanta vendor NEC 0x0409 NEC vendor KODAK 0x040a Eastman Kodak vendor WELTREND 0x040b Weltrend vendor VIA 0x040d VIA vendor MCCI 0x040e MCCI vendor MELCO 0x0411 Melco vendor LEADTEK 0x0413 Leadtek vendor WINBOND 0x0416 Winbond vendor PHOENIX 0x041a Phoenix vendor CREATIVE 0x041e Creative Labs vendor NOKIA 0x0421 Nokia vendor ADI 0x0422 ADI Systems vendor CATC 0x0423 Computer Access Technology vendor SMC2 0x0424 Standard Microsystems vendor MOTOROLA_HK 0x0425 Motorola HK vendor GRAVIS 0x0428 Advanced Gravis Computer vendor CIRRUSLOGIC 0x0429 Cirrus Logic vendor INNOVATIVE 0x042c Innovative Semiconductors vendor MOLEX 0x042f Molex vendor SUN 0x0430 Sun Microsystems vendor UNISYS 0x0432 Unisys vendor TAUGA 0x0436 Taugagreining HF vendor AMD 0x0438 Advanced Micro Devices vendor LEXMARK 0x043d Lexmark International vendor LG 0x043e LG Electronics vendor NANAO 0x0440 NANAO vendor GATEWAY 0x0443 Gateway 2000 vendor NMB 0x0446 NMB vendor ALPS 0x044e Alps Electric vendor THRUST 0x044f Thrustmaster vendor TI 0x0451 Texas Instruments vendor ANALOGDEVICES 0x0456 Analog Devices vendor SIS 0x0457 Silicon Integrated Systems Corp. vendor KYE 0x0458 KYE Systems vendor DIAMOND2 0x045a Diamond (Supra) vendor RENESAS 0x045b Renesas vendor MICROSOFT 0x045e Microsoft vendor PRIMAX 0x0461 Primax Electronics vendor MGE 0x0463 MGE UPS Systems vendor AMP 0x0464 AMP vendor CHERRY 0x046a Cherry Mikroschalter vendor MEGATRENDS 0x046b American Megatrends vendor LOGITECH 0x046d Logitech vendor BTC 0x046e Behavior Tech. Computer vendor PHILIPS 0x0471 Philips vendor SUN2 0x0472 Sun Microsystems (offical) vendor SANYO 0x0474 Sanyo Electric vendor SEAGATE 0x0477 Seagate vendor CONNECTIX 0x0478 Connectix vendor SEMTECH 0x047a Semtech vendor KENSINGTON 0x047d Kensington vendor LUCENT 0x047e Lucent vendor PLANTRONICS 0x047f Plantronics vendor KYOCERA 0x0482 Kyocera Wireless Corp. vendor STMICRO 0x0483 STMicroelectronics vendor FOXCONN 0x0489 Foxconn vendor MEIZU 0x0492 Meizu Electronics vendor YAMAHA 0x0499 YAMAHA vendor COMPAQ 0x049f Compaq vendor HITACHI 0x04a4 Hitachi vendor ACERP 0x04a5 Acer Peripherals vendor DAVICOM 0x04a6 Davicom vendor VISIONEER 0x04a7 Visioneer vendor CANON 0x04a9 Canon vendor NIKON 0x04b0 Nikon vendor PAN 0x04b1 Pan International vendor IBM 0x04b3 IBM vendor CYPRESS 0x04b4 Cypress Semiconductor vendor ROHM 0x04b5 ROHM vendor COMPAL 0x04b7 Compal vendor EPSON 0x04b8 Seiko Epson vendor RAINBOW 0x04b9 Rainbow Technologies vendor IODATA 0x04bb I-O Data vendor TDK 0x04bf TDK vendor 3COMUSR 0x04c1 U.S. Robotics vendor METHODE 0x04c2 Methode Electronics Far East vendor MAXISWITCH 0x04c3 Maxi Switch vendor LOCKHEEDMER 0x04c4 Lockheed Martin Energy Research vendor FUJITSU 0x04c5 Fujitsu vendor TOSHIBAAM 0x04c6 Toshiba America vendor MICROMACRO 0x04c7 Micro Macro Technologies vendor KONICA 0x04c8 Konica vendor LITEON 0x04ca Lite-On Technology vendor FUJIPHOTO 0x04cb Fuji Photo Film vendor PHILIPSSEMI 0x04cc Philips Semiconductors vendor TATUNG 0x04cd Tatung Co. Of America vendor SCANLOGIC 0x04ce ScanLogic vendor MYSON 0x04cf Myson Technology vendor DIGI2 0x04d0 Digi vendor ITTCANON 0x04d1 ITT Canon vendor ALTEC 0x04d2 Altec Lansing vendor LSI 0x04d4 LSI vendor MENTORGRAPHICS 0x04d6 Mentor Graphics vendor ITUNERNET 0x04d8 I-Tuner Networks vendor HOLTEK 0x04d9 Holtek Semiconductor, Inc. vendor PANASONIC 0x04da Panasonic (Matsushita) vendor HUANHSIN 0x04dc Huan Hsin vendor SHARP 0x04dd Sharp vendor IIYAMA 0x04e1 Iiyama vendor SHUTTLE 0x04e6 Shuttle Technology vendor ELO 0x04e7 Elo TouchSystems vendor SAMSUNG 0x04e8 Samsung Electronics vendor NORTHSTAR 0x04eb Northstar vendor TOKYOELECTRON 0x04ec Tokyo Electron vendor ANNABOOKS 0x04ed Annabooks vendor JVC 0x04f1 JVC vendor CHICONY 0x04f2 Chicony Electronics vendor ELAN 0x04f3 Elan vendor NEWNEX 0x04f7 Newnex vendor BROTHER 0x04f9 Brother Industries vendor DALLAS 0x04fa Dallas Semiconductor vendor AIPTEK2 0x04fc AIPTEK International vendor PFU 0x04fe PFU vendor FUJIKURA 0x0501 Fujikura/DDK vendor ACER 0x0502 Acer vendor 3COM 0x0506 3Com vendor HOSIDEN 0x0507 Hosiden Corporation vendor AZTECH 0x0509 Aztech Systems vendor BELKIN 0x050d Belkin Components vendor KAWATSU 0x050f Kawatsu Semiconductor vendor FCI 0x0514 FCI vendor LONGWELL 0x0516 Longwell vendor COMPOSITE 0x0518 Composite vendor STAR 0x0519 Star Micronics vendor APC 0x051d American Power Conversion vendor SCIATLANTA 0x051e Scientific Atlanta vendor TSM 0x0520 TSM vendor CONNECTEK 0x0522 Advanced Connectek USA vendor NETCHIP 0x0525 NetChip Technology vendor ALTRA 0x0527 ALTRA vendor ATI 0x0528 ATI Technologies vendor AKS 0x0529 Aladdin Knowledge Systems vendor TEKOM 0x052b Tekom vendor CANONDEV 0x052c Canon vendor WACOMTECH 0x0531 Wacom vendor INVENTEC 0x0537 Inventec vendor SHYHSHIUN 0x0539 Shyh Shiun Terminals vendor PREHWERKE 0x053a Preh Werke Gmbh & Co. KG vendor SYNOPSYS 0x053f Synopsys vendor UNIACCESS 0x0540 Universal Access vendor VIEWSONIC 0x0543 ViewSonic vendor XIRLINK 0x0545 Xirlink vendor ANCHOR 0x0547 Anchor Chips vendor SONY 0x054c Sony vendor FUJIXEROX 0x0550 Fuji Xerox vendor VISION 0x0553 VLSI Vision vendor ASAHIKASEI 0x0556 Asahi Kasei Microsystems vendor ATEN 0x0557 ATEN International vendor SAMSUNG2 0x055d Samsung Electronics vendor MUSTEK 0x055f Mustek Systems vendor TELEX 0x0562 Telex Communications vendor CHINON 0x0564 Chinon vendor PERACOM 0x0565 Peracom Networks vendor ALCOR2 0x0566 Alcor Micro vendor XYRATEX 0x0567 Xyratex vendor WACOM 0x056a WACOM vendor ETEK 0x056c e-TEK Labs vendor EIZO 0x056d EIZO vendor ELECOM 0x056e Elecom vendor CONEXANT 0x0572 Conexant vendor HAUPPAUGE 0x0573 Hauppauge Computer Works vendor BAFO 0x0576 BAFO/Quality Computer Accessories vendor YEDATA 0x057b Y-E Data vendor AVM 0x057c AVM vendor QUICKSHOT 0x057f Quickshot vendor ROLAND 0x0582 Roland vendor ROCKFIRE 0x0583 Rockfire vendor RATOC 0x0584 RATOC Systems vendor ZYXEL 0x0586 ZyXEL Communication vendor INFINEON 0x058b Infineon vendor MICREL 0x058d Micrel vendor ALCOR 0x058f Alcor Micro vendor OMRON 0x0590 OMRON vendor ZORAN 0x0595 Zoran Microelectronics vendor NIIGATA 0x0598 Niigata vendor IOMEGA 0x059b Iomega vendor ATREND 0x059c A-Trend Technology vendor AID 0x059d Advanced Input Devices vendor LACIE 0x059f LaCie vendor FUJIFILM 0x05a2 Fuji Film vendor ARC 0x05a3 ARC vendor ORTEK 0x05a4 Ortek vendor CISCOLINKSYS3 0x05a6 Cisco-Linksys vendor BOSE 0x05a7 Bose vendor OMNIVISION 0x05a9 OmniVision vendor INSYSTEM 0x05ab In-System Design vendor APPLE 0x05ac Apple Computer vendor YCCABLE 0x05ad Y.C. Cable vendor DIGITALPERSONA 0x05ba DigitalPersona vendor 3G 0x05bc 3G Green Green Globe vendor RAFI 0x05bd RAFI vendor TYCO 0x05be Tyco vendor KAWASAKI 0x05c1 Kawasaki vendor DIGI 0x05c5 Digi International vendor QUALCOMM2 0x05c6 Qualcomm vendor QTRONIX 0x05c7 Qtronix vendor FOXLINK 0x05c8 Foxlink vendor RICOH 0x05ca Ricoh vendor ELSA 0x05cc ELSA vendor SCIWORX 0x05ce sci-worx vendor BRAINBOXES 0x05d1 Brainboxes Limited vendor ULTIMA 0x05d8 Ultima vendor AXIOHM 0x05d9 Axiohm Transaction Solutions vendor MICROTEK 0x05da Microtek vendor SUNTAC 0x05db SUN Corporation vendor LEXAR 0x05dc Lexar Media vendor ADDTRON 0x05dd Addtron vendor SYMBOL 0x05e0 Symbol Technologies vendor SYNTEK 0x05e1 Syntek vendor GENESYS 0x05e3 Genesys Logic vendor FUJI 0x05e5 Fuji Electric vendor KEITHLEY 0x05e6 Keithley Instruments vendor EIZONANAO 0x05e7 EIZO Nanao vendor KLSI 0x05e9 Kawasaki LSI vendor FFC 0x05eb FFC vendor ANKO 0x05ef Anko Electronic vendor PIENGINEERING 0x05f3 P.I. Engineering vendor AOC 0x05f6 AOC International vendor CHIC 0x05fe Chic Technology vendor BARCO 0x0600 Barco Display Systems vendor BRIDGE 0x0607 Bridge Information vendor SOLIDYEAR 0x060b Solid Year vendor BIORAD 0x0614 Bio-Rad Laboratories vendor MACALLY 0x0618 Macally vendor ACTLABS 0x061c Act Labs vendor ALARIS 0x0620 Alaris vendor APEX 0x0624 Apex vendor CREATIVE3 0x062a Creative Labs vendor MICRON 0x0634 Micron Technology vendor VIVITAR 0x0636 Vivitar vendor GUNZE 0x0637 Gunze Electronics USA vendor AVISION 0x0638 Avision vendor TEAC 0x0644 TEAC vendor ACTON 0x0647 Acton Research Corp. vendor OPTO 0x065a Optoelectronics Co., Ltd vendor SGI 0x065e Silicon Graphics vendor SANWASUPPLY 0x0663 Sanwa Supply vendor MEGATEC 0x0665 Megatec vendor LINKSYS 0x066b Linksys vendor ACERSA 0x066e Acer Semiconductor America vendor SIGMATEL 0x066f Sigmatel vendor DRAYTEK 0x0675 DrayTek vendor AIWA 0x0677 Aiwa vendor ACARD 0x0678 ACARD Technology vendor PROLIFIC 0x067b Prolific Technology vendor SIEMENS 0x067c Siemens vendor AVANCELOGIC 0x0680 Avance Logic vendor SIEMENS2 0x0681 Siemens vendor MINOLTA 0x0686 Minolta vendor CHPRODUCTS 0x068e CH Products vendor HAGIWARA 0x0693 Hagiwara Sys-Com vendor CTX 0x0698 Chuntex vendor ASKEY 0x069a Askey Computer vendor SAITEK 0x06a3 Saitek vendor ALCATELT 0x06b9 Alcatel Telecom vendor AGFA 0x06bd AGFA-Gevaert vendor ASIAMD 0x06be Asia Microelectronic Development vendor BIZLINK 0x06c4 Bizlink International vendor KEYSPAN 0x06cd Keyspan / InnoSys Inc. vendor CONTEC 0x06ce Contec products vendor AASHIMA 0x06d6 Aashima Technology vendor LIEBERT 0x06da Liebert vendor MULTITECH 0x06e0 MultiTech vendor ADS 0x06e1 ADS Technologies vendor ALCATELM 0x06e4 Alcatel Microelectronics vendor SIRIUS 0x06ea Sirius Technologies vendor GUILLEMOT 0x06f8 Guillemot vendor BOSTON 0x06fd Boston Acoustics vendor SMC 0x0707 Standard Microsystems vendor PUTERCOM 0x0708 Putercom vendor MCT 0x0711 MCT vendor IMATION 0x0718 Imation vendor TECLAST 0x071b Teclast vendor SONYERICSSON 0x0731 Sony Ericsson vendor EICON 0x0734 Eicon Networks vendor SYNTECH 0x0745 Syntech Information vendor DIGITALSTREAM 0x074e Digital Stream vendor AUREAL 0x0755 Aureal Semiconductor vendor MAUDIO 0x0763 M-Audio vendor CYBERPOWER 0x0764 Cyber Power Systems, Inc. vendor SURECOM 0x0769 Surecom Technology vendor HIDGLOBAL 0x076b HID Global vendor LINKSYS2 0x077b Linksys vendor GRIFFIN 0x077d Griffin Technology vendor SANDISK 0x0781 SanDisk vendor JENOPTIK 0x0784 Jenoptik vendor LOGITEC 0x0789 Logitec vendor NOKIA2 0x078b Nokia vendor BRIMAX 0x078e Brimax vendor AXIS 0x0792 Axis Communications vendor ABL 0x0794 ABL Electronics vendor SAGEM 0x079b Sagem vendor SUNCOMM 0x079c Sun Communications, Inc. vendor ALFADATA 0x079d Alfadata Computer vendor NATIONALTECH 0x07a2 National Technical Systems vendor ONNTO 0x07a3 Onnto vendor BE 0x07a4 Be vendor ADMTEK 0x07a6 ADMtek vendor COREGA 0x07aa Corega vendor FREECOM 0x07ab Freecom vendor MICROTECH 0x07af Microtech vendor GENERALINSTMNTS 0x07b2 General Instruments (Motorola) vendor OLYMPUS 0x07b4 Olympus vendor ABOCOM 0x07b8 AboCom Systems vendor KEISOKUGIKEN 0x07c1 Keisokugiken vendor ONSPEC 0x07c4 OnSpec vendor APG 0x07c5 APG Cash Drawer vendor BUG 0x07c8 B.U.G. vendor ALLIEDTELESYN 0x07c9 Allied Telesyn International vendor AVERMEDIA 0x07ca AVerMedia Technologies vendor SIIG 0x07cc SIIG vendor CASIO 0x07cf CASIO vendor DLINK2 0x07d1 D-Link vendor APTIO 0x07d2 Aptio Products vendor ARASAN 0x07da Arasan Chip Systems vendor ALLIEDCABLE 0x07e6 Allied Cable vendor STSN 0x07ef STSN vendor CENTURY 0x07f7 Century Corp vendor NEWLINK 0x07ff NEWlink vendor MAGTEK 0x0801 Mag-Tek vendor ZOOM 0x0803 Zoom Telephonics vendor PCS 0x0810 Personal Communication Systems vendor ALPHASMART 0x081e AlphaSmart, Inc. vendor BROADLOGIC 0x0827 BroadLogic vendor HANDSPRING 0x082d Handspring vendor PALM 0x0830 Palm Computing vendor SOURCENEXT 0x0833 SOURCENEXT vendor ACTIONSTAR 0x0835 Action Star Enterprise vendor SAMSUNG_TECHWIN 0x0839 Samsung Techwin vendor ACCTON 0x083a Accton Technology vendor DIAMOND 0x0841 Diamond vendor NETGEAR 0x0846 BayNETGEAR vendor TOPRE 0x0853 Topre Corporation vendor ACTIVEWIRE 0x0854 ActiveWire vendor BBELECTRONICS 0x0856 B&B Electronics vendor PORTGEAR 0x085a PortGear vendor NETGEAR2 0x0864 Netgear vendor SYSTEMTALKS 0x086e System Talks vendor METRICOM 0x0870 Metricom vendor ADESSOKBTEK 0x087c ADESSO/Kbtek America vendor JATON 0x087d Jaton vendor APT 0x0880 APT Technologies vendor BOCARESEARCH 0x0885 Boca Research vendor ANDREA 0x08a8 Andrea Electronics vendor BURRBROWN 0x08bb Burr-Brown Japan vendor 2WIRE 0x08c8 2Wire vendor AIPTEK 0x08ca AIPTEK International vendor SMARTBRIDGES 0x08d1 SmartBridges vendor FUJITSUSIEMENS 0x08d4 Fujitsu-Siemens vendor BILLIONTON 0x08dd Billionton Systems vendor GEMALTO 0x08e6 Gemalto SA vendor EXTENDED 0x08e9 Extended Systems vendor MSYSTEMS 0x08ec M-Systems vendor DIGIANSWER 0x08fd Digianswer vendor AUTHENTEC 0x08ff AuthenTec vendor AUDIOTECHNICA 0x0909 Audio-Technica vendor TRUMPION 0x090a Trumpion Microelectronics vendor FEIYA 0x090c Feiya vendor ALATION 0x0910 Alation Systems vendor GLOBESPAN 0x0915 Globespan vendor CONCORDCAMERA 0x0919 Concord Camera vendor GARMIN 0x091e Garmin International vendor GOHUBS 0x0921 GoHubs vendor XEROX 0x0924 Xerox vendor BIOMETRIC 0x0929 American Biometric Company vendor TOSHIBA 0x0930 Toshiba vendor PLEXTOR 0x093b Plextor vendor INTREPIDCS 0x093c Intrepid vendor YANO 0x094f Yano vendor KINGSTON 0x0951 Kingston Technology vendor BLUEWATER 0x0956 BlueWater Systems vendor AGILENT 0x0957 Agilent Technologies vendor GUDE 0x0959 Gude ADS vendor PORTSMITH 0x095a Portsmith vendor ACERW 0x0967 Acer vendor ADIRONDACK 0x0976 Adirondack Wire & Cable vendor BECKHOFF 0x0978 Beckhoff vendor MINDSATWORK 0x097a Minds At Work vendor POINTCHIPS 0x09a6 PointChips vendor INTERSIL 0x09aa Intersil vendor ALTIUS 0x09b3 Altius Solutions vendor ARRIS 0x09c1 Arris Interactive vendor ACTIVCARD 0x09c3 ACTIVCARD vendor ACTISYS 0x09c4 ACTiSYS vendor NOVATEL2 0x09d7 Novatel Wireless vendor AFOURTECH 0x09da A-FOUR TECH vendor AIMEX 0x09dc AIMEX vendor ADDONICS 0x09df Addonics Technologies vendor AKAI 0x09e8 AKAI professional M.I. vendor ARESCOM 0x09f5 ARESCOM vendor BAY 0x09f9 Bay Associates vendor ALTERA 0x09fb Altera vendor CSR 0x0a12 Cambridge Silicon Radio vendor TREK 0x0a16 Trek Technology vendor ASAHIOPTICAL 0x0a17 Asahi Optical vendor BOCASYSTEMS 0x0a43 Boca Systems vendor SHANTOU 0x0a46 ShanTou vendor MEDIAGEAR 0x0a48 MediaGear vendor BROADCOM 0x0a5c Broadcom vendor GREENHOUSE 0x0a6b GREENHOUSE vendor MEDELI 0x0a67 Medeli vendor GEOCAST 0x0a79 Geocast Network Systems vendor EGO 0x0a92 EGO systems vendor IDQUANTIQUE 0x0aba ID Quantique vendor IDTECH 0x0acd ID TECH vendor ZYDAS 0x0ace Zydas Technology Corporation vendor NEODIO 0x0aec Neodio vendor OPTION 0x0af0 Option N.V. vendor ASUS 0x0b05 ASUSTeK Computer vendor TODOS 0x0b0c Todos Data System vendor SIIG2 0x0b39 SIIG vendor TEKRAM 0x0b3b Tekram Technology vendor HAL 0x0b41 HAL Corporation vendor EMS 0x0b43 EMS Production vendor NEC2 0x0b62 NEC vendor ADLINK 0x0b63 ADLINK Technoligy, Inc. vendor ATI2 0x0b6f ATI vendor ZEEVO 0x0b7a Zeevo, Inc. vendor KURUSUGAWA 0x0b7e Kurusugawa Electronics, Inc. vendor SMART 0x0b8c Smart Technologies vendor ASIX 0x0b95 ASIX Electronics vendor O2MICRO 0x0b97 O2 Micro, Inc. vendor USR 0x0baf U.S. Robotics vendor AMBIT 0x0bb2 Ambit Microsystems vendor HTC 0x0bb4 HTC vendor REALTEK 0x0bda Realtek vendor ERICSSON2 0x0bdb Ericsson vendor MEI 0x0bed MEI vendor ADDONICS2 0x0bf6 Addonics Technology vendor FSC 0x0bf8 Fujitsu Siemens Computers vendor AGATE 0x0c08 Agate Technologies vendor DMI 0x0c0b DMI vendor CANYON 0x0c10 Canyon vendor ICOM 0x0c26 Icom Inc. vendor GNOTOMETRICS 0x0c33 GN Otometrics vendor CHICONY2 0x0c45 Chicony vendor REINERSCT 0x0c4b Reiner-SCT vendor SEALEVEL 0x0c52 Sealevel System vendor JETI 0x0c6c Jeti vendor LUWEN 0x0c76 Luwen vendor ELEKTOR 0x0c7d ELEKTOR Electronics vendor KYOCERA2 0x0c88 Kyocera Wireless Corp. vendor ZCOM 0x0cde Z-Com vendor ATHEROS2 0x0cf3 Atheros Communications vendor POSIFLEX 0x0d3a POSIFLEX vendor TANGTOP 0x0d3d Tangtop vendor KOBIL 0x0d46 KOBIL vendor SMC3 0x0d5c Standard Microsystems vendor ADDON 0x0d7d Add-on Technology vendor ACDC 0x0d7e American Computer & Digital Components vendor CMEDIA 0x0d8c CMEDIA vendor CONCEPTRONIC 0x0d8e Conceptronic vendor SKANHEX 0x0d96 Skanhex Technology, Inc. vendor MSI 0x0db0 Micro Star International vendor ELCON 0x0db7 ELCON Systemtechnik vendor UNKNOWN4 0x0dcd Unknown vendor vendor NETAC 0x0dd8 Netac vendor SITECOMEU 0x0df6 Sitecom Europe vendor MOBILEACTION 0x0df7 Mobile Action vendor AMIGO 0x0e0b Amigo Technology vendor SPEEDDRAGON 0x0e55 Speed Dragon Multimedia vendor HAWKING 0x0e66 Hawking vendor FOSSIL 0x0e67 Fossil, Inc vendor GMATE 0x0e7e G.Mate, Inc vendor MEDIATEK 0x0e8d MediaTek, Inc. vendor OTI 0x0ea0 Ours Technology vendor YISO 0x0eab Yiso Wireless Co. vendor PILOTECH 0x0eaf Pilotech vendor NOVATECH 0x0eb0 NovaTech vendor ITEGNO 0x0eba iTegno vendor WINMAXGROUP 0x0ed1 WinMaxGroup vendor TOD 0x0ede TOD vendor EGALAX 0x0eef eGalax, Inc. vendor AIRPRIME 0x0f3d AirPrime, Inc. vendor MICROTUNE 0x0f4d Microtune vendor VTECH 0x0f88 VTech vendor FALCOM 0x0f94 Falcom Wireless Communications GmbH vendor RIM 0x0fca Research In Motion vendor DYNASTREAM 0x0fcf Dynastream Innovations vendor LARSENBRUSGAARD 0x0fd8 Larsen and Brusgaard vendor OWL 0x0fde OWL vendor KONTRON 0x0fe6 Kontron AG vendor QUALCOMM 0x1004 Qualcomm vendor APACER 0x1005 Apacer vendor MOTOROLA4 0x100d Motorola vendor HP3 0x103c Hewlett Packard vendor AIRPLUS 0x1011 Airplus vendor DESKNOTE 0x1019 Desknote vendor NEC3 0x1033 NEC vendor TTI 0x103e Thurlby Thandar Instruments vendor GIGABYTE 0x1044 GIGABYTE vendor WESTERN 0x1058 Western Digital vendor MOTOROLA 0x1063 Motorola vendor CCYU 0x1065 CCYU Technology vendor CURITEL 0x106c Curitel Communications Inc vendor SILABS2 0x10a6 SILABS2 vendor USI 0x10ab USI vendor LIEBERT2 0x10af Liebert vendor PLX 0x10b5 PLX vendor ASANTE 0x10bd Asante vendor SILABS 0x10c4 Silicon Labs vendor SILABS3 0x10c5 Silicon Labs vendor SILABS4 0x10ce Silicon Labs vendor ACTIONS 0x10d6 Actions vendor ANALOG 0x1110 Analog Devices vendor TENX 0x1130 Ten X Technology, Inc. vendor ISSC 0x1131 Integrated System Solution Corp. vendor JRC 0x1145 Japan Radio Company vendor SPHAIRON 0x114b Sphairon Access Systems GmbH vendor DELORME 0x1163 DeLorme vendor SERVERWORKS 0x1166 ServerWorks vendor DLINK3 0x1186 Dlink vendor ACERCM 0x1189 Acer Communications & Multimedia vendor SIERRA 0x1199 Sierra Wireless vendor SANWA 0x11ad Sanwa Electric Instrument Co., Ltd. vendor TOPFIELD 0x11db Topfield Co., Ltd vendor SIEMENS3 0x11f5 Siemens vendor NETINDEX 0x11f6 NetIndex vendor ALCATEL 0x11f7 Alcatel vendor INTERBIOMETRICS 0x1209 Interbiometrics vendor UNKNOWN3 0x1233 Unknown vendor vendor TSUNAMI 0x1241 Tsunami vendor PHEENET 0x124a Pheenet vendor TARGUS 0x1267 Targus vendor TWINMOS 0x126f TwinMOS vendor TENDA 0x1286 Tenda vendor TESTO 0x128d Testo products vendor CREATIVE2 0x1292 Creative Labs vendor BELKIN2 0x1293 Belkin Components vendor CYBERTAN 0x129b CyberTAN Technology vendor HUAWEI 0x12d1 Huawei Technologies vendor ARANEUS 0x12d8 Araneus Information Systems vendor TAPWAVE 0x12ef Tapwave vendor AINCOMM 0x12fd Aincomm vendor MOBILITY 0x1342 Mobility vendor DICKSMITH 0x1371 Dick Smith Electronics vendor NETGEAR3 0x1385 Netgear vendor BALTECH 0x13ad Baltech vendor CISCOLINKSYS 0x13b1 Cisco-Linksys vendor SHARK 0x13d2 Shark vendor AZUREWAVE 0x13d3 AsureWave vendor INITIO 0x13fd Initio Corporation vendor EMTEC 0x13fe Emtec vendor NOVATEL 0x1410 Novatel Wireless vendor MERLIN 0x1416 Merlin vendor REDOCTANE 0x1430 RedOctane vendor WISTRONNEWEB 0x1435 Wistron NeWeb vendor RADIOSHACK 0x1453 Radio Shack vendor FIC 0x1457 FIC / OpenMoko vendor HUAWEI3COM 0x1472 Huawei-3Com vendor ABOCOM2 0x1482 AboCom Systems vendor SILICOM 0x1485 Silicom vendor RALINK 0x148f Ralink Technology vendor IMAGINATION 0x149a Imagination Technologies vendor ATP 0x14af ATP Electronics vendor CONCEPTRONIC2 0x14b2 Conceptronic vendor SUPERTOP 0x14cd Super Top vendor PLANEX3 0x14ea Planex Communications vendor SILICONPORTALS 0x1527 Silicon Portals vendor UBIQUAM 0x1529 UBIQUAM Co., Ltd. vendor JMICRON 0x152d JMicron vendor UBLOX 0x1546 U-blox vendor PNY 0x154b PNY vendor OWEN 0x1555 Owen vendor OQO 0x1557 OQO vendor UMEDIA 0x157e U-MEDIA Communications vendor FIBERLINE 0x1582 Fiberline vendor FREESCALE 0x15a2 Freescale Semiconductor, Inc. vendor AFATECH 0x15a4 Afatech Technologies, Inc. vendor SPARKLAN 0x15a9 SparkLAN vendor OLIMEX 0x15ba Olimex vendor SOUNDGRAPH 0x15c2 Soundgraph, Inc. vendor AMIT2 0x15c5 AMIT vendor TEXTECH 0x15ca Textech International Ltd. vendor SOHOWARE 0x15e8 SOHOware vendor UMAX 0x1606 UMAX Data Systems vendor INSIDEOUT 0x1608 Inside Out Networks vendor AMOI 0x1614 Amoi Electronics vendor GOODWAY 0x1631 Good Way Technology vendor ENTREGA 0x1645 Entrega vendor ACTIONTEC 0x1668 Actiontec Electronics vendor CLIPSAL 0x166a Clipsal vendor CISCOLINKSYS2 0x167b Cisco-Linksys vendor ATHEROS 0x168c Atheros Communications vendor GIGASET 0x1690 Gigaset vendor GLOBALSUN 0x16ab Global Sun Technology vendor ANYDATA 0x16d5 AnyDATA Corporation vendor JABLOTRON 0x16d6 Jablotron vendor CMOTECH 0x16d8 C-motech vendor WIENERPLEINBAUS 0x16dc WIENER Plein & Baus GmbH. vendor AXESSTEL 0x1726 Axesstel Co., Ltd. vendor LINKSYS4 0x1737 Linksys vendor SENAO 0x1740 Senao vendor ASUS2 0x1761 ASUS vendor SWEEX2 0x177f Sweex vendor METAGEEK 0x1781 MetaGeek vendor KAMSTRUP 0x17a8 Kamstrup A/S vendor LENOVO 0x17ef Lenovo vendor WAVESENSE 0x17f4 WaveSense vendor VAISALA 0x1843 Vaisala vendor AMIT 0x18c5 AMIT vendor GOOGLE 0x18d1 Google vendor QCOM 0x18e8 Qcom vendor ELV 0x18ef ELV vendor LINKSYS3 0x1915 Linksys vendor QUALCOMMINC 0x19d2 Qualcomm, Incorporated +vendor QUALCOMM3 0x19f5 Qualcomm, Inc. vendor BAYER 0x1a79 Bayer vendor WCH2 0x1a86 QinHeng Electronics vendor STELERA 0x1a8d Stelera Wireless vendor SEL 0x1adb Schweitzer Engineering Laboratories vendor CORSAIR 0x1b1c Corsair vendor MATRIXORBITAL 0x1b3d Matrix Orbital vendor OVISLINK 0x1b75 OvisLink vendor TML 0x1b91 The Mobility Lab vendor TCTMOBILE 0x1bbb TCT Mobile vendor ALTI2 0x1bc9 Alti-2 products vendor SUNPLUS 0x1bcf Sunplus Innovation Technology Inc. vendor WAGO 0x1be3 WAGO Kontakttechnik GmbH. vendor TELIT 0x1bc7 Telit vendor IONICS 0x1c0c Ionics PlugComputer vendor LONGCHEER 0x1c9e Longcheer Holdings, Ltd. vendor MPMAN 0x1cae MpMan vendor DRESDENELEKTRONIK 0x1cf1 dresden elektronik vendor NEOTEL 0x1d09 Neotel vendor DREAMLINK 0x1d34 Dream Link vendor PEGATRON 0x1d4d Pegatron vendor QISDA 0x1da5 Qisda vendor METAGEEK2 0x1dd5 MetaGeek vendor ALINK 0x1e0e Alink vendor AIRTIES 0x1eda AirTies vendor FESTO 0x1e29 Festo vendor LAKESHORE 0x1fb9 Lake Shore Cryotronics, Inc. vendor VERTEX 0x1fe7 Vertex Wireless Co., Ltd. vendor DLINK 0x2001 D-Link vendor PLANEX2 0x2019 Planex Communications vendor HAUPPAUGE2 0x2040 Hauppauge Computer Works vendor TLAYTECH 0x20b9 Tlay Tech vendor ENCORE 0x203d Encore vendor QIHARDWARE 0x20b7 QI-hardware vendor PARA 0x20b8 PARA Industrial vendor SIMTEC 0x20df Simtec Electronics vendor TRENDNET 0x20f4 TRENDnet vendor RTSYSTEMS 0x2100 RTSYSTEMS vendor VIALABS 0x2109 VIA Labs vendor ERICSSON 0x2282 Ericsson vendor MOTOROLA2 0x22b8 Motorola vendor WETELECOM 0x22de WeTelecom vendor WESTMOUNTAIN 0x2405 West Mountain Radio vendor TRIPPLITE 0x2478 Tripp-Lite vendor HIROSE 0x2631 Hirose Electric vendor NHJ 0x2770 NHJ vendor PLANEX 0x2c02 Planex Communications vendor VIDZMEDIA 0x3275 VidzMedia Pte Ltd vendor LINKINSTRUMENTS 0x3195 Link Instruments Inc. vendor AEI 0x3334 AEI vendor HANK 0x3353 Hank Connection vendor PQI 0x3538 PQI vendor DAISY 0x3579 Daisy Technology vendor NI 0x3923 National Instruments vendor MICRONET 0x3980 Micronet Communications vendor IODATA2 0x40bb I-O Data vendor IRIVER 0x4102 iRiver vendor DELL 0x413c Dell vendor WCH 0x4348 QinHeng Electronics vendor ACEECA 0x4766 Aceeca vendor FEIXUN 0x4855 FeiXun Communication vendor PAPOUCH 0x5050 Papouch products vendor AVERATEC 0x50c2 Averatec vendor SWEEX 0x5173 Sweex vendor PROLIFIC2 0x5372 Prolific Technologies vendor ONSPEC2 0x55aa OnSpec Electronic Inc. vendor ZINWELL 0x5a57 Zinwell vendor SITECOM 0x6189 Sitecom vendor ARKMICRO 0x6547 Arkmicro Technologies Inc. vendor 3COM2 0x6891 3Com vendor EDIMAX 0x7392 Edimax vendor INTEL 0x8086 Intel vendor INTEL2 0x8087 Intel vendor ALLWIN 0x8516 ALLWIN Tech vendor SITECOM2 0x9016 Sitecom vendor MOSCHIP 0x9710 MosChip Semiconductor vendor NETGEAR4 0x9846 Netgear vendor MARVELL 0x9e88 Marvell Technology Group Ltd. vendor 3COM3 0xa727 3Com vendor EVOLUTION 0xdeee Evolution Robotics products vendor DATAAPEX 0xdaae DataApex vendor HP2 0xf003 Hewlett Packard vendor LOGILINK 0xfc08 LogiLink vendor USRP 0xfffe GNU Radio USRP /* * List of known products. Grouped by vendor. */ /* 3Com products */ product 3COM HOMECONN 0x009d HomeConnect Camera product 3COM 3CREB96 0x00a0 Bluetooth USB Adapter product 3COM 3C19250 0x03e8 3C19250 Ethernet Adapter product 3COM 3CRSHEW696 0x0a01 3CRSHEW696 Wireless Adapter product 3COM 3C460 0x11f8 HomeConnect 3C460 product 3COM USR56K 0x3021 U.S.Robotics 56000 Voice FaxModem Pro product 3COM 3C460B 0x4601 HomeConnect 3C460B product 3COM2 3CRUSB10075 0xa727 3CRUSB10075 product 3COM3 AR5523_1 0x6893 AR5523 product 3COM3 AR5523_2 0x6895 AR5523 product 3COM3 AR5523_3 0x6897 AR5523 product 3COMUSR OFFICECONN 0x0082 3Com OfficeConnect Analog Modem product 3COMUSR USRISDN 0x008f 3Com U.S. Robotics Pro ISDN TA product 3COMUSR HOMECONN 0x009d 3Com HomeConnect Camera product 3COMUSR USR56K 0x3021 U.S. Robotics 56000 Voice FaxModem Pro /* AboCom products */ product ABOCOM XX1 0x110c XX1 product ABOCOM XX2 0x200c XX2 product ABOCOM RT2770 0x2770 RT2770 product ABOCOM RT2870 0x2870 RT2870 product ABOCOM RT3070 0x3070 RT3070 product ABOCOM RT3071 0x3071 RT3071 product ABOCOM RT3072 0x3072 RT3072 product ABOCOM2 RT2870_1 0x3c09 RT2870 product ABOCOM URE450 0x4000 URE450 Ethernet Adapter product ABOCOM UFE1000 0x4002 UFE1000 Fast Ethernet Adapter product ABOCOM DSB650TX_PNA 0x4003 1/10/100 Ethernet Adapter product ABOCOM XX4 0x4004 XX4 product ABOCOM XX5 0x4007 XX5 product ABOCOM XX6 0x400b XX6 product ABOCOM XX7 0x400c XX7 product ABOCOM RTL8151 0x401a RTL8151 product ABOCOM XX8 0x4102 XX8 product ABOCOM XX9 0x4104 XX9 product ABOCOM UF200 0x420a UF200 Ethernet product ABOCOM WL54 0x6001 WL54 product ABOCOM XX10 0xabc1 XX10 product ABOCOM BWU613 0xb000 BWU613 product ABOCOM HWU54DM 0xb21b HWU54DM product ABOCOM RT2573_2 0xb21c RT2573 product ABOCOM RT2573_3 0xb21d RT2573 product ABOCOM RT2573_4 0xb21e RT2573 product ABOCOM RTL8188CU_1 0x8188 RTL8188CU product ABOCOM RTL8188CU_2 0x8189 RTL8188CU product ABOCOM RTL8192CU 0x8178 RTL8192CU product ABOCOM WUG2700 0xb21f WUG2700 /* Acton Research Corp. */ product ACTON SPECTRAPRO 0x0100 FTDI compatible adapter /* Accton products */ product ACCTON USB320_EC 0x1046 USB320-EC Ethernet Adapter product ACCTON 2664W 0x3501 2664W product ACCTON 111 0x3503 T-Sinus 111 Wireless Adapter product ACCTON SMCWUSBG_NF 0x4505 SMCWUSB-G (no firmware) product ACCTON SMCWUSBG 0x4506 SMCWUSB-G product ACCTON SMCWUSBTG2_NF 0x4507 SMCWUSBT-G2 (no firmware) product ACCTON SMCWUSBTG2 0x4508 SMCWUSBT-G2 product ACCTON PRISM_GT 0x4521 PrismGT USB 2.0 WLAN product ACCTON SS1001 0x5046 SpeedStream Ethernet Adapter product ACCTON RT2870_2 0x6618 RT2870 product ACCTON RT3070 0x7511 RT3070 product ACCTON RT2770 0x7512 RT2770 product ACCTON RT2870_3 0x7522 RT2870 product ACCTON RT2870_5 0x8522 RT2870 product ACCTON RT3070_4 0xa512 RT3070 product ACCTON RT2870_4 0xa618 RT2870 product ACCTON RT3070_1 0xa701 RT3070 product ACCTON RT3070_2 0xa702 RT3070 product ACCTON RT2870_1 0xb522 RT2870 product ACCTON RT3070_3 0xc522 RT3070 product ACCTON RT3070_5 0xd522 RT3070 product ACCTON RTL8192SU 0xc512 RTL8192SU product ACCTON ZD1211B 0xe501 ZD1211B /* Aceeca products */ product ACEECA MEZ1000 0x0001 MEZ1000 RDA /* Acer Communications & Multimedia (oemd by Surecom) */ product ACERCM EP1427X2 0x0893 EP-1427X-2 Ethernet Adapter /* Acer Labs products */ product ACERLABS M5632 0x5632 USB 2.0 Data Link /* Acer Peripherals, Inc. products */ product ACERP ACERSCAN_C310U 0x12a6 Acerscan C310U product ACERP ACERSCAN_320U 0x2022 Acerscan 320U product ACERP ACERSCAN_640U 0x2040 Acerscan 640U product ACERP ACERSCAN_620U 0x2060 Acerscan 620U product ACERP ACERSCAN_4300U 0x20b0 Benq 3300U/4300U product ACERP ACERSCAN_640BT 0x20be Acerscan 640BT product ACERP ACERSCAN_1240U 0x20c0 Acerscan 1240U product ACERP S81 0x4027 BenQ S81 phone product ACERP H10 0x4068 AWL400 Wireless Adapter product ACERP ATAPI 0x6003 ATA/ATAPI Adapter product ACERP AWL300 0x9000 AWL300 Wireless Adapter product ACERP AWL400 0x9001 AWL400 Wireless Adapter /* Acer Warp products */ product ACERW WARPLINK 0x0204 Warplink /* Actions products */ product ACTIONS MP4 0x1101 Actions MP4 Player /* Actiontec, Inc. products */ product ACTIONTEC PRISM_25 0x0408 Prism2.5 Wireless Adapter product ACTIONTEC PRISM_25A 0x0421 Prism2.5 Wireless Adapter A product ACTIONTEC FREELAN 0x6106 ROPEX FreeLan 802.11b product ACTIONTEC UAT1 0x7605 UAT1 Wireless Ethernet Adapter /* ACTiSYS products */ product ACTISYS IR2000U 0x0011 ACT-IR2000U FIR /* ActiveWire, Inc. products */ product ACTIVEWIRE IOBOARD 0x0100 I/O Board product ACTIVEWIRE IOBOARD_FW1 0x0101 I/O Board, rev. 1 firmware /* Adaptec products */ product ADAPTEC AWN8020 0x0020 AWN-8020 WLAN /* Addtron products */ product ADDTRON AWU120 0xff31 AWU-120 /* ADLINK Texhnology products */ product ADLINK ND6530 0x6530 ND-6530 USB-Serial /* ADMtek products */ product ADMTEK PEGASUSII_4 0x07c2 AN986A Ethernet product ADMTEK PEGASUS 0x0986 AN986 Ethernet product ADMTEK PEGASUSII 0x8511 AN8511 Ethernet product ADMTEK PEGASUSII_2 0x8513 AN8513 Ethernet product ADMTEK PEGASUSII_3 0x8515 AN8515 Ethernet /* ADDON products */ /* PNY OEMs these */ product ADDON ATTACHE 0x1300 USB 2.0 Flash Drive product ADDON ATTACHE 0x1300 USB 2.0 Flash Drive product ADDON A256MB 0x1400 Attache 256MB USB 2.0 Flash Drive product ADDON DISKPRO512 0x1420 USB 2.0 Flash Drive (DANE-ELEC zMate 512MB USB flash drive) /* Addonics products */ product ADDONICS2 CABLE_205 0xa001 Cable 205 /* ADS products */ product ADS UBS10BT 0x0008 UBS-10BT Ethernet product ADS UBS10BTX 0x0009 UBS-10BT Ethernet /* AEI products */ product AEI FASTETHERNET 0x1701 Fast Ethernet /* Afatech Technologies, Inc. */ product AFATECH AFATECH1336 0x1336 Flash Card Reader /* Agate Technologies products */ product AGATE QDRIVE 0x0378 Q-Drive /* AGFA products */ product AGFA SNAPSCAN1212U 0x0001 SnapScan 1212U product AGFA SNAPSCAN1236U 0x0002 SnapScan 1236U product AGFA SNAPSCANTOUCH 0x0100 SnapScan Touch product AGFA SNAPSCAN1212U2 0x2061 SnapScan 1212U product AGFA SNAPSCANE40 0x208d SnapScan e40 product AGFA SNAPSCANE50 0x208f SnapScan e50 product AGFA SNAPSCANE20 0x2091 SnapScan e20 product AGFA SNAPSCANE25 0x2095 SnapScan e25 product AGFA SNAPSCANE26 0x2097 SnapScan e26 product AGFA SNAPSCANE52 0x20fd SnapScan e52 /* Ain Communication Technology products */ product AINCOMM AWU2000B 0x1001 AWU2000B Wireless Adapter /* AIPTEK products */ product AIPTEK POCKETCAM3M 0x2011 PocketCAM 3Mega product AIPTEK2 PENCAM_MEGA_1_3 0x504a PenCam Mega 1.3 product AIPTEK2 SUNPLUS_TECH 0x0c15 Sunplus Technology Inc. /* AirPlis products */ product AIRPLUS MCD650 0x3198 MCD650 modem /* AirPrime products */ product AIRPRIME PC5220 0x0112 CDMA Wireless PC Card product AIRPRIME USB308 0x68A3 USB308 HSPA+ USB Modem product AIRPRIME AC313U 0x68aa Sierra Wireless AirCard 313U /* AirTies products */ product AIRTIES RT3070 0x2310 RT3070 /* AKS products */ product AKS USBHASP 0x0001 USB-HASP 0.06 /* Alcatel products */ product ALCATEL OT535 0x02df One Touch 535/735 /* Alcor Micro, Inc. products */ product ALCOR2 KBD_HUB 0x2802 Kbd Hub product ALCOR DUMMY 0x0000 Dummy product product ALCOR SDCR_6335 0x6335 SD/MMC Card Reader product ALCOR SDCR_6362 0x6362 SD/MMC Card Reader product ALCOR SDCR_6366 0x6366 SD/MMC Card Reader product ALCOR TRANSCEND 0x6387 Transcend JetFlash Drive product ALCOR MA_KBD_HUB 0x9213 MacAlly Kbd Hub product ALCOR AU9814 0x9215 AU9814 Hub product ALCOR UMCR_9361 0x9361 USB Multimedia Card Reader product ALCOR SM_KBD 0x9410 MicroConnectors/StrongMan Keyboard product ALCOR NEC_KBD_HUB 0x9472 NEC Kbd Hub product ALCOR AU9720 0x9720 USB2 - RS-232 product ALCOR AU6390 0x6390 AU6390 USB-IDE converter /* Alink products */ product ALINK DWM652U5 0xce16 DWM-652 product ALINK 3G 0x9000 3G modem product ALINK 3GU 0x9200 3G modem /* Altec Lansing products */ product ALTEC ADA70 0x0070 ADA70 Speakers product ALTEC ASC495 0xff05 ASC495 Speakers /* Alti-2 products */ product ALTI2 N3 0x6001 FTDI compatible adapter /* Allied Telesyn International products */ product ALLIEDTELESYN ATUSB100 0xb100 AT-USB100 /* ALLWIN Tech products */ product ALLWIN RT2070 0x2070 RT2070 product ALLWIN RT2770 0x2770 RT2770 product ALLWIN RT2870 0x2870 RT2870 product ALLWIN RT3070 0x3070 RT3070 product ALLWIN RT3071 0x3071 RT3071 product ALLWIN RT3072 0x3072 RT3072 product ALLWIN RT3572 0x3572 RT3572 /* AlphaSmart, Inc. products */ product ALPHASMART DANA_KB 0xdbac AlphaSmart Dana Keyboard product ALPHASMART DANA_SYNC 0xdf00 AlphaSmart Dana HotSync /* Amoi products */ product AMOI H01 0x0800 H01 3G modem product AMOI H01A 0x7002 H01A 3G modem product AMOI H02 0x0802 H02 3G modem /* American Power Conversion products */ product APC UPS 0x0002 Uninterruptible Power Supply /* Ambit Microsystems products */ product AMBIT WLAN 0x0302 WLAN product AMBIT NTL_250 0x6098 NTL 250 cable modem /* Apacer products */ product APACER HT202 0xb113 USB 2.0 Flash Drive /* American Power Conversion products */ product APC UPS 0x0002 Uninterruptible Power Supply /* Amigo Technology products */ product AMIGO RT2870_1 0x9031 RT2870 product AMIGO RT2870_2 0x9041 RT2870 /* AMIT products */ product AMIT CGWLUSB2GO 0x0002 CG-WLUSB2GO product AMIT CGWLUSB2GNR 0x0008 CG-WLUSB2GNR product AMIT RT2870_1 0x0012 RT2870 /* AMIT(2) products */ product AMIT2 RT2870 0x0008 RT2870 /* Analog Devices products */ product ANALOGDEVICES GNICE 0xf000 FTDI compatible adapter product ANALOGDEVICES GNICEPLUS 0xf001 FTDI compatible adapter /* Anchor products */ product ANCHOR SERIAL 0x2008 Serial product ANCHOR EZUSB 0x2131 EZUSB product ANCHOR EZLINK 0x2720 EZLINK /* AnyData products */ product ANYDATA ADU_620UW 0x6202 CDMA 2000 EV-DO USB Modem product ANYDATA ADU_E100X 0x6501 CDMA 2000 1xRTT/EV-DO USB Modem product ANYDATA ADU_500A 0x6502 CDMA 2000 EV-DO USB Modem /* AOX, Inc. products */ product AOX USB101 0x0008 Ethernet /* American Power Conversion products */ product APC UPS 0x0002 Uninterruptible Power Supply /* Apple Computer products */ product APPLE DUMMY 0x0000 Dummy product product APPLE IMAC_KBD 0x0201 USB iMac Keyboard product APPLE KBD 0x0202 USB Keyboard M2452 product APPLE EXT_KBD 0x020c Apple Extended USB Keyboard /* MacbookAir, aka wellspring */ product APPLE WELLSPRING_ANSI 0x0223 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING_ISO 0x0224 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING_JIS 0x0225 Apple Internal Keyboard/Trackpad /* MacbookProPenryn, aka wellspring2 */ product APPLE WELLSPRING2_ANSI 0x0230 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING2_ISO 0x0231 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING2_JIS 0x0232 Apple Internal Keyboard/Trackpad /* Macbook5,1 (unibody), aka wellspring3 */ product APPLE WELLSPRING3_ANSI 0x0236 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING3_ISO 0x0237 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING3_JIS 0x0238 Apple Internal Keyboard/Trackpad /* MacbookAir3,2 (unibody), aka wellspring4 */ product APPLE WELLSPRING4_ANSI 0x023f Apple Internal Keyboard/Trackpad product APPLE WELLSPRING4_ISO 0x0240 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING4_JIS 0x0241 Apple Internal Keyboard/Trackpad /* MacbookAir3,1 (unibody), aka wellspring4 */ product APPLE WELLSPRING4A_ANSI 0x0242 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING4A_ISO 0x0243 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING4A_JIS 0x0244 Apple Internal Keyboard/Trackpad /* Macbook8 (unibody, March 2011) */ product APPLE WELLSPRING5_ANSI 0x0245 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING5_ISO 0x0246 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING5_JIS 0x0247 Apple Internal Keyboard/Trackpad /* MacbookAir4,1 (unibody, July 2011) */ product APPLE WELLSPRING6A_ANSI 0x0249 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING6A_ISO 0x024a Apple Internal Keyboard/Trackpad product APPLE WELLSPRING6A_JIS 0x024b Apple Internal Keyboard/Trackpad /* MacbookAir4,2 (unibody, July 2011) */ product APPLE WELLSPRING6_ANSI 0x024c Apple Internal Keyboard/Trackpad product APPLE WELLSPRING6_ISO 0x024d Apple Internal Keyboard/Trackpad product APPLE WELLSPRING6_JIS 0x024e Apple Internal Keyboard/Trackpad /* Macbook8,2 (unibody) */ product APPLE WELLSPRING5A_ANSI 0x0252 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING5A_ISO 0x0253 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING5A_JIS 0x0254 Apple Internal Keyboard/Trackpad /* MacbookPro10,1 (unibody, June 2012) */ product APPLE WELLSPRING7_ANSI 0x0262 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING7_ISO 0x0263 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING7_JIS 0x0264 Apple Internal Keyboard/Trackpad /* MacbookPro10,2 (unibody, October 2012) */ product APPLE WELLSPRING7A_ANSI 0x0259 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING7A_ISO 0x025a Apple Internal Keyboard/Trackpad product APPLE WELLSPRING7A_JIS 0x025b Apple Internal Keyboard/Trackpad /* MacbookAir6,2 (unibody, June 2013) */ product APPLE WELLSPRING8_ANSI 0x0290 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING8_ISO 0x0291 Apple Internal Keyboard/Trackpad product APPLE WELLSPRING8_JIS 0x0292 Apple Internal Keyboard/Trackpad product APPLE MOUSE 0x0301 Mouse M4848 product APPLE OPTMOUSE 0x0302 Optical mouse product APPLE MIGHTYMOUSE 0x0304 Mighty Mouse product APPLE KBD_HUB 0x1001 Hub in Apple USB Keyboard product APPLE EXT_KBD_HUB 0x1003 Hub in Apple Extended USB Keyboard product APPLE SPEAKERS 0x1101 Speakers product APPLE IPOD 0x1201 iPod product APPLE IPOD2G 0x1202 iPod 2G product APPLE IPOD3G 0x1203 iPod 3G product APPLE IPOD_04 0x1204 iPod '04' product APPLE IPODMINI 0x1205 iPod Mini product APPLE IPOD_06 0x1206 iPod '06' product APPLE IPOD_07 0x1207 iPod '07' product APPLE IPOD_08 0x1208 iPod '08' product APPLE IPODVIDEO 0x1209 iPod Video product APPLE IPODNANO 0x120a iPod Nano product APPLE IPHONE 0x1290 iPhone product APPLE IPOD_TOUCH 0x1291 iPod Touch product APPLE IPHONE_3G 0x1292 iPhone 3G product APPLE IPHONE_3GS 0x1294 iPhone 3GS product APPLE IPHONE_4 0x1297 iPhone 4 product APPLE IPHONE_4S 0x12a0 iPhone 4S product APPLE IPHONE_5 0x12a8 iPhone 5 product APPLE IPAD 0x129a iPad product APPLE ETHERNET 0x1402 Ethernet A1277 /* Arkmicro Technologies */ product ARKMICRO ARK3116 0x0232 ARK3116 Serial /* Asahi Optical products */ product ASAHIOPTICAL OPTIO230 0x0004 Digital camera product ASAHIOPTICAL OPTIO330 0x0006 Digital camera /* Asante products */ product ASANTE EA 0x1427 Ethernet /* ASIX Electronics products */ product ASIX AX88172 0x1720 10/100 Ethernet product ASIX AX88178 0x1780 AX88178 product ASIX AX88178A 0x178a AX88178A USB 2.0 10/100/1000 Ethernet product ASIX AX88179 0x1790 AX88179 USB 3.0 10/100/1000 Ethernet product ASIX AX88772 0x7720 AX88772 product ASIX AX88772A 0x772a AX88772A USB 2.0 10/100 Ethernet product ASIX AX88772B 0x772b AX88772B USB 2.0 10/100 Ethernet product ASIX AX88772B_1 0x7e2b AX88772B USB 2.0 10/100 Ethernet /* ASUS products */ product ASUS2 USBN11 0x0b05 USB-N11 product ASUS RT2570 0x1706 RT2500USB Wireless Adapter product ASUS WL167G 0x1707 WL-167g Wireless Adapter product ASUS WL159G 0x170c WL-159g product ASUS A9T_WIFI 0x171b A9T wireless product ASUS P5B_WIFI 0x171d P5B wireless product ASUS RT2573_1 0x1723 RT2573 product ASUS RT2573_2 0x1724 RT2573 product ASUS LCM 0x1726 LCM display product ASUS RT2870_1 0x1731 RT2870 product ASUS RT2870_2 0x1732 RT2870 product ASUS RT2870_3 0x1742 RT2870 product ASUS RT2870_4 0x1760 RT2870 product ASUS RT2870_5 0x1761 RT2870 product ASUS USBN13 0x1784 USB-N13 product ASUS USBN10 0x1786 USB-N10 product ASUS RT3070_1 0x1790 RT3070 product ASUS RTL8192SU 0x1791 RTL8192SU product ASUS USB_N53 0x179d ASUS Black Diamond Dual Band USB-N53 product ASUS RTL8192CU 0x17ab RTL8192CU product ASUS USBN66 0x17ad USB-N66 product ASUS USBN10NANO 0x17ba USB-N10 Nano product ASUS USBAC51 0x17d1 USB-AC51 product ASUS A730W 0x4202 ASUS MyPal A730W product ASUS P535 0x420f ASUS P535 PDA product ASUS GMSC 0x422f ASUS Generic Mass Storage /* ATen products */ product ATEN UC1284 0x2001 Parallel printer product ATEN UC10T 0x2002 10Mbps Ethernet product ATEN UC110T 0x2007 UC-110T Ethernet product ATEN UC232A 0x2008 Serial product ATEN UC210T 0x2009 UC-210T Ethernet product ATEN DSB650C 0x4000 DSB-650C /* ATP Electronics products */ product ATP EUSB 0xaf01 ATP IG eUSB SSD /* Atheros Communications products */ product ATHEROS AR5523 0x0001 AR5523 product ATHEROS AR5523_NF 0x0002 AR5523 (no firmware) product ATHEROS2 AR5523_1 0x0001 AR5523 product ATHEROS2 AR5523_1_NF 0x0002 AR5523 (no firmware) product ATHEROS2 AR5523_2 0x0003 AR5523 product ATHEROS2 AR5523_2_NF 0x0004 AR5523 (no firmware) product ATHEROS2 AR5523_3 0x0005 AR5523 product ATHEROS2 AR5523_3_NF 0x0006 AR5523 (no firmware) /* Atmel Comp. products */ product ATMEL STK541 0x2109 Zigbee Controller product ATMEL UHB124 0x3301 UHB124 hub product ATMEL DWL120 0x7603 DWL-120 Wireless Adapter product ATMEL BW002 0x7605 BW002 Wireless Adapter product ATMEL WL1130USB 0x7613 WL-1130 USB product ATMEL AT76C505A 0x7614 AT76c505a Wireless Adapter /* AuthenTec products */ product AUTHENTEC AES1610 0x1600 AES1610 Fingerprint Sensor /* Avision products */ product AVISION 1200U 0x0268 1200U scanner /* Axesstel products */ product AXESSTEL DATAMODEM 0x1000 Data Modem /* AsureWave products */ product AZUREWAVE RT2870_1 0x3247 RT2870 product AZUREWAVE RT2870_2 0x3262 RT2870 product AZUREWAVE RT3070_1 0x3273 RT3070 product AZUREWAVE RT3070_2 0x3284 RT3070 product AZUREWAVE RT3070_3 0x3305 RT3070 product AZUREWAVE RTL8188CU 0x3357 RTL8188CU product AZUREWAVE RTL8188CE_1 0x3358 RTL8188CE product AZUREWAVE RTL8188CE_2 0x3359 RTL8188CE product AZUREWAVE RTL8192SU_1 0x3306 RTL8192SU product AZUREWAVE RTL8192SU_2 0x3309 RTL8192SU product AZUREWAVE RTL8192SU_3 0x3310 RTL8192SU product AZUREWAVE RTL8192SU_4 0x3311 RTL8192SU product AZUREWAVE RTL8192SU_5 0x3325 RTL8192SU /* Baltech products */ product BALTECH CARDREADER 0x9999 Card reader /* Bayer products */ product BAYER CONTOUR_CABLE 0x6001 FTDI compatible adapter /* B&B Electronics products */ product BBELECTRONICS USOTL4 0xAC01 RS-422/485 product BBELECTRONICS 232USB9M 0xac27 FTDI compatible adapter product BBELECTRONICS 485USB9F_2W 0xac25 FTDI compatible adapter product BBELECTRONICS 485USB9F_4W 0xac26 FTDI compatible adapter product BBELECTRONICS 485USBTB_2W 0xac33 FTDI compatible adapter product BBELECTRONICS 485USBTB_4W 0xac34 FTDI compatible adapter product BBELECTRONICS TTL3USB9M 0xac50 FTDI compatible adapter product BBELECTRONICS TTL5USB9M 0xac49 FTDI compatible adapter product BBELECTRONICS USO9ML2 0xac03 FTDI compatible adapter product BBELECTRONICS USO9ML2DR 0xac17 FTDI compatible adapter product BBELECTRONICS USO9ML2DR_2 0xac16 FTDI compatible adapter product BBELECTRONICS USOPTL4 0xac11 FTDI compatible adapter product BBELECTRONICS USOPTL4DR 0xac19 FTDI compatible adapter product BBELECTRONICS USOPTL4DR2 0xac18 FTDI compatible adapter product BBELECTRONICS USPTL4 0xac12 FTDI compatible adapter product BBELECTRONICS USTL4 0xac02 FTDI compatible adapter product BBELECTRONICS ZZ_PROG1_USB 0xba02 FTDI compatible adapter /* Belkin products */ /*product BELKIN F5U111 0x???? F5U111 Ethernet*/ product BELKIN F5D6050 0x0050 F5D6050 802.11b Wireless Adapter product BELKIN FBT001V 0x0081 FBT001v2 Bluetooth product BELKIN FBT003V 0x0084 FBT003v2 Bluetooth product BELKIN F5U103 0x0103 F5U103 Serial product BELKIN F5U109 0x0109 F5U109 Serial product BELKIN USB2SCSI 0x0115 USB to SCSI product BELKIN F8T012 0x0121 F8T012xx1 Bluetooth USB Adapter product BELKIN USB2LAN 0x0121 USB to LAN product BELKIN F5U208 0x0208 F5U208 VideoBus II product BELKIN F5U237 0x0237 F5U237 USB 2.0 7-Port Hub product BELKIN F5U257 0x0257 F5U257 Serial product BELKIN F5U409 0x0409 F5U409 Serial product BELKIN F6C550AVR 0x0551 F6C550-AVR UPS product BELKIN F5U120 0x1203 F5U120-PC Hub product BELKIN RTL8188CU 0x1102 RTL8188CU Wireless Adapter product BELKIN F9L1103 0x1103 F9L1103 Wireless Adapter product BELKIN RTL8192CU 0x2102 RTL8192CU Wireless Adapter product BELKIN F7D2102 0x2103 F7D2102 Wireless Adapter product BELKIN ZD1211B 0x4050 ZD1211B product BELKIN F5D5055 0x5055 F5D5055 product BELKIN F5D7050 0x7050 F5D7050 Wireless Adapter product BELKIN F5D7051 0x7051 F5D7051 54g USB Network Adapter product BELKIN F5D7050A 0x705a F5D7050A Wireless Adapter /* Also sold as 'Ativa 802.11g wireless card' */ product BELKIN F5D7050_V4000 0x705c F5D7050 v4000 Wireless Adapter product BELKIN F5D7050E 0x705e F5D7050E Wireless Adapter product BELKIN RT2870_1 0x8053 RT2870 product BELKIN RT2870_2 0x805c RT2870 product BELKIN F5D8053V3 0x815c F5D8053 v3 product BELKIN RTL8192SU_1 0x815f RTL8192SU product BELKIN RTL8192SU_2 0x845a RTL8192SU product BELKIN RTL8192SU_3 0x945a RTL8192SU product BELKIN F5D8055 0x825a F5D8055 product BELKIN F5D8055V2 0x825b F5D8055 v2 product BELKIN F5D9050V3 0x905b F5D9050 ver 3 Wireless Adapter product BELKIN2 F5U002 0x0002 F5U002 Parallel printer product BELKIN F6D4050V1 0x935a F6D4050 v1 product BELKIN F6D4050V2 0x935b F6D4050 v2 /* Billionton products */ product BILLIONTON USB100 0x0986 USB100N 10/100 FastEthernet product BILLIONTON USBLP100 0x0987 USB100LP product BILLIONTON USBEL100 0x0988 USB100EL product BILLIONTON USBE100 0x8511 USBE100 product BILLIONTON USB2AR 0x90ff USB2AR Ethernet /* Broadcom products */ product BROADCOM BCM2033 0x2033 BCM2033 Bluetooth USB dongle /* Brother Industries products */ product BROTHER HL1050 0x0002 HL-1050 laser printer product BROTHER MFC8600_9650 0x0100 MFC8600/9650 multifunction device /* Behavior Technology Computer products */ product BTC BTC6100 0x5550 6100C Keyboard product BTC BTC7932 0x6782 Keyboard with mouse port /* Canon, Inc. products */ product CANON N656U 0x2206 CanoScan N656U product CANON N1220U 0x2207 CanoScan N1220U product CANON D660U 0x2208 CanoScan D660U product CANON N676U 0x220d CanoScan N676U product CANON N1240U 0x220e CanoScan N1240U product CANON LIDE25 0x2220 CanoScan LIDE 25 product CANON S10 0x3041 PowerShot S10 product CANON S100 0x3045 PowerShot S100 product CANON S200 0x3065 PowerShot S200 product CANON REBELXT 0x30ef Digital Rebel XT /* CATC products */ product CATC NETMATE 0x000a Netmate Ethernet product CATC NETMATE2 0x000c Netmate2 Ethernet product CATC CHIEF 0x000d USB Chief Bus & Protocol Analyzer product CATC ANDROMEDA 0x1237 Andromeda hub /* CASIO products */ product CASIO QV_DIGICAM 0x1001 QV DigiCam product CASIO EXS880 0x1105 Exilim EX-S880 product CASIO BE300 0x2002 BE-300 PDA product CASIO NAMELAND 0x4001 CASIO Nameland EZ-USB /* CCYU products */ product CCYU ED1064 0x2136 EasyDisk ED1064 /* Century products */ product CENTURY EX35QUAT 0x011e Century USB Disk Enclosure product CENTURY EX35SW4_SB4 0x011f Century USB Disk Enclosure /* Cherry products */ product CHERRY MY3000KBD 0x0001 My3000 keyboard product CHERRY MY3000HUB 0x0003 My3000 hub product CHERRY CYBOARD 0x0004 CyBoard Keyboard /* Chic Technology products */ product CHIC MOUSE1 0x0001 mouse product CHIC CYPRESS 0x0003 Cypress USB Mouse /* Chicony products */ product CHICONY KB8933 0x0001 KB-8933 keyboard product CHICONY KU0325 0x0116 KU-0325 keyboard product CHICONY CNF7129 0xb071 Notebook Web Camera product CHICONY HDUVCCAM 0xb40a HD UVC WebCam product CHICONY RTL8188CUS_1 0xaff7 RTL8188CUS product CHICONY RTL8188CUS_2 0xaff8 RTL8188CUS product CHICONY RTL8188CUS_3 0xaff9 RTL8188CUS product CHICONY RTL8188CUS_4 0xaffa RTL8188CUS product CHICONY RTL8188CUS_5 0xaffa RTL8188CUS product CHICONY2 TWINKLECAM 0x600d TwinkleCam USB camera /* CH Products */ product CHPRODUCTS PROTHROTTLE 0x00f1 Pro Throttle product CHPRODUCTS PROPEDALS 0x00f2 Pro Pedals product CHPRODUCTS FIGHTERSTICK 0x00f3 Fighterstick product CHPRODUCTS FLIGHTYOKE 0x00ff Flight Sim Yoke /* Cisco-Linksys products */ product CISCOLINKSYS WUSB54AG 0x000c WUSB54AG Wireless Adapter product CISCOLINKSYS WUSB54G 0x000d WUSB54G Wireless Adapter product CISCOLINKSYS WUSB54GP 0x0011 WUSB54GP Wireless Adapter product CISCOLINKSYS USB200MV2 0x0018 USB200M v2 product CISCOLINKSYS HU200TS 0x001a HU200TS Wireless Adapter product CISCOLINKSYS WUSB54GC 0x0020 WUSB54GC product CISCOLINKSYS WUSB54GR 0x0023 WUSB54GR product CISCOLINKSYS WUSBF54G 0x0024 WUSBF54G product CISCOLINKSYS AE1000 0x002f AE1000 product CISCOLINKSYS2 RT3070 0x4001 RT3070 product CISCOLINKSYS3 RT3070 0x0101 RT3070 /* Clipsal products */ product CLIPSAL 560884 0x0101 560884 C-Bus Audio Matrix Switch product CLIPSAL 5500PACA 0x0201 5500PACA C-Bus Pascal Automation Controller product CLIPSAL 5800PC 0x0301 5800PC C-Bus Wireless Interface product CLIPSAL 5500PCU 0x0303 5500PCU C-Bus Interface product CLIPSAL 5000CT2 0x0304 5000CT2 C-Bus Touch Screen product CLIPSAL C5000CT2 0x0305 C5000CT2 C-Bus Touch Screen product CLIPSAL L51xx 0x0401 L51xx C-Bus Dimmer /* CMOTECH products */ product CMOTECH CNU510 0x5141 CDMA Technologies USB modem product CMOTECH CNU550 0x5543 CDMA 2000 1xRTT/1xEVDO USB modem product CMOTECH CGU628 0x6006 CGU-628 product CMOTECH CDMA_MODEM1 0x6280 CDMA Technologies USB modem product CMOTECH DISK 0xf000 disk mode /* Compaq products */ product COMPAQ IPAQPOCKETPC 0x0003 iPAQ PocketPC product COMPAQ PJB100 0x504a Personal Jukebox PJB100 product COMPAQ IPAQLINUX 0x505a iPAQ Linux /* Composite Corp products looks the same as "TANGTOP" */ product COMPOSITE USBPS2 0x0001 USB to PS2 Adaptor /* Conceptronic products */ product CONCEPTRONIC PRISM_GT 0x3762 PrismGT USB 2.0 WLAN product CONCEPTRONIC C11U 0x7100 C11U product CONCEPTRONIC WL210 0x7110 WL-210 product CONCEPTRONIC AR5523_1 0x7801 AR5523 product CONCEPTRONIC AR5523_1_NF 0x7802 AR5523 (no firmware) product CONCEPTRONIC AR5523_2 0x7811 AR5523 product CONCEPTRONIC AR5523_2_NF 0x7812 AR5523 (no firmware) product CONCEPTRONIC2 RTL8192SU_1 0x3300 RTL8192SU product CONCEPTRONIC2 RTL8192SU_2 0x3301 RTL8192SU product CONCEPTRONIC2 RTL8192SU_3 0x3302 RTL8192SU product CONCEPTRONIC2 C54RU 0x3c02 C54RU WLAN product CONCEPTRONIC2 C54RU2 0x3c22 C54RU product CONCEPTRONIC2 RT3070_1 0x3c08 RT3070 product CONCEPTRONIC2 RT3070_2 0x3c11 RT3070 product CONCEPTRONIC2 VIGORN61 0x3c25 VIGORN61 product CONCEPTRONIC2 RT2870_1 0x3c06 RT2870 product CONCEPTRONIC2 RT2870_2 0x3c07 RT2870 product CONCEPTRONIC2 RT2870_7 0x3c09 RT2870 product CONCEPTRONIC2 RT2870_8 0x3c12 RT2870 product CONCEPTRONIC2 RT2870_3 0x3c23 RT2870 product CONCEPTRONIC2 RT2870_4 0x3c25 RT2870 product CONCEPTRONIC2 RT2870_5 0x3c27 RT2870 product CONCEPTRONIC2 RT2870_6 0x3c28 RT2870 /* Connectix products */ product CONNECTIX QUICKCAM 0x0001 QuickCam /* Conect products */ product CONTEC COM1USBH 0x8311 FTDI compatible adapter /* Corega products */ product COREGA ETHER_USB_T 0x0001 Ether USB-T product COREGA FETHER_USB_TX 0x0004 FEther USB-TX product COREGA WLAN_USB_USB_11 0x000c WirelessLAN USB-11 product COREGA FETHER_USB_TXS 0x000d FEther USB-TXS product COREGA WLANUSB 0x0012 Wireless LAN Stick-11 product COREGA FETHER_USB2_TX 0x0017 FEther USB2-TX product COREGA WLUSB_11_KEY 0x001a ULUSB-11 Key product COREGA CGUSBRS232R 0x002a CG-USBRS232R product COREGA CGWLUSB2GL 0x002d CG-WLUSB2GL product COREGA CGWLUSB2GPX 0x002e CG-WLUSB2GPX product COREGA RT2870_1 0x002f RT2870 product COREGA RT2870_2 0x003c RT2870 product COREGA RT2870_3 0x003f RT2870 product COREGA RT3070 0x0041 RT3070 product COREGA CGWLUSB300GNM 0x0042 CG-WLUSB300GNM product COREGA RTL8192SU 0x0047 RTL8192SU product COREGA RTL8192CU 0x0056 RTL8192CU product COREGA WLUSB_11_STICK 0x7613 WLAN USB Stick 11 product COREGA FETHER_USB_TXC 0x9601 FEther USB-TXC /* Corsair products */ product CORSAIR K60 0x0a60 Corsair Vengeance K60 keyboard product CORSAIR K70 0x1b09 Corsair Vengeance K70 keyboard /* Creative products */ product CREATIVE NOMAD_II 0x1002 Nomad II MP3 player product CREATIVE NOMAD_IIMG 0x4004 Nomad II MG product CREATIVE NOMAD 0x4106 Nomad product CREATIVE2 VOIP_BLASTER 0x0258 Voip Blaster product CREATIVE3 OPTICAL_MOUSE 0x0001 Notebook Optical Mouse /* Cambridge Silicon Radio Ltd. products */ product CSR BT_DONGLE 0x0001 Bluetooth USB dongle product CSR CSRDFU 0xffff USB Bluetooth Device in DFU State /* Chipsbank Microelectronics Co., Ltd */ product CHIPSBANK USBMEMSTICK 0x6025 CBM2080 Flash drive controller product CHIPSBANK USBMEMSTICK1 0x6026 CBM1180 Flash drive controller /* CTX products */ product CTX EX1300 0x9999 Ex1300 hub /* Curitel products */ product CURITEL HX550C 0x1101 CDMA 2000 1xRTT USB modem (HX-550C) product CURITEL HX57XB 0x2101 CDMA 2000 1xRTT USB modem (HX-570/575B/PR-600) product CURITEL PC5740 0x3701 Broadband Wireless modem product CURITEL UM150 0x3711 EVDO modem product CURITEL UM175 0x3714 EVDO modem /* CyberPower products */ product CYBERPOWER 1500CAVRLCD 0x0501 1500CAVRLCD /* CyberTAN Technology products */ product CYBERTAN TG54USB 0x1666 TG54USB product CYBERTAN RT2870 0x1828 RT2870 /* Cypress Semiconductor products */ product CYPRESS MOUSE 0x0001 mouse product CYPRESS THERMO 0x0002 thermometer product CYPRESS WISPY1A 0x0bad MetaGeek Wi-Spy product CYPRESS KBDHUB 0x0101 Keyboard/Hub product CYPRESS FMRADIO 0x1002 FM Radio product CYPRESS IKARILASER 0x121f Ikari Laser SteelSeries ApS product CYPRESS USBRS232 0x5500 USB-RS232 Interface product CYPRESS SLIM_HUB 0x6560 Slim Hub product CYPRESS XX6830XX 0x6830 PATA Storage Device product CYPRESS SILVERSHIELD 0xfd13 Gembird Silver Shield PM /* Daisy Technology products */ product DAISY DMC 0x6901 USB MultiMedia Reader /* Dallas Semiconductor products */ product DALLAS J6502 0x4201 J-6502 speakers /* DataApex products */ product DATAAPEX MULTICOM 0xead6 MultiCom /* Dell products */ product DELL PORT 0x0058 Port Replicator product DELL AIO926 0x5115 Photo AIO Printer 926 product DELL BC02 0x8000 BC02 Bluetooth USB Adapter product DELL PRISM_GT_1 0x8102 PrismGT USB 2.0 WLAN product DELL TM350 0x8103 TrueMobile 350 Bluetooth USB Adapter product DELL PRISM_GT_2 0x8104 PrismGT USB 2.0 WLAN product DELL U5700 0x8114 Dell 5700 3G product DELL U5500 0x8115 Dell 5500 3G product DELL U5505 0x8116 Dell 5505 3G product DELL U5700_2 0x8117 Dell 5700 3G product DELL U5510 0x8118 Dell 5510 3G product DELL U5700_3 0x8128 Dell 5700 3G product DELL U5700_4 0x8129 Dell 5700 3G product DELL U5720 0x8133 Dell 5720 3G product DELL U5720_2 0x8134 Dell 5720 3G product DELL U740 0x8135 Dell U740 CDMA product DELL U5520 0x8136 Dell 5520 3G product DELL U5520_2 0x8137 Dell 5520 3G product DELL U5520_3 0x8138 Dell 5520 3G product DELL U5730 0x8180 Dell 5730 3G product DELL U5730_2 0x8181 Dell 5730 3G product DELL U5730_3 0x8182 Dell 5730 3G product DELL DW700 0x9500 Dell DW700 GPS /* Delorme Paublishing products */ product DELORME EARTHMATE 0x0100 Earthmate GPS /* Desknote products */ product DESKNOTE UCR_61S2B 0x0c55 UCR-61S2B /* Diamond products */ product DIAMOND RIO500USB 0x0001 Rio 500 USB /* Dick Smith Electronics (really C-Net) products */ product DICKSMITH RT2573 0x9022 RT2573 product DICKSMITH CWD854F 0x9032 C-Net CWD-854 rev F /* Digi International products */ product DIGI ACCELEPORT2 0x0002 AccelePort USB 2 product DIGI ACCELEPORT4 0x0004 AccelePort USB 4 product DIGI ACCELEPORT8 0x0008 AccelePort USB 8 /* Digianswer A/S products */ product DIGIANSWER ZIGBEE802154 0x000a ZigBee/802.15.4 MAC /* D-Link products */ /*product DLINK DSBS25 0x0100 DSB-S25 serial*/ product DLINK DUBE100 0x1a00 10/100 Ethernet product DLINK DUBE100C1 0x1a02 DUB-E100 rev C1 product DLINK DSB650TX4 0x200c 10/100 Ethernet product DLINK DWL120E 0x3200 DWL-120 rev E product DLINK DWA125D1 0x330f DWA-125 rev D1 product DLINK DWA123D1 0x3310 DWA-123 rev D1 product DLINK DWL122 0x3700 DWL-122 product DLINK DWLG120 0x3701 DWL-G120 product DLINK DWL120F 0x3702 DWL-120 rev F product DLINK DWLAG132 0x3a00 DWL-AG132 product DLINK DWLAG132_NF 0x3a01 DWL-AG132 (no firmware) product DLINK DWLG132 0x3a02 DWL-G132 product DLINK DWLG132_NF 0x3a03 DWL-G132 (no firmware) product DLINK DWLAG122 0x3a04 DWL-AG122 product DLINK DWLAG122_NF 0x3a05 DWL-AG122 (no firmware) product DLINK DWLG122 0x3c00 DWL-G122 b1 Wireless Adapter product DLINK DUBE100B1 0x3c05 DUB-E100 rev B1 product DLINK RT2870 0x3c09 RT2870 product DLINK RT3072 0x3c0a RT3072 product DLINK DWA140B3 0x3c15 DWA-140 rev B3 product DLINK DWA160B2 0x3c1a DWA-160 rev B2 product DLINK DWA127 0x3c1b DWA-127 Wireless Adapter product DLINK DWA162 0x3c1f DWA-162 Wireless Adapter product DLINK DWA140D1 0x3c20 DWA-140 rev D1 product DLINK DSB650C 0x4000 10Mbps Ethernet product DLINK DSB650TX1 0x4001 10/100 Ethernet product DLINK DSB650TX 0x4002 10/100 Ethernet product DLINK DSB650TX_PNA 0x4003 1/10/100 Ethernet product DLINK DSB650TX3 0x400b 10/100 Ethernet product DLINK DSB650TX2 0x4102 10/100 Ethernet product DLINK DUB1312 0x4a00 10/100/1000 Ethernet product DLINK DSB650 0xabc1 10/100 Ethernet product DLINK DUBH7 0xf103 DUB-H7 USB 2.0 7-Port Hub product DLINK DWR510_CD 0xa805 DWR-510 CD-ROM Mode product DLINK DWR510 0x7e12 DWR-510 product DLINK RTL8188CU 0x3308 RTL8188CU product DLINK RTL8192CU_1 0x3307 RTL8192CU product DLINK RTL8192CU_2 0x3309 RTL8192CU product DLINK RTL8192CU_3 0x330a RTL8192CU product DLINK DWA131B 0x330d DWA-131 rev B product DLINK2 RTL8192SU_1 0x3300 RTL8192SU product DLINK2 RTL8192SU_2 0x3302 RTL8192SU product DLINK2 DWA131A1 0x3303 DWA-131 A1 product DLINK2 DWA120 0x3a0c DWA-120 product DLINK2 DWA120_NF 0x3a0d DWA-120 (no firmware) product DLINK2 DWLG122C1 0x3c03 DWL-G122 c1 product DLINK2 WUA1340 0x3c04 WUA-1340 product DLINK2 DWA111 0x3c06 DWA-111 product DLINK2 RT2870_1 0x3c09 RT2870 product DLINK2 DWA110 0x3c07 DWA-110 product DLINK2 RT3072 0x3c0a RT3072 product DLINK2 RT3072_1 0x3c0b RT3072 product DLINK2 RT3070_1 0x3c0d RT3070 product DLINK2 RT3070_2 0x3c0e RT3070 product DLINK2 RT3070_3 0x3c0f RT3070 product DLINK2 RT2870_2 0x3c11 RT2870 product DLINK2 DWA130 0x3c13 DWA-130 product DLINK2 RT3070_4 0x3c15 RT3070 product DLINK2 RT3070_5 0x3c16 RT3070 product DLINK3 DWM652 0x3e04 DWM-652 /* DMI products */ product DMI CFSM_RW 0xa109 CF/SM Reader/Writer product DMI DISK 0x2bcf Generic Disk /* DrayTek products */ product DRAYTEK VIGOR550 0x0550 Vigor550 /* Dream Link products */ product DREAMLINK DL100B 0x0004 USB Webmail Notifier /* dresden elektronik products */ product DRESDENELEKTRONIK SENSORTERMINALBOARD 0x0001 SensorTerminalBoard product DRESDENELEKTRONIK WIRELESSHANDHELDTERMINAL 0x0004 Wireless Handheld Terminal product DRESDENELEKTRONIK DE_RFNODE 0x001c deRFnode product DRESDENELEKTRONIK LEVELSHIFTERSTICKLOWCOST 0x0022 Levelshifter Stick Low Cost /* Dynastream Innovations */ product DYNASTREAM ANTDEVBOARD 0x1003 ANT dev board product DYNASTREAM ANT2USB 0x1004 ANT2USB product DYNASTREAM ANTDEVBOARD2 0x1006 ANT dev board /* Edimax products */ product EDIMAX EW7318USG 0x7318 USB Wireless dongle product EDIMAX RTL8192SU_1 0x7611 RTL8192SU product EDIMAX RTL8192SU_2 0x7612 RTL8192SU product EDIMAX EW7622UMN 0x7622 EW-7622UMn product EDIMAX RT2870_1 0x7711 RT2870 product EDIMAX EW7717 0x7717 EW-7717 product EDIMAX EW7718 0x7718 EW-7718 product EDIMAX EW7733UND 0x7733 EW-7733UnD product EDIMAX EW7811UN 0x7811 EW-7811Un product EDIMAX RTL8192CU 0x7822 RTL8192CU /* eGalax Products */ product EGALAX TPANEL 0x0001 Touch Panel product EGALAX TPANEL2 0x0002 Touch Panel product EGALAX2 TPANEL 0x0001 Touch Panel /* EGO Products */ product EGO DUMMY 0x0000 Dummy Product product EGO M4U 0x1020 ESI M4U /* Eicon Networks */ product EICON DIVA852 0x4905 Diva 852 ISDN TA /* EIZO products */ product EIZO HUB 0x0000 hub product EIZO MONITOR 0x0001 monitor /* ELCON Systemtechnik products */ product ELCON PLAN 0x0002 Goldpfeil P-LAN /* Elecom products */ product ELECOM MOUSE29UO 0x0002 mouse 29UO product ELECOM LDUSBTX0 0x200c LD-USB/TX product ELECOM LDUSBTX1 0x4002 LD-USB/TX product ELECOM LDUSBLTX 0x4005 LD-USBL/TX product ELECOM WDC150SU2M 0x4008 WDC-150SU2M product ELECOM LDUSBTX2 0x400b LD-USB/TX product ELECOM LDUSB20 0x4010 LD-USB20 product ELECOM UCSGT 0x5003 UC-SGT product ELECOM UCSGT0 0x5004 UC-SGT product ELECOM LDUSBTX3 0xabc1 LD-USB/TX /* Elektor products */ product ELEKTOR FT323R 0x0005 FTDI compatible adapter /* Elsa products */ product ELSA MODEM1 0x2265 ELSA Modem Board product ELSA USB2ETHERNET 0x3000 Microlink USB2Ethernet /* ELV products */ product ELV USBI2C 0xe00f USB-I2C interface /* EMS products */ product EMS DUAL_SHOOTER 0x0003 PSX gun controller converter /* Emtec products */ product EMTEC RUF2PS 0x2240 Flash Drive /* Encore products */ product ENCORE RT3070_1 0x1480 RT3070 product ENCORE RT3070_2 0x14a1 RT3070 product ENCORE RT3070_3 0x14a9 RT3070 /* Entrega products */ product ENTREGA 1S 0x0001 1S serial product ENTREGA 2S 0x0002 2S serial product ENTREGA 1S25 0x0003 1S25 serial product ENTREGA 4S 0x0004 4S serial product ENTREGA E45 0x0005 E45 Ethernet product ENTREGA CENTRONICS 0x0006 Parallel Port product ENTREGA XX1 0x0008 Ethernet product ENTREGA 1S9 0x0093 1S9 serial product ENTREGA EZUSB 0x8000 EZ-USB /*product ENTREGA SERIAL 0x8001 DB25 Serial*/ product ENTREGA 2U4S 0x8004 2U4S serial/usb hub product ENTREGA XX2 0x8005 Ethernet /*product ENTREGA SERIAL_DB9 0x8093 DB9 Serial*/ /* Epson products */ product EPSON PRINTER1 0x0001 USB Printer product EPSON PRINTER2 0x0002 ISD USB Smart Cable for Mac product EPSON PRINTER3 0x0003 ISD USB Smart Cable product EPSON PRINTER5 0x0005 USB Printer product EPSON 636 0x0101 Perfection 636U / 636Photo scanner product EPSON 610 0x0103 Perfection 610 scanner product EPSON 1200 0x0104 Perfection 1200U / 1200Photo scanner product EPSON 1600 0x0107 Expression 1600 scanner product EPSON 1640 0x010a Perfection 1640SU scanner product EPSON 1240 0x010b Perfection 1240U / 1240Photo scanner product EPSON 640U 0x010c Perfection 640U scanner product EPSON 1250 0x010f Perfection 1250U / 1250Photo scanner product EPSON 1650 0x0110 Perfection 1650 scanner product EPSON GT9700F 0x0112 GT-9700F scanner product EPSON GT9300UF 0x011b GT-9300UF scanner product EPSON 3200 0x011c Perfection 3200 scanner product EPSON 1260 0x011d Perfection 1260 scanner product EPSON 1660 0x011e Perfection 1660 scanner product EPSON 1670 0x011f Perfection 1670 scanner product EPSON 1270 0x0120 Perfection 1270 scanner product EPSON 2480 0x0121 Perfection 2480 scanner product EPSON 3590 0x0122 Perfection 3590 scanner product EPSON 4990 0x012a Perfection 4990 Photo scanner product EPSON CRESSI_EDY 0x0521 Cressi Edy diving computer product EPSON N2ITION3 0x0522 Zeagle N2iTion3 diving computer product EPSON STYLUS_875DC 0x0601 Stylus Photo 875DC Card Reader product EPSON STYLUS_895 0x0602 Stylus Photo 895 Card Reader product EPSON CX5400 0x0808 CX5400 scanner product EPSON 3500 0x080e CX-3500/3600/3650 MFP product EPSON RX425 0x080f Stylus Photo RX425 scanner product EPSON DX3800 0x0818 CX3700/CX3800/DX38x0 MFP scanner product EPSON 4800 0x0819 CX4700/CX4800/DX48x0 MFP scanner product EPSON 4200 0x0820 CX4100/CX4200/DX4200 MFP scanner product EPSON 5000 0x082b CX4900/CX5000/DX50x0 MFP scanner product EPSON 6000 0x082e CX5900/CX6000/DX60x0 MFP scanner product EPSON DX4000 0x082f DX4000 MFP scanner product EPSON DX7400 0x0838 CX7300/CX7400/DX7400 MFP scanner product EPSON DX8400 0x0839 CX8300/CX8400/DX8400 MFP scanner product EPSON SX100 0x0841 SX100/NX100 MFP scanner product EPSON NX300 0x0848 NX300 MFP scanner product EPSON SX200 0x0849 SX200/SX205 MFP scanner product EPSON SX400 0x084a SX400/NX400/TX400 MFP scanner /* e-TEK Labs products */ product ETEK 1COM 0x8007 Serial /* Evolution products */ product EVOLUTION ER1 0x0300 FTDI compatible adapter product EVOLUTION HYBRID 0x0302 FTDI compatible adapter product EVOLUTION RCM4 0x0303 FTDI compatible adapter /* Extended Systems products */ product EXTENDED XTNDACCESS 0x0100 XTNDAccess IrDA /* Falcom products */ product FALCOM TWIST 0x0001 USB GSM/GPRS Modem product FALCOM SAMBA 0x0005 FTDI compatible adapter /* FEIYA products */ product FEIYA DUMMY 0x0000 Dummy product product FEIYA 5IN1 0x1132 5-in-1 Card Reader product FEIYA ELANGO 0x6200 MicroSDHC Card Reader product FEIYA AC110 0x6300 AC-110 Card Reader /* FeiXun Communication products */ product FEIXUN RTL8188CU 0x0090 RTL8188CU product FEIXUN RTL8192CU 0x0091 RTL8192CU /* Festo */ product FESTO CPX_USB 0x0102 CPX-USB product FESTO CMSP 0x0501 CMSP /* Fiberline */ product FIBERLINE WL430U 0x6003 WL-430U /* FIC / OpenMoko */ product FIC NEO1973_DEBUG 0x5118 FTDI compatible adapter /* Fossil, Inc products */ product FOSSIL WRISTPDA 0x0002 Wrist PDA /* Foxconn products */ product FOXCONN TCOM_TC_300 0xe000 T-Com TC 300 product FOXCONN PIRELLI_DP_L10 0xe003 Pirelli DP-L10 /* Freecom products */ product FREECOM DVD 0xfc01 DVD drive product FREECOM HDD 0xfc05 Classic SL Hard Drive /* Fujitsu Siemens Computers products */ product FSC E5400 0x1009 PrismGT USB 2.0 WLAN /* Future Technology Devices products */ product FTDI SERIAL_8U100AX 0x8372 8U100AX Serial product FTDI SERIAL_8U232AM 0x6001 8U232AM Serial product FTDI SERIAL_8U232AM4 0x6004 8U232AM Serial product FTDI SERIAL_232RL 0x6006 FT232RL Serial product FTDI SERIAL_2232C 0x6010 FT2232C Dual port Serial product FTDI 232H 0x6014 FTDI compatible adapter product FTDI 232EX 0x6015 FTDI compatible adapter product FTDI SERIAL_2232D 0x9e90 FT2232D Dual port Serial product FTDI SERIAL_4232H 0x6011 FT4232H Quad port Serial product FTDI XDS100V2 0xa6d0 TI XDS100V1/V2 and early Beaglebones product FTDI XDS100V3 0xa6d1 TI XDS100V3 product FTDI KTLINK 0xbbe2 KT-LINK Embedded Hackers Multitool product FTDI TURTELIZER2 0xbdc8 egnite Turtelizer 2 JTAG/RS232 Adapter /* Gude Analog- und Digitalsysteme products also uses FTDI's id: */ product FTDI TACTRIX_OPENPORT_13M 0xcc48 OpenPort 1.3 Mitsubishi product FTDI TACTRIX_OPENPORT_13S 0xcc49 OpenPort 1.3 Subaru product FTDI TACTRIX_OPENPORT_13U 0xcc4a OpenPort 1.3 Universal product FTDI GAMMASCOUT 0xd678 Gamma-Scout product FTDI KBS 0xe6c8 Pyramid KBS USB LCD product FTDI EISCOU 0xe888 Expert ISDN Control USB product FTDI UOPTBR 0xe889 USB-RS232 OptoBridge product FTDI EMCU2D 0xe88a Expert mouseCLOCK USB II product FTDI PCMSFU 0xe88b Precision Clock MSF USB product FTDI EMCU2H 0xe88c Expert mouseCLOCK USB II HBG product FTDI MAXSTREAM 0xee18 Maxstream PKG-U product FTDI USB_UIRT 0xf850 USB-UIRT product FTDI USBSERIAL 0xfa00 Matrix Orbital USB Serial product FTDI MX2_3 0xfa01 Matrix Orbital MX2 or MX3 product FTDI MX4_5 0xfa02 Matrix Orbital MX4 or MX5 product FTDI LK202 0xfa03 Matrix Orbital VK/LK202 Family product FTDI LK204 0xfa04 Matrix Orbital VK/LK204 Family product FTDI CFA_632 0xfc08 Crystalfontz CFA-632 USB LCD product FTDI CFA_634 0xfc09 Crystalfontz CFA-634 USB LCD product FTDI CFA_633 0xfc0b Crystalfontz CFA-633 USB LCD product FTDI CFA_631 0xfc0c Crystalfontz CFA-631 USB LCD product FTDI CFA_635 0xfc0d Crystalfontz CFA-635 USB LCD product FTDI SEMC_DSS20 0xfc82 SEMC DSS-20 SyncStation /* Commerzielle und Technische Informationssysteme GmbH products */ product FTDI CTI_USB_NANO_485 0xf60b CTI USB-Nano 485 product FTDI CTI_USB_MINI_485 0xf608 CTI USB-Mini 485 /* Other products */ product FTDI 232RL 0xfbfa FTDI compatible adapter product FTDI 4N_GALAXY_DE_1 0xf3c0 FTDI compatible adapter product FTDI 4N_GALAXY_DE_2 0xf3c1 FTDI compatible adapter product FTDI 4N_GALAXY_DE_3 0xf3c2 FTDI compatible adapter product FTDI 8U232AM_ALT 0x6006 FTDI compatible adapter product FTDI ACCESSO 0xfad0 FTDI compatible adapter product FTDI ACG_HFDUAL 0xdd20 FTDI compatible adapter product FTDI ACTIVE_ROBOTS 0xe548 FTDI compatible adapter product FTDI ACTZWAVE 0xf2d0 FTDI compatible adapter product FTDI AMC232 0xff00 FTDI compatible adapter product FTDI ARTEMIS 0xdf28 FTDI compatible adapter product FTDI ASK_RDR400 0xc991 FTDI compatible adapter product FTDI ATIK_ATK16 0xdf30 FTDI compatible adapter product FTDI ATIK_ATK16C 0xdf32 FTDI compatible adapter product FTDI ATIK_ATK16HR 0xdf31 FTDI compatible adapter product FTDI ATIK_ATK16HRC 0xdf33 FTDI compatible adapter product FTDI ATIK_ATK16IC 0xdf35 FTDI compatible adapter product FTDI BCS_SE923 0xfb99 FTDI compatible adapter product FTDI CANDAPTER 0x9f80 FTDI compatible adapter product FTDI CANUSB 0xffa8 FTDI compatible adapter product FTDI CCSICDU20_0 0xf9d0 FTDI compatible adapter product FTDI CCSICDU40_1 0xf9d1 FTDI compatible adapter product FTDI CCSICDU64_4 0xf9d4 FTDI compatible adapter product FTDI CCSLOAD_N_GO_3 0xf9d3 FTDI compatible adapter product FTDI CCSMACHX_2 0xf9d2 FTDI compatible adapter product FTDI CCSPRIME8_5 0xf9d5 FTDI compatible adapter product FTDI CHAMSYS_24_MASTER_WING 0xdaf8 FTDI compatible adapter product FTDI CHAMSYS_MAXI_WING 0xdafd FTDI compatible adapter product FTDI CHAMSYS_MEDIA_WING 0xdafe FTDI compatible adapter product FTDI CHAMSYS_MIDI_TIMECODE 0xdafb FTDI compatible adapter product FTDI CHAMSYS_MINI_WING 0xdafc FTDI compatible adapter product FTDI CHAMSYS_PC_WING 0xdaf9 FTDI compatible adapter product FTDI CHAMSYS_USB_DMX 0xdafa FTDI compatible adapter product FTDI CHAMSYS_WING 0xdaff FTDI compatible adapter product FTDI COM4SM 0xd578 FTDI compatible adapter product FTDI CONVERTER_0 0xd388 FTDI compatible adapter product FTDI CONVERTER_1 0xd389 FTDI compatible adapter product FTDI CONVERTER_2 0xd38a FTDI compatible adapter product FTDI CONVERTER_3 0xd38b FTDI compatible adapter product FTDI CONVERTER_4 0xd38c FTDI compatible adapter product FTDI CONVERTER_5 0xd38d FTDI compatible adapter product FTDI CONVERTER_6 0xd38e FTDI compatible adapter product FTDI CONVERTER_7 0xd38f FTDI compatible adapter product FTDI DMX4ALL 0xc850 FTDI compatible adapter product FTDI DOMINTELL_DGQG 0xef50 FTDI compatible adapter product FTDI DOMINTELL_DUSB 0xef51 FTDI compatible adapter product FTDI DOTEC 0x9868 FTDI compatible adapter product FTDI ECLO_COM_1WIRE 0xea90 FTDI compatible adapter product FTDI ECO_PRO_CDS 0xe520 FTDI compatible adapter product FTDI ELSTER_UNICOM 0xe700 FTDI compatible adapter product FTDI ELV_ALC8500 0xf06e FTDI compatible adapter product FTDI ELV_CLI7000 0xfb59 FTDI compatible adapter product FTDI ELV_CSI8 0xe0f0 FTDI compatible adapter product FTDI ELV_EC3000 0xe006 FTDI compatible adapter product FTDI ELV_EM1000DL 0xe0f1 FTDI compatible adapter product FTDI ELV_EM1010PC 0xe0ef FTDI compatible adapter product FTDI ELV_FEM 0xe00a FTDI compatible adapter product FTDI ELV_FHZ1000PC 0xf06f FTDI compatible adapter product FTDI ELV_FHZ1300PC 0xe0e8 FTDI compatible adapter product FTDI ELV_FM3RX 0xe0ed FTDI compatible adapter product FTDI ELV_FS20SIG 0xe0f4 FTDI compatible adapter product FTDI ELV_HS485 0xe0ea FTDI compatible adapter product FTDI ELV_KL100 0xe002 FTDI compatible adapter product FTDI ELV_MSM1 0xe001 FTDI compatible adapter product FTDI ELV_PCD200 0xf06c FTDI compatible adapter product FTDI ELV_PCK100 0xe0f2 FTDI compatible adapter product FTDI ELV_PPS7330 0xfb5c FTDI compatible adapter product FTDI ELV_RFP500 0xe0f3 FTDI compatible adapter product FTDI ELV_T1100 0xf06b FTDI compatible adapter product FTDI ELV_TFD128 0xe0ec FTDI compatible adapter product FTDI ELV_TFM100 0xfb5d FTDI compatible adapter product FTDI ELV_TWS550 0xe009 FTDI compatible adapter product FTDI ELV_UAD8 0xf068 FTDI compatible adapter product FTDI ELV_UDA7 0xf069 FTDI compatible adapter product FTDI ELV_UDF77 0xfb5e FTDI compatible adapter product FTDI ELV_UIO88 0xfb5f FTDI compatible adapter product FTDI ELV_ULA200 0xf06d FTDI compatible adapter product FTDI ELV_UM100 0xfb5a FTDI compatible adapter product FTDI ELV_UMS100 0xe0eb FTDI compatible adapter product FTDI ELV_UO100 0xfb5b FTDI compatible adapter product FTDI ELV_UR100 0xfb58 FTDI compatible adapter product FTDI ELV_USI2 0xf06a FTDI compatible adapter product FTDI ELV_USR 0xe000 FTDI compatible adapter product FTDI ELV_UTP8 0xe0f5 FTDI compatible adapter product FTDI ELV_WS300PC 0xe0f6 FTDI compatible adapter product FTDI ELV_WS444PC 0xe0f7 FTDI compatible adapter product FTDI ELV_WS500 0xe0e9 FTDI compatible adapter product FTDI ELV_WS550 0xe004 FTDI compatible adapter product FTDI ELV_WS777 0xe0ee FTDI compatible adapter product FTDI ELV_WS888 0xe008 FTDI compatible adapter product FTDI FUTURE_0 0xf44a FTDI compatible adapter product FTDI FUTURE_1 0xf44b FTDI compatible adapter product FTDI FUTURE_2 0xf44c FTDI compatible adapter product FTDI GENERIC 0x9378 FTDI compatible adapter product FTDI GUDEADS_E808 0xe808 FTDI compatible adapter product FTDI GUDEADS_E809 0xe809 FTDI compatible adapter product FTDI GUDEADS_E80A 0xe80a FTDI compatible adapter product FTDI GUDEADS_E80B 0xe80b FTDI compatible adapter product FTDI GUDEADS_E80C 0xe80c FTDI compatible adapter product FTDI GUDEADS_E80D 0xe80d FTDI compatible adapter product FTDI GUDEADS_E80E 0xe80e FTDI compatible adapter product FTDI GUDEADS_E80F 0xe80f FTDI compatible adapter product FTDI GUDEADS_E88D 0xe88d FTDI compatible adapter product FTDI GUDEADS_E88E 0xe88e FTDI compatible adapter product FTDI GUDEADS_E88F 0xe88f FTDI compatible adapter product FTDI HD_RADIO 0x937c FTDI compatible adapter product FTDI HO720 0xed72 FTDI compatible adapter product FTDI HO730 0xed73 FTDI compatible adapter product FTDI HO820 0xed74 FTDI compatible adapter product FTDI HO870 0xed71 FTDI compatible adapter product FTDI IBS_APP70 0xff3d FTDI compatible adapter product FTDI IBS_PCMCIA 0xff3a FTDI compatible adapter product FTDI IBS_PEDO 0xff3e FTDI compatible adapter product FTDI IBS_PICPRO 0xff39 FTDI compatible adapter product FTDI IBS_PK1 0xff3b FTDI compatible adapter product FTDI IBS_PROD 0xff3f FTDI compatible adapter product FTDI IBS_RS232MON 0xff3c FTDI compatible adapter product FTDI IBS_US485 0xff38 FTDI compatible adapter product FTDI IPLUS 0xd070 FTDI compatible adapter product FTDI IPLUS2 0xd071 FTDI compatible adapter product FTDI IRTRANS 0xfc60 FTDI compatible adapter product FTDI LENZ_LIUSB 0xd780 FTDI compatible adapter product FTDI LM3S_DEVEL_BOARD 0xbcd8 FTDI compatible adapter product FTDI LM3S_EVAL_BOARD 0xbcd9 FTDI compatible adapter product FTDI LM3S_ICDI_B_BOARD 0xbcda FTDI compatible adapter product FTDI MASTERDEVEL2 0xf449 FTDI compatible adapter product FTDI MHAM_DB9 0xeeed FTDI compatible adapter product FTDI MHAM_IC 0xeeec FTDI compatible adapter product FTDI MHAM_KW 0xeee8 FTDI compatible adapter product FTDI MHAM_RS232 0xeeee FTDI compatible adapter product FTDI MHAM_Y6 0xeeea FTDI compatible adapter product FTDI MHAM_Y8 0xeeeb FTDI compatible adapter product FTDI MHAM_Y9 0xeeef FTDI compatible adapter product FTDI MHAM_YS 0xeee9 FTDI compatible adapter product FTDI MICRO_CHAMELEON 0xcaa0 FTDI compatible adapter product FTDI MTXORB_5 0xfa05 FTDI compatible adapter product FTDI MTXORB_6 0xfa06 FTDI compatible adapter product FTDI NXTCAM 0xabb8 FTDI compatible adapter product FTDI OCEANIC 0xf460 FTDI compatible adapter product FTDI OOCDLINK 0xbaf8 FTDI compatible adapter product FTDI OPENDCC 0xbfd8 FTDI compatible adapter product FTDI OPENDCC_GATEWAY 0xbfdb FTDI compatible adapter product FTDI OPENDCC_GBM 0xbfdc FTDI compatible adapter product FTDI OPENDCC_SNIFFER 0xbfd9 FTDI compatible adapter product FTDI OPENDCC_THROTTLE 0xbfda FTDI compatible adapter product FTDI PCDJ_DAC2 0xfa88 FTDI compatible adapter product FTDI PERLE_ULTRAPORT 0xf0c0 FTDI compatible adapter product FTDI PHI_FISCO 0xe40b FTDI compatible adapter product FTDI PIEGROUP 0xf208 FTDI compatible adapter product FTDI PROPOX_JTAGCABLEII 0xd738 FTDI compatible adapter product FTDI R2000KU_TRUE_RNG 0xfb80 FTDI compatible adapter product FTDI R2X0 0xfc71 FTDI compatible adapter product FTDI RELAIS 0xfa10 FTDI compatible adapter product FTDI REU_TINY 0xed22 FTDI compatible adapter product FTDI RMP200 0xe729 FTDI compatible adapter product FTDI RM_CANVIEW 0xfd60 FTDI compatible adapter product FTDI RRCIRKITS_LOCOBUFFER 0xc7d0 FTDI compatible adapter product FTDI SCIENCESCOPE_HS_LOGBOOK 0xff1d FTDI compatible adapter product FTDI SCIENCESCOPE_LOGBOOKML 0xff18 FTDI compatible adapter product FTDI SCIENCESCOPE_LS_LOGBOOK 0xff1c FTDI compatible adapter product FTDI SCS_DEVICE_0 0xd010 FTDI compatible adapter product FTDI SCS_DEVICE_1 0xd011 FTDI compatible adapter product FTDI SCS_DEVICE_2 0xd012 FTDI compatible adapter product FTDI SCS_DEVICE_3 0xd013 FTDI compatible adapter product FTDI SCS_DEVICE_4 0xd014 FTDI compatible adapter product FTDI SCS_DEVICE_5 0xd015 FTDI compatible adapter product FTDI SCS_DEVICE_6 0xd016 FTDI compatible adapter product FTDI SCS_DEVICE_7 0xd017 FTDI compatible adapter product FTDI SDMUSBQSS 0xf448 FTDI compatible adapter product FTDI SIGNALYZER_SH2 0xbca2 FTDI compatible adapter product FTDI SIGNALYZER_SH4 0xbca4 FTDI compatible adapter product FTDI SIGNALYZER_SLITE 0xbca1 FTDI compatible adapter product FTDI SIGNALYZER_ST 0xbca0 FTDI compatible adapter product FTDI SPECIAL_1 0xfc70 FTDI compatible adapter product FTDI SPECIAL_3 0xfc72 FTDI compatible adapter product FTDI SPECIAL_4 0xfc73 FTDI compatible adapter product FTDI SPROG_II 0xf0c8 FTDI compatible adapter product FTDI SR_RADIO 0x9379 FTDI compatible adapter product FTDI SUUNTO_SPORTS 0xf680 FTDI compatible adapter product FTDI TAVIR_STK500 0xfa33 FTDI compatible adapter product FTDI TERATRONIK_D2XX 0xec89 FTDI compatible adapter product FTDI TERATRONIK_VCP 0xec88 FTDI compatible adapter product FTDI THORLABS 0xfaf0 FTDI compatible adapter product FTDI TNC_X 0xebe0 FTDI compatible adapter product FTDI TTUSB 0xff20 FTDI compatible adapter product FTDI USBX_707 0xf857 FTDI compatible adapter product FTDI USINT_CAT 0xb810 FTDI compatible adapter product FTDI USINT_RS232 0xb812 FTDI compatible adapter product FTDI USINT_WKEY 0xb811 FTDI compatible adapter product FTDI VARDAAN 0xf070 FTDI compatible adapter product FTDI VNHCPCUSB_D 0xfe38 FTDI compatible adapter product FTDI WESTREX_MODEL_777 0xdc00 FTDI compatible adapter product FTDI WESTREX_MODEL_8900F 0xdc01 FTDI compatible adapter product FTDI XF_547 0xfc0a FTDI compatible adapter product FTDI XF_640 0xfc0e FTDI compatible adapter product FTDI XF_642 0xfc0f FTDI compatible adapter product FTDI XM_RADIO 0x937a FTDI compatible adapter product FTDI YEI_SERVOCENTER31 0xe050 FTDI compatible adapter /* Fuji photo products */ product FUJIPHOTO MASS0100 0x0100 Mass Storage /* Fujitsu protducts */ product FUJITSU AH_F401U 0x105b AH-F401U Air H device /* Fujitsu-Siemens protducts */ product FUJITSUSIEMENS SCR 0x0009 Fujitsu-Siemens SCR USB Reader /* Garmin products */ product GARMIN IQUE_3600 0x0004 iQue 3600 /* Gemalto products */ product GEMALTO PROXPU 0x5501 Prox-PU/CU RFID Card Reader /* General Instruments (Motorola) products */ product GENERALINSTMNTS SB5100 0x5100 SURFboard SB5100 Cable modem /* Genesys Logic products */ product GENESYS GL620USB 0x0501 GL620USB Host-Host interface product GENESYS GL650 0x0604 GL650 HUB product GENESYS GL606 0x0606 USB 2.0 HUB product GENESYS GL641USB 0x0700 GL641USB CompactFlash Card Reader product GENESYS GL641USB2IDE_2 0x0701 GL641USB USB-IDE Bridge No 2 product GENESYS GL641USB2IDE 0x0702 GL641USB USB-IDE Bridge product GENESYS GL641USB_2 0x0760 GL641USB 6-in-1 Card Reader /* GIGABYTE products */ product GIGABYTE GN54G 0x8001 GN-54G product GIGABYTE GNBR402W 0x8002 GN-BR402W product GIGABYTE GNWLBM101 0x8003 GN-WLBM101 product GIGABYTE GNWBKG 0x8007 GN-WBKG product GIGABYTE GNWB01GS 0x8008 GN-WB01GS product GIGABYTE GNWI05GS 0x800a GN-WI05GS /* Gigaset products */ product GIGASET WLAN 0x0701 WLAN product GIGASET SMCWUSBTG 0x0710 SMCWUSBT-G product GIGASET SMCWUSBTG_NF 0x0711 SMCWUSBT-G (no firmware) product GIGASET AR5523 0x0712 AR5523 product GIGASET AR5523_NF 0x0713 AR5523 (no firmware) product GIGASET RT2573 0x0722 RT2573 product GIGASET RT3070_1 0x0740 RT3070 product GIGASET RT3070_2 0x0744 RT3070 product GIGABYTE RT2870_1 0x800b RT2870 product GIGABYTE GNWB31N 0x800c GN-WB31N product GIGABYTE GNWB32L 0x800d GN-WB32L /* Global Sun Technology product */ product GLOBALSUN AR5523_1 0x7801 AR5523 product GLOBALSUN AR5523_1_NF 0x7802 AR5523 (no firmware) product GLOBALSUN AR5523_2 0x7811 AR5523 product GLOBALSUN AR5523_2_NF 0x7812 AR5523 (no firmware) /* Globespan products */ product GLOBESPAN PRISM_GT_1 0x2000 PrismGT USB 2.0 WLAN product GLOBESPAN PRISM_GT_2 0x2002 PrismGT USB 2.0 WLAN /* G.Mate, Inc products */ product GMATE YP3X00 0x1001 YP3X00 PDA /* GN Otometrics */ product GNOTOMETRICS USB 0x0010 FTDI compatible adapter /* GoHubs products */ product GOHUBS GOCOM232 0x1001 GoCOM232 Serial /* Good Way Technology products */ product GOODWAY GWUSB2E 0x6200 GWUSB2E product GOODWAY RT2573 0xc019 RT2573 /* Google products */ product GOOGLE NEXUSONE 0x4e11 Nexus One /* Gravis products */ product GRAVIS GAMEPADPRO 0x4001 GamePad Pro /* GREENHOUSE products */ product GREENHOUSE KANA21 0x0001 CF-writer with MP3 /* Griffin Technology */ product GRIFFIN IMATE 0x0405 iMate, ADB Adapter /* Guillemot Corporation */ product GUILLEMOT DALEADER 0xa300 DA Leader product GUILLEMOT HWGUSB254 0xe000 HWGUSB2-54 WLAN product GUILLEMOT HWGUSB254LB 0xe010 HWGUSB2-54-LB product GUILLEMOT HWGUSB254V2AP 0xe020 HWGUSB2-54V2-AP product GUILLEMOT HWNU300 0xe030 HWNU-300 product GUILLEMOT HWNUM300 0xe031 HWNUm-300 product GUILLEMOT HWGUN54 0xe032 HWGUn-54 product GUILLEMOT HWNUP150 0xe033 HWNUP-150 /* Hagiwara products */ product HAGIWARA FGSM 0x0002 FlashGate SmartMedia Card Reader product HAGIWARA FGCF 0x0003 FlashGate CompactFlash Card Reader product HAGIWARA FG 0x0005 FlashGate /* HAL Corporation products */ product HAL IMR001 0x0011 Crossam2+USB IR commander /* Handspring, Inc. */ product HANDSPRING VISOR 0x0100 Handspring Visor product HANDSPRING TREO 0x0200 Handspring Treo product HANDSPRING TREO600 0x0300 Handspring Treo 600 /* Hauppauge Computer Works */ product HAUPPAUGE WINTV_USB_FM 0x4d12 WinTV USB FM product HAUPPAUGE2 NOVAT500 0x9580 NovaT 500Stick /* Hawking Technologies products */ product HAWKING RT2870_1 0x0001 RT2870 product HAWKING RT2870_2 0x0003 RT2870 product HAWKING HWUN2 0x0009 HWUN2 product HAWKING RT3070 0x000b RT3070 product HAWKING RTL8192CU 0x0019 RTL8192CU product HAWKING UF100 0x400c 10/100 USB Ethernet product HAWKING RTL8192SU_1 0x0015 RTL8192SU product HAWKING RTL8192SU_2 0x0016 RTL8192SU /* HID Global GmbH products */ product HIDGLOBAL CM2020 0x0596 Omnikey Cardman 2020 product HIDGLOBAL CM6020 0x1784 Omnikey Cardman 6020 /* Hitachi, Ltd. products */ product HITACHI DVDCAM_DZ_MV100A 0x0004 DVD-CAM DZ-MV100A Camcorder product HITACHI DVDCAM_USB 0x001e DVDCAM USB HS Interface /* HP products */ product HP 895C 0x0004 DeskJet 895C product HP 4100C 0x0101 Scanjet 4100C product HP S20 0x0102 Photosmart S20 product HP 880C 0x0104 DeskJet 880C product HP 4200C 0x0105 ScanJet 4200C product HP CDWRITERPLUS 0x0107 CD-Writer Plus product HP KBDHUB 0x010c Multimedia Keyboard Hub product HP G55XI 0x0111 OfficeJet G55xi product HP HN210W 0x011c HN210W 802.11b WLAN product HP 49GPLUS 0x0121 49g+ graphing calculator product HP 6200C 0x0201 ScanJet 6200C product HP S20b 0x0202 PhotoSmart S20 product HP 815C 0x0204 DeskJet 815C product HP 3300C 0x0205 ScanJet 3300C product HP CDW8200 0x0207 CD-Writer Plus 8200e product HP MMKEYB 0x020c Multimedia keyboard product HP 1220C 0x0212 DeskJet 1220C product HP UN2420_QDL 0x241d UN2420 QDL Firmware Loader product HP UN2420 0x251d UN2420 WWAN/GPS Module product HP 810C 0x0304 DeskJet 810C/812C product HP 4300C 0x0305 Scanjet 4300C product HP CDW4E 0x0307 CD-Writer+ CD-4e product HP G85XI 0x0311 OfficeJet G85xi product HP 1200 0x0317 LaserJet 1200 product HP 5200C 0x0401 Scanjet 5200C product HP 830C 0x0404 DeskJet 830C product HP 3400CSE 0x0405 ScanJet 3400cse product HP 6300C 0x0601 Scanjet 6300C product HP 840C 0x0604 DeskJet 840c product HP 2200C 0x0605 ScanJet 2200C product HP 5300C 0x0701 Scanjet 5300C product HP 4400C 0x0705 Scanjet 4400C product HP 4470C 0x0805 Scanjet 4470C product HP 82x0C 0x0b01 Scanjet 82x0C product HP 2300D 0x0b17 Laserjet 2300d product HP 970CSE 0x1004 Deskjet 970Cse product HP 5400C 0x1005 Scanjet 5400C product HP 2215 0x1016 iPAQ 22xx/Jornada 548 product HP 568J 0x1116 Jornada 568 product HP 930C 0x1204 DeskJet 930c product HP3 RTL8188CU 0x1629 RTL8188CU product HP P2000U 0x1801 Inkjet P-2000U product HP HS2300 0x1e1d HS2300 HSDPA (aka MC8775) product HP 640C 0x2004 DeskJet 640c product HP 4670V 0x3005 ScanJet 4670v product HP P1100 0x3102 Photosmart P1100 product HP LD220 0x3524 LD220 POS Display product HP OJ4215 0x3d11 OfficeJet 4215 product HP HN210E 0x811c Ethernet HN210E product HP2 C500 0x6002 PhotoSmart C500 product HP EV2200 0x1b1d ev2200 HSDPA (aka MC5720) product HP HS2300 0x1e1d hs2300 HSDPA (aka MC8775) /* HTC products */ product HTC WINMOBILE 0x00ce HTC USB Sync product HTC PPC6700MODEM 0x00cf PPC6700 Modem product HTC SMARTPHONE 0x0a51 SmartPhone USB Sync product HTC WIZARD 0x0bce HTC Wizard USB Sync product HTC LEGENDSYNC 0x0c97 HTC Legend USB Sync product HTC LEGEND 0x0ff9 HTC Legend product HTC LEGENDINTERNET 0x0ffe HTC Legend Internet Sharing /* HUAWEI products */ product HUAWEI MOBILE 0x1001 Huawei Mobile product HUAWEI E220 0x1003 HSDPA modem product HUAWEI E220BIS 0x1004 HSDPA modem product HUAWEI E1401 0x1401 3G modem product HUAWEI E1402 0x1402 3G modem product HUAWEI E1403 0x1403 3G modem product HUAWEI E1404 0x1404 3G modem product HUAWEI E1405 0x1405 3G modem product HUAWEI E1406 0x1406 3G modem product HUAWEI E1407 0x1407 3G modem product HUAWEI E1408 0x1408 3G modem product HUAWEI E1409 0x1409 3G modem product HUAWEI E140A 0x140a 3G modem product HUAWEI E140B 0x140b 3G modem product HUAWEI E180V 0x140c E180V product HUAWEI E140D 0x140d 3G modem product HUAWEI E140E 0x140e 3G modem product HUAWEI E140F 0x140f 3G modem product HUAWEI E1410 0x1410 3G modem product HUAWEI E1411 0x1411 3G modem product HUAWEI E1412 0x1412 3G modem product HUAWEI E1413 0x1413 3G modem product HUAWEI E1414 0x1414 3G modem product HUAWEI E1415 0x1415 3G modem product HUAWEI E1416 0x1416 3G modem product HUAWEI E1417 0x1417 3G modem product HUAWEI E1418 0x1418 3G modem product HUAWEI E1419 0x1419 3G modem product HUAWEI E141A 0x141a 3G modem product HUAWEI E141B 0x141b 3G modem product HUAWEI E141C 0x141c 3G modem product HUAWEI E141D 0x141d 3G modem product HUAWEI E141E 0x141e 3G modem product HUAWEI E141F 0x141f 3G modem product HUAWEI E1420 0x1420 3G modem product HUAWEI E1421 0x1421 3G modem product HUAWEI E1422 0x1422 3G modem product HUAWEI E1423 0x1423 3G modem product HUAWEI E1424 0x1424 3G modem product HUAWEI E1425 0x1425 3G modem product HUAWEI E1426 0x1426 3G modem product HUAWEI E1427 0x1427 3G modem product HUAWEI E1428 0x1428 3G modem product HUAWEI E1429 0x1429 3G modem product HUAWEI E142A 0x142a 3G modem product HUAWEI E142B 0x142b 3G modem product HUAWEI E142C 0x142c 3G modem product HUAWEI E142D 0x142d 3G modem product HUAWEI E142E 0x142e 3G modem product HUAWEI E142F 0x142f 3G modem product HUAWEI E1430 0x1430 3G modem product HUAWEI E1431 0x1431 3G modem product HUAWEI E1432 0x1432 3G modem product HUAWEI E1433 0x1433 3G modem product HUAWEI E1434 0x1434 3G modem product HUAWEI E1435 0x1435 3G modem product HUAWEI E1436 0x1436 3G modem product HUAWEI E1437 0x1437 3G modem product HUAWEI E1438 0x1438 3G modem product HUAWEI E1439 0x1439 3G modem product HUAWEI E143A 0x143a 3G modem product HUAWEI E143B 0x143b 3G modem product HUAWEI E143C 0x143c 3G modem product HUAWEI E143D 0x143d 3G modem product HUAWEI E143E 0x143e 3G modem product HUAWEI E143F 0x143f 3G modem product HUAWEI E1752 0x1446 3G modem product HUAWEI K4505 0x1464 3G modem product HUAWEI K3765 0x1465 3G modem product HUAWEI E1820 0x14ac E1820 HSPA+ USB Slider product HUAWEI K3770 0x14c9 3G modem product HUAWEI K3772 0x14cf K3772 product HUAWEI K3770_INIT 0x14d1 K3770 Initial product HUAWEI E3131_INIT 0x14fe 3G modem initial product HUAWEI E392 0x1505 LTE modem product HUAWEI E3131 0x1506 3G modem product HUAWEI K3765_INIT 0x1520 K3765 Initial product HUAWEI K4505_INIT 0x1521 K4505 Initial product HUAWEI K3772_INIT 0x1526 K3772 Initial product HUAWEI E3272_INIT 0x155b LTE modem initial product HUAWEI R215_INIT 0x1582 LTE modem initial product HUAWEI R215 0x1588 LTE modem product HUAWEI ETS2055 0x1803 CDMA modem product HUAWEI E173 0x1c05 3G modem product HUAWEI E173_INIT 0x1c0b 3G modem initial product HUAWEI E3272 0x1c1e LTE modem /* HUAWEI 3com products */ product HUAWEI3COM WUB320G 0x0009 Aolynk WUB320g /* IBM Corporation */ product IBM USBCDROMDRIVE 0x4427 USB CD-ROM Drive /* Icom products */ product ICOM SP1 0x0004 FTDI compatible adapter product ICOM OPC_U_UC 0x0018 FTDI compatible adapter product ICOM RP2C1 0x0009 FTDI compatible adapter product ICOM RP2C2 0x000a FTDI compatible adapter product ICOM RP2D 0x000b FTDI compatible adapter product ICOM RP2KVR 0x0013 FTDI compatible adapter product ICOM RP2KVT 0x0012 FTDI compatible adapter product ICOM RP2VR 0x000d FTDI compatible adapter product ICOM RP2VT 0x000c FTDI compatible adapter product ICOM RP4KVR 0x0011 FTDI compatible adapter product ICOM RP4KVT 0x0010 FTDI compatible adapter /* ID-tech products */ product IDTECH IDT1221U 0x0300 FTDI compatible adapter /* Imagination Technologies products */ product IMAGINATION DBX1 0x2107 DBX1 DSP core /* Initio Corporation products */ product INITIO DUMMY 0x0000 Dummy product product INITIO INIC_1610P 0x1e40 USB to SATA Bridge /* Inside Out Networks products */ product INSIDEOUT EDGEPORT4 0x0001 EdgePort/4 serial ports /* In-System products */ product INSYSTEM F5U002 0x0002 Parallel printer product INSYSTEM ATAPI 0x0031 ATAPI Adapter product INSYSTEM ISD110 0x0200 IDE Adapter ISD110 product INSYSTEM ISD105 0x0202 IDE Adapter ISD105 product INSYSTEM USBCABLE 0x081a USB cable product INSYSTEM STORAGE_V2 0x5701 USB Storage Adapter V2 /* Intel products */ product INTEL EASYPC_CAMERA 0x0110 Easy PC Camera product INTEL TESTBOARD 0x9890 82930 test board product INTEL2 IRMH 0x0020 Integrated Rate Matching Hub product INTEL2 IRMH2 0x0024 Integrated Rate Matching Hub /* Interbiometric products */ product INTERBIOMETRICS IOBOARD 0x1002 FTDI compatible adapter product INTERBIOMETRICS MINI_IOBOARD 0x1006 FTDI compatible adapter /* Intersil products */ product INTERSIL PRISM_GT 0x1000 PrismGT USB 2.0 WLAN product INTERSIL PRISM_2X 0x3642 Prism2.x or Atmel WLAN /* Interpid Control Systems products */ product INTREPIDCS VALUECAN 0x0601 ValueCAN CAN bus interface product INTREPIDCS NEOVI 0x0701 NeoVI Blue vehicle bus interface /* I/O DATA products */ product IODATA IU_CD2 0x0204 DVD Multi-plus unit iU-CD2 product IODATA DVR_UEH8 0x0206 DVD Multi-plus unit DVR-UEH8 product IODATA USBSSMRW 0x0314 USB-SSMRW SD-card product IODATA USBSDRW 0x031e USB-SDRW SD-card product IODATA USBETT 0x0901 USB ETT product IODATA USBETTX 0x0904 USB ETTX product IODATA USBETTXS 0x0913 USB ETTX product IODATA USBWNB11A 0x0919 USB WN-B11 product IODATA USBWNB11 0x0922 USB Airport WN-B11 product IODATA ETGUS2 0x0930 ETG-US2 product IODATA RT3072_1 0x0944 RT3072 product IODATA RT3072_2 0x0945 RT3072 product IODATA RT3072_3 0x0947 RT3072 product IODATA RT3072_4 0x0948 RT3072 product IODATA USBRSAQ 0x0a03 Serial USB-RSAQ1 product IODATA USBRSAQ5 0x0a0e Serial USB-RSAQ5 product IODATA2 USB2SC 0x0a09 USB2.0-SCSI Bridge USB2-SC /* Iomega products */ product IOMEGA ZIP100 0x0001 Zip 100 product IOMEGA ZIP250 0x0030 Zip 250 /* Ionic products */ product IONICS PLUGCOMPUTER 0x0102 FTDI compatible adapter /* Integrated System Solution Corp. products */ product ISSC ISSCBTA 0x1001 Bluetooth USB Adapter /* iTegno products */ product ITEGNO WM1080A 0x1080 WM1080A GSM/GPRS modem product ITEGNO WM2080A 0x2080 WM2080A CDMA modem /* Ituner networks products */ product ITUNERNET USBLCD2X20 0x0002 USB-LCD 2x20 product ITUNERNET USBLCD4X20 0xc001 USB-LCD 4x20 /* Jablotron products */ product JABLOTRON PC60B 0x0001 PC-60B /* Jaton products */ product JATON EDA 0x5704 Ethernet /* Jeti products */ product JETI SPC1201 0x04b2 FTDI compatible adapter /* JMicron products */ product JMICRON JM20336 0x2336 USB to SATA Bridge product JMICRON JM20337 0x2338 USB to ATA/ATAPI Bridge /* JVC products */ product JVC GR_DX95 0x000a GR-DX95 product JVC MP_PRX1 0x3008 MP-PRX1 Ethernet /* JRC products */ product JRC AH_J3001V_J3002V 0x0001 AirH PHONE AH-J3001V/J3002V /* Kamstrrup products */ product KAMSTRUP OPTICALEYE 0x0001 Optical Eye/3-wire product KAMSTRUP MBUS_250D 0x0005 M-Bus Master MultiPort 250D /* Kawatsu products */ product KAWATSU MH4000P 0x0003 MiniHub 4000P /* Keisokugiken Corp. products */ product KEISOKUGIKEN USBDAQ 0x0068 HKS-0200 USBDAQ /* Kensington products */ product KENSINGTON ORBIT 0x1003 Orbit USB/PS2 trackball product KENSINGTON TURBOBALL 0x1005 TurboBall /* Keyspan products */ product KEYSPAN USA28_NF 0x0101 USA-28 serial Adapter (no firmware) product KEYSPAN USA28X_NF 0x0102 USA-28X serial Adapter (no firmware) product KEYSPAN USA19_NF 0x0103 USA-19 serial Adapter (no firmware) product KEYSPAN USA18_NF 0x0104 USA-18 serial Adapter (no firmware) product KEYSPAN USA18X_NF 0x0105 USA-18X serial Adapter (no firmware) product KEYSPAN USA19W_NF 0x0106 USA-19W serial Adapter (no firmware) product KEYSPAN USA19 0x0107 USA-19 serial Adapter product KEYSPAN USA19W 0x0108 USA-19W serial Adapter product KEYSPAN USA49W_NF 0x0109 USA-49W serial Adapter (no firmware) product KEYSPAN USA49W 0x010a USA-49W serial Adapter product KEYSPAN USA19QI_NF 0x010b USA-19QI serial Adapter (no firmware) product KEYSPAN USA19QI 0x010c USA-19QI serial Adapter product KEYSPAN USA19Q_NF 0x010d USA-19Q serial Adapter (no firmware) product KEYSPAN USA19Q 0x010e USA-19Q serial Adapter product KEYSPAN USA28 0x010f USA-28 serial Adapter product KEYSPAN USA28XXB 0x0110 USA-28X/XB serial Adapter product KEYSPAN USA18 0x0111 USA-18 serial Adapter product KEYSPAN USA18X 0x0112 USA-18X serial Adapter product KEYSPAN USA28XB_NF 0x0113 USA-28XB serial Adapter (no firmware) product KEYSPAN USA28XA_NF 0x0114 USA-28XB serial Adapter (no firmware) product KEYSPAN USA28XA 0x0115 USA-28XA serial Adapter product KEYSPAN USA18XA_NF 0x0116 USA-18XA serial Adapter (no firmware) product KEYSPAN USA18XA 0x0117 USA-18XA serial Adapter product KEYSPAN USA19QW_NF 0x0118 USA-19WQ serial Adapter (no firmware) product KEYSPAN USA19QW 0x0119 USA-19WQ serial Adapter product KEYSPAN USA19HA 0x0121 USA-19HS serial Adapter product KEYSPAN UIA10 0x0201 UIA-10 remote control product KEYSPAN UIA11 0x0202 UIA-11 remote control /* Kingston products */ product KINGSTON XX1 0x0008 Ethernet product KINGSTON KNU101TX 0x000a KNU101TX USB Ethernet product KINGSTON HYPERX3_0 0x162b DT HyperX 3.0 /* Kawasaki products */ product KLSI DUH3E10BT 0x0008 USB Ethernet product KLSI DUH3E10BTN 0x0009 USB Ethernet /* Kobil products */ product KOBIL CONV_B1 0x2020 FTDI compatible adapter product KOBIL CONV_KAAN 0x2021 FTDI compatible adapter /* Kodak products */ product KODAK DC220 0x0100 Digital Science DC220 product KODAK DC260 0x0110 Digital Science DC260 product KODAK DC265 0x0111 Digital Science DC265 product KODAK DC290 0x0112 Digital Science DC290 product KODAK DC240 0x0120 Digital Science DC240 product KODAK DC280 0x0130 Digital Science DC280 /* Kontron AG products */ product KONTRON DM9601 0x8101 USB Ethernet product KONTRON JP1082 0x9700 USB Ethernet /* Konica Corp. Products */ product KONICA CAMERA 0x0720 Digital Color Camera /* KYE products */ product KYE NICHE 0x0001 Niche mouse product KYE NETSCROLL 0x0003 Genius NetScroll mouse product KYE FLIGHT2000 0x1004 Flight 2000 joystick product KYE VIVIDPRO 0x2001 ColorPage Vivid-Pro scanner /* Kyocera products */ product KYOCERA FINECAM_S3X 0x0100 Finecam S3x product KYOCERA FINECAM_S4 0x0101 Finecam S4 product KYOCERA FINECAM_S5 0x0103 Finecam S5 product KYOCERA FINECAM_L3 0x0105 Finecam L3 product KYOCERA AHK3001V 0x0203 AH-K3001V product KYOCERA2 CDMA_MSM_K 0x17da Qualcomm Kyocera CDMA Technologies MSM product KYOCERA2 KPC680 0x180a Qualcomm Kyocera CDMA Technologies MSM /* LaCie products */ product LACIE HD 0xa601 Hard Disk product LACIE CDRW 0xa602 CD R/W /* Lake Shore Cryotronics products */ product LAKESHORE 121 0x0100 121 Current Source product LAKESHORE 218A 0x0200 218A Temperature Monitor product LAKESHORE 219 0x0201 219 Temperature Monitor product LAKESHORE 233 0x0202 233 Temperature Transmitter product LAKESHORE 235 0x0203 235 Temperature Transmitter product LAKESHORE 335 0x0300 335 Temperature Controller product LAKESHORE 336 0x0301 336 Temperature Controller product LAKESHORE 350 0x0302 350 Temperature Controller product LAKESHORE 371 0x0303 371 AC Bridge product LAKESHORE 411 0x0400 411 Handheld Gaussmeter product LAKESHORE 425 0x0401 425 Gaussmeter product LAKESHORE 455A 0x0402 455A DSP Gaussmeter product LAKESHORE 475A 0x0403 475A DSP Gaussmeter product LAKESHORE 465 0x0404 465 Gaussmeter product LAKESHORE 625A 0x0600 625A Magnet PSU product LAKESHORE 642A 0x0601 642A Magnet PSU product LAKESHORE 648 0x0602 648 Magnet PSU product LAKESHORE 737 0x0700 737 VSM Controller product LAKESHORE 776 0x0701 776 Matrix Switch /* Larsen and Brusgaard products */ product LARSENBRUSGAARD ALTITRACK 0x0001 FTDI compatible adapter /* Leadtek products */ product LEADTEK 9531 0x2101 9531 GPS /* Lenovo products */ product LENOVO ETHERNET 0x7203 USB 2.0 Ethernet /* Lexar products */ product LEXAR JUMPSHOT 0x0001 jumpSHOT CompactFlash Reader product LEXAR CF_READER 0xb002 USB CF Reader product LEXAR JUMPDRIVE 0xa833 USB Jumpdrive Flash Drive /* Lexmark products */ product LEXMARK S2450 0x0009 Optra S 2450 /* Liebert products */ product LIEBERT POWERSURE_PXT 0xffff PowerSure Personal XT product LIEBERT2 PSI1000 0x0004 UPS PSI 1000 FW:08 /* Link Instruments Inc. products */ product LINKINSTRUMENTS MSO19 0xf190 Link Instruments MSO-19 product LINKINSTRUMENTS MSO28 0xf280 Link Instruments MSO-28 product LINKINSTRUMENTS MSO28_2 0xf281 Link Instruments MSO-28 /* Linksys products */ product LINKSYS MAUSB2 0x0105 Camedia MAUSB-2 product LINKSYS USB10TX1 0x200c USB10TX product LINKSYS USB10T 0x2202 USB10T Ethernet product LINKSYS USB100TX 0x2203 USB100TX Ethernet product LINKSYS USB100H1 0x2204 USB100H1 Ethernet/HPNA product LINKSYS USB10TA 0x2206 USB10TA Ethernet product LINKSYS USB10TX2 0x400b USB10TX product LINKSYS2 WUSB11 0x2219 WUSB11 Wireless Adapter product LINKSYS2 USB200M 0x2226 USB 2.0 10/100 Ethernet product LINKSYS3 WUSB11v28 0x2233 WUSB11 v2.8 Wireless Adapter product LINKSYS4 USB1000 0x0039 USB1000 product LINKSYS4 WUSB100 0x0070 WUSB100 product LINKSYS4 WUSB600N 0x0071 WUSB600N product LINKSYS4 WUSB54GCV2 0x0073 WUSB54GC v2 product LINKSYS4 WUSB54GCV3 0x0077 WUSB54GC v3 product LINKSYS4 RT3070 0x0078 RT3070 product LINKSYS4 WUSB600NV2 0x0079 WUSB600N v2 /* Logilink products */ product LOGILINK DUMMY 0x0000 Dummy product product LOGILINK U2M 0x0101 LogiLink USB MIDI Cable /* Logitech products */ product LOGITECH M2452 0x0203 M2452 keyboard product LOGITECH M4848 0x0301 M4848 mouse product LOGITECH PAGESCAN 0x040f PageScan product LOGITECH QUICKCAMWEB 0x0801 QuickCam Web product LOGITECH QUICKCAMPRO 0x0810 QuickCam Pro product LOGITECH WEBCAMC100 0X0817 Webcam C100 product LOGITECH QUICKCAMEXP 0x0840 QuickCam Express product LOGITECH QUICKCAM 0x0850 QuickCam product LOGITECH QUICKCAMPRO3 0x0990 QuickCam Pro 9000 product LOGITECH N43 0xc000 N43 product LOGITECH N48 0xc001 N48 mouse product LOGITECH MBA47 0xc002 M-BA47 mouse product LOGITECH WMMOUSE 0xc004 WingMan Gaming Mouse product LOGITECH BD58 0xc00c BD58 mouse product LOGITECH UN58A 0xc030 iFeel Mouse product LOGITECH UN53B 0xc032 iFeel MouseMan product LOGITECH WMPAD 0xc208 WingMan GamePad Extreme product LOGITECH WMRPAD 0xc20a WingMan RumblePad product LOGITECH WMJOY 0xc281 WingMan Force joystick product LOGITECH BB13 0xc401 USB-PS/2 Trackball product LOGITECH RK53 0xc501 Cordless mouse product LOGITECH RB6 0xc503 Cordless keyboard product LOGITECH MX700 0xc506 Cordless optical mouse product LOGITECH QUICKCAMPRO2 0xd001 QuickCam Pro /* Logitec Corp. products */ product LOGITEC LDR_H443SU2 0x0033 DVD Multi-plus unit LDR-H443SU2 product LOGITEC LDR_H443U2 0x00b3 DVD Multi-plus unit LDR-H443U2 product LOGITEC LAN_GTJU2A 0x0160 LAN-GTJ/U2A Ethernet product LOGITEC RT2870_1 0x0162 RT2870 product LOGITEC RT2870_2 0x0163 RT2870 product LOGITEC RT2870_3 0x0164 RT2870 product LOGITEC LANW300NU2 0x0166 LAN-W300N/U2 product LOGITEC LANW150NU2 0x0168 LAN-W150N/U2 product LOGITEC LANW300NU2S 0x0169 LAN-W300N/U2S /* Longcheer Holdings, Ltd. products */ product LONGCHEER WM66 0x6061 Longcheer WM66 HSDPA product LONGCHEER W14 0x9603 Mobilcom W14 product LONGCHEER DISK 0xf000 Driver disk product LONGCHEER XSSTICK 0x9605 4G Systems XSStick P14 /* Lucent products */ product LUCENT EVALKIT 0x1001 USS-720 evaluation kit /* Luwen products */ product LUWEN EASYDISK 0x0005 EasyDisc /* Macally products */ product MACALLY MOUSE1 0x0101 mouse /* Mag-Tek products */ product MAGTEK USBSWIPE 0x0002 USB Mag Stripe Swipe Reader /* Marvell Technology Group, Ltd. products */ product MARVELL SHEEVAPLUG 0x9e8f SheevaPlug serial interface /* Matrix Orbital products */ product MATRIXORBITAL FTDI_RANGE_0100 0x0100 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0101 0x0101 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0102 0x0102 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0103 0x0103 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0104 0x0104 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0105 0x0105 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0106 0x0106 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0107 0x0107 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0108 0x0108 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0109 0x0109 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_010A 0x010a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_010B 0x010b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_010C 0x010c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_010D 0x010d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_010E 0x010e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_010F 0x010f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0110 0x0110 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0111 0x0111 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0112 0x0112 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0113 0x0113 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0114 0x0114 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0115 0x0115 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0116 0x0116 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0117 0x0117 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0118 0x0118 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0119 0x0119 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_011A 0x011a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_011B 0x011b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_011C 0x011c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_011D 0x011d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_011E 0x011e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_011F 0x011f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0120 0x0120 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0121 0x0121 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0122 0x0122 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0123 0x0123 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0124 0x0124 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0125 0x0125 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0126 0x0126 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0128 0x0128 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0129 0x0129 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_012A 0x012a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_012B 0x012b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_012D 0x012d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_012E 0x012e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_012F 0x012f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0130 0x0130 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0131 0x0131 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0132 0x0132 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0133 0x0133 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0134 0x0134 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0135 0x0135 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0136 0x0136 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0137 0x0137 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0138 0x0138 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0139 0x0139 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_013A 0x013a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_013B 0x013b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_013C 0x013c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_013D 0x013d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_013E 0x013e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_013F 0x013f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0140 0x0140 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0141 0x0141 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0142 0x0142 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0143 0x0143 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0144 0x0144 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0145 0x0145 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0146 0x0146 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0147 0x0147 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0148 0x0148 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0149 0x0149 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_014A 0x014a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_014B 0x014b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_014C 0x014c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_014D 0x014d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_014E 0x014e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_014F 0x014f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0150 0x0150 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0151 0x0151 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0152 0x0152 FTDI compatible adapter product MATRIXORBITAL MOUA 0x0153 Martrix Orbital MOU-Axxxx LCD displays product MATRIXORBITAL FTDI_RANGE_0159 0x0159 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_015A 0x015a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_015B 0x015b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_015C 0x015c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_015D 0x015d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_015E 0x015e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_015F 0x015f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0160 0x0160 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0161 0x0161 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0162 0x0162 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0163 0x0163 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0164 0x0164 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0165 0x0165 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0166 0x0166 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0167 0x0167 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0168 0x0168 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0169 0x0169 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_016A 0x016a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_016B 0x016b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_016C 0x016c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_016D 0x016d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_016E 0x016e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_016F 0x016f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0170 0x0170 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0171 0x0171 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0172 0x0172 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0173 0x0173 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0174 0x0174 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0175 0x0175 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0176 0x0176 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0177 0x0177 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0178 0x0178 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0179 0x0179 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_017A 0x017a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_017B 0x017b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_017C 0x017c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_017D 0x017d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_017E 0x017e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_017F 0x017f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0180 0x0180 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0181 0x0181 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0182 0x0182 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0183 0x0183 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0184 0x0184 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0185 0x0185 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0186 0x0186 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0187 0x0187 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0188 0x0188 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0189 0x0189 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_018A 0x018a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_018B 0x018b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_018C 0x018c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_018D 0x018d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_018E 0x018e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_018F 0x018f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0190 0x0190 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0191 0x0191 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0192 0x0192 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0193 0x0193 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0194 0x0194 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0195 0x0195 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0196 0x0196 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0197 0x0197 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0198 0x0198 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_0199 0x0199 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_019A 0x019a FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_019B 0x019b FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_019C 0x019c FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_019D 0x019d FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_019E 0x019e FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_019F 0x019f FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A0 0x01a0 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A1 0x01a1 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A2 0x01a2 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A3 0x01a3 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A4 0x01a4 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A5 0x01a5 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A6 0x01a6 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A7 0x01a7 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A8 0x01a8 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01A9 0x01a9 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01AA 0x01aa FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01AB 0x01ab FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01AC 0x01ac FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01AD 0x01ad FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01AE 0x01ae FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01AF 0x01af FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B0 0x01b0 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B1 0x01b1 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B2 0x01b2 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B3 0x01b3 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B4 0x01b4 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B5 0x01b5 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B6 0x01b6 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B7 0x01b7 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B8 0x01b8 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01B9 0x01b9 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01BA 0x01ba FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01BB 0x01bb FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01BC 0x01bc FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01BD 0x01bd FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01BE 0x01be FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01BF 0x01bf FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C0 0x01c0 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C1 0x01c1 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C2 0x01c2 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C3 0x01c3 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C4 0x01c4 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C5 0x01c5 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C6 0x01c6 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C7 0x01c7 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C8 0x01c8 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01C9 0x01c9 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01CA 0x01ca FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01CB 0x01cb FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01CC 0x01cc FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01CD 0x01cd FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01CE 0x01ce FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01CF 0x01cf FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D0 0x01d0 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D1 0x01d1 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D2 0x01d2 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D3 0x01d3 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D4 0x01d4 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D5 0x01d5 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D6 0x01d6 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D7 0x01d7 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D8 0x01d8 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01D9 0x01d9 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01DA 0x01da FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01DB 0x01db FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01DC 0x01dc FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01DD 0x01dd FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01DE 0x01de FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01DF 0x01df FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E0 0x01e0 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E1 0x01e1 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E2 0x01e2 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E3 0x01e3 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E4 0x01e4 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E5 0x01e5 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E6 0x01e6 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E7 0x01e7 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E8 0x01e8 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01E9 0x01e9 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01EA 0x01ea FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01EB 0x01eb FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01EC 0x01ec FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01ED 0x01ed FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01EE 0x01ee FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01EF 0x01ef FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F0 0x01f0 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F1 0x01f1 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F2 0x01f2 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F3 0x01f3 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F4 0x01f4 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F5 0x01f5 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F6 0x01f6 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F7 0x01f7 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F8 0x01f8 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01F9 0x01f9 FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01FA 0x01fa FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01FB 0x01fb FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01FC 0x01fc FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01FD 0x01fd FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01FE 0x01fe FTDI compatible adapter product MATRIXORBITAL FTDI_RANGE_01FF 0x01ff FTDI compatible adapter /* MCT Corp. */ product MCT HUB0100 0x0100 Hub product MCT DU_H3SP_USB232 0x0200 D-Link DU-H3SP USB BAY Hub product MCT USB232 0x0210 USB-232 Interface product MCT SITECOM_USB232 0x0230 Sitecom USB-232 Products /* Medeli */ product MEDELI DD305 0x5011 DD305 Digital Drum Set /* MediaTek, Inc. */ product MEDIATEK MTK3329 0x3329 MTK II GPS Receiver /* Meizu Electronics */ product MEIZU M6_SL 0x0140 MiniPlayer M6 (SL) /* Melco, Inc products */ product MELCO LUATX1 0x0001 LUA-TX Ethernet product MELCO LUATX5 0x0005 LUA-TX Ethernet product MELCO LUA2TX5 0x0009 LUA2-TX Ethernet product MELCO LUAKTX 0x0012 LUA-KTX Ethernet product MELCO DUBPXXG 0x001c DUB-PxxG product MELCO LUAU2KTX 0x003d LUA-U2-KTX Ethernet product MELCO KG54YB 0x005e WLI-U2-KG54-YB WLAN product MELCO KG54 0x0066 WLI-U2-KG54 WLAN product MELCO KG54AI 0x0067 WLI-U2-KG54-AI WLAN product MELCO LUA3U2AGT 0x006e LUA3-U2-AGT product MELCO NINWIFI 0x008b Nintendo Wi-Fi product MELCO PCOPRS1 0x00b3 PC-OP-RS1 RemoteStation product MELCO SG54HP 0x00d8 WLI-U2-SG54HP product MELCO G54HP 0x00d9 WLI-U2-G54HP product MELCO KG54L 0x00da WLI-U2-KG54L product MELCO WLIUCG300N 0x00e8 WLI-UC-G300N product MELCO SG54HG 0x00f4 WLI-U2-SG54HG product MELCO WLRUCG 0x0116 WLR-UC-G product MELCO WLRUCGAOSS 0x0119 WLR-UC-G-AOSS product MELCO WLIUCAG300N 0x012e WLI-UC-AG300N product MELCO WLIUCG 0x0137 WLI-UC-G product MELCO RT2870_1 0x0148 RT2870 product MELCO RT2870_2 0x0150 RT2870 product MELCO WLIUCGN 0x015d WLI-UC-GN product MELCO WLIUCG301N 0x016f WLI-UC-G301N product MELCO WLIUCGNM 0x01a2 WLI-UC-GNM product MELCO WLIUCGNM2 0x01ee WLI-UC-GNM2 /* Merlin products */ product MERLIN V620 0x1110 Merlin V620 /* MetaGeek products */ product METAGEEK TELLSTICK 0x0c30 FTDI compatible adapter product METAGEEK WISPY1B 0x083e MetaGeek Wi-Spy product METAGEEK WISPY24X 0x083f MetaGeek Wi-Spy 2.4x product METAGEEK2 WISPYDBX 0x5000 MetaGeek Wi-Spy DBx /* Metricom products */ product METRICOM RICOCHET_GS 0x0001 Ricochet GS /* MGE UPS Systems */ product MGE UPS1 0x0001 MGE UPS SYSTEMS PROTECTIONCENTER 1 product MGE UPS2 0xffff MGE UPS SYSTEMS PROTECTIONCENTER 2 /* MEI products */ product MEI CASHFLOW_SC 0x1100 Cashflow-SC Cash Acceptor product MEI S2000 0x1101 Series 2000 Combo Acceptor /* Micro Star International products */ product MSI BT_DONGLE 0x1967 Bluetooth USB dongle product MSI RT3070_1 0x3820 RT3070 product MSI RT3070_2 0x3821 RT3070 product MSI RT3070_8 0x3822 RT3070 product MSI RT3070_3 0x3870 RT3070 product MSI RT3070_9 0x3871 RT3070 product MSI UB11B 0x6823 UB11B product MSI RT2570 0x6861 RT2570 product MSI RT2570_2 0x6865 RT2570 product MSI RT2570_3 0x6869 RT2570 product MSI RT2573_1 0x6874 RT2573 product MSI RT2573_2 0x6877 RT2573 product MSI RT3070_4 0x6899 RT3070 product MSI RT3070_5 0x821a RT3070 product MSI RT3070_10 0x822a RT3070 product MSI RT3070_6 0x870a RT3070 product MSI RT3070_11 0x871a RT3070 product MSI RT3070_7 0x899a RT3070 product MSI RT2573_3 0xa861 RT2573 product MSI RT2573_4 0xa874 RT2573 /* Micron products */ product MICRON REALSSD 0x0655 Real SSD eUSB /* Microsoft products */ product MICROSOFT SIDEPREC 0x0008 SideWinder Precision Pro product MICROSOFT INTELLIMOUSE 0x0009 IntelliMouse product MICROSOFT NATURALKBD 0x000b Natural Keyboard Elite product MICROSOFT DDS80 0x0014 Digital Sound System 80 product MICROSOFT SIDEWINDER 0x001a Sidewinder Precision Racing Wheel product MICROSOFT INETPRO 0x001c Internet Keyboard Pro product MICROSOFT TBEXPLORER 0x0024 Trackball Explorer product MICROSOFT INTELLIEYE 0x0025 IntelliEye mouse product MICROSOFT INETPRO2 0x002b Internet Keyboard Pro product MICROSOFT INTELLIMOUSE5 0x0039 IntelliMouse 1.1 5-Button Mouse product MICROSOFT WHEELMOUSE 0x0040 Wheel Mouse Optical product MICROSOFT MN510 0x006e MN510 Wireless product MICROSOFT 700WX 0x0079 Palm 700WX product MICROSOFT MN110 0x007a 10/100 USB NIC product MICROSOFT WLINTELLIMOUSE 0x008c Wireless Optical IntelliMouse product MICROSOFT WLNOTEBOOK 0x00b9 Wireless Optical Mouse (Model 1023) product MICROSOFT COMFORT3000 0x00d1 Comfort Optical Mouse 3000 (Model 1043) product MICROSOFT WLNOTEBOOK3 0x00d2 Wireless Optical Mouse 3000 (Model 1049) product MICROSOFT NATURAL4000 0x00db Natural Ergonomic Keyboard 4000 product MICROSOFT WLNOTEBOOK2 0x00e1 Wireless Optical Mouse 3000 (Model 1056) product MICROSOFT XBOX360 0x0292 XBOX 360 WLAN /* Microtech products */ product MICROTECH SCSIDB25 0x0004 USB-SCSI-DB25 product MICROTECH SCSIHD50 0x0005 USB-SCSI-HD50 product MICROTECH DPCM 0x0006 USB CameraMate product MICROTECH FREECOM 0xfc01 Freecom USB-IDE /* Microtek products */ product MICROTEK 336CX 0x0094 Phantom 336CX - C3 scanner product MICROTEK X6U 0x0099 ScanMaker X6 - X6U product MICROTEK C6 0x009a Phantom C6 scanner product MICROTEK 336CX2 0x00a0 Phantom 336CX - C3 scanner product MICROTEK V6USL 0x00a3 ScanMaker V6USL product MICROTEK V6USL2 0x80a3 ScanMaker V6USL product MICROTEK V6UL 0x80ac ScanMaker V6UL /* Microtune, Inc. products */ product MICROTUNE BT_DONGLE 0x1000 Bluetooth USB dongle /* Midiman products */ product MAUDIO MIDISPORT2X2 0x1001 Midisport 2x2 product MAUDIO FASTTRACKULTRA 0x2080 Fast Track Ultra product MAUDIO FASTTRACKULTRA8R 0x2081 Fast Track Ultra 8R /* MindsAtWork products */ product MINDSATWORK WALLET 0x0001 Digital Wallet /* Minolta Co., Ltd. */ product MINOLTA 2300 0x4001 Dimage 2300 product MINOLTA S304 0x4007 Dimage S304 product MINOLTA X 0x4009 Dimage X product MINOLTA 5400 0x400e Dimage 5400 product MINOLTA F300 0x4011 Dimage F300 product MINOLTA E223 0x4017 Dimage E223 /* Mitsumi products */ product MITSUMI CDRRW 0x0000 CD-R/RW Drive product MITSUMI BT_DONGLE 0x641f Bluetooth USB dongle product MITSUMI FDD 0x6901 USB FDD /* Mobile Action products */ product MOBILEACTION MA620 0x0620 MA-620 Infrared Adapter /* Mobility products */ product MOBILITY USB_SERIAL 0x0202 FTDI compatible adapter product MOBILITY EA 0x0204 Ethernet product MOBILITY EASIDOCK 0x0304 EasiDock Ethernet /* MosChip products */ product MOSCHIP MCS7703 0x7703 MCS7703 Serial Port Adapter product MOSCHIP MCS7730 0x7730 MCS7730 Ethernet product MOSCHIP MCS7820 0x7820 MCS7820 Serial Port Adapter product MOSCHIP MCS7830 0x7830 MCS7830 Ethernet product MOSCHIP MCS7832 0x7832 MCS7832 Ethernet product MOSCHIP MCS7840 0x7840 MCS7840 Serial Port Adapter /* Motorola products */ product MOTOROLA MC141555 0x1555 MC141555 hub controller product MOTOROLA SB4100 0x4100 SB4100 USB Cable Modem product MOTOROLA2 T720C 0x2822 T720c product MOTOROLA2 A41XV32X 0x2a22 A41x/V32x Mobile Phones product MOTOROLA2 E398 0x4810 E398 Mobile Phone product MOTOROLA2 USBLAN 0x600c USBLAN product MOTOROLA2 USBLAN2 0x6027 USBLAN product MOTOROLA2 MB886 0x710f MB886 Mobile Phone (Atria HD) product MOTOROLA4 RT2770 0x9031 RT2770 product MOTOROLA4 RT3070 0x9032 RT3070 /* MpMan products */ product MPMAN MPF400_2 0x25a8 MPF400 Music Player 2Go product MPMAN MPF400_1 0x36d0 MPF400 Music Player 1Go /* MultiTech products */ product MULTITECH ATLAS 0xf101 MT5634ZBA-USB modem /* Mustek products */ product MUSTEK 1200CU 0x0001 1200 CU scanner product MUSTEK 600CU 0x0002 600 CU scanner product MUSTEK 1200USB 0x0003 1200 USB scanner product MUSTEK 1200UB 0x0006 1200 UB scanner product MUSTEK 1200USBPLUS 0x0007 1200 USB Plus scanner product MUSTEK 1200CUPLUS 0x0008 1200 CU Plus scanner product MUSTEK BEARPAW1200F 0x0010 BearPaw 1200F scanner product MUSTEK BEARPAW2400TA 0x0218 BearPaw 2400TA scanner product MUSTEK BEARPAW1200TA 0x021e BearPaw 1200TA scanner product MUSTEK 600USB 0x0873 600 USB scanner product MUSTEK MDC800 0xa800 MDC-800 digital camera /* M-Systems products */ product MSYSTEMS DISKONKEY 0x0010 DiskOnKey product MSYSTEMS DISKONKEY2 0x0011 DiskOnKey /* Myson products */ product MYSON HEDEN_8813 0x8813 USB-IDE product MYSON HEDEN 0x8818 USB-IDE product MYSON HUBREADER 0x8819 COMBO Card reader with USB HUB product MYSON STARREADER 0x9920 USB flash card adapter /* National Semiconductor */ product NATIONAL BEARPAW1200 0x1000 BearPaw 1200 product NATIONAL BEARPAW2400 0x1001 BearPaw 2400 /* NEC products */ product NEC HUB_0050 0x0050 USB 2.0 7-Port Hub product NEC HUB_005A 0x005a USB 2.0 4-Port Hub product NEC HUB 0x55aa hub product NEC HUB_B 0x55ab hub /* NEODIO products */ product NEODIO ND3260 0x3260 8-in-1 Multi-format Flash Controller product NEODIO ND5010 0x5010 Multi-format Flash Controller /* Neotel products */ product NEOTEL PRIME 0x4000 Prime USB modem /* Netac products */ product NETAC CF_CARD 0x1060 USB-CF-Card product NETAC ONLYDISK 0x0003 OnlyDisk /* NetChip Technology Products */ product NETCHIP TURBOCONNECT 0x1080 Turbo-Connect product NETCHIP CLIK_40 0xa140 USB Clik! 40 product NETCHIP GADGETZERO 0xa4a0 Linux Gadget Zero product NETCHIP ETHERNETGADGET 0xa4a2 Linux Ethernet/RNDIS gadget on pxa210/25x/26x product NETCHIP POCKETBOOK 0xa4a5 PocketBook /* Netgear products */ product NETGEAR EA101 0x1001 Ethernet product NETGEAR EA101X 0x1002 Ethernet product NETGEAR FA101 0x1020 Ethernet 10/100, USB1.1 product NETGEAR FA120 0x1040 USB 2.0 Ethernet product NETGEAR M4100 0x1100 M4100/M5300/M7100 series switch -product NETGEAR WG111V2_2 0x4240 PrismGT USB 2.0 WLAN +product NETGEAR WG111V1_2 0x4240 PrismGT USB 2.0 WLAN product NETGEAR WG111V3 0x4260 WG111v3 product NETGEAR WG111U 0x4300 WG111U product NETGEAR WG111U_NF 0x4301 WG111U (no firmware) product NETGEAR WG111V2 0x6a00 WG111V2 product NETGEAR RTL8192CU 0x9021 RTL8192CU product NETGEAR WNA1000M 0x9041 WNA1000M product NETGEAR2 MA101 0x4100 MA101 product NETGEAR2 MA101B 0x4102 MA101 Rev B product NETGEAR3 WG111T 0x4250 WG111T product NETGEAR3 WG111T_NF 0x4251 WG111T (no firmware) product NETGEAR3 WPN111 0x5f00 WPN111 product NETGEAR3 WPN111_NF 0x5f01 WPN111 (no firmware) product NETGEAR3 WPN111_2 0x5f02 WPN111 product NETGEAR4 RTL8188CU 0x9041 RTL8188CU /* NetIndex products */ product NETINDEX WS002IN 0x2001 Willcom WS002IN /* NEWlink */ product NEWLINK USB2IDEBRIDGE 0x00ff USB 2.0 Hard Drive Enclosure /* Nikon products */ product NIKON E990 0x0102 Digital Camera E990 product NIKON LS40 0x4000 CoolScan LS40 ED product NIKON D300 0x041a Digital Camera D300 /* NovaTech Products */ product NOVATECH NV902 0x9020 NovaTech NV-902W product NOVATECH RT2573 0x9021 RT2573 product NOVATECH RTL8188CU 0x9071 RTL8188CU /* Nokia products */ product NOKIA N958GB 0x0070 Nokia N95 8GBc product NOKIA2 CA42 0x1234 CA-42 cable /* Novatel Wireless products */ product NOVATEL V640 0x1100 Merlin V620 product NOVATEL CDMA_MODEM 0x1110 Novatel Wireless Merlin CDMA product NOVATEL V620 0x1110 Merlin V620 product NOVATEL V740 0x1120 Merlin V740 product NOVATEL V720 0x1130 Merlin V720 product NOVATEL U740 0x1400 Merlin U740 product NOVATEL U740_2 0x1410 Merlin U740 product NOVATEL U870 0x1420 Merlin U870 product NOVATEL XU870 0x1430 Merlin XU870 product NOVATEL X950D 0x1450 Merlin X950D product NOVATEL ES620 0x2100 Expedite ES620 product NOVATEL E725 0x2120 Expedite E725 product NOVATEL ES620_2 0x2130 Expedite ES620 product NOVATEL ES620 0x2100 ES620 CDMA product NOVATEL U720 0x2110 Merlin U720 product NOVATEL EU730 0x2400 Expedite EU730 product NOVATEL EU740 0x2410 Expedite EU740 product NOVATEL EU870D 0x2420 Expedite EU870D product NOVATEL U727 0x4100 Merlin U727 CDMA product NOVATEL MC950D 0x4400 Novatel MC950D HSUPA product NOVATEL MC990D 0x7001 Novatel MC990D product NOVATEL ZEROCD 0x5010 Novatel ZeroCD product NOVATEL MIFI2200V 0x5020 Novatel MiFi 2200 CDMA Virgin Mobile product NOVATEL ZEROCD2 0x5030 Novatel ZeroCD product NOVATEL MIFI2200 0x5041 Novatel MiFi 2200 CDMA product NOVATEL U727_2 0x5100 Merlin U727 CDMA product NOVATEL U760 0x6000 Novatel U760 product NOVATEL MC760 0x6002 Novatel MC760 product NOVATEL MC547 0x7042 Novatel MC547 product NOVATEL MC679 0x7031 Novatel MC679 product NOVATEL2 FLEXPACKGPS 0x0100 NovAtel FlexPack GPS receiver /* Merlin products */ product MERLIN V620 0x1110 Merlin V620 /* O2Micro products */ product O2MICRO OZ776_HUB 0x7761 OZ776 hub product O2MICRO OZ776_CCID_SC 0x7772 OZ776 CCID SC Reader /* Olimex products */ product OLIMEX ARM_USB_OCD 0x0003 FTDI compatible adapter product OLIMEX ARM_USB_OCD_H 0x002b FTDI compatible adapter /* Olympus products */ product OLYMPUS C1 0x0102 C-1 Digital Camera product OLYMPUS C700 0x0105 C-700 Ultra Zoom /* OmniVision Technologies, Inc. products */ product OMNIVISION OV511 0x0511 OV511 Camera product OMNIVISION OV511PLUS 0xa511 OV511+ Camera /* OnSpec Electronic, Inc. */ product ONSPEC SDS_HOTFIND_D 0x0400 SDS-infrared.com Hotfind-D Infrared Camera product ONSPEC MDCFE_B_CF_READER 0xa000 MDCFE-B USB CF Reader product ONSPEC CFMS_RW 0xa001 SIIG/Datafab Memory Stick+CF Reader/Writer product ONSPEC READER 0xa003 Datafab-based Reader product ONSPEC CFSM_READER 0xa005 PNY/Datafab CF+SM Reader product ONSPEC CFSM_READER2 0xa006 Simple Tech/Datafab CF+SM Reader product ONSPEC MDSM_B_READER 0xa103 MDSM-B reader product ONSPEC CFSM_COMBO 0xa109 USB to CF + SM Combo (LC1) product ONSPEC UCF100 0xa400 FlashLink UCF-100 CompactFlash Reader product ONSPEC2 IMAGEMATE_SDDR55 0xa103 ImageMate SDDR55 /* Option products */ product OPTION VODAFONEMC3G 0x5000 Vodafone Mobile Connect 3G datacard product OPTION GT3G 0x6000 GlobeTrotter 3G datacard product OPTION GT3GQUAD 0x6300 GlobeTrotter 3G QUAD datacard product OPTION GT3GPLUS 0x6600 GlobeTrotter 3G+ datacard product OPTION GTICON322 0xd033 GlobeTrotter Icon322 storage product OPTION GTMAX36 0x6701 GlobeTrotter Max 3.6 Modem product OPTION GTMAX72 0x6711 GlobeTrotter Max 7.2 HSDPA product OPTION GTHSDPA 0x6971 GlobeTrotter HSDPA product OPTION GTMAXHSUPA 0x7001 GlobeTrotter HSUPA product OPTION GTMAXHSUPAE 0x6901 GlobeTrotter HSUPA PCIe product OPTION GTMAX380HSUPAE 0x7211 GlobeTrotter 380HSUPA PCIe product OPTION GT3G_1 0x6050 3G modem product OPTION GT3G_2 0x6100 3G modem product OPTION GT3G_3 0x6150 3G modem product OPTION GT3G_4 0x6200 3G modem product OPTION GT3G_5 0x6250 3G modem product OPTION GT3G_6 0x6350 3G modem product OPTION E6500 0x6500 3G modem product OPTION E6501 0x6501 3G modem product OPTION E6601 0x6601 3G modem product OPTION E6721 0x6721 3G modem product OPTION E6741 0x6741 3G modem product OPTION E6761 0x6761 3G modem product OPTION E6800 0x6800 3G modem product OPTION E7021 0x7021 3G modem product OPTION E7041 0x7041 3G modem product OPTION E7061 0x7061 3G modem product OPTION E7100 0x7100 3G modem product OPTION GTM380 0x7201 3G modem product OPTION GE40X 0x7601 Globetrotter HSUPA product OPTION GSICON72 0x6911 GlobeSurfer iCON product OPTION GSICONHSUPA 0x7251 Globetrotter HSUPA product OPTION ICON401 0x7401 GlobeSurfer iCON 401 product OPTION GTHSUPA 0x7011 Globetrotter HSUPA product OPTION GMT382 0x7501 Globetrotter HSUPA product OPTION GE40X_1 0x7301 Globetrotter HSUPA product OPTION GE40X_2 0x7361 Globetrotter HSUPA product OPTION GE40X_3 0x7381 Globetrotter HSUPA product OPTION GTM661W 0x9000 GTM661W product OPTION ICONEDGE 0xc031 GlobeSurfer iCON EDGE product OPTION MODHSXPA 0xd013 Globetrotter HSUPA product OPTION ICON321 0xd031 Globetrotter HSUPA product OPTION ICON505 0xd055 Globetrotter iCON 505 product OPTION ICON452 0x7901 Globetrotter iCON 452 /* Optoelectronics Co., Ltd */ product OPTO BARCODE 0x0001 Barcode Reader product OPTO OPTICONCODE 0x0009 Opticon Code Reader product OPTO BARCODE_1 0xa002 Barcode Reader product OPTO CRD7734 0xc000 USB Cradle CRD-7734-RU product OPTO CRD7734_1 0xc001 USB Cradle CRD-7734-RU /* OvisLink product */ product OVISLINK RT3072 0x3072 RT3072 /* OQO */ product OQO WIFI01 0x0002 model 01 WiFi interface product OQO BT01 0x0003 model 01 Bluetooth interface product OQO ETHER01PLUS 0x7720 model 01+ Ethernet product OQO ETHER01 0x8150 model 01 Ethernet interface /* Ours Technology Inc. */ product OTI DKU5 0x6858 DKU-5 Serial /* Owen.ru products */ product OWEN AC4 0x0004 AC4 USB-RS485 converter /* OWL producs */ product OWL CM_160 0xca05 OWL CM-160 power monitor /* Palm Computing, Inc. product */ product PALM SERIAL 0x0080 USB Serial product PALM M500 0x0001 Palm m500 product PALM M505 0x0002 Palm m505 product PALM M515 0x0003 Palm m515 product PALM I705 0x0020 Palm i705 product PALM TUNGSTEN_Z 0x0031 Palm Tungsten Z product PALM M125 0x0040 Palm m125 product PALM M130 0x0050 Palm m130 product PALM TUNGSTEN_T 0x0060 Palm Tungsten T product PALM ZIRE31 0x0061 Palm Zire 31 product PALM ZIRE 0x0070 Palm Zire /* Panasonic products */ product PANASONIC LS120CAM 0x0901 LS-120 Camera product PANASONIC KXL840AN 0x0d01 CD-R Drive KXL-840AN product PANASONIC KXLRW32AN 0x0d09 CD-R Drive KXL-RW32AN product PANASONIC KXLCB20AN 0x0d0a CD-R Drive KXL-CB20AN product PANASONIC KXLCB35AN 0x0d0e DVD-ROM & CD-R/RW product PANASONIC SDCAAE 0x1b00 MultiMediaCard product PANASONIC TYTP50P6S 0x3900 TY-TP50P6-S 50in Touch Panel /* Papouch products */ product PAPOUCH AD4USB 0x8003 FTDI compatible adapter product PAPOUCH AP485 0x0101 FTDI compatible adapter product PAPOUCH AP485_2 0x0104 FTDI compatible adapter product PAPOUCH DRAK5 0x0700 FTDI compatible adapter product PAPOUCH DRAK6 0x1000 FTDI compatible adapter product PAPOUCH GMSR 0x8005 FTDI compatible adapter product PAPOUCH GMUX 0x8004 FTDI compatible adapter product PAPOUCH IRAMP 0x0500 FTDI compatible adapter product PAPOUCH LEC 0x0300 FTDI compatible adapter product PAPOUCH MU 0x8001 FTDI compatible adapter product PAPOUCH QUIDO10X1 0x0b00 FTDI compatible adapter product PAPOUCH QUIDO2X16 0x0e00 FTDI compatible adapter product PAPOUCH QUIDO2X2 0x0a00 FTDI compatible adapter product PAPOUCH QUIDO30X3 0x0c00 FTDI compatible adapter product PAPOUCH QUIDO3X32 0x0f00 FTDI compatible adapter product PAPOUCH QUIDO4X4 0x0900 FTDI compatible adapter product PAPOUCH QUIDO60X3 0x0d00 FTDI compatible adapter product PAPOUCH QUIDO8X8 0x0800 FTDI compatible adapter product PAPOUCH SB232 0x0301 FTDI compatible adapter product PAPOUCH SB422 0x0102 FTDI compatible adapter product PAPOUCH SB422_2 0x0105 FTDI compatible adapter product PAPOUCH SB485 0x0100 FTDI compatible adapter product PAPOUCH SB485C 0x0107 FTDI compatible adapter product PAPOUCH SB485S 0x0106 FTDI compatible adapter product PAPOUCH SB485_2 0x0103 FTDI compatible adapter product PAPOUCH SIMUKEY 0x8002 FTDI compatible adapter product PAPOUCH TMU 0x0400 FTDI compatible adapter product PAPOUCH UPSUSB 0x8000 FTDI compatible adapter /* PARA Industrial products */ product PARA RT3070 0x8888 RT3070 /* Simtec Electronics products */ product SIMTEC ENTROPYKEY 0x0001 Entropy Key /* Pegatron products */ product PEGATRON RT2870 0x0002 RT2870 product PEGATRON RT3070 0x000c RT3070 product PEGATRON RT3070_2 0x000e RT3070 product PEGATRON RT3070_3 0x0010 RT3070 /* Peracom products */ product PERACOM SERIAL1 0x0001 Serial product PERACOM ENET 0x0002 Ethernet product PERACOM ENET3 0x0003 At Home Ethernet product PERACOM ENET2 0x0005 Ethernet /* Philips products */ product PHILIPS DSS350 0x0101 DSS 350 Digital Speaker System product PHILIPS DSS 0x0104 DSS XXX Digital Speaker System product PHILIPS HUB 0x0201 hub product PHILIPS PCA646VC 0x0303 PCA646VC PC Camera product PHILIPS PCVC680K 0x0308 PCVC680K Vesta Pro PC Camera product PHILIPS DSS150 0x0471 DSS 150 Digital Speaker System product PHILIPS ACE1001 0x066a AKTAKOM ACE-1001 cable product PHILIPS SPE3030CC 0x083a USB 2.0 External Disk product PHILIPS SNU5600 0x1236 SNU5600 product PHILIPS UM10016 0x1552 ISP 1581 Hi-Speed USB MPEG2 Encoder Reference Kit product PHILIPS DIVAUSB 0x1801 DIVA USB mp3 player product PHILIPS RT2870 0x200f RT2870 /* Philips Semiconductor products */ product PHILIPSSEMI HUB1122 0x1122 HUB /* Megatec */ product MEGATEC UPS 0x5161 Phoenixtec protocol based UPS /* P.I. Engineering products */ product PIENGINEERING PS2USB 0x020b PS2 to Mac USB Adapter /* Planex Communications products */ product PLANEX GW_US11H 0x14ea GW-US11H WLAN product PLANEX2 RTL8188CUS 0x1201 RTL8188CUS product PLANEX2 GW_US11S 0x3220 GW-US11S WLAN product PLANEX2 GW_US54GXS 0x5303 GW-US54GXS WLAN product PLANEX2 RTL8188CU_1 0xab2a RTL8188CU product PLANEX2 RTL8188CU_2 0xed17 RTL8188CU product PLANEX2 RTL8188CU_3 0x4902 RTL8188CU product PLANEX2 RTL8188CU_4 0xab2e RTL8188CU product PLANEX2 RTL8192CU 0xab2b RTL8192CU product PLANEX2 GWUS54HP 0xab01 GW-US54HP product PLANEX2 GWUS300MINIS 0xab24 GW-US300MiniS product PLANEX2 RT3070 0xab25 RT3070 product PLANEX2 MZKUE150N 0xab2f MZK-UE150N product PLANEX2 GWUS54MINI2 0xab50 GW-US54Mini2 product PLANEX2 GWUS54SG 0xc002 GW-US54SG product PLANEX2 GWUS54GZL 0xc007 GW-US54GZL product PLANEX2 GWUS54GD 0xed01 GW-US54GD product PLANEX2 GWUSMM 0xed02 GW-USMM product PLANEX2 RT2870 0xed06 RT2870 product PLANEX2 GWUSMICRON 0xed14 GW-USMicroN product PLANEX2 GWUSVALUEEZ 0xed17 GW-USValue-EZ product PLANEX3 GWUS54GZ 0xab10 GW-US54GZ product PLANEX3 GU1000T 0xab11 GU-1000T product PLANEX3 GWUS54MINI 0xab13 GW-US54Mini product PLANEX2 GWUSNANO 0xab28 GW-USNano /* Plextor Corp. */ product PLEXTOR 40_12_40U 0x0011 PlexWriter 40/12/40U /* PLX products */ product PLX TESTBOARD 0x9060 test board product PLX CA42 0xac70 CA-42 /* PNY products */ product PNY ATTACHE2 0x0010 USB 2.0 Flash Drive /* PortGear products */ product PORTGEAR EA8 0x0008 Ethernet product PORTGEAR EA9 0x0009 Ethernet /* Portsmith products */ product PORTSMITH EEA 0x3003 Express Ethernet /* Posiflex products */ product POSIFLEX PP7000 0x0300 FTDI compatible adapter /* Primax products */ product PRIMAX G2X300 0x0300 G2-200 scanner product PRIMAX G2E300 0x0301 G2E-300 scanner product PRIMAX G2300 0x0302 G2-300 scanner product PRIMAX G2E3002 0x0303 G2E-300 scanner product PRIMAX 9600 0x0340 Colorado USB 9600 scanner product PRIMAX 600U 0x0341 Colorado 600u scanner product PRIMAX 6200 0x0345 Visioneer 6200 scanner product PRIMAX 19200 0x0360 Colorado USB 19200 scanner product PRIMAX 1200U 0x0361 Colorado 1200u scanner product PRIMAX G600 0x0380 G2-600 scanner product PRIMAX 636I 0x0381 ReadyScan 636i product PRIMAX G2600 0x0382 G2-600 scanner product PRIMAX G2E600 0x0383 G2E-600 scanner product PRIMAX COMFORT 0x4d01 Comfort product PRIMAX MOUSEINABOX 0x4d02 Mouse-in-a-Box product PRIMAX PCGAUMS1 0x4d04 Sony PCGA-UMS1 product PRIMAX HP_RH304AA 0x4d17 HP RH304AA mouse /* Prolific products */ product PROLIFIC PL2301 0x0000 PL2301 Host-Host interface product PROLIFIC PL2302 0x0001 PL2302 Host-Host interface product PROLIFIC MOTOROLA 0x0307 Motorola Cable product PROLIFIC RSAQ2 0x04bb PL2303 Serial (IODATA USB-RSAQ2) product PROLIFIC ALLTRONIX_GPRS 0x0609 Alltronix ACM003U00 modem product PROLIFIC ALDIGA_AL11U 0x0611 AlDiga AL-11U modem product PROLIFIC MICROMAX_610U 0x0612 Micromax 610U product PROLIFIC DCU11 0x1234 DCU-11 Phone Cable product PROLIFIC UIC_MSR206 0x206a UIC MSR206 Card Reader product PROLIFIC PL2303 0x2303 PL2303 Serial (ATEN/IOGEAR UC232A) product PROLIFIC PL2305 0x2305 Parallel printer product PROLIFIC ATAPI4 0x2307 ATAPI-4 Controller product PROLIFIC PL2501 0x2501 PL2501 Host-Host interface product PROLIFIC PL2506 0x2506 PL2506 USB to IDE Bridge product PROLIFIC HCR331 0x331a HCR331 Hybrid Card Reader product PROLIFIC PHAROS 0xaaa0 Prolific Pharos product PROLIFIC RSAQ3 0xaaa2 PL2303 Serial Adapter (IODATA USB-RSAQ3) product PROLIFIC2 PL2303 0x2303 PL2303 Serial Adapter /* Putercom products */ product PUTERCOM UPA100 0x047e USB-1284 BRIDGE /* Qcom products */ product QCOM RT2573 0x6196 RT2573 product QCOM RT2573_2 0x6229 RT2573 product QCOM RT2573_3 0x6238 RT2573 product QCOM RT2870 0x6259 RT2870 /* QI-hardware */ product QIHARDWARE JTAGSERIAL 0x0713 FTDI compatible adapter /* Qisda products */ product QISDA H21_1 0x4512 3G modem product QISDA H21_2 0x4523 3G modem product QISDA H20_1 0x4515 3G modem product QISDA H20_2 0x4519 3G modem /* Qualcomm products */ product QUALCOMM CDMA_MSM 0x6000 CDMA Technologies MSM phone product QUALCOMM NTT_L02C_MODEM 0x618f NTT DOCOMO L-02C product QUALCOMM NTT_L02C_STORAGE 0x61dd NTT DOCOMO L-02C product QUALCOMM2 MF330 0x6613 MF330 product QUALCOMM2 RWT_FCT 0x3100 RWT FCT-CDMA 2000 1xRTT modem product QUALCOMM2 CDMA_MSM 0x3196 CDMA Technologies MSM modem product QUALCOMM2 AC8700 0x6000 AC8700 product QUALCOMM2 VW110L 0x1000 Vertex Wireless 110L modem product QUALCOMM2 SIM5218 0x9000 SIM5218 product QUALCOMM2 GOBI2000_QDL 0x9204 Qualcomm Gobi 2000 QDL product QUALCOMM2 GOBI2000 0x9205 Qualcomm Gobi 2000 modem +product QUALCOMM2 VT80N 0x6500 Venus VT80N +product QUALCOMM3 VFAST2 0x9909 Venus Fast2 modem product QUALCOMMINC CDMA_MSM 0x0001 CDMA Technologies MSM modem product QUALCOMMINC E0002 0x0002 3G modem product QUALCOMMINC E0003 0x0003 3G modem product QUALCOMMINC E0004 0x0004 3G modem product QUALCOMMINC E0005 0x0005 3G modem product QUALCOMMINC E0006 0x0006 3G modem product QUALCOMMINC E0007 0x0007 3G modem product QUALCOMMINC E0008 0x0008 3G modem product QUALCOMMINC E0009 0x0009 3G modem product QUALCOMMINC E000A 0x000a 3G modem product QUALCOMMINC E000B 0x000b 3G modem product QUALCOMMINC E000C 0x000c 3G modem product QUALCOMMINC E000D 0x000d 3G modem product QUALCOMMINC E000E 0x000e 3G modem product QUALCOMMINC E000F 0x000f 3G modem product QUALCOMMINC E0010 0x0010 3G modem product QUALCOMMINC E0011 0x0011 3G modem product QUALCOMMINC E0012 0x0012 3G modem product QUALCOMMINC E0013 0x0013 3G modem product QUALCOMMINC E0014 0x0014 3G modem product QUALCOMMINC MF628 0x0015 3G modem product QUALCOMMINC MF633R 0x0016 ZTE WCDMA modem product QUALCOMMINC E0017 0x0017 3G modem product QUALCOMMINC E0018 0x0018 3G modem product QUALCOMMINC E0019 0x0019 3G modem product QUALCOMMINC E0020 0x0020 3G modem product QUALCOMMINC E0021 0x0021 3G modem product QUALCOMMINC E0022 0x0022 3G modem product QUALCOMMINC E0023 0x0023 3G modem product QUALCOMMINC E0024 0x0024 3G modem product QUALCOMMINC E0025 0x0025 3G modem product QUALCOMMINC E0026 0x0026 3G modem product QUALCOMMINC E0027 0x0027 3G modem product QUALCOMMINC E0028 0x0028 3G modem product QUALCOMMINC E0029 0x0029 3G modem product QUALCOMMINC E0030 0x0030 3G modem product QUALCOMMINC MF626 0x0031 3G modem product QUALCOMMINC E0032 0x0032 3G modem product QUALCOMMINC E0033 0x0033 3G modem product QUALCOMMINC E0037 0x0037 3G modem product QUALCOMMINC E0039 0x0039 3G modem product QUALCOMMINC E0042 0x0042 3G modem product QUALCOMMINC E0043 0x0043 3G modem product QUALCOMMINC E0048 0x0048 3G modem product QUALCOMMINC E0049 0x0049 3G modem product QUALCOMMINC E0051 0x0051 3G modem product QUALCOMMINC E0052 0x0052 3G modem product QUALCOMMINC ZTE_STOR2 0x0053 USB ZTE Storage product QUALCOMMINC E0054 0x0054 3G modem product QUALCOMMINC E0055 0x0055 3G modem product QUALCOMMINC E0057 0x0057 3G modem product QUALCOMMINC E0058 0x0058 3G modem product QUALCOMMINC E0059 0x0059 3G modem product QUALCOMMINC E0060 0x0060 3G modem product QUALCOMMINC E0061 0x0061 3G modem product QUALCOMMINC E0062 0x0062 3G modem product QUALCOMMINC E0063 0x0063 3G modem product QUALCOMMINC E0064 0x0064 3G modem product QUALCOMMINC E0066 0x0066 3G modem product QUALCOMMINC E0069 0x0069 3G modem product QUALCOMMINC E0070 0x0070 3G modem product QUALCOMMINC E0073 0x0073 3G modem product QUALCOMMINC E0076 0x0076 3G modem product QUALCOMMINC E0078 0x0078 3G modem product QUALCOMMINC E0082 0x0082 3G modem product QUALCOMMINC E0086 0x0086 3G modem product QUALCOMMINC SURFSTICK 0x0117 1&1 Surf Stick product QUALCOMMINC K3772_Z_INIT 0x1179 K3772-Z Initial product QUALCOMMINC K3772_Z 0x1181 K3772-Z product QUALCOMMINC MF195E_INIT 0x1514 MF195E initial product QUALCOMMINC MF195E 0x1516 MF195E product QUALCOMMINC ZTE_STOR 0x2000 USB ZTE Storage product QUALCOMMINC E2002 0x2002 3G modem product QUALCOMMINC E2003 0x2003 3G modem +product QUALCOMMINC AC682 0xffdd CDMA 1xEVDO USB modem +product QUALCOMMINC AC682_INIT 0xffde CDMA 1xEVDO USB modem (initial) product QUALCOMMINC AC8710 0xfff1 3G modem product QUALCOMMINC AC2726 0xfff5 3G modem product QUALCOMMINC AC8700 0xfffe CDMA 1xEVDO USB modem /* Quanta products */ product QUANTA RW6815_1 0x00ce HP iPAQ rw6815 product QUANTA RT3070 0x0304 RT3070 product QUANTA Q101_STOR 0x1000 USB Q101 Storage product QUANTA Q101 0xea02 HSDPA modem product QUANTA Q111 0xea03 HSDPA modem product QUANTA GLX 0xea04 HSDPA modem product QUANTA GKE 0xea05 HSDPA modem product QUANTA GLE 0xea06 HSDPA modem product QUANTA RW6815R 0xf003 HP iPAQ rw6815 RNDIS /* Qtronix products */ product QTRONIX 980N 0x2011 Scorpion-980N keyboard /* Quickshot products */ product QUICKSHOT STRIKEPAD 0x6238 USB StrikePad /* Radio Shack */ product RADIOSHACK USBCABLE 0x4026 USB to Serial Cable /* Rainbow Technologies products */ product RAINBOW IKEY2000 0x1200 i-Key 2000 /* Ralink Technology products */ product RALINK RT2570 0x1706 RT2500USB Wireless Adapter product RALINK RT2070 0x2070 RT2070 product RALINK RT2570_2 0x2570 RT2500USB Wireless Adapter product RALINK RT2573 0x2573 RT2501USB Wireless Adapter product RALINK RT2671 0x2671 RT2601USB Wireless Adapter product RALINK RT2770 0x2770 RT2770 product RALINK RT2870 0x2870 RT2870 product RALINK RT_STOR 0x2878 USB Storage product RALINK RT3070 0x3070 RT3070 product RALINK RT3071 0x3071 RT3071 product RALINK RT3072 0x3072 RT3072 product RALINK RT3370 0x3370 RT3370 product RALINK RT3572 0x3572 RT3572 product RALINK RT3573 0x3573 RT3573 product RALINK RT5370 0x5370 RT5370 product RALINK RT5572 0x5572 RT5572 product RALINK RT8070 0x8070 RT8070 product RALINK RT2570_3 0x9020 RT2500USB Wireless Adapter product RALINK RT2573_2 0x9021 RT2501USB Wireless Adapter /* RATOC Systems products */ product RATOC REXUSB60 0xb000 USB serial adapter REX-USB60 product RATOC REXUSB60F 0xb020 USB serial adapter REX-USB60F /* Realtek products */ /* Green House and CompUSA OEM this part */ product REALTEK DUMMY 0x0000 Dummy product product REALTEK USB20CRW 0x0158 USB20CRW Card Reader product REALTEK RTL8188ETV 0x0179 RTL8188ETV product REALTEK RTL8188CTV 0x018a RTL8188CTV product REALTEK USBKR100 0x8150 USBKR100 USB Ethernet product REALTEK RTL8188CE_0 0x8170 RTL8188CE product REALTEK RTL8171 0x8171 RTL8171 product REALTEK RTL8172 0x8172 RTL8172 product REALTEK RTL8173 0x8173 RTL8173 product REALTEK RTL8174 0x8174 RTL8174 product REALTEK RTL8188CU_0 0x8176 RTL8188CU product REALTEK RTL8188EU 0x8179 RTL8188EU product REALTEK RTL8188CE_1 0x817e RTL8188CE product REALTEK RTL8188CU_1 0x817a RTL8188CU product REALTEK RTL8188CU_2 0x817b RTL8188CU product REALTEK RTL8187 0x8187 RTL8187 Wireless Adapter product REALTEK RTL8187B_0 0x8189 RTL8187B Wireless Adapter product REALTEK RTL8196EU 0x8196 RTL8196EU product REALTEK RTL8187B_1 0x8197 RTL8187B Wireless Adapter product REALTEK RTL8187B_2 0x8198 RTL8187B Wireless Adapter product REALTEK RTL8188CUS 0x818a RTL8188CUS product REALTEK RTL8188CU_COMBO 0x8754 RTL8188CU product REALTEK RTL8191CU 0x8177 RTL8191CU product REALTEK RTL8192CU 0x8178 RTL8192CU product REALTEK RTL8192CE 0x817c RTL8192CE product REALTEK RTL8188RU_1 0x817d RTL8188RU product REALTEK RTL8188RU_3 0x817f RTL8188RU product REALTEK RTL8712 0x8712 RTL8712 product REALTEK RTL8713 0x8712 RTL8713 product REALTEK RTL8188RU_2 0x317f RTL8188RU product REALTEK RTL8192SU 0xc512 RTL8192SU /* RedOctane products */ product REDOCTANE DUMMY 0x0000 Dummy product product REDOCTANE GHMIDI 0x474b GH MIDI INTERFACE /* Renesas products */ product RENESAS RX610 0x0053 RX610 RX-Stick /* Ricoh products */ product RICOH VGPVCC2 0x1830 VGP-VCC2 Camera product RICOH VGPVCC3 0x1832 VGP-VCC3 Camera product RICOH VGPVCC2_2 0x1833 VGP-VCC2 Camera product RICOH VGPVCC2_3 0x1834 VGP-VCC2 Camera product RICOH VGPVCC7 0x183a VGP-VCC7 Camera product RICOH VGPVCC8 0x183b VGP-VCC8 Camera /* Reiner-SCT products */ product REINERSCT CYBERJACK_ECOM 0x0100 e-com cyberJack /* Roland products */ product ROLAND UA100 0x0000 UA-100 Audio I/F product ROLAND UM4 0x0002 UM-4 MIDI I/F product ROLAND SC8850 0x0003 SC-8850 MIDI Synth product ROLAND U8 0x0004 U-8 Audio I/F product ROLAND UM2 0x0005 UM-2 MIDI I/F product ROLAND SC8820 0x0007 SC-8820 MIDI Synth product ROLAND PC300 0x0008 PC-300 MIDI Keyboard product ROLAND UM1 0x0009 UM-1 MIDI I/F product ROLAND SK500 0x000b SK-500 MIDI Keyboard product ROLAND SCD70 0x000c SC-D70 MIDI Synth product ROLAND UM880N 0x0014 EDIROL UM-880 MIDI I/F (native) product ROLAND UM880G 0x0015 EDIROL UM-880 MIDI I/F (generic) product ROLAND SD90 0x0016 SD-90 MIDI Synth product ROLAND UM550 0x0023 UM-550 MIDI I/F product ROLAND SD20 0x0027 SD-20 MIDI Synth product ROLAND SD80 0x0029 SD-80 MIDI Synth product ROLAND UA700 0x002b UA-700 Audio I/F /* Rockfire products */ product ROCKFIRE GAMEPAD 0x2033 gamepad 203USB /* RATOC Systems products */ product RATOC REXUSB60 0xb000 REX-USB60 product RATOC REXUSB60F 0xb020 REX-USB60F /* RT system products */ product RTSYSTEMS CT29B 0x9e54 FTDI compatible adapter product RTSYSTEMS SERIAL_VX7 0x9e52 FTDI compatible adapter /* Sagem products */ product SAGEM USBSERIAL 0x0027 USB-Serial Controller product SAGEM XG760A 0x004a XG-760A product SAGEM XG76NA 0x0062 XG-76NA /* Samsung products */ product SAMSUNG WIS09ABGN 0x2018 WIS09ABGN Wireless LAN adapter product SAMSUNG ML6060 0x3008 ML-6060 laser printer product SAMSUNG YP_U2 0x5050 YP-U2 MP3 Player product SAMSUNG YP_U4 0x5092 YP-U4 MP3 Player product SAMSUNG I500 0x6601 I500 Palm USB Phone product SAMSUNG I330 0x8001 I330 phone cradle product SAMSUNG2 RT2870_1 0x2018 RT2870 /* Samsung Techwin products */ product SAMSUNG_TECHWIN DIGIMAX_410 0x000a Digimax 410 /* SanDisk products */ product SANDISK SDDR05A 0x0001 ImageMate SDDR-05a product SANDISK SDDR31 0x0002 ImageMate SDDR-31 product SANDISK SDDR05 0x0005 ImageMate SDDR-05 product SANDISK SDDR12 0x0100 ImageMate SDDR-12 product SANDISK SDDR09 0x0200 ImageMate SDDR-09 product SANDISK SDDR75 0x0810 ImageMate SDDR-75 product SANDISK SDCZ2_128 0x7100 Cruzer Mini 128MB product SANDISK SDCZ2_256 0x7104 Cruzer Mini 256MB product SANDISK SDCZ4_128 0x7112 Cruzer Micro 128MB product SANDISK SDCZ4_256 0x7113 Cruzer Micro 256MB product SANDISK IMAGEMATE_SDDR289 0xb6ba ImageMate SDDR-289 /* Sanwa Electric Instrument Co., Ltd. products */ product SANWA KB_USB2 0x0701 KB-USB2 multimeter cable /* Sanyo Electric products */ product SANYO SCP4900 0x0701 Sanyo SCP-4900 USB Phone /* ScanLogic products */ product SCANLOGIC SL11R 0x0002 SL11R IDE Adapter product SCANLOGIC 336CX 0x0300 Phantom 336CX - C3 scanner /* Schweitzer Engineering Laboratories products */ product SEL C662 0x0001 C662 Cable /* Sealevel products */ product SEALEVEL 2101 0x2101 FTDI compatible adapter product SEALEVEL 2102 0x2102 FTDI compatible adapter product SEALEVEL 2103 0x2103 FTDI compatible adapter product SEALEVEL 2104 0x2104 FTDI compatible adapter product SEALEVEL 2106 0x9020 FTDI compatible adapter product SEALEVEL 2201_1 0x2211 FTDI compatible adapter product SEALEVEL 2201_2 0x2221 FTDI compatible adapter product SEALEVEL 2202_1 0x2212 FTDI compatible adapter product SEALEVEL 2202_2 0x2222 FTDI compatible adapter product SEALEVEL 2203_1 0x2213 FTDI compatible adapter product SEALEVEL 2203_2 0x2223 FTDI compatible adapter product SEALEVEL 2401_1 0x2411 FTDI compatible adapter product SEALEVEL 2401_2 0x2421 FTDI compatible adapter product SEALEVEL 2401_3 0x2431 FTDI compatible adapter product SEALEVEL 2401_4 0x2441 FTDI compatible adapter product SEALEVEL 2402_1 0x2412 FTDI compatible adapter product SEALEVEL 2402_2 0x2422 FTDI compatible adapter product SEALEVEL 2402_3 0x2432 FTDI compatible adapter product SEALEVEL 2402_4 0x2442 FTDI compatible adapter product SEALEVEL 2403_1 0x2413 FTDI compatible adapter product SEALEVEL 2403_2 0x2423 FTDI compatible adapter product SEALEVEL 2403_3 0x2433 FTDI compatible adapter product SEALEVEL 2403_4 0x2443 FTDI compatible adapter product SEALEVEL 2801_1 0x2811 FTDI compatible adapter product SEALEVEL 2801_2 0x2821 FTDI compatible adapter product SEALEVEL 2801_3 0x2831 FTDI compatible adapter product SEALEVEL 2801_4 0x2841 FTDI compatible adapter product SEALEVEL 2801_5 0x2851 FTDI compatible adapter product SEALEVEL 2801_6 0x2861 FTDI compatible adapter product SEALEVEL 2801_7 0x2871 FTDI compatible adapter product SEALEVEL 2801_8 0x2881 FTDI compatible adapter product SEALEVEL 2802_1 0x2812 FTDI compatible adapter product SEALEVEL 2802_2 0x2822 FTDI compatible adapter product SEALEVEL 2802_3 0x2832 FTDI compatible adapter product SEALEVEL 2802_4 0x2842 FTDI compatible adapter product SEALEVEL 2802_5 0x2852 FTDI compatible adapter product SEALEVEL 2802_6 0x2862 FTDI compatible adapter product SEALEVEL 2802_7 0x2872 FTDI compatible adapter product SEALEVEL 2802_8 0x2882 FTDI compatible adapter product SEALEVEL 2803_1 0x2813 FTDI compatible adapter product SEALEVEL 2803_2 0x2823 FTDI compatible adapter product SEALEVEL 2803_3 0x2833 FTDI compatible adapter product SEALEVEL 2803_4 0x2843 FTDI compatible adapter product SEALEVEL 2803_5 0x2853 FTDI compatible adapter product SEALEVEL 2803_6 0x2863 FTDI compatible adapter product SEALEVEL 2803_7 0x2873 FTDI compatible adapter product SEALEVEL 2803_8 0x2883 FTDI compatible adapter /* Senao products */ product SENAO RT2870_3 0x0605 RT2870 product SENAO RT2870_4 0x0615 RT2870 product SENAO NUB8301 0x2000 NUB-8301 product SENAO RT2870_1 0x9701 RT2870 product SENAO RT2870_2 0x9702 RT2870 product SENAO RT3070 0x9703 RT3070 product SENAO RT3071 0x9705 RT3071 product SENAO RT3072_1 0x9706 RT3072 product SENAO RT3072_2 0x9707 RT3072 product SENAO RT3072_3 0x9708 RT3072 product SENAO RT3072_4 0x9709 RT3072 product SENAO RT3072_5 0x9801 RT3072 product SENAO RTL8192SU_1 0x9603 RTL8192SU product SENAO RTL8192SU_2 0x9605 RTL8192SU /* ShanTou products */ product SHANTOU ST268 0x0268 ST268 product SHANTOU DM9601 0x9601 DM 9601 product SHANTOU ADM8515 0x8515 ADM8515 /* Shark products */ product SHARK PA 0x0400 Pocket Adapter /* Sharp products */ product SHARP SL5500 0x8004 Zaurus SL-5500 PDA product SHARP SLA300 0x8005 Zaurus SL-A300 PDA product SHARP SL5600 0x8006 Zaurus SL-5600 PDA product SHARP SLC700 0x8007 Zaurus SL-C700 PDA product SHARP SLC750 0x9031 Zaurus SL-C750 PDA product SHARP WZERO3ES 0x9123 W-ZERO3 ES Smartphone product SHARP WZERO3ADES 0x91ac Advanced W-ZERO3 ES Smartphone product SHARP WILLCOM03 0x9242 WILLCOM03 /* Shuttle Technology products */ product SHUTTLE EUSB 0x0001 E-USB Bridge product SHUTTLE EUSCSI 0x0002 eUSCSI Bridge product SHUTTLE SDDR09 0x0003 ImageMate SDDR09 product SHUTTLE EUSBCFSM 0x0005 eUSB SmartMedia / CompactFlash Adapter product SHUTTLE ZIOMMC 0x0006 eUSB MultiMediaCard Adapter product SHUTTLE HIFD 0x0007 Sony Hifd product SHUTTLE EUSBATAPI 0x0009 eUSB ATA/ATAPI Adapter product SHUTTLE CF 0x000a eUSB CompactFlash Adapter product SHUTTLE EUSCSI_B 0x000b eUSCSI Bridge product SHUTTLE EUSCSI_C 0x000c eUSCSI Bridge product SHUTTLE CDRW 0x0101 CD-RW Device product SHUTTLE EUSBORCA 0x0325 eUSB ORCA Quad Reader /* Siemens products */ product SIEMENS SPEEDSTREAM 0x1001 SpeedStream product SIEMENS SPEEDSTREAM22 0x1022 SpeedStream 1022 product SIEMENS2 WLL013 0x001b WLL013 product SIEMENS2 ES75 0x0034 GSM module MC35 product SIEMENS2 WL54G 0x3c06 54g USB Network Adapter product SIEMENS3 SX1 0x0001 SX1 product SIEMENS3 X65 0x0003 X65 product SIEMENS3 X75 0x0004 X75 product SIEMENS3 EF81 0x0005 EF81 /* Sierra Wireless products */ product SIERRA EM5625 0x0017 EM5625 product SIERRA MC5720_2 0x0018 MC5720 product SIERRA MC5725 0x0020 MC5725 product SIERRA AIRCARD580 0x0112 Sierra Wireless AirCard 580 product SIERRA AIRCARD595 0x0019 Sierra Wireless AirCard 595 product SIERRA AC595U 0x0120 Sierra Wireless AirCard 595U product SIERRA AC597E 0x0021 Sierra Wireless AirCard 597E product SIERRA EM5725 0x0022 EM5725 product SIERRA C597 0x0023 Sierra Wireless Compass 597 product SIERRA MC5727 0x0024 MC5727 product SIERRA T598 0x0025 T598 product SIERRA T11 0x0026 T11 product SIERRA AC402 0x0027 AC402 product SIERRA MC5728 0x0028 MC5728 product SIERRA E0029 0x0029 E0029 product SIERRA AIRCARD580 0x0112 Sierra Wireless AirCard 580 product SIERRA AC595U 0x0120 Sierra Wireless AirCard 595U product SIERRA MC5720 0x0218 MC5720 Wireless Modem product SIERRA MINI5725 0x0220 Sierra Wireless miniPCI 5275 product SIERRA MC5727_2 0x0224 MC5727 product SIERRA MC8755_2 0x6802 MC8755 product SIERRA MC8765 0x6803 MC8765 product SIERRA MC8755 0x6804 MC8755 product SIERRA MC8765_2 0x6805 MC8765 product SIERRA MC8755_4 0x6808 MC8755 product SIERRA MC8765_3 0x6809 MC8765 product SIERRA AC875U 0x6812 AC875U HSDPA USB Modem product SIERRA MC8755_3 0x6813 MC8755 HSDPA product SIERRA MC8775_2 0x6815 MC8775 product SIERRA MC8775 0x6816 MC8775 product SIERRA AC875 0x6820 Sierra Wireless AirCard 875 product SIERRA AC875U_2 0x6821 AC875U product SIERRA AC875E 0x6822 AC875E product SIERRA MC8780 0x6832 MC8780 product SIERRA MC8781 0x6833 MC8781 product SIERRA MC8780_2 0x6834 MC8780 product SIERRA MC8781_2 0x6835 MC8781 product SIERRA MC8780_3 0x6838 MC8780 product SIERRA MC8781_3 0x6839 MC8781 product SIERRA MC8785 0x683A MC8785 product SIERRA MC8785_2 0x683B MC8785 product SIERRA MC8790 0x683C MC8790 product SIERRA MC8791 0x683D MC8791 product SIERRA MC8792 0x683E MC8792 product SIERRA AC880 0x6850 Sierra Wireless AirCard 880 product SIERRA AC881 0x6851 Sierra Wireless AirCard 881 product SIERRA AC880E 0x6852 Sierra Wireless AirCard 880E product SIERRA AC881E 0x6853 Sierra Wireless AirCard 881E product SIERRA AC880U 0x6855 Sierra Wireless AirCard 880U product SIERRA AC881U 0x6856 Sierra Wireless AirCard 881U product SIERRA AC885E 0x6859 AC885E product SIERRA AC885E_2 0x685A AC885E product SIERRA AC885U 0x6880 Sierra Wireless AirCard 885U product SIERRA C888 0x6890 C888 product SIERRA C22 0x6891 C22 product SIERRA E6892 0x6892 E6892 product SIERRA E6893 0x6893 E6893 product SIERRA MC8700 0x68A3 MC8700 product SIERRA AIRCARD875 0x6820 Aircard 875 HSDPA product SIERRA AC313U 0x68aa Sierra Wireless AirCard 313U product SIERRA TRUINSTALL 0x0fff Aircard Tru Installer /* Sigmatel products */ product SIGMATEL WBT_3052 0x4200 WBT-3052 IrDA/USB Bridge product SIGMATEL I_BEAD100 0x8008 i-Bead 100 MP3 Player /* SIIG products */ /* Also: Omnidirectional Control Technology products */ product SIIG DIGIFILMREADER 0x0004 DigiFilm-Combo Reader product SIIG WINTERREADER 0x0330 WINTERREADER Reader product SIIG2 DK201 0x0103 FTDI compatible adapter product SIIG2 USBTOETHER 0x0109 USB TO Ethernet product SIIG2 US2308 0x0421 Serial /* Silicom products */ product SILICOM U2E 0x0001 U2E product SILICOM GPE 0x0002 Psion Gold Port Ethernet /* SI Labs */ product SILABS VSTABI 0x0f91 VStabi Controller product SILABS ARKHAM_DS101_M 0x1101 Arkham DS101 Monitor product SILABS ARKHAM_DS101_A 0x1601 Arkham DS101 Adapter product SILABS BSM7DUSB 0x800a SPORTident BSM7-D USB product SILABS POLOLU 0x803b Pololu Serial product SILABS CYGNAL_DEBUG 0x8044 Cygnal Debug Adapter product SILABS SB_PARAMOUNT_ME 0x8043 Software Bisque Paramount ME product SILABS SAEL 0x8053 SA-EL USB product SILABS GSM2228 0x8054 Enfora GSM2228 USB product SILABS ARGUSISP 0x8066 Argussoft ISP product SILABS IMS_USB_RS422 0x806f IMS USB-RS422 product SILABS CRUMB128 0x807a Crumb128 board product SILABS OPTRIS_MSPRO 0x80c4 Optris MSpro LT Thermometer product SILABS DEGREE 0x80ca Degree Controls Inc product SILABS TRACIENT 0x80dd Tracient RFID product SILABS TRAQMATE 0x80ed Track Systems Traqmate product SILABS SUUNTO 0x80f6 Suunto Sports Instrument product SILABS ARYGON_MIFARE 0x8115 Arygon Mifare RFID reader product SILABS BURNSIDE 0x813d Burnside Telecon Deskmobile product SILABS TAMSMASTER 0x813f Tams Master Easy Control product SILABS WMRBATT 0x814a WMR RIGblaster Plug&Play product SILABS WMRRIGBLASTER 0x814a WMR RIGblaster Plug&Play product SILABS WMRRIGTALK 0x814b WMR RIGtalk RT1 product SILABS B_G_H3000 0x8156 B&G H3000 Data Cable product SILABS HELICOM 0x815e Helicomm IP-Link 1220-DVM product SILABS HAMLINKUSB 0x815f Timewave HamLinkUSB product SILABS AVIT_USB_TTL 0x818b AVIT Research USB-TTL product SILABS MJS_TOSLINK 0x819f MJS USB-TOSLINK product SILABS WAVIT 0x81a6 ThinkOptics WavIt product SILABS MULTIPLEX_RC 0x81a9 Multiplex RC adapter product SILABS MSD_DASHHAWK 0x81ac MSD DashHawk product SILABS INSYS_MODEM 0x81ad INSYS Modem product SILABS LIPOWSKY_JTAG 0x81c8 Lipowsky Baby-JTAG product SILABS LIPOWSKY_LIN 0x81e2 Lipowsky Baby-LIN product SILABS AEROCOMM 0x81e7 Aerocomm Radio product SILABS ZEPHYR_BIO 0x81e8 Zephyr Bioharness product SILABS EMS_C1007 0x81f2 EMS C1007 HF RFID controller product SILABS LIPOWSKY_HARP 0x8218 Lipowsky HARP-1 product SILABS C2_EDGE_MODEM 0x822b Commander 2 EDGE(GSM) Modem product SILABS CYGNAL_GPS 0x826b Cygnal Fasttrax GPS product SILABS TELEGESIS_ETRX2 0x8293 Telegesis ETRX2USB product SILABS PROCYON_AVS 0x82f9 Procyon AVS product SILABS MC35PU 0x8341 MC35pu product SILABS CYGNAL 0x8382 Cygnal product SILABS AMBER_AMB2560 0x83a8 Amber Wireless AMB2560 product SILABS DEKTEK_DTAPLUS 0x83d8 DekTec DTA Plus VHF/UHF Booster product SILABS KYOCERA_GPS 0x8411 Kyocera GPS product SILABS IRZ_SG10 0x8418 IRZ SG-10 GSM/GPRS Modem product SILABS BEI_VCP 0x846e BEI USB Sensor (VCP) product SILABS BALLUFF_RFID 0x8477 Balluff RFID reader product SILABS AC_SERV_IBUS 0x85ea AC-Services IBUS Interface product SILABS AC_SERV_CIS 0x85eb AC-Services CIS-IBUS product SILABS V_PREON32 0x85f8 Virtenio Preon32 product SILABS AC_SERV_CAN 0x8664 AC-Services CAN Interface product SILABS AC_SERV_OBD 0x8665 AC-Services OBD Interface product SILABS MMB_ZIGBEE 0x88a4 MMB Networks ZigBee product SILABS INGENI_ZIGBEE 0x88a5 Planet Innovation Ingeni ZigBee product SILABS CP2102 0xea60 SILABS USB UART product SILABS CP210X_2 0xea61 CP210x Serial product SILABS CP210X_3 0xea70 CP210x Serial product SILABS CP210X_4 0xea80 CP210x Serial product SILABS INFINITY_MIC 0xea71 Infinity GPS-MIC-1 Radio Monophone product SILABS USBSCOPE50 0xf001 USBscope50 product SILABS USBWAVE12 0xf002 USBwave12 product SILABS USBPULSE100 0xf003 USBpulse100 product SILABS USBCOUNT50 0xf004 USBcount50 product SILABS2 DCU11CLONE 0xaa26 DCU-11 clone product SILABS3 GPRS_MODEM 0xea61 GPRS Modem product SILABS4 100EU_MODEM 0xea61 GPRS Modem 100EU /* Silicon Portals Inc. */ product SILICONPORTALS YAPPH_NF 0x0200 YAP Phone (no firmware) product SILICONPORTALS YAPPHONE 0x0201 YAP Phone /* Sirius Technologies products */ product SIRIUS ROADSTER 0x0001 NetComm Roadster II 56 USB /* Sitecom products */ product SITECOM LN029 0x182d USB 2.0 Ethernet product SITECOM SERIAL 0x2068 USB to serial cable (v2) product SITECOM2 WL022 0x182d WL-022 /* Sitecom Europe products */ product SITECOMEU RT2870_1 0x0017 RT2870 product SITECOMEU WL168V1 0x000d WL-168 v1 product SITECOMEU LN030 0x0021 MCS7830 product SITECOMEU WL168V4 0x0028 WL-168 v4 product SITECOMEU RT2870_2 0x002b RT2870 product SITECOMEU RT2870_3 0x002c RT2870 product SITECOMEU RT2870_4 0x002d RT2870 product SITECOMEU RT2770 0x0039 RT2770 product SITECOMEU RT3070_2 0x003b RT3070 product SITECOMEU RT3070_3 0x003c RT3070 product SITECOMEU RT3070_4 0x003d RT3070 product SITECOMEU RT3070 0x003e RT3070 product SITECOMEU WL608 0x003f WL-608 product SITECOMEU RT3071 0x0040 RT3071 product SITECOMEU RT3072_1 0x0041 RT3072 product SITECOMEU RT3072_2 0x0042 RT3072 product SITECOMEU WL353 0x0045 WL-353 product SITECOMEU RT3072_3 0x0047 RT3072 product SITECOMEU RT3072_4 0x0048 RT3072 product SITECOMEU RT3072_5 0x004a RT3072 product SITECOMEU WL349V1 0x004b WL-349 v1 product SITECOMEU RT3072_6 0x004d RT3072 product SITECOMEU RTL8188CU_1 0x0052 RTL8188CU product SITECOMEU RTL8188CU_2 0x005c RTL8188CU product SITECOMEU RTL8192CU 0x0061 RTL8192CU product SITECOMEU LN032 0x0072 LN-032 product SITECOMEU LN028 0x061c LN-028 product SITECOMEU WL113 0x9071 WL-113 product SITECOMEU ZD1211B 0x9075 ZD1211B product SITECOMEU WL172 0x90ac WL-172 product SITECOMEU WL113R2 0x9712 WL-113 rev 2 /* Skanhex Technology products */ product SKANHEX MD_7425 0x410a MD 7425 Camera product SKANHEX SX_520Z 0x5200 SX 520z Camera /* Smart Technologies products */ product SMART PL2303 0x2303 Serial adapter /* SmartBridges products */ product SMARTBRIDGES SMARTLINK 0x0001 SmartLink USB Ethernet product SMARTBRIDGES SMARTNIC 0x0003 smartNIC 2 PnP Ethernet /* SMC products */ product SMC 2102USB 0x0100 10Mbps Ethernet product SMC 2202USB 0x0200 10/100 Ethernet product SMC 2206USB 0x0201 EZ Connect USB Ethernet product SMC 2862WG 0xee13 EZ Connect Wireless Adapter product SMC2 2020HUB 0x2020 USB Hub product SMC2 2514HUB 0x2514 USB Hub product SMC3 2662WUSB 0xa002 2662W-AR Wireless product SMC2 LAN9500_ETH 0x9500 USB/Ethernet product SMC2 LAN9505_ETH 0x9505 USB/Ethernet product SMC2 LAN9530_ETH 0x9530 USB/Ethernet product SMC2 LAN9730_ETH 0x9730 USB/Ethernet product SMC2 LAN9500_SAL10 0x9900 USB/Ethernet product SMC2 LAN9505_SAL10 0x9901 USB/Ethernet product SMC2 LAN9500A_SAL10 0x9902 USB/Ethernet product SMC2 LAN9505A_SAL10 0x9903 USB/Ethernet product SMC2 LAN9514_SAL10 0x9904 USB/Ethernet product SMC2 LAN9500A_HAL 0x9905 USB/Ethernet product SMC2 LAN9505A_HAL 0x9906 USB/Ethernet product SMC2 LAN9500_ETH_2 0x9907 USB/Ethernet product SMC2 LAN9500A_ETH_2 0x9908 USB/Ethernet product SMC2 LAN9514_ETH_2 0x9909 USB/Ethernet product SMC2 LAN9500A_ETH 0x9e00 USB/Ethernet product SMC2 LAN9505A_ETH 0x9e01 USB/Ethernet product SMC2 LAN89530_ETH 0x9e08 USB/Ethernet product SMC2 LAN9514_ETH 0xec00 USB/Ethernet /* SOHOware products */ product SOHOWARE NUB100 0x9100 10/100 USB Ethernet product SOHOWARE NUB110 0x9110 10/100 USB Ethernet /* SOLID YEAR products */ product SOLIDYEAR KEYBOARD 0x2101 Solid Year USB keyboard /* SONY products */ product SONY DSC 0x0010 DSC cameras product SONY MS_NW_MS7 0x0025 Memorystick NW-MS7 product SONY PORTABLE_HDD_V2 0x002b Portable USB Harddrive V2 product SONY MSACUS1 0x002d Memorystick MSAC-US1 product SONY HANDYCAM 0x002e Handycam product SONY MSC 0x0032 MSC memory stick slot product SONY CLIE_35 0x0038 Sony Clie v3.5 product SONY MS_PEG_N760C 0x0058 PEG N760c Memorystick product SONY CLIE_40 0x0066 Sony Clie v4.0 product SONY MS_MSC_U03 0x0069 Memorystick MSC-U03 product SONY CLIE_40_MS 0x006d Sony Clie v4.0 Memory Stick slot product SONY CLIE_S360 0x0095 Sony Clie s360 product SONY CLIE_41_MS 0x0099 Sony Clie v4.1 Memory Stick slot product SONY CLIE_41 0x009a Sony Clie v4.1 product SONY CLIE_NX60 0x00da Sony Clie nx60 product SONY CLIE_TH55 0x0144 Sony Clie th55 product SONY CLIE_TJ37 0x0169 Sony Clie tj37 product SONY RF_RECEIVER 0x01db Sony RF mouse/kbd Receiver VGP-WRC1 product SONY QN3 0x0437 Sony QN3 CMD-Jxx phone cable /* Sony Ericsson products */ product SONYERICSSON DCU10 0x0528 DCU-10 Phone Data Cable product SONYERICSSON DATAPILOT 0x2003 Datapilot Phone Cable /* SOURCENEXT products */ product SOURCENEXT KEIKAI8 0x039f KeikaiDenwa 8 product SOURCENEXT KEIKAI8_CHG 0x012e KeikaiDenwa 8 with charger /* SparkLAN products */ product SPARKLAN RT2573 0x0004 RT2573 product SPARKLAN RT2870_1 0x0006 RT2870 product SPARKLAN RT3070 0x0010 RT3070 /* Soundgraph products */ product SOUNDGRAPH IMON_VFD 0x0044 Antec Veris Elite VFD Panel, Knob, and Remote product SOUNDGRAPH SSTONE_LC16 0xffdc Silverstone LC16 VFD Panel, Knob, and Remote /* Speed Dragon Multimedia products */ product SPEEDDRAGON MS3303H 0x110b MS3303H Serial /* Sphairon Access Systems GmbH products */ product SPHAIRON UB801R 0x0110 UB801R /* Stelera Wireless products */ product STELERA ZEROCD 0x1000 Zerocd Installer product STELERA C105 0x1002 Stelera/Bandrish C105 USB product STELERA E1003 0x1003 3G modem product STELERA E1004 0x1004 3G modem product STELERA E1005 0x1005 3G modem product STELERA E1006 0x1006 3G modem product STELERA E1007 0x1007 3G modem product STELERA E1008 0x1008 3G modem product STELERA E1009 0x1009 3G modem product STELERA E100A 0x100a 3G modem product STELERA E100B 0x100b 3G modem product STELERA E100C 0x100c 3G modem product STELERA E100D 0x100d 3G modem product STELERA E100E 0x100e 3G modem product STELERA E100F 0x100f 3G modem product STELERA E1010 0x1010 3G modem product STELERA E1011 0x1011 3G modem product STELERA E1012 0x1012 3G modem /* STMicroelectronics products */ product STMICRO BIOCPU 0x2016 Biometric Coprocessor product STMICRO COMMUNICATOR 0x7554 USB Communicator product STMICRO ST72682 0xfada USB 2.0 Flash drive controller /* STSN products */ product STSN STSN0001 0x0001 Internet Access Device /* SUN Corporation products */ product SUNTAC DS96L 0x0003 SUNTAC U-Cable type D2 product SUNTAC PS64P1 0x0005 SUNTAC U-Cable type P1 product SUNTAC VS10U 0x0009 SUNTAC Slipper U product SUNTAC IS96U 0x000a SUNTAC Ir-Trinity product SUNTAC AS64LX 0x000b SUNTAC U-Cable type A3 product SUNTAC AS144L4 0x0011 SUNTAC U-Cable type A4 /* Sun Microsystems products */ product SUN KEYBOARD_TYPE_6 0x0005 Type 6 USB keyboard product SUN KEYBOARD_TYPE_7 0x00a2 Type 7 USB keyboard /* XXX The above is a North American PC style keyboard possibly */ product SUN MOUSE 0x0100 Type 6 USB mouse product SUN KBD_HUB 0x100e Kbd Hub /* Sunplus Innovation Technology Inc. products */ product SUNPLUS USBMOUSE 0x0007 USB Optical Mouse /* Super Top products */ product SUPERTOP IDE 0x6600 USB-IDE product SUPERTOP FLASHDRIVE 0x121c extrememory Snippy /* Syntech products */ product SYNTECH CPT8001C 0x0001 CPT-8001C Barcode scanner product SYNTECH CYPHERLAB100 0x1000 CipherLab USB Barcode Scanner /* Teclast products */ product TECLAST TLC300 0x3203 USB Media Player /* Testo products */ product TESTO USB_INTERFACE 0x0001 FTDI compatible adapter /* TexTech products */ product TEXTECH DUMMY 0x0000 Dummy product product TEXTECH U2M_1 0x0101 Textech USB MIDI cable product TEXTECH U2M_2 0x1806 Textech USB MIDI cable /* The Mobility Lab products */ product TML USB_SERIAL 0x0064 FTDI compatible adapter /* Thurlby Thandar Instrument products */ product TTI QL355P 0x03e8 FTDI compatible adapter /* Supra products */ product DIAMOND2 SUPRAEXPRESS56K 0x07da Supra Express 56K modem product DIAMOND2 SUPRA2890 0x0b4a SupraMax 2890 56K Modem product DIAMOND2 RIO600USB 0x5001 Rio 600 USB product DIAMOND2 RIO800USB 0x5002 Rio 800 USB /* Surecom Technology products */ product SURECOM EP9001G2A 0x11f2 EP-9001-G rev 2A product SURECOM RT2570 0x11f3 RT2570 product SURECOM RT2573 0x31f3 RT2573 /* Sweex products */ product SWEEX ZD1211 0x1809 ZD1211 product SWEEX2 LW153 0x0153 LW153 product SWEEX2 LW154 0x0154 LW154 product SWEEX2 LW303 0x0302 LW303 product SWEEX2 LW313 0x0313 LW313 /* System TALKS, Inc. */ product SYSTEMTALKS SGCX2UL 0x1920 SGC-X2UL /* Tapwave products */ product TAPWAVE ZODIAC 0x0100 Zodiac /* Taugagreining products */ product TAUGA CAMERAMATE 0x0005 CameraMate (DPCM_USB) /* TCTMobile products */ product TCTMOBILE X060S 0x0000 X060S 3G modem product TCTMOBILE X080S 0xf000 X080S 3G modem /* TDK products */ product TDK UPA9664 0x0115 USB-PDC Adapter UPA9664 product TDK UCA1464 0x0116 USB-cdmaOne Adapter UCA1464 product TDK UHA6400 0x0117 USB-PHS Adapter UHA6400 product TDK UPA6400 0x0118 USB-PHS Adapter UPA6400 product TDK BT_DONGLE 0x0309 Bluetooth USB dongle /* TEAC products */ product TEAC FD05PUB 0x0000 FD-05PUB floppy /* Tekram Technology products */ product TEKRAM QUICKWLAN 0x1630 QuickWLAN product TEKRAM ZD1211_1 0x5630 ZD1211 product TEKRAM ZD1211_2 0x6630 ZD1211 /* Telex Communications products */ product TELEX MIC1 0x0001 Enhanced USB Microphone /* Telit products */ product TELIT UC864E 0x1003 UC864E 3G modem product TELIT UC864G 0x1004 UC864G 3G modem /* Ten X Technology, Inc. */ product TENX UAUDIO0 0xf211 USB audio headset /* Texas Intel products */ product TI UTUSB41 0x1446 UT-USB41 hub product TI TUSB2046 0x2046 TUSB2046 hub /* Thrustmaster products */ product THRUST FUSION_PAD 0xa0a3 Fusion Digital Gamepad /* TLayTech products */ product TLAYTECH TEU800 0x1682 TEU800 3G modem /* Topre Corporation products */ product TOPRE HHKB 0x0100 HHKB Professional /* Toshiba Corporation products */ product TOSHIBA POCKETPC_E740 0x0706 PocketPC e740 product TOSHIBA RT3070 0x0a07 RT3070 product TOSHIBA G450 0x0d45 G450 modem product TOSHIBA HSDPA 0x1302 G450 modem product TOSHIBA TRANSMEMORY 0x6545 USB ThumbDrive /* Trek Technology products */ product TREK THUMBDRIVE 0x1111 ThumbDrive product TREK MEMKEY 0x8888 IBM USB Memory Key product TREK THUMBDRIVE_8MB 0x9988 ThumbDrive_8MB /* TRENDnet products */ product TRENDNET RTL8192CU 0x624d RTL8192CU product TRENDNET TEW646UBH 0x646b TEW-646UBH product TRENDNET RTL8188CU 0x648b RTL8188CU /* Tripp-Lite products */ product TRIPPLITE U209 0x2008 Serial /* Trumpion products */ product TRUMPION T33520 0x1001 T33520 USB Flash Card Controller product TRUMPION C3310 0x1100 Comotron C3310 MP3 player product TRUMPION MP3 0x1200 MP3 player /* TwinMOS */ product TWINMOS G240 0xa006 G240 product TWINMOS MDIV 0x1325 Memory Disk IV /* Ubiquam products */ product UBIQUAM UALL 0x3100 CDMA 1xRTT USB Modem (U-100/105/200/300/520) /* Ultima products */ product ULTIMA 1200UBPLUS 0x4002 1200 UB Plus scanner /* UMAX products */ product UMAX ASTRA1236U 0x0002 Astra 1236U Scanner product UMAX ASTRA1220U 0x0010 Astra 1220U Scanner product UMAX ASTRA2000U 0x0030 Astra 2000U Scanner product UMAX ASTRA2100U 0x0130 Astra 2100U Scanner product UMAX ASTRA2200U 0x0230 Astra 2200U Scanner product UMAX ASTRA3400 0x0060 Astra 3400 Scanner /* U-MEDIA Communications products */ product UMEDIA TEW444UBEU 0x3006 TEW-444UB EU product UMEDIA TEW444UBEU_NF 0x3007 TEW-444UB EU (no firmware) product UMEDIA TEW429UB_A 0x300a TEW-429UB_A product UMEDIA TEW429UB 0x300b TEW-429UB product UMEDIA TEW429UBC1 0x300d TEW-429UB C1 product UMEDIA RT2870_1 0x300e RT2870 product UMEDIA ALL0298V2 0x3204 ALL0298 v2 product UMEDIA AR5523_2 0x3205 AR5523 product UMEDIA AR5523_2_NF 0x3206 AR5523 (no firmware) /* Universal Access products */ product UNIACCESS PANACHE 0x0101 Panache Surf USB ISDN Adapter /* Unknown products */ product UNKNOWN4 NF_RIC 0x0001 FTDI compatible adapter /* USI products */ product USI MC60 0x10c5 MC60 Serial /* U.S. Robotics products */ product USR USR5422 0x0118 USR5422 WLAN product USR USR5423 0x0121 USR5423 WLAN /* VIA Technologies products */ product VIA USB2IDEBRIDGE 0x6204 USB 2.0 IDE Bridge /* VIA Labs */ product VIALABS USB30SATABRIDGE 0x0700 USB 3.0 SATA Bridge /* Vaisala products */ product VAISALA CABLE 0x0200 USB Interface cable /* Vertex products */ product VERTEX VW110L 0x0100 Vertex VW110L modem /* VidzMedia products */ product VIDZMEDIA MONSTERTV 0x4fb1 MonsterTV P2H /* Vision products */ product VISION VC6452V002 0x0002 CPiA Camera /* Visioneer products */ product VISIONEER 7600 0x0211 OneTouch 7600 product VISIONEER 5300 0x0221 OneTouch 5300 product VISIONEER 3000 0x0224 Scanport 3000 product VISIONEER 6100 0x0231 OneTouch 6100 product VISIONEER 6200 0x0311 OneTouch 6200 product VISIONEER 8100 0x0321 OneTouch 8100 product VISIONEER 8600 0x0331 OneTouch 8600 /* Vivitar products */ product VIVITAR 35XX 0x0003 Vivicam 35Xx /* VTech products */ product VTECH RT2570 0x3012 RT2570 product VTECH ZD1211B 0x3014 ZD1211B /* Wacom products */ product WACOM CT0405U 0x0000 CT-0405-U Tablet product WACOM GRAPHIRE 0x0010 Graphire product WACOM GRAPHIRE3_4X5 0x0013 Graphire 3 4x5 product WACOM INTUOSA5 0x0021 Intuos A5 product WACOM GD0912U 0x0022 Intuos 9x12 Graphics Tablet /* WAGO Kontakttechnik GmbH products */ product WAGO SERVICECABLE 0x07a6 USB Service Cable 750-923 /* WaveSense products */ product WAVESENSE JAZZ 0xaaaa Jazz blood glucose meter /* WCH products */ product WCH CH341SER 0x5523 CH341/CH340 USB-Serial Bridge product WCH2 DUMMY 0x0000 Dummy product product WCH2 CH341SER_2 0x5523 CH341/CH340 USB-Serial Bridge product WCH2 CH341SER 0x7523 CH341/CH340 USB-Serial Bridge product WCH2 U2M 0X752d CH345 USB2.0-MIDI /* West Mountain Radio products */ product WESTMOUNTAIN RIGBLASTER_ADVANTAGE 0x0003 RIGblaster Advantage /* Western Digital products */ product WESTERN COMBO 0x0200 Firewire USB Combo product WESTERN EXTHDD 0x0400 External HDD product WESTERN HUB 0x0500 USB HUB product WESTERN MYBOOK 0x0901 MyBook External HDD -product WESTERN MYPASSWORD 0x0704 MyPassword External HDD -product WESTERN MYPASSPORT 0x0748 MyPassport External HDD +product WESTERN MYPASSPORT_00 0x0704 MyPassport External HDD +product WESTERN MYPASSPORT_01 0x0746 MyPassport External HDD +product WESTERN MYPASSPORT_02 0x0748 MyPassport External HDD +product WESTERN MYPASSPORT_03 0x074A MyPassport External HDD +product WESTERN MYPASSPORT_04 0x074C MyPassport External HDD +product WESTERN MYPASSPORT_05 0x074E MyPassport External HDD +product WESTERN MYPASSPORT_06 0x07A6 MyPassport External HDD +product WESTERN MYPASSPORT_07 0x07A8 MyPassport External HDD +product WESTERN MYPASSPORT_08 0x07AA MyPassport External HDD +product WESTERN MYPASSPORT_09 0x07AC MyPassport External HDD +product WESTERN MYPASSPORT_10 0x07AE MyPassport External HDD +product WESTERN MYPASSPORTES_00 0x070A MyPassport Essential External HDD +product WESTERN MYPASSPORTES_01 0x071A MyPassport Essential External HDD +product WESTERN MYPASSPORTES_02 0x0730 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_03 0x0732 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_04 0x0740 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_05 0x0742 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_06 0x0750 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_07 0x0752 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_08 0x07A0 MyPassport Essential External HDD +product WESTERN MYPASSPORTES_09 0x07A2 MyPassport Essential External HDD /* WeTelecom products */ product WETELECOM WM_D200 0x6801 WM-D200 /* WIENER Plein & Baus GmbH products */ product WIENERPLEINBAUS PL512 0x0010 PL512 PSU product WIENERPLEINBAUS RCM 0x0011 RCM Remote Control product WIENERPLEINBAUS MPOD 0x0012 MPOD PSU product WIENERPLEINBAUS CML 0x0015 CML Data Logger /* Windbond Electronics */ product WINBOND UH104 0x5518 4-port USB Hub /* WinMaxGroup products */ product WINMAXGROUP FLASH64MC 0x6660 USB Flash Disk 64M-C /* Wistron NeWeb products */ product WISTRONNEWEB UR045G 0x0427 PrismGT USB 2.0 WLAN product WISTRONNEWEB UR055G 0x0711 UR055G product WISTRONNEWEB AR5523_1 0x0826 AR5523 product WISTRONNEWEB AR5523_1_NF 0x0827 AR5523 (no firmware) product WISTRONNEWEB AR5523_2 0x082a AR5523 product WISTRONNEWEB AR5523_2_NF 0x0829 AR5523 (no firmware) /* Xerox products */ product XEROX WCM15 0xffef WorkCenter M15 /* Xirlink products */ product XIRLINK PCCAM 0x8080 IBM PC Camera /* Xyratex products */ product XYRATEX PRISM_GT_1 0x2000 PrismGT USB 2.0 WLAN product XYRATEX PRISM_GT_2 0x2002 PrismGT USB 2.0 WLAN /* Yamaha products */ product YAMAHA UX256 0x1000 UX256 MIDI I/F product YAMAHA UX96 0x1008 UX96 MIDI I/F product YAMAHA RPU200 0x3104 RP-U200 product YAMAHA RTA54I 0x4000 NetVolante RTA54i Broadband&ISDN Router product YAMAHA RTW65B 0x4001 NetVolante RTW65b Broadband Wireless Router product YAMAHA RTW65I 0x4002 NetVolante RTW65i Broadband&ISDN Wireless Router product YAMAHA RTA55I 0x4004 NetVolante RTA55i Broadband VoIP Router /* Yano products */ product YANO U640MO 0x0101 U640MO-03 product YANO FW800HD 0x05fc METALWEAR-HDD /* Y.C. Cable products */ product YCCABLE PL2303 0x0fba PL2303 Serial /* Y-E Data products */ product YEDATA FLASHBUSTERU 0x0000 Flashbuster-U /* Yiso Wireless Co. products */ product YISO C893 0xc893 CDMA 2000 1xEVDO PC Card /* Z-Com products */ product ZCOM M4Y750 0x0001 M4Y-750 product ZCOM XI725 0x0002 XI-725/726 product ZCOM XI735 0x0005 XI-735 product ZCOM XG703A 0x0008 PrismGT USB 2.0 WLAN product ZCOM ZD1211 0x0011 ZD1211 product ZCOM AR5523 0x0012 AR5523 product ZCOM AR5523_NF 0x0013 AR5523 driver (no firmware) product ZCOM XM142 0x0015 XM-142 product ZCOM ZD1211B 0x001a ZD1211B product ZCOM RT2870_1 0x0022 RT2870 product ZCOM RT2870_2 0x0025 RT2870 /* Zinwell products */ product ZINWELL RT2570 0x0260 RT2570 product ZINWELL RT2870_1 0x0280 RT2870 product ZINWELL RT2870_2 0x0282 RT2870 product ZINWELL RT3072_1 0x0283 RT3072 product ZINWELL RT3072_2 0x0284 RT3072 product ZINWELL RT3070 0x5257 RT3070 /* Zoom Telephonics, Inc. products */ product ZOOM 2986L 0x9700 2986L Fax modem /* Zoran Microelectronics products */ product ZORAN EX20DSC 0x4343 Digital Camera EX-20 DSC /* Zydas Technology Corporation products */ product ZYDAS ZD1211 0x1211 ZD1211 WLAN abg product ZYDAS ZD1211B 0x1215 ZD1211B /* ZyXEL Communication Co. products */ product ZYXEL OMNI56K 0x1500 Omni 56K Plus product ZYXEL 980N 0x2011 Scorpion-980N keyboard product ZYXEL ZYAIRG220 0x3401 ZyAIR G-220 product ZYXEL G200V2 0x3407 G-200 v2 product ZYXEL AG225H 0x3409 AG-225H product ZYXEL M202 0x340a M-202 product ZYXEL G220V2 0x340f G-220 v2 product ZYXEL G202 0x3410 G-202 product ZYXEL RT2870_1 0x3416 RT2870 product ZYXEL RT2870_2 0x341a RT2870 product ZYXEL RT3070 0x341e NWD2105 product ZYXEL RTL8192CU 0x341f RTL8192CU product ZYXEL NWD2705 0x3421 NWD2705 Index: projects/building-blocks/sys/dev/usb/wlan/if_upgt.c =================================================================== --- projects/building-blocks/sys/dev/usb/wlan/if_upgt.c (revision 275663) +++ projects/building-blocks/sys/dev/usb/wlan/if_upgt.c (revision 275664) @@ -1,2452 +1,2452 @@ /* $OpenBSD: if_upgt.c,v 1.35 2008/04/16 18:32:15 damien Exp $ */ /* $FreeBSD$ */ /* * Copyright (c) 2007 Marcus Glocker * * Permission to use, copy, modify, and distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "usbdevs.h" #include /* * Driver for the USB PrismGT devices. * * For now just USB 2.0 devices with the GW3887 chipset are supported. * The driver has been written based on the firmware version 2.13.1.0_LM87. * * TODO's: * - MONITOR mode test. * - Add HOSTAP mode. * - Add IBSS mode. * - Support the USB 1.0 devices (NET2280, ISL3880, ISL3886 chipsets). * * Parts of this driver has been influenced by reading the p54u driver * written by Jean-Baptiste Note and * Sebastien Bourdeauducq . */ static SYSCTL_NODE(_hw, OID_AUTO, upgt, CTLFLAG_RD, 0, "USB PrismGT GW3887 driver parameters"); #ifdef UPGT_DEBUG int upgt_debug = 0; SYSCTL_INT(_hw_upgt, OID_AUTO, debug, CTLFLAG_RWTUN, &upgt_debug, 0, "control debugging printfs"); enum { UPGT_DEBUG_XMIT = 0x00000001, /* basic xmit operation */ UPGT_DEBUG_RECV = 0x00000002, /* basic recv operation */ UPGT_DEBUG_RESET = 0x00000004, /* reset processing */ UPGT_DEBUG_INTR = 0x00000008, /* INTR */ UPGT_DEBUG_TX_PROC = 0x00000010, /* tx ISR proc */ UPGT_DEBUG_RX_PROC = 0x00000020, /* rx ISR proc */ UPGT_DEBUG_STATE = 0x00000040, /* 802.11 state transitions */ UPGT_DEBUG_STAT = 0x00000080, /* statistic */ UPGT_DEBUG_FW = 0x00000100, /* firmware */ UPGT_DEBUG_ANY = 0xffffffff }; #define DPRINTF(sc, m, fmt, ...) do { \ if (sc->sc_debug & (m)) \ printf(fmt, __VA_ARGS__); \ } while (0) #else #define DPRINTF(sc, m, fmt, ...) do { \ (void) sc; \ } while (0) #endif /* * Prototypes. */ static device_probe_t upgt_match; static device_attach_t upgt_attach; static device_detach_t upgt_detach; static int upgt_alloc_tx(struct upgt_softc *); static int upgt_alloc_rx(struct upgt_softc *); static int upgt_device_reset(struct upgt_softc *); static void upgt_bulk_tx(struct upgt_softc *, struct upgt_data *); static int upgt_fw_verify(struct upgt_softc *); static int upgt_mem_init(struct upgt_softc *); static int upgt_fw_load(struct upgt_softc *); static int upgt_fw_copy(const uint8_t *, char *, int); static uint32_t upgt_crc32_le(const void *, size_t); static struct mbuf * upgt_rxeof(struct usb_xfer *, struct upgt_data *, int *); static struct mbuf * upgt_rx(struct upgt_softc *, uint8_t *, int, int *); static void upgt_txeof(struct usb_xfer *, struct upgt_data *); static int upgt_eeprom_read(struct upgt_softc *); static int upgt_eeprom_parse(struct upgt_softc *); static void upgt_eeprom_parse_hwrx(struct upgt_softc *, uint8_t *); static void upgt_eeprom_parse_freq3(struct upgt_softc *, uint8_t *, int); static void upgt_eeprom_parse_freq4(struct upgt_softc *, uint8_t *, int); static void upgt_eeprom_parse_freq6(struct upgt_softc *, uint8_t *, int); static uint32_t upgt_chksum_le(const uint32_t *, size_t); static void upgt_tx_done(struct upgt_softc *, uint8_t *); static void upgt_init(void *); static void upgt_init_locked(struct upgt_softc *); static int upgt_ioctl(struct ifnet *, u_long, caddr_t); static void upgt_start(struct ifnet *); static int upgt_raw_xmit(struct ieee80211_node *, struct mbuf *, const struct ieee80211_bpf_params *); static void upgt_scan_start(struct ieee80211com *); static void upgt_scan_end(struct ieee80211com *); static void upgt_set_channel(struct ieee80211com *); static struct ieee80211vap *upgt_vap_create(struct ieee80211com *, const char [IFNAMSIZ], int, enum ieee80211_opmode, int, const uint8_t [IEEE80211_ADDR_LEN], const uint8_t [IEEE80211_ADDR_LEN]); static void upgt_vap_delete(struct ieee80211vap *); static void upgt_update_mcast(struct ifnet *); static uint8_t upgt_rx_rate(struct upgt_softc *, const int); static void upgt_set_multi(void *); static void upgt_stop(struct upgt_softc *); static void upgt_setup_rates(struct ieee80211vap *, struct ieee80211com *); static int upgt_set_macfilter(struct upgt_softc *, uint8_t); static int upgt_newstate(struct ieee80211vap *, enum ieee80211_state, int); static void upgt_set_chan(struct upgt_softc *, struct ieee80211_channel *); static void upgt_set_led(struct upgt_softc *, int); static void upgt_set_led_blink(void *); static void upgt_get_stats(struct upgt_softc *); static void upgt_mem_free(struct upgt_softc *, uint32_t); static uint32_t upgt_mem_alloc(struct upgt_softc *); static void upgt_free_tx(struct upgt_softc *); static void upgt_free_rx(struct upgt_softc *); static void upgt_watchdog(void *); static void upgt_abort_xfers(struct upgt_softc *); static void upgt_abort_xfers_locked(struct upgt_softc *); static void upgt_sysctl_node(struct upgt_softc *); static struct upgt_data * upgt_getbuf(struct upgt_softc *); static struct upgt_data * upgt_gettxbuf(struct upgt_softc *); static int upgt_tx_start(struct upgt_softc *, struct mbuf *, struct ieee80211_node *, struct upgt_data *); static const char *upgt_fwname = "upgt-gw3887"; static const STRUCT_USB_HOST_ID upgt_devs[] = { #define UPGT_DEV(v,p) { USB_VP(USB_VENDOR_##v, USB_PRODUCT_##v##_##p) } /* version 2 devices */ UPGT_DEV(ACCTON, PRISM_GT), UPGT_DEV(BELKIN, F5D7050), UPGT_DEV(CISCOLINKSYS, WUSB54AG), UPGT_DEV(CONCEPTRONIC, PRISM_GT), UPGT_DEV(DELL, PRISM_GT_1), UPGT_DEV(DELL, PRISM_GT_2), UPGT_DEV(FSC, E5400), UPGT_DEV(GLOBESPAN, PRISM_GT_1), UPGT_DEV(GLOBESPAN, PRISM_GT_2), - UPGT_DEV(NETGEAR, WG111V2_2), + UPGT_DEV(NETGEAR, WG111V1_2), UPGT_DEV(INTERSIL, PRISM_GT), UPGT_DEV(SMC, 2862WG), UPGT_DEV(USR, USR5422), UPGT_DEV(WISTRONNEWEB, UR045G), UPGT_DEV(XYRATEX, PRISM_GT_1), UPGT_DEV(XYRATEX, PRISM_GT_2), UPGT_DEV(ZCOM, XG703A), UPGT_DEV(ZCOM, XM142) }; static usb_callback_t upgt_bulk_rx_callback; static usb_callback_t upgt_bulk_tx_callback; static const struct usb_config upgt_config[UPGT_N_XFERS] = { [UPGT_BULK_TX] = { .type = UE_BULK, .endpoint = UE_ADDR_ANY, .direction = UE_DIR_OUT, .bufsize = MCLBYTES * UPGT_TX_MAXCOUNT, .flags = { .force_short_xfer = 1, .pipe_bof = 1 }, .callback = upgt_bulk_tx_callback, .timeout = UPGT_USB_TIMEOUT, /* ms */ }, [UPGT_BULK_RX] = { .type = UE_BULK, .endpoint = UE_ADDR_ANY, .direction = UE_DIR_IN, .bufsize = MCLBYTES * UPGT_RX_MAXCOUNT, .flags = { .pipe_bof = 1, .short_xfer_ok = 1 }, .callback = upgt_bulk_rx_callback, }, }; static int upgt_match(device_t dev) { struct usb_attach_arg *uaa = device_get_ivars(dev); if (uaa->usb_mode != USB_MODE_HOST) return (ENXIO); if (uaa->info.bConfigIndex != UPGT_CONFIG_INDEX) return (ENXIO); if (uaa->info.bIfaceIndex != UPGT_IFACE_INDEX) return (ENXIO); return (usbd_lookup_id_by_uaa(upgt_devs, sizeof(upgt_devs), uaa)); } static int upgt_attach(device_t dev) { int error; struct ieee80211com *ic; struct ifnet *ifp; struct upgt_softc *sc = device_get_softc(dev); struct usb_attach_arg *uaa = device_get_ivars(dev); uint8_t bands, iface_index = UPGT_IFACE_INDEX; sc->sc_dev = dev; sc->sc_udev = uaa->device; #ifdef UPGT_DEBUG sc->sc_debug = upgt_debug; #endif device_set_usb_desc(dev); mtx_init(&sc->sc_mtx, device_get_nameunit(sc->sc_dev), MTX_NETWORK_LOCK, MTX_DEF); callout_init(&sc->sc_led_ch, 0); callout_init(&sc->sc_watchdog_ch, 0); error = usbd_transfer_setup(uaa->device, &iface_index, sc->sc_xfer, upgt_config, UPGT_N_XFERS, sc, &sc->sc_mtx); if (error) { device_printf(dev, "could not allocate USB transfers, " "err=%s\n", usbd_errstr(error)); goto fail1; } sc->sc_rx_dma_buf = usbd_xfer_get_frame_buffer( sc->sc_xfer[UPGT_BULK_RX], 0); sc->sc_tx_dma_buf = usbd_xfer_get_frame_buffer( sc->sc_xfer[UPGT_BULK_TX], 0); /* Setup TX and RX buffers */ error = upgt_alloc_tx(sc); if (error) goto fail2; error = upgt_alloc_rx(sc); if (error) goto fail3; ifp = sc->sc_ifp = if_alloc(IFT_IEEE80211); if (ifp == NULL) { device_printf(dev, "can not if_alloc()\n"); goto fail4; } /* Initialize the device. */ error = upgt_device_reset(sc); if (error) goto fail5; /* Verify the firmware. */ error = upgt_fw_verify(sc); if (error) goto fail5; /* Calculate device memory space. */ if (sc->sc_memaddr_frame_start == 0 || sc->sc_memaddr_frame_end == 0) { device_printf(dev, "could not find memory space addresses on FW\n"); error = EIO; goto fail5; } sc->sc_memaddr_frame_end -= UPGT_MEMSIZE_RX + 1; sc->sc_memaddr_rx_start = sc->sc_memaddr_frame_end + 1; DPRINTF(sc, UPGT_DEBUG_FW, "memory address frame start=0x%08x\n", sc->sc_memaddr_frame_start); DPRINTF(sc, UPGT_DEBUG_FW, "memory address frame end=0x%08x\n", sc->sc_memaddr_frame_end); DPRINTF(sc, UPGT_DEBUG_FW, "memory address rx start=0x%08x\n", sc->sc_memaddr_rx_start); upgt_mem_init(sc); /* Load the firmware. */ error = upgt_fw_load(sc); if (error) goto fail5; /* Read the whole EEPROM content and parse it. */ error = upgt_eeprom_read(sc); if (error) goto fail5; error = upgt_eeprom_parse(sc); if (error) goto fail5; /* all works related with the device have done here. */ upgt_abort_xfers(sc); /* Setup the 802.11 device. */ ifp->if_softc = sc; if_initname(ifp, "upgt", device_get_unit(sc->sc_dev)); ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST; ifp->if_init = upgt_init; ifp->if_ioctl = upgt_ioctl; ifp->if_start = upgt_start; IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen); IFQ_SET_READY(&ifp->if_snd); ic = ifp->if_l2com; ic->ic_ifp = ifp; ic->ic_phytype = IEEE80211_T_OFDM; /* not only, but not used */ ic->ic_opmode = IEEE80211_M_STA; /* set device capabilities */ ic->ic_caps = IEEE80211_C_STA /* station mode */ | IEEE80211_C_MONITOR /* monitor mode */ | IEEE80211_C_SHPREAMBLE /* short preamble supported */ | IEEE80211_C_SHSLOT /* short slot time supported */ | IEEE80211_C_BGSCAN /* capable of bg scanning */ | IEEE80211_C_WPA /* 802.11i */ ; bands = 0; setbit(&bands, IEEE80211_MODE_11B); setbit(&bands, IEEE80211_MODE_11G); ieee80211_init_channels(ic, NULL, &bands); ieee80211_ifattach(ic, sc->sc_myaddr); ic->ic_raw_xmit = upgt_raw_xmit; ic->ic_scan_start = upgt_scan_start; ic->ic_scan_end = upgt_scan_end; ic->ic_set_channel = upgt_set_channel; ic->ic_vap_create = upgt_vap_create; ic->ic_vap_delete = upgt_vap_delete; ic->ic_update_mcast = upgt_update_mcast; ieee80211_radiotap_attach(ic, &sc->sc_txtap.wt_ihdr, sizeof(sc->sc_txtap), UPGT_TX_RADIOTAP_PRESENT, &sc->sc_rxtap.wr_ihdr, sizeof(sc->sc_rxtap), UPGT_RX_RADIOTAP_PRESENT); upgt_sysctl_node(sc); if (bootverbose) ieee80211_announce(ic); return (0); fail5: if_free(ifp); fail4: upgt_free_rx(sc); fail3: upgt_free_tx(sc); fail2: usbd_transfer_unsetup(sc->sc_xfer, UPGT_N_XFERS); fail1: mtx_destroy(&sc->sc_mtx); return (error); } static void upgt_txeof(struct usb_xfer *xfer, struct upgt_data *data) { struct upgt_softc *sc = usbd_xfer_softc(xfer); struct ifnet *ifp = sc->sc_ifp; struct mbuf *m; UPGT_ASSERT_LOCKED(sc); /* * Do any tx complete callback. Note this must be done before releasing * the node reference. */ if (data->m) { m = data->m; if (m->m_flags & M_TXCB) { /* XXX status? */ ieee80211_process_callback(data->ni, m, 0); } m_freem(m); data->m = NULL; } if (data->ni) { ieee80211_free_node(data->ni); data->ni = NULL; } if_inc_counter(ifp, IFCOUNTER_OPACKETS, 1); } static void upgt_get_stats(struct upgt_softc *sc) { struct upgt_data *data_cmd; struct upgt_lmac_mem *mem; struct upgt_lmac_stats *stats; data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { device_printf(sc->sc_dev, "%s: out of buffers.\n", __func__); return; } /* * Transmit the URB containing the CMD data. */ memset(data_cmd->buf, 0, MCLBYTES); mem = (struct upgt_lmac_mem *)data_cmd->buf; mem->addr = htole32(sc->sc_memaddr_frame_start + UPGT_MEMSIZE_FRAME_HEAD); stats = (struct upgt_lmac_stats *)(mem + 1); stats->header1.flags = 0; stats->header1.type = UPGT_H1_TYPE_CTRL; stats->header1.len = htole16( sizeof(struct upgt_lmac_stats) - sizeof(struct upgt_lmac_header)); stats->header2.reqid = htole32(sc->sc_memaddr_frame_start); stats->header2.type = htole16(UPGT_H2_TYPE_STATS); stats->header2.flags = 0; data_cmd->buflen = sizeof(*mem) + sizeof(*stats); mem->chksum = upgt_chksum_le((uint32_t *)stats, data_cmd->buflen - sizeof(*mem)); upgt_bulk_tx(sc, data_cmd); } static int upgt_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data) { struct upgt_softc *sc = ifp->if_softc; struct ieee80211com *ic = ifp->if_l2com; struct ifreq *ifr = (struct ifreq *) data; int error; int startall = 0; UPGT_LOCK(sc); error = (sc->sc_flags & UPGT_FLAG_DETACHED) ? ENXIO : 0; UPGT_UNLOCK(sc); if (error) return (error); switch (cmd) { case SIOCSIFFLAGS: if (ifp->if_flags & IFF_UP) { if (ifp->if_drv_flags & IFF_DRV_RUNNING) { if ((ifp->if_flags ^ sc->sc_if_flags) & (IFF_ALLMULTI | IFF_PROMISC)) upgt_set_multi(sc); } else { upgt_init(sc); startall = 1; } } else { if (ifp->if_drv_flags & IFF_DRV_RUNNING) upgt_stop(sc); } sc->sc_if_flags = ifp->if_flags; if (startall) ieee80211_start_all(ic); break; case SIOCGIFMEDIA: error = ifmedia_ioctl(ifp, ifr, &ic->ic_media, cmd); break; case SIOCGIFADDR: error = ether_ioctl(ifp, cmd, data); break; default: error = EINVAL; break; } return error; } static void upgt_stop_locked(struct upgt_softc *sc) { struct ifnet *ifp = sc->sc_ifp; UPGT_ASSERT_LOCKED(sc); if (ifp->if_drv_flags & IFF_DRV_RUNNING) upgt_set_macfilter(sc, IEEE80211_S_INIT); upgt_abort_xfers_locked(sc); } static void upgt_stop(struct upgt_softc *sc) { struct ifnet *ifp = sc->sc_ifp; UPGT_LOCK(sc); upgt_stop_locked(sc); UPGT_UNLOCK(sc); /* device down */ sc->sc_tx_timer = 0; ifp->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE); sc->sc_flags &= ~UPGT_FLAG_INITDONE; } static void upgt_set_led(struct upgt_softc *sc, int action) { struct upgt_data *data_cmd; struct upgt_lmac_mem *mem; struct upgt_lmac_led *led; data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { device_printf(sc->sc_dev, "%s: out of buffers.\n", __func__); return; } /* * Transmit the URB containing the CMD data. */ memset(data_cmd->buf, 0, MCLBYTES); mem = (struct upgt_lmac_mem *)data_cmd->buf; mem->addr = htole32(sc->sc_memaddr_frame_start + UPGT_MEMSIZE_FRAME_HEAD); led = (struct upgt_lmac_led *)(mem + 1); led->header1.flags = UPGT_H1_FLAGS_TX_NO_CALLBACK; led->header1.type = UPGT_H1_TYPE_CTRL; led->header1.len = htole16( sizeof(struct upgt_lmac_led) - sizeof(struct upgt_lmac_header)); led->header2.reqid = htole32(sc->sc_memaddr_frame_start); led->header2.type = htole16(UPGT_H2_TYPE_LED); led->header2.flags = 0; switch (action) { case UPGT_LED_OFF: led->mode = htole16(UPGT_LED_MODE_SET); led->action_fix = 0; led->action_tmp = htole16(UPGT_LED_ACTION_OFF); led->action_tmp_dur = 0; break; case UPGT_LED_ON: led->mode = htole16(UPGT_LED_MODE_SET); led->action_fix = 0; led->action_tmp = htole16(UPGT_LED_ACTION_ON); led->action_tmp_dur = 0; break; case UPGT_LED_BLINK: if (sc->sc_state != IEEE80211_S_RUN) { STAILQ_INSERT_TAIL(&sc->sc_tx_inactive, data_cmd, next); return; } if (sc->sc_led_blink) { /* previous blink was not finished */ STAILQ_INSERT_TAIL(&sc->sc_tx_inactive, data_cmd, next); return; } led->mode = htole16(UPGT_LED_MODE_SET); led->action_fix = htole16(UPGT_LED_ACTION_OFF); led->action_tmp = htole16(UPGT_LED_ACTION_ON); led->action_tmp_dur = htole16(UPGT_LED_ACTION_TMP_DUR); /* lock blink */ sc->sc_led_blink = 1; callout_reset(&sc->sc_led_ch, hz, upgt_set_led_blink, sc); break; default: STAILQ_INSERT_TAIL(&sc->sc_tx_inactive, data_cmd, next); return; } data_cmd->buflen = sizeof(*mem) + sizeof(*led); mem->chksum = upgt_chksum_le((uint32_t *)led, data_cmd->buflen - sizeof(*mem)); upgt_bulk_tx(sc, data_cmd); } static void upgt_set_led_blink(void *arg) { struct upgt_softc *sc = arg; /* blink finished, we are ready for a next one */ sc->sc_led_blink = 0; } static void upgt_init(void *priv) { struct upgt_softc *sc = priv; struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; UPGT_LOCK(sc); upgt_init_locked(sc); UPGT_UNLOCK(sc); if (ifp->if_drv_flags & IFF_DRV_RUNNING) ieee80211_start_all(ic); /* start all vap's */ } static void upgt_init_locked(struct upgt_softc *sc) { struct ifnet *ifp = sc->sc_ifp; UPGT_ASSERT_LOCKED(sc); if (ifp->if_drv_flags & IFF_DRV_RUNNING) upgt_stop_locked(sc); usbd_transfer_start(sc->sc_xfer[UPGT_BULK_RX]); (void)upgt_set_macfilter(sc, IEEE80211_S_SCAN); ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; ifp->if_drv_flags |= IFF_DRV_RUNNING; sc->sc_flags |= UPGT_FLAG_INITDONE; callout_reset(&sc->sc_watchdog_ch, hz, upgt_watchdog, sc); } static int upgt_set_macfilter(struct upgt_softc *sc, uint8_t state) { struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps); struct ieee80211_node *ni; struct upgt_data *data_cmd; struct upgt_lmac_mem *mem; struct upgt_lmac_filter *filter; uint8_t broadcast[] = { 0xff, 0xff, 0xff, 0xff, 0xff, 0xff }; UPGT_ASSERT_LOCKED(sc); data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { device_printf(sc->sc_dev, "out of TX buffers.\n"); return (ENOBUFS); } /* * Transmit the URB containing the CMD data. */ memset(data_cmd->buf, 0, MCLBYTES); mem = (struct upgt_lmac_mem *)data_cmd->buf; mem->addr = htole32(sc->sc_memaddr_frame_start + UPGT_MEMSIZE_FRAME_HEAD); filter = (struct upgt_lmac_filter *)(mem + 1); filter->header1.flags = UPGT_H1_FLAGS_TX_NO_CALLBACK; filter->header1.type = UPGT_H1_TYPE_CTRL; filter->header1.len = htole16( sizeof(struct upgt_lmac_filter) - sizeof(struct upgt_lmac_header)); filter->header2.reqid = htole32(sc->sc_memaddr_frame_start); filter->header2.type = htole16(UPGT_H2_TYPE_MACFILTER); filter->header2.flags = 0; switch (state) { case IEEE80211_S_INIT: DPRINTF(sc, UPGT_DEBUG_STATE, "%s: set MAC filter to INIT\n", __func__); filter->type = htole16(UPGT_FILTER_TYPE_RESET); break; case IEEE80211_S_SCAN: DPRINTF(sc, UPGT_DEBUG_STATE, "set MAC filter to SCAN (bssid %s)\n", ether_sprintf(broadcast)); filter->type = htole16(UPGT_FILTER_TYPE_NONE); IEEE80211_ADDR_COPY(filter->dst, sc->sc_myaddr); IEEE80211_ADDR_COPY(filter->src, broadcast); filter->unknown1 = htole16(UPGT_FILTER_UNKNOWN1); filter->rxaddr = htole32(sc->sc_memaddr_rx_start); filter->unknown2 = htole16(UPGT_FILTER_UNKNOWN2); filter->rxhw = htole32(sc->sc_eeprom_hwrx); filter->unknown3 = htole16(UPGT_FILTER_UNKNOWN3); break; case IEEE80211_S_RUN: ni = ieee80211_ref_node(vap->iv_bss); /* XXX monitor mode isn't tested yet. */ if (vap->iv_opmode == IEEE80211_M_MONITOR) { filter->type = htole16(UPGT_FILTER_TYPE_MONITOR); IEEE80211_ADDR_COPY(filter->dst, sc->sc_myaddr); IEEE80211_ADDR_COPY(filter->src, ni->ni_bssid); filter->unknown1 = htole16(UPGT_FILTER_MONITOR_UNKNOWN1); filter->rxaddr = htole32(sc->sc_memaddr_rx_start); filter->unknown2 = htole16(UPGT_FILTER_MONITOR_UNKNOWN2); filter->rxhw = htole32(sc->sc_eeprom_hwrx); filter->unknown3 = htole16(UPGT_FILTER_MONITOR_UNKNOWN3); } else { DPRINTF(sc, UPGT_DEBUG_STATE, "set MAC filter to RUN (bssid %s)\n", ether_sprintf(ni->ni_bssid)); filter->type = htole16(UPGT_FILTER_TYPE_STA); IEEE80211_ADDR_COPY(filter->dst, sc->sc_myaddr); IEEE80211_ADDR_COPY(filter->src, ni->ni_bssid); filter->unknown1 = htole16(UPGT_FILTER_UNKNOWN1); filter->rxaddr = htole32(sc->sc_memaddr_rx_start); filter->unknown2 = htole16(UPGT_FILTER_UNKNOWN2); filter->rxhw = htole32(sc->sc_eeprom_hwrx); filter->unknown3 = htole16(UPGT_FILTER_UNKNOWN3); } ieee80211_free_node(ni); break; default: device_printf(sc->sc_dev, "MAC filter does not know that state\n"); break; } data_cmd->buflen = sizeof(*mem) + sizeof(*filter); mem->chksum = upgt_chksum_le((uint32_t *)filter, data_cmd->buflen - sizeof(*mem)); upgt_bulk_tx(sc, data_cmd); return (0); } static void upgt_setup_rates(struct ieee80211vap *vap, struct ieee80211com *ic) { struct ifnet *ifp = ic->ic_ifp; struct upgt_softc *sc = ifp->if_softc; const struct ieee80211_txparam *tp; /* * 0x01 = OFMD6 0x10 = DS1 * 0x04 = OFDM9 0x11 = DS2 * 0x06 = OFDM12 0x12 = DS5 * 0x07 = OFDM18 0x13 = DS11 * 0x08 = OFDM24 * 0x09 = OFDM36 * 0x0a = OFDM48 * 0x0b = OFDM54 */ const uint8_t rateset_auto_11b[] = { 0x13, 0x13, 0x12, 0x11, 0x11, 0x10, 0x10, 0x10 }; const uint8_t rateset_auto_11g[] = { 0x0b, 0x0a, 0x09, 0x08, 0x07, 0x06, 0x04, 0x01 }; const uint8_t rateset_fix_11bg[] = { 0x10, 0x11, 0x12, 0x13, 0x01, 0x04, 0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b }; tp = &vap->iv_txparms[ieee80211_chan2mode(ic->ic_curchan)]; /* XXX */ if (tp->ucastrate == IEEE80211_FIXED_RATE_NONE) { /* * Automatic rate control is done by the device. * We just pass the rateset from which the device * will pickup a rate. */ if (ic->ic_curmode == IEEE80211_MODE_11B) memcpy(sc->sc_cur_rateset, rateset_auto_11b, sizeof(sc->sc_cur_rateset)); if (ic->ic_curmode == IEEE80211_MODE_11G || ic->ic_curmode == IEEE80211_MODE_AUTO) memcpy(sc->sc_cur_rateset, rateset_auto_11g, sizeof(sc->sc_cur_rateset)); } else { /* set a fixed rate */ memset(sc->sc_cur_rateset, rateset_fix_11bg[tp->ucastrate], sizeof(sc->sc_cur_rateset)); } } static void upgt_set_multi(void *arg) { struct upgt_softc *sc = arg; struct ifnet *ifp = sc->sc_ifp; if (!(ifp->if_flags & IFF_UP)) return; /* * XXX don't know how to set a device. Lack of docs. Just try to set * IFF_ALLMULTI flag here. */ ifp->if_flags |= IFF_ALLMULTI; } static void upgt_start(struct ifnet *ifp) { struct upgt_softc *sc = ifp->if_softc; struct upgt_data *data_tx; struct ieee80211_node *ni; struct mbuf *m; if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) return; UPGT_LOCK(sc); for (;;) { IFQ_DRV_DEQUEUE(&ifp->if_snd, m); if (m == NULL) break; data_tx = upgt_gettxbuf(sc); if (data_tx == NULL) { IFQ_DRV_PREPEND(&ifp->if_snd, m); break; } ni = (struct ieee80211_node *)m->m_pkthdr.rcvif; m->m_pkthdr.rcvif = NULL; if (upgt_tx_start(sc, m, ni, data_tx) != 0) { STAILQ_INSERT_HEAD(&sc->sc_tx_inactive, data_tx, next); UPGT_STAT_INC(sc, st_tx_inactive); ieee80211_free_node(ni); if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); continue; } sc->sc_tx_timer = 5; } UPGT_UNLOCK(sc); } static int upgt_raw_xmit(struct ieee80211_node *ni, struct mbuf *m, const struct ieee80211_bpf_params *params) { struct ieee80211com *ic = ni->ni_ic; struct ifnet *ifp = ic->ic_ifp; struct upgt_softc *sc = ifp->if_softc; struct upgt_data *data_tx = NULL; /* prevent management frames from being sent if we're not ready */ if (!(ifp->if_drv_flags & IFF_DRV_RUNNING)) { m_freem(m); ieee80211_free_node(ni); return ENETDOWN; } UPGT_LOCK(sc); data_tx = upgt_gettxbuf(sc); if (data_tx == NULL) { ieee80211_free_node(ni); m_freem(m); UPGT_UNLOCK(sc); return (ENOBUFS); } if (upgt_tx_start(sc, m, ni, data_tx) != 0) { STAILQ_INSERT_HEAD(&sc->sc_tx_inactive, data_tx, next); UPGT_STAT_INC(sc, st_tx_inactive); ieee80211_free_node(ni); if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); UPGT_UNLOCK(sc); return (EIO); } UPGT_UNLOCK(sc); sc->sc_tx_timer = 5; return (0); } static void upgt_watchdog(void *arg) { struct upgt_softc *sc = arg; struct ifnet *ifp = sc->sc_ifp; if (sc->sc_tx_timer > 0) { if (--sc->sc_tx_timer == 0) { device_printf(sc->sc_dev, "watchdog timeout\n"); /* upgt_init(ifp); XXX needs a process context ? */ if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); return; } callout_reset(&sc->sc_watchdog_ch, hz, upgt_watchdog, sc); } } static uint32_t upgt_mem_alloc(struct upgt_softc *sc) { int i; for (i = 0; i < sc->sc_memory.pages; i++) { if (sc->sc_memory.page[i].used == 0) { sc->sc_memory.page[i].used = 1; return (sc->sc_memory.page[i].addr); } } return (0); } static void upgt_scan_start(struct ieee80211com *ic) { /* do nothing. */ } static void upgt_scan_end(struct ieee80211com *ic) { /* do nothing. */ } static void upgt_set_channel(struct ieee80211com *ic) { struct upgt_softc *sc = ic->ic_ifp->if_softc; UPGT_LOCK(sc); upgt_set_chan(sc, ic->ic_curchan); UPGT_UNLOCK(sc); } static void upgt_set_chan(struct upgt_softc *sc, struct ieee80211_channel *c) { struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; struct upgt_data *data_cmd; struct upgt_lmac_mem *mem; struct upgt_lmac_channel *chan; int channel; UPGT_ASSERT_LOCKED(sc); channel = ieee80211_chan2ieee(ic, c); if (channel == 0 || channel == IEEE80211_CHAN_ANY) { /* XXX should NEVER happen */ device_printf(sc->sc_dev, "%s: invalid channel %x\n", __func__, channel); return; } DPRINTF(sc, UPGT_DEBUG_STATE, "%s: channel %d\n", __func__, channel); data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { device_printf(sc->sc_dev, "%s: out of buffers.\n", __func__); return; } /* * Transmit the URB containing the CMD data. */ memset(data_cmd->buf, 0, MCLBYTES); mem = (struct upgt_lmac_mem *)data_cmd->buf; mem->addr = htole32(sc->sc_memaddr_frame_start + UPGT_MEMSIZE_FRAME_HEAD); chan = (struct upgt_lmac_channel *)(mem + 1); chan->header1.flags = UPGT_H1_FLAGS_TX_NO_CALLBACK; chan->header1.type = UPGT_H1_TYPE_CTRL; chan->header1.len = htole16( sizeof(struct upgt_lmac_channel) - sizeof(struct upgt_lmac_header)); chan->header2.reqid = htole32(sc->sc_memaddr_frame_start); chan->header2.type = htole16(UPGT_H2_TYPE_CHANNEL); chan->header2.flags = 0; chan->unknown1 = htole16(UPGT_CHANNEL_UNKNOWN1); chan->unknown2 = htole16(UPGT_CHANNEL_UNKNOWN2); chan->freq6 = sc->sc_eeprom_freq6[channel]; chan->settings = sc->sc_eeprom_freq6_settings; chan->unknown3 = UPGT_CHANNEL_UNKNOWN3; memcpy(chan->freq3_1, &sc->sc_eeprom_freq3[channel].data, sizeof(chan->freq3_1)); memcpy(chan->freq4, &sc->sc_eeprom_freq4[channel], sizeof(sc->sc_eeprom_freq4[channel])); memcpy(chan->freq3_2, &sc->sc_eeprom_freq3[channel].data, sizeof(chan->freq3_2)); data_cmd->buflen = sizeof(*mem) + sizeof(*chan); mem->chksum = upgt_chksum_le((uint32_t *)chan, data_cmd->buflen - sizeof(*mem)); upgt_bulk_tx(sc, data_cmd); } static struct ieee80211vap * upgt_vap_create(struct ieee80211com *ic, const char name[IFNAMSIZ], int unit, enum ieee80211_opmode opmode, int flags, const uint8_t bssid[IEEE80211_ADDR_LEN], const uint8_t mac[IEEE80211_ADDR_LEN]) { struct upgt_vap *uvp; struct ieee80211vap *vap; if (!TAILQ_EMPTY(&ic->ic_vaps)) /* only one at a time */ return NULL; uvp = (struct upgt_vap *) malloc(sizeof(struct upgt_vap), M_80211_VAP, M_NOWAIT | M_ZERO); if (uvp == NULL) return NULL; vap = &uvp->vap; /* enable s/w bmiss handling for sta mode */ if (ieee80211_vap_setup(ic, vap, name, unit, opmode, flags | IEEE80211_CLONE_NOBEACONS, bssid, mac) != 0) { /* out of memory */ free(uvp, M_80211_VAP); return (NULL); } /* override state transition machine */ uvp->newstate = vap->iv_newstate; vap->iv_newstate = upgt_newstate; /* setup device rates */ upgt_setup_rates(vap, ic); /* complete setup */ ieee80211_vap_attach(vap, ieee80211_media_change, ieee80211_media_status); ic->ic_opmode = opmode; return vap; } static int upgt_newstate(struct ieee80211vap *vap, enum ieee80211_state nstate, int arg) { struct upgt_vap *uvp = UPGT_VAP(vap); struct ieee80211com *ic = vap->iv_ic; struct upgt_softc *sc = ic->ic_ifp->if_softc; /* do it in a process context */ sc->sc_state = nstate; IEEE80211_UNLOCK(ic); UPGT_LOCK(sc); callout_stop(&sc->sc_led_ch); callout_stop(&sc->sc_watchdog_ch); switch (nstate) { case IEEE80211_S_INIT: /* do not accept any frames if the device is down */ (void)upgt_set_macfilter(sc, sc->sc_state); upgt_set_led(sc, UPGT_LED_OFF); break; case IEEE80211_S_SCAN: upgt_set_chan(sc, ic->ic_curchan); break; case IEEE80211_S_AUTH: upgt_set_chan(sc, ic->ic_curchan); break; case IEEE80211_S_ASSOC: break; case IEEE80211_S_RUN: upgt_set_macfilter(sc, sc->sc_state); upgt_set_led(sc, UPGT_LED_ON); break; default: break; } UPGT_UNLOCK(sc); IEEE80211_LOCK(ic); return (uvp->newstate(vap, nstate, arg)); } static void upgt_vap_delete(struct ieee80211vap *vap) { struct upgt_vap *uvp = UPGT_VAP(vap); ieee80211_vap_detach(vap); free(uvp, M_80211_VAP); } static void upgt_update_mcast(struct ifnet *ifp) { struct upgt_softc *sc = ifp->if_softc; upgt_set_multi(sc); } static int upgt_eeprom_parse(struct upgt_softc *sc) { struct upgt_eeprom_header *eeprom_header; struct upgt_eeprom_option *eeprom_option; uint16_t option_len; uint16_t option_type; uint16_t preamble_len; int option_end = 0; /* calculate eeprom options start offset */ eeprom_header = (struct upgt_eeprom_header *)sc->sc_eeprom; preamble_len = le16toh(eeprom_header->preamble_len); eeprom_option = (struct upgt_eeprom_option *)(sc->sc_eeprom + (sizeof(struct upgt_eeprom_header) + preamble_len)); while (!option_end) { /* sanity check */ if (eeprom_option >= (struct upgt_eeprom_option *) (sc->sc_eeprom + UPGT_EEPROM_SIZE)) { return (EINVAL); } /* the eeprom option length is stored in words */ option_len = (le16toh(eeprom_option->len) - 1) * sizeof(uint16_t); option_type = le16toh(eeprom_option->type); /* sanity check */ if (option_len == 0 || option_len >= UPGT_EEPROM_SIZE) return (EINVAL); switch (option_type) { case UPGT_EEPROM_TYPE_NAME: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM name len=%d\n", option_len); break; case UPGT_EEPROM_TYPE_SERIAL: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM serial len=%d\n", option_len); break; case UPGT_EEPROM_TYPE_MAC: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM mac len=%d\n", option_len); IEEE80211_ADDR_COPY(sc->sc_myaddr, eeprom_option->data); break; case UPGT_EEPROM_TYPE_HWRX: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM hwrx len=%d\n", option_len); upgt_eeprom_parse_hwrx(sc, eeprom_option->data); break; case UPGT_EEPROM_TYPE_CHIP: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM chip len=%d\n", option_len); break; case UPGT_EEPROM_TYPE_FREQ3: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM freq3 len=%d\n", option_len); upgt_eeprom_parse_freq3(sc, eeprom_option->data, option_len); break; case UPGT_EEPROM_TYPE_FREQ4: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM freq4 len=%d\n", option_len); upgt_eeprom_parse_freq4(sc, eeprom_option->data, option_len); break; case UPGT_EEPROM_TYPE_FREQ5: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM freq5 len=%d\n", option_len); break; case UPGT_EEPROM_TYPE_FREQ6: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM freq6 len=%d\n", option_len); upgt_eeprom_parse_freq6(sc, eeprom_option->data, option_len); break; case UPGT_EEPROM_TYPE_END: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM end len=%d\n", option_len); option_end = 1; break; case UPGT_EEPROM_TYPE_OFF: DPRINTF(sc, UPGT_DEBUG_FW, "%s: EEPROM off without end option\n", __func__); return (EIO); default: DPRINTF(sc, UPGT_DEBUG_FW, "EEPROM unknown type 0x%04x len=%d\n", option_type, option_len); break; } /* jump to next EEPROM option */ eeprom_option = (struct upgt_eeprom_option *) (eeprom_option->data + option_len); } return (0); } static void upgt_eeprom_parse_freq3(struct upgt_softc *sc, uint8_t *data, int len) { struct upgt_eeprom_freq3_header *freq3_header; struct upgt_lmac_freq3 *freq3; int i; int elements; int flags; unsigned channel; freq3_header = (struct upgt_eeprom_freq3_header *)data; freq3 = (struct upgt_lmac_freq3 *)(freq3_header + 1); flags = freq3_header->flags; elements = freq3_header->elements; DPRINTF(sc, UPGT_DEBUG_FW, "flags=0x%02x elements=%d\n", flags, elements); if (elements >= (int)(UPGT_EEPROM_SIZE / sizeof(freq3[0]))) return; for (i = 0; i < elements; i++) { channel = ieee80211_mhz2ieee(le16toh(freq3[i].freq), 0); if (channel >= IEEE80211_CHAN_MAX) continue; sc->sc_eeprom_freq3[channel] = freq3[i]; DPRINTF(sc, UPGT_DEBUG_FW, "frequence=%d, channel=%d\n", le16toh(sc->sc_eeprom_freq3[channel].freq), channel); } } void upgt_eeprom_parse_freq4(struct upgt_softc *sc, uint8_t *data, int len) { struct upgt_eeprom_freq4_header *freq4_header; struct upgt_eeprom_freq4_1 *freq4_1; struct upgt_eeprom_freq4_2 *freq4_2; int i; int j; int elements; int settings; int flags; unsigned channel; freq4_header = (struct upgt_eeprom_freq4_header *)data; freq4_1 = (struct upgt_eeprom_freq4_1 *)(freq4_header + 1); flags = freq4_header->flags; elements = freq4_header->elements; settings = freq4_header->settings; /* we need this value later */ sc->sc_eeprom_freq6_settings = freq4_header->settings; DPRINTF(sc, UPGT_DEBUG_FW, "flags=0x%02x elements=%d settings=%d\n", flags, elements, settings); if (elements >= (int)(UPGT_EEPROM_SIZE / sizeof(freq4_1[0]))) return; for (i = 0; i < elements; i++) { channel = ieee80211_mhz2ieee(le16toh(freq4_1[i].freq), 0); if (channel >= IEEE80211_CHAN_MAX) continue; freq4_2 = (struct upgt_eeprom_freq4_2 *)freq4_1[i].data; for (j = 0; j < settings; j++) { sc->sc_eeprom_freq4[channel][j].cmd = freq4_2[j]; sc->sc_eeprom_freq4[channel][j].pad = 0; } DPRINTF(sc, UPGT_DEBUG_FW, "frequence=%d, channel=%d\n", le16toh(freq4_1[i].freq), channel); } } void upgt_eeprom_parse_freq6(struct upgt_softc *sc, uint8_t *data, int len) { struct upgt_lmac_freq6 *freq6; int i; int elements; unsigned channel; freq6 = (struct upgt_lmac_freq6 *)data; elements = len / sizeof(struct upgt_lmac_freq6); DPRINTF(sc, UPGT_DEBUG_FW, "elements=%d\n", elements); if (elements >= (int)(UPGT_EEPROM_SIZE / sizeof(freq6[0]))) return; for (i = 0; i < elements; i++) { channel = ieee80211_mhz2ieee(le16toh(freq6[i].freq), 0); if (channel >= IEEE80211_CHAN_MAX) continue; sc->sc_eeprom_freq6[channel] = freq6[i]; DPRINTF(sc, UPGT_DEBUG_FW, "frequence=%d, channel=%d\n", le16toh(sc->sc_eeprom_freq6[channel].freq), channel); } } static void upgt_eeprom_parse_hwrx(struct upgt_softc *sc, uint8_t *data) { struct upgt_eeprom_option_hwrx *option_hwrx; option_hwrx = (struct upgt_eeprom_option_hwrx *)data; sc->sc_eeprom_hwrx = option_hwrx->rxfilter - UPGT_EEPROM_RX_CONST; DPRINTF(sc, UPGT_DEBUG_FW, "hwrx option value=0x%04x\n", sc->sc_eeprom_hwrx); } static int upgt_eeprom_read(struct upgt_softc *sc) { struct upgt_data *data_cmd; struct upgt_lmac_mem *mem; struct upgt_lmac_eeprom *eeprom; int block, error, offset; UPGT_LOCK(sc); usb_pause_mtx(&sc->sc_mtx, 100); offset = 0; block = UPGT_EEPROM_BLOCK_SIZE; while (offset < UPGT_EEPROM_SIZE) { DPRINTF(sc, UPGT_DEBUG_FW, "request EEPROM block (offset=%d, len=%d)\n", offset, block); data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { UPGT_UNLOCK(sc); return (ENOBUFS); } /* * Transmit the URB containing the CMD data. */ memset(data_cmd->buf, 0, MCLBYTES); mem = (struct upgt_lmac_mem *)data_cmd->buf; mem->addr = htole32(sc->sc_memaddr_frame_start + UPGT_MEMSIZE_FRAME_HEAD); eeprom = (struct upgt_lmac_eeprom *)(mem + 1); eeprom->header1.flags = 0; eeprom->header1.type = UPGT_H1_TYPE_CTRL; eeprom->header1.len = htole16(( sizeof(struct upgt_lmac_eeprom) - sizeof(struct upgt_lmac_header)) + block); eeprom->header2.reqid = htole32(sc->sc_memaddr_frame_start); eeprom->header2.type = htole16(UPGT_H2_TYPE_EEPROM); eeprom->header2.flags = 0; eeprom->offset = htole16(offset); eeprom->len = htole16(block); data_cmd->buflen = sizeof(*mem) + sizeof(*eeprom) + block; mem->chksum = upgt_chksum_le((uint32_t *)eeprom, data_cmd->buflen - sizeof(*mem)); upgt_bulk_tx(sc, data_cmd); error = mtx_sleep(sc, &sc->sc_mtx, 0, "eeprom_request", hz); if (error != 0) { device_printf(sc->sc_dev, "timeout while waiting for EEPROM data\n"); UPGT_UNLOCK(sc); return (EIO); } offset += block; if (UPGT_EEPROM_SIZE - offset < block) block = UPGT_EEPROM_SIZE - offset; } UPGT_UNLOCK(sc); return (0); } /* * When a rx data came in the function returns a mbuf and a rssi values. */ static struct mbuf * upgt_rxeof(struct usb_xfer *xfer, struct upgt_data *data, int *rssi) { struct mbuf *m = NULL; struct upgt_softc *sc = usbd_xfer_softc(xfer); struct upgt_lmac_header *header; struct upgt_lmac_eeprom *eeprom; uint8_t h1_type; uint16_t h2_type; int actlen, sumlen; usbd_xfer_status(xfer, &actlen, &sumlen, NULL, NULL); UPGT_ASSERT_LOCKED(sc); if (actlen < 1) return (NULL); /* Check only at the very beginning. */ if (!(sc->sc_flags & UPGT_FLAG_FWLOADED) && (memcmp(data->buf, "OK", 2) == 0)) { sc->sc_flags |= UPGT_FLAG_FWLOADED; wakeup_one(sc); return (NULL); } if (actlen < (int)UPGT_RX_MINSZ) return (NULL); /* * Check what type of frame came in. */ header = (struct upgt_lmac_header *)(data->buf + 4); h1_type = header->header1.type; h2_type = le16toh(header->header2.type); if (h1_type == UPGT_H1_TYPE_CTRL && h2_type == UPGT_H2_TYPE_EEPROM) { eeprom = (struct upgt_lmac_eeprom *)(data->buf + 4); uint16_t eeprom_offset = le16toh(eeprom->offset); uint16_t eeprom_len = le16toh(eeprom->len); DPRINTF(sc, UPGT_DEBUG_FW, "received EEPROM block (offset=%d, len=%d)\n", eeprom_offset, eeprom_len); memcpy(sc->sc_eeprom + eeprom_offset, data->buf + sizeof(struct upgt_lmac_eeprom) + 4, eeprom_len); /* EEPROM data has arrived in time, wakeup. */ wakeup(sc); } else if (h1_type == UPGT_H1_TYPE_CTRL && h2_type == UPGT_H2_TYPE_TX_DONE) { DPRINTF(sc, UPGT_DEBUG_XMIT, "%s: received 802.11 TX done\n", __func__); upgt_tx_done(sc, data->buf + 4); } else if (h1_type == UPGT_H1_TYPE_RX_DATA || h1_type == UPGT_H1_TYPE_RX_DATA_MGMT) { DPRINTF(sc, UPGT_DEBUG_RECV, "%s: received 802.11 RX data\n", __func__); m = upgt_rx(sc, data->buf + 4, le16toh(header->header1.len), rssi); } else if (h1_type == UPGT_H1_TYPE_CTRL && h2_type == UPGT_H2_TYPE_STATS) { DPRINTF(sc, UPGT_DEBUG_STAT, "%s: received statistic data\n", __func__); /* TODO: what could we do with the statistic data? */ } else { /* ignore unknown frame types */ DPRINTF(sc, UPGT_DEBUG_INTR, "received unknown frame type 0x%02x\n", header->header1.type); } return (m); } /* * The firmware awaits a checksum for each frame we send to it. * The algorithm used therefor is uncommon but somehow similar to CRC32. */ static uint32_t upgt_chksum_le(const uint32_t *buf, size_t size) { size_t i; uint32_t crc = 0; for (i = 0; i < size; i += sizeof(uint32_t)) { crc = htole32(crc ^ *buf++); crc = htole32((crc >> 5) ^ (crc << 3)); } return (crc); } static struct mbuf * upgt_rx(struct upgt_softc *sc, uint8_t *data, int pkglen, int *rssi) { struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; struct upgt_lmac_rx_desc *rxdesc; struct mbuf *m; /* * don't pass packets to the ieee80211 framework if the driver isn't * RUNNING. */ if (!(ifp->if_drv_flags & IFF_DRV_RUNNING)) return (NULL); /* access RX packet descriptor */ rxdesc = (struct upgt_lmac_rx_desc *)data; /* create mbuf which is suitable for strict alignment archs */ KASSERT((pkglen + ETHER_ALIGN) < MCLBYTES, ("A current mbuf storage is small (%d)", pkglen + ETHER_ALIGN)); m = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR); if (m == NULL) { device_printf(sc->sc_dev, "could not create RX mbuf\n"); return (NULL); } m_adj(m, ETHER_ALIGN); memcpy(mtod(m, char *), rxdesc->data, pkglen); /* trim FCS */ m->m_len = m->m_pkthdr.len = pkglen - IEEE80211_CRC_LEN; m->m_pkthdr.rcvif = ifp; if (ieee80211_radiotap_active(ic)) { struct upgt_rx_radiotap_header *tap = &sc->sc_rxtap; tap->wr_flags = 0; tap->wr_rate = upgt_rx_rate(sc, rxdesc->rate); tap->wr_antsignal = rxdesc->rssi; } if_inc_counter(ifp, IFCOUNTER_IPACKETS, 1); DPRINTF(sc, UPGT_DEBUG_RX_PROC, "%s: RX done\n", __func__); *rssi = rxdesc->rssi; return (m); } static uint8_t upgt_rx_rate(struct upgt_softc *sc, const int rate) { struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; static const uint8_t cck_upgt2rate[4] = { 2, 4, 11, 22 }; static const uint8_t ofdm_upgt2rate[12] = { 2, 4, 11, 22, 12, 18, 24, 36, 48, 72, 96, 108 }; if (ic->ic_curmode == IEEE80211_MODE_11B && !(rate < 0 || rate > 3)) return cck_upgt2rate[rate & 0xf]; if (ic->ic_curmode == IEEE80211_MODE_11G && !(rate < 0 || rate > 11)) return ofdm_upgt2rate[rate & 0xf]; return (0); } static void upgt_tx_done(struct upgt_softc *sc, uint8_t *data) { struct ifnet *ifp = sc->sc_ifp; struct upgt_lmac_tx_done_desc *desc; int i, freed = 0; UPGT_ASSERT_LOCKED(sc); desc = (struct upgt_lmac_tx_done_desc *)data; for (i = 0; i < UPGT_TX_MAXCOUNT; i++) { struct upgt_data *data_tx = &sc->sc_tx_data[i]; if (data_tx->addr == le32toh(desc->header2.reqid)) { upgt_mem_free(sc, data_tx->addr); data_tx->ni = NULL; data_tx->addr = 0; data_tx->m = NULL; DPRINTF(sc, UPGT_DEBUG_TX_PROC, "TX done: memaddr=0x%08x, status=0x%04x, rssi=%d, ", le32toh(desc->header2.reqid), le16toh(desc->status), le16toh(desc->rssi)); DPRINTF(sc, UPGT_DEBUG_TX_PROC, "seq=%d\n", le16toh(desc->seq)); freed++; } } if (freed != 0) { sc->sc_tx_timer = 0; ifp->if_drv_flags &= ~IFF_DRV_OACTIVE; UPGT_UNLOCK(sc); upgt_start(ifp); UPGT_LOCK(sc); } } static void upgt_mem_free(struct upgt_softc *sc, uint32_t addr) { int i; for (i = 0; i < sc->sc_memory.pages; i++) { if (sc->sc_memory.page[i].addr == addr) { sc->sc_memory.page[i].used = 0; return; } } device_printf(sc->sc_dev, "could not free memory address 0x%08x\n", addr); } static int upgt_fw_load(struct upgt_softc *sc) { const struct firmware *fw; struct upgt_data *data_cmd; struct upgt_fw_x2_header *x2; char start_fwload_cmd[] = { 0x3c, 0x0d }; int error = 0; size_t offset; int bsize; int n; uint32_t crc32; fw = firmware_get(upgt_fwname); if (fw == NULL) { device_printf(sc->sc_dev, "could not read microcode %s\n", upgt_fwname); return (EIO); } UPGT_LOCK(sc); /* send firmware start load command */ data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { error = ENOBUFS; goto fail; } data_cmd->buflen = sizeof(start_fwload_cmd); memcpy(data_cmd->buf, start_fwload_cmd, data_cmd->buflen); upgt_bulk_tx(sc, data_cmd); /* send X2 header */ data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { error = ENOBUFS; goto fail; } data_cmd->buflen = sizeof(struct upgt_fw_x2_header); x2 = (struct upgt_fw_x2_header *)data_cmd->buf; memcpy(x2->signature, UPGT_X2_SIGNATURE, UPGT_X2_SIGNATURE_SIZE); x2->startaddr = htole32(UPGT_MEMADDR_FIRMWARE_START); x2->len = htole32(fw->datasize); x2->crc = upgt_crc32_le((uint8_t *)data_cmd->buf + UPGT_X2_SIGNATURE_SIZE, sizeof(struct upgt_fw_x2_header) - UPGT_X2_SIGNATURE_SIZE - sizeof(uint32_t)); upgt_bulk_tx(sc, data_cmd); /* download firmware */ for (offset = 0; offset < fw->datasize; offset += bsize) { if (fw->datasize - offset > UPGT_FW_BLOCK_SIZE) bsize = UPGT_FW_BLOCK_SIZE; else bsize = fw->datasize - offset; data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { error = ENOBUFS; goto fail; } n = upgt_fw_copy((const uint8_t *)fw->data + offset, data_cmd->buf, bsize); data_cmd->buflen = bsize; upgt_bulk_tx(sc, data_cmd); DPRINTF(sc, UPGT_DEBUG_FW, "FW offset=%d, read=%d, sent=%d\n", offset, n, bsize); bsize = n; } DPRINTF(sc, UPGT_DEBUG_FW, "%s: firmware downloaded\n", __func__); /* load firmware */ data_cmd = upgt_getbuf(sc); if (data_cmd == NULL) { error = ENOBUFS; goto fail; } crc32 = upgt_crc32_le(fw->data, fw->datasize); *((uint32_t *)(data_cmd->buf) ) = crc32; *((uint8_t *)(data_cmd->buf) + 4) = 'g'; *((uint8_t *)(data_cmd->buf) + 5) = '\r'; data_cmd->buflen = 6; upgt_bulk_tx(sc, data_cmd); /* waiting 'OK' response. */ usbd_transfer_start(sc->sc_xfer[UPGT_BULK_RX]); error = mtx_sleep(sc, &sc->sc_mtx, 0, "upgtfw", 2 * hz); if (error != 0) { device_printf(sc->sc_dev, "firmware load failed\n"); error = EIO; } DPRINTF(sc, UPGT_DEBUG_FW, "%s: firmware loaded\n", __func__); fail: UPGT_UNLOCK(sc); firmware_put(fw, FIRMWARE_UNLOAD); return (error); } static uint32_t upgt_crc32_le(const void *buf, size_t size) { uint32_t crc; crc = ether_crc32_le(buf, size); /* apply final XOR value as common for CRC-32 */ crc = htole32(crc ^ 0xffffffffU); return (crc); } /* * While copying the version 2 firmware, we need to replace two characters: * * 0x7e -> 0x7d 0x5e * 0x7d -> 0x7d 0x5d */ static int upgt_fw_copy(const uint8_t *src, char *dst, int size) { int i, j; for (i = 0, j = 0; i < size && j < size; i++) { switch (src[i]) { case 0x7e: dst[j] = 0x7d; j++; dst[j] = 0x5e; j++; break; case 0x7d: dst[j] = 0x7d; j++; dst[j] = 0x5d; j++; break; default: dst[j] = src[i]; j++; break; } } return (i); } static int upgt_mem_init(struct upgt_softc *sc) { int i; for (i = 0; i < UPGT_MEMORY_MAX_PAGES; i++) { sc->sc_memory.page[i].used = 0; if (i == 0) { /* * The first memory page is always reserved for * command data. */ sc->sc_memory.page[i].addr = sc->sc_memaddr_frame_start + MCLBYTES; } else { sc->sc_memory.page[i].addr = sc->sc_memory.page[i - 1].addr + MCLBYTES; } if (sc->sc_memory.page[i].addr + MCLBYTES >= sc->sc_memaddr_frame_end) break; DPRINTF(sc, UPGT_DEBUG_FW, "memory address page %d=0x%08x\n", i, sc->sc_memory.page[i].addr); } sc->sc_memory.pages = i; DPRINTF(sc, UPGT_DEBUG_FW, "memory pages=%d\n", sc->sc_memory.pages); return (0); } static int upgt_fw_verify(struct upgt_softc *sc) { const struct firmware *fw; const struct upgt_fw_bra_option *bra_opt; const struct upgt_fw_bra_descr *descr; const uint8_t *p; const uint32_t *uc; uint32_t bra_option_type, bra_option_len; size_t offset; int bra_end = 0; int error = 0; fw = firmware_get(upgt_fwname); if (fw == NULL) { device_printf(sc->sc_dev, "could not read microcode %s\n", upgt_fwname); return EIO; } /* * Seek to beginning of Boot Record Area (BRA). */ for (offset = 0; offset < fw->datasize; offset += sizeof(*uc)) { uc = (const uint32_t *)((const uint8_t *)fw->data + offset); if (*uc == 0) break; } for (; offset < fw->datasize; offset += sizeof(*uc)) { uc = (const uint32_t *)((const uint8_t *)fw->data + offset); if (*uc != 0) break; } if (offset == fw->datasize) { device_printf(sc->sc_dev, "firmware Boot Record Area not found\n"); error = EIO; goto fail; } DPRINTF(sc, UPGT_DEBUG_FW, "firmware Boot Record Area found at offset %d\n", offset); /* * Parse Boot Record Area (BRA) options. */ while (offset < fw->datasize && bra_end == 0) { /* get current BRA option */ p = (const uint8_t *)fw->data + offset; bra_opt = (const struct upgt_fw_bra_option *)p; bra_option_type = le32toh(bra_opt->type); bra_option_len = le32toh(bra_opt->len) * sizeof(*uc); switch (bra_option_type) { case UPGT_BRA_TYPE_FW: DPRINTF(sc, UPGT_DEBUG_FW, "UPGT_BRA_TYPE_FW len=%d\n", bra_option_len); if (bra_option_len != UPGT_BRA_FWTYPE_SIZE) { device_printf(sc->sc_dev, "wrong UPGT_BRA_TYPE_FW len\n"); error = EIO; goto fail; } if (memcmp(UPGT_BRA_FWTYPE_LM86, bra_opt->data, bra_option_len) == 0) { sc->sc_fw_type = UPGT_FWTYPE_LM86; break; } if (memcmp(UPGT_BRA_FWTYPE_LM87, bra_opt->data, bra_option_len) == 0) { sc->sc_fw_type = UPGT_FWTYPE_LM87; break; } device_printf(sc->sc_dev, "unsupported firmware type\n"); error = EIO; goto fail; case UPGT_BRA_TYPE_VERSION: DPRINTF(sc, UPGT_DEBUG_FW, "UPGT_BRA_TYPE_VERSION len=%d\n", bra_option_len); break; case UPGT_BRA_TYPE_DEPIF: DPRINTF(sc, UPGT_DEBUG_FW, "UPGT_BRA_TYPE_DEPIF len=%d\n", bra_option_len); break; case UPGT_BRA_TYPE_EXPIF: DPRINTF(sc, UPGT_DEBUG_FW, "UPGT_BRA_TYPE_EXPIF len=%d\n", bra_option_len); break; case UPGT_BRA_TYPE_DESCR: DPRINTF(sc, UPGT_DEBUG_FW, "UPGT_BRA_TYPE_DESCR len=%d\n", bra_option_len); descr = (const struct upgt_fw_bra_descr *)bra_opt->data; sc->sc_memaddr_frame_start = le32toh(descr->memaddr_space_start); sc->sc_memaddr_frame_end = le32toh(descr->memaddr_space_end); DPRINTF(sc, UPGT_DEBUG_FW, "memory address space start=0x%08x\n", sc->sc_memaddr_frame_start); DPRINTF(sc, UPGT_DEBUG_FW, "memory address space end=0x%08x\n", sc->sc_memaddr_frame_end); break; case UPGT_BRA_TYPE_END: DPRINTF(sc, UPGT_DEBUG_FW, "UPGT_BRA_TYPE_END len=%d\n", bra_option_len); bra_end = 1; break; default: DPRINTF(sc, UPGT_DEBUG_FW, "unknown BRA option len=%d\n", bra_option_len); error = EIO; goto fail; } /* jump to next BRA option */ offset += sizeof(struct upgt_fw_bra_option) + bra_option_len; } DPRINTF(sc, UPGT_DEBUG_FW, "%s: firmware verified", __func__); fail: firmware_put(fw, FIRMWARE_UNLOAD); return (error); } static void upgt_bulk_tx(struct upgt_softc *sc, struct upgt_data *data) { UPGT_ASSERT_LOCKED(sc); STAILQ_INSERT_TAIL(&sc->sc_tx_pending, data, next); UPGT_STAT_INC(sc, st_tx_pending); usbd_transfer_start(sc->sc_xfer[UPGT_BULK_TX]); } static int upgt_device_reset(struct upgt_softc *sc) { struct upgt_data *data; char init_cmd[] = { 0x7e, 0x7e, 0x7e, 0x7e }; UPGT_LOCK(sc); data = upgt_getbuf(sc); if (data == NULL) { UPGT_UNLOCK(sc); return (ENOBUFS); } memcpy(data->buf, init_cmd, sizeof(init_cmd)); data->buflen = sizeof(init_cmd); upgt_bulk_tx(sc, data); usb_pause_mtx(&sc->sc_mtx, 100); UPGT_UNLOCK(sc); DPRINTF(sc, UPGT_DEBUG_FW, "%s: device initialized\n", __func__); return (0); } static int upgt_alloc_tx(struct upgt_softc *sc) { int i; STAILQ_INIT(&sc->sc_tx_active); STAILQ_INIT(&sc->sc_tx_inactive); STAILQ_INIT(&sc->sc_tx_pending); for (i = 0; i < UPGT_TX_MAXCOUNT; i++) { struct upgt_data *data = &sc->sc_tx_data[i]; data->buf = ((uint8_t *)sc->sc_tx_dma_buf) + (i * MCLBYTES); STAILQ_INSERT_TAIL(&sc->sc_tx_inactive, data, next); UPGT_STAT_INC(sc, st_tx_inactive); } return (0); } static int upgt_alloc_rx(struct upgt_softc *sc) { int i; STAILQ_INIT(&sc->sc_rx_active); STAILQ_INIT(&sc->sc_rx_inactive); for (i = 0; i < UPGT_RX_MAXCOUNT; i++) { struct upgt_data *data = &sc->sc_rx_data[i]; data->buf = ((uint8_t *)sc->sc_rx_dma_buf) + (i * MCLBYTES); STAILQ_INSERT_TAIL(&sc->sc_rx_inactive, data, next); } return (0); } static int upgt_detach(device_t dev) { struct upgt_softc *sc = device_get_softc(dev); struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; unsigned int x; /* * Prevent further allocations from RX/TX/CMD * data lists and ioctls */ UPGT_LOCK(sc); sc->sc_flags |= UPGT_FLAG_DETACHED; STAILQ_INIT(&sc->sc_tx_active); STAILQ_INIT(&sc->sc_tx_inactive); STAILQ_INIT(&sc->sc_tx_pending); STAILQ_INIT(&sc->sc_rx_active); STAILQ_INIT(&sc->sc_rx_inactive); UPGT_UNLOCK(sc); upgt_stop(sc); callout_drain(&sc->sc_led_ch); callout_drain(&sc->sc_watchdog_ch); /* drain USB transfers */ for (x = 0; x != UPGT_N_XFERS; x++) usbd_transfer_drain(sc->sc_xfer[x]); /* free data buffers */ UPGT_LOCK(sc); upgt_free_rx(sc); upgt_free_tx(sc); UPGT_UNLOCK(sc); /* free USB transfers and some data buffers */ usbd_transfer_unsetup(sc->sc_xfer, UPGT_N_XFERS); ieee80211_ifdetach(ic); if_free(ifp); mtx_destroy(&sc->sc_mtx); return (0); } static void upgt_free_rx(struct upgt_softc *sc) { int i; for (i = 0; i < UPGT_RX_MAXCOUNT; i++) { struct upgt_data *data = &sc->sc_rx_data[i]; data->buf = NULL; data->ni = NULL; } } static void upgt_free_tx(struct upgt_softc *sc) { int i; for (i = 0; i < UPGT_TX_MAXCOUNT; i++) { struct upgt_data *data = &sc->sc_tx_data[i]; if (data->ni != NULL) ieee80211_free_node(data->ni); data->buf = NULL; data->ni = NULL; } } static void upgt_abort_xfers_locked(struct upgt_softc *sc) { int i; UPGT_ASSERT_LOCKED(sc); /* abort any pending transfers */ for (i = 0; i < UPGT_N_XFERS; i++) usbd_transfer_stop(sc->sc_xfer[i]); } static void upgt_abort_xfers(struct upgt_softc *sc) { UPGT_LOCK(sc); upgt_abort_xfers_locked(sc); UPGT_UNLOCK(sc); } #define UPGT_SYSCTL_STAT_ADD32(c, h, n, p, d) \ SYSCTL_ADD_UINT(c, h, OID_AUTO, n, CTLFLAG_RD, p, 0, d) static void upgt_sysctl_node(struct upgt_softc *sc) { struct sysctl_ctx_list *ctx; struct sysctl_oid_list *child; struct sysctl_oid *tree; struct upgt_stat *stats; stats = &sc->sc_stat; ctx = device_get_sysctl_ctx(sc->sc_dev); child = SYSCTL_CHILDREN(device_get_sysctl_tree(sc->sc_dev)); tree = SYSCTL_ADD_NODE(ctx, child, OID_AUTO, "stats", CTLFLAG_RD, NULL, "UPGT statistics"); child = SYSCTL_CHILDREN(tree); UPGT_SYSCTL_STAT_ADD32(ctx, child, "tx_active", &stats->st_tx_active, "Active numbers in TX queue"); UPGT_SYSCTL_STAT_ADD32(ctx, child, "tx_inactive", &stats->st_tx_inactive, "Inactive numbers in TX queue"); UPGT_SYSCTL_STAT_ADD32(ctx, child, "tx_pending", &stats->st_tx_pending, "Pending numbers in TX queue"); } #undef UPGT_SYSCTL_STAT_ADD32 static struct upgt_data * _upgt_getbuf(struct upgt_softc *sc) { struct upgt_data *bf; bf = STAILQ_FIRST(&sc->sc_tx_inactive); if (bf != NULL) { STAILQ_REMOVE_HEAD(&sc->sc_tx_inactive, next); UPGT_STAT_DEC(sc, st_tx_inactive); } else bf = NULL; if (bf == NULL) DPRINTF(sc, UPGT_DEBUG_XMIT, "%s: %s\n", __func__, "out of xmit buffers"); return (bf); } static struct upgt_data * upgt_getbuf(struct upgt_softc *sc) { struct upgt_data *bf; UPGT_ASSERT_LOCKED(sc); bf = _upgt_getbuf(sc); if (bf == NULL) { struct ifnet *ifp = sc->sc_ifp; DPRINTF(sc, UPGT_DEBUG_XMIT, "%s: stop queue\n", __func__); ifp->if_drv_flags |= IFF_DRV_OACTIVE; } return (bf); } static struct upgt_data * upgt_gettxbuf(struct upgt_softc *sc) { struct upgt_data *bf; UPGT_ASSERT_LOCKED(sc); bf = upgt_getbuf(sc); if (bf == NULL) return (NULL); bf->addr = upgt_mem_alloc(sc); if (bf->addr == 0) { struct ifnet *ifp = sc->sc_ifp; DPRINTF(sc, UPGT_DEBUG_XMIT, "%s: no free prism memory!\n", __func__); STAILQ_INSERT_HEAD(&sc->sc_tx_inactive, bf, next); UPGT_STAT_INC(sc, st_tx_inactive); if (!(ifp->if_drv_flags & IFF_DRV_OACTIVE)) ifp->if_drv_flags |= IFF_DRV_OACTIVE; return (NULL); } return (bf); } static int upgt_tx_start(struct upgt_softc *sc, struct mbuf *m, struct ieee80211_node *ni, struct upgt_data *data) { struct ieee80211vap *vap = ni->ni_vap; int error = 0, len; struct ieee80211_frame *wh; struct ieee80211_key *k; struct ifnet *ifp = sc->sc_ifp; struct upgt_lmac_mem *mem; struct upgt_lmac_tx_desc *txdesc; UPGT_ASSERT_LOCKED(sc); upgt_set_led(sc, UPGT_LED_BLINK); /* * Software crypto. */ wh = mtod(m, struct ieee80211_frame *); if (wh->i_fc[1] & IEEE80211_FC1_PROTECTED) { k = ieee80211_crypto_encap(ni, m); if (k == NULL) { device_printf(sc->sc_dev, "ieee80211_crypto_encap returns NULL.\n"); error = EIO; goto done; } /* in case packet header moved, reset pointer */ wh = mtod(m, struct ieee80211_frame *); } /* Transmit the URB containing the TX data. */ memset(data->buf, 0, MCLBYTES); mem = (struct upgt_lmac_mem *)data->buf; mem->addr = htole32(data->addr); txdesc = (struct upgt_lmac_tx_desc *)(mem + 1); if ((wh->i_fc[0] & IEEE80211_FC0_TYPE_MASK) == IEEE80211_FC0_TYPE_MGT) { /* mgmt frames */ txdesc->header1.flags = UPGT_H1_FLAGS_TX_MGMT; /* always send mgmt frames at lowest rate (DS1) */ memset(txdesc->rates, 0x10, sizeof(txdesc->rates)); } else { /* data frames */ txdesc->header1.flags = UPGT_H1_FLAGS_TX_DATA; memcpy(txdesc->rates, sc->sc_cur_rateset, sizeof(txdesc->rates)); } txdesc->header1.type = UPGT_H1_TYPE_TX_DATA; txdesc->header1.len = htole16(m->m_pkthdr.len); txdesc->header2.reqid = htole32(data->addr); txdesc->header2.type = htole16(UPGT_H2_TYPE_TX_ACK_YES); txdesc->header2.flags = htole16(UPGT_H2_FLAGS_TX_ACK_YES); txdesc->type = htole32(UPGT_TX_DESC_TYPE_DATA); txdesc->pad3[0] = UPGT_TX_DESC_PAD3_SIZE; if (ieee80211_radiotap_active_vap(vap)) { struct upgt_tx_radiotap_header *tap = &sc->sc_txtap; tap->wt_flags = 0; tap->wt_rate = 0; /* XXX where to get from? */ ieee80211_radiotap_tx(vap, m); } /* copy frame below our TX descriptor header */ m_copydata(m, 0, m->m_pkthdr.len, data->buf + (sizeof(*mem) + sizeof(*txdesc))); /* calculate frame size */ len = sizeof(*mem) + sizeof(*txdesc) + m->m_pkthdr.len; /* we need to align the frame to a 4 byte boundary */ len = (len + 3) & ~3; /* calculate frame checksum */ mem->chksum = upgt_chksum_le((uint32_t *)txdesc, len - sizeof(*mem)); data->ni = ni; data->m = m; data->buflen = len; DPRINTF(sc, UPGT_DEBUG_XMIT, "%s: TX start data sending (%d bytes)\n", __func__, len); KASSERT(len <= MCLBYTES, ("mbuf is small for saving data")); upgt_bulk_tx(sc, data); done: /* * If we don't regulary read the device statistics, the RX queue * will stall. It's strange, but it works, so we keep reading * the statistics here. *shrug* */ if (!(ifp->if_get_counter(ifp, IFCOUNTER_OPACKETS) % UPGT_TX_STAT_INTERVAL)) upgt_get_stats(sc); return (error); } static void upgt_bulk_rx_callback(struct usb_xfer *xfer, usb_error_t error) { struct upgt_softc *sc = usbd_xfer_softc(xfer); struct ifnet *ifp = sc->sc_ifp; struct ieee80211com *ic = ifp->if_l2com; struct ieee80211_frame *wh; struct ieee80211_node *ni; struct mbuf *m = NULL; struct upgt_data *data; int8_t nf; int rssi = -1; UPGT_ASSERT_LOCKED(sc); switch (USB_GET_STATE(xfer)) { case USB_ST_TRANSFERRED: data = STAILQ_FIRST(&sc->sc_rx_active); if (data == NULL) goto setup; STAILQ_REMOVE_HEAD(&sc->sc_rx_active, next); m = upgt_rxeof(xfer, data, &rssi); STAILQ_INSERT_TAIL(&sc->sc_rx_inactive, data, next); /* FALLTHROUGH */ case USB_ST_SETUP: setup: data = STAILQ_FIRST(&sc->sc_rx_inactive); if (data == NULL) return; STAILQ_REMOVE_HEAD(&sc->sc_rx_inactive, next); STAILQ_INSERT_TAIL(&sc->sc_rx_active, data, next); usbd_xfer_set_frame_data(xfer, 0, data->buf, MCLBYTES); usbd_transfer_submit(xfer); /* * To avoid LOR we should unlock our private mutex here to call * ieee80211_input() because here is at the end of a USB * callback and safe to unlock. */ UPGT_UNLOCK(sc); if (m != NULL) { wh = mtod(m, struct ieee80211_frame *); ni = ieee80211_find_rxnode(ic, (struct ieee80211_frame_min *)wh); nf = -95; /* XXX */ if (ni != NULL) { (void) ieee80211_input(ni, m, rssi, nf); /* node is no longer needed */ ieee80211_free_node(ni); } else (void) ieee80211_input_all(ic, m, rssi, nf); m = NULL; } if ((ifp->if_drv_flags & IFF_DRV_OACTIVE) == 0 && !IFQ_IS_EMPTY(&ifp->if_snd)) upgt_start(ifp); UPGT_LOCK(sc); break; default: /* needs it to the inactive queue due to a error. */ data = STAILQ_FIRST(&sc->sc_rx_active); if (data != NULL) { STAILQ_REMOVE_HEAD(&sc->sc_rx_active, next); STAILQ_INSERT_TAIL(&sc->sc_rx_inactive, data, next); } if (error != USB_ERR_CANCELLED) { usbd_xfer_set_stall(xfer); if_inc_counter(ifp, IFCOUNTER_IERRORS, 1); goto setup; } break; } } static void upgt_bulk_tx_callback(struct usb_xfer *xfer, usb_error_t error) { struct upgt_softc *sc = usbd_xfer_softc(xfer); struct ifnet *ifp = sc->sc_ifp; struct upgt_data *data; UPGT_ASSERT_LOCKED(sc); switch (USB_GET_STATE(xfer)) { case USB_ST_TRANSFERRED: data = STAILQ_FIRST(&sc->sc_tx_active); if (data == NULL) goto setup; STAILQ_REMOVE_HEAD(&sc->sc_tx_active, next); UPGT_STAT_DEC(sc, st_tx_active); upgt_txeof(xfer, data); STAILQ_INSERT_TAIL(&sc->sc_tx_inactive, data, next); UPGT_STAT_INC(sc, st_tx_inactive); /* FALLTHROUGH */ case USB_ST_SETUP: setup: data = STAILQ_FIRST(&sc->sc_tx_pending); if (data == NULL) { DPRINTF(sc, UPGT_DEBUG_XMIT, "%s: empty pending queue\n", __func__); return; } STAILQ_REMOVE_HEAD(&sc->sc_tx_pending, next); UPGT_STAT_DEC(sc, st_tx_pending); STAILQ_INSERT_TAIL(&sc->sc_tx_active, data, next); UPGT_STAT_INC(sc, st_tx_active); usbd_xfer_set_frame_data(xfer, 0, data->buf, data->buflen); usbd_transfer_submit(xfer); UPGT_UNLOCK(sc); upgt_start(ifp); UPGT_LOCK(sc); break; default: data = STAILQ_FIRST(&sc->sc_tx_active); if (data == NULL) goto setup; if (data->ni != NULL) { ieee80211_free_node(data->ni); data->ni = NULL; if_inc_counter(ifp, IFCOUNTER_OERRORS, 1); } if (error != USB_ERR_CANCELLED) { usbd_xfer_set_stall(xfer); goto setup; } break; } } static device_method_t upgt_methods[] = { /* Device interface */ DEVMETHOD(device_probe, upgt_match), DEVMETHOD(device_attach, upgt_attach), DEVMETHOD(device_detach, upgt_detach), DEVMETHOD_END }; static driver_t upgt_driver = { .name = "upgt", .methods = upgt_methods, .size = sizeof(struct upgt_softc) }; static devclass_t upgt_devclass; DRIVER_MODULE(if_upgt, uhub, upgt_driver, upgt_devclass, NULL, 0); MODULE_VERSION(if_upgt, 1); MODULE_DEPEND(if_upgt, usb, 1, 1, 1); MODULE_DEPEND(if_upgt, wlan, 1, 1, 1); MODULE_DEPEND(if_upgt, upgtfw_fw, 1, 1, 1); Index: projects/building-blocks/sys/dev/virtio/mmio/virtio_mmio.c =================================================================== --- projects/building-blocks/sys/dev/virtio/mmio/virtio_mmio.c (revision 275663) +++ projects/building-blocks/sys/dev/virtio/mmio/virtio_mmio.c (revision 275664) @@ -1,800 +1,840 @@ /*- * Copyright (c) 2014 Ruslan Bukin + * Copyright (c) 2014 The FreeBSD Foundation * All rights reserved. * * This software was developed by SRI International and the University of * Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) * ("CTSRD"), as part of the DARPA CRASH research programme. * + * Portions of this software were developed by Andrew Turner + * under sponsorship from the FreeBSD Foundation. + * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * VirtIO MMIO interface. * This driver is heavily based on VirtIO PCI interface driver. */ /* * FDT example: * virtio_block@1000 { * compatible = "virtio,mmio"; * reg = <0x1000 0x100>; * interrupts = <63>; * interrupt-parent = <&GIC>; * }; */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "virtio_mmio_if.h" #include "virtio_bus_if.h" #include "virtio_if.h" #define PAGE_SHIFT 12 struct vtmmio_virtqueue { struct virtqueue *vtv_vq; int vtv_no_intr; }; struct vtmmio_softc { device_t dev; device_t platform; struct resource *res[2]; uint64_t vtmmio_features; uint32_t vtmmio_flags; /* This "bus" will only ever have one child. */ device_t vtmmio_child_dev; struct virtio_feature_desc *vtmmio_child_feat_desc; int vtmmio_nvqs; struct vtmmio_virtqueue *vtmmio_vqs; void *ih; }; static int vtmmio_probe(device_t); static int vtmmio_attach(device_t); static int vtmmio_detach(device_t); static int vtmmio_suspend(device_t); static int vtmmio_resume(device_t); static int vtmmio_shutdown(device_t); static void vtmmio_driver_added(device_t, driver_t *); static void vtmmio_child_detached(device_t, device_t); static int vtmmio_read_ivar(device_t, device_t, int, uintptr_t *); static int vtmmio_write_ivar(device_t, device_t, int, uintptr_t); static uint64_t vtmmio_negotiate_features(device_t, uint64_t); static int vtmmio_with_feature(device_t, uint64_t); static int vtmmio_alloc_virtqueues(device_t, int, int, struct vq_alloc_info *); static int vtmmio_setup_intr(device_t, enum intr_type); static void vtmmio_stop(device_t); static int vtmmio_reinit(device_t, uint64_t); static void vtmmio_reinit_complete(device_t); static void vtmmio_notify_virtqueue(device_t, uint16_t); static uint8_t vtmmio_get_status(device_t); static void vtmmio_set_status(device_t, uint8_t); static void vtmmio_read_dev_config(device_t, bus_size_t, void *, int); static void vtmmio_write_dev_config(device_t, bus_size_t, void *, int); static void vtmmio_describe_features(struct vtmmio_softc *, const char *, uint64_t); static void vtmmio_probe_and_attach_child(struct vtmmio_softc *); static int vtmmio_reinit_virtqueue(struct vtmmio_softc *, int); static void vtmmio_free_interrupts(struct vtmmio_softc *); static void vtmmio_free_virtqueues(struct vtmmio_softc *); static void vtmmio_release_child_resources(struct vtmmio_softc *); static void vtmmio_reset(struct vtmmio_softc *); static void vtmmio_select_virtqueue(struct vtmmio_softc *, int); static void vtmmio_vq_intr(void *); /* * I/O port read/write wrappers. */ -#define vtmmio_write_config_1(sc, o, v) \ - bus_write_1((sc)->res[0], (o), (v)); \ - VIRTIO_MMIO_NOTE(sc->platform, (o)) -#define vtmmio_write_config_2(sc, o, v) \ - bus_write_2((sc)->res[0], (o), (v)); \ - VIRTIO_MMIO_NOTE(sc->platform, (o)) -#define vtmmio_write_config_4(sc, o, v) \ - bus_write_4((sc)->res[0], (o), (v)); \ - VIRTIO_MMIO_NOTE(sc->platform, (o)) +#define vtmmio_write_config_1(sc, o, v) \ +do { \ + bus_write_1((sc)->res[0], (o), (v)); \ + if (sc->platform != NULL) \ + VIRTIO_MMIO_NOTE(sc->platform, (o), (v)); \ +} while (0) +#define vtmmio_write_config_2(sc, o, v) \ +do { \ + bus_write_2((sc)->res[0], (o), (v)); \ + if (sc->platform != NULL) \ + VIRTIO_MMIO_NOTE(sc->platform, (o), (v)); \ +} while (0) +#define vtmmio_write_config_4(sc, o, v) \ +do { \ + bus_write_4((sc)->res[0], (o), (v)); \ + if (sc->platform != NULL) \ + VIRTIO_MMIO_NOTE(sc->platform, (o), (v)); \ +} while (0) #define vtmmio_read_config_1(sc, o) \ bus_read_1((sc)->res[0], (o)) #define vtmmio_read_config_2(sc, o) \ bus_read_2((sc)->res[0], (o)) #define vtmmio_read_config_4(sc, o) \ bus_read_4((sc)->res[0], (o)) static device_method_t vtmmio_methods[] = { /* Device interface. */ DEVMETHOD(device_probe, vtmmio_probe), DEVMETHOD(device_attach, vtmmio_attach), DEVMETHOD(device_detach, vtmmio_detach), DEVMETHOD(device_suspend, vtmmio_suspend), DEVMETHOD(device_resume, vtmmio_resume), DEVMETHOD(device_shutdown, vtmmio_shutdown), /* Bus interface. */ DEVMETHOD(bus_driver_added, vtmmio_driver_added), DEVMETHOD(bus_child_detached, vtmmio_child_detached), DEVMETHOD(bus_read_ivar, vtmmio_read_ivar), DEVMETHOD(bus_write_ivar, vtmmio_write_ivar), /* VirtIO bus interface. */ DEVMETHOD(virtio_bus_negotiate_features, vtmmio_negotiate_features), DEVMETHOD(virtio_bus_with_feature, vtmmio_with_feature), DEVMETHOD(virtio_bus_alloc_virtqueues, vtmmio_alloc_virtqueues), DEVMETHOD(virtio_bus_setup_intr, vtmmio_setup_intr), DEVMETHOD(virtio_bus_stop, vtmmio_stop), DEVMETHOD(virtio_bus_reinit, vtmmio_reinit), DEVMETHOD(virtio_bus_reinit_complete, vtmmio_reinit_complete), DEVMETHOD(virtio_bus_notify_vq, vtmmio_notify_virtqueue), DEVMETHOD(virtio_bus_read_device_config, vtmmio_read_dev_config), DEVMETHOD(virtio_bus_write_device_config, vtmmio_write_dev_config), DEVMETHOD_END }; static driver_t vtmmio_driver = { "virtio_mmio", vtmmio_methods, sizeof(struct vtmmio_softc) }; devclass_t vtmmio_devclass; DRIVER_MODULE(virtio_mmio, simplebus, vtmmio_driver, vtmmio_devclass, 0, 0); MODULE_VERSION(virtio_mmio, 1); MODULE_DEPEND(virtio_mmio, simplebus, 1, 1, 1); MODULE_DEPEND(virtio_mmio, virtio, 1, 1, 1); static int vtmmio_setup_intr(device_t dev, enum intr_type type) { struct vtmmio_softc *sc; int rid; int err; sc = device_get_softc(dev); - err = VIRTIO_MMIO_SETUP_INTR(sc->platform, sc->dev, - vtmmio_vq_intr, sc); - if (err == 0) { - /* Okay we have backend-specific interrupts */ - return (0); + if (sc->platform != NULL) { + err = VIRTIO_MMIO_SETUP_INTR(sc->platform, sc->dev, + vtmmio_vq_intr, sc); + if (err == 0) { + /* Okay we have backend-specific interrupts */ + return (0); + } } rid = 0; sc->res[1] = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid, RF_ACTIVE); if (!sc->res[1]) { device_printf(dev, "Can't allocate interrupt\n"); return (ENXIO); } if (bus_setup_intr(dev, sc->res[1], INTR_TYPE_MISC | INTR_MPSAFE, NULL, vtmmio_vq_intr, sc, &sc->ih)) { device_printf(dev, "Can't setup the interrupt\n"); return (ENXIO); } return (0); } static int vtmmio_probe(device_t dev) { if (!ofw_bus_status_okay(dev)) return (ENXIO); if (!ofw_bus_is_compatible(dev, "virtio,mmio")) return (ENXIO); device_set_desc(dev, "VirtIO MMIO adapter"); return (BUS_PROBE_DEFAULT); } static int vtmmio_setup_platform(struct vtmmio_softc *sc) { phandle_t platform_node; struct fdt_ic *ic; phandle_t xref; phandle_t node; sc->platform = NULL; if ((node = ofw_bus_get_node(sc->dev)) == -1) return (ENXIO); if (OF_searchencprop(node, "platform", &xref, sizeof(xref)) == -1) { return (ENXIO); } platform_node = OF_node_from_xref(xref); SLIST_FOREACH(ic, &fdt_ic_list_head, fdt_ics) { if (ic->iph == platform_node) { sc->platform = ic->dev; break; } } if (sc->platform == NULL) { /* No platform-specific device. Ignore it. */ } return (0); } static int vtmmio_attach(device_t dev) { struct vtmmio_softc *sc; device_t child; int rid; sc = device_get_softc(dev); sc->dev = dev; vtmmio_setup_platform(sc); rid = 0; sc->res[0] = bus_alloc_resource_any(dev, SYS_RES_MEMORY, &rid, RF_ACTIVE); if (!sc->res[0]) { device_printf(dev, "Cannot allocate memory window.\n"); return (ENXIO); } vtmmio_reset(sc); /* Tell the host we've noticed this device. */ vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_ACK); if ((child = device_add_child(dev, NULL, -1)) == NULL) { device_printf(dev, "Cannot create child device.\n"); vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_FAILED); vtmmio_detach(dev); return (ENOMEM); } sc->vtmmio_child_dev = child; vtmmio_probe_and_attach_child(sc); return (0); } static int vtmmio_detach(device_t dev) { struct vtmmio_softc *sc; device_t child; int error; sc = device_get_softc(dev); if ((child = sc->vtmmio_child_dev) != NULL) { error = device_delete_child(dev, child); if (error) return (error); sc->vtmmio_child_dev = NULL; } vtmmio_reset(sc); if (sc->res[0] != NULL) { bus_release_resource(dev, SYS_RES_MEMORY, 0, sc->res[0]); sc->res[0] = NULL; } return (0); } static int vtmmio_suspend(device_t dev) { return (bus_generic_suspend(dev)); } static int vtmmio_resume(device_t dev) { return (bus_generic_resume(dev)); } static int vtmmio_shutdown(device_t dev) { (void) bus_generic_shutdown(dev); /* Forcibly stop the host device. */ vtmmio_stop(dev); return (0); } static void vtmmio_driver_added(device_t dev, driver_t *driver) { struct vtmmio_softc *sc; sc = device_get_softc(dev); vtmmio_probe_and_attach_child(sc); } static void vtmmio_child_detached(device_t dev, device_t child) { struct vtmmio_softc *sc; sc = device_get_softc(dev); vtmmio_reset(sc); vtmmio_release_child_resources(sc); } static int vtmmio_read_ivar(device_t dev, device_t child, int index, uintptr_t *result) { struct vtmmio_softc *sc; sc = device_get_softc(dev); if (sc->vtmmio_child_dev != child) return (ENOENT); switch (index) { case VIRTIO_IVAR_DEVTYPE: case VIRTIO_IVAR_SUBDEVICE: *result = vtmmio_read_config_4(sc, VIRTIO_MMIO_DEVICE_ID); break; case VIRTIO_IVAR_VENDOR: *result = vtmmio_read_config_4(sc, VIRTIO_MMIO_VENDOR_ID); break; default: return (ENOENT); } return (0); } static int vtmmio_write_ivar(device_t dev, device_t child, int index, uintptr_t value) { struct vtmmio_softc *sc; sc = device_get_softc(dev); if (sc->vtmmio_child_dev != child) return (ENOENT); switch (index) { case VIRTIO_IVAR_FEATURE_DESC: sc->vtmmio_child_feat_desc = (void *) value; break; default: return (ENOENT); } return (0); } static uint64_t vtmmio_negotiate_features(device_t dev, uint64_t child_features) { struct vtmmio_softc *sc; uint64_t host_features, features; sc = device_get_softc(dev); host_features = vtmmio_read_config_4(sc, VIRTIO_MMIO_HOST_FEATURES); vtmmio_describe_features(sc, "host", host_features); /* * Limit negotiated features to what the driver, virtqueue, and * host all support. */ features = host_features & child_features; features = virtqueue_filter_features(features); sc->vtmmio_features = features; vtmmio_describe_features(sc, "negotiated", features); vtmmio_write_config_4(sc, VIRTIO_MMIO_GUEST_FEATURES, features); return (features); } static int vtmmio_with_feature(device_t dev, uint64_t feature) { struct vtmmio_softc *sc; sc = device_get_softc(dev); return ((sc->vtmmio_features & feature) != 0); } static int vtmmio_alloc_virtqueues(device_t dev, int flags, int nvqs, struct vq_alloc_info *vq_info) { struct vtmmio_virtqueue *vqx; struct vq_alloc_info *info; struct vtmmio_softc *sc; struct virtqueue *vq; + uint32_t size; int idx, error; - uint16_t size; sc = device_get_softc(dev); if (sc->vtmmio_nvqs != 0) return (EALREADY); if (nvqs <= 0) return (EINVAL); sc->vtmmio_vqs = malloc(nvqs * sizeof(struct vtmmio_virtqueue), M_DEVBUF, M_NOWAIT | M_ZERO); if (sc->vtmmio_vqs == NULL) return (ENOMEM); for (idx = 0; idx < nvqs; idx++) { vqx = &sc->vtmmio_vqs[idx]; info = &vq_info[idx]; + vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_SEL, idx); + vtmmio_select_virtqueue(sc, idx); - size = vtmmio_read_config_2(sc, VIRTIO_MMIO_QUEUE_NUM); + size = vtmmio_read_config_4(sc, VIRTIO_MMIO_QUEUE_NUM_MAX); error = virtqueue_alloc(dev, idx, size, VIRTIO_MMIO_VRING_ALIGN, 0xFFFFFFFFUL, info, &vq); if (error) { device_printf(dev, "cannot allocate virtqueue %d: %d\n", idx, error); break; } + + vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_NUM, size); + vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_ALIGN, + VIRTIO_MMIO_VRING_ALIGN); #if 0 device_printf(dev, "virtqueue paddr 0x%08lx\n", (uint64_t)virtqueue_paddr(vq)); #endif vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_PFN, virtqueue_paddr(vq) >> PAGE_SHIFT); vqx->vtv_vq = *info->vqai_vq = vq; vqx->vtv_no_intr = info->vqai_intr == NULL; sc->vtmmio_nvqs++; } if (error) vtmmio_free_virtqueues(sc); return (error); } static void vtmmio_stop(device_t dev) { vtmmio_reset(device_get_softc(dev)); } static int vtmmio_reinit(device_t dev, uint64_t features) { struct vtmmio_softc *sc; int idx, error; sc = device_get_softc(dev); if (vtmmio_get_status(dev) != VIRTIO_CONFIG_STATUS_RESET) vtmmio_stop(dev); /* * Quickly drive the status through ACK and DRIVER. The device * does not become usable again until vtmmio_reinit_complete(). */ vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_ACK); vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_DRIVER); vtmmio_negotiate_features(dev, features); for (idx = 0; idx < sc->vtmmio_nvqs; idx++) { error = vtmmio_reinit_virtqueue(sc, idx); if (error) return (error); } return (0); } static void vtmmio_reinit_complete(device_t dev) { vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_DRIVER_OK); } static void vtmmio_notify_virtqueue(device_t dev, uint16_t queue) { struct vtmmio_softc *sc; sc = device_get_softc(dev); - vtmmio_write_config_2(sc, VIRTIO_MMIO_QUEUE_NOTIFY, queue); + vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_NOTIFY, queue); } static uint8_t vtmmio_get_status(device_t dev) { struct vtmmio_softc *sc; sc = device_get_softc(dev); - return (vtmmio_read_config_1(sc, VIRTIO_MMIO_STATUS)); + return (vtmmio_read_config_4(sc, VIRTIO_MMIO_STATUS)); } static void vtmmio_set_status(device_t dev, uint8_t status) { struct vtmmio_softc *sc; sc = device_get_softc(dev); if (status != VIRTIO_CONFIG_STATUS_RESET) status |= vtmmio_get_status(dev); - vtmmio_write_config_1(sc, VIRTIO_MMIO_STATUS, status); + vtmmio_write_config_4(sc, VIRTIO_MMIO_STATUS, status); } static void vtmmio_read_dev_config(device_t dev, bus_size_t offset, void *dst, int length) { struct vtmmio_softc *sc; bus_size_t off; uint8_t *d; int size; sc = device_get_softc(dev); off = VIRTIO_MMIO_CONFIG + offset; for (d = dst; length > 0; d += size, off += size, length -= size) { +#ifdef ALLOW_WORD_ALIGNED_ACCESS if (length >= 4) { size = 4; *(uint32_t *)d = vtmmio_read_config_4(sc, off); } else if (length >= 2) { size = 2; *(uint16_t *)d = vtmmio_read_config_2(sc, off); - } else { + } else +#endif + { size = 1; *d = vtmmio_read_config_1(sc, off); } } } static void vtmmio_write_dev_config(device_t dev, bus_size_t offset, void *src, int length) { struct vtmmio_softc *sc; bus_size_t off; uint8_t *s; int size; sc = device_get_softc(dev); off = VIRTIO_MMIO_CONFIG + offset; for (s = src; length > 0; s += size, off += size, length -= size) { +#ifdef ALLOW_WORD_ALIGNED_ACCESS if (length >= 4) { size = 4; vtmmio_write_config_4(sc, off, *(uint32_t *)s); } else if (length >= 2) { size = 2; vtmmio_write_config_2(sc, off, *(uint16_t *)s); - } else { + } else +#endif + { size = 1; vtmmio_write_config_1(sc, off, *s); } } } static void vtmmio_describe_features(struct vtmmio_softc *sc, const char *msg, uint64_t features) { device_t dev, child; dev = sc->dev; child = sc->vtmmio_child_dev; if (device_is_attached(child) && bootverbose == 0) return; virtio_describe(dev, msg, features, sc->vtmmio_child_feat_desc); } static void vtmmio_probe_and_attach_child(struct vtmmio_softc *sc) { device_t dev, child; dev = sc->dev; child = sc->vtmmio_child_dev; if (child == NULL) return; if (device_get_state(child) != DS_NOTPRESENT) { return; } if (device_probe(child) != 0) { return; } vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_DRIVER); if (device_attach(child) != 0) { vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_FAILED); vtmmio_reset(sc); vtmmio_release_child_resources(sc); /* Reset status for future attempt. */ vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_ACK); } else { vtmmio_set_status(dev, VIRTIO_CONFIG_STATUS_DRIVER_OK); VIRTIO_ATTACH_COMPLETED(child); } } static int vtmmio_reinit_virtqueue(struct vtmmio_softc *sc, int idx) { struct vtmmio_virtqueue *vqx; struct virtqueue *vq; int error; uint16_t size; vqx = &sc->vtmmio_vqs[idx]; vq = vqx->vtv_vq; KASSERT(vq != NULL, ("%s: vq %d not allocated", __func__, idx)); vtmmio_select_virtqueue(sc, idx); - size = vtmmio_read_config_2(sc, VIRTIO_MMIO_QUEUE_NUM); + size = vtmmio_read_config_4(sc, VIRTIO_MMIO_QUEUE_NUM_MAX); error = virtqueue_reinit(vq, size); if (error) return (error); vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_PFN, virtqueue_paddr(vq) >> PAGE_SHIFT); return (0); } static void vtmmio_free_interrupts(struct vtmmio_softc *sc) { if (sc->ih != NULL) bus_teardown_intr(sc->dev, sc->res[1], sc->ih); if (sc->res[1] != NULL) bus_release_resource(sc->dev, SYS_RES_IRQ, 0, sc->res[1]); } static void vtmmio_free_virtqueues(struct vtmmio_softc *sc) { struct vtmmio_virtqueue *vqx; int idx; for (idx = 0; idx < sc->vtmmio_nvqs; idx++) { vqx = &sc->vtmmio_vqs[idx]; vtmmio_select_virtqueue(sc, idx); vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_PFN, 0); virtqueue_free(vqx->vtv_vq); vqx->vtv_vq = NULL; } free(sc->vtmmio_vqs, M_DEVBUF); sc->vtmmio_vqs = NULL; sc->vtmmio_nvqs = 0; } static void vtmmio_release_child_resources(struct vtmmio_softc *sc) { vtmmio_free_interrupts(sc); vtmmio_free_virtqueues(sc); } static void vtmmio_reset(struct vtmmio_softc *sc) { /* * Setting the status to RESET sets the host device to * the original, uninitialized state. */ vtmmio_set_status(sc->dev, VIRTIO_CONFIG_STATUS_RESET); } static void vtmmio_select_virtqueue(struct vtmmio_softc *sc, int idx) { - vtmmio_write_config_2(sc, VIRTIO_MMIO_QUEUE_SEL, idx); + vtmmio_write_config_4(sc, VIRTIO_MMIO_QUEUE_SEL, idx); } static void vtmmio_vq_intr(void *arg) { struct vtmmio_virtqueue *vqx; struct vtmmio_softc *sc; struct virtqueue *vq; + uint32_t status; int idx; sc = arg; + status = vtmmio_read_config_4(sc, VIRTIO_MMIO_INTERRUPT_STATUS); + vtmmio_write_config_4(sc, VIRTIO_MMIO_INTERRUPT_ACK, status); + + /* The config changed */ + if (status & VIRTIO_MMIO_INT_CONFIG) + if (sc->vtmmio_child_dev != NULL) + VIRTIO_CONFIG_CHANGE(sc->vtmmio_child_dev); + /* Notify all virtqueues. */ - for (idx = 0; idx < sc->vtmmio_nvqs; idx++) { - vqx = &sc->vtmmio_vqs[idx]; - vq = vqx->vtv_vq; - virtqueue_intr(vq); - }; + if (status & VIRTIO_MMIO_INT_VRING) { + for (idx = 0; idx < sc->vtmmio_nvqs; idx++) { + vqx = &sc->vtmmio_vqs[idx]; + if (vqx->vtv_no_intr == 0) { + vq = vqx->vtv_vq; + virtqueue_intr(vq); + } + } + } } Index: projects/building-blocks/sys/dev/virtio/mmio/virtio_mmio_if.m =================================================================== --- projects/building-blocks/sys/dev/virtio/mmio/virtio_mmio_if.m (revision 275663) +++ projects/building-blocks/sys/dev/virtio/mmio/virtio_mmio_if.m (revision 275664) @@ -1,76 +1,77 @@ #- # Copyright (c) 2014 Ruslan Bukin # All rights reserved. # # This software was developed by SRI International and the University of # Cambridge Computer Laboratory under DARPA/AFRL contract (FA8750-10-C-0237) # ("CTSRD"), as part of the DARPA CRASH research programme. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # 1. Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # 2. Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # # THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE # ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS # OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) # HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT # LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY # OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF # SUCH DAMAGE. # # $FreeBSD$ # #include # # This is optional interface to virtio mmio backend. # Useful when backend is implemented not by the hardware but software, e.g. # by using another cpu core. # INTERFACE virtio_mmio; CODE { static int - virtio_mmio_write(device_t dev, size_t offset) + virtio_mmio_note(device_t dev, size_t offset, int val) { return (1); } static int virtio_mmio_setup_intr(device_t dev, device_t mmio_dev, void *handler, void *ih_user) { return (1); } }; # # Inform backend we have data wrotten to offset. # METHOD int note { device_t dev; size_t offset; -} DEFAULT virtio_mmio_write; + int val; +} DEFAULT virtio_mmio_note; # # Setup backend-specific interrupts. # METHOD int setup_intr { device_t dev; device_t mmio_dev; void *handler; void *ih_user; } DEFAULT virtio_mmio_setup_intr; Index: projects/building-blocks/sys/dev/xen/pci/xen_acpi_pci.c =================================================================== --- projects/building-blocks/sys/dev/xen/pci/xen_acpi_pci.c (nonexistent) +++ projects/building-blocks/sys/dev/xen/pci/xen_acpi_pci.c (revision 275664) @@ -0,0 +1,88 @@ +/* + * Copyright (c) 2014 Roger Pau Monné + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include +#include +#include +#include +#include + +#include +#include + +#include + +#include +#include +#include +#include + +#include +#include +#include + +#include "pcib_if.h" +#include "pci_if.h" + +static int +xen_acpi_pci_probe(device_t dev) +{ + + if (!xen_pv_domain()) + return (ENXIO); + if (acpi_get_handle(dev) == NULL) + return (ENXIO); + + device_set_desc(dev, "Xen ACPI PCI bus"); + + return (BUS_PROBE_SPECIFIC); +} + +static device_method_t xen_acpi_pci_methods[] = { + /* Device interface */ + DEVMETHOD(device_probe, xen_acpi_pci_probe), + + /* PCI interface overwrites */ + DEVMETHOD(pci_enable_msi, xen_pci_enable_msi_method), + DEVMETHOD(pci_disable_msi, xen_pci_disable_msi_method), + DEVMETHOD(pci_child_added, xen_pci_child_added_method), + + DEVMETHOD_END +}; + +static devclass_t pci_devclass; + +DECLARE_CLASS(acpi_pci_driver); +DEFINE_CLASS_1(pci, xen_acpi_pci_driver, xen_acpi_pci_methods, + sizeof(struct pci_softc), acpi_pci_driver); +DRIVER_MODULE(xen_acpi_pci, pcib, xen_acpi_pci_driver, pci_devclass, 0, 0); +MODULE_DEPEND(xen_acpi_pci, pci, 1, 1, 1); +MODULE_DEPEND(xen_acpi_pci, acpi, 1, 1, 1); +MODULE_VERSION(xen_acpi_pci, 1); Property changes on: projects/building-blocks/sys/dev/xen/pci/xen_acpi_pci.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/sys/dev/xen/pci/xen_pci.c =================================================================== --- projects/building-blocks/sys/dev/xen/pci/xen_pci.c (nonexistent) +++ projects/building-blocks/sys/dev/xen/pci/xen_pci.c (revision 275664) @@ -0,0 +1,80 @@ +/* + * Copyright (c) 2014 Roger Pau Monné + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include +#include +#include +#include +#include + +#include +#include +#include +#include + +#include +#include +#include + +#include "pcib_if.h" +#include "pci_if.h" + +static int +xen_pci_probe(device_t dev) +{ + + if (!xen_pv_domain()) + return (ENXIO); + + device_set_desc(dev, "Xen PCI bus"); + + return (BUS_PROBE_DEFAULT); +} + +static device_method_t xen_pci_methods[] = { + /* Device interface */ + DEVMETHOD(device_probe, xen_pci_probe), + + /* PCI interface overwrites */ + DEVMETHOD(pci_enable_msi, xen_pci_enable_msi_method), + DEVMETHOD(pci_disable_msi, xen_pci_disable_msi_method), + DEVMETHOD(pci_child_added, xen_pci_child_added_method), + + DEVMETHOD_END +}; + +static devclass_t pci_devclass; + +DECLARE_CLASS(pci_driver); +DEFINE_CLASS_1(pci, xen_pci_driver, xen_pci_methods, sizeof(struct pci_softc), + pci_driver); +DRIVER_MODULE(xen_pci, pcib, xen_pci_driver, pci_devclass, 0, 0); +MODULE_DEPEND(xen_pci, pci, 1, 1, 1); +MODULE_VERSION(xen_pci, 1); Property changes on: projects/building-blocks/sys/dev/xen/pci/xen_pci.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/sys/fs/ext2fs/ext2_inode.c =================================================================== --- projects/building-blocks/sys/fs/ext2fs/ext2_inode.c (revision 275663) +++ projects/building-blocks/sys/fs/ext2fs/ext2_inode.c (revision 275664) @@ -1,515 +1,525 @@ /*- * modified for Lites 1.1 * * Aug 1995, Godmar Back (gback@cs.utah.edu) * University of Utah, Department of Computer Science */ /*- * Copyright (c) 1982, 1986, 1989, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)ffs_inode.c 8.5 (Berkeley) 12/30/93 * $FreeBSD$ */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static int ext2_indirtrunc(struct inode *, daddr_t, daddr_t, daddr_t, int, e4fs_daddr_t *); /* * Update the access, modified, and inode change times as specified by the * IN_ACCESS, IN_UPDATE, and IN_CHANGE flags respectively. Write the inode * to disk if the IN_MODIFIED flag is set (it may be set initially, or by * the timestamp update). The IN_LAZYMOD flag is set to force a write * later if not now. If we write now, then clear both IN_MODIFIED and * IN_LAZYMOD to reflect the presumably successful write, and if waitfor is * set, then wait for the write to complete. */ int ext2_update(struct vnode *vp, int waitfor) { struct m_ext2fs *fs; struct buf *bp; struct inode *ip; int error; ASSERT_VOP_ELOCKED(vp, "ext2_update"); ext2_itimes(vp); ip = VTOI(vp); if ((ip->i_flag & IN_MODIFIED) == 0 && waitfor == 0) return (0); ip->i_flag &= ~(IN_LAZYACCESS | IN_LAZYMOD | IN_MODIFIED); fs = ip->i_e2fs; if(fs->e2fs_ronly) return (0); if ((error = bread(ip->i_devvp, fsbtodb(fs, ino_to_fsba(fs, ip->i_number)), (int)fs->e2fs_bsize, NOCRED, &bp)) != 0) { brelse(bp); return (error); } ext2_i2ei(ip, (struct ext2fs_dinode *)((char *)bp->b_data + EXT2_INODE_SIZE(fs) * ino_to_fsbo(fs, ip->i_number))); if (waitfor && !DOINGASYNC(vp)) return (bwrite(bp)); else { bdwrite(bp); return (0); } } #define SINGLE 0 /* index of single indirect block */ #define DOUBLE 1 /* index of double indirect block */ #define TRIPLE 2 /* index of triple indirect block */ /* * Truncate the inode oip to at most length size, freeing the * disk blocks. */ int ext2_truncate(struct vnode *vp, off_t length, int flags, struct ucred *cred, struct thread *td) { struct vnode *ovp = vp; int32_t lastblock; struct inode *oip; int32_t bn, lbn, lastiblock[NIADDR], indir_lbn[NIADDR]; uint32_t oldblks[NDADDR + NIADDR], newblks[NDADDR + NIADDR]; struct bufobj *bo; struct m_ext2fs *fs; struct buf *bp; int offset, size, level; e4fs_daddr_t count, nblocks, blocksreleased = 0; int error, i, allerror; off_t osize; oip = VTOI(ovp); bo = &ovp->v_bufobj; ASSERT_VOP_LOCKED(vp, "ext2_truncate"); if (length < 0) return (EINVAL); if (ovp->v_type == VLNK && oip->i_size < ovp->v_mount->mnt_maxsymlinklen) { #ifdef INVARIANTS if (length != 0) panic("ext2_truncate: partial truncate of symlink"); #endif bzero((char *)&oip->i_shortlink, (u_int)oip->i_size); oip->i_size = 0; oip->i_flag |= IN_CHANGE | IN_UPDATE; return (ext2_update(ovp, 1)); } if (oip->i_size == length) { oip->i_flag |= IN_CHANGE | IN_UPDATE; return (ext2_update(ovp, 0)); } fs = oip->i_e2fs; osize = oip->i_size; /* * Lengthen the size of the file. We must ensure that the * last byte of the file is allocated. Since the smallest * value of osize is 0, length will be at least 1. */ if (osize < length) { if (length > oip->i_e2fs->e2fs_maxfilesize) return (EFBIG); vnode_pager_setsize(ovp, length); offset = blkoff(fs, length - 1); lbn = lblkno(fs, length - 1); flags |= BA_CLRBUF; error = ext2_balloc(oip, lbn, offset + 1, cred, &bp, flags); if (error) { vnode_pager_setsize(vp, osize); return (error); } oip->i_size = length; if (bp->b_bufsize == fs->e2fs_bsize) bp->b_flags |= B_CLUSTEROK; if (flags & IO_SYNC) bwrite(bp); else if (DOINGASYNC(ovp)) bdwrite(bp); else bawrite(bp); oip->i_flag |= IN_CHANGE | IN_UPDATE; return (ext2_update(ovp, !DOINGASYNC(ovp))); } /* * Shorten the size of the file. If the file is not being * truncated to a block boundry, the contents of the * partial block following the end of the file must be * zero'ed in case it ever become accessible again because * of subsequent file growth. */ /* I don't understand the comment above */ offset = blkoff(fs, length); if (offset == 0) { oip->i_size = length; } else { lbn = lblkno(fs, length); flags |= BA_CLRBUF; error = ext2_balloc(oip, lbn, offset, cred, &bp, flags); if (error) return (error); oip->i_size = length; size = blksize(fs, oip, lbn); bzero((char *)bp->b_data + offset, (u_int)(size - offset)); allocbuf(bp, size); if (bp->b_bufsize == fs->e2fs_bsize) bp->b_flags |= B_CLUSTEROK; if (flags & IO_SYNC) bwrite(bp); else if (DOINGASYNC(ovp)) bdwrite(bp); else bawrite(bp); } /* * Calculate index into inode's block list of * last direct and indirect blocks (if any) * which we want to keep. Lastblock is -1 when * the file is truncated to 0. */ lastblock = lblkno(fs, length + fs->e2fs_bsize - 1) - 1; lastiblock[SINGLE] = lastblock - NDADDR; lastiblock[DOUBLE] = lastiblock[SINGLE] - NINDIR(fs); lastiblock[TRIPLE] = lastiblock[DOUBLE] - NINDIR(fs) * NINDIR(fs); nblocks = btodb(fs->e2fs_bsize); /* * Update file and block pointers on disk before we start freeing * blocks. If we crash before free'ing blocks below, the blocks * will be returned to the free list. lastiblock values are also * normalized to -1 for calls to ext2_indirtrunc below. */ - bcopy((caddr_t)&oip->i_db[0], (caddr_t)oldblks, sizeof(oldblks)); - for (level = TRIPLE; level >= SINGLE; level--) + for (level = TRIPLE; level >= SINGLE; level--) { + oldblks[NDADDR + level] = oip->i_ib[level]; if (lastiblock[level] < 0) { oip->i_ib[level] = 0; lastiblock[level] = -1; } - for (i = NDADDR - 1; i > lastblock; i--) - oip->i_db[i] = 0; + } + for (i = 0; i < NDADDR; i++) { + oldblks[i] = oip->i_db[i]; + if (i > lastblock) + oip->i_db[i] = 0; + } oip->i_flag |= IN_CHANGE | IN_UPDATE; allerror = ext2_update(ovp, !DOINGASYNC(ovp)); /* * Having written the new inode to disk, save its new configuration * and put back the old block pointers long enough to process them. * Note that we save the new block configuration so we can check it * when we are done. */ - bcopy((caddr_t)&oip->i_db[0], (caddr_t)newblks, sizeof(newblks)); - bcopy((caddr_t)oldblks, (caddr_t)&oip->i_db[0], sizeof(oldblks)); + for (i = 0; i < NDADDR; i++) { + newblks[i] = oip->i_db[i]; + oip->i_db[i] = oldblks[i]; + } + for (i = 0; i < NIADDR; i++) { + newblks[NDADDR + i] = oip->i_ib[i]; + oip->i_ib[i] = oldblks[NDADDR + i]; + } oip->i_size = osize; error = vtruncbuf(ovp, cred, length, (int)fs->e2fs_bsize); if (error && (allerror == 0)) allerror = error; vnode_pager_setsize(ovp, length); /* * Indirect blocks first. */ indir_lbn[SINGLE] = -NDADDR; indir_lbn[DOUBLE] = indir_lbn[SINGLE] - NINDIR(fs) - 1; indir_lbn[TRIPLE] = indir_lbn[DOUBLE] - NINDIR(fs) * NINDIR(fs) - 1; for (level = TRIPLE; level >= SINGLE; level--) { bn = oip->i_ib[level]; if (bn != 0) { error = ext2_indirtrunc(oip, indir_lbn[level], fsbtodb(fs, bn), lastiblock[level], level, &count); if (error) allerror = error; blocksreleased += count; if (lastiblock[level] < 0) { oip->i_ib[level] = 0; ext2_blkfree(oip, bn, fs->e2fs_fsize); blocksreleased += nblocks; } } if (lastiblock[level] >= 0) goto done; } /* * All whole direct blocks or frags. */ for (i = NDADDR - 1; i > lastblock; i--) { long bsize; bn = oip->i_db[i]; if (bn == 0) continue; oip->i_db[i] = 0; bsize = blksize(fs, oip, i); ext2_blkfree(oip, bn, bsize); blocksreleased += btodb(bsize); } if (lastblock < 0) goto done; /* * Finally, look for a change in size of the * last direct block; release any frags. */ bn = oip->i_db[lastblock]; if (bn != 0) { long oldspace, newspace; /* * Calculate amount of space we're giving * back as old block size minus new block size. */ oldspace = blksize(fs, oip, lastblock); oip->i_size = length; newspace = blksize(fs, oip, lastblock); if (newspace == 0) panic("ext2_truncate: newspace"); if (oldspace - newspace > 0) { /* * Block number of space to be free'd is * the old block # plus the number of frags * required for the storage we're keeping. */ bn += numfrags(fs, newspace); ext2_blkfree(oip, bn, oldspace - newspace); blocksreleased += btodb(oldspace - newspace); } } done: #ifdef INVARIANTS for (level = SINGLE; level <= TRIPLE; level++) if (newblks[NDADDR + level] != oip->i_ib[level]) panic("itrunc1"); for (i = 0; i < NDADDR; i++) if (newblks[i] != oip->i_db[i]) panic("itrunc2"); BO_LOCK(bo); if (length == 0 && (bo->bo_dirty.bv_cnt != 0 || bo->bo_clean.bv_cnt != 0)) panic("itrunc3"); BO_UNLOCK(bo); #endif /* INVARIANTS */ /* * Put back the real size. */ oip->i_size = length; if (oip->i_blocks >= blocksreleased) oip->i_blocks -= blocksreleased; else /* sanity */ oip->i_blocks = 0; oip->i_flag |= IN_CHANGE; vnode_pager_setsize(ovp, length); return (allerror); } /* * Release blocks associated with the inode ip and stored in the indirect * block bn. Blocks are free'd in LIFO order up to (but not including) * lastbn. If level is greater than SINGLE, the block is an indirect block * and recursive calls to indirtrunc must be used to cleanse other indirect * blocks. * * NB: triple indirect blocks are untested. */ static int ext2_indirtrunc(struct inode *ip, daddr_t lbn, daddr_t dbn, daddr_t lastbn, int level, e4fs_daddr_t *countp) { struct buf *bp; struct m_ext2fs *fs = ip->i_e2fs; struct vnode *vp; e2fs_daddr_t *bap, *copy; int i, nblocks, error = 0, allerror = 0; e2fs_lbn_t nb, nlbn, last; e4fs_daddr_t blkcount, factor, blocksreleased = 0; /* * Calculate index in current block of last * block to be kept. -1 indicates the entire * block so we need not calculate the index. */ factor = 1; for (i = SINGLE; i < level; i++) factor *= NINDIR(fs); last = lastbn; if (lastbn > 0) last /= factor; nblocks = btodb(fs->e2fs_bsize); /* * Get buffer of block pointers, zero those entries corresponding * to blocks to be free'd, and update on disk copy first. Since * double(triple) indirect before single(double) indirect, calls * to bmap on these blocks will fail. However, we already have * the on disk address, so we have to set the b_blkno field * explicitly instead of letting bread do everything for us. */ vp = ITOV(ip); bp = getblk(vp, lbn, (int)fs->e2fs_bsize, 0, 0, 0); if ((bp->b_flags & (B_DONE | B_DELWRI)) == 0) { bp->b_iocmd = BIO_READ; if (bp->b_bcount > bp->b_bufsize) panic("ext2_indirtrunc: bad buffer size"); bp->b_blkno = dbn; vfs_busy_pages(bp, 0); bp->b_iooffset = dbtob(bp->b_blkno); bstrategy(bp); error = bufwait(bp); } if (error) { brelse(bp); *countp = 0; return (error); } bap = (e2fs_daddr_t *)bp->b_data; copy = malloc(fs->e2fs_bsize, M_TEMP, M_WAITOK); bcopy((caddr_t)bap, (caddr_t)copy, (u_int)fs->e2fs_bsize); bzero((caddr_t)&bap[last + 1], (NINDIR(fs) - (last + 1)) * sizeof(e2fs_daddr_t)); if (last == -1) bp->b_flags |= B_INVAL; if (DOINGASYNC(vp)) { bdwrite(bp); } else { error = bwrite(bp); if (error) allerror = error; } bap = copy; /* * Recursively free totally unused blocks. */ for (i = NINDIR(fs) - 1, nlbn = lbn + 1 - i * factor; i > last; i--, nlbn += factor) { nb = bap[i]; if (nb == 0) continue; if (level > SINGLE) { if ((error = ext2_indirtrunc(ip, nlbn, fsbtodb(fs, nb), (int32_t)-1, level - 1, &blkcount)) != 0) allerror = error; blocksreleased += blkcount; } ext2_blkfree(ip, nb, fs->e2fs_bsize); blocksreleased += nblocks; } /* * Recursively free last partial block. */ if (level > SINGLE && lastbn >= 0) { last = lastbn % factor; nb = bap[i]; if (nb != 0) { if ((error = ext2_indirtrunc(ip, nlbn, fsbtodb(fs, nb), last, level - 1, &blkcount)) != 0) allerror = error; blocksreleased += blkcount; } } free(copy, M_TEMP); *countp = blocksreleased; return (allerror); } /* * discard preallocated blocks */ int ext2_inactive(struct vop_inactive_args *ap) { struct vnode *vp = ap->a_vp; struct inode *ip = VTOI(vp); struct thread *td = ap->a_td; int mode, error = 0; /* * Ignore inodes related to stale file handles. */ if (ip->i_mode == 0) goto out; if (ip->i_nlink <= 0) { error = ext2_truncate(vp, (off_t)0, 0, NOCRED, td); ip->i_rdev = 0; mode = ip->i_mode; ip->i_mode = 0; ip->i_flag |= IN_CHANGE | IN_UPDATE; ext2_vfree(vp, ip->i_number, mode); } if (ip->i_flag & (IN_ACCESS | IN_CHANGE | IN_MODIFIED | IN_UPDATE)) ext2_update(vp, 0); out: /* * If we are done with the inode, reclaim it * so that it can be reused immediately. */ if (ip->i_mode == 0) vrecycle(vp); return (error); } /* * Reclaim an inode so that it can be used for other purposes. */ int ext2_reclaim(struct vop_reclaim_args *ap) { struct inode *ip; struct vnode *vp = ap->a_vp; ip = VTOI(vp); if (ip->i_flag & IN_LAZYMOD) { ip->i_flag |= IN_MODIFIED; ext2_update(vp, 0); } vfs_hash_remove(vp); free(vp->v_data, M_EXT2NODE); vp->v_data = 0; vnode_destroy_vobject(vp); return (0); } Index: projects/building-blocks/sys/fs/msdosfs/msdosfs_vfsops.c =================================================================== --- projects/building-blocks/sys/fs/msdosfs/msdosfs_vfsops.c (revision 275663) +++ projects/building-blocks/sys/fs/msdosfs/msdosfs_vfsops.c (revision 275664) @@ -1,1031 +1,1035 @@ /* $FreeBSD$ */ /* $NetBSD: msdosfs_vfsops.c,v 1.51 1997/11/17 15:36:58 ws Exp $ */ /*- * Copyright (C) 1994, 1995, 1997 Wolfgang Solfrank. * Copyright (C) 1994, 1995, 1997 TooLs GmbH. * All rights reserved. * Original code by Paul Popelka (paulp@uts.amdahl.com) (see below). * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by TooLs GmbH. * 4. The name of TooLs GmbH may not be used to endorse or promote products * derived from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY TOOLS GMBH ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL TOOLS GMBH BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; * OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, * WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR * OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF * ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ /*- * Written by Paul Popelka (paulp@uts.amdahl.com) * * You can do anything you want with this software, just don't say you wrote * it, and don't remove this notice. * * This software is provided "as is". * * The author supplies this software to be publicly redistributed on the * understanding that the author is not responsible for the correct * functioning of this software in any circumstances and is not liable for * any damages caused by this software. * * October 1992 */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static const char msdosfs_lock_msg[] = "fatlk"; /* Mount options that we support. */ static const char *msdosfs_opts[] = { "async", "noatime", "noclusterr", "noclusterw", "export", "force", "from", "sync", "cs_dos", "cs_local", "cs_win", "dirmask", "gid", "kiconv", "large", "longname", "longnames", "mask", "shortname", "shortnames", "uid", "win95", "nowin95", NULL }; #if 1 /*def PC98*/ /* * XXX - The boot signature formatted by NEC PC-98 DOS looks like a * garbage or a random value :-{ * If you want to use that broken-signatured media, define the * following symbol even though PC/AT. * (ex. mount PC-98 DOS formatted FD on PC/AT) */ #define MSDOSFS_NOCHECKSIG #endif MALLOC_DEFINE(M_MSDOSFSMNT, "msdosfs_mount", "MSDOSFS mount structure"); static MALLOC_DEFINE(M_MSDOSFSFAT, "msdosfs_fat", "MSDOSFS file allocation table"); struct iconv_functions *msdosfs_iconv; static int update_mp(struct mount *mp, struct thread *td); static int mountmsdosfs(struct vnode *devvp, struct mount *mp); static vfs_fhtovp_t msdosfs_fhtovp; static vfs_mount_t msdosfs_mount; static vfs_root_t msdosfs_root; static vfs_statfs_t msdosfs_statfs; static vfs_sync_t msdosfs_sync; static vfs_unmount_t msdosfs_unmount; /* Maximum length of a character set name (arbitrary). */ #define MAXCSLEN 64 static int update_mp(struct mount *mp, struct thread *td) { struct msdosfsmount *pmp = VFSTOMSDOSFS(mp); void *dos, *win, *local; int error, v; if (!vfs_getopt(mp->mnt_optnew, "kiconv", NULL, NULL)) { if (msdosfs_iconv != NULL) { error = vfs_getopt(mp->mnt_optnew, "cs_win", &win, NULL); if (!error) error = vfs_getopt(mp->mnt_optnew, "cs_local", &local, NULL); if (!error) error = vfs_getopt(mp->mnt_optnew, "cs_dos", &dos, NULL); if (!error) { msdosfs_iconv->open(win, local, &pmp->pm_u2w); msdosfs_iconv->open(local, win, &pmp->pm_w2u); msdosfs_iconv->open(dos, local, &pmp->pm_u2d); msdosfs_iconv->open(local, dos, &pmp->pm_d2u); } if (error != 0) return (error); } else { pmp->pm_w2u = NULL; pmp->pm_u2w = NULL; pmp->pm_d2u = NULL; pmp->pm_u2d = NULL; } } if (vfs_scanopt(mp->mnt_optnew, "gid", "%d", &v) == 1) pmp->pm_gid = v; if (vfs_scanopt(mp->mnt_optnew, "uid", "%d", &v) == 1) pmp->pm_uid = v; if (vfs_scanopt(mp->mnt_optnew, "mask", "%d", &v) == 1) pmp->pm_mask = v & ALLPERMS; if (vfs_scanopt(mp->mnt_optnew, "dirmask", "%d", &v) == 1) pmp->pm_dirmask = v & ALLPERMS; vfs_flagopt(mp->mnt_optnew, "shortname", &pmp->pm_flags, MSDOSFSMNT_SHORTNAME); vfs_flagopt(mp->mnt_optnew, "shortnames", &pmp->pm_flags, MSDOSFSMNT_SHORTNAME); vfs_flagopt(mp->mnt_optnew, "longname", &pmp->pm_flags, MSDOSFSMNT_LONGNAME); vfs_flagopt(mp->mnt_optnew, "longnames", &pmp->pm_flags, MSDOSFSMNT_LONGNAME); vfs_flagopt(mp->mnt_optnew, "kiconv", &pmp->pm_flags, MSDOSFSMNT_KICONV); if (vfs_getopt(mp->mnt_optnew, "nowin95", NULL, NULL) == 0) pmp->pm_flags |= MSDOSFSMNT_NOWIN95; else pmp->pm_flags &= ~MSDOSFSMNT_NOWIN95; if (pmp->pm_flags & MSDOSFSMNT_NOWIN95) pmp->pm_flags |= MSDOSFSMNT_SHORTNAME; else if (!(pmp->pm_flags & (MSDOSFSMNT_SHORTNAME | MSDOSFSMNT_LONGNAME))) { struct vnode *rootvp; /* * Try to divine whether to support Win'95 long filenames */ if (FAT32(pmp)) pmp->pm_flags |= MSDOSFSMNT_LONGNAME; else { if ((error = msdosfs_root(mp, LK_EXCLUSIVE, &rootvp)) != 0) return error; pmp->pm_flags |= findwin95(VTODE(rootvp)) ? MSDOSFSMNT_LONGNAME : MSDOSFSMNT_SHORTNAME; vput(rootvp); } } return 0; } static int msdosfs_cmount(struct mntarg *ma, void *data, uint64_t flags) { struct msdosfs_args args; struct export_args exp; int error; if (data == NULL) return (EINVAL); error = copyin(data, &args, sizeof args); if (error) return (error); vfs_oexport_conv(&args.export, &exp); ma = mount_argsu(ma, "from", args.fspec, MAXPATHLEN); ma = mount_arg(ma, "export", &exp, sizeof(exp)); ma = mount_argf(ma, "uid", "%d", args.uid); ma = mount_argf(ma, "gid", "%d", args.gid); ma = mount_argf(ma, "mask", "%d", args.mask); ma = mount_argf(ma, "dirmask", "%d", args.dirmask); ma = mount_argb(ma, args.flags & MSDOSFSMNT_SHORTNAME, "noshortname"); ma = mount_argb(ma, args.flags & MSDOSFSMNT_LONGNAME, "nolongname"); ma = mount_argb(ma, !(args.flags & MSDOSFSMNT_NOWIN95), "nowin95"); ma = mount_argb(ma, args.flags & MSDOSFSMNT_KICONV, "nokiconv"); ma = mount_argsu(ma, "cs_win", args.cs_win, MAXCSLEN); ma = mount_argsu(ma, "cs_dos", args.cs_dos, MAXCSLEN); ma = mount_argsu(ma, "cs_local", args.cs_local, MAXCSLEN); error = kernel_mount(ma, flags); return (error); } /* * mp - path - addr in user space of mount point (ie /usr or whatever) * data - addr in user space of mount params including the name of the block * special file to treat as a filesystem. */ static int msdosfs_mount(struct mount *mp) { struct vnode *devvp; /* vnode for blk device to mount */ struct thread *td; /* msdosfs specific mount control block */ struct msdosfsmount *pmp = NULL; struct nameidata ndp; int error, flags; accmode_t accmode; char *from; td = curthread; if (vfs_filteropt(mp->mnt_optnew, msdosfs_opts)) return (EINVAL); /* * If updating, check whether changing from read-only to * read/write; if there is no device name, that's all we do. */ if (mp->mnt_flag & MNT_UPDATE) { pmp = VFSTOMSDOSFS(mp); if (vfs_flagopt(mp->mnt_optnew, "export", NULL, 0)) { /* * Forbid export requests if filesystem has * MSDOSFS_LARGEFS flag set. */ if ((pmp->pm_flags & MSDOSFS_LARGEFS) != 0) { vfs_mount_error(mp, "MSDOSFS_LARGEFS flag set, cannot export"); return (EOPNOTSUPP); } } if (!(pmp->pm_flags & MSDOSFSMNT_RONLY) && vfs_flagopt(mp->mnt_optnew, "ro", NULL, 0)) { error = VFS_SYNC(mp, MNT_WAIT); if (error) return (error); flags = WRITECLOSE; if (mp->mnt_flag & MNT_FORCE) flags |= FORCECLOSE; error = vflush(mp, 0, flags, td); if (error) return (error); /* * Now the volume is clean. Mark it so while the * device is still rw. */ error = markvoldirty(pmp, 0); if (error) { (void)markvoldirty(pmp, 1); return (error); } /* Downgrade the device from rw to ro. */ DROP_GIANT(); g_topology_lock(); error = g_access(pmp->pm_cp, 0, -1, 0); g_topology_unlock(); PICKUP_GIANT(); if (error) { (void)markvoldirty(pmp, 1); return (error); } /* * Backing out after an error was painful in the * above. Now we are committed to succeeding. */ pmp->pm_fmod = 0; pmp->pm_flags |= MSDOSFSMNT_RONLY; MNT_ILOCK(mp); mp->mnt_flag |= MNT_RDONLY; MNT_IUNLOCK(mp); } else if ((pmp->pm_flags & MSDOSFSMNT_RONLY) && !vfs_flagopt(mp->mnt_optnew, "ro", NULL, 0)) { /* * If upgrade to read-write by non-root, then verify * that user has necessary permissions on the device. */ devvp = pmp->pm_devvp; vn_lock(devvp, LK_EXCLUSIVE | LK_RETRY); error = VOP_ACCESS(devvp, VREAD | VWRITE, td->td_ucred, td); if (error) error = priv_check(td, PRIV_VFS_MOUNT_PERM); if (error) { VOP_UNLOCK(devvp, 0); return (error); } VOP_UNLOCK(devvp, 0); DROP_GIANT(); g_topology_lock(); error = g_access(pmp->pm_cp, 0, 1, 0); g_topology_unlock(); PICKUP_GIANT(); if (error) return (error); pmp->pm_fmod = 1; pmp->pm_flags &= ~MSDOSFSMNT_RONLY; MNT_ILOCK(mp); mp->mnt_flag &= ~MNT_RDONLY; MNT_IUNLOCK(mp); /* Now that the volume is modifiable, mark it dirty. */ error = markvoldirty(pmp, 1); if (error) return (error); } } /* * Not an update, or updating the name: look up the name * and verify that it refers to a sensible disk device. */ if (vfs_getopt(mp->mnt_optnew, "from", (void **)&from, NULL)) return (EINVAL); NDINIT(&ndp, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, from, td); error = namei(&ndp); if (error) return (error); devvp = ndp.ni_vp; NDFREE(&ndp, NDF_ONLY_PNBUF); if (!vn_isdisk(devvp, &error)) { vput(devvp); return (error); } /* * If mount by non-root, then verify that user has necessary * permissions on the device. */ accmode = VREAD; if ((mp->mnt_flag & MNT_RDONLY) == 0) accmode |= VWRITE; error = VOP_ACCESS(devvp, accmode, td->td_ucred, td); if (error) error = priv_check(td, PRIV_VFS_MOUNT_PERM); if (error) { vput(devvp); return (error); } if ((mp->mnt_flag & MNT_UPDATE) == 0) { error = mountmsdosfs(devvp, mp); #ifdef MSDOSFS_DEBUG /* only needed for the printf below */ pmp = VFSTOMSDOSFS(mp); #endif } else { vput(devvp); if (devvp != pmp->pm_devvp) return (EINVAL); /* XXX needs translation */ } if (error) { vrele(devvp); return (error); } error = update_mp(mp, td); if (error) { if ((mp->mnt_flag & MNT_UPDATE) == 0) msdosfs_unmount(mp, MNT_FORCE); return error; } if (devvp->v_type == VCHR && devvp->v_rdev != NULL) devvp->v_rdev->si_mountpt = mp; vfs_mountedfrom(mp, from); #ifdef MSDOSFS_DEBUG printf("msdosfs_mount(): mp %p, pmp %p, inusemap %p\n", mp, pmp, pmp->pm_inusemap); #endif return (0); } static int mountmsdosfs(struct vnode *devvp, struct mount *mp) { struct msdosfsmount *pmp; struct buf *bp; struct cdev *dev; union bootsector *bsp; struct byte_bpb33 *b33; struct byte_bpb50 *b50; struct byte_bpb710 *b710; u_int8_t SecPerClust; u_long clusters; int ronly, error; struct g_consumer *cp; struct bufobj *bo; bp = NULL; /* This and pmp both used in error_exit. */ pmp = NULL; ronly = (mp->mnt_flag & MNT_RDONLY) != 0; dev = devvp->v_rdev; dev_ref(dev); DROP_GIANT(); g_topology_lock(); error = g_vfs_open(devvp, &cp, "msdosfs", ronly ? 0 : 1); g_topology_unlock(); PICKUP_GIANT(); VOP_UNLOCK(devvp, 0); if (error) goto error_exit; bo = &devvp->v_bufobj; /* * Read the boot sector of the filesystem, and then check the * boot signature. If not a dos boot sector then error out. * * NOTE: 8192 is a magic size that works for ffs. */ error = bread(devvp, 0, 8192, NOCRED, &bp); if (error) goto error_exit; bp->b_flags |= B_AGE; bsp = (union bootsector *)bp->b_data; b33 = (struct byte_bpb33 *)bsp->bs33.bsBPB; b50 = (struct byte_bpb50 *)bsp->bs50.bsBPB; b710 = (struct byte_bpb710 *)bsp->bs710.bsBPB; #ifndef MSDOSFS_NOCHECKSIG if (bsp->bs50.bsBootSectSig0 != BOOTSIG0 || bsp->bs50.bsBootSectSig1 != BOOTSIG1) { error = EINVAL; goto error_exit; } #endif pmp = malloc(sizeof *pmp, M_MSDOSFSMNT, M_WAITOK | M_ZERO); pmp->pm_mountp = mp; pmp->pm_cp = cp; pmp->pm_bo = bo; lockinit(&pmp->pm_fatlock, 0, msdosfs_lock_msg, 0, 0); /* * Initialize ownerships and permissions, since nothing else will * initialize them iff we are mounting root. */ pmp->pm_uid = UID_ROOT; pmp->pm_gid = GID_WHEEL; pmp->pm_mask = pmp->pm_dirmask = S_IXUSR | S_IXGRP | S_IXOTH | S_IRUSR | S_IRGRP | S_IROTH | S_IWUSR; /* * Experimental support for large MS-DOS filesystems. * WARNING: This uses at least 32 bytes of kernel memory (which is not * reclaimed until the FS is unmounted) for each file on disk to map * between the 32-bit inode numbers used by VFS and the 64-bit * pseudo-inode numbers used internally by msdosfs. This is only * safe to use in certain controlled situations (e.g. read-only FS * with less than 1 million files). * Since the mappings do not persist across unmounts (or reboots), these * filesystems are not suitable for exporting through NFS, or any other * application that requires fixed inode numbers. */ vfs_flagopt(mp->mnt_optnew, "large", &pmp->pm_flags, MSDOSFS_LARGEFS); /* * Compute several useful quantities from the bpb in the * bootsector. Copy in the dos 5 variant of the bpb then fix up * the fields that are different between dos 5 and dos 3.3. */ SecPerClust = b50->bpbSecPerClust; pmp->pm_BytesPerSec = getushort(b50->bpbBytesPerSec); if (pmp->pm_BytesPerSec < DEV_BSIZE) { error = EINVAL; goto error_exit; } pmp->pm_ResSectors = getushort(b50->bpbResSectors); pmp->pm_FATs = b50->bpbFATs; pmp->pm_RootDirEnts = getushort(b50->bpbRootDirEnts); pmp->pm_Sectors = getushort(b50->bpbSectors); pmp->pm_FATsecs = getushort(b50->bpbFATsecs); pmp->pm_SecPerTrack = getushort(b50->bpbSecPerTrack); pmp->pm_Heads = getushort(b50->bpbHeads); pmp->pm_Media = b50->bpbMedia; /* calculate the ratio of sector size to DEV_BSIZE */ pmp->pm_BlkPerSec = pmp->pm_BytesPerSec / DEV_BSIZE; /* * We don't check pm_Heads nor pm_SecPerTrack, because * these may not be set for EFI file systems. We don't * use these anyway, so we're unaffected if they are * invalid. */ if (!pmp->pm_BytesPerSec || !SecPerClust) { error = EINVAL; goto error_exit; } if (pmp->pm_Sectors == 0) { pmp->pm_HiddenSects = getulong(b50->bpbHiddenSecs); pmp->pm_HugeSectors = getulong(b50->bpbHugeSectors); } else { pmp->pm_HiddenSects = getushort(b33->bpbHiddenSecs); pmp->pm_HugeSectors = pmp->pm_Sectors; } if (!(pmp->pm_flags & MSDOSFS_LARGEFS)) { if (pmp->pm_HugeSectors > 0xffffffff / (pmp->pm_BytesPerSec / sizeof(struct direntry)) + 1) { /* * We cannot deal currently with this size of disk * due to fileid limitations (see msdosfs_getattr and * msdosfs_readdir) */ error = EINVAL; vfs_mount_error(mp, "Disk too big, try '-o large' mount option"); goto error_exit; } } if (pmp->pm_RootDirEnts == 0) { if (pmp->pm_FATsecs || getushort(b710->bpbFSVers)) { error = EINVAL; #ifdef MSDOSFS_DEBUG printf("mountmsdosfs(): bad FAT32 filesystem\n"); #endif goto error_exit; } pmp->pm_fatmask = FAT32_MASK; pmp->pm_fatmult = 4; pmp->pm_fatdiv = 1; pmp->pm_FATsecs = getulong(b710->bpbBigFATsecs); if (getushort(b710->bpbExtFlags) & FATMIRROR) pmp->pm_curfat = getushort(b710->bpbExtFlags) & FATNUM; else pmp->pm_flags |= MSDOSFS_FATMIRROR; } else pmp->pm_flags |= MSDOSFS_FATMIRROR; /* * Check a few values (could do some more): * - logical sector size: power of 2, >= block size * - sectors per cluster: power of 2, >= 1 * - number of sectors: >= 1, <= size of partition * - number of FAT sectors: >= 1 */ if ( (SecPerClust == 0) || (SecPerClust & (SecPerClust - 1)) || (pmp->pm_BytesPerSec < DEV_BSIZE) || (pmp->pm_BytesPerSec & (pmp->pm_BytesPerSec - 1)) || (pmp->pm_HugeSectors == 0) || (pmp->pm_FATsecs == 0) || (SecPerClust * pmp->pm_BlkPerSec > MAXBSIZE / DEV_BSIZE) ) { error = EINVAL; goto error_exit; } pmp->pm_HugeSectors *= pmp->pm_BlkPerSec; pmp->pm_HiddenSects *= pmp->pm_BlkPerSec; /* XXX not used? */ pmp->pm_FATsecs *= pmp->pm_BlkPerSec; SecPerClust *= pmp->pm_BlkPerSec; pmp->pm_fatblk = pmp->pm_ResSectors * pmp->pm_BlkPerSec; if (FAT32(pmp)) { pmp->pm_rootdirblk = getulong(b710->bpbRootClust); pmp->pm_firstcluster = pmp->pm_fatblk + (pmp->pm_FATs * pmp->pm_FATsecs); pmp->pm_fsinfo = getushort(b710->bpbFSInfo) * pmp->pm_BlkPerSec; } else { pmp->pm_rootdirblk = pmp->pm_fatblk + (pmp->pm_FATs * pmp->pm_FATsecs); pmp->pm_rootdirsize = (pmp->pm_RootDirEnts * sizeof(struct direntry) + DEV_BSIZE - 1) / DEV_BSIZE; /* in blocks */ pmp->pm_firstcluster = pmp->pm_rootdirblk + pmp->pm_rootdirsize; } pmp->pm_maxcluster = (pmp->pm_HugeSectors - pmp->pm_firstcluster) / SecPerClust + 1; pmp->pm_fatsize = pmp->pm_FATsecs * DEV_BSIZE; /* XXX not used? */ if (pmp->pm_fatmask == 0) { if (pmp->pm_maxcluster <= ((CLUST_RSRVD - CLUST_FIRST) & FAT12_MASK)) { /* * This will usually be a floppy disk. This size makes * sure that one fat entry will not be split across * multiple blocks. */ pmp->pm_fatmask = FAT12_MASK; pmp->pm_fatmult = 3; pmp->pm_fatdiv = 2; } else { pmp->pm_fatmask = FAT16_MASK; pmp->pm_fatmult = 2; pmp->pm_fatdiv = 1; } } clusters = (pmp->pm_fatsize / pmp->pm_fatmult) * pmp->pm_fatdiv; if (pmp->pm_maxcluster >= clusters) { #ifdef MSDOSFS_DEBUG printf("Warning: number of clusters (%ld) exceeds FAT " "capacity (%ld)\n", pmp->pm_maxcluster + 1, clusters); #endif pmp->pm_maxcluster = clusters - 1; } if (FAT12(pmp)) pmp->pm_fatblocksize = 3 * 512; else pmp->pm_fatblocksize = PAGE_SIZE; pmp->pm_fatblocksize = roundup(pmp->pm_fatblocksize, pmp->pm_BytesPerSec); pmp->pm_fatblocksec = pmp->pm_fatblocksize / DEV_BSIZE; pmp->pm_bnshift = ffs(DEV_BSIZE) - 1; /* * Compute mask and shift value for isolating cluster relative byte * offsets and cluster numbers from a file offset. */ pmp->pm_bpcluster = SecPerClust * DEV_BSIZE; pmp->pm_crbomask = pmp->pm_bpcluster - 1; pmp->pm_cnshift = ffs(pmp->pm_bpcluster) - 1; /* * Check for valid cluster size * must be a power of 2 */ if (pmp->pm_bpcluster ^ (1 << pmp->pm_cnshift)) { error = EINVAL; goto error_exit; } /* * Release the bootsector buffer. */ brelse(bp); bp = NULL; /* * Check the fsinfo sector if we have one. Silently fix up our * in-core copy of fp->fsinxtfree if it is unknown (0xffffffff) * or too large. Ignore fp->fsinfree for now, since we need to * read the entire FAT anyway to fill the inuse map. */ if (pmp->pm_fsinfo) { struct fsinfo *fp; if ((error = bread(devvp, pmp->pm_fsinfo, pmp->pm_BytesPerSec, NOCRED, &bp)) != 0) goto error_exit; fp = (struct fsinfo *)bp->b_data; if (!bcmp(fp->fsisig1, "RRaA", 4) && !bcmp(fp->fsisig2, "rrAa", 4) && !bcmp(fp->fsisig3, "\0\0\125\252", 4)) { pmp->pm_nxtfree = getulong(fp->fsinxtfree); if (pmp->pm_nxtfree > pmp->pm_maxcluster) pmp->pm_nxtfree = CLUST_FIRST; } else pmp->pm_fsinfo = 0; brelse(bp); bp = NULL; } /* * Finish initializing pmp->pm_nxtfree (just in case the first few * sectors aren't properly reserved in the FAT). This completes * the fixup for fp->fsinxtfree, and fixes up the zero-initialized * value if there is no fsinfo. We will use pmp->pm_nxtfree * internally even if there is no fsinfo. */ if (pmp->pm_nxtfree < CLUST_FIRST) pmp->pm_nxtfree = CLUST_FIRST; /* * Allocate memory for the bitmap of allocated clusters, and then * fill it in. */ pmp->pm_inusemap = malloc(howmany(pmp->pm_maxcluster + 1, N_INUSEBITS) * sizeof(*pmp->pm_inusemap), M_MSDOSFSFAT, M_WAITOK); /* * fillinusemap() needs pm_devvp. */ pmp->pm_devvp = devvp; pmp->pm_dev = dev; /* * Have the inuse map filled in. */ MSDOSFS_LOCK_MP(pmp); error = fillinusemap(pmp); MSDOSFS_UNLOCK_MP(pmp); if (error != 0) goto error_exit; /* * If they want fat updates to be synchronous then let them suffer * the performance degradation in exchange for the on disk copy of * the fat being correct just about all the time. I suppose this * would be a good thing to turn on if the kernel is still flakey. */ if (mp->mnt_flag & MNT_SYNCHRONOUS) pmp->pm_flags |= MSDOSFSMNT_WAITONFAT; /* * Finish up. */ if (ronly) pmp->pm_flags |= MSDOSFSMNT_RONLY; else { if ((error = markvoldirty(pmp, 1)) != 0) { (void)markvoldirty(pmp, 0); goto error_exit; } pmp->pm_fmod = 1; } mp->mnt_data = pmp; mp->mnt_stat.f_fsid.val[0] = dev2udev(dev); mp->mnt_stat.f_fsid.val[1] = mp->mnt_vfc->vfc_typenum; MNT_ILOCK(mp); mp->mnt_flag |= MNT_LOCAL; MNT_IUNLOCK(mp); if (pmp->pm_flags & MSDOSFS_LARGEFS) msdosfs_fileno_init(mp); return 0; error_exit: if (bp) brelse(bp); if (cp != NULL) { DROP_GIANT(); g_topology_lock(); g_vfs_close(cp); g_topology_unlock(); PICKUP_GIANT(); } if (pmp) { lockdestroy(&pmp->pm_fatlock); if (pmp->pm_inusemap) free(pmp->pm_inusemap, M_MSDOSFSFAT); free(pmp, M_MSDOSFSMNT); mp->mnt_data = NULL; } dev_rel(dev); return (error); } /* * Unmount the filesystem described by mp. */ static int msdosfs_unmount(struct mount *mp, int mntflags) { struct msdosfsmount *pmp; int error, flags; flags = 0; - if (mntflags & MNT_FORCE) + error = msdosfs_sync(mp, MNT_WAIT); + if ((mntflags & MNT_FORCE) != 0) { flags |= FORCECLOSE; + } else if (error != 0) { + return (error); + } error = vflush(mp, 0, flags, curthread); - if (error && error != ENXIO) - return error; + if (error != 0 && error != ENXIO) + return (error); pmp = VFSTOMSDOSFS(mp); if ((pmp->pm_flags & MSDOSFSMNT_RONLY) == 0) { error = markvoldirty(pmp, 0); if (error && error != ENXIO) { (void)markvoldirty(pmp, 1); return (error); } } if (pmp->pm_flags & MSDOSFSMNT_KICONV && msdosfs_iconv) { if (pmp->pm_w2u) msdosfs_iconv->close(pmp->pm_w2u); if (pmp->pm_u2w) msdosfs_iconv->close(pmp->pm_u2w); if (pmp->pm_d2u) msdosfs_iconv->close(pmp->pm_d2u); if (pmp->pm_u2d) msdosfs_iconv->close(pmp->pm_u2d); } #ifdef MSDOSFS_DEBUG { struct vnode *vp = pmp->pm_devvp; struct bufobj *bo; bo = &vp->v_bufobj; BO_LOCK(bo); VI_LOCK(vp); vn_printf(vp, "msdosfs_umount(): just before calling VOP_CLOSE()\n"); printf("freef %p, freeb %p, mount %p\n", TAILQ_NEXT(vp, v_actfreelist), vp->v_actfreelist.tqe_prev, vp->v_mount); printf("cleanblkhd %p, dirtyblkhd %p, numoutput %ld, type %d\n", TAILQ_FIRST(&vp->v_bufobj.bo_clean.bv_hd), TAILQ_FIRST(&vp->v_bufobj.bo_dirty.bv_hd), vp->v_bufobj.bo_numoutput, vp->v_type); VI_UNLOCK(vp); BO_UNLOCK(bo); } #endif DROP_GIANT(); if (pmp->pm_devvp->v_type == VCHR && pmp->pm_devvp->v_rdev != NULL) pmp->pm_devvp->v_rdev->si_mountpt = NULL; g_topology_lock(); g_vfs_close(pmp->pm_cp); g_topology_unlock(); PICKUP_GIANT(); vrele(pmp->pm_devvp); dev_rel(pmp->pm_dev); free(pmp->pm_inusemap, M_MSDOSFSFAT); if (pmp->pm_flags & MSDOSFS_LARGEFS) msdosfs_fileno_free(mp); lockdestroy(&pmp->pm_fatlock); free(pmp, M_MSDOSFSMNT); mp->mnt_data = NULL; MNT_ILOCK(mp); mp->mnt_flag &= ~MNT_LOCAL; MNT_IUNLOCK(mp); return (error); } static int msdosfs_root(struct mount *mp, int flags, struct vnode **vpp) { struct msdosfsmount *pmp = VFSTOMSDOSFS(mp); struct denode *ndep; int error; #ifdef MSDOSFS_DEBUG printf("msdosfs_root(); mp %p, pmp %p\n", mp, pmp); #endif error = deget(pmp, MSDOSFSROOT, MSDOSFSROOT_OFS, &ndep); if (error) return (error); *vpp = DETOV(ndep); return (0); } static int msdosfs_statfs(struct mount *mp, struct statfs *sbp) { struct msdosfsmount *pmp; pmp = VFSTOMSDOSFS(mp); sbp->f_bsize = pmp->pm_bpcluster; sbp->f_iosize = pmp->pm_bpcluster; sbp->f_blocks = pmp->pm_maxcluster + 1; sbp->f_bfree = pmp->pm_freeclustercount; sbp->f_bavail = pmp->pm_freeclustercount; sbp->f_files = pmp->pm_RootDirEnts; /* XXX */ sbp->f_ffree = 0; /* what to put in here? */ return (0); } /* * If we have an FSInfo block, update it. */ static int msdosfs_fsiflush(struct msdosfsmount *pmp, int waitfor) { struct fsinfo *fp; struct buf *bp; int error; MSDOSFS_LOCK_MP(pmp); if (pmp->pm_fsinfo == 0 || (pmp->pm_flags & MSDOSFS_FSIMOD) == 0) { error = 0; goto unlock; } error = bread(pmp->pm_devvp, pmp->pm_fsinfo, pmp->pm_BytesPerSec, NOCRED, &bp); if (error != 0) { brelse(bp); goto unlock; } fp = (struct fsinfo *)bp->b_data; putulong(fp->fsinfree, pmp->pm_freeclustercount); putulong(fp->fsinxtfree, pmp->pm_nxtfree); pmp->pm_flags &= ~MSDOSFS_FSIMOD; if (waitfor == MNT_WAIT) error = bwrite(bp); else bawrite(bp); unlock: MSDOSFS_UNLOCK_MP(pmp); return (error); } static int msdosfs_sync(struct mount *mp, int waitfor) { struct vnode *vp, *nvp; struct thread *td; struct denode *dep; struct msdosfsmount *pmp = VFSTOMSDOSFS(mp); int error, allerror = 0; td = curthread; /* * If we ever switch to not updating all of the fats all the time, * this would be the place to update them from the first one. */ if (pmp->pm_fmod != 0) { if (pmp->pm_flags & MSDOSFSMNT_RONLY) panic("msdosfs_sync: rofs mod"); else { /* update fats here */ } } /* * Write back each (modified) denode. */ loop: MNT_VNODE_FOREACH_ALL(vp, mp, nvp) { if (vp->v_type == VNON) { VI_UNLOCK(vp); continue; } dep = VTODE(vp); if ((dep->de_flag & (DE_ACCESS | DE_CREATE | DE_UPDATE | DE_MODIFIED)) == 0 && (vp->v_bufobj.bo_dirty.bv_cnt == 0 || waitfor == MNT_LAZY)) { VI_UNLOCK(vp); continue; } error = vget(vp, LK_EXCLUSIVE | LK_NOWAIT | LK_INTERLOCK, td); if (error) { if (error == ENOENT) goto loop; continue; } error = VOP_FSYNC(vp, waitfor, td); if (error) allerror = error; VOP_UNLOCK(vp, 0); vrele(vp); } /* * Flush filesystem control info. */ if (waitfor != MNT_LAZY) { vn_lock(pmp->pm_devvp, LK_EXCLUSIVE | LK_RETRY); error = VOP_FSYNC(pmp->pm_devvp, waitfor, td); if (error) allerror = error; VOP_UNLOCK(pmp->pm_devvp, 0); } error = msdosfs_fsiflush(pmp, waitfor); if (error != 0) allerror = error; return (allerror); } static int msdosfs_fhtovp(struct mount *mp, struct fid *fhp, int flags, struct vnode **vpp) { struct msdosfsmount *pmp = VFSTOMSDOSFS(mp); struct defid *defhp = (struct defid *) fhp; struct denode *dep; int error; error = deget(pmp, defhp->defid_dirclust, defhp->defid_dirofs, &dep); if (error) { *vpp = NULLVP; return (error); } *vpp = DETOV(dep); vnode_create_vobject(*vpp, dep->de_FileSize, curthread); return (0); } static struct vfsops msdosfs_vfsops = { .vfs_fhtovp = msdosfs_fhtovp, .vfs_mount = msdosfs_mount, .vfs_cmount = msdosfs_cmount, .vfs_root = msdosfs_root, .vfs_statfs = msdosfs_statfs, .vfs_sync = msdosfs_sync, .vfs_unmount = msdosfs_unmount, }; VFS_SET(msdosfs_vfsops, msdosfs, 0); MODULE_VERSION(msdosfs, 1); Index: projects/building-blocks/sys/kern/kern_exit.c =================================================================== --- projects/building-blocks/sys/kern/kern_exit.c (revision 275663) +++ projects/building-blocks/sys/kern/kern_exit.c (revision 275664) @@ -1,1307 +1,1314 @@ /*- * Copyright (c) 1982, 1986, 1989, 1991, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)kern_exit.c 8.7 (Berkeley) 2/12/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_compat.h" #include "opt_ktrace.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* for acct_process() function prototype */ #include #include #include #include #ifdef KTRACE #include #endif #include #include #include #include #include #include #include #include #include #ifdef KDTRACE_HOOKS #include dtrace_execexit_func_t dtrace_fasttrap_exit; #endif SDT_PROVIDER_DECLARE(proc); SDT_PROBE_DEFINE1(proc, kernel, , exit, "int"); /* Hook for NFS teardown procedure. */ void (*nlminfo_release_p)(struct proc *p); struct proc * proc_realparent(struct proc *child) { struct proc *p, *parent; sx_assert(&proctree_lock, SX_LOCKED); if ((child->p_treeflag & P_TREE_ORPHANED) == 0) { if (child->p_oppid == 0 || child->p_pptr->p_pid == child->p_oppid) parent = child->p_pptr; else parent = initproc; return (parent); } for (p = child; (p->p_treeflag & P_TREE_FIRST_ORPHAN) == 0;) { /* Cannot use LIST_PREV(), since the list head is not known. */ p = __containerof(p->p_orphan.le_prev, struct proc, p_orphan.le_next); KASSERT((p->p_treeflag & P_TREE_ORPHANED) != 0, ("missing P_ORPHAN %p", p)); } parent = __containerof(p->p_orphan.le_prev, struct proc, p_orphans.lh_first); return (parent); } static void clear_orphan(struct proc *p) { struct proc *p1; sx_assert(&proctree_lock, SA_XLOCKED); if ((p->p_treeflag & P_TREE_ORPHANED) == 0) return; if ((p->p_treeflag & P_TREE_FIRST_ORPHAN) != 0) { p1 = LIST_NEXT(p, p_orphan); if (p1 != NULL) p1->p_treeflag |= P_TREE_FIRST_ORPHAN; p->p_treeflag &= ~P_TREE_FIRST_ORPHAN; } LIST_REMOVE(p, p_orphan); p->p_treeflag &= ~P_TREE_ORPHANED; } /* * exit -- death of process. */ void sys_sys_exit(struct thread *td, struct sys_exit_args *uap) { exit1(td, W_EXITCODE(uap->rval, 0)); /* NOTREACHED */ } /* * Exit: deallocate address space and other resources, change proc state to * zombie, and unlink proc from allproc and parent's lists. Save exit status * and rusage for wait(). Check for child processes and orphan them. */ void exit1(struct thread *td, int rv) { struct proc *p, *nq, *q, *t; struct thread *tdt; struct vnode *ttyvp = NULL; mtx_assert(&Giant, MA_NOTOWNED); p = td->td_proc; /* * XXX in case we're rebooting we just let init die in order to * work around an unsolved stack overflow seen very late during * shutdown on sparc64 when the gmirror worker process exists. */ if (p == initproc && rebooting == 0) { printf("init died (signal %d, exit %d)\n", WTERMSIG(rv), WEXITSTATUS(rv)); panic("Going nowhere without my init!"); } /* * MUST abort all other threads before proceeding past here. */ PROC_LOCK(p); + /* + * First check if some other thread or external request got + * here before us. If so, act appropriately: exit or suspend. + * We must ensure that stop requests are handled before we set + * P_WEXIT. + */ + thread_suspend_check(0); while (p->p_flag & P_HADTHREADS) { /* - * First check if some other thread got here before us. - * If so, act appropriately: exit or suspend. - */ - thread_suspend_check(0); - - /* * Kill off the other threads. This requires * some co-operation from other parts of the kernel * so it may not be instantaneous. With this state set * any thread entering the kernel from userspace will * thread_exit() in trap(). Any thread attempting to * sleep will return immediately with EINTR or EWOULDBLOCK * which will hopefully force them to back out to userland * freeing resources as they go. Any thread attempting * to return to userland will thread_exit() from userret(). * thread_exit() will unsuspend us when the last of the * other threads exits. * If there is already a thread singler after resumption, * calling thread_single will fail; in that case, we just * re-check all suspension request, the thread should * either be suspended there or exit. */ if (!thread_single(SINGLE_EXIT)) + /* + * All other activity in this process is now + * stopped. Threading support has been turned + * off. + */ break; - /* - * All other activity in this process is now stopped. - * Threading support has been turned off. + * Recheck for new stop or suspend requests which + * might appear while process lock was dropped in + * thread_single(). */ + thread_suspend_check(0); } KASSERT(p->p_numthreads == 1, ("exit1: proc %p exiting with %d threads", p, p->p_numthreads)); racct_sub(p, RACCT_NTHR, 1); /* * Wakeup anyone in procfs' PIOCWAIT. They should have a hold * on our vmspace, so we should block below until they have * released their reference to us. Note that if they have * requested S_EXIT stops we will block here until they ack * via PIOCCONT. */ _STOPEVENT(p, S_EXIT, rv); /* * Ignore any pending request to stop due to a stop signal. * Once P_WEXIT is set, future requests will be ignored as * well. */ p->p_flag &= ~P_STOPPED_SIG; KASSERT(!P_SHOULDSTOP(p), ("exiting process is stopped")); /* * Note that we are exiting and do another wakeup of anyone in * PIOCWAIT in case they aren't listening for S_EXIT stops or * decided to wait again after we told them we are exiting. */ p->p_flag |= P_WEXIT; wakeup(&p->p_stype); /* * Wait for any processes that have a hold on our vmspace to * release their reference. */ while (p->p_lock > 0) msleep(&p->p_lock, &p->p_mtx, PWAIT, "exithold", 0); p->p_xstat = rv; /* Let event handler change exit status */ PROC_UNLOCK(p); /* Drain the limit callout while we don't have the proc locked */ callout_drain(&p->p_limco); #ifdef AUDIT /* * The Sun BSM exit token contains two components: an exit status as * passed to exit(), and a return value to indicate what sort of exit * it was. The exit status is WEXITSTATUS(rv), but it's not clear * what the return value is. */ AUDIT_ARG_EXIT(WEXITSTATUS(rv), 0); AUDIT_SYSCALL_EXIT(0, td); #endif /* Are we a task leader with peers? */ if (p->p_peers != NULL && p == p->p_leader) { mtx_lock(&ppeers_lock); q = p->p_peers; while (q != NULL) { PROC_LOCK(q); kern_psignal(q, SIGKILL); PROC_UNLOCK(q); q = q->p_peers; } while (p->p_peers != NULL) msleep(p, &ppeers_lock, PWAIT, "exit1", 0); mtx_unlock(&ppeers_lock); } /* * Check if any loadable modules need anything done at process exit. * E.g. SYSV IPC stuff * XXX what if one of these generates an error? */ EVENTHANDLER_INVOKE(process_exit, p); /* * If parent is waiting for us to exit or exec, * P_PPWAIT is set; we will wakeup the parent below. */ PROC_LOCK(p); rv = p->p_xstat; /* Event handler could change exit status */ stopprofclock(p); p->p_flag &= ~(P_TRACED | P_PPWAIT | P_PPTRACE); /* * Stop the real interval timer. If the handler is currently * executing, prevent it from rearming itself and let it finish. */ if (timevalisset(&p->p_realtimer.it_value) && callout_stop(&p->p_itcallout) == 0) { timevalclear(&p->p_realtimer.it_interval); msleep(&p->p_itcallout, &p->p_mtx, PWAIT, "ritwait", 0); KASSERT(!timevalisset(&p->p_realtimer.it_value), ("realtime timer is still armed")); } PROC_UNLOCK(p); /* * Reset any sigio structures pointing to us as a result of * F_SETOWN with our pid. */ funsetownlst(&p->p_sigiolst); /* * If this process has an nlminfo data area (for lockd), release it */ if (nlminfo_release_p != NULL && p->p_nlminfo != NULL) (*nlminfo_release_p)(p); /* * Close open files and release open-file table. * This may block! */ fdescfree(td); /* * If this thread tickled GEOM, we need to wait for the giggling to * stop before we return to userland */ if (td->td_pflags & TDP_GEOM) g_waitidle(); /* * Remove ourself from our leader's peer list and wake our leader. */ if (p->p_leader->p_peers != NULL) { mtx_lock(&ppeers_lock); if (p->p_leader->p_peers != NULL) { q = p->p_leader; while (q->p_peers != p) q = q->p_peers; q->p_peers = p->p_peers; wakeup(p->p_leader); } mtx_unlock(&ppeers_lock); } vmspace_exit(td); sx_xlock(&proctree_lock); if (SESS_LEADER(p)) { struct session *sp = p->p_session; struct tty *tp; /* * s_ttyp is not zero'd; we use this to indicate that * the session once had a controlling terminal. (for * logging and informational purposes) */ SESS_LOCK(sp); ttyvp = sp->s_ttyvp; tp = sp->s_ttyp; sp->s_ttyvp = NULL; sp->s_ttydp = NULL; sp->s_leader = NULL; SESS_UNLOCK(sp); /* * Signal foreground pgrp and revoke access to * controlling terminal if it has not been revoked * already. * * Because the TTY may have been revoked in the mean * time and could already have a new session associated * with it, make sure we don't send a SIGHUP to a * foreground process group that does not belong to this * session. */ if (tp != NULL) { tty_lock(tp); if (tp->t_session == sp) tty_signal_pgrp(tp, SIGHUP); tty_unlock(tp); } if (ttyvp != NULL) { sx_xunlock(&proctree_lock); if (vn_lock(ttyvp, LK_EXCLUSIVE) == 0) { VOP_REVOKE(ttyvp, REVOKEALL); VOP_UNLOCK(ttyvp, 0); } sx_xlock(&proctree_lock); } } fixjobc(p, p->p_pgrp, 0); sx_xunlock(&proctree_lock); (void)acct_process(td); /* Release the TTY now we've unlocked everything. */ if (ttyvp != NULL) vrele(ttyvp); #ifdef KTRACE ktrprocexit(td); #endif /* * Release reference to text vnode */ if (p->p_textvp != NULL) { vrele(p->p_textvp); p->p_textvp = NULL; } /* * Release our limits structure. */ lim_free(p->p_limit); p->p_limit = NULL; tidhash_remove(td); /* * Remove proc from allproc queue and pidhash chain. * Place onto zombproc. Unlink from parent's child list. */ sx_xlock(&allproc_lock); LIST_REMOVE(p, p_list); LIST_INSERT_HEAD(&zombproc, p, p_list); LIST_REMOVE(p, p_hash); sx_xunlock(&allproc_lock); /* * Call machine-dependent code to release any * machine-dependent resources other than the address space. * The address space is released by "vmspace_exitfree(p)" in * vm_waitproc(). */ cpu_exit(td); WITNESS_WARN(WARN_PANIC, NULL, "process (pid %d) exiting", p->p_pid); /* * Reparent all children processes: * - traced ones to the original parent (or init if we are that parent) * - the rest to init */ sx_xlock(&proctree_lock); q = LIST_FIRST(&p->p_children); if (q != NULL) /* only need this if any child is S_ZOMB */ wakeup(initproc); for (; q != NULL; q = nq) { nq = LIST_NEXT(q, p_sibling); PROC_LOCK(q); q->p_sigparent = SIGCHLD; if (!(q->p_flag & P_TRACED)) { proc_reparent(q, initproc); } else { /* * Traced processes are killed since their existence * means someone is screwing up. */ t = proc_realparent(q); if (t == p) { proc_reparent(q, initproc); } else { PROC_LOCK(t); proc_reparent(q, t); PROC_UNLOCK(t); } /* * Since q was found on our children list, the * proc_reparent() call moved q to the orphan * list due to present P_TRACED flag. Clear * orphan link for q now while q is locked. */ clear_orphan(q); q->p_flag &= ~(P_TRACED | P_STOPPED_TRACE); FOREACH_THREAD_IN_PROC(q, tdt) tdt->td_dbgflags &= ~TDB_SUSPEND; kern_psignal(q, SIGKILL); } PROC_UNLOCK(q); } /* * Also get rid of our orphans. */ while ((q = LIST_FIRST(&p->p_orphans)) != NULL) { PROC_LOCK(q); clear_orphan(q); PROC_UNLOCK(q); } /* Save exit status. */ PROC_LOCK(p); p->p_xthread = td; /* Tell the prison that we are gone. */ prison_proc_free(p->p_ucred->cr_prison); #ifdef KDTRACE_HOOKS /* * Tell the DTrace fasttrap provider about the exit if it * has declared an interest. */ if (dtrace_fasttrap_exit) dtrace_fasttrap_exit(p); #endif /* * Notify interested parties of our demise. */ KNOTE_LOCKED(&p->p_klist, NOTE_EXIT); #ifdef KDTRACE_HOOKS int reason = CLD_EXITED; if (WCOREDUMP(rv)) reason = CLD_DUMPED; else if (WIFSIGNALED(rv)) reason = CLD_KILLED; SDT_PROBE(proc, kernel, , exit, reason, 0, 0, 0, 0); #endif /* * Just delete all entries in the p_klist. At this point we won't * report any more events, and there are nasty race conditions that * can beat us if we don't. */ knlist_clear(&p->p_klist, 1); /* * If this is a process with a descriptor, we may not need to deliver * a signal to the parent. proctree_lock is held over * procdesc_exit() to serialize concurrent calls to close() and * exit(). */ if (p->p_procdesc == NULL || procdesc_exit(p)) { /* * Notify parent that we're gone. If parent has the * PS_NOCLDWAIT flag set, or if the handler is set to SIG_IGN, * notify process 1 instead (and hope it will handle this * situation). */ PROC_LOCK(p->p_pptr); mtx_lock(&p->p_pptr->p_sigacts->ps_mtx); if (p->p_pptr->p_sigacts->ps_flag & (PS_NOCLDWAIT | PS_CLDSIGIGN)) { struct proc *pp; mtx_unlock(&p->p_pptr->p_sigacts->ps_mtx); pp = p->p_pptr; PROC_UNLOCK(pp); proc_reparent(p, initproc); p->p_sigparent = SIGCHLD; PROC_LOCK(p->p_pptr); /* * Notify parent, so in case he was wait(2)ing or * executing waitpid(2) with our pid, he will * continue. */ wakeup(pp); } else mtx_unlock(&p->p_pptr->p_sigacts->ps_mtx); if (p->p_pptr == initproc) kern_psignal(p->p_pptr, SIGCHLD); else if (p->p_sigparent != 0) { if (p->p_sigparent == SIGCHLD) childproc_exited(p); else /* LINUX thread */ kern_psignal(p->p_pptr, p->p_sigparent); } } else PROC_LOCK(p->p_pptr); sx_xunlock(&proctree_lock); /* * The state PRS_ZOMBIE prevents other proesses from sending * signal to the process, to avoid memory leak, we free memory * for signal queue at the time when the state is set. */ sigqueue_flush(&p->p_sigqueue); sigqueue_flush(&td->td_sigqueue); /* * We have to wait until after acquiring all locks before * changing p_state. We need to avoid all possible context * switches (including ones from blocking on a mutex) while * marked as a zombie. We also have to set the zombie state * before we release the parent process' proc lock to avoid * a lost wakeup. So, we first call wakeup, then we grab the * sched lock, update the state, and release the parent process' * proc lock. */ wakeup(p->p_pptr); cv_broadcast(&p->p_pwait); sched_exit(p->p_pptr, td); PROC_SLOCK(p); p->p_state = PRS_ZOMBIE; PROC_UNLOCK(p->p_pptr); /* * Hopefully no one will try to deliver a signal to the process this * late in the game. */ knlist_destroy(&p->p_klist); /* * Save our children's rusage information in our exit rusage. */ PROC_STATLOCK(p); ruadd(&p->p_ru, &p->p_rux, &p->p_stats->p_cru, &p->p_crux); PROC_STATUNLOCK(p); /* * Make sure the scheduler takes this thread out of its tables etc. * This will also release this thread's reference to the ucred. * Other thread parts to release include pcb bits and such. */ thread_exit(); } #ifndef _SYS_SYSPROTO_H_ struct abort2_args { char *why; int nargs; void **args; }; #endif int sys_abort2(struct thread *td, struct abort2_args *uap) { struct proc *p = td->td_proc; struct sbuf *sb; void *uargs[16]; int error, i, sig; /* * Do it right now so we can log either proper call of abort2(), or * note, that invalid argument was passed. 512 is big enough to * handle 16 arguments' descriptions with additional comments. */ sb = sbuf_new(NULL, NULL, 512, SBUF_FIXEDLEN); sbuf_clear(sb); sbuf_printf(sb, "%s(pid %d uid %d) aborted: ", p->p_comm, p->p_pid, td->td_ucred->cr_uid); /* * Since we can't return from abort2(), send SIGKILL in cases, where * abort2() was called improperly */ sig = SIGKILL; /* Prevent from DoSes from user-space. */ if (uap->nargs < 0 || uap->nargs > 16) goto out; if (uap->nargs > 0) { if (uap->args == NULL) goto out; error = copyin(uap->args, uargs, uap->nargs * sizeof(void *)); if (error != 0) goto out; } /* * Limit size of 'reason' string to 128. Will fit even when * maximal number of arguments was chosen to be logged. */ if (uap->why != NULL) { error = sbuf_copyin(sb, uap->why, 128); if (error < 0) goto out; } else { sbuf_printf(sb, "(null)"); } if (uap->nargs > 0) { sbuf_printf(sb, "("); for (i = 0;i < uap->nargs; i++) sbuf_printf(sb, "%s%p", i == 0 ? "" : ", ", uargs[i]); sbuf_printf(sb, ")"); } /* * Final stage: arguments were proper, string has been * successfully copied from userspace, and copying pointers * from user-space succeed. */ sig = SIGABRT; out: if (sig == SIGKILL) { sbuf_trim(sb); sbuf_printf(sb, " (Reason text inaccessible)"); } sbuf_cat(sb, "\n"); sbuf_finish(sb); log(LOG_INFO, "%s", sbuf_data(sb)); sbuf_delete(sb); exit1(td, W_EXITCODE(0, sig)); return (0); } #ifdef COMPAT_43 /* * The dirty work is handled by kern_wait(). */ int owait(struct thread *td, struct owait_args *uap __unused) { int error, status; error = kern_wait(td, WAIT_ANY, &status, 0, NULL); if (error == 0) td->td_retval[1] = status; return (error); } #endif /* COMPAT_43 */ /* * The dirty work is handled by kern_wait(). */ int sys_wait4(struct thread *td, struct wait4_args *uap) { struct rusage ru, *rup; int error, status; if (uap->rusage != NULL) rup = &ru; else rup = NULL; error = kern_wait(td, uap->pid, &status, uap->options, rup); if (uap->status != NULL && error == 0) error = copyout(&status, uap->status, sizeof(status)); if (uap->rusage != NULL && error == 0) error = copyout(&ru, uap->rusage, sizeof(struct rusage)); return (error); } int sys_wait6(struct thread *td, struct wait6_args *uap) { struct __wrusage wru, *wrup; siginfo_t si, *sip; idtype_t idtype; id_t id; int error, status; idtype = uap->idtype; id = uap->id; if (uap->wrusage != NULL) wrup = &wru; else wrup = NULL; if (uap->info != NULL) { sip = &si; bzero(sip, sizeof(*sip)); } else sip = NULL; /* * We expect all callers of wait6() to know about WEXITED and * WTRAPPED. */ error = kern_wait6(td, idtype, id, &status, uap->options, wrup, sip); if (uap->status != NULL && error == 0) error = copyout(&status, uap->status, sizeof(status)); if (uap->wrusage != NULL && error == 0) error = copyout(&wru, uap->wrusage, sizeof(wru)); if (uap->info != NULL && error == 0) error = copyout(&si, uap->info, sizeof(si)); return (error); } /* * Reap the remains of a zombie process and optionally return status and * rusage. Asserts and will release both the proctree_lock and the process * lock as part of its work. */ void proc_reap(struct thread *td, struct proc *p, int *status, int options) { struct proc *q, *t; sx_assert(&proctree_lock, SA_XLOCKED); PROC_LOCK_ASSERT(p, MA_OWNED); PROC_SLOCK_ASSERT(p, MA_OWNED); KASSERT(p->p_state == PRS_ZOMBIE, ("proc_reap: !PRS_ZOMBIE")); q = td->td_proc; PROC_SUNLOCK(p); td->td_retval[0] = p->p_pid; if (status) *status = p->p_xstat; /* convert to int */ if (options & WNOWAIT) { /* * Only poll, returning the status. Caller does not wish to * release the proc struct just yet. */ PROC_UNLOCK(p); sx_xunlock(&proctree_lock); return; } PROC_LOCK(q); sigqueue_take(p->p_ksi); PROC_UNLOCK(q); PROC_UNLOCK(p); /* * If we got the child via a ptrace 'attach', we need to give it back * to the old parent. */ if (p->p_oppid != 0) { t = proc_realparent(p); PROC_LOCK(t); PROC_LOCK(p); proc_reparent(p, t); p->p_oppid = 0; PROC_UNLOCK(p); pksignal(t, SIGCHLD, p->p_ksi); wakeup(t); cv_broadcast(&p->p_pwait); PROC_UNLOCK(t); sx_xunlock(&proctree_lock); return; } /* * Remove other references to this process to ensure we have an * exclusive reference. */ sx_xlock(&allproc_lock); LIST_REMOVE(p, p_list); /* off zombproc */ sx_xunlock(&allproc_lock); LIST_REMOVE(p, p_sibling); PROC_LOCK(p); clear_orphan(p); PROC_UNLOCK(p); leavepgrp(p); if (p->p_procdesc != NULL) procdesc_reap(p); sx_xunlock(&proctree_lock); /* * As a side effect of this lock, we know that all other writes to * this proc are visible now, so no more locking is needed for p. */ PROC_LOCK(p); p->p_xstat = 0; /* XXX: why? */ PROC_UNLOCK(p); PROC_LOCK(q); ruadd(&q->p_stats->p_cru, &q->p_crux, &p->p_ru, &p->p_rux); PROC_UNLOCK(q); /* * Decrement the count of procs running with this uid. */ (void)chgproccnt(p->p_ucred->cr_ruidinfo, -1, 0); /* * Destroy resource accounting information associated with the process. */ #ifdef RACCT PROC_LOCK(p); racct_sub(p, RACCT_NPROC, 1); PROC_UNLOCK(p); #endif racct_proc_exit(p); /* * Free credentials, arguments, and sigacts. */ crfree(p->p_ucred); p->p_ucred = NULL; pargs_drop(p->p_args); p->p_args = NULL; sigacts_free(p->p_sigacts); p->p_sigacts = NULL; /* * Do any thread-system specific cleanups. */ thread_wait(p); /* * Give vm and machine-dependent layer a chance to free anything that * cpu_exit couldn't release while still running in process context. */ vm_waitproc(p); #ifdef MAC mac_proc_destroy(p); #endif KASSERT(FIRST_THREAD_IN_PROC(p), ("proc_reap: no residual thread!")); uma_zfree(proc_zone, p); sx_xlock(&allproc_lock); nprocs--; sx_xunlock(&allproc_lock); } static int proc_to_reap(struct thread *td, struct proc *p, idtype_t idtype, id_t id, int *status, int options, struct __wrusage *wrusage, siginfo_t *siginfo) { struct proc *q; struct rusage *rup; sx_assert(&proctree_lock, SA_XLOCKED); q = td->td_proc; PROC_LOCK(p); switch (idtype) { case P_ALL: break; case P_PID: if (p->p_pid != (pid_t)id) { PROC_UNLOCK(p); return (0); } break; case P_PGID: if (p->p_pgid != (pid_t)id) { PROC_UNLOCK(p); return (0); } break; case P_SID: if (p->p_session->s_sid != (pid_t)id) { PROC_UNLOCK(p); return (0); } break; case P_UID: if (p->p_ucred->cr_uid != (uid_t)id) { PROC_UNLOCK(p); return (0); } break; case P_GID: if (p->p_ucred->cr_gid != (gid_t)id) { PROC_UNLOCK(p); return (0); } break; case P_JAILID: if (p->p_ucred->cr_prison->pr_id != (int)id) { PROC_UNLOCK(p); return (0); } break; /* * It seems that the thread structures get zeroed out * at process exit. This makes it impossible to * support P_SETID, P_CID or P_CPUID. */ default: PROC_UNLOCK(p); return (0); } if (p_canwait(td, p)) { PROC_UNLOCK(p); return (0); } if (((options & WEXITED) == 0) && (p->p_state == PRS_ZOMBIE)) { PROC_UNLOCK(p); return (0); } /* * This special case handles a kthread spawned by linux_clone * (see linux_misc.c). The linux_wait4 and linux_waitpid * functions need to be able to distinguish between waiting * on a process and waiting on a thread. It is a thread if * p_sigparent is not SIGCHLD, and the WLINUXCLONE option * signifies we want to wait for threads and not processes. */ if ((p->p_sigparent != SIGCHLD) ^ ((options & WLINUXCLONE) != 0)) { PROC_UNLOCK(p); return (0); } if (siginfo != NULL) { bzero(siginfo, sizeof(*siginfo)); siginfo->si_errno = 0; /* * SUSv4 requires that the si_signo value is always * SIGCHLD. Obey it despite the rfork(2) interface * allows to request other signal for child exit * notification. */ siginfo->si_signo = SIGCHLD; /* * This is still a rough estimate. We will fix the * cases TRAPPED, STOPPED, and CONTINUED later. */ if (WCOREDUMP(p->p_xstat)) { siginfo->si_code = CLD_DUMPED; siginfo->si_status = WTERMSIG(p->p_xstat); } else if (WIFSIGNALED(p->p_xstat)) { siginfo->si_code = CLD_KILLED; siginfo->si_status = WTERMSIG(p->p_xstat); } else { siginfo->si_code = CLD_EXITED; siginfo->si_status = WEXITSTATUS(p->p_xstat); } siginfo->si_pid = p->p_pid; siginfo->si_uid = p->p_ucred->cr_uid; /* * The si_addr field would be useful additional * detail, but apparently the PC value may be lost * when we reach this point. bzero() above sets * siginfo->si_addr to NULL. */ } /* * There should be no reason to limit resources usage info to * exited processes only. A snapshot about any resources used * by a stopped process may be exactly what is needed. */ if (wrusage != NULL) { rup = &wrusage->wru_self; *rup = p->p_ru; PROC_STATLOCK(p); calcru(p, &rup->ru_utime, &rup->ru_stime); PROC_STATUNLOCK(p); rup = &wrusage->wru_children; *rup = p->p_stats->p_cru; calccru(p, &rup->ru_utime, &rup->ru_stime); } if (p->p_state == PRS_ZOMBIE) { PROC_SLOCK(p); proc_reap(td, p, status, options); return (-1); } PROC_UNLOCK(p); return (1); } int kern_wait(struct thread *td, pid_t pid, int *status, int options, struct rusage *rusage) { struct __wrusage wru, *wrup; idtype_t idtype; id_t id; int ret; /* * Translate the special pid values into the (idtype, pid) * pair for kern_wait6. The WAIT_MYPGRP case is handled by * kern_wait6() on its own. */ if (pid == WAIT_ANY) { idtype = P_ALL; id = 0; } else if (pid < 0) { idtype = P_PGID; id = (id_t)-pid; } else { idtype = P_PID; id = (id_t)pid; } if (rusage != NULL) wrup = &wru; else wrup = NULL; /* * For backward compatibility we implicitly add flags WEXITED * and WTRAPPED here. */ options |= WEXITED | WTRAPPED; ret = kern_wait6(td, idtype, id, status, options, wrup, NULL); if (rusage != NULL) *rusage = wru.wru_self; return (ret); } int kern_wait6(struct thread *td, idtype_t idtype, id_t id, int *status, int options, struct __wrusage *wrusage, siginfo_t *siginfo) { struct proc *p, *q; int error, nfound, ret; AUDIT_ARG_VALUE((int)idtype); /* XXX - This is likely wrong! */ AUDIT_ARG_PID((pid_t)id); /* XXX - This may be wrong! */ AUDIT_ARG_VALUE(options); q = td->td_proc; if ((pid_t)id == WAIT_MYPGRP && (idtype == P_PID || idtype == P_PGID)) { PROC_LOCK(q); id = (id_t)q->p_pgid; PROC_UNLOCK(q); idtype = P_PGID; } /* If we don't know the option, just return. */ if ((options & ~(WUNTRACED | WNOHANG | WCONTINUED | WNOWAIT | WEXITED | WTRAPPED | WLINUXCLONE)) != 0) return (EINVAL); if ((options & (WEXITED | WUNTRACED | WCONTINUED | WTRAPPED)) == 0) { /* * We will be unable to find any matching processes, * because there are no known events to look for. * Prefer to return error instead of blocking * indefinitely. */ return (EINVAL); } loop: if (q->p_flag & P_STATCHILD) { PROC_LOCK(q); q->p_flag &= ~P_STATCHILD; PROC_UNLOCK(q); } nfound = 0; sx_xlock(&proctree_lock); LIST_FOREACH(p, &q->p_children, p_sibling) { ret = proc_to_reap(td, p, idtype, id, status, options, wrusage, siginfo); if (ret == 0) continue; else if (ret == 1) nfound++; else return (0); PROC_LOCK(p); PROC_SLOCK(p); if ((options & WTRAPPED) != 0 && (p->p_flag & P_TRACED) != 0 && (p->p_flag & (P_STOPPED_TRACE | P_STOPPED_SIG)) != 0 && (p->p_suspcount == p->p_numthreads) && ((p->p_flag & P_WAITED) == 0)) { PROC_SUNLOCK(p); if ((options & WNOWAIT) == 0) p->p_flag |= P_WAITED; sx_xunlock(&proctree_lock); td->td_retval[0] = p->p_pid; if (status != NULL) *status = W_STOPCODE(p->p_xstat); if (siginfo != NULL) { siginfo->si_status = p->p_xstat; siginfo->si_code = CLD_TRAPPED; } if ((options & WNOWAIT) == 0) { PROC_LOCK(q); sigqueue_take(p->p_ksi); PROC_UNLOCK(q); } PROC_UNLOCK(p); return (0); } if ((options & WUNTRACED) != 0 && (p->p_flag & P_STOPPED_SIG) != 0 && (p->p_suspcount == p->p_numthreads) && ((p->p_flag & P_WAITED) == 0)) { PROC_SUNLOCK(p); if ((options & WNOWAIT) == 0) p->p_flag |= P_WAITED; sx_xunlock(&proctree_lock); td->td_retval[0] = p->p_pid; if (status != NULL) *status = W_STOPCODE(p->p_xstat); if (siginfo != NULL) { siginfo->si_status = p->p_xstat; siginfo->si_code = CLD_STOPPED; } if ((options & WNOWAIT) == 0) { PROC_LOCK(q); sigqueue_take(p->p_ksi); PROC_UNLOCK(q); } PROC_UNLOCK(p); return (0); } PROC_SUNLOCK(p); if ((options & WCONTINUED) != 0 && (p->p_flag & P_CONTINUED) != 0) { sx_xunlock(&proctree_lock); td->td_retval[0] = p->p_pid; if ((options & WNOWAIT) == 0) { p->p_flag &= ~P_CONTINUED; PROC_LOCK(q); sigqueue_take(p->p_ksi); PROC_UNLOCK(q); } PROC_UNLOCK(p); if (status != NULL) *status = SIGCONT; if (siginfo != NULL) { siginfo->si_status = SIGCONT; siginfo->si_code = CLD_CONTINUED; } return (0); } PROC_UNLOCK(p); } /* * Look in the orphans list too, to allow the parent to * collect it's child exit status even if child is being * debugged. * * Debugger detaches from the parent upon successful * switch-over from parent to child. At this point due to * re-parenting the parent loses the child to debugger and a * wait4(2) call would report that it has no children to wait * for. By maintaining a list of orphans we allow the parent * to successfully wait until the child becomes a zombie. */ LIST_FOREACH(p, &q->p_orphans, p_orphan) { ret = proc_to_reap(td, p, idtype, id, status, options, wrusage, siginfo); if (ret == 0) continue; else if (ret == 1) nfound++; else return (0); } if (nfound == 0) { sx_xunlock(&proctree_lock); return (ECHILD); } if (options & WNOHANG) { sx_xunlock(&proctree_lock); td->td_retval[0] = 0; return (0); } PROC_LOCK(q); sx_xunlock(&proctree_lock); if (q->p_flag & P_STATCHILD) { q->p_flag &= ~P_STATCHILD; error = 0; } else error = msleep(q, &q->p_mtx, PWAIT | PCATCH, "wait", 0); PROC_UNLOCK(q); if (error) return (error); goto loop; } /* * Make process 'parent' the new parent of process 'child'. * Must be called with an exclusive hold of proctree lock. */ void proc_reparent(struct proc *child, struct proc *parent) { sx_assert(&proctree_lock, SX_XLOCKED); PROC_LOCK_ASSERT(child, MA_OWNED); if (child->p_pptr == parent) return; PROC_LOCK(child->p_pptr); sigqueue_take(child->p_ksi); PROC_UNLOCK(child->p_pptr); LIST_REMOVE(child, p_sibling); LIST_INSERT_HEAD(&parent->p_children, child, p_sibling); clear_orphan(child); if (child->p_flag & P_TRACED) { if (LIST_EMPTY(&child->p_pptr->p_orphans)) { child->p_treeflag |= P_TREE_FIRST_ORPHAN; LIST_INSERT_HEAD(&child->p_pptr->p_orphans, child, p_orphan); } else { LIST_INSERT_AFTER(LIST_FIRST(&child->p_pptr->p_orphans), child, p_orphan); } child->p_treeflag |= P_TREE_ORPHANED; } child->p_pptr = parent; } Index: projects/building-blocks/sys/kern/kern_thread.c =================================================================== --- projects/building-blocks/sys/kern/kern_thread.c (revision 275663) +++ projects/building-blocks/sys/kern/kern_thread.c (revision 275664) @@ -1,1051 +1,1069 @@ /*- * Copyright (C) 2001 Julian Elischer . * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice(s), this list of conditions and the following disclaimer as * the first lines of this file unmodified other than the possible * addition of one or more copyright notices. * 2. Redistributions in binary form must reproduce the above copyright * notice(s), this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH * DAMAGE. */ #include "opt_witness.h" #include "opt_hwpmc_hooks.h" #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef HWPMC_HOOKS #include #endif #include #include #include #include #include SDT_PROVIDER_DECLARE(proc); SDT_PROBE_DEFINE(proc, , , lwp__exit); - /* * thread related storage. */ static uma_zone_t thread_zone; TAILQ_HEAD(, thread) zombie_threads = TAILQ_HEAD_INITIALIZER(zombie_threads); static struct mtx zombie_lock; MTX_SYSINIT(zombie_lock, &zombie_lock, "zombie lock", MTX_SPIN); static void thread_zombie(struct thread *); #define TID_BUFFER_SIZE 1024 struct mtx tid_lock; static struct unrhdr *tid_unrhdr; static lwpid_t tid_buffer[TID_BUFFER_SIZE]; static int tid_head, tid_tail; static MALLOC_DEFINE(M_TIDHASH, "tidhash", "thread hash"); struct tidhashhead *tidhashtbl; u_long tidhash; struct rwlock tidhash_lock; static lwpid_t tid_alloc(void) { lwpid_t tid; tid = alloc_unr(tid_unrhdr); if (tid != -1) return (tid); mtx_lock(&tid_lock); if (tid_head == tid_tail) { mtx_unlock(&tid_lock); return (-1); } tid = tid_buffer[tid_head]; tid_head = (tid_head + 1) % TID_BUFFER_SIZE; mtx_unlock(&tid_lock); return (tid); } static void tid_free(lwpid_t tid) { lwpid_t tmp_tid = -1; mtx_lock(&tid_lock); if ((tid_tail + 1) % TID_BUFFER_SIZE == tid_head) { tmp_tid = tid_buffer[tid_head]; tid_head = (tid_head + 1) % TID_BUFFER_SIZE; } tid_buffer[tid_tail] = tid; tid_tail = (tid_tail + 1) % TID_BUFFER_SIZE; mtx_unlock(&tid_lock); if (tmp_tid != -1) free_unr(tid_unrhdr, tmp_tid); } /* * Prepare a thread for use. */ static int thread_ctor(void *mem, int size, void *arg, int flags) { struct thread *td; td = (struct thread *)mem; td->td_state = TDS_INACTIVE; td->td_oncpu = NOCPU; td->td_tid = tid_alloc(); /* * Note that td_critnest begins life as 1 because the thread is not * running and is thereby implicitly waiting to be on the receiving * end of a context switch. */ td->td_critnest = 1; td->td_lend_user_pri = PRI_MAX; EVENTHANDLER_INVOKE(thread_ctor, td); #ifdef AUDIT audit_thread_alloc(td); #endif umtx_thread_alloc(td); return (0); } /* * Reclaim a thread after use. */ static void thread_dtor(void *mem, int size, void *arg) { struct thread *td; td = (struct thread *)mem; #ifdef INVARIANTS /* Verify that this thread is in a safe state to free. */ switch (td->td_state) { case TDS_INHIBITED: case TDS_RUNNING: case TDS_CAN_RUN: case TDS_RUNQ: /* * We must never unlink a thread that is in one of * these states, because it is currently active. */ panic("bad state for thread unlinking"); /* NOTREACHED */ case TDS_INACTIVE: break; default: panic("bad thread state"); /* NOTREACHED */ } #endif #ifdef AUDIT audit_thread_free(td); #endif /* Free all OSD associated to this thread. */ osd_thread_exit(td); EVENTHANDLER_INVOKE(thread_dtor, td); tid_free(td->td_tid); } /* * Initialize type-stable parts of a thread (when newly created). */ static int thread_init(void *mem, int size, int flags) { struct thread *td; td = (struct thread *)mem; td->td_sleepqueue = sleepq_alloc(); td->td_turnstile = turnstile_alloc(); td->td_rlqe = NULL; EVENTHANDLER_INVOKE(thread_init, td); td->td_sched = (struct td_sched *)&td[1]; umtx_thread_init(td); td->td_kstack = 0; return (0); } /* * Tear down type-stable parts of a thread (just before being discarded). */ static void thread_fini(void *mem, int size) { struct thread *td; td = (struct thread *)mem; EVENTHANDLER_INVOKE(thread_fini, td); rlqentry_free(td->td_rlqe); turnstile_free(td->td_turnstile); sleepq_free(td->td_sleepqueue); umtx_thread_fini(td); seltdfini(td); } /* * For a newly created process, * link up all the structures and its initial threads etc. * called from: * {arch}/{arch}/machdep.c {arch}_init(), init386() etc. * proc_dtor() (should go away) * proc_init() */ void proc_linkup0(struct proc *p, struct thread *td) { TAILQ_INIT(&p->p_threads); /* all threads in proc */ proc_linkup(p, td); } void proc_linkup(struct proc *p, struct thread *td) { sigqueue_init(&p->p_sigqueue, p); p->p_ksi = ksiginfo_alloc(1); if (p->p_ksi != NULL) { /* XXX p_ksi may be null if ksiginfo zone is not ready */ p->p_ksi->ksi_flags = KSI_EXT | KSI_INS; } LIST_INIT(&p->p_mqnotifier); p->p_numthreads = 0; thread_link(td, p); } /* * Initialize global thread allocation resources. */ void threadinit(void) { mtx_init(&tid_lock, "TID lock", NULL, MTX_DEF); /* * pid_max cannot be greater than PID_MAX. * leave one number for thread0. */ tid_unrhdr = new_unrhdr(PID_MAX + 2, INT_MAX, &tid_lock); thread_zone = uma_zcreate("THREAD", sched_sizeof_thread(), thread_ctor, thread_dtor, thread_init, thread_fini, 16 - 1, 0); tidhashtbl = hashinit(maxproc / 2, M_TIDHASH, &tidhash); rw_init(&tidhash_lock, "tidhash"); } /* * Place an unused thread on the zombie list. * Use the slpq as that must be unused by now. */ void thread_zombie(struct thread *td) { mtx_lock_spin(&zombie_lock); TAILQ_INSERT_HEAD(&zombie_threads, td, td_slpq); mtx_unlock_spin(&zombie_lock); } /* * Release a thread that has exited after cpu_throw(). */ void thread_stash(struct thread *td) { atomic_subtract_rel_int(&td->td_proc->p_exitthreads, 1); thread_zombie(td); } /* * Reap zombie resources. */ void thread_reap(void) { struct thread *td_first, *td_next; /* * Don't even bother to lock if none at this instant, * we really don't care about the next instant.. */ if (!TAILQ_EMPTY(&zombie_threads)) { mtx_lock_spin(&zombie_lock); td_first = TAILQ_FIRST(&zombie_threads); if (td_first) TAILQ_INIT(&zombie_threads); mtx_unlock_spin(&zombie_lock); while (td_first) { td_next = TAILQ_NEXT(td_first, td_slpq); if (td_first->td_ucred) crfree(td_first->td_ucred); thread_free(td_first); td_first = td_next; } } } /* * Allocate a thread. */ struct thread * thread_alloc(int pages) { struct thread *td; thread_reap(); /* check if any zombies to get */ td = (struct thread *)uma_zalloc(thread_zone, M_WAITOK); KASSERT(td->td_kstack == 0, ("thread_alloc got thread with kstack")); if (!vm_thread_new(td, pages)) { uma_zfree(thread_zone, td); return (NULL); } cpu_thread_alloc(td); return (td); } int thread_alloc_stack(struct thread *td, int pages) { KASSERT(td->td_kstack == 0, ("thread_alloc_stack called on a thread with kstack")); if (!vm_thread_new(td, pages)) return (0); cpu_thread_alloc(td); return (1); } /* * Deallocate a thread. */ void thread_free(struct thread *td) { lock_profile_thread_exit(td); if (td->td_cpuset) cpuset_rel(td->td_cpuset); td->td_cpuset = NULL; cpu_thread_free(td); if (td->td_kstack != 0) vm_thread_dispose(td); uma_zfree(thread_zone, td); } /* * Discard the current thread and exit from its context. * Always called with scheduler locked. * * Because we can't free a thread while we're operating under its context, * push the current thread into our CPU's deadthread holder. This means * we needn't worry about someone else grabbing our context before we * do a cpu_throw(). */ void thread_exit(void) { uint64_t runtime, new_switchtime; struct thread *td; struct thread *td2; struct proc *p; int wakeup_swapper; td = curthread; p = td->td_proc; PROC_SLOCK_ASSERT(p, MA_OWNED); mtx_assert(&Giant, MA_NOTOWNED); PROC_LOCK_ASSERT(p, MA_OWNED); KASSERT(p != NULL, ("thread exiting without a process")); CTR3(KTR_PROC, "thread_exit: thread %p (pid %ld, %s)", td, (long)p->p_pid, td->td_name); KASSERT(TAILQ_EMPTY(&td->td_sigqueue.sq_list), ("signal pending")); #ifdef AUDIT AUDIT_SYSCALL_EXIT(0, td); #endif umtx_thread_exit(td); /* * drop FPU & debug register state storage, or any other * architecture specific resources that * would not be on a new untouched process. */ cpu_thread_exit(td); /* XXXSMP */ /* * The last thread is left attached to the process * So that the whole bundle gets recycled. Skip * all this stuff if we never had threads. * EXIT clears all sign of other threads when * it goes to single threading, so the last thread always * takes the short path. */ if (p->p_flag & P_HADTHREADS) { if (p->p_numthreads > 1) { atomic_add_int(&td->td_proc->p_exitthreads, 1); thread_unlink(td); td2 = FIRST_THREAD_IN_PROC(p); sched_exit_thread(td2, td); /* * The test below is NOT true if we are the * sole exiting thread. P_STOPPED_SINGLE is unset * in exit1() after it is the only survivor. */ if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) { if (p->p_numthreads == p->p_suspcount) { thread_lock(p->p_singlethread); wakeup_swapper = thread_unsuspend_one( p->p_singlethread); thread_unlock(p->p_singlethread); if (wakeup_swapper) kick_proc0(); } } PCPU_SET(deadthread, td); } else { /* * The last thread is exiting.. but not through exit() */ panic ("thread_exit: Last thread exiting on its own"); } } #ifdef HWPMC_HOOKS /* * If this thread is part of a process that is being tracked by hwpmc(4), * inform the module of the thread's impending exit. */ if (PMC_PROC_IS_USING_PMCS(td->td_proc)) PMC_SWITCH_CONTEXT(td, PMC_FN_CSW_OUT); #endif PROC_UNLOCK(p); PROC_STATLOCK(p); thread_lock(td); PROC_SUNLOCK(p); /* Do the same timestamp bookkeeping that mi_switch() would do. */ new_switchtime = cpu_ticks(); runtime = new_switchtime - PCPU_GET(switchtime); td->td_runtime += runtime; td->td_incruntime += runtime; PCPU_SET(switchtime, new_switchtime); PCPU_SET(switchticks, ticks); PCPU_INC(cnt.v_swtch); /* Save our resource usage in our process. */ td->td_ru.ru_nvcsw++; ruxagg(p, td); rucollect(&p->p_ru, &td->td_ru); PROC_STATUNLOCK(p); td->td_state = TDS_INACTIVE; #ifdef WITNESS witness_thread_exit(td); #endif CTR1(KTR_PROC, "thread_exit: cpu_throw() thread %p", td); sched_throw(td); panic("I'm a teapot!"); /* NOTREACHED */ } /* * Do any thread specific cleanups that may be needed in wait() * called with Giant, proc and schedlock not held. */ void thread_wait(struct proc *p) { struct thread *td; mtx_assert(&Giant, MA_NOTOWNED); KASSERT(p->p_numthreads == 1, ("multiple threads in thread_wait()")); KASSERT(p->p_exitthreads == 0, ("p_exitthreads leaking")); td = FIRST_THREAD_IN_PROC(p); /* Lock the last thread so we spin until it exits cpu_throw(). */ thread_lock(td); thread_unlock(td); lock_profile_thread_exit(td); cpuset_rel(td->td_cpuset); td->td_cpuset = NULL; cpu_thread_clean(td); crfree(td->td_ucred); thread_reap(); /* check for zombie threads etc. */ } /* * Link a thread to a process. * set up anything that needs to be initialized for it to * be used by the process. */ void thread_link(struct thread *td, struct proc *p) { /* * XXX This can't be enabled because it's called for proc0 before * its lock has been created. * PROC_LOCK_ASSERT(p, MA_OWNED); */ td->td_state = TDS_INACTIVE; td->td_proc = p; td->td_flags = TDF_INMEM; LIST_INIT(&td->td_contested); LIST_INIT(&td->td_lprof[0]); LIST_INIT(&td->td_lprof[1]); sigqueue_init(&td->td_sigqueue, p); callout_init(&td->td_slpcallout, CALLOUT_MPSAFE); TAILQ_INSERT_TAIL(&p->p_threads, td, td_plist); p->p_numthreads++; } /* * Called from: * thread_exit() */ void thread_unlink(struct thread *td) { struct proc *p = td->td_proc; PROC_LOCK_ASSERT(p, MA_OWNED); TAILQ_REMOVE(&p->p_threads, td, td_plist); p->p_numthreads--; /* could clear a few other things here */ /* Must NOT clear links to proc! */ } static int calc_remaining(struct proc *p, int mode) { int remaining; PROC_LOCK_ASSERT(p, MA_OWNED); PROC_SLOCK_ASSERT(p, MA_OWNED); if (mode == SINGLE_EXIT) remaining = p->p_numthreads; else if (mode == SINGLE_BOUNDARY) remaining = p->p_numthreads - p->p_boundary_count; else if (mode == SINGLE_NO_EXIT) remaining = p->p_numthreads - p->p_suspcount; else panic("calc_remaining: wrong mode %d", mode); return (remaining); } +static int +remain_for_mode(int mode) +{ + + return (1); +} + +static int +weed_inhib(int mode, struct thread *td2, struct proc *p) +{ + int wakeup_swapper; + + PROC_LOCK_ASSERT(p, MA_OWNED); + PROC_SLOCK_ASSERT(p, MA_OWNED); + THREAD_LOCK_ASSERT(td2, MA_OWNED); + + wakeup_swapper = 0; + switch (mode) { + case SINGLE_EXIT: + if (TD_IS_SUSPENDED(td2)) + wakeup_swapper |= thread_unsuspend_one(td2); + if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0) + wakeup_swapper |= sleepq_abort(td2, EINTR); + break; + case SINGLE_BOUNDARY: + if (TD_IS_SUSPENDED(td2) && (td2->td_flags & TDF_BOUNDARY) == 0) + wakeup_swapper |= thread_unsuspend_one(td2); + if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0) + wakeup_swapper |= sleepq_abort(td2, ERESTART); + break; + case SINGLE_NO_EXIT: + if (TD_IS_SUSPENDED(td2) && (td2->td_flags & TDF_BOUNDARY) == 0) + wakeup_swapper |= thread_unsuspend_one(td2); + if (TD_ON_SLEEPQ(td2) && (td2->td_flags & TDF_SINTR) != 0) + wakeup_swapper |= sleepq_abort(td2, ERESTART); + break; + } + return (wakeup_swapper); +} + /* * Enforce single-threading. * * Returns 1 if the caller must abort (another thread is waiting to * exit the process or similar). Process is locked! * Returns 0 when you are successfully the only thread running. * A process has successfully single threaded in the suspend mode when * There are no threads in user mode. Threads in the kernel must be * allowed to continue until they get to the user boundary. They may even * copy out their return values and data before suspending. They may however be * accelerated in reaching the user boundary as we will wake up * any sleeping threads that are interruptable. (PCATCH). */ int thread_single(int mode) { struct thread *td; struct thread *td2; struct proc *p; int remaining, wakeup_swapper; td = curthread; p = td->td_proc; mtx_assert(&Giant, MA_NOTOWNED); PROC_LOCK_ASSERT(p, MA_OWNED); if ((p->p_flag & P_HADTHREADS) == 0) return (0); /* Is someone already single threading? */ if (p->p_singlethread != NULL && p->p_singlethread != td) return (1); if (mode == SINGLE_EXIT) { p->p_flag |= P_SINGLE_EXIT; p->p_flag &= ~P_SINGLE_BOUNDARY; } else { p->p_flag &= ~P_SINGLE_EXIT; if (mode == SINGLE_BOUNDARY) p->p_flag |= P_SINGLE_BOUNDARY; else p->p_flag &= ~P_SINGLE_BOUNDARY; } p->p_flag |= P_STOPPED_SINGLE; PROC_SLOCK(p); p->p_singlethread = td; remaining = calc_remaining(p, mode); - while (remaining != 1) { + while (remaining != remain_for_mode(mode)) { if (P_SHOULDSTOP(p) != P_STOPPED_SINGLE) goto stopme; wakeup_swapper = 0; FOREACH_THREAD_IN_PROC(p, td2) { if (td2 == td) continue; thread_lock(td2); td2->td_flags |= TDF_ASTPENDING | TDF_NEEDSUSPCHK; - if (TD_IS_INHIBITED(td2)) { - switch (mode) { - case SINGLE_EXIT: - if (TD_IS_SUSPENDED(td2)) - wakeup_swapper |= - thread_unsuspend_one(td2); - if (TD_ON_SLEEPQ(td2) && - (td2->td_flags & TDF_SINTR)) - wakeup_swapper |= - sleepq_abort(td2, EINTR); - break; - case SINGLE_BOUNDARY: - if (TD_IS_SUSPENDED(td2) && - !(td2->td_flags & TDF_BOUNDARY)) - wakeup_swapper |= - thread_unsuspend_one(td2); - if (TD_ON_SLEEPQ(td2) && - (td2->td_flags & TDF_SINTR)) - wakeup_swapper |= - sleepq_abort(td2, ERESTART); - break; - case SINGLE_NO_EXIT: - if (TD_IS_SUSPENDED(td2) && - !(td2->td_flags & TDF_BOUNDARY)) - wakeup_swapper |= - thread_unsuspend_one(td2); - if (TD_ON_SLEEPQ(td2) && - (td2->td_flags & TDF_SINTR)) - wakeup_swapper |= - sleepq_abort(td2, ERESTART); - break; - default: - break; - } - } + if (TD_IS_INHIBITED(td2)) + wakeup_swapper |= weed_inhib(mode, td2, p); #ifdef SMP else if (TD_IS_RUNNING(td2) && td != td2) { forward_signal(td2); } #endif thread_unlock(td2); } if (wakeup_swapper) kick_proc0(); remaining = calc_remaining(p, mode); /* * Maybe we suspended some threads.. was it enough? */ - if (remaining == 1) + if (remaining == remain_for_mode(mode)) break; stopme: /* * Wake us up when everyone else has suspended. * In the mean time we suspend as well. */ thread_suspend_switch(td); remaining = calc_remaining(p, mode); } if (mode == SINGLE_EXIT) { /* * Convert the process to an unthreaded process. The * SINGLE_EXIT is called by exit1() or execve(), in * both cases other threads must be retired. */ KASSERT(p->p_numthreads == 1, ("Unthreading with >1 threads")); p->p_singlethread = NULL; p->p_flag &= ~(P_STOPPED_SINGLE | P_SINGLE_EXIT | P_HADTHREADS); /* * Wait for any remaining threads to exit cpu_throw(). */ while (p->p_exitthreads != 0) { PROC_SUNLOCK(p); PROC_UNLOCK(p); sched_relinquish(td); PROC_LOCK(p); PROC_SLOCK(p); } } PROC_SUNLOCK(p); return (0); } +bool +thread_suspend_check_needed(void) +{ + struct proc *p; + struct thread *td; + + td = curthread; + p = td->td_proc; + PROC_LOCK_ASSERT(p, MA_OWNED); + return (P_SHOULDSTOP(p) || ((p->p_flag & P_TRACED) != 0 && + (td->td_dbgflags & TDB_SUSPEND) != 0)); +} + /* * Called in from locations that can safely check to see * whether we have to suspend or at least throttle for a * single-thread event (e.g. fork). * * Such locations include userret(). * If the "return_instead" argument is non zero, the thread must be able to * accept 0 (caller may continue), or 1 (caller must abort) as a result. * * The 'return_instead' argument tells the function if it may do a * thread_exit() or suspend, or whether the caller must abort and back * out instead. * * If the thread that set the single_threading request has set the * P_SINGLE_EXIT bit in the process flags then this call will never return * if 'return_instead' is false, but will exit. * * P_SINGLE_EXIT | return_instead == 0| return_instead != 0 *---------------+--------------------+--------------------- * 0 | returns 0 | returns 0 or 1 * | when ST ends | immediately *---------------+--------------------+--------------------- * 1 | thread exits | returns 1 * | | immediately * 0 = thread_exit() or suspension ok, * other = return error instead of stopping the thread. * * While a full suspension is under effect, even a single threading * thread would be suspended if it made this call (but it shouldn't). * This call should only be made from places where * thread_exit() would be safe as that may be the outcome unless * return_instead is set. */ int thread_suspend_check(int return_instead) { struct thread *td; struct proc *p; int wakeup_swapper; td = curthread; p = td->td_proc; mtx_assert(&Giant, MA_NOTOWNED); PROC_LOCK_ASSERT(p, MA_OWNED); - while (P_SHOULDSTOP(p) || - ((p->p_flag & P_TRACED) && (td->td_dbgflags & TDB_SUSPEND))) { + while (thread_suspend_check_needed()) { if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) { KASSERT(p->p_singlethread != NULL, ("singlethread not set")); /* * The only suspension in action is a * single-threading. Single threader need not stop. * XXX Should be safe to access unlocked * as it can only be set to be true by us. */ if (p->p_singlethread == td) return (0); /* Exempt from stopping. */ } if ((p->p_flag & P_SINGLE_EXIT) && return_instead) return (EINTR); /* Should we goto user boundary if we didn't come from there? */ if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE && (p->p_flag & P_SINGLE_BOUNDARY) && return_instead) return (ERESTART); /* * Ignore suspend requests for stop signals if they * are deferred. */ if (P_SHOULDSTOP(p) == P_STOPPED_SIG && td->td_flags & TDF_SBDRY) { KASSERT(return_instead, ("TDF_SBDRY set for unsafe thread_suspend_check")); return (0); } /* * If the process is waiting for us to exit, * this thread should just suicide. * Assumes that P_SINGLE_EXIT implies P_STOPPED_SINGLE. */ if ((p->p_flag & P_SINGLE_EXIT) && (p->p_singlethread != td)) { PROC_UNLOCK(p); tidhash_remove(td); PROC_LOCK(p); tdsigcleanup(td); PROC_SLOCK(p); thread_stopped(p); thread_exit(); } PROC_SLOCK(p); thread_stopped(p); if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) { if (p->p_numthreads == p->p_suspcount + 1) { thread_lock(p->p_singlethread); wakeup_swapper = thread_unsuspend_one(p->p_singlethread); thread_unlock(p->p_singlethread); if (wakeup_swapper) kick_proc0(); } } PROC_UNLOCK(p); thread_lock(td); /* * When a thread suspends, it just * gets taken off all queues. */ thread_suspend_one(td); if (return_instead == 0) { p->p_boundary_count++; td->td_flags |= TDF_BOUNDARY; } PROC_SUNLOCK(p); mi_switch(SW_INVOL | SWT_SUSPEND, NULL); if (return_instead == 0) td->td_flags &= ~TDF_BOUNDARY; thread_unlock(td); PROC_LOCK(p); if (return_instead == 0) { PROC_SLOCK(p); p->p_boundary_count--; PROC_SUNLOCK(p); } } return (0); } void thread_suspend_switch(struct thread *td) { struct proc *p; p = td->td_proc; KASSERT(!TD_IS_SUSPENDED(td), ("already suspended")); PROC_LOCK_ASSERT(p, MA_OWNED); PROC_SLOCK_ASSERT(p, MA_OWNED); /* * We implement thread_suspend_one in stages here to avoid * dropping the proc lock while the thread lock is owned. */ thread_stopped(p); p->p_suspcount++; PROC_UNLOCK(p); thread_lock(td); td->td_flags &= ~TDF_NEEDSUSPCHK; TD_SET_SUSPENDED(td); sched_sleep(td, 0); PROC_SUNLOCK(p); DROP_GIANT(); mi_switch(SW_VOL | SWT_SUSPEND, NULL); thread_unlock(td); PICKUP_GIANT(); PROC_LOCK(p); PROC_SLOCK(p); } void thread_suspend_one(struct thread *td) { struct proc *p = td->td_proc; PROC_SLOCK_ASSERT(p, MA_OWNED); THREAD_LOCK_ASSERT(td, MA_OWNED); KASSERT(!TD_IS_SUSPENDED(td), ("already suspended")); p->p_suspcount++; td->td_flags &= ~TDF_NEEDSUSPCHK; TD_SET_SUSPENDED(td); sched_sleep(td, 0); } int thread_unsuspend_one(struct thread *td) { struct proc *p = td->td_proc; PROC_SLOCK_ASSERT(p, MA_OWNED); THREAD_LOCK_ASSERT(td, MA_OWNED); KASSERT(TD_IS_SUSPENDED(td), ("Thread not suspended")); TD_CLR_SUSPENDED(td); p->p_suspcount--; return (setrunnable(td)); } /* * Allow all threads blocked by single threading to continue running. */ void thread_unsuspend(struct proc *p) { struct thread *td; int wakeup_swapper; PROC_LOCK_ASSERT(p, MA_OWNED); PROC_SLOCK_ASSERT(p, MA_OWNED); wakeup_swapper = 0; if (!P_SHOULDSTOP(p)) { FOREACH_THREAD_IN_PROC(p, td) { thread_lock(td); if (TD_IS_SUSPENDED(td)) { wakeup_swapper |= thread_unsuspend_one(td); } thread_unlock(td); } } else if ((P_SHOULDSTOP(p) == P_STOPPED_SINGLE) && (p->p_numthreads == p->p_suspcount)) { /* * Stopping everything also did the job for the single * threading request. Now we've downgraded to single-threaded, * let it continue. */ thread_lock(p->p_singlethread); wakeup_swapper = thread_unsuspend_one(p->p_singlethread); thread_unlock(p->p_singlethread); } if (wakeup_swapper) kick_proc0(); } /* * End the single threading mode.. */ void thread_single_end(void) { struct thread *td; struct proc *p; int wakeup_swapper; - td = curthread; - p = td->td_proc; + p = curproc; PROC_LOCK_ASSERT(p, MA_OWNED); p->p_flag &= ~(P_STOPPED_SINGLE | P_SINGLE_EXIT | P_SINGLE_BOUNDARY); PROC_SLOCK(p); p->p_singlethread = NULL; wakeup_swapper = 0; /* * If there are other threads they may now run, * unless of course there is a blanket 'stop order' * on the process. The single threader must be allowed * to continue however as this is a bad place to stop. */ - if ((p->p_numthreads != 1) && (!P_SHOULDSTOP(p))) { + if (p->p_numthreads != remain_for_mode(SINGLE_EXIT) && + !P_SHOULDSTOP(p)) { FOREACH_THREAD_IN_PROC(p, td) { thread_lock(td); if (TD_IS_SUSPENDED(td)) { wakeup_swapper |= thread_unsuspend_one(td); } thread_unlock(td); } } PROC_SUNLOCK(p); if (wakeup_swapper) kick_proc0(); } struct thread * thread_find(struct proc *p, lwpid_t tid) { struct thread *td; PROC_LOCK_ASSERT(p, MA_OWNED); FOREACH_THREAD_IN_PROC(p, td) { if (td->td_tid == tid) break; } return (td); } /* Locate a thread by number; return with proc lock held. */ struct thread * tdfind(lwpid_t tid, pid_t pid) { #define RUN_THRESH 16 struct thread *td; int run = 0; rw_rlock(&tidhash_lock); LIST_FOREACH(td, TIDHASH(tid), td_hash) { if (td->td_tid == tid) { if (pid != -1 && td->td_proc->p_pid != pid) { td = NULL; break; } PROC_LOCK(td->td_proc); if (td->td_proc->p_state == PRS_NEW) { PROC_UNLOCK(td->td_proc); td = NULL; break; } if (run > RUN_THRESH) { if (rw_try_upgrade(&tidhash_lock)) { LIST_REMOVE(td, td_hash); LIST_INSERT_HEAD(TIDHASH(td->td_tid), td, td_hash); rw_wunlock(&tidhash_lock); return (td); } } break; } run++; } rw_runlock(&tidhash_lock); return (td); } void tidhash_add(struct thread *td) { rw_wlock(&tidhash_lock); LIST_INSERT_HEAD(TIDHASH(td->td_tid), td, td_hash); rw_wunlock(&tidhash_lock); } void tidhash_remove(struct thread *td) { rw_wlock(&tidhash_lock); LIST_REMOVE(td, td_hash); rw_wunlock(&tidhash_lock); } Index: projects/building-blocks/sys/kern/subr_syscall.c =================================================================== --- projects/building-blocks/sys/kern/subr_syscall.c (revision 275663) +++ projects/building-blocks/sys/kern/subr_syscall.c (revision 275664) @@ -1,234 +1,245 @@ /*- * Copyright (C) 1994, David Greenman * Copyright (c) 1990, 1993 * The Regents of the University of California. All rights reserved. * Copyright (C) 2010 Konstantin Belousov * * This code is derived from software contributed to Berkeley by * the University of Utah, and William Jolitz. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * This product includes software developed by the University of * California, Berkeley and its contributors. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * from: @(#)trap.c 7.4 (Berkeley) 5/13/91 */ #include "opt_capsicum.h" #include "opt_ktrace.h" __FBSDID("$FreeBSD$"); #include #include #ifdef KTRACE #include #include #endif #include static inline int syscallenter(struct thread *td, struct syscall_args *sa) { struct proc *p; int error, traced; PCPU_INC(cnt.v_syscall); p = td->td_proc; td->td_pticks = 0; if (td->td_ucred != p->p_ucred) cred_update_thread(td); if (p->p_flag & P_TRACED) { traced = 1; PROC_LOCK(p); td->td_dbgflags &= ~TDB_USERWR; td->td_dbgflags |= TDB_SCE; PROC_UNLOCK(p); } else traced = 0; error = (p->p_sysent->sv_fetch_syscall_args)(td, sa); #ifdef KTRACE if (KTRPOINT(td, KTR_SYSCALL)) ktrsyscall(sa->code, sa->narg, sa->args); #endif KTR_START4(KTR_SYSC, "syscall", syscallname(p, sa->code), (uintptr_t)td, "pid:%d", td->td_proc->p_pid, "arg0:%p", sa->args[0], "arg1:%p", sa->args[1], "arg2:%p", sa->args[2]); if (error == 0) { STOPEVENT(p, S_SCE, sa->narg); if (p->p_flag & P_TRACED && p->p_stops & S_PT_SCE) { PROC_LOCK(p); ptracestop((td), SIGTRAP); PROC_UNLOCK(p); } if (td->td_dbgflags & TDB_USERWR) { /* * Reread syscall number and arguments if * debugger modified registers or memory. */ error = (p->p_sysent->sv_fetch_syscall_args)(td, sa); #ifdef KTRACE if (KTRPOINT(td, KTR_SYSCALL)) ktrsyscall(sa->code, sa->narg, sa->args); #endif if (error != 0) goto retval; } #ifdef CAPABILITY_MODE /* * In capability mode, we only allow access to system calls * flagged with SYF_CAPENABLED. */ if (IN_CAPABILITY_MODE(td) && !(sa->callp->sy_flags & SYF_CAPENABLED)) { error = ECAPMODE; goto retval; } #endif error = syscall_thread_enter(td, sa->callp); if (error != 0) goto retval; #ifdef KDTRACE_HOOKS /* * If the systrace module has registered it's probe * callback and if there is a probe active for the * syscall 'entry', process the probe. */ if (systrace_probe_func != NULL && sa->callp->sy_entry != 0) (*systrace_probe_func)(sa->callp->sy_entry, sa->code, sa->callp, sa->args, 0); #endif AUDIT_SYSCALL_ENTER(sa->code, td); error = (sa->callp->sy_call)(td, sa->args); AUDIT_SYSCALL_EXIT(error, td); /* Save the latest error return value. */ if ((td->td_pflags & TDP_NERRNO) == 0) td->td_errno = error; #ifdef KDTRACE_HOOKS /* * If the systrace module has registered it's probe * callback and if there is a probe active for the * syscall 'return', process the probe. */ if (systrace_probe_func != NULL && sa->callp->sy_return != 0) (*systrace_probe_func)(sa->callp->sy_return, sa->code, sa->callp, NULL, (error) ? -1 : td->td_retval[0]); #endif syscall_thread_exit(td, sa->callp); } retval: KTR_STOP4(KTR_SYSC, "syscall", syscallname(p, sa->code), (uintptr_t)td, "pid:%d", td->td_proc->p_pid, "error:%d", error, "retval0:%#lx", td->td_retval[0], "retval1:%#lx", td->td_retval[1]); if (traced) { PROC_LOCK(p); td->td_dbgflags &= ~TDB_SCE; PROC_UNLOCK(p); } (p->p_sysent->sv_set_syscall_retval)(td, error); return (error); } static inline void syscallret(struct thread *td, int error, struct syscall_args *sa __unused) { struct proc *p, *p2; int traced; p = td->td_proc; /* * Handle reschedule and other end-of-syscall issues */ userret(td, td->td_frame); #ifdef KTRACE if (KTRPOINT(td, KTR_SYSRET)) { ktrsysret(sa->code, (td->td_pflags & TDP_NERRNO) == 0 ? error : td->td_errno, td->td_retval[0]); } #endif td->td_pflags &= ~TDP_NERRNO; if (p->p_flag & P_TRACED) { traced = 1; PROC_LOCK(p); td->td_dbgflags |= TDB_SCX; PROC_UNLOCK(p); } else traced = 0; /* * This works because errno is findable through the * register set. If we ever support an emulation where this * is not the case, this code will need to be revisited. */ STOPEVENT(p, S_SCX, sa->code); if (traced || (td->td_dbgflags & (TDB_EXEC | TDB_FORK)) != 0) { PROC_LOCK(p); /* * If tracing the execed process, trap to the debugger * so that breakpoints can be set before the program * executes. If debugger requested tracing of syscall * returns, do it now too. */ if (traced && ((td->td_dbgflags & (TDB_FORK | TDB_EXEC)) != 0 || (p->p_stops & S_PT_SCX) != 0)) ptracestop(td, SIGTRAP); td->td_dbgflags &= ~(TDB_SCX | TDB_EXEC | TDB_FORK); PROC_UNLOCK(p); } if (td->td_pflags & TDP_RFPPWAIT) { /* * Preserve synchronization semantics of vfork. If * waiting for child to exec or exit, fork set * P_PPWAIT on child, and there we sleep on our proc * (in case of exit). * * Do it after the ptracestop() above is finished, to * not block our debugger until child execs or exits * to finish vfork wait. */ td->td_pflags &= ~TDP_RFPPWAIT; p2 = td->td_rfppwait_p; +again: PROC_LOCK(p2); - while (p2->p_flag & P_PPWAIT) - cv_wait(&p2->p_pwait, &p2->p_mtx); + while (p2->p_flag & P_PPWAIT) { + PROC_LOCK(p); + if (thread_suspend_check_needed()) { + PROC_UNLOCK(p2); + thread_suspend_check(0); + PROC_UNLOCK(p); + goto again; + } else { + PROC_UNLOCK(p); + } + cv_timedwait(&p2->p_pwait, &p2->p_mtx, hz); + } PROC_UNLOCK(p2); } } Index: projects/building-blocks/sys/kern/vfs_bio.c =================================================================== --- projects/building-blocks/sys/kern/vfs_bio.c (revision 275663) +++ projects/building-blocks/sys/kern/vfs_bio.c (revision 275664) @@ -1,4630 +1,4634 @@ /*- * Copyright (c) 2004 Poul-Henning Kamp * Copyright (c) 1994,1997 John S. Dyson * Copyright (c) 2013 The FreeBSD Foundation * All rights reserved. * * Portions of this software were developed by Konstantin Belousov * under sponsorship from the FreeBSD Foundation. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ /* * this file contains a new buffer I/O scheme implementing a coherent * VM object and buffer cache scheme. Pains have been taken to make * sure that the performance degradation associated with schemes such * as this is not realized. * * Author: John S. Dyson * Significant help during the development and debugging phases * had been provided by David Greenman, also of the FreeBSD core team. * * see man buf(9) for more info. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "opt_compat.h" #include "opt_swap.h" static MALLOC_DEFINE(M_BIOBUF, "biobuf", "BIO buffer"); struct bio_ops bioops; /* I/O operation notification */ struct buf_ops buf_ops_bio = { .bop_name = "buf_ops_bio", .bop_write = bufwrite, .bop_strategy = bufstrategy, .bop_sync = bufsync, .bop_bdflush = bufbdflush, }; /* * XXX buf is global because kern_shutdown.c and ffs_checkoverlap has * carnal knowledge of buffers. This knowledge should be moved to vfs_bio.c. */ struct buf *buf; /* buffer header pool */ caddr_t unmapped_buf; /* Used below and for softdep flushing threads in ufs/ffs/ffs_softdep.c */ struct proc *bufdaemonproc; static int inmem(struct vnode *vp, daddr_t blkno); static void vm_hold_free_pages(struct buf *bp, int newbsize); static void vm_hold_load_pages(struct buf *bp, vm_offset_t from, vm_offset_t to); static void vfs_page_set_valid(struct buf *bp, vm_ooffset_t off, vm_page_t m); static void vfs_page_set_validclean(struct buf *bp, vm_ooffset_t off, vm_page_t m); static void vfs_clean_pages_dirty_buf(struct buf *bp); static void vfs_setdirty_locked_object(struct buf *bp); static void vfs_vmio_release(struct buf *bp); static int vfs_bio_clcheck(struct vnode *vp, int size, daddr_t lblkno, daddr_t blkno); static int buf_flush(int); static int flushbufqueues(int, int); static void buf_daemon(void); static void bremfreel(struct buf *bp); static __inline void bd_wakeup(void); static int sysctl_runningspace(SYSCTL_HANDLER_ARGS); #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \ defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7) static int sysctl_bufspace(SYSCTL_HANDLER_ARGS); #endif int vmiodirenable = TRUE; SYSCTL_INT(_vfs, OID_AUTO, vmiodirenable, CTLFLAG_RW, &vmiodirenable, 0, "Use the VM system for directory writes"); long runningbufspace; SYSCTL_LONG(_vfs, OID_AUTO, runningbufspace, CTLFLAG_RD, &runningbufspace, 0, "Amount of presently outstanding async buffer io"); static long bufspace; #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \ defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7) SYSCTL_PROC(_vfs, OID_AUTO, bufspace, CTLTYPE_LONG|CTLFLAG_MPSAFE|CTLFLAG_RD, &bufspace, 0, sysctl_bufspace, "L", "Virtual memory used for buffers"); #else SYSCTL_LONG(_vfs, OID_AUTO, bufspace, CTLFLAG_RD, &bufspace, 0, "Virtual memory used for buffers"); #endif static long unmapped_bufspace; SYSCTL_LONG(_vfs, OID_AUTO, unmapped_bufspace, CTLFLAG_RD, &unmapped_bufspace, 0, "Amount of unmapped buffers, inclusive in the bufspace"); static long maxbufspace; SYSCTL_LONG(_vfs, OID_AUTO, maxbufspace, CTLFLAG_RD, &maxbufspace, 0, "Maximum allowed value of bufspace (including buf_daemon)"); static long bufmallocspace; SYSCTL_LONG(_vfs, OID_AUTO, bufmallocspace, CTLFLAG_RD, &bufmallocspace, 0, "Amount of malloced memory for buffers"); static long maxbufmallocspace; SYSCTL_LONG(_vfs, OID_AUTO, maxmallocbufspace, CTLFLAG_RW, &maxbufmallocspace, 0, "Maximum amount of malloced memory for buffers"); static long lobufspace; SYSCTL_LONG(_vfs, OID_AUTO, lobufspace, CTLFLAG_RD, &lobufspace, 0, "Minimum amount of buffers we want to have"); long hibufspace; SYSCTL_LONG(_vfs, OID_AUTO, hibufspace, CTLFLAG_RD, &hibufspace, 0, "Maximum allowed value of bufspace (excluding buf_daemon)"); static int bufreusecnt; SYSCTL_INT(_vfs, OID_AUTO, bufreusecnt, CTLFLAG_RW, &bufreusecnt, 0, "Number of times we have reused a buffer"); static int buffreekvacnt; SYSCTL_INT(_vfs, OID_AUTO, buffreekvacnt, CTLFLAG_RW, &buffreekvacnt, 0, "Number of times we have freed the KVA space from some buffer"); static int bufdefragcnt; SYSCTL_INT(_vfs, OID_AUTO, bufdefragcnt, CTLFLAG_RW, &bufdefragcnt, 0, "Number of times we have had to repeat buffer allocation to defragment"); static long lorunningspace; SYSCTL_PROC(_vfs, OID_AUTO, lorunningspace, CTLTYPE_LONG | CTLFLAG_MPSAFE | CTLFLAG_RW, &lorunningspace, 0, sysctl_runningspace, "L", "Minimum preferred space used for in-progress I/O"); static long hirunningspace; SYSCTL_PROC(_vfs, OID_AUTO, hirunningspace, CTLTYPE_LONG | CTLFLAG_MPSAFE | CTLFLAG_RW, &hirunningspace, 0, sysctl_runningspace, "L", "Maximum amount of space to use for in-progress I/O"); int dirtybufferflushes; SYSCTL_INT(_vfs, OID_AUTO, dirtybufferflushes, CTLFLAG_RW, &dirtybufferflushes, 0, "Number of bdwrite to bawrite conversions to limit dirty buffers"); int bdwriteskip; SYSCTL_INT(_vfs, OID_AUTO, bdwriteskip, CTLFLAG_RW, &bdwriteskip, 0, "Number of buffers supplied to bdwrite with snapshot deadlock risk"); int altbufferflushes; SYSCTL_INT(_vfs, OID_AUTO, altbufferflushes, CTLFLAG_RW, &altbufferflushes, 0, "Number of fsync flushes to limit dirty buffers"); static int recursiveflushes; SYSCTL_INT(_vfs, OID_AUTO, recursiveflushes, CTLFLAG_RW, &recursiveflushes, 0, "Number of flushes skipped due to being recursive"); static int numdirtybuffers; SYSCTL_INT(_vfs, OID_AUTO, numdirtybuffers, CTLFLAG_RD, &numdirtybuffers, 0, "Number of buffers that are dirty (has unwritten changes) at the moment"); static int lodirtybuffers; SYSCTL_INT(_vfs, OID_AUTO, lodirtybuffers, CTLFLAG_RW, &lodirtybuffers, 0, "How many buffers we want to have free before bufdaemon can sleep"); static int hidirtybuffers; SYSCTL_INT(_vfs, OID_AUTO, hidirtybuffers, CTLFLAG_RW, &hidirtybuffers, 0, "When the number of dirty buffers is considered severe"); int dirtybufthresh; SYSCTL_INT(_vfs, OID_AUTO, dirtybufthresh, CTLFLAG_RW, &dirtybufthresh, 0, "Number of bdwrite to bawrite conversions to clear dirty buffers"); static int numfreebuffers; SYSCTL_INT(_vfs, OID_AUTO, numfreebuffers, CTLFLAG_RD, &numfreebuffers, 0, "Number of free buffers"); static int lofreebuffers; SYSCTL_INT(_vfs, OID_AUTO, lofreebuffers, CTLFLAG_RW, &lofreebuffers, 0, "XXX Unused"); static int hifreebuffers; SYSCTL_INT(_vfs, OID_AUTO, hifreebuffers, CTLFLAG_RW, &hifreebuffers, 0, "XXX Complicatedly unused"); static int getnewbufcalls; SYSCTL_INT(_vfs, OID_AUTO, getnewbufcalls, CTLFLAG_RW, &getnewbufcalls, 0, "Number of calls to getnewbuf"); static int getnewbufrestarts; SYSCTL_INT(_vfs, OID_AUTO, getnewbufrestarts, CTLFLAG_RW, &getnewbufrestarts, 0, "Number of times getnewbuf has had to restart a buffer aquisition"); static int mappingrestarts; SYSCTL_INT(_vfs, OID_AUTO, mappingrestarts, CTLFLAG_RW, &mappingrestarts, 0, "Number of times getblk has had to restart a buffer mapping for " "unmapped buffer"); static int flushbufqtarget = 100; SYSCTL_INT(_vfs, OID_AUTO, flushbufqtarget, CTLFLAG_RW, &flushbufqtarget, 0, "Amount of work to do in flushbufqueues when helping bufdaemon"); static long notbufdflushes; SYSCTL_LONG(_vfs, OID_AUTO, notbufdflushes, CTLFLAG_RD, ¬bufdflushes, 0, "Number of dirty buffer flushes done by the bufdaemon helpers"); static long barrierwrites; SYSCTL_LONG(_vfs, OID_AUTO, barrierwrites, CTLFLAG_RW, &barrierwrites, 0, "Number of barrier writes"); SYSCTL_INT(_vfs, OID_AUTO, unmapped_buf_allowed, CTLFLAG_RD, &unmapped_buf_allowed, 0, "Permit the use of the unmapped i/o"); /* * Lock for the non-dirty bufqueues */ static struct mtx_padalign bqclean; /* * Lock for the dirty queue. */ static struct mtx_padalign bqdirty; /* * This lock synchronizes access to bd_request. */ static struct mtx_padalign bdlock; /* * This lock protects the runningbufreq and synchronizes runningbufwakeup and * waitrunningbufspace(). */ static struct mtx_padalign rbreqlock; /* * Lock that protects needsbuffer and the sleeps/wakeups surrounding it. */ static struct rwlock_padalign nblock; /* * Lock that protects bdirtywait. */ static struct mtx_padalign bdirtylock; /* * Wakeup point for bufdaemon, as well as indicator of whether it is already * active. Set to 1 when the bufdaemon is already "on" the queue, 0 when it * is idling. */ static int bd_request; /* * Request for the buf daemon to write more buffers than is indicated by * lodirtybuf. This may be necessary to push out excess dependencies or * defragment the address space where a simple count of the number of dirty * buffers is insufficient to characterize the demand for flushing them. */ static int bd_speedupreq; /* * bogus page -- for I/O to/from partially complete buffers * this is a temporary solution to the problem, but it is not * really that bad. it would be better to split the buffer * for input in the case of buffers partially already in memory, * but the code is intricate enough already. */ vm_page_t bogus_page; /* * Synchronization (sleep/wakeup) variable for active buffer space requests. * Set when wait starts, cleared prior to wakeup(). * Used in runningbufwakeup() and waitrunningbufspace(). */ static int runningbufreq; /* * Synchronization (sleep/wakeup) variable for buffer requests. * Can contain the VFS_BIO_NEED flags defined below; setting/clearing is done * by and/or. * Used in numdirtywakeup(), bufspacewakeup(), bufcountadd(), bwillwrite(), * getnewbuf(), and getblk(). */ static volatile int needsbuffer; /* * Synchronization for bwillwrite() waiters. */ static int bdirtywait; /* * Definitions for the buffer free lists. */ #define BUFFER_QUEUES 5 /* number of free buffer queues */ #define QUEUE_NONE 0 /* on no queue */ #define QUEUE_CLEAN 1 /* non-B_DELWRI buffers */ #define QUEUE_DIRTY 2 /* B_DELWRI buffers */ #define QUEUE_EMPTYKVA 3 /* empty buffer headers w/KVA assignment */ #define QUEUE_EMPTY 4 /* empty buffer headers */ #define QUEUE_SENTINEL 1024 /* not an queue index, but mark for sentinel */ /* Queues for free buffers with various properties */ static TAILQ_HEAD(bqueues, buf) bufqueues[BUFFER_QUEUES] = { { 0 } }; #ifdef INVARIANTS static int bq_len[BUFFER_QUEUES]; #endif /* * Single global constant for BUF_WMESG, to avoid getting multiple references. * buf_wmesg is referred from macros. */ const char *buf_wmesg = BUF_WMESG; #define VFS_BIO_NEED_ANY 0x01 /* any freeable buffer */ #define VFS_BIO_NEED_FREE 0x04 /* wait for free bufs, hi hysteresis */ #define VFS_BIO_NEED_BUFSPACE 0x08 /* wait for buf space, lo hysteresis */ static int sysctl_runningspace(SYSCTL_HANDLER_ARGS) { long value; int error; value = *(long *)arg1; error = sysctl_handle_long(oidp, &value, 0, req); if (error != 0 || req->newptr == NULL) return (error); mtx_lock(&rbreqlock); if (arg1 == &hirunningspace) { if (value < lorunningspace) error = EINVAL; else hirunningspace = value; } else { KASSERT(arg1 == &lorunningspace, ("%s: unknown arg1", __func__)); if (value > hirunningspace) error = EINVAL; else lorunningspace = value; } mtx_unlock(&rbreqlock); return (error); } #if defined(COMPAT_FREEBSD4) || defined(COMPAT_FREEBSD5) || \ defined(COMPAT_FREEBSD6) || defined(COMPAT_FREEBSD7) static int sysctl_bufspace(SYSCTL_HANDLER_ARGS) { long lvalue; int ivalue; if (sizeof(int) == sizeof(long) || req->oldlen >= sizeof(long)) return (sysctl_handle_long(oidp, arg1, arg2, req)); lvalue = *(long *)arg1; if (lvalue > INT_MAX) /* On overflow, still write out a long to trigger ENOMEM. */ return (sysctl_handle_long(oidp, &lvalue, 0, req)); ivalue = lvalue; return (sysctl_handle_int(oidp, &ivalue, 0, req)); } #endif /* * bqlock: * * Return the appropriate queue lock based on the index. */ static inline struct mtx * bqlock(int qindex) { if (qindex == QUEUE_DIRTY) return (struct mtx *)(&bqdirty); return (struct mtx *)(&bqclean); } /* * bdirtywakeup: * * Wakeup any bwillwrite() waiters. */ static void bdirtywakeup(void) { mtx_lock(&bdirtylock); if (bdirtywait) { bdirtywait = 0; wakeup(&bdirtywait); } mtx_unlock(&bdirtylock); } /* * bdirtysub: * * Decrement the numdirtybuffers count by one and wakeup any * threads blocked in bwillwrite(). */ static void bdirtysub(void) { if (atomic_fetchadd_int(&numdirtybuffers, -1) == (lodirtybuffers + hidirtybuffers) / 2) bdirtywakeup(); } /* * bdirtyadd: * * Increment the numdirtybuffers count by one and wakeup the buf * daemon if needed. */ static void bdirtyadd(void) { /* * Only do the wakeup once as we cross the boundary. The * buf daemon will keep running until the condition clears. */ if (atomic_fetchadd_int(&numdirtybuffers, 1) == (lodirtybuffers + hidirtybuffers) / 2) bd_wakeup(); } /* * bufspacewakeup: * * Called when buffer space is potentially available for recovery. * getnewbuf() will block on this flag when it is unable to free * sufficient buffer space. Buffer space becomes recoverable when * bp's get placed back in the queues. */ static __inline void bufspacewakeup(void) { int need_wakeup, on; /* * If someone is waiting for BUF space, wake them up. Even * though we haven't freed the kva space yet, the waiting * process will be able to now. */ rw_rlock(&nblock); for (;;) { need_wakeup = 0; on = needsbuffer; if ((on & VFS_BIO_NEED_BUFSPACE) == 0) break; need_wakeup = 1; if (atomic_cmpset_rel_int(&needsbuffer, on, on & ~VFS_BIO_NEED_BUFSPACE)) break; } if (need_wakeup) wakeup(__DEVOLATILE(void *, &needsbuffer)); rw_runlock(&nblock); } /* * runningwakeup: * * Wake up processes that are waiting on asynchronous writes to fall * below lorunningspace. */ static void runningwakeup(void) { mtx_lock(&rbreqlock); if (runningbufreq) { runningbufreq = 0; wakeup(&runningbufreq); } mtx_unlock(&rbreqlock); } /* * runningbufwakeup: * * Decrement the outstanding write count according. */ void runningbufwakeup(struct buf *bp) { long space, bspace; bspace = bp->b_runningbufspace; if (bspace == 0) return; space = atomic_fetchadd_long(&runningbufspace, -bspace); KASSERT(space >= bspace, ("runningbufspace underflow %ld %ld", space, bspace)); bp->b_runningbufspace = 0; /* * Only acquire the lock and wakeup on the transition from exceeding * the threshold to falling below it. */ if (space < lorunningspace) return; if (space - bspace > lorunningspace) return; runningwakeup(); } /* * bufcountadd: * * Called when a buffer has been added to one of the free queues to * account for the buffer and to wakeup anyone waiting for free buffers. * This typically occurs when large amounts of metadata are being handled * by the buffer cache ( else buffer space runs out first, usually ). */ static __inline void bufcountadd(struct buf *bp) { int mask, need_wakeup, old, on; KASSERT((bp->b_flags & B_INFREECNT) == 0, ("buf %p already counted as free", bp)); bp->b_flags |= B_INFREECNT; old = atomic_fetchadd_int(&numfreebuffers, 1); KASSERT(old >= 0 && old < nbuf, ("numfreebuffers climbed to %d", old + 1)); mask = VFS_BIO_NEED_ANY; if (numfreebuffers >= hifreebuffers) mask |= VFS_BIO_NEED_FREE; rw_rlock(&nblock); for (;;) { need_wakeup = 0; on = needsbuffer; if (on == 0) break; need_wakeup = 1; if (atomic_cmpset_rel_int(&needsbuffer, on, on & ~mask)) break; } if (need_wakeup) wakeup(__DEVOLATILE(void *, &needsbuffer)); rw_runlock(&nblock); } /* * bufcountsub: * * Decrement the numfreebuffers count as needed. */ static void bufcountsub(struct buf *bp) { int old; /* * Fixup numfreebuffers count. If the buffer is invalid or not * delayed-write, the buffer was free and we must decrement * numfreebuffers. */ if ((bp->b_flags & B_INVAL) || (bp->b_flags & B_DELWRI) == 0) { KASSERT((bp->b_flags & B_INFREECNT) != 0, ("buf %p not counted in numfreebuffers", bp)); bp->b_flags &= ~B_INFREECNT; old = atomic_fetchadd_int(&numfreebuffers, -1); KASSERT(old > 0, ("numfreebuffers dropped to %d", old - 1)); } } /* * waitrunningbufspace() * * runningbufspace is a measure of the amount of I/O currently * running. This routine is used in async-write situations to * prevent creating huge backups of pending writes to a device. * Only asynchronous writes are governed by this function. * * This does NOT turn an async write into a sync write. It waits * for earlier writes to complete and generally returns before the * caller's write has reached the device. */ void waitrunningbufspace(void) { mtx_lock(&rbreqlock); while (runningbufspace > hirunningspace) { runningbufreq = 1; msleep(&runningbufreq, &rbreqlock, PVM, "wdrain", 0); } mtx_unlock(&rbreqlock); } /* * vfs_buf_test_cache: * * Called when a buffer is extended. This function clears the B_CACHE * bit if the newly extended portion of the buffer does not contain * valid data. */ static __inline void vfs_buf_test_cache(struct buf *bp, vm_ooffset_t foff, vm_offset_t off, vm_offset_t size, vm_page_t m) { VM_OBJECT_ASSERT_LOCKED(m->object); if (bp->b_flags & B_CACHE) { int base = (foff + off) & PAGE_MASK; if (vm_page_is_valid(m, base, size) == 0) bp->b_flags &= ~B_CACHE; } } /* Wake up the buffer daemon if necessary */ static __inline void bd_wakeup(void) { mtx_lock(&bdlock); if (bd_request == 0) { bd_request = 1; wakeup(&bd_request); } mtx_unlock(&bdlock); } /* * bd_speedup - speedup the buffer cache flushing code */ void bd_speedup(void) { int needwake; mtx_lock(&bdlock); needwake = 0; if (bd_speedupreq == 0 || bd_request == 0) needwake = 1; bd_speedupreq = 1; bd_request = 1; if (needwake) wakeup(&bd_request); mtx_unlock(&bdlock); } #ifndef NSWBUF_MIN #define NSWBUF_MIN 16 #endif #ifdef __i386__ #define TRANSIENT_DENOM 5 #else #define TRANSIENT_DENOM 10 #endif /* * Calculating buffer cache scaling values and reserve space for buffer * headers. This is called during low level kernel initialization and * may be called more then once. We CANNOT write to the memory area * being reserved at this time. */ caddr_t kern_vfs_bio_buffer_alloc(caddr_t v, long physmem_est) { int tuned_nbuf; long maxbuf, maxbuf_sz, buf_sz, biotmap_sz; /* * physmem_est is in pages. Convert it to kilobytes (assumes * PAGE_SIZE is >= 1K) */ physmem_est = physmem_est * (PAGE_SIZE / 1024); /* * The nominal buffer size (and minimum KVA allocation) is BKVASIZE. * For the first 64MB of ram nominally allocate sufficient buffers to * cover 1/4 of our ram. Beyond the first 64MB allocate additional * buffers to cover 1/10 of our ram over 64MB. When auto-sizing * the buffer cache we limit the eventual kva reservation to * maxbcache bytes. * * factor represents the 1/4 x ram conversion. */ if (nbuf == 0) { int factor = 4 * BKVASIZE / 1024; nbuf = 50; if (physmem_est > 4096) nbuf += min((physmem_est - 4096) / factor, 65536 / factor); if (physmem_est > 65536) nbuf += min((physmem_est - 65536) * 2 / (factor * 5), 32 * 1024 * 1024 / (factor * 5)); if (maxbcache && nbuf > maxbcache / BKVASIZE) nbuf = maxbcache / BKVASIZE; tuned_nbuf = 1; } else tuned_nbuf = 0; /* XXX Avoid unsigned long overflows later on with maxbufspace. */ maxbuf = (LONG_MAX / 3) / BKVASIZE; if (nbuf > maxbuf) { if (!tuned_nbuf) printf("Warning: nbufs lowered from %d to %ld\n", nbuf, maxbuf); nbuf = maxbuf; } /* * Ideal allocation size for the transient bio submap is 10% * of the maximal space buffer map. This roughly corresponds * to the amount of the buffer mapped for typical UFS load. * * Clip the buffer map to reserve space for the transient * BIOs, if its extent is bigger than 90% (80% on i386) of the * maximum buffer map extent on the platform. * * The fall-back to the maxbuf in case of maxbcache unset, * allows to not trim the buffer KVA for the architectures * with ample KVA space. */ if (bio_transient_maxcnt == 0 && unmapped_buf_allowed) { maxbuf_sz = maxbcache != 0 ? maxbcache : maxbuf * BKVASIZE; buf_sz = (long)nbuf * BKVASIZE; if (buf_sz < maxbuf_sz / TRANSIENT_DENOM * (TRANSIENT_DENOM - 1)) { /* * There is more KVA than memory. Do not * adjust buffer map size, and assign the rest * of maxbuf to transient map. */ biotmap_sz = maxbuf_sz - buf_sz; } else { /* * Buffer map spans all KVA we could afford on * this platform. Give 10% (20% on i386) of * the buffer map to the transient bio map. */ biotmap_sz = buf_sz / TRANSIENT_DENOM; buf_sz -= biotmap_sz; } if (biotmap_sz / INT_MAX > MAXPHYS) bio_transient_maxcnt = INT_MAX; else bio_transient_maxcnt = biotmap_sz / MAXPHYS; /* * Artifically limit to 1024 simultaneous in-flight I/Os * using the transient mapping. */ if (bio_transient_maxcnt > 1024) bio_transient_maxcnt = 1024; if (tuned_nbuf) nbuf = buf_sz / BKVASIZE; } /* * swbufs are used as temporary holders for I/O, such as paging I/O. * We have no less then 16 and no more then 256. */ nswbuf = min(nbuf / 4, 256); TUNABLE_INT_FETCH("kern.nswbuf", &nswbuf); if (nswbuf < NSWBUF_MIN) nswbuf = NSWBUF_MIN; /* * Reserve space for the buffer cache buffers */ swbuf = (void *)v; v = (caddr_t)(swbuf + nswbuf); buf = (void *)v; v = (caddr_t)(buf + nbuf); return(v); } /* Initialize the buffer subsystem. Called before use of any buffers. */ void bufinit(void) { struct buf *bp; int i; mtx_init(&bqclean, "bufq clean lock", NULL, MTX_DEF); mtx_init(&bqdirty, "bufq dirty lock", NULL, MTX_DEF); mtx_init(&rbreqlock, "runningbufspace lock", NULL, MTX_DEF); rw_init(&nblock, "needsbuffer lock"); mtx_init(&bdlock, "buffer daemon lock", NULL, MTX_DEF); mtx_init(&bdirtylock, "dirty buf lock", NULL, MTX_DEF); /* next, make a null set of free lists */ for (i = 0; i < BUFFER_QUEUES; i++) TAILQ_INIT(&bufqueues[i]); /* finally, initialize each buffer header and stick on empty q */ for (i = 0; i < nbuf; i++) { bp = &buf[i]; bzero(bp, sizeof *bp); bp->b_flags = B_INVAL | B_INFREECNT; bp->b_rcred = NOCRED; bp->b_wcred = NOCRED; bp->b_qindex = QUEUE_EMPTY; bp->b_xflags = 0; LIST_INIT(&bp->b_dep); BUF_LOCKINIT(bp); TAILQ_INSERT_TAIL(&bufqueues[QUEUE_EMPTY], bp, b_freelist); #ifdef INVARIANTS bq_len[QUEUE_EMPTY]++; #endif } /* * maxbufspace is the absolute maximum amount of buffer space we are * allowed to reserve in KVM and in real terms. The absolute maximum * is nominally used by buf_daemon. hibufspace is the nominal maximum * used by most other processes. The differential is required to * ensure that buf_daemon is able to run when other processes might * be blocked waiting for buffer space. * * maxbufspace is based on BKVASIZE. Allocating buffers larger then * this may result in KVM fragmentation which is not handled optimally * by the system. */ maxbufspace = (long)nbuf * BKVASIZE; hibufspace = lmax(3 * maxbufspace / 4, maxbufspace - MAXBSIZE * 10); lobufspace = hibufspace - MAXBSIZE; /* * Note: The 16 MiB upper limit for hirunningspace was chosen * arbitrarily and may need further tuning. It corresponds to * 128 outstanding write IO requests (if IO size is 128 KiB), * which fits with many RAID controllers' tagged queuing limits. * The lower 1 MiB limit is the historical upper limit for * hirunningspace. */ hirunningspace = lmax(lmin(roundup(hibufspace / 64, MAXBSIZE), 16 * 1024 * 1024), 1024 * 1024); lorunningspace = roundup((hirunningspace * 2) / 3, MAXBSIZE); /* * Limit the amount of malloc memory since it is wired permanently into * the kernel space. Even though this is accounted for in the buffer * allocation, we don't want the malloced region to grow uncontrolled. * The malloc scheme improves memory utilization significantly on average * (small) directories. */ maxbufmallocspace = hibufspace / 20; /* * Reduce the chance of a deadlock occuring by limiting the number * of delayed-write dirty buffers we allow to stack up. */ hidirtybuffers = nbuf / 4 + 20; dirtybufthresh = hidirtybuffers * 9 / 10; numdirtybuffers = 0; /* * To support extreme low-memory systems, make sure hidirtybuffers cannot * eat up all available buffer space. This occurs when our minimum cannot * be met. We try to size hidirtybuffers to 3/4 our buffer space assuming * BKVASIZE'd buffers. */ while ((long)hidirtybuffers * BKVASIZE > 3 * hibufspace / 4) { hidirtybuffers >>= 1; } lodirtybuffers = hidirtybuffers / 2; /* * Try to keep the number of free buffers in the specified range, * and give special processes (e.g. like buf_daemon) access to an * emergency reserve. */ lofreebuffers = nbuf / 18 + 5; hifreebuffers = 2 * lofreebuffers; numfreebuffers = nbuf; bogus_page = vm_page_alloc(NULL, 0, VM_ALLOC_NOOBJ | VM_ALLOC_NORMAL | VM_ALLOC_WIRED); unmapped_buf = (caddr_t)kva_alloc(MAXPHYS); } #ifdef INVARIANTS static inline void vfs_buf_check_mapped(struct buf *bp) { KASSERT((bp->b_flags & B_UNMAPPED) == 0, ("mapped buf %p %x", bp, bp->b_flags)); KASSERT(bp->b_kvabase != unmapped_buf, ("mapped buf: b_kvabase was not updated %p", bp)); KASSERT(bp->b_data != unmapped_buf, ("mapped buf: b_data was not updated %p", bp)); } static inline void vfs_buf_check_unmapped(struct buf *bp) { KASSERT((bp->b_flags & B_UNMAPPED) == B_UNMAPPED, ("unmapped buf %p %x", bp, bp->b_flags)); KASSERT(bp->b_kvabase == unmapped_buf, ("unmapped buf: corrupted b_kvabase %p", bp)); KASSERT(bp->b_data == unmapped_buf, ("unmapped buf: corrupted b_data %p", bp)); } #define BUF_CHECK_MAPPED(bp) vfs_buf_check_mapped(bp) #define BUF_CHECK_UNMAPPED(bp) vfs_buf_check_unmapped(bp) #else #define BUF_CHECK_MAPPED(bp) do {} while (0) #define BUF_CHECK_UNMAPPED(bp) do {} while (0) #endif static void bpmap_qenter(struct buf *bp) { BUF_CHECK_MAPPED(bp); /* * bp->b_data is relative to bp->b_offset, but * bp->b_offset may be offset into the first page. */ bp->b_data = (caddr_t)trunc_page((vm_offset_t)bp->b_data); pmap_qenter((vm_offset_t)bp->b_data, bp->b_pages, bp->b_npages); bp->b_data = (caddr_t)((vm_offset_t)bp->b_data | (vm_offset_t)(bp->b_offset & PAGE_MASK)); } /* * bfreekva() - free the kva allocation for a buffer. * * Since this call frees up buffer space, we call bufspacewakeup(). */ static void bfreekva(struct buf *bp) { if (bp->b_kvasize == 0) return; atomic_add_int(&buffreekvacnt, 1); atomic_subtract_long(&bufspace, bp->b_kvasize); if ((bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); vmem_free(buffer_arena, (vm_offset_t)bp->b_kvabase, bp->b_kvasize); } else { BUF_CHECK_UNMAPPED(bp); if ((bp->b_flags & B_KVAALLOC) != 0) { vmem_free(buffer_arena, (vm_offset_t)bp->b_kvaalloc, bp->b_kvasize); } atomic_subtract_long(&unmapped_bufspace, bp->b_kvasize); bp->b_flags &= ~(B_UNMAPPED | B_KVAALLOC); } bp->b_kvasize = 0; bufspacewakeup(); } /* * binsfree: * * Insert the buffer into the appropriate free list. */ static void binsfree(struct buf *bp, int qindex) { struct mtx *olock, *nlock; BUF_ASSERT_XLOCKED(bp); olock = bqlock(bp->b_qindex); nlock = bqlock(qindex); mtx_lock(olock); /* Handle delayed bremfree() processing. */ if (bp->b_flags & B_REMFREE) bremfreel(bp); if (bp->b_qindex != QUEUE_NONE) panic("binsfree: free buffer onto another queue???"); bp->b_qindex = qindex; if (olock != nlock) { mtx_unlock(olock); mtx_lock(nlock); } if (bp->b_flags & B_AGE) TAILQ_INSERT_HEAD(&bufqueues[bp->b_qindex], bp, b_freelist); else TAILQ_INSERT_TAIL(&bufqueues[bp->b_qindex], bp, b_freelist); #ifdef INVARIANTS bq_len[bp->b_qindex]++; #endif mtx_unlock(nlock); /* * Something we can maybe free or reuse. */ if (bp->b_bufsize && !(bp->b_flags & B_DELWRI)) bufspacewakeup(); if ((bp->b_flags & B_INVAL) || !(bp->b_flags & B_DELWRI)) bufcountadd(bp); } /* * bremfree: * * Mark the buffer for removal from the appropriate free list. * */ void bremfree(struct buf *bp) { CTR3(KTR_BUF, "bremfree(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT((bp->b_flags & B_REMFREE) == 0, ("bremfree: buffer %p already marked for delayed removal.", bp)); KASSERT(bp->b_qindex != QUEUE_NONE, ("bremfree: buffer %p not on a queue.", bp)); BUF_ASSERT_XLOCKED(bp); bp->b_flags |= B_REMFREE; bufcountsub(bp); } /* * bremfreef: * * Force an immediate removal from a free list. Used only in nfs when * it abuses the b_freelist pointer. */ void bremfreef(struct buf *bp) { struct mtx *qlock; qlock = bqlock(bp->b_qindex); mtx_lock(qlock); bremfreel(bp); mtx_unlock(qlock); } /* * bremfreel: * * Removes a buffer from the free list, must be called with the * correct qlock held. */ static void bremfreel(struct buf *bp) { CTR3(KTR_BUF, "bremfreel(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(bp->b_qindex != QUEUE_NONE, ("bremfreel: buffer %p not on a queue.", bp)); BUF_ASSERT_XLOCKED(bp); mtx_assert(bqlock(bp->b_qindex), MA_OWNED); TAILQ_REMOVE(&bufqueues[bp->b_qindex], bp, b_freelist); #ifdef INVARIANTS KASSERT(bq_len[bp->b_qindex] >= 1, ("queue %d underflow", bp->b_qindex)); bq_len[bp->b_qindex]--; #endif bp->b_qindex = QUEUE_NONE; /* * If this was a delayed bremfree() we only need to remove the buffer * from the queue and return the stats are already done. */ if (bp->b_flags & B_REMFREE) { bp->b_flags &= ~B_REMFREE; return; } bufcountsub(bp); } /* * Attempt to initiate asynchronous I/O on read-ahead blocks. We must * clear BIO_ERROR and B_INVAL prior to initiating I/O . If B_CACHE is set, * the buffer is valid and we do not have to do anything. */ void breada(struct vnode * vp, daddr_t * rablkno, int * rabsize, int cnt, struct ucred * cred) { struct buf *rabp; int i; for (i = 0; i < cnt; i++, rablkno++, rabsize++) { if (inmem(vp, *rablkno)) continue; rabp = getblk(vp, *rablkno, *rabsize, 0, 0, 0); if ((rabp->b_flags & B_CACHE) == 0) { if (!TD_IS_IDLETHREAD(curthread)) curthread->td_ru.ru_inblock++; rabp->b_flags |= B_ASYNC; rabp->b_flags &= ~B_INVAL; rabp->b_ioflags &= ~BIO_ERROR; rabp->b_iocmd = BIO_READ; if (rabp->b_rcred == NOCRED && cred != NOCRED) rabp->b_rcred = crhold(cred); vfs_busy_pages(rabp, 0); BUF_KERNPROC(rabp); rabp->b_iooffset = dbtob(rabp->b_blkno); bstrategy(rabp); } else { brelse(rabp); } } } /* * Entry point for bread() and breadn() via #defines in sys/buf.h. * * Get a buffer with the specified data. Look in the cache first. We * must clear BIO_ERROR and B_INVAL prior to initiating I/O. If B_CACHE * is set, the buffer is valid and we do not have to do anything, see * getblk(). Also starts asynchronous I/O on read-ahead blocks. */ int breadn_flags(struct vnode *vp, daddr_t blkno, int size, daddr_t *rablkno, int *rabsize, int cnt, struct ucred *cred, int flags, struct buf **bpp) { struct buf *bp; int rv = 0, readwait = 0; CTR3(KTR_BUF, "breadn(%p, %jd, %d)", vp, blkno, size); /* * Can only return NULL if GB_LOCK_NOWAIT flag is specified. */ *bpp = bp = getblk(vp, blkno, size, 0, 0, flags); if (bp == NULL) return (EBUSY); /* if not found in cache, do some I/O */ if ((bp->b_flags & B_CACHE) == 0) { if (!TD_IS_IDLETHREAD(curthread)) curthread->td_ru.ru_inblock++; bp->b_iocmd = BIO_READ; bp->b_flags &= ~B_INVAL; bp->b_ioflags &= ~BIO_ERROR; if (bp->b_rcred == NOCRED && cred != NOCRED) bp->b_rcred = crhold(cred); vfs_busy_pages(bp, 0); bp->b_iooffset = dbtob(bp->b_blkno); bstrategy(bp); ++readwait; } breada(vp, rablkno, rabsize, cnt, cred); if (readwait) { rv = bufwait(bp); } return (rv); } /* * Write, release buffer on completion. (Done by iodone * if async). Do not bother writing anything if the buffer * is invalid. * * Note that we set B_CACHE here, indicating that buffer is * fully valid and thus cacheable. This is true even of NFS * now so we set it generally. This could be set either here * or in biodone() since the I/O is synchronous. We put it * here. */ int bufwrite(struct buf *bp) { int oldflags; struct vnode *vp; long space; int vp_md; CTR3(KTR_BUF, "bufwrite(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); if (bp->b_flags & B_INVAL) { brelse(bp); return (0); } if (bp->b_flags & B_BARRIER) barrierwrites++; oldflags = bp->b_flags; BUF_ASSERT_HELD(bp); if (bp->b_pin_count > 0) bunpin_wait(bp); KASSERT(!(bp->b_vflags & BV_BKGRDINPROG), ("FFS background buffer should not get here %p", bp)); vp = bp->b_vp; if (vp) vp_md = vp->v_vflag & VV_MD; else vp_md = 0; /* * Mark the buffer clean. Increment the bufobj write count * before bundirty() call, to prevent other thread from seeing * empty dirty list and zero counter for writes in progress, * falsely indicating that the bufobj is clean. */ bufobj_wref(bp->b_bufobj); bundirty(bp); bp->b_flags &= ~B_DONE; bp->b_ioflags &= ~BIO_ERROR; bp->b_flags |= B_CACHE; bp->b_iocmd = BIO_WRITE; vfs_busy_pages(bp, 1); /* * Normal bwrites pipeline writes */ bp->b_runningbufspace = bp->b_bufsize; space = atomic_fetchadd_long(&runningbufspace, bp->b_runningbufspace); if (!TD_IS_IDLETHREAD(curthread)) curthread->td_ru.ru_oublock++; if (oldflags & B_ASYNC) BUF_KERNPROC(bp); bp->b_iooffset = dbtob(bp->b_blkno); bstrategy(bp); if ((oldflags & B_ASYNC) == 0) { int rtval = bufwait(bp); brelse(bp); return (rtval); } else if (space > hirunningspace) { /* * don't allow the async write to saturate the I/O * system. We will not deadlock here because * we are blocking waiting for I/O that is already in-progress * to complete. We do not block here if it is the update * or syncer daemon trying to clean up as that can lead * to deadlock. */ if ((curthread->td_pflags & TDP_NORUNNINGBUF) == 0 && !vp_md) waitrunningbufspace(); } return (0); } void bufbdflush(struct bufobj *bo, struct buf *bp) { struct buf *nbp; if (bo->bo_dirty.bv_cnt > dirtybufthresh + 10) { (void) VOP_FSYNC(bp->b_vp, MNT_NOWAIT, curthread); altbufferflushes++; } else if (bo->bo_dirty.bv_cnt > dirtybufthresh) { BO_LOCK(bo); /* * Try to find a buffer to flush. */ TAILQ_FOREACH(nbp, &bo->bo_dirty.bv_hd, b_bobufs) { if ((nbp->b_vflags & BV_BKGRDINPROG) || BUF_LOCK(nbp, LK_EXCLUSIVE | LK_NOWAIT, NULL)) continue; if (bp == nbp) panic("bdwrite: found ourselves"); BO_UNLOCK(bo); /* Don't countdeps with the bo lock held. */ if (buf_countdeps(nbp, 0)) { BO_LOCK(bo); BUF_UNLOCK(nbp); continue; } if (nbp->b_flags & B_CLUSTEROK) { vfs_bio_awrite(nbp); } else { bremfree(nbp); bawrite(nbp); } dirtybufferflushes++; break; } if (nbp == NULL) BO_UNLOCK(bo); } } /* * Delayed write. (Buffer is marked dirty). Do not bother writing * anything if the buffer is marked invalid. * * Note that since the buffer must be completely valid, we can safely * set B_CACHE. In fact, we have to set B_CACHE here rather then in * biodone() in order to prevent getblk from writing the buffer * out synchronously. */ void bdwrite(struct buf *bp) { struct thread *td = curthread; struct vnode *vp; struct bufobj *bo; CTR3(KTR_BUF, "bdwrite(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(bp->b_bufobj != NULL, ("No b_bufobj %p", bp)); KASSERT((bp->b_flags & B_BARRIER) == 0, ("Barrier request in delayed write %p", bp)); BUF_ASSERT_HELD(bp); if (bp->b_flags & B_INVAL) { brelse(bp); return; } /* * If we have too many dirty buffers, don't create any more. * If we are wildly over our limit, then force a complete * cleanup. Otherwise, just keep the situation from getting * out of control. Note that we have to avoid a recursive * disaster and not try to clean up after our own cleanup! */ vp = bp->b_vp; bo = bp->b_bufobj; if ((td->td_pflags & (TDP_COWINPROGRESS|TDP_INBDFLUSH)) == 0) { td->td_pflags |= TDP_INBDFLUSH; BO_BDFLUSH(bo, bp); td->td_pflags &= ~TDP_INBDFLUSH; } else recursiveflushes++; bdirty(bp); /* * Set B_CACHE, indicating that the buffer is fully valid. This is * true even of NFS now. */ bp->b_flags |= B_CACHE; /* * This bmap keeps the system from needing to do the bmap later, * perhaps when the system is attempting to do a sync. Since it * is likely that the indirect block -- or whatever other datastructure * that the filesystem needs is still in memory now, it is a good * thing to do this. Note also, that if the pageout daemon is * requesting a sync -- there might not be enough memory to do * the bmap then... So, this is important to do. */ if (vp->v_type != VCHR && bp->b_lblkno == bp->b_blkno) { VOP_BMAP(vp, bp->b_lblkno, NULL, &bp->b_blkno, NULL, NULL); } /* * Set the *dirty* buffer range based upon the VM system dirty * pages. * * Mark the buffer pages as clean. We need to do this here to * satisfy the vnode_pager and the pageout daemon, so that it * thinks that the pages have been "cleaned". Note that since * the pages are in a delayed write buffer -- the VFS layer * "will" see that the pages get written out on the next sync, * or perhaps the cluster will be completed. */ vfs_clean_pages_dirty_buf(bp); bqrelse(bp); /* * note: we cannot initiate I/O from a bdwrite even if we wanted to, * due to the softdep code. */ } /* * bdirty: * * Turn buffer into delayed write request. We must clear BIO_READ and * B_RELBUF, and we must set B_DELWRI. We reassign the buffer to * itself to properly update it in the dirty/clean lists. We mark it * B_DONE to ensure that any asynchronization of the buffer properly * clears B_DONE ( else a panic will occur later ). * * bdirty() is kinda like bdwrite() - we have to clear B_INVAL which * might have been set pre-getblk(). Unlike bwrite/bdwrite, bdirty() * should only be called if the buffer is known-good. * * Since the buffer is not on a queue, we do not update the numfreebuffers * count. * * The buffer must be on QUEUE_NONE. */ void bdirty(struct buf *bp) { CTR3(KTR_BUF, "bdirty(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(bp->b_bufobj != NULL, ("No b_bufobj %p", bp)); KASSERT(bp->b_flags & B_REMFREE || bp->b_qindex == QUEUE_NONE, ("bdirty: buffer %p still on queue %d", bp, bp->b_qindex)); BUF_ASSERT_HELD(bp); bp->b_flags &= ~(B_RELBUF); bp->b_iocmd = BIO_WRITE; if ((bp->b_flags & B_DELWRI) == 0) { bp->b_flags |= /* XXX B_DONE | */ B_DELWRI; reassignbuf(bp); bdirtyadd(); } } /* * bundirty: * * Clear B_DELWRI for buffer. * * Since the buffer is not on a queue, we do not update the numfreebuffers * count. * * The buffer must be on QUEUE_NONE. */ void bundirty(struct buf *bp) { CTR3(KTR_BUF, "bundirty(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(bp->b_bufobj != NULL, ("No b_bufobj %p", bp)); KASSERT(bp->b_flags & B_REMFREE || bp->b_qindex == QUEUE_NONE, ("bundirty: buffer %p still on queue %d", bp, bp->b_qindex)); BUF_ASSERT_HELD(bp); if (bp->b_flags & B_DELWRI) { bp->b_flags &= ~B_DELWRI; reassignbuf(bp); bdirtysub(); } /* * Since it is now being written, we can clear its deferred write flag. */ bp->b_flags &= ~B_DEFERRED; } /* * bawrite: * * Asynchronous write. Start output on a buffer, but do not wait for * it to complete. The buffer is released when the output completes. * * bwrite() ( or the VOP routine anyway ) is responsible for handling * B_INVAL buffers. Not us. */ void bawrite(struct buf *bp) { bp->b_flags |= B_ASYNC; (void) bwrite(bp); } /* * babarrierwrite: * * Asynchronous barrier write. Start output on a buffer, but do not * wait for it to complete. Place a write barrier after this write so * that this buffer and all buffers written before it are committed to * the disk before any buffers written after this write are committed * to the disk. The buffer is released when the output completes. */ void babarrierwrite(struct buf *bp) { bp->b_flags |= B_ASYNC | B_BARRIER; (void) bwrite(bp); } /* * bbarrierwrite: * * Synchronous barrier write. Start output on a buffer and wait for * it to complete. Place a write barrier after this write so that * this buffer and all buffers written before it are committed to * the disk before any buffers written after this write are committed * to the disk. The buffer is released when the output completes. */ int bbarrierwrite(struct buf *bp) { bp->b_flags |= B_BARRIER; return (bwrite(bp)); } /* * bwillwrite: * * Called prior to the locking of any vnodes when we are expecting to * write. We do not want to starve the buffer cache with too many * dirty buffers so we block here. By blocking prior to the locking * of any vnodes we attempt to avoid the situation where a locked vnode * prevents the various system daemons from flushing related buffers. */ void bwillwrite(void) { if (numdirtybuffers >= hidirtybuffers) { mtx_lock(&bdirtylock); while (numdirtybuffers >= hidirtybuffers) { bdirtywait = 1; msleep(&bdirtywait, &bdirtylock, (PRIBIO + 4), "flswai", 0); } mtx_unlock(&bdirtylock); } } /* * Return true if we have too many dirty buffers. */ int buf_dirty_count_severe(void) { return(numdirtybuffers >= hidirtybuffers); } static __noinline int buf_vm_page_count_severe(void) { KFAIL_POINT_CODE(DEBUG_FP, buf_pressure, return 1); return vm_page_count_severe(); } /* * brelse: * * Release a busy buffer and, if requested, free its resources. The * buffer will be stashed in the appropriate bufqueue[] allowing it * to be accessed later as a cache entity or reused for other purposes. */ void brelse(struct buf *bp) { int qindex; CTR3(KTR_BUF, "brelse(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(!(bp->b_flags & (B_CLUSTER|B_PAGING)), ("brelse: inappropriate B_PAGING or B_CLUSTER bp %p", bp)); if (BUF_LOCKRECURSED(bp)) { /* * Do not process, in particular, do not handle the * B_INVAL/B_RELBUF and do not release to free list. */ BUF_UNLOCK(bp); return; } if (bp->b_flags & B_MANAGED) { bqrelse(bp); return; } if (bp->b_iocmd == BIO_WRITE && (bp->b_ioflags & BIO_ERROR) && bp->b_error == EIO && !(bp->b_flags & B_INVAL)) { /* * Failed write, redirty. Must clear BIO_ERROR to prevent * pages from being scrapped. If the error is anything * other than an I/O error (EIO), assume that retrying * is futile. */ bp->b_ioflags &= ~BIO_ERROR; bdirty(bp); } else if ((bp->b_flags & (B_NOCACHE | B_INVAL)) || (bp->b_ioflags & BIO_ERROR) || (bp->b_bufsize <= 0)) { /* * Either a failed I/O or we were asked to free or not * cache the buffer. */ bp->b_flags |= B_INVAL; if (!LIST_EMPTY(&bp->b_dep)) buf_deallocate(bp); if (bp->b_flags & B_DELWRI) bdirtysub(); bp->b_flags &= ~(B_DELWRI | B_CACHE); if ((bp->b_flags & B_VMIO) == 0) { if (bp->b_bufsize) allocbuf(bp, 0); if (bp->b_vp) brelvp(bp); } } /* * We must clear B_RELBUF if B_DELWRI is set. If vfs_vmio_release() * is called with B_DELWRI set, the underlying pages may wind up * getting freed causing a previous write (bdwrite()) to get 'lost' * because pages associated with a B_DELWRI bp are marked clean. * * We still allow the B_INVAL case to call vfs_vmio_release(), even * if B_DELWRI is set. * * If B_DELWRI is not set we may have to set B_RELBUF if we are low * on pages to return pages to the VM page queues. */ if (bp->b_flags & B_DELWRI) bp->b_flags &= ~B_RELBUF; else if (buf_vm_page_count_severe()) { /* * BKGRDINPROG can only be set with the buf and bufobj * locks both held. We tolerate a race to clear it here. */ if (!(bp->b_vflags & BV_BKGRDINPROG)) bp->b_flags |= B_RELBUF; } /* * VMIO buffer rundown. It is not very necessary to keep a VMIO buffer * constituted, not even NFS buffers now. Two flags effect this. If * B_INVAL, the struct buf is invalidated but the VM object is kept * around ( i.e. so it is trivial to reconstitute the buffer later ). * * If BIO_ERROR or B_NOCACHE is set, pages in the VM object will be * invalidated. BIO_ERROR cannot be set for a failed write unless the * buffer is also B_INVAL because it hits the re-dirtying code above. * * Normally we can do this whether a buffer is B_DELWRI or not. If * the buffer is an NFS buffer, it is tracking piecemeal writes or * the commit state and we cannot afford to lose the buffer. If the * buffer has a background write in progress, we need to keep it * around to prevent it from being reconstituted and starting a second * background write. */ if ((bp->b_flags & B_VMIO) && !(bp->b_vp->v_mount != NULL && (bp->b_vp->v_mount->mnt_vfc->vfc_flags & VFCF_NETWORK) != 0 && !vn_isdisk(bp->b_vp, NULL) && (bp->b_flags & B_DELWRI)) ) { int i, j, resid; vm_page_t m; off_t foff; vm_pindex_t poff; vm_object_t obj; obj = bp->b_bufobj->bo_object; /* * Get the base offset and length of the buffer. Note that * in the VMIO case if the buffer block size is not * page-aligned then b_data pointer may not be page-aligned. * But our b_pages[] array *IS* page aligned. * * block sizes less then DEV_BSIZE (usually 512) are not * supported due to the page granularity bits (m->valid, * m->dirty, etc...). * * See man buf(9) for more information */ resid = bp->b_bufsize; foff = bp->b_offset; for (i = 0; i < bp->b_npages; i++) { int had_bogus = 0; m = bp->b_pages[i]; /* * If we hit a bogus page, fixup *all* the bogus pages * now. */ if (m == bogus_page) { poff = OFF_TO_IDX(bp->b_offset); had_bogus = 1; VM_OBJECT_RLOCK(obj); for (j = i; j < bp->b_npages; j++) { vm_page_t mtmp; mtmp = bp->b_pages[j]; if (mtmp == bogus_page) { mtmp = vm_page_lookup(obj, poff + j); if (!mtmp) { panic("brelse: page missing\n"); } bp->b_pages[j] = mtmp; } } VM_OBJECT_RUNLOCK(obj); if ((bp->b_flags & (B_INVAL | B_UNMAPPED)) == 0) { BUF_CHECK_MAPPED(bp); pmap_qenter( trunc_page((vm_offset_t)bp->b_data), bp->b_pages, bp->b_npages); } m = bp->b_pages[i]; } if ((bp->b_flags & B_NOCACHE) || (bp->b_ioflags & BIO_ERROR && bp->b_iocmd == BIO_READ)) { int poffset = foff & PAGE_MASK; int presid = resid > (PAGE_SIZE - poffset) ? (PAGE_SIZE - poffset) : resid; KASSERT(presid >= 0, ("brelse: extra page")); VM_OBJECT_WLOCK(obj); while (vm_page_xbusied(m)) { vm_page_lock(m); VM_OBJECT_WUNLOCK(obj); vm_page_busy_sleep(m, "mbncsh"); VM_OBJECT_WLOCK(obj); } if (pmap_page_wired_mappings(m) == 0) vm_page_set_invalid(m, poffset, presid); VM_OBJECT_WUNLOCK(obj); if (had_bogus) printf("avoided corruption bug in bogus_page/brelse code\n"); } resid -= PAGE_SIZE - (foff & PAGE_MASK); foff = (foff + PAGE_SIZE) & ~(off_t)PAGE_MASK; } if (bp->b_flags & (B_INVAL | B_RELBUF)) vfs_vmio_release(bp); } else if (bp->b_flags & B_VMIO) { if (bp->b_flags & (B_INVAL | B_RELBUF)) { vfs_vmio_release(bp); } } else if ((bp->b_flags & (B_INVAL | B_RELBUF)) != 0) { if (bp->b_bufsize != 0) allocbuf(bp, 0); if (bp->b_vp != NULL) brelvp(bp); } /* * If the buffer has junk contents signal it and eventually * clean up B_DELWRI and diassociate the vnode so that gbincore() * doesn't find it. */ if (bp->b_bufsize == 0 || (bp->b_ioflags & BIO_ERROR) != 0 || (bp->b_flags & (B_INVAL | B_NOCACHE | B_RELBUF)) != 0) bp->b_flags |= B_INVAL; if (bp->b_flags & B_INVAL) { if (bp->b_flags & B_DELWRI) bundirty(bp); if (bp->b_vp) brelvp(bp); } /* buffers with no memory */ if (bp->b_bufsize == 0) { bp->b_xflags &= ~(BX_BKGRDWRITE | BX_ALTDATA); if (bp->b_vflags & BV_BKGRDINPROG) panic("losing buffer 1"); if (bp->b_kvasize) qindex = QUEUE_EMPTYKVA; else qindex = QUEUE_EMPTY; bp->b_flags |= B_AGE; /* buffers with junk contents */ } else if (bp->b_flags & (B_INVAL | B_NOCACHE | B_RELBUF) || (bp->b_ioflags & BIO_ERROR)) { bp->b_xflags &= ~(BX_BKGRDWRITE | BX_ALTDATA); if (bp->b_vflags & BV_BKGRDINPROG) panic("losing buffer 2"); qindex = QUEUE_CLEAN; bp->b_flags |= B_AGE; /* remaining buffers */ } else if (bp->b_flags & B_DELWRI) qindex = QUEUE_DIRTY; else qindex = QUEUE_CLEAN; binsfree(bp, qindex); bp->b_flags &= ~(B_ASYNC | B_NOCACHE | B_AGE | B_RELBUF | B_DIRECT); if ((bp->b_flags & B_DELWRI) == 0 && (bp->b_xflags & BX_VNDIRTY)) panic("brelse: not dirty"); /* unlock */ BUF_UNLOCK(bp); } /* * Release a buffer back to the appropriate queue but do not try to free * it. The buffer is expected to be used again soon. * * bqrelse() is used by bdwrite() to requeue a delayed write, and used by * biodone() to requeue an async I/O on completion. It is also used when * known good buffers need to be requeued but we think we may need the data * again soon. * * XXX we should be able to leave the B_RELBUF hint set on completion. */ void bqrelse(struct buf *bp) { int qindex; CTR3(KTR_BUF, "bqrelse(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(!(bp->b_flags & (B_CLUSTER|B_PAGING)), ("bqrelse: inappropriate B_PAGING or B_CLUSTER bp %p", bp)); if (BUF_LOCKRECURSED(bp)) { /* do not release to free list */ BUF_UNLOCK(bp); return; } bp->b_flags &= ~(B_ASYNC | B_NOCACHE | B_AGE | B_RELBUF); if (bp->b_flags & B_MANAGED) { if (bp->b_flags & B_REMFREE) bremfreef(bp); goto out; } /* buffers with stale but valid contents */ if (bp->b_flags & B_DELWRI) { qindex = QUEUE_DIRTY; } else { if ((bp->b_flags & B_DELWRI) == 0 && (bp->b_xflags & BX_VNDIRTY)) panic("bqrelse: not dirty"); /* * BKGRDINPROG can only be set with the buf and bufobj * locks both held. We tolerate a race to clear it here. */ if (buf_vm_page_count_severe() && (bp->b_vflags & BV_BKGRDINPROG) == 0) { /* * We are too low on memory, we have to try to free * the buffer (most importantly: the wired pages * making up its backing store) *now*. */ brelse(bp); return; } qindex = QUEUE_CLEAN; } binsfree(bp, qindex); out: /* unlock */ BUF_UNLOCK(bp); } /* Give pages used by the bp back to the VM system (where possible) */ static void vfs_vmio_release(struct buf *bp) { - int i; + vm_object_t obj; vm_page_t m; + int i; if ((bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); pmap_qremove(trunc_page((vm_offset_t)bp->b_data), bp->b_npages); } else BUF_CHECK_UNMAPPED(bp); - VM_OBJECT_WLOCK(bp->b_bufobj->bo_object); + obj = bp->b_bufobj->bo_object; + if (obj != NULL) + VM_OBJECT_WLOCK(obj); for (i = 0; i < bp->b_npages; i++) { m = bp->b_pages[i]; bp->b_pages[i] = NULL; /* * In order to keep page LRU ordering consistent, put * everything on the inactive queue. */ vm_page_lock(m); vm_page_unwire(m, PQ_INACTIVE); /* * Might as well free the page if we can and it has * no valid data. We also free the page if the * buffer was used for direct I/O */ if ((bp->b_flags & B_ASYNC) == 0 && !m->valid) { if (m->wire_count == 0 && !vm_page_busied(m)) vm_page_free(m); } else if (bp->b_flags & B_DIRECT) vm_page_try_to_free(m); else if (buf_vm_page_count_severe()) vm_page_try_to_cache(m); vm_page_unlock(m); } - VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object); + if (obj != NULL) + VM_OBJECT_WUNLOCK(obj); if (bp->b_bufsize) { bufspacewakeup(); bp->b_bufsize = 0; } bp->b_npages = 0; bp->b_flags &= ~B_VMIO; if (bp->b_vp) brelvp(bp); } /* * Check to see if a block at a particular lbn is available for a clustered * write. */ static int vfs_bio_clcheck(struct vnode *vp, int size, daddr_t lblkno, daddr_t blkno) { struct buf *bpa; int match; match = 0; /* If the buf isn't in core skip it */ if ((bpa = gbincore(&vp->v_bufobj, lblkno)) == NULL) return (0); /* If the buf is busy we don't want to wait for it */ if (BUF_LOCK(bpa, LK_EXCLUSIVE | LK_NOWAIT, NULL) != 0) return (0); /* Only cluster with valid clusterable delayed write buffers */ if ((bpa->b_flags & (B_DELWRI | B_CLUSTEROK | B_INVAL)) != (B_DELWRI | B_CLUSTEROK)) goto done; if (bpa->b_bufsize != size) goto done; /* * Check to see if it is in the expected place on disk and that the * block has been mapped. */ if ((bpa->b_blkno != bpa->b_lblkno) && (bpa->b_blkno == blkno)) match = 1; done: BUF_UNLOCK(bpa); return (match); } /* * vfs_bio_awrite: * * Implement clustered async writes for clearing out B_DELWRI buffers. * This is much better then the old way of writing only one buffer at * a time. Note that we may not be presented with the buffers in the * correct order, so we search for the cluster in both directions. */ int vfs_bio_awrite(struct buf *bp) { struct bufobj *bo; int i; int j; daddr_t lblkno = bp->b_lblkno; struct vnode *vp = bp->b_vp; int ncl; int nwritten; int size; int maxcl; int gbflags; bo = &vp->v_bufobj; gbflags = (bp->b_flags & B_UNMAPPED) != 0 ? GB_UNMAPPED : 0; /* * right now we support clustered writing only to regular files. If * we find a clusterable block we could be in the middle of a cluster * rather then at the beginning. */ if ((vp->v_type == VREG) && (vp->v_mount != 0) && /* Only on nodes that have the size info */ (bp->b_flags & (B_CLUSTEROK | B_INVAL)) == B_CLUSTEROK) { size = vp->v_mount->mnt_stat.f_iosize; maxcl = MAXPHYS / size; BO_RLOCK(bo); for (i = 1; i < maxcl; i++) if (vfs_bio_clcheck(vp, size, lblkno + i, bp->b_blkno + ((i * size) >> DEV_BSHIFT)) == 0) break; for (j = 1; i + j <= maxcl && j <= lblkno; j++) if (vfs_bio_clcheck(vp, size, lblkno - j, bp->b_blkno - ((j * size) >> DEV_BSHIFT)) == 0) break; BO_RUNLOCK(bo); --j; ncl = i + j; /* * this is a possible cluster write */ if (ncl != 1) { BUF_UNLOCK(bp); nwritten = cluster_wbuild(vp, size, lblkno - j, ncl, gbflags); return (nwritten); } } bremfree(bp); bp->b_flags |= B_ASYNC; /* * default (old) behavior, writing out only one block * * XXX returns b_bufsize instead of b_bcount for nwritten? */ nwritten = bp->b_bufsize; (void) bwrite(bp); return (nwritten); } static void setbufkva(struct buf *bp, vm_offset_t addr, int maxsize, int gbflags) { KASSERT((bp->b_flags & (B_UNMAPPED | B_KVAALLOC)) == 0 && bp->b_kvasize == 0, ("call bfreekva(%p)", bp)); if ((gbflags & GB_UNMAPPED) == 0) { bp->b_kvabase = (caddr_t)addr; } else if ((gbflags & GB_KVAALLOC) != 0) { KASSERT((gbflags & GB_UNMAPPED) != 0, ("GB_KVAALLOC without GB_UNMAPPED")); bp->b_kvaalloc = (caddr_t)addr; bp->b_flags |= B_UNMAPPED | B_KVAALLOC; atomic_add_long(&unmapped_bufspace, bp->b_kvasize); } bp->b_kvasize = maxsize; } /* * Allocate the buffer KVA and set b_kvasize. Also set b_kvabase if * needed. */ static int allocbufkva(struct buf *bp, int maxsize, int gbflags) { vm_offset_t addr; bfreekva(bp); addr = 0; if (vmem_alloc(buffer_arena, maxsize, M_BESTFIT | M_NOWAIT, &addr)) { /* * Buffer map is too fragmented. Request the caller * to defragment the map. */ atomic_add_int(&bufdefragcnt, 1); return (1); } setbufkva(bp, addr, maxsize, gbflags); atomic_add_long(&bufspace, bp->b_kvasize); return (0); } /* * Ask the bufdaemon for help, or act as bufdaemon itself, when a * locked vnode is supplied. */ static void getnewbuf_bufd_help(struct vnode *vp, int gbflags, int slpflag, int slptimeo, int defrag) { struct thread *td; char *waitmsg; int cnt, error, flags, norunbuf, wait; mtx_assert(&bqclean, MA_OWNED); if (defrag) { flags = VFS_BIO_NEED_BUFSPACE; waitmsg = "nbufkv"; } else if (bufspace >= hibufspace) { waitmsg = "nbufbs"; flags = VFS_BIO_NEED_BUFSPACE; } else { waitmsg = "newbuf"; flags = VFS_BIO_NEED_ANY; } atomic_set_int(&needsbuffer, flags); mtx_unlock(&bqclean); bd_speedup(); /* heeeelp */ if ((gbflags & GB_NOWAIT_BD) != 0) return; td = curthread; cnt = 0; wait = MNT_NOWAIT; rw_wlock(&nblock); while ((needsbuffer & flags) != 0) { if (vp != NULL && vp->v_type != VCHR && (td->td_pflags & TDP_BUFNEED) == 0) { rw_wunlock(&nblock); /* * getblk() is called with a vnode locked, and * some majority of the dirty buffers may as * well belong to the vnode. Flushing the * buffers there would make a progress that * cannot be achieved by the buf_daemon, that * cannot lock the vnode. */ if (cnt++ > 2) wait = MNT_WAIT; ASSERT_VOP_LOCKED(vp, "bufd_helper"); error = VOP_ISLOCKED(vp) == LK_EXCLUSIVE ? 0 : vn_lock(vp, LK_TRYUPGRADE); if (error == 0) { /* play bufdaemon */ norunbuf = curthread_pflags_set(TDP_BUFNEED | TDP_NORUNNINGBUF); VOP_FSYNC(vp, wait, td); atomic_add_long(¬bufdflushes, 1); curthread_pflags_restore(norunbuf); } rw_wlock(&nblock); if ((needsbuffer & flags) == 0) break; } error = rw_sleep(__DEVOLATILE(void *, &needsbuffer), &nblock, (PRIBIO + 4) | slpflag, waitmsg, slptimeo); if (error != 0) break; } rw_wunlock(&nblock); } static void getnewbuf_reuse_bp(struct buf *bp, int qindex) { CTR6(KTR_BUF, "getnewbuf(%p) vp %p flags %X kvasize %d bufsize %d " "queue %d (recycling)", bp, bp->b_vp, bp->b_flags, bp->b_kvasize, bp->b_bufsize, qindex); mtx_assert(&bqclean, MA_NOTOWNED); /* * Note: we no longer distinguish between VMIO and non-VMIO * buffers. */ KASSERT((bp->b_flags & B_DELWRI) == 0, ("delwri buffer %p found in queue %d", bp, qindex)); if (qindex == QUEUE_CLEAN) { if (bp->b_flags & B_VMIO) { bp->b_flags &= ~B_ASYNC; vfs_vmio_release(bp); } if (bp->b_vp != NULL) brelvp(bp); } /* * Get the rest of the buffer freed up. b_kva* is still valid * after this operation. */ if (bp->b_rcred != NOCRED) { crfree(bp->b_rcred); bp->b_rcred = NOCRED; } if (bp->b_wcred != NOCRED) { crfree(bp->b_wcred); bp->b_wcred = NOCRED; } if (!LIST_EMPTY(&bp->b_dep)) buf_deallocate(bp); if (bp->b_vflags & BV_BKGRDINPROG) panic("losing buffer 3"); KASSERT(bp->b_vp == NULL, ("bp: %p still has vnode %p. qindex: %d", bp, bp->b_vp, qindex)); KASSERT((bp->b_xflags & (BX_VNCLEAN|BX_VNDIRTY)) == 0, ("bp: %p still on a buffer list. xflags %X", bp, bp->b_xflags)); if (bp->b_bufsize) allocbuf(bp, 0); bp->b_flags &= B_UNMAPPED | B_KVAALLOC; bp->b_ioflags = 0; bp->b_xflags = 0; KASSERT((bp->b_flags & B_INFREECNT) == 0, ("buf %p still counted as free?", bp)); bp->b_vflags = 0; bp->b_vp = NULL; bp->b_blkno = bp->b_lblkno = 0; bp->b_offset = NOOFFSET; bp->b_iodone = 0; bp->b_error = 0; bp->b_resid = 0; bp->b_bcount = 0; bp->b_npages = 0; bp->b_dirtyoff = bp->b_dirtyend = 0; bp->b_bufobj = NULL; bp->b_pin_count = 0; bp->b_fsprivate1 = NULL; bp->b_fsprivate2 = NULL; bp->b_fsprivate3 = NULL; LIST_INIT(&bp->b_dep); } static int flushingbufs; static struct buf * getnewbuf_scan(int maxsize, int defrag, int unmapped, int metadata) { struct buf *bp, *nbp; int nqindex, qindex, pass; KASSERT(!unmapped || !defrag, ("both unmapped and defrag")); pass = 1; restart: atomic_add_int(&getnewbufrestarts, 1); /* * Setup for scan. If we do not have enough free buffers, * we setup a degenerate case that immediately fails. Note * that if we are specially marked process, we are allowed to * dip into our reserves. * * The scanning sequence is nominally: EMPTY->EMPTYKVA->CLEAN * for the allocation of the mapped buffer. For unmapped, the * easiest is to start with EMPTY outright. * * We start with EMPTYKVA. If the list is empty we backup to EMPTY. * However, there are a number of cases (defragging, reusing, ...) * where we cannot backup. */ nbp = NULL; mtx_lock(&bqclean); if (!defrag && unmapped) { nqindex = QUEUE_EMPTY; nbp = TAILQ_FIRST(&bufqueues[QUEUE_EMPTY]); } if (nbp == NULL) { nqindex = QUEUE_EMPTYKVA; nbp = TAILQ_FIRST(&bufqueues[QUEUE_EMPTYKVA]); } /* * If no EMPTYKVA buffers and we are either defragging or * reusing, locate a CLEAN buffer to free or reuse. If * bufspace useage is low skip this step so we can allocate a * new buffer. */ if (nbp == NULL && (defrag || bufspace >= lobufspace)) { nqindex = QUEUE_CLEAN; nbp = TAILQ_FIRST(&bufqueues[QUEUE_CLEAN]); } /* * If we could not find or were not allowed to reuse a CLEAN * buffer, check to see if it is ok to use an EMPTY buffer. * We can only use an EMPTY buffer if allocating its KVA would * not otherwise run us out of buffer space. No KVA is needed * for the unmapped allocation. */ if (nbp == NULL && defrag == 0 && (bufspace + maxsize < hibufspace || metadata)) { nqindex = QUEUE_EMPTY; nbp = TAILQ_FIRST(&bufqueues[QUEUE_EMPTY]); } /* * All available buffers might be clean, retry ignoring the * lobufspace as the last resort. */ if (nbp == NULL && !TAILQ_EMPTY(&bufqueues[QUEUE_CLEAN])) { nqindex = QUEUE_CLEAN; nbp = TAILQ_FIRST(&bufqueues[QUEUE_CLEAN]); } /* * Run scan, possibly freeing data and/or kva mappings on the fly * depending. */ while ((bp = nbp) != NULL) { qindex = nqindex; /* * Calculate next bp (we can only use it if we do not * block or do other fancy things). */ if ((nbp = TAILQ_NEXT(bp, b_freelist)) == NULL) { switch (qindex) { case QUEUE_EMPTY: nqindex = QUEUE_EMPTYKVA; nbp = TAILQ_FIRST(&bufqueues[QUEUE_EMPTYKVA]); if (nbp != NULL) break; /* FALLTHROUGH */ case QUEUE_EMPTYKVA: nqindex = QUEUE_CLEAN; nbp = TAILQ_FIRST(&bufqueues[QUEUE_CLEAN]); if (nbp != NULL) break; /* FALLTHROUGH */ case QUEUE_CLEAN: if (metadata && pass == 1) { pass = 2; nqindex = QUEUE_EMPTY; nbp = TAILQ_FIRST( &bufqueues[QUEUE_EMPTY]); } /* * nbp is NULL. */ break; } } /* * If we are defragging then we need a buffer with * b_kvasize != 0. XXX this situation should no longer * occur, if defrag is non-zero the buffer's b_kvasize * should also be non-zero at this point. XXX */ if (defrag && bp->b_kvasize == 0) { printf("Warning: defrag empty buffer %p\n", bp); continue; } /* * Start freeing the bp. This is somewhat involved. nbp * remains valid only for QUEUE_EMPTY[KVA] bp's. */ if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_NOWAIT, NULL) != 0) continue; /* * BKGRDINPROG can only be set with the buf and bufobj * locks both held. We tolerate a race to clear it here. */ if (bp->b_vflags & BV_BKGRDINPROG) { BUF_UNLOCK(bp); continue; } KASSERT(bp->b_qindex == qindex, ("getnewbuf: inconsistent queue %d bp %p", qindex, bp)); bremfreel(bp); mtx_unlock(&bqclean); /* * NOTE: nbp is now entirely invalid. We can only restart * the scan from this point on. */ getnewbuf_reuse_bp(bp, qindex); mtx_assert(&bqclean, MA_NOTOWNED); /* * If we are defragging then free the buffer. */ if (defrag) { bp->b_flags |= B_INVAL; bfreekva(bp); brelse(bp); defrag = 0; goto restart; } /* * Notify any waiters for the buffer lock about * identity change by freeing the buffer. */ if (qindex == QUEUE_CLEAN && BUF_LOCKWAITERS(bp)) { bp->b_flags |= B_INVAL; bfreekva(bp); brelse(bp); goto restart; } if (metadata) break; /* * If we are overcomitted then recover the buffer and its * KVM space. This occurs in rare situations when multiple * processes are blocked in getnewbuf() or allocbuf(). */ if (bufspace >= hibufspace) flushingbufs = 1; if (flushingbufs && bp->b_kvasize != 0) { bp->b_flags |= B_INVAL; bfreekva(bp); brelse(bp); goto restart; } if (bufspace < lobufspace) flushingbufs = 0; break; } return (bp); } /* * getnewbuf: * * Find and initialize a new buffer header, freeing up existing buffers * in the bufqueues as necessary. The new buffer is returned locked. * * Important: B_INVAL is not set. If the caller wishes to throw the * buffer away, the caller must set B_INVAL prior to calling brelse(). * * We block if: * We have insufficient buffer headers * We have insufficient buffer space * buffer_arena is too fragmented ( space reservation fails ) * If we have to flush dirty buffers ( but we try to avoid this ) */ static struct buf * getnewbuf(struct vnode *vp, int slpflag, int slptimeo, int size, int maxsize, int gbflags) { struct buf *bp; int defrag, metadata; KASSERT((gbflags & (GB_UNMAPPED | GB_KVAALLOC)) != GB_KVAALLOC, ("GB_KVAALLOC only makes sense with GB_UNMAPPED")); if (!unmapped_buf_allowed) gbflags &= ~(GB_UNMAPPED | GB_KVAALLOC); defrag = 0; if (vp == NULL || (vp->v_vflag & (VV_MD | VV_SYSTEM)) != 0 || vp->v_type == VCHR) metadata = 1; else metadata = 0; /* * We can't afford to block since we might be holding a vnode lock, * which may prevent system daemons from running. We deal with * low-memory situations by proactively returning memory and running * async I/O rather then sync I/O. */ atomic_add_int(&getnewbufcalls, 1); atomic_subtract_int(&getnewbufrestarts, 1); restart: bp = getnewbuf_scan(maxsize, defrag, (gbflags & (GB_UNMAPPED | GB_KVAALLOC)) == GB_UNMAPPED, metadata); if (bp != NULL) defrag = 0; /* * If we exhausted our list, sleep as appropriate. We may have to * wakeup various daemons and write out some dirty buffers. * * Generally we are sleeping due to insufficient buffer space. */ if (bp == NULL) { mtx_assert(&bqclean, MA_OWNED); getnewbuf_bufd_help(vp, gbflags, slpflag, slptimeo, defrag); mtx_assert(&bqclean, MA_NOTOWNED); } else if ((gbflags & (GB_UNMAPPED | GB_KVAALLOC)) == GB_UNMAPPED) { mtx_assert(&bqclean, MA_NOTOWNED); bfreekva(bp); bp->b_flags |= B_UNMAPPED; bp->b_kvabase = bp->b_data = unmapped_buf; bp->b_kvasize = maxsize; atomic_add_long(&bufspace, bp->b_kvasize); atomic_add_long(&unmapped_bufspace, bp->b_kvasize); atomic_add_int(&bufreusecnt, 1); } else { mtx_assert(&bqclean, MA_NOTOWNED); /* * We finally have a valid bp. We aren't quite out of the * woods, we still have to reserve kva space. In order * to keep fragmentation sane we only allocate kva in * BKVASIZE chunks. */ maxsize = (maxsize + BKVAMASK) & ~BKVAMASK; if (maxsize != bp->b_kvasize || (bp->b_flags & (B_UNMAPPED | B_KVAALLOC)) == B_UNMAPPED) { if (allocbufkva(bp, maxsize, gbflags)) { defrag = 1; bp->b_flags |= B_INVAL; brelse(bp); goto restart; } atomic_add_int(&bufreusecnt, 1); } else if ((bp->b_flags & B_KVAALLOC) != 0 && (gbflags & (GB_UNMAPPED | GB_KVAALLOC)) == 0) { /* * If the reused buffer has KVA allocated, * reassign b_kvaalloc to b_kvabase. */ bp->b_kvabase = bp->b_kvaalloc; bp->b_flags &= ~B_KVAALLOC; atomic_subtract_long(&unmapped_bufspace, bp->b_kvasize); atomic_add_int(&bufreusecnt, 1); } else if ((bp->b_flags & (B_UNMAPPED | B_KVAALLOC)) == 0 && (gbflags & (GB_UNMAPPED | GB_KVAALLOC)) == (GB_UNMAPPED | GB_KVAALLOC)) { /* * The case of reused buffer already have KVA * mapped, but the request is for unmapped * buffer with KVA allocated. */ bp->b_kvaalloc = bp->b_kvabase; bp->b_data = bp->b_kvabase = unmapped_buf; bp->b_flags |= B_UNMAPPED | B_KVAALLOC; atomic_add_long(&unmapped_bufspace, bp->b_kvasize); atomic_add_int(&bufreusecnt, 1); } if ((gbflags & GB_UNMAPPED) == 0) { bp->b_saveaddr = bp->b_kvabase; bp->b_data = bp->b_saveaddr; bp->b_flags &= ~B_UNMAPPED; BUF_CHECK_MAPPED(bp); } } return (bp); } /* * buf_daemon: * * buffer flushing daemon. Buffers are normally flushed by the * update daemon but if it cannot keep up this process starts to * take the load in an attempt to prevent getnewbuf() from blocking. */ static struct kproc_desc buf_kp = { "bufdaemon", buf_daemon, &bufdaemonproc }; SYSINIT(bufdaemon, SI_SUB_KTHREAD_BUF, SI_ORDER_FIRST, kproc_start, &buf_kp); static int buf_flush(int target) { int flushed; flushed = flushbufqueues(target, 0); if (flushed == 0) { /* * Could not find any buffers without rollback * dependencies, so just write the first one * in the hopes of eventually making progress. */ flushed = flushbufqueues(target, 1); } return (flushed); } static void buf_daemon() { int lodirty; /* * This process needs to be suspended prior to shutdown sync. */ EVENTHANDLER_REGISTER(shutdown_pre_sync, kproc_shutdown, bufdaemonproc, SHUTDOWN_PRI_LAST); /* * This process is allowed to take the buffer cache to the limit */ curthread->td_pflags |= TDP_NORUNNINGBUF | TDP_BUFNEED; mtx_lock(&bdlock); for (;;) { bd_request = 0; mtx_unlock(&bdlock); kproc_suspend_check(bufdaemonproc); lodirty = lodirtybuffers; if (bd_speedupreq) { lodirty = numdirtybuffers / 2; bd_speedupreq = 0; } /* * Do the flush. Limit the amount of in-transit I/O we * allow to build up, otherwise we would completely saturate * the I/O system. */ while (numdirtybuffers > lodirty) { if (buf_flush(numdirtybuffers - lodirty) == 0) break; kern_yield(PRI_USER); } /* * Only clear bd_request if we have reached our low water * mark. The buf_daemon normally waits 1 second and * then incrementally flushes any dirty buffers that have * built up, within reason. * * If we were unable to hit our low water mark and couldn't * find any flushable buffers, we sleep for a short period * to avoid endless loops on unlockable buffers. */ mtx_lock(&bdlock); if (numdirtybuffers <= lodirtybuffers) { /* * We reached our low water mark, reset the * request and sleep until we are needed again. * The sleep is just so the suspend code works. */ bd_request = 0; /* * Do an extra wakeup in case dirty threshold * changed via sysctl and the explicit transition * out of shortfall was missed. */ bdirtywakeup(); if (runningbufspace <= lorunningspace) runningwakeup(); msleep(&bd_request, &bdlock, PVM, "psleep", hz); } else { /* * We couldn't find any flushable dirty buffers but * still have too many dirty buffers, we * have to sleep and try again. (rare) */ msleep(&bd_request, &bdlock, PVM, "qsleep", hz / 10); } } } /* * flushbufqueues: * * Try to flush a buffer in the dirty queue. We must be careful to * free up B_INVAL buffers instead of write them, which NFS is * particularly sensitive to. */ static int flushwithdeps = 0; SYSCTL_INT(_vfs, OID_AUTO, flushwithdeps, CTLFLAG_RW, &flushwithdeps, 0, "Number of buffers flushed with dependecies that require rollbacks"); static int flushbufqueues(int target, int flushdeps) { struct buf *sentinel; struct vnode *vp; struct mount *mp; struct buf *bp; int hasdeps; int flushed; int queue; int error; flushed = 0; queue = QUEUE_DIRTY; bp = NULL; sentinel = malloc(sizeof(struct buf), M_TEMP, M_WAITOK | M_ZERO); sentinel->b_qindex = QUEUE_SENTINEL; mtx_lock(&bqdirty); TAILQ_INSERT_HEAD(&bufqueues[queue], sentinel, b_freelist); mtx_unlock(&bqdirty); while (flushed != target) { maybe_yield(); mtx_lock(&bqdirty); bp = TAILQ_NEXT(sentinel, b_freelist); if (bp != NULL) { TAILQ_REMOVE(&bufqueues[queue], sentinel, b_freelist); TAILQ_INSERT_AFTER(&bufqueues[queue], bp, sentinel, b_freelist); } else { mtx_unlock(&bqdirty); break; } KASSERT(bp->b_qindex != QUEUE_SENTINEL, ("parallel calls to flushbufqueues() bp %p", bp)); error = BUF_LOCK(bp, LK_EXCLUSIVE | LK_NOWAIT, NULL); mtx_unlock(&bqdirty); if (error != 0) continue; if (bp->b_pin_count > 0) { BUF_UNLOCK(bp); continue; } /* * BKGRDINPROG can only be set with the buf and bufobj * locks both held. We tolerate a race to clear it here. */ if ((bp->b_vflags & BV_BKGRDINPROG) != 0 || (bp->b_flags & B_DELWRI) == 0) { BUF_UNLOCK(bp); continue; } if (bp->b_flags & B_INVAL) { bremfreef(bp); brelse(bp); flushed++; continue; } if (!LIST_EMPTY(&bp->b_dep) && buf_countdeps(bp, 0)) { if (flushdeps == 0) { BUF_UNLOCK(bp); continue; } hasdeps = 1; } else hasdeps = 0; /* * We must hold the lock on a vnode before writing * one of its buffers. Otherwise we may confuse, or * in the case of a snapshot vnode, deadlock the * system. * * The lock order here is the reverse of the normal * of vnode followed by buf lock. This is ok because * the NOWAIT will prevent deadlock. */ vp = bp->b_vp; if (vn_start_write(vp, &mp, V_NOWAIT) != 0) { BUF_UNLOCK(bp); continue; } error = vn_lock(vp, LK_EXCLUSIVE | LK_NOWAIT); if (error == 0) { CTR3(KTR_BUF, "flushbufqueue(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); vfs_bio_awrite(bp); vn_finished_write(mp); VOP_UNLOCK(vp, 0); flushwithdeps += hasdeps; flushed++; if (runningbufspace > hirunningspace) waitrunningbufspace(); continue; } vn_finished_write(mp); BUF_UNLOCK(bp); } mtx_lock(&bqdirty); TAILQ_REMOVE(&bufqueues[queue], sentinel, b_freelist); mtx_unlock(&bqdirty); free(sentinel, M_TEMP); return (flushed); } /* * Check to see if a block is currently memory resident. */ struct buf * incore(struct bufobj *bo, daddr_t blkno) { struct buf *bp; BO_RLOCK(bo); bp = gbincore(bo, blkno); BO_RUNLOCK(bo); return (bp); } /* * Returns true if no I/O is needed to access the * associated VM object. This is like incore except * it also hunts around in the VM system for the data. */ static int inmem(struct vnode * vp, daddr_t blkno) { vm_object_t obj; vm_offset_t toff, tinc, size; vm_page_t m; vm_ooffset_t off; ASSERT_VOP_LOCKED(vp, "inmem"); if (incore(&vp->v_bufobj, blkno)) return 1; if (vp->v_mount == NULL) return 0; obj = vp->v_object; if (obj == NULL) return (0); size = PAGE_SIZE; if (size > vp->v_mount->mnt_stat.f_iosize) size = vp->v_mount->mnt_stat.f_iosize; off = (vm_ooffset_t)blkno * (vm_ooffset_t)vp->v_mount->mnt_stat.f_iosize; VM_OBJECT_RLOCK(obj); for (toff = 0; toff < vp->v_mount->mnt_stat.f_iosize; toff += tinc) { m = vm_page_lookup(obj, OFF_TO_IDX(off + toff)); if (!m) goto notinmem; tinc = size; if (tinc > PAGE_SIZE - ((toff + off) & PAGE_MASK)) tinc = PAGE_SIZE - ((toff + off) & PAGE_MASK); if (vm_page_is_valid(m, (vm_offset_t) ((toff + off) & PAGE_MASK), tinc) == 0) goto notinmem; } VM_OBJECT_RUNLOCK(obj); return 1; notinmem: VM_OBJECT_RUNLOCK(obj); return (0); } /* * Set the dirty range for a buffer based on the status of the dirty * bits in the pages comprising the buffer. The range is limited * to the size of the buffer. * * Tell the VM system that the pages associated with this buffer * are clean. This is used for delayed writes where the data is * going to go to disk eventually without additional VM intevention. * * Note that while we only really need to clean through to b_bcount, we * just go ahead and clean through to b_bufsize. */ static void vfs_clean_pages_dirty_buf(struct buf *bp) { vm_ooffset_t foff, noff, eoff; vm_page_t m; int i; if ((bp->b_flags & B_VMIO) == 0 || bp->b_bufsize == 0) return; foff = bp->b_offset; KASSERT(bp->b_offset != NOOFFSET, ("vfs_clean_pages_dirty_buf: no buffer offset")); VM_OBJECT_WLOCK(bp->b_bufobj->bo_object); vfs_drain_busy_pages(bp); vfs_setdirty_locked_object(bp); for (i = 0; i < bp->b_npages; i++) { noff = (foff + PAGE_SIZE) & ~(off_t)PAGE_MASK; eoff = noff; if (eoff > bp->b_offset + bp->b_bufsize) eoff = bp->b_offset + bp->b_bufsize; m = bp->b_pages[i]; vfs_page_set_validclean(bp, foff, m); /* vm_page_clear_dirty(m, foff & PAGE_MASK, eoff - foff); */ foff = noff; } VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object); } static void vfs_setdirty_locked_object(struct buf *bp) { vm_object_t object; int i; object = bp->b_bufobj->bo_object; VM_OBJECT_ASSERT_WLOCKED(object); /* * We qualify the scan for modified pages on whether the * object has been flushed yet. */ if ((object->flags & OBJ_MIGHTBEDIRTY) != 0) { vm_offset_t boffset; vm_offset_t eoffset; /* * test the pages to see if they have been modified directly * by users through the VM system. */ for (i = 0; i < bp->b_npages; i++) vm_page_test_dirty(bp->b_pages[i]); /* * Calculate the encompassing dirty range, boffset and eoffset, * (eoffset - boffset) bytes. */ for (i = 0; i < bp->b_npages; i++) { if (bp->b_pages[i]->dirty) break; } boffset = (i << PAGE_SHIFT) - (bp->b_offset & PAGE_MASK); for (i = bp->b_npages - 1; i >= 0; --i) { if (bp->b_pages[i]->dirty) { break; } } eoffset = ((i + 1) << PAGE_SHIFT) - (bp->b_offset & PAGE_MASK); /* * Fit it to the buffer. */ if (eoffset > bp->b_bcount) eoffset = bp->b_bcount; /* * If we have a good dirty range, merge with the existing * dirty range. */ if (boffset < eoffset) { if (bp->b_dirtyoff > boffset) bp->b_dirtyoff = boffset; if (bp->b_dirtyend < eoffset) bp->b_dirtyend = eoffset; } } } /* * Allocate the KVA mapping for an existing buffer. It handles the * cases of both B_UNMAPPED buffer, and buffer with the preallocated * KVA which is not mapped (B_KVAALLOC). */ static void bp_unmapped_get_kva(struct buf *bp, daddr_t blkno, int size, int gbflags) { struct buf *scratch_bp; int bsize, maxsize, need_mapping, need_kva; off_t offset; need_mapping = (bp->b_flags & B_UNMAPPED) != 0 && (gbflags & GB_UNMAPPED) == 0; need_kva = (bp->b_flags & (B_KVAALLOC | B_UNMAPPED)) == B_UNMAPPED && (gbflags & GB_KVAALLOC) != 0; if (!need_mapping && !need_kva) return; BUF_CHECK_UNMAPPED(bp); if (need_mapping && (bp->b_flags & B_KVAALLOC) != 0) { /* * Buffer is not mapped, but the KVA was already * reserved at the time of the instantiation. Use the * allocated space. */ bp->b_flags &= ~B_KVAALLOC; KASSERT(bp->b_kvaalloc != 0, ("kvaalloc == 0")); bp->b_kvabase = bp->b_kvaalloc; atomic_subtract_long(&unmapped_bufspace, bp->b_kvasize); goto has_addr; } /* * Calculate the amount of the address space we would reserve * if the buffer was mapped. */ bsize = vn_isdisk(bp->b_vp, NULL) ? DEV_BSIZE : bp->b_bufobj->bo_bsize; KASSERT(bsize != 0, ("bsize == 0, check bo->bo_bsize")); offset = blkno * bsize; maxsize = size + (offset & PAGE_MASK); maxsize = imax(maxsize, bsize); mapping_loop: if (allocbufkva(bp, maxsize, gbflags)) { /* * Request defragmentation. getnewbuf() returns us the * allocated space by the scratch buffer KVA. */ scratch_bp = getnewbuf(bp->b_vp, 0, 0, size, maxsize, gbflags | (GB_UNMAPPED | GB_KVAALLOC)); if (scratch_bp == NULL) { if ((gbflags & GB_NOWAIT_BD) != 0) { /* * XXXKIB: defragmentation cannot * succeed, not sure what else to do. */ panic("GB_NOWAIT_BD and B_UNMAPPED %p", bp); } atomic_add_int(&mappingrestarts, 1); goto mapping_loop; } KASSERT((scratch_bp->b_flags & B_KVAALLOC) != 0, ("scratch bp !B_KVAALLOC %p", scratch_bp)); setbufkva(bp, (vm_offset_t)scratch_bp->b_kvaalloc, scratch_bp->b_kvasize, gbflags); /* Get rid of the scratch buffer. */ scratch_bp->b_kvasize = 0; scratch_bp->b_flags |= B_INVAL; scratch_bp->b_flags &= ~(B_UNMAPPED | B_KVAALLOC); brelse(scratch_bp); } if (!need_mapping) return; has_addr: bp->b_saveaddr = bp->b_kvabase; bp->b_data = bp->b_saveaddr; /* b_offset is handled by bpmap_qenter */ bp->b_flags &= ~B_UNMAPPED; BUF_CHECK_MAPPED(bp); bpmap_qenter(bp); } /* * getblk: * * Get a block given a specified block and offset into a file/device. * The buffers B_DONE bit will be cleared on return, making it almost * ready for an I/O initiation. B_INVAL may or may not be set on * return. The caller should clear B_INVAL prior to initiating a * READ. * * For a non-VMIO buffer, B_CACHE is set to the opposite of B_INVAL for * an existing buffer. * * For a VMIO buffer, B_CACHE is modified according to the backing VM. * If getblk()ing a previously 0-sized invalid buffer, B_CACHE is set * and then cleared based on the backing VM. If the previous buffer is * non-0-sized but invalid, B_CACHE will be cleared. * * If getblk() must create a new buffer, the new buffer is returned with * both B_INVAL and B_CACHE clear unless it is a VMIO buffer, in which * case it is returned with B_INVAL clear and B_CACHE set based on the * backing VM. * * getblk() also forces a bwrite() for any B_DELWRI buffer whos * B_CACHE bit is clear. * * What this means, basically, is that the caller should use B_CACHE to * determine whether the buffer is fully valid or not and should clear * B_INVAL prior to issuing a read. If the caller intends to validate * the buffer by loading its data area with something, the caller needs * to clear B_INVAL. If the caller does this without issuing an I/O, * the caller should set B_CACHE ( as an optimization ), else the caller * should issue the I/O and biodone() will set B_CACHE if the I/O was * a write attempt or if it was a successfull read. If the caller * intends to issue a READ, the caller must clear B_INVAL and BIO_ERROR * prior to issuing the READ. biodone() will *not* clear B_INVAL. */ struct buf * getblk(struct vnode *vp, daddr_t blkno, int size, int slpflag, int slptimeo, int flags) { struct buf *bp; struct bufobj *bo; int bsize, error, maxsize, vmio; off_t offset; CTR3(KTR_BUF, "getblk(%p, %ld, %d)", vp, (long)blkno, size); KASSERT((flags & (GB_UNMAPPED | GB_KVAALLOC)) != GB_KVAALLOC, ("GB_KVAALLOC only makes sense with GB_UNMAPPED")); ASSERT_VOP_LOCKED(vp, "getblk"); if (size > MAXBSIZE) panic("getblk: size(%d) > MAXBSIZE(%d)\n", size, MAXBSIZE); if (!unmapped_buf_allowed) flags &= ~(GB_UNMAPPED | GB_KVAALLOC); bo = &vp->v_bufobj; loop: BO_RLOCK(bo); bp = gbincore(bo, blkno); if (bp != NULL) { int lockflags; /* * Buffer is in-core. If the buffer is not busy nor managed, * it must be on a queue. */ lockflags = LK_EXCLUSIVE | LK_SLEEPFAIL | LK_INTERLOCK; if (flags & GB_LOCK_NOWAIT) lockflags |= LK_NOWAIT; error = BUF_TIMELOCK(bp, lockflags, BO_LOCKPTR(bo), "getblk", slpflag, slptimeo); /* * If we slept and got the lock we have to restart in case * the buffer changed identities. */ if (error == ENOLCK) goto loop; /* We timed out or were interrupted. */ else if (error) return (NULL); /* If recursed, assume caller knows the rules. */ else if (BUF_LOCKRECURSED(bp)) goto end; /* * The buffer is locked. B_CACHE is cleared if the buffer is * invalid. Otherwise, for a non-VMIO buffer, B_CACHE is set * and for a VMIO buffer B_CACHE is adjusted according to the * backing VM cache. */ if (bp->b_flags & B_INVAL) bp->b_flags &= ~B_CACHE; else if ((bp->b_flags & (B_VMIO | B_INVAL)) == 0) bp->b_flags |= B_CACHE; if (bp->b_flags & B_MANAGED) MPASS(bp->b_qindex == QUEUE_NONE); else bremfree(bp); /* * check for size inconsistencies for non-VMIO case. */ if (bp->b_bcount != size) { if ((bp->b_flags & B_VMIO) == 0 || (size > bp->b_kvasize)) { if (bp->b_flags & B_DELWRI) { /* * If buffer is pinned and caller does * not want sleep waiting for it to be * unpinned, bail out * */ if (bp->b_pin_count > 0) { if (flags & GB_LOCK_NOWAIT) { bqrelse(bp); return (NULL); } else { bunpin_wait(bp); } } bp->b_flags |= B_NOCACHE; bwrite(bp); } else { if (LIST_EMPTY(&bp->b_dep)) { bp->b_flags |= B_RELBUF; brelse(bp); } else { bp->b_flags |= B_NOCACHE; bwrite(bp); } } goto loop; } } /* * Handle the case of unmapped buffer which should * become mapped, or the buffer for which KVA * reservation is requested. */ bp_unmapped_get_kva(bp, blkno, size, flags); /* * If the size is inconsistant in the VMIO case, we can resize * the buffer. This might lead to B_CACHE getting set or * cleared. If the size has not changed, B_CACHE remains * unchanged from its previous state. */ if (bp->b_bcount != size) allocbuf(bp, size); KASSERT(bp->b_offset != NOOFFSET, ("getblk: no buffer offset")); /* * A buffer with B_DELWRI set and B_CACHE clear must * be committed before we can return the buffer in * order to prevent the caller from issuing a read * ( due to B_CACHE not being set ) and overwriting * it. * * Most callers, including NFS and FFS, need this to * operate properly either because they assume they * can issue a read if B_CACHE is not set, or because * ( for example ) an uncached B_DELWRI might loop due * to softupdates re-dirtying the buffer. In the latter * case, B_CACHE is set after the first write completes, * preventing further loops. * NOTE! b*write() sets B_CACHE. If we cleared B_CACHE * above while extending the buffer, we cannot allow the * buffer to remain with B_CACHE set after the write * completes or it will represent a corrupt state. To * deal with this we set B_NOCACHE to scrap the buffer * after the write. * * We might be able to do something fancy, like setting * B_CACHE in bwrite() except if B_DELWRI is already set, * so the below call doesn't set B_CACHE, but that gets real * confusing. This is much easier. */ if ((bp->b_flags & (B_CACHE|B_DELWRI)) == B_DELWRI) { bp->b_flags |= B_NOCACHE; bwrite(bp); goto loop; } bp->b_flags &= ~B_DONE; } else { /* * Buffer is not in-core, create new buffer. The buffer * returned by getnewbuf() is locked. Note that the returned * buffer is also considered valid (not marked B_INVAL). */ BO_RUNLOCK(bo); /* * If the user does not want us to create the buffer, bail out * here. */ if (flags & GB_NOCREAT) return NULL; if (numfreebuffers == 0 && TD_IS_IDLETHREAD(curthread)) return NULL; bsize = vn_isdisk(vp, NULL) ? DEV_BSIZE : bo->bo_bsize; KASSERT(bsize != 0, ("bsize == 0, check bo->bo_bsize")); offset = blkno * bsize; vmio = vp->v_object != NULL; if (vmio) { maxsize = size + (offset & PAGE_MASK); } else { maxsize = size; /* Do not allow non-VMIO notmapped buffers. */ flags &= ~GB_UNMAPPED; } maxsize = imax(maxsize, bsize); bp = getnewbuf(vp, slpflag, slptimeo, size, maxsize, flags); if (bp == NULL) { if (slpflag || slptimeo) return NULL; goto loop; } /* * This code is used to make sure that a buffer is not * created while the getnewbuf routine is blocked. * This can be a problem whether the vnode is locked or not. * If the buffer is created out from under us, we have to * throw away the one we just created. * * Note: this must occur before we associate the buffer * with the vp especially considering limitations in * the splay tree implementation when dealing with duplicate * lblkno's. */ BO_LOCK(bo); if (gbincore(bo, blkno)) { BO_UNLOCK(bo); bp->b_flags |= B_INVAL; brelse(bp); goto loop; } /* * Insert the buffer into the hash, so that it can * be found by incore. */ bp->b_blkno = bp->b_lblkno = blkno; bp->b_offset = offset; bgetvp(vp, bp); BO_UNLOCK(bo); /* * set B_VMIO bit. allocbuf() the buffer bigger. Since the * buffer size starts out as 0, B_CACHE will be set by * allocbuf() for the VMIO case prior to it testing the * backing store for validity. */ if (vmio) { bp->b_flags |= B_VMIO; KASSERT(vp->v_object == bp->b_bufobj->bo_object, ("ARGH! different b_bufobj->bo_object %p %p %p\n", bp, vp->v_object, bp->b_bufobj->bo_object)); } else { bp->b_flags &= ~B_VMIO; KASSERT(bp->b_bufobj->bo_object == NULL, ("ARGH! has b_bufobj->bo_object %p %p\n", bp, bp->b_bufobj->bo_object)); BUF_CHECK_MAPPED(bp); } allocbuf(bp, size); bp->b_flags &= ~B_DONE; } CTR4(KTR_BUF, "getblk(%p, %ld, %d) = %p", vp, (long)blkno, size, bp); BUF_ASSERT_HELD(bp); end: KASSERT(bp->b_bufobj == bo, ("bp %p wrong b_bufobj %p should be %p", bp, bp->b_bufobj, bo)); return (bp); } /* * Get an empty, disassociated buffer of given size. The buffer is initially * set to B_INVAL. */ struct buf * geteblk(int size, int flags) { struct buf *bp; int maxsize; maxsize = (size + BKVAMASK) & ~BKVAMASK; while ((bp = getnewbuf(NULL, 0, 0, size, maxsize, flags)) == NULL) { if ((flags & GB_NOWAIT_BD) && (curthread->td_pflags & TDP_BUFNEED) != 0) return (NULL); } allocbuf(bp, size); bp->b_flags |= B_INVAL; /* b_dep cleared by getnewbuf() */ BUF_ASSERT_HELD(bp); return (bp); } /* * This code constitutes the buffer memory from either anonymous system * memory (in the case of non-VMIO operations) or from an associated * VM object (in the case of VMIO operations). This code is able to * resize a buffer up or down. * * Note that this code is tricky, and has many complications to resolve * deadlock or inconsistant data situations. Tread lightly!!! * There are B_CACHE and B_DELWRI interactions that must be dealt with by * the caller. Calling this code willy nilly can result in the loss of data. * * allocbuf() only adjusts B_CACHE for VMIO buffers. getblk() deals with * B_CACHE for the non-VMIO case. */ int allocbuf(struct buf *bp, int size) { int newbsize, mbsize; int i; BUF_ASSERT_HELD(bp); if (bp->b_kvasize < size) panic("allocbuf: buffer too small"); if ((bp->b_flags & B_VMIO) == 0) { caddr_t origbuf; int origbufsize; /* * Just get anonymous memory from the kernel. Don't * mess with B_CACHE. */ mbsize = (size + DEV_BSIZE - 1) & ~(DEV_BSIZE - 1); if (bp->b_flags & B_MALLOC) newbsize = mbsize; else newbsize = round_page(size); if (newbsize < bp->b_bufsize) { /* * malloced buffers are not shrunk */ if (bp->b_flags & B_MALLOC) { if (newbsize) { bp->b_bcount = size; } else { free(bp->b_data, M_BIOBUF); if (bp->b_bufsize) { atomic_subtract_long( &bufmallocspace, bp->b_bufsize); bufspacewakeup(); bp->b_bufsize = 0; } bp->b_saveaddr = bp->b_kvabase; bp->b_data = bp->b_saveaddr; bp->b_bcount = 0; bp->b_flags &= ~B_MALLOC; } return 1; } vm_hold_free_pages(bp, newbsize); } else if (newbsize > bp->b_bufsize) { /* * We only use malloced memory on the first allocation. * and revert to page-allocated memory when the buffer * grows. */ /* * There is a potential smp race here that could lead * to bufmallocspace slightly passing the max. It * is probably extremely rare and not worth worrying * over. */ if ( (bufmallocspace < maxbufmallocspace) && (bp->b_bufsize == 0) && (mbsize <= PAGE_SIZE/2)) { bp->b_data = malloc(mbsize, M_BIOBUF, M_WAITOK); bp->b_bufsize = mbsize; bp->b_bcount = size; bp->b_flags |= B_MALLOC; atomic_add_long(&bufmallocspace, mbsize); return 1; } origbuf = NULL; origbufsize = 0; /* * If the buffer is growing on its other-than-first allocation, * then we revert to the page-allocation scheme. */ if (bp->b_flags & B_MALLOC) { origbuf = bp->b_data; origbufsize = bp->b_bufsize; bp->b_data = bp->b_kvabase; if (bp->b_bufsize) { atomic_subtract_long(&bufmallocspace, bp->b_bufsize); bufspacewakeup(); bp->b_bufsize = 0; } bp->b_flags &= ~B_MALLOC; newbsize = round_page(newbsize); } vm_hold_load_pages( bp, (vm_offset_t) bp->b_data + bp->b_bufsize, (vm_offset_t) bp->b_data + newbsize); if (origbuf) { bcopy(origbuf, bp->b_data, origbufsize); free(origbuf, M_BIOBUF); } } } else { int desiredpages; newbsize = (size + DEV_BSIZE - 1) & ~(DEV_BSIZE - 1); desiredpages = (size == 0) ? 0 : num_pages((bp->b_offset & PAGE_MASK) + newbsize); if (bp->b_flags & B_MALLOC) panic("allocbuf: VMIO buffer can't be malloced"); /* * Set B_CACHE initially if buffer is 0 length or will become * 0-length. */ if (size == 0 || bp->b_bufsize == 0) bp->b_flags |= B_CACHE; if (newbsize < bp->b_bufsize) { /* * DEV_BSIZE aligned new buffer size is less then the * DEV_BSIZE aligned existing buffer size. Figure out * if we have to remove any pages. */ if (desiredpages < bp->b_npages) { vm_page_t m; if ((bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); pmap_qremove((vm_offset_t)trunc_page( (vm_offset_t)bp->b_data) + (desiredpages << PAGE_SHIFT), (bp->b_npages - desiredpages)); } else BUF_CHECK_UNMAPPED(bp); VM_OBJECT_WLOCK(bp->b_bufobj->bo_object); for (i = desiredpages; i < bp->b_npages; i++) { /* * the page is not freed here -- it * is the responsibility of * vnode_pager_setsize */ m = bp->b_pages[i]; KASSERT(m != bogus_page, ("allocbuf: bogus page found")); while (vm_page_sleep_if_busy(m, "biodep")) continue; bp->b_pages[i] = NULL; vm_page_lock(m); vm_page_unwire(m, PQ_INACTIVE); vm_page_unlock(m); } VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object); bp->b_npages = desiredpages; } } else if (size > bp->b_bcount) { /* * We are growing the buffer, possibly in a * byte-granular fashion. */ vm_object_t obj; vm_offset_t toff; vm_offset_t tinc; /* * Step 1, bring in the VM pages from the object, * allocating them if necessary. We must clear * B_CACHE if these pages are not valid for the * range covered by the buffer. */ obj = bp->b_bufobj->bo_object; VM_OBJECT_WLOCK(obj); while (bp->b_npages < desiredpages) { vm_page_t m; /* * We must allocate system pages since blocking * here could interfere with paging I/O, no * matter which process we are. * * Only exclusive busy can be tested here. * Blocking on shared busy might lead to * deadlocks once allocbuf() is called after * pages are vfs_busy_pages(). */ m = vm_page_grab(obj, OFF_TO_IDX(bp->b_offset) + bp->b_npages, VM_ALLOC_NOBUSY | VM_ALLOC_SYSTEM | VM_ALLOC_WIRED | VM_ALLOC_IGN_SBUSY | VM_ALLOC_COUNT(desiredpages - bp->b_npages)); if (m->valid == 0) bp->b_flags &= ~B_CACHE; bp->b_pages[bp->b_npages] = m; ++bp->b_npages; } /* * Step 2. We've loaded the pages into the buffer, * we have to figure out if we can still have B_CACHE * set. Note that B_CACHE is set according to the * byte-granular range ( bcount and size ), new the * aligned range ( newbsize ). * * The VM test is against m->valid, which is DEV_BSIZE * aligned. Needless to say, the validity of the data * needs to also be DEV_BSIZE aligned. Note that this * fails with NFS if the server or some other client * extends the file's EOF. If our buffer is resized, * B_CACHE may remain set! XXX */ toff = bp->b_bcount; tinc = PAGE_SIZE - ((bp->b_offset + toff) & PAGE_MASK); while ((bp->b_flags & B_CACHE) && toff < size) { vm_pindex_t pi; if (tinc > (size - toff)) tinc = size - toff; pi = ((bp->b_offset & PAGE_MASK) + toff) >> PAGE_SHIFT; vfs_buf_test_cache( bp, bp->b_offset, toff, tinc, bp->b_pages[pi] ); toff += tinc; tinc = PAGE_SIZE; } VM_OBJECT_WUNLOCK(obj); /* * Step 3, fixup the KVM pmap. */ if ((bp->b_flags & B_UNMAPPED) == 0) bpmap_qenter(bp); else BUF_CHECK_UNMAPPED(bp); } } if (newbsize < bp->b_bufsize) bufspacewakeup(); bp->b_bufsize = newbsize; /* actual buffer allocation */ bp->b_bcount = size; /* requested buffer size */ return 1; } extern int inflight_transient_maps; void biodone(struct bio *bp) { struct mtx *mtxp; void (*done)(struct bio *); vm_offset_t start, end; if ((bp->bio_flags & BIO_TRANSIENT_MAPPING) != 0) { bp->bio_flags &= ~BIO_TRANSIENT_MAPPING; bp->bio_flags |= BIO_UNMAPPED; start = trunc_page((vm_offset_t)bp->bio_data); end = round_page((vm_offset_t)bp->bio_data + bp->bio_length); pmap_qremove(start, OFF_TO_IDX(end - start)); vmem_free(transient_arena, start, end - start); atomic_add_int(&inflight_transient_maps, -1); } done = bp->bio_done; if (done == NULL) { mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); bp->bio_flags |= BIO_DONE; wakeup(bp); mtx_unlock(mtxp); } else { bp->bio_flags |= BIO_DONE; done(bp); } } /* * Wait for a BIO to finish. */ int biowait(struct bio *bp, const char *wchan) { struct mtx *mtxp; mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); while ((bp->bio_flags & BIO_DONE) == 0) msleep(bp, mtxp, PRIBIO, wchan, 0); mtx_unlock(mtxp); if (bp->bio_error != 0) return (bp->bio_error); if (!(bp->bio_flags & BIO_ERROR)) return (0); return (EIO); } void biofinish(struct bio *bp, struct devstat *stat, int error) { if (error) { bp->bio_error = error; bp->bio_flags |= BIO_ERROR; } if (stat != NULL) devstat_end_transaction_bio(stat, bp); biodone(bp); } /* * bufwait: * * Wait for buffer I/O completion, returning error status. The buffer * is left locked and B_DONE on return. B_EINTR is converted into an EINTR * error and cleared. */ int bufwait(struct buf *bp) { if (bp->b_iocmd == BIO_READ) bwait(bp, PRIBIO, "biord"); else bwait(bp, PRIBIO, "biowr"); if (bp->b_flags & B_EINTR) { bp->b_flags &= ~B_EINTR; return (EINTR); } if (bp->b_ioflags & BIO_ERROR) { return (bp->b_error ? bp->b_error : EIO); } else { return (0); } } /* * Call back function from struct bio back up to struct buf. */ static void bufdonebio(struct bio *bip) { struct buf *bp; bp = bip->bio_caller2; bp->b_resid = bip->bio_resid; bp->b_ioflags = bip->bio_flags; bp->b_error = bip->bio_error; if (bp->b_error) bp->b_ioflags |= BIO_ERROR; bufdone(bp); g_destroy_bio(bip); } void dev_strategy(struct cdev *dev, struct buf *bp) { struct cdevsw *csw; int ref; KASSERT(dev->si_refcount > 0, ("dev_strategy on un-referenced struct cdev *(%s) %p", devtoname(dev), dev)); csw = dev_refthread(dev, &ref); dev_strategy_csw(dev, csw, bp); dev_relthread(dev, ref); } void dev_strategy_csw(struct cdev *dev, struct cdevsw *csw, struct buf *bp) { struct bio *bip; KASSERT(bp->b_iocmd == BIO_READ || bp->b_iocmd == BIO_WRITE, ("b_iocmd botch")); KASSERT(((dev->si_flags & SI_ETERNAL) != 0 && csw != NULL) || dev->si_threadcount > 0, ("dev_strategy_csw threadcount cdev *(%s) %p", devtoname(dev), dev)); if (csw == NULL) { bp->b_error = ENXIO; bp->b_ioflags = BIO_ERROR; bufdone(bp); return; } for (;;) { bip = g_new_bio(); if (bip != NULL) break; /* Try again later */ tsleep(&bp, PRIBIO, "dev_strat", hz/10); } bip->bio_cmd = bp->b_iocmd; bip->bio_offset = bp->b_iooffset; bip->bio_length = bp->b_bcount; bip->bio_bcount = bp->b_bcount; /* XXX: remove */ bdata2bio(bp, bip); bip->bio_done = bufdonebio; bip->bio_caller2 = bp; bip->bio_dev = dev; (*csw->d_strategy)(bip); } /* * bufdone: * * Finish I/O on a buffer, optionally calling a completion function. * This is usually called from an interrupt so process blocking is * not allowed. * * biodone is also responsible for setting B_CACHE in a B_VMIO bp. * In a non-VMIO bp, B_CACHE will be set on the next getblk() * assuming B_INVAL is clear. * * For the VMIO case, we set B_CACHE if the op was a read and no * read error occured, or if the op was a write. B_CACHE is never * set if the buffer is invalid or otherwise uncacheable. * * biodone does not mess with B_INVAL, allowing the I/O routine or the * initiator to leave B_INVAL set to brelse the buffer out of existance * in the biodone routine. */ void bufdone(struct buf *bp) { struct bufobj *dropobj; void (*biodone)(struct buf *); CTR3(KTR_BUF, "bufdone(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); dropobj = NULL; KASSERT(!(bp->b_flags & B_DONE), ("biodone: bp %p already done", bp)); BUF_ASSERT_HELD(bp); runningbufwakeup(bp); if (bp->b_iocmd == BIO_WRITE) dropobj = bp->b_bufobj; /* call optional completion function if requested */ if (bp->b_iodone != NULL) { biodone = bp->b_iodone; bp->b_iodone = NULL; (*biodone) (bp); if (dropobj) bufobj_wdrop(dropobj); return; } bufdone_finish(bp); if (dropobj) bufobj_wdrop(dropobj); } void bufdone_finish(struct buf *bp) { BUF_ASSERT_HELD(bp); if (!LIST_EMPTY(&bp->b_dep)) buf_complete(bp); if (bp->b_flags & B_VMIO) { vm_ooffset_t foff; vm_page_t m; vm_object_t obj; struct vnode *vp; int bogus, i, iosize; obj = bp->b_bufobj->bo_object; KASSERT(obj->paging_in_progress >= bp->b_npages, ("biodone_finish: paging in progress(%d) < b_npages(%d)", obj->paging_in_progress, bp->b_npages)); vp = bp->b_vp; KASSERT(vp->v_holdcnt > 0, ("biodone_finish: vnode %p has zero hold count", vp)); KASSERT(vp->v_object != NULL, ("biodone_finish: vnode %p has no vm_object", vp)); foff = bp->b_offset; KASSERT(bp->b_offset != NOOFFSET, ("biodone_finish: bp %p has no buffer offset", bp)); /* * Set B_CACHE if the op was a normal read and no error * occured. B_CACHE is set for writes in the b*write() * routines. */ iosize = bp->b_bcount - bp->b_resid; if (bp->b_iocmd == BIO_READ && !(bp->b_flags & (B_INVAL|B_NOCACHE)) && !(bp->b_ioflags & BIO_ERROR)) { bp->b_flags |= B_CACHE; } bogus = 0; VM_OBJECT_WLOCK(obj); for (i = 0; i < bp->b_npages; i++) { int bogusflag = 0; int resid; resid = ((foff + PAGE_SIZE) & ~(off_t)PAGE_MASK) - foff; if (resid > iosize) resid = iosize; /* * cleanup bogus pages, restoring the originals */ m = bp->b_pages[i]; if (m == bogus_page) { bogus = bogusflag = 1; m = vm_page_lookup(obj, OFF_TO_IDX(foff)); if (m == NULL) panic("biodone: page disappeared!"); bp->b_pages[i] = m; } KASSERT(OFF_TO_IDX(foff) == m->pindex, ("biodone_finish: foff(%jd)/pindex(%ju) mismatch", (intmax_t)foff, (uintmax_t)m->pindex)); /* * In the write case, the valid and clean bits are * already changed correctly ( see bdwrite() ), so we * only need to do this here in the read case. */ if ((bp->b_iocmd == BIO_READ) && !bogusflag && resid > 0) { KASSERT((m->dirty & vm_page_bits(foff & PAGE_MASK, resid)) == 0, ("bufdone_finish:" " page %p has unexpected dirty bits", m)); vfs_page_set_valid(bp, foff, m); } vm_page_sunbusy(m); vm_object_pip_subtract(obj, 1); foff = (foff + PAGE_SIZE) & ~(off_t)PAGE_MASK; iosize -= resid; } vm_object_pip_wakeupn(obj, 0); VM_OBJECT_WUNLOCK(obj); if (bogus && (bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); pmap_qenter(trunc_page((vm_offset_t)bp->b_data), bp->b_pages, bp->b_npages); } } /* * For asynchronous completions, release the buffer now. The brelse * will do a wakeup there if necessary - so no need to do a wakeup * here in the async case. The sync case always needs to do a wakeup. */ if (bp->b_flags & B_ASYNC) { if ((bp->b_flags & (B_NOCACHE | B_INVAL | B_RELBUF)) || (bp->b_ioflags & BIO_ERROR)) brelse(bp); else bqrelse(bp); } else bdone(bp); } /* * This routine is called in lieu of iodone in the case of * incomplete I/O. This keeps the busy status for pages * consistant. */ void vfs_unbusy_pages(struct buf *bp) { int i; vm_object_t obj; vm_page_t m; runningbufwakeup(bp); if (!(bp->b_flags & B_VMIO)) return; obj = bp->b_bufobj->bo_object; VM_OBJECT_WLOCK(obj); for (i = 0; i < bp->b_npages; i++) { m = bp->b_pages[i]; if (m == bogus_page) { m = vm_page_lookup(obj, OFF_TO_IDX(bp->b_offset) + i); if (!m) panic("vfs_unbusy_pages: page missing\n"); bp->b_pages[i] = m; if ((bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); pmap_qenter(trunc_page((vm_offset_t)bp->b_data), bp->b_pages, bp->b_npages); } else BUF_CHECK_UNMAPPED(bp); } vm_object_pip_subtract(obj, 1); vm_page_sunbusy(m); } vm_object_pip_wakeupn(obj, 0); VM_OBJECT_WUNLOCK(obj); } /* * vfs_page_set_valid: * * Set the valid bits in a page based on the supplied offset. The * range is restricted to the buffer's size. * * This routine is typically called after a read completes. */ static void vfs_page_set_valid(struct buf *bp, vm_ooffset_t off, vm_page_t m) { vm_ooffset_t eoff; /* * Compute the end offset, eoff, such that [off, eoff) does not span a * page boundary and eoff is not greater than the end of the buffer. * The end of the buffer, in this case, is our file EOF, not the * allocation size of the buffer. */ eoff = (off + PAGE_SIZE) & ~(vm_ooffset_t)PAGE_MASK; if (eoff > bp->b_offset + bp->b_bcount) eoff = bp->b_offset + bp->b_bcount; /* * Set valid range. This is typically the entire buffer and thus the * entire page. */ if (eoff > off) vm_page_set_valid_range(m, off & PAGE_MASK, eoff - off); } /* * vfs_page_set_validclean: * * Set the valid bits and clear the dirty bits in a page based on the * supplied offset. The range is restricted to the buffer's size. */ static void vfs_page_set_validclean(struct buf *bp, vm_ooffset_t off, vm_page_t m) { vm_ooffset_t soff, eoff; /* * Start and end offsets in buffer. eoff - soff may not cross a * page boundry or cross the end of the buffer. The end of the * buffer, in this case, is our file EOF, not the allocation size * of the buffer. */ soff = off; eoff = (off + PAGE_SIZE) & ~(off_t)PAGE_MASK; if (eoff > bp->b_offset + bp->b_bcount) eoff = bp->b_offset + bp->b_bcount; /* * Set valid range. This is typically the entire buffer and thus the * entire page. */ if (eoff > soff) { vm_page_set_validclean( m, (vm_offset_t) (soff & PAGE_MASK), (vm_offset_t) (eoff - soff) ); } } /* * Ensure that all buffer pages are not exclusive busied. If any page is * exclusive busy, drain it. */ void vfs_drain_busy_pages(struct buf *bp) { vm_page_t m; int i, last_busied; VM_OBJECT_ASSERT_WLOCKED(bp->b_bufobj->bo_object); last_busied = 0; for (i = 0; i < bp->b_npages; i++) { m = bp->b_pages[i]; if (vm_page_xbusied(m)) { for (; last_busied < i; last_busied++) vm_page_sbusy(bp->b_pages[last_busied]); while (vm_page_xbusied(m)) { vm_page_lock(m); VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object); vm_page_busy_sleep(m, "vbpage"); VM_OBJECT_WLOCK(bp->b_bufobj->bo_object); } } } for (i = 0; i < last_busied; i++) vm_page_sunbusy(bp->b_pages[i]); } /* * This routine is called before a device strategy routine. * It is used to tell the VM system that paging I/O is in * progress, and treat the pages associated with the buffer * almost as being exclusive busy. Also the object paging_in_progress * flag is handled to make sure that the object doesn't become * inconsistant. * * Since I/O has not been initiated yet, certain buffer flags * such as BIO_ERROR or B_INVAL may be in an inconsistant state * and should be ignored. */ void vfs_busy_pages(struct buf *bp, int clear_modify) { int i, bogus; vm_object_t obj; vm_ooffset_t foff; vm_page_t m; if (!(bp->b_flags & B_VMIO)) return; obj = bp->b_bufobj->bo_object; foff = bp->b_offset; KASSERT(bp->b_offset != NOOFFSET, ("vfs_busy_pages: no buffer offset")); VM_OBJECT_WLOCK(obj); vfs_drain_busy_pages(bp); if (bp->b_bufsize != 0) vfs_setdirty_locked_object(bp); bogus = 0; for (i = 0; i < bp->b_npages; i++) { m = bp->b_pages[i]; if ((bp->b_flags & B_CLUSTER) == 0) { vm_object_pip_add(obj, 1); vm_page_sbusy(m); } /* * When readying a buffer for a read ( i.e * clear_modify == 0 ), it is important to do * bogus_page replacement for valid pages in * partially instantiated buffers. Partially * instantiated buffers can, in turn, occur when * reconstituting a buffer from its VM backing store * base. We only have to do this if B_CACHE is * clear ( which causes the I/O to occur in the * first place ). The replacement prevents the read * I/O from overwriting potentially dirty VM-backed * pages. XXX bogus page replacement is, uh, bogus. * It may not work properly with small-block devices. * We need to find a better way. */ if (clear_modify) { pmap_remove_write(m); vfs_page_set_validclean(bp, foff, m); } else if (m->valid == VM_PAGE_BITS_ALL && (bp->b_flags & B_CACHE) == 0) { bp->b_pages[i] = bogus_page; bogus++; } foff = (foff + PAGE_SIZE) & ~(off_t)PAGE_MASK; } VM_OBJECT_WUNLOCK(obj); if (bogus && (bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); pmap_qenter(trunc_page((vm_offset_t)bp->b_data), bp->b_pages, bp->b_npages); } } /* * vfs_bio_set_valid: * * Set the range within the buffer to valid. The range is * relative to the beginning of the buffer, b_offset. Note that * b_offset itself may be offset from the beginning of the first * page. */ void vfs_bio_set_valid(struct buf *bp, int base, int size) { int i, n; vm_page_t m; if (!(bp->b_flags & B_VMIO)) return; /* * Fixup base to be relative to beginning of first page. * Set initial n to be the maximum number of bytes in the * first page that can be validated. */ base += (bp->b_offset & PAGE_MASK); n = PAGE_SIZE - (base & PAGE_MASK); VM_OBJECT_WLOCK(bp->b_bufobj->bo_object); for (i = base / PAGE_SIZE; size > 0 && i < bp->b_npages; ++i) { m = bp->b_pages[i]; if (n > size) n = size; vm_page_set_valid_range(m, base & PAGE_MASK, n); base += n; size -= n; n = PAGE_SIZE; } VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object); } /* * vfs_bio_clrbuf: * * If the specified buffer is a non-VMIO buffer, clear the entire * buffer. If the specified buffer is a VMIO buffer, clear and * validate only the previously invalid portions of the buffer. * This routine essentially fakes an I/O, so we need to clear * BIO_ERROR and B_INVAL. * * Note that while we only theoretically need to clear through b_bcount, * we go ahead and clear through b_bufsize. */ void vfs_bio_clrbuf(struct buf *bp) { int i, j, mask, sa, ea, slide; if ((bp->b_flags & (B_VMIO | B_MALLOC)) != B_VMIO) { clrbuf(bp); return; } bp->b_flags &= ~B_INVAL; bp->b_ioflags &= ~BIO_ERROR; VM_OBJECT_WLOCK(bp->b_bufobj->bo_object); if ((bp->b_npages == 1) && (bp->b_bufsize < PAGE_SIZE) && (bp->b_offset & PAGE_MASK) == 0) { if (bp->b_pages[0] == bogus_page) goto unlock; mask = (1 << (bp->b_bufsize / DEV_BSIZE)) - 1; VM_OBJECT_ASSERT_WLOCKED(bp->b_pages[0]->object); if ((bp->b_pages[0]->valid & mask) == mask) goto unlock; if ((bp->b_pages[0]->valid & mask) == 0) { pmap_zero_page_area(bp->b_pages[0], 0, bp->b_bufsize); bp->b_pages[0]->valid |= mask; goto unlock; } } sa = bp->b_offset & PAGE_MASK; slide = 0; for (i = 0; i < bp->b_npages; i++, sa = 0) { slide = imin(slide + PAGE_SIZE, bp->b_offset + bp->b_bufsize); ea = slide & PAGE_MASK; if (ea == 0) ea = PAGE_SIZE; if (bp->b_pages[i] == bogus_page) continue; j = sa / DEV_BSIZE; mask = ((1 << ((ea - sa) / DEV_BSIZE)) - 1) << j; VM_OBJECT_ASSERT_WLOCKED(bp->b_pages[i]->object); if ((bp->b_pages[i]->valid & mask) == mask) continue; if ((bp->b_pages[i]->valid & mask) == 0) pmap_zero_page_area(bp->b_pages[i], sa, ea - sa); else { for (; sa < ea; sa += DEV_BSIZE, j++) { if ((bp->b_pages[i]->valid & (1 << j)) == 0) { pmap_zero_page_area(bp->b_pages[i], sa, DEV_BSIZE); } } } bp->b_pages[i]->valid |= mask; } unlock: VM_OBJECT_WUNLOCK(bp->b_bufobj->bo_object); bp->b_resid = 0; } void vfs_bio_bzero_buf(struct buf *bp, int base, int size) { vm_page_t m; int i, n; if ((bp->b_flags & B_UNMAPPED) == 0) { BUF_CHECK_MAPPED(bp); bzero(bp->b_data + base, size); } else { BUF_CHECK_UNMAPPED(bp); n = PAGE_SIZE - (base & PAGE_MASK); for (i = base / PAGE_SIZE; size > 0 && i < bp->b_npages; ++i) { m = bp->b_pages[i]; if (n > size) n = size; pmap_zero_page_area(m, base & PAGE_MASK, n); base += n; size -= n; n = PAGE_SIZE; } } } /* * vm_hold_load_pages and vm_hold_free_pages get pages into * a buffers address space. The pages are anonymous and are * not associated with a file object. */ static void vm_hold_load_pages(struct buf *bp, vm_offset_t from, vm_offset_t to) { vm_offset_t pg; vm_page_t p; int index; BUF_CHECK_MAPPED(bp); to = round_page(to); from = round_page(from); index = (from - trunc_page((vm_offset_t)bp->b_data)) >> PAGE_SHIFT; for (pg = from; pg < to; pg += PAGE_SIZE, index++) { tryagain: /* * note: must allocate system pages since blocking here * could interfere with paging I/O, no matter which * process we are. */ p = vm_page_alloc(NULL, 0, VM_ALLOC_SYSTEM | VM_ALLOC_NOOBJ | VM_ALLOC_WIRED | VM_ALLOC_COUNT((to - pg) >> PAGE_SHIFT)); if (p == NULL) { VM_WAIT; goto tryagain; } pmap_qenter(pg, &p, 1); bp->b_pages[index] = p; } bp->b_npages = index; } /* Return pages associated with this buf to the vm system */ static void vm_hold_free_pages(struct buf *bp, int newbsize) { vm_offset_t from; vm_page_t p; int index, newnpages; BUF_CHECK_MAPPED(bp); from = round_page((vm_offset_t)bp->b_data + newbsize); newnpages = (from - trunc_page((vm_offset_t)bp->b_data)) >> PAGE_SHIFT; if (bp->b_npages > newnpages) pmap_qremove(from, bp->b_npages - newnpages); for (index = newnpages; index < bp->b_npages; index++) { p = bp->b_pages[index]; bp->b_pages[index] = NULL; if (vm_page_sbusied(p)) printf("vm_hold_free_pages: blkno: %jd, lblkno: %jd\n", (intmax_t)bp->b_blkno, (intmax_t)bp->b_lblkno); p->wire_count--; vm_page_free(p); atomic_subtract_int(&vm_cnt.v_wire_count, 1); } bp->b_npages = newnpages; } /* * Map an IO request into kernel virtual address space. * * All requests are (re)mapped into kernel VA space. * Notice that we use b_bufsize for the size of the buffer * to be mapped. b_bcount might be modified by the driver. * * Note that even if the caller determines that the address space should * be valid, a race or a smaller-file mapped into a larger space may * actually cause vmapbuf() to fail, so all callers of vmapbuf() MUST * check the return value. */ int vmapbuf(struct buf *bp, int mapbuf) { caddr_t kva; vm_prot_t prot; int pidx; if (bp->b_bufsize < 0) return (-1); prot = VM_PROT_READ; if (bp->b_iocmd == BIO_READ) prot |= VM_PROT_WRITE; /* Less backwards than it looks */ if ((pidx = vm_fault_quick_hold_pages(&curproc->p_vmspace->vm_map, (vm_offset_t)bp->b_data, bp->b_bufsize, prot, bp->b_pages, btoc(MAXPHYS))) < 0) return (-1); bp->b_npages = pidx; if (mapbuf || !unmapped_buf_allowed) { pmap_qenter((vm_offset_t)bp->b_saveaddr, bp->b_pages, pidx); kva = bp->b_saveaddr; bp->b_saveaddr = bp->b_data; bp->b_data = kva + (((vm_offset_t)bp->b_data) & PAGE_MASK); bp->b_flags &= ~B_UNMAPPED; } else { bp->b_flags |= B_UNMAPPED; bp->b_offset = ((vm_offset_t)bp->b_data) & PAGE_MASK; bp->b_saveaddr = bp->b_data; bp->b_data = unmapped_buf; } return(0); } /* * Free the io map PTEs associated with this IO operation. * We also invalidate the TLB entries and restore the original b_addr. */ void vunmapbuf(struct buf *bp) { int npages; npages = bp->b_npages; if (bp->b_flags & B_UNMAPPED) bp->b_flags &= ~B_UNMAPPED; else pmap_qremove(trunc_page((vm_offset_t)bp->b_data), npages); vm_page_unhold_pages(bp->b_pages, npages); bp->b_data = bp->b_saveaddr; } void bdone(struct buf *bp) { struct mtx *mtxp; mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); bp->b_flags |= B_DONE; wakeup(bp); mtx_unlock(mtxp); } void bwait(struct buf *bp, u_char pri, const char *wchan) { struct mtx *mtxp; mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); while ((bp->b_flags & B_DONE) == 0) msleep(bp, mtxp, pri, wchan, 0); mtx_unlock(mtxp); } int bufsync(struct bufobj *bo, int waitfor) { return (VOP_FSYNC(bo->__bo_vnode, waitfor, curthread)); } void bufstrategy(struct bufobj *bo, struct buf *bp) { int i = 0; struct vnode *vp; vp = bp->b_vp; KASSERT(vp == bo->bo_private, ("Inconsistent vnode bufstrategy")); KASSERT(vp->v_type != VCHR && vp->v_type != VBLK, ("Wrong vnode in bufstrategy(bp=%p, vp=%p)", bp, vp)); i = VOP_STRATEGY(vp, bp); KASSERT(i == 0, ("VOP_STRATEGY failed bp=%p vp=%p", bp, bp->b_vp)); } void bufobj_wrefl(struct bufobj *bo) { KASSERT(bo != NULL, ("NULL bo in bufobj_wref")); ASSERT_BO_WLOCKED(bo); bo->bo_numoutput++; } void bufobj_wref(struct bufobj *bo) { KASSERT(bo != NULL, ("NULL bo in bufobj_wref")); BO_LOCK(bo); bo->bo_numoutput++; BO_UNLOCK(bo); } void bufobj_wdrop(struct bufobj *bo) { KASSERT(bo != NULL, ("NULL bo in bufobj_wdrop")); BO_LOCK(bo); KASSERT(bo->bo_numoutput > 0, ("bufobj_wdrop non-positive count")); if ((--bo->bo_numoutput == 0) && (bo->bo_flag & BO_WWAIT)) { bo->bo_flag &= ~BO_WWAIT; wakeup(&bo->bo_numoutput); } BO_UNLOCK(bo); } int bufobj_wwait(struct bufobj *bo, int slpflag, int timeo) { int error; KASSERT(bo != NULL, ("NULL bo in bufobj_wwait")); ASSERT_BO_WLOCKED(bo); error = 0; while (bo->bo_numoutput) { bo->bo_flag |= BO_WWAIT; error = msleep(&bo->bo_numoutput, BO_LOCKPTR(bo), slpflag | (PRIBIO + 1), "bo_wwait", timeo); if (error) break; } return (error); } void bpin(struct buf *bp) { struct mtx *mtxp; mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); bp->b_pin_count++; mtx_unlock(mtxp); } void bunpin(struct buf *bp) { struct mtx *mtxp; mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); if (--bp->b_pin_count == 0) wakeup(bp); mtx_unlock(mtxp); } void bunpin_wait(struct buf *bp) { struct mtx *mtxp; mtxp = mtx_pool_find(mtxpool_sleep, bp); mtx_lock(mtxp); while (bp->b_pin_count > 0) msleep(bp, mtxp, PRIBIO, "bwunpin", 0); mtx_unlock(mtxp); } /* * Set bio_data or bio_ma for struct bio from the struct buf. */ void bdata2bio(struct buf *bp, struct bio *bip) { if ((bp->b_flags & B_UNMAPPED) != 0) { KASSERT(unmapped_buf_allowed, ("unmapped")); bip->bio_ma = bp->b_pages; bip->bio_ma_n = bp->b_npages; bip->bio_data = unmapped_buf; bip->bio_ma_offset = (vm_offset_t)bp->b_offset & PAGE_MASK; bip->bio_flags |= BIO_UNMAPPED; KASSERT(round_page(bip->bio_ma_offset + bip->bio_length) / PAGE_SIZE == bp->b_npages, ("Buffer %p too short: %d %lld %d", bp, bip->bio_ma_offset, (long long)bip->bio_length, bip->bio_ma_n)); } else { bip->bio_data = bp->b_data; bip->bio_ma = NULL; } } #include "opt_ddb.h" #ifdef DDB #include /* DDB command to show buffer data */ DB_SHOW_COMMAND(buffer, db_show_buffer) { /* get args */ struct buf *bp = (struct buf *)addr; if (!have_addr) { db_printf("usage: show buffer \n"); return; } db_printf("buf at %p\n", bp); db_printf("b_flags = 0x%b, b_xflags=0x%b, b_vflags=0x%b\n", (u_int)bp->b_flags, PRINT_BUF_FLAGS, (u_int)bp->b_xflags, PRINT_BUF_XFLAGS, (u_int)bp->b_vflags, PRINT_BUF_VFLAGS); db_printf( "b_error = %d, b_bufsize = %ld, b_bcount = %ld, b_resid = %ld\n" "b_bufobj = (%p), b_data = %p, b_blkno = %jd, b_lblkno = %jd, " "b_dep = %p\n", bp->b_error, bp->b_bufsize, bp->b_bcount, bp->b_resid, bp->b_bufobj, bp->b_data, (intmax_t)bp->b_blkno, (intmax_t)bp->b_lblkno, bp->b_dep.lh_first); if (bp->b_npages) { int i; db_printf("b_npages = %d, pages(OBJ, IDX, PA): ", bp->b_npages); for (i = 0; i < bp->b_npages; i++) { vm_page_t m; m = bp->b_pages[i]; db_printf("(%p, 0x%lx, 0x%lx)", (void *)m->object, (u_long)m->pindex, (u_long)VM_PAGE_TO_PHYS(m)); if ((i + 1) < bp->b_npages) db_printf(","); } db_printf("\n"); } db_printf(" "); BUF_LOCKPRINTINFO(bp); } DB_SHOW_COMMAND(lockedbufs, lockedbufs) { struct buf *bp; int i; for (i = 0; i < nbuf; i++) { bp = &buf[i]; if (BUF_ISLOCKED(bp)) { db_show_buffer((uintptr_t)bp, 1, 0, NULL); db_printf("\n"); } } } DB_SHOW_COMMAND(vnodebufs, db_show_vnodebufs) { struct vnode *vp; struct buf *bp; if (!have_addr) { db_printf("usage: show vnodebufs \n"); return; } vp = (struct vnode *)addr; db_printf("Clean buffers:\n"); TAILQ_FOREACH(bp, &vp->v_bufobj.bo_clean.bv_hd, b_bobufs) { db_show_buffer((uintptr_t)bp, 1, 0, NULL); db_printf("\n"); } db_printf("Dirty buffers:\n"); TAILQ_FOREACH(bp, &vp->v_bufobj.bo_dirty.bv_hd, b_bobufs) { db_show_buffer((uintptr_t)bp, 1, 0, NULL); db_printf("\n"); } } DB_COMMAND(countfreebufs, db_coundfreebufs) { struct buf *bp; int i, used = 0, nfree = 0; if (have_addr) { db_printf("usage: countfreebufs\n"); return; } for (i = 0; i < nbuf; i++) { bp = &buf[i]; if ((bp->b_flags & B_INFREECNT) != 0) nfree++; else used++; } db_printf("Counted %d free, %d used (%d tot)\n", nfree, used, nfree + used); db_printf("numfreebuffers is %d\n", numfreebuffers); } #endif /* DDB */ Index: projects/building-blocks/sys/kern/vfs_mount.c =================================================================== --- projects/building-blocks/sys/kern/vfs_mount.c (revision 275663) +++ projects/building-blocks/sys/kern/vfs_mount.c (revision 275664) @@ -1,1935 +1,1935 @@ /*- * Copyright (c) 1999-2004 Poul-Henning Kamp * Copyright (c) 1999 Michael Smith * Copyright (c) 1989, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define VFS_MOUNTARG_SIZE_MAX (1024 * 64) static int vfs_domount(struct thread *td, const char *fstype, char *fspath, uint64_t fsflags, struct vfsoptlist **optlist); static void free_mntarg(struct mntarg *ma); static int usermount = 0; SYSCTL_INT(_vfs, OID_AUTO, usermount, CTLFLAG_RW, &usermount, 0, "Unprivileged users may mount and unmount file systems"); MALLOC_DEFINE(M_MOUNT, "mount", "vfs mount structure"); static uma_zone_t mount_zone; /* List of mounted filesystems. */ struct mntlist mountlist = TAILQ_HEAD_INITIALIZER(mountlist); /* For any iteration/modification of mountlist */ struct mtx mountlist_mtx; MTX_SYSINIT(mountlist, &mountlist_mtx, "mountlist", MTX_DEF); /* * Global opts, taken by all filesystems */ static const char *global_opts[] = { "errmsg", "fstype", "fspath", "ro", "rw", "nosuid", "noexec", NULL }; static int mount_init(void *mem, int size, int flags) { struct mount *mp; mp = (struct mount *)mem; mtx_init(&mp->mnt_mtx, "struct mount mtx", NULL, MTX_DEF); lockinit(&mp->mnt_explock, PVFS, "explock", 0, 0); return (0); } static void mount_fini(void *mem, int size) { struct mount *mp; mp = (struct mount *)mem; lockdestroy(&mp->mnt_explock); mtx_destroy(&mp->mnt_mtx); } static void vfs_mount_init(void *dummy __unused) { mount_zone = uma_zcreate("Mountpoints", sizeof(struct mount), NULL, NULL, mount_init, mount_fini, UMA_ALIGN_PTR, UMA_ZONE_NOFREE); } SYSINIT(vfs_mount, SI_SUB_VFS, SI_ORDER_ANY, vfs_mount_init, NULL); /* * --------------------------------------------------------------------- * Functions for building and sanitizing the mount options */ /* Remove one mount option. */ static void vfs_freeopt(struct vfsoptlist *opts, struct vfsopt *opt) { TAILQ_REMOVE(opts, opt, link); free(opt->name, M_MOUNT); if (opt->value != NULL) free(opt->value, M_MOUNT); free(opt, M_MOUNT); } /* Release all resources related to the mount options. */ void vfs_freeopts(struct vfsoptlist *opts) { struct vfsopt *opt; while (!TAILQ_EMPTY(opts)) { opt = TAILQ_FIRST(opts); vfs_freeopt(opts, opt); } free(opts, M_MOUNT); } void vfs_deleteopt(struct vfsoptlist *opts, const char *name) { struct vfsopt *opt, *temp; if (opts == NULL) return; TAILQ_FOREACH_SAFE(opt, opts, link, temp) { if (strcmp(opt->name, name) == 0) vfs_freeopt(opts, opt); } } static int vfs_isopt_ro(const char *opt) { if (strcmp(opt, "ro") == 0 || strcmp(opt, "rdonly") == 0 || strcmp(opt, "norw") == 0) return (1); return (0); } static int vfs_isopt_rw(const char *opt) { if (strcmp(opt, "rw") == 0 || strcmp(opt, "noro") == 0) return (1); return (0); } /* * Check if options are equal (with or without the "no" prefix). */ static int vfs_equalopts(const char *opt1, const char *opt2) { char *p; /* "opt" vs. "opt" or "noopt" vs. "noopt" */ if (strcmp(opt1, opt2) == 0) return (1); /* "noopt" vs. "opt" */ if (strncmp(opt1, "no", 2) == 0 && strcmp(opt1 + 2, opt2) == 0) return (1); /* "opt" vs. "noopt" */ if (strncmp(opt2, "no", 2) == 0 && strcmp(opt1, opt2 + 2) == 0) return (1); while ((p = strchr(opt1, '.')) != NULL && !strncmp(opt1, opt2, ++p - opt1)) { opt2 += p - opt1; opt1 = p; /* "foo.noopt" vs. "foo.opt" */ if (strncmp(opt1, "no", 2) == 0 && strcmp(opt1 + 2, opt2) == 0) return (1); /* "foo.opt" vs. "foo.noopt" */ if (strncmp(opt2, "no", 2) == 0 && strcmp(opt1, opt2 + 2) == 0) return (1); } /* "ro" / "rdonly" / "norw" / "rw" / "noro" */ if ((vfs_isopt_ro(opt1) || vfs_isopt_rw(opt1)) && (vfs_isopt_ro(opt2) || vfs_isopt_rw(opt2))) return (1); return (0); } /* * If a mount option is specified several times, * (with or without the "no" prefix) only keep * the last occurrence of it. */ static void vfs_sanitizeopts(struct vfsoptlist *opts) { struct vfsopt *opt, *opt2, *tmp; TAILQ_FOREACH_REVERSE(opt, opts, vfsoptlist, link) { opt2 = TAILQ_PREV(opt, vfsoptlist, link); while (opt2 != NULL) { if (vfs_equalopts(opt->name, opt2->name)) { tmp = TAILQ_PREV(opt2, vfsoptlist, link); vfs_freeopt(opts, opt2); opt2 = tmp; } else { opt2 = TAILQ_PREV(opt2, vfsoptlist, link); } } } } /* * Build a linked list of mount options from a struct uio. */ int vfs_buildopts(struct uio *auio, struct vfsoptlist **options) { struct vfsoptlist *opts; struct vfsopt *opt; size_t memused, namelen, optlen; unsigned int i, iovcnt; int error; opts = malloc(sizeof(struct vfsoptlist), M_MOUNT, M_WAITOK); TAILQ_INIT(opts); memused = 0; iovcnt = auio->uio_iovcnt; for (i = 0; i < iovcnt; i += 2) { namelen = auio->uio_iov[i].iov_len; optlen = auio->uio_iov[i + 1].iov_len; memused += sizeof(struct vfsopt) + optlen + namelen; /* * Avoid consuming too much memory, and attempts to overflow * memused. */ if (memused > VFS_MOUNTARG_SIZE_MAX || optlen > VFS_MOUNTARG_SIZE_MAX || namelen > VFS_MOUNTARG_SIZE_MAX) { error = EINVAL; goto bad; } opt = malloc(sizeof(struct vfsopt), M_MOUNT, M_WAITOK); opt->name = malloc(namelen, M_MOUNT, M_WAITOK); opt->value = NULL; opt->len = 0; opt->pos = i / 2; opt->seen = 0; /* * Do this early, so jumps to "bad" will free the current * option. */ TAILQ_INSERT_TAIL(opts, opt, link); if (auio->uio_segflg == UIO_SYSSPACE) { bcopy(auio->uio_iov[i].iov_base, opt->name, namelen); } else { error = copyin(auio->uio_iov[i].iov_base, opt->name, namelen); if (error) goto bad; } /* Ensure names are null-terminated strings. */ if (namelen == 0 || opt->name[namelen - 1] != '\0') { error = EINVAL; goto bad; } if (optlen != 0) { opt->len = optlen; opt->value = malloc(optlen, M_MOUNT, M_WAITOK); if (auio->uio_segflg == UIO_SYSSPACE) { bcopy(auio->uio_iov[i + 1].iov_base, opt->value, optlen); } else { error = copyin(auio->uio_iov[i + 1].iov_base, opt->value, optlen); if (error) goto bad; } } } vfs_sanitizeopts(opts); *options = opts; return (0); bad: vfs_freeopts(opts); return (error); } /* * Merge the old mount options with the new ones passed * in the MNT_UPDATE case. * * XXX: This function will keep a "nofoo" option in the new * options. E.g, if the option's canonical name is "foo", * "nofoo" ends up in the mount point's active options. */ static void vfs_mergeopts(struct vfsoptlist *toopts, struct vfsoptlist *oldopts) { struct vfsopt *opt, *new; TAILQ_FOREACH(opt, oldopts, link) { new = malloc(sizeof(struct vfsopt), M_MOUNT, M_WAITOK); new->name = strdup(opt->name, M_MOUNT); if (opt->len != 0) { new->value = malloc(opt->len, M_MOUNT, M_WAITOK); bcopy(opt->value, new->value, opt->len); } else new->value = NULL; new->len = opt->len; new->seen = opt->seen; TAILQ_INSERT_HEAD(toopts, new, link); } vfs_sanitizeopts(toopts); } /* * Mount a filesystem. */ int sys_nmount(td, uap) struct thread *td; struct nmount_args /* { struct iovec *iovp; unsigned int iovcnt; int flags; } */ *uap; { struct uio *auio; int error; u_int iovcnt; uint64_t flags; /* * Mount flags are now 64-bits. On 32-bit archtectures only * 32-bits are passed in, but from here on everything handles * 64-bit flags correctly. */ flags = uap->flags; AUDIT_ARG_FFLAGS(flags); CTR4(KTR_VFS, "%s: iovp %p with iovcnt %d and flags %d", __func__, uap->iovp, uap->iovcnt, flags); /* * Filter out MNT_ROOTFS. We do not want clients of nmount() in * userspace to set this flag, but we must filter it out if we want * MNT_UPDATE on the root file system to work. * MNT_ROOTFS should only be set by the kernel when mounting its * root file system. */ flags &= ~MNT_ROOTFS; iovcnt = uap->iovcnt; /* * Check that we have an even number of iovec's * and that we have at least two options. */ if ((iovcnt & 1) || (iovcnt < 4)) { CTR2(KTR_VFS, "%s: failed for invalid iovcnt %d", __func__, uap->iovcnt); return (EINVAL); } error = copyinuio(uap->iovp, iovcnt, &auio); if (error) { CTR2(KTR_VFS, "%s: failed for invalid uio op with %d errno", __func__, error); return (error); } error = vfs_donmount(td, flags, auio); free(auio, M_IOV); return (error); } /* * --------------------------------------------------------------------- * Various utility functions */ void vfs_ref(struct mount *mp) { CTR2(KTR_VFS, "%s: mp %p", __func__, mp); MNT_ILOCK(mp); MNT_REF(mp); MNT_IUNLOCK(mp); } void vfs_rel(struct mount *mp) { CTR2(KTR_VFS, "%s: mp %p", __func__, mp); MNT_ILOCK(mp); MNT_REL(mp); MNT_IUNLOCK(mp); } /* * Allocate and initialize the mount point struct. */ struct mount * vfs_mount_alloc(struct vnode *vp, struct vfsconf *vfsp, const char *fspath, struct ucred *cred) { struct mount *mp; mp = uma_zalloc(mount_zone, M_WAITOK); bzero(&mp->mnt_startzero, __rangeof(struct mount, mnt_startzero, mnt_endzero)); TAILQ_INIT(&mp->mnt_nvnodelist); mp->mnt_nvnodelistsize = 0; TAILQ_INIT(&mp->mnt_activevnodelist); mp->mnt_activevnodelistsize = 0; mp->mnt_ref = 0; (void) vfs_busy(mp, MBF_NOWAIT); atomic_add_acq_int(&vfsp->vfc_refcount, 1); mp->mnt_op = vfsp->vfc_vfsops; mp->mnt_vfc = vfsp; mp->mnt_stat.f_type = vfsp->vfc_typenum; mp->mnt_gen++; strlcpy(mp->mnt_stat.f_fstypename, vfsp->vfc_name, MFSNAMELEN); mp->mnt_vnodecovered = vp; mp->mnt_cred = crdup(cred); mp->mnt_stat.f_owner = cred->cr_uid; strlcpy(mp->mnt_stat.f_mntonname, fspath, MNAMELEN); mp->mnt_iosize_max = DFLTPHYS; #ifdef MAC mac_mount_init(mp); mac_mount_create(cred, mp); #endif arc4rand(&mp->mnt_hashseed, sizeof mp->mnt_hashseed, 0); TAILQ_INIT(&mp->mnt_uppers); return (mp); } /* * Destroy the mount struct previously allocated by vfs_mount_alloc(). */ void vfs_mount_destroy(struct mount *mp) { MNT_ILOCK(mp); mp->mnt_kern_flag |= MNTK_REFEXPIRE; if (mp->mnt_kern_flag & MNTK_MWAIT) { mp->mnt_kern_flag &= ~MNTK_MWAIT; wakeup(mp); } while (mp->mnt_ref) msleep(mp, MNT_MTX(mp), PVFS, "mntref", 0); KASSERT(mp->mnt_ref == 0, ("%s: invalid refcount in the drain path @ %s:%d", __func__, __FILE__, __LINE__)); if (mp->mnt_writeopcount != 0) panic("vfs_mount_destroy: nonzero writeopcount"); if (mp->mnt_secondary_writes != 0) panic("vfs_mount_destroy: nonzero secondary_writes"); atomic_subtract_rel_int(&mp->mnt_vfc->vfc_refcount, 1); if (!TAILQ_EMPTY(&mp->mnt_nvnodelist)) { struct vnode *vp; TAILQ_FOREACH(vp, &mp->mnt_nvnodelist, v_nmntvnodes) vprint("", vp); panic("unmount: dangling vnode"); } KASSERT(TAILQ_EMPTY(&mp->mnt_uppers), ("mnt_uppers")); if (mp->mnt_nvnodelistsize != 0) panic("vfs_mount_destroy: nonzero nvnodelistsize"); if (mp->mnt_activevnodelistsize != 0) panic("vfs_mount_destroy: nonzero activevnodelistsize"); if (mp->mnt_lockref != 0) panic("vfs_mount_destroy: nonzero lock refcount"); MNT_IUNLOCK(mp); #ifdef MAC mac_mount_destroy(mp); #endif if (mp->mnt_opt != NULL) vfs_freeopts(mp->mnt_opt); crfree(mp->mnt_cred); uma_zfree(mount_zone, mp); } int vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions) { struct vfsoptlist *optlist; struct vfsopt *opt, *tmp_opt; char *fstype, *fspath, *errmsg; int error, fstypelen, fspathlen, errmsg_len, errmsg_pos; errmsg = fspath = NULL; errmsg_len = fspathlen = 0; errmsg_pos = -1; error = vfs_buildopts(fsoptions, &optlist); if (error) return (error); if (vfs_getopt(optlist, "errmsg", (void **)&errmsg, &errmsg_len) == 0) errmsg_pos = vfs_getopt_pos(optlist, "errmsg"); /* * We need these two options before the others, * and they are mandatory for any filesystem. * Ensure they are NUL terminated as well. */ fstypelen = 0; error = vfs_getopt(optlist, "fstype", (void **)&fstype, &fstypelen); if (error || fstype[fstypelen - 1] != '\0') { error = EINVAL; if (errmsg != NULL) strncpy(errmsg, "Invalid fstype", errmsg_len); goto bail; } fspathlen = 0; error = vfs_getopt(optlist, "fspath", (void **)&fspath, &fspathlen); if (error || fspath[fspathlen - 1] != '\0') { error = EINVAL; if (errmsg != NULL) strncpy(errmsg, "Invalid fspath", errmsg_len); goto bail; } /* * We need to see if we have the "update" option * before we call vfs_domount(), since vfs_domount() has special * logic based on MNT_UPDATE. This is very important * when we want to update the root filesystem. */ TAILQ_FOREACH_SAFE(opt, optlist, link, tmp_opt) { if (strcmp(opt->name, "update") == 0) { fsflags |= MNT_UPDATE; vfs_freeopt(optlist, opt); } else if (strcmp(opt->name, "async") == 0) fsflags |= MNT_ASYNC; else if (strcmp(opt->name, "force") == 0) { fsflags |= MNT_FORCE; vfs_freeopt(optlist, opt); } else if (strcmp(opt->name, "reload") == 0) { fsflags |= MNT_RELOAD; vfs_freeopt(optlist, opt); } else if (strcmp(opt->name, "multilabel") == 0) fsflags |= MNT_MULTILABEL; else if (strcmp(opt->name, "noasync") == 0) fsflags &= ~MNT_ASYNC; else if (strcmp(opt->name, "noatime") == 0) fsflags |= MNT_NOATIME; else if (strcmp(opt->name, "atime") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("nonoatime", M_MOUNT); } else if (strcmp(opt->name, "noclusterr") == 0) fsflags |= MNT_NOCLUSTERR; else if (strcmp(opt->name, "clusterr") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("nonoclusterr", M_MOUNT); } else if (strcmp(opt->name, "noclusterw") == 0) fsflags |= MNT_NOCLUSTERW; else if (strcmp(opt->name, "clusterw") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("nonoclusterw", M_MOUNT); } else if (strcmp(opt->name, "noexec") == 0) fsflags |= MNT_NOEXEC; else if (strcmp(opt->name, "exec") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("nonoexec", M_MOUNT); } else if (strcmp(opt->name, "nosuid") == 0) fsflags |= MNT_NOSUID; else if (strcmp(opt->name, "suid") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("nonosuid", M_MOUNT); } else if (strcmp(opt->name, "nosymfollow") == 0) fsflags |= MNT_NOSYMFOLLOW; else if (strcmp(opt->name, "symfollow") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("nonosymfollow", M_MOUNT); } else if (strcmp(opt->name, "noro") == 0) fsflags &= ~MNT_RDONLY; else if (strcmp(opt->name, "rw") == 0) fsflags &= ~MNT_RDONLY; else if (strcmp(opt->name, "ro") == 0) fsflags |= MNT_RDONLY; else if (strcmp(opt->name, "rdonly") == 0) { free(opt->name, M_MOUNT); opt->name = strdup("ro", M_MOUNT); fsflags |= MNT_RDONLY; } else if (strcmp(opt->name, "suiddir") == 0) fsflags |= MNT_SUIDDIR; else if (strcmp(opt->name, "sync") == 0) fsflags |= MNT_SYNCHRONOUS; else if (strcmp(opt->name, "union") == 0) fsflags |= MNT_UNION; else if (strcmp(opt->name, "automounted") == 0) { fsflags |= MNT_AUTOMOUNTED; vfs_freeopt(optlist, opt); } } /* * Be ultra-paranoid about making sure the type and fspath * variables will fit in our mp buffers, including the * terminating NUL. */ if (fstypelen > MFSNAMELEN || fspathlen > MNAMELEN) { error = ENAMETOOLONG; goto bail; } error = vfs_domount(td, fstype, fspath, fsflags, &optlist); bail: /* copyout the errmsg */ if (errmsg_pos != -1 && ((2 * errmsg_pos + 1) < fsoptions->uio_iovcnt) && errmsg_len > 0 && errmsg != NULL) { if (fsoptions->uio_segflg == UIO_SYSSPACE) { bcopy(errmsg, fsoptions->uio_iov[2 * errmsg_pos + 1].iov_base, fsoptions->uio_iov[2 * errmsg_pos + 1].iov_len); } else { copyout(errmsg, fsoptions->uio_iov[2 * errmsg_pos + 1].iov_base, fsoptions->uio_iov[2 * errmsg_pos + 1].iov_len); } } if (optlist != NULL) vfs_freeopts(optlist); return (error); } /* * Old mount API. */ #ifndef _SYS_SYSPROTO_H_ struct mount_args { char *type; char *path; int flags; caddr_t data; }; #endif /* ARGSUSED */ int sys_mount(td, uap) struct thread *td; struct mount_args /* { char *type; char *path; int flags; caddr_t data; } */ *uap; { char *fstype; struct vfsconf *vfsp = NULL; struct mntarg *ma = NULL; uint64_t flags; int error; /* * Mount flags are now 64-bits. On 32-bit architectures only * 32-bits are passed in, but from here on everything handles * 64-bit flags correctly. */ flags = uap->flags; AUDIT_ARG_FFLAGS(flags); /* * Filter out MNT_ROOTFS. We do not want clients of mount() in * userspace to set this flag, but we must filter it out if we want * MNT_UPDATE on the root file system to work. * MNT_ROOTFS should only be set by the kernel when mounting its * root file system. */ flags &= ~MNT_ROOTFS; fstype = malloc(MFSNAMELEN, M_TEMP, M_WAITOK); error = copyinstr(uap->type, fstype, MFSNAMELEN, NULL); if (error) { free(fstype, M_TEMP); return (error); } AUDIT_ARG_TEXT(fstype); vfsp = vfs_byname_kld(fstype, td, &error); free(fstype, M_TEMP); if (vfsp == NULL) return (ENOENT); if (vfsp->vfc_vfsops->vfs_cmount == NULL) return (EOPNOTSUPP); ma = mount_argsu(ma, "fstype", uap->type, MFSNAMELEN); ma = mount_argsu(ma, "fspath", uap->path, MNAMELEN); ma = mount_argb(ma, flags & MNT_RDONLY, "noro"); ma = mount_argb(ma, !(flags & MNT_NOSUID), "nosuid"); ma = mount_argb(ma, !(flags & MNT_NOEXEC), "noexec"); error = vfsp->vfc_vfsops->vfs_cmount(ma, uap->data, flags); return (error); } /* * vfs_domount_first(): first file system mount (not update) */ static int vfs_domount_first( struct thread *td, /* Calling thread. */ struct vfsconf *vfsp, /* File system type. */ char *fspath, /* Mount path. */ struct vnode *vp, /* Vnode to be covered. */ uint64_t fsflags, /* Flags common to all filesystems. */ struct vfsoptlist **optlist /* Options local to the filesystem. */ ) { struct vattr va; struct mount *mp; struct vnode *newdp; int error; ASSERT_VOP_ELOCKED(vp, __func__); KASSERT((fsflags & MNT_UPDATE) == 0, ("MNT_UPDATE shouldn't be here")); /* * If the user is not root, ensure that they own the directory * onto which we are attempting to mount. */ error = VOP_GETATTR(vp, &va, td->td_ucred); if (error == 0 && va.va_uid != td->td_ucred->cr_uid) error = priv_check_cred(td->td_ucred, PRIV_VFS_ADMIN, 0); if (error == 0) error = vinvalbuf(vp, V_SAVE, 0, 0); if (error == 0 && vp->v_type != VDIR) error = ENOTDIR; if (error == 0) { VI_LOCK(vp); if ((vp->v_iflag & VI_MOUNT) == 0 && vp->v_mountedhere == NULL) vp->v_iflag |= VI_MOUNT; else error = EBUSY; VI_UNLOCK(vp); } if (error != 0) { vput(vp); return (error); } VOP_UNLOCK(vp, 0); /* Allocate and initialize the filesystem. */ mp = vfs_mount_alloc(vp, vfsp, fspath, td->td_ucred); /* XXXMAC: pass to vfs_mount_alloc? */ mp->mnt_optnew = *optlist; /* Set the mount level flags. */ mp->mnt_flag = (fsflags & (MNT_UPDATEMASK | MNT_ROOTFS | MNT_RDONLY)); /* * Mount the filesystem. * XXX The final recipients of VFS_MOUNT just overwrite the ndp they * get. No freeing of cn_pnbuf. */ error = VFS_MOUNT(mp); if (error != 0) { vfs_unbusy(mp); vfs_mount_destroy(mp); VI_LOCK(vp); vp->v_iflag &= ~VI_MOUNT; VI_UNLOCK(vp); vrele(vp); return (error); } if (mp->mnt_opt != NULL) vfs_freeopts(mp->mnt_opt); mp->mnt_opt = mp->mnt_optnew; *optlist = NULL; (void)VFS_STATFS(mp, &mp->mnt_stat); /* * Prevent external consumers of mount options from reading mnt_optnew. */ mp->mnt_optnew = NULL; MNT_ILOCK(mp); if ((mp->mnt_flag & MNT_ASYNC) != 0 && (mp->mnt_kern_flag & MNTK_NOASYNC) == 0) mp->mnt_kern_flag |= MNTK_ASYNC; else mp->mnt_kern_flag &= ~MNTK_ASYNC; MNT_IUNLOCK(mp); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); cache_purge(vp); VI_LOCK(vp); vp->v_iflag &= ~VI_MOUNT; VI_UNLOCK(vp); vp->v_mountedhere = mp; /* Place the new filesystem at the end of the mount list. */ mtx_lock(&mountlist_mtx); TAILQ_INSERT_TAIL(&mountlist, mp, mnt_list); mtx_unlock(&mountlist_mtx); vfs_event_signal(NULL, VQ_MOUNT, 0); if (VFS_ROOT(mp, LK_EXCLUSIVE, &newdp)) panic("mount: lost mount"); VOP_UNLOCK(vp, 0); EVENTHANDLER_INVOKE(vfs_mounted, mp, newdp, td); VOP_UNLOCK(newdp, 0); mountcheckdirs(vp, newdp); vrele(newdp); if ((mp->mnt_flag & MNT_RDONLY) == 0) vfs_allocate_syncvnode(mp); vfs_unbusy(mp); return (0); } /* * vfs_domount_update(): update of mounted file system */ static int vfs_domount_update( struct thread *td, /* Calling thread. */ struct vnode *vp, /* Mount point vnode. */ uint64_t fsflags, /* Flags common to all filesystems. */ struct vfsoptlist **optlist /* Options local to the filesystem. */ ) { struct oexport_args oexport; struct export_args export; struct mount *mp; int error, export_error; uint64_t flag; ASSERT_VOP_ELOCKED(vp, __func__); KASSERT((fsflags & MNT_UPDATE) != 0, ("MNT_UPDATE should be here")); if ((vp->v_vflag & VV_ROOT) == 0) { vput(vp); return (EINVAL); } mp = vp->v_mount; /* * We only allow the filesystem to be reloaded if it * is currently mounted read-only. */ flag = mp->mnt_flag; if ((fsflags & MNT_RELOAD) != 0 && (flag & MNT_RDONLY) == 0) { vput(vp); return (EOPNOTSUPP); /* Needs translation */ } /* * Only privileged root, or (if MNT_USER is set) the user that * did the original mount is permitted to update it. */ error = vfs_suser(mp, td); if (error != 0) { vput(vp); return (error); } if (vfs_busy(mp, MBF_NOWAIT)) { vput(vp); return (EBUSY); } VI_LOCK(vp); if ((vp->v_iflag & VI_MOUNT) != 0 || vp->v_mountedhere != NULL) { VI_UNLOCK(vp); vfs_unbusy(mp); vput(vp); return (EBUSY); } vp->v_iflag |= VI_MOUNT; VI_UNLOCK(vp); VOP_UNLOCK(vp, 0); MNT_ILOCK(mp); mp->mnt_flag &= ~MNT_UPDATEMASK; mp->mnt_flag |= fsflags & (MNT_RELOAD | MNT_FORCE | MNT_UPDATE | MNT_SNAPSHOT | MNT_ROOTFS | MNT_UPDATEMASK | MNT_RDONLY); if ((mp->mnt_flag & MNT_ASYNC) == 0) mp->mnt_kern_flag &= ~MNTK_ASYNC; MNT_IUNLOCK(mp); mp->mnt_optnew = *optlist; vfs_mergeopts(mp->mnt_optnew, mp->mnt_opt); /* * Mount the filesystem. * XXX The final recipients of VFS_MOUNT just overwrite the ndp they * get. No freeing of cn_pnbuf. */ error = VFS_MOUNT(mp); export_error = 0; if (error == 0) { /* Process the export option. */ if (vfs_copyopt(mp->mnt_optnew, "export", &export, sizeof(export)) == 0) { export_error = vfs_export(mp, &export); } else if (vfs_copyopt(mp->mnt_optnew, "export", &oexport, sizeof(oexport)) == 0) { export.ex_flags = oexport.ex_flags; export.ex_root = oexport.ex_root; export.ex_anon = oexport.ex_anon; export.ex_addr = oexport.ex_addr; export.ex_addrlen = oexport.ex_addrlen; export.ex_mask = oexport.ex_mask; export.ex_masklen = oexport.ex_masklen; export.ex_indexfile = oexport.ex_indexfile; export.ex_numsecflavors = 0; export_error = vfs_export(mp, &export); } } MNT_ILOCK(mp); if (error == 0) { mp->mnt_flag &= ~(MNT_UPDATE | MNT_RELOAD | MNT_FORCE | MNT_SNAPSHOT); } else { /* * If we fail, restore old mount flags. MNT_QUOTA is special, * because it is not part of MNT_UPDATEMASK, but it could have * changed in the meantime if quotactl(2) was called. * All in all we want current value of MNT_QUOTA, not the old * one. */ mp->mnt_flag = (mp->mnt_flag & MNT_QUOTA) | (flag & ~MNT_QUOTA); } if ((mp->mnt_flag & MNT_ASYNC) != 0 && (mp->mnt_kern_flag & MNTK_NOASYNC) == 0) mp->mnt_kern_flag |= MNTK_ASYNC; else mp->mnt_kern_flag &= ~MNTK_ASYNC; MNT_IUNLOCK(mp); if (error != 0) goto end; if (mp->mnt_opt != NULL) vfs_freeopts(mp->mnt_opt); mp->mnt_opt = mp->mnt_optnew; *optlist = NULL; (void)VFS_STATFS(mp, &mp->mnt_stat); /* * Prevent external consumers of mount options from reading * mnt_optnew. */ mp->mnt_optnew = NULL; if ((mp->mnt_flag & MNT_RDONLY) == 0) vfs_allocate_syncvnode(mp); else vfs_deallocate_syncvnode(mp); end: vfs_unbusy(mp); VI_LOCK(vp); vp->v_iflag &= ~VI_MOUNT; VI_UNLOCK(vp); vrele(vp); return (error != 0 ? error : export_error); } /* * vfs_domount(): actually attempt a filesystem mount. */ static int vfs_domount( struct thread *td, /* Calling thread. */ const char *fstype, /* Filesystem type. */ char *fspath, /* Mount path. */ uint64_t fsflags, /* Flags common to all filesystems. */ struct vfsoptlist **optlist /* Options local to the filesystem. */ ) { struct vfsconf *vfsp; struct nameidata nd; struct vnode *vp; char *pathbuf; int error; /* * Be ultra-paranoid about making sure the type and fspath * variables will fit in our mp buffers, including the * terminating NUL. */ if (strlen(fstype) >= MFSNAMELEN || strlen(fspath) >= MNAMELEN) return (ENAMETOOLONG); if (jailed(td->td_ucred) || usermount == 0) { if ((error = priv_check(td, PRIV_VFS_MOUNT)) != 0) return (error); } /* * Do not allow NFS export or MNT_SUIDDIR by unprivileged users. */ if (fsflags & MNT_EXPORTED) { error = priv_check(td, PRIV_VFS_MOUNT_EXPORTED); if (error) return (error); } if (fsflags & MNT_SUIDDIR) { error = priv_check(td, PRIV_VFS_MOUNT_SUIDDIR); if (error) return (error); } /* * Silently enforce MNT_NOSUID and MNT_USER for unprivileged users. */ if ((fsflags & (MNT_NOSUID | MNT_USER)) != (MNT_NOSUID | MNT_USER)) { if (priv_check(td, PRIV_VFS_MOUNT_NONUSER) != 0) fsflags |= MNT_NOSUID | MNT_USER; } /* Load KLDs before we lock the covered vnode to avoid reversals. */ vfsp = NULL; if ((fsflags & MNT_UPDATE) == 0) { /* Don't try to load KLDs if we're mounting the root. */ if (fsflags & MNT_ROOTFS) vfsp = vfs_byname(fstype); else vfsp = vfs_byname_kld(fstype, td, &error); if (vfsp == NULL) return (ENODEV); if (jailed(td->td_ucred) && !(vfsp->vfc_flags & VFCF_JAIL)) return (EPERM); } /* * Get vnode to be covered or mount point's vnode in case of MNT_UPDATE. */ NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | AUDITVNODE1, UIO_SYSSPACE, fspath, td); error = namei(&nd); if (error != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); vp = nd.ni_vp; if ((fsflags & MNT_UPDATE) == 0) { pathbuf = malloc(MNAMELEN, M_TEMP, M_WAITOK); strcpy(pathbuf, fspath); error = vn_path_to_global_path(td, vp, pathbuf, MNAMELEN); /* debug.disablefullpath == 1 results in ENODEV */ if (error == 0 || error == ENODEV) { error = vfs_domount_first(td, vfsp, pathbuf, vp, fsflags, optlist); } free(pathbuf, M_TEMP); } else error = vfs_domount_update(td, vp, fsflags, optlist); ASSERT_VI_UNLOCKED(vp, __func__); ASSERT_VOP_UNLOCKED(vp, __func__); return (error); } /* * Unmount a filesystem. * * Note: unmount takes a path to the vnode mounted on as argument, not * special file (as before). */ #ifndef _SYS_SYSPROTO_H_ struct unmount_args { char *path; int flags; }; #endif /* ARGSUSED */ int sys_unmount(td, uap) struct thread *td; register struct unmount_args /* { char *path; int flags; } */ *uap; { struct nameidata nd; struct mount *mp; char *pathbuf; int error, id0, id1; AUDIT_ARG_VALUE(uap->flags); if (jailed(td->td_ucred) || usermount == 0) { error = priv_check(td, PRIV_VFS_UNMOUNT); if (error) return (error); } pathbuf = malloc(MNAMELEN, M_TEMP, M_WAITOK); error = copyinstr(uap->path, pathbuf, MNAMELEN, NULL); if (error) { free(pathbuf, M_TEMP); return (error); } if (uap->flags & MNT_BYFSID) { AUDIT_ARG_TEXT(pathbuf); /* Decode the filesystem ID. */ if (sscanf(pathbuf, "FSID:%d:%d", &id0, &id1) != 2) { free(pathbuf, M_TEMP); return (EINVAL); } mtx_lock(&mountlist_mtx); TAILQ_FOREACH_REVERSE(mp, &mountlist, mntlist, mnt_list) { if (mp->mnt_stat.f_fsid.val[0] == id0 && mp->mnt_stat.f_fsid.val[1] == id1) break; } mtx_unlock(&mountlist_mtx); } else { /* * Try to find global path for path argument. */ NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | AUDITVNODE1, UIO_SYSSPACE, pathbuf, td); if (namei(&nd) == 0) { NDFREE(&nd, NDF_ONLY_PNBUF); error = vn_path_to_global_path(td, nd.ni_vp, pathbuf, MNAMELEN); if (error == 0 || error == ENODEV) vput(nd.ni_vp); } mtx_lock(&mountlist_mtx); TAILQ_FOREACH_REVERSE(mp, &mountlist, mntlist, mnt_list) { if (strcmp(mp->mnt_stat.f_mntonname, pathbuf) == 0) break; } mtx_unlock(&mountlist_mtx); } free(pathbuf, M_TEMP); if (mp == NULL) { /* * Previously we returned ENOENT for a nonexistent path and * EINVAL for a non-mountpoint. We cannot tell these apart * now, so in the !MNT_BYFSID case return the more likely * EINVAL for compatibility. */ return ((uap->flags & MNT_BYFSID) ? ENOENT : EINVAL); } /* * Don't allow unmounting the root filesystem. */ if (mp->mnt_flag & MNT_ROOTFS) return (EINVAL); error = dounmount(mp, uap->flags, td); return (error); } /* * Do the actual filesystem unmount. */ int dounmount(mp, flags, td) struct mount *mp; int flags; struct thread *td; { struct vnode *coveredvp, *fsrootvp; int error; uint64_t async_flag; int mnt_gen_r; if ((coveredvp = mp->mnt_vnodecovered) != NULL) { mnt_gen_r = mp->mnt_gen; VI_LOCK(coveredvp); vholdl(coveredvp); vn_lock(coveredvp, LK_EXCLUSIVE | LK_INTERLOCK | LK_RETRY); vdrop(coveredvp); /* * Check for mp being unmounted while waiting for the * covered vnode lock. */ if (coveredvp->v_mountedhere != mp || coveredvp->v_mountedhere->mnt_gen != mnt_gen_r) { VOP_UNLOCK(coveredvp, 0); return (EBUSY); } } /* * Only privileged root, or (if MNT_USER is set) the user that did the * original mount is permitted to unmount this filesystem. */ error = vfs_suser(mp, td); if (error) { if (coveredvp) VOP_UNLOCK(coveredvp, 0); return (error); } vn_start_write(NULL, &mp, V_WAIT); MNT_ILOCK(mp); if ((mp->mnt_kern_flag & MNTK_UNMOUNT) != 0 || !TAILQ_EMPTY(&mp->mnt_uppers)) { MNT_IUNLOCK(mp); if (coveredvp) VOP_UNLOCK(coveredvp, 0); vn_finished_write(mp); return (EBUSY); } mp->mnt_kern_flag |= MNTK_UNMOUNT | MNTK_NOINSMNTQ; /* Allow filesystems to detect that a forced unmount is in progress. */ if (flags & MNT_FORCE) { mp->mnt_kern_flag |= MNTK_UNMOUNTF; MNT_IUNLOCK(mp); /* * Must be done after setting MNTK_UNMOUNTF and before * waiting for mnt_lockref to become 0. */ VFS_PURGE(mp); MNT_ILOCK(mp); } error = 0; if (mp->mnt_lockref) { mp->mnt_kern_flag |= MNTK_DRAINING; error = msleep(&mp->mnt_lockref, MNT_MTX(mp), PVFS, "mount drain", 0); } MNT_IUNLOCK(mp); KASSERT(mp->mnt_lockref == 0, ("%s: invalid lock refcount in the drain path @ %s:%d", __func__, __FILE__, __LINE__)); KASSERT(error == 0, ("%s: invalid return value for msleep in the drain path @ %s:%d", __func__, __FILE__, __LINE__)); if (mp->mnt_flag & MNT_EXPUBLIC) vfs_setpublicfs(NULL, NULL, NULL); vfs_msync(mp, MNT_WAIT); MNT_ILOCK(mp); async_flag = mp->mnt_flag & MNT_ASYNC; mp->mnt_flag &= ~MNT_ASYNC; mp->mnt_kern_flag &= ~MNTK_ASYNC; MNT_IUNLOCK(mp); cache_purgevfs(mp); /* remove cache entries for this file sys */ vfs_deallocate_syncvnode(mp); /* * For forced unmounts, move process cdir/rdir refs on the fs root * vnode to the covered vnode. For non-forced unmounts we want * such references to cause an EBUSY error. */ if ((flags & MNT_FORCE) && VFS_ROOT(mp, LK_EXCLUSIVE, &fsrootvp) == 0) { if (mp->mnt_vnodecovered != NULL) mountcheckdirs(fsrootvp, mp->mnt_vnodecovered); if (fsrootvp == rootvnode) { vrele(rootvnode); rootvnode = NULL; } vput(fsrootvp); } - if (((mp->mnt_flag & MNT_RDONLY) || - (error = VFS_SYNC(mp, MNT_WAIT)) == 0) || (flags & MNT_FORCE) != 0) + if ((mp->mnt_flag & MNT_RDONLY) != 0 || (flags & MNT_FORCE) != 0 || + (error = VFS_SYNC(mp, MNT_WAIT)) == 0) error = VFS_UNMOUNT(mp, flags); vn_finished_write(mp); /* * If we failed to flush the dirty blocks for this mount point, * undo all the cdir/rdir and rootvnode changes we made above. * Unless we failed to do so because the device is reporting that * it doesn't exist anymore. */ if (error && error != ENXIO) { if ((flags & MNT_FORCE) && VFS_ROOT(mp, LK_EXCLUSIVE, &fsrootvp) == 0) { if (mp->mnt_vnodecovered != NULL) mountcheckdirs(mp->mnt_vnodecovered, fsrootvp); if (rootvnode == NULL) { rootvnode = fsrootvp; vref(rootvnode); } vput(fsrootvp); } MNT_ILOCK(mp); mp->mnt_kern_flag &= ~MNTK_NOINSMNTQ; if ((mp->mnt_flag & MNT_RDONLY) == 0) { MNT_IUNLOCK(mp); vfs_allocate_syncvnode(mp); MNT_ILOCK(mp); } mp->mnt_kern_flag &= ~(MNTK_UNMOUNT | MNTK_UNMOUNTF); mp->mnt_flag |= async_flag; if ((mp->mnt_flag & MNT_ASYNC) != 0 && (mp->mnt_kern_flag & MNTK_NOASYNC) == 0) mp->mnt_kern_flag |= MNTK_ASYNC; if (mp->mnt_kern_flag & MNTK_MWAIT) { mp->mnt_kern_flag &= ~MNTK_MWAIT; wakeup(mp); } MNT_IUNLOCK(mp); if (coveredvp) VOP_UNLOCK(coveredvp, 0); return (error); } mtx_lock(&mountlist_mtx); TAILQ_REMOVE(&mountlist, mp, mnt_list); mtx_unlock(&mountlist_mtx); EVENTHANDLER_INVOKE(vfs_unmounted, mp, td); if (coveredvp != NULL) { coveredvp->v_mountedhere = NULL; vput(coveredvp); } vfs_event_signal(NULL, VQ_UNMOUNT, 0); vfs_mount_destroy(mp); return (0); } /* * Report errors during filesystem mounting. */ void vfs_mount_error(struct mount *mp, const char *fmt, ...) { struct vfsoptlist *moptlist = mp->mnt_optnew; va_list ap; int error, len; char *errmsg; error = vfs_getopt(moptlist, "errmsg", (void **)&errmsg, &len); if (error || errmsg == NULL || len <= 0) return; va_start(ap, fmt); vsnprintf(errmsg, (size_t)len, fmt, ap); va_end(ap); } void vfs_opterror(struct vfsoptlist *opts, const char *fmt, ...) { va_list ap; int error, len; char *errmsg; error = vfs_getopt(opts, "errmsg", (void **)&errmsg, &len); if (error || errmsg == NULL || len <= 0) return; va_start(ap, fmt); vsnprintf(errmsg, (size_t)len, fmt, ap); va_end(ap); } /* * --------------------------------------------------------------------- * Functions for querying mount options/arguments from filesystems. */ /* * Check that no unknown options are given */ int vfs_filteropt(struct vfsoptlist *opts, const char **legal) { struct vfsopt *opt; char errmsg[255]; const char **t, *p, *q; int ret = 0; TAILQ_FOREACH(opt, opts, link) { p = opt->name; q = NULL; if (p[0] == 'n' && p[1] == 'o') q = p + 2; for(t = global_opts; *t != NULL; t++) { if (strcmp(*t, p) == 0) break; if (q != NULL) { if (strcmp(*t, q) == 0) break; } } if (*t != NULL) continue; for(t = legal; *t != NULL; t++) { if (strcmp(*t, p) == 0) break; if (q != NULL) { if (strcmp(*t, q) == 0) break; } } if (*t != NULL) continue; snprintf(errmsg, sizeof(errmsg), "mount option <%s> is unknown", p); ret = EINVAL; } if (ret != 0) { TAILQ_FOREACH(opt, opts, link) { if (strcmp(opt->name, "errmsg") == 0) { strncpy((char *)opt->value, errmsg, opt->len); break; } } if (opt == NULL) printf("%s\n", errmsg); } return (ret); } /* * Get a mount option by its name. * * Return 0 if the option was found, ENOENT otherwise. * If len is non-NULL it will be filled with the length * of the option. If buf is non-NULL, it will be filled * with the address of the option. */ int vfs_getopt(opts, name, buf, len) struct vfsoptlist *opts; const char *name; void **buf; int *len; { struct vfsopt *opt; KASSERT(opts != NULL, ("vfs_getopt: caller passed 'opts' as NULL")); TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) == 0) { opt->seen = 1; if (len != NULL) *len = opt->len; if (buf != NULL) *buf = opt->value; return (0); } } return (ENOENT); } int vfs_getopt_pos(struct vfsoptlist *opts, const char *name) { struct vfsopt *opt; if (opts == NULL) return (-1); TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) == 0) { opt->seen = 1; return (opt->pos); } } return (-1); } int vfs_getopt_size(struct vfsoptlist *opts, const char *name, off_t *value) { char *opt_value, *vtp; quad_t iv; int error, opt_len; error = vfs_getopt(opts, name, (void **)&opt_value, &opt_len); if (error != 0) return (error); if (opt_len == 0 || opt_value == NULL) return (EINVAL); if (opt_value[0] == '\0' || opt_value[opt_len - 1] != '\0') return (EINVAL); iv = strtoq(opt_value, &vtp, 0); if (vtp == opt_value || (vtp[0] != '\0' && vtp[1] != '\0')) return (EINVAL); if (iv < 0) return (EINVAL); switch (vtp[0]) { case 't': case 'T': iv *= 1024; case 'g': case 'G': iv *= 1024; case 'm': case 'M': iv *= 1024; case 'k': case 'K': iv *= 1024; case '\0': break; default: return (EINVAL); } *value = iv; return (0); } char * vfs_getopts(struct vfsoptlist *opts, const char *name, int *error) { struct vfsopt *opt; *error = 0; TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) != 0) continue; opt->seen = 1; if (opt->len == 0 || ((char *)opt->value)[opt->len - 1] != '\0') { *error = EINVAL; return (NULL); } return (opt->value); } *error = ENOENT; return (NULL); } int vfs_flagopt(struct vfsoptlist *opts, const char *name, uint64_t *w, uint64_t val) { struct vfsopt *opt; TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) == 0) { opt->seen = 1; if (w != NULL) *w |= val; return (1); } } if (w != NULL) *w &= ~val; return (0); } int vfs_scanopt(struct vfsoptlist *opts, const char *name, const char *fmt, ...) { va_list ap; struct vfsopt *opt; int ret; KASSERT(opts != NULL, ("vfs_getopt: caller passed 'opts' as NULL")); TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) != 0) continue; opt->seen = 1; if (opt->len == 0 || opt->value == NULL) return (0); if (((char *)opt->value)[opt->len - 1] != '\0') return (0); va_start(ap, fmt); ret = vsscanf(opt->value, fmt, ap); va_end(ap); return (ret); } return (0); } int vfs_setopt(struct vfsoptlist *opts, const char *name, void *value, int len) { struct vfsopt *opt; TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) != 0) continue; opt->seen = 1; if (opt->value == NULL) opt->len = len; else { if (opt->len != len) return (EINVAL); bcopy(value, opt->value, len); } return (0); } return (ENOENT); } int vfs_setopt_part(struct vfsoptlist *opts, const char *name, void *value, int len) { struct vfsopt *opt; TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) != 0) continue; opt->seen = 1; if (opt->value == NULL) opt->len = len; else { if (opt->len < len) return (EINVAL); opt->len = len; bcopy(value, opt->value, len); } return (0); } return (ENOENT); } int vfs_setopts(struct vfsoptlist *opts, const char *name, const char *value) { struct vfsopt *opt; TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) != 0) continue; opt->seen = 1; if (opt->value == NULL) opt->len = strlen(value) + 1; else if (strlcpy(opt->value, value, opt->len) >= opt->len) return (EINVAL); return (0); } return (ENOENT); } /* * Find and copy a mount option. * * The size of the buffer has to be specified * in len, if it is not the same length as the * mount option, EINVAL is returned. * Returns ENOENT if the option is not found. */ int vfs_copyopt(opts, name, dest, len) struct vfsoptlist *opts; const char *name; void *dest; int len; { struct vfsopt *opt; KASSERT(opts != NULL, ("vfs_copyopt: caller passed 'opts' as NULL")); TAILQ_FOREACH(opt, opts, link) { if (strcmp(name, opt->name) == 0) { opt->seen = 1; if (len != opt->len) return (EINVAL); bcopy(opt->value, dest, opt->len); return (0); } } return (ENOENT); } int __vfs_statfs(struct mount *mp, struct statfs *sbp) { int error; error = mp->mnt_op->vfs_statfs(mp, &mp->mnt_stat); if (sbp != &mp->mnt_stat) *sbp = mp->mnt_stat; return (error); } void vfs_mountedfrom(struct mount *mp, const char *from) { bzero(mp->mnt_stat.f_mntfromname, sizeof mp->mnt_stat.f_mntfromname); strlcpy(mp->mnt_stat.f_mntfromname, from, sizeof mp->mnt_stat.f_mntfromname); } /* * --------------------------------------------------------------------- * This is the api for building mount args and mounting filesystems from * inside the kernel. * * The API works by accumulation of individual args. First error is * latched. * * XXX: should be documented in new manpage kernel_mount(9) */ /* A memory allocation which must be freed when we are done */ struct mntaarg { SLIST_ENTRY(mntaarg) next; }; /* The header for the mount arguments */ struct mntarg { struct iovec *v; int len; int error; SLIST_HEAD(, mntaarg) list; }; /* * Add a boolean argument. * * flag is the boolean value. * name must start with "no". */ struct mntarg * mount_argb(struct mntarg *ma, int flag, const char *name) { KASSERT(name[0] == 'n' && name[1] == 'o', ("mount_argb(...,%s): name must start with 'no'", name)); return (mount_arg(ma, name + (flag ? 2 : 0), NULL, 0)); } /* * Add an argument printf style */ struct mntarg * mount_argf(struct mntarg *ma, const char *name, const char *fmt, ...) { va_list ap; struct mntaarg *maa; struct sbuf *sb; int len; if (ma == NULL) { ma = malloc(sizeof *ma, M_MOUNT, M_WAITOK | M_ZERO); SLIST_INIT(&ma->list); } if (ma->error) return (ma); ma->v = realloc(ma->v, sizeof *ma->v * (ma->len + 2), M_MOUNT, M_WAITOK); ma->v[ma->len].iov_base = (void *)(uintptr_t)name; ma->v[ma->len].iov_len = strlen(name) + 1; ma->len++; sb = sbuf_new_auto(); va_start(ap, fmt); sbuf_vprintf(sb, fmt, ap); va_end(ap); sbuf_finish(sb); len = sbuf_len(sb) + 1; maa = malloc(sizeof *maa + len, M_MOUNT, M_WAITOK | M_ZERO); SLIST_INSERT_HEAD(&ma->list, maa, next); bcopy(sbuf_data(sb), maa + 1, len); sbuf_delete(sb); ma->v[ma->len].iov_base = maa + 1; ma->v[ma->len].iov_len = len; ma->len++; return (ma); } /* * Add an argument which is a userland string. */ struct mntarg * mount_argsu(struct mntarg *ma, const char *name, const void *val, int len) { struct mntaarg *maa; char *tbuf; if (val == NULL) return (ma); if (ma == NULL) { ma = malloc(sizeof *ma, M_MOUNT, M_WAITOK | M_ZERO); SLIST_INIT(&ma->list); } if (ma->error) return (ma); maa = malloc(sizeof *maa + len, M_MOUNT, M_WAITOK | M_ZERO); SLIST_INSERT_HEAD(&ma->list, maa, next); tbuf = (void *)(maa + 1); ma->error = copyinstr(val, tbuf, len, NULL); return (mount_arg(ma, name, tbuf, -1)); } /* * Plain argument. * * If length is -1, treat value as a C string. */ struct mntarg * mount_arg(struct mntarg *ma, const char *name, const void *val, int len) { if (ma == NULL) { ma = malloc(sizeof *ma, M_MOUNT, M_WAITOK | M_ZERO); SLIST_INIT(&ma->list); } if (ma->error) return (ma); ma->v = realloc(ma->v, sizeof *ma->v * (ma->len + 2), M_MOUNT, M_WAITOK); ma->v[ma->len].iov_base = (void *)(uintptr_t)name; ma->v[ma->len].iov_len = strlen(name) + 1; ma->len++; ma->v[ma->len].iov_base = (void *)(uintptr_t)val; if (len < 0) ma->v[ma->len].iov_len = strlen(val) + 1; else ma->v[ma->len].iov_len = len; ma->len++; return (ma); } /* * Free a mntarg structure */ static void free_mntarg(struct mntarg *ma) { struct mntaarg *maa; while (!SLIST_EMPTY(&ma->list)) { maa = SLIST_FIRST(&ma->list); SLIST_REMOVE_HEAD(&ma->list, next); free(maa, M_MOUNT); } free(ma->v, M_MOUNT); free(ma, M_MOUNT); } /* * Mount a filesystem */ int kernel_mount(struct mntarg *ma, uint64_t flags) { struct uio auio; int error; KASSERT(ma != NULL, ("kernel_mount NULL ma")); KASSERT(ma->v != NULL, ("kernel_mount NULL ma->v")); KASSERT(!(ma->len & 1), ("kernel_mount odd ma->len (%d)", ma->len)); auio.uio_iov = ma->v; auio.uio_iovcnt = ma->len; auio.uio_segflg = UIO_SYSSPACE; error = ma->error; if (!error) error = vfs_donmount(curthread, flags, &auio); free_mntarg(ma); return (error); } /* * A printflike function to mount a filesystem. */ int kernel_vmount(int flags, ...) { struct mntarg *ma = NULL; va_list ap; const char *cp; const void *vp; int error; va_start(ap, flags); for (;;) { cp = va_arg(ap, const char *); if (cp == NULL) break; vp = va_arg(ap, const void *); ma = mount_arg(ma, cp, vp, (vp != NULL ? -1 : 0)); } va_end(ap); error = kernel_mount(ma, flags); return (error); } void vfs_oexport_conv(const struct oexport_args *oexp, struct export_args *exp) { bcopy(oexp, exp, sizeof(*oexp)); exp->ex_numsecflavors = 0; } Index: projects/building-blocks/sys/kern/vfs_subr.c =================================================================== --- projects/building-blocks/sys/kern/vfs_subr.c (revision 275663) +++ projects/building-blocks/sys/kern/vfs_subr.c (revision 275664) @@ -1,4845 +1,4864 @@ /*- * Copyright (c) 1989, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)vfs_subr.c 8.31 (Berkeley) 5/26/95 */ /* * External virtual filesystem routines */ #include __FBSDID("$FreeBSD$"); #include "opt_compat.h" #include "opt_ddb.h" #include "opt_watchdog.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef DDB #include #endif static void delmntque(struct vnode *vp); static int flushbuflist(struct bufv *bufv, int flags, struct bufobj *bo, int slpflag, int slptimeo); static void syncer_shutdown(void *arg, int howto); static int vtryrecycle(struct vnode *vp); static void v_incr_usecount(struct vnode *); static void v_decr_usecount(struct vnode *); static void v_decr_useonly(struct vnode *); static void v_upgrade_usecount(struct vnode *); static void vnlru_free(int); static void vgonel(struct vnode *); static void vfs_knllock(void *arg); static void vfs_knlunlock(void *arg); static void vfs_knl_assert_locked(void *arg); static void vfs_knl_assert_unlocked(void *arg); static void destroy_vpollinfo(struct vpollinfo *vi); /* * Number of vnodes in existence. Increased whenever getnewvnode() * allocates a new vnode, decreased in vdropl() for VI_DOOMED vnode. */ static unsigned long numvnodes; SYSCTL_ULONG(_vfs, OID_AUTO, numvnodes, CTLFLAG_RD, &numvnodes, 0, "Number of vnodes in existence"); /* * Conversion tables for conversion from vnode types to inode formats * and back. */ enum vtype iftovt_tab[16] = { VNON, VFIFO, VCHR, VNON, VDIR, VNON, VBLK, VNON, VREG, VNON, VLNK, VNON, VSOCK, VNON, VNON, VBAD, }; int vttoif_tab[10] = { 0, S_IFREG, S_IFDIR, S_IFBLK, S_IFCHR, S_IFLNK, S_IFSOCK, S_IFIFO, S_IFMT, S_IFMT }; /* * List of vnodes that are ready for recycling. */ static TAILQ_HEAD(freelst, vnode) vnode_free_list; /* * Free vnode target. Free vnodes may simply be files which have been stat'd * but not read. This is somewhat common, and a small cache of such files * should be kept to avoid recreation costs. */ static u_long wantfreevnodes; SYSCTL_ULONG(_vfs, OID_AUTO, wantfreevnodes, CTLFLAG_RW, &wantfreevnodes, 0, ""); /* Number of vnodes in the free list. */ static u_long freevnodes; SYSCTL_ULONG(_vfs, OID_AUTO, freevnodes, CTLFLAG_RD, &freevnodes, 0, "Number of vnodes in the free list"); static int vlru_allow_cache_src; SYSCTL_INT(_vfs, OID_AUTO, vlru_allow_cache_src, CTLFLAG_RW, &vlru_allow_cache_src, 0, "Allow vlru to reclaim source vnode"); /* * Various variables used for debugging the new implementation of * reassignbuf(). * XXX these are probably of (very) limited utility now. */ static int reassignbufcalls; SYSCTL_INT(_vfs, OID_AUTO, reassignbufcalls, CTLFLAG_RW, &reassignbufcalls, 0, "Number of calls to reassignbuf"); /* * Cache for the mount type id assigned to NFS. This is used for * special checks in nfs/nfs_nqlease.c and vm/vnode_pager.c. */ int nfs_mount_type = -1; /* To keep more than one thread at a time from running vfs_getnewfsid */ static struct mtx mntid_mtx; /* * Lock for any access to the following: * vnode_free_list * numvnodes * freevnodes */ static struct mtx vnode_free_list_mtx; /* Publicly exported FS */ struct nfs_public nfs_pub; static uma_zone_t buf_trie_zone; /* Zone for allocation of new vnodes - used exclusively by getnewvnode() */ static uma_zone_t vnode_zone; static uma_zone_t vnodepoll_zone; /* * The workitem queue. * * It is useful to delay writes of file data and filesystem metadata * for tens of seconds so that quickly created and deleted files need * not waste disk bandwidth being created and removed. To realize this, * we append vnodes to a "workitem" queue. When running with a soft * updates implementation, most pending metadata dependencies should * not wait for more than a few seconds. Thus, mounted on block devices * are delayed only about a half the time that file data is delayed. * Similarly, directory updates are more critical, so are only delayed * about a third the time that file data is delayed. Thus, there are * SYNCER_MAXDELAY queues that are processed round-robin at a rate of * one each second (driven off the filesystem syncer process). The * syncer_delayno variable indicates the next queue that is to be processed. * Items that need to be processed soon are placed in this queue: * * syncer_workitem_pending[syncer_delayno] * * A delay of fifteen seconds is done by placing the request fifteen * entries later in the queue: * * syncer_workitem_pending[(syncer_delayno + 15) & syncer_mask] * */ static int syncer_delayno; static long syncer_mask; LIST_HEAD(synclist, bufobj); static struct synclist *syncer_workitem_pending; /* * The sync_mtx protects: * bo->bo_synclist * sync_vnode_count * syncer_delayno * syncer_state * syncer_workitem_pending * syncer_worklist_len * rushjob */ static struct mtx sync_mtx; static struct cv sync_wakeup; #define SYNCER_MAXDELAY 32 static int syncer_maxdelay = SYNCER_MAXDELAY; /* maximum delay time */ static int syncdelay = 30; /* max time to delay syncing data */ static int filedelay = 30; /* time to delay syncing files */ SYSCTL_INT(_kern, OID_AUTO, filedelay, CTLFLAG_RW, &filedelay, 0, "Time to delay syncing files (in seconds)"); static int dirdelay = 29; /* time to delay syncing directories */ SYSCTL_INT(_kern, OID_AUTO, dirdelay, CTLFLAG_RW, &dirdelay, 0, "Time to delay syncing directories (in seconds)"); static int metadelay = 28; /* time to delay syncing metadata */ SYSCTL_INT(_kern, OID_AUTO, metadelay, CTLFLAG_RW, &metadelay, 0, "Time to delay syncing metadata (in seconds)"); static int rushjob; /* number of slots to run ASAP */ static int stat_rush_requests; /* number of times I/O speeded up */ SYSCTL_INT(_debug, OID_AUTO, rush_requests, CTLFLAG_RW, &stat_rush_requests, 0, "Number of times I/O speeded up (rush requests)"); /* * When shutting down the syncer, run it at four times normal speed. */ #define SYNCER_SHUTDOWN_SPEEDUP 4 static int sync_vnode_count; static int syncer_worklist_len; static enum { SYNCER_RUNNING, SYNCER_SHUTTING_DOWN, SYNCER_FINAL_DELAY } syncer_state; /* * Number of vnodes we want to exist at any one time. This is mostly used * to size hash tables in vnode-related code. It is normally not used in * getnewvnode(), as wantfreevnodes is normally nonzero.) * * XXX desiredvnodes is historical cruft and should not exist. */ int desiredvnodes; SYSCTL_INT(_kern, KERN_MAXVNODES, maxvnodes, CTLFLAG_RW, &desiredvnodes, 0, "Maximum number of vnodes"); SYSCTL_ULONG(_kern, OID_AUTO, minvnodes, CTLFLAG_RW, &wantfreevnodes, 0, "Minimum number of vnodes (legacy)"); static int vnlru_nowhere; SYSCTL_INT(_debug, OID_AUTO, vnlru_nowhere, CTLFLAG_RW, &vnlru_nowhere, 0, "Number of times the vnlru process ran without success"); /* Shift count for (uintptr_t)vp to initialize vp->v_hash. */ static int vnsz2log; /* * Support for the bufobj clean & dirty pctrie. */ static void * buf_trie_alloc(struct pctrie *ptree) { return uma_zalloc(buf_trie_zone, M_NOWAIT); } static void buf_trie_free(struct pctrie *ptree, void *node) { uma_zfree(buf_trie_zone, node); } PCTRIE_DEFINE(BUF, buf, b_lblkno, buf_trie_alloc, buf_trie_free); /* * Initialize the vnode management data structures. * * Reevaluate the following cap on the number of vnodes after the physical * memory size exceeds 512GB. In the limit, as the physical memory size * grows, the ratio of physical pages to vnodes approaches sixteen to one. */ #ifndef MAXVNODES_MAX #define MAXVNODES_MAX (512 * (1024 * 1024 * 1024 / (int)PAGE_SIZE / 16)) #endif static void vntblinit(void *dummy __unused) { u_int i; int physvnodes, virtvnodes; /* * Desiredvnodes is a function of the physical memory size and the * kernel's heap size. Generally speaking, it scales with the * physical memory size. The ratio of desiredvnodes to physical pages * is one to four until desiredvnodes exceeds 98,304. Thereafter, the * marginal ratio of desiredvnodes to physical pages is one to * sixteen. However, desiredvnodes is limited by the kernel's heap * size. The memory required by desiredvnodes vnodes and vm objects * may not exceed one seventh of the kernel's heap size. */ physvnodes = maxproc + vm_cnt.v_page_count / 16 + 3 * min(98304 * 4, vm_cnt.v_page_count) / 16; virtvnodes = vm_kmem_size / (7 * (sizeof(struct vm_object) + sizeof(struct vnode))); desiredvnodes = min(physvnodes, virtvnodes); if (desiredvnodes > MAXVNODES_MAX) { if (bootverbose) printf("Reducing kern.maxvnodes %d -> %d\n", desiredvnodes, MAXVNODES_MAX); desiredvnodes = MAXVNODES_MAX; } wantfreevnodes = desiredvnodes / 4; mtx_init(&mntid_mtx, "mntid", NULL, MTX_DEF); TAILQ_INIT(&vnode_free_list); mtx_init(&vnode_free_list_mtx, "vnode_free_list", NULL, MTX_DEF); vnode_zone = uma_zcreate("VNODE", sizeof (struct vnode), NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0); vnodepoll_zone = uma_zcreate("VNODEPOLL", sizeof (struct vpollinfo), NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0); /* * Preallocate enough nodes to support one-per buf so that * we can not fail an insert. reassignbuf() callers can not * tolerate the insertion failure. */ buf_trie_zone = uma_zcreate("BUF TRIE", pctrie_node_size(), NULL, NULL, pctrie_zone_init, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE | UMA_ZONE_VM); uma_prealloc(buf_trie_zone, nbuf); /* * Initialize the filesystem syncer. */ syncer_workitem_pending = hashinit(syncer_maxdelay, M_VNODE, &syncer_mask); syncer_maxdelay = syncer_mask + 1; mtx_init(&sync_mtx, "Syncer mtx", NULL, MTX_DEF); cv_init(&sync_wakeup, "syncer"); for (i = 1; i <= sizeof(struct vnode); i <<= 1) vnsz2log++; vnsz2log--; } SYSINIT(vfs, SI_SUB_VFS, SI_ORDER_FIRST, vntblinit, NULL); /* * Mark a mount point as busy. Used to synchronize access and to delay * unmounting. Eventually, mountlist_mtx is not released on failure. * * vfs_busy() is a custom lock, it can block the caller. * vfs_busy() only sleeps if the unmount is active on the mount point. * For a mountpoint mp, vfs_busy-enforced lock is before lock of any * vnode belonging to mp. * * Lookup uses vfs_busy() to traverse mount points. * root fs var fs * / vnode lock A / vnode lock (/var) D * /var vnode lock B /log vnode lock(/var/log) E * vfs_busy lock C vfs_busy lock F * * Within each file system, the lock order is C->A->B and F->D->E. * * When traversing across mounts, the system follows that lock order: * * C->A->B * | * +->F->D->E * * The lookup() process for namei("/var") illustrates the process: * VOP_LOOKUP() obtains B while A is held * vfs_busy() obtains a shared lock on F while A and B are held * vput() releases lock on B * vput() releases lock on A * VFS_ROOT() obtains lock on D while shared lock on F is held * vfs_unbusy() releases shared lock on F * vn_lock() obtains lock on deadfs vnode vp_crossmp instead of A. * Attempt to lock A (instead of vp_crossmp) while D is held would * violate the global order, causing deadlocks. * * dounmount() locks B while F is drained. */ int vfs_busy(struct mount *mp, int flags) { MPASS((flags & ~MBF_MASK) == 0); CTR3(KTR_VFS, "%s: mp %p with flags %d", __func__, mp, flags); MNT_ILOCK(mp); MNT_REF(mp); /* * If mount point is currenly being unmounted, sleep until the * mount point fate is decided. If thread doing the unmounting fails, * it will clear MNTK_UNMOUNT flag before waking us up, indicating * that this mount point has survived the unmount attempt and vfs_busy * should retry. Otherwise the unmounter thread will set MNTK_REFEXPIRE * flag in addition to MNTK_UNMOUNT, indicating that mount point is * about to be really destroyed. vfs_busy needs to release its * reference on the mount point in this case and return with ENOENT, * telling the caller that mount mount it tried to busy is no longer * valid. */ while (mp->mnt_kern_flag & MNTK_UNMOUNT) { if (flags & MBF_NOWAIT || mp->mnt_kern_flag & MNTK_REFEXPIRE) { MNT_REL(mp); MNT_IUNLOCK(mp); CTR1(KTR_VFS, "%s: failed busying before sleeping", __func__); return (ENOENT); } if (flags & MBF_MNTLSTLOCK) mtx_unlock(&mountlist_mtx); mp->mnt_kern_flag |= MNTK_MWAIT; msleep(mp, MNT_MTX(mp), PVFS | PDROP, "vfs_busy", 0); if (flags & MBF_MNTLSTLOCK) mtx_lock(&mountlist_mtx); MNT_ILOCK(mp); } if (flags & MBF_MNTLSTLOCK) mtx_unlock(&mountlist_mtx); mp->mnt_lockref++; MNT_IUNLOCK(mp); return (0); } /* * Free a busy filesystem. */ void vfs_unbusy(struct mount *mp) { CTR2(KTR_VFS, "%s: mp %p", __func__, mp); MNT_ILOCK(mp); MNT_REL(mp); KASSERT(mp->mnt_lockref > 0, ("negative mnt_lockref")); mp->mnt_lockref--; if (mp->mnt_lockref == 0 && (mp->mnt_kern_flag & MNTK_DRAINING) != 0) { MPASS(mp->mnt_kern_flag & MNTK_UNMOUNT); CTR1(KTR_VFS, "%s: waking up waiters", __func__); mp->mnt_kern_flag &= ~MNTK_DRAINING; wakeup(&mp->mnt_lockref); } MNT_IUNLOCK(mp); } /* * Lookup a mount point by filesystem identifier. */ struct mount * vfs_getvfs(fsid_t *fsid) { struct mount *mp; CTR2(KTR_VFS, "%s: fsid %p", __func__, fsid); mtx_lock(&mountlist_mtx); TAILQ_FOREACH(mp, &mountlist, mnt_list) { if (mp->mnt_stat.f_fsid.val[0] == fsid->val[0] && mp->mnt_stat.f_fsid.val[1] == fsid->val[1]) { vfs_ref(mp); mtx_unlock(&mountlist_mtx); return (mp); } } mtx_unlock(&mountlist_mtx); CTR2(KTR_VFS, "%s: lookup failed for %p id", __func__, fsid); return ((struct mount *) 0); } /* * Lookup a mount point by filesystem identifier, busying it before * returning. * * To avoid congestion on mountlist_mtx, implement simple direct-mapped * cache for popular filesystem identifiers. The cache is lockess, using * the fact that struct mount's are never freed. In worst case we may * get pointer to unmounted or even different filesystem, so we have to * check what we got, and go slow way if so. */ struct mount * vfs_busyfs(fsid_t *fsid) { #define FSID_CACHE_SIZE 256 typedef struct mount * volatile vmp_t; static vmp_t cache[FSID_CACHE_SIZE]; struct mount *mp; int error; uint32_t hash; CTR2(KTR_VFS, "%s: fsid %p", __func__, fsid); hash = fsid->val[0] ^ fsid->val[1]; hash = (hash >> 16 ^ hash) & (FSID_CACHE_SIZE - 1); mp = cache[hash]; if (mp == NULL || mp->mnt_stat.f_fsid.val[0] != fsid->val[0] || mp->mnt_stat.f_fsid.val[1] != fsid->val[1]) goto slow; if (vfs_busy(mp, 0) != 0) { cache[hash] = NULL; goto slow; } if (mp->mnt_stat.f_fsid.val[0] == fsid->val[0] && mp->mnt_stat.f_fsid.val[1] == fsid->val[1]) return (mp); else vfs_unbusy(mp); slow: mtx_lock(&mountlist_mtx); TAILQ_FOREACH(mp, &mountlist, mnt_list) { if (mp->mnt_stat.f_fsid.val[0] == fsid->val[0] && mp->mnt_stat.f_fsid.val[1] == fsid->val[1]) { error = vfs_busy(mp, MBF_MNTLSTLOCK); if (error) { cache[hash] = NULL; mtx_unlock(&mountlist_mtx); return (NULL); } cache[hash] = mp; return (mp); } } CTR2(KTR_VFS, "%s: lookup failed for %p id", __func__, fsid); mtx_unlock(&mountlist_mtx); return ((struct mount *) 0); } /* * Check if a user can access privileged mount options. */ int vfs_suser(struct mount *mp, struct thread *td) { int error; /* * If the thread is jailed, but this is not a jail-friendly file * system, deny immediately. */ if (!(mp->mnt_vfc->vfc_flags & VFCF_JAIL) && jailed(td->td_ucred)) return (EPERM); /* * If the file system was mounted outside the jail of the calling * thread, deny immediately. */ if (prison_check(td->td_ucred, mp->mnt_cred) != 0) return (EPERM); /* * If file system supports delegated administration, we don't check * for the PRIV_VFS_MOUNT_OWNER privilege - it will be better verified * by the file system itself. * If this is not the user that did original mount, we check for * the PRIV_VFS_MOUNT_OWNER privilege. */ if (!(mp->mnt_vfc->vfc_flags & VFCF_DELEGADMIN) && mp->mnt_cred->cr_uid != td->td_ucred->cr_uid) { if ((error = priv_check(td, PRIV_VFS_MOUNT_OWNER)) != 0) return (error); } return (0); } /* * Get a new unique fsid. Try to make its val[0] unique, since this value * will be used to create fake device numbers for stat(). Also try (but * not so hard) make its val[0] unique mod 2^16, since some emulators only * support 16-bit device numbers. We end up with unique val[0]'s for the * first 2^16 calls and unique val[0]'s mod 2^16 for the first 2^8 calls. * * Keep in mind that several mounts may be running in parallel. Starting * the search one past where the previous search terminated is both a * micro-optimization and a defense against returning the same fsid to * different mounts. */ void vfs_getnewfsid(struct mount *mp) { static uint16_t mntid_base; struct mount *nmp; fsid_t tfsid; int mtype; CTR2(KTR_VFS, "%s: mp %p", __func__, mp); mtx_lock(&mntid_mtx); mtype = mp->mnt_vfc->vfc_typenum; tfsid.val[1] = mtype; mtype = (mtype & 0xFF) << 24; for (;;) { tfsid.val[0] = makedev(255, mtype | ((mntid_base & 0xFF00) << 8) | (mntid_base & 0xFF)); mntid_base++; if ((nmp = vfs_getvfs(&tfsid)) == NULL) break; vfs_rel(nmp); } mp->mnt_stat.f_fsid.val[0] = tfsid.val[0]; mp->mnt_stat.f_fsid.val[1] = tfsid.val[1]; mtx_unlock(&mntid_mtx); } /* * Knob to control the precision of file timestamps: * * 0 = seconds only; nanoseconds zeroed. * 1 = seconds and nanoseconds, accurate within 1/HZ. * 2 = seconds and nanoseconds, truncated to microseconds. * >=3 = seconds and nanoseconds, maximum precision. */ enum { TSP_SEC, TSP_HZ, TSP_USEC, TSP_NSEC }; static int timestamp_precision = TSP_SEC; SYSCTL_INT(_vfs, OID_AUTO, timestamp_precision, CTLFLAG_RW, ×tamp_precision, 0, "File timestamp precision (0: seconds, " "1: sec + ns accurate to 1/HZ, 2: sec + ns truncated to ms, " "3+: sec + ns (max. precision))"); /* * Get a current timestamp. */ void vfs_timestamp(struct timespec *tsp) { struct timeval tv; switch (timestamp_precision) { case TSP_SEC: tsp->tv_sec = time_second; tsp->tv_nsec = 0; break; case TSP_HZ: getnanotime(tsp); break; case TSP_USEC: microtime(&tv); TIMEVAL_TO_TIMESPEC(&tv, tsp); break; case TSP_NSEC: default: nanotime(tsp); break; } } /* * Set vnode attributes to VNOVAL */ void vattr_null(struct vattr *vap) { vap->va_type = VNON; vap->va_size = VNOVAL; vap->va_bytes = VNOVAL; vap->va_mode = VNOVAL; vap->va_nlink = VNOVAL; vap->va_uid = VNOVAL; vap->va_gid = VNOVAL; vap->va_fsid = VNOVAL; vap->va_fileid = VNOVAL; vap->va_blocksize = VNOVAL; vap->va_rdev = VNOVAL; vap->va_atime.tv_sec = VNOVAL; vap->va_atime.tv_nsec = VNOVAL; vap->va_mtime.tv_sec = VNOVAL; vap->va_mtime.tv_nsec = VNOVAL; vap->va_ctime.tv_sec = VNOVAL; vap->va_ctime.tv_nsec = VNOVAL; vap->va_birthtime.tv_sec = VNOVAL; vap->va_birthtime.tv_nsec = VNOVAL; vap->va_flags = VNOVAL; vap->va_gen = VNOVAL; vap->va_vaflags = 0; } /* * This routine is called when we have too many vnodes. It attempts * to free vnodes and will potentially free vnodes that still * have VM backing store (VM backing store is typically the cause * of a vnode blowout so we want to do this). Therefore, this operation * is not considered cheap. * * A number of conditions may prevent a vnode from being reclaimed. * the buffer cache may have references on the vnode, a directory * vnode may still have references due to the namei cache representing * underlying files, or the vnode may be in active use. It is not * desireable to reuse such vnodes. These conditions may cause the * number of vnodes to reach some minimum value regardless of what * you set kern.maxvnodes to. Do not set kern.maxvnodes too low. */ static int vlrureclaim(struct mount *mp) { struct vnode *vp; int done; int trigger; int usevnodes; int count; /* * Calculate the trigger point, don't allow user * screwups to blow us up. This prevents us from * recycling vnodes with lots of resident pages. We * aren't trying to free memory, we are trying to * free vnodes. */ usevnodes = desiredvnodes; if (usevnodes <= 0) usevnodes = 1; trigger = vm_cnt.v_page_count * 2 / usevnodes; done = 0; vn_start_write(NULL, &mp, V_WAIT); MNT_ILOCK(mp); count = mp->mnt_nvnodelistsize / 10 + 1; while (count != 0) { vp = TAILQ_FIRST(&mp->mnt_nvnodelist); while (vp != NULL && vp->v_type == VMARKER) vp = TAILQ_NEXT(vp, v_nmntvnodes); if (vp == NULL) break; TAILQ_REMOVE(&mp->mnt_nvnodelist, vp, v_nmntvnodes); TAILQ_INSERT_TAIL(&mp->mnt_nvnodelist, vp, v_nmntvnodes); --count; if (!VI_TRYLOCK(vp)) goto next_iter; /* * If it's been deconstructed already, it's still * referenced, or it exceeds the trigger, skip it. */ if (vp->v_usecount || (!vlru_allow_cache_src && !LIST_EMPTY(&(vp)->v_cache_src)) || (vp->v_iflag & VI_DOOMED) != 0 || (vp->v_object != NULL && vp->v_object->resident_page_count > trigger)) { VI_UNLOCK(vp); goto next_iter; } MNT_IUNLOCK(mp); vholdl(vp); if (VOP_LOCK(vp, LK_INTERLOCK|LK_EXCLUSIVE|LK_NOWAIT)) { vdrop(vp); goto next_iter_mntunlocked; } VI_LOCK(vp); /* * v_usecount may have been bumped after VOP_LOCK() dropped * the vnode interlock and before it was locked again. * * It is not necessary to recheck VI_DOOMED because it can * only be set by another thread that holds both the vnode * lock and vnode interlock. If another thread has the * vnode lock before we get to VOP_LOCK() and obtains the * vnode interlock after VOP_LOCK() drops the vnode * interlock, the other thread will be unable to drop the * vnode lock before our VOP_LOCK() call fails. */ if (vp->v_usecount || (!vlru_allow_cache_src && !LIST_EMPTY(&(vp)->v_cache_src)) || (vp->v_object != NULL && vp->v_object->resident_page_count > trigger)) { VOP_UNLOCK(vp, LK_INTERLOCK); vdrop(vp); goto next_iter_mntunlocked; } KASSERT((vp->v_iflag & VI_DOOMED) == 0, ("VI_DOOMED unexpectedly detected in vlrureclaim()")); vgonel(vp); VOP_UNLOCK(vp, 0); vdropl(vp); done++; next_iter_mntunlocked: if (!should_yield()) goto relock_mnt; goto yield; next_iter: if (!should_yield()) continue; MNT_IUNLOCK(mp); yield: kern_yield(PRI_USER); relock_mnt: MNT_ILOCK(mp); } MNT_IUNLOCK(mp); vn_finished_write(mp); return done; } /* * Attempt to keep the free list at wantfreevnodes length. */ static void vnlru_free(int count) { struct vnode *vp; mtx_assert(&vnode_free_list_mtx, MA_OWNED); for (; count > 0; count--) { vp = TAILQ_FIRST(&vnode_free_list); /* * The list can be modified while the free_list_mtx * has been dropped and vp could be NULL here. */ if (!vp) break; VNASSERT(vp->v_op != NULL, vp, ("vnlru_free: vnode already reclaimed.")); KASSERT((vp->v_iflag & VI_FREE) != 0, ("Removing vnode not on freelist")); KASSERT((vp->v_iflag & VI_ACTIVE) == 0, ("Mangling active vnode")); TAILQ_REMOVE(&vnode_free_list, vp, v_actfreelist); /* * Don't recycle if we can't get the interlock. */ if (!VI_TRYLOCK(vp)) { TAILQ_INSERT_TAIL(&vnode_free_list, vp, v_actfreelist); continue; } VNASSERT((vp->v_iflag & VI_FREE) != 0 && vp->v_holdcnt == 0, vp, ("vp inconsistent on freelist")); /* * The clear of VI_FREE prevents activation of the * vnode. There is no sense in putting the vnode on * the mount point active list, only to remove it * later during recycling. Inline the relevant part * of vholdl(), to avoid triggering assertions or * activating. */ freevnodes--; vp->v_iflag &= ~VI_FREE; vp->v_holdcnt++; mtx_unlock(&vnode_free_list_mtx); VI_UNLOCK(vp); vtryrecycle(vp); /* * If the recycled succeeded this vdrop will actually free * the vnode. If not it will simply place it back on * the free list. */ vdrop(vp); mtx_lock(&vnode_free_list_mtx); } } /* * Attempt to recycle vnodes in a context that is always safe to block. * Calling vlrurecycle() from the bowels of filesystem code has some * interesting deadlock problems. */ static struct proc *vnlruproc; static int vnlruproc_sig; static void vnlru_proc(void) { struct mount *mp, *nmp; int done; struct proc *p = vnlruproc; EVENTHANDLER_REGISTER(shutdown_pre_sync, kproc_shutdown, p, SHUTDOWN_PRI_FIRST); for (;;) { kproc_suspend_check(p); mtx_lock(&vnode_free_list_mtx); if (freevnodes > wantfreevnodes) vnlru_free(freevnodes - wantfreevnodes); if (numvnodes <= desiredvnodes * 9 / 10) { vnlruproc_sig = 0; wakeup(&vnlruproc_sig); msleep(vnlruproc, &vnode_free_list_mtx, PVFS|PDROP, "vlruwt", hz); continue; } mtx_unlock(&vnode_free_list_mtx); done = 0; mtx_lock(&mountlist_mtx); for (mp = TAILQ_FIRST(&mountlist); mp != NULL; mp = nmp) { if (vfs_busy(mp, MBF_NOWAIT | MBF_MNTLSTLOCK)) { nmp = TAILQ_NEXT(mp, mnt_list); continue; } done += vlrureclaim(mp); mtx_lock(&mountlist_mtx); nmp = TAILQ_NEXT(mp, mnt_list); vfs_unbusy(mp); } mtx_unlock(&mountlist_mtx); if (done == 0) { #if 0 /* These messages are temporary debugging aids */ if (vnlru_nowhere < 5) printf("vnlru process getting nowhere..\n"); else if (vnlru_nowhere == 5) printf("vnlru process messages stopped.\n"); #endif vnlru_nowhere++; tsleep(vnlruproc, PPAUSE, "vlrup", hz * 3); } else kern_yield(PRI_USER); } } static struct kproc_desc vnlru_kp = { "vnlru", vnlru_proc, &vnlruproc }; SYSINIT(vnlru, SI_SUB_KTHREAD_UPDATE, SI_ORDER_FIRST, kproc_start, &vnlru_kp); /* * Routines having to do with the management of the vnode table. */ /* * Try to recycle a freed vnode. We abort if anyone picks up a reference * before we actually vgone(). This function must be called with the vnode * held to prevent the vnode from being returned to the free list midway * through vgone(). */ static int vtryrecycle(struct vnode *vp) { struct mount *vnmp; CTR2(KTR_VFS, "%s: vp %p", __func__, vp); VNASSERT(vp->v_holdcnt, vp, ("vtryrecycle: Recycling vp %p without a reference.", vp)); /* * This vnode may found and locked via some other list, if so we * can't recycle it yet. */ if (VOP_LOCK(vp, LK_EXCLUSIVE | LK_NOWAIT) != 0) { CTR2(KTR_VFS, "%s: impossible to recycle, vp %p lock is already held", __func__, vp); return (EWOULDBLOCK); } /* * Don't recycle if its filesystem is being suspended. */ if (vn_start_write(vp, &vnmp, V_NOWAIT) != 0) { VOP_UNLOCK(vp, 0); CTR2(KTR_VFS, "%s: impossible to recycle, cannot start the write for %p", __func__, vp); return (EBUSY); } /* * If we got this far, we need to acquire the interlock and see if * anyone picked up this vnode from another list. If not, we will * mark it with DOOMED via vgonel() so that anyone who does find it * will skip over it. */ VI_LOCK(vp); if (vp->v_usecount) { VOP_UNLOCK(vp, LK_INTERLOCK); vn_finished_write(vnmp); CTR2(KTR_VFS, "%s: impossible to recycle, %p is already referenced", __func__, vp); return (EBUSY); } if ((vp->v_iflag & VI_DOOMED) == 0) vgonel(vp); VOP_UNLOCK(vp, LK_INTERLOCK); vn_finished_write(vnmp); return (0); } /* * Wait for available vnodes. */ static int getnewvnode_wait(int suspended) { mtx_assert(&vnode_free_list_mtx, MA_OWNED); if (numvnodes > desiredvnodes) { if (suspended) { /* * File system is beeing suspended, we cannot risk a * deadlock here, so allocate new vnode anyway. */ if (freevnodes > wantfreevnodes) vnlru_free(freevnodes - wantfreevnodes); return (0); } if (vnlruproc_sig == 0) { vnlruproc_sig = 1; /* avoid unnecessary wakeups */ wakeup(vnlruproc); } msleep(&vnlruproc_sig, &vnode_free_list_mtx, PVFS, "vlruwk", hz); } return (numvnodes > desiredvnodes ? ENFILE : 0); } void getnewvnode_reserve(u_int count) { struct thread *td; td = curthread; /* First try to be quick and racy. */ if (atomic_fetchadd_long(&numvnodes, count) + count <= desiredvnodes) { td->td_vp_reserv += count; return; } else atomic_subtract_long(&numvnodes, count); mtx_lock(&vnode_free_list_mtx); while (count > 0) { if (getnewvnode_wait(0) == 0) { count--; td->td_vp_reserv++; atomic_add_long(&numvnodes, 1); } } mtx_unlock(&vnode_free_list_mtx); } void getnewvnode_drop_reserve(void) { struct thread *td; td = curthread; atomic_subtract_long(&numvnodes, td->td_vp_reserv); td->td_vp_reserv = 0; } /* * Return the next vnode from the free list. */ int getnewvnode(const char *tag, struct mount *mp, struct vop_vector *vops, struct vnode **vpp) { struct vnode *vp; struct bufobj *bo; struct thread *td; int error; CTR3(KTR_VFS, "%s: mp %p with tag %s", __func__, mp, tag); vp = NULL; td = curthread; if (td->td_vp_reserv > 0) { td->td_vp_reserv -= 1; goto alloc; } mtx_lock(&vnode_free_list_mtx); /* * Lend our context to reclaim vnodes if they've exceeded the max. */ if (freevnodes > wantfreevnodes) vnlru_free(1); error = getnewvnode_wait(mp != NULL && (mp->mnt_kern_flag & MNTK_SUSPEND)); #if 0 /* XXX Not all VFS_VGET/ffs_vget callers check returns. */ if (error != 0) { mtx_unlock(&vnode_free_list_mtx); return (error); } #endif atomic_add_long(&numvnodes, 1); mtx_unlock(&vnode_free_list_mtx); alloc: vp = (struct vnode *) uma_zalloc(vnode_zone, M_WAITOK|M_ZERO); /* * Setup locks. */ vp->v_vnlock = &vp->v_lock; mtx_init(&vp->v_interlock, "vnode interlock", NULL, MTX_DEF); /* * By default, don't allow shared locks unless filesystems * opt-in. */ lockinit(vp->v_vnlock, PVFS, tag, VLKTIMEOUT, LK_NOSHARE | LK_IS_VNODE); /* * Initialize bufobj. */ bo = &vp->v_bufobj; bo->__bo_vnode = vp; rw_init(BO_LOCKPTR(bo), "bufobj interlock"); bo->bo_ops = &buf_ops_bio; bo->bo_private = vp; TAILQ_INIT(&bo->bo_clean.bv_hd); TAILQ_INIT(&bo->bo_dirty.bv_hd); /* * Initialize namecache. */ LIST_INIT(&vp->v_cache_src); TAILQ_INIT(&vp->v_cache_dst); /* * Finalize various vnode identity bits. */ vp->v_type = VNON; vp->v_tag = tag; vp->v_op = vops; v_incr_usecount(vp); vp->v_data = NULL; #ifdef MAC mac_vnode_init(vp); if (mp != NULL && (mp->mnt_flag & MNT_MULTILABEL) == 0) mac_vnode_associate_singlelabel(mp, vp); else if (mp == NULL && vops != &dead_vnodeops) printf("NULL mp in getnewvnode()\n"); #endif if (mp != NULL) { bo->bo_bsize = mp->mnt_stat.f_iosize; if ((mp->mnt_kern_flag & MNTK_NOKNOTE) != 0) vp->v_vflag |= VV_NOKNOTE; } rangelock_init(&vp->v_rl); /* * For the filesystems which do not use vfs_hash_insert(), * still initialize v_hash to have vfs_hash_index() useful. * E.g., nullfs uses vfs_hash_index() on the lower vnode for * its own hashing. */ vp->v_hash = (uintptr_t)vp >> vnsz2log; *vpp = vp; return (0); } /* * Delete from old mount point vnode list, if on one. */ static void delmntque(struct vnode *vp) { struct mount *mp; int active; mp = vp->v_mount; if (mp == NULL) return; MNT_ILOCK(mp); VI_LOCK(vp); KASSERT(mp->mnt_activevnodelistsize <= mp->mnt_nvnodelistsize, ("Active vnode list size %d > Vnode list size %d", mp->mnt_activevnodelistsize, mp->mnt_nvnodelistsize)); active = vp->v_iflag & VI_ACTIVE; vp->v_iflag &= ~VI_ACTIVE; if (active) { mtx_lock(&vnode_free_list_mtx); TAILQ_REMOVE(&mp->mnt_activevnodelist, vp, v_actfreelist); mp->mnt_activevnodelistsize--; mtx_unlock(&vnode_free_list_mtx); } vp->v_mount = NULL; VI_UNLOCK(vp); VNASSERT(mp->mnt_nvnodelistsize > 0, vp, ("bad mount point vnode list size")); TAILQ_REMOVE(&mp->mnt_nvnodelist, vp, v_nmntvnodes); mp->mnt_nvnodelistsize--; MNT_REL(mp); MNT_IUNLOCK(mp); } static void insmntque_stddtr(struct vnode *vp, void *dtr_arg) { vp->v_data = NULL; vp->v_op = &dead_vnodeops; vgone(vp); vput(vp); } /* * Insert into list of vnodes for the new mount point, if available. */ int insmntque1(struct vnode *vp, struct mount *mp, void (*dtr)(struct vnode *, void *), void *dtr_arg) { KASSERT(vp->v_mount == NULL, ("insmntque: vnode already on per mount vnode list")); VNASSERT(mp != NULL, vp, ("Don't call insmntque(foo, NULL)")); ASSERT_VOP_ELOCKED(vp, "insmntque: non-locked vp"); /* * We acquire the vnode interlock early to ensure that the * vnode cannot be recycled by another process releasing a * holdcnt on it before we get it on both the vnode list * and the active vnode list. The mount mutex protects only * manipulation of the vnode list and the vnode freelist * mutex protects only manipulation of the active vnode list. * Hence the need to hold the vnode interlock throughout. */ MNT_ILOCK(mp); VI_LOCK(vp); if (((mp->mnt_kern_flag & MNTK_NOINSMNTQ) != 0 && ((mp->mnt_kern_flag & MNTK_UNMOUNTF) != 0 || mp->mnt_nvnodelistsize == 0)) && (vp->v_vflag & VV_FORCEINSMQ) == 0) { VI_UNLOCK(vp); MNT_IUNLOCK(mp); if (dtr != NULL) dtr(vp, dtr_arg); return (EBUSY); } vp->v_mount = mp; MNT_REF(mp); TAILQ_INSERT_TAIL(&mp->mnt_nvnodelist, vp, v_nmntvnodes); VNASSERT(mp->mnt_nvnodelistsize >= 0, vp, ("neg mount point vnode list size")); mp->mnt_nvnodelistsize++; KASSERT((vp->v_iflag & VI_ACTIVE) == 0, ("Activating already active vnode")); vp->v_iflag |= VI_ACTIVE; mtx_lock(&vnode_free_list_mtx); TAILQ_INSERT_HEAD(&mp->mnt_activevnodelist, vp, v_actfreelist); mp->mnt_activevnodelistsize++; mtx_unlock(&vnode_free_list_mtx); VI_UNLOCK(vp); MNT_IUNLOCK(mp); return (0); } int insmntque(struct vnode *vp, struct mount *mp) { return (insmntque1(vp, mp, insmntque_stddtr, NULL)); } /* * Flush out and invalidate all buffers associated with a bufobj * Called with the underlying object locked. */ int bufobj_invalbuf(struct bufobj *bo, int flags, int slpflag, int slptimeo) { int error; BO_LOCK(bo); if (flags & V_SAVE) { error = bufobj_wwait(bo, slpflag, slptimeo); if (error) { BO_UNLOCK(bo); return (error); } if (bo->bo_dirty.bv_cnt > 0) { BO_UNLOCK(bo); if ((error = BO_SYNC(bo, MNT_WAIT)) != 0) return (error); /* * XXX We could save a lock/unlock if this was only * enabled under INVARIANTS */ BO_LOCK(bo); if (bo->bo_numoutput > 0 || bo->bo_dirty.bv_cnt > 0) panic("vinvalbuf: dirty bufs"); } } /* * If you alter this loop please notice that interlock is dropped and * reacquired in flushbuflist. Special care is needed to ensure that * no race conditions occur from this. */ do { error = flushbuflist(&bo->bo_clean, flags, bo, slpflag, slptimeo); if (error == 0 && !(flags & V_CLEANONLY)) error = flushbuflist(&bo->bo_dirty, flags, bo, slpflag, slptimeo); if (error != 0 && error != EAGAIN) { BO_UNLOCK(bo); return (error); } } while (error != 0); /* * Wait for I/O to complete. XXX needs cleaning up. The vnode can * have write I/O in-progress but if there is a VM object then the * VM object can also have read-I/O in-progress. */ do { bufobj_wwait(bo, 0, 0); BO_UNLOCK(bo); if (bo->bo_object != NULL) { VM_OBJECT_WLOCK(bo->bo_object); vm_object_pip_wait(bo->bo_object, "bovlbx"); VM_OBJECT_WUNLOCK(bo->bo_object); } BO_LOCK(bo); } while (bo->bo_numoutput > 0); BO_UNLOCK(bo); /* * Destroy the copy in the VM cache, too. */ if (bo->bo_object != NULL && (flags & (V_ALT | V_NORMAL | V_CLEANONLY)) == 0) { VM_OBJECT_WLOCK(bo->bo_object); vm_object_page_remove(bo->bo_object, 0, 0, (flags & V_SAVE) ? OBJPR_CLEANONLY : 0); VM_OBJECT_WUNLOCK(bo->bo_object); } #ifdef INVARIANTS BO_LOCK(bo); if ((flags & (V_ALT | V_NORMAL | V_CLEANONLY)) == 0 && (bo->bo_dirty.bv_cnt > 0 || bo->bo_clean.bv_cnt > 0)) panic("vinvalbuf: flush failed"); BO_UNLOCK(bo); #endif return (0); } /* * Flush out and invalidate all buffers associated with a vnode. * Called with the underlying object locked. */ int vinvalbuf(struct vnode *vp, int flags, int slpflag, int slptimeo) { CTR3(KTR_VFS, "%s: vp %p with flags %d", __func__, vp, flags); ASSERT_VOP_LOCKED(vp, "vinvalbuf"); if (vp->v_object != NULL && vp->v_object->handle != vp) return (0); return (bufobj_invalbuf(&vp->v_bufobj, flags, slpflag, slptimeo)); } /* * Flush out buffers on the specified list. * */ static int flushbuflist(struct bufv *bufv, int flags, struct bufobj *bo, int slpflag, int slptimeo) { struct buf *bp, *nbp; int retval, error; daddr_t lblkno; b_xflags_t xflags; ASSERT_BO_WLOCKED(bo); retval = 0; TAILQ_FOREACH_SAFE(bp, &bufv->bv_hd, b_bobufs, nbp) { if (((flags & V_NORMAL) && (bp->b_xflags & BX_ALTDATA)) || ((flags & V_ALT) && (bp->b_xflags & BX_ALTDATA) == 0)) { continue; } lblkno = 0; xflags = 0; if (nbp != NULL) { lblkno = nbp->b_lblkno; xflags = nbp->b_xflags & (BX_VNDIRTY | BX_VNCLEAN); } retval = EAGAIN; error = BUF_TIMELOCK(bp, LK_EXCLUSIVE | LK_SLEEPFAIL | LK_INTERLOCK, BO_LOCKPTR(bo), "flushbuf", slpflag, slptimeo); if (error) { BO_LOCK(bo); return (error != ENOLCK ? error : EAGAIN); } KASSERT(bp->b_bufobj == bo, ("bp %p wrong b_bufobj %p should be %p", bp, bp->b_bufobj, bo)); if (bp->b_bufobj != bo) { /* XXX: necessary ? */ BUF_UNLOCK(bp); BO_LOCK(bo); return (EAGAIN); } /* * XXX Since there are no node locks for NFS, I * believe there is a slight chance that a delayed * write will occur while sleeping just above, so * check for it. */ if (((bp->b_flags & (B_DELWRI | B_INVAL)) == B_DELWRI) && (flags & V_SAVE)) { bremfree(bp); bp->b_flags |= B_ASYNC; bwrite(bp); BO_LOCK(bo); return (EAGAIN); /* XXX: why not loop ? */ } bremfree(bp); bp->b_flags |= (B_INVAL | B_RELBUF); bp->b_flags &= ~B_ASYNC; brelse(bp); BO_LOCK(bo); if (nbp != NULL && (nbp->b_bufobj != bo || nbp->b_lblkno != lblkno || (nbp->b_xflags & (BX_VNDIRTY | BX_VNCLEAN)) != xflags)) break; /* nbp invalid */ } return (retval); } /* * Truncate a file's buffer and pages to a specified length. This * is in lieu of the old vinvalbuf mechanism, which performed unneeded * sync activity. */ int vtruncbuf(struct vnode *vp, struct ucred *cred, off_t length, int blksize) { struct buf *bp, *nbp; int anyfreed; int trunclbn; struct bufobj *bo; CTR5(KTR_VFS, "%s: vp %p with cred %p and block %d:%ju", __func__, vp, cred, blksize, (uintmax_t)length); /* * Round up to the *next* lbn. */ trunclbn = (length + blksize - 1) / blksize; ASSERT_VOP_LOCKED(vp, "vtruncbuf"); restart: bo = &vp->v_bufobj; BO_LOCK(bo); anyfreed = 1; for (;anyfreed;) { anyfreed = 0; TAILQ_FOREACH_SAFE(bp, &bo->bo_clean.bv_hd, b_bobufs, nbp) { if (bp->b_lblkno < trunclbn) continue; if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_SLEEPFAIL | LK_INTERLOCK, BO_LOCKPTR(bo)) == ENOLCK) goto restart; bremfree(bp); bp->b_flags |= (B_INVAL | B_RELBUF); bp->b_flags &= ~B_ASYNC; brelse(bp); anyfreed = 1; BO_LOCK(bo); if (nbp != NULL && (((nbp->b_xflags & BX_VNCLEAN) == 0) || (nbp->b_vp != vp) || (nbp->b_flags & B_DELWRI))) { BO_UNLOCK(bo); goto restart; } } TAILQ_FOREACH_SAFE(bp, &bo->bo_dirty.bv_hd, b_bobufs, nbp) { if (bp->b_lblkno < trunclbn) continue; if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_SLEEPFAIL | LK_INTERLOCK, BO_LOCKPTR(bo)) == ENOLCK) goto restart; bremfree(bp); bp->b_flags |= (B_INVAL | B_RELBUF); bp->b_flags &= ~B_ASYNC; brelse(bp); anyfreed = 1; BO_LOCK(bo); if (nbp != NULL && (((nbp->b_xflags & BX_VNDIRTY) == 0) || (nbp->b_vp != vp) || (nbp->b_flags & B_DELWRI) == 0)) { BO_UNLOCK(bo); goto restart; } } } if (length > 0) { restartsync: TAILQ_FOREACH_SAFE(bp, &bo->bo_dirty.bv_hd, b_bobufs, nbp) { if (bp->b_lblkno > 0) continue; /* * Since we hold the vnode lock this should only * fail if we're racing with the buf daemon. */ if (BUF_LOCK(bp, LK_EXCLUSIVE | LK_SLEEPFAIL | LK_INTERLOCK, BO_LOCKPTR(bo)) == ENOLCK) { goto restart; } VNASSERT((bp->b_flags & B_DELWRI), vp, ("buf(%p) on dirty queue without DELWRI", bp)); bremfree(bp); bawrite(bp); BO_LOCK(bo); goto restartsync; } } bufobj_wwait(bo, 0, 0); BO_UNLOCK(bo); vnode_pager_setsize(vp, length); return (0); } static void buf_vlist_remove(struct buf *bp) { struct bufv *bv; KASSERT(bp->b_bufobj != NULL, ("No b_bufobj %p", bp)); ASSERT_BO_WLOCKED(bp->b_bufobj); KASSERT((bp->b_xflags & (BX_VNDIRTY|BX_VNCLEAN)) != (BX_VNDIRTY|BX_VNCLEAN), ("buf_vlist_remove: Buf %p is on two lists", bp)); if (bp->b_xflags & BX_VNDIRTY) bv = &bp->b_bufobj->bo_dirty; else bv = &bp->b_bufobj->bo_clean; BUF_PCTRIE_REMOVE(&bv->bv_root, bp->b_lblkno); TAILQ_REMOVE(&bv->bv_hd, bp, b_bobufs); bv->bv_cnt--; bp->b_xflags &= ~(BX_VNDIRTY | BX_VNCLEAN); } /* * Add the buffer to the sorted clean or dirty block list. * * NOTE: xflags is passed as a constant, optimizing this inline function! */ static void buf_vlist_add(struct buf *bp, struct bufobj *bo, b_xflags_t xflags) { struct bufv *bv; struct buf *n; int error; ASSERT_BO_WLOCKED(bo); KASSERT((bp->b_xflags & (BX_VNDIRTY|BX_VNCLEAN)) == 0, ("buf_vlist_add: Buf %p has existing xflags %d", bp, bp->b_xflags)); bp->b_xflags |= xflags; if (xflags & BX_VNDIRTY) bv = &bo->bo_dirty; else bv = &bo->bo_clean; /* * Keep the list ordered. Optimize empty list insertion. Assume * we tend to grow at the tail so lookup_le should usually be cheaper * than _ge. */ if (bv->bv_cnt == 0 || bp->b_lblkno > TAILQ_LAST(&bv->bv_hd, buflists)->b_lblkno) TAILQ_INSERT_TAIL(&bv->bv_hd, bp, b_bobufs); else if ((n = BUF_PCTRIE_LOOKUP_LE(&bv->bv_root, bp->b_lblkno)) == NULL) TAILQ_INSERT_HEAD(&bv->bv_hd, bp, b_bobufs); else TAILQ_INSERT_AFTER(&bv->bv_hd, n, bp, b_bobufs); error = BUF_PCTRIE_INSERT(&bv->bv_root, bp); if (error) panic("buf_vlist_add: Preallocated nodes insufficient."); bv->bv_cnt++; } /* * Lookup a buffer using the splay tree. Note that we specifically avoid * shadow buffers used in background bitmap writes. * * This code isn't quite efficient as it could be because we are maintaining * two sorted lists and do not know which list the block resides in. * * During a "make buildworld" the desired buffer is found at one of * the roots more than 60% of the time. Thus, checking both roots * before performing either splay eliminates unnecessary splays on the * first tree splayed. */ struct buf * gbincore(struct bufobj *bo, daddr_t lblkno) { struct buf *bp; ASSERT_BO_LOCKED(bo); bp = BUF_PCTRIE_LOOKUP(&bo->bo_clean.bv_root, lblkno); if (bp != NULL) return (bp); return BUF_PCTRIE_LOOKUP(&bo->bo_dirty.bv_root, lblkno); } /* * Associate a buffer with a vnode. */ void bgetvp(struct vnode *vp, struct buf *bp) { struct bufobj *bo; bo = &vp->v_bufobj; ASSERT_BO_WLOCKED(bo); VNASSERT(bp->b_vp == NULL, bp->b_vp, ("bgetvp: not free")); CTR3(KTR_BUF, "bgetvp(%p) vp %p flags %X", bp, vp, bp->b_flags); VNASSERT((bp->b_xflags & (BX_VNDIRTY|BX_VNCLEAN)) == 0, vp, ("bgetvp: bp already attached! %p", bp)); vhold(vp); bp->b_vp = vp; bp->b_bufobj = bo; /* * Insert onto list for new vnode. */ buf_vlist_add(bp, bo, BX_VNCLEAN); } /* * Disassociate a buffer from a vnode. */ void brelvp(struct buf *bp) { struct bufobj *bo; struct vnode *vp; CTR3(KTR_BUF, "brelvp(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); KASSERT(bp->b_vp != NULL, ("brelvp: NULL")); /* * Delete from old vnode list, if on one. */ vp = bp->b_vp; /* XXX */ bo = bp->b_bufobj; BO_LOCK(bo); if (bp->b_xflags & (BX_VNDIRTY | BX_VNCLEAN)) buf_vlist_remove(bp); else panic("brelvp: Buffer %p not on queue.", bp); if ((bo->bo_flag & BO_ONWORKLST) && bo->bo_dirty.bv_cnt == 0) { bo->bo_flag &= ~BO_ONWORKLST; mtx_lock(&sync_mtx); LIST_REMOVE(bo, bo_synclist); syncer_worklist_len--; mtx_unlock(&sync_mtx); } bp->b_vp = NULL; bp->b_bufobj = NULL; BO_UNLOCK(bo); vdrop(vp); } /* * Add an item to the syncer work queue. */ static void vn_syncer_add_to_worklist(struct bufobj *bo, int delay) { int slot; ASSERT_BO_WLOCKED(bo); mtx_lock(&sync_mtx); if (bo->bo_flag & BO_ONWORKLST) LIST_REMOVE(bo, bo_synclist); else { bo->bo_flag |= BO_ONWORKLST; syncer_worklist_len++; } if (delay > syncer_maxdelay - 2) delay = syncer_maxdelay - 2; slot = (syncer_delayno + delay) & syncer_mask; LIST_INSERT_HEAD(&syncer_workitem_pending[slot], bo, bo_synclist); mtx_unlock(&sync_mtx); } static int sysctl_vfs_worklist_len(SYSCTL_HANDLER_ARGS) { int error, len; mtx_lock(&sync_mtx); len = syncer_worklist_len - sync_vnode_count; mtx_unlock(&sync_mtx); error = SYSCTL_OUT(req, &len, sizeof(len)); return (error); } SYSCTL_PROC(_vfs, OID_AUTO, worklist_len, CTLTYPE_INT | CTLFLAG_RD, NULL, 0, sysctl_vfs_worklist_len, "I", "Syncer thread worklist length"); static struct proc *updateproc; static void sched_sync(void); static struct kproc_desc up_kp = { "syncer", sched_sync, &updateproc }; SYSINIT(syncer, SI_SUB_KTHREAD_UPDATE, SI_ORDER_FIRST, kproc_start, &up_kp); static int sync_vnode(struct synclist *slp, struct bufobj **bo, struct thread *td) { struct vnode *vp; struct mount *mp; *bo = LIST_FIRST(slp); if (*bo == NULL) return (0); vp = (*bo)->__bo_vnode; /* XXX */ if (VOP_ISLOCKED(vp) != 0 || VI_TRYLOCK(vp) == 0) return (1); /* * We use vhold in case the vnode does not * successfully sync. vhold prevents the vnode from * going away when we unlock the sync_mtx so that * we can acquire the vnode interlock. */ vholdl(vp); mtx_unlock(&sync_mtx); VI_UNLOCK(vp); if (vn_start_write(vp, &mp, V_NOWAIT) != 0) { vdrop(vp); mtx_lock(&sync_mtx); return (*bo == LIST_FIRST(slp)); } vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); (void) VOP_FSYNC(vp, MNT_LAZY, td); VOP_UNLOCK(vp, 0); vn_finished_write(mp); BO_LOCK(*bo); if (((*bo)->bo_flag & BO_ONWORKLST) != 0) { /* * Put us back on the worklist. The worklist * routine will remove us from our current * position and then add us back in at a later * position. */ vn_syncer_add_to_worklist(*bo, syncdelay); } BO_UNLOCK(*bo); vdrop(vp); mtx_lock(&sync_mtx); return (0); } +static int first_printf = 1; + /* * System filesystem synchronizer daemon. */ static void sched_sync(void) { struct synclist *next, *slp; struct bufobj *bo; long starttime; struct thread *td = curthread; int last_work_seen; int net_worklist_len; int syncer_final_iter; - int first_printf; int error; last_work_seen = 0; syncer_final_iter = 0; - first_printf = 1; syncer_state = SYNCER_RUNNING; starttime = time_uptime; td->td_pflags |= TDP_NORUNNINGBUF; EVENTHANDLER_REGISTER(shutdown_pre_sync, syncer_shutdown, td->td_proc, SHUTDOWN_PRI_LAST); mtx_lock(&sync_mtx); for (;;) { if (syncer_state == SYNCER_FINAL_DELAY && syncer_final_iter == 0) { mtx_unlock(&sync_mtx); kproc_suspend_check(td->td_proc); mtx_lock(&sync_mtx); } net_worklist_len = syncer_worklist_len - sync_vnode_count; if (syncer_state != SYNCER_RUNNING && starttime != time_uptime) { if (first_printf) { printf("\nSyncing disks, vnodes remaining..."); first_printf = 0; } printf("%d ", net_worklist_len); } starttime = time_uptime; /* * Push files whose dirty time has expired. Be careful * of interrupt race on slp queue. * * Skip over empty worklist slots when shutting down. */ do { slp = &syncer_workitem_pending[syncer_delayno]; syncer_delayno += 1; if (syncer_delayno == syncer_maxdelay) syncer_delayno = 0; next = &syncer_workitem_pending[syncer_delayno]; /* * If the worklist has wrapped since the * it was emptied of all but syncer vnodes, * switch to the FINAL_DELAY state and run * for one more second. */ if (syncer_state == SYNCER_SHUTTING_DOWN && net_worklist_len == 0 && last_work_seen == syncer_delayno) { syncer_state = SYNCER_FINAL_DELAY; syncer_final_iter = SYNCER_SHUTDOWN_SPEEDUP; } } while (syncer_state != SYNCER_RUNNING && LIST_EMPTY(slp) && syncer_worklist_len > 0); /* * Keep track of the last time there was anything * on the worklist other than syncer vnodes. * Return to the SHUTTING_DOWN state if any * new work appears. */ if (net_worklist_len > 0 || syncer_state == SYNCER_RUNNING) last_work_seen = syncer_delayno; if (net_worklist_len > 0 && syncer_state == SYNCER_FINAL_DELAY) syncer_state = SYNCER_SHUTTING_DOWN; while (!LIST_EMPTY(slp)) { error = sync_vnode(slp, &bo, td); if (error == 1) { LIST_REMOVE(bo, bo_synclist); LIST_INSERT_HEAD(next, bo, bo_synclist); continue; } if (first_printf == 0) { /* * Drop the sync mutex, because some watchdog * drivers need to sleep while patting */ mtx_unlock(&sync_mtx); wdog_kern_pat(WD_LASTVAL); mtx_lock(&sync_mtx); } } if (syncer_state == SYNCER_FINAL_DELAY && syncer_final_iter > 0) syncer_final_iter--; /* * The variable rushjob allows the kernel to speed up the * processing of the filesystem syncer process. A rushjob * value of N tells the filesystem syncer to process the next * N seconds worth of work on its queue ASAP. Currently rushjob * is used by the soft update code to speed up the filesystem * syncer process when the incore state is getting so far * ahead of the disk that the kernel memory pool is being * threatened with exhaustion. */ if (rushjob > 0) { rushjob -= 1; continue; } /* * Just sleep for a short period of time between * iterations when shutting down to allow some I/O * to happen. * * If it has taken us less than a second to process the * current work, then wait. Otherwise start right over * again. We can still lose time if any single round * takes more than two seconds, but it does not really * matter as we are just trying to generally pace the * filesystem activity. */ if (syncer_state != SYNCER_RUNNING || time_uptime == starttime) { thread_lock(td); sched_prio(td, PPAUSE); thread_unlock(td); } if (syncer_state != SYNCER_RUNNING) cv_timedwait(&sync_wakeup, &sync_mtx, hz / SYNCER_SHUTDOWN_SPEEDUP); else if (time_uptime == starttime) cv_timedwait(&sync_wakeup, &sync_mtx, hz); } } /* * Request the syncer daemon to speed up its work. * We never push it to speed up more than half of its * normal turn time, otherwise it could take over the cpu. */ int speedup_syncer(void) { int ret = 0; mtx_lock(&sync_mtx); if (rushjob < syncdelay / 2) { rushjob += 1; stat_rush_requests += 1; ret = 1; } mtx_unlock(&sync_mtx); cv_broadcast(&sync_wakeup); return (ret); } /* * Tell the syncer to speed up its work and run though its work * list several times, then tell it to shut down. */ static void syncer_shutdown(void *arg, int howto) { if (howto & RB_NOSYNC) return; mtx_lock(&sync_mtx); syncer_state = SYNCER_SHUTTING_DOWN; rushjob = 0; mtx_unlock(&sync_mtx); cv_broadcast(&sync_wakeup); kproc_shutdown(arg, howto); +} + +void +syncer_suspend(void) +{ + + syncer_shutdown(updateproc, 0); +} + +void +syncer_resume(void) +{ + + mtx_lock(&sync_mtx); + first_printf = 1; + syncer_state = SYNCER_RUNNING; + mtx_unlock(&sync_mtx); + cv_broadcast(&sync_wakeup); + kproc_resume(updateproc); } /* * Reassign a buffer from one vnode to another. * Used to assign file specific control information * (indirect blocks) to the vnode to which they belong. */ void reassignbuf(struct buf *bp) { struct vnode *vp; struct bufobj *bo; int delay; #ifdef INVARIANTS struct bufv *bv; #endif vp = bp->b_vp; bo = bp->b_bufobj; ++reassignbufcalls; CTR3(KTR_BUF, "reassignbuf(%p) vp %p flags %X", bp, bp->b_vp, bp->b_flags); /* * B_PAGING flagged buffers cannot be reassigned because their vp * is not fully linked in. */ if (bp->b_flags & B_PAGING) panic("cannot reassign paging buffer"); /* * Delete from old vnode list, if on one. */ BO_LOCK(bo); if (bp->b_xflags & (BX_VNDIRTY | BX_VNCLEAN)) buf_vlist_remove(bp); else panic("reassignbuf: Buffer %p not on queue.", bp); /* * If dirty, put on list of dirty buffers; otherwise insert onto list * of clean buffers. */ if (bp->b_flags & B_DELWRI) { if ((bo->bo_flag & BO_ONWORKLST) == 0) { switch (vp->v_type) { case VDIR: delay = dirdelay; break; case VCHR: delay = metadelay; break; default: delay = filedelay; } vn_syncer_add_to_worklist(bo, delay); } buf_vlist_add(bp, bo, BX_VNDIRTY); } else { buf_vlist_add(bp, bo, BX_VNCLEAN); if ((bo->bo_flag & BO_ONWORKLST) && bo->bo_dirty.bv_cnt == 0) { mtx_lock(&sync_mtx); LIST_REMOVE(bo, bo_synclist); syncer_worklist_len--; mtx_unlock(&sync_mtx); bo->bo_flag &= ~BO_ONWORKLST; } } #ifdef INVARIANTS bv = &bo->bo_clean; bp = TAILQ_FIRST(&bv->bv_hd); KASSERT(bp == NULL || bp->b_bufobj == bo, ("bp %p wrong b_bufobj %p should be %p", bp, bp->b_bufobj, bo)); bp = TAILQ_LAST(&bv->bv_hd, buflists); KASSERT(bp == NULL || bp->b_bufobj == bo, ("bp %p wrong b_bufobj %p should be %p", bp, bp->b_bufobj, bo)); bv = &bo->bo_dirty; bp = TAILQ_FIRST(&bv->bv_hd); KASSERT(bp == NULL || bp->b_bufobj == bo, ("bp %p wrong b_bufobj %p should be %p", bp, bp->b_bufobj, bo)); bp = TAILQ_LAST(&bv->bv_hd, buflists); KASSERT(bp == NULL || bp->b_bufobj == bo, ("bp %p wrong b_bufobj %p should be %p", bp, bp->b_bufobj, bo)); #endif BO_UNLOCK(bo); } /* * Increment the use and hold counts on the vnode, taking care to reference * the driver's usecount if this is a chardev. The vholdl() will remove * the vnode from the free list if it is presently free. Requires the * vnode interlock and returns with it held. */ static void v_incr_usecount(struct vnode *vp) { CTR2(KTR_VFS, "%s: vp %p", __func__, vp); vholdl(vp); vp->v_usecount++; if (vp->v_type == VCHR && vp->v_rdev != NULL) { dev_lock(); vp->v_rdev->si_usecount++; dev_unlock(); } } /* * Turn a holdcnt into a use+holdcnt such that only one call to * v_decr_usecount is needed. */ static void v_upgrade_usecount(struct vnode *vp) { CTR2(KTR_VFS, "%s: vp %p", __func__, vp); vp->v_usecount++; if (vp->v_type == VCHR && vp->v_rdev != NULL) { dev_lock(); vp->v_rdev->si_usecount++; dev_unlock(); } } /* * Decrement the vnode use and hold count along with the driver's usecount * if this is a chardev. The vdropl() below releases the vnode interlock * as it may free the vnode. */ static void v_decr_usecount(struct vnode *vp) { ASSERT_VI_LOCKED(vp, __FUNCTION__); VNASSERT(vp->v_usecount > 0, vp, ("v_decr_usecount: negative usecount")); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); vp->v_usecount--; if (vp->v_type == VCHR && vp->v_rdev != NULL) { dev_lock(); vp->v_rdev->si_usecount--; dev_unlock(); } vdropl(vp); } /* * Decrement only the use count and driver use count. This is intended to * be paired with a follow on vdropl() to release the remaining hold count. * In this way we may vgone() a vnode with a 0 usecount without risk of * having it end up on a free list because the hold count is kept above 0. */ static void v_decr_useonly(struct vnode *vp) { ASSERT_VI_LOCKED(vp, __FUNCTION__); VNASSERT(vp->v_usecount > 0, vp, ("v_decr_useonly: negative usecount")); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); vp->v_usecount--; if (vp->v_type == VCHR && vp->v_rdev != NULL) { dev_lock(); vp->v_rdev->si_usecount--; dev_unlock(); } } /* * Grab a particular vnode from the free list, increment its * reference count and lock it. VI_DOOMED is set if the vnode * is being destroyed. Only callers who specify LK_RETRY will * see doomed vnodes. If inactive processing was delayed in * vput try to do it here. */ int vget(struct vnode *vp, int flags, struct thread *td) { int error; error = 0; VNASSERT((flags & LK_TYPE_MASK) != 0, vp, ("vget: invalid lock operation")); CTR3(KTR_VFS, "%s: vp %p with flags %d", __func__, vp, flags); if ((flags & LK_INTERLOCK) == 0) VI_LOCK(vp); vholdl(vp); if ((error = vn_lock(vp, flags | LK_INTERLOCK)) != 0) { vdrop(vp); CTR2(KTR_VFS, "%s: impossible to lock vnode %p", __func__, vp); return (error); } if (vp->v_iflag & VI_DOOMED && (flags & LK_RETRY) == 0) panic("vget: vn_lock failed to return ENOENT\n"); VI_LOCK(vp); /* Upgrade our holdcnt to a usecount. */ v_upgrade_usecount(vp); /* * We don't guarantee that any particular close will * trigger inactive processing so just make a best effort * here at preventing a reference to a removed file. If * we don't succeed no harm is done. */ if (vp->v_iflag & VI_OWEINACT) { if (VOP_ISLOCKED(vp) == LK_EXCLUSIVE && (flags & LK_NOWAIT) == 0) vinactive(vp, td); vp->v_iflag &= ~VI_OWEINACT; } VI_UNLOCK(vp); return (0); } /* * Increase the reference count of a vnode. */ void vref(struct vnode *vp) { CTR2(KTR_VFS, "%s: vp %p", __func__, vp); VI_LOCK(vp); v_incr_usecount(vp); VI_UNLOCK(vp); } /* * Return reference count of a vnode. * * The results of this call are only guaranteed when some mechanism other * than the VI lock is used to stop other processes from gaining references * to the vnode. This may be the case if the caller holds the only reference. * This is also useful when stale data is acceptable as race conditions may * be accounted for by some other means. */ int vrefcnt(struct vnode *vp) { int usecnt; VI_LOCK(vp); usecnt = vp->v_usecount; VI_UNLOCK(vp); return (usecnt); } #define VPUTX_VRELE 1 #define VPUTX_VPUT 2 #define VPUTX_VUNREF 3 static void vputx(struct vnode *vp, int func) { int error; KASSERT(vp != NULL, ("vputx: null vp")); if (func == VPUTX_VUNREF) ASSERT_VOP_LOCKED(vp, "vunref"); else if (func == VPUTX_VPUT) ASSERT_VOP_LOCKED(vp, "vput"); else KASSERT(func == VPUTX_VRELE, ("vputx: wrong func")); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); VI_LOCK(vp); /* Skip this v_writecount check if we're going to panic below. */ VNASSERT(vp->v_writecount < vp->v_usecount || vp->v_usecount < 1, vp, ("vputx: missed vn_close")); error = 0; if (vp->v_usecount > 1 || ((vp->v_iflag & VI_DOINGINACT) && vp->v_usecount == 1)) { if (func == VPUTX_VPUT) VOP_UNLOCK(vp, 0); v_decr_usecount(vp); return; } if (vp->v_usecount != 1) { vprint("vputx: negative ref count", vp); panic("vputx: negative ref cnt"); } CTR2(KTR_VFS, "%s: return vnode %p to the freelist", __func__, vp); /* * We want to hold the vnode until the inactive finishes to * prevent vgone() races. We drop the use count here and the * hold count below when we're done. */ v_decr_useonly(vp); /* * We must call VOP_INACTIVE with the node locked. Mark * as VI_DOINGINACT to avoid recursion. */ vp->v_iflag |= VI_OWEINACT; switch (func) { case VPUTX_VRELE: error = vn_lock(vp, LK_EXCLUSIVE | LK_INTERLOCK); VI_LOCK(vp); break; case VPUTX_VPUT: if (VOP_ISLOCKED(vp) != LK_EXCLUSIVE) { error = VOP_LOCK(vp, LK_UPGRADE | LK_INTERLOCK | LK_NOWAIT); VI_LOCK(vp); } break; case VPUTX_VUNREF: if (VOP_ISLOCKED(vp) != LK_EXCLUSIVE) { error = VOP_LOCK(vp, LK_TRYUPGRADE | LK_INTERLOCK); VI_LOCK(vp); } break; } if (vp->v_usecount > 0) vp->v_iflag &= ~VI_OWEINACT; if (error == 0) { if (vp->v_iflag & VI_OWEINACT) vinactive(vp, curthread); if (func != VPUTX_VUNREF) VOP_UNLOCK(vp, 0); } vdropl(vp); } /* * Vnode put/release. * If count drops to zero, call inactive routine and return to freelist. */ void vrele(struct vnode *vp) { vputx(vp, VPUTX_VRELE); } /* * Release an already locked vnode. This give the same effects as * unlock+vrele(), but takes less time and avoids releasing and * re-aquiring the lock (as vrele() acquires the lock internally.) */ void vput(struct vnode *vp) { vputx(vp, VPUTX_VPUT); } /* * Release an exclusively locked vnode. Do not unlock the vnode lock. */ void vunref(struct vnode *vp) { vputx(vp, VPUTX_VUNREF); } /* * Somebody doesn't want the vnode recycled. */ void vhold(struct vnode *vp) { VI_LOCK(vp); vholdl(vp); VI_UNLOCK(vp); } /* * Increase the hold count and activate if this is the first reference. */ void vholdl(struct vnode *vp) { struct mount *mp; CTR2(KTR_VFS, "%s: vp %p", __func__, vp); #ifdef INVARIANTS /* getnewvnode() calls v_incr_usecount() without holding interlock. */ if (vp->v_type != VNON || vp->v_data != NULL) { ASSERT_VI_LOCKED(vp, "vholdl"); VNASSERT(vp->v_holdcnt > 0 || (vp->v_iflag & VI_FREE) != 0, vp, ("vholdl: free vnode is held")); } #endif vp->v_holdcnt++; if ((vp->v_iflag & VI_FREE) == 0) return; VNASSERT(vp->v_holdcnt == 1, vp, ("vholdl: wrong hold count")); VNASSERT(vp->v_op != NULL, vp, ("vholdl: vnode already reclaimed.")); /* * Remove a vnode from the free list, mark it as in use, * and put it on the active list. */ mtx_lock(&vnode_free_list_mtx); TAILQ_REMOVE(&vnode_free_list, vp, v_actfreelist); freevnodes--; vp->v_iflag &= ~(VI_FREE|VI_AGE); KASSERT((vp->v_iflag & VI_ACTIVE) == 0, ("Activating already active vnode")); vp->v_iflag |= VI_ACTIVE; mp = vp->v_mount; TAILQ_INSERT_HEAD(&mp->mnt_activevnodelist, vp, v_actfreelist); mp->mnt_activevnodelistsize++; mtx_unlock(&vnode_free_list_mtx); } /* * Note that there is one less who cares about this vnode. * vdrop() is the opposite of vhold(). */ void vdrop(struct vnode *vp) { VI_LOCK(vp); vdropl(vp); } /* * Drop the hold count of the vnode. If this is the last reference to * the vnode we place it on the free list unless it has been vgone'd * (marked VI_DOOMED) in which case we will free it. */ void vdropl(struct vnode *vp) { struct bufobj *bo; struct mount *mp; int active; ASSERT_VI_LOCKED(vp, "vdropl"); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); if (vp->v_holdcnt <= 0) panic("vdrop: holdcnt %d", vp->v_holdcnt); vp->v_holdcnt--; VNASSERT(vp->v_holdcnt >= vp->v_usecount, vp, ("hold count less than use count")); if (vp->v_holdcnt > 0) { VI_UNLOCK(vp); return; } if ((vp->v_iflag & VI_DOOMED) == 0) { /* * Mark a vnode as free: remove it from its active list * and put it up for recycling on the freelist. */ VNASSERT(vp->v_op != NULL, vp, ("vdropl: vnode already reclaimed.")); VNASSERT((vp->v_iflag & VI_FREE) == 0, vp, ("vnode already free")); VNASSERT(vp->v_holdcnt == 0, vp, ("vdropl: freeing when we shouldn't")); active = vp->v_iflag & VI_ACTIVE; vp->v_iflag &= ~VI_ACTIVE; mp = vp->v_mount; mtx_lock(&vnode_free_list_mtx); if (active) { TAILQ_REMOVE(&mp->mnt_activevnodelist, vp, v_actfreelist); mp->mnt_activevnodelistsize--; } if (vp->v_iflag & VI_AGE) { TAILQ_INSERT_HEAD(&vnode_free_list, vp, v_actfreelist); } else { TAILQ_INSERT_TAIL(&vnode_free_list, vp, v_actfreelist); } freevnodes++; vp->v_iflag &= ~VI_AGE; vp->v_iflag |= VI_FREE; mtx_unlock(&vnode_free_list_mtx); VI_UNLOCK(vp); return; } /* * The vnode has been marked for destruction, so free it. */ CTR2(KTR_VFS, "%s: destroying the vnode %p", __func__, vp); atomic_subtract_long(&numvnodes, 1); bo = &vp->v_bufobj; VNASSERT((vp->v_iflag & VI_FREE) == 0, vp, ("cleaned vnode still on the free list.")); VNASSERT(vp->v_data == NULL, vp, ("cleaned vnode isn't")); VNASSERT(vp->v_holdcnt == 0, vp, ("Non-zero hold count")); VNASSERT(vp->v_usecount == 0, vp, ("Non-zero use count")); VNASSERT(vp->v_writecount == 0, vp, ("Non-zero write count")); VNASSERT(bo->bo_numoutput == 0, vp, ("Clean vnode has pending I/O's")); VNASSERT(bo->bo_clean.bv_cnt == 0, vp, ("cleanbufcnt not 0")); VNASSERT(pctrie_is_empty(&bo->bo_clean.bv_root), vp, ("clean blk trie not empty")); VNASSERT(bo->bo_dirty.bv_cnt == 0, vp, ("dirtybufcnt not 0")); VNASSERT(pctrie_is_empty(&bo->bo_dirty.bv_root), vp, ("dirty blk trie not empty")); VNASSERT(TAILQ_EMPTY(&vp->v_cache_dst), vp, ("vp has namecache dst")); VNASSERT(LIST_EMPTY(&vp->v_cache_src), vp, ("vp has namecache src")); VNASSERT(vp->v_cache_dd == NULL, vp, ("vp has namecache for ..")); VI_UNLOCK(vp); #ifdef MAC mac_vnode_destroy(vp); #endif if (vp->v_pollinfo != NULL) destroy_vpollinfo(vp->v_pollinfo); #ifdef INVARIANTS /* XXX Elsewhere we detect an already freed vnode via NULL v_op. */ vp->v_op = NULL; #endif rangelock_destroy(&vp->v_rl); lockdestroy(vp->v_vnlock); mtx_destroy(&vp->v_interlock); rw_destroy(BO_LOCKPTR(bo)); uma_zfree(vnode_zone, vp); } /* * Call VOP_INACTIVE on the vnode and manage the DOINGINACT and OWEINACT * flags. DOINGINACT prevents us from recursing in calls to vinactive. * OWEINACT tracks whether a vnode missed a call to inactive due to a * failed lock upgrade. */ void vinactive(struct vnode *vp, struct thread *td) { struct vm_object *obj; ASSERT_VOP_ELOCKED(vp, "vinactive"); ASSERT_VI_LOCKED(vp, "vinactive"); VNASSERT((vp->v_iflag & VI_DOINGINACT) == 0, vp, ("vinactive: recursed on VI_DOINGINACT")); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); vp->v_iflag |= VI_DOINGINACT; vp->v_iflag &= ~VI_OWEINACT; VI_UNLOCK(vp); /* * Before moving off the active list, we must be sure that any * modified pages are on the vnode's dirty list since these will * no longer be checked once the vnode is on the inactive list. * Because the vnode vm object keeps a hold reference on the vnode * if there is at least one resident non-cached page, the vnode * cannot leave the active list without the page cleanup done. */ obj = vp->v_object; if (obj != NULL && (obj->flags & OBJ_MIGHTBEDIRTY) != 0) { VM_OBJECT_WLOCK(obj); vm_object_page_clean(obj, 0, 0, OBJPC_NOSYNC); VM_OBJECT_WUNLOCK(obj); } VOP_INACTIVE(vp, td); VI_LOCK(vp); VNASSERT(vp->v_iflag & VI_DOINGINACT, vp, ("vinactive: lost VI_DOINGINACT")); vp->v_iflag &= ~VI_DOINGINACT; } /* * Remove any vnodes in the vnode table belonging to mount point mp. * * If FORCECLOSE is not specified, there should not be any active ones, * return error if any are found (nb: this is a user error, not a * system error). If FORCECLOSE is specified, detach any active vnodes * that are found. * * If WRITECLOSE is set, only flush out regular file vnodes open for * writing. * * SKIPSYSTEM causes any vnodes marked VV_SYSTEM to be skipped. * * `rootrefs' specifies the base reference count for the root vnode * of this filesystem. The root vnode is considered busy if its * v_usecount exceeds this value. On a successful return, vflush(, td) * will call vrele() on the root vnode exactly rootrefs times. * If the SKIPSYSTEM or WRITECLOSE flags are specified, rootrefs must * be zero. */ #ifdef DIAGNOSTIC static int busyprt = 0; /* print out busy vnodes */ SYSCTL_INT(_debug, OID_AUTO, busyprt, CTLFLAG_RW, &busyprt, 0, "Print out busy vnodes"); #endif int vflush(struct mount *mp, int rootrefs, int flags, struct thread *td) { struct vnode *vp, *mvp, *rootvp = NULL; struct vattr vattr; int busy = 0, error; CTR4(KTR_VFS, "%s: mp %p with rootrefs %d and flags %d", __func__, mp, rootrefs, flags); if (rootrefs > 0) { KASSERT((flags & (SKIPSYSTEM | WRITECLOSE)) == 0, ("vflush: bad args")); /* * Get the filesystem root vnode. We can vput() it * immediately, since with rootrefs > 0, it won't go away. */ if ((error = VFS_ROOT(mp, LK_EXCLUSIVE, &rootvp)) != 0) { CTR2(KTR_VFS, "%s: vfs_root lookup failed with %d", __func__, error); return (error); } vput(rootvp); } loop: MNT_VNODE_FOREACH_ALL(vp, mp, mvp) { vholdl(vp); error = vn_lock(vp, LK_INTERLOCK | LK_EXCLUSIVE); if (error) { vdrop(vp); MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp); goto loop; } /* * Skip over a vnodes marked VV_SYSTEM. */ if ((flags & SKIPSYSTEM) && (vp->v_vflag & VV_SYSTEM)) { VOP_UNLOCK(vp, 0); vdrop(vp); continue; } /* * If WRITECLOSE is set, flush out unlinked but still open * files (even if open only for reading) and regular file * vnodes open for writing. */ if (flags & WRITECLOSE) { if (vp->v_object != NULL) { VM_OBJECT_WLOCK(vp->v_object); vm_object_page_clean(vp->v_object, 0, 0, 0); VM_OBJECT_WUNLOCK(vp->v_object); } error = VOP_FSYNC(vp, MNT_WAIT, td); if (error != 0) { VOP_UNLOCK(vp, 0); vdrop(vp); MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp); return (error); } error = VOP_GETATTR(vp, &vattr, td->td_ucred); VI_LOCK(vp); if ((vp->v_type == VNON || (error == 0 && vattr.va_nlink > 0)) && (vp->v_writecount == 0 || vp->v_type != VREG)) { VOP_UNLOCK(vp, 0); vdropl(vp); continue; } } else VI_LOCK(vp); /* * With v_usecount == 0, all we need to do is clear out the * vnode data structures and we are done. * * If FORCECLOSE is set, forcibly close the vnode. */ if (vp->v_usecount == 0 || (flags & FORCECLOSE)) { VNASSERT(vp->v_usecount == 0 || (vp->v_type != VCHR && vp->v_type != VBLK), vp, ("device VNODE %p is FORCECLOSED", vp)); vgonel(vp); } else { busy++; #ifdef DIAGNOSTIC if (busyprt) vprint("vflush: busy vnode", vp); #endif } VOP_UNLOCK(vp, 0); vdropl(vp); } if (rootrefs > 0 && (flags & FORCECLOSE) == 0) { /* * If just the root vnode is busy, and if its refcount * is equal to `rootrefs', then go ahead and kill it. */ VI_LOCK(rootvp); KASSERT(busy > 0, ("vflush: not busy")); VNASSERT(rootvp->v_usecount >= rootrefs, rootvp, ("vflush: usecount %d < rootrefs %d", rootvp->v_usecount, rootrefs)); if (busy == 1 && rootvp->v_usecount == rootrefs) { VOP_LOCK(rootvp, LK_EXCLUSIVE|LK_INTERLOCK); vgone(rootvp); VOP_UNLOCK(rootvp, 0); busy = 0; } else VI_UNLOCK(rootvp); } if (busy) { CTR2(KTR_VFS, "%s: failing as %d vnodes are busy", __func__, busy); return (EBUSY); } for (; rootrefs > 0; rootrefs--) vrele(rootvp); return (0); } /* * Recycle an unused vnode to the front of the free list. */ int vrecycle(struct vnode *vp) { int recycled; ASSERT_VOP_ELOCKED(vp, "vrecycle"); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); recycled = 0; VI_LOCK(vp); if (vp->v_usecount == 0) { recycled = 1; vgonel(vp); } VI_UNLOCK(vp); return (recycled); } /* * Eliminate all activity associated with a vnode * in preparation for reuse. */ void vgone(struct vnode *vp) { VI_LOCK(vp); vgonel(vp); VI_UNLOCK(vp); } static void notify_lowervp_vfs_dummy(struct mount *mp __unused, struct vnode *lowervp __unused) { } /* * Notify upper mounts about reclaimed or unlinked vnode. */ void vfs_notify_upper(struct vnode *vp, int event) { static struct vfsops vgonel_vfsops = { .vfs_reclaim_lowervp = notify_lowervp_vfs_dummy, .vfs_unlink_lowervp = notify_lowervp_vfs_dummy, }; struct mount *mp, *ump, *mmp; mp = vp->v_mount; if (mp == NULL) return; MNT_ILOCK(mp); if (TAILQ_EMPTY(&mp->mnt_uppers)) goto unlock; MNT_IUNLOCK(mp); mmp = malloc(sizeof(struct mount), M_TEMP, M_WAITOK | M_ZERO); mmp->mnt_op = &vgonel_vfsops; mmp->mnt_kern_flag |= MNTK_MARKER; MNT_ILOCK(mp); mp->mnt_kern_flag |= MNTK_VGONE_UPPER; for (ump = TAILQ_FIRST(&mp->mnt_uppers); ump != NULL;) { if ((ump->mnt_kern_flag & MNTK_MARKER) != 0) { ump = TAILQ_NEXT(ump, mnt_upper_link); continue; } TAILQ_INSERT_AFTER(&mp->mnt_uppers, ump, mmp, mnt_upper_link); MNT_IUNLOCK(mp); switch (event) { case VFS_NOTIFY_UPPER_RECLAIM: VFS_RECLAIM_LOWERVP(ump, vp); break; case VFS_NOTIFY_UPPER_UNLINK: VFS_UNLINK_LOWERVP(ump, vp); break; default: KASSERT(0, ("invalid event %d", event)); break; } MNT_ILOCK(mp); ump = TAILQ_NEXT(mmp, mnt_upper_link); TAILQ_REMOVE(&mp->mnt_uppers, mmp, mnt_upper_link); } free(mmp, M_TEMP); mp->mnt_kern_flag &= ~MNTK_VGONE_UPPER; if ((mp->mnt_kern_flag & MNTK_VGONE_WAITER) != 0) { mp->mnt_kern_flag &= ~MNTK_VGONE_WAITER; wakeup(&mp->mnt_uppers); } unlock: MNT_IUNLOCK(mp); } /* * vgone, with the vp interlock held. */ void vgonel(struct vnode *vp) { struct thread *td; int oweinact; int active; struct mount *mp; ASSERT_VOP_ELOCKED(vp, "vgonel"); ASSERT_VI_LOCKED(vp, "vgonel"); VNASSERT(vp->v_holdcnt, vp, ("vgonel: vp %p has no reference.", vp)); CTR2(KTR_VFS, "%s: vp %p", __func__, vp); td = curthread; /* * Don't vgonel if we're already doomed. */ if (vp->v_iflag & VI_DOOMED) return; vp->v_iflag |= VI_DOOMED; /* * Check to see if the vnode is in use. If so, we have to call * VOP_CLOSE() and VOP_INACTIVE(). */ active = vp->v_usecount; oweinact = (vp->v_iflag & VI_OWEINACT); VI_UNLOCK(vp); vfs_notify_upper(vp, VFS_NOTIFY_UPPER_RECLAIM); /* * Clean out any buffers associated with the vnode. * If the flush fails, just toss the buffers. */ mp = NULL; if (!TAILQ_EMPTY(&vp->v_bufobj.bo_dirty.bv_hd)) (void) vn_start_secondary_write(vp, &mp, V_WAIT); if (vinvalbuf(vp, V_SAVE, 0, 0) != 0) vinvalbuf(vp, 0, 0, 0); /* * If purging an active vnode, it must be closed and * deactivated before being reclaimed. */ if (active) VOP_CLOSE(vp, FNONBLOCK, NOCRED, td); if (oweinact || active) { VI_LOCK(vp); if ((vp->v_iflag & VI_DOINGINACT) == 0) vinactive(vp, td); VI_UNLOCK(vp); } if (vp->v_type == VSOCK) vfs_unp_reclaim(vp); /* * Reclaim the vnode. */ if (VOP_RECLAIM(vp, td)) panic("vgone: cannot reclaim"); if (mp != NULL) vn_finished_secondary_write(mp); VNASSERT(vp->v_object == NULL, vp, ("vop_reclaim left v_object vp=%p, tag=%s", vp, vp->v_tag)); /* * Clear the advisory locks and wake up waiting threads. */ (void)VOP_ADVLOCKPURGE(vp); /* * Delete from old mount point vnode list. */ delmntque(vp); cache_purge(vp); /* * Done with purge, reset to the standard lock and invalidate * the vnode. */ VI_LOCK(vp); vp->v_vnlock = &vp->v_lock; vp->v_op = &dead_vnodeops; vp->v_tag = "none"; vp->v_type = VBAD; } /* * Calculate the total number of references to a special device. */ int vcount(struct vnode *vp) { int count; dev_lock(); count = vp->v_rdev->si_usecount; dev_unlock(); return (count); } /* * Same as above, but using the struct cdev *as argument */ int count_dev(struct cdev *dev) { int count; dev_lock(); count = dev->si_usecount; dev_unlock(); return(count); } /* * Print out a description of a vnode. */ static char *typename[] = {"VNON", "VREG", "VDIR", "VBLK", "VCHR", "VLNK", "VSOCK", "VFIFO", "VBAD", "VMARKER"}; void vn_printf(struct vnode *vp, const char *fmt, ...) { va_list ap; char buf[256], buf2[16]; u_long flags; va_start(ap, fmt); vprintf(fmt, ap); va_end(ap); printf("%p: ", (void *)vp); printf("tag %s, type %s\n", vp->v_tag, typename[vp->v_type]); printf(" usecount %d, writecount %d, refcount %d mountedhere %p\n", vp->v_usecount, vp->v_writecount, vp->v_holdcnt, vp->v_mountedhere); buf[0] = '\0'; buf[1] = '\0'; if (vp->v_vflag & VV_ROOT) strlcat(buf, "|VV_ROOT", sizeof(buf)); if (vp->v_vflag & VV_ISTTY) strlcat(buf, "|VV_ISTTY", sizeof(buf)); if (vp->v_vflag & VV_NOSYNC) strlcat(buf, "|VV_NOSYNC", sizeof(buf)); if (vp->v_vflag & VV_ETERNALDEV) strlcat(buf, "|VV_ETERNALDEV", sizeof(buf)); if (vp->v_vflag & VV_CACHEDLABEL) strlcat(buf, "|VV_CACHEDLABEL", sizeof(buf)); if (vp->v_vflag & VV_TEXT) strlcat(buf, "|VV_TEXT", sizeof(buf)); if (vp->v_vflag & VV_COPYONWRITE) strlcat(buf, "|VV_COPYONWRITE", sizeof(buf)); if (vp->v_vflag & VV_SYSTEM) strlcat(buf, "|VV_SYSTEM", sizeof(buf)); if (vp->v_vflag & VV_PROCDEP) strlcat(buf, "|VV_PROCDEP", sizeof(buf)); if (vp->v_vflag & VV_NOKNOTE) strlcat(buf, "|VV_NOKNOTE", sizeof(buf)); if (vp->v_vflag & VV_DELETED) strlcat(buf, "|VV_DELETED", sizeof(buf)); if (vp->v_vflag & VV_MD) strlcat(buf, "|VV_MD", sizeof(buf)); if (vp->v_vflag & VV_FORCEINSMQ) strlcat(buf, "|VV_FORCEINSMQ", sizeof(buf)); flags = vp->v_vflag & ~(VV_ROOT | VV_ISTTY | VV_NOSYNC | VV_ETERNALDEV | VV_CACHEDLABEL | VV_TEXT | VV_COPYONWRITE | VV_SYSTEM | VV_PROCDEP | VV_NOKNOTE | VV_DELETED | VV_MD | VV_FORCEINSMQ); if (flags != 0) { snprintf(buf2, sizeof(buf2), "|VV(0x%lx)", flags); strlcat(buf, buf2, sizeof(buf)); } if (vp->v_iflag & VI_MOUNT) strlcat(buf, "|VI_MOUNT", sizeof(buf)); if (vp->v_iflag & VI_AGE) strlcat(buf, "|VI_AGE", sizeof(buf)); if (vp->v_iflag & VI_DOOMED) strlcat(buf, "|VI_DOOMED", sizeof(buf)); if (vp->v_iflag & VI_FREE) strlcat(buf, "|VI_FREE", sizeof(buf)); if (vp->v_iflag & VI_ACTIVE) strlcat(buf, "|VI_ACTIVE", sizeof(buf)); if (vp->v_iflag & VI_DOINGINACT) strlcat(buf, "|VI_DOINGINACT", sizeof(buf)); if (vp->v_iflag & VI_OWEINACT) strlcat(buf, "|VI_OWEINACT", sizeof(buf)); flags = vp->v_iflag & ~(VI_MOUNT | VI_AGE | VI_DOOMED | VI_FREE | VI_ACTIVE | VI_DOINGINACT | VI_OWEINACT); if (flags != 0) { snprintf(buf2, sizeof(buf2), "|VI(0x%lx)", flags); strlcat(buf, buf2, sizeof(buf)); } printf(" flags (%s)\n", buf + 1); if (mtx_owned(VI_MTX(vp))) printf(" VI_LOCKed"); if (vp->v_object != NULL) printf(" v_object %p ref %d pages %d " "cleanbuf %d dirtybuf %d\n", vp->v_object, vp->v_object->ref_count, vp->v_object->resident_page_count, vp->v_bufobj.bo_dirty.bv_cnt, vp->v_bufobj.bo_clean.bv_cnt); printf(" "); lockmgr_printinfo(vp->v_vnlock); if (vp->v_data != NULL) VOP_PRINT(vp); } #ifdef DDB /* * List all of the locked vnodes in the system. * Called when debugging the kernel. */ DB_SHOW_COMMAND(lockedvnods, lockedvnodes) { struct mount *mp; struct vnode *vp; /* * Note: because this is DDB, we can't obey the locking semantics * for these structures, which means we could catch an inconsistent * state and dereference a nasty pointer. Not much to be done * about that. */ db_printf("Locked vnodes\n"); TAILQ_FOREACH(mp, &mountlist, mnt_list) { TAILQ_FOREACH(vp, &mp->mnt_nvnodelist, v_nmntvnodes) { if (vp->v_type != VMARKER && VOP_ISLOCKED(vp)) vprint("", vp); } } } /* * Show details about the given vnode. */ DB_SHOW_COMMAND(vnode, db_show_vnode) { struct vnode *vp; if (!have_addr) return; vp = (struct vnode *)addr; vn_printf(vp, "vnode "); } /* * Show details about the given mount point. */ DB_SHOW_COMMAND(mount, db_show_mount) { struct mount *mp; struct vfsopt *opt; struct statfs *sp; struct vnode *vp; char buf[512]; uint64_t mflags; u_int flags; if (!have_addr) { /* No address given, print short info about all mount points. */ TAILQ_FOREACH(mp, &mountlist, mnt_list) { db_printf("%p %s on %s (%s)\n", mp, mp->mnt_stat.f_mntfromname, mp->mnt_stat.f_mntonname, mp->mnt_stat.f_fstypename); if (db_pager_quit) break; } db_printf("\nMore info: show mount \n"); return; } mp = (struct mount *)addr; db_printf("%p %s on %s (%s)\n", mp, mp->mnt_stat.f_mntfromname, mp->mnt_stat.f_mntonname, mp->mnt_stat.f_fstypename); buf[0] = '\0'; mflags = mp->mnt_flag; #define MNT_FLAG(flag) do { \ if (mflags & (flag)) { \ if (buf[0] != '\0') \ strlcat(buf, ", ", sizeof(buf)); \ strlcat(buf, (#flag) + 4, sizeof(buf)); \ mflags &= ~(flag); \ } \ } while (0) MNT_FLAG(MNT_RDONLY); MNT_FLAG(MNT_SYNCHRONOUS); MNT_FLAG(MNT_NOEXEC); MNT_FLAG(MNT_NOSUID); MNT_FLAG(MNT_NFS4ACLS); MNT_FLAG(MNT_UNION); MNT_FLAG(MNT_ASYNC); MNT_FLAG(MNT_SUIDDIR); MNT_FLAG(MNT_SOFTDEP); MNT_FLAG(MNT_NOSYMFOLLOW); MNT_FLAG(MNT_GJOURNAL); MNT_FLAG(MNT_MULTILABEL); MNT_FLAG(MNT_ACLS); MNT_FLAG(MNT_NOATIME); MNT_FLAG(MNT_NOCLUSTERR); MNT_FLAG(MNT_NOCLUSTERW); MNT_FLAG(MNT_SUJ); MNT_FLAG(MNT_EXRDONLY); MNT_FLAG(MNT_EXPORTED); MNT_FLAG(MNT_DEFEXPORTED); MNT_FLAG(MNT_EXPORTANON); MNT_FLAG(MNT_EXKERB); MNT_FLAG(MNT_EXPUBLIC); MNT_FLAG(MNT_LOCAL); MNT_FLAG(MNT_QUOTA); MNT_FLAG(MNT_ROOTFS); MNT_FLAG(MNT_USER); MNT_FLAG(MNT_IGNORE); MNT_FLAG(MNT_UPDATE); MNT_FLAG(MNT_DELEXPORT); MNT_FLAG(MNT_RELOAD); MNT_FLAG(MNT_FORCE); MNT_FLAG(MNT_SNAPSHOT); MNT_FLAG(MNT_BYFSID); #undef MNT_FLAG if (mflags != 0) { if (buf[0] != '\0') strlcat(buf, ", ", sizeof(buf)); snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), "0x%016jx", mflags); } db_printf(" mnt_flag = %s\n", buf); buf[0] = '\0'; flags = mp->mnt_kern_flag; #define MNT_KERN_FLAG(flag) do { \ if (flags & (flag)) { \ if (buf[0] != '\0') \ strlcat(buf, ", ", sizeof(buf)); \ strlcat(buf, (#flag) + 5, sizeof(buf)); \ flags &= ~(flag); \ } \ } while (0) MNT_KERN_FLAG(MNTK_UNMOUNTF); MNT_KERN_FLAG(MNTK_ASYNC); MNT_KERN_FLAG(MNTK_SOFTDEP); MNT_KERN_FLAG(MNTK_NOINSMNTQ); MNT_KERN_FLAG(MNTK_DRAINING); MNT_KERN_FLAG(MNTK_REFEXPIRE); MNT_KERN_FLAG(MNTK_EXTENDED_SHARED); MNT_KERN_FLAG(MNTK_SHARED_WRITES); MNT_KERN_FLAG(MNTK_NO_IOPF); MNT_KERN_FLAG(MNTK_VGONE_UPPER); MNT_KERN_FLAG(MNTK_VGONE_WAITER); MNT_KERN_FLAG(MNTK_LOOKUP_EXCL_DOTDOT); MNT_KERN_FLAG(MNTK_MARKER); MNT_KERN_FLAG(MNTK_NOASYNC); MNT_KERN_FLAG(MNTK_UNMOUNT); MNT_KERN_FLAG(MNTK_MWAIT); MNT_KERN_FLAG(MNTK_SUSPEND); MNT_KERN_FLAG(MNTK_SUSPEND2); MNT_KERN_FLAG(MNTK_SUSPENDED); MNT_KERN_FLAG(MNTK_LOOKUP_SHARED); MNT_KERN_FLAG(MNTK_NOKNOTE); #undef MNT_KERN_FLAG if (flags != 0) { if (buf[0] != '\0') strlcat(buf, ", ", sizeof(buf)); snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), "0x%08x", flags); } db_printf(" mnt_kern_flag = %s\n", buf); db_printf(" mnt_opt = "); opt = TAILQ_FIRST(mp->mnt_opt); if (opt != NULL) { db_printf("%s", opt->name); opt = TAILQ_NEXT(opt, link); while (opt != NULL) { db_printf(", %s", opt->name); opt = TAILQ_NEXT(opt, link); } } db_printf("\n"); sp = &mp->mnt_stat; db_printf(" mnt_stat = { version=%u type=%u flags=0x%016jx " "bsize=%ju iosize=%ju blocks=%ju bfree=%ju bavail=%jd files=%ju " "ffree=%jd syncwrites=%ju asyncwrites=%ju syncreads=%ju " "asyncreads=%ju namemax=%u owner=%u fsid=[%d, %d] }\n", (u_int)sp->f_version, (u_int)sp->f_type, (uintmax_t)sp->f_flags, (uintmax_t)sp->f_bsize, (uintmax_t)sp->f_iosize, (uintmax_t)sp->f_blocks, (uintmax_t)sp->f_bfree, (intmax_t)sp->f_bavail, (uintmax_t)sp->f_files, (intmax_t)sp->f_ffree, (uintmax_t)sp->f_syncwrites, (uintmax_t)sp->f_asyncwrites, (uintmax_t)sp->f_syncreads, (uintmax_t)sp->f_asyncreads, (u_int)sp->f_namemax, (u_int)sp->f_owner, (int)sp->f_fsid.val[0], (int)sp->f_fsid.val[1]); db_printf(" mnt_cred = { uid=%u ruid=%u", (u_int)mp->mnt_cred->cr_uid, (u_int)mp->mnt_cred->cr_ruid); if (jailed(mp->mnt_cred)) db_printf(", jail=%d", mp->mnt_cred->cr_prison->pr_id); db_printf(" }\n"); db_printf(" mnt_ref = %d\n", mp->mnt_ref); db_printf(" mnt_gen = %d\n", mp->mnt_gen); db_printf(" mnt_nvnodelistsize = %d\n", mp->mnt_nvnodelistsize); db_printf(" mnt_activevnodelistsize = %d\n", mp->mnt_activevnodelistsize); db_printf(" mnt_writeopcount = %d\n", mp->mnt_writeopcount); db_printf(" mnt_maxsymlinklen = %d\n", mp->mnt_maxsymlinklen); db_printf(" mnt_iosize_max = %d\n", mp->mnt_iosize_max); db_printf(" mnt_hashseed = %u\n", mp->mnt_hashseed); db_printf(" mnt_secondary_writes = %d\n", mp->mnt_secondary_writes); db_printf(" mnt_secondary_accwrites = %d\n", mp->mnt_secondary_accwrites); db_printf(" mnt_gjprovider = %s\n", mp->mnt_gjprovider != NULL ? mp->mnt_gjprovider : "NULL"); db_printf("\n\nList of active vnodes\n"); TAILQ_FOREACH(vp, &mp->mnt_activevnodelist, v_actfreelist) { if (vp->v_type != VMARKER) { vn_printf(vp, "vnode "); if (db_pager_quit) break; } } db_printf("\n\nList of inactive vnodes\n"); TAILQ_FOREACH(vp, &mp->mnt_nvnodelist, v_nmntvnodes) { if (vp->v_type != VMARKER && (vp->v_iflag & VI_ACTIVE) == 0) { vn_printf(vp, "vnode "); if (db_pager_quit) break; } } } #endif /* DDB */ /* * Fill in a struct xvfsconf based on a struct vfsconf. */ static int vfsconf2x(struct sysctl_req *req, struct vfsconf *vfsp) { struct xvfsconf xvfsp; bzero(&xvfsp, sizeof(xvfsp)); strcpy(xvfsp.vfc_name, vfsp->vfc_name); xvfsp.vfc_typenum = vfsp->vfc_typenum; xvfsp.vfc_refcount = vfsp->vfc_refcount; xvfsp.vfc_flags = vfsp->vfc_flags; /* * These are unused in userland, we keep them * to not break binary compatibility. */ xvfsp.vfc_vfsops = NULL; xvfsp.vfc_next = NULL; return (SYSCTL_OUT(req, &xvfsp, sizeof(xvfsp))); } #ifdef COMPAT_FREEBSD32 struct xvfsconf32 { uint32_t vfc_vfsops; char vfc_name[MFSNAMELEN]; int32_t vfc_typenum; int32_t vfc_refcount; int32_t vfc_flags; uint32_t vfc_next; }; static int vfsconf2x32(struct sysctl_req *req, struct vfsconf *vfsp) { struct xvfsconf32 xvfsp; strcpy(xvfsp.vfc_name, vfsp->vfc_name); xvfsp.vfc_typenum = vfsp->vfc_typenum; xvfsp.vfc_refcount = vfsp->vfc_refcount; xvfsp.vfc_flags = vfsp->vfc_flags; xvfsp.vfc_vfsops = 0; xvfsp.vfc_next = 0; return (SYSCTL_OUT(req, &xvfsp, sizeof(xvfsp))); } #endif /* * Top level filesystem related information gathering. */ static int sysctl_vfs_conflist(SYSCTL_HANDLER_ARGS) { struct vfsconf *vfsp; int error; error = 0; vfsconf_slock(); TAILQ_FOREACH(vfsp, &vfsconf, vfc_list) { #ifdef COMPAT_FREEBSD32 if (req->flags & SCTL_MASK32) error = vfsconf2x32(req, vfsp); else #endif error = vfsconf2x(req, vfsp); if (error) break; } vfsconf_sunlock(); return (error); } SYSCTL_PROC(_vfs, OID_AUTO, conflist, CTLTYPE_OPAQUE | CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, 0, sysctl_vfs_conflist, "S,xvfsconf", "List of all configured filesystems"); #ifndef BURN_BRIDGES static int sysctl_ovfs_conf(SYSCTL_HANDLER_ARGS); static int vfs_sysctl(SYSCTL_HANDLER_ARGS) { int *name = (int *)arg1 - 1; /* XXX */ u_int namelen = arg2 + 1; /* XXX */ struct vfsconf *vfsp; log(LOG_WARNING, "userland calling deprecated sysctl, " "please rebuild world\n"); #if 1 || defined(COMPAT_PRELITE2) /* Resolve ambiguity between VFS_VFSCONF and VFS_GENERIC. */ if (namelen == 1) return (sysctl_ovfs_conf(oidp, arg1, arg2, req)); #endif switch (name[1]) { case VFS_MAXTYPENUM: if (namelen != 2) return (ENOTDIR); return (SYSCTL_OUT(req, &maxvfsconf, sizeof(int))); case VFS_CONF: if (namelen != 3) return (ENOTDIR); /* overloaded */ vfsconf_slock(); TAILQ_FOREACH(vfsp, &vfsconf, vfc_list) { if (vfsp->vfc_typenum == name[2]) break; } vfsconf_sunlock(); if (vfsp == NULL) return (EOPNOTSUPP); #ifdef COMPAT_FREEBSD32 if (req->flags & SCTL_MASK32) return (vfsconf2x32(req, vfsp)); else #endif return (vfsconf2x(req, vfsp)); } return (EOPNOTSUPP); } static SYSCTL_NODE(_vfs, VFS_GENERIC, generic, CTLFLAG_RD | CTLFLAG_SKIP | CTLFLAG_MPSAFE, vfs_sysctl, "Generic filesystem"); #if 1 || defined(COMPAT_PRELITE2) static int sysctl_ovfs_conf(SYSCTL_HANDLER_ARGS) { int error; struct vfsconf *vfsp; struct ovfsconf ovfs; vfsconf_slock(); TAILQ_FOREACH(vfsp, &vfsconf, vfc_list) { bzero(&ovfs, sizeof(ovfs)); ovfs.vfc_vfsops = vfsp->vfc_vfsops; /* XXX used as flag */ strcpy(ovfs.vfc_name, vfsp->vfc_name); ovfs.vfc_index = vfsp->vfc_typenum; ovfs.vfc_refcount = vfsp->vfc_refcount; ovfs.vfc_flags = vfsp->vfc_flags; error = SYSCTL_OUT(req, &ovfs, sizeof ovfs); if (error != 0) { vfsconf_sunlock(); return (error); } } vfsconf_sunlock(); return (0); } #endif /* 1 || COMPAT_PRELITE2 */ #endif /* !BURN_BRIDGES */ #define KINFO_VNODESLOP 10 #ifdef notyet /* * Dump vnode list (via sysctl). */ /* ARGSUSED */ static int sysctl_vnode(SYSCTL_HANDLER_ARGS) { struct xvnode *xvn; struct mount *mp; struct vnode *vp; int error, len, n; /* * Stale numvnodes access is not fatal here. */ req->lock = 0; len = (numvnodes + KINFO_VNODESLOP) * sizeof *xvn; if (!req->oldptr) /* Make an estimate */ return (SYSCTL_OUT(req, 0, len)); error = sysctl_wire_old_buffer(req, 0); if (error != 0) return (error); xvn = malloc(len, M_TEMP, M_ZERO | M_WAITOK); n = 0; mtx_lock(&mountlist_mtx); TAILQ_FOREACH(mp, &mountlist, mnt_list) { if (vfs_busy(mp, MBF_NOWAIT | MBF_MNTLSTLOCK)) continue; MNT_ILOCK(mp); TAILQ_FOREACH(vp, &mp->mnt_nvnodelist, v_nmntvnodes) { if (n == len) break; vref(vp); xvn[n].xv_size = sizeof *xvn; xvn[n].xv_vnode = vp; xvn[n].xv_id = 0; /* XXX compat */ #define XV_COPY(field) xvn[n].xv_##field = vp->v_##field XV_COPY(usecount); XV_COPY(writecount); XV_COPY(holdcnt); XV_COPY(mount); XV_COPY(numoutput); XV_COPY(type); #undef XV_COPY xvn[n].xv_flag = vp->v_vflag; switch (vp->v_type) { case VREG: case VDIR: case VLNK: break; case VBLK: case VCHR: if (vp->v_rdev == NULL) { vrele(vp); continue; } xvn[n].xv_dev = dev2udev(vp->v_rdev); break; case VSOCK: xvn[n].xv_socket = vp->v_socket; break; case VFIFO: xvn[n].xv_fifo = vp->v_fifoinfo; break; case VNON: case VBAD: default: /* shouldn't happen? */ vrele(vp); continue; } vrele(vp); ++n; } MNT_IUNLOCK(mp); mtx_lock(&mountlist_mtx); vfs_unbusy(mp); if (n == len) break; } mtx_unlock(&mountlist_mtx); error = SYSCTL_OUT(req, xvn, n * sizeof *xvn); free(xvn, M_TEMP); return (error); } SYSCTL_PROC(_kern, KERN_VNODE, vnode, CTLTYPE_OPAQUE | CTLFLAG_RD | CTLFLAG_MPSAFE, 0, 0, sysctl_vnode, "S,xvnode", ""); #endif /* * Unmount all filesystems. The list is traversed in reverse order * of mounting to avoid dependencies. */ void vfs_unmountall(void) { struct mount *mp; struct thread *td; int error; CTR1(KTR_VFS, "%s: unmounting all filesystems", __func__); td = curthread; /* * Since this only runs when rebooting, it is not interlocked. */ while(!TAILQ_EMPTY(&mountlist)) { mp = TAILQ_LAST(&mountlist, mntlist); error = dounmount(mp, MNT_FORCE, td); if (error) { TAILQ_REMOVE(&mountlist, mp, mnt_list); /* * XXX: Due to the way in which we mount the root * file system off of devfs, devfs will generate a * "busy" warning when we try to unmount it before * the root. Don't print a warning as a result in * order to avoid false positive errors that may * cause needless upset. */ if (strcmp(mp->mnt_vfc->vfc_name, "devfs") != 0) { printf("unmount of %s failed (", mp->mnt_stat.f_mntonname); if (error == EBUSY) printf("BUSY)\n"); else printf("%d)\n", error); } } else { /* The unmount has removed mp from the mountlist */ } } } /* * perform msync on all vnodes under a mount point * the mount point must be locked. */ void vfs_msync(struct mount *mp, int flags) { struct vnode *vp, *mvp; struct vm_object *obj; CTR2(KTR_VFS, "%s: mp %p", __func__, mp); MNT_VNODE_FOREACH_ACTIVE(vp, mp, mvp) { obj = vp->v_object; if (obj != NULL && (obj->flags & OBJ_MIGHTBEDIRTY) != 0 && (flags == MNT_WAIT || VOP_ISLOCKED(vp) == 0)) { if (!vget(vp, LK_EXCLUSIVE | LK_RETRY | LK_INTERLOCK, curthread)) { if (vp->v_vflag & VV_NOSYNC) { /* unlinked */ vput(vp); continue; } obj = vp->v_object; if (obj != NULL) { VM_OBJECT_WLOCK(obj); vm_object_page_clean(obj, 0, 0, flags == MNT_WAIT ? OBJPC_SYNC : OBJPC_NOSYNC); VM_OBJECT_WUNLOCK(obj); } vput(vp); } } else VI_UNLOCK(vp); } } static void destroy_vpollinfo_free(struct vpollinfo *vi) { knlist_destroy(&vi->vpi_selinfo.si_note); mtx_destroy(&vi->vpi_lock); uma_zfree(vnodepoll_zone, vi); } static void destroy_vpollinfo(struct vpollinfo *vi) { knlist_clear(&vi->vpi_selinfo.si_note, 1); seldrain(&vi->vpi_selinfo); destroy_vpollinfo_free(vi); } /* * Initalize per-vnode helper structure to hold poll-related state. */ void v_addpollinfo(struct vnode *vp) { struct vpollinfo *vi; if (vp->v_pollinfo != NULL) return; vi = uma_zalloc(vnodepoll_zone, M_WAITOK); mtx_init(&vi->vpi_lock, "vnode pollinfo", NULL, MTX_DEF); knlist_init(&vi->vpi_selinfo.si_note, vp, vfs_knllock, vfs_knlunlock, vfs_knl_assert_locked, vfs_knl_assert_unlocked); VI_LOCK(vp); if (vp->v_pollinfo != NULL) { VI_UNLOCK(vp); destroy_vpollinfo_free(vi); return; } vp->v_pollinfo = vi; VI_UNLOCK(vp); } /* * Record a process's interest in events which might happen to * a vnode. Because poll uses the historic select-style interface * internally, this routine serves as both the ``check for any * pending events'' and the ``record my interest in future events'' * functions. (These are done together, while the lock is held, * to avoid race conditions.) */ int vn_pollrecord(struct vnode *vp, struct thread *td, int events) { v_addpollinfo(vp); mtx_lock(&vp->v_pollinfo->vpi_lock); if (vp->v_pollinfo->vpi_revents & events) { /* * This leaves events we are not interested * in available for the other process which * which presumably had requested them * (otherwise they would never have been * recorded). */ events &= vp->v_pollinfo->vpi_revents; vp->v_pollinfo->vpi_revents &= ~events; mtx_unlock(&vp->v_pollinfo->vpi_lock); return (events); } vp->v_pollinfo->vpi_events |= events; selrecord(td, &vp->v_pollinfo->vpi_selinfo); mtx_unlock(&vp->v_pollinfo->vpi_lock); return (0); } /* * Routine to create and manage a filesystem syncer vnode. */ #define sync_close ((int (*)(struct vop_close_args *))nullop) static int sync_fsync(struct vop_fsync_args *); static int sync_inactive(struct vop_inactive_args *); static int sync_reclaim(struct vop_reclaim_args *); static struct vop_vector sync_vnodeops = { .vop_bypass = VOP_EOPNOTSUPP, .vop_close = sync_close, /* close */ .vop_fsync = sync_fsync, /* fsync */ .vop_inactive = sync_inactive, /* inactive */ .vop_reclaim = sync_reclaim, /* reclaim */ .vop_lock1 = vop_stdlock, /* lock */ .vop_unlock = vop_stdunlock, /* unlock */ .vop_islocked = vop_stdislocked, /* islocked */ }; /* * Create a new filesystem syncer vnode for the specified mount point. */ void vfs_allocate_syncvnode(struct mount *mp) { struct vnode *vp; struct bufobj *bo; static long start, incr, next; int error; /* Allocate a new vnode */ error = getnewvnode("syncer", mp, &sync_vnodeops, &vp); if (error != 0) panic("vfs_allocate_syncvnode: getnewvnode() failed"); vp->v_type = VNON; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); vp->v_vflag |= VV_FORCEINSMQ; error = insmntque(vp, mp); if (error != 0) panic("vfs_allocate_syncvnode: insmntque() failed"); vp->v_vflag &= ~VV_FORCEINSMQ; VOP_UNLOCK(vp, 0); /* * Place the vnode onto the syncer worklist. We attempt to * scatter them about on the list so that they will go off * at evenly distributed times even if all the filesystems * are mounted at once. */ next += incr; if (next == 0 || next > syncer_maxdelay) { start /= 2; incr /= 2; if (start == 0) { start = syncer_maxdelay / 2; incr = syncer_maxdelay; } next = start; } bo = &vp->v_bufobj; BO_LOCK(bo); vn_syncer_add_to_worklist(bo, syncdelay > 0 ? next % syncdelay : 0); /* XXX - vn_syncer_add_to_worklist() also grabs and drops sync_mtx. */ mtx_lock(&sync_mtx); sync_vnode_count++; if (mp->mnt_syncer == NULL) { mp->mnt_syncer = vp; vp = NULL; } mtx_unlock(&sync_mtx); BO_UNLOCK(bo); if (vp != NULL) { vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); vgone(vp); vput(vp); } } void vfs_deallocate_syncvnode(struct mount *mp) { struct vnode *vp; mtx_lock(&sync_mtx); vp = mp->mnt_syncer; if (vp != NULL) mp->mnt_syncer = NULL; mtx_unlock(&sync_mtx); if (vp != NULL) vrele(vp); } /* * Do a lazy sync of the filesystem. */ static int sync_fsync(struct vop_fsync_args *ap) { struct vnode *syncvp = ap->a_vp; struct mount *mp = syncvp->v_mount; int error, save; struct bufobj *bo; /* * We only need to do something if this is a lazy evaluation. */ if (ap->a_waitfor != MNT_LAZY) return (0); /* * Move ourselves to the back of the sync list. */ bo = &syncvp->v_bufobj; BO_LOCK(bo); vn_syncer_add_to_worklist(bo, syncdelay); BO_UNLOCK(bo); /* * Walk the list of vnodes pushing all that are dirty and * not already on the sync list. */ if (vfs_busy(mp, MBF_NOWAIT) != 0) return (0); if (vn_start_write(NULL, &mp, V_NOWAIT) != 0) { vfs_unbusy(mp); return (0); } save = curthread_pflags_set(TDP_SYNCIO); vfs_msync(mp, MNT_NOWAIT); error = VFS_SYNC(mp, MNT_LAZY); curthread_pflags_restore(save); vn_finished_write(mp); vfs_unbusy(mp); return (error); } /* * The syncer vnode is no referenced. */ static int sync_inactive(struct vop_inactive_args *ap) { vgone(ap->a_vp); return (0); } /* * The syncer vnode is no longer needed and is being decommissioned. * * Modifications to the worklist must be protected by sync_mtx. */ static int sync_reclaim(struct vop_reclaim_args *ap) { struct vnode *vp = ap->a_vp; struct bufobj *bo; bo = &vp->v_bufobj; BO_LOCK(bo); mtx_lock(&sync_mtx); if (vp->v_mount->mnt_syncer == vp) vp->v_mount->mnt_syncer = NULL; if (bo->bo_flag & BO_ONWORKLST) { LIST_REMOVE(bo, bo_synclist); syncer_worklist_len--; sync_vnode_count--; bo->bo_flag &= ~BO_ONWORKLST; } mtx_unlock(&sync_mtx); BO_UNLOCK(bo); return (0); } /* * Check if vnode represents a disk device */ int vn_isdisk(struct vnode *vp, int *errp) { int error; if (vp->v_type != VCHR) { error = ENOTBLK; goto out; } error = 0; dev_lock(); if (vp->v_rdev == NULL) error = ENXIO; else if (vp->v_rdev->si_devsw == NULL) error = ENXIO; else if (!(vp->v_rdev->si_devsw->d_flags & D_DISK)) error = ENOTBLK; dev_unlock(); out: if (errp != NULL) *errp = error; return (error == 0); } /* * Common filesystem object access control check routine. Accepts a * vnode's type, "mode", uid and gid, requested access mode, credentials, * and optional call-by-reference privused argument allowing vaccess() * to indicate to the caller whether privilege was used to satisfy the * request (obsoleted). Returns 0 on success, or an errno on failure. */ int vaccess(enum vtype type, mode_t file_mode, uid_t file_uid, gid_t file_gid, accmode_t accmode, struct ucred *cred, int *privused) { accmode_t dac_granted; accmode_t priv_granted; KASSERT((accmode & ~(VEXEC | VWRITE | VREAD | VADMIN | VAPPEND)) == 0, ("invalid bit in accmode")); KASSERT((accmode & VAPPEND) == 0 || (accmode & VWRITE), ("VAPPEND without VWRITE")); /* * Look for a normal, non-privileged way to access the file/directory * as requested. If it exists, go with that. */ if (privused != NULL) *privused = 0; dac_granted = 0; /* Check the owner. */ if (cred->cr_uid == file_uid) { dac_granted |= VADMIN; if (file_mode & S_IXUSR) dac_granted |= VEXEC; if (file_mode & S_IRUSR) dac_granted |= VREAD; if (file_mode & S_IWUSR) dac_granted |= (VWRITE | VAPPEND); if ((accmode & dac_granted) == accmode) return (0); goto privcheck; } /* Otherwise, check the groups (first match) */ if (groupmember(file_gid, cred)) { if (file_mode & S_IXGRP) dac_granted |= VEXEC; if (file_mode & S_IRGRP) dac_granted |= VREAD; if (file_mode & S_IWGRP) dac_granted |= (VWRITE | VAPPEND); if ((accmode & dac_granted) == accmode) return (0); goto privcheck; } /* Otherwise, check everyone else. */ if (file_mode & S_IXOTH) dac_granted |= VEXEC; if (file_mode & S_IROTH) dac_granted |= VREAD; if (file_mode & S_IWOTH) dac_granted |= (VWRITE | VAPPEND); if ((accmode & dac_granted) == accmode) return (0); privcheck: /* * Build a privilege mask to determine if the set of privileges * satisfies the requirements when combined with the granted mask * from above. For each privilege, if the privilege is required, * bitwise or the request type onto the priv_granted mask. */ priv_granted = 0; if (type == VDIR) { /* * For directories, use PRIV_VFS_LOOKUP to satisfy VEXEC * requests, instead of PRIV_VFS_EXEC. */ if ((accmode & VEXEC) && ((dac_granted & VEXEC) == 0) && !priv_check_cred(cred, PRIV_VFS_LOOKUP, 0)) priv_granted |= VEXEC; } else { /* * Ensure that at least one execute bit is on. Otherwise, * a privileged user will always succeed, and we don't want * this to happen unless the file really is executable. */ if ((accmode & VEXEC) && ((dac_granted & VEXEC) == 0) && (file_mode & (S_IXUSR | S_IXGRP | S_IXOTH)) != 0 && !priv_check_cred(cred, PRIV_VFS_EXEC, 0)) priv_granted |= VEXEC; } if ((accmode & VREAD) && ((dac_granted & VREAD) == 0) && !priv_check_cred(cred, PRIV_VFS_READ, 0)) priv_granted |= VREAD; if ((accmode & VWRITE) && ((dac_granted & VWRITE) == 0) && !priv_check_cred(cred, PRIV_VFS_WRITE, 0)) priv_granted |= (VWRITE | VAPPEND); if ((accmode & VADMIN) && ((dac_granted & VADMIN) == 0) && !priv_check_cred(cred, PRIV_VFS_ADMIN, 0)) priv_granted |= VADMIN; if ((accmode & (priv_granted | dac_granted)) == accmode) { /* XXX audit: privilege used */ if (privused != NULL) *privused = 1; return (0); } return ((accmode & VADMIN) ? EPERM : EACCES); } /* * Credential check based on process requesting service, and per-attribute * permissions. */ int extattr_check_cred(struct vnode *vp, int attrnamespace, struct ucred *cred, struct thread *td, accmode_t accmode) { /* * Kernel-invoked always succeeds. */ if (cred == NOCRED) return (0); /* * Do not allow privileged processes in jail to directly manipulate * system attributes. */ switch (attrnamespace) { case EXTATTR_NAMESPACE_SYSTEM: /* Potentially should be: return (EPERM); */ return (priv_check_cred(cred, PRIV_VFS_EXTATTR_SYSTEM, 0)); case EXTATTR_NAMESPACE_USER: return (VOP_ACCESS(vp, accmode, cred, td)); default: return (EPERM); } } #ifdef DEBUG_VFS_LOCKS /* * This only exists to supress warnings from unlocked specfs accesses. It is * no longer ok to have an unlocked VFS. */ #define IGNORE_LOCK(vp) (panicstr != NULL || (vp) == NULL || \ (vp)->v_type == VCHR || (vp)->v_type == VBAD) int vfs_badlock_ddb = 1; /* Drop into debugger on violation. */ SYSCTL_INT(_debug, OID_AUTO, vfs_badlock_ddb, CTLFLAG_RW, &vfs_badlock_ddb, 0, "Drop into debugger on lock violation"); int vfs_badlock_mutex = 1; /* Check for interlock across VOPs. */ SYSCTL_INT(_debug, OID_AUTO, vfs_badlock_mutex, CTLFLAG_RW, &vfs_badlock_mutex, 0, "Check for interlock across VOPs"); int vfs_badlock_print = 1; /* Print lock violations. */ SYSCTL_INT(_debug, OID_AUTO, vfs_badlock_print, CTLFLAG_RW, &vfs_badlock_print, 0, "Print lock violations"); #ifdef KDB int vfs_badlock_backtrace = 1; /* Print backtrace at lock violations. */ SYSCTL_INT(_debug, OID_AUTO, vfs_badlock_backtrace, CTLFLAG_RW, &vfs_badlock_backtrace, 0, "Print backtrace at lock violations"); #endif static void vfs_badlock(const char *msg, const char *str, struct vnode *vp) { #ifdef KDB if (vfs_badlock_backtrace) kdb_backtrace(); #endif if (vfs_badlock_print) printf("%s: %p %s\n", str, (void *)vp, msg); if (vfs_badlock_ddb) kdb_enter(KDB_WHY_VFSLOCK, "lock violation"); } void assert_vi_locked(struct vnode *vp, const char *str) { if (vfs_badlock_mutex && !mtx_owned(VI_MTX(vp))) vfs_badlock("interlock is not locked but should be", str, vp); } void assert_vi_unlocked(struct vnode *vp, const char *str) { if (vfs_badlock_mutex && mtx_owned(VI_MTX(vp))) vfs_badlock("interlock is locked but should not be", str, vp); } void assert_vop_locked(struct vnode *vp, const char *str) { int locked; if (!IGNORE_LOCK(vp)) { locked = VOP_ISLOCKED(vp); if (locked == 0 || locked == LK_EXCLOTHER) vfs_badlock("is not locked but should be", str, vp); } } void assert_vop_unlocked(struct vnode *vp, const char *str) { if (!IGNORE_LOCK(vp) && VOP_ISLOCKED(vp) == LK_EXCLUSIVE) vfs_badlock("is locked but should not be", str, vp); } void assert_vop_elocked(struct vnode *vp, const char *str) { if (!IGNORE_LOCK(vp) && VOP_ISLOCKED(vp) != LK_EXCLUSIVE) vfs_badlock("is not exclusive locked but should be", str, vp); } #if 0 void assert_vop_elocked_other(struct vnode *vp, const char *str) { if (!IGNORE_LOCK(vp) && VOP_ISLOCKED(vp) != LK_EXCLOTHER) vfs_badlock("is not exclusive locked by another thread", str, vp); } void assert_vop_slocked(struct vnode *vp, const char *str) { if (!IGNORE_LOCK(vp) && VOP_ISLOCKED(vp) != LK_SHARED) vfs_badlock("is not locked shared but should be", str, vp); } #endif /* 0 */ #endif /* DEBUG_VFS_LOCKS */ void vop_rename_fail(struct vop_rename_args *ap) { if (ap->a_tvp != NULL) vput(ap->a_tvp); if (ap->a_tdvp == ap->a_tvp) vrele(ap->a_tdvp); else vput(ap->a_tdvp); vrele(ap->a_fdvp); vrele(ap->a_fvp); } void vop_rename_pre(void *ap) { struct vop_rename_args *a = ap; #ifdef DEBUG_VFS_LOCKS if (a->a_tvp) ASSERT_VI_UNLOCKED(a->a_tvp, "VOP_RENAME"); ASSERT_VI_UNLOCKED(a->a_tdvp, "VOP_RENAME"); ASSERT_VI_UNLOCKED(a->a_fvp, "VOP_RENAME"); ASSERT_VI_UNLOCKED(a->a_fdvp, "VOP_RENAME"); /* Check the source (from). */ if (a->a_tdvp->v_vnlock != a->a_fdvp->v_vnlock && (a->a_tvp == NULL || a->a_tvp->v_vnlock != a->a_fdvp->v_vnlock)) ASSERT_VOP_UNLOCKED(a->a_fdvp, "vop_rename: fdvp locked"); if (a->a_tvp == NULL || a->a_tvp->v_vnlock != a->a_fvp->v_vnlock) ASSERT_VOP_UNLOCKED(a->a_fvp, "vop_rename: fvp locked"); /* Check the target. */ if (a->a_tvp) ASSERT_VOP_LOCKED(a->a_tvp, "vop_rename: tvp not locked"); ASSERT_VOP_LOCKED(a->a_tdvp, "vop_rename: tdvp not locked"); #endif if (a->a_tdvp != a->a_fdvp) vhold(a->a_fdvp); if (a->a_tvp != a->a_fvp) vhold(a->a_fvp); vhold(a->a_tdvp); if (a->a_tvp) vhold(a->a_tvp); } void vop_strategy_pre(void *ap) { #ifdef DEBUG_VFS_LOCKS struct vop_strategy_args *a; struct buf *bp; a = ap; bp = a->a_bp; /* * Cluster ops lock their component buffers but not the IO container. */ if ((bp->b_flags & B_CLUSTER) != 0) return; if (panicstr == NULL && !BUF_ISLOCKED(bp)) { if (vfs_badlock_print) printf( "VOP_STRATEGY: bp is not locked but should be\n"); if (vfs_badlock_ddb) kdb_enter(KDB_WHY_VFSLOCK, "lock violation"); } #endif } void vop_lock_pre(void *ap) { #ifdef DEBUG_VFS_LOCKS struct vop_lock1_args *a = ap; if ((a->a_flags & LK_INTERLOCK) == 0) ASSERT_VI_UNLOCKED(a->a_vp, "VOP_LOCK"); else ASSERT_VI_LOCKED(a->a_vp, "VOP_LOCK"); #endif } void vop_lock_post(void *ap, int rc) { #ifdef DEBUG_VFS_LOCKS struct vop_lock1_args *a = ap; ASSERT_VI_UNLOCKED(a->a_vp, "VOP_LOCK"); if (rc == 0 && (a->a_flags & LK_EXCLOTHER) == 0) ASSERT_VOP_LOCKED(a->a_vp, "VOP_LOCK"); #endif } void vop_unlock_pre(void *ap) { #ifdef DEBUG_VFS_LOCKS struct vop_unlock_args *a = ap; if (a->a_flags & LK_INTERLOCK) ASSERT_VI_LOCKED(a->a_vp, "VOP_UNLOCK"); ASSERT_VOP_LOCKED(a->a_vp, "VOP_UNLOCK"); #endif } void vop_unlock_post(void *ap, int rc) { #ifdef DEBUG_VFS_LOCKS struct vop_unlock_args *a = ap; if (a->a_flags & LK_INTERLOCK) ASSERT_VI_UNLOCKED(a->a_vp, "VOP_UNLOCK"); #endif } void vop_create_post(void *ap, int rc) { struct vop_create_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_dvp, NOTE_WRITE); } void vop_deleteextattr_post(void *ap, int rc) { struct vop_deleteextattr_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_vp, NOTE_ATTRIB); } void vop_link_post(void *ap, int rc) { struct vop_link_args *a = ap; if (!rc) { VFS_KNOTE_LOCKED(a->a_vp, NOTE_LINK); VFS_KNOTE_LOCKED(a->a_tdvp, NOTE_WRITE); } } void vop_mkdir_post(void *ap, int rc) { struct vop_mkdir_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_dvp, NOTE_WRITE | NOTE_LINK); } void vop_mknod_post(void *ap, int rc) { struct vop_mknod_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_dvp, NOTE_WRITE); } void vop_remove_post(void *ap, int rc) { struct vop_remove_args *a = ap; if (!rc) { VFS_KNOTE_LOCKED(a->a_dvp, NOTE_WRITE); VFS_KNOTE_LOCKED(a->a_vp, NOTE_DELETE); } } void vop_rename_post(void *ap, int rc) { struct vop_rename_args *a = ap; if (!rc) { VFS_KNOTE_UNLOCKED(a->a_fdvp, NOTE_WRITE); VFS_KNOTE_UNLOCKED(a->a_tdvp, NOTE_WRITE); VFS_KNOTE_UNLOCKED(a->a_fvp, NOTE_RENAME); if (a->a_tvp) VFS_KNOTE_UNLOCKED(a->a_tvp, NOTE_DELETE); } if (a->a_tdvp != a->a_fdvp) vdrop(a->a_fdvp); if (a->a_tvp != a->a_fvp) vdrop(a->a_fvp); vdrop(a->a_tdvp); if (a->a_tvp) vdrop(a->a_tvp); } void vop_rmdir_post(void *ap, int rc) { struct vop_rmdir_args *a = ap; if (!rc) { VFS_KNOTE_LOCKED(a->a_dvp, NOTE_WRITE | NOTE_LINK); VFS_KNOTE_LOCKED(a->a_vp, NOTE_DELETE); } } void vop_setattr_post(void *ap, int rc) { struct vop_setattr_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_vp, NOTE_ATTRIB); } void vop_setextattr_post(void *ap, int rc) { struct vop_setextattr_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_vp, NOTE_ATTRIB); } void vop_symlink_post(void *ap, int rc) { struct vop_symlink_args *a = ap; if (!rc) VFS_KNOTE_LOCKED(a->a_dvp, NOTE_WRITE); } static struct knlist fs_knlist; static void vfs_event_init(void *arg) { knlist_init_mtx(&fs_knlist, NULL); } /* XXX - correct order? */ SYSINIT(vfs_knlist, SI_SUB_VFS, SI_ORDER_ANY, vfs_event_init, NULL); void vfs_event_signal(fsid_t *fsid, uint32_t event, intptr_t data __unused) { KNOTE_UNLOCKED(&fs_knlist, event); } static int filt_fsattach(struct knote *kn); static void filt_fsdetach(struct knote *kn); static int filt_fsevent(struct knote *kn, long hint); struct filterops fs_filtops = { .f_isfd = 0, .f_attach = filt_fsattach, .f_detach = filt_fsdetach, .f_event = filt_fsevent }; static int filt_fsattach(struct knote *kn) { kn->kn_flags |= EV_CLEAR; knlist_add(&fs_knlist, kn, 0); return (0); } static void filt_fsdetach(struct knote *kn) { knlist_remove(&fs_knlist, kn, 0); } static int filt_fsevent(struct knote *kn, long hint) { kn->kn_fflags |= hint; return (kn->kn_fflags != 0); } static int sysctl_vfs_ctl(SYSCTL_HANDLER_ARGS) { struct vfsidctl vc; int error; struct mount *mp; error = SYSCTL_IN(req, &vc, sizeof(vc)); if (error) return (error); if (vc.vc_vers != VFS_CTL_VERS1) return (EINVAL); mp = vfs_getvfs(&vc.vc_fsid); if (mp == NULL) return (ENOENT); /* ensure that a specific sysctl goes to the right filesystem. */ if (strcmp(vc.vc_fstypename, "*") != 0 && strcmp(vc.vc_fstypename, mp->mnt_vfc->vfc_name) != 0) { vfs_rel(mp); return (EINVAL); } VCTLTOREQ(&vc, req); error = VFS_SYSCTL(mp, vc.vc_op, req); vfs_rel(mp); return (error); } SYSCTL_PROC(_vfs, OID_AUTO, ctl, CTLTYPE_OPAQUE | CTLFLAG_WR, NULL, 0, sysctl_vfs_ctl, "", "Sysctl by fsid"); /* * Function to initialize a va_filerev field sensibly. * XXX: Wouldn't a random number make a lot more sense ?? */ u_quad_t init_va_filerev(void) { struct bintime bt; getbinuptime(&bt); return (((u_quad_t)bt.sec << 32LL) | (bt.frac >> 32LL)); } static int filt_vfsread(struct knote *kn, long hint); static int filt_vfswrite(struct knote *kn, long hint); static int filt_vfsvnode(struct knote *kn, long hint); static void filt_vfsdetach(struct knote *kn); static struct filterops vfsread_filtops = { .f_isfd = 1, .f_detach = filt_vfsdetach, .f_event = filt_vfsread }; static struct filterops vfswrite_filtops = { .f_isfd = 1, .f_detach = filt_vfsdetach, .f_event = filt_vfswrite }; static struct filterops vfsvnode_filtops = { .f_isfd = 1, .f_detach = filt_vfsdetach, .f_event = filt_vfsvnode }; static void vfs_knllock(void *arg) { struct vnode *vp = arg; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); } static void vfs_knlunlock(void *arg) { struct vnode *vp = arg; VOP_UNLOCK(vp, 0); } static void vfs_knl_assert_locked(void *arg) { #ifdef DEBUG_VFS_LOCKS struct vnode *vp = arg; ASSERT_VOP_LOCKED(vp, "vfs_knl_assert_locked"); #endif } static void vfs_knl_assert_unlocked(void *arg) { #ifdef DEBUG_VFS_LOCKS struct vnode *vp = arg; ASSERT_VOP_UNLOCKED(vp, "vfs_knl_assert_unlocked"); #endif } int vfs_kqfilter(struct vop_kqfilter_args *ap) { struct vnode *vp = ap->a_vp; struct knote *kn = ap->a_kn; struct knlist *knl; switch (kn->kn_filter) { case EVFILT_READ: kn->kn_fop = &vfsread_filtops; break; case EVFILT_WRITE: kn->kn_fop = &vfswrite_filtops; break; case EVFILT_VNODE: kn->kn_fop = &vfsvnode_filtops; break; default: return (EINVAL); } kn->kn_hook = (caddr_t)vp; v_addpollinfo(vp); if (vp->v_pollinfo == NULL) return (ENOMEM); knl = &vp->v_pollinfo->vpi_selinfo.si_note; vhold(vp); knlist_add(knl, kn, 0); return (0); } /* * Detach knote from vnode */ static void filt_vfsdetach(struct knote *kn) { struct vnode *vp = (struct vnode *)kn->kn_hook; KASSERT(vp->v_pollinfo != NULL, ("Missing v_pollinfo")); knlist_remove(&vp->v_pollinfo->vpi_selinfo.si_note, kn, 0); vdrop(vp); } /*ARGSUSED*/ static int filt_vfsread(struct knote *kn, long hint) { struct vnode *vp = (struct vnode *)kn->kn_hook; struct vattr va; int res; /* * filesystem is gone, so set the EOF flag and schedule * the knote for deletion. */ if (hint == NOTE_REVOKE) { VI_LOCK(vp); kn->kn_flags |= (EV_EOF | EV_ONESHOT); VI_UNLOCK(vp); return (1); } if (VOP_GETATTR(vp, &va, curthread->td_ucred)) return (0); VI_LOCK(vp); kn->kn_data = va.va_size - kn->kn_fp->f_offset; res = (kn->kn_data != 0); VI_UNLOCK(vp); return (res); } /*ARGSUSED*/ static int filt_vfswrite(struct knote *kn, long hint) { struct vnode *vp = (struct vnode *)kn->kn_hook; VI_LOCK(vp); /* * filesystem is gone, so set the EOF flag and schedule * the knote for deletion. */ if (hint == NOTE_REVOKE) kn->kn_flags |= (EV_EOF | EV_ONESHOT); kn->kn_data = 0; VI_UNLOCK(vp); return (1); } static int filt_vfsvnode(struct knote *kn, long hint) { struct vnode *vp = (struct vnode *)kn->kn_hook; int res; VI_LOCK(vp); if (kn->kn_sfflags & hint) kn->kn_fflags |= hint; if (hint == NOTE_REVOKE) { kn->kn_flags |= EV_EOF; VI_UNLOCK(vp); return (1); } res = (kn->kn_fflags != 0); VI_UNLOCK(vp); return (res); } int vfs_read_dirent(struct vop_readdir_args *ap, struct dirent *dp, off_t off) { int error; if (dp->d_reclen > ap->a_uio->uio_resid) return (ENAMETOOLONG); error = uiomove(dp, dp->d_reclen, ap->a_uio); if (error) { if (ap->a_ncookies != NULL) { if (ap->a_cookies != NULL) free(ap->a_cookies, M_TEMP); ap->a_cookies = NULL; *ap->a_ncookies = 0; } return (error); } if (ap->a_ncookies == NULL) return (0); KASSERT(ap->a_cookies, ("NULL ap->a_cookies value with non-NULL ap->a_ncookies!")); *ap->a_cookies = realloc(*ap->a_cookies, (*ap->a_ncookies + 1) * sizeof(u_long), M_TEMP, M_WAITOK | M_ZERO); (*ap->a_cookies)[*ap->a_ncookies] = off; return (0); } /* * Mark for update the access time of the file if the filesystem * supports VOP_MARKATIME. This functionality is used by execve and * mmap, so we want to avoid the I/O implied by directly setting * va_atime for the sake of efficiency. */ void vfs_mark_atime(struct vnode *vp, struct ucred *cred) { struct mount *mp; mp = vp->v_mount; ASSERT_VOP_LOCKED(vp, "vfs_mark_atime"); if (mp != NULL && (mp->mnt_flag & (MNT_NOATIME | MNT_RDONLY)) == 0) (void)VOP_MARKATIME(vp); } /* * The purpose of this routine is to remove granularity from accmode_t, * reducing it into standard unix access bits - VEXEC, VREAD, VWRITE, * VADMIN and VAPPEND. * * If it returns 0, the caller is supposed to continue with the usual * access checks using 'accmode' as modified by this routine. If it * returns nonzero value, the caller is supposed to return that value * as errno. * * Note that after this routine runs, accmode may be zero. */ int vfs_unixify_accmode(accmode_t *accmode) { /* * There is no way to specify explicit "deny" rule using * file mode or POSIX.1e ACLs. */ if (*accmode & VEXPLICIT_DENY) { *accmode = 0; return (0); } /* * None of these can be translated into usual access bits. * Also, the common case for NFSv4 ACLs is to not contain * either of these bits. Caller should check for VWRITE * on the containing directory instead. */ if (*accmode & (VDELETE_CHILD | VDELETE)) return (EPERM); if (*accmode & VADMIN_PERMS) { *accmode &= ~VADMIN_PERMS; *accmode |= VADMIN; } /* * There is no way to deny VREAD_ATTRIBUTES, VREAD_ACL * or VSYNCHRONIZE using file mode or POSIX.1e ACL. */ *accmode &= ~(VSTAT_PERMS | VSYNCHRONIZE); return (0); } /* * These are helper functions for filesystems to traverse all * their vnodes. See MNT_VNODE_FOREACH_ALL() in sys/mount.h. * * This interface replaces MNT_VNODE_FOREACH. */ MALLOC_DEFINE(M_VNODE_MARKER, "vnodemarker", "vnode marker"); struct vnode * __mnt_vnode_next_all(struct vnode **mvp, struct mount *mp) { struct vnode *vp; if (should_yield()) kern_yield(PRI_USER); MNT_ILOCK(mp); KASSERT((*mvp)->v_mount == mp, ("marker vnode mount list mismatch")); vp = TAILQ_NEXT(*mvp, v_nmntvnodes); while (vp != NULL && (vp->v_type == VMARKER || (vp->v_iflag & VI_DOOMED) != 0)) vp = TAILQ_NEXT(vp, v_nmntvnodes); /* Check if we are done */ if (vp == NULL) { __mnt_vnode_markerfree_all(mvp, mp); /* MNT_IUNLOCK(mp); -- done in above function */ mtx_assert(MNT_MTX(mp), MA_NOTOWNED); return (NULL); } TAILQ_REMOVE(&mp->mnt_nvnodelist, *mvp, v_nmntvnodes); TAILQ_INSERT_AFTER(&mp->mnt_nvnodelist, vp, *mvp, v_nmntvnodes); VI_LOCK(vp); MNT_IUNLOCK(mp); return (vp); } struct vnode * __mnt_vnode_first_all(struct vnode **mvp, struct mount *mp) { struct vnode *vp; *mvp = malloc(sizeof(struct vnode), M_VNODE_MARKER, M_WAITOK | M_ZERO); MNT_ILOCK(mp); MNT_REF(mp); (*mvp)->v_type = VMARKER; vp = TAILQ_FIRST(&mp->mnt_nvnodelist); while (vp != NULL && (vp->v_type == VMARKER || (vp->v_iflag & VI_DOOMED) != 0)) vp = TAILQ_NEXT(vp, v_nmntvnodes); /* Check if we are done */ if (vp == NULL) { MNT_REL(mp); MNT_IUNLOCK(mp); free(*mvp, M_VNODE_MARKER); *mvp = NULL; return (NULL); } (*mvp)->v_mount = mp; TAILQ_INSERT_AFTER(&mp->mnt_nvnodelist, vp, *mvp, v_nmntvnodes); VI_LOCK(vp); MNT_IUNLOCK(mp); return (vp); } void __mnt_vnode_markerfree_all(struct vnode **mvp, struct mount *mp) { if (*mvp == NULL) { MNT_IUNLOCK(mp); return; } mtx_assert(MNT_MTX(mp), MA_OWNED); KASSERT((*mvp)->v_mount == mp, ("marker vnode mount list mismatch")); TAILQ_REMOVE(&mp->mnt_nvnodelist, *mvp, v_nmntvnodes); MNT_REL(mp); MNT_IUNLOCK(mp); free(*mvp, M_VNODE_MARKER); *mvp = NULL; } /* * These are helper functions for filesystems to traverse their * active vnodes. See MNT_VNODE_FOREACH_ACTIVE() in sys/mount.h */ static void mnt_vnode_markerfree_active(struct vnode **mvp, struct mount *mp) { KASSERT((*mvp)->v_mount == mp, ("marker vnode mount list mismatch")); MNT_ILOCK(mp); MNT_REL(mp); MNT_IUNLOCK(mp); free(*mvp, M_VNODE_MARKER); *mvp = NULL; } static struct vnode * mnt_vnode_next_active(struct vnode **mvp, struct mount *mp) { struct vnode *vp, *nvp; mtx_assert(&vnode_free_list_mtx, MA_OWNED); KASSERT((*mvp)->v_mount == mp, ("marker vnode mount list mismatch")); restart: vp = TAILQ_NEXT(*mvp, v_actfreelist); TAILQ_REMOVE(&mp->mnt_activevnodelist, *mvp, v_actfreelist); while (vp != NULL) { if (vp->v_type == VMARKER) { vp = TAILQ_NEXT(vp, v_actfreelist); continue; } if (!VI_TRYLOCK(vp)) { if (mp_ncpus == 1 || should_yield()) { TAILQ_INSERT_BEFORE(vp, *mvp, v_actfreelist); mtx_unlock(&vnode_free_list_mtx); pause("vnacti", 1); mtx_lock(&vnode_free_list_mtx); goto restart; } continue; } KASSERT(vp->v_type != VMARKER, ("locked marker %p", vp)); KASSERT(vp->v_mount == mp || vp->v_mount == NULL, ("alien vnode on the active list %p %p", vp, mp)); if (vp->v_mount == mp && (vp->v_iflag & VI_DOOMED) == 0) break; nvp = TAILQ_NEXT(vp, v_actfreelist); VI_UNLOCK(vp); vp = nvp; } /* Check if we are done */ if (vp == NULL) { mtx_unlock(&vnode_free_list_mtx); mnt_vnode_markerfree_active(mvp, mp); return (NULL); } TAILQ_INSERT_AFTER(&mp->mnt_activevnodelist, vp, *mvp, v_actfreelist); mtx_unlock(&vnode_free_list_mtx); ASSERT_VI_LOCKED(vp, "active iter"); KASSERT((vp->v_iflag & VI_ACTIVE) != 0, ("Non-active vp %p", vp)); return (vp); } struct vnode * __mnt_vnode_next_active(struct vnode **mvp, struct mount *mp) { if (should_yield()) kern_yield(PRI_USER); mtx_lock(&vnode_free_list_mtx); return (mnt_vnode_next_active(mvp, mp)); } struct vnode * __mnt_vnode_first_active(struct vnode **mvp, struct mount *mp) { struct vnode *vp; *mvp = malloc(sizeof(struct vnode), M_VNODE_MARKER, M_WAITOK | M_ZERO); MNT_ILOCK(mp); MNT_REF(mp); MNT_IUNLOCK(mp); (*mvp)->v_type = VMARKER; (*mvp)->v_mount = mp; mtx_lock(&vnode_free_list_mtx); vp = TAILQ_FIRST(&mp->mnt_activevnodelist); if (vp == NULL) { mtx_unlock(&vnode_free_list_mtx); mnt_vnode_markerfree_active(mvp, mp); return (NULL); } TAILQ_INSERT_BEFORE(vp, *mvp, v_actfreelist); return (mnt_vnode_next_active(mvp, mp)); } void __mnt_vnode_markerfree_active(struct vnode **mvp, struct mount *mp) { if (*mvp == NULL) return; mtx_lock(&vnode_free_list_mtx); TAILQ_REMOVE(&mp->mnt_activevnodelist, *mvp, v_actfreelist); mtx_unlock(&vnode_free_list_mtx); mnt_vnode_markerfree_active(mvp, mp); } Index: projects/building-blocks/sys/netinet6/nd6.c =================================================================== --- projects/building-blocks/sys/netinet6/nd6.c (revision 275663) +++ projects/building-blocks/sys/netinet6/nd6.c (revision 275664) @@ -1,2476 +1,2478 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: nd6.c,v 1.144 2001/05/24 07:44:00 itojun Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define L3_ADDR_SIN6(le) ((struct sockaddr_in6 *) L3_ADDR(le)) #include #include #include #include #include #include #include #include #include #include #include #define ND6_SLOWTIMER_INTERVAL (60 * 60) /* 1 hour */ #define ND6_RECALC_REACHTM_INTERVAL (60 * 120) /* 2 hours */ #define SIN6(s) ((const struct sockaddr_in6 *)(s)) /* timer values */ VNET_DEFINE(int, nd6_prune) = 1; /* walk list every 1 seconds */ VNET_DEFINE(int, nd6_delay) = 5; /* delay first probe time 5 second */ VNET_DEFINE(int, nd6_umaxtries) = 3; /* maximum unicast query */ VNET_DEFINE(int, nd6_mmaxtries) = 3; /* maximum multicast query */ VNET_DEFINE(int, nd6_useloopback) = 1; /* use loopback interface for * local traffic */ VNET_DEFINE(int, nd6_gctimer) = (60 * 60 * 24); /* 1 day: garbage * collection timer */ /* preventing too many loops in ND option parsing */ static VNET_DEFINE(int, nd6_maxndopt) = 10; /* max # of ND options allowed */ VNET_DEFINE(int, nd6_maxnudhint) = 0; /* max # of subsequent upper * layer hints */ static VNET_DEFINE(int, nd6_maxqueuelen) = 1; /* max pkts cached in unresolved * ND entries */ #define V_nd6_maxndopt VNET(nd6_maxndopt) #define V_nd6_maxqueuelen VNET(nd6_maxqueuelen) #ifdef ND6_DEBUG VNET_DEFINE(int, nd6_debug) = 1; #else VNET_DEFINE(int, nd6_debug) = 0; #endif /* for debugging? */ #if 0 static int nd6_inuse, nd6_allocated; #endif VNET_DEFINE(struct nd_drhead, nd_defrouter); VNET_DEFINE(struct nd_prhead, nd_prefix); VNET_DEFINE(int, nd6_recalc_reachtm_interval) = ND6_RECALC_REACHTM_INTERVAL; #define V_nd6_recalc_reachtm_interval VNET(nd6_recalc_reachtm_interval) int (*send_sendso_input_hook)(struct mbuf *, struct ifnet *, int, int); static int nd6_is_new_addr_neighbor(struct sockaddr_in6 *, struct ifnet *); static void nd6_setmtu0(struct ifnet *, struct nd_ifinfo *); static void nd6_slowtimo(void *); static int regen_tmpaddr(struct in6_ifaddr *); static struct llentry *nd6_free(struct llentry *, int); static void nd6_llinfo_timer(void *); static void clear_llinfo_pqueue(struct llentry *); static void nd6_rtrequest(int, struct rtentry *, struct rt_addrinfo *); static VNET_DEFINE(struct callout, nd6_slowtimo_ch); #define V_nd6_slowtimo_ch VNET(nd6_slowtimo_ch) VNET_DEFINE(struct callout, nd6_timer_ch); void nd6_init(void) { LIST_INIT(&V_nd_prefix); /* initialization of the default router list */ TAILQ_INIT(&V_nd_defrouter); /* start timer */ callout_init(&V_nd6_slowtimo_ch, 0); callout_reset(&V_nd6_slowtimo_ch, ND6_SLOWTIMER_INTERVAL * hz, nd6_slowtimo, curvnet); + + nd6_dad_init(); } #ifdef VIMAGE void nd6_destroy() { callout_drain(&V_nd6_slowtimo_ch); callout_drain(&V_nd6_timer_ch); } #endif struct nd_ifinfo * nd6_ifattach(struct ifnet *ifp) { struct nd_ifinfo *nd; nd = (struct nd_ifinfo *)malloc(sizeof(*nd), M_IP6NDP, M_WAITOK|M_ZERO); nd->initialized = 1; nd->chlim = IPV6_DEFHLIM; nd->basereachable = REACHABLE_TIME; nd->reachable = ND_COMPUTE_RTIME(nd->basereachable); nd->retrans = RETRANS_TIMER; nd->flags = ND6_IFF_PERFORMNUD; /* A loopback interface always has ND6_IFF_AUTO_LINKLOCAL. * XXXHRS: Clear ND6_IFF_AUTO_LINKLOCAL on an IFT_BRIDGE interface by * default regardless of the V_ip6_auto_linklocal configuration to * give a reasonable default behavior. */ if ((V_ip6_auto_linklocal && ifp->if_type != IFT_BRIDGE) || (ifp->if_flags & IFF_LOOPBACK)) nd->flags |= ND6_IFF_AUTO_LINKLOCAL; /* * A loopback interface does not need to accept RTADV. * XXXHRS: Clear ND6_IFF_ACCEPT_RTADV on an IFT_BRIDGE interface by * default regardless of the V_ip6_accept_rtadv configuration to * prevent the interface from accepting RA messages arrived * on one of the member interfaces with ND6_IFF_ACCEPT_RTADV. */ if (V_ip6_accept_rtadv && !(ifp->if_flags & IFF_LOOPBACK) && (ifp->if_type != IFT_BRIDGE)) nd->flags |= ND6_IFF_ACCEPT_RTADV; if (V_ip6_no_radr && !(ifp->if_flags & IFF_LOOPBACK)) nd->flags |= ND6_IFF_NO_RADR; /* XXX: we cannot call nd6_setmtu since ifp is not fully initialized */ nd6_setmtu0(ifp, nd); return nd; } void nd6_ifdetach(struct nd_ifinfo *nd) { free(nd, M_IP6NDP); } /* * Reset ND level link MTU. This function is called when the physical MTU * changes, which means we might have to adjust the ND level MTU. */ void nd6_setmtu(struct ifnet *ifp) { nd6_setmtu0(ifp, ND_IFINFO(ifp)); } /* XXX todo: do not maintain copy of ifp->if_mtu in ndi->maxmtu */ void nd6_setmtu0(struct ifnet *ifp, struct nd_ifinfo *ndi) { u_int32_t omaxmtu; omaxmtu = ndi->maxmtu; switch (ifp->if_type) { case IFT_ARCNET: ndi->maxmtu = MIN(ARC_PHDS_MAXMTU, ifp->if_mtu); /* RFC2497 */ break; case IFT_FDDI: ndi->maxmtu = MIN(FDDIIPMTU, ifp->if_mtu); /* RFC2467 */ break; case IFT_ISO88025: ndi->maxmtu = MIN(ISO88025_MAX_MTU, ifp->if_mtu); break; default: ndi->maxmtu = ifp->if_mtu; break; } /* * Decreasing the interface MTU under IPV6 minimum MTU may cause * undesirable situation. We thus notify the operator of the change * explicitly. The check for omaxmtu is necessary to restrict the * log to the case of changing the MTU, not initializing it. */ if (omaxmtu >= IPV6_MMTU && ndi->maxmtu < IPV6_MMTU) { log(LOG_NOTICE, "nd6_setmtu0: " "new link MTU on %s (%lu) is too small for IPv6\n", if_name(ifp), (unsigned long)ndi->maxmtu); } if (ndi->maxmtu > V_in6_maxmtu) in6_setmaxmtu(); /* check all interfaces just in case */ } void nd6_option_init(void *opt, int icmp6len, union nd_opts *ndopts) { bzero(ndopts, sizeof(*ndopts)); ndopts->nd_opts_search = (struct nd_opt_hdr *)opt; ndopts->nd_opts_last = (struct nd_opt_hdr *)(((u_char *)opt) + icmp6len); if (icmp6len == 0) { ndopts->nd_opts_done = 1; ndopts->nd_opts_search = NULL; } } /* * Take one ND option. */ struct nd_opt_hdr * nd6_option(union nd_opts *ndopts) { struct nd_opt_hdr *nd_opt; int olen; KASSERT(ndopts != NULL, ("%s: ndopts == NULL", __func__)); KASSERT(ndopts->nd_opts_last != NULL, ("%s: uninitialized ndopts", __func__)); if (ndopts->nd_opts_search == NULL) return NULL; if (ndopts->nd_opts_done) return NULL; nd_opt = ndopts->nd_opts_search; /* make sure nd_opt_len is inside the buffer */ if ((caddr_t)&nd_opt->nd_opt_len >= (caddr_t)ndopts->nd_opts_last) { bzero(ndopts, sizeof(*ndopts)); return NULL; } olen = nd_opt->nd_opt_len << 3; if (olen == 0) { /* * Message validation requires that all included * options have a length that is greater than zero. */ bzero(ndopts, sizeof(*ndopts)); return NULL; } ndopts->nd_opts_search = (struct nd_opt_hdr *)((caddr_t)nd_opt + olen); if (ndopts->nd_opts_search > ndopts->nd_opts_last) { /* option overruns the end of buffer, invalid */ bzero(ndopts, sizeof(*ndopts)); return NULL; } else if (ndopts->nd_opts_search == ndopts->nd_opts_last) { /* reached the end of options chain */ ndopts->nd_opts_done = 1; ndopts->nd_opts_search = NULL; } return nd_opt; } /* * Parse multiple ND options. * This function is much easier to use, for ND routines that do not need * multiple options of the same type. */ int nd6_options(union nd_opts *ndopts) { struct nd_opt_hdr *nd_opt; int i = 0; KASSERT(ndopts != NULL, ("%s: ndopts == NULL", __func__)); KASSERT(ndopts->nd_opts_last != NULL, ("%s: uninitialized ndopts", __func__)); if (ndopts->nd_opts_search == NULL) return 0; while (1) { nd_opt = nd6_option(ndopts); if (nd_opt == NULL && ndopts->nd_opts_last == NULL) { /* * Message validation requires that all included * options have a length that is greater than zero. */ ICMP6STAT_INC(icp6s_nd_badopt); bzero(ndopts, sizeof(*ndopts)); return -1; } if (nd_opt == NULL) goto skip1; switch (nd_opt->nd_opt_type) { case ND_OPT_SOURCE_LINKADDR: case ND_OPT_TARGET_LINKADDR: case ND_OPT_MTU: case ND_OPT_REDIRECTED_HEADER: if (ndopts->nd_opt_array[nd_opt->nd_opt_type]) { nd6log((LOG_INFO, "duplicated ND6 option found (type=%d)\n", nd_opt->nd_opt_type)); /* XXX bark? */ } else { ndopts->nd_opt_array[nd_opt->nd_opt_type] = nd_opt; } break; case ND_OPT_PREFIX_INFORMATION: if (ndopts->nd_opt_array[nd_opt->nd_opt_type] == 0) { ndopts->nd_opt_array[nd_opt->nd_opt_type] = nd_opt; } ndopts->nd_opts_pi_end = (struct nd_opt_prefix_info *)nd_opt; break; /* What about ND_OPT_ROUTE_INFO? RFC 4191 */ case ND_OPT_RDNSS: /* RFC 6106 */ case ND_OPT_DNSSL: /* RFC 6106 */ /* * Silently ignore options we know and do not care about * in the kernel. */ break; default: /* * Unknown options must be silently ignored, * to accomodate future extension to the protocol. */ nd6log((LOG_DEBUG, "nd6_options: unsupported option %d - " "option ignored\n", nd_opt->nd_opt_type)); } skip1: i++; if (i > V_nd6_maxndopt) { ICMP6STAT_INC(icp6s_nd_toomanyopt); nd6log((LOG_INFO, "too many loop in nd opt\n")); break; } if (ndopts->nd_opts_done) break; } return 0; } /* * ND6 timer routine to handle ND6 entries */ void nd6_llinfo_settimer_locked(struct llentry *ln, long tick) { int canceled; LLE_WLOCK_ASSERT(ln); if (tick < 0) { ln->la_expire = 0; ln->ln_ntick = 0; canceled = callout_stop(&ln->ln_timer_ch); } else { ln->la_expire = time_uptime + tick / hz; LLE_ADDREF(ln); if (tick > INT_MAX) { ln->ln_ntick = tick - INT_MAX; canceled = callout_reset(&ln->ln_timer_ch, INT_MAX, nd6_llinfo_timer, ln); } else { ln->ln_ntick = 0; canceled = callout_reset(&ln->ln_timer_ch, tick, nd6_llinfo_timer, ln); } } if (canceled) LLE_REMREF(ln); } void nd6_llinfo_settimer(struct llentry *ln, long tick) { LLE_WLOCK(ln); nd6_llinfo_settimer_locked(ln, tick); LLE_WUNLOCK(ln); } static void nd6_llinfo_timer(void *arg) { struct llentry *ln; struct in6_addr *dst; struct ifnet *ifp; struct nd_ifinfo *ndi = NULL; KASSERT(arg != NULL, ("%s: arg NULL", __func__)); ln = (struct llentry *)arg; LLE_WLOCK_ASSERT(ln); ifp = ln->lle_tbl->llt_ifp; CURVNET_SET(ifp->if_vnet); if (ln->ln_ntick > 0) { if (ln->ln_ntick > INT_MAX) { ln->ln_ntick -= INT_MAX; nd6_llinfo_settimer_locked(ln, INT_MAX); } else { ln->ln_ntick = 0; nd6_llinfo_settimer_locked(ln, ln->ln_ntick); } goto done; } ndi = ND_IFINFO(ifp); dst = &L3_ADDR_SIN6(ln)->sin6_addr; if (ln->la_flags & LLE_STATIC) { goto done; } if (ln->la_flags & LLE_DELETED) { (void)nd6_free(ln, 0); ln = NULL; goto done; } switch (ln->ln_state) { case ND6_LLINFO_INCOMPLETE: if (ln->la_asked < V_nd6_mmaxtries) { ln->la_asked++; nd6_llinfo_settimer_locked(ln, (long)ndi->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, NULL, dst, ln, 0); LLE_WLOCK(ln); } else { struct mbuf *m = ln->la_hold; if (m) { struct mbuf *m0; /* * assuming every packet in la_hold has the * same IP header. Send error after unlock. */ m0 = m->m_nextpkt; m->m_nextpkt = NULL; ln->la_hold = m0; clear_llinfo_pqueue(ln); } EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_TIMEDOUT); (void)nd6_free(ln, 0); ln = NULL; if (m != NULL) icmp6_error2(m, ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR, 0, ifp); } break; case ND6_LLINFO_REACHABLE: if (!ND6_LLINFO_PERMANENT(ln)) { ln->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } break; case ND6_LLINFO_STALE: /* Garbage Collection(RFC 2461 5.3) */ if (!ND6_LLINFO_PERMANENT(ln)) { EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_EXPIRED); (void)nd6_free(ln, 1); ln = NULL; } break; case ND6_LLINFO_DELAY: if (ndi && (ndi->flags & ND6_IFF_PERFORMNUD) != 0) { /* We need NUD */ ln->la_asked = 1; ln->ln_state = ND6_LLINFO_PROBE; nd6_llinfo_settimer_locked(ln, (long)ndi->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, dst, dst, ln, 0); LLE_WLOCK(ln); } else { ln->ln_state = ND6_LLINFO_STALE; /* XXX */ nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } break; case ND6_LLINFO_PROBE: if (ln->la_asked < V_nd6_umaxtries) { ln->la_asked++; nd6_llinfo_settimer_locked(ln, (long)ndi->retrans * hz / 1000); LLE_WUNLOCK(ln); nd6_ns_output(ifp, dst, dst, ln, 0); LLE_WLOCK(ln); } else { EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_EXPIRED); (void)nd6_free(ln, 0); ln = NULL; } break; default: panic("%s: paths in a dark night can be confusing: %d", __func__, ln->ln_state); } done: if (ln != NULL) LLE_FREE_LOCKED(ln); CURVNET_RESTORE(); } /* * ND6 timer routine to expire default route list and prefix list */ void nd6_timer(void *arg) { CURVNET_SET((struct vnet *) arg); struct nd_defrouter *dr, *ndr; struct nd_prefix *pr, *npr; struct in6_ifaddr *ia6, *nia6; callout_reset(&V_nd6_timer_ch, V_nd6_prune * hz, nd6_timer, curvnet); /* expire default router list */ TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, ndr) { if (dr->expire && dr->expire < time_uptime) defrtrlist_del(dr); } /* * expire interface addresses. * in the past the loop was inside prefix expiry processing. * However, from a stricter speci-confrmance standpoint, we should * rather separate address lifetimes and prefix lifetimes. * * XXXRW: in6_ifaddrhead locking. */ addrloop: TAILQ_FOREACH_SAFE(ia6, &V_in6_ifaddrhead, ia_link, nia6) { /* check address lifetime */ if (IFA6_IS_INVALID(ia6)) { int regen = 0; /* * If the expiring address is temporary, try * regenerating a new one. This would be useful when * we suspended a laptop PC, then turned it on after a * period that could invalidate all temporary * addresses. Although we may have to restart the * loop (see below), it must be after purging the * address. Otherwise, we'd see an infinite loop of * regeneration. */ if (V_ip6_use_tempaddr && (ia6->ia6_flags & IN6_IFF_TEMPORARY) != 0) { if (regen_tmpaddr(ia6) == 0) regen = 1; } in6_purgeaddr(&ia6->ia_ifa); if (regen) goto addrloop; /* XXX: see below */ } else if (IFA6_IS_DEPRECATED(ia6)) { int oldflags = ia6->ia6_flags; ia6->ia6_flags |= IN6_IFF_DEPRECATED; /* * If a temporary address has just become deprecated, * regenerate a new one if possible. */ if (V_ip6_use_tempaddr && (ia6->ia6_flags & IN6_IFF_TEMPORARY) != 0 && (oldflags & IN6_IFF_DEPRECATED) == 0) { if (regen_tmpaddr(ia6) == 0) { /* * A new temporary address is * generated. * XXX: this means the address chain * has changed while we are still in * the loop. Although the change * would not cause disaster (because * it's not a deletion, but an * addition,) we'd rather restart the * loop just for safety. Or does this * significantly reduce performance?? */ goto addrloop; } } } else { /* * A new RA might have made a deprecated address * preferred. */ ia6->ia6_flags &= ~IN6_IFF_DEPRECATED; } } /* expire prefix list */ LIST_FOREACH_SAFE(pr, &V_nd_prefix, ndpr_entry, npr) { /* * check prefix lifetime. * since pltime is just for autoconf, pltime processing for * prefix is not necessary. */ if (pr->ndpr_vltime != ND6_INFINITE_LIFETIME && time_uptime - pr->ndpr_lastupdate > pr->ndpr_vltime) { /* * address expiration and prefix expiration are * separate. NEVER perform in6_purgeaddr here. */ prelist_remove(pr); } } CURVNET_RESTORE(); } /* * ia6 - deprecated/invalidated temporary address */ static int regen_tmpaddr(struct in6_ifaddr *ia6) { struct ifaddr *ifa; struct ifnet *ifp; struct in6_ifaddr *public_ifa6 = NULL; ifp = ia6->ia_ifa.ifa_ifp; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { struct in6_ifaddr *it6; if (ifa->ifa_addr->sa_family != AF_INET6) continue; it6 = (struct in6_ifaddr *)ifa; /* ignore no autoconf addresses. */ if ((it6->ia6_flags & IN6_IFF_AUTOCONF) == 0) continue; /* ignore autoconf addresses with different prefixes. */ if (it6->ia6_ndpr == NULL || it6->ia6_ndpr != ia6->ia6_ndpr) continue; /* * Now we are looking at an autoconf address with the same * prefix as ours. If the address is temporary and is still * preferred, do not create another one. It would be rare, but * could happen, for example, when we resume a laptop PC after * a long period. */ if ((it6->ia6_flags & IN6_IFF_TEMPORARY) != 0 && !IFA6_IS_DEPRECATED(it6)) { public_ifa6 = NULL; break; } /* * This is a public autoconf address that has the same prefix * as ours. If it is preferred, keep it. We can't break the * loop here, because there may be a still-preferred temporary * address with the prefix. */ if (!IFA6_IS_DEPRECATED(it6)) public_ifa6 = it6; if (public_ifa6 != NULL) ifa_ref(&public_ifa6->ia_ifa); } IF_ADDR_RUNLOCK(ifp); if (public_ifa6 != NULL) { int e; if ((e = in6_tmpifadd(public_ifa6, 0, 0)) != 0) { ifa_free(&public_ifa6->ia_ifa); log(LOG_NOTICE, "regen_tmpaddr: failed to create a new" " tmp addr,errno=%d\n", e); return (-1); } ifa_free(&public_ifa6->ia_ifa); return (0); } return (-1); } /* * Nuke neighbor cache/prefix/default router management table, right before * ifp goes away. */ void nd6_purge(struct ifnet *ifp) { struct nd_defrouter *dr, *ndr; struct nd_prefix *pr, *npr; /* * Nuke default router list entries toward ifp. * We defer removal of default router list entries that is installed * in the routing table, in order to keep additional side effects as * small as possible. */ TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, ndr) { if (dr->installed) continue; if (dr->ifp == ifp) defrtrlist_del(dr); } TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, ndr) { if (!dr->installed) continue; if (dr->ifp == ifp) defrtrlist_del(dr); } /* Nuke prefix list entries toward ifp */ LIST_FOREACH_SAFE(pr, &V_nd_prefix, ndpr_entry, npr) { if (pr->ndpr_ifp == ifp) { /* * Because if_detach() does *not* release prefixes * while purging addresses the reference count will * still be above zero. We therefore reset it to * make sure that the prefix really gets purged. */ pr->ndpr_refcnt = 0; /* * Previously, pr->ndpr_addr is removed as well, * but I strongly believe we don't have to do it. * nd6_purge() is only called from in6_ifdetach(), * which removes all the associated interface addresses * by itself. * (jinmei@kame.net 20010129) */ prelist_remove(pr); } } /* cancel default outgoing interface setting */ if (V_nd6_defifindex == ifp->if_index) nd6_setdefaultiface(0); if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) { /* Refresh default router list. */ defrouter_select(); } /* XXXXX * We do not nuke the neighbor cache entries here any more * because the neighbor cache is kept in if_afdata[AF_INET6]. * nd6_purge() is invoked by in6_ifdetach() which is called * from if_detach() where everything gets purged. So let * in6_domifdetach() do the actual L2 table purging work. */ } /* * the caller acquires and releases the lock on the lltbls * Returns the llentry locked */ struct llentry * nd6_lookup(struct in6_addr *addr6, int flags, struct ifnet *ifp) { struct sockaddr_in6 sin6; struct llentry *ln; int llflags; bzero(&sin6, sizeof(sin6)); sin6.sin6_len = sizeof(struct sockaddr_in6); sin6.sin6_family = AF_INET6; sin6.sin6_addr = *addr6; IF_AFDATA_LOCK_ASSERT(ifp); llflags = 0; if (flags & ND6_CREATE) llflags |= LLE_CREATE; if (flags & ND6_EXCLUSIVE) llflags |= LLE_EXCLUSIVE; ln = lla_lookup(LLTABLE6(ifp), llflags, (struct sockaddr *)&sin6); if ((ln != NULL) && (llflags & LLE_CREATE)) ln->ln_state = ND6_LLINFO_NOSTATE; return (ln); } /* * Test whether a given IPv6 address is a neighbor or not, ignoring * the actual neighbor cache. The neighbor cache is ignored in order * to not reenter the routing code from within itself. */ static int nd6_is_new_addr_neighbor(struct sockaddr_in6 *addr, struct ifnet *ifp) { struct nd_prefix *pr; struct ifaddr *dstaddr; /* * A link-local address is always a neighbor. * XXX: a link does not necessarily specify a single interface. */ if (IN6_IS_ADDR_LINKLOCAL(&addr->sin6_addr)) { struct sockaddr_in6 sin6_copy; u_int32_t zone; /* * We need sin6_copy since sa6_recoverscope() may modify the * content (XXX). */ sin6_copy = *addr; if (sa6_recoverscope(&sin6_copy)) return (0); /* XXX: should be impossible */ if (in6_setscope(&sin6_copy.sin6_addr, ifp, &zone)) return (0); if (sin6_copy.sin6_scope_id == zone) return (1); else return (0); } /* * If the address matches one of our addresses, * it should be a neighbor. * If the address matches one of our on-link prefixes, it should be a * neighbor. */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { if (pr->ndpr_ifp != ifp) continue; if (!(pr->ndpr_stateflags & NDPRF_ONLINK)) { struct rtentry *rt; /* Always use the default FIB here. */ rt = in6_rtalloc1((struct sockaddr *)&pr->ndpr_prefix, 0, 0, RT_DEFAULT_FIB); if (rt == NULL) continue; /* * This is the case where multiple interfaces * have the same prefix, but only one is installed * into the routing table and that prefix entry * is not the one being examined here. In the case * where RADIX_MPATH is enabled, multiple route * entries (of the same rt_key value) will be * installed because the interface addresses all * differ. */ if (!IN6_ARE_ADDR_EQUAL(&pr->ndpr_prefix.sin6_addr, &((struct sockaddr_in6 *)rt_key(rt))->sin6_addr)) { RTFREE_LOCKED(rt); continue; } RTFREE_LOCKED(rt); } if (IN6_ARE_MASKED_ADDR_EQUAL(&pr->ndpr_prefix.sin6_addr, &addr->sin6_addr, &pr->ndpr_mask)) return (1); } /* * If the address is assigned on the node of the other side of * a p2p interface, the address should be a neighbor. */ dstaddr = ifa_ifwithdstaddr((struct sockaddr *)addr, RT_ALL_FIBS); if (dstaddr != NULL) { if (dstaddr->ifa_ifp == ifp) { ifa_free(dstaddr); return (1); } ifa_free(dstaddr); } /* * If the default router list is empty, all addresses are regarded * as on-link, and thus, as a neighbor. */ if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV && TAILQ_EMPTY(&V_nd_defrouter) && V_nd6_defifindex == ifp->if_index) { return (1); } return (0); } /* * Detect if a given IPv6 address identifies a neighbor on a given link. * XXX: should take care of the destination of a p2p link? */ int nd6_is_addr_neighbor(struct sockaddr_in6 *addr, struct ifnet *ifp) { struct llentry *lle; int rc = 0; IF_AFDATA_UNLOCK_ASSERT(ifp); if (nd6_is_new_addr_neighbor(addr, ifp)) return (1); /* * Even if the address matches none of our addresses, it might be * in the neighbor cache. */ IF_AFDATA_RLOCK(ifp); if ((lle = nd6_lookup(&addr->sin6_addr, 0, ifp)) != NULL) { LLE_RUNLOCK(lle); rc = 1; } IF_AFDATA_RUNLOCK(ifp); return (rc); } /* * Free an nd6 llinfo entry. * Since the function would cause significant changes in the kernel, DO NOT * make it global, unless you have a strong reason for the change, and are sure * that the change is safe. */ static struct llentry * nd6_free(struct llentry *ln, int gc) { struct llentry *next; struct nd_defrouter *dr; struct ifnet *ifp; LLE_WLOCK_ASSERT(ln); /* * we used to have pfctlinput(PRC_HOSTDEAD) here. * even though it is not harmful, it was not really necessary. */ /* cancel timer */ nd6_llinfo_settimer_locked(ln, -1); ifp = ln->lle_tbl->llt_ifp; if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) { dr = defrouter_lookup(&L3_ADDR_SIN6(ln)->sin6_addr, ifp); if (dr != NULL && dr->expire && ln->ln_state == ND6_LLINFO_STALE && gc) { /* * If the reason for the deletion is just garbage * collection, and the neighbor is an active default * router, do not delete it. Instead, reset the GC * timer using the router's lifetime. * Simply deleting the entry would affect default * router selection, which is not necessarily a good * thing, especially when we're using router preference * values. * XXX: the check for ln_state would be redundant, * but we intentionally keep it just in case. */ if (dr->expire > time_uptime) nd6_llinfo_settimer_locked(ln, (dr->expire - time_uptime) * hz); else nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); next = LIST_NEXT(ln, lle_next); LLE_REMREF(ln); LLE_WUNLOCK(ln); return (next); } if (dr) { /* * Unreachablity of a router might affect the default * router selection and on-link detection of advertised * prefixes. */ /* * Temporarily fake the state to choose a new default * router and to perform on-link determination of * prefixes correctly. * Below the state will be set correctly, * or the entry itself will be deleted. */ ln->ln_state = ND6_LLINFO_INCOMPLETE; } if (ln->ln_router || dr) { /* * We need to unlock to avoid a LOR with rt6_flush() with the * rnh and for the calls to pfxlist_onlink_check() and * defrouter_select() in the block further down for calls * into nd6_lookup(). We still hold a ref. */ LLE_WUNLOCK(ln); /* * rt6_flush must be called whether or not the neighbor * is in the Default Router List. * See a corresponding comment in nd6_na_input(). */ rt6_flush(&L3_ADDR_SIN6(ln)->sin6_addr, ifp); } if (dr) { /* * Since defrouter_select() does not affect the * on-link determination and MIP6 needs the check * before the default router selection, we perform * the check now. */ pfxlist_onlink_check(); /* * Refresh default router list. */ defrouter_select(); } if (ln->ln_router || dr) LLE_WLOCK(ln); } /* * Before deleting the entry, remember the next entry as the * return value. We need this because pfxlist_onlink_check() above * might have freed other entries (particularly the old next entry) as * a side effect (XXX). */ next = LIST_NEXT(ln, lle_next); /* * Save to unlock. We still hold an extra reference and will not * free(9) in llentry_free() if someone else holds one as well. */ LLE_WUNLOCK(ln); IF_AFDATA_LOCK(ifp); LLE_WLOCK(ln); /* Guard against race with other llentry_free(). */ if (ln->la_flags & LLE_LINKED) { LLE_REMREF(ln); llentry_free(ln); } else LLE_FREE_LOCKED(ln); IF_AFDATA_UNLOCK(ifp); return (next); } /* * Upper-layer reachability hint for Neighbor Unreachability Detection. * * XXX cost-effective methods? */ void nd6_nud_hint(struct rtentry *rt, struct in6_addr *dst6, int force) { struct llentry *ln; struct ifnet *ifp; if ((dst6 == NULL) || (rt == NULL)) return; ifp = rt->rt_ifp; IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(dst6, ND6_EXCLUSIVE, NULL); IF_AFDATA_RUNLOCK(ifp); if (ln == NULL) return; if (ln->ln_state < ND6_LLINFO_REACHABLE) goto done; /* * if we get upper-layer reachability confirmation many times, * it is possible we have false information. */ if (!force) { ln->ln_byhint++; if (ln->ln_byhint > V_nd6_maxnudhint) { goto done; } } ln->ln_state = ND6_LLINFO_REACHABLE; if (!ND6_LLINFO_PERMANENT(ln)) { nd6_llinfo_settimer_locked(ln, (long)ND_IFINFO(rt->rt_ifp)->reachable * hz); } done: LLE_WUNLOCK(ln); } /* * Rejuvenate this function for routing operations related * processing. */ void nd6_rtrequest(int req, struct rtentry *rt, struct rt_addrinfo *info) { struct sockaddr_in6 *gateway; struct nd_defrouter *dr; struct ifnet *ifp; gateway = (struct sockaddr_in6 *)rt->rt_gateway; ifp = rt->rt_ifp; switch (req) { case RTM_ADD: break; case RTM_DELETE: if (!ifp) return; /* * Only indirect routes are interesting. */ if ((rt->rt_flags & RTF_GATEWAY) == 0) return; /* * check for default route */ if (IN6_ARE_ADDR_EQUAL(&in6addr_any, &SIN6(rt_key(rt))->sin6_addr)) { dr = defrouter_lookup(&gateway->sin6_addr, ifp); if (dr != NULL) dr->installed = 0; } break; } } int nd6_ioctl(u_long cmd, caddr_t data, struct ifnet *ifp) { struct in6_drlist *drl = (struct in6_drlist *)data; struct in6_oprlist *oprl = (struct in6_oprlist *)data; struct in6_ndireq *ndi = (struct in6_ndireq *)data; struct in6_nbrinfo *nbi = (struct in6_nbrinfo *)data; struct in6_ndifreq *ndif = (struct in6_ndifreq *)data; struct nd_defrouter *dr; struct nd_prefix *pr; int i = 0, error = 0; if (ifp->if_afdata[AF_INET6] == NULL) return (EPFNOSUPPORT); switch (cmd) { case SIOCGDRLST_IN6: /* * obsolete API, use sysctl under net.inet6.icmp6 */ bzero(drl, sizeof(*drl)); TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { if (i >= DRLSTSIZ) break; drl->defrouter[i].rtaddr = dr->rtaddr; in6_clearscope(&drl->defrouter[i].rtaddr); drl->defrouter[i].flags = dr->flags; drl->defrouter[i].rtlifetime = dr->rtlifetime; drl->defrouter[i].expire = dr->expire + (time_second - time_uptime); drl->defrouter[i].if_index = dr->ifp->if_index; i++; } break; case SIOCGPRLST_IN6: /* * obsolete API, use sysctl under net.inet6.icmp6 * * XXX the structure in6_prlist was changed in backward- * incompatible manner. in6_oprlist is used for SIOCGPRLST_IN6, * in6_prlist is used for nd6_sysctl() - fill_prlist(). */ /* * XXX meaning of fields, especialy "raflags", is very * differnet between RA prefix list and RR/static prefix list. * how about separating ioctls into two? */ bzero(oprl, sizeof(*oprl)); LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { struct nd_pfxrouter *pfr; int j; if (i >= PRLSTSIZ) break; oprl->prefix[i].prefix = pr->ndpr_prefix.sin6_addr; oprl->prefix[i].raflags = pr->ndpr_raf; oprl->prefix[i].prefixlen = pr->ndpr_plen; oprl->prefix[i].vltime = pr->ndpr_vltime; oprl->prefix[i].pltime = pr->ndpr_pltime; oprl->prefix[i].if_index = pr->ndpr_ifp->if_index; if (pr->ndpr_vltime == ND6_INFINITE_LIFETIME) oprl->prefix[i].expire = 0; else { time_t maxexpire; /* XXX: we assume time_t is signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (pr->ndpr_vltime < maxexpire - pr->ndpr_lastupdate) { oprl->prefix[i].expire = pr->ndpr_lastupdate + pr->ndpr_vltime + (time_second - time_uptime); } else oprl->prefix[i].expire = maxexpire; } j = 0; LIST_FOREACH(pfr, &pr->ndpr_advrtrs, pfr_entry) { if (j < DRLSTSIZ) { #define RTRADDR oprl->prefix[i].advrtr[j] RTRADDR = pfr->router->rtaddr; in6_clearscope(&RTRADDR); #undef RTRADDR } j++; } oprl->prefix[i].advrtrs = j; oprl->prefix[i].origin = PR_ORIG_RA; i++; } break; case OSIOCGIFINFO_IN6: #define ND ndi->ndi /* XXX: old ndp(8) assumes a positive value for linkmtu. */ bzero(&ND, sizeof(ND)); ND.linkmtu = IN6_LINKMTU(ifp); ND.maxmtu = ND_IFINFO(ifp)->maxmtu; ND.basereachable = ND_IFINFO(ifp)->basereachable; ND.reachable = ND_IFINFO(ifp)->reachable; ND.retrans = ND_IFINFO(ifp)->retrans; ND.flags = ND_IFINFO(ifp)->flags; ND.recalctm = ND_IFINFO(ifp)->recalctm; ND.chlim = ND_IFINFO(ifp)->chlim; break; case SIOCGIFINFO_IN6: ND = *ND_IFINFO(ifp); break; case SIOCSIFINFO_IN6: /* * used to change host variables from userland. * intented for a use on router to reflect RA configurations. */ /* 0 means 'unspecified' */ if (ND.linkmtu != 0) { if (ND.linkmtu < IPV6_MMTU || ND.linkmtu > IN6_LINKMTU(ifp)) { error = EINVAL; break; } ND_IFINFO(ifp)->linkmtu = ND.linkmtu; } if (ND.basereachable != 0) { int obasereachable = ND_IFINFO(ifp)->basereachable; ND_IFINFO(ifp)->basereachable = ND.basereachable; if (ND.basereachable != obasereachable) ND_IFINFO(ifp)->reachable = ND_COMPUTE_RTIME(ND.basereachable); } if (ND.retrans != 0) ND_IFINFO(ifp)->retrans = ND.retrans; if (ND.chlim != 0) ND_IFINFO(ifp)->chlim = ND.chlim; /* FALLTHROUGH */ case SIOCSIFINFO_FLAGS: { struct ifaddr *ifa; struct in6_ifaddr *ia; if ((ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) && !(ND.flags & ND6_IFF_IFDISABLED)) { /* ifdisabled 1->0 transision */ /* * If the interface is marked as ND6_IFF_IFDISABLED and * has an link-local address with IN6_IFF_DUPLICATED, * do not clear ND6_IFF_IFDISABLED. * See RFC 4862, Section 5.4.5. */ int duplicated_linklocal = 0; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; if ((ia->ia6_flags & IN6_IFF_DUPLICATED) && IN6_IS_ADDR_LINKLOCAL(IA6_IN6(ia))) { duplicated_linklocal = 1; break; } } IF_ADDR_RUNLOCK(ifp); if (duplicated_linklocal) { ND.flags |= ND6_IFF_IFDISABLED; log(LOG_ERR, "Cannot enable an interface" " with a link-local address marked" " duplicate.\n"); } else { ND_IFINFO(ifp)->flags &= ~ND6_IFF_IFDISABLED; if (ifp->if_flags & IFF_UP) in6_if_up(ifp); } } else if (!(ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) && (ND.flags & ND6_IFF_IFDISABLED)) { /* ifdisabled 0->1 transision */ /* Mark all IPv6 address as tentative. */ ND_IFINFO(ifp)->flags |= ND6_IFF_IFDISABLED; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; ia->ia6_flags |= IN6_IFF_TENTATIVE; } IF_ADDR_RUNLOCK(ifp); } if (ND.flags & ND6_IFF_AUTO_LINKLOCAL) { if (!(ND_IFINFO(ifp)->flags & ND6_IFF_AUTO_LINKLOCAL)) { /* auto_linklocal 0->1 transision */ /* If no link-local address on ifp, configure */ ND_IFINFO(ifp)->flags |= ND6_IFF_AUTO_LINKLOCAL; in6_ifattach(ifp, NULL); } else if (!(ND.flags & ND6_IFF_IFDISABLED) && ifp->if_flags & IFF_UP) { /* * When the IF already has * ND6_IFF_AUTO_LINKLOCAL, no link-local * address is assigned, and IFF_UP, try to * assign one. */ int haslinklocal = 0; IF_ADDR_RLOCK(ifp); TAILQ_FOREACH(ifa, &ifp->if_addrhead, ifa_link) { if (ifa->ifa_addr->sa_family != AF_INET6) continue; ia = (struct in6_ifaddr *)ifa; if (IN6_IS_ADDR_LINKLOCAL(IA6_IN6(ia))) { haslinklocal = 1; break; } } IF_ADDR_RUNLOCK(ifp); if (!haslinklocal) in6_ifattach(ifp, NULL); } } } ND_IFINFO(ifp)->flags = ND.flags; break; #undef ND case SIOCSNDFLUSH_IN6: /* XXX: the ioctl name is confusing... */ /* sync kernel routing table with the default router list */ defrouter_reset(); defrouter_select(); break; case SIOCSPFXFLUSH_IN6: { /* flush all the prefix advertised by routers */ struct nd_prefix *pr, *next; LIST_FOREACH_SAFE(pr, &V_nd_prefix, ndpr_entry, next) { struct in6_ifaddr *ia, *ia_next; if (IN6_IS_ADDR_LINKLOCAL(&pr->ndpr_prefix.sin6_addr)) continue; /* XXX */ /* do we really have to remove addresses as well? */ /* XXXRW: in6_ifaddrhead locking. */ TAILQ_FOREACH_SAFE(ia, &V_in6_ifaddrhead, ia_link, ia_next) { if ((ia->ia6_flags & IN6_IFF_AUTOCONF) == 0) continue; if (ia->ia6_ndpr == pr) in6_purgeaddr(&ia->ia_ifa); } prelist_remove(pr); } break; } case SIOCSRTRFLUSH_IN6: { /* flush all the default routers */ struct nd_defrouter *dr, *next; defrouter_reset(); TAILQ_FOREACH_SAFE(dr, &V_nd_defrouter, dr_entry, next) { defrtrlist_del(dr); } defrouter_select(); break; } case SIOCGNBRINFO_IN6: { struct llentry *ln; struct in6_addr nb_addr = nbi->addr; /* make local for safety */ if ((error = in6_setscope(&nb_addr, ifp, NULL)) != 0) return (error); IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(&nb_addr, 0, ifp); IF_AFDATA_RUNLOCK(ifp); if (ln == NULL) { error = EINVAL; break; } nbi->state = ln->ln_state; nbi->asked = ln->la_asked; nbi->isrouter = ln->ln_router; if (ln->la_expire == 0) nbi->expire = 0; else nbi->expire = ln->la_expire + (time_second - time_uptime); LLE_RUNLOCK(ln); break; } case SIOCGDEFIFACE_IN6: /* XXX: should be implemented as a sysctl? */ ndif->ifindex = V_nd6_defifindex; break; case SIOCSDEFIFACE_IN6: /* XXX: should be implemented as a sysctl? */ return (nd6_setdefaultiface(ndif->ifindex)); } return (error); } /* * Create neighbor cache entry and cache link-layer address, * on reception of inbound ND6 packets. (RS/RA/NS/redirect) * * type - ICMP6 type * code - type dependent information * * XXXXX * The caller of this function already acquired the ndp * cache table lock because the cache entry is returned. */ struct llentry * nd6_cache_lladdr(struct ifnet *ifp, struct in6_addr *from, char *lladdr, int lladdrlen, int type, int code) { struct llentry *ln = NULL; int is_newentry; int do_update; int olladdr; int llchange; int flags; int newstate = 0; uint16_t router = 0; struct sockaddr_in6 sin6; struct mbuf *chain = NULL; int static_route = 0; IF_AFDATA_UNLOCK_ASSERT(ifp); KASSERT(ifp != NULL, ("%s: ifp == NULL", __func__)); KASSERT(from != NULL, ("%s: from == NULL", __func__)); /* nothing must be updated for unspecified address */ if (IN6_IS_ADDR_UNSPECIFIED(from)) return NULL; /* * Validation about ifp->if_addrlen and lladdrlen must be done in * the caller. * * XXX If the link does not have link-layer adderss, what should * we do? (ifp->if_addrlen == 0) * Spec says nothing in sections for RA, RS and NA. There's small * description on it in NS section (RFC 2461 7.2.3). */ flags = lladdr ? ND6_EXCLUSIVE : 0; IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(from, flags, ifp); IF_AFDATA_RUNLOCK(ifp); if (ln == NULL) { flags |= ND6_EXCLUSIVE; IF_AFDATA_LOCK(ifp); ln = nd6_lookup(from, flags | ND6_CREATE, ifp); IF_AFDATA_UNLOCK(ifp); is_newentry = 1; } else { /* do nothing if static ndp is set */ if (ln->la_flags & LLE_STATIC) { static_route = 1; goto done; } is_newentry = 0; } if (ln == NULL) return (NULL); olladdr = (ln->la_flags & LLE_VALID) ? 1 : 0; if (olladdr && lladdr) { llchange = bcmp(lladdr, &ln->ll_addr, ifp->if_addrlen); } else llchange = 0; /* * newentry olladdr lladdr llchange (*=record) * 0 n n -- (1) * 0 y n -- (2) * 0 n y -- (3) * STALE * 0 y y n (4) * * 0 y y y (5) * STALE * 1 -- n -- (6) NOSTATE(= PASSIVE) * 1 -- y -- (7) * STALE */ if (lladdr) { /* (3-5) and (7) */ /* * Record source link-layer address * XXX is it dependent to ifp->if_type? */ bcopy(lladdr, &ln->ll_addr, ifp->if_addrlen); ln->la_flags |= LLE_VALID; EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_RESOLVED); } if (!is_newentry) { if ((!olladdr && lladdr != NULL) || /* (3) */ (olladdr && lladdr != NULL && llchange)) { /* (5) */ do_update = 1; newstate = ND6_LLINFO_STALE; } else /* (1-2,4) */ do_update = 0; } else { do_update = 1; if (lladdr == NULL) /* (6) */ newstate = ND6_LLINFO_NOSTATE; else /* (7) */ newstate = ND6_LLINFO_STALE; } if (do_update) { /* * Update the state of the neighbor cache. */ ln->ln_state = newstate; if (ln->ln_state == ND6_LLINFO_STALE) { /* * XXX: since nd6_output() below will cause * state tansition to DELAY and reset the timer, * we must set the timer now, although it is actually * meaningless. */ nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); if (ln->la_hold) { struct mbuf *m_hold, *m_hold_next; /* * reset the la_hold in advance, to explicitly * prevent a la_hold lookup in nd6_output() * (wouldn't happen, though...) */ for (m_hold = ln->la_hold, ln->la_hold = NULL; m_hold; m_hold = m_hold_next) { m_hold_next = m_hold->m_nextpkt; m_hold->m_nextpkt = NULL; /* * we assume ifp is not a p2p here, so * just set the 2nd argument as the * 1st one. */ nd6_output_lle(ifp, ifp, m_hold, L3_ADDR_SIN6(ln), NULL, ln, &chain); } /* * If we have mbufs in the chain we need to do * deferred transmit. Copy the address from the * llentry before dropping the lock down below. */ if (chain != NULL) memcpy(&sin6, L3_ADDR_SIN6(ln), sizeof(sin6)); } } else if (ln->ln_state == ND6_LLINFO_INCOMPLETE) { /* probe right away */ nd6_llinfo_settimer_locked((void *)ln, 0); } } /* * ICMP6 type dependent behavior. * * NS: clear IsRouter if new entry * RS: clear IsRouter * RA: set IsRouter if there's lladdr * redir: clear IsRouter if new entry * * RA case, (1): * The spec says that we must set IsRouter in the following cases: * - If lladdr exist, set IsRouter. This means (1-5). * - If it is old entry (!newentry), set IsRouter. This means (7). * So, based on the spec, in (1-5) and (7) cases we must set IsRouter. * A quetion arises for (1) case. (1) case has no lladdr in the * neighbor cache, this is similar to (6). * This case is rare but we figured that we MUST NOT set IsRouter. * * newentry olladdr lladdr llchange NS RS RA redir * D R * 0 n n -- (1) c ? s * 0 y n -- (2) c s s * 0 n y -- (3) c s s * 0 y y n (4) c s s * 0 y y y (5) c s s * 1 -- n -- (6) c c c s * 1 -- y -- (7) c c s c s * * (c=clear s=set) */ switch (type & 0xff) { case ND_NEIGHBOR_SOLICIT: /* * New entry must have is_router flag cleared. */ if (is_newentry) /* (6-7) */ ln->ln_router = 0; break; case ND_REDIRECT: /* * If the icmp is a redirect to a better router, always set the * is_router flag. Otherwise, if the entry is newly created, * clear the flag. [RFC 2461, sec 8.3] */ if (code == ND_REDIRECT_ROUTER) ln->ln_router = 1; else if (is_newentry) /* (6-7) */ ln->ln_router = 0; break; case ND_ROUTER_SOLICIT: /* * is_router flag must always be cleared. */ ln->ln_router = 0; break; case ND_ROUTER_ADVERT: /* * Mark an entry with lladdr as a router. */ if ((!is_newentry && (olladdr || lladdr)) || /* (2-5) */ (is_newentry && lladdr)) { /* (7) */ ln->ln_router = 1; } break; } if (ln != NULL) { static_route = (ln->la_flags & LLE_STATIC); router = ln->ln_router; if (flags & ND6_EXCLUSIVE) LLE_WUNLOCK(ln); else LLE_RUNLOCK(ln); if (static_route) ln = NULL; } if (chain) nd6_output_flush(ifp, ifp, chain, &sin6); /* * When the link-layer address of a router changes, select the * best router again. In particular, when the neighbor entry is newly * created, it might affect the selection policy. * Question: can we restrict the first condition to the "is_newentry" * case? * XXX: when we hear an RA from a new router with the link-layer * address option, defrouter_select() is called twice, since * defrtrlist_update called the function as well. However, I believe * we can compromise the overhead, since it only happens the first * time. * XXX: although defrouter_select() should not have a bad effect * for those are not autoconfigured hosts, we explicitly avoid such * cases for safety. */ if (do_update && router && ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV) { /* * guaranteed recursion */ defrouter_select(); } return (ln); done: if (ln != NULL) { if (flags & ND6_EXCLUSIVE) LLE_WUNLOCK(ln); else LLE_RUNLOCK(ln); if (static_route) ln = NULL; } return (ln); } static void nd6_slowtimo(void *arg) { CURVNET_SET((struct vnet *) arg); struct nd_ifinfo *nd6if; struct ifnet *ifp; callout_reset(&V_nd6_slowtimo_ch, ND6_SLOWTIMER_INTERVAL * hz, nd6_slowtimo, curvnet); IFNET_RLOCK_NOSLEEP(); TAILQ_FOREACH(ifp, &V_ifnet, if_link) { if (ifp->if_afdata[AF_INET6] == NULL) continue; nd6if = ND_IFINFO(ifp); if (nd6if->basereachable && /* already initialized */ (nd6if->recalctm -= ND6_SLOWTIMER_INTERVAL) <= 0) { /* * Since reachable time rarely changes by router * advertisements, we SHOULD insure that a new random * value gets recomputed at least once every few hours. * (RFC 2461, 6.3.4) */ nd6if->recalctm = V_nd6_recalc_reachtm_interval; nd6if->reachable = ND_COMPUTE_RTIME(nd6if->basereachable); } } IFNET_RUNLOCK_NOSLEEP(); CURVNET_RESTORE(); } /* * IPv6 packet output - light version. * Checks if destination LLE exists and is in proper state * (e.g no modification required). If not true, fall back to * "heavy" version. */ int nd6_output(struct ifnet *ifp, struct ifnet *origifp, struct mbuf *m, struct sockaddr_in6 *dst, struct rtentry *rt0) { struct llentry *ln = NULL; int error = 0; /* discard the packet if IPv6 operation is disabled on the interface */ if ((ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED)) { m_freem(m); return (ENETDOWN); /* better error? */ } if (IN6_IS_ADDR_MULTICAST(&dst->sin6_addr)) goto sendpkt; if (nd6_need_cache(ifp) == 0) goto sendpkt; IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(&dst->sin6_addr, 0, ifp); IF_AFDATA_RUNLOCK(ifp); /* * Perform fast path for the following cases: * 1) lle state is REACHABLE * 2) lle state is DELAY (NS message sentNS message sent) * * Every other case involves lle modification, so we handle * them separately. */ if (ln == NULL || (ln->ln_state != ND6_LLINFO_REACHABLE && ln->ln_state != ND6_LLINFO_DELAY)) { /* Fall back to slow processing path */ if (ln != NULL) LLE_RUNLOCK(ln); return (nd6_output_lle(ifp, origifp, m, dst, rt0, NULL, NULL)); } sendpkt: if (ln != NULL) LLE_RUNLOCK(ln); #ifdef MAC mac_netinet6_nd6_send(ifp, m); #endif /* * If called from nd6_ns_output() (NS), nd6_na_output() (NA), * icmp6_redirect_output() (REDIRECT) or from rip6_output() (RS, RA * as handled by rtsol and rtadvd), mbufs will be tagged for SeND * to be diverted to user space. When re-injected into the kernel, * send_output() will directly dispatch them to the outgoing interface. */ if (send_sendso_input_hook != NULL) { struct m_tag *mtag; struct ip6_hdr *ip6; int ip6len; mtag = m_tag_find(m, PACKET_TAG_ND_OUTGOING, NULL); if (mtag != NULL) { ip6 = mtod(m, struct ip6_hdr *); ip6len = sizeof(struct ip6_hdr) + ntohs(ip6->ip6_plen); /* Use the SEND socket */ error = send_sendso_input_hook(m, ifp, SND_OUT, ip6len); /* -1 == no app on SEND socket */ if (error == 0 || error != -1) return (error); } } m_clrprotoflags(m); /* Avoid confusing lower layers. */ IP_PROBE(send, NULL, NULL, mtod(m, struct ip6_hdr *), ifp, NULL, mtod(m, struct ip6_hdr *)); if ((ifp->if_flags & IFF_LOOPBACK) == 0) origifp = ifp; error = (*ifp->if_output)(origifp, m, (struct sockaddr *)dst, NULL); return (error); } /* * Output IPv6 packet - heavy version. * Function assume that either * 1) destination LLE does not exist, is invalid or stale, so * ND6_EXCLUSIVE lock needs to be acquired * 2) destination lle is provided (with ND6_EXCLUSIVE lock), * in that case packets are queued in &chain. * */ int nd6_output_lle(struct ifnet *ifp, struct ifnet *origifp, struct mbuf *m, struct sockaddr_in6 *dst, struct rtentry *rt0, struct llentry *lle, struct mbuf **chain) { struct m_tag *mtag; struct ip6_hdr *ip6; int error = 0; int flags = 0; int has_lle = 0; int ip6len; #ifdef INVARIANTS if (lle != NULL) { LLE_WLOCK_ASSERT(lle); KASSERT(chain != NULL, (" lle locked but no mbuf chain pointer passed")); } #endif KASSERT(m != NULL, ("NULL mbuf, nothing to send")); /* discard the packet if IPv6 operation is disabled on the interface */ if ((ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED)) { m_freem(m); return (ENETDOWN); /* better error? */ } if (lle != NULL) has_lle = 1; if (IN6_IS_ADDR_MULTICAST(&dst->sin6_addr)) goto sendpkt; if (nd6_need_cache(ifp) == 0) goto sendpkt; /* * Address resolution or Neighbor Unreachability Detection * for the next hop. * At this point, the destination of the packet must be a unicast * or an anycast address(i.e. not a multicast). */ if (lle == NULL) { IF_AFDATA_RLOCK(ifp); lle = nd6_lookup(&dst->sin6_addr, ND6_EXCLUSIVE, ifp); IF_AFDATA_RUNLOCK(ifp); if ((lle == NULL) && nd6_is_addr_neighbor(dst, ifp)) { /* * Since nd6_is_addr_neighbor() internally calls nd6_lookup(), * the condition below is not very efficient. But we believe * it is tolerable, because this should be a rare case. */ flags = ND6_CREATE | ND6_EXCLUSIVE; IF_AFDATA_LOCK(ifp); lle = nd6_lookup(&dst->sin6_addr, flags, ifp); IF_AFDATA_UNLOCK(ifp); } } if (lle == NULL) { if ((ifp->if_flags & IFF_POINTOPOINT) == 0 && !(ND_IFINFO(ifp)->flags & ND6_IFF_PERFORMNUD)) { char ip6buf[INET6_ADDRSTRLEN]; log(LOG_DEBUG, "nd6_output: can't allocate llinfo for %s " "(ln=%p)\n", ip6_sprintf(ip6buf, &dst->sin6_addr), lle); m_freem(m); return (ENOBUFS); } goto sendpkt; /* send anyway */ } LLE_WLOCK_ASSERT(lle); /* We don't have to do link-layer address resolution on a p2p link. */ if ((ifp->if_flags & IFF_POINTOPOINT) != 0 && lle->ln_state < ND6_LLINFO_REACHABLE) { lle->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(lle, (long)V_nd6_gctimer * hz); } /* * The first time we send a packet to a neighbor whose entry is * STALE, we have to change the state to DELAY and a sets a timer to * expire in DELAY_FIRST_PROBE_TIME seconds to ensure do * neighbor unreachability detection on expiration. * (RFC 2461 7.3.3) */ if (lle->ln_state == ND6_LLINFO_STALE) { lle->la_asked = 0; lle->ln_state = ND6_LLINFO_DELAY; nd6_llinfo_settimer_locked(lle, (long)V_nd6_delay * hz); } /* * If the neighbor cache entry has a state other than INCOMPLETE * (i.e. its link-layer address is already resolved), just * send the packet. */ if (lle->ln_state > ND6_LLINFO_INCOMPLETE) goto sendpkt; /* * There is a neighbor cache entry, but no ethernet address * response yet. Append this latest packet to the end of the * packet queue in the mbuf, unless the number of the packet * does not exceed nd6_maxqueuelen. When it exceeds nd6_maxqueuelen, * the oldest packet in the queue will be removed. */ if (lle->ln_state == ND6_LLINFO_NOSTATE) lle->ln_state = ND6_LLINFO_INCOMPLETE; if (lle->la_hold != NULL) { struct mbuf *m_hold; int i; i = 0; for (m_hold = lle->la_hold; m_hold; m_hold = m_hold->m_nextpkt){ i++; if (m_hold->m_nextpkt == NULL) { m_hold->m_nextpkt = m; break; } } while (i >= V_nd6_maxqueuelen) { m_hold = lle->la_hold; lle->la_hold = lle->la_hold->m_nextpkt; m_freem(m_hold); i--; } } else { lle->la_hold = m; } /* * If there has been no NS for the neighbor after entering the * INCOMPLETE state, send the first solicitation. */ if (!ND6_LLINFO_PERMANENT(lle) && lle->la_asked == 0) { lle->la_asked++; nd6_llinfo_settimer_locked(lle, (long)ND_IFINFO(ifp)->retrans * hz / 1000); LLE_WUNLOCK(lle); nd6_ns_output(ifp, NULL, &dst->sin6_addr, lle, 0); if (has_lle != 0) LLE_WLOCK(lle); } else if (has_lle == 0) { /* * We did the lookup (no lle arg) so we * need to do the unlock here. */ LLE_WUNLOCK(lle); } return (0); sendpkt: /* * ln is valid and the caller did not pass in * an llentry */ if (lle != NULL && has_lle == 0) LLE_WUNLOCK(lle); #ifdef MAC mac_netinet6_nd6_send(ifp, m); #endif /* * If called from nd6_ns_output() (NS), nd6_na_output() (NA), * icmp6_redirect_output() (REDIRECT) or from rip6_output() (RS, RA * as handled by rtsol and rtadvd), mbufs will be tagged for SeND * to be diverted to user space. When re-injected into the kernel, * send_output() will directly dispatch them to the outgoing interface. */ if (send_sendso_input_hook != NULL) { mtag = m_tag_find(m, PACKET_TAG_ND_OUTGOING, NULL); if (mtag != NULL) { ip6 = mtod(m, struct ip6_hdr *); ip6len = sizeof(struct ip6_hdr) + ntohs(ip6->ip6_plen); /* Use the SEND socket */ error = send_sendso_input_hook(m, ifp, SND_OUT, ip6len); /* -1 == no app on SEND socket */ if (error == 0 || error != -1) return (error); } } /* * We were passed in a pointer to an lle with the lock held * this means that we can't call if_output as we will * recurse on the lle lock - so what we do is we create * a list of mbufs to send and transmit them in the caller * after the lock is dropped */ if (has_lle != 0) { if (*chain == NULL) *chain = m; else { struct mbuf *mb; /* * append mbuf to end of deferred chain */ mb = *chain; while (mb->m_nextpkt != NULL) mb = mb->m_nextpkt; mb->m_nextpkt = m; } return (error); } m_clrprotoflags(m); /* Avoid confusing lower layers. */ IP_PROBE(send, NULL, NULL, mtod(m, struct ip6_hdr *), ifp, NULL, mtod(m, struct ip6_hdr *)); if ((ifp->if_flags & IFF_LOOPBACK) == 0) origifp = ifp; error = (*ifp->if_output)(origifp, m, (struct sockaddr *)dst, NULL); return (error); } int nd6_output_flush(struct ifnet *ifp, struct ifnet *origifp, struct mbuf *chain, struct sockaddr_in6 *dst) { struct mbuf *m, *m_head; struct ifnet *outifp; int error = 0; m_head = chain; if ((ifp->if_flags & IFF_LOOPBACK) != 0) outifp = origifp; else outifp = ifp; while (m_head) { m = m_head; m_head = m_head->m_nextpkt; error = (*ifp->if_output)(ifp, m, (struct sockaddr *)dst, NULL); } /* * XXX * note that intermediate errors are blindly ignored - but this is * the same convention as used with nd6_output when called by * nd6_cache_lladdr */ return (error); } int nd6_need_cache(struct ifnet *ifp) { /* * XXX: we currently do not make neighbor cache on any interface * other than ARCnet, Ethernet, FDDI and GIF. * * RFC2893 says: * - unidirectional tunnels needs no ND */ switch (ifp->if_type) { case IFT_ARCNET: case IFT_ETHER: case IFT_FDDI: case IFT_IEEE1394: #ifdef IFT_L2VLAN case IFT_L2VLAN: #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: #endif case IFT_INFINIBAND: case IFT_BRIDGE: case IFT_PROPVIRTUAL: return (1); default: return (0); } } /* * Add pernament ND6 link-layer record for given * interface address. * * Very similar to IPv4 arp_ifinit(), but: * 1) IPv6 DAD is performed in different place * 2) It is called by IPv6 protocol stack in contrast to * arp_ifinit() which is typically called in SIOCSIFADDR * driver ioctl handler. * */ int nd6_add_ifa_lle(struct in6_ifaddr *ia) { struct ifnet *ifp; struct llentry *ln; ifp = ia->ia_ifa.ifa_ifp; if (nd6_need_cache(ifp) == 0) return (0); IF_AFDATA_LOCK(ifp); ia->ia_ifa.ifa_rtrequest = nd6_rtrequest; ln = lla_lookup(LLTABLE6(ifp), (LLE_CREATE | LLE_IFADDR | LLE_EXCLUSIVE), (struct sockaddr *)&ia->ia_addr); IF_AFDATA_UNLOCK(ifp); if (ln != NULL) { ln->la_expire = 0; /* for IPv6 this means permanent */ ln->ln_state = ND6_LLINFO_REACHABLE; LLE_WUNLOCK(ln); in6_newaddrmsg(ia, RTM_ADD); return (0); } return (ENOBUFS); } /* * Removes ALL lle records for interface address prefix. * XXXME: That's probably not we really want to do, we need * to remove address record only and keep other records * until we determine if given prefix is really going * to be removed. */ void nd6_rem_ifa_lle(struct in6_ifaddr *ia) { struct sockaddr_in6 mask, addr; struct ifnet *ifp; in6_newaddrmsg(ia, RTM_DELETE); ifp = ia->ia_ifa.ifa_ifp; memcpy(&addr, &ia->ia_addr, sizeof(ia->ia_addr)); memcpy(&mask, &ia->ia_prefixmask, sizeof(ia->ia_prefixmask)); lltable_prefix_free(AF_INET6, (struct sockaddr *)&addr, (struct sockaddr *)&mask, LLE_STATIC); } /* * the callers of this function need to be re-worked to drop * the lle lock, drop here for now */ int nd6_storelladdr(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, u_char *desten, uint32_t *pflags) { struct llentry *ln; if (pflags != NULL) *pflags = 0; IF_AFDATA_UNLOCK_ASSERT(ifp); if (m != NULL && m->m_flags & M_MCAST) { int i; switch (ifp->if_type) { case IFT_ETHER: case IFT_FDDI: #ifdef IFT_L2VLAN case IFT_L2VLAN: #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: #endif case IFT_BRIDGE: case IFT_ISO88025: ETHER_MAP_IPV6_MULTICAST(&SIN6(dst)->sin6_addr, desten); return (0); case IFT_IEEE1394: /* * netbsd can use if_broadcastaddr, but we don't do so * to reduce # of ifdef. */ for (i = 0; i < ifp->if_addrlen; i++) desten[i] = ~0; return (0); case IFT_ARCNET: *desten = 0; return (0); default: m_freem(m); return (EAFNOSUPPORT); } } /* * the entry should have been created in nd6_store_lladdr */ IF_AFDATA_RLOCK(ifp); ln = lla_lookup(LLTABLE6(ifp), 0, dst); IF_AFDATA_RUNLOCK(ifp); if ((ln == NULL) || !(ln->la_flags & LLE_VALID)) { if (ln != NULL) LLE_RUNLOCK(ln); /* this could happen, if we could not allocate memory */ m_freem(m); return (1); } bcopy(&ln->ll_addr, desten, ifp->if_addrlen); if (pflags != NULL) *pflags = ln->la_flags; LLE_RUNLOCK(ln); /* * A *small* use after free race exists here */ return (0); } static void clear_llinfo_pqueue(struct llentry *ln) { struct mbuf *m_hold, *m_hold_next; for (m_hold = ln->la_hold; m_hold; m_hold = m_hold_next) { m_hold_next = m_hold->m_nextpkt; m_freem(m_hold); } ln->la_hold = NULL; return; } static int nd6_sysctl_drlist(SYSCTL_HANDLER_ARGS); static int nd6_sysctl_prlist(SYSCTL_HANDLER_ARGS); #ifdef SYSCTL_DECL SYSCTL_DECL(_net_inet6_icmp6); #endif SYSCTL_NODE(_net_inet6_icmp6, ICMPV6CTL_ND6_DRLIST, nd6_drlist, CTLFLAG_RD, nd6_sysctl_drlist, ""); SYSCTL_NODE(_net_inet6_icmp6, ICMPV6CTL_ND6_PRLIST, nd6_prlist, CTLFLAG_RD, nd6_sysctl_prlist, ""); SYSCTL_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_MAXQLEN, nd6_maxqueuelen, CTLFLAG_VNET | CTLFLAG_RW, &VNET_NAME(nd6_maxqueuelen), 1, ""); SYSCTL_INT(_net_inet6_icmp6, OID_AUTO, nd6_gctimer, CTLFLAG_VNET | CTLFLAG_RW, &VNET_NAME(nd6_gctimer), (60 * 60 * 24), ""); static int nd6_sysctl_drlist(SYSCTL_HANDLER_ARGS) { struct in6_defrouter d; struct nd_defrouter *dr; int error; if (req->newptr) return (EPERM); bzero(&d, sizeof(d)); d.rtaddr.sin6_family = AF_INET6; d.rtaddr.sin6_len = sizeof(d.rtaddr); /* * XXX locking */ TAILQ_FOREACH(dr, &V_nd_defrouter, dr_entry) { d.rtaddr.sin6_addr = dr->rtaddr; error = sa6_recoverscope(&d.rtaddr); if (error != 0) return (error); d.flags = dr->flags; d.rtlifetime = dr->rtlifetime; d.expire = dr->expire + (time_second - time_uptime); d.if_index = dr->ifp->if_index; error = SYSCTL_OUT(req, &d, sizeof(d)); if (error != 0) return (error); } return (0); } static int nd6_sysctl_prlist(SYSCTL_HANDLER_ARGS) { struct in6_prefix p; struct sockaddr_in6 s6; struct nd_prefix *pr; struct nd_pfxrouter *pfr; time_t maxexpire; int error; char ip6buf[INET6_ADDRSTRLEN]; if (req->newptr) return (EPERM); bzero(&p, sizeof(p)); p.origin = PR_ORIG_RA; bzero(&s6, sizeof(s6)); s6.sin6_family = AF_INET6; s6.sin6_len = sizeof(s6); /* * XXX locking */ LIST_FOREACH(pr, &V_nd_prefix, ndpr_entry) { p.prefix = pr->ndpr_prefix; if (sa6_recoverscope(&p.prefix)) { log(LOG_ERR, "scope error in prefix list (%s)\n", ip6_sprintf(ip6buf, &p.prefix.sin6_addr)); /* XXX: press on... */ } p.raflags = pr->ndpr_raf; p.prefixlen = pr->ndpr_plen; p.vltime = pr->ndpr_vltime; p.pltime = pr->ndpr_pltime; p.if_index = pr->ndpr_ifp->if_index; if (pr->ndpr_vltime == ND6_INFINITE_LIFETIME) p.expire = 0; else { /* XXX: we assume time_t is signed. */ maxexpire = (-1) & ~((time_t)1 << ((sizeof(maxexpire) * 8) - 1)); if (pr->ndpr_vltime < maxexpire - pr->ndpr_lastupdate) p.expire = pr->ndpr_lastupdate + pr->ndpr_vltime + (time_second - time_uptime); else p.expire = maxexpire; } p.refcnt = pr->ndpr_refcnt; p.flags = pr->ndpr_stateflags; p.advrtrs = 0; LIST_FOREACH(pfr, &pr->ndpr_advrtrs, pfr_entry) p.advrtrs++; error = SYSCTL_OUT(req, &p, sizeof(p)); if (error != 0) return (error); LIST_FOREACH(pfr, &pr->ndpr_advrtrs, pfr_entry) { s6.sin6_addr = pfr->router->rtaddr; if (sa6_recoverscope(&s6)) log(LOG_ERR, "scope error in prefix list (%s)\n", ip6_sprintf(ip6buf, &pfr->router->rtaddr)); error = SYSCTL_OUT(req, &s6, sizeof(s6)); if (error != 0) return (error); } } return (0); } Index: projects/building-blocks/sys/netinet6/nd6.h =================================================================== --- projects/building-blocks/sys/netinet6/nd6.h (revision 275663) +++ projects/building-blocks/sys/netinet6/nd6.h (revision 275664) @@ -1,453 +1,454 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: nd6.h,v 1.76 2001/12/18 02:10:31 itojun Exp $ * $FreeBSD$ */ #ifndef _NETINET6_ND6_H_ #define _NETINET6_ND6_H_ /* see net/route.h, or net/if_inarp.h */ #ifndef RTF_ANNOUNCE #define RTF_ANNOUNCE RTF_PROTO2 #endif #include #include struct llentry; #define ND6_LLINFO_NOSTATE -2 /* * We don't need the WAITDELETE state any more, but we keep the definition * in a comment line instead of removing it. This is necessary to avoid * unintentionally reusing the value for another purpose, which might * affect backward compatibility with old applications. * (20000711 jinmei@kame.net) */ /* #define ND6_LLINFO_WAITDELETE -1 */ #define ND6_LLINFO_INCOMPLETE 0 #define ND6_LLINFO_REACHABLE 1 #define ND6_LLINFO_STALE 2 #define ND6_LLINFO_DELAY 3 #define ND6_LLINFO_PROBE 4 #define ND6_IS_LLINFO_PROBREACH(n) ((n)->ln_state > ND6_LLINFO_INCOMPLETE) #define ND6_LLINFO_PERMANENT(n) (((n)->la_expire == 0) && ((n)->ln_state > ND6_LLINFO_INCOMPLETE)) struct nd_ifinfo { u_int32_t linkmtu; /* LinkMTU */ u_int32_t maxmtu; /* Upper bound of LinkMTU */ u_int32_t basereachable; /* BaseReachableTime */ u_int32_t reachable; /* Reachable Time */ u_int32_t retrans; /* Retrans Timer */ u_int32_t flags; /* Flags */ int recalctm; /* BaseReacable re-calculation timer */ u_int8_t chlim; /* CurHopLimit */ u_int8_t initialized; /* Flag to see the entry is initialized */ /* the following 3 members are for privacy extension for addrconf */ u_int8_t randomseed0[8]; /* upper 64 bits of MD5 digest */ u_int8_t randomseed1[8]; /* lower 64 bits (usually the EUI64 IFID) */ u_int8_t randomid[8]; /* current random ID */ }; #define ND6_IFF_PERFORMNUD 0x1 #define ND6_IFF_ACCEPT_RTADV 0x2 #define ND6_IFF_PREFER_SOURCE 0x4 /* Not used in FreeBSD. */ #define ND6_IFF_IFDISABLED 0x8 /* IPv6 operation is disabled due to * DAD failure. (XXX: not ND-specific) */ #define ND6_IFF_DONT_SET_IFROUTE 0x10 #define ND6_IFF_AUTO_LINKLOCAL 0x20 #define ND6_IFF_NO_RADR 0x40 #define ND6_IFF_NO_PREFER_IFACE 0x80 /* XXX: not related to ND. */ #define ND6_CREATE LLE_CREATE #define ND6_EXCLUSIVE LLE_EXCLUSIVE #ifdef _KERNEL #define ND_IFINFO(ifp) \ (((struct in6_ifextra *)(ifp)->if_afdata[AF_INET6])->nd_ifinfo) #define IN6_LINKMTU(ifp) \ ((ND_IFINFO(ifp)->linkmtu && ND_IFINFO(ifp)->linkmtu < (ifp)->if_mtu) \ ? ND_IFINFO(ifp)->linkmtu \ : ((ND_IFINFO(ifp)->maxmtu && ND_IFINFO(ifp)->maxmtu < (ifp)->if_mtu) \ ? ND_IFINFO(ifp)->maxmtu : (ifp)->if_mtu)) #endif struct in6_nbrinfo { char ifname[IFNAMSIZ]; /* if name, e.g. "en0" */ struct in6_addr addr; /* IPv6 address of the neighbor */ long asked; /* number of queries already sent for this addr */ int isrouter; /* if it acts as a router */ int state; /* reachability state */ int expire; /* lifetime for NDP state transition */ }; #define DRLSTSIZ 10 #define PRLSTSIZ 10 struct in6_drlist { char ifname[IFNAMSIZ]; struct { struct in6_addr rtaddr; u_char flags; u_short rtlifetime; u_long expire; u_short if_index; } defrouter[DRLSTSIZ]; }; struct in6_defrouter { struct sockaddr_in6 rtaddr; u_char flags; u_short rtlifetime; u_long expire; u_short if_index; }; #ifdef _KERNEL struct in6_oprlist { char ifname[IFNAMSIZ]; struct { struct in6_addr prefix; struct prf_ra raflags; u_char prefixlen; u_char origin; u_long vltime; u_long pltime; u_long expire; u_short if_index; u_short advrtrs; /* number of advertisement routers */ struct in6_addr advrtr[DRLSTSIZ]; /* XXX: explicit limit */ } prefix[PRLSTSIZ]; }; #endif struct in6_prlist { char ifname[IFNAMSIZ]; struct { struct in6_addr prefix; struct prf_ra raflags; u_char prefixlen; u_char origin; u_int32_t vltime; u_int32_t pltime; time_t expire; u_short if_index; u_short advrtrs; /* number of advertisement routers */ struct in6_addr advrtr[DRLSTSIZ]; /* XXX: explicit limit */ } prefix[PRLSTSIZ]; }; struct in6_prefix { struct sockaddr_in6 prefix; struct prf_ra raflags; u_char prefixlen; u_char origin; u_int32_t vltime; u_int32_t pltime; time_t expire; u_int32_t flags; int refcnt; u_short if_index; u_short advrtrs; /* number of advertisement routers */ /* struct sockaddr_in6 advrtr[] */ }; #ifdef _KERNEL struct in6_ondireq { char ifname[IFNAMSIZ]; struct { u_int32_t linkmtu; /* LinkMTU */ u_int32_t maxmtu; /* Upper bound of LinkMTU */ u_int32_t basereachable; /* BaseReachableTime */ u_int32_t reachable; /* Reachable Time */ u_int32_t retrans; /* Retrans Timer */ u_int32_t flags; /* Flags */ int recalctm; /* BaseReacable re-calculation timer */ u_int8_t chlim; /* CurHopLimit */ u_int8_t receivedra; } ndi; }; #endif struct in6_ndireq { char ifname[IFNAMSIZ]; struct nd_ifinfo ndi; }; struct in6_ndifreq { char ifname[IFNAMSIZ]; u_long ifindex; }; /* Prefix status */ #define NDPRF_ONLINK 0x1 #define NDPRF_DETACHED 0x2 /* protocol constants */ #define MAX_RTR_SOLICITATION_DELAY 1 /* 1sec */ #define RTR_SOLICITATION_INTERVAL 4 /* 4sec */ #define MAX_RTR_SOLICITATIONS 3 #define ND6_INFINITE_LIFETIME 0xffffffff #ifdef _KERNEL /* node constants */ #define MAX_REACHABLE_TIME 3600000 /* msec */ #define REACHABLE_TIME 30000 /* msec */ #define RETRANS_TIMER 1000 /* msec */ #define MIN_RANDOM_FACTOR 512 /* 1024 * 0.5 */ #define MAX_RANDOM_FACTOR 1536 /* 1024 * 1.5 */ #define DEF_TEMP_VALID_LIFETIME 604800 /* 1 week */ #define DEF_TEMP_PREFERRED_LIFETIME 86400 /* 1 day */ #define TEMPADDR_REGEN_ADVANCE 5 /* sec */ #define MAX_TEMP_DESYNC_FACTOR 600 /* 10 min */ #define ND_COMPUTE_RTIME(x) \ (((MIN_RANDOM_FACTOR * (x >> 10)) + (arc4random() & \ ((MAX_RANDOM_FACTOR - MIN_RANDOM_FACTOR) * (x >> 10)))) /1000) TAILQ_HEAD(nd_drhead, nd_defrouter); struct nd_defrouter { TAILQ_ENTRY(nd_defrouter) dr_entry; struct in6_addr rtaddr; u_char flags; /* flags on RA message */ u_short rtlifetime; u_long expire; struct ifnet *ifp; int installed; /* is installed into kernel routing table */ }; struct nd_prefixctl { struct ifnet *ndpr_ifp; /* prefix */ struct sockaddr_in6 ndpr_prefix; u_char ndpr_plen; u_int32_t ndpr_vltime; /* advertised valid lifetime */ u_int32_t ndpr_pltime; /* advertised preferred lifetime */ struct prf_ra ndpr_flags; }; struct nd_prefix { struct ifnet *ndpr_ifp; LIST_ENTRY(nd_prefix) ndpr_entry; struct sockaddr_in6 ndpr_prefix; /* prefix */ struct in6_addr ndpr_mask; /* netmask derived from the prefix */ u_int32_t ndpr_vltime; /* advertised valid lifetime */ u_int32_t ndpr_pltime; /* advertised preferred lifetime */ time_t ndpr_expire; /* expiration time of the prefix */ time_t ndpr_preferred; /* preferred time of the prefix */ time_t ndpr_lastupdate; /* reception time of last advertisement */ struct prf_ra ndpr_flags; u_int32_t ndpr_stateflags; /* actual state flags */ /* list of routers that advertise the prefix: */ LIST_HEAD(pr_rtrhead, nd_pfxrouter) ndpr_advrtrs; u_char ndpr_plen; int ndpr_refcnt; /* reference couter from addresses */ }; #define ndpr_raf ndpr_flags #define ndpr_raf_onlink ndpr_flags.onlink #define ndpr_raf_auto ndpr_flags.autonomous #define ndpr_raf_router ndpr_flags.router /* * Message format for use in obtaining information about prefixes * from inet6 sysctl function */ struct inet6_ndpr_msghdr { u_short inpm_msglen; /* to skip over non-understood messages */ u_char inpm_version; /* future binary compatibility */ u_char inpm_type; /* message type */ struct in6_addr inpm_prefix; u_long prm_vltim; u_long prm_pltime; u_long prm_expire; u_long prm_preferred; struct in6_prflags prm_flags; u_short prm_index; /* index for associated ifp */ u_char prm_plen; /* length of prefix in bits */ }; #define prm_raf_onlink prm_flags.prf_ra.onlink #define prm_raf_auto prm_flags.prf_ra.autonomous #define prm_statef_onlink prm_flags.prf_state.onlink #define prm_rrf_decrvalid prm_flags.prf_rr.decrvalid #define prm_rrf_decrprefd prm_flags.prf_rr.decrprefd struct nd_pfxrouter { LIST_ENTRY(nd_pfxrouter) pfr_entry; struct nd_defrouter *router; }; LIST_HEAD(nd_prhead, nd_prefix); /* nd6.c */ VNET_DECLARE(int, nd6_prune); VNET_DECLARE(int, nd6_delay); VNET_DECLARE(int, nd6_umaxtries); VNET_DECLARE(int, nd6_mmaxtries); VNET_DECLARE(int, nd6_useloopback); VNET_DECLARE(int, nd6_maxnudhint); VNET_DECLARE(int, nd6_gctimer); VNET_DECLARE(struct nd_drhead, nd_defrouter); VNET_DECLARE(struct nd_prhead, nd_prefix); VNET_DECLARE(int, nd6_debug); VNET_DECLARE(int, nd6_onlink_ns_rfc4861); #define V_nd6_prune VNET(nd6_prune) #define V_nd6_delay VNET(nd6_delay) #define V_nd6_umaxtries VNET(nd6_umaxtries) #define V_nd6_mmaxtries VNET(nd6_mmaxtries) #define V_nd6_useloopback VNET(nd6_useloopback) #define V_nd6_maxnudhint VNET(nd6_maxnudhint) #define V_nd6_gctimer VNET(nd6_gctimer) #define V_nd_defrouter VNET(nd_defrouter) #define V_nd_prefix VNET(nd_prefix) #define V_nd6_debug VNET(nd6_debug) #define V_nd6_onlink_ns_rfc4861 VNET(nd6_onlink_ns_rfc4861) #define nd6log(x) do { if (V_nd6_debug) log x; } while (/*CONSTCOND*/ 0) VNET_DECLARE(struct callout, nd6_timer_ch); #define V_nd6_timer_ch VNET(nd6_timer_ch) /* nd6_rtr.c */ VNET_DECLARE(int, nd6_defifindex); VNET_DECLARE(int, ip6_desync_factor); /* seconds */ VNET_DECLARE(u_int32_t, ip6_temp_preferred_lifetime); /* seconds */ VNET_DECLARE(u_int32_t, ip6_temp_valid_lifetime); /* seconds */ VNET_DECLARE(int, ip6_temp_regen_advance); /* seconds */ #define V_nd6_defifindex VNET(nd6_defifindex) #define V_ip6_desync_factor VNET(ip6_desync_factor) #define V_ip6_temp_preferred_lifetime VNET(ip6_temp_preferred_lifetime) #define V_ip6_temp_valid_lifetime VNET(ip6_temp_valid_lifetime) #define V_ip6_temp_regen_advance VNET(ip6_temp_regen_advance) union nd_opts { struct nd_opt_hdr *nd_opt_array[8]; /* max = target address list */ struct { struct nd_opt_hdr *zero; struct nd_opt_hdr *src_lladdr; struct nd_opt_hdr *tgt_lladdr; struct nd_opt_prefix_info *pi_beg; /* multiple opts, start */ struct nd_opt_rd_hdr *rh; struct nd_opt_mtu *mtu; struct nd_opt_hdr *search; /* multiple opts */ struct nd_opt_hdr *last; /* multiple opts */ int done; struct nd_opt_prefix_info *pi_end;/* multiple opts, end */ } nd_opt_each; }; #define nd_opts_src_lladdr nd_opt_each.src_lladdr #define nd_opts_tgt_lladdr nd_opt_each.tgt_lladdr #define nd_opts_pi nd_opt_each.pi_beg #define nd_opts_pi_end nd_opt_each.pi_end #define nd_opts_rh nd_opt_each.rh #define nd_opts_mtu nd_opt_each.mtu #define nd_opts_search nd_opt_each.search #define nd_opts_last nd_opt_each.last #define nd_opts_done nd_opt_each.done /* XXX: need nd6_var.h?? */ /* nd6.c */ void nd6_init(void); #ifdef VIMAGE void nd6_destroy(void); #endif struct nd_ifinfo *nd6_ifattach(struct ifnet *); void nd6_ifdetach(struct nd_ifinfo *); int nd6_is_addr_neighbor(struct sockaddr_in6 *, struct ifnet *); void nd6_option_init(void *, int, union nd_opts *); struct nd_opt_hdr *nd6_option(union nd_opts *); int nd6_options(union nd_opts *); struct llentry *nd6_lookup(struct in6_addr *, int, struct ifnet *); void nd6_setmtu(struct ifnet *); void nd6_llinfo_settimer(struct llentry *, long); void nd6_llinfo_settimer_locked(struct llentry *, long); void nd6_timer(void *); void nd6_purge(struct ifnet *); void nd6_nud_hint(struct rtentry *, struct in6_addr *, int); int nd6_resolve(struct ifnet *, struct rtentry *, struct mbuf *, struct sockaddr *, u_char *); int nd6_ioctl(u_long, caddr_t, struct ifnet *); struct llentry *nd6_cache_lladdr(struct ifnet *, struct in6_addr *, char *, int, int, int); int nd6_output(struct ifnet *, struct ifnet *, struct mbuf *, struct sockaddr_in6 *, struct rtentry *); int nd6_output_lle(struct ifnet *, struct ifnet *, struct mbuf *, struct sockaddr_in6 *, struct rtentry *, struct llentry *, struct mbuf **); int nd6_output_flush(struct ifnet *, struct ifnet *, struct mbuf *, struct sockaddr_in6 *); int nd6_need_cache(struct ifnet *); int nd6_add_ifa_lle(struct in6_ifaddr *); void nd6_rem_ifa_lle(struct in6_ifaddr *); int nd6_storelladdr(struct ifnet *, struct mbuf *, const struct sockaddr *, u_char *, uint32_t *); /* nd6_nbr.c */ void nd6_na_input(struct mbuf *, int, int); void nd6_na_output(struct ifnet *, const struct in6_addr *, const struct in6_addr *, u_long, int, struct sockaddr *); void nd6_ns_input(struct mbuf *, int, int); void nd6_ns_output(struct ifnet *, const struct in6_addr *, const struct in6_addr *, struct llentry *, int); caddr_t nd6_ifptomac(struct ifnet *); +void nd6_dad_init(void); void nd6_dad_start(struct ifaddr *, int); void nd6_dad_stop(struct ifaddr *); /* nd6_rtr.c */ void nd6_rs_input(struct mbuf *, int, int); void nd6_ra_input(struct mbuf *, int, int); void prelist_del(struct nd_prefix *); void defrouter_reset(void); void defrouter_select(void); void defrtrlist_del(struct nd_defrouter *); void prelist_remove(struct nd_prefix *); int nd6_prelist_add(struct nd_prefixctl *, struct nd_defrouter *, struct nd_prefix **); void pfxlist_onlink_check(void); struct nd_defrouter *defrouter_lookup(struct in6_addr *, struct ifnet *); struct nd_prefix *nd6_prefix_lookup(struct nd_prefixctl *); void rt6_flush(struct in6_addr *, struct ifnet *); int nd6_setdefaultiface(int); int in6_tmpifadd(const struct in6_ifaddr *, int, int); #endif /* _KERNEL */ #endif /* _NETINET6_ND6_H_ */ Index: projects/building-blocks/sys/netinet6/nd6_nbr.c =================================================================== --- projects/building-blocks/sys/netinet6/nd6_nbr.c (revision 275663) +++ projects/building-blocks/sys/netinet6/nd6_nbr.c (revision 275664) @@ -1,1595 +1,1566 @@ /*- * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the project nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * $KAME: nd6_nbr.c,v 1.86 2002/01/21 02:33:04 jinmei Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_ipsec.h" #include "opt_mpath.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include #include #include #include #ifdef RADIX_MPATH #include #endif #include #include #include #define L3_ADDR_SIN6(le) ((struct sockaddr_in6 *) L3_ADDR(le)) #include #include #include #include #include #include #include #include #include #define SDL(s) ((struct sockaddr_dl *)s) struct dadq; static struct dadq *nd6_dad_find(struct ifaddr *); static void nd6_dad_add(struct dadq *dp); static void nd6_dad_del(struct dadq *dp); +static void nd6_dad_rele(struct dadq *); static void nd6_dad_starttimer(struct dadq *, int); static void nd6_dad_stoptimer(struct dadq *); static void nd6_dad_timer(struct dadq *); static void nd6_dad_duplicated(struct ifaddr *, struct dadq *); static void nd6_dad_ns_output(struct dadq *, struct ifaddr *); static void nd6_dad_ns_input(struct ifaddr *); static void nd6_dad_na_input(struct ifaddr *); static void nd6_na_output_fib(struct ifnet *, const struct in6_addr *, const struct in6_addr *, u_long, int, struct sockaddr *, u_int); static VNET_DEFINE(int, dad_ignore_ns) = 0; /* ignore NS in DAD - specwise incorrect */ static VNET_DEFINE(int, dad_maxtry) = 15; /* max # of *tries* to transmit DAD packet */ #define V_dad_ignore_ns VNET(dad_ignore_ns) #define V_dad_maxtry VNET(dad_maxtry) /* * Input a Neighbor Solicitation Message. * * Based on RFC 2461 * Based on RFC 2462 (duplicate address detection) */ void nd6_ns_input(struct mbuf *m, int off, int icmp6len) { struct ifnet *ifp = m->m_pkthdr.rcvif; struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct nd_neighbor_solicit *nd_ns; struct in6_addr saddr6 = ip6->ip6_src; struct in6_addr daddr6 = ip6->ip6_dst; struct in6_addr taddr6; struct in6_addr myaddr6; char *lladdr = NULL; struct ifaddr *ifa = NULL; int lladdrlen = 0; int anycast = 0, proxy = 0, tentative = 0; int tlladdr; int rflag; union nd_opts ndopts; struct sockaddr_dl proxydl; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; rflag = (V_ip6_forwarding) ? ND_NA_FLAG_ROUTER : 0; if (ND_IFINFO(ifp)->flags & ND6_IFF_ACCEPT_RTADV && V_ip6_norbit_raif) rflag = 0; #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, icmp6len,); nd_ns = (struct nd_neighbor_solicit *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(nd_ns, struct nd_neighbor_solicit *, m, off, icmp6len); if (nd_ns == NULL) { ICMP6STAT_INC(icp6s_tooshort); return; } #endif ip6 = mtod(m, struct ip6_hdr *); /* adjust pointer for safety */ taddr6 = nd_ns->nd_ns_target; if (in6_setscope(&taddr6, ifp, NULL) != 0) goto bad; if (ip6->ip6_hlim != 255) { nd6log((LOG_ERR, "nd6_ns_input: invalid hlim (%d) from %s to %s on %s\n", ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp))); goto bad; } if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) { /* dst has to be a solicited node multicast address. */ if (daddr6.s6_addr16[0] == IPV6_ADDR_INT16_MLL && /* don't check ifindex portion */ daddr6.s6_addr32[1] == 0 && daddr6.s6_addr32[2] == IPV6_ADDR_INT32_ONE && daddr6.s6_addr8[12] == 0xff) { ; /* good */ } else { nd6log((LOG_INFO, "nd6_ns_input: bad DAD packet " "(wrong ip6 dst)\n")); goto bad; } } else if (!V_nd6_onlink_ns_rfc4861) { struct sockaddr_in6 src_sa6; /* * According to recent IETF discussions, it is not a good idea * to accept a NS from an address which would not be deemed * to be a neighbor otherwise. This point is expected to be * clarified in future revisions of the specification. */ bzero(&src_sa6, sizeof(src_sa6)); src_sa6.sin6_family = AF_INET6; src_sa6.sin6_len = sizeof(src_sa6); src_sa6.sin6_addr = saddr6; if (nd6_is_addr_neighbor(&src_sa6, ifp) == 0) { nd6log((LOG_INFO, "nd6_ns_input: " "NS packet from non-neighbor\n")); goto bad; } } if (IN6_IS_ADDR_MULTICAST(&taddr6)) { nd6log((LOG_INFO, "nd6_ns_input: bad NS target (multicast)\n")); goto bad; } icmp6len -= sizeof(*nd_ns); nd6_option_init(nd_ns + 1, icmp6len, &ndopts); if (nd6_options(&ndopts) < 0) { nd6log((LOG_INFO, "nd6_ns_input: invalid ND option, ignored\n")); /* nd6_options have incremented stats */ goto freeit; } if (ndopts.nd_opts_src_lladdr) { lladdr = (char *)(ndopts.nd_opts_src_lladdr + 1); lladdrlen = ndopts.nd_opts_src_lladdr->nd_opt_len << 3; } if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) && lladdr) { nd6log((LOG_INFO, "nd6_ns_input: bad DAD packet " "(link-layer address option)\n")); goto bad; } /* * Attaching target link-layer address to the NA? * (RFC 2461 7.2.4) * * NS IP dst is unicast/anycast MUST NOT add * NS IP dst is solicited-node multicast MUST add * * In implementation, we add target link-layer address by default. * We do not add one in MUST NOT cases. */ if (!IN6_IS_ADDR_MULTICAST(&daddr6)) tlladdr = 0; else tlladdr = 1; /* * Target address (taddr6) must be either: * (1) Valid unicast/anycast address for my receiving interface, * (2) Unicast address for which I'm offering proxy service, or * (3) "tentative" address on which DAD is being performed. */ /* (1) and (3) check. */ if (ifp->if_carp) ifa = (*carp_iamatch6_p)(ifp, &taddr6); else ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &taddr6); /* (2) check. */ if (ifa == NULL) { struct route_in6 ro; int need_proxy; bzero(&ro, sizeof(ro)); ro.ro_dst.sin6_len = sizeof(struct sockaddr_in6); ro.ro_dst.sin6_family = AF_INET6; ro.ro_dst.sin6_addr = taddr6; /* Always use the default FIB. */ #ifdef RADIX_MPATH rtalloc_mpath_fib((struct route *)&ro, ntohl(taddr6.s6_addr32[3]), RT_DEFAULT_FIB); #else in6_rtalloc(&ro, RT_DEFAULT_FIB); #endif need_proxy = (ro.ro_rt && (ro.ro_rt->rt_flags & RTF_ANNOUNCE) != 0 && ro.ro_rt->rt_gateway->sa_family == AF_LINK); if (ro.ro_rt != NULL) { if (need_proxy) proxydl = *SDL(ro.ro_rt->rt_gateway); RTFREE(ro.ro_rt); } if (need_proxy) { /* * proxy NDP for single entry */ ifa = (struct ifaddr *)in6ifa_ifpforlinklocal(ifp, IN6_IFF_NOTREADY|IN6_IFF_ANYCAST); if (ifa) proxy = 1; } } if (ifa == NULL) { /* * We've got an NS packet, and we don't have that adddress * assigned for us. We MUST silently ignore it. * See RFC2461 7.2.3. */ goto freeit; } myaddr6 = *IFA_IN6(ifa); anycast = ((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_ANYCAST; tentative = ((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_TENTATIVE; if (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_DUPLICATED) goto freeit; if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { nd6log((LOG_INFO, "nd6_ns_input: lladdrlen mismatch for %s " "(if %d, NS packet %d)\n", ip6_sprintf(ip6bufs, &taddr6), ifp->if_addrlen, lladdrlen - 2)); goto bad; } if (IN6_ARE_ADDR_EQUAL(&myaddr6, &saddr6)) { nd6log((LOG_INFO, "nd6_ns_input: duplicate IP6 address %s\n", ip6_sprintf(ip6bufs, &saddr6))); goto freeit; } /* * We have neighbor solicitation packet, with target address equals to * one of my tentative address. * * src addr how to process? * --- --- * multicast of course, invalid (rejected in ip6_input) * unicast somebody is doing address resolution -> ignore * unspec dup address detection * * The processing is defined in RFC 2462. */ if (tentative) { /* * If source address is unspecified address, it is for * duplicate address detection. * * If not, the packet is for addess resolution; * silently ignore it. */ if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) nd6_dad_ns_input(ifa); goto freeit; } /* * If the source address is unspecified address, entries must not * be created or updated. * It looks that sender is performing DAD. Output NA toward * all-node multicast address, to tell the sender that I'm using * the address. * S bit ("solicited") must be zero. */ if (IN6_IS_ADDR_UNSPECIFIED(&saddr6)) { struct in6_addr in6_all; in6_all = in6addr_linklocal_allnodes; if (in6_setscope(&in6_all, ifp, NULL) != 0) goto bad; nd6_na_output_fib(ifp, &in6_all, &taddr6, ((anycast || proxy || !tlladdr) ? 0 : ND_NA_FLAG_OVERRIDE) | rflag, tlladdr, proxy ? (struct sockaddr *)&proxydl : NULL, M_GETFIB(m)); goto freeit; } nd6_cache_lladdr(ifp, &saddr6, lladdr, lladdrlen, ND_NEIGHBOR_SOLICIT, 0); nd6_na_output_fib(ifp, &saddr6, &taddr6, ((anycast || proxy || !tlladdr) ? 0 : ND_NA_FLAG_OVERRIDE) | rflag | ND_NA_FLAG_SOLICITED, tlladdr, proxy ? (struct sockaddr *)&proxydl : NULL, M_GETFIB(m)); freeit: if (ifa != NULL) ifa_free(ifa); m_freem(m); return; bad: nd6log((LOG_ERR, "nd6_ns_input: src=%s\n", ip6_sprintf(ip6bufs, &saddr6))); nd6log((LOG_ERR, "nd6_ns_input: dst=%s\n", ip6_sprintf(ip6bufs, &daddr6))); nd6log((LOG_ERR, "nd6_ns_input: tgt=%s\n", ip6_sprintf(ip6bufs, &taddr6))); ICMP6STAT_INC(icp6s_badns); if (ifa != NULL) ifa_free(ifa); m_freem(m); } /* * Output a Neighbor Solicitation Message. Caller specifies: * - ICMP6 header source IP6 address * - ND6 header target IP6 address * - ND6 header source datalink address * * Based on RFC 2461 * Based on RFC 2462 (duplicate address detection) * * ln - for source address determination * dad - duplicate address detection */ void nd6_ns_output(struct ifnet *ifp, const struct in6_addr *daddr6, const struct in6_addr *taddr6, struct llentry *ln, int dad) { struct mbuf *m; struct m_tag *mtag; struct ip6_hdr *ip6; struct nd_neighbor_solicit *nd_ns; struct ip6_moptions im6o; int icmp6len; int maxlen; caddr_t mac; struct route_in6 ro; if (IN6_IS_ADDR_MULTICAST(taddr6)) return; /* estimate the size of message */ maxlen = sizeof(*ip6) + sizeof(*nd_ns); maxlen += (sizeof(struct nd_opt_hdr) + ifp->if_addrlen + 7) & ~7; if (max_linkhdr + maxlen >= MCLBYTES) { #ifdef DIAGNOSTIC printf("nd6_ns_output: max_linkhdr + maxlen >= MCLBYTES " "(%d + %d > %d)\n", max_linkhdr, maxlen, MCLBYTES); #endif return; } if (max_linkhdr + maxlen > MHLEN) m = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR); else m = m_gethdr(M_NOWAIT, MT_DATA); if (m == NULL) return; bzero(&ro, sizeof(ro)); if (daddr6 == NULL || IN6_IS_ADDR_MULTICAST(daddr6)) { m->m_flags |= M_MCAST; im6o.im6o_multicast_ifp = ifp; im6o.im6o_multicast_hlim = 255; im6o.im6o_multicast_loop = 0; } icmp6len = sizeof(*nd_ns); m->m_pkthdr.len = m->m_len = sizeof(*ip6) + icmp6len; m->m_data += max_linkhdr; /* or MH_ALIGN() equivalent? */ /* fill neighbor solicitation packet */ ip6 = mtod(m, struct ip6_hdr *); ip6->ip6_flow = 0; ip6->ip6_vfc &= ~IPV6_VERSION_MASK; ip6->ip6_vfc |= IPV6_VERSION; /* ip6->ip6_plen will be set later */ ip6->ip6_nxt = IPPROTO_ICMPV6; ip6->ip6_hlim = 255; if (daddr6) ip6->ip6_dst = *daddr6; else { ip6->ip6_dst.s6_addr16[0] = IPV6_ADDR_INT16_MLL; ip6->ip6_dst.s6_addr16[1] = 0; ip6->ip6_dst.s6_addr32[1] = 0; ip6->ip6_dst.s6_addr32[2] = IPV6_ADDR_INT32_ONE; ip6->ip6_dst.s6_addr32[3] = taddr6->s6_addr32[3]; ip6->ip6_dst.s6_addr8[12] = 0xff; if (in6_setscope(&ip6->ip6_dst, ifp, NULL) != 0) goto bad; } if (!dad) { struct ifaddr *ifa; /* * RFC2461 7.2.2: * "If the source address of the packet prompting the * solicitation is the same as one of the addresses assigned * to the outgoing interface, that address SHOULD be placed * in the IP Source Address of the outgoing solicitation. * Otherwise, any one of the addresses assigned to the * interface should be used." * * We use the source address for the prompting packet * (saddr6), if: * - saddr6 is given from the caller (by giving "ln"), and * - saddr6 belongs to the outgoing interface. * Otherwise, we perform the source address selection as usual. */ struct in6_addr *hsrc; hsrc = NULL; if (ln != NULL) { LLE_RLOCK(ln); if (ln->la_hold != NULL) { struct ip6_hdr *hip6; /* hold ip6 */ /* * assuming every packet in la_hold has the same IP * header */ hip6 = mtod(ln->la_hold, struct ip6_hdr *); /* XXX pullup? */ if (sizeof(*hip6) < ln->la_hold->m_len) { ip6->ip6_src = hip6->ip6_src; hsrc = &hip6->ip6_src; } } LLE_RUNLOCK(ln); } if (hsrc && (ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, hsrc)) != NULL) { /* ip6_src set already. */ ifa_free(ifa); } else { int error; struct sockaddr_in6 dst_sa; struct in6_addr src_in; struct ifnet *oifp; bzero(&dst_sa, sizeof(dst_sa)); dst_sa.sin6_family = AF_INET6; dst_sa.sin6_len = sizeof(dst_sa); dst_sa.sin6_addr = ip6->ip6_dst; oifp = ifp; error = in6_selectsrc(&dst_sa, NULL, NULL, &ro, NULL, &oifp, &src_in); if (error) { char ip6buf[INET6_ADDRSTRLEN]; nd6log((LOG_DEBUG, "nd6_ns_output: source can't be " "determined: dst=%s, error=%d\n", ip6_sprintf(ip6buf, &dst_sa.sin6_addr), error)); goto bad; } ip6->ip6_src = src_in; } } else { /* * Source address for DAD packet must always be IPv6 * unspecified address. (0::0) * We actually don't have to 0-clear the address (we did it * above), but we do so here explicitly to make the intention * clearer. */ bzero(&ip6->ip6_src, sizeof(ip6->ip6_src)); } nd_ns = (struct nd_neighbor_solicit *)(ip6 + 1); nd_ns->nd_ns_type = ND_NEIGHBOR_SOLICIT; nd_ns->nd_ns_code = 0; nd_ns->nd_ns_reserved = 0; nd_ns->nd_ns_target = *taddr6; in6_clearscope(&nd_ns->nd_ns_target); /* XXX */ /* * Add source link-layer address option. * * spec implementation * --- --- * DAD packet MUST NOT do not add the option * there's no link layer address: * impossible do not add the option * there's link layer address: * Multicast NS MUST add one add the option * Unicast NS SHOULD add one add the option */ if (!dad && (mac = nd6_ifptomac(ifp))) { int optlen = sizeof(struct nd_opt_hdr) + ifp->if_addrlen; struct nd_opt_hdr *nd_opt = (struct nd_opt_hdr *)(nd_ns + 1); /* 8 byte alignments... */ optlen = (optlen + 7) & ~7; m->m_pkthdr.len += optlen; m->m_len += optlen; icmp6len += optlen; bzero((caddr_t)nd_opt, optlen); nd_opt->nd_opt_type = ND_OPT_SOURCE_LINKADDR; nd_opt->nd_opt_len = optlen >> 3; bcopy(mac, (caddr_t)(nd_opt + 1), ifp->if_addrlen); } ip6->ip6_plen = htons((u_short)icmp6len); nd_ns->nd_ns_cksum = 0; nd_ns->nd_ns_cksum = in6_cksum(m, IPPROTO_ICMPV6, sizeof(*ip6), icmp6len); if (send_sendso_input_hook != NULL) { mtag = m_tag_get(PACKET_TAG_ND_OUTGOING, sizeof(unsigned short), M_NOWAIT); if (mtag == NULL) goto bad; *(unsigned short *)(mtag + 1) = nd_ns->nd_ns_type; m_tag_prepend(m, mtag); } ip6_output(m, NULL, &ro, dad ? IPV6_UNSPECSRC : 0, &im6o, NULL, NULL); icmp6_ifstat_inc(ifp, ifs6_out_msg); icmp6_ifstat_inc(ifp, ifs6_out_neighborsolicit); ICMP6STAT_INC(icp6s_outhist[ND_NEIGHBOR_SOLICIT]); /* We don't cache this route. */ RO_RTFREE(&ro); return; bad: if (ro.ro_rt) { RTFREE(ro.ro_rt); } m_freem(m); return; } /* * Neighbor advertisement input handling. * * Based on RFC 2461 * Based on RFC 2462 (duplicate address detection) * * the following items are not implemented yet: * - proxy advertisement delay rule (RFC2461 7.2.8, last paragraph, SHOULD) * - anycast advertisement delay rule (RFC2461 7.2.7, SHOULD) */ void nd6_na_input(struct mbuf *m, int off, int icmp6len) { struct ifnet *ifp = m->m_pkthdr.rcvif; struct ip6_hdr *ip6 = mtod(m, struct ip6_hdr *); struct nd_neighbor_advert *nd_na; struct in6_addr daddr6 = ip6->ip6_dst; struct in6_addr taddr6; int flags; int is_router; int is_solicited; int is_override; char *lladdr = NULL; int lladdrlen = 0; int checklink = 0; struct ifaddr *ifa; struct llentry *ln = NULL; union nd_opts ndopts; struct mbuf *chain = NULL; struct m_tag *mtag; struct sockaddr_in6 sin6; char ip6bufs[INET6_ADDRSTRLEN], ip6bufd[INET6_ADDRSTRLEN]; if (ip6->ip6_hlim != 255) { nd6log((LOG_ERR, "nd6_na_input: invalid hlim (%d) from %s to %s on %s\n", ip6->ip6_hlim, ip6_sprintf(ip6bufs, &ip6->ip6_src), ip6_sprintf(ip6bufd, &ip6->ip6_dst), if_name(ifp))); goto bad; } #ifndef PULLDOWN_TEST IP6_EXTHDR_CHECK(m, off, icmp6len,); nd_na = (struct nd_neighbor_advert *)((caddr_t)ip6 + off); #else IP6_EXTHDR_GET(nd_na, struct nd_neighbor_advert *, m, off, icmp6len); if (nd_na == NULL) { ICMP6STAT_INC(icp6s_tooshort); return; } #endif flags = nd_na->nd_na_flags_reserved; is_router = ((flags & ND_NA_FLAG_ROUTER) != 0); is_solicited = ((flags & ND_NA_FLAG_SOLICITED) != 0); is_override = ((flags & ND_NA_FLAG_OVERRIDE) != 0); taddr6 = nd_na->nd_na_target; if (in6_setscope(&taddr6, ifp, NULL)) goto bad; /* XXX: impossible */ if (IN6_IS_ADDR_MULTICAST(&taddr6)) { nd6log((LOG_ERR, "nd6_na_input: invalid target address %s\n", ip6_sprintf(ip6bufs, &taddr6))); goto bad; } if (IN6_IS_ADDR_MULTICAST(&daddr6)) if (is_solicited) { nd6log((LOG_ERR, "nd6_na_input: a solicited adv is multicasted\n")); goto bad; } icmp6len -= sizeof(*nd_na); nd6_option_init(nd_na + 1, icmp6len, &ndopts); if (nd6_options(&ndopts) < 0) { nd6log((LOG_INFO, "nd6_na_input: invalid ND option, ignored\n")); /* nd6_options have incremented stats */ goto freeit; } if (ndopts.nd_opts_tgt_lladdr) { lladdr = (char *)(ndopts.nd_opts_tgt_lladdr + 1); lladdrlen = ndopts.nd_opts_tgt_lladdr->nd_opt_len << 3; } /* * This effectively disables the DAD check on a non-master CARP * address. */ if (ifp->if_carp) ifa = (*carp_iamatch6_p)(ifp, &taddr6); else ifa = (struct ifaddr *)in6ifa_ifpwithaddr(ifp, &taddr6); /* * Target address matches one of my interface address. * * If my address is tentative, this means that there's somebody * already using the same address as mine. This indicates DAD failure. * This is defined in RFC 2462. * * Otherwise, process as defined in RFC 2461. */ if (ifa && (((struct in6_ifaddr *)ifa)->ia6_flags & IN6_IFF_TENTATIVE)) { ifa_free(ifa); nd6_dad_na_input(ifa); goto freeit; } /* Just for safety, maybe unnecessary. */ if (ifa) { ifa_free(ifa); log(LOG_ERR, "nd6_na_input: duplicate IP6 address %s\n", ip6_sprintf(ip6bufs, &taddr6)); goto freeit; } if (lladdr && ((ifp->if_addrlen + 2 + 7) & ~7) != lladdrlen) { nd6log((LOG_INFO, "nd6_na_input: lladdrlen mismatch for %s " "(if %d, NA packet %d)\n", ip6_sprintf(ip6bufs, &taddr6), ifp->if_addrlen, lladdrlen - 2)); goto bad; } /* * If no neighbor cache entry is found, NA SHOULD silently be * discarded. */ IF_AFDATA_RLOCK(ifp); ln = nd6_lookup(&taddr6, LLE_EXCLUSIVE, ifp); IF_AFDATA_RUNLOCK(ifp); if (ln == NULL) { goto freeit; } if (ln->ln_state == ND6_LLINFO_INCOMPLETE) { /* * If the link-layer has address, and no lladdr option came, * discard the packet. */ if (ifp->if_addrlen && lladdr == NULL) { goto freeit; } /* * Record link-layer address, and update the state. */ bcopy(lladdr, &ln->ll_addr, ifp->if_addrlen); ln->la_flags |= LLE_VALID; EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_RESOLVED); if (is_solicited) { ln->ln_state = ND6_LLINFO_REACHABLE; ln->ln_byhint = 0; if (!ND6_LLINFO_PERMANENT(ln)) { nd6_llinfo_settimer_locked(ln, (long)ND_IFINFO(ln->lle_tbl->llt_ifp)->reachable * hz); } } else { ln->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } if ((ln->ln_router = is_router) != 0) { /* * This means a router's state has changed from * non-reachable to probably reachable, and might * affect the status of associated prefixes.. */ checklink = 1; } } else { int llchange; /* * Check if the link-layer address has changed or not. */ if (lladdr == NULL) llchange = 0; else { if (ln->la_flags & LLE_VALID) { if (bcmp(lladdr, &ln->ll_addr, ifp->if_addrlen)) llchange = 1; else llchange = 0; } else llchange = 1; } /* * This is VERY complex. Look at it with care. * * override solicit lladdr llchange action * (L: record lladdr) * * 0 0 n -- (2c) * 0 0 y n (2b) L * 0 0 y y (1) REACHABLE->STALE * 0 1 n -- (2c) *->REACHABLE * 0 1 y n (2b) L *->REACHABLE * 0 1 y y (1) REACHABLE->STALE * 1 0 n -- (2a) * 1 0 y n (2a) L * 1 0 y y (2a) L *->STALE * 1 1 n -- (2a) *->REACHABLE * 1 1 y n (2a) L *->REACHABLE * 1 1 y y (2a) L *->REACHABLE */ if (!is_override && (lladdr != NULL && llchange)) { /* (1) */ /* * If state is REACHABLE, make it STALE. * no other updates should be done. */ if (ln->ln_state == ND6_LLINFO_REACHABLE) { ln->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } goto freeit; } else if (is_override /* (2a) */ || (!is_override && (lladdr != NULL && !llchange)) /* (2b) */ || lladdr == NULL) { /* (2c) */ /* * Update link-local address, if any. */ if (lladdr != NULL) { bcopy(lladdr, &ln->ll_addr, ifp->if_addrlen); ln->la_flags |= LLE_VALID; EVENTHANDLER_INVOKE(lle_event, ln, LLENTRY_RESOLVED); } /* * If solicited, make the state REACHABLE. * If not solicited and the link-layer address was * changed, make it STALE. */ if (is_solicited) { ln->ln_state = ND6_LLINFO_REACHABLE; ln->ln_byhint = 0; if (!ND6_LLINFO_PERMANENT(ln)) { nd6_llinfo_settimer_locked(ln, (long)ND_IFINFO(ifp)->reachable * hz); } } else { if (lladdr != NULL && llchange) { ln->ln_state = ND6_LLINFO_STALE; nd6_llinfo_settimer_locked(ln, (long)V_nd6_gctimer * hz); } } } if (ln->ln_router && !is_router) { /* * The peer dropped the router flag. * Remove the sender from the Default Router List and * update the Destination Cache entries. */ struct nd_defrouter *dr; struct in6_addr *in6; in6 = &L3_ADDR_SIN6(ln)->sin6_addr; /* * Lock to protect the default router list. * XXX: this might be unnecessary, since this function * is only called under the network software interrupt * context. However, we keep it just for safety. */ dr = defrouter_lookup(in6, ln->lle_tbl->llt_ifp); if (dr) defrtrlist_del(dr); else if (ND_IFINFO(ln->lle_tbl->llt_ifp)->flags & ND6_IFF_ACCEPT_RTADV) { /* * Even if the neighbor is not in the default * router list, the neighbor may be used * as a next hop for some destinations * (e.g. redirect case). So we must * call rt6_flush explicitly. */ rt6_flush(&ip6->ip6_src, ifp); } } ln->ln_router = is_router; } /* XXX - QL * Does this matter? * rt->rt_flags &= ~RTF_REJECT; */ ln->la_asked = 0; if (ln->la_hold) { struct mbuf *m_hold, *m_hold_next; /* * reset the la_hold in advance, to explicitly * prevent a la_hold lookup in nd6_output() * (wouldn't happen, though...) */ for (m_hold = ln->la_hold, ln->la_hold = NULL; m_hold; m_hold = m_hold_next) { m_hold_next = m_hold->m_nextpkt; m_hold->m_nextpkt = NULL; /* * we assume ifp is not a loopback here, so just set * the 2nd argument as the 1st one. */ if (send_sendso_input_hook != NULL) { mtag = m_tag_get(PACKET_TAG_ND_OUTGOING, sizeof(unsigned short), M_NOWAIT); if (mtag == NULL) goto bad; m_tag_prepend(m, mtag); } nd6_output_lle(ifp, ifp, m_hold, L3_ADDR_SIN6(ln), NULL, ln, &chain); } } freeit: if (ln != NULL) { if (chain) memcpy(&sin6, L3_ADDR_SIN6(ln), sizeof(sin6)); LLE_WUNLOCK(ln); if (chain) nd6_output_flush(ifp, ifp, chain, &sin6); } if (checklink) pfxlist_onlink_check(); m_freem(m); return; bad: if (ln != NULL) LLE_WUNLOCK(ln); ICMP6STAT_INC(icp6s_badna); m_freem(m); } /* * Neighbor advertisement output handling. * * Based on RFC 2461 * * the following items are not implemented yet: * - proxy advertisement delay rule (RFC2461 7.2.8, last paragraph, SHOULD) * - anycast advertisement delay rule (RFC2461 7.2.7, SHOULD) * * tlladdr - 1 if include target link-layer address * sdl0 - sockaddr_dl (= proxy NA) or NULL */ static void nd6_na_output_fib(struct ifnet *ifp, const struct in6_addr *daddr6_0, const struct in6_addr *taddr6, u_long flags, int tlladdr, struct sockaddr *sdl0, u_int fibnum) { struct mbuf *m; struct m_tag *mtag; struct ifnet *oifp; struct ip6_hdr *ip6; struct nd_neighbor_advert *nd_na; struct ip6_moptions im6o; struct in6_addr src, daddr6; struct sockaddr_in6 dst_sa; int icmp6len, maxlen, error; caddr_t mac = NULL; struct route_in6 ro; bzero(&ro, sizeof(ro)); daddr6 = *daddr6_0; /* make a local copy for modification */ /* estimate the size of message */ maxlen = sizeof(*ip6) + sizeof(*nd_na); maxlen += (sizeof(struct nd_opt_hdr) + ifp->if_addrlen + 7) & ~7; if (max_linkhdr + maxlen >= MCLBYTES) { #ifdef DIAGNOSTIC printf("nd6_na_output: max_linkhdr + maxlen >= MCLBYTES " "(%d + %d > %d)\n", max_linkhdr, maxlen, MCLBYTES); #endif return; } if (max_linkhdr + maxlen > MHLEN) m = m_getcl(M_NOWAIT, MT_DATA, M_PKTHDR); else m = m_gethdr(M_NOWAIT, MT_DATA); if (m == NULL) return; M_SETFIB(m, fibnum); if (IN6_IS_ADDR_MULTICAST(&daddr6)) { m->m_flags |= M_MCAST; im6o.im6o_multicast_ifp = ifp; im6o.im6o_multicast_hlim = 255; im6o.im6o_multicast_loop = 0; } icmp6len = sizeof(*nd_na); m->m_pkthdr.len = m->m_len = sizeof(struct ip6_hdr) + icmp6len; m->m_data += max_linkhdr; /* or MH_ALIGN() equivalent? */ /* fill neighbor advertisement packet */ ip6 = mtod(m, struct ip6_hdr *); ip6->ip6_flow = 0; ip6->ip6_vfc &= ~IPV6_VERSION_MASK; ip6->ip6_vfc |= IPV6_VERSION; ip6->ip6_nxt = IPPROTO_ICMPV6; ip6->ip6_hlim = 255; if (IN6_IS_ADDR_UNSPECIFIED(&daddr6)) { /* reply to DAD */ daddr6.s6_addr16[0] = IPV6_ADDR_INT16_MLL; daddr6.s6_addr16[1] = 0; daddr6.s6_addr32[1] = 0; daddr6.s6_addr32[2] = 0; daddr6.s6_addr32[3] = IPV6_ADDR_INT32_ONE; if (in6_setscope(&daddr6, ifp, NULL)) goto bad; flags &= ~ND_NA_FLAG_SOLICITED; } ip6->ip6_dst = daddr6; bzero(&dst_sa, sizeof(struct sockaddr_in6)); dst_sa.sin6_family = AF_INET6; dst_sa.sin6_len = sizeof(struct sockaddr_in6); dst_sa.sin6_addr = daddr6; /* * Select a source whose scope is the same as that of the dest. */ bcopy(&dst_sa, &ro.ro_dst, sizeof(dst_sa)); oifp = ifp; error = in6_selectsrc(&dst_sa, NULL, NULL, &ro, NULL, &oifp, &src); if (error) { char ip6buf[INET6_ADDRSTRLEN]; nd6log((LOG_DEBUG, "nd6_na_output: source can't be " "determined: dst=%s, error=%d\n", ip6_sprintf(ip6buf, &dst_sa.sin6_addr), error)); goto bad; } ip6->ip6_src = src; nd_na = (struct nd_neighbor_advert *)(ip6 + 1); nd_na->nd_na_type = ND_NEIGHBOR_ADVERT; nd_na->nd_na_code = 0; nd_na->nd_na_target = *taddr6; in6_clearscope(&nd_na->nd_na_target); /* XXX */ /* * "tlladdr" indicates NS's condition for adding tlladdr or not. * see nd6_ns_input() for details. * Basically, if NS packet is sent to unicast/anycast addr, * target lladdr option SHOULD NOT be included. */ if (tlladdr) { /* * sdl0 != NULL indicates proxy NA. If we do proxy, use * lladdr in sdl0. If we are not proxying (sending NA for * my address) use lladdr configured for the interface. */ if (sdl0 == NULL) { if (ifp->if_carp) mac = (*carp_macmatch6_p)(ifp, m, taddr6); if (mac == NULL) mac = nd6_ifptomac(ifp); } else if (sdl0->sa_family == AF_LINK) { struct sockaddr_dl *sdl; sdl = (struct sockaddr_dl *)sdl0; if (sdl->sdl_alen == ifp->if_addrlen) mac = LLADDR(sdl); } } if (tlladdr && mac) { int optlen = sizeof(struct nd_opt_hdr) + ifp->if_addrlen; struct nd_opt_hdr *nd_opt = (struct nd_opt_hdr *)(nd_na + 1); /* roundup to 8 bytes alignment! */ optlen = (optlen + 7) & ~7; m->m_pkthdr.len += optlen; m->m_len += optlen; icmp6len += optlen; bzero((caddr_t)nd_opt, optlen); nd_opt->nd_opt_type = ND_OPT_TARGET_LINKADDR; nd_opt->nd_opt_len = optlen >> 3; bcopy(mac, (caddr_t)(nd_opt + 1), ifp->if_addrlen); } else flags &= ~ND_NA_FLAG_OVERRIDE; ip6->ip6_plen = htons((u_short)icmp6len); nd_na->nd_na_flags_reserved = flags; nd_na->nd_na_cksum = 0; nd_na->nd_na_cksum = in6_cksum(m, IPPROTO_ICMPV6, sizeof(struct ip6_hdr), icmp6len); if (send_sendso_input_hook != NULL) { mtag = m_tag_get(PACKET_TAG_ND_OUTGOING, sizeof(unsigned short), M_NOWAIT); if (mtag == NULL) goto bad; *(unsigned short *)(mtag + 1) = nd_na->nd_na_type; m_tag_prepend(m, mtag); } ip6_output(m, NULL, &ro, 0, &im6o, NULL, NULL); icmp6_ifstat_inc(ifp, ifs6_out_msg); icmp6_ifstat_inc(ifp, ifs6_out_neighboradvert); ICMP6STAT_INC(icp6s_outhist[ND_NEIGHBOR_ADVERT]); /* We don't cache this route. */ RO_RTFREE(&ro); return; bad: if (ro.ro_rt) { RTFREE(ro.ro_rt); } m_freem(m); return; } #ifndef BURN_BRIDGES void nd6_na_output(struct ifnet *ifp, const struct in6_addr *daddr6_0, const struct in6_addr *taddr6, u_long flags, int tlladdr, struct sockaddr *sdl0) { nd6_na_output_fib(ifp, daddr6_0, taddr6, flags, tlladdr, sdl0, RT_DEFAULT_FIB); } #endif caddr_t nd6_ifptomac(struct ifnet *ifp) { switch (ifp->if_type) { case IFT_ARCNET: case IFT_ETHER: case IFT_FDDI: case IFT_IEEE1394: #ifdef IFT_L2VLAN case IFT_L2VLAN: #endif #ifdef IFT_IEEE80211 case IFT_IEEE80211: #endif case IFT_INFINIBAND: case IFT_BRIDGE: case IFT_ISO88025: return IF_LLADDR(ifp); default: return NULL; } } struct dadq { TAILQ_ENTRY(dadq) dad_list; struct ifaddr *dad_ifa; int dad_count; /* max NS to send */ int dad_ns_tcount; /* # of trials to send NS */ int dad_ns_ocount; /* NS sent so far */ int dad_ns_icount; int dad_na_icount; struct callout dad_timer_ch; struct vnet *dad_vnet; + u_int dad_refcnt; }; static VNET_DEFINE(TAILQ_HEAD(, dadq), dadq); static VNET_DEFINE(struct rwlock, dad_rwlock); #define V_dadq VNET(dadq) #define V_dad_rwlock VNET(dad_rwlock) -#define DADQ_LOCK_INIT() rw_init(&V_dad_rwlock, "nd6 DAD queue") -#define DADQ_LOCK_DESTROY() rw_destroy(&V_dad_rwlock) -#define DADQ_LOCK_INITIALIZED() rw_initialized(&V_dad_rwlock) #define DADQ_RLOCK() rw_rlock(&V_dad_rwlock) #define DADQ_RUNLOCK() rw_runlock(&V_dad_rwlock) #define DADQ_WLOCK() rw_wlock(&V_dad_rwlock) #define DADQ_WUNLOCK() rw_wunlock(&V_dad_rwlock) static void nd6_dad_add(struct dadq *dp) { - ifa_ref(dp->dad_ifa); /* just for safety */ DADQ_WLOCK(); - TAILQ_INSERT_TAIL(&V_dadq, (struct dadq *)dp, dad_list); + TAILQ_INSERT_TAIL(&V_dadq, dp, dad_list); DADQ_WUNLOCK(); } static void nd6_dad_del(struct dadq *dp) { - ifa_free(dp->dad_ifa); DADQ_WLOCK(); - TAILQ_REMOVE(&V_dadq, (struct dadq *)dp, dad_list); + TAILQ_REMOVE(&V_dadq, dp, dad_list); DADQ_WUNLOCK(); + nd6_dad_rele(dp); } static struct dadq * nd6_dad_find(struct ifaddr *ifa) { struct dadq *dp; DADQ_RLOCK(); TAILQ_FOREACH(dp, &V_dadq, dad_list) - if (dp->dad_ifa == ifa) + if (dp->dad_ifa == ifa) { + refcount_acquire(&dp->dad_refcnt); break; + } DADQ_RUNLOCK(); return (dp); } static void nd6_dad_starttimer(struct dadq *dp, int ticks) { callout_reset(&dp->dad_timer_ch, ticks, (void (*)(void *))nd6_dad_timer, (void *)dp); } static void nd6_dad_stoptimer(struct dadq *dp) { - callout_stop(&dp->dad_timer_ch); + callout_drain(&dp->dad_timer_ch); } +static void +nd6_dad_rele(struct dadq *dp) +{ + + if (refcount_release(&dp->dad_refcnt)) { + ifa_free(dp->dad_ifa); + free(dp, M_IP6NDP); + } +} + +void +nd6_dad_init(void) +{ + + rw_init(&V_dad_rwlock, "nd6 DAD queue"); + TAILQ_INIT(&V_dadq); +} + /* * Start Duplicate Address Detection (DAD) for specified interface address. */ void nd6_dad_start(struct ifaddr *ifa, int delay) { struct in6_ifaddr *ia = (struct in6_ifaddr *)ifa; struct dadq *dp; char ip6buf[INET6_ADDRSTRLEN]; - if (DADQ_LOCK_INITIALIZED() == 0) { - DADQ_LOCK_INIT(); - TAILQ_INIT(&V_dadq); - } - /* * If we don't need DAD, don't do it. * There are several cases: * - DAD is disabled (ip6_dad_count == 0) * - the interface address is anycast */ if (!(ia->ia6_flags & IN6_IFF_TENTATIVE)) { log(LOG_DEBUG, "nd6_dad_start: called with non-tentative address " "%s(%s)\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr), ifa->ifa_ifp ? if_name(ifa->ifa_ifp) : "???"); return; } if (ia->ia6_flags & IN6_IFF_ANYCAST) { ia->ia6_flags &= ~IN6_IFF_TENTATIVE; return; } if (!V_ip6_dad_count) { ia->ia6_flags &= ~IN6_IFF_TENTATIVE; return; } if (ifa->ifa_ifp == NULL) panic("nd6_dad_start: ifa->ifa_ifp == NULL"); if (!(ifa->ifa_ifp->if_flags & IFF_UP)) { return; } if (ND_IFINFO(ifa->ifa_ifp)->flags & ND6_IFF_IFDISABLED) return; - if (nd6_dad_find(ifa) != NULL) { + if ((dp = nd6_dad_find(ifa)) != NULL) { /* DAD already in progress */ + nd6_dad_rele(dp); return; } - dp = malloc(sizeof(*dp), M_IP6NDP, M_NOWAIT); + dp = malloc(sizeof(*dp), M_IP6NDP, M_NOWAIT | M_ZERO); if (dp == NULL) { log(LOG_ERR, "nd6_dad_start: memory allocation failed for " "%s(%s)\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr), ifa->ifa_ifp ? if_name(ifa->ifa_ifp) : "???"); return; } - bzero(dp, sizeof(*dp)); callout_init(&dp->dad_timer_ch, 0); #ifdef VIMAGE dp->dad_vnet = curvnet; #endif nd6log((LOG_DEBUG, "%s: starting DAD for %s\n", if_name(ifa->ifa_ifp), ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr))); /* * Send NS packet for DAD, ip6_dad_count times. * Note that we must delay the first transmission, if this is the * first packet to be sent from the interface after interface * (re)initialization. */ dp->dad_ifa = ifa; + ifa_ref(dp->dad_ifa); dp->dad_count = V_ip6_dad_count; dp->dad_ns_icount = dp->dad_na_icount = 0; dp->dad_ns_ocount = dp->dad_ns_tcount = 0; + refcount_init(&dp->dad_refcnt, 1); nd6_dad_add(dp); if (delay == 0) { nd6_dad_ns_output(dp, ifa); nd6_dad_starttimer(dp, (long)ND_IFINFO(ifa->ifa_ifp)->retrans * hz / 1000); } else { nd6_dad_starttimer(dp, delay); } } /* * terminate DAD unconditionally. used for address removals. */ void nd6_dad_stop(struct ifaddr *ifa) { struct dadq *dp; - if (DADQ_LOCK_INITIALIZED() == 0) - return; dp = nd6_dad_find(ifa); if (!dp) { /* DAD wasn't started yet */ return; } nd6_dad_stoptimer(dp); + /* + * The DAD queue entry may have been removed by nd6_dad_timer() while + * we were waiting for it to stop, so re-do the lookup. + */ + nd6_dad_rele(dp); + if (nd6_dad_find(ifa) == NULL) + return; + nd6_dad_del(dp); - free(dp, M_IP6NDP); + nd6_dad_rele(dp); } static void nd6_dad_timer(struct dadq *dp) { CURVNET_SET(dp->dad_vnet); struct ifaddr *ifa = dp->dad_ifa; struct ifnet *ifp = dp->dad_ifa->ifa_ifp; struct in6_ifaddr *ia = (struct in6_ifaddr *)ifa; char ip6buf[INET6_ADDRSTRLEN]; /* Sanity check */ if (ia == NULL) { log(LOG_ERR, "nd6_dad_timer: called with null parameter\n"); - goto done; + goto err; } if (ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) { /* Do not need DAD for ifdisabled interface. */ - TAILQ_REMOVE(&V_dadq, (struct dadq *)dp, dad_list); log(LOG_ERR, "nd6_dad_timer: cancel DAD on %s because of " "ND6_IFF_IFDISABLED.\n", ifp->if_xname); - free(dp, M_IP6NDP); - dp = NULL; - ifa_free(ifa); - goto done; + goto err; } if (ia->ia6_flags & IN6_IFF_DUPLICATED) { log(LOG_ERR, "nd6_dad_timer: called with duplicated address " "%s(%s)\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr), ifa->ifa_ifp ? if_name(ifa->ifa_ifp) : "???"); - goto done; + goto err; } if ((ia->ia6_flags & IN6_IFF_TENTATIVE) == 0) { log(LOG_ERR, "nd6_dad_timer: called with non-tentative address " "%s(%s)\n", ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr), ifa->ifa_ifp ? if_name(ifa->ifa_ifp) : "???"); - goto done; + goto err; } /* timeouted with IFF_{RUNNING,UP} check */ if (dp->dad_ns_tcount > V_dad_maxtry) { nd6log((LOG_INFO, "%s: could not run DAD, driver problem?\n", if_name(ifa->ifa_ifp))); - - nd6_dad_del(dp); - free(dp, M_IP6NDP); - dp = NULL; - goto done; + goto err; } /* Need more checks? */ if (dp->dad_ns_ocount < dp->dad_count) { /* * We have more NS to go. Send NS packet for DAD. */ nd6_dad_ns_output(dp, ifa); nd6_dad_starttimer(dp, (long)ND_IFINFO(ifa->ifa_ifp)->retrans * hz / 1000); + goto done; } else { /* * We have transmitted sufficient number of DAD packets. * See what we've got. */ - int duplicate; - - duplicate = 0; - - if (dp->dad_na_icount) { - /* - * the check is in nd6_dad_na_input(), - * but just in case - */ - duplicate++; - } - - if (dp->dad_ns_icount) { - /* We've seen NS, means DAD has failed. */ - duplicate++; - } - - if (duplicate) { - /* (*dp) will be freed in nd6_dad_duplicated() */ + if (dp->dad_ns_icount > 0 || dp->dad_na_icount > 0) + /* We've seen NS or NA, means DAD has failed. */ nd6_dad_duplicated(ifa, dp); - dp = NULL; - } else { + else { /* * We are done with DAD. No NA came, no NS came. * No duplicate address found. Check IFDISABLED flag * again in case that it is changed between the * beginning of this function and here. */ if ((ND_IFINFO(ifp)->flags & ND6_IFF_IFDISABLED) == 0) ia->ia6_flags &= ~IN6_IFF_TENTATIVE; nd6log((LOG_DEBUG, "%s: DAD complete for %s - no duplicates found\n", if_name(ifa->ifa_ifp), ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr))); - - nd6_dad_del(dp); - free(dp, M_IP6NDP); - dp = NULL; } } - +err: + nd6_dad_del(dp); done: CURVNET_RESTORE(); } -void +static void nd6_dad_duplicated(struct ifaddr *ifa, struct dadq *dp) { struct in6_ifaddr *ia = (struct in6_ifaddr *)ifa; struct ifnet *ifp; char ip6buf[INET6_ADDRSTRLEN]; log(LOG_ERR, "%s: DAD detected duplicate IPv6 address %s: " "NS in/out=%d/%d, NA in=%d\n", if_name(ifa->ifa_ifp), ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr), dp->dad_ns_icount, dp->dad_ns_ocount, dp->dad_na_icount); ia->ia6_flags &= ~IN6_IFF_TENTATIVE; ia->ia6_flags |= IN6_IFF_DUPLICATED; - /* We are done with DAD, with duplicate address found. (failure) */ - nd6_dad_stoptimer(dp); - ifp = ifa->ifa_ifp; log(LOG_ERR, "%s: DAD complete for %s - duplicate found\n", if_name(ifp), ip6_sprintf(ip6buf, &ia->ia_addr.sin6_addr)); log(LOG_ERR, "%s: manual intervention required\n", if_name(ifp)); /* * If the address is a link-local address formed from an interface * identifier based on the hardware address which is supposed to be * uniquely assigned (e.g., EUI-64 for an Ethernet interface), IP * operation on the interface SHOULD be disabled. * [RFC 4862, Section 5.4.5] */ if (IN6_IS_ADDR_LINKLOCAL(&ia->ia_addr.sin6_addr)) { struct in6_addr in6; /* * To avoid over-reaction, we only apply this logic when we are * very sure that hardware addresses are supposed to be unique. */ switch (ifp->if_type) { case IFT_ETHER: case IFT_FDDI: case IFT_ATM: case IFT_IEEE1394: #ifdef IFT_IEEE80211 case IFT_IEEE80211: #endif case IFT_INFINIBAND: in6 = ia->ia_addr.sin6_addr; if (in6_get_hw_ifid(ifp, &in6) == 0 && IN6_ARE_ADDR_EQUAL(&ia->ia_addr.sin6_addr, &in6)) { ND_IFINFO(ifp)->flags |= ND6_IFF_IFDISABLED; log(LOG_ERR, "%s: possible hardware address " "duplication detected, disable IPv6\n", if_name(ifp)); } break; } } - - nd6_dad_del(dp); - free(dp, M_IP6NDP); } static void nd6_dad_ns_output(struct dadq *dp, struct ifaddr *ifa) { struct in6_ifaddr *ia = (struct in6_ifaddr *)ifa; struct ifnet *ifp = ifa->ifa_ifp; dp->dad_ns_tcount++; if ((ifp->if_flags & IFF_UP) == 0) { return; } if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0) { return; } dp->dad_ns_ocount++; nd6_ns_output(ifp, NULL, &ia->ia_addr.sin6_addr, NULL, 1); } static void nd6_dad_ns_input(struct ifaddr *ifa) { struct in6_ifaddr *ia; struct ifnet *ifp; const struct in6_addr *taddr6; struct dadq *dp; - int duplicate; if (ifa == NULL) panic("ifa == NULL in nd6_dad_ns_input"); ia = (struct in6_ifaddr *)ifa; ifp = ifa->ifa_ifp; taddr6 = &ia->ia_addr.sin6_addr; - duplicate = 0; dp = nd6_dad_find(ifa); + if (dp == NULL) + return; /* Quickhack - completely ignore DAD NS packets */ if (V_dad_ignore_ns) { char ip6buf[INET6_ADDRSTRLEN]; nd6log((LOG_INFO, "nd6_dad_ns_input: ignoring DAD NS packet for " "address %s(%s)\n", ip6_sprintf(ip6buf, taddr6), if_name(ifa->ifa_ifp))); return; } - /* - * if I'm yet to start DAD, someone else started using this address - * first. I have a duplicate and you win. - */ - if (dp == NULL || dp->dad_ns_ocount == 0) - duplicate++; - /* XXX more checks for loopback situation - see nd6_dad_timer too */ - if (duplicate) { - nd6_dad_duplicated(ifa, dp); - dp = NULL; /* will be freed in nd6_dad_duplicated() */ - } else { - /* - * not sure if I got a duplicate. - * increment ns count and see what happens. - */ - if (dp) - dp->dad_ns_icount++; - } + dp->dad_ns_icount++; + nd6_dad_rele(dp); } static void nd6_dad_na_input(struct ifaddr *ifa) { struct dadq *dp; if (ifa == NULL) panic("ifa == NULL in nd6_dad_na_input"); dp = nd6_dad_find(ifa); - if (dp) + if (dp != NULL) { dp->dad_na_icount++; - - /* remove the address. */ - nd6_dad_duplicated(ifa, dp); + nd6_dad_rele(dp); + } } Index: projects/building-blocks/sys/ofed/include/linux/list.h =================================================================== --- projects/building-blocks/sys/ofed/include/linux/list.h (revision 275663) +++ projects/building-blocks/sys/ofed/include/linux/list.h (revision 275664) @@ -1,398 +1,400 @@ /*- * Copyright (c) 2010 Isilon Systems, Inc. * Copyright (c) 2010 iX Systems, Inc. * Copyright (c) 2010 Panasas, Inc. * Copyright (c) 2013, 2014 Mellanox Technologies, Ltd. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice unmodified, this list of conditions, and the following * disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #ifndef _LINUX_LIST_H_ #define _LINUX_LIST_H_ /* * Since LIST_HEAD conflicts with the linux definition we must include any * FreeBSD header which requires it here so it is resolved with the correct * definition prior to the undef. */ #include #include #include #include #include +#include #include #include #include #include #include #include #include #include #include #include #include #include +#include #include #include #include #include #include #include #define prefetch(x) struct list_head { struct list_head *next; struct list_head *prev; }; static inline void INIT_LIST_HEAD(struct list_head *list) { list->next = list->prev = list; } static inline int list_empty(const struct list_head *head) { return (head->next == head); } static inline void list_del(struct list_head *entry) { entry->next->prev = entry->prev; entry->prev->next = entry->next; } static inline void _list_add(struct list_head *new, struct list_head *prev, struct list_head *next) { next->prev = new; new->next = next; new->prev = prev; prev->next = new; } static inline void list_del_init(struct list_head *entry) { list_del(entry); INIT_LIST_HEAD(entry); } #define list_entry(ptr, type, field) container_of(ptr, type, field) #define list_first_entry(ptr, type, member) \ list_entry((ptr)->next, type, member) #define list_for_each(p, head) \ for (p = (head)->next; p != (head); p = p->next) #define list_for_each_safe(p, n, head) \ for (p = (head)->next, n = p->next; p != (head); p = n, n = p->next) #define list_for_each_entry(p, h, field) \ for (p = list_entry((h)->next, typeof(*p), field); &p->field != (h); \ p = list_entry(p->field.next, typeof(*p), field)) #define list_for_each_entry_safe(p, n, h, field) \ for (p = list_entry((h)->next, typeof(*p), field), \ n = list_entry(p->field.next, typeof(*p), field); &p->field != (h);\ p = n, n = list_entry(n->field.next, typeof(*n), field)) #define list_for_each_entry_reverse(p, h, field) \ for (p = list_entry((h)->prev, typeof(*p), field); &p->field != (h); \ p = list_entry(p->field.prev, typeof(*p), field)) #define list_for_each_prev(p, h) for (p = (h)->prev; p != (h); p = p->prev) static inline void list_add(struct list_head *new, struct list_head *head) { _list_add(new, head, head->next); } static inline void list_add_tail(struct list_head *new, struct list_head *head) { _list_add(new, head->prev, head); } static inline void list_move(struct list_head *list, struct list_head *head) { list_del(list); list_add(list, head); } static inline void list_move_tail(struct list_head *entry, struct list_head *head) { list_del(entry); list_add_tail(entry, head); } static inline void _list_splice(const struct list_head *list, struct list_head *prev, struct list_head *next) { struct list_head *first; struct list_head *last; if (list_empty(list)) return; first = list->next; last = list->prev; first->prev = prev; prev->next = first; last->next = next; next->prev = last; } static inline void list_splice(const struct list_head *list, struct list_head *head) { _list_splice(list, head, head->next); } static inline void list_splice_tail(struct list_head *list, struct list_head *head) { _list_splice(list, head->prev, head); } static inline void list_splice_init(struct list_head *list, struct list_head *head) { _list_splice(list, head, head->next); INIT_LIST_HEAD(list); } static inline void list_splice_tail_init(struct list_head *list, struct list_head *head) { _list_splice(list, head->prev, head); INIT_LIST_HEAD(list); } #undef LIST_HEAD #define LIST_HEAD(name) struct list_head name = { &(name), &(name) } struct hlist_head { struct hlist_node *first; }; struct hlist_node { struct hlist_node *next, **pprev; }; #define HLIST_HEAD_INIT { } #define HLIST_HEAD(name) struct hlist_head name = HLIST_HEAD_INIT #define INIT_HLIST_HEAD(head) (head)->first = NULL #define INIT_HLIST_NODE(node) \ do { \ (node)->next = NULL; \ (node)->pprev = NULL; \ } while (0) static inline int hlist_unhashed(const struct hlist_node *h) { return !h->pprev; } static inline int hlist_empty(const struct hlist_head *h) { return !h->first; } static inline void hlist_del(struct hlist_node *n) { if (n->next) n->next->pprev = n->pprev; *n->pprev = n->next; } static inline void hlist_del_init(struct hlist_node *n) { if (hlist_unhashed(n)) return; hlist_del(n); INIT_HLIST_NODE(n); } static inline void hlist_add_head(struct hlist_node *n, struct hlist_head *h) { n->next = h->first; if (h->first) h->first->pprev = &n->next; h->first = n; n->pprev = &h->first; } static inline void hlist_add_before(struct hlist_node *n, struct hlist_node *next) { n->pprev = next->pprev; n->next = next; next->pprev = &n->next; *(n->pprev) = n; } static inline void hlist_add_after(struct hlist_node *n, struct hlist_node *next) { next->next = n->next; n->next = next; next->pprev = &n->next; if (next->next) next->next->pprev = &next->next; } static inline void hlist_move_list(struct hlist_head *old, struct hlist_head *new) { new->first = old->first; if (new->first) new->first->pprev = &new->first; old->first = NULL; } /** * list_is_singular - tests whether a list has just one entry. * @head: the list to test. */ static inline int list_is_singular(const struct list_head *head) { return !list_empty(head) && (head->next == head->prev); } static inline void __list_cut_position(struct list_head *list, struct list_head *head, struct list_head *entry) { struct list_head *new_first = entry->next; list->next = head->next; list->next->prev = list; list->prev = entry; entry->next = list; head->next = new_first; new_first->prev = head; } /** * list_cut_position - cut a list into two * @list: a new list to add all removed entries * @head: a list with entries * @entry: an entry within head, could be the head itself * and if so we won't cut the list * * This helper moves the initial part of @head, up to and * including @entry, from @head to @list. You should * pass on @entry an element you know is on @head. @list * should be an empty list or a list you do not care about * losing its data. * */ static inline void list_cut_position(struct list_head *list, struct list_head *head, struct list_head *entry) { if (list_empty(head)) return; if (list_is_singular(head) && (head->next != entry && head != entry)) return; if (entry == head) INIT_LIST_HEAD(list); else __list_cut_position(list, head, entry); } /** * list_is_last - tests whether @list is the last entry in list @head * @list: the entry to test * @head: the head of the list */ static inline int list_is_last(const struct list_head *list, const struct list_head *head) { return list->next == head; } #define hlist_entry(ptr, type, field) container_of(ptr, type, field) #define hlist_for_each(p, head) \ for (p = (head)->first; p; p = p->next) #define hlist_for_each_safe(p, n, head) \ for (p = (head)->first; p && ({ n = p->next; 1; }); p = n) #define hlist_for_each_entry(tp, p, head, field) \ for (p = (head)->first; \ p ? (tp = hlist_entry(p, typeof(*tp), field)): NULL; p = p->next) #define hlist_for_each_entry_continue(tp, p, field) \ for (p = (p)->next; \ p ? (tp = hlist_entry(p, typeof(*tp), field)): NULL; p = p->next) #define hlist_for_each_entry_from(tp, p, field) \ for (; p ? (tp = hlist_entry(p, typeof(*tp), field)): NULL; p = p->next) #define hlist_for_each_entry_safe(tpos, pos, n, head, member) \ for (pos = (head)->first; \ (pos) != 0 && ({ n = (pos)->next; \ tpos = hlist_entry((pos), typeof(*(tpos)), member); 1;}); \ pos = (n)) #endif /* _LINUX_LIST_H_ */ Index: projects/building-blocks/sys/ofed/include/linux/module.h =================================================================== --- projects/building-blocks/sys/ofed/include/linux/module.h (revision 275663) +++ projects/building-blocks/sys/ofed/include/linux/module.h (revision 275664) @@ -1,99 +1,99 @@ /*- * Copyright (c) 2010 Isilon Systems, Inc. * Copyright (c) 2010 iX Systems, Inc. * Copyright (c) 2010 Panasas, Inc. * Copyright (c) 2013, 2014 Mellanox Technologies, Ltd. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice unmodified, this list of conditions, and the following * disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #ifndef _LINUX_MODULE_H_ #define _LINUX_MODULE_H_ #include #include #include #include #include #define MODULE_AUTHOR(name) #define MODULE_DESCRIPTION(name) #define MODULE_LICENSE(name) #ifndef MODULE_VERSION #define MODULE_VERSION(name) #endif #define THIS_MODULE ((struct module *)0) #define EXPORT_SYMBOL(name) #define EXPORT_SYMBOL_GPL(name) /* OFED pre-module initialization */ -#define SI_SUB_OFED_PREINIT (SI_SUB_KTHREAD_INIT - 2) +#define SI_SUB_OFED_PREINIT (SI_SUB_ROOT_CONF - 2) /* OFED default module initialization */ -#define SI_SUB_OFED_MODINIT (SI_SUB_KTHREAD_INIT - 1) +#define SI_SUB_OFED_MODINIT (SI_SUB_ROOT_CONF - 1) #include static inline void _module_run(void *arg) { void (*fn)(void); #ifdef OFED_DEBUG_INIT char name[1024]; caddr_t pc; long offset; pc = (caddr_t)arg; if (linker_search_symbol_name(pc, name, sizeof(name), &offset) != 0) printf("Running ??? (%p)\n", pc); else printf("Running %s (%p)\n", name, pc); #endif fn = arg; DROP_GIANT(); fn(); PICKUP_GIANT(); } #define module_init(fn) \ SYSINIT(fn, SI_SUB_OFED_MODINIT, SI_ORDER_FIRST, _module_run, (fn)) #define module_exit(fn) \ SYSUNINIT(fn, SI_SUB_OFED_MODINIT, SI_ORDER_SECOND, _module_run, (fn)) /* * The following two macros are a workaround for not having a module * load and unload order resolver: */ #define module_init_order(fn, order) \ SYSINIT(fn, SI_SUB_OFED_MODINIT, (order), _module_run, (fn)) #define module_exit_order(fn, order) \ SYSUNINIT(fn, SI_SUB_OFED_MODINIT, (order), _module_run, (fn)) #define module_get(module) #define module_put(module) #define try_module_get(module) 1 #endif /* _LINUX_MODULE_H_ */ Index: projects/building-blocks/sys/ofed/include/net/ip.h =================================================================== --- projects/building-blocks/sys/ofed/include/net/ip.h (revision 275663) +++ projects/building-blocks/sys/ofed/include/net/ip.h (revision 275664) @@ -1,82 +1,84 @@ /*- * Copyright (c) 2010 Isilon Systems, Inc. * Copyright (c) 2010 iX Systems, Inc. * Copyright (c) 2010 Panasas, Inc. * Copyright (c) 2013, 2014 Mellanox Technologies, Ltd. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice unmodified, this list of conditions, and the following * disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #ifndef _LINUX_NET_IP_H_ #define _LINUX_NET_IP_H_ #include "opt_inet.h" #include #include #include #include #include #include #include #ifdef INET static inline void inet_get_local_port_range(int *low, int *high) { + CURVNET_SET_QUIET(TD_TO_VNET(curthread)); *low = V_ipport_firstauto; *high = V_ipport_lastauto; + CURVNET_RESTORE(); } static inline void ip_ib_mc_map(uint32_t addr, const unsigned char *bcast, char *buf) { unsigned char scope; addr = ntohl(addr); scope = bcast[5] & 0xF; buf[0] = 0; buf[1] = 0xff; buf[2] = 0xff; buf[3] = 0xff; buf[4] = 0xff; buf[5] = 0x10 | scope; buf[6] = 0x40; buf[7] = 0x1b; buf[8] = bcast[8]; buf[9] = bcast[9]; buf[10] = 0; buf[11] = 0; buf[12] = 0; buf[13] = 0; buf[14] = 0; buf[15] = 0; buf[16] = (addr >> 24) & 0x0f; buf[17] = (addr >> 16) & 0xff; buf[18] = (addr >> 8) & 0xff; buf[19] = addr & 0xff; } #endif #endif /* _LINUX_NET_IP_H_ */ Index: projects/building-blocks/sys/rpc/svc.c =================================================================== --- projects/building-blocks/sys/rpc/svc.c (revision 275663) +++ projects/building-blocks/sys/rpc/svc.c (revision 275664) @@ -1,1422 +1,1434 @@ /* $NetBSD: svc.c,v 1.21 2000/07/06 03:10:35 christos Exp $ */ /*- * Copyright (c) 2009, Sun Microsystems, Inc. * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are met: * - Redistributions of source code must retain the above copyright notice, * this list of conditions and the following disclaimer. * - Redistributions in binary form must reproduce the above copyright notice, * this list of conditions and the following disclaimer in the documentation * and/or other materials provided with the distribution. * - Neither the name of Sun Microsystems, Inc. nor the names of its * contributors may be used to endorse or promote products derived * from this software without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE * POSSIBILITY OF SUCH DAMAGE. */ #if defined(LIBC_SCCS) && !defined(lint) static char *sccsid2 = "@(#)svc.c 1.44 88/02/08 Copyr 1984 Sun Micro"; static char *sccsid = "@(#)svc.c 2.4 88/08/11 4.0 RPCSRC"; #endif #include __FBSDID("$FreeBSD$"); /* * svc.c, Server-side remote procedure call interface. * * There are two sets of procedures here. The xprt routines are * for handling transport handles. The svc routines handle the * list of service routines. * * Copyright (C) 1984, Sun Microsystems, Inc. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define SVC_VERSQUIET 0x0001 /* keep quiet about vers mismatch */ #define version_keepquiet(xp) (SVC_EXT(xp)->xp_flags & SVC_VERSQUIET) static struct svc_callout *svc_find(SVCPOOL *pool, rpcprog_t, rpcvers_t, char *); static void svc_new_thread(SVCGROUP *grp); static void xprt_unregister_locked(SVCXPRT *xprt); static void svc_change_space_used(SVCPOOL *pool, int delta); static bool_t svc_request_space_available(SVCPOOL *pool); /* *************** SVCXPRT related stuff **************** */ static int svcpool_minthread_sysctl(SYSCTL_HANDLER_ARGS); static int svcpool_maxthread_sysctl(SYSCTL_HANDLER_ARGS); static int svcpool_threads_sysctl(SYSCTL_HANDLER_ARGS); SVCPOOL* svcpool_create(const char *name, struct sysctl_oid_list *sysctl_base) { SVCPOOL *pool; SVCGROUP *grp; int g; pool = malloc(sizeof(SVCPOOL), M_RPC, M_WAITOK|M_ZERO); mtx_init(&pool->sp_lock, "sp_lock", NULL, MTX_DEF); pool->sp_name = name; pool->sp_state = SVCPOOL_INIT; pool->sp_proc = NULL; TAILQ_INIT(&pool->sp_callouts); TAILQ_INIT(&pool->sp_lcallouts); pool->sp_minthreads = 1; pool->sp_maxthreads = 1; pool->sp_groupcount = 1; for (g = 0; g < SVC_MAXGROUPS; g++) { grp = &pool->sp_groups[g]; mtx_init(&grp->sg_lock, "sg_lock", NULL, MTX_DEF); grp->sg_pool = pool; grp->sg_state = SVCPOOL_ACTIVE; TAILQ_INIT(&grp->sg_xlist); TAILQ_INIT(&grp->sg_active); LIST_INIT(&grp->sg_idlethreads); grp->sg_minthreads = 1; grp->sg_maxthreads = 1; } /* * Don't use more than a quarter of mbuf clusters or more than * 45Mb buffering requests. */ pool->sp_space_high = nmbclusters * MCLBYTES / 4; if (pool->sp_space_high > 45 << 20) pool->sp_space_high = 45 << 20; pool->sp_space_low = 2 * pool->sp_space_high / 3; sysctl_ctx_init(&pool->sp_sysctl); if (sysctl_base) { SYSCTL_ADD_PROC(&pool->sp_sysctl, sysctl_base, OID_AUTO, "minthreads", CTLTYPE_INT | CTLFLAG_RW, pool, 0, svcpool_minthread_sysctl, "I", "Minimal number of threads"); SYSCTL_ADD_PROC(&pool->sp_sysctl, sysctl_base, OID_AUTO, "maxthreads", CTLTYPE_INT | CTLFLAG_RW, pool, 0, svcpool_maxthread_sysctl, "I", "Maximal number of threads"); SYSCTL_ADD_PROC(&pool->sp_sysctl, sysctl_base, OID_AUTO, "threads", CTLTYPE_INT | CTLFLAG_RD, pool, 0, svcpool_threads_sysctl, "I", "Current number of threads"); SYSCTL_ADD_INT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "groups", CTLFLAG_RD, &pool->sp_groupcount, 0, "Number of thread groups"); SYSCTL_ADD_UINT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "request_space_used", CTLFLAG_RD, &pool->sp_space_used, 0, "Space in parsed but not handled requests."); SYSCTL_ADD_UINT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "request_space_used_highest", CTLFLAG_RD, &pool->sp_space_used_highest, 0, "Highest space used since reboot."); SYSCTL_ADD_UINT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "request_space_high", CTLFLAG_RW, &pool->sp_space_high, 0, "Maximum space in parsed but not handled requests."); SYSCTL_ADD_UINT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "request_space_low", CTLFLAG_RW, &pool->sp_space_low, 0, "Low water mark for request space."); SYSCTL_ADD_INT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "request_space_throttled", CTLFLAG_RD, &pool->sp_space_throttled, 0, "Whether nfs requests are currently throttled"); SYSCTL_ADD_INT(&pool->sp_sysctl, sysctl_base, OID_AUTO, "request_space_throttle_count", CTLFLAG_RD, &pool->sp_space_throttle_count, 0, "Count of times throttling based on request space has occurred"); } return pool; } void svcpool_destroy(SVCPOOL *pool) { SVCGROUP *grp; SVCXPRT *xprt, *nxprt; struct svc_callout *s; struct svc_loss_callout *sl; struct svcxprt_list cleanup; int g; TAILQ_INIT(&cleanup); for (g = 0; g < SVC_MAXGROUPS; g++) { grp = &pool->sp_groups[g]; mtx_lock(&grp->sg_lock); while ((xprt = TAILQ_FIRST(&grp->sg_xlist)) != NULL) { xprt_unregister_locked(xprt); TAILQ_INSERT_TAIL(&cleanup, xprt, xp_link); } mtx_unlock(&grp->sg_lock); } TAILQ_FOREACH_SAFE(xprt, &cleanup, xp_link, nxprt) { SVC_RELEASE(xprt); } mtx_lock(&pool->sp_lock); while ((s = TAILQ_FIRST(&pool->sp_callouts)) != NULL) { mtx_unlock(&pool->sp_lock); svc_unreg(pool, s->sc_prog, s->sc_vers); mtx_lock(&pool->sp_lock); } while ((sl = TAILQ_FIRST(&pool->sp_lcallouts)) != NULL) { mtx_unlock(&pool->sp_lock); svc_loss_unreg(pool, sl->slc_dispatch); mtx_lock(&pool->sp_lock); } mtx_unlock(&pool->sp_lock); for (g = 0; g < SVC_MAXGROUPS; g++) { grp = &pool->sp_groups[g]; mtx_destroy(&grp->sg_lock); } mtx_destroy(&pool->sp_lock); if (pool->sp_rcache) replay_freecache(pool->sp_rcache); sysctl_ctx_free(&pool->sp_sysctl); free(pool, M_RPC); } /* * Sysctl handler to get the present thread count on a pool */ static int svcpool_threads_sysctl(SYSCTL_HANDLER_ARGS) { SVCPOOL *pool; int threads, error, g; pool = oidp->oid_arg1; threads = 0; mtx_lock(&pool->sp_lock); for (g = 0; g < pool->sp_groupcount; g++) threads += pool->sp_groups[g].sg_threadcount; mtx_unlock(&pool->sp_lock); error = sysctl_handle_int(oidp, &threads, 0, req); return (error); } /* * Sysctl handler to set the minimum thread count on a pool */ static int svcpool_minthread_sysctl(SYSCTL_HANDLER_ARGS) { SVCPOOL *pool; int newminthreads, error, g; pool = oidp->oid_arg1; newminthreads = pool->sp_minthreads; error = sysctl_handle_int(oidp, &newminthreads, 0, req); if (error == 0 && newminthreads != pool->sp_minthreads) { if (newminthreads > pool->sp_maxthreads) return (EINVAL); mtx_lock(&pool->sp_lock); pool->sp_minthreads = newminthreads; for (g = 0; g < pool->sp_groupcount; g++) { pool->sp_groups[g].sg_minthreads = max(1, pool->sp_minthreads / pool->sp_groupcount); } mtx_unlock(&pool->sp_lock); } return (error); } /* * Sysctl handler to set the maximum thread count on a pool */ static int svcpool_maxthread_sysctl(SYSCTL_HANDLER_ARGS) { SVCPOOL *pool; int newmaxthreads, error, g; pool = oidp->oid_arg1; newmaxthreads = pool->sp_maxthreads; error = sysctl_handle_int(oidp, &newmaxthreads, 0, req); if (error == 0 && newmaxthreads != pool->sp_maxthreads) { if (newmaxthreads < pool->sp_minthreads) return (EINVAL); mtx_lock(&pool->sp_lock); pool->sp_maxthreads = newmaxthreads; for (g = 0; g < pool->sp_groupcount; g++) { pool->sp_groups[g].sg_maxthreads = max(1, pool->sp_maxthreads / pool->sp_groupcount); } mtx_unlock(&pool->sp_lock); } return (error); } /* * Activate a transport handle. */ void xprt_register(SVCXPRT *xprt) { SVCPOOL *pool = xprt->xp_pool; SVCGROUP *grp; int g; SVC_ACQUIRE(xprt); g = atomic_fetchadd_int(&pool->sp_nextgroup, 1) % pool->sp_groupcount; xprt->xp_group = grp = &pool->sp_groups[g]; mtx_lock(&grp->sg_lock); xprt->xp_registered = TRUE; xprt->xp_active = FALSE; TAILQ_INSERT_TAIL(&grp->sg_xlist, xprt, xp_link); mtx_unlock(&grp->sg_lock); } /* * De-activate a transport handle. Note: the locked version doesn't * release the transport - caller must do that after dropping the pool * lock. */ static void xprt_unregister_locked(SVCXPRT *xprt) { SVCGROUP *grp = xprt->xp_group; mtx_assert(&grp->sg_lock, MA_OWNED); KASSERT(xprt->xp_registered == TRUE, ("xprt_unregister_locked: not registered")); xprt_inactive_locked(xprt); TAILQ_REMOVE(&grp->sg_xlist, xprt, xp_link); xprt->xp_registered = FALSE; } void xprt_unregister(SVCXPRT *xprt) { SVCGROUP *grp = xprt->xp_group; mtx_lock(&grp->sg_lock); if (xprt->xp_registered == FALSE) { /* Already unregistered by another thread */ mtx_unlock(&grp->sg_lock); return; } xprt_unregister_locked(xprt); mtx_unlock(&grp->sg_lock); SVC_RELEASE(xprt); } /* * Attempt to assign a service thread to this transport. */ static int xprt_assignthread(SVCXPRT *xprt) { SVCGROUP *grp = xprt->xp_group; SVCTHREAD *st; mtx_assert(&grp->sg_lock, MA_OWNED); st = LIST_FIRST(&grp->sg_idlethreads); if (st) { LIST_REMOVE(st, st_ilink); SVC_ACQUIRE(xprt); xprt->xp_thread = st; st->st_xprt = xprt; cv_signal(&st->st_cond); return (TRUE); } else { /* * See if we can create a new thread. The * actual thread creation happens in * svc_run_internal because our locking state * is poorly defined (we are typically called * from a socket upcall). Don't create more * than one thread per second. */ if (grp->sg_state == SVCPOOL_ACTIVE && grp->sg_lastcreatetime < time_uptime && grp->sg_threadcount < grp->sg_maxthreads) { grp->sg_state = SVCPOOL_THREADWANTED; } } return (FALSE); } void xprt_active(SVCXPRT *xprt) { SVCGROUP *grp = xprt->xp_group; mtx_lock(&grp->sg_lock); if (!xprt->xp_registered) { /* * Race with xprt_unregister - we lose. */ mtx_unlock(&grp->sg_lock); return; } if (!xprt->xp_active) { xprt->xp_active = TRUE; if (xprt->xp_thread == NULL) { if (!svc_request_space_available(xprt->xp_pool) || !xprt_assignthread(xprt)) TAILQ_INSERT_TAIL(&grp->sg_active, xprt, xp_alink); } } mtx_unlock(&grp->sg_lock); } void xprt_inactive_locked(SVCXPRT *xprt) { SVCGROUP *grp = xprt->xp_group; mtx_assert(&grp->sg_lock, MA_OWNED); if (xprt->xp_active) { if (xprt->xp_thread == NULL) TAILQ_REMOVE(&grp->sg_active, xprt, xp_alink); xprt->xp_active = FALSE; } } void xprt_inactive(SVCXPRT *xprt) { SVCGROUP *grp = xprt->xp_group; mtx_lock(&grp->sg_lock); xprt_inactive_locked(xprt); mtx_unlock(&grp->sg_lock); } /* * Variant of xprt_inactive() for use only when sure that port is * assigned to thread. For example, withing receive handlers. */ void xprt_inactive_self(SVCXPRT *xprt) { KASSERT(xprt->xp_thread != NULL, ("xprt_inactive_self(%p) with NULL xp_thread", xprt)); xprt->xp_active = FALSE; } /* * Add a service program to the callout list. * The dispatch routine will be called when a rpc request for this * program number comes in. */ bool_t svc_reg(SVCXPRT *xprt, const rpcprog_t prog, const rpcvers_t vers, void (*dispatch)(struct svc_req *, SVCXPRT *), const struct netconfig *nconf) { SVCPOOL *pool = xprt->xp_pool; struct svc_callout *s; char *netid = NULL; int flag = 0; /* VARIABLES PROTECTED BY svc_lock: s, svc_head */ if (xprt->xp_netid) { netid = strdup(xprt->xp_netid, M_RPC); flag = 1; } else if (nconf && nconf->nc_netid) { netid = strdup(nconf->nc_netid, M_RPC); flag = 1; } /* must have been created with svc_raw_create */ if ((netid == NULL) && (flag == 1)) { return (FALSE); } mtx_lock(&pool->sp_lock); if ((s = svc_find(pool, prog, vers, netid)) != NULL) { if (netid) free(netid, M_RPC); if (s->sc_dispatch == dispatch) goto rpcb_it; /* he is registering another xptr */ mtx_unlock(&pool->sp_lock); return (FALSE); } s = malloc(sizeof (struct svc_callout), M_RPC, M_NOWAIT); if (s == NULL) { if (netid) free(netid, M_RPC); mtx_unlock(&pool->sp_lock); return (FALSE); } s->sc_prog = prog; s->sc_vers = vers; s->sc_dispatch = dispatch; s->sc_netid = netid; TAILQ_INSERT_TAIL(&pool->sp_callouts, s, sc_link); if ((xprt->xp_netid == NULL) && (flag == 1) && netid) ((SVCXPRT *) xprt)->xp_netid = strdup(netid, M_RPC); rpcb_it: mtx_unlock(&pool->sp_lock); /* now register the information with the local binder service */ if (nconf) { bool_t dummy; struct netconfig tnc; struct netbuf nb; tnc = *nconf; nb.buf = &xprt->xp_ltaddr; nb.len = xprt->xp_ltaddr.ss_len; dummy = rpcb_set(prog, vers, &tnc, &nb); return (dummy); } return (TRUE); } /* * Remove a service program from the callout list. */ void svc_unreg(SVCPOOL *pool, const rpcprog_t prog, const rpcvers_t vers) { struct svc_callout *s; /* unregister the information anyway */ (void) rpcb_unset(prog, vers, NULL); mtx_lock(&pool->sp_lock); while ((s = svc_find(pool, prog, vers, NULL)) != NULL) { TAILQ_REMOVE(&pool->sp_callouts, s, sc_link); if (s->sc_netid) mem_free(s->sc_netid, sizeof (s->sc_netid) + 1); mem_free(s, sizeof (struct svc_callout)); } mtx_unlock(&pool->sp_lock); } /* * Add a service connection loss program to the callout list. * The dispatch routine will be called when some port in ths pool die. */ bool_t svc_loss_reg(SVCXPRT *xprt, void (*dispatch)(SVCXPRT *)) { SVCPOOL *pool = xprt->xp_pool; struct svc_loss_callout *s; mtx_lock(&pool->sp_lock); TAILQ_FOREACH(s, &pool->sp_lcallouts, slc_link) { if (s->slc_dispatch == dispatch) break; } if (s != NULL) { mtx_unlock(&pool->sp_lock); return (TRUE); } s = malloc(sizeof (struct svc_callout), M_RPC, M_NOWAIT); if (s == NULL) { mtx_unlock(&pool->sp_lock); return (FALSE); } s->slc_dispatch = dispatch; TAILQ_INSERT_TAIL(&pool->sp_lcallouts, s, slc_link); mtx_unlock(&pool->sp_lock); return (TRUE); } /* * Remove a service connection loss program from the callout list. */ void svc_loss_unreg(SVCPOOL *pool, void (*dispatch)(SVCXPRT *)) { struct svc_loss_callout *s; mtx_lock(&pool->sp_lock); TAILQ_FOREACH(s, &pool->sp_lcallouts, slc_link) { if (s->slc_dispatch == dispatch) { TAILQ_REMOVE(&pool->sp_lcallouts, s, slc_link); free(s, M_RPC); break; } } mtx_unlock(&pool->sp_lock); } /* ********************** CALLOUT list related stuff ************* */ /* * Search the callout list for a program number, return the callout * struct. */ static struct svc_callout * svc_find(SVCPOOL *pool, rpcprog_t prog, rpcvers_t vers, char *netid) { struct svc_callout *s; mtx_assert(&pool->sp_lock, MA_OWNED); TAILQ_FOREACH(s, &pool->sp_callouts, sc_link) { if (s->sc_prog == prog && s->sc_vers == vers && (netid == NULL || s->sc_netid == NULL || strcmp(netid, s->sc_netid) == 0)) break; } return (s); } /* ******************* REPLY GENERATION ROUTINES ************ */ static bool_t svc_sendreply_common(struct svc_req *rqstp, struct rpc_msg *rply, struct mbuf *body) { SVCXPRT *xprt = rqstp->rq_xprt; bool_t ok; if (rqstp->rq_args) { m_freem(rqstp->rq_args); rqstp->rq_args = NULL; } if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, rply, svc_getrpccaller(rqstp), body); if (!SVCAUTH_WRAP(&rqstp->rq_auth, &body)) return (FALSE); ok = SVC_REPLY(xprt, rply, rqstp->rq_addr, body, &rqstp->rq_reply_seq); if (rqstp->rq_addr) { free(rqstp->rq_addr, M_SONAME); rqstp->rq_addr = NULL; } return (ok); } /* * Send a reply to an rpc request */ bool_t svc_sendreply(struct svc_req *rqstp, xdrproc_t xdr_results, void * xdr_location) { struct rpc_msg rply; struct mbuf *m; XDR xdrs; bool_t ok; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = SUCCESS; rply.acpted_rply.ar_results.where = NULL; rply.acpted_rply.ar_results.proc = (xdrproc_t) xdr_void; m = m_getcl(M_WAITOK, MT_DATA, 0); xdrmbuf_create(&xdrs, m, XDR_ENCODE); ok = xdr_results(&xdrs, xdr_location); XDR_DESTROY(&xdrs); if (ok) { return (svc_sendreply_common(rqstp, &rply, m)); } else { m_freem(m); return (FALSE); } } bool_t svc_sendreply_mbuf(struct svc_req *rqstp, struct mbuf *m) { struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = SUCCESS; rply.acpted_rply.ar_results.where = NULL; rply.acpted_rply.ar_results.proc = (xdrproc_t) xdr_void; return (svc_sendreply_common(rqstp, &rply, m)); } /* * No procedure error reply */ void svcerr_noproc(struct svc_req *rqstp) { SVCXPRT *xprt = rqstp->rq_xprt; struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = PROC_UNAVAIL; if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, &rply, svc_getrpccaller(rqstp), NULL); svc_sendreply_common(rqstp, &rply, NULL); } /* * Can't decode args error reply */ void svcerr_decode(struct svc_req *rqstp) { SVCXPRT *xprt = rqstp->rq_xprt; struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = GARBAGE_ARGS; if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, &rply, (struct sockaddr *) &xprt->xp_rtaddr, NULL); svc_sendreply_common(rqstp, &rply, NULL); } /* * Some system error */ void svcerr_systemerr(struct svc_req *rqstp) { SVCXPRT *xprt = rqstp->rq_xprt; struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = SYSTEM_ERR; if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, &rply, svc_getrpccaller(rqstp), NULL); svc_sendreply_common(rqstp, &rply, NULL); } /* * Authentication error reply */ void svcerr_auth(struct svc_req *rqstp, enum auth_stat why) { SVCXPRT *xprt = rqstp->rq_xprt; struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_DENIED; rply.rjcted_rply.rj_stat = AUTH_ERROR; rply.rjcted_rply.rj_why = why; if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, &rply, svc_getrpccaller(rqstp), NULL); svc_sendreply_common(rqstp, &rply, NULL); } /* * Auth too weak error reply */ void svcerr_weakauth(struct svc_req *rqstp) { svcerr_auth(rqstp, AUTH_TOOWEAK); } /* * Program unavailable error reply */ void svcerr_noprog(struct svc_req *rqstp) { SVCXPRT *xprt = rqstp->rq_xprt; struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = PROG_UNAVAIL; if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, &rply, svc_getrpccaller(rqstp), NULL); svc_sendreply_common(rqstp, &rply, NULL); } /* * Program version mismatch error reply */ void svcerr_progvers(struct svc_req *rqstp, rpcvers_t low_vers, rpcvers_t high_vers) { SVCXPRT *xprt = rqstp->rq_xprt; struct rpc_msg rply; rply.rm_xid = rqstp->rq_xid; rply.rm_direction = REPLY; rply.rm_reply.rp_stat = MSG_ACCEPTED; rply.acpted_rply.ar_verf = rqstp->rq_verf; rply.acpted_rply.ar_stat = PROG_MISMATCH; rply.acpted_rply.ar_vers.low = (uint32_t)low_vers; rply.acpted_rply.ar_vers.high = (uint32_t)high_vers; if (xprt->xp_pool->sp_rcache) replay_setreply(xprt->xp_pool->sp_rcache, &rply, svc_getrpccaller(rqstp), NULL); svc_sendreply_common(rqstp, &rply, NULL); } /* * Allocate a new server transport structure. All fields are * initialized to zero and xp_p3 is initialized to point at an * extension structure to hold various flags and authentication * parameters. */ SVCXPRT * svc_xprt_alloc() { SVCXPRT *xprt; SVCXPRT_EXT *ext; xprt = mem_alloc(sizeof(SVCXPRT)); memset(xprt, 0, sizeof(SVCXPRT)); ext = mem_alloc(sizeof(SVCXPRT_EXT)); memset(ext, 0, sizeof(SVCXPRT_EXT)); xprt->xp_p3 = ext; refcount_init(&xprt->xp_refs, 1); return (xprt); } /* * Free a server transport structure. */ void svc_xprt_free(xprt) SVCXPRT *xprt; { mem_free(xprt->xp_p3, sizeof(SVCXPRT_EXT)); mem_free(xprt, sizeof(SVCXPRT)); } /* ******************* SERVER INPUT STUFF ******************* */ /* * Read RPC requests from a transport and queue them to be * executed. We handle authentication and replay cache replies here. * Actually dispatching the RPC is deferred till svc_executereq. */ static enum xprt_stat svc_getreq(SVCXPRT *xprt, struct svc_req **rqstp_ret) { SVCPOOL *pool = xprt->xp_pool; struct svc_req *r; struct rpc_msg msg; struct mbuf *args; struct svc_loss_callout *s; enum xprt_stat stat; /* now receive msgs from xprtprt (support batch calls) */ r = malloc(sizeof(*r), M_RPC, M_WAITOK|M_ZERO); msg.rm_call.cb_cred.oa_base = r->rq_credarea; msg.rm_call.cb_verf.oa_base = &r->rq_credarea[MAX_AUTH_BYTES]; r->rq_clntcred = &r->rq_credarea[2*MAX_AUTH_BYTES]; if (SVC_RECV(xprt, &msg, &r->rq_addr, &args)) { enum auth_stat why; /* * Handle replays and authenticate before queuing the * request to be executed. */ SVC_ACQUIRE(xprt); r->rq_xprt = xprt; if (pool->sp_rcache) { struct rpc_msg repmsg; struct mbuf *repbody; enum replay_state rs; rs = replay_find(pool->sp_rcache, &msg, svc_getrpccaller(r), &repmsg, &repbody); switch (rs) { case RS_NEW: break; case RS_DONE: SVC_REPLY(xprt, &repmsg, r->rq_addr, repbody, &r->rq_reply_seq); if (r->rq_addr) { free(r->rq_addr, M_SONAME); r->rq_addr = NULL; } m_freem(args); goto call_done; default: m_freem(args); goto call_done; } } r->rq_xid = msg.rm_xid; r->rq_prog = msg.rm_call.cb_prog; r->rq_vers = msg.rm_call.cb_vers; r->rq_proc = msg.rm_call.cb_proc; r->rq_size = sizeof(*r) + m_length(args, NULL); r->rq_args = args; if ((why = _authenticate(r, &msg)) != AUTH_OK) { /* * RPCSEC_GSS uses this return code * for requests that form part of its * context establishment protocol and * should not be dispatched to the * application. */ if (why != RPCSEC_GSS_NODISPATCH) svcerr_auth(r, why); goto call_done; } if (!SVCAUTH_UNWRAP(&r->rq_auth, &r->rq_args)) { svcerr_decode(r); goto call_done; } /* * Everything checks out, return request to caller. */ *rqstp_ret = r; r = NULL; } call_done: if (r) { svc_freereq(r); r = NULL; } if ((stat = SVC_STAT(xprt)) == XPRT_DIED) { TAILQ_FOREACH(s, &pool->sp_lcallouts, slc_link) (*s->slc_dispatch)(xprt); xprt_unregister(xprt); } return (stat); } static void svc_executereq(struct svc_req *rqstp) { SVCXPRT *xprt = rqstp->rq_xprt; SVCPOOL *pool = xprt->xp_pool; int prog_found; rpcvers_t low_vers; rpcvers_t high_vers; struct svc_callout *s; /* now match message with a registered service*/ prog_found = FALSE; low_vers = (rpcvers_t) -1L; high_vers = (rpcvers_t) 0L; TAILQ_FOREACH(s, &pool->sp_callouts, sc_link) { if (s->sc_prog == rqstp->rq_prog) { if (s->sc_vers == rqstp->rq_vers) { /* * We hand ownership of r to the * dispatch method - they must call * svc_freereq. */ (*s->sc_dispatch)(rqstp, xprt); return; } /* found correct version */ prog_found = TRUE; if (s->sc_vers < low_vers) low_vers = s->sc_vers; if (s->sc_vers > high_vers) high_vers = s->sc_vers; } /* found correct program */ } /* * if we got here, the program or version * is not served ... */ if (prog_found) svcerr_progvers(rqstp, low_vers, high_vers); else svcerr_noprog(rqstp); svc_freereq(rqstp); } static void svc_checkidle(SVCGROUP *grp) { SVCXPRT *xprt, *nxprt; time_t timo; struct svcxprt_list cleanup; TAILQ_INIT(&cleanup); TAILQ_FOREACH_SAFE(xprt, &grp->sg_xlist, xp_link, nxprt) { /* * Only some transports have idle timers. Don't time * something out which is just waking up. */ if (!xprt->xp_idletimeout || xprt->xp_thread) continue; timo = xprt->xp_lastactive + xprt->xp_idletimeout; if (time_uptime > timo) { xprt_unregister_locked(xprt); TAILQ_INSERT_TAIL(&cleanup, xprt, xp_link); } } mtx_unlock(&grp->sg_lock); TAILQ_FOREACH_SAFE(xprt, &cleanup, xp_link, nxprt) { SVC_RELEASE(xprt); } mtx_lock(&grp->sg_lock); } static void svc_assign_waiting_sockets(SVCPOOL *pool) { SVCGROUP *grp; SVCXPRT *xprt; int g; for (g = 0; g < pool->sp_groupcount; g++) { grp = &pool->sp_groups[g]; mtx_lock(&grp->sg_lock); while ((xprt = TAILQ_FIRST(&grp->sg_active)) != NULL) { if (xprt_assignthread(xprt)) TAILQ_REMOVE(&grp->sg_active, xprt, xp_alink); else break; } mtx_unlock(&grp->sg_lock); } } static void svc_change_space_used(SVCPOOL *pool, int delta) { unsigned int value; value = atomic_fetchadd_int(&pool->sp_space_used, delta) + delta; if (delta > 0) { if (value >= pool->sp_space_high && !pool->sp_space_throttled) { pool->sp_space_throttled = TRUE; pool->sp_space_throttle_count++; } if (value > pool->sp_space_used_highest) pool->sp_space_used_highest = value; } else { if (value < pool->sp_space_low && pool->sp_space_throttled) { pool->sp_space_throttled = FALSE; svc_assign_waiting_sockets(pool); } } } static bool_t svc_request_space_available(SVCPOOL *pool) { if (pool->sp_space_throttled) return (FALSE); return (TRUE); } static void svc_run_internal(SVCGROUP *grp, bool_t ismaster) { SVCPOOL *pool = grp->sg_pool; SVCTHREAD *st, *stpref; SVCXPRT *xprt; enum xprt_stat stat; struct svc_req *rqstp; + struct proc *p; size_t sz; int error; st = mem_alloc(sizeof(*st)); mtx_init(&st->st_lock, "st_lock", NULL, MTX_DEF); st->st_pool = pool; st->st_xprt = NULL; STAILQ_INIT(&st->st_reqs); cv_init(&st->st_cond, "rpcsvc"); mtx_lock(&grp->sg_lock); /* * If we are a new thread which was spawned to cope with * increased load, set the state back to SVCPOOL_ACTIVE. */ if (grp->sg_state == SVCPOOL_THREADSTARTING) grp->sg_state = SVCPOOL_ACTIVE; while (grp->sg_state != SVCPOOL_CLOSING) { /* * Create new thread if requested. */ if (grp->sg_state == SVCPOOL_THREADWANTED) { grp->sg_state = SVCPOOL_THREADSTARTING; grp->sg_lastcreatetime = time_uptime; mtx_unlock(&grp->sg_lock); svc_new_thread(grp); mtx_lock(&grp->sg_lock); continue; } /* * Check for idle transports once per second. */ if (time_uptime > grp->sg_lastidlecheck) { grp->sg_lastidlecheck = time_uptime; svc_checkidle(grp); } xprt = st->st_xprt; if (!xprt) { /* * Enforce maxthreads count. */ if (grp->sg_threadcount > grp->sg_maxthreads) break; /* * Before sleeping, see if we can find an * active transport which isn't being serviced * by a thread. */ if (svc_request_space_available(pool) && (xprt = TAILQ_FIRST(&grp->sg_active)) != NULL) { TAILQ_REMOVE(&grp->sg_active, xprt, xp_alink); SVC_ACQUIRE(xprt); xprt->xp_thread = st; st->st_xprt = xprt; continue; } LIST_INSERT_HEAD(&grp->sg_idlethreads, st, st_ilink); if (ismaster || (!ismaster && grp->sg_threadcount > grp->sg_minthreads)) error = cv_timedwait_sig(&st->st_cond, &grp->sg_lock, 5 * hz); else error = cv_wait_sig(&st->st_cond, &grp->sg_lock); if (st->st_xprt == NULL) LIST_REMOVE(st, st_ilink); /* * Reduce worker thread count when idle. */ if (error == EWOULDBLOCK) { if (!ismaster && (grp->sg_threadcount > grp->sg_minthreads) && !st->st_xprt) break; - } else if (error) { + } else if (error != 0) { + KASSERT(error == EINTR || error == ERESTART, + ("non-signal error %d", error)); mtx_unlock(&grp->sg_lock); - svc_exit(pool); - mtx_lock(&grp->sg_lock); - break; + p = curproc; + PROC_LOCK(p); + if (P_SHOULDSTOP(p)) { + thread_suspend_check(0); + PROC_UNLOCK(p); + mtx_lock(&grp->sg_lock); + } else { + PROC_UNLOCK(p); + svc_exit(pool); + mtx_lock(&grp->sg_lock); + break; + } } continue; } mtx_unlock(&grp->sg_lock); /* * Drain the transport socket and queue up any RPCs. */ xprt->xp_lastactive = time_uptime; do { if (!svc_request_space_available(pool)) break; rqstp = NULL; stat = svc_getreq(xprt, &rqstp); if (rqstp) { svc_change_space_used(pool, rqstp->rq_size); /* * See if the application has a preference * for some other thread. */ if (pool->sp_assign) { stpref = pool->sp_assign(st, rqstp); rqstp->rq_thread = stpref; STAILQ_INSERT_TAIL(&stpref->st_reqs, rqstp, rq_link); mtx_unlock(&stpref->st_lock); if (stpref != st) rqstp = NULL; } else { rqstp->rq_thread = st; STAILQ_INSERT_TAIL(&st->st_reqs, rqstp, rq_link); } } } while (rqstp == NULL && stat == XPRT_MOREREQS && grp->sg_state != SVCPOOL_CLOSING); /* * Move this transport to the end of the active list to * ensure fairness when multiple transports are active. * If this was the last queued request, svc_getreq will end * up calling xprt_inactive to remove from the active list. */ mtx_lock(&grp->sg_lock); xprt->xp_thread = NULL; st->st_xprt = NULL; if (xprt->xp_active) { if (!svc_request_space_available(pool) || !xprt_assignthread(xprt)) TAILQ_INSERT_TAIL(&grp->sg_active, xprt, xp_alink); } mtx_unlock(&grp->sg_lock); SVC_RELEASE(xprt); /* * Execute what we have queued. */ sz = 0; mtx_lock(&st->st_lock); while ((rqstp = STAILQ_FIRST(&st->st_reqs)) != NULL) { STAILQ_REMOVE_HEAD(&st->st_reqs, rq_link); mtx_unlock(&st->st_lock); sz += rqstp->rq_size; svc_executereq(rqstp); mtx_lock(&st->st_lock); } mtx_unlock(&st->st_lock); svc_change_space_used(pool, -sz); mtx_lock(&grp->sg_lock); } if (st->st_xprt) { xprt = st->st_xprt; st->st_xprt = NULL; SVC_RELEASE(xprt); } KASSERT(STAILQ_EMPTY(&st->st_reqs), ("stray reqs on exit")); mtx_destroy(&st->st_lock); cv_destroy(&st->st_cond); mem_free(st, sizeof(*st)); grp->sg_threadcount--; if (!ismaster) wakeup(grp); mtx_unlock(&grp->sg_lock); } static void svc_thread_start(void *arg) { svc_run_internal((SVCGROUP *) arg, FALSE); kthread_exit(); } static void svc_new_thread(SVCGROUP *grp) { SVCPOOL *pool = grp->sg_pool; struct thread *td; grp->sg_threadcount++; kthread_add(svc_thread_start, grp, pool->sp_proc, &td, 0, 0, "%s: service", pool->sp_name); } void svc_run(SVCPOOL *pool) { int g, i; struct proc *p; struct thread *td; SVCGROUP *grp; p = curproc; td = curthread; snprintf(td->td_name, sizeof(td->td_name), "%s: master", pool->sp_name); pool->sp_state = SVCPOOL_ACTIVE; pool->sp_proc = p; /* Choose group count based on number of threads and CPUs. */ pool->sp_groupcount = max(1, min(SVC_MAXGROUPS, min(pool->sp_maxthreads / 2, mp_ncpus) / 6)); for (g = 0; g < pool->sp_groupcount; g++) { grp = &pool->sp_groups[g]; grp->sg_minthreads = max(1, pool->sp_minthreads / pool->sp_groupcount); grp->sg_maxthreads = max(1, pool->sp_maxthreads / pool->sp_groupcount); grp->sg_lastcreatetime = time_uptime; } /* Starting threads */ for (g = 0; g < pool->sp_groupcount; g++) { grp = &pool->sp_groups[g]; for (i = ((g == 0) ? 1 : 0); i < grp->sg_minthreads; i++) svc_new_thread(grp); } pool->sp_groups[0].sg_threadcount++; svc_run_internal(&pool->sp_groups[0], TRUE); /* Waiting for threads to stop. */ for (g = 0; g < pool->sp_groupcount; g++) { grp = &pool->sp_groups[g]; mtx_lock(&grp->sg_lock); while (grp->sg_threadcount > 0) msleep(grp, &grp->sg_lock, 0, "svcexit", 0); mtx_unlock(&grp->sg_lock); } } void svc_exit(SVCPOOL *pool) { SVCGROUP *grp; SVCTHREAD *st; int g; pool->sp_state = SVCPOOL_CLOSING; for (g = 0; g < pool->sp_groupcount; g++) { grp = &pool->sp_groups[g]; mtx_lock(&grp->sg_lock); if (grp->sg_state != SVCPOOL_CLOSING) { grp->sg_state = SVCPOOL_CLOSING; LIST_FOREACH(st, &grp->sg_idlethreads, st_ilink) cv_signal(&st->st_cond); } mtx_unlock(&grp->sg_lock); } } bool_t svc_getargs(struct svc_req *rqstp, xdrproc_t xargs, void *args) { struct mbuf *m; XDR xdrs; bool_t stat; m = rqstp->rq_args; rqstp->rq_args = NULL; xdrmbuf_create(&xdrs, m, XDR_DECODE); stat = xargs(&xdrs, args); XDR_DESTROY(&xdrs); return (stat); } bool_t svc_freeargs(struct svc_req *rqstp, xdrproc_t xargs, void *args) { XDR xdrs; if (rqstp->rq_addr) { free(rqstp->rq_addr, M_SONAME); rqstp->rq_addr = NULL; } xdrs.x_op = XDR_FREE; return (xargs(&xdrs, args)); } void svc_freereq(struct svc_req *rqstp) { SVCTHREAD *st; SVCPOOL *pool; st = rqstp->rq_thread; if (st) { pool = st->st_pool; if (pool->sp_done) pool->sp_done(st, rqstp); } if (rqstp->rq_auth.svc_ah_ops) SVCAUTH_RELEASE(&rqstp->rq_auth); if (rqstp->rq_xprt) { SVC_RELEASE(rqstp->rq_xprt); } if (rqstp->rq_addr) free(rqstp->rq_addr, M_SONAME); if (rqstp->rq_args) m_freem(rqstp->rq_args); free(rqstp, M_RPC); } Index: projects/building-blocks/sys/sys/mount.h =================================================================== --- projects/building-blocks/sys/sys/mount.h (revision 275663) +++ projects/building-blocks/sys/sys/mount.h (revision 275664) @@ -1,946 +1,949 @@ /*- * Copyright (c) 1989, 1991, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)mount.h 8.21 (Berkeley) 5/20/95 * $FreeBSD$ */ #ifndef _SYS_MOUNT_H_ #define _SYS_MOUNT_H_ #include #include #ifdef _KERNEL #include #include #include #include #endif /* * NOTE: When changing statfs structure, mount structure, MNT_* flags or * MNTK_* flags also update DDB show mount command in vfs_subr.c. */ typedef struct fsid { int32_t val[2]; } fsid_t; /* filesystem id type */ /* * File identifier. * These are unique per filesystem on a single machine. */ #define MAXFIDSZ 16 struct fid { u_short fid_len; /* length of data in bytes */ u_short fid_data0; /* force longword alignment */ char fid_data[MAXFIDSZ]; /* data (variable length) */ }; /* * filesystem statistics */ #define MFSNAMELEN 16 /* length of type name including null */ #define MNAMELEN 88 /* size of on/from name bufs */ #define STATFS_VERSION 0x20030518 /* current version number */ struct statfs { uint32_t f_version; /* structure version number */ uint32_t f_type; /* type of filesystem */ uint64_t f_flags; /* copy of mount exported flags */ uint64_t f_bsize; /* filesystem fragment size */ uint64_t f_iosize; /* optimal transfer block size */ uint64_t f_blocks; /* total data blocks in filesystem */ uint64_t f_bfree; /* free blocks in filesystem */ int64_t f_bavail; /* free blocks avail to non-superuser */ uint64_t f_files; /* total file nodes in filesystem */ int64_t f_ffree; /* free nodes avail to non-superuser */ uint64_t f_syncwrites; /* count of sync writes since mount */ uint64_t f_asyncwrites; /* count of async writes since mount */ uint64_t f_syncreads; /* count of sync reads since mount */ uint64_t f_asyncreads; /* count of async reads since mount */ uint64_t f_spare[10]; /* unused spare */ uint32_t f_namemax; /* maximum filename length */ uid_t f_owner; /* user that mounted the filesystem */ fsid_t f_fsid; /* filesystem id */ char f_charspare[80]; /* spare string space */ char f_fstypename[MFSNAMELEN]; /* filesystem type name */ char f_mntfromname[MNAMELEN]; /* mounted filesystem */ char f_mntonname[MNAMELEN]; /* directory on which mounted */ }; #ifdef _KERNEL #define OMFSNAMELEN 16 /* length of fs type name, including null */ #define OMNAMELEN (88 - 2 * sizeof(long)) /* size of on/from name bufs */ /* XXX getfsstat.2 is out of date with write and read counter changes here. */ /* XXX statfs.2 is out of date with read counter changes here. */ struct ostatfs { long f_spare2; /* placeholder */ long f_bsize; /* fundamental filesystem block size */ long f_iosize; /* optimal transfer block size */ long f_blocks; /* total data blocks in filesystem */ long f_bfree; /* free blocks in fs */ long f_bavail; /* free blocks avail to non-superuser */ long f_files; /* total file nodes in filesystem */ long f_ffree; /* free file nodes in fs */ fsid_t f_fsid; /* filesystem id */ uid_t f_owner; /* user that mounted the filesystem */ int f_type; /* type of filesystem */ int f_flags; /* copy of mount exported flags */ long f_syncwrites; /* count of sync writes since mount */ long f_asyncwrites; /* count of async writes since mount */ char f_fstypename[OMFSNAMELEN]; /* fs type name */ char f_mntonname[OMNAMELEN]; /* directory on which mounted */ long f_syncreads; /* count of sync reads since mount */ long f_asyncreads; /* count of async reads since mount */ short f_spares1; /* unused spare */ char f_mntfromname[OMNAMELEN];/* mounted filesystem */ short f_spares2; /* unused spare */ /* * XXX on machines where longs are aligned to 8-byte boundaries, there * is an unnamed int32_t here. This spare was after the apparent end * of the struct until we bit off the read counters from f_mntonname. */ long f_spare[2]; /* unused spare */ }; TAILQ_HEAD(vnodelst, vnode); /* Mount options list */ TAILQ_HEAD(vfsoptlist, vfsopt); struct vfsopt { TAILQ_ENTRY(vfsopt) link; char *name; void *value; int len; int pos; int seen; }; /* * Structure per mounted filesystem. Each mounted filesystem has an * array of operations and an instance record. The filesystems are * put on a doubly linked list. * * Lock reference: * m - mountlist_mtx * i - interlock * v - vnode freelist mutex * * Unmarked fields are considered stable as long as a ref is held. * */ struct mount { struct mtx mnt_mtx; /* mount structure interlock */ int mnt_gen; /* struct mount generation */ #define mnt_startzero mnt_list TAILQ_ENTRY(mount) mnt_list; /* (m) mount list */ struct vfsops *mnt_op; /* operations on fs */ struct vfsconf *mnt_vfc; /* configuration info */ struct vnode *mnt_vnodecovered; /* vnode we mounted on */ struct vnode *mnt_syncer; /* syncer vnode */ int mnt_ref; /* (i) Reference count */ struct vnodelst mnt_nvnodelist; /* (i) list of vnodes */ int mnt_nvnodelistsize; /* (i) # of vnodes */ struct vnodelst mnt_activevnodelist; /* (v) list of active vnodes */ int mnt_activevnodelistsize;/* (v) # of active vnodes */ int mnt_writeopcount; /* (i) write syscalls pending */ int mnt_kern_flag; /* (i) kernel only flags */ uint64_t mnt_flag; /* (i) flags shared with user */ struct vfsoptlist *mnt_opt; /* current mount options */ struct vfsoptlist *mnt_optnew; /* new options passed to fs */ int mnt_maxsymlinklen; /* max size of short symlink */ struct statfs mnt_stat; /* cache of filesystem stats */ struct ucred *mnt_cred; /* credentials of mounter */ void * mnt_data; /* private data */ time_t mnt_time; /* last time written*/ int mnt_iosize_max; /* max size for clusters, etc */ struct netexport *mnt_export; /* export list */ struct label *mnt_label; /* MAC label for the fs */ u_int mnt_hashseed; /* Random seed for vfs_hash */ int mnt_lockref; /* (i) Lock reference count */ int mnt_secondary_writes; /* (i) # of secondary writes */ int mnt_secondary_accwrites;/* (i) secondary wr. starts */ struct thread *mnt_susp_owner; /* (i) thread owning suspension */ #define mnt_endzero mnt_gjprovider char *mnt_gjprovider; /* gjournal provider name */ struct lock mnt_explock; /* vfs_export walkers lock */ TAILQ_ENTRY(mount) mnt_upper_link; /* (m) we in the all uppers */ TAILQ_HEAD(, mount) mnt_uppers; /* (m) upper mounts over us*/ }; /* * Definitions for MNT_VNODE_FOREACH_ALL. */ struct vnode *__mnt_vnode_next_all(struct vnode **mvp, struct mount *mp); struct vnode *__mnt_vnode_first_all(struct vnode **mvp, struct mount *mp); void __mnt_vnode_markerfree_all(struct vnode **mvp, struct mount *mp); #define MNT_VNODE_FOREACH_ALL(vp, mp, mvp) \ for (vp = __mnt_vnode_first_all(&(mvp), (mp)); \ (vp) != NULL; vp = __mnt_vnode_next_all(&(mvp), (mp))) #define MNT_VNODE_FOREACH_ALL_ABORT(mp, mvp) \ do { \ MNT_ILOCK(mp); \ __mnt_vnode_markerfree_all(&(mvp), (mp)); \ /* MNT_IUNLOCK(mp); -- done in above function */ \ mtx_assert(MNT_MTX(mp), MA_NOTOWNED); \ } while (0) /* * Definitions for MNT_VNODE_FOREACH_ACTIVE. */ struct vnode *__mnt_vnode_next_active(struct vnode **mvp, struct mount *mp); struct vnode *__mnt_vnode_first_active(struct vnode **mvp, struct mount *mp); void __mnt_vnode_markerfree_active(struct vnode **mvp, struct mount *); #define MNT_VNODE_FOREACH_ACTIVE(vp, mp, mvp) \ for (vp = __mnt_vnode_first_active(&(mvp), (mp)); \ (vp) != NULL; vp = __mnt_vnode_next_active(&(mvp), (mp))) #define MNT_VNODE_FOREACH_ACTIVE_ABORT(mp, mvp) \ __mnt_vnode_markerfree_active(&(mvp), (mp)) #define MNT_ILOCK(mp) mtx_lock(&(mp)->mnt_mtx) #define MNT_ITRYLOCK(mp) mtx_trylock(&(mp)->mnt_mtx) #define MNT_IUNLOCK(mp) mtx_unlock(&(mp)->mnt_mtx) #define MNT_MTX(mp) (&(mp)->mnt_mtx) #define MNT_REF(mp) (mp)->mnt_ref++ #define MNT_REL(mp) do { \ KASSERT((mp)->mnt_ref > 0, ("negative mnt_ref")); \ (mp)->mnt_ref--; \ if ((mp)->mnt_ref == 0) \ wakeup((mp)); \ } while (0) #endif /* _KERNEL */ /* * User specifiable flags, stored in mnt_flag. */ #define MNT_RDONLY 0x0000000000000001ULL /* read only filesystem */ #define MNT_SYNCHRONOUS 0x0000000000000002ULL /* fs written synchronously */ #define MNT_NOEXEC 0x0000000000000004ULL /* can't exec from filesystem */ #define MNT_NOSUID 0x0000000000000008ULL /* don't honor setuid fs bits */ #define MNT_NFS4ACLS 0x0000000000000010ULL /* enable NFS version 4 ACLs */ #define MNT_UNION 0x0000000000000020ULL /* union with underlying fs */ #define MNT_ASYNC 0x0000000000000040ULL /* fs written asynchronously */ #define MNT_SUIDDIR 0x0000000000100000ULL /* special SUID dir handling */ #define MNT_SOFTDEP 0x0000000000200000ULL /* using soft updates */ #define MNT_NOSYMFOLLOW 0x0000000000400000ULL /* do not follow symlinks */ #define MNT_GJOURNAL 0x0000000002000000ULL /* GEOM journal support enabled */ #define MNT_MULTILABEL 0x0000000004000000ULL /* MAC support for objects */ #define MNT_ACLS 0x0000000008000000ULL /* ACL support enabled */ #define MNT_NOATIME 0x0000000010000000ULL /* dont update file access time */ #define MNT_NOCLUSTERR 0x0000000040000000ULL /* disable cluster read */ #define MNT_NOCLUSTERW 0x0000000080000000ULL /* disable cluster write */ #define MNT_SUJ 0x0000000100000000ULL /* using journaled soft updates */ #define MNT_AUTOMOUNTED 0x0000000200000000ULL /* mounted by automountd(8) */ /* * NFS export related mount flags. */ #define MNT_EXRDONLY 0x0000000000000080ULL /* exported read only */ #define MNT_EXPORTED 0x0000000000000100ULL /* filesystem is exported */ #define MNT_DEFEXPORTED 0x0000000000000200ULL /* exported to the world */ #define MNT_EXPORTANON 0x0000000000000400ULL /* anon uid mapping for all */ #define MNT_EXKERB 0x0000000000000800ULL /* exported with Kerberos */ #define MNT_EXPUBLIC 0x0000000020000000ULL /* public export (WebNFS) */ /* * Flags set by internal operations, * but visible to the user. * XXX some of these are not quite right.. (I've never seen the root flag set) */ #define MNT_LOCAL 0x0000000000001000ULL /* filesystem is stored locally */ #define MNT_QUOTA 0x0000000000002000ULL /* quotas are enabled on fs */ #define MNT_ROOTFS 0x0000000000004000ULL /* identifies the root fs */ #define MNT_USER 0x0000000000008000ULL /* mounted by a user */ #define MNT_IGNORE 0x0000000000800000ULL /* do not show entry in df */ /* * Mask of flags that are visible to statfs(). * XXX I think that this could now become (~(MNT_CMDFLAGS)) * but the 'mount' program may need changing to handle this. */ #define MNT_VISFLAGMASK (MNT_RDONLY | MNT_SYNCHRONOUS | MNT_NOEXEC | \ MNT_NOSUID | MNT_UNION | MNT_SUJ | \ MNT_ASYNC | MNT_EXRDONLY | MNT_EXPORTED | \ MNT_DEFEXPORTED | MNT_EXPORTANON| MNT_EXKERB | \ MNT_LOCAL | MNT_USER | MNT_QUOTA | \ MNT_ROOTFS | MNT_NOATIME | MNT_NOCLUSTERR| \ MNT_NOCLUSTERW | MNT_SUIDDIR | MNT_SOFTDEP | \ MNT_IGNORE | MNT_EXPUBLIC | MNT_NOSYMFOLLOW | \ MNT_GJOURNAL | MNT_MULTILABEL | MNT_ACLS | \ MNT_NFS4ACLS | MNT_AUTOMOUNTED) /* Mask of flags that can be updated. */ #define MNT_UPDATEMASK (MNT_NOSUID | MNT_NOEXEC | \ MNT_SYNCHRONOUS | MNT_UNION | MNT_ASYNC | \ MNT_NOATIME | \ MNT_NOSYMFOLLOW | MNT_IGNORE | \ MNT_NOCLUSTERR | MNT_NOCLUSTERW | MNT_SUIDDIR | \ MNT_ACLS | MNT_USER | MNT_NFS4ACLS | \ MNT_AUTOMOUNTED) /* * External filesystem command modifier flags. * Unmount can use the MNT_FORCE flag. * XXX: These are not STATES and really should be somewhere else. * XXX: MNT_BYFSID collides with MNT_ACLS, but because MNT_ACLS is only used for * mount(2) and MNT_BYFSID is only used for unmount(2) it's harmless. */ #define MNT_UPDATE 0x0000000000010000ULL /* not real mount, just update */ #define MNT_DELEXPORT 0x0000000000020000ULL /* delete export host lists */ #define MNT_RELOAD 0x0000000000040000ULL /* reload filesystem data */ #define MNT_FORCE 0x0000000000080000ULL /* force unmount or readonly */ #define MNT_SNAPSHOT 0x0000000001000000ULL /* snapshot the filesystem */ #define MNT_BYFSID 0x0000000008000000ULL /* specify filesystem by ID. */ #define MNT_CMDFLAGS (MNT_UPDATE | MNT_DELEXPORT | MNT_RELOAD | \ MNT_FORCE | MNT_SNAPSHOT | MNT_BYFSID) /* * Internal filesystem control flags stored in mnt_kern_flag. * * MNTK_UNMOUNT locks the mount entry so that name lookup cannot proceed * past the mount point. This keeps the subtree stable during mounts * and unmounts. * * MNTK_UNMOUNTF permits filesystems to detect a forced unmount while * dounmount() is still waiting to lock the mountpoint. This allows * the filesystem to cancel operations that might otherwise deadlock * with the unmount attempt (used by NFS). * * MNTK_NOINSMNTQ is strict subset of MNTK_UNMOUNT. They are separated * to allow for failed unmount attempt to restore the syncer vnode for * the mount. */ #define MNTK_UNMOUNTF 0x00000001 /* forced unmount in progress */ #define MNTK_ASYNC 0x00000002 /* filtered async flag */ #define MNTK_SOFTDEP 0x00000004 /* async disabled by softdep */ #define MNTK_NOINSMNTQ 0x00000008 /* insmntque is not allowed */ #define MNTK_DRAINING 0x00000010 /* lock draining is happening */ #define MNTK_REFEXPIRE 0x00000020 /* refcount expiring is happening */ #define MNTK_EXTENDED_SHARED 0x00000040 /* Allow shared locking for more ops */ #define MNTK_SHARED_WRITES 0x00000080 /* Allow shared locking for writes */ #define MNTK_NO_IOPF 0x00000100 /* Disallow page faults during reads and writes. Filesystem shall properly handle i/o state on EFAULT. */ #define MNTK_VGONE_UPPER 0x00000200 #define MNTK_VGONE_WAITER 0x00000400 #define MNTK_LOOKUP_EXCL_DOTDOT 0x00000800 #define MNTK_MARKER 0x00001000 #define MNTK_UNMAPPED_BUFS 0x00002000 #define MNTK_NOASYNC 0x00800000 /* disable async */ #define MNTK_UNMOUNT 0x01000000 /* unmount in progress */ #define MNTK_MWAIT 0x02000000 /* waiting for unmount to finish */ #define MNTK_SUSPEND 0x08000000 /* request write suspension */ #define MNTK_SUSPEND2 0x04000000 /* block secondary writes */ #define MNTK_SUSPENDED 0x10000000 /* write operations are suspended */ #define MNTK_SUSPENDABLE 0x20000000 /* writes can be suspended */ #define MNTK_LOOKUP_SHARED 0x40000000 /* FS supports shared lock lookups */ #define MNTK_NOKNOTE 0x80000000 /* Don't send KNOTEs from VOP hooks */ #ifdef _KERNEL static inline int MNT_SHARED_WRITES(struct mount *mp) { return (mp != NULL && (mp->mnt_kern_flag & MNTK_SHARED_WRITES) != 0); } static inline int MNT_EXTENDED_SHARED(struct mount *mp) { return (mp != NULL && (mp->mnt_kern_flag & MNTK_EXTENDED_SHARED) != 0); } #endif /* * Sysctl CTL_VFS definitions. * * Second level identifier specifies which filesystem. Second level * identifier VFS_VFSCONF returns information about all filesystems. * Second level identifier VFS_GENERIC is non-terminal. */ #define VFS_VFSCONF 0 /* get configured filesystems */ #define VFS_GENERIC 0 /* generic filesystem information */ /* * Third level identifiers for VFS_GENERIC are given below; third * level identifiers for specific filesystems are given in their * mount specific header files. */ #define VFS_MAXTYPENUM 1 /* int: highest defined filesystem type */ #define VFS_CONF 2 /* struct: vfsconf for filesystem given as next argument */ /* * Flags for various system call interfaces. * * waitfor flags to vfs_sync() and getfsstat() */ #define MNT_WAIT 1 /* synchronously wait for I/O to complete */ #define MNT_NOWAIT 2 /* start all I/O, but do not wait for it */ #define MNT_LAZY 3 /* push data not written by filesystem syncer */ #define MNT_SUSPEND 4 /* Suspend file system after sync */ /* * Generic file handle */ struct fhandle { fsid_t fh_fsid; /* Filesystem id of mount point */ struct fid fh_fid; /* Filesys specific id */ }; typedef struct fhandle fhandle_t; /* * Old export arguments without security flavor list */ struct oexport_args { int ex_flags; /* export related flags */ uid_t ex_root; /* mapping for root uid */ struct xucred ex_anon; /* mapping for anonymous user */ struct sockaddr *ex_addr; /* net address to which exported */ u_char ex_addrlen; /* and the net address length */ struct sockaddr *ex_mask; /* mask of valid bits in saddr */ u_char ex_masklen; /* and the smask length */ char *ex_indexfile; /* index file for WebNFS URLs */ }; /* * Export arguments for local filesystem mount calls. */ #define MAXSECFLAVORS 5 struct export_args { int ex_flags; /* export related flags */ uid_t ex_root; /* mapping for root uid */ struct xucred ex_anon; /* mapping for anonymous user */ struct sockaddr *ex_addr; /* net address to which exported */ u_char ex_addrlen; /* and the net address length */ struct sockaddr *ex_mask; /* mask of valid bits in saddr */ u_char ex_masklen; /* and the smask length */ char *ex_indexfile; /* index file for WebNFS URLs */ int ex_numsecflavors; /* security flavor count */ int ex_secflavors[MAXSECFLAVORS]; /* list of security flavors */ }; /* * Structure holding information for a publicly exported filesystem * (WebNFS). Currently the specs allow just for one such filesystem. */ struct nfs_public { int np_valid; /* Do we hold valid information */ fhandle_t np_handle; /* Filehandle for pub fs (internal) */ struct mount *np_mount; /* Mountpoint of exported fs */ char *np_index; /* Index file */ }; /* * Filesystem configuration information. One of these exists for each * type of filesystem supported by the kernel. These are searched at * mount time to identify the requested filesystem. * * XXX: Never change the first two arguments! */ struct vfsconf { u_int vfc_version; /* ABI version number */ char vfc_name[MFSNAMELEN]; /* filesystem type name */ struct vfsops *vfc_vfsops; /* filesystem operations vector */ int vfc_typenum; /* historic filesystem type number */ int vfc_refcount; /* number mounted of this type */ int vfc_flags; /* permanent flags */ struct vfsoptdecl *vfc_opts; /* mount options */ TAILQ_ENTRY(vfsconf) vfc_list; /* list of vfscons */ }; /* Userland version of the struct vfsconf. */ struct xvfsconf { struct vfsops *vfc_vfsops; /* filesystem operations vector */ char vfc_name[MFSNAMELEN]; /* filesystem type name */ int vfc_typenum; /* historic filesystem type number */ int vfc_refcount; /* number mounted of this type */ int vfc_flags; /* permanent flags */ struct vfsconf *vfc_next; /* next in list */ }; #ifndef BURN_BRIDGES struct ovfsconf { void *vfc_vfsops; char vfc_name[32]; int vfc_index; int vfc_refcount; int vfc_flags; }; #endif /* * NB: these flags refer to IMPLEMENTATION properties, not properties of * any actual mounts; i.e., it does not make sense to change the flags. */ #define VFCF_STATIC 0x00010000 /* statically compiled into kernel */ #define VFCF_NETWORK 0x00020000 /* may get data over the network */ #define VFCF_READONLY 0x00040000 /* writes are not implemented */ #define VFCF_SYNTHETIC 0x00080000 /* data does not represent real files */ #define VFCF_LOOPBACK 0x00100000 /* aliases some other mounted FS */ #define VFCF_UNICODE 0x00200000 /* stores file names as Unicode */ #define VFCF_JAIL 0x00400000 /* can be mounted from within a jail */ #define VFCF_DELEGADMIN 0x00800000 /* supports delegated administration */ #define VFCF_SBDRY 0x01000000 /* defer stop requests */ typedef uint32_t fsctlop_t; struct vfsidctl { int vc_vers; /* should be VFSIDCTL_VERS1 (below) */ fsid_t vc_fsid; /* fsid to operate on */ char vc_fstypename[MFSNAMELEN]; /* type of fs 'nfs' or '*' */ fsctlop_t vc_op; /* operation VFS_CTL_* (below) */ void *vc_ptr; /* pointer to data structure */ size_t vc_len; /* sizeof said structure */ u_int32_t vc_spare[12]; /* spare (must be zero) */ }; /* vfsidctl API version. */ #define VFS_CTL_VERS1 0x01 /* * New style VFS sysctls, do not reuse/conflict with the namespace for * private sysctls. * All "global" sysctl ops have the 33rd bit set: * 0x...1.... * Private sysctl ops should have the 33rd bit unset. */ #define VFS_CTL_QUERY 0x00010001 /* anything wrong? (vfsquery) */ #define VFS_CTL_TIMEO 0x00010002 /* set timeout for vfs notification */ #define VFS_CTL_NOLOCKS 0x00010003 /* disable file locking */ struct vfsquery { u_int32_t vq_flags; u_int32_t vq_spare[31]; }; /* vfsquery flags */ #define VQ_NOTRESP 0x0001 /* server down */ #define VQ_NEEDAUTH 0x0002 /* server bad auth */ #define VQ_LOWDISK 0x0004 /* we're low on space */ #define VQ_MOUNT 0x0008 /* new filesystem arrived */ #define VQ_UNMOUNT 0x0010 /* filesystem has left */ #define VQ_DEAD 0x0020 /* filesystem is dead, needs force unmount */ #define VQ_ASSIST 0x0040 /* filesystem needs assistance from external program */ #define VQ_NOTRESPLOCK 0x0080 /* server lockd down */ #define VQ_FLAG0100 0x0100 /* placeholder */ #define VQ_FLAG0200 0x0200 /* placeholder */ #define VQ_FLAG0400 0x0400 /* placeholder */ #define VQ_FLAG0800 0x0800 /* placeholder */ #define VQ_FLAG1000 0x1000 /* placeholder */ #define VQ_FLAG2000 0x2000 /* placeholder */ #define VQ_FLAG4000 0x4000 /* placeholder */ #define VQ_FLAG8000 0x8000 /* placeholder */ #ifdef _KERNEL /* Point a sysctl request at a vfsidctl's data. */ #define VCTLTOREQ(vc, req) \ do { \ (req)->newptr = (vc)->vc_ptr; \ (req)->newlen = (vc)->vc_len; \ (req)->newidx = 0; \ } while (0) #endif struct iovec; struct uio; #ifdef _KERNEL /* * vfs_busy specific flags and mask. */ #define MBF_NOWAIT 0x01 #define MBF_MNTLSTLOCK 0x02 #define MBF_MASK (MBF_NOWAIT | MBF_MNTLSTLOCK) #ifdef MALLOC_DECLARE MALLOC_DECLARE(M_MOUNT); #endif extern int maxvfsconf; /* highest defined filesystem type */ extern int nfs_mount_type; /* vfc_typenum for nfs, or -1 */ TAILQ_HEAD(vfsconfhead, vfsconf); extern struct vfsconfhead vfsconf; /* * Operations supported on mounted filesystem. */ struct mount_args; struct nameidata; struct sysctl_req; struct mntarg; typedef int vfs_cmount_t(struct mntarg *ma, void *data, uint64_t flags); typedef int vfs_unmount_t(struct mount *mp, int mntflags); typedef int vfs_root_t(struct mount *mp, int flags, struct vnode **vpp); typedef int vfs_quotactl_t(struct mount *mp, int cmds, uid_t uid, void *arg); typedef int vfs_statfs_t(struct mount *mp, struct statfs *sbp); typedef int vfs_sync_t(struct mount *mp, int waitfor); typedef int vfs_vget_t(struct mount *mp, ino_t ino, int flags, struct vnode **vpp); typedef int vfs_fhtovp_t(struct mount *mp, struct fid *fhp, int flags, struct vnode **vpp); typedef int vfs_checkexp_t(struct mount *mp, struct sockaddr *nam, int *extflagsp, struct ucred **credanonp, int *numsecflavors, int **secflavors); typedef int vfs_init_t(struct vfsconf *); typedef int vfs_uninit_t(struct vfsconf *); typedef int vfs_extattrctl_t(struct mount *mp, int cmd, struct vnode *filename_vp, int attrnamespace, const char *attrname); typedef int vfs_mount_t(struct mount *mp); typedef int vfs_sysctl_t(struct mount *mp, fsctlop_t op, struct sysctl_req *req); typedef void vfs_susp_clean_t(struct mount *mp); typedef void vfs_notify_lowervp_t(struct mount *mp, struct vnode *lowervp); typedef void vfs_purge_t(struct mount *mp); struct vfsops { vfs_mount_t *vfs_mount; vfs_cmount_t *vfs_cmount; vfs_unmount_t *vfs_unmount; vfs_root_t *vfs_root; vfs_quotactl_t *vfs_quotactl; vfs_statfs_t *vfs_statfs; vfs_sync_t *vfs_sync; vfs_vget_t *vfs_vget; vfs_fhtovp_t *vfs_fhtovp; vfs_checkexp_t *vfs_checkexp; vfs_init_t *vfs_init; vfs_uninit_t *vfs_uninit; vfs_extattrctl_t *vfs_extattrctl; vfs_sysctl_t *vfs_sysctl; vfs_susp_clean_t *vfs_susp_clean; vfs_notify_lowervp_t *vfs_reclaim_lowervp; vfs_notify_lowervp_t *vfs_unlink_lowervp; vfs_purge_t *vfs_purge; vfs_mount_t *vfs_spare[6]; /* spares for ABI compat */ }; vfs_statfs_t __vfs_statfs; #define VFS_PROLOGUE(MP) do { \ struct mount *mp__; \ int _enable_stops; \ \ mp__ = (MP); \ _enable_stops = (mp__ != NULL && \ (mp__->mnt_vfc->vfc_flags & VFCF_SBDRY) && sigdeferstop()) #define VFS_EPILOGUE(MP) \ if (_enable_stops) \ sigallowstop(); \ } while (0) #define VFS_MOUNT(MP) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_mount)(MP); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_UNMOUNT(MP, FORCE) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_unmount)(MP, FORCE); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_ROOT(MP, FLAGS, VPP) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_root)(MP, FLAGS, VPP); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_QUOTACTL(MP, C, U, A) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_quotactl)(MP, C, U, A); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_STATFS(MP, SBP) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = __vfs_statfs((MP), (SBP)); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_SYNC(MP, WAIT) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_sync)(MP, WAIT); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_VGET(MP, INO, FLAGS, VPP) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_vget)(MP, INO, FLAGS, VPP); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_FHTOVP(MP, FIDP, FLAGS, VPP) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_fhtovp)(MP, FIDP, FLAGS, VPP); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_CHECKEXP(MP, NAM, EXFLG, CRED, NUMSEC, SEC) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_checkexp)(MP, NAM, EXFLG, CRED, NUMSEC,\ SEC); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_EXTATTRCTL(MP, C, FN, NS, N) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_extattrctl)(MP, C, FN, NS, N); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_SYSCTL(MP, OP, REQ) ({ \ int _rc; \ \ VFS_PROLOGUE(MP); \ _rc = (*(MP)->mnt_op->vfs_sysctl)(MP, OP, REQ); \ VFS_EPILOGUE(MP); \ _rc; }) #define VFS_SUSP_CLEAN(MP) do { \ if (*(MP)->mnt_op->vfs_susp_clean != NULL) { \ VFS_PROLOGUE(MP); \ (*(MP)->mnt_op->vfs_susp_clean)(MP); \ VFS_EPILOGUE(MP); \ } \ } while (0) #define VFS_RECLAIM_LOWERVP(MP, VP) do { \ if (*(MP)->mnt_op->vfs_reclaim_lowervp != NULL) { \ VFS_PROLOGUE(MP); \ (*(MP)->mnt_op->vfs_reclaim_lowervp)((MP), (VP)); \ VFS_EPILOGUE(MP); \ } \ } while (0) #define VFS_UNLINK_LOWERVP(MP, VP) do { \ if (*(MP)->mnt_op->vfs_unlink_lowervp != NULL) { \ VFS_PROLOGUE(MP); \ (*(MP)->mnt_op->vfs_unlink_lowervp)((MP), (VP)); \ VFS_EPILOGUE(MP); \ } \ } while (0) #define VFS_PURGE(MP) do { \ if (*(MP)->mnt_op->vfs_purge != NULL) { \ VFS_PROLOGUE(MP); \ (*(MP)->mnt_op->vfs_purge)(MP); \ VFS_EPILOGUE(MP); \ } \ } while (0) #define VFS_KNOTE_LOCKED(vp, hint) do \ { \ if (((vp)->v_vflag & VV_NOKNOTE) == 0) \ VN_KNOTE((vp), (hint), KNF_LISTLOCKED); \ } while (0) #define VFS_KNOTE_UNLOCKED(vp, hint) do \ { \ if (((vp)->v_vflag & VV_NOKNOTE) == 0) \ VN_KNOTE((vp), (hint), 0); \ } while (0) #define VFS_NOTIFY_UPPER_RECLAIM 1 #define VFS_NOTIFY_UPPER_UNLINK 2 #include /* * Version numbers. */ #define VFS_VERSION_00 0x19660120 #define VFS_VERSION_01 0x20121030 #define VFS_VERSION VFS_VERSION_01 #define VFS_SET(vfsops, fsname, flags) \ static struct vfsconf fsname ## _vfsconf = { \ .vfc_version = VFS_VERSION, \ .vfc_name = #fsname, \ .vfc_vfsops = &vfsops, \ .vfc_typenum = -1, \ .vfc_flags = flags, \ }; \ static moduledata_t fsname ## _mod = { \ #fsname, \ vfs_modevent, \ & fsname ## _vfsconf \ }; \ DECLARE_MODULE(fsname, fsname ## _mod, SI_SUB_VFS, SI_ORDER_MIDDLE) /* * exported vnode operations */ int dounmount(struct mount *, int, struct thread *); int kernel_mount(struct mntarg *ma, uint64_t flags); int kernel_vmount(int flags, ...); struct mntarg *mount_arg(struct mntarg *ma, const char *name, const void *val, int len); struct mntarg *mount_argb(struct mntarg *ma, int flag, const char *name); struct mntarg *mount_argf(struct mntarg *ma, const char *name, const char *fmt, ...); struct mntarg *mount_argsu(struct mntarg *ma, const char *name, const void *val, int len); void statfs_scale_blocks(struct statfs *sf, long max_size); struct vfsconf *vfs_byname(const char *); struct vfsconf *vfs_byname_kld(const char *, struct thread *td, int *); void vfs_mount_destroy(struct mount *); void vfs_event_signal(fsid_t *, u_int32_t, intptr_t); void vfs_freeopts(struct vfsoptlist *opts); void vfs_deleteopt(struct vfsoptlist *opts, const char *name); int vfs_buildopts(struct uio *auio, struct vfsoptlist **options); int vfs_flagopt(struct vfsoptlist *opts, const char *name, uint64_t *w, uint64_t val); int vfs_getopt(struct vfsoptlist *, const char *, void **, int *); int vfs_getopt_pos(struct vfsoptlist *opts, const char *name); int vfs_getopt_size(struct vfsoptlist *opts, const char *name, off_t *value); char *vfs_getopts(struct vfsoptlist *, const char *, int *error); int vfs_copyopt(struct vfsoptlist *, const char *, void *, int); int vfs_filteropt(struct vfsoptlist *, const char **legal); void vfs_opterror(struct vfsoptlist *opts, const char *fmt, ...); int vfs_scanopt(struct vfsoptlist *opts, const char *name, const char *fmt, ...); int vfs_setopt(struct vfsoptlist *opts, const char *name, void *value, int len); int vfs_setopt_part(struct vfsoptlist *opts, const char *name, void *value, int len); int vfs_setopts(struct vfsoptlist *opts, const char *name, const char *value); int vfs_setpublicfs /* set publicly exported fs */ (struct mount *, struct netexport *, struct export_args *); void vfs_msync(struct mount *, int); int vfs_busy(struct mount *, int); int vfs_export /* process mount export info */ (struct mount *, struct export_args *); void vfs_allocate_syncvnode(struct mount *); void vfs_deallocate_syncvnode(struct mount *); int vfs_donmount(struct thread *td, uint64_t fsflags, struct uio *fsoptions); void vfs_getnewfsid(struct mount *); struct cdev *vfs_getrootfsid(struct mount *); struct mount *vfs_getvfs(fsid_t *); /* return vfs given fsid */ struct mount *vfs_busyfs(fsid_t *); int vfs_modevent(module_t, int, void *); void vfs_mount_error(struct mount *, const char *, ...); void vfs_mountroot(void); /* mount our root filesystem */ void vfs_mountedfrom(struct mount *, const char *from); void vfs_notify_upper(struct vnode *, int); void vfs_oexport_conv(const struct oexport_args *oexp, struct export_args *exp); void vfs_ref(struct mount *); void vfs_rel(struct mount *); struct mount *vfs_mount_alloc(struct vnode *, struct vfsconf *, const char *, struct ucred *); int vfs_suser(struct mount *, struct thread *); void vfs_unbusy(struct mount *); void vfs_unmountall(void); extern TAILQ_HEAD(mntlist, mount) mountlist; /* mounted filesystem list */ extern struct mtx mountlist_mtx; extern struct nfs_public nfs_pub; extern struct sx vfsconf_sx; #define vfsconf_lock() sx_xlock(&vfsconf_sx) #define vfsconf_unlock() sx_xunlock(&vfsconf_sx) #define vfsconf_slock() sx_slock(&vfsconf_sx) #define vfsconf_sunlock() sx_sunlock(&vfsconf_sx) /* * Declarations for these vfs default operations are located in * kern/vfs_default.c. They will be automatically used to replace * null entries in VFS ops tables when registering a new filesystem * type in the global table. */ vfs_root_t vfs_stdroot; vfs_quotactl_t vfs_stdquotactl; vfs_statfs_t vfs_stdstatfs; vfs_sync_t vfs_stdsync; vfs_sync_t vfs_stdnosync; vfs_vget_t vfs_stdvget; vfs_fhtovp_t vfs_stdfhtovp; vfs_checkexp_t vfs_stdcheckexp; vfs_init_t vfs_stdinit; vfs_uninit_t vfs_stduninit; vfs_extattrctl_t vfs_stdextattrctl; vfs_sysctl_t vfs_stdsysctl; +void syncer_suspend(void); +void syncer_resume(void); + #else /* !_KERNEL */ #include struct stat; __BEGIN_DECLS int fhopen(const struct fhandle *, int); int fhstat(const struct fhandle *, struct stat *); int fhstatfs(const struct fhandle *, struct statfs *); int fstatfs(int, struct statfs *); int getfh(const char *, fhandle_t *); int getfsstat(struct statfs *, long, int); int getmntinfo(struct statfs **, int); int lgetfh(const char *, fhandle_t *); int mount(const char *, const char *, int, void *); int nmount(struct iovec *, unsigned int, int); int statfs(const char *, struct statfs *); int unmount(const char *, int); /* C library stuff */ int getvfsbyname(const char *, struct xvfsconf *); __END_DECLS #endif /* _KERNEL */ #endif /* !_SYS_MOUNT_H_ */ Index: projects/building-blocks/sys/sys/param.h =================================================================== --- projects/building-blocks/sys/sys/param.h (revision 275663) +++ projects/building-blocks/sys/sys/param.h (revision 275664) @@ -1,348 +1,348 @@ /*- * Copyright (c) 1982, 1986, 1989, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)param.h 8.3 (Berkeley) 4/4/95 * $FreeBSD$ */ #ifndef _SYS_PARAM_H_ #define _SYS_PARAM_H_ #include #define BSD 199506 /* System version (year & month). */ #define BSD4_3 1 #define BSD4_4 1 /* * __FreeBSD_version numbers are documented in the Porter's Handbook. * If you bump the version for any reason, you should update the documentation * there. * Currently this lives here in the doc/ repository: * * head/en_US.ISO8859-1/books/porters-handbook/book.xml * * scheme is: Rxx * 'R' is in the range 0 to 4 if this is a release branch or * x.0-CURRENT before RELENG_*_0 is created, otherwise 'R' is * in the range 5 to 9. */ #undef __FreeBSD_version -#define __FreeBSD_version 1100048 /* Master, propagated to newvers */ +#define __FreeBSD_version 1100049 /* Master, propagated to newvers */ /* * __FreeBSD_kernel__ indicates that this system uses the kernel of FreeBSD, * which by definition is always true on FreeBSD. This macro is also defined * on other systems that use the kernel of FreeBSD, such as GNU/kFreeBSD. * * It is tempting to use this macro in userland code when we want to enable * kernel-specific routines, and in fact it's fine to do this in code that * is part of FreeBSD itself. However, be aware that as presence of this * macro is still not widespread (e.g. older FreeBSD versions, 3rd party * compilers, etc), it is STRONGLY DISCOURAGED to check for this macro in * external applications without also checking for __FreeBSD__ as an * alternative. */ #undef __FreeBSD_kernel__ #define __FreeBSD_kernel__ #ifdef _KERNEL #define P_OSREL_SIGWAIT 700000 #define P_OSREL_SIGSEGV 700004 #define P_OSREL_MAP_ANON 800104 #define P_OSREL_MAP_FSTRICT 1100036 #define P_OSREL_MAJOR(x) ((x) / 100000) #endif #ifndef LOCORE #include #endif /* * Machine-independent constants (some used in following include files). * Redefined constants are from POSIX 1003.1 limits file. * * MAXCOMLEN should be >= sizeof(ac_comm) (see ) */ #include #define MAXCOMLEN 19 /* max command name remembered */ #define MAXINTERP PATH_MAX /* max interpreter file name length */ #define MAXLOGNAME 33 /* max login name length (incl. NUL) */ #define MAXUPRC CHILD_MAX /* max simultaneous processes */ #define NCARGS ARG_MAX /* max bytes for an exec function */ #define NGROUPS (NGROUPS_MAX+1) /* max number groups */ #define NOFILE OPEN_MAX /* max open files per process */ #define NOGROUP 65535 /* marker for empty group set member */ #define MAXHOSTNAMELEN 256 /* max hostname size */ #define SPECNAMELEN 63 /* max length of devicename */ /* More types and definitions used throughout the kernel. */ #ifdef _KERNEL #include #include #ifndef LOCORE #include #include #endif #ifndef FALSE #define FALSE 0 #endif #ifndef TRUE #define TRUE 1 #endif #endif #ifndef _KERNEL /* Signals. */ #include #endif /* Machine type dependent parameters. */ #include #ifndef _KERNEL #include #endif #ifndef DEV_BSHIFT #define DEV_BSHIFT 9 /* log2(DEV_BSIZE) */ #endif #define DEV_BSIZE (1<>PAGE_SHIFT) #endif /* * btodb() is messy and perhaps slow because `bytes' may be an off_t. We * want to shift an unsigned type to avoid sign extension and we don't * want to widen `bytes' unnecessarily. Assume that the result fits in * a daddr_t. */ #ifndef btodb #define btodb(bytes) /* calculates (bytes / DEV_BSIZE) */ \ (sizeof (bytes) > sizeof(long) \ ? (daddr_t)((unsigned long long)(bytes) >> DEV_BSHIFT) \ : (daddr_t)((unsigned long)(bytes) >> DEV_BSHIFT)) #endif #ifndef dbtob #define dbtob(db) /* calculates (db * DEV_BSIZE) */ \ ((off_t)(db) << DEV_BSHIFT) #endif #define PRIMASK 0x0ff #define PCATCH 0x100 /* OR'd with pri for tsleep to check signals */ #define PDROP 0x200 /* OR'd with pri to stop re-entry of interlock mutex */ #define NZERO 0 /* default "nice" */ #define NBBY 8 /* number of bits in a byte */ #define NBPW sizeof(int) /* number of bytes per word (integer) */ #define CMASK 022 /* default file mask: S_IWGRP|S_IWOTH */ #define NODEV (dev_t)(-1) /* non-existent device */ /* * File system parameters and macros. * * MAXBSIZE - Filesystems are made out of blocks of at most MAXBSIZE bytes * per block. MAXBSIZE may be made larger without effecting * any existing filesystems as long as it does not exceed MAXPHYS, * and may be made smaller at the risk of not being able to use * filesystems which require a block size exceeding MAXBSIZE. * * BKVASIZE - Nominal buffer space per buffer, in bytes. BKVASIZE is the * minimum KVM memory reservation the kernel is willing to make. * Filesystems can of course request smaller chunks. Actual * backing memory uses a chunk size of a page (PAGE_SIZE). * * If you make BKVASIZE too small you risk seriously fragmenting * the buffer KVM map which may slow things down a bit. If you * make it too big the kernel will not be able to optimally use * the KVM memory reserved for the buffer cache and will wind * up with too-few buffers. * * The default is 16384, roughly 2x the block size used by a * normal UFS filesystem. */ #define MAXBSIZE 65536 /* must be power of 2 */ #define BKVASIZE 16384 /* must be power of 2 */ #define BKVAMASK (BKVASIZE-1) /* * MAXPATHLEN defines the longest permissible path length after expanding * symbolic links. It is used to allocate a temporary buffer from the buffer * pool in which to do the name expansion, hence should be a power of two, * and must be less than or equal to MAXBSIZE. MAXSYMLINKS defines the * maximum number of symbolic links that may be expanded in a path name. * It should be set high enough to allow all legitimate uses, but halt * infinite loops reasonably quickly. */ #define MAXPATHLEN PATH_MAX #define MAXSYMLINKS 32 /* Bit map related macros. */ #define setbit(a,i) (((unsigned char *)(a))[(i)/NBBY] |= 1<<((i)%NBBY)) #define clrbit(a,i) (((unsigned char *)(a))[(i)/NBBY] &= ~(1<<((i)%NBBY))) #define isset(a,i) \ (((const unsigned char *)(a))[(i)/NBBY] & (1<<((i)%NBBY))) #define isclr(a,i) \ ((((const unsigned char *)(a))[(i)/NBBY] & (1<<((i)%NBBY))) == 0) /* Macros for counting and rounding. */ #ifndef howmany #define howmany(x, y) (((x)+((y)-1))/(y)) #endif #define nitems(x) (sizeof((x)) / sizeof((x)[0])) #define rounddown(x, y) (((x)/(y))*(y)) #define rounddown2(x, y) ((x)&(~((y)-1))) /* if y is power of two */ #define roundup(x, y) ((((x)+((y)-1))/(y))*(y)) /* to any y */ #define roundup2(x, y) (((x)+((y)-1))&(~((y)-1))) /* if y is powers of two */ #define powerof2(x) ((((x)-1)&(x))==0) /* Macros for min/max. */ #define MIN(a,b) (((a)<(b))?(a):(b)) #define MAX(a,b) (((a)>(b))?(a):(b)) #ifdef _KERNEL /* * Basic byte order function prototypes for non-inline functions. */ #ifndef LOCORE #ifndef _BYTEORDER_PROTOTYPED #define _BYTEORDER_PROTOTYPED __BEGIN_DECLS __uint32_t htonl(__uint32_t); __uint16_t htons(__uint16_t); __uint32_t ntohl(__uint32_t); __uint16_t ntohs(__uint16_t); __END_DECLS #endif #endif #ifndef lint #ifndef _BYTEORDER_FUNC_DEFINED #define _BYTEORDER_FUNC_DEFINED #define htonl(x) __htonl(x) #define htons(x) __htons(x) #define ntohl(x) __ntohl(x) #define ntohs(x) __ntohs(x) #endif /* !_BYTEORDER_FUNC_DEFINED */ #endif /* lint */ #endif /* _KERNEL */ /* * Scale factor for scaled integers used to count %cpu time and load avgs. * * The number of CPU `tick's that map to a unique `%age' can be expressed * by the formula (1 / (2 ^ (FSHIFT - 11))). The maximum load average that * can be calculated (assuming 32 bits) can be closely approximated using * the formula (2 ^ (2 * (16 - FSHIFT))) for (FSHIFT < 15). * * For the scheduler to maintain a 1:1 mapping of CPU `tick' to `%age', * FSHIFT must be at least 11; this gives us a maximum load avg of ~1024. */ #define FSHIFT 11 /* bits to right of fixed binary point */ #define FSCALE (1<> (PAGE_SHIFT - DEV_BSHIFT)) #define ctodb(db) /* calculates pages to devblks */ \ ((db) << (PAGE_SHIFT - DEV_BSHIFT)) /* * Old spelling of __containerof(). */ #define member2struct(s, m, x) \ ((struct s *)(void *)((char *)(x) - offsetof(struct s, m))) /* * Access a variable length array that has been declared as a fixed * length array. */ #define __PAST_END(array, offset) (((__typeof__(*(array)) *)(array))[offset]) #endif /* _SYS_PARAM_H_ */ Index: projects/building-blocks/sys/sys/proc.h =================================================================== --- projects/building-blocks/sys/sys/proc.h (revision 275663) +++ projects/building-blocks/sys/sys/proc.h (revision 275664) @@ -1,1002 +1,1003 @@ /*- * Copyright (c) 1986, 1989, 1991, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. * All or some portions of this file are derived from material licensed * to the University of California by American Telephone and Telegraph * Co. or Unix System Laboratories, Inc. and are reproduced herein with * the permission of UNIX System Laboratories, Inc. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 4. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * @(#)proc.h 8.15 (Berkeley) 5/19/95 * $FreeBSD$ */ #ifndef _SYS_PROC_H_ #define _SYS_PROC_H_ #include /* For struct callout. */ #include /* For struct klist. */ #include #ifndef _KERNEL #include #endif #include #include #include #include #include #include #include /* XXX. */ #include #include #include #include #include #ifndef _KERNEL #include /* For structs itimerval, timeval. */ #else #include #endif #include #include #include /* Machine-dependent proc substruct. */ /* * One structure allocated per session. * * List of locks * (m) locked by s_mtx mtx * (e) locked by proctree_lock sx * (c) const until freeing */ struct session { u_int s_count; /* Ref cnt; pgrps in session - atomic. */ struct proc *s_leader; /* (m + e) Session leader. */ struct vnode *s_ttyvp; /* (m) Vnode of controlling tty. */ struct cdev_priv *s_ttydp; /* (m) Device of controlling tty. */ struct tty *s_ttyp; /* (e) Controlling tty. */ pid_t s_sid; /* (c) Session ID. */ /* (m) Setlogin() name: */ char s_login[roundup(MAXLOGNAME, sizeof(long))]; struct mtx s_mtx; /* Mutex to protect members. */ }; /* * One structure allocated per process group. * * List of locks * (m) locked by pg_mtx mtx * (e) locked by proctree_lock sx * (c) const until freeing */ struct pgrp { LIST_ENTRY(pgrp) pg_hash; /* (e) Hash chain. */ LIST_HEAD(, proc) pg_members; /* (m + e) Pointer to pgrp members. */ struct session *pg_session; /* (c) Pointer to session. */ struct sigiolst pg_sigiolst; /* (m) List of sigio sources. */ pid_t pg_id; /* (c) Process group id. */ int pg_jobc; /* (m) Job control process count. */ struct mtx pg_mtx; /* Mutex to protect members */ }; /* * pargs, used to hold a copy of the command line, if it had a sane length. */ struct pargs { u_int ar_ref; /* Reference count. */ u_int ar_length; /* Length. */ u_char ar_args[1]; /* Arguments. */ }; /*- * Description of a process. * * This structure contains the information needed to manage a thread of * control, known in UN*X as a process; it has references to substructures * containing descriptions of things that the process uses, but may share * with related processes. The process structure and the substructures * are always addressable except for those marked "(CPU)" below, * which might be addressable only on a processor on which the process * is running. * * Below is a key of locks used to protect each member of struct proc. The * lock is indicated by a reference to a specific character in parens in the * associated comment. * * - not yet protected * a - only touched by curproc or parent during fork/wait * b - created at fork, never changes * (exception aiods switch vmspaces, but they are also * marked 'P_SYSTEM' so hopefully it will be left alone) * c - locked by proc mtx * d - locked by allproc_lock lock * e - locked by proctree_lock lock * f - session mtx * g - process group mtx * h - callout_lock mtx * i - by curproc or the master session mtx * j - locked by proc slock * k - only accessed by curthread * k*- only accessed by curthread and from an interrupt * l - the attaching proc or attaching proc parent * m - Giant * n - not locked, lazy * o - ktrace lock * q - td_contested lock * r - p_peers lock * t - thread lock * u - process stat lock * w - process timer lock * x - created at fork, only changes during single threading in exec * y - created at first aio, doesn't change until exit or exec at which * point we are single-threaded and only curthread changes it * z - zombie threads lock * * If the locking key specifies two identifiers (for example, p_pptr) then * either lock is sufficient for read access, but both locks must be held * for write access. */ struct cpuset; struct kaioinfo; struct kaudit_record; struct kdtrace_proc; struct kdtrace_thread; struct mqueue_notifier; struct nlminfo; struct p_sched; struct proc; struct procdesc; struct racct; struct sbuf; struct sleepqueue; struct td_sched; struct thread; struct trapframe; struct turnstile; /* * XXX: Does this belong in resource.h or resourcevar.h instead? * Resource usage extension. The times in rusage structs in the kernel are * never up to date. The actual times are kept as runtimes and tick counts * (with control info in the "previous" times), and are converted when * userland asks for rusage info. Backwards compatibility prevents putting * this directly in the user-visible rusage struct. * * Locking for p_rux: (cu) means (u) for p_rux and (c) for p_crux. * Locking for td_rux: (t) for all fields. */ struct rusage_ext { uint64_t rux_runtime; /* (cu) Real time. */ uint64_t rux_uticks; /* (cu) Statclock hits in user mode. */ uint64_t rux_sticks; /* (cu) Statclock hits in sys mode. */ uint64_t rux_iticks; /* (cu) Statclock hits in intr mode. */ uint64_t rux_uu; /* (c) Previous user time in usec. */ uint64_t rux_su; /* (c) Previous sys time in usec. */ uint64_t rux_tu; /* (c) Previous total time in usec. */ }; /* * Kernel runnable context (thread). * This is what is put to sleep and reactivated. * Thread context. Processes may have multiple threads. */ struct thread { struct mtx *volatile td_lock; /* replaces sched lock */ struct proc *td_proc; /* (*) Associated process. */ TAILQ_ENTRY(thread) td_plist; /* (*) All threads in this proc. */ TAILQ_ENTRY(thread) td_runq; /* (t) Run queue. */ TAILQ_ENTRY(thread) td_slpq; /* (t) Sleep queue. */ TAILQ_ENTRY(thread) td_lockq; /* (t) Lock queue. */ LIST_ENTRY(thread) td_hash; /* (d) Hash chain. */ struct cpuset *td_cpuset; /* (t) CPU affinity mask. */ struct seltd *td_sel; /* Select queue/channel. */ struct sleepqueue *td_sleepqueue; /* (k) Associated sleep queue. */ struct turnstile *td_turnstile; /* (k) Associated turnstile. */ struct rl_q_entry *td_rlqe; /* (k) Associated range lock entry. */ struct umtx_q *td_umtxq; /* (c?) Link for when we're blocked. */ lwpid_t td_tid; /* (b) Thread ID. */ sigqueue_t td_sigqueue; /* (c) Sigs arrived, not delivered. */ #define td_siglist td_sigqueue.sq_signals u_char td_lend_user_pri; /* (t) Lend user pri. */ /* Cleared during fork1() */ #define td_startzero td_flags int td_flags; /* (t) TDF_* flags. */ int td_inhibitors; /* (t) Why can not run. */ int td_pflags; /* (k) Private thread (TDP_*) flags. */ int td_dupfd; /* (k) Ret value from fdopen. XXX */ int td_sqqueue; /* (t) Sleepqueue queue blocked on. */ void *td_wchan; /* (t) Sleep address. */ const char *td_wmesg; /* (t) Reason for sleep. */ int td_lastcpu; /* (t) Last cpu we were on. */ int td_oncpu; /* (t) Which cpu we are on. */ volatile u_char td_owepreempt; /* (k*) Preempt on last critical_exit */ u_char td_tsqueue; /* (t) Turnstile queue blocked on. */ short td_locks; /* (k) Count of non-spin locks. */ short td_rw_rlocks; /* (k) Count of rwlock read locks. */ short td_lk_slocks; /* (k) Count of lockmgr shared locks. */ short td_stopsched; /* (k) Scheduler stopped. */ struct turnstile *td_blocked; /* (t) Lock thread is blocked on. */ const char *td_lockname; /* (t) Name of lock blocked on. */ LIST_HEAD(, turnstile) td_contested; /* (q) Contested locks. */ struct lock_list_entry *td_sleeplocks; /* (k) Held sleep locks. */ int td_intr_nesting_level; /* (k) Interrupt recursion. */ int td_pinned; /* (k) Temporary cpu pin count. */ struct ucred *td_ucred; /* (k) Reference to credentials. */ u_int td_estcpu; /* (t) estimated cpu utilization */ int td_slptick; /* (t) Time at sleep. */ int td_blktick; /* (t) Time spent blocked. */ int td_swvoltick; /* (t) Time at last SW_VOL switch. */ u_int td_cow; /* (*) Number of copy-on-write faults */ struct rusage td_ru; /* (t) rusage information. */ struct rusage_ext td_rux; /* (t) Internal rusage information. */ uint64_t td_incruntime; /* (t) Cpu ticks to transfer to proc. */ uint64_t td_runtime; /* (t) How many cpu ticks we've run. */ u_int td_pticks; /* (t) Statclock hits for profiling */ u_int td_sticks; /* (t) Statclock hits in system mode. */ u_int td_iticks; /* (t) Statclock hits in intr mode. */ u_int td_uticks; /* (t) Statclock hits in user mode. */ int td_intrval; /* (t) Return value for sleepq. */ sigset_t td_oldsigmask; /* (k) Saved mask from pre sigpause. */ volatile u_int td_generation; /* (k) For detection of preemption */ stack_t td_sigstk; /* (k) Stack ptr and on-stack flag. */ int td_xsig; /* (c) Signal for ptrace */ u_long td_profil_addr; /* (k) Temporary addr until AST. */ u_int td_profil_ticks; /* (k) Temporary ticks until AST. */ char td_name[MAXCOMLEN + 1]; /* (*) Thread name. */ struct file *td_fpop; /* (k) file referencing cdev under op */ int td_dbgflags; /* (c) Userland debugger flags */ struct ksiginfo td_dbgksi; /* (c) ksi reflected to debugger. */ int td_ng_outbound; /* (k) Thread entered ng from above. */ struct osd td_osd; /* (k) Object specific data. */ struct vm_map_entry *td_map_def_user; /* (k) Deferred entries. */ pid_t td_dbg_forked; /* (c) Child pid for debugger. */ u_int td_vp_reserv; /* (k) Count of reserved vnodes. */ int td_no_sleeping; /* (k) Sleeping disabled count. */ int td_dom_rr_idx; /* (k) RR Numa domain selection. */ #define td_endzero td_sigmask /* Copied during fork1() or create_thread(). */ #define td_startcopy td_endzero sigset_t td_sigmask; /* (c) Current signal mask. */ u_char td_rqindex; /* (t) Run queue index. */ u_char td_base_pri; /* (t) Thread base kernel priority. */ u_char td_priority; /* (t) Thread active priority. */ u_char td_pri_class; /* (t) Scheduling class. */ u_char td_user_pri; /* (t) User pri from estcpu and nice. */ u_char td_base_user_pri; /* (t) Base user pri */ #define td_endcopy td_pcb /* * Fields that must be manually set in fork1() or create_thread() * or already have been set in the allocator, constructor, etc. */ struct pcb *td_pcb; /* (k) Kernel VA of pcb and kstack. */ enum { TDS_INACTIVE = 0x0, TDS_INHIBITED, TDS_CAN_RUN, TDS_RUNQ, TDS_RUNNING } td_state; /* (t) thread state */ union { register_t tdu_retval[2]; off_t tdu_off; } td_uretoff; /* (k) Syscall aux returns. */ #define td_retval td_uretoff.tdu_retval struct callout td_slpcallout; /* (h) Callout for sleep. */ struct trapframe *td_frame; /* (k) */ struct vm_object *td_kstack_obj;/* (a) Kstack object. */ vm_offset_t td_kstack; /* (a) Kernel VA of kstack. */ int td_kstack_pages; /* (a) Size of the kstack. */ volatile u_int td_critnest; /* (k*) Critical section nest level. */ struct mdthread td_md; /* (k) Any machine-dependent fields. */ struct td_sched *td_sched; /* (*) Scheduler-specific data. */ struct kaudit_record *td_ar; /* (k) Active audit record, if any. */ struct lpohead td_lprof[2]; /* (a) lock profiling objects. */ struct kdtrace_thread *td_dtrace; /* (*) DTrace-specific data. */ int td_errno; /* Error returned by last syscall. */ struct vnet *td_vnet; /* (k) Effective vnet. */ const char *td_vnet_lpush; /* (k) Debugging vnet push / pop. */ struct trapframe *td_intr_frame;/* (k) Frame of the current irq */ struct proc *td_rfppwait_p; /* (k) The vforked child */ struct vm_page **td_ma; /* (k) uio pages held */ int td_ma_cnt; /* (k) size of *td_ma */ }; struct mtx *thread_lock_block(struct thread *); void thread_lock_unblock(struct thread *, struct mtx *); void thread_lock_set(struct thread *, struct mtx *); #define THREAD_LOCK_ASSERT(td, type) \ do { \ struct mtx *__m = (td)->td_lock; \ if (__m != &blocked_lock) \ mtx_assert(__m, (type)); \ } while (0) #ifdef INVARIANTS #define THREAD_LOCKPTR_ASSERT(td, lock) \ do { \ struct mtx *__m = (td)->td_lock; \ KASSERT((__m == &blocked_lock || __m == (lock)), \ ("Thread %p lock %p does not match %p", td, __m, (lock))); \ } while (0) #else #define THREAD_LOCKPTR_ASSERT(td, lock) #endif /* * Flags kept in td_flags: * To change these you MUST have the scheduler lock. */ #define TDF_BORROWING 0x00000001 /* Thread is borrowing pri from another. */ #define TDF_INPANIC 0x00000002 /* Caused a panic, let it drive crashdump. */ #define TDF_INMEM 0x00000004 /* Thread's stack is in memory. */ #define TDF_SINTR 0x00000008 /* Sleep is interruptible. */ #define TDF_TIMEOUT 0x00000010 /* Timing out during sleep. */ #define TDF_IDLETD 0x00000020 /* This is a per-CPU idle thread. */ #define TDF_CANSWAP 0x00000040 /* Thread can be swapped. */ #define TDF_SLEEPABORT 0x00000080 /* sleepq_abort was called. */ #define TDF_KTH_SUSP 0x00000100 /* kthread is suspended */ #define TDF_UNUSED09 0x00000200 /* --available-- */ #define TDF_BOUNDARY 0x00000400 /* Thread suspended at user boundary */ #define TDF_ASTPENDING 0x00000800 /* Thread has some asynchronous events. */ #define TDF_TIMOFAIL 0x00001000 /* Timeout from sleep after we were awake. */ #define TDF_SBDRY 0x00002000 /* Stop only on usermode boundary. */ #define TDF_UPIBLOCKED 0x00004000 /* Thread blocked on user PI mutex. */ #define TDF_NEEDSUSPCHK 0x00008000 /* Thread may need to suspend. */ #define TDF_NEEDRESCHED 0x00010000 /* Thread needs to yield. */ #define TDF_NEEDSIGCHK 0x00020000 /* Thread may need signal delivery. */ #define TDF_NOLOAD 0x00040000 /* Ignore during load avg calculations. */ #define TDF_UNUSED19 0x00080000 /* --available-- */ #define TDF_THRWAKEUP 0x00100000 /* Libthr thread must not suspend itself. */ #define TDF_UNUSED21 0x00200000 /* --available-- */ #define TDF_SWAPINREQ 0x00400000 /* Swapin request due to wakeup. */ #define TDF_UNUSED23 0x00800000 /* --available-- */ #define TDF_SCHED0 0x01000000 /* Reserved for scheduler private use */ #define TDF_SCHED1 0x02000000 /* Reserved for scheduler private use */ #define TDF_SCHED2 0x04000000 /* Reserved for scheduler private use */ #define TDF_SCHED3 0x08000000 /* Reserved for scheduler private use */ #define TDF_ALRMPEND 0x10000000 /* Pending SIGVTALRM needs to be posted. */ #define TDF_PROFPEND 0x20000000 /* Pending SIGPROF needs to be posted. */ #define TDF_MACPEND 0x40000000 /* AST-based MAC event pending. */ /* Userland debug flags */ #define TDB_SUSPEND 0x00000001 /* Thread is suspended by debugger */ #define TDB_XSIG 0x00000002 /* Thread is exchanging signal under trace */ #define TDB_USERWR 0x00000004 /* Debugger modified memory or registers */ #define TDB_SCE 0x00000008 /* Thread performs syscall enter */ #define TDB_SCX 0x00000010 /* Thread performs syscall exit */ #define TDB_EXEC 0x00000020 /* TDB_SCX from exec(2) family */ #define TDB_FORK 0x00000040 /* TDB_SCX from fork(2) that created new process */ #define TDB_STOPATFORK 0x00000080 /* Stop at the return from fork (child only) */ #define TDB_CHILD 0x00000100 /* New child indicator for ptrace() */ /* * "Private" flags kept in td_pflags: * These are only written by curthread and thus need no locking. */ #define TDP_OLDMASK 0x00000001 /* Need to restore mask after suspend. */ #define TDP_INKTR 0x00000002 /* Thread is currently in KTR code. */ #define TDP_INKTRACE 0x00000004 /* Thread is currently in KTRACE code. */ #define TDP_BUFNEED 0x00000008 /* Do not recurse into the buf flush */ #define TDP_COWINPROGRESS 0x00000010 /* Snapshot copy-on-write in progress. */ #define TDP_ALTSTACK 0x00000020 /* Have alternate signal stack. */ #define TDP_DEADLKTREAT 0x00000040 /* Lock aquisition - deadlock treatment. */ #define TDP_NOFAULTING 0x00000080 /* Do not handle page faults. */ #define TDP_UNUSED9 0x00000100 /* --available-- */ #define TDP_OWEUPC 0x00000200 /* Call addupc() at next AST. */ #define TDP_ITHREAD 0x00000400 /* Thread is an interrupt thread. */ #define TDP_SYNCIO 0x00000800 /* Local override, disable async i/o. */ #define TDP_SCHED1 0x00001000 /* Reserved for scheduler private use */ #define TDP_SCHED2 0x00002000 /* Reserved for scheduler private use */ #define TDP_SCHED3 0x00004000 /* Reserved for scheduler private use */ #define TDP_SCHED4 0x00008000 /* Reserved for scheduler private use */ #define TDP_GEOM 0x00010000 /* Settle GEOM before finishing syscall */ #define TDP_SOFTDEP 0x00020000 /* Stuck processing softdep worklist */ #define TDP_NORUNNINGBUF 0x00040000 /* Ignore runningbufspace check */ #define TDP_WAKEUP 0x00080000 /* Don't sleep in umtx cond_wait */ #define TDP_INBDFLUSH 0x00100000 /* Already in BO_BDFLUSH, do not recurse */ #define TDP_KTHREAD 0x00200000 /* This is an official kernel thread */ #define TDP_CALLCHAIN 0x00400000 /* Capture thread's callchain */ #define TDP_IGNSUSP 0x00800000 /* Permission to ignore the MNTK_SUSPEND* */ #define TDP_AUDITREC 0x01000000 /* Audit record pending on thread */ #define TDP_RFPPWAIT 0x02000000 /* Handle RFPPWAIT on syscall exit */ #define TDP_RESETSPUR 0x04000000 /* Reset spurious page fault history. */ #define TDP_NERRNO 0x08000000 /* Last errno is already in td_errno */ #define TDP_UIOHELD 0x10000000 /* Current uio has pages held in td_ma */ #define TDP_DEVMEMIO 0x20000000 /* Accessing memory for /dev/mem */ #define TDP_EXECVMSPC 0x40000000 /* Execve destroyed old vmspace */ /* * Reasons that the current thread can not be run yet. * More than one may apply. */ #define TDI_SUSPENDED 0x0001 /* On suspension queue. */ #define TDI_SLEEPING 0x0002 /* Actually asleep! (tricky). */ #define TDI_SWAPPED 0x0004 /* Stack not in mem. Bad juju if run. */ #define TDI_LOCK 0x0008 /* Stopped on a lock. */ #define TDI_IWAIT 0x0010 /* Awaiting interrupt. */ #define TD_IS_SLEEPING(td) ((td)->td_inhibitors & TDI_SLEEPING) #define TD_ON_SLEEPQ(td) ((td)->td_wchan != NULL) #define TD_IS_SUSPENDED(td) ((td)->td_inhibitors & TDI_SUSPENDED) #define TD_IS_SWAPPED(td) ((td)->td_inhibitors & TDI_SWAPPED) #define TD_ON_LOCK(td) ((td)->td_inhibitors & TDI_LOCK) #define TD_AWAITING_INTR(td) ((td)->td_inhibitors & TDI_IWAIT) #define TD_IS_RUNNING(td) ((td)->td_state == TDS_RUNNING) #define TD_ON_RUNQ(td) ((td)->td_state == TDS_RUNQ) #define TD_CAN_RUN(td) ((td)->td_state == TDS_CAN_RUN) #define TD_IS_INHIBITED(td) ((td)->td_state == TDS_INHIBITED) #define TD_ON_UPILOCK(td) ((td)->td_flags & TDF_UPIBLOCKED) #define TD_IS_IDLETHREAD(td) ((td)->td_flags & TDF_IDLETD) #define TD_SET_INHIB(td, inhib) do { \ (td)->td_state = TDS_INHIBITED; \ (td)->td_inhibitors |= (inhib); \ } while (0) #define TD_CLR_INHIB(td, inhib) do { \ if (((td)->td_inhibitors & (inhib)) && \ (((td)->td_inhibitors &= ~(inhib)) == 0)) \ (td)->td_state = TDS_CAN_RUN; \ } while (0) #define TD_SET_SLEEPING(td) TD_SET_INHIB((td), TDI_SLEEPING) #define TD_SET_SWAPPED(td) TD_SET_INHIB((td), TDI_SWAPPED) #define TD_SET_LOCK(td) TD_SET_INHIB((td), TDI_LOCK) #define TD_SET_SUSPENDED(td) TD_SET_INHIB((td), TDI_SUSPENDED) #define TD_SET_IWAIT(td) TD_SET_INHIB((td), TDI_IWAIT) #define TD_SET_EXITING(td) TD_SET_INHIB((td), TDI_EXITING) #define TD_CLR_SLEEPING(td) TD_CLR_INHIB((td), TDI_SLEEPING) #define TD_CLR_SWAPPED(td) TD_CLR_INHIB((td), TDI_SWAPPED) #define TD_CLR_LOCK(td) TD_CLR_INHIB((td), TDI_LOCK) #define TD_CLR_SUSPENDED(td) TD_CLR_INHIB((td), TDI_SUSPENDED) #define TD_CLR_IWAIT(td) TD_CLR_INHIB((td), TDI_IWAIT) #define TD_SET_RUNNING(td) (td)->td_state = TDS_RUNNING #define TD_SET_RUNQ(td) (td)->td_state = TDS_RUNQ #define TD_SET_CAN_RUN(td) (td)->td_state = TDS_CAN_RUN /* * Process structure. */ struct proc { LIST_ENTRY(proc) p_list; /* (d) List of all processes. */ TAILQ_HEAD(, thread) p_threads; /* (c) all threads. */ struct mtx p_slock; /* process spin lock */ struct ucred *p_ucred; /* (c) Process owner's identity. */ struct filedesc *p_fd; /* (b) Open files. */ struct filedesc_to_leader *p_fdtol; /* (b) Tracking node */ struct pstats *p_stats; /* (b) Accounting/statistics (CPU). */ struct plimit *p_limit; /* (c) Process limits. */ struct callout p_limco; /* (c) Limit callout handle */ struct sigacts *p_sigacts; /* (x) Signal actions, state (CPU). */ int p_flag; /* (c) P_* flags. */ int p_flag2; /* (c) P2_* flags. */ enum { PRS_NEW = 0, /* In creation */ PRS_NORMAL, /* threads can be run. */ PRS_ZOMBIE } p_state; /* (j/c) Process status. */ pid_t p_pid; /* (b) Process identifier. */ LIST_ENTRY(proc) p_hash; /* (d) Hash chain. */ LIST_ENTRY(proc) p_pglist; /* (g + e) List of processes in pgrp. */ struct proc *p_pptr; /* (c + e) Pointer to parent process. */ LIST_ENTRY(proc) p_sibling; /* (e) List of sibling processes. */ LIST_HEAD(, proc) p_children; /* (e) Pointer to list of children. */ struct mtx p_mtx; /* (n) Lock for this struct. */ struct mtx p_statmtx; /* Lock for the stats */ struct mtx p_itimmtx; /* Lock for the virt/prof timers */ struct mtx p_profmtx; /* Lock for the profiling */ struct ksiginfo *p_ksi; /* Locked by parent proc lock */ sigqueue_t p_sigqueue; /* (c) Sigs not delivered to a td. */ #define p_siglist p_sigqueue.sq_signals /* The following fields are all zeroed upon creation in fork. */ #define p_startzero p_oppid pid_t p_oppid; /* (c + e) Save ppid in ptrace. XXX */ struct vmspace *p_vmspace; /* (b) Address space. */ u_int p_swtick; /* (c) Tick when swapped in or out. */ struct itimerval p_realtimer; /* (c) Alarm timer. */ struct rusage p_ru; /* (a) Exit information. */ struct rusage_ext p_rux; /* (cu) Internal resource usage. */ struct rusage_ext p_crux; /* (c) Internal child resource usage. */ int p_profthreads; /* (c) Num threads in addupc_task. */ volatile int p_exitthreads; /* (j) Number of threads exiting */ int p_traceflag; /* (o) Kernel trace points. */ struct vnode *p_tracevp; /* (c + o) Trace to vnode. */ struct ucred *p_tracecred; /* (o) Credentials to trace with. */ struct vnode *p_textvp; /* (b) Vnode of executable. */ u_int p_lock; /* (c) Proclock (prevent swap) count. */ struct sigiolst p_sigiolst; /* (c) List of sigio sources. */ int p_sigparent; /* (c) Signal to parent on exit. */ int p_sig; /* (n) For core dump/debugger XXX. */ u_long p_code; /* (n) For core dump/debugger XXX. */ u_int p_stops; /* (c) Stop event bitmask. */ u_int p_stype; /* (c) Stop event type. */ char p_step; /* (c) Process is stopped. */ u_char p_pfsflags; /* (c) Procfs flags. */ struct nlminfo *p_nlminfo; /* (?) Only used by/for lockd. */ struct kaioinfo *p_aioinfo; /* (y) ASYNC I/O info. */ struct thread *p_singlethread;/* (c + j) If single threading this is it */ int p_suspcount; /* (j) Num threads in suspended mode. */ struct thread *p_xthread; /* (c) Trap thread */ int p_boundary_count;/* (j) Num threads at user boundary */ int p_pendingcnt; /* how many signals are pending */ struct itimers *p_itimers; /* (c) POSIX interval timers. */ struct procdesc *p_procdesc; /* (e) Process descriptor, if any. */ u_int p_treeflag; /* (e) P_TREE flags */ /* End area that is zeroed on creation. */ #define p_endzero p_magic /* The following fields are all copied upon creation in fork. */ #define p_startcopy p_endzero u_int p_magic; /* (b) Magic number. */ int p_osrel; /* (x) osreldate for the binary (from ELF note, if any) */ char p_comm[MAXCOMLEN + 1]; /* (b) Process name. */ struct pgrp *p_pgrp; /* (c + e) Pointer to process group. */ struct sysentvec *p_sysent; /* (b) Syscall dispatch info. */ struct pargs *p_args; /* (c) Process arguments. */ rlim_t p_cpulimit; /* (c) Current CPU limit in seconds. */ signed char p_nice; /* (c) Process "nice" value. */ int p_fibnum; /* in this routing domain XXX MRT */ /* End area that is copied on creation. */ #define p_endcopy p_xstat u_short p_xstat; /* (c) Exit status; also stop sig. */ struct knlist p_klist; /* (c) Knotes attached to this proc. */ int p_numthreads; /* (c) Number of threads. */ struct mdproc p_md; /* Any machine-dependent fields. */ struct callout p_itcallout; /* (h + c) Interval timer callout. */ u_short p_acflag; /* (c) Accounting flags. */ struct proc *p_peers; /* (r) */ struct proc *p_leader; /* (b) */ void *p_emuldata; /* (c) Emulator state data. */ struct label *p_label; /* (*) Proc (not subject) MAC label. */ struct p_sched *p_sched; /* (*) Scheduler-specific data. */ STAILQ_HEAD(, ktr_request) p_ktr; /* (o) KTR event queue. */ LIST_HEAD(, mqueue_notifier) p_mqnotifier; /* (c) mqueue notifiers.*/ struct kdtrace_proc *p_dtrace; /* (*) DTrace-specific data. */ struct cv p_pwait; /* (*) wait cv for exit/exec. */ struct cv p_dbgwait; /* (*) wait cv for debugger attach after fork. */ uint64_t p_prev_runtime; /* (c) Resource usage accounting. */ struct racct *p_racct; /* (b) Resource accounting. */ u_char p_throttled; /* (c) Flag for racct pcpu throttling */ /* * An orphan is the child that has beed re-parented to the * debugger as a result of attaching to it. Need to keep * track of them for parent to be able to collect the exit * status of what used to be children. */ LIST_ENTRY(proc) p_orphan; /* (e) List of orphan processes. */ LIST_HEAD(, proc) p_orphans; /* (e) Pointer to list of orphans. */ }; #define p_session p_pgrp->pg_session #define p_pgid p_pgrp->pg_id #define NOCPU (-1) /* For when we aren't on a CPU. */ #define NOCPU_OLD (255) #define MAXCPU_OLD (254) #define PROC_SLOCK(p) mtx_lock_spin(&(p)->p_slock) #define PROC_SUNLOCK(p) mtx_unlock_spin(&(p)->p_slock) #define PROC_SLOCK_ASSERT(p, type) mtx_assert(&(p)->p_slock, (type)) #define PROC_STATLOCK(p) mtx_lock_spin(&(p)->p_statmtx) #define PROC_STATUNLOCK(p) mtx_unlock_spin(&(p)->p_statmtx) #define PROC_STATLOCK_ASSERT(p, type) mtx_assert(&(p)->p_statmtx, (type)) #define PROC_ITIMLOCK(p) mtx_lock_spin(&(p)->p_itimmtx) #define PROC_ITIMUNLOCK(p) mtx_unlock_spin(&(p)->p_itimmtx) #define PROC_ITIMLOCK_ASSERT(p, type) mtx_assert(&(p)->p_itimmtx, (type)) #define PROC_PROFLOCK(p) mtx_lock_spin(&(p)->p_profmtx) #define PROC_PROFUNLOCK(p) mtx_unlock_spin(&(p)->p_profmtx) #define PROC_PROFLOCK_ASSERT(p, type) mtx_assert(&(p)->p_profmtx, (type)) /* These flags are kept in p_flag. */ #define P_ADVLOCK 0x00001 /* Process may hold a POSIX advisory lock. */ #define P_CONTROLT 0x00002 /* Has a controlling terminal. */ #define P_KTHREAD 0x00004 /* Kernel thread (*). */ #define P_FOLLOWFORK 0x00008 /* Attach parent debugger to children. */ #define P_PPWAIT 0x00010 /* Parent is waiting for child to exec/exit. */ #define P_PROFIL 0x00020 /* Has started profiling. */ #define P_STOPPROF 0x00040 /* Has thread requesting to stop profiling. */ #define P_HADTHREADS 0x00080 /* Has had threads (no cleanup shortcuts) */ #define P_SUGID 0x00100 /* Had set id privileges since last exec. */ #define P_SYSTEM 0x00200 /* System proc: no sigs, stats or swapping. */ #define P_SINGLE_EXIT 0x00400 /* Threads suspending should exit, not wait. */ #define P_TRACED 0x00800 /* Debugged process being traced. */ #define P_WAITED 0x01000 /* Someone is waiting for us. */ #define P_WEXIT 0x02000 /* Working on exiting. */ #define P_EXEC 0x04000 /* Process called exec. */ #define P_WKILLED 0x08000 /* Killed, go to kernel/user boundary ASAP. */ #define P_CONTINUED 0x10000 /* Proc has continued from a stopped state. */ #define P_STOPPED_SIG 0x20000 /* Stopped due to SIGSTOP/SIGTSTP. */ #define P_STOPPED_TRACE 0x40000 /* Stopped because of tracing. */ #define P_STOPPED_SINGLE 0x80000 /* Only 1 thread can continue (not to user). */ #define P_PROTECTED 0x100000 /* Do not kill on memory overcommit. */ #define P_SIGEVENT 0x200000 /* Process pending signals changed. */ #define P_SINGLE_BOUNDARY 0x400000 /* Threads should suspend at user boundary. */ #define P_HWPMC 0x800000 /* Process is using HWPMCs */ #define P_JAILED 0x1000000 /* Process is in jail. */ #define P_UNUSED1 0x2000000 #define P_INEXEC 0x4000000 /* Process is in execve(). */ #define P_STATCHILD 0x8000000 /* Child process stopped or exited. */ #define P_INMEM 0x10000000 /* Loaded into memory. */ #define P_SWAPPINGOUT 0x20000000 /* Process is being swapped out. */ #define P_SWAPPINGIN 0x40000000 /* Process is being swapped in. */ #define P_PPTRACE 0x80000000 /* PT_TRACEME by vforked child. */ #define P_STOPPED (P_STOPPED_SIG|P_STOPPED_SINGLE|P_STOPPED_TRACE) #define P_SHOULDSTOP(p) ((p)->p_flag & P_STOPPED) #define P_KILLED(p) ((p)->p_flag & P_WKILLED) /* These flags are kept in p_flag2. */ #define P2_INHERIT_PROTECTED 0x00000001 /* New children get P_PROTECTED. */ /* Flags protected by proctree_lock, kept in p_treeflags. */ #define P_TREE_ORPHANED 0x00000001 /* Reparented, on orphan list */ #define P_TREE_FIRST_ORPHAN 0x00000002 /* First element of orphan list */ /* * These were process status values (p_stat), now they are only used in * legacy conversion code. */ #define SIDL 1 /* Process being created by fork. */ #define SRUN 2 /* Currently runnable. */ #define SSLEEP 3 /* Sleeping on an address. */ #define SSTOP 4 /* Process debugging or suspension. */ #define SZOMB 5 /* Awaiting collection by parent. */ #define SWAIT 6 /* Waiting for interrupt. */ #define SLOCK 7 /* Blocked on a lock. */ #define P_MAGIC 0xbeefface #ifdef _KERNEL /* Types and flags for mi_switch(). */ #define SW_TYPE_MASK 0xff /* First 8 bits are switch type */ #define SWT_NONE 0 /* Unspecified switch. */ #define SWT_PREEMPT 1 /* Switching due to preemption. */ #define SWT_OWEPREEMPT 2 /* Switching due to opepreempt. */ #define SWT_TURNSTILE 3 /* Turnstile contention. */ #define SWT_SLEEPQ 4 /* Sleepq wait. */ #define SWT_SLEEPQTIMO 5 /* Sleepq timeout wait. */ #define SWT_RELINQUISH 6 /* yield call. */ #define SWT_NEEDRESCHED 7 /* NEEDRESCHED was set. */ #define SWT_IDLE 8 /* Switching from the idle thread. */ #define SWT_IWAIT 9 /* Waiting for interrupts. */ #define SWT_SUSPEND 10 /* Thread suspended. */ #define SWT_REMOTEPREEMPT 11 /* Remote processor preempted. */ #define SWT_REMOTEWAKEIDLE 12 /* Remote processor preempted idle. */ #define SWT_COUNT 13 /* Number of switch types. */ /* Flags */ #define SW_VOL 0x0100 /* Voluntary switch. */ #define SW_INVOL 0x0200 /* Involuntary switch. */ #define SW_PREEMPT 0x0400 /* The invol switch is a preemption */ /* How values for thread_single(). */ #define SINGLE_NO_EXIT 0 #define SINGLE_EXIT 1 #define SINGLE_BOUNDARY 2 #ifdef MALLOC_DECLARE MALLOC_DECLARE(M_PARGS); MALLOC_DECLARE(M_PGRP); MALLOC_DECLARE(M_SESSION); MALLOC_DECLARE(M_SUBPROC); #endif #define FOREACH_PROC_IN_SYSTEM(p) \ LIST_FOREACH((p), &allproc, p_list) #define FOREACH_THREAD_IN_PROC(p, td) \ TAILQ_FOREACH((td), &(p)->p_threads, td_plist) #define FIRST_THREAD_IN_PROC(p) TAILQ_FIRST(&(p)->p_threads) /* * We use process IDs <= pid_max <= PID_MAX; PID_MAX + 1 must also fit * in a pid_t, as it is used to represent "no process group". */ #define PID_MAX 99999 #define NO_PID 100000 extern pid_t pid_max; #define SESS_LEADER(p) ((p)->p_session->s_leader == (p)) #define STOPEVENT(p, e, v) do { \ if ((p)->p_stops & (e)) { \ PROC_LOCK(p); \ stopevent((p), (e), (v)); \ PROC_UNLOCK(p); \ } \ } while (0) #define _STOPEVENT(p, e, v) do { \ PROC_LOCK_ASSERT(p, MA_OWNED); \ WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, &p->p_mtx.lock_object, \ "checking stopevent %d", (e)); \ if ((p)->p_stops & (e)) \ stopevent((p), (e), (v)); \ } while (0) /* Lock and unlock a process. */ #define PROC_LOCK(p) mtx_lock(&(p)->p_mtx) #define PROC_TRYLOCK(p) mtx_trylock(&(p)->p_mtx) #define PROC_UNLOCK(p) mtx_unlock(&(p)->p_mtx) #define PROC_LOCKED(p) mtx_owned(&(p)->p_mtx) #define PROC_LOCK_ASSERT(p, type) mtx_assert(&(p)->p_mtx, (type)) /* Lock and unlock a process group. */ #define PGRP_LOCK(pg) mtx_lock(&(pg)->pg_mtx) #define PGRP_UNLOCK(pg) mtx_unlock(&(pg)->pg_mtx) #define PGRP_LOCKED(pg) mtx_owned(&(pg)->pg_mtx) #define PGRP_LOCK_ASSERT(pg, type) mtx_assert(&(pg)->pg_mtx, (type)) #define PGRP_LOCK_PGSIGNAL(pg) do { \ if ((pg) != NULL) \ PGRP_LOCK(pg); \ } while (0) #define PGRP_UNLOCK_PGSIGNAL(pg) do { \ if ((pg) != NULL) \ PGRP_UNLOCK(pg); \ } while (0) /* Lock and unlock a session. */ #define SESS_LOCK(s) mtx_lock(&(s)->s_mtx) #define SESS_UNLOCK(s) mtx_unlock(&(s)->s_mtx) #define SESS_LOCKED(s) mtx_owned(&(s)->s_mtx) #define SESS_LOCK_ASSERT(s, type) mtx_assert(&(s)->s_mtx, (type)) /* Hold process U-area in memory, normally for ptrace/procfs work. */ #define PHOLD(p) do { \ PROC_LOCK(p); \ _PHOLD(p); \ PROC_UNLOCK(p); \ } while (0) #define _PHOLD(p) do { \ PROC_LOCK_ASSERT((p), MA_OWNED); \ KASSERT(!((p)->p_flag & P_WEXIT) || (p) == curproc, \ ("PHOLD of exiting process")); \ (p)->p_lock++; \ if (((p)->p_flag & P_INMEM) == 0) \ faultin((p)); \ } while (0) #define PROC_ASSERT_HELD(p) do { \ KASSERT((p)->p_lock > 0, ("process not held")); \ } while (0) #define PRELE(p) do { \ PROC_LOCK((p)); \ _PRELE((p)); \ PROC_UNLOCK((p)); \ } while (0) #define _PRELE(p) do { \ PROC_LOCK_ASSERT((p), MA_OWNED); \ PROC_ASSERT_HELD(p); \ (--(p)->p_lock); \ if (((p)->p_flag & P_WEXIT) && (p)->p_lock == 0) \ wakeup(&(p)->p_lock); \ } while (0) #define PROC_ASSERT_NOT_HELD(p) do { \ KASSERT((p)->p_lock == 0, ("process held")); \ } while (0) /* Check whether a thread is safe to be swapped out. */ #define thread_safetoswapout(td) ((td)->td_flags & TDF_CANSWAP) /* Control whether or not it is safe for curthread to sleep. */ #define THREAD_NO_SLEEPING() ((curthread)->td_no_sleeping++) #define THREAD_SLEEPING_OK() ((curthread)->td_no_sleeping--) #define THREAD_CAN_SLEEP() ((curthread)->td_no_sleeping == 0) #define PIDHASH(pid) (&pidhashtbl[(pid) & pidhash]) extern LIST_HEAD(pidhashhead, proc) *pidhashtbl; extern u_long pidhash; #define TIDHASH(tid) (&tidhashtbl[(tid) & tidhash]) extern LIST_HEAD(tidhashhead, thread) *tidhashtbl; extern u_long tidhash; extern struct rwlock tidhash_lock; #define PGRPHASH(pgid) (&pgrphashtbl[(pgid) & pgrphash]) extern LIST_HEAD(pgrphashhead, pgrp) *pgrphashtbl; extern u_long pgrphash; extern struct sx allproc_lock; extern struct sx proctree_lock; extern struct mtx ppeers_lock; extern struct proc proc0; /* Process slot for swapper. */ extern struct thread thread0; /* Primary thread in proc0. */ extern struct vmspace vmspace0; /* VM space for proc0. */ extern int hogticks; /* Limit on kernel cpu hogs. */ extern int lastpid; extern int nprocs, maxproc; /* Current and max number of procs. */ extern int maxprocperuid; /* Max procs per uid. */ extern u_long ps_arg_cache_limit; LIST_HEAD(proclist, proc); TAILQ_HEAD(procqueue, proc); TAILQ_HEAD(threadqueue, thread); extern struct proclist allproc; /* List of all processes. */ extern struct proclist zombproc; /* List of zombie processes. */ extern struct proc *initproc, *pageproc; /* Process slots for init, pager. */ extern struct uma_zone *proc_zone; struct proc *pfind(pid_t); /* Find process by id. */ struct proc *pfind_locked(pid_t pid); struct pgrp *pgfind(pid_t); /* Find process group by id. */ struct proc *zpfind(pid_t); /* Find zombie process by id. */ /* * pget() flags. */ #define PGET_HOLD 0x00001 /* Hold the process. */ #define PGET_CANSEE 0x00002 /* Check against p_cansee(). */ #define PGET_CANDEBUG 0x00004 /* Check against p_candebug(). */ #define PGET_ISCURRENT 0x00008 /* Check that the found process is current. */ #define PGET_NOTWEXIT 0x00010 /* Check that the process is not in P_WEXIT. */ #define PGET_NOTINEXEC 0x00020 /* Check that the process is not in P_INEXEC. */ #define PGET_NOTID 0x00040 /* Do not assume tid if pid > PID_MAX. */ #define PGET_WANTREAD (PGET_HOLD | PGET_CANDEBUG | PGET_NOTWEXIT) int pget(pid_t pid, int flags, struct proc **pp); void ast(struct trapframe *framep); struct thread *choosethread(void); int cr_cansignal(struct ucred *cred, struct proc *proc, int signum); int enterpgrp(struct proc *p, pid_t pgid, struct pgrp *pgrp, struct session *sess); int enterthispgrp(struct proc *p, struct pgrp *pgrp); void faultin(struct proc *p); void fixjobc(struct proc *p, struct pgrp *pgrp, int entering); int fork1(struct thread *, int, int, struct proc **, int *, int); void fork_exit(void (*)(void *, struct trapframe *), void *, struct trapframe *); void fork_return(struct thread *, struct trapframe *); int inferior(struct proc *p); void kern_yield(int); void kick_proc0(void); int leavepgrp(struct proc *p); int maybe_preempt(struct thread *td); void maybe_yield(void); void mi_switch(int flags, struct thread *newtd); int p_candebug(struct thread *td, struct proc *p); int p_cansee(struct thread *td, struct proc *p); int p_cansched(struct thread *td, struct proc *p); int p_cansignal(struct thread *td, struct proc *p, int signum); int p_canwait(struct thread *td, struct proc *p); struct pargs *pargs_alloc(int len); void pargs_drop(struct pargs *pa); void pargs_hold(struct pargs *pa); int proc_getargv(struct thread *td, struct proc *p, struct sbuf *sb); int proc_getauxv(struct thread *td, struct proc *p, struct sbuf *sb); int proc_getenvv(struct thread *td, struct proc *p, struct sbuf *sb); void procinit(void); void proc_linkup0(struct proc *p, struct thread *td); void proc_linkup(struct proc *p, struct thread *td); struct proc *proc_realparent(struct proc *child); void proc_reap(struct thread *td, struct proc *p, int *status, int options); void proc_reparent(struct proc *child, struct proc *newparent); struct pstats *pstats_alloc(void); void pstats_fork(struct pstats *src, struct pstats *dst); void pstats_free(struct pstats *ps); int securelevel_ge(struct ucred *cr, int level); int securelevel_gt(struct ucred *cr, int level); void sess_hold(struct session *); void sess_release(struct session *); int setrunnable(struct thread *); void setsugid(struct proc *p); int should_yield(void); int sigonstack(size_t sp); void stopevent(struct proc *, u_int, u_int); struct thread *tdfind(lwpid_t, pid_t); void threadinit(void); void tidhash_add(struct thread *); void tidhash_remove(struct thread *); void cpu_idle(int); int cpu_idle_wakeup(int); extern void (*cpu_idle_hook)(sbintime_t); /* Hook to machdep CPU idler. */ void cpu_switch(struct thread *, struct thread *, struct mtx *); void cpu_throw(struct thread *, struct thread *) __dead2; void unsleep(struct thread *); void userret(struct thread *, struct trapframe *); void cpu_exit(struct thread *); void exit1(struct thread *, int) __dead2; struct syscall_args; int cpu_fetch_syscall_args(struct thread *td, struct syscall_args *sa); void cpu_fork(struct thread *, struct proc *, struct thread *, int); void cpu_set_fork_handler(struct thread *, void (*)(void *), void *); void cpu_set_syscall_retval(struct thread *, int); void cpu_set_upcall(struct thread *td, struct thread *td0); void cpu_set_upcall_kse(struct thread *, void (*)(void *), void *, stack_t *); int cpu_set_user_tls(struct thread *, void *tls_base); void cpu_thread_alloc(struct thread *); void cpu_thread_clean(struct thread *); void cpu_thread_exit(struct thread *); void cpu_thread_free(struct thread *); void cpu_thread_swapin(struct thread *); void cpu_thread_swapout(struct thread *); struct thread *thread_alloc(int pages); int thread_alloc_stack(struct thread *, int pages); void thread_exit(void) __dead2; void thread_free(struct thread *td); void thread_link(struct thread *td, struct proc *p); void thread_reap(void); int thread_single(int how); void thread_single_end(void); void thread_stash(struct thread *td); void thread_stopped(struct proc *p); void childproc_stopped(struct proc *child, int reason); void childproc_continued(struct proc *child); void childproc_exited(struct proc *child); int thread_suspend_check(int how); +bool thread_suspend_check_needed(void); void thread_suspend_switch(struct thread *); void thread_suspend_one(struct thread *td); void thread_unlink(struct thread *td); void thread_unsuspend(struct proc *p); int thread_unsuspend_one(struct thread *td); void thread_wait(struct proc *p); struct thread *thread_find(struct proc *p, lwpid_t tid); static __inline int curthread_pflags_set(int flags) { struct thread *td; int save; td = curthread; save = ~flags | (td->td_pflags & flags); td->td_pflags |= flags; return (save); } static __inline void curthread_pflags_restore(int save) { curthread->td_pflags &= save; } #endif /* _KERNEL */ #endif /* !_SYS_PROC_H_ */ Index: projects/building-blocks/sys/x86/xen/xen_pci.c =================================================================== --- projects/building-blocks/sys/x86/xen/xen_pci.c (revision 275663) +++ projects/building-blocks/sys/x86/xen/xen_pci.c (nonexistent) @@ -1,132 +0,0 @@ -/* - * Copyright (c) 2014 Roger Pau Monné - * All rights reserved. - * - * Redistribution and use in source and binary forms, with or without - * modification, are permitted provided that the following conditions - * are met: - * 1. Redistributions of source code must retain the above copyright - * notice, this list of conditions and the following disclaimer. - * 2. Redistributions in binary form must reproduce the above copyright - * notice, this list of conditions and the following disclaimer in the - * documentation and/or other materials provided with the distribution. - * - * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS AS IS'' AND - * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE - * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE - * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE - * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL - * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS - * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) - * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT - * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY - * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF - * SUCH DAMAGE. - */ - -#include -__FBSDID("$FreeBSD$"); - -#include -#include -#include -#include -#include -#include - -#include -#include -#include -#include - -#include -#include - -#include "pcib_if.h" -#include "pci_if.h" - -static int xen_pci_probe(device_t dev); - -static void xen_pci_enable_msi_method(device_t dev, device_t child, - uint64_t address, uint16_t data); -static void xen_pci_disable_msi_method(device_t dev, device_t child); -static void xen_pci_child_added_method(device_t dev, device_t child); - -static device_method_t xen_pci_methods[] = { - /* Device interface */ - DEVMETHOD(device_probe, xen_pci_probe), - - /* PCI interface overwrites */ - DEVMETHOD(pci_enable_msi, xen_pci_enable_msi_method), - DEVMETHOD(pci_disable_msi, xen_pci_disable_msi_method), - DEVMETHOD(pci_child_added, xen_pci_child_added_method), - - DEVMETHOD_END -}; - -static devclass_t pci_devclass; - -DECLARE_CLASS(acpi_pci_driver); -DEFINE_CLASS_1(pci, xen_pci_driver, xen_pci_methods, sizeof(struct pci_softc), - acpi_pci_driver); -DRIVER_MODULE(xen_pci, pcib, xen_pci_driver, pci_devclass, 0, 0); -MODULE_DEPEND(xen_pci, pci, 1, 1, 1); -MODULE_DEPEND(xen_pci, acpi, 1, 1, 1); -MODULE_VERSION(xen_pci, 1); - -static int -xen_pci_probe(device_t dev) -{ - - device_set_desc(dev, "Xen PCI bus"); - - if (!xen_pv_domain()) - return (ENXIO); - - return (BUS_PROBE_SPECIFIC); -} - -static void -xen_pci_enable_msi_method(device_t dev, device_t child, uint64_t address, - uint16_t data) -{ - struct pci_devinfo *dinfo = device_get_ivars(child); - struct pcicfg_msi *msi = &dinfo->cfg.msi; - - /* Enable MSI in the control register. */ - msi->msi_ctrl |= PCIM_MSICTRL_MSI_ENABLE; - pci_write_config(child, msi->msi_location + PCIR_MSI_CTRL, - msi->msi_ctrl, 2); -} - -static void -xen_pci_disable_msi_method(device_t dev, device_t child) -{ - struct pci_devinfo *dinfo = device_get_ivars(child); - struct pcicfg_msi *msi = &dinfo->cfg.msi; - - msi->msi_ctrl &= ~PCIM_MSICTRL_MSI_ENABLE; - pci_write_config(child, msi->msi_location + PCIR_MSI_CTRL, - msi->msi_ctrl, 2); -} - -static void -xen_pci_child_added_method(device_t dev, device_t child) -{ - struct pci_devinfo *dinfo; - struct physdev_pci_device_add add_pci; - int error; - - dinfo = device_get_ivars(child); - KASSERT((dinfo != NULL), - ("xen_pci_add_child_method called with NULL dinfo")); - - bzero(&add_pci, sizeof(add_pci)); - add_pci.seg = dinfo->cfg.domain; - add_pci.bus = dinfo->cfg.bus; - add_pci.devfn = (dinfo->cfg.slot << 3) | dinfo->cfg.func; - error = HYPERVISOR_physdev_op(PHYSDEVOP_pci_device_add, &add_pci); - if (error) - panic("unable to add device bus %u devfn %u error: %d\n", - add_pci.bus, add_pci.devfn, error); -} Property changes on: projects/building-blocks/sys/x86/xen/xen_pci.c ___________________________________________________________________ Deleted: svn:eol-style ## -1 +0,0 ## -native \ No newline at end of property Deleted: svn:keywords ## -1 +0,0 ## -FreeBSD=%H \ No newline at end of property Deleted: svn:mime-type ## -1 +0,0 ## -text/plain \ No newline at end of property Index: projects/building-blocks/sys/x86/xen/pvcpu_enum.c =================================================================== --- projects/building-blocks/sys/x86/xen/pvcpu_enum.c (revision 275663) +++ projects/building-blocks/sys/x86/xen/pvcpu_enum.c (revision 275664) @@ -1,227 +1,236 @@ /*- * Copyright (c) 2003 John Baldwin * Copyright (c) 2013 Roger Pau Monné * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static int xenpv_probe(void); static int xenpv_probe_cpus(void); static int xenpv_setup_local(void); static int xenpv_setup_io(void); static ACPI_TABLE_MADT *madt; static vm_paddr_t madt_physaddr; static vm_offset_t madt_length; static struct apic_enumerator xenpv_enumerator = { "Xen PV", xenpv_probe, xenpv_probe_cpus, xenpv_setup_local, xenpv_setup_io }; /*--------------------- Helper functions to parse MADT -----------------------*/ /* * Parse an interrupt source override for an ISA interrupt. */ static void madt_parse_interrupt_override(ACPI_MADT_INTERRUPT_OVERRIDE *intr) { enum intr_trigger trig; enum intr_polarity pol; if (acpi_quirks & ACPI_Q_MADT_IRQ0 && intr->SourceIrq == 0 && intr->GlobalIrq == 2) { if (bootverbose) printf("MADT: Skipping timer override\n"); return; } madt_parse_interrupt_values(intr, &trig, &pol); + /* Remap the IRQ if it is mapped to a different interrupt vector. */ + if (intr->SourceIrq != intr->GlobalIrq && intr->GlobalIrq > 15 && + intr->SourceIrq == AcpiGbl_FADT.SciInterrupt) + /* + * If the SCI is remapped to a non-ISA global interrupt, + * then override the vector we use to setup. + */ + acpi_OverrideInterruptLevel(intr->GlobalIrq); + /* Register the IRQ with the polarity and trigger mode found. */ xen_register_pirq(intr->GlobalIrq, trig, pol); } /* * Call the handler routine for each entry in the MADT table. */ static void madt_walk_table(acpi_subtable_handler *handler, void *arg) { acpi_walk_subtables(madt + 1, (char *)madt + madt->Header.Length, handler, arg); } /* * Parse interrupt entries. */ static void madt_parse_ints(ACPI_SUBTABLE_HEADER *entry, void *arg __unused) { if (entry->Type == ACPI_MADT_TYPE_INTERRUPT_OVERRIDE) madt_parse_interrupt_override( (ACPI_MADT_INTERRUPT_OVERRIDE *)entry); } /*---------------------------- Xen PV enumerator -----------------------------*/ /* * This enumerator will only be registered on PVH */ static int xenpv_probe(void) { return (0); } /* * Test each possible vCPU in order to find the number of vCPUs */ static int xenpv_probe_cpus(void) { #ifdef SMP int i, ret; for (i = 0; i < MAXCPU; i++) { ret = HYPERVISOR_vcpu_op(VCPUOP_is_up, i, NULL); if (ret >= 0) lapic_create((i * 2), (i == 0)); } #endif return (0); } /* * Initialize the vCPU id of the BSP */ static int xenpv_setup_local(void) { PCPU_SET(vcpu_id, 0); lapic_init(0); return (0); } /* * On PVH guests there's no IO APIC */ static int xenpv_setup_io(void) { if (xen_initial_domain()) { int i; /* Map MADT */ madt_physaddr = acpi_find_table(ACPI_SIG_MADT); madt = acpi_map_table(madt_physaddr, ACPI_SIG_MADT); madt_length = madt->Header.Length; /* Try to initialize ACPI so that we can access the FADT. */ i = acpi_Startup(); if (ACPI_FAILURE(i)) { printf("MADT: ACPI Startup failed with %s\n", AcpiFormatException(i)); printf("Try disabling either ACPI or apic support.\n"); panic("Using MADT but ACPI doesn't work"); } /* Run through the table to see if there are any overrides. */ madt_walk_table(madt_parse_ints, NULL); /* * If there was not an explicit override entry for the SCI, * force it to use level trigger and active-low polarity. */ if (!madt_found_sci_override) { printf( "MADT: Forcing active-low polarity and level trigger for SCI\n"); xen_register_pirq(AcpiGbl_FADT.SciInterrupt, INTR_TRIGGER_LEVEL, INTR_POLARITY_LOW); } acpi_SetDefaultIntrModel(ACPI_INTR_APIC); } return (0); } static void xenpv_register(void *dummy __unused) { if (xen_pv_domain()) { apic_register_enumerator(&xenpv_enumerator); } } SYSINIT(xenpv_register, SI_SUB_TUNABLES - 1, SI_ORDER_FIRST, xenpv_register, NULL); /* * Setup per-CPU vCPU IDs */ static void xenpv_set_ids(void *dummy) { struct pcpu *pc; int i; CPU_FOREACH(i) { pc = pcpu_find(i); pc->pc_vcpu_id = i; } } SYSINIT(xenpv_set_ids, SI_SUB_CPU, SI_ORDER_MIDDLE, xenpv_set_ids, NULL); Index: projects/building-blocks/sys/x86/xen/xen_pci_bus.c =================================================================== --- projects/building-blocks/sys/x86/xen/xen_pci_bus.c (nonexistent) +++ projects/building-blocks/sys/x86/xen/xen_pci_bus.c (revision 275664) @@ -0,0 +1,90 @@ +/* + * Copyright (c) 2014 Roger Pau Monné + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include +__FBSDID("$FreeBSD$"); + +#include +#include +#include +#include + +#include +#include +#include +#include + +#include +#include +#include + +#include "pcib_if.h" +#include "pci_if.h" + +void +xen_pci_enable_msi_method(device_t dev, device_t child, uint64_t address, + uint16_t data) +{ + struct pci_devinfo *dinfo = device_get_ivars(child); + struct pcicfg_msi *msi = &dinfo->cfg.msi; + + /* Enable MSI in the control register. */ + msi->msi_ctrl |= PCIM_MSICTRL_MSI_ENABLE; + pci_write_config(child, msi->msi_location + PCIR_MSI_CTRL, + msi->msi_ctrl, 2); +} + +void +xen_pci_disable_msi_method(device_t dev, device_t child) +{ + struct pci_devinfo *dinfo = device_get_ivars(child); + struct pcicfg_msi *msi = &dinfo->cfg.msi; + + msi->msi_ctrl &= ~PCIM_MSICTRL_MSI_ENABLE; + pci_write_config(child, msi->msi_location + PCIR_MSI_CTRL, + msi->msi_ctrl, 2); +} + +void +xen_pci_child_added_method(device_t dev, device_t child) +{ + struct pci_devinfo *dinfo; + struct physdev_pci_device_add add_pci; + int error; + + dinfo = device_get_ivars(child); + KASSERT((dinfo != NULL), + ("xen_pci_add_child_method called with NULL dinfo")); + + bzero(&add_pci, sizeof(add_pci)); + add_pci.seg = dinfo->cfg.domain; + add_pci.bus = dinfo->cfg.bus; + add_pci.devfn = (dinfo->cfg.slot << 3) | dinfo->cfg.func; + error = HYPERVISOR_physdev_op(PHYSDEVOP_pci_device_add, &add_pci); + if (error) + panic("unable to add device bus %u devfn %u error: %d\n", + add_pci.bus, add_pci.devfn, error); +} Property changes on: projects/building-blocks/sys/x86/xen/xen_pci_bus.c ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/sys/xen/xen_pci.h =================================================================== --- projects/building-blocks/sys/xen/xen_pci.h (nonexistent) +++ projects/building-blocks/sys/xen/xen_pci.h (revision 275664) @@ -0,0 +1,37 @@ +/* + * Copyright (c) 2014 Roger Pau Monné + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * $FreeBSD$ + */ + +#ifndef __XEN_PCI_H__ +#define __XEN_PCI_H__ + +void xen_pci_enable_msi_method(device_t dev, device_t child, uint64_t address, + uint16_t data); +void xen_pci_disable_msi_method(device_t dev, device_t child); +void xen_pci_child_added_method(device_t dev, device_t child); + +#endif /* !__XEN_PCI_H__ */ Property changes on: projects/building-blocks/sys/xen/xen_pci.h ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/sys =================================================================== --- projects/building-blocks/sys (revision 275663) +++ projects/building-blocks/sys (revision 275664) Property changes on: projects/building-blocks/sys ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head/sys:r275590-275663 Index: projects/building-blocks/tools/bsdbox/Makefile =================================================================== --- projects/building-blocks/tools/bsdbox/Makefile (revision 275663) +++ projects/building-blocks/tools/bsdbox/Makefile (revision 275664) @@ -1,112 +1,112 @@ #$FreeBSD$ # @(#)Makefile 8.1 (Berkeley) 6/2/93 MAN= .include MK_SSP= no PROG= bsdbox BINDIR?=/sbin ################################################################# # # General notes: # # A number of Make variables are used to generate the crunchgen config file. # # CRUNCH_SRCDIRS: lists directories to search for included programs # CRUNCH_PROGS: lists programs to be included # CRUNCH_LIBS: libraries to statically link with # CRUNCH_SHLIBS: libraries to dynamically link with # CRUNCH_BUILDOPTS: generic build options to be added to every program # CRUNCH_BUILDTOOLS: lists programs that need build tools built in the # local architecture. # # Special options can be specified for individual programs # CRUNCH_SRCDIR_$(P): base source directory for program $(P) # CRUNCH_BUILDOPTS_$(P): additional build options for $(P) # CRUNCH_ALIAS_$(P): additional names to be used for $(P) # # By default, any name appearing in CRUNCH_PROGS or CRUNCH_ALIAS_${P} # will be used to generate a hard link to the resulting binary. # Specific links can be suppressed by setting # CRUNCH_SUPPRESS_LINK_$(NAME) to 1. # # Define Makefile variable RESCUE CRUNCH_BUILDOPTS+= -DRESCUE # Don't do symlinks as part of the install CRUNCH_GENERATE_LINKS= no # Which sources have local-arch build tools? # Define as blank; othrs need to override CRUNCH_BUILDTOOLS= ################################################################### # Programs from stock /bin # # WARNING: Changing this list may require adjusting # /usr/include/paths.h as well! You were warned! # CRUNCH_SRCDIRS+= bin # These are required to be shared so login and su can run as # setuid binaries - they use these libraries. PAM needs to be # built dynamically or it tries to build _all_ of the modules # statically - and that ends very badly. -CRUNCH_SHLIBS+= -lc -lutil -lcrypt +CRUNCH_SHLIBS+= -lc -lutil -lcrypt -lxo -lgpio CRUNCH_LIBS+= -lkvm -lmemstat -lnetgraph CRUNCH_LIBS+= -lcrypt -ledit -ll -ltermcapw ################################################################### # Programs from standard /sbin # # WARNING: Changing this list may require adjusting # /usr/include/paths.h as well! You were warned! # # Note that mdmfs have their own private 'pathnames.h' # headers in addition to the standard 'paths.h' header. # CRUNCH_SRCDIRS+= sbin CRUNCH_LIBS+= -lalias -lcam -lcursesw -ldevstat -lipsec # Don't forget this - ifconfig, etc -adrian CRUNCH_LIBS+= -lgeom -lbsdxml -ljail -lkiconv -lmd -lsbuf -lufs ################################################################## # Programs from stock /usr/bin # CRUNCH_SRCDIRS+= usr.bin # grep CRUNCH_LIBS+= -lbz2 ################################################################## # Programs from stock /usr/sbin # CRUNCH_SRCDIRS+= usr.sbin ################################################################## CRUNCH_SRCDIRS+= libexec CRUNCH_LIBS+= -lm .include "Makefile.base" .include "Makefile.net" .include "Makefile.hostapd" #.include "Makefile.textproc" .include "Makefile.login" .include "Makefile.kld" # telnet/telnetd are too broken to include as a crunchgen'ed binary, # thanks to some of the horrible layering violations going on. # .include "Makefile.telnetd" .include "Makefile.fs" CRUNCH_LIBS+= -lcrypto -lssl -lz # the crunchgen build environment .include # and since it creates a program.. .include Index: projects/building-blocks/usr.bin/patch/common.h =================================================================== --- projects/building-blocks/usr.bin/patch/common.h (revision 275663) +++ projects/building-blocks/usr.bin/patch/common.h (revision 275664) @@ -1,119 +1,120 @@ /*- * Copyright 1986, Larry Wall * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following condition is met: * 1. Redistributions of source code must retain the above copyright notice, * this condition and the following disclaimer. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ANY * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE * DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * patch - a program to apply diffs to original files * * -C option added in 1998, original code by Marc Espie, based on FreeBSD * behaviour * * $OpenBSD: common.h,v 1.26 2006/03/11 19:41:30 otto Exp $ * $FreeBSD$ */ #include #include #include #define DEBUGGING /* constants */ #define MAXHUNKSIZE 200000 /* is this enough lines? */ #define INITHUNKMAX 125 /* initial dynamic allocation size */ #define INITLINELEN 4096 #define BUFFERSIZE 4096 +#define LINENUM_MAX LONG_MAX #define SCCSPREFIX "s." #define GET "get -e %s" #define SCCSDIFF "get -p %s | diff - %s >/dev/null" #define RCSSUFFIX ",v" #define CHECKOUT "co -l %s" #define RCSDIFF "rcsdiff %s > /dev/null" #define ORIGEXT ".orig" #define REJEXT ".rej" /* handy definitions */ #define strNE(s1,s2) (strcmp(s1, s2)) #define strEQ(s1,s2) (!strcmp(s1, s2)) #define strnNE(s1,s2,l) (strncmp(s1, s2, l)) #define strnEQ(s1,s2,l) (!strncmp(s1, s2, l)) /* typedefs */ typedef long LINENUM; /* must be signed */ /* globals */ extern mode_t filemode; extern char *buf; /* general purpose buffer */ extern size_t buf_size; /* size of general purpose buffer */ extern bool using_plan_a; /* try to keep everything in memory */ extern bool out_of_mem; /* ran out of memory in plan a */ #define MAXFILEC 2 extern char *filearg[MAXFILEC]; extern bool ok_to_create_file; extern char *outname; extern char *origprae; extern char *TMPOUTNAME; extern char *TMPINNAME; extern char *TMPREJNAME; extern char *TMPPATNAME; extern bool toutkeep; extern bool trejkeep; #ifdef DEBUGGING extern int debug; #endif extern bool force; extern bool batch; extern bool verbose; extern bool reverse; extern bool noreverse; extern bool skip_rest_of_patch; extern int strippath; extern bool canonicalize; /* TRUE if -C was specified on command line. */ extern bool check_only; extern bool warn_on_invalid_line; extern bool last_line_missing_eol; #define CONTEXT_DIFF 1 #define NORMAL_DIFF 2 #define ED_DIFF 3 #define NEW_CONTEXT_DIFF 4 #define UNI_DIFF 5 extern int diff_type; extern char *revision; /* prerequisite revision, if any */ extern LINENUM input_lines; /* how long is input file in lines */ extern int posix; Index: projects/building-blocks/usr.bin/patch/pch.c =================================================================== --- projects/building-blocks/usr.bin/patch/pch.c (revision 275663) +++ projects/building-blocks/usr.bin/patch/pch.c (revision 275664) @@ -1,1584 +1,1621 @@ /*- * Copyright 1986, Larry Wall * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following condition is met: * 1. Redistributions of source code must retain the above copyright notice, * this condition and the following disclaimer. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND ANY * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE * DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE FOR * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. * * patch - a program to apply diffs to original files * * -C option added in 1998, original code by Marc Espie, based on FreeBSD * behaviour * * $OpenBSD: pch.c,v 1.43 2014/11/18 17:03:35 tobias Exp $ * $FreeBSD$ */ #include #include #include #include #include #include #include #include #include #include "common.h" #include "util.h" #include "pch.h" #include "pathnames.h" /* Patch (diff listing) abstract type. */ static off_t p_filesize; /* size of the patch file */ static LINENUM p_first; /* 1st line number */ static LINENUM p_newfirst; /* 1st line number of replacement */ static LINENUM p_ptrn_lines; /* # lines in pattern */ static LINENUM p_repl_lines; /* # lines in replacement text */ static LINENUM p_end = -1; /* last line in hunk */ static LINENUM p_max; /* max allowed value of p_end */ static LINENUM p_context = 3; /* # of context lines */ static LINENUM p_input_line = 0; /* current line # from patch file */ static char **p_line = NULL;/* the text of the hunk */ static unsigned short *p_len = NULL; /* length of each line */ static char *p_char = NULL; /* +, -, and ! */ static int hunkmax = INITHUNKMAX; /* size of above arrays to begin with */ static int p_indent; /* indent to patch */ static off_t p_base; /* where to intuit this time */ static LINENUM p_bline; /* line # of p_base */ static off_t p_start; /* where intuit found a patch */ static LINENUM p_sline; /* and the line number for it */ static LINENUM p_hunk_beg; /* line number of current hunk */ static LINENUM p_efake = -1; /* end of faked up lines--don't free */ static LINENUM p_bfake = -1; /* beg of faked up lines */ static FILE *pfp = NULL; /* patch file pointer */ static char *bestguess = NULL; /* guess at correct filename */ static void grow_hunkmax(void); static int intuit_diff_type(void); static void next_intuit_at(off_t, LINENUM); static void skip_to(off_t, LINENUM); static size_t pgets(bool _do_indent); static char *best_name(const struct file_name *, bool); static char *posix_name(const struct file_name *, bool); static size_t num_components(const char *); +static LINENUM strtolinenum(char *, char **); /* * Prepare to look for the next patch in the patch file. */ void re_patch(void) { p_first = 0; p_newfirst = 0; p_ptrn_lines = 0; p_repl_lines = 0; p_end = (LINENUM) - 1; p_max = 0; p_indent = 0; } /* * Open the patch file at the beginning of time. */ void open_patch_file(const char *filename) { struct stat filestat; int nr, nw; if (filename == NULL || *filename == '\0' || strEQ(filename, "-")) { pfp = fopen(TMPPATNAME, "w"); if (pfp == NULL) pfatal("can't create %s", TMPPATNAME); while ((nr = fread(buf, 1, buf_size, stdin)) > 0) { nw = fwrite(buf, 1, nr, pfp); if (nr != nw) pfatal("write error to %s", TMPPATNAME); } if (ferror(pfp) || fclose(pfp)) pfatal("can't write %s", TMPPATNAME); filename = TMPPATNAME; } pfp = fopen(filename, "r"); if (pfp == NULL) pfatal("patch file %s not found", filename); if (fstat(fileno(pfp), &filestat)) pfatal("can't stat %s", filename); p_filesize = filestat.st_size; next_intuit_at(0, 1L); /* start at the beginning */ set_hunkmax(); } /* * Make sure our dynamically realloced tables are malloced to begin with. */ void set_hunkmax(void) { if (p_line == NULL) p_line = malloc(hunkmax * sizeof(char *)); if (p_len == NULL) p_len = malloc(hunkmax * sizeof(unsigned short)); if (p_char == NULL) p_char = malloc(hunkmax * sizeof(char)); } /* * Enlarge the arrays containing the current hunk of patch. */ static void grow_hunkmax(void) { int new_hunkmax = hunkmax * 2; if (p_line == NULL || p_len == NULL || p_char == NULL) fatal("Internal memory allocation error\n"); p_line = reallocf(p_line, new_hunkmax * sizeof(char *)); p_len = reallocf(p_len, new_hunkmax * sizeof(unsigned short)); p_char = reallocf(p_char, new_hunkmax * sizeof(char)); if (p_line != NULL && p_len != NULL && p_char != NULL) { hunkmax = new_hunkmax; return; } if (!using_plan_a) fatal("out of memory\n"); out_of_mem = true; /* whatever is null will be allocated again */ /* from within plan_a(), of all places */ } /* True if the remainder of the patch file contains a diff of some sort. */ bool there_is_another_patch(void) { bool exists = false; if (p_base != 0 && p_base >= p_filesize) { if (verbose) say("done\n"); return false; } if (verbose) say("Hmm..."); diff_type = intuit_diff_type(); if (!diff_type) { if (p_base != 0) { if (verbose) say(" Ignoring the trailing garbage.\ndone\n"); } else say(" I can't seem to find a patch in there anywhere.\n"); return false; } if (verbose) say(" %sooks like %s to me...\n", (p_base == 0 ? "L" : "The next patch l"), diff_type == UNI_DIFF ? "a unified diff" : diff_type == CONTEXT_DIFF ? "a context diff" : diff_type == NEW_CONTEXT_DIFF ? "a new-style context diff" : diff_type == NORMAL_DIFF ? "a normal diff" : "an ed script"); if (p_indent && verbose) say("(Patch is indented %d space%s.)\n", p_indent, p_indent == 1 ? "" : "s"); skip_to(p_start, p_sline); while (filearg[0] == NULL) { if (force || batch) { say("No file to patch. Skipping...\n"); filearg[0] = savestr(bestguess); skip_rest_of_patch = true; return true; } ask("File to patch: "); if (*buf != '\n') { free(bestguess); bestguess = savestr(buf); filearg[0] = fetchname(buf, &exists, 0); } if (!exists) { ask("No file found--skip this patch? [n] "); if (*buf != 'y') continue; if (verbose) say("Skipping patch...\n"); free(filearg[0]); filearg[0] = fetchname(bestguess, &exists, 0); skip_rest_of_patch = true; return true; } } return true; } static void p4_fetchname(struct file_name *name, char *str) { char *t, *h; /* Skip leading whitespace. */ while (isspace((unsigned char)*str)) str++; /* Remove the file revision number. */ for (t = str, h = NULL; *t != '\0' && !isspace((unsigned char)*t); t++) if (*t == '#') h = t; if (h != NULL) *h = '\0'; name->path = fetchname(str, &name->exists, strippath); } /* Determine what kind of diff is in the remaining part of the patch file. */ static int intuit_diff_type(void) { off_t this_line = 0, previous_line; off_t first_command_line = -1; LINENUM fcl_line = -1; bool last_line_was_command = false, this_is_a_command = false; bool stars_last_line = false, stars_this_line = false; char *s, *t; int indent, retval; struct file_name names[MAX_FILE]; memset(names, 0, sizeof(names)); ok_to_create_file = false; fseeko(pfp, p_base, SEEK_SET); p_input_line = p_bline - 1; for (;;) { previous_line = this_line; last_line_was_command = this_is_a_command; stars_last_line = stars_this_line; this_line = ftello(pfp); indent = 0; p_input_line++; if (pgets(false) == 0) { if (first_command_line >= 0) { /* nothing but deletes!? */ p_start = first_command_line; p_sline = fcl_line; retval = ED_DIFF; goto scan_exit; } else { p_start = this_line; p_sline = p_input_line; retval = 0; goto scan_exit; } } for (s = buf; *s == ' ' || *s == '\t' || *s == 'X'; s++) { if (*s == '\t') indent += 8 - (indent % 8); else indent++; } for (t = s; isdigit((unsigned char)*t) || *t == ','; t++) ; this_is_a_command = (isdigit((unsigned char)*s) && (*t == 'd' || *t == 'c' || *t == 'a')); if (first_command_line < 0 && this_is_a_command) { first_command_line = this_line; fcl_line = p_input_line; p_indent = indent; /* assume this for now */ } if (!stars_last_line && strnEQ(s, "*** ", 4)) names[OLD_FILE].path = fetchname(s + 4, &names[OLD_FILE].exists, strippath); else if (strnEQ(s, "--- ", 4)) names[NEW_FILE].path = fetchname(s + 4, &names[NEW_FILE].exists, strippath); else if (strnEQ(s, "+++ ", 4)) /* pretend it is the old name */ names[OLD_FILE].path = fetchname(s + 4, &names[OLD_FILE].exists, strippath); else if (strnEQ(s, "Index:", 6)) names[INDEX_FILE].path = fetchname(s + 6, &names[INDEX_FILE].exists, strippath); else if (strnEQ(s, "Prereq:", 7)) { for (t = s + 7; isspace((unsigned char)*t); t++) ; revision = savestr(t); for (t = revision; *t && !isspace((unsigned char)*t); t++) ; *t = '\0'; if (*revision == '\0') { free(revision); revision = NULL; } } else if (strnEQ(s, "==== ", 5)) { /* Perforce-style diffs. */ if ((t = strstr(s + 5, " - ")) != NULL) p4_fetchname(&names[NEW_FILE], t + 3); p4_fetchname(&names[OLD_FILE], s + 5); } if ((!diff_type || diff_type == ED_DIFF) && first_command_line >= 0 && strEQ(s, ".\n")) { p_indent = indent; p_start = first_command_line; p_sline = fcl_line; retval = ED_DIFF; goto scan_exit; } if ((!diff_type || diff_type == UNI_DIFF) && strnEQ(s, "@@ -", 4)) { if (strnEQ(s + 4, "0,0", 3)) ok_to_create_file = true; p_indent = indent; p_start = this_line; p_sline = p_input_line; retval = UNI_DIFF; goto scan_exit; } stars_this_line = strnEQ(s, "********", 8); if ((!diff_type || diff_type == CONTEXT_DIFF) && stars_last_line && strnEQ(s, "*** ", 4)) { - if (atol(s + 4) == 0) + if (strtolinenum(s + 4, &s) == 0) ok_to_create_file = true; /* * If this is a new context diff the character just * at the end of the line is a '*'. */ while (*s && *s != '\n') s++; p_indent = indent; p_start = previous_line; p_sline = p_input_line - 1; retval = (*(s - 1) == '*' ? NEW_CONTEXT_DIFF : CONTEXT_DIFF); goto scan_exit; } if ((!diff_type || diff_type == NORMAL_DIFF) && last_line_was_command && (strnEQ(s, "< ", 2) || strnEQ(s, "> ", 2))) { p_start = previous_line; p_sline = p_input_line - 1; p_indent = indent; retval = NORMAL_DIFF; goto scan_exit; } } scan_exit: if (retval == UNI_DIFF) { /* unswap old and new */ struct file_name tmp = names[OLD_FILE]; names[OLD_FILE] = names[NEW_FILE]; names[NEW_FILE] = tmp; } if (filearg[0] == NULL) { if (posix) filearg[0] = posix_name(names, ok_to_create_file); else { /* Ignore the Index: name for context diffs, like GNU */ if (names[OLD_FILE].path != NULL || names[NEW_FILE].path != NULL) { free(names[INDEX_FILE].path); names[INDEX_FILE].path = NULL; } filearg[0] = best_name(names, ok_to_create_file); } } free(bestguess); bestguess = NULL; if (filearg[0] != NULL) bestguess = savestr(filearg[0]); else if (!ok_to_create_file) { /* * We don't want to create a new file but we need a * filename to set bestguess. Avoid setting filearg[0] * so the file is not created automatically. */ if (posix) bestguess = posix_name(names, true); else bestguess = best_name(names, true); } free(names[OLD_FILE].path); free(names[NEW_FILE].path); free(names[INDEX_FILE].path); return retval; } /* * Remember where this patch ends so we know where to start up again. */ static void next_intuit_at(off_t file_pos, LINENUM file_line) { p_base = file_pos; p_bline = file_line; } /* * Basically a verbose fseeko() to the actual diff listing. */ static void skip_to(off_t file_pos, LINENUM file_line) { size_t len; if (p_base > file_pos) fatal("Internal error: seek %lld>%lld\n", (long long)p_base, (long long)file_pos); if (verbose && p_base < file_pos) { fseeko(pfp, p_base, SEEK_SET); say("The text leading up to this was:\n--------------------------\n"); while (ftello(pfp) < file_pos) { len = pgets(false); if (len == 0) fatal("Unexpected end of file\n"); say("|%s", buf); } say("--------------------------\n"); } else fseeko(pfp, file_pos, SEEK_SET); p_input_line = file_line - 1; } /* Make this a function for better debugging. */ static void malformed(void) { fatal("malformed patch at line %ld: %s", p_input_line, buf); /* about as informative as "Syntax error" in C */ } /* * True if the line has been discarded (i.e. it is a line saying * "\ No newline at end of file".) */ static bool remove_special_line(void) { int c; c = fgetc(pfp); if (c == '\\') { do { c = fgetc(pfp); } while (c != EOF && c != '\n'); return true; } if (c != EOF) fseeko(pfp, -1, SEEK_CUR); return false; } /* * True if there is more of the current diff listing to process. */ bool another_hunk(void) { off_t line_beginning; /* file pos of the current line */ LINENUM repl_beginning; /* index of --- line */ LINENUM fillcnt; /* #lines of missing ptrn or repl */ LINENUM fillsrc; /* index of first line to copy */ LINENUM filldst; /* index of first missing line */ bool ptrn_spaces_eaten; /* ptrn was slightly misformed */ bool repl_could_be_missing; /* no + or ! lines in this hunk */ bool repl_missing; /* we are now backtracking */ off_t repl_backtrack_position; /* file pos of first repl line */ LINENUM repl_patch_line; /* input line number for same */ LINENUM ptrn_copiable; /* # of copiable lines in ptrn */ char *s; size_t len; int context = 0; while (p_end >= 0) { if (p_end == p_efake) p_end = p_bfake; /* don't free twice */ else free(p_line[p_end]); p_end--; } p_efake = -1; p_max = hunkmax; /* gets reduced when --- found */ if (diff_type == CONTEXT_DIFF || diff_type == NEW_CONTEXT_DIFF) { line_beginning = ftello(pfp); repl_beginning = 0; fillcnt = 0; fillsrc = 0; filldst = 0; ptrn_spaces_eaten = false; repl_could_be_missing = true; repl_missing = false; repl_backtrack_position = 0; repl_patch_line = 0; ptrn_copiable = 0; len = pgets(true); p_input_line++; if (len == 0 || strnNE(buf, "********", 8)) { next_intuit_at(line_beginning, p_input_line); return false; } p_context = 100; p_hunk_beg = p_input_line + 1; while (p_end < p_max) { line_beginning = ftello(pfp); len = pgets(true); p_input_line++; if (len == 0) { if (p_max - p_end < 4) { /* assume blank lines got chopped */ strlcpy(buf, " \n", buf_size); } else { if (repl_beginning && repl_could_be_missing) { repl_missing = true; goto hunk_done; } fatal("unexpected end of file in patch\n"); } } p_end++; if (p_end >= hunkmax) fatal("Internal error: hunk larger than hunk " "buffer size"); p_char[p_end] = *buf; p_line[p_end] = NULL; switch (*buf) { case '*': if (strnEQ(buf, "********", 8)) { if (repl_beginning && repl_could_be_missing) { repl_missing = true; goto hunk_done; } else fatal("unexpected end of hunk " "at line %ld\n", p_input_line); } if (p_end != 0) { if (repl_beginning && repl_could_be_missing) { repl_missing = true; goto hunk_done; } fatal("unexpected *** at line %ld: %s", p_input_line, buf); } context = 0; p_line[p_end] = savestr(buf); if (out_of_mem) { p_end--; return false; } for (s = buf; *s && !isdigit((unsigned char)*s); s++) ; if (!*s) malformed(); if (strnEQ(s, "0,0", 3)) memmove(s, s + 2, strlen(s + 2) + 1); - p_first = (LINENUM) atol(s); - while (isdigit((unsigned char)*s)) - s++; + p_first = strtolinenum(s, &s); if (*s == ',') { for (; *s && !isdigit((unsigned char)*s); s++) ; if (!*s) malformed(); - p_ptrn_lines = ((LINENUM) atol(s)) - p_first + 1; + p_ptrn_lines = strtolinenum(s, &s) - p_first + 1; + if (p_ptrn_lines < 0) + malformed(); } else if (p_first) p_ptrn_lines = 1; else { p_ptrn_lines = 0; p_first = 1; } + if (p_first >= LINENUM_MAX - p_ptrn_lines || + p_ptrn_lines >= LINENUM_MAX - 6) + malformed(); /* we need this much at least */ p_max = p_ptrn_lines + 6; while (p_max >= hunkmax) grow_hunkmax(); p_max = hunkmax; break; case '-': if (buf[1] == '-') { if (repl_beginning || (p_end != p_ptrn_lines + 1 + (p_char[p_end - 1] == '\n'))) { if (p_end == 1) { /* * `old' lines were omitted; * set up to fill them in * from 'new' context lines. */ p_end = p_ptrn_lines + 1; fillsrc = p_end + 1; filldst = 1; fillcnt = p_ptrn_lines; } else { if (repl_beginning) { if (repl_could_be_missing) { repl_missing = true; goto hunk_done; } fatal("duplicate \"---\" at line %ld--check line numbers at line %ld\n", p_input_line, p_hunk_beg + repl_beginning); } else { fatal("%s \"---\" at line %ld--check line numbers at line %ld\n", (p_end <= p_ptrn_lines ? "Premature" : "Overdue"), p_input_line, p_hunk_beg); } } } repl_beginning = p_end; repl_backtrack_position = ftello(pfp); repl_patch_line = p_input_line; p_line[p_end] = savestr(buf); if (out_of_mem) { p_end--; return false; } p_char[p_end] = '='; for (s = buf; *s && !isdigit((unsigned char)*s); s++) ; if (!*s) malformed(); - p_newfirst = (LINENUM) atol(s); - while (isdigit((unsigned char)*s)) - s++; + p_newfirst = strtolinenum(s, &s); if (*s == ',') { for (; *s && !isdigit((unsigned char)*s); s++) ; if (!*s) malformed(); - p_repl_lines = ((LINENUM) atol(s)) - + p_repl_lines = strtolinenum(s, &s) - p_newfirst + 1; + if (p_repl_lines < 0) + malformed(); } else if (p_newfirst) p_repl_lines = 1; else { p_repl_lines = 0; p_newfirst = 1; } + if (p_newfirst >= LINENUM_MAX - p_repl_lines || + p_repl_lines >= LINENUM_MAX - p_end) + malformed(); p_max = p_repl_lines + p_end; if (p_max > MAXHUNKSIZE) fatal("hunk too large (%ld lines) at line %ld: %s", p_max, p_input_line, buf); while (p_max >= hunkmax) grow_hunkmax(); if (p_repl_lines != ptrn_copiable && (p_context != 0 || p_repl_lines != 1)) repl_could_be_missing = false; break; } goto change_line; case '+': case '!': repl_could_be_missing = false; change_line: if (buf[1] == '\n' && canonicalize) strlcpy(buf + 1, " \n", buf_size - 1); if (!isspace((unsigned char)buf[1]) && buf[1] != '>' && buf[1] != '<' && repl_beginning && repl_could_be_missing) { repl_missing = true; goto hunk_done; } if (context >= 0) { if (context < p_context) p_context = context; context = -1000; } p_line[p_end] = savestr(buf + 2); if (out_of_mem) { p_end--; return false; } if (p_end == p_ptrn_lines) { if (remove_special_line()) { int l; l = strlen(p_line[p_end]) - 1; (p_line[p_end])[l] = 0; } } break; case '\t': case '\n': /* assume the 2 spaces got eaten */ if (repl_beginning && repl_could_be_missing && (!ptrn_spaces_eaten || diff_type == NEW_CONTEXT_DIFF)) { repl_missing = true; goto hunk_done; } p_line[p_end] = savestr(buf); if (out_of_mem) { p_end--; return false; } if (p_end != p_ptrn_lines + 1) { ptrn_spaces_eaten |= (repl_beginning != 0); context++; if (!repl_beginning) ptrn_copiable++; p_char[p_end] = ' '; } break; case ' ': if (!isspace((unsigned char)buf[1]) && repl_beginning && repl_could_be_missing) { repl_missing = true; goto hunk_done; } context++; if (!repl_beginning) ptrn_copiable++; p_line[p_end] = savestr(buf + 2); if (out_of_mem) { p_end--; return false; } break; default: if (repl_beginning && repl_could_be_missing) { repl_missing = true; goto hunk_done; } malformed(); } /* set up p_len for strncmp() so we don't have to */ /* assume null termination */ if (p_line[p_end]) p_len[p_end] = strlen(p_line[p_end]); else p_len[p_end] = 0; } hunk_done: if (p_end >= 0 && !repl_beginning) fatal("no --- found in patch at line %ld\n", pch_hunk_beg()); if (repl_missing) { /* reset state back to just after --- */ p_input_line = repl_patch_line; for (p_end--; p_end > repl_beginning; p_end--) free(p_line[p_end]); fseeko(pfp, repl_backtrack_position, SEEK_SET); /* redundant 'new' context lines were omitted - set */ /* up to fill them in from the old file context */ if (!p_context && p_repl_lines == 1) { p_repl_lines = 0; p_max--; } fillsrc = 1; filldst = repl_beginning + 1; fillcnt = p_repl_lines; p_end = p_max; } else if (!p_context && fillcnt == 1) { /* the first hunk was a null hunk with no context */ /* and we were expecting one line -- fix it up. */ while (filldst < p_end) { p_line[filldst] = p_line[filldst + 1]; p_char[filldst] = p_char[filldst + 1]; p_len[filldst] = p_len[filldst + 1]; filldst++; } #if 0 repl_beginning--; /* this doesn't need to be fixed */ #endif p_end--; p_first++; /* do append rather than insert */ fillcnt = 0; p_ptrn_lines = 0; } if (diff_type == CONTEXT_DIFF && (fillcnt || (p_first > 1 && ptrn_copiable > 2 * p_context))) { if (verbose) say("%s\n%s\n%s\n", "(Fascinating--this is really a new-style context diff but without", "the telltale extra asterisks on the *** line that usually indicate", "the new style...)"); diff_type = NEW_CONTEXT_DIFF; } /* if there were omitted context lines, fill them in now */ if (fillcnt) { p_bfake = filldst; /* remember where not to free() */ p_efake = filldst + fillcnt - 1; while (fillcnt-- > 0) { while (fillsrc <= p_end && p_char[fillsrc] != ' ') fillsrc++; if (fillsrc > p_end) fatal("replacement text or line numbers mangled in hunk at line %ld\n", p_hunk_beg); p_line[filldst] = p_line[fillsrc]; p_char[filldst] = p_char[fillsrc]; p_len[filldst] = p_len[fillsrc]; fillsrc++; filldst++; } while (fillsrc <= p_end && fillsrc != repl_beginning && p_char[fillsrc] != ' ') fillsrc++; #ifdef DEBUGGING if (debug & 64) printf("fillsrc %ld, filldst %ld, rb %ld, e+1 %ld\n", fillsrc, filldst, repl_beginning, p_end + 1); #endif if (fillsrc != p_end + 1 && fillsrc != repl_beginning) malformed(); if (filldst != p_end + 1 && filldst != repl_beginning) malformed(); } if (p_line[p_end] != NULL) { if (remove_special_line()) { p_len[p_end] -= 1; (p_line[p_end])[p_len[p_end]] = 0; } } } else if (diff_type == UNI_DIFF) { LINENUM fillold; /* index of old lines */ LINENUM fillnew; /* index of new lines */ char ch; line_beginning = ftello(pfp); /* file pos of the current line */ len = pgets(true); p_input_line++; if (len == 0 || strnNE(buf, "@@ -", 4)) { next_intuit_at(line_beginning, p_input_line); return false; } s = buf + 4; if (!*s) malformed(); - p_first = (LINENUM) atol(s); - while (isdigit((unsigned char)*s)) - s++; + p_first = strtolinenum(s, &s); if (*s == ',') { - p_ptrn_lines = (LINENUM) atol(++s); - while (isdigit((unsigned char)*s)) - s++; + p_ptrn_lines = strtolinenum(s + 1, &s); } else p_ptrn_lines = 1; if (*s == ' ') s++; if (*s != '+' || !*++s) malformed(); - p_newfirst = (LINENUM) atol(s); - while (isdigit((unsigned char)*s)) - s++; + p_newfirst = strtolinenum(s, &s); if (*s == ',') { - p_repl_lines = (LINENUM) atol(++s); - while (isdigit((unsigned char)*s)) - s++; + p_repl_lines = strtolinenum(s + 1, &s); } else p_repl_lines = 1; if (*s == ' ') s++; if (*s != '@') malformed(); + if (p_first >= LINENUM_MAX - p_ptrn_lines || + p_newfirst > LINENUM_MAX - p_repl_lines || + p_ptrn_lines >= LINENUM_MAX - p_repl_lines - 1) + malformed(); if (!p_ptrn_lines) p_first++; /* do append rather than insert */ p_max = p_ptrn_lines + p_repl_lines + 1; while (p_max >= hunkmax) grow_hunkmax(); fillold = 1; fillnew = fillold + p_ptrn_lines; p_end = fillnew + p_repl_lines; snprintf(buf, buf_size, "*** %ld,%ld ****\n", p_first, p_first + p_ptrn_lines - 1); p_line[0] = savestr(buf); if (out_of_mem) { p_end = -1; return false; } p_char[0] = '*'; snprintf(buf, buf_size, "--- %ld,%ld ----\n", p_newfirst, p_newfirst + p_repl_lines - 1); p_line[fillnew] = savestr(buf); if (out_of_mem) { p_end = 0; return false; } p_char[fillnew++] = '='; p_context = 100; context = 0; p_hunk_beg = p_input_line + 1; while (fillold <= p_ptrn_lines || fillnew <= p_end) { line_beginning = ftello(pfp); len = pgets(true); p_input_line++; if (len == 0) { if (p_max - fillnew < 3) { /* assume blank lines got chopped */ strlcpy(buf, " \n", buf_size); } else { fatal("unexpected end of file in patch\n"); } } if (*buf == '\t' || *buf == '\n') { ch = ' '; /* assume the space got eaten */ s = savestr(buf); } else { ch = *buf; s = savestr(buf + 1); } if (out_of_mem) { while (--fillnew > p_ptrn_lines) free(p_line[fillnew]); p_end = fillold - 1; return false; } switch (ch) { case '-': if (fillold > p_ptrn_lines) { free(s); p_end = fillnew - 1; malformed(); } p_char[fillold] = ch; p_line[fillold] = s; p_len[fillold++] = strlen(s); if (fillold > p_ptrn_lines) { if (remove_special_line()) { p_len[fillold - 1] -= 1; s[p_len[fillold - 1]] = 0; } } break; case '=': ch = ' '; /* FALL THROUGH */ case ' ': if (fillold > p_ptrn_lines) { free(s); while (--fillnew > p_ptrn_lines) free(p_line[fillnew]); p_end = fillold - 1; malformed(); } context++; p_char[fillold] = ch; p_line[fillold] = s; p_len[fillold++] = strlen(s); s = savestr(s); if (out_of_mem) { while (--fillnew > p_ptrn_lines) free(p_line[fillnew]); p_end = fillold - 1; return false; } if (fillold > p_ptrn_lines) { if (remove_special_line()) { p_len[fillold - 1] -= 1; s[p_len[fillold - 1]] = 0; } } /* FALL THROUGH */ case '+': if (fillnew > p_end) { free(s); while (--fillnew > p_ptrn_lines) free(p_line[fillnew]); p_end = fillold - 1; malformed(); } p_char[fillnew] = ch; p_line[fillnew] = s; p_len[fillnew++] = strlen(s); if (fillold > p_ptrn_lines) { if (remove_special_line()) { p_len[fillnew - 1] -= 1; s[p_len[fillnew - 1]] = 0; } } break; default: p_end = fillnew; malformed(); } if (ch != ' ' && context > 0) { if (context < p_context) p_context = context; context = -1000; } } /* while */ } else { /* normal diff--fake it up */ char hunk_type; int i; LINENUM min, max; line_beginning = ftello(pfp); p_context = 0; len = pgets(true); p_input_line++; if (len == 0 || !isdigit((unsigned char)*buf)) { next_intuit_at(line_beginning, p_input_line); return false; } - p_first = (LINENUM) atol(buf); - for (s = buf; isdigit((unsigned char)*s); s++) - ; + p_first = strtolinenum(buf, &s); if (*s == ',') { - p_ptrn_lines = (LINENUM) atol(++s) - p_first + 1; - while (isdigit((unsigned char)*s)) - s++; + p_ptrn_lines = strtolinenum(s + 1, &s) - p_first + 1; + if (p_ptrn_lines < 0) + malformed(); } else p_ptrn_lines = (*s != 'a'); hunk_type = *s; if (hunk_type == 'a') p_first++; /* do append rather than insert */ - min = (LINENUM) atol(++s); - for (; isdigit((unsigned char)*s); s++) - ; + min = strtolinenum(s + 1, &s); if (*s == ',') - max = (LINENUM) atol(++s); + max = strtolinenum(s + 1, &s); else max = min; + if (min < 0 || min > max || max - min == LINENUM_MAX) + malformed(); if (hunk_type == 'd') min++; - p_end = p_ptrn_lines + 1 + max - min + 1; + p_newfirst = min; + p_repl_lines = max - min + 1; + if (p_newfirst > LINENUM_MAX - p_repl_lines || + p_ptrn_lines >= LINENUM_MAX - p_repl_lines - 1) + malformed(); + p_end = p_ptrn_lines + p_repl_lines + 1; if (p_end > MAXHUNKSIZE) fatal("hunk too large (%ld lines) at line %ld: %s", p_end, p_input_line, buf); while (p_end >= hunkmax) grow_hunkmax(); - p_newfirst = min; - p_repl_lines = max - min + 1; snprintf(buf, buf_size, "*** %ld,%ld\n", p_first, p_first + p_ptrn_lines - 1); p_line[0] = savestr(buf); if (out_of_mem) { p_end = -1; return false; } p_char[0] = '*'; for (i = 1; i <= p_ptrn_lines; i++) { len = pgets(true); p_input_line++; if (len == 0) fatal("unexpected end of file in patch at line %ld\n", p_input_line); if (*buf != '<') fatal("< expected at line %ld of patch\n", p_input_line); p_line[i] = savestr(buf + 2); if (out_of_mem) { p_end = i - 1; return false; } p_len[i] = strlen(p_line[i]); p_char[i] = '-'; } if (remove_special_line()) { p_len[i - 1] -= 1; (p_line[i - 1])[p_len[i - 1]] = 0; } if (hunk_type == 'c') { len = pgets(true); p_input_line++; if (len == 0) fatal("unexpected end of file in patch at line %ld\n", p_input_line); if (*buf != '-') fatal("--- expected at line %ld of patch\n", p_input_line); } snprintf(buf, buf_size, "--- %ld,%ld\n", min, max); p_line[i] = savestr(buf); if (out_of_mem) { p_end = i - 1; return false; } p_char[i] = '='; for (i++; i <= p_end; i++) { len = pgets(true); p_input_line++; if (len == 0) fatal("unexpected end of file in patch at line %ld\n", p_input_line); if (*buf != '>') fatal("> expected at line %ld of patch\n", p_input_line); p_line[i] = savestr(buf + 2); if (out_of_mem) { p_end = i - 1; return false; } p_len[i] = strlen(p_line[i]); p_char[i] = '+'; } if (remove_special_line()) { p_len[i - 1] -= 1; (p_line[i - 1])[p_len[i - 1]] = 0; } } if (reverse) /* backwards patch? */ if (!pch_swap()) say("Not enough memory to swap next hunk!\n"); #ifdef DEBUGGING if (debug & 2) { int i; char special; for (i = 0; i <= p_end; i++) { if (i == p_ptrn_lines) special = '^'; else special = ' '; fprintf(stderr, "%3d %c %c %s", i, p_char[i], special, p_line[i]); fflush(stderr); } } #endif if (p_end + 1 < hunkmax)/* paranoia reigns supreme... */ p_char[p_end + 1] = '^'; /* add a stopper for apply_hunk */ return true; } /* * Input a line from the patch file. * Worry about indentation if do_indent is true. * The line is read directly into the buf global variable which * is resized if necessary in order to hold the complete line. * Returns the number of characters read including the terminating * '\n', if any. */ size_t pgets(bool do_indent) { char *line; size_t len; int indent = 0, skipped = 0; line = fgetln(pfp, &len); if (line != NULL) { if (len + 1 > buf_size) { while (len + 1 > buf_size) buf_size *= 2; free(buf); buf = malloc(buf_size); if (buf == NULL) fatal("out of memory\n"); } if (do_indent == 1 && p_indent) { for (; indent < p_indent && (*line == ' ' || *line == '\t' || *line == 'X'); line++, skipped++) { if (*line == '\t') indent += 8 - (indent %7); else indent++; } } memcpy(buf, line, len - skipped); buf[len - skipped] = '\0'; } return len; } /* * Reverse the old and new portions of the current hunk. */ bool pch_swap(void) { char **tp_line; /* the text of the hunk */ unsigned short *tp_len;/* length of each line */ char *tp_char; /* +, -, and ! */ LINENUM i; LINENUM n; bool blankline = false; char *s; i = p_first; p_first = p_newfirst; p_newfirst = i; /* make a scratch copy */ tp_line = p_line; tp_len = p_len; tp_char = p_char; p_line = NULL; /* force set_hunkmax to allocate again */ p_len = NULL; p_char = NULL; set_hunkmax(); if (p_line == NULL || p_len == NULL || p_char == NULL) { free(p_line); p_line = tp_line; free(p_len); p_len = tp_len; free(p_char); p_char = tp_char; return false; /* not enough memory to swap hunk! */ } /* now turn the new into the old */ i = p_ptrn_lines + 1; if (tp_char[i] == '\n') { /* account for possible blank line */ blankline = true; i++; } if (p_efake >= 0) { /* fix non-freeable ptr range */ if (p_efake <= i) n = p_end - i + 1; else n = -i; p_efake += n; p_bfake += n; } for (n = 0; i <= p_end; i++, n++) { p_line[n] = tp_line[i]; p_char[n] = tp_char[i]; if (p_char[n] == '+') p_char[n] = '-'; p_len[n] = tp_len[i]; } if (blankline) { i = p_ptrn_lines + 1; p_line[n] = tp_line[i]; p_char[n] = tp_char[i]; p_len[n] = tp_len[i]; n++; } if (p_char[0] != '=') fatal("Malformed patch at line %ld: expected '=' found '%c'\n", p_input_line, p_char[0]); p_char[0] = '*'; for (s = p_line[0]; *s; s++) if (*s == '-') *s = '*'; /* now turn the old into the new */ if (p_char[0] != '*') fatal("Malformed patch at line %ld: expected '*' found '%c'\n", p_input_line, p_char[0]); tp_char[0] = '='; for (s = tp_line[0]; *s; s++) if (*s == '*') *s = '-'; for (i = 0; n <= p_end; i++, n++) { p_line[n] = tp_line[i]; p_char[n] = tp_char[i]; if (p_char[n] == '-') p_char[n] = '+'; p_len[n] = tp_len[i]; } if (i != p_ptrn_lines + 1) fatal("Malformed patch at line %ld: expected %ld lines, " "got %ld\n", p_input_line, p_ptrn_lines + 1, i); i = p_ptrn_lines; p_ptrn_lines = p_repl_lines; p_repl_lines = i; free(tp_line); free(tp_len); free(tp_char); return true; } /* * Return the specified line position in the old file of the old context. */ LINENUM pch_first(void) { return p_first; } /* * Return the number of lines of old context. */ LINENUM pch_ptrn_lines(void) { return p_ptrn_lines; } /* * Return the probable line position in the new file of the first line. */ LINENUM pch_newfirst(void) { return p_newfirst; } /* * Return the number of lines in the replacement text including context. */ LINENUM pch_repl_lines(void) { return p_repl_lines; } /* * Return the number of lines in the whole hunk. */ LINENUM pch_end(void) { return p_end; } /* * Return the number of context lines before the first changed line. */ LINENUM pch_context(void) { return p_context; } /* * Return the length of a particular patch line. */ unsigned short pch_line_len(LINENUM line) { return p_len[line]; } /* * Return the control character (+, -, *, !, etc) for a patch line. */ char pch_char(LINENUM line) { return p_char[line]; } /* * Return a pointer to a particular patch line. */ char * pfetch(LINENUM line) { return p_line[line]; } /* * Return where in the patch file this hunk began, for error messages. */ LINENUM pch_hunk_beg(void) { return p_hunk_beg; } /* * Apply an ed script by feeding ed itself. */ void do_ed_script(void) { char *t; off_t beginning_of_this_line; FILE *pipefp = NULL; if (!skip_rest_of_patch) { if (copy_file(filearg[0], TMPOUTNAME) < 0) { unlink(TMPOUTNAME); fatal("can't create temp file %s", TMPOUTNAME); } snprintf(buf, buf_size, "%s%s%s", _PATH_ED, verbose ? " " : " -s ", TMPOUTNAME); pipefp = popen(buf, "w"); } for (;;) { beginning_of_this_line = ftello(pfp); if (pgets(true) == 0) { next_intuit_at(beginning_of_this_line, p_input_line); break; } p_input_line++; for (t = buf; isdigit((unsigned char)*t) || *t == ','; t++) ; /* POSIX defines allowed commands as {a,c,d,i,s} */ if (isdigit((unsigned char)*buf) && (*t == 'a' || *t == 'c' || *t == 'd' || *t == 'i' || *t == 's')) { if (pipefp != NULL) fputs(buf, pipefp); if (*t != 'd') { while (pgets(true)) { p_input_line++; if (pipefp != NULL) fputs(buf, pipefp); if (strEQ(buf, ".\n")) break; } } } else { next_intuit_at(beginning_of_this_line, p_input_line); break; } } if (pipefp == NULL) return; fprintf(pipefp, "w\n"); fprintf(pipefp, "q\n"); fflush(pipefp); pclose(pipefp); ignore_signals(); if (!check_only) { if (move_file(TMPOUTNAME, outname) < 0) { toutkeep = true; chmod(TMPOUTNAME, filemode); } else chmod(outname, filemode); } set_signals(1); } /* * Choose the name of the file to be patched based on POSIX rules. * NOTE: the POSIX rules are amazingly stupid and we only follow them * if the user specified --posix or set POSIXLY_CORRECT. */ static char * posix_name(const struct file_name *names, bool assume_exists) { char *path = NULL; int i; /* * POSIX states that the filename will be chosen from one * of the old, new and index names (in that order) if * the file exists relative to CWD after -p stripping. */ for (i = 0; i < MAX_FILE; i++) { if (names[i].path != NULL && names[i].exists) { path = names[i].path; break; } } if (path == NULL && !assume_exists) { /* * No files found, look for something we can checkout from * RCS/SCCS dirs. Same order as above. */ for (i = 0; i < MAX_FILE; i++) { if (names[i].path != NULL && (path = checked_in(names[i].path)) != NULL) break; } /* * Still no match? Check to see if the diff could be creating * a new file. */ if (path == NULL && ok_to_create_file && names[NEW_FILE].path != NULL) path = names[NEW_FILE].path; } return path ? savestr(path) : NULL; } static char * compare_names(const struct file_name *names, bool assume_exists, int phase) { size_t min_components, min_baselen, min_len, tmp; char *best = NULL; char *path; int i; /* * The "best" name is the one with the fewest number of path * components, the shortest basename length, and the shortest * overall length (in that order). We only use the Index: file * if neither of the old or new files could be intuited from * the diff header. */ min_components = min_baselen = min_len = SIZE_MAX; for (i = INDEX_FILE; i >= OLD_FILE; i--) { path = names[i].path; if (path == NULL || (phase == 1 && !names[i].exists && !assume_exists) || (phase == 2 && checked_in(path) == NULL)) continue; if ((tmp = num_components(path)) > min_components) continue; if (tmp < min_components) { min_components = tmp; best = path; } if ((tmp = strlen(basename(path))) > min_baselen) continue; if (tmp < min_baselen) { min_baselen = tmp; best = path; } if ((tmp = strlen(path)) > min_len) continue; min_len = tmp; best = path; } return best; } /* * Choose the name of the file to be patched based the "best" one * available. */ static char * best_name(const struct file_name *names, bool assume_exists) { char *best; best = compare_names(names, assume_exists, 1); if (best == NULL) { best = compare_names(names, assume_exists, 2); /* * Still no match? Check to see if the diff could be creating * a new file. */ if (best == NULL && ok_to_create_file && names[NEW_FILE].path != NULL) best = names[NEW_FILE].path; } return best ? savestr(best) : NULL; } static size_t num_components(const char *path) { size_t n; const char *cp; for (n = 0, cp = path; (cp = strchr(cp, '/')) != NULL; n++, cp++) { while (*cp == '/') cp++; /* skip consecutive slashes */ } return n; +} + +/* + * Convert number at NPTR into LINENUM and save address of first + * character that is not a digit in ENDPTR. If conversion is not + * possible, call fatal. + */ +static LINENUM +strtolinenum(char *nptr, char **endptr) +{ + LINENUM rv; + char c; + char *p; + const char *errstr; + + for (p = nptr; isdigit((unsigned char)*p); p++) + ; + + if (p == nptr) + malformed(); + + c = *p; + *p = '\0'; + + rv = strtonum(nptr, 0, LINENUM_MAX, &errstr); + if (errstr != NULL) + fatal("invalid line number at line %ld: `%s' is %s\n", + p_input_line, nptr, errstr); + + *p = c; + *endptr = p; + + return rv; } Index: projects/building-blocks/usr.sbin/mtree/verify.c =================================================================== --- projects/building-blocks/usr.sbin/mtree/verify.c (revision 275663) +++ projects/building-blocks/usr.sbin/mtree/verify.c (revision 275664) @@ -1,259 +1,259 @@ /*- * Copyright (c) 1990, 1993 * The Regents of the University of California. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. Neither the name of the University nor the names of its contributors * may be used to endorse or promote products derived from this software * without specific prior written permission. * * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF * SUCH DAMAGE. */ #if 0 #ifndef lint static char sccsid[] = "@(#)verify.c 8.1 (Berkeley) 6/6/93"; #endif /* not lint */ #endif #include __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include "mtree.h" #include "extern.h" static NODE *root; static char path[MAXPATHLEN]; static void miss(NODE *, char *); static int vwalk(void); int mtree_verifyspec(FILE *fi) { int rval; root = mtree_readspec(fi); rval = vwalk(); miss(root, path); return (rval); } static int nsort(const FTSENT * const *a, const FTSENT * const *b) { return (strcmp((*a)->fts_name, (*b)->fts_name)); } static int vwalk(void) { FTS *t; FTSENT *p; NODE *ep, *level; int specdepth, rval; char *argv[2]; char dot[] = "."; argv[0] = dot; argv[1] = NULL; if ((t = fts_open(argv, ftsoptions, nsort)) == NULL) err(1, "line %d: fts_open", lineno); level = root; specdepth = rval = 0; while ((p = fts_read(t))) { if (check_excludes(p->fts_name, p->fts_path)) { fts_set(t, p, FTS_SKIP); continue; } switch(p->fts_info) { case FTS_D: case FTS_SL: break; case FTS_DP: if (specdepth > p->fts_level) { for (level = level->parent; level->prev; level = level->prev); --specdepth; } continue; case FTS_DNR: case FTS_ERR: case FTS_NS: warnx("%s: %s", RP(p), strerror(p->fts_errno)); continue; default: if (dflag) continue; } if (specdepth != p->fts_level) goto extra; for (ep = level; ep; ep = ep->next) if ((ep->flags & F_MAGIC && !fnmatch(ep->name, p->fts_name, FNM_PATHNAME)) || !strcmp(ep->name, p->fts_name)) { ep->flags |= F_VISIT; if ((ep->flags & F_NOCHANGE) == 0 && compare(ep->name, ep, p)) rval = MISMATCHEXIT; if (ep->flags & F_IGN) (void)fts_set(t, p, FTS_SKIP); else if (ep->child && ep->type == F_DIR && p->fts_info == FTS_D) { level = ep->child; ++specdepth; } break; } if (ep) continue; extra: if (!eflag) { (void)printf("%s extra", RP(p)); if (rflag) { if ((S_ISDIR(p->fts_statp->st_mode) ? rmdir : unlink)(p->fts_accpath)) { (void)printf(", not removed: %s", strerror(errno)); } else (void)printf(", removed"); } (void)putchar('\n'); } (void)fts_set(t, p, FTS_SKIP); } (void)fts_close(t); if (sflag) warnx("%s checksum: %lu", fullpath, (unsigned long)crc_total); return (rval); } static void miss(NODE *p, char *tail) { int create; char *tp; const char *type, *what; int serr; for (; p; p = p->next) { if (p->flags & F_OPT && !(p->flags & F_VISIT)) continue; if (p->type != F_DIR && (dflag || p->flags & F_VISIT)) continue; (void)strcpy(tail, p->name); if (!(p->flags & F_VISIT)) { /* Don't print missing message if file exists as a symbolic link and the -q flag is set. */ struct stat statbuf; if (qflag && stat(path, &statbuf) == 0) p->flags |= F_VISIT; else (void)printf("%s missing", path); } if (p->type != F_DIR && p->type != F_LINK) { putchar('\n'); continue; } create = 0; if (p->type == F_LINK) type = "symlink"; else type = "directory"; if (!(p->flags & F_VISIT) && uflag) { if (!(p->flags & (F_UID | F_UNAME))) (void)printf(" (%s not created: user not specified)", type); else if (!(p->flags & (F_GID | F_GNAME))) (void)printf(" (%s not created: group not specified)", type); else if (p->type == F_LINK) { if (symlink(p->slink, path)) (void)printf(" (symlink not created: %s)\n", strerror(errno)); else (void)printf(" (created)\n"); if (lchown(path, p->st_uid, p->st_gid) == -1) { serr = errno; if (p->st_uid == (uid_t)-1) what = "group"; else if (lchown(path, (uid_t)-1, p->st_gid) == -1) what = "user & group"; else { what = "user"; errno = serr; } (void)printf("%s: %s not modified: %s" "\n", path, what, strerror(errno)); } continue; } else if (!(p->flags & F_MODE)) (void)printf(" (directory not created: mode not specified)"); else if (mkdir(path, S_IRWXU)) (void)printf(" (directory not created: %s)", strerror(errno)); else { create = 1; (void)printf(" (created)"); } } if (!(p->flags & F_VISIT)) (void)putchar('\n'); for (tp = tail; *tp; ++tp); *tp = '/'; miss(p->child, tp + 1); *tp = '\0'; - if (!create) + if (!create && !uflag) continue; if (chown(path, p->st_uid, p->st_gid) == -1) { serr = errno; if (p->st_uid == (uid_t)-1) what = "group"; else if (chown(path, (uid_t)-1, p->st_gid) == -1) what = "user & group"; else { what = "user"; errno = serr; } (void)printf("%s: %s not modified: %s\n", path, what, strerror(errno)); } if (chmod(path, p->st_mode)) (void)printf("%s: permissions not set: %s\n", path, strerror(errno)); if ((p->flags & F_FLAGS) && p->st_flags && chflags(path, p->st_flags)) (void)printf("%s: file flags not set: %s\n", path, strerror(errno)); } } Index: projects/building-blocks/usr.sbin/pw/tests/pw_delete.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_delete.sh (revision 275663) +++ projects/building-blocks/usr.sbin/pw/tests/pw_delete.sh (nonexistent) @@ -1,47 +0,0 @@ -# $FreeBSD$ - -# Import helper functions -. $(atf_get_srcdir)/helper_functions.shin - -# Test that a user can be deleted when another user is part of this -# user's default group and does not go into an infinate loop. -# PR: 191427 -atf_test_case rmuser_seperate_group cleanup -rmuser_seperate_group_head() { - atf_set "timeout" "30" -} -rmuser_seperate_group_body() { - populate_etc_skel - ${PW} useradd test || atf_fail "Creating test user" - ${PW} groupmod test -M 'test,root' || \ - atf_fail "Modifying the group" - ${PW} userdel test || atf_fail "Delete the test user" -} - -atf_test_case group_do_not_delete_wheel_if_group_unknown -group_do_not_delete_wheel_if_group_unknown_head() { - atf_set "descr" "Make sure we do not consider gid 0 an unknown group" -} - -group_do_not_delete_wheel_if_group_unknown_body() { - populate_etc_skel - atf_check -s exit:0 -o inline:"wheel:*:0:root\n" -x ${PW} groupshow wheel - atf_check -e inline:"pw: -g expects a number\n" -s exit:64 -x ${PW} groupdel -g I_do_not_exist - atf_check -s exit:0 -o inline:"wheel:*:0:root\n" -x ${PW} groupshow wheel -} - -atf_test_case user_do_not_try_to_delete_root_if_user_unknown -user_do_not_try_to_delete_root_if_user_unknown_head() { - atf_set "descr" "Make sure not to try to remove root if deleting an unknown user" -} - -user_do_not_try_to_delete_root_if_user_unknown_body() { - populate_etc_skel - atf_check -e inline:"pw: -u expects a number\n" -s exit:64 -x ${PW} userdel -u plop -} - -atf_init_test_cases() { - atf_add_test_case rmuser_seperate_group - atf_add_test_case group_do_not_delete_wheel_if_group_unknown - atf_add_test_case user_do_not_try_to_delete_root_if_user_unknown -} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_delete.sh ___________________________________________________________________ Deleted: svn:eol-style ## -1 +0,0 ## -native \ No newline at end of property Deleted: svn:executable ## -1 +0,0 ## -* \ No newline at end of property Deleted: svn:keywords ## -1 +0,0 ## -FreeBSD=%H \ No newline at end of property Deleted: svn:mime-type ## -1 +0,0 ## -text/plain \ No newline at end of property Index: projects/building-blocks/usr.sbin/pw/tests/pw_modify.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_modify.sh (revision 275663) +++ projects/building-blocks/usr.sbin/pw/tests/pw_modify.sh (nonexistent) @@ -1,80 +0,0 @@ -# $FreeBSD$ - -# Import helper functions -. $(atf_get_srcdir)/helper_functions.shin - - -# Test adding & removing a user from a group -atf_test_case groupmod_user -groupmod_user_body() { - populate_etc_skel - atf_check -s exit:0 ${PW} addgroup test - atf_check -s exit:0 ${PW} groupmod test -m root - atf_check -s exit:0 -o match:"^test:\*:1001:root$" \ - grep "^test:\*:.*:root$" $HOME/group - atf_check -s exit:0 ${PW} groupmod test -d root - atf_check -s exit:0 -o match:"^test:\*:1001:$" \ - grep "^test:\*:.*:$" $HOME/group -} - - -# Test adding and removing a user that does not exist -atf_test_case groupmod_invalid_user -groupmod_invalid_user_body() { - populate_etc_skel - atf_check -s exit:0 ${PW} addgroup test - atf_check -s exit:67 -e match:"does not exist" ${PW} groupmod test -m foo - atf_check -s exit:0 ${PW} groupmod test -d foo -} - -atf_test_case groupmod_bug_193704 -groupmod_bug_193704_head() { - atf_set "descr" "Regression test for the #193704 bug" -} -groupmod_bug_193704_body() { - populate_etc_skel - atf_check -s exit:0 -x ${PW} groupadd test - atf_check -s exit:0 -x ${PW} groupmod test -l newgroupname - atf_check -s exit:65 -e match:"^pw: unknown group" -x ${PW} groupshow test -} - -atf_test_case usermod_bug_185666 -usermod_bug_185666_head() { - atf_set "descr" "Regression test for the #185666 bug" -} - -usermod_bug_185666_body() { - populate_etc_skel - atf_check -s exit:0 -x ${PW} useradd testuser - atf_check -s exit:0 -x ${PW} groupadd testgroup - atf_check -s exit:0 -x ${PW} groupadd testgroup2 - atf_check -s exit:0 -x ${PW} usermod testuser -G testgroup - atf_check -o inline:"testuser:*:1001:\n" -x ${PW} groupshow testuser - atf_check -o inline:"testgroup:*:1002:testuser\n" -x ${PW} groupshow testgroup - atf_check -o inline:"testgroup2:*:1003:\n" -x ${PW} groupshow testgroup2 - atf_check -s exit:0 -x ${PW} usermod testuser -G testgroup2 - atf_check -o inline:"testuser:*:1001:\n" -x ${PW} groupshow testuser - atf_check -o inline:"testgroup:*:1002:\n" -x ${PW} groupshow testgroup - atf_check -o inline:"testgroup2:*:1003:testuser\n" -x ${PW} groupshow testgroup2 -} - -atf_test_case do_not_duplicate_group_on_gid_change -do_not_duplicate_group_on_gid_change_head() { - atf_set "descr" "Do not duplicate group on gid change" -} - -do_not_duplicate_group_on_gid_change_body() { - populate_etc_skel - atf_check -s exit:0 -x ${PW} groupadd testgroup - atf_check -s exit:0 -x ${PW} groupmod testgroup -g 12345 - # use grep to see if the entry has not be duplicated - atf_check -o inline:"testgroup:*:12345:\n" -s exit:0 -x grep "^testgroup" ${HOME}/group -} - -atf_init_test_cases() { - atf_add_test_case groupmod_user - atf_add_test_case groupmod_invalid_user - atf_add_test_case groupmod_bug_193704 - atf_add_test_case usermod_bug_185666 - atf_add_test_case do_not_duplicate_group_on_gid_change -} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_modify.sh ___________________________________________________________________ Deleted: svn:eol-style ## -1 +0,0 ## -native \ No newline at end of property Deleted: svn:executable ## -1 +0,0 ## -* \ No newline at end of property Deleted: svn:keywords ## -1 +0,0 ## -FreeBSD=%H \ No newline at end of property Deleted: svn:mime-type ## -1 +0,0 ## -text/plain \ No newline at end of property Index: projects/building-blocks/usr.sbin/pw/tests/Makefile =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/Makefile (revision 275663) +++ projects/building-blocks/usr.sbin/pw/tests/Makefile (revision 275664) @@ -1,24 +1,31 @@ # $FreeBSD$ TESTSRC= ${.CURDIR}/../../../contrib/netbsd-tests/usr.sbin/useradd .PATH: ${TESTSRC} TESTSDIR= ${TESTSBASE}/usr.sbin/pw -ATF_TESTS_SH= pw_delete pw_lock pw_modify pw_etcdir +ATF_TESTS_SH= pw_etcdir \ + pw_lock \ + pw_groupdel \ + pw_groupmod \ + pw_useradd \ + pw_userdel \ + pw_usermod -TEST_METADATA.pw_delete+= required_user="root" -TEST_METADATA.pw_modify+= required_user="root" +.for tp in ${ATF_TESTS_SH} +TEST_METADATA.${tp}+= required_user="root" +.endfor FILES= group helper_functions.shin master.passwd FILESDIR= ${TESTSDIR} ATF_TESTS_SH+= pw_test # - user{add,del} does not exist on FreeBSD; use pw user{add,del} instead # - The command passes on FreeBSD ATF_TESTS_SH_SED_pw_test= -e 's/useradd /pw useradd /' ATF_TESTS_SH_SED_pw_test+= -e 's/userdel /pw userdel /' ATF_TESTS_SH_SED_pw_test+= -e '/atf_expect_fail "PR bin\/39546"/d' ATF_TESTS_SH_SRC_pw_test= t_useradd.sh .include Index: projects/building-blocks/usr.sbin/pw/tests/pw_groupdel.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_groupdel.sh (nonexistent) +++ projects/building-blocks/usr.sbin/pw/tests/pw_groupdel.sh (revision 275664) @@ -0,0 +1,24 @@ +# $FreeBSD$ + +# Import helper functions +. $(atf_get_srcdir)/helper_functions.shin + + +# Test to make sure we do not accidentially delete wheel when trying to delete +# an unknown group +atf_test_case group_do_not_delete_wheel_if_group_unknown +group_do_not_delete_wheel_if_group_unknown_head() { + atf_set "descr" "Make sure we do not consider gid 0 an unknown group" +} +group_do_not_delete_wheel_if_group_unknown_body() { + populate_etc_skel + atf_check -s exit:0 -o inline:"wheel:*:0:root\n" -x ${PW} groupshow wheel + atf_check -e inline:"pw: -g expects a number\n" -s exit:64 -x \ + ${PW} groupdel -g I_do_not_exist + atf_check -s exit:0 -o inline:"wheel:*:0:root\n" -x ${PW} groupshow wheel +} + + +atf_init_test_cases() { + atf_add_test_case group_do_not_delete_wheel_if_group_unknown +} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_groupdel.sh ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:executable ## -0,0 +1 ## +* \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/usr.sbin/pw/tests/pw_groupmod.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_groupmod.sh (nonexistent) +++ projects/building-blocks/usr.sbin/pw/tests/pw_groupmod.sh (revision 275664) @@ -0,0 +1,80 @@ +# $FreeBSD$ + +# Import helper functions +. $(atf_get_srcdir)/helper_functions.shin + + +# Test adding & removing a user from a group +atf_test_case groupmod_user +groupmod_user_body() { + populate_etc_skel + atf_check -s exit:0 ${PW} addgroup test + atf_check -s exit:0 ${PW} groupmod test -m root + atf_check -s exit:0 -o match:"^test:\*:1001:root$" \ + grep "^test:\*:.*:root$" $HOME/group + atf_check -s exit:0 ${PW} groupmod test -d root + atf_check -s exit:0 -o match:"^test:\*:1001:$" \ + grep "^test:\*:.*:$" $HOME/group +} + + +# Test adding and removing a user that does not exist +atf_test_case groupmod_invalid_user +groupmod_invalid_user_body() { + populate_etc_skel + atf_check -s exit:0 ${PW} addgroup test + atf_check -s exit:67 -e match:"does not exist" ${PW} groupmod test -m foo + atf_check -s exit:0 ${PW} groupmod test -d foo +} + +atf_test_case groupmod_bug_193704 +groupmod_bug_193704_head() { + atf_set "descr" "Regression test for the #193704 bug" +} +groupmod_bug_193704_body() { + populate_etc_skel + atf_check -s exit:0 -x ${PW} groupadd test + atf_check -s exit:0 -x ${PW} groupmod test -l newgroupname + atf_check -s exit:65 -e match:"^pw: unknown group" -x ${PW} groupshow test +} + +atf_test_case usermod_bug_185666 +usermod_bug_185666_head() { + atf_set "descr" "Regression test for the #185666 bug" +} + +usermod_bug_185666_body() { + populate_etc_skel + atf_check -s exit:0 -x ${PW} useradd testuser + atf_check -s exit:0 -x ${PW} groupadd testgroup + atf_check -s exit:0 -x ${PW} groupadd testgroup2 + atf_check -s exit:0 -x ${PW} usermod testuser -G testgroup + atf_check -o inline:"testuser:*:1001:\n" -x ${PW} groupshow testuser + atf_check -o inline:"testgroup:*:1002:testuser\n" -x ${PW} groupshow testgroup + atf_check -o inline:"testgroup2:*:1003:\n" -x ${PW} groupshow testgroup2 + atf_check -s exit:0 -x ${PW} usermod testuser -G testgroup2 + atf_check -o inline:"testuser:*:1001:\n" -x ${PW} groupshow testuser + atf_check -o inline:"testgroup:*:1002:\n" -x ${PW} groupshow testgroup + atf_check -o inline:"testgroup2:*:1003:testuser\n" -x ${PW} groupshow testgroup2 +} + +atf_test_case do_not_duplicate_group_on_gid_change +do_not_duplicate_group_on_gid_change_head() { + atf_set "descr" "Do not duplicate group on gid change" +} + +do_not_duplicate_group_on_gid_change_body() { + populate_etc_skel + atf_check -s exit:0 -x ${PW} groupadd testgroup + atf_check -s exit:0 -x ${PW} groupmod testgroup -g 12345 + # use grep to see if the entry has not be duplicated + atf_check -o inline:"testgroup:*:12345:\n" -s exit:0 -x grep "^testgroup" ${HOME}/group +} + +atf_init_test_cases() { + atf_add_test_case groupmod_user + atf_add_test_case groupmod_invalid_user + atf_add_test_case groupmod_bug_193704 + atf_add_test_case usermod_bug_185666 + atf_add_test_case do_not_duplicate_group_on_gid_change +} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_groupmod.sh ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:executable ## -0,0 +1 ## +* \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/usr.sbin/pw/tests/pw_useradd.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_useradd.sh (nonexistent) +++ projects/building-blocks/usr.sbin/pw/tests/pw_useradd.sh (revision 275664) @@ -0,0 +1,40 @@ +# $FreeBSD$ + +# Import helper functions +. $(atf_get_srcdir)/helper_functions.shin + +# Test add user +atf_test_case user_add +user_add_body() { + populate_etc_skel + + atf_check -s exit:0 ${PW} useradd test + atf_check -s exit:0 -o match:"^test:.*" \ + grep "^test:.*" $HOME/master.passwd +} + + +atf_test_case user_add_comments +user_add_comments_body() { + populate_etc_skel + + atf_check -s exit:0 ${PW} useradd test -c "Test User,work,123,456" + atf_check -s exit:0 -o match:"^test:.*:Test User,work,123,456:" \ + grep "^test:.*:Test User,work,123,456:" $HOME/master.passwd +} + +atf_test_case user_add_comments_invalid +user_add_comments_invalid_body() { + populate_etc_skel + + atf_check -s exit:65 -e match:"invalid character" \ + ${PW} useradd test -c "Test User,work,123:456,456" + atf_check -s exit:1 -o empty \ + grep "^test:.*:Test User,work,123:456,456:" $HOME/master.passwd +} + +atf_init_test_cases() { + atf_add_test_case user_add + atf_add_test_case user_add_comments + atf_add_test_case user_add_comments_invalid +} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_useradd.sh ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:executable ## -0,0 +1 ## +* \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/usr.sbin/pw/tests/pw_userdel.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_userdel.sh (nonexistent) +++ projects/building-blocks/usr.sbin/pw/tests/pw_userdel.sh (revision 275664) @@ -0,0 +1,37 @@ +# $FreeBSD$ + +# Import helper functions +. $(atf_get_srcdir)/helper_functions.shin + + +# Test that a user can be deleted when another user is part of this +# user's default group and does not go into an infinate loop. +# PR: 191427 +atf_test_case rmuser_seperate_group cleanup +rmuser_seperate_group_head() { + atf_set "timeout" "30" +} +rmuser_seperate_group_body() { + populate_etc_skel + ${PW} useradd test || atf_fail "Creating test user" + ${PW} groupmod test -M 'test,root' || \ + atf_fail "Modifying the group" + ${PW} userdel test || atf_fail "Delete the test user" +} + + +atf_test_case user_do_not_try_to_delete_root_if_user_unknown +user_do_not_try_to_delete_root_if_user_unknown_head() { + atf_set "descr" \ + "Make sure not to try to remove root if deleting an unknown user" +} +user_do_not_try_to_delete_root_if_user_unknown_body() { + populate_etc_skel + atf_check -e inline:"pw: -u expects a number\n" -s exit:64 -x \ + ${PW} userdel -u plop +} + +atf_init_test_cases() { + atf_add_test_case rmuser_seperate_group + atf_add_test_case user_do_not_try_to_delete_root_if_user_unknown +} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_userdel.sh ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:executable ## -0,0 +1 ## +* \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks/usr.sbin/pw/tests/pw_usermod.sh =================================================================== --- projects/building-blocks/usr.sbin/pw/tests/pw_usermod.sh (nonexistent) +++ projects/building-blocks/usr.sbin/pw/tests/pw_usermod.sh (revision 275664) @@ -0,0 +1,56 @@ +# $FreeBSD$ + +# Import helper functions +. $(atf_get_srcdir)/helper_functions.shin + +# Test modifying a user +atf_test_case user_mod +user_mod_body() { + populate_etc_skel + + atf_check -s exit:67 -e match:"no such user" ${PW} usermod test + atf_check -s exit:0 ${PW} useradd test + atf_check -s exit:0 ${PW} usermod test + atf_check -s exit:0 -o match:"^test:.*" \ + grep "^test:.*" $HOME/master.passwd +} + +# Test modifying a user with comments +atf_test_case user_mod_comments +user_mod_comments_body() { + populate_etc_skel + + atf_check -s exit:0 ${PW} useradd test -c "Test User,home,123,456" + atf_check -s exit:0 ${PW} usermod test -c "Test User,work,123,456" + atf_check -s exit:0 -o match:"^test:.*:Test User,work,123,456:" \ + grep "^test:.*:Test User,work,123,456:" $HOME/master.passwd +} + +# Test modifying a user with invalid comments +atf_test_case user_mod_comments_invalid +user_mod_comments_invalid_body() { + populate_etc_skel + + atf_check -s exit:0 ${PW} useradd test + atf_check -s exit:65 -e match:"invalid character" \ + ${PW} usermod test -c "Test User,work,123:456,456" + atf_check -s exit:1 -o empty \ + grep "^test:.*:Test User,work,123:456,456:" $HOME/master.passwd +} + +# Test modifying a user name with -l +atf_test_case user_mod_name +user_mod_name_body() { + populate_etc_skel + + atf_check -s exit:0 ${PW} useradd foo + atf_check -s exit:0 ${PW} usermod foo -l "bar" + atf_check -s exit:0 -o match:"^bar:.*" \ + grep "^bar:.*" $HOME/master.passwd +} +atf_init_test_cases() { + atf_add_test_case user_mod + atf_add_test_case user_mod_comments + atf_add_test_case user_mod_comments_invalid + atf_add_test_case user_mod_name +} Property changes on: projects/building-blocks/usr.sbin/pw/tests/pw_usermod.sh ___________________________________________________________________ Added: svn:eol-style ## -0,0 +1 ## +native \ No newline at end of property Added: svn:executable ## -0,0 +1 ## +* \ No newline at end of property Added: svn:keywords ## -0,0 +1 ## +FreeBSD=%H \ No newline at end of property Added: svn:mime-type ## -0,0 +1 ## +text/plain \ No newline at end of property Index: projects/building-blocks =================================================================== --- projects/building-blocks (revision 275663) +++ projects/building-blocks (revision 275664) Property changes on: projects/building-blocks ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head:r275590-275663