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.
15 Clang and llvm have been upgraded to 3.4 release.
18 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
19 been deprecated and will not be part of FreeBSD 11.0 and later
20 releases. If you use this driver, please consider switching to
21 the nfe(4) driver instead.
24 The behavior of gss_pseudo_random() for the krb5 mechanism
25 has changed, for applications requesting a longer random string
26 than produced by the underlying enctype's pseudo-random() function.
27 In particular, the random string produced from a session key of
28 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
29 be different at the 17th octet and later, after this change.
30 The counter used in the PRF+ construction is now encoded as a
31 big-endian integer in accordance with RFC 4402.
32 __FreeBSD_version is bumped to 902505.
38 Behavior of devfs rules path matching has been changed.
39 Pattern is now always matched against fully qualified devfs
40 path and slash characters must be explicitly matched by
41 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
42 subdirectories must be reviewed.
45 hastctl(8)'s `status' command output changed to terse one-liner format.
46 Scripts using this should switch to `list' command or be rewritten.
49 Fix a bug that allowed a tracing process (e.g. gdb) to write
50 to a memory-mapped file in the traced process's address space
51 even if neither the traced process nor the tracing process had
52 write access to that file.
55 Added ZFS TRIM support which is enabled by default. To disable
56 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
58 Creating new ZFS pools and adding new devices to existing pools
59 first performs a full device level TRIM which can take a significant
60 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
61 to disable this behaviour.
63 ZFS TRIM requires the underlying device support BIO_DELETE which
64 is currently provided by methods such as ATA TRIM and SCSI UNMAP
65 via CAM, which are typically supported by SSD's.
67 Stats for ZFS TRIM can be monitored by looking at the sysctl's
68 under kstat.zfs.misc.zio_trim.
71 `list' command has been added to hastctl(8). For now, it is full
72 equivalent of `status' command.
73 WARNING: in the near future the output of hastctl's status command
74 will change to more terse format. If you use `hastctl status'
75 for parsing in your scripts, switch to `hastctl list'.
78 The mergemaster command now uses the default MAKEOBJDIRPREFIX
79 rather than creating it's own in the temporary directory in
80 order allow access to bootstrapped versions of tools such as
81 install and mtree. When upgrading from version of FreeBSD where
82 the install command does not support -l, you will need to
83 install a new mergemaster command if mergemaster -p is required.
84 This can be accomplished with the command (cd src/usr.sbin/mergemaster
87 Due to the use of the new -l option to install(1) during build
88 and install, you must take care not to directly set the INSTALL
89 make variable in your /etc/make.conf, /etc/src.conf, or on the
90 command line. If you wish to use the -C flag for all installs
91 you may be able to add INSTALL+=-C to /etc/make.conf or
95 Fix a bug that allows NFS clients to issue READDIR on files.
98 The install(1) option -M has changed meaning and now takes an
99 argument that is a file or path to append logs to. In the
100 unlikely event that -M was the last option on the command line
101 and the command line contained at least two files and a target
102 directory the first file will have logs appended to it. The -M
103 option served little practical purpose in the last decade so it's
104 used expected to be extremely rare.
107 A new compression method (lz4) has been merged to. Please refer to
108 zpool-features(7) for more information.
110 Please refer to the "ZFS notes" section of this file for information
111 on upgrading boot ZFS pools.
114 The VFS KBI was changed with the merge of several nullfs
115 optimizations and fixes. All filesystem modules must be
119 With the addition of auditdistd(8), a new auditdistd user is now
120 depended on during installworld. "mergemaster -p" can be used to add
121 the user prior to installworld, as documented in the handbook.
127 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
128 Starting with this version the old system of ZFS pool versioning
129 is superseded by "feature flags". This concept enables forward
130 compatibility against certain future changes in functionality of ZFS
131 pools. The first two read-only compatible "feature flags" for ZFS
132 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
133 For more information read the new zpool-features(7) manual page.
134 Please refer to the "ZFS notes" section of this file for information
135 on upgrading boot ZFS pools.
138 The commit introducing bsd.compiler.mk breaks the traditional
139 building of kernels before this point. Add -m ${SRC}/share/mk
140 (for the right value of SRC) to your command lines to work
141 around; update your useland to a point after this; or use the
142 buildkernel/installkernel top-level targets. See also 20120829.
145 The IPFIREWALL_FORWARD kernel option has been removed. Its
146 functionality now turned on by default.
149 The random(4) support for the VIA hardware random number
150 generator (`PADLOCK') is no longer enabled unconditionally.
151 Add the PADLOCK_RNG option in the custom kernel config if
152 needed. The GENERIC kernels on i386 and amd64 do include the
153 option, so the change only affects the custom kernel
157 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
158 the traditional method, you must update your system with an installworld
159 before the kernel will build. The documented make buildkernel/installkernel
160 interfaces (coupled with fresh make kernel-toolchain) continue to work.
163 The sparc64 ZFS loader has been changed to no longer try to auto-
164 detect ZFS providers based on diskN aliases but now requires these
165 to be explicitly listed in the OFW boot-device environment variable.
168 Now unix domain sockets behave "as expected" on nullfs(5). Previously
169 nullfs(5) did not pass through all behaviours to the underlying layer,
170 as a result if we bound to a socket on the lower layer we could connect
171 only to the lower path; if we bound to the upper layer we could connect
172 only to the upper path. The new behavior is one can connect to both the
173 lower and the upper paths regardless what layer path one binds to.
176 The acpi_wmi(4) status device /dev/wmistat has been renamed to
180 A new VOP_ADVISE() was added to support posix_fadvise(2). All
181 filesystem modules must be recompiled.
184 The interface of the VOP_VPTOCNP(9) changed, now the returned
185 vnode shall be referenced, previously it was required to be
186 only held. All in-tree filesystems are converted.
192 The broken amd(4) driver has been replaced with esp(4) in the amd64,
193 i386 and pc98 GENERIC kernel configuration files.
196 This commit modifies vfs_register() so that it uses a hash
197 calculation to set vfc_typenum, which is enabled by default.
198 The first time a system is booted after this change, the
199 vfc_typenum values will change for all file systems. The
200 main effect of this is a change to the NFS server file handles
201 for file systems that use vfc_typenum in their fsid, such as ZFS.
202 It will, however, prevent vfc_typenum from changing when file
203 systems are loaded in a different order for subsequent reboots.
204 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
205 until you are ready to remount all NFS clients after a reboot.
208 Bump the shared library version numbers for libraries that
209 do not use symbol versioning, have changed the ABI compared
210 to stable/8 and which shared library version was not bumped.
211 Done as part of 9.0-RELEASE cycle.
214 During the merge of Capsicum features, the fget(9) KPI was modified.
215 This may require the rebuilding of out-of-tree device drivers --
216 issues have been reported specifically with the nVidia device driver.
217 __FreeBSD_version is bumped to 900041.
219 Also, there is a period between 20110811 and 20110814 where the
220 special devices /dev/{stdin,stdout,stderr} did not work correctly.
221 Building world from a kernel during that window may not work.
224 The packet filter (pf) code has been updated to OpenBSD 4.5.
225 You need to update userland tools to be in sync with kernel.
226 This update breaks backward compatibility with earlier pfsync(4)
227 versions. Care must be taken when updating redundant firewall setups.
230 The following sysctls and tunables are retired on x86 platforms:
232 machdep.hlt_logical_cpus
233 The following sysctl is retired:
234 machdep.hyperthreading_allowed
235 The sysctls were supposed to provide a way to dynamically offline and
236 online selected CPUs on x86 platforms, but the implementation has not
237 been reliable especially with SCHED_ULE scheduler.
238 machdep.hyperthreading_allowed tunable is still available to ignore
239 hyperthreading CPUs at OS level.
240 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
241 where X is an APIC ID of a CPU. Be advised, though, that disabling
242 CPUs in non-uniform fashion will result in non-uniform topology and
243 may lead to sub-optimal system performance with SCHED_ULE, which is
247 cpumask_t type is retired and cpuset_t is used in order to describe
251 Changes to ifconfig(8) for dynamic address family detection mandate
252 that you are running a kernel of 20110525 or later. Make sure to
253 follow the update procedure to boot a new kernel before installing
257 Support for sun4v architecture is officially dropped
260 Several KPI breaking changes have been committed to the mii(4) layer,
261 the PHY drivers and consequently some Ethernet drivers using mii(4).
262 This means that miibus.ko and the modules of the affected Ethernet
263 drivers need to be recompiled.
265 Note to kernel developers: Given that the OUI bit reversion problem
266 was fixed as part of these changes all mii(4) commits related to OUIs,
267 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
268 handling, no longer can be merged verbatim to stable/8 and previous
272 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
273 into their kernel configurations along with 'device pci'.
276 The default NFS client is now the new NFS client, so fstype "newnfs"
277 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
278 Although mounts via fstype "nfs" will usually work without userland
279 changes, it is recommended that the mount(8) and mount_nfs(8)
280 commands be rebuilt from sources and that a link to mount_nfs called
281 mount_oldnfs be created. The new client is compiled into the
282 kernel with "options NFSCL" and this is needed for diskless root
283 file systems. The GENERIC kernel configs have been changed to use
284 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
285 To use the regular/old client, you can "mount -t oldnfs ...". For
286 a diskless root file system, you must also include a line like:
288 vfs.root.mountfrom="oldnfs:"
290 in the boot/loader.conf on the root fs on the NFS server to make
291 a diskless root fs use the old client.
294 The GENERIC kernels for all architectures now default to the new
295 CAM-based ATA stack. It means that all legacy ATA drivers were
296 removed and replaced by respective CAM drivers. If you are using
297 ATA device names in /etc/fstab or other places, make sure to update
298 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
299 where 'Y's are the sequential numbers starting from zero for each type
300 in order of detection, unless configured otherwise with tunables,
301 see cam(4)). There will be symbolic links created in /dev/ to map
302 old adX devices to the respective adaY. They should provide basic
303 compatibility for file systems mounting in most cases, but they do
304 not support old user-level APIs and do not have respective providers
305 in GEOM. Consider using updated management tools with new device names.
307 It is possible to load devices ahci, ata, siis and mvs as modules,
308 but option ATA_CAM should remain in kernel configuration to make ata
309 module work as CAM driver supporting legacy ATA controllers. Device ata
310 still can be used in modular fashion (atacore + ...). Modules atadisk
311 and atapi* are not used and won't affect operation in ATA_CAM mode.
312 Note that to use CAM-based ATA kernel should include CAM devices
313 scbus, pass, da (or explicitly ada), cd and optionally others. All of
314 them are parts of the cam module.
316 ataraid(4) functionality is now supported by the RAID GEOM class.
317 To use it you can load geom_raid kernel module and use graid(8) tool
318 for management. Instead of /dev/arX device names, use /dev/raid/rX.
320 No kernel config options or code have been removed, so if a problem
321 arises, please report it and optionally revert to the old ATA stack.
322 In order to do it you can remove from the kernel config:
327 , and instead add back:
328 device atadisk # ATA disk drives
329 device ataraid # ATA RAID drives
330 device atapicd # ATAPI CDROM drives
331 device atapifd # ATAPI floppy drives
332 device atapist # ATAPI tape drives
335 The default NFS server has been changed to the new server, which
336 was referred to as the experimental server. If you need to switch
337 back to the old NFS server, you must now put the "-o" option on
338 both the mountd and nfsd commands. This can be done using the
339 mountd_flags and nfs_server_flags rc.conf variables until an
340 update to the rc scripts is committed, which is coming soon.
343 The GNU Objective-C runtime library (libobjc), and other Objective-C
344 related components have been removed from the base system. If you
345 require an Objective-C library, please use one of the available ports.
348 ath(4) has been split into bus- and device- modules. if_ath contains
349 the HAL, the TX rate control and the network device code. if_ath_pci
350 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
351 contains the AHB glue. Users need to load both if_ath_pci and if_ath
352 in order to use ath on everything else.
354 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
355 need to load if_ath and if_ath_pci for ath(4) operation.
358 As part of the replacement of sysinstall, the process of building
359 release media has changed significantly. For details, please re-read
360 release(7), which has been updated to reflect the new build process.
363 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
364 is the last available version under GPLv2. It brings a number of new
365 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
366 4.1 and SSE 4.2), better support for powerpc64, a number of new
367 directives, and lots of other small improvements. See the ChangeLog
368 file in contrib/binutils for the full details.
371 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
372 compliant, and will now use half of hash for authentication.
373 This will break interoperability with all stacks (including all
374 actual FreeBSD versions) who implement
375 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
377 The only workaround with such peers is to use another HMAC
378 algorithm for IPsec ("phase 2") authentication.
381 Remove the uio_yield prototype and symbol. This function has
382 been misnamed since it was introduced and should not be
383 globally exposed with this name. The equivalent functionality
384 is now available using kern_yield(curthread->td_user_pri).
385 The function remains undocumented.
388 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
389 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
390 for scalar sysctls is defined but disabled. Code that needs
391 UQUAD to pass the type checking that must compile on older
392 systems where the define is not present can check against
393 __FreeBSD_version >= 900030.
395 The system dialog(1) has been replaced with a new version previously
396 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
397 with the previous version, but the libdialog associated with it has
398 a largely incompatible API. As such, the original version of libdialog
399 will be kept temporarily as libodialog, until its base system consumers
400 are replaced or updated. Bump __FreeBSD_version to 900030.
403 If you are trying to run make universe on a -stable system, and you get
404 the following warning:
405 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
406 unknown. config(8) likely too old."
407 or something similar to it, then you must upgrade your -stable system
408 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
409 stable/8) or build the config from the latest stable/8 branch and
410 install it on your system.
412 Prior to this date, building a current universe on 8-stable system from
413 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
414 error in the first kernel build phase. A new config on those old
415 systems will fix that problem for older versions of -current.
418 The TCP stack has been modified to allow Khelp modules to interact with
419 it via helper hook points and store per-connection data in the TCP
420 control block. Bump __FreeBSD_version to 900029. User space tools that
421 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
425 Generic IEEE 802.3 annex 31B full duplex flow control support has been
426 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
427 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
428 to take advantage of it instead of using custom implementations. This
429 means that these drivers now no longer unconditionally advertise
430 support for flow control but only do so if flow control is a selected
431 media option. This was implemented in the generic support that way in
432 order to allow flow control to be switched on and off via ifconfig(8)
433 with the PHY specific default to typically off in order to protect
434 from unwanted effects. Consequently, if you used flow control with
435 one of the above mentioned drivers you now need to explicitly enable
437 ifconfig bge0 media auto mediaopt flowcontrol
439 Along with the above mentioned changes generic support for setting
440 1000baseT master mode also has been added and brgphy(4), ciphy(4),
441 e1000phy(4) as well as ip1000phy(4) have been converted to take
442 advantage of it. This means that these drivers now no longer take the
443 link0 parameter for selecting master mode but the master media option
444 has to be used instead, for example like in the following:
445 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
447 Selection of master mode now is also available with all other PHY
448 drivers supporting 1000baseT.
451 The TCP stack has received a significant update to add support for
452 modularised congestion control and generally improve the clarity of
453 congestion control decisions. Bump __FreeBSD_version to 900025. User
454 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
455 sockstat) need to be recompiled.
458 The man(1) utility has been replaced by a new version that no longer
459 uses /etc/manpath.config. Please consult man.conf(5) for how to
460 migrate local entries to the new format.
463 The copyright strings printed by login(1) and sshd(8) at the time of a
464 new connection have been removed to follow other operating systems and
468 A workaround for a fixed ld bug has been removed in kernel code,
469 so make sure that your system ld is built from sources after
470 revision 210245 from 2010-07-19 (r211583 if building head kernel
471 on stable/8, r211584 for stable/7; both from 2010-08-21).
472 A symptom of incorrect ld version is different addresses for
473 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
476 The $ipv6_prefer variable in rc.conf(5) has been split into
477 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
479 The $ip6addrctl_policy is a variable to choose a pre-defined
480 address selection policy set by ip6addrctl(8). A value
481 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
484 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
485 flag (see an entry of 20090926) is set on an interface with no
486 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
487 security reason. If you want IPv6 link-local address on all
488 interfaces by default, set this to "YES".
490 The old ipv6_prefer="YES" is equivalent to
491 ipv6_activate_all_interfaces="YES" and
492 ip6addrctl_policy="ipv6_prefer".
495 DTrace has grown support for userland tracing. Due to this, DTrace is
496 now i386 and amd64 only.
497 dtruss(1) is now installed by default on those systems and a new
498 kernel module is needed for userland tracing: fasttrap.
499 No changes to your kernel config file are necessary to enable
500 userland tracing, but you might consider adding 'STRIP=' and
501 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
502 to have informative userland stack traces in DTrace (ustack).
505 The acpi_aiboost(4) driver has been removed in favor of the new
506 aibs(4) driver. You should update your kernel configuration file.
509 BSD grep has been imported to the base system and it is built by
510 default. It is completely BSD licensed, highly GNU-compatible, uses
511 less memory than its GNU counterpart and has a small codebase.
512 However, it is slower than its GNU counterpart, which is mostly
513 noticeable for larger searches, for smaller ones it is measurable
514 but not significant. The reason is complex, the most important factor
515 is that we lack a modern and efficient regex library and GNU
516 overcomes this by optimizing the searches internally. Future work
517 on improving the regex performance is planned, for the meantime,
518 users that need better performance, can build GNU grep instead by
519 setting the WITH_GNU_GREP knob.
522 Due to the import of powerpc64 support, all existing powerpc kernel
523 configuration files must be updated with a machine directive like this:
524 machine powerpc powerpc
526 In addition, an updated config(8) is required to build powerpc kernels
530 A new version of ZFS (version 15) has been merged to -HEAD.
531 This version uses a python library for the following subcommands:
532 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
533 For full functionality of these commands the following port must
534 be installed: sysutils/py-zfs
537 'vm_page's are now hashed by physical address to an array of mutexes.
538 Currently this is only used to serialize access to hold_count. Over
539 time the page queue mutex will be peeled away. This changes the size
540 of pmap on every architecture. And requires all callers of vm_page_hold
541 and vm_page_unhold to be updated.
544 WITH_CTF can now be specified in src.conf (not recommended, there
545 are some problems with static executables), make.conf (would also
546 affect ports which do not use GNU make and do not override the
547 compile targets) or in the kernel config (via "makeoptions
549 When WITH_CTF was specified there before this was silently ignored,
550 so make sure that WITH_CTF is not used in places which could lead
551 to unwanted behavior.
554 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
555 to allow 32-bit compatibility on non-x86 platforms. All kernel
556 configurations on amd64 and ia64 platforms using these options must
557 be modified accordingly.
560 The utmp user accounting database has been replaced with utmpx,
561 the user accounting interface standardized by POSIX.
562 Unfortunately the semantics of utmp and utmpx don't match,
563 making it practically impossible to support both interfaces.
564 The user accounting database is used by tools like finger(1),
565 last(1), talk(1), w(1) and ac(8).
567 All applications in the base system use utmpx. This means only
568 local binaries (e.g. from the ports tree) may still use these
569 utmp database files. These applications must be rebuilt to make
572 After the system has been upgraded, it is safe to remove the old
573 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
574 assuming their contents is of no importance anymore. Old wtmp
575 databases can only be used by last(1) and ac(8) after they have
576 been converted to the new format using wtmpcvt(1).
579 Introduce the kernel thread "deadlock resolver" (which can be enabled
580 via the DEADLKRES option, see NOTES for more details) and the
581 sleepq_type() function for sleepqueues.
584 The rc.firewall and rc.firewall6 were unified, and
585 rc.firewall6 and rc.d/ip6fw were removed.
586 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
587 variables are obsoleted. Instead, the following new rc
588 variables are added to rc.d/ipfw:
590 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
591 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
592 firewall_simple_onet_ipv6, firewall_trusted_ipv6
594 The meanings correspond to the relevant IPv4 variables.
600 The default terminal emulation for syscons(4) has been changed
601 from cons25 to xterm on all platforms except pc98. This means
602 that the /etc/ttys file needs to be updated to ensure correct
603 operation of applications on the console.
605 The terminal emulation style can be toggled per window by using
606 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
607 options can be used to change the compile-time default back to
610 To prevent graphical artifacts, make sure the TERM environment
611 variable is set to match the terminal emulation that is being
612 performed by syscons(4).
615 The layout of the structure ieee80211req_scan_result has changed.
616 Applications that require wireless scan results (e.g. ifconfig(8))
617 from net80211 need to be recompiled.
619 Applications such as wpa_supplicant(8) may require a full world
620 build without using NO_CLEAN in order to get synchronized with the
624 The iwn(4) driver has been updated to support the 5000 and 5150 series.
625 There's one kernel module for each firmware. Adding "device iwnfw"
626 to the kernel configuration file means including all three firmware
627 images inside the kernel. If you want to include just the one for
628 your wireless card, use the devices iwn4965fw, iwn5000fw or
632 The rc.d/network_ipv6, IPv6 configuration script has been integrated
633 into rc.d/netif. The changes are the following:
635 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
636 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
637 Note that both variables need the "inet6" keyword at the head.
639 Do not set $ipv6_network_interfaces manually if you do not
640 understand what you are doing. It is not needed in most cases.
642 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
645 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
646 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
648 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
649 all configured interfaces have "inet6 accept_rtadv" in the
650 $ifconfig_IF_ipv6. These are for backward compatibility.
652 3. A new variable $ipv6_prefer has been added. If NO, IPv6
653 functionality of interfaces with no corresponding
654 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
655 and the default address selection policy of ip6addrctl(8)
656 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
657 Note that if you want to configure IPv6 functionality on the
658 disabled interfaces after boot, first you need to clear the flag by
659 using ifconfig(8) like:
661 ifconfig em0 inet6 -ifdisabled
663 If YES, the default address selection policy is set as
666 The default value of $ipv6_prefer is NO.
668 4. If your system need to receive Router Advertisement messages,
669 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
670 scripts automatically invoke rtsol(8) when the interface becomes
671 UP. The Router Advertisement messages are used for SLAAC
672 (State-Less Address AutoConfiguration).
675 802.11s D3.03 support was committed. This is incompatible with the
676 previous code, which was based on D3.0.
679 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
680 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
681 control whether accepting Router Advertisement messages or not.
682 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
683 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
684 The ifconfig(8) utility now supports these flags.
687 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
688 mount(8) and -a option for df(1) to see them.
691 The old tunable hw.bus.devctl_disable has been superseded by
692 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
693 replaced by hw.bus.devctl_queue=0. The default for this new tunable
697 Remove the option STOP_NMI. The default action is now to use NMI only
698 for KDB via the newly introduced function stop_cpus_hard() and
699 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
702 The stable/8 branch created in subversion. This corresponds to the
703 RELENG_8 branch in CVS.
706 Bump the shared library version numbers for all libraries that do not
707 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
708 __FreeBSD_version to 800105.
711 Due to changes in the implementation of virtual network stack support,
712 all network-related kernel modules must be recompiled. As this change
713 breaks the ABI, bump __FreeBSD_version to 800104.
716 The TOE interface to the TCP syncache has been modified to remove
717 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
718 The cxgb driver is the only TOE consumer affected by this change, and
719 needs to be recompiled along with the kernel. As this change breaks
720 the ABI, bump __FreeBSD_version to 800103.
723 Padding has been added to struct tcpcb, sackhint and tcpstat in
724 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
725 maintaining the ABI. However, this change breaks the ABI, so bump
726 __FreeBSD_version to 800102. User space tools that rely on the size of
727 any of these structs (e.g. sockstat) need to be recompiled.
730 The NFS_LEGACYRPC option has been removed along with the old kernel
731 RPC implementation that this option selected. Kernel configurations
732 may need to be adjusted.
735 The network interface device nodes at /dev/net/<interface> have been
736 removed. All ioctl operations can be performed the normal way using
737 routing sockets. The kqueue functionality can generally be replaced
738 with routing sockets.
741 The documentation from the FreeBSD Documentation Project (Handbook,
742 FAQ, etc.) is now installed via packages by sysinstall(8) and under
743 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
746 The ABI of various structures related to the SYSV IPC API have been
747 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
748 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
752 Layout of struct vnet has changed as routing related variables were
753 moved to their own Vimage module. Modules need to be recompiled. Bump
754 __FreeBSD_version to 800099.
757 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
758 respectively. As long as no more than 16 groups per process are used,
759 no changes should be visible. When more than 16 groups are used, old
760 binaries may fail if they call getgroups() or getgrouplist() with
761 statically sized storage. Recompiling will work around this, but
762 applications should be modified to use dynamically allocated storage
763 for group arrays as POSIX.1-2008 does not cap an implementation's
764 number of supported groups at NGROUPS_MAX+1 as previous versions did.
766 NFS and portalfs mounts may also be affected as the list of groups is
767 truncated to 16. Users of NFS who use more than 16 groups, should
768 take care that negative group permissions are not used on the exported
769 file systems as they will not be reliable unless a GSSAPI based
770 authentication method is used.
773 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
774 option compiles in the support for adaptive spinning for lockmgrs
775 which want to enable it. The lockinit() function now accepts the flag
776 LK_ADAPTIVE in order to make the lock object subject to adaptive
777 spinning when both held in write and read mode.
780 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
781 changed. User applications that use this ioctl need to be rebuilt.
784 The layout of struct thread has changed. Kernel and modules need to
788 The layout of structs ifnet, domain, protosw and vnet_net has changed.
789 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
792 window(1) has been removed from the base system. It can now be
793 installed from ports. The port is called misc/window.
796 The way we are storing and accessing `routing table' entries has
797 changed. Programs reading the FIB, like netstat, need to be
801 A new netisr implementation has been added for FreeBSD 8. Network
802 file system modules, such as igmp, ipdivert, and others, should be
804 Bump __FreeBSD_version to 800096.
807 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
811 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
812 Bump __FreeBSD_version to 800094.
815 Add mnt_xflag field to 'struct mount'. File system modules need to be
817 Bump __FreeBSD_version to 800093.
820 The compiling option ADAPTIVE_SX has been retired while it has been
821 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
822 The KPI for sx_init_flags() changes as accepting flags:
823 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
824 been introduced in order to handle the reversed logic.
825 Bump __FreeBSD_version to 800092.
828 Add support for hierarchical jails. Remove global securelevel.
829 Bump __FreeBSD_version to 800091.
832 The layout of struct vnet_net has changed, therefore modules
834 Bump __FreeBSD_version to 800090.
837 The newly imported zic(8) produces a new format in the output. Please
838 run tzsetup(8) to install the newly created data to /etc/localtime.
841 The sysctl tree for the usb stack has renamed from hw.usb2.* to
842 hw.usb.* and is now consistent again with previous releases.
845 802.11 monitor mode support was revised and driver api's were changed.
846 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
847 of DLT_IEEE802_11. No user-visible data structures were changed but
848 applications that use DLT_IEEE802_11 may require changes.
849 Bump __FreeBSD_version to 800088.
852 The layout of the following structs has changed: sysctl_oid,
853 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
854 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
855 panics may be experienced. World rebuild is required for
856 correctly checking networking state from userland.
857 Bump __FreeBSD_version to 800085.
860 MLDv2 and Source-Specific Multicast (SSM) have been merged
861 to the IPv6 stack. VIMAGE hooks are in but not yet used.
862 The implementation of SSM within FreeBSD's IPv6 stack closely
863 follows the IPv4 implementation.
865 For kernel developers:
867 * The most important changes are that the ip6_output() and
868 ip6_input() paths no longer take the IN6_MULTI_LOCK,
869 and this lock has been downgraded to a non-recursive mutex.
871 * As with the changes to the IPv4 stack to support SSM, filtering
872 of inbound multicast traffic must now be performed by transport
873 protocols within the IPv6 stack. This does not apply to TCP and
874 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
876 * The KPIs used by IPv6 multicast are similar to those used by
877 the IPv4 stack, with the following differences:
878 * im6o_mc_filter() is analogous to imo_multicast_filter().
879 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
880 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
881 * IN6_LOOKUP_MULTI() has been deprecated and removed.
882 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
883 for MLDv1 have an additional 'timer' argument which is used to
884 jitter the initial membership report for the solicited-node
885 multicast membership on-link.
886 * This is not strictly needed for MLDv2, which already jitters
887 its report transmissions. However, the 'timer' argument is
888 preserved in case MLDv1 is active on the interface.
890 * The KAME linked-list based IPv6 membership implementation has
891 been refactored to use a vector similar to that used by the IPv4
893 Code which maintains a list of its own multicast memberships
894 internally, e.g. carp, has been updated to reflect the new
897 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
898 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
899 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
900 implementation constraint which needs to be addressed in HEAD.
902 For application developers:
904 * The changes are broadly similar to those made for the IPv4
907 * The use of IPv4 and IPv6 multicast socket options on the same
908 socket, using mapped addresses, HAS NOT been tested or supported.
910 * There are a number of issues with the implementation of various
911 IPv6 multicast APIs which need to be resolved in the API surface
912 before the implementation is fully compatible with KAME userland
913 use, and these are mostly to do with interface index treatment.
915 * The literature available discusses the use of either the delta / ASM
916 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
917 using setsourcefilter(3)/getsourcefilter(3). For more information
918 please refer to RFC 3768, 'Socket Interface Extensions for
919 Multicast Source Filters'.
921 * Applications which use the published RFC 3678 APIs should be fine.
923 For systems administrators:
925 * The mtest(8) utility has been refactored to support IPv6, in
926 addition to IPv4. Interface addresses are no longer accepted
927 as arguments, their names must be used instead. The utility
928 will map the interface name to its first IPv4 address as
929 returned by getifaddrs(3).
931 * The ifmcstat(8) utility has also been updated to print the MLDv2
932 endpoint state and source filter lists via sysctl(3).
934 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
935 loopback of IPv6 multicast datagrams by default; it defaults to 1
936 to preserve the existing behaviour. Disabling multicast loopback is
937 recommended for optimal system performance.
939 * The IPv6 MROUTING code has been changed to examine this sysctl
940 instead of attempting to perform a group lookup before looping
941 back forwarded datagrams.
943 Bump __FreeBSD_version to 800084.
946 Implement low-level Bluetooth HCI API.
947 Bump __FreeBSD_version to 800083.
950 The layout of struct malloc_type, used by modules to register new
951 memory allocation types, has changed. Most modules will need to
952 be rebuilt or panics may be experienced.
953 Bump __FreeBSD_version to 800081.
956 Anticipate overflowing inp_flags - add inp_flags2.
957 This changes most offsets in inpcb, so checking v4 connection
958 state will require a world rebuild.
959 Bump __FreeBSD_version to 800080.
962 Add an llentry to struct route and struct route_in6. Modules
963 embedding a struct route will need to be recompiled.
964 Bump __FreeBSD_version to 800079.
967 The size of rt_metrics_lite and by extension rtentry has changed.
968 Networking administration apps will need to be recompiled.
969 The route command now supports show as an alias for get, weighting
970 of routes, sticky and nostick flags to alter the behavior of stateful
972 Bump __FreeBSD_version to 800078.
975 Do not use Giant for kbdmux(4) locking. This is wrong and
976 apparently causing more problems than it solves. This will
977 re-open the issue where interrupt handlers may race with
978 kbdmux(4) in polling mode. Typical symptoms include (but
979 not limited to) duplicated and/or missing characters when
980 low level console functions (such as gets) are used while
981 interrupts are enabled (for example geli password prompt,
982 mountroot prompt etc.). Disabling kbdmux(4) may help.
985 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
986 kernel modules referencing any of the above need to be recompiled.
987 Bump __FreeBSD_version to 800075.
990 GEOM_PART has become the default partition slicer for storage devices,
991 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
992 introduces some changes:
994 MSDOS/EBR: the devices created from MSDOS extended partition entries
995 (EBR) can be named differently than with GEOM_MBR and are now symlinks
996 to devices with offset-based names. fstabs may need to be modified.
998 BSD: the "geometry does not match label" warning is harmless in most
999 cases but it points to problems in file system misalignment with
1000 disk geometry. The "c" partition is now implicit, covers the whole
1001 top-level drive and cannot be (mis)used by users.
1003 General: Kernel dumps are now not allowed to be written to devices
1004 whose partition types indicate they are meant to be used for file
1005 systems (or, in case of MSDOS partitions, as something else than
1008 Most of these changes date approximately from 200812.
1011 The uscanner(4) driver has been removed from the kernel. This follows
1012 Linux removing theirs in 2.6 and making libusb the default interface
1013 (supported by sane).
1016 The multicast forwarding code has been cleaned up. netstat(1)
1017 only relies on KVM now for printing bandwidth upcall meters.
1018 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1019 ip6_mroute_mod respectively. The config(5) options for statically
1020 compiling this code remain the same, i.e. 'options MROUTING'.
1023 Support for the IFF_NEEDSGIANT network interface flag has been
1024 removed, which means that non-MPSAFE network device drivers are no
1025 longer supported. In particular, if_ar, if_sr, and network device
1026 drivers from the old (legacy) USB stack can no longer be built or
1030 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1031 a bunch of new language catalog files have also been added.
1032 This means that some common libc messages are now localized and
1033 they depend on the LC_MESSAGES environmental variable.
1036 The k8temp(4) driver has been renamed to amdtemp(4) since
1037 support for Family 10 and Family 11 CPU families was added.
1040 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1041 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1043 For kernel developers, the most important changes are that the
1044 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1045 and this lock has been downgraded to a non-recursive mutex.
1047 Transport protocols (UDP, Raw IP) are now responsible for filtering
1048 inbound multicast traffic according to group membership and source
1049 filters. The imo_multicast_filter() KPI exists for this purpose.
1050 Transports which do not use multicast (SCTP, TCP) already reject
1051 multicast by default. Forwarding and receive performance may improve
1052 as a mutex acquisition is no longer needed in the ip_input()
1053 low-level input path. in_addmulti() and in_delmulti() are shimmed
1054 to new KPIs which exist to support SSM in-kernel.
1056 For application developers, it is recommended that loopback of
1057 multicast datagrams be disabled for best performance, as this
1058 will still cause the lock to be taken for each looped-back
1059 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1060 be tuned to 0 to disable loopback by default; it defaults to 1
1061 to preserve the existing behaviour.
1063 For systems administrators, to obtain best performance with
1064 multicast reception and multiple groups, it is always recommended
1065 that a card with a suitably precise hash filter is used. Hash
1066 collisions will still result in the lock being taken within the
1067 transport protocol input path to check group membership.
1069 If deploying FreeBSD in an environment with IGMP snooping switches,
1070 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1071 enabled; this forces 224.0.0.0/24 group membership to be announced
1074 The size of 'struct igmpstat' has changed; netstat needs to be
1075 recompiled to reflect this.
1076 Bump __FreeBSD_version to 800070.
1079 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1080 updated to use it. This requires a buildworld/installworld in order to
1081 update the library and dependencies (usbconfig, etc). Its advisable to
1082 rebuild all ports which uses libusb. More specific directions are given
1083 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1084 libusb are no longer required and can be removed.
1087 A workaround is committed to allow the creation of System V shared
1088 memory segment of size > 2 GB on the 64-bit architectures.
1089 Due to a limitation of the existing ABI, the shm_segsz member
1090 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1091 wrong for large segments. Note that limits must be explicitly
1092 raised to allow such segments to be created.
1095 The layout of struct ifnet has changed, requiring a rebuild of all
1096 network device driver modules.
1099 The /dev handling for the new USB stack has changed, a
1100 buildworld/installworld is required for libusb20.
1103 The new USB2 stack has now been permanently moved in and all kernel and
1104 module names reverted to their previous values (eg, usb, ehci, ohci,
1105 ums, ...). The old usb stack can be compiled in by prefixing the name
1106 with the letter 'o', the old usb modules have been removed.
1107 Updating entry 20090216 for xorg and 20090215 for libmap may still
1111 The rc.conf(5) option if_up_delay has been renamed to
1112 defaultroute_delay to better reflect its purpose. If you have
1113 customized this setting in /etc/rc.conf you need to update it to
1117 xorg 7.4 wants to configure its input devices via hald which does not
1118 yet work with USB2. If the keyboard/mouse does not work in xorg then
1120 Option "AllowEmptyInput" "off"
1121 to your ServerLayout section. This will cause X to use the configured
1122 kbd and mouse sections from your xorg.conf.
1125 The GENERIC kernels for all architectures now default to the new USB2
1126 stack. No kernel config options or code have been removed so if a
1127 problem arises please report it and optionally revert to the old USB
1128 stack. If you are loading USB kernel modules or have a custom kernel
1129 that includes GENERIC then ensure that usb names are also changed over,
1130 eg uftdi -> usb2_serial_ftdi.
1132 Older programs linked against the ports libusb 0.1 need to be
1133 redirected to the new stack's libusb20. /etc/libmap.conf can
1135 # Map old usb library to new one for usb2 stack
1136 libusb-0.1.so.8 libusb20.so.1
1139 All USB ethernet devices now attach as interfaces under the name ueN
1140 (eg. ue0). This is to provide a predictable name as vendors often
1141 change usb chipsets in a product without notice.
1144 The ichsmb(4) driver has been changed to require SMBus slave
1145 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1146 All of the other SMBus controller drivers require left-justified
1147 slave addresses, so this change makes all the drivers provide the
1151 INET6 statistics (struct ip6stat) was updated.
1152 netstat(1) needs to be recompiled.
1155 NTFS has been removed from GENERIC kernel on amd64 to match
1156 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1157 will load ntfs.ko module automatically when NTFS support is
1158 actually needed, unless ntfs.ko is not installed or security
1159 level prohibits loading kernel modules. If either is the case,
1160 "options NTFS" has to be added into kernel config.
1163 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1164 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1165 800061. User space tools that rely on the size of struct tcpcb in
1166 tcp_var.h (e.g. sockstat) need to be recompiled.
1169 ng_tty(4) module updated to match the new TTY subsystem.
1170 Due to API change, user-level applications must be updated.
1171 New API support added to mpd5 CVS and expected to be present
1172 in next mpd5.3 release.
1175 With __FreeBSD_version 800060 the makefs tool is part of
1176 the base system (it was a port).
1179 The afdata and ifnet locks have been changed from mutexes to
1180 rwlocks, network modules will need to be re-compiled.
1183 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1184 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1185 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1186 architecture and 40 bytes on 64-bit architecture. The userland
1187 applications "arp" and "ndp" have been updated accordingly.
1188 The output from "netstat -r" shows only routing entries and
1189 none of the L2 information.
1192 __FreeBSD_version 800057 marks the switchover from the
1193 binary ath hal to source code. Users must add the line:
1195 options AH_SUPPORT_AR5416
1197 to their kernel config files when specifying:
1201 The ath_hal module no longer exists; the code is now compiled
1202 together with the driver in the ath module. It is now
1203 possible to tailor chip support (i.e. reduce the set of chips
1204 and thereby the code size); consult ath_hal(4) for details.
1207 __FreeBSD_version 800054 adds memory barriers to
1208 <machine/atomic.h>, new interfaces to ifnet to facilitate
1209 multiple hardware transmit queues for cards that support
1210 them, and a lock-less ring-buffer implementation to
1211 enable drivers to more efficiently manage queueing of
1215 A new version of ZFS (version 13) has been merged to -HEAD.
1216 This version has zpool attribute "listsnapshots" off by
1217 default, which means "zfs list" does not show snapshots,
1218 and is the same as Solaris behavior.
1221 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1224 The uhci, ohci, ehci and slhci USB Host controller drivers have
1225 been put into separate modules. If you load the usb module
1226 separately through loader.conf you will need to load the
1227 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1228 controller add the following to loader.conf:
1234 The ABI used by the PMC toolset has changed. Please keep
1235 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1239 atapci kernel module now includes only generic PCI ATA
1240 driver. AHCI driver moved to ataahci kernel module.
1241 All vendor-specific code moved into separate kernel modules:
1242 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1243 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1244 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1245 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1248 The TTY subsystem of the kernel has been replaced by a new
1249 implementation, which provides better scalability and an
1250 improved driver model. Most common drivers have been migrated to
1251 the new TTY subsystem, while others have not. The following
1252 drivers have not yet been ported to the new TTY layer:
1255 cy, digi, rc, rp, sio
1261 ng_h4, ng_tty, ppp, sl, snp
1263 Adding these drivers to your kernel configuration file shall
1264 cause compilation to fail.
1267 ntpd has been upgraded to 4.2.4p5.
1270 OpenSSH has been upgraded to 5.1p1.
1272 For many years, FreeBSD's version of OpenSSH preferred DSA
1273 over RSA for host and user authentication keys. With this
1274 upgrade, we've switched to the vendor's default of RSA over
1275 DSA. This may cause upgraded clients to warn about unknown
1276 host keys even for previously known hosts. Users should
1277 follow the usual procedure for verifying host keys before
1278 accepting the RSA key.
1280 This can be circumvented by setting the "HostKeyAlgorithms"
1281 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1284 Please note that the sequence of keys offered for
1285 authentication has been changed as well. You may want to
1286 specify IdentityFile in a different order to revert this
1290 The sio(4) driver has been removed from the i386 and amd64
1291 kernel configuration files. This means uart(4) is now the
1292 default serial port driver on those platforms as well.
1294 To prevent collisions with the sio(4) driver, the uart(4) driver
1295 uses different names for its device nodes. This means the
1296 onboard serial port will now most likely be called "ttyu0"
1297 instead of "ttyd0". You may need to reconfigure applications to
1298 use the new device names.
1300 When using the serial port as a boot console, be sure to update
1301 /boot/device.hints and /etc/ttys before booting the new kernel.
1302 If you forget to do so, you can still manually specify the hints
1303 at the loader prompt:
1305 set hint.uart.0.at="isa"
1306 set hint.uart.0.port="0x3F8"
1307 set hint.uart.0.flags="0x10"
1308 set hint.uart.0.irq="4"
1312 The gpt(8) utility has been removed. Use gpart(8) to partition
1316 The version that Linuxulator emulates was changed from 2.4.2
1317 to 2.6.16. If you experience any problems with Linux binaries
1318 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1319 if it fixes the problem contact emulation mailing list.
1322 ISDN4BSD (I4B) was removed from the src tree. You may need to
1323 update a your kernel configuration and remove relevant entries.
1326 I have checked in code to support multiple routing tables.
1327 See the man pages setfib(1) and setfib(2).
1328 This is a hopefully backwards compatible version,
1329 but to make use of it you need to compile your kernel
1330 with options ROUTETABLES=2 (or more up to 16).
1333 The 802.11 wireless support was redone to enable multi-bss
1334 operation on devices that are capable. The underlying device
1335 is no longer used directly but instead wlanX devices are
1336 cloned with ifconfig. This requires changes to rc.conf files.
1337 For example, change:
1338 ifconfig_ath0="WPA DHCP"
1341 ifconfig_wlan0="WPA DHCP"
1342 see rc.conf(5) for more details. In addition, mergemaster of
1343 /etc/rc.d is highly recommended. Simultaneous update of userland
1344 and kernel wouldn't hurt either.
1346 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1347 modules were merged into the base wlan module. All references
1348 to these modules (e.g. in kernel config files) must be removed.
1351 psm(4) has gained write(2) support in native operation level.
1352 Arbitrary commands can be written to /dev/psm%d and status can
1353 be read back from it. Therefore, an application is responsible
1354 for status validation and error recovery. It is a no-op in
1355 other operation levels.
1358 Support for KSE threading has been removed from the kernel. To
1359 run legacy applications linked against KSE libmap.conf may
1360 be used. The following libmap.conf may be used to ensure
1361 compatibility with any prior release:
1363 libpthread.so.1 libthr.so.1
1364 libpthread.so.2 libthr.so.2
1365 libkse.so.3 libthr.so.3
1368 The layout of struct vmspace has changed. This affects libkvm
1369 and any executables that link against libkvm and use the
1370 kvm_getprocs() function. In particular, but not exclusively,
1371 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1372 The effects are minimal, but it's advisable to upgrade world
1376 The latest em driver no longer has support in it for the
1377 82575 adapter, this is now moved to the igb driver. The
1378 split was done to make new features that are incompatible
1379 with older hardware easier to do.
1382 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1383 likewise the kernel option is now GEOM_LINUX_LVM.
1386 The default NFS mount mode has changed from UDP to TCP for
1387 increased reliability. If you rely on (insecurely) NFS
1388 mounting across a firewall you may need to update your
1392 Belatedly note the addition of m_collapse for compacting
1396 The fts(3) structures have been changed to use adequate
1397 integer types for their members and so to be able to cope
1398 with huge file trees. The old fts(3) ABI is preserved
1399 through symbol versioning in libc, so third-party binaries
1400 using fts(3) should still work, although they will not take
1401 advantage of the extended types. At the same time, some
1402 third-party software might fail to build after this change
1403 due to unportable assumptions made in its source code about
1404 fts(3) structure members. Such software should be fixed
1405 by its vendor or, in the worst case, in the ports tree.
1406 FreeBSD_version 800015 marks this change for the unlikely
1407 case that a portable fix is impossible.
1410 To upgrade to -current after this date, you must be running
1411 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1412 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1415 The ADAPTIVE_GIANT kernel option has been retired because its
1416 functionality is the default now.
1419 The AT keyboard emulation of sunkbd(4) has been turned on
1420 by default. In order to make the special symbols of the Sun
1421 keyboards driven by sunkbd(4) work under X these now have
1422 to be configured the same way as Sun USB keyboards driven
1423 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1425 Option "XkbLayout" "us"
1426 Option "XkbRules" "xorg"
1427 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1430 It has been decided that it is desirable to provide ABI
1431 backwards compatibility to the FreeBSD 4/5/6 versions of the
1432 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1433 broken with the introduction of PCI domain support (see the
1434 20070930 entry). Unfortunately, this required the ABI of
1435 PCIOCGETCONF to be broken again in order to be able to
1436 provide backwards compatibility to the old version of that
1437 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1438 again. As for prominent ports this affects neither pciutils
1439 nor xorg-server this time, the hal port needs to be rebuilt
1443 The misnamed kthread_create() and friends have been renamed
1444 to kproc_create() etc. Many of the callers already
1445 used kproc_start()..
1446 I will return kthread_create() and friends in a while
1447 with implementations that actually create threads, not procs.
1448 Renaming corresponds with version 800002.
1457 Avoid using make -j when upgrading. While generally safe, there are
1458 sometimes problems using -j to upgrade. If your upgrade fails with
1459 -j, please try again without -j. From time to time in the past there
1460 have been problems using -j with buildworld and/or installworld. This
1461 is especially true when upgrading between "distant" versions (eg one
1462 that cross a major release boundary or several minor releases, or when
1463 several months have passed on the -current branch).
1465 Sometimes, obscure build problems are the result of environment
1466 poisoning. This can happen because the make utility reads its
1467 environment when searching for values for global variables. To run
1468 your build attempts in an "environmental clean room", prefix all make
1469 commands with 'env -i '. See the env(1) manual page for more details.
1471 When upgrading from one major version to another it is generally best
1472 to upgrade to the latest code in the currently installed branch first,
1473 then do an upgrade to the new branch. This is the best-tested upgrade
1474 path, and has the highest probability of being successful. Please try
1475 this approach before reporting problems with a major version upgrade.
1479 When upgrading the boot ZFS pool to a new version, always follow
1482 1.) recompile and reinstall the ZFS boot loader and boot block
1483 (this is part of "make buildworld" and "make installworld")
1485 2.) update the ZFS boot block on your boot drive
1487 The following example updates the ZFS boot block on the first
1488 partition (freebsd-boot) of a GPT partitioned drive ad0:
1489 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1491 Non-boot pools do not need these updates.
1495 If you are updating from a prior version of FreeBSD (even one just
1496 a few days old), you should follow this procedure. It is the most
1497 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1499 make kernel-toolchain
1500 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1501 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1503 To test a kernel once
1504 ---------------------
1505 If you just want to boot a kernel once (because you are not sure
1506 if it works, or if you want to boot a known bad kernel to provide
1507 debugging information) run
1508 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1509 nextboot -k testkernel
1511 To just build a kernel when you know that it won't mess you up
1512 --------------------------------------------------------------
1513 This assumes you are already running a CURRENT system. Replace
1514 ${arch} with the architecture of your machine (e.g. "i386",
1515 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1517 cd src/sys/${arch}/conf
1518 config KERNEL_NAME_HERE
1519 cd ../compile/KERNEL_NAME_HERE
1524 If this fails, go to the "To build a kernel" section.
1526 To rebuild everything and install it on the current system.
1527 -----------------------------------------------------------
1528 # Note: sometimes if you are running current you gotta do more than
1529 # is listed here if you are upgrading from a really old current.
1531 <make sure you have good level 0 dumps>
1533 make kernel KERNCONF=YOUR_KERNEL_HERE
1535 <reboot in single user> [3]
1543 To cross-install current onto a separate partition
1544 --------------------------------------------------
1545 # In this approach we use a separate partition to hold
1546 # current's root, 'usr', and 'var' directories. A partition
1547 # holding "/", "/usr" and "/var" should be about 2GB in
1550 <make sure you have good level 0 dumps>
1553 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1554 <maybe newfs current's root partition>
1555 <mount current's root partition on directory ${CURRENT_ROOT}>
1556 make installworld DESTDIR=${CURRENT_ROOT}
1557 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1558 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1559 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1560 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1561 <reboot into current>
1562 <do a "native" rebuild/install as described in the previous section>
1563 <maybe install compatibility libraries from ports/misc/compat*>
1567 To upgrade in-place from stable to current
1568 ----------------------------------------------
1569 <make sure you have good level 0 dumps>
1571 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1573 <reboot in single user> [3]
1580 Make sure that you've read the UPDATING file to understand the
1581 tweaks to various things you need. At this point in the life
1582 cycle of current, things change often and you are on your own
1583 to cope. The defaults can also change, so please read ALL of
1584 the UPDATING entries.
1586 Also, if you are tracking -current, you must be subscribed to
1587 freebsd-current@freebsd.org. Make sure that before you update
1588 your sources that you have read and understood all the recent
1589 messages there. If in doubt, please track -stable which has
1590 much fewer pitfalls.
1592 [1] If you have third party modules, such as vmware, you
1593 should disable them at this point so they don't crash your
1596 [3] From the bootblocks, boot -s, and then do
1601 adjkerntz -i # if CMOS is wall time
1602 Also, when doing a major release upgrade, it is required that
1603 you boot into single user mode to do the installworld.
1605 [4] Note: This step is non-optional. Failure to do this step
1606 can result in a significant reduction in the functionality of the
1607 system. Attempting to do it by hand is not recommended and those
1608 that pursue this avenue should read this file carefully, as well
1609 as the archives of freebsd-current and freebsd-hackers mailing lists
1610 for potential gotchas. The -U option is also useful to consider.
1611 See mergemaster(8) for more information.
1613 [5] Usually this step is a noop. However, from time to time
1614 you may need to do this if you get unknown user in the following
1615 step. It never hurts to do it all the time. You may need to
1616 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1617 install) after the buildworld before this step if you last updated
1618 from [78]-stable or 9-stable before 20130430.
1620 [6] This only deletes old files and directories. Old libraries
1621 can be deleted by "make delete-old-libs", but you have to make
1622 sure that no program is using those libraries anymore.
1624 [8] In order to have a kernel that can run the 4.x binaries needed to
1625 do an installworld, you must include the COMPAT_FREEBSD4 option in
1626 your kernel. Failure to do so may leave you with a system that is
1627 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1628 required to run the 5.x binaries on more recent kernels. And so on
1629 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1631 Make sure that you merge any new devices from GENERIC since the
1632 last time you updated your kernel config file.
1634 [9] When checking out sources, you must include the -P flag to have
1635 cvs prune empty directories.
1637 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1638 "?=" instead of the "=" assignment operator, so that buildworld can
1639 override the CPUTYPE if it needs to.
1641 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1642 not on the command line, or in /etc/make.conf. buildworld will
1643 warn if it is improperly defined.
1646 This file contains a list, in reverse chronological order, of major
1647 breakages in tracking -current. It is not guaranteed to be a complete
1648 list of such breakages, and only contains entries since October 10, 2007.
1649 If you need to see UPDATING entries from before that date, you will need
1650 to fetch an UPDATING file from an older FreeBSD release.
1652 Copyright information:
1654 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1656 Redistribution, publication, translation and use, with or without
1657 modification, in full or in part, in any form or format of this
1658 document are permitted without further permission from the author.
1660 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1661 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1662 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1663 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1664 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1665 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1666 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1667 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1668 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1669 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1670 POSSIBILITY OF SUCH DAMAGE.
1672 Contact Warner Losh if you have any questions about your use of