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 20131026: p1 FreeBSD-EN-13:04.freebsd-update
15 Fix multiple freebsd-update bugs that break upgrading to
19 hastctl(8)'s `status' command output changed to terse one-liner format.
20 Scripts using this should switch to `list' command or be rewritten.
23 Fix a bug that allowed a tracing process (e.g. gdb) to write
24 to a memory-mapped file in the traced process's address space
25 even if neither the traced process nor the tracing process had
26 write access to that file.
29 Added ZFS TRIM support which is enabled by default. To disable
30 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
32 Creating new ZFS pools and adding new devices to existing pools
33 first performs a full device level TRIM which can take a significant
34 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
35 to disable this behaviour.
37 ZFS TRIM requires the underlying device support BIO_DELETE which
38 is currently provided by methods such as ATA TRIM and SCSI UNMAP
39 via CAM, which are typically supported by SSD's.
41 Stats for ZFS TRIM can be monitored by looking at the sysctl's
42 under kstat.zfs.misc.zio_trim.
45 `list' command has been added to hastctl(8). For now, it is full
46 equivalent of `status' command.
47 WARNING: in the near future the output of hastctl's status command
48 will change to more terse format. If you use `hastctl status'
49 for parsing in your scripts, switch to `hastctl list'.
52 The mergemaster command now uses the default MAKEOBJDIRPREFIX
53 rather than creating it's own in the temporary directory in
54 order allow access to bootstrapped versions of tools such as
55 install and mtree. When upgrading from version of FreeBSD where
56 the install command does not support -l, you will need to
57 install a new mergemaster command if mergemaster -p is required.
58 This can be accomplished with the command (cd src/usr.sbin/mergemaster
61 Due to the use of the new -l option to install(1) during build
62 and install, you must take care not to directly set the INSTALL
63 make variable in your /etc/make.conf, /etc/src.conf, or on the
64 command line. If you wish to use the -C flag for all installs
65 you may be able to add INSTALL+=-C to /etc/make.conf or
69 Fix a bug that allows NFS clients to issue READDIR on files.
72 The install(1) option -M has changed meaning and now takes an
73 argument that is a file or path to append logs to. In the
74 unlikely event that -M was the last option on the command line
75 and the command line contained at least two files and a target
76 directory the first file will have logs appended to it. The -M
77 option served little practical purpose in the last decade so it's
78 used expected to be extremely rare.
81 A new compression method (lz4) has been merged to. Please refer to
82 zpool-features(7) for more information.
84 Please refer to the "ZFS notes" section of this file for information
85 on upgrading boot ZFS pools.
88 The VFS KBI was changed with the merge of several nullfs
89 optimizations and fixes. All filesystem modules must be
93 With the addition of auditdistd(8), a new auditdistd user is now
94 depended on during installworld. "mergemaster -p" can be used to add
95 the user prior to installworld, as documented in the handbook.
101 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
102 Starting with this version the old system of ZFS pool versioning
103 is superseded by "feature flags". This concept enables forward
104 compatibility against certain future changes in functionality of ZFS
105 pools. The first two read-only compatible "feature flags" for ZFS
106 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
107 For more information read the new zpool-features(7) manual page.
108 Please refer to the "ZFS notes" section of this file for information
109 on upgrading boot ZFS pools.
112 The commit introducing bsd.compiler.mk breaks the traditional
113 building of kernels before this point. Add -m ${SRC}/share/mk
114 (for the right value of SRC) to your command lines to work
115 around; update your useland to a point after this; or use the
116 buildkernel/installkernel top-level targets. See also 20120829.
119 The IPFIREWALL_FORWARD kernel option has been removed. Its
120 functionality now turned on by default.
123 The random(4) support for the VIA hardware random number
124 generator (`PADLOCK') is no longer enabled unconditionally.
125 Add the PADLOCK_RNG option in the custom kernel config if
126 needed. The GENERIC kernels on i386 and amd64 do include the
127 option, so the change only affects the custom kernel
131 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
132 the traditional method, you must update your system with an installworld
133 before the kernel will build. The documented make buildkernel/installkernel
134 interfaces (coupled with fresh make kernel-toolchain) continue to work.
137 The sparc64 ZFS loader has been changed to no longer try to auto-
138 detect ZFS providers based on diskN aliases but now requires these
139 to be explicitly listed in the OFW boot-device environment variable.
142 Now unix domain sockets behave "as expected" on nullfs(5). Previously
143 nullfs(5) did not pass through all behaviours to the underlying layer,
144 as a result if we bound to a socket on the lower layer we could connect
145 only to the lower path; if we bound to the upper layer we could connect
146 only to the upper path. The new behavior is one can connect to both the
147 lower and the upper paths regardless what layer path one binds to.
150 The acpi_wmi(4) status device /dev/wmistat has been renamed to
154 A new VOP_ADVISE() was added to support posix_fadvise(2). All
155 filesystem modules must be recompiled.
158 The interface of the VOP_VPTOCNP(9) changed, now the returned
159 vnode shall be referenced, previously it was required to be
160 only held. All in-tree filesystems are converted.
166 The broken amd(4) driver has been replaced with esp(4) in the amd64,
167 i386 and pc98 GENERIC kernel configuration files.
170 This commit modifies vfs_register() so that it uses a hash
171 calculation to set vfc_typenum, which is enabled by default.
172 The first time a system is booted after this change, the
173 vfc_typenum values will change for all file systems. The
174 main effect of this is a change to the NFS server file handles
175 for file systems that use vfc_typenum in their fsid, such as ZFS.
176 It will, however, prevent vfc_typenum from changing when file
177 systems are loaded in a different order for subsequent reboots.
178 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
179 until you are ready to remount all NFS clients after a reboot.
182 Bump the shared library version numbers for libraries that
183 do not use symbol versioning, have changed the ABI compared
184 to stable/8 and which shared library version was not bumped.
185 Done as part of 9.0-RELEASE cycle.
188 During the merge of Capsicum features, the fget(9) KPI was modified.
189 This may require the rebuilding of out-of-tree device drivers --
190 issues have been reported specifically with the nVidia device driver.
191 __FreeBSD_version is bumped to 900041.
193 Also, there is a period between 20110811 and 20110814 where the
194 special devices /dev/{stdin,stdout,stderr} did not work correctly.
195 Building world from a kernel during that window may not work.
198 The packet filter (pf) code has been updated to OpenBSD 4.5.
199 You need to update userland tools to be in sync with kernel.
200 This update breaks backward compatibility with earlier pfsync(4)
201 versions. Care must be taken when updating redundant firewall setups.
204 The following sysctls and tunables are retired on x86 platforms:
206 machdep.hlt_logical_cpus
207 The following sysctl is retired:
208 machdep.hyperthreading_allowed
209 The sysctls were supposed to provide a way to dynamically offline and
210 online selected CPUs on x86 platforms, but the implementation has not
211 been reliable especially with SCHED_ULE scheduler.
212 machdep.hyperthreading_allowed tunable is still available to ignore
213 hyperthreading CPUs at OS level.
214 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
215 where X is an APIC ID of a CPU. Be advised, though, that disabling
216 CPUs in non-uniform fashion will result in non-uniform topology and
217 may lead to sub-optimal system performance with SCHED_ULE, which is
221 cpumask_t type is retired and cpuset_t is used in order to describe
225 Changes to ifconfig(8) for dynamic address family detection mandate
226 that you are running a kernel of 20110525 or later. Make sure to
227 follow the update procedure to boot a new kernel before installing
231 Support for sun4v architecture is officially dropped
234 Several KPI breaking changes have been committed to the mii(4) layer,
235 the PHY drivers and consequently some Ethernet drivers using mii(4).
236 This means that miibus.ko and the modules of the affected Ethernet
237 drivers need to be recompiled.
239 Note to kernel developers: Given that the OUI bit reversion problem
240 was fixed as part of these changes all mii(4) commits related to OUIs,
241 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
242 handling, no longer can be merged verbatim to stable/8 and previous
246 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
247 into their kernel configurations along with 'device pci'.
250 The default NFS client is now the new NFS client, so fstype "newnfs"
251 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
252 Although mounts via fstype "nfs" will usually work without userland
253 changes, it is recommended that the mount(8) and mount_nfs(8)
254 commands be rebuilt from sources and that a link to mount_nfs called
255 mount_oldnfs be created. The new client is compiled into the
256 kernel with "options NFSCL" and this is needed for diskless root
257 file systems. The GENERIC kernel configs have been changed to use
258 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
259 To use the regular/old client, you can "mount -t oldnfs ...". For
260 a diskless root file system, you must also include a line like:
262 vfs.root.mountfrom="oldnfs:"
264 in the boot/loader.conf on the root fs on the NFS server to make
265 a diskless root fs use the old client.
268 The GENERIC kernels for all architectures now default to the new
269 CAM-based ATA stack. It means that all legacy ATA drivers were
270 removed and replaced by respective CAM drivers. If you are using
271 ATA device names in /etc/fstab or other places, make sure to update
272 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
273 where 'Y's are the sequential numbers starting from zero for each type
274 in order of detection, unless configured otherwise with tunables,
275 see cam(4)). There will be symbolic links created in /dev/ to map
276 old adX devices to the respective adaY. They should provide basic
277 compatibility for file systems mounting in most cases, but they do
278 not support old user-level APIs and do not have respective providers
279 in GEOM. Consider using updated management tools with new device names.
281 It is possible to load devices ahci, ata, siis and mvs as modules,
282 but option ATA_CAM should remain in kernel configuration to make ata
283 module work as CAM driver supporting legacy ATA controllers. Device ata
284 still can be used in modular fashion (atacore + ...). Modules atadisk
285 and atapi* are not used and won't affect operation in ATA_CAM mode.
286 Note that to use CAM-based ATA kernel should include CAM devices
287 scbus, pass, da (or explicitly ada), cd and optionally others. All of
288 them are parts of the cam module.
290 ataraid(4) functionality is now supported by the RAID GEOM class.
291 To use it you can load geom_raid kernel module and use graid(8) tool
292 for management. Instead of /dev/arX device names, use /dev/raid/rX.
294 No kernel config options or code have been removed, so if a problem
295 arises, please report it and optionally revert to the old ATA stack.
296 In order to do it you can remove from the kernel config:
301 , and instead add back:
302 device atadisk # ATA disk drives
303 device ataraid # ATA RAID drives
304 device atapicd # ATAPI CDROM drives
305 device atapifd # ATAPI floppy drives
306 device atapist # ATAPI tape drives
309 The default NFS server has been changed to the new server, which
310 was referred to as the experimental server. If you need to switch
311 back to the old NFS server, you must now put the "-o" option on
312 both the mountd and nfsd commands. This can be done using the
313 mountd_flags and nfs_server_flags rc.conf variables until an
314 update to the rc scripts is committed, which is coming soon.
317 The GNU Objective-C runtime library (libobjc), and other Objective-C
318 related components have been removed from the base system. If you
319 require an Objective-C library, please use one of the available ports.
322 ath(4) has been split into bus- and device- modules. if_ath contains
323 the HAL, the TX rate control and the network device code. if_ath_pci
324 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
325 contains the AHB glue. Users need to load both if_ath_pci and if_ath
326 in order to use ath on everything else.
328 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
329 need to load if_ath and if_ath_pci for ath(4) operation.
332 As part of the replacement of sysinstall, the process of building
333 release media has changed significantly. For details, please re-read
334 release(7), which has been updated to reflect the new build process.
337 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
338 is the last available version under GPLv2. It brings a number of new
339 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
340 4.1 and SSE 4.2), better support for powerpc64, a number of new
341 directives, and lots of other small improvements. See the ChangeLog
342 file in contrib/binutils for the full details.
345 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
346 compliant, and will now use half of hash for authentication.
347 This will break interoperability with all stacks (including all
348 actual FreeBSD versions) who implement
349 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
351 The only workaround with such peers is to use another HMAC
352 algorithm for IPsec ("phase 2") authentication.
355 Remove the uio_yield prototype and symbol. This function has
356 been misnamed since it was introduced and should not be
357 globally exposed with this name. The equivalent functionality
358 is now available using kern_yield(curthread->td_user_pri).
359 The function remains undocumented.
362 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
363 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
364 for scalar sysctls is defined but disabled. Code that needs
365 UQUAD to pass the type checking that must compile on older
366 systems where the define is not present can check against
367 __FreeBSD_version >= 900030.
369 The system dialog(1) has been replaced with a new version previously
370 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
371 with the previous version, but the libdialog associated with it has
372 a largely incompatible API. As such, the original version of libdialog
373 will be kept temporarily as libodialog, until its base system consumers
374 are replaced or updated. Bump __FreeBSD_version to 900030.
377 If you are trying to run make universe on a -stable system, and you get
378 the following warning:
379 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
380 unknown. config(8) likely too old."
381 or something similar to it, then you must upgrade your -stable system
382 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
383 stable/8) or build the config from the latest stable/8 branch and
384 install it on your system.
386 Prior to this date, building a current universe on 8-stable system from
387 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
388 error in the first kernel build phase. A new config on those old
389 systems will fix that problem for older versions of -current.
392 The TCP stack has been modified to allow Khelp modules to interact with
393 it via helper hook points and store per-connection data in the TCP
394 control block. Bump __FreeBSD_version to 900029. User space tools that
395 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
399 Generic IEEE 802.3 annex 31B full duplex flow control support has been
400 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
401 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
402 to take advantage of it instead of using custom implementations. This
403 means that these drivers now no longer unconditionally advertise
404 support for flow control but only do so if flow control is a selected
405 media option. This was implemented in the generic support that way in
406 order to allow flow control to be switched on and off via ifconfig(8)
407 with the PHY specific default to typically off in order to protect
408 from unwanted effects. Consequently, if you used flow control with
409 one of the above mentioned drivers you now need to explicitly enable
411 ifconfig bge0 media auto mediaopt flowcontrol
413 Along with the above mentioned changes generic support for setting
414 1000baseT master mode also has been added and brgphy(4), ciphy(4),
415 e1000phy(4) as well as ip1000phy(4) have been converted to take
416 advantage of it. This means that these drivers now no longer take the
417 link0 parameter for selecting master mode but the master media option
418 has to be used instead, for example like in the following:
419 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
421 Selection of master mode now is also available with all other PHY
422 drivers supporting 1000baseT.
425 The TCP stack has received a significant update to add support for
426 modularised congestion control and generally improve the clarity of
427 congestion control decisions. Bump __FreeBSD_version to 900025. User
428 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
429 sockstat) need to be recompiled.
432 The man(1) utility has been replaced by a new version that no longer
433 uses /etc/manpath.config. Please consult man.conf(5) for how to
434 migrate local entries to the new format.
437 The copyright strings printed by login(1) and sshd(8) at the time of a
438 new connection have been removed to follow other operating systems and
442 A workaround for a fixed ld bug has been removed in kernel code,
443 so make sure that your system ld is built from sources after
444 revision 210245 from 2010-07-19 (r211583 if building head kernel
445 on stable/8, r211584 for stable/7; both from 2010-08-21).
446 A symptom of incorrect ld version is different addresses for
447 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
450 The $ipv6_prefer variable in rc.conf(5) has been split into
451 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
453 The $ip6addrctl_policy is a variable to choose a pre-defined
454 address selection policy set by ip6addrctl(8). A value
455 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
458 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
459 flag (see an entry of 20090926) is set on an interface with no
460 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
461 security reason. If you want IPv6 link-local address on all
462 interfaces by default, set this to "YES".
464 The old ipv6_prefer="YES" is equivalent to
465 ipv6_activate_all_interfaces="YES" and
466 ip6addrctl_policy="ipv6_prefer".
469 DTrace has grown support for userland tracing. Due to this, DTrace is
470 now i386 and amd64 only.
471 dtruss(1) is now installed by default on those systems and a new
472 kernel module is needed for userland tracing: fasttrap.
473 No changes to your kernel config file are necessary to enable
474 userland tracing, but you might consider adding 'STRIP=' and
475 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
476 to have informative userland stack traces in DTrace (ustack).
479 The acpi_aiboost(4) driver has been removed in favor of the new
480 aibs(4) driver. You should update your kernel configuration file.
483 BSD grep has been imported to the base system and it is built by
484 default. It is completely BSD licensed, highly GNU-compatible, uses
485 less memory than its GNU counterpart and has a small codebase.
486 However, it is slower than its GNU counterpart, which is mostly
487 noticeable for larger searches, for smaller ones it is measurable
488 but not significant. The reason is complex, the most important factor
489 is that we lack a modern and efficient regex library and GNU
490 overcomes this by optimizing the searches internally. Future work
491 on improving the regex performance is planned, for the meantime,
492 users that need better performance, can build GNU grep instead by
493 setting the WITH_GNU_GREP knob.
496 Due to the import of powerpc64 support, all existing powerpc kernel
497 configuration files must be updated with a machine directive like this:
498 machine powerpc powerpc
500 In addition, an updated config(8) is required to build powerpc kernels
504 A new version of ZFS (version 15) has been merged to -HEAD.
505 This version uses a python library for the following subcommands:
506 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
507 For full functionality of these commands the following port must
508 be installed: sysutils/py-zfs
511 'vm_page's are now hashed by physical address to an array of mutexes.
512 Currently this is only used to serialize access to hold_count. Over
513 time the page queue mutex will be peeled away. This changes the size
514 of pmap on every architecture. And requires all callers of vm_page_hold
515 and vm_page_unhold to be updated.
518 WITH_CTF can now be specified in src.conf (not recommended, there
519 are some problems with static executables), make.conf (would also
520 affect ports which do not use GNU make and do not override the
521 compile targets) or in the kernel config (via "makeoptions
523 When WITH_CTF was specified there before this was silently ignored,
524 so make sure that WITH_CTF is not used in places which could lead
525 to unwanted behavior.
528 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
529 to allow 32-bit compatibility on non-x86 platforms. All kernel
530 configurations on amd64 and ia64 platforms using these options must
531 be modified accordingly.
534 The utmp user accounting database has been replaced with utmpx,
535 the user accounting interface standardized by POSIX.
536 Unfortunately the semantics of utmp and utmpx don't match,
537 making it practically impossible to support both interfaces.
538 The user accounting database is used by tools like finger(1),
539 last(1), talk(1), w(1) and ac(8).
541 All applications in the base system use utmpx. This means only
542 local binaries (e.g. from the ports tree) may still use these
543 utmp database files. These applications must be rebuilt to make
546 After the system has been upgraded, it is safe to remove the old
547 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
548 assuming their contents is of no importance anymore. Old wtmp
549 databases can only be used by last(1) and ac(8) after they have
550 been converted to the new format using wtmpcvt(1).
553 Introduce the kernel thread "deadlock resolver" (which can be enabled
554 via the DEADLKRES option, see NOTES for more details) and the
555 sleepq_type() function for sleepqueues.
558 The rc.firewall and rc.firewall6 were unified, and
559 rc.firewall6 and rc.d/ip6fw were removed.
560 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
561 variables are obsoleted. Instead, the following new rc
562 variables are added to rc.d/ipfw:
564 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
565 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
566 firewall_simple_onet_ipv6, firewall_trusted_ipv6
568 The meanings correspond to the relevant IPv4 variables.
574 The default terminal emulation for syscons(4) has been changed
575 from cons25 to xterm on all platforms except pc98. This means
576 that the /etc/ttys file needs to be updated to ensure correct
577 operation of applications on the console.
579 The terminal emulation style can be toggled per window by using
580 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
581 options can be used to change the compile-time default back to
584 To prevent graphical artifacts, make sure the TERM environment
585 variable is set to match the terminal emulation that is being
586 performed by syscons(4).
589 The layout of the structure ieee80211req_scan_result has changed.
590 Applications that require wireless scan results (e.g. ifconfig(8))
591 from net80211 need to be recompiled.
593 Applications such as wpa_supplicant(8) may require a full world
594 build without using NO_CLEAN in order to get synchronized with the
598 The iwn(4) driver has been updated to support the 5000 and 5150 series.
599 There's one kernel module for each firmware. Adding "device iwnfw"
600 to the kernel configuration file means including all three firmware
601 images inside the kernel. If you want to include just the one for
602 your wireless card, use the devices iwn4965fw, iwn5000fw or
606 The rc.d/network_ipv6, IPv6 configuration script has been integrated
607 into rc.d/netif. The changes are the following:
609 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
610 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
611 Note that both variables need the "inet6" keyword at the head.
613 Do not set $ipv6_network_interfaces manually if you do not
614 understand what you are doing. It is not needed in most cases.
616 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
619 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
620 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
622 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
623 all configured interfaces have "inet6 accept_rtadv" in the
624 $ifconfig_IF_ipv6. These are for backward compatibility.
626 3. A new variable $ipv6_prefer has been added. If NO, IPv6
627 functionality of interfaces with no corresponding
628 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
629 and the default address selection policy of ip6addrctl(8)
630 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
631 Note that if you want to configure IPv6 functionality on the
632 disabled interfaces after boot, first you need to clear the flag by
633 using ifconfig(8) like:
635 ifconfig em0 inet6 -ifdisabled
637 If YES, the default address selection policy is set as
640 The default value of $ipv6_prefer is NO.
642 4. If your system need to receive Router Advertisement messages,
643 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
644 scripts automatically invoke rtsol(8) when the interface becomes
645 UP. The Router Advertisement messages are used for SLAAC
646 (State-Less Address AutoConfiguration).
649 802.11s D3.03 support was committed. This is incompatible with the
650 previous code, which was based on D3.0.
653 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
654 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
655 control whether accepting Router Advertisement messages or not.
656 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
657 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
658 The ifconfig(8) utility now supports these flags.
661 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
662 mount(8) and -a option for df(1) to see them.
665 The old tunable hw.bus.devctl_disable has been superseded by
666 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
667 replaced by hw.bus.devctl_queue=0. The default for this new tunable
671 Remove the option STOP_NMI. The default action is now to use NMI only
672 for KDB via the newly introduced function stop_cpus_hard() and
673 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
676 The stable/8 branch created in subversion. This corresponds to the
677 RELENG_8 branch in CVS.
680 Bump the shared library version numbers for all libraries that do not
681 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
682 __FreeBSD_version to 800105.
685 Due to changes in the implementation of virtual network stack support,
686 all network-related kernel modules must be recompiled. As this change
687 breaks the ABI, bump __FreeBSD_version to 800104.
690 The TOE interface to the TCP syncache has been modified to remove
691 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
692 The cxgb driver is the only TOE consumer affected by this change, and
693 needs to be recompiled along with the kernel. As this change breaks
694 the ABI, bump __FreeBSD_version to 800103.
697 Padding has been added to struct tcpcb, sackhint and tcpstat in
698 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
699 maintaining the ABI. However, this change breaks the ABI, so bump
700 __FreeBSD_version to 800102. User space tools that rely on the size of
701 any of these structs (e.g. sockstat) need to be recompiled.
704 The NFS_LEGACYRPC option has been removed along with the old kernel
705 RPC implementation that this option selected. Kernel configurations
706 may need to be adjusted.
709 The network interface device nodes at /dev/net/<interface> have been
710 removed. All ioctl operations can be performed the normal way using
711 routing sockets. The kqueue functionality can generally be replaced
712 with routing sockets.
715 The documentation from the FreeBSD Documentation Project (Handbook,
716 FAQ, etc.) is now installed via packages by sysinstall(8) and under
717 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
720 The ABI of various structures related to the SYSV IPC API have been
721 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
722 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
726 Layout of struct vnet has changed as routing related variables were
727 moved to their own Vimage module. Modules need to be recompiled. Bump
728 __FreeBSD_version to 800099.
731 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
732 respectively. As long as no more than 16 groups per process are used,
733 no changes should be visible. When more than 16 groups are used, old
734 binaries may fail if they call getgroups() or getgrouplist() with
735 statically sized storage. Recompiling will work around this, but
736 applications should be modified to use dynamically allocated storage
737 for group arrays as POSIX.1-2008 does not cap an implementation's
738 number of supported groups at NGROUPS_MAX+1 as previous versions did.
740 NFS and portalfs mounts may also be affected as the list of groups is
741 truncated to 16. Users of NFS who use more than 16 groups, should
742 take care that negative group permissions are not used on the exported
743 file systems as they will not be reliable unless a GSSAPI based
744 authentication method is used.
747 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
748 option compiles in the support for adaptive spinning for lockmgrs
749 which want to enable it. The lockinit() function now accepts the flag
750 LK_ADAPTIVE in order to make the lock object subject to adaptive
751 spinning when both held in write and read mode.
754 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
755 changed. User applications that use this ioctl need to be rebuilt.
758 The layout of struct thread has changed. Kernel and modules need to
762 The layout of structs ifnet, domain, protosw and vnet_net has changed.
763 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
766 window(1) has been removed from the base system. It can now be
767 installed from ports. The port is called misc/window.
770 The way we are storing and accessing `routing table' entries has
771 changed. Programs reading the FIB, like netstat, need to be
775 A new netisr implementation has been added for FreeBSD 8. Network
776 file system modules, such as igmp, ipdivert, and others, should be
778 Bump __FreeBSD_version to 800096.
781 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
785 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
786 Bump __FreeBSD_version to 800094.
789 Add mnt_xflag field to 'struct mount'. File system modules need to be
791 Bump __FreeBSD_version to 800093.
794 The compiling option ADAPTIVE_SX has been retired while it has been
795 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
796 The KPI for sx_init_flags() changes as accepting flags:
797 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
798 been introduced in order to handle the reversed logic.
799 Bump __FreeBSD_version to 800092.
802 Add support for hierarchical jails. Remove global securelevel.
803 Bump __FreeBSD_version to 800091.
806 The layout of struct vnet_net has changed, therefore modules
808 Bump __FreeBSD_version to 800090.
811 The newly imported zic(8) produces a new format in the output. Please
812 run tzsetup(8) to install the newly created data to /etc/localtime.
815 The sysctl tree for the usb stack has renamed from hw.usb2.* to
816 hw.usb.* and is now consistent again with previous releases.
819 802.11 monitor mode support was revised and driver api's were changed.
820 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
821 of DLT_IEEE802_11. No user-visible data structures were changed but
822 applications that use DLT_IEEE802_11 may require changes.
823 Bump __FreeBSD_version to 800088.
826 The layout of the following structs has changed: sysctl_oid,
827 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
828 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
829 panics may be experienced. World rebuild is required for
830 correctly checking networking state from userland.
831 Bump __FreeBSD_version to 800085.
834 MLDv2 and Source-Specific Multicast (SSM) have been merged
835 to the IPv6 stack. VIMAGE hooks are in but not yet used.
836 The implementation of SSM within FreeBSD's IPv6 stack closely
837 follows the IPv4 implementation.
839 For kernel developers:
841 * The most important changes are that the ip6_output() and
842 ip6_input() paths no longer take the IN6_MULTI_LOCK,
843 and this lock has been downgraded to a non-recursive mutex.
845 * As with the changes to the IPv4 stack to support SSM, filtering
846 of inbound multicast traffic must now be performed by transport
847 protocols within the IPv6 stack. This does not apply to TCP and
848 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
850 * The KPIs used by IPv6 multicast are similar to those used by
851 the IPv4 stack, with the following differences:
852 * im6o_mc_filter() is analogous to imo_multicast_filter().
853 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
854 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
855 * IN6_LOOKUP_MULTI() has been deprecated and removed.
856 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
857 for MLDv1 have an additional 'timer' argument which is used to
858 jitter the initial membership report for the solicited-node
859 multicast membership on-link.
860 * This is not strictly needed for MLDv2, which already jitters
861 its report transmissions. However, the 'timer' argument is
862 preserved in case MLDv1 is active on the interface.
864 * The KAME linked-list based IPv6 membership implementation has
865 been refactored to use a vector similar to that used by the IPv4
867 Code which maintains a list of its own multicast memberships
868 internally, e.g. carp, has been updated to reflect the new
871 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
872 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
873 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
874 implementation constraint which needs to be addressed in HEAD.
876 For application developers:
878 * The changes are broadly similar to those made for the IPv4
881 * The use of IPv4 and IPv6 multicast socket options on the same
882 socket, using mapped addresses, HAS NOT been tested or supported.
884 * There are a number of issues with the implementation of various
885 IPv6 multicast APIs which need to be resolved in the API surface
886 before the implementation is fully compatible with KAME userland
887 use, and these are mostly to do with interface index treatment.
889 * The literature available discusses the use of either the delta / ASM
890 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
891 using setsourcefilter(3)/getsourcefilter(3). For more information
892 please refer to RFC 3768, 'Socket Interface Extensions for
893 Multicast Source Filters'.
895 * Applications which use the published RFC 3678 APIs should be fine.
897 For systems administrators:
899 * The mtest(8) utility has been refactored to support IPv6, in
900 addition to IPv4. Interface addresses are no longer accepted
901 as arguments, their names must be used instead. The utility
902 will map the interface name to its first IPv4 address as
903 returned by getifaddrs(3).
905 * The ifmcstat(8) utility has also been updated to print the MLDv2
906 endpoint state and source filter lists via sysctl(3).
908 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
909 loopback of IPv6 multicast datagrams by default; it defaults to 1
910 to preserve the existing behaviour. Disabling multicast loopback is
911 recommended for optimal system performance.
913 * The IPv6 MROUTING code has been changed to examine this sysctl
914 instead of attempting to perform a group lookup before looping
915 back forwarded datagrams.
917 Bump __FreeBSD_version to 800084.
920 Implement low-level Bluetooth HCI API.
921 Bump __FreeBSD_version to 800083.
924 The layout of struct malloc_type, used by modules to register new
925 memory allocation types, has changed. Most modules will need to
926 be rebuilt or panics may be experienced.
927 Bump __FreeBSD_version to 800081.
930 Anticipate overflowing inp_flags - add inp_flags2.
931 This changes most offsets in inpcb, so checking v4 connection
932 state will require a world rebuild.
933 Bump __FreeBSD_version to 800080.
936 Add an llentry to struct route and struct route_in6. Modules
937 embedding a struct route will need to be recompiled.
938 Bump __FreeBSD_version to 800079.
941 The size of rt_metrics_lite and by extension rtentry has changed.
942 Networking administration apps will need to be recompiled.
943 The route command now supports show as an alias for get, weighting
944 of routes, sticky and nostick flags to alter the behavior of stateful
946 Bump __FreeBSD_version to 800078.
949 Do not use Giant for kbdmux(4) locking. This is wrong and
950 apparently causing more problems than it solves. This will
951 re-open the issue where interrupt handlers may race with
952 kbdmux(4) in polling mode. Typical symptoms include (but
953 not limited to) duplicated and/or missing characters when
954 low level console functions (such as gets) are used while
955 interrupts are enabled (for example geli password prompt,
956 mountroot prompt etc.). Disabling kbdmux(4) may help.
959 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
960 kernel modules referencing any of the above need to be recompiled.
961 Bump __FreeBSD_version to 800075.
964 GEOM_PART has become the default partition slicer for storage devices,
965 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
966 introduces some changes:
968 MSDOS/EBR: the devices created from MSDOS extended partition entries
969 (EBR) can be named differently than with GEOM_MBR and are now symlinks
970 to devices with offset-based names. fstabs may need to be modified.
972 BSD: the "geometry does not match label" warning is harmless in most
973 cases but it points to problems in file system misalignment with
974 disk geometry. The "c" partition is now implicit, covers the whole
975 top-level drive and cannot be (mis)used by users.
977 General: Kernel dumps are now not allowed to be written to devices
978 whose partition types indicate they are meant to be used for file
979 systems (or, in case of MSDOS partitions, as something else than
982 Most of these changes date approximately from 200812.
985 The uscanner(4) driver has been removed from the kernel. This follows
986 Linux removing theirs in 2.6 and making libusb the default interface
990 The multicast forwarding code has been cleaned up. netstat(1)
991 only relies on KVM now for printing bandwidth upcall meters.
992 The IPv4 and IPv6 modules are split into ip_mroute_mod and
993 ip6_mroute_mod respectively. The config(5) options for statically
994 compiling this code remain the same, i.e. 'options MROUTING'.
997 Support for the IFF_NEEDSGIANT network interface flag has been
998 removed, which means that non-MPSAFE network device drivers are no
999 longer supported. In particular, if_ar, if_sr, and network device
1000 drivers from the old (legacy) USB stack can no longer be built or
1004 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1005 a bunch of new language catalog files have also been added.
1006 This means that some common libc messages are now localized and
1007 they depend on the LC_MESSAGES environmental variable.
1010 The k8temp(4) driver has been renamed to amdtemp(4) since
1011 support for Family 10 and Family 11 CPU families was added.
1014 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1015 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1017 For kernel developers, the most important changes are that the
1018 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1019 and this lock has been downgraded to a non-recursive mutex.
1021 Transport protocols (UDP, Raw IP) are now responsible for filtering
1022 inbound multicast traffic according to group membership and source
1023 filters. The imo_multicast_filter() KPI exists for this purpose.
1024 Transports which do not use multicast (SCTP, TCP) already reject
1025 multicast by default. Forwarding and receive performance may improve
1026 as a mutex acquisition is no longer needed in the ip_input()
1027 low-level input path. in_addmulti() and in_delmulti() are shimmed
1028 to new KPIs which exist to support SSM in-kernel.
1030 For application developers, it is recommended that loopback of
1031 multicast datagrams be disabled for best performance, as this
1032 will still cause the lock to be taken for each looped-back
1033 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1034 be tuned to 0 to disable loopback by default; it defaults to 1
1035 to preserve the existing behaviour.
1037 For systems administrators, to obtain best performance with
1038 multicast reception and multiple groups, it is always recommended
1039 that a card with a suitably precise hash filter is used. Hash
1040 collisions will still result in the lock being taken within the
1041 transport protocol input path to check group membership.
1043 If deploying FreeBSD in an environment with IGMP snooping switches,
1044 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1045 enabled; this forces 224.0.0.0/24 group membership to be announced
1048 The size of 'struct igmpstat' has changed; netstat needs to be
1049 recompiled to reflect this.
1050 Bump __FreeBSD_version to 800070.
1053 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1054 updated to use it. This requires a buildworld/installworld in order to
1055 update the library and dependencies (usbconfig, etc). Its advisable to
1056 rebuild all ports which uses libusb. More specific directions are given
1057 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1058 libusb are no longer required and can be removed.
1061 A workaround is committed to allow the creation of System V shared
1062 memory segment of size > 2 GB on the 64-bit architectures.
1063 Due to a limitation of the existing ABI, the shm_segsz member
1064 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1065 wrong for large segments. Note that limits must be explicitly
1066 raised to allow such segments to be created.
1069 The layout of struct ifnet has changed, requiring a rebuild of all
1070 network device driver modules.
1073 The /dev handling for the new USB stack has changed, a
1074 buildworld/installworld is required for libusb20.
1077 The new USB2 stack has now been permanently moved in and all kernel and
1078 module names reverted to their previous values (eg, usb, ehci, ohci,
1079 ums, ...). The old usb stack can be compiled in by prefixing the name
1080 with the letter 'o', the old usb modules have been removed.
1081 Updating entry 20090216 for xorg and 20090215 for libmap may still
1085 The rc.conf(5) option if_up_delay has been renamed to
1086 defaultroute_delay to better reflect its purpose. If you have
1087 customized this setting in /etc/rc.conf you need to update it to
1091 xorg 7.4 wants to configure its input devices via hald which does not
1092 yet work with USB2. If the keyboard/mouse does not work in xorg then
1094 Option "AllowEmptyInput" "off"
1095 to your ServerLayout section. This will cause X to use the configured
1096 kbd and mouse sections from your xorg.conf.
1099 The GENERIC kernels for all architectures now default to the new USB2
1100 stack. No kernel config options or code have been removed so if a
1101 problem arises please report it and optionally revert to the old USB
1102 stack. If you are loading USB kernel modules or have a custom kernel
1103 that includes GENERIC then ensure that usb names are also changed over,
1104 eg uftdi -> usb2_serial_ftdi.
1106 Older programs linked against the ports libusb 0.1 need to be
1107 redirected to the new stack's libusb20. /etc/libmap.conf can
1109 # Map old usb library to new one for usb2 stack
1110 libusb-0.1.so.8 libusb20.so.1
1113 All USB ethernet devices now attach as interfaces under the name ueN
1114 (eg. ue0). This is to provide a predictable name as vendors often
1115 change usb chipsets in a product without notice.
1118 The ichsmb(4) driver has been changed to require SMBus slave
1119 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1120 All of the other SMBus controller drivers require left-justified
1121 slave addresses, so this change makes all the drivers provide the
1125 INET6 statistics (struct ip6stat) was updated.
1126 netstat(1) needs to be recompiled.
1129 NTFS has been removed from GENERIC kernel on amd64 to match
1130 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1131 will load ntfs.ko module automatically when NTFS support is
1132 actually needed, unless ntfs.ko is not installed or security
1133 level prohibits loading kernel modules. If either is the case,
1134 "options NTFS" has to be added into kernel config.
1137 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1138 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1139 800061. User space tools that rely on the size of struct tcpcb in
1140 tcp_var.h (e.g. sockstat) need to be recompiled.
1143 ng_tty(4) module updated to match the new TTY subsystem.
1144 Due to API change, user-level applications must be updated.
1145 New API support added to mpd5 CVS and expected to be present
1146 in next mpd5.3 release.
1149 With __FreeBSD_version 800060 the makefs tool is part of
1150 the base system (it was a port).
1153 The afdata and ifnet locks have been changed from mutexes to
1154 rwlocks, network modules will need to be re-compiled.
1157 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1158 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1159 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1160 architecture and 40 bytes on 64-bit architecture. The userland
1161 applications "arp" and "ndp" have been updated accordingly.
1162 The output from "netstat -r" shows only routing entries and
1163 none of the L2 information.
1166 __FreeBSD_version 800057 marks the switchover from the
1167 binary ath hal to source code. Users must add the line:
1169 options AH_SUPPORT_AR5416
1171 to their kernel config files when specifying:
1175 The ath_hal module no longer exists; the code is now compiled
1176 together with the driver in the ath module. It is now
1177 possible to tailor chip support (i.e. reduce the set of chips
1178 and thereby the code size); consult ath_hal(4) for details.
1181 __FreeBSD_version 800054 adds memory barriers to
1182 <machine/atomic.h>, new interfaces to ifnet to facilitate
1183 multiple hardware transmit queues for cards that support
1184 them, and a lock-less ring-buffer implementation to
1185 enable drivers to more efficiently manage queueing of
1189 A new version of ZFS (version 13) has been merged to -HEAD.
1190 This version has zpool attribute "listsnapshots" off by
1191 default, which means "zfs list" does not show snapshots,
1192 and is the same as Solaris behavior.
1195 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1198 The uhci, ohci, ehci and slhci USB Host controller drivers have
1199 been put into separate modules. If you load the usb module
1200 separately through loader.conf you will need to load the
1201 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1202 controller add the following to loader.conf:
1208 The ABI used by the PMC toolset has changed. Please keep
1209 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1213 atapci kernel module now includes only generic PCI ATA
1214 driver. AHCI driver moved to ataahci kernel module.
1215 All vendor-specific code moved into separate kernel modules:
1216 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1217 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1218 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1219 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1222 The TTY subsystem of the kernel has been replaced by a new
1223 implementation, which provides better scalability and an
1224 improved driver model. Most common drivers have been migrated to
1225 the new TTY subsystem, while others have not. The following
1226 drivers have not yet been ported to the new TTY layer:
1229 cy, digi, rc, rp, sio
1235 ng_h4, ng_tty, ppp, sl, snp
1237 Adding these drivers to your kernel configuration file shall
1238 cause compilation to fail.
1241 ntpd has been upgraded to 4.2.4p5.
1244 OpenSSH has been upgraded to 5.1p1.
1246 For many years, FreeBSD's version of OpenSSH preferred DSA
1247 over RSA for host and user authentication keys. With this
1248 upgrade, we've switched to the vendor's default of RSA over
1249 DSA. This may cause upgraded clients to warn about unknown
1250 host keys even for previously known hosts. Users should
1251 follow the usual procedure for verifying host keys before
1252 accepting the RSA key.
1254 This can be circumvented by setting the "HostKeyAlgorithms"
1255 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1258 Please note that the sequence of keys offered for
1259 authentication has been changed as well. You may want to
1260 specify IdentityFile in a different order to revert this
1264 The sio(4) driver has been removed from the i386 and amd64
1265 kernel configuration files. This means uart(4) is now the
1266 default serial port driver on those platforms as well.
1268 To prevent collisions with the sio(4) driver, the uart(4) driver
1269 uses different names for its device nodes. This means the
1270 onboard serial port will now most likely be called "ttyu0"
1271 instead of "ttyd0". You may need to reconfigure applications to
1272 use the new device names.
1274 When using the serial port as a boot console, be sure to update
1275 /boot/device.hints and /etc/ttys before booting the new kernel.
1276 If you forget to do so, you can still manually specify the hints
1277 at the loader prompt:
1279 set hint.uart.0.at="isa"
1280 set hint.uart.0.port="0x3F8"
1281 set hint.uart.0.flags="0x10"
1282 set hint.uart.0.irq="4"
1286 The gpt(8) utility has been removed. Use gpart(8) to partition
1290 The version that Linuxulator emulates was changed from 2.4.2
1291 to 2.6.16. If you experience any problems with Linux binaries
1292 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1293 if it fixes the problem contact emulation mailing list.
1296 ISDN4BSD (I4B) was removed from the src tree. You may need to
1297 update a your kernel configuration and remove relevant entries.
1300 I have checked in code to support multiple routing tables.
1301 See the man pages setfib(1) and setfib(2).
1302 This is a hopefully backwards compatible version,
1303 but to make use of it you need to compile your kernel
1304 with options ROUTETABLES=2 (or more up to 16).
1307 The 802.11 wireless support was redone to enable multi-bss
1308 operation on devices that are capable. The underlying device
1309 is no longer used directly but instead wlanX devices are
1310 cloned with ifconfig. This requires changes to rc.conf files.
1311 For example, change:
1312 ifconfig_ath0="WPA DHCP"
1315 ifconfig_wlan0="WPA DHCP"
1316 see rc.conf(5) for more details. In addition, mergemaster of
1317 /etc/rc.d is highly recommended. Simultaneous update of userland
1318 and kernel wouldn't hurt either.
1320 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1321 modules were merged into the base wlan module. All references
1322 to these modules (e.g. in kernel config files) must be removed.
1325 psm(4) has gained write(2) support in native operation level.
1326 Arbitrary commands can be written to /dev/psm%d and status can
1327 be read back from it. Therefore, an application is responsible
1328 for status validation and error recovery. It is a no-op in
1329 other operation levels.
1332 Support for KSE threading has been removed from the kernel. To
1333 run legacy applications linked against KSE libmap.conf may
1334 be used. The following libmap.conf may be used to ensure
1335 compatibility with any prior release:
1337 libpthread.so.1 libthr.so.1
1338 libpthread.so.2 libthr.so.2
1339 libkse.so.3 libthr.so.3
1342 The layout of struct vmspace has changed. This affects libkvm
1343 and any executables that link against libkvm and use the
1344 kvm_getprocs() function. In particular, but not exclusively,
1345 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1346 The effects are minimal, but it's advisable to upgrade world
1350 The latest em driver no longer has support in it for the
1351 82575 adapter, this is now moved to the igb driver. The
1352 split was done to make new features that are incompatible
1353 with older hardware easier to do.
1356 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1357 likewise the kernel option is now GEOM_LINUX_LVM.
1360 The default NFS mount mode has changed from UDP to TCP for
1361 increased reliability. If you rely on (insecurely) NFS
1362 mounting across a firewall you may need to update your
1366 Belatedly note the addition of m_collapse for compacting
1370 The fts(3) structures have been changed to use adequate
1371 integer types for their members and so to be able to cope
1372 with huge file trees. The old fts(3) ABI is preserved
1373 through symbol versioning in libc, so third-party binaries
1374 using fts(3) should still work, although they will not take
1375 advantage of the extended types. At the same time, some
1376 third-party software might fail to build after this change
1377 due to unportable assumptions made in its source code about
1378 fts(3) structure members. Such software should be fixed
1379 by its vendor or, in the worst case, in the ports tree.
1380 FreeBSD_version 800015 marks this change for the unlikely
1381 case that a portable fix is impossible.
1384 To upgrade to -current after this date, you must be running
1385 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1386 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1389 The ADAPTIVE_GIANT kernel option has been retired because its
1390 functionality is the default now.
1393 The AT keyboard emulation of sunkbd(4) has been turned on
1394 by default. In order to make the special symbols of the Sun
1395 keyboards driven by sunkbd(4) work under X these now have
1396 to be configured the same way as Sun USB keyboards driven
1397 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1399 Option "XkbLayout" "us"
1400 Option "XkbRules" "xorg"
1401 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1404 It has been decided that it is desirable to provide ABI
1405 backwards compatibility to the FreeBSD 4/5/6 versions of the
1406 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1407 broken with the introduction of PCI domain support (see the
1408 20070930 entry). Unfortunately, this required the ABI of
1409 PCIOCGETCONF to be broken again in order to be able to
1410 provide backwards compatibility to the old version of that
1411 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1412 again. As for prominent ports this affects neither pciutils
1413 nor xorg-server this time, the hal port needs to be rebuilt
1417 The misnamed kthread_create() and friends have been renamed
1418 to kproc_create() etc. Many of the callers already
1419 used kproc_start()..
1420 I will return kthread_create() and friends in a while
1421 with implementations that actually create threads, not procs.
1422 Renaming corresponds with version 800002.
1431 Avoid using make -j when upgrading. While generally safe, there are
1432 sometimes problems using -j to upgrade. If your upgrade fails with
1433 -j, please try again without -j. From time to time in the past there
1434 have been problems using -j with buildworld and/or installworld. This
1435 is especially true when upgrading between "distant" versions (eg one
1436 that cross a major release boundary or several minor releases, or when
1437 several months have passed on the -current branch).
1439 Sometimes, obscure build problems are the result of environment
1440 poisoning. This can happen because the make utility reads its
1441 environment when searching for values for global variables. To run
1442 your build attempts in an "environmental clean room", prefix all make
1443 commands with 'env -i '. See the env(1) manual page for more details.
1445 When upgrading from one major version to another it is generally best
1446 to upgrade to the latest code in the currently installed branch first,
1447 then do an upgrade to the new branch. This is the best-tested upgrade
1448 path, and has the highest probability of being successful. Please try
1449 this approach before reporting problems with a major version upgrade.
1453 When upgrading the boot ZFS pool to a new version, always follow
1456 1.) recompile and reinstall the ZFS boot loader and boot block
1457 (this is part of "make buildworld" and "make installworld")
1459 2.) update the ZFS boot block on your boot drive
1461 The following example updates the ZFS boot block on the first
1462 partition (freebsd-boot) of a GPT partitioned drive ad0:
1463 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1465 Non-boot pools do not need these updates.
1469 If you are updating from a prior version of FreeBSD (even one just
1470 a few days old), you should follow this procedure. It is the most
1471 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1473 make kernel-toolchain
1474 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1475 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1477 To test a kernel once
1478 ---------------------
1479 If you just want to boot a kernel once (because you are not sure
1480 if it works, or if you want to boot a known bad kernel to provide
1481 debugging information) run
1482 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1483 nextboot -k testkernel
1485 To just build a kernel when you know that it won't mess you up
1486 --------------------------------------------------------------
1487 This assumes you are already running a CURRENT system. Replace
1488 ${arch} with the architecture of your machine (e.g. "i386",
1489 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1491 cd src/sys/${arch}/conf
1492 config KERNEL_NAME_HERE
1493 cd ../compile/KERNEL_NAME_HERE
1498 If this fails, go to the "To build a kernel" section.
1500 To rebuild everything and install it on the current system.
1501 -----------------------------------------------------------
1502 # Note: sometimes if you are running current you gotta do more than
1503 # is listed here if you are upgrading from a really old current.
1505 <make sure you have good level 0 dumps>
1507 make kernel KERNCONF=YOUR_KERNEL_HERE
1509 <reboot in single user> [3]
1517 To cross-install current onto a separate partition
1518 --------------------------------------------------
1519 # In this approach we use a separate partition to hold
1520 # current's root, 'usr', and 'var' directories. A partition
1521 # holding "/", "/usr" and "/var" should be about 2GB in
1524 <make sure you have good level 0 dumps>
1527 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1528 <maybe newfs current's root partition>
1529 <mount current's root partition on directory ${CURRENT_ROOT}>
1530 make installworld DESTDIR=${CURRENT_ROOT}
1531 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1532 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1533 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1534 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1535 <reboot into current>
1536 <do a "native" rebuild/install as described in the previous section>
1537 <maybe install compatibility libraries from ports/misc/compat*>
1541 To upgrade in-place from stable to current
1542 ----------------------------------------------
1543 <make sure you have good level 0 dumps>
1545 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1547 <reboot in single user> [3]
1554 Make sure that you've read the UPDATING file to understand the
1555 tweaks to various things you need. At this point in the life
1556 cycle of current, things change often and you are on your own
1557 to cope. The defaults can also change, so please read ALL of
1558 the UPDATING entries.
1560 Also, if you are tracking -current, you must be subscribed to
1561 freebsd-current@freebsd.org. Make sure that before you update
1562 your sources that you have read and understood all the recent
1563 messages there. If in doubt, please track -stable which has
1564 much fewer pitfalls.
1566 [1] If you have third party modules, such as vmware, you
1567 should disable them at this point so they don't crash your
1570 [3] From the bootblocks, boot -s, and then do
1575 adjkerntz -i # if CMOS is wall time
1576 Also, when doing a major release upgrade, it is required that
1577 you boot into single user mode to do the installworld.
1579 [4] Note: This step is non-optional. Failure to do this step
1580 can result in a significant reduction in the functionality of the
1581 system. Attempting to do it by hand is not recommended and those
1582 that pursue this avenue should read this file carefully, as well
1583 as the archives of freebsd-current and freebsd-hackers mailing lists
1584 for potential gotchas. The -U option is also useful to consider.
1585 See mergemaster(8) for more information.
1587 [5] Usually this step is a noop. However, from time to time
1588 you may need to do this if you get unknown user in the following
1589 step. It never hurts to do it all the time. You may need to
1590 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1591 install) after the buildworld before this step if you last updated
1592 from [78]-stable or 9-stable before 20130430.
1594 [6] This only deletes old files and directories. Old libraries
1595 can be deleted by "make delete-old-libs", but you have to make
1596 sure that no program is using those libraries anymore.
1598 [8] In order to have a kernel that can run the 4.x binaries needed to
1599 do an installworld, you must include the COMPAT_FREEBSD4 option in
1600 your kernel. Failure to do so may leave you with a system that is
1601 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1602 required to run the 5.x binaries on more recent kernels. And so on
1603 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1605 Make sure that you merge any new devices from GENERIC since the
1606 last time you updated your kernel config file.
1608 [9] When checking out sources, you must include the -P flag to have
1609 cvs prune empty directories.
1611 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1612 "?=" instead of the "=" assignment operator, so that buildworld can
1613 override the CPUTYPE if it needs to.
1615 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1616 not on the command line, or in /etc/make.conf. buildworld will
1617 warn if it is improperly defined.
1620 This file contains a list, in reverse chronological order, of major
1621 breakages in tracking -current. It is not guaranteed to be a complete
1622 list of such breakages, and only contains entries since October 10, 2007.
1623 If you need to see UPDATING entries from before that date, you will need
1624 to fetch an UPDATING file from an older FreeBSD release.
1626 Copyright information:
1628 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1630 Redistribution, publication, translation and use, with or without
1631 modification, in full or in part, in any form or format of this
1632 document are permitted without further permission from the author.
1634 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1635 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1636 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1637 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1638 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1639 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1640 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1641 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1642 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1643 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1644 POSSIBILITY OF SUCH DAMAGE.
1646 Contact Warner Losh if you have any questions about your use of