Index: head/include/unistd.h =================================================================== --- head/include/unistd.h (revision 345981) +++ head/include/unistd.h (revision 345982) @@ -1,596 +1,597 @@ /*- * SPDX-License-Identifier: BSD-3-Clause * * Copyright (c) 1991, 1993, 1994 * 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. * * @(#)unistd.h 8.12 (Berkeley) 4/27/95 * $FreeBSD$ */ #ifndef _UNISTD_H_ #define _UNISTD_H_ #include #include /* XXX adds too much pollution. */ #include #include #include #ifndef _GID_T_DECLARED typedef __gid_t gid_t; #define _GID_T_DECLARED #endif #ifndef _OFF_T_DECLARED typedef __off_t off_t; #define _OFF_T_DECLARED #endif #ifndef _PID_T_DECLARED typedef __pid_t pid_t; #define _PID_T_DECLARED #endif #ifndef _SIZE_T_DECLARED typedef __size_t size_t; #define _SIZE_T_DECLARED #endif #ifndef _SSIZE_T_DECLARED typedef __ssize_t ssize_t; #define _SSIZE_T_DECLARED #endif #ifndef _UID_T_DECLARED typedef __uid_t uid_t; #define _UID_T_DECLARED #endif #ifndef _USECONDS_T_DECLARED typedef __useconds_t useconds_t; #define _USECONDS_T_DECLARED #endif #define STDIN_FILENO 0 /* standard input file descriptor */ #define STDOUT_FILENO 1 /* standard output file descriptor */ #define STDERR_FILENO 2 /* standard error file descriptor */ #if __XSI_VISIBLE || __POSIX_VISIBLE >= 200112 #define F_ULOCK 0 /* unlock locked section */ #define F_LOCK 1 /* lock a section for exclusive use */ #define F_TLOCK 2 /* test and lock a section for exclusive use */ #define F_TEST 3 /* test a section for locks by other procs */ #endif /* * POSIX options and option groups we unconditionally do or don't * implement. This list includes those options which are exclusively * implemented (or not) in user mode. Please keep this list in * alphabetical order. * * Anything which is defined as zero below **must** have an * implementation for the corresponding sysconf() which is able to * determine conclusively whether or not the feature is supported. * Anything which is defined as other than -1 below **must** have * complete headers, types, and function declarations as specified by * the POSIX standard; however, if the relevant sysconf() function * returns -1, the functions may be stubbed out. */ #define _POSIX_BARRIERS 200112L #define _POSIX_CPUTIME 200112L #define _POSIX_READER_WRITER_LOCKS 200112L #define _POSIX_REGEXP 1 #define _POSIX_SHELL 1 #define _POSIX_SPAWN 200112L #define _POSIX_SPIN_LOCKS 200112L #define _POSIX_THREAD_ATTR_STACKADDR 200112L #define _POSIX_THREAD_ATTR_STACKSIZE 200112L #define _POSIX_THREAD_CPUTIME 200112L #define _POSIX_THREAD_PRIO_INHERIT 200112L #define _POSIX_THREAD_PRIO_PROTECT 200112L #define _POSIX_THREAD_PRIORITY_SCHEDULING 200112L #define _POSIX_THREAD_PROCESS_SHARED 200112L #define _POSIX_THREAD_SAFE_FUNCTIONS -1 #define _POSIX_THREAD_SPORADIC_SERVER -1 #define _POSIX_THREADS 200112L #define _POSIX_TRACE -1 #define _POSIX_TRACE_EVENT_FILTER -1 #define _POSIX_TRACE_INHERIT -1 #define _POSIX_TRACE_LOG -1 #define _POSIX2_C_BIND 200112L /* mandatory */ #define _POSIX2_C_DEV -1 /* need c99 utility */ #define _POSIX2_CHAR_TERM 1 #define _POSIX2_FORT_DEV -1 /* need fort77 utility */ #define _POSIX2_FORT_RUN 200112L #define _POSIX2_LOCALEDEF -1 #define _POSIX2_PBS -1 #define _POSIX2_PBS_ACCOUNTING -1 #define _POSIX2_PBS_CHECKPOINT -1 #define _POSIX2_PBS_LOCATE -1 #define _POSIX2_PBS_MESSAGE -1 #define _POSIX2_PBS_TRACK -1 #define _POSIX2_SW_DEV -1 /* XXX ??? */ #define _POSIX2_UPE 200112L #define _V6_ILP32_OFF32 -1 #define _V6_ILP32_OFFBIG 0 #define _V6_LP64_OFF64 0 #define _V6_LPBIG_OFFBIG -1 #if __XSI_VISIBLE #define _XOPEN_CRYPT -1 /* XXX ??? */ #define _XOPEN_ENH_I18N -1 /* mandatory in XSI */ #define _XOPEN_LEGACY -1 #define _XOPEN_REALTIME -1 #define _XOPEN_REALTIME_THREADS -1 #define _XOPEN_UNIX -1 #endif /* Define the POSIX.2 version we target for compliance. */ #define _POSIX2_VERSION 199212L /* * POSIX-style system configuration variable accessors (for the * sysconf function). The kernel does not directly implement the * sysconf() interface; rather, a C library stub translates references * to sysconf() into calls to sysctl() using a giant switch statement. * Those that are marked `user' are implemented entirely in the C * library and never query the kernel. pathconf() is implemented * directly by the kernel so those are not defined here. */ #define _SC_ARG_MAX 1 #define _SC_CHILD_MAX 2 #define _SC_CLK_TCK 3 #define _SC_NGROUPS_MAX 4 #define _SC_OPEN_MAX 5 #define _SC_JOB_CONTROL 6 #define _SC_SAVED_IDS 7 #define _SC_VERSION 8 #define _SC_BC_BASE_MAX 9 /* user */ #define _SC_BC_DIM_MAX 10 /* user */ #define _SC_BC_SCALE_MAX 11 /* user */ #define _SC_BC_STRING_MAX 12 /* user */ #define _SC_COLL_WEIGHTS_MAX 13 /* user */ #define _SC_EXPR_NEST_MAX 14 /* user */ #define _SC_LINE_MAX 15 /* user */ #define _SC_RE_DUP_MAX 16 /* user */ #define _SC_2_VERSION 17 /* user */ #define _SC_2_C_BIND 18 /* user */ #define _SC_2_C_DEV 19 /* user */ #define _SC_2_CHAR_TERM 20 /* user */ #define _SC_2_FORT_DEV 21 /* user */ #define _SC_2_FORT_RUN 22 /* user */ #define _SC_2_LOCALEDEF 23 /* user */ #define _SC_2_SW_DEV 24 /* user */ #define _SC_2_UPE 25 /* user */ #define _SC_STREAM_MAX 26 /* user */ #define _SC_TZNAME_MAX 27 /* user */ #if __POSIX_VISIBLE >= 199309 #define _SC_ASYNCHRONOUS_IO 28 #define _SC_MAPPED_FILES 29 #define _SC_MEMLOCK 30 #define _SC_MEMLOCK_RANGE 31 #define _SC_MEMORY_PROTECTION 32 #define _SC_MESSAGE_PASSING 33 #define _SC_PRIORITIZED_IO 34 #define _SC_PRIORITY_SCHEDULING 35 #define _SC_REALTIME_SIGNALS 36 #define _SC_SEMAPHORES 37 #define _SC_FSYNC 38 #define _SC_SHARED_MEMORY_OBJECTS 39 #define _SC_SYNCHRONIZED_IO 40 #define _SC_TIMERS 41 #define _SC_AIO_LISTIO_MAX 42 #define _SC_AIO_MAX 43 #define _SC_AIO_PRIO_DELTA_MAX 44 #define _SC_DELAYTIMER_MAX 45 #define _SC_MQ_OPEN_MAX 46 #define _SC_PAGESIZE 47 #define _SC_RTSIG_MAX 48 #define _SC_SEM_NSEMS_MAX 49 #define _SC_SEM_VALUE_MAX 50 #define _SC_SIGQUEUE_MAX 51 #define _SC_TIMER_MAX 52 #endif #if __POSIX_VISIBLE >= 200112 #define _SC_2_PBS 59 /* user */ #define _SC_2_PBS_ACCOUNTING 60 /* user */ #define _SC_2_PBS_CHECKPOINT 61 /* user */ #define _SC_2_PBS_LOCATE 62 /* user */ #define _SC_2_PBS_MESSAGE 63 /* user */ #define _SC_2_PBS_TRACK 64 /* user */ #define _SC_ADVISORY_INFO 65 #define _SC_BARRIERS 66 /* user */ #define _SC_CLOCK_SELECTION 67 #define _SC_CPUTIME 68 #define _SC_FILE_LOCKING 69 #define _SC_GETGR_R_SIZE_MAX 70 /* user */ #define _SC_GETPW_R_SIZE_MAX 71 /* user */ #define _SC_HOST_NAME_MAX 72 #define _SC_LOGIN_NAME_MAX 73 #define _SC_MONOTONIC_CLOCK 74 #define _SC_MQ_PRIO_MAX 75 #define _SC_READER_WRITER_LOCKS 76 /* user */ #define _SC_REGEXP 77 /* user */ #define _SC_SHELL 78 /* user */ #define _SC_SPAWN 79 /* user */ #define _SC_SPIN_LOCKS 80 /* user */ #define _SC_SPORADIC_SERVER 81 #define _SC_THREAD_ATTR_STACKADDR 82 /* user */ #define _SC_THREAD_ATTR_STACKSIZE 83 /* user */ #define _SC_THREAD_CPUTIME 84 /* user */ #define _SC_THREAD_DESTRUCTOR_ITERATIONS 85 /* user */ #define _SC_THREAD_KEYS_MAX 86 /* user */ #define _SC_THREAD_PRIO_INHERIT 87 /* user */ #define _SC_THREAD_PRIO_PROTECT 88 /* user */ #define _SC_THREAD_PRIORITY_SCHEDULING 89 /* user */ #define _SC_THREAD_PROCESS_SHARED 90 /* user */ #define _SC_THREAD_SAFE_FUNCTIONS 91 /* user */ #define _SC_THREAD_SPORADIC_SERVER 92 /* user */ #define _SC_THREAD_STACK_MIN 93 /* user */ #define _SC_THREAD_THREADS_MAX 94 /* user */ #define _SC_TIMEOUTS 95 /* user */ #define _SC_THREADS 96 /* user */ #define _SC_TRACE 97 /* user */ #define _SC_TRACE_EVENT_FILTER 98 /* user */ #define _SC_TRACE_INHERIT 99 /* user */ #define _SC_TRACE_LOG 100 /* user */ #define _SC_TTY_NAME_MAX 101 /* user */ #define _SC_TYPED_MEMORY_OBJECTS 102 #define _SC_V6_ILP32_OFF32 103 /* user */ #define _SC_V6_ILP32_OFFBIG 104 /* user */ #define _SC_V6_LP64_OFF64 105 /* user */ #define _SC_V6_LPBIG_OFFBIG 106 /* user */ #define _SC_IPV6 118 #define _SC_RAW_SOCKETS 119 #define _SC_SYMLOOP_MAX 120 #endif #if __XSI_VISIBLE #define _SC_ATEXIT_MAX 107 /* user */ #define _SC_IOV_MAX 56 #define _SC_PAGE_SIZE _SC_PAGESIZE #define _SC_XOPEN_CRYPT 108 /* user */ #define _SC_XOPEN_ENH_I18N 109 /* user */ #define _SC_XOPEN_LEGACY 110 /* user */ #define _SC_XOPEN_REALTIME 111 #define _SC_XOPEN_REALTIME_THREADS 112 #define _SC_XOPEN_SHM 113 #define _SC_XOPEN_STREAMS 114 #define _SC_XOPEN_UNIX 115 #define _SC_XOPEN_VERSION 116 #define _SC_XOPEN_XCU_VERSION 117 /* user */ #endif #if __BSD_VISIBLE #define _SC_NPROCESSORS_CONF 57 #define _SC_NPROCESSORS_ONLN 58 #define _SC_CPUSET_SIZE 122 #endif /* Extensions found in Solaris and Linux. */ #define _SC_PHYS_PAGES 121 /* Keys for the confstr(3) function. */ #if __POSIX_VISIBLE >= 199209 #define _CS_PATH 1 /* default value of PATH */ #endif #if __POSIX_VISIBLE >= 200112 #define _CS_POSIX_V6_ILP32_OFF32_CFLAGS 2 #define _CS_POSIX_V6_ILP32_OFF32_LDFLAGS 3 #define _CS_POSIX_V6_ILP32_OFF32_LIBS 4 #define _CS_POSIX_V6_ILP32_OFFBIG_CFLAGS 5 #define _CS_POSIX_V6_ILP32_OFFBIG_LDFLAGS 6 #define _CS_POSIX_V6_ILP32_OFFBIG_LIBS 7 #define _CS_POSIX_V6_LP64_OFF64_CFLAGS 8 #define _CS_POSIX_V6_LP64_OFF64_LDFLAGS 9 #define _CS_POSIX_V6_LP64_OFF64_LIBS 10 #define _CS_POSIX_V6_LPBIG_OFFBIG_CFLAGS 11 #define _CS_POSIX_V6_LPBIG_OFFBIG_LDFLAGS 12 #define _CS_POSIX_V6_LPBIG_OFFBIG_LIBS 13 #define _CS_POSIX_V6_WIDTH_RESTRICTED_ENVS 14 #endif __BEGIN_DECLS /* 1003.1-1990 */ void _exit(int) __dead2; int access(const char *, int); unsigned int alarm(unsigned int); int chdir(const char *); int chown(const char *, uid_t, gid_t); int close(int); void closefrom(int); int dup(int); int dup2(int, int); int execl(const char *, const char *, ...) __null_sentinel; int execle(const char *, const char *, ...); int execlp(const char *, const char *, ...) __null_sentinel; int execv(const char *, char * const *); int execve(const char *, char * const *, char * const *); int execvp(const char *, char * const *); pid_t fork(void); long fpathconf(int, int); char *getcwd(char *, size_t); gid_t getegid(void); uid_t geteuid(void); gid_t getgid(void); int getgroups(int, gid_t []); char *getlogin(void); pid_t getpgrp(void); pid_t getpid(void); pid_t getppid(void); uid_t getuid(void); int isatty(int); int link(const char *, const char *); #ifndef _LSEEK_DECLARED #define _LSEEK_DECLARED off_t lseek(int, off_t, int); #endif long pathconf(const char *, int); int pause(void); int pipe(int *); ssize_t read(int, void *, size_t); int rmdir(const char *); int setgid(gid_t); int setpgid(pid_t, pid_t); pid_t setsid(void); int setuid(uid_t); unsigned int sleep(unsigned int); long sysconf(int); pid_t tcgetpgrp(int); int tcsetpgrp(int, pid_t); char *ttyname(int); int ttyname_r(int, char *, size_t); int unlink(const char *); ssize_t write(int, const void *, size_t); /* 1003.2-1992 */ #if __POSIX_VISIBLE >= 199209 || __XSI_VISIBLE size_t confstr(int, char *, size_t); #ifndef _GETOPT_DECLARED #define _GETOPT_DECLARED int getopt(int, char * const [], const char *); extern char *optarg; /* getopt(3) external variables */ extern int optind, opterr, optopt; #endif /* _GETOPT_DECLARED */ #endif /* ISO/IEC 9945-1: 1996 */ #if __POSIX_VISIBLE >= 199506 || __XSI_VISIBLE int fsync(int); int fdatasync(int); /* * ftruncate() was in the POSIX Realtime Extension (it's used for shared * memory), but truncate() was not. */ #ifndef _FTRUNCATE_DECLARED #define _FTRUNCATE_DECLARED int ftruncate(int, off_t); #endif #endif #if __POSIX_VISIBLE >= 199506 int getlogin_r(char *, int); #endif /* 1003.1-2001 */ #if __POSIX_VISIBLE >= 200112 || __XSI_VISIBLE int fchown(int, uid_t, gid_t); ssize_t readlink(const char * __restrict, char * __restrict, size_t); #endif #if __POSIX_VISIBLE >= 200112 int gethostname(char *, size_t); int setegid(gid_t); int seteuid(uid_t); #endif /* 1003.1-2008 */ #if __POSIX_VISIBLE >= 200809 || __XSI_VISIBLE int getsid(pid_t _pid); int fchdir(int); int getpgid(pid_t _pid); int lchown(const char *, uid_t, gid_t); ssize_t pread(int, void *, size_t, off_t); ssize_t pwrite(int, const void *, size_t, off_t); /* See comment at ftruncate() above. */ #ifndef _TRUNCATE_DECLARED #define _TRUNCATE_DECLARED int truncate(const char *, off_t); #endif #endif /* __POSIX_VISIBLE >= 200809 || __XSI_VISIBLE */ #if __POSIX_VISIBLE >= 200809 int faccessat(int, const char *, int, int); int fchownat(int, const char *, uid_t, gid_t, int); int fexecve(int, char *const [], char *const []); int linkat(int, const char *, int, const char *, int); ssize_t readlinkat(int, const char * __restrict, char * __restrict, size_t); int symlinkat(const char *, int, const char *); int unlinkat(int, const char *, int); #endif /* __POSIX_VISIBLE >= 200809 */ /* * symlink() was originally in POSIX.1a, which was withdrawn after * being overtaken by events (1003.1-2001). It was in XPG4.2, and of * course has been in BSD since 4.2. */ #if __POSIX_VISIBLE >= 200112 || __XSI_VISIBLE >= 402 int symlink(const char * __restrict, const char * __restrict); #endif /* X/Open System Interfaces */ #if __XSI_VISIBLE char *crypt(const char *, const char *); long gethostid(void); int lockf(int, int, off_t); int nice(int); int setregid(gid_t, gid_t); int setreuid(uid_t, uid_t); #ifndef _SWAB_DECLARED #define _SWAB_DECLARED void swab(const void * __restrict, void * __restrict, ssize_t); #endif /* _SWAB_DECLARED */ void sync(void); #endif /* __XSI_VISIBLE */ #if (__XSI_VISIBLE && __XSI_VISIBLE <= 500) || __BSD_VISIBLE int brk(const void *); int chroot(const char *); int getdtablesize(void); int getpagesize(void) __pure2; char *getpass(const char *); void *sbrk(intptr_t); #endif #if (__XSI_VISIBLE && __XSI_VISIBLE <= 600) || __BSD_VISIBLE char *getwd(char *); /* obsoleted by getcwd() */ useconds_t ualarm(useconds_t, useconds_t); int usleep(useconds_t); pid_t vfork(void) __returns_twice; #endif #if __BSD_VISIBLE struct timeval; /* select(2) */ struct crypt_data { int initialized; /* For compatibility with glibc. */ char __buf[256]; /* Buffer returned by crypt_r(). */ }; int acct(const char *); int async_daemon(void); int check_utility_compat(const char *); const char * crypt_get_format(void); char *crypt_r(const char *, const char *, struct crypt_data *); int crypt_set_format(const char *); int dup3(int, int, int); int eaccess(const char *, int); void endusershell(void); int exect(const char *, char * const *, char * const *); int execvP(const char *, const char *, char * const *); int feature_present(const char *); char *fflagstostr(u_long); int getdomainname(char *, int); int getentropy(void *, size_t); int getgrouplist(const char *, gid_t, gid_t *, int *); int getloginclass(char *, size_t); mode_t getmode(const void *, mode_t); int getosreldate(void); int getpeereid(int, uid_t *, gid_t *); int getresgid(gid_t *, gid_t *, gid_t *); int getresuid(uid_t *, uid_t *, uid_t *); char *getusershell(void); int initgroups(const char *, gid_t); int iruserok(unsigned long, int, const char *, const char *); int iruserok_sa(const void *, int, int, const char *, const char *); int issetugid(void); void __FreeBSD_libc_enter_restricted_mode(void); long lpathconf(const char *, int); #ifndef _MKDTEMP_DECLARED char *mkdtemp(char *); #define _MKDTEMP_DECLARED #endif #ifndef _MKNOD_DECLARED int mknod(const char *, mode_t, dev_t); #define _MKNOD_DECLARED #endif #ifndef _MKSTEMP_DECLARED int mkstemp(char *); #define _MKSTEMP_DECLARED #endif int mkstemps(char *, int); #ifndef _MKTEMP_DECLARED char *mktemp(char *); #define _MKTEMP_DECLARED #endif int nfssvc(int, void *); int nlm_syscall(int, int, int, char **); int pipe2(int *, int); int profil(char *, size_t, vm_offset_t, int); int rcmd(char **, int, const char *, const char *, const char *, int *); int rcmd_af(char **, int, const char *, const char *, const char *, int *, int); int rcmdsh(char **, int, const char *, const char *, const char *, const char *); char *re_comp(const char *); int re_exec(const char *); int reboot(int); int revoke(const char *); pid_t rfork(int); pid_t rfork_thread(int, void *, int (*)(void *), void *); int rresvport(int *); int rresvport_af(int *, int); int ruserok(const char *, int, const char *, const char *); #if __BSD_VISIBLE #ifndef _SELECT_DECLARED #define _SELECT_DECLARED int select(int, fd_set *, fd_set *, fd_set *, struct timeval *); #endif #endif int setdomainname(const char *, int); int setgroups(int, const gid_t *); void sethostid(long); int sethostname(const char *, int); int setlogin(const char *); int setloginclass(const char *); void *setmode(const char *); int setpgrp(pid_t, pid_t); /* obsoleted by setpgid() */ void setproctitle(const char *_fmt, ...) __printf0like(1, 2); void setproctitle_fast(const char *_fmt, ...) __printf0like(1, 2); int setresgid(gid_t, gid_t, gid_t); int setresuid(uid_t, uid_t, uid_t); int setrgid(gid_t); int setruid(uid_t); void setusershell(void); int strtofflags(char **, u_long *, u_long *); int swapon(const char *); int swapoff(const char *); int syscall(int, ...); off_t __syscall(quad_t, ...); int undelete(const char *); int unwhiteout(const char *); void *valloc(size_t); /* obsoleted by malloc() */ +int funlinkat(int, const char *, int, int); #ifndef _OPTRESET_DECLARED #define _OPTRESET_DECLARED extern int optreset; /* getopt(3) external variable */ #endif #endif /* __BSD_VISIBLE */ __END_DECLS #endif /* !_UNISTD_H_ */ Index: head/lib/libc/sys/Makefile.inc =================================================================== --- head/lib/libc/sys/Makefile.inc (revision 345981) +++ head/lib/libc/sys/Makefile.inc (revision 345982) @@ -1,499 +1,500 @@ # @(#)Makefile.inc 8.3 (Berkeley) 10/24/94 # $FreeBSD$ # sys sources .PATH: ${LIBC_SRCTOP}/${LIBC_ARCH}/sys ${LIBC_SRCTOP}/sys # Include the generated makefile containing the *complete* list # of syscall names in MIASM. .include "${SRCTOP}/sys/sys/syscall.mk" # Include machine dependent definitions. # # MDASM names override the default syscall names in MIASM. # NOASM will prevent the default syscall code from being generated. # PSEUDO generates _() and __sys_() symbols, but not (). # # While historically machine dependent, all architectures have the following # declarations in common: # NOASM= exit.o \ getlogin.o \ sstk.o \ yield.o PSEUDO= _exit.o \ _getlogin.o .sinclude "${LIBC_SRCTOP}/${LIBC_ARCH}/sys/Makefile.inc" SRCS+= clock_gettime.c gettimeofday.c __vdso_gettimeofday.c NOASM+= clock_gettime.o gettimeofday.o PSEUDO+= _clock_gettime.o _gettimeofday.o # Sources common to both syscall interfaces: SRCS+= \ __error.c \ interposing_table.c SRCS+= getdents.c lstat.c mknod.c stat.c SRCS+= fstat.c fstatat.c fstatfs.c getfsstat.c statfs.c NOASM+= fstat.o fstatat.o fstatfs.o getfsstat.o statfs.o PSEUDO+= _fstat.o _fstatat.o _fstatfs.o _getfsstat.o _statfs.o SRCS+= getdirentries.c NOASM+= getdirentries.o PSEUDO+= _getdirentries.o SRCS+= brk.c SRCS+= pipe.c SRCS+= vadvise.c SRCS+= compat-stub.c INTERPOSED = \ accept \ accept4 \ aio_suspend \ clock_nanosleep \ close \ connect \ fcntl \ fdatasync \ fsync \ fork \ kevent \ msync \ nanosleep \ open \ openat \ poll \ ppoll \ pselect \ ptrace \ read \ readv \ recvfrom \ recvmsg \ select \ sendmsg \ sendto \ setcontext \ sigprocmask \ sigsuspend \ sigtimedwait \ sigwait \ sigwaitinfo \ swapcontext \ wait4 \ wait6 \ write \ writev .if ${MACHINE_CPUARCH} == "sparc64" SRCS+= sigaction.c NOASM+= sigaction.o .else INTERPOSED+= sigaction .endif SRCS+= ${INTERPOSED:S/$/.c/} NOASM+= ${INTERPOSED:S/$/.o/} PSEUDO+= ${INTERPOSED:C/^.*$/_&.o/} # Add machine dependent asm sources: SRCS+=${MDASM} # Look though the complete list of syscalls (MIASM) for names that are # not defined with machine dependent implementations (MDASM) and are # not declared for no generation of default code (NOASM). Add each # syscall that satisfies these conditions to the ASM list. .for _asm in ${MIASM} .if (${MDASM:R:M${_asm:R}} == "") .if (${NOASM:R:M${_asm:R}} == "") ASM+=$(_asm) .endif .endif .endfor SASM= ${ASM:S/.o/.S/} SPSEUDO= ${PSEUDO:S/.o/.S/} SRCS+= ${SASM} ${SPSEUDO} SYM_MAPS+= ${LIBC_SRCTOP}/sys/Symbol.map # Generated files CLEANFILES+= ${SASM} ${SPSEUDO} .if ${MACHINE_CPUARCH} == "amd64" || ${MACHINE_CPUARCH} == "i386" || \ ${MACHINE_CPUARCH} == "powerpc" || ${MACHINE_ARCH:Marmv[67]*} NOTE_GNU_STACK='\t.section .note.GNU-stack,"",%%progbits\n' .else NOTE_GNU_STACK='' .endif ${SASM}: printf '#include "compat.h"\n' > ${.TARGET} printf '#include "SYS.h"\nRSYSCALL(${.PREFIX})\n' >> ${.TARGET} printf ${NOTE_GNU_STACK} >>${.TARGET} ${SPSEUDO}: printf '#include "compat.h"\n' > ${.TARGET} printf '#include "SYS.h"\nPSEUDO(${.PREFIX:S/_//})\n' \ >> ${.TARGET} printf ${NOTE_GNU_STACK} >>${.TARGET} MAN+= abort2.2 \ accept.2 \ access.2 \ acct.2 \ adjtime.2 \ aio_cancel.2 \ aio_error.2 \ aio_fsync.2 \ aio_mlock.2 \ aio_read.2 \ aio_return.2 \ aio_suspend.2 \ aio_waitcomplete.2 \ aio_write.2 \ bind.2 \ bindat.2 \ brk.2 \ cap_enter.2 \ cap_fcntls_limit.2 \ cap_ioctls_limit.2 \ cap_rights_limit.2 \ chdir.2 \ chflags.2 \ chmod.2 \ chown.2 \ chroot.2 \ clock_gettime.2 \ close.2 \ closefrom.2 \ connect.2 \ connectat.2 \ cpuset.2 \ cpuset_getaffinity.2 \ cpuset_getdomain.2 \ dup.2 \ execve.2 \ _exit.2 \ extattr_get_file.2 \ fcntl.2 \ ffclock.2 \ fhlink.2 \ fhopen.2 \ fhreadlink.2 \ flock.2 \ fork.2 \ fsync.2 \ getdirentries.2 \ getdtablesize.2 \ getfh.2 \ getfsstat.2 \ getgid.2 \ getgroups.2 \ getitimer.2 \ getlogin.2 \ getloginclass.2 \ getpeername.2 \ getpgrp.2 \ getpid.2 \ getpriority.2 \ getrandom.2 \ getrlimit.2 \ getrusage.2 \ getsid.2 \ getsockname.2 \ getsockopt.2 \ gettimeofday.2 \ getuid.2 \ intro.2 \ ioctl.2 \ issetugid.2 \ jail.2 \ kenv.2 \ kill.2 \ kldfind.2 \ kldfirstmod.2 \ kldload.2 \ kldnext.2 \ kldstat.2 \ kldsym.2 \ kldunload.2 \ kqueue.2 \ ktrace.2 \ link.2 \ lio_listio.2 \ listen.2 \ lseek.2 \ madvise.2 \ mincore.2 \ minherit.2 \ mkdir.2 \ mkfifo.2 \ mknod.2 \ mlock.2 \ mlockall.2 \ mmap.2 \ modfind.2 \ modnext.2 \ modstat.2 \ mount.2 \ mprotect.2 \ mq_close.2 \ mq_getattr.2 \ mq_notify.2 \ mq_open.2 \ mq_receive.2 \ mq_send.2 \ mq_setattr.2 \ msgctl.2 \ msgget.2 \ msgrcv.2 \ msgsnd.2 \ msync.2 \ munmap.2 \ nanosleep.2 \ nfssvc.2 \ ntp_adjtime.2 \ open.2 \ pathconf.2 \ pdfork.2 \ pipe.2 \ poll.2 \ posix_fadvise.2 \ posix_fallocate.2 \ posix_openpt.2 \ procctl.2 \ profil.2 \ pselect.2 \ ptrace.2 \ quotactl.2 \ rctl_add_rule.2 \ read.2 \ readlink.2 \ reboot.2 \ recv.2 \ rename.2 \ revoke.2 \ rfork.2 \ rmdir.2 \ rtprio.2 .if !defined(NO_P1003_1B) MAN+= sched_get_priority_max.2 \ sched_setparam.2 \ sched_setscheduler.2 \ sched_yield.2 .endif MAN+= sctp_generic_recvmsg.2 \ sctp_generic_sendmsg.2 \ sctp_peeloff.2 \ select.2 \ semctl.2 \ semget.2 \ semop.2 \ send.2 \ setfib.2 \ sendfile.2 \ setgroups.2 \ setpgid.2 \ setregid.2 \ setresuid.2 \ setreuid.2 \ setsid.2 \ setuid.2 \ shmat.2 \ shmctl.2 \ shmget.2 \ shm_open.2 \ shutdown.2 \ sigaction.2 \ sigaltstack.2 \ sigpending.2 \ sigprocmask.2 \ sigqueue.2 \ sigreturn.2 \ sigstack.2 \ sigsuspend.2 \ sigwait.2 \ sigwaitinfo.2 \ socket.2 \ socketpair.2 \ stat.2 \ statfs.2 \ swapon.2 \ symlink.2 \ sync.2 \ sysarch.2 \ syscall.2 \ thr_exit.2 \ thr_kill.2 \ thr_new.2 \ thr_self.2 \ thr_set_name.2 \ thr_suspend.2 \ thr_wake.2 \ timer_create.2 \ timer_delete.2 \ timer_settime.2 \ truncate.2 \ umask.2 \ undelete.2 \ unlink.2 \ utimensat.2 \ utimes.2 \ utrace.2 \ uuidgen.2 \ vfork.2 \ wait.2 \ write.2 \ _umtx_op.2 MLINKS+=accept.2 accept4.2 MLINKS+=access.2 eaccess.2 \ access.2 faccessat.2 MLINKS+=brk.2 sbrk.2 MLINKS+=cap_enter.2 cap_getmode.2 MLINKS+=cap_fcntls_limit.2 cap_fcntls_get.2 MLINKS+=cap_ioctls_limit.2 cap_ioctls_get.2 MLINKS+=chdir.2 fchdir.2 MLINKS+=chflags.2 chflagsat.2 \ chflags.2 fchflags.2 \ chflags.2 lchflags.2 MLINKS+=chmod.2 fchmod.2 \ chmod.2 fchmodat.2 \ chmod.2 lchmod.2 MLINKS+=chown.2 fchown.2 \ chown.2 fchownat.2 \ chown.2 lchown.2 MLINKS+=clock_gettime.2 clock_getres.2 \ clock_gettime.2 clock_settime.2 MLINKS+=nanosleep.2 clock_nanosleep.2 MLINKS+=cpuset.2 cpuset_getid.2 \ cpuset.2 cpuset_setid.2 MLINKS+=cpuset_getaffinity.2 cpuset_setaffinity.2 MLINKS+=cpuset_getdomain.2 cpuset_setdomain.2 MLINKS+=dup.2 dup2.2 MLINKS+=execve.2 fexecve.2 MLINKS+=extattr_get_file.2 extattr.2 \ extattr_get_file.2 extattr_delete_fd.2 \ extattr_get_file.2 extattr_delete_file.2 \ extattr_get_file.2 extattr_delete_link.2 \ extattr_get_file.2 extattr_get_fd.2 \ extattr_get_file.2 extattr_get_link.2 \ extattr_get_file.2 extattr_list_fd.2 \ extattr_get_file.2 extattr_list_file.2 \ extattr_get_file.2 extattr_list_link.2 \ extattr_get_file.2 extattr_set_fd.2 \ extattr_get_file.2 extattr_set_file.2 \ extattr_get_file.2 extattr_set_link.2 MLINKS+=ffclock.2 ffclock_getcounter.2 \ ffclock.2 ffclock_getestimate.2 \ ffclock.2 ffclock_setestimate.2 MLINKS+=fhopen.2 fhstat.2 fhopen.2 fhstatfs.2 MLINKS+=fsync.2 fdatasync.2 MLINKS+=getdirentries.2 getdents.2 MLINKS+=getfh.2 lgetfh.2 \ getfh.2 getfhat.2 MLINKS+=getgid.2 getegid.2 MLINKS+=getitimer.2 setitimer.2 MLINKS+=getlogin.2 getlogin_r.3 MLINKS+=getlogin.2 setlogin.2 MLINKS+=getloginclass.2 setloginclass.2 MLINKS+=getpgrp.2 getpgid.2 MLINKS+=getpid.2 getppid.2 MLINKS+=getpriority.2 setpriority.2 MLINKS+=getrlimit.2 setrlimit.2 MLINKS+=getsockopt.2 setsockopt.2 MLINKS+=gettimeofday.2 settimeofday.2 MLINKS+=getuid.2 geteuid.2 MLINKS+=intro.2 errno.2 MLINKS+=jail.2 jail_attach.2 \ jail.2 jail_get.2 \ jail.2 jail_remove.2 \ jail.2 jail_set.2 MLINKS+=kldunload.2 kldunloadf.2 MLINKS+=kqueue.2 kevent.2 \ kqueue.2 EV_SET.3 MLINKS+=link.2 linkat.2 MLINKS+=madvise.2 posix_madvise.2 MLINKS+=mkdir.2 mkdirat.2 MLINKS+=mkfifo.2 mkfifoat.2 MLINKS+=mknod.2 mknodat.2 MLINKS+=mlock.2 munlock.2 MLINKS+=mlockall.2 munlockall.2 MLINKS+=modnext.2 modfnext.2 MLINKS+=mount.2 nmount.2 \ mount.2 unmount.2 MLINKS+=mq_receive.2 mq_timedreceive.2 MLINKS+=mq_send.2 mq_timedsend.2 MLINKS+=ntp_adjtime.2 ntp_gettime.2 MLINKS+=open.2 openat.2 MLINKS+=pathconf.2 fpathconf.2 MLINKS+=pathconf.2 lpathconf.2 MLINKS+=pdfork.2 pdgetpid.2\ pdfork.2 pdkill.2 MLINKS+=pipe.2 pipe2.2 MLINKS+=poll.2 ppoll.2 MLINKS+=rctl_add_rule.2 rctl_get_limits.2 \ rctl_add_rule.2 rctl_get_racct.2 \ rctl_add_rule.2 rctl_get_rules.2 \ rctl_add_rule.2 rctl_remove_rule.2 MLINKS+=read.2 pread.2 \ read.2 preadv.2 \ read.2 readv.2 MLINKS+=readlink.2 readlinkat.2 MLINKS+=recv.2 recvfrom.2 \ recv.2 recvmsg.2 MLINKS+=rename.2 renameat.2 MLINKS+=rtprio.2 rtprio_thread.2 .if !defined(NO_P1003_1B) MLINKS+=sched_get_priority_max.2 sched_get_priority_min.2 \ sched_get_priority_max.2 sched_rr_get_interval.2 MLINKS+=sched_setparam.2 sched_getparam.2 MLINKS+=sched_setscheduler.2 sched_getscheduler.2 .endif MLINKS+=select.2 FD_CLR.3 \ select.2 FD_ISSET.3 \ select.2 FD_SET.3 \ select.2 FD_ZERO.3 MLINKS+=send.2 sendmsg.2 \ send.2 sendto.2 MLINKS+=setpgid.2 setpgrp.2 MLINKS+=setresuid.2 getresgid.2 \ setresuid.2 getresuid.2 \ setresuid.2 setresgid.2 MLINKS+=setuid.2 setegid.2 \ setuid.2 seteuid.2 \ setuid.2 setgid.2 MLINKS+=shmat.2 shmdt.2 MLINKS+=shm_open.2 shm_unlink.2 MLINKS+=sigwaitinfo.2 sigtimedwait.2 MLINKS+=stat.2 fstat.2 \ stat.2 fstatat.2 \ stat.2 lstat.2 MLINKS+=statfs.2 fstatfs.2 MLINKS+=swapon.2 swapoff.2 MLINKS+=symlink.2 symlinkat.2 MLINKS+=syscall.2 __syscall.2 MLINKS+=timer_settime.2 timer_getoverrun.2 \ timer_settime.2 timer_gettime.2 MLINKS+=thr_kill.2 thr_kill2.2 MLINKS+=truncate.2 ftruncate.2 MLINKS+=unlink.2 unlinkat.2 +MLINKS+=unlink.2 funlinkat.2 MLINKS+=utimensat.2 futimens.2 MLINKS+=utimes.2 futimes.2 \ utimes.2 futimesat.2 \ utimes.2 lutimes.2 MLINKS+=wait.2 wait3.2 \ wait.2 wait4.2 \ wait.2 waitpid.2 \ wait.2 waitid.2 \ wait.2 wait6.2 MLINKS+=write.2 pwrite.2 \ write.2 pwritev.2 \ write.2 writev.2 Index: head/lib/libc/sys/Symbol.map =================================================================== --- head/lib/libc/sys/Symbol.map (revision 345981) +++ head/lib/libc/sys/Symbol.map (revision 345982) @@ -1,1031 +1,1032 @@ /* * $FreeBSD$ */ /* * It'd be nice to automatically generate the syscall symbols, but we * don't know to what version they will eventually belong to, so for now * it has to be manual. */ FBSD_1.0 { __acl_aclcheck_fd; __acl_aclcheck_file; __acl_aclcheck_link; __acl_delete_fd; __acl_delete_file; __acl_delete_link; __acl_get_fd; __acl_get_file; __acl_get_link; __acl_set_fd; __acl_set_file; __acl_set_link; __getcwd; __mac_execve; __mac_get_fd; __mac_get_file; __mac_get_link; __mac_get_pid; __mac_get_proc; __mac_set_fd; __mac_set_file; __mac_set_link; __mac_set_proc; __setugid; __syscall; __sysctl; _umtx_op; abort2; accept; access; acct; adjtime; aio_cancel; aio_error; aio_fsync; aio_read; aio_return; aio_suspend; aio_waitcomplete; aio_write; audit; auditctl; auditon; bind; chdir; chflags; chmod; chown; chroot; clock_getres; clock_gettime; clock_settime; close; connect; dup; dup2; eaccess; execve; extattr_delete_fd; extattr_delete_file; extattr_delete_link; extattr_get_fd; extattr_get_file; extattr_get_link; extattr_list_fd; extattr_list_file; extattr_list_link; extattr_set_fd; extattr_set_file; extattr_set_link; extattrctl; fchdir; fchflags; fchmod; fchown; fcntl; fhopen; flock; fork; fpathconf; fsync; futimes; getaudit; getaudit_addr; getauid; getcontext; getdtablesize; getegid; geteuid; getfh; getgid; getgroups; getitimer; getpeername; getpgid; getpgrp; getpid; getppid; getpriority; getresgid; getresuid; getrlimit; getrusage; getsid; getsockname; getsockopt; gettimeofday; getuid; ioctl; issetugid; jail; jail_attach; kenv; kill; kldfind; kldfirstmod; kldload; kldnext; kldstat; kldsym; kldunload; kldunloadf; kqueue; kmq_notify; /* Do we want these to be public interfaces? */ kmq_open; /* librt uses them to provide mq_xxx. */ kmq_setattr; kmq_timedreceive; kmq_timedsend; kmq_unlink; ksem_close; ksem_destroy; ksem_getvalue; ksem_init; ksem_open; ksem_post; ksem_timedwait; ksem_trywait; ksem_unlink; ksem_wait; ktrace; lchflags; lchmod; lchown; lgetfh; link; lio_listio; listen; lutimes; mac_syscall; madvise; mincore; minherit; mkdir; mkfifo; mlock; mlockall; modfind; modfnext; modnext; modstat; mount; mprotect; msgget; msgrcv; msgsnd; msgsys; msync; munlock; munlockall; munmap; nanosleep; nfssvc; nmount; ntp_adjtime; ntp_gettime; open; pathconf; pipe; poll; posix_openpt; preadv; profil; pselect; ptrace; pwritev; quotactl; read; readlink; readv; reboot; recvfrom; recvmsg; rename; revoke; rfork; rmdir; rtprio; rtprio_thread; sched_get_priority_max; sched_get_priority_min; sched_getparam; sched_getscheduler; sched_rr_get_interval; sched_setparam; sched_setscheduler; sched_yield; select; semget; semop; semsys; sendfile; sendmsg; sendto; setaudit; setaudit_addr; setauid; setegid; seteuid; setgid; setgroups; setitimer; setlogin; setpgid; setpriority; setregid; setresgid; setresuid; setreuid; setrlimit; setsid; setsockopt; settimeofday; setuid; shm_open; shm_unlink; shmat; shmdt; shmget; shmsys; shutdown; sigaction; sigaltstack; sigpending; sigprocmask; sigqueue; sigreturn; sigsuspend; sigtimedwait; sigwait; sigwaitinfo; socket; socketpair; swapoff; swapon; symlink; sync; sysarch; syscall; thr_create; thr_exit; thr_kill; thr_kill2; thr_new; thr_self; thr_set_name; thr_suspend; thr_wake; ktimer_create; /* Do we want these to be public interfaces? */ ktimer_delete; /* librt uses them to provide timer_xxx. */ ktimer_getoverrun; ktimer_gettime; ktimer_settime; umask; undelete; unlink; unmount; utimes; utrace; uuidgen; vadvise; wait4; write; writev; __error; ftruncate; lseek; mmap; pread; pwrite; truncate; }; FBSD_1.1 { __semctl; closefrom; cpuset; cpuset_getid; cpuset_setid; cpuset_getaffinity; cpuset_setaffinity; faccessat; fchmodat; fchownat; fexecve; futimesat; jail_get; jail_set; jail_remove; linkat; lpathconf; mkdirat; mkfifoat; msgctl; readlinkat; renameat; setfib; shmctl; symlinkat; unlinkat; }; FBSD_1.2 { cap_enter; cap_getmode; getloginclass; pdfork; pdgetpid; pdkill; posix_fallocate; rctl_get_racct; rctl_get_rules; rctl_get_limits; rctl_add_rule; rctl_remove_rule; setloginclass; }; FBSD_1.3 { accept4; aio_mlock; bindat; cap_fcntls_get; cap_fcntls_limit; cap_ioctls_get; cap_ioctls_limit; __cap_rights_get; cap_rights_limit; cap_sandboxed; chflagsat; clock_getcpuclockid2; connectat; ffclock_getcounter; ffclock_getestimate; ffclock_setestimate; pipe2; posix_fadvise; procctl; wait6; }; FBSD_1.4 { futimens; ppoll; utimensat; numa_setaffinity; numa_getaffinity; sendmmsg; recvmmsg; }; FBSD_1.5 { clock_nanosleep; fdatasync; fhstat; fhstatfs; fstat; fstatat; fstatfs; getdents; getdirentries; getfsstat; getrandom; kevent; lstat; mknod; mknodat; stat; statfs; cpuset_getdomain; cpuset_setdomain; }; FBSD_1.6 { fhlink; fhlinkat; fhreadlink; getfhat; + funlinkat; }; FBSDprivate_1.0 { ___acl_aclcheck_fd; __sys___acl_aclcheck_fd; ___acl_aclcheck_file; __sys___acl_aclcheck_file; ___acl_aclcheck_link; __sys___acl_aclcheck_link; ___acl_delete_fd; __sys___acl_delete_fd; ___acl_delete_file; __sys___acl_delete_file; ___acl_delete_link; __sys___acl_delete_link; ___acl_get_fd; __sys___acl_get_fd; ___acl_get_file; __sys___acl_get_file; ___acl_get_link; __sys___acl_get_link; ___acl_set_fd; __sys___acl_set_fd; ___acl_set_file; __sys___acl_set_file; ___acl_set_link; __sys___acl_set_link; ___getcwd; __sys___getcwd; ___mac_execve; __sys___mac_execve; ___mac_get_fd; __sys___mac_get_fd; ___mac_get_file; __sys___mac_get_file; ___mac_get_link; __sys___mac_get_link; ___mac_get_pid; __sys___mac_get_pid; ___mac_get_proc; __sys___mac_get_proc; ___mac_set_fd; __sys___mac_set_fd; ___mac_set_file; __sys___mac_set_file; ___mac_set_link; __sys___mac_set_link; ___mac_set_proc; __sys___mac_set_proc; ___semctl; __sys___semctl; ___setugid; __sys___setugid; ___syscall; __sys___syscall; ___sysctl; __sys___sysctl; __umtx_op; __sys__umtx_op; _abort2; __sys_abort2; _accept; __sys_accept; _accept4; __sys_accept4; _access; __sys_access; _acct; __sys_acct; _adjtime; __sys_adjtime; __sys_aio_cancel; __sys_aio_error; __sys_aio_fsync; __sys_aio_read; __sys_aio_return; __sys_aio_suspend; __sys_aio_waitcomplete; __sys_aio_write; _audit; __sys_audit; _auditctl; __sys_auditctl; _auditon; __sys_auditon; _bind; __sys_bind; _chdir; __sys_chdir; _chflags; __sys_chflags; _chmod; __sys_chmod; _chown; __sys_chown; _chroot; __sys_chroot; _clock_getcpuclockid2; __sys_clock_getcpuclockid2; _clock_getres; __sys_clock_getres; _clock_gettime; __sys_clock_gettime; __sys_clock_nanosleep; _clock_settime; __sys_clock_settime; _close; __sys_close; _closefrom; __sys_closefrom; _connect; __sys_connect; _cpuset; __sys_cpuset; _cpuset_getid; __sys_cpuset_getid; _cpuset_setid; __sys_cpuset_setid; _cpuset_getaffinity; __sys_cpuset_getaffinity; _cpuset_setaffinity; __sys_cpuset_setaffinity; _dup; __sys_dup; _dup2; __sys_dup2; _eaccess; __sys_eaccess; _execve; __sys_execve; _extattr_delete_fd; __sys_extattr_delete_fd; _extattr_delete_file; __sys_extattr_delete_file; _extattr_delete_link; __sys_extattr_delete_link; _extattr_get_fd; __sys_extattr_get_fd; _extattr_get_file; __sys_extattr_get_file; _extattr_get_link; __sys_extattr_get_link; _extattr_list_fd; __sys_extattr_list_fd; _extattr_list_file; __sys_extattr_list_file; _extattr_list_link; __sys_extattr_list_link; _extattr_set_fd; __sys_extattr_set_fd; _extattr_set_file; __sys_extattr_set_file; _extattr_set_link; __sys_extattr_set_link; _extattrctl; __sys_extattrctl; _fchdir; __sys_fchdir; _fchflags; __sys_fchflags; _fchmod; __sys_fchmod; _fchown; __sys_fchown; _fcntl; __sys_fcntl; __fcntl_compat; _fhopen; __sys_fhopen; _fhstat; __sys_fhstat; _fhstatfs; __sys_fhstatfs; _flock; __sys_flock; _fork; __sys_fork; _fpathconf; __sys_fpathconf; _fstat; __sys_fstat; _fstatfs; __sys_fstatfs; _fsync; __sys_fsync; _fdatasync; __sys_fdatasync; _futimes; __sys_futimes; _getaudit; __sys_getaudit; _getaudit_addr; __sys_getaudit_addr; _getauid; __sys_getauid; _getcontext; __sys_getcontext; _getdirentries; __sys_getdirentries; _getdtablesize; __sys_getdtablesize; _getegid; __sys_getegid; _geteuid; __sys_geteuid; _getfh; __sys_getfh; _getfsstat; __sys_getfsstat; _getgid; __sys_getgid; _getgroups; __sys_getgroups; _getitimer; __sys_getitimer; _getpeername; __sys_getpeername; _getpgid; __sys_getpgid; _getpgrp; __sys_getpgrp; _getpid; __sys_getpid; _getppid; __sys_getppid; _getpriority; __sys_getpriority; _getresgid; __sys_getresgid; _getresuid; __sys_getresuid; _getrlimit; __sys_getrlimit; _getrusage; __sys_getrusage; _getsid; __sys_getsid; _getsockname; __sys_getsockname; _getsockopt; __sys_getsockopt; _gettimeofday; __sys_gettimeofday; _getuid; __sys_getuid; _ioctl; __sys_ioctl; _issetugid; __sys_issetugid; _jail; __sys_jail; _jail_attach; __sys_jail_attach; _kenv; __sys_kenv; _kevent; __sys_kevent; _kill; __sys_kill; _kldfind; __sys_kldfind; _kldfirstmod; __sys_kldfirstmod; _kldload; __sys_kldload; _kldnext; __sys_kldnext; _kldstat; __sys_kldstat; _kldsym; __sys_kldsym; _kldunload; __sys_kldunload; _kldunloadf; __sys_kldunloadf; _kmq_notify; __sys_kmq_notify; _kmq_open; __sys_kmq_open; _kmq_setattr; __sys_kmq_setattr; _kmq_timedreceive; __sys_kmq_timedreceive; _kmq_timedsend; __sys_kmq_timedsend; _kmq_unlink; __sys_kmq_unlink; _kqueue; __sys_kqueue; _ksem_close; __sys_ksem_close; _ksem_destroy; __sys_ksem_destroy; _ksem_getvalue; __sys_ksem_getvalue; _ksem_init; __sys_ksem_init; _ksem_open; __sys_ksem_open; _ksem_post; __sys_ksem_post; _ksem_timedwait; __sys_ksem_timedwait; _ksem_trywait; __sys_ksem_trywait; _ksem_unlink; __sys_ksem_unlink; _ksem_wait; __sys_ksem_wait; _ktrace; __sys_ktrace; _lchflags; __sys_lchflags; _lchmod; __sys_lchmod; _lchown; __sys_lchown; _lgetfh; __sys_lgetfh; _link; __sys_link; __sys_lio_listio; _listen; __sys_listen; _lutimes; __sys_lutimes; _mac_syscall; __sys_mac_syscall; _madvise; __sys_madvise; _mincore; __sys_mincore; _minherit; __sys_minherit; _mkdir; __sys_mkdir; _mkfifo; __sys_mkfifo; _mknod; __sys_mknod; _mlock; __sys_mlock; _mlockall; __sys_mlockall; _modfind; __sys_modfind; _modfnext; __sys_modfnext; _modnext; __sys_modnext; _modstat; __sys_modstat; _mount; __sys_mount; _mprotect; __sys_mprotect; _msgctl; __sys_msgctl; _msgget; __sys_msgget; _msgrcv; __sys_msgrcv; _msgsnd; __sys_msgsnd; _msgsys; __sys_msgsys; _msync; __sys_msync; _munlock; __sys_munlock; _munlockall; __sys_munlockall; _munmap; __sys_munmap; _nanosleep; __sys_nanosleep; _nfssvc; __sys_nfssvc; _nmount; __sys_nmount; _ntp_adjtime; __sys_ntp_adjtime; _ntp_gettime; __sys_ntp_gettime; _open; __sys_open; _openat; __sys_openat; _pathconf; __sys_pathconf; _pipe; __sys_pipe; _poll; __sys_poll; _ppoll; __sys_ppoll; _preadv; __sys_preadv; _procctl; __sys_procctl; _profil; __sys_profil; _pselect; __sys_pselect; _ptrace; __sys_ptrace; _pwritev; __sys_pwritev; _quotactl; __sys_quotactl; _read; __sys_read; _readlink; __sys_readlink; _readv; __sys_readv; _reboot; __sys_reboot; _recvfrom; __sys_recvfrom; _recvmsg; __sys_recvmsg; _rename; __sys_rename; _revoke; __sys_revoke; _rfork; __sys_rfork; _rmdir; __sys_rmdir; _rtprio; __sys_rtprio; _rtprio_thread; __sys_rtprio_thread; _sched_get_priority_max; __sys_sched_get_priority_max; _sched_get_priority_min; __sys_sched_get_priority_min; _sched_getparam; __sys_sched_getparam; _sched_getscheduler; __sys_sched_getscheduler; _sched_rr_get_interval; __sys_sched_rr_get_interval; _sched_setparam; __sys_sched_setparam; _sched_setscheduler; __sys_sched_setscheduler; _sched_yield; __sys_sched_yield; _select; __sys_select; _semget; __sys_semget; _semop; __sys_semop; _semsys; __sys_semsys; _sendfile; __sys_sendfile; _sendmsg; __sys_sendmsg; _sendto; __sys_sendto; _setaudit; __sys_setaudit; _setaudit_addr; __sys_setaudit_addr; _setauid; __sys_setauid; _setcontext; __sys_setcontext; _setegid; __sys_setegid; _seteuid; __sys_seteuid; _setgid; __sys_setgid; _setgroups; __sys_setgroups; _setitimer; __sys_setitimer; _setlogin; __sys_setlogin; _setpgid; __sys_setpgid; _setpriority; __sys_setpriority; _setregid; __sys_setregid; _setresgid; __sys_setresgid; _setresuid; __sys_setresuid; _setreuid; __sys_setreuid; _setrlimit; __sys_setrlimit; _setsid; __sys_setsid; _setsockopt; __sys_setsockopt; _settimeofday; __sys_settimeofday; _setuid; __sys_setuid; _shm_open; __sys_shm_open; _shm_unlink; __sys_shm_unlink; _shmat; __sys_shmat; _shmctl; __sys_shmctl; _shmdt; __sys_shmdt; _shmget; __sys_shmget; _shmsys; __sys_shmsys; _shutdown; __sys_shutdown; _sigaction; __sys_sigaction; _sigaltstack; __sys_sigaltstack; _sigpending; __sys_sigpending; _sigprocmask; __sys_sigprocmask; _sigqueue; __sys_sigqueue; _sigreturn; __sys_sigreturn; _sigsuspend; __sys_sigsuspend; _sigtimedwait; __sys_sigtimedwait; _sigwait; __sigwait; __sys_sigwait; _sigwaitinfo; __sys_sigwaitinfo; _socket; __sys_socket; _socketpair; __sys_socketpair; _statfs; __sys_statfs; _swapcontext; __sys_swapcontext; _swapoff; __sys_swapoff; _swapon; __sys_swapon; _symlink; __sys_symlink; _sync; __sys_sync; _sysarch; __sys_sysarch; _syscall; __sys_syscall; _thr_create; __sys_thr_create; _thr_exit; __sys_thr_exit; _thr_kill; __sys_thr_kill; _thr_kill2; __sys_thr_kill2; _thr_new; __sys_thr_new; _thr_self; __sys_thr_self; _thr_set_name; __sys_thr_set_name; _thr_suspend; __sys_thr_suspend; _thr_wake; __sys_thr_wake; _ktimer_create; __sys_ktimer_create; _ktimer_delete; __sys_ktimer_delete; _ktimer_getoverrun; __sys_ktimer_getoverrun; _ktimer_gettime; __sys_ktimer_gettime; _ktimer_settime; __sys_ktimer_settime; _umask; __sys_umask; _undelete; __sys_undelete; _unlink; __sys_unlink; _unmount; __sys_unmount; _utimes; __sys_utimes; _utrace; __sys_utrace; _uuidgen; __sys_uuidgen; _wait4; __sys_wait4; _wait6; __sys_wait6; _write; __sys_write; _writev; __sys_writev; __set_error_selector; nlm_syscall; gssd_syscall; __libc_interposing_slot; __libc_sigwait; _cpuset_getdomain; __sys_cpuset_getdomain; _cpuset_setdomain; __sys_cpuset_setdomain; }; Index: head/lib/libc/sys/unlink.2 =================================================================== --- head/lib/libc/sys/unlink.2 (revision 345981) +++ head/lib/libc/sys/unlink.2 (revision 345982) @@ -1,254 +1,289 @@ .\" Copyright (c) 1980, 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. .\" 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. .\" .\" @(#)unlink.2 8.1 (Berkeley) 6/4/93 .\" $FreeBSD$ .\" -.Dd November 11, 2018 +.Dd April 6, 2019 .Dt UNLINK 2 .Os .Sh NAME .Nm unlink , .Nm unlinkat .Nd remove directory entry .Sh LIBRARY .Lb libc .Sh SYNOPSIS .In unistd.h .Ft int .Fn unlink "const char *path" .Ft int -.Fn unlinkat "int fd" "const char *path" "int flag" +.Fn unlinkat "int dfd" "const char *path" "int flag" +.Ft int +.Fn funlinkat "int dfd" "const char *path" "int fd" "int flag" .Sh DESCRIPTION The .Fn unlink system call removes the link named by .Fa path from its directory and decrements the link count of the file which was referenced by the link. If that decrement reduces the link count of the file to zero, and no process has the file open, then all resources associated with the file are reclaimed. If one or more process have the file open when the last link is removed, the link is removed, but the removal of the file is delayed until all references to it have been closed. The .Fa path argument may not be a directory. .Pp The .Fn unlinkat system call is equivalent to .Fn unlink or .Fn rmdir except in the case where .Fa path specifies a relative path. In this case the directory entry to be removed is determined relative to the directory associated with the file descriptor -.Fa fd +.Fa dfd instead of the current working directory. .Pp The values for .Fa flag are constructed by a bitwise-inclusive OR of flags from the following list, defined in .In fcntl.h : .Bl -tag -width indent .It Dv AT_REMOVEDIR Remove the directory entry specified by .Fa fd and .Fa path as a directory, not a normal file. .It Dv AT_BENEATH Only unlink files and directories which are beneath of the topping directory. See the description of the .Dv O_BENEATH flag in the .Xr open 2 manual page. .El .Pp If .Fn unlinkat is passed the special value .Dv AT_FDCWD in the .Fa fd parameter, the current working directory is used and the behavior is identical to a call to .Fa unlink or .Fa rmdir respectively, depending on whether or not the .Dv AT_REMOVEDIR bit is set in flag. +.Pp +The +.Fn funlinkat +system call can be used to unlink an already-opened file, unless that +file has been replaced since it was opened. +It is equivalent to +.Fn unlinkat +in the case where +.Fa path +is already open as the file descriptor +.Fa fd . +Otherwise, the path will not be removed and an error will be returned. +The +.Fa fd +can be set the +.Dv FD_NONE . +In that case +.Fn funlinkat +behaves exactly like +.Fn unlinkat . .Sh RETURN VALUES .Rv -std unlink .Sh ERRORS The .Fn unlink succeeds unless: .Bl -tag -width Er .It Bq Er ENOTDIR A component of the path prefix is not a directory. .It Bq Er EISDIR The named file is a directory. .It Bq Er ENAMETOOLONG A component of a pathname exceeded 255 characters, or an entire path name exceeded 1023 characters. .It Bq Er ENOENT The named file does not exist. .It Bq Er EACCES Search permission is denied for a component of the path prefix. .It Bq Er EACCES Write permission is denied on the directory containing the link to be removed. .It Bq Er ELOOP Too many symbolic links were encountered in translating the pathname. .It Bq Er EPERM The named file is a directory. .It Bq Er EPERM The named file has its immutable, undeletable or append-only flag set, see the .Xr chflags 2 manual page for more information. .It Bq Er EPERM The parent directory of the named file has its immutable or append-only flag set. .It Bq Er EPERM The directory containing the file is marked sticky, and neither the containing directory nor the file to be removed are owned by the effective user ID. .It Bq Er EIO An I/O error occurred while deleting the directory entry or deallocating the inode. .It Bq Er EROFS The named file resides on a read-only file system. .It Bq Er EFAULT The .Fa path argument points outside the process's allocated address space. .It Bq Er ENOSPC On file systems supporting copy-on-write or snapshots, there was not enough free space to record metadata for the delete operation of the file. .El .Pp In addition to the errors returned by the .Fn unlink , the .Fn unlinkat may fail if: .Bl -tag -width Er .It Bq Er EBADF The .Fa path argument does not specify an absolute path and the .Fa fd argument is neither .Dv AT_FDCWD nor a valid file descriptor open for searching. .It Bq Er ENOTEMPTY The .Fa flag parameter has the .Dv AT_REMOVEDIR bit set and the .Fa path argument names a directory that is not an empty directory, or there are hard links to the directory other than dot or a single entry in dot-dot. .It Bq Er ENOTDIR The .Fa flag parameter has the .Dv AT_REMOVEDIR bit set and .Fa path does not name a directory. .It Bq Er EINVAL The value of the .Fa flag argument is not valid. .It Bq Er ENOTDIR The .Fa path argument is not an absolute path and .Fa fd is neither .Dv AT_FDCWD nor a file descriptor associated with a directory. .It Bq Er ENOTCAPABLE .Fa path is an absolute path, or contained a ".." component leading to a directory outside of the directory hierarchy specified by .Fa fd , and the process is in capability mode. .It Bq Er ENOTCAPABLE The .Dv AT_BENEATH flag was provided to .Fn unlinkat , and the absolute .Fa path does not have its tail fully contained under the topping directory, or the relative .Fa path escapes it. .El +.Pp +In addition to the errors returned by +.Fn unlinkat , +.Fn funlinkat +may fail if: +.Bl -tag -width Er +.It Bq Er EDEADLK +The file descriptor is not associated with the path. +.El .Sh SEE ALSO .Xr chflags 2 , .Xr close 2 , .Xr link 2 , .Xr rmdir 2 , .Xr symlink 7 .Sh STANDARDS The .Fn unlinkat system call follows The Open Group Extended API Set 2 specification. .Sh HISTORY The .Fn unlink function appeared in .At v1 . The .Fn unlinkat system call appeared in .Fx 8.0 . +The +.Fn funlinkat +system call appeared in +.Fx 13.0 . .Pp The .Fn unlink system call traditionally allows the super-user to unlink directories which can damage the file system integrity. This implementation no longer permits it. Index: head/sys/cddl/compat/opensolaris/sys/vnode.h =================================================================== --- head/sys/cddl/compat/opensolaris/sys/vnode.h (revision 345981) +++ head/sys/cddl/compat/opensolaris/sys/vnode.h (revision 345982) @@ -1,286 +1,287 @@ /*- * Copyright (c) 2007 Pawel Jakub Dawidek * 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 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 AUTHORS 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 _OPENSOLARIS_SYS_VNODE_H_ #define _OPENSOLARIS_SYS_VNODE_H_ #ifdef _KERNEL struct vnode; struct vattr; typedef struct vnode vnode_t; typedef struct vattr vattr_t; typedef enum vtype vtype_t; #include enum symfollow { NO_FOLLOW = NOFOLLOW }; #include #include_next #include #include #include #include #include #include typedef struct vop_vector vnodeops_t; #define VOP_FID VOP_VPTOFH #define vop_fid vop_vptofh #define vop_fid_args vop_vptofh_args #define a_fid a_fhp #define IS_XATTRDIR(dvp) (0) #define v_count v_usecount #define V_APPEND VAPPEND #define rootvfs (rootvnode == NULL ? NULL : rootvnode->v_mount) static __inline int vn_is_readonly(vnode_t *vp) { return (vp->v_mount->mnt_flag & MNT_RDONLY); } #define vn_vfswlock(vp) (0) #define vn_vfsunlock(vp) do { } while (0) #define vn_ismntpt(vp) ((vp)->v_type == VDIR && (vp)->v_mountedhere != NULL) #define vn_mountedvfs(vp) ((vp)->v_mountedhere) #define vn_has_cached_data(vp) \ ((vp)->v_object != NULL && \ (vp)->v_object->resident_page_count > 0) #define vn_exists(vp) do { } while (0) #define vn_invalid(vp) do { } while (0) #define vn_renamepath(tdvp, svp, tnm, lentnm) do { } while (0) #define vn_free(vp) do { } while (0) #define vn_matchops(vp, vops) ((vp)->v_op == &(vops)) #define VN_HOLD(v) vref(v) #define VN_RELE(v) vrele(v) #define VN_URELE(v) vput(v) #define vnevent_create(vp, ct) do { } while (0) #define vnevent_link(vp, ct) do { } while (0) #define vnevent_remove(vp, dvp, name, ct) do { } while (0) #define vnevent_rmdir(vp, dvp, name, ct) do { } while (0) #define vnevent_rename_src(vp, dvp, name, ct) do { } while (0) #define vnevent_rename_dest(vp, dvp, name, ct) do { } while (0) #define vnevent_rename_dest_dir(vp, ct) do { } while (0) #define specvp(vp, rdev, type, cr) (VN_HOLD(vp), (vp)) #define MANDMODE(mode) (0) #define MANDLOCK(vp, mode) (0) #define chklock(vp, op, offset, size, mode, ct) (0) #define cleanlocks(vp, pid, foo) do { } while (0) #define cleanshares(vp, pid) do { } while (0) /* * We will use va_spare is place of Solaris' va_mask. * This field is initialized in zfs_setattr(). */ #define va_mask va_spare /* TODO: va_fileid is shorter than va_nodeid !!! */ #define va_nodeid va_fileid /* TODO: This field needs conversion! */ #define va_nblocks va_bytes #define va_blksize va_blocksize #define va_seq va_gen #define MAXOFFSET_T OFF_MAX #define EXCL 0 #define ACCESSED (AT_ATIME) #define STATE_CHANGED (AT_CTIME) #define CONTENT_MODIFIED (AT_MTIME | AT_CTIME) static __inline void vattr_init_mask(vattr_t *vap) { vap->va_mask = 0; if (vap->va_type != VNON) vap->va_mask |= AT_TYPE; if (vap->va_uid != (uid_t)VNOVAL) vap->va_mask |= AT_UID; if (vap->va_gid != (gid_t)VNOVAL) vap->va_mask |= AT_GID; if (vap->va_size != (u_quad_t)VNOVAL) vap->va_mask |= AT_SIZE; if (vap->va_atime.tv_sec != VNOVAL) vap->va_mask |= AT_ATIME; if (vap->va_mtime.tv_sec != VNOVAL) vap->va_mask |= AT_MTIME; if (vap->va_mode != (u_short)VNOVAL) vap->va_mask |= AT_MODE; if (vap->va_flags != VNOVAL) vap->va_mask |= AT_XVATTR; } #define FCREAT O_CREAT #define FTRUNC O_TRUNC #define FEXCL O_EXCL #define FDSYNC FFSYNC #define FRSYNC FFSYNC #define FSYNC FFSYNC #define FOFFMAX 0x00 #define FIGNORECASE 0x00 static __inline int vn_openat(char *pnamep, enum uio_seg seg, int filemode, int createmode, vnode_t **vpp, enum create crwhy, mode_t umask, struct vnode *startvp, int fd) { struct thread *td = curthread; struct nameidata nd; int error, operation; ASSERT(seg == UIO_SYSSPACE); if ((filemode & FCREAT) != 0) { ASSERT(filemode == (FWRITE | FCREAT | FTRUNC | FOFFMAX)); ASSERT(crwhy == CRCREAT); operation = CREATE; } else { ASSERT(filemode == (FREAD | FOFFMAX) || filemode == (FREAD | FWRITE | FOFFMAX)); ASSERT(crwhy == 0); operation = LOOKUP; } ASSERT(umask == 0); pwd_ensure_dirs(); if (startvp != NULL) vref(startvp); NDINIT_ATVP(&nd, operation, 0, UIO_SYSSPACE, pnamep, startvp, td); filemode |= O_NOFOLLOW; error = vn_open_cred(&nd, &filemode, createmode, 0, td->td_ucred, NULL); NDFREE(&nd, NDF_ONLY_PNBUF); if (error == 0) { /* We just unlock so we hold a reference. */ VOP_UNLOCK(nd.ni_vp, 0); *vpp = nd.ni_vp; } return (error); } static __inline int zfs_vn_open(char *pnamep, enum uio_seg seg, int filemode, int createmode, vnode_t **vpp, enum create crwhy, mode_t umask) { return (vn_openat(pnamep, seg, filemode, createmode, vpp, crwhy, umask, NULL, -1)); } #define vn_open(pnamep, seg, filemode, createmode, vpp, crwhy, umask) \ zfs_vn_open((pnamep), (seg), (filemode), (createmode), (vpp), (crwhy), (umask)) #define RLIM64_INFINITY 0 static __inline int zfs_vn_rdwr(enum uio_rw rw, vnode_t *vp, caddr_t base, ssize_t len, offset_t offset, enum uio_seg seg, int ioflag, int ulimit, cred_t *cr, ssize_t *residp) { struct thread *td = curthread; int error; ssize_t resid; ASSERT(ioflag == 0); ASSERT(ulimit == RLIM64_INFINITY); if (rw == UIO_WRITE) { ioflag = IO_SYNC; } else { ioflag = IO_DIRECT; } error = vn_rdwr(rw, vp, base, len, offset, seg, ioflag, cr, NOCRED, &resid, td); if (residp != NULL) *residp = (ssize_t)resid; return (error); } #define vn_rdwr(rw, vp, base, len, offset, seg, ioflag, ulimit, cr, residp) \ zfs_vn_rdwr((rw), (vp), (base), (len), (offset), (seg), (ioflag), (ulimit), (cr), (residp)) static __inline int zfs_vop_fsync(vnode_t *vp, int flag, cred_t *cr) { struct mount *mp; int error; ASSERT(flag == FSYNC); if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) goto drop; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); error = VOP_FSYNC(vp, MNT_WAIT, curthread); VOP_UNLOCK(vp, 0); vn_finished_write(mp); drop: return (error); } #define VOP_FSYNC(vp, flag, cr, ct) zfs_vop_fsync((vp), (flag), (cr)) static __inline int zfs_vop_close(vnode_t *vp, int flag, int count, offset_t offset, cred_t *cr) { int error; ASSERT(count == 1); ASSERT(offset == 0); error = vn_close(vp, flag, cr, curthread); return (error); } #define VOP_CLOSE(vp, oflags, count, offset, cr, ct) \ zfs_vop_close((vp), (oflags), (count), (offset), (cr)) static __inline int vn_rename(char *from, char *to, enum uio_seg seg) { ASSERT(seg == UIO_SYSSPACE); return (kern_renameat(curthread, AT_FDCWD, from, AT_FDCWD, to, seg)); } static __inline int vn_remove(char *fnamep, enum uio_seg seg, enum rm dirflag) { ASSERT(seg == UIO_SYSSPACE); ASSERT(dirflag == RMFILE); - return (kern_unlinkat(curthread, AT_FDCWD, fnamep, seg, 0, 0)); + return (kern_funlinkat(curthread, AT_FDCWD, fnamep, FD_NONE, seg, 0, + 0)); } #endif /* _KERNEL */ #endif /* _OPENSOLARIS_SYS_VNODE_H_ */ Index: head/sys/compat/cloudabi/cloudabi_file.c =================================================================== --- head/sys/compat/cloudabi/cloudabi_file.c (revision 345981) +++ head/sys/compat/cloudabi/cloudabi_file.c (revision 345982) @@ -1,760 +1,762 @@ /*- * Copyright (c) 2015 Nuxi, https://nuxi.nl/ * * 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 static MALLOC_DEFINE(M_CLOUDABI_PATH, "cloudabipath", "CloudABI pathnames"); /* * Copying pathnames from userspace to kernelspace. * * Unlike most operating systems, CloudABI doesn't use null-terminated * pathname strings. Processes always pass pathnames to the kernel by * providing a base pointer and a length. This has a couple of reasons: * * - It makes it easier to use CloudABI in combination with programming * languages other than C, that may use non-null terminated strings. * - It allows for calling system calls on individual components of the * pathname without modifying the input string. * * The function below copies in pathname strings and null-terminates it. * It also ensure that the string itself does not contain any null * bytes. * * TODO(ed): Add an abstraction to vfs_lookup.c that allows us to pass * in unterminated pathname strings, so we can do away with * the copying. */ static int copyin_path(const char *uaddr, size_t len, char **result) { char *buf; int error; if (len >= PATH_MAX) return (ENAMETOOLONG); buf = malloc(len + 1, M_CLOUDABI_PATH, M_WAITOK); error = copyin(uaddr, buf, len); if (error != 0) { free(buf, M_CLOUDABI_PATH); return (error); } if (memchr(buf, '\0', len) != NULL) { free(buf, M_CLOUDABI_PATH); return (EINVAL); } buf[len] = '\0'; *result = buf; return (0); } static void cloudabi_freestr(char *buf) { free(buf, M_CLOUDABI_PATH); } int cloudabi_sys_file_advise(struct thread *td, struct cloudabi_sys_file_advise_args *uap) { int advice; switch (uap->advice) { case CLOUDABI_ADVICE_DONTNEED: advice = POSIX_FADV_DONTNEED; break; case CLOUDABI_ADVICE_NOREUSE: advice = POSIX_FADV_NOREUSE; break; case CLOUDABI_ADVICE_NORMAL: advice = POSIX_FADV_NORMAL; break; case CLOUDABI_ADVICE_RANDOM: advice = POSIX_FADV_RANDOM; break; case CLOUDABI_ADVICE_SEQUENTIAL: advice = POSIX_FADV_SEQUENTIAL; break; case CLOUDABI_ADVICE_WILLNEED: advice = POSIX_FADV_WILLNEED; break; default: return (EINVAL); } return (kern_posix_fadvise(td, uap->fd, uap->offset, uap->len, advice)); } int cloudabi_sys_file_allocate(struct thread *td, struct cloudabi_sys_file_allocate_args *uap) { return (kern_posix_fallocate(td, uap->fd, uap->offset, uap->len)); } int cloudabi_sys_file_create(struct thread *td, struct cloudabi_sys_file_create_args *uap) { char *path; int error; error = copyin_path(uap->path, uap->path_len, &path); if (error != 0) return (error); /* * CloudABI processes cannot interact with UNIX credentials and * permissions. Depend on the umask that is set prior to * execution to restrict the file permissions. */ switch (uap->type) { case CLOUDABI_FILETYPE_DIRECTORY: error = kern_mkdirat(td, uap->fd, path, UIO_SYSSPACE, 0777); break; default: error = EINVAL; break; } cloudabi_freestr(path); return (error); } int cloudabi_sys_file_link(struct thread *td, struct cloudabi_sys_file_link_args *uap) { char *path1, *path2; int error; error = copyin_path(uap->path1, uap->path1_len, &path1); if (error != 0) return (error); error = copyin_path(uap->path2, uap->path2_len, &path2); if (error != 0) { cloudabi_freestr(path1); return (error); } error = kern_linkat(td, uap->fd1.fd, uap->fd2, path1, path2, UIO_SYSSPACE, (uap->fd1.flags & CLOUDABI_LOOKUP_SYMLINK_FOLLOW) ? FOLLOW : NOFOLLOW); cloudabi_freestr(path1); cloudabi_freestr(path2); return (error); } int cloudabi_sys_file_open(struct thread *td, struct cloudabi_sys_file_open_args *uap) { cloudabi_fdstat_t fds; cap_rights_t rights; struct filecaps fcaps = {}; struct nameidata nd; struct file *fp; struct vnode *vp; char *path; int error, fd, fflags; bool read, write; error = copyin(uap->fds, &fds, sizeof(fds)); if (error != 0) return (error); /* All the requested rights should be set on the descriptor. */ error = cloudabi_convert_rights( fds.fs_rights_base | fds.fs_rights_inheriting, &rights); if (error != 0) return (error); cap_rights_set(&rights, CAP_LOOKUP); /* Convert rights to corresponding access mode. */ read = (fds.fs_rights_base & (CLOUDABI_RIGHT_FD_READ | CLOUDABI_RIGHT_FILE_READDIR | CLOUDABI_RIGHT_MEM_MAP_EXEC)) != 0; write = (fds.fs_rights_base & (CLOUDABI_RIGHT_FD_DATASYNC | CLOUDABI_RIGHT_FD_WRITE | CLOUDABI_RIGHT_FILE_ALLOCATE | CLOUDABI_RIGHT_FILE_STAT_FPUT_SIZE)) != 0; fflags = write ? read ? FREAD | FWRITE : FWRITE : FREAD; /* Convert open flags. */ if ((uap->oflags & CLOUDABI_O_CREAT) != 0) { fflags |= O_CREAT; cap_rights_set(&rights, CAP_CREATE); } if ((uap->oflags & CLOUDABI_O_DIRECTORY) != 0) fflags |= O_DIRECTORY; if ((uap->oflags & CLOUDABI_O_EXCL) != 0) fflags |= O_EXCL; if ((uap->oflags & CLOUDABI_O_TRUNC) != 0) { fflags |= O_TRUNC; cap_rights_set(&rights, CAP_FTRUNCATE); } if ((fds.fs_flags & CLOUDABI_FDFLAG_APPEND) != 0) fflags |= O_APPEND; if ((fds.fs_flags & CLOUDABI_FDFLAG_NONBLOCK) != 0) fflags |= O_NONBLOCK; if ((fds.fs_flags & (CLOUDABI_FDFLAG_SYNC | CLOUDABI_FDFLAG_DSYNC | CLOUDABI_FDFLAG_RSYNC)) != 0) { fflags |= O_SYNC; cap_rights_set(&rights, CAP_FSYNC); } if ((uap->dirfd.flags & CLOUDABI_LOOKUP_SYMLINK_FOLLOW) == 0) fflags |= O_NOFOLLOW; if (write && (fflags & (O_APPEND | O_TRUNC)) == 0) cap_rights_set(&rights, CAP_SEEK); /* Allocate new file descriptor. */ error = falloc_noinstall(td, &fp); if (error != 0) return (error); fp->f_flag = fflags & FMASK; /* Open path. */ error = copyin_path(uap->path, uap->path_len, &path); if (error != 0) { fdrop(fp, td); return (error); } NDINIT_ATRIGHTS(&nd, LOOKUP, FOLLOW, UIO_SYSSPACE, path, uap->dirfd.fd, &rights, td); error = vn_open(&nd, &fflags, 0777 & ~td->td_proc->p_fd->fd_cmask, fp); cloudabi_freestr(path); if (error != 0) { /* Custom operations provided. */ if (error == ENXIO && fp->f_ops != &badfileops) goto success; /* * POSIX compliance: return ELOOP in case openat() is * called on a symbolic link and O_NOFOLLOW is set. */ if (error == EMLINK) error = ELOOP; fdrop(fp, td); return (error); } NDFREE(&nd, NDF_ONLY_PNBUF); filecaps_free(&nd.ni_filecaps); fp->f_vnode = vp = nd.ni_vp; /* Install vnode operations if no custom operations are provided. */ if (fp->f_ops == &badfileops) { fp->f_seqcount = 1; finit(fp, (fflags & FMASK) | (fp->f_flag & FHASLOCK), DTYPE_VNODE, vp, &vnops); } VOP_UNLOCK(vp, 0); /* Truncate file. */ if (fflags & O_TRUNC) { error = fo_truncate(fp, 0, td->td_ucred, td); if (error != 0) { fdrop(fp, td); return (error); } } success: /* Determine which Capsicum rights to set on the file descriptor. */ cloudabi_remove_conflicting_rights(cloudabi_convert_filetype(fp), &fds.fs_rights_base, &fds.fs_rights_inheriting); cloudabi_convert_rights(fds.fs_rights_base | fds.fs_rights_inheriting, &fcaps.fc_rights); if (cap_rights_is_set(&fcaps.fc_rights)) fcaps.fc_fcntls = CAP_FCNTL_SETFL; error = finstall(td, fp, &fd, fflags, &fcaps); fdrop(fp, td); if (error != 0) return (error); td->td_retval[0] = fd; return (0); } /* Converts a FreeBSD directory entry structure and writes it to userspace. */ static int write_dirent(struct dirent *bde, cloudabi_dircookie_t cookie, struct uio *uio) { cloudabi_dirent_t cde = { .d_next = cookie, .d_ino = bde->d_fileno, .d_namlen = bde->d_namlen, }; size_t len; int error; /* Convert file type. */ switch (bde->d_type) { case DT_BLK: cde.d_type = CLOUDABI_FILETYPE_BLOCK_DEVICE; break; case DT_CHR: cde.d_type = CLOUDABI_FILETYPE_CHARACTER_DEVICE; break; case DT_DIR: cde.d_type = CLOUDABI_FILETYPE_DIRECTORY; break; case DT_FIFO: cde.d_type = CLOUDABI_FILETYPE_SOCKET_STREAM; break; case DT_LNK: cde.d_type = CLOUDABI_FILETYPE_SYMBOLIC_LINK; break; case DT_REG: cde.d_type = CLOUDABI_FILETYPE_REGULAR_FILE; break; case DT_SOCK: /* The exact socket type cannot be derived. */ cde.d_type = CLOUDABI_FILETYPE_SOCKET_STREAM; break; default: cde.d_type = CLOUDABI_FILETYPE_UNKNOWN; break; } /* Write directory entry structure. */ len = sizeof(cde) < uio->uio_resid ? sizeof(cde) : uio->uio_resid; error = uiomove(&cde, len, uio); if (error != 0) return (error); /* Write filename. */ len = bde->d_namlen < uio->uio_resid ? bde->d_namlen : uio->uio_resid; return (uiomove(bde->d_name, len, uio)); } int cloudabi_sys_file_readdir(struct thread *td, struct cloudabi_sys_file_readdir_args *uap) { struct iovec iov = { .iov_base = uap->buf, .iov_len = uap->buf_len }; struct uio uio = { .uio_iov = &iov, .uio_iovcnt = 1, .uio_resid = iov.iov_len, .uio_segflg = UIO_USERSPACE, .uio_rw = UIO_READ, .uio_td = td }; struct file *fp; struct vnode *vp; void *readbuf; cloudabi_dircookie_t offset; int error; /* Obtain directory vnode. */ error = getvnode(td, uap->fd, &cap_read_rights, &fp); if (error != 0) { if (error == EINVAL) return (ENOTDIR); return (error); } if ((fp->f_flag & FREAD) == 0) { fdrop(fp, td); return (EBADF); } /* * Call VOP_READDIR() and convert resulting data until the user * provided buffer is filled. */ readbuf = malloc(MAXBSIZE, M_TEMP, M_WAITOK); offset = uap->cookie; vp = fp->f_vnode; while (uio.uio_resid > 0) { struct iovec readiov = { .iov_base = readbuf, .iov_len = MAXBSIZE }; struct uio readuio = { .uio_iov = &readiov, .uio_iovcnt = 1, .uio_rw = UIO_READ, .uio_segflg = UIO_SYSSPACE, .uio_td = td, .uio_resid = MAXBSIZE, .uio_offset = offset }; struct dirent *bde; unsigned long *cookies, *cookie; size_t readbuflen; int eof, ncookies; /* Validate file type. */ vn_lock(vp, LK_SHARED | LK_RETRY); if (vp->v_type != VDIR) { VOP_UNLOCK(vp, 0); error = ENOTDIR; goto done; } #ifdef MAC error = mac_vnode_check_readdir(td->td_ucred, vp); if (error != 0) { VOP_UNLOCK(vp, 0); goto done; } #endif /* MAC */ /* Read new directory entries. */ cookies = NULL; ncookies = 0; error = VOP_READDIR(vp, &readuio, fp->f_cred, &eof, &ncookies, &cookies); VOP_UNLOCK(vp, 0); if (error != 0) goto done; /* Convert entries to CloudABI's format. */ readbuflen = MAXBSIZE - readuio.uio_resid; bde = readbuf; cookie = cookies; while (readbuflen >= offsetof(struct dirent, d_name) && uio.uio_resid > 0 && ncookies > 0) { /* Ensure that the returned offset always increases. */ if (readbuflen >= bde->d_reclen && bde->d_fileno != 0 && *cookie > offset) { error = write_dirent(bde, *cookie, &uio); if (error != 0) { free(cookies, M_TEMP); goto done; } } if (offset < *cookie) offset = *cookie; ++cookie; --ncookies; readbuflen -= bde->d_reclen; bde = (struct dirent *)((char *)bde + bde->d_reclen); } free(cookies, M_TEMP); if (eof) break; } done: fdrop(fp, td); free(readbuf, M_TEMP); if (error != 0) return (error); /* Return number of bytes copied to userspace. */ td->td_retval[0] = uap->buf_len - uio.uio_resid; return (0); } int cloudabi_sys_file_readlink(struct thread *td, struct cloudabi_sys_file_readlink_args *uap) { char *path; int error; error = copyin_path(uap->path, uap->path_len, &path); if (error != 0) return (error); error = kern_readlinkat(td, uap->fd, path, UIO_SYSSPACE, uap->buf, UIO_USERSPACE, uap->buf_len); cloudabi_freestr(path); return (error); } int cloudabi_sys_file_rename(struct thread *td, struct cloudabi_sys_file_rename_args *uap) { char *old, *new; int error; error = copyin_path(uap->path1, uap->path1_len, &old); if (error != 0) return (error); error = copyin_path(uap->path2, uap->path2_len, &new); if (error != 0) { cloudabi_freestr(old); return (error); } error = kern_renameat(td, uap->fd1, old, uap->fd2, new, UIO_SYSSPACE); cloudabi_freestr(old); cloudabi_freestr(new); return (error); } /* Converts a FreeBSD stat structure to a CloudABI stat structure. */ static void convert_stat(const struct stat *sb, cloudabi_filestat_t *csb) { cloudabi_filestat_t res = { .st_dev = sb->st_dev, .st_ino = sb->st_ino, .st_nlink = sb->st_nlink, .st_size = sb->st_size, }; cloudabi_convert_timespec(&sb->st_atim, &res.st_atim); cloudabi_convert_timespec(&sb->st_mtim, &res.st_mtim); cloudabi_convert_timespec(&sb->st_ctim, &res.st_ctim); *csb = res; } int cloudabi_sys_file_stat_fget(struct thread *td, struct cloudabi_sys_file_stat_fget_args *uap) { struct stat sb; cloudabi_filestat_t csb; struct file *fp; cloudabi_filetype_t filetype; int error; memset(&csb, 0, sizeof(csb)); /* Fetch file descriptor attributes. */ error = fget(td, uap->fd, &cap_fstat_rights, &fp); if (error != 0) return (error); error = fo_stat(fp, &sb, td->td_ucred, td); if (error != 0) { fdrop(fp, td); return (error); } filetype = cloudabi_convert_filetype(fp); fdrop(fp, td); /* Convert attributes to CloudABI's format. */ convert_stat(&sb, &csb); csb.st_filetype = filetype; return (copyout(&csb, uap->buf, sizeof(csb))); } /* Converts timestamps to arguments to futimens() and utimensat(). */ static void convert_utimens_arguments(const cloudabi_filestat_t *fs, cloudabi_fsflags_t flags, struct timespec *ts) { if ((flags & CLOUDABI_FILESTAT_ATIM_NOW) != 0) { ts[0].tv_nsec = UTIME_NOW; } else if ((flags & CLOUDABI_FILESTAT_ATIM) != 0) { ts[0].tv_sec = fs->st_atim / 1000000000; ts[0].tv_nsec = fs->st_atim % 1000000000; } else { ts[0].tv_nsec = UTIME_OMIT; } if ((flags & CLOUDABI_FILESTAT_MTIM_NOW) != 0) { ts[1].tv_nsec = UTIME_NOW; } else if ((flags & CLOUDABI_FILESTAT_MTIM) != 0) { ts[1].tv_sec = fs->st_mtim / 1000000000; ts[1].tv_nsec = fs->st_mtim % 1000000000; } else { ts[1].tv_nsec = UTIME_OMIT; } } int cloudabi_sys_file_stat_fput(struct thread *td, struct cloudabi_sys_file_stat_fput_args *uap) { cloudabi_filestat_t fs; struct timespec ts[2]; int error; error = copyin(uap->buf, &fs, sizeof(fs)); if (error != 0) return (error); /* * Only support truncation and timestamp modification separately * for now, to prevent unnecessary code duplication. */ if ((uap->flags & CLOUDABI_FILESTAT_SIZE) != 0) { /* Call into kern_ftruncate() for file truncation. */ if ((uap->flags & ~CLOUDABI_FILESTAT_SIZE) != 0) return (EINVAL); return (kern_ftruncate(td, uap->fd, fs.st_size)); } else if ((uap->flags & (CLOUDABI_FILESTAT_ATIM | CLOUDABI_FILESTAT_ATIM_NOW | CLOUDABI_FILESTAT_MTIM | CLOUDABI_FILESTAT_MTIM_NOW)) != 0) { /* Call into kern_futimens() for timestamp modification. */ if ((uap->flags & ~(CLOUDABI_FILESTAT_ATIM | CLOUDABI_FILESTAT_ATIM_NOW | CLOUDABI_FILESTAT_MTIM | CLOUDABI_FILESTAT_MTIM_NOW)) != 0) return (EINVAL); convert_utimens_arguments(&fs, uap->flags, ts); return (kern_futimens(td, uap->fd, ts, UIO_SYSSPACE)); } return (EINVAL); } int cloudabi_sys_file_stat_get(struct thread *td, struct cloudabi_sys_file_stat_get_args *uap) { struct stat sb; cloudabi_filestat_t csb; char *path; int error; memset(&csb, 0, sizeof(csb)); error = copyin_path(uap->path, uap->path_len, &path); if (error != 0) return (error); error = kern_statat(td, (uap->fd.flags & CLOUDABI_LOOKUP_SYMLINK_FOLLOW) != 0 ? 0 : AT_SYMLINK_NOFOLLOW, uap->fd.fd, path, UIO_SYSSPACE, &sb, NULL); cloudabi_freestr(path); if (error != 0) return (error); /* Convert results and return them. */ convert_stat(&sb, &csb); if (S_ISBLK(sb.st_mode)) csb.st_filetype = CLOUDABI_FILETYPE_BLOCK_DEVICE; else if (S_ISCHR(sb.st_mode)) csb.st_filetype = CLOUDABI_FILETYPE_CHARACTER_DEVICE; else if (S_ISDIR(sb.st_mode)) csb.st_filetype = CLOUDABI_FILETYPE_DIRECTORY; else if (S_ISFIFO(sb.st_mode)) csb.st_filetype = CLOUDABI_FILETYPE_SOCKET_STREAM; else if (S_ISREG(sb.st_mode)) csb.st_filetype = CLOUDABI_FILETYPE_REGULAR_FILE; else if (S_ISSOCK(sb.st_mode)) { /* Inaccurate, but the best that we can do. */ csb.st_filetype = CLOUDABI_FILETYPE_SOCKET_STREAM; } else if (S_ISLNK(sb.st_mode)) csb.st_filetype = CLOUDABI_FILETYPE_SYMBOLIC_LINK; else csb.st_filetype = CLOUDABI_FILETYPE_UNKNOWN; return (copyout(&csb, uap->buf, sizeof(csb))); } int cloudabi_sys_file_stat_put(struct thread *td, struct cloudabi_sys_file_stat_put_args *uap) { cloudabi_filestat_t fs; struct timespec ts[2]; char *path; int error; /* * Only support timestamp modification for now, as there is no * truncateat(). */ if ((uap->flags & ~(CLOUDABI_FILESTAT_ATIM | CLOUDABI_FILESTAT_ATIM_NOW | CLOUDABI_FILESTAT_MTIM | CLOUDABI_FILESTAT_MTIM_NOW)) != 0) return (EINVAL); error = copyin(uap->buf, &fs, sizeof(fs)); if (error != 0) return (error); error = copyin_path(uap->path, uap->path_len, &path); if (error != 0) return (error); convert_utimens_arguments(&fs, uap->flags, ts); error = kern_utimensat(td, uap->fd.fd, path, UIO_SYSSPACE, ts, UIO_SYSSPACE, (uap->fd.flags & CLOUDABI_LOOKUP_SYMLINK_FOLLOW) ? 0 : AT_SYMLINK_NOFOLLOW); cloudabi_freestr(path); return (error); } int cloudabi_sys_file_symlink(struct thread *td, struct cloudabi_sys_file_symlink_args *uap) { char *path1, *path2; int error; error = copyin_path(uap->path1, uap->path1_len, &path1); if (error != 0) return (error); error = copyin_path(uap->path2, uap->path2_len, &path2); if (error != 0) { cloudabi_freestr(path1); return (error); } error = kern_symlinkat(td, path1, uap->fd, path2, UIO_SYSSPACE); cloudabi_freestr(path1); cloudabi_freestr(path2); return (error); } int cloudabi_sys_file_unlink(struct thread *td, struct cloudabi_sys_file_unlink_args *uap) { char *path; int error; error = copyin_path(uap->path, uap->path_len, &path); if (error != 0) return (error); if (uap->flags & CLOUDABI_UNLINK_REMOVEDIR) - error = kern_rmdirat(td, uap->fd, path, UIO_SYSSPACE, 0); + error = kern_frmdirat(td, uap->fd, path, FD_NONE, + UIO_SYSSPACE, 0); else - error = kern_unlinkat(td, uap->fd, path, UIO_SYSSPACE, 0, 0); + error = kern_funlinkat(td, uap->fd, path, FD_NONE, + UIO_SYSSPACE, 0, 0); cloudabi_freestr(path); return (error); } Index: head/sys/compat/freebsd32/syscalls.master =================================================================== --- head/sys/compat/freebsd32/syscalls.master (revision 345981) +++ head/sys/compat/freebsd32/syscalls.master (revision 345982) @@ -1,1149 +1,1151 @@ $FreeBSD$ ; from: @(#)syscalls.master 8.2 (Berkeley) 1/13/94 ; from: src/sys/kern/syscalls.master 1.107 ; ; System call name/number master file. ; Processed to created init_sysent.c, syscalls.c and syscall.h. ; Columns: number audit type name alt{name,tag,rtyp}/comments ; number system call number, must be in order ; audit the audit event associated with the system call ; A value of AUE_NULL means no auditing, but it also means that ; there is no audit event for the call at this time. For the ; case where the event exists, but we don't want auditing, the ; event should be #defined to AUE_NULL in audit_kevents.h. ; type one of STD, OBSOL, UNIMPL, COMPAT, COMPAT4, COMPAT6, ; COMPAT7, COMPAT11, NODEF, NOARGS, NOPROTO, NOSTD ; The COMPAT* options may be combined with one or more NO* ; options separated by '|' with no spaces (e.g. COMPAT|NOARGS) ; name pseudo-prototype of syscall routine ; If one of the following alts is different, then all appear: ; altname name of system call if different ; alttag name of args struct tag if different from [o]`name'"_args" ; altrtyp return type if not int (bogus - syscalls always return int) ; for UNIMPL/OBSOL, name continues with comments ; types: ; STD always included ; COMPAT included on COMPAT #ifdef ; COMPAT4 included on COMPAT_FREEBSD4 #ifdef (FreeBSD 4 compat) ; COMPAT6 included on COMPAT_FREEBSD6 #ifdef (FreeBSD 6 compat) ; COMPAT7 included on COMPAT_FREEBSD7 #ifdef (FreeBSD 7 compat) ; COMPAT10 included on COMPAT_FREEBSD10 #ifdef (FreeBSD 10 compat) ; COMPAT11 included on COMPAT_FREEBSD11 #ifdef (FreeBSD 11 compat) ; OBSOL obsolete, not included in system, only specifies name ; UNIMPL not implemented, placeholder only ; NOSTD implemented but as a lkm that can be statically ; compiled in; sysent entry will be filled with lkmressys ; so the SYSCALL_MODULE macro works ; NOARGS same as STD except do not create structure in sys/sysproto.h ; NODEF same as STD except only have the entry in the syscall table ; added. Meaning - do not create structure or function ; prototype in sys/sysproto.h ; NOPROTO same as STD except do not create structure or ; function prototype in sys/sysproto.h. Does add a ; definition to syscall.h besides adding a sysent. ; #ifdef's, etc. may be included, and are copied to the output files. #include #include #include #include #include #include #include #if !defined(PAD64_REQUIRED) && !defined(__amd64__) #define PAD64_REQUIRED #endif ; Reserved/unimplemented system calls in the range 0-150 inclusive ; are reserved for use in future Berkeley releases. ; Additional system calls implemented in vendor and other ; redistributions should be placed in the reserved range at the end ; of the current calls. 0 AUE_NULL NOPROTO { int nosys(void); } syscall nosys_args int 1 AUE_EXIT NOPROTO { void sys_exit(int rval); } exit \ sys_exit_args void 2 AUE_FORK NOPROTO { int fork(void); } 3 AUE_READ NOPROTO { ssize_t read(int fd, void *buf, \ size_t nbyte); } 4 AUE_WRITE NOPROTO { ssize_t write(int fd, const void *buf, \ size_t nbyte); } 5 AUE_OPEN_RWTC NOPROTO { int open(const char *path, int flags, \ mode_t mode); } 6 AUE_CLOSE NOPROTO { int close(int fd); } 7 AUE_WAIT4 STD { int freebsd32_wait4(int pid, int *status, \ int options, struct rusage32 *rusage); } 8 AUE_CREAT OBSOL old creat 9 AUE_LINK NOPROTO { int link(const char *path, \ const char *link); } 10 AUE_UNLINK NOPROTO { int unlink(const char *path); } 11 AUE_NULL OBSOL execv 12 AUE_CHDIR NOPROTO { int chdir(const char *path); } 13 AUE_FCHDIR NOPROTO { int fchdir(int fd); } 14 AUE_MKNOD COMPAT11|NOPROTO { int mknod(const char *path, \ int mode, uint32_t dev); } 15 AUE_CHMOD NOPROTO { int chmod(const char *path, mode_t mode); } 16 AUE_CHOWN NOPROTO { int chown(const char *path, int uid, int gid); } 17 AUE_NULL NOPROTO { void *break(char *nsize); } 18 AUE_GETFSSTAT COMPAT4 { int freebsd32_getfsstat( \ struct statfs32 *buf, long bufsize, \ int mode); } 19 AUE_LSEEK COMPAT { int freebsd32_lseek(int fd, int offset, \ int whence); } 20 AUE_GETPID NOPROTO { pid_t getpid(void); } 21 AUE_MOUNT NOPROTO { int mount(const char *type, \ const char *path, \ int flags, void *data); } 22 AUE_UMOUNT NOPROTO { int unmount(const char *path, int flags); } 23 AUE_SETUID NOPROTO { int setuid(uid_t uid); } 24 AUE_GETUID NOPROTO { uid_t getuid(void); } 25 AUE_GETEUID NOPROTO { uid_t geteuid(void); } 26 AUE_PTRACE NOPROTO { int ptrace(int req, pid_t pid, \ caddr_t addr, int data); } 27 AUE_RECVMSG STD { int freebsd32_recvmsg(int s, struct msghdr32 *msg, \ int flags); } 28 AUE_SENDMSG STD { int freebsd32_sendmsg(int s, struct msghdr32 *msg, \ int flags); } 29 AUE_RECVFROM STD { int freebsd32_recvfrom(int s, void *buf, \ uint32_t len, int flags, \ struct sockaddr *from, \ uint32_t fromlenaddr); } 30 AUE_ACCEPT NOPROTO { int accept(int s, struct sockaddr *name, \ int *anamelen); } 31 AUE_GETPEERNAME NOPROTO { int getpeername(int fdes, \ struct sockaddr *asa, \ int *alen); } 32 AUE_GETSOCKNAME NOPROTO { int getsockname(int fdes, \ struct sockaddr *asa, \ int *alen); } 33 AUE_ACCESS NOPROTO { int access(const char *path, int amode); } 34 AUE_CHFLAGS NOPROTO { int chflags(const char *path, u_long flags); } 35 AUE_FCHFLAGS NOPROTO { int fchflags(int fd, u_long flags); } 36 AUE_SYNC NOPROTO { int sync(void); } 37 AUE_KILL NOPROTO { int kill(int pid, int signum); } 38 AUE_STAT COMPAT { int freebsd32_stat(const char *path, \ struct ostat32 *ub); } 39 AUE_GETPPID NOPROTO { pid_t getppid(void); } 40 AUE_LSTAT COMPAT { int freebsd32_lstat(const char *path, \ struct ostat *ub); } 41 AUE_DUP NOPROTO { int dup(u_int fd); } 42 AUE_PIPE COMPAT10 { int freebsd32_pipe(void); } 43 AUE_GETEGID NOPROTO { gid_t getegid(void); } 44 AUE_PROFILE NOPROTO { int profil(char *samples, size_t size, \ size_t offset, u_int scale); } 45 AUE_KTRACE NOPROTO { int ktrace(const char *fname, int ops, \ int facs, int pid); } 46 AUE_SIGACTION COMPAT { int freebsd32_sigaction( int signum, \ struct osigaction32 *nsa, \ struct osigaction32 *osa); } 47 AUE_GETGID NOPROTO { gid_t getgid(void); } 48 AUE_SIGPROCMASK COMPAT { int freebsd32_sigprocmask(int how, \ osigset_t mask); } 49 AUE_GETLOGIN NOPROTO { int getlogin(char *namebuf, \ u_int namelen); } 50 AUE_SETLOGIN NOPROTO { int setlogin(const char *namebuf); } 51 AUE_ACCT NOPROTO { int acct(const char *path); } 52 AUE_SIGPENDING COMPAT { int freebsd32_sigpending(void); } 53 AUE_SIGALTSTACK STD { int freebsd32_sigaltstack( \ struct sigaltstack32 *ss, \ struct sigaltstack32 *oss); } 54 AUE_IOCTL STD { int freebsd32_ioctl(int fd, uint32_t com, \ struct md_ioctl32 *data); } 55 AUE_REBOOT NOPROTO { int reboot(int opt); } 56 AUE_REVOKE NOPROTO { int revoke(const char *path); } 57 AUE_SYMLINK NOPROTO { int symlink(const char *path, \ const char *link); } 58 AUE_READLINK NOPROTO { ssize_t readlink(const char *path, char *buf, \ size_t count); } 59 AUE_EXECVE STD { int freebsd32_execve(const char *fname, \ uint32_t *argv, uint32_t *envv); } 60 AUE_UMASK NOPROTO { int umask(mode_t newmask); } 61 AUE_CHROOT NOPROTO { int chroot(const char *path); } 62 AUE_FSTAT COMPAT { int freebsd32_fstat(int fd, \ struct ostat32 *ub); } 63 AUE_NULL OBSOL ogetkerninfo 64 AUE_NULL COMPAT { int freebsd32_getpagesize( \ int32_t dummy); } 65 AUE_MSYNC NOPROTO { int msync(void *addr, size_t len, \ int flags); } 66 AUE_VFORK NOPROTO { int vfork(void); } 67 AUE_NULL OBSOL vread 68 AUE_NULL OBSOL vwrite 69 AUE_SBRK NOPROTO { int sbrk(int incr); } 70 AUE_SSTK NOPROTO { int sstk(int incr); } 71 AUE_MMAP COMPAT|NOPROTO { void *mmap(void *addr, int len, \ int prot, int flags, int fd, int pos); } 72 AUE_O_VADVISE COMPAT11|NOPROTO { int vadvise(int anom); } 73 AUE_MUNMAP NOPROTO { int munmap(void *addr, size_t len); } 74 AUE_MPROTECT STD { int freebsd32_mprotect(void *addr, \ size_t len, int prot); } 75 AUE_MADVISE NOPROTO { int madvise(void *addr, size_t len, \ int behav); } 76 AUE_NULL OBSOL vhangup 77 AUE_NULL OBSOL vlimit 78 AUE_MINCORE NOPROTO { int mincore(const void *addr, size_t len, \ char *vec); } 79 AUE_GETGROUPS NOPROTO { int getgroups(u_int gidsetsize, \ gid_t *gidset); } 80 AUE_SETGROUPS NOPROTO { int setgroups(u_int gidsetsize, \ gid_t *gidset); } 81 AUE_GETPGRP NOPROTO { int getpgrp(void); } 82 AUE_SETPGRP NOPROTO { int setpgid(int pid, int pgid); } 83 AUE_SETITIMER STD { int freebsd32_setitimer(u_int which, \ struct itimerval32 *itv, \ struct itimerval32 *oitv); } 84 AUE_NULL OBSOL owait ; XXX implement 85 AUE_SWAPON NOPROTO { int swapon(const char *name); } 86 AUE_GETITIMER STD { int freebsd32_getitimer(u_int which, \ struct itimerval32 *itv); } 87 AUE_O_GETHOSTNAME OBSOL ogethostname 88 AUE_O_SETHOSTNAME OBSOL osethostname 89 AUE_GETDTABLESIZE NOPROTO { int getdtablesize(void); } 90 AUE_DUP2 NOPROTO { int dup2(u_int from, u_int to); } 91 AUE_NULL UNIMPL getdopt 92 AUE_FCNTL STD { int freebsd32_fcntl(int fd, int cmd, \ int arg); } 93 AUE_SELECT STD { int freebsd32_select(int nd, fd_set *in, \ fd_set *ou, fd_set *ex, \ struct timeval32 *tv); } 94 AUE_NULL UNIMPL setdopt 95 AUE_FSYNC NOPROTO { int fsync(int fd); } 96 AUE_SETPRIORITY NOPROTO { int setpriority(int which, int who, \ int prio); } 97 AUE_SOCKET NOPROTO { int socket(int domain, int type, \ int protocol); } 98 AUE_CONNECT NOPROTO { int connect(int s, \ const struct sockaddr *name, \ int namelen); } 99 AUE_NULL OBSOL oaccept 100 AUE_GETPRIORITY NOPROTO { int getpriority(int which, int who); } 101 AUE_NULL OBSOL osend 102 AUE_NULL OBSOL orecv 103 AUE_SIGRETURN COMPAT { int freebsd32_sigreturn( \ struct ia32_sigcontext3 *sigcntxp); } 104 AUE_BIND NOPROTO { int bind(int s, const struct sockaddr *name, \ int namelen); } 105 AUE_SETSOCKOPT NOPROTO { int setsockopt(int s, int level, \ int name, const void *val, int valsize); } 106 AUE_LISTEN NOPROTO { int listen(int s, int backlog); } 107 AUE_NULL OBSOL vtimes 108 AUE_O_SIGVEC COMPAT { int freebsd32_sigvec(int signum, \ struct sigvec32 *nsv, \ struct sigvec32 *osv); } 109 AUE_O_SIGBLOCK COMPAT { int freebsd32_sigblock(int mask); } 110 AUE_O_SIGSETMASK COMPAT { int freebsd32_sigsetmask( int mask); } 111 AUE_SIGSUSPEND COMPAT { int freebsd32_sigsuspend( int mask); } 112 AUE_O_SIGSTACK COMPAT { int freebsd32_sigstack( \ struct sigstack32 *nss, \ struct sigstack32 *oss); } 113 AUE_NULL OBSOL orecvmsg 114 AUE_NULL OBSOL osendmsg 115 AUE_NULL OBSOL vtrace 116 AUE_GETTIMEOFDAY STD { int freebsd32_gettimeofday( \ struct timeval32 *tp, \ struct timezone *tzp); } 117 AUE_GETRUSAGE STD { int freebsd32_getrusage(int who, \ struct rusage32 *rusage); } 118 AUE_GETSOCKOPT NOPROTO { int getsockopt(int s, int level, \ int name, void *val, int *avalsize); } 119 AUE_NULL UNIMPL resuba (BSD/OS 2.x) 120 AUE_READV STD { int freebsd32_readv(int fd, \ struct iovec32 *iovp, u_int iovcnt); } 121 AUE_WRITEV STD { int freebsd32_writev(int fd, \ struct iovec32 *iovp, u_int iovcnt); } 122 AUE_SETTIMEOFDAY STD { int freebsd32_settimeofday( \ struct timeval32 *tv, \ struct timezone *tzp); } 123 AUE_FCHOWN NOPROTO { int fchown(int fd, int uid, int gid); } 124 AUE_FCHMOD NOPROTO { int fchmod(int fd, mode_t mode); } 125 AUE_RECVFROM OBSOL orecvfrom 126 AUE_SETREUID NOPROTO { int setreuid(int ruid, int euid); } 127 AUE_SETREGID NOPROTO { int setregid(int rgid, int egid); } 128 AUE_RENAME NOPROTO { int rename(const char *from, \ const char *to); } 129 AUE_TRUNCATE COMPAT|NOPROTO { int truncate(const char *path, \ int length); } 130 AUE_FTRUNCATE COMPAT|NOPROTO { int ftruncate(int fd, int length); } 131 AUE_FLOCK NOPROTO { int flock(int fd, int how); } 132 AUE_MKFIFO NOPROTO { int mkfifo(const char *path, mode_t mode); } 133 AUE_SENDTO NOPROTO { int sendto(int s, const void *buf, \ size_t len, int flags, \ const struct sockaddr *to, \ int tolen); } 134 AUE_SHUTDOWN NOPROTO { int shutdown(int s, int how); } 135 AUE_SOCKETPAIR NOPROTO { int socketpair(int domain, int type, \ int protocol, int *rsv); } 136 AUE_MKDIR NOPROTO { int mkdir(const char *path, mode_t mode); } 137 AUE_RMDIR NOPROTO { int rmdir(const char *path); } 138 AUE_UTIMES STD { int freebsd32_utimes(const char *path, \ struct timeval32 *tptr); } 139 AUE_NULL OBSOL 4.2 sigreturn 140 AUE_ADJTIME STD { int freebsd32_adjtime( \ struct timeval32 *delta, \ struct timeval32 *olddelta); } 141 AUE_GETPEERNAME OBSOL ogetpeername 142 AUE_SYSCTL OBSOL ogethostid 143 AUE_SYSCTL OBSOL sethostid 144 AUE_GETRLIMIT OBSOL getrlimit 145 AUE_SETRLIMIT OBSOL setrlimit 146 AUE_KILLPG OBSOL killpg 147 AUE_SETSID NOPROTO { int setsid(void); } 148 AUE_QUOTACTL NOPROTO { int quotactl(const char *path, int cmd, \ int uid, void *arg); } 149 AUE_O_QUOTA OBSOL oquota 150 AUE_GETSOCKNAME OBSOL ogetsockname ; Syscalls 151-180 inclusive are reserved for vendor-specific ; system calls. (This includes various calls added for compatibity ; with other Unix variants.) ; Some of these calls are now supported by BSD... 151 AUE_NULL UNIMPL sem_lock (BSD/OS 2.x) 152 AUE_NULL UNIMPL sem_wakeup (BSD/OS 2.x) 153 AUE_NULL UNIMPL asyncdaemon (BSD/OS 2.x) ; 154 is initialised by the NLM code, if present. 154 AUE_NULL UNIMPL nlm_syscall ; 155 is initialized by the NFS code, if present. ; XXX this is a problem!!! 155 AUE_NFS_SVC UNIMPL nfssvc 156 AUE_GETDIRENTRIES COMPAT { int freebsd32_getdirentries(int fd, \ char *buf, u_int count, uint32_t *basep); } 157 AUE_STATFS COMPAT4 { int freebsd32_statfs(const char *path, \ struct statfs32 *buf); } 158 AUE_FSTATFS COMPAT4 { int freebsd32_fstatfs(int fd, \ struct statfs32 *buf); } 159 AUE_NULL UNIMPL nosys 160 AUE_LGETFH UNIMPL lgetfh 161 AUE_NFS_GETFH NOPROTO { int getfh(const char *fname, \ struct fhandle *fhp); } 162 AUE_SYSCTL OBSOL getdomainname 163 AUE_SYSCTL OBSOL setdomainname 164 AUE_NULL OBSOL uname 165 AUE_SYSARCH STD { int freebsd32_sysarch(int op, char *parms); } 166 AUE_RTPRIO NOPROTO { int rtprio(int function, pid_t pid, \ struct rtprio *rtp); } 167 AUE_NULL UNIMPL nosys 168 AUE_NULL UNIMPL nosys 169 AUE_SEMSYS NOSTD { int freebsd32_semsys(int which, int a2, \ int a3, int a4, int a5); } 170 AUE_MSGSYS NOSTD { int freebsd32_msgsys(int which, int a2, \ int a3, int a4, int a5, int a6); } 171 AUE_SHMSYS NOSTD { int freebsd32_shmsys(uint32_t which, uint32_t a2, \ uint32_t a3, uint32_t a4); } 172 AUE_NULL UNIMPL nosys 173 AUE_PREAD COMPAT6 { ssize_t freebsd32_pread(int fd, void *buf, \ size_t nbyte, int pad, \ uint32_t offset1, uint32_t offset2); } 174 AUE_PWRITE COMPAT6 { ssize_t freebsd32_pwrite(int fd, \ const void *buf, size_t nbyte, int pad, \ uint32_t offset1, uint32_t offset2); } 175 AUE_NULL UNIMPL nosys 176 AUE_NTP_ADJTIME NOPROTO { int ntp_adjtime(struct timex *tp); } 177 AUE_NULL UNIMPL sfork (BSD/OS 2.x) 178 AUE_NULL UNIMPL getdescriptor (BSD/OS 2.x) 179 AUE_NULL UNIMPL setdescriptor (BSD/OS 2.x) 180 AUE_NULL UNIMPL nosys ; Syscalls 181-199 are used by/reserved for BSD 181 AUE_SETGID NOPROTO { int setgid(gid_t gid); } 182 AUE_SETEGID NOPROTO { int setegid(gid_t egid); } 183 AUE_SETEUID NOPROTO { int seteuid(uid_t euid); } 184 AUE_NULL OBSOL lfs_bmapv 185 AUE_NULL OBSOL lfs_markv 186 AUE_NULL OBSOL lfs_segclean 187 AUE_NULL OBSOL lfs_segwait 188 AUE_STAT COMPAT11 { int freebsd32_stat(const char *path, \ struct freebsd11_stat32 *ub); } 189 AUE_FSTAT COMPAT11 { int freebsd32_fstat(int fd, \ struct freebsd11_stat32 *ub); } 190 AUE_LSTAT COMPAT11 { int freebsd32_lstat(const char *path, \ struct freebsd11_stat32 *ub); } 191 AUE_PATHCONF NOPROTO { int pathconf(const char *path, int name); } 192 AUE_FPATHCONF NOPROTO { int fpathconf(int fd, int name); } 193 AUE_NULL UNIMPL nosys 194 AUE_GETRLIMIT NOPROTO { int getrlimit(u_int which, \ struct rlimit *rlp); } getrlimit \ __getrlimit_args int 195 AUE_SETRLIMIT NOPROTO { int setrlimit(u_int which, \ struct rlimit *rlp); } setrlimit \ __setrlimit_args int 196 AUE_GETDIRENTRIES COMPAT11 { int freebsd32_getdirentries(int fd, \ char *buf, u_int count, int32_t *basep); } 197 AUE_MMAP COMPAT6 { void *freebsd32_mmap(void *addr, \ size_t len, int prot, int flags, int fd, \ int pad, uint32_t pos1, uint32_t pos2); } 198 AUE_NULL NOPROTO { int nosys(void); } __syscall \ __syscall_args int 199 AUE_LSEEK COMPAT6 { off_t freebsd32_lseek(int fd, int pad, \ uint32_t offset1, uint32_t offset2, \ int whence); } 200 AUE_TRUNCATE COMPAT6 { int freebsd32_truncate(const char *path, \ int pad, uint32_t length1, \ uint32_t length2); } 201 AUE_FTRUNCATE COMPAT6 { int freebsd32_ftruncate(int fd, int pad, \ uint32_t length1, uint32_t length2); } 202 AUE_SYSCTL STD { int freebsd32___sysctl(int *name, \ u_int namelen, void *old, \ uint32_t *oldlenp, const void *new, \ uint32_t newlen); } 203 AUE_MLOCK NOPROTO { int mlock(const void *addr, \ size_t len); } 204 AUE_MUNLOCK NOPROTO { int munlock(const void *addr, \ size_t len); } 205 AUE_UNDELETE NOPROTO { int undelete(const char *path); } 206 AUE_FUTIMES STD { int freebsd32_futimes(int fd, \ struct timeval32 *tptr); } 207 AUE_GETPGID NOPROTO { int getpgid(pid_t pid); } 208 AUE_NULL UNIMPL nosys 209 AUE_POLL NOPROTO { int poll(struct pollfd *fds, u_int nfds, \ int timeout); } ; ; The following are reserved for loadable syscalls ; 210 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 211 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 212 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 213 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 214 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 215 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 216 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 217 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 218 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 219 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 220 AUE_SEMCTL COMPAT7|NOSTD { int freebsd32_semctl( \ int semid, int semnum, \ int cmd, union semun32 *arg); } 221 AUE_SEMGET NOSTD|NOPROTO { int semget(key_t key, int nsems, \ int semflg); } 222 AUE_SEMOP NOSTD|NOPROTO { int semop(int semid, \ struct sembuf *sops, u_int nsops); } 223 AUE_NULL OBSOL semconfig 224 AUE_MSGCTL COMPAT7|NOSTD { int freebsd32_msgctl( \ int msqid, int cmd, \ struct msqid_ds32_old *buf); } 225 AUE_MSGGET NOSTD|NOPROTO { int msgget(key_t key, int msgflg); } 226 AUE_MSGSND NOSTD { int freebsd32_msgsnd(int msqid, void *msgp, \ size_t msgsz, int msgflg); } 227 AUE_MSGRCV NOSTD { int freebsd32_msgrcv(int msqid, void *msgp, \ size_t msgsz, long msgtyp, int msgflg); } 228 AUE_SHMAT NOSTD|NOPROTO { void *shmat(int shmid, void *shmaddr, \ int shmflg); } 229 AUE_SHMCTL COMPAT7|NOSTD { int freebsd32_shmctl( \ int shmid, int cmd, \ struct shmid_ds32_old *buf); } 230 AUE_SHMDT NOSTD|NOPROTO { int shmdt(void *shmaddr); } 231 AUE_SHMGET NOSTD|NOPROTO { int shmget(key_t key, int size, \ int shmflg); } ; 232 AUE_NULL STD { int freebsd32_clock_gettime(clockid_t clock_id, \ struct timespec32 *tp); } 233 AUE_CLOCK_SETTIME STD { int freebsd32_clock_settime(clockid_t clock_id, \ const struct timespec32 *tp); } 234 AUE_NULL STD { int freebsd32_clock_getres(clockid_t clock_id, \ struct timespec32 *tp); } 235 AUE_NULL STD { int freebsd32_ktimer_create(\ clockid_t clock_id, \ struct sigevent32 *evp, int *timerid); } 236 AUE_NULL NOPROTO { int ktimer_delete(int timerid); } 237 AUE_NULL STD { int freebsd32_ktimer_settime(int timerid,\ int flags, \ const struct itimerspec32 *value, \ struct itimerspec32 *ovalue); } 238 AUE_NULL STD { int freebsd32_ktimer_gettime(int timerid,\ struct itimerspec32 *value); } 239 AUE_NULL NOPROTO { int ktimer_getoverrun(int timerid); } 240 AUE_NULL STD { int freebsd32_nanosleep( \ const struct timespec32 *rqtp, \ struct timespec32 *rmtp); } 241 AUE_NULL NOPROTO { int ffclock_getcounter(ffcounter *ffcount); } 242 AUE_NULL NOPROTO { int ffclock_setestimate( \ struct ffclock_estimate *cest); } 243 AUE_NULL NOPROTO { int ffclock_getestimate( \ struct ffclock_estimate *cest); } 244 AUE_NULL STD { int freebsd32_clock_nanosleep( \ clockid_t clock_id, int flags, \ const struct timespec32 *rqtp, \ struct timespec32 *rmtp); } 245 AUE_NULL UNIMPL nosys 246 AUE_NULL UNIMPL nosys 247 AUE_NULL STD { int freebsd32_clock_getcpuclockid2(\ uint32_t id1, uint32_t id2,\ int which, clockid_t *clock_id); } 248 AUE_NULL UNIMPL ntp_gettime 249 AUE_NULL UNIMPL nosys 250 AUE_MINHERIT NOPROTO { int minherit(void *addr, size_t len, \ int inherit); } 251 AUE_RFORK NOPROTO { int rfork(int flags); } 252 AUE_POLL OBSOL openbsd_poll 253 AUE_ISSETUGID NOPROTO { int issetugid(void); } 254 AUE_LCHOWN NOPROTO { int lchown(const char *path, int uid, \ int gid); } 255 AUE_AIO_READ STD { int freebsd32_aio_read( \ struct aiocb32 *aiocbp); } 256 AUE_AIO_WRITE STD { int freebsd32_aio_write( \ struct aiocb32 *aiocbp); } 257 AUE_LIO_LISTIO STD { int freebsd32_lio_listio(int mode, \ struct aiocb32 * const *acb_list, \ int nent, struct sigevent32 *sig); } 258 AUE_NULL UNIMPL nosys 259 AUE_NULL UNIMPL nosys 260 AUE_NULL UNIMPL nosys 261 AUE_NULL UNIMPL nosys 262 AUE_NULL UNIMPL nosys 263 AUE_NULL UNIMPL nosys 264 AUE_NULL UNIMPL nosys 265 AUE_NULL UNIMPL nosys 266 AUE_NULL UNIMPL nosys 267 AUE_NULL UNIMPL nosys 268 AUE_NULL UNIMPL nosys 269 AUE_NULL UNIMPL nosys 270 AUE_NULL UNIMPL nosys 271 AUE_NULL UNIMPL nosys 272 AUE_O_GETDENTS COMPAT11 { int freebsd32_getdents(int fd, char *buf, \ int count); } 273 AUE_NULL UNIMPL nosys 274 AUE_LCHMOD NOPROTO { int lchmod(const char *path, mode_t mode); } 275 AUE_NULL OBSOL netbsd_lchown 276 AUE_LUTIMES STD { int freebsd32_lutimes(const char *path, \ struct timeval32 *tptr); } 277 AUE_NULL OBSOL netbsd_msync 278 AUE_STAT COMPAT11|NOPROTO { int nstat(const char *path, \ struct nstat *ub); } 279 AUE_FSTAT COMPAT11|NOPROTO { int nfstat(int fd, struct nstat *sb); } 280 AUE_LSTAT COMPAT11|NOPROTO { int nlstat(const char *path, \ struct nstat *ub); } 281 AUE_NULL UNIMPL nosys 282 AUE_NULL UNIMPL nosys 283 AUE_NULL UNIMPL nosys 284 AUE_NULL UNIMPL nosys 285 AUE_NULL UNIMPL nosys 286 AUE_NULL UNIMPL nosys 287 AUE_NULL UNIMPL nosys 288 AUE_NULL UNIMPL nosys 289 AUE_PREADV STD { ssize_t freebsd32_preadv(int fd, \ struct iovec32 *iovp, \ u_int iovcnt, \ uint32_t offset1, uint32_t offset2); } 290 AUE_PWRITEV STD { ssize_t freebsd32_pwritev(int fd, \ struct iovec32 *iovp, \ u_int iovcnt, \ uint32_t offset1, uint32_t offset2); } 291 AUE_NULL UNIMPL nosys 292 AUE_NULL UNIMPL nosys 293 AUE_NULL UNIMPL nosys 294 AUE_NULL UNIMPL nosys 295 AUE_NULL UNIMPL nosys 296 AUE_NULL UNIMPL nosys 297 AUE_FHSTATFS COMPAT4 { int freebsd32_fhstatfs( \ const struct fhandle *u_fhp, \ struct statfs32 *buf); } 298 AUE_FHOPEN NOPROTO { int fhopen(const struct fhandle *u_fhp, \ int flags); } 299 AUE_FHSTAT COMPAT11 { int freebsd32_fhstat( \ const struct fhandle *u_fhp, \ struct freebsd11_stat32 *sb); } ; syscall numbers for FreeBSD 300 AUE_NULL NOPROTO { int modnext(int modid); } 301 AUE_NULL STD { int freebsd32_modstat(int modid, \ struct module_stat32* stat); } 302 AUE_NULL NOPROTO { int modfnext(int modid); } 303 AUE_NULL NOPROTO { int modfind(const char *name); } 304 AUE_MODLOAD NOPROTO { int kldload(const char *file); } 305 AUE_MODUNLOAD NOPROTO { int kldunload(int fileid); } 306 AUE_NULL NOPROTO { int kldfind(const char *file); } 307 AUE_NULL NOPROTO { int kldnext(int fileid); } 308 AUE_NULL STD { int freebsd32_kldstat(int fileid, \ struct kld32_file_stat* stat); } 309 AUE_NULL NOPROTO { int kldfirstmod(int fileid); } 310 AUE_GETSID NOPROTO { int getsid(pid_t pid); } 311 AUE_SETRESUID NOPROTO { int setresuid(uid_t ruid, uid_t euid, \ uid_t suid); } 312 AUE_SETRESGID NOPROTO { int setresgid(gid_t rgid, gid_t egid, \ gid_t sgid); } 313 AUE_NULL OBSOL signanosleep 314 AUE_AIO_RETURN STD { int freebsd32_aio_return( \ struct aiocb32 *aiocbp); } 315 AUE_AIO_SUSPEND STD { int freebsd32_aio_suspend( \ struct aiocb32 * const * aiocbp, int nent, \ const struct timespec32 *timeout); } 316 AUE_AIO_CANCEL NOPROTO { int aio_cancel(int fd, \ struct aiocb *aiocbp); } 317 AUE_AIO_ERROR STD { int freebsd32_aio_error( \ struct aiocb32 *aiocbp); } 318 AUE_AIO_READ COMPAT6 { int freebsd32_aio_read( \ struct oaiocb32 *aiocbp); } 319 AUE_AIO_WRITE COMPAT6 { int freebsd32_aio_write( \ struct oaiocb32 *aiocbp); } 320 AUE_LIO_LISTIO COMPAT6 { int freebsd32_lio_listio(int mode, \ struct oaiocb32 * const *acb_list, \ int nent, struct osigevent32 *sig); } 321 AUE_NULL NOPROTO { int yield(void); } 322 AUE_NULL OBSOL thr_sleep 323 AUE_NULL OBSOL thr_wakeup 324 AUE_MLOCKALL NOPROTO { int mlockall(int how); } 325 AUE_MUNLOCKALL NOPROTO { int munlockall(void); } 326 AUE_GETCWD NOPROTO { int __getcwd(char *buf, size_t buflen); } 327 AUE_NULL NOPROTO { int sched_setparam (pid_t pid, \ const struct sched_param *param); } 328 AUE_NULL NOPROTO { int sched_getparam (pid_t pid, \ struct sched_param *param); } 329 AUE_NULL NOPROTO { int sched_setscheduler (pid_t pid, \ int policy, \ const struct sched_param *param); } 330 AUE_NULL NOPROTO { int sched_getscheduler (pid_t pid); } 331 AUE_NULL NOPROTO { int sched_yield (void); } 332 AUE_NULL NOPROTO { int sched_get_priority_max (int policy); } 333 AUE_NULL NOPROTO { int sched_get_priority_min (int policy); } 334 AUE_NULL STD { int freebsd32_sched_rr_get_interval ( \ pid_t pid, \ struct timespec32 *interval); } 335 AUE_NULL NOPROTO { int utrace(const void *addr, size_t len); } 336 AUE_SENDFILE COMPAT4 { int freebsd32_sendfile(int fd, int s, \ uint32_t offset1, uint32_t offset2, \ size_t nbytes, struct sf_hdtr32 *hdtr, \ off_t *sbytes, int flags); } 337 AUE_NULL NOPROTO { int kldsym(int fileid, int cmd, \ void *data); } 338 AUE_JAIL STD { int freebsd32_jail(struct jail32 *jail); } 339 AUE_NULL UNIMPL pioctl 340 AUE_SIGPROCMASK NOPROTO { int sigprocmask(int how, \ const sigset_t *set, sigset_t *oset); } 341 AUE_SIGSUSPEND NOPROTO { int sigsuspend(const sigset_t *sigmask); } 342 AUE_SIGACTION COMPAT4 { int freebsd32_sigaction(int sig, \ struct sigaction32 *act, \ struct sigaction32 *oact); } 343 AUE_SIGPENDING NOPROTO { int sigpending(sigset_t *set); } 344 AUE_SIGRETURN COMPAT4 { int freebsd32_sigreturn( \ const struct freebsd4_freebsd32_ucontext *sigcntxp); } 345 AUE_SIGWAIT STD { int freebsd32_sigtimedwait(const sigset_t *set, \ siginfo_t *info, \ const struct timespec *timeout); } 346 AUE_NULL STD { int freebsd32_sigwaitinfo(const sigset_t *set, \ siginfo_t *info); } 347 AUE_ACL_GET_FILE NOPROTO { int __acl_get_file(const char *path, \ acl_type_t type, struct acl *aclp); } 348 AUE_ACL_SET_FILE NOPROTO { int __acl_set_file(const char *path, \ acl_type_t type, struct acl *aclp); } 349 AUE_ACL_GET_FD NOPROTO { int __acl_get_fd(int filedes, \ acl_type_t type, struct acl *aclp); } 350 AUE_ACL_SET_FD NOPROTO { int __acl_set_fd(int filedes, \ acl_type_t type, struct acl *aclp); } 351 AUE_ACL_DELETE_FILE NOPROTO { int __acl_delete_file(const char *path, \ acl_type_t type); } 352 AUE_ACL_DELETE_FD NOPROTO { int __acl_delete_fd(int filedes, \ acl_type_t type); } 353 AUE_ACL_CHECK_FILE NOPROTO { int __acl_aclcheck_file(const char *path, \ acl_type_t type, struct acl *aclp); } 354 AUE_ACL_CHECK_FD NOPROTO { int __acl_aclcheck_fd(int filedes, \ acl_type_t type, struct acl *aclp); } 355 AUE_EXTATTRCTL NOPROTO { int extattrctl(const char *path, int cmd, \ const char *filename, int attrnamespace, \ const char *attrname); } 356 AUE_EXTATTR_SET_FILE NOPROTO { ssize_t extattr_set_file( \ const char *path, int attrnamespace, \ const char *attrname, void *data, \ size_t nbytes); } 357 AUE_EXTATTR_GET_FILE NOPROTO { ssize_t extattr_get_file( \ const char *path, int attrnamespace, \ const char *attrname, void *data, \ size_t nbytes); } 358 AUE_EXTATTR_DELETE_FILE NOPROTO { int extattr_delete_file( \ const char *path, int attrnamespace, \ const char *attrname); } 359 AUE_AIO_WAITCOMPLETE STD { int freebsd32_aio_waitcomplete( \ struct aiocb32 **aiocbp, \ struct timespec32 *timeout); } 360 AUE_GETRESUID NOPROTO { int getresuid(uid_t *ruid, uid_t *euid, \ uid_t *suid); } 361 AUE_GETRESGID NOPROTO { int getresgid(gid_t *rgid, gid_t *egid, \ gid_t *sgid); } 362 AUE_KQUEUE NOPROTO { int kqueue(void); } 363 AUE_KEVENT COMPAT11 { int freebsd32_kevent(int fd, \ const struct kevent32_freebsd11 * \ changelist, \ int nchanges, \ struct kevent32_freebsd11 *eventlist, \ int nevents, \ const struct timespec32 *timeout); } 364 AUE_NULL OBSOL __cap_get_proc 365 AUE_NULL OBSOL __cap_set_proc 366 AUE_NULL OBSOL __cap_get_fd 367 AUE_NULL OBSOL __cap_get_file 368 AUE_NULL OBSOL __cap_set_fd 369 AUE_NULL OBSOL __cap_set_file 370 AUE_NULL UNIMPL nosys 371 AUE_EXTATTR_SET_FD NOPROTO { ssize_t extattr_set_fd(int fd, \ int attrnamespace, const char *attrname, \ void *data, size_t nbytes); } 372 AUE_EXTATTR_GET_FD NOPROTO { ssize_t extattr_get_fd(int fd, \ int attrnamespace, const char *attrname, \ void *data, size_t nbytes); } 373 AUE_EXTATTR_DELETE_FD NOPROTO { int extattr_delete_fd(int fd, \ int attrnamespace, \ const char *attrname); } 374 AUE_SETUGID NOPROTO { int __setugid(int flag); } 375 AUE_NULL OBSOL nfsclnt 376 AUE_EACCESS NOPROTO { int eaccess(const char *path, int amode); } 377 AUE_NULL UNIMPL afs_syscall 378 AUE_NMOUNT STD { int freebsd32_nmount(struct iovec32 *iovp, \ unsigned int iovcnt, int flags); } 379 AUE_NULL OBSOL kse_exit 380 AUE_NULL OBSOL kse_wakeup 381 AUE_NULL OBSOL kse_create 382 AUE_NULL OBSOL kse_thr_interrupt 383 AUE_NULL OBSOL kse_release 384 AUE_NULL UNIMPL __mac_get_proc 385 AUE_NULL UNIMPL __mac_set_proc 386 AUE_NULL UNIMPL __mac_get_fd 387 AUE_NULL UNIMPL __mac_get_file 388 AUE_NULL UNIMPL __mac_set_fd 389 AUE_NULL UNIMPL __mac_set_file 390 AUE_NULL NOPROTO { int kenv(int what, const char *name, \ char *value, int len); } 391 AUE_LCHFLAGS NOPROTO { int lchflags(const char *path, \ u_long flags); } 392 AUE_NULL NOPROTO { int uuidgen(struct uuid *store, \ int count); } 393 AUE_SENDFILE STD { int freebsd32_sendfile(int fd, int s, \ uint32_t offset1, uint32_t offset2, \ size_t nbytes, struct sf_hdtr32 *hdtr, \ off_t *sbytes, int flags); } 394 AUE_NULL UNIMPL mac_syscall 395 AUE_GETFSSTAT COMPAT11|NOPROTO { int getfsstat( \ struct freebsd11_statfs *buf, \ long bufsize, int mode); } 396 AUE_STATFS COMPAT11|NOPROTO { int statfs(const char *path, \ struct statfs *buf); } 397 AUE_FSTATFS COMPAT11|NOPROTO { int fstatfs(int fd, \ struct freebsd11_statfs *buf); } 398 AUE_FHSTATFS COMPAT11|NOPROTO { int fhstatfs( \ const struct fhandle *u_fhp, \ struct freebsd11_statfs *buf); } 399 AUE_NULL UNIMPL nosys 400 AUE_SEMCLOSE NOSTD|NOPROTO { int ksem_close(semid_t id); } 401 AUE_SEMPOST NOSTD|NOPROTO { int ksem_post(semid_t id); } 402 AUE_SEMWAIT NOSTD|NOPROTO { int ksem_wait(semid_t id); } 403 AUE_SEMTRYWAIT NOSTD|NOPROTO { int ksem_trywait(semid_t id); } 404 AUE_SEMINIT NOSTD { int freebsd32_ksem_init(semid_t *idp, \ unsigned int value); } 405 AUE_SEMOPEN NOSTD { int freebsd32_ksem_open(semid_t *idp, \ const char *name, int oflag, \ mode_t mode, unsigned int value); } 406 AUE_SEMUNLINK NOSTD|NOPROTO { int ksem_unlink(const char *name); } 407 AUE_SEMGETVALUE NOSTD|NOPROTO { int ksem_getvalue(semid_t id, \ int *val); } 408 AUE_SEMDESTROY NOSTD|NOPROTO { int ksem_destroy(semid_t id); } 409 AUE_NULL UNIMPL __mac_get_pid 410 AUE_NULL UNIMPL __mac_get_link 411 AUE_NULL UNIMPL __mac_set_link 412 AUE_EXTATTR_SET_LINK NOPROTO { ssize_t extattr_set_link( \ const char *path, int attrnamespace, \ const char *attrname, void *data, \ size_t nbytes); } 413 AUE_EXTATTR_GET_LINK NOPROTO { ssize_t extattr_get_link( \ const char *path, int attrnamespace, \ const char *attrname, void *data, \ size_t nbytes); } 414 AUE_EXTATTR_DELETE_LINK NOPROTO { int extattr_delete_link( \ const char *path, int attrnamespace, \ const char *attrname); } 415 AUE_NULL UNIMPL __mac_execve 416 AUE_SIGACTION STD { int freebsd32_sigaction(int sig, \ struct sigaction32 *act, \ struct sigaction32 *oact); } 417 AUE_SIGRETURN STD { int freebsd32_sigreturn( \ const struct freebsd32_ucontext *sigcntxp); } 418 AUE_NULL UNIMPL __xstat 419 AUE_NULL UNIMPL __xfstat 420 AUE_NULL UNIMPL __xlstat 421 AUE_NULL STD { int freebsd32_getcontext( \ struct freebsd32_ucontext *ucp); } 422 AUE_NULL STD { int freebsd32_setcontext( \ const struct freebsd32_ucontext *ucp); } 423 AUE_NULL STD { int freebsd32_swapcontext( \ struct freebsd32_ucontext *oucp, \ const struct freebsd32_ucontext *ucp); } 424 AUE_SWAPOFF UNIMPL swapoff 425 AUE_ACL_GET_LINK NOPROTO { int __acl_get_link(const char *path, \ acl_type_t type, struct acl *aclp); } 426 AUE_ACL_SET_LINK NOPROTO { int __acl_set_link(const char *path, \ acl_type_t type, struct acl *aclp); } 427 AUE_ACL_DELETE_LINK NOPROTO { int __acl_delete_link(const char *path, \ acl_type_t type); } 428 AUE_ACL_CHECK_LINK NOPROTO { int __acl_aclcheck_link(const char *path, \ acl_type_t type, struct acl *aclp); } 429 AUE_SIGWAIT NOPROTO { int sigwait(const sigset_t *set, \ int *sig); } 430 AUE_THR_CREATE UNIMPL thr_create; 431 AUE_THR_EXIT NOPROTO { void thr_exit(long *state); } 432 AUE_NULL NOPROTO { int thr_self(long *id); } 433 AUE_THR_KILL NOPROTO { int thr_kill(long id, int sig); } 434 AUE_NULL UNIMPL nosys 435 AUE_NULL UNIMPL nosys 436 AUE_JAIL_ATTACH NOPROTO { int jail_attach(int jid); } 437 AUE_EXTATTR_LIST_FD NOPROTO { ssize_t extattr_list_fd(int fd, \ int attrnamespace, void *data, \ size_t nbytes); } 438 AUE_EXTATTR_LIST_FILE NOPROTO { ssize_t extattr_list_file( \ const char *path, int attrnamespace, \ void *data, size_t nbytes); } 439 AUE_EXTATTR_LIST_LINK NOPROTO { ssize_t extattr_list_link( \ const char *path, int attrnamespace, \ void *data, size_t nbytes); } 440 AUE_NULL OBSOL kse_switchin 441 AUE_SEMWAIT NOSTD { int freebsd32_ksem_timedwait(semid_t id, \ const struct timespec32 *abstime); } 442 AUE_NULL STD { int freebsd32_thr_suspend( \ const struct timespec32 *timeout); } 443 AUE_NULL NOPROTO { int thr_wake(long id); } 444 AUE_MODUNLOAD NOPROTO { int kldunloadf(int fileid, int flags); } 445 AUE_AUDIT NOPROTO { int audit(const void *record, \ u_int length); } 446 AUE_AUDITON NOPROTO { int auditon(int cmd, void *data, \ u_int length); } 447 AUE_GETAUID NOPROTO { int getauid(uid_t *auid); } 448 AUE_SETAUID NOPROTO { int setauid(uid_t *auid); } 449 AUE_GETAUDIT NOPROTO { int getaudit(struct auditinfo *auditinfo); } 450 AUE_SETAUDIT NOPROTO { int setaudit(struct auditinfo *auditinfo); } 451 AUE_GETAUDIT_ADDR NOPROTO { int getaudit_addr( \ struct auditinfo_addr *auditinfo_addr, \ u_int length); } 452 AUE_SETAUDIT_ADDR NOPROTO { int setaudit_addr( \ struct auditinfo_addr *auditinfo_addr, \ u_int length); } 453 AUE_AUDITCTL NOPROTO { int auditctl(const char *path); } 454 AUE_NULL STD { int freebsd32__umtx_op(void *obj, int op,\ u_long val, void *uaddr, \ void *uaddr2); } 455 AUE_THR_NEW STD { int freebsd32_thr_new( \ struct thr_param32 *param, \ int param_size); } 456 AUE_NULL STD { int freebsd32_sigqueue(pid_t pid, \ int signum, int value); } 457 AUE_MQ_OPEN NOSTD { int freebsd32_kmq_open( \ const char *path, int flags, mode_t mode, \ const struct mq_attr32 *attr); } 458 AUE_MQ_SETATTR NOSTD { int freebsd32_kmq_setattr(int mqd, \ const struct mq_attr32 *attr, \ struct mq_attr32 *oattr); } 459 AUE_MQ_TIMEDRECEIVE NOSTD { int freebsd32_kmq_timedreceive(int mqd, \ char *msg_ptr, size_t msg_len, \ unsigned *msg_prio, \ const struct timespec32 *abs_timeout); } 460 AUE_MQ_TIMEDSEND NOSTD { int freebsd32_kmq_timedsend(int mqd, \ const char *msg_ptr, size_t msg_len,\ unsigned msg_prio, \ const struct timespec32 *abs_timeout);} 461 AUE_MQ_NOTIFY NOSTD { int freebsd32_kmq_notify(int mqd, \ const struct sigevent32 *sigev); } 462 AUE_MQ_UNLINK NOPROTO|NOSTD { int kmq_unlink(const char *path); } 463 AUE_NULL NOPROTO { int abort2(const char *why, int nargs, void **args); } 464 AUE_NULL NOPROTO { int thr_set_name(long id, const char *name); } 465 AUE_AIO_FSYNC STD { int freebsd32_aio_fsync(int op, \ struct aiocb32 *aiocbp); } 466 AUE_RTPRIO NOPROTO { int rtprio_thread(int function, \ lwpid_t lwpid, struct rtprio *rtp); } 467 AUE_NULL UNIMPL nosys 468 AUE_NULL UNIMPL nosys 469 AUE_NULL UNIMPL __getpath_fromfd 470 AUE_NULL UNIMPL __getpath_fromaddr 471 AUE_SCTP_PEELOFF NOPROTO|NOSTD { int sctp_peeloff(int sd, uint32_t name); } 472 AUE_SCTP_GENERIC_SENDMSG NOPROTO|NOSTD { int sctp_generic_sendmsg( \ int sd, void *msg, int mlen, \ struct sockaddr *to, __socklen_t tolen, \ struct sctp_sndrcvinfo *sinfo, int flags); } 473 AUE_SCTP_GENERIC_SENDMSG_IOV NOPROTO|NOSTD { int sctp_generic_sendmsg_iov(int sd, struct iovec *iov, int iovlen, \ struct sockaddr *to, __socklen_t tolen, \ struct sctp_sndrcvinfo *sinfo, int flags); } 474 AUE_SCTP_GENERIC_RECVMSG NOPROTO|NOSTD { int sctp_generic_recvmsg(int sd, struct iovec *iov, int iovlen, \ struct sockaddr * from, __socklen_t *fromlenaddr, \ struct sctp_sndrcvinfo *sinfo, int *msg_flags); } #ifdef PAD64_REQUIRED 475 AUE_PREAD STD { ssize_t freebsd32_pread(int fd, \ void *buf,size_t nbyte, \ int pad, \ uint32_t offset1, uint32_t offset2); } 476 AUE_PWRITE STD { ssize_t freebsd32_pwrite(int fd, \ const void *buf, size_t nbyte, \ int pad, \ uint32_t offset1, uint32_t offset2); } 477 AUE_MMAP STD { void *freebsd32_mmap(void *addr, \ size_t len, int prot, int flags, int fd, \ int pad, \ uint32_t pos1, uint32_t pos2); } 478 AUE_LSEEK STD { off_t freebsd32_lseek(int fd, \ int pad, \ uint32_t offset1, uint32_t offset2, \ int whence); } 479 AUE_TRUNCATE STD { int freebsd32_truncate(const char *path, \ int pad, \ uint32_t length1, uint32_t length2); } 480 AUE_FTRUNCATE STD { int freebsd32_ftruncate(int fd, \ int pad, \ uint32_t length1, uint32_t length2); } #else 475 AUE_PREAD STD { ssize_t freebsd32_pread(int fd, \ void *buf,size_t nbyte, \ uint32_t offset1, uint32_t offset2); } 476 AUE_PWRITE STD { ssize_t freebsd32_pwrite(int fd, \ const void *buf, size_t nbyte, \ uint32_t offset1, uint32_t offset2); } 477 AUE_MMAP STD { void *freebsd32_mmap(void *addr, \ size_t len, int prot, int flags, int fd, \ uint32_t pos1, uint32_t pos2); } 478 AUE_LSEEK STD { off_t freebsd32_lseek(int fd, \ uint32_t offset1, uint32_t offset2, \ int whence); } 479 AUE_TRUNCATE STD { int freebsd32_truncate(const char *path, \ uint32_t length1, uint32_t length2); } 480 AUE_FTRUNCATE STD { int freebsd32_ftruncate(int fd, \ uint32_t length1, uint32_t length2); } #endif 481 AUE_THR_KILL2 NOPROTO { int thr_kill2(pid_t pid, long id, int sig); } 482 AUE_SHMOPEN NOPROTO { int shm_open(const char *path, int flags, \ mode_t mode); } 483 AUE_SHMUNLINK NOPROTO { int shm_unlink(const char *path); } 484 AUE_NULL NOPROTO { int cpuset(cpusetid_t *setid); } #ifdef PAD64_REQUIRED 485 AUE_NULL STD { int freebsd32_cpuset_setid(cpuwhich_t which, \ int pad, \ uint32_t id1, uint32_t id2, \ cpusetid_t setid); } #else 485 AUE_NULL STD { int freebsd32_cpuset_setid(cpuwhich_t which, \ uint32_t id1, uint32_t id2, \ cpusetid_t setid); } #endif 486 AUE_NULL STD { int freebsd32_cpuset_getid(cpulevel_t level, \ cpuwhich_t which, \ uint32_t id1, uint32_t id2, \ cpusetid_t *setid); } 487 AUE_NULL STD { int freebsd32_cpuset_getaffinity( \ cpulevel_t level, cpuwhich_t which, \ uint32_t id1, uint32_t id2, \ size_t cpusetsize, \ cpuset_t *mask); } 488 AUE_NULL STD { int freebsd32_cpuset_setaffinity( \ cpulevel_t level, cpuwhich_t which, \ uint32_t id1, uint32_t id2, \ size_t cpusetsize, \ const cpuset_t *mask); } 489 AUE_FACCESSAT NOPROTO { int faccessat(int fd, const char *path, \ int amode, int flag); } 490 AUE_FCHMODAT NOPROTO { int fchmodat(int fd, const char *path, \ mode_t mode, int flag); } 491 AUE_FCHOWNAT NOPROTO { int fchownat(int fd, const char *path, \ uid_t uid, gid_t gid, int flag); } 492 AUE_FEXECVE STD { int freebsd32_fexecve(int fd, \ uint32_t *argv, uint32_t *envv); } 493 AUE_FSTATAT COMPAT11 { int freebsd32_fstatat(int fd, \ const char *path, \ struct freebsd11_stat32 *buf, \ int flag); } 494 AUE_FUTIMESAT STD { int freebsd32_futimesat(int fd, \ const char *path, \ struct timeval *times); } 495 AUE_LINKAT NOPROTO { int linkat(int fd1, const char *path1, \ int fd2, const char *path2, int flag); } 496 AUE_MKDIRAT NOPROTO { int mkdirat(int fd, const char *path, \ mode_t mode); } 497 AUE_MKFIFOAT NOPROTO { int mkfifoat(int fd, const char *path, \ mode_t mode); } 498 AUE_MKNODAT COMPAT11|NOPROTO { int mknodat(int fd, \ const char *path, mode_t mode, \ uint32_t dev); } 499 AUE_OPENAT_RWTC NOPROTO { int openat(int fd, const char *path, \ int flag, mode_t mode); } 500 AUE_READLINKAT NOPROTO { int readlinkat(int fd, const char *path, \ char *buf, size_t bufsize); } 501 AUE_RENAMEAT NOPROTO { int renameat(int oldfd, const char *old, \ int newfd, const char *new); } 502 AUE_SYMLINKAT NOPROTO { int symlinkat(const char *path1, int fd, \ const char *path2); } 503 AUE_UNLINKAT NOPROTO { int unlinkat(int fd, const char *path, \ int flag); } 504 AUE_POSIX_OPENPT NOPROTO { int posix_openpt(int flags); } ; 505 is initialised by the kgssapi code, if present. 505 AUE_NULL UNIMPL gssd_syscall 506 AUE_JAIL_GET STD { int freebsd32_jail_get(struct iovec32 *iovp, \ unsigned int iovcnt, int flags); } 507 AUE_JAIL_SET STD { int freebsd32_jail_set(struct iovec32 *iovp, \ unsigned int iovcnt, int flags); } 508 AUE_JAIL_REMOVE NOPROTO { int jail_remove(int jid); } 509 AUE_CLOSEFROM NOPROTO { int closefrom(int lowfd); } 510 AUE_SEMCTL NOSTD { int freebsd32_semctl(int semid, int semnum, \ int cmd, union semun32 *arg); } 511 AUE_MSGCTL NOSTD { int freebsd32_msgctl(int msqid, int cmd, \ struct msqid_ds32 *buf); } 512 AUE_SHMCTL NOSTD { int freebsd32_shmctl(int shmid, int cmd, \ struct shmid_ds32 *buf); } 513 AUE_LPATHCONF NOPROTO { int lpathconf(const char *path, int name); } 514 AUE_NULL OBSOL cap_new 515 AUE_CAP_RIGHTS_GET NOPROTO { int __cap_rights_get(int version, \ int fd, cap_rights_t *rightsp); } 516 AUE_CAP_ENTER NOPROTO { int cap_enter(void); } 517 AUE_CAP_GETMODE NOPROTO { int cap_getmode(u_int *modep); } 518 AUE_PDFORK NOPROTO { int pdfork(int *fdp, int flags); } 519 AUE_PDKILL NOPROTO { int pdkill(int fd, int signum); } 520 AUE_PDGETPID NOPROTO { int pdgetpid(int fd, pid_t *pidp); } 521 AUE_PDWAIT UNIMPL pdwait4 522 AUE_SELECT STD { int freebsd32_pselect(int nd, fd_set *in, \ fd_set *ou, fd_set *ex, \ const struct timespec32 *ts, \ const sigset_t *sm); } 523 AUE_GETLOGINCLASS NOPROTO { int getloginclass(char *namebuf, \ size_t namelen); } 524 AUE_SETLOGINCLASS NOPROTO { int setloginclass(const char *namebuf); } 525 AUE_NULL NOPROTO { int rctl_get_racct(const void *inbufp, \ size_t inbuflen, void *outbufp, \ size_t outbuflen); } 526 AUE_NULL NOPROTO { int rctl_get_rules(const void *inbufp, \ size_t inbuflen, void *outbufp, \ size_t outbuflen); } 527 AUE_NULL NOPROTO { int rctl_get_limits(const void *inbufp, \ size_t inbuflen, void *outbufp, \ size_t outbuflen); } 528 AUE_NULL NOPROTO { int rctl_add_rule(const void *inbufp, \ size_t inbuflen, void *outbufp, \ size_t outbuflen); } 529 AUE_NULL NOPROTO { int rctl_remove_rule(const void *inbufp, \ size_t inbuflen, void *outbufp, \ size_t outbuflen); } #ifdef PAD64_REQUIRED 530 AUE_POSIX_FALLOCATE STD { int freebsd32_posix_fallocate(int fd, \ int pad, \ uint32_t offset1, uint32_t offset2,\ uint32_t len1, uint32_t len2); } 531 AUE_POSIX_FADVISE STD { int freebsd32_posix_fadvise(int fd, \ int pad, \ uint32_t offset1, uint32_t offset2,\ uint32_t len1, uint32_t len2, \ int advice); } 532 AUE_WAIT6 STD { int freebsd32_wait6(int idtype, int pad, \ uint32_t id1, uint32_t id2, \ int *status, int options, \ struct wrusage32 *wrusage, \ siginfo_t *info); } #else 530 AUE_POSIX_FALLOCATE STD { int freebsd32_posix_fallocate(int fd,\ uint32_t offset1, uint32_t offset2,\ uint32_t len1, uint32_t len2); } 531 AUE_POSIX_FADVISE STD { int freebsd32_posix_fadvise(int fd, \ uint32_t offset1, uint32_t offset2,\ uint32_t len1, uint32_t len2, \ int advice); } 532 AUE_WAIT6 STD { int freebsd32_wait6(int idtype, \ uint32_t id1, uint32_t id2, \ int *status, int options, \ struct wrusage32 *wrusage, \ siginfo_t *info); } #endif 533 AUE_CAP_RIGHTS_LIMIT NOPROTO { \ int cap_rights_limit(int fd, \ cap_rights_t *rightsp); } 534 AUE_CAP_IOCTLS_LIMIT STD { \ int freebsd32_cap_ioctls_limit(int fd, \ const uint32_t *cmds, size_t ncmds); } 535 AUE_CAP_IOCTLS_GET STD { \ ssize_t freebsd32_cap_ioctls_get(int fd, \ uint32_t *cmds, size_t maxcmds); } 536 AUE_CAP_FCNTLS_LIMIT NOPROTO { int cap_fcntls_limit(int fd, \ uint32_t fcntlrights); } 537 AUE_CAP_FCNTLS_GET NOPROTO { int cap_fcntls_get(int fd, \ uint32_t *fcntlrightsp); } 538 AUE_BINDAT NOPROTO { int bindat(int fd, int s, \ const struct sockaddr *name, \ int namelen); } 539 AUE_CONNECTAT NOPROTO { int connectat(int fd, int s, \ const struct sockaddr *name, \ int namelen); } 540 AUE_CHFLAGSAT NOPROTO { int chflagsat(int fd, const char *path, \ u_long flags, int atflag); } 541 AUE_ACCEPT NOPROTO { int accept4(int s, \ struct sockaddr *name, \ __socklen_t *anamelen, \ int flags); } 542 AUE_PIPE NOPROTO { int pipe2(int *fildes, int flags); } 543 AUE_AIO_MLOCK STD { int freebsd32_aio_mlock( \ struct aiocb32 *aiocbp); } #ifdef PAD64_REQUIRED 544 AUE_PROCCTL STD { int freebsd32_procctl(int idtype, int pad, \ uint32_t id1, uint32_t id2, int com, \ void *data); } #else 544 AUE_PROCCTL STD { int freebsd32_procctl(int idtype, \ uint32_t id1, uint32_t id2, int com, \ void *data); } #endif 545 AUE_POLL STD { int freebsd32_ppoll(struct pollfd *fds, \ u_int nfds, const struct timespec32 *ts, \ const sigset_t *set); } 546 AUE_FUTIMES STD { int freebsd32_futimens(int fd, \ struct timespec *times); } 547 AUE_FUTIMESAT STD { int freebsd32_utimensat(int fd, \ const char *path, \ struct timespec *times, int flag); } 548 AUE_NULL OBSOL numa_getaffinity 549 AUE_NULL OBSOL numa_setaffinity 550 AUE_FSYNC NOPROTO { int fdatasync(int fd); } 551 AUE_FSTAT STD { int freebsd32_fstat(int fd, \ struct stat32 *ub); } 552 AUE_FSTATAT STD { int freebsd32_fstatat(int fd, \ const char *path, struct stat32 *buf, \ int flag); } 553 AUE_FHSTAT STD { int freebsd32_fhstat( \ const struct fhandle *u_fhp, \ struct stat32 *sb); } 554 AUE_GETDIRENTRIES NOPROTO { ssize_t getdirentries( \ int fd, char *buf, size_t count, \ off_t *basep); } 555 AUE_STATFS NOPROTO { int statfs(const char *path, \ struct statfs32 *buf); } 556 AUE_FSTATFS NOPROTO { int fstatfs(int fd, struct statfs32 *buf); } 557 AUE_GETFSSTAT NOPROTO { int getfsstat(struct statfs32 *buf, \ long bufsize, int mode); } 558 AUE_FHSTATFS NOPROTO { int fhstatfs(const struct fhandle *u_fhp, \ struct statfs32 *buf); } #ifdef PAD64_REQUIRED 559 AUE_MKNODAT STD { int freebsd32_mknodat(int fd, \ const char *path, mode_t mode, \ int pad, uint32_t dev1, uint32_t dev2); } #else 559 AUE_MKNODAT STD { int freebsd32_mknodat(int fd, \ const char *path, mode_t mode, \ uint32_t dev1, uint32_t dev2); } #endif 560 AUE_KEVENT STD { int freebsd32_kevent(int fd, \ const struct kevent32 *changelist, \ int nchanges, \ struct kevent32 *eventlist, \ int nevents, \ const struct timespec32 *timeout); } 561 AUE_NULL STD { int freebsd32_cpuset_getdomain(cpulevel_t level, \ cpuwhich_t which, uint32_t id1, uint32_t id2, \ size_t domainsetsize, domainset_t *mask, \ int *policy); } 562 AUE_NULL STD { int freebsd32_cpuset_setdomain(cpulevel_t level, \ cpuwhich_t which, uint32_t id1, uint32_t id2, \ size_t domainsetsize, domainset_t *mask, \ int policy); } 563 AUE_NULL NOPROTO { int getrandom(void *buf, size_t buflen, \ unsigned int flags); } 564 AUE_NULL NOPROTO { int getfhat( int fd, char *path, \ struct fhandle *fhp, int flags); } 565 AUE_NULL NOPROTO { int fhlink( struct fhandle *fhp, const char *to ); } 566 AUE_NULL NOPROTO { int fhlinkat( struct fhandle *fhp, int tofd, \ const char *to); } 567 AUE_NULL NOPROTO { int fhreadlink( struct fhandle *fhp, char *buf, \ size_t bufsize); } +568 AUE_UNLINKAT NOPROTO { int funlinkat(int dfd, const char *path, int fd, \ + int flag); } ; vim: syntax=off Index: head/sys/compat/linux/linux_file.c =================================================================== --- head/sys/compat/linux/linux_file.c (revision 345981) +++ head/sys/compat/linux/linux_file.c (revision 345982) @@ -1,1690 +1,1691 @@ /*- * SPDX-License-Identifier: BSD-2-Clause-FreeBSD * * Copyright (c) 1994-1995 Søren Schmidt * 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 "opt_compat.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef COMPAT_LINUX32 #include #include #else #include #include #endif #include #include #include static int linux_common_open(struct thread *, int, char *, int, int); static int linux_getdents_error(struct thread *, int, int); #ifdef LINUX_LEGACY_SYSCALLS int linux_creat(struct thread *td, struct linux_creat_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(creat)) printf(ARGS(creat, "%s, %d"), path, args->mode); #endif error = kern_openat(td, AT_FDCWD, path, UIO_SYSSPACE, O_WRONLY | O_CREAT | O_TRUNC, args->mode); LFREEPATH(path); return (error); } #endif static int linux_common_open(struct thread *td, int dirfd, char *path, int l_flags, int mode) { struct proc *p = td->td_proc; struct file *fp; int fd; int bsd_flags, error; bsd_flags = 0; switch (l_flags & LINUX_O_ACCMODE) { case LINUX_O_WRONLY: bsd_flags |= O_WRONLY; break; case LINUX_O_RDWR: bsd_flags |= O_RDWR; break; default: bsd_flags |= O_RDONLY; } if (l_flags & LINUX_O_NDELAY) bsd_flags |= O_NONBLOCK; if (l_flags & LINUX_O_APPEND) bsd_flags |= O_APPEND; if (l_flags & LINUX_O_SYNC) bsd_flags |= O_FSYNC; if (l_flags & LINUX_O_NONBLOCK) bsd_flags |= O_NONBLOCK; if (l_flags & LINUX_FASYNC) bsd_flags |= O_ASYNC; if (l_flags & LINUX_O_CREAT) bsd_flags |= O_CREAT; if (l_flags & LINUX_O_TRUNC) bsd_flags |= O_TRUNC; if (l_flags & LINUX_O_EXCL) bsd_flags |= O_EXCL; if (l_flags & LINUX_O_NOCTTY) bsd_flags |= O_NOCTTY; if (l_flags & LINUX_O_DIRECT) bsd_flags |= O_DIRECT; if (l_flags & LINUX_O_NOFOLLOW) bsd_flags |= O_NOFOLLOW; if (l_flags & LINUX_O_DIRECTORY) bsd_flags |= O_DIRECTORY; /* XXX LINUX_O_NOATIME: unable to be easily implemented. */ error = kern_openat(td, dirfd, path, UIO_SYSSPACE, bsd_flags, mode); if (error != 0) goto done; if (bsd_flags & O_NOCTTY) goto done; /* * XXX In between kern_openat() and fget(), another process * having the same filedesc could use that fd without * checking below. */ fd = td->td_retval[0]; if (fget(td, fd, &cap_ioctl_rights, &fp) == 0) { if (fp->f_type != DTYPE_VNODE) { fdrop(fp, td); goto done; } sx_slock(&proctree_lock); PROC_LOCK(p); if (SESS_LEADER(p) && !(p->p_flag & P_CONTROLT)) { PROC_UNLOCK(p); sx_sunlock(&proctree_lock); /* XXXPJD: Verify if TIOCSCTTY is allowed. */ (void) fo_ioctl(fp, TIOCSCTTY, (caddr_t) 0, td->td_ucred, td); } else { PROC_UNLOCK(p); sx_sunlock(&proctree_lock); } fdrop(fp, td); } done: #ifdef DEBUG #ifdef LINUX_LEGACY_SYSCALLS if (ldebug(open)) #else if (ldebug(openat)) #endif printf(LMSG("open returns error %d"), error); #endif LFREEPATH(path); return (error); } int linux_openat(struct thread *td, struct linux_openat_args *args) { char *path; int dfd; dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; if (args->flags & LINUX_O_CREAT) LCONVPATH_AT(td, args->filename, &path, 1, dfd); else LCONVPATH_AT(td, args->filename, &path, 0, dfd); #ifdef DEBUG if (ldebug(openat)) printf(ARGS(openat, "%i, %s, 0x%x, 0x%x"), args->dfd, path, args->flags, args->mode); #endif return (linux_common_open(td, dfd, path, args->flags, args->mode)); } #ifdef LINUX_LEGACY_SYSCALLS int linux_open(struct thread *td, struct linux_open_args *args) { char *path; if (args->flags & LINUX_O_CREAT) LCONVPATHCREAT(td, args->path, &path); else LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(open)) printf(ARGS(open, "%s, 0x%x, 0x%x"), path, args->flags, args->mode); #endif return (linux_common_open(td, AT_FDCWD, path, args->flags, args->mode)); } #endif int linux_lseek(struct thread *td, struct linux_lseek_args *args) { #ifdef DEBUG if (ldebug(lseek)) printf(ARGS(lseek, "%d, %ld, %d"), args->fdes, (long)args->off, args->whence); #endif return (kern_lseek(td, args->fdes, args->off, args->whence)); } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) int linux_llseek(struct thread *td, struct linux_llseek_args *args) { int error; off_t off; #ifdef DEBUG if (ldebug(llseek)) printf(ARGS(llseek, "%d, %d:%d, %d"), args->fd, args->ohigh, args->olow, args->whence); #endif off = (args->olow) | (((off_t) args->ohigh) << 32); error = kern_lseek(td, args->fd, off, args->whence); if (error != 0) return (error); error = copyout(td->td_retval, args->res, sizeof(off_t)); if (error != 0) return (error); td->td_retval[0] = 0; return (0); } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ /* * Note that linux_getdents(2) and linux_getdents64(2) have the same * arguments. They only differ in the definition of struct dirent they * operate on. * Note that linux_readdir(2) is a special case of linux_getdents(2) * where count is always equals 1, meaning that the buffer is one * dirent-structure in size and that the code can't handle more anyway. * Note that linux_readdir(2) can't be implemented by means of linux_getdents(2) * as in case when the *dent buffer size is equal to 1 linux_getdents(2) will * trash user stack. */ static int linux_getdents_error(struct thread *td, int fd, int err) { struct vnode *vp; struct file *fp; int error; /* Linux return ENOTDIR in case when fd is not a directory. */ error = getvnode(td, fd, &cap_read_rights, &fp); if (error != 0) return (error); vp = fp->f_vnode; if (vp->v_type != VDIR) { fdrop(fp, td); return (ENOTDIR); } fdrop(fp, td); return (err); } struct l_dirent { l_ulong d_ino; l_off_t d_off; l_ushort d_reclen; char d_name[LINUX_NAME_MAX + 1]; }; struct l_dirent64 { uint64_t d_ino; int64_t d_off; l_ushort d_reclen; u_char d_type; char d_name[LINUX_NAME_MAX + 1]; }; /* * Linux uses the last byte in the dirent buffer to store d_type, * at least glibc-2.7 requires it. That is why l_dirent is padded with 2 bytes. */ #define LINUX_RECLEN(namlen) \ roundup(offsetof(struct l_dirent, d_name) + (namlen) + 2, sizeof(l_ulong)) #define LINUX_RECLEN64(namlen) \ roundup(offsetof(struct l_dirent64, d_name) + (namlen) + 1, \ sizeof(uint64_t)) #ifdef LINUX_LEGACY_SYSCALLS int linux_getdents(struct thread *td, struct linux_getdents_args *args) { struct dirent *bdp; caddr_t inp, buf; /* BSD-format */ int len, reclen; /* BSD-format */ caddr_t outp; /* Linux-format */ int resid, linuxreclen; /* Linux-format */ caddr_t lbuf; /* Linux-format */ off_t base; struct l_dirent *linux_dirent; int buflen, error; size_t retval; #ifdef DEBUG if (ldebug(getdents)) printf(ARGS(getdents, "%d, *, %d"), args->fd, args->count); #endif buflen = min(args->count, MAXBSIZE); buf = malloc(buflen, M_TEMP, M_WAITOK); error = kern_getdirentries(td, args->fd, buf, buflen, &base, NULL, UIO_SYSSPACE); if (error != 0) { error = linux_getdents_error(td, args->fd, error); goto out1; } lbuf = malloc(LINUX_RECLEN(LINUX_NAME_MAX), M_TEMP, M_WAITOK | M_ZERO); len = td->td_retval[0]; inp = buf; outp = (caddr_t)args->dent; resid = args->count; retval = 0; while (len > 0) { bdp = (struct dirent *) inp; reclen = bdp->d_reclen; linuxreclen = LINUX_RECLEN(bdp->d_namlen); /* * No more space in the user supplied dirent buffer. * Return EINVAL. */ if (resid < linuxreclen) { error = EINVAL; goto out; } linux_dirent = (struct l_dirent*)lbuf; linux_dirent->d_ino = bdp->d_fileno; linux_dirent->d_off = base + reclen; linux_dirent->d_reclen = linuxreclen; /* * Copy d_type to last byte of l_dirent buffer */ lbuf[linuxreclen - 1] = bdp->d_type; strlcpy(linux_dirent->d_name, bdp->d_name, linuxreclen - offsetof(struct l_dirent, d_name)-1); error = copyout(linux_dirent, outp, linuxreclen); if (error != 0) goto out; inp += reclen; base += reclen; len -= reclen; retval += linuxreclen; outp += linuxreclen; resid -= linuxreclen; } td->td_retval[0] = retval; out: free(lbuf, M_TEMP); out1: free(buf, M_TEMP); return (error); } #endif int linux_getdents64(struct thread *td, struct linux_getdents64_args *args) { struct dirent *bdp; caddr_t inp, buf; /* BSD-format */ int len, reclen; /* BSD-format */ caddr_t outp; /* Linux-format */ int resid, linuxreclen; /* Linux-format */ caddr_t lbuf; /* Linux-format */ off_t base; struct l_dirent64 *linux_dirent64; int buflen, error; size_t retval; #ifdef DEBUG if (ldebug(getdents64)) uprintf(ARGS(getdents64, "%d, *, %d"), args->fd, args->count); #endif buflen = min(args->count, MAXBSIZE); buf = malloc(buflen, M_TEMP, M_WAITOK); error = kern_getdirentries(td, args->fd, buf, buflen, &base, NULL, UIO_SYSSPACE); if (error != 0) { error = linux_getdents_error(td, args->fd, error); goto out1; } lbuf = malloc(LINUX_RECLEN64(LINUX_NAME_MAX), M_TEMP, M_WAITOK | M_ZERO); len = td->td_retval[0]; inp = buf; outp = (caddr_t)args->dirent; resid = args->count; retval = 0; while (len > 0) { bdp = (struct dirent *) inp; reclen = bdp->d_reclen; linuxreclen = LINUX_RECLEN64(bdp->d_namlen); /* * No more space in the user supplied dirent buffer. * Return EINVAL. */ if (resid < linuxreclen) { error = EINVAL; goto out; } linux_dirent64 = (struct l_dirent64*)lbuf; linux_dirent64->d_ino = bdp->d_fileno; linux_dirent64->d_off = base + reclen; linux_dirent64->d_reclen = linuxreclen; linux_dirent64->d_type = bdp->d_type; strlcpy(linux_dirent64->d_name, bdp->d_name, linuxreclen - offsetof(struct l_dirent64, d_name)); error = copyout(linux_dirent64, outp, linuxreclen); if (error != 0) goto out; inp += reclen; base += reclen; len -= reclen; retval += linuxreclen; outp += linuxreclen; resid -= linuxreclen; } td->td_retval[0] = retval; out: free(lbuf, M_TEMP); out1: free(buf, M_TEMP); return (error); } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) int linux_readdir(struct thread *td, struct linux_readdir_args *args) { struct dirent *bdp; caddr_t buf; /* BSD-format */ int linuxreclen; /* Linux-format */ caddr_t lbuf; /* Linux-format */ off_t base; struct l_dirent *linux_dirent; int buflen, error; #ifdef DEBUG if (ldebug(readdir)) printf(ARGS(readdir, "%d, *"), args->fd); #endif buflen = LINUX_RECLEN(LINUX_NAME_MAX); buf = malloc(buflen, M_TEMP, M_WAITOK); error = kern_getdirentries(td, args->fd, buf, buflen, &base, NULL, UIO_SYSSPACE); if (error != 0) { error = linux_getdents_error(td, args->fd, error); goto out; } if (td->td_retval[0] == 0) goto out; lbuf = malloc(LINUX_RECLEN(LINUX_NAME_MAX), M_TEMP, M_WAITOK | M_ZERO); bdp = (struct dirent *) buf; linuxreclen = LINUX_RECLEN(bdp->d_namlen); linux_dirent = (struct l_dirent*)lbuf; linux_dirent->d_ino = bdp->d_fileno; linux_dirent->d_off = linuxreclen; linux_dirent->d_reclen = bdp->d_namlen; strlcpy(linux_dirent->d_name, bdp->d_name, linuxreclen - offsetof(struct l_dirent, d_name)); error = copyout(linux_dirent, args->dent, linuxreclen); if (error == 0) td->td_retval[0] = linuxreclen; free(lbuf, M_TEMP); out: free(buf, M_TEMP); return (error); } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ /* * These exist mainly for hooks for doing /compat/linux translation. */ #ifdef LINUX_LEGACY_SYSCALLS int linux_access(struct thread *td, struct linux_access_args *args) { char *path; int error; /* Linux convention. */ if (args->amode & ~(F_OK | X_OK | W_OK | R_OK)) return (EINVAL); LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(access)) printf(ARGS(access, "%s, %d"), path, args->amode); #endif error = kern_accessat(td, AT_FDCWD, path, UIO_SYSSPACE, 0, args->amode); LFREEPATH(path); return (error); } #endif int linux_faccessat(struct thread *td, struct linux_faccessat_args *args) { char *path; int error, dfd; /* Linux convention. */ if (args->amode & ~(F_OK | X_OK | W_OK | R_OK)) return (EINVAL); dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; LCONVPATHEXIST_AT(td, args->filename, &path, dfd); #ifdef DEBUG if (ldebug(faccessat)) printf(ARGS(access, "%s, %d"), path, args->amode); #endif error = kern_accessat(td, dfd, path, UIO_SYSSPACE, 0, args->amode); LFREEPATH(path); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_unlink(struct thread *td, struct linux_unlink_args *args) { char *path; int error; struct stat st; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(unlink)) printf(ARGS(unlink, "%s"), path); #endif - error = kern_unlinkat(td, AT_FDCWD, path, UIO_SYSSPACE, 0, 0); + error = kern_funlinkat(td, AT_FDCWD, path, FD_NONE, UIO_SYSSPACE, 0, 0); if (error == EPERM) { /* Introduce POSIX noncompliant behaviour of Linux */ if (kern_statat(td, 0, AT_FDCWD, path, UIO_SYSSPACE, &st, NULL) == 0) { if (S_ISDIR(st.st_mode)) error = EISDIR; } } LFREEPATH(path); return (error); } #endif int linux_unlinkat(struct thread *td, struct linux_unlinkat_args *args) { char *path; int error, dfd; struct stat st; if (args->flag & ~LINUX_AT_REMOVEDIR) return (EINVAL); dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; LCONVPATHEXIST_AT(td, args->pathname, &path, dfd); #ifdef DEBUG if (ldebug(unlinkat)) printf(ARGS(unlinkat, "%s"), path); #endif if (args->flag & LINUX_AT_REMOVEDIR) - error = kern_rmdirat(td, dfd, path, UIO_SYSSPACE, 0); + error = kern_frmdirat(td, dfd, path, FD_NONE, UIO_SYSSPACE, 0); else - error = kern_unlinkat(td, dfd, path, UIO_SYSSPACE, 0, 0); + error = kern_funlinkat(td, dfd, path, FD_NONE, UIO_SYSSPACE, 0, + 0); if (error == EPERM && !(args->flag & LINUX_AT_REMOVEDIR)) { /* Introduce POSIX noncompliant behaviour of Linux */ if (kern_statat(td, AT_SYMLINK_NOFOLLOW, dfd, path, UIO_SYSSPACE, &st, NULL) == 0 && S_ISDIR(st.st_mode)) error = EISDIR; } LFREEPATH(path); return (error); } int linux_chdir(struct thread *td, struct linux_chdir_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(chdir)) printf(ARGS(chdir, "%s"), path); #endif error = kern_chdir(td, path, UIO_SYSSPACE); LFREEPATH(path); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_chmod(struct thread *td, struct linux_chmod_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(chmod)) printf(ARGS(chmod, "%s, %d"), path, args->mode); #endif error = kern_fchmodat(td, AT_FDCWD, path, UIO_SYSSPACE, args->mode, 0); LFREEPATH(path); return (error); } #endif int linux_fchmodat(struct thread *td, struct linux_fchmodat_args *args) { char *path; int error, dfd; dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; LCONVPATHEXIST_AT(td, args->filename, &path, dfd); #ifdef DEBUG if (ldebug(fchmodat)) printf(ARGS(fchmodat, "%s, %d"), path, args->mode); #endif error = kern_fchmodat(td, dfd, path, UIO_SYSSPACE, args->mode, 0); LFREEPATH(path); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_mkdir(struct thread *td, struct linux_mkdir_args *args) { char *path; int error; LCONVPATHCREAT(td, args->path, &path); #ifdef DEBUG if (ldebug(mkdir)) printf(ARGS(mkdir, "%s, %d"), path, args->mode); #endif error = kern_mkdirat(td, AT_FDCWD, path, UIO_SYSSPACE, args->mode); LFREEPATH(path); return (error); } #endif int linux_mkdirat(struct thread *td, struct linux_mkdirat_args *args) { char *path; int error, dfd; dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; LCONVPATHCREAT_AT(td, args->pathname, &path, dfd); #ifdef DEBUG if (ldebug(mkdirat)) printf(ARGS(mkdirat, "%s, %d"), path, args->mode); #endif error = kern_mkdirat(td, dfd, path, UIO_SYSSPACE, args->mode); LFREEPATH(path); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_rmdir(struct thread *td, struct linux_rmdir_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(rmdir)) printf(ARGS(rmdir, "%s"), path); #endif - error = kern_rmdirat(td, AT_FDCWD, path, UIO_SYSSPACE, 0); + error = kern_frmdirat(td, AT_FDCWD, path, FD_NONE, UIO_SYSSPACE, 0); LFREEPATH(path); return (error); } int linux_rename(struct thread *td, struct linux_rename_args *args) { char *from, *to; int error; LCONVPATHEXIST(td, args->from, &from); /* Expand LCONVPATHCREATE so that `from' can be freed on errors */ error = linux_emul_convpath(td, args->to, UIO_USERSPACE, &to, 1, AT_FDCWD); if (to == NULL) { LFREEPATH(from); return (error); } #ifdef DEBUG if (ldebug(rename)) printf(ARGS(rename, "%s, %s"), from, to); #endif error = kern_renameat(td, AT_FDCWD, from, AT_FDCWD, to, UIO_SYSSPACE); LFREEPATH(from); LFREEPATH(to); return (error); } #endif int linux_renameat(struct thread *td, struct linux_renameat_args *args) { char *from, *to; int error, olddfd, newdfd; olddfd = (args->olddfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->olddfd; newdfd = (args->newdfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->newdfd; LCONVPATHEXIST_AT(td, args->oldname, &from, olddfd); /* Expand LCONVPATHCREATE so that `from' can be freed on errors */ error = linux_emul_convpath(td, args->newname, UIO_USERSPACE, &to, 1, newdfd); if (to == NULL) { LFREEPATH(from); return (error); } #ifdef DEBUG if (ldebug(renameat)) printf(ARGS(renameat, "%s, %s"), from, to); #endif error = kern_renameat(td, olddfd, from, newdfd, to, UIO_SYSSPACE); LFREEPATH(from); LFREEPATH(to); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_symlink(struct thread *td, struct linux_symlink_args *args) { char *path, *to; int error; LCONVPATHEXIST(td, args->path, &path); /* Expand LCONVPATHCREATE so that `path' can be freed on errors */ error = linux_emul_convpath(td, args->to, UIO_USERSPACE, &to, 1, AT_FDCWD); if (to == NULL) { LFREEPATH(path); return (error); } #ifdef DEBUG if (ldebug(symlink)) printf(ARGS(symlink, "%s, %s"), path, to); #endif error = kern_symlinkat(td, path, AT_FDCWD, to, UIO_SYSSPACE); LFREEPATH(path); LFREEPATH(to); return (error); } #endif int linux_symlinkat(struct thread *td, struct linux_symlinkat_args *args) { char *path, *to; int error, dfd; dfd = (args->newdfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->newdfd; LCONVPATHEXIST(td, args->oldname, &path); /* Expand LCONVPATHCREATE so that `path' can be freed on errors */ error = linux_emul_convpath(td, args->newname, UIO_USERSPACE, &to, 1, dfd); if (to == NULL) { LFREEPATH(path); return (error); } #ifdef DEBUG if (ldebug(symlinkat)) printf(ARGS(symlinkat, "%s, %s"), path, to); #endif error = kern_symlinkat(td, path, dfd, to, UIO_SYSSPACE); LFREEPATH(path); LFREEPATH(to); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_readlink(struct thread *td, struct linux_readlink_args *args) { char *name; int error; LCONVPATHEXIST(td, args->name, &name); #ifdef DEBUG if (ldebug(readlink)) printf(ARGS(readlink, "%s, %p, %d"), name, (void *)args->buf, args->count); #endif error = kern_readlinkat(td, AT_FDCWD, name, UIO_SYSSPACE, args->buf, UIO_USERSPACE, args->count); LFREEPATH(name); return (error); } #endif int linux_readlinkat(struct thread *td, struct linux_readlinkat_args *args) { char *name; int error, dfd; dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; LCONVPATHEXIST_AT(td, args->path, &name, dfd); #ifdef DEBUG if (ldebug(readlinkat)) printf(ARGS(readlinkat, "%s, %p, %d"), name, (void *)args->buf, args->bufsiz); #endif error = kern_readlinkat(td, dfd, name, UIO_SYSSPACE, args->buf, UIO_USERSPACE, args->bufsiz); LFREEPATH(name); return (error); } int linux_truncate(struct thread *td, struct linux_truncate_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(truncate)) printf(ARGS(truncate, "%s, %ld"), path, (long)args->length); #endif error = kern_truncate(td, path, UIO_SYSSPACE, args->length); LFREEPATH(path); return (error); } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) int linux_truncate64(struct thread *td, struct linux_truncate64_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(truncate64)) printf(ARGS(truncate64, "%s, %jd"), path, args->length); #endif error = kern_truncate(td, path, UIO_SYSSPACE, args->length); LFREEPATH(path); return (error); } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ int linux_ftruncate(struct thread *td, struct linux_ftruncate_args *args) { return (kern_ftruncate(td, args->fd, args->length)); } #ifdef LINUX_LEGACY_SYSCALLS int linux_link(struct thread *td, struct linux_link_args *args) { char *path, *to; int error; LCONVPATHEXIST(td, args->path, &path); /* Expand LCONVPATHCREATE so that `path' can be freed on errors */ error = linux_emul_convpath(td, args->to, UIO_USERSPACE, &to, 1, AT_FDCWD); if (to == NULL) { LFREEPATH(path); return (error); } #ifdef DEBUG if (ldebug(link)) printf(ARGS(link, "%s, %s"), path, to); #endif error = kern_linkat(td, AT_FDCWD, AT_FDCWD, path, to, UIO_SYSSPACE, FOLLOW); LFREEPATH(path); LFREEPATH(to); return (error); } #endif int linux_linkat(struct thread *td, struct linux_linkat_args *args) { char *path, *to; int error, olddfd, newdfd, follow; if (args->flag & ~LINUX_AT_SYMLINK_FOLLOW) return (EINVAL); olddfd = (args->olddfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->olddfd; newdfd = (args->newdfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->newdfd; LCONVPATHEXIST_AT(td, args->oldname, &path, olddfd); /* Expand LCONVPATHCREATE so that `path' can be freed on errors */ error = linux_emul_convpath(td, args->newname, UIO_USERSPACE, &to, 1, newdfd); if (to == NULL) { LFREEPATH(path); return (error); } #ifdef DEBUG if (ldebug(linkat)) printf(ARGS(linkat, "%i, %s, %i, %s, %i"), args->olddfd, path, args->newdfd, to, args->flag); #endif follow = (args->flag & LINUX_AT_SYMLINK_FOLLOW) == 0 ? NOFOLLOW : FOLLOW; error = kern_linkat(td, olddfd, newdfd, path, to, UIO_SYSSPACE, follow); LFREEPATH(path); LFREEPATH(to); return (error); } int linux_fdatasync(td, uap) struct thread *td; struct linux_fdatasync_args *uap; { return (kern_fsync(td, uap->fd, false)); } int linux_pread(struct thread *td, struct linux_pread_args *uap) { struct vnode *vp; int error; error = kern_pread(td, uap->fd, uap->buf, uap->nbyte, uap->offset); if (error == 0) { /* This seems to violate POSIX but Linux does it. */ error = fgetvp(td, uap->fd, &cap_pread_rights, &vp); if (error != 0) return (error); if (vp->v_type == VDIR) { vrele(vp); return (EISDIR); } vrele(vp); } return (error); } int linux_pwrite(struct thread *td, struct linux_pwrite_args *uap) { return (kern_pwrite(td, uap->fd, uap->buf, uap->nbyte, uap->offset)); } int linux_preadv(struct thread *td, struct linux_preadv_args *uap) { struct uio *auio; int error; off_t offset; /* * According http://man7.org/linux/man-pages/man2/preadv.2.html#NOTES * pos_l and pos_h, respectively, contain the * low order and high order 32 bits of offset. */ offset = (((off_t)uap->pos_h << (sizeof(offset) * 4)) << (sizeof(offset) * 4)) | uap->pos_l; if (offset < 0) return (EINVAL); #ifdef COMPAT_LINUX32 error = linux32_copyinuio(PTRIN(uap->vec), uap->vlen, &auio); #else error = copyinuio(uap->vec, uap->vlen, &auio); #endif if (error != 0) return (error); error = kern_preadv(td, uap->fd, auio, offset); free(auio, M_IOV); return (error); } int linux_pwritev(struct thread *td, struct linux_pwritev_args *uap) { struct uio *auio; int error; off_t offset; /* * According http://man7.org/linux/man-pages/man2/pwritev.2.html#NOTES * pos_l and pos_h, respectively, contain the * low order and high order 32 bits of offset. */ offset = (((off_t)uap->pos_h << (sizeof(offset) * 4)) << (sizeof(offset) * 4)) | uap->pos_l; if (offset < 0) return (EINVAL); #ifdef COMPAT_LINUX32 error = linux32_copyinuio(PTRIN(uap->vec), uap->vlen, &auio); #else error = copyinuio(uap->vec, uap->vlen, &auio); #endif if (error != 0) return (error); error = kern_pwritev(td, uap->fd, auio, offset); free(auio, M_IOV); return (error); } int linux_mount(struct thread *td, struct linux_mount_args *args) { char fstypename[MFSNAMELEN]; char *mntonname, *mntfromname; int error, fsflags; mntonname = malloc(MNAMELEN, M_TEMP, M_WAITOK); mntfromname = malloc(MNAMELEN, M_TEMP, M_WAITOK); error = copyinstr(args->filesystemtype, fstypename, MFSNAMELEN - 1, NULL); if (error != 0) goto out; error = copyinstr(args->specialfile, mntfromname, MNAMELEN - 1, NULL); if (error != 0) goto out; error = copyinstr(args->dir, mntonname, MNAMELEN - 1, NULL); if (error != 0) goto out; #ifdef DEBUG if (ldebug(mount)) printf(ARGS(mount, "%s, %s, %s"), fstypename, mntfromname, mntonname); #endif if (strcmp(fstypename, "ext2") == 0) { strcpy(fstypename, "ext2fs"); } else if (strcmp(fstypename, "proc") == 0) { strcpy(fstypename, "linprocfs"); } else if (strcmp(fstypename, "vfat") == 0) { strcpy(fstypename, "msdosfs"); } fsflags = 0; if ((args->rwflag & 0xffff0000) == 0xc0ed0000) { /* * Linux SYNC flag is not included; the closest equivalent * FreeBSD has is !ASYNC, which is our default. */ if (args->rwflag & LINUX_MS_RDONLY) fsflags |= MNT_RDONLY; if (args->rwflag & LINUX_MS_NOSUID) fsflags |= MNT_NOSUID; if (args->rwflag & LINUX_MS_NOEXEC) fsflags |= MNT_NOEXEC; if (args->rwflag & LINUX_MS_REMOUNT) fsflags |= MNT_UPDATE; } error = kernel_vmount(fsflags, "fstype", fstypename, "fspath", mntonname, "from", mntfromname, NULL); out: free(mntonname, M_TEMP); free(mntfromname, M_TEMP); return (error); } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) int linux_oldumount(struct thread *td, struct linux_oldumount_args *args) { struct linux_umount_args args2; args2.path = args->path; args2.flags = 0; return (linux_umount(td, &args2)); } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ #ifdef LINUX_LEGACY_SYSCALLS int linux_umount(struct thread *td, struct linux_umount_args *args) { struct unmount_args bsd; bsd.path = args->path; bsd.flags = args->flags; /* XXX correct? */ return (sys_unmount(td, &bsd)); } #endif /* * fcntl family of syscalls */ struct l_flock { l_short l_type; l_short l_whence; l_off_t l_start; l_off_t l_len; l_pid_t l_pid; } #if defined(__amd64__) && defined(COMPAT_LINUX32) __packed #endif ; static void linux_to_bsd_flock(struct l_flock *linux_flock, struct flock *bsd_flock) { switch (linux_flock->l_type) { case LINUX_F_RDLCK: bsd_flock->l_type = F_RDLCK; break; case LINUX_F_WRLCK: bsd_flock->l_type = F_WRLCK; break; case LINUX_F_UNLCK: bsd_flock->l_type = F_UNLCK; break; default: bsd_flock->l_type = -1; break; } bsd_flock->l_whence = linux_flock->l_whence; bsd_flock->l_start = (off_t)linux_flock->l_start; bsd_flock->l_len = (off_t)linux_flock->l_len; bsd_flock->l_pid = (pid_t)linux_flock->l_pid; bsd_flock->l_sysid = 0; } static void bsd_to_linux_flock(struct flock *bsd_flock, struct l_flock *linux_flock) { switch (bsd_flock->l_type) { case F_RDLCK: linux_flock->l_type = LINUX_F_RDLCK; break; case F_WRLCK: linux_flock->l_type = LINUX_F_WRLCK; break; case F_UNLCK: linux_flock->l_type = LINUX_F_UNLCK; break; } linux_flock->l_whence = bsd_flock->l_whence; linux_flock->l_start = (l_off_t)bsd_flock->l_start; linux_flock->l_len = (l_off_t)bsd_flock->l_len; linux_flock->l_pid = (l_pid_t)bsd_flock->l_pid; } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) struct l_flock64 { l_short l_type; l_short l_whence; l_loff_t l_start; l_loff_t l_len; l_pid_t l_pid; } #if defined(__amd64__) && defined(COMPAT_LINUX32) __packed #endif ; static void linux_to_bsd_flock64(struct l_flock64 *linux_flock, struct flock *bsd_flock) { switch (linux_flock->l_type) { case LINUX_F_RDLCK: bsd_flock->l_type = F_RDLCK; break; case LINUX_F_WRLCK: bsd_flock->l_type = F_WRLCK; break; case LINUX_F_UNLCK: bsd_flock->l_type = F_UNLCK; break; default: bsd_flock->l_type = -1; break; } bsd_flock->l_whence = linux_flock->l_whence; bsd_flock->l_start = (off_t)linux_flock->l_start; bsd_flock->l_len = (off_t)linux_flock->l_len; bsd_flock->l_pid = (pid_t)linux_flock->l_pid; bsd_flock->l_sysid = 0; } static void bsd_to_linux_flock64(struct flock *bsd_flock, struct l_flock64 *linux_flock) { switch (bsd_flock->l_type) { case F_RDLCK: linux_flock->l_type = LINUX_F_RDLCK; break; case F_WRLCK: linux_flock->l_type = LINUX_F_WRLCK; break; case F_UNLCK: linux_flock->l_type = LINUX_F_UNLCK; break; } linux_flock->l_whence = bsd_flock->l_whence; linux_flock->l_start = (l_loff_t)bsd_flock->l_start; linux_flock->l_len = (l_loff_t)bsd_flock->l_len; linux_flock->l_pid = (l_pid_t)bsd_flock->l_pid; } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ static int fcntl_common(struct thread *td, struct linux_fcntl_args *args) { struct l_flock linux_flock; struct flock bsd_flock; struct file *fp; long arg; int error, result; switch (args->cmd) { case LINUX_F_DUPFD: return (kern_fcntl(td, args->fd, F_DUPFD, args->arg)); case LINUX_F_GETFD: return (kern_fcntl(td, args->fd, F_GETFD, 0)); case LINUX_F_SETFD: return (kern_fcntl(td, args->fd, F_SETFD, args->arg)); case LINUX_F_GETFL: error = kern_fcntl(td, args->fd, F_GETFL, 0); result = td->td_retval[0]; td->td_retval[0] = 0; if (result & O_RDONLY) td->td_retval[0] |= LINUX_O_RDONLY; if (result & O_WRONLY) td->td_retval[0] |= LINUX_O_WRONLY; if (result & O_RDWR) td->td_retval[0] |= LINUX_O_RDWR; if (result & O_NDELAY) td->td_retval[0] |= LINUX_O_NONBLOCK; if (result & O_APPEND) td->td_retval[0] |= LINUX_O_APPEND; if (result & O_FSYNC) td->td_retval[0] |= LINUX_O_SYNC; if (result & O_ASYNC) td->td_retval[0] |= LINUX_FASYNC; #ifdef LINUX_O_NOFOLLOW if (result & O_NOFOLLOW) td->td_retval[0] |= LINUX_O_NOFOLLOW; #endif #ifdef LINUX_O_DIRECT if (result & O_DIRECT) td->td_retval[0] |= LINUX_O_DIRECT; #endif return (error); case LINUX_F_SETFL: arg = 0; if (args->arg & LINUX_O_NDELAY) arg |= O_NONBLOCK; if (args->arg & LINUX_O_APPEND) arg |= O_APPEND; if (args->arg & LINUX_O_SYNC) arg |= O_FSYNC; if (args->arg & LINUX_FASYNC) arg |= O_ASYNC; #ifdef LINUX_O_NOFOLLOW if (args->arg & LINUX_O_NOFOLLOW) arg |= O_NOFOLLOW; #endif #ifdef LINUX_O_DIRECT if (args->arg & LINUX_O_DIRECT) arg |= O_DIRECT; #endif return (kern_fcntl(td, args->fd, F_SETFL, arg)); case LINUX_F_GETLK: error = copyin((void *)args->arg, &linux_flock, sizeof(linux_flock)); if (error) return (error); linux_to_bsd_flock(&linux_flock, &bsd_flock); error = kern_fcntl(td, args->fd, F_GETLK, (intptr_t)&bsd_flock); if (error) return (error); bsd_to_linux_flock(&bsd_flock, &linux_flock); return (copyout(&linux_flock, (void *)args->arg, sizeof(linux_flock))); case LINUX_F_SETLK: error = copyin((void *)args->arg, &linux_flock, sizeof(linux_flock)); if (error) return (error); linux_to_bsd_flock(&linux_flock, &bsd_flock); return (kern_fcntl(td, args->fd, F_SETLK, (intptr_t)&bsd_flock)); case LINUX_F_SETLKW: error = copyin((void *)args->arg, &linux_flock, sizeof(linux_flock)); if (error) return (error); linux_to_bsd_flock(&linux_flock, &bsd_flock); return (kern_fcntl(td, args->fd, F_SETLKW, (intptr_t)&bsd_flock)); case LINUX_F_GETOWN: return (kern_fcntl(td, args->fd, F_GETOWN, 0)); case LINUX_F_SETOWN: /* * XXX some Linux applications depend on F_SETOWN having no * significant effect for pipes (SIGIO is not delivered for * pipes under Linux-2.2.35 at least). */ error = fget(td, args->fd, &cap_fcntl_rights, &fp); if (error) return (error); if (fp->f_type == DTYPE_PIPE) { fdrop(fp, td); return (EINVAL); } fdrop(fp, td); return (kern_fcntl(td, args->fd, F_SETOWN, args->arg)); case LINUX_F_DUPFD_CLOEXEC: return (kern_fcntl(td, args->fd, F_DUPFD_CLOEXEC, args->arg)); } return (EINVAL); } int linux_fcntl(struct thread *td, struct linux_fcntl_args *args) { #ifdef DEBUG if (ldebug(fcntl)) printf(ARGS(fcntl, "%d, %08x, *"), args->fd, args->cmd); #endif return (fcntl_common(td, args)); } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) int linux_fcntl64(struct thread *td, struct linux_fcntl64_args *args) { struct l_flock64 linux_flock; struct flock bsd_flock; struct linux_fcntl_args fcntl_args; int error; #ifdef DEBUG if (ldebug(fcntl64)) printf(ARGS(fcntl64, "%d, %08x, *"), args->fd, args->cmd); #endif switch (args->cmd) { case LINUX_F_GETLK64: error = copyin((void *)args->arg, &linux_flock, sizeof(linux_flock)); if (error) return (error); linux_to_bsd_flock64(&linux_flock, &bsd_flock); error = kern_fcntl(td, args->fd, F_GETLK, (intptr_t)&bsd_flock); if (error) return (error); bsd_to_linux_flock64(&bsd_flock, &linux_flock); return (copyout(&linux_flock, (void *)args->arg, sizeof(linux_flock))); case LINUX_F_SETLK64: error = copyin((void *)args->arg, &linux_flock, sizeof(linux_flock)); if (error) return (error); linux_to_bsd_flock64(&linux_flock, &bsd_flock); return (kern_fcntl(td, args->fd, F_SETLK, (intptr_t)&bsd_flock)); case LINUX_F_SETLKW64: error = copyin((void *)args->arg, &linux_flock, sizeof(linux_flock)); if (error) return (error); linux_to_bsd_flock64(&linux_flock, &bsd_flock); return (kern_fcntl(td, args->fd, F_SETLKW, (intptr_t)&bsd_flock)); } fcntl_args.fd = args->fd; fcntl_args.cmd = args->cmd; fcntl_args.arg = args->arg; return (fcntl_common(td, &fcntl_args)); } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ #ifdef LINUX_LEGACY_SYSCALLS int linux_chown(struct thread *td, struct linux_chown_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(chown)) printf(ARGS(chown, "%s, %d, %d"), path, args->uid, args->gid); #endif error = kern_fchownat(td, AT_FDCWD, path, UIO_SYSSPACE, args->uid, args->gid, 0); LFREEPATH(path); return (error); } #endif int linux_fchownat(struct thread *td, struct linux_fchownat_args *args) { char *path; int error, dfd, flag; if (args->flag & ~LINUX_AT_SYMLINK_NOFOLLOW) return (EINVAL); dfd = (args->dfd == LINUX_AT_FDCWD) ? AT_FDCWD : args->dfd; LCONVPATHEXIST_AT(td, args->filename, &path, dfd); #ifdef DEBUG if (ldebug(fchownat)) printf(ARGS(fchownat, "%s, %d, %d"), path, args->uid, args->gid); #endif flag = (args->flag & LINUX_AT_SYMLINK_NOFOLLOW) == 0 ? 0 : AT_SYMLINK_NOFOLLOW; error = kern_fchownat(td, dfd, path, UIO_SYSSPACE, args->uid, args->gid, flag); LFREEPATH(path); return (error); } #ifdef LINUX_LEGACY_SYSCALLS int linux_lchown(struct thread *td, struct linux_lchown_args *args) { char *path; int error; LCONVPATHEXIST(td, args->path, &path); #ifdef DEBUG if (ldebug(lchown)) printf(ARGS(lchown, "%s, %d, %d"), path, args->uid, args->gid); #endif error = kern_fchownat(td, AT_FDCWD, path, UIO_SYSSPACE, args->uid, args->gid, AT_SYMLINK_NOFOLLOW); LFREEPATH(path); return (error); } #endif static int convert_fadvice(int advice) { switch (advice) { case LINUX_POSIX_FADV_NORMAL: return (POSIX_FADV_NORMAL); case LINUX_POSIX_FADV_RANDOM: return (POSIX_FADV_RANDOM); case LINUX_POSIX_FADV_SEQUENTIAL: return (POSIX_FADV_SEQUENTIAL); case LINUX_POSIX_FADV_WILLNEED: return (POSIX_FADV_WILLNEED); case LINUX_POSIX_FADV_DONTNEED: return (POSIX_FADV_DONTNEED); case LINUX_POSIX_FADV_NOREUSE: return (POSIX_FADV_NOREUSE); default: return (-1); } } int linux_fadvise64(struct thread *td, struct linux_fadvise64_args *args) { int advice; advice = convert_fadvice(args->advice); if (advice == -1) return (EINVAL); return (kern_posix_fadvise(td, args->fd, args->offset, args->len, advice)); } #if defined(__i386__) || (defined(__amd64__) && defined(COMPAT_LINUX32)) int linux_fadvise64_64(struct thread *td, struct linux_fadvise64_64_args *args) { int advice; advice = convert_fadvice(args->advice); if (advice == -1) return (EINVAL); return (kern_posix_fadvise(td, args->fd, args->offset, args->len, advice)); } #endif /* __i386__ || (__amd64__ && COMPAT_LINUX32) */ #ifdef LINUX_LEGACY_SYSCALLS int linux_pipe(struct thread *td, struct linux_pipe_args *args) { int fildes[2]; int error; #ifdef DEBUG if (ldebug(pipe)) printf(ARGS(pipe, "*")); #endif error = kern_pipe(td, fildes, 0, NULL, NULL); if (error != 0) return (error); error = copyout(fildes, args->pipefds, sizeof(fildes)); if (error != 0) { (void)kern_close(td, fildes[0]); (void)kern_close(td, fildes[1]); } return (error); } #endif int linux_pipe2(struct thread *td, struct linux_pipe2_args *args) { int fildes[2]; int error, flags; #ifdef DEBUG if (ldebug(pipe2)) printf(ARGS(pipe2, "*, %d"), args->flags); #endif if ((args->flags & ~(LINUX_O_NONBLOCK | LINUX_O_CLOEXEC)) != 0) return (EINVAL); flags = 0; if ((args->flags & LINUX_O_NONBLOCK) != 0) flags |= O_NONBLOCK; if ((args->flags & LINUX_O_CLOEXEC) != 0) flags |= O_CLOEXEC; error = kern_pipe(td, fildes, flags, NULL, NULL); if (error != 0) return (error); error = copyout(fildes, args->pipefds, sizeof(fildes)); if (error != 0) { (void)kern_close(td, fildes[0]); (void)kern_close(td, fildes[1]); } return (error); } int linux_dup3(struct thread *td, struct linux_dup3_args *args) { int cmd; intptr_t newfd; if (args->oldfd == args->newfd) return (EINVAL); if ((args->flags & ~LINUX_O_CLOEXEC) != 0) return (EINVAL); if (args->flags & LINUX_O_CLOEXEC) cmd = F_DUP2FD_CLOEXEC; else cmd = F_DUP2FD; newfd = args->newfd; return (kern_fcntl(td, args->oldfd, cmd, newfd)); } int linux_fallocate(struct thread *td, struct linux_fallocate_args *args) { /* * We emulate only posix_fallocate system call for which * mode should be 0. */ if (args->mode != 0) return (ENOSYS); return (kern_posix_fallocate(td, args->fd, args->offset, args->len)); } Index: head/sys/kern/capabilities.conf =================================================================== --- head/sys/kern/capabilities.conf (revision 345981) +++ head/sys/kern/capabilities.conf (revision 345982) @@ -1,752 +1,753 @@ ## ## Copyright (c) 2008-2010 Robert N. M. Watson ## All rights reserved. ## ## This software was developed at the University of Cambridge Computer ## Laboratory with support from a grant from Google, 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. ## ## 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. ## ## List of system calls enabled in capability mode, one name per line. ## ## Notes: ## - sys_exit(2), abort2(2) and close(2) are very important. ## - Sorted alphabetically, please keep it that way. ## ## $FreeBSD$ ## ## ## Allow ACL and MAC label operations by file descriptor, subject to ## capability rights. Allow MAC label operations on the current process but ## we will need to scope __mac_get_pid(2). ## __acl_aclcheck_fd __acl_delete_fd __acl_get_fd __acl_set_fd __mac_get_fd #__mac_get_pid __mac_get_proc __mac_set_fd __mac_set_proc ## ## Allow sysctl(2) as we scope internal to the call; this is a global ## namespace, but there are several critical sysctls required for almost ## anything to run, such as hw.pagesize. For now that policy lives in the ## kernel for performance and simplicity, but perhaps it could move to a ## proxying daemon in userspace. ## __sysctl ## ## Allow umtx operations as these are scoped by address space. ## ## XXRW: Need to check this very carefully. ## _umtx_op ## ## Allow process termination using abort2(2). ## abort2 ## ## Allow accept(2) since it doesn't manipulate namespaces directly, rather ## relies on existing bindings on a socket, subject to capability rights. ## accept accept4 ## ## Allow AIO operations by file descriptor, subject to capability rights. ## aio_cancel aio_error aio_fsync aio_read aio_return aio_suspend aio_waitcomplete aio_write ## ## audit(2) is a global operation, submitting to the global trail, but it is ## controlled by privilege, and it might be useful to be able to submit ## records from sandboxes. For now, disallow, but we may want to think about ## providing some sort of proxy service for this. ## #audit ## ## Allow bindat(2). ## bindat ## ## Allow capability mode and capability system calls. ## cap_enter cap_fcntls_get cap_fcntls_limit cap_getmode cap_ioctls_get cap_ioctls_limit __cap_rights_get cap_rights_limit ## ## Allow read-only clock operations. ## clock_getres clock_gettime ## ## Always allow file descriptor close(2). ## close closefrom ## ## Allow connectat(2). ## connectat ## ## cpuset(2) and related calls are limited to caller's own process/thread. ## #cpuset cpuset_getaffinity #cpuset_getid cpuset_setaffinity #cpuset_setid ## ## Always allow dup(2) and dup2(2) manipulation of the file descriptor table. ## dup dup2 ## ## Allow extended attribute operations by file descriptor, subject to ## capability rights. ## extattr_delete_fd extattr_get_fd extattr_list_fd extattr_set_fd ## ## Allow changing file flags, mode, and owner by file descriptor, subject to ## capability rights. ## fchflags fchmod fchown ## ## For now, allow fcntl(2), subject to capability rights, but this probably ## needs additional scoping. ## fcntl ## ## Allow fexecve(2), subject to capability rights. We perform some scoping, ## such as disallowing privilege escalation. ## fexecve ## ## Allow flock(2), subject to capability rights. ## flock ## ## Allow fork(2), even though it returns pids -- some applications seem to ## prefer this interface. ## fork ## ## Allow fpathconf(2), subject to capability rights. ## fpathconf ## ## Allow various file descriptor-based I/O operations, subject to capability ## rights. ## freebsd11_fstat freebsd11_fstatat freebsd11_getdirentries freebsd11_fstatfs freebsd11_mknodat freebsd6_ftruncate freebsd6_lseek freebsd6_mmap freebsd6_pread freebsd6_pwrite ## ## Allow querying file and file system state with fstat(2) and fstatfs(2), ## subject to capability rights. ## fstat fstatfs ## ## Allow further file descriptor-based I/O operations, subject to capability ## rights. ## fsync ftruncate ## ## Allow futimens(2) and futimes(2), subject to capability rights. ## futimens futimes ## ## Allow querying process audit state, subject to normal access control. ## getaudit getaudit_addr getauid ## ## Allow thread context management with getcontext(2). ## getcontext ## ## Allow directory I/O on a file descriptor, subject to capability rights. ## Originally we had separate capabilities for directory-specific read ## operations, but on BSD we allow reading the raw directory data, so we just ## rely on CAP_READ now. ## getdents getdirentries ## ## Allow querying certain trivial global state. ## getdomainname ## ## Allow querying certain per-process resource limit state. ## getdtablesize ## ## Allow querying current process credential state. ## getegid geteuid ## ## Allow querying certain trivial global state. ## gethostid gethostname ## ## Allow querying per-process timer. ## getitimer ## ## Allow querying current process credential state. ## getgid getgroups getlogin ## ## Allow querying certain trivial global state. ## getpagesize getpeername ## ## Allow querying certain per-process scheduling, resource limit, and ## credential state. ## ## XXXRW: getpgid(2) needs scoping. It's not clear if it's worth scoping ## getppid(2). getpriority(2) needs scoping. getrusage(2) needs scoping. ## getsid(2) needs scoping. ## getpgid getpgrp getpid getppid getpriority getresgid getresuid getrlimit getrusage getsid ## ## Allow getrandom ## getrandom ## ## Allow querying socket state, subject to capability rights. ## ## XXXRW: getsockopt(2) may need more attention. ## getsockname getsockopt ## ## Allow querying the global clock. ## gettimeofday ## ## Allow querying current process credential state. ## getuid ## ## Allow ioctl(2), which hopefully will be limited by applications only to ## required commands with cap_ioctls_limit(2) syscall. ## ioctl ## ## Allow querying current process credential state. ## issetugid ## ## Allow kevent(2), as we will authorize based on capability rights on the ## target descriptor. ## kevent ## ## Allow kill(2), as we allow the process to send signals only to himself. ## kill ## ## Allow message queue operations on file descriptors, subject to capability ## rights. ## NOTE: Corresponding sysents are initialized in sys/kern/uipc_mqueue.c with ## SYF_CAPENABLED. ## kmq_notify kmq_setattr kmq_timedreceive kmq_timedsend ## ## Allow kqueue(2), we will control use. ## kqueue ## ## Allow managing per-process timers. ## ktimer_create ktimer_delete ktimer_getoverrun ktimer_gettime ktimer_settime ## ## We can't allow ktrace(2) because it relies on a global namespace, but we ## might want to introduce an fktrace(2) of some sort. ## #ktrace ## ## Allow AIO operations by file descriptor, subject to capability rights. ## lio_listio ## ## Allow listen(2), subject to capability rights. ## ## XXXRW: One might argue this manipulates a global namespace. ## listen ## ## Allow I/O-related file descriptors, subject to capability rights. ## lseek ## ## Allow simple VM operations on the current process. ## madvise mincore minherit mlock mlockall ## ## Allow memory mapping a file descriptor, and updating protections, subject ## to capability rights. ## mmap mprotect ## ## Allow simple VM operations on the current process. ## msync munlock munlockall munmap ## ## Allow the current process to sleep. ## nanosleep ## ## Allow querying the global clock. ## ntp_gettime ## ## Allow AIO operations by file descriptor, subject to capability rights. ## oaio_read oaio_write ## ## Allow simple VM operations on the current process. ## break ## ## Allow AIO operations by file descriptor, subject to capability rights. ## olio_listio ## ## Operations relative to directory capabilities. ## chflagsat faccessat fchmodat fchownat fstatat futimesat linkat mkdirat mkfifoat mknodat openat readlinkat renameat symlinkat unlinkat +funlinkat utimensat ## ## Process descriptor-related system calls are allowed. ## pdfork pdgetpid pdkill #pdwait4 # not yet implemented ## ## Allow pipe(2). ## pipe pipe2 ## ## Allow poll(2), which will be scoped by capability rights. ## poll ppoll ## ## Allow I/O-related file descriptors, subject to capability rights. ## posix_fallocate pread preadv ## ## Allow access to profiling state on the current process. ## profil ## ## Disallow ptrace(2) for now, but we do need debugging facilities in ## capability mode, so we will want to revisit this, possibly by scoping its ## operation. ## #ptrace ## ## Allow I/O-related file descriptors, subject to capability rights. ## pwrite pwritev read readv recv recvfrom recvmsg ## ## Allow real-time scheduling primitives to be used. ## ## XXXRW: These require scoping. ## rtprio rtprio_thread ## ## Allow simple VM operations on the current process. ## sbrk ## ## Allow querying trivial global scheduler state. ## sched_get_priority_max sched_get_priority_min ## ## Allow various thread/process scheduler operations. ## ## XXXRW: Some of these require further scoping. ## sched_getparam sched_getscheduler sched_rr_get_interval sched_setparam sched_setscheduler sched_yield ## ## Allow I/O-related file descriptors, subject to capability rights. ## NOTE: Corresponding sysents are initialized in sys/netinet/sctp_syscalls.c ## with SYF_CAPENABLED. ## sctp_generic_recvmsg sctp_generic_sendmsg sctp_generic_sendmsg_iov sctp_peeloff ## ## Allow pselect(2) and select(2), which will be scoped by capability rights. ## ## XXXRW: But is it? ## pselect select ## ## Allow I/O-related file descriptors, subject to capability rights. Use of ## explicit addresses here is restricted by the system calls themselves. ## send sendfile sendmsg sendto ## ## Allow setting per-process audit state, which is controlled separately by ## privileges. ## setaudit setaudit_addr setauid ## ## Allow setting thread context. ## setcontext ## ## Allow setting current process credential state, which is controlled ## separately by privilege. ## setegid seteuid setgid ## ## Allow use of the process interval timer. ## setitimer ## ## Allow setpriority(2). ## ## XXXRW: Requires scoping. ## setpriority ## ## Allow setting current process credential state, which is controlled ## separately by privilege. ## setregid setresgid setresuid setreuid ## ## Allow setting process resource limits with setrlimit(2). ## setrlimit ## ## Allow creating a new session with setsid(2). ## setsid ## ## Allow setting socket options with setsockopt(2), subject to capability ## rights. ## ## XXXRW: Might require scoping. ## setsockopt ## ## Allow setting current process credential state, which is controlled ## separately by privilege. ## setuid ## ## shm_open(2) is scoped so as to allow only access to new anonymous objects. ## shm_open ## ## Allow I/O-related file descriptors, subject to capability rights. ## shutdown ## ## Allow signal control on current process. ## sigaction sigaltstack sigblock sigpending sigprocmask sigqueue sigreturn sigsetmask sigstack sigsuspend sigtimedwait sigvec sigwaitinfo sigwait ## ## Allow creating new socket pairs with socket(2) and socketpair(2). ## socket socketpair ## ## Allow simple VM operations on the current process. ## ## XXXRW: Kernel doesn't implement this, so drop? ## sstk ## ## Do allow sync(2) for now, but possibly shouldn't. ## sync ## ## Always allow process termination with sys_exit(2). ## sys_exit ## ## sysarch(2) does rather diverse things, but is required on at least i386 ## in order to configure per-thread data. As such, it's scoped on each ## architecture. ## sysarch ## ## Allow thread operations operating only on current process. ## thr_create thr_exit thr_kill ## ## Disallow thr_kill2(2), as it may operate beyond the current process. ## ## XXXRW: Requires scoping. ## #thr_kill2 ## ## Allow thread operations operating only on current process. ## thr_new thr_self thr_set_name thr_suspend thr_wake ## ## Allow manipulation of the current process umask with umask(2). ## umask ## ## Allow submitting of process trace entries with utrace(2). ## utrace ## ## Allow generating UUIDs with uuidgen(2). ## uuidgen ## ## Allow I/O-related file descriptors, subject to capability rights. ## write writev ## ## Allow processes to yield(2). ## yield Index: head/sys/kern/syscalls.master =================================================================== --- head/sys/kern/syscalls.master (revision 345981) +++ head/sys/kern/syscalls.master (revision 345982) @@ -1,3173 +1,3181 @@ $FreeBSD$ ; from: @(#)syscalls.master 8.2 (Berkeley) 1/13/94 ; ; System call name/number master file. ; Processed to created init_sysent.c, syscalls.c and syscall.h. ; Columns: number audit type name alt{name,tag,rtyp}/comments ; number system call number, must be in order ; audit the audit event associated with the system call ; A value of AUE_NULL means no auditing, but it also means that ; there is no audit event for the call at this time. For the ; case where the event exists, but we don't want auditing, the ; event should be #defined to AUE_NULL in audit_kevents.h. ; type one of STD, OBSOL, UNIMPL, COMPAT, COMPAT4, COMPAT6, ; COMPAT7, COMPAT11, NODEF, NOARGS, NOPROTO, NOSTD ; The COMPAT* options may be combined with one or more NO* ; options separated by '|' with no spaces (e.g. COMPAT|NOARGS) ; name pseudo-prototype of syscall routine ; If one of the following alts is different, then all appear: ; altname name of system call if different ; alttag name of args struct tag if different from [o]`name'"_args" ; altrtyp return type if not int (bogus - syscalls always return int) ; for UNIMPL/OBSOL, name continues with comments ; types: ; STD always included ; COMPAT included on COMPAT #ifdef ; COMPAT4 included on COMPAT_FREEBSD4 #ifdef (FreeBSD 4 compat) ; COMPAT6 included on COMPAT_FREEBSD6 #ifdef (FreeBSD 6 compat) ; COMPAT7 included on COMPAT_FREEBSD7 #ifdef (FreeBSD 7 compat) ; COMPAT10 included on COMPAT_FREEBSD10 #ifdef (FreeBSD 10 compat) ; COMPAT11 included on COMPAT_FREEBSD11 #ifdef (FreeBSD 11 compat) ; OBSOL obsolete, not included in system, only specifies name ; UNIMPL not implemented, placeholder only ; NOSTD implemented but as a lkm that can be statically ; compiled in; sysent entry will be filled with lkmressys ; so the SYSCALL_MODULE macro works ; NOARGS same as STD except do not create structure in sys/sysproto.h ; NODEF same as STD except only have the entry in the syscall table ; added. Meaning - do not create structure or function ; prototype in sys/sysproto.h ; NOPROTO same as STD except do not create structure or ; function prototype in sys/sysproto.h. Does add a ; definition to syscall.h besides adding a sysent. ; NOTSTATIC syscall is loadable ; annotations: ; SAL 2.0 annotations are used to specify how system calls treat ; arguments that are passed using pointers. There are three basic ; annotations. ; ; _In_ Object pointed to will be read and not modified. ; _Out_ Object pointed to will be written and not read. ; _Inout_ Object pointed to will be written and read. ; ; These annotations are used alone when the pointer refers to a single ; object i.e. scalar types, structs, and pointers, and not NULL. Adding ; the _opt_ suffix, e.g. _In_opt_, implies that the pointer may also ; refer to NULL. ; ; For pointers to arrays, additional suffixes are added: ; ; _In_z_, _Out_z_, _Inout_z_: ; for a NUL terminated array e.g. a string. ; _In_reads_z_(n),_Out_writes_z_(n), _Inout_updates_z_(n): ; for a NUL terminated array e.g. a string, of known length n bytes. ; _In_reads_(n),_Out_writes_(n),_Inout_updates_(n): ; for an array of n elements. ; _In_reads_bytes_(n), _Out_writes_bytes_(n), _Inout_updates_bytes(n): ; for a buffer of n-bytes. ; Please copy any additions and changes to the following compatability tables: ; sys/compat/freebsd32/syscalls.master ; #ifdef's, etc. may be included, and are copied to the output files. #include #include #include ; Reserved/unimplemented system calls in the range 0-150 inclusive ; are reserved for use in future Berkeley releases. ; Additional system calls implemented in vendor and other ; redistributions should be placed in the reserved range at the end ; of the current calls. 0 AUE_NULL STD { int nosys(void); } syscall nosys_args int 1 AUE_EXIT STD { void sys_exit( int rval ); } exit sys_exit_args void 2 AUE_FORK STD { int fork(void); } 3 AUE_READ STD { ssize_t read( int fd, _Out_writes_bytes_(nbyte) void *buf, size_t nbyte ); } 4 AUE_WRITE STD { ssize_t write( int fd, _In_reads_bytes_(nbyte) const void *buf, size_t nbyte ); } 5 AUE_OPEN_RWTC STD { int open( _In_z_ const char *path, int flags, mode_t mode ); } ; XXX should be { int open(const char *path, int flags, ...); } ; but we're not ready for varargs. 6 AUE_CLOSE STD { int close( int fd ); } 7 AUE_WAIT4 STD { int wait4( int pid, _Out_opt_ int *status, int options, _Out_opt_ struct rusage *rusage ); } 8 AUE_CREAT COMPAT { int creat( _In_z_ const char *path, int mode ); } 9 AUE_LINK STD { int link( _In_z_ const char *path, _In_z_ const char *link ); } 10 AUE_UNLINK STD { int unlink( _In_z_ const char *path ); } 11 AUE_NULL OBSOL execv 12 AUE_CHDIR STD { int chdir( _In_z_ const char *path ); } 13 AUE_FCHDIR STD { int fchdir( int fd ); } 14 AUE_MKNOD COMPAT11 { int mknod( _In_z_ const char *path, int mode, uint32_t dev ); } 15 AUE_CHMOD STD { int chmod( _In_z_ const char *path, mode_t mode ); } 16 AUE_CHOWN STD { int chown( _In_z_ const char *path, int uid, int gid ); } 17 AUE_NULL STD { void *break( _In_ char *nsize ); } 18 AUE_GETFSSTAT COMPAT4 { int getfsstat( _Out_writes_bytes_opt_(bufsize) struct ostatfs *buf, long bufsize, int mode ); } 19 AUE_LSEEK COMPAT { long lseek( int fd, long offset, int whence ); } 20 AUE_GETPID STD { pid_t getpid(void); } 21 AUE_MOUNT STD { int mount( _In_z_ const char *type, _In_z_ const char *path, int flags, _In_opt_ void *data ); } 22 AUE_UMOUNT STD { int unmount( _In_z_ const char *path, int flags ); } 23 AUE_SETUID STD { int setuid( uid_t uid ); } 24 AUE_GETUID STD { uid_t getuid(void); } 25 AUE_GETEUID STD { uid_t geteuid(void); } 26 AUE_PTRACE STD { int ptrace( int req, pid_t pid, _Inout_opt_ caddr_t addr, int data ); } 27 AUE_RECVMSG STD { int recvmsg( int s, _Inout_ struct msghdr *msg, int flags ); } 28 AUE_SENDMSG STD { int sendmsg( int s, _In_ struct msghdr *msg, int flags ); } 29 AUE_RECVFROM STD { int recvfrom( int s, _Out_writes_bytes_(len) void *buf, size_t len, int flags, _Out_writes_bytes_opt_(*fromlenaddr) struct sockaddr *from, _Inout_opt_ __socklen_t *fromlenaddr ); } 30 AUE_ACCEPT STD { int accept( int s, _Out_writes_bytes_opt_(*anamelen) struct sockaddr *name, _Inout_opt_ __socklen_t *anamelen ); } 31 AUE_GETPEERNAME STD { int getpeername( int fdes, _Out_writes_bytes_(*alen) struct sockaddr *asa, _Inout_opt_ __socklen_t *alen ); } 32 AUE_GETSOCKNAME STD { int getsockname( int fdes, _Out_writes_bytes_(*alen) struct sockaddr *asa, _Inout_ __socklen_t *alen ); } 33 AUE_ACCESS STD { int access( _In_z_ const char *path, int amode ); } 34 AUE_CHFLAGS STD { int chflags( _In_z_ const char *path, u_long flags ); } 35 AUE_FCHFLAGS STD { int fchflags( int fd, u_long flags ); } 36 AUE_SYNC STD { int sync(void); } 37 AUE_KILL STD { int kill( int pid, int signum ); } 38 AUE_STAT COMPAT { int stat( _In_z_ const char *path, _Out_ struct ostat *ub ); } 39 AUE_GETPPID STD { pid_t getppid(void); } 40 AUE_LSTAT COMPAT { int lstat( _In_z_ const char *path, _Out_ struct ostat *ub ); } 41 AUE_DUP STD { int dup( u_int fd ); } 42 AUE_PIPE COMPAT10 { int pipe(void); } 43 AUE_GETEGID STD { gid_t getegid(void); } 44 AUE_PROFILE STD { int profil( _Out_writes_bytes_(size) char *samples, size_t size, size_t offset, u_int scale ); } 45 AUE_KTRACE STD { int ktrace( _In_z_ const char *fname, int ops, int facs, int pid ); } 46 AUE_SIGACTION COMPAT { int sigaction( int signum, _In_opt_ struct osigaction *nsa, _Out_opt_ struct osigaction *osa ); } 47 AUE_GETGID STD { gid_t getgid(void); } 48 AUE_SIGPROCMASK COMPAT { int sigprocmask( int how, osigset_t mask ); } ; XXX note nonstandard (bogus) calling convention - the libc stub passes ; us the mask, not a pointer to it, and we return the old mask as the ; (int) return value. 49 AUE_GETLOGIN STD { int getlogin( _Out_writes_z_(namelen) char *namebuf, u_int namelen ); } 50 AUE_SETLOGIN STD { int setlogin( _In_z_ const char *namebuf ); } 51 AUE_ACCT STD { int acct( _In_z_ const char *path ); } 52 AUE_SIGPENDING COMPAT { int sigpending(void); } 53 AUE_SIGALTSTACK STD { int sigaltstack( _In_opt_ stack_t *ss, _Out_opt_ stack_t *oss ); } 54 AUE_IOCTL STD { int ioctl( int fd, u_long com, _Inout_opt_ char *data ); } 55 AUE_REBOOT STD { int reboot( int opt ); } 56 AUE_REVOKE STD { int revoke( _In_z_ const char *path ); } 57 AUE_SYMLINK STD { int symlink( _In_z_ const char *path, _In_z_ const char *link ); } 58 AUE_READLINK STD { ssize_t readlink( _In_z_ const char *path, _Out_writes_z_(count) char *buf, size_t count ); } 59 AUE_EXECVE STD { int execve( _In_z_ const char *fname, _In_z_ char **argv, _In_z_ char **envv ); } 60 AUE_UMASK STD { int umask( mode_t newmask ); } 61 AUE_CHROOT STD { int chroot( _In_z_ const char *path ); } 62 AUE_FSTAT COMPAT { int fstat( int fd, _Out_ struct ostat *sb ); } 63 AUE_NULL COMPAT { int getkerninfo( int op, _Out_writes_bytes_opt( *size) char *where, _Inout_opt_ size_t *size, int arg ); } 64 AUE_NULL COMPAT { int getpagesize(void); } 65 AUE_MSYNC STD { int msync( _In_ void *addr, size_t len, int flags ); } 66 AUE_VFORK STD { int vfork(void); } 67 AUE_NULL OBSOL vread 68 AUE_NULL OBSOL vwrite 69 AUE_SBRK STD { int sbrk( int incr ); } 70 AUE_SSTK STD { int sstk( int incr ); } 71 AUE_MMAP COMPAT { void *mmap( _In_ void *addr, int len, int prot, int flags, int fd, long pos ); } 72 AUE_O_VADVISE COMPAT11 { int vadvise( int anom ); } 73 AUE_MUNMAP STD { int munmap( _In_ void *addr, size_t len ); } 74 AUE_MPROTECT STD { int mprotect( _In_ void *addr, size_t len, int prot ); } 75 AUE_MADVISE STD { int madvise( _In_ void *addr, size_t len, int behav ); } 76 AUE_NULL OBSOL vhangup 77 AUE_NULL OBSOL vlimit 78 AUE_MINCORE STD { int mincore( _In_ const void *addr, size_t len, _Out_writes_bytes_(len/PAGE_SIZE) char *vec ); } 79 AUE_GETGROUPS STD { int getgroups( u_int gidsetsize, _Out_writes_opt_(gidsetsize) gid_t *gidset ); } 80 AUE_SETGROUPS STD { int setgroups( u_int gidsetsize, _In_reads_(gidsetsize) gid_t *gidset ); } 81 AUE_GETPGRP STD { int getpgrp(void); } 82 AUE_SETPGRP STD { int setpgid( int pid, int pgid ); } 83 AUE_SETITIMER STD { int setitimer( u_int which, _In_ struct itimerval *itv, _Out_opt_ struct itimerval *oitv ); } 84 AUE_WAIT4 COMPAT { int wait(void); } 85 AUE_SWAPON STD { int swapon( _In_z_ const char *name ); } 86 AUE_GETITIMER STD { int getitimer( u_int which, _Out_ struct itimerval *itv ); } 87 AUE_SYSCTL COMPAT { int gethostname( _Out_writes_z_(len) char *hostname, u_int len ); } 88 AUE_SYSCTL COMPAT { int sethostname( _In_reads_z_(len) char *hostname, u_int len ); } 89 AUE_GETDTABLESIZE STD { int getdtablesize(void); } 90 AUE_DUP2 STD { int dup2( u_int from, u_int to ); } 91 AUE_NULL UNIMPL getdopt 92 AUE_FCNTL STD { int fcntl( int fd, int cmd, long arg ); } ; XXX should be { int fcntl(int fd, int cmd, ...); } ; but we're not ready for varargs. 93 AUE_SELECT STD { int select( int nd, _Inout_opt_ fd_set *in, _Inout_opt_ fd_set *ou, _Inout_opt_ fd_set *ex, _In_opt_ struct timeval *tv ); } 94 AUE_NULL UNIMPL setdopt 95 AUE_FSYNC STD { int fsync( int fd ); } 96 AUE_SETPRIORITY STD { int setpriority( int which, int who, int prio ); } 97 AUE_SOCKET STD { int socket( int domain, int type, int protocol ); } 98 AUE_CONNECT STD { int connect( int s, _In_reads_bytes_(namelen) const struct sockaddr *name, int namelen ); } 99 AUE_ACCEPT COMPAT { int accept( int s, _Out_writes_bytes_opt_(*anamelen) struct sockaddr *name, int *anamelen ); } 100 AUE_GETPRIORITY STD { int getpriority( int which, int who ); } 101 AUE_SEND COMPAT { int send( int s, _In_reads_bytes_(len) const void *buf, int len, int flags ); } 102 AUE_RECV COMPAT { int recv( int s, _Out_writes_bytes_(len) void *buf, int len, int flags ); } 103 AUE_SIGRETURN COMPAT { int sigreturn( _In_ struct osigcontext *sigcntxp ); } 104 AUE_BIND STD { int bind( int s, _In_reads_bytes_(namelen) const struct sockaddr *name, int namelen ); } 105 AUE_SETSOCKOPT STD { int setsockopt( int s, int level, int name, _In_reads_bytes_opt_(valsize) const void *val, int valsize ); } 106 AUE_LISTEN STD { int listen( int s, int backlog ); } 107 AUE_NULL OBSOL vtimes 108 AUE_NULL COMPAT { int sigvec( int signum, _In_opt_ struct sigvec *nsv, _Out_opt_ struct sigvec *osv ); } 109 AUE_NULL COMPAT { int sigblock( int mask ); } 110 AUE_NULL COMPAT { int sigsetmask( int mask ); } 111 AUE_NULL COMPAT { int sigsuspend( osigset_t mask ); } ; XXX note nonstandard (bogus) calling convention - the libc stub passes ; us the mask, not a pointer to it. 112 AUE_NULL COMPAT { int sigstack( _In_opt_ struct sigstack *nss, _Out_opt_ struct sigstack *oss ); } 113 AUE_RECVMSG COMPAT { int recvmsg( int s, _Inout_ struct omsghdr *msg, int flags ); } 114 AUE_SENDMSG COMPAT { int sendmsg( int s, _In_ const void *msg, int flags ); } 115 AUE_NULL OBSOL vtrace 116 AUE_GETTIMEOFDAY STD { int gettimeofday( _Out_ struct timeval *tp, _Out_opt_ struct timezone *tzp ); } 117 AUE_GETRUSAGE STD { int getrusage( int who, _Out_ struct rusage *rusage ); } 118 AUE_GETSOCKOPT STD { int getsockopt( int s, int level, int name, _Out_writes_bytes_opt_(*avalsize) void *val, _Inout_ int *avalsize ); } 119 AUE_NULL UNIMPL resuba (BSD/OS 2.x) 120 AUE_READV STD { int readv( int fd, _Inout_updates_(iovcnt) struct iovec *iovp, u_int iovcnt ); } 121 AUE_WRITEV STD { int writev( int fd, _In_reads_opt_(iovcnt) struct iovec *iovp, u_int iovcnt ); } 122 AUE_SETTIMEOFDAY STD { int settimeofday( _In_ struct timeval *tv, _In_opt_ struct timezone *tzp ); } 123 AUE_FCHOWN STD { int fchown( int fd, int uid, int gid ); } 124 AUE_FCHMOD STD { int fchmod( int fd, mode_t mode ); } 125 AUE_RECVFROM COMPAT|NOARGS { int recvfrom( int s, _Out_writes_(len) void *buf, size_t len, int flags, _Out_writes_bytes_(*fromlenaddr) struct sockaddr *from, _Inout_ int *fromlenaddr ); } recvfrom recvfrom_args int 126 AUE_SETREUID STD { int setreuid( int ruid, int euid ); } 127 AUE_SETREGID STD { int setregid( int rgid, int egid ); } 128 AUE_RENAME STD { int rename( _In_z_ const char *from, _In_z_ const char *to ); } 129 AUE_TRUNCATE COMPAT { int truncate( _In_z_ const char *path, long length ); } 130 AUE_FTRUNCATE COMPAT { int ftruncate( int fd, long length ); } 131 AUE_FLOCK STD { int flock( int fd, int how ); } 132 AUE_MKFIFO STD { int mkfifo( _In_z_ const char *path, mode_t mode ); } 133 AUE_SENDTO STD { int sendto( int s, _In_reads_bytes_(len) const void *buf, size_t len, int flags, _In_reads_bytes_opt_(tolen) const struct sockaddr *to, int tolen ); } 134 AUE_SHUTDOWN STD { int shutdown( int s, int how ); } 135 AUE_SOCKETPAIR STD { int socketpair( int domain, int type, int protocol, _Out_writes_(2) int *rsv ); } 136 AUE_MKDIR STD { int mkdir( _In_z_ const char *path, mode_t mode ); } 137 AUE_RMDIR STD { int rmdir( _In_z_ const char *path ); } 138 AUE_UTIMES STD { int utimes( _In_z_ const char *path, _In_ struct timeval *tptr ); } 139 AUE_NULL OBSOL 4.2 sigreturn 140 AUE_ADJTIME STD { int adjtime( _In_ struct timeval *delta, _Out_opt_ struct timeval *olddelta ); } 141 AUE_GETPEERNAME COMPAT { int getpeername( int fdes, _Out_writes_bytes_(*alen) struct sockaddr *asa, _Inout_opt_ int *alen ); } 142 AUE_SYSCTL COMPAT { long gethostid(void); } 143 AUE_SYSCTL COMPAT { int sethostid( long hostid ); } 144 AUE_GETRLIMIT COMPAT { int getrlimit( u_int which, _Out_ struct orlimit *rlp ); } 145 AUE_SETRLIMIT COMPAT { int setrlimit( u_int which, _Out_ struct orlimit *rlp ); } 146 AUE_KILLPG COMPAT { int killpg( int pgid, int signum ); } 147 AUE_SETSID STD { int setsid(void); } 148 AUE_QUOTACTL STD { int quotactl( _In_z_ const char *path, int cmd, int uid, _In_ void *arg ); } 149 AUE_O_QUOTA COMPAT { int quota(void); } 150 AUE_GETSOCKNAME COMPAT|NOARGS { int getsockname( int fdec, _Out_writes_bytes_(*alen) struct sockaddr *asa, _Inout_ int *alen ); } getsockname getsockname_args int ; Syscalls 151-180 inclusive are reserved for vendor-specific ; system calls. (This includes various calls added for compatibity ; with other Unix variants.) ; Some of these calls are now supported by BSD... 151 AUE_NULL UNIMPL sem_lock (BSD/OS 2.x) 152 AUE_NULL UNIMPL sem_wakeup (BSD/OS 2.x) 153 AUE_NULL UNIMPL asyncdaemon (BSD/OS 2.x) ; 154 is initialised by the NLM code, if present. 154 AUE_NULL NOSTD { int nlm_syscall( int debug_level, int grace_period, int addr_count, _In_reads_(addr_count) char **addrs ); } ; 155 is initialized by the NFS code, if present. 155 AUE_NFS_SVC NOSTD { int nfssvc( int flag, _In_ void *argp ); } 156 AUE_GETDIRENTRIES COMPAT { int getdirentries( int fd, _Out_writes_bytes_(count) char *buf, u_int count, _Out_ long *basep ); } 157 AUE_STATFS COMPAT4 { int statfs( _In_z_ const char *path, _Out_ struct ostatfs *buf ); } 158 AUE_FSTATFS COMPAT4 { int fstatfs( int fd, _Out_ struct ostatfs *buf ); } 159 AUE_NULL UNIMPL nosys 160 AUE_LGETFH STD { int lgetfh( _In_z_ const char *fname, _Out_ struct fhandle *fhp ); } 161 AUE_NFS_GETFH STD { int getfh( _In_z_ const char *fname, _Out_ struct fhandle *fhp ); } 162 AUE_SYSCTL COMPAT4 { int getdomainname( _Out_writes_z_(len) char *domainname, int len ); } 163 AUE_SYSCTL COMPAT4 { int setdomainname( _In_reads_z_(len) char *domainname, int len ); } 164 AUE_NULL COMPAT4 { int uname( _Out_ struct utsname *name ); } 165 AUE_SYSARCH STD { int sysarch( int op, _In_z_ char *parms ); } 166 AUE_RTPRIO STD { int rtprio( int function, pid_t pid, _Inout_ struct rtprio *rtp ); } 167 AUE_NULL UNIMPL nosys 168 AUE_NULL UNIMPL nosys 169 AUE_SEMSYS NOSTD { int semsys( int which, int a2, int a3, int a4, int a5 ); } ; XXX should be { int semsys(int which, ...); } 170 AUE_MSGSYS NOSTD { int msgsys( int which, int a2, int a3, int a4, int a5, int a6 ); } ; XXX should be { int msgsys(int which, ...); } 171 AUE_SHMSYS NOSTD { int shmsys( int which, int a2, int a3, int a4 ); } ; XXX should be { int shmsys(int which, ...); } 172 AUE_NULL UNIMPL nosys 173 AUE_PREAD COMPAT6 { ssize_t pread( int fd, _Out_writes_bytes_(nbyte) void *buf, size_t nbyte, int pad, off_t offset ); } 174 AUE_PWRITE COMPAT6 { ssize_t pwrite( int fd, _In_reads_bytes_(nbyte) const void *buf, size_t nbyte, int pad, off_t offset ); } 175 AUE_SETFIB STD { int setfib( int fibnum ); } 176 AUE_NTP_ADJTIME STD { int ntp_adjtime( _Inout_ struct timex *tp ); } 177 AUE_NULL UNIMPL sfork (BSD/OS 2.x) 178 AUE_NULL UNIMPL getdescriptor (BSD/OS 2.x) 179 AUE_NULL UNIMPL setdescriptor (BSD/OS 2.x) 180 AUE_NULL UNIMPL nosys ; Syscalls 181-199 are used by/reserved for BSD 181 AUE_SETGID STD { int setgid( gid_t gid ); } 182 AUE_SETEGID STD { int setegid( gid_t egid ); } 183 AUE_SETEUID STD { int seteuid( uid_t euid ); } 184 AUE_NULL OBSOL lfs_bmapv 185 AUE_NULL OBSOL lfs_markv 186 AUE_NULL OBSOL lfs_segclean 187 AUE_NULL OBSOL lfs_segwait 188 AUE_STAT COMPAT11 { int stat( _In_z_ const char *path, _Out_ struct freebsd11_stat *ub ); } 189 AUE_FSTAT COMPAT11 { int fstat( int fd, _Out_ struct freebsd11_stat *sb ); } 190 AUE_LSTAT COMPAT11 { int lstat( _In_z_ const char *path, _Out_ struct freebsd11_stat *ub ); } 191 AUE_PATHCONF STD { int pathconf( _In_z_ const char *path, int name ); } 192 AUE_FPATHCONF STD { int fpathconf( int fd, int name ); } 193 AUE_NULL UNIMPL nosys 194 AUE_GETRLIMIT STD { int getrlimit( u_int which, _Out_ struct rlimit *rlp ); } getrlimit __getrlimit_args int 195 AUE_SETRLIMIT STD { int setrlimit( u_int which, _In_ struct rlimit *rlp ); } setrlimit __setrlimit_args int 196 AUE_GETDIRENTRIES COMPAT11 { int getdirentries( int fd, _Out_writes_bytes_(count) char *buf, u_int count, _Out_ long *basep ); } 197 AUE_MMAP COMPAT6 { void *mmap( _In_ void *addr, size_t len, int prot, int flags, int fd, int pad, off_t pos ); } 198 AUE_NULL NOPROTO { int nosys(void); } __syscall __syscall_args int 199 AUE_LSEEK COMPAT6 { off_t lseek( int fd, int pad, off_t offset, int whence ); } 200 AUE_TRUNCATE COMPAT6 { int truncate( _In_z_ const char *path, int pad, off_t length ); } 201 AUE_FTRUNCATE COMPAT6 { int ftruncate( int fd, int pad, off_t length ); } 202 AUE_SYSCTL STD { int __sysctl( _In_reads_(namelen) int *name, u_int namelen, _Out_writes_bytes_opt_(*oldlenp) void *old, _Inout_opt_ size_t *oldlenp, _In_reads_bytes_opt_(newlen) const void *new, size_t newlen ); } __sysctl sysctl_args int 203 AUE_MLOCK STD { int mlock( _In_ const void *addr, size_t len ); } 204 AUE_MUNLOCK STD { int munlock( _In_ const void *addr, size_t len ); } 205 AUE_UNDELETE STD { int undelete( _In_z_ const char *path ); } 206 AUE_FUTIMES STD { int futimes( int fd, _In_reads_(2) struct timeval *tptr ); } 207 AUE_GETPGID STD { int getpgid( pid_t pid ); } 208 AUE_NULL UNIMPL nosys 209 AUE_POLL STD { int poll( _Inout_updates_(nfds) struct pollfd *fds, u_int nfds, int timeout ); } ; ; The following are reserved for loadable syscalls ; 210 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 211 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 212 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 213 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 214 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 215 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 216 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 217 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 218 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 219 AUE_NULL NODEF|NOTSTATIC lkmnosys lkmnosys nosys_args int 220 AUE_SEMCTL COMPAT7|NOSTD { int __semctl( int semid, int semnum, int cmd, union semun_old *arg ); } 221 AUE_SEMGET NOSTD { int semget( key_t key, int nsems, int semflg ); } 222 AUE_SEMOP NOSTD { int semop( int semid, _In_reads_(nsops) struct sembuf *sops, size_t nsops ); } 223 AUE_NULL OBSOL semconfig 224 AUE_MSGCTL COMPAT7|NOSTD { int msgctl( int msqid, int cmd, struct msqid_ds_old *buf ); } 225 AUE_MSGGET NOSTD { int msgget( key_t key, int msgflg ); } 226 AUE_MSGSND NOSTD { int msgsnd( int msqid, _In_reads_bytes_(msgsz) const void *msgp, size_t msgsz, int msgflg ); } 227 AUE_MSGRCV NOSTD { ssize_t msgrcv( int msqid, _Out_writes_bytes_(msgsz) void *msgp, size_t msgsz, long msgtyp, int msgflg ); } 228 AUE_SHMAT NOSTD { void *shmat( int shmid, _In_ const void *shmaddr, int shmflg ); } 229 AUE_SHMCTL COMPAT7|NOSTD { int shmctl( int shmid, int cmd, struct shmid_ds_old *buf ); } 230 AUE_SHMDT NOSTD { int shmdt( _In_ const void *shmaddr ); } 231 AUE_SHMGET NOSTD { int shmget( key_t key, size_t size, int shmflg ); } 232 AUE_NULL STD { int clock_gettime( clockid_t clock_id, _Out_ struct timespec *tp ); } 233 AUE_CLOCK_SETTIME STD { int clock_settime( clockid_t clock_id, _In_ const struct timespec *tp ); } 234 AUE_NULL STD { int clock_getres( clockid_t clock_id, _Out_ struct timespec *tp ); } 235 AUE_NULL STD { int ktimer_create( clockid_t clock_id, _In_ struct sigevent *evp, _Out_ int *timerid ); } 236 AUE_NULL STD { int ktimer_delete( int timerid ); } 237 AUE_NULL STD { int ktimer_settime( int timerid, int flags, _In_ const struct itimerspec *value, _Out_opt_ struct itimerspec *ovalue ); } 238 AUE_NULL STD { int ktimer_gettime( int timerid, _Out_ struct itimerspec *value ); } 239 AUE_NULL STD { int ktimer_getoverrun( int timerid ); } 240 AUE_NULL STD { int nanosleep( _In_ const struct timespec *rqtp, _Out_opt_ struct timespec *rmtp ); } 241 AUE_NULL STD { int ffclock_getcounter( _Out_ ffcounter *ffcount ); } 242 AUE_NULL STD { int ffclock_setestimate( _In_ struct ffclock_estimate *cest ); } 243 AUE_NULL STD { int ffclock_getestimate( _Out_ struct ffclock_estimate *cest ); } 244 AUE_NULL STD { int clock_nanosleep( clockid_t clock_id, int flags, _In_ const struct timespec *rqtp, _Out_opt_ struct timespec *rmtp ); } 245-246 AUE_NULL UNIMPL nosys 247 AUE_NULL STD { int clock_getcpuclockid2( id_t id, int which, _Out_ clockid_t *clock_id ); } 248 AUE_NULL STD { int ntp_gettime( _Out_ struct ntptimeval *ntvp ); } 249 AUE_NULL UNIMPL nosys ; syscall numbers initially used in OpenBSD 250 AUE_MINHERIT STD { int minherit( _In_ void *addr, size_t len, int inherit ); } 251 AUE_RFORK STD { int rfork( int flags ); } 252 AUE_POLL OBSOL openbsd_poll 253 AUE_ISSETUGID STD { int issetugid(void); } 254 AUE_LCHOWN STD { int lchown( _In_z_ const char *path, int uid, int gid ); } 255 AUE_AIO_READ STD { int aio_read( _Inout_ struct aiocb *aiocbp ); } 256 AUE_AIO_WRITE STD { int aio_write( _Inout_ struct aiocb *aiocbp ); } 257 AUE_LIO_LISTIO STD { int lio_listio( int mode, _Inout_updates_(nent) struct aiocb* const *acb_list, int nent, _In_opt_ struct sigevent *sig ); } 258-271 AUE_NULL UNIMPL nosys 272 AUE_O_GETDENTS COMPAT11 { int getdents( int fd, _Out_writes_bytes_(count) char *buf, size_t count ); } 273 AUE_NULL UNIMPL nosys 274 AUE_LCHMOD STD { int lchmod( _In_z_ const char *path, mode_t mode ); } 275 AUE_NULL OBSOL netbsd_lchown 276 AUE_LUTIMES STD { int lutimes( _In_z_ const char *path, _In_ struct timeval *tptr ); } 277 AUE_NULL OBSOL netbsd_msync 278 AUE_STAT COMPAT11 { int nstat( _In_z_ const char *path, _Out_ struct nstat *ub ); } 279 AUE_FSTAT COMPAT11 { int nfstat( int fd, _Out_ struct nstat *sb ); } 280 AUE_LSTAT COMPAT11 { int nlstat( _In_z_ const char *path, _Out_ struct nstat *ub ); } 281-288 AUE_NULL UNIMPL nosys 289 AUE_PREADV STD { ssize_t preadv( int fd, _In_reads_(iovcnt) struct iovec *iovp, u_int iovcnt, off_t offset ); } 290 AUE_PWRITEV STD { ssize_t pwritev( int fd, _In_reads_(iovcnt) struct iovec *iovp, u_int iovcnt, off_t offset ); } 291-296 AUE_NULL UNIMPL nosys 297 AUE_FHSTATFS COMPAT4 { int fhstatfs( _In_ const struct fhandle *u_fhp, _Out_ struct ostatfs *buf ); } 298 AUE_FHOPEN STD { int fhopen( _In_ const struct fhandle *u_fhp, int flags ); } 299 AUE_FHSTAT COMPAT11 { int fhstat( _In_ const struct fhandle *u_fhp, _Out_ struct freebsd11_stat *sb ); } 300 AUE_NULL STD { int modnext( int modid ); } 301 AUE_NULL STD { int modstat( int modid, _Out_ struct module_stat* stat ); } 302 AUE_NULL STD { int modfnext( int modid ); } 303 AUE_NULL STD { int modfind( _In_z_ const char *name ); } 304 AUE_MODLOAD STD { int kldload( _In_z_ const char *file ); } 305 AUE_MODUNLOAD STD { int kldunload( int fileid ); } 306 AUE_NULL STD { int kldfind( _In_z_ const char *file ); } 307 AUE_NULL STD { int kldnext( int fileid ); } 308 AUE_NULL STD { int kldstat( int fileid, _Out_ struct kld_file_stat *stat ); } 309 AUE_NULL STD { int kldfirstmod( int fileid ); } 310 AUE_GETSID STD { int getsid( pid_t pid ); } 311 AUE_SETRESUID STD { int setresuid( uid_t ruid, uid_t euid, uid_t suid ); } 312 AUE_SETRESGID STD { int setresgid( gid_t rgid, gid_t egid, gid_t sgid ); } 313 AUE_NULL OBSOL signanosleep 314 AUE_AIO_RETURN STD { ssize_t aio_return( _Inout_ struct aiocb *aiocbp ); } 315 AUE_AIO_SUSPEND STD { int aio_suspend( _Inout_updates_(nent) struct aiocb * const * aiocbp, int nent, _In_opt_ const struct timespec *timeout ); } 316 AUE_AIO_CANCEL STD { int aio_cancel( int fd, _In_opt_ struct aiocb *aiocbp ); } 317 AUE_AIO_ERROR STD { int aio_error( _In_ struct aiocb *aiocbp ); } 318 AUE_AIO_READ COMPAT6 { int aio_read( _Inout_ struct oaiocb *aiocbp ); } 319 AUE_AIO_WRITE COMPAT6 { int aio_write( _Inout_ struct oaiocb *aiocbp ); } 320 AUE_LIO_LISTIO COMPAT6 { int lio_listio( int mode, _Inout_updates_(nent) struct oaiocb * const *acb_list, int nent, _In_opt_ struct osigevent *sig ); } 321 AUE_NULL STD { int yield(void); } 322 AUE_NULL OBSOL thr_sleep 323 AUE_NULL OBSOL thr_wakeup 324 AUE_MLOCKALL STD { int mlockall( int how ); } 325 AUE_MUNLOCKALL STD { int munlockall(void); } 326 AUE_GETCWD STD { int __getcwd( _Out_writes_z_(buflen) char *buf, size_t buflen ); } 327 AUE_NULL STD { int sched_setparam( pid_t pid, _In_ const struct sched_param *param ); } 328 AUE_NULL STD { int sched_getparam( pid_t pid, _Out_ struct sched_param *param ); } 329 AUE_NULL STD { int sched_setscheduler( pid_t pid, int policy, _In_ const struct sched_param *param ); } 330 AUE_NULL STD { int sched_getscheduler( pid_t pid ); } 331 AUE_NULL STD { int sched_yield(void); } 332 AUE_NULL STD { int sched_get_priority_max( int policy ); } 333 AUE_NULL STD { int sched_get_priority_min( int policy ); } 334 AUE_NULL STD { int sched_rr_get_interval( pid_t pid, _Out_ struct timespec *interval ); } 335 AUE_NULL STD { int utrace( _In_reads_bytes_(len) const void *addr, size_t len ); } 336 AUE_SENDFILE COMPAT4 { int sendfile( int fd, int s, off_t offset, size_t nbytes, _In_opt_ struct sf_hdtr *hdtr, _Out_opt_ off_t *sbytes, int flags ); } 337 AUE_NULL STD { int kldsym( int fileid, int cmd, _In_ void *data ); } 338 AUE_JAIL STD { int jail( _In_ struct jail *jail ); } 339 AUE_NULL NOSTD|NOTSTATIC { int nnpfs_syscall( int operation, char *a_pathP, int a_opcode, void *a_paramsP, int a_followSymlinks ); } 340 AUE_SIGPROCMASK STD { int sigprocmask( int how, _In_opt_ const sigset_t *set, _Out_opt_ sigset_t *oset ); } 341 AUE_SIGSUSPEND STD { int sigsuspend( _In_ const sigset_t *sigmask ); } 342 AUE_SIGACTION COMPAT4 { int sigaction( int sig, _In_opt_ const struct sigaction *act, _Out_opt_ struct sigaction *oact ); } 343 AUE_SIGPENDING STD { int sigpending( _In_ sigset_t *set ); } 344 AUE_SIGRETURN COMPAT4 { int sigreturn( _In_ const struct ucontext4 *sigcntxp ); } 345 AUE_SIGWAIT STD { int sigtimedwait( _In_ const sigset_t *set, _Out_opt_ siginfo_t *info, _In_opt_ const struct timespec *timeout ); } 346 AUE_NULL STD { int sigwaitinfo( _In_ const sigset_t *set, _Out_opt_ siginfo_t *info ); } 347 AUE_ACL_GET_FILE STD { int __acl_get_file( _In_z_ const char *path, acl_type_t type, _Out_ struct acl *aclp ); } 348 AUE_ACL_SET_FILE STD { int __acl_set_file( _In_z_ const char *path, acl_type_t type, _In_ struct acl *aclp ); } 349 AUE_ACL_GET_FD STD { int __acl_get_fd( int filedes, acl_type_t type, _Out_ struct acl *aclp ); } 350 AUE_ACL_SET_FD STD { int __acl_set_fd( int filedes, acl_type_t type, _In_ struct acl *aclp ); } 351 AUE_ACL_DELETE_FILE STD { int __acl_delete_file( _In_z_ const char *path, acl_type_t type ); } 352 AUE_ACL_DELETE_FD STD { int __acl_delete_fd( int filedes, acl_type_t type ); } 353 AUE_ACL_CHECK_FILE STD { int __acl_aclcheck_file( _In_z_ const char *path, acl_type_t type, _In_ struct acl *aclp ); } 354 AUE_ACL_CHECK_FD STD { int __acl_aclcheck_fd( int filedes, acl_type_t type, _In_ struct acl *aclp ); } 355 AUE_EXTATTRCTL STD { int extattrctl( _In_z_ const char *path, int cmd, _In_z_opt_ const char *filename, int attrnamespace, _In_z_ const char *attrname ); } 356 AUE_EXTATTR_SET_FILE STD { ssize_t extattr_set_file( _In_z_ const char *path, int attrnamespace, _In_z_ const char *attrname, _In_reads_bytes_(nbytes) void *data, size_t nbytes ); } 357 AUE_EXTATTR_GET_FILE STD { ssize_t extattr_get_file( _In_z_ const char *path, int attrnamespace, _In_z_ const char *attrname, _Out_writes_bytes_(nbytes) void *data, size_t nbytes ); } 358 AUE_EXTATTR_DELETE_FILE STD { int extattr_delete_file( _In_z_ const char *path, int attrnamespace, _In_z_ const char *attrname ); } 359 AUE_AIO_WAITCOMPLETE STD { ssize_t aio_waitcomplete( _Outptr_result_maybenull_ struct aiocb **aiocbp, _In_opt_ struct timespec *timeout ); } 360 AUE_GETRESUID STD { int getresuid( _Out_opt_ uid_t *ruid, _Out_opt_ uid_t *euid, _Out_opt_ uid_t *suid ); } 361 AUE_GETRESGID STD { int getresgid( _Out_opt_ gid_t *rgid, _Out_opt_ gid_t *egid, _Out_opt_ gid_t *sgid ); } 362 AUE_KQUEUE STD { int kqueue(void); } 363 AUE_KEVENT COMPAT11 { int kevent( int fd, _In_reads_opt_(nchanges) struct kevent_freebsd11 *changelist, int nchanges, _Out_writes_opt_(nevents) struct kevent_freebsd11 *eventlist, int nevents, _In_opt_ const struct timespec *timeout ); } 364 AUE_NULL OBSOL __cap_get_proc 365 AUE_NULL OBSOL __cap_set_proc 366 AUE_NULL OBSOL __cap_get_fd 367 AUE_NULL OBSOL __cap_get_file 368 AUE_NULL OBSOL __cap_set_fd 369 AUE_NULL OBSOL __cap_set_file 370 AUE_NULL UNIMPL nosys 371 AUE_EXTATTR_SET_FD STD { ssize_t extattr_set_fd( int fd, int attrnamespace, _In_z_ const char *attrname, _In_reads_bytes_(nbytes) void *data, size_t nbytes ); } 372 AUE_EXTATTR_GET_FD STD { ssize_t extattr_get_fd( int fd, int attrnamespace, _In_z_ const char *attrname, _Out_writes_bytes_(nbytes) void *data, size_t nbytes ); } 373 AUE_EXTATTR_DELETE_FD STD { int extattr_delete_fd( int fd, int attrnamespace, _In_z_ const char *attrname ); } 374 AUE_SETUGID STD { int __setugid( int flag ); } 375 AUE_NULL OBSOL nfsclnt 376 AUE_EACCESS STD { int eaccess( _In_z_ const char *path, int amode ); } 377 AUE_NULL NOSTD|NOTSTATIC { int afs3_syscall( long syscall, long parm1, long parm2, long parm3, long parm4, long parm5, long parm6 ); } 378 AUE_NMOUNT STD { int nmount( _In_reads_(iovcnt) struct iovec *iovp, unsigned int iovcnt, int flags ); } 379 AUE_NULL OBSOL kse_exit 380 AUE_NULL OBSOL kse_wakeup 381 AUE_NULL OBSOL kse_create 382 AUE_NULL OBSOL kse_thr_interrupt 383 AUE_NULL OBSOL kse_release 384 AUE_NULL STD { int __mac_get_proc( _In_ struct mac *mac_p ); } 385 AUE_NULL STD { int __mac_set_proc( _In_ struct mac *mac_p ); } 386 AUE_NULL STD { int __mac_get_fd( int fd, _In_ struct mac *mac_p ); } 387 AUE_NULL STD { int __mac_get_file( _In_z_ const char *path_p, _In_ struct mac *mac_p ); } 388 AUE_NULL STD { int __mac_set_fd( int fd, _In_ struct mac *mac_p ); } 389 AUE_NULL STD { int __mac_set_file( _In_z_ const char *path_p, _In_ struct mac *mac_p ); } 390 AUE_NULL STD { int kenv( int what, _In_z_opt_ const char *name, _Inout_updates_opt_(len) char *value, int len ); } 391 AUE_LCHFLAGS STD { int lchflags( _In_z_ const char *path, u_long flags ); } 392 AUE_NULL STD { int uuidgen( _Out_writes_(count) struct uuid *store, int count ); } 393 AUE_SENDFILE STD { int sendfile( int fd, int s, off_t offset, size_t nbytes, _In_opt_ struct sf_hdtr *hdtr, _Out_opt_ off_t *sbytes, int flags ); } 394 AUE_NULL STD { int mac_syscall( _In_z_ const char *policy, int call, _In_opt_ void *arg ); } 395 AUE_GETFSSTAT COMPAT11 { int getfsstat( _Out_writes_bytes_opt_(bufsize) struct freebsd11_statfs *buf, long bufsize, int mode ); } 396 AUE_STATFS COMPAT11 { int statfs( _In_z_ const char *path, _Out_ struct freebsd11_statfs *buf ); } 397 AUE_FSTATFS COMPAT11 { int fstatfs( int fd, _Out_ struct freebsd11_statfs *buf ); } 398 AUE_FHSTATFS COMPAT11 { int fhstatfs( _In_ const struct fhandle *u_fhp, _Out_ struct freebsd11_statfs *buf ); } 399 AUE_NULL UNIMPL nosys 400 AUE_SEMCLOSE NOSTD { int ksem_close( semid_t id ); } 401 AUE_SEMPOST NOSTD { int ksem_post( semid_t id ); } 402 AUE_SEMWAIT NOSTD { int ksem_wait( semid_t id ); } 403 AUE_SEMTRYWAIT NOSTD { int ksem_trywait( semid_t id ); } 404 AUE_SEMINIT NOSTD { int ksem_init( _Out_ semid_t *idp, unsigned int value ); } 405 AUE_SEMOPEN NOSTD { int ksem_open( _Out_ semid_t *idp, _In_z_ const char *name, int oflag, mode_t mode, unsigned int value ); } 406 AUE_SEMUNLINK NOSTD { int ksem_unlink( _In_z_ const char *name ); } 407 AUE_SEMGETVALUE NOSTD { int ksem_getvalue( semid_t id, _Out_ int *val ); } 408 AUE_SEMDESTROY NOSTD { int ksem_destroy( semid_t id ); } 409 AUE_NULL STD { int __mac_get_pid( pid_t pid, _In_ struct mac *mac_p ); } 410 AUE_NULL STD { int __mac_get_link( _In_z_ const char *path_p, _In_ struct mac *mac_p ); } 411 AUE_NULL STD { int __mac_set_link( _In_z_ const char *path_p, _In_ struct mac *mac_p ); } 412 AUE_EXTATTR_SET_LINK STD { ssize_t extattr_set_link( _In_z_ const char *path, int attrnamespace, _In_z_ const char *attrname, _In_reads_bytes_(nbytes) void *data, size_t nbytes ); } 413 AUE_EXTATTR_GET_LINK STD { ssize_t extattr_get_link( _In_z_ const char *path, int attrnamespace, _In_z_ const char *attrname, _Out_writes_bytes_(nbytes) void *data, size_t nbytes ); } 414 AUE_EXTATTR_DELETE_LINK STD { int extattr_delete_link( _In_z_ const char *path, int attrnamespace, _In_z_ const char *attrname ); } 415 AUE_NULL STD { int __mac_execve( _In_z_ const char *fname, _In_ char **argv, _In_ char **envv, _In_ struct mac *mac_p ); } 416 AUE_SIGACTION STD { int sigaction( int sig, _In_opt_ const struct sigaction *act, _Out_opt_ struct sigaction *oact ); } 417 AUE_SIGRETURN STD { int sigreturn( _In_ const struct __ucontext *sigcntxp ); } 418 AUE_NULL UNIMPL __xstat 419 AUE_NULL UNIMPL __xfstat 420 AUE_NULL UNIMPL __xlstat 421 AUE_NULL STD { int getcontext( _Out_ struct __ucontext *ucp ); } 422 AUE_NULL STD { int setcontext( _In_ const struct __ucontext *ucp ); } 423 AUE_NULL STD { int swapcontext( _Out_ struct __ucontext *oucp, _In_ const struct __ucontext *ucp ); } 424 AUE_SWAPOFF STD { int swapoff( _In_z_ const char *name ); } 425 AUE_ACL_GET_LINK STD { int __acl_get_link( _In_z_ const char *path, acl_type_t type, _Out_ struct acl *aclp ); } 426 AUE_ACL_SET_LINK STD { int __acl_set_link( _In_z_ const char *path, acl_type_t type, _In_ struct acl *aclp ); } 427 AUE_ACL_DELETE_LINK STD { int __acl_delete_link( _In_z_ const char *path, acl_type_t type ); } 428 AUE_ACL_CHECK_LINK STD { int __acl_aclcheck_link( _In_z_ const char *path, acl_type_t type, _In_ struct acl *aclp ); } 429 AUE_SIGWAIT STD { int sigwait( _In_ const sigset_t *set, _Out_ int *sig ); } 430 AUE_THR_CREATE STD { int thr_create( _In_ ucontext_t *ctx, _Out_ long *id, int flags ); } 431 AUE_THR_EXIT STD { void thr_exit( _Out_opt_ long *state ); } 432 AUE_NULL STD { int thr_self( _Out_ long *id ); } 433 AUE_THR_KILL STD { int thr_kill( long id, int sig ); } 434-435 AUE_NULL UNIMPL nosys 436 AUE_JAIL_ATTACH STD { int jail_attach( int jid ); } 437 AUE_EXTATTR_LIST_FD STD { ssize_t extattr_list_fd( int fd, int attrnamespace, _Out_writes_bytes_opt_(nbytes) void *data, size_t nbytes ); } 438 AUE_EXTATTR_LIST_FILE STD { ssize_t extattr_list_file( _In_z_ const char *path, int attrnamespace, _Out_writes_bytes_opt_(nbytes) void *data, size_t nbytes ); } 439 AUE_EXTATTR_LIST_LINK STD { ssize_t extattr_list_link( _In_z_ const char *path, int attrnamespace, _Out_writes_bytes_opt_(nbytes) void *data, size_t nbytes ); } 440 AUE_NULL OBSOL kse_switchin 441 AUE_SEMWAIT NOSTD { int ksem_timedwait( semid_t id, _In_opt_ const struct timespec *abstime ); } 442 AUE_NULL STD { int thr_suspend( _In_opt_ const struct timespec *timeout ); } 443 AUE_NULL STD { int thr_wake( long id ); } 444 AUE_MODUNLOAD STD { int kldunloadf( int fileid, int flags ); } 445 AUE_AUDIT STD { int audit( _In_reads_bytes_(length) const void *record, u_int length ); } 446 AUE_AUDITON STD { int auditon( int cmd, _In_opt_ void *data, u_int length ); } 447 AUE_GETAUID STD { int getauid( _Out_ uid_t *auid ); } 448 AUE_SETAUID STD { int setauid( _In_ uid_t *auid ); } 449 AUE_GETAUDIT STD { int getaudit( _Out_ struct auditinfo *auditinfo ); } 450 AUE_SETAUDIT STD { int setaudit( _In_ struct auditinfo *auditinfo ); } 451 AUE_GETAUDIT_ADDR STD { int getaudit_addr( _Out_writes_bytes_(length) struct auditinfo_addr *auditinfo_addr, u_int length ); } 452 AUE_SETAUDIT_ADDR STD { int setaudit_addr( _In_reads_bytes_(length) struct auditinfo_addr *auditinfo_addr, u_int length ); } 453 AUE_AUDITCTL STD { int auditctl( _In_z_ const char *path ); } 454 AUE_NULL STD { int _umtx_op( _Inout_ void *obj, int op, u_long val, _In_ void *uaddr1, _In_ void *uaddr2 ); } 455 AUE_THR_NEW STD { int thr_new( _In_ struct thr_param *param, int param_size ); } 456 AUE_NULL STD { int sigqueue( pid_t pid, int signum, _In_ void *value ); } 457 AUE_MQ_OPEN NOSTD { int kmq_open( _In_z_ const char *path, int flags, mode_t mode, _In_opt_ const struct mq_attr *attr ); } 458 AUE_MQ_SETATTR NOSTD { int kmq_setattr( int mqd, _In_opt_ const struct mq_attr *attr, _Out_opt_ struct mq_attr *oattr ); } 459 AUE_MQ_TIMEDRECEIVE NOSTD { int kmq_timedreceive( int mqd, _Out_writes_bytes_(msg_len) char *msg_ptr, size_t msg_len, _Out_opt_ unsigned *msg_prio, _In_opt_ const struct timespec *abs_timeout ); } 460 AUE_MQ_TIMEDSEND NOSTD { int kmq_timedsend( int mqd, _In_reads_bytes_(msg_len) const char *msg_ptr, size_t msg_len, unsigned msg_prio, _In_opt_ const struct timespec *abs_timeout ); } 461 AUE_MQ_NOTIFY NOSTD { int kmq_notify( int mqd, _In_opt_ const struct sigevent *sigev ); } 462 AUE_MQ_UNLINK NOSTD { int kmq_unlink( _In_z_ const char *path ); } 463 AUE_NULL STD { int abort2( _In_z_ const char *why, int nargs, _In_reads_(nargs) void **args ); } 464 AUE_NULL STD { int thr_set_name( long id, _In_z_ const char *name ); } 465 AUE_AIO_FSYNC STD { int aio_fsync( int op, _In_ struct aiocb *aiocbp ); } 466 AUE_RTPRIO STD { int rtprio_thread( int function, lwpid_t lwpid, _Inout_ struct rtprio *rtp ); } 467-468 AUE_NULL UNIMPL nosys 469 AUE_NULL UNIMPL __getpath_fromfd 470 AUE_NULL UNIMPL __getpath_fromaddr 471 AUE_SCTP_PEELOFF NOSTD { int sctp_peeloff( int sd, uint32_t name ); } 472 AUE_SCTP_GENERIC_SENDMSG NOSTD { int sctp_generic_sendmsg( int sd, _In_reads_bytes_(mlen) void *msg, int mlen, _In_reads_bytes_(tolen) struct sockaddr *to, __socklen_t tolen, _In_opt_ struct sctp_sndrcvinfo *sinfo, int flags ); } 473 AUE_SCTP_GENERIC_SENDMSG_IOV NOSTD { int sctp_generic_sendmsg_iov( int sd, _In_reads_(iovlen) struct iovec *iov, int iovlen, _In_reads_bytes_(tolen) struct sockaddr *to, __socklen_t tolen, _In_opt_ struct sctp_sndrcvinfo *sinfo, int flags ); } 474 AUE_SCTP_GENERIC_RECVMSG NOSTD { int sctp_generic_recvmsg( int sd, _In_reads_(iovlen) struct iovec *iov, int iovlen, _Out_writes_bytes_(*fromlenaddr) struct sockaddr *from, _Out_ __socklen_t *fromlenaddr, _In_opt_ struct sctp_sndrcvinfo *sinfo, _Out_opt_ int *msg_flags ); } 475 AUE_PREAD STD { ssize_t pread( int fd, _Out_writes_bytes_(nbyte) void *buf, size_t nbyte, off_t offset ); } 476 AUE_PWRITE STD { ssize_t pwrite( int fd, _In_reads_bytes_(nbyte) const void *buf, size_t nbyte, off_t offset ); } 477 AUE_MMAP STD { void *mmap( _In_ void *addr, size_t len, int prot, int flags, int fd, off_t pos ); } 478 AUE_LSEEK STD { off_t lseek( int fd, off_t offset, int whence ); } 479 AUE_TRUNCATE STD { int truncate( _In_z_ const char *path, off_t length ); } 480 AUE_FTRUNCATE STD { int ftruncate( int fd, off_t length ); } 481 AUE_THR_KILL2 STD { int thr_kill2( pid_t pid, long id, int sig ); } 482 AUE_SHMOPEN STD { int shm_open( _In_z_ const char *path, int flags, mode_t mode ); } 483 AUE_SHMUNLINK STD { int shm_unlink( _In_z_ const char *path ); } 484 AUE_NULL STD { int cpuset( _Out_ cpusetid_t *setid ); } 485 AUE_NULL STD { int cpuset_setid( cpuwhich_t which, id_t id, cpusetid_t setid ); } 486 AUE_NULL STD { int cpuset_getid( cpulevel_t level, cpuwhich_t which, id_t id, _Out_ cpusetid_t *setid ); } 487 AUE_NULL STD { int cpuset_getaffinity( cpulevel_t level, cpuwhich_t which, id_t id, size_t cpusetsize, _Out_ cpuset_t *mask ); } 488 AUE_NULL STD { int cpuset_setaffinity( cpulevel_t level, cpuwhich_t which, id_t id, size_t cpusetsize, _Out_ const cpuset_t *mask ); } 489 AUE_FACCESSAT STD { int faccessat( int fd, _In_z_ const char *path, int amode, int flag ); } 490 AUE_FCHMODAT STD { int fchmodat( int fd, _In_z_ const char *path, mode_t mode, int flag ); } 491 AUE_FCHOWNAT STD { int fchownat( int fd, _In_z_ const char *path, uid_t uid, gid_t gid, int flag ); } 492 AUE_FEXECVE STD { int fexecve( int fd, _In_ char **argv, _In_ char **envv ); } 493 AUE_FSTATAT COMPAT11 { int fstatat( int fd, _In_z_ const char *path, _Out_ struct freebsd11_stat *buf, int flag ); } 494 AUE_FUTIMESAT STD { int futimesat( int fd, _In_z_ const char *path, _In_reads_(2) struct timeval *times ); } 495 AUE_LINKAT STD { int linkat( int fd1, _In_z_ const char *path1, int fd2, _In_z_ const char *path2, int flag ); } 496 AUE_MKDIRAT STD { int mkdirat( int fd, _In_z_ const char *path, mode_t mode ); } 497 AUE_MKFIFOAT STD { int mkfifoat( int fd, _In_z_ const char *path, mode_t mode ); } 498 AUE_MKNODAT COMPAT11 { int mknodat( int fd, _In_z_ const char *path, mode_t mode, uint32_t dev ); } ; XXX: see the comment for open 499 AUE_OPENAT_RWTC STD { int openat( int fd, _In_z_ const char *path, int flag, mode_t mode ); } 500 AUE_READLINKAT STD { int readlinkat( int fd, _In_z_ const char *path, _Out_writes_bytes_(bufsize) char *buf, size_t bufsize ); } 501 AUE_RENAMEAT STD { int renameat( int oldfd, _In_z_ const char *old, int newfd, _In_z_ const char *new ); } 502 AUE_SYMLINKAT STD { int symlinkat( _In_z_ const char *path1, int fd, _In_z_ const char *path2 ); } 503 AUE_UNLINKAT STD { int unlinkat( int fd, _In_z_ const char *path, int flag ); } 504 AUE_POSIX_OPENPT STD { int posix_openpt( int flags ); } ; 505 is initialised by the kgssapi code, if present. 505 AUE_NULL NOSTD { int gssd_syscall( _In_z_ const char *path ); } 506 AUE_JAIL_GET STD { int jail_get( _In_reads_(iovcnt) struct iovec *iovp, unsigned int iovcnt, int flags ); } 507 AUE_JAIL_SET STD { int jail_set( _In_reads_(iovcnt) struct iovec *iovp, unsigned int iovcnt, int flags ); } 508 AUE_JAIL_REMOVE STD { int jail_remove( int jid ); } 509 AUE_CLOSEFROM STD { int closefrom( int lowfd ); } 510 AUE_SEMCTL NOSTD { int __semctl( int semid, int semnum, int cmd, _Inout_ union semun *arg ); } 511 AUE_MSGCTL NOSTD { int msgctl( int msqid, int cmd, _Inout_opt_ struct msqid_ds *buf ); } 512 AUE_SHMCTL NOSTD { int shmctl( int shmid, int cmd, _Inout_opt_ struct shmid_ds *buf ); } 513 AUE_LPATHCONF STD { int lpathconf( _In_z_ const char *path, int name ); } 514 AUE_NULL OBSOL cap_new 515 AUE_CAP_RIGHTS_GET STD { int __cap_rights_get( int version, int fd, _Out_ cap_rights_t *rightsp ); } 516 AUE_CAP_ENTER STD { int cap_enter(void); } 517 AUE_CAP_GETMODE STD { int cap_getmode( _Out_ u_int *modep ); } 518 AUE_PDFORK STD { int pdfork( _Out_ int *fdp, int flags ); } 519 AUE_PDKILL STD { int pdkill( int fd, int signum ); } 520 AUE_PDGETPID STD { int pdgetpid( int fd, _Out_ pid_t *pidp ); } 521 AUE_PDWAIT UNIMPL pdwait4 522 AUE_SELECT STD { int pselect( int nd, _Inout_opt_ fd_set *in, _Inout_opt_ fd_set *ou, _Inout_opt_ fd_set *ex, _In_opt_ const struct timespec *ts, _In_opt_ const sigset_t *sm ); } 523 AUE_GETLOGINCLASS STD { int getloginclass( _Out_writes_z_(namelen) char *namebuf, size_t namelen ); } 524 AUE_SETLOGINCLASS STD { int setloginclass( _In_z_ const char *namebuf ); } 525 AUE_NULL STD { int rctl_get_racct( _In_reads_bytes_(inbuflen) const void *inbufp, size_t inbuflen, _Out_writes_bytes_(outbuflen) void *outbufp, size_t outbuflen ); } 526 AUE_NULL STD { int rctl_get_rules( _In_reads_bytes_(inbuflen) const void *inbufp, size_t inbuflen, _Out_writes_bytes_(outbuflen) void *outbufp, size_t outbuflen ); } 527 AUE_NULL STD { int rctl_get_limits( _In_reads_bytes_(inbuflen) const void *inbufp, size_t inbuflen, _Out_writes_bytes_(outbuflen) void *outbufp, size_t outbuflen ); } 528 AUE_NULL STD { int rctl_add_rule( _In_reads_bytes_(inbuflen) const void *inbufp, size_t inbuflen, _Out_writes_bytes_(outbuflen) void *outbufp, size_t outbuflen ); } 529 AUE_NULL STD { int rctl_remove_rule( _In_reads_bytes_(inbuflen) const void *inbufp, size_t inbuflen, _Out_writes_bytes_(outbuflen) void *outbufp, size_t outbuflen ); } 530 AUE_POSIX_FALLOCATE STD { int posix_fallocate( int fd, off_t offset, off_t len ); } 531 AUE_POSIX_FADVISE STD { int posix_fadvise( int fd, off_t offset, off_t len, int advice ); } 532 AUE_WAIT6 STD { int wait6( idtype_t idtype, id_t id, _Out_opt_ int *status, int options, _Out_opt_ struct __wrusage *wrusage, _Out_opt_ siginfo_t *info ); } 533 AUE_CAP_RIGHTS_LIMIT STD { int cap_rights_limit( int fd, _In_ cap_rights_t *rightsp ); } 534 AUE_CAP_IOCTLS_LIMIT STD { int cap_ioctls_limit( int fd, _In_reads_(ncmds) const u_long *cmds, size_t ncmds ); } 535 AUE_CAP_IOCTLS_GET STD { ssize_t cap_ioctls_get( int fd, _Out_writes_(maxcmds) u_long *cmds, size_t maxcmds ); } 536 AUE_CAP_FCNTLS_LIMIT STD { int cap_fcntls_limit( int fd, uint32_t fcntlrights ); } 537 AUE_CAP_FCNTLS_GET STD { int cap_fcntls_get( int fd, _Out_ uint32_t *fcntlrightsp ); } 538 AUE_BINDAT STD { int bindat( int fd, int s, _In_reads_bytes_(namelen) const struct sockaddr *name, int namelen ); } 539 AUE_CONNECTAT STD { int connectat( int fd, int s, _In_reads_bytes_(namelen) const struct sockaddr *name, int namelen ); } 540 AUE_CHFLAGSAT STD { int chflagsat( int fd, _In_z_ const char *path, u_long flags, int atflag ); } 541 AUE_ACCEPT STD { int accept4( int s, _Out_writes_bytes_opt_(*anamelen) struct sockaddr *name, _Inout_opt_ __socklen_t *anamelen, int flags ); } 542 AUE_PIPE STD { int pipe2( _Out_writes_(2) int *fildes, int flags ); } 543 AUE_AIO_MLOCK STD { int aio_mlock( _In_ struct aiocb *aiocbp ); } 544 AUE_PROCCTL STD { int procctl( idtype_t idtype, id_t id, int com, _In_opt_ void *data ); } 545 AUE_POLL STD { int ppoll( _Inout_updates_(nfds) struct pollfd *fds, u_int nfds, _In_opt_ const struct timespec *ts, _In_opt_ const sigset_t *set ); } 546 AUE_FUTIMES STD { int futimens( int fd, _In_reads_(2) struct timespec *times ); } 547 AUE_FUTIMESAT STD { int utimensat( int fd, _In_z_ const char *path, _In_reads_(2) struct timespec *times, int flag ); } 548 AUE_NULL OBSOL numa_getaffinity 549 AUE_NULL OBSOL numa_setaffinity 550 AUE_FSYNC STD { int fdatasync( int fd ); } 551 AUE_FSTAT STD { int fstat( int fd, _Out_ struct stat *sb ); } 552 AUE_FSTATAT STD { int fstatat( int fd, _In_z_ const char *path, _Out_ struct stat *buf, int flag ); } 553 AUE_FHSTAT STD { int fhstat( _In_ const struct fhandle *u_fhp, _Out_ struct stat *sb ); } 554 AUE_GETDIRENTRIES STD { ssize_t getdirentries( int fd, _Out_writes_bytes_(count) char *buf, size_t count, _Out_ off_t *basep ); } 555 AUE_STATFS STD { int statfs( _In_z_ const char *path, _Out_ struct statfs *buf ); } 556 AUE_FSTATFS STD { int fstatfs( int fd, _Out_ struct statfs *buf ); } 557 AUE_GETFSSTAT STD { int getfsstat( _Out_writes_bytes_opt_(bufsize) struct statfs *buf, long bufsize, int mode ); } 558 AUE_FHSTATFS STD { int fhstatfs( _In_ const struct fhandle *u_fhp, _Out_ struct statfs *buf ); } 559 AUE_MKNODAT STD { int mknodat( int fd, _In_z_ const char *path, mode_t mode, dev_t dev ); } 560 AUE_KEVENT STD { int kevent( int fd, _In_reads_opt_(nchanges) struct kevent *changelist, int nchanges, _Out_writes_opt_(nevents) struct kevent *eventlist, int nevents, _In_opt_ const struct timespec *timeout ); } 561 AUE_NULL STD { int cpuset_getdomain( cpulevel_t level, cpuwhich_t which, id_t id, size_t domainsetsize, _Out_writes_bytes_(domainsetsize) domainset_t *mask, _Out_ int *policy ); } 562 AUE_NULL STD { int cpuset_setdomain( cpulevel_t level, cpuwhich_t which, id_t id, size_t domainsetsize, _In_ domainset_t *mask, int policy ); } 563 AUE_NULL STD { int getrandom( _Out_writes_bytes_(buflen) void *buf, size_t buflen, unsigned int flags ); } 564 AUE_NULL STD { int getfhat( int fd, _In_z_ char *path, _Out_ struct fhandle *fhp, int flags ); } 565 AUE_NULL STD { int fhlink( _In_ struct fhandle *fhp, _In_z_ const char *to ); } 566 AUE_NULL STD { int fhlinkat( _In_ struct fhandle *fhp, int tofd, _In_z_ const char *to, ); } 567 AUE_NULL STD { int fhreadlink( _In_ struct fhandle *fhp, _Out_writes_(bufsize) char *buf, size_t bufsize ); } +568 AUE_UNLINKAT STD { + int funlinkat( + int dfd, + _In_z_ const char *path, + int fd, + int flag + ); + } ; Please copy any additions and changes to the following compatability tables: ; sys/compat/freebsd32/syscalls.master ; vim: syntax=off Index: head/sys/kern/vfs_mountroot.c =================================================================== --- head/sys/kern/vfs_mountroot.c (revision 345981) +++ head/sys/kern/vfs_mountroot.c (revision 345982) @@ -1,1132 +1,1132 @@ /*- * SPDX-License-Identifier: BSD-3-Clause * * Copyright (c) 2010 Marcel Moolenaar * 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. * 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 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 "opt_rootdevname.h" #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 /* * The root filesystem is detailed in the kernel environment variable * vfs.root.mountfrom, which is expected to be in the general format * * :[][ :[] ...] * vfsname := the name of a VFS known to the kernel and capable * of being mounted as root * path := disk device name or other data used by the filesystem * to locate its physical store * * If the environment variable vfs.root.mountfrom is a space separated list, * each list element is tried in turn and the root filesystem will be mounted * from the first one that succeeds. * * The environment variable vfs.root.mountfrom.options is a comma delimited * set of string mount options. These mount options must be parseable * by nmount() in the kernel. */ static int parse_mount(char **); static struct mntarg *parse_mountroot_options(struct mntarg *, const char *); static int sysctl_vfs_root_mount_hold(SYSCTL_HANDLER_ARGS); static void vfs_mountroot_wait(void); static int vfs_mountroot_wait_if_neccessary(const char *fs, const char *dev); /* * The vnode of the system's root (/ in the filesystem, without chroot * active.) */ struct vnode *rootvnode; /* * Mount of the system's /dev. */ struct mount *rootdevmp; char *rootdevnames[2] = {NULL, NULL}; struct mtx root_holds_mtx; MTX_SYSINIT(root_holds, &root_holds_mtx, "root_holds", MTX_DEF); struct root_hold_token { const char *who; LIST_ENTRY(root_hold_token) list; }; static LIST_HEAD(, root_hold_token) root_holds = LIST_HEAD_INITIALIZER(root_holds); enum action { A_CONTINUE, A_PANIC, A_REBOOT, A_RETRY }; static enum action root_mount_onfail = A_CONTINUE; static int root_mount_mddev; static int root_mount_complete; /* By default wait up to 3 seconds for devices to appear. */ static int root_mount_timeout = 3; TUNABLE_INT("vfs.mountroot.timeout", &root_mount_timeout); static int root_mount_always_wait = 0; SYSCTL_INT(_vfs, OID_AUTO, root_mount_always_wait, CTLFLAG_RDTUN, &root_mount_always_wait, 0, "Wait for root mount holds even if the root device already exists"); SYSCTL_PROC(_vfs, OID_AUTO, root_mount_hold, CTLTYPE_STRING | CTLFLAG_RD | CTLFLAG_MPSAFE, NULL, 0, sysctl_vfs_root_mount_hold, "A", "List of root mount hold tokens"); static int sysctl_vfs_root_mount_hold(SYSCTL_HANDLER_ARGS) { struct sbuf sb; struct root_hold_token *h; int error; sbuf_new(&sb, NULL, 256, SBUF_AUTOEXTEND | SBUF_INCLUDENUL); mtx_lock(&root_holds_mtx); LIST_FOREACH(h, &root_holds, list) { if (h != LIST_FIRST(&root_holds)) sbuf_putc(&sb, ' '); sbuf_printf(&sb, "%s", h->who); } mtx_unlock(&root_holds_mtx); error = sbuf_finish(&sb); if (error == 0) error = SYSCTL_OUT(req, sbuf_data(&sb), sbuf_len(&sb)); sbuf_delete(&sb); return (error); } struct root_hold_token * root_mount_hold(const char *identifier) { struct root_hold_token *h; h = malloc(sizeof *h, M_DEVBUF, M_ZERO | M_WAITOK); h->who = identifier; mtx_lock(&root_holds_mtx); TSHOLD("root mount"); LIST_INSERT_HEAD(&root_holds, h, list); mtx_unlock(&root_holds_mtx); return (h); } void root_mount_rel(struct root_hold_token *h) { if (h == NULL) return; mtx_lock(&root_holds_mtx); LIST_REMOVE(h, list); TSRELEASE("root mount"); wakeup(&root_holds); mtx_unlock(&root_holds_mtx); free(h, M_DEVBUF); } int root_mounted(void) { /* No mutex is acquired here because int stores are atomic. */ return (root_mount_complete); } static void set_rootvnode(void) { struct proc *p; if (VFS_ROOT(TAILQ_FIRST(&mountlist), LK_EXCLUSIVE, &rootvnode)) panic("set_rootvnode: Cannot find root vnode"); VOP_UNLOCK(rootvnode, 0); p = curthread->td_proc; FILEDESC_XLOCK(p->p_fd); if (p->p_fd->fd_cdir != NULL) vrele(p->p_fd->fd_cdir); p->p_fd->fd_cdir = rootvnode; VREF(rootvnode); if (p->p_fd->fd_rdir != NULL) vrele(p->p_fd->fd_rdir); p->p_fd->fd_rdir = rootvnode; VREF(rootvnode); FILEDESC_XUNLOCK(p->p_fd); } static int vfs_mountroot_devfs(struct thread *td, struct mount **mpp) { struct vfsoptlist *opts; struct vfsconf *vfsp; struct mount *mp; int error; *mpp = NULL; if (rootdevmp != NULL) { /* * Already have /dev; this happens during rerooting. */ error = vfs_busy(rootdevmp, 0); if (error != 0) return (error); *mpp = rootdevmp; } else { vfsp = vfs_byname("devfs"); KASSERT(vfsp != NULL, ("Could not find devfs by name")); if (vfsp == NULL) return (ENOENT); mp = vfs_mount_alloc(NULLVP, vfsp, "/dev", td->td_ucred); error = VFS_MOUNT(mp); KASSERT(error == 0, ("VFS_MOUNT(devfs) failed %d", error)); if (error) return (error); opts = malloc(sizeof(struct vfsoptlist), M_MOUNT, M_WAITOK); TAILQ_INIT(opts); mp->mnt_opt = opts; mtx_lock(&mountlist_mtx); TAILQ_INSERT_HEAD(&mountlist, mp, mnt_list); mtx_unlock(&mountlist_mtx); *mpp = mp; rootdevmp = mp; } set_rootvnode(); error = kern_symlinkat(td, "/", AT_FDCWD, "dev", UIO_SYSSPACE); if (error) printf("kern_symlink /dev -> / returns %d\n", error); return (error); } static void vfs_mountroot_shuffle(struct thread *td, struct mount *mpdevfs) { struct nameidata nd; struct mount *mporoot, *mpnroot; struct vnode *vp, *vporoot, *vpdevfs; char *fspath; int error; mpnroot = TAILQ_NEXT(mpdevfs, mnt_list); /* Shuffle the mountlist. */ mtx_lock(&mountlist_mtx); mporoot = TAILQ_FIRST(&mountlist); TAILQ_REMOVE(&mountlist, mpdevfs, mnt_list); if (mporoot != mpdevfs) { TAILQ_REMOVE(&mountlist, mpnroot, mnt_list); TAILQ_INSERT_HEAD(&mountlist, mpnroot, mnt_list); } TAILQ_INSERT_TAIL(&mountlist, mpdevfs, mnt_list); mtx_unlock(&mountlist_mtx); cache_purgevfs(mporoot, true); if (mporoot != mpdevfs) cache_purgevfs(mpdevfs, true); if (VFS_ROOT(mporoot, LK_EXCLUSIVE, &vporoot)) panic("vfs_mountroot_shuffle: Cannot find root vnode"); VI_LOCK(vporoot); vporoot->v_iflag &= ~VI_MOUNT; VI_UNLOCK(vporoot); vporoot->v_mountedhere = NULL; mporoot->mnt_flag &= ~MNT_ROOTFS; mporoot->mnt_vnodecovered = NULL; vput(vporoot); /* Set up the new rootvnode, and purge the cache */ mpnroot->mnt_vnodecovered = NULL; set_rootvnode(); cache_purgevfs(rootvnode->v_mount, true); if (mporoot != mpdevfs) { /* Remount old root under /.mount or /mnt */ fspath = "/.mount"; NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, fspath, td); error = namei(&nd); if (error) { NDFREE(&nd, NDF_ONLY_PNBUF); fspath = "/mnt"; NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, fspath, td); error = namei(&nd); } if (!error) { vp = nd.ni_vp; error = (vp->v_type == VDIR) ? 0 : ENOTDIR; if (!error) error = vinvalbuf(vp, V_SAVE, 0, 0); if (!error) { cache_purge(vp); mporoot->mnt_vnodecovered = vp; vp->v_mountedhere = mporoot; strlcpy(mporoot->mnt_stat.f_mntonname, fspath, MNAMELEN); VOP_UNLOCK(vp, 0); } else vput(vp); } NDFREE(&nd, NDF_ONLY_PNBUF); if (error) printf("mountroot: unable to remount previous root " "under /.mount or /mnt (error %d)\n", error); } /* Remount devfs under /dev */ NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, "/dev", td); error = namei(&nd); if (!error) { vp = nd.ni_vp; error = (vp->v_type == VDIR) ? 0 : ENOTDIR; if (!error) error = vinvalbuf(vp, V_SAVE, 0, 0); if (!error) { vpdevfs = mpdevfs->mnt_vnodecovered; if (vpdevfs != NULL) { cache_purge(vpdevfs); vpdevfs->v_mountedhere = NULL; vrele(vpdevfs); } mpdevfs->mnt_vnodecovered = vp; vp->v_mountedhere = mpdevfs; VOP_UNLOCK(vp, 0); } else vput(vp); } if (error) printf("mountroot: unable to remount devfs under /dev " "(error %d)\n", error); NDFREE(&nd, NDF_ONLY_PNBUF); if (mporoot == mpdevfs) { vfs_unbusy(mpdevfs); /* Unlink the no longer needed /dev/dev -> / symlink */ - error = kern_unlinkat(td, AT_FDCWD, "/dev/dev", + error = kern_funlinkat(td, AT_FDCWD, "/dev/dev", FD_NONE, UIO_SYSSPACE, 0, 0); if (error) printf("mountroot: unable to unlink /dev/dev " "(error %d)\n", error); } } /* * Configuration parser. */ /* Parser character classes. */ #define CC_WHITESPACE -1 #define CC_NONWHITESPACE -2 /* Parse errors. */ #define PE_EOF -1 #define PE_EOL -2 static __inline int parse_peek(char **conf) { return (**conf); } static __inline void parse_poke(char **conf, int c) { **conf = c; } static __inline void parse_advance(char **conf) { (*conf)++; } static int parse_skipto(char **conf, int mc) { int c, match; while (1) { c = parse_peek(conf); if (c == 0) return (PE_EOF); switch (mc) { case CC_WHITESPACE: match = (c == ' ' || c == '\t' || c == '\n') ? 1 : 0; break; case CC_NONWHITESPACE: if (c == '\n') return (PE_EOL); match = (c != ' ' && c != '\t') ? 1 : 0; break; default: match = (c == mc) ? 1 : 0; break; } if (match) break; parse_advance(conf); } return (0); } static int parse_token(char **conf, char **tok) { char *p; size_t len; int error; *tok = NULL; error = parse_skipto(conf, CC_NONWHITESPACE); if (error) return (error); p = *conf; error = parse_skipto(conf, CC_WHITESPACE); len = *conf - p; *tok = malloc(len + 1, M_TEMP, M_WAITOK | M_ZERO); bcopy(p, *tok, len); return (0); } static void parse_dir_ask_printenv(const char *var) { char *val; val = kern_getenv(var); if (val != NULL) { printf(" %s=%s\n", var, val); freeenv(val); } } static int parse_dir_ask(char **conf) { char name[80]; char *mnt; int error; vfs_mountroot_wait(); printf("\nLoader variables:\n"); parse_dir_ask_printenv("vfs.root.mountfrom"); parse_dir_ask_printenv("vfs.root.mountfrom.options"); printf("\nManual root filesystem specification:\n"); printf(" : [options]\n"); printf(" Mount using filesystem \n"); printf(" and with the specified (optional) option list.\n"); printf("\n"); printf(" eg. ufs:/dev/da0s1a\n"); printf(" zfs:zroot/ROOT/default\n"); printf(" cd9660:/dev/cd0 ro\n"); printf(" (which is equivalent to: "); printf("mount -t cd9660 -o ro /dev/cd0 /)\n"); printf("\n"); printf(" ? List valid disk boot devices\n"); printf(" . Yield 1 second (for background tasks)\n"); printf(" Abort manual input\n"); do { error = EINVAL; printf("\nmountroot> "); cngets(name, sizeof(name), GETS_ECHO); if (name[0] == '\0') break; if (name[0] == '?' && name[1] == '\0') { printf("\nList of GEOM managed disk devices:\n "); g_dev_print(); continue; } if (name[0] == '.' && name[1] == '\0') { pause("rmask", hz); continue; } mnt = name; error = parse_mount(&mnt); if (error == -1) printf("Invalid file system specification.\n"); } while (error != 0); return (error); } static int parse_dir_md(char **conf) { struct stat sb; struct thread *td; struct md_ioctl *mdio; char *path, *tok; int error, fd, len; td = curthread; error = parse_token(conf, &tok); if (error) return (error); len = strlen(tok); mdio = malloc(sizeof(*mdio) + len + 1, M_TEMP, M_WAITOK | M_ZERO); path = (void *)(mdio + 1); bcopy(tok, path, len); free(tok, M_TEMP); /* Get file status. */ error = kern_statat(td, 0, AT_FDCWD, path, UIO_SYSSPACE, &sb, NULL); if (error) goto out; /* Open /dev/mdctl so that we can attach/detach. */ error = kern_openat(td, AT_FDCWD, "/dev/" MDCTL_NAME, UIO_SYSSPACE, O_RDWR, 0); if (error) goto out; fd = td->td_retval[0]; mdio->md_version = MDIOVERSION; mdio->md_type = MD_VNODE; if (root_mount_mddev != -1) { mdio->md_unit = root_mount_mddev; (void)kern_ioctl(td, fd, MDIOCDETACH, (void *)mdio); /* Ignore errors. We don't care. */ root_mount_mddev = -1; } mdio->md_file = (void *)(mdio + 1); mdio->md_options = MD_AUTOUNIT | MD_READONLY; mdio->md_mediasize = sb.st_size; mdio->md_unit = 0; error = kern_ioctl(td, fd, MDIOCATTACH, (void *)mdio); if (error) goto out; if (mdio->md_unit > 9) { printf("rootmount: too many md units\n"); mdio->md_file = NULL; mdio->md_options = 0; mdio->md_mediasize = 0; error = kern_ioctl(td, fd, MDIOCDETACH, (void *)mdio); /* Ignore errors. We don't care. */ error = ERANGE; goto out; } root_mount_mddev = mdio->md_unit; printf(MD_NAME "%u attached to %s\n", root_mount_mddev, mdio->md_file); error = kern_close(td, fd); out: free(mdio, M_TEMP); return (error); } static int parse_dir_onfail(char **conf) { char *action; int error; error = parse_token(conf, &action); if (error) return (error); if (!strcmp(action, "continue")) root_mount_onfail = A_CONTINUE; else if (!strcmp(action, "panic")) root_mount_onfail = A_PANIC; else if (!strcmp(action, "reboot")) root_mount_onfail = A_REBOOT; else if (!strcmp(action, "retry")) root_mount_onfail = A_RETRY; else { printf("rootmount: %s: unknown action\n", action); error = EINVAL; } free(action, M_TEMP); return (0); } static int parse_dir_timeout(char **conf) { char *tok, *endtok; long secs; int error; error = parse_token(conf, &tok); if (error) return (error); secs = strtol(tok, &endtok, 0); error = (secs < 0 || *endtok != '\0') ? EINVAL : 0; if (!error) root_mount_timeout = secs; free(tok, M_TEMP); return (error); } static int parse_directive(char **conf) { char *dir; int error; error = parse_token(conf, &dir); if (error) return (error); if (strcmp(dir, ".ask") == 0) error = parse_dir_ask(conf); else if (strcmp(dir, ".md") == 0) error = parse_dir_md(conf); else if (strcmp(dir, ".onfail") == 0) error = parse_dir_onfail(conf); else if (strcmp(dir, ".timeout") == 0) error = parse_dir_timeout(conf); else { printf("mountroot: invalid directive `%s'\n", dir); /* Ignore the rest of the line. */ (void)parse_skipto(conf, '\n'); error = EINVAL; } free(dir, M_TEMP); return (error); } static int parse_mount_dev_present(const char *dev) { struct nameidata nd; int error; NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF, UIO_SYSSPACE, dev, curthread); error = namei(&nd); if (!error) vput(nd.ni_vp); NDFREE(&nd, NDF_ONLY_PNBUF); return (error != 0) ? 0 : 1; } #define ERRMSGL 255 static int parse_mount(char **conf) { char *errmsg; struct mntarg *ma; char *dev, *fs, *opts, *tok; int delay, error, timeout; error = parse_token(conf, &tok); if (error) return (error); fs = tok; error = parse_skipto(&tok, ':'); if (error) { free(fs, M_TEMP); return (error); } parse_poke(&tok, '\0'); parse_advance(&tok); dev = tok; if (root_mount_mddev != -1) { /* Handle substitution for the md unit number. */ tok = strstr(dev, "md#"); if (tok != NULL) tok[2] = '0' + root_mount_mddev; } /* Parse options. */ error = parse_token(conf, &tok); opts = (error == 0) ? tok : NULL; printf("Trying to mount root from %s:%s [%s]...\n", fs, dev, (opts != NULL) ? opts : ""); errmsg = malloc(ERRMSGL, M_TEMP, M_WAITOK | M_ZERO); if (vfs_byname(fs) == NULL) { strlcpy(errmsg, "unknown file system", ERRMSGL); error = ENOENT; goto out; } error = vfs_mountroot_wait_if_neccessary(fs, dev); if (error != 0) goto out; delay = hz / 10; timeout = root_mount_timeout * hz; for (;;) { ma = NULL; ma = mount_arg(ma, "fstype", fs, -1); ma = mount_arg(ma, "fspath", "/", -1); ma = mount_arg(ma, "from", dev, -1); ma = mount_arg(ma, "errmsg", errmsg, ERRMSGL); ma = mount_arg(ma, "ro", NULL, 0); ma = parse_mountroot_options(ma, opts); error = kernel_mount(ma, MNT_ROOTFS); if (error == 0 || timeout <= 0) break; if (root_mount_timeout * hz == timeout || (bootverbose && timeout % hz == 0)) { printf("Mounting from %s:%s failed with error %d; " "retrying for %d more second%s\n", fs, dev, error, timeout / hz, (timeout / hz > 1) ? "s" : ""); } pause("rmretry", delay); timeout -= delay; } out: if (error) { printf("Mounting from %s:%s failed with error %d", fs, dev, error); if (errmsg[0] != '\0') printf(": %s", errmsg); printf(".\n"); } free(fs, M_TEMP); free(errmsg, M_TEMP); if (opts != NULL) free(opts, M_TEMP); /* kernel_mount can return -1 on error. */ return ((error < 0) ? EDOOFUS : error); } #undef ERRMSGL static int vfs_mountroot_parse(struct sbuf *sb, struct mount *mpdevfs) { struct mount *mp; char *conf; int error; root_mount_mddev = -1; retry: conf = sbuf_data(sb); mp = TAILQ_NEXT(mpdevfs, mnt_list); error = (mp == NULL) ? 0 : EDOOFUS; root_mount_onfail = A_CONTINUE; while (mp == NULL) { error = parse_skipto(&conf, CC_NONWHITESPACE); if (error == PE_EOL) { parse_advance(&conf); continue; } if (error < 0) break; switch (parse_peek(&conf)) { case '#': error = parse_skipto(&conf, '\n'); break; case '.': error = parse_directive(&conf); break; default: error = parse_mount(&conf); if (error == -1) { printf("mountroot: invalid file system " "specification.\n"); error = 0; } break; } if (error < 0) break; /* Ignore any trailing garbage on the line. */ if (parse_peek(&conf) != '\n') { printf("mountroot: advancing to next directive...\n"); (void)parse_skipto(&conf, '\n'); } mp = TAILQ_NEXT(mpdevfs, mnt_list); } if (mp != NULL) return (0); /* * We failed to mount (a new) root. */ switch (root_mount_onfail) { case A_CONTINUE: break; case A_PANIC: panic("mountroot: unable to (re-)mount root."); /* NOTREACHED */ case A_RETRY: goto retry; case A_REBOOT: kern_reboot(RB_NOSYNC); /* NOTREACHED */ } return (error); } static void vfs_mountroot_conf0(struct sbuf *sb) { char *s, *tok, *mnt, *opt; int error; sbuf_printf(sb, ".onfail panic\n"); sbuf_printf(sb, ".timeout %d\n", root_mount_timeout); if (boothowto & RB_ASKNAME) sbuf_printf(sb, ".ask\n"); #ifdef ROOTDEVNAME if (boothowto & RB_DFLTROOT) sbuf_printf(sb, "%s\n", ROOTDEVNAME); #endif if (boothowto & RB_CDROM) { sbuf_printf(sb, "cd9660:/dev/cd0 ro\n"); sbuf_printf(sb, ".timeout 0\n"); sbuf_printf(sb, "cd9660:/dev/cd1 ro\n"); sbuf_printf(sb, ".timeout %d\n", root_mount_timeout); } s = kern_getenv("vfs.root.mountfrom"); if (s != NULL) { opt = kern_getenv("vfs.root.mountfrom.options"); tok = s; error = parse_token(&tok, &mnt); while (!error) { sbuf_printf(sb, "%s %s\n", mnt, (opt != NULL) ? opt : ""); free(mnt, M_TEMP); error = parse_token(&tok, &mnt); } if (opt != NULL) freeenv(opt); freeenv(s); } if (rootdevnames[0] != NULL) sbuf_printf(sb, "%s\n", rootdevnames[0]); if (rootdevnames[1] != NULL) sbuf_printf(sb, "%s\n", rootdevnames[1]); #ifdef ROOTDEVNAME if (!(boothowto & RB_DFLTROOT)) sbuf_printf(sb, "%s\n", ROOTDEVNAME); #endif if (!(boothowto & RB_ASKNAME)) sbuf_printf(sb, ".ask\n"); } static int vfs_mountroot_readconf(struct thread *td, struct sbuf *sb) { static char buf[128]; struct nameidata nd; off_t ofs; ssize_t resid; int error, flags, len; NDINIT(&nd, LOOKUP, FOLLOW, UIO_SYSSPACE, "/.mount.conf", td); flags = FREAD; error = vn_open(&nd, &flags, 0, NULL); if (error) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); ofs = 0; len = sizeof(buf) - 1; while (1) { error = vn_rdwr(UIO_READ, nd.ni_vp, buf, len, ofs, UIO_SYSSPACE, IO_NODELOCKED, td->td_ucred, NOCRED, &resid, td); if (error) break; if (resid == len) break; buf[len - resid] = 0; sbuf_printf(sb, "%s", buf); ofs += len - resid; } VOP_UNLOCK(nd.ni_vp, 0); vn_close(nd.ni_vp, FREAD, td->td_ucred, td); return (error); } static void vfs_mountroot_wait(void) { struct root_hold_token *h; struct timeval lastfail; int curfail; TSENTER(); curfail = 0; while (1) { g_waitidle(); mtx_lock(&root_holds_mtx); if (LIST_EMPTY(&root_holds)) { mtx_unlock(&root_holds_mtx); break; } if (ppsratecheck(&lastfail, &curfail, 1)) { printf("Root mount waiting for:"); LIST_FOREACH(h, &root_holds, list) printf(" %s", h->who); printf("\n"); } TSWAIT("root mount"); msleep(&root_holds, &root_holds_mtx, PZERO | PDROP, "roothold", hz); TSUNWAIT("root mount"); } TSEXIT(); } static int vfs_mountroot_wait_if_neccessary(const char *fs, const char *dev) { int delay, timeout; /* * In case of ZFS and NFS we don't have a way to wait for * specific device. Also do the wait if the user forced that * behaviour by setting vfs.root_mount_always_wait=1. */ if (strcmp(fs, "zfs") == 0 || strstr(fs, "nfs") != NULL || dev[0] == '\0' || root_mount_always_wait != 0) { vfs_mountroot_wait(); return (0); } /* * Otherwise, no point in waiting if the device is already there. * Note that we must wait for GEOM to finish reconfiguring itself, * eg for geom_part(4) to finish tasting. */ g_waitidle(); if (parse_mount_dev_present(dev)) return (0); /* * No luck. Let's wait. This code looks weird, but it's that way * to behave exactly as it used to work before. */ vfs_mountroot_wait(); printf("mountroot: waiting for device %s...\n", dev); delay = hz / 10; timeout = root_mount_timeout * hz; do { pause("rmdev", delay); timeout -= delay; } while (timeout > 0 && !parse_mount_dev_present(dev)); if (timeout <= 0) return (ENODEV); return (0); } void vfs_mountroot(void) { struct mount *mp; struct sbuf *sb; struct thread *td; time_t timebase; int error; mtx_assert(&Giant, MA_NOTOWNED); TSENTER(); td = curthread; sb = sbuf_new_auto(); vfs_mountroot_conf0(sb); sbuf_finish(sb); error = vfs_mountroot_devfs(td, &mp); while (!error) { error = vfs_mountroot_parse(sb, mp); if (!error) { vfs_mountroot_shuffle(td, mp); sbuf_clear(sb); error = vfs_mountroot_readconf(td, sb); sbuf_finish(sb); } } sbuf_delete(sb); /* * Iterate over all currently mounted file systems and use * the time stamp found to check and/or initialize the RTC. * Call inittodr() only once and pass it the largest of the * timestamps we encounter. */ timebase = 0; mtx_lock(&mountlist_mtx); mp = TAILQ_FIRST(&mountlist); while (mp != NULL) { if (mp->mnt_time > timebase) timebase = mp->mnt_time; mp = TAILQ_NEXT(mp, mnt_list); } mtx_unlock(&mountlist_mtx); inittodr(timebase); /* Keep prison0's root in sync with the global rootvnode. */ mtx_lock(&prison0.pr_mtx); prison0.pr_root = rootvnode; vref(prison0.pr_root); mtx_unlock(&prison0.pr_mtx); mtx_lock(&root_holds_mtx); atomic_store_rel_int(&root_mount_complete, 1); wakeup(&root_mount_complete); mtx_unlock(&root_holds_mtx); EVENTHANDLER_INVOKE(mountroot); TSEXIT(); } static struct mntarg * parse_mountroot_options(struct mntarg *ma, const char *options) { char *p; char *name, *name_arg; char *val, *val_arg; char *opts; if (options == NULL || options[0] == '\0') return (ma); p = opts = strdup(options, M_MOUNT); if (opts == NULL) { return (ma); } while((name = strsep(&p, ",")) != NULL) { if (name[0] == '\0') break; val = strchr(name, '='); if (val != NULL) { *val = '\0'; ++val; } if( strcmp(name, "rw") == 0 || strcmp(name, "noro") == 0) { /* * The first time we mount the root file system, * we need to mount 'ro', so We need to ignore * 'rw' and 'noro' mount options. */ continue; } name_arg = strdup(name, M_MOUNT); val_arg = NULL; if (val != NULL) val_arg = strdup(val, M_MOUNT); ma = mount_arg(ma, name_arg, val_arg, (val_arg != NULL ? -1 : 0)); } free(opts, M_MOUNT); return (ma); } Index: head/sys/kern/vfs_syscalls.c =================================================================== --- head/sys/kern/vfs_syscalls.c (revision 345981) +++ head/sys/kern/vfs_syscalls.c (revision 345982) @@ -1,4753 +1,4818 @@ /*- * SPDX-License-Identifier: BSD-3-Clause * * 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. * 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. * * @(#)vfs_syscalls.c 8.13 (Berkeley) 4/15/94 */ #include __FBSDID("$FreeBSD$"); #include "opt_capsicum.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 #include #include #include #include #ifdef KTRACE #include #endif #include #include #include #include #include #include #include #include MALLOC_DEFINE(M_FADVISE, "fadvise", "posix_fadvise(2) information"); SDT_PROVIDER_DEFINE(vfs); SDT_PROBE_DEFINE2(vfs, , stat, mode, "char *", "int"); SDT_PROBE_DEFINE2(vfs, , stat, reg, "char *", "int"); static int kern_chflagsat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, u_long flags, int atflag); static int setfflags(struct thread *td, struct vnode *, u_long); static int getutimes(const struct timeval *, enum uio_seg, struct timespec *); static int getutimens(const struct timespec *, enum uio_seg, struct timespec *, int *); static int setutimes(struct thread *td, struct vnode *, const struct timespec *, int, int); static int vn_access(struct vnode *vp, int user_flags, struct ucred *cred, struct thread *td); static int kern_fhlinkat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, fhandle_t *fhp); static int kern_getfhat(struct thread *td, int flags, int fd, const char *path, enum uio_seg pathseg, fhandle_t *fhp); static int kern_readlink_vp(struct vnode *vp, char *buf, enum uio_seg bufseg, size_t count, struct thread *td); static int kern_linkat_vp(struct thread *td, struct vnode *vp, int fd, const char *path, enum uio_seg segflag); /* * Sync each mounted filesystem. */ #ifndef _SYS_SYSPROTO_H_ struct sync_args { int dummy; }; #endif /* ARGSUSED */ int sys_sync(struct thread *td, struct sync_args *uap) { struct mount *mp, *nmp; int save; 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; } if ((mp->mnt_flag & MNT_RDONLY) == 0 && vn_start_write(NULL, &mp, V_NOWAIT) == 0) { save = curthread_pflags_set(TDP_SYNCIO); vfs_msync(mp, MNT_NOWAIT); VFS_SYNC(mp, MNT_NOWAIT); curthread_pflags_restore(save); vn_finished_write(mp); } mtx_lock(&mountlist_mtx); nmp = TAILQ_NEXT(mp, mnt_list); vfs_unbusy(mp); } mtx_unlock(&mountlist_mtx); return (0); } /* * Change filesystem quotas. */ #ifndef _SYS_SYSPROTO_H_ struct quotactl_args { char *path; int cmd; int uid; caddr_t arg; }; #endif int sys_quotactl(struct thread *td, struct quotactl_args *uap) { struct mount *mp; struct nameidata nd; int error; AUDIT_ARG_CMD(uap->cmd); AUDIT_ARG_UID(uap->uid); if (!prison_allow(td->td_ucred, PR_ALLOW_QUOTAS)) return (EPERM); NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | AUDITVNODE1, UIO_USERSPACE, uap->path, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); mp = nd.ni_vp->v_mount; vfs_ref(mp); vput(nd.ni_vp); error = vfs_busy(mp, 0); vfs_rel(mp); if (error != 0) return (error); error = VFS_QUOTACTL(mp, uap->cmd, uap->uid, uap->arg); /* * Since quota on operation typically needs to open quota * file, the Q_QUOTAON handler needs to unbusy the mount point * before calling into namei. Otherwise, unmount might be * started between two vfs_busy() invocations (first is our, * second is from mount point cross-walk code in lookup()), * causing deadlock. * * Require that Q_QUOTAON handles the vfs_busy() reference on * its own, always returning with ubusied mount point. */ if ((uap->cmd >> SUBCMDSHIFT) != Q_QUOTAON && (uap->cmd >> SUBCMDSHIFT) != Q_QUOTAOFF) vfs_unbusy(mp); return (error); } /* * Used by statfs conversion routines to scale the block size up if * necessary so that all of the block counts are <= 'max_size'. Note * that 'max_size' should be a bitmask, i.e. 2^n - 1 for some non-zero * value of 'n'. */ void statfs_scale_blocks(struct statfs *sf, long max_size) { uint64_t count; int shift; KASSERT(powerof2(max_size + 1), ("%s: invalid max_size", __func__)); /* * Attempt to scale the block counts to give a more accurate * overview to userland of the ratio of free space to used * space. To do this, find the largest block count and compute * a divisor that lets it fit into a signed integer <= max_size. */ if (sf->f_bavail < 0) count = -sf->f_bavail; else count = sf->f_bavail; count = MAX(sf->f_blocks, MAX(sf->f_bfree, count)); if (count <= max_size) return; count >>= flsl(max_size); shift = 0; while (count > 0) { shift++; count >>=1; } sf->f_bsize <<= shift; sf->f_blocks >>= shift; sf->f_bfree >>= shift; sf->f_bavail >>= shift; } static int kern_do_statfs(struct thread *td, struct mount *mp, struct statfs *buf) { struct statfs *sp; int error; if (mp == NULL) return (EBADF); error = vfs_busy(mp, 0); vfs_rel(mp); if (error != 0) return (error); #ifdef MAC error = mac_mount_check_stat(td->td_ucred, mp); if (error != 0) goto out; #endif /* * Set these in case the underlying filesystem fails to do so. */ sp = &mp->mnt_stat; sp->f_version = STATFS_VERSION; sp->f_namemax = NAME_MAX; sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK; error = VFS_STATFS(mp, sp); if (error != 0) goto out; *buf = *sp; if (priv_check(td, PRIV_VFS_GENERATION)) { buf->f_fsid.val[0] = buf->f_fsid.val[1] = 0; prison_enforce_statfs(td->td_ucred, mp, buf); } out: vfs_unbusy(mp); return (error); } /* * Get filesystem statistics. */ #ifndef _SYS_SYSPROTO_H_ struct statfs_args { char *path; struct statfs *buf; }; #endif int sys_statfs(struct thread *td, struct statfs_args *uap) { struct statfs *sfp; int error; sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_statfs(td, uap->path, UIO_USERSPACE, sfp); if (error == 0) error = copyout(sfp, uap->buf, sizeof(struct statfs)); free(sfp, M_STATFS); return (error); } int kern_statfs(struct thread *td, const char *path, enum uio_seg pathseg, struct statfs *buf) { struct mount *mp; struct nameidata nd; int error; NDINIT(&nd, LOOKUP, FOLLOW | LOCKSHARED | LOCKLEAF | AUDITVNODE1, pathseg, path, td); error = namei(&nd); if (error != 0) return (error); mp = nd.ni_vp->v_mount; vfs_ref(mp); NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_vp); return (kern_do_statfs(td, mp, buf)); } /* * Get filesystem statistics. */ #ifndef _SYS_SYSPROTO_H_ struct fstatfs_args { int fd; struct statfs *buf; }; #endif int sys_fstatfs(struct thread *td, struct fstatfs_args *uap) { struct statfs *sfp; int error; sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_fstatfs(td, uap->fd, sfp); if (error == 0) error = copyout(sfp, uap->buf, sizeof(struct statfs)); free(sfp, M_STATFS); return (error); } int kern_fstatfs(struct thread *td, int fd, struct statfs *buf) { struct file *fp; struct mount *mp; struct vnode *vp; int error; AUDIT_ARG_FD(fd); error = getvnode(td, fd, &cap_fstatfs_rights, &fp); if (error != 0) return (error); vp = fp->f_vnode; vn_lock(vp, LK_SHARED | LK_RETRY); #ifdef AUDIT AUDIT_ARG_VNODE1(vp); #endif mp = vp->v_mount; if (mp != NULL) vfs_ref(mp); VOP_UNLOCK(vp, 0); fdrop(fp, td); return (kern_do_statfs(td, mp, buf)); } /* * Get statistics on all filesystems. */ #ifndef _SYS_SYSPROTO_H_ struct getfsstat_args { struct statfs *buf; long bufsize; int mode; }; #endif int sys_getfsstat(struct thread *td, struct getfsstat_args *uap) { size_t count; int error; if (uap->bufsize < 0 || uap->bufsize > SIZE_MAX) return (EINVAL); error = kern_getfsstat(td, &uap->buf, uap->bufsize, &count, UIO_USERSPACE, uap->mode); if (error == 0) td->td_retval[0] = count; return (error); } /* * If (bufsize > 0 && bufseg == UIO_SYSSPACE) * The caller is responsible for freeing memory which will be allocated * in '*buf'. */ int kern_getfsstat(struct thread *td, struct statfs **buf, size_t bufsize, size_t *countp, enum uio_seg bufseg, int mode) { struct mount *mp, *nmp; struct statfs *sfsp, *sp, *sptmp, *tofree; size_t count, maxcount; int error; switch (mode) { case MNT_WAIT: case MNT_NOWAIT: break; default: if (bufseg == UIO_SYSSPACE) *buf = NULL; return (EINVAL); } restart: maxcount = bufsize / sizeof(struct statfs); if (bufsize == 0) { sfsp = NULL; tofree = NULL; } else if (bufseg == UIO_USERSPACE) { sfsp = *buf; tofree = NULL; } else /* if (bufseg == UIO_SYSSPACE) */ { count = 0; mtx_lock(&mountlist_mtx); TAILQ_FOREACH(mp, &mountlist, mnt_list) { count++; } mtx_unlock(&mountlist_mtx); if (maxcount > count) maxcount = count; tofree = sfsp = *buf = malloc(maxcount * sizeof(struct statfs), M_STATFS, M_WAITOK); } count = 0; mtx_lock(&mountlist_mtx); for (mp = TAILQ_FIRST(&mountlist); mp != NULL; mp = nmp) { if (prison_canseemount(td->td_ucred, mp) != 0) { nmp = TAILQ_NEXT(mp, mnt_list); continue; } #ifdef MAC if (mac_mount_check_stat(td->td_ucred, mp) != 0) { nmp = TAILQ_NEXT(mp, mnt_list); continue; } #endif if (mode == MNT_WAIT) { if (vfs_busy(mp, MBF_MNTLSTLOCK) != 0) { /* * If vfs_busy() failed, and MBF_NOWAIT * wasn't passed, then the mp is gone. * Furthermore, because of MBF_MNTLSTLOCK, * the mountlist_mtx was dropped. We have * no other choice than to start over. */ mtx_unlock(&mountlist_mtx); free(tofree, M_STATFS); goto restart; } } else { if (vfs_busy(mp, MBF_NOWAIT | MBF_MNTLSTLOCK) != 0) { nmp = TAILQ_NEXT(mp, mnt_list); continue; } } if (sfsp != NULL && count < maxcount) { sp = &mp->mnt_stat; /* * Set these in case the underlying filesystem * fails to do so. */ sp->f_version = STATFS_VERSION; sp->f_namemax = NAME_MAX; sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK; /* * If MNT_NOWAIT is specified, do not refresh * the fsstat cache. */ if (mode != MNT_NOWAIT) { error = VFS_STATFS(mp, sp); if (error != 0) { mtx_lock(&mountlist_mtx); nmp = TAILQ_NEXT(mp, mnt_list); vfs_unbusy(mp); continue; } } if (priv_check(td, PRIV_VFS_GENERATION)) { sptmp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); *sptmp = *sp; sptmp->f_fsid.val[0] = sptmp->f_fsid.val[1] = 0; prison_enforce_statfs(td->td_ucred, mp, sptmp); sp = sptmp; } else sptmp = NULL; if (bufseg == UIO_SYSSPACE) { bcopy(sp, sfsp, sizeof(*sp)); free(sptmp, M_STATFS); } else /* if (bufseg == UIO_USERSPACE) */ { error = copyout(sp, sfsp, sizeof(*sp)); free(sptmp, M_STATFS); if (error != 0) { vfs_unbusy(mp); return (error); } } sfsp++; } count++; mtx_lock(&mountlist_mtx); nmp = TAILQ_NEXT(mp, mnt_list); vfs_unbusy(mp); } mtx_unlock(&mountlist_mtx); if (sfsp != NULL && count > maxcount) *countp = maxcount; else *countp = count; return (0); } #ifdef COMPAT_FREEBSD4 /* * Get old format filesystem statistics. */ static void freebsd4_cvtstatfs(struct statfs *, struct ostatfs *); #ifndef _SYS_SYSPROTO_H_ struct freebsd4_statfs_args { char *path; struct ostatfs *buf; }; #endif int freebsd4_statfs(struct thread *td, struct freebsd4_statfs_args *uap) { struct ostatfs osb; struct statfs *sfp; int error; sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_statfs(td, uap->path, UIO_USERSPACE, sfp); if (error == 0) { freebsd4_cvtstatfs(sfp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); } free(sfp, M_STATFS); return (error); } /* * Get filesystem statistics. */ #ifndef _SYS_SYSPROTO_H_ struct freebsd4_fstatfs_args { int fd; struct ostatfs *buf; }; #endif int freebsd4_fstatfs(struct thread *td, struct freebsd4_fstatfs_args *uap) { struct ostatfs osb; struct statfs *sfp; int error; sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_fstatfs(td, uap->fd, sfp); if (error == 0) { freebsd4_cvtstatfs(sfp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); } free(sfp, M_STATFS); return (error); } /* * Get statistics on all filesystems. */ #ifndef _SYS_SYSPROTO_H_ struct freebsd4_getfsstat_args { struct ostatfs *buf; long bufsize; int mode; }; #endif int freebsd4_getfsstat(struct thread *td, struct freebsd4_getfsstat_args *uap) { struct statfs *buf, *sp; struct ostatfs osb; size_t count, size; int error; if (uap->bufsize < 0) return (EINVAL); count = uap->bufsize / sizeof(struct ostatfs); if (count > SIZE_MAX / sizeof(struct statfs)) return (EINVAL); size = count * sizeof(struct statfs); error = kern_getfsstat(td, &buf, size, &count, UIO_SYSSPACE, uap->mode); if (error == 0) td->td_retval[0] = count; if (size != 0) { sp = buf; while (count != 0 && error == 0) { freebsd4_cvtstatfs(sp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); sp++; uap->buf++; count--; } free(buf, M_STATFS); } return (error); } /* * Implement fstatfs() for (NFS) file handles. */ #ifndef _SYS_SYSPROTO_H_ struct freebsd4_fhstatfs_args { struct fhandle *u_fhp; struct ostatfs *buf; }; #endif int freebsd4_fhstatfs(struct thread *td, struct freebsd4_fhstatfs_args *uap) { struct ostatfs osb; struct statfs *sfp; fhandle_t fh; int error; error = copyin(uap->u_fhp, &fh, sizeof(fhandle_t)); if (error != 0) return (error); sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_fhstatfs(td, fh, sfp); if (error == 0) { freebsd4_cvtstatfs(sfp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); } free(sfp, M_STATFS); return (error); } /* * Convert a new format statfs structure to an old format statfs structure. */ static void freebsd4_cvtstatfs(struct statfs *nsp, struct ostatfs *osp) { statfs_scale_blocks(nsp, LONG_MAX); bzero(osp, sizeof(*osp)); osp->f_bsize = nsp->f_bsize; osp->f_iosize = MIN(nsp->f_iosize, LONG_MAX); osp->f_blocks = nsp->f_blocks; osp->f_bfree = nsp->f_bfree; osp->f_bavail = nsp->f_bavail; osp->f_files = MIN(nsp->f_files, LONG_MAX); osp->f_ffree = MIN(nsp->f_ffree, LONG_MAX); osp->f_owner = nsp->f_owner; osp->f_type = nsp->f_type; osp->f_flags = nsp->f_flags; osp->f_syncwrites = MIN(nsp->f_syncwrites, LONG_MAX); osp->f_asyncwrites = MIN(nsp->f_asyncwrites, LONG_MAX); osp->f_syncreads = MIN(nsp->f_syncreads, LONG_MAX); osp->f_asyncreads = MIN(nsp->f_asyncreads, LONG_MAX); strlcpy(osp->f_fstypename, nsp->f_fstypename, MIN(MFSNAMELEN, OMFSNAMELEN)); strlcpy(osp->f_mntonname, nsp->f_mntonname, MIN(MNAMELEN, OMNAMELEN)); strlcpy(osp->f_mntfromname, nsp->f_mntfromname, MIN(MNAMELEN, OMNAMELEN)); osp->f_fsid = nsp->f_fsid; } #endif /* COMPAT_FREEBSD4 */ #if defined(COMPAT_FREEBSD11) /* * Get old format filesystem statistics. */ static void freebsd11_cvtstatfs(struct statfs *, struct freebsd11_statfs *); int freebsd11_statfs(struct thread *td, struct freebsd11_statfs_args *uap) { struct freebsd11_statfs osb; struct statfs *sfp; int error; sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_statfs(td, uap->path, UIO_USERSPACE, sfp); if (error == 0) { freebsd11_cvtstatfs(sfp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); } free(sfp, M_STATFS); return (error); } /* * Get filesystem statistics. */ int freebsd11_fstatfs(struct thread *td, struct freebsd11_fstatfs_args *uap) { struct freebsd11_statfs osb; struct statfs *sfp; int error; sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_fstatfs(td, uap->fd, sfp); if (error == 0) { freebsd11_cvtstatfs(sfp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); } free(sfp, M_STATFS); return (error); } /* * Get statistics on all filesystems. */ int freebsd11_getfsstat(struct thread *td, struct freebsd11_getfsstat_args *uap) { struct freebsd11_statfs osb; struct statfs *buf, *sp; size_t count, size; int error; count = uap->bufsize / sizeof(struct ostatfs); size = count * sizeof(struct statfs); error = kern_getfsstat(td, &buf, size, &count, UIO_SYSSPACE, uap->mode); if (error == 0) td->td_retval[0] = count; if (size > 0) { sp = buf; while (count > 0 && error == 0) { freebsd11_cvtstatfs(sp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); sp++; uap->buf++; count--; } free(buf, M_STATFS); } return (error); } /* * Implement fstatfs() for (NFS) file handles. */ int freebsd11_fhstatfs(struct thread *td, struct freebsd11_fhstatfs_args *uap) { struct freebsd11_statfs osb; struct statfs *sfp; fhandle_t fh; int error; error = copyin(uap->u_fhp, &fh, sizeof(fhandle_t)); if (error) return (error); sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_fhstatfs(td, fh, sfp); if (error == 0) { freebsd11_cvtstatfs(sfp, &osb); error = copyout(&osb, uap->buf, sizeof(osb)); } free(sfp, M_STATFS); return (error); } /* * Convert a new format statfs structure to an old format statfs structure. */ static void freebsd11_cvtstatfs(struct statfs *nsp, struct freebsd11_statfs *osp) { bzero(osp, sizeof(*osp)); osp->f_version = FREEBSD11_STATFS_VERSION; osp->f_type = nsp->f_type; osp->f_flags = nsp->f_flags; osp->f_bsize = nsp->f_bsize; osp->f_iosize = nsp->f_iosize; osp->f_blocks = nsp->f_blocks; osp->f_bfree = nsp->f_bfree; osp->f_bavail = nsp->f_bavail; osp->f_files = nsp->f_files; osp->f_ffree = nsp->f_ffree; osp->f_syncwrites = nsp->f_syncwrites; osp->f_asyncwrites = nsp->f_asyncwrites; osp->f_syncreads = nsp->f_syncreads; osp->f_asyncreads = nsp->f_asyncreads; osp->f_namemax = nsp->f_namemax; osp->f_owner = nsp->f_owner; osp->f_fsid = nsp->f_fsid; strlcpy(osp->f_fstypename, nsp->f_fstypename, MIN(MFSNAMELEN, sizeof(osp->f_fstypename))); strlcpy(osp->f_mntonname, nsp->f_mntonname, MIN(MNAMELEN, sizeof(osp->f_mntonname))); strlcpy(osp->f_mntfromname, nsp->f_mntfromname, MIN(MNAMELEN, sizeof(osp->f_mntfromname))); } #endif /* COMPAT_FREEBSD11 */ /* * Change current working directory to a given file descriptor. */ #ifndef _SYS_SYSPROTO_H_ struct fchdir_args { int fd; }; #endif int sys_fchdir(struct thread *td, struct fchdir_args *uap) { struct vnode *vp, *tdp; struct mount *mp; struct file *fp; int error; AUDIT_ARG_FD(uap->fd); error = getvnode(td, uap->fd, &cap_fchdir_rights, &fp); if (error != 0) return (error); vp = fp->f_vnode; vrefact(vp); fdrop(fp, td); vn_lock(vp, LK_SHARED | LK_RETRY); AUDIT_ARG_VNODE1(vp); error = change_dir(vp, td); while (!error && (mp = vp->v_mountedhere) != NULL) { if (vfs_busy(mp, 0)) continue; error = VFS_ROOT(mp, LK_SHARED, &tdp); vfs_unbusy(mp); if (error != 0) break; vput(vp); vp = tdp; } if (error != 0) { vput(vp); return (error); } VOP_UNLOCK(vp, 0); pwd_chdir(td, vp); return (0); } /* * Change current working directory (``.''). */ #ifndef _SYS_SYSPROTO_H_ struct chdir_args { char *path; }; #endif int sys_chdir(struct thread *td, struct chdir_args *uap) { return (kern_chdir(td, uap->path, UIO_USERSPACE)); } int kern_chdir(struct thread *td, const char *path, enum uio_seg pathseg) { struct nameidata nd; int error; NDINIT(&nd, LOOKUP, FOLLOW | LOCKSHARED | LOCKLEAF | AUDITVNODE1, pathseg, path, td); if ((error = namei(&nd)) != 0) return (error); if ((error = change_dir(nd.ni_vp, td)) != 0) { vput(nd.ni_vp); NDFREE(&nd, NDF_ONLY_PNBUF); return (error); } VOP_UNLOCK(nd.ni_vp, 0); NDFREE(&nd, NDF_ONLY_PNBUF); pwd_chdir(td, nd.ni_vp); return (0); } /* * Change notion of root (``/'') directory. */ #ifndef _SYS_SYSPROTO_H_ struct chroot_args { char *path; }; #endif int sys_chroot(struct thread *td, struct chroot_args *uap) { struct nameidata nd; int error; error = priv_check(td, PRIV_VFS_CHROOT); if (error != 0) return (error); NDINIT(&nd, LOOKUP, FOLLOW | LOCKSHARED | LOCKLEAF | AUDITVNODE1, UIO_USERSPACE, uap->path, td); error = namei(&nd); if (error != 0) goto error; error = change_dir(nd.ni_vp, td); if (error != 0) goto e_vunlock; #ifdef MAC error = mac_vnode_check_chroot(td->td_ucred, nd.ni_vp); if (error != 0) goto e_vunlock; #endif VOP_UNLOCK(nd.ni_vp, 0); error = pwd_chroot(td, nd.ni_vp); vrele(nd.ni_vp); NDFREE(&nd, NDF_ONLY_PNBUF); return (error); e_vunlock: vput(nd.ni_vp); error: NDFREE(&nd, NDF_ONLY_PNBUF); return (error); } /* * Common routine for chroot and chdir. Callers must provide a locked vnode * instance. */ int change_dir(struct vnode *vp, struct thread *td) { #ifdef MAC int error; #endif ASSERT_VOP_LOCKED(vp, "change_dir(): vp not locked"); if (vp->v_type != VDIR) return (ENOTDIR); #ifdef MAC error = mac_vnode_check_chdir(td->td_ucred, vp); if (error != 0) return (error); #endif return (VOP_ACCESS(vp, VEXEC, td->td_ucred, td)); } static __inline void flags_to_rights(int flags, cap_rights_t *rightsp) { if (flags & O_EXEC) { cap_rights_set(rightsp, CAP_FEXECVE); } else { switch ((flags & O_ACCMODE)) { case O_RDONLY: cap_rights_set(rightsp, CAP_READ); break; case O_RDWR: cap_rights_set(rightsp, CAP_READ); /* FALLTHROUGH */ case O_WRONLY: cap_rights_set(rightsp, CAP_WRITE); if (!(flags & (O_APPEND | O_TRUNC))) cap_rights_set(rightsp, CAP_SEEK); break; } } if (flags & O_CREAT) cap_rights_set(rightsp, CAP_CREATE); if (flags & O_TRUNC) cap_rights_set(rightsp, CAP_FTRUNCATE); if (flags & (O_SYNC | O_FSYNC)) cap_rights_set(rightsp, CAP_FSYNC); if (flags & (O_EXLOCK | O_SHLOCK)) cap_rights_set(rightsp, CAP_FLOCK); } /* * Check permissions, allocate an open file structure, and call the device * open routine if any. */ #ifndef _SYS_SYSPROTO_H_ struct open_args { char *path; int flags; int mode; }; #endif int sys_open(struct thread *td, struct open_args *uap) { return (kern_openat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->flags, uap->mode)); } #ifndef _SYS_SYSPROTO_H_ struct openat_args { int fd; char *path; int flag; int mode; }; #endif int sys_openat(struct thread *td, struct openat_args *uap) { AUDIT_ARG_FD(uap->fd); return (kern_openat(td, uap->fd, uap->path, UIO_USERSPACE, uap->flag, uap->mode)); } int kern_openat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int flags, int mode) { struct proc *p = td->td_proc; struct filedesc *fdp = p->p_fd; struct file *fp; struct vnode *vp; struct nameidata nd; cap_rights_t rights; int cmode, error, indx; indx = -1; AUDIT_ARG_FFLAGS(flags); AUDIT_ARG_MODE(mode); cap_rights_init(&rights, CAP_LOOKUP); flags_to_rights(flags, &rights); /* * Only one of the O_EXEC, O_RDONLY, O_WRONLY and O_RDWR flags * may be specified. */ if (flags & O_EXEC) { if (flags & O_ACCMODE) return (EINVAL); } else if ((flags & O_ACCMODE) == O_ACCMODE) { return (EINVAL); } else { flags = FFLAGS(flags); } /* * Allocate a file structure. The descriptor to reference it * is allocated and set by finstall() below. */ error = falloc_noinstall(td, &fp); if (error != 0) return (error); /* * An extra reference on `fp' has been held for us by * falloc_noinstall(). */ /* Set the flags early so the finit in devfs can pick them up. */ fp->f_flag = flags & FMASK; cmode = ((mode & ~fdp->fd_cmask) & ALLPERMS) & ~S_ISTXT; NDINIT_ATRIGHTS(&nd, LOOKUP, FOLLOW | AUDITVNODE1, pathseg, path, fd, &rights, td); td->td_dupfd = -1; /* XXX check for fdopen */ error = vn_open(&nd, &flags, cmode, fp); if (error != 0) { /* * If the vn_open replaced the method vector, something * wonderous happened deep below and we just pass it up * pretending we know what we do. */ if (error == ENXIO && fp->f_ops != &badfileops) goto success; /* * Handle special fdopen() case. bleh. * * Don't do this for relative (capability) lookups; we don't * understand exactly what would happen, and we don't think * that it ever should. */ if ((nd.ni_lcf & NI_LCF_STRICTRELATIVE) == 0 && (error == ENODEV || error == ENXIO) && td->td_dupfd >= 0) { error = dupfdopen(td, fdp, td->td_dupfd, flags, error, &indx); if (error == 0) goto success; } goto bad; } td->td_dupfd = 0; NDFREE(&nd, NDF_ONLY_PNBUF); vp = nd.ni_vp; /* * Store the vnode, for any f_type. Typically, the vnode use * count is decremented by direct call to vn_closefile() for * files that switched type in the cdevsw fdopen() method. */ fp->f_vnode = vp; /* * If the file wasn't claimed by devfs bind it to the normal * vnode operations here. */ if (fp->f_ops == &badfileops) { KASSERT(vp->v_type != VFIFO, ("Unexpected fifo.")); fp->f_seqcount = 1; finit(fp, (flags & FMASK) | (fp->f_flag & FHASLOCK), DTYPE_VNODE, vp, &vnops); } VOP_UNLOCK(vp, 0); if (flags & O_TRUNC) { error = fo_truncate(fp, 0, td->td_ucred, td); if (error != 0) goto bad; } success: /* * If we haven't already installed the FD (for dupfdopen), do so now. */ if (indx == -1) { struct filecaps *fcaps; #ifdef CAPABILITIES if ((nd.ni_lcf & NI_LCF_STRICTRELATIVE) != 0) fcaps = &nd.ni_filecaps; else #endif fcaps = NULL; error = finstall(td, fp, &indx, flags, fcaps); /* On success finstall() consumes fcaps. */ if (error != 0) { filecaps_free(&nd.ni_filecaps); goto bad; } } else { filecaps_free(&nd.ni_filecaps); } /* * Release our private reference, leaving the one associated with * the descriptor table intact. */ fdrop(fp, td); td->td_retval[0] = indx; return (0); bad: KASSERT(indx == -1, ("indx=%d, should be -1", indx)); fdrop(fp, td); return (error); } #ifdef COMPAT_43 /* * Create a file. */ #ifndef _SYS_SYSPROTO_H_ struct ocreat_args { char *path; int mode; }; #endif int ocreat(struct thread *td, struct ocreat_args *uap) { return (kern_openat(td, AT_FDCWD, uap->path, UIO_USERSPACE, O_WRONLY | O_CREAT | O_TRUNC, uap->mode)); } #endif /* COMPAT_43 */ /* * Create a special file. */ #ifndef _SYS_SYSPROTO_H_ struct mknodat_args { int fd; char *path; mode_t mode; dev_t dev; }; #endif int sys_mknodat(struct thread *td, struct mknodat_args *uap) { return (kern_mknodat(td, uap->fd, uap->path, UIO_USERSPACE, uap->mode, uap->dev)); } #if defined(COMPAT_FREEBSD11) int freebsd11_mknod(struct thread *td, struct freebsd11_mknod_args *uap) { return (kern_mknodat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->mode, uap->dev)); } int freebsd11_mknodat(struct thread *td, struct freebsd11_mknodat_args *uap) { return (kern_mknodat(td, uap->fd, uap->path, UIO_USERSPACE, uap->mode, uap->dev)); } #endif /* COMPAT_FREEBSD11 */ int kern_mknodat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int mode, dev_t dev) { struct vnode *vp; struct mount *mp; struct vattr vattr; struct nameidata nd; int error, whiteout = 0; AUDIT_ARG_MODE(mode); AUDIT_ARG_DEV(dev); switch (mode & S_IFMT) { case S_IFCHR: case S_IFBLK: error = priv_check(td, PRIV_VFS_MKNOD_DEV); if (error == 0 && dev == VNOVAL) error = EINVAL; break; case S_IFWHT: error = priv_check(td, PRIV_VFS_MKNOD_WHT); break; case S_IFIFO: if (dev == 0) return (kern_mkfifoat(td, fd, path, pathseg, mode)); /* FALLTHROUGH */ default: error = EINVAL; break; } if (error != 0) return (error); restart: bwillwrite(); NDINIT_ATRIGHTS(&nd, CREATE, LOCKPARENT | SAVENAME | AUDITVNODE1 | NOCACHE, pathseg, path, fd, &cap_mknodat_rights, td); if ((error = namei(&nd)) != 0) return (error); vp = nd.ni_vp; if (vp != NULL) { NDFREE(&nd, NDF_ONLY_PNBUF); if (vp == nd.ni_dvp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); vrele(vp); return (EEXIST); } else { VATTR_NULL(&vattr); vattr.va_mode = (mode & ALLPERMS) & ~td->td_proc->p_fd->fd_cmask; vattr.va_rdev = dev; whiteout = 0; switch (mode & S_IFMT) { case S_IFCHR: vattr.va_type = VCHR; break; case S_IFBLK: vattr.va_type = VBLK; break; case S_IFWHT: whiteout = 1; break; default: panic("kern_mknod: invalid mode"); } } if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0) return (error); goto restart; } #ifdef MAC if (error == 0 && !whiteout) error = mac_vnode_check_create(td->td_ucred, nd.ni_dvp, &nd.ni_cnd, &vattr); #endif if (error == 0) { if (whiteout) error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, CREATE); else { error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr); if (error == 0) vput(nd.ni_vp); } } NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); vn_finished_write(mp); return (error); } /* * Create a named pipe. */ #ifndef _SYS_SYSPROTO_H_ struct mkfifo_args { char *path; int mode; }; #endif int sys_mkfifo(struct thread *td, struct mkfifo_args *uap) { return (kern_mkfifoat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->mode)); } #ifndef _SYS_SYSPROTO_H_ struct mkfifoat_args { int fd; char *path; mode_t mode; }; #endif int sys_mkfifoat(struct thread *td, struct mkfifoat_args *uap) { return (kern_mkfifoat(td, uap->fd, uap->path, UIO_USERSPACE, uap->mode)); } int kern_mkfifoat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int mode) { struct mount *mp; struct vattr vattr; struct nameidata nd; int error; AUDIT_ARG_MODE(mode); restart: bwillwrite(); NDINIT_ATRIGHTS(&nd, CREATE, LOCKPARENT | SAVENAME | AUDITVNODE1 | NOCACHE, pathseg, path, fd, &cap_mkfifoat_rights, td); if ((error = namei(&nd)) != 0) return (error); if (nd.ni_vp != NULL) { NDFREE(&nd, NDF_ONLY_PNBUF); if (nd.ni_vp == nd.ni_dvp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); vrele(nd.ni_vp); return (EEXIST); } if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0) return (error); goto restart; } VATTR_NULL(&vattr); vattr.va_type = VFIFO; vattr.va_mode = (mode & ALLPERMS) & ~td->td_proc->p_fd->fd_cmask; #ifdef MAC error = mac_vnode_check_create(td->td_ucred, nd.ni_dvp, &nd.ni_cnd, &vattr); if (error != 0) goto out; #endif error = VOP_MKNOD(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr); if (error == 0) vput(nd.ni_vp); #ifdef MAC out: #endif vput(nd.ni_dvp); vn_finished_write(mp); NDFREE(&nd, NDF_ONLY_PNBUF); return (error); } /* * Make a hard file link. */ #ifndef _SYS_SYSPROTO_H_ struct link_args { char *path; char *link; }; #endif int sys_link(struct thread *td, struct link_args *uap) { return (kern_linkat(td, AT_FDCWD, AT_FDCWD, uap->path, uap->link, UIO_USERSPACE, FOLLOW)); } #ifndef _SYS_SYSPROTO_H_ struct linkat_args { int fd1; char *path1; int fd2; char *path2; int flag; }; #endif int sys_linkat(struct thread *td, struct linkat_args *uap) { int flag; flag = uap->flag; if ((flag & ~(AT_SYMLINK_FOLLOW | AT_BENEATH)) != 0) return (EINVAL); return (kern_linkat(td, uap->fd1, uap->fd2, uap->path1, uap->path2, UIO_USERSPACE, ((flag & AT_SYMLINK_FOLLOW) != 0 ? FOLLOW : NOFOLLOW) | ((flag & AT_BENEATH) != 0 ? BENEATH : 0))); } int hardlink_check_uid = 0; SYSCTL_INT(_security_bsd, OID_AUTO, hardlink_check_uid, CTLFLAG_RW, &hardlink_check_uid, 0, "Unprivileged processes cannot create hard links to files owned by other " "users"); static int hardlink_check_gid = 0; SYSCTL_INT(_security_bsd, OID_AUTO, hardlink_check_gid, CTLFLAG_RW, &hardlink_check_gid, 0, "Unprivileged processes cannot create hard links to files owned by other " "groups"); static int can_hardlink(struct vnode *vp, struct ucred *cred) { struct vattr va; int error; if (!hardlink_check_uid && !hardlink_check_gid) return (0); error = VOP_GETATTR(vp, &va, cred); if (error != 0) return (error); if (hardlink_check_uid && cred->cr_uid != va.va_uid) { error = priv_check_cred(cred, PRIV_VFS_LINK); if (error != 0) return (error); } if (hardlink_check_gid && !groupmember(va.va_gid, cred)) { error = priv_check_cred(cred, PRIV_VFS_LINK); if (error != 0) return (error); } return (0); } int kern_linkat(struct thread *td, int fd1, int fd2, const char *path1, const char *path2, enum uio_seg segflag, int follow) { struct nameidata nd; int error; do { bwillwrite(); NDINIT_ATRIGHTS(&nd, LOOKUP, follow | AUDITVNODE1, segflag, path1, fd1, &cap_linkat_source_rights, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = kern_linkat_vp(td, nd.ni_vp, fd2, path2, segflag); } while (error == EAGAIN); return (error); } static int kern_linkat_vp(struct thread *td, struct vnode *vp, int fd, const char *path, enum uio_seg segflag) { struct nameidata nd; struct mount *mp; int error; if (vp->v_type == VDIR) { vrele(vp); return (EPERM); /* POSIX */ } NDINIT_ATRIGHTS(&nd, CREATE, LOCKPARENT | SAVENAME | AUDITVNODE2 | NOCACHE, segflag, path, fd, &cap_linkat_target_rights, td); if ((error = namei(&nd)) == 0) { if (nd.ni_vp != NULL) { NDFREE(&nd, NDF_ONLY_PNBUF); if (nd.ni_dvp == nd.ni_vp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); vrele(nd.ni_vp); vrele(vp); return (EEXIST); } else if (nd.ni_dvp->v_mount != vp->v_mount) { /* * Cross-device link. No need to recheck * vp->v_type, since it cannot change, except * to VBAD. */ NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); vrele(vp); return (EXDEV); } else if ((error = vn_lock(vp, LK_EXCLUSIVE)) == 0) { error = can_hardlink(vp, td->td_ucred); #ifdef MAC if (error == 0) error = mac_vnode_check_link(td->td_ucred, nd.ni_dvp, vp, &nd.ni_cnd); #endif if (error != 0) { vput(vp); vput(nd.ni_dvp); NDFREE(&nd, NDF_ONLY_PNBUF); return (error); } error = vn_start_write(vp, &mp, V_NOWAIT); if (error != 0) { vput(vp); vput(nd.ni_dvp); NDFREE(&nd, NDF_ONLY_PNBUF); error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH); if (error != 0) return (error); return (EAGAIN); } error = VOP_LINK(nd.ni_dvp, vp, &nd.ni_cnd); VOP_UNLOCK(vp, 0); vput(nd.ni_dvp); vn_finished_write(mp); NDFREE(&nd, NDF_ONLY_PNBUF); } else { vput(nd.ni_dvp); NDFREE(&nd, NDF_ONLY_PNBUF); vrele(vp); return (EAGAIN); } } vrele(vp); return (error); } /* * Make a symbolic link. */ #ifndef _SYS_SYSPROTO_H_ struct symlink_args { char *path; char *link; }; #endif int sys_symlink(struct thread *td, struct symlink_args *uap) { return (kern_symlinkat(td, uap->path, AT_FDCWD, uap->link, UIO_USERSPACE)); } #ifndef _SYS_SYSPROTO_H_ struct symlinkat_args { char *path; int fd; char *path2; }; #endif int sys_symlinkat(struct thread *td, struct symlinkat_args *uap) { return (kern_symlinkat(td, uap->path1, uap->fd, uap->path2, UIO_USERSPACE)); } int kern_symlinkat(struct thread *td, const char *path1, int fd, const char *path2, enum uio_seg segflg) { struct mount *mp; struct vattr vattr; const char *syspath; char *tmppath; struct nameidata nd; int error; if (segflg == UIO_SYSSPACE) { syspath = path1; } else { tmppath = uma_zalloc(namei_zone, M_WAITOK); if ((error = copyinstr(path1, tmppath, MAXPATHLEN, NULL)) != 0) goto out; syspath = tmppath; } AUDIT_ARG_TEXT(syspath); restart: bwillwrite(); NDINIT_ATRIGHTS(&nd, CREATE, LOCKPARENT | SAVENAME | AUDITVNODE1 | NOCACHE, segflg, path2, fd, &cap_symlinkat_rights, td); if ((error = namei(&nd)) != 0) goto out; if (nd.ni_vp) { NDFREE(&nd, NDF_ONLY_PNBUF); if (nd.ni_vp == nd.ni_dvp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); vrele(nd.ni_vp); error = EEXIST; goto out; } if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0) goto out; goto restart; } VATTR_NULL(&vattr); vattr.va_mode = ACCESSPERMS &~ td->td_proc->p_fd->fd_cmask; #ifdef MAC vattr.va_type = VLNK; error = mac_vnode_check_create(td->td_ucred, nd.ni_dvp, &nd.ni_cnd, &vattr); if (error != 0) goto out2; #endif error = VOP_SYMLINK(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr, syspath); if (error == 0) vput(nd.ni_vp); #ifdef MAC out2: #endif NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); vn_finished_write(mp); out: if (segflg != UIO_SYSSPACE) uma_zfree(namei_zone, tmppath); return (error); } /* * Delete a whiteout from the filesystem. */ #ifndef _SYS_SYSPROTO_H_ struct undelete_args { char *path; }; #endif int sys_undelete(struct thread *td, struct undelete_args *uap) { struct mount *mp; struct nameidata nd; int error; restart: bwillwrite(); NDINIT(&nd, DELETE, LOCKPARENT | DOWHITEOUT | AUDITVNODE1, UIO_USERSPACE, uap->path, td); error = namei(&nd); if (error != 0) return (error); if (nd.ni_vp != NULLVP || !(nd.ni_cnd.cn_flags & ISWHITEOUT)) { NDFREE(&nd, NDF_ONLY_PNBUF); if (nd.ni_vp == nd.ni_dvp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); if (nd.ni_vp) vrele(nd.ni_vp); return (EEXIST); } if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0) return (error); goto restart; } error = VOP_WHITEOUT(nd.ni_dvp, &nd.ni_cnd, DELETE); NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); vn_finished_write(mp); return (error); } /* * Delete a name from the filesystem. */ #ifndef _SYS_SYSPROTO_H_ struct unlink_args { char *path; }; #endif int sys_unlink(struct thread *td, struct unlink_args *uap) { - return (kern_unlinkat(td, AT_FDCWD, uap->path, UIO_USERSPACE, 0, 0)); + return (kern_funlinkat(td, AT_FDCWD, uap->path, FD_NONE, UIO_USERSPACE, + 0, 0)); } +static int +kern_funlinkat_ex(struct thread *td, int dfd, const char *path, int fd, + int flag, enum uio_seg pathseg, ino_t oldinum) +{ + + if ((flag & ~AT_REMOVEDIR) != 0) + return (EINVAL); + + if ((flag & AT_REMOVEDIR) != 0) + return (kern_frmdirat(td, dfd, path, fd, UIO_USERSPACE, 0)); + + return (kern_funlinkat(td, dfd, path, fd, UIO_USERSPACE, 0, 0)); +} + #ifndef _SYS_SYSPROTO_H_ struct unlinkat_args { int fd; char *path; int flag; }; #endif int sys_unlinkat(struct thread *td, struct unlinkat_args *uap) { - int fd, flag; - const char *path; - flag = uap->flag; - fd = uap->fd; - path = uap->path; + return (kern_funlinkat_ex(td, uap->fd, uap->path, FD_NONE, uap->flag, + UIO_USERSPACE, 0)); +} - if ((flag & ~(AT_REMOVEDIR | AT_BENEATH)) != 0) - return (EINVAL); +#ifndef _SYS_SYSPROTO_H_ +struct funlinkat_args { + int dfd; + const char *path; + int fd; + int flag; +}; +#endif +int +sys_funlinkat(struct thread *td, struct funlinkat_args *uap) +{ - if ((uap->flag & AT_REMOVEDIR) != 0) - return (kern_rmdirat(td, fd, path, UIO_USERSPACE, flag)); - else - return (kern_unlinkat(td, fd, path, UIO_USERSPACE, flag, 0)); + return (kern_funlinkat_ex(td, uap->dfd, uap->path, uap->fd, uap->flag, + UIO_USERSPACE, 0)); } int -kern_unlinkat(struct thread *td, int fd, const char *path, +kern_funlinkat(struct thread *td, int dfd, const char *path, int fd, enum uio_seg pathseg, int flag, ino_t oldinum) { struct mount *mp; + struct file *fp; struct vnode *vp; struct nameidata nd; struct stat sb; + cap_rights_t rights; int error; + fp = NULL; + if (fd != FD_NONE) { + error = getvnode(td, fd, cap_rights_init(&rights, CAP_LOOKUP), + &fp); + if (error != 0) + return (error); + } + restart: bwillwrite(); NDINIT_ATRIGHTS(&nd, DELETE, LOCKPARENT | LOCKLEAF | AUDITVNODE1 | ((flag & AT_BENEATH) != 0 ? BENEATH : 0), - pathseg, path, fd, &cap_unlinkat_rights, td); - if ((error = namei(&nd)) != 0) - return (error == EINVAL ? EPERM : error); + pathseg, path, dfd, &cap_unlinkat_rights, td); + if ((error = namei(&nd)) != 0) { + if (error == EINVAL) + error = EPERM; + goto fdout; + } vp = nd.ni_vp; if (vp->v_type == VDIR && oldinum == 0) { error = EPERM; /* POSIX */ } else if (oldinum != 0 && ((error = vn_stat(vp, &sb, td->td_ucred, NOCRED, td)) == 0) && sb.st_ino != oldinum) { - error = EIDRM; /* Identifier removed */ + error = EIDRM; /* Identifier removed */ + } else if (fp != NULL && fp->f_vnode != vp) { + if ((fp->f_vnode->v_iflag & VI_DOOMED) != 0) + error = EBADF; + else + error = EDEADLK; } else { /* * The root of a mounted filesystem cannot be deleted. * * XXX: can this only be a VDIR case? */ if (vp->v_vflag & VV_ROOT) error = EBUSY; } if (error == 0) { if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if (vp == nd.ni_dvp) vrele(vp); else vput(vp); if ((error = vn_start_write(NULL, &mp, - V_XSLEEP | PCATCH)) != 0) - return (error); + V_XSLEEP | PCATCH)) != 0) { + goto fdout; + } goto restart; } #ifdef MAC error = mac_vnode_check_unlink(td->td_ucred, nd.ni_dvp, vp, &nd.ni_cnd); if (error != 0) goto out; #endif vfs_notify_upper(vp, VFS_NOTIFY_UPPER_UNLINK); error = VOP_REMOVE(nd.ni_dvp, vp, &nd.ni_cnd); #ifdef MAC out: #endif vn_finished_write(mp); } NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if (vp == nd.ni_dvp) vrele(vp); else vput(vp); +fdout: + if (fp != NULL) + fdrop(fp, td); return (error); } /* * Reposition read/write file offset. */ #ifndef _SYS_SYSPROTO_H_ struct lseek_args { int fd; int pad; off_t offset; int whence; }; #endif int sys_lseek(struct thread *td, struct lseek_args *uap) { return (kern_lseek(td, uap->fd, uap->offset, uap->whence)); } int kern_lseek(struct thread *td, int fd, off_t offset, int whence) { struct file *fp; int error; AUDIT_ARG_FD(fd); error = fget(td, fd, &cap_seek_rights, &fp); if (error != 0) return (error); error = (fp->f_ops->fo_flags & DFLAG_SEEKABLE) != 0 ? fo_seek(fp, offset, whence, td) : ESPIPE; fdrop(fp, td); return (error); } #if defined(COMPAT_43) /* * Reposition read/write file offset. */ #ifndef _SYS_SYSPROTO_H_ struct olseek_args { int fd; long offset; int whence; }; #endif int olseek(struct thread *td, struct olseek_args *uap) { return (kern_lseek(td, uap->fd, uap->offset, uap->whence)); } #endif /* COMPAT_43 */ #if defined(COMPAT_FREEBSD6) /* Version with the 'pad' argument */ int freebsd6_lseek(struct thread *td, struct freebsd6_lseek_args *uap) { return (kern_lseek(td, uap->fd, uap->offset, uap->whence)); } #endif /* * Check access permissions using passed credentials. */ static int vn_access(struct vnode *vp, int user_flags, struct ucred *cred, struct thread *td) { accmode_t accmode; int error; /* Flags == 0 means only check for existence. */ if (user_flags == 0) return (0); accmode = 0; if (user_flags & R_OK) accmode |= VREAD; if (user_flags & W_OK) accmode |= VWRITE; if (user_flags & X_OK) accmode |= VEXEC; #ifdef MAC error = mac_vnode_check_access(cred, vp, accmode); if (error != 0) return (error); #endif if ((accmode & VWRITE) == 0 || (error = vn_writechk(vp)) == 0) error = VOP_ACCESS(vp, accmode, cred, td); return (error); } /* * Check access permissions using "real" credentials. */ #ifndef _SYS_SYSPROTO_H_ struct access_args { char *path; int amode; }; #endif int sys_access(struct thread *td, struct access_args *uap) { return (kern_accessat(td, AT_FDCWD, uap->path, UIO_USERSPACE, 0, uap->amode)); } #ifndef _SYS_SYSPROTO_H_ struct faccessat_args { int dirfd; char *path; int amode; int flag; } #endif int sys_faccessat(struct thread *td, struct faccessat_args *uap) { return (kern_accessat(td, uap->fd, uap->path, UIO_USERSPACE, uap->flag, uap->amode)); } int kern_accessat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int flag, int amode) { struct ucred *cred, *usecred; struct vnode *vp; struct nameidata nd; int error; if ((flag & ~(AT_EACCESS | AT_BENEATH)) != 0) return (EINVAL); if (amode != F_OK && (amode & ~(R_OK | W_OK | X_OK)) != 0) return (EINVAL); /* * Create and modify a temporary credential instead of one that * is potentially shared (if we need one). */ cred = td->td_ucred; if ((flag & AT_EACCESS) == 0 && ((cred->cr_uid != cred->cr_ruid || cred->cr_rgid != cred->cr_groups[0]))) { usecred = crdup(cred); usecred->cr_uid = cred->cr_ruid; usecred->cr_groups[0] = cred->cr_rgid; td->td_ucred = usecred; } else usecred = cred; AUDIT_ARG_VALUE(amode); NDINIT_ATRIGHTS(&nd, LOOKUP, FOLLOW | LOCKSHARED | LOCKLEAF | AUDITVNODE1 | ((flag & AT_BENEATH) != 0 ? BENEATH : 0), pathseg, path, fd, &cap_fstat_rights, td); if ((error = namei(&nd)) != 0) goto out; vp = nd.ni_vp; error = vn_access(vp, amode, usecred, td); NDFREE(&nd, NDF_ONLY_PNBUF); vput(vp); out: if (usecred != cred) { td->td_ucred = cred; crfree(usecred); } return (error); } /* * Check access permissions using "effective" credentials. */ #ifndef _SYS_SYSPROTO_H_ struct eaccess_args { char *path; int amode; }; #endif int sys_eaccess(struct thread *td, struct eaccess_args *uap) { return (kern_accessat(td, AT_FDCWD, uap->path, UIO_USERSPACE, AT_EACCESS, uap->amode)); } #if defined(COMPAT_43) /* * Get file status; this version follows links. */ #ifndef _SYS_SYSPROTO_H_ struct ostat_args { char *path; struct ostat *ub; }; #endif int ostat(struct thread *td, struct ostat_args *uap) { struct stat sb; struct ostat osb; int error; error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); cvtstat(&sb, &osb); return (copyout(&osb, uap->ub, sizeof (osb))); } /* * Get file status; this version does not follow links. */ #ifndef _SYS_SYSPROTO_H_ struct olstat_args { char *path; struct ostat *ub; }; #endif int olstat(struct thread *td, struct olstat_args *uap) { struct stat sb; struct ostat osb; int error; error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); cvtstat(&sb, &osb); return (copyout(&osb, uap->ub, sizeof (osb))); } /* * Convert from an old to a new stat structure. * XXX: many values are blindly truncated. */ void cvtstat(struct stat *st, struct ostat *ost) { bzero(ost, sizeof(*ost)); ost->st_dev = st->st_dev; ost->st_ino = st->st_ino; ost->st_mode = st->st_mode; ost->st_nlink = st->st_nlink; ost->st_uid = st->st_uid; ost->st_gid = st->st_gid; ost->st_rdev = st->st_rdev; ost->st_size = MIN(st->st_size, INT32_MAX); ost->st_atim = st->st_atim; ost->st_mtim = st->st_mtim; ost->st_ctim = st->st_ctim; ost->st_blksize = st->st_blksize; ost->st_blocks = st->st_blocks; ost->st_flags = st->st_flags; ost->st_gen = st->st_gen; } #endif /* COMPAT_43 */ #if defined(COMPAT_43) || defined(COMPAT_FREEBSD11) int ino64_trunc_error; SYSCTL_INT(_vfs, OID_AUTO, ino64_trunc_error, CTLFLAG_RW, &ino64_trunc_error, 0, "Error on truncation of device, file or inode number, or link count"); int freebsd11_cvtstat(struct stat *st, struct freebsd11_stat *ost) { ost->st_dev = st->st_dev; if (ost->st_dev != st->st_dev) { switch (ino64_trunc_error) { default: /* * Since dev_t is almost raw, don't clamp to the * maximum for case 2, but ignore the error. */ break; case 1: return (EOVERFLOW); } } ost->st_ino = st->st_ino; if (ost->st_ino != st->st_ino) { switch (ino64_trunc_error) { default: case 0: break; case 1: return (EOVERFLOW); case 2: ost->st_ino = UINT32_MAX; break; } } ost->st_mode = st->st_mode; ost->st_nlink = st->st_nlink; if (ost->st_nlink != st->st_nlink) { switch (ino64_trunc_error) { default: case 0: break; case 1: return (EOVERFLOW); case 2: ost->st_nlink = UINT16_MAX; break; } } ost->st_uid = st->st_uid; ost->st_gid = st->st_gid; ost->st_rdev = st->st_rdev; if (ost->st_rdev != st->st_rdev) { switch (ino64_trunc_error) { default: break; case 1: return (EOVERFLOW); } } ost->st_atim = st->st_atim; ost->st_mtim = st->st_mtim; ost->st_ctim = st->st_ctim; ost->st_size = st->st_size; ost->st_blocks = st->st_blocks; ost->st_blksize = st->st_blksize; ost->st_flags = st->st_flags; ost->st_gen = st->st_gen; ost->st_lspare = 0; ost->st_birthtim = st->st_birthtim; bzero((char *)&ost->st_birthtim + sizeof(ost->st_birthtim), sizeof(*ost) - offsetof(struct freebsd11_stat, st_birthtim) - sizeof(ost->st_birthtim)); return (0); } int freebsd11_stat(struct thread *td, struct freebsd11_stat_args* uap) { struct stat sb; struct freebsd11_stat osb; int error; error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); error = freebsd11_cvtstat(&sb, &osb); if (error == 0) error = copyout(&osb, uap->ub, sizeof(osb)); return (error); } int freebsd11_lstat(struct thread *td, struct freebsd11_lstat_args* uap) { struct stat sb; struct freebsd11_stat osb; int error; error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); error = freebsd11_cvtstat(&sb, &osb); if (error == 0) error = copyout(&osb, uap->ub, sizeof(osb)); return (error); } int freebsd11_fhstat(struct thread *td, struct freebsd11_fhstat_args* uap) { struct fhandle fh; struct stat sb; struct freebsd11_stat osb; int error; error = copyin(uap->u_fhp, &fh, sizeof(fhandle_t)); if (error != 0) return (error); error = kern_fhstat(td, fh, &sb); if (error != 0) return (error); error = freebsd11_cvtstat(&sb, &osb); if (error == 0) error = copyout(&osb, uap->sb, sizeof(osb)); return (error); } int freebsd11_fstatat(struct thread *td, struct freebsd11_fstatat_args* uap) { struct stat sb; struct freebsd11_stat osb; int error; error = kern_statat(td, uap->flag, uap->fd, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); error = freebsd11_cvtstat(&sb, &osb); if (error == 0) error = copyout(&osb, uap->buf, sizeof(osb)); return (error); } #endif /* COMPAT_FREEBSD11 */ /* * Get file status */ #ifndef _SYS_SYSPROTO_H_ struct fstatat_args { int fd; char *path; struct stat *buf; int flag; } #endif int sys_fstatat(struct thread *td, struct fstatat_args *uap) { struct stat sb; int error; error = kern_statat(td, uap->flag, uap->fd, uap->path, UIO_USERSPACE, &sb, NULL); if (error == 0) error = copyout(&sb, uap->buf, sizeof (sb)); return (error); } int kern_statat(struct thread *td, int flag, int fd, const char *path, enum uio_seg pathseg, struct stat *sbp, void (*hook)(struct vnode *vp, struct stat *sbp)) { struct nameidata nd; int error; if ((flag & ~(AT_SYMLINK_NOFOLLOW | AT_BENEATH)) != 0) return (EINVAL); NDINIT_ATRIGHTS(&nd, LOOKUP, ((flag & AT_SYMLINK_NOFOLLOW) != 0 ? NOFOLLOW : FOLLOW) | ((flag & AT_BENEATH) != 0 ? BENEATH : 0) | LOCKSHARED | LOCKLEAF | AUDITVNODE1, pathseg, path, fd, &cap_fstat_rights, td); if ((error = namei(&nd)) != 0) return (error); error = vn_stat(nd.ni_vp, sbp, td->td_ucred, NOCRED, td); if (error == 0) { SDT_PROBE2(vfs, , stat, mode, path, sbp->st_mode); if (S_ISREG(sbp->st_mode)) SDT_PROBE2(vfs, , stat, reg, path, pathseg); if (__predict_false(hook != NULL)) hook(nd.ni_vp, sbp); } NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_vp); if (error != 0) return (error); #ifdef __STAT_TIME_T_EXT sbp->st_atim_ext = 0; sbp->st_mtim_ext = 0; sbp->st_ctim_ext = 0; sbp->st_btim_ext = 0; #endif #ifdef KTRACE if (KTRPOINT(td, KTR_STRUCT)) ktrstat(sbp); #endif return (0); } #if defined(COMPAT_FREEBSD11) /* * Implementation of the NetBSD [l]stat() functions. */ void freebsd11_cvtnstat(struct stat *sb, struct nstat *nsb) { bzero(nsb, sizeof(*nsb)); nsb->st_dev = sb->st_dev; nsb->st_ino = sb->st_ino; nsb->st_mode = sb->st_mode; nsb->st_nlink = sb->st_nlink; nsb->st_uid = sb->st_uid; nsb->st_gid = sb->st_gid; nsb->st_rdev = sb->st_rdev; nsb->st_atim = sb->st_atim; nsb->st_mtim = sb->st_mtim; nsb->st_ctim = sb->st_ctim; nsb->st_size = sb->st_size; nsb->st_blocks = sb->st_blocks; nsb->st_blksize = sb->st_blksize; nsb->st_flags = sb->st_flags; nsb->st_gen = sb->st_gen; nsb->st_birthtim = sb->st_birthtim; } #ifndef _SYS_SYSPROTO_H_ struct freebsd11_nstat_args { char *path; struct nstat *ub; }; #endif int freebsd11_nstat(struct thread *td, struct freebsd11_nstat_args *uap) { struct stat sb; struct nstat nsb; int error; error = kern_statat(td, 0, AT_FDCWD, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); freebsd11_cvtnstat(&sb, &nsb); return (copyout(&nsb, uap->ub, sizeof (nsb))); } /* * NetBSD lstat. Get file status; this version does not follow links. */ #ifndef _SYS_SYSPROTO_H_ struct freebsd11_nlstat_args { char *path; struct nstat *ub; }; #endif int freebsd11_nlstat(struct thread *td, struct freebsd11_nlstat_args *uap) { struct stat sb; struct nstat nsb; int error; error = kern_statat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->path, UIO_USERSPACE, &sb, NULL); if (error != 0) return (error); freebsd11_cvtnstat(&sb, &nsb); return (copyout(&nsb, uap->ub, sizeof (nsb))); } #endif /* COMPAT_FREEBSD11 */ /* * Get configurable pathname variables. */ #ifndef _SYS_SYSPROTO_H_ struct pathconf_args { char *path; int name; }; #endif int sys_pathconf(struct thread *td, struct pathconf_args *uap) { long value; int error; error = kern_pathconf(td, uap->path, UIO_USERSPACE, uap->name, FOLLOW, &value); if (error == 0) td->td_retval[0] = value; return (error); } #ifndef _SYS_SYSPROTO_H_ struct lpathconf_args { char *path; int name; }; #endif int sys_lpathconf(struct thread *td, struct lpathconf_args *uap) { long value; int error; error = kern_pathconf(td, uap->path, UIO_USERSPACE, uap->name, NOFOLLOW, &value); if (error == 0) td->td_retval[0] = value; return (error); } int kern_pathconf(struct thread *td, const char *path, enum uio_seg pathseg, int name, u_long flags, long *valuep) { struct nameidata nd; int error; NDINIT(&nd, LOOKUP, LOCKSHARED | LOCKLEAF | AUDITVNODE1 | flags, pathseg, path, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = VOP_PATHCONF(nd.ni_vp, name, valuep); vput(nd.ni_vp); return (error); } /* * Return target name of a symbolic link. */ #ifndef _SYS_SYSPROTO_H_ struct readlink_args { char *path; char *buf; size_t count; }; #endif int sys_readlink(struct thread *td, struct readlink_args *uap) { return (kern_readlinkat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->buf, UIO_USERSPACE, uap->count)); } #ifndef _SYS_SYSPROTO_H_ struct readlinkat_args { int fd; char *path; char *buf; size_t bufsize; }; #endif int sys_readlinkat(struct thread *td, struct readlinkat_args *uap) { return (kern_readlinkat(td, uap->fd, uap->path, UIO_USERSPACE, uap->buf, UIO_USERSPACE, uap->bufsize)); } int kern_readlinkat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, char *buf, enum uio_seg bufseg, size_t count) { struct vnode *vp; struct nameidata nd; int error; if (count > IOSIZE_MAX) return (EINVAL); NDINIT_AT(&nd, LOOKUP, NOFOLLOW | LOCKSHARED | LOCKLEAF | AUDITVNODE1, pathseg, path, fd, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); vp = nd.ni_vp; error = kern_readlink_vp(vp, buf, bufseg, count, td); vput(vp); return (error); } /* * Helper function to readlink from a vnode */ static int kern_readlink_vp(struct vnode *vp, char *buf, enum uio_seg bufseg, size_t count, struct thread *td) { struct iovec aiov; struct uio auio; int error; ASSERT_VOP_LOCKED(vp, "kern_readlink_vp(): vp not locked"); #ifdef MAC error = mac_vnode_check_readlink(td->td_ucred, vp); if (error != 0) return (error); #endif if (vp->v_type != VLNK && (vp->v_vflag & VV_READLINK) == 0) return (EINVAL); aiov.iov_base = buf; aiov.iov_len = count; auio.uio_iov = &aiov; auio.uio_iovcnt = 1; auio.uio_offset = 0; auio.uio_rw = UIO_READ; auio.uio_segflg = bufseg; auio.uio_td = td; auio.uio_resid = count; error = VOP_READLINK(vp, &auio, td->td_ucred); td->td_retval[0] = count - auio.uio_resid; return (error); } /* * Common implementation code for chflags() and fchflags(). */ static int setfflags(struct thread *td, struct vnode *vp, u_long flags) { struct mount *mp; struct vattr vattr; int error; /* We can't support the value matching VNOVAL. */ if (flags == VNOVAL) return (EOPNOTSUPP); /* * Prevent non-root users from setting flags on devices. When * a device is reused, users can retain ownership of the device * if they are allowed to set flags and programs assume that * chown can't fail when done as root. */ if (vp->v_type == VCHR || vp->v_type == VBLK) { error = priv_check(td, PRIV_VFS_CHFLAGS_DEV); if (error != 0) return (error); } if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) return (error); VATTR_NULL(&vattr); vattr.va_flags = flags; vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); #ifdef MAC error = mac_vnode_check_setflags(td->td_ucred, vp, vattr.va_flags); if (error == 0) #endif error = VOP_SETATTR(vp, &vattr, td->td_ucred); VOP_UNLOCK(vp, 0); vn_finished_write(mp); return (error); } /* * Change flags of a file given a path name. */ #ifndef _SYS_SYSPROTO_H_ struct chflags_args { const char *path; u_long flags; }; #endif int sys_chflags(struct thread *td, struct chflags_args *uap) { return (kern_chflagsat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->flags, 0)); } #ifndef _SYS_SYSPROTO_H_ struct chflagsat_args { int fd; const char *path; u_long flags; int atflag; } #endif int sys_chflagsat(struct thread *td, struct chflagsat_args *uap) { if ((uap->atflag & ~(AT_SYMLINK_NOFOLLOW | AT_BENEATH)) != 0) return (EINVAL); return (kern_chflagsat(td, uap->fd, uap->path, UIO_USERSPACE, uap->flags, uap->atflag)); } /* * Same as chflags() but doesn't follow symlinks. */ #ifndef _SYS_SYSPROTO_H_ struct lchflags_args { const char *path; u_long flags; }; #endif int sys_lchflags(struct thread *td, struct lchflags_args *uap) { return (kern_chflagsat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->flags, AT_SYMLINK_NOFOLLOW)); } static int kern_chflagsat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, u_long flags, int atflag) { struct nameidata nd; int error, follow; AUDIT_ARG_FFLAGS(flags); follow = (atflag & AT_SYMLINK_NOFOLLOW) ? NOFOLLOW : FOLLOW; follow |= (atflag & AT_BENEATH) != 0 ? BENEATH : 0; NDINIT_ATRIGHTS(&nd, LOOKUP, follow | AUDITVNODE1, pathseg, path, fd, &cap_fchflags_rights, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = setfflags(td, nd.ni_vp, flags); vrele(nd.ni_vp); return (error); } /* * Change flags of a file given a file descriptor. */ #ifndef _SYS_SYSPROTO_H_ struct fchflags_args { int fd; u_long flags; }; #endif int sys_fchflags(struct thread *td, struct fchflags_args *uap) { struct file *fp; int error; AUDIT_ARG_FD(uap->fd); AUDIT_ARG_FFLAGS(uap->flags); error = getvnode(td, uap->fd, &cap_fchflags_rights, &fp); if (error != 0) return (error); #ifdef AUDIT vn_lock(fp->f_vnode, LK_SHARED | LK_RETRY); AUDIT_ARG_VNODE1(fp->f_vnode); VOP_UNLOCK(fp->f_vnode, 0); #endif error = setfflags(td, fp->f_vnode, uap->flags); fdrop(fp, td); return (error); } /* * Common implementation code for chmod(), lchmod() and fchmod(). */ int setfmode(struct thread *td, struct ucred *cred, struct vnode *vp, int mode) { struct mount *mp; struct vattr vattr; int error; if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) return (error); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); VATTR_NULL(&vattr); vattr.va_mode = mode & ALLPERMS; #ifdef MAC error = mac_vnode_check_setmode(cred, vp, vattr.va_mode); if (error == 0) #endif error = VOP_SETATTR(vp, &vattr, cred); VOP_UNLOCK(vp, 0); vn_finished_write(mp); return (error); } /* * Change mode of a file given path name. */ #ifndef _SYS_SYSPROTO_H_ struct chmod_args { char *path; int mode; }; #endif int sys_chmod(struct thread *td, struct chmod_args *uap) { return (kern_fchmodat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->mode, 0)); } #ifndef _SYS_SYSPROTO_H_ struct fchmodat_args { int dirfd; char *path; mode_t mode; int flag; } #endif int sys_fchmodat(struct thread *td, struct fchmodat_args *uap) { if ((uap->flag & ~(AT_SYMLINK_NOFOLLOW | AT_BENEATH)) != 0) return (EINVAL); return (kern_fchmodat(td, uap->fd, uap->path, UIO_USERSPACE, uap->mode, uap->flag)); } /* * Change mode of a file given path name (don't follow links.) */ #ifndef _SYS_SYSPROTO_H_ struct lchmod_args { char *path; int mode; }; #endif int sys_lchmod(struct thread *td, struct lchmod_args *uap) { return (kern_fchmodat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->mode, AT_SYMLINK_NOFOLLOW)); } int kern_fchmodat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, mode_t mode, int flag) { struct nameidata nd; int error, follow; AUDIT_ARG_MODE(mode); follow = (flag & AT_SYMLINK_NOFOLLOW) != 0 ? NOFOLLOW : FOLLOW; follow |= (flag & AT_BENEATH) != 0 ? BENEATH : 0; NDINIT_ATRIGHTS(&nd, LOOKUP, follow | AUDITVNODE1, pathseg, path, fd, &cap_fchmod_rights, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = setfmode(td, td->td_ucred, nd.ni_vp, mode); vrele(nd.ni_vp); return (error); } /* * Change mode of a file given a file descriptor. */ #ifndef _SYS_SYSPROTO_H_ struct fchmod_args { int fd; int mode; }; #endif int sys_fchmod(struct thread *td, struct fchmod_args *uap) { struct file *fp; int error; AUDIT_ARG_FD(uap->fd); AUDIT_ARG_MODE(uap->mode); error = fget(td, uap->fd, &cap_fchmod_rights, &fp); if (error != 0) return (error); error = fo_chmod(fp, uap->mode, td->td_ucred, td); fdrop(fp, td); return (error); } /* * Common implementation for chown(), lchown(), and fchown() */ int setfown(struct thread *td, struct ucred *cred, struct vnode *vp, uid_t uid, gid_t gid) { struct mount *mp; struct vattr vattr; int error; if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) return (error); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); VATTR_NULL(&vattr); vattr.va_uid = uid; vattr.va_gid = gid; #ifdef MAC error = mac_vnode_check_setowner(cred, vp, vattr.va_uid, vattr.va_gid); if (error == 0) #endif error = VOP_SETATTR(vp, &vattr, cred); VOP_UNLOCK(vp, 0); vn_finished_write(mp); return (error); } /* * Set ownership given a path name. */ #ifndef _SYS_SYSPROTO_H_ struct chown_args { char *path; int uid; int gid; }; #endif int sys_chown(struct thread *td, struct chown_args *uap) { return (kern_fchownat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->uid, uap->gid, 0)); } #ifndef _SYS_SYSPROTO_H_ struct fchownat_args { int fd; const char * path; uid_t uid; gid_t gid; int flag; }; #endif int sys_fchownat(struct thread *td, struct fchownat_args *uap) { if ((uap->flag & ~(AT_SYMLINK_NOFOLLOW | AT_BENEATH)) != 0) return (EINVAL); return (kern_fchownat(td, uap->fd, uap->path, UIO_USERSPACE, uap->uid, uap->gid, uap->flag)); } int kern_fchownat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int uid, int gid, int flag) { struct nameidata nd; int error, follow; AUDIT_ARG_OWNER(uid, gid); follow = (flag & AT_SYMLINK_NOFOLLOW) ? NOFOLLOW : FOLLOW; follow |= (flag & AT_BENEATH) != 0 ? BENEATH : 0; NDINIT_ATRIGHTS(&nd, LOOKUP, follow | AUDITVNODE1, pathseg, path, fd, &cap_fchown_rights, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = setfown(td, td->td_ucred, nd.ni_vp, uid, gid); vrele(nd.ni_vp); return (error); } /* * Set ownership given a path name, do not cross symlinks. */ #ifndef _SYS_SYSPROTO_H_ struct lchown_args { char *path; int uid; int gid; }; #endif int sys_lchown(struct thread *td, struct lchown_args *uap) { return (kern_fchownat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->uid, uap->gid, AT_SYMLINK_NOFOLLOW)); } /* * Set ownership given a file descriptor. */ #ifndef _SYS_SYSPROTO_H_ struct fchown_args { int fd; int uid; int gid; }; #endif int sys_fchown(struct thread *td, struct fchown_args *uap) { struct file *fp; int error; AUDIT_ARG_FD(uap->fd); AUDIT_ARG_OWNER(uap->uid, uap->gid); error = fget(td, uap->fd, &cap_fchown_rights, &fp); if (error != 0) return (error); error = fo_chown(fp, uap->uid, uap->gid, td->td_ucred, td); fdrop(fp, td); return (error); } /* * Common implementation code for utimes(), lutimes(), and futimes(). */ static int getutimes(const struct timeval *usrtvp, enum uio_seg tvpseg, struct timespec *tsp) { struct timeval tv[2]; const struct timeval *tvp; int error; if (usrtvp == NULL) { vfs_timestamp(&tsp[0]); tsp[1] = tsp[0]; } else { if (tvpseg == UIO_SYSSPACE) { tvp = usrtvp; } else { if ((error = copyin(usrtvp, tv, sizeof(tv))) != 0) return (error); tvp = tv; } if (tvp[0].tv_usec < 0 || tvp[0].tv_usec >= 1000000 || tvp[1].tv_usec < 0 || tvp[1].tv_usec >= 1000000) return (EINVAL); TIMEVAL_TO_TIMESPEC(&tvp[0], &tsp[0]); TIMEVAL_TO_TIMESPEC(&tvp[1], &tsp[1]); } return (0); } /* * Common implementation code for futimens(), utimensat(). */ #define UTIMENS_NULL 0x1 #define UTIMENS_EXIT 0x2 static int getutimens(const struct timespec *usrtsp, enum uio_seg tspseg, struct timespec *tsp, int *retflags) { struct timespec tsnow; int error; vfs_timestamp(&tsnow); *retflags = 0; if (usrtsp == NULL) { tsp[0] = tsnow; tsp[1] = tsnow; *retflags |= UTIMENS_NULL; return (0); } if (tspseg == UIO_SYSSPACE) { tsp[0] = usrtsp[0]; tsp[1] = usrtsp[1]; } else if ((error = copyin(usrtsp, tsp, sizeof(*tsp) * 2)) != 0) return (error); if (tsp[0].tv_nsec == UTIME_OMIT && tsp[1].tv_nsec == UTIME_OMIT) *retflags |= UTIMENS_EXIT; if (tsp[0].tv_nsec == UTIME_NOW && tsp[1].tv_nsec == UTIME_NOW) *retflags |= UTIMENS_NULL; if (tsp[0].tv_nsec == UTIME_OMIT) tsp[0].tv_sec = VNOVAL; else if (tsp[0].tv_nsec == UTIME_NOW) tsp[0] = tsnow; else if (tsp[0].tv_nsec < 0 || tsp[0].tv_nsec >= 1000000000L) return (EINVAL); if (tsp[1].tv_nsec == UTIME_OMIT) tsp[1].tv_sec = VNOVAL; else if (tsp[1].tv_nsec == UTIME_NOW) tsp[1] = tsnow; else if (tsp[1].tv_nsec < 0 || tsp[1].tv_nsec >= 1000000000L) return (EINVAL); return (0); } /* * Common implementation code for utimes(), lutimes(), futimes(), futimens(), * and utimensat(). */ static int setutimes(struct thread *td, struct vnode *vp, const struct timespec *ts, int numtimes, int nullflag) { struct mount *mp; struct vattr vattr; int error, setbirthtime; if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) return (error); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); setbirthtime = 0; if (numtimes < 3 && !VOP_GETATTR(vp, &vattr, td->td_ucred) && timespeccmp(&ts[1], &vattr.va_birthtime, < )) setbirthtime = 1; VATTR_NULL(&vattr); vattr.va_atime = ts[0]; vattr.va_mtime = ts[1]; if (setbirthtime) vattr.va_birthtime = ts[1]; if (numtimes > 2) vattr.va_birthtime = ts[2]; if (nullflag) vattr.va_vaflags |= VA_UTIMES_NULL; #ifdef MAC error = mac_vnode_check_setutimes(td->td_ucred, vp, vattr.va_atime, vattr.va_mtime); #endif if (error == 0) error = VOP_SETATTR(vp, &vattr, td->td_ucred); VOP_UNLOCK(vp, 0); vn_finished_write(mp); return (error); } /* * Set the access and modification times of a file. */ #ifndef _SYS_SYSPROTO_H_ struct utimes_args { char *path; struct timeval *tptr; }; #endif int sys_utimes(struct thread *td, struct utimes_args *uap) { return (kern_utimesat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->tptr, UIO_USERSPACE)); } #ifndef _SYS_SYSPROTO_H_ struct futimesat_args { int fd; const char * path; const struct timeval * times; }; #endif int sys_futimesat(struct thread *td, struct futimesat_args *uap) { return (kern_utimesat(td, uap->fd, uap->path, UIO_USERSPACE, uap->times, UIO_USERSPACE)); } int kern_utimesat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, struct timeval *tptr, enum uio_seg tptrseg) { struct nameidata nd; struct timespec ts[2]; int error; if ((error = getutimes(tptr, tptrseg, ts)) != 0) return (error); NDINIT_ATRIGHTS(&nd, LOOKUP, FOLLOW | AUDITVNODE1, pathseg, path, fd, &cap_futimes_rights, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = setutimes(td, nd.ni_vp, ts, 2, tptr == NULL); vrele(nd.ni_vp); return (error); } /* * Set the access and modification times of a file. */ #ifndef _SYS_SYSPROTO_H_ struct lutimes_args { char *path; struct timeval *tptr; }; #endif int sys_lutimes(struct thread *td, struct lutimes_args *uap) { return (kern_lutimes(td, uap->path, UIO_USERSPACE, uap->tptr, UIO_USERSPACE)); } int kern_lutimes(struct thread *td, const char *path, enum uio_seg pathseg, struct timeval *tptr, enum uio_seg tptrseg) { struct timespec ts[2]; struct nameidata nd; int error; if ((error = getutimes(tptr, tptrseg, ts)) != 0) return (error); NDINIT(&nd, LOOKUP, NOFOLLOW | AUDITVNODE1, pathseg, path, td); if ((error = namei(&nd)) != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); error = setutimes(td, nd.ni_vp, ts, 2, tptr == NULL); vrele(nd.ni_vp); return (error); } /* * Set the access and modification times of a file. */ #ifndef _SYS_SYSPROTO_H_ struct futimes_args { int fd; struct timeval *tptr; }; #endif int sys_futimes(struct thread *td, struct futimes_args *uap) { return (kern_futimes(td, uap->fd, uap->tptr, UIO_USERSPACE)); } int kern_futimes(struct thread *td, int fd, struct timeval *tptr, enum uio_seg tptrseg) { struct timespec ts[2]; struct file *fp; int error; AUDIT_ARG_FD(fd); error = getutimes(tptr, tptrseg, ts); if (error != 0) return (error); error = getvnode(td, fd, &cap_futimes_rights, &fp); if (error != 0) return (error); #ifdef AUDIT vn_lock(fp->f_vnode, LK_SHARED | LK_RETRY); AUDIT_ARG_VNODE1(fp->f_vnode); VOP_UNLOCK(fp->f_vnode, 0); #endif error = setutimes(td, fp->f_vnode, ts, 2, tptr == NULL); fdrop(fp, td); return (error); } int sys_futimens(struct thread *td, struct futimens_args *uap) { return (kern_futimens(td, uap->fd, uap->times, UIO_USERSPACE)); } int kern_futimens(struct thread *td, int fd, struct timespec *tptr, enum uio_seg tptrseg) { struct timespec ts[2]; struct file *fp; int error, flags; AUDIT_ARG_FD(fd); error = getutimens(tptr, tptrseg, ts, &flags); if (error != 0) return (error); if (flags & UTIMENS_EXIT) return (0); error = getvnode(td, fd, &cap_futimes_rights, &fp); if (error != 0) return (error); #ifdef AUDIT vn_lock(fp->f_vnode, LK_SHARED | LK_RETRY); AUDIT_ARG_VNODE1(fp->f_vnode); VOP_UNLOCK(fp->f_vnode, 0); #endif error = setutimes(td, fp->f_vnode, ts, 2, flags & UTIMENS_NULL); fdrop(fp, td); return (error); } int sys_utimensat(struct thread *td, struct utimensat_args *uap) { return (kern_utimensat(td, uap->fd, uap->path, UIO_USERSPACE, uap->times, UIO_USERSPACE, uap->flag)); } int kern_utimensat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, struct timespec *tptr, enum uio_seg tptrseg, int flag) { struct nameidata nd; struct timespec ts[2]; int error, flags; if ((flag & ~(AT_SYMLINK_NOFOLLOW | AT_BENEATH)) != 0) return (EINVAL); if ((error = getutimens(tptr, tptrseg, ts, &flags)) != 0) return (error); NDINIT_ATRIGHTS(&nd, LOOKUP, ((flag & AT_SYMLINK_NOFOLLOW) ? NOFOLLOW : FOLLOW) | ((flag & AT_BENEATH) != 0 ? BENEATH : 0) | AUDITVNODE1, pathseg, path, fd, &cap_futimes_rights, td); if ((error = namei(&nd)) != 0) return (error); /* * We are allowed to call namei() regardless of 2xUTIME_OMIT. * POSIX states: * "If both tv_nsec fields are UTIME_OMIT... EACCESS may be detected." * "Search permission is denied by a component of the path prefix." */ NDFREE(&nd, NDF_ONLY_PNBUF); if ((flags & UTIMENS_EXIT) == 0) error = setutimes(td, nd.ni_vp, ts, 2, flags & UTIMENS_NULL); vrele(nd.ni_vp); return (error); } /* * Truncate a file given its path name. */ #ifndef _SYS_SYSPROTO_H_ struct truncate_args { char *path; int pad; off_t length; }; #endif int sys_truncate(struct thread *td, struct truncate_args *uap) { return (kern_truncate(td, uap->path, UIO_USERSPACE, uap->length)); } int kern_truncate(struct thread *td, const char *path, enum uio_seg pathseg, off_t length) { struct mount *mp; struct vnode *vp; void *rl_cookie; struct vattr vattr; struct nameidata nd; int error; if (length < 0) return(EINVAL); NDINIT(&nd, LOOKUP, FOLLOW | AUDITVNODE1, pathseg, path, td); if ((error = namei(&nd)) != 0) return (error); vp = nd.ni_vp; rl_cookie = vn_rangelock_wlock(vp, 0, OFF_MAX); if ((error = vn_start_write(vp, &mp, V_WAIT | PCATCH)) != 0) { vn_rangelock_unlock(vp, rl_cookie); vrele(vp); return (error); } NDFREE(&nd, NDF_ONLY_PNBUF); vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); if (vp->v_type == VDIR) error = EISDIR; #ifdef MAC else if ((error = mac_vnode_check_write(td->td_ucred, NOCRED, vp))) { } #endif else if ((error = vn_writechk(vp)) == 0 && (error = VOP_ACCESS(vp, VWRITE, td->td_ucred, td)) == 0) { VATTR_NULL(&vattr); vattr.va_size = length; error = VOP_SETATTR(vp, &vattr, td->td_ucred); } VOP_UNLOCK(vp, 0); vn_finished_write(mp); vn_rangelock_unlock(vp, rl_cookie); vrele(vp); return (error); } #if defined(COMPAT_43) /* * Truncate a file given its path name. */ #ifndef _SYS_SYSPROTO_H_ struct otruncate_args { char *path; long length; }; #endif int otruncate(struct thread *td, struct otruncate_args *uap) { return (kern_truncate(td, uap->path, UIO_USERSPACE, uap->length)); } #endif /* COMPAT_43 */ #if defined(COMPAT_FREEBSD6) /* Versions with the pad argument */ int freebsd6_truncate(struct thread *td, struct freebsd6_truncate_args *uap) { return (kern_truncate(td, uap->path, UIO_USERSPACE, uap->length)); } int freebsd6_ftruncate(struct thread *td, struct freebsd6_ftruncate_args *uap) { return (kern_ftruncate(td, uap->fd, uap->length)); } #endif int kern_fsync(struct thread *td, int fd, bool fullsync) { struct vnode *vp; struct mount *mp; struct file *fp; int error, lock_flags; AUDIT_ARG_FD(fd); error = getvnode(td, fd, &cap_fsync_rights, &fp); if (error != 0) return (error); vp = fp->f_vnode; #if 0 if (!fullsync) /* XXXKIB: compete outstanding aio writes */; #endif error = vn_start_write(vp, &mp, V_WAIT | PCATCH); if (error != 0) goto drop; if (MNT_SHARED_WRITES(mp) || ((mp == NULL) && MNT_SHARED_WRITES(vp->v_mount))) { lock_flags = LK_SHARED; } else { lock_flags = LK_EXCLUSIVE; } vn_lock(vp, lock_flags | LK_RETRY); AUDIT_ARG_VNODE1(vp); 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 = fullsync ? VOP_FSYNC(vp, MNT_WAIT, td) : VOP_FDATASYNC(vp, td); VOP_UNLOCK(vp, 0); vn_finished_write(mp); drop: fdrop(fp, td); return (error); } /* * Sync an open file. */ #ifndef _SYS_SYSPROTO_H_ struct fsync_args { int fd; }; #endif int sys_fsync(struct thread *td, struct fsync_args *uap) { return (kern_fsync(td, uap->fd, true)); } int sys_fdatasync(struct thread *td, struct fdatasync_args *uap) { return (kern_fsync(td, uap->fd, false)); } /* * Rename files. Source and destination must either both be directories, or * both not be directories. If target is a directory, it must be empty. */ #ifndef _SYS_SYSPROTO_H_ struct rename_args { char *from; char *to; }; #endif int sys_rename(struct thread *td, struct rename_args *uap) { return (kern_renameat(td, AT_FDCWD, uap->from, AT_FDCWD, uap->to, UIO_USERSPACE)); } #ifndef _SYS_SYSPROTO_H_ struct renameat_args { int oldfd; char *old; int newfd; char *new; }; #endif int sys_renameat(struct thread *td, struct renameat_args *uap) { return (kern_renameat(td, uap->oldfd, uap->old, uap->newfd, uap->new, UIO_USERSPACE)); } int kern_renameat(struct thread *td, int oldfd, const char *old, int newfd, const char *new, enum uio_seg pathseg) { struct mount *mp = NULL; struct vnode *tvp, *fvp, *tdvp; struct nameidata fromnd, tond; int error; again: bwillwrite(); #ifdef MAC NDINIT_ATRIGHTS(&fromnd, DELETE, LOCKPARENT | LOCKLEAF | SAVESTART | AUDITVNODE1, pathseg, old, oldfd, &cap_renameat_source_rights, td); #else NDINIT_ATRIGHTS(&fromnd, DELETE, WANTPARENT | SAVESTART | AUDITVNODE1, pathseg, old, oldfd, &cap_renameat_source_rights, td); #endif if ((error = namei(&fromnd)) != 0) return (error); #ifdef MAC error = mac_vnode_check_rename_from(td->td_ucred, fromnd.ni_dvp, fromnd.ni_vp, &fromnd.ni_cnd); VOP_UNLOCK(fromnd.ni_dvp, 0); if (fromnd.ni_dvp != fromnd.ni_vp) VOP_UNLOCK(fromnd.ni_vp, 0); #endif fvp = fromnd.ni_vp; NDINIT_ATRIGHTS(&tond, RENAME, LOCKPARENT | LOCKLEAF | NOCACHE | SAVESTART | AUDITVNODE2, pathseg, new, newfd, &cap_renameat_target_rights, td); if (fromnd.ni_vp->v_type == VDIR) tond.ni_cnd.cn_flags |= WILLBEDIR; if ((error = namei(&tond)) != 0) { /* Translate error code for rename("dir1", "dir2/."). */ if (error == EISDIR && fvp->v_type == VDIR) error = EINVAL; NDFREE(&fromnd, NDF_ONLY_PNBUF); vrele(fromnd.ni_dvp); vrele(fvp); goto out1; } tdvp = tond.ni_dvp; tvp = tond.ni_vp; error = vn_start_write(fvp, &mp, V_NOWAIT); if (error != 0) { NDFREE(&fromnd, NDF_ONLY_PNBUF); NDFREE(&tond, NDF_ONLY_PNBUF); if (tvp != NULL) vput(tvp); if (tdvp == tvp) vrele(tdvp); else vput(tdvp); vrele(fromnd.ni_dvp); vrele(fvp); vrele(tond.ni_startdir); if (fromnd.ni_startdir != NULL) vrele(fromnd.ni_startdir); error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH); if (error != 0) return (error); goto again; } if (tvp != NULL) { if (fvp->v_type == VDIR && tvp->v_type != VDIR) { error = ENOTDIR; goto out; } else if (fvp->v_type != VDIR && tvp->v_type == VDIR) { error = EISDIR; goto out; } #ifdef CAPABILITIES if (newfd != AT_FDCWD && (tond.ni_resflags & NIRES_ABS) == 0) { /* * If the target already exists we require CAP_UNLINKAT * from 'newfd', when newfd was used for the lookup. */ error = cap_check(&tond.ni_filecaps.fc_rights, &cap_unlinkat_rights); if (error != 0) goto out; } #endif } if (fvp == tdvp) { error = EINVAL; goto out; } /* * If the source is the same as the destination (that is, if they * are links to the same vnode), then there is nothing to do. */ if (fvp == tvp) error = -1; #ifdef MAC else error = mac_vnode_check_rename_to(td->td_ucred, tdvp, tond.ni_vp, fromnd.ni_dvp == tdvp, &tond.ni_cnd); #endif out: if (error == 0) { error = VOP_RENAME(fromnd.ni_dvp, fromnd.ni_vp, &fromnd.ni_cnd, tond.ni_dvp, tond.ni_vp, &tond.ni_cnd); NDFREE(&fromnd, NDF_ONLY_PNBUF); NDFREE(&tond, NDF_ONLY_PNBUF); } else { NDFREE(&fromnd, NDF_ONLY_PNBUF); NDFREE(&tond, NDF_ONLY_PNBUF); if (tvp != NULL) vput(tvp); if (tdvp == tvp) vrele(tdvp); else vput(tdvp); vrele(fromnd.ni_dvp); vrele(fvp); } vrele(tond.ni_startdir); vn_finished_write(mp); out1: if (fromnd.ni_startdir) vrele(fromnd.ni_startdir); if (error == -1) return (0); return (error); } /* * Make a directory file. */ #ifndef _SYS_SYSPROTO_H_ struct mkdir_args { char *path; int mode; }; #endif int sys_mkdir(struct thread *td, struct mkdir_args *uap) { return (kern_mkdirat(td, AT_FDCWD, uap->path, UIO_USERSPACE, uap->mode)); } #ifndef _SYS_SYSPROTO_H_ struct mkdirat_args { int fd; char *path; mode_t mode; }; #endif int sys_mkdirat(struct thread *td, struct mkdirat_args *uap) { return (kern_mkdirat(td, uap->fd, uap->path, UIO_USERSPACE, uap->mode)); } int kern_mkdirat(struct thread *td, int fd, const char *path, enum uio_seg segflg, int mode) { struct mount *mp; struct vnode *vp; struct vattr vattr; struct nameidata nd; int error; AUDIT_ARG_MODE(mode); restart: bwillwrite(); NDINIT_ATRIGHTS(&nd, CREATE, LOCKPARENT | SAVENAME | AUDITVNODE1 | NOCACHE, segflg, path, fd, &cap_mkdirat_rights, td); nd.ni_cnd.cn_flags |= WILLBEDIR; if ((error = namei(&nd)) != 0) return (error); vp = nd.ni_vp; if (vp != NULL) { NDFREE(&nd, NDF_ONLY_PNBUF); /* * XXX namei called with LOCKPARENT but not LOCKLEAF has * the strange behaviour of leaving the vnode unlocked * if the target is the same vnode as the parent. */ if (vp == nd.ni_dvp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); vrele(vp); return (EEXIST); } if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0) return (error); goto restart; } VATTR_NULL(&vattr); vattr.va_type = VDIR; vattr.va_mode = (mode & ACCESSPERMS) &~ td->td_proc->p_fd->fd_cmask; #ifdef MAC error = mac_vnode_check_create(td->td_ucred, nd.ni_dvp, &nd.ni_cnd, &vattr); if (error != 0) goto out; #endif error = VOP_MKDIR(nd.ni_dvp, &nd.ni_vp, &nd.ni_cnd, &vattr); #ifdef MAC out: #endif NDFREE(&nd, NDF_ONLY_PNBUF); vput(nd.ni_dvp); if (error == 0) vput(nd.ni_vp); vn_finished_write(mp); return (error); } /* * Remove a directory file. */ #ifndef _SYS_SYSPROTO_H_ struct rmdir_args { char *path; }; #endif int sys_rmdir(struct thread *td, struct rmdir_args *uap) { - return (kern_rmdirat(td, AT_FDCWD, uap->path, UIO_USERSPACE, 0)); + return (kern_frmdirat(td, AT_FDCWD, uap->path, FD_NONE, UIO_USERSPACE, + 0)); } int -kern_rmdirat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, - int flag) +kern_frmdirat(struct thread *td, int dfd, const char *path, int fd, + enum uio_seg pathseg, int flag) { struct mount *mp; struct vnode *vp; + struct file *fp; struct nameidata nd; + cap_rights_t rights; int error; + fp = NULL; + if (fd != FD_NONE) { + error = getvnode(td, fd, cap_rights_init(&rights, CAP_LOOKUP), + &fp); + if (error != 0) + return (error); + } + restart: bwillwrite(); NDINIT_ATRIGHTS(&nd, DELETE, LOCKPARENT | LOCKLEAF | AUDITVNODE1 | ((flag & AT_BENEATH) != 0 ? BENEATH : 0), - pathseg, path, fd, &cap_unlinkat_rights, td); + pathseg, path, dfd, &cap_unlinkat_rights, td); if ((error = namei(&nd)) != 0) - return (error); + goto fdout; vp = nd.ni_vp; if (vp->v_type != VDIR) { error = ENOTDIR; goto out; } /* * No rmdir "." please. */ if (nd.ni_dvp == vp) { error = EINVAL; goto out; } /* * The root of a mounted filesystem cannot be deleted. */ if (vp->v_vflag & VV_ROOT) { error = EBUSY; goto out; } + + if (fp != NULL && fp->f_vnode != vp) { + if ((fp->f_vnode->v_iflag & VI_DOOMED) != 0) + error = EBADF; + else + error = EDEADLK; + goto out; + } + #ifdef MAC error = mac_vnode_check_unlink(td->td_ucred, nd.ni_dvp, vp, &nd.ni_cnd); if (error != 0) goto out; #endif if (vn_start_write(nd.ni_dvp, &mp, V_NOWAIT) != 0) { NDFREE(&nd, NDF_ONLY_PNBUF); vput(vp); if (nd.ni_dvp == vp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); if ((error = vn_start_write(NULL, &mp, V_XSLEEP | PCATCH)) != 0) - return (error); + goto fdout; goto restart; } vfs_notify_upper(vp, VFS_NOTIFY_UPPER_UNLINK); error = VOP_RMDIR(nd.ni_dvp, nd.ni_vp, &nd.ni_cnd); vn_finished_write(mp); out: NDFREE(&nd, NDF_ONLY_PNBUF); vput(vp); if (nd.ni_dvp == vp) vrele(nd.ni_dvp); else vput(nd.ni_dvp); +fdout: + if (fp != NULL) + fdrop(fp, td); return (error); } #if defined(COMPAT_43) || defined(COMPAT_FREEBSD11) int freebsd11_kern_getdirentries(struct thread *td, int fd, char *ubuf, u_int count, long *basep, void (*func)(struct freebsd11_dirent *)) { struct freebsd11_dirent dstdp; struct dirent *dp, *edp; char *dirbuf; off_t base; ssize_t resid, ucount; int error; /* XXX arbitrary sanity limit on `count'. */ count = min(count, 64 * 1024); dirbuf = malloc(count, M_TEMP, M_WAITOK); error = kern_getdirentries(td, fd, dirbuf, count, &base, &resid, UIO_SYSSPACE); if (error != 0) goto done; if (basep != NULL) *basep = base; ucount = 0; for (dp = (struct dirent *)dirbuf, edp = (struct dirent *)&dirbuf[count - resid]; ucount < count && dp < edp; ) { if (dp->d_reclen == 0) break; MPASS(dp->d_reclen >= _GENERIC_DIRLEN(0)); if (dp->d_namlen >= sizeof(dstdp.d_name)) continue; dstdp.d_type = dp->d_type; dstdp.d_namlen = dp->d_namlen; dstdp.d_fileno = dp->d_fileno; /* truncate */ if (dstdp.d_fileno != dp->d_fileno) { switch (ino64_trunc_error) { default: case 0: break; case 1: error = EOVERFLOW; goto done; case 2: dstdp.d_fileno = UINT32_MAX; break; } } dstdp.d_reclen = sizeof(dstdp) - sizeof(dstdp.d_name) + ((dp->d_namlen + 1 + 3) &~ 3); bcopy(dp->d_name, dstdp.d_name, dstdp.d_namlen); bzero(dstdp.d_name + dstdp.d_namlen, dstdp.d_reclen - offsetof(struct freebsd11_dirent, d_name) - dstdp.d_namlen); MPASS(dstdp.d_reclen <= dp->d_reclen); MPASS(ucount + dstdp.d_reclen <= count); if (func != NULL) func(&dstdp); error = copyout(&dstdp, ubuf + ucount, dstdp.d_reclen); if (error != 0) break; dp = (struct dirent *)((char *)dp + dp->d_reclen); ucount += dstdp.d_reclen; } done: free(dirbuf, M_TEMP); if (error == 0) td->td_retval[0] = ucount; return (error); } #endif /* COMPAT */ #ifdef COMPAT_43 static void ogetdirentries_cvt(struct freebsd11_dirent *dp) { #if (BYTE_ORDER == LITTLE_ENDIAN) /* * The expected low byte of dp->d_namlen is our dp->d_type. * The high MBZ byte of dp->d_namlen is our dp->d_namlen. */ dp->d_type = dp->d_namlen; dp->d_namlen = 0; #else /* * The dp->d_type is the high byte of the expected dp->d_namlen, * so must be zero'ed. */ dp->d_type = 0; #endif } /* * Read a block of directory entries in a filesystem independent format. */ #ifndef _SYS_SYSPROTO_H_ struct ogetdirentries_args { int fd; char *buf; u_int count; long *basep; }; #endif int ogetdirentries(struct thread *td, struct ogetdirentries_args *uap) { long loff; int error; error = kern_ogetdirentries(td, uap, &loff); if (error == 0) error = copyout(&loff, uap->basep, sizeof(long)); return (error); } int kern_ogetdirentries(struct thread *td, struct ogetdirentries_args *uap, long *ploff) { long base; int error; /* XXX arbitrary sanity limit on `count'. */ if (uap->count > 64 * 1024) return (EINVAL); error = freebsd11_kern_getdirentries(td, uap->fd, uap->buf, uap->count, &base, ogetdirentries_cvt); if (error == 0 && uap->basep != NULL) error = copyout(&base, uap->basep, sizeof(long)); return (error); } #endif /* COMPAT_43 */ #if defined(COMPAT_FREEBSD11) #ifndef _SYS_SYSPROTO_H_ struct freebsd11_getdirentries_args { int fd; char *buf; u_int count; long *basep; }; #endif int freebsd11_getdirentries(struct thread *td, struct freebsd11_getdirentries_args *uap) { long base; int error; error = freebsd11_kern_getdirentries(td, uap->fd, uap->buf, uap->count, &base, NULL); if (error == 0 && uap->basep != NULL) error = copyout(&base, uap->basep, sizeof(long)); return (error); } int freebsd11_getdents(struct thread *td, struct freebsd11_getdents_args *uap) { struct freebsd11_getdirentries_args ap; ap.fd = uap->fd; ap.buf = uap->buf; ap.count = uap->count; ap.basep = NULL; return (freebsd11_getdirentries(td, &ap)); } #endif /* COMPAT_FREEBSD11 */ /* * Read a block of directory entries in a filesystem independent format. */ int sys_getdirentries(struct thread *td, struct getdirentries_args *uap) { off_t base; int error; error = kern_getdirentries(td, uap->fd, uap->buf, uap->count, &base, NULL, UIO_USERSPACE); if (error != 0) return (error); if (uap->basep != NULL) error = copyout(&base, uap->basep, sizeof(off_t)); return (error); } int kern_getdirentries(struct thread *td, int fd, char *buf, size_t count, off_t *basep, ssize_t *residp, enum uio_seg bufseg) { struct vnode *vp; struct file *fp; struct uio auio; struct iovec aiov; off_t loff; int error, eofflag; off_t foffset; AUDIT_ARG_FD(fd); if (count > IOSIZE_MAX) return (EINVAL); auio.uio_resid = count; error = getvnode(td, fd, &cap_read_rights, &fp); if (error != 0) return (error); if ((fp->f_flag & FREAD) == 0) { fdrop(fp, td); return (EBADF); } vp = fp->f_vnode; foffset = foffset_lock(fp, 0); unionread: if (vp->v_type != VDIR) { error = EINVAL; goto fail; } aiov.iov_base = buf; aiov.iov_len = count; auio.uio_iov = &aiov; auio.uio_iovcnt = 1; auio.uio_rw = UIO_READ; auio.uio_segflg = bufseg; auio.uio_td = td; vn_lock(vp, LK_SHARED | LK_RETRY); AUDIT_ARG_VNODE1(vp); loff = auio.uio_offset = foffset; #ifdef MAC error = mac_vnode_check_readdir(td->td_ucred, vp); if (error == 0) #endif error = VOP_READDIR(vp, &auio, fp->f_cred, &eofflag, NULL, NULL); foffset = auio.uio_offset; if (error != 0) { VOP_UNLOCK(vp, 0); goto fail; } if (count == auio.uio_resid && (vp->v_vflag & VV_ROOT) && (vp->v_mount->mnt_flag & MNT_UNION)) { struct vnode *tvp = vp; vp = vp->v_mount->mnt_vnodecovered; VREF(vp); fp->f_vnode = vp; fp->f_data = vp; foffset = 0; vput(tvp); goto unionread; } VOP_UNLOCK(vp, 0); *basep = loff; if (residp != NULL) *residp = auio.uio_resid; td->td_retval[0] = count - auio.uio_resid; fail: foffset_unlock(fp, foffset, 0); fdrop(fp, td); return (error); } /* * Set the mode mask for creation of filesystem nodes. */ #ifndef _SYS_SYSPROTO_H_ struct umask_args { int newmask; }; #endif int sys_umask(struct thread *td, struct umask_args *uap) { struct filedesc *fdp; fdp = td->td_proc->p_fd; FILEDESC_XLOCK(fdp); td->td_retval[0] = fdp->fd_cmask; fdp->fd_cmask = uap->newmask & ALLPERMS; FILEDESC_XUNLOCK(fdp); return (0); } /* * Void all references to file by ripping underlying filesystem away from * vnode. */ #ifndef _SYS_SYSPROTO_H_ struct revoke_args { char *path; }; #endif int sys_revoke(struct thread *td, struct revoke_args *uap) { struct vnode *vp; struct vattr vattr; struct nameidata nd; int error; NDINIT(&nd, LOOKUP, FOLLOW | LOCKLEAF | AUDITVNODE1, UIO_USERSPACE, uap->path, td); if ((error = namei(&nd)) != 0) return (error); vp = nd.ni_vp; NDFREE(&nd, NDF_ONLY_PNBUF); if (vp->v_type != VCHR || vp->v_rdev == NULL) { error = EINVAL; goto out; } #ifdef MAC error = mac_vnode_check_revoke(td->td_ucred, vp); if (error != 0) goto out; #endif error = VOP_GETATTR(vp, &vattr, td->td_ucred); if (error != 0) goto out; if (td->td_ucred->cr_uid != vattr.va_uid) { error = priv_check(td, PRIV_VFS_ADMIN); if (error != 0) goto out; } if (vcount(vp) > 1) VOP_REVOKE(vp, REVOKEALL); out: vput(vp); return (error); } /* * Convert a user file descriptor to a kernel file entry and check that, if it * is a capability, the correct rights are present. A reference on the file * entry is held upon returning. */ int getvnode(struct thread *td, int fd, cap_rights_t *rightsp, struct file **fpp) { struct file *fp; int error; error = fget_unlocked(td->td_proc->p_fd, fd, rightsp, &fp, NULL); if (error != 0) return (error); /* * The file could be not of the vnode type, or it may be not * yet fully initialized, in which case the f_vnode pointer * may be set, but f_ops is still badfileops. E.g., * devfs_open() transiently create such situation to * facilitate csw d_fdopen(). * * Dupfdopen() handling in kern_openat() installs the * half-baked file into the process descriptor table, allowing * other thread to dereference it. Guard against the race by * checking f_ops. */ if (fp->f_vnode == NULL || fp->f_ops == &badfileops) { fdrop(fp, td); return (EINVAL); } *fpp = fp; return (0); } /* * Get an (NFS) file handle. */ #ifndef _SYS_SYSPROTO_H_ struct lgetfh_args { char *fname; fhandle_t *fhp; }; #endif int sys_lgetfh(struct thread *td, struct lgetfh_args *uap) { return (kern_getfhat(td, AT_SYMLINK_NOFOLLOW, AT_FDCWD, uap->fname, UIO_USERSPACE, uap->fhp)); } #ifndef _SYS_SYSPROTO_H_ struct getfh_args { char *fname; fhandle_t *fhp; }; #endif int sys_getfh(struct thread *td, struct getfh_args *uap) { return (kern_getfhat(td, 0, AT_FDCWD, uap->fname, UIO_USERSPACE, uap->fhp)); } /* * syscall for the rpc.lockd to use to translate an open descriptor into * a NFS file handle. * * warning: do not remove the priv_check() call or this becomes one giant * security hole. */ #ifndef _SYS_SYSPROTO_H_ struct getfhat_args { int fd; char *path; fhandle_t *fhp; int flags; }; #endif int sys_getfhat(struct thread *td, struct getfhat_args *uap) { if ((uap->flags & ~(AT_SYMLINK_NOFOLLOW | AT_BENEATH)) != 0) return (EINVAL); return (kern_getfhat(td, uap->flags, uap->fd, uap->path, UIO_USERSPACE, uap->fhp)); } static int kern_getfhat(struct thread *td, int flags, int fd, const char *path, enum uio_seg pathseg, fhandle_t *fhp) { struct nameidata nd; fhandle_t fh; struct vnode *vp; int error; error = priv_check(td, PRIV_VFS_GETFH); if (error != 0) return (error); NDINIT_AT(&nd, LOOKUP, ((flags & AT_SYMLINK_NOFOLLOW) != 0 ? NOFOLLOW : FOLLOW) | ((flags & AT_BENEATH) != 0 ? BENEATH : 0) | LOCKLEAF | AUDITVNODE1, pathseg, path, fd, td); error = namei(&nd); if (error != 0) return (error); NDFREE(&nd, NDF_ONLY_PNBUF); vp = nd.ni_vp; bzero(&fh, sizeof(fh)); fh.fh_fsid = vp->v_mount->mnt_stat.f_fsid; error = VOP_VPTOFH(vp, &fh.fh_fid); vput(vp); if (error == 0) error = copyout(&fh, fhp, sizeof (fh)); return (error); } #ifndef _SYS_SYSPROTO_H_ struct fhlink_args { fhandle_t *fhp; const char *to; }; #endif int sys_fhlink(struct thread *td, struct fhlink_args *uap) { return (kern_fhlinkat(td, AT_FDCWD, uap->to, UIO_USERSPACE, uap->fhp)); } #ifndef _SYS_SYSPROTO_H_ struct fhlinkat_args { fhandle_t *fhp; int tofd; const char *to; }; #endif int sys_fhlinkat(struct thread *td, struct fhlinkat_args *uap) { return (kern_fhlinkat(td, uap->tofd, uap->to, UIO_USERSPACE, uap->fhp)); } static int kern_fhlinkat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, fhandle_t *fhp) { fhandle_t fh; struct mount *mp; struct vnode *vp; int error; error = priv_check(td, PRIV_VFS_GETFH); if (error != 0) return (error); error = copyin(fhp, &fh, sizeof(fh)); if (error != 0) return (error); do { bwillwrite(); if ((mp = vfs_busyfs(&fh.fh_fsid)) == NULL) return (ESTALE); error = VFS_FHTOVP(mp, &fh.fh_fid, LK_SHARED, &vp); vfs_unbusy(mp); if (error != 0) return (error); VOP_UNLOCK(vp, 0); } while ((error = kern_linkat_vp(td, vp, fd, path, pathseg)) == EAGAIN); return (error); } #ifndef _SYS_SYSPROTO_H_ struct fhreadlink_args { fhandle_t *fhp; char *buf; size_t bufsize; }; #endif int sys_fhreadlink(struct thread *td, struct fhreadlink_args *uap) { fhandle_t fh; struct mount *mp; struct vnode *vp; int error; error = priv_check(td, PRIV_VFS_GETFH); if (error != 0) return (error); if (uap->bufsize > IOSIZE_MAX) return (EINVAL); error = copyin(uap->fhp, &fh, sizeof(fh)); if (error != 0) return (error); if ((mp = vfs_busyfs(&fh.fh_fsid)) == NULL) return (ESTALE); error = VFS_FHTOVP(mp, &fh.fh_fid, LK_SHARED, &vp); vfs_unbusy(mp); if (error != 0) return (error); error = kern_readlink_vp(vp, uap->buf, UIO_USERSPACE, uap->bufsize, td); vput(vp); return (error); } /* * syscall for the rpc.lockd to use to translate a NFS file handle into an * open descriptor. * * warning: do not remove the priv_check() call or this becomes one giant * security hole. */ #ifndef _SYS_SYSPROTO_H_ struct fhopen_args { const struct fhandle *u_fhp; int flags; }; #endif int sys_fhopen(struct thread *td, struct fhopen_args *uap) { struct mount *mp; struct vnode *vp; struct fhandle fhp; struct file *fp; int fmode, error; int indx; error = priv_check(td, PRIV_VFS_FHOPEN); if (error != 0) return (error); indx = -1; fmode = FFLAGS(uap->flags); /* why not allow a non-read/write open for our lockd? */ if (((fmode & (FREAD | FWRITE)) == 0) || (fmode & O_CREAT)) return (EINVAL); error = copyin(uap->u_fhp, &fhp, sizeof(fhp)); if (error != 0) return(error); /* find the mount point */ mp = vfs_busyfs(&fhp.fh_fsid); if (mp == NULL) return (ESTALE); /* now give me my vnode, it gets returned to me locked */ error = VFS_FHTOVP(mp, &fhp.fh_fid, LK_EXCLUSIVE, &vp); vfs_unbusy(mp); if (error != 0) return (error); error = falloc_noinstall(td, &fp); if (error != 0) { vput(vp); return (error); } /* * An extra reference on `fp' has been held for us by * falloc_noinstall(). */ #ifdef INVARIANTS td->td_dupfd = -1; #endif error = vn_open_vnode(vp, fmode, td->td_ucred, td, fp); if (error != 0) { KASSERT(fp->f_ops == &badfileops, ("VOP_OPEN in fhopen() set f_ops")); KASSERT(td->td_dupfd < 0, ("fhopen() encountered fdopen()")); vput(vp); goto bad; } #ifdef INVARIANTS td->td_dupfd = 0; #endif fp->f_vnode = vp; fp->f_seqcount = 1; finit(fp, (fmode & FMASK) | (fp->f_flag & FHASLOCK), DTYPE_VNODE, vp, &vnops); VOP_UNLOCK(vp, 0); if ((fmode & O_TRUNC) != 0) { error = fo_truncate(fp, 0, td->td_ucred, td); if (error != 0) goto bad; } error = finstall(td, fp, &indx, fmode, NULL); bad: fdrop(fp, td); td->td_retval[0] = indx; return (error); } /* * Stat an (NFS) file handle. */ #ifndef _SYS_SYSPROTO_H_ struct fhstat_args { struct fhandle *u_fhp; struct stat *sb; }; #endif int sys_fhstat(struct thread *td, struct fhstat_args *uap) { struct stat sb; struct fhandle fh; int error; error = copyin(uap->u_fhp, &fh, sizeof(fh)); if (error != 0) return (error); error = kern_fhstat(td, fh, &sb); if (error == 0) error = copyout(&sb, uap->sb, sizeof(sb)); return (error); } int kern_fhstat(struct thread *td, struct fhandle fh, struct stat *sb) { struct mount *mp; struct vnode *vp; int error; error = priv_check(td, PRIV_VFS_FHSTAT); if (error != 0) return (error); if ((mp = vfs_busyfs(&fh.fh_fsid)) == NULL) return (ESTALE); error = VFS_FHTOVP(mp, &fh.fh_fid, LK_EXCLUSIVE, &vp); vfs_unbusy(mp); if (error != 0) return (error); error = vn_stat(vp, sb, td->td_ucred, NOCRED, td); vput(vp); return (error); } /* * Implement fstatfs() for (NFS) file handles. */ #ifndef _SYS_SYSPROTO_H_ struct fhstatfs_args { struct fhandle *u_fhp; struct statfs *buf; }; #endif int sys_fhstatfs(struct thread *td, struct fhstatfs_args *uap) { struct statfs *sfp; fhandle_t fh; int error; error = copyin(uap->u_fhp, &fh, sizeof(fhandle_t)); if (error != 0) return (error); sfp = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK); error = kern_fhstatfs(td, fh, sfp); if (error == 0) error = copyout(sfp, uap->buf, sizeof(*sfp)); free(sfp, M_STATFS); return (error); } int kern_fhstatfs(struct thread *td, fhandle_t fh, struct statfs *buf) { struct statfs *sp; struct mount *mp; struct vnode *vp; int error; error = priv_check(td, PRIV_VFS_FHSTATFS); if (error != 0) return (error); if ((mp = vfs_busyfs(&fh.fh_fsid)) == NULL) return (ESTALE); error = VFS_FHTOVP(mp, &fh.fh_fid, LK_EXCLUSIVE, &vp); if (error != 0) { vfs_unbusy(mp); return (error); } vput(vp); error = prison_canseemount(td->td_ucred, mp); if (error != 0) goto out; #ifdef MAC error = mac_mount_check_stat(td->td_ucred, mp); if (error != 0) goto out; #endif /* * Set these in case the underlying filesystem fails to do so. */ sp = &mp->mnt_stat; sp->f_version = STATFS_VERSION; sp->f_namemax = NAME_MAX; sp->f_flags = mp->mnt_flag & MNT_VISFLAGMASK; error = VFS_STATFS(mp, sp); if (error == 0) *buf = *sp; out: vfs_unbusy(mp); return (error); } int kern_posix_fallocate(struct thread *td, int fd, off_t offset, off_t len) { struct file *fp; struct mount *mp; struct vnode *vp; off_t olen, ooffset; int error; #ifdef AUDIT int audited_vnode1 = 0; #endif AUDIT_ARG_FD(fd); if (offset < 0 || len <= 0) return (EINVAL); /* Check for wrap. */ if (offset > OFF_MAX - len) return (EFBIG); AUDIT_ARG_FD(fd); error = fget(td, fd, &cap_pwrite_rights, &fp); if (error != 0) return (error); AUDIT_ARG_FILE(td->td_proc, fp); if ((fp->f_ops->fo_flags & DFLAG_SEEKABLE) == 0) { error = ESPIPE; goto out; } if ((fp->f_flag & FWRITE) == 0) { error = EBADF; goto out; } if (fp->f_type != DTYPE_VNODE) { error = ENODEV; goto out; } vp = fp->f_vnode; if (vp->v_type != VREG) { error = ENODEV; goto out; } /* Allocating blocks may take a long time, so iterate. */ for (;;) { olen = len; ooffset = offset; bwillwrite(); mp = NULL; error = vn_start_write(vp, &mp, V_WAIT | PCATCH); if (error != 0) break; error = vn_lock(vp, LK_EXCLUSIVE); if (error != 0) { vn_finished_write(mp); break; } #ifdef AUDIT if (!audited_vnode1) { AUDIT_ARG_VNODE1(vp); audited_vnode1 = 1; } #endif #ifdef MAC error = mac_vnode_check_write(td->td_ucred, fp->f_cred, vp); if (error == 0) #endif error = VOP_ALLOCATE(vp, &offset, &len); VOP_UNLOCK(vp, 0); vn_finished_write(mp); if (olen + ooffset != offset + len) { panic("offset + len changed from %jx/%jx to %jx/%jx", ooffset, olen, offset, len); } if (error != 0 || len == 0) break; KASSERT(olen > len, ("Iteration did not make progress?")); maybe_yield(); } out: fdrop(fp, td); return (error); } int sys_posix_fallocate(struct thread *td, struct posix_fallocate_args *uap) { int error; error = kern_posix_fallocate(td, uap->fd, uap->offset, uap->len); return (kern_posix_error(td, error)); } /* * Unlike madvise(2), we do not make a best effort to remember every * possible caching hint. Instead, we remember the last setting with * the exception that we will allow POSIX_FADV_NORMAL to adjust the * region of any current setting. */ int kern_posix_fadvise(struct thread *td, int fd, off_t offset, off_t len, int advice) { struct fadvise_info *fa, *new; struct file *fp; struct vnode *vp; off_t end; int error; if (offset < 0 || len < 0 || offset > OFF_MAX - len) return (EINVAL); AUDIT_ARG_VALUE(advice); switch (advice) { case POSIX_FADV_SEQUENTIAL: case POSIX_FADV_RANDOM: case POSIX_FADV_NOREUSE: new = malloc(sizeof(*fa), M_FADVISE, M_WAITOK); break; case POSIX_FADV_NORMAL: case POSIX_FADV_WILLNEED: case POSIX_FADV_DONTNEED: new = NULL; break; default: return (EINVAL); } /* XXX: CAP_POSIX_FADVISE? */ AUDIT_ARG_FD(fd); error = fget(td, fd, &cap_no_rights, &fp); if (error != 0) goto out; AUDIT_ARG_FILE(td->td_proc, fp); if ((fp->f_ops->fo_flags & DFLAG_SEEKABLE) == 0) { error = ESPIPE; goto out; } if (fp->f_type != DTYPE_VNODE) { error = ENODEV; goto out; } vp = fp->f_vnode; if (vp->v_type != VREG) { error = ENODEV; goto out; } if (len == 0) end = OFF_MAX; else end = offset + len - 1; switch (advice) { case POSIX_FADV_SEQUENTIAL: case POSIX_FADV_RANDOM: case POSIX_FADV_NOREUSE: /* * Try to merge any existing non-standard region with * this new region if possible, otherwise create a new * non-standard region for this request. */ mtx_pool_lock(mtxpool_sleep, fp); fa = fp->f_advice; if (fa != NULL && fa->fa_advice == advice && ((fa->fa_start <= end && fa->fa_end >= offset) || (end != OFF_MAX && fa->fa_start == end + 1) || (fa->fa_end != OFF_MAX && fa->fa_end + 1 == offset))) { if (offset < fa->fa_start) fa->fa_start = offset; if (end > fa->fa_end) fa->fa_end = end; } else { new->fa_advice = advice; new->fa_start = offset; new->fa_end = end; fp->f_advice = new; new = fa; } mtx_pool_unlock(mtxpool_sleep, fp); break; case POSIX_FADV_NORMAL: /* * If a the "normal" region overlaps with an existing * non-standard region, trim or remove the * non-standard region. */ mtx_pool_lock(mtxpool_sleep, fp); fa = fp->f_advice; if (fa != NULL) { if (offset <= fa->fa_start && end >= fa->fa_end) { new = fa; fp->f_advice = NULL; } else if (offset <= fa->fa_start && end >= fa->fa_start) fa->fa_start = end + 1; else if (offset <= fa->fa_end && end >= fa->fa_end) fa->fa_end = offset - 1; else if (offset >= fa->fa_start && end <= fa->fa_end) { /* * If the "normal" region is a middle * portion of the existing * non-standard region, just remove * the whole thing rather than picking * one side or the other to * preserve. */ new = fa; fp->f_advice = NULL; } } mtx_pool_unlock(mtxpool_sleep, fp); break; case POSIX_FADV_WILLNEED: case POSIX_FADV_DONTNEED: error = VOP_ADVISE(vp, offset, end, advice); break; } out: if (fp != NULL) fdrop(fp, td); free(new, M_FADVISE); return (error); } int sys_posix_fadvise(struct thread *td, struct posix_fadvise_args *uap) { int error; error = kern_posix_fadvise(td, uap->fd, uap->offset, uap->len, uap->advice); return (kern_posix_error(td, error)); } Index: head/sys/sys/fcntl.h =================================================================== --- head/sys/sys/fcntl.h (revision 345981) +++ head/sys/sys/fcntl.h (revision 345982) @@ -1,337 +1,347 @@ /*- * SPDX-License-Identifier: BSD-3-Clause * * Copyright (c) 1983, 1990, 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. * 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. * * @(#)fcntl.h 8.3 (Berkeley) 1/21/94 * $FreeBSD$ */ #ifndef _SYS_FCNTL_H_ #define _SYS_FCNTL_H_ /* * This file includes the definitions for open and fcntl * described by POSIX for ; it also includes * related kernel definitions. */ #include #include #ifndef _MODE_T_DECLARED typedef __mode_t mode_t; #define _MODE_T_DECLARED #endif #ifndef _OFF_T_DECLARED typedef __off_t off_t; #define _OFF_T_DECLARED #endif #ifndef _PID_T_DECLARED typedef __pid_t pid_t; #define _PID_T_DECLARED #endif /* * File status flags: these are used by open(2), fcntl(2). * They are also used (indirectly) in the kernel file structure f_flags, * which is a superset of the open/fcntl flags. Open flags and f_flags * are inter-convertible using OFLAGS(fflags) and FFLAGS(oflags). * Open/fcntl flags begin with O_; kernel-internal flags begin with F. */ /* open-only flags */ #define O_RDONLY 0x0000 /* open for reading only */ #define O_WRONLY 0x0001 /* open for writing only */ #define O_RDWR 0x0002 /* open for reading and writing */ #define O_ACCMODE 0x0003 /* mask for above modes */ /* * Kernel encoding of open mode; separate read and write bits that are * independently testable: 1 greater than the above. * * XXX * FREAD and FWRITE are excluded from the #ifdef _KERNEL so that TIOCFLUSH, * which was documented to use FREAD/FWRITE, continues to work. */ #if __BSD_VISIBLE #define FREAD 0x0001 #define FWRITE 0x0002 #endif #define O_NONBLOCK 0x0004 /* no delay */ #define O_APPEND 0x0008 /* set append mode */ #if __BSD_VISIBLE #define O_SHLOCK 0x0010 /* open with shared file lock */ #define O_EXLOCK 0x0020 /* open with exclusive file lock */ #define O_ASYNC 0x0040 /* signal pgrp when data ready */ #define O_FSYNC 0x0080 /* synchronous writes */ #endif #define O_SYNC 0x0080 /* POSIX synonym for O_FSYNC */ #if __POSIX_VISIBLE >= 200809 #define O_NOFOLLOW 0x0100 /* don't follow symlinks */ #endif #define O_CREAT 0x0200 /* create if nonexistent */ #define O_TRUNC 0x0400 /* truncate to zero length */ #define O_EXCL 0x0800 /* error if already exists */ #ifdef _KERNEL #define FHASLOCK 0x4000 /* descriptor holds advisory lock */ #endif /* Defined by POSIX 1003.1; BSD default, but must be distinct from O_RDONLY. */ #define O_NOCTTY 0x8000 /* don't assign controlling terminal */ #if __BSD_VISIBLE /* Attempt to bypass buffer cache */ #define O_DIRECT 0x00010000 #endif #if __POSIX_VISIBLE >= 200809 #define O_DIRECTORY 0x00020000 /* Fail if not directory */ #define O_EXEC 0x00040000 /* Open for execute only */ #endif #ifdef _KERNEL #define FEXEC O_EXEC #endif #if __POSIX_VISIBLE >= 200809 /* Defined by POSIX 1003.1-2008; BSD default, but reserve for future use. */ #define O_TTY_INIT 0x00080000 /* Restore default termios attributes */ #define O_CLOEXEC 0x00100000 #endif #if __BSD_VISIBLE #define O_VERIFY 0x00200000 /* open only after verification */ #define O_BENEATH 0x00400000 /* Fail if not under cwd */ #endif /* * XXX missing O_DSYNC, O_RSYNC. */ #ifdef _KERNEL /* Only for devfs d_close() flags. */ #define FLASTCLOSE O_DIRECTORY #define FREVOKE O_VERIFY /* Only for fo_close() from half-succeeded open */ #define FOPENFAILED O_TTY_INIT /* convert from open() flags to/from fflags; convert O_RD/WR to FREAD/FWRITE */ #define FFLAGS(oflags) ((oflags) & O_EXEC ? (oflags) : (oflags) + 1) #define OFLAGS(fflags) ((fflags) & O_EXEC ? (fflags) : (fflags) - 1) /* bits to save after open */ #define FMASK (FREAD|FWRITE|FAPPEND|FASYNC|FFSYNC|FNONBLOCK|O_DIRECT|FEXEC) /* bits settable by fcntl(F_SETFL, ...) */ #define FCNTLFLAGS (FAPPEND|FASYNC|FFSYNC|FNONBLOCK|FRDAHEAD|O_DIRECT) #if defined(COMPAT_FREEBSD7) || defined(COMPAT_FREEBSD6) || \ defined(COMPAT_FREEBSD5) || defined(COMPAT_FREEBSD4) /* * Set by shm_open(3) in older libc's to get automatic MAP_ASYNC * behavior for POSIX shared memory objects (which are otherwise * implemented as plain files). */ #define FPOSIXSHM O_NOFOLLOW #undef FCNTLFLAGS #define FCNTLFLAGS (FAPPEND|FASYNC|FFSYNC|FNONBLOCK|FPOSIXSHM|FRDAHEAD| \ O_DIRECT) #endif #endif /* * The O_* flags used to have only F* names, which were used in the kernel * and by fcntl. We retain the F* names for the kernel f_flag field * and for backward compatibility for fcntl. These flags are deprecated. */ #if __BSD_VISIBLE #define FAPPEND O_APPEND /* kernel/compat */ #define FASYNC O_ASYNC /* kernel/compat */ #define FFSYNC O_FSYNC /* kernel */ #define FNONBLOCK O_NONBLOCK /* kernel */ #define FNDELAY O_NONBLOCK /* compat */ #define O_NDELAY O_NONBLOCK /* compat */ #endif /* * We are out of bits in f_flag (which is a short). However, * the flag bits not set in FMASK are only meaningful in the * initial open syscall. Those bits can thus be given a * different meaning for fcntl(2). */ #if __BSD_VISIBLE /* Read ahead */ #define FRDAHEAD O_CREAT #endif #if __POSIX_VISIBLE >= 200809 /* * Magic value that specify the use of the current working directory * to determine the target of relative file paths in the openat() and * similar syscalls. */ #define AT_FDCWD -100 /* * Miscellaneous flags for the *at() syscalls. */ #define AT_EACCESS 0x0100 /* Check access using effective user and group ID */ #define AT_SYMLINK_NOFOLLOW 0x0200 /* Do not follow symbolic links */ #define AT_SYMLINK_FOLLOW 0x0400 /* Follow symbolic link */ #define AT_REMOVEDIR 0x0800 /* Remove directory instead of file */ #define AT_BENEATH 0x1000 /* Fail if not under dirfd */ #endif /* * Constants used for fcntl(2) */ /* command values */ #define F_DUPFD 0 /* duplicate file descriptor */ #define F_GETFD 1 /* get file descriptor flags */ #define F_SETFD 2 /* set file descriptor flags */ #define F_GETFL 3 /* get file status flags */ #define F_SETFL 4 /* set file status flags */ #if __XSI_VISIBLE || __POSIX_VISIBLE >= 200112 #define F_GETOWN 5 /* get SIGIO/SIGURG proc/pgrp */ #define F_SETOWN 6 /* set SIGIO/SIGURG proc/pgrp */ #endif #if __BSD_VISIBLE #define F_OGETLK 7 /* get record locking information */ #define F_OSETLK 8 /* set record locking information */ #define F_OSETLKW 9 /* F_SETLK; wait if blocked */ #define F_DUP2FD 10 /* duplicate file descriptor to arg */ #endif #define F_GETLK 11 /* get record locking information */ #define F_SETLK 12 /* set record locking information */ #define F_SETLKW 13 /* F_SETLK; wait if blocked */ #if __BSD_VISIBLE #define F_SETLK_REMOTE 14 /* debugging support for remote locks */ #define F_READAHEAD 15 /* read ahead */ #define F_RDAHEAD 16 /* Darwin compatible read ahead */ #endif #if __POSIX_VISIBLE >= 200809 #define F_DUPFD_CLOEXEC 17 /* Like F_DUPFD, but FD_CLOEXEC is set */ #endif #if __BSD_VISIBLE #define F_DUP2FD_CLOEXEC 18 /* Like F_DUP2FD, but FD_CLOEXEC is set */ #endif /* file descriptor flags (F_GETFD, F_SETFD) */ #define FD_CLOEXEC 1 /* close-on-exec flag */ /* record locking flags (F_GETLK, F_SETLK, F_SETLKW) */ #define F_RDLCK 1 /* shared or read lock */ #define F_UNLCK 2 /* unlock */ #define F_WRLCK 3 /* exclusive or write lock */ #if __BSD_VISIBLE #define F_UNLCKSYS 4 /* purge locks for a given system ID */ #define F_CANCEL 5 /* cancel an async lock request */ #endif #ifdef _KERNEL #define F_WAIT 0x010 /* Wait until lock is granted */ #define F_FLOCK 0x020 /* Use flock(2) semantics for lock */ #define F_POSIX 0x040 /* Use POSIX semantics for lock */ #define F_REMOTE 0x080 /* Lock owner is remote NFS client */ #define F_NOINTR 0x100 /* Ignore signals when waiting */ #endif /* * Advisory file segment locking data type - * information passed to system by user */ struct flock { off_t l_start; /* starting offset */ off_t l_len; /* len = 0 means until end of file */ pid_t l_pid; /* lock owner */ short l_type; /* lock type: read/write, etc. */ short l_whence; /* type of l_start */ int l_sysid; /* remote system id or zero for local */ }; #if __BSD_VISIBLE /* * Old advisory file segment locking data type, * before adding l_sysid. */ struct __oflock { off_t l_start; /* starting offset */ off_t l_len; /* len = 0 means until end of file */ pid_t l_pid; /* lock owner */ short l_type; /* lock type: read/write, etc. */ short l_whence; /* type of l_start */ }; #endif #if __BSD_VISIBLE /* lock operations for flock(2) */ #define LOCK_SH 0x01 /* shared file lock */ #define LOCK_EX 0x02 /* exclusive file lock */ #define LOCK_NB 0x04 /* don't block when locking */ #define LOCK_UN 0x08 /* unlock file */ #endif #if __POSIX_VISIBLE >= 200112 /* * Advice to posix_fadvise */ #define POSIX_FADV_NORMAL 0 /* no special treatment */ #define POSIX_FADV_RANDOM 1 /* expect random page references */ #define POSIX_FADV_SEQUENTIAL 2 /* expect sequential page references */ #define POSIX_FADV_WILLNEED 3 /* will need these pages */ #define POSIX_FADV_DONTNEED 4 /* dont need these pages */ #define POSIX_FADV_NOREUSE 5 /* access data only once */ #endif + +#ifdef __BSD_VISIBLE +/* + * Magic value that specify that corresponding file descriptor to filename + * is unknown and sanitary check should be omitted in the funlinkat() and + * similar syscalls. + */ +#define FD_NONE -200 +#endif + #ifndef _KERNEL __BEGIN_DECLS int open(const char *, int, ...); int creat(const char *, mode_t); int fcntl(int, int, ...); #if __BSD_VISIBLE int flock(int, int); #endif #if __POSIX_VISIBLE >= 200809 int openat(int, const char *, int, ...); #endif #if __POSIX_VISIBLE >= 200112 int posix_fadvise(int, off_t, off_t, int); int posix_fallocate(int, off_t, off_t); #endif __END_DECLS #endif #endif /* !_SYS_FCNTL_H_ */ Index: head/sys/sys/syscallsubr.h =================================================================== --- head/sys/sys/syscallsubr.h (revision 345981) +++ head/sys/sys/syscallsubr.h (revision 345982) @@ -1,312 +1,312 @@ /*- * SPDX-License-Identifier: BSD-2-Clause-FreeBSD * * Copyright (c) 2002 Ian Dowse. 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 _SYS_SYSCALLSUBR_H_ #define _SYS_SYSCALLSUBR_H_ #include #include #include #include #include #include #include struct __wrusage; struct file; struct filecaps; enum idtype; struct itimerval; struct image_args; struct jail; struct kevent; struct kevent_copyops; struct kld_file_stat; struct ksiginfo; struct mbuf; struct msghdr; struct msqid_ds; struct pollfd; struct ogetdirentries_args; struct rlimit; struct rusage; struct sched_param; union semun; struct sockaddr; struct stat; struct thr_param; struct uio; int kern___getcwd(struct thread *td, char *buf, enum uio_seg bufseg, size_t buflen, size_t path_max); int kern_accept(struct thread *td, int s, struct sockaddr **name, socklen_t *namelen, struct file **fp); int kern_accept4(struct thread *td, int s, struct sockaddr **name, socklen_t *namelen, int flags, struct file **fp); int kern_accessat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int flags, int mode); int kern_adjtime(struct thread *td, struct timeval *delta, struct timeval *olddelta); int kern_alternate_path(struct thread *td, const char *prefix, const char *path, enum uio_seg pathseg, char **pathbuf, int create, int dirfd); int kern_bindat(struct thread *td, int dirfd, int fd, struct sockaddr *sa); int kern_break(struct thread *td, uintptr_t *addr); int kern_cap_ioctls_limit(struct thread *td, int fd, u_long *cmds, size_t ncmds); int kern_cap_rights_limit(struct thread *td, int fd, cap_rights_t *rights); int kern_chdir(struct thread *td, const char *path, enum uio_seg pathseg); int kern_clock_getcpuclockid2(struct thread *td, id_t id, int which, clockid_t *clk_id); int kern_clock_getres(struct thread *td, clockid_t clock_id, struct timespec *ts); int kern_clock_gettime(struct thread *td, clockid_t clock_id, struct timespec *ats); int kern_clock_nanosleep(struct thread *td, clockid_t clock_id, int flags, const struct timespec *rqtp, struct timespec *rmtp); int kern_clock_settime(struct thread *td, clockid_t clock_id, struct timespec *ats); int kern_close(struct thread *td, int fd); int kern_connectat(struct thread *td, int dirfd, int fd, struct sockaddr *sa); int kern_cpuset_getaffinity(struct thread *td, cpulevel_t level, cpuwhich_t which, id_t id, size_t cpusetsize, cpuset_t *maskp); int kern_cpuset_setaffinity(struct thread *td, cpulevel_t level, cpuwhich_t which, id_t id, size_t cpusetsize, const cpuset_t *maskp); int kern_cpuset_getdomain(struct thread *td, cpulevel_t level, cpuwhich_t which, id_t id, size_t domainsetsize, domainset_t *maskp, int *policyp); int kern_cpuset_setdomain(struct thread *td, cpulevel_t level, cpuwhich_t which, id_t id, size_t domainsetsize, const domainset_t *maskp, int policy); int kern_cpuset_getid(struct thread *td, cpulevel_t level, cpuwhich_t which, id_t id, cpusetid_t *setid); int kern_cpuset_setid(struct thread *td, cpuwhich_t which, id_t id, cpusetid_t setid); int kern_dup(struct thread *td, u_int mode, int flags, int old, int new); int kern_execve(struct thread *td, struct image_args *args, struct mac *mac_p); int kern_fchmodat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, mode_t mode, int flag); int kern_fchownat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int uid, int gid, int flag); int kern_fcntl(struct thread *td, int fd, int cmd, intptr_t arg); int kern_fcntl_freebsd(struct thread *td, int fd, int cmd, long arg); int kern_fhstat(struct thread *td, fhandle_t fh, struct stat *buf); int kern_fhstatfs(struct thread *td, fhandle_t fh, struct statfs *buf); int kern_fpathconf(struct thread *td, int fd, int name, long *valuep); int kern_fstat(struct thread *td, int fd, struct stat *sbp); int kern_fstatfs(struct thread *td, int fd, struct statfs *buf); int kern_fsync(struct thread *td, int fd, bool fullsync); int kern_ftruncate(struct thread *td, int fd, off_t length); int kern_futimes(struct thread *td, int fd, struct timeval *tptr, enum uio_seg tptrseg); int kern_futimens(struct thread *td, int fd, struct timespec *tptr, enum uio_seg tptrseg); int kern_getdirentries(struct thread *td, int fd, char *buf, size_t count, off_t *basep, ssize_t *residp, enum uio_seg bufseg); int kern_getfsstat(struct thread *td, struct statfs **buf, size_t bufsize, size_t *countp, enum uio_seg bufseg, int mode); int kern_getitimer(struct thread *, u_int, struct itimerval *); int kern_getppid(struct thread *); int kern_getpeername(struct thread *td, int fd, struct sockaddr **sa, socklen_t *alen); int kern_getrusage(struct thread *td, int who, struct rusage *rup); int kern_getsockname(struct thread *td, int fd, struct sockaddr **sa, socklen_t *alen); int kern_getsockopt(struct thread *td, int s, int level, int name, void *optval, enum uio_seg valseg, socklen_t *valsize); int kern_ioctl(struct thread *td, int fd, u_long com, caddr_t data); int kern_jail(struct thread *td, struct jail *j); int kern_jail_get(struct thread *td, struct uio *options, int flags); int kern_jail_set(struct thread *td, struct uio *options, int flags); int kern_kevent(struct thread *td, int fd, int nchanges, int nevents, struct kevent_copyops *k_ops, const struct timespec *timeout); int kern_kevent_anonymous(struct thread *td, int nevents, struct kevent_copyops *k_ops); int kern_kevent_fp(struct thread *td, struct file *fp, int nchanges, int nevents, struct kevent_copyops *k_ops, const struct timespec *timeout); int kern_kqueue(struct thread *td, int flags, struct filecaps *fcaps); int kern_kldload(struct thread *td, const char *file, int *fileid); int kern_kldstat(struct thread *td, int fileid, struct kld_file_stat *stat); int kern_kldunload(struct thread *td, int fileid, int flags); int kern_linkat(struct thread *td, int fd1, int fd2, const char *path1, const char *path2, enum uio_seg segflg, int follow); int kern_listen(struct thread *td, int s, int backlog); int kern_lseek(struct thread *td, int fd, off_t offset, int whence); int kern_lutimes(struct thread *td, const char *path, enum uio_seg pathseg, struct timeval *tptr, enum uio_seg tptrseg); int kern_madvise(struct thread *td, uintptr_t addr, size_t len, int behav); int kern_mincore(struct thread *td, uintptr_t addr, size_t len, char *vec); int kern_mkdirat(struct thread *td, int fd, const char *path, enum uio_seg segflg, int mode); int kern_mkfifoat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int mode); int kern_mknodat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int mode, dev_t dev); int kern_mlock(struct proc *proc, struct ucred *cred, uintptr_t addr, size_t len); int kern_mmap(struct thread *td, uintptr_t addr, size_t size, int prot, int flags, int fd, off_t pos); int kern_mprotect(struct thread *td, uintptr_t addr, size_t size, int prot); int kern_msgctl(struct thread *, int, int, struct msqid_ds *); int kern_msgrcv(struct thread *, int, void *, size_t, long, int, long *); int kern_msgsnd(struct thread *, int, const void *, size_t, int, long); int kern_msync(struct thread *td, uintptr_t addr, size_t size, int flags); int kern_munlock(struct thread *td, uintptr_t addr, size_t size); int kern_munmap(struct thread *td, uintptr_t addr, size_t size); int kern_nanosleep(struct thread *td, struct timespec *rqt, struct timespec *rmt); int kern_ogetdirentries(struct thread *td, struct ogetdirentries_args *uap, long *ploff); int kern_openat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, int flags, int mode); int kern_pathconf(struct thread *td, const char *path, enum uio_seg pathseg, int name, u_long flags, long *valuep); int kern_pipe(struct thread *td, int fildes[2], int flags, struct filecaps *fcaps1, struct filecaps *fcaps2); int kern_poll(struct thread *td, struct pollfd *fds, u_int nfds, struct timespec *tsp, sigset_t *uset); int kern_posix_error(struct thread *td, int error); int kern_posix_fadvise(struct thread *td, int fd, off_t offset, off_t len, int advice); int kern_posix_fallocate(struct thread *td, int fd, off_t offset, off_t len); int kern_procctl(struct thread *td, enum idtype idtype, id_t id, int com, void *data); int kern_pread(struct thread *td, int fd, void *buf, size_t nbyte, off_t offset); int kern_preadv(struct thread *td, int fd, struct uio *auio, off_t offset); int kern_pselect(struct thread *td, int nd, fd_set *in, fd_set *ou, fd_set *ex, struct timeval *tvp, sigset_t *uset, int abi_nfdbits); int kern_ptrace(struct thread *td, int req, pid_t pid, void *addr, int data); int kern_pwrite(struct thread *td, int fd, const void *buf, size_t nbyte, off_t offset); int kern_pwritev(struct thread *td, int fd, struct uio *auio, off_t offset); int kern_readlinkat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, char *buf, enum uio_seg bufseg, size_t count); int kern_readv(struct thread *td, int fd, struct uio *auio); int kern_recvit(struct thread *td, int s, struct msghdr *mp, enum uio_seg fromseg, struct mbuf **controlp); int kern_renameat(struct thread *td, int oldfd, const char *old, int newfd, const char *new, enum uio_seg pathseg); -int kern_rmdirat(struct thread *td, int fd, const char *path, +int kern_frmdirat(struct thread *td, int dfd, const char *path, int fd, enum uio_seg pathseg, int flag); int kern_sched_getparam(struct thread *td, struct thread *targettd, struct sched_param *param); int kern_sched_getscheduler(struct thread *td, struct thread *targettd, int *policy); int kern_sched_setparam(struct thread *td, struct thread *targettd, struct sched_param *param); int kern_sched_setscheduler(struct thread *td, struct thread *targettd, int policy, struct sched_param *param); int kern_sched_rr_get_interval(struct thread *td, pid_t pid, struct timespec *ts); int kern_sched_rr_get_interval_td(struct thread *td, struct thread *targettd, struct timespec *ts); int kern_semctl(struct thread *td, int semid, int semnum, int cmd, union semun *arg, register_t *rval); int kern_select(struct thread *td, int nd, fd_set *fd_in, fd_set *fd_ou, fd_set *fd_ex, struct timeval *tvp, int abi_nfdbits); int kern_sendit(struct thread *td, int s, struct msghdr *mp, int flags, struct mbuf *control, enum uio_seg segflg); int kern_setgroups(struct thread *td, u_int ngrp, gid_t *groups); int kern_setitimer(struct thread *, u_int, struct itimerval *, struct itimerval *); int kern_setrlimit(struct thread *, u_int, struct rlimit *); int kern_setsockopt(struct thread *td, int s, int level, int name, const void *optval, enum uio_seg valseg, socklen_t valsize); int kern_settimeofday(struct thread *td, struct timeval *tv, struct timezone *tzp); int kern_shm_open(struct thread *td, const char *userpath, int flags, mode_t mode, struct filecaps *fcaps); int kern_shmat(struct thread *td, int shmid, const void *shmaddr, int shmflg); int kern_shmctl(struct thread *td, int shmid, int cmd, void *buf, size_t *bufsz); int kern_shutdown(struct thread *td, int s, int how); int kern_sigaction(struct thread *td, int sig, const struct sigaction *act, struct sigaction *oact, int flags); int kern_sigaltstack(struct thread *td, stack_t *ss, stack_t *oss); int kern_sigprocmask(struct thread *td, int how, sigset_t *set, sigset_t *oset, int flags); int kern_sigsuspend(struct thread *td, sigset_t mask); int kern_sigtimedwait(struct thread *td, sigset_t waitset, struct ksiginfo *ksi, struct timespec *timeout); int kern_sigqueue(struct thread *td, pid_t pid, int signum, union sigval *value); int kern_socket(struct thread *td, int domain, int type, int protocol); int kern_statat(struct thread *td, int flag, int fd, const char *path, enum uio_seg pathseg, struct stat *sbp, void (*hook)(struct vnode *vp, struct stat *sbp)); int kern_statfs(struct thread *td, const char *path, enum uio_seg pathseg, struct statfs *buf); int kern_symlinkat(struct thread *td, const char *path1, int fd, const char *path2, enum uio_seg segflg); int kern_ktimer_create(struct thread *td, clockid_t clock_id, struct sigevent *evp, int *timerid, int preset_id); int kern_ktimer_delete(struct thread *, int); int kern_ktimer_settime(struct thread *td, int timer_id, int flags, struct itimerspec *val, struct itimerspec *oval); int kern_ktimer_gettime(struct thread *td, int timer_id, struct itimerspec *val); int kern_ktimer_getoverrun(struct thread *td, int timer_id); int kern_thr_alloc(struct proc *, int pages, struct thread **); int kern_thr_exit(struct thread *td); int kern_thr_new(struct thread *td, struct thr_param *param); int kern_thr_suspend(struct thread *td, struct timespec *tsp); int kern_truncate(struct thread *td, const char *path, enum uio_seg pathseg, off_t length); -int kern_unlinkat(struct thread *td, int fd, const char *path, +int kern_funlinkat(struct thread *td, int dfd, const char *path, int fd, enum uio_seg pathseg, int flag, ino_t oldinum); int kern_utimesat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, struct timeval *tptr, enum uio_seg tptrseg); int kern_utimensat(struct thread *td, int fd, const char *path, enum uio_seg pathseg, struct timespec *tptr, enum uio_seg tptrseg, int follow); int kern_wait(struct thread *td, pid_t pid, int *status, int options, struct rusage *rup); int kern_wait6(struct thread *td, enum idtype idtype, id_t id, int *status, int options, struct __wrusage *wrup, siginfo_t *sip); int kern_writev(struct thread *td, int fd, struct uio *auio); int kern_socketpair(struct thread *td, int domain, int type, int protocol, int *rsv); /* flags for kern_sigaction */ #define KSA_OSIGSET 0x0001 /* uses osigact_t */ #define KSA_FREEBSD4 0x0002 /* uses ucontext4 */ struct freebsd11_dirent; int freebsd11_kern_getdirentries(struct thread *td, int fd, char *ubuf, u_int count, long *basep, void (*func)(struct freebsd11_dirent *)); #endif /* !_SYS_SYSCALLSUBR_H_ */ Index: head/sys/ufs/ffs/ffs_alloc.c =================================================================== --- head/sys/ufs/ffs/ffs_alloc.c (revision 345981) +++ head/sys/ufs/ffs/ffs_alloc.c (revision 345982) @@ -1,3612 +1,3613 @@ /*- * SPDX-License-Identifier: (BSD-2-Clause-FreeBSD AND BSD-3-Clause) * * Copyright (c) 2002 Networks Associates Technology, Inc. * All rights reserved. * * This software was developed for the FreeBSD Project by Marshall * Kirk McKusick and Network Associates Laboratories, the Security * Research Division of Network Associates, Inc. under DARPA/SPAWAR * contract N66001-01-C-8035 ("CBOSS"), as part of the DARPA CHATS * research program * * 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. * * 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. * 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. * * @(#)ffs_alloc.c 8.18 (Berkeley) 5/26/95 */ #include __FBSDID("$FreeBSD$"); #include "opt_quota.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 typedef ufs2_daddr_t allocfcn_t(struct inode *ip, u_int cg, ufs2_daddr_t bpref, int size, int rsize); static ufs2_daddr_t ffs_alloccg(struct inode *, u_int, ufs2_daddr_t, int, int); static ufs2_daddr_t ffs_alloccgblk(struct inode *, struct buf *, ufs2_daddr_t, int); static void ffs_blkfree_cg(struct ufsmount *, struct fs *, struct vnode *, ufs2_daddr_t, long, ino_t, struct workhead *); #ifdef INVARIANTS static int ffs_checkblk(struct inode *, ufs2_daddr_t, long); #endif static ufs2_daddr_t ffs_clusteralloc(struct inode *, u_int, ufs2_daddr_t, int); static ino_t ffs_dirpref(struct inode *); static ufs2_daddr_t ffs_fragextend(struct inode *, u_int, ufs2_daddr_t, int, int); static ufs2_daddr_t ffs_hashalloc (struct inode *, u_int, ufs2_daddr_t, int, int, allocfcn_t *); static ufs2_daddr_t ffs_nodealloccg(struct inode *, u_int, ufs2_daddr_t, int, int); static ufs1_daddr_t ffs_mapsearch(struct fs *, struct cg *, ufs2_daddr_t, int); static int ffs_reallocblks_ufs1(struct vop_reallocblks_args *); static int ffs_reallocblks_ufs2(struct vop_reallocblks_args *); static void ffs_ckhash_cg(struct buf *); /* * Allocate a block in the filesystem. * * The size of the requested block is given, which must be some * multiple of fs_fsize and <= fs_bsize. * A preference may be optionally specified. If a preference is given * the following hierarchy is used to allocate a block: * 1) allocate the requested block. * 2) allocate a rotationally optimal block in the same cylinder. * 3) allocate a block in the same cylinder group. * 4) quadradically rehash into other cylinder groups, until an * available block is located. * If no block preference is given the following hierarchy is used * to allocate a block: * 1) allocate a block in the cylinder group that contains the * inode for the file. * 2) quadradically rehash into other cylinder groups, until an * available block is located. */ int ffs_alloc(ip, lbn, bpref, size, flags, cred, bnp) struct inode *ip; ufs2_daddr_t lbn, bpref; int size, flags; struct ucred *cred; ufs2_daddr_t *bnp; { struct fs *fs; struct ufsmount *ump; ufs2_daddr_t bno; u_int cg, reclaimed; static struct timeval lastfail; static int curfail; int64_t delta; #ifdef QUOTA int error; #endif *bnp = 0; ump = ITOUMP(ip); fs = ump->um_fs; mtx_assert(UFS_MTX(ump), MA_OWNED); #ifdef INVARIANTS if ((u_int)size > fs->fs_bsize || fragoff(fs, size) != 0) { printf("dev = %s, bsize = %ld, size = %d, fs = %s\n", devtoname(ump->um_dev), (long)fs->fs_bsize, size, fs->fs_fsmnt); panic("ffs_alloc: bad size"); } if (cred == NOCRED) panic("ffs_alloc: missing credential"); #endif /* INVARIANTS */ reclaimed = 0; retry: #ifdef QUOTA UFS_UNLOCK(ump); error = chkdq(ip, btodb(size), cred, 0); if (error) return (error); UFS_LOCK(ump); #endif if (size == fs->fs_bsize && fs->fs_cstotal.cs_nbfree == 0) goto nospace; if (priv_check_cred(cred, PRIV_VFS_BLOCKRESERVE) && freespace(fs, fs->fs_minfree) - numfrags(fs, size) < 0) goto nospace; if (bpref >= fs->fs_size) bpref = 0; if (bpref == 0) cg = ino_to_cg(fs, ip->i_number); else cg = dtog(fs, bpref); bno = ffs_hashalloc(ip, cg, bpref, size, size, ffs_alloccg); if (bno > 0) { delta = btodb(size); DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + delta); if (flags & IO_EXT) ip->i_flag |= IN_CHANGE; else ip->i_flag |= IN_CHANGE | IN_UPDATE; *bnp = bno; return (0); } nospace: #ifdef QUOTA UFS_UNLOCK(ump); /* * Restore user's disk quota because allocation failed. */ (void) chkdq(ip, -btodb(size), cred, FORCE); UFS_LOCK(ump); #endif if (reclaimed == 0 && (flags & IO_BUFLOCKED) == 0) { reclaimed = 1; softdep_request_cleanup(fs, ITOV(ip), cred, FLUSH_BLOCKS_WAIT); goto retry; } UFS_UNLOCK(ump); if (reclaimed > 0 && ppsratecheck(&lastfail, &curfail, 1)) { ffs_fserr(fs, ip->i_number, "filesystem full"); uprintf("\n%s: write failed, filesystem is full\n", fs->fs_fsmnt); } return (ENOSPC); } /* * Reallocate a fragment to a bigger size * * The number and size of the old block is given, and a preference * and new size is also specified. The allocator attempts to extend * the original block. Failing that, the regular block allocator is * invoked to get an appropriate block. */ int ffs_realloccg(ip, lbprev, bprev, bpref, osize, nsize, flags, cred, bpp) struct inode *ip; ufs2_daddr_t lbprev; ufs2_daddr_t bprev; ufs2_daddr_t bpref; int osize, nsize, flags; struct ucred *cred; struct buf **bpp; { struct vnode *vp; struct fs *fs; struct buf *bp; struct ufsmount *ump; u_int cg, request, reclaimed; int error, gbflags; ufs2_daddr_t bno; static struct timeval lastfail; static int curfail; int64_t delta; vp = ITOV(ip); ump = ITOUMP(ip); fs = ump->um_fs; bp = NULL; gbflags = (flags & BA_UNMAPPED) != 0 ? GB_UNMAPPED : 0; mtx_assert(UFS_MTX(ump), MA_OWNED); #ifdef INVARIANTS if (vp->v_mount->mnt_kern_flag & MNTK_SUSPENDED) panic("ffs_realloccg: allocation on suspended filesystem"); if ((u_int)osize > fs->fs_bsize || fragoff(fs, osize) != 0 || (u_int)nsize > fs->fs_bsize || fragoff(fs, nsize) != 0) { printf( "dev = %s, bsize = %ld, osize = %d, nsize = %d, fs = %s\n", devtoname(ump->um_dev), (long)fs->fs_bsize, osize, nsize, fs->fs_fsmnt); panic("ffs_realloccg: bad size"); } if (cred == NOCRED) panic("ffs_realloccg: missing credential"); #endif /* INVARIANTS */ reclaimed = 0; retry: if (priv_check_cred(cred, PRIV_VFS_BLOCKRESERVE) && freespace(fs, fs->fs_minfree) - numfrags(fs, nsize - osize) < 0) { goto nospace; } if (bprev == 0) { printf("dev = %s, bsize = %ld, bprev = %jd, fs = %s\n", devtoname(ump->um_dev), (long)fs->fs_bsize, (intmax_t)bprev, fs->fs_fsmnt); panic("ffs_realloccg: bad bprev"); } UFS_UNLOCK(ump); /* * Allocate the extra space in the buffer. */ error = bread_gb(vp, lbprev, osize, NOCRED, gbflags, &bp); if (error) { brelse(bp); return (error); } if (bp->b_blkno == bp->b_lblkno) { if (lbprev >= UFS_NDADDR) panic("ffs_realloccg: lbprev out of range"); bp->b_blkno = fsbtodb(fs, bprev); } #ifdef QUOTA error = chkdq(ip, btodb(nsize - osize), cred, 0); if (error) { brelse(bp); return (error); } #endif /* * Check for extension in the existing location. */ *bpp = NULL; cg = dtog(fs, bprev); UFS_LOCK(ump); bno = ffs_fragextend(ip, cg, bprev, osize, nsize); if (bno) { if (bp->b_blkno != fsbtodb(fs, bno)) panic("ffs_realloccg: bad blockno"); delta = btodb(nsize - osize); DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + delta); if (flags & IO_EXT) ip->i_flag |= IN_CHANGE; else ip->i_flag |= IN_CHANGE | IN_UPDATE; allocbuf(bp, nsize); bp->b_flags |= B_DONE; vfs_bio_bzero_buf(bp, osize, nsize - osize); if ((bp->b_flags & (B_MALLOC | B_VMIO)) == B_VMIO) vfs_bio_set_valid(bp, osize, nsize - osize); *bpp = bp; return (0); } /* * Allocate a new disk location. */ if (bpref >= fs->fs_size) bpref = 0; switch ((int)fs->fs_optim) { case FS_OPTSPACE: /* * Allocate an exact sized fragment. Although this makes * best use of space, we will waste time relocating it if * the file continues to grow. If the fragmentation is * less than half of the minimum free reserve, we choose * to begin optimizing for time. */ request = nsize; if (fs->fs_minfree <= 5 || fs->fs_cstotal.cs_nffree > (off_t)fs->fs_dsize * fs->fs_minfree / (2 * 100)) break; log(LOG_NOTICE, "%s: optimization changed from SPACE to TIME\n", fs->fs_fsmnt); fs->fs_optim = FS_OPTTIME; break; case FS_OPTTIME: /* * At this point we have discovered a file that is trying to * grow a small fragment to a larger fragment. To save time, * we allocate a full sized block, then free the unused portion. * If the file continues to grow, the `ffs_fragextend' call * above will be able to grow it in place without further * copying. If aberrant programs cause disk fragmentation to * grow within 2% of the free reserve, we choose to begin * optimizing for space. */ request = fs->fs_bsize; if (fs->fs_cstotal.cs_nffree < (off_t)fs->fs_dsize * (fs->fs_minfree - 2) / 100) break; log(LOG_NOTICE, "%s: optimization changed from TIME to SPACE\n", fs->fs_fsmnt); fs->fs_optim = FS_OPTSPACE; break; default: printf("dev = %s, optim = %ld, fs = %s\n", devtoname(ump->um_dev), (long)fs->fs_optim, fs->fs_fsmnt); panic("ffs_realloccg: bad optim"); /* NOTREACHED */ } bno = ffs_hashalloc(ip, cg, bpref, request, nsize, ffs_alloccg); if (bno > 0) { bp->b_blkno = fsbtodb(fs, bno); if (!DOINGSOFTDEP(vp)) /* * The usual case is that a smaller fragment that * was just allocated has been replaced with a bigger * fragment or a full-size block. If it is marked as * B_DELWRI, the current contents have not been written * to disk. It is possible that the block was written * earlier, but very uncommon. If the block has never * been written, there is no need to send a BIO_DELETE * for it when it is freed. The gain from avoiding the * TRIMs for the common case of unwritten blocks far * exceeds the cost of the write amplification for the * uncommon case of failing to send a TRIM for a block * that had been written. */ ffs_blkfree(ump, fs, ump->um_devvp, bprev, (long)osize, ip->i_number, vp->v_type, NULL, (bp->b_flags & B_DELWRI) != 0 ? NOTRIM_KEY : SINGLETON_KEY); delta = btodb(nsize - osize); DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + delta); if (flags & IO_EXT) ip->i_flag |= IN_CHANGE; else ip->i_flag |= IN_CHANGE | IN_UPDATE; allocbuf(bp, nsize); bp->b_flags |= B_DONE; vfs_bio_bzero_buf(bp, osize, nsize - osize); if ((bp->b_flags & (B_MALLOC | B_VMIO)) == B_VMIO) vfs_bio_set_valid(bp, osize, nsize - osize); *bpp = bp; return (0); } #ifdef QUOTA UFS_UNLOCK(ump); /* * Restore user's disk quota because allocation failed. */ (void) chkdq(ip, -btodb(nsize - osize), cred, FORCE); UFS_LOCK(ump); #endif nospace: /* * no space available */ if (reclaimed == 0 && (flags & IO_BUFLOCKED) == 0) { reclaimed = 1; UFS_UNLOCK(ump); if (bp) { brelse(bp); bp = NULL; } UFS_LOCK(ump); softdep_request_cleanup(fs, vp, cred, FLUSH_BLOCKS_WAIT); goto retry; } UFS_UNLOCK(ump); if (bp) brelse(bp); if (reclaimed > 0 && ppsratecheck(&lastfail, &curfail, 1)) { ffs_fserr(fs, ip->i_number, "filesystem full"); uprintf("\n%s: write failed, filesystem is full\n", fs->fs_fsmnt); } return (ENOSPC); } /* * Reallocate a sequence of blocks into a contiguous sequence of blocks. * * The vnode and an array of buffer pointers for a range of sequential * logical blocks to be made contiguous is given. The allocator attempts * to find a range of sequential blocks starting as close as possible * from the end of the allocation for the logical block immediately * preceding the current range. If successful, the physical block numbers * in the buffer pointers and in the inode are changed to reflect the new * allocation. If unsuccessful, the allocation is left unchanged. The * success in doing the reallocation is returned. Note that the error * return is not reflected back to the user. Rather the previous block * allocation will be used. */ SYSCTL_NODE(_vfs, OID_AUTO, ffs, CTLFLAG_RW, 0, "FFS filesystem"); static int doasyncfree = 1; SYSCTL_INT(_vfs_ffs, OID_AUTO, doasyncfree, CTLFLAG_RW, &doasyncfree, 0, "do not force synchronous writes when blocks are reallocated"); static int doreallocblks = 1; SYSCTL_INT(_vfs_ffs, OID_AUTO, doreallocblks, CTLFLAG_RW, &doreallocblks, 0, "enable block reallocation"); static int dotrimcons = 1; SYSCTL_INT(_vfs_ffs, OID_AUTO, dotrimcons, CTLFLAG_RWTUN, &dotrimcons, 0, "enable BIO_DELETE / TRIM consolidation"); static int maxclustersearch = 10; SYSCTL_INT(_vfs_ffs, OID_AUTO, maxclustersearch, CTLFLAG_RW, &maxclustersearch, 0, "max number of cylinder group to search for contigous blocks"); #ifdef DEBUG static volatile int prtrealloc = 0; #endif int ffs_reallocblks(ap) struct vop_reallocblks_args /* { struct vnode *a_vp; struct cluster_save *a_buflist; } */ *ap; { struct ufsmount *ump; /* * We used to skip reallocating the blocks of a file into a * contiguous sequence if the underlying flash device requested * BIO_DELETE notifications, because devices that benefit from * BIO_DELETE also benefit from not moving the data. However, * the destination for the data is usually moved before the data * is written to the initially allocated location, so we rarely * suffer the penalty of extra writes. With the addition of the * consolidation of contiguous blocks into single BIO_DELETE * operations, having fewer but larger contiguous blocks reduces * the number of (slow and expensive) BIO_DELETE operations. So * when doing BIO_DELETE consolidation, we do block reallocation. * * Skip if reallocblks has been disabled globally. */ ump = ap->a_vp->v_mount->mnt_data; if ((((ump->um_flags) & UM_CANDELETE) != 0 && dotrimcons == 0) || doreallocblks == 0) return (ENOSPC); /* * We can't wait in softdep prealloc as it may fsync and recurse * here. Instead we simply fail to reallocate blocks if this * rare condition arises. */ if (DOINGSOFTDEP(ap->a_vp)) if (softdep_prealloc(ap->a_vp, MNT_NOWAIT) != 0) return (ENOSPC); if (ump->um_fstype == UFS1) return (ffs_reallocblks_ufs1(ap)); return (ffs_reallocblks_ufs2(ap)); } static int ffs_reallocblks_ufs1(ap) struct vop_reallocblks_args /* { struct vnode *a_vp; struct cluster_save *a_buflist; } */ *ap; { struct fs *fs; struct inode *ip; struct vnode *vp; struct buf *sbp, *ebp, *bp; ufs1_daddr_t *bap, *sbap, *ebap; struct cluster_save *buflist; struct ufsmount *ump; ufs_lbn_t start_lbn, end_lbn; ufs1_daddr_t soff, newblk, blkno; ufs2_daddr_t pref; struct indir start_ap[UFS_NIADDR + 1], end_ap[UFS_NIADDR + 1], *idp; int i, cg, len, start_lvl, end_lvl, ssize; vp = ap->a_vp; ip = VTOI(vp); ump = ITOUMP(ip); fs = ump->um_fs; /* * If we are not tracking block clusters or if we have less than 4% * free blocks left, then do not attempt to cluster. Running with * less than 5% free block reserve is not recommended and those that * choose to do so do not expect to have good file layout. */ if (fs->fs_contigsumsize <= 0 || freespace(fs, 4) < 0) return (ENOSPC); buflist = ap->a_buflist; len = buflist->bs_nchildren; start_lbn = buflist->bs_children[0]->b_lblkno; end_lbn = start_lbn + len - 1; #ifdef INVARIANTS for (i = 0; i < len; i++) if (!ffs_checkblk(ip, dbtofsb(fs, buflist->bs_children[i]->b_blkno), fs->fs_bsize)) panic("ffs_reallocblks: unallocated block 1"); for (i = 1; i < len; i++) if (buflist->bs_children[i]->b_lblkno != start_lbn + i) panic("ffs_reallocblks: non-logical cluster"); blkno = buflist->bs_children[0]->b_blkno; ssize = fsbtodb(fs, fs->fs_frag); for (i = 1; i < len - 1; i++) if (buflist->bs_children[i]->b_blkno != blkno + (i * ssize)) panic("ffs_reallocblks: non-physical cluster %d", i); #endif /* * If the cluster crosses the boundary for the first indirect * block, leave space for the indirect block. Indirect blocks * are initially laid out in a position after the last direct * block. Block reallocation would usually destroy locality by * moving the indirect block out of the way to make room for * data blocks if we didn't compensate here. We should also do * this for other indirect block boundaries, but it is only * important for the first one. */ if (start_lbn < UFS_NDADDR && end_lbn >= UFS_NDADDR) return (ENOSPC); /* * If the latest allocation is in a new cylinder group, assume that * the filesystem has decided to move and do not force it back to * the previous cylinder group. */ if (dtog(fs, dbtofsb(fs, buflist->bs_children[0]->b_blkno)) != dtog(fs, dbtofsb(fs, buflist->bs_children[len - 1]->b_blkno))) return (ENOSPC); if (ufs_getlbns(vp, start_lbn, start_ap, &start_lvl) || ufs_getlbns(vp, end_lbn, end_ap, &end_lvl)) return (ENOSPC); /* * Get the starting offset and block map for the first block. */ if (start_lvl == 0) { sbap = &ip->i_din1->di_db[0]; soff = start_lbn; } else { idp = &start_ap[start_lvl - 1]; if (bread(vp, idp->in_lbn, (int)fs->fs_bsize, NOCRED, &sbp)) { brelse(sbp); return (ENOSPC); } sbap = (ufs1_daddr_t *)sbp->b_data; soff = idp->in_off; } /* * If the block range spans two block maps, get the second map. */ ebap = NULL; if (end_lvl == 0 || (idp = &end_ap[end_lvl - 1])->in_off + 1 >= len) { ssize = len; } else { #ifdef INVARIANTS if (start_lvl > 0 && start_ap[start_lvl - 1].in_lbn == idp->in_lbn) panic("ffs_reallocblk: start == end"); #endif ssize = len - (idp->in_off + 1); if (bread(vp, idp->in_lbn, (int)fs->fs_bsize, NOCRED, &ebp)) goto fail; ebap = (ufs1_daddr_t *)ebp->b_data; } /* * Find the preferred location for the cluster. If we have not * previously failed at this endeavor, then follow our standard * preference calculation. If we have failed at it, then pick up * where we last ended our search. */ UFS_LOCK(ump); if (ip->i_nextclustercg == -1) pref = ffs_blkpref_ufs1(ip, start_lbn, soff, sbap); else pref = cgdata(fs, ip->i_nextclustercg); /* * Search the block map looking for an allocation of the desired size. * To avoid wasting too much time, we limit the number of cylinder * groups that we will search. */ cg = dtog(fs, pref); for (i = min(maxclustersearch, fs->fs_ncg); i > 0; i--) { if ((newblk = ffs_clusteralloc(ip, cg, pref, len)) != 0) break; cg += 1; if (cg >= fs->fs_ncg) cg = 0; } /* * If we have failed in our search, record where we gave up for * next time. Otherwise, fall back to our usual search citerion. */ if (newblk == 0) { ip->i_nextclustercg = cg; UFS_UNLOCK(ump); goto fail; } ip->i_nextclustercg = -1; /* * We have found a new contiguous block. * * First we have to replace the old block pointers with the new * block pointers in the inode and indirect blocks associated * with the file. */ #ifdef DEBUG if (prtrealloc) printf("realloc: ino %ju, lbns %jd-%jd\n\told:", (uintmax_t)ip->i_number, (intmax_t)start_lbn, (intmax_t)end_lbn); #endif blkno = newblk; for (bap = &sbap[soff], i = 0; i < len; i++, blkno += fs->fs_frag) { if (i == ssize) { bap = ebap; soff = -i; } #ifdef INVARIANTS if (!ffs_checkblk(ip, dbtofsb(fs, buflist->bs_children[i]->b_blkno), fs->fs_bsize)) panic("ffs_reallocblks: unallocated block 2"); if (dbtofsb(fs, buflist->bs_children[i]->b_blkno) != *bap) panic("ffs_reallocblks: alloc mismatch"); #endif #ifdef DEBUG if (prtrealloc) printf(" %d,", *bap); #endif if (DOINGSOFTDEP(vp)) { if (sbap == &ip->i_din1->di_db[0] && i < ssize) softdep_setup_allocdirect(ip, start_lbn + i, blkno, *bap, fs->fs_bsize, fs->fs_bsize, buflist->bs_children[i]); else softdep_setup_allocindir_page(ip, start_lbn + i, i < ssize ? sbp : ebp, soff + i, blkno, *bap, buflist->bs_children[i]); } *bap++ = blkno; } /* * Next we must write out the modified inode and indirect blocks. * For strict correctness, the writes should be synchronous since * the old block values may have been written to disk. In practise * they are almost never written, but if we are concerned about * strict correctness, the `doasyncfree' flag should be set to zero. * * The test on `doasyncfree' should be changed to test a flag * that shows whether the associated buffers and inodes have * been written. The flag should be set when the cluster is * started and cleared whenever the buffer or inode is flushed. * We can then check below to see if it is set, and do the * synchronous write only when it has been cleared. */ if (sbap != &ip->i_din1->di_db[0]) { if (doasyncfree) bdwrite(sbp); else bwrite(sbp); } else { ip->i_flag |= IN_CHANGE | IN_UPDATE; if (!doasyncfree) ffs_update(vp, 1); } if (ssize < len) { if (doasyncfree) bdwrite(ebp); else bwrite(ebp); } /* * Last, free the old blocks and assign the new blocks to the buffers. */ #ifdef DEBUG if (prtrealloc) printf("\n\tnew:"); #endif for (blkno = newblk, i = 0; i < len; i++, blkno += fs->fs_frag) { bp = buflist->bs_children[i]; if (!DOINGSOFTDEP(vp)) /* * The usual case is that a set of N-contiguous blocks * that was just allocated has been replaced with a * set of N+1-contiguous blocks. If they are marked as * B_DELWRI, the current contents have not been written * to disk. It is possible that the blocks were written * earlier, but very uncommon. If the blocks have never * been written, there is no need to send a BIO_DELETE * for them when they are freed. The gain from avoiding * the TRIMs for the common case of unwritten blocks * far exceeds the cost of the write amplification for * the uncommon case of failing to send a TRIM for the * blocks that had been written. */ ffs_blkfree(ump, fs, ump->um_devvp, dbtofsb(fs, bp->b_blkno), fs->fs_bsize, ip->i_number, vp->v_type, NULL, (bp->b_flags & B_DELWRI) != 0 ? NOTRIM_KEY : SINGLETON_KEY); bp->b_blkno = fsbtodb(fs, blkno); #ifdef INVARIANTS if (!ffs_checkblk(ip, dbtofsb(fs, bp->b_blkno), fs->fs_bsize)) panic("ffs_reallocblks: unallocated block 3"); #endif #ifdef DEBUG if (prtrealloc) printf(" %d,", blkno); #endif } #ifdef DEBUG if (prtrealloc) { prtrealloc--; printf("\n"); } #endif return (0); fail: if (ssize < len) brelse(ebp); if (sbap != &ip->i_din1->di_db[0]) brelse(sbp); return (ENOSPC); } static int ffs_reallocblks_ufs2(ap) struct vop_reallocblks_args /* { struct vnode *a_vp; struct cluster_save *a_buflist; } */ *ap; { struct fs *fs; struct inode *ip; struct vnode *vp; struct buf *sbp, *ebp, *bp; ufs2_daddr_t *bap, *sbap, *ebap; struct cluster_save *buflist; struct ufsmount *ump; ufs_lbn_t start_lbn, end_lbn; ufs2_daddr_t soff, newblk, blkno, pref; struct indir start_ap[UFS_NIADDR + 1], end_ap[UFS_NIADDR + 1], *idp; int i, cg, len, start_lvl, end_lvl, ssize; vp = ap->a_vp; ip = VTOI(vp); ump = ITOUMP(ip); fs = ump->um_fs; /* * If we are not tracking block clusters or if we have less than 4% * free blocks left, then do not attempt to cluster. Running with * less than 5% free block reserve is not recommended and those that * choose to do so do not expect to have good file layout. */ if (fs->fs_contigsumsize <= 0 || freespace(fs, 4) < 0) return (ENOSPC); buflist = ap->a_buflist; len = buflist->bs_nchildren; start_lbn = buflist->bs_children[0]->b_lblkno; end_lbn = start_lbn + len - 1; #ifdef INVARIANTS for (i = 0; i < len; i++) if (!ffs_checkblk(ip, dbtofsb(fs, buflist->bs_children[i]->b_blkno), fs->fs_bsize)) panic("ffs_reallocblks: unallocated block 1"); for (i = 1; i < len; i++) if (buflist->bs_children[i]->b_lblkno != start_lbn + i) panic("ffs_reallocblks: non-logical cluster"); blkno = buflist->bs_children[0]->b_blkno; ssize = fsbtodb(fs, fs->fs_frag); for (i = 1; i < len - 1; i++) if (buflist->bs_children[i]->b_blkno != blkno + (i * ssize)) panic("ffs_reallocblks: non-physical cluster %d", i); #endif /* * If the cluster crosses the boundary for the first indirect * block, do not move anything in it. Indirect blocks are * usually initially laid out in a position between the data * blocks. Block reallocation would usually destroy locality by * moving the indirect block out of the way to make room for * data blocks if we didn't compensate here. We should also do * this for other indirect block boundaries, but it is only * important for the first one. */ if (start_lbn < UFS_NDADDR && end_lbn >= UFS_NDADDR) return (ENOSPC); /* * If the latest allocation is in a new cylinder group, assume that * the filesystem has decided to move and do not force it back to * the previous cylinder group. */ if (dtog(fs, dbtofsb(fs, buflist->bs_children[0]->b_blkno)) != dtog(fs, dbtofsb(fs, buflist->bs_children[len - 1]->b_blkno))) return (ENOSPC); if (ufs_getlbns(vp, start_lbn, start_ap, &start_lvl) || ufs_getlbns(vp, end_lbn, end_ap, &end_lvl)) return (ENOSPC); /* * Get the starting offset and block map for the first block. */ if (start_lvl == 0) { sbap = &ip->i_din2->di_db[0]; soff = start_lbn; } else { idp = &start_ap[start_lvl - 1]; if (bread(vp, idp->in_lbn, (int)fs->fs_bsize, NOCRED, &sbp)) { brelse(sbp); return (ENOSPC); } sbap = (ufs2_daddr_t *)sbp->b_data; soff = idp->in_off; } /* * If the block range spans two block maps, get the second map. */ ebap = NULL; if (end_lvl == 0 || (idp = &end_ap[end_lvl - 1])->in_off + 1 >= len) { ssize = len; } else { #ifdef INVARIANTS if (start_lvl > 0 && start_ap[start_lvl - 1].in_lbn == idp->in_lbn) panic("ffs_reallocblk: start == end"); #endif ssize = len - (idp->in_off + 1); if (bread(vp, idp->in_lbn, (int)fs->fs_bsize, NOCRED, &ebp)) goto fail; ebap = (ufs2_daddr_t *)ebp->b_data; } /* * Find the preferred location for the cluster. If we have not * previously failed at this endeavor, then follow our standard * preference calculation. If we have failed at it, then pick up * where we last ended our search. */ UFS_LOCK(ump); if (ip->i_nextclustercg == -1) pref = ffs_blkpref_ufs2(ip, start_lbn, soff, sbap); else pref = cgdata(fs, ip->i_nextclustercg); /* * Search the block map looking for an allocation of the desired size. * To avoid wasting too much time, we limit the number of cylinder * groups that we will search. */ cg = dtog(fs, pref); for (i = min(maxclustersearch, fs->fs_ncg); i > 0; i--) { if ((newblk = ffs_clusteralloc(ip, cg, pref, len)) != 0) break; cg += 1; if (cg >= fs->fs_ncg) cg = 0; } /* * If we have failed in our search, record where we gave up for * next time. Otherwise, fall back to our usual search citerion. */ if (newblk == 0) { ip->i_nextclustercg = cg; UFS_UNLOCK(ump); goto fail; } ip->i_nextclustercg = -1; /* * We have found a new contiguous block. * * First we have to replace the old block pointers with the new * block pointers in the inode and indirect blocks associated * with the file. */ #ifdef DEBUG if (prtrealloc) printf("realloc: ino %ju, lbns %jd-%jd\n\told:", (uintmax_t)ip->i_number, (intmax_t)start_lbn, (intmax_t)end_lbn); #endif blkno = newblk; for (bap = &sbap[soff], i = 0; i < len; i++, blkno += fs->fs_frag) { if (i == ssize) { bap = ebap; soff = -i; } #ifdef INVARIANTS if (!ffs_checkblk(ip, dbtofsb(fs, buflist->bs_children[i]->b_blkno), fs->fs_bsize)) panic("ffs_reallocblks: unallocated block 2"); if (dbtofsb(fs, buflist->bs_children[i]->b_blkno) != *bap) panic("ffs_reallocblks: alloc mismatch"); #endif #ifdef DEBUG if (prtrealloc) printf(" %jd,", (intmax_t)*bap); #endif if (DOINGSOFTDEP(vp)) { if (sbap == &ip->i_din2->di_db[0] && i < ssize) softdep_setup_allocdirect(ip, start_lbn + i, blkno, *bap, fs->fs_bsize, fs->fs_bsize, buflist->bs_children[i]); else softdep_setup_allocindir_page(ip, start_lbn + i, i < ssize ? sbp : ebp, soff + i, blkno, *bap, buflist->bs_children[i]); } *bap++ = blkno; } /* * Next we must write out the modified inode and indirect blocks. * For strict correctness, the writes should be synchronous since * the old block values may have been written to disk. In practise * they are almost never written, but if we are concerned about * strict correctness, the `doasyncfree' flag should be set to zero. * * The test on `doasyncfree' should be changed to test a flag * that shows whether the associated buffers and inodes have * been written. The flag should be set when the cluster is * started and cleared whenever the buffer or inode is flushed. * We can then check below to see if it is set, and do the * synchronous write only when it has been cleared. */ if (sbap != &ip->i_din2->di_db[0]) { if (doasyncfree) bdwrite(sbp); else bwrite(sbp); } else { ip->i_flag |= IN_CHANGE | IN_UPDATE; if (!doasyncfree) ffs_update(vp, 1); } if (ssize < len) { if (doasyncfree) bdwrite(ebp); else bwrite(ebp); } /* * Last, free the old blocks and assign the new blocks to the buffers. */ #ifdef DEBUG if (prtrealloc) printf("\n\tnew:"); #endif for (blkno = newblk, i = 0; i < len; i++, blkno += fs->fs_frag) { bp = buflist->bs_children[i]; if (!DOINGSOFTDEP(vp)) /* * The usual case is that a set of N-contiguous blocks * that was just allocated has been replaced with a * set of N+1-contiguous blocks. If they are marked as * B_DELWRI, the current contents have not been written * to disk. It is possible that the blocks were written * earlier, but very uncommon. If the blocks have never * been written, there is no need to send a BIO_DELETE * for them when they are freed. The gain from avoiding * the TRIMs for the common case of unwritten blocks * far exceeds the cost of the write amplification for * the uncommon case of failing to send a TRIM for the * blocks that had been written. */ ffs_blkfree(ump, fs, ump->um_devvp, dbtofsb(fs, bp->b_blkno), fs->fs_bsize, ip->i_number, vp->v_type, NULL, (bp->b_flags & B_DELWRI) != 0 ? NOTRIM_KEY : SINGLETON_KEY); bp->b_blkno = fsbtodb(fs, blkno); #ifdef INVARIANTS if (!ffs_checkblk(ip, dbtofsb(fs, bp->b_blkno), fs->fs_bsize)) panic("ffs_reallocblks: unallocated block 3"); #endif #ifdef DEBUG if (prtrealloc) printf(" %jd,", (intmax_t)blkno); #endif } #ifdef DEBUG if (prtrealloc) { prtrealloc--; printf("\n"); } #endif return (0); fail: if (ssize < len) brelse(ebp); if (sbap != &ip->i_din2->di_db[0]) brelse(sbp); return (ENOSPC); } /* * Allocate an inode in the filesystem. * * If allocating a directory, use ffs_dirpref to select the inode. * If allocating in a directory, the following hierarchy is followed: * 1) allocate the preferred inode. * 2) allocate an inode in the same cylinder group. * 3) quadradically rehash into other cylinder groups, until an * available inode is located. * If no inode preference is given the following hierarchy is used * to allocate an inode: * 1) allocate an inode in cylinder group 0. * 2) quadradically rehash into other cylinder groups, until an * available inode is located. */ int ffs_valloc(pvp, mode, cred, vpp) struct vnode *pvp; int mode; struct ucred *cred; struct vnode **vpp; { struct inode *pip; struct fs *fs; struct inode *ip; struct timespec ts; struct ufsmount *ump; ino_t ino, ipref; u_int cg; int error, error1, reclaimed; static struct timeval lastfail; static int curfail; *vpp = NULL; pip = VTOI(pvp); ump = ITOUMP(pip); fs = ump->um_fs; UFS_LOCK(ump); reclaimed = 0; retry: if (fs->fs_cstotal.cs_nifree == 0) goto noinodes; if ((mode & IFMT) == IFDIR) ipref = ffs_dirpref(pip); else ipref = pip->i_number; if (ipref >= fs->fs_ncg * fs->fs_ipg) ipref = 0; cg = ino_to_cg(fs, ipref); /* * Track number of dirs created one after another * in a same cg without intervening by files. */ if ((mode & IFMT) == IFDIR) { if (fs->fs_contigdirs[cg] < 255) fs->fs_contigdirs[cg]++; } else { if (fs->fs_contigdirs[cg] > 0) fs->fs_contigdirs[cg]--; } ino = (ino_t)ffs_hashalloc(pip, cg, ipref, mode, 0, (allocfcn_t *)ffs_nodealloccg); if (ino == 0) goto noinodes; error = ffs_vget(pvp->v_mount, ino, LK_EXCLUSIVE, vpp); if (error) { error1 = ffs_vgetf(pvp->v_mount, ino, LK_EXCLUSIVE, vpp, FFSV_FORCEINSMQ); ffs_vfree(pvp, ino, mode); if (error1 == 0) { ip = VTOI(*vpp); if (ip->i_mode) goto dup_alloc; ip->i_flag |= IN_MODIFIED; vput(*vpp); } return (error); } ip = VTOI(*vpp); if (ip->i_mode) { dup_alloc: printf("mode = 0%o, inum = %ju, fs = %s\n", ip->i_mode, (uintmax_t)ip->i_number, fs->fs_fsmnt); panic("ffs_valloc: dup alloc"); } if (DIP(ip, i_blocks) && (fs->fs_flags & FS_UNCLEAN) == 0) { /* XXX */ printf("free inode %s/%lu had %ld blocks\n", fs->fs_fsmnt, (u_long)ino, (long)DIP(ip, i_blocks)); DIP_SET(ip, i_blocks, 0); } ip->i_flags = 0; DIP_SET(ip, i_flags, 0); /* * Set up a new generation number for this inode. */ while (ip->i_gen == 0 || ++ip->i_gen == 0) ip->i_gen = arc4random(); DIP_SET(ip, i_gen, ip->i_gen); if (fs->fs_magic == FS_UFS2_MAGIC) { vfs_timestamp(&ts); ip->i_din2->di_birthtime = ts.tv_sec; ip->i_din2->di_birthnsec = ts.tv_nsec; } ufs_prepare_reclaim(*vpp); ip->i_flag = 0; (*vpp)->v_vflag = 0; (*vpp)->v_type = VNON; if (fs->fs_magic == FS_UFS2_MAGIC) { (*vpp)->v_op = &ffs_vnodeops2; ip->i_flag |= IN_UFS2; } else { (*vpp)->v_op = &ffs_vnodeops1; } return (0); noinodes: if (reclaimed == 0) { reclaimed = 1; softdep_request_cleanup(fs, pvp, cred, FLUSH_INODES_WAIT); goto retry; } UFS_UNLOCK(ump); if (ppsratecheck(&lastfail, &curfail, 1)) { ffs_fserr(fs, pip->i_number, "out of inodes"); uprintf("\n%s: create/symlink failed, no inodes free\n", fs->fs_fsmnt); } return (ENOSPC); } /* * Find a cylinder group to place a directory. * * The policy implemented by this algorithm is to allocate a * directory inode in the same cylinder group as its parent * directory, but also to reserve space for its files inodes * and data. Restrict the number of directories which may be * allocated one after another in the same cylinder group * without intervening allocation of files. * * If we allocate a first level directory then force allocation * in another cylinder group. */ static ino_t ffs_dirpref(pip) struct inode *pip; { struct fs *fs; int cg, prefcg, dirsize, cgsize; u_int avgifree, avgbfree, avgndir, curdirsize; u_int minifree, minbfree, maxndir; u_int mincg, minndir; u_int maxcontigdirs; mtx_assert(UFS_MTX(ITOUMP(pip)), MA_OWNED); fs = ITOFS(pip); avgifree = fs->fs_cstotal.cs_nifree / fs->fs_ncg; avgbfree = fs->fs_cstotal.cs_nbfree / fs->fs_ncg; avgndir = fs->fs_cstotal.cs_ndir / fs->fs_ncg; /* * Force allocation in another cg if creating a first level dir. */ ASSERT_VOP_LOCKED(ITOV(pip), "ffs_dirpref"); if (ITOV(pip)->v_vflag & VV_ROOT) { prefcg = arc4random() % fs->fs_ncg; mincg = prefcg; minndir = fs->fs_ipg; for (cg = prefcg; cg < fs->fs_ncg; cg++) if (fs->fs_cs(fs, cg).cs_ndir < minndir && fs->fs_cs(fs, cg).cs_nifree >= avgifree && fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { mincg = cg; minndir = fs->fs_cs(fs, cg).cs_ndir; } for (cg = 0; cg < prefcg; cg++) if (fs->fs_cs(fs, cg).cs_ndir < minndir && fs->fs_cs(fs, cg).cs_nifree >= avgifree && fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { mincg = cg; minndir = fs->fs_cs(fs, cg).cs_ndir; } return ((ino_t)(fs->fs_ipg * mincg)); } /* * Count various limits which used for * optimal allocation of a directory inode. */ maxndir = min(avgndir + fs->fs_ipg / 16, fs->fs_ipg); minifree = avgifree - avgifree / 4; if (minifree < 1) minifree = 1; minbfree = avgbfree - avgbfree / 4; if (minbfree < 1) minbfree = 1; cgsize = fs->fs_fsize * fs->fs_fpg; dirsize = fs->fs_avgfilesize * fs->fs_avgfpdir; curdirsize = avgndir ? (cgsize - avgbfree * fs->fs_bsize) / avgndir : 0; if (dirsize < curdirsize) dirsize = curdirsize; if (dirsize <= 0) maxcontigdirs = 0; /* dirsize overflowed */ else maxcontigdirs = min((avgbfree * fs->fs_bsize) / dirsize, 255); if (fs->fs_avgfpdir > 0) maxcontigdirs = min(maxcontigdirs, fs->fs_ipg / fs->fs_avgfpdir); if (maxcontigdirs == 0) maxcontigdirs = 1; /* * Limit number of dirs in one cg and reserve space for * regular files, but only if we have no deficit in * inodes or space. * * We are trying to find a suitable cylinder group nearby * our preferred cylinder group to place a new directory. * We scan from our preferred cylinder group forward looking * for a cylinder group that meets our criterion. If we get * to the final cylinder group and do not find anything, * we start scanning forwards from the beginning of the * filesystem. While it might seem sensible to start scanning * backwards or even to alternate looking forward and backward, * this approach fails badly when the filesystem is nearly full. * Specifically, we first search all the areas that have no space * and finally try the one preceding that. We repeat this on * every request and in the case of the final block end up * searching the entire filesystem. By jumping to the front * of the filesystem, our future forward searches always look * in new cylinder groups so finds every possible block after * one pass over the filesystem. */ prefcg = ino_to_cg(fs, pip->i_number); for (cg = prefcg; cg < fs->fs_ncg; cg++) if (fs->fs_cs(fs, cg).cs_ndir < maxndir && fs->fs_cs(fs, cg).cs_nifree >= minifree && fs->fs_cs(fs, cg).cs_nbfree >= minbfree) { if (fs->fs_contigdirs[cg] < maxcontigdirs) return ((ino_t)(fs->fs_ipg * cg)); } for (cg = 0; cg < prefcg; cg++) if (fs->fs_cs(fs, cg).cs_ndir < maxndir && fs->fs_cs(fs, cg).cs_nifree >= minifree && fs->fs_cs(fs, cg).cs_nbfree >= minbfree) { if (fs->fs_contigdirs[cg] < maxcontigdirs) return ((ino_t)(fs->fs_ipg * cg)); } /* * This is a backstop when we have deficit in space. */ for (cg = prefcg; cg < fs->fs_ncg; cg++) if (fs->fs_cs(fs, cg).cs_nifree >= avgifree) return ((ino_t)(fs->fs_ipg * cg)); for (cg = 0; cg < prefcg; cg++) if (fs->fs_cs(fs, cg).cs_nifree >= avgifree) break; return ((ino_t)(fs->fs_ipg * cg)); } /* * Select the desired position for the next block in a file. The file is * logically divided into sections. The first section is composed of the * direct blocks and the next fs_maxbpg blocks. Each additional section * contains fs_maxbpg blocks. * * If no blocks have been allocated in the first section, the policy is to * request a block in the same cylinder group as the inode that describes * the file. The first indirect is allocated immediately following the last * direct block and the data blocks for the first indirect immediately * follow it. * * If no blocks have been allocated in any other section, the indirect * block(s) are allocated in the same cylinder group as its inode in an * area reserved immediately following the inode blocks. The policy for * the data blocks is to place them in a cylinder group with a greater than * average number of free blocks. An appropriate cylinder group is found * by using a rotor that sweeps the cylinder groups. When a new group of * blocks is needed, the sweep begins in the cylinder group following the * cylinder group from which the previous allocation was made. The sweep * continues until a cylinder group with greater than the average number * of free blocks is found. If the allocation is for the first block in an * indirect block or the previous block is a hole, then the information on * the previous allocation is unavailable; here a best guess is made based * on the logical block number being allocated. * * If a section is already partially allocated, the policy is to * allocate blocks contiguously within the section if possible. */ ufs2_daddr_t ffs_blkpref_ufs1(ip, lbn, indx, bap) struct inode *ip; ufs_lbn_t lbn; int indx; ufs1_daddr_t *bap; { struct fs *fs; u_int cg, inocg; u_int avgbfree, startcg; ufs2_daddr_t pref; KASSERT(indx <= 0 || bap != NULL, ("need non-NULL bap")); mtx_assert(UFS_MTX(ITOUMP(ip)), MA_OWNED); fs = ITOFS(ip); /* * Allocation of indirect blocks is indicated by passing negative * values in indx: -1 for single indirect, -2 for double indirect, * -3 for triple indirect. As noted below, we attempt to allocate * the first indirect inline with the file data. For all later * indirect blocks, the data is often allocated in other cylinder * groups. However to speed random file access and to speed up * fsck, the filesystem reserves the first fs_metaspace blocks * (typically half of fs_minfree) of the data area of each cylinder * group to hold these later indirect blocks. */ inocg = ino_to_cg(fs, ip->i_number); if (indx < 0) { /* * Our preference for indirect blocks is the zone at the * beginning of the inode's cylinder group data area that * we try to reserve for indirect blocks. */ pref = cgmeta(fs, inocg); /* * If we are allocating the first indirect block, try to * place it immediately following the last direct block. */ if (indx == -1 && lbn < UFS_NDADDR + NINDIR(fs) && ip->i_din1->di_db[UFS_NDADDR - 1] != 0) pref = ip->i_din1->di_db[UFS_NDADDR - 1] + fs->fs_frag; return (pref); } /* * If we are allocating the first data block in the first indirect * block and the indirect has been allocated in the data block area, * try to place it immediately following the indirect block. */ if (lbn == UFS_NDADDR) { pref = ip->i_din1->di_ib[0]; if (pref != 0 && pref >= cgdata(fs, inocg) && pref < cgbase(fs, inocg + 1)) return (pref + fs->fs_frag); } /* * If we are at the beginning of a file, or we have already allocated * the maximum number of blocks per cylinder group, or we do not * have a block allocated immediately preceding us, then we need * to decide where to start allocating new blocks. */ if (indx % fs->fs_maxbpg == 0 || bap[indx - 1] == 0) { /* * If we are allocating a directory data block, we want * to place it in the metadata area. */ if ((ip->i_mode & IFMT) == IFDIR) return (cgmeta(fs, inocg)); /* * Until we fill all the direct and all the first indirect's * blocks, we try to allocate in the data area of the inode's * cylinder group. */ if (lbn < UFS_NDADDR + NINDIR(fs)) return (cgdata(fs, inocg)); /* * Find a cylinder with greater than average number of * unused data blocks. */ if (indx == 0 || bap[indx - 1] == 0) startcg = inocg + lbn / fs->fs_maxbpg; else startcg = dtog(fs, bap[indx - 1]) + 1; startcg %= fs->fs_ncg; avgbfree = fs->fs_cstotal.cs_nbfree / fs->fs_ncg; for (cg = startcg; cg < fs->fs_ncg; cg++) if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { fs->fs_cgrotor = cg; return (cgdata(fs, cg)); } for (cg = 0; cg <= startcg; cg++) if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { fs->fs_cgrotor = cg; return (cgdata(fs, cg)); } return (0); } /* * Otherwise, we just always try to lay things out contiguously. */ return (bap[indx - 1] + fs->fs_frag); } /* * Same as above, but for UFS2 */ ufs2_daddr_t ffs_blkpref_ufs2(ip, lbn, indx, bap) struct inode *ip; ufs_lbn_t lbn; int indx; ufs2_daddr_t *bap; { struct fs *fs; u_int cg, inocg; u_int avgbfree, startcg; ufs2_daddr_t pref; KASSERT(indx <= 0 || bap != NULL, ("need non-NULL bap")); mtx_assert(UFS_MTX(ITOUMP(ip)), MA_OWNED); fs = ITOFS(ip); /* * Allocation of indirect blocks is indicated by passing negative * values in indx: -1 for single indirect, -2 for double indirect, * -3 for triple indirect. As noted below, we attempt to allocate * the first indirect inline with the file data. For all later * indirect blocks, the data is often allocated in other cylinder * groups. However to speed random file access and to speed up * fsck, the filesystem reserves the first fs_metaspace blocks * (typically half of fs_minfree) of the data area of each cylinder * group to hold these later indirect blocks. */ inocg = ino_to_cg(fs, ip->i_number); if (indx < 0) { /* * Our preference for indirect blocks is the zone at the * beginning of the inode's cylinder group data area that * we try to reserve for indirect blocks. */ pref = cgmeta(fs, inocg); /* * If we are allocating the first indirect block, try to * place it immediately following the last direct block. */ if (indx == -1 && lbn < UFS_NDADDR + NINDIR(fs) && ip->i_din2->di_db[UFS_NDADDR - 1] != 0) pref = ip->i_din2->di_db[UFS_NDADDR - 1] + fs->fs_frag; return (pref); } /* * If we are allocating the first data block in the first indirect * block and the indirect has been allocated in the data block area, * try to place it immediately following the indirect block. */ if (lbn == UFS_NDADDR) { pref = ip->i_din2->di_ib[0]; if (pref != 0 && pref >= cgdata(fs, inocg) && pref < cgbase(fs, inocg + 1)) return (pref + fs->fs_frag); } /* * If we are at the beginning of a file, or we have already allocated * the maximum number of blocks per cylinder group, or we do not * have a block allocated immediately preceding us, then we need * to decide where to start allocating new blocks. */ if (indx % fs->fs_maxbpg == 0 || bap[indx - 1] == 0) { /* * If we are allocating a directory data block, we want * to place it in the metadata area. */ if ((ip->i_mode & IFMT) == IFDIR) return (cgmeta(fs, inocg)); /* * Until we fill all the direct and all the first indirect's * blocks, we try to allocate in the data area of the inode's * cylinder group. */ if (lbn < UFS_NDADDR + NINDIR(fs)) return (cgdata(fs, inocg)); /* * Find a cylinder with greater than average number of * unused data blocks. */ if (indx == 0 || bap[indx - 1] == 0) startcg = inocg + lbn / fs->fs_maxbpg; else startcg = dtog(fs, bap[indx - 1]) + 1; startcg %= fs->fs_ncg; avgbfree = fs->fs_cstotal.cs_nbfree / fs->fs_ncg; for (cg = startcg; cg < fs->fs_ncg; cg++) if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { fs->fs_cgrotor = cg; return (cgdata(fs, cg)); } for (cg = 0; cg <= startcg; cg++) if (fs->fs_cs(fs, cg).cs_nbfree >= avgbfree) { fs->fs_cgrotor = cg; return (cgdata(fs, cg)); } return (0); } /* * Otherwise, we just always try to lay things out contiguously. */ return (bap[indx - 1] + fs->fs_frag); } /* * Implement the cylinder overflow algorithm. * * The policy implemented by this algorithm is: * 1) allocate the block in its requested cylinder group. * 2) quadradically rehash on the cylinder group number. * 3) brute force search for a free block. * * Must be called with the UFS lock held. Will release the lock on success * and return with it held on failure. */ /*VARARGS5*/ static ufs2_daddr_t ffs_hashalloc(ip, cg, pref, size, rsize, allocator) struct inode *ip; u_int cg; ufs2_daddr_t pref; int size; /* Search size for data blocks, mode for inodes */ int rsize; /* Real allocated size. */ allocfcn_t *allocator; { struct fs *fs; ufs2_daddr_t result; u_int i, icg = cg; mtx_assert(UFS_MTX(ITOUMP(ip)), MA_OWNED); #ifdef INVARIANTS if (ITOV(ip)->v_mount->mnt_kern_flag & MNTK_SUSPENDED) panic("ffs_hashalloc: allocation on suspended filesystem"); #endif fs = ITOFS(ip); /* * 1: preferred cylinder group */ result = (*allocator)(ip, cg, pref, size, rsize); if (result) return (result); /* * 2: quadratic rehash */ for (i = 1; i < fs->fs_ncg; i *= 2) { cg += i; if (cg >= fs->fs_ncg) cg -= fs->fs_ncg; result = (*allocator)(ip, cg, 0, size, rsize); if (result) return (result); } /* * 3: brute force search * Note that we start at i == 2, since 0 was checked initially, * and 1 is always checked in the quadratic rehash. */ cg = (icg + 2) % fs->fs_ncg; for (i = 2; i < fs->fs_ncg; i++) { result = (*allocator)(ip, cg, 0, size, rsize); if (result) return (result); cg++; if (cg == fs->fs_ncg) cg = 0; } return (0); } /* * Determine whether a fragment can be extended. * * Check to see if the necessary fragments are available, and * if they are, allocate them. */ static ufs2_daddr_t ffs_fragextend(ip, cg, bprev, osize, nsize) struct inode *ip; u_int cg; ufs2_daddr_t bprev; int osize, nsize; { struct fs *fs; struct cg *cgp; struct buf *bp; struct ufsmount *ump; int nffree; long bno; int frags, bbase; int i, error; u_int8_t *blksfree; ump = ITOUMP(ip); fs = ump->um_fs; if (fs->fs_cs(fs, cg).cs_nffree < numfrags(fs, nsize - osize)) return (0); frags = numfrags(fs, nsize); bbase = fragnum(fs, bprev); if (bbase > fragnum(fs, (bprev + frags - 1))) { /* cannot extend across a block boundary */ return (0); } UFS_UNLOCK(ump); if ((error = ffs_getcg(fs, ump->um_devvp, cg, &bp, &cgp)) != 0) goto fail; bno = dtogd(fs, bprev); blksfree = cg_blksfree(cgp); for (i = numfrags(fs, osize); i < frags; i++) if (isclr(blksfree, bno + i)) goto fail; /* * the current fragment can be extended * deduct the count on fragment being extended into * increase the count on the remaining fragment (if any) * allocate the extended piece */ for (i = frags; i < fs->fs_frag - bbase; i++) if (isclr(blksfree, bno + i)) break; cgp->cg_frsum[i - numfrags(fs, osize)]--; if (i != frags) cgp->cg_frsum[i - frags]++; for (i = numfrags(fs, osize), nffree = 0; i < frags; i++) { clrbit(blksfree, bno + i); cgp->cg_cs.cs_nffree--; nffree++; } UFS_LOCK(ump); fs->fs_cstotal.cs_nffree -= nffree; fs->fs_cs(fs, cg).cs_nffree -= nffree; fs->fs_fmod = 1; ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); if (DOINGSOFTDEP(ITOV(ip))) softdep_setup_blkmapdep(bp, UFSTOVFS(ump), bprev, frags, numfrags(fs, osize)); bdwrite(bp); return (bprev); fail: brelse(bp); UFS_LOCK(ump); return (0); } /* * Determine whether a block can be allocated. * * Check to see if a block of the appropriate size is available, * and if it is, allocate it. */ static ufs2_daddr_t ffs_alloccg(ip, cg, bpref, size, rsize) struct inode *ip; u_int cg; ufs2_daddr_t bpref; int size; int rsize; { struct fs *fs; struct cg *cgp; struct buf *bp; struct ufsmount *ump; ufs1_daddr_t bno; ufs2_daddr_t blkno; int i, allocsiz, error, frags; u_int8_t *blksfree; ump = ITOUMP(ip); fs = ump->um_fs; if (fs->fs_cs(fs, cg).cs_nbfree == 0 && size == fs->fs_bsize) return (0); UFS_UNLOCK(ump); if ((error = ffs_getcg(fs, ump->um_devvp, cg, &bp, &cgp)) != 0 || (cgp->cg_cs.cs_nbfree == 0 && size == fs->fs_bsize)) goto fail; if (size == fs->fs_bsize) { UFS_LOCK(ump); blkno = ffs_alloccgblk(ip, bp, bpref, rsize); ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); bdwrite(bp); return (blkno); } /* * check to see if any fragments are already available * allocsiz is the size which will be allocated, hacking * it down to a smaller size if necessary */ blksfree = cg_blksfree(cgp); frags = numfrags(fs, size); for (allocsiz = frags; allocsiz < fs->fs_frag; allocsiz++) if (cgp->cg_frsum[allocsiz] != 0) break; if (allocsiz == fs->fs_frag) { /* * no fragments were available, so a block will be * allocated, and hacked up */ if (cgp->cg_cs.cs_nbfree == 0) goto fail; UFS_LOCK(ump); blkno = ffs_alloccgblk(ip, bp, bpref, rsize); ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); bdwrite(bp); return (blkno); } KASSERT(size == rsize, ("ffs_alloccg: size(%d) != rsize(%d)", size, rsize)); bno = ffs_mapsearch(fs, cgp, bpref, allocsiz); if (bno < 0) goto fail; for (i = 0; i < frags; i++) clrbit(blksfree, bno + i); cgp->cg_cs.cs_nffree -= frags; cgp->cg_frsum[allocsiz]--; if (frags != allocsiz) cgp->cg_frsum[allocsiz - frags]++; UFS_LOCK(ump); fs->fs_cstotal.cs_nffree -= frags; fs->fs_cs(fs, cg).cs_nffree -= frags; fs->fs_fmod = 1; blkno = cgbase(fs, cg) + bno; ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); if (DOINGSOFTDEP(ITOV(ip))) softdep_setup_blkmapdep(bp, UFSTOVFS(ump), blkno, frags, 0); bdwrite(bp); return (blkno); fail: brelse(bp); UFS_LOCK(ump); return (0); } /* * Allocate a block in a cylinder group. * * This algorithm implements the following policy: * 1) allocate the requested block. * 2) allocate a rotationally optimal block in the same cylinder. * 3) allocate the next available block on the block rotor for the * specified cylinder group. * Note that this routine only allocates fs_bsize blocks; these * blocks may be fragmented by the routine that allocates them. */ static ufs2_daddr_t ffs_alloccgblk(ip, bp, bpref, size) struct inode *ip; struct buf *bp; ufs2_daddr_t bpref; int size; { struct fs *fs; struct cg *cgp; struct ufsmount *ump; ufs1_daddr_t bno; ufs2_daddr_t blkno; u_int8_t *blksfree; int i, cgbpref; ump = ITOUMP(ip); fs = ump->um_fs; mtx_assert(UFS_MTX(ump), MA_OWNED); cgp = (struct cg *)bp->b_data; blksfree = cg_blksfree(cgp); if (bpref == 0) { bpref = cgbase(fs, cgp->cg_cgx) + cgp->cg_rotor + fs->fs_frag; } else if ((cgbpref = dtog(fs, bpref)) != cgp->cg_cgx) { /* map bpref to correct zone in this cg */ if (bpref < cgdata(fs, cgbpref)) bpref = cgmeta(fs, cgp->cg_cgx); else bpref = cgdata(fs, cgp->cg_cgx); } /* * if the requested block is available, use it */ bno = dtogd(fs, blknum(fs, bpref)); if (ffs_isblock(fs, blksfree, fragstoblks(fs, bno))) goto gotit; /* * Take the next available block in this cylinder group. */ bno = ffs_mapsearch(fs, cgp, bpref, (int)fs->fs_frag); if (bno < 0) return (0); /* Update cg_rotor only if allocated from the data zone */ if (bno >= dtogd(fs, cgdata(fs, cgp->cg_cgx))) cgp->cg_rotor = bno; gotit: blkno = fragstoblks(fs, bno); ffs_clrblock(fs, blksfree, (long)blkno); ffs_clusteracct(fs, cgp, blkno, -1); cgp->cg_cs.cs_nbfree--; fs->fs_cstotal.cs_nbfree--; fs->fs_cs(fs, cgp->cg_cgx).cs_nbfree--; fs->fs_fmod = 1; blkno = cgbase(fs, cgp->cg_cgx) + bno; /* * If the caller didn't want the whole block free the frags here. */ size = numfrags(fs, size); if (size != fs->fs_frag) { bno = dtogd(fs, blkno); for (i = size; i < fs->fs_frag; i++) setbit(blksfree, bno + i); i = fs->fs_frag - size; cgp->cg_cs.cs_nffree += i; fs->fs_cstotal.cs_nffree += i; fs->fs_cs(fs, cgp->cg_cgx).cs_nffree += i; fs->fs_fmod = 1; cgp->cg_frsum[i]++; } /* XXX Fixme. */ UFS_UNLOCK(ump); if (DOINGSOFTDEP(ITOV(ip))) softdep_setup_blkmapdep(bp, UFSTOVFS(ump), blkno, size, 0); UFS_LOCK(ump); return (blkno); } /* * Determine whether a cluster can be allocated. * * We do not currently check for optimal rotational layout if there * are multiple choices in the same cylinder group. Instead we just * take the first one that we find following bpref. */ static ufs2_daddr_t ffs_clusteralloc(ip, cg, bpref, len) struct inode *ip; u_int cg; ufs2_daddr_t bpref; int len; { struct fs *fs; struct cg *cgp; struct buf *bp; struct ufsmount *ump; int i, run, bit, map, got, error; ufs2_daddr_t bno; u_char *mapp; int32_t *lp; u_int8_t *blksfree; ump = ITOUMP(ip); fs = ump->um_fs; if (fs->fs_maxcluster[cg] < len) return (0); UFS_UNLOCK(ump); if ((error = ffs_getcg(fs, ump->um_devvp, cg, &bp, &cgp)) != 0) { UFS_LOCK(ump); return (0); } /* * Check to see if a cluster of the needed size (or bigger) is * available in this cylinder group. */ lp = &cg_clustersum(cgp)[len]; for (i = len; i <= fs->fs_contigsumsize; i++) if (*lp++ > 0) break; if (i > fs->fs_contigsumsize) { /* * This is the first time looking for a cluster in this * cylinder group. Update the cluster summary information * to reflect the true maximum sized cluster so that * future cluster allocation requests can avoid reading * the cylinder group map only to find no clusters. */ lp = &cg_clustersum(cgp)[len - 1]; for (i = len - 1; i > 0; i--) if (*lp-- > 0) break; UFS_LOCK(ump); fs->fs_maxcluster[cg] = i; brelse(bp); return (0); } /* * Search the cluster map to find a big enough cluster. * We take the first one that we find, even if it is larger * than we need as we prefer to get one close to the previous * block allocation. We do not search before the current * preference point as we do not want to allocate a block * that is allocated before the previous one (as we will * then have to wait for another pass of the elevator * algorithm before it will be read). We prefer to fail and * be recalled to try an allocation in the next cylinder group. */ if (dtog(fs, bpref) != cg) bpref = cgdata(fs, cg); else bpref = blknum(fs, bpref); bpref = fragstoblks(fs, dtogd(fs, bpref)); mapp = &cg_clustersfree(cgp)[bpref / NBBY]; map = *mapp++; bit = 1 << (bpref % NBBY); for (run = 0, got = bpref; got < cgp->cg_nclusterblks; got++) { if ((map & bit) == 0) { run = 0; } else { run++; if (run == len) break; } if ((got & (NBBY - 1)) != (NBBY - 1)) { bit <<= 1; } else { map = *mapp++; bit = 1; } } if (got >= cgp->cg_nclusterblks) { UFS_LOCK(ump); brelse(bp); return (0); } /* * Allocate the cluster that we have found. */ blksfree = cg_blksfree(cgp); for (i = 1; i <= len; i++) if (!ffs_isblock(fs, blksfree, got - run + i)) panic("ffs_clusteralloc: map mismatch"); bno = cgbase(fs, cg) + blkstofrags(fs, got - run + 1); if (dtog(fs, bno) != cg) panic("ffs_clusteralloc: allocated out of group"); len = blkstofrags(fs, len); UFS_LOCK(ump); for (i = 0; i < len; i += fs->fs_frag) if (ffs_alloccgblk(ip, bp, bno + i, fs->fs_bsize) != bno + i) panic("ffs_clusteralloc: lost block"); ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); bdwrite(bp); return (bno); } static inline struct buf * getinobuf(struct inode *ip, u_int cg, u_int32_t cginoblk, int gbflags) { struct fs *fs; fs = ITOFS(ip); return (getblk(ITODEVVP(ip), fsbtodb(fs, ino_to_fsba(fs, cg * fs->fs_ipg + cginoblk)), (int)fs->fs_bsize, 0, 0, gbflags)); } /* * Synchronous inode initialization is needed only when barrier writes do not * work as advertised, and will impose a heavy cost on file creation in a newly * created filesystem. */ static int doasyncinodeinit = 1; SYSCTL_INT(_vfs_ffs, OID_AUTO, doasyncinodeinit, CTLFLAG_RWTUN, &doasyncinodeinit, 0, "Perform inode block initialization using asynchronous writes"); /* * Determine whether an inode can be allocated. * * Check to see if an inode is available, and if it is, * allocate it using the following policy: * 1) allocate the requested inode. * 2) allocate the next available inode after the requested * inode in the specified cylinder group. */ static ufs2_daddr_t ffs_nodealloccg(ip, cg, ipref, mode, unused) struct inode *ip; u_int cg; ufs2_daddr_t ipref; int mode; int unused; { struct fs *fs; struct cg *cgp; struct buf *bp, *ibp; struct ufsmount *ump; u_int8_t *inosused, *loc; struct ufs2_dinode *dp2; int error, start, len, i; u_int32_t old_initediblk; ump = ITOUMP(ip); fs = ump->um_fs; check_nifree: if (fs->fs_cs(fs, cg).cs_nifree == 0) return (0); UFS_UNLOCK(ump); if ((error = ffs_getcg(fs, ump->um_devvp, cg, &bp, &cgp)) != 0) { UFS_LOCK(ump); return (0); } restart: if (cgp->cg_cs.cs_nifree == 0) { brelse(bp); UFS_LOCK(ump); return (0); } inosused = cg_inosused(cgp); if (ipref) { ipref %= fs->fs_ipg; if (isclr(inosused, ipref)) goto gotit; } start = cgp->cg_irotor / NBBY; len = howmany(fs->fs_ipg - cgp->cg_irotor, NBBY); loc = memcchr(&inosused[start], 0xff, len); if (loc == NULL) { len = start + 1; start = 0; loc = memcchr(&inosused[start], 0xff, len); if (loc == NULL) { printf("cg = %d, irotor = %ld, fs = %s\n", cg, (long)cgp->cg_irotor, fs->fs_fsmnt); panic("ffs_nodealloccg: map corrupted"); /* NOTREACHED */ } } ipref = (loc - inosused) * NBBY + ffs(~*loc) - 1; gotit: /* * Check to see if we need to initialize more inodes. */ if (fs->fs_magic == FS_UFS2_MAGIC && ipref + INOPB(fs) > cgp->cg_initediblk && cgp->cg_initediblk < cgp->cg_niblk) { old_initediblk = cgp->cg_initediblk; /* * Free the cylinder group lock before writing the * initialized inode block. Entering the * babarrierwrite() with the cylinder group lock * causes lock order violation between the lock and * snaplk. * * Another thread can decide to initialize the same * inode block, but whichever thread first gets the * cylinder group lock after writing the newly * allocated inode block will update it and the other * will realize that it has lost and leave the * cylinder group unchanged. */ ibp = getinobuf(ip, cg, old_initediblk, GB_LOCK_NOWAIT); brelse(bp); if (ibp == NULL) { /* * The inode block buffer is already owned by * another thread, which must initialize it. * Wait on the buffer to allow another thread * to finish the updates, with dropped cg * buffer lock, then retry. */ ibp = getinobuf(ip, cg, old_initediblk, 0); brelse(ibp); UFS_LOCK(ump); goto check_nifree; } bzero(ibp->b_data, (int)fs->fs_bsize); dp2 = (struct ufs2_dinode *)(ibp->b_data); for (i = 0; i < INOPB(fs); i++) { while (dp2->di_gen == 0) dp2->di_gen = arc4random(); dp2++; } /* * Rather than adding a soft updates dependency to ensure * that the new inode block is written before it is claimed * by the cylinder group map, we just do a barrier write * here. The barrier write will ensure that the inode block * gets written before the updated cylinder group map can be * written. The barrier write should only slow down bulk * loading of newly created filesystems. */ if (doasyncinodeinit) babarrierwrite(ibp); else bwrite(ibp); /* * After the inode block is written, try to update the * cg initediblk pointer. If another thread beat us * to it, then leave it unchanged as the other thread * has already set it correctly. */ error = ffs_getcg(fs, ump->um_devvp, cg, &bp, &cgp); UFS_LOCK(ump); ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); if (error != 0) return (error); if (cgp->cg_initediblk == old_initediblk) cgp->cg_initediblk += INOPB(fs); goto restart; } cgp->cg_irotor = ipref; UFS_LOCK(ump); ACTIVECLEAR(fs, cg); setbit(inosused, ipref); cgp->cg_cs.cs_nifree--; fs->fs_cstotal.cs_nifree--; fs->fs_cs(fs, cg).cs_nifree--; fs->fs_fmod = 1; if ((mode & IFMT) == IFDIR) { cgp->cg_cs.cs_ndir++; fs->fs_cstotal.cs_ndir++; fs->fs_cs(fs, cg).cs_ndir++; } UFS_UNLOCK(ump); if (DOINGSOFTDEP(ITOV(ip))) softdep_setup_inomapdep(bp, ip, cg * fs->fs_ipg + ipref, mode); bdwrite(bp); return ((ino_t)(cg * fs->fs_ipg + ipref)); } /* * Free a block or fragment. * * The specified block or fragment is placed back in the * free map. If a fragment is deallocated, a possible * block reassembly is checked. */ static void ffs_blkfree_cg(ump, fs, devvp, bno, size, inum, dephd) struct ufsmount *ump; struct fs *fs; struct vnode *devvp; ufs2_daddr_t bno; long size; ino_t inum; struct workhead *dephd; { struct mount *mp; struct cg *cgp; struct buf *bp; ufs1_daddr_t fragno, cgbno; int i, blk, frags, bbase, error; u_int cg; u_int8_t *blksfree; struct cdev *dev; cg = dtog(fs, bno); if (devvp->v_type == VREG) { /* devvp is a snapshot */ MPASS(devvp->v_mount->mnt_data == ump); dev = ump->um_devvp->v_rdev; } else if (devvp->v_type == VCHR) { /* devvp is a normal disk device */ dev = devvp->v_rdev; ASSERT_VOP_LOCKED(devvp, "ffs_blkfree_cg"); } else return; #ifdef INVARIANTS if ((u_int)size > fs->fs_bsize || fragoff(fs, size) != 0 || fragnum(fs, bno) + numfrags(fs, size) > fs->fs_frag) { printf("dev=%s, bno = %jd, bsize = %ld, size = %ld, fs = %s\n", devtoname(dev), (intmax_t)bno, (long)fs->fs_bsize, size, fs->fs_fsmnt); panic("ffs_blkfree_cg: bad size"); } #endif if ((u_int)bno >= fs->fs_size) { printf("bad block %jd, ino %lu\n", (intmax_t)bno, (u_long)inum); ffs_fserr(fs, inum, "bad block"); return; } if ((error = ffs_getcg(fs, devvp, cg, &bp, &cgp)) != 0) return; cgbno = dtogd(fs, bno); blksfree = cg_blksfree(cgp); UFS_LOCK(ump); if (size == fs->fs_bsize) { fragno = fragstoblks(fs, cgbno); if (!ffs_isfreeblock(fs, blksfree, fragno)) { if (devvp->v_type == VREG) { UFS_UNLOCK(ump); /* devvp is a snapshot */ brelse(bp); return; } printf("dev = %s, block = %jd, fs = %s\n", devtoname(dev), (intmax_t)bno, fs->fs_fsmnt); panic("ffs_blkfree_cg: freeing free block"); } ffs_setblock(fs, blksfree, fragno); ffs_clusteracct(fs, cgp, fragno, 1); cgp->cg_cs.cs_nbfree++; fs->fs_cstotal.cs_nbfree++; fs->fs_cs(fs, cg).cs_nbfree++; } else { bbase = cgbno - fragnum(fs, cgbno); /* * decrement the counts associated with the old frags */ blk = blkmap(fs, blksfree, bbase); ffs_fragacct(fs, blk, cgp->cg_frsum, -1); /* * deallocate the fragment */ frags = numfrags(fs, size); for (i = 0; i < frags; i++) { if (isset(blksfree, cgbno + i)) { printf("dev = %s, block = %jd, fs = %s\n", devtoname(dev), (intmax_t)(bno + i), fs->fs_fsmnt); panic("ffs_blkfree_cg: freeing free frag"); } setbit(blksfree, cgbno + i); } cgp->cg_cs.cs_nffree += i; fs->fs_cstotal.cs_nffree += i; fs->fs_cs(fs, cg).cs_nffree += i; /* * add back in counts associated with the new frags */ blk = blkmap(fs, blksfree, bbase); ffs_fragacct(fs, blk, cgp->cg_frsum, 1); /* * if a complete block has been reassembled, account for it */ fragno = fragstoblks(fs, bbase); if (ffs_isblock(fs, blksfree, fragno)) { cgp->cg_cs.cs_nffree -= fs->fs_frag; fs->fs_cstotal.cs_nffree -= fs->fs_frag; fs->fs_cs(fs, cg).cs_nffree -= fs->fs_frag; ffs_clusteracct(fs, cgp, fragno, 1); cgp->cg_cs.cs_nbfree++; fs->fs_cstotal.cs_nbfree++; fs->fs_cs(fs, cg).cs_nbfree++; } } fs->fs_fmod = 1; ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); mp = UFSTOVFS(ump); if (MOUNTEDSOFTDEP(mp) && devvp->v_type == VCHR) softdep_setup_blkfree(UFSTOVFS(ump), bp, bno, numfrags(fs, size), dephd); bdwrite(bp); } /* * Structures and routines associated with trim management. * * The following requests are passed to trim_lookup to indicate * the actions that should be taken. */ #define NEW 1 /* if found, error else allocate and hash it */ #define OLD 2 /* if not found, error, else return it */ #define REPLACE 3 /* if not found, error else unhash and reallocate it */ #define DONE 4 /* if not found, error else unhash and return it */ #define SINGLE 5 /* don't look up, just allocate it and don't hash it */ MALLOC_DEFINE(M_TRIM, "ufs_trim", "UFS trim structures"); #define TRIMLIST_HASH(ump, key) \ (&(ump)->um_trimhash[(key) & (ump)->um_trimlisthashsize]) /* * These structures describe each of the block free requests aggregated * together to make up a trim request. */ struct trim_blkreq { TAILQ_ENTRY(trim_blkreq) blkreqlist; ufs2_daddr_t bno; long size; struct workhead *pdephd; struct workhead dephd; }; /* * Description of a trim request. */ struct ffs_blkfree_trim_params { TAILQ_HEAD(, trim_blkreq) blklist; LIST_ENTRY(ffs_blkfree_trim_params) hashlist; struct task task; struct ufsmount *ump; struct vnode *devvp; ino_t inum; ufs2_daddr_t bno; long size; long key; }; static void ffs_blkfree_trim_completed(struct buf *); static void ffs_blkfree_trim_task(void *ctx, int pending __unused); static struct ffs_blkfree_trim_params *trim_lookup(struct ufsmount *, struct vnode *, ufs2_daddr_t, long, ino_t, u_long, int); static void ffs_blkfree_sendtrim(struct ffs_blkfree_trim_params *); /* * Called on trim completion to start a task to free the associated block(s). */ static void ffs_blkfree_trim_completed(bp) struct buf *bp; { struct ffs_blkfree_trim_params *tp; tp = bp->b_fsprivate1; free(bp, M_TRIM); TASK_INIT(&tp->task, 0, ffs_blkfree_trim_task, tp); taskqueue_enqueue(tp->ump->um_trim_tq, &tp->task); } /* * Trim completion task that free associated block(s). */ static void ffs_blkfree_trim_task(ctx, pending) void *ctx; int pending; { struct ffs_blkfree_trim_params *tp; struct trim_blkreq *blkelm; struct ufsmount *ump; tp = ctx; ump = tp->ump; while ((blkelm = TAILQ_FIRST(&tp->blklist)) != NULL) { ffs_blkfree_cg(ump, ump->um_fs, tp->devvp, blkelm->bno, blkelm->size, tp->inum, blkelm->pdephd); TAILQ_REMOVE(&tp->blklist, blkelm, blkreqlist); free(blkelm, M_TRIM); } vn_finished_secondary_write(UFSTOVFS(ump)); UFS_LOCK(ump); ump->um_trim_inflight -= 1; ump->um_trim_inflight_blks -= numfrags(ump->um_fs, tp->size); UFS_UNLOCK(ump); free(tp, M_TRIM); } /* * Lookup a trim request by inode number. * Allocate if requested (NEW, REPLACE, SINGLE). */ static struct ffs_blkfree_trim_params * trim_lookup(ump, devvp, bno, size, inum, key, alloctype) struct ufsmount *ump; struct vnode *devvp; ufs2_daddr_t bno; long size; ino_t inum; u_long key; int alloctype; { struct trimlist_hashhead *tphashhead; struct ffs_blkfree_trim_params *tp, *ntp; ntp = malloc(sizeof(struct ffs_blkfree_trim_params), M_TRIM, M_WAITOK); if (alloctype != SINGLE) { KASSERT(key >= FIRST_VALID_KEY, ("trim_lookup: invalid key")); UFS_LOCK(ump); tphashhead = TRIMLIST_HASH(ump, key); LIST_FOREACH(tp, tphashhead, hashlist) if (key == tp->key) break; } switch (alloctype) { case NEW: KASSERT(tp == NULL, ("trim_lookup: found trim")); break; case OLD: KASSERT(tp != NULL, ("trim_lookup: missing call to ffs_blkrelease_start()")); UFS_UNLOCK(ump); free(ntp, M_TRIM); return (tp); case REPLACE: KASSERT(tp != NULL, ("trim_lookup: missing REPLACE trim")); LIST_REMOVE(tp, hashlist); /* tp will be freed by caller */ break; case DONE: KASSERT(tp != NULL, ("trim_lookup: missing DONE trim")); LIST_REMOVE(tp, hashlist); UFS_UNLOCK(ump); free(ntp, M_TRIM); return (tp); } TAILQ_INIT(&ntp->blklist); ntp->ump = ump; ntp->devvp = devvp; ntp->bno = bno; ntp->size = size; ntp->inum = inum; ntp->key = key; if (alloctype != SINGLE) { LIST_INSERT_HEAD(tphashhead, ntp, hashlist); UFS_UNLOCK(ump); } return (ntp); } /* * Dispatch a trim request. */ static void ffs_blkfree_sendtrim(tp) struct ffs_blkfree_trim_params *tp; { struct ufsmount *ump; struct mount *mp; struct buf *bp; /* * Postpone the set of the free bit in the cg bitmap until the * BIO_DELETE is completed. Otherwise, due to disk queue * reordering, TRIM might be issued after we reuse the block * and write some new data into it. */ ump = tp->ump; bp = malloc(sizeof(*bp), M_TRIM, M_WAITOK | M_ZERO); bp->b_iocmd = BIO_DELETE; bp->b_iooffset = dbtob(fsbtodb(ump->um_fs, tp->bno)); bp->b_iodone = ffs_blkfree_trim_completed; bp->b_bcount = tp->size; bp->b_fsprivate1 = tp; UFS_LOCK(ump); ump->um_trim_total += 1; ump->um_trim_inflight += 1; ump->um_trim_inflight_blks += numfrags(ump->um_fs, tp->size); ump->um_trim_total_blks += numfrags(ump->um_fs, tp->size); UFS_UNLOCK(ump); mp = UFSTOVFS(ump); vn_start_secondary_write(NULL, &mp, 0); g_vfs_strategy(ump->um_bo, bp); } /* * Allocate a new key to use to identify a range of blocks. */ u_long ffs_blkrelease_start(ump, devvp, inum) struct ufsmount *ump; struct vnode *devvp; ino_t inum; { static u_long masterkey; u_long key; if (((ump->um_flags & UM_CANDELETE) == 0) || dotrimcons == 0) return (SINGLETON_KEY); do { key = atomic_fetchadd_long(&masterkey, 1); } while (key < FIRST_VALID_KEY); (void) trim_lookup(ump, devvp, 0, 0, inum, key, NEW); return (key); } /* * Deallocate a key that has been used to identify a range of blocks. */ void ffs_blkrelease_finish(ump, key) struct ufsmount *ump; u_long key; { struct ffs_blkfree_trim_params *tp; if (((ump->um_flags & UM_CANDELETE) == 0) || dotrimcons == 0) return; /* * If the vfs.ffs.dotrimcons sysctl option is enabled while * a file deletion is active, specifically after a call * to ffs_blkrelease_start() but before the call to * ffs_blkrelease_finish(), ffs_blkrelease_start() will * have handed out SINGLETON_KEY rather than starting a * collection sequence. Thus if we get a SINGLETON_KEY * passed to ffs_blkrelease_finish(), we just return rather * than trying to finish the nonexistent sequence. */ if (key == SINGLETON_KEY) { #ifdef INVARIANTS printf("%s: vfs.ffs.dotrimcons enabled on active filesystem\n", ump->um_mountp->mnt_stat.f_mntonname); #endif return; } /* * We are done with sending blocks using this key. Look up the key * using the DONE alloctype (in tp) to request that it be unhashed * as we will not be adding to it. If the key has never been used, * tp->size will be zero, so we can just free tp. Otherwise the call * to ffs_blkfree_sendtrim(tp) causes the block range described by * tp to be issued (and then tp to be freed). */ tp = trim_lookup(ump, NULL, 0, 0, 0, key, DONE); if (tp->size == 0) free(tp, M_TRIM); else ffs_blkfree_sendtrim(tp); } /* * Setup to free a block or fragment. * * Check for snapshots that might want to claim the block. * If trims are requested, prepare a trim request. Attempt to * aggregate consecutive blocks into a single trim request. */ void ffs_blkfree(ump, fs, devvp, bno, size, inum, vtype, dephd, key) struct ufsmount *ump; struct fs *fs; struct vnode *devvp; ufs2_daddr_t bno; long size; ino_t inum; enum vtype vtype; struct workhead *dephd; u_long key; { struct ffs_blkfree_trim_params *tp, *ntp; struct trim_blkreq *blkelm; /* * Check to see if a snapshot wants to claim the block. * Check that devvp is a normal disk device, not a snapshot, * it has a snapshot(s) associated with it, and one of the * snapshots wants to claim the block. */ if (devvp->v_type == VCHR && (devvp->v_vflag & VV_COPYONWRITE) && ffs_snapblkfree(fs, devvp, bno, size, inum, vtype, dephd)) { return; } /* * Nothing to delay if TRIM is not required for this block or TRIM * is disabled or the operation is performed on a snapshot. */ if (key == NOTRIM_KEY || ((ump->um_flags & UM_CANDELETE) == 0) || devvp->v_type == VREG) { ffs_blkfree_cg(ump, fs, devvp, bno, size, inum, dephd); return; } blkelm = malloc(sizeof(struct trim_blkreq), M_TRIM, M_WAITOK); blkelm->bno = bno; blkelm->size = size; if (dephd == NULL) { blkelm->pdephd = NULL; } else { LIST_INIT(&blkelm->dephd); LIST_SWAP(dephd, &blkelm->dephd, worklist, wk_list); blkelm->pdephd = &blkelm->dephd; } if (key == SINGLETON_KEY) { /* * Just a single non-contiguous piece. Use the SINGLE * alloctype to return a trim request that will not be * hashed for future lookup. */ tp = trim_lookup(ump, devvp, bno, size, inum, key, SINGLE); TAILQ_INSERT_HEAD(&tp->blklist, blkelm, blkreqlist); ffs_blkfree_sendtrim(tp); return; } /* * The callers of this function are not tracking whether or not * the blocks are contiguous. They are just saying that they * are freeing a set of blocks. It is this code that determines * the pieces of that range that are actually contiguous. * * Calling ffs_blkrelease_start() will have created an entry * that we will use. */ tp = trim_lookup(ump, devvp, bno, size, inum, key, OLD); if (tp->size == 0) { /* * First block of a potential range, set block and size * for the trim block. */ tp->bno = bno; tp->size = size; TAILQ_INSERT_HEAD(&tp->blklist, blkelm, blkreqlist); return; } /* * If this block is a continuation of the range (either * follows at the end or preceeds in the front) then we * add it to the front or back of the list and return. * * If it is not a continuation of the trim that we were * building, using the REPLACE alloctype, we request that * the old trim request (still in tp) be unhashed and a * new range started (in ntp). The ffs_blkfree_sendtrim(tp) * call causes the block range described by tp to be issued * (and then tp to be freed). */ if (bno + numfrags(fs, size) == tp->bno) { TAILQ_INSERT_HEAD(&tp->blklist, blkelm, blkreqlist); tp->bno = bno; tp->size += size; return; } else if (bno == tp->bno + numfrags(fs, tp->size)) { TAILQ_INSERT_TAIL(&tp->blklist, blkelm, blkreqlist); tp->size += size; return; } ntp = trim_lookup(ump, devvp, bno, size, inum, key, REPLACE); TAILQ_INSERT_HEAD(&ntp->blklist, blkelm, blkreqlist); ffs_blkfree_sendtrim(tp); } #ifdef INVARIANTS /* * Verify allocation of a block or fragment. Returns true if block or * fragment is allocated, false if it is free. */ static int ffs_checkblk(ip, bno, size) struct inode *ip; ufs2_daddr_t bno; long size; { struct fs *fs; struct cg *cgp; struct buf *bp; ufs1_daddr_t cgbno; int i, error, frags, free; u_int8_t *blksfree; fs = ITOFS(ip); if ((u_int)size > fs->fs_bsize || fragoff(fs, size) != 0) { printf("bsize = %ld, size = %ld, fs = %s\n", (long)fs->fs_bsize, size, fs->fs_fsmnt); panic("ffs_checkblk: bad size"); } if ((u_int)bno >= fs->fs_size) panic("ffs_checkblk: bad block %jd", (intmax_t)bno); error = ffs_getcg(fs, ITODEVVP(ip), dtog(fs, bno), &bp, &cgp); if (error) panic("ffs_checkblk: cylinder group read failed"); blksfree = cg_blksfree(cgp); cgbno = dtogd(fs, bno); if (size == fs->fs_bsize) { free = ffs_isblock(fs, blksfree, fragstoblks(fs, cgbno)); } else { frags = numfrags(fs, size); for (free = 0, i = 0; i < frags; i++) if (isset(blksfree, cgbno + i)) free++; if (free != 0 && free != frags) panic("ffs_checkblk: partially free fragment"); } brelse(bp); return (!free); } #endif /* INVARIANTS */ /* * Free an inode. */ int ffs_vfree(pvp, ino, mode) struct vnode *pvp; ino_t ino; int mode; { struct ufsmount *ump; if (DOINGSOFTDEP(pvp)) { softdep_freefile(pvp, ino, mode); return (0); } ump = VFSTOUFS(pvp->v_mount); return (ffs_freefile(ump, ump->um_fs, ump->um_devvp, ino, mode, NULL)); } /* * Do the actual free operation. * The specified inode is placed back in the free map. */ int ffs_freefile(ump, fs, devvp, ino, mode, wkhd) struct ufsmount *ump; struct fs *fs; struct vnode *devvp; ino_t ino; int mode; struct workhead *wkhd; { struct cg *cgp; struct buf *bp; int error; u_int cg; u_int8_t *inosused; struct cdev *dev; cg = ino_to_cg(fs, ino); if (devvp->v_type == VREG) { /* devvp is a snapshot */ MPASS(devvp->v_mount->mnt_data == ump); dev = ump->um_devvp->v_rdev; } else if (devvp->v_type == VCHR) { /* devvp is a normal disk device */ dev = devvp->v_rdev; } else { bp = NULL; return (0); } if (ino >= fs->fs_ipg * fs->fs_ncg) panic("ffs_freefile: range: dev = %s, ino = %ju, fs = %s", devtoname(dev), (uintmax_t)ino, fs->fs_fsmnt); if ((error = ffs_getcg(fs, devvp, cg, &bp, &cgp)) != 0) return (error); inosused = cg_inosused(cgp); ino %= fs->fs_ipg; if (isclr(inosused, ino)) { printf("dev = %s, ino = %ju, fs = %s\n", devtoname(dev), (uintmax_t)(ino + cg * fs->fs_ipg), fs->fs_fsmnt); if (fs->fs_ronly == 0) panic("ffs_freefile: freeing free inode"); } clrbit(inosused, ino); if (ino < cgp->cg_irotor) cgp->cg_irotor = ino; cgp->cg_cs.cs_nifree++; UFS_LOCK(ump); fs->fs_cstotal.cs_nifree++; fs->fs_cs(fs, cg).cs_nifree++; if ((mode & IFMT) == IFDIR) { cgp->cg_cs.cs_ndir--; fs->fs_cstotal.cs_ndir--; fs->fs_cs(fs, cg).cs_ndir--; } fs->fs_fmod = 1; ACTIVECLEAR(fs, cg); UFS_UNLOCK(ump); if (MOUNTEDSOFTDEP(UFSTOVFS(ump)) && devvp->v_type == VCHR) softdep_setup_inofree(UFSTOVFS(ump), bp, ino + cg * fs->fs_ipg, wkhd); bdwrite(bp); return (0); } /* * Check to see if a file is free. * Used to check for allocated files in snapshots. */ int ffs_checkfreefile(fs, devvp, ino) struct fs *fs; struct vnode *devvp; ino_t ino; { struct cg *cgp; struct buf *bp; int ret, error; u_int cg; u_int8_t *inosused; cg = ino_to_cg(fs, ino); if ((devvp->v_type != VREG) && (devvp->v_type != VCHR)) return (1); if (ino >= fs->fs_ipg * fs->fs_ncg) return (1); if ((error = ffs_getcg(fs, devvp, cg, &bp, &cgp)) != 0) return (1); inosused = cg_inosused(cgp); ino %= fs->fs_ipg; ret = isclr(inosused, ino); brelse(bp); return (ret); } /* * Find a block of the specified size in the specified cylinder group. * * It is a panic if a request is made to find a block if none are * available. */ static ufs1_daddr_t ffs_mapsearch(fs, cgp, bpref, allocsiz) struct fs *fs; struct cg *cgp; ufs2_daddr_t bpref; int allocsiz; { ufs1_daddr_t bno; int start, len, loc, i; int blk, field, subfield, pos; u_int8_t *blksfree; /* * find the fragment by searching through the free block * map for an appropriate bit pattern */ if (bpref) start = dtogd(fs, bpref) / NBBY; else start = cgp->cg_frotor / NBBY; blksfree = cg_blksfree(cgp); len = howmany(fs->fs_fpg, NBBY) - start; loc = scanc((u_int)len, (u_char *)&blksfree[start], fragtbl[fs->fs_frag], (u_char)(1 << (allocsiz - 1 + (fs->fs_frag % NBBY)))); if (loc == 0) { len = start + 1; start = 0; loc = scanc((u_int)len, (u_char *)&blksfree[0], fragtbl[fs->fs_frag], (u_char)(1 << (allocsiz - 1 + (fs->fs_frag % NBBY)))); if (loc == 0) { printf("start = %d, len = %d, fs = %s\n", start, len, fs->fs_fsmnt); panic("ffs_alloccg: map corrupted"); /* NOTREACHED */ } } bno = (start + len - loc) * NBBY; cgp->cg_frotor = bno; /* * found the byte in the map * sift through the bits to find the selected frag */ for (i = bno + NBBY; bno < i; bno += fs->fs_frag) { blk = blkmap(fs, blksfree, bno); blk <<= 1; field = around[allocsiz]; subfield = inside[allocsiz]; for (pos = 0; pos <= fs->fs_frag - allocsiz; pos++) { if ((blk & field) == subfield) return (bno + pos); field <<= 1; subfield <<= 1; } } printf("bno = %lu, fs = %s\n", (u_long)bno, fs->fs_fsmnt); panic("ffs_alloccg: block not in map"); return (-1); } static const struct statfs * ffs_getmntstat(struct vnode *devvp) { if (devvp->v_type == VCHR) return (&devvp->v_rdev->si_mountpt->mnt_stat); return (ffs_getmntstat(VFSTOUFS(devvp->v_mount)->um_devvp)); } /* * Fetch and verify a cylinder group. */ int ffs_getcg(fs, devvp, cg, bpp, cgpp) struct fs *fs; struct vnode *devvp; u_int cg; struct buf **bpp; struct cg **cgpp; { struct buf *bp; struct cg *cgp; const struct statfs *sfs; int flags, error; *bpp = NULL; *cgpp = NULL; flags = 0; if ((fs->fs_metackhash & CK_CYLGRP) != 0) flags |= GB_CKHASH; error = breadn_flags(devvp, devvp->v_type == VREG ? fragstoblks(fs, cgtod(fs, cg)) : fsbtodb(fs, cgtod(fs, cg)), (int)fs->fs_cgsize, NULL, NULL, 0, NOCRED, flags, ffs_ckhash_cg, &bp); if (error != 0) return (error); cgp = (struct cg *)bp->b_data; if ((fs->fs_metackhash & CK_CYLGRP) != 0 && (bp->b_flags & B_CKHASH) != 0 && cgp->cg_ckhash != bp->b_ckhash) { sfs = ffs_getmntstat(devvp); printf("UFS %s%s (%s) cylinder checksum failed: cg %u, cgp: " "0x%x != bp: 0x%jx\n", devvp->v_type == VCHR ? "" : "snapshot of ", sfs->f_mntfromname, sfs->f_mntonname, cg, cgp->cg_ckhash, (uintmax_t)bp->b_ckhash); bp->b_flags &= ~B_CKHASH; bp->b_flags |= B_INVAL | B_NOCACHE; brelse(bp); return (EIO); } if (!cg_chkmagic(cgp) || cgp->cg_cgx != cg) { sfs = ffs_getmntstat(devvp); printf("UFS %s%s (%s)", devvp->v_type == VCHR ? "" : "snapshot of ", sfs->f_mntfromname, sfs->f_mntonname); if (!cg_chkmagic(cgp)) printf(" cg %u: bad magic number 0x%x should be 0x%x\n", cg, cgp->cg_magic, CG_MAGIC); else printf(": wrong cylinder group cg %u != cgx %u\n", cg, cgp->cg_cgx); bp->b_flags &= ~B_CKHASH; bp->b_flags |= B_INVAL | B_NOCACHE; brelse(bp); return (EIO); } bp->b_flags &= ~B_CKHASH; bp->b_xflags |= BX_BKGRDWRITE; /* * If we are using check hashes on the cylinder group then we want * to limit changing the cylinder group time to when we are actually * going to write it to disk so that its check hash remains correct * in memory. If the CK_CYLGRP flag is set the time is updated in * ffs_bufwrite() as the buffer is queued for writing. Otherwise we * update the time here as we have done historically. */ if ((fs->fs_metackhash & CK_CYLGRP) != 0) bp->b_xflags |= BX_CYLGRP; else cgp->cg_old_time = cgp->cg_time = time_second; *bpp = bp; *cgpp = cgp; return (0); } static void ffs_ckhash_cg(bp) struct buf *bp; { uint32_t ckhash; struct cg *cgp; cgp = (struct cg *)bp->b_data; ckhash = cgp->cg_ckhash; cgp->cg_ckhash = 0; bp->b_ckhash = calculate_crc32c(~0L, bp->b_data, bp->b_bcount); cgp->cg_ckhash = ckhash; } /* * Fserr prints the name of a filesystem with an error diagnostic. * * The form of the error message is: * fs: error message */ void ffs_fserr(fs, inum, cp) struct fs *fs; ino_t inum; char *cp; { struct thread *td = curthread; /* XXX */ struct proc *p = td->td_proc; log(LOG_ERR, "pid %d (%s), uid %d inumber %ju on %s: %s\n", p->p_pid, p->p_comm, td->td_ucred->cr_uid, (uintmax_t)inum, fs->fs_fsmnt, cp); } /* * This function provides the capability for the fsck program to * update an active filesystem. Fourteen operations are provided: * * adjrefcnt(inode, amt) - adjusts the reference count on the * specified inode by the specified amount. Under normal * operation the count should always go down. Decrementing * the count to zero will cause the inode to be freed. * adjblkcnt(inode, amt) - adjust the number of blocks used by the * inode by the specified amount. * adjsize(inode, size) - set the size of the inode to the * specified size. * adjndir, adjbfree, adjifree, adjffree, adjnumclusters(amt) - * adjust the superblock summary. * freedirs(inode, count) - directory inodes [inode..inode + count - 1] * are marked as free. Inodes should never have to be marked * as in use. * freefiles(inode, count) - file inodes [inode..inode + count - 1] * are marked as free. Inodes should never have to be marked * as in use. * freeblks(blockno, size) - blocks [blockno..blockno + size - 1] * are marked as free. Blocks should never have to be marked * as in use. * setflags(flags, set/clear) - the fs_flags field has the specified * flags set (second parameter +1) or cleared (second parameter -1). * setcwd(dirinode) - set the current directory to dirinode in the * filesystem associated with the snapshot. * setdotdot(oldvalue, newvalue) - Verify that the inode number for ".." * in the current directory is oldvalue then change it to newvalue. * unlink(nameptr, oldvalue) - Verify that the inode number associated * with nameptr in the current directory is oldvalue then unlink it. * * The following functions may only be used on a quiescent filesystem * by the soft updates journal. They are not safe to be run on an active * filesystem. * * setinode(inode, dip) - the specified disk inode is replaced with the * contents pointed to by dip. * setbufoutput(fd, flags) - output associated with the specified file * descriptor (which must reference the character device supporting * the filesystem) switches from using physio to running through the * buffer cache when flags is set to 1. The descriptor reverts to * physio for output when flags is set to zero. */ static int sysctl_ffs_fsck(SYSCTL_HANDLER_ARGS); SYSCTL_PROC(_vfs_ffs, FFS_ADJ_REFCNT, adjrefcnt, CTLFLAG_WR|CTLTYPE_STRUCT, 0, 0, sysctl_ffs_fsck, "S,fsck", "Adjust Inode Reference Count"); static SYSCTL_NODE(_vfs_ffs, FFS_ADJ_BLKCNT, adjblkcnt, CTLFLAG_WR, sysctl_ffs_fsck, "Adjust Inode Used Blocks Count"); static SYSCTL_NODE(_vfs_ffs, FFS_SET_SIZE, setsize, CTLFLAG_WR, sysctl_ffs_fsck, "Set the inode size"); static SYSCTL_NODE(_vfs_ffs, FFS_ADJ_NDIR, adjndir, CTLFLAG_WR, sysctl_ffs_fsck, "Adjust number of directories"); static SYSCTL_NODE(_vfs_ffs, FFS_ADJ_NBFREE, adjnbfree, CTLFLAG_WR, sysctl_ffs_fsck, "Adjust number of free blocks"); static SYSCTL_NODE(_vfs_ffs, FFS_ADJ_NIFREE, adjnifree, CTLFLAG_WR, sysctl_ffs_fsck, "Adjust number of free inodes"); static SYSCTL_NODE(_vfs_ffs, FFS_ADJ_NFFREE, adjnffree, CTLFLAG_WR, sysctl_ffs_fsck, "Adjust number of free frags"); static SYSCTL_NODE(_vfs_ffs, FFS_ADJ_NUMCLUSTERS, adjnumclusters, CTLFLAG_WR, sysctl_ffs_fsck, "Adjust number of free clusters"); static SYSCTL_NODE(_vfs_ffs, FFS_DIR_FREE, freedirs, CTLFLAG_WR, sysctl_ffs_fsck, "Free Range of Directory Inodes"); static SYSCTL_NODE(_vfs_ffs, FFS_FILE_FREE, freefiles, CTLFLAG_WR, sysctl_ffs_fsck, "Free Range of File Inodes"); static SYSCTL_NODE(_vfs_ffs, FFS_BLK_FREE, freeblks, CTLFLAG_WR, sysctl_ffs_fsck, "Free Range of Blocks"); static SYSCTL_NODE(_vfs_ffs, FFS_SET_FLAGS, setflags, CTLFLAG_WR, sysctl_ffs_fsck, "Change Filesystem Flags"); static SYSCTL_NODE(_vfs_ffs, FFS_SET_CWD, setcwd, CTLFLAG_WR, sysctl_ffs_fsck, "Set Current Working Directory"); static SYSCTL_NODE(_vfs_ffs, FFS_SET_DOTDOT, setdotdot, CTLFLAG_WR, sysctl_ffs_fsck, "Change Value of .. Entry"); static SYSCTL_NODE(_vfs_ffs, FFS_UNLINK, unlink, CTLFLAG_WR, sysctl_ffs_fsck, "Unlink a Duplicate Name"); static SYSCTL_NODE(_vfs_ffs, FFS_SET_INODE, setinode, CTLFLAG_WR, sysctl_ffs_fsck, "Update an On-Disk Inode"); static SYSCTL_NODE(_vfs_ffs, FFS_SET_BUFOUTPUT, setbufoutput, CTLFLAG_WR, sysctl_ffs_fsck, "Set Buffered Writing for Descriptor"); #define DEBUG 1 #ifdef DEBUG static int fsckcmds = 0; SYSCTL_INT(_debug, OID_AUTO, fsckcmds, CTLFLAG_RW, &fsckcmds, 0, ""); #endif /* DEBUG */ static int buffered_write(struct file *, struct uio *, struct ucred *, int, struct thread *); static int sysctl_ffs_fsck(SYSCTL_HANDLER_ARGS) { struct thread *td = curthread; struct fsck_cmd cmd; struct ufsmount *ump; struct vnode *vp, *dvp, *fdvp; struct inode *ip, *dp; struct mount *mp; struct fs *fs; ufs2_daddr_t blkno; long blkcnt, blksize; u_long key; struct file *fp, *vfp; cap_rights_t rights; int filetype, error; static struct fileops *origops, bufferedops; if (req->newlen > sizeof cmd) return (EBADRPC); if ((error = SYSCTL_IN(req, &cmd, sizeof cmd)) != 0) return (error); if (cmd.version != FFS_CMD_VERSION) return (ERPCMISMATCH); if ((error = getvnode(td, cmd.handle, cap_rights_init(&rights, CAP_FSCK), &fp)) != 0) return (error); vp = fp->f_data; if (vp->v_type != VREG && vp->v_type != VDIR) { fdrop(fp, td); return (EINVAL); } vn_start_write(vp, &mp, V_WAIT); if (mp == NULL || strncmp(mp->mnt_stat.f_fstypename, "ufs", MFSNAMELEN)) { vn_finished_write(mp); fdrop(fp, td); return (EINVAL); } ump = VFSTOUFS(mp); if ((mp->mnt_flag & MNT_RDONLY) && ump->um_fsckpid != td->td_proc->p_pid) { vn_finished_write(mp); fdrop(fp, td); return (EROFS); } fs = ump->um_fs; filetype = IFREG; switch (oidp->oid_number) { case FFS_SET_FLAGS: #ifdef DEBUG if (fsckcmds) printf("%s: %s flags\n", mp->mnt_stat.f_mntonname, cmd.size > 0 ? "set" : "clear"); #endif /* DEBUG */ if (cmd.size > 0) fs->fs_flags |= (long)cmd.value; else fs->fs_flags &= ~(long)cmd.value; break; case FFS_ADJ_REFCNT: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust inode %jd link count by %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value, (intmax_t)cmd.size); } #endif /* DEBUG */ if ((error = ffs_vget(mp, (ino_t)cmd.value, LK_EXCLUSIVE, &vp))) break; ip = VTOI(vp); ip->i_nlink += cmd.size; DIP_SET(ip, i_nlink, ip->i_nlink); ip->i_effnlink += cmd.size; ip->i_flag |= IN_CHANGE | IN_MODIFIED; error = ffs_update(vp, 1); if (DOINGSOFTDEP(vp)) softdep_change_linkcnt(ip); vput(vp); break; case FFS_ADJ_BLKCNT: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust inode %jd block count by %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value, (intmax_t)cmd.size); } #endif /* DEBUG */ if ((error = ffs_vget(mp, (ino_t)cmd.value, LK_EXCLUSIVE, &vp))) break; ip = VTOI(vp); DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + cmd.size); ip->i_flag |= IN_CHANGE | IN_MODIFIED; error = ffs_update(vp, 1); vput(vp); break; case FFS_SET_SIZE: #ifdef DEBUG if (fsckcmds) { printf("%s: set inode %jd size to %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value, (intmax_t)cmd.size); } #endif /* DEBUG */ if ((error = ffs_vget(mp, (ino_t)cmd.value, LK_EXCLUSIVE, &vp))) break; ip = VTOI(vp); DIP_SET(ip, i_size, cmd.size); ip->i_flag |= IN_CHANGE | IN_MODIFIED; error = ffs_update(vp, 1); vput(vp); break; case FFS_DIR_FREE: filetype = IFDIR; /* fall through */ case FFS_FILE_FREE: #ifdef DEBUG if (fsckcmds) { if (cmd.size == 1) printf("%s: free %s inode %ju\n", mp->mnt_stat.f_mntonname, filetype == IFDIR ? "directory" : "file", (uintmax_t)cmd.value); else printf("%s: free %s inodes %ju-%ju\n", mp->mnt_stat.f_mntonname, filetype == IFDIR ? "directory" : "file", (uintmax_t)cmd.value, (uintmax_t)(cmd.value + cmd.size - 1)); } #endif /* DEBUG */ while (cmd.size > 0) { if ((error = ffs_freefile(ump, fs, ump->um_devvp, cmd.value, filetype, NULL))) break; cmd.size -= 1; cmd.value += 1; } break; case FFS_BLK_FREE: #ifdef DEBUG if (fsckcmds) { if (cmd.size == 1) printf("%s: free block %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); else printf("%s: free blocks %jd-%jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value, (intmax_t)cmd.value + cmd.size - 1); } #endif /* DEBUG */ blkno = cmd.value; blkcnt = cmd.size; blksize = fs->fs_frag - (blkno % fs->fs_frag); key = ffs_blkrelease_start(ump, ump->um_devvp, UFS_ROOTINO); while (blkcnt > 0) { if (blkcnt < blksize) blksize = blkcnt; ffs_blkfree(ump, fs, ump->um_devvp, blkno, blksize * fs->fs_fsize, UFS_ROOTINO, VDIR, NULL, key); blkno += blksize; blkcnt -= blksize; blksize = fs->fs_frag; } ffs_blkrelease_finish(ump, key); break; /* * Adjust superblock summaries. fsck(8) is expected to * submit deltas when necessary. */ case FFS_ADJ_NDIR: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust number of directories by %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ fs->fs_cstotal.cs_ndir += cmd.value; break; case FFS_ADJ_NBFREE: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust number of free blocks by %+jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ fs->fs_cstotal.cs_nbfree += cmd.value; break; case FFS_ADJ_NIFREE: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust number of free inodes by %+jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ fs->fs_cstotal.cs_nifree += cmd.value; break; case FFS_ADJ_NFFREE: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust number of free frags by %+jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ fs->fs_cstotal.cs_nffree += cmd.value; break; case FFS_ADJ_NUMCLUSTERS: #ifdef DEBUG if (fsckcmds) { printf("%s: adjust number of free clusters by %+jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ fs->fs_cstotal.cs_numclusters += cmd.value; break; case FFS_SET_CWD: #ifdef DEBUG if (fsckcmds) { printf("%s: set current directory to inode %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ if ((error = ffs_vget(mp, (ino_t)cmd.value, LK_SHARED, &vp))) break; AUDIT_ARG_VNODE1(vp); if ((error = change_dir(vp, td)) != 0) { vput(vp); break; } VOP_UNLOCK(vp, 0); pwd_chdir(td, vp); break; case FFS_SET_DOTDOT: #ifdef DEBUG if (fsckcmds) { printf("%s: change .. in cwd from %jd to %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value, (intmax_t)cmd.size); } #endif /* DEBUG */ /* * First we have to get and lock the parent directory * to which ".." points. */ error = ffs_vget(mp, (ino_t)cmd.value, LK_EXCLUSIVE, &fdvp); if (error) break; /* * Now we get and lock the child directory containing "..". */ FILEDESC_SLOCK(td->td_proc->p_fd); dvp = td->td_proc->p_fd->fd_cdir; FILEDESC_SUNLOCK(td->td_proc->p_fd); if ((error = vget(dvp, LK_EXCLUSIVE, td)) != 0) { vput(fdvp); break; } dp = VTOI(dvp); dp->i_offset = 12; /* XXX mastertemplate.dot_reclen */ error = ufs_dirrewrite(dp, VTOI(fdvp), (ino_t)cmd.size, DT_DIR, 0); cache_purge(fdvp); cache_purge(dvp); vput(dvp); vput(fdvp); break; case FFS_UNLINK: #ifdef DEBUG if (fsckcmds) { char buf[32]; if (copyinstr((char *)(intptr_t)cmd.value, buf,32,NULL)) strncpy(buf, "Name_too_long", 32); printf("%s: unlink %s (inode %jd)\n", mp->mnt_stat.f_mntonname, buf, (intmax_t)cmd.size); } #endif /* DEBUG */ /* - * kern_unlinkat will do its own start/finish writes and + * kern_funlinkat will do its own start/finish writes and * they do not nest, so drop ours here. Setting mp == NULL * indicates that vn_finished_write is not needed down below. */ vn_finished_write(mp); mp = NULL; - error = kern_unlinkat(td, AT_FDCWD, (char *)(intptr_t)cmd.value, - UIO_USERSPACE, 0, (ino_t)cmd.size); + error = kern_funlinkat(td, AT_FDCWD, + (char *)(intptr_t)cmd.value, FD_NONE, UIO_USERSPACE, + 0, (ino_t)cmd.size); break; case FFS_SET_INODE: if (ump->um_fsckpid != td->td_proc->p_pid) { error = EPERM; break; } #ifdef DEBUG if (fsckcmds) { printf("%s: update inode %jd\n", mp->mnt_stat.f_mntonname, (intmax_t)cmd.value); } #endif /* DEBUG */ if ((error = ffs_vget(mp, (ino_t)cmd.value, LK_EXCLUSIVE, &vp))) break; AUDIT_ARG_VNODE1(vp); ip = VTOI(vp); if (I_IS_UFS1(ip)) error = copyin((void *)(intptr_t)cmd.size, ip->i_din1, sizeof(struct ufs1_dinode)); else error = copyin((void *)(intptr_t)cmd.size, ip->i_din2, sizeof(struct ufs2_dinode)); if (error) { vput(vp); break; } ip->i_flag |= IN_CHANGE | IN_MODIFIED; error = ffs_update(vp, 1); vput(vp); break; case FFS_SET_BUFOUTPUT: if (ump->um_fsckpid != td->td_proc->p_pid) { error = EPERM; break; } if (ITOUMP(VTOI(vp)) != ump) { error = EINVAL; break; } #ifdef DEBUG if (fsckcmds) { printf("%s: %s buffered output for descriptor %jd\n", mp->mnt_stat.f_mntonname, cmd.size == 1 ? "enable" : "disable", (intmax_t)cmd.value); } #endif /* DEBUG */ if ((error = getvnode(td, cmd.value, cap_rights_init(&rights, CAP_FSCK), &vfp)) != 0) break; if (vfp->f_vnode->v_type != VCHR) { fdrop(vfp, td); error = EINVAL; break; } if (origops == NULL) { origops = vfp->f_ops; bcopy((void *)origops, (void *)&bufferedops, sizeof(bufferedops)); bufferedops.fo_write = buffered_write; } if (cmd.size == 1) atomic_store_rel_ptr((volatile uintptr_t *)&vfp->f_ops, (uintptr_t)&bufferedops); else atomic_store_rel_ptr((volatile uintptr_t *)&vfp->f_ops, (uintptr_t)origops); fdrop(vfp, td); break; default: #ifdef DEBUG if (fsckcmds) { printf("Invalid request %d from fsck\n", oidp->oid_number); } #endif /* DEBUG */ error = EINVAL; break; } fdrop(fp, td); vn_finished_write(mp); return (error); } /* * Function to switch a descriptor to use the buffer cache to stage * its I/O. This is needed so that writes to the filesystem device * will give snapshots a chance to copy modified blocks for which it * needs to retain copies. */ static int buffered_write(fp, uio, active_cred, flags, td) struct file *fp; struct uio *uio; struct ucred *active_cred; int flags; struct thread *td; { struct vnode *devvp, *vp; struct inode *ip; struct buf *bp; struct fs *fs; struct filedesc *fdp; int error; daddr_t lbn; /* * The devvp is associated with the /dev filesystem. To discover * the filesystem with which the device is associated, we depend * on the application setting the current directory to a location * within the filesystem being written. Yes, this is an ugly hack. */ devvp = fp->f_vnode; if (!vn_isdisk(devvp, NULL)) return (EINVAL); fdp = td->td_proc->p_fd; FILEDESC_SLOCK(fdp); vp = fdp->fd_cdir; vref(vp); FILEDESC_SUNLOCK(fdp); vn_lock(vp, LK_SHARED | LK_RETRY); /* * Check that the current directory vnode indeed belongs to * UFS before trying to dereference UFS-specific v_data fields. */ if (vp->v_op != &ffs_vnodeops1 && vp->v_op != &ffs_vnodeops2) { vput(vp); return (EINVAL); } ip = VTOI(vp); if (ITODEVVP(ip) != devvp) { vput(vp); return (EINVAL); } fs = ITOFS(ip); vput(vp); foffset_lock_uio(fp, uio, flags); vn_lock(devvp, LK_EXCLUSIVE | LK_RETRY); #ifdef DEBUG if (fsckcmds) { printf("%s: buffered write for block %jd\n", fs->fs_fsmnt, (intmax_t)btodb(uio->uio_offset)); } #endif /* DEBUG */ /* * All I/O must be contained within a filesystem block, start on * a fragment boundary, and be a multiple of fragments in length. */ if (uio->uio_resid > fs->fs_bsize - (uio->uio_offset % fs->fs_bsize) || fragoff(fs, uio->uio_offset) != 0 || fragoff(fs, uio->uio_resid) != 0) { error = EINVAL; goto out; } lbn = numfrags(fs, uio->uio_offset); bp = getblk(devvp, lbn, uio->uio_resid, 0, 0, 0); bp->b_flags |= B_RELBUF; if ((error = uiomove((char *)bp->b_data, uio->uio_resid, uio)) != 0) { brelse(bp); goto out; } error = bwrite(bp); out: VOP_UNLOCK(devvp, 0); foffset_unlock_uio(fp, uio, flags | FOF_NEXTOFF); return (error); }