1 Updating Information for FreeBSD current users
3 This file is maintained and copyrighted by M. Warner Losh <imp@freebsd.org>.
4 See end of file for further details. For commonly done items, please see the
5 COMMON ITEMS: section later in the file. These instructions assume that you
6 basically know what you are doing. If not, then please consult the FreeBSD
9 http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html
11 Items affecting the ports and packages system can be found in
12 /usr/ports/UPDATING. Please read that file before running portupgrade.
14 20150127: p9 FreeBSD-SA-15:02.kmem
17 Fix SCTP SCTP_SS_VALUE kernel memory corruption and disclosure
18 vulnerability. [SA-15:02]
20 Fix SCTP stream reset vulnerability. [SA-15:03]
22 20150114: p8 FreeBSD-SA-15:01.openssl
23 Fix multiple vulnerabilities in OpenSSL. [SA-15:01]
25 20141223: p7 FreeBSD-SA-14:31.ntp
26 FreeBSD-EN-14:13.freebsd-update
28 Fix multiple vulnerabilities in NTP suite. [SA-14:31]
29 Fix directory deletion issue in freebsd-update. [EN-14:13]
31 20141210: p6 FreeBSD-SA-14:28.file
34 Fix multiple vulnerabilities in file(1) and libmagic(3).
37 Fix BIND remote denial of service vulnerability. [SA-14:29]
39 20141104: p5 FreeBSD-SA-14:25.setlogin
43 Fix kernel stack disclosure in setlogin(2) / getlogin(2).
46 Fix remote command execution in ftp(1). [SA-14:26]
48 Fix NFSv4 and ZFS cache consistency issue. [EN-14:12]
50 20141022: p4 FreeBSD-EN-14:10.tzdata
51 FreeBSD-EN-14:11.crypt
53 Time zone data file update. [EN-14:10]
55 Change crypt(3) default hashing algorithm back to DES. [EN-14:11]
57 20141021: p3 FreeBSD-SA-14:20.rtsold
58 FreeBSD-SA-14:21.routed
59 FreeBSD-SA-14:22.namei
60 FreeBSD-SA-14:23.openssl
62 Fix rtsold(8) remote buffer overflow vulnerability. [SA-14:20]
64 Fix routed(8) remote denial of service vulnerability. [SA-14:21]
66 Fix memory leak in sandboxed namei lookup. [SA-14:22]
68 Fix OpenSSL multiple vulnerabilities. [SA-14:23]
70 20140916: p2 FreeBSD-SA-14:19.tcp
71 Fix Denial of Service in TCP packet processing. [SA-14:19]
73 20140909: p1 FreeBSD-SA-14:18.openssl
74 Fix OpenSSL multiple vulnerabilities. [SA-14:18]
80 On i386 and amd64 systems, the onifconsole flag is now set by default
81 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
82 as a login TTY if it is set in the bootloader as an active kernel
83 console. No changes in behavior should result otherwise. To revert to
84 the previous behavior, set ttyu0 to "off" in /etc/ttys.
87 Clang and llvm have been upgraded to 3.4.1 release.
90 Clang and llvm have been upgraded to 3.4 release.
93 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
94 been deprecated and will not be part of FreeBSD 11.0 and later
95 releases. If you use this driver, please consider switching to
96 the nfe(4) driver instead.
99 The behavior of gss_pseudo_random() for the krb5 mechanism
100 has changed, for applications requesting a longer random string
101 than produced by the underlying enctype's pseudo-random() function.
102 In particular, the random string produced from a session key of
103 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
104 be different at the 17th octet and later, after this change.
105 The counter used in the PRF+ construction is now encoded as a
106 big-endian integer in accordance with RFC 4402.
107 __FreeBSD_version is bumped to 902505.
113 Behavior of devfs rules path matching has been changed.
114 Pattern is now always matched against fully qualified devfs
115 path and slash characters must be explicitly matched by
116 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
117 subdirectories must be reviewed.
120 hastctl(8)'s `status' command output changed to terse one-liner format.
121 Scripts using this should switch to `list' command or be rewritten.
124 Fix a bug that allowed a tracing process (e.g. gdb) to write
125 to a memory-mapped file in the traced process's address space
126 even if neither the traced process nor the tracing process had
127 write access to that file.
130 Added ZFS TRIM support which is enabled by default. To disable
131 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
133 Creating new ZFS pools and adding new devices to existing pools
134 first performs a full device level TRIM which can take a significant
135 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
136 to disable this behaviour.
138 ZFS TRIM requires the underlying device support BIO_DELETE which
139 is currently provided by methods such as ATA TRIM and SCSI UNMAP
140 via CAM, which are typically supported by SSD's.
142 Stats for ZFS TRIM can be monitored by looking at the sysctl's
143 under kstat.zfs.misc.zio_trim.
146 `list' command has been added to hastctl(8). For now, it is full
147 equivalent of `status' command.
148 WARNING: in the near future the output of hastctl's status command
149 will change to more terse format. If you use `hastctl status'
150 for parsing in your scripts, switch to `hastctl list'.
153 The mergemaster command now uses the default MAKEOBJDIRPREFIX
154 rather than creating it's own in the temporary directory in
155 order allow access to bootstrapped versions of tools such as
156 install and mtree. When upgrading from version of FreeBSD where
157 the install command does not support -l, you will need to
158 install a new mergemaster command if mergemaster -p is required.
159 This can be accomplished with the command (cd src/usr.sbin/mergemaster
162 Due to the use of the new -l option to install(1) during build
163 and install, you must take care not to directly set the INSTALL
164 make variable in your /etc/make.conf, /etc/src.conf, or on the
165 command line. If you wish to use the -C flag for all installs
166 you may be able to add INSTALL+=-C to /etc/make.conf or
170 Fix a bug that allows NFS clients to issue READDIR on files.
173 The install(1) option -M has changed meaning and now takes an
174 argument that is a file or path to append logs to. In the
175 unlikely event that -M was the last option on the command line
176 and the command line contained at least two files and a target
177 directory the first file will have logs appended to it. The -M
178 option served little practical purpose in the last decade so it's
179 used expected to be extremely rare.
182 A new compression method (lz4) has been merged to. Please refer to
183 zpool-features(7) for more information.
185 Please refer to the "ZFS notes" section of this file for information
186 on upgrading boot ZFS pools.
189 The VFS KBI was changed with the merge of several nullfs
190 optimizations and fixes. All filesystem modules must be
194 With the addition of auditdistd(8), a new auditdistd user is now
195 depended on during installworld. "mergemaster -p" can be used to add
196 the user prior to installworld, as documented in the handbook.
202 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
203 Starting with this version the old system of ZFS pool versioning
204 is superseded by "feature flags". This concept enables forward
205 compatibility against certain future changes in functionality of ZFS
206 pools. The first two read-only compatible "feature flags" for ZFS
207 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
208 For more information read the new zpool-features(7) manual page.
209 Please refer to the "ZFS notes" section of this file for information
210 on upgrading boot ZFS pools.
213 The commit introducing bsd.compiler.mk breaks the traditional
214 building of kernels before this point. Add -m ${SRC}/share/mk
215 (for the right value of SRC) to your command lines to work
216 around; update your useland to a point after this; or use the
217 buildkernel/installkernel top-level targets. See also 20120829.
220 The IPFIREWALL_FORWARD kernel option has been removed. Its
221 functionality now turned on by default.
224 The random(4) support for the VIA hardware random number
225 generator (`PADLOCK') is no longer enabled unconditionally.
226 Add the PADLOCK_RNG option in the custom kernel config if
227 needed. The GENERIC kernels on i386 and amd64 do include the
228 option, so the change only affects the custom kernel
232 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
233 the traditional method, you must update your system with an installworld
234 before the kernel will build. The documented make buildkernel/installkernel
235 interfaces (coupled with fresh make kernel-toolchain) continue to work.
238 The sparc64 ZFS loader has been changed to no longer try to auto-
239 detect ZFS providers based on diskN aliases but now requires these
240 to be explicitly listed in the OFW boot-device environment variable.
243 Now unix domain sockets behave "as expected" on nullfs(5). Previously
244 nullfs(5) did not pass through all behaviours to the underlying layer,
245 as a result if we bound to a socket on the lower layer we could connect
246 only to the lower path; if we bound to the upper layer we could connect
247 only to the upper path. The new behavior is one can connect to both the
248 lower and the upper paths regardless what layer path one binds to.
251 The acpi_wmi(4) status device /dev/wmistat has been renamed to
255 A new VOP_ADVISE() was added to support posix_fadvise(2). All
256 filesystem modules must be recompiled.
259 The interface of the VOP_VPTOCNP(9) changed, now the returned
260 vnode shall be referenced, previously it was required to be
261 only held. All in-tree filesystems are converted.
267 The broken amd(4) driver has been replaced with esp(4) in the amd64,
268 i386 and pc98 GENERIC kernel configuration files.
271 This commit modifies vfs_register() so that it uses a hash
272 calculation to set vfc_typenum, which is enabled by default.
273 The first time a system is booted after this change, the
274 vfc_typenum values will change for all file systems. The
275 main effect of this is a change to the NFS server file handles
276 for file systems that use vfc_typenum in their fsid, such as ZFS.
277 It will, however, prevent vfc_typenum from changing when file
278 systems are loaded in a different order for subsequent reboots.
279 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
280 until you are ready to remount all NFS clients after a reboot.
283 Bump the shared library version numbers for libraries that
284 do not use symbol versioning, have changed the ABI compared
285 to stable/8 and which shared library version was not bumped.
286 Done as part of 9.0-RELEASE cycle.
289 During the merge of Capsicum features, the fget(9) KPI was modified.
290 This may require the rebuilding of out-of-tree device drivers --
291 issues have been reported specifically with the nVidia device driver.
292 __FreeBSD_version is bumped to 900041.
294 Also, there is a period between 20110811 and 20110814 where the
295 special devices /dev/{stdin,stdout,stderr} did not work correctly.
296 Building world from a kernel during that window may not work.
299 The packet filter (pf) code has been updated to OpenBSD 4.5.
300 You need to update userland tools to be in sync with kernel.
301 This update breaks backward compatibility with earlier pfsync(4)
302 versions. Care must be taken when updating redundant firewall setups.
305 The following sysctls and tunables are retired on x86 platforms:
307 machdep.hlt_logical_cpus
308 The following sysctl is retired:
309 machdep.hyperthreading_allowed
310 The sysctls were supposed to provide a way to dynamically offline and
311 online selected CPUs on x86 platforms, but the implementation has not
312 been reliable especially with SCHED_ULE scheduler.
313 machdep.hyperthreading_allowed tunable is still available to ignore
314 hyperthreading CPUs at OS level.
315 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
316 where X is an APIC ID of a CPU. Be advised, though, that disabling
317 CPUs in non-uniform fashion will result in non-uniform topology and
318 may lead to sub-optimal system performance with SCHED_ULE, which is
322 cpumask_t type is retired and cpuset_t is used in order to describe
326 Changes to ifconfig(8) for dynamic address family detection mandate
327 that you are running a kernel of 20110525 or later. Make sure to
328 follow the update procedure to boot a new kernel before installing
332 Support for sun4v architecture is officially dropped
335 Several KPI breaking changes have been committed to the mii(4) layer,
336 the PHY drivers and consequently some Ethernet drivers using mii(4).
337 This means that miibus.ko and the modules of the affected Ethernet
338 drivers need to be recompiled.
340 Note to kernel developers: Given that the OUI bit reversion problem
341 was fixed as part of these changes all mii(4) commits related to OUIs,
342 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
343 handling, no longer can be merged verbatim to stable/8 and previous
347 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
348 into their kernel configurations along with 'device pci'.
351 The default NFS client is now the new NFS client, so fstype "newnfs"
352 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
353 Although mounts via fstype "nfs" will usually work without userland
354 changes, it is recommended that the mount(8) and mount_nfs(8)
355 commands be rebuilt from sources and that a link to mount_nfs called
356 mount_oldnfs be created. The new client is compiled into the
357 kernel with "options NFSCL" and this is needed for diskless root
358 file systems. The GENERIC kernel configs have been changed to use
359 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
360 To use the regular/old client, you can "mount -t oldnfs ...". For
361 a diskless root file system, you must also include a line like:
363 vfs.root.mountfrom="oldnfs:"
365 in the boot/loader.conf on the root fs on the NFS server to make
366 a diskless root fs use the old client.
369 The GENERIC kernels for all architectures now default to the new
370 CAM-based ATA stack. It means that all legacy ATA drivers were
371 removed and replaced by respective CAM drivers. If you are using
372 ATA device names in /etc/fstab or other places, make sure to update
373 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
374 where 'Y's are the sequential numbers starting from zero for each type
375 in order of detection, unless configured otherwise with tunables,
376 see cam(4)). There will be symbolic links created in /dev/ to map
377 old adX devices to the respective adaY. They should provide basic
378 compatibility for file systems mounting in most cases, but they do
379 not support old user-level APIs and do not have respective providers
380 in GEOM. Consider using updated management tools with new device names.
382 It is possible to load devices ahci, ata, siis and mvs as modules,
383 but option ATA_CAM should remain in kernel configuration to make ata
384 module work as CAM driver supporting legacy ATA controllers. Device ata
385 still can be used in modular fashion (atacore + ...). Modules atadisk
386 and atapi* are not used and won't affect operation in ATA_CAM mode.
387 Note that to use CAM-based ATA kernel should include CAM devices
388 scbus, pass, da (or explicitly ada), cd and optionally others. All of
389 them are parts of the cam module.
391 ataraid(4) functionality is now supported by the RAID GEOM class.
392 To use it you can load geom_raid kernel module and use graid(8) tool
393 for management. Instead of /dev/arX device names, use /dev/raid/rX.
395 No kernel config options or code have been removed, so if a problem
396 arises, please report it and optionally revert to the old ATA stack.
397 In order to do it you can remove from the kernel config:
402 , and instead add back:
403 device atadisk # ATA disk drives
404 device ataraid # ATA RAID drives
405 device atapicd # ATAPI CDROM drives
406 device atapifd # ATAPI floppy drives
407 device atapist # ATAPI tape drives
410 The default NFS server has been changed to the new server, which
411 was referred to as the experimental server. If you need to switch
412 back to the old NFS server, you must now put the "-o" option on
413 both the mountd and nfsd commands. This can be done using the
414 mountd_flags and nfs_server_flags rc.conf variables until an
415 update to the rc scripts is committed, which is coming soon.
418 The GNU Objective-C runtime library (libobjc), and other Objective-C
419 related components have been removed from the base system. If you
420 require an Objective-C library, please use one of the available ports.
423 ath(4) has been split into bus- and device- modules. if_ath contains
424 the HAL, the TX rate control and the network device code. if_ath_pci
425 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
426 contains the AHB glue. Users need to load both if_ath_pci and if_ath
427 in order to use ath on everything else.
429 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
430 need to load if_ath and if_ath_pci for ath(4) operation.
433 As part of the replacement of sysinstall, the process of building
434 release media has changed significantly. For details, please re-read
435 release(7), which has been updated to reflect the new build process.
438 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
439 is the last available version under GPLv2. It brings a number of new
440 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
441 4.1 and SSE 4.2), better support for powerpc64, a number of new
442 directives, and lots of other small improvements. See the ChangeLog
443 file in contrib/binutils for the full details.
446 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
447 compliant, and will now use half of hash for authentication.
448 This will break interoperability with all stacks (including all
449 actual FreeBSD versions) who implement
450 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
452 The only workaround with such peers is to use another HMAC
453 algorithm for IPsec ("phase 2") authentication.
456 Remove the uio_yield prototype and symbol. This function has
457 been misnamed since it was introduced and should not be
458 globally exposed with this name. The equivalent functionality
459 is now available using kern_yield(curthread->td_user_pri).
460 The function remains undocumented.
463 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
464 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
465 for scalar sysctls is defined but disabled. Code that needs
466 UQUAD to pass the type checking that must compile on older
467 systems where the define is not present can check against
468 __FreeBSD_version >= 900030.
470 The system dialog(1) has been replaced with a new version previously
471 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
472 with the previous version, but the libdialog associated with it has
473 a largely incompatible API. As such, the original version of libdialog
474 will be kept temporarily as libodialog, until its base system consumers
475 are replaced or updated. Bump __FreeBSD_version to 900030.
478 If you are trying to run make universe on a -stable system, and you get
479 the following warning:
480 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
481 unknown. config(8) likely too old."
482 or something similar to it, then you must upgrade your -stable system
483 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
484 stable/8) or build the config from the latest stable/8 branch and
485 install it on your system.
487 Prior to this date, building a current universe on 8-stable system from
488 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
489 error in the first kernel build phase. A new config on those old
490 systems will fix that problem for older versions of -current.
493 The TCP stack has been modified to allow Khelp modules to interact with
494 it via helper hook points and store per-connection data in the TCP
495 control block. Bump __FreeBSD_version to 900029. User space tools that
496 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
500 Generic IEEE 802.3 annex 31B full duplex flow control support has been
501 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
502 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
503 to take advantage of it instead of using custom implementations. This
504 means that these drivers now no longer unconditionally advertise
505 support for flow control but only do so if flow control is a selected
506 media option. This was implemented in the generic support that way in
507 order to allow flow control to be switched on and off via ifconfig(8)
508 with the PHY specific default to typically off in order to protect
509 from unwanted effects. Consequently, if you used flow control with
510 one of the above mentioned drivers you now need to explicitly enable
512 ifconfig bge0 media auto mediaopt flowcontrol
514 Along with the above mentioned changes generic support for setting
515 1000baseT master mode also has been added and brgphy(4), ciphy(4),
516 e1000phy(4) as well as ip1000phy(4) have been converted to take
517 advantage of it. This means that these drivers now no longer take the
518 link0 parameter for selecting master mode but the master media option
519 has to be used instead, for example like in the following:
520 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
522 Selection of master mode now is also available with all other PHY
523 drivers supporting 1000baseT.
526 The TCP stack has received a significant update to add support for
527 modularised congestion control and generally improve the clarity of
528 congestion control decisions. Bump __FreeBSD_version to 900025. User
529 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
530 sockstat) need to be recompiled.
533 The man(1) utility has been replaced by a new version that no longer
534 uses /etc/manpath.config. Please consult man.conf(5) for how to
535 migrate local entries to the new format.
538 The copyright strings printed by login(1) and sshd(8) at the time of a
539 new connection have been removed to follow other operating systems and
543 A workaround for a fixed ld bug has been removed in kernel code,
544 so make sure that your system ld is built from sources after
545 revision 210245 from 2010-07-19 (r211583 if building head kernel
546 on stable/8, r211584 for stable/7; both from 2010-08-21).
547 A symptom of incorrect ld version is different addresses for
548 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
551 The $ipv6_prefer variable in rc.conf(5) has been split into
552 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
554 The $ip6addrctl_policy is a variable to choose a pre-defined
555 address selection policy set by ip6addrctl(8). A value
556 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
559 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
560 flag (see an entry of 20090926) is set on an interface with no
561 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
562 security reason. If you want IPv6 link-local address on all
563 interfaces by default, set this to "YES".
565 The old ipv6_prefer="YES" is equivalent to
566 ipv6_activate_all_interfaces="YES" and
567 ip6addrctl_policy="ipv6_prefer".
570 DTrace has grown support for userland tracing. Due to this, DTrace is
571 now i386 and amd64 only.
572 dtruss(1) is now installed by default on those systems and a new
573 kernel module is needed for userland tracing: fasttrap.
574 No changes to your kernel config file are necessary to enable
575 userland tracing, but you might consider adding 'STRIP=' and
576 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
577 to have informative userland stack traces in DTrace (ustack).
580 The acpi_aiboost(4) driver has been removed in favor of the new
581 aibs(4) driver. You should update your kernel configuration file.
584 BSD grep has been imported to the base system and it is built by
585 default. It is completely BSD licensed, highly GNU-compatible, uses
586 less memory than its GNU counterpart and has a small codebase.
587 However, it is slower than its GNU counterpart, which is mostly
588 noticeable for larger searches, for smaller ones it is measurable
589 but not significant. The reason is complex, the most important factor
590 is that we lack a modern and efficient regex library and GNU
591 overcomes this by optimizing the searches internally. Future work
592 on improving the regex performance is planned, for the meantime,
593 users that need better performance, can build GNU grep instead by
594 setting the WITH_GNU_GREP knob.
597 Due to the import of powerpc64 support, all existing powerpc kernel
598 configuration files must be updated with a machine directive like this:
599 machine powerpc powerpc
601 In addition, an updated config(8) is required to build powerpc kernels
605 A new version of ZFS (version 15) has been merged to -HEAD.
606 This version uses a python library for the following subcommands:
607 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
608 For full functionality of these commands the following port must
609 be installed: sysutils/py-zfs
612 'vm_page's are now hashed by physical address to an array of mutexes.
613 Currently this is only used to serialize access to hold_count. Over
614 time the page queue mutex will be peeled away. This changes the size
615 of pmap on every architecture. And requires all callers of vm_page_hold
616 and vm_page_unhold to be updated.
619 WITH_CTF can now be specified in src.conf (not recommended, there
620 are some problems with static executables), make.conf (would also
621 affect ports which do not use GNU make and do not override the
622 compile targets) or in the kernel config (via "makeoptions
624 When WITH_CTF was specified there before this was silently ignored,
625 so make sure that WITH_CTF is not used in places which could lead
626 to unwanted behavior.
629 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
630 to allow 32-bit compatibility on non-x86 platforms. All kernel
631 configurations on amd64 and ia64 platforms using these options must
632 be modified accordingly.
635 The utmp user accounting database has been replaced with utmpx,
636 the user accounting interface standardized by POSIX.
637 Unfortunately the semantics of utmp and utmpx don't match,
638 making it practically impossible to support both interfaces.
639 The user accounting database is used by tools like finger(1),
640 last(1), talk(1), w(1) and ac(8).
642 All applications in the base system use utmpx. This means only
643 local binaries (e.g. from the ports tree) may still use these
644 utmp database files. These applications must be rebuilt to make
647 After the system has been upgraded, it is safe to remove the old
648 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
649 assuming their contents is of no importance anymore. Old wtmp
650 databases can only be used by last(1) and ac(8) after they have
651 been converted to the new format using wtmpcvt(1).
654 Introduce the kernel thread "deadlock resolver" (which can be enabled
655 via the DEADLKRES option, see NOTES for more details) and the
656 sleepq_type() function for sleepqueues.
659 The rc.firewall and rc.firewall6 were unified, and
660 rc.firewall6 and rc.d/ip6fw were removed.
661 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
662 variables are obsoleted. Instead, the following new rc
663 variables are added to rc.d/ipfw:
665 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
666 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
667 firewall_simple_onet_ipv6, firewall_trusted_ipv6
669 The meanings correspond to the relevant IPv4 variables.
675 The default terminal emulation for syscons(4) has been changed
676 from cons25 to xterm on all platforms except pc98. This means
677 that the /etc/ttys file needs to be updated to ensure correct
678 operation of applications on the console.
680 The terminal emulation style can be toggled per window by using
681 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
682 options can be used to change the compile-time default back to
685 To prevent graphical artifacts, make sure the TERM environment
686 variable is set to match the terminal emulation that is being
687 performed by syscons(4).
690 The layout of the structure ieee80211req_scan_result has changed.
691 Applications that require wireless scan results (e.g. ifconfig(8))
692 from net80211 need to be recompiled.
694 Applications such as wpa_supplicant(8) may require a full world
695 build without using NO_CLEAN in order to get synchronized with the
699 The iwn(4) driver has been updated to support the 5000 and 5150 series.
700 There's one kernel module for each firmware. Adding "device iwnfw"
701 to the kernel configuration file means including all three firmware
702 images inside the kernel. If you want to include just the one for
703 your wireless card, use the devices iwn4965fw, iwn5000fw or
707 The rc.d/network_ipv6, IPv6 configuration script has been integrated
708 into rc.d/netif. The changes are the following:
710 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
711 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
712 Note that both variables need the "inet6" keyword at the head.
714 Do not set $ipv6_network_interfaces manually if you do not
715 understand what you are doing. It is not needed in most cases.
717 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
720 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
721 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
723 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
724 all configured interfaces have "inet6 accept_rtadv" in the
725 $ifconfig_IF_ipv6. These are for backward compatibility.
727 3. A new variable $ipv6_prefer has been added. If NO, IPv6
728 functionality of interfaces with no corresponding
729 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
730 and the default address selection policy of ip6addrctl(8)
731 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
732 Note that if you want to configure IPv6 functionality on the
733 disabled interfaces after boot, first you need to clear the flag by
734 using ifconfig(8) like:
736 ifconfig em0 inet6 -ifdisabled
738 If YES, the default address selection policy is set as
741 The default value of $ipv6_prefer is NO.
743 4. If your system need to receive Router Advertisement messages,
744 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
745 scripts automatically invoke rtsol(8) when the interface becomes
746 UP. The Router Advertisement messages are used for SLAAC
747 (State-Less Address AutoConfiguration).
750 802.11s D3.03 support was committed. This is incompatible with the
751 previous code, which was based on D3.0.
754 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
755 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
756 control whether accepting Router Advertisement messages or not.
757 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
758 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
759 The ifconfig(8) utility now supports these flags.
762 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
763 mount(8) and -a option for df(1) to see them.
766 The old tunable hw.bus.devctl_disable has been superseded by
767 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
768 replaced by hw.bus.devctl_queue=0. The default for this new tunable
772 Remove the option STOP_NMI. The default action is now to use NMI only
773 for KDB via the newly introduced function stop_cpus_hard() and
774 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
777 The stable/8 branch created in subversion. This corresponds to the
778 RELENG_8 branch in CVS.
781 Bump the shared library version numbers for all libraries that do not
782 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
783 __FreeBSD_version to 800105.
786 Due to changes in the implementation of virtual network stack support,
787 all network-related kernel modules must be recompiled. As this change
788 breaks the ABI, bump __FreeBSD_version to 800104.
791 The TOE interface to the TCP syncache has been modified to remove
792 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
793 The cxgb driver is the only TOE consumer affected by this change, and
794 needs to be recompiled along with the kernel. As this change breaks
795 the ABI, bump __FreeBSD_version to 800103.
798 Padding has been added to struct tcpcb, sackhint and tcpstat in
799 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
800 maintaining the ABI. However, this change breaks the ABI, so bump
801 __FreeBSD_version to 800102. User space tools that rely on the size of
802 any of these structs (e.g. sockstat) need to be recompiled.
805 The NFS_LEGACYRPC option has been removed along with the old kernel
806 RPC implementation that this option selected. Kernel configurations
807 may need to be adjusted.
810 The network interface device nodes at /dev/net/<interface> have been
811 removed. All ioctl operations can be performed the normal way using
812 routing sockets. The kqueue functionality can generally be replaced
813 with routing sockets.
816 The documentation from the FreeBSD Documentation Project (Handbook,
817 FAQ, etc.) is now installed via packages by sysinstall(8) and under
818 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
821 The ABI of various structures related to the SYSV IPC API have been
822 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
823 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
827 Layout of struct vnet has changed as routing related variables were
828 moved to their own Vimage module. Modules need to be recompiled. Bump
829 __FreeBSD_version to 800099.
832 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
833 respectively. As long as no more than 16 groups per process are used,
834 no changes should be visible. When more than 16 groups are used, old
835 binaries may fail if they call getgroups() or getgrouplist() with
836 statically sized storage. Recompiling will work around this, but
837 applications should be modified to use dynamically allocated storage
838 for group arrays as POSIX.1-2008 does not cap an implementation's
839 number of supported groups at NGROUPS_MAX+1 as previous versions did.
841 NFS and portalfs mounts may also be affected as the list of groups is
842 truncated to 16. Users of NFS who use more than 16 groups, should
843 take care that negative group permissions are not used on the exported
844 file systems as they will not be reliable unless a GSSAPI based
845 authentication method is used.
848 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
849 option compiles in the support for adaptive spinning for lockmgrs
850 which want to enable it. The lockinit() function now accepts the flag
851 LK_ADAPTIVE in order to make the lock object subject to adaptive
852 spinning when both held in write and read mode.
855 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
856 changed. User applications that use this ioctl need to be rebuilt.
859 The layout of struct thread has changed. Kernel and modules need to
863 The layout of structs ifnet, domain, protosw and vnet_net has changed.
864 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
867 window(1) has been removed from the base system. It can now be
868 installed from ports. The port is called misc/window.
871 The way we are storing and accessing `routing table' entries has
872 changed. Programs reading the FIB, like netstat, need to be
876 A new netisr implementation has been added for FreeBSD 8. Network
877 file system modules, such as igmp, ipdivert, and others, should be
879 Bump __FreeBSD_version to 800096.
882 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
886 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
887 Bump __FreeBSD_version to 800094.
890 Add mnt_xflag field to 'struct mount'. File system modules need to be
892 Bump __FreeBSD_version to 800093.
895 The compiling option ADAPTIVE_SX has been retired while it has been
896 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
897 The KPI for sx_init_flags() changes as accepting flags:
898 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
899 been introduced in order to handle the reversed logic.
900 Bump __FreeBSD_version to 800092.
903 Add support for hierarchical jails. Remove global securelevel.
904 Bump __FreeBSD_version to 800091.
907 The layout of struct vnet_net has changed, therefore modules
909 Bump __FreeBSD_version to 800090.
912 The newly imported zic(8) produces a new format in the output. Please
913 run tzsetup(8) to install the newly created data to /etc/localtime.
916 The sysctl tree for the usb stack has renamed from hw.usb2.* to
917 hw.usb.* and is now consistent again with previous releases.
920 802.11 monitor mode support was revised and driver api's were changed.
921 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
922 of DLT_IEEE802_11. No user-visible data structures were changed but
923 applications that use DLT_IEEE802_11 may require changes.
924 Bump __FreeBSD_version to 800088.
927 The layout of the following structs has changed: sysctl_oid,
928 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
929 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
930 panics may be experienced. World rebuild is required for
931 correctly checking networking state from userland.
932 Bump __FreeBSD_version to 800085.
935 MLDv2 and Source-Specific Multicast (SSM) have been merged
936 to the IPv6 stack. VIMAGE hooks are in but not yet used.
937 The implementation of SSM within FreeBSD's IPv6 stack closely
938 follows the IPv4 implementation.
940 For kernel developers:
942 * The most important changes are that the ip6_output() and
943 ip6_input() paths no longer take the IN6_MULTI_LOCK,
944 and this lock has been downgraded to a non-recursive mutex.
946 * As with the changes to the IPv4 stack to support SSM, filtering
947 of inbound multicast traffic must now be performed by transport
948 protocols within the IPv6 stack. This does not apply to TCP and
949 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
951 * The KPIs used by IPv6 multicast are similar to those used by
952 the IPv4 stack, with the following differences:
953 * im6o_mc_filter() is analogous to imo_multicast_filter().
954 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
955 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
956 * IN6_LOOKUP_MULTI() has been deprecated and removed.
957 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
958 for MLDv1 have an additional 'timer' argument which is used to
959 jitter the initial membership report for the solicited-node
960 multicast membership on-link.
961 * This is not strictly needed for MLDv2, which already jitters
962 its report transmissions. However, the 'timer' argument is
963 preserved in case MLDv1 is active on the interface.
965 * The KAME linked-list based IPv6 membership implementation has
966 been refactored to use a vector similar to that used by the IPv4
968 Code which maintains a list of its own multicast memberships
969 internally, e.g. carp, has been updated to reflect the new
972 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
973 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
974 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
975 implementation constraint which needs to be addressed in HEAD.
977 For application developers:
979 * The changes are broadly similar to those made for the IPv4
982 * The use of IPv4 and IPv6 multicast socket options on the same
983 socket, using mapped addresses, HAS NOT been tested or supported.
985 * There are a number of issues with the implementation of various
986 IPv6 multicast APIs which need to be resolved in the API surface
987 before the implementation is fully compatible with KAME userland
988 use, and these are mostly to do with interface index treatment.
990 * The literature available discusses the use of either the delta / ASM
991 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
992 using setsourcefilter(3)/getsourcefilter(3). For more information
993 please refer to RFC 3768, 'Socket Interface Extensions for
994 Multicast Source Filters'.
996 * Applications which use the published RFC 3678 APIs should be fine.
998 For systems administrators:
1000 * The mtest(8) utility has been refactored to support IPv6, in
1001 addition to IPv4. Interface addresses are no longer accepted
1002 as arguments, their names must be used instead. The utility
1003 will map the interface name to its first IPv4 address as
1004 returned by getifaddrs(3).
1006 * The ifmcstat(8) utility has also been updated to print the MLDv2
1007 endpoint state and source filter lists via sysctl(3).
1009 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
1010 loopback of IPv6 multicast datagrams by default; it defaults to 1
1011 to preserve the existing behaviour. Disabling multicast loopback is
1012 recommended for optimal system performance.
1014 * The IPv6 MROUTING code has been changed to examine this sysctl
1015 instead of attempting to perform a group lookup before looping
1016 back forwarded datagrams.
1018 Bump __FreeBSD_version to 800084.
1021 Implement low-level Bluetooth HCI API.
1022 Bump __FreeBSD_version to 800083.
1025 The layout of struct malloc_type, used by modules to register new
1026 memory allocation types, has changed. Most modules will need to
1027 be rebuilt or panics may be experienced.
1028 Bump __FreeBSD_version to 800081.
1031 Anticipate overflowing inp_flags - add inp_flags2.
1032 This changes most offsets in inpcb, so checking v4 connection
1033 state will require a world rebuild.
1034 Bump __FreeBSD_version to 800080.
1037 Add an llentry to struct route and struct route_in6. Modules
1038 embedding a struct route will need to be recompiled.
1039 Bump __FreeBSD_version to 800079.
1042 The size of rt_metrics_lite and by extension rtentry has changed.
1043 Networking administration apps will need to be recompiled.
1044 The route command now supports show as an alias for get, weighting
1045 of routes, sticky and nostick flags to alter the behavior of stateful
1047 Bump __FreeBSD_version to 800078.
1050 Do not use Giant for kbdmux(4) locking. This is wrong and
1051 apparently causing more problems than it solves. This will
1052 re-open the issue where interrupt handlers may race with
1053 kbdmux(4) in polling mode. Typical symptoms include (but
1054 not limited to) duplicated and/or missing characters when
1055 low level console functions (such as gets) are used while
1056 interrupts are enabled (for example geli password prompt,
1057 mountroot prompt etc.). Disabling kbdmux(4) may help.
1060 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1061 kernel modules referencing any of the above need to be recompiled.
1062 Bump __FreeBSD_version to 800075.
1065 GEOM_PART has become the default partition slicer for storage devices,
1066 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1067 introduces some changes:
1069 MSDOS/EBR: the devices created from MSDOS extended partition entries
1070 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1071 to devices with offset-based names. fstabs may need to be modified.
1073 BSD: the "geometry does not match label" warning is harmless in most
1074 cases but it points to problems in file system misalignment with
1075 disk geometry. The "c" partition is now implicit, covers the whole
1076 top-level drive and cannot be (mis)used by users.
1078 General: Kernel dumps are now not allowed to be written to devices
1079 whose partition types indicate they are meant to be used for file
1080 systems (or, in case of MSDOS partitions, as something else than
1083 Most of these changes date approximately from 200812.
1086 The uscanner(4) driver has been removed from the kernel. This follows
1087 Linux removing theirs in 2.6 and making libusb the default interface
1088 (supported by sane).
1091 The multicast forwarding code has been cleaned up. netstat(1)
1092 only relies on KVM now for printing bandwidth upcall meters.
1093 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1094 ip6_mroute_mod respectively. The config(5) options for statically
1095 compiling this code remain the same, i.e. 'options MROUTING'.
1098 Support for the IFF_NEEDSGIANT network interface flag has been
1099 removed, which means that non-MPSAFE network device drivers are no
1100 longer supported. In particular, if_ar, if_sr, and network device
1101 drivers from the old (legacy) USB stack can no longer be built or
1105 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1106 a bunch of new language catalog files have also been added.
1107 This means that some common libc messages are now localized and
1108 they depend on the LC_MESSAGES environmental variable.
1111 The k8temp(4) driver has been renamed to amdtemp(4) since
1112 support for Family 10 and Family 11 CPU families was added.
1115 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1116 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1118 For kernel developers, the most important changes are that the
1119 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1120 and this lock has been downgraded to a non-recursive mutex.
1122 Transport protocols (UDP, Raw IP) are now responsible for filtering
1123 inbound multicast traffic according to group membership and source
1124 filters. The imo_multicast_filter() KPI exists for this purpose.
1125 Transports which do not use multicast (SCTP, TCP) already reject
1126 multicast by default. Forwarding and receive performance may improve
1127 as a mutex acquisition is no longer needed in the ip_input()
1128 low-level input path. in_addmulti() and in_delmulti() are shimmed
1129 to new KPIs which exist to support SSM in-kernel.
1131 For application developers, it is recommended that loopback of
1132 multicast datagrams be disabled for best performance, as this
1133 will still cause the lock to be taken for each looped-back
1134 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1135 be tuned to 0 to disable loopback by default; it defaults to 1
1136 to preserve the existing behaviour.
1138 For systems administrators, to obtain best performance with
1139 multicast reception and multiple groups, it is always recommended
1140 that a card with a suitably precise hash filter is used. Hash
1141 collisions will still result in the lock being taken within the
1142 transport protocol input path to check group membership.
1144 If deploying FreeBSD in an environment with IGMP snooping switches,
1145 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1146 enabled; this forces 224.0.0.0/24 group membership to be announced
1149 The size of 'struct igmpstat' has changed; netstat needs to be
1150 recompiled to reflect this.
1151 Bump __FreeBSD_version to 800070.
1154 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1155 updated to use it. This requires a buildworld/installworld in order to
1156 update the library and dependencies (usbconfig, etc). Its advisable to
1157 rebuild all ports which uses libusb. More specific directions are given
1158 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1159 libusb are no longer required and can be removed.
1162 A workaround is committed to allow the creation of System V shared
1163 memory segment of size > 2 GB on the 64-bit architectures.
1164 Due to a limitation of the existing ABI, the shm_segsz member
1165 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1166 wrong for large segments. Note that limits must be explicitly
1167 raised to allow such segments to be created.
1170 The layout of struct ifnet has changed, requiring a rebuild of all
1171 network device driver modules.
1174 The /dev handling for the new USB stack has changed, a
1175 buildworld/installworld is required for libusb20.
1178 The new USB2 stack has now been permanently moved in and all kernel and
1179 module names reverted to their previous values (eg, usb, ehci, ohci,
1180 ums, ...). The old usb stack can be compiled in by prefixing the name
1181 with the letter 'o', the old usb modules have been removed.
1182 Updating entry 20090216 for xorg and 20090215 for libmap may still
1186 The rc.conf(5) option if_up_delay has been renamed to
1187 defaultroute_delay to better reflect its purpose. If you have
1188 customized this setting in /etc/rc.conf you need to update it to
1192 xorg 7.4 wants to configure its input devices via hald which does not
1193 yet work with USB2. If the keyboard/mouse does not work in xorg then
1195 Option "AllowEmptyInput" "off"
1196 to your ServerLayout section. This will cause X to use the configured
1197 kbd and mouse sections from your xorg.conf.
1200 The GENERIC kernels for all architectures now default to the new USB2
1201 stack. No kernel config options or code have been removed so if a
1202 problem arises please report it and optionally revert to the old USB
1203 stack. If you are loading USB kernel modules or have a custom kernel
1204 that includes GENERIC then ensure that usb names are also changed over,
1205 eg uftdi -> usb2_serial_ftdi.
1207 Older programs linked against the ports libusb 0.1 need to be
1208 redirected to the new stack's libusb20. /etc/libmap.conf can
1210 # Map old usb library to new one for usb2 stack
1211 libusb-0.1.so.8 libusb20.so.1
1214 All USB ethernet devices now attach as interfaces under the name ueN
1215 (eg. ue0). This is to provide a predictable name as vendors often
1216 change usb chipsets in a product without notice.
1219 The ichsmb(4) driver has been changed to require SMBus slave
1220 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1221 All of the other SMBus controller drivers require left-justified
1222 slave addresses, so this change makes all the drivers provide the
1226 INET6 statistics (struct ip6stat) was updated.
1227 netstat(1) needs to be recompiled.
1230 NTFS has been removed from GENERIC kernel on amd64 to match
1231 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1232 will load ntfs.ko module automatically when NTFS support is
1233 actually needed, unless ntfs.ko is not installed or security
1234 level prohibits loading kernel modules. If either is the case,
1235 "options NTFS" has to be added into kernel config.
1238 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1239 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1240 800061. User space tools that rely on the size of struct tcpcb in
1241 tcp_var.h (e.g. sockstat) need to be recompiled.
1244 ng_tty(4) module updated to match the new TTY subsystem.
1245 Due to API change, user-level applications must be updated.
1246 New API support added to mpd5 CVS and expected to be present
1247 in next mpd5.3 release.
1250 With __FreeBSD_version 800060 the makefs tool is part of
1251 the base system (it was a port).
1254 The afdata and ifnet locks have been changed from mutexes to
1255 rwlocks, network modules will need to be re-compiled.
1258 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1259 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1260 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1261 architecture and 40 bytes on 64-bit architecture. The userland
1262 applications "arp" and "ndp" have been updated accordingly.
1263 The output from "netstat -r" shows only routing entries and
1264 none of the L2 information.
1267 __FreeBSD_version 800057 marks the switchover from the
1268 binary ath hal to source code. Users must add the line:
1270 options AH_SUPPORT_AR5416
1272 to their kernel config files when specifying:
1276 The ath_hal module no longer exists; the code is now compiled
1277 together with the driver in the ath module. It is now
1278 possible to tailor chip support (i.e. reduce the set of chips
1279 and thereby the code size); consult ath_hal(4) for details.
1282 __FreeBSD_version 800054 adds memory barriers to
1283 <machine/atomic.h>, new interfaces to ifnet to facilitate
1284 multiple hardware transmit queues for cards that support
1285 them, and a lock-less ring-buffer implementation to
1286 enable drivers to more efficiently manage queueing of
1290 A new version of ZFS (version 13) has been merged to -HEAD.
1291 This version has zpool attribute "listsnapshots" off by
1292 default, which means "zfs list" does not show snapshots,
1293 and is the same as Solaris behavior.
1296 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1299 The uhci, ohci, ehci and slhci USB Host controller drivers have
1300 been put into separate modules. If you load the usb module
1301 separately through loader.conf you will need to load the
1302 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1303 controller add the following to loader.conf:
1309 The ABI used by the PMC toolset has changed. Please keep
1310 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1314 atapci kernel module now includes only generic PCI ATA
1315 driver. AHCI driver moved to ataahci kernel module.
1316 All vendor-specific code moved into separate kernel modules:
1317 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1318 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1319 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1320 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1323 The TTY subsystem of the kernel has been replaced by a new
1324 implementation, which provides better scalability and an
1325 improved driver model. Most common drivers have been migrated to
1326 the new TTY subsystem, while others have not. The following
1327 drivers have not yet been ported to the new TTY layer:
1330 cy, digi, rc, rp, sio
1336 ng_h4, ng_tty, ppp, sl, snp
1338 Adding these drivers to your kernel configuration file shall
1339 cause compilation to fail.
1342 ntpd has been upgraded to 4.2.4p5.
1345 OpenSSH has been upgraded to 5.1p1.
1347 For many years, FreeBSD's version of OpenSSH preferred DSA
1348 over RSA for host and user authentication keys. With this
1349 upgrade, we've switched to the vendor's default of RSA over
1350 DSA. This may cause upgraded clients to warn about unknown
1351 host keys even for previously known hosts. Users should
1352 follow the usual procedure for verifying host keys before
1353 accepting the RSA key.
1355 This can be circumvented by setting the "HostKeyAlgorithms"
1356 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1359 Please note that the sequence of keys offered for
1360 authentication has been changed as well. You may want to
1361 specify IdentityFile in a different order to revert this
1365 The sio(4) driver has been removed from the i386 and amd64
1366 kernel configuration files. This means uart(4) is now the
1367 default serial port driver on those platforms as well.
1369 To prevent collisions with the sio(4) driver, the uart(4) driver
1370 uses different names for its device nodes. This means the
1371 onboard serial port will now most likely be called "ttyu0"
1372 instead of "ttyd0". You may need to reconfigure applications to
1373 use the new device names.
1375 When using the serial port as a boot console, be sure to update
1376 /boot/device.hints and /etc/ttys before booting the new kernel.
1377 If you forget to do so, you can still manually specify the hints
1378 at the loader prompt:
1380 set hint.uart.0.at="isa"
1381 set hint.uart.0.port="0x3F8"
1382 set hint.uart.0.flags="0x10"
1383 set hint.uart.0.irq="4"
1387 The gpt(8) utility has been removed. Use gpart(8) to partition
1391 The version that Linuxulator emulates was changed from 2.4.2
1392 to 2.6.16. If you experience any problems with Linux binaries
1393 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1394 if it fixes the problem contact emulation mailing list.
1397 ISDN4BSD (I4B) was removed from the src tree. You may need to
1398 update a your kernel configuration and remove relevant entries.
1401 I have checked in code to support multiple routing tables.
1402 See the man pages setfib(1) and setfib(2).
1403 This is a hopefully backwards compatible version,
1404 but to make use of it you need to compile your kernel
1405 with options ROUTETABLES=2 (or more up to 16).
1408 The 802.11 wireless support was redone to enable multi-bss
1409 operation on devices that are capable. The underlying device
1410 is no longer used directly but instead wlanX devices are
1411 cloned with ifconfig. This requires changes to rc.conf files.
1412 For example, change:
1413 ifconfig_ath0="WPA DHCP"
1416 ifconfig_wlan0="WPA DHCP"
1417 see rc.conf(5) for more details. In addition, mergemaster of
1418 /etc/rc.d is highly recommended. Simultaneous update of userland
1419 and kernel wouldn't hurt either.
1421 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1422 modules were merged into the base wlan module. All references
1423 to these modules (e.g. in kernel config files) must be removed.
1426 psm(4) has gained write(2) support in native operation level.
1427 Arbitrary commands can be written to /dev/psm%d and status can
1428 be read back from it. Therefore, an application is responsible
1429 for status validation and error recovery. It is a no-op in
1430 other operation levels.
1433 Support for KSE threading has been removed from the kernel. To
1434 run legacy applications linked against KSE libmap.conf may
1435 be used. The following libmap.conf may be used to ensure
1436 compatibility with any prior release:
1438 libpthread.so.1 libthr.so.1
1439 libpthread.so.2 libthr.so.2
1440 libkse.so.3 libthr.so.3
1443 The layout of struct vmspace has changed. This affects libkvm
1444 and any executables that link against libkvm and use the
1445 kvm_getprocs() function. In particular, but not exclusively,
1446 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1447 The effects are minimal, but it's advisable to upgrade world
1451 The latest em driver no longer has support in it for the
1452 82575 adapter, this is now moved to the igb driver. The
1453 split was done to make new features that are incompatible
1454 with older hardware easier to do.
1457 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1458 likewise the kernel option is now GEOM_LINUX_LVM.
1461 The default NFS mount mode has changed from UDP to TCP for
1462 increased reliability. If you rely on (insecurely) NFS
1463 mounting across a firewall you may need to update your
1467 Belatedly note the addition of m_collapse for compacting
1471 The fts(3) structures have been changed to use adequate
1472 integer types for their members and so to be able to cope
1473 with huge file trees. The old fts(3) ABI is preserved
1474 through symbol versioning in libc, so third-party binaries
1475 using fts(3) should still work, although they will not take
1476 advantage of the extended types. At the same time, some
1477 third-party software might fail to build after this change
1478 due to unportable assumptions made in its source code about
1479 fts(3) structure members. Such software should be fixed
1480 by its vendor or, in the worst case, in the ports tree.
1481 FreeBSD_version 800015 marks this change for the unlikely
1482 case that a portable fix is impossible.
1485 To upgrade to -current after this date, you must be running
1486 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1487 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1490 The ADAPTIVE_GIANT kernel option has been retired because its
1491 functionality is the default now.
1494 The AT keyboard emulation of sunkbd(4) has been turned on
1495 by default. In order to make the special symbols of the Sun
1496 keyboards driven by sunkbd(4) work under X these now have
1497 to be configured the same way as Sun USB keyboards driven
1498 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1500 Option "XkbLayout" "us"
1501 Option "XkbRules" "xorg"
1502 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1505 It has been decided that it is desirable to provide ABI
1506 backwards compatibility to the FreeBSD 4/5/6 versions of the
1507 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1508 broken with the introduction of PCI domain support (see the
1509 20070930 entry). Unfortunately, this required the ABI of
1510 PCIOCGETCONF to be broken again in order to be able to
1511 provide backwards compatibility to the old version of that
1512 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1513 again. As for prominent ports this affects neither pciutils
1514 nor xorg-server this time, the hal port needs to be rebuilt
1518 The misnamed kthread_create() and friends have been renamed
1519 to kproc_create() etc. Many of the callers already
1520 used kproc_start()..
1521 I will return kthread_create() and friends in a while
1522 with implementations that actually create threads, not procs.
1523 Renaming corresponds with version 800002.
1532 Avoid using make -j when upgrading. While generally safe, there are
1533 sometimes problems using -j to upgrade. If your upgrade fails with
1534 -j, please try again without -j. From time to time in the past there
1535 have been problems using -j with buildworld and/or installworld. This
1536 is especially true when upgrading between "distant" versions (eg one
1537 that cross a major release boundary or several minor releases, or when
1538 several months have passed on the -current branch).
1540 Sometimes, obscure build problems are the result of environment
1541 poisoning. This can happen because the make utility reads its
1542 environment when searching for values for global variables. To run
1543 your build attempts in an "environmental clean room", prefix all make
1544 commands with 'env -i '. See the env(1) manual page for more details.
1546 When upgrading from one major version to another it is generally best
1547 to upgrade to the latest code in the currently installed branch first,
1548 then do an upgrade to the new branch. This is the best-tested upgrade
1549 path, and has the highest probability of being successful. Please try
1550 this approach before reporting problems with a major version upgrade.
1554 When upgrading the boot ZFS pool to a new version, always follow
1557 1.) recompile and reinstall the ZFS boot loader and boot block
1558 (this is part of "make buildworld" and "make installworld")
1560 2.) update the ZFS boot block on your boot drive
1562 The following example updates the ZFS boot block on the first
1563 partition (freebsd-boot) of a GPT partitioned drive ad0:
1564 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1566 Non-boot pools do not need these updates.
1570 If you are updating from a prior version of FreeBSD (even one just
1571 a few days old), you should follow this procedure. It is the most
1572 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1574 make kernel-toolchain
1575 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1576 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1578 To test a kernel once
1579 ---------------------
1580 If you just want to boot a kernel once (because you are not sure
1581 if it works, or if you want to boot a known bad kernel to provide
1582 debugging information) run
1583 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1584 nextboot -k testkernel
1586 To just build a kernel when you know that it won't mess you up
1587 --------------------------------------------------------------
1588 This assumes you are already running a CURRENT system. Replace
1589 ${arch} with the architecture of your machine (e.g. "i386",
1590 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1592 cd src/sys/${arch}/conf
1593 config KERNEL_NAME_HERE
1594 cd ../compile/KERNEL_NAME_HERE
1599 If this fails, go to the "To build a kernel" section.
1601 To rebuild everything and install it on the current system.
1602 -----------------------------------------------------------
1603 # Note: sometimes if you are running current you gotta do more than
1604 # is listed here if you are upgrading from a really old current.
1606 <make sure you have good level 0 dumps>
1608 make kernel KERNCONF=YOUR_KERNEL_HERE
1610 <reboot in single user> [3]
1618 To cross-install current onto a separate partition
1619 --------------------------------------------------
1620 # In this approach we use a separate partition to hold
1621 # current's root, 'usr', and 'var' directories. A partition
1622 # holding "/", "/usr" and "/var" should be about 2GB in
1625 <make sure you have good level 0 dumps>
1628 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1629 <maybe newfs current's root partition>
1630 <mount current's root partition on directory ${CURRENT_ROOT}>
1631 make installworld DESTDIR=${CURRENT_ROOT}
1632 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1633 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1634 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1635 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1636 <reboot into current>
1637 <do a "native" rebuild/install as described in the previous section>
1638 <maybe install compatibility libraries from ports/misc/compat*>
1642 To upgrade in-place from stable to current
1643 ----------------------------------------------
1644 <make sure you have good level 0 dumps>
1646 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1648 <reboot in single user> [3]
1655 Make sure that you've read the UPDATING file to understand the
1656 tweaks to various things you need. At this point in the life
1657 cycle of current, things change often and you are on your own
1658 to cope. The defaults can also change, so please read ALL of
1659 the UPDATING entries.
1661 Also, if you are tracking -current, you must be subscribed to
1662 freebsd-current@freebsd.org. Make sure that before you update
1663 your sources that you have read and understood all the recent
1664 messages there. If in doubt, please track -stable which has
1665 much fewer pitfalls.
1667 [1] If you have third party modules, such as vmware, you
1668 should disable them at this point so they don't crash your
1671 [3] From the bootblocks, boot -s, and then do
1676 adjkerntz -i # if CMOS is wall time
1677 Also, when doing a major release upgrade, it is required that
1678 you boot into single user mode to do the installworld.
1680 [4] Note: This step is non-optional. Failure to do this step
1681 can result in a significant reduction in the functionality of the
1682 system. Attempting to do it by hand is not recommended and those
1683 that pursue this avenue should read this file carefully, as well
1684 as the archives of freebsd-current and freebsd-hackers mailing lists
1685 for potential gotchas. The -U option is also useful to consider.
1686 See mergemaster(8) for more information.
1688 [5] Usually this step is a noop. However, from time to time
1689 you may need to do this if you get unknown user in the following
1690 step. It never hurts to do it all the time. You may need to
1691 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1692 install) after the buildworld before this step if you last updated
1693 from [78]-stable or 9-stable before 20130430.
1695 [6] This only deletes old files and directories. Old libraries
1696 can be deleted by "make delete-old-libs", but you have to make
1697 sure that no program is using those libraries anymore.
1699 [8] In order to have a kernel that can run the 4.x binaries needed to
1700 do an installworld, you must include the COMPAT_FREEBSD4 option in
1701 your kernel. Failure to do so may leave you with a system that is
1702 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1703 required to run the 5.x binaries on more recent kernels. And so on
1704 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1706 Make sure that you merge any new devices from GENERIC since the
1707 last time you updated your kernel config file.
1709 [9] When checking out sources, you must include the -P flag to have
1710 cvs prune empty directories.
1712 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1713 "?=" instead of the "=" assignment operator, so that buildworld can
1714 override the CPUTYPE if it needs to.
1716 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1717 not on the command line, or in /etc/make.conf. buildworld will
1718 warn if it is improperly defined.
1721 This file contains a list, in reverse chronological order, of major
1722 breakages in tracking -current. It is not guaranteed to be a complete
1723 list of such breakages, and only contains entries since October 10, 2007.
1724 If you need to see UPDATING entries from before that date, you will need
1725 to fetch an UPDATING file from an older FreeBSD release.
1727 Copyright information:
1729 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1731 Redistribution, publication, translation and use, with or without
1732 modification, in full or in part, in any form or format of this
1733 document are permitted without further permission from the author.
1735 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1736 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1737 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1738 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1739 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1740 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1741 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1742 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1743 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1744 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1745 POSSIBILITY OF SUCH DAMAGE.
1747 Contact Warner Losh if you have any questions about your use of