Index: releng/9.3/UPDATING =================================================================== --- releng/9.3/UPDATING (revision 296610) +++ releng/9.3/UPDATING (revision 296611) @@ -1,1931 +1,1938 @@ Updating Information for FreeBSD current users This file is maintained and copyrighted by M. Warner Losh . See end of file for further details. For commonly done items, please see the COMMON ITEMS: section later in the file. These instructions assume that you basically know what you are doing. If not, then please consult the FreeBSD handbook: http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html Items affecting the ports and packages system can be found in /usr/ports/UPDATING. Please read that file before running portupgrade. +20160310 p38 FreeBSD-SA-16:13.bind + FreeBSD-SA-16:12.openssl [revised] + + Fix multiple vulnerabilities of BIND. [SA-16:13] + + Fix a regression with OpenSSL patch. [SA-16:12] + 20160303 p37 FreeBSD-SA-16:12.openssl Fix multiple vulnerabilities of OpenSSL. 20160130 p36 FreeBSD-SA-16:11.openssl Fix OpenSSL SSLv2 ciphersuite downgrade vulnerability. [SA-16:11] 20160127 p35 FreeBSD-SA-16:08.bind FreeBSD-SA-16:09.ntp FreeBSD-SA-16:10.linux Fix BIND remote denial of service vulnerability. [SA-16:08] Fix multiple vulnerabilities of ntp. [SA-16:09] Fix Linux compatibility layer issetugid(2) system call vulnerability. [SA-16:10] 20160114 p34 FreeBSD-SA-16:07.openssh Fix OpenSSH client information leak. [SA-16:07] 20160114 p33 FreeBSD-EN-16:02.pf FreeBSD-EN-16:03.yplib FreeBSD-SA-16:01.sctp FreeBSD-SA-16:02.ntp FreeBSD-SA-16:03.linux FreeBSD-SA-16:04.linux FreeBSD-SA-16:05.tcp FreeBSD-SA-16:06.bsnmpd Fix pf(4) generating bad TCP checksums. [EN-16:02] Fix infinite loop in YP/NIS client library. [EN-16:03] Fix remote denial of service in SCTP. [SA-16:01] Update NTP to 4.2.8p5. [SA-16:02] Fix kernel memory diclosure in Linux compatibility layer. [SA-16:03] Fix kernel memory overwrite in Linux compatibility layer. [SA-16:04] Fix crash in TCP MD5 signatures. [SA-16:05] Fix insecure default permissions for snmpd.config. [SA-16:06] 20151216 p32 FreeBSD-SA-15:27.bind Fix BIND remote denial of service vulnerability. [SA-15:27] 20151205 p31 FreeBSD-SA-15:26.openssl Fix OpenSSL X509_ATTRIBUTE memory leak. [SA-15:26] 20151104 p30 FreeBSD-SA-15:25.ntp [revised] FreeBSD-EN-15:19.kqueue FreeBSD-EN-15:20.vm Fix regression of ntpq(8) utility exiting due to trap 6 in 9.3-RELEASE-p29. [SA-15:25] Fix regression in ntpd(8) lacking support for RAWDCF reference clock in 9.3-RELEASE-p29. [SA-15:25] Fix kqueue write events never fired for files greater 2GB. [EN-15:19] Fix applications exiting due to segmentation violation on a correct memory address. [EN-15:20] 20151026: p29 FreeBSD-SA-15:25.ntp Fix multiple NTP vulnerabilities. New NTP version is 4.2.8p4. The configuration file syntax has been changed, thus mergemaster run is recommended. Now the "kod" parameter requires "limited" parameter. If the ntp.conf is not updated, the ntpd will run successfully, but with KoD disabled. 20151002: p28 FreeBSD-SA-15:24.rpcbind [revised] Revised patch to address a regression that prevents NIS from working. 20150929: p27 FreeBSD-SA-15:24.rpcbind Fix rpcbind(8) remote denial of service. [SA-15:24] 20150916: p26 FreeBSD-EN-15:18.pkg Implement pubkey support for pkg(7) bootstrap. [EN-15:18] 20150902: p25 FreeBSD-SA-15:23.bind Fix BIND remote denial of service vulnerability. [SA-15:23] 20150825: p24 FreeBSD-SA-15:21.amd64 FreeBSD-SA-15:22.openssh FreeBSD-EN-15:15.pkg Fix local privilege escalation in IRET handler. [SA-15:21] Fix OpenSSH multiple vulnerabilities. [SA-15:22] Fix insufficient check of unsupported pkg(7) signature methods. [EN-15:15] 20150818: p23 FreeBSD-SA-15:20.expat Fix multiple integer overflows in expat (libbsdxml) XML parser. [SA-15:20] 20150805: p22 FreeBSD-SA-15:19.routed Fix routed remote denial of service vulnerability. 20150728: p21 FreeBSD-SA-15:15.tcp FreeBSD-SA-15:16.openssh FreeBSD-SA-15:17.bind Fix resource exhaustion in TCP reassembly. [SA-15:15] Fix OpenSSH multiple vulnerabilities. [SA-15:16] Fix BIND remote denial of service vulnerability. [SA-15:17] 20150721: p20 FreeBSD-SA-15:13.tcp Fix resource exhaustion due to sessions stuck in LAST_ACK state. [SA-15:13] 20150707: p19 FreeBSD-SA-15:11.bind Fix BIND resolver remote denial of service when validating. 20150630: p18 FreeBSD-EN-15:08.sendmail [revised] FreeBSD-EN-15:09.xlocale Improvements to sendmail TLS/DH interoperability. [EN-15:08] Fix inconsistency between locale and rune locale states. [EN-15:09] 20150618: p17 FreeBSD-EN-15:08.sendmail Improvements to sendmail TLS/DH interoperability. [EN-15:08] 20150612: p16 FreeBSD-SA-15:10.openssl Fix multiple vulnerabilities in OpenSSL. [SA-15:10] 20150609: p15 FreeBSD-EN-15:06.file Updated base system file(1) to 5.22 to address multiple denial of service issues. 20150513: p14 FreeBSD-EN-15:04.freebsd-update Fix bug with freebsd-update(8) that does not ensure the previous upgrade was completed. [EN-15:04] 20150407: p13 FreeBSD-SA-15:04.igmp [revised] FreeBSD-SA-15:07.ntp FreeBSD-SA-15:09.ipv6 Improved patch for SA-15:04.igmp. Fix multiple vulnerabilities of ntp. [SA-15:07] Fix Denial of Service with IPv6 Router Advertisements. [SA-15:09] 20150320: p12 Fix patch for SA-15:06.openssl. 20150319: p11 FreeBSD-SA-15:06.openssl Fix multiple vulnerabilities in OpenSSL. [SA-15:06] 20150225: p10 FreeBSD-SA-15:04.igmp FreeBSD-SA-15:05.bind FreeBSD-EN-15:01.vt FreeBSD-EN-15:02.openssl FreeBSD-EN-15:03.freebsd-update Fix integer overflow in IGMP protocol. [SA-15:04] Fix BIND remote denial of service vulnerability. [SA-15:05] Fix vt(4) crash with improper ioctl parameters. [EN-15:01] Updated base system OpenSSL to 0.9.8zd. [EN-15:02] Fix freebsd-update libraries update ordering issue. [EN-15:03] 20150127: p9 FreeBSD-SA-15:02.kmem FreeBSD-SA-15:03.sctp Fix SCTP SCTP_SS_VALUE kernel memory corruption and disclosure vulnerability. [SA-15:02] Fix SCTP stream reset vulnerability. [SA-15:03] 20150114: p8 FreeBSD-SA-15:01.openssl Fix multiple vulnerabilities in OpenSSL. [SA-15:01] 20141223: p7 FreeBSD-SA-14:31.ntp FreeBSD-EN-14:13.freebsd-update Fix multiple vulnerabilities in NTP suite. [SA-14:31] Fix directory deletion issue in freebsd-update. [EN-14:13] 20141210: p6 FreeBSD-SA-14:28.file FreeBSD-SA-14:29.bind Fix multiple vulnerabilities in file(1) and libmagic(3). [SA-14:28] Fix BIND remote denial of service vulnerability. [SA-14:29] 20141104: p5 FreeBSD-SA-14:25.setlogin FreeBSD-SA-14:26.ftp FreeBSD-EN-14:12.zfs Fix kernel stack disclosure in setlogin(2) / getlogin(2). [SA-14:25] Fix remote command execution in ftp(1). [SA-14:26] Fix NFSv4 and ZFS cache consistency issue. [EN-14:12] 20141022: p4 FreeBSD-EN-14:10.tzdata FreeBSD-EN-14:11.crypt Time zone data file update. [EN-14:10] Change crypt(3) default hashing algorithm back to DES. [EN-14:11] 20141021: p3 FreeBSD-SA-14:20.rtsold FreeBSD-SA-14:21.routed FreeBSD-SA-14:22.namei FreeBSD-SA-14:23.openssl Fix rtsold(8) remote buffer overflow vulnerability. [SA-14:20] Fix routed(8) remote denial of service vulnerability. [SA-14:21] Fix memory leak in sandboxed namei lookup. [SA-14:22] Fix OpenSSL multiple vulnerabilities. [SA-14:23] 20140916: p2 FreeBSD-SA-14:19.tcp Fix Denial of Service in TCP packet processing. [SA-14:19] 20140909: p1 FreeBSD-SA-14:18.openssl Fix OpenSSL multiple vulnerabilities. [SA-14:18] 20140716: 9.3-RELEASE. 20140608: On i386 and amd64 systems, the onifconsole flag is now set by default in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled as a login TTY if it is set in the bootloader as an active kernel console. No changes in behavior should result otherwise. To revert to the previous behavior, set ttyu0 to "off" in /etc/ttys. 20140512: Clang and llvm have been upgraded to 3.4.1 release. 20140321: Clang and llvm have been upgraded to 3.4 release. 20140216: The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has been deprecated and will not be part of FreeBSD 11.0 and later releases. If you use this driver, please consider switching to the nfe(4) driver instead. 20131216: The behavior of gss_pseudo_random() for the krb5 mechanism has changed, for applications requesting a longer random string than produced by the underlying enctype's pseudo-random() function. In particular, the random string produced from a session key of enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will be different at the 17th octet and later, after this change. The counter used in the PRF+ construction is now encoded as a big-endian integer in accordance with RFC 4402. __FreeBSD_version is bumped to 902505. 20130930: 9.2-RELEASE. 20130823: Behavior of devfs rules path matching has been changed. Pattern is now always matched against fully qualified devfs path and slash characters must be explicitly matched by slashes in pattern (FNM_PATHNAME). Rulesets involving devfs subdirectories must be reviewed. 20130705: hastctl(8)'s `status' command output changed to terse one-liner format. Scripts using this should switch to `list' command or be rewritten. 20130618: Fix a bug that allowed a tracing process (e.g. gdb) to write to a memory-mapped file in the traced process's address space even if neither the traced process nor the tracing process had write access to that file. 20130605: Added ZFS TRIM support which is enabled by default. To disable ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf. Creating new ZFS pools and adding new devices to existing pools first performs a full device level TRIM which can take a significant amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0 to disable this behaviour. ZFS TRIM requires the underlying device support BIO_DELETE which is currently provided by methods such as ATA TRIM and SCSI UNMAP via CAM, which are typically supported by SSD's. Stats for ZFS TRIM can be monitored by looking at the sysctl's under kstat.zfs.misc.zio_trim. 20130524: `list' command has been added to hastctl(8). For now, it is full equivalent of `status' command. WARNING: in the near future the output of hastctl's status command will change to more terse format. If you use `hastctl status' for parsing in your scripts, switch to `hastctl list'. 20130430: The mergemaster command now uses the default MAKEOBJDIRPREFIX rather than creating it's own in the temporary directory in order allow access to bootstrapped versions of tools such as install and mtree. When upgrading from version of FreeBSD where the install command does not support -l, you will need to install a new mergemaster command if mergemaster -p is required. This can be accomplished with the command (cd src/usr.sbin/mergemaster && make install). Due to the use of the new -l option to install(1) during build and install, you must take care not to directly set the INSTALL make variable in your /etc/make.conf, /etc/src.conf, or on the command line. If you wish to use the -C flag for all installs you may be able to add INSTALL+=-C to /etc/make.conf or /etc/src.conf. 20130429: Fix a bug that allows NFS clients to issue READDIR on files. 20130315: The install(1) option -M has changed meaning and now takes an argument that is a file or path to append logs to. In the unlikely event that -M was the last option on the command line and the command line contained at least two files and a target directory the first file will have logs appended to it. The -M option served little practical purpose in the last decade so it's used expected to be extremely rare. 20130225: A new compression method (lz4) has been merged to. Please refer to zpool-features(7) for more information. Please refer to the "ZFS notes" section of this file for information on upgrading boot ZFS pools. 20121224: The VFS KBI was changed with the merge of several nullfs optimizations and fixes. All filesystem modules must be recompiled. 20121218: With the addition of auditdistd(8), a new auditdistd user is now depended on during installworld. "mergemaster -p" can be used to add the user prior to installworld, as documented in the handbook. 20121205: 9.1-RELEASE. 20121129: A new version of ZFS (pool version 5000) has been merged to 9-STABLE. Starting with this version the old system of ZFS pool versioning is superseded by "feature flags". This concept enables forward compatibility against certain future changes in functionality of ZFS pools. The first two read-only compatible "feature flags" for ZFS pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj". For more information read the new zpool-features(7) manual page. Please refer to the "ZFS notes" section of this file for information on upgrading boot ZFS pools. 20121114: The commit introducing bsd.compiler.mk breaks the traditional building of kernels before this point. Add -m ${SRC}/share/mk (for the right value of SRC) to your command lines to work around; update your useland to a point after this; or use the buildkernel/installkernel top-level targets. See also 20120829. 20121102: The IPFIREWALL_FORWARD kernel option has been removed. Its functionality now turned on by default. 20120913: The random(4) support for the VIA hardware random number generator (`PADLOCK') is no longer enabled unconditionally. Add the PADLOCK_RNG option in the custom kernel config if needed. The GENERIC kernels on i386 and amd64 do include the option, so the change only affects the custom kernel configurations. 20120829: The amd64 kernel now uses xsetbv, xrstor instructions. To compile with the traditional method, you must update your system with an installworld before the kernel will build. The documented make buildkernel/installkernel interfaces (coupled with fresh make kernel-toolchain) continue to work. 20120727: The sparc64 ZFS loader has been changed to no longer try to auto- detect ZFS providers based on diskN aliases but now requires these to be explicitly listed in the OFW boot-device environment variable. 20120422: Now unix domain sockets behave "as expected" on nullfs(5). Previously nullfs(5) did not pass through all behaviours to the underlying layer, as a result if we bound to a socket on the lower layer we could connect only to the lower path; if we bound to the upper layer we could connect only to the upper path. The new behavior is one can connect to both the lower and the upper paths regardless what layer path one binds to. 20120109: The acpi_wmi(4) status device /dev/wmistat has been renamed to /dev/wmistat0. 20120106: A new VOP_ADVISE() was added to support posix_fadvise(2). All filesystem modules must be recompiled. 20120106: The interface of the VOP_VPTOCNP(9) changed, now the returned vnode shall be referenced, previously it was required to be only held. All in-tree filesystems are converted. 20120106: 9.0-RELEASE. 20111101: The broken amd(4) driver has been replaced with esp(4) in the amd64, i386 and pc98 GENERIC kernel configuration files. 20110913: This commit modifies vfs_register() so that it uses a hash calculation to set vfc_typenum, which is enabled by default. The first time a system is booted after this change, the vfc_typenum values will change for all file systems. The main effect of this is a change to the NFS server file handles for file systems that use vfc_typenum in their fsid, such as ZFS. It will, however, prevent vfc_typenum from changing when file systems are loaded in a different order for subsequent reboots. To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf until you are ready to remount all NFS clients after a reboot. 20110828: Bump the shared library version numbers for libraries that do not use symbol versioning, have changed the ABI compared to stable/8 and which shared library version was not bumped. Done as part of 9.0-RELEASE cycle. 20110815: During the merge of Capsicum features, the fget(9) KPI was modified. This may require the rebuilding of out-of-tree device drivers -- issues have been reported specifically with the nVidia device driver. __FreeBSD_version is bumped to 900041. Also, there is a period between 20110811 and 20110814 where the special devices /dev/{stdin,stdout,stderr} did not work correctly. Building world from a kernel during that window may not work. 20110628: The packet filter (pf) code has been updated to OpenBSD 4.5. You need to update userland tools to be in sync with kernel. This update breaks backward compatibility with earlier pfsync(4) versions. Care must be taken when updating redundant firewall setups. 20110608: The following sysctls and tunables are retired on x86 platforms: machdep.hlt_cpus machdep.hlt_logical_cpus The following sysctl is retired: machdep.hyperthreading_allowed The sysctls were supposed to provide a way to dynamically offline and online selected CPUs on x86 platforms, but the implementation has not been reliable especially with SCHED_ULE scheduler. machdep.hyperthreading_allowed tunable is still available to ignore hyperthreading CPUs at OS level. Individual CPUs can be disabled using hint.lapic.X.disabled tunable, where X is an APIC ID of a CPU. Be advised, though, that disabling CPUs in non-uniform fashion will result in non-uniform topology and may lead to sub-optimal system performance with SCHED_ULE, which is a default scheduler. 20110607: cpumask_t type is retired and cpuset_t is used in order to describe a mask of CPUs. 20110531: Changes to ifconfig(8) for dynamic address family detection mandate that you are running a kernel of 20110525 or later. Make sure to follow the update procedure to boot a new kernel before installing world. 20110513: Support for sun4v architecture is officially dropped 20110503: Several KPI breaking changes have been committed to the mii(4) layer, the PHY drivers and consequently some Ethernet drivers using mii(4). This means that miibus.ko and the modules of the affected Ethernet drivers need to be recompiled. Note to kernel developers: Given that the OUI bit reversion problem was fixed as part of these changes all mii(4) commits related to OUIs, i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific handling, no longer can be merged verbatim to stable/8 and previous branches. 20110430: Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci' into their kernel configurations along with 'device pci'. 20110427: The default NFS client is now the new NFS client, so fstype "newnfs" is now "nfs" and the regular/old NFS client is now fstype "oldnfs". Although mounts via fstype "nfs" will usually work without userland changes, it is recommended that the mount(8) and mount_nfs(8) commands be rebuilt from sources and that a link to mount_nfs called mount_oldnfs be created. The new client is compiled into the kernel with "options NFSCL" and this is needed for diskless root file systems. The GENERIC kernel configs have been changed to use NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER. To use the regular/old client, you can "mount -t oldnfs ...". For a diskless root file system, you must also include a line like: vfs.root.mountfrom="oldnfs:" in the boot/loader.conf on the root fs on the NFS server to make a diskless root fs use the old client. 20110424: The GENERIC kernels for all architectures now default to the new CAM-based ATA stack. It means that all legacy ATA drivers were removed and replaced by respective CAM drivers. If you are using ATA device names in /etc/fstab or other places, make sure to update them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY, where 'Y's are the sequential numbers starting from zero for each type in order of detection, unless configured otherwise with tunables, see cam(4)). There will be symbolic links created in /dev/ to map old adX devices to the respective adaY. They should provide basic compatibility for file systems mounting in most cases, but they do not support old user-level APIs and do not have respective providers in GEOM. Consider using updated management tools with new device names. It is possible to load devices ahci, ata, siis and mvs as modules, but option ATA_CAM should remain in kernel configuration to make ata module work as CAM driver supporting legacy ATA controllers. Device ata still can be used in modular fashion (atacore + ...). Modules atadisk and atapi* are not used and won't affect operation in ATA_CAM mode. Note that to use CAM-based ATA kernel should include CAM devices scbus, pass, da (or explicitly ada), cd and optionally others. All of them are parts of the cam module. ataraid(4) functionality is now supported by the RAID GEOM class. To use it you can load geom_raid kernel module and use graid(8) tool for management. Instead of /dev/arX device names, use /dev/raid/rX. No kernel config options or code have been removed, so if a problem arises, please report it and optionally revert to the old ATA stack. In order to do it you can remove from the kernel config: options ATA_CAM device ahci device mvs device siis , and instead add back: device atadisk # ATA disk drives device ataraid # ATA RAID drives device atapicd # ATAPI CDROM drives device atapifd # ATAPI floppy drives device atapist # ATAPI tape drives 20110423: The default NFS server has been changed to the new server, which was referred to as the experimental server. If you need to switch back to the old NFS server, you must now put the "-o" option on both the mountd and nfsd commands. This can be done using the mountd_flags and nfs_server_flags rc.conf variables until an update to the rc scripts is committed, which is coming soon. 20110418: The GNU Objective-C runtime library (libobjc), and other Objective-C related components have been removed from the base system. If you require an Objective-C library, please use one of the available ports. 20110331: ath(4) has been split into bus- and device- modules. if_ath contains the HAL, the TX rate control and the network device code. if_ath_pci contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb contains the AHB glue. Users need to load both if_ath_pci and if_ath in order to use ath on everything else. TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only need to load if_ath and if_ath_pci for ath(4) operation. 20110314: As part of the replacement of sysinstall, the process of building release media has changed significantly. For details, please re-read release(7), which has been updated to reflect the new build process. 20110218: GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This is the last available version under GPLv2. It brings a number of new features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE 4.1 and SSE 4.2), better support for powerpc64, a number of new directives, and lots of other small improvements. See the ChangeLog file in contrib/binutils for the full details. 20110218: IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868 compliant, and will now use half of hash for authentication. This will break interoperability with all stacks (including all actual FreeBSD versions) who implement draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for authentication). The only workaround with such peers is to use another HMAC algorithm for IPsec ("phase 2") authentication. 20110207: Remove the uio_yield prototype and symbol. This function has been misnamed since it was introduced and should not be globally exposed with this name. The equivalent functionality is now available using kern_yield(curthread->td_user_pri). The function remains undocumented. 20110112: A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers, symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking for scalar sysctls is defined but disabled. Code that needs UQUAD to pass the type checking that must compile on older systems where the define is not present can check against __FreeBSD_version >= 900030. The system dialog(1) has been replaced with a new version previously in ports as devel/cdialog. dialog(1) is mostly command-line compatible with the previous version, but the libdialog associated with it has a largely incompatible API. As such, the original version of libdialog will be kept temporarily as libodialog, until its base system consumers are replaced or updated. Bump __FreeBSD_version to 900030. 20110103: If you are trying to run make universe on a -stable system, and you get the following warning: "Makefile", line 356: "Target architecture for i386/conf/GENERIC unknown. config(8) likely too old." or something similar to it, then you must upgrade your -stable system to 8.2-Release or newer (really, any time after r210146 7/15/2010 in stable/8) or build the config from the latest stable/8 branch and install it on your system. Prior to this date, building a current universe on 8-stable system from between 7/15/2010 and 1/2/2011 would result in a weird shell parsing error in the first kernel build phase. A new config on those old systems will fix that problem for older versions of -current. 20101228: The TCP stack has been modified to allow Khelp modules to interact with it via helper hook points and store per-connection data in the TCP control block. Bump __FreeBSD_version to 900029. User space tools that rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to be recompiled. 20101114: Generic IEEE 802.3 annex 31B full duplex flow control support has been added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted to take advantage of it instead of using custom implementations. This means that these drivers now no longer unconditionally advertise support for flow control but only do so if flow control is a selected media option. This was implemented in the generic support that way in order to allow flow control to be switched on and off via ifconfig(8) with the PHY specific default to typically off in order to protect from unwanted effects. Consequently, if you used flow control with one of the above mentioned drivers you now need to explicitly enable it, for example via: ifconfig bge0 media auto mediaopt flowcontrol Along with the above mentioned changes generic support for setting 1000baseT master mode also has been added and brgphy(4), ciphy(4), e1000phy(4) as well as ip1000phy(4) have been converted to take advantage of it. This means that these drivers now no longer take the link0 parameter for selecting master mode but the master media option has to be used instead, for example like in the following: ifconfig bge0 media 1000baseT mediaopt full-duplex,master Selection of master mode now is also available with all other PHY drivers supporting 1000baseT. 20101111: The TCP stack has received a significant update to add support for modularised congestion control and generally improve the clarity of congestion control decisions. Bump __FreeBSD_version to 900025. User space tools that rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to be recompiled. 20101002: The man(1) utility has been replaced by a new version that no longer uses /etc/manpath.config. Please consult man.conf(5) for how to migrate local entries to the new format. 20100928: The copyright strings printed by login(1) and sshd(8) at the time of a new connection have been removed to follow other operating systems and upstream sshd. 20100915: A workaround for a fixed ld bug has been removed in kernel code, so make sure that your system ld is built from sources after revision 210245 from 2010-07-19 (r211583 if building head kernel on stable/8, r211584 for stable/7; both from 2010-08-21). A symptom of incorrect ld version is different addresses for set_pcpu section and __start_set_pcpu symbol in kernel and/or modules. 20100913: The $ipv6_prefer variable in rc.conf(5) has been split into $ip6addrctl_policy and $ipv6_activate_all_interfaces. The $ip6addrctl_policy is a variable to choose a pre-defined address selection policy set by ip6addrctl(8). A value "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The default is "AUTO". The $ipv6_activate_all_interfaces specifies whether IFDISABLED flag (see an entry of 20090926) is set on an interface with no corresponding $ifconfig_IF_ipv6 line. The default is "NO" for security reason. If you want IPv6 link-local address on all interfaces by default, set this to "YES". The old ipv6_prefer="YES" is equivalent to ipv6_activate_all_interfaces="YES" and ip6addrctl_policy="ipv6_prefer". 20100913: DTrace has grown support for userland tracing. Due to this, DTrace is now i386 and amd64 only. dtruss(1) is now installed by default on those systems and a new kernel module is needed for userland tracing: fasttrap. No changes to your kernel config file are necessary to enable userland tracing, but you might consider adding 'STRIP=' and 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want to have informative userland stack traces in DTrace (ustack). 20100725: The acpi_aiboost(4) driver has been removed in favor of the new aibs(4) driver. You should update your kernel configuration file. 20100722: BSD grep has been imported to the base system and it is built by default. It is completely BSD licensed, highly GNU-compatible, uses less memory than its GNU counterpart and has a small codebase. However, it is slower than its GNU counterpart, which is mostly noticeable for larger searches, for smaller ones it is measurable but not significant. The reason is complex, the most important factor is that we lack a modern and efficient regex library and GNU overcomes this by optimizing the searches internally. Future work on improving the regex performance is planned, for the meantime, users that need better performance, can build GNU grep instead by setting the WITH_GNU_GREP knob. 20100713: Due to the import of powerpc64 support, all existing powerpc kernel configuration files must be updated with a machine directive like this: machine powerpc powerpc In addition, an updated config(8) is required to build powerpc kernels after this change. 20100713: A new version of ZFS (version 15) has been merged to -HEAD. This version uses a python library for the following subcommands: zfs allow, zfs unallow, zfs groupspace, zfs userspace. For full functionality of these commands the following port must be installed: sysutils/py-zfs 20100429: 'vm_page's are now hashed by physical address to an array of mutexes. Currently this is only used to serialize access to hold_count. Over time the page queue mutex will be peeled away. This changes the size of pmap on every architecture. And requires all callers of vm_page_hold and vm_page_unhold to be updated. 20100402: WITH_CTF can now be specified in src.conf (not recommended, there are some problems with static executables), make.conf (would also affect ports which do not use GNU make and do not override the compile targets) or in the kernel config (via "makeoptions WITH_CTF=yes"). When WITH_CTF was specified there before this was silently ignored, so make sure that WITH_CTF is not used in places which could lead to unwanted behavior. 20100311: The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32 to allow 32-bit compatibility on non-x86 platforms. All kernel configurations on amd64 and ia64 platforms using these options must be modified accordingly. 20100113: The utmp user accounting database has been replaced with utmpx, the user accounting interface standardized by POSIX. Unfortunately the semantics of utmp and utmpx don't match, making it practically impossible to support both interfaces. The user accounting database is used by tools like finger(1), last(1), talk(1), w(1) and ac(8). All applications in the base system use utmpx. This means only local binaries (e.g. from the ports tree) may still use these utmp database files. These applications must be rebuilt to make use of utmpx. After the system has been upgraded, it is safe to remove the old log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*), assuming their contents is of no importance anymore. Old wtmp databases can only be used by last(1) and ac(8) after they have been converted to the new format using wtmpcvt(1). 20100108: Introduce the kernel thread "deadlock resolver" (which can be enabled via the DEADLKRES option, see NOTES for more details) and the sleepq_type() function for sleepqueues. 20091202: The rc.firewall and rc.firewall6 were unified, and rc.firewall6 and rc.d/ip6fw were removed. According to the removal of rc.d/ip6fw, ipv6_firewall_* rc variables are obsoleted. Instead, the following new rc variables are added to rc.d/ipfw: firewall_client_net_ipv6, firewall_simple_iif_ipv6, firewall_simple_inet_ipv6, firewall_simple_oif_ipv6, firewall_simple_onet_ipv6, firewall_trusted_ipv6 The meanings correspond to the relevant IPv4 variables. 20091125: 8.0-RELEASE. 20091113: The default terminal emulation for syscons(4) has been changed from cons25 to xterm on all platforms except pc98. This means that the /etc/ttys file needs to be updated to ensure correct operation of applications on the console. The terminal emulation style can be toggled per window by using vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration options can be used to change the compile-time default back to cons25. To prevent graphical artifacts, make sure the TERM environment variable is set to match the terminal emulation that is being performed by syscons(4). 20091109: The layout of the structure ieee80211req_scan_result has changed. Applications that require wireless scan results (e.g. ifconfig(8)) from net80211 need to be recompiled. Applications such as wpa_supplicant(8) may require a full world build without using NO_CLEAN in order to get synchronized with the new structure. 20091025: The iwn(4) driver has been updated to support the 5000 and 5150 series. There's one kernel module for each firmware. Adding "device iwnfw" to the kernel configuration file means including all three firmware images inside the kernel. If you want to include just the one for your wireless card, use the devices iwn4965fw, iwn5000fw or iwn5150fw. 20090926: The rc.d/network_ipv6, IPv6 configuration script has been integrated into rc.d/netif. The changes are the following: 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF for IPv4. For aliases, $ifconfig_IF_aliasN should be used. Note that both variables need the "inet6" keyword at the head. Do not set $ipv6_network_interfaces manually if you do not understand what you are doing. It is not needed in most cases. $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but they are obsolete. 2. $ipv6_enable is obsolete. Use $ipv6_prefer and "inet6 accept_rtadv" keyword in ifconfig(8) instead. If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and all configured interfaces have "inet6 accept_rtadv" in the $ifconfig_IF_ipv6. These are for backward compatibility. 3. A new variable $ipv6_prefer has been added. If NO, IPv6 functionality of interfaces with no corresponding $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag, and the default address selection policy of ip6addrctl(8) is the IPv4-preferred one (see rc.d/ip6addrctl for more details). Note that if you want to configure IPv6 functionality on the disabled interfaces after boot, first you need to clear the flag by using ifconfig(8) like: ifconfig em0 inet6 -ifdisabled If YES, the default address selection policy is set as IPv6-preferred. The default value of $ipv6_prefer is NO. 4. If your system need to receive Router Advertisement messages, define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8) scripts automatically invoke rtsol(8) when the interface becomes UP. The Router Advertisement messages are used for SLAAC (State-Less Address AutoConfiguration). 20090922: 802.11s D3.03 support was committed. This is incompatible with the previous code, which was based on D3.0. 20090912: A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to control whether accepting Router Advertisement messages or not. Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and a sysctl variable net.inet6.ip6.auto_linklocal is its default value. The ifconfig(8) utility now supports these flags. 20090910: ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for mount(8) and -a option for df(1) to see them. 20090825: The old tunable hw.bus.devctl_disable has been superseded by hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be replaced by hw.bus.devctl_queue=0. The default for this new tunable is 1000. 20090813: Remove the option STOP_NMI. The default action is now to use NMI only for KDB via the newly introduced function stop_cpus_hard() and maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64. 20090803: The stable/8 branch created in subversion. This corresponds to the RELENG_8 branch in CVS. 20090719: Bump the shared library version numbers for all libraries that do not use symbol versioning as part of the 8.0-RELEASE cycle. Bump __FreeBSD_version to 800105. 20090714: Due to changes in the implementation of virtual network stack support, all network-related kernel modules must be recompiled. As this change breaks the ABI, bump __FreeBSD_version to 800104. 20090713: The TOE interface to the TCP syncache has been modified to remove struct tcpopt () from the ABI of the network stack. The cxgb driver is the only TOE consumer affected by this change, and needs to be recompiled along with the kernel. As this change breaks the ABI, bump __FreeBSD_version to 800103. 20090712: Padding has been added to struct tcpcb, sackhint and tcpstat in to facilitate future MFCs and bug fixes whilst maintaining the ABI. However, this change breaks the ABI, so bump __FreeBSD_version to 800102. User space tools that rely on the size of any of these structs (e.g. sockstat) need to be recompiled. 20090630: The NFS_LEGACYRPC option has been removed along with the old kernel RPC implementation that this option selected. Kernel configurations may need to be adjusted. 20090629: The network interface device nodes at /dev/net/ have been removed. All ioctl operations can be performed the normal way using routing sockets. The kqueue functionality can generally be replaced with routing sockets. 20090628: The documentation from the FreeBSD Documentation Project (Handbook, FAQ, etc.) is now installed via packages by sysinstall(8) and under the /usr/local/share/doc/freebsd directory instead of /usr/share/doc. 20090624: The ABI of various structures related to the SYSV IPC API have been changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to 800100. 20090622: Layout of struct vnet has changed as routing related variables were moved to their own Vimage module. Modules need to be recompiled. Bump __FreeBSD_version to 800099. 20090619: NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024 respectively. As long as no more than 16 groups per process are used, no changes should be visible. When more than 16 groups are used, old binaries may fail if they call getgroups() or getgrouplist() with statically sized storage. Recompiling will work around this, but applications should be modified to use dynamically allocated storage for group arrays as POSIX.1-2008 does not cap an implementation's number of supported groups at NGROUPS_MAX+1 as previous versions did. NFS and portalfs mounts may also be affected as the list of groups is truncated to 16. Users of NFS who use more than 16 groups, should take care that negative group permissions are not used on the exported file systems as they will not be reliable unless a GSSAPI based authentication method is used. 20090616: The compiling option ADAPTIVE_LOCKMGRS has been introduced. This option compiles in the support for adaptive spinning for lockmgrs which want to enable it. The lockinit() function now accepts the flag LK_ADAPTIVE in order to make the lock object subject to adaptive spinning when both held in write and read mode. 20090613: The layout of the structure returned by IEEE80211_IOC_STA_INFO has changed. User applications that use this ioctl need to be rebuilt. 20090611: The layout of struct thread has changed. Kernel and modules need to be rebuilt. 20090608: The layout of structs ifnet, domain, protosw and vnet_net has changed. Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097. 20090602: window(1) has been removed from the base system. It can now be installed from ports. The port is called misc/window. 20090601: The way we are storing and accessing `routing table' entries has changed. Programs reading the FIB, like netstat, need to be re-compiled. 20090601: A new netisr implementation has been added for FreeBSD 8. Network file system modules, such as igmp, ipdivert, and others, should be rebuilt. Bump __FreeBSD_version to 800096. 20090530: Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no more valid. 20090530: Add VOP_ACCESSX(9). File system modules need to be rebuilt. Bump __FreeBSD_version to 800094. 20090529: Add mnt_xflag field to 'struct mount'. File system modules need to be rebuilt. Bump __FreeBSD_version to 800093. 20090528: The compiling option ADAPTIVE_SX has been retired while it has been introduced the option NO_ADAPTIVE_SX which handles the reversed logic. The KPI for sx_init_flags() changes as accepting flags: SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has been introduced in order to handle the reversed logic. Bump __FreeBSD_version to 800092. 20090527: Add support for hierarchical jails. Remove global securelevel. Bump __FreeBSD_version to 800091. 20090523: The layout of struct vnet_net has changed, therefore modules need to be rebuilt. Bump __FreeBSD_version to 800090. 20090523: The newly imported zic(8) produces a new format in the output. Please run tzsetup(8) to install the newly created data to /etc/localtime. 20090520: The sysctl tree for the usb stack has renamed from hw.usb2.* to hw.usb.* and is now consistent again with previous releases. 20090520: 802.11 monitor mode support was revised and driver api's were changed. Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead of DLT_IEEE802_11. No user-visible data structures were changed but applications that use DLT_IEEE802_11 may require changes. Bump __FreeBSD_version to 800088. 20090430: The layout of the following structs has changed: sysctl_oid, socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet, vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or panics may be experienced. World rebuild is required for correctly checking networking state from userland. Bump __FreeBSD_version to 800085. 20090429: MLDv2 and Source-Specific Multicast (SSM) have been merged to the IPv6 stack. VIMAGE hooks are in but not yet used. The implementation of SSM within FreeBSD's IPv6 stack closely follows the IPv4 implementation. For kernel developers: * The most important changes are that the ip6_output() and ip6_input() paths no longer take the IN6_MULTI_LOCK, and this lock has been downgraded to a non-recursive mutex. * As with the changes to the IPv4 stack to support SSM, filtering of inbound multicast traffic must now be performed by transport protocols within the IPv6 stack. This does not apply to TCP and SCTP, however, it does apply to UDP in IPv6 and raw IPv6. * The KPIs used by IPv6 multicast are similar to those used by the IPv4 stack, with the following differences: * im6o_mc_filter() is analogous to imo_multicast_filter(). * The legacy KAME entry points in6_joingroup and in6_leavegroup() are shimmed to in6_mc_join() and in6_mc_leave() respectively. * IN6_LOOKUP_MULTI() has been deprecated and removed. * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs for MLDv1 have an additional 'timer' argument which is used to jitter the initial membership report for the solicited-node multicast membership on-link. * This is not strictly needed for MLDv2, which already jitters its report transmissions. However, the 'timer' argument is preserved in case MLDv1 is active on the interface. * The KAME linked-list based IPv6 membership implementation has been refactored to use a vector similar to that used by the IPv4 stack. Code which maintains a list of its own multicast memberships internally, e.g. carp, has been updated to reflect the new semantics. * There is a known Lock Order Reversal (LOR) due to in6_setscope() acquiring the IF_AFDATA_LOCK and being called within ip6_output(). Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an implementation constraint which needs to be addressed in HEAD. For application developers: * The changes are broadly similar to those made for the IPv4 stack. * The use of IPv4 and IPv6 multicast socket options on the same socket, using mapped addresses, HAS NOT been tested or supported. * There are a number of issues with the implementation of various IPv6 multicast APIs which need to be resolved in the API surface before the implementation is fully compatible with KAME userland use, and these are mostly to do with interface index treatment. * The literature available discusses the use of either the delta / ASM API with setsockopt(2)/getsockopt(2), or the full-state / ASM API using setsourcefilter(3)/getsourcefilter(3). For more information please refer to RFC 3768, 'Socket Interface Extensions for Multicast Source Filters'. * Applications which use the published RFC 3678 APIs should be fine. For systems administrators: * The mtest(8) utility has been refactored to support IPv6, in addition to IPv4. Interface addresses are no longer accepted as arguments, their names must be used instead. The utility will map the interface name to its first IPv4 address as returned by getifaddrs(3). * The ifmcstat(8) utility has also been updated to print the MLDv2 endpoint state and source filter lists via sysctl(3). * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable loopback of IPv6 multicast datagrams by default; it defaults to 1 to preserve the existing behaviour. Disabling multicast loopback is recommended for optimal system performance. * The IPv6 MROUTING code has been changed to examine this sysctl instead of attempting to perform a group lookup before looping back forwarded datagrams. Bump __FreeBSD_version to 800084. 20090422: Implement low-level Bluetooth HCI API. Bump __FreeBSD_version to 800083. 20090419: The layout of struct malloc_type, used by modules to register new memory allocation types, has changed. Most modules will need to be rebuilt or panics may be experienced. Bump __FreeBSD_version to 800081. 20090415: Anticipate overflowing inp_flags - add inp_flags2. This changes most offsets in inpcb, so checking v4 connection state will require a world rebuild. Bump __FreeBSD_version to 800080. 20090415: Add an llentry to struct route and struct route_in6. Modules embedding a struct route will need to be recompiled. Bump __FreeBSD_version to 800079. 20090414: The size of rt_metrics_lite and by extension rtentry has changed. Networking administration apps will need to be recompiled. The route command now supports show as an alias for get, weighting of routes, sticky and nostick flags to alter the behavior of stateful load balancing. Bump __FreeBSD_version to 800078. 20090408: Do not use Giant for kbdmux(4) locking. This is wrong and apparently causing more problems than it solves. This will re-open the issue where interrupt handlers may race with kbdmux(4) in polling mode. Typical symptoms include (but not limited to) duplicated and/or missing characters when low level console functions (such as gets) are used while interrupts are enabled (for example geli password prompt, mountroot prompt etc.). Disabling kbdmux(4) may help. 20090407: The size of structs vnet_net, vnet_inet and vnet_ipfw has changed; kernel modules referencing any of the above need to be recompiled. Bump __FreeBSD_version to 800075. 20090320: GEOM_PART has become the default partition slicer for storage devices, replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It introduces some changes: MSDOS/EBR: the devices created from MSDOS extended partition entries (EBR) can be named differently than with GEOM_MBR and are now symlinks to devices with offset-based names. fstabs may need to be modified. BSD: the "geometry does not match label" warning is harmless in most cases but it points to problems in file system misalignment with disk geometry. The "c" partition is now implicit, covers the whole top-level drive and cannot be (mis)used by users. General: Kernel dumps are now not allowed to be written to devices whose partition types indicate they are meant to be used for file systems (or, in case of MSDOS partitions, as something else than the "386BSD" type). Most of these changes date approximately from 200812. 20090319: The uscanner(4) driver has been removed from the kernel. This follows Linux removing theirs in 2.6 and making libusb the default interface (supported by sane). 20090319: The multicast forwarding code has been cleaned up. netstat(1) only relies on KVM now for printing bandwidth upcall meters. The IPv4 and IPv6 modules are split into ip_mroute_mod and ip6_mroute_mod respectively. The config(5) options for statically compiling this code remain the same, i.e. 'options MROUTING'. 20090315: Support for the IFF_NEEDSGIANT network interface flag has been removed, which means that non-MPSAFE network device drivers are no longer supported. In particular, if_ar, if_sr, and network device drivers from the old (legacy) USB stack can no longer be built or used. 20090313: POSIX.1 Native Language Support (NLS) has been enabled in libc and a bunch of new language catalog files have also been added. This means that some common libc messages are now localized and they depend on the LC_MESSAGES environmental variable. 20090313: The k8temp(4) driver has been renamed to amdtemp(4) since support for Family 10 and Family 11 CPU families was added. 20090309: IGMPv3 and Source-Specific Multicast (SSM) have been merged to the IPv4 stack. VIMAGE hooks are in but not yet used. For kernel developers, the most important changes are that the ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(), and this lock has been downgraded to a non-recursive mutex. Transport protocols (UDP, Raw IP) are now responsible for filtering inbound multicast traffic according to group membership and source filters. The imo_multicast_filter() KPI exists for this purpose. Transports which do not use multicast (SCTP, TCP) already reject multicast by default. Forwarding and receive performance may improve as a mutex acquisition is no longer needed in the ip_input() low-level input path. in_addmulti() and in_delmulti() are shimmed to new KPIs which exist to support SSM in-kernel. For application developers, it is recommended that loopback of multicast datagrams be disabled for best performance, as this will still cause the lock to be taken for each looped-back datagram transmission. The net.inet.ip.mcast.loop sysctl may be tuned to 0 to disable loopback by default; it defaults to 1 to preserve the existing behaviour. For systems administrators, to obtain best performance with multicast reception and multiple groups, it is always recommended that a card with a suitably precise hash filter is used. Hash collisions will still result in the lock being taken within the transport protocol input path to check group membership. If deploying FreeBSD in an environment with IGMP snooping switches, it is recommended that the net.inet.igmp.sendlocal sysctl remain enabled; this forces 224.0.0.0/24 group membership to be announced via IGMP. The size of 'struct igmpstat' has changed; netstat needs to be recompiled to reflect this. Bump __FreeBSD_version to 800070. 20090309: libusb20.so.1 is now installed as libusb.so.1 and the ports system updated to use it. This requires a buildworld/installworld in order to update the library and dependencies (usbconfig, etc). Its advisable to rebuild all ports which uses libusb. More specific directions are given in the ports collection UPDATING file. Any /etc/libmap.conf entries for libusb are no longer required and can be removed. 20090302: A workaround is committed to allow the creation of System V shared memory segment of size > 2 GB on the 64-bit architectures. Due to a limitation of the existing ABI, the shm_segsz member of the struct shmid_ds, returned by shmctl(IPC_STAT) call is wrong for large segments. Note that limits must be explicitly raised to allow such segments to be created. 20090301: The layout of struct ifnet has changed, requiring a rebuild of all network device driver modules. 20090227: The /dev handling for the new USB stack has changed, a buildworld/installworld is required for libusb20. 20090223: The new USB2 stack has now been permanently moved in and all kernel and module names reverted to their previous values (eg, usb, ehci, ohci, ums, ...). The old usb stack can be compiled in by prefixing the name with the letter 'o', the old usb modules have been removed. Updating entry 20090216 for xorg and 20090215 for libmap may still apply. 20090217: The rc.conf(5) option if_up_delay has been renamed to defaultroute_delay to better reflect its purpose. If you have customized this setting in /etc/rc.conf you need to update it to use the new name. 20090216: xorg 7.4 wants to configure its input devices via hald which does not yet work with USB2. If the keyboard/mouse does not work in xorg then add Option "AllowEmptyInput" "off" to your ServerLayout section. This will cause X to use the configured kbd and mouse sections from your xorg.conf. 20090215: The GENERIC kernels for all architectures now default to the new USB2 stack. No kernel config options or code have been removed so if a problem arises please report it and optionally revert to the old USB stack. If you are loading USB kernel modules or have a custom kernel that includes GENERIC then ensure that usb names are also changed over, eg uftdi -> usb2_serial_ftdi. Older programs linked against the ports libusb 0.1 need to be redirected to the new stack's libusb20. /etc/libmap.conf can be used for this: # Map old usb library to new one for usb2 stack libusb-0.1.so.8 libusb20.so.1 20090209: All USB ethernet devices now attach as interfaces under the name ueN (eg. ue0). This is to provide a predictable name as vendors often change usb chipsets in a product without notice. 20090203: The ichsmb(4) driver has been changed to require SMBus slave addresses be left-justified (xxxxxxx0b) rather than right-justified. All of the other SMBus controller drivers require left-justified slave addresses, so this change makes all the drivers provide the same interface. 20090201: INET6 statistics (struct ip6stat) was updated. netstat(1) needs to be recompiled. 20090119: NTFS has been removed from GENERIC kernel on amd64 to match GENERIC on i386. Should not cause any issues since mount_ntfs(8) will load ntfs.ko module automatically when NTFS support is actually needed, unless ntfs.ko is not installed or security level prohibits loading kernel modules. If either is the case, "options NTFS" has to be added into kernel config. 20090115: TCP Appropriate Byte Counting (RFC 3465) support added to kernel. New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to 800061. User space tools that rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to be recompiled. 20081225: ng_tty(4) module updated to match the new TTY subsystem. Due to API change, user-level applications must be updated. New API support added to mpd5 CVS and expected to be present in next mpd5.3 release. 20081219: With __FreeBSD_version 800060 the makefs tool is part of the base system (it was a port). 20081216: The afdata and ifnet locks have been changed from mutexes to rwlocks, network modules will need to be re-compiled. 20081214: __FreeBSD_version 800059 incorporates the new arp-v2 rewrite. RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated. The new code reduced struct rtentry{} by 16 bytes on 32-bit architecture and 40 bytes on 64-bit architecture. The userland applications "arp" and "ndp" have been updated accordingly. The output from "netstat -r" shows only routing entries and none of the L2 information. 20081130: __FreeBSD_version 800057 marks the switchover from the binary ath hal to source code. Users must add the line: options AH_SUPPORT_AR5416 to their kernel config files when specifying: device ath_hal The ath_hal module no longer exists; the code is now compiled together with the driver in the ath module. It is now possible to tailor chip support (i.e. reduce the set of chips and thereby the code size); consult ath_hal(4) for details. 20081121: __FreeBSD_version 800054 adds memory barriers to , new interfaces to ifnet to facilitate multiple hardware transmit queues for cards that support them, and a lock-less ring-buffer implementation to enable drivers to more efficiently manage queueing of packets. 20081117: A new version of ZFS (version 13) has been merged to -HEAD. This version has zpool attribute "listsnapshots" off by default, which means "zfs list" does not show snapshots, and is the same as Solaris behavior. 20081028: dummynet(4) ABI has changed. ipfw(8) needs to be recompiled. 20081009: The uhci, ohci, ehci and slhci USB Host controller drivers have been put into separate modules. If you load the usb module separately through loader.conf you will need to load the appropriate *hci module as well. E.g. for a UHCI-based USB 2.0 controller add the following to loader.conf: uhci_load="YES" ehci_load="YES" 20081009: The ABI used by the PMC toolset has changed. Please keep userland (libpmc(3)) and the kernel module (hwpmc(4)) in sync. 20081009: atapci kernel module now includes only generic PCI ATA driver. AHCI driver moved to ataahci kernel module. All vendor-specific code moved into separate kernel modules: ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia 20080820: The TTY subsystem of the kernel has been replaced by a new implementation, which provides better scalability and an improved driver model. Most common drivers have been migrated to the new TTY subsystem, while others have not. The following drivers have not yet been ported to the new TTY layer: PCI/ISA: cy, digi, rc, rp, sio USB: ubser, ucycom Line disciplines: ng_h4, ng_tty, ppp, sl, snp Adding these drivers to your kernel configuration file shall cause compilation to fail. 20080818: ntpd has been upgraded to 4.2.4p5. 20080801: OpenSSH has been upgraded to 5.1p1. For many years, FreeBSD's version of OpenSSH preferred DSA over RSA for host and user authentication keys. With this upgrade, we've switched to the vendor's default of RSA over DSA. This may cause upgraded clients to warn about unknown host keys even for previously known hosts. Users should follow the usual procedure for verifying host keys before accepting the RSA key. This can be circumvented by setting the "HostKeyAlgorithms" option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh command line. Please note that the sequence of keys offered for authentication has been changed as well. You may want to specify IdentityFile in a different order to revert this behavior. 20080713: The sio(4) driver has been removed from the i386 and amd64 kernel configuration files. This means uart(4) is now the default serial port driver on those platforms as well. To prevent collisions with the sio(4) driver, the uart(4) driver uses different names for its device nodes. This means the onboard serial port will now most likely be called "ttyu0" instead of "ttyd0". You may need to reconfigure applications to use the new device names. When using the serial port as a boot console, be sure to update /boot/device.hints and /etc/ttys before booting the new kernel. If you forget to do so, you can still manually specify the hints at the loader prompt: set hint.uart.0.at="isa" set hint.uart.0.port="0x3F8" set hint.uart.0.flags="0x10" set hint.uart.0.irq="4" boot -s 20080609: The gpt(8) utility has been removed. Use gpart(8) to partition disks instead. 20080603: The version that Linuxulator emulates was changed from 2.4.2 to 2.6.16. If you experience any problems with Linux binaries please try to set sysctl compat.linux.osrelease to 2.4.2 and if it fixes the problem contact emulation mailing list. 20080525: ISDN4BSD (I4B) was removed from the src tree. You may need to update a your kernel configuration and remove relevant entries. 20080509: I have checked in code to support multiple routing tables. See the man pages setfib(1) and setfib(2). This is a hopefully backwards compatible version, but to make use of it you need to compile your kernel with options ROUTETABLES=2 (or more up to 16). 20080420: The 802.11 wireless support was redone to enable multi-bss operation on devices that are capable. The underlying device is no longer used directly but instead wlanX devices are cloned with ifconfig. This requires changes to rc.conf files. For example, change: ifconfig_ath0="WPA DHCP" to wlans_ath0=wlan0 ifconfig_wlan0="WPA DHCP" see rc.conf(5) for more details. In addition, mergemaster of /etc/rc.d is highly recommended. Simultaneous update of userland and kernel wouldn't hurt either. As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta modules were merged into the base wlan module. All references to these modules (e.g. in kernel config files) must be removed. 20080408: psm(4) has gained write(2) support in native operation level. Arbitrary commands can be written to /dev/psm%d and status can be read back from it. Therefore, an application is responsible for status validation and error recovery. It is a no-op in other operation levels. 20080312: Support for KSE threading has been removed from the kernel. To run legacy applications linked against KSE libmap.conf may be used. The following libmap.conf may be used to ensure compatibility with any prior release: libpthread.so.1 libthr.so.1 libpthread.so.2 libthr.so.2 libkse.so.3 libthr.so.3 20080301: The layout of struct vmspace has changed. This affects libkvm and any executables that link against libkvm and use the kvm_getprocs() function. In particular, but not exclusively, it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1). The effects are minimal, but it's advisable to upgrade world nonetheless. 20080229: The latest em driver no longer has support in it for the 82575 adapter, this is now moved to the igb driver. The split was done to make new features that are incompatible with older hardware easier to do. 20080220: The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4), likewise the kernel option is now GEOM_LINUX_LVM. 20080211: The default NFS mount mode has changed from UDP to TCP for increased reliability. If you rely on (insecurely) NFS mounting across a firewall you may need to update your firewall rules. 20080208: Belatedly note the addition of m_collapse for compacting mbuf chains. 20080126: The fts(3) structures have been changed to use adequate integer types for their members and so to be able to cope with huge file trees. The old fts(3) ABI is preserved through symbol versioning in libc, so third-party binaries using fts(3) should still work, although they will not take advantage of the extended types. At the same time, some third-party software might fail to build after this change due to unportable assumptions made in its source code about fts(3) structure members. Such software should be fixed by its vendor or, in the worst case, in the ports tree. FreeBSD_version 800015 marks this change for the unlikely case that a portable fix is impossible. 20080123: To upgrade to -current after this date, you must be running FreeBSD not older than 6.0-RELEASE. Upgrading to -current from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems. 20071128: The ADAPTIVE_GIANT kernel option has been retired because its functionality is the default now. 20071118: The AT keyboard emulation of sunkbd(4) has been turned on by default. In order to make the special symbols of the Sun keyboards driven by sunkbd(4) work under X these now have to be configured the same way as Sun USB keyboards driven by ukbd(4) (which also does AT keyboard emulation), f.e.: Option "XkbLayout" "us" Option "XkbRules" "xorg" Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us" 20071024: It has been decided that it is desirable to provide ABI backwards compatibility to the FreeBSD 4/5/6 versions of the PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was broken with the introduction of PCI domain support (see the 20070930 entry). Unfortunately, this required the ABI of PCIOCGETCONF to be broken again in order to be able to provide backwards compatibility to the old version of that IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled again. As for prominent ports this affects neither pciutils nor xorg-server this time, the hal port needs to be rebuilt however. 20071020: The misnamed kthread_create() and friends have been renamed to kproc_create() etc. Many of the callers already used kproc_start().. I will return kthread_create() and friends in a while with implementations that actually create threads, not procs. Renaming corresponds with version 800002. 20071010: RELENG_7 branched. COMMON ITEMS: General Notes ------------- Avoid using make -j when upgrading. While generally safe, there are sometimes problems using -j to upgrade. If your upgrade fails with -j, please try again without -j. From time to time in the past there have been problems using -j with buildworld and/or installworld. This is especially true when upgrading between "distant" versions (eg one that cross a major release boundary or several minor releases, or when several months have passed on the -current branch). Sometimes, obscure build problems are the result of environment poisoning. This can happen because the make utility reads its environment when searching for values for global variables. To run your build attempts in an "environmental clean room", prefix all make commands with 'env -i '. See the env(1) manual page for more details. When upgrading from one major version to another it is generally best to upgrade to the latest code in the currently installed branch first, then do an upgrade to the new branch. This is the best-tested upgrade path, and has the highest probability of being successful. Please try this approach before reporting problems with a major version upgrade. ZFS notes --------- When upgrading the boot ZFS pool to a new version, always follow these two steps: 1.) recompile and reinstall the ZFS boot loader and boot block (this is part of "make buildworld" and "make installworld") 2.) update the ZFS boot block on your boot drive The following example updates the ZFS boot block on the first partition (freebsd-boot) of a GPT partitioned drive ad0: "gpart bootcode -p /boot/gptzfsboot -i 1 ad0" Non-boot pools do not need these updates. To build a kernel ----------------- If you are updating from a prior version of FreeBSD (even one just a few days old), you should follow this procedure. It is the most failsafe as it uses a /usr/obj tree with a fresh mini-buildworld, make kernel-toolchain make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE To test a kernel once --------------------- If you just want to boot a kernel once (because you are not sure if it works, or if you want to boot a known bad kernel to provide debugging information) run make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel nextboot -k testkernel To just build a kernel when you know that it won't mess you up -------------------------------------------------------------- This assumes you are already running a CURRENT system. Replace ${arch} with the architecture of your machine (e.g. "i386", "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc). cd src/sys/${arch}/conf config KERNEL_NAME_HERE cd ../compile/KERNEL_NAME_HERE make depend make make install If this fails, go to the "To build a kernel" section. To rebuild everything and install it on the current system. ----------------------------------------------------------- # Note: sometimes if you are running current you gotta do more than # is listed here if you are upgrading from a really old current. make buildworld make kernel KERNCONF=YOUR_KERNEL_HERE [1] [3] mergemaster -p [5] make installworld mergemaster -i [4] make delete-old [6] To cross-install current onto a separate partition -------------------------------------------------- # In this approach we use a separate partition to hold # current's root, 'usr', and 'var' directories. A partition # holding "/", "/usr" and "/var" should be about 2GB in # size. make buildworld make buildkernel KERNCONF=YOUR_KERNEL_HERE make installworld DESTDIR=${CURRENT_ROOT} make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT} cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd To upgrade in-place from stable to current ---------------------------------------------- make buildworld [9] make kernel KERNCONF=YOUR_KERNEL_HERE [8] [1] [3] mergemaster -p [5] make installworld mergemaster -i [4] make delete-old [6] Make sure that you've read the UPDATING file to understand the tweaks to various things you need. At this point in the life cycle of current, things change often and you are on your own to cope. The defaults can also change, so please read ALL of the UPDATING entries. Also, if you are tracking -current, you must be subscribed to freebsd-current@freebsd.org. Make sure that before you update your sources that you have read and understood all the recent messages there. If in doubt, please track -stable which has much fewer pitfalls. [1] If you have third party modules, such as vmware, you should disable them at this point so they don't crash your system on reboot. [3] From the bootblocks, boot -s, and then do fsck -p mount -u / mount -a cd src adjkerntz -i # if CMOS is wall time Also, when doing a major release upgrade, it is required that you boot into single user mode to do the installworld. [4] Note: This step is non-optional. Failure to do this step can result in a significant reduction in the functionality of the system. Attempting to do it by hand is not recommended and those that pursue this avenue should read this file carefully, as well as the archives of freebsd-current and freebsd-hackers mailing lists for potential gotchas. The -U option is also useful to consider. See mergemaster(8) for more information. [5] Usually this step is a noop. However, from time to time you may need to do this if you get unknown user in the following step. It never hurts to do it all the time. You may need to install a new mergemaster (cd src/usr.sbin/mergemaster && make install) after the buildworld before this step if you last updated from [78]-stable or 9-stable before 20130430. [6] This only deletes old files and directories. Old libraries can be deleted by "make delete-old-libs", but you have to make sure that no program is using those libraries anymore. [8] In order to have a kernel that can run the 4.x binaries needed to do an installworld, you must include the COMPAT_FREEBSD4 option in your kernel. Failure to do so may leave you with a system that is hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is required to run the 5.x binaries on more recent kernels. And so on for COMPAT_FREEBSD6 and COMPAT_FREEBSD7. Make sure that you merge any new devices from GENERIC since the last time you updated your kernel config file. [9] When checking out sources, you must include the -P flag to have cvs prune empty directories. If CPUTYPE is defined in your /etc/make.conf, make sure to use the "?=" instead of the "=" assignment operator, so that buildworld can override the CPUTYPE if it needs to. MAKEOBJDIRPREFIX must be defined in an environment variable, and not on the command line, or in /etc/make.conf. buildworld will warn if it is improperly defined. FORMAT: This file contains a list, in reverse chronological order, of major breakages in tracking -current. It is not guaranteed to be a complete list of such breakages, and only contains entries since October 10, 2007. If you need to see UPDATING entries from before that date, you will need to fetch an UPDATING file from an older FreeBSD release. Copyright information: Copyright 1998-2009 M. Warner Losh. All Rights Reserved. Redistribution, publication, translation and use, with or without modification, in full or in part, in any form or format of this document are permitted without further permission from the author. THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``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 WARNER LOSH 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. Contact Warner Losh if you have any questions about your use of this document. $FreeBSD$ Index: releng/9.3/contrib/bind9/bin/named/control.c =================================================================== --- releng/9.3/contrib/bind9/bin/named/control.c (revision 296610) +++ releng/9.3/contrib/bind9/bin/named/control.c (revision 296611) @@ -1,220 +1,220 @@ /* * Copyright (C) 2004-2007, 2009-2013 Internet Systems Consortium, Inc. ("ISC") * Copyright (C) 2001-2003 Internet Software Consortium. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH * REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY * AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT, * INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR * PERFORMANCE OF THIS SOFTWARE. */ /* $Id$ */ /*! \file */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef HAVE_LIBSCF #include #endif static isc_boolean_t command_compare(const char *text, const char *command) { unsigned int commandlen = strlen(command); if (strncasecmp(text, command, commandlen) == 0 && (text[commandlen] == '\0' || text[commandlen] == ' ' || text[commandlen] == '\t')) return (ISC_TRUE); return (ISC_FALSE); } /*% * This function is called to process the incoming command * when a control channel message is received. */ isc_result_t ns_control_docommand(isccc_sexpr_t *message, isc_buffer_t *text) { isccc_sexpr_t *data; char *command = NULL; isc_result_t result; int log_level; #ifdef HAVE_LIBSCF ns_smf_want_disable = 0; #endif data = isccc_alist_lookup(message, "_data"); - if (data == NULL) { + if (!isccc_alist_alistp(data)) { /* * No data section. */ return (ISC_R_FAILURE); } result = isccc_cc_lookupstring(data, "type", &command); if (result != ISC_R_SUCCESS) { /* * We have no idea what this is. */ return (result); } /* * Compare the 'command' parameter against all known control commands. */ if (command_compare(command, NS_COMMAND_NULL) || command_compare(command, NS_COMMAND_STATUS)) { log_level = ISC_LOG_DEBUG(1); } else { log_level = ISC_LOG_INFO; } isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, log_level, "received control channel command '%s'", command); if (command_compare(command, NS_COMMAND_RELOAD)) { result = ns_server_reloadcommand(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_RECONFIG)) { result = ns_server_reconfigcommand(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_REFRESH)) { result = ns_server_refreshcommand(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_RETRANSFER)) { result = ns_server_retransfercommand(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_HALT)) { #ifdef HAVE_LIBSCF /* * If we are managed by smf(5), AND in chroot, then * we cannot connect to the smf repository, so just * return with an appropriate message back to rndc. */ if (ns_smf_got_instance == 1 && ns_smf_chroot == 1) { result = ns_smf_add_message(text); return (result); } /* * If we are managed by smf(5) but not in chroot, * try to disable ourselves the smf way. */ if (ns_smf_got_instance == 1 && ns_smf_chroot == 0) ns_smf_want_disable = 1; /* * If ns_smf_got_instance = 0, ns_smf_chroot * is not relevant and we fall through to * isc_app_shutdown below. */ #endif /* Do not flush master files */ ns_server_flushonshutdown(ns_g_server, ISC_FALSE); ns_os_shutdownmsg(command, text); isc_app_shutdown(); result = ISC_R_SUCCESS; } else if (command_compare(command, NS_COMMAND_STOP)) { /* * "stop" is the same as "halt" except it does * flush master files. */ #ifdef HAVE_LIBSCF if (ns_smf_got_instance == 1 && ns_smf_chroot == 1) { result = ns_smf_add_message(text); return (result); } if (ns_smf_got_instance == 1 && ns_smf_chroot == 0) ns_smf_want_disable = 1; #endif ns_server_flushonshutdown(ns_g_server, ISC_TRUE); ns_os_shutdownmsg(command, text); isc_app_shutdown(); result = ISC_R_SUCCESS; } else if (command_compare(command, NS_COMMAND_DUMPSTATS)) { result = ns_server_dumpstats(ns_g_server); } else if (command_compare(command, NS_COMMAND_QUERYLOG)) { result = ns_server_togglequerylog(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_DUMPDB)) { ns_server_dumpdb(ns_g_server, command); result = ISC_R_SUCCESS; } else if (command_compare(command, NS_COMMAND_SECROOTS)) { result = ns_server_dumpsecroots(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_TRACE)) { result = ns_server_setdebuglevel(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_NOTRACE)) { ns_g_debuglevel = 0; isc_log_setdebuglevel(ns_g_lctx, ns_g_debuglevel); result = ISC_R_SUCCESS; } else if (command_compare(command, NS_COMMAND_FLUSH)) { result = ns_server_flushcache(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_FLUSHNAME)) { result = ns_server_flushnode(ns_g_server, command, ISC_FALSE); } else if (command_compare(command, NS_COMMAND_FLUSHTREE)) { result = ns_server_flushnode(ns_g_server, command, ISC_TRUE); } else if (command_compare(command, NS_COMMAND_STATUS)) { result = ns_server_status(ns_g_server, text); } else if (command_compare(command, NS_COMMAND_TSIGLIST)) { result = ns_server_tsiglist(ns_g_server, text); } else if (command_compare(command, NS_COMMAND_TSIGDELETE)) { result = ns_server_tsigdelete(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_FREEZE)) { result = ns_server_freeze(ns_g_server, ISC_TRUE, command, text); } else if (command_compare(command, NS_COMMAND_UNFREEZE) || command_compare(command, NS_COMMAND_THAW)) { result = ns_server_freeze(ns_g_server, ISC_FALSE, command, text); } else if (command_compare(command, NS_COMMAND_SYNC)) { result = ns_server_sync(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_RECURSING)) { result = ns_server_dumprecursing(ns_g_server); } else if (command_compare(command, NS_COMMAND_TIMERPOKE)) { result = ISC_R_SUCCESS; isc_timermgr_poke(ns_g_timermgr); } else if (command_compare(command, NS_COMMAND_NULL)) { result = ISC_R_SUCCESS; } else if (command_compare(command, NS_COMMAND_NOTIFY)) { result = ns_server_notifycommand(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_VALIDATION)) { result = ns_server_validation(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_SIGN) || command_compare(command, NS_COMMAND_LOADKEYS)) { result = ns_server_rekey(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_ADDZONE)) { result = ns_server_add_zone(ns_g_server, command); } else if (command_compare(command, NS_COMMAND_DELZONE)) { result = ns_server_del_zone(ns_g_server, command, text); } else if (command_compare(command, NS_COMMAND_SIGNING)) { result = ns_server_signing(ns_g_server, command, text); } else { isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_WARNING, "unknown control channel command '%s'", command); result = DNS_R_UNKNOWNCOMMAND; } return (result); } Index: releng/9.3/contrib/bind9/bin/named/controlconf.c =================================================================== --- releng/9.3/contrib/bind9/bin/named/controlconf.c (revision 296610) +++ releng/9.3/contrib/bind9/bin/named/controlconf.c (revision 296611) @@ -1,1458 +1,1458 @@ /* - * Copyright (C) 2004-2008, 2011-2014 Internet Systems Consortium, Inc. ("ISC") + * Copyright (C) 2004-2008, 2011-2014, 2016 Internet Systems Consortium, Inc. ("ISC") * Copyright (C) 2001-2003 Internet Software Consortium. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH * REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY * AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT, * INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR * PERFORMANCE OF THIS SOFTWARE. */ /* $Id: controlconf.c,v 1.63 2011/12/22 08:07:48 marka Exp $ */ /*! \file */ #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 /* * Note: Listeners and connections are not locked. All event handlers are * executed by the server task, and all callers of exported routines must * be running under the server task. */ typedef struct controlkey controlkey_t; typedef ISC_LIST(controlkey_t) controlkeylist_t; typedef struct controlconnection controlconnection_t; typedef ISC_LIST(controlconnection_t) controlconnectionlist_t; typedef struct controllistener controllistener_t; typedef ISC_LIST(controllistener_t) controllistenerlist_t; struct controlkey { char * keyname; isc_region_t secret; ISC_LINK(controlkey_t) link; }; struct controlconnection { isc_socket_t * sock; isccc_ccmsg_t ccmsg; isc_boolean_t ccmsg_valid; isc_boolean_t sending; isc_timer_t * timer; unsigned char buffer[2048]; controllistener_t * listener; isc_uint32_t nonce; ISC_LINK(controlconnection_t) link; }; struct controllistener { ns_controls_t * controls; isc_mem_t * mctx; isc_task_t * task; isc_sockaddr_t address; isc_socket_t * sock; dns_acl_t * acl; isc_boolean_t listening; isc_boolean_t exiting; controlkeylist_t keys; controlconnectionlist_t connections; isc_sockettype_t type; isc_uint32_t perm; isc_uint32_t owner; isc_uint32_t group; ISC_LINK(controllistener_t) link; }; struct ns_controls { ns_server_t *server; controllistenerlist_t listeners; isc_boolean_t shuttingdown; isccc_symtab_t *symtab; }; static void control_newconn(isc_task_t *task, isc_event_t *event); static void control_recvmessage(isc_task_t *task, isc_event_t *event); #define CLOCKSKEW 300 static void free_controlkey(controlkey_t *key, isc_mem_t *mctx) { if (key->keyname != NULL) isc_mem_free(mctx, key->keyname); if (key->secret.base != NULL) isc_mem_put(mctx, key->secret.base, key->secret.length); isc_mem_put(mctx, key, sizeof(*key)); } static void free_controlkeylist(controlkeylist_t *keylist, isc_mem_t *mctx) { while (!ISC_LIST_EMPTY(*keylist)) { controlkey_t *key = ISC_LIST_HEAD(*keylist); ISC_LIST_UNLINK(*keylist, key, link); free_controlkey(key, mctx); } } static void free_listener(controllistener_t *listener) { INSIST(listener->exiting); INSIST(!listener->listening); INSIST(ISC_LIST_EMPTY(listener->connections)); if (listener->sock != NULL) isc_socket_detach(&listener->sock); free_controlkeylist(&listener->keys, listener->mctx); if (listener->acl != NULL) dns_acl_detach(&listener->acl); isc_mem_putanddetach(&listener->mctx, listener, sizeof(*listener)); } static void maybe_free_listener(controllistener_t *listener) { if (listener->exiting && !listener->listening && ISC_LIST_EMPTY(listener->connections)) free_listener(listener); } static void maybe_free_connection(controlconnection_t *conn) { controllistener_t *listener = conn->listener; if (conn->timer != NULL) isc_timer_detach(&conn->timer); if (conn->ccmsg_valid) { isccc_ccmsg_cancelread(&conn->ccmsg); return; } if (conn->sending) { isc_socket_cancel(conn->sock, listener->task, ISC_SOCKCANCEL_SEND); return; } ISC_LIST_UNLINK(listener->connections, conn, link); isc_mem_put(listener->mctx, conn, sizeof(*conn)); } static void shutdown_listener(controllistener_t *listener) { controlconnection_t *conn; controlconnection_t *next; if (!listener->exiting) { char socktext[ISC_SOCKADDR_FORMATSIZE]; ISC_LIST_UNLINK(listener->controls->listeners, listener, link); isc_sockaddr_format(&listener->address, socktext, sizeof(socktext)); isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_NOTICE, "stopping command channel on %s", socktext); if (listener->type == isc_sockettype_unix) isc_socket_cleanunix(&listener->address, ISC_TRUE); listener->exiting = ISC_TRUE; } for (conn = ISC_LIST_HEAD(listener->connections); conn != NULL; conn = next) { next = ISC_LIST_NEXT(conn, link); maybe_free_connection(conn); } if (listener->listening) isc_socket_cancel(listener->sock, listener->task, ISC_SOCKCANCEL_ACCEPT); maybe_free_listener(listener); } static isc_boolean_t address_ok(isc_sockaddr_t *sockaddr, dns_acl_t *acl) { isc_netaddr_t netaddr; isc_result_t result; int match; isc_netaddr_fromsockaddr(&netaddr, sockaddr); result = dns_acl_match(&netaddr, NULL, acl, &ns_g_server->aclenv, &match, NULL); if (result != ISC_R_SUCCESS || match <= 0) return (ISC_FALSE); else return (ISC_TRUE); } static isc_result_t control_accept(controllistener_t *listener) { isc_result_t result; result = isc_socket_accept(listener->sock, listener->task, control_newconn, listener); if (result != ISC_R_SUCCESS) UNEXPECTED_ERROR(__FILE__, __LINE__, "isc_socket_accept() failed: %s", isc_result_totext(result)); else listener->listening = ISC_TRUE; return (result); } static isc_result_t control_listen(controllistener_t *listener) { isc_result_t result; result = isc_socket_listen(listener->sock, 0); if (result != ISC_R_SUCCESS) UNEXPECTED_ERROR(__FILE__, __LINE__, "isc_socket_listen() failed: %s", isc_result_totext(result)); return (result); } static void control_next(controllistener_t *listener) { (void)control_accept(listener); } static void control_senddone(isc_task_t *task, isc_event_t *event) { isc_socketevent_t *sevent = (isc_socketevent_t *) event; controlconnection_t *conn = event->ev_arg; controllistener_t *listener = conn->listener; isc_socket_t *sock = (isc_socket_t *)sevent->ev_sender; isc_result_t result; REQUIRE(conn->sending); UNUSED(task); conn->sending = ISC_FALSE; if (sevent->result != ISC_R_SUCCESS && sevent->result != ISC_R_CANCELED) { char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_t peeraddr; (void)isc_socket_getpeername(sock, &peeraddr); isc_sockaddr_format(&peeraddr, socktext, sizeof(socktext)); isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_WARNING, "error sending command response to %s: %s", socktext, isc_result_totext(sevent->result)); } isc_event_free(&event); result = isccc_ccmsg_readmessage(&conn->ccmsg, listener->task, control_recvmessage, conn); if (result != ISC_R_SUCCESS) { isc_socket_detach(&conn->sock); maybe_free_connection(conn); maybe_free_listener(listener); } } static inline void log_invalid(isccc_ccmsg_t *ccmsg, isc_result_t result) { char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_t peeraddr; (void)isc_socket_getpeername(ccmsg->sock, &peeraddr); isc_sockaddr_format(&peeraddr, socktext, sizeof(socktext)); isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_ERROR, "invalid command from %s: %s", socktext, isc_result_totext(result)); } static void control_recvmessage(isc_task_t *task, isc_event_t *event) { controlconnection_t *conn; controllistener_t *listener; controlkey_t *key; isccc_sexpr_t *request = NULL; isccc_sexpr_t *response = NULL; isccc_region_t ccregion; isccc_region_t secret; isc_stdtime_t now; isc_buffer_t b; isc_region_t r; isc_uint32_t len; isc_buffer_t text; char textarray[1024]; isc_result_t result; isc_result_t eresult; isccc_sexpr_t *_ctrl; isccc_time_t sent; isccc_time_t exp; isc_uint32_t nonce; REQUIRE(event->ev_type == ISCCC_EVENT_CCMSG); conn = event->ev_arg; listener = conn->listener; secret.rstart = NULL; /* Is the server shutting down? */ if (listener->controls->shuttingdown) goto cleanup; if (conn->ccmsg.result != ISC_R_SUCCESS) { if (conn->ccmsg.result != ISC_R_CANCELED && conn->ccmsg.result != ISC_R_EOF) log_invalid(&conn->ccmsg, conn->ccmsg.result); goto cleanup; } request = NULL; for (key = ISC_LIST_HEAD(listener->keys); key != NULL; key = ISC_LIST_NEXT(key, link)) { ccregion.rstart = isc_buffer_base(&conn->ccmsg.buffer); ccregion.rend = isc_buffer_used(&conn->ccmsg.buffer); secret.rstart = isc_mem_get(listener->mctx, key->secret.length); if (secret.rstart == NULL) goto cleanup; memmove(secret.rstart, key->secret.base, key->secret.length); secret.rend = secret.rstart + key->secret.length; result = isccc_cc_fromwire(&ccregion, &request, &secret); if (result == ISC_R_SUCCESS) break; isc_mem_put(listener->mctx, secret.rstart, REGION_SIZE(secret)); if (result != ISCCC_R_BADAUTH) { log_invalid(&conn->ccmsg, result); goto cleanup; } } if (key == NULL) { log_invalid(&conn->ccmsg, ISCCC_R_BADAUTH); goto cleanup; } /* We shouldn't be getting a reply. */ if (isccc_cc_isreply(request)) { log_invalid(&conn->ccmsg, ISC_R_FAILURE); goto cleanup_request; } isc_stdtime_get(&now); /* * Limit exposure to replay attacks. */ _ctrl = isccc_alist_lookup(request, "_ctrl"); - if (_ctrl == NULL) { + if (!isccc_alist_alistp(_ctrl)) { log_invalid(&conn->ccmsg, ISC_R_FAILURE); goto cleanup_request; } if (isccc_cc_lookupuint32(_ctrl, "_tim", &sent) == ISC_R_SUCCESS) { if ((sent + CLOCKSKEW) < now || (sent - CLOCKSKEW) > now) { log_invalid(&conn->ccmsg, ISCCC_R_CLOCKSKEW); goto cleanup_request; } } else { log_invalid(&conn->ccmsg, ISC_R_FAILURE); goto cleanup_request; } /* * Expire messages that are too old. */ if (isccc_cc_lookupuint32(_ctrl, "_exp", &exp) == ISC_R_SUCCESS && now > exp) { log_invalid(&conn->ccmsg, ISCCC_R_EXPIRED); goto cleanup_request; } /* * Duplicate suppression (required for UDP). */ isccc_cc_cleansymtab(listener->controls->symtab, now); result = isccc_cc_checkdup(listener->controls->symtab, request, now); if (result != ISC_R_SUCCESS) { if (result == ISC_R_EXISTS) result = ISCCC_R_DUPLICATE; log_invalid(&conn->ccmsg, result); goto cleanup_request; } if (conn->nonce != 0 && (isccc_cc_lookupuint32(_ctrl, "_nonce", &nonce) != ISC_R_SUCCESS || conn->nonce != nonce)) { log_invalid(&conn->ccmsg, ISCCC_R_BADAUTH); goto cleanup_request; } /* * Establish nonce. */ while (conn->nonce == 0) isc_random_get(&conn->nonce); isc_buffer_init(&text, textarray, sizeof(textarray)); eresult = ns_control_docommand(request, &text); result = isccc_cc_createresponse(request, now, now + 60, &response); if (result != ISC_R_SUCCESS) goto cleanup_request; if (eresult != ISC_R_SUCCESS) { isccc_sexpr_t *data; data = isccc_alist_lookup(response, "_data"); if (data != NULL) { const char *estr = isc_result_totext(eresult); if (isccc_cc_definestring(data, "err", estr) == NULL) goto cleanup_response; } } if (isc_buffer_usedlength(&text) > 0) { isccc_sexpr_t *data; data = isccc_alist_lookup(response, "_data"); if (data != NULL) { char *str = (char *)isc_buffer_base(&text); if (isccc_cc_definestring(data, "text", str) == NULL) goto cleanup_response; } } _ctrl = isccc_alist_lookup(response, "_ctrl"); if (_ctrl == NULL || isccc_cc_defineuint32(_ctrl, "_nonce", conn->nonce) == NULL) goto cleanup_response; ccregion.rstart = conn->buffer + 4; ccregion.rend = conn->buffer + sizeof(conn->buffer); result = isccc_cc_towire(response, &ccregion, &secret); if (result != ISC_R_SUCCESS) goto cleanup_response; isc_buffer_init(&b, conn->buffer, 4); len = sizeof(conn->buffer) - REGION_SIZE(ccregion); isc_buffer_putuint32(&b, len - 4); r.base = conn->buffer; r.length = len; result = isc_socket_send(conn->sock, &r, task, control_senddone, conn); if (result != ISC_R_SUCCESS) goto cleanup_response; conn->sending = ISC_TRUE; isc_mem_put(listener->mctx, secret.rstart, REGION_SIZE(secret)); isccc_sexpr_free(&request); isccc_sexpr_free(&response); return; cleanup_response: isccc_sexpr_free(&response); cleanup_request: isccc_sexpr_free(&request); isc_mem_put(listener->mctx, secret.rstart, REGION_SIZE(secret)); cleanup: isc_socket_detach(&conn->sock); isccc_ccmsg_invalidate(&conn->ccmsg); conn->ccmsg_valid = ISC_FALSE; maybe_free_connection(conn); maybe_free_listener(listener); } static void control_timeout(isc_task_t *task, isc_event_t *event) { controlconnection_t *conn = event->ev_arg; UNUSED(task); isc_timer_detach(&conn->timer); maybe_free_connection(conn); isc_event_free(&event); } static isc_result_t newconnection(controllistener_t *listener, isc_socket_t *sock) { controlconnection_t *conn; isc_interval_t interval; isc_result_t result; conn = isc_mem_get(listener->mctx, sizeof(*conn)); if (conn == NULL) return (ISC_R_NOMEMORY); conn->sock = sock; isccc_ccmsg_init(listener->mctx, sock, &conn->ccmsg); conn->ccmsg_valid = ISC_TRUE; conn->sending = ISC_FALSE; conn->timer = NULL; isc_interval_set(&interval, 60, 0); result = isc_timer_create(ns_g_timermgr, isc_timertype_once, NULL, &interval, listener->task, control_timeout, conn, &conn->timer); if (result != ISC_R_SUCCESS) goto cleanup; conn->listener = listener; conn->nonce = 0; ISC_LINK_INIT(conn, link); result = isccc_ccmsg_readmessage(&conn->ccmsg, listener->task, control_recvmessage, conn); if (result != ISC_R_SUCCESS) goto cleanup; isccc_ccmsg_setmaxsize(&conn->ccmsg, 2048); ISC_LIST_APPEND(listener->connections, conn, link); return (ISC_R_SUCCESS); cleanup: isccc_ccmsg_invalidate(&conn->ccmsg); if (conn->timer != NULL) isc_timer_detach(&conn->timer); isc_mem_put(listener->mctx, conn, sizeof(*conn)); return (result); } static void control_newconn(isc_task_t *task, isc_event_t *event) { isc_socket_newconnev_t *nevent = (isc_socket_newconnev_t *)event; controllistener_t *listener = event->ev_arg; isc_socket_t *sock; isc_sockaddr_t peeraddr; isc_result_t result; UNUSED(task); listener->listening = ISC_FALSE; if (nevent->result != ISC_R_SUCCESS) { if (nevent->result == ISC_R_CANCELED) { shutdown_listener(listener); goto cleanup; } goto restart; } sock = nevent->newsocket; isc_socket_setname(sock, "control", NULL); (void)isc_socket_getpeername(sock, &peeraddr); if (listener->type == isc_sockettype_tcp && !address_ok(&peeraddr, listener->acl)) { char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_format(&peeraddr, socktext, sizeof(socktext)); isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_WARNING, "rejected command channel message from %s", socktext); isc_socket_detach(&sock); goto restart; } result = newconnection(listener, sock); if (result != ISC_R_SUCCESS) { char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_format(&peeraddr, socktext, sizeof(socktext)); isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_WARNING, "dropped command channel from %s: %s", socktext, isc_result_totext(result)); isc_socket_detach(&sock); goto restart; } restart: control_next(listener); cleanup: isc_event_free(&event); } static void controls_shutdown(ns_controls_t *controls) { controllistener_t *listener; controllistener_t *next; for (listener = ISC_LIST_HEAD(controls->listeners); listener != NULL; listener = next) { /* * This is asynchronous. As listeners shut down, they will * call their callbacks. */ next = ISC_LIST_NEXT(listener, link); shutdown_listener(listener); } } void ns_controls_shutdown(ns_controls_t *controls) { controls_shutdown(controls); controls->shuttingdown = ISC_TRUE; } static isc_result_t cfgkeylist_find(const cfg_obj_t *keylist, const char *keyname, const cfg_obj_t **objp) { const cfg_listelt_t *element; const char *str; const cfg_obj_t *obj; for (element = cfg_list_first(keylist); element != NULL; element = cfg_list_next(element)) { obj = cfg_listelt_value(element); str = cfg_obj_asstring(cfg_map_getname(obj)); if (strcasecmp(str, keyname) == 0) break; } if (element == NULL) return (ISC_R_NOTFOUND); obj = cfg_listelt_value(element); *objp = obj; return (ISC_R_SUCCESS); } static isc_result_t controlkeylist_fromcfg(const cfg_obj_t *keylist, isc_mem_t *mctx, controlkeylist_t *keyids) { const cfg_listelt_t *element; char *newstr = NULL; const char *str; const cfg_obj_t *obj; controlkey_t *key; for (element = cfg_list_first(keylist); element != NULL; element = cfg_list_next(element)) { obj = cfg_listelt_value(element); str = cfg_obj_asstring(obj); newstr = isc_mem_strdup(mctx, str); if (newstr == NULL) goto cleanup; key = isc_mem_get(mctx, sizeof(*key)); if (key == NULL) goto cleanup; key->keyname = newstr; key->secret.base = NULL; key->secret.length = 0; ISC_LINK_INIT(key, link); ISC_LIST_APPEND(*keyids, key, link); newstr = NULL; } return (ISC_R_SUCCESS); cleanup: if (newstr != NULL) isc_mem_free(mctx, newstr); free_controlkeylist(keyids, mctx); return (ISC_R_NOMEMORY); } static void register_keys(const cfg_obj_t *control, const cfg_obj_t *keylist, controlkeylist_t *keyids, isc_mem_t *mctx, const char *socktext) { controlkey_t *keyid, *next; const cfg_obj_t *keydef; char secret[1024]; isc_buffer_t b; isc_result_t result; /* * Find the keys corresponding to the keyids used by this listener. */ for (keyid = ISC_LIST_HEAD(*keyids); keyid != NULL; keyid = next) { next = ISC_LIST_NEXT(keyid, link); result = cfgkeylist_find(keylist, keyid->keyname, &keydef); if (result != ISC_R_SUCCESS) { cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "couldn't find key '%s' for use with " "command channel %s", keyid->keyname, socktext); ISC_LIST_UNLINK(*keyids, keyid, link); free_controlkey(keyid, mctx); } else { const cfg_obj_t *algobj = NULL; const cfg_obj_t *secretobj = NULL; const char *algstr = NULL; const char *secretstr = NULL; (void)cfg_map_get(keydef, "algorithm", &algobj); (void)cfg_map_get(keydef, "secret", &secretobj); INSIST(algobj != NULL && secretobj != NULL); algstr = cfg_obj_asstring(algobj); secretstr = cfg_obj_asstring(secretobj); if (ns_config_getkeyalgorithm(algstr, NULL, NULL) != ISC_R_SUCCESS) { cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "unsupported algorithm '%s' in " "key '%s' for use with command " "channel %s", algstr, keyid->keyname, socktext); ISC_LIST_UNLINK(*keyids, keyid, link); free_controlkey(keyid, mctx); continue; } isc_buffer_init(&b, secret, sizeof(secret)); result = isc_base64_decodestring(secretstr, &b); if (result != ISC_R_SUCCESS) { cfg_obj_log(keydef, ns_g_lctx, ISC_LOG_WARNING, "secret for key '%s' on " "command channel %s: %s", keyid->keyname, socktext, isc_result_totext(result)); ISC_LIST_UNLINK(*keyids, keyid, link); free_controlkey(keyid, mctx); continue; } keyid->secret.length = isc_buffer_usedlength(&b); keyid->secret.base = isc_mem_get(mctx, keyid->secret.length); if (keyid->secret.base == NULL) { cfg_obj_log(keydef, ns_g_lctx, ISC_LOG_WARNING, "couldn't register key '%s': " "out of memory", keyid->keyname); ISC_LIST_UNLINK(*keyids, keyid, link); free_controlkey(keyid, mctx); break; } memmove(keyid->secret.base, isc_buffer_base(&b), keyid->secret.length); } } } #define CHECK(x) \ do { \ result = (x); \ if (result != ISC_R_SUCCESS) \ goto cleanup; \ } while (0) static isc_result_t get_rndckey(isc_mem_t *mctx, controlkeylist_t *keyids) { isc_result_t result; cfg_parser_t *pctx = NULL; cfg_obj_t *config = NULL; const cfg_obj_t *key = NULL; const cfg_obj_t *algobj = NULL; const cfg_obj_t *secretobj = NULL; const char *algstr = NULL; const char *secretstr = NULL; controlkey_t *keyid = NULL; char secret[1024]; isc_buffer_t b; CHECK(cfg_parser_create(mctx, ns_g_lctx, &pctx)); CHECK(cfg_parse_file(pctx, ns_g_keyfile, &cfg_type_rndckey, &config)); CHECK(cfg_map_get(config, "key", &key)); keyid = isc_mem_get(mctx, sizeof(*keyid)); if (keyid == NULL) CHECK(ISC_R_NOMEMORY); keyid->keyname = isc_mem_strdup(mctx, cfg_obj_asstring(cfg_map_getname(key))); keyid->secret.base = NULL; keyid->secret.length = 0; ISC_LINK_INIT(keyid, link); if (keyid->keyname == NULL) CHECK(ISC_R_NOMEMORY); CHECK(bind9_check_key(key, ns_g_lctx)); (void)cfg_map_get(key, "algorithm", &algobj); (void)cfg_map_get(key, "secret", &secretobj); INSIST(algobj != NULL && secretobj != NULL); algstr = cfg_obj_asstring(algobj); secretstr = cfg_obj_asstring(secretobj); if (ns_config_getkeyalgorithm(algstr, NULL, NULL) != ISC_R_SUCCESS) { cfg_obj_log(key, ns_g_lctx, ISC_LOG_WARNING, "unsupported algorithm '%s' in " "key '%s' for use with command " "channel", algstr, keyid->keyname); goto cleanup; } isc_buffer_init(&b, secret, sizeof(secret)); result = isc_base64_decodestring(secretstr, &b); if (result != ISC_R_SUCCESS) { cfg_obj_log(key, ns_g_lctx, ISC_LOG_WARNING, "secret for key '%s' on command channel: %s", keyid->keyname, isc_result_totext(result)); goto cleanup; } keyid->secret.length = isc_buffer_usedlength(&b); keyid->secret.base = isc_mem_get(mctx, keyid->secret.length); if (keyid->secret.base == NULL) { cfg_obj_log(key, ns_g_lctx, ISC_LOG_WARNING, "couldn't register key '%s': " "out of memory", keyid->keyname); CHECK(ISC_R_NOMEMORY); } memmove(keyid->secret.base, isc_buffer_base(&b), keyid->secret.length); ISC_LIST_APPEND(*keyids, keyid, link); keyid = NULL; result = ISC_R_SUCCESS; cleanup: if (keyid != NULL) free_controlkey(keyid, mctx); if (config != NULL) cfg_obj_destroy(pctx, &config); if (pctx != NULL) cfg_parser_destroy(&pctx); return (result); } /* * Ensures that both '*global_keylistp' and '*control_keylistp' are * valid or both are NULL. */ static void get_key_info(const cfg_obj_t *config, const cfg_obj_t *control, const cfg_obj_t **global_keylistp, const cfg_obj_t **control_keylistp) { isc_result_t result; const cfg_obj_t *control_keylist = NULL; const cfg_obj_t *global_keylist = NULL; REQUIRE(global_keylistp != NULL && *global_keylistp == NULL); REQUIRE(control_keylistp != NULL && *control_keylistp == NULL); control_keylist = cfg_tuple_get(control, "keys"); if (!cfg_obj_isvoid(control_keylist) && cfg_list_first(control_keylist) != NULL) { result = cfg_map_get(config, "key", &global_keylist); if (result == ISC_R_SUCCESS) { *global_keylistp = global_keylist; *control_keylistp = control_keylist; } } } static void update_listener(ns_controls_t *cp, controllistener_t **listenerp, const cfg_obj_t *control, const cfg_obj_t *config, isc_sockaddr_t *addr, cfg_aclconfctx_t *aclconfctx, const char *socktext, isc_sockettype_t type) { controllistener_t *listener; const cfg_obj_t *allow; const cfg_obj_t *global_keylist = NULL; const cfg_obj_t *control_keylist = NULL; dns_acl_t *new_acl = NULL; controlkeylist_t keys; isc_result_t result = ISC_R_SUCCESS; for (listener = ISC_LIST_HEAD(cp->listeners); listener != NULL; listener = ISC_LIST_NEXT(listener, link)) if (isc_sockaddr_equal(addr, &listener->address)) break; if (listener == NULL) { *listenerp = NULL; return; } /* * There is already a listener for this sockaddr. * Update the access list and key information. * * First try to deal with the key situation. There are a few * possibilities: * (a) It had an explicit keylist and still has an explicit keylist. * (b) It had an automagic key and now has an explicit keylist. * (c) It had an explicit keylist and now needs an automagic key. * (d) It has an automagic key and still needs the automagic key. * * (c) and (d) are the annoying ones. The caller needs to know * that it should use the automagic configuration for key information * in place of the named.conf configuration. * * XXXDCL There is one other hazard that has not been dealt with, * the problem that if a key change is being caused by a control * channel reload, then the response will be with the new key * and not able to be decrypted by the client. */ if (control != NULL) get_key_info(config, control, &global_keylist, &control_keylist); if (control_keylist != NULL) { INSIST(global_keylist != NULL); ISC_LIST_INIT(keys); result = controlkeylist_fromcfg(control_keylist, listener->mctx, &keys); if (result == ISC_R_SUCCESS) { free_controlkeylist(&listener->keys, listener->mctx); listener->keys = keys; register_keys(control, global_keylist, &listener->keys, listener->mctx, socktext); } } else { free_controlkeylist(&listener->keys, listener->mctx); result = get_rndckey(listener->mctx, &listener->keys); } if (result != ISC_R_SUCCESS && global_keylist != NULL) { /* * This message might be a little misleading since the * "new keys" might in fact be identical to the old ones, * but tracking whether they are identical just for the * sake of avoiding this message would be too much trouble. */ if (control != NULL) cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "couldn't install new keys for " "command channel %s: %s", socktext, isc_result_totext(result)); else isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_WARNING, "couldn't install new keys for " "command channel %s: %s", socktext, isc_result_totext(result)); } /* * Now, keep the old access list unless a new one can be made. */ if (control != NULL && type == isc_sockettype_tcp) { allow = cfg_tuple_get(control, "allow"); result = cfg_acl_fromconfig(allow, config, ns_g_lctx, aclconfctx, listener->mctx, 0, &new_acl); } else { result = dns_acl_any(listener->mctx, &new_acl); } if (result == ISC_R_SUCCESS) { dns_acl_detach(&listener->acl); dns_acl_attach(new_acl, &listener->acl); dns_acl_detach(&new_acl); /* XXXDCL say the old acl is still used? */ } else if (control != NULL) cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "couldn't install new acl for " "command channel %s: %s", socktext, isc_result_totext(result)); else isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_WARNING, "couldn't install new acl for " "command channel %s: %s", socktext, isc_result_totext(result)); if (result == ISC_R_SUCCESS && type == isc_sockettype_unix) { isc_uint32_t perm, owner, group; perm = cfg_obj_asuint32(cfg_tuple_get(control, "perm")); owner = cfg_obj_asuint32(cfg_tuple_get(control, "owner")); group = cfg_obj_asuint32(cfg_tuple_get(control, "group")); result = ISC_R_SUCCESS; if (listener->perm != perm || listener->owner != owner || listener->group != group) result = isc_socket_permunix(&listener->address, perm, owner, group); if (result == ISC_R_SUCCESS) { listener->perm = perm; listener->owner = owner; listener->group = group; } else if (control != NULL) cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "couldn't update ownership/permission for " "command channel %s", socktext); } *listenerp = listener; } static void add_listener(ns_controls_t *cp, controllistener_t **listenerp, const cfg_obj_t *control, const cfg_obj_t *config, isc_sockaddr_t *addr, cfg_aclconfctx_t *aclconfctx, const char *socktext, isc_sockettype_t type) { isc_mem_t *mctx = cp->server->mctx; controllistener_t *listener; const cfg_obj_t *allow; const cfg_obj_t *global_keylist = NULL; const cfg_obj_t *control_keylist = NULL; dns_acl_t *new_acl = NULL; isc_result_t result = ISC_R_SUCCESS; listener = isc_mem_get(mctx, sizeof(*listener)); if (listener == NULL) result = ISC_R_NOMEMORY; if (result == ISC_R_SUCCESS) { listener->mctx = NULL; isc_mem_attach(mctx, &listener->mctx); listener->controls = cp; listener->task = cp->server->task; listener->address = *addr; listener->sock = NULL; listener->listening = ISC_FALSE; listener->exiting = ISC_FALSE; listener->acl = NULL; listener->type = type; listener->perm = 0; listener->owner = 0; listener->group = 0; ISC_LINK_INIT(listener, link); ISC_LIST_INIT(listener->keys); ISC_LIST_INIT(listener->connections); /* * Make the acl. */ if (control != NULL && type == isc_sockettype_tcp) { allow = cfg_tuple_get(control, "allow"); result = cfg_acl_fromconfig(allow, config, ns_g_lctx, aclconfctx, mctx, 0, &new_acl); } else { result = dns_acl_any(mctx, &new_acl); } } if (result == ISC_R_SUCCESS) { dns_acl_attach(new_acl, &listener->acl); dns_acl_detach(&new_acl); if (config != NULL) get_key_info(config, control, &global_keylist, &control_keylist); if (control_keylist != NULL) { result = controlkeylist_fromcfg(control_keylist, listener->mctx, &listener->keys); if (result == ISC_R_SUCCESS) register_keys(control, global_keylist, &listener->keys, listener->mctx, socktext); } else result = get_rndckey(mctx, &listener->keys); if (result != ISC_R_SUCCESS && control != NULL) cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "couldn't install keys for " "command channel %s: %s", socktext, isc_result_totext(result)); } if (result == ISC_R_SUCCESS) { int pf = isc_sockaddr_pf(&listener->address); if ((pf == AF_INET && isc_net_probeipv4() != ISC_R_SUCCESS) || #ifdef ISC_PLATFORM_HAVESYSUNH (pf == AF_UNIX && isc_net_probeunix() != ISC_R_SUCCESS) || #endif (pf == AF_INET6 && isc_net_probeipv6() != ISC_R_SUCCESS)) result = ISC_R_FAMILYNOSUPPORT; } if (result == ISC_R_SUCCESS && type == isc_sockettype_unix) isc_socket_cleanunix(&listener->address, ISC_FALSE); if (result == ISC_R_SUCCESS) result = isc_socket_create(ns_g_socketmgr, isc_sockaddr_pf(&listener->address), type, &listener->sock); if (result == ISC_R_SUCCESS) isc_socket_setname(listener->sock, "control", NULL); #ifndef ISC_ALLOW_MAPPED if (result == ISC_R_SUCCESS) isc_socket_ipv6only(listener->sock, ISC_TRUE); #endif if (result == ISC_R_SUCCESS) result = isc_socket_bind(listener->sock, &listener->address, ISC_SOCKET_REUSEADDRESS); if (result == ISC_R_SUCCESS && type == isc_sockettype_unix) { listener->perm = cfg_obj_asuint32(cfg_tuple_get(control, "perm")); listener->owner = cfg_obj_asuint32(cfg_tuple_get(control, "owner")); listener->group = cfg_obj_asuint32(cfg_tuple_get(control, "group")); result = isc_socket_permunix(&listener->address, listener->perm, listener->owner, listener->group); } if (result == ISC_R_SUCCESS) result = control_listen(listener); if (result == ISC_R_SUCCESS) result = control_accept(listener); if (result == ISC_R_SUCCESS) { isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_NOTICE, "command channel listening on %s", socktext); *listenerp = listener; } else { if (listener != NULL) { listener->exiting = ISC_TRUE; free_listener(listener); } if (control != NULL) cfg_obj_log(control, ns_g_lctx, ISC_LOG_WARNING, "couldn't add command channel %s: %s", socktext, isc_result_totext(result)); else isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_NOTICE, "couldn't add command channel %s: %s", socktext, isc_result_totext(result)); *listenerp = NULL; } /* XXXDCL return error results? fail hard? */ } isc_result_t ns_controls_configure(ns_controls_t *cp, const cfg_obj_t *config, cfg_aclconfctx_t *aclconfctx) { controllistener_t *listener; controllistenerlist_t new_listeners; const cfg_obj_t *controlslist = NULL; const cfg_listelt_t *element, *element2; char socktext[ISC_SOCKADDR_FORMATSIZE]; ISC_LIST_INIT(new_listeners); /* * Get the list of named.conf 'controls' statements. */ (void)cfg_map_get(config, "controls", &controlslist); /* * Run through the new control channel list, noting sockets that * are already being listened on and moving them to the new list. * * Identifying duplicate addr/port combinations is left to either * the underlying config code, or to the bind attempt getting an * address-in-use error. */ if (controlslist != NULL) { for (element = cfg_list_first(controlslist); element != NULL; element = cfg_list_next(element)) { const cfg_obj_t *controls; const cfg_obj_t *inetcontrols = NULL; controls = cfg_listelt_value(element); (void)cfg_map_get(controls, "inet", &inetcontrols); if (inetcontrols == NULL) continue; for (element2 = cfg_list_first(inetcontrols); element2 != NULL; element2 = cfg_list_next(element2)) { const cfg_obj_t *control; const cfg_obj_t *obj; isc_sockaddr_t addr; /* * The parser handles BIND 8 configuration file * syntax, so it allows unix phrases as well * inet phrases with no keys{} clause. */ control = cfg_listelt_value(element2); obj = cfg_tuple_get(control, "address"); addr = *cfg_obj_assockaddr(obj); if (isc_sockaddr_getport(&addr) == 0) isc_sockaddr_setport(&addr, NS_CONTROL_PORT); isc_sockaddr_format(&addr, socktext, sizeof(socktext)); isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_DEBUG(9), "processing control channel %s", socktext); update_listener(cp, &listener, control, config, &addr, aclconfctx, socktext, isc_sockettype_tcp); if (listener != NULL) /* * Remove the listener from the old * list, so it won't be shut down. */ ISC_LIST_UNLINK(cp->listeners, listener, link); else /* * This is a new listener. */ add_listener(cp, &listener, control, config, &addr, aclconfctx, socktext, isc_sockettype_tcp); if (listener != NULL) ISC_LIST_APPEND(new_listeners, listener, link); } } for (element = cfg_list_first(controlslist); element != NULL; element = cfg_list_next(element)) { const cfg_obj_t *controls; const cfg_obj_t *unixcontrols = NULL; controls = cfg_listelt_value(element); (void)cfg_map_get(controls, "unix", &unixcontrols); if (unixcontrols == NULL) continue; for (element2 = cfg_list_first(unixcontrols); element2 != NULL; element2 = cfg_list_next(element2)) { const cfg_obj_t *control; const cfg_obj_t *path; isc_sockaddr_t addr; isc_result_t result; /* * The parser handles BIND 8 configuration file * syntax, so it allows unix phrases as well * inet phrases with no keys{} clause. */ control = cfg_listelt_value(element2); path = cfg_tuple_get(control, "path"); result = isc_sockaddr_frompath(&addr, cfg_obj_asstring(path)); if (result != ISC_R_SUCCESS) { isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_DEBUG(9), "control channel '%s': %s", cfg_obj_asstring(path), isc_result_totext(result)); continue; } isc_log_write(ns_g_lctx, NS_LOGCATEGORY_GENERAL, NS_LOGMODULE_CONTROL, ISC_LOG_DEBUG(9), "processing control channel '%s'", cfg_obj_asstring(path)); update_listener(cp, &listener, control, config, &addr, aclconfctx, cfg_obj_asstring(path), isc_sockettype_unix); if (listener != NULL) /* * Remove the listener from the old * list, so it won't be shut down. */ ISC_LIST_UNLINK(cp->listeners, listener, link); else /* * This is a new listener. */ add_listener(cp, &listener, control, config, &addr, aclconfctx, cfg_obj_asstring(path), isc_sockettype_unix); if (listener != NULL) ISC_LIST_APPEND(new_listeners, listener, link); } } } else { int i; for (i = 0; i < 2; i++) { isc_sockaddr_t addr; if (i == 0) { struct in_addr localhost; if (isc_net_probeipv4() != ISC_R_SUCCESS) continue; localhost.s_addr = htonl(INADDR_LOOPBACK); isc_sockaddr_fromin(&addr, &localhost, 0); } else { if (isc_net_probeipv6() != ISC_R_SUCCESS) continue; isc_sockaddr_fromin6(&addr, &in6addr_loopback, 0); } isc_sockaddr_setport(&addr, NS_CONTROL_PORT); isc_sockaddr_format(&addr, socktext, sizeof(socktext)); update_listener(cp, &listener, NULL, NULL, &addr, NULL, socktext, isc_sockettype_tcp); if (listener != NULL) /* * Remove the listener from the old * list, so it won't be shut down. */ ISC_LIST_UNLINK(cp->listeners, listener, link); else /* * This is a new listener. */ add_listener(cp, &listener, NULL, NULL, &addr, NULL, socktext, isc_sockettype_tcp); if (listener != NULL) ISC_LIST_APPEND(new_listeners, listener, link); } } /* * ns_control_shutdown() will stop whatever is on the global * listeners list, which currently only has whatever sockaddrs * were in the previous configuration (if any) that do not * remain in the current configuration. */ controls_shutdown(cp); /* * Put all of the valid listeners on the listeners list. * Anything already on listeners in the process of shutting * down will be taken care of by listen_done(). */ ISC_LIST_APPENDLIST(cp->listeners, new_listeners, link); return (ISC_R_SUCCESS); } isc_result_t ns_controls_create(ns_server_t *server, ns_controls_t **ctrlsp) { isc_mem_t *mctx = server->mctx; isc_result_t result; ns_controls_t *controls = isc_mem_get(mctx, sizeof(*controls)); if (controls == NULL) return (ISC_R_NOMEMORY); controls->server = server; ISC_LIST_INIT(controls->listeners); controls->shuttingdown = ISC_FALSE; controls->symtab = NULL; result = isccc_cc_createsymtab(&controls->symtab); if (result != ISC_R_SUCCESS) { isc_mem_put(server->mctx, controls, sizeof(*controls)); return (result); } *ctrlsp = controls; return (ISC_R_SUCCESS); } void ns_controls_destroy(ns_controls_t **ctrlsp) { ns_controls_t *controls = *ctrlsp; REQUIRE(ISC_LIST_EMPTY(controls->listeners)); isccc_symtab_destroy(&controls->symtab); isc_mem_put(controls->server->mctx, controls, sizeof(*controls)); *ctrlsp = NULL; } Index: releng/9.3/contrib/bind9/bin/rndc/rndc.c =================================================================== --- releng/9.3/contrib/bind9/bin/rndc/rndc.c (revision 296610) +++ releng/9.3/contrib/bind9/bin/rndc/rndc.c (revision 296611) @@ -1,909 +1,909 @@ /* * Copyright (C) 2004-2014 Internet Systems Consortium, Inc. ("ISC") * Copyright (C) 2000-2003 Internet Software Consortium. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH * REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY * AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT, * INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR * PERFORMANCE OF THIS SOFTWARE. */ /* $Id$ */ /*! \file */ /* * Principal Author: DCL */ #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 "util.h" #define SERVERADDRS 10 const char *progname; isc_boolean_t verbose; static const char *admin_conffile; static const char *admin_keyfile; static const char *version = VERSION; static const char *servername = NULL; static isc_sockaddr_t serveraddrs[SERVERADDRS]; static isc_sockaddr_t local4, local6; static isc_boolean_t local4set = ISC_FALSE, local6set = ISC_FALSE; static int nserveraddrs; static int currentaddr = 0; static unsigned int remoteport = 0; static isc_socketmgr_t *socketmgr = NULL; static unsigned char databuf[2048]; static isccc_ccmsg_t ccmsg; static isccc_region_t secret; static isc_boolean_t failed = ISC_FALSE; static isc_boolean_t c_flag = ISC_FALSE; static isc_mem_t *mctx; static int sends, recvs, connects; static char *command; static char *args; static char program[256]; static isc_socket_t *sock = NULL; static isc_uint32_t serial; static void rndc_startconnect(isc_sockaddr_t *addr, isc_task_t *task); ISC_PLATFORM_NORETURN_PRE static void usage(int status) ISC_PLATFORM_NORETURN_POST; static void usage(int status) { fprintf(stderr, "\ Usage: %s [-b address] [-c config] [-s server] [-p port]\n\ [-k key-file ] [-y key] [-V] command\n\ \n\ command is one of the following:\n\ \n\ reload Reload configuration file and zones.\n\ reload zone [class [view]]\n\ Reload a single zone.\n\ refresh zone [class [view]]\n\ Schedule immediate maintenance for a zone.\n\ retransfer zone [class [view]]\n\ Retransfer a single zone without checking serial number.\n\ freeze Suspend updates to all dynamic zones.\n\ freeze zone [class [view]]\n\ Suspend updates to a dynamic zone.\n\ thaw Enable updates to all dynamic zones and reload them.\n\ thaw zone [class [view]]\n\ Enable updates to a frozen dynamic zone and reload it.\n\ sync [-clean] Dump changes to all dynamic zones to disk, and optionally\n\ remove their journal files.\n\ sync [-clean] zone [class [view]]\n\ Dump a single zone's changes to disk, and optionally\n\ remove its journal file.\n\ notify zone [class [view]]\n\ Resend NOTIFY messages for the zone.\n\ reconfig Reload configuration file and new zones only.\n\ sign zone [class [view]]\n\ Update zone keys, and sign as needed.\n\ loadkeys zone [class [view]]\n\ Update keys without signing immediately.\n\ stats Write server statistics to the statistics file.\n\ querylog newstate\n\ Enable / disable query logging.\n\ dumpdb [-all|-cache|-zones] [view ...]\n\ Dump cache(s) to the dump file (named_dump.db).\n\ secroots [view ...]\n\ Write security roots to the secroots file.\n\ stop Save pending updates to master files and stop the server.\n\ stop -p Save pending updates to master files and stop the server\n\ reporting process id.\n\ halt Stop the server without saving pending updates.\n\ halt -p Stop the server without saving pending updates reporting\n\ process id.\n\ trace Increment debugging level by one.\n\ trace level Change the debugging level.\n\ notrace Set debugging level to 0.\n\ flush Flushes all of the server's caches.\n\ flush [view] Flushes the server's cache for a view.\n\ flushname name [view]\n\ Flush the given name from the server's cache(s)\n\ flushtree name [view]\n\ Flush all names under the given name from the server's cache(s)\n\ status Display status of the server.\n\ recursing Dump the queries that are currently recursing (named.recursing)\n\ tsig-list List all currently active TSIG keys, including both statically\n\ configured and TKEY-negotiated keys.\n\ tsig-delete keyname [view] \n\ Delete a TKEY-negotiated TSIG key.\n\ validation newstate [view]\n\ Enable / disable DNSSEC validation.\n\ addzone [\"file\"] zone [class [view]] { zone-options }\n\ Add zone to given view. Requires new-zone-file option.\n\ delzone [\"file\"] zone [class [view]]\n\ Removes zone from given view. Requires new-zone-file option.\n\ signing -list zone [class [view]]\n\ List the private records showing the state of DNSSEC\n\ signing in the given zone.\n\ signing -clear / zone [class [view]]\n\ Remove the private record that indicating the given key\n\ has finished signing the given zone.\n\ signing -clear all zone [class [view]]\n\ Remove the private records for all keys that have\n\ finished signing the given zone.\n\ signing -nsec3param none zone [class [view]]\n\ Remove NSEC3 chains from zone.\n\ signing -nsec3param hash flags iterations salt zone [class [view]]\n\ Add NSEC3 chain to zone if already signed.\n\ Prime zone with NSEC3 chain if not yet signed.\n\ *restart Restart the server.\n\ \n\ * == not yet implemented\n\ Version: %s\n", progname, version); exit(status); } static void get_addresses(const char *host, in_port_t port) { isc_result_t result; int found = 0, count; if (*host == '/') { result = isc_sockaddr_frompath(&serveraddrs[nserveraddrs], host); if (result == ISC_R_SUCCESS) nserveraddrs++; } else { count = SERVERADDRS - nserveraddrs; result = bind9_getaddresses(host, port, &serveraddrs[nserveraddrs], count, &found); nserveraddrs += found; } if (result != ISC_R_SUCCESS) fatal("couldn't get address for '%s': %s", host, isc_result_totext(result)); INSIST(nserveraddrs > 0); } static void rndc_senddone(isc_task_t *task, isc_event_t *event) { isc_socketevent_t *sevent = (isc_socketevent_t *)event; UNUSED(task); sends--; if (sevent->result != ISC_R_SUCCESS) fatal("send failed: %s", isc_result_totext(sevent->result)); isc_event_free(&event); if (sends == 0 && recvs == 0) { isc_socket_detach(&sock); isc_task_shutdown(task); RUNTIME_CHECK(isc_app_shutdown() == ISC_R_SUCCESS); } } static void rndc_recvdone(isc_task_t *task, isc_event_t *event) { isccc_sexpr_t *response = NULL; isccc_sexpr_t *data; isccc_region_t source; char *errormsg = NULL; char *textmsg = NULL; isc_result_t result; recvs--; if (ccmsg.result == ISC_R_EOF) fatal("connection to remote host closed\n" "This may indicate that\n" "* the remote server is using an older version of" " the command protocol,\n" "* this host is not authorized to connect,\n" "* the clocks are not synchronized, or\n" "* the key is invalid."); if (ccmsg.result != ISC_R_SUCCESS) fatal("recv failed: %s", isc_result_totext(ccmsg.result)); source.rstart = isc_buffer_base(&ccmsg.buffer); source.rend = isc_buffer_used(&ccmsg.buffer); DO("parse message", isccc_cc_fromwire(&source, &response, &secret)); data = isccc_alist_lookup(response, "_data"); - if (data == NULL) - fatal("no data section in response"); + if (!isccc_alist_alistp(data)) + fatal("bad or missing data section in response"); result = isccc_cc_lookupstring(data, "err", &errormsg); if (result == ISC_R_SUCCESS) { failed = ISC_TRUE; fprintf(stderr, "%s: '%s' failed: %s\n", progname, command, errormsg); } else if (result != ISC_R_NOTFOUND) fprintf(stderr, "%s: parsing response failed: %s\n", progname, isc_result_totext(result)); result = isccc_cc_lookupstring(data, "text", &textmsg); if (result == ISC_R_SUCCESS) { if (strlen(textmsg) != 0U) printf("%s\n", textmsg); } else if (result != ISC_R_NOTFOUND) fprintf(stderr, "%s: parsing response failed: %s\n", progname, isc_result_totext(result)); isc_event_free(&event); isccc_sexpr_free(&response); if (sends == 0 && recvs == 0) { isc_socket_detach(&sock); isc_task_shutdown(task); RUNTIME_CHECK(isc_app_shutdown() == ISC_R_SUCCESS); } } static void rndc_recvnonce(isc_task_t *task, isc_event_t *event) { isccc_sexpr_t *response = NULL; isccc_sexpr_t *_ctrl; isccc_region_t source; isc_result_t result; isc_uint32_t nonce; isccc_sexpr_t *request = NULL; isccc_time_t now; isc_region_t r; isccc_sexpr_t *data; isccc_region_t message; isc_uint32_t len; isc_buffer_t b; recvs--; if (ccmsg.result == ISC_R_EOF) fatal("connection to remote host closed\n" "This may indicate that\n" "* the remote server is using an older version of" " the command protocol,\n" "* this host is not authorized to connect,\n" "* the clocks are not synchronized, or\n" "* the key is invalid."); if (ccmsg.result != ISC_R_SUCCESS) fatal("recv failed: %s", isc_result_totext(ccmsg.result)); source.rstart = isc_buffer_base(&ccmsg.buffer); source.rend = isc_buffer_used(&ccmsg.buffer); DO("parse message", isccc_cc_fromwire(&source, &response, &secret)); _ctrl = isccc_alist_lookup(response, "_ctrl"); - if (_ctrl == NULL) - fatal("_ctrl section missing"); + if (!isccc_alist_alistp(_ctrl)) + fatal("bad or missing ctrl section in response"); nonce = 0; if (isccc_cc_lookupuint32(_ctrl, "_nonce", &nonce) != ISC_R_SUCCESS) nonce = 0; isc_stdtime_get(&now); DO("create message", isccc_cc_createmessage(1, NULL, NULL, ++serial, now, now + 60, &request)); data = isccc_alist_lookup(request, "_data"); if (data == NULL) fatal("_data section missing"); if (isccc_cc_definestring(data, "type", args) == NULL) fatal("out of memory"); if (nonce != 0) { _ctrl = isccc_alist_lookup(request, "_ctrl"); if (_ctrl == NULL) fatal("_ctrl section missing"); if (isccc_cc_defineuint32(_ctrl, "_nonce", nonce) == NULL) fatal("out of memory"); } message.rstart = databuf + 4; message.rend = databuf + sizeof(databuf); DO("render message", isccc_cc_towire(request, &message, &secret)); len = sizeof(databuf) - REGION_SIZE(message); isc_buffer_init(&b, databuf, 4); isc_buffer_putuint32(&b, len - 4); r.length = len; r.base = databuf; isccc_ccmsg_cancelread(&ccmsg); DO("schedule recv", isccc_ccmsg_readmessage(&ccmsg, task, rndc_recvdone, NULL)); recvs++; DO("send message", isc_socket_send(sock, &r, task, rndc_senddone, NULL)); sends++; isc_event_free(&event); isccc_sexpr_free(&response); return; } static void rndc_connected(isc_task_t *task, isc_event_t *event) { char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_socketevent_t *sevent = (isc_socketevent_t *)event; isccc_sexpr_t *request = NULL; isccc_sexpr_t *data; isccc_time_t now; isccc_region_t message; isc_region_t r; isc_uint32_t len; isc_buffer_t b; isc_result_t result; connects--; if (sevent->result != ISC_R_SUCCESS) { isc_sockaddr_format(&serveraddrs[currentaddr], socktext, sizeof(socktext)); if (sevent->result != ISC_R_CANCELED && ++currentaddr < nserveraddrs) { notify("connection failed: %s: %s", socktext, isc_result_totext(sevent->result)); isc_socket_detach(&sock); isc_event_free(&event); rndc_startconnect(&serveraddrs[currentaddr], task); return; } else fatal("connect failed: %s: %s", socktext, isc_result_totext(sevent->result)); } isc_stdtime_get(&now); DO("create message", isccc_cc_createmessage(1, NULL, NULL, ++serial, now, now + 60, &request)); data = isccc_alist_lookup(request, "_data"); if (data == NULL) fatal("_data section missing"); if (isccc_cc_definestring(data, "type", "null") == NULL) fatal("out of memory"); message.rstart = databuf + 4; message.rend = databuf + sizeof(databuf); DO("render message", isccc_cc_towire(request, &message, &secret)); len = sizeof(databuf) - REGION_SIZE(message); isc_buffer_init(&b, databuf, 4); isc_buffer_putuint32(&b, len - 4); r.length = len; r.base = databuf; isccc_ccmsg_init(mctx, sock, &ccmsg); isccc_ccmsg_setmaxsize(&ccmsg, 1024 * 1024); DO("schedule recv", isccc_ccmsg_readmessage(&ccmsg, task, rndc_recvnonce, NULL)); recvs++; DO("send message", isc_socket_send(sock, &r, task, rndc_senddone, NULL)); sends++; isc_event_free(&event); } static void rndc_startconnect(isc_sockaddr_t *addr, isc_task_t *task) { isc_result_t result; int pf; isc_sockettype_t type; char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_format(addr, socktext, sizeof(socktext)); notify("using server %s (%s)", servername, socktext); pf = isc_sockaddr_pf(addr); if (pf == AF_INET || pf == AF_INET6) type = isc_sockettype_tcp; else type = isc_sockettype_unix; DO("create socket", isc_socket_create(socketmgr, pf, type, &sock)); switch (isc_sockaddr_pf(addr)) { case AF_INET: DO("bind socket", isc_socket_bind(sock, &local4, 0)); break; case AF_INET6: DO("bind socket", isc_socket_bind(sock, &local6, 0)); break; default: break; } DO("connect", isc_socket_connect(sock, addr, task, rndc_connected, NULL)); connects++; } static void rndc_start(isc_task_t *task, isc_event_t *event) { isc_event_free(&event); currentaddr = 0; rndc_startconnect(&serveraddrs[currentaddr], task); } static void parse_config(isc_mem_t *mctx, isc_log_t *log, const char *keyname, cfg_parser_t **pctxp, cfg_obj_t **configp) { isc_result_t result; const char *conffile = admin_conffile; const cfg_obj_t *addresses = NULL; const cfg_obj_t *defkey = NULL; const cfg_obj_t *options = NULL; const cfg_obj_t *servers = NULL; const cfg_obj_t *server = NULL; const cfg_obj_t *keys = NULL; const cfg_obj_t *key = NULL; const cfg_obj_t *defport = NULL; const cfg_obj_t *secretobj = NULL; const cfg_obj_t *algorithmobj = NULL; cfg_obj_t *config = NULL; const cfg_obj_t *address = NULL; const cfg_listelt_t *elt; const char *secretstr; const char *algorithm; static char secretarray[1024]; const cfg_type_t *conftype = &cfg_type_rndcconf; isc_boolean_t key_only = ISC_FALSE; const cfg_listelt_t *element; if (! isc_file_exists(conffile)) { conffile = admin_keyfile; conftype = &cfg_type_rndckey; if (c_flag) fatal("%s does not exist", admin_conffile); if (! isc_file_exists(conffile)) fatal("neither %s nor %s was found", admin_conffile, admin_keyfile); key_only = ISC_TRUE; } else if (! c_flag && isc_file_exists(admin_keyfile)) { fprintf(stderr, "WARNING: key file (%s) exists, but using " "default configuration file (%s)\n", admin_keyfile, admin_conffile); } DO("create parser", cfg_parser_create(mctx, log, pctxp)); /* * The parser will output its own errors, so DO() is not used. */ result = cfg_parse_file(*pctxp, conffile, conftype, &config); if (result != ISC_R_SUCCESS) fatal("could not load rndc configuration"); if (!key_only) (void)cfg_map_get(config, "options", &options); if (key_only && servername == NULL) servername = "127.0.0.1"; else if (servername == NULL && options != NULL) { const cfg_obj_t *defserverobj = NULL; (void)cfg_map_get(options, "default-server", &defserverobj); if (defserverobj != NULL) servername = cfg_obj_asstring(defserverobj); } if (servername == NULL) fatal("no server specified and no default"); if (!key_only) { (void)cfg_map_get(config, "server", &servers); if (servers != NULL) { for (elt = cfg_list_first(servers); elt != NULL; elt = cfg_list_next(elt)) { const char *name; server = cfg_listelt_value(elt); name = cfg_obj_asstring(cfg_map_getname(server)); if (strcasecmp(name, servername) == 0) break; server = NULL; } } } /* * Look for the name of the key to use. */ if (keyname != NULL) ; /* Was set on command line, do nothing. */ else if (server != NULL) { DO("get key for server", cfg_map_get(server, "key", &defkey)); keyname = cfg_obj_asstring(defkey); } else if (options != NULL) { DO("get default key", cfg_map_get(options, "default-key", &defkey)); keyname = cfg_obj_asstring(defkey); } else if (!key_only) fatal("no key for server and no default"); /* * Get the key's definition. */ if (key_only) DO("get key", cfg_map_get(config, "key", &key)); else { DO("get config key list", cfg_map_get(config, "key", &keys)); for (elt = cfg_list_first(keys); elt != NULL; elt = cfg_list_next(elt)) { key = cfg_listelt_value(elt); if (strcasecmp(cfg_obj_asstring(cfg_map_getname(key)), keyname) == 0) break; } if (elt == NULL) fatal("no key definition for name %s", keyname); } (void)cfg_map_get(key, "secret", &secretobj); (void)cfg_map_get(key, "algorithm", &algorithmobj); if (secretobj == NULL || algorithmobj == NULL) fatal("key must have algorithm and secret"); secretstr = cfg_obj_asstring(secretobj); algorithm = cfg_obj_asstring(algorithmobj); if (strcasecmp(algorithm, "hmac-md5") != 0) fatal("unsupported algorithm: %s", algorithm); secret.rstart = (unsigned char *)secretarray; secret.rend = (unsigned char *)secretarray + sizeof(secretarray); DO("decode base64 secret", isccc_base64_decode(secretstr, &secret)); secret.rend = secret.rstart; secret.rstart = (unsigned char *)secretarray; /* * Find the port to connect to. */ if (remoteport != 0) ; /* Was set on command line, do nothing. */ else { if (server != NULL) (void)cfg_map_get(server, "port", &defport); if (defport == NULL && options != NULL) (void)cfg_map_get(options, "default-port", &defport); } if (defport != NULL) { remoteport = cfg_obj_asuint32(defport); if (remoteport > 65535 || remoteport == 0) fatal("port %u out of range", remoteport); } else if (remoteport == 0) remoteport = NS_CONTROL_PORT; if (server != NULL) result = cfg_map_get(server, "addresses", &addresses); else result = ISC_R_NOTFOUND; if (result == ISC_R_SUCCESS) { for (element = cfg_list_first(addresses); element != NULL; element = cfg_list_next(element)) { isc_sockaddr_t sa; address = cfg_listelt_value(element); if (!cfg_obj_issockaddr(address)) { unsigned int myport; const char *name; const cfg_obj_t *obj; obj = cfg_tuple_get(address, "name"); name = cfg_obj_asstring(obj); obj = cfg_tuple_get(address, "port"); if (cfg_obj_isuint32(obj)) { myport = cfg_obj_asuint32(obj); if (myport > ISC_UINT16_MAX || myport == 0) fatal("port %u out of range", myport); } else myport = remoteport; if (nserveraddrs < SERVERADDRS) get_addresses(name, (in_port_t) myport); else fprintf(stderr, "too many address: " "%s: dropped\n", name); continue; } sa = *cfg_obj_assockaddr(address); if (isc_sockaddr_getport(&sa) == 0) isc_sockaddr_setport(&sa, remoteport); if (nserveraddrs < SERVERADDRS) serveraddrs[nserveraddrs++] = sa; else { char socktext[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_format(&sa, socktext, sizeof(socktext)); fprintf(stderr, "too many address: %s: dropped\n", socktext); } } } if (!local4set && server != NULL) { address = NULL; cfg_map_get(server, "source-address", &address); if (address != NULL) { local4 = *cfg_obj_assockaddr(address); local4set = ISC_TRUE; } } if (!local4set && options != NULL) { address = NULL; cfg_map_get(options, "default-source-address", &address); if (address != NULL) { local4 = *cfg_obj_assockaddr(address); local4set = ISC_TRUE; } } if (!local6set && server != NULL) { address = NULL; cfg_map_get(server, "source-address-v6", &address); if (address != NULL) { local6 = *cfg_obj_assockaddr(address); local6set = ISC_TRUE; } } if (!local6set && options != NULL) { address = NULL; cfg_map_get(options, "default-source-address-v6", &address); if (address != NULL) { local6 = *cfg_obj_assockaddr(address); local6set = ISC_TRUE; } } *configp = config; } int main(int argc, char **argv) { isc_boolean_t show_final_mem = ISC_FALSE; isc_result_t result = ISC_R_SUCCESS; isc_taskmgr_t *taskmgr = NULL; isc_task_t *task = NULL; isc_log_t *log = NULL; isc_logconfig_t *logconfig = NULL; isc_logdestination_t logdest; cfg_parser_t *pctx = NULL; cfg_obj_t *config = NULL; const char *keyname = NULL; struct in_addr in; struct in6_addr in6; char *p; size_t argslen; int ch; int i; result = isc_file_progname(*argv, program, sizeof(program)); if (result != ISC_R_SUCCESS) memmove(program, "rndc", 5); progname = program; admin_conffile = RNDC_CONFFILE; admin_keyfile = RNDC_KEYFILE; isc_sockaddr_any(&local4); isc_sockaddr_any6(&local6); result = isc_app_start(); if (result != ISC_R_SUCCESS) fatal("isc_app_start() failed: %s", isc_result_totext(result)); isc_commandline_errprint = ISC_FALSE; while ((ch = isc_commandline_parse(argc, argv, "b:c:hk:Mmp:s:Vy:")) != -1) { switch (ch) { case 'b': if (inet_pton(AF_INET, isc_commandline_argument, &in) == 1) { isc_sockaddr_fromin(&local4, &in, 0); local4set = ISC_TRUE; } else if (inet_pton(AF_INET6, isc_commandline_argument, &in6) == 1) { isc_sockaddr_fromin6(&local6, &in6, 0); local6set = ISC_TRUE; } break; case 'c': admin_conffile = isc_commandline_argument; c_flag = ISC_TRUE; break; case 'k': admin_keyfile = isc_commandline_argument; break; case 'M': isc_mem_debugging = ISC_MEM_DEBUGTRACE; break; case 'm': show_final_mem = ISC_TRUE; break; case 'p': remoteport = atoi(isc_commandline_argument); if (remoteport > 65535 || remoteport == 0) fatal("port '%s' out of range", isc_commandline_argument); break; case 's': servername = isc_commandline_argument; break; case 'V': verbose = ISC_TRUE; break; case 'y': keyname = isc_commandline_argument; break; case '?': if (isc_commandline_option != '?') { fprintf(stderr, "%s: invalid argument -%c\n", program, isc_commandline_option); usage(1); } /* FALLTHROUGH */ case 'h': usage(0); break; default: fprintf(stderr, "%s: unhandled option -%c\n", program, isc_commandline_option); exit(1); } } argc -= isc_commandline_index; argv += isc_commandline_index; if (argc < 1) usage(1); isc_random_get(&serial); DO("create memory context", isc_mem_create(0, 0, &mctx)); DO("create socket manager", isc_socketmgr_create(mctx, &socketmgr)); DO("create task manager", isc_taskmgr_create(mctx, 1, 0, &taskmgr)); DO("create task", isc_task_create(taskmgr, 0, &task)); DO("create logging context", isc_log_create(mctx, &log, &logconfig)); isc_log_setcontext(log); DO("setting log tag", isc_log_settag(logconfig, progname)); logdest.file.stream = stderr; logdest.file.name = NULL; logdest.file.versions = ISC_LOG_ROLLNEVER; logdest.file.maximum_size = 0; DO("creating log channel", isc_log_createchannel(logconfig, "stderr", ISC_LOG_TOFILEDESC, ISC_LOG_INFO, &logdest, ISC_LOG_PRINTTAG|ISC_LOG_PRINTLEVEL)); DO("enabling log channel", isc_log_usechannel(logconfig, "stderr", NULL, NULL)); parse_config(mctx, log, keyname, &pctx, &config); isccc_result_register(); command = *argv; /* * Convert argc/argv into a space-delimited command string * similar to what the user might enter in interactive mode * (if that were implemented). */ argslen = 0; for (i = 0; i < argc; i++) argslen += strlen(argv[i]) + 1; args = isc_mem_get(mctx, argslen); if (args == NULL) DO("isc_mem_get", ISC_R_NOMEMORY); p = args; for (i = 0; i < argc; i++) { size_t len = strlen(argv[i]); memmove(p, argv[i], len); p += len; *p++ = ' '; } p--; *p++ = '\0'; INSIST(p == args + argslen); notify("%s", command); if (strcmp(command, "restart") == 0) fatal("'%s' is not implemented", command); if (nserveraddrs == 0) get_addresses(servername, (in_port_t) remoteport); DO("post event", isc_app_onrun(mctx, task, rndc_start, NULL)); result = isc_app_run(); if (result != ISC_R_SUCCESS) fatal("isc_app_run() failed: %s", isc_result_totext(result)); if (connects > 0 || sends > 0 || recvs > 0) isc_socket_cancel(sock, task, ISC_SOCKCANCEL_ALL); isc_task_detach(&task); isc_taskmgr_destroy(&taskmgr); isc_socketmgr_destroy(&socketmgr); isc_log_destroy(&log); isc_log_setcontext(NULL); cfg_obj_destroy(pctx, &config); cfg_parser_destroy(&pctx); isc_mem_put(mctx, args, argslen); isccc_ccmsg_invalidate(&ccmsg); dns_name_destroy(); if (show_final_mem) isc_mem_stats(mctx, stderr); isc_mem_destroy(&mctx); if (failed) return (1); return (0); } Index: releng/9.3/contrib/bind9/lib/dns/resolver.c =================================================================== --- releng/9.3/contrib/bind9/lib/dns/resolver.c (revision 296610) +++ releng/9.3/contrib/bind9/lib/dns/resolver.c (revision 296611) @@ -1,9130 +1,9127 @@ /* * Copyright (C) 2004-2014 Internet Systems Consortium, Inc. ("ISC") * Copyright (C) 1999-2003 Internet Software Consortium. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND ISC DISCLAIMS ALL WARRANTIES WITH * REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY * AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY SPECIAL, DIRECT, * INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM * LOSS OF USE, DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE * OR OTHER TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR * PERFORMANCE OF THIS SOFTWARE. */ /* $Id$ */ /*! \file */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define DNS_RESOLVER_TRACE #ifdef DNS_RESOLVER_TRACE #define RTRACE(m) isc_log_write(dns_lctx, \ DNS_LOGCATEGORY_RESOLVER, \ DNS_LOGMODULE_RESOLVER, \ ISC_LOG_DEBUG(3), \ "res %p: %s", res, (m)) #define RRTRACE(r, m) isc_log_write(dns_lctx, \ DNS_LOGCATEGORY_RESOLVER, \ DNS_LOGMODULE_RESOLVER, \ ISC_LOG_DEBUG(3), \ "res %p: %s", (r), (m)) #define FCTXTRACE(m) isc_log_write(dns_lctx, \ DNS_LOGCATEGORY_RESOLVER, \ DNS_LOGMODULE_RESOLVER, \ ISC_LOG_DEBUG(3), \ "fctx %p(%s): %s", fctx, fctx->info, (m)) #define FCTXTRACE2(m1, m2) \ isc_log_write(dns_lctx, \ DNS_LOGCATEGORY_RESOLVER, \ DNS_LOGMODULE_RESOLVER, \ ISC_LOG_DEBUG(3), \ "fctx %p(%s): %s %s", \ fctx, fctx->info, (m1), (m2)) #define FTRACE(m) isc_log_write(dns_lctx, \ DNS_LOGCATEGORY_RESOLVER, \ DNS_LOGMODULE_RESOLVER, \ ISC_LOG_DEBUG(3), \ "fetch %p (fctx %p(%s)): %s", \ fetch, fetch->private, \ fetch->private->info, (m)) #define QTRACE(m) isc_log_write(dns_lctx, \ DNS_LOGCATEGORY_RESOLVER, \ DNS_LOGMODULE_RESOLVER, \ ISC_LOG_DEBUG(3), \ "resquery %p (fctx %p(%s)): %s", \ query, query->fctx, \ query->fctx->info, (m)) #else #define RTRACE(m) #define RRTRACE(r, m) #define FCTXTRACE(m) #define FCTXTRACE2(m1, m2) #define FTRACE(m) #define QTRACE(m) #endif #define US_PER_SEC 1000000U /* * The maximum time we will wait for a single query. */ #define MAX_SINGLE_QUERY_TIMEOUT 9U #define MAX_SINGLE_QUERY_TIMEOUT_US (MAX_SINGLE_QUERY_TIMEOUT*US_PER_SEC) /* * We need to allow a individual query time to complete / timeout. */ #define MINIMUM_QUERY_TIMEOUT (MAX_SINGLE_QUERY_TIMEOUT + 1U) /* The default time in seconds for the whole query to live. */ #ifndef DEFAULT_QUERY_TIMEOUT #define DEFAULT_QUERY_TIMEOUT MINIMUM_QUERY_TIMEOUT #endif #ifndef MAXIMUM_QUERY_TIMEOUT #define MAXIMUM_QUERY_TIMEOUT 30 /* The maximum time in seconds for the whole query to live. */ #endif /* The default maximum number of recursions to follow before giving up. */ #ifndef DEFAULT_RECURSION_DEPTH #define DEFAULT_RECURSION_DEPTH 7 #endif /* The default maximum number of iterative queries to allow before giving up. */ #ifndef DEFAULT_MAX_QUERIES #define DEFAULT_MAX_QUERIES 50 #endif /*% * Maximum EDNS0 input packet size. */ #define RECV_BUFFER_SIZE 4096 /* XXXRTH Constant. */ #define EDNSOPTS 2 /*% * This defines the maximum number of timeouts we will permit before we * disable EDNS0 on the query. */ #define MAX_EDNS0_TIMEOUTS 3 typedef struct fetchctx fetchctx_t; typedef struct query { /* Locked by task event serialization. */ unsigned int magic; fetchctx_t * fctx; isc_mem_t * mctx; dns_dispatchmgr_t * dispatchmgr; dns_dispatch_t * dispatch; isc_boolean_t exclusivesocket; dns_adbaddrinfo_t * addrinfo; isc_socket_t * tcpsocket; isc_time_t start; dns_messageid_t id; dns_dispentry_t * dispentry; ISC_LINK(struct query) link; isc_buffer_t buffer; isc_buffer_t *tsig; dns_tsigkey_t *tsigkey; isc_socketevent_t sendevent; unsigned int options; unsigned int attributes; unsigned int sends; unsigned int connects; unsigned char data[512]; } resquery_t; #define QUERY_MAGIC ISC_MAGIC('Q', '!', '!', '!') #define VALID_QUERY(query) ISC_MAGIC_VALID(query, QUERY_MAGIC) #define RESQUERY_ATTR_CANCELED 0x02 #define RESQUERY_CONNECTING(q) ((q)->connects > 0) #define RESQUERY_CANCELED(q) (((q)->attributes & \ RESQUERY_ATTR_CANCELED) != 0) #define RESQUERY_SENDING(q) ((q)->sends > 0) typedef enum { fetchstate_init = 0, /*%< Start event has not run yet. */ fetchstate_active, fetchstate_done /*%< FETCHDONE events posted. */ } fetchstate; typedef enum { badns_unreachable = 0, badns_response, badns_validation } badnstype_t; struct fetchctx { /*% Not locked. */ unsigned int magic; dns_resolver_t * res; dns_name_t name; dns_rdatatype_t type; unsigned int options; unsigned int bucketnum; char * info; isc_mem_t * mctx; /*% Locked by appropriate bucket lock. */ fetchstate state; isc_boolean_t want_shutdown; isc_boolean_t cloned; isc_boolean_t spilled; unsigned int references; isc_event_t control_event; ISC_LINK(struct fetchctx) link; ISC_LIST(dns_fetchevent_t) events; /*% Locked by task event serialization. */ dns_name_t domain; dns_rdataset_t nameservers; unsigned int attributes; isc_timer_t * timer; isc_time_t expires; isc_interval_t interval; dns_message_t * qmessage; dns_message_t * rmessage; ISC_LIST(resquery_t) queries; dns_adbfindlist_t finds; dns_adbfind_t * find; dns_adbfindlist_t altfinds; dns_adbfind_t * altfind; dns_adbaddrinfolist_t forwaddrs; dns_adbaddrinfolist_t altaddrs; isc_sockaddrlist_t forwarders; dns_fwdpolicy_t fwdpolicy; isc_sockaddrlist_t bad; isc_sockaddrlist_t edns; isc_sockaddrlist_t edns512; isc_sockaddrlist_t bad_edns; dns_validator_t *validator; ISC_LIST(dns_validator_t) validators; dns_db_t * cache; dns_adb_t * adb; isc_boolean_t ns_ttl_ok; isc_uint32_t ns_ttl; /*% * The number of events we're waiting for. */ unsigned int pending; /*% * The number of times we've "restarted" the current * nameserver set. This acts as a failsafe to prevent * us from pounding constantly on a particular set of * servers that, for whatever reason, are not giving * us useful responses, but are responding in such a * way that they are not marked "bad". */ unsigned int restarts; /*% * The number of timeouts that have occurred since we * last successfully received a response packet. This * is used for EDNS0 black hole detection. */ unsigned int timeouts; /*% * Look aside state for DS lookups. */ dns_name_t nsname; dns_fetch_t * nsfetch; dns_rdataset_t nsrrset; /*% * Number of queries that reference this context. */ unsigned int nqueries; /*% * The reason to print when logging a successful * response to a query. */ const char * reason; /*% * Random numbers to use for mixing up server addresses. */ isc_uint32_t rand_buf; isc_uint32_t rand_bits; /*% * Fetch-local statistics for detailed logging. */ isc_result_t result; /*%< fetch result */ isc_result_t vresult; /*%< validation result */ int exitline; isc_time_t start; isc_uint64_t duration; isc_boolean_t logged; unsigned int querysent; unsigned int totalqueries; unsigned int referrals; unsigned int lamecount; unsigned int neterr; unsigned int badresp; unsigned int adberr; unsigned int findfail; unsigned int valfail; isc_boolean_t timeout; dns_adbaddrinfo_t *addrinfo; isc_sockaddr_t *client; unsigned int depth; }; #define FCTX_MAGIC ISC_MAGIC('F', '!', '!', '!') #define VALID_FCTX(fctx) ISC_MAGIC_VALID(fctx, FCTX_MAGIC) #define FCTX_ATTR_HAVEANSWER 0x0001 #define FCTX_ATTR_GLUING 0x0002 #define FCTX_ATTR_ADDRWAIT 0x0004 #define FCTX_ATTR_SHUTTINGDOWN 0x0008 #define FCTX_ATTR_WANTCACHE 0x0010 #define FCTX_ATTR_WANTNCACHE 0x0020 #define FCTX_ATTR_NEEDEDNS0 0x0040 #define FCTX_ATTR_TRIEDFIND 0x0080 #define FCTX_ATTR_TRIEDALT 0x0100 #define HAVE_ANSWER(f) (((f)->attributes & FCTX_ATTR_HAVEANSWER) != \ 0) #define GLUING(f) (((f)->attributes & FCTX_ATTR_GLUING) != \ 0) #define ADDRWAIT(f) (((f)->attributes & FCTX_ATTR_ADDRWAIT) != \ 0) #define SHUTTINGDOWN(f) (((f)->attributes & FCTX_ATTR_SHUTTINGDOWN) \ != 0) #define WANTCACHE(f) (((f)->attributes & FCTX_ATTR_WANTCACHE) != 0) #define WANTNCACHE(f) (((f)->attributes & FCTX_ATTR_WANTNCACHE) != 0) #define NEEDEDNS0(f) (((f)->attributes & FCTX_ATTR_NEEDEDNS0) != 0) #define TRIEDFIND(f) (((f)->attributes & FCTX_ATTR_TRIEDFIND) != 0) #define TRIEDALT(f) (((f)->attributes & FCTX_ATTR_TRIEDALT) != 0) typedef struct { dns_adbaddrinfo_t * addrinfo; fetchctx_t * fctx; } dns_valarg_t; struct dns_fetch { unsigned int magic; fetchctx_t * private; }; #define DNS_FETCH_MAGIC ISC_MAGIC('F', 't', 'c', 'h') #define DNS_FETCH_VALID(fetch) ISC_MAGIC_VALID(fetch, DNS_FETCH_MAGIC) typedef struct fctxbucket { isc_task_t * task; isc_mutex_t lock; ISC_LIST(fetchctx_t) fctxs; isc_boolean_t exiting; isc_mem_t * mctx; } fctxbucket_t; typedef struct alternate { isc_boolean_t isaddress; union { isc_sockaddr_t addr; struct { dns_name_t name; in_port_t port; } _n; } _u; ISC_LINK(struct alternate) link; } alternate_t; typedef struct dns_badcache dns_badcache_t; struct dns_badcache { dns_badcache_t * next; dns_rdatatype_t type; isc_time_t expire; unsigned int hashval; dns_name_t name; }; #define DNS_BADCACHE_SIZE 1021 #define DNS_BADCACHE_TTL(fctx) \ (((fctx)->res->lame_ttl > 30 ) ? (fctx)->res->lame_ttl : 30) struct dns_resolver { /* Unlocked. */ unsigned int magic; isc_mem_t * mctx; isc_mutex_t lock; isc_mutex_t nlock; isc_mutex_t primelock; dns_rdataclass_t rdclass; isc_socketmgr_t * socketmgr; isc_timermgr_t * timermgr; isc_taskmgr_t * taskmgr; dns_view_t * view; isc_boolean_t frozen; unsigned int options; dns_dispatchmgr_t * dispatchmgr; dns_dispatchset_t * dispatches4; isc_boolean_t exclusivev4; dns_dispatchset_t * dispatches6; isc_boolean_t exclusivev6; unsigned int nbuckets; fctxbucket_t * buckets; isc_uint32_t lame_ttl; ISC_LIST(alternate_t) alternates; isc_uint16_t udpsize; #if USE_ALGLOCK isc_rwlock_t alglock; #endif dns_rbt_t * algorithms; #if USE_MBSLOCK isc_rwlock_t mbslock; #endif dns_rbt_t * mustbesecure; unsigned int spillatmax; unsigned int spillatmin; isc_timer_t * spillattimer; isc_boolean_t zero_no_soa_ttl; unsigned int query_timeout; unsigned int maxdepth; /* Locked by lock. */ unsigned int references; isc_boolean_t exiting; isc_eventlist_t whenshutdown; unsigned int activebuckets; isc_boolean_t priming; unsigned int spillat; /* clients-per-query */ /* Bad cache. */ dns_badcache_t ** badcache; unsigned int badcount; unsigned int badhash; unsigned int badsweep; /* Locked by primelock. */ dns_fetch_t * primefetch; /* Locked by nlock. */ unsigned int nfctx; }; #define RES_MAGIC ISC_MAGIC('R', 'e', 's', '!') #define VALID_RESOLVER(res) ISC_MAGIC_VALID(res, RES_MAGIC) /*% * Private addrinfo flags. These must not conflict with DNS_FETCHOPT_NOEDNS0, * which we also use as an addrinfo flag. */ #define FCTX_ADDRINFO_MARK 0x0001 #define FCTX_ADDRINFO_FORWARDER 0x1000 #define FCTX_ADDRINFO_TRIED 0x2000 #define UNMARKED(a) (((a)->flags & FCTX_ADDRINFO_MARK) \ == 0) #define ISFORWARDER(a) (((a)->flags & \ FCTX_ADDRINFO_FORWARDER) != 0) #define TRIED(a) (((a)->flags & \ FCTX_ADDRINFO_TRIED) != 0) #define NXDOMAIN(r) (((r)->attributes & DNS_RDATASETATTR_NXDOMAIN) != 0) #define NEGATIVE(r) (((r)->attributes & DNS_RDATASETATTR_NEGATIVE) != 0) static void destroy(dns_resolver_t *res); static void empty_bucket(dns_resolver_t *res); static isc_result_t resquery_send(resquery_t *query); static void resquery_response(isc_task_t *task, isc_event_t *event); static void resquery_connected(isc_task_t *task, isc_event_t *event); static void fctx_try(fetchctx_t *fctx, isc_boolean_t retrying, isc_boolean_t badcache); static void fctx_destroy(fetchctx_t *fctx); static isc_boolean_t fctx_unlink(fetchctx_t *fctx); static isc_result_t ncache_adderesult(dns_message_t *message, dns_db_t *cache, dns_dbnode_t *node, dns_rdatatype_t covers, isc_stdtime_t now, dns_ttl_t maxttl, isc_boolean_t optout, isc_boolean_t secure, dns_rdataset_t *ardataset, isc_result_t *eresultp); static void validated(isc_task_t *task, isc_event_t *event); static isc_boolean_t maybe_destroy(fetchctx_t *fctx, isc_boolean_t locked); static void add_bad(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo, isc_result_t reason, badnstype_t badtype); static inline isc_result_t findnoqname(fetchctx_t *fctx, dns_name_t *name, dns_rdatatype_t type, dns_name_t **noqname); /*% * Increment resolver-related statistics counters. */ static inline void inc_stats(dns_resolver_t *res, isc_statscounter_t counter) { if (res->view->resstats != NULL) isc_stats_increment(res->view->resstats, counter); } static isc_result_t valcreate(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo, dns_name_t *name, dns_rdatatype_t type, dns_rdataset_t *rdataset, dns_rdataset_t *sigrdataset, unsigned int valoptions, isc_task_t *task) { dns_validator_t *validator = NULL; dns_valarg_t *valarg; isc_result_t result; valarg = isc_mem_get(fctx->mctx, sizeof(*valarg)); if (valarg == NULL) return (ISC_R_NOMEMORY); valarg->fctx = fctx; valarg->addrinfo = addrinfo; if (!ISC_LIST_EMPTY(fctx->validators)) INSIST((valoptions & DNS_VALIDATOR_DEFER) != 0); result = dns_validator_create(fctx->res->view, name, type, rdataset, sigrdataset, fctx->rmessage, valoptions, task, validated, valarg, &validator); if (result == ISC_R_SUCCESS) { inc_stats(fctx->res, dns_resstatscounter_val); if ((valoptions & DNS_VALIDATOR_DEFER) == 0) { INSIST(fctx->validator == NULL); fctx->validator = validator; } ISC_LIST_APPEND(fctx->validators, validator, link); } else isc_mem_put(fctx->mctx, valarg, sizeof(*valarg)); return (result); } static isc_boolean_t rrsig_fromchildzone(fetchctx_t *fctx, dns_rdataset_t *rdataset) { dns_namereln_t namereln; dns_rdata_rrsig_t rrsig; dns_rdata_t rdata = DNS_RDATA_INIT; int order; isc_result_t result; unsigned int labels; for (result = dns_rdataset_first(rdataset); result == ISC_R_SUCCESS; result = dns_rdataset_next(rdataset)) { dns_rdataset_current(rdataset, &rdata); result = dns_rdata_tostruct(&rdata, &rrsig, NULL); RUNTIME_CHECK(result == ISC_R_SUCCESS); namereln = dns_name_fullcompare(&rrsig.signer, &fctx->domain, &order, &labels); if (namereln == dns_namereln_subdomain) return (ISC_TRUE); dns_rdata_reset(&rdata); } return (ISC_FALSE); } static isc_boolean_t fix_mustbedelegationornxdomain(dns_message_t *message, fetchctx_t *fctx) { dns_name_t *name; dns_name_t *domain = &fctx->domain; dns_rdataset_t *rdataset; dns_rdatatype_t type; isc_result_t result; isc_boolean_t keep_auth = ISC_FALSE; if (message->rcode == dns_rcode_nxdomain) return (ISC_FALSE); /* * A DS RRset can appear anywhere in a zone, even for a delegation-only * zone. So a response to an explicit query for this type should be * excluded from delegation-only fixup. * * SOA, NS, and DNSKEY can only exist at a zone apex, so a postive * response to a query for these types can never violate the * delegation-only assumption: if the query name is below a * zone cut, the response should normally be a referral, which should * be accepted; if the query name is below a zone cut but the server * happens to have authority for the zone of the query name, the * response is a (non-referral) answer. But this does not violate * delegation-only because the query name must be in a different zone * due to the "apex-only" nature of these types. Note that if the * remote server happens to have authority for a child zone of a * delegation-only zone, we may still incorrectly "fix" the response * with NXDOMAIN for queries for other types. Unfortunately it's * generally impossible to differentiate this case from violation of * the delegation-only assumption. Once the resolver learns the * correct zone cut, possibly via a separate query for an "apex-only" * type, queries for other types will be resolved correctly. * * A query for type ANY will be accepted if it hits an exceptional * type above in the answer section as it should be from a child * zone. * * Also accept answers with RRSIG records from the child zone. * Direct queries for RRSIG records should not be answered from * the parent zone. */ if (message->counts[DNS_SECTION_ANSWER] != 0 && (fctx->type == dns_rdatatype_ns || fctx->type == dns_rdatatype_ds || fctx->type == dns_rdatatype_soa || fctx->type == dns_rdatatype_any || fctx->type == dns_rdatatype_rrsig || fctx->type == dns_rdatatype_dnskey)) { result = dns_message_firstname(message, DNS_SECTION_ANSWER); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(message, DNS_SECTION_ANSWER, &name); for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { if (!dns_name_equal(name, &fctx->name)) continue; type = rdataset->type; /* * RRsig from child? */ if (type == dns_rdatatype_rrsig && rrsig_fromchildzone(fctx, rdataset)) return (ISC_FALSE); /* * Direct query for apex records or DS. */ if (fctx->type == type && (type == dns_rdatatype_ds || type == dns_rdatatype_ns || type == dns_rdatatype_soa || type == dns_rdatatype_dnskey)) return (ISC_FALSE); /* * Indirect query for apex records or DS. */ if (fctx->type == dns_rdatatype_any && (type == dns_rdatatype_ns || type == dns_rdatatype_ds || type == dns_rdatatype_soa || type == dns_rdatatype_dnskey)) return (ISC_FALSE); } result = dns_message_nextname(message, DNS_SECTION_ANSWER); } } /* * A NODATA response to a DS query? */ if (fctx->type == dns_rdatatype_ds && message->counts[DNS_SECTION_ANSWER] == 0) return (ISC_FALSE); /* Look for referral or indication of answer from child zone? */ if (message->counts[DNS_SECTION_AUTHORITY] == 0) goto munge; result = dns_message_firstname(message, DNS_SECTION_AUTHORITY); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(message, DNS_SECTION_AUTHORITY, &name); for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { type = rdataset->type; if (type == dns_rdatatype_soa && dns_name_equal(name, domain)) keep_auth = ISC_TRUE; if (type != dns_rdatatype_ns && type != dns_rdatatype_soa && type != dns_rdatatype_rrsig) continue; if (type == dns_rdatatype_rrsig) { if (rrsig_fromchildzone(fctx, rdataset)) return (ISC_FALSE); else continue; } /* NS or SOA records. */ if (dns_name_equal(name, domain)) { /* * If a query for ANY causes a negative * response, we can be sure that this is * an empty node. For other type of queries * we cannot differentiate an empty node * from a node that just doesn't have that * type of record. We only accept the former * case. */ if (message->counts[DNS_SECTION_ANSWER] == 0 && fctx->type == dns_rdatatype_any) return (ISC_FALSE); } else if (dns_name_issubdomain(name, domain)) { /* Referral or answer from child zone. */ return (ISC_FALSE); } } result = dns_message_nextname(message, DNS_SECTION_AUTHORITY); } munge: message->rcode = dns_rcode_nxdomain; message->counts[DNS_SECTION_ANSWER] = 0; if (!keep_auth) message->counts[DNS_SECTION_AUTHORITY] = 0; message->counts[DNS_SECTION_ADDITIONAL] = 0; return (ISC_TRUE); } static inline isc_result_t fctx_starttimer(fetchctx_t *fctx) { /* * Start the lifetime timer for fctx. * * This is also used for stopping the idle timer; in that * case we must purge events already posted to ensure that * no further idle events are delivered. */ return (isc_timer_reset(fctx->timer, isc_timertype_once, &fctx->expires, NULL, ISC_TRUE)); } static inline void fctx_stoptimer(fetchctx_t *fctx) { isc_result_t result; /* * We don't return a result if resetting the timer to inactive fails * since there's nothing to be done about it. Resetting to inactive * should never fail anyway, since the code as currently written * cannot fail in that case. */ result = isc_timer_reset(fctx->timer, isc_timertype_inactive, NULL, NULL, ISC_TRUE); if (result != ISC_R_SUCCESS) { UNEXPECTED_ERROR(__FILE__, __LINE__, "isc_timer_reset(): %s", isc_result_totext(result)); } } static inline isc_result_t fctx_startidletimer(fetchctx_t *fctx, isc_interval_t *interval) { /* * Start the idle timer for fctx. The lifetime timer continues * to be in effect. */ return (isc_timer_reset(fctx->timer, isc_timertype_once, &fctx->expires, interval, ISC_FALSE)); } /* * Stopping the idle timer is equivalent to calling fctx_starttimer(), but * we use fctx_stopidletimer for readability in the code below. */ #define fctx_stopidletimer fctx_starttimer static inline void resquery_destroy(resquery_t **queryp) { resquery_t *query; REQUIRE(queryp != NULL); query = *queryp; REQUIRE(!ISC_LINK_LINKED(query, link)); INSIST(query->tcpsocket == NULL); query->fctx->nqueries--; if (SHUTTINGDOWN(query->fctx)) { dns_resolver_t *res = query->fctx->res; if (maybe_destroy(query->fctx, ISC_FALSE)) empty_bucket(res); } query->magic = 0; isc_mem_put(query->mctx, query, sizeof(*query)); *queryp = NULL; } static void fctx_cancelquery(resquery_t **queryp, dns_dispatchevent_t **deventp, isc_time_t *finish, isc_boolean_t no_response) { fetchctx_t *fctx; resquery_t *query; unsigned int rtt, rttms; unsigned int factor; dns_adbfind_t *find; dns_adbaddrinfo_t *addrinfo; isc_socket_t *socket; query = *queryp; fctx = query->fctx; FCTXTRACE("cancelquery"); REQUIRE(!RESQUERY_CANCELED(query)); query->attributes |= RESQUERY_ATTR_CANCELED; /* * Should we update the RTT? */ if (finish != NULL || no_response) { if (finish != NULL) { /* * We have both the start and finish times for this * packet, so we can compute a real RTT. */ rtt = (unsigned int)isc_time_microdiff(finish, &query->start); factor = DNS_ADB_RTTADJDEFAULT; rttms = rtt / 1000; if (rttms < DNS_RESOLVER_QRYRTTCLASS0) { inc_stats(fctx->res, dns_resstatscounter_queryrtt0); } else if (rttms < DNS_RESOLVER_QRYRTTCLASS1) { inc_stats(fctx->res, dns_resstatscounter_queryrtt1); } else if (rttms < DNS_RESOLVER_QRYRTTCLASS2) { inc_stats(fctx->res, dns_resstatscounter_queryrtt2); } else if (rttms < DNS_RESOLVER_QRYRTTCLASS3) { inc_stats(fctx->res, dns_resstatscounter_queryrtt3); } else if (rttms < DNS_RESOLVER_QRYRTTCLASS4) { inc_stats(fctx->res, dns_resstatscounter_queryrtt4); } else { inc_stats(fctx->res, dns_resstatscounter_queryrtt5); } } else { /* * We don't have an RTT for this query. Maybe the * packet was lost, or maybe this server is very * slow. We don't know. Increase the RTT. */ INSIST(no_response); rtt = query->addrinfo->srtt + 200000; if (rtt > MAX_SINGLE_QUERY_TIMEOUT_US) rtt = MAX_SINGLE_QUERY_TIMEOUT_US; /* * Replace the current RTT with our value. */ factor = DNS_ADB_RTTADJREPLACE; } dns_adb_adjustsrtt(fctx->adb, query->addrinfo, rtt, factor); } /* Remember that the server has been tried. */ if (!TRIED(query->addrinfo)) { dns_adb_changeflags(fctx->adb, query->addrinfo, FCTX_ADDRINFO_TRIED, FCTX_ADDRINFO_TRIED); } /* * Age RTTs of servers not tried. */ factor = DNS_ADB_RTTADJAGE; if (finish != NULL) for (addrinfo = ISC_LIST_HEAD(fctx->forwaddrs); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) if (UNMARKED(addrinfo)) dns_adb_adjustsrtt(fctx->adb, addrinfo, 0, factor); if (finish != NULL && TRIEDFIND(fctx)) for (find = ISC_LIST_HEAD(fctx->finds); find != NULL; find = ISC_LIST_NEXT(find, publink)) for (addrinfo = ISC_LIST_HEAD(find->list); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) if (UNMARKED(addrinfo)) dns_adb_adjustsrtt(fctx->adb, addrinfo, 0, factor); if (finish != NULL && TRIEDALT(fctx)) { for (addrinfo = ISC_LIST_HEAD(fctx->altaddrs); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) if (UNMARKED(addrinfo)) dns_adb_adjustsrtt(fctx->adb, addrinfo, 0, factor); for (find = ISC_LIST_HEAD(fctx->altfinds); find != NULL; find = ISC_LIST_NEXT(find, publink)) for (addrinfo = ISC_LIST_HEAD(find->list); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) if (UNMARKED(addrinfo)) dns_adb_adjustsrtt(fctx->adb, addrinfo, 0, factor); } /* * Check for any outstanding socket events. If they exist, cancel * them and let the event handlers finish the cleanup. The resolver * only needs to worry about managing the connect and send events; * the dispatcher manages the recv events. */ if (RESQUERY_CONNECTING(query)) { /* * Cancel the connect. */ if (query->tcpsocket != NULL) { isc_socket_cancel(query->tcpsocket, NULL, ISC_SOCKCANCEL_CONNECT); } else if (query->dispentry != NULL) { INSIST(query->exclusivesocket); socket = dns_dispatch_getentrysocket(query->dispentry); if (socket != NULL) isc_socket_cancel(socket, NULL, ISC_SOCKCANCEL_CONNECT); } } else if (RESQUERY_SENDING(query)) { /* * Cancel the pending send. */ if (query->exclusivesocket && query->dispentry != NULL) socket = dns_dispatch_getentrysocket(query->dispentry); else socket = dns_dispatch_getsocket(query->dispatch); if (socket != NULL) isc_socket_cancel(socket, NULL, ISC_SOCKCANCEL_SEND); } if (query->dispentry != NULL) dns_dispatch_removeresponse(&query->dispentry, deventp); ISC_LIST_UNLINK(fctx->queries, query, link); if (query->tsig != NULL) isc_buffer_free(&query->tsig); if (query->tsigkey != NULL) dns_tsigkey_detach(&query->tsigkey); if (query->dispatch != NULL) dns_dispatch_detach(&query->dispatch); if (! (RESQUERY_CONNECTING(query) || RESQUERY_SENDING(query))) /* * It's safe to destroy the query now. */ resquery_destroy(&query); } static void fctx_cancelqueries(fetchctx_t *fctx, isc_boolean_t no_response) { resquery_t *query, *next_query; FCTXTRACE("cancelqueries"); for (query = ISC_LIST_HEAD(fctx->queries); query != NULL; query = next_query) { next_query = ISC_LIST_NEXT(query, link); fctx_cancelquery(&query, NULL, NULL, no_response); } } static void fctx_cleanupfinds(fetchctx_t *fctx) { dns_adbfind_t *find, *next_find; REQUIRE(ISC_LIST_EMPTY(fctx->queries)); for (find = ISC_LIST_HEAD(fctx->finds); find != NULL; find = next_find) { next_find = ISC_LIST_NEXT(find, publink); ISC_LIST_UNLINK(fctx->finds, find, publink); dns_adb_destroyfind(&find); } fctx->find = NULL; } static void fctx_cleanupaltfinds(fetchctx_t *fctx) { dns_adbfind_t *find, *next_find; REQUIRE(ISC_LIST_EMPTY(fctx->queries)); for (find = ISC_LIST_HEAD(fctx->altfinds); find != NULL; find = next_find) { next_find = ISC_LIST_NEXT(find, publink); ISC_LIST_UNLINK(fctx->altfinds, find, publink); dns_adb_destroyfind(&find); } fctx->altfind = NULL; } static void fctx_cleanupforwaddrs(fetchctx_t *fctx) { dns_adbaddrinfo_t *addr, *next_addr; REQUIRE(ISC_LIST_EMPTY(fctx->queries)); for (addr = ISC_LIST_HEAD(fctx->forwaddrs); addr != NULL; addr = next_addr) { next_addr = ISC_LIST_NEXT(addr, publink); ISC_LIST_UNLINK(fctx->forwaddrs, addr, publink); dns_adb_freeaddrinfo(fctx->adb, &addr); } } static void fctx_cleanupaltaddrs(fetchctx_t *fctx) { dns_adbaddrinfo_t *addr, *next_addr; REQUIRE(ISC_LIST_EMPTY(fctx->queries)); for (addr = ISC_LIST_HEAD(fctx->altaddrs); addr != NULL; addr = next_addr) { next_addr = ISC_LIST_NEXT(addr, publink); ISC_LIST_UNLINK(fctx->altaddrs, addr, publink); dns_adb_freeaddrinfo(fctx->adb, &addr); } } static inline void fctx_stopeverything(fetchctx_t *fctx, isc_boolean_t no_response) { FCTXTRACE("stopeverything"); fctx_cancelqueries(fctx, no_response); fctx_cleanupfinds(fctx); fctx_cleanupaltfinds(fctx); fctx_cleanupforwaddrs(fctx); fctx_cleanupaltaddrs(fctx); fctx_stoptimer(fctx); } static inline void fctx_sendevents(fetchctx_t *fctx, isc_result_t result, int line) { dns_fetchevent_t *event, *next_event; isc_task_t *task; unsigned int count = 0; isc_interval_t i; isc_boolean_t logit = ISC_FALSE; isc_time_t now; unsigned int old_spillat; unsigned int new_spillat = 0; /* initialized to silence compiler warnings */ /* * Caller must be holding the appropriate bucket lock. */ REQUIRE(fctx->state == fetchstate_done); FCTXTRACE("sendevents"); /* * Keep some record of fetch result for logging later (if required). */ fctx->result = result; fctx->exitline = line; TIME_NOW(&now); fctx->duration = isc_time_microdiff(&now, &fctx->start); for (event = ISC_LIST_HEAD(fctx->events); event != NULL; event = next_event) { next_event = ISC_LIST_NEXT(event, ev_link); ISC_LIST_UNLINK(fctx->events, event, ev_link); task = event->ev_sender; event->ev_sender = fctx; event->vresult = fctx->vresult; if (!HAVE_ANSWER(fctx)) event->result = result; INSIST(result != ISC_R_SUCCESS || dns_rdataset_isassociated(event->rdataset) || fctx->type == dns_rdatatype_any || fctx->type == dns_rdatatype_rrsig || fctx->type == dns_rdatatype_sig); /* * Negative results must be indicated in event->result. */ if (dns_rdataset_isassociated(event->rdataset) && NEGATIVE(event->rdataset)) { INSIST(event->result == DNS_R_NCACHENXDOMAIN || event->result == DNS_R_NCACHENXRRSET); } event->qtotal = fctx->totalqueries; isc_task_sendanddetach(&task, ISC_EVENT_PTR(&event)); count++; } if ((fctx->attributes & FCTX_ATTR_HAVEANSWER) != 0 && fctx->spilled && (count < fctx->res->spillatmax || fctx->res->spillatmax == 0)) { LOCK(&fctx->res->lock); if (count == fctx->res->spillat && !fctx->res->exiting) { old_spillat = fctx->res->spillat; fctx->res->spillat += 5; if (fctx->res->spillat > fctx->res->spillatmax && fctx->res->spillatmax != 0) fctx->res->spillat = fctx->res->spillatmax; new_spillat = fctx->res->spillat; if (new_spillat != old_spillat) { logit = ISC_TRUE; } isc_interval_set(&i, 20 * 60, 0); result = isc_timer_reset(fctx->res->spillattimer, isc_timertype_ticker, NULL, &i, ISC_TRUE); RUNTIME_CHECK(result == ISC_R_SUCCESS); } UNLOCK(&fctx->res->lock); if (logit) isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "clients-per-query increased to %u", new_spillat); } } static inline void log_edns(fetchctx_t *fctx) { char domainbuf[DNS_NAME_FORMATSIZE]; if (fctx->reason == NULL) return; /* * We do not know if fctx->domain is the actual domain the record * lives in or a parent domain so we have a '?' after it. */ dns_name_format(&fctx->domain, domainbuf, sizeof(domainbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_EDNS_DISABLED, DNS_LOGMODULE_RESOLVER, ISC_LOG_INFO, "success resolving '%s' (in '%s'?) after %s", fctx->info, domainbuf, fctx->reason); fctx->reason = NULL; } static void fctx_done(fetchctx_t *fctx, isc_result_t result, int line) { dns_resolver_t *res; isc_boolean_t no_response; REQUIRE(line >= 0); FCTXTRACE("done"); res = fctx->res; if (result == ISC_R_SUCCESS) { /*% * Log any deferred EDNS timeout messages. */ log_edns(fctx); no_response = ISC_TRUE; } else no_response = ISC_FALSE; fctx->reason = NULL; fctx_stopeverything(fctx, no_response); LOCK(&res->buckets[fctx->bucketnum].lock); fctx->state = fetchstate_done; fctx->attributes &= ~FCTX_ATTR_ADDRWAIT; fctx_sendevents(fctx, result, line); UNLOCK(&res->buckets[fctx->bucketnum].lock); } static void process_sendevent(resquery_t *query, isc_event_t *event) { isc_socketevent_t *sevent = (isc_socketevent_t *)event; isc_boolean_t retry = ISC_FALSE; isc_result_t result; fetchctx_t *fctx; fctx = query->fctx; if (RESQUERY_CANCELED(query)) { if (query->sends == 0 && query->connects == 0) { /* * This query was canceled while the * isc_socket_sendto/connect() was in progress. */ if (query->tcpsocket != NULL) isc_socket_detach(&query->tcpsocket); resquery_destroy(&query); } } else { switch (sevent->result) { case ISC_R_SUCCESS: break; case ISC_R_HOSTUNREACH: case ISC_R_NETUNREACH: case ISC_R_NOPERM: case ISC_R_ADDRNOTAVAIL: case ISC_R_CONNREFUSED: /* * No route to remote. */ add_bad(fctx, query->addrinfo, sevent->result, badns_unreachable); fctx_cancelquery(&query, NULL, NULL, ISC_TRUE); retry = ISC_TRUE; break; default: fctx_cancelquery(&query, NULL, NULL, ISC_FALSE); break; } } if (event->ev_type == ISC_SOCKEVENT_CONNECT) isc_event_free(&event); if (retry) { /* * Behave as if the idle timer has expired. For TCP * this may not actually reflect the latest timer. */ fctx->attributes &= ~FCTX_ATTR_ADDRWAIT; result = fctx_stopidletimer(fctx); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else fctx_try(fctx, ISC_TRUE, ISC_FALSE); } } static void resquery_udpconnected(isc_task_t *task, isc_event_t *event) { resquery_t *query = event->ev_arg; REQUIRE(event->ev_type == ISC_SOCKEVENT_CONNECT); QTRACE("udpconnected"); UNUSED(task); INSIST(RESQUERY_CONNECTING(query)); query->connects--; process_sendevent(query, event); } static void resquery_senddone(isc_task_t *task, isc_event_t *event) { resquery_t *query = event->ev_arg; REQUIRE(event->ev_type == ISC_SOCKEVENT_SENDDONE); QTRACE("senddone"); /* * XXXRTH * * Currently we don't wait for the senddone event before retrying * a query. This means that if we get really behind, we may end * up doing extra work! */ UNUSED(task); INSIST(RESQUERY_SENDING(query)); query->sends--; process_sendevent(query, event); } static inline isc_result_t fctx_addopt(dns_message_t *message, unsigned int version, isc_uint16_t udpsize, dns_ednsopt_t *ednsopts, size_t count) { dns_rdataset_t *rdataset = NULL; isc_result_t result; result = dns_message_buildopt(message, &rdataset, version, udpsize, DNS_MESSAGEEXTFLAG_DO, ednsopts, count); if (result != ISC_R_SUCCESS) return (result); return (dns_message_setopt(message, rdataset)); } static inline void fctx_setretryinterval(fetchctx_t *fctx, unsigned int rtt) { unsigned int seconds; unsigned int us; /* * We retry every .8 seconds the first two times through the address * list, and then we do exponential back-off. */ if (fctx->restarts < 3) us = 800000; else us = (800000 << (fctx->restarts - 2)); /* * Add a fudge factor to the expected rtt based on the current * estimate. */ if (rtt < 50000) rtt += 50000; else if (rtt < 100000) rtt += 100000; else rtt += 200000; /* * Always wait for at least the expected rtt. */ if (us < rtt) us = rtt; /* * But don't ever wait for more than 10 seconds. */ if (us > MAX_SINGLE_QUERY_TIMEOUT_US) us = MAX_SINGLE_QUERY_TIMEOUT_US; seconds = us / US_PER_SEC; us -= seconds * US_PER_SEC; isc_interval_set(&fctx->interval, seconds, us * 1000); } static isc_result_t fctx_query(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo, unsigned int options) { dns_resolver_t *res; isc_task_t *task; isc_result_t result; resquery_t *query; isc_sockaddr_t addr; isc_boolean_t have_addr = ISC_FALSE; unsigned int srtt; FCTXTRACE("query"); res = fctx->res; task = res->buckets[fctx->bucketnum].task; srtt = addrinfo->srtt; /* * A forwarder needs to make multiple queries. Give it at least * a second to do these in. */ if (ISFORWARDER(addrinfo) && srtt < 1000000) srtt = 1000000; fctx_setretryinterval(fctx, srtt); result = fctx_startidletimer(fctx, &fctx->interval); if (result != ISC_R_SUCCESS) return (result); INSIST(ISC_LIST_EMPTY(fctx->validators)); dns_message_reset(fctx->rmessage, DNS_MESSAGE_INTENTPARSE); query = isc_mem_get(fctx->mctx, sizeof(*query)); if (query == NULL) { result = ISC_R_NOMEMORY; goto stop_idle_timer; } query->mctx = fctx->mctx; query->options = options; query->attributes = 0; query->sends = 0; query->connects = 0; /* * Note that the caller MUST guarantee that 'addrinfo' will remain * valid until this query is canceled. */ query->addrinfo = addrinfo; TIME_NOW(&query->start); /* * If this is a TCP query, then we need to make a socket and * a dispatch for it here. Otherwise we use the resolver's * shared dispatch. */ query->dispatchmgr = res->dispatchmgr; query->dispatch = NULL; query->exclusivesocket = ISC_FALSE; query->tcpsocket = NULL; if (res->view->peers != NULL) { dns_peer_t *peer = NULL; isc_netaddr_t dstip; isc_netaddr_fromsockaddr(&dstip, &addrinfo->sockaddr); result = dns_peerlist_peerbyaddr(res->view->peers, &dstip, &peer); if (result == ISC_R_SUCCESS) { result = dns_peer_getquerysource(peer, &addr); if (result == ISC_R_SUCCESS) have_addr = ISC_TRUE; } } if ((query->options & DNS_FETCHOPT_TCP) != 0) { int pf; pf = isc_sockaddr_pf(&addrinfo->sockaddr); if (!have_addr) { switch (pf) { case PF_INET: result = dns_dispatch_getlocaladdress( res->dispatches4->dispatches[0], &addr); break; case PF_INET6: result = dns_dispatch_getlocaladdress( res->dispatches6->dispatches[0], &addr); break; default: result = ISC_R_NOTIMPLEMENTED; break; } if (result != ISC_R_SUCCESS) goto cleanup_query; } isc_sockaddr_setport(&addr, 0); result = isc_socket_create(res->socketmgr, pf, isc_sockettype_tcp, &query->tcpsocket); if (result != ISC_R_SUCCESS) goto cleanup_query; #ifndef BROKEN_TCP_BIND_BEFORE_CONNECT result = isc_socket_bind(query->tcpsocket, &addr, 0); if (result != ISC_R_SUCCESS) goto cleanup_socket; #endif /* * A dispatch will be created once the connect succeeds. */ } else { if (have_addr) { unsigned int attrs, attrmask; attrs = DNS_DISPATCHATTR_UDP; switch (isc_sockaddr_pf(&addr)) { case AF_INET: attrs |= DNS_DISPATCHATTR_IPV4; break; case AF_INET6: attrs |= DNS_DISPATCHATTR_IPV6; break; default: result = ISC_R_NOTIMPLEMENTED; goto cleanup_query; } attrmask = DNS_DISPATCHATTR_UDP; attrmask |= DNS_DISPATCHATTR_TCP; attrmask |= DNS_DISPATCHATTR_IPV4; attrmask |= DNS_DISPATCHATTR_IPV6; result = dns_dispatch_getudp(res->dispatchmgr, res->socketmgr, res->taskmgr, &addr, 4096, 1000, 32768, 16411, 16433, attrs, attrmask, &query->dispatch); if (result != ISC_R_SUCCESS) goto cleanup_query; } else { switch (isc_sockaddr_pf(&addrinfo->sockaddr)) { case PF_INET: dns_dispatch_attach( dns_resolver_dispatchv4(res), &query->dispatch); query->exclusivesocket = res->exclusivev4; break; case PF_INET6: dns_dispatch_attach( dns_resolver_dispatchv6(res), &query->dispatch); query->exclusivesocket = res->exclusivev6; break; default: result = ISC_R_NOTIMPLEMENTED; goto cleanup_query; } } /* * We should always have a valid dispatcher here. If we * don't support a protocol family, then its dispatcher * will be NULL, but we shouldn't be finding addresses for * protocol types we don't support, so the dispatcher * we found should never be NULL. */ INSIST(query->dispatch != NULL); } query->dispentry = NULL; query->fctx = fctx; query->tsig = NULL; query->tsigkey = NULL; ISC_LINK_INIT(query, link); query->magic = QUERY_MAGIC; if ((query->options & DNS_FETCHOPT_TCP) != 0) { /* * Connect to the remote server. * * XXXRTH Should we attach to the socket? */ result = isc_socket_connect(query->tcpsocket, &addrinfo->sockaddr, task, resquery_connected, query); if (result != ISC_R_SUCCESS) goto cleanup_socket; query->connects++; QTRACE("connecting via TCP"); } else { result = resquery_send(query); if (result != ISC_R_SUCCESS) goto cleanup_dispatch; } fctx->querysent++; fctx->totalqueries++; ISC_LIST_APPEND(fctx->queries, query, link); query->fctx->nqueries++; if (isc_sockaddr_pf(&addrinfo->sockaddr) == PF_INET) inc_stats(res, dns_resstatscounter_queryv4); else inc_stats(res, dns_resstatscounter_queryv6); if (res->view->resquerystats != NULL) dns_rdatatypestats_increment(res->view->resquerystats, fctx->type); return (ISC_R_SUCCESS); cleanup_socket: isc_socket_detach(&query->tcpsocket); cleanup_dispatch: if (query->dispatch != NULL) dns_dispatch_detach(&query->dispatch); cleanup_query: if (query->connects == 0) { query->magic = 0; isc_mem_put(fctx->mctx, query, sizeof(*query)); } stop_idle_timer: RUNTIME_CHECK(fctx_stopidletimer(fctx) == ISC_R_SUCCESS); return (result); } static isc_boolean_t bad_edns(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; for (sa = ISC_LIST_HEAD(fctx->bad_edns); sa != NULL; sa = ISC_LIST_NEXT(sa, link)) { if (isc_sockaddr_equal(sa, address)) return (ISC_TRUE); } return (ISC_FALSE); } static void add_bad_edns(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; if (bad_edns(fctx, address)) return; sa = isc_mem_get(fctx->mctx, sizeof(*sa)); if (sa == NULL) return; *sa = *address; ISC_LIST_INITANDAPPEND(fctx->bad_edns, sa, link); } static isc_boolean_t triededns(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; for (sa = ISC_LIST_HEAD(fctx->edns); sa != NULL; sa = ISC_LIST_NEXT(sa, link)) { if (isc_sockaddr_equal(sa, address)) return (ISC_TRUE); } return (ISC_FALSE); } static void add_triededns(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; if (triededns(fctx, address)) return; sa = isc_mem_get(fctx->mctx, sizeof(*sa)); if (sa == NULL) return; *sa = *address; ISC_LIST_INITANDAPPEND(fctx->edns, sa, link); } static isc_boolean_t triededns512(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; for (sa = ISC_LIST_HEAD(fctx->edns512); sa != NULL; sa = ISC_LIST_NEXT(sa, link)) { if (isc_sockaddr_equal(sa, address)) return (ISC_TRUE); } return (ISC_FALSE); } static void add_triededns512(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; if (triededns512(fctx, address)) return; sa = isc_mem_get(fctx->mctx, sizeof(*sa)); if (sa == NULL) return; *sa = *address; ISC_LIST_INITANDAPPEND(fctx->edns512, sa, link); } static isc_result_t resquery_send(resquery_t *query) { fetchctx_t *fctx; isc_result_t result; dns_name_t *qname = NULL; dns_rdataset_t *qrdataset = NULL; isc_region_t r; dns_resolver_t *res; isc_task_t *task; isc_socket_t *socket; isc_buffer_t tcpbuffer; isc_sockaddr_t *address; isc_buffer_t *buffer; isc_netaddr_t ipaddr; dns_tsigkey_t *tsigkey = NULL; dns_peer_t *peer = NULL; isc_boolean_t useedns; dns_compress_t cctx; isc_boolean_t cleanup_cctx = ISC_FALSE; isc_boolean_t secure_domain; isc_boolean_t connecting = ISC_FALSE; dns_ednsopt_t ednsopts[EDNSOPTS]; unsigned ednsopt = 0; fctx = query->fctx; QTRACE("send"); res = fctx->res; task = res->buckets[fctx->bucketnum].task; address = NULL; if ((query->options & DNS_FETCHOPT_TCP) != 0) { /* * Reserve space for the TCP message length. */ isc_buffer_init(&tcpbuffer, query->data, sizeof(query->data)); isc_buffer_init(&query->buffer, query->data + 2, sizeof(query->data) - 2); buffer = &tcpbuffer; } else { isc_buffer_init(&query->buffer, query->data, sizeof(query->data)); buffer = &query->buffer; } result = dns_message_gettempname(fctx->qmessage, &qname); if (result != ISC_R_SUCCESS) goto cleanup_temps; result = dns_message_gettemprdataset(fctx->qmessage, &qrdataset); if (result != ISC_R_SUCCESS) goto cleanup_temps; /* * Get a query id from the dispatch. */ result = dns_dispatch_addresponse2(query->dispatch, &query->addrinfo->sockaddr, task, resquery_response, query, &query->id, &query->dispentry, res->socketmgr); if (result != ISC_R_SUCCESS) goto cleanup_temps; fctx->qmessage->opcode = dns_opcode_query; /* * Set up question. */ dns_name_init(qname, NULL); dns_name_clone(&fctx->name, qname); dns_rdataset_init(qrdataset); dns_rdataset_makequestion(qrdataset, res->rdclass, fctx->type); ISC_LIST_APPEND(qname->list, qrdataset, link); dns_message_addname(fctx->qmessage, qname, DNS_SECTION_QUESTION); qname = NULL; qrdataset = NULL; /* * Set RD if the client has requested that we do a recursive query, * or if we're sending to a forwarder. */ if ((query->options & DNS_FETCHOPT_RECURSIVE) != 0 || ISFORWARDER(query->addrinfo)) fctx->qmessage->flags |= DNS_MESSAGEFLAG_RD; /* * Set CD if the client says don't validate or the question is * under a secure entry point. */ if ((query->options & DNS_FETCHOPT_NOVALIDATE) != 0) { fctx->qmessage->flags |= DNS_MESSAGEFLAG_CD; } else if (res->view->enablevalidation) { result = dns_view_issecuredomain(res->view, &fctx->name, &secure_domain); if (result != ISC_R_SUCCESS) secure_domain = ISC_FALSE; if (res->view->dlv != NULL) secure_domain = ISC_TRUE; if (secure_domain) fctx->qmessage->flags |= DNS_MESSAGEFLAG_CD; } /* * We don't have to set opcode because it defaults to query. */ fctx->qmessage->id = query->id; /* * Convert the question to wire format. */ result = dns_compress_init(&cctx, -1, fctx->res->mctx); if (result != ISC_R_SUCCESS) goto cleanup_message; cleanup_cctx = ISC_TRUE; result = dns_message_renderbegin(fctx->qmessage, &cctx, &query->buffer); if (result != ISC_R_SUCCESS) goto cleanup_message; result = dns_message_rendersection(fctx->qmessage, DNS_SECTION_QUESTION, 0); if (result != ISC_R_SUCCESS) goto cleanup_message; peer = NULL; isc_netaddr_fromsockaddr(&ipaddr, &query->addrinfo->sockaddr); (void) dns_peerlist_peerbyaddr(fctx->res->view->peers, &ipaddr, &peer); /* * The ADB does not know about servers with "edns no". Check this, * and then inform the ADB for future use. */ if ((query->addrinfo->flags & DNS_FETCHOPT_NOEDNS0) == 0 && peer != NULL && dns_peer_getsupportedns(peer, &useedns) == ISC_R_SUCCESS && !useedns) { query->options |= DNS_FETCHOPT_NOEDNS0; dns_adb_changeflags(fctx->adb, query->addrinfo, DNS_FETCHOPT_NOEDNS0, DNS_FETCHOPT_NOEDNS0); } /* Sync NOEDNS0 flag in addrinfo->flags and options now. */ if ((query->addrinfo->flags & DNS_FETCHOPT_NOEDNS0) != 0) query->options |= DNS_FETCHOPT_NOEDNS0; /* * Handle timeouts by reducing the UDP response size to 512 bytes * then if that doesn't work disabling EDNS (includes DO) and CD. * * These timeout can be due to: * * broken nameservers that don't respond to EDNS queries. * * broken/misconfigured firewalls and NAT implementations * that don't handle IP fragmentation. * * broken/misconfigured firewalls that don't handle responses * greater than 512 bytes. * * broken/misconfigured firewalls that don't handle EDNS, DO * or CD. * * packet loss / link outage. */ if (fctx->timeout) { if ((triededns512(fctx, &query->addrinfo->sockaddr) || fctx->timeouts >= (MAX_EDNS0_TIMEOUTS * 2)) && (query->options & DNS_FETCHOPT_NOEDNS0) == 0) { query->options |= DNS_FETCHOPT_NOEDNS0; fctx->reason = "disabling EDNS"; } else if ((triededns(fctx, &query->addrinfo->sockaddr) || fctx->timeouts >= MAX_EDNS0_TIMEOUTS) && (query->options & DNS_FETCHOPT_NOEDNS0) == 0) { query->options |= DNS_FETCHOPT_EDNS512; fctx->reason = "reducing the advertised EDNS UDP " "packet size to 512 octets"; } fctx->timeout = ISC_FALSE; } /* * Use EDNS0, unless the caller doesn't want it, or we know that * the remote server doesn't like it. */ if ((query->options & DNS_FETCHOPT_NOEDNS0) == 0) { if ((query->addrinfo->flags & DNS_FETCHOPT_NOEDNS0) == 0) { unsigned int version = 0; /* Default version. */ unsigned int flags; isc_uint16_t udpsize = res->udpsize; isc_boolean_t reqnsid = res->view->requestnsid; flags = query->addrinfo->flags; if ((flags & DNS_FETCHOPT_EDNSVERSIONSET) != 0) { version = flags & DNS_FETCHOPT_EDNSVERSIONMASK; version >>= DNS_FETCHOPT_EDNSVERSIONSHIFT; } if ((query->options & DNS_FETCHOPT_EDNS512) != 0) udpsize = 512; else if (peer != NULL) (void)dns_peer_getudpsize(peer, &udpsize); /* request NSID for current view or peer? */ if (peer != NULL) (void) dns_peer_getrequestnsid(peer, &reqnsid); if (reqnsid) { INSIST(ednsopt < EDNSOPTS); ednsopts[ednsopt].code = DNS_OPT_NSID; ednsopts[ednsopt].length = 0; ednsopts[ednsopt].value = NULL; ednsopt++; } result = fctx_addopt(fctx->qmessage, version, udpsize, ednsopts, ednsopt); if (reqnsid && result == ISC_R_SUCCESS) { query->options |= DNS_FETCHOPT_WANTNSID; } else if (result != ISC_R_SUCCESS) { /* * We couldn't add the OPT, but we'll press on. * We're not using EDNS0, so set the NOEDNS0 * bit. */ query->options |= DNS_FETCHOPT_NOEDNS0; } } else { /* * We know this server doesn't like EDNS0, so we * won't use it. Set the NOEDNS0 bit since we're * not using EDNS0. */ query->options |= DNS_FETCHOPT_NOEDNS0; } } /* * If we need EDNS0 to do this query and aren't using it, we lose. */ if (NEEDEDNS0(fctx) && (query->options & DNS_FETCHOPT_NOEDNS0) != 0) { result = DNS_R_SERVFAIL; goto cleanup_message; } if ((query->options & DNS_FETCHOPT_NOEDNS0) == 0) add_triededns(fctx, &query->addrinfo->sockaddr); if ((query->options & DNS_FETCHOPT_EDNS512) != 0) add_triededns512(fctx, &query->addrinfo->sockaddr); /* * Clear CD if EDNS is not in use. */ if ((query->options & DNS_FETCHOPT_NOEDNS0) != 0) fctx->qmessage->flags &= ~DNS_MESSAGEFLAG_CD; /* * Add TSIG record tailored to the current recipient. */ result = dns_view_getpeertsig(fctx->res->view, &ipaddr, &tsigkey); if (result != ISC_R_SUCCESS && result != ISC_R_NOTFOUND) goto cleanup_message; if (tsigkey != NULL) { result = dns_message_settsigkey(fctx->qmessage, tsigkey); dns_tsigkey_detach(&tsigkey); if (result != ISC_R_SUCCESS) goto cleanup_message; } result = dns_message_rendersection(fctx->qmessage, DNS_SECTION_ADDITIONAL, 0); if (result != ISC_R_SUCCESS) goto cleanup_message; result = dns_message_renderend(fctx->qmessage); if (result != ISC_R_SUCCESS) goto cleanup_message; dns_compress_invalidate(&cctx); cleanup_cctx = ISC_FALSE; if (dns_message_gettsigkey(fctx->qmessage) != NULL) { dns_tsigkey_attach(dns_message_gettsigkey(fctx->qmessage), &query->tsigkey); result = dns_message_getquerytsig(fctx->qmessage, fctx->res->mctx, &query->tsig); if (result != ISC_R_SUCCESS) goto cleanup_message; } /* * If using TCP, write the length of the message at the beginning * of the buffer. */ if ((query->options & DNS_FETCHOPT_TCP) != 0) { isc_buffer_usedregion(&query->buffer, &r); isc_buffer_putuint16(&tcpbuffer, (isc_uint16_t)r.length); isc_buffer_add(&tcpbuffer, r.length); } /* * We're now done with the query message. */ dns_message_reset(fctx->qmessage, DNS_MESSAGE_INTENTRENDER); if (query->exclusivesocket) socket = dns_dispatch_getentrysocket(query->dispentry); else socket = dns_dispatch_getsocket(query->dispatch); /* * Send the query! */ if ((query->options & DNS_FETCHOPT_TCP) == 0) { address = &query->addrinfo->sockaddr; if (query->exclusivesocket) { result = isc_socket_connect(socket, address, task, resquery_udpconnected, query); if (result != ISC_R_SUCCESS) goto cleanup_message; connecting = ISC_TRUE; query->connects++; } } isc_buffer_usedregion(buffer, &r); /* * XXXRTH Make sure we don't send to ourselves! We should probably * prune out these addresses when we get them from the ADB. */ ISC_EVENT_INIT(&query->sendevent, sizeof(query->sendevent), 0, NULL, ISC_SOCKEVENT_SENDDONE, resquery_senddone, query, NULL, NULL, NULL); result = isc_socket_sendto2(socket, &r, task, address, NULL, &query->sendevent, 0); if (result != ISC_R_SUCCESS) { if (connecting) { /* * This query is still connecting. * Mark it as canceled so that it will just be * cleaned up when the connected event is received. * Keep fctx around until the event is processed. */ query->fctx->nqueries++; query->attributes |= RESQUERY_ATTR_CANCELED; } goto cleanup_message; } query->sends++; QTRACE("sent"); return (ISC_R_SUCCESS); cleanup_message: if (cleanup_cctx) dns_compress_invalidate(&cctx); dns_message_reset(fctx->qmessage, DNS_MESSAGE_INTENTRENDER); /* * Stop the dispatcher from listening. */ dns_dispatch_removeresponse(&query->dispentry, NULL); cleanup_temps: if (qname != NULL) dns_message_puttempname(fctx->qmessage, &qname); if (qrdataset != NULL) dns_message_puttemprdataset(fctx->qmessage, &qrdataset); return (result); } static void resquery_connected(isc_task_t *task, isc_event_t *event) { isc_socketevent_t *sevent = (isc_socketevent_t *)event; resquery_t *query = event->ev_arg; isc_boolean_t retry = ISC_FALSE; isc_interval_t interval; isc_result_t result; unsigned int attrs; fetchctx_t *fctx; REQUIRE(event->ev_type == ISC_SOCKEVENT_CONNECT); REQUIRE(VALID_QUERY(query)); QTRACE("connected"); UNUSED(task); /* * XXXRTH * * Currently we don't wait for the connect event before retrying * a query. This means that if we get really behind, we may end * up doing extra work! */ query->connects--; fctx = query->fctx; if (RESQUERY_CANCELED(query)) { /* * This query was canceled while the connect() was in * progress. */ isc_socket_detach(&query->tcpsocket); resquery_destroy(&query); } else { switch (sevent->result) { case ISC_R_SUCCESS: /* * Extend the idle timer for TCP. 20 seconds * should be long enough for a TCP connection to be * established, a single DNS request to be sent, * and the response received. */ isc_interval_set(&interval, 20, 0); result = fctx_startidletimer(query->fctx, &interval); if (result != ISC_R_SUCCESS) { fctx_cancelquery(&query, NULL, NULL, ISC_FALSE); fctx_done(fctx, result, __LINE__); break; } /* * We are connected. Create a dispatcher and * send the query. */ attrs = 0; attrs |= DNS_DISPATCHATTR_TCP; attrs |= DNS_DISPATCHATTR_PRIVATE; attrs |= DNS_DISPATCHATTR_CONNECTED; if (isc_sockaddr_pf(&query->addrinfo->sockaddr) == AF_INET) attrs |= DNS_DISPATCHATTR_IPV4; else attrs |= DNS_DISPATCHATTR_IPV6; attrs |= DNS_DISPATCHATTR_MAKEQUERY; result = dns_dispatch_createtcp(query->dispatchmgr, query->tcpsocket, query->fctx->res->taskmgr, 4096, 2, 1, 1, 3, attrs, &query->dispatch); /* * Regardless of whether dns_dispatch_create() * succeeded or not, we don't need our reference * to the socket anymore. */ isc_socket_detach(&query->tcpsocket); if (result == ISC_R_SUCCESS) result = resquery_send(query); if (result != ISC_R_SUCCESS) { fctx_cancelquery(&query, NULL, NULL, ISC_FALSE); fctx_done(fctx, result, __LINE__); } break; case ISC_R_NETUNREACH: case ISC_R_HOSTUNREACH: case ISC_R_CONNREFUSED: case ISC_R_NOPERM: case ISC_R_ADDRNOTAVAIL: case ISC_R_CONNECTIONRESET: /* * No route to remote. */ isc_socket_detach(&query->tcpsocket); fctx_cancelquery(&query, NULL, NULL, ISC_TRUE); retry = ISC_TRUE; break; default: isc_socket_detach(&query->tcpsocket); fctx_cancelquery(&query, NULL, NULL, ISC_FALSE); break; } } isc_event_free(&event); if (retry) { /* * Behave as if the idle timer has expired. For TCP * connections this may not actually reflect the latest timer. */ fctx->attributes &= ~FCTX_ATTR_ADDRWAIT; result = fctx_stopidletimer(fctx); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else fctx_try(fctx, ISC_TRUE, ISC_FALSE); } } static void fctx_finddone(isc_task_t *task, isc_event_t *event) { fetchctx_t *fctx; dns_adbfind_t *find; dns_resolver_t *res; isc_boolean_t want_try = ISC_FALSE; isc_boolean_t want_done = ISC_FALSE; isc_boolean_t bucket_empty = ISC_FALSE; unsigned int bucketnum; isc_boolean_t destroy = ISC_FALSE; find = event->ev_sender; fctx = event->ev_arg; REQUIRE(VALID_FCTX(fctx)); res = fctx->res; UNUSED(task); FCTXTRACE("finddone"); bucketnum = fctx->bucketnum; LOCK(&res->buckets[bucketnum].lock); INSIST(fctx->pending > 0); fctx->pending--; if (ADDRWAIT(fctx)) { /* * The fetch is waiting for a name to be found. */ INSIST(!SHUTTINGDOWN(fctx)); fctx->attributes &= ~FCTX_ATTR_ADDRWAIT; if (event->ev_type == DNS_EVENT_ADBMOREADDRESSES) { want_try = ISC_TRUE; fctx->totalqueries += find->qtotal; } else { fctx->findfail++; if (fctx->pending == 0) { /* * We've got nothing else to wait for and don't * know the answer. There's nothing to do but * fail the fctx. */ want_done = ISC_TRUE; } } } else if (SHUTTINGDOWN(fctx) && fctx->pending == 0 && fctx->nqueries == 0 && ISC_LIST_EMPTY(fctx->validators)) { if (fctx->references == 0) { bucket_empty = fctx_unlink(fctx); destroy = ISC_TRUE; } } UNLOCK(&res->buckets[bucketnum].lock); isc_event_free(&event); dns_adb_destroyfind(&find); if (want_try) fctx_try(fctx, ISC_TRUE, ISC_FALSE); else if (want_done) fctx_done(fctx, ISC_R_FAILURE, __LINE__); else if (destroy) { fctx_destroy(fctx); if (bucket_empty) empty_bucket(res); } } static inline isc_boolean_t bad_server(fetchctx_t *fctx, isc_sockaddr_t *address) { isc_sockaddr_t *sa; for (sa = ISC_LIST_HEAD(fctx->bad); sa != NULL; sa = ISC_LIST_NEXT(sa, link)) { if (isc_sockaddr_equal(sa, address)) return (ISC_TRUE); } return (ISC_FALSE); } static inline isc_boolean_t mark_bad(fetchctx_t *fctx) { dns_adbfind_t *curr; dns_adbaddrinfo_t *addrinfo; isc_boolean_t all_bad = ISC_TRUE; /* * Mark all known bad servers, so we don't try to talk to them * again. */ /* * Mark any bad nameservers. */ for (curr = ISC_LIST_HEAD(fctx->finds); curr != NULL; curr = ISC_LIST_NEXT(curr, publink)) { for (addrinfo = ISC_LIST_HEAD(curr->list); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (bad_server(fctx, &addrinfo->sockaddr)) addrinfo->flags |= FCTX_ADDRINFO_MARK; else all_bad = ISC_FALSE; } } /* * Mark any bad forwarders. */ for (addrinfo = ISC_LIST_HEAD(fctx->forwaddrs); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (bad_server(fctx, &addrinfo->sockaddr)) addrinfo->flags |= FCTX_ADDRINFO_MARK; else all_bad = ISC_FALSE; } /* * Mark any bad alternates. */ for (curr = ISC_LIST_HEAD(fctx->altfinds); curr != NULL; curr = ISC_LIST_NEXT(curr, publink)) { for (addrinfo = ISC_LIST_HEAD(curr->list); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (bad_server(fctx, &addrinfo->sockaddr)) addrinfo->flags |= FCTX_ADDRINFO_MARK; else all_bad = ISC_FALSE; } } for (addrinfo = ISC_LIST_HEAD(fctx->altaddrs); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (bad_server(fctx, &addrinfo->sockaddr)) addrinfo->flags |= FCTX_ADDRINFO_MARK; else all_bad = ISC_FALSE; } return (all_bad); } static void add_bad(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo, isc_result_t reason, badnstype_t badtype) { char namebuf[DNS_NAME_FORMATSIZE]; char addrbuf[ISC_SOCKADDR_FORMATSIZE]; char classbuf[64]; char typebuf[64]; char code[64]; isc_buffer_t b; isc_sockaddr_t *sa; const char *spc = ""; isc_sockaddr_t *address = &addrinfo->sockaddr; if (reason == DNS_R_LAME) fctx->lamecount++; else { switch (badtype) { case badns_unreachable: fctx->neterr++; break; case badns_response: fctx->badresp++; break; case badns_validation: break; /* counted as 'valfail' */ } } if (bad_server(fctx, address)) { /* * We already know this server is bad. */ return; } FCTXTRACE("add_bad"); sa = isc_mem_get(fctx->mctx, sizeof(*sa)); if (sa == NULL) return; *sa = *address; ISC_LIST_INITANDAPPEND(fctx->bad, sa, link); if (reason == DNS_R_LAME) /* already logged */ return; if (reason == DNS_R_UNEXPECTEDRCODE && fctx->rmessage->rcode == dns_rcode_servfail && ISFORWARDER(addrinfo)) return; if (reason == DNS_R_UNEXPECTEDRCODE) { isc_buffer_init(&b, code, sizeof(code) - 1); dns_rcode_totext(fctx->rmessage->rcode, &b); code[isc_buffer_usedlength(&b)] = '\0'; spc = " "; } else if (reason == DNS_R_UNEXPECTEDOPCODE) { isc_buffer_init(&b, code, sizeof(code) - 1); dns_opcode_totext((dns_opcode_t)fctx->rmessage->opcode, &b); code[isc_buffer_usedlength(&b)] = '\0'; spc = " "; } else { code[0] = '\0'; } dns_name_format(&fctx->name, namebuf, sizeof(namebuf)); dns_rdatatype_format(fctx->type, typebuf, sizeof(typebuf)); dns_rdataclass_format(fctx->res->rdclass, classbuf, sizeof(classbuf)); isc_sockaddr_format(address, addrbuf, sizeof(addrbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_LAME_SERVERS, DNS_LOGMODULE_RESOLVER, ISC_LOG_INFO, "error (%s%s%s) resolving '%s/%s/%s': %s", dns_result_totext(reason), spc, code, namebuf, typebuf, classbuf, addrbuf); } /* * Sort addrinfo list by RTT. */ static void sort_adbfind(dns_adbfind_t *find) { dns_adbaddrinfo_t *best, *curr; dns_adbaddrinfolist_t sorted; /* Lame N^2 bubble sort. */ ISC_LIST_INIT(sorted); while (!ISC_LIST_EMPTY(find->list)) { best = ISC_LIST_HEAD(find->list); curr = ISC_LIST_NEXT(best, publink); while (curr != NULL) { if (curr->srtt < best->srtt) best = curr; curr = ISC_LIST_NEXT(curr, publink); } ISC_LIST_UNLINK(find->list, best, publink); ISC_LIST_APPEND(sorted, best, publink); } find->list = sorted; } /* * Sort a list of finds by server RTT. */ static void sort_finds(dns_adbfindlist_t *findlist) { dns_adbfind_t *best, *curr; dns_adbfindlist_t sorted; dns_adbaddrinfo_t *addrinfo, *bestaddrinfo; /* Sort each find's addrinfo list by SRTT. */ for (curr = ISC_LIST_HEAD(*findlist); curr != NULL; curr = ISC_LIST_NEXT(curr, publink)) sort_adbfind(curr); /* Lame N^2 bubble sort. */ ISC_LIST_INIT(sorted); while (!ISC_LIST_EMPTY(*findlist)) { best = ISC_LIST_HEAD(*findlist); bestaddrinfo = ISC_LIST_HEAD(best->list); INSIST(bestaddrinfo != NULL); curr = ISC_LIST_NEXT(best, publink); while (curr != NULL) { addrinfo = ISC_LIST_HEAD(curr->list); INSIST(addrinfo != NULL); if (addrinfo->srtt < bestaddrinfo->srtt) { best = curr; bestaddrinfo = addrinfo; } curr = ISC_LIST_NEXT(curr, publink); } ISC_LIST_UNLINK(*findlist, best, publink); ISC_LIST_APPEND(sorted, best, publink); } *findlist = sorted; } static void findname(fetchctx_t *fctx, dns_name_t *name, in_port_t port, unsigned int options, unsigned int flags, isc_stdtime_t now, isc_boolean_t *need_alternate) { dns_adbaddrinfo_t *ai; dns_adbfind_t *find; dns_resolver_t *res; isc_boolean_t unshared; isc_result_t result; res = fctx->res; unshared = ISC_TF((fctx->options & DNS_FETCHOPT_UNSHARED) != 0); /* * If this name is a subdomain of the query domain, tell * the ADB to start looking using zone/hint data. This keeps us * from getting stuck if the nameserver is beneath the zone cut * and we don't know its address (e.g. because the A record has * expired). */ if (dns_name_issubdomain(name, &fctx->domain)) options |= DNS_ADBFIND_STARTATZONE; options |= DNS_ADBFIND_GLUEOK; options |= DNS_ADBFIND_HINTOK; /* * See what we know about this address. */ find = NULL; result = dns_adb_createfind2(fctx->adb, res->buckets[fctx->bucketnum].task, fctx_finddone, fctx, name, &fctx->name, fctx->type, options, now, NULL, res->view->dstport, fctx->depth + 1, &find); if (result != ISC_R_SUCCESS) { if (result == DNS_R_ALIAS) { /* * XXXRTH Follow the CNAME/DNAME chain? */ dns_adb_destroyfind(&find); fctx->adberr++; } } else if (!ISC_LIST_EMPTY(find->list)) { /* * We have at least some of the addresses for the * name. */ INSIST((find->options & DNS_ADBFIND_WANTEVENT) == 0); if (flags != 0 || port != 0) { for (ai = ISC_LIST_HEAD(find->list); ai != NULL; ai = ISC_LIST_NEXT(ai, publink)) { ai->flags |= flags; if (port != 0) isc_sockaddr_setport(&ai->sockaddr, port); } } if ((flags & FCTX_ADDRINFO_FORWARDER) != 0) ISC_LIST_APPEND(fctx->altfinds, find, publink); else ISC_LIST_APPEND(fctx->finds, find, publink); } else { /* * We don't know any of the addresses for this * name. */ if ((find->options & DNS_ADBFIND_WANTEVENT) != 0) { /* * We're looking for them and will get an * event about it later. */ fctx->pending++; /* * Bootstrap. */ if (need_alternate != NULL && !*need_alternate && unshared && ((res->dispatches4 == NULL && find->result_v6 != DNS_R_NXDOMAIN) || (res->dispatches6 == NULL && find->result_v4 != DNS_R_NXDOMAIN))) *need_alternate = ISC_TRUE; } else { if ((find->options & DNS_ADBFIND_LAMEPRUNED) != 0) fctx->lamecount++; /* cached lame server */ else fctx->adberr++; /* unreachable server, etc. */ /* * If we know there are no addresses for * the family we are using then try to add * an alternative server. */ if (need_alternate != NULL && !*need_alternate && ((res->dispatches4 == NULL && find->result_v6 == DNS_R_NXRRSET) || (res->dispatches6 == NULL && find->result_v4 == DNS_R_NXRRSET))) *need_alternate = ISC_TRUE; dns_adb_destroyfind(&find); } } } static isc_boolean_t isstrictsubdomain(dns_name_t *name1, dns_name_t *name2) { int order; unsigned int nlabels; dns_namereln_t namereln; namereln = dns_name_fullcompare(name1, name2, &order, &nlabels); return (ISC_TF(namereln == dns_namereln_subdomain)); } static isc_result_t fctx_getaddresses(fetchctx_t *fctx, isc_boolean_t badcache) { dns_rdata_t rdata = DNS_RDATA_INIT; isc_result_t result; dns_resolver_t *res; isc_stdtime_t now; unsigned int stdoptions = 0; isc_sockaddr_t *sa; dns_adbaddrinfo_t *ai; isc_boolean_t all_bad; dns_rdata_ns_t ns; isc_boolean_t need_alternate = ISC_FALSE; FCTXTRACE("getaddresses"); /* * Don't pound on remote servers. (Failsafe!) */ fctx->restarts++; if (fctx->restarts > 10) { FCTXTRACE("too many restarts"); return (DNS_R_SERVFAIL); } res = fctx->res; if (fctx->depth > res->maxdepth) { FCTXTRACE("too much NS indirection"); return (DNS_R_SERVFAIL); } /* * Forwarders. */ INSIST(ISC_LIST_EMPTY(fctx->forwaddrs)); INSIST(ISC_LIST_EMPTY(fctx->altaddrs)); /* * If this fctx has forwarders, use them; otherwise use any * selective forwarders specified in the view; otherwise use the * resolver's forwarders (if any). */ sa = ISC_LIST_HEAD(fctx->forwarders); if (sa == NULL) { dns_forwarders_t *forwarders = NULL; dns_name_t *name = &fctx->name; dns_name_t suffix; unsigned int labels; dns_fixedname_t fixed; dns_name_t *domain; /* * DS records are found in the parent server. * Strip label to get the correct forwarder (if any). */ if (dns_rdatatype_atparent(fctx->type) && dns_name_countlabels(name) > 1) { dns_name_init(&suffix, NULL); labels = dns_name_countlabels(name); dns_name_getlabelsequence(name, 1, labels - 1, &suffix); name = &suffix; } dns_fixedname_init(&fixed); domain = dns_fixedname_name(&fixed); result = dns_fwdtable_find2(fctx->res->view->fwdtable, name, domain, &forwarders); if (result == ISC_R_SUCCESS) { sa = ISC_LIST_HEAD(forwarders->addrs); fctx->fwdpolicy = forwarders->fwdpolicy; if (fctx->fwdpolicy == dns_fwdpolicy_only && isstrictsubdomain(domain, &fctx->domain)) { dns_name_free(&fctx->domain, fctx->mctx); dns_name_init(&fctx->domain, NULL); result = dns_name_dup(domain, fctx->mctx, &fctx->domain); if (result != ISC_R_SUCCESS) return (result); } } } while (sa != NULL) { if ((isc_sockaddr_pf(sa) == AF_INET && fctx->res->dispatches4 == NULL) || (isc_sockaddr_pf(sa) == AF_INET6 && fctx->res->dispatches6 == NULL)) { sa = ISC_LIST_NEXT(sa, link); continue; } ai = NULL; result = dns_adb_findaddrinfo(fctx->adb, sa, &ai, 0); /* XXXMLG */ if (result == ISC_R_SUCCESS) { dns_adbaddrinfo_t *cur; ai->flags |= FCTX_ADDRINFO_FORWARDER; cur = ISC_LIST_HEAD(fctx->forwaddrs); while (cur != NULL && cur->srtt < ai->srtt) cur = ISC_LIST_NEXT(cur, publink); if (cur != NULL) ISC_LIST_INSERTBEFORE(fctx->forwaddrs, cur, ai, publink); else ISC_LIST_APPEND(fctx->forwaddrs, ai, publink); } sa = ISC_LIST_NEXT(sa, link); } /* * If the forwarding policy is "only", we don't need the addresses * of the nameservers. */ if (fctx->fwdpolicy == dns_fwdpolicy_only) goto out; /* * Normal nameservers. */ stdoptions = DNS_ADBFIND_WANTEVENT | DNS_ADBFIND_EMPTYEVENT; if (fctx->restarts == 1) { /* * To avoid sending out a flood of queries likely to * result in NXRRSET, we suppress fetches for address * families we don't have the first time through, * provided that we have addresses in some family we * can use. * * We don't want to set this option all the time, since * if fctx->restarts > 1, we've clearly been having trouble * with the addresses we had, so getting more could help. */ stdoptions |= DNS_ADBFIND_AVOIDFETCHES; } if (res->dispatches4 != NULL) stdoptions |= DNS_ADBFIND_INET; if (res->dispatches6 != NULL) stdoptions |= DNS_ADBFIND_INET6; isc_stdtime_get(&now); INSIST(ISC_LIST_EMPTY(fctx->finds)); INSIST(ISC_LIST_EMPTY(fctx->altfinds)); for (result = dns_rdataset_first(&fctx->nameservers); result == ISC_R_SUCCESS; result = dns_rdataset_next(&fctx->nameservers)) { dns_rdataset_current(&fctx->nameservers, &rdata); /* * Extract the name from the NS record. */ result = dns_rdata_tostruct(&rdata, &ns, NULL); if (result != ISC_R_SUCCESS) continue; findname(fctx, &ns.name, 0, stdoptions, 0, now, &need_alternate); dns_rdata_reset(&rdata); dns_rdata_freestruct(&ns); } if (result != ISC_R_NOMORE) return (result); /* * Do we need to use 6 to 4? */ if (need_alternate) { int family; alternate_t *a; family = (res->dispatches6 != NULL) ? AF_INET6 : AF_INET; for (a = ISC_LIST_HEAD(fctx->res->alternates); a != NULL; a = ISC_LIST_NEXT(a, link)) { if (!a->isaddress) { findname(fctx, &a->_u._n.name, a->_u._n.port, stdoptions, FCTX_ADDRINFO_FORWARDER, now, NULL); continue; } if (isc_sockaddr_pf(&a->_u.addr) != family) continue; ai = NULL; result = dns_adb_findaddrinfo(fctx->adb, &a->_u.addr, &ai, 0); if (result == ISC_R_SUCCESS) { dns_adbaddrinfo_t *cur; ai->flags |= FCTX_ADDRINFO_FORWARDER; cur = ISC_LIST_HEAD(fctx->altaddrs); while (cur != NULL && cur->srtt < ai->srtt) cur = ISC_LIST_NEXT(cur, publink); if (cur != NULL) ISC_LIST_INSERTBEFORE(fctx->altaddrs, cur, ai, publink); else ISC_LIST_APPEND(fctx->altaddrs, ai, publink); } } } out: /* * Mark all known bad servers. */ all_bad = mark_bad(fctx); /* * How are we doing? */ if (all_bad) { /* * We've got no addresses. */ if (fctx->pending > 0) { /* * We're fetching the addresses, but don't have any * yet. Tell the caller to wait for an answer. */ result = DNS_R_WAIT; } else { isc_time_t expire; isc_interval_t i; /* * We've lost completely. We don't know any * addresses, and the ADB has told us it can't get * them. */ FCTXTRACE("no addresses"); isc_interval_set(&i, DNS_BADCACHE_TTL(fctx), 0); result = isc_time_nowplusinterval(&expire, &i); if (badcache && (fctx->type == dns_rdatatype_dnskey || fctx->type == dns_rdatatype_dlv || fctx->type == dns_rdatatype_ds) && result == ISC_R_SUCCESS) dns_resolver_addbadcache(fctx->res, &fctx->name, fctx->type, &expire); result = ISC_R_FAILURE; } } else { /* * We've found some addresses. We might still be looking * for more addresses. */ sort_finds(&fctx->finds); sort_finds(&fctx->altfinds); result = ISC_R_SUCCESS; } return (result); } static inline void possibly_mark(fetchctx_t *fctx, dns_adbaddrinfo_t *addr) { isc_netaddr_t na; char buf[ISC_NETADDR_FORMATSIZE]; isc_sockaddr_t *sa; isc_boolean_t aborted = ISC_FALSE; isc_boolean_t bogus; dns_acl_t *blackhole; isc_netaddr_t ipaddr; dns_peer_t *peer = NULL; dns_resolver_t *res; const char *msg = NULL; sa = &addr->sockaddr; res = fctx->res; isc_netaddr_fromsockaddr(&ipaddr, sa); blackhole = dns_dispatchmgr_getblackhole(res->dispatchmgr); (void) dns_peerlist_peerbyaddr(res->view->peers, &ipaddr, &peer); if (blackhole != NULL) { int match; if (dns_acl_match(&ipaddr, NULL, blackhole, &res->view->aclenv, &match, NULL) == ISC_R_SUCCESS && match > 0) aborted = ISC_TRUE; } if (peer != NULL && dns_peer_getbogus(peer, &bogus) == ISC_R_SUCCESS && bogus) aborted = ISC_TRUE; if (aborted) { addr->flags |= FCTX_ADDRINFO_MARK; msg = "ignoring blackholed / bogus server: "; } else if (isc_sockaddr_ismulticast(sa)) { addr->flags |= FCTX_ADDRINFO_MARK; msg = "ignoring multicast address: "; } else if (isc_sockaddr_isexperimental(sa)) { addr->flags |= FCTX_ADDRINFO_MARK; msg = "ignoring experimental address: "; } else if (sa->type.sa.sa_family != AF_INET6) { return; } else if (IN6_IS_ADDR_V4MAPPED(&sa->type.sin6.sin6_addr)) { addr->flags |= FCTX_ADDRINFO_MARK; msg = "ignoring IPv6 mapped IPV4 address: "; } else if (IN6_IS_ADDR_V4COMPAT(&sa->type.sin6.sin6_addr)) { addr->flags |= FCTX_ADDRINFO_MARK; msg = "ignoring IPv6 compatibility IPV4 address: "; } else return; if (!isc_log_wouldlog(dns_lctx, ISC_LOG_DEBUG(3))) return; isc_netaddr_fromsockaddr(&na, sa); isc_netaddr_format(&na, buf, sizeof(buf)); FCTXTRACE2(msg, buf); } static inline dns_adbaddrinfo_t * fctx_nextaddress(fetchctx_t *fctx) { dns_adbfind_t *find, *start; dns_adbaddrinfo_t *addrinfo; dns_adbaddrinfo_t *faddrinfo; /* * Return the next untried address, if any. */ /* * Find the first unmarked forwarder (if any). */ for (addrinfo = ISC_LIST_HEAD(fctx->forwaddrs); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (!UNMARKED(addrinfo)) continue; possibly_mark(fctx, addrinfo); if (UNMARKED(addrinfo)) { addrinfo->flags |= FCTX_ADDRINFO_MARK; fctx->find = NULL; return (addrinfo); } } /* * No forwarders. Move to the next find. */ fctx->attributes |= FCTX_ATTR_TRIEDFIND; find = fctx->find; if (find == NULL) find = ISC_LIST_HEAD(fctx->finds); else { find = ISC_LIST_NEXT(find, publink); if (find == NULL) find = ISC_LIST_HEAD(fctx->finds); } /* * Find the first unmarked addrinfo. */ addrinfo = NULL; if (find != NULL) { start = find; do { for (addrinfo = ISC_LIST_HEAD(find->list); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (!UNMARKED(addrinfo)) continue; possibly_mark(fctx, addrinfo); if (UNMARKED(addrinfo)) { addrinfo->flags |= FCTX_ADDRINFO_MARK; break; } } if (addrinfo != NULL) break; find = ISC_LIST_NEXT(find, publink); if (find == NULL) find = ISC_LIST_HEAD(fctx->finds); } while (find != start); } fctx->find = find; if (addrinfo != NULL) return (addrinfo); /* * No nameservers left. Try alternates. */ fctx->attributes |= FCTX_ATTR_TRIEDALT; find = fctx->altfind; if (find == NULL) find = ISC_LIST_HEAD(fctx->altfinds); else { find = ISC_LIST_NEXT(find, publink); if (find == NULL) find = ISC_LIST_HEAD(fctx->altfinds); } /* * Find the first unmarked addrinfo. */ addrinfo = NULL; if (find != NULL) { start = find; do { for (addrinfo = ISC_LIST_HEAD(find->list); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (!UNMARKED(addrinfo)) continue; possibly_mark(fctx, addrinfo); if (UNMARKED(addrinfo)) { addrinfo->flags |= FCTX_ADDRINFO_MARK; break; } } if (addrinfo != NULL) break; find = ISC_LIST_NEXT(find, publink); if (find == NULL) find = ISC_LIST_HEAD(fctx->altfinds); } while (find != start); } faddrinfo = addrinfo; /* * See if we have a better alternate server by address. */ for (addrinfo = ISC_LIST_HEAD(fctx->altaddrs); addrinfo != NULL; addrinfo = ISC_LIST_NEXT(addrinfo, publink)) { if (!UNMARKED(addrinfo)) continue; possibly_mark(fctx, addrinfo); if (UNMARKED(addrinfo) && (faddrinfo == NULL || addrinfo->srtt < faddrinfo->srtt)) { if (faddrinfo != NULL) faddrinfo->flags &= ~FCTX_ADDRINFO_MARK; addrinfo->flags |= FCTX_ADDRINFO_MARK; break; } } if (addrinfo == NULL) { addrinfo = faddrinfo; fctx->altfind = find; } return (addrinfo); } static void fctx_try(fetchctx_t *fctx, isc_boolean_t retrying, isc_boolean_t badcache) { isc_result_t result; dns_adbaddrinfo_t *addrinfo; FCTXTRACE("try"); REQUIRE(!ADDRWAIT(fctx)); if (fctx->totalqueries > DEFAULT_MAX_QUERIES) fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); addrinfo = fctx_nextaddress(fctx); if (addrinfo == NULL) { /* * We have no more addresses. Start over. */ fctx_cancelqueries(fctx, ISC_TRUE); fctx_cleanupfinds(fctx); fctx_cleanupaltfinds(fctx); fctx_cleanupforwaddrs(fctx); fctx_cleanupaltaddrs(fctx); result = fctx_getaddresses(fctx, badcache); if (result == DNS_R_WAIT) { /* * Sleep waiting for addresses. */ FCTXTRACE("addrwait"); fctx->attributes |= FCTX_ATTR_ADDRWAIT; return; } else if (result != ISC_R_SUCCESS) { /* * Something bad happened. */ fctx_done(fctx, result, __LINE__); return; } addrinfo = fctx_nextaddress(fctx); /* * While we may have addresses from the ADB, they * might be bad ones. In this case, return SERVFAIL. */ if (addrinfo == NULL) { fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); return; } } result = fctx_query(fctx, addrinfo, fctx->options); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else if (retrying) inc_stats(fctx->res, dns_resstatscounter_retry); } static isc_boolean_t fctx_unlink(fetchctx_t *fctx) { dns_resolver_t *res; unsigned int bucketnum; /* * Caller must be holding the bucket lock. */ REQUIRE(VALID_FCTX(fctx)); REQUIRE(fctx->state == fetchstate_done || fctx->state == fetchstate_init); REQUIRE(ISC_LIST_EMPTY(fctx->events)); REQUIRE(ISC_LIST_EMPTY(fctx->queries)); REQUIRE(ISC_LIST_EMPTY(fctx->finds)); REQUIRE(ISC_LIST_EMPTY(fctx->altfinds)); REQUIRE(fctx->pending == 0); REQUIRE(fctx->references == 0); REQUIRE(ISC_LIST_EMPTY(fctx->validators)); FCTXTRACE("unlink"); res = fctx->res; bucketnum = fctx->bucketnum; ISC_LIST_UNLINK(res->buckets[bucketnum].fctxs, fctx, link); LOCK(&res->nlock); res->nfctx--; UNLOCK(&res->nlock); if (res->buckets[bucketnum].exiting && ISC_LIST_EMPTY(res->buckets[bucketnum].fctxs)) return (ISC_TRUE); return (ISC_FALSE); } static void fctx_destroy(fetchctx_t *fctx) { isc_sockaddr_t *sa, *next_sa; REQUIRE(VALID_FCTX(fctx)); REQUIRE(fctx->state == fetchstate_done || fctx->state == fetchstate_init); REQUIRE(ISC_LIST_EMPTY(fctx->events)); REQUIRE(ISC_LIST_EMPTY(fctx->queries)); REQUIRE(ISC_LIST_EMPTY(fctx->finds)); REQUIRE(ISC_LIST_EMPTY(fctx->altfinds)); REQUIRE(fctx->pending == 0); REQUIRE(fctx->references == 0); REQUIRE(ISC_LIST_EMPTY(fctx->validators)); REQUIRE(!ISC_LINK_LINKED(fctx, link)); FCTXTRACE("destroy"); /* * Free bad. */ for (sa = ISC_LIST_HEAD(fctx->bad); sa != NULL; sa = next_sa) { next_sa = ISC_LIST_NEXT(sa, link); ISC_LIST_UNLINK(fctx->bad, sa, link); isc_mem_put(fctx->mctx, sa, sizeof(*sa)); } for (sa = ISC_LIST_HEAD(fctx->edns); sa != NULL; sa = next_sa) { next_sa = ISC_LIST_NEXT(sa, link); ISC_LIST_UNLINK(fctx->edns, sa, link); isc_mem_put(fctx->mctx, sa, sizeof(*sa)); } for (sa = ISC_LIST_HEAD(fctx->edns512); sa != NULL; sa = next_sa) { next_sa = ISC_LIST_NEXT(sa, link); ISC_LIST_UNLINK(fctx->edns512, sa, link); isc_mem_put(fctx->mctx, sa, sizeof(*sa)); } for (sa = ISC_LIST_HEAD(fctx->bad_edns); sa != NULL; sa = next_sa) { next_sa = ISC_LIST_NEXT(sa, link); ISC_LIST_UNLINK(fctx->bad_edns, sa, link); isc_mem_put(fctx->mctx, sa, sizeof(*sa)); } isc_timer_detach(&fctx->timer); dns_message_destroy(&fctx->rmessage); dns_message_destroy(&fctx->qmessage); if (dns_name_countlabels(&fctx->domain) > 0) dns_name_free(&fctx->domain, fctx->mctx); if (dns_rdataset_isassociated(&fctx->nameservers)) dns_rdataset_disassociate(&fctx->nameservers); dns_name_free(&fctx->name, fctx->mctx); dns_db_detach(&fctx->cache); dns_adb_detach(&fctx->adb); isc_mem_free(fctx->mctx, fctx->info); isc_mem_putanddetach(&fctx->mctx, fctx, sizeof(*fctx)); } /* * Fetch event handlers. */ static void fctx_timeout(isc_task_t *task, isc_event_t *event) { fetchctx_t *fctx = event->ev_arg; isc_timerevent_t *tevent = (isc_timerevent_t *)event; resquery_t *query; REQUIRE(VALID_FCTX(fctx)); UNUSED(task); FCTXTRACE("timeout"); inc_stats(fctx->res, dns_resstatscounter_querytimeout); if (event->ev_type == ISC_TIMEREVENT_LIFE) { fctx->reason = NULL; fctx_done(fctx, ISC_R_TIMEDOUT, __LINE__); } else { isc_result_t result; fctx->timeouts++; fctx->timeout = ISC_TRUE; /* * We could cancel the running queries here, or we could let * them keep going. Since we normally use separate sockets for * different queries, we adopt the former approach to reduce * the number of open sockets: cancel the oldest query if it * expired after the query had started (this is usually the * case but is not always so, depending on the task schedule * timing). */ query = ISC_LIST_HEAD(fctx->queries); if (query != NULL && isc_time_compare(&tevent->due, &query->start) >= 0) { fctx_cancelquery(&query, NULL, NULL, ISC_TRUE); } fctx->attributes &= ~FCTX_ATTR_ADDRWAIT; /* * Our timer has triggered. Reestablish the fctx lifetime * timer. */ result = fctx_starttimer(fctx); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else /* * Keep trying. */ fctx_try(fctx, ISC_TRUE, ISC_FALSE); } isc_event_free(&event); } static void fctx_shutdown(fetchctx_t *fctx) { isc_event_t *cevent; /* * Start the shutdown process for fctx, if it isn't already underway. */ FCTXTRACE("shutdown"); /* * The caller must be holding the appropriate bucket lock. */ if (fctx->want_shutdown) return; fctx->want_shutdown = ISC_TRUE; /* * Unless we're still initializing (in which case the * control event is still outstanding), we need to post * the control event to tell the fetch we want it to * exit. */ if (fctx->state != fetchstate_init) { cevent = &fctx->control_event; isc_task_send(fctx->res->buckets[fctx->bucketnum].task, &cevent); } } static void fctx_doshutdown(isc_task_t *task, isc_event_t *event) { fetchctx_t *fctx = event->ev_arg; isc_boolean_t bucket_empty = ISC_FALSE; dns_resolver_t *res; unsigned int bucketnum; dns_validator_t *validator; isc_boolean_t destroy = ISC_FALSE; REQUIRE(VALID_FCTX(fctx)); UNUSED(task); res = fctx->res; bucketnum = fctx->bucketnum; FCTXTRACE("doshutdown"); /* * An fctx that is shutting down is no longer in ADDRWAIT mode. */ fctx->attributes &= ~FCTX_ATTR_ADDRWAIT; /* * Cancel all pending validators. Note that this must be done * without the bucket lock held, since that could cause deadlock. */ validator = ISC_LIST_HEAD(fctx->validators); while (validator != NULL) { dns_validator_cancel(validator); validator = ISC_LIST_NEXT(validator, link); } if (fctx->nsfetch != NULL) dns_resolver_cancelfetch(fctx->nsfetch); /* * Shut down anything that is still running on behalf of this * fetch. To avoid deadlock with the ADB, we must do this * before we lock the bucket lock. */ fctx_stopeverything(fctx, ISC_FALSE); LOCK(&res->buckets[bucketnum].lock); fctx->attributes |= FCTX_ATTR_SHUTTINGDOWN; INSIST(fctx->state == fetchstate_active || fctx->state == fetchstate_done); INSIST(fctx->want_shutdown); if (fctx->state != fetchstate_done) { fctx->state = fetchstate_done; fctx_sendevents(fctx, ISC_R_CANCELED, __LINE__); } if (fctx->references == 0 && fctx->pending == 0 && fctx->nqueries == 0 && ISC_LIST_EMPTY(fctx->validators)) { bucket_empty = fctx_unlink(fctx); destroy = ISC_TRUE; } UNLOCK(&res->buckets[bucketnum].lock); if (destroy) { fctx_destroy(fctx); if (bucket_empty) empty_bucket(res); } } static void fctx_start(isc_task_t *task, isc_event_t *event) { fetchctx_t *fctx = event->ev_arg; isc_boolean_t done = ISC_FALSE, bucket_empty = ISC_FALSE; dns_resolver_t *res; unsigned int bucketnum; isc_boolean_t destroy = ISC_FALSE; REQUIRE(VALID_FCTX(fctx)); UNUSED(task); res = fctx->res; bucketnum = fctx->bucketnum; FCTXTRACE("start"); LOCK(&res->buckets[bucketnum].lock); INSIST(fctx->state == fetchstate_init); if (fctx->want_shutdown) { /* * We haven't started this fctx yet, and we've been requested * to shut it down. */ fctx->attributes |= FCTX_ATTR_SHUTTINGDOWN; fctx->state = fetchstate_done; fctx_sendevents(fctx, ISC_R_CANCELED, __LINE__); /* * Since we haven't started, we INSIST that we have no * pending ADB finds and no pending validations. */ INSIST(fctx->pending == 0); INSIST(fctx->nqueries == 0); INSIST(ISC_LIST_EMPTY(fctx->validators)); if (fctx->references == 0) { /* * It's now safe to destroy this fctx. */ bucket_empty = fctx_unlink(fctx); destroy = ISC_TRUE; } done = ISC_TRUE; } else { /* * Normal fctx startup. */ fctx->state = fetchstate_active; fctx->totalqueries = 0; /* * Reset the control event for later use in shutting down * the fctx. */ ISC_EVENT_INIT(event, sizeof(*event), 0, NULL, DNS_EVENT_FETCHCONTROL, fctx_doshutdown, fctx, NULL, NULL, NULL); } UNLOCK(&res->buckets[bucketnum].lock); if (!done) { isc_result_t result; INSIST(!destroy); /* * All is well. Start working on the fetch. */ result = fctx_starttimer(fctx); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else fctx_try(fctx, ISC_FALSE, ISC_FALSE); } else if (destroy) { fctx_destroy(fctx); if (bucket_empty) empty_bucket(res); } } /* * Fetch Creation, Joining, and Cancelation. */ static inline isc_result_t fctx_join(fetchctx_t *fctx, isc_task_t *task, isc_sockaddr_t *client, dns_messageid_t id, isc_taskaction_t action, void *arg, dns_rdataset_t *rdataset, dns_rdataset_t *sigrdataset, dns_fetch_t *fetch) { isc_task_t *clone; dns_fetchevent_t *event; FCTXTRACE("join"); /* * We store the task we're going to send this event to in the * sender field. We'll make the fetch the sender when we actually * send the event. */ clone = NULL; isc_task_attach(task, &clone); event = (dns_fetchevent_t *) isc_event_allocate(fctx->res->mctx, clone, DNS_EVENT_FETCHDONE, action, arg, sizeof(*event)); if (event == NULL) { isc_task_detach(&clone); return (ISC_R_NOMEMORY); } event->result = DNS_R_SERVFAIL; event->qtype = fctx->type; event->db = NULL; event->node = NULL; event->rdataset = rdataset; event->sigrdataset = sigrdataset; event->fetch = fetch; event->client = client; event->id = id; event->qtotal = 0; dns_fixedname_init(&event->foundname); /* * Make sure that we can store the sigrdataset in the * first event if it is needed by any of the events. */ if (event->sigrdataset != NULL) ISC_LIST_PREPEND(fctx->events, event, ev_link); else ISC_LIST_APPEND(fctx->events, event, ev_link); fctx->references++; fctx->client = client; fetch->magic = DNS_FETCH_MAGIC; fetch->private = fctx; return (ISC_R_SUCCESS); } static inline void log_ns_ttl(fetchctx_t *fctx, const char *where) { char namebuf[DNS_NAME_FORMATSIZE]; char domainbuf[DNS_NAME_FORMATSIZE]; dns_name_format(&fctx->name, namebuf, sizeof(namebuf)); dns_name_format(&fctx->domain, domainbuf, sizeof(domainbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_DEBUG(10), "log_ns_ttl: fctx %p: %s: %s (in '%s'?): %u %u", fctx, where, namebuf, domainbuf, fctx->ns_ttl_ok, fctx->ns_ttl); } static isc_result_t fctx_create(dns_resolver_t *res, dns_name_t *name, dns_rdatatype_t type, dns_name_t *domain, dns_rdataset_t *nameservers, unsigned int options, unsigned int bucketnum, unsigned int depth, fetchctx_t **fctxp) { fetchctx_t *fctx; isc_result_t result; isc_result_t iresult; isc_interval_t interval; dns_fixedname_t fixed; unsigned int findoptions = 0; char buf[DNS_NAME_FORMATSIZE + DNS_RDATATYPE_FORMATSIZE]; char typebuf[DNS_RDATATYPE_FORMATSIZE]; dns_name_t suffix; isc_mem_t *mctx; /* * Caller must be holding the lock for bucket number 'bucketnum'. */ REQUIRE(fctxp != NULL && *fctxp == NULL); mctx = res->buckets[bucketnum].mctx; fctx = isc_mem_get(mctx, sizeof(*fctx)); if (fctx == NULL) return (ISC_R_NOMEMORY); dns_name_format(name, buf, sizeof(buf)); dns_rdatatype_format(type, typebuf, sizeof(typebuf)); strcat(buf, "/"); /* checked */ strcat(buf, typebuf); /* checked */ fctx->info = isc_mem_strdup(mctx, buf); if (fctx->info == NULL) { result = ISC_R_NOMEMORY; goto cleanup_fetch; } FCTXTRACE("create"); dns_name_init(&fctx->name, NULL); result = dns_name_dup(name, mctx, &fctx->name); if (result != ISC_R_SUCCESS) goto cleanup_info; dns_name_init(&fctx->domain, NULL); dns_rdataset_init(&fctx->nameservers); fctx->type = type; fctx->options = options; /* * Note! We do not attach to the task. We are relying on the * resolver to ensure that this task doesn't go away while we are * using it. */ fctx->res = res; fctx->references = 0; fctx->bucketnum = bucketnum; fctx->state = fetchstate_init; fctx->want_shutdown = ISC_FALSE; fctx->cloned = ISC_FALSE; fctx->depth = depth; ISC_LIST_INIT(fctx->queries); ISC_LIST_INIT(fctx->finds); ISC_LIST_INIT(fctx->altfinds); ISC_LIST_INIT(fctx->forwaddrs); ISC_LIST_INIT(fctx->altaddrs); ISC_LIST_INIT(fctx->forwarders); fctx->fwdpolicy = dns_fwdpolicy_none; ISC_LIST_INIT(fctx->bad); ISC_LIST_INIT(fctx->edns); ISC_LIST_INIT(fctx->edns512); ISC_LIST_INIT(fctx->bad_edns); ISC_LIST_INIT(fctx->validators); fctx->validator = NULL; fctx->find = NULL; fctx->altfind = NULL; fctx->pending = 0; fctx->restarts = 0; fctx->querysent = 0; fctx->totalqueries = 0; fctx->referrals = 0; TIME_NOW(&fctx->start); fctx->timeouts = 0; fctx->lamecount = 0; fctx->adberr = 0; fctx->neterr = 0; fctx->badresp = 0; fctx->findfail = 0; fctx->valfail = 0; fctx->result = ISC_R_FAILURE; fctx->vresult = ISC_R_SUCCESS; fctx->exitline = -1; /* sentinel */ fctx->logged = ISC_FALSE; fctx->attributes = 0; fctx->spilled = ISC_FALSE; fctx->nqueries = 0; fctx->reason = NULL; fctx->rand_buf = 0; fctx->rand_bits = 0; fctx->timeout = ISC_FALSE; fctx->addrinfo = NULL; fctx->client = NULL; fctx->ns_ttl = 0; fctx->ns_ttl_ok = ISC_FALSE; dns_name_init(&fctx->nsname, NULL); fctx->nsfetch = NULL; dns_rdataset_init(&fctx->nsrrset); if (domain == NULL) { dns_forwarders_t *forwarders = NULL; unsigned int labels; dns_name_t *fwdname = name; /* * DS records are found in the parent server. * Strip label to get the correct forwarder (if any). */ if (dns_rdatatype_atparent(fctx->type) && dns_name_countlabels(name) > 1) { dns_name_init(&suffix, NULL); labels = dns_name_countlabels(name); dns_name_getlabelsequence(name, 1, labels - 1, &suffix); fwdname = &suffix; } dns_fixedname_init(&fixed); domain = dns_fixedname_name(&fixed); result = dns_fwdtable_find2(fctx->res->view->fwdtable, fwdname, domain, &forwarders); if (result == ISC_R_SUCCESS) fctx->fwdpolicy = forwarders->fwdpolicy; if (fctx->fwdpolicy != dns_fwdpolicy_only) { /* * The caller didn't supply a query domain and * nameservers, and we're not in forward-only mode, * so find the best nameservers to use. */ if (dns_rdatatype_atparent(fctx->type)) findoptions |= DNS_DBFIND_NOEXACT; result = dns_view_findzonecut(res->view, fwdname, domain, 0, findoptions, ISC_TRUE, &fctx->nameservers, NULL); if (result != ISC_R_SUCCESS) goto cleanup_name; result = dns_name_dup(domain, mctx, &fctx->domain); if (result != ISC_R_SUCCESS) { dns_rdataset_disassociate(&fctx->nameservers); goto cleanup_name; } fctx->ns_ttl = fctx->nameservers.ttl; fctx->ns_ttl_ok = ISC_TRUE; } else { /* * We're in forward-only mode. Set the query domain. */ result = dns_name_dup(domain, mctx, &fctx->domain); if (result != ISC_R_SUCCESS) goto cleanup_name; } } else { result = dns_name_dup(domain, mctx, &fctx->domain); if (result != ISC_R_SUCCESS) goto cleanup_name; dns_rdataset_clone(nameservers, &fctx->nameservers); fctx->ns_ttl = fctx->nameservers.ttl; fctx->ns_ttl_ok = ISC_TRUE; } log_ns_ttl(fctx, "fctx_create"); INSIST(dns_name_issubdomain(&fctx->name, &fctx->domain)); fctx->qmessage = NULL; result = dns_message_create(mctx, DNS_MESSAGE_INTENTRENDER, &fctx->qmessage); if (result != ISC_R_SUCCESS) goto cleanup_domain; fctx->rmessage = NULL; result = dns_message_create(mctx, DNS_MESSAGE_INTENTPARSE, &fctx->rmessage); if (result != ISC_R_SUCCESS) goto cleanup_qmessage; /* * Compute an expiration time for the entire fetch. */ isc_interval_set(&interval, res->query_timeout, 0); iresult = isc_time_nowplusinterval(&fctx->expires, &interval); if (iresult != ISC_R_SUCCESS) { UNEXPECTED_ERROR(__FILE__, __LINE__, "isc_time_nowplusinterval: %s", isc_result_totext(iresult)); result = ISC_R_UNEXPECTED; goto cleanup_rmessage; } /* * Default retry interval initialization. We set the interval now * mostly so it won't be uninitialized. It will be set to the * correct value before a query is issued. */ isc_interval_set(&fctx->interval, 2, 0); /* * Create an inactive timer. It will be made active when the fetch * is actually started. */ fctx->timer = NULL; iresult = isc_timer_create(res->timermgr, isc_timertype_inactive, NULL, NULL, res->buckets[bucketnum].task, fctx_timeout, fctx, &fctx->timer); if (iresult != ISC_R_SUCCESS) { UNEXPECTED_ERROR(__FILE__, __LINE__, "isc_timer_create: %s", isc_result_totext(iresult)); result = ISC_R_UNEXPECTED; goto cleanup_rmessage; } /* * Attach to the view's cache and adb. */ fctx->cache = NULL; dns_db_attach(res->view->cachedb, &fctx->cache); fctx->adb = NULL; dns_adb_attach(res->view->adb, &fctx->adb); fctx->mctx = NULL; isc_mem_attach(mctx, &fctx->mctx); ISC_LIST_INIT(fctx->events); ISC_LINK_INIT(fctx, link); fctx->magic = FCTX_MAGIC; ISC_LIST_APPEND(res->buckets[bucketnum].fctxs, fctx, link); LOCK(&res->nlock); res->nfctx++; UNLOCK(&res->nlock); *fctxp = fctx; return (ISC_R_SUCCESS); cleanup_rmessage: dns_message_destroy(&fctx->rmessage); cleanup_qmessage: dns_message_destroy(&fctx->qmessage); cleanup_domain: if (dns_name_countlabels(&fctx->domain) > 0) dns_name_free(&fctx->domain, mctx); if (dns_rdataset_isassociated(&fctx->nameservers)) dns_rdataset_disassociate(&fctx->nameservers); cleanup_name: dns_name_free(&fctx->name, mctx); cleanup_info: isc_mem_free(mctx, fctx->info); cleanup_fetch: isc_mem_put(mctx, fctx, sizeof(*fctx)); return (result); } /* * Handle Responses */ static inline isc_boolean_t is_lame(fetchctx_t *fctx) { dns_message_t *message = fctx->rmessage; dns_name_t *name; dns_rdataset_t *rdataset; isc_result_t result; if (message->rcode != dns_rcode_noerror && message->rcode != dns_rcode_nxdomain) return (ISC_FALSE); if (message->counts[DNS_SECTION_ANSWER] != 0) return (ISC_FALSE); if (message->counts[DNS_SECTION_AUTHORITY] == 0) return (ISC_FALSE); result = dns_message_firstname(message, DNS_SECTION_AUTHORITY); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(message, DNS_SECTION_AUTHORITY, &name); for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { dns_namereln_t namereln; int order; unsigned int labels; if (rdataset->type != dns_rdatatype_ns) continue; namereln = dns_name_fullcompare(name, &fctx->domain, &order, &labels); if (namereln == dns_namereln_equal && (message->flags & DNS_MESSAGEFLAG_AA) != 0) return (ISC_FALSE); if (namereln == dns_namereln_subdomain) return (ISC_FALSE); return (ISC_TRUE); } result = dns_message_nextname(message, DNS_SECTION_AUTHORITY); } return (ISC_FALSE); } static inline void log_lame(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo) { char namebuf[DNS_NAME_FORMATSIZE]; char domainbuf[DNS_NAME_FORMATSIZE]; char addrbuf[ISC_SOCKADDR_FORMATSIZE]; dns_name_format(&fctx->name, namebuf, sizeof(namebuf)); dns_name_format(&fctx->domain, domainbuf, sizeof(domainbuf)); isc_sockaddr_format(&addrinfo->sockaddr, addrbuf, sizeof(addrbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_LAME_SERVERS, DNS_LOGMODULE_RESOLVER, ISC_LOG_INFO, "lame server resolving '%s' (in '%s'?): %s", namebuf, domainbuf, addrbuf); } static inline void log_formerr(fetchctx_t *fctx, const char *format, ...) { char nsbuf[ISC_SOCKADDR_FORMATSIZE]; char clbuf[ISC_SOCKADDR_FORMATSIZE]; const char *clmsg = ""; char msgbuf[2048]; va_list args; va_start(args, format); vsnprintf(msgbuf, sizeof(msgbuf), format, args); va_end(args); isc_sockaddr_format(&fctx->addrinfo->sockaddr, nsbuf, sizeof(nsbuf)); if (fctx->client != NULL) { clmsg = " for client "; isc_sockaddr_format(fctx->client, clbuf, sizeof(clbuf)); } else { clbuf[0] = '\0'; } isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "DNS format error from %s resolving %s%s%s: %s", nsbuf, fctx->info, clmsg, clbuf, msgbuf); } static inline isc_result_t same_question(fetchctx_t *fctx) { isc_result_t result; dns_message_t *message = fctx->rmessage; dns_name_t *name; dns_rdataset_t *rdataset; /* * Caller must be holding the fctx lock. */ /* * XXXRTH Currently we support only one question. */ if (message->counts[DNS_SECTION_QUESTION] != 1) { log_formerr(fctx, "too many questions"); return (DNS_R_FORMERR); } result = dns_message_firstname(message, DNS_SECTION_QUESTION); if (result != ISC_R_SUCCESS) return (result); name = NULL; dns_message_currentname(message, DNS_SECTION_QUESTION, &name); rdataset = ISC_LIST_HEAD(name->list); INSIST(rdataset != NULL); INSIST(ISC_LIST_NEXT(rdataset, link) == NULL); if (fctx->type != rdataset->type || fctx->res->rdclass != rdataset->rdclass || !dns_name_equal(&fctx->name, name)) { char namebuf[DNS_NAME_FORMATSIZE]; char class[DNS_RDATACLASS_FORMATSIZE]; char type[DNS_RDATATYPE_FORMATSIZE]; dns_name_format(name, namebuf, sizeof(namebuf)); dns_rdataclass_format(rdataset->rdclass, class, sizeof(class)); dns_rdatatype_format(rdataset->type, type, sizeof(type)); log_formerr(fctx, "question section mismatch: got %s/%s/%s", namebuf, class, type); return (DNS_R_FORMERR); } return (ISC_R_SUCCESS); } static void clone_results(fetchctx_t *fctx) { dns_fetchevent_t *event, *hevent; isc_result_t result; dns_name_t *name, *hname; FCTXTRACE("clone_results"); /* * Set up any other events to have the same data as the first * event. * * Caller must be holding the appropriate lock. */ fctx->cloned = ISC_TRUE; hevent = ISC_LIST_HEAD(fctx->events); if (hevent == NULL) return; hname = dns_fixedname_name(&hevent->foundname); for (event = ISC_LIST_NEXT(hevent, ev_link); event != NULL; event = ISC_LIST_NEXT(event, ev_link)) { name = dns_fixedname_name(&event->foundname); result = dns_name_copy(hname, name, NULL); if (result != ISC_R_SUCCESS) event->result = result; else event->result = hevent->result; dns_db_attach(hevent->db, &event->db); dns_db_attachnode(hevent->db, hevent->node, &event->node); INSIST(hevent->rdataset != NULL); INSIST(event->rdataset != NULL); if (dns_rdataset_isassociated(hevent->rdataset)) dns_rdataset_clone(hevent->rdataset, event->rdataset); INSIST(! (hevent->sigrdataset == NULL && event->sigrdataset != NULL)); if (hevent->sigrdataset != NULL && dns_rdataset_isassociated(hevent->sigrdataset) && event->sigrdataset != NULL) dns_rdataset_clone(hevent->sigrdataset, event->sigrdataset); } } #define CACHE(r) (((r)->attributes & DNS_RDATASETATTR_CACHE) != 0) #define ANSWER(r) (((r)->attributes & DNS_RDATASETATTR_ANSWER) != 0) #define ANSWERSIG(r) (((r)->attributes & DNS_RDATASETATTR_ANSWERSIG) != 0) #define EXTERNAL(r) (((r)->attributes & DNS_RDATASETATTR_EXTERNAL) != 0) #define CHAINING(r) (((r)->attributes & DNS_RDATASETATTR_CHAINING) != 0) #define CHASE(r) (((r)->attributes & DNS_RDATASETATTR_CHASE) != 0) #define CHECKNAMES(r) (((r)->attributes & DNS_RDATASETATTR_CHECKNAMES) != 0) /* * Destroy '*fctx' if it is ready to be destroyed (i.e., if it has * no references and is no longer waiting for any events). * * Requires: * '*fctx' is shutting down. * * Returns: * true if the resolver is exiting and this is the last fctx in the bucket. */ static isc_boolean_t maybe_destroy(fetchctx_t *fctx, isc_boolean_t locked) { unsigned int bucketnum; isc_boolean_t bucket_empty = ISC_FALSE; dns_resolver_t *res = fctx->res; dns_validator_t *validator, *next_validator; isc_boolean_t destroy = ISC_FALSE; REQUIRE(SHUTTINGDOWN(fctx)); bucketnum = fctx->bucketnum; if (!locked) LOCK(&res->buckets[bucketnum].lock); if (fctx->pending != 0 || fctx->nqueries != 0) goto unlock; for (validator = ISC_LIST_HEAD(fctx->validators); validator != NULL; validator = next_validator) { next_validator = ISC_LIST_NEXT(validator, link); dns_validator_cancel(validator); } if (fctx->references == 0 && ISC_LIST_EMPTY(fctx->validators)) { bucket_empty = fctx_unlink(fctx); destroy = ISC_TRUE; } unlock: if (!locked) UNLOCK(&res->buckets[bucketnum].lock); if (destroy) fctx_destroy(fctx); return (bucket_empty); } /* * The validator has finished. */ static void validated(isc_task_t *task, isc_event_t *event) { dns_adbaddrinfo_t *addrinfo; dns_dbnode_t *node = NULL; dns_dbnode_t *nsnode = NULL; dns_fetchevent_t *hevent; dns_name_t *name; dns_rdataset_t *ardataset = NULL; dns_rdataset_t *asigrdataset = NULL; dns_rdataset_t *rdataset; dns_rdataset_t *sigrdataset; dns_resolver_t *res; dns_valarg_t *valarg; dns_validatorevent_t *vevent; fetchctx_t *fctx; isc_boolean_t chaining; isc_boolean_t negative; isc_boolean_t sentresponse; isc_result_t eresult = ISC_R_SUCCESS; isc_result_t result = ISC_R_SUCCESS; isc_stdtime_t now; isc_uint32_t ttl; UNUSED(task); /* for now */ REQUIRE(event->ev_type == DNS_EVENT_VALIDATORDONE); valarg = event->ev_arg; fctx = valarg->fctx; res = fctx->res; addrinfo = valarg->addrinfo; REQUIRE(VALID_FCTX(fctx)); REQUIRE(!ISC_LIST_EMPTY(fctx->validators)); vevent = (dns_validatorevent_t *)event; fctx->vresult = vevent->result; FCTXTRACE("received validation completion event"); LOCK(&res->buckets[fctx->bucketnum].lock); ISC_LIST_UNLINK(fctx->validators, vevent->validator, link); fctx->validator = NULL; /* * Destroy the validator early so that we can * destroy the fctx if necessary. */ dns_validator_destroy(&vevent->validator); isc_mem_put(fctx->mctx, valarg, sizeof(*valarg)); negative = ISC_TF(vevent->rdataset == NULL); sentresponse = ISC_TF((fctx->options & DNS_FETCHOPT_NOVALIDATE) != 0); /* * If shutting down, ignore the results. Check to see if we're * done waiting for validator completions and ADB pending events; if * so, destroy the fctx. */ if (SHUTTINGDOWN(fctx) && !sentresponse) { isc_uint32_t bucketnum = fctx->bucketnum; isc_boolean_t bucket_empty; bucket_empty = maybe_destroy(fctx, ISC_TRUE); UNLOCK(&res->buckets[bucketnum].lock); if (bucket_empty) empty_bucket(res); goto cleanup_event; } isc_stdtime_get(&now); /* * If chaining, we need to make sure that the right result code is * returned, and that the rdatasets are bound. */ if (vevent->result == ISC_R_SUCCESS && !negative && vevent->rdataset != NULL && CHAINING(vevent->rdataset)) { if (vevent->rdataset->type == dns_rdatatype_cname) eresult = DNS_R_CNAME; else { INSIST(vevent->rdataset->type == dns_rdatatype_dname); eresult = DNS_R_DNAME; } chaining = ISC_TRUE; } else chaining = ISC_FALSE; /* * Either we're not shutting down, or we are shutting down but want * to cache the result anyway (if this was a validation started by * a query with cd set) */ hevent = ISC_LIST_HEAD(fctx->events); if (hevent != NULL) { if (!negative && !chaining && (fctx->type == dns_rdatatype_any || fctx->type == dns_rdatatype_rrsig || fctx->type == dns_rdatatype_sig)) { /* * Don't bind rdatasets; the caller * will iterate the node. */ } else { ardataset = hevent->rdataset; asigrdataset = hevent->sigrdataset; } } if (vevent->result != ISC_R_SUCCESS) { FCTXTRACE("validation failed"); inc_stats(res, dns_resstatscounter_valfail); fctx->valfail++; fctx->vresult = vevent->result; if (fctx->vresult != DNS_R_BROKENCHAIN) { result = ISC_R_NOTFOUND; if (vevent->rdataset != NULL) result = dns_db_findnode(fctx->cache, vevent->name, ISC_TRUE, &node); if (result == ISC_R_SUCCESS) (void)dns_db_deleterdataset(fctx->cache, node, NULL, vevent->type, 0); if (result == ISC_R_SUCCESS && vevent->sigrdataset != NULL) (void)dns_db_deleterdataset(fctx->cache, node, NULL, dns_rdatatype_rrsig, vevent->type); if (result == ISC_R_SUCCESS) dns_db_detachnode(fctx->cache, &node); } if (fctx->vresult == DNS_R_BROKENCHAIN && !negative) { /* * Cache the data as pending for later validation. */ result = ISC_R_NOTFOUND; if (vevent->rdataset != NULL) result = dns_db_findnode(fctx->cache, vevent->name, ISC_TRUE, &node); if (result == ISC_R_SUCCESS) { (void)dns_db_addrdataset(fctx->cache, node, NULL, now, vevent->rdataset, 0, NULL); } if (result == ISC_R_SUCCESS && vevent->sigrdataset != NULL) (void)dns_db_addrdataset(fctx->cache, node, NULL, now, vevent->sigrdataset, 0, NULL); if (result == ISC_R_SUCCESS) dns_db_detachnode(fctx->cache, &node); } result = fctx->vresult; add_bad(fctx, addrinfo, result, badns_validation); isc_event_free(&event); UNLOCK(&res->buckets[fctx->bucketnum].lock); INSIST(fctx->validator == NULL); fctx->validator = ISC_LIST_HEAD(fctx->validators); if (fctx->validator != NULL) dns_validator_send(fctx->validator); else if (sentresponse) fctx_done(fctx, result, __LINE__); /* Locks bucket. */ else if (result == DNS_R_BROKENCHAIN) { isc_result_t tresult; isc_time_t expire; isc_interval_t i; isc_interval_set(&i, DNS_BADCACHE_TTL(fctx), 0); tresult = isc_time_nowplusinterval(&expire, &i); if (negative && (fctx->type == dns_rdatatype_dnskey || fctx->type == dns_rdatatype_dlv || fctx->type == dns_rdatatype_ds) && tresult == ISC_R_SUCCESS) dns_resolver_addbadcache(res, &fctx->name, fctx->type, &expire); fctx_done(fctx, result, __LINE__); /* Locks bucket. */ } else fctx_try(fctx, ISC_TRUE, ISC_TRUE); /* Locks bucket. */ return; } if (negative) { dns_rdatatype_t covers; FCTXTRACE("nonexistence validation OK"); inc_stats(res, dns_resstatscounter_valnegsuccess); if (fctx->rmessage->rcode == dns_rcode_nxdomain) covers = dns_rdatatype_any; else covers = fctx->type; result = dns_db_findnode(fctx->cache, vevent->name, ISC_TRUE, &node); if (result != ISC_R_SUCCESS) goto noanswer_response; /* * If we are asking for a SOA record set the cache time * to zero to facilitate locating the containing zone of * a arbitrary zone. */ ttl = res->view->maxncachettl; if (fctx->type == dns_rdatatype_soa && covers == dns_rdatatype_any && res->zero_no_soa_ttl) ttl = 0; result = ncache_adderesult(fctx->rmessage, fctx->cache, node, covers, now, ttl, vevent->optout, vevent->secure, ardataset, &eresult); if (result != ISC_R_SUCCESS) goto noanswer_response; goto answer_response; } else inc_stats(res, dns_resstatscounter_valsuccess); FCTXTRACE("validation OK"); if (vevent->proofs[DNS_VALIDATOR_NOQNAMEPROOF] != NULL) { result = dns_rdataset_addnoqname(vevent->rdataset, vevent->proofs[DNS_VALIDATOR_NOQNAMEPROOF]); RUNTIME_CHECK(result == ISC_R_SUCCESS); INSIST(vevent->sigrdataset != NULL); vevent->sigrdataset->ttl = vevent->rdataset->ttl; if (vevent->proofs[DNS_VALIDATOR_CLOSESTENCLOSER] != NULL) { result = dns_rdataset_addclosest(vevent->rdataset, vevent->proofs[DNS_VALIDATOR_CLOSESTENCLOSER]); RUNTIME_CHECK(result == ISC_R_SUCCESS); } } else if (vevent->rdataset->trust == dns_trust_answer && vevent->rdataset->type != dns_rdatatype_rrsig) { isc_result_t tresult; dns_name_t *noqname = NULL; tresult = findnoqname(fctx, vevent->name, vevent->rdataset->type, &noqname); if (tresult == ISC_R_SUCCESS && noqname != NULL) { tresult = dns_rdataset_addnoqname(vevent->rdataset, noqname); RUNTIME_CHECK(tresult == ISC_R_SUCCESS); } } /* * The data was already cached as pending data. * Re-cache it as secure and bind the cached * rdatasets to the first event on the fetch * event list. */ result = dns_db_findnode(fctx->cache, vevent->name, ISC_TRUE, &node); if (result != ISC_R_SUCCESS) goto noanswer_response; result = dns_db_addrdataset(fctx->cache, node, NULL, now, vevent->rdataset, 0, ardataset); if (result != ISC_R_SUCCESS && result != DNS_R_UNCHANGED) goto noanswer_response; if (ardataset != NULL && NEGATIVE(ardataset)) { if (NXDOMAIN(ardataset)) eresult = DNS_R_NCACHENXDOMAIN; else eresult = DNS_R_NCACHENXRRSET; } else if (vevent->sigrdataset != NULL) { result = dns_db_addrdataset(fctx->cache, node, NULL, now, vevent->sigrdataset, 0, asigrdataset); if (result != ISC_R_SUCCESS && result != DNS_R_UNCHANGED) goto noanswer_response; } if (sentresponse) { isc_boolean_t bucket_empty = ISC_FALSE; /* * If we only deferred the destroy because we wanted to cache * the data, destroy now. */ dns_db_detachnode(fctx->cache, &node); if (SHUTTINGDOWN(fctx)) bucket_empty = maybe_destroy(fctx, ISC_TRUE); UNLOCK(&res->buckets[fctx->bucketnum].lock); if (bucket_empty) empty_bucket(res); goto cleanup_event; } if (!ISC_LIST_EMPTY(fctx->validators)) { INSIST(!negative); INSIST(fctx->type == dns_rdatatype_any || fctx->type == dns_rdatatype_rrsig || fctx->type == dns_rdatatype_sig); /* * Don't send a response yet - we have * more rdatasets that still need to * be validated. */ dns_db_detachnode(fctx->cache, &node); UNLOCK(&res->buckets[fctx->bucketnum].lock); dns_validator_send(ISC_LIST_HEAD(fctx->validators)); goto cleanup_event; } answer_response: /* * Cache any NS/NSEC records that happened to be validated. */ result = dns_message_firstname(fctx->rmessage, DNS_SECTION_AUTHORITY); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(fctx->rmessage, DNS_SECTION_AUTHORITY, &name); for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { if ((rdataset->type != dns_rdatatype_ns && rdataset->type != dns_rdatatype_nsec) || rdataset->trust != dns_trust_secure) continue; for (sigrdataset = ISC_LIST_HEAD(name->list); sigrdataset != NULL; sigrdataset = ISC_LIST_NEXT(sigrdataset, link)) { if (sigrdataset->type != dns_rdatatype_rrsig || sigrdataset->covers != rdataset->type) continue; break; } if (sigrdataset == NULL || sigrdataset->trust != dns_trust_secure) continue; result = dns_db_findnode(fctx->cache, name, ISC_TRUE, &nsnode); if (result != ISC_R_SUCCESS) continue; result = dns_db_addrdataset(fctx->cache, nsnode, NULL, now, rdataset, 0, NULL); if (result == ISC_R_SUCCESS) result = dns_db_addrdataset(fctx->cache, nsnode, NULL, now, sigrdataset, 0, NULL); dns_db_detachnode(fctx->cache, &nsnode); if (result != ISC_R_SUCCESS) continue; } result = dns_message_nextname(fctx->rmessage, DNS_SECTION_AUTHORITY); } result = ISC_R_SUCCESS; /* * Respond with an answer, positive or negative, * as opposed to an error. 'node' must be non-NULL. */ fctx->attributes |= FCTX_ATTR_HAVEANSWER; if (hevent != NULL) { /* * Negative results must be indicated in event->result. */ if (dns_rdataset_isassociated(hevent->rdataset) && NEGATIVE(hevent->rdataset)) { INSIST(eresult == DNS_R_NCACHENXDOMAIN || eresult == DNS_R_NCACHENXRRSET); } hevent->result = eresult; RUNTIME_CHECK(dns_name_copy(vevent->name, dns_fixedname_name(&hevent->foundname), NULL) == ISC_R_SUCCESS); dns_db_attach(fctx->cache, &hevent->db); dns_db_transfernode(fctx->cache, &node, &hevent->node); clone_results(fctx); } noanswer_response: if (node != NULL) dns_db_detachnode(fctx->cache, &node); UNLOCK(&res->buckets[fctx->bucketnum].lock); fctx_done(fctx, result, __LINE__); /* Locks bucket. */ cleanup_event: INSIST(node == NULL); isc_event_free(&event); } static void fctx_log(void *arg, int level, const char *fmt, ...) { char msgbuf[2048]; va_list args; fetchctx_t *fctx = arg; va_start(args, fmt); vsnprintf(msgbuf, sizeof(msgbuf), fmt, args); va_end(args); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, level, "fctx %p(%s): %s", fctx, fctx->info, msgbuf); } static inline isc_result_t findnoqname(fetchctx_t *fctx, dns_name_t *name, dns_rdatatype_t type, dns_name_t **noqnamep) { dns_rdataset_t *nrdataset, *next, *sigrdataset; dns_rdata_rrsig_t rrsig; isc_result_t result; unsigned int labels; dns_section_t section; dns_name_t *zonename; dns_fixedname_t fzonename; dns_name_t *closest; dns_fixedname_t fclosest; dns_name_t *nearest; dns_fixedname_t fnearest; dns_rdatatype_t found = dns_rdatatype_none; dns_name_t *noqname = NULL; FCTXTRACE("findnoqname"); REQUIRE(noqnamep != NULL && *noqnamep == NULL); /* * Find the SIG for this rdataset, if we have it. */ for (sigrdataset = ISC_LIST_HEAD(name->list); sigrdataset != NULL; sigrdataset = ISC_LIST_NEXT(sigrdataset, link)) { if (sigrdataset->type == dns_rdatatype_rrsig && sigrdataset->covers == type) break; } if (sigrdataset == NULL) return (ISC_R_NOTFOUND); labels = dns_name_countlabels(name); for (result = dns_rdataset_first(sigrdataset); result == ISC_R_SUCCESS; result = dns_rdataset_next(sigrdataset)) { dns_rdata_t rdata = DNS_RDATA_INIT; dns_rdataset_current(sigrdataset, &rdata); result = dns_rdata_tostruct(&rdata, &rrsig, NULL); RUNTIME_CHECK(result == ISC_R_SUCCESS); /* Wildcard has rrsig.labels < labels - 1. */ if (rrsig.labels + 1U >= labels) continue; break; } if (result == ISC_R_NOMORE) return (ISC_R_NOTFOUND); if (result != ISC_R_SUCCESS) return (result); dns_fixedname_init(&fzonename); zonename = dns_fixedname_name(&fzonename); dns_fixedname_init(&fclosest); closest = dns_fixedname_name(&fclosest); dns_fixedname_init(&fnearest); nearest = dns_fixedname_name(&fnearest); #define NXND(x) ((x) == ISC_R_SUCCESS) section = DNS_SECTION_AUTHORITY; for (result = dns_message_firstname(fctx->rmessage, section); result == ISC_R_SUCCESS; result = dns_message_nextname(fctx->rmessage, section)) { dns_name_t *nsec = NULL; dns_message_currentname(fctx->rmessage, section, &nsec); for (nrdataset = ISC_LIST_HEAD(nsec->list); nrdataset != NULL; nrdataset = next) { isc_boolean_t data = ISC_FALSE, exists = ISC_FALSE; isc_boolean_t optout = ISC_FALSE, unknown = ISC_FALSE; isc_boolean_t setclosest = ISC_FALSE; isc_boolean_t setnearest = ISC_FALSE; next = ISC_LIST_NEXT(nrdataset, link); if (nrdataset->type != dns_rdatatype_nsec && nrdataset->type != dns_rdatatype_nsec3) continue; if (nrdataset->type == dns_rdatatype_nsec && NXND(dns_nsec_noexistnodata(type, name, nsec, nrdataset, &exists, &data, NULL, fctx_log, fctx))) { if (!exists) { noqname = nsec; found = dns_rdatatype_nsec; } } if (nrdataset->type == dns_rdatatype_nsec3 && NXND(dns_nsec3_noexistnodata(type, name, nsec, nrdataset, zonename, &exists, &data, &optout, &unknown, &setclosest, &setnearest, closest, nearest, fctx_log, fctx))) { if (!exists && setnearest) { noqname = nsec; found = dns_rdatatype_nsec3; } } } } if (result == ISC_R_NOMORE) result = ISC_R_SUCCESS; if (noqname != NULL) { for (sigrdataset = ISC_LIST_HEAD(noqname->list); sigrdataset != NULL; sigrdataset = ISC_LIST_NEXT(sigrdataset, link)) { if (sigrdataset->type == dns_rdatatype_rrsig && sigrdataset->covers == found) break; } if (sigrdataset != NULL) *noqnamep = noqname; } return (result); } static inline isc_result_t cache_name(fetchctx_t *fctx, dns_name_t *name, dns_adbaddrinfo_t *addrinfo, isc_stdtime_t now) { dns_rdataset_t *rdataset, *sigrdataset; dns_rdataset_t *addedrdataset, *ardataset, *asigrdataset; dns_rdataset_t *valrdataset = NULL, *valsigrdataset = NULL; dns_dbnode_t *node, **anodep; dns_db_t **adbp; dns_name_t *aname; dns_resolver_t *res; isc_boolean_t need_validation, secure_domain, have_answer; isc_result_t result, eresult; dns_fetchevent_t *event; unsigned int options; isc_task_t *task; isc_boolean_t fail; unsigned int valoptions = 0; /* * The appropriate bucket lock must be held. */ res = fctx->res; need_validation = ISC_FALSE; POST(need_validation); secure_domain = ISC_FALSE; have_answer = ISC_FALSE; eresult = ISC_R_SUCCESS; task = res->buckets[fctx->bucketnum].task; /* * Is DNSSEC validation required for this name? */ if (res->view->enablevalidation) { result = dns_view_issecuredomain(res->view, name, &secure_domain); if (result != ISC_R_SUCCESS) return (result); if (!secure_domain && res->view->dlv != NULL) { valoptions = DNS_VALIDATOR_DLV; secure_domain = ISC_TRUE; } } if ((fctx->options & DNS_FETCHOPT_NOVALIDATE) != 0) need_validation = ISC_FALSE; else need_validation = secure_domain; adbp = NULL; aname = NULL; anodep = NULL; ardataset = NULL; asigrdataset = NULL; event = NULL; if ((name->attributes & DNS_NAMEATTR_ANSWER) != 0 && !need_validation) { have_answer = ISC_TRUE; event = ISC_LIST_HEAD(fctx->events); if (event != NULL) { adbp = &event->db; aname = dns_fixedname_name(&event->foundname); result = dns_name_copy(name, aname, NULL); if (result != ISC_R_SUCCESS) return (result); anodep = &event->node; /* * If this is an ANY, SIG or RRSIG query, we're not * going to return any rdatasets, unless we encountered * a CNAME or DNAME as "the answer". In this case, * we're going to return DNS_R_CNAME or DNS_R_DNAME * and we must set up the rdatasets. */ if ((fctx->type != dns_rdatatype_any && fctx->type != dns_rdatatype_rrsig && fctx->type != dns_rdatatype_sig) || (name->attributes & DNS_NAMEATTR_CHAINING) != 0) { ardataset = event->rdataset; asigrdataset = event->sigrdataset; } } } /* * Find or create the cache node. */ node = NULL; result = dns_db_findnode(fctx->cache, name, ISC_TRUE, &node); if (result != ISC_R_SUCCESS) return (result); /* * Cache or validate each cacheable rdataset. */ fail = ISC_TF((fctx->res->options & DNS_RESOLVER_CHECKNAMESFAIL) != 0); for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { if (!CACHE(rdataset)) continue; if (CHECKNAMES(rdataset)) { char namebuf[DNS_NAME_FORMATSIZE]; char typebuf[DNS_RDATATYPE_FORMATSIZE]; char classbuf[DNS_RDATATYPE_FORMATSIZE]; dns_name_format(name, namebuf, sizeof(namebuf)); dns_rdatatype_format(rdataset->type, typebuf, sizeof(typebuf)); dns_rdataclass_format(rdataset->rdclass, classbuf, sizeof(classbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "check-names %s %s/%s/%s", fail ? "failure" : "warning", namebuf, typebuf, classbuf); if (fail) { if (ANSWER(rdataset)) { dns_db_detachnode(fctx->cache, &node); return (DNS_R_BADNAME); } continue; } } /* * Enforce the configure maximum cache TTL. */ if (rdataset->ttl > res->view->maxcachettl) rdataset->ttl = res->view->maxcachettl; /* * Find the SIG for this rdataset, if we have it. */ for (sigrdataset = ISC_LIST_HEAD(name->list); sigrdataset != NULL; sigrdataset = ISC_LIST_NEXT(sigrdataset, link)) { if (sigrdataset->type == dns_rdatatype_rrsig && sigrdataset->covers == rdataset->type) break; } /* * If this RRset is in a secure domain, is in bailiwick, * and is not glue, attempt DNSSEC validation. (We do not * attempt to validate glue or out-of-bailiwick data--even * though there might be some performance benefit to doing * so--because it makes it simpler and safer to ensure that * records from a secure domain are only cached if validated * within the context of a query to the domain that owns * them.) */ if (secure_domain && rdataset->trust != dns_trust_glue && !EXTERNAL(rdataset)) { dns_trust_t trust; /* * RRSIGs are validated as part of validating the * type they cover. */ if (rdataset->type == dns_rdatatype_rrsig) continue; if (sigrdataset == NULL) { if (!ANSWER(rdataset) && need_validation) { /* * Ignore non-answer rdatasets that * are missing signatures. */ continue; } } /* * Normalize the rdataset and sigrdataset TTLs. */ if (sigrdataset != NULL) { rdataset->ttl = ISC_MIN(rdataset->ttl, sigrdataset->ttl); sigrdataset->ttl = rdataset->ttl; } /* * Cache this rdataset/sigrdataset pair as * pending data. Track whether it was additional * or not. */ if (rdataset->trust == dns_trust_additional) trust = dns_trust_pending_additional; else trust = dns_trust_pending_answer; rdataset->trust = trust; if (sigrdataset != NULL) sigrdataset->trust = trust; if (!need_validation || !ANSWER(rdataset)) { if (ANSWER(rdataset) && rdataset->type != dns_rdatatype_rrsig) { isc_result_t tresult; dns_name_t *noqname = NULL; tresult = findnoqname(fctx, name, rdataset->type, &noqname); if (tresult == ISC_R_SUCCESS && noqname != NULL) { tresult = dns_rdataset_addnoqname( rdataset, noqname); RUNTIME_CHECK(tresult == ISC_R_SUCCESS); } } addedrdataset = ardataset; result = dns_db_addrdataset(fctx->cache, node, NULL, now, rdataset, 0, addedrdataset); if (result == DNS_R_UNCHANGED) { result = ISC_R_SUCCESS; if (!need_validation && ardataset != NULL && NEGATIVE(ardataset)) { /* * The answer in the cache is * better than the answer we * found, and is a negative * cache entry, so we must set * eresult appropriately. */ if (NXDOMAIN(ardataset)) eresult = DNS_R_NCACHENXDOMAIN; else eresult = DNS_R_NCACHENXRRSET; /* * We have a negative response * from the cache so don't * attempt to add the RRSIG * rrset. */ continue; } } if (result != ISC_R_SUCCESS) break; if (sigrdataset != NULL) { addedrdataset = asigrdataset; result = dns_db_addrdataset(fctx->cache, node, NULL, now, sigrdataset, 0, addedrdataset); if (result == DNS_R_UNCHANGED) result = ISC_R_SUCCESS; if (result != ISC_R_SUCCESS) break; } else if (!ANSWER(rdataset)) continue; } if (ANSWER(rdataset) && need_validation) { if (fctx->type != dns_rdatatype_any && fctx->type != dns_rdatatype_rrsig && fctx->type != dns_rdatatype_sig) { /* * This is The Answer. We will * validate it, but first we cache * the rest of the response - it may * contain useful keys. */ INSIST(valrdataset == NULL && valsigrdataset == NULL); valrdataset = rdataset; valsigrdataset = sigrdataset; } else { /* * This is one of (potentially) * multiple answers to an ANY * or SIG query. To keep things * simple, we just start the * validator right away rather * than caching first and * having to remember which * rdatasets needed validation. */ result = valcreate(fctx, addrinfo, name, rdataset->type, rdataset, sigrdataset, valoptions, task); /* * Defer any further validations. * This prevents multiple validators * from manipulating fctx->rmessage * simultaneously. */ valoptions |= DNS_VALIDATOR_DEFER; } } else if (CHAINING(rdataset)) { if (rdataset->type == dns_rdatatype_cname) eresult = DNS_R_CNAME; else { INSIST(rdataset->type == dns_rdatatype_dname); eresult = DNS_R_DNAME; } } } else if (!EXTERNAL(rdataset)) { /* * It's OK to cache this rdataset now. */ if (ANSWER(rdataset)) addedrdataset = ardataset; else if (ANSWERSIG(rdataset)) addedrdataset = asigrdataset; else addedrdataset = NULL; if (CHAINING(rdataset)) { if (rdataset->type == dns_rdatatype_cname) eresult = DNS_R_CNAME; else { INSIST(rdataset->type == dns_rdatatype_dname); eresult = DNS_R_DNAME; } } if (rdataset->trust == dns_trust_glue && (rdataset->type == dns_rdatatype_ns || (rdataset->type == dns_rdatatype_rrsig && rdataset->covers == dns_rdatatype_ns))) { /* * If the trust level is 'dns_trust_glue' * then we are adding data from a referral * we got while executing the search algorithm. * New referral data always takes precedence * over the existing cache contents. */ options = DNS_DBADD_FORCE; } else options = 0; if (ANSWER(rdataset) && rdataset->type != dns_rdatatype_rrsig) { isc_result_t tresult; dns_name_t *noqname = NULL; tresult = findnoqname(fctx, name, rdataset->type, &noqname); if (tresult == ISC_R_SUCCESS && noqname != NULL) { tresult = dns_rdataset_addnoqname( rdataset, noqname); RUNTIME_CHECK(tresult == ISC_R_SUCCESS); } } /* * Now we can add the rdataset. */ result = dns_db_addrdataset(fctx->cache, node, NULL, now, rdataset, options, addedrdataset); if (result == DNS_R_UNCHANGED) { if (ANSWER(rdataset) && ardataset != NULL && NEGATIVE(ardataset)) { /* * The answer in the cache is better * than the answer we found, and is * a negative cache entry, so we * must set eresult appropriately. */ if (NXDOMAIN(ardataset)) eresult = DNS_R_NCACHENXDOMAIN; else eresult = DNS_R_NCACHENXRRSET; } result = ISC_R_SUCCESS; } else if (result != ISC_R_SUCCESS) break; } } if (valrdataset != NULL) result = valcreate(fctx, addrinfo, name, fctx->type, valrdataset, valsigrdataset, valoptions, task); if (result == ISC_R_SUCCESS && have_answer) { fctx->attributes |= FCTX_ATTR_HAVEANSWER; if (event != NULL) { /* * Negative results must be indicated in event->result. */ if (dns_rdataset_isassociated(event->rdataset) && NEGATIVE(event->rdataset)) { INSIST(eresult == DNS_R_NCACHENXDOMAIN || eresult == DNS_R_NCACHENXRRSET); } event->result = eresult; dns_db_attach(fctx->cache, adbp); dns_db_transfernode(fctx->cache, &node, anodep); clone_results(fctx); } } if (node != NULL) dns_db_detachnode(fctx->cache, &node); return (result); } static inline isc_result_t cache_message(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo, isc_stdtime_t now) { isc_result_t result; dns_section_t section; dns_name_t *name; FCTXTRACE("cache_message"); fctx->attributes &= ~FCTX_ATTR_WANTCACHE; LOCK(&fctx->res->buckets[fctx->bucketnum].lock); for (section = DNS_SECTION_ANSWER; section <= DNS_SECTION_ADDITIONAL; section++) { result = dns_message_firstname(fctx->rmessage, section); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(fctx->rmessage, section, &name); if ((name->attributes & DNS_NAMEATTR_CACHE) != 0) { result = cache_name(fctx, name, addrinfo, now); if (result != ISC_R_SUCCESS) break; } result = dns_message_nextname(fctx->rmessage, section); } if (result != ISC_R_NOMORE) break; } if (result == ISC_R_NOMORE) result = ISC_R_SUCCESS; UNLOCK(&fctx->res->buckets[fctx->bucketnum].lock); return (result); } /* * Do what dns_ncache_addoptout() does, and then compute an appropriate eresult. */ static isc_result_t ncache_adderesult(dns_message_t *message, dns_db_t *cache, dns_dbnode_t *node, dns_rdatatype_t covers, isc_stdtime_t now, dns_ttl_t maxttl, isc_boolean_t optout, isc_boolean_t secure, dns_rdataset_t *ardataset, isc_result_t *eresultp) { isc_result_t result; dns_rdataset_t rdataset; if (ardataset == NULL) { dns_rdataset_init(&rdataset); ardataset = &rdataset; } if (secure) result = dns_ncache_addoptout(message, cache, node, covers, now, maxttl, optout, ardataset); else result = dns_ncache_add(message, cache, node, covers, now, maxttl, ardataset); if (result == DNS_R_UNCHANGED || result == ISC_R_SUCCESS) { /* * If the cache now contains a negative entry and we * care about whether it is DNS_R_NCACHENXDOMAIN or * DNS_R_NCACHENXRRSET then extract it. */ if (NEGATIVE(ardataset)) { /* * The cache data is a negative cache entry. */ if (NXDOMAIN(ardataset)) *eresultp = DNS_R_NCACHENXDOMAIN; else *eresultp = DNS_R_NCACHENXRRSET; } else { /* * Either we don't care about the nature of the * cache rdataset (because no fetch is interested * in the outcome), or the cache rdataset is not * a negative cache entry. Whichever case it is, * we can return success. * * XXXRTH There's a CNAME/DNAME problem here. */ *eresultp = ISC_R_SUCCESS; } result = ISC_R_SUCCESS; } if (ardataset == &rdataset && dns_rdataset_isassociated(ardataset)) dns_rdataset_disassociate(ardataset); return (result); } static inline isc_result_t ncache_message(fetchctx_t *fctx, dns_adbaddrinfo_t *addrinfo, dns_rdatatype_t covers, isc_stdtime_t now) { isc_result_t result, eresult; dns_name_t *name; dns_resolver_t *res; dns_db_t **adbp; dns_dbnode_t *node, **anodep; dns_rdataset_t *ardataset; isc_boolean_t need_validation, secure_domain; dns_name_t *aname; dns_fetchevent_t *event; isc_uint32_t ttl; unsigned int valoptions = 0; FCTXTRACE("ncache_message"); fctx->attributes &= ~FCTX_ATTR_WANTNCACHE; res = fctx->res; need_validation = ISC_FALSE; POST(need_validation); secure_domain = ISC_FALSE; eresult = ISC_R_SUCCESS; name = &fctx->name; node = NULL; /* * XXXMPA remove when we follow cnames and adjust the setting * of FCTX_ATTR_WANTNCACHE in noanswer_response(). */ INSIST(fctx->rmessage->counts[DNS_SECTION_ANSWER] == 0); /* * Is DNSSEC validation required for this name? */ if (fctx->res->view->enablevalidation) { result = dns_view_issecuredomain(res->view, name, &secure_domain); if (result != ISC_R_SUCCESS) return (result); if (!secure_domain && res->view->dlv != NULL) { valoptions = DNS_VALIDATOR_DLV; secure_domain = ISC_TRUE; } } if ((fctx->options & DNS_FETCHOPT_NOVALIDATE) != 0) need_validation = ISC_FALSE; else need_validation = secure_domain; if (secure_domain) { /* * Mark all rdatasets as pending. */ dns_rdataset_t *trdataset; dns_name_t *tname; result = dns_message_firstname(fctx->rmessage, DNS_SECTION_AUTHORITY); while (result == ISC_R_SUCCESS) { tname = NULL; dns_message_currentname(fctx->rmessage, DNS_SECTION_AUTHORITY, &tname); for (trdataset = ISC_LIST_HEAD(tname->list); trdataset != NULL; trdataset = ISC_LIST_NEXT(trdataset, link)) trdataset->trust = dns_trust_pending_answer; result = dns_message_nextname(fctx->rmessage, DNS_SECTION_AUTHORITY); } if (result != ISC_R_NOMORE) return (result); } if (need_validation) { /* * Do negative response validation. */ result = valcreate(fctx, addrinfo, name, fctx->type, NULL, NULL, valoptions, res->buckets[fctx->bucketnum].task); /* * If validation is necessary, return now. Otherwise continue * to process the message, letting the validation complete * in its own good time. */ return (result); } LOCK(&res->buckets[fctx->bucketnum].lock); adbp = NULL; aname = NULL; anodep = NULL; ardataset = NULL; if (!HAVE_ANSWER(fctx)) { event = ISC_LIST_HEAD(fctx->events); if (event != NULL) { adbp = &event->db; aname = dns_fixedname_name(&event->foundname); result = dns_name_copy(name, aname, NULL); if (result != ISC_R_SUCCESS) goto unlock; anodep = &event->node; ardataset = event->rdataset; } } else event = NULL; result = dns_db_findnode(fctx->cache, name, ISC_TRUE, &node); if (result != ISC_R_SUCCESS) goto unlock; /* * If we are asking for a SOA record set the cache time * to zero to facilitate locating the containing zone of * a arbitrary zone. */ ttl = fctx->res->view->maxncachettl; if (fctx->type == dns_rdatatype_soa && covers == dns_rdatatype_any && fctx->res->zero_no_soa_ttl) ttl = 0; result = ncache_adderesult(fctx->rmessage, fctx->cache, node, covers, now, ttl, ISC_FALSE, ISC_FALSE, ardataset, &eresult); if (result != ISC_R_SUCCESS) goto unlock; if (!HAVE_ANSWER(fctx)) { fctx->attributes |= FCTX_ATTR_HAVEANSWER; if (event != NULL) { event->result = eresult; dns_db_attach(fctx->cache, adbp); dns_db_transfernode(fctx->cache, &node, anodep); clone_results(fctx); } } unlock: UNLOCK(&res->buckets[fctx->bucketnum].lock); if (node != NULL) dns_db_detachnode(fctx->cache, &node); return (result); } static inline void mark_related(dns_name_t *name, dns_rdataset_t *rdataset, isc_boolean_t external, isc_boolean_t gluing) { name->attributes |= DNS_NAMEATTR_CACHE; if (gluing) { rdataset->trust = dns_trust_glue; /* * Glue with 0 TTL causes problems. We force the TTL to * 1 second to prevent this. */ if (rdataset->ttl == 0) rdataset->ttl = 1; } else rdataset->trust = dns_trust_additional; /* * Avoid infinite loops by only marking new rdatasets. */ if (!CACHE(rdataset)) { name->attributes |= DNS_NAMEATTR_CHASE; rdataset->attributes |= DNS_RDATASETATTR_CHASE; } rdataset->attributes |= DNS_RDATASETATTR_CACHE; if (external) rdataset->attributes |= DNS_RDATASETATTR_EXTERNAL; } static isc_result_t check_section(void *arg, dns_name_t *addname, dns_rdatatype_t type, dns_section_t section) { fetchctx_t *fctx = arg; isc_result_t result; dns_name_t *name; dns_rdataset_t *rdataset; isc_boolean_t external; dns_rdatatype_t rtype; isc_boolean_t gluing; REQUIRE(VALID_FCTX(fctx)); #if CHECK_FOR_GLUE_IN_ANSWER if (section == DNS_SECTION_ANSWER && type != dns_rdatatype_a) return (ISC_R_SUCCESS); #endif if (GLUING(fctx)) gluing = ISC_TRUE; else gluing = ISC_FALSE; name = NULL; rdataset = NULL; result = dns_message_findname(fctx->rmessage, section, addname, dns_rdatatype_any, 0, &name, NULL); if (result == ISC_R_SUCCESS) { external = ISC_TF(!dns_name_issubdomain(name, &fctx->domain)); if (type == dns_rdatatype_a) { for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { if (rdataset->type == dns_rdatatype_rrsig) rtype = rdataset->covers; else rtype = rdataset->type; if (rtype == dns_rdatatype_a || rtype == dns_rdatatype_aaaa) mark_related(name, rdataset, external, gluing); } } else { result = dns_message_findtype(name, type, 0, &rdataset); if (result == ISC_R_SUCCESS) { mark_related(name, rdataset, external, gluing); /* * Do we have its SIG too? */ rdataset = NULL; result = dns_message_findtype(name, dns_rdatatype_rrsig, type, &rdataset); if (result == ISC_R_SUCCESS) mark_related(name, rdataset, external, gluing); } } } return (ISC_R_SUCCESS); } static isc_result_t check_related(void *arg, dns_name_t *addname, dns_rdatatype_t type) { return (check_section(arg, addname, type, DNS_SECTION_ADDITIONAL)); } #ifndef CHECK_FOR_GLUE_IN_ANSWER #define CHECK_FOR_GLUE_IN_ANSWER 0 #endif #if CHECK_FOR_GLUE_IN_ANSWER static isc_result_t check_answer(void *arg, dns_name_t *addname, dns_rdatatype_t type) { return (check_section(arg, addname, type, DNS_SECTION_ANSWER)); } #endif static void chase_additional(fetchctx_t *fctx) { isc_boolean_t rescan; dns_section_t section = DNS_SECTION_ADDITIONAL; isc_result_t result; again: rescan = ISC_FALSE; for (result = dns_message_firstname(fctx->rmessage, section); result == ISC_R_SUCCESS; result = dns_message_nextname(fctx->rmessage, section)) { dns_name_t *name = NULL; dns_rdataset_t *rdataset; dns_message_currentname(fctx->rmessage, DNS_SECTION_ADDITIONAL, &name); if ((name->attributes & DNS_NAMEATTR_CHASE) == 0) continue; name->attributes &= ~DNS_NAMEATTR_CHASE; for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { if (CHASE(rdataset)) { rdataset->attributes &= ~DNS_RDATASETATTR_CHASE; (void)dns_rdataset_additionaldata(rdataset, check_related, fctx); rescan = ISC_TRUE; } } } if (rescan) goto again; } static inline isc_result_t cname_target(dns_rdataset_t *rdataset, dns_name_t *tname) { isc_result_t result; dns_rdata_t rdata = DNS_RDATA_INIT; dns_rdata_cname_t cname; result = dns_rdataset_first(rdataset); if (result != ISC_R_SUCCESS) return (result); dns_rdataset_current(rdataset, &rdata); result = dns_rdata_tostruct(&rdata, &cname, NULL); if (result != ISC_R_SUCCESS) return (result); dns_name_init(tname, NULL); dns_name_clone(&cname.cname, tname); dns_rdata_freestruct(&cname); return (ISC_R_SUCCESS); } static inline isc_result_t -dname_target(fetchctx_t *fctx, dns_rdataset_t *rdataset, dns_name_t *qname, - dns_name_t *oname, dns_fixedname_t *fixeddname) +dname_target(dns_rdataset_t *rdataset, dns_name_t *qname, + unsigned int nlabels, dns_fixedname_t *fixeddname) { isc_result_t result; dns_rdata_t rdata = DNS_RDATA_INIT; - unsigned int nlabels; - int order; - dns_namereln_t namereln; dns_rdata_dname_t dname; dns_fixedname_t prefix; /* * Get the target name of the DNAME. */ result = dns_rdataset_first(rdataset); if (result != ISC_R_SUCCESS) return (result); dns_rdataset_current(rdataset, &rdata); result = dns_rdata_tostruct(&rdata, &dname, NULL); if (result != ISC_R_SUCCESS) return (result); - /* - * Get the prefix of qname. - */ - namereln = dns_name_fullcompare(qname, oname, &order, &nlabels); - if (namereln != dns_namereln_subdomain) { - char qbuf[DNS_NAME_FORMATSIZE]; - char obuf[DNS_NAME_FORMATSIZE]; - - dns_rdata_freestruct(&dname); - dns_name_format(qname, qbuf, sizeof(qbuf)); - dns_name_format(oname, obuf, sizeof(obuf)); - log_formerr(fctx, "unrelated DNAME in answer: " - "%s is not in %s", qbuf, obuf); - return (DNS_R_FORMERR); - } dns_fixedname_init(&prefix); dns_name_split(qname, nlabels, dns_fixedname_name(&prefix), NULL); dns_fixedname_init(fixeddname); result = dns_name_concatenate(dns_fixedname_name(&prefix), &dname.dname, dns_fixedname_name(fixeddname), NULL); dns_rdata_freestruct(&dname); return (result); } static isc_boolean_t is_answeraddress_allowed(dns_view_t *view, dns_name_t *name, dns_rdataset_t *rdataset) { isc_result_t result; dns_rdata_t rdata = DNS_RDATA_INIT; struct in_addr ina; struct in6_addr in6a; isc_netaddr_t netaddr; char addrbuf[ISC_NETADDR_FORMATSIZE]; char namebuf[DNS_NAME_FORMATSIZE]; char classbuf[64]; char typebuf[64]; int match; /* By default, we allow any addresses. */ if (view->denyansweracl == NULL) return (ISC_TRUE); /* * If the owner name matches one in the exclusion list, either exactly * or partially, allow it. */ if (view->answeracl_exclude != NULL) { dns_rbtnode_t *node = NULL; result = dns_rbt_findnode(view->answeracl_exclude, name, NULL, &node, NULL, 0, NULL, NULL); if (result == ISC_R_SUCCESS || result == DNS_R_PARTIALMATCH) return (ISC_TRUE); } /* * Otherwise, search the filter list for a match for each address * record. If a match is found, the address should be filtered, * so should the entire answer. */ for (result = dns_rdataset_first(rdataset); result == ISC_R_SUCCESS; result = dns_rdataset_next(rdataset)) { dns_rdata_reset(&rdata); dns_rdataset_current(rdataset, &rdata); if (rdataset->type == dns_rdatatype_a) { INSIST(rdata.length == sizeof(ina.s_addr)); memmove(&ina.s_addr, rdata.data, sizeof(ina.s_addr)); isc_netaddr_fromin(&netaddr, &ina); } else { INSIST(rdata.length == sizeof(in6a.s6_addr)); memmove(in6a.s6_addr, rdata.data, sizeof(in6a.s6_addr)); isc_netaddr_fromin6(&netaddr, &in6a); } result = dns_acl_match(&netaddr, NULL, view->denyansweracl, &view->aclenv, &match, NULL); if (result == ISC_R_SUCCESS && match > 0) { isc_netaddr_format(&netaddr, addrbuf, sizeof(addrbuf)); dns_name_format(name, namebuf, sizeof(namebuf)); dns_rdatatype_format(rdataset->type, typebuf, sizeof(typebuf)); dns_rdataclass_format(rdataset->rdclass, classbuf, sizeof(classbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "answer address %s denied for %s/%s/%s", addrbuf, namebuf, typebuf, classbuf); return (ISC_FALSE); } } return (ISC_TRUE); } static isc_boolean_t is_answertarget_allowed(dns_view_t *view, dns_name_t *name, dns_rdatatype_t type, dns_name_t *tname, dns_name_t *domain) { isc_result_t result; dns_rbtnode_t *node = NULL; char qnamebuf[DNS_NAME_FORMATSIZE]; char tnamebuf[DNS_NAME_FORMATSIZE]; char classbuf[64]; char typebuf[64]; /* By default, we allow any target name. */ if (view->denyanswernames == NULL) return (ISC_TRUE); /* * If the owner name matches one in the exclusion list, either exactly * or partially, allow it. */ if (view->answernames_exclude != NULL) { result = dns_rbt_findnode(view->answernames_exclude, name, NULL, &node, NULL, 0, NULL, NULL); if (result == ISC_R_SUCCESS || result == DNS_R_PARTIALMATCH) return (ISC_TRUE); } /* * If the target name is a subdomain of the search domain, allow it. */ if (dns_name_issubdomain(tname, domain)) return (ISC_TRUE); /* * Otherwise, apply filters. */ result = dns_rbt_findnode(view->denyanswernames, tname, NULL, &node, NULL, 0, NULL, NULL); if (result == ISC_R_SUCCESS || result == DNS_R_PARTIALMATCH) { dns_name_format(name, qnamebuf, sizeof(qnamebuf)); dns_name_format(tname, tnamebuf, sizeof(tnamebuf)); dns_rdatatype_format(type, typebuf, sizeof(typebuf)); dns_rdataclass_format(view->rdclass, classbuf, sizeof(classbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "%s target %s denied for %s/%s", typebuf, tnamebuf, qnamebuf, classbuf); return (ISC_FALSE); } return (ISC_TRUE); } static void trim_ns_ttl(fetchctx_t *fctx, dns_name_t *name, dns_rdataset_t *rdataset) { char ns_namebuf[DNS_NAME_FORMATSIZE]; char namebuf[DNS_NAME_FORMATSIZE]; char tbuf[DNS_RDATATYPE_FORMATSIZE]; if (fctx->ns_ttl_ok && rdataset->ttl > fctx->ns_ttl) { dns_name_format(name, ns_namebuf, sizeof(ns_namebuf)); dns_name_format(&fctx->name, namebuf, sizeof(namebuf)); dns_rdatatype_format(fctx->type, tbuf, sizeof(tbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_DEBUG(10), "fctx %p: trimming ttl of %s/NS for %s/%s: " "%u -> %u", fctx, ns_namebuf, namebuf, tbuf, rdataset->ttl, fctx->ns_ttl); rdataset->ttl = fctx->ns_ttl; } } /* * Handle a no-answer response (NXDOMAIN, NXRRSET, or referral). * If look_in_options has LOOK_FOR_NS_IN_ANSWER then we look in the answer * section for the NS RRset if the query type is NS; if it has * LOOK_FOR_GLUE_IN_ANSWER we look for glue incorrectly returned in the answer * section for A and AAAA queries. */ #define LOOK_FOR_NS_IN_ANSWER 0x1 #define LOOK_FOR_GLUE_IN_ANSWER 0x2 static isc_result_t noanswer_response(fetchctx_t *fctx, dns_name_t *oqname, unsigned int look_in_options) { isc_result_t result; dns_message_t *message; dns_name_t *name, *qname, *ns_name, *soa_name, *ds_name, *save_name; dns_rdataset_t *rdataset, *ns_rdataset; isc_boolean_t aa, negative_response; dns_rdatatype_t type, save_type; dns_section_t section; FCTXTRACE("noanswer_response"); if ((look_in_options & LOOK_FOR_NS_IN_ANSWER) != 0) { INSIST(fctx->type == dns_rdatatype_ns); section = DNS_SECTION_ANSWER; } else section = DNS_SECTION_AUTHORITY; message = fctx->rmessage; /* * Setup qname. */ if (oqname == NULL) { /* * We have a normal, non-chained negative response or * referral. */ if ((message->flags & DNS_MESSAGEFLAG_AA) != 0) aa = ISC_TRUE; else aa = ISC_FALSE; qname = &fctx->name; } else { /* * We're being invoked by answer_response() after it has * followed a CNAME/DNAME chain. */ qname = oqname; aa = ISC_FALSE; /* * If the current qname is not a subdomain of the query * domain, there's no point in looking at the authority * section without doing DNSSEC validation. * * Until we do that validation, we'll just return success * in this case. */ if (!dns_name_issubdomain(qname, &fctx->domain)) return (ISC_R_SUCCESS); } /* * We have to figure out if this is a negative response, or a * referral. */ /* * Sometimes we can tell if its a negative response by looking at * the message header. */ negative_response = ISC_FALSE; if (message->rcode == dns_rcode_nxdomain || (message->counts[DNS_SECTION_ANSWER] == 0 && message->counts[DNS_SECTION_AUTHORITY] == 0)) negative_response = ISC_TRUE; /* * Process the authority section. */ ns_name = NULL; ns_rdataset = NULL; soa_name = NULL; ds_name = NULL; save_name = NULL; save_type = dns_rdatatype_none; result = dns_message_firstname(message, section); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(message, section, &name); if (dns_name_issubdomain(name, &fctx->domain)) { /* * Look for NS/SOA RRsets first. */ for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { type = rdataset->type; if (type == dns_rdatatype_rrsig) type = rdataset->covers; if (((type == dns_rdatatype_ns || type == dns_rdatatype_soa) && !dns_name_issubdomain(qname, name))) { char qbuf[DNS_NAME_FORMATSIZE]; char nbuf[DNS_NAME_FORMATSIZE]; char tbuf[DNS_RDATATYPE_FORMATSIZE]; dns_rdatatype_format(fctx->type, tbuf, sizeof(tbuf)); dns_name_format(name, nbuf, sizeof(nbuf)); dns_name_format(qname, qbuf, sizeof(qbuf)); log_formerr(fctx, "unrelated %s %s in " "%s authority section", tbuf, qbuf, nbuf); return (DNS_R_FORMERR); } if (type == dns_rdatatype_ns) { /* * NS or RRSIG NS. * * Only one set of NS RRs is allowed. */ if (rdataset->type == dns_rdatatype_ns) { if (ns_name != NULL && name != ns_name) { log_formerr(fctx, "multiple NS " "RRsets in " "authority " "section"); return (DNS_R_FORMERR); } ns_name = name; ns_rdataset = rdataset; } name->attributes |= DNS_NAMEATTR_CACHE; rdataset->attributes |= DNS_RDATASETATTR_CACHE; rdataset->trust = dns_trust_glue; } if (type == dns_rdatatype_soa) { /* * SOA, or RRSIG SOA. * * Only one SOA is allowed. */ if (rdataset->type == dns_rdatatype_soa) { if (soa_name != NULL && name != soa_name) { log_formerr(fctx, "multiple SOA " "RRs in " "authority " "section"); return (DNS_R_FORMERR); } soa_name = name; } name->attributes |= DNS_NAMEATTR_NCACHE; rdataset->attributes |= DNS_RDATASETATTR_NCACHE; if (aa) rdataset->trust = dns_trust_authauthority; else if (ISFORWARDER(fctx->addrinfo)) rdataset->trust = dns_trust_answer; else rdataset->trust = dns_trust_additional; } } } result = dns_message_nextname(message, section); if (result == ISC_R_NOMORE) break; else if (result != ISC_R_SUCCESS) return (result); } log_ns_ttl(fctx, "noanswer_response"); if (ns_rdataset != NULL && dns_name_equal(&fctx->domain, ns_name) && !dns_name_equal(ns_name, dns_rootname)) trim_ns_ttl(fctx, ns_name, ns_rdataset); /* * A negative response has a SOA record (Type 2) * and a optional NS RRset (Type 1) or it has neither * a SOA or a NS RRset (Type 3, handled above) or * rcode is NXDOMAIN (handled above) in which case * the NS RRset is allowed (Type 4). */ if (soa_name != NULL) negative_response = ISC_TRUE; result = dns_message_firstname(message, section); while (result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(message, section, &name); if (dns_name_issubdomain(name, &fctx->domain)) { for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { type = rdataset->type; if (type == dns_rdatatype_rrsig) type = rdataset->covers; if (type == dns_rdatatype_nsec || type == dns_rdatatype_nsec3) { /* * NSEC or RRSIG NSEC. */ if (negative_response) { name->attributes |= DNS_NAMEATTR_NCACHE; rdataset->attributes |= DNS_RDATASETATTR_NCACHE; } else if (type == dns_rdatatype_nsec) { name->attributes |= DNS_NAMEATTR_CACHE; rdataset->attributes |= DNS_RDATASETATTR_CACHE; } if (aa) rdataset->trust = dns_trust_authauthority; else if (ISFORWARDER(fctx->addrinfo)) rdataset->trust = dns_trust_answer; else rdataset->trust = dns_trust_additional; /* * No additional data needs to be * marked. */ } else if (type == dns_rdatatype_ds) { /* * DS or SIG DS. * * These should only be here if * this is a referral, and there * should only be one DS RRset. */ if (ns_name == NULL) { log_formerr(fctx, "DS with no " "referral"); return (DNS_R_FORMERR); } if (rdataset->type == dns_rdatatype_ds) { if (ds_name != NULL && name != ds_name) { log_formerr(fctx, "DS doesn't " "match " "referral " "(NS)"); return (DNS_R_FORMERR); } ds_name = name; } name->attributes |= DNS_NAMEATTR_CACHE; rdataset->attributes |= DNS_RDATASETATTR_CACHE; if (aa) rdataset->trust = dns_trust_authauthority; else if (ISFORWARDER(fctx->addrinfo)) rdataset->trust = dns_trust_answer; else rdataset->trust = dns_trust_additional; } } } else { save_name = name; save_type = ISC_LIST_HEAD(name->list)->type; } result = dns_message_nextname(message, section); if (result == ISC_R_NOMORE) break; else if (result != ISC_R_SUCCESS) return (result); } /* * Trigger lookups for DNS nameservers. */ if (negative_response && message->rcode == dns_rcode_noerror && fctx->type == dns_rdatatype_ds && soa_name != NULL && dns_name_equal(soa_name, qname) && !dns_name_equal(qname, dns_rootname)) return (DNS_R_CHASEDSSERVERS); /* * Did we find anything? */ if (!negative_response && ns_name == NULL) { /* * Nope. */ if (oqname != NULL) { /* * We've already got a partial CNAME/DNAME chain, * and haven't found else anything useful here, but * no error has occurred since we have an answer. */ return (ISC_R_SUCCESS); } else { /* * The responder is insane. */ if (save_name == NULL) { log_formerr(fctx, "invalid response"); return (DNS_R_FORMERR); } if (!dns_name_issubdomain(save_name, &fctx->domain)) { char nbuf[DNS_NAME_FORMATSIZE]; char dbuf[DNS_NAME_FORMATSIZE]; char tbuf[DNS_RDATATYPE_FORMATSIZE]; dns_rdatatype_format(save_type, tbuf, sizeof(tbuf)); dns_name_format(save_name, nbuf, sizeof(nbuf)); dns_name_format(&fctx->domain, dbuf, sizeof(dbuf)); log_formerr(fctx, "Name %s (%s) not subdomain" " of zone %s -- invalid response", nbuf, tbuf, dbuf); } else { log_formerr(fctx, "invalid response"); } return (DNS_R_FORMERR); } } /* * If we found both NS and SOA, they should be the same name. */ if (ns_name != NULL && soa_name != NULL && ns_name != soa_name) { log_formerr(fctx, "NS/SOA mismatch"); return (DNS_R_FORMERR); } /* * Do we have a referral? (We only want to follow a referral if * we're not following a chain.) */ if (!negative_response && ns_name != NULL && oqname == NULL) { /* * We already know ns_name is a subdomain of fctx->domain. * If ns_name is equal to fctx->domain, we're not making * progress. We return DNS_R_FORMERR so that we'll keep * trying other servers. */ if (dns_name_equal(ns_name, &fctx->domain)) { log_formerr(fctx, "non-improving referral"); return (DNS_R_FORMERR); } /* * If the referral name is not a parent of the query * name, consider the responder insane. */ if (! dns_name_issubdomain(&fctx->name, ns_name)) { /* Logged twice */ log_formerr(fctx, "referral to non-parent"); FCTXTRACE("referral to non-parent"); return (DNS_R_FORMERR); } /* * Mark any additional data related to this rdataset. * It's important that we do this before we change the * query domain. */ INSIST(ns_rdataset != NULL); fctx->attributes |= FCTX_ATTR_GLUING; (void)dns_rdataset_additionaldata(ns_rdataset, check_related, fctx); #if CHECK_FOR_GLUE_IN_ANSWER /* * Look in the answer section for "glue" that is incorrectly * returned as a answer. This is needed if the server also * minimizes the response size by not adding records to the * additional section that are in the answer section or if * the record gets dropped due to message size constraints. */ if ((look_in_options & LOOK_FOR_GLUE_IN_ANSWER) != 0 && (fctx->type == dns_rdatatype_aaaa || fctx->type == dns_rdatatype_a)) (void)dns_rdataset_additionaldata(ns_rdataset, check_answer, fctx); #endif fctx->attributes &= ~FCTX_ATTR_GLUING; /* * NS rdatasets with 0 TTL cause problems. * dns_view_findzonecut() will not find them when we * try to follow the referral, and we'll SERVFAIL * because the best nameservers are now above QDOMAIN. * We force the TTL to 1 second to prevent this. */ if (ns_rdataset->ttl == 0) ns_rdataset->ttl = 1; /* * Set the current query domain to the referral name. * * XXXRTH We should check if we're in forward-only mode, and * if so we should bail out. */ INSIST(dns_name_countlabels(&fctx->domain) > 0); dns_name_free(&fctx->domain, fctx->mctx); if (dns_rdataset_isassociated(&fctx->nameservers)) dns_rdataset_disassociate(&fctx->nameservers); dns_name_init(&fctx->domain, NULL); result = dns_name_dup(ns_name, fctx->mctx, &fctx->domain); if (result != ISC_R_SUCCESS) return (result); fctx->attributes |= FCTX_ATTR_WANTCACHE; fctx->ns_ttl_ok = ISC_FALSE; log_ns_ttl(fctx, "DELEGATION"); return (DNS_R_DELEGATION); } /* * Since we're not doing a referral, we don't want to cache any * NS RRs we may have found. */ if (ns_name != NULL) ns_name->attributes &= ~DNS_NAMEATTR_CACHE; if (negative_response && oqname == NULL) fctx->attributes |= FCTX_ATTR_WANTNCACHE; return (ISC_R_SUCCESS); } static isc_result_t answer_response(fetchctx_t *fctx) { isc_result_t result; dns_message_t *message; - dns_name_t *name, *qname, tname, *ns_name; + dns_name_t *name, *dname = NULL, *qname, tname, *ns_name; dns_rdataset_t *rdataset, *ns_rdataset; isc_boolean_t done, external, chaining, aa, found, want_chaining; isc_boolean_t have_answer, found_cname, found_type, wanted_chaining; unsigned int aflag; dns_rdatatype_t type; - dns_fixedname_t dname, fqname; + dns_fixedname_t fdname, fqname; dns_view_t *view; FCTXTRACE("answer_response"); message = fctx->rmessage; /* * Examine the answer section, marking those rdatasets which are * part of the answer and should be cached. */ done = ISC_FALSE; found_cname = ISC_FALSE; found_type = ISC_FALSE; chaining = ISC_FALSE; have_answer = ISC_FALSE; want_chaining = ISC_FALSE; POST(want_chaining); if ((message->flags & DNS_MESSAGEFLAG_AA) != 0) aa = ISC_TRUE; else aa = ISC_FALSE; qname = &fctx->name; type = fctx->type; view = fctx->res->view; result = dns_message_firstname(message, DNS_SECTION_ANSWER); while (!done && result == ISC_R_SUCCESS) { + dns_namereln_t namereln; + int order; + unsigned int nlabels; + name = NULL; dns_message_currentname(message, DNS_SECTION_ANSWER, &name); external = ISC_TF(!dns_name_issubdomain(name, &fctx->domain)); - if (dns_name_equal(name, qname)) { + namereln = dns_name_fullcompare(qname, name, &order, &nlabels); + if (namereln == dns_namereln_equal) { wanted_chaining = ISC_FALSE; for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { found = ISC_FALSE; want_chaining = ISC_FALSE; aflag = 0; if (rdataset->type == dns_rdatatype_nsec3) { /* * NSEC3 records are not allowed to * appear in the answer section. */ log_formerr(fctx, "NSEC3 in answer"); return (DNS_R_FORMERR); } /* * Apply filters, if given, on answers to reject * a malicious attempt of rebinding. */ if ((rdataset->type == dns_rdatatype_a || rdataset->type == dns_rdatatype_aaaa) && !is_answeraddress_allowed(view, name, rdataset)) { return (DNS_R_SERVFAIL); } if (rdataset->type == type && !found_cname) { /* * We've found an ordinary answer. */ found = ISC_TRUE; found_type = ISC_TRUE; done = ISC_TRUE; aflag = DNS_RDATASETATTR_ANSWER; } else if (type == dns_rdatatype_any) { /* * We've found an answer matching * an ANY query. There may be * more. */ found = ISC_TRUE; aflag = DNS_RDATASETATTR_ANSWER; } else if (rdataset->type == dns_rdatatype_rrsig && rdataset->covers == type && !found_cname) { /* * We've found a signature that * covers the type we're looking for. */ found = ISC_TRUE; found_type = ISC_TRUE; aflag = DNS_RDATASETATTR_ANSWERSIG; } else if (rdataset->type == dns_rdatatype_cname && !found_type) { /* * We're looking for something else, * but we found a CNAME. * * Getting a CNAME response for some * query types is an error, see * RFC 4035, Section 2.5. */ if (type == dns_rdatatype_rrsig || type == dns_rdatatype_key || type == dns_rdatatype_nsec) { char buf[DNS_RDATATYPE_FORMATSIZE]; dns_rdatatype_format(fctx->type, buf, sizeof(buf)); log_formerr(fctx, "CNAME response " "for %s RR", buf); return (DNS_R_FORMERR); } found = ISC_TRUE; found_cname = ISC_TRUE; want_chaining = ISC_TRUE; aflag = DNS_RDATASETATTR_ANSWER; result = cname_target(rdataset, &tname); if (result != ISC_R_SUCCESS) return (result); /* Apply filters on the target name. */ if (!is_answertarget_allowed(view, name, rdataset->type, &tname, &fctx->domain)) { return (DNS_R_SERVFAIL); } } else if (rdataset->type == dns_rdatatype_rrsig && rdataset->covers == dns_rdatatype_cname && !found_type) { /* * We're looking for something else, * but we found a SIG CNAME. */ found = ISC_TRUE; found_cname = ISC_TRUE; aflag = DNS_RDATASETATTR_ANSWERSIG; } if (found) { /* * We've found an answer to our * question. */ name->attributes |= DNS_NAMEATTR_CACHE; rdataset->attributes |= DNS_RDATASETATTR_CACHE; rdataset->trust = dns_trust_answer; if (!chaining) { /* * This data is "the" answer * to our question only if * we're not chaining (i.e. * if we haven't followed * a CNAME or DNAME). */ INSIST(!external); if (aflag == - DNS_RDATASETATTR_ANSWER) + DNS_RDATASETATTR_ANSWER) { have_answer = ISC_TRUE; - name->attributes |= - DNS_NAMEATTR_ANSWER; + name->attributes |= + DNS_NAMEATTR_ANSWER; + } rdataset->attributes |= aflag; if (aa) rdataset->trust = dns_trust_authanswer; } else if (external) { /* * This data is outside of * our query domain, and * may not be cached. */ rdataset->attributes |= DNS_RDATASETATTR_EXTERNAL; } /* * Mark any additional data related * to this rdataset. */ (void)dns_rdataset_additionaldata( rdataset, check_related, fctx); /* * CNAME chaining. */ if (want_chaining) { wanted_chaining = ISC_TRUE; name->attributes |= DNS_NAMEATTR_CHAINING; rdataset->attributes |= DNS_RDATASETATTR_CHAINING; qname = &tname; } } /* * We could add an "else" clause here and * log that we're ignoring this rdataset. */ } /* * If wanted_chaining is true, we've done * some chaining as the result of processing * this node, and thus we need to set * chaining to true. * * We don't set chaining inside of the * rdataset loop because doing that would * cause us to ignore the signatures of * CNAMEs. */ if (wanted_chaining) chaining = ISC_TRUE; } else { + dns_rdataset_t *dnameset = NULL; + /* * Look for a DNAME (or its SIG). Anything else is * ignored. */ wanted_chaining = ISC_FALSE; for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; - rdataset = ISC_LIST_NEXT(rdataset, link)) { - isc_boolean_t found_dname = ISC_FALSE; - dns_name_t *dname_name; + rdataset = ISC_LIST_NEXT(rdataset, link)) + { + /* + * Only pass DNAME or RRSIG(DNAME). + */ + if (rdataset->type != dns_rdatatype_dname && + (rdataset->type != dns_rdatatype_rrsig || + rdataset->covers != dns_rdatatype_dname)) + continue; - found = ISC_FALSE; + /* + * If we're not chaining, then the DNAME and + * its signature should not be external. + */ + if (!chaining && external) { + char qbuf[DNS_NAME_FORMATSIZE]; + char obuf[DNS_NAME_FORMATSIZE]; + + dns_name_format(name, qbuf, + sizeof(qbuf)); + dns_name_format(&fctx->domain, obuf, + sizeof(obuf)); + log_formerr(fctx, "external DNAME or " + "RRSIG covering DNAME " + "in answer: %s is " + "not in %s", qbuf, obuf); + return (DNS_R_FORMERR); + } + + if (namereln != dns_namereln_subdomain) { + char qbuf[DNS_NAME_FORMATSIZE]; + char obuf[DNS_NAME_FORMATSIZE]; + + dns_name_format(qname, qbuf, + sizeof(qbuf)); + dns_name_format(name, obuf, + sizeof(obuf)); + log_formerr(fctx, "unrelated DNAME " + "in answer: %s is " + "not in %s", qbuf, obuf); + return (DNS_R_FORMERR); + } + aflag = 0; if (rdataset->type == dns_rdatatype_dname) { - /* - * We're looking for something else, - * but we found a DNAME. - * - * If we're not chaining, then the - * DNAME should not be external. - */ - if (!chaining && external) { - log_formerr(fctx, - "external DNAME"); - return (DNS_R_FORMERR); - } - found = ISC_TRUE; want_chaining = ISC_TRUE; POST(want_chaining); aflag = DNS_RDATASETATTR_ANSWER; - result = dname_target(fctx, rdataset, - qname, name, - &dname); + result = dname_target(rdataset, qname, + nlabels, &fdname); if (result == ISC_R_NOSPACE) { /* * We can't construct the * DNAME target. Do not * try to continue. */ want_chaining = ISC_FALSE; POST(want_chaining); } else if (result != ISC_R_SUCCESS) return (result); else - found_dname = ISC_TRUE; + dnameset = rdataset; - dname_name = dns_fixedname_name(&dname); + dname = dns_fixedname_name(&fdname); if (!is_answertarget_allowed(view, - qname, - rdataset->type, - dname_name, - &fctx->domain)) { + qname, rdataset->type, + dname, &fctx->domain)) { return (DNS_R_SERVFAIL); } - } else if (rdataset->type == dns_rdatatype_rrsig - && rdataset->covers == - dns_rdatatype_dname) { + } else { /* * We've found a signature that * covers the DNAME. */ - found = ISC_TRUE; aflag = DNS_RDATASETATTR_ANSWERSIG; } - if (found) { + /* + * We've found an answer to our + * question. + */ + name->attributes |= DNS_NAMEATTR_CACHE; + rdataset->attributes |= DNS_RDATASETATTR_CACHE; + rdataset->trust = dns_trust_answer; + if (!chaining) { /* - * We've found an answer to our - * question. + * This data is "the" answer to + * our question only if we're + * not chaining. */ - name->attributes |= - DNS_NAMEATTR_CACHE; - rdataset->attributes |= - DNS_RDATASETATTR_CACHE; - rdataset->trust = dns_trust_answer; - if (!chaining) { - /* - * This data is "the" answer - * to our question only if - * we're not chaining. - */ - INSIST(!external); - if (aflag == - DNS_RDATASETATTR_ANSWER) - have_answer = ISC_TRUE; + INSIST(!external); + if (aflag == DNS_RDATASETATTR_ANSWER) { + have_answer = ISC_TRUE; name->attributes |= DNS_NAMEATTR_ANSWER; - rdataset->attributes |= aflag; - if (aa) - rdataset->trust = - dns_trust_authanswer; - } else if (external) { - rdataset->attributes |= - DNS_RDATASETATTR_EXTERNAL; } - - /* - * DNAME chaining. - */ - if (found_dname) { - /* - * Copy the dname into the - * qname fixed name. - * - * Although we check for - * failure of the copy - * operation, in practice it - * should never fail since - * we already know that the - * result fits in a fixedname. - */ - dns_fixedname_init(&fqname); - result = dns_name_copy( - dns_fixedname_name(&dname), - dns_fixedname_name(&fqname), - NULL); - if (result != ISC_R_SUCCESS) - return (result); - wanted_chaining = ISC_TRUE; - name->attributes |= - DNS_NAMEATTR_CHAINING; - rdataset->attributes |= - DNS_RDATASETATTR_CHAINING; - qname = dns_fixedname_name( - &fqname); - } + rdataset->attributes |= aflag; + if (aa) + rdataset->trust = + dns_trust_authanswer; + } else if (external) { + rdataset->attributes |= + DNS_RDATASETATTR_EXTERNAL; } + } + + /* + * DNAME chaining. + */ + if (dnameset != NULL) { + /* + * Copy the dname into the qname fixed name. + * + * Although we check for failure of the copy + * operation, in practice it should never fail + * since we already know that the result fits + * in a fixedname. + */ + dns_fixedname_init(&fqname); + qname = dns_fixedname_name(&fqname); + result = dns_name_copy(dname, qname, NULL); + if (result != ISC_R_SUCCESS) + return (result); + wanted_chaining = ISC_TRUE; + name->attributes |= DNS_NAMEATTR_CHAINING; + dnameset->attributes |= + DNS_RDATASETATTR_CHAINING; } if (wanted_chaining) chaining = ISC_TRUE; } result = dns_message_nextname(message, DNS_SECTION_ANSWER); } if (result == ISC_R_NOMORE) result = ISC_R_SUCCESS; if (result != ISC_R_SUCCESS) return (result); /* * We should have found an answer. */ if (!have_answer) { log_formerr(fctx, "reply has no answer"); return (DNS_R_FORMERR); } /* * This response is now potentially cacheable. */ fctx->attributes |= FCTX_ATTR_WANTCACHE; /* * Did chaining end before we got the final answer? */ if (chaining) { /* * Yes. This may be a negative reply, so hand off * authority section processing to the noanswer code. * If it isn't a noanswer response, no harm will be * done. */ return (noanswer_response(fctx, qname, 0)); } /* * We didn't end with an incomplete chain, so the rcode should be * "no error". */ if (message->rcode != dns_rcode_noerror) { log_formerr(fctx, "CNAME/DNAME chain complete, but RCODE " "indicates error"); return (DNS_R_FORMERR); } /* * Examine the authority section (if there is one). * * We expect there to be only one owner name for all the rdatasets * in this section, and we expect that it is not external. */ done = ISC_FALSE; ns_name = NULL; ns_rdataset = NULL; result = dns_message_firstname(message, DNS_SECTION_AUTHORITY); while (!done && result == ISC_R_SUCCESS) { name = NULL; dns_message_currentname(message, DNS_SECTION_AUTHORITY, &name); external = ISC_TF(!dns_name_issubdomain(name, &fctx->domain)); if (!external) { /* * We expect to find NS or SIG NS rdatasets, and * nothing else. */ for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { if (rdataset->type == dns_rdatatype_ns || (rdataset->type == dns_rdatatype_rrsig && rdataset->covers == dns_rdatatype_ns)) { name->attributes |= DNS_NAMEATTR_CACHE; rdataset->attributes |= DNS_RDATASETATTR_CACHE; if (aa && !chaining) rdataset->trust = dns_trust_authauthority; else rdataset->trust = dns_trust_additional; if (rdataset->type == dns_rdatatype_ns) { ns_name = name; ns_rdataset = rdataset; } /* * Mark any additional data related * to this rdataset. */ (void)dns_rdataset_additionaldata( rdataset, check_related, fctx); done = ISC_TRUE; } } } result = dns_message_nextname(message, DNS_SECTION_AUTHORITY); } if (result == ISC_R_NOMORE) result = ISC_R_SUCCESS; log_ns_ttl(fctx, "answer_response"); if (ns_rdataset != NULL && dns_name_equal(&fctx->domain, ns_name) && !dns_name_equal(ns_name, dns_rootname)) trim_ns_ttl(fctx, ns_name, ns_rdataset); return (result); } static isc_boolean_t fctx_decreference(fetchctx_t *fctx) { isc_boolean_t bucket_empty = ISC_FALSE; INSIST(fctx->references > 0); fctx->references--; if (fctx->references == 0) { /* * No one cares about the result of this fetch anymore. */ if (fctx->pending == 0 && fctx->nqueries == 0 && ISC_LIST_EMPTY(fctx->validators) && SHUTTINGDOWN(fctx)) { /* * This fctx is already shutdown; we were just * waiting for the last reference to go away. */ bucket_empty = fctx_unlink(fctx); fctx_destroy(fctx); } else { /* * Initiate shutdown. */ fctx_shutdown(fctx); } } return (bucket_empty); } static void resume_dslookup(isc_task_t *task, isc_event_t *event) { dns_fetchevent_t *fevent; dns_resolver_t *res; fetchctx_t *fctx; isc_result_t result; isc_boolean_t bucket_empty; isc_boolean_t locked = ISC_FALSE; unsigned int bucketnum; dns_rdataset_t nameservers; dns_fixedname_t fixed; dns_name_t *domain; REQUIRE(event->ev_type == DNS_EVENT_FETCHDONE); fevent = (dns_fetchevent_t *)event; fctx = event->ev_arg; REQUIRE(VALID_FCTX(fctx)); res = fctx->res; UNUSED(task); FCTXTRACE("resume_dslookup"); if (fevent->node != NULL) dns_db_detachnode(fevent->db, &fevent->node); if (fevent->db != NULL) dns_db_detach(&fevent->db); dns_rdataset_init(&nameservers); bucketnum = fctx->bucketnum; if (fevent->result == ISC_R_CANCELED) { dns_resolver_destroyfetch(&fctx->nsfetch); fctx_done(fctx, ISC_R_CANCELED, __LINE__); } else if (fevent->result == ISC_R_SUCCESS) { FCTXTRACE("resuming DS lookup"); dns_resolver_destroyfetch(&fctx->nsfetch); if (dns_rdataset_isassociated(&fctx->nameservers)) dns_rdataset_disassociate(&fctx->nameservers); dns_rdataset_clone(fevent->rdataset, &fctx->nameservers); fctx->ns_ttl = fctx->nameservers.ttl; fctx->ns_ttl_ok = ISC_TRUE; log_ns_ttl(fctx, "resume_dslookup"); dns_name_free(&fctx->domain, fctx->mctx); dns_name_init(&fctx->domain, NULL); result = dns_name_dup(&fctx->nsname, fctx->mctx, &fctx->domain); if (result != ISC_R_SUCCESS) { fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); goto cleanup; } /* * Try again. */ fctx_try(fctx, ISC_TRUE, ISC_FALSE); } else { unsigned int n; dns_rdataset_t *nsrdataset = NULL; /* * Retrieve state from fctx->nsfetch before we destroy it. */ dns_fixedname_init(&fixed); domain = dns_fixedname_name(&fixed); dns_name_copy(&fctx->nsfetch->private->domain, domain, NULL); if (dns_name_equal(&fctx->nsname, domain)) { fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); dns_resolver_destroyfetch(&fctx->nsfetch); goto cleanup; } if (dns_rdataset_isassociated( &fctx->nsfetch->private->nameservers)) { dns_rdataset_clone( &fctx->nsfetch->private->nameservers, &nameservers); nsrdataset = &nameservers; } else domain = NULL; dns_resolver_destroyfetch(&fctx->nsfetch); n = dns_name_countlabels(&fctx->nsname); dns_name_getlabelsequence(&fctx->nsname, 1, n - 1, &fctx->nsname); if (dns_rdataset_isassociated(fevent->rdataset)) dns_rdataset_disassociate(fevent->rdataset); FCTXTRACE("continuing to look for parent's NS records"); result = dns_resolver_createfetch(fctx->res, &fctx->nsname, dns_rdatatype_ns, domain, nsrdataset, NULL, 0, task, resume_dslookup, fctx, &fctx->nsrrset, NULL, &fctx->nsfetch); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else { LOCK(&res->buckets[bucketnum].lock); locked = ISC_TRUE; fctx->references++; } } cleanup: if (dns_rdataset_isassociated(&nameservers)) dns_rdataset_disassociate(&nameservers); if (dns_rdataset_isassociated(fevent->rdataset)) dns_rdataset_disassociate(fevent->rdataset); INSIST(fevent->sigrdataset == NULL); isc_event_free(&event); if (!locked) LOCK(&res->buckets[bucketnum].lock); bucket_empty = fctx_decreference(fctx); UNLOCK(&res->buckets[bucketnum].lock); if (bucket_empty) empty_bucket(res); } static inline void checknamessection(dns_message_t *message, dns_section_t section) { isc_result_t result; dns_name_t *name; dns_rdata_t rdata = DNS_RDATA_INIT; dns_rdataset_t *rdataset; for (result = dns_message_firstname(message, section); result == ISC_R_SUCCESS; result = dns_message_nextname(message, section)) { name = NULL; dns_message_currentname(message, section, &name); for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) { for (result = dns_rdataset_first(rdataset); result == ISC_R_SUCCESS; result = dns_rdataset_next(rdataset)) { dns_rdataset_current(rdataset, &rdata); if (!dns_rdata_checkowner(name, rdata.rdclass, rdata.type, ISC_FALSE) || !dns_rdata_checknames(&rdata, name, NULL)) { rdataset->attributes |= DNS_RDATASETATTR_CHECKNAMES; } dns_rdata_reset(&rdata); } } } } static void checknames(dns_message_t *message) { checknamessection(message, DNS_SECTION_ANSWER); checknamessection(message, DNS_SECTION_AUTHORITY); checknamessection(message, DNS_SECTION_ADDITIONAL); } /* * Log server NSID at log level 'level' */ static void log_nsid(isc_buffer_t *opt, size_t nsid_len, resquery_t *query, int level, isc_mem_t *mctx) { static const char hex[17] = "0123456789abcdef"; char addrbuf[ISC_SOCKADDR_FORMATSIZE]; isc_uint16_t buflen, i; unsigned char *p, *buf, *nsid; /* Allocate buffer for storing hex version of the NSID */ buflen = (isc_uint16_t)nsid_len * 2 + 1; buf = isc_mem_get(mctx, buflen); if (buf == NULL) return; /* Convert to hex */ p = buf; nsid = isc_buffer_current(opt); for (i = 0; i < nsid_len; i++) { *p++ = hex[(nsid[0] >> 4) & 0xf]; *p++ = hex[nsid[0] & 0xf]; nsid++; } *p = '\0'; isc_sockaddr_format(&query->addrinfo->sockaddr, addrbuf, sizeof(addrbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, level, "received NSID '%s' from %s", buf, addrbuf); /* Clean up */ isc_mem_put(mctx, buf, buflen); return; } static void log_packet(dns_message_t *message, int level, isc_mem_t *mctx) { isc_buffer_t buffer; char *buf = NULL; int len = 1024; isc_result_t result; if (! isc_log_wouldlog(dns_lctx, level)) return; /* * Note that these are multiline debug messages. We want a newline * to appear in the log after each message. */ do { buf = isc_mem_get(mctx, len); if (buf == NULL) break; isc_buffer_init(&buffer, buf, len); result = dns_message_totext(message, &dns_master_style_debug, 0, &buffer); if (result == ISC_R_NOSPACE) { isc_mem_put(mctx, buf, len); len += 1024; } else if (result == ISC_R_SUCCESS) isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, level, "received packet:\n%.*s", (int)isc_buffer_usedlength(&buffer), buf); } while (result == ISC_R_NOSPACE); if (buf != NULL) isc_mem_put(mctx, buf, len); } static isc_boolean_t iscname(fetchctx_t *fctx) { isc_result_t result; result = dns_message_findname(fctx->rmessage, DNS_SECTION_ANSWER, &fctx->name, dns_rdatatype_cname, 0, NULL, NULL); return (result == ISC_R_SUCCESS ? ISC_TRUE : ISC_FALSE); } static isc_boolean_t betterreferral(fetchctx_t *fctx) { isc_result_t result; dns_name_t *name; dns_rdataset_t *rdataset; dns_message_t *message = fctx->rmessage; for (result = dns_message_firstname(message, DNS_SECTION_AUTHORITY); result == ISC_R_SUCCESS; result = dns_message_nextname(message, DNS_SECTION_AUTHORITY)) { name = NULL; dns_message_currentname(message, DNS_SECTION_AUTHORITY, &name); if (!isstrictsubdomain(name, &fctx->domain)) continue; for (rdataset = ISC_LIST_HEAD(name->list); rdataset != NULL; rdataset = ISC_LIST_NEXT(rdataset, link)) if (rdataset->type == dns_rdatatype_ns) return (ISC_TRUE); } return (ISC_FALSE); } static void process_opt(resquery_t *query, dns_rdataset_t *opt) { dns_rdata_t rdata; isc_buffer_t optbuf; isc_result_t result; isc_uint16_t optcode; isc_uint16_t optlen; result = dns_rdataset_first(opt); if (result == ISC_R_SUCCESS) { dns_rdata_init(&rdata); dns_rdataset_current(opt, &rdata); isc_buffer_init(&optbuf, rdata.data, rdata.length); isc_buffer_add(&optbuf, rdata.length); while (isc_buffer_remaininglength(&optbuf) >= 4) { optcode = isc_buffer_getuint16(&optbuf); optlen = isc_buffer_getuint16(&optbuf); INSIST(optlen <= isc_buffer_remaininglength(&optbuf)); switch (optcode) { case DNS_OPT_NSID: if (query->options & DNS_FETCHOPT_WANTNSID) log_nsid(&optbuf, optlen, query, ISC_LOG_INFO, query->fctx->res->mctx); isc_buffer_forward(&optbuf, optlen); break; default: isc_buffer_forward(&optbuf, optlen); break; } } INSIST(isc_buffer_remaininglength(&optbuf) == 0U); } } static void resquery_response(isc_task_t *task, isc_event_t *event) { isc_result_t result = ISC_R_SUCCESS; resquery_t *query = event->ev_arg; dns_dispatchevent_t *devent = (dns_dispatchevent_t *)event; isc_boolean_t keep_trying, get_nameservers, resend; isc_boolean_t truncated; dns_message_t *message; dns_rdataset_t *opt; fetchctx_t *fctx; dns_name_t *fname; dns_fixedname_t foundname; isc_stdtime_t now; isc_time_t tnow, *finish; dns_adbaddrinfo_t *addrinfo; unsigned int options; unsigned int findoptions; isc_result_t broken_server; badnstype_t broken_type = badns_response; isc_boolean_t no_response; REQUIRE(VALID_QUERY(query)); fctx = query->fctx; options = query->options; REQUIRE(VALID_FCTX(fctx)); REQUIRE(event->ev_type == DNS_EVENT_DISPATCH); QTRACE("response"); if (isc_sockaddr_pf(&query->addrinfo->sockaddr) == PF_INET) inc_stats(fctx->res, dns_resstatscounter_responsev4); else inc_stats(fctx->res, dns_resstatscounter_responsev6); (void)isc_timer_touch(fctx->timer); keep_trying = ISC_FALSE; broken_server = ISC_R_SUCCESS; get_nameservers = ISC_FALSE; resend = ISC_FALSE; truncated = ISC_FALSE; finish = NULL; no_response = ISC_FALSE; if (fctx->res->exiting) { result = ISC_R_SHUTTINGDOWN; goto done; } fctx->timeouts = 0; fctx->timeout = ISC_FALSE; fctx->addrinfo = query->addrinfo; /* * XXXRTH We should really get the current time just once. We * need a routine to convert from an isc_time_t to an * isc_stdtime_t. */ TIME_NOW(&tnow); finish = &tnow; isc_stdtime_get(&now); /* * Did the dispatcher have a problem? */ if (devent->result != ISC_R_SUCCESS) { if (devent->result == ISC_R_EOF && (query->options & DNS_FETCHOPT_NOEDNS0) == 0) { /* * The problem might be that they * don't understand EDNS0. Turn it * off and try again. */ options |= DNS_FETCHOPT_NOEDNS0; resend = ISC_TRUE; add_bad_edns(fctx, &query->addrinfo->sockaddr); } else { /* * There's no hope for this query. */ keep_trying = ISC_TRUE; /* * If this is a network error on an exclusive query * socket, mark the server as bad so that we won't try * it for this fetch again. Also adjust finish and * no_response so that we penalize this address in SRTT * adjustment later. */ if (query->exclusivesocket && (devent->result == ISC_R_HOSTUNREACH || devent->result == ISC_R_NETUNREACH || devent->result == ISC_R_CONNREFUSED || devent->result == ISC_R_CANCELED)) { broken_server = devent->result; broken_type = badns_unreachable; finish = NULL; no_response = ISC_TRUE; } } goto done; } message = fctx->rmessage; if (query->tsig != NULL) { result = dns_message_setquerytsig(message, query->tsig); if (result != ISC_R_SUCCESS) goto done; } if (query->tsigkey) { result = dns_message_settsigkey(message, query->tsigkey); if (result != ISC_R_SUCCESS) goto done; } dns_message_setclass(message, fctx->res->rdclass); result = dns_message_parse(message, &devent->buffer, 0); if (result != ISC_R_SUCCESS) { switch (result) { case ISC_R_UNEXPECTEDEND: if (!message->question_ok || (message->flags & DNS_MESSAGEFLAG_TC) == 0 || (options & DNS_FETCHOPT_TCP) != 0) { /* * Either the message ended prematurely, * and/or wasn't marked as being truncated, * and/or this is a response to a query we * sent over TCP. In all of these cases, * something is wrong with the remote * server and we don't want to retry using * TCP. */ if ((query->options & DNS_FETCHOPT_NOEDNS0) == 0) { /* * The problem might be that they * don't understand EDNS0. Turn it * off and try again. */ options |= DNS_FETCHOPT_NOEDNS0; resend = ISC_TRUE; add_bad_edns(fctx, &query->addrinfo->sockaddr); inc_stats(fctx->res, dns_resstatscounter_edns0fail); } else { broken_server = result; keep_trying = ISC_TRUE; } goto done; } /* * We defer retrying via TCP for a bit so we can * check out this message further. */ truncated = ISC_TRUE; break; case DNS_R_FORMERR: if ((query->options & DNS_FETCHOPT_NOEDNS0) == 0) { /* * The problem might be that they * don't understand EDNS0. Turn it * off and try again. */ options |= DNS_FETCHOPT_NOEDNS0; resend = ISC_TRUE; add_bad_edns(fctx, &query->addrinfo->sockaddr); inc_stats(fctx->res, dns_resstatscounter_edns0fail); } else { broken_server = DNS_R_UNEXPECTEDRCODE; keep_trying = ISC_TRUE; } goto done; default: /* * Something bad has happened. */ goto done; } } /* * Log the incoming packet. */ log_packet(message, ISC_LOG_DEBUG(10), fctx->res->mctx); if (message->rdclass != fctx->res->rdclass) { resend = ISC_TRUE; FCTXTRACE("bad class"); goto done; } /* * Process receive opt record. */ opt = dns_message_getopt(message); if (opt != NULL) process_opt(query, opt); /* * If the message is signed, check the signature. If not, this * returns success anyway. */ result = dns_message_checksig(message, fctx->res->view); if (result != ISC_R_SUCCESS) goto done; /* * The dispatcher should ensure we only get responses with QR set. */ INSIST((message->flags & DNS_MESSAGEFLAG_QR) != 0); /* * INSIST() that the message comes from the place we sent it to, * since the dispatch code should ensure this. * * INSIST() that the message id is correct (this should also be * ensured by the dispatch code). */ /* * We have an affirmative response to the query and we have * previously got a response from this server which indicated * EDNS may not be supported so we can now cache the lack of * EDNS support. */ if (opt == NULL && (message->rcode == dns_rcode_noerror || message->rcode == dns_rcode_nxdomain || message->rcode == dns_rcode_refused || message->rcode == dns_rcode_yxdomain) && bad_edns(fctx, &query->addrinfo->sockaddr)) { char addrbuf[ISC_SOCKADDR_FORMATSIZE]; isc_sockaddr_format(&query->addrinfo->sockaddr, addrbuf, sizeof(addrbuf)); dns_adb_changeflags(fctx->adb, query->addrinfo, DNS_FETCHOPT_NOEDNS0, DNS_FETCHOPT_NOEDNS0); } /* * Deal with truncated responses by retrying using TCP. */ if ((message->flags & DNS_MESSAGEFLAG_TC) != 0) truncated = ISC_TRUE; if (truncated) { inc_stats(fctx->res, dns_resstatscounter_truncated); if ((options & DNS_FETCHOPT_TCP) != 0) { broken_server = DNS_R_TRUNCATEDTCP; keep_trying = ISC_TRUE; } else { options |= DNS_FETCHOPT_TCP; resend = ISC_TRUE; } goto done; } /* * Is it a query response? */ if (message->opcode != dns_opcode_query) { /* XXXRTH Log */ broken_server = DNS_R_UNEXPECTEDOPCODE; keep_trying = ISC_TRUE; goto done; } /* * Update statistics about erroneous responses. */ if (message->rcode != dns_rcode_noerror) { switch (message->rcode) { case dns_rcode_nxdomain: inc_stats(fctx->res, dns_resstatscounter_nxdomain); break; case dns_rcode_servfail: inc_stats(fctx->res, dns_resstatscounter_servfail); break; case dns_rcode_formerr: inc_stats(fctx->res, dns_resstatscounter_formerr); break; default: inc_stats(fctx->res, dns_resstatscounter_othererror); break; } } /* * Is the remote server broken, or does it dislike us? */ if (message->rcode != dns_rcode_noerror && message->rcode != dns_rcode_nxdomain) { if (((message->rcode == dns_rcode_formerr || message->rcode == dns_rcode_notimp) || (message->rcode == dns_rcode_servfail && dns_message_getopt(message) == NULL)) && (query->options & DNS_FETCHOPT_NOEDNS0) == 0) { /* * It's very likely they don't like EDNS0. * If the response code is SERVFAIL, also check if the * response contains an OPT RR and don't cache the * failure since it can be returned for various other * reasons. * * XXXRTH We should check if the question * we're asking requires EDNS0, and * if so, we should bail out. */ options |= DNS_FETCHOPT_NOEDNS0; resend = ISC_TRUE; /* * Remember that they may not like EDNS0. */ add_bad_edns(fctx, &query->addrinfo->sockaddr); inc_stats(fctx->res, dns_resstatscounter_edns0fail); } else if (message->rcode == dns_rcode_formerr) { if (ISFORWARDER(query->addrinfo)) { /* * This forwarder doesn't understand us, * but other forwarders might. Keep trying. */ broken_server = DNS_R_REMOTEFORMERR; keep_trying = ISC_TRUE; } else { /* * The server doesn't understand us. Since * all servers for a zone need similar * capabilities, we assume that we will get * FORMERR from all servers, and thus we * cannot make any more progress with this * fetch. */ log_formerr(fctx, "server sent FORMERR"); result = DNS_R_FORMERR; } } else if (message->rcode == dns_rcode_yxdomain) { /* * DNAME mapping failed because the new name * was too long. There's no chance of success * for this fetch. */ result = DNS_R_YXDOMAIN; } else if (message->rcode == dns_rcode_badvers) { unsigned int flags, mask; unsigned int version; resend = ISC_TRUE; INSIST(opt != NULL); version = (opt->ttl >> 16) & 0xff; flags = (version << DNS_FETCHOPT_EDNSVERSIONSHIFT) | DNS_FETCHOPT_EDNSVERSIONSET; mask = DNS_FETCHOPT_EDNSVERSIONMASK | DNS_FETCHOPT_EDNSVERSIONSET; switch (version) { case 0: dns_adb_changeflags(fctx->adb, query->addrinfo, flags, mask); break; default: broken_server = DNS_R_BADVERS; keep_trying = ISC_TRUE; break; } } else { /* * XXXRTH log. */ broken_server = DNS_R_UNEXPECTEDRCODE; INSIST(broken_server != ISC_R_SUCCESS); keep_trying = ISC_TRUE; } goto done; } /* * Is the question the same as the one we asked? */ result = same_question(fctx); if (result != ISC_R_SUCCESS) { /* XXXRTH Log */ if (result == DNS_R_FORMERR) keep_trying = ISC_TRUE; goto done; } /* * Is the server lame? */ if (fctx->res->lame_ttl != 0 && !ISFORWARDER(query->addrinfo) && is_lame(fctx)) { inc_stats(fctx->res, dns_resstatscounter_lame); log_lame(fctx, query->addrinfo); result = dns_adb_marklame(fctx->adb, query->addrinfo, &fctx->name, fctx->type, now + fctx->res->lame_ttl); if (result != ISC_R_SUCCESS) isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_ERROR, "could not mark server as lame: %s", isc_result_totext(result)); broken_server = DNS_R_LAME; keep_trying = ISC_TRUE; goto done; } /* * Enforce delegations only zones like NET and COM. */ if (!ISFORWARDER(query->addrinfo) && dns_view_isdelegationonly(fctx->res->view, &fctx->domain) && !dns_name_equal(&fctx->domain, &fctx->name) && fix_mustbedelegationornxdomain(message, fctx)) { char namebuf[DNS_NAME_FORMATSIZE]; char domainbuf[DNS_NAME_FORMATSIZE]; char addrbuf[ISC_SOCKADDR_FORMATSIZE]; char classbuf[64]; char typebuf[64]; dns_name_format(&fctx->name, namebuf, sizeof(namebuf)); dns_name_format(&fctx->domain, domainbuf, sizeof(domainbuf)); dns_rdatatype_format(fctx->type, typebuf, sizeof(typebuf)); dns_rdataclass_format(fctx->res->rdclass, classbuf, sizeof(classbuf)); isc_sockaddr_format(&query->addrinfo->sockaddr, addrbuf, sizeof(addrbuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_DELEGATION_ONLY, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "enforced delegation-only for '%s' (%s/%s/%s) " "from %s", domainbuf, namebuf, typebuf, classbuf, addrbuf); } if ((fctx->res->options & DNS_RESOLVER_CHECKNAMES) != 0) checknames(message); /* * Clear cache bits. */ fctx->attributes &= ~(FCTX_ATTR_WANTNCACHE | FCTX_ATTR_WANTCACHE); /* * Did we get any answers? */ if (message->counts[DNS_SECTION_ANSWER] > 0 && (message->rcode == dns_rcode_noerror || message->rcode == dns_rcode_nxdomain)) { /* * [normal case] * We've got answers. If it has an authoritative answer or an * answer from a forwarder, we're done. */ if ((message->flags & DNS_MESSAGEFLAG_AA) != 0 || ISFORWARDER(query->addrinfo)) result = answer_response(fctx); else if (iscname(fctx) && fctx->type != dns_rdatatype_any && fctx->type != dns_rdatatype_cname) { /* * A BIND8 server could return a non-authoritative * answer when a CNAME is followed. We should treat * it as a valid answer. */ result = answer_response(fctx); } else if (fctx->type != dns_rdatatype_ns && !betterreferral(fctx)) { /* * Lame response !!!. */ result = answer_response(fctx); } else { if (fctx->type == dns_rdatatype_ns) { /* * A BIND 8 server could incorrectly return a * non-authoritative answer to an NS query * instead of a referral. Since this answer * lacks the SIGs necessary to do DNSSEC * validation, we must invoke the following * special kludge to treat it as a referral. */ result = noanswer_response(fctx, NULL, LOOK_FOR_NS_IN_ANSWER); } else { /* * Some other servers may still somehow include * an answer when it should return a referral * with an empty answer. Check to see if we can * treat this as a referral by ignoring the * answer. Further more, there may be an * implementation that moves A/AAAA glue records * to the answer section for that type of * delegation when the query is for that glue * record. LOOK_FOR_GLUE_IN_ANSWER will handle * such a corner case. */ result = noanswer_response(fctx, NULL, LOOK_FOR_GLUE_IN_ANSWER); } if (result != DNS_R_DELEGATION) { /* * At this point, AA is not set, the response * is not a referral, and the server is not a * forwarder. It is technically lame and it's * easier to treat it as such than to figure out * some more elaborate course of action. */ broken_server = DNS_R_LAME; keep_trying = ISC_TRUE; goto done; } goto force_referral; } if (result != ISC_R_SUCCESS) { if (result == DNS_R_FORMERR) keep_trying = ISC_TRUE; goto done; } } else if (message->counts[DNS_SECTION_AUTHORITY] > 0 || message->rcode == dns_rcode_noerror || message->rcode == dns_rcode_nxdomain) { /* * NXDOMAIN, NXRDATASET, or referral. */ result = noanswer_response(fctx, NULL, 0); switch (result) { case ISC_R_SUCCESS: case DNS_R_CHASEDSSERVERS: break; case DNS_R_DELEGATION: force_referral: /* * We don't have the answer, but we know a better * place to look. */ get_nameservers = ISC_TRUE; keep_trying = ISC_TRUE; /* * We have a new set of name servers, and it * has not experienced any restarts yet. */ fctx->restarts = 0; /* * Update local statistics counters collected for each * new zone. */ fctx->referrals++; fctx->querysent = 0; fctx->lamecount = 0; fctx->neterr = 0; fctx->badresp = 0; fctx->adberr = 0; result = ISC_R_SUCCESS; break; default: /* * Something has gone wrong. */ if (result == DNS_R_FORMERR) keep_trying = ISC_TRUE; goto done; } } else { /* * The server is insane. */ /* XXXRTH Log */ broken_server = DNS_R_UNEXPECTEDRCODE; keep_trying = ISC_TRUE; goto done; } /* * Follow additional section data chains. */ chase_additional(fctx); /* * Cache the cacheable parts of the message. This may also cause * work to be queued to the DNSSEC validator. */ if (WANTCACHE(fctx)) { result = cache_message(fctx, query->addrinfo, now); if (result != ISC_R_SUCCESS) goto done; } /* * Ncache the negatively cacheable parts of the message. This may * also cause work to be queued to the DNSSEC validator. */ if (WANTNCACHE(fctx)) { dns_rdatatype_t covers; if (message->rcode == dns_rcode_nxdomain) covers = dns_rdatatype_any; else covers = fctx->type; /* * Cache any negative cache entries in the message. */ result = ncache_message(fctx, query->addrinfo, covers, now); } done: /* * Remember the query's addrinfo, in case we need to mark the * server as broken. */ addrinfo = query->addrinfo; /* * Cancel the query. * * XXXRTH Don't cancel the query if waiting for validation? */ fctx_cancelquery(&query, &devent, finish, no_response); if (keep_trying) { if (result == DNS_R_FORMERR) broken_server = DNS_R_FORMERR; if (broken_server != ISC_R_SUCCESS) { /* * Add this server to the list of bad servers for * this fctx. */ add_bad(fctx, addrinfo, broken_server, broken_type); } if (get_nameservers) { dns_name_t *name; dns_fixedname_init(&foundname); fname = dns_fixedname_name(&foundname); if (result != ISC_R_SUCCESS) { fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); return; } findoptions = 0; if (dns_rdatatype_atparent(fctx->type)) findoptions |= DNS_DBFIND_NOEXACT; if ((options & DNS_FETCHOPT_UNSHARED) == 0) name = &fctx->name; else name = &fctx->domain; result = dns_view_findzonecut(fctx->res->view, name, fname, now, findoptions, ISC_TRUE, &fctx->nameservers, NULL); if (result != ISC_R_SUCCESS) { FCTXTRACE("couldn't find a zonecut"); fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); return; } if (!dns_name_issubdomain(fname, &fctx->domain)) { /* * The best nameservers are now above our * QDOMAIN. */ FCTXTRACE("nameservers now above QDOMAIN"); fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); return; } dns_name_free(&fctx->domain, fctx->mctx); dns_name_init(&fctx->domain, NULL); result = dns_name_dup(fname, fctx->mctx, &fctx->domain); if (result != ISC_R_SUCCESS) { fctx_done(fctx, DNS_R_SERVFAIL, __LINE__); return; } fctx->ns_ttl = fctx->nameservers.ttl; fctx->ns_ttl_ok = ISC_TRUE; fctx_cancelqueries(fctx, ISC_TRUE); fctx_cleanupfinds(fctx); fctx_cleanupaltfinds(fctx); fctx_cleanupforwaddrs(fctx); fctx_cleanupaltaddrs(fctx); } /* * Try again. */ fctx_try(fctx, !get_nameservers, ISC_FALSE); } else if (resend) { /* * Resend (probably with changed options). */ FCTXTRACE("resend"); inc_stats(fctx->res, dns_resstatscounter_retry); result = fctx_query(fctx, addrinfo, options); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); } else if (result == ISC_R_SUCCESS && !HAVE_ANSWER(fctx)) { /* * All has gone well so far, but we are waiting for the * DNSSEC validator to validate the answer. */ FCTXTRACE("wait for validator"); fctx_cancelqueries(fctx, ISC_TRUE); /* * We must not retransmit while the validator is working; * it has references to the current rmessage. */ result = fctx_stopidletimer(fctx); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); } else if (result == DNS_R_CHASEDSSERVERS) { unsigned int n; add_bad(fctx, addrinfo, result, broken_type); fctx_cancelqueries(fctx, ISC_TRUE); fctx_cleanupfinds(fctx); fctx_cleanupforwaddrs(fctx); n = dns_name_countlabels(&fctx->name); dns_name_getlabelsequence(&fctx->name, 1, n - 1, &fctx->nsname); FCTXTRACE("suspending DS lookup to find parent's NS records"); result = dns_resolver_createfetch(fctx->res, &fctx->nsname, dns_rdatatype_ns, NULL, NULL, NULL, 0, task, resume_dslookup, fctx, &fctx->nsrrset, NULL, &fctx->nsfetch); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); else { LOCK(&fctx->res->buckets[fctx->bucketnum].lock); fctx->references++; UNLOCK(&fctx->res->buckets[fctx->bucketnum].lock); result = fctx_stopidletimer(fctx); if (result != ISC_R_SUCCESS) fctx_done(fctx, result, __LINE__); } } else { /* * We're done. */ fctx_done(fctx, result, __LINE__); } } /*** *** Resolver Methods ***/ static void destroy_badcache(dns_resolver_t *res) { dns_badcache_t *bad, *next; unsigned int i; if (res->badcache != NULL) { for (i = 0; i < res->badhash; i++) for (bad = res->badcache[i]; bad != NULL; bad = next) { next = bad->next; isc_mem_put(res->mctx, bad, sizeof(*bad) + bad->name.length); res->badcount--; } isc_mem_put(res->mctx, res->badcache, sizeof(*res->badcache) * res->badhash); res->badcache = NULL; res->badhash = 0; INSIST(res->badcount == 0); } } static void destroy(dns_resolver_t *res) { unsigned int i; alternate_t *a; REQUIRE(res->references == 0); REQUIRE(!res->priming); REQUIRE(res->primefetch == NULL); RTRACE("destroy"); INSIST(res->nfctx == 0); DESTROYLOCK(&res->primelock); DESTROYLOCK(&res->nlock); DESTROYLOCK(&res->lock); for (i = 0; i < res->nbuckets; i++) { INSIST(ISC_LIST_EMPTY(res->buckets[i].fctxs)); isc_task_shutdown(res->buckets[i].task); isc_task_detach(&res->buckets[i].task); DESTROYLOCK(&res->buckets[i].lock); isc_mem_detach(&res->buckets[i].mctx); } isc_mem_put(res->mctx, res->buckets, res->nbuckets * sizeof(fctxbucket_t)); if (res->dispatches4 != NULL) dns_dispatchset_destroy(&res->dispatches4); if (res->dispatches6 != NULL) dns_dispatchset_destroy(&res->dispatches6); while ((a = ISC_LIST_HEAD(res->alternates)) != NULL) { ISC_LIST_UNLINK(res->alternates, a, link); if (!a->isaddress) dns_name_free(&a->_u._n.name, res->mctx); isc_mem_put(res->mctx, a, sizeof(*a)); } dns_resolver_reset_algorithms(res); destroy_badcache(res); dns_resolver_resetmustbesecure(res); #if USE_ALGLOCK isc_rwlock_destroy(&res->alglock); #endif #if USE_MBSLOCK isc_rwlock_destroy(&res->mbslock); #endif isc_timer_detach(&res->spillattimer); res->magic = 0; isc_mem_put(res->mctx, res, sizeof(*res)); } static void send_shutdown_events(dns_resolver_t *res) { isc_event_t *event, *next_event; isc_task_t *etask; /* * Caller must be holding the resolver lock. */ for (event = ISC_LIST_HEAD(res->whenshutdown); event != NULL; event = next_event) { next_event = ISC_LIST_NEXT(event, ev_link); ISC_LIST_UNLINK(res->whenshutdown, event, ev_link); etask = event->ev_sender; event->ev_sender = res; isc_task_sendanddetach(&etask, &event); } } static void empty_bucket(dns_resolver_t *res) { RTRACE("empty_bucket"); LOCK(&res->lock); INSIST(res->activebuckets > 0); res->activebuckets--; if (res->activebuckets == 0) send_shutdown_events(res); UNLOCK(&res->lock); } static void spillattimer_countdown(isc_task_t *task, isc_event_t *event) { dns_resolver_t *res = event->ev_arg; isc_result_t result; unsigned int count; isc_boolean_t logit = ISC_FALSE; REQUIRE(VALID_RESOLVER(res)); UNUSED(task); LOCK(&res->lock); INSIST(!res->exiting); if (res->spillat > res->spillatmin) { res->spillat--; logit = ISC_TRUE; } if (res->spillat <= res->spillatmin) { result = isc_timer_reset(res->spillattimer, isc_timertype_inactive, NULL, NULL, ISC_TRUE); RUNTIME_CHECK(result == ISC_R_SUCCESS); } count = res->spillat; UNLOCK(&res->lock); if (logit) isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, ISC_LOG_NOTICE, "clients-per-query decreased to %u", count); isc_event_free(&event); } isc_result_t dns_resolver_create(dns_view_t *view, isc_taskmgr_t *taskmgr, unsigned int ntasks, unsigned int ndisp, isc_socketmgr_t *socketmgr, isc_timermgr_t *timermgr, unsigned int options, dns_dispatchmgr_t *dispatchmgr, dns_dispatch_t *dispatchv4, dns_dispatch_t *dispatchv6, dns_resolver_t **resp) { dns_resolver_t *res; isc_result_t result = ISC_R_SUCCESS; unsigned int i, buckets_created = 0; isc_task_t *task = NULL; char name[16]; unsigned dispattr; /* * Create a resolver. */ REQUIRE(DNS_VIEW_VALID(view)); REQUIRE(ntasks > 0); REQUIRE(ndisp > 0); REQUIRE(resp != NULL && *resp == NULL); REQUIRE(dispatchmgr != NULL); REQUIRE(dispatchv4 != NULL || dispatchv6 != NULL); res = isc_mem_get(view->mctx, sizeof(*res)); if (res == NULL) return (ISC_R_NOMEMORY); RTRACE("create"); res->mctx = view->mctx; res->rdclass = view->rdclass; res->socketmgr = socketmgr; res->timermgr = timermgr; res->taskmgr = taskmgr; res->dispatchmgr = dispatchmgr; res->view = view; res->options = options; res->lame_ttl = 0; ISC_LIST_INIT(res->alternates); res->udpsize = RECV_BUFFER_SIZE; res->algorithms = NULL; res->badcache = NULL; res->badcount = 0; res->badhash = 0; res->badsweep = 0; res->mustbesecure = NULL; res->spillatmin = res->spillat = 10; res->spillatmax = 100; res->spillattimer = NULL; res->zero_no_soa_ttl = ISC_FALSE; res->query_timeout = DEFAULT_QUERY_TIMEOUT; res->maxdepth = DEFAULT_RECURSION_DEPTH; res->nbuckets = ntasks; res->activebuckets = ntasks; res->buckets = isc_mem_get(view->mctx, ntasks * sizeof(fctxbucket_t)); if (res->buckets == NULL) { result = ISC_R_NOMEMORY; goto cleanup_res; } for (i = 0; i < ntasks; i++) { result = isc_mutex_init(&res->buckets[i].lock); if (result != ISC_R_SUCCESS) goto cleanup_buckets; res->buckets[i].task = NULL; result = isc_task_create(taskmgr, 0, &res->buckets[i].task); if (result != ISC_R_SUCCESS) { DESTROYLOCK(&res->buckets[i].lock); goto cleanup_buckets; } res->buckets[i].mctx = NULL; snprintf(name, sizeof(name), "res%u", i); #ifdef ISC_PLATFORM_USETHREADS /* * Use a separate memory context for each bucket to reduce * contention among multiple threads. Do this only when * enabling threads because it will be require more memory. */ result = isc_mem_create(0, 0, &res->buckets[i].mctx); if (result != ISC_R_SUCCESS) { isc_task_detach(&res->buckets[i].task); DESTROYLOCK(&res->buckets[i].lock); goto cleanup_buckets; } isc_mem_setname(res->buckets[i].mctx, name, NULL); #else isc_mem_attach(view->mctx, &res->buckets[i].mctx); #endif isc_task_setname(res->buckets[i].task, name, res); ISC_LIST_INIT(res->buckets[i].fctxs); res->buckets[i].exiting = ISC_FALSE; buckets_created++; } res->dispatches4 = NULL; if (dispatchv4 != NULL) { dns_dispatchset_create(view->mctx, socketmgr, taskmgr, dispatchv4, &res->dispatches4, ndisp); dispattr = dns_dispatch_getattributes(dispatchv4); res->exclusivev4 = ISC_TF((dispattr & DNS_DISPATCHATTR_EXCLUSIVE) != 0); } res->dispatches6 = NULL; if (dispatchv6 != NULL) { dns_dispatchset_create(view->mctx, socketmgr, taskmgr, dispatchv6, &res->dispatches6, ndisp); dispattr = dns_dispatch_getattributes(dispatchv6); res->exclusivev6 = ISC_TF((dispattr & DNS_DISPATCHATTR_EXCLUSIVE) != 0); } res->references = 1; res->exiting = ISC_FALSE; res->frozen = ISC_FALSE; ISC_LIST_INIT(res->whenshutdown); res->priming = ISC_FALSE; res->primefetch = NULL; res->nfctx = 0; result = isc_mutex_init(&res->lock); if (result != ISC_R_SUCCESS) goto cleanup_dispatches; result = isc_mutex_init(&res->nlock); if (result != ISC_R_SUCCESS) goto cleanup_lock; result = isc_mutex_init(&res->primelock); if (result != ISC_R_SUCCESS) goto cleanup_nlock; task = NULL; result = isc_task_create(taskmgr, 0, &task); if (result != ISC_R_SUCCESS) goto cleanup_primelock; result = isc_timer_create(timermgr, isc_timertype_inactive, NULL, NULL, task, spillattimer_countdown, res, &res->spillattimer); isc_task_detach(&task); if (result != ISC_R_SUCCESS) goto cleanup_primelock; #if USE_ALGLOCK result = isc_rwlock_init(&res->alglock, 0, 0); if (result != ISC_R_SUCCESS) goto cleanup_spillattimer; #endif #if USE_MBSLOCK result = isc_rwlock_init(&res->mbslock, 0, 0); if (result != ISC_R_SUCCESS) goto cleanup_alglock; #endif res->magic = RES_MAGIC; *resp = res; return (ISC_R_SUCCESS); #if USE_MBSLOCK cleanup_alglock: #if USE_ALGLOCK isc_rwlock_destroy(&res->alglock); #endif #endif #if USE_ALGLOCK || USE_MBSLOCK cleanup_spillattimer: isc_timer_detach(&res->spillattimer); #endif cleanup_primelock: DESTROYLOCK(&res->primelock); cleanup_nlock: DESTROYLOCK(&res->nlock); cleanup_lock: DESTROYLOCK(&res->lock); cleanup_dispatches: if (res->dispatches6 != NULL) dns_dispatchset_destroy(&res->dispatches6); if (res->dispatches4 != NULL) dns_dispatchset_destroy(&res->dispatches4); cleanup_buckets: for (i = 0; i < buckets_created; i++) { isc_mem_detach(&res->buckets[i].mctx); DESTROYLOCK(&res->buckets[i].lock); isc_task_shutdown(res->buckets[i].task); isc_task_detach(&res->buckets[i].task); } isc_mem_put(view->mctx, res->buckets, res->nbuckets * sizeof(fctxbucket_t)); cleanup_res: isc_mem_put(view->mctx, res, sizeof(*res)); return (result); } #ifdef BIND9 static void prime_done(isc_task_t *task, isc_event_t *event) { dns_resolver_t *res; dns_fetchevent_t *fevent; dns_fetch_t *fetch; dns_db_t *db = NULL; REQUIRE(event->ev_type == DNS_EVENT_FETCHDONE); fevent = (dns_fetchevent_t *)event; res = event->ev_arg; REQUIRE(VALID_RESOLVER(res)); UNUSED(task); LOCK(&res->lock); INSIST(res->priming); res->priming = ISC_FALSE; LOCK(&res->primelock); fetch = res->primefetch; res->primefetch = NULL; UNLOCK(&res->primelock); UNLOCK(&res->lock); if (fevent->result == ISC_R_SUCCESS && res->view->cache != NULL && res->view->hints != NULL) { dns_cache_attachdb(res->view->cache, &db); dns_root_checkhints(res->view, res->view->hints, db); dns_db_detach(&db); } if (fevent->node != NULL) dns_db_detachnode(fevent->db, &fevent->node); if (fevent->db != NULL) dns_db_detach(&fevent->db); if (dns_rdataset_isassociated(fevent->rdataset)) dns_rdataset_disassociate(fevent->rdataset); INSIST(fevent->sigrdataset == NULL); isc_mem_put(res->mctx, fevent->rdataset, sizeof(*fevent->rdataset)); isc_event_free(&event); dns_resolver_destroyfetch(&fetch); } void dns_resolver_prime(dns_resolver_t *res) { isc_boolean_t want_priming = ISC_FALSE; dns_rdataset_t *rdataset; isc_result_t result; REQUIRE(VALID_RESOLVER(res)); REQUIRE(res->frozen); RTRACE("dns_resolver_prime"); LOCK(&res->lock); if (!res->exiting && !res->priming) { INSIST(res->primefetch == NULL); res->priming = ISC_TRUE; want_priming = ISC_TRUE; } UNLOCK(&res->lock); if (want_priming) { /* * To avoid any possible recursive locking problems, we * start the priming fetch like any other fetch, and holding * no resolver locks. No one else will try to start it * because we're the ones who set res->priming to true. * Any other callers of dns_resolver_prime() while we're * running will see that res->priming is already true and * do nothing. */ RTRACE("priming"); rdataset = isc_mem_get(res->mctx, sizeof(*rdataset)); if (rdataset == NULL) { LOCK(&res->lock); INSIST(res->priming); INSIST(res->primefetch == NULL); res->priming = ISC_FALSE; UNLOCK(&res->lock); return; } dns_rdataset_init(rdataset); LOCK(&res->primelock); result = dns_resolver_createfetch(res, dns_rootname, dns_rdatatype_ns, NULL, NULL, NULL, 0, res->buckets[0].task, prime_done, res, rdataset, NULL, &res->primefetch); UNLOCK(&res->primelock); if (result != ISC_R_SUCCESS) { LOCK(&res->lock); INSIST(res->priming); res->priming = ISC_FALSE; UNLOCK(&res->lock); } } } #endif /* BIND9 */ void dns_resolver_freeze(dns_resolver_t *res) { /* * Freeze resolver. */ REQUIRE(VALID_RESOLVER(res)); res->frozen = ISC_TRUE; } void dns_resolver_attach(dns_resolver_t *source, dns_resolver_t **targetp) { REQUIRE(VALID_RESOLVER(source)); REQUIRE(targetp != NULL && *targetp == NULL); RRTRACE(source, "attach"); LOCK(&source->lock); REQUIRE(!source->exiting); INSIST(source->references > 0); source->references++; INSIST(source->references != 0); UNLOCK(&source->lock); *targetp = source; } void dns_resolver_whenshutdown(dns_resolver_t *res, isc_task_t *task, isc_event_t **eventp) { isc_task_t *clone; isc_event_t *event; REQUIRE(VALID_RESOLVER(res)); REQUIRE(eventp != NULL); event = *eventp; *eventp = NULL; LOCK(&res->lock); if (res->exiting && res->activebuckets == 0) { /* * We're already shutdown. Send the event. */ event->ev_sender = res; isc_task_send(task, &event); } else { clone = NULL; isc_task_attach(task, &clone); event->ev_sender = clone; ISC_LIST_APPEND(res->whenshutdown, event, ev_link); } UNLOCK(&res->lock); } void dns_resolver_shutdown(dns_resolver_t *res) { unsigned int i; fetchctx_t *fctx; isc_result_t result; REQUIRE(VALID_RESOLVER(res)); RTRACE("shutdown"); LOCK(&res->lock); if (!res->exiting) { RTRACE("exiting"); res->exiting = ISC_TRUE; for (i = 0; i < res->nbuckets; i++) { LOCK(&res->buckets[i].lock); for (fctx = ISC_LIST_HEAD(res->buckets[i].fctxs); fctx != NULL; fctx = ISC_LIST_NEXT(fctx, link)) fctx_shutdown(fctx); if (res->dispatches4 != NULL && !res->exclusivev4) { dns_dispatchset_cancelall(res->dispatches4, res->buckets[i].task); } if (res->dispatches6 != NULL && !res->exclusivev6) { dns_dispatchset_cancelall(res->dispatches6, res->buckets[i].task); } res->buckets[i].exiting = ISC_TRUE; if (ISC_LIST_EMPTY(res->buckets[i].fctxs)) { INSIST(res->activebuckets > 0); res->activebuckets--; } UNLOCK(&res->buckets[i].lock); } if (res->activebuckets == 0) send_shutdown_events(res); result = isc_timer_reset(res->spillattimer, isc_timertype_inactive, NULL, NULL, ISC_TRUE); RUNTIME_CHECK(result == ISC_R_SUCCESS); } UNLOCK(&res->lock); } void dns_resolver_detach(dns_resolver_t **resp) { dns_resolver_t *res; isc_boolean_t need_destroy = ISC_FALSE; REQUIRE(resp != NULL); res = *resp; REQUIRE(VALID_RESOLVER(res)); RTRACE("detach"); LOCK(&res->lock); INSIST(res->references > 0); res->references--; if (res->references == 0) { INSIST(res->exiting && res->activebuckets == 0); need_destroy = ISC_TRUE; } UNLOCK(&res->lock); if (need_destroy) destroy(res); *resp = NULL; } static inline isc_boolean_t fctx_match(fetchctx_t *fctx, dns_name_t *name, dns_rdatatype_t type, unsigned int options) { /* * Don't match fetch contexts that are shutting down. */ if (fctx->cloned || fctx->state == fetchstate_done || ISC_LIST_EMPTY(fctx->events)) return (ISC_FALSE); if (fctx->type != type || fctx->options != options) return (ISC_FALSE); return (dns_name_equal(&fctx->name, name)); } static inline void log_fetch(dns_name_t *name, dns_rdatatype_t type) { char namebuf[DNS_NAME_FORMATSIZE]; char typebuf[DNS_RDATATYPE_FORMATSIZE]; int level = ISC_LOG_DEBUG(1); if (! isc_log_wouldlog(dns_lctx, level)) return; dns_name_format(name, namebuf, sizeof(namebuf)); dns_rdatatype_format(type, typebuf, sizeof(typebuf)); isc_log_write(dns_lctx, DNS_LOGCATEGORY_RESOLVER, DNS_LOGMODULE_RESOLVER, level, "createfetch: %s %s", namebuf, typebuf); } isc_result_t dns_resolver_createfetch(dns_resolver_t *res, dns_name_t *name, dns_rdatatype_t type, dns_name_t *domain, dns_rdataset_t *nameservers, dns_forwarders_t *forwarders, unsigned int options, isc_task_t *task, isc_taskaction_t action, void *arg, dns_rdataset_t *rdataset, dns_rdataset_t *sigrdataset, dns_fetch_t **fetchp) { return (dns_resolver_createfetch3(res, name, type, domain, nameservers, forwarders, NULL, 0, options, 0, task, action, arg, rdataset, sigrdataset, fetchp)); } isc_result_t dns_resolver_createfetch2(dns_resolver_t *res, dns_name_t *name, dns_rdatatype_t type, dns_name_t *domain, dns_rdataset_t *nameservers, dns_forwarders_t *forwarders, isc_sockaddr_t *client, dns_messageid_t id, unsigned int options, isc_task_t *task, isc_taskaction_t action, void *arg, dns_rdataset_t *rdataset, dns_rdataset_t *sigrdataset, dns_fetch_t **fetchp) { return (dns_resolver_createfetch3(res, name, type, domain, nameservers, forwarders, client, id, options, 0, task, action, arg, rdataset, sigrdataset, fetchp)); } isc_result_t dns_resolver_createfetch3(dns_resolver_t *res, dns_name_t *name, dns_rdatatype_t type, dns_name_t *domain, dns_rdataset_t *nameservers, dns_forwarders_t *forwarders, isc_sockaddr_t *client, dns_messageid_t id, unsigned int options, unsigned int depth, isc_task_t *task, isc_taskaction_t action, void *arg, dns_rdataset_t *rdataset, dns_rdataset_t *sigrdataset, dns_fetch_t **fetchp) { dns_fetch_t *fetch; fetchctx_t *fctx = NULL; isc_result_t result = ISC_R_SUCCESS; unsigned int bucketnum; isc_boolean_t new_fctx = ISC_FALSE; isc_event_t *event; unsigned int count = 0; unsigned int spillat; unsigned int spillatmin; isc_boolean_t destroy = ISC_FALSE; UNUSED(forwarders); REQUIRE(VALID_RESOLVER(res)); REQUIRE(res->frozen); /* XXXRTH Check for meta type */ if (domain != NULL) { REQUIRE(DNS_RDATASET_VALID(nameservers)); REQUIRE(nameservers->type == dns_rdatatype_ns); } else REQUIRE(nameservers == NULL); REQUIRE(forwarders == NULL); REQUIRE(!dns_rdataset_isassociated(rdataset)); REQUIRE(sigrdataset == NULL || !dns_rdataset_isassociated(sigrdataset)); REQUIRE(fetchp != NULL && *fetchp == NULL); log_fetch(name, type); /* * XXXRTH use a mempool? */ fetch = isc_mem_get(res->mctx, sizeof(*fetch)); if (fetch == NULL) return (ISC_R_NOMEMORY); bucketnum = dns_name_fullhash(name, ISC_FALSE) % res->nbuckets; LOCK(&res->lock); spillat = res->spillat; spillatmin = res->spillatmin; UNLOCK(&res->lock); LOCK(&res->buckets[bucketnum].lock); if (res->buckets[bucketnum].exiting) { result = ISC_R_SHUTTINGDOWN; goto unlock; } if ((options & DNS_FETCHOPT_UNSHARED) == 0) { for (fctx = ISC_LIST_HEAD(res->buckets[bucketnum].fctxs); fctx != NULL; fctx = ISC_LIST_NEXT(fctx, link)) { if (fctx_match(fctx, name, type, options)) break; } } /* * Is this a duplicate? */ if (fctx != NULL && client != NULL) { dns_fetchevent_t *fevent; for (fevent = ISC_LIST_HEAD(fctx->events); fevent != NULL; fevent = ISC_LIST_NEXT(fevent, ev_link)) { if (fevent->client != NULL && fevent->id == id && isc_sockaddr_equal(fevent->client, client)) { result = DNS_R_DUPLICATE; goto unlock; } count++; } } if (count >= spillatmin && spillatmin != 0) { INSIST(fctx != NULL); if (count >= spillat) fctx->spilled = ISC_TRUE; if (fctx->spilled) { result = DNS_R_DROP; goto unlock; } } if (fctx == NULL) { result = fctx_create(res, name, type, domain, nameservers, options, bucketnum, depth, &fctx); if (result != ISC_R_SUCCESS) goto unlock; new_fctx = ISC_TRUE; } else if (fctx->depth > depth) fctx->depth = depth; result = fctx_join(fctx, task, client, id, action, arg, rdataset, sigrdataset, fetch); if (new_fctx) { if (result == ISC_R_SUCCESS) { /* * Launch this fctx. */ event = &fctx->control_event; ISC_EVENT_INIT(event, sizeof(*event), 0, NULL, DNS_EVENT_FETCHCONTROL, fctx_start, fctx, NULL, NULL, NULL); isc_task_send(res->buckets[bucketnum].task, &event); } else { /* * We don't care about the result of fctx_unlink() * since we know we're not exiting. */ (void)fctx_unlink(fctx); destroy = ISC_TRUE; } } unlock: UNLOCK(&res->buckets[bucketnum].lock); if (destroy) fctx_destroy(fctx); if (result == ISC_R_SUCCESS) { FTRACE("created"); *fetchp = fetch; } else isc_mem_put(res->mctx, fetch, sizeof(*fetch)); return (result); } void dns_resolver_cancelfetch(dns_fetch_t *fetch) { fetchctx_t *fctx; dns_resolver_t *res; dns_fetchevent_t *event, *next_event; isc_task_t *etask; REQUIRE(DNS_FETCH_VALID(fetch)); fctx = fetch->private; REQUIRE(VALID_FCTX(fctx)); res = fctx->res; FTRACE("cancelfetch"); LOCK(&res->buckets[fctx->bucketnum].lock); /* * Find the completion event for this fetch (as opposed * to those for other fetches that have joined the same * fctx) and send it with result = ISC_R_CANCELED. */ event = NULL; if (fctx->state != fetchstate_done) { for (event = ISC_LIST_HEAD(fctx->events); event != NULL; event = next_event) { next_event = ISC_LIST_NEXT(event, ev_link); if (event->fetch == fetch) { ISC_LIST_UNLINK(fctx->events, event, ev_link); break; } } } if (event != NULL) { etask = event->ev_sender; event->ev_sender = fctx; event->result = ISC_R_CANCELED; isc_task_sendanddetach(&etask, ISC_EVENT_PTR(&event)); } /* * The fctx continues running even if no fetches remain; * the answer is still cached. */ UNLOCK(&res->buckets[fctx->bucketnum].lock); } void dns_resolver_destroyfetch(dns_fetch_t **fetchp) { dns_fetch_t *fetch; dns_resolver_t *res; dns_fetchevent_t *event, *next_event; fetchctx_t *fctx; unsigned int bucketnum; isc_boolean_t bucket_empty; REQUIRE(fetchp != NULL); fetch = *fetchp; REQUIRE(DNS_FETCH_VALID(fetch)); fctx = fetch->private; REQUIRE(VALID_FCTX(fctx)); res = fctx->res; FTRACE("destroyfetch"); bucketnum = fctx->bucketnum; LOCK(&res->buckets[bucketnum].lock); /* * Sanity check: the caller should have gotten its event before * trying to destroy the fetch. */ event = NULL; if (fctx->state != fetchstate_done) { for (event = ISC_LIST_HEAD(fctx->events); event != NULL; event = next_event) { next_event = ISC_LIST_NEXT(event, ev_link); RUNTIME_CHECK(event->fetch != fetch); } } bucket_empty = fctx_decreference(fctx); UNLOCK(&res->buckets[bucketnum].lock); isc_mem_put(res->mctx, fetch, sizeof(*fetch)); *fetchp = NULL; if (bucket_empty) empty_bucket(res); } void dns_resolver_logfetch(dns_fetch_t *fetch, isc_log_t *lctx, isc_logcategory_t *category, isc_logmodule_t *module, int level, isc_boolean_t duplicateok) { fetchctx_t *fctx; dns_resolver_t *res; char domainbuf[DNS_NAME_FORMATSIZE]; REQUIRE(DNS_FETCH_VALID(fetch)); fctx = fetch->private; REQUIRE(VALID_FCTX(fctx)); res = fctx->res; LOCK(&res->buckets[fctx->bucketnum].lock); INSIST(fctx->exitline >= 0); if (!fctx->logged || duplicateok) { dns_name_format(&fctx->domain, domainbuf, sizeof(domainbuf)); isc_log_write(lctx, category, module, level, "fetch completed at %s:%d for %s in " "%" ISC_PRINT_QUADFORMAT "u." "%06" ISC_PRINT_QUADFORMAT "u: %s/%s " "[domain:%s,referral:%u,restart:%u,qrysent:%u," "timeout:%u,lame:%u,neterr:%u,badresp:%u," "adberr:%u,findfail:%u,valfail:%u]", __FILE__, fctx->exitline, fctx->info, fctx->duration / US_PER_SEC, fctx->duration % US_PER_SEC, isc_result_totext(fctx->result), isc_result_totext(fctx->vresult), domainbuf, fctx->referrals, fctx->restarts, fctx->querysent, fctx->timeouts, fctx->lamecount, fctx->neterr, fctx->badresp, fctx->adberr, fctx->findfail, fctx->valfail); fctx->logged = ISC_TRUE; } UNLOCK(&res->buckets[fctx->bucketnum].lock); } dns_dispatchmgr_t * dns_resolver_dispatchmgr(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->dispatchmgr); } dns_dispatch_t * dns_resolver_dispatchv4(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (dns_dispatchset_get(resolver->dispatches4)); } dns_dispatch_t * dns_resolver_dispatchv6(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (dns_dispatchset_get(resolver->dispatches6)); } isc_socketmgr_t * dns_resolver_socketmgr(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->socketmgr); } isc_taskmgr_t * dns_resolver_taskmgr(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->taskmgr); } isc_uint32_t dns_resolver_getlamettl(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->lame_ttl); } void dns_resolver_setlamettl(dns_resolver_t *resolver, isc_uint32_t lame_ttl) { REQUIRE(VALID_RESOLVER(resolver)); resolver->lame_ttl = lame_ttl; } unsigned int dns_resolver_nrunning(dns_resolver_t *resolver) { unsigned int n; LOCK(&resolver->nlock); n = resolver->nfctx; UNLOCK(&resolver->nlock); return (n); } isc_result_t dns_resolver_addalternate(dns_resolver_t *resolver, isc_sockaddr_t *alt, dns_name_t *name, in_port_t port) { alternate_t *a; isc_result_t result; REQUIRE(VALID_RESOLVER(resolver)); REQUIRE(!resolver->frozen); REQUIRE((alt == NULL) ^ (name == NULL)); a = isc_mem_get(resolver->mctx, sizeof(*a)); if (a == NULL) return (ISC_R_NOMEMORY); if (alt != NULL) { a->isaddress = ISC_TRUE; a->_u.addr = *alt; } else { a->isaddress = ISC_FALSE; a->_u._n.port = port; dns_name_init(&a->_u._n.name, NULL); result = dns_name_dup(name, resolver->mctx, &a->_u._n.name); if (result != ISC_R_SUCCESS) { isc_mem_put(resolver->mctx, a, sizeof(*a)); return (result); } } ISC_LINK_INIT(a, link); ISC_LIST_APPEND(resolver->alternates, a, link); return (ISC_R_SUCCESS); } void dns_resolver_setudpsize(dns_resolver_t *resolver, isc_uint16_t udpsize) { REQUIRE(VALID_RESOLVER(resolver)); resolver->udpsize = udpsize; } isc_uint16_t dns_resolver_getudpsize(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->udpsize); } void dns_resolver_flushbadcache(dns_resolver_t *resolver, dns_name_t *name) { unsigned int i; dns_badcache_t *bad, *prev, *next; REQUIRE(VALID_RESOLVER(resolver)); LOCK(&resolver->lock); if (resolver->badcache == NULL) goto unlock; if (name != NULL) { isc_time_t now; isc_result_t result; result = isc_time_now(&now); if (result != ISC_R_SUCCESS) isc_time_settoepoch(&now); i = dns_name_hash(name, ISC_FALSE) % resolver->badhash; prev = NULL; for (bad = resolver->badcache[i]; bad != NULL; bad = next) { int n; next = bad->next; n = isc_time_compare(&bad->expire, &now); if (n < 0 || dns_name_equal(name, &bad->name)) { if (prev == NULL) resolver->badcache[i] = bad->next; else prev->next = bad->next; isc_mem_put(resolver->mctx, bad, sizeof(*bad) + bad->name.length); resolver->badcount--; } else prev = bad; } } else destroy_badcache(resolver); unlock: UNLOCK(&resolver->lock); } static void resizehash(dns_resolver_t *resolver, isc_time_t *now, isc_boolean_t grow) { unsigned int newsize; dns_badcache_t **new, *bad, *next; unsigned int i; if (grow) newsize = resolver->badhash * 2 + 1; else newsize = (resolver->badhash - 1) / 2; new = isc_mem_get(resolver->mctx, sizeof(*resolver->badcache) * newsize); if (new == NULL) return; memset(new, 0, sizeof(*resolver->badcache) * newsize); for (i = 0; i < resolver->badhash; i++) { for (bad = resolver->badcache[i]; bad != NULL; bad = next) { next = bad->next; if (isc_time_compare(&bad->expire, now) < 0) { isc_mem_put(resolver->mctx, bad, sizeof(*bad) + bad->name.length); resolver->badcount--; } else { bad->next = new[bad->hashval % newsize]; new[bad->hashval % newsize] = bad; } } } isc_mem_put(resolver->mctx, resolver->badcache, sizeof(*resolver->badcache) * resolver->badhash); resolver->badhash = newsize; resolver->badcache = new; } void dns_resolver_addbadcache(dns_resolver_t *resolver, dns_name_t *name, dns_rdatatype_t type, isc_time_t *expire) { isc_time_t now; isc_result_t result = ISC_R_SUCCESS; unsigned int i, hashval; dns_badcache_t *bad, *prev, *next; REQUIRE(VALID_RESOLVER(resolver)); LOCK(&resolver->lock); if (resolver->badcache == NULL) { resolver->badcache = isc_mem_get(resolver->mctx, sizeof(*resolver->badcache) * DNS_BADCACHE_SIZE); if (resolver->badcache == NULL) goto cleanup; resolver->badhash = DNS_BADCACHE_SIZE; memset(resolver->badcache, 0, sizeof(*resolver->badcache) * resolver->badhash); } result = isc_time_now(&now); if (result != ISC_R_SUCCESS) isc_time_settoepoch(&now); hashval = dns_name_hash(name, ISC_FALSE); i = hashval % resolver->badhash; prev = NULL; for (bad = resolver->badcache[i]; bad != NULL; bad = next) { next = bad->next; if (bad->type == type && dns_name_equal(name, &bad->name)) break; if (isc_time_compare(&bad->expire, &now) < 0) { if (prev == NULL) resolver->badcache[i] = bad->next; else prev->next = bad->next; isc_mem_put(resolver->mctx, bad, sizeof(*bad) + bad->name.length); resolver->badcount--; } else prev = bad; } if (bad == NULL) { isc_buffer_t buffer; bad = isc_mem_get(resolver->mctx, sizeof(*bad) + name->length); if (bad == NULL) goto cleanup; bad->type = type; bad->hashval = hashval; bad->expire = *expire; isc_buffer_init(&buffer, bad + 1, name->length); dns_name_init(&bad->name, NULL); dns_name_copy(name, &bad->name, &buffer); bad->next = resolver->badcache[i]; resolver->badcache[i] = bad; resolver->badcount++; if (resolver->badcount > resolver->badhash * 8) resizehash(resolver, &now, ISC_TRUE); if (resolver->badcount < resolver->badhash * 2 && resolver->badhash > DNS_BADCACHE_SIZE) resizehash(resolver, &now, ISC_FALSE); } else bad->expire = *expire; cleanup: UNLOCK(&resolver->lock); } isc_boolean_t dns_resolver_getbadcache(dns_resolver_t *resolver, dns_name_t *name, dns_rdatatype_t type, isc_time_t *now) { dns_badcache_t *bad, *prev, *next; isc_boolean_t answer = ISC_FALSE; unsigned int i; REQUIRE(VALID_RESOLVER(resolver)); LOCK(&resolver->lock); if (resolver->badcache == NULL) goto unlock; i = dns_name_hash(name, ISC_FALSE) % resolver->badhash; prev = NULL; for (bad = resolver->badcache[i]; bad != NULL; bad = next) { next = bad->next; /* * Search the hash list. Clean out expired records as we go. */ if (isc_time_compare(&bad->expire, now) < 0) { if (prev != NULL) prev->next = bad->next; else resolver->badcache[i] = bad->next; isc_mem_put(resolver->mctx, bad, sizeof(*bad) + bad->name.length); resolver->badcount--; continue; } if (bad->type == type && dns_name_equal(name, &bad->name)) { answer = ISC_TRUE; break; } prev = bad; } /* * Slow sweep to clean out stale records. */ i = resolver->badsweep++ % resolver->badhash; bad = resolver->badcache[i]; if (bad != NULL && isc_time_compare(&bad->expire, now) < 0) { resolver->badcache[i] = bad->next; isc_mem_put(resolver->mctx, bad, sizeof(*bad) + bad->name.length); resolver->badcount--; } unlock: UNLOCK(&resolver->lock); return (answer); } void dns_resolver_printbadcache(dns_resolver_t *resolver, FILE *fp) { char namebuf[DNS_NAME_FORMATSIZE]; char typebuf[DNS_RDATATYPE_FORMATSIZE]; dns_badcache_t *bad, *next, *prev; isc_time_t now; unsigned int i; isc_uint64_t t; LOCK(&resolver->lock); fprintf(fp, ";\n; Bad cache\n;\n"); if (resolver->badcache == NULL) goto unlock; TIME_NOW(&now); for (i = 0; i < resolver->badhash; i++) { prev = NULL; for (bad = resolver->badcache[i]; bad != NULL; bad = next) { next = bad->next; if (isc_time_compare(&bad->expire, &now) < 0) { if (prev != NULL) prev->next = bad->next; else resolver->badcache[i] = bad->next; isc_mem_put(resolver->mctx, bad, sizeof(*bad) + bad->name.length); resolver->badcount--; continue; } prev = bad; dns_name_format(&bad->name, namebuf, sizeof(namebuf)); dns_rdatatype_format(bad->type, typebuf, sizeof(typebuf)); t = isc_time_microdiff(&bad->expire, &now); t /= 1000; fprintf(fp, "; %s/%s [ttl " "%" ISC_PLATFORM_QUADFORMAT "u]\n", namebuf, typebuf, t); } } unlock: UNLOCK(&resolver->lock); } static void free_algorithm(void *node, void *arg) { unsigned char *algorithms = node; isc_mem_t *mctx = arg; isc_mem_put(mctx, algorithms, *algorithms); } void dns_resolver_reset_algorithms(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); #if USE_ALGLOCK RWLOCK(&resolver->alglock, isc_rwlocktype_write); #endif if (resolver->algorithms != NULL) dns_rbt_destroy(&resolver->algorithms); #if USE_ALGLOCK RWUNLOCK(&resolver->alglock, isc_rwlocktype_write); #endif } isc_result_t dns_resolver_disable_algorithm(dns_resolver_t *resolver, dns_name_t *name, unsigned int alg) { unsigned int len, mask; unsigned char *new; unsigned char *algorithms; isc_result_t result; dns_rbtnode_t *node = NULL; REQUIRE(VALID_RESOLVER(resolver)); if (alg > 255) return (ISC_R_RANGE); #if USE_ALGLOCK RWLOCK(&resolver->alglock, isc_rwlocktype_write); #endif if (resolver->algorithms == NULL) { result = dns_rbt_create(resolver->mctx, free_algorithm, resolver->mctx, &resolver->algorithms); if (result != ISC_R_SUCCESS) goto cleanup; } len = alg/8 + 2; mask = 1 << (alg%8); result = dns_rbt_addnode(resolver->algorithms, name, &node); if (result == ISC_R_SUCCESS || result == ISC_R_EXISTS) { algorithms = node->data; if (algorithms == NULL || len > *algorithms) { new = isc_mem_get(resolver->mctx, len); if (new == NULL) { result = ISC_R_NOMEMORY; goto cleanup; } memset(new, 0, len); if (algorithms != NULL) memmove(new, algorithms, *algorithms); new[len-1] |= mask; *new = len; node->data = new; if (algorithms != NULL) isc_mem_put(resolver->mctx, algorithms, *algorithms); } else algorithms[len-1] |= mask; } result = ISC_R_SUCCESS; cleanup: #if USE_ALGLOCK RWUNLOCK(&resolver->alglock, isc_rwlocktype_write); #endif return (result); } isc_boolean_t dns_resolver_algorithm_supported(dns_resolver_t *resolver, dns_name_t *name, unsigned int alg) { unsigned int len, mask; unsigned char *algorithms; void *data = NULL; isc_result_t result; isc_boolean_t found = ISC_FALSE; REQUIRE(VALID_RESOLVER(resolver)); /* * DH is unsupported for DNSKEYs, see RFC 4034 sec. A.1. */ if ((alg == DST_ALG_DH) || (alg == DST_ALG_INDIRECT)) return (ISC_FALSE); #if USE_ALGLOCK RWLOCK(&resolver->alglock, isc_rwlocktype_read); #endif if (resolver->algorithms == NULL) goto unlock; result = dns_rbt_findname(resolver->algorithms, name, 0, NULL, &data); if (result == ISC_R_SUCCESS || result == DNS_R_PARTIALMATCH) { len = alg/8 + 2; mask = 1 << (alg%8); algorithms = data; if (len <= *algorithms && (algorithms[len-1] & mask) != 0) found = ISC_TRUE; } unlock: #if USE_ALGLOCK RWUNLOCK(&resolver->alglock, isc_rwlocktype_read); #endif if (found) return (ISC_FALSE); return (dst_algorithm_supported(alg)); } isc_boolean_t dns_resolver_digest_supported(dns_resolver_t *resolver, unsigned int digest) { UNUSED(resolver); return (dns_ds_digest_supported(digest)); } void dns_resolver_resetmustbesecure(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); #if USE_MBSLOCK RWLOCK(&resolver->mbslock, isc_rwlocktype_write); #endif if (resolver->mustbesecure != NULL) dns_rbt_destroy(&resolver->mustbesecure); #if USE_MBSLOCK RWUNLOCK(&resolver->mbslock, isc_rwlocktype_write); #endif } static isc_boolean_t yes = ISC_TRUE, no = ISC_FALSE; isc_result_t dns_resolver_setmustbesecure(dns_resolver_t *resolver, dns_name_t *name, isc_boolean_t value) { isc_result_t result; REQUIRE(VALID_RESOLVER(resolver)); #if USE_MBSLOCK RWLOCK(&resolver->mbslock, isc_rwlocktype_write); #endif if (resolver->mustbesecure == NULL) { result = dns_rbt_create(resolver->mctx, NULL, NULL, &resolver->mustbesecure); if (result != ISC_R_SUCCESS) goto cleanup; } result = dns_rbt_addname(resolver->mustbesecure, name, value ? &yes : &no); cleanup: #if USE_MBSLOCK RWUNLOCK(&resolver->mbslock, isc_rwlocktype_write); #endif return (result); } isc_boolean_t dns_resolver_getmustbesecure(dns_resolver_t *resolver, dns_name_t *name) { void *data = NULL; isc_boolean_t value = ISC_FALSE; isc_result_t result; REQUIRE(VALID_RESOLVER(resolver)); #if USE_MBSLOCK RWLOCK(&resolver->mbslock, isc_rwlocktype_read); #endif if (resolver->mustbesecure == NULL) goto unlock; result = dns_rbt_findname(resolver->mustbesecure, name, 0, NULL, &data); if (result == ISC_R_SUCCESS || result == DNS_R_PARTIALMATCH) value = *(isc_boolean_t*)data; unlock: #if USE_MBSLOCK RWUNLOCK(&resolver->mbslock, isc_rwlocktype_read); #endif return (value); } void dns_resolver_getclientsperquery(dns_resolver_t *resolver, isc_uint32_t *cur, isc_uint32_t *min, isc_uint32_t *max) { REQUIRE(VALID_RESOLVER(resolver)); LOCK(&resolver->lock); if (cur != NULL) *cur = resolver->spillat; if (min != NULL) *min = resolver->spillatmin; if (max != NULL) *max = resolver->spillatmax; UNLOCK(&resolver->lock); } void dns_resolver_setclientsperquery(dns_resolver_t *resolver, isc_uint32_t min, isc_uint32_t max) { REQUIRE(VALID_RESOLVER(resolver)); LOCK(&resolver->lock); resolver->spillatmin = resolver->spillat = min; resolver->spillatmax = max; UNLOCK(&resolver->lock); } isc_boolean_t dns_resolver_getzeronosoattl(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->zero_no_soa_ttl); } void dns_resolver_setzeronosoattl(dns_resolver_t *resolver, isc_boolean_t state) { REQUIRE(VALID_RESOLVER(resolver)); resolver->zero_no_soa_ttl = state; } unsigned int dns_resolver_getoptions(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->options); } unsigned int dns_resolver_gettimeout(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->query_timeout); } void dns_resolver_settimeout(dns_resolver_t *resolver, unsigned int seconds) { REQUIRE(VALID_RESOLVER(resolver)); if (seconds == 0) seconds = DEFAULT_QUERY_TIMEOUT; if (seconds > MAXIMUM_QUERY_TIMEOUT) seconds = MAXIMUM_QUERY_TIMEOUT; if (seconds < MINIMUM_QUERY_TIMEOUT) seconds = MINIMUM_QUERY_TIMEOUT; resolver->query_timeout = seconds; } void dns_resolver_setmaxdepth(dns_resolver_t *resolver, unsigned int maxdepth) { REQUIRE(VALID_RESOLVER(resolver)); resolver->maxdepth = maxdepth; } unsigned int dns_resolver_getmaxdepth(dns_resolver_t *resolver) { REQUIRE(VALID_RESOLVER(resolver)); return (resolver->maxdepth); } Index: releng/9.3/contrib/bind9/lib/isccc/cc.c =================================================================== --- releng/9.3/contrib/bind9/lib/isccc/cc.c (revision 296610) +++ releng/9.3/contrib/bind9/lib/isccc/cc.c (revision 296611) @@ -1,857 +1,857 @@ /* * Portions Copyright (C) 2004-2007, 2012, 2013 Internet Systems Consortium, Inc. ("ISC") * Portions Copyright (C) 2001-2003 Internet Software Consortium. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND ISC AND NOMINUM DISCLAIMS ALL * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY * SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. * * Portions Copyright (C) 2001 Nominum, Inc. * * Permission to use, copy, modify, and/or distribute this software for any * purpose with or without fee is hereby granted, provided that the above * copyright notice and this permission notice appear in all copies. * * THE SOFTWARE IS PROVIDED "AS IS" AND ISC AND NOMINUM DISCLAIMS ALL * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES * OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL ISC BE LIABLE FOR ANY * SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ /* $Id: cc.c,v 1.18 2007/08/28 07:20:43 tbox Exp $ */ /*! \file */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #define MAX_TAGS 256 #define DUP_LIFETIME 900 typedef isccc_sexpr_t *sexpr_ptr; static unsigned char auth_hmd5[] = { 0x05, 0x5f, 0x61, 0x75, 0x74, 0x68, /*%< len + _auth */ ISCCC_CCMSGTYPE_TABLE, /*%< message type */ 0x00, 0x00, 0x00, 0x20, /*%< length == 32 */ 0x04, 0x68, 0x6d, 0x64, 0x35, /*%< len + hmd5 */ ISCCC_CCMSGTYPE_BINARYDATA, /*%< message type */ 0x00, 0x00, 0x00, 0x16, /*%< length == 22 */ /* * The base64 encoding of one of our HMAC-MD5 signatures is * 22 bytes. */ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 }; #define HMD5_OFFSET 21 /*%< 21 = 6 + 1 + 4 + 5 + 1 + 4 */ #define HMD5_LENGTH 22 static isc_result_t table_towire(isccc_sexpr_t *alist, isccc_region_t *target); static isc_result_t list_towire(isccc_sexpr_t *alist, isccc_region_t *target); static isc_result_t value_towire(isccc_sexpr_t *elt, isccc_region_t *target) { unsigned int len; unsigned char *lenp; isccc_region_t *vr; isc_result_t result; if (isccc_sexpr_binaryp(elt)) { vr = isccc_sexpr_tobinary(elt); len = REGION_SIZE(*vr); if (REGION_SIZE(*target) < 1 + 4 + len) return (ISC_R_NOSPACE); PUT8(ISCCC_CCMSGTYPE_BINARYDATA, target->rstart); PUT32(len, target->rstart); if (REGION_SIZE(*target) < len) return (ISC_R_NOSPACE); PUT_MEM(vr->rstart, len, target->rstart); } else if (isccc_alist_alistp(elt)) { if (REGION_SIZE(*target) < 1 + 4) return (ISC_R_NOSPACE); PUT8(ISCCC_CCMSGTYPE_TABLE, target->rstart); /* * Emit a placeholder length. */ lenp = target->rstart; PUT32(0, target->rstart); /* * Emit the table. */ result = table_towire(elt, target); if (result != ISC_R_SUCCESS) return (result); len = (unsigned int)(target->rstart - lenp); /* * 'len' is 4 bytes too big, since it counts * the placeholder length too. Adjust and * emit. */ INSIST(len >= 4U); len -= 4; PUT32(len, lenp); } else if (isccc_sexpr_listp(elt)) { if (REGION_SIZE(*target) < 1 + 4) return (ISC_R_NOSPACE); PUT8(ISCCC_CCMSGTYPE_LIST, target->rstart); /* * Emit a placeholder length and count. */ lenp = target->rstart; PUT32(0, target->rstart); /* * Emit the list. */ result = list_towire(elt, target); if (result != ISC_R_SUCCESS) return (result); len = (unsigned int)(target->rstart - lenp); /* * 'len' is 4 bytes too big, since it counts * the placeholder length. Adjust and emit. */ INSIST(len >= 4U); len -= 4; PUT32(len, lenp); } return (ISC_R_SUCCESS); } static isc_result_t table_towire(isccc_sexpr_t *alist, isccc_region_t *target) { isccc_sexpr_t *kv, *elt, *k, *v; char *ks; isc_result_t result; size_t len; for (elt = isccc_alist_first(alist); elt != NULL; elt = ISCCC_SEXPR_CDR(elt)) { kv = ISCCC_SEXPR_CAR(elt); k = ISCCC_SEXPR_CAR(kv); ks = isccc_sexpr_tostring(k); v = ISCCC_SEXPR_CDR(kv); len = strlen(ks); INSIST(len <= 255U); /* * Emit the key name. */ if (REGION_SIZE(*target) < 1 + len) return (ISC_R_NOSPACE); PUT8(len, target->rstart); PUT_MEM(ks, len, target->rstart); /* * Emit the value. */ result = value_towire(v, target); if (result != ISC_R_SUCCESS) return (result); } return (ISC_R_SUCCESS); } static isc_result_t list_towire(isccc_sexpr_t *list, isccc_region_t *target) { isc_result_t result; while (list != NULL) { result = value_towire(ISCCC_SEXPR_CAR(list), target); if (result != ISC_R_SUCCESS) return (result); list = ISCCC_SEXPR_CDR(list); } return (ISC_R_SUCCESS); } static isc_result_t sign(unsigned char *data, unsigned int length, unsigned char *hmd5, isccc_region_t *secret) { isc_hmacmd5_t ctx; isc_result_t result; isccc_region_t source, target; unsigned char digest[ISC_MD5_DIGESTLENGTH]; unsigned char digestb64[ISC_MD5_DIGESTLENGTH * 4]; isc_hmacmd5_init(&ctx, secret->rstart, REGION_SIZE(*secret)); isc_hmacmd5_update(&ctx, data, length); isc_hmacmd5_sign(&ctx, digest); source.rstart = digest; source.rend = digest + ISC_MD5_DIGESTLENGTH; target.rstart = digestb64; target.rend = digestb64 + ISC_MD5_DIGESTLENGTH * 4; result = isccc_base64_encode(&source, 64, "", &target); if (result != ISC_R_SUCCESS) return (result); PUT_MEM(digestb64, HMD5_LENGTH, hmd5); return (ISC_R_SUCCESS); } isc_result_t isccc_cc_towire(isccc_sexpr_t *alist, isccc_region_t *target, isccc_region_t *secret) { unsigned char *hmd5_rstart, *signed_rstart; isc_result_t result; if (REGION_SIZE(*target) < 4 + sizeof(auth_hmd5)) return (ISC_R_NOSPACE); /* * Emit protocol version. */ PUT32(1, target->rstart); if (secret != NULL) { /* * Emit _auth section with zeroed HMAC-MD5 signature. * We'll replace the zeros with the real signature once * we know what it is. */ hmd5_rstart = target->rstart + HMD5_OFFSET; PUT_MEM(auth_hmd5, sizeof(auth_hmd5), target->rstart); } else hmd5_rstart = NULL; signed_rstart = target->rstart; /* * Delete any existing _auth section so that we don't try * to encode it. */ isccc_alist_delete(alist, "_auth"); /* * Emit the message. */ result = table_towire(alist, target); if (result != ISC_R_SUCCESS) return (result); if (secret != NULL) return (sign(signed_rstart, (unsigned int)(target->rstart - signed_rstart), hmd5_rstart, secret)); return (ISC_R_SUCCESS); } static isc_result_t verify(isccc_sexpr_t *alist, unsigned char *data, unsigned int length, isccc_region_t *secret) { isc_hmacmd5_t ctx; isccc_region_t source; isccc_region_t target; isc_result_t result; isccc_sexpr_t *_auth, *hmd5; unsigned char digest[ISC_MD5_DIGESTLENGTH]; unsigned char digestb64[ISC_MD5_DIGESTLENGTH * 4]; /* * Extract digest. */ _auth = isccc_alist_lookup(alist, "_auth"); - if (_auth == NULL) + if (!isccc_alist_alistp(_auth)) return (ISC_R_FAILURE); hmd5 = isccc_alist_lookup(_auth, "hmd5"); - if (hmd5 == NULL) + if (!isccc_sexpr_binaryp(hmd5)) return (ISC_R_FAILURE); /* * Compute digest. */ isc_hmacmd5_init(&ctx, secret->rstart, REGION_SIZE(*secret)); isc_hmacmd5_update(&ctx, data, length); isc_hmacmd5_sign(&ctx, digest); source.rstart = digest; source.rend = digest + ISC_MD5_DIGESTLENGTH; target.rstart = digestb64; target.rend = digestb64 + ISC_MD5_DIGESTLENGTH * 4; result = isccc_base64_encode(&source, 64, "", &target); if (result != ISC_R_SUCCESS) return (result); /* * Strip trailing == and NUL terminate target. */ target.rstart -= 2; *target.rstart++ = '\0'; /* * Verify. */ if (!isc_safe_memcmp((unsigned char *) isccc_sexpr_tostring(hmd5), digestb64, HMD5_LENGTH)) return (ISCCC_R_BADAUTH); return (ISC_R_SUCCESS); } static isc_result_t table_fromwire(isccc_region_t *source, isccc_region_t *secret, isccc_sexpr_t **alistp); static isc_result_t list_fromwire(isccc_region_t *source, isccc_sexpr_t **listp); static isc_result_t value_fromwire(isccc_region_t *source, isccc_sexpr_t **valuep) { unsigned int msgtype; isc_uint32_t len; isccc_sexpr_t *value; isccc_region_t active; isc_result_t result; if (REGION_SIZE(*source) < 1 + 4) return (ISC_R_UNEXPECTEDEND); GET8(msgtype, source->rstart); GET32(len, source->rstart); if (REGION_SIZE(*source) < len) return (ISC_R_UNEXPECTEDEND); active.rstart = source->rstart; active.rend = active.rstart + len; source->rstart = active.rend; if (msgtype == ISCCC_CCMSGTYPE_BINARYDATA) { value = isccc_sexpr_frombinary(&active); if (value != NULL) { *valuep = value; result = ISC_R_SUCCESS; } else result = ISC_R_NOMEMORY; } else if (msgtype == ISCCC_CCMSGTYPE_TABLE) result = table_fromwire(&active, NULL, valuep); else if (msgtype == ISCCC_CCMSGTYPE_LIST) result = list_fromwire(&active, valuep); else result = ISCCC_R_SYNTAX; return (result); } static isc_result_t table_fromwire(isccc_region_t *source, isccc_region_t *secret, isccc_sexpr_t **alistp) { char key[256]; isc_uint32_t len; isc_result_t result; isccc_sexpr_t *alist, *value; isc_boolean_t first_tag; unsigned char *checksum_rstart; REQUIRE(alistp != NULL && *alistp == NULL); checksum_rstart = NULL; first_tag = ISC_TRUE; alist = isccc_alist_create(); if (alist == NULL) return (ISC_R_NOMEMORY); while (!REGION_EMPTY(*source)) { GET8(len, source->rstart); if (REGION_SIZE(*source) < len) { result = ISC_R_UNEXPECTEDEND; goto bad; } GET_MEM(key, len, source->rstart); key[len] = '\0'; /* Ensure NUL termination. */ value = NULL; result = value_fromwire(source, &value); if (result != ISC_R_SUCCESS) goto bad; if (isccc_alist_define(alist, key, value) == NULL) { result = ISC_R_NOMEMORY; goto bad; } if (first_tag && secret != NULL && strcmp(key, "_auth") == 0) checksum_rstart = source->rstart; first_tag = ISC_FALSE; } if (secret != NULL) { if (checksum_rstart != NULL) result = verify(alist, checksum_rstart, (unsigned int) (source->rend - checksum_rstart), secret); else result = ISCCC_R_BADAUTH; } else result = ISC_R_SUCCESS; bad: if (result == ISC_R_SUCCESS) *alistp = alist; else isccc_sexpr_free(&alist); return (result); } static isc_result_t list_fromwire(isccc_region_t *source, isccc_sexpr_t **listp) { isccc_sexpr_t *list, *value; isc_result_t result; list = NULL; while (!REGION_EMPTY(*source)) { value = NULL; result = value_fromwire(source, &value); if (result != ISC_R_SUCCESS) { isccc_sexpr_free(&list); return (result); } if (isccc_sexpr_addtolist(&list, value) == NULL) { isccc_sexpr_free(&value); isccc_sexpr_free(&list); return (result); } } *listp = list; return (ISC_R_SUCCESS); } isc_result_t isccc_cc_fromwire(isccc_region_t *source, isccc_sexpr_t **alistp, isccc_region_t *secret) { unsigned int size; isc_uint32_t version; size = REGION_SIZE(*source); if (size < 4) return (ISC_R_UNEXPECTEDEND); GET32(version, source->rstart); if (version != 1) return (ISCCC_R_UNKNOWNVERSION); return (table_fromwire(source, secret, alistp)); } static isc_result_t createmessage(isc_uint32_t version, const char *from, const char *to, isc_uint32_t serial, isccc_time_t now, isccc_time_t expires, isccc_sexpr_t **alistp, isc_boolean_t want_expires) { isccc_sexpr_t *alist, *_ctrl, *_data; isc_result_t result; REQUIRE(alistp != NULL && *alistp == NULL); if (version != 1) return (ISCCC_R_UNKNOWNVERSION); alist = isccc_alist_create(); if (alist == NULL) return (ISC_R_NOMEMORY); result = ISC_R_NOMEMORY; _ctrl = isccc_alist_create(); if (_ctrl == NULL) goto bad; if (isccc_alist_define(alist, "_ctrl", _ctrl) == NULL) { isccc_sexpr_free(&_ctrl); goto bad; } _data = isccc_alist_create(); if (_data == NULL) goto bad; if (isccc_alist_define(alist, "_data", _data) == NULL) { isccc_sexpr_free(&_data); goto bad; } if (isccc_cc_defineuint32(_ctrl, "_ser", serial) == NULL || isccc_cc_defineuint32(_ctrl, "_tim", now) == NULL || (want_expires && isccc_cc_defineuint32(_ctrl, "_exp", expires) == NULL)) goto bad; if (from != NULL && isccc_cc_definestring(_ctrl, "_frm", from) == NULL) goto bad; if (to != NULL && isccc_cc_definestring(_ctrl, "_to", to) == NULL) goto bad; *alistp = alist; return (ISC_R_SUCCESS); bad: isccc_sexpr_free(&alist); return (result); } isc_result_t isccc_cc_createmessage(isc_uint32_t version, const char *from, const char *to, isc_uint32_t serial, isccc_time_t now, isccc_time_t expires, isccc_sexpr_t **alistp) { return (createmessage(version, from, to, serial, now, expires, alistp, ISC_TRUE)); } isc_result_t isccc_cc_createack(isccc_sexpr_t *message, isc_boolean_t ok, isccc_sexpr_t **ackp) { char *_frm, *_to; isc_uint32_t serial; isccc_sexpr_t *ack, *_ctrl; isc_result_t result; isccc_time_t t; REQUIRE(ackp != NULL && *ackp == NULL); _ctrl = isccc_alist_lookup(message, "_ctrl"); - if (_ctrl == NULL || + if (!isccc_alist_alistp(_ctrl) || isccc_cc_lookupuint32(_ctrl, "_ser", &serial) != ISC_R_SUCCESS || isccc_cc_lookupuint32(_ctrl, "_tim", &t) != ISC_R_SUCCESS) return (ISC_R_FAILURE); /* * _frm and _to are optional. */ _frm = NULL; (void)isccc_cc_lookupstring(_ctrl, "_frm", &_frm); _to = NULL; (void)isccc_cc_lookupstring(_ctrl, "_to", &_to); /* * Create the ack. */ ack = NULL; result = createmessage(1, _to, _frm, serial, t, 0, &ack, ISC_FALSE); if (result != ISC_R_SUCCESS) return (result); _ctrl = isccc_alist_lookup(ack, "_ctrl"); if (_ctrl == NULL) { result = ISC_R_FAILURE; goto bad; } if (isccc_cc_definestring(ack, "_ack", (ok) ? "1" : "0") == NULL) { result = ISC_R_NOMEMORY; goto bad; } *ackp = ack; return (ISC_R_SUCCESS); bad: isccc_sexpr_free(&ack); return (result); } isc_boolean_t isccc_cc_isack(isccc_sexpr_t *message) { isccc_sexpr_t *_ctrl; _ctrl = isccc_alist_lookup(message, "_ctrl"); - if (_ctrl == NULL) + if (!isccc_alist_alistp(_ctrl)) return (ISC_FALSE); if (isccc_cc_lookupstring(_ctrl, "_ack", NULL) == ISC_R_SUCCESS) return (ISC_TRUE); return (ISC_FALSE); } isc_boolean_t isccc_cc_isreply(isccc_sexpr_t *message) { isccc_sexpr_t *_ctrl; _ctrl = isccc_alist_lookup(message, "_ctrl"); - if (_ctrl == NULL) + if (!isccc_alist_alistp(_ctrl)) return (ISC_FALSE); if (isccc_cc_lookupstring(_ctrl, "_rpl", NULL) == ISC_R_SUCCESS) return (ISC_TRUE); return (ISC_FALSE); } isc_result_t isccc_cc_createresponse(isccc_sexpr_t *message, isccc_time_t now, isccc_time_t expires, isccc_sexpr_t **alistp) { char *_frm, *_to, *type = NULL; isc_uint32_t serial; isccc_sexpr_t *alist, *_ctrl, *_data; isc_result_t result; REQUIRE(alistp != NULL && *alistp == NULL); _ctrl = isccc_alist_lookup(message, "_ctrl"); _data = isccc_alist_lookup(message, "_data"); - if (_ctrl == NULL || _data == NULL || + if (!isccc_alist_alistp(_ctrl) || !isccc_alist_alistp(_data) || isccc_cc_lookupuint32(_ctrl, "_ser", &serial) != ISC_R_SUCCESS || isccc_cc_lookupstring(_data, "type", &type) != ISC_R_SUCCESS) return (ISC_R_FAILURE); /* * _frm and _to are optional. */ _frm = NULL; (void)isccc_cc_lookupstring(_ctrl, "_frm", &_frm); _to = NULL; (void)isccc_cc_lookupstring(_ctrl, "_to", &_to); /* * Create the response. */ alist = NULL; result = isccc_cc_createmessage(1, _to, _frm, serial, now, expires, &alist); if (result != ISC_R_SUCCESS) return (result); _ctrl = isccc_alist_lookup(alist, "_ctrl"); if (_ctrl == NULL) { result = ISC_R_FAILURE; goto bad; } _data = isccc_alist_lookup(alist, "_data"); if (_data == NULL) { result = ISC_R_FAILURE; goto bad; } if (isccc_cc_definestring(_ctrl, "_rpl", "1") == NULL || isccc_cc_definestring(_data, "type", type) == NULL) { result = ISC_R_NOMEMORY; goto bad; } *alistp = alist; return (ISC_R_SUCCESS); bad: isccc_sexpr_free(&alist); return (result); } isccc_sexpr_t * isccc_cc_definestring(isccc_sexpr_t *alist, const char *key, const char *str) { size_t len; isccc_region_t r; len = strlen(str); DE_CONST(str, r.rstart); r.rend = r.rstart + len; return (isccc_alist_definebinary(alist, key, &r)); } isccc_sexpr_t * isccc_cc_defineuint32(isccc_sexpr_t *alist, const char *key, isc_uint32_t i) { char b[100]; size_t len; isccc_region_t r; snprintf(b, sizeof(b), "%u", i); len = strlen(b); r.rstart = (unsigned char *)b; r.rend = (unsigned char *)b + len; return (isccc_alist_definebinary(alist, key, &r)); } isc_result_t isccc_cc_lookupstring(isccc_sexpr_t *alist, const char *key, char **strp) { isccc_sexpr_t *kv, *v; REQUIRE(strp == NULL || *strp == NULL); kv = isccc_alist_assq(alist, key); if (kv != NULL) { v = ISCCC_SEXPR_CDR(kv); if (isccc_sexpr_binaryp(v)) { if (strp != NULL) *strp = isccc_sexpr_tostring(v); return (ISC_R_SUCCESS); } else return (ISC_R_EXISTS); } return (ISC_R_NOTFOUND); } isc_result_t isccc_cc_lookupuint32(isccc_sexpr_t *alist, const char *key, isc_uint32_t *uintp) { isccc_sexpr_t *kv, *v; kv = isccc_alist_assq(alist, key); if (kv != NULL) { v = ISCCC_SEXPR_CDR(kv); if (isccc_sexpr_binaryp(v)) { if (uintp != NULL) *uintp = (isc_uint32_t) strtoul(isccc_sexpr_tostring(v), NULL, 10); return (ISC_R_SUCCESS); } else return (ISC_R_EXISTS); } return (ISC_R_NOTFOUND); } static void symtab_undefine(char *key, unsigned int type, isccc_symvalue_t value, void *arg) { UNUSED(type); UNUSED(value); UNUSED(arg); free(key); } static isc_boolean_t symtab_clean(char *key, unsigned int type, isccc_symvalue_t value, void *arg) { isccc_time_t *now; UNUSED(key); UNUSED(type); now = arg; if (*now < value.as_uinteger) return (ISC_FALSE); if ((*now - value.as_uinteger) < DUP_LIFETIME) return (ISC_FALSE); return (ISC_TRUE); } isc_result_t isccc_cc_createsymtab(isccc_symtab_t **symtabp) { return (isccc_symtab_create(11897, symtab_undefine, NULL, ISC_FALSE, symtabp)); } void isccc_cc_cleansymtab(isccc_symtab_t *symtab, isccc_time_t now) { isccc_symtab_foreach(symtab, symtab_clean, &now); } static isc_boolean_t has_whitespace(const char *str) { char c; if (str == NULL) return (ISC_FALSE); while ((c = *str++) != '\0') { if (c == ' ' || c == '\t' || c == '\n') return (ISC_TRUE); } return (ISC_FALSE); } isc_result_t isccc_cc_checkdup(isccc_symtab_t *symtab, isccc_sexpr_t *message, isccc_time_t now) { const char *_frm; const char *_to; char *_ser = NULL, *_tim = NULL, *tmp; isc_result_t result; char *key; size_t len; isccc_symvalue_t value; isccc_sexpr_t *_ctrl; _ctrl = isccc_alist_lookup(message, "_ctrl"); - if (_ctrl == NULL || + if (!isccc_alist_alistp(_ctrl) || isccc_cc_lookupstring(_ctrl, "_ser", &_ser) != ISC_R_SUCCESS || isccc_cc_lookupstring(_ctrl, "_tim", &_tim) != ISC_R_SUCCESS) return (ISC_R_FAILURE); INSIST(_ser != NULL); INSIST(_tim != NULL); /* * _frm and _to are optional. */ tmp = NULL; if (isccc_cc_lookupstring(_ctrl, "_frm", &tmp) != ISC_R_SUCCESS) _frm = ""; else _frm = tmp; tmp = NULL; if (isccc_cc_lookupstring(_ctrl, "_to", &tmp) != ISC_R_SUCCESS) _to = ""; else _to = tmp; /* * Ensure there is no newline in any of the strings. This is so * we can write them to a file later. */ if (has_whitespace(_frm) || has_whitespace(_to) || has_whitespace(_ser) || has_whitespace(_tim)) return (ISC_R_FAILURE); len = strlen(_frm) + strlen(_to) + strlen(_ser) + strlen(_tim) + 4; key = malloc(len); if (key == NULL) return (ISC_R_NOMEMORY); snprintf(key, len, "%s;%s;%s;%s", _frm, _to, _ser, _tim); value.as_uinteger = now; result = isccc_symtab_define(symtab, key, ISCCC_SYMTYPE_CCDUP, value, isccc_symexists_reject); if (result != ISC_R_SUCCESS) { free(key); return (result); } return (ISC_R_SUCCESS); } Index: releng/9.3/crypto/openssl/crypto/bn/bn_exp.c =================================================================== --- releng/9.3/crypto/openssl/crypto/bn/bn_exp.c (revision 296610) +++ releng/9.3/crypto/openssl/crypto/bn/bn_exp.c (revision 296611) @@ -1,1046 +1,1046 @@ /* crypto/bn/bn_exp.c */ /* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com) * All rights reserved. * * This package is an SSL implementation written * by Eric Young (eay@cryptsoft.com). * The implementation was written so as to conform with Netscapes SSL. * * This library is free for commercial and non-commercial use as long as * the following conditions are aheared to. The following conditions * apply to all code found in this distribution, be it the RC4, RSA, * lhash, DES, etc., code; not just the SSL code. The SSL documentation * included with this distribution is covered by the same copyright terms * except that the holder is Tim Hudson (tjh@cryptsoft.com). * * Copyright remains Eric Young's, and as such any Copyright notices in * the code are not to be removed. * If this package is used in a product, Eric Young should be given attribution * as the author of the parts of the library used. * This can be in the form of a textual message at program startup or * in documentation (online or textual) provided with the package. * * 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 copyright * notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in the * documentation and/or other materials provided with the distribution. * 3. All advertising materials mentioning features or use of this software * must display the following acknowledgement: * "This product includes cryptographic software written by * Eric Young (eay@cryptsoft.com)" * The word 'cryptographic' can be left out if the rouines from the library * being used are not cryptographic related :-). * 4. If you include any Windows specific code (or a derivative thereof) from * the apps directory (application code) you must include an acknowledgement: * "This product includes software written by Tim Hudson (tjh@cryptsoft.com)" * * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``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. * * The licence and distribution terms for any publically available version or * derivative of this code cannot be changed. i.e. this code cannot simply be * copied and put under another distribution licence * [including the GNU Public Licence.] */ /* ==================================================================== * Copyright (c) 1998-2005 The OpenSSL Project. All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * 1. Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * 2. Redistributions in binary form must reproduce the above copyright * notice, this list of conditions and the following disclaimer in * the documentation and/or other materials provided with the * distribution. * * 3. All advertising materials mentioning features or use of this * software must display the following acknowledgment: * "This product includes software developed by the OpenSSL Project * for use in the OpenSSL Toolkit. (http://www.openssl.org/)" * * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to * endorse or promote products derived from this software without * prior written permission. For written permission, please contact * openssl-core@openssl.org. * * 5. Products derived from this software may not be called "OpenSSL" * nor may "OpenSSL" appear in their names without prior written * permission of the OpenSSL Project. * * 6. Redistributions of any form whatsoever must retain the following * acknowledgment: * "This product includes software developed by the OpenSSL Project * for use in the OpenSSL Toolkit (http://www.openssl.org/)" * * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY * EXPRESSED 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 OpenSSL PROJECT OR * ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED * OF THE POSSIBILITY OF SUCH DAMAGE. * ==================================================================== * * This product includes cryptographic software written by Eric Young * (eay@cryptsoft.com). This product includes software written by Tim * Hudson (tjh@cryptsoft.com). * - */ - -#include "cryptlib.h" -#include "constant_time_locl.h" -#include "bn_lcl.h" - -/* maximum precomputation table size for *variable* sliding windows */ + */ + +#include "cryptlib.h" +#include "constant_time_locl.h" +#include "bn_lcl.h" + +/* maximum precomputation table size for *variable* sliding windows */ #define TABLE_SIZE 32 /* this one works - simple but works */ int BN_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx) { int i, bits, ret = 0; BIGNUM *v, *rr; if (BN_get_flags(p, BN_FLG_CONSTTIME) != 0) { /* BN_FLG_CONSTTIME only supported by BN_mod_exp_mont() */ BNerr(BN_F_BN_EXP, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED); return -1; } BN_CTX_start(ctx); if ((r == a) || (r == p)) rr = BN_CTX_get(ctx); else rr = r; v = BN_CTX_get(ctx); if (rr == NULL || v == NULL) goto err; if (BN_copy(v, a) == NULL) goto err; bits = BN_num_bits(p); if (BN_is_odd(p)) { if (BN_copy(rr, a) == NULL) goto err; } else { if (!BN_one(rr)) goto err; } for (i = 1; i < bits; i++) { if (!BN_sqr(v, v, ctx)) goto err; if (BN_is_bit_set(p, i)) { if (!BN_mul(rr, rr, v, ctx)) goto err; } } ret = 1; err: if (r != rr) BN_copy(r, rr); BN_CTX_end(ctx); bn_check_top(r); return (ret); } int BN_mod_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, const BIGNUM *m, BN_CTX *ctx) { int ret; bn_check_top(a); bn_check_top(p); bn_check_top(m); /*- * For even modulus m = 2^k*m_odd, it might make sense to compute * a^p mod m_odd and a^p mod 2^k separately (with Montgomery * exponentiation for the odd part), using appropriate exponent * reductions, and combine the results using the CRT. * * For now, we use Montgomery only if the modulus is odd; otherwise, * exponentiation using the reciprocal-based quick remaindering * algorithm is used. * * (Timing obtained with expspeed.c [computations a^p mod m * where a, p, m are of the same length: 256, 512, 1024, 2048, * 4096, 8192 bits], compared to the running time of the * standard algorithm: * * BN_mod_exp_mont 33 .. 40 % [AMD K6-2, Linux, debug configuration] * 55 .. 77 % [UltraSparc processor, but * debug-solaris-sparcv8-gcc conf.] * * BN_mod_exp_recp 50 .. 70 % [AMD K6-2, Linux, debug configuration] * 62 .. 118 % [UltraSparc, debug-solaris-sparcv8-gcc] * * On the Sparc, BN_mod_exp_recp was faster than BN_mod_exp_mont * at 2048 and more bits, but at 512 and 1024 bits, it was * slower even than the standard algorithm! * * "Real" timings [linux-elf, solaris-sparcv9-gcc configurations] * should be obtained when the new Montgomery reduction code * has been integrated into OpenSSL.) */ #define MONT_MUL_MOD #define MONT_EXP_WORD #define RECP_MUL_MOD #ifdef MONT_MUL_MOD /* * I have finally been able to take out this pre-condition of the top bit * being set. It was caused by an error in BN_div with negatives. There * was also another problem when for a^b%m a >= m. eay 07-May-97 */ /* if ((m->d[m->top-1]&BN_TBIT) && BN_is_odd(m)) */ if (BN_is_odd(m)) { # ifdef MONT_EXP_WORD if (a->top == 1 && !a->neg && (BN_get_flags(p, BN_FLG_CONSTTIME) == 0)) { BN_ULONG A = a->d[0]; ret = BN_mod_exp_mont_word(r, A, p, m, ctx, NULL); } else # endif ret = BN_mod_exp_mont(r, a, p, m, ctx, NULL); } else #endif #ifdef RECP_MUL_MOD { ret = BN_mod_exp_recp(r, a, p, m, ctx); } #else { ret = BN_mod_exp_simple(r, a, p, m, ctx); } #endif bn_check_top(r); return (ret); } int BN_mod_exp_recp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, const BIGNUM *m, BN_CTX *ctx) { int i, j, bits, ret = 0, wstart, wend, window, wvalue; int start = 1; BIGNUM *aa; /* Table of variables obtained from 'ctx' */ BIGNUM *val[TABLE_SIZE]; BN_RECP_CTX recp; if (BN_get_flags(p, BN_FLG_CONSTTIME) != 0) { /* BN_FLG_CONSTTIME only supported by BN_mod_exp_mont() */ BNerr(BN_F_BN_MOD_EXP_RECP, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED); return -1; } bits = BN_num_bits(p); if (bits == 0) { ret = BN_one(r); return ret; } BN_CTX_start(ctx); aa = BN_CTX_get(ctx); val[0] = BN_CTX_get(ctx); if (!aa || !val[0]) goto err; BN_RECP_CTX_init(&recp); if (m->neg) { /* ignore sign of 'm' */ if (!BN_copy(aa, m)) goto err; aa->neg = 0; if (BN_RECP_CTX_set(&recp, aa, ctx) <= 0) goto err; } else { if (BN_RECP_CTX_set(&recp, m, ctx) <= 0) goto err; } if (!BN_nnmod(val[0], a, m, ctx)) goto err; /* 1 */ if (BN_is_zero(val[0])) { BN_zero(r); ret = 1; goto err; } window = BN_window_bits_for_exponent_size(bits); if (window > 1) { if (!BN_mod_mul_reciprocal(aa, val[0], val[0], &recp, ctx)) goto err; /* 2 */ j = 1 << (window - 1); for (i = 1; i < j; i++) { if (((val[i] = BN_CTX_get(ctx)) == NULL) || !BN_mod_mul_reciprocal(val[i], val[i - 1], aa, &recp, ctx)) goto err; } } start = 1; /* This is used to avoid multiplication etc * when there is only the value '1' in the * buffer. */ wvalue = 0; /* The 'value' of the window */ wstart = bits - 1; /* The top bit of the window */ wend = 0; /* The bottom bit of the window */ if (!BN_one(r)) goto err; for (;;) { if (BN_is_bit_set(p, wstart) == 0) { if (!start) if (!BN_mod_mul_reciprocal(r, r, r, &recp, ctx)) goto err; if (wstart == 0) break; wstart--; continue; } /* * We now have wstart on a 'set' bit, we now need to work out how bit * a window to do. To do this we need to scan forward until the last * set bit before the end of the window */ j = wstart; wvalue = 1; wend = 0; for (i = 1; i < window; i++) { if (wstart - i < 0) break; if (BN_is_bit_set(p, wstart - i)) { wvalue <<= (i - wend); wvalue |= 1; wend = i; } } /* wend is the size of the current window */ j = wend + 1; /* add the 'bytes above' */ if (!start) for (i = 0; i < j; i++) { if (!BN_mod_mul_reciprocal(r, r, r, &recp, ctx)) goto err; } /* wvalue will be an odd number < 2^window */ if (!BN_mod_mul_reciprocal(r, r, val[wvalue >> 1], &recp, ctx)) goto err; /* move the 'window' down further */ wstart -= wend + 1; wvalue = 0; start = 0; if (wstart < 0) break; } ret = 1; err: BN_CTX_end(ctx); BN_RECP_CTX_free(&recp); bn_check_top(r); return (ret); } int BN_mod_exp_mont(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p, const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *in_mont) { int i, j, bits, ret = 0, wstart, wend, window, wvalue; int start = 1; BIGNUM *d, *r; const BIGNUM *aa; /* Table of variables obtained from 'ctx' */ BIGNUM *val[TABLE_SIZE]; BN_MONT_CTX *mont = NULL; if (BN_get_flags(p, BN_FLG_CONSTTIME) != 0) { return BN_mod_exp_mont_consttime(rr, a, p, m, ctx, in_mont); } bn_check_top(a); bn_check_top(p); bn_check_top(m); if (!BN_is_odd(m)) { BNerr(BN_F_BN_MOD_EXP_MONT, BN_R_CALLED_WITH_EVEN_MODULUS); return (0); } bits = BN_num_bits(p); if (bits == 0) { ret = BN_one(rr); return ret; } BN_CTX_start(ctx); d = BN_CTX_get(ctx); r = BN_CTX_get(ctx); val[0] = BN_CTX_get(ctx); if (!d || !r || !val[0]) goto err; /* * If this is not done, things will break in the montgomery part */ if (in_mont != NULL) mont = in_mont; else { if ((mont = BN_MONT_CTX_new()) == NULL) goto err; if (!BN_MONT_CTX_set(mont, m, ctx)) goto err; } if (a->neg || BN_ucmp(a, m) >= 0) { if (!BN_nnmod(val[0], a, m, ctx)) goto err; aa = val[0]; } else aa = a; if (BN_is_zero(aa)) { BN_zero(rr); ret = 1; goto err; } if (!BN_to_montgomery(val[0], aa, mont, ctx)) goto err; /* 1 */ window = BN_window_bits_for_exponent_size(bits); if (window > 1) { if (!BN_mod_mul_montgomery(d, val[0], val[0], mont, ctx)) goto err; /* 2 */ j = 1 << (window - 1); for (i = 1; i < j; i++) { if (((val[i] = BN_CTX_get(ctx)) == NULL) || !BN_mod_mul_montgomery(val[i], val[i - 1], d, mont, ctx)) goto err; } } start = 1; /* This is used to avoid multiplication etc * when there is only the value '1' in the * buffer. */ wvalue = 0; /* The 'value' of the window */ wstart = bits - 1; /* The top bit of the window */ wend = 0; /* The bottom bit of the window */ if (!BN_to_montgomery(r, BN_value_one(), mont, ctx)) goto err; for (;;) { if (BN_is_bit_set(p, wstart) == 0) { if (!start) { if (!BN_mod_mul_montgomery(r, r, r, mont, ctx)) goto err; } if (wstart == 0) break; wstart--; continue; } /* * We now have wstart on a 'set' bit, we now need to work out how bit * a window to do. To do this we need to scan forward until the last * set bit before the end of the window */ j = wstart; wvalue = 1; wend = 0; for (i = 1; i < window; i++) { if (wstart - i < 0) break; if (BN_is_bit_set(p, wstart - i)) { wvalue <<= (i - wend); wvalue |= 1; wend = i; } } /* wend is the size of the current window */ j = wend + 1; /* add the 'bytes above' */ if (!start) for (i = 0; i < j; i++) { if (!BN_mod_mul_montgomery(r, r, r, mont, ctx)) goto err; } /* wvalue will be an odd number < 2^window */ if (!BN_mod_mul_montgomery(r, r, val[wvalue >> 1], mont, ctx)) goto err; /* move the 'window' down further */ wstart -= wend + 1; wvalue = 0; start = 0; if (wstart < 0) break; } if (!BN_from_montgomery(rr, r, mont, ctx)) goto err; ret = 1; err: if ((in_mont == NULL) && (mont != NULL)) BN_MONT_CTX_free(mont); BN_CTX_end(ctx); bn_check_top(rr); return (ret); } /* * BN_mod_exp_mont_consttime() stores the precomputed powers in a specific * layout so that accessing any of these table values shows the same access * pattern as far as cache lines are concerned. The following functions are * used to transfer a BIGNUM from/to that table. */ - -static int MOD_EXP_CTIME_COPY_TO_PREBUF(BIGNUM *b, int top, - unsigned char *buf, int idx, - int window) -{ - int i, j; - int width = 1 << window; - BN_ULONG *table = (BN_ULONG *)buf; - - if (bn_wexpand(b, top) == NULL) - return 0; + +static int MOD_EXP_CTIME_COPY_TO_PREBUF(BIGNUM *b, int top, + unsigned char *buf, int idx, + int window) +{ + int i, j; + int width = 1 << window; + BN_ULONG *table = (BN_ULONG *)buf; + + if (bn_wexpand(b, top) == NULL) + return 0; while (b->top < top) { - b->d[b->top++] = 0; - } - - for (i = 0, j = idx; i < top; i++, j += width) { - table[j] = b->d[i]; - } - - bn_correct_top(b); + b->d[b->top++] = 0; + } + + for (i = 0, j = idx; i < top; i++, j += width) { + table[j] = b->d[i]; + } + + bn_correct_top(b); return 1; } - -static int MOD_EXP_CTIME_COPY_FROM_PREBUF(BIGNUM *b, int top, - unsigned char *buf, int idx, - int window) -{ - int i, j; - int width = 1 << window; - volatile BN_ULONG *table = (volatile BN_ULONG *)buf; - - if (bn_wexpand(b, top) == NULL) - return 0; - - if (window <= 3) { - for (i = 0; i < top; i++, table += width) { - BN_ULONG acc = 0; - - for (j = 0; j < width; j++) { - acc |= table[j] & - ((BN_ULONG)0 - (constant_time_eq_int(j,idx)&1)); - } - - b->d[i] = acc; - } - } else { - int xstride = 1 << (window - 2); - BN_ULONG y0, y1, y2, y3; - - i = idx >> (window - 2); /* equivalent of idx / xstride */ - idx &= xstride - 1; /* equivalent of idx % xstride */ - - y0 = (BN_ULONG)0 - (constant_time_eq_int(i,0)&1); - y1 = (BN_ULONG)0 - (constant_time_eq_int(i,1)&1); - y2 = (BN_ULONG)0 - (constant_time_eq_int(i,2)&1); - y3 = (BN_ULONG)0 - (constant_time_eq_int(i,3)&1); - - for (i = 0; i < top; i++, table += width) { - BN_ULONG acc = 0; - - for (j = 0; j < xstride; j++) { - acc |= ( (table[j + 0 * xstride] & y0) | - (table[j + 1 * xstride] & y1) | - (table[j + 2 * xstride] & y2) | - (table[j + 3 * xstride] & y3) ) - & ((BN_ULONG)0 - (constant_time_eq_int(j,idx)&1)); - } - - b->d[i] = acc; - } - } - - b->top = top; + +static int MOD_EXP_CTIME_COPY_FROM_PREBUF(BIGNUM *b, int top, + unsigned char *buf, int idx, + int window) +{ + int i, j; + int width = 1 << window; + volatile BN_ULONG *table = (volatile BN_ULONG *)buf; + + if (bn_wexpand(b, top) == NULL) + return 0; + + if (window <= 3) { + for (i = 0; i < top; i++, table += width) { + BN_ULONG acc = 0; + + for (j = 0; j < width; j++) { + acc |= table[j] & + ((BN_ULONG)0 - (constant_time_eq_int(j,idx)&1)); + } + + b->d[i] = acc; + } + } else { + int xstride = 1 << (window - 2); + BN_ULONG y0, y1, y2, y3; + + i = idx >> (window - 2); /* equivalent of idx / xstride */ + idx &= xstride - 1; /* equivalent of idx % xstride */ + + y0 = (BN_ULONG)0 - (constant_time_eq_int(i,0)&1); + y1 = (BN_ULONG)0 - (constant_time_eq_int(i,1)&1); + y2 = (BN_ULONG)0 - (constant_time_eq_int(i,2)&1); + y3 = (BN_ULONG)0 - (constant_time_eq_int(i,3)&1); + + for (i = 0; i < top; i++, table += width) { + BN_ULONG acc = 0; + + for (j = 0; j < xstride; j++) { + acc |= ( (table[j + 0 * xstride] & y0) | + (table[j + 1 * xstride] & y1) | + (table[j + 2 * xstride] & y2) | + (table[j + 3 * xstride] & y3) ) + & ((BN_ULONG)0 - (constant_time_eq_int(j,idx)&1)); + } + + b->d[i] = acc; + } + } + + b->top = top; bn_correct_top(b); return 1; } /* * Given a pointer value, compute the next address that is a cache line * multiple. */ #define MOD_EXP_CTIME_ALIGN(x_) \ ((unsigned char*)(x_) + (MOD_EXP_CTIME_MIN_CACHE_LINE_WIDTH - (((BN_ULONG)(x_)) & (MOD_EXP_CTIME_MIN_CACHE_LINE_MASK)))) /* * This variant of BN_mod_exp_mont() uses fixed windows and the special * precomputation memory layout to limit data-dependency to a minimum to * protect secret exponents (cf. the hyper-threading timing attacks pointed * out by Colin Percival, * http://www.daemong-consideredperthreading-considered-harmful/) */ int BN_mod_exp_mont_consttime(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p, const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *in_mont) { int i, bits, ret = 0, idx, window, wvalue; int top; BIGNUM *r; const BIGNUM *aa; BN_MONT_CTX *mont = NULL; int numPowers; unsigned char *powerbufFree = NULL; int powerbufLen = 0; unsigned char *powerbuf = NULL; BIGNUM *computeTemp = NULL, *am = NULL; bn_check_top(a); bn_check_top(p); bn_check_top(m); top = m->top; if (!(m->d[0] & 1)) { BNerr(BN_F_BN_MOD_EXP_MONT_CONSTTIME, BN_R_CALLED_WITH_EVEN_MODULUS); return (0); } bits = BN_num_bits(p); if (bits == 0) { ret = BN_one(rr); return ret; } /* Initialize BIGNUM context and allocate intermediate result */ BN_CTX_start(ctx); r = BN_CTX_get(ctx); if (r == NULL) goto err; /* * Allocate a montgomery context if it was not supplied by the caller. If * this is not done, things will break in the montgomery part. */ if (in_mont != NULL) mont = in_mont; else { if ((mont = BN_MONT_CTX_new()) == NULL) goto err; if (!BN_MONT_CTX_set(mont, m, ctx)) goto err; } /* Get the window size to use with size of p. */ window = BN_window_bits_for_ctime_exponent_size(bits); /* * Allocate a buffer large enough to hold all of the pre-computed powers * of a. */ numPowers = 1 << window; powerbufLen = sizeof(m->d[0]) * top * numPowers; if ((powerbufFree = (unsigned char *)OPENSSL_malloc(powerbufLen + MOD_EXP_CTIME_MIN_CACHE_LINE_WIDTH)) == NULL) goto err; powerbuf = MOD_EXP_CTIME_ALIGN(powerbufFree); memset(powerbuf, 0, powerbufLen); /* * Initialize the intermediate result. Do this early to save double * conversion, once each for a^0 and intermediate result. - */ - if (!BN_to_montgomery(r, BN_value_one(), mont, ctx)) - goto err; - if (!MOD_EXP_CTIME_COPY_TO_PREBUF(r, top, powerbuf, 0, window)) - goto err; - - /* Initialize computeTemp as a^1 with montgomery precalcs */ + */ + if (!BN_to_montgomery(r, BN_value_one(), mont, ctx)) + goto err; + if (!MOD_EXP_CTIME_COPY_TO_PREBUF(r, top, powerbuf, 0, window)) + goto err; + + /* Initialize computeTemp as a^1 with montgomery precalcs */ computeTemp = BN_CTX_get(ctx); am = BN_CTX_get(ctx); if (computeTemp == NULL || am == NULL) goto err; if (a->neg || BN_ucmp(a, m) >= 0) { if (!BN_mod(am, a, m, ctx)) goto err; aa = am; } else aa = a; if (!BN_to_montgomery(am, aa, mont, ctx)) - goto err; - if (!BN_copy(computeTemp, am)) - goto err; - if (!MOD_EXP_CTIME_COPY_TO_PREBUF(am, top, powerbuf, 1, window)) - goto err; - - /* + goto err; + if (!BN_copy(computeTemp, am)) + goto err; + if (!MOD_EXP_CTIME_COPY_TO_PREBUF(am, top, powerbuf, 1, window)) + goto err; + + /* * If the window size is greater than 1, then calculate * val[i=2..2^winsize-1]. Powers are computed as a*a^(i-1) (even powers * could instead be computed as (a^(i/2))^2 to use the slight performance * advantage of sqr over mul). */ if (window > 1) { for (i = 2; i < numPowers; i++) { /* Calculate a^i = a^(i-1) * a */ - if (!BN_mod_mul_montgomery - (computeTemp, am, computeTemp, mont, ctx)) - goto err; - if (!MOD_EXP_CTIME_COPY_TO_PREBUF(computeTemp, top, powerbuf, i, - window)) - goto err; - } - } + if (!BN_mod_mul_montgomery + (computeTemp, am, computeTemp, mont, ctx)) + goto err; + if (!MOD_EXP_CTIME_COPY_TO_PREBUF(computeTemp, top, powerbuf, i, + window)) + goto err; + } + } /* * Adjust the number of bits up to a multiple of the window size. If the * exponent length is not a multiple of the window size, then this pads * the most significant bits with zeros to normalize the scanning loop to * there's no special cases. * NOTE: Making the window size a power of * two less than the native * word size ensures that the padded bits * won't go past the last * word in the internal BIGNUM structure. Going * past the end will * still produce the correct result, but causes a * different branch * to be taken in the BN_is_bit_set function. */ bits = ((bits + window - 1) / window) * window; idx = bits - 1; /* The top bit of the window */ /* * Scan the exponent one window at a time starting from the most * significant bits. */ while (idx >= 0) { wvalue = 0; /* The 'value' of the window */ /* Scan the window, squaring the result as we go */ for (i = 0; i < window; i++, idx--) { if (!BN_mod_mul_montgomery(r, r, r, mont, ctx)) goto err; wvalue = (wvalue << 1) + BN_is_bit_set(p, idx); } /* * Fetch the appropriate pre-computed value from the pre-buf */ if (!MOD_EXP_CTIME_COPY_FROM_PREBUF - (computeTemp, top, powerbuf, wvalue, numPowers)) + (computeTemp, top, powerbuf, wvalue, window)) goto err; /* Multiply the result into the intermediate result */ if (!BN_mod_mul_montgomery(r, r, computeTemp, mont, ctx)) goto err; } /* Convert the final result from montgomery to standard format */ if (!BN_from_montgomery(rr, r, mont, ctx)) goto err; ret = 1; err: if ((in_mont == NULL) && (mont != NULL)) BN_MONT_CTX_free(mont); if (powerbuf != NULL) { OPENSSL_cleanse(powerbuf, powerbufLen); OPENSSL_free(powerbufFree); } if (am != NULL) BN_clear(am); if (computeTemp != NULL) BN_clear(computeTemp); BN_CTX_end(ctx); return (ret); } int BN_mod_exp_mont_word(BIGNUM *rr, BN_ULONG a, const BIGNUM *p, const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *in_mont) { BN_MONT_CTX *mont = NULL; int b, bits, ret = 0; int r_is_one; BN_ULONG w, next_w; BIGNUM *d, *r, *t; BIGNUM *swap_tmp; #define BN_MOD_MUL_WORD(r, w, m) \ (BN_mul_word(r, (w)) && \ (/* BN_ucmp(r, (m)) < 0 ? 1 :*/ \ (BN_mod(t, r, m, ctx) && (swap_tmp = r, r = t, t = swap_tmp, 1)))) /* * BN_MOD_MUL_WORD is only used with 'w' large, so the BN_ucmp test is * probably more overhead than always using BN_mod (which uses BN_copy if * a similar test returns true). */ /* * We can use BN_mod and do not need BN_nnmod because our accumulator is * never negative (the result of BN_mod does not depend on the sign of * the modulus). */ #define BN_TO_MONTGOMERY_WORD(r, w, mont) \ (BN_set_word(r, (w)) && BN_to_montgomery(r, r, (mont), ctx)) if (BN_get_flags(p, BN_FLG_CONSTTIME) != 0) { /* BN_FLG_CONSTTIME only supported by BN_mod_exp_mont() */ BNerr(BN_F_BN_MOD_EXP_MONT_WORD, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED); return -1; } bn_check_top(p); bn_check_top(m); if (!BN_is_odd(m)) { BNerr(BN_F_BN_MOD_EXP_MONT_WORD, BN_R_CALLED_WITH_EVEN_MODULUS); return (0); } if (m->top == 1) a %= m->d[0]; /* make sure that 'a' is reduced */ bits = BN_num_bits(p); if (bits == 0) { /* x**0 mod 1 is still zero. */ if (BN_is_one(m)) { ret = 1; BN_zero(rr); } else ret = BN_one(rr); return ret; } if (a == 0) { BN_zero(rr); ret = 1; return ret; } BN_CTX_start(ctx); d = BN_CTX_get(ctx); r = BN_CTX_get(ctx); t = BN_CTX_get(ctx); if (d == NULL || r == NULL || t == NULL) goto err; if (in_mont != NULL) mont = in_mont; else { if ((mont = BN_MONT_CTX_new()) == NULL) goto err; if (!BN_MONT_CTX_set(mont, m, ctx)) goto err; } r_is_one = 1; /* except for Montgomery factor */ /* bits-1 >= 0 */ /* The result is accumulated in the product r*w. */ w = a; /* bit 'bits-1' of 'p' is always set */ for (b = bits - 2; b >= 0; b--) { /* First, square r*w. */ next_w = w * w; if ((next_w / w) != w) { /* overflow */ if (r_is_one) { if (!BN_TO_MONTGOMERY_WORD(r, w, mont)) goto err; r_is_one = 0; } else { if (!BN_MOD_MUL_WORD(r, w, m)) goto err; } next_w = 1; } w = next_w; if (!r_is_one) { if (!BN_mod_mul_montgomery(r, r, r, mont, ctx)) goto err; } /* Second, multiply r*w by 'a' if exponent bit is set. */ if (BN_is_bit_set(p, b)) { next_w = w * a; if ((next_w / a) != w) { /* overflow */ if (r_is_one) { if (!BN_TO_MONTGOMERY_WORD(r, w, mont)) goto err; r_is_one = 0; } else { if (!BN_MOD_MUL_WORD(r, w, m)) goto err; } next_w = a; } w = next_w; } } /* Finally, set r:=r*w. */ if (w != 1) { if (r_is_one) { if (!BN_TO_MONTGOMERY_WORD(r, w, mont)) goto err; r_is_one = 0; } else { if (!BN_MOD_MUL_WORD(r, w, m)) goto err; } } if (r_is_one) { /* can happen only if a == 1 */ if (!BN_one(rr)) goto err; } else { if (!BN_from_montgomery(rr, r, mont, ctx)) goto err; } ret = 1; err: if ((in_mont == NULL) && (mont != NULL)) BN_MONT_CTX_free(mont); BN_CTX_end(ctx); bn_check_top(rr); return (ret); } /* The old fallback, simple version :-) */ int BN_mod_exp_simple(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, const BIGNUM *m, BN_CTX *ctx) { int i, j, bits, ret = 0, wstart, wend, window, wvalue; int start = 1; BIGNUM *d; /* Table of variables obtained from 'ctx' */ BIGNUM *val[TABLE_SIZE]; if (BN_get_flags(p, BN_FLG_CONSTTIME) != 0) { /* BN_FLG_CONSTTIME only supported by BN_mod_exp_mont() */ BNerr(BN_F_BN_MOD_EXP_SIMPLE, ERR_R_SHOULD_NOT_HAVE_BEEN_CALLED); return -1; } bits = BN_num_bits(p); if (bits == 0) { ret = BN_one(r); return ret; } BN_CTX_start(ctx); d = BN_CTX_get(ctx); val[0] = BN_CTX_get(ctx); if (!d || !val[0]) goto err; if (!BN_nnmod(val[0], a, m, ctx)) goto err; /* 1 */ if (BN_is_zero(val[0])) { BN_zero(r); ret = 1; goto err; } window = BN_window_bits_for_exponent_size(bits); if (window > 1) { if (!BN_mod_mul(d, val[0], val[0], m, ctx)) goto err; /* 2 */ j = 1 << (window - 1); for (i = 1; i < j; i++) { if (((val[i] = BN_CTX_get(ctx)) == NULL) || !BN_mod_mul(val[i], val[i - 1], d, m, ctx)) goto err; } } start = 1; /* This is used to avoid multiplication etc * when there is only the value '1' in the * buffer. */ wvalue = 0; /* The 'value' of the window */ wstart = bits - 1; /* The top bit of the window */ wend = 0; /* The bottom bit of the window */ if (!BN_one(r)) goto err; for (;;) { if (BN_is_bit_set(p, wstart) == 0) { if (!start) if (!BN_mod_mul(r, r, r, m, ctx)) goto err; if (wstart == 0) break; wstart--; continue; } /* * We now have wstart on a 'set' bit, we now need to work out how bit * a window to do. To do this we need to scan forward until the last * set bit before the end of the window */ j = wstart; wvalue = 1; wend = 0; for (i = 1; i < window; i++) { if (wstart - i < 0) break; if (BN_is_bit_set(p, wstart - i)) { wvalue <<= (i - wend); wvalue |= 1; wend = i; } } /* wend is the size of the current window */ j = wend + 1; /* add the 'bytes above' */ if (!start) for (i = 0; i < j; i++) { if (!BN_mod_mul(r, r, r, m, ctx)) goto err; } /* wvalue will be an odd number < 2^window */ if (!BN_mod_mul(r, r, val[wvalue >> 1], m, ctx)) goto err; /* move the 'window' down further */ wstart -= wend + 1; wvalue = 0; start = 0; if (wstart < 0) break; } ret = 1; err: BN_CTX_end(ctx); bn_check_top(r); return (ret); } Index: releng/9.3/sys/conf/newvers.sh =================================================================== --- releng/9.3/sys/conf/newvers.sh (revision 296610) +++ releng/9.3/sys/conf/newvers.sh (revision 296611) @@ -1,160 +1,160 @@ #!/bin/sh - # # Copyright (c) 1984, 1986, 1990, 1993 # The Regents of the University of California. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # 1. Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # 2. Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # 4. Neither the name of the University nor the names of its contributors # may be used to endorse or promote products derived from this software # without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND # ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE # ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE # FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL # DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS # OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) # HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT # LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY # OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF # SUCH DAMAGE. # # @(#)newvers.sh 8.1 (Berkeley) 4/20/94 # $FreeBSD$ TYPE="FreeBSD" REVISION="9.3" -BRANCH="RELEASE-p37" +BRANCH="RELEASE-p38" if [ "X${BRANCH_OVERRIDE}" != "X" ]; then BRANCH=${BRANCH_OVERRIDE} fi RELEASE="${REVISION}-${BRANCH}" VERSION="${TYPE} ${RELEASE}" if [ "X${SYSDIR}" = "X" ]; then SYSDIR=$(dirname $0)/.. fi if [ "X${PARAMFILE}" != "X" ]; then RELDATE=$(awk '/__FreeBSD_version.*propagated to newvers/ {print $3}' \ ${PARAMFILE}) else RELDATE=$(awk '/__FreeBSD_version.*propagated to newvers/ {print $3}' \ ${SYSDIR}/sys/param.h) fi b=share/examples/etc/bsd-style-copyright year=`date '+%Y'` # look for copyright template for bsd_copyright in ../$b ../../$b ../../../$b /usr/src/$b /usr/$b do if [ -r "$bsd_copyright" ]; then COPYRIGHT=`sed \ -e "s/\[year\]/1992-$year/" \ -e 's/\[your name here\]\.* /The FreeBSD Project./' \ -e 's/\[your name\]\.*/The FreeBSD Project./' \ -e '/\[id for your version control system, if any\]/d' \ $bsd_copyright` break fi done # no copyright found, use a dummy if [ X"$COPYRIGHT" = X ]; then COPYRIGHT="/*- * Copyright (c) 1992-$year The FreeBSD Project. * All rights reserved. * */" fi # add newline COPYRIGHT="$COPYRIGHT " LC_ALL=C; export LC_ALL if [ ! -r version ] then echo 0 > version fi touch version v=`cat version` u=${USER:-root} d=`pwd` h=${HOSTNAME:-`hostname`} t=`date` i=`${MAKE:-make} -V KERN_IDENT` compiler_v=$($(${MAKE:-make} -V CC) -v 2>&1 | grep 'version') for dir in /bin /usr/bin /usr/local/bin; do if [ -x "${dir}/svnversion" ] ; then svnversion=${dir}/svnversion break fi done if [ -d "${SYSDIR}/../.git" ] ; then for dir in /bin /usr/bin /usr/local/bin; do if [ -x "${dir}/git" ] ; then git_cmd="${dir}/git --git-dir=${SYSDIR}/../.git" break fi done fi if [ -n "$svnversion" ] ; then echo "$svnversion" svn=`cd ${SYSDIR} && $svnversion` case "$svn" in [0-9]*) svn=" r${svn}" ;; *) unset svn ;; esac fi if [ -n "$git_cmd" ] ; then git=`$git_cmd rev-parse --verify --short HEAD 2>/dev/null` svn=`$git_cmd svn find-rev $git 2>/dev/null` if [ -n "$svn" ] ; then svn=" r${svn}" git="=${git}" else svn=`$git_cmd log | fgrep 'git-svn-id:' | head -1 | \ sed -n 's/^.*@\([0-9][0-9]*\).*$/\1/p'` if [ -z "$svn" ] ; then svn=`$git_cmd log --format='format:%N' | \ grep '^svn ' | head -1 | \ sed -n 's/^.*revision=\([0-9][0-9]*\).*$/\1/p'` fi if [ -n "$svn" ] ; then svn=" r${svn}" git="+${git}" else git=" ${git}" fi fi if $git_cmd --work-tree=${SYSDIR}/.. diff-index \ --name-only HEAD | read dummy; then git="${git}-dirty" fi fi cat << EOF > vers.c $COPYRIGHT #define SCCSSTR "@(#)${VERSION} #${v}${svn}${git}: ${t}" #define VERSTR "${VERSION} #${v}${svn}${git}: ${t}\\n ${u}@${h}:${d}\\n" #define RELSTR "${RELEASE}" char sccs[sizeof(SCCSSTR) > 128 ? sizeof(SCCSSTR) : 128] = SCCSSTR; char version[sizeof(VERSTR) > 256 ? sizeof(VERSTR) : 256] = VERSTR; char compiler_version[] = "${compiler_v}"; char ostype[] = "${TYPE}"; char osrelease[sizeof(RELSTR) > 32 ? sizeof(RELSTR) : 32] = RELSTR; int osreldate = ${RELDATE}; char kern_ident[] = "${i}"; EOF echo $((v + 1)) > version