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 20150225: p10 FreeBSD-SA-15:04.igmp
17 FreeBSD-EN-15:02.openssl
18 FreeBSD-EN-15:03.freebsd-update
20 Fix integer overflow in IGMP protocol. [SA-15:04]
22 Fix BIND remote denial of service vulnerability. [SA-15:05]
24 Fix vt(4) crash with improper ioctl parameters. [EN-15:01]
26 Updated base system OpenSSL to 0.9.8zd. [EN-15:02]
28 Fix freebsd-update libraries update ordering issue. [EN-15:03]
30 20150127: p9 FreeBSD-SA-15:02.kmem
33 Fix SCTP SCTP_SS_VALUE kernel memory corruption and disclosure
34 vulnerability. [SA-15:02]
36 Fix SCTP stream reset vulnerability. [SA-15:03]
38 20150114: p8 FreeBSD-SA-15:01.openssl
39 Fix multiple vulnerabilities in OpenSSL. [SA-15:01]
41 20141223: p7 FreeBSD-SA-14:31.ntp
42 FreeBSD-EN-14:13.freebsd-update
44 Fix multiple vulnerabilities in NTP suite. [SA-14:31]
45 Fix directory deletion issue in freebsd-update. [EN-14:13]
47 20141210: p6 FreeBSD-SA-14:28.file
50 Fix multiple vulnerabilities in file(1) and libmagic(3).
53 Fix BIND remote denial of service vulnerability. [SA-14:29]
55 20141104: p5 FreeBSD-SA-14:25.setlogin
59 Fix kernel stack disclosure in setlogin(2) / getlogin(2).
62 Fix remote command execution in ftp(1). [SA-14:26]
64 Fix NFSv4 and ZFS cache consistency issue. [EN-14:12]
66 20141022: p4 FreeBSD-EN-14:10.tzdata
67 FreeBSD-EN-14:11.crypt
69 Time zone data file update. [EN-14:10]
71 Change crypt(3) default hashing algorithm back to DES. [EN-14:11]
73 20141021: p3 FreeBSD-SA-14:20.rtsold
74 FreeBSD-SA-14:21.routed
75 FreeBSD-SA-14:22.namei
76 FreeBSD-SA-14:23.openssl
78 Fix rtsold(8) remote buffer overflow vulnerability. [SA-14:20]
80 Fix routed(8) remote denial of service vulnerability. [SA-14:21]
82 Fix memory leak in sandboxed namei lookup. [SA-14:22]
84 Fix OpenSSL multiple vulnerabilities. [SA-14:23]
86 20140916: p2 FreeBSD-SA-14:19.tcp
87 Fix Denial of Service in TCP packet processing. [SA-14:19]
89 20140909: p1 FreeBSD-SA-14:18.openssl
90 Fix OpenSSL multiple vulnerabilities. [SA-14:18]
96 On i386 and amd64 systems, the onifconsole flag is now set by default
97 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
98 as a login TTY if it is set in the bootloader as an active kernel
99 console. No changes in behavior should result otherwise. To revert to
100 the previous behavior, set ttyu0 to "off" in /etc/ttys.
103 Clang and llvm have been upgraded to 3.4.1 release.
106 Clang and llvm have been upgraded to 3.4 release.
109 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
110 been deprecated and will not be part of FreeBSD 11.0 and later
111 releases. If you use this driver, please consider switching to
112 the nfe(4) driver instead.
115 The behavior of gss_pseudo_random() for the krb5 mechanism
116 has changed, for applications requesting a longer random string
117 than produced by the underlying enctype's pseudo-random() function.
118 In particular, the random string produced from a session key of
119 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
120 be different at the 17th octet and later, after this change.
121 The counter used in the PRF+ construction is now encoded as a
122 big-endian integer in accordance with RFC 4402.
123 __FreeBSD_version is bumped to 902505.
129 Behavior of devfs rules path matching has been changed.
130 Pattern is now always matched against fully qualified devfs
131 path and slash characters must be explicitly matched by
132 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
133 subdirectories must be reviewed.
136 hastctl(8)'s `status' command output changed to terse one-liner format.
137 Scripts using this should switch to `list' command or be rewritten.
140 Fix a bug that allowed a tracing process (e.g. gdb) to write
141 to a memory-mapped file in the traced process's address space
142 even if neither the traced process nor the tracing process had
143 write access to that file.
146 Added ZFS TRIM support which is enabled by default. To disable
147 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
149 Creating new ZFS pools and adding new devices to existing pools
150 first performs a full device level TRIM which can take a significant
151 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
152 to disable this behaviour.
154 ZFS TRIM requires the underlying device support BIO_DELETE which
155 is currently provided by methods such as ATA TRIM and SCSI UNMAP
156 via CAM, which are typically supported by SSD's.
158 Stats for ZFS TRIM can be monitored by looking at the sysctl's
159 under kstat.zfs.misc.zio_trim.
162 `list' command has been added to hastctl(8). For now, it is full
163 equivalent of `status' command.
164 WARNING: in the near future the output of hastctl's status command
165 will change to more terse format. If you use `hastctl status'
166 for parsing in your scripts, switch to `hastctl list'.
169 The mergemaster command now uses the default MAKEOBJDIRPREFIX
170 rather than creating it's own in the temporary directory in
171 order allow access to bootstrapped versions of tools such as
172 install and mtree. When upgrading from version of FreeBSD where
173 the install command does not support -l, you will need to
174 install a new mergemaster command if mergemaster -p is required.
175 This can be accomplished with the command (cd src/usr.sbin/mergemaster
178 Due to the use of the new -l option to install(1) during build
179 and install, you must take care not to directly set the INSTALL
180 make variable in your /etc/make.conf, /etc/src.conf, or on the
181 command line. If you wish to use the -C flag for all installs
182 you may be able to add INSTALL+=-C to /etc/make.conf or
186 Fix a bug that allows NFS clients to issue READDIR on files.
189 The install(1) option -M has changed meaning and now takes an
190 argument that is a file or path to append logs to. In the
191 unlikely event that -M was the last option on the command line
192 and the command line contained at least two files and a target
193 directory the first file will have logs appended to it. The -M
194 option served little practical purpose in the last decade so it's
195 used expected to be extremely rare.
198 A new compression method (lz4) has been merged to. Please refer to
199 zpool-features(7) for more information.
201 Please refer to the "ZFS notes" section of this file for information
202 on upgrading boot ZFS pools.
205 The VFS KBI was changed with the merge of several nullfs
206 optimizations and fixes. All filesystem modules must be
210 With the addition of auditdistd(8), a new auditdistd user is now
211 depended on during installworld. "mergemaster -p" can be used to add
212 the user prior to installworld, as documented in the handbook.
218 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
219 Starting with this version the old system of ZFS pool versioning
220 is superseded by "feature flags". This concept enables forward
221 compatibility against certain future changes in functionality of ZFS
222 pools. The first two read-only compatible "feature flags" for ZFS
223 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
224 For more information read the new zpool-features(7) manual page.
225 Please refer to the "ZFS notes" section of this file for information
226 on upgrading boot ZFS pools.
229 The commit introducing bsd.compiler.mk breaks the traditional
230 building of kernels before this point. Add -m ${SRC}/share/mk
231 (for the right value of SRC) to your command lines to work
232 around; update your useland to a point after this; or use the
233 buildkernel/installkernel top-level targets. See also 20120829.
236 The IPFIREWALL_FORWARD kernel option has been removed. Its
237 functionality now turned on by default.
240 The random(4) support for the VIA hardware random number
241 generator (`PADLOCK') is no longer enabled unconditionally.
242 Add the PADLOCK_RNG option in the custom kernel config if
243 needed. The GENERIC kernels on i386 and amd64 do include the
244 option, so the change only affects the custom kernel
248 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
249 the traditional method, you must update your system with an installworld
250 before the kernel will build. The documented make buildkernel/installkernel
251 interfaces (coupled with fresh make kernel-toolchain) continue to work.
254 The sparc64 ZFS loader has been changed to no longer try to auto-
255 detect ZFS providers based on diskN aliases but now requires these
256 to be explicitly listed in the OFW boot-device environment variable.
259 Now unix domain sockets behave "as expected" on nullfs(5). Previously
260 nullfs(5) did not pass through all behaviours to the underlying layer,
261 as a result if we bound to a socket on the lower layer we could connect
262 only to the lower path; if we bound to the upper layer we could connect
263 only to the upper path. The new behavior is one can connect to both the
264 lower and the upper paths regardless what layer path one binds to.
267 The acpi_wmi(4) status device /dev/wmistat has been renamed to
271 A new VOP_ADVISE() was added to support posix_fadvise(2). All
272 filesystem modules must be recompiled.
275 The interface of the VOP_VPTOCNP(9) changed, now the returned
276 vnode shall be referenced, previously it was required to be
277 only held. All in-tree filesystems are converted.
283 The broken amd(4) driver has been replaced with esp(4) in the amd64,
284 i386 and pc98 GENERIC kernel configuration files.
287 This commit modifies vfs_register() so that it uses a hash
288 calculation to set vfc_typenum, which is enabled by default.
289 The first time a system is booted after this change, the
290 vfc_typenum values will change for all file systems. The
291 main effect of this is a change to the NFS server file handles
292 for file systems that use vfc_typenum in their fsid, such as ZFS.
293 It will, however, prevent vfc_typenum from changing when file
294 systems are loaded in a different order for subsequent reboots.
295 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
296 until you are ready to remount all NFS clients after a reboot.
299 Bump the shared library version numbers for libraries that
300 do not use symbol versioning, have changed the ABI compared
301 to stable/8 and which shared library version was not bumped.
302 Done as part of 9.0-RELEASE cycle.
305 During the merge of Capsicum features, the fget(9) KPI was modified.
306 This may require the rebuilding of out-of-tree device drivers --
307 issues have been reported specifically with the nVidia device driver.
308 __FreeBSD_version is bumped to 900041.
310 Also, there is a period between 20110811 and 20110814 where the
311 special devices /dev/{stdin,stdout,stderr} did not work correctly.
312 Building world from a kernel during that window may not work.
315 The packet filter (pf) code has been updated to OpenBSD 4.5.
316 You need to update userland tools to be in sync with kernel.
317 This update breaks backward compatibility with earlier pfsync(4)
318 versions. Care must be taken when updating redundant firewall setups.
321 The following sysctls and tunables are retired on x86 platforms:
323 machdep.hlt_logical_cpus
324 The following sysctl is retired:
325 machdep.hyperthreading_allowed
326 The sysctls were supposed to provide a way to dynamically offline and
327 online selected CPUs on x86 platforms, but the implementation has not
328 been reliable especially with SCHED_ULE scheduler.
329 machdep.hyperthreading_allowed tunable is still available to ignore
330 hyperthreading CPUs at OS level.
331 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
332 where X is an APIC ID of a CPU. Be advised, though, that disabling
333 CPUs in non-uniform fashion will result in non-uniform topology and
334 may lead to sub-optimal system performance with SCHED_ULE, which is
338 cpumask_t type is retired and cpuset_t is used in order to describe
342 Changes to ifconfig(8) for dynamic address family detection mandate
343 that you are running a kernel of 20110525 or later. Make sure to
344 follow the update procedure to boot a new kernel before installing
348 Support for sun4v architecture is officially dropped
351 Several KPI breaking changes have been committed to the mii(4) layer,
352 the PHY drivers and consequently some Ethernet drivers using mii(4).
353 This means that miibus.ko and the modules of the affected Ethernet
354 drivers need to be recompiled.
356 Note to kernel developers: Given that the OUI bit reversion problem
357 was fixed as part of these changes all mii(4) commits related to OUIs,
358 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
359 handling, no longer can be merged verbatim to stable/8 and previous
363 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
364 into their kernel configurations along with 'device pci'.
367 The default NFS client is now the new NFS client, so fstype "newnfs"
368 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
369 Although mounts via fstype "nfs" will usually work without userland
370 changes, it is recommended that the mount(8) and mount_nfs(8)
371 commands be rebuilt from sources and that a link to mount_nfs called
372 mount_oldnfs be created. The new client is compiled into the
373 kernel with "options NFSCL" and this is needed for diskless root
374 file systems. The GENERIC kernel configs have been changed to use
375 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
376 To use the regular/old client, you can "mount -t oldnfs ...". For
377 a diskless root file system, you must also include a line like:
379 vfs.root.mountfrom="oldnfs:"
381 in the boot/loader.conf on the root fs on the NFS server to make
382 a diskless root fs use the old client.
385 The GENERIC kernels for all architectures now default to the new
386 CAM-based ATA stack. It means that all legacy ATA drivers were
387 removed and replaced by respective CAM drivers. If you are using
388 ATA device names in /etc/fstab or other places, make sure to update
389 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
390 where 'Y's are the sequential numbers starting from zero for each type
391 in order of detection, unless configured otherwise with tunables,
392 see cam(4)). There will be symbolic links created in /dev/ to map
393 old adX devices to the respective adaY. They should provide basic
394 compatibility for file systems mounting in most cases, but they do
395 not support old user-level APIs and do not have respective providers
396 in GEOM. Consider using updated management tools with new device names.
398 It is possible to load devices ahci, ata, siis and mvs as modules,
399 but option ATA_CAM should remain in kernel configuration to make ata
400 module work as CAM driver supporting legacy ATA controllers. Device ata
401 still can be used in modular fashion (atacore + ...). Modules atadisk
402 and atapi* are not used and won't affect operation in ATA_CAM mode.
403 Note that to use CAM-based ATA kernel should include CAM devices
404 scbus, pass, da (or explicitly ada), cd and optionally others. All of
405 them are parts of the cam module.
407 ataraid(4) functionality is now supported by the RAID GEOM class.
408 To use it you can load geom_raid kernel module and use graid(8) tool
409 for management. Instead of /dev/arX device names, use /dev/raid/rX.
411 No kernel config options or code have been removed, so if a problem
412 arises, please report it and optionally revert to the old ATA stack.
413 In order to do it you can remove from the kernel config:
418 , and instead add back:
419 device atadisk # ATA disk drives
420 device ataraid # ATA RAID drives
421 device atapicd # ATAPI CDROM drives
422 device atapifd # ATAPI floppy drives
423 device atapist # ATAPI tape drives
426 The default NFS server has been changed to the new server, which
427 was referred to as the experimental server. If you need to switch
428 back to the old NFS server, you must now put the "-o" option on
429 both the mountd and nfsd commands. This can be done using the
430 mountd_flags and nfs_server_flags rc.conf variables until an
431 update to the rc scripts is committed, which is coming soon.
434 The GNU Objective-C runtime library (libobjc), and other Objective-C
435 related components have been removed from the base system. If you
436 require an Objective-C library, please use one of the available ports.
439 ath(4) has been split into bus- and device- modules. if_ath contains
440 the HAL, the TX rate control and the network device code. if_ath_pci
441 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
442 contains the AHB glue. Users need to load both if_ath_pci and if_ath
443 in order to use ath on everything else.
445 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
446 need to load if_ath and if_ath_pci for ath(4) operation.
449 As part of the replacement of sysinstall, the process of building
450 release media has changed significantly. For details, please re-read
451 release(7), which has been updated to reflect the new build process.
454 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
455 is the last available version under GPLv2. It brings a number of new
456 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
457 4.1 and SSE 4.2), better support for powerpc64, a number of new
458 directives, and lots of other small improvements. See the ChangeLog
459 file in contrib/binutils for the full details.
462 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
463 compliant, and will now use half of hash for authentication.
464 This will break interoperability with all stacks (including all
465 actual FreeBSD versions) who implement
466 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
468 The only workaround with such peers is to use another HMAC
469 algorithm for IPsec ("phase 2") authentication.
472 Remove the uio_yield prototype and symbol. This function has
473 been misnamed since it was introduced and should not be
474 globally exposed with this name. The equivalent functionality
475 is now available using kern_yield(curthread->td_user_pri).
476 The function remains undocumented.
479 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
480 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
481 for scalar sysctls is defined but disabled. Code that needs
482 UQUAD to pass the type checking that must compile on older
483 systems where the define is not present can check against
484 __FreeBSD_version >= 900030.
486 The system dialog(1) has been replaced with a new version previously
487 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
488 with the previous version, but the libdialog associated with it has
489 a largely incompatible API. As such, the original version of libdialog
490 will be kept temporarily as libodialog, until its base system consumers
491 are replaced or updated. Bump __FreeBSD_version to 900030.
494 If you are trying to run make universe on a -stable system, and you get
495 the following warning:
496 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
497 unknown. config(8) likely too old."
498 or something similar to it, then you must upgrade your -stable system
499 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
500 stable/8) or build the config from the latest stable/8 branch and
501 install it on your system.
503 Prior to this date, building a current universe on 8-stable system from
504 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
505 error in the first kernel build phase. A new config on those old
506 systems will fix that problem for older versions of -current.
509 The TCP stack has been modified to allow Khelp modules to interact with
510 it via helper hook points and store per-connection data in the TCP
511 control block. Bump __FreeBSD_version to 900029. User space tools that
512 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
516 Generic IEEE 802.3 annex 31B full duplex flow control support has been
517 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
518 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
519 to take advantage of it instead of using custom implementations. This
520 means that these drivers now no longer unconditionally advertise
521 support for flow control but only do so if flow control is a selected
522 media option. This was implemented in the generic support that way in
523 order to allow flow control to be switched on and off via ifconfig(8)
524 with the PHY specific default to typically off in order to protect
525 from unwanted effects. Consequently, if you used flow control with
526 one of the above mentioned drivers you now need to explicitly enable
528 ifconfig bge0 media auto mediaopt flowcontrol
530 Along with the above mentioned changes generic support for setting
531 1000baseT master mode also has been added and brgphy(4), ciphy(4),
532 e1000phy(4) as well as ip1000phy(4) have been converted to take
533 advantage of it. This means that these drivers now no longer take the
534 link0 parameter for selecting master mode but the master media option
535 has to be used instead, for example like in the following:
536 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
538 Selection of master mode now is also available with all other PHY
539 drivers supporting 1000baseT.
542 The TCP stack has received a significant update to add support for
543 modularised congestion control and generally improve the clarity of
544 congestion control decisions. Bump __FreeBSD_version to 900025. User
545 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
546 sockstat) need to be recompiled.
549 The man(1) utility has been replaced by a new version that no longer
550 uses /etc/manpath.config. Please consult man.conf(5) for how to
551 migrate local entries to the new format.
554 The copyright strings printed by login(1) and sshd(8) at the time of a
555 new connection have been removed to follow other operating systems and
559 A workaround for a fixed ld bug has been removed in kernel code,
560 so make sure that your system ld is built from sources after
561 revision 210245 from 2010-07-19 (r211583 if building head kernel
562 on stable/8, r211584 for stable/7; both from 2010-08-21).
563 A symptom of incorrect ld version is different addresses for
564 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
567 The $ipv6_prefer variable in rc.conf(5) has been split into
568 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
570 The $ip6addrctl_policy is a variable to choose a pre-defined
571 address selection policy set by ip6addrctl(8). A value
572 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
575 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
576 flag (see an entry of 20090926) is set on an interface with no
577 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
578 security reason. If you want IPv6 link-local address on all
579 interfaces by default, set this to "YES".
581 The old ipv6_prefer="YES" is equivalent to
582 ipv6_activate_all_interfaces="YES" and
583 ip6addrctl_policy="ipv6_prefer".
586 DTrace has grown support for userland tracing. Due to this, DTrace is
587 now i386 and amd64 only.
588 dtruss(1) is now installed by default on those systems and a new
589 kernel module is needed for userland tracing: fasttrap.
590 No changes to your kernel config file are necessary to enable
591 userland tracing, but you might consider adding 'STRIP=' and
592 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
593 to have informative userland stack traces in DTrace (ustack).
596 The acpi_aiboost(4) driver has been removed in favor of the new
597 aibs(4) driver. You should update your kernel configuration file.
600 BSD grep has been imported to the base system and it is built by
601 default. It is completely BSD licensed, highly GNU-compatible, uses
602 less memory than its GNU counterpart and has a small codebase.
603 However, it is slower than its GNU counterpart, which is mostly
604 noticeable for larger searches, for smaller ones it is measurable
605 but not significant. The reason is complex, the most important factor
606 is that we lack a modern and efficient regex library and GNU
607 overcomes this by optimizing the searches internally. Future work
608 on improving the regex performance is planned, for the meantime,
609 users that need better performance, can build GNU grep instead by
610 setting the WITH_GNU_GREP knob.
613 Due to the import of powerpc64 support, all existing powerpc kernel
614 configuration files must be updated with a machine directive like this:
615 machine powerpc powerpc
617 In addition, an updated config(8) is required to build powerpc kernels
621 A new version of ZFS (version 15) has been merged to -HEAD.
622 This version uses a python library for the following subcommands:
623 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
624 For full functionality of these commands the following port must
625 be installed: sysutils/py-zfs
628 'vm_page's are now hashed by physical address to an array of mutexes.
629 Currently this is only used to serialize access to hold_count. Over
630 time the page queue mutex will be peeled away. This changes the size
631 of pmap on every architecture. And requires all callers of vm_page_hold
632 and vm_page_unhold to be updated.
635 WITH_CTF can now be specified in src.conf (not recommended, there
636 are some problems with static executables), make.conf (would also
637 affect ports which do not use GNU make and do not override the
638 compile targets) or in the kernel config (via "makeoptions
640 When WITH_CTF was specified there before this was silently ignored,
641 so make sure that WITH_CTF is not used in places which could lead
642 to unwanted behavior.
645 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
646 to allow 32-bit compatibility on non-x86 platforms. All kernel
647 configurations on amd64 and ia64 platforms using these options must
648 be modified accordingly.
651 The utmp user accounting database has been replaced with utmpx,
652 the user accounting interface standardized by POSIX.
653 Unfortunately the semantics of utmp and utmpx don't match,
654 making it practically impossible to support both interfaces.
655 The user accounting database is used by tools like finger(1),
656 last(1), talk(1), w(1) and ac(8).
658 All applications in the base system use utmpx. This means only
659 local binaries (e.g. from the ports tree) may still use these
660 utmp database files. These applications must be rebuilt to make
663 After the system has been upgraded, it is safe to remove the old
664 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
665 assuming their contents is of no importance anymore. Old wtmp
666 databases can only be used by last(1) and ac(8) after they have
667 been converted to the new format using wtmpcvt(1).
670 Introduce the kernel thread "deadlock resolver" (which can be enabled
671 via the DEADLKRES option, see NOTES for more details) and the
672 sleepq_type() function for sleepqueues.
675 The rc.firewall and rc.firewall6 were unified, and
676 rc.firewall6 and rc.d/ip6fw were removed.
677 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
678 variables are obsoleted. Instead, the following new rc
679 variables are added to rc.d/ipfw:
681 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
682 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
683 firewall_simple_onet_ipv6, firewall_trusted_ipv6
685 The meanings correspond to the relevant IPv4 variables.
691 The default terminal emulation for syscons(4) has been changed
692 from cons25 to xterm on all platforms except pc98. This means
693 that the /etc/ttys file needs to be updated to ensure correct
694 operation of applications on the console.
696 The terminal emulation style can be toggled per window by using
697 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
698 options can be used to change the compile-time default back to
701 To prevent graphical artifacts, make sure the TERM environment
702 variable is set to match the terminal emulation that is being
703 performed by syscons(4).
706 The layout of the structure ieee80211req_scan_result has changed.
707 Applications that require wireless scan results (e.g. ifconfig(8))
708 from net80211 need to be recompiled.
710 Applications such as wpa_supplicant(8) may require a full world
711 build without using NO_CLEAN in order to get synchronized with the
715 The iwn(4) driver has been updated to support the 5000 and 5150 series.
716 There's one kernel module for each firmware. Adding "device iwnfw"
717 to the kernel configuration file means including all three firmware
718 images inside the kernel. If you want to include just the one for
719 your wireless card, use the devices iwn4965fw, iwn5000fw or
723 The rc.d/network_ipv6, IPv6 configuration script has been integrated
724 into rc.d/netif. The changes are the following:
726 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
727 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
728 Note that both variables need the "inet6" keyword at the head.
730 Do not set $ipv6_network_interfaces manually if you do not
731 understand what you are doing. It is not needed in most cases.
733 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
736 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
737 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
739 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
740 all configured interfaces have "inet6 accept_rtadv" in the
741 $ifconfig_IF_ipv6. These are for backward compatibility.
743 3. A new variable $ipv6_prefer has been added. If NO, IPv6
744 functionality of interfaces with no corresponding
745 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
746 and the default address selection policy of ip6addrctl(8)
747 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
748 Note that if you want to configure IPv6 functionality on the
749 disabled interfaces after boot, first you need to clear the flag by
750 using ifconfig(8) like:
752 ifconfig em0 inet6 -ifdisabled
754 If YES, the default address selection policy is set as
757 The default value of $ipv6_prefer is NO.
759 4. If your system need to receive Router Advertisement messages,
760 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
761 scripts automatically invoke rtsol(8) when the interface becomes
762 UP. The Router Advertisement messages are used for SLAAC
763 (State-Less Address AutoConfiguration).
766 802.11s D3.03 support was committed. This is incompatible with the
767 previous code, which was based on D3.0.
770 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
771 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
772 control whether accepting Router Advertisement messages or not.
773 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
774 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
775 The ifconfig(8) utility now supports these flags.
778 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
779 mount(8) and -a option for df(1) to see them.
782 The old tunable hw.bus.devctl_disable has been superseded by
783 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
784 replaced by hw.bus.devctl_queue=0. The default for this new tunable
788 Remove the option STOP_NMI. The default action is now to use NMI only
789 for KDB via the newly introduced function stop_cpus_hard() and
790 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
793 The stable/8 branch created in subversion. This corresponds to the
794 RELENG_8 branch in CVS.
797 Bump the shared library version numbers for all libraries that do not
798 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
799 __FreeBSD_version to 800105.
802 Due to changes in the implementation of virtual network stack support,
803 all network-related kernel modules must be recompiled. As this change
804 breaks the ABI, bump __FreeBSD_version to 800104.
807 The TOE interface to the TCP syncache has been modified to remove
808 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
809 The cxgb driver is the only TOE consumer affected by this change, and
810 needs to be recompiled along with the kernel. As this change breaks
811 the ABI, bump __FreeBSD_version to 800103.
814 Padding has been added to struct tcpcb, sackhint and tcpstat in
815 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
816 maintaining the ABI. However, this change breaks the ABI, so bump
817 __FreeBSD_version to 800102. User space tools that rely on the size of
818 any of these structs (e.g. sockstat) need to be recompiled.
821 The NFS_LEGACYRPC option has been removed along with the old kernel
822 RPC implementation that this option selected. Kernel configurations
823 may need to be adjusted.
826 The network interface device nodes at /dev/net/<interface> have been
827 removed. All ioctl operations can be performed the normal way using
828 routing sockets. The kqueue functionality can generally be replaced
829 with routing sockets.
832 The documentation from the FreeBSD Documentation Project (Handbook,
833 FAQ, etc.) is now installed via packages by sysinstall(8) and under
834 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
837 The ABI of various structures related to the SYSV IPC API have been
838 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
839 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
843 Layout of struct vnet has changed as routing related variables were
844 moved to their own Vimage module. Modules need to be recompiled. Bump
845 __FreeBSD_version to 800099.
848 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
849 respectively. As long as no more than 16 groups per process are used,
850 no changes should be visible. When more than 16 groups are used, old
851 binaries may fail if they call getgroups() or getgrouplist() with
852 statically sized storage. Recompiling will work around this, but
853 applications should be modified to use dynamically allocated storage
854 for group arrays as POSIX.1-2008 does not cap an implementation's
855 number of supported groups at NGROUPS_MAX+1 as previous versions did.
857 NFS and portalfs mounts may also be affected as the list of groups is
858 truncated to 16. Users of NFS who use more than 16 groups, should
859 take care that negative group permissions are not used on the exported
860 file systems as they will not be reliable unless a GSSAPI based
861 authentication method is used.
864 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
865 option compiles in the support for adaptive spinning for lockmgrs
866 which want to enable it. The lockinit() function now accepts the flag
867 LK_ADAPTIVE in order to make the lock object subject to adaptive
868 spinning when both held in write and read mode.
871 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
872 changed. User applications that use this ioctl need to be rebuilt.
875 The layout of struct thread has changed. Kernel and modules need to
879 The layout of structs ifnet, domain, protosw and vnet_net has changed.
880 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
883 window(1) has been removed from the base system. It can now be
884 installed from ports. The port is called misc/window.
887 The way we are storing and accessing `routing table' entries has
888 changed. Programs reading the FIB, like netstat, need to be
892 A new netisr implementation has been added for FreeBSD 8. Network
893 file system modules, such as igmp, ipdivert, and others, should be
895 Bump __FreeBSD_version to 800096.
898 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
902 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
903 Bump __FreeBSD_version to 800094.
906 Add mnt_xflag field to 'struct mount'. File system modules need to be
908 Bump __FreeBSD_version to 800093.
911 The compiling option ADAPTIVE_SX has been retired while it has been
912 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
913 The KPI for sx_init_flags() changes as accepting flags:
914 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
915 been introduced in order to handle the reversed logic.
916 Bump __FreeBSD_version to 800092.
919 Add support for hierarchical jails. Remove global securelevel.
920 Bump __FreeBSD_version to 800091.
923 The layout of struct vnet_net has changed, therefore modules
925 Bump __FreeBSD_version to 800090.
928 The newly imported zic(8) produces a new format in the output. Please
929 run tzsetup(8) to install the newly created data to /etc/localtime.
932 The sysctl tree for the usb stack has renamed from hw.usb2.* to
933 hw.usb.* and is now consistent again with previous releases.
936 802.11 monitor mode support was revised and driver api's were changed.
937 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
938 of DLT_IEEE802_11. No user-visible data structures were changed but
939 applications that use DLT_IEEE802_11 may require changes.
940 Bump __FreeBSD_version to 800088.
943 The layout of the following structs has changed: sysctl_oid,
944 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
945 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
946 panics may be experienced. World rebuild is required for
947 correctly checking networking state from userland.
948 Bump __FreeBSD_version to 800085.
951 MLDv2 and Source-Specific Multicast (SSM) have been merged
952 to the IPv6 stack. VIMAGE hooks are in but not yet used.
953 The implementation of SSM within FreeBSD's IPv6 stack closely
954 follows the IPv4 implementation.
956 For kernel developers:
958 * The most important changes are that the ip6_output() and
959 ip6_input() paths no longer take the IN6_MULTI_LOCK,
960 and this lock has been downgraded to a non-recursive mutex.
962 * As with the changes to the IPv4 stack to support SSM, filtering
963 of inbound multicast traffic must now be performed by transport
964 protocols within the IPv6 stack. This does not apply to TCP and
965 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
967 * The KPIs used by IPv6 multicast are similar to those used by
968 the IPv4 stack, with the following differences:
969 * im6o_mc_filter() is analogous to imo_multicast_filter().
970 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
971 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
972 * IN6_LOOKUP_MULTI() has been deprecated and removed.
973 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
974 for MLDv1 have an additional 'timer' argument which is used to
975 jitter the initial membership report for the solicited-node
976 multicast membership on-link.
977 * This is not strictly needed for MLDv2, which already jitters
978 its report transmissions. However, the 'timer' argument is
979 preserved in case MLDv1 is active on the interface.
981 * The KAME linked-list based IPv6 membership implementation has
982 been refactored to use a vector similar to that used by the IPv4
984 Code which maintains a list of its own multicast memberships
985 internally, e.g. carp, has been updated to reflect the new
988 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
989 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
990 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
991 implementation constraint which needs to be addressed in HEAD.
993 For application developers:
995 * The changes are broadly similar to those made for the IPv4
998 * The use of IPv4 and IPv6 multicast socket options on the same
999 socket, using mapped addresses, HAS NOT been tested or supported.
1001 * There are a number of issues with the implementation of various
1002 IPv6 multicast APIs which need to be resolved in the API surface
1003 before the implementation is fully compatible with KAME userland
1004 use, and these are mostly to do with interface index treatment.
1006 * The literature available discusses the use of either the delta / ASM
1007 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
1008 using setsourcefilter(3)/getsourcefilter(3). For more information
1009 please refer to RFC 3768, 'Socket Interface Extensions for
1010 Multicast Source Filters'.
1012 * Applications which use the published RFC 3678 APIs should be fine.
1014 For systems administrators:
1016 * The mtest(8) utility has been refactored to support IPv6, in
1017 addition to IPv4. Interface addresses are no longer accepted
1018 as arguments, their names must be used instead. The utility
1019 will map the interface name to its first IPv4 address as
1020 returned by getifaddrs(3).
1022 * The ifmcstat(8) utility has also been updated to print the MLDv2
1023 endpoint state and source filter lists via sysctl(3).
1025 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
1026 loopback of IPv6 multicast datagrams by default; it defaults to 1
1027 to preserve the existing behaviour. Disabling multicast loopback is
1028 recommended for optimal system performance.
1030 * The IPv6 MROUTING code has been changed to examine this sysctl
1031 instead of attempting to perform a group lookup before looping
1032 back forwarded datagrams.
1034 Bump __FreeBSD_version to 800084.
1037 Implement low-level Bluetooth HCI API.
1038 Bump __FreeBSD_version to 800083.
1041 The layout of struct malloc_type, used by modules to register new
1042 memory allocation types, has changed. Most modules will need to
1043 be rebuilt or panics may be experienced.
1044 Bump __FreeBSD_version to 800081.
1047 Anticipate overflowing inp_flags - add inp_flags2.
1048 This changes most offsets in inpcb, so checking v4 connection
1049 state will require a world rebuild.
1050 Bump __FreeBSD_version to 800080.
1053 Add an llentry to struct route and struct route_in6. Modules
1054 embedding a struct route will need to be recompiled.
1055 Bump __FreeBSD_version to 800079.
1058 The size of rt_metrics_lite and by extension rtentry has changed.
1059 Networking administration apps will need to be recompiled.
1060 The route command now supports show as an alias for get, weighting
1061 of routes, sticky and nostick flags to alter the behavior of stateful
1063 Bump __FreeBSD_version to 800078.
1066 Do not use Giant for kbdmux(4) locking. This is wrong and
1067 apparently causing more problems than it solves. This will
1068 re-open the issue where interrupt handlers may race with
1069 kbdmux(4) in polling mode. Typical symptoms include (but
1070 not limited to) duplicated and/or missing characters when
1071 low level console functions (such as gets) are used while
1072 interrupts are enabled (for example geli password prompt,
1073 mountroot prompt etc.). Disabling kbdmux(4) may help.
1076 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1077 kernel modules referencing any of the above need to be recompiled.
1078 Bump __FreeBSD_version to 800075.
1081 GEOM_PART has become the default partition slicer for storage devices,
1082 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1083 introduces some changes:
1085 MSDOS/EBR: the devices created from MSDOS extended partition entries
1086 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1087 to devices with offset-based names. fstabs may need to be modified.
1089 BSD: the "geometry does not match label" warning is harmless in most
1090 cases but it points to problems in file system misalignment with
1091 disk geometry. The "c" partition is now implicit, covers the whole
1092 top-level drive and cannot be (mis)used by users.
1094 General: Kernel dumps are now not allowed to be written to devices
1095 whose partition types indicate they are meant to be used for file
1096 systems (or, in case of MSDOS partitions, as something else than
1099 Most of these changes date approximately from 200812.
1102 The uscanner(4) driver has been removed from the kernel. This follows
1103 Linux removing theirs in 2.6 and making libusb the default interface
1104 (supported by sane).
1107 The multicast forwarding code has been cleaned up. netstat(1)
1108 only relies on KVM now for printing bandwidth upcall meters.
1109 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1110 ip6_mroute_mod respectively. The config(5) options for statically
1111 compiling this code remain the same, i.e. 'options MROUTING'.
1114 Support for the IFF_NEEDSGIANT network interface flag has been
1115 removed, which means that non-MPSAFE network device drivers are no
1116 longer supported. In particular, if_ar, if_sr, and network device
1117 drivers from the old (legacy) USB stack can no longer be built or
1121 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1122 a bunch of new language catalog files have also been added.
1123 This means that some common libc messages are now localized and
1124 they depend on the LC_MESSAGES environmental variable.
1127 The k8temp(4) driver has been renamed to amdtemp(4) since
1128 support for Family 10 and Family 11 CPU families was added.
1131 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1132 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1134 For kernel developers, the most important changes are that the
1135 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1136 and this lock has been downgraded to a non-recursive mutex.
1138 Transport protocols (UDP, Raw IP) are now responsible for filtering
1139 inbound multicast traffic according to group membership and source
1140 filters. The imo_multicast_filter() KPI exists for this purpose.
1141 Transports which do not use multicast (SCTP, TCP) already reject
1142 multicast by default. Forwarding and receive performance may improve
1143 as a mutex acquisition is no longer needed in the ip_input()
1144 low-level input path. in_addmulti() and in_delmulti() are shimmed
1145 to new KPIs which exist to support SSM in-kernel.
1147 For application developers, it is recommended that loopback of
1148 multicast datagrams be disabled for best performance, as this
1149 will still cause the lock to be taken for each looped-back
1150 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1151 be tuned to 0 to disable loopback by default; it defaults to 1
1152 to preserve the existing behaviour.
1154 For systems administrators, to obtain best performance with
1155 multicast reception and multiple groups, it is always recommended
1156 that a card with a suitably precise hash filter is used. Hash
1157 collisions will still result in the lock being taken within the
1158 transport protocol input path to check group membership.
1160 If deploying FreeBSD in an environment with IGMP snooping switches,
1161 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1162 enabled; this forces 224.0.0.0/24 group membership to be announced
1165 The size of 'struct igmpstat' has changed; netstat needs to be
1166 recompiled to reflect this.
1167 Bump __FreeBSD_version to 800070.
1170 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1171 updated to use it. This requires a buildworld/installworld in order to
1172 update the library and dependencies (usbconfig, etc). Its advisable to
1173 rebuild all ports which uses libusb. More specific directions are given
1174 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1175 libusb are no longer required and can be removed.
1178 A workaround is committed to allow the creation of System V shared
1179 memory segment of size > 2 GB on the 64-bit architectures.
1180 Due to a limitation of the existing ABI, the shm_segsz member
1181 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1182 wrong for large segments. Note that limits must be explicitly
1183 raised to allow such segments to be created.
1186 The layout of struct ifnet has changed, requiring a rebuild of all
1187 network device driver modules.
1190 The /dev handling for the new USB stack has changed, a
1191 buildworld/installworld is required for libusb20.
1194 The new USB2 stack has now been permanently moved in and all kernel and
1195 module names reverted to their previous values (eg, usb, ehci, ohci,
1196 ums, ...). The old usb stack can be compiled in by prefixing the name
1197 with the letter 'o', the old usb modules have been removed.
1198 Updating entry 20090216 for xorg and 20090215 for libmap may still
1202 The rc.conf(5) option if_up_delay has been renamed to
1203 defaultroute_delay to better reflect its purpose. If you have
1204 customized this setting in /etc/rc.conf you need to update it to
1208 xorg 7.4 wants to configure its input devices via hald which does not
1209 yet work with USB2. If the keyboard/mouse does not work in xorg then
1211 Option "AllowEmptyInput" "off"
1212 to your ServerLayout section. This will cause X to use the configured
1213 kbd and mouse sections from your xorg.conf.
1216 The GENERIC kernels for all architectures now default to the new USB2
1217 stack. No kernel config options or code have been removed so if a
1218 problem arises please report it and optionally revert to the old USB
1219 stack. If you are loading USB kernel modules or have a custom kernel
1220 that includes GENERIC then ensure that usb names are also changed over,
1221 eg uftdi -> usb2_serial_ftdi.
1223 Older programs linked against the ports libusb 0.1 need to be
1224 redirected to the new stack's libusb20. /etc/libmap.conf can
1226 # Map old usb library to new one for usb2 stack
1227 libusb-0.1.so.8 libusb20.so.1
1230 All USB ethernet devices now attach as interfaces under the name ueN
1231 (eg. ue0). This is to provide a predictable name as vendors often
1232 change usb chipsets in a product without notice.
1235 The ichsmb(4) driver has been changed to require SMBus slave
1236 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1237 All of the other SMBus controller drivers require left-justified
1238 slave addresses, so this change makes all the drivers provide the
1242 INET6 statistics (struct ip6stat) was updated.
1243 netstat(1) needs to be recompiled.
1246 NTFS has been removed from GENERIC kernel on amd64 to match
1247 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1248 will load ntfs.ko module automatically when NTFS support is
1249 actually needed, unless ntfs.ko is not installed or security
1250 level prohibits loading kernel modules. If either is the case,
1251 "options NTFS" has to be added into kernel config.
1254 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1255 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1256 800061. User space tools that rely on the size of struct tcpcb in
1257 tcp_var.h (e.g. sockstat) need to be recompiled.
1260 ng_tty(4) module updated to match the new TTY subsystem.
1261 Due to API change, user-level applications must be updated.
1262 New API support added to mpd5 CVS and expected to be present
1263 in next mpd5.3 release.
1266 With __FreeBSD_version 800060 the makefs tool is part of
1267 the base system (it was a port).
1270 The afdata and ifnet locks have been changed from mutexes to
1271 rwlocks, network modules will need to be re-compiled.
1274 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1275 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1276 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1277 architecture and 40 bytes on 64-bit architecture. The userland
1278 applications "arp" and "ndp" have been updated accordingly.
1279 The output from "netstat -r" shows only routing entries and
1280 none of the L2 information.
1283 __FreeBSD_version 800057 marks the switchover from the
1284 binary ath hal to source code. Users must add the line:
1286 options AH_SUPPORT_AR5416
1288 to their kernel config files when specifying:
1292 The ath_hal module no longer exists; the code is now compiled
1293 together with the driver in the ath module. It is now
1294 possible to tailor chip support (i.e. reduce the set of chips
1295 and thereby the code size); consult ath_hal(4) for details.
1298 __FreeBSD_version 800054 adds memory barriers to
1299 <machine/atomic.h>, new interfaces to ifnet to facilitate
1300 multiple hardware transmit queues for cards that support
1301 them, and a lock-less ring-buffer implementation to
1302 enable drivers to more efficiently manage queueing of
1306 A new version of ZFS (version 13) has been merged to -HEAD.
1307 This version has zpool attribute "listsnapshots" off by
1308 default, which means "zfs list" does not show snapshots,
1309 and is the same as Solaris behavior.
1312 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1315 The uhci, ohci, ehci and slhci USB Host controller drivers have
1316 been put into separate modules. If you load the usb module
1317 separately through loader.conf you will need to load the
1318 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1319 controller add the following to loader.conf:
1325 The ABI used by the PMC toolset has changed. Please keep
1326 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1330 atapci kernel module now includes only generic PCI ATA
1331 driver. AHCI driver moved to ataahci kernel module.
1332 All vendor-specific code moved into separate kernel modules:
1333 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1334 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1335 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1336 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1339 The TTY subsystem of the kernel has been replaced by a new
1340 implementation, which provides better scalability and an
1341 improved driver model. Most common drivers have been migrated to
1342 the new TTY subsystem, while others have not. The following
1343 drivers have not yet been ported to the new TTY layer:
1346 cy, digi, rc, rp, sio
1352 ng_h4, ng_tty, ppp, sl, snp
1354 Adding these drivers to your kernel configuration file shall
1355 cause compilation to fail.
1358 ntpd has been upgraded to 4.2.4p5.
1361 OpenSSH has been upgraded to 5.1p1.
1363 For many years, FreeBSD's version of OpenSSH preferred DSA
1364 over RSA for host and user authentication keys. With this
1365 upgrade, we've switched to the vendor's default of RSA over
1366 DSA. This may cause upgraded clients to warn about unknown
1367 host keys even for previously known hosts. Users should
1368 follow the usual procedure for verifying host keys before
1369 accepting the RSA key.
1371 This can be circumvented by setting the "HostKeyAlgorithms"
1372 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1375 Please note that the sequence of keys offered for
1376 authentication has been changed as well. You may want to
1377 specify IdentityFile in a different order to revert this
1381 The sio(4) driver has been removed from the i386 and amd64
1382 kernel configuration files. This means uart(4) is now the
1383 default serial port driver on those platforms as well.
1385 To prevent collisions with the sio(4) driver, the uart(4) driver
1386 uses different names for its device nodes. This means the
1387 onboard serial port will now most likely be called "ttyu0"
1388 instead of "ttyd0". You may need to reconfigure applications to
1389 use the new device names.
1391 When using the serial port as a boot console, be sure to update
1392 /boot/device.hints and /etc/ttys before booting the new kernel.
1393 If you forget to do so, you can still manually specify the hints
1394 at the loader prompt:
1396 set hint.uart.0.at="isa"
1397 set hint.uart.0.port="0x3F8"
1398 set hint.uart.0.flags="0x10"
1399 set hint.uart.0.irq="4"
1403 The gpt(8) utility has been removed. Use gpart(8) to partition
1407 The version that Linuxulator emulates was changed from 2.4.2
1408 to 2.6.16. If you experience any problems with Linux binaries
1409 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1410 if it fixes the problem contact emulation mailing list.
1413 ISDN4BSD (I4B) was removed from the src tree. You may need to
1414 update a your kernel configuration and remove relevant entries.
1417 I have checked in code to support multiple routing tables.
1418 See the man pages setfib(1) and setfib(2).
1419 This is a hopefully backwards compatible version,
1420 but to make use of it you need to compile your kernel
1421 with options ROUTETABLES=2 (or more up to 16).
1424 The 802.11 wireless support was redone to enable multi-bss
1425 operation on devices that are capable. The underlying device
1426 is no longer used directly but instead wlanX devices are
1427 cloned with ifconfig. This requires changes to rc.conf files.
1428 For example, change:
1429 ifconfig_ath0="WPA DHCP"
1432 ifconfig_wlan0="WPA DHCP"
1433 see rc.conf(5) for more details. In addition, mergemaster of
1434 /etc/rc.d is highly recommended. Simultaneous update of userland
1435 and kernel wouldn't hurt either.
1437 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1438 modules were merged into the base wlan module. All references
1439 to these modules (e.g. in kernel config files) must be removed.
1442 psm(4) has gained write(2) support in native operation level.
1443 Arbitrary commands can be written to /dev/psm%d and status can
1444 be read back from it. Therefore, an application is responsible
1445 for status validation and error recovery. It is a no-op in
1446 other operation levels.
1449 Support for KSE threading has been removed from the kernel. To
1450 run legacy applications linked against KSE libmap.conf may
1451 be used. The following libmap.conf may be used to ensure
1452 compatibility with any prior release:
1454 libpthread.so.1 libthr.so.1
1455 libpthread.so.2 libthr.so.2
1456 libkse.so.3 libthr.so.3
1459 The layout of struct vmspace has changed. This affects libkvm
1460 and any executables that link against libkvm and use the
1461 kvm_getprocs() function. In particular, but not exclusively,
1462 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1463 The effects are minimal, but it's advisable to upgrade world
1467 The latest em driver no longer has support in it for the
1468 82575 adapter, this is now moved to the igb driver. The
1469 split was done to make new features that are incompatible
1470 with older hardware easier to do.
1473 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1474 likewise the kernel option is now GEOM_LINUX_LVM.
1477 The default NFS mount mode has changed from UDP to TCP for
1478 increased reliability. If you rely on (insecurely) NFS
1479 mounting across a firewall you may need to update your
1483 Belatedly note the addition of m_collapse for compacting
1487 The fts(3) structures have been changed to use adequate
1488 integer types for their members and so to be able to cope
1489 with huge file trees. The old fts(3) ABI is preserved
1490 through symbol versioning in libc, so third-party binaries
1491 using fts(3) should still work, although they will not take
1492 advantage of the extended types. At the same time, some
1493 third-party software might fail to build after this change
1494 due to unportable assumptions made in its source code about
1495 fts(3) structure members. Such software should be fixed
1496 by its vendor or, in the worst case, in the ports tree.
1497 FreeBSD_version 800015 marks this change for the unlikely
1498 case that a portable fix is impossible.
1501 To upgrade to -current after this date, you must be running
1502 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1503 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1506 The ADAPTIVE_GIANT kernel option has been retired because its
1507 functionality is the default now.
1510 The AT keyboard emulation of sunkbd(4) has been turned on
1511 by default. In order to make the special symbols of the Sun
1512 keyboards driven by sunkbd(4) work under X these now have
1513 to be configured the same way as Sun USB keyboards driven
1514 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1516 Option "XkbLayout" "us"
1517 Option "XkbRules" "xorg"
1518 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1521 It has been decided that it is desirable to provide ABI
1522 backwards compatibility to the FreeBSD 4/5/6 versions of the
1523 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1524 broken with the introduction of PCI domain support (see the
1525 20070930 entry). Unfortunately, this required the ABI of
1526 PCIOCGETCONF to be broken again in order to be able to
1527 provide backwards compatibility to the old version of that
1528 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1529 again. As for prominent ports this affects neither pciutils
1530 nor xorg-server this time, the hal port needs to be rebuilt
1534 The misnamed kthread_create() and friends have been renamed
1535 to kproc_create() etc. Many of the callers already
1536 used kproc_start()..
1537 I will return kthread_create() and friends in a while
1538 with implementations that actually create threads, not procs.
1539 Renaming corresponds with version 800002.
1548 Avoid using make -j when upgrading. While generally safe, there are
1549 sometimes problems using -j to upgrade. If your upgrade fails with
1550 -j, please try again without -j. From time to time in the past there
1551 have been problems using -j with buildworld and/or installworld. This
1552 is especially true when upgrading between "distant" versions (eg one
1553 that cross a major release boundary or several minor releases, or when
1554 several months have passed on the -current branch).
1556 Sometimes, obscure build problems are the result of environment
1557 poisoning. This can happen because the make utility reads its
1558 environment when searching for values for global variables. To run
1559 your build attempts in an "environmental clean room", prefix all make
1560 commands with 'env -i '. See the env(1) manual page for more details.
1562 When upgrading from one major version to another it is generally best
1563 to upgrade to the latest code in the currently installed branch first,
1564 then do an upgrade to the new branch. This is the best-tested upgrade
1565 path, and has the highest probability of being successful. Please try
1566 this approach before reporting problems with a major version upgrade.
1570 When upgrading the boot ZFS pool to a new version, always follow
1573 1.) recompile and reinstall the ZFS boot loader and boot block
1574 (this is part of "make buildworld" and "make installworld")
1576 2.) update the ZFS boot block on your boot drive
1578 The following example updates the ZFS boot block on the first
1579 partition (freebsd-boot) of a GPT partitioned drive ad0:
1580 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1582 Non-boot pools do not need these updates.
1586 If you are updating from a prior version of FreeBSD (even one just
1587 a few days old), you should follow this procedure. It is the most
1588 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1590 make kernel-toolchain
1591 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1592 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1594 To test a kernel once
1595 ---------------------
1596 If you just want to boot a kernel once (because you are not sure
1597 if it works, or if you want to boot a known bad kernel to provide
1598 debugging information) run
1599 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1600 nextboot -k testkernel
1602 To just build a kernel when you know that it won't mess you up
1603 --------------------------------------------------------------
1604 This assumes you are already running a CURRENT system. Replace
1605 ${arch} with the architecture of your machine (e.g. "i386",
1606 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1608 cd src/sys/${arch}/conf
1609 config KERNEL_NAME_HERE
1610 cd ../compile/KERNEL_NAME_HERE
1615 If this fails, go to the "To build a kernel" section.
1617 To rebuild everything and install it on the current system.
1618 -----------------------------------------------------------
1619 # Note: sometimes if you are running current you gotta do more than
1620 # is listed here if you are upgrading from a really old current.
1622 <make sure you have good level 0 dumps>
1624 make kernel KERNCONF=YOUR_KERNEL_HERE
1626 <reboot in single user> [3]
1634 To cross-install current onto a separate partition
1635 --------------------------------------------------
1636 # In this approach we use a separate partition to hold
1637 # current's root, 'usr', and 'var' directories. A partition
1638 # holding "/", "/usr" and "/var" should be about 2GB in
1641 <make sure you have good level 0 dumps>
1644 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1645 <maybe newfs current's root partition>
1646 <mount current's root partition on directory ${CURRENT_ROOT}>
1647 make installworld DESTDIR=${CURRENT_ROOT}
1648 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1649 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1650 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1651 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1652 <reboot into current>
1653 <do a "native" rebuild/install as described in the previous section>
1654 <maybe install compatibility libraries from ports/misc/compat*>
1658 To upgrade in-place from stable to current
1659 ----------------------------------------------
1660 <make sure you have good level 0 dumps>
1662 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1664 <reboot in single user> [3]
1671 Make sure that you've read the UPDATING file to understand the
1672 tweaks to various things you need. At this point in the life
1673 cycle of current, things change often and you are on your own
1674 to cope. The defaults can also change, so please read ALL of
1675 the UPDATING entries.
1677 Also, if you are tracking -current, you must be subscribed to
1678 freebsd-current@freebsd.org. Make sure that before you update
1679 your sources that you have read and understood all the recent
1680 messages there. If in doubt, please track -stable which has
1681 much fewer pitfalls.
1683 [1] If you have third party modules, such as vmware, you
1684 should disable them at this point so they don't crash your
1687 [3] From the bootblocks, boot -s, and then do
1692 adjkerntz -i # if CMOS is wall time
1693 Also, when doing a major release upgrade, it is required that
1694 you boot into single user mode to do the installworld.
1696 [4] Note: This step is non-optional. Failure to do this step
1697 can result in a significant reduction in the functionality of the
1698 system. Attempting to do it by hand is not recommended and those
1699 that pursue this avenue should read this file carefully, as well
1700 as the archives of freebsd-current and freebsd-hackers mailing lists
1701 for potential gotchas. The -U option is also useful to consider.
1702 See mergemaster(8) for more information.
1704 [5] Usually this step is a noop. However, from time to time
1705 you may need to do this if you get unknown user in the following
1706 step. It never hurts to do it all the time. You may need to
1707 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1708 install) after the buildworld before this step if you last updated
1709 from [78]-stable or 9-stable before 20130430.
1711 [6] This only deletes old files and directories. Old libraries
1712 can be deleted by "make delete-old-libs", but you have to make
1713 sure that no program is using those libraries anymore.
1715 [8] In order to have a kernel that can run the 4.x binaries needed to
1716 do an installworld, you must include the COMPAT_FREEBSD4 option in
1717 your kernel. Failure to do so may leave you with a system that is
1718 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1719 required to run the 5.x binaries on more recent kernels. And so on
1720 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1722 Make sure that you merge any new devices from GENERIC since the
1723 last time you updated your kernel config file.
1725 [9] When checking out sources, you must include the -P flag to have
1726 cvs prune empty directories.
1728 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1729 "?=" instead of the "=" assignment operator, so that buildworld can
1730 override the CPUTYPE if it needs to.
1732 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1733 not on the command line, or in /etc/make.conf. buildworld will
1734 warn if it is improperly defined.
1737 This file contains a list, in reverse chronological order, of major
1738 breakages in tracking -current. It is not guaranteed to be a complete
1739 list of such breakages, and only contains entries since October 10, 2007.
1740 If you need to see UPDATING entries from before that date, you will need
1741 to fetch an UPDATING file from an older FreeBSD release.
1743 Copyright information:
1745 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1747 Redistribution, publication, translation and use, with or without
1748 modification, in full or in part, in any form or format of this
1749 document are permitted without further permission from the author.
1751 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1752 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1753 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1754 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1755 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1756 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1757 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1758 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1759 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1760 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1761 POSSIBILITY OF SUCH DAMAGE.
1763 Contact Warner Losh if you have any questions about your use of