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 20151026: p29 FreeBSD-SA-15:25.ntp
16 Fix multiple NTP vulnerabilities. New NTP version is 4.2.8p4.
18 The configuration file syntax has been changed, thus mergemaster
19 run is recommended. Now the "kod" parameter requires "limited"
20 parameter. If the ntp.conf is not updated, the ntpd will run
21 successfully, but with KoD disabled.
23 20151002: p28 FreeBSD-SA-15:24.rpcbind [revised]
24 Revised patch to address a regression that prevents NIS from working.
26 20150929: p27 FreeBSD-SA-15:24.rpcbind
28 Fix rpcbind(8) remote denial of service. [SA-15:24]
30 20150916: p26 FreeBSD-EN-15:18.pkg
32 Implement pubkey support for pkg(7) bootstrap. [EN-15:18]
34 20150902: p25 FreeBSD-SA-15:23.bind
36 Fix BIND remote denial of service vulnerability. [SA-15:23]
38 20150825: p24 FreeBSD-SA-15:21.amd64
39 FreeBSD-SA-15:22.openssh
42 Fix local privilege escalation in IRET handler. [SA-15:21]
44 Fix OpenSSH multiple vulnerabilities. [SA-15:22]
46 Fix insufficient check of unsupported pkg(7) signature methods.
49 20150818: p23 FreeBSD-SA-15:20.expat
51 Fix multiple integer overflows in expat (libbsdxml) XML parser.
54 20150805: p22 FreeBSD-SA-15:19.routed
56 Fix routed remote denial of service vulnerability.
58 20150728: p21 FreeBSD-SA-15:15.tcp
59 FreeBSD-SA-15:16.openssh
62 Fix resource exhaustion in TCP reassembly. [SA-15:15]
64 Fix OpenSSH multiple vulnerabilities. [SA-15:16]
66 Fix BIND remote denial of service vulnerability. [SA-15:17]
68 20150721: p20 FreeBSD-SA-15:13.tcp
70 Fix resource exhaustion due to sessions stuck in LAST_ACK state.
73 20150707: p19 FreeBSD-SA-15:11.bind
74 Fix BIND resolver remote denial of service when validating.
76 20150630: p18 FreeBSD-EN-15:08.sendmail [revised]
77 FreeBSD-EN-15:09.xlocale
79 Improvements to sendmail TLS/DH interoperability. [EN-15:08]
81 Fix inconsistency between locale and rune locale states.
84 20150618: p17 FreeBSD-EN-15:08.sendmail
85 Improvements to sendmail TLS/DH interoperability. [EN-15:08]
87 20150612: p16 FreeBSD-SA-15:10.openssl
88 Fix multiple vulnerabilities in OpenSSL. [SA-15:10]
90 20150609: p15 FreeBSD-EN-15:06.file
92 Updated base system file(1) to 5.22 to address multiple denial
95 20150513: p14 FreeBSD-EN-15:04.freebsd-update
97 Fix bug with freebsd-update(8) that does not ensure the previous
98 upgrade was completed. [EN-15:04]
100 20150407: p13 FreeBSD-SA-15:04.igmp [revised]
102 FreeBSD-SA-15:09.ipv6
104 Improved patch for SA-15:04.igmp.
106 Fix multiple vulnerabilities of ntp. [SA-15:07]
108 Fix Denial of Service with IPv6 Router Advertisements. [SA-15:09]
111 Fix patch for SA-15:06.openssl.
113 20150319: p11 FreeBSD-SA-15:06.openssl
114 Fix multiple vulnerabilities in OpenSSL. [SA-15:06]
116 20150225: p10 FreeBSD-SA-15:04.igmp
117 FreeBSD-SA-15:05.bind
119 FreeBSD-EN-15:02.openssl
120 FreeBSD-EN-15:03.freebsd-update
122 Fix integer overflow in IGMP protocol. [SA-15:04]
124 Fix BIND remote denial of service vulnerability. [SA-15:05]
126 Fix vt(4) crash with improper ioctl parameters. [EN-15:01]
128 Updated base system OpenSSL to 0.9.8zd. [EN-15:02]
130 Fix freebsd-update libraries update ordering issue. [EN-15:03]
132 20150127: p9 FreeBSD-SA-15:02.kmem
133 FreeBSD-SA-15:03.sctp
135 Fix SCTP SCTP_SS_VALUE kernel memory corruption and disclosure
136 vulnerability. [SA-15:02]
138 Fix SCTP stream reset vulnerability. [SA-15:03]
140 20150114: p8 FreeBSD-SA-15:01.openssl
141 Fix multiple vulnerabilities in OpenSSL. [SA-15:01]
143 20141223: p7 FreeBSD-SA-14:31.ntp
144 FreeBSD-EN-14:13.freebsd-update
146 Fix multiple vulnerabilities in NTP suite. [SA-14:31]
147 Fix directory deletion issue in freebsd-update. [EN-14:13]
149 20141210: p6 FreeBSD-SA-14:28.file
150 FreeBSD-SA-14:29.bind
152 Fix multiple vulnerabilities in file(1) and libmagic(3).
155 Fix BIND remote denial of service vulnerability. [SA-14:29]
157 20141104: p5 FreeBSD-SA-14:25.setlogin
161 Fix kernel stack disclosure in setlogin(2) / getlogin(2).
164 Fix remote command execution in ftp(1). [SA-14:26]
166 Fix NFSv4 and ZFS cache consistency issue. [EN-14:12]
168 20141022: p4 FreeBSD-EN-14:10.tzdata
169 FreeBSD-EN-14:11.crypt
171 Time zone data file update. [EN-14:10]
173 Change crypt(3) default hashing algorithm back to DES. [EN-14:11]
175 20141021: p3 FreeBSD-SA-14:20.rtsold
176 FreeBSD-SA-14:21.routed
177 FreeBSD-SA-14:22.namei
178 FreeBSD-SA-14:23.openssl
180 Fix rtsold(8) remote buffer overflow vulnerability. [SA-14:20]
182 Fix routed(8) remote denial of service vulnerability. [SA-14:21]
184 Fix memory leak in sandboxed namei lookup. [SA-14:22]
186 Fix OpenSSL multiple vulnerabilities. [SA-14:23]
188 20140916: p2 FreeBSD-SA-14:19.tcp
189 Fix Denial of Service in TCP packet processing. [SA-14:19]
191 20140909: p1 FreeBSD-SA-14:18.openssl
192 Fix OpenSSL multiple vulnerabilities. [SA-14:18]
198 On i386 and amd64 systems, the onifconsole flag is now set by default
199 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
200 as a login TTY if it is set in the bootloader as an active kernel
201 console. No changes in behavior should result otherwise. To revert to
202 the previous behavior, set ttyu0 to "off" in /etc/ttys.
205 Clang and llvm have been upgraded to 3.4.1 release.
208 Clang and llvm have been upgraded to 3.4 release.
211 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
212 been deprecated and will not be part of FreeBSD 11.0 and later
213 releases. If you use this driver, please consider switching to
214 the nfe(4) driver instead.
217 The behavior of gss_pseudo_random() for the krb5 mechanism
218 has changed, for applications requesting a longer random string
219 than produced by the underlying enctype's pseudo-random() function.
220 In particular, the random string produced from a session key of
221 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
222 be different at the 17th octet and later, after this change.
223 The counter used in the PRF+ construction is now encoded as a
224 big-endian integer in accordance with RFC 4402.
225 __FreeBSD_version is bumped to 902505.
231 Behavior of devfs rules path matching has been changed.
232 Pattern is now always matched against fully qualified devfs
233 path and slash characters must be explicitly matched by
234 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
235 subdirectories must be reviewed.
238 hastctl(8)'s `status' command output changed to terse one-liner format.
239 Scripts using this should switch to `list' command or be rewritten.
242 Fix a bug that allowed a tracing process (e.g. gdb) to write
243 to a memory-mapped file in the traced process's address space
244 even if neither the traced process nor the tracing process had
245 write access to that file.
248 Added ZFS TRIM support which is enabled by default. To disable
249 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
251 Creating new ZFS pools and adding new devices to existing pools
252 first performs a full device level TRIM which can take a significant
253 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
254 to disable this behaviour.
256 ZFS TRIM requires the underlying device support BIO_DELETE which
257 is currently provided by methods such as ATA TRIM and SCSI UNMAP
258 via CAM, which are typically supported by SSD's.
260 Stats for ZFS TRIM can be monitored by looking at the sysctl's
261 under kstat.zfs.misc.zio_trim.
264 `list' command has been added to hastctl(8). For now, it is full
265 equivalent of `status' command.
266 WARNING: in the near future the output of hastctl's status command
267 will change to more terse format. If you use `hastctl status'
268 for parsing in your scripts, switch to `hastctl list'.
271 The mergemaster command now uses the default MAKEOBJDIRPREFIX
272 rather than creating it's own in the temporary directory in
273 order allow access to bootstrapped versions of tools such as
274 install and mtree. When upgrading from version of FreeBSD where
275 the install command does not support -l, you will need to
276 install a new mergemaster command if mergemaster -p is required.
277 This can be accomplished with the command (cd src/usr.sbin/mergemaster
280 Due to the use of the new -l option to install(1) during build
281 and install, you must take care not to directly set the INSTALL
282 make variable in your /etc/make.conf, /etc/src.conf, or on the
283 command line. If you wish to use the -C flag for all installs
284 you may be able to add INSTALL+=-C to /etc/make.conf or
288 Fix a bug that allows NFS clients to issue READDIR on files.
291 The install(1) option -M has changed meaning and now takes an
292 argument that is a file or path to append logs to. In the
293 unlikely event that -M was the last option on the command line
294 and the command line contained at least two files and a target
295 directory the first file will have logs appended to it. The -M
296 option served little practical purpose in the last decade so it's
297 used expected to be extremely rare.
300 A new compression method (lz4) has been merged to. Please refer to
301 zpool-features(7) for more information.
303 Please refer to the "ZFS notes" section of this file for information
304 on upgrading boot ZFS pools.
307 The VFS KBI was changed with the merge of several nullfs
308 optimizations and fixes. All filesystem modules must be
312 With the addition of auditdistd(8), a new auditdistd user is now
313 depended on during installworld. "mergemaster -p" can be used to add
314 the user prior to installworld, as documented in the handbook.
320 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
321 Starting with this version the old system of ZFS pool versioning
322 is superseded by "feature flags". This concept enables forward
323 compatibility against certain future changes in functionality of ZFS
324 pools. The first two read-only compatible "feature flags" for ZFS
325 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
326 For more information read the new zpool-features(7) manual page.
327 Please refer to the "ZFS notes" section of this file for information
328 on upgrading boot ZFS pools.
331 The commit introducing bsd.compiler.mk breaks the traditional
332 building of kernels before this point. Add -m ${SRC}/share/mk
333 (for the right value of SRC) to your command lines to work
334 around; update your useland to a point after this; or use the
335 buildkernel/installkernel top-level targets. See also 20120829.
338 The IPFIREWALL_FORWARD kernel option has been removed. Its
339 functionality now turned on by default.
342 The random(4) support for the VIA hardware random number
343 generator (`PADLOCK') is no longer enabled unconditionally.
344 Add the PADLOCK_RNG option in the custom kernel config if
345 needed. The GENERIC kernels on i386 and amd64 do include the
346 option, so the change only affects the custom kernel
350 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
351 the traditional method, you must update your system with an installworld
352 before the kernel will build. The documented make buildkernel/installkernel
353 interfaces (coupled with fresh make kernel-toolchain) continue to work.
356 The sparc64 ZFS loader has been changed to no longer try to auto-
357 detect ZFS providers based on diskN aliases but now requires these
358 to be explicitly listed in the OFW boot-device environment variable.
361 Now unix domain sockets behave "as expected" on nullfs(5). Previously
362 nullfs(5) did not pass through all behaviours to the underlying layer,
363 as a result if we bound to a socket on the lower layer we could connect
364 only to the lower path; if we bound to the upper layer we could connect
365 only to the upper path. The new behavior is one can connect to both the
366 lower and the upper paths regardless what layer path one binds to.
369 The acpi_wmi(4) status device /dev/wmistat has been renamed to
373 A new VOP_ADVISE() was added to support posix_fadvise(2). All
374 filesystem modules must be recompiled.
377 The interface of the VOP_VPTOCNP(9) changed, now the returned
378 vnode shall be referenced, previously it was required to be
379 only held. All in-tree filesystems are converted.
385 The broken amd(4) driver has been replaced with esp(4) in the amd64,
386 i386 and pc98 GENERIC kernel configuration files.
389 This commit modifies vfs_register() so that it uses a hash
390 calculation to set vfc_typenum, which is enabled by default.
391 The first time a system is booted after this change, the
392 vfc_typenum values will change for all file systems. The
393 main effect of this is a change to the NFS server file handles
394 for file systems that use vfc_typenum in their fsid, such as ZFS.
395 It will, however, prevent vfc_typenum from changing when file
396 systems are loaded in a different order for subsequent reboots.
397 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
398 until you are ready to remount all NFS clients after a reboot.
401 Bump the shared library version numbers for libraries that
402 do not use symbol versioning, have changed the ABI compared
403 to stable/8 and which shared library version was not bumped.
404 Done as part of 9.0-RELEASE cycle.
407 During the merge of Capsicum features, the fget(9) KPI was modified.
408 This may require the rebuilding of out-of-tree device drivers --
409 issues have been reported specifically with the nVidia device driver.
410 __FreeBSD_version is bumped to 900041.
412 Also, there is a period between 20110811 and 20110814 where the
413 special devices /dev/{stdin,stdout,stderr} did not work correctly.
414 Building world from a kernel during that window may not work.
417 The packet filter (pf) code has been updated to OpenBSD 4.5.
418 You need to update userland tools to be in sync with kernel.
419 This update breaks backward compatibility with earlier pfsync(4)
420 versions. Care must be taken when updating redundant firewall setups.
423 The following sysctls and tunables are retired on x86 platforms:
425 machdep.hlt_logical_cpus
426 The following sysctl is retired:
427 machdep.hyperthreading_allowed
428 The sysctls were supposed to provide a way to dynamically offline and
429 online selected CPUs on x86 platforms, but the implementation has not
430 been reliable especially with SCHED_ULE scheduler.
431 machdep.hyperthreading_allowed tunable is still available to ignore
432 hyperthreading CPUs at OS level.
433 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
434 where X is an APIC ID of a CPU. Be advised, though, that disabling
435 CPUs in non-uniform fashion will result in non-uniform topology and
436 may lead to sub-optimal system performance with SCHED_ULE, which is
440 cpumask_t type is retired and cpuset_t is used in order to describe
444 Changes to ifconfig(8) for dynamic address family detection mandate
445 that you are running a kernel of 20110525 or later. Make sure to
446 follow the update procedure to boot a new kernel before installing
450 Support for sun4v architecture is officially dropped
453 Several KPI breaking changes have been committed to the mii(4) layer,
454 the PHY drivers and consequently some Ethernet drivers using mii(4).
455 This means that miibus.ko and the modules of the affected Ethernet
456 drivers need to be recompiled.
458 Note to kernel developers: Given that the OUI bit reversion problem
459 was fixed as part of these changes all mii(4) commits related to OUIs,
460 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
461 handling, no longer can be merged verbatim to stable/8 and previous
465 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
466 into their kernel configurations along with 'device pci'.
469 The default NFS client is now the new NFS client, so fstype "newnfs"
470 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
471 Although mounts via fstype "nfs" will usually work without userland
472 changes, it is recommended that the mount(8) and mount_nfs(8)
473 commands be rebuilt from sources and that a link to mount_nfs called
474 mount_oldnfs be created. The new client is compiled into the
475 kernel with "options NFSCL" and this is needed for diskless root
476 file systems. The GENERIC kernel configs have been changed to use
477 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
478 To use the regular/old client, you can "mount -t oldnfs ...". For
479 a diskless root file system, you must also include a line like:
481 vfs.root.mountfrom="oldnfs:"
483 in the boot/loader.conf on the root fs on the NFS server to make
484 a diskless root fs use the old client.
487 The GENERIC kernels for all architectures now default to the new
488 CAM-based ATA stack. It means that all legacy ATA drivers were
489 removed and replaced by respective CAM drivers. If you are using
490 ATA device names in /etc/fstab or other places, make sure to update
491 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
492 where 'Y's are the sequential numbers starting from zero for each type
493 in order of detection, unless configured otherwise with tunables,
494 see cam(4)). There will be symbolic links created in /dev/ to map
495 old adX devices to the respective adaY. They should provide basic
496 compatibility for file systems mounting in most cases, but they do
497 not support old user-level APIs and do not have respective providers
498 in GEOM. Consider using updated management tools with new device names.
500 It is possible to load devices ahci, ata, siis and mvs as modules,
501 but option ATA_CAM should remain in kernel configuration to make ata
502 module work as CAM driver supporting legacy ATA controllers. Device ata
503 still can be used in modular fashion (atacore + ...). Modules atadisk
504 and atapi* are not used and won't affect operation in ATA_CAM mode.
505 Note that to use CAM-based ATA kernel should include CAM devices
506 scbus, pass, da (or explicitly ada), cd and optionally others. All of
507 them are parts of the cam module.
509 ataraid(4) functionality is now supported by the RAID GEOM class.
510 To use it you can load geom_raid kernel module and use graid(8) tool
511 for management. Instead of /dev/arX device names, use /dev/raid/rX.
513 No kernel config options or code have been removed, so if a problem
514 arises, please report it and optionally revert to the old ATA stack.
515 In order to do it you can remove from the kernel config:
520 , and instead add back:
521 device atadisk # ATA disk drives
522 device ataraid # ATA RAID drives
523 device atapicd # ATAPI CDROM drives
524 device atapifd # ATAPI floppy drives
525 device atapist # ATAPI tape drives
528 The default NFS server has been changed to the new server, which
529 was referred to as the experimental server. If you need to switch
530 back to the old NFS server, you must now put the "-o" option on
531 both the mountd and nfsd commands. This can be done using the
532 mountd_flags and nfs_server_flags rc.conf variables until an
533 update to the rc scripts is committed, which is coming soon.
536 The GNU Objective-C runtime library (libobjc), and other Objective-C
537 related components have been removed from the base system. If you
538 require an Objective-C library, please use one of the available ports.
541 ath(4) has been split into bus- and device- modules. if_ath contains
542 the HAL, the TX rate control and the network device code. if_ath_pci
543 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
544 contains the AHB glue. Users need to load both if_ath_pci and if_ath
545 in order to use ath on everything else.
547 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
548 need to load if_ath and if_ath_pci for ath(4) operation.
551 As part of the replacement of sysinstall, the process of building
552 release media has changed significantly. For details, please re-read
553 release(7), which has been updated to reflect the new build process.
556 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
557 is the last available version under GPLv2. It brings a number of new
558 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
559 4.1 and SSE 4.2), better support for powerpc64, a number of new
560 directives, and lots of other small improvements. See the ChangeLog
561 file in contrib/binutils for the full details.
564 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
565 compliant, and will now use half of hash for authentication.
566 This will break interoperability with all stacks (including all
567 actual FreeBSD versions) who implement
568 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
570 The only workaround with such peers is to use another HMAC
571 algorithm for IPsec ("phase 2") authentication.
574 Remove the uio_yield prototype and symbol. This function has
575 been misnamed since it was introduced and should not be
576 globally exposed with this name. The equivalent functionality
577 is now available using kern_yield(curthread->td_user_pri).
578 The function remains undocumented.
581 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
582 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
583 for scalar sysctls is defined but disabled. Code that needs
584 UQUAD to pass the type checking that must compile on older
585 systems where the define is not present can check against
586 __FreeBSD_version >= 900030.
588 The system dialog(1) has been replaced with a new version previously
589 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
590 with the previous version, but the libdialog associated with it has
591 a largely incompatible API. As such, the original version of libdialog
592 will be kept temporarily as libodialog, until its base system consumers
593 are replaced or updated. Bump __FreeBSD_version to 900030.
596 If you are trying to run make universe on a -stable system, and you get
597 the following warning:
598 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
599 unknown. config(8) likely too old."
600 or something similar to it, then you must upgrade your -stable system
601 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
602 stable/8) or build the config from the latest stable/8 branch and
603 install it on your system.
605 Prior to this date, building a current universe on 8-stable system from
606 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
607 error in the first kernel build phase. A new config on those old
608 systems will fix that problem for older versions of -current.
611 The TCP stack has been modified to allow Khelp modules to interact with
612 it via helper hook points and store per-connection data in the TCP
613 control block. Bump __FreeBSD_version to 900029. User space tools that
614 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
618 Generic IEEE 802.3 annex 31B full duplex flow control support has been
619 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
620 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
621 to take advantage of it instead of using custom implementations. This
622 means that these drivers now no longer unconditionally advertise
623 support for flow control but only do so if flow control is a selected
624 media option. This was implemented in the generic support that way in
625 order to allow flow control to be switched on and off via ifconfig(8)
626 with the PHY specific default to typically off in order to protect
627 from unwanted effects. Consequently, if you used flow control with
628 one of the above mentioned drivers you now need to explicitly enable
630 ifconfig bge0 media auto mediaopt flowcontrol
632 Along with the above mentioned changes generic support for setting
633 1000baseT master mode also has been added and brgphy(4), ciphy(4),
634 e1000phy(4) as well as ip1000phy(4) have been converted to take
635 advantage of it. This means that these drivers now no longer take the
636 link0 parameter for selecting master mode but the master media option
637 has to be used instead, for example like in the following:
638 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
640 Selection of master mode now is also available with all other PHY
641 drivers supporting 1000baseT.
644 The TCP stack has received a significant update to add support for
645 modularised congestion control and generally improve the clarity of
646 congestion control decisions. Bump __FreeBSD_version to 900025. User
647 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
648 sockstat) need to be recompiled.
651 The man(1) utility has been replaced by a new version that no longer
652 uses /etc/manpath.config. Please consult man.conf(5) for how to
653 migrate local entries to the new format.
656 The copyright strings printed by login(1) and sshd(8) at the time of a
657 new connection have been removed to follow other operating systems and
661 A workaround for a fixed ld bug has been removed in kernel code,
662 so make sure that your system ld is built from sources after
663 revision 210245 from 2010-07-19 (r211583 if building head kernel
664 on stable/8, r211584 for stable/7; both from 2010-08-21).
665 A symptom of incorrect ld version is different addresses for
666 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
669 The $ipv6_prefer variable in rc.conf(5) has been split into
670 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
672 The $ip6addrctl_policy is a variable to choose a pre-defined
673 address selection policy set by ip6addrctl(8). A value
674 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
677 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
678 flag (see an entry of 20090926) is set on an interface with no
679 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
680 security reason. If you want IPv6 link-local address on all
681 interfaces by default, set this to "YES".
683 The old ipv6_prefer="YES" is equivalent to
684 ipv6_activate_all_interfaces="YES" and
685 ip6addrctl_policy="ipv6_prefer".
688 DTrace has grown support for userland tracing. Due to this, DTrace is
689 now i386 and amd64 only.
690 dtruss(1) is now installed by default on those systems and a new
691 kernel module is needed for userland tracing: fasttrap.
692 No changes to your kernel config file are necessary to enable
693 userland tracing, but you might consider adding 'STRIP=' and
694 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
695 to have informative userland stack traces in DTrace (ustack).
698 The acpi_aiboost(4) driver has been removed in favor of the new
699 aibs(4) driver. You should update your kernel configuration file.
702 BSD grep has been imported to the base system and it is built by
703 default. It is completely BSD licensed, highly GNU-compatible, uses
704 less memory than its GNU counterpart and has a small codebase.
705 However, it is slower than its GNU counterpart, which is mostly
706 noticeable for larger searches, for smaller ones it is measurable
707 but not significant. The reason is complex, the most important factor
708 is that we lack a modern and efficient regex library and GNU
709 overcomes this by optimizing the searches internally. Future work
710 on improving the regex performance is planned, for the meantime,
711 users that need better performance, can build GNU grep instead by
712 setting the WITH_GNU_GREP knob.
715 Due to the import of powerpc64 support, all existing powerpc kernel
716 configuration files must be updated with a machine directive like this:
717 machine powerpc powerpc
719 In addition, an updated config(8) is required to build powerpc kernels
723 A new version of ZFS (version 15) has been merged to -HEAD.
724 This version uses a python library for the following subcommands:
725 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
726 For full functionality of these commands the following port must
727 be installed: sysutils/py-zfs
730 'vm_page's are now hashed by physical address to an array of mutexes.
731 Currently this is only used to serialize access to hold_count. Over
732 time the page queue mutex will be peeled away. This changes the size
733 of pmap on every architecture. And requires all callers of vm_page_hold
734 and vm_page_unhold to be updated.
737 WITH_CTF can now be specified in src.conf (not recommended, there
738 are some problems with static executables), make.conf (would also
739 affect ports which do not use GNU make and do not override the
740 compile targets) or in the kernel config (via "makeoptions
742 When WITH_CTF was specified there before this was silently ignored,
743 so make sure that WITH_CTF is not used in places which could lead
744 to unwanted behavior.
747 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
748 to allow 32-bit compatibility on non-x86 platforms. All kernel
749 configurations on amd64 and ia64 platforms using these options must
750 be modified accordingly.
753 The utmp user accounting database has been replaced with utmpx,
754 the user accounting interface standardized by POSIX.
755 Unfortunately the semantics of utmp and utmpx don't match,
756 making it practically impossible to support both interfaces.
757 The user accounting database is used by tools like finger(1),
758 last(1), talk(1), w(1) and ac(8).
760 All applications in the base system use utmpx. This means only
761 local binaries (e.g. from the ports tree) may still use these
762 utmp database files. These applications must be rebuilt to make
765 After the system has been upgraded, it is safe to remove the old
766 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
767 assuming their contents is of no importance anymore. Old wtmp
768 databases can only be used by last(1) and ac(8) after they have
769 been converted to the new format using wtmpcvt(1).
772 Introduce the kernel thread "deadlock resolver" (which can be enabled
773 via the DEADLKRES option, see NOTES for more details) and the
774 sleepq_type() function for sleepqueues.
777 The rc.firewall and rc.firewall6 were unified, and
778 rc.firewall6 and rc.d/ip6fw were removed.
779 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
780 variables are obsoleted. Instead, the following new rc
781 variables are added to rc.d/ipfw:
783 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
784 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
785 firewall_simple_onet_ipv6, firewall_trusted_ipv6
787 The meanings correspond to the relevant IPv4 variables.
793 The default terminal emulation for syscons(4) has been changed
794 from cons25 to xterm on all platforms except pc98. This means
795 that the /etc/ttys file needs to be updated to ensure correct
796 operation of applications on the console.
798 The terminal emulation style can be toggled per window by using
799 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
800 options can be used to change the compile-time default back to
803 To prevent graphical artifacts, make sure the TERM environment
804 variable is set to match the terminal emulation that is being
805 performed by syscons(4).
808 The layout of the structure ieee80211req_scan_result has changed.
809 Applications that require wireless scan results (e.g. ifconfig(8))
810 from net80211 need to be recompiled.
812 Applications such as wpa_supplicant(8) may require a full world
813 build without using NO_CLEAN in order to get synchronized with the
817 The iwn(4) driver has been updated to support the 5000 and 5150 series.
818 There's one kernel module for each firmware. Adding "device iwnfw"
819 to the kernel configuration file means including all three firmware
820 images inside the kernel. If you want to include just the one for
821 your wireless card, use the devices iwn4965fw, iwn5000fw or
825 The rc.d/network_ipv6, IPv6 configuration script has been integrated
826 into rc.d/netif. The changes are the following:
828 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
829 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
830 Note that both variables need the "inet6" keyword at the head.
832 Do not set $ipv6_network_interfaces manually if you do not
833 understand what you are doing. It is not needed in most cases.
835 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
838 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
839 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
841 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
842 all configured interfaces have "inet6 accept_rtadv" in the
843 $ifconfig_IF_ipv6. These are for backward compatibility.
845 3. A new variable $ipv6_prefer has been added. If NO, IPv6
846 functionality of interfaces with no corresponding
847 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
848 and the default address selection policy of ip6addrctl(8)
849 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
850 Note that if you want to configure IPv6 functionality on the
851 disabled interfaces after boot, first you need to clear the flag by
852 using ifconfig(8) like:
854 ifconfig em0 inet6 -ifdisabled
856 If YES, the default address selection policy is set as
859 The default value of $ipv6_prefer is NO.
861 4. If your system need to receive Router Advertisement messages,
862 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
863 scripts automatically invoke rtsol(8) when the interface becomes
864 UP. The Router Advertisement messages are used for SLAAC
865 (State-Less Address AutoConfiguration).
868 802.11s D3.03 support was committed. This is incompatible with the
869 previous code, which was based on D3.0.
872 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
873 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
874 control whether accepting Router Advertisement messages or not.
875 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
876 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
877 The ifconfig(8) utility now supports these flags.
880 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
881 mount(8) and -a option for df(1) to see them.
884 The old tunable hw.bus.devctl_disable has been superseded by
885 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
886 replaced by hw.bus.devctl_queue=0. The default for this new tunable
890 Remove the option STOP_NMI. The default action is now to use NMI only
891 for KDB via the newly introduced function stop_cpus_hard() and
892 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
895 The stable/8 branch created in subversion. This corresponds to the
896 RELENG_8 branch in CVS.
899 Bump the shared library version numbers for all libraries that do not
900 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
901 __FreeBSD_version to 800105.
904 Due to changes in the implementation of virtual network stack support,
905 all network-related kernel modules must be recompiled. As this change
906 breaks the ABI, bump __FreeBSD_version to 800104.
909 The TOE interface to the TCP syncache has been modified to remove
910 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
911 The cxgb driver is the only TOE consumer affected by this change, and
912 needs to be recompiled along with the kernel. As this change breaks
913 the ABI, bump __FreeBSD_version to 800103.
916 Padding has been added to struct tcpcb, sackhint and tcpstat in
917 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
918 maintaining the ABI. However, this change breaks the ABI, so bump
919 __FreeBSD_version to 800102. User space tools that rely on the size of
920 any of these structs (e.g. sockstat) need to be recompiled.
923 The NFS_LEGACYRPC option has been removed along with the old kernel
924 RPC implementation that this option selected. Kernel configurations
925 may need to be adjusted.
928 The network interface device nodes at /dev/net/<interface> have been
929 removed. All ioctl operations can be performed the normal way using
930 routing sockets. The kqueue functionality can generally be replaced
931 with routing sockets.
934 The documentation from the FreeBSD Documentation Project (Handbook,
935 FAQ, etc.) is now installed via packages by sysinstall(8) and under
936 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
939 The ABI of various structures related to the SYSV IPC API have been
940 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
941 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
945 Layout of struct vnet has changed as routing related variables were
946 moved to their own Vimage module. Modules need to be recompiled. Bump
947 __FreeBSD_version to 800099.
950 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
951 respectively. As long as no more than 16 groups per process are used,
952 no changes should be visible. When more than 16 groups are used, old
953 binaries may fail if they call getgroups() or getgrouplist() with
954 statically sized storage. Recompiling will work around this, but
955 applications should be modified to use dynamically allocated storage
956 for group arrays as POSIX.1-2008 does not cap an implementation's
957 number of supported groups at NGROUPS_MAX+1 as previous versions did.
959 NFS and portalfs mounts may also be affected as the list of groups is
960 truncated to 16. Users of NFS who use more than 16 groups, should
961 take care that negative group permissions are not used on the exported
962 file systems as they will not be reliable unless a GSSAPI based
963 authentication method is used.
966 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
967 option compiles in the support for adaptive spinning for lockmgrs
968 which want to enable it. The lockinit() function now accepts the flag
969 LK_ADAPTIVE in order to make the lock object subject to adaptive
970 spinning when both held in write and read mode.
973 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
974 changed. User applications that use this ioctl need to be rebuilt.
977 The layout of struct thread has changed. Kernel and modules need to
981 The layout of structs ifnet, domain, protosw and vnet_net has changed.
982 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
985 window(1) has been removed from the base system. It can now be
986 installed from ports. The port is called misc/window.
989 The way we are storing and accessing `routing table' entries has
990 changed. Programs reading the FIB, like netstat, need to be
994 A new netisr implementation has been added for FreeBSD 8. Network
995 file system modules, such as igmp, ipdivert, and others, should be
997 Bump __FreeBSD_version to 800096.
1000 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
1004 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
1005 Bump __FreeBSD_version to 800094.
1008 Add mnt_xflag field to 'struct mount'. File system modules need to be
1010 Bump __FreeBSD_version to 800093.
1013 The compiling option ADAPTIVE_SX has been retired while it has been
1014 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
1015 The KPI for sx_init_flags() changes as accepting flags:
1016 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
1017 been introduced in order to handle the reversed logic.
1018 Bump __FreeBSD_version to 800092.
1021 Add support for hierarchical jails. Remove global securelevel.
1022 Bump __FreeBSD_version to 800091.
1025 The layout of struct vnet_net has changed, therefore modules
1027 Bump __FreeBSD_version to 800090.
1030 The newly imported zic(8) produces a new format in the output. Please
1031 run tzsetup(8) to install the newly created data to /etc/localtime.
1034 The sysctl tree for the usb stack has renamed from hw.usb2.* to
1035 hw.usb.* and is now consistent again with previous releases.
1038 802.11 monitor mode support was revised and driver api's were changed.
1039 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
1040 of DLT_IEEE802_11. No user-visible data structures were changed but
1041 applications that use DLT_IEEE802_11 may require changes.
1042 Bump __FreeBSD_version to 800088.
1045 The layout of the following structs has changed: sysctl_oid,
1046 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
1047 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
1048 panics may be experienced. World rebuild is required for
1049 correctly checking networking state from userland.
1050 Bump __FreeBSD_version to 800085.
1053 MLDv2 and Source-Specific Multicast (SSM) have been merged
1054 to the IPv6 stack. VIMAGE hooks are in but not yet used.
1055 The implementation of SSM within FreeBSD's IPv6 stack closely
1056 follows the IPv4 implementation.
1058 For kernel developers:
1060 * The most important changes are that the ip6_output() and
1061 ip6_input() paths no longer take the IN6_MULTI_LOCK,
1062 and this lock has been downgraded to a non-recursive mutex.
1064 * As with the changes to the IPv4 stack to support SSM, filtering
1065 of inbound multicast traffic must now be performed by transport
1066 protocols within the IPv6 stack. This does not apply to TCP and
1067 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
1069 * The KPIs used by IPv6 multicast are similar to those used by
1070 the IPv4 stack, with the following differences:
1071 * im6o_mc_filter() is analogous to imo_multicast_filter().
1072 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
1073 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
1074 * IN6_LOOKUP_MULTI() has been deprecated and removed.
1075 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
1076 for MLDv1 have an additional 'timer' argument which is used to
1077 jitter the initial membership report for the solicited-node
1078 multicast membership on-link.
1079 * This is not strictly needed for MLDv2, which already jitters
1080 its report transmissions. However, the 'timer' argument is
1081 preserved in case MLDv1 is active on the interface.
1083 * The KAME linked-list based IPv6 membership implementation has
1084 been refactored to use a vector similar to that used by the IPv4
1086 Code which maintains a list of its own multicast memberships
1087 internally, e.g. carp, has been updated to reflect the new
1090 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
1091 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
1092 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
1093 implementation constraint which needs to be addressed in HEAD.
1095 For application developers:
1097 * The changes are broadly similar to those made for the IPv4
1100 * The use of IPv4 and IPv6 multicast socket options on the same
1101 socket, using mapped addresses, HAS NOT been tested or supported.
1103 * There are a number of issues with the implementation of various
1104 IPv6 multicast APIs which need to be resolved in the API surface
1105 before the implementation is fully compatible with KAME userland
1106 use, and these are mostly to do with interface index treatment.
1108 * The literature available discusses the use of either the delta / ASM
1109 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
1110 using setsourcefilter(3)/getsourcefilter(3). For more information
1111 please refer to RFC 3768, 'Socket Interface Extensions for
1112 Multicast Source Filters'.
1114 * Applications which use the published RFC 3678 APIs should be fine.
1116 For systems administrators:
1118 * The mtest(8) utility has been refactored to support IPv6, in
1119 addition to IPv4. Interface addresses are no longer accepted
1120 as arguments, their names must be used instead. The utility
1121 will map the interface name to its first IPv4 address as
1122 returned by getifaddrs(3).
1124 * The ifmcstat(8) utility has also been updated to print the MLDv2
1125 endpoint state and source filter lists via sysctl(3).
1127 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
1128 loopback of IPv6 multicast datagrams by default; it defaults to 1
1129 to preserve the existing behaviour. Disabling multicast loopback is
1130 recommended for optimal system performance.
1132 * The IPv6 MROUTING code has been changed to examine this sysctl
1133 instead of attempting to perform a group lookup before looping
1134 back forwarded datagrams.
1136 Bump __FreeBSD_version to 800084.
1139 Implement low-level Bluetooth HCI API.
1140 Bump __FreeBSD_version to 800083.
1143 The layout of struct malloc_type, used by modules to register new
1144 memory allocation types, has changed. Most modules will need to
1145 be rebuilt or panics may be experienced.
1146 Bump __FreeBSD_version to 800081.
1149 Anticipate overflowing inp_flags - add inp_flags2.
1150 This changes most offsets in inpcb, so checking v4 connection
1151 state will require a world rebuild.
1152 Bump __FreeBSD_version to 800080.
1155 Add an llentry to struct route and struct route_in6. Modules
1156 embedding a struct route will need to be recompiled.
1157 Bump __FreeBSD_version to 800079.
1160 The size of rt_metrics_lite and by extension rtentry has changed.
1161 Networking administration apps will need to be recompiled.
1162 The route command now supports show as an alias for get, weighting
1163 of routes, sticky and nostick flags to alter the behavior of stateful
1165 Bump __FreeBSD_version to 800078.
1168 Do not use Giant for kbdmux(4) locking. This is wrong and
1169 apparently causing more problems than it solves. This will
1170 re-open the issue where interrupt handlers may race with
1171 kbdmux(4) in polling mode. Typical symptoms include (but
1172 not limited to) duplicated and/or missing characters when
1173 low level console functions (such as gets) are used while
1174 interrupts are enabled (for example geli password prompt,
1175 mountroot prompt etc.). Disabling kbdmux(4) may help.
1178 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1179 kernel modules referencing any of the above need to be recompiled.
1180 Bump __FreeBSD_version to 800075.
1183 GEOM_PART has become the default partition slicer for storage devices,
1184 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1185 introduces some changes:
1187 MSDOS/EBR: the devices created from MSDOS extended partition entries
1188 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1189 to devices with offset-based names. fstabs may need to be modified.
1191 BSD: the "geometry does not match label" warning is harmless in most
1192 cases but it points to problems in file system misalignment with
1193 disk geometry. The "c" partition is now implicit, covers the whole
1194 top-level drive and cannot be (mis)used by users.
1196 General: Kernel dumps are now not allowed to be written to devices
1197 whose partition types indicate they are meant to be used for file
1198 systems (or, in case of MSDOS partitions, as something else than
1201 Most of these changes date approximately from 200812.
1204 The uscanner(4) driver has been removed from the kernel. This follows
1205 Linux removing theirs in 2.6 and making libusb the default interface
1206 (supported by sane).
1209 The multicast forwarding code has been cleaned up. netstat(1)
1210 only relies on KVM now for printing bandwidth upcall meters.
1211 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1212 ip6_mroute_mod respectively. The config(5) options for statically
1213 compiling this code remain the same, i.e. 'options MROUTING'.
1216 Support for the IFF_NEEDSGIANT network interface flag has been
1217 removed, which means that non-MPSAFE network device drivers are no
1218 longer supported. In particular, if_ar, if_sr, and network device
1219 drivers from the old (legacy) USB stack can no longer be built or
1223 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1224 a bunch of new language catalog files have also been added.
1225 This means that some common libc messages are now localized and
1226 they depend on the LC_MESSAGES environmental variable.
1229 The k8temp(4) driver has been renamed to amdtemp(4) since
1230 support for Family 10 and Family 11 CPU families was added.
1233 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1234 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1236 For kernel developers, the most important changes are that the
1237 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1238 and this lock has been downgraded to a non-recursive mutex.
1240 Transport protocols (UDP, Raw IP) are now responsible for filtering
1241 inbound multicast traffic according to group membership and source
1242 filters. The imo_multicast_filter() KPI exists for this purpose.
1243 Transports which do not use multicast (SCTP, TCP) already reject
1244 multicast by default. Forwarding and receive performance may improve
1245 as a mutex acquisition is no longer needed in the ip_input()
1246 low-level input path. in_addmulti() and in_delmulti() are shimmed
1247 to new KPIs which exist to support SSM in-kernel.
1249 For application developers, it is recommended that loopback of
1250 multicast datagrams be disabled for best performance, as this
1251 will still cause the lock to be taken for each looped-back
1252 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1253 be tuned to 0 to disable loopback by default; it defaults to 1
1254 to preserve the existing behaviour.
1256 For systems administrators, to obtain best performance with
1257 multicast reception and multiple groups, it is always recommended
1258 that a card with a suitably precise hash filter is used. Hash
1259 collisions will still result in the lock being taken within the
1260 transport protocol input path to check group membership.
1262 If deploying FreeBSD in an environment with IGMP snooping switches,
1263 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1264 enabled; this forces 224.0.0.0/24 group membership to be announced
1267 The size of 'struct igmpstat' has changed; netstat needs to be
1268 recompiled to reflect this.
1269 Bump __FreeBSD_version to 800070.
1272 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1273 updated to use it. This requires a buildworld/installworld in order to
1274 update the library and dependencies (usbconfig, etc). Its advisable to
1275 rebuild all ports which uses libusb. More specific directions are given
1276 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1277 libusb are no longer required and can be removed.
1280 A workaround is committed to allow the creation of System V shared
1281 memory segment of size > 2 GB on the 64-bit architectures.
1282 Due to a limitation of the existing ABI, the shm_segsz member
1283 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1284 wrong for large segments. Note that limits must be explicitly
1285 raised to allow such segments to be created.
1288 The layout of struct ifnet has changed, requiring a rebuild of all
1289 network device driver modules.
1292 The /dev handling for the new USB stack has changed, a
1293 buildworld/installworld is required for libusb20.
1296 The new USB2 stack has now been permanently moved in and all kernel and
1297 module names reverted to their previous values (eg, usb, ehci, ohci,
1298 ums, ...). The old usb stack can be compiled in by prefixing the name
1299 with the letter 'o', the old usb modules have been removed.
1300 Updating entry 20090216 for xorg and 20090215 for libmap may still
1304 The rc.conf(5) option if_up_delay has been renamed to
1305 defaultroute_delay to better reflect its purpose. If you have
1306 customized this setting in /etc/rc.conf you need to update it to
1310 xorg 7.4 wants to configure its input devices via hald which does not
1311 yet work with USB2. If the keyboard/mouse does not work in xorg then
1313 Option "AllowEmptyInput" "off"
1314 to your ServerLayout section. This will cause X to use the configured
1315 kbd and mouse sections from your xorg.conf.
1318 The GENERIC kernels for all architectures now default to the new USB2
1319 stack. No kernel config options or code have been removed so if a
1320 problem arises please report it and optionally revert to the old USB
1321 stack. If you are loading USB kernel modules or have a custom kernel
1322 that includes GENERIC then ensure that usb names are also changed over,
1323 eg uftdi -> usb2_serial_ftdi.
1325 Older programs linked against the ports libusb 0.1 need to be
1326 redirected to the new stack's libusb20. /etc/libmap.conf can
1328 # Map old usb library to new one for usb2 stack
1329 libusb-0.1.so.8 libusb20.so.1
1332 All USB ethernet devices now attach as interfaces under the name ueN
1333 (eg. ue0). This is to provide a predictable name as vendors often
1334 change usb chipsets in a product without notice.
1337 The ichsmb(4) driver has been changed to require SMBus slave
1338 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1339 All of the other SMBus controller drivers require left-justified
1340 slave addresses, so this change makes all the drivers provide the
1344 INET6 statistics (struct ip6stat) was updated.
1345 netstat(1) needs to be recompiled.
1348 NTFS has been removed from GENERIC kernel on amd64 to match
1349 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1350 will load ntfs.ko module automatically when NTFS support is
1351 actually needed, unless ntfs.ko is not installed or security
1352 level prohibits loading kernel modules. If either is the case,
1353 "options NTFS" has to be added into kernel config.
1356 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1357 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1358 800061. User space tools that rely on the size of struct tcpcb in
1359 tcp_var.h (e.g. sockstat) need to be recompiled.
1362 ng_tty(4) module updated to match the new TTY subsystem.
1363 Due to API change, user-level applications must be updated.
1364 New API support added to mpd5 CVS and expected to be present
1365 in next mpd5.3 release.
1368 With __FreeBSD_version 800060 the makefs tool is part of
1369 the base system (it was a port).
1372 The afdata and ifnet locks have been changed from mutexes to
1373 rwlocks, network modules will need to be re-compiled.
1376 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1377 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1378 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1379 architecture and 40 bytes on 64-bit architecture. The userland
1380 applications "arp" and "ndp" have been updated accordingly.
1381 The output from "netstat -r" shows only routing entries and
1382 none of the L2 information.
1385 __FreeBSD_version 800057 marks the switchover from the
1386 binary ath hal to source code. Users must add the line:
1388 options AH_SUPPORT_AR5416
1390 to their kernel config files when specifying:
1394 The ath_hal module no longer exists; the code is now compiled
1395 together with the driver in the ath module. It is now
1396 possible to tailor chip support (i.e. reduce the set of chips
1397 and thereby the code size); consult ath_hal(4) for details.
1400 __FreeBSD_version 800054 adds memory barriers to
1401 <machine/atomic.h>, new interfaces to ifnet to facilitate
1402 multiple hardware transmit queues for cards that support
1403 them, and a lock-less ring-buffer implementation to
1404 enable drivers to more efficiently manage queueing of
1408 A new version of ZFS (version 13) has been merged to -HEAD.
1409 This version has zpool attribute "listsnapshots" off by
1410 default, which means "zfs list" does not show snapshots,
1411 and is the same as Solaris behavior.
1414 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1417 The uhci, ohci, ehci and slhci USB Host controller drivers have
1418 been put into separate modules. If you load the usb module
1419 separately through loader.conf you will need to load the
1420 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1421 controller add the following to loader.conf:
1427 The ABI used by the PMC toolset has changed. Please keep
1428 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1432 atapci kernel module now includes only generic PCI ATA
1433 driver. AHCI driver moved to ataahci kernel module.
1434 All vendor-specific code moved into separate kernel modules:
1435 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1436 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1437 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1438 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1441 The TTY subsystem of the kernel has been replaced by a new
1442 implementation, which provides better scalability and an
1443 improved driver model. Most common drivers have been migrated to
1444 the new TTY subsystem, while others have not. The following
1445 drivers have not yet been ported to the new TTY layer:
1448 cy, digi, rc, rp, sio
1454 ng_h4, ng_tty, ppp, sl, snp
1456 Adding these drivers to your kernel configuration file shall
1457 cause compilation to fail.
1460 ntpd has been upgraded to 4.2.4p5.
1463 OpenSSH has been upgraded to 5.1p1.
1465 For many years, FreeBSD's version of OpenSSH preferred DSA
1466 over RSA for host and user authentication keys. With this
1467 upgrade, we've switched to the vendor's default of RSA over
1468 DSA. This may cause upgraded clients to warn about unknown
1469 host keys even for previously known hosts. Users should
1470 follow the usual procedure for verifying host keys before
1471 accepting the RSA key.
1473 This can be circumvented by setting the "HostKeyAlgorithms"
1474 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1477 Please note that the sequence of keys offered for
1478 authentication has been changed as well. You may want to
1479 specify IdentityFile in a different order to revert this
1483 The sio(4) driver has been removed from the i386 and amd64
1484 kernel configuration files. This means uart(4) is now the
1485 default serial port driver on those platforms as well.
1487 To prevent collisions with the sio(4) driver, the uart(4) driver
1488 uses different names for its device nodes. This means the
1489 onboard serial port will now most likely be called "ttyu0"
1490 instead of "ttyd0". You may need to reconfigure applications to
1491 use the new device names.
1493 When using the serial port as a boot console, be sure to update
1494 /boot/device.hints and /etc/ttys before booting the new kernel.
1495 If you forget to do so, you can still manually specify the hints
1496 at the loader prompt:
1498 set hint.uart.0.at="isa"
1499 set hint.uart.0.port="0x3F8"
1500 set hint.uart.0.flags="0x10"
1501 set hint.uart.0.irq="4"
1505 The gpt(8) utility has been removed. Use gpart(8) to partition
1509 The version that Linuxulator emulates was changed from 2.4.2
1510 to 2.6.16. If you experience any problems with Linux binaries
1511 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1512 if it fixes the problem contact emulation mailing list.
1515 ISDN4BSD (I4B) was removed from the src tree. You may need to
1516 update a your kernel configuration and remove relevant entries.
1519 I have checked in code to support multiple routing tables.
1520 See the man pages setfib(1) and setfib(2).
1521 This is a hopefully backwards compatible version,
1522 but to make use of it you need to compile your kernel
1523 with options ROUTETABLES=2 (or more up to 16).
1526 The 802.11 wireless support was redone to enable multi-bss
1527 operation on devices that are capable. The underlying device
1528 is no longer used directly but instead wlanX devices are
1529 cloned with ifconfig. This requires changes to rc.conf files.
1530 For example, change:
1531 ifconfig_ath0="WPA DHCP"
1534 ifconfig_wlan0="WPA DHCP"
1535 see rc.conf(5) for more details. In addition, mergemaster of
1536 /etc/rc.d is highly recommended. Simultaneous update of userland
1537 and kernel wouldn't hurt either.
1539 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1540 modules were merged into the base wlan module. All references
1541 to these modules (e.g. in kernel config files) must be removed.
1544 psm(4) has gained write(2) support in native operation level.
1545 Arbitrary commands can be written to /dev/psm%d and status can
1546 be read back from it. Therefore, an application is responsible
1547 for status validation and error recovery. It is a no-op in
1548 other operation levels.
1551 Support for KSE threading has been removed from the kernel. To
1552 run legacy applications linked against KSE libmap.conf may
1553 be used. The following libmap.conf may be used to ensure
1554 compatibility with any prior release:
1556 libpthread.so.1 libthr.so.1
1557 libpthread.so.2 libthr.so.2
1558 libkse.so.3 libthr.so.3
1561 The layout of struct vmspace has changed. This affects libkvm
1562 and any executables that link against libkvm and use the
1563 kvm_getprocs() function. In particular, but not exclusively,
1564 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1565 The effects are minimal, but it's advisable to upgrade world
1569 The latest em driver no longer has support in it for the
1570 82575 adapter, this is now moved to the igb driver. The
1571 split was done to make new features that are incompatible
1572 with older hardware easier to do.
1575 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1576 likewise the kernel option is now GEOM_LINUX_LVM.
1579 The default NFS mount mode has changed from UDP to TCP for
1580 increased reliability. If you rely on (insecurely) NFS
1581 mounting across a firewall you may need to update your
1585 Belatedly note the addition of m_collapse for compacting
1589 The fts(3) structures have been changed to use adequate
1590 integer types for their members and so to be able to cope
1591 with huge file trees. The old fts(3) ABI is preserved
1592 through symbol versioning in libc, so third-party binaries
1593 using fts(3) should still work, although they will not take
1594 advantage of the extended types. At the same time, some
1595 third-party software might fail to build after this change
1596 due to unportable assumptions made in its source code about
1597 fts(3) structure members. Such software should be fixed
1598 by its vendor or, in the worst case, in the ports tree.
1599 FreeBSD_version 800015 marks this change for the unlikely
1600 case that a portable fix is impossible.
1603 To upgrade to -current after this date, you must be running
1604 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1605 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1608 The ADAPTIVE_GIANT kernel option has been retired because its
1609 functionality is the default now.
1612 The AT keyboard emulation of sunkbd(4) has been turned on
1613 by default. In order to make the special symbols of the Sun
1614 keyboards driven by sunkbd(4) work under X these now have
1615 to be configured the same way as Sun USB keyboards driven
1616 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1618 Option "XkbLayout" "us"
1619 Option "XkbRules" "xorg"
1620 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1623 It has been decided that it is desirable to provide ABI
1624 backwards compatibility to the FreeBSD 4/5/6 versions of the
1625 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1626 broken with the introduction of PCI domain support (see the
1627 20070930 entry). Unfortunately, this required the ABI of
1628 PCIOCGETCONF to be broken again in order to be able to
1629 provide backwards compatibility to the old version of that
1630 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1631 again. As for prominent ports this affects neither pciutils
1632 nor xorg-server this time, the hal port needs to be rebuilt
1636 The misnamed kthread_create() and friends have been renamed
1637 to kproc_create() etc. Many of the callers already
1638 used kproc_start()..
1639 I will return kthread_create() and friends in a while
1640 with implementations that actually create threads, not procs.
1641 Renaming corresponds with version 800002.
1650 Avoid using make -j when upgrading. While generally safe, there are
1651 sometimes problems using -j to upgrade. If your upgrade fails with
1652 -j, please try again without -j. From time to time in the past there
1653 have been problems using -j with buildworld and/or installworld. This
1654 is especially true when upgrading between "distant" versions (eg one
1655 that cross a major release boundary or several minor releases, or when
1656 several months have passed on the -current branch).
1658 Sometimes, obscure build problems are the result of environment
1659 poisoning. This can happen because the make utility reads its
1660 environment when searching for values for global variables. To run
1661 your build attempts in an "environmental clean room", prefix all make
1662 commands with 'env -i '. See the env(1) manual page for more details.
1664 When upgrading from one major version to another it is generally best
1665 to upgrade to the latest code in the currently installed branch first,
1666 then do an upgrade to the new branch. This is the best-tested upgrade
1667 path, and has the highest probability of being successful. Please try
1668 this approach before reporting problems with a major version upgrade.
1672 When upgrading the boot ZFS pool to a new version, always follow
1675 1.) recompile and reinstall the ZFS boot loader and boot block
1676 (this is part of "make buildworld" and "make installworld")
1678 2.) update the ZFS boot block on your boot drive
1680 The following example updates the ZFS boot block on the first
1681 partition (freebsd-boot) of a GPT partitioned drive ad0:
1682 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1684 Non-boot pools do not need these updates.
1688 If you are updating from a prior version of FreeBSD (even one just
1689 a few days old), you should follow this procedure. It is the most
1690 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1692 make kernel-toolchain
1693 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1694 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1696 To test a kernel once
1697 ---------------------
1698 If you just want to boot a kernel once (because you are not sure
1699 if it works, or if you want to boot a known bad kernel to provide
1700 debugging information) run
1701 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1702 nextboot -k testkernel
1704 To just build a kernel when you know that it won't mess you up
1705 --------------------------------------------------------------
1706 This assumes you are already running a CURRENT system. Replace
1707 ${arch} with the architecture of your machine (e.g. "i386",
1708 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1710 cd src/sys/${arch}/conf
1711 config KERNEL_NAME_HERE
1712 cd ../compile/KERNEL_NAME_HERE
1717 If this fails, go to the "To build a kernel" section.
1719 To rebuild everything and install it on the current system.
1720 -----------------------------------------------------------
1721 # Note: sometimes if you are running current you gotta do more than
1722 # is listed here if you are upgrading from a really old current.
1724 <make sure you have good level 0 dumps>
1726 make kernel KERNCONF=YOUR_KERNEL_HERE
1728 <reboot in single user> [3]
1736 To cross-install current onto a separate partition
1737 --------------------------------------------------
1738 # In this approach we use a separate partition to hold
1739 # current's root, 'usr', and 'var' directories. A partition
1740 # holding "/", "/usr" and "/var" should be about 2GB in
1743 <make sure you have good level 0 dumps>
1746 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1747 <maybe newfs current's root partition>
1748 <mount current's root partition on directory ${CURRENT_ROOT}>
1749 make installworld DESTDIR=${CURRENT_ROOT}
1750 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1751 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1752 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1753 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1754 <reboot into current>
1755 <do a "native" rebuild/install as described in the previous section>
1756 <maybe install compatibility libraries from ports/misc/compat*>
1760 To upgrade in-place from stable to current
1761 ----------------------------------------------
1762 <make sure you have good level 0 dumps>
1764 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1766 <reboot in single user> [3]
1773 Make sure that you've read the UPDATING file to understand the
1774 tweaks to various things you need. At this point in the life
1775 cycle of current, things change often and you are on your own
1776 to cope. The defaults can also change, so please read ALL of
1777 the UPDATING entries.
1779 Also, if you are tracking -current, you must be subscribed to
1780 freebsd-current@freebsd.org. Make sure that before you update
1781 your sources that you have read and understood all the recent
1782 messages there. If in doubt, please track -stable which has
1783 much fewer pitfalls.
1785 [1] If you have third party modules, such as vmware, you
1786 should disable them at this point so they don't crash your
1789 [3] From the bootblocks, boot -s, and then do
1794 adjkerntz -i # if CMOS is wall time
1795 Also, when doing a major release upgrade, it is required that
1796 you boot into single user mode to do the installworld.
1798 [4] Note: This step is non-optional. Failure to do this step
1799 can result in a significant reduction in the functionality of the
1800 system. Attempting to do it by hand is not recommended and those
1801 that pursue this avenue should read this file carefully, as well
1802 as the archives of freebsd-current and freebsd-hackers mailing lists
1803 for potential gotchas. The -U option is also useful to consider.
1804 See mergemaster(8) for more information.
1806 [5] Usually this step is a noop. However, from time to time
1807 you may need to do this if you get unknown user in the following
1808 step. It never hurts to do it all the time. You may need to
1809 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1810 install) after the buildworld before this step if you last updated
1811 from [78]-stable or 9-stable before 20130430.
1813 [6] This only deletes old files and directories. Old libraries
1814 can be deleted by "make delete-old-libs", but you have to make
1815 sure that no program is using those libraries anymore.
1817 [8] In order to have a kernel that can run the 4.x binaries needed to
1818 do an installworld, you must include the COMPAT_FREEBSD4 option in
1819 your kernel. Failure to do so may leave you with a system that is
1820 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1821 required to run the 5.x binaries on more recent kernels. And so on
1822 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1824 Make sure that you merge any new devices from GENERIC since the
1825 last time you updated your kernel config file.
1827 [9] When checking out sources, you must include the -P flag to have
1828 cvs prune empty directories.
1830 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1831 "?=" instead of the "=" assignment operator, so that buildworld can
1832 override the CPUTYPE if it needs to.
1834 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1835 not on the command line, or in /etc/make.conf. buildworld will
1836 warn if it is improperly defined.
1839 This file contains a list, in reverse chronological order, of major
1840 breakages in tracking -current. It is not guaranteed to be a complete
1841 list of such breakages, and only contains entries since October 10, 2007.
1842 If you need to see UPDATING entries from before that date, you will need
1843 to fetch an UPDATING file from an older FreeBSD release.
1845 Copyright information:
1847 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1849 Redistribution, publication, translation and use, with or without
1850 modification, in full or in part, in any form or format of this
1851 document are permitted without further permission from the author.
1853 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1854 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1855 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1856 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1857 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1858 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1859 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1860 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1861 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1862 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1863 POSSIBILITY OF SUCH DAMAGE.
1865 Contact Warner Losh if you have any questions about your use of