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 20140916: p2 FreeBSD-SA-14:19.tcp
15 Fix Denial of Service in TCP packet processing. [SA-14:19]
17 20140909: p1 FreeBSD-SA-14:18.openssl
18 Fix OpenSSL multiple vulnerabilities. [SA-14:18]
24 On i386 and amd64 systems, the onifconsole flag is now set by default
25 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
26 as a login TTY if it is set in the bootloader as an active kernel
27 console. No changes in behavior should result otherwise. To revert to
28 the previous behavior, set ttyu0 to "off" in /etc/ttys.
31 Clang and llvm have been upgraded to 3.4.1 release.
34 Clang and llvm have been upgraded to 3.4 release.
37 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
38 been deprecated and will not be part of FreeBSD 11.0 and later
39 releases. If you use this driver, please consider switching to
40 the nfe(4) driver instead.
43 The behavior of gss_pseudo_random() for the krb5 mechanism
44 has changed, for applications requesting a longer random string
45 than produced by the underlying enctype's pseudo-random() function.
46 In particular, the random string produced from a session key of
47 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
48 be different at the 17th octet and later, after this change.
49 The counter used in the PRF+ construction is now encoded as a
50 big-endian integer in accordance with RFC 4402.
51 __FreeBSD_version is bumped to 902505.
57 Behavior of devfs rules path matching has been changed.
58 Pattern is now always matched against fully qualified devfs
59 path and slash characters must be explicitly matched by
60 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
61 subdirectories must be reviewed.
64 hastctl(8)'s `status' command output changed to terse one-liner format.
65 Scripts using this should switch to `list' command or be rewritten.
68 Fix a bug that allowed a tracing process (e.g. gdb) to write
69 to a memory-mapped file in the traced process's address space
70 even if neither the traced process nor the tracing process had
71 write access to that file.
74 Added ZFS TRIM support which is enabled by default. To disable
75 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
77 Creating new ZFS pools and adding new devices to existing pools
78 first performs a full device level TRIM which can take a significant
79 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
80 to disable this behaviour.
82 ZFS TRIM requires the underlying device support BIO_DELETE which
83 is currently provided by methods such as ATA TRIM and SCSI UNMAP
84 via CAM, which are typically supported by SSD's.
86 Stats for ZFS TRIM can be monitored by looking at the sysctl's
87 under kstat.zfs.misc.zio_trim.
90 `list' command has been added to hastctl(8). For now, it is full
91 equivalent of `status' command.
92 WARNING: in the near future the output of hastctl's status command
93 will change to more terse format. If you use `hastctl status'
94 for parsing in your scripts, switch to `hastctl list'.
97 The mergemaster command now uses the default MAKEOBJDIRPREFIX
98 rather than creating it's own in the temporary directory in
99 order allow access to bootstrapped versions of tools such as
100 install and mtree. When upgrading from version of FreeBSD where
101 the install command does not support -l, you will need to
102 install a new mergemaster command if mergemaster -p is required.
103 This can be accomplished with the command (cd src/usr.sbin/mergemaster
106 Due to the use of the new -l option to install(1) during build
107 and install, you must take care not to directly set the INSTALL
108 make variable in your /etc/make.conf, /etc/src.conf, or on the
109 command line. If you wish to use the -C flag for all installs
110 you may be able to add INSTALL+=-C to /etc/make.conf or
114 Fix a bug that allows NFS clients to issue READDIR on files.
117 The install(1) option -M has changed meaning and now takes an
118 argument that is a file or path to append logs to. In the
119 unlikely event that -M was the last option on the command line
120 and the command line contained at least two files and a target
121 directory the first file will have logs appended to it. The -M
122 option served little practical purpose in the last decade so it's
123 used expected to be extremely rare.
126 A new compression method (lz4) has been merged to. Please refer to
127 zpool-features(7) for more information.
129 Please refer to the "ZFS notes" section of this file for information
130 on upgrading boot ZFS pools.
133 The VFS KBI was changed with the merge of several nullfs
134 optimizations and fixes. All filesystem modules must be
138 With the addition of auditdistd(8), a new auditdistd user is now
139 depended on during installworld. "mergemaster -p" can be used to add
140 the user prior to installworld, as documented in the handbook.
146 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
147 Starting with this version the old system of ZFS pool versioning
148 is superseded by "feature flags". This concept enables forward
149 compatibility against certain future changes in functionality of ZFS
150 pools. The first two read-only compatible "feature flags" for ZFS
151 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
152 For more information read the new zpool-features(7) manual page.
153 Please refer to the "ZFS notes" section of this file for information
154 on upgrading boot ZFS pools.
157 The commit introducing bsd.compiler.mk breaks the traditional
158 building of kernels before this point. Add -m ${SRC}/share/mk
159 (for the right value of SRC) to your command lines to work
160 around; update your useland to a point after this; or use the
161 buildkernel/installkernel top-level targets. See also 20120829.
164 The IPFIREWALL_FORWARD kernel option has been removed. Its
165 functionality now turned on by default.
168 The random(4) support for the VIA hardware random number
169 generator (`PADLOCK') is no longer enabled unconditionally.
170 Add the PADLOCK_RNG option in the custom kernel config if
171 needed. The GENERIC kernels on i386 and amd64 do include the
172 option, so the change only affects the custom kernel
176 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
177 the traditional method, you must update your system with an installworld
178 before the kernel will build. The documented make buildkernel/installkernel
179 interfaces (coupled with fresh make kernel-toolchain) continue to work.
182 The sparc64 ZFS loader has been changed to no longer try to auto-
183 detect ZFS providers based on diskN aliases but now requires these
184 to be explicitly listed in the OFW boot-device environment variable.
187 Now unix domain sockets behave "as expected" on nullfs(5). Previously
188 nullfs(5) did not pass through all behaviours to the underlying layer,
189 as a result if we bound to a socket on the lower layer we could connect
190 only to the lower path; if we bound to the upper layer we could connect
191 only to the upper path. The new behavior is one can connect to both the
192 lower and the upper paths regardless what layer path one binds to.
195 The acpi_wmi(4) status device /dev/wmistat has been renamed to
199 A new VOP_ADVISE() was added to support posix_fadvise(2). All
200 filesystem modules must be recompiled.
203 The interface of the VOP_VPTOCNP(9) changed, now the returned
204 vnode shall be referenced, previously it was required to be
205 only held. All in-tree filesystems are converted.
211 The broken amd(4) driver has been replaced with esp(4) in the amd64,
212 i386 and pc98 GENERIC kernel configuration files.
215 This commit modifies vfs_register() so that it uses a hash
216 calculation to set vfc_typenum, which is enabled by default.
217 The first time a system is booted after this change, the
218 vfc_typenum values will change for all file systems. The
219 main effect of this is a change to the NFS server file handles
220 for file systems that use vfc_typenum in their fsid, such as ZFS.
221 It will, however, prevent vfc_typenum from changing when file
222 systems are loaded in a different order for subsequent reboots.
223 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
224 until you are ready to remount all NFS clients after a reboot.
227 Bump the shared library version numbers for libraries that
228 do not use symbol versioning, have changed the ABI compared
229 to stable/8 and which shared library version was not bumped.
230 Done as part of 9.0-RELEASE cycle.
233 During the merge of Capsicum features, the fget(9) KPI was modified.
234 This may require the rebuilding of out-of-tree device drivers --
235 issues have been reported specifically with the nVidia device driver.
236 __FreeBSD_version is bumped to 900041.
238 Also, there is a period between 20110811 and 20110814 where the
239 special devices /dev/{stdin,stdout,stderr} did not work correctly.
240 Building world from a kernel during that window may not work.
243 The packet filter (pf) code has been updated to OpenBSD 4.5.
244 You need to update userland tools to be in sync with kernel.
245 This update breaks backward compatibility with earlier pfsync(4)
246 versions. Care must be taken when updating redundant firewall setups.
249 The following sysctls and tunables are retired on x86 platforms:
251 machdep.hlt_logical_cpus
252 The following sysctl is retired:
253 machdep.hyperthreading_allowed
254 The sysctls were supposed to provide a way to dynamically offline and
255 online selected CPUs on x86 platforms, but the implementation has not
256 been reliable especially with SCHED_ULE scheduler.
257 machdep.hyperthreading_allowed tunable is still available to ignore
258 hyperthreading CPUs at OS level.
259 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
260 where X is an APIC ID of a CPU. Be advised, though, that disabling
261 CPUs in non-uniform fashion will result in non-uniform topology and
262 may lead to sub-optimal system performance with SCHED_ULE, which is
266 cpumask_t type is retired and cpuset_t is used in order to describe
270 Changes to ifconfig(8) for dynamic address family detection mandate
271 that you are running a kernel of 20110525 or later. Make sure to
272 follow the update procedure to boot a new kernel before installing
276 Support for sun4v architecture is officially dropped
279 Several KPI breaking changes have been committed to the mii(4) layer,
280 the PHY drivers and consequently some Ethernet drivers using mii(4).
281 This means that miibus.ko and the modules of the affected Ethernet
282 drivers need to be recompiled.
284 Note to kernel developers: Given that the OUI bit reversion problem
285 was fixed as part of these changes all mii(4) commits related to OUIs,
286 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
287 handling, no longer can be merged verbatim to stable/8 and previous
291 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
292 into their kernel configurations along with 'device pci'.
295 The default NFS client is now the new NFS client, so fstype "newnfs"
296 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
297 Although mounts via fstype "nfs" will usually work without userland
298 changes, it is recommended that the mount(8) and mount_nfs(8)
299 commands be rebuilt from sources and that a link to mount_nfs called
300 mount_oldnfs be created. The new client is compiled into the
301 kernel with "options NFSCL" and this is needed for diskless root
302 file systems. The GENERIC kernel configs have been changed to use
303 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
304 To use the regular/old client, you can "mount -t oldnfs ...". For
305 a diskless root file system, you must also include a line like:
307 vfs.root.mountfrom="oldnfs:"
309 in the boot/loader.conf on the root fs on the NFS server to make
310 a diskless root fs use the old client.
313 The GENERIC kernels for all architectures now default to the new
314 CAM-based ATA stack. It means that all legacy ATA drivers were
315 removed and replaced by respective CAM drivers. If you are using
316 ATA device names in /etc/fstab or other places, make sure to update
317 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
318 where 'Y's are the sequential numbers starting from zero for each type
319 in order of detection, unless configured otherwise with tunables,
320 see cam(4)). There will be symbolic links created in /dev/ to map
321 old adX devices to the respective adaY. They should provide basic
322 compatibility for file systems mounting in most cases, but they do
323 not support old user-level APIs and do not have respective providers
324 in GEOM. Consider using updated management tools with new device names.
326 It is possible to load devices ahci, ata, siis and mvs as modules,
327 but option ATA_CAM should remain in kernel configuration to make ata
328 module work as CAM driver supporting legacy ATA controllers. Device ata
329 still can be used in modular fashion (atacore + ...). Modules atadisk
330 and atapi* are not used and won't affect operation in ATA_CAM mode.
331 Note that to use CAM-based ATA kernel should include CAM devices
332 scbus, pass, da (or explicitly ada), cd and optionally others. All of
333 them are parts of the cam module.
335 ataraid(4) functionality is now supported by the RAID GEOM class.
336 To use it you can load geom_raid kernel module and use graid(8) tool
337 for management. Instead of /dev/arX device names, use /dev/raid/rX.
339 No kernel config options or code have been removed, so if a problem
340 arises, please report it and optionally revert to the old ATA stack.
341 In order to do it you can remove from the kernel config:
346 , and instead add back:
347 device atadisk # ATA disk drives
348 device ataraid # ATA RAID drives
349 device atapicd # ATAPI CDROM drives
350 device atapifd # ATAPI floppy drives
351 device atapist # ATAPI tape drives
354 The default NFS server has been changed to the new server, which
355 was referred to as the experimental server. If you need to switch
356 back to the old NFS server, you must now put the "-o" option on
357 both the mountd and nfsd commands. This can be done using the
358 mountd_flags and nfs_server_flags rc.conf variables until an
359 update to the rc scripts is committed, which is coming soon.
362 The GNU Objective-C runtime library (libobjc), and other Objective-C
363 related components have been removed from the base system. If you
364 require an Objective-C library, please use one of the available ports.
367 ath(4) has been split into bus- and device- modules. if_ath contains
368 the HAL, the TX rate control and the network device code. if_ath_pci
369 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
370 contains the AHB glue. Users need to load both if_ath_pci and if_ath
371 in order to use ath on everything else.
373 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
374 need to load if_ath and if_ath_pci for ath(4) operation.
377 As part of the replacement of sysinstall, the process of building
378 release media has changed significantly. For details, please re-read
379 release(7), which has been updated to reflect the new build process.
382 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
383 is the last available version under GPLv2. It brings a number of new
384 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
385 4.1 and SSE 4.2), better support for powerpc64, a number of new
386 directives, and lots of other small improvements. See the ChangeLog
387 file in contrib/binutils for the full details.
390 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
391 compliant, and will now use half of hash for authentication.
392 This will break interoperability with all stacks (including all
393 actual FreeBSD versions) who implement
394 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
396 The only workaround with such peers is to use another HMAC
397 algorithm for IPsec ("phase 2") authentication.
400 Remove the uio_yield prototype and symbol. This function has
401 been misnamed since it was introduced and should not be
402 globally exposed with this name. The equivalent functionality
403 is now available using kern_yield(curthread->td_user_pri).
404 The function remains undocumented.
407 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
408 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
409 for scalar sysctls is defined but disabled. Code that needs
410 UQUAD to pass the type checking that must compile on older
411 systems where the define is not present can check against
412 __FreeBSD_version >= 900030.
414 The system dialog(1) has been replaced with a new version previously
415 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
416 with the previous version, but the libdialog associated with it has
417 a largely incompatible API. As such, the original version of libdialog
418 will be kept temporarily as libodialog, until its base system consumers
419 are replaced or updated. Bump __FreeBSD_version to 900030.
422 If you are trying to run make universe on a -stable system, and you get
423 the following warning:
424 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
425 unknown. config(8) likely too old."
426 or something similar to it, then you must upgrade your -stable system
427 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
428 stable/8) or build the config from the latest stable/8 branch and
429 install it on your system.
431 Prior to this date, building a current universe on 8-stable system from
432 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
433 error in the first kernel build phase. A new config on those old
434 systems will fix that problem for older versions of -current.
437 The TCP stack has been modified to allow Khelp modules to interact with
438 it via helper hook points and store per-connection data in the TCP
439 control block. Bump __FreeBSD_version to 900029. User space tools that
440 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
444 Generic IEEE 802.3 annex 31B full duplex flow control support has been
445 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
446 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
447 to take advantage of it instead of using custom implementations. This
448 means that these drivers now no longer unconditionally advertise
449 support for flow control but only do so if flow control is a selected
450 media option. This was implemented in the generic support that way in
451 order to allow flow control to be switched on and off via ifconfig(8)
452 with the PHY specific default to typically off in order to protect
453 from unwanted effects. Consequently, if you used flow control with
454 one of the above mentioned drivers you now need to explicitly enable
456 ifconfig bge0 media auto mediaopt flowcontrol
458 Along with the above mentioned changes generic support for setting
459 1000baseT master mode also has been added and brgphy(4), ciphy(4),
460 e1000phy(4) as well as ip1000phy(4) have been converted to take
461 advantage of it. This means that these drivers now no longer take the
462 link0 parameter for selecting master mode but the master media option
463 has to be used instead, for example like in the following:
464 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
466 Selection of master mode now is also available with all other PHY
467 drivers supporting 1000baseT.
470 The TCP stack has received a significant update to add support for
471 modularised congestion control and generally improve the clarity of
472 congestion control decisions. Bump __FreeBSD_version to 900025. User
473 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
474 sockstat) need to be recompiled.
477 The man(1) utility has been replaced by a new version that no longer
478 uses /etc/manpath.config. Please consult man.conf(5) for how to
479 migrate local entries to the new format.
482 The copyright strings printed by login(1) and sshd(8) at the time of a
483 new connection have been removed to follow other operating systems and
487 A workaround for a fixed ld bug has been removed in kernel code,
488 so make sure that your system ld is built from sources after
489 revision 210245 from 2010-07-19 (r211583 if building head kernel
490 on stable/8, r211584 for stable/7; both from 2010-08-21).
491 A symptom of incorrect ld version is different addresses for
492 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
495 The $ipv6_prefer variable in rc.conf(5) has been split into
496 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
498 The $ip6addrctl_policy is a variable to choose a pre-defined
499 address selection policy set by ip6addrctl(8). A value
500 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
503 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
504 flag (see an entry of 20090926) is set on an interface with no
505 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
506 security reason. If you want IPv6 link-local address on all
507 interfaces by default, set this to "YES".
509 The old ipv6_prefer="YES" is equivalent to
510 ipv6_activate_all_interfaces="YES" and
511 ip6addrctl_policy="ipv6_prefer".
514 DTrace has grown support for userland tracing. Due to this, DTrace is
515 now i386 and amd64 only.
516 dtruss(1) is now installed by default on those systems and a new
517 kernel module is needed for userland tracing: fasttrap.
518 No changes to your kernel config file are necessary to enable
519 userland tracing, but you might consider adding 'STRIP=' and
520 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
521 to have informative userland stack traces in DTrace (ustack).
524 The acpi_aiboost(4) driver has been removed in favor of the new
525 aibs(4) driver. You should update your kernel configuration file.
528 BSD grep has been imported to the base system and it is built by
529 default. It is completely BSD licensed, highly GNU-compatible, uses
530 less memory than its GNU counterpart and has a small codebase.
531 However, it is slower than its GNU counterpart, which is mostly
532 noticeable for larger searches, for smaller ones it is measurable
533 but not significant. The reason is complex, the most important factor
534 is that we lack a modern and efficient regex library and GNU
535 overcomes this by optimizing the searches internally. Future work
536 on improving the regex performance is planned, for the meantime,
537 users that need better performance, can build GNU grep instead by
538 setting the WITH_GNU_GREP knob.
541 Due to the import of powerpc64 support, all existing powerpc kernel
542 configuration files must be updated with a machine directive like this:
543 machine powerpc powerpc
545 In addition, an updated config(8) is required to build powerpc kernels
549 A new version of ZFS (version 15) has been merged to -HEAD.
550 This version uses a python library for the following subcommands:
551 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
552 For full functionality of these commands the following port must
553 be installed: sysutils/py-zfs
556 'vm_page's are now hashed by physical address to an array of mutexes.
557 Currently this is only used to serialize access to hold_count. Over
558 time the page queue mutex will be peeled away. This changes the size
559 of pmap on every architecture. And requires all callers of vm_page_hold
560 and vm_page_unhold to be updated.
563 WITH_CTF can now be specified in src.conf (not recommended, there
564 are some problems with static executables), make.conf (would also
565 affect ports which do not use GNU make and do not override the
566 compile targets) or in the kernel config (via "makeoptions
568 When WITH_CTF was specified there before this was silently ignored,
569 so make sure that WITH_CTF is not used in places which could lead
570 to unwanted behavior.
573 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
574 to allow 32-bit compatibility on non-x86 platforms. All kernel
575 configurations on amd64 and ia64 platforms using these options must
576 be modified accordingly.
579 The utmp user accounting database has been replaced with utmpx,
580 the user accounting interface standardized by POSIX.
581 Unfortunately the semantics of utmp and utmpx don't match,
582 making it practically impossible to support both interfaces.
583 The user accounting database is used by tools like finger(1),
584 last(1), talk(1), w(1) and ac(8).
586 All applications in the base system use utmpx. This means only
587 local binaries (e.g. from the ports tree) may still use these
588 utmp database files. These applications must be rebuilt to make
591 After the system has been upgraded, it is safe to remove the old
592 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
593 assuming their contents is of no importance anymore. Old wtmp
594 databases can only be used by last(1) and ac(8) after they have
595 been converted to the new format using wtmpcvt(1).
598 Introduce the kernel thread "deadlock resolver" (which can be enabled
599 via the DEADLKRES option, see NOTES for more details) and the
600 sleepq_type() function for sleepqueues.
603 The rc.firewall and rc.firewall6 were unified, and
604 rc.firewall6 and rc.d/ip6fw were removed.
605 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
606 variables are obsoleted. Instead, the following new rc
607 variables are added to rc.d/ipfw:
609 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
610 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
611 firewall_simple_onet_ipv6, firewall_trusted_ipv6
613 The meanings correspond to the relevant IPv4 variables.
619 The default terminal emulation for syscons(4) has been changed
620 from cons25 to xterm on all platforms except pc98. This means
621 that the /etc/ttys file needs to be updated to ensure correct
622 operation of applications on the console.
624 The terminal emulation style can be toggled per window by using
625 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
626 options can be used to change the compile-time default back to
629 To prevent graphical artifacts, make sure the TERM environment
630 variable is set to match the terminal emulation that is being
631 performed by syscons(4).
634 The layout of the structure ieee80211req_scan_result has changed.
635 Applications that require wireless scan results (e.g. ifconfig(8))
636 from net80211 need to be recompiled.
638 Applications such as wpa_supplicant(8) may require a full world
639 build without using NO_CLEAN in order to get synchronized with the
643 The iwn(4) driver has been updated to support the 5000 and 5150 series.
644 There's one kernel module for each firmware. Adding "device iwnfw"
645 to the kernel configuration file means including all three firmware
646 images inside the kernel. If you want to include just the one for
647 your wireless card, use the devices iwn4965fw, iwn5000fw or
651 The rc.d/network_ipv6, IPv6 configuration script has been integrated
652 into rc.d/netif. The changes are the following:
654 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
655 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
656 Note that both variables need the "inet6" keyword at the head.
658 Do not set $ipv6_network_interfaces manually if you do not
659 understand what you are doing. It is not needed in most cases.
661 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
664 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
665 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
667 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
668 all configured interfaces have "inet6 accept_rtadv" in the
669 $ifconfig_IF_ipv6. These are for backward compatibility.
671 3. A new variable $ipv6_prefer has been added. If NO, IPv6
672 functionality of interfaces with no corresponding
673 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
674 and the default address selection policy of ip6addrctl(8)
675 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
676 Note that if you want to configure IPv6 functionality on the
677 disabled interfaces after boot, first you need to clear the flag by
678 using ifconfig(8) like:
680 ifconfig em0 inet6 -ifdisabled
682 If YES, the default address selection policy is set as
685 The default value of $ipv6_prefer is NO.
687 4. If your system need to receive Router Advertisement messages,
688 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
689 scripts automatically invoke rtsol(8) when the interface becomes
690 UP. The Router Advertisement messages are used for SLAAC
691 (State-Less Address AutoConfiguration).
694 802.11s D3.03 support was committed. This is incompatible with the
695 previous code, which was based on D3.0.
698 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
699 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
700 control whether accepting Router Advertisement messages or not.
701 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
702 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
703 The ifconfig(8) utility now supports these flags.
706 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
707 mount(8) and -a option for df(1) to see them.
710 The old tunable hw.bus.devctl_disable has been superseded by
711 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
712 replaced by hw.bus.devctl_queue=0. The default for this new tunable
716 Remove the option STOP_NMI. The default action is now to use NMI only
717 for KDB via the newly introduced function stop_cpus_hard() and
718 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
721 The stable/8 branch created in subversion. This corresponds to the
722 RELENG_8 branch in CVS.
725 Bump the shared library version numbers for all libraries that do not
726 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
727 __FreeBSD_version to 800105.
730 Due to changes in the implementation of virtual network stack support,
731 all network-related kernel modules must be recompiled. As this change
732 breaks the ABI, bump __FreeBSD_version to 800104.
735 The TOE interface to the TCP syncache has been modified to remove
736 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
737 The cxgb driver is the only TOE consumer affected by this change, and
738 needs to be recompiled along with the kernel. As this change breaks
739 the ABI, bump __FreeBSD_version to 800103.
742 Padding has been added to struct tcpcb, sackhint and tcpstat in
743 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
744 maintaining the ABI. However, this change breaks the ABI, so bump
745 __FreeBSD_version to 800102. User space tools that rely on the size of
746 any of these structs (e.g. sockstat) need to be recompiled.
749 The NFS_LEGACYRPC option has been removed along with the old kernel
750 RPC implementation that this option selected. Kernel configurations
751 may need to be adjusted.
754 The network interface device nodes at /dev/net/<interface> have been
755 removed. All ioctl operations can be performed the normal way using
756 routing sockets. The kqueue functionality can generally be replaced
757 with routing sockets.
760 The documentation from the FreeBSD Documentation Project (Handbook,
761 FAQ, etc.) is now installed via packages by sysinstall(8) and under
762 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
765 The ABI of various structures related to the SYSV IPC API have been
766 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
767 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
771 Layout of struct vnet has changed as routing related variables were
772 moved to their own Vimage module. Modules need to be recompiled. Bump
773 __FreeBSD_version to 800099.
776 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
777 respectively. As long as no more than 16 groups per process are used,
778 no changes should be visible. When more than 16 groups are used, old
779 binaries may fail if they call getgroups() or getgrouplist() with
780 statically sized storage. Recompiling will work around this, but
781 applications should be modified to use dynamically allocated storage
782 for group arrays as POSIX.1-2008 does not cap an implementation's
783 number of supported groups at NGROUPS_MAX+1 as previous versions did.
785 NFS and portalfs mounts may also be affected as the list of groups is
786 truncated to 16. Users of NFS who use more than 16 groups, should
787 take care that negative group permissions are not used on the exported
788 file systems as they will not be reliable unless a GSSAPI based
789 authentication method is used.
792 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
793 option compiles in the support for adaptive spinning for lockmgrs
794 which want to enable it. The lockinit() function now accepts the flag
795 LK_ADAPTIVE in order to make the lock object subject to adaptive
796 spinning when both held in write and read mode.
799 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
800 changed. User applications that use this ioctl need to be rebuilt.
803 The layout of struct thread has changed. Kernel and modules need to
807 The layout of structs ifnet, domain, protosw and vnet_net has changed.
808 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
811 window(1) has been removed from the base system. It can now be
812 installed from ports. The port is called misc/window.
815 The way we are storing and accessing `routing table' entries has
816 changed. Programs reading the FIB, like netstat, need to be
820 A new netisr implementation has been added for FreeBSD 8. Network
821 file system modules, such as igmp, ipdivert, and others, should be
823 Bump __FreeBSD_version to 800096.
826 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
830 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
831 Bump __FreeBSD_version to 800094.
834 Add mnt_xflag field to 'struct mount'. File system modules need to be
836 Bump __FreeBSD_version to 800093.
839 The compiling option ADAPTIVE_SX has been retired while it has been
840 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
841 The KPI for sx_init_flags() changes as accepting flags:
842 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
843 been introduced in order to handle the reversed logic.
844 Bump __FreeBSD_version to 800092.
847 Add support for hierarchical jails. Remove global securelevel.
848 Bump __FreeBSD_version to 800091.
851 The layout of struct vnet_net has changed, therefore modules
853 Bump __FreeBSD_version to 800090.
856 The newly imported zic(8) produces a new format in the output. Please
857 run tzsetup(8) to install the newly created data to /etc/localtime.
860 The sysctl tree for the usb stack has renamed from hw.usb2.* to
861 hw.usb.* and is now consistent again with previous releases.
864 802.11 monitor mode support was revised and driver api's were changed.
865 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
866 of DLT_IEEE802_11. No user-visible data structures were changed but
867 applications that use DLT_IEEE802_11 may require changes.
868 Bump __FreeBSD_version to 800088.
871 The layout of the following structs has changed: sysctl_oid,
872 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
873 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
874 panics may be experienced. World rebuild is required for
875 correctly checking networking state from userland.
876 Bump __FreeBSD_version to 800085.
879 MLDv2 and Source-Specific Multicast (SSM) have been merged
880 to the IPv6 stack. VIMAGE hooks are in but not yet used.
881 The implementation of SSM within FreeBSD's IPv6 stack closely
882 follows the IPv4 implementation.
884 For kernel developers:
886 * The most important changes are that the ip6_output() and
887 ip6_input() paths no longer take the IN6_MULTI_LOCK,
888 and this lock has been downgraded to a non-recursive mutex.
890 * As with the changes to the IPv4 stack to support SSM, filtering
891 of inbound multicast traffic must now be performed by transport
892 protocols within the IPv6 stack. This does not apply to TCP and
893 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
895 * The KPIs used by IPv6 multicast are similar to those used by
896 the IPv4 stack, with the following differences:
897 * im6o_mc_filter() is analogous to imo_multicast_filter().
898 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
899 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
900 * IN6_LOOKUP_MULTI() has been deprecated and removed.
901 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
902 for MLDv1 have an additional 'timer' argument which is used to
903 jitter the initial membership report for the solicited-node
904 multicast membership on-link.
905 * This is not strictly needed for MLDv2, which already jitters
906 its report transmissions. However, the 'timer' argument is
907 preserved in case MLDv1 is active on the interface.
909 * The KAME linked-list based IPv6 membership implementation has
910 been refactored to use a vector similar to that used by the IPv4
912 Code which maintains a list of its own multicast memberships
913 internally, e.g. carp, has been updated to reflect the new
916 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
917 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
918 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
919 implementation constraint which needs to be addressed in HEAD.
921 For application developers:
923 * The changes are broadly similar to those made for the IPv4
926 * The use of IPv4 and IPv6 multicast socket options on the same
927 socket, using mapped addresses, HAS NOT been tested or supported.
929 * There are a number of issues with the implementation of various
930 IPv6 multicast APIs which need to be resolved in the API surface
931 before the implementation is fully compatible with KAME userland
932 use, and these are mostly to do with interface index treatment.
934 * The literature available discusses the use of either the delta / ASM
935 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
936 using setsourcefilter(3)/getsourcefilter(3). For more information
937 please refer to RFC 3768, 'Socket Interface Extensions for
938 Multicast Source Filters'.
940 * Applications which use the published RFC 3678 APIs should be fine.
942 For systems administrators:
944 * The mtest(8) utility has been refactored to support IPv6, in
945 addition to IPv4. Interface addresses are no longer accepted
946 as arguments, their names must be used instead. The utility
947 will map the interface name to its first IPv4 address as
948 returned by getifaddrs(3).
950 * The ifmcstat(8) utility has also been updated to print the MLDv2
951 endpoint state and source filter lists via sysctl(3).
953 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
954 loopback of IPv6 multicast datagrams by default; it defaults to 1
955 to preserve the existing behaviour. Disabling multicast loopback is
956 recommended for optimal system performance.
958 * The IPv6 MROUTING code has been changed to examine this sysctl
959 instead of attempting to perform a group lookup before looping
960 back forwarded datagrams.
962 Bump __FreeBSD_version to 800084.
965 Implement low-level Bluetooth HCI API.
966 Bump __FreeBSD_version to 800083.
969 The layout of struct malloc_type, used by modules to register new
970 memory allocation types, has changed. Most modules will need to
971 be rebuilt or panics may be experienced.
972 Bump __FreeBSD_version to 800081.
975 Anticipate overflowing inp_flags - add inp_flags2.
976 This changes most offsets in inpcb, so checking v4 connection
977 state will require a world rebuild.
978 Bump __FreeBSD_version to 800080.
981 Add an llentry to struct route and struct route_in6. Modules
982 embedding a struct route will need to be recompiled.
983 Bump __FreeBSD_version to 800079.
986 The size of rt_metrics_lite and by extension rtentry has changed.
987 Networking administration apps will need to be recompiled.
988 The route command now supports show as an alias for get, weighting
989 of routes, sticky and nostick flags to alter the behavior of stateful
991 Bump __FreeBSD_version to 800078.
994 Do not use Giant for kbdmux(4) locking. This is wrong and
995 apparently causing more problems than it solves. This will
996 re-open the issue where interrupt handlers may race with
997 kbdmux(4) in polling mode. Typical symptoms include (but
998 not limited to) duplicated and/or missing characters when
999 low level console functions (such as gets) are used while
1000 interrupts are enabled (for example geli password prompt,
1001 mountroot prompt etc.). Disabling kbdmux(4) may help.
1004 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1005 kernel modules referencing any of the above need to be recompiled.
1006 Bump __FreeBSD_version to 800075.
1009 GEOM_PART has become the default partition slicer for storage devices,
1010 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1011 introduces some changes:
1013 MSDOS/EBR: the devices created from MSDOS extended partition entries
1014 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1015 to devices with offset-based names. fstabs may need to be modified.
1017 BSD: the "geometry does not match label" warning is harmless in most
1018 cases but it points to problems in file system misalignment with
1019 disk geometry. The "c" partition is now implicit, covers the whole
1020 top-level drive and cannot be (mis)used by users.
1022 General: Kernel dumps are now not allowed to be written to devices
1023 whose partition types indicate they are meant to be used for file
1024 systems (or, in case of MSDOS partitions, as something else than
1027 Most of these changes date approximately from 200812.
1030 The uscanner(4) driver has been removed from the kernel. This follows
1031 Linux removing theirs in 2.6 and making libusb the default interface
1032 (supported by sane).
1035 The multicast forwarding code has been cleaned up. netstat(1)
1036 only relies on KVM now for printing bandwidth upcall meters.
1037 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1038 ip6_mroute_mod respectively. The config(5) options for statically
1039 compiling this code remain the same, i.e. 'options MROUTING'.
1042 Support for the IFF_NEEDSGIANT network interface flag has been
1043 removed, which means that non-MPSAFE network device drivers are no
1044 longer supported. In particular, if_ar, if_sr, and network device
1045 drivers from the old (legacy) USB stack can no longer be built or
1049 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1050 a bunch of new language catalog files have also been added.
1051 This means that some common libc messages are now localized and
1052 they depend on the LC_MESSAGES environmental variable.
1055 The k8temp(4) driver has been renamed to amdtemp(4) since
1056 support for Family 10 and Family 11 CPU families was added.
1059 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1060 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1062 For kernel developers, the most important changes are that the
1063 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1064 and this lock has been downgraded to a non-recursive mutex.
1066 Transport protocols (UDP, Raw IP) are now responsible for filtering
1067 inbound multicast traffic according to group membership and source
1068 filters. The imo_multicast_filter() KPI exists for this purpose.
1069 Transports which do not use multicast (SCTP, TCP) already reject
1070 multicast by default. Forwarding and receive performance may improve
1071 as a mutex acquisition is no longer needed in the ip_input()
1072 low-level input path. in_addmulti() and in_delmulti() are shimmed
1073 to new KPIs which exist to support SSM in-kernel.
1075 For application developers, it is recommended that loopback of
1076 multicast datagrams be disabled for best performance, as this
1077 will still cause the lock to be taken for each looped-back
1078 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1079 be tuned to 0 to disable loopback by default; it defaults to 1
1080 to preserve the existing behaviour.
1082 For systems administrators, to obtain best performance with
1083 multicast reception and multiple groups, it is always recommended
1084 that a card with a suitably precise hash filter is used. Hash
1085 collisions will still result in the lock being taken within the
1086 transport protocol input path to check group membership.
1088 If deploying FreeBSD in an environment with IGMP snooping switches,
1089 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1090 enabled; this forces 224.0.0.0/24 group membership to be announced
1093 The size of 'struct igmpstat' has changed; netstat needs to be
1094 recompiled to reflect this.
1095 Bump __FreeBSD_version to 800070.
1098 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1099 updated to use it. This requires a buildworld/installworld in order to
1100 update the library and dependencies (usbconfig, etc). Its advisable to
1101 rebuild all ports which uses libusb. More specific directions are given
1102 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1103 libusb are no longer required and can be removed.
1106 A workaround is committed to allow the creation of System V shared
1107 memory segment of size > 2 GB on the 64-bit architectures.
1108 Due to a limitation of the existing ABI, the shm_segsz member
1109 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1110 wrong for large segments. Note that limits must be explicitly
1111 raised to allow such segments to be created.
1114 The layout of struct ifnet has changed, requiring a rebuild of all
1115 network device driver modules.
1118 The /dev handling for the new USB stack has changed, a
1119 buildworld/installworld is required for libusb20.
1122 The new USB2 stack has now been permanently moved in and all kernel and
1123 module names reverted to their previous values (eg, usb, ehci, ohci,
1124 ums, ...). The old usb stack can be compiled in by prefixing the name
1125 with the letter 'o', the old usb modules have been removed.
1126 Updating entry 20090216 for xorg and 20090215 for libmap may still
1130 The rc.conf(5) option if_up_delay has been renamed to
1131 defaultroute_delay to better reflect its purpose. If you have
1132 customized this setting in /etc/rc.conf you need to update it to
1136 xorg 7.4 wants to configure its input devices via hald which does not
1137 yet work with USB2. If the keyboard/mouse does not work in xorg then
1139 Option "AllowEmptyInput" "off"
1140 to your ServerLayout section. This will cause X to use the configured
1141 kbd and mouse sections from your xorg.conf.
1144 The GENERIC kernels for all architectures now default to the new USB2
1145 stack. No kernel config options or code have been removed so if a
1146 problem arises please report it and optionally revert to the old USB
1147 stack. If you are loading USB kernel modules or have a custom kernel
1148 that includes GENERIC then ensure that usb names are also changed over,
1149 eg uftdi -> usb2_serial_ftdi.
1151 Older programs linked against the ports libusb 0.1 need to be
1152 redirected to the new stack's libusb20. /etc/libmap.conf can
1154 # Map old usb library to new one for usb2 stack
1155 libusb-0.1.so.8 libusb20.so.1
1158 All USB ethernet devices now attach as interfaces under the name ueN
1159 (eg. ue0). This is to provide a predictable name as vendors often
1160 change usb chipsets in a product without notice.
1163 The ichsmb(4) driver has been changed to require SMBus slave
1164 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1165 All of the other SMBus controller drivers require left-justified
1166 slave addresses, so this change makes all the drivers provide the
1170 INET6 statistics (struct ip6stat) was updated.
1171 netstat(1) needs to be recompiled.
1174 NTFS has been removed from GENERIC kernel on amd64 to match
1175 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1176 will load ntfs.ko module automatically when NTFS support is
1177 actually needed, unless ntfs.ko is not installed or security
1178 level prohibits loading kernel modules. If either is the case,
1179 "options NTFS" has to be added into kernel config.
1182 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1183 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1184 800061. User space tools that rely on the size of struct tcpcb in
1185 tcp_var.h (e.g. sockstat) need to be recompiled.
1188 ng_tty(4) module updated to match the new TTY subsystem.
1189 Due to API change, user-level applications must be updated.
1190 New API support added to mpd5 CVS and expected to be present
1191 in next mpd5.3 release.
1194 With __FreeBSD_version 800060 the makefs tool is part of
1195 the base system (it was a port).
1198 The afdata and ifnet locks have been changed from mutexes to
1199 rwlocks, network modules will need to be re-compiled.
1202 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1203 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1204 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1205 architecture and 40 bytes on 64-bit architecture. The userland
1206 applications "arp" and "ndp" have been updated accordingly.
1207 The output from "netstat -r" shows only routing entries and
1208 none of the L2 information.
1211 __FreeBSD_version 800057 marks the switchover from the
1212 binary ath hal to source code. Users must add the line:
1214 options AH_SUPPORT_AR5416
1216 to their kernel config files when specifying:
1220 The ath_hal module no longer exists; the code is now compiled
1221 together with the driver in the ath module. It is now
1222 possible to tailor chip support (i.e. reduce the set of chips
1223 and thereby the code size); consult ath_hal(4) for details.
1226 __FreeBSD_version 800054 adds memory barriers to
1227 <machine/atomic.h>, new interfaces to ifnet to facilitate
1228 multiple hardware transmit queues for cards that support
1229 them, and a lock-less ring-buffer implementation to
1230 enable drivers to more efficiently manage queueing of
1234 A new version of ZFS (version 13) has been merged to -HEAD.
1235 This version has zpool attribute "listsnapshots" off by
1236 default, which means "zfs list" does not show snapshots,
1237 and is the same as Solaris behavior.
1240 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1243 The uhci, ohci, ehci and slhci USB Host controller drivers have
1244 been put into separate modules. If you load the usb module
1245 separately through loader.conf you will need to load the
1246 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1247 controller add the following to loader.conf:
1253 The ABI used by the PMC toolset has changed. Please keep
1254 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1258 atapci kernel module now includes only generic PCI ATA
1259 driver. AHCI driver moved to ataahci kernel module.
1260 All vendor-specific code moved into separate kernel modules:
1261 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1262 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1263 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1264 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1267 The TTY subsystem of the kernel has been replaced by a new
1268 implementation, which provides better scalability and an
1269 improved driver model. Most common drivers have been migrated to
1270 the new TTY subsystem, while others have not. The following
1271 drivers have not yet been ported to the new TTY layer:
1274 cy, digi, rc, rp, sio
1280 ng_h4, ng_tty, ppp, sl, snp
1282 Adding these drivers to your kernel configuration file shall
1283 cause compilation to fail.
1286 ntpd has been upgraded to 4.2.4p5.
1289 OpenSSH has been upgraded to 5.1p1.
1291 For many years, FreeBSD's version of OpenSSH preferred DSA
1292 over RSA for host and user authentication keys. With this
1293 upgrade, we've switched to the vendor's default of RSA over
1294 DSA. This may cause upgraded clients to warn about unknown
1295 host keys even for previously known hosts. Users should
1296 follow the usual procedure for verifying host keys before
1297 accepting the RSA key.
1299 This can be circumvented by setting the "HostKeyAlgorithms"
1300 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1303 Please note that the sequence of keys offered for
1304 authentication has been changed as well. You may want to
1305 specify IdentityFile in a different order to revert this
1309 The sio(4) driver has been removed from the i386 and amd64
1310 kernel configuration files. This means uart(4) is now the
1311 default serial port driver on those platforms as well.
1313 To prevent collisions with the sio(4) driver, the uart(4) driver
1314 uses different names for its device nodes. This means the
1315 onboard serial port will now most likely be called "ttyu0"
1316 instead of "ttyd0". You may need to reconfigure applications to
1317 use the new device names.
1319 When using the serial port as a boot console, be sure to update
1320 /boot/device.hints and /etc/ttys before booting the new kernel.
1321 If you forget to do so, you can still manually specify the hints
1322 at the loader prompt:
1324 set hint.uart.0.at="isa"
1325 set hint.uart.0.port="0x3F8"
1326 set hint.uart.0.flags="0x10"
1327 set hint.uart.0.irq="4"
1331 The gpt(8) utility has been removed. Use gpart(8) to partition
1335 The version that Linuxulator emulates was changed from 2.4.2
1336 to 2.6.16. If you experience any problems with Linux binaries
1337 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1338 if it fixes the problem contact emulation mailing list.
1341 ISDN4BSD (I4B) was removed from the src tree. You may need to
1342 update a your kernel configuration and remove relevant entries.
1345 I have checked in code to support multiple routing tables.
1346 See the man pages setfib(1) and setfib(2).
1347 This is a hopefully backwards compatible version,
1348 but to make use of it you need to compile your kernel
1349 with options ROUTETABLES=2 (or more up to 16).
1352 The 802.11 wireless support was redone to enable multi-bss
1353 operation on devices that are capable. The underlying device
1354 is no longer used directly but instead wlanX devices are
1355 cloned with ifconfig. This requires changes to rc.conf files.
1356 For example, change:
1357 ifconfig_ath0="WPA DHCP"
1360 ifconfig_wlan0="WPA DHCP"
1361 see rc.conf(5) for more details. In addition, mergemaster of
1362 /etc/rc.d is highly recommended. Simultaneous update of userland
1363 and kernel wouldn't hurt either.
1365 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1366 modules were merged into the base wlan module. All references
1367 to these modules (e.g. in kernel config files) must be removed.
1370 psm(4) has gained write(2) support in native operation level.
1371 Arbitrary commands can be written to /dev/psm%d and status can
1372 be read back from it. Therefore, an application is responsible
1373 for status validation and error recovery. It is a no-op in
1374 other operation levels.
1377 Support for KSE threading has been removed from the kernel. To
1378 run legacy applications linked against KSE libmap.conf may
1379 be used. The following libmap.conf may be used to ensure
1380 compatibility with any prior release:
1382 libpthread.so.1 libthr.so.1
1383 libpthread.so.2 libthr.so.2
1384 libkse.so.3 libthr.so.3
1387 The layout of struct vmspace has changed. This affects libkvm
1388 and any executables that link against libkvm and use the
1389 kvm_getprocs() function. In particular, but not exclusively,
1390 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1391 The effects are minimal, but it's advisable to upgrade world
1395 The latest em driver no longer has support in it for the
1396 82575 adapter, this is now moved to the igb driver. The
1397 split was done to make new features that are incompatible
1398 with older hardware easier to do.
1401 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1402 likewise the kernel option is now GEOM_LINUX_LVM.
1405 The default NFS mount mode has changed from UDP to TCP for
1406 increased reliability. If you rely on (insecurely) NFS
1407 mounting across a firewall you may need to update your
1411 Belatedly note the addition of m_collapse for compacting
1415 The fts(3) structures have been changed to use adequate
1416 integer types for their members and so to be able to cope
1417 with huge file trees. The old fts(3) ABI is preserved
1418 through symbol versioning in libc, so third-party binaries
1419 using fts(3) should still work, although they will not take
1420 advantage of the extended types. At the same time, some
1421 third-party software might fail to build after this change
1422 due to unportable assumptions made in its source code about
1423 fts(3) structure members. Such software should be fixed
1424 by its vendor or, in the worst case, in the ports tree.
1425 FreeBSD_version 800015 marks this change for the unlikely
1426 case that a portable fix is impossible.
1429 To upgrade to -current after this date, you must be running
1430 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1431 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1434 The ADAPTIVE_GIANT kernel option has been retired because its
1435 functionality is the default now.
1438 The AT keyboard emulation of sunkbd(4) has been turned on
1439 by default. In order to make the special symbols of the Sun
1440 keyboards driven by sunkbd(4) work under X these now have
1441 to be configured the same way as Sun USB keyboards driven
1442 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1444 Option "XkbLayout" "us"
1445 Option "XkbRules" "xorg"
1446 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1449 It has been decided that it is desirable to provide ABI
1450 backwards compatibility to the FreeBSD 4/5/6 versions of the
1451 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1452 broken with the introduction of PCI domain support (see the
1453 20070930 entry). Unfortunately, this required the ABI of
1454 PCIOCGETCONF to be broken again in order to be able to
1455 provide backwards compatibility to the old version of that
1456 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1457 again. As for prominent ports this affects neither pciutils
1458 nor xorg-server this time, the hal port needs to be rebuilt
1462 The misnamed kthread_create() and friends have been renamed
1463 to kproc_create() etc. Many of the callers already
1464 used kproc_start()..
1465 I will return kthread_create() and friends in a while
1466 with implementations that actually create threads, not procs.
1467 Renaming corresponds with version 800002.
1476 Avoid using make -j when upgrading. While generally safe, there are
1477 sometimes problems using -j to upgrade. If your upgrade fails with
1478 -j, please try again without -j. From time to time in the past there
1479 have been problems using -j with buildworld and/or installworld. This
1480 is especially true when upgrading between "distant" versions (eg one
1481 that cross a major release boundary or several minor releases, or when
1482 several months have passed on the -current branch).
1484 Sometimes, obscure build problems are the result of environment
1485 poisoning. This can happen because the make utility reads its
1486 environment when searching for values for global variables. To run
1487 your build attempts in an "environmental clean room", prefix all make
1488 commands with 'env -i '. See the env(1) manual page for more details.
1490 When upgrading from one major version to another it is generally best
1491 to upgrade to the latest code in the currently installed branch first,
1492 then do an upgrade to the new branch. This is the best-tested upgrade
1493 path, and has the highest probability of being successful. Please try
1494 this approach before reporting problems with a major version upgrade.
1498 When upgrading the boot ZFS pool to a new version, always follow
1501 1.) recompile and reinstall the ZFS boot loader and boot block
1502 (this is part of "make buildworld" and "make installworld")
1504 2.) update the ZFS boot block on your boot drive
1506 The following example updates the ZFS boot block on the first
1507 partition (freebsd-boot) of a GPT partitioned drive ad0:
1508 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1510 Non-boot pools do not need these updates.
1514 If you are updating from a prior version of FreeBSD (even one just
1515 a few days old), you should follow this procedure. It is the most
1516 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1518 make kernel-toolchain
1519 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1520 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1522 To test a kernel once
1523 ---------------------
1524 If you just want to boot a kernel once (because you are not sure
1525 if it works, or if you want to boot a known bad kernel to provide
1526 debugging information) run
1527 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1528 nextboot -k testkernel
1530 To just build a kernel when you know that it won't mess you up
1531 --------------------------------------------------------------
1532 This assumes you are already running a CURRENT system. Replace
1533 ${arch} with the architecture of your machine (e.g. "i386",
1534 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1536 cd src/sys/${arch}/conf
1537 config KERNEL_NAME_HERE
1538 cd ../compile/KERNEL_NAME_HERE
1543 If this fails, go to the "To build a kernel" section.
1545 To rebuild everything and install it on the current system.
1546 -----------------------------------------------------------
1547 # Note: sometimes if you are running current you gotta do more than
1548 # is listed here if you are upgrading from a really old current.
1550 <make sure you have good level 0 dumps>
1552 make kernel KERNCONF=YOUR_KERNEL_HERE
1554 <reboot in single user> [3]
1562 To cross-install current onto a separate partition
1563 --------------------------------------------------
1564 # In this approach we use a separate partition to hold
1565 # current's root, 'usr', and 'var' directories. A partition
1566 # holding "/", "/usr" and "/var" should be about 2GB in
1569 <make sure you have good level 0 dumps>
1572 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1573 <maybe newfs current's root partition>
1574 <mount current's root partition on directory ${CURRENT_ROOT}>
1575 make installworld DESTDIR=${CURRENT_ROOT}
1576 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1577 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1578 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1579 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1580 <reboot into current>
1581 <do a "native" rebuild/install as described in the previous section>
1582 <maybe install compatibility libraries from ports/misc/compat*>
1586 To upgrade in-place from stable to current
1587 ----------------------------------------------
1588 <make sure you have good level 0 dumps>
1590 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1592 <reboot in single user> [3]
1599 Make sure that you've read the UPDATING file to understand the
1600 tweaks to various things you need. At this point in the life
1601 cycle of current, things change often and you are on your own
1602 to cope. The defaults can also change, so please read ALL of
1603 the UPDATING entries.
1605 Also, if you are tracking -current, you must be subscribed to
1606 freebsd-current@freebsd.org. Make sure that before you update
1607 your sources that you have read and understood all the recent
1608 messages there. If in doubt, please track -stable which has
1609 much fewer pitfalls.
1611 [1] If you have third party modules, such as vmware, you
1612 should disable them at this point so they don't crash your
1615 [3] From the bootblocks, boot -s, and then do
1620 adjkerntz -i # if CMOS is wall time
1621 Also, when doing a major release upgrade, it is required that
1622 you boot into single user mode to do the installworld.
1624 [4] Note: This step is non-optional. Failure to do this step
1625 can result in a significant reduction in the functionality of the
1626 system. Attempting to do it by hand is not recommended and those
1627 that pursue this avenue should read this file carefully, as well
1628 as the archives of freebsd-current and freebsd-hackers mailing lists
1629 for potential gotchas. The -U option is also useful to consider.
1630 See mergemaster(8) for more information.
1632 [5] Usually this step is a noop. However, from time to time
1633 you may need to do this if you get unknown user in the following
1634 step. It never hurts to do it all the time. You may need to
1635 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1636 install) after the buildworld before this step if you last updated
1637 from [78]-stable or 9-stable before 20130430.
1639 [6] This only deletes old files and directories. Old libraries
1640 can be deleted by "make delete-old-libs", but you have to make
1641 sure that no program is using those libraries anymore.
1643 [8] In order to have a kernel that can run the 4.x binaries needed to
1644 do an installworld, you must include the COMPAT_FREEBSD4 option in
1645 your kernel. Failure to do so may leave you with a system that is
1646 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1647 required to run the 5.x binaries on more recent kernels. And so on
1648 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1650 Make sure that you merge any new devices from GENERIC since the
1651 last time you updated your kernel config file.
1653 [9] When checking out sources, you must include the -P flag to have
1654 cvs prune empty directories.
1656 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1657 "?=" instead of the "=" assignment operator, so that buildworld can
1658 override the CPUTYPE if it needs to.
1660 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1661 not on the command line, or in /etc/make.conf. buildworld will
1662 warn if it is improperly defined.
1665 This file contains a list, in reverse chronological order, of major
1666 breakages in tracking -current. It is not guaranteed to be a complete
1667 list of such breakages, and only contains entries since October 10, 2007.
1668 If you need to see UPDATING entries from before that date, you will need
1669 to fetch an UPDATING file from an older FreeBSD release.
1671 Copyright information:
1673 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1675 Redistribution, publication, translation and use, with or without
1676 modification, in full or in part, in any form or format of this
1677 document are permitted without further permission from the author.
1679 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1680 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1681 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1682 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1683 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1684 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1685 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1686 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1687 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1688 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1689 POSSIBILITY OF SUCH DAMAGE.
1691 Contact Warner Losh if you have any questions about your use of