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.
18 Behavior of devfs rules path matching has been changed.
19 Pattern is now always matched against fully qualified devfs
20 path and slash characters must be explicitly matched by
21 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
22 subdirectories must be reviewed.
25 hastctl(8)'s `status' command output changed to terse one-liner format.
26 Scripts using this should switch to `list' command or be rewritten.
29 Fix a bug that allowed a tracing process (e.g. gdb) to write
30 to a memory-mapped file in the traced process's address space
31 even if neither the traced process nor the tracing process had
32 write access to that file.
35 Added ZFS TRIM support which is enabled by default. To disable
36 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
38 Creating new ZFS pools and adding new devices to existing pools
39 first performs a full device level TRIM which can take a significant
40 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
41 to disable this behaviour.
43 ZFS TRIM requires the underlying device support BIO_DELETE which
44 is currently provided by methods such as ATA TRIM and SCSI UNMAP
45 via CAM, which are typically supported by SSD's.
47 Stats for ZFS TRIM can be monitored by looking at the sysctl's
48 under kstat.zfs.misc.zio_trim.
51 `list' command has been added to hastctl(8). For now, it is full
52 equivalent of `status' command.
53 WARNING: in the near future the output of hastctl's status command
54 will change to more terse format. If you use `hastctl status'
55 for parsing in your scripts, switch to `hastctl list'.
58 The mergemaster command now uses the default MAKEOBJDIRPREFIX
59 rather than creating it's own in the temporary directory in
60 order allow access to bootstrapped versions of tools such as
61 install and mtree. When upgrading from version of FreeBSD where
62 the install command does not support -l, you will need to
63 install a new mergemaster command if mergemaster -p is required.
64 This can be accomplished with the command (cd src/usr.sbin/mergemaster
67 Due to the use of the new -l option to install(1) during build
68 and install, you must take care not to directly set the INSTALL
69 make variable in your /etc/make.conf, /etc/src.conf, or on the
70 command line. If you wish to use the -C flag for all installs
71 you may be able to add INSTALL+=-C to /etc/make.conf or
75 Fix a bug that allows NFS clients to issue READDIR on files.
78 The install(1) option -M has changed meaning and now takes an
79 argument that is a file or path to append logs to. In the
80 unlikely event that -M was the last option on the command line
81 and the command line contained at least two files and a target
82 directory the first file will have logs appended to it. The -M
83 option served little practical purpose in the last decade so it's
84 used expected to be extremely rare.
87 A new compression method (lz4) has been merged to. Please refer to
88 zpool-features(7) for more information.
90 Please refer to the "ZFS notes" section of this file for information
91 on upgrading boot ZFS pools.
94 The VFS KBI was changed with the merge of several nullfs
95 optimizations and fixes. All filesystem modules must be
99 With the addition of auditdistd(8), a new auditdistd user is now
100 depended on during installworld. "mergemaster -p" can be used to add
101 the user prior to installworld, as documented in the handbook.
107 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
108 Starting with this version the old system of ZFS pool versioning
109 is superseded by "feature flags". This concept enables forward
110 compatibility against certain future changes in functionality of ZFS
111 pools. The first two read-only compatible "feature flags" for ZFS
112 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
113 For more information read the new zpool-features(7) manual page.
114 Please refer to the "ZFS notes" section of this file for information
115 on upgrading boot ZFS pools.
118 The commit introducing bsd.compiler.mk breaks the traditional
119 building of kernels before this point. Add -m ${SRC}/share/mk
120 (for the right value of SRC) to your command lines to work
121 around; update your useland to a point after this; or use the
122 buildkernel/installkernel top-level targets. See also 20120829.
125 The IPFIREWALL_FORWARD kernel option has been removed. Its
126 functionality now turned on by default.
129 The random(4) support for the VIA hardware random number
130 generator (`PADLOCK') is no longer enabled unconditionally.
131 Add the PADLOCK_RNG option in the custom kernel config if
132 needed. The GENERIC kernels on i386 and amd64 do include the
133 option, so the change only affects the custom kernel
137 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
138 the traditional method, you must update your system with an installworld
139 before the kernel will build. The documented make buildkernel/installkernel
140 interfaces (coupled with fresh make kernel-toolchain) continue to work.
143 The sparc64 ZFS loader has been changed to no longer try to auto-
144 detect ZFS providers based on diskN aliases but now requires these
145 to be explicitly listed in the OFW boot-device environment variable.
148 Now unix domain sockets behave "as expected" on nullfs(5). Previously
149 nullfs(5) did not pass through all behaviours to the underlying layer,
150 as a result if we bound to a socket on the lower layer we could connect
151 only to the lower path; if we bound to the upper layer we could connect
152 only to the upper path. The new behavior is one can connect to both the
153 lower and the upper paths regardless what layer path one binds to.
156 The acpi_wmi(4) status device /dev/wmistat has been renamed to
160 A new VOP_ADVISE() was added to support posix_fadvise(2). All
161 filesystem modules must be recompiled.
164 The interface of the VOP_VPTOCNP(9) changed, now the returned
165 vnode shall be referenced, previously it was required to be
166 only held. All in-tree filesystems are converted.
172 The broken amd(4) driver has been replaced with esp(4) in the amd64,
173 i386 and pc98 GENERIC kernel configuration files.
176 This commit modifies vfs_register() so that it uses a hash
177 calculation to set vfc_typenum, which is enabled by default.
178 The first time a system is booted after this change, the
179 vfc_typenum values will change for all file systems. The
180 main effect of this is a change to the NFS server file handles
181 for file systems that use vfc_typenum in their fsid, such as ZFS.
182 It will, however, prevent vfc_typenum from changing when file
183 systems are loaded in a different order for subsequent reboots.
184 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
185 until you are ready to remount all NFS clients after a reboot.
188 Bump the shared library version numbers for libraries that
189 do not use symbol versioning, have changed the ABI compared
190 to stable/8 and which shared library version was not bumped.
191 Done as part of 9.0-RELEASE cycle.
194 During the merge of Capsicum features, the fget(9) KPI was modified.
195 This may require the rebuilding of out-of-tree device drivers --
196 issues have been reported specifically with the nVidia device driver.
197 __FreeBSD_version is bumped to 900041.
199 Also, there is a period between 20110811 and 20110814 where the
200 special devices /dev/{stdin,stdout,stderr} did not work correctly.
201 Building world from a kernel during that window may not work.
204 The packet filter (pf) code has been updated to OpenBSD 4.5.
205 You need to update userland tools to be in sync with kernel.
206 This update breaks backward compatibility with earlier pfsync(4)
207 versions. Care must be taken when updating redundant firewall setups.
210 The following sysctls and tunables are retired on x86 platforms:
212 machdep.hlt_logical_cpus
213 The following sysctl is retired:
214 machdep.hyperthreading_allowed
215 The sysctls were supposed to provide a way to dynamically offline and
216 online selected CPUs on x86 platforms, but the implementation has not
217 been reliable especially with SCHED_ULE scheduler.
218 machdep.hyperthreading_allowed tunable is still available to ignore
219 hyperthreading CPUs at OS level.
220 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
221 where X is an APIC ID of a CPU. Be advised, though, that disabling
222 CPUs in non-uniform fashion will result in non-uniform topology and
223 may lead to sub-optimal system performance with SCHED_ULE, which is
227 cpumask_t type is retired and cpuset_t is used in order to describe
231 Changes to ifconfig(8) for dynamic address family detection mandate
232 that you are running a kernel of 20110525 or later. Make sure to
233 follow the update procedure to boot a new kernel before installing
237 Support for sun4v architecture is officially dropped
240 Several KPI breaking changes have been committed to the mii(4) layer,
241 the PHY drivers and consequently some Ethernet drivers using mii(4).
242 This means that miibus.ko and the modules of the affected Ethernet
243 drivers need to be recompiled.
245 Note to kernel developers: Given that the OUI bit reversion problem
246 was fixed as part of these changes all mii(4) commits related to OUIs,
247 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
248 handling, no longer can be merged verbatim to stable/8 and previous
252 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
253 into their kernel configurations along with 'device pci'.
256 The default NFS client is now the new NFS client, so fstype "newnfs"
257 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
258 Although mounts via fstype "nfs" will usually work without userland
259 changes, it is recommended that the mount(8) and mount_nfs(8)
260 commands be rebuilt from sources and that a link to mount_nfs called
261 mount_oldnfs be created. The new client is compiled into the
262 kernel with "options NFSCL" and this is needed for diskless root
263 file systems. The GENERIC kernel configs have been changed to use
264 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
265 To use the regular/old client, you can "mount -t oldnfs ...". For
266 a diskless root file system, you must also include a line like:
268 vfs.root.mountfrom="oldnfs:"
270 in the boot/loader.conf on the root fs on the NFS server to make
271 a diskless root fs use the old client.
274 The GENERIC kernels for all architectures now default to the new
275 CAM-based ATA stack. It means that all legacy ATA drivers were
276 removed and replaced by respective CAM drivers. If you are using
277 ATA device names in /etc/fstab or other places, make sure to update
278 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
279 where 'Y's are the sequential numbers starting from zero for each type
280 in order of detection, unless configured otherwise with tunables,
281 see cam(4)). There will be symbolic links created in /dev/ to map
282 old adX devices to the respective adaY. They should provide basic
283 compatibility for file systems mounting in most cases, but they do
284 not support old user-level APIs and do not have respective providers
285 in GEOM. Consider using updated management tools with new device names.
287 It is possible to load devices ahci, ata, siis and mvs as modules,
288 but option ATA_CAM should remain in kernel configuration to make ata
289 module work as CAM driver supporting legacy ATA controllers. Device ata
290 still can be used in modular fashion (atacore + ...). Modules atadisk
291 and atapi* are not used and won't affect operation in ATA_CAM mode.
292 Note that to use CAM-based ATA kernel should include CAM devices
293 scbus, pass, da (or explicitly ada), cd and optionally others. All of
294 them are parts of the cam module.
296 ataraid(4) functionality is now supported by the RAID GEOM class.
297 To use it you can load geom_raid kernel module and use graid(8) tool
298 for management. Instead of /dev/arX device names, use /dev/raid/rX.
300 No kernel config options or code have been removed, so if a problem
301 arises, please report it and optionally revert to the old ATA stack.
302 In order to do it you can remove from the kernel config:
307 , and instead add back:
308 device atadisk # ATA disk drives
309 device ataraid # ATA RAID drives
310 device atapicd # ATAPI CDROM drives
311 device atapifd # ATAPI floppy drives
312 device atapist # ATAPI tape drives
315 The default NFS server has been changed to the new server, which
316 was referred to as the experimental server. If you need to switch
317 back to the old NFS server, you must now put the "-o" option on
318 both the mountd and nfsd commands. This can be done using the
319 mountd_flags and nfs_server_flags rc.conf variables until an
320 update to the rc scripts is committed, which is coming soon.
323 The GNU Objective-C runtime library (libobjc), and other Objective-C
324 related components have been removed from the base system. If you
325 require an Objective-C library, please use one of the available ports.
328 ath(4) has been split into bus- and device- modules. if_ath contains
329 the HAL, the TX rate control and the network device code. if_ath_pci
330 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
331 contains the AHB glue. Users need to load both if_ath_pci and if_ath
332 in order to use ath on everything else.
334 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
335 need to load if_ath and if_ath_pci for ath(4) operation.
338 As part of the replacement of sysinstall, the process of building
339 release media has changed significantly. For details, please re-read
340 release(7), which has been updated to reflect the new build process.
343 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
344 is the last available version under GPLv2. It brings a number of new
345 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
346 4.1 and SSE 4.2), better support for powerpc64, a number of new
347 directives, and lots of other small improvements. See the ChangeLog
348 file in contrib/binutils for the full details.
351 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
352 compliant, and will now use half of hash for authentication.
353 This will break interoperability with all stacks (including all
354 actual FreeBSD versions) who implement
355 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
357 The only workaround with such peers is to use another HMAC
358 algorithm for IPsec ("phase 2") authentication.
361 Remove the uio_yield prototype and symbol. This function has
362 been misnamed since it was introduced and should not be
363 globally exposed with this name. The equivalent functionality
364 is now available using kern_yield(curthread->td_user_pri).
365 The function remains undocumented.
368 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
369 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
370 for scalar sysctls is defined but disabled. Code that needs
371 UQUAD to pass the type checking that must compile on older
372 systems where the define is not present can check against
373 __FreeBSD_version >= 900030.
375 The system dialog(1) has been replaced with a new version previously
376 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
377 with the previous version, but the libdialog associated with it has
378 a largely incompatible API. As such, the original version of libdialog
379 will be kept temporarily as libodialog, until its base system consumers
380 are replaced or updated. Bump __FreeBSD_version to 900030.
383 If you are trying to run make universe on a -stable system, and you get
384 the following warning:
385 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
386 unknown. config(8) likely too old."
387 or something similar to it, then you must upgrade your -stable system
388 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
389 stable/8) or build the config from the latest stable/8 branch and
390 install it on your system.
392 Prior to this date, building a current universe on 8-stable system from
393 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
394 error in the first kernel build phase. A new config on those old
395 systems will fix that problem for older versions of -current.
398 The TCP stack has been modified to allow Khelp modules to interact with
399 it via helper hook points and store per-connection data in the TCP
400 control block. Bump __FreeBSD_version to 900029. User space tools that
401 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
405 Generic IEEE 802.3 annex 31B full duplex flow control support has been
406 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
407 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
408 to take advantage of it instead of using custom implementations. This
409 means that these drivers now no longer unconditionally advertise
410 support for flow control but only do so if flow control is a selected
411 media option. This was implemented in the generic support that way in
412 order to allow flow control to be switched on and off via ifconfig(8)
413 with the PHY specific default to typically off in order to protect
414 from unwanted effects. Consequently, if you used flow control with
415 one of the above mentioned drivers you now need to explicitly enable
417 ifconfig bge0 media auto mediaopt flowcontrol
419 Along with the above mentioned changes generic support for setting
420 1000baseT master mode also has been added and brgphy(4), ciphy(4),
421 e1000phy(4) as well as ip1000phy(4) have been converted to take
422 advantage of it. This means that these drivers now no longer take the
423 link0 parameter for selecting master mode but the master media option
424 has to be used instead, for example like in the following:
425 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
427 Selection of master mode now is also available with all other PHY
428 drivers supporting 1000baseT.
431 The TCP stack has received a significant update to add support for
432 modularised congestion control and generally improve the clarity of
433 congestion control decisions. Bump __FreeBSD_version to 900025. User
434 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
435 sockstat) need to be recompiled.
438 The man(1) utility has been replaced by a new version that no longer
439 uses /etc/manpath.config. Please consult man.conf(5) for how to
440 migrate local entries to the new format.
443 The copyright strings printed by login(1) and sshd(8) at the time of a
444 new connection have been removed to follow other operating systems and
448 A workaround for a fixed ld bug has been removed in kernel code,
449 so make sure that your system ld is built from sources after
450 revision 210245 from 2010-07-19 (r211583 if building head kernel
451 on stable/8, r211584 for stable/7; both from 2010-08-21).
452 A symptom of incorrect ld version is different addresses for
453 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
456 The $ipv6_prefer variable in rc.conf(5) has been split into
457 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
459 The $ip6addrctl_policy is a variable to choose a pre-defined
460 address selection policy set by ip6addrctl(8). A value
461 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
464 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
465 flag (see an entry of 20090926) is set on an interface with no
466 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
467 security reason. If you want IPv6 link-local address on all
468 interfaces by default, set this to "YES".
470 The old ipv6_prefer="YES" is equivalent to
471 ipv6_activate_all_interfaces="YES" and
472 ip6addrctl_policy="ipv6_prefer".
475 DTrace has grown support for userland tracing. Due to this, DTrace is
476 now i386 and amd64 only.
477 dtruss(1) is now installed by default on those systems and a new
478 kernel module is needed for userland tracing: fasttrap.
479 No changes to your kernel config file are necessary to enable
480 userland tracing, but you might consider adding 'STRIP=' and
481 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
482 to have informative userland stack traces in DTrace (ustack).
485 The acpi_aiboost(4) driver has been removed in favor of the new
486 aibs(4) driver. You should update your kernel configuration file.
489 BSD grep has been imported to the base system and it is built by
490 default. It is completely BSD licensed, highly GNU-compatible, uses
491 less memory than its GNU counterpart and has a small codebase.
492 However, it is slower than its GNU counterpart, which is mostly
493 noticeable for larger searches, for smaller ones it is measurable
494 but not significant. The reason is complex, the most important factor
495 is that we lack a modern and efficient regex library and GNU
496 overcomes this by optimizing the searches internally. Future work
497 on improving the regex performance is planned, for the meantime,
498 users that need better performance, can build GNU grep instead by
499 setting the WITH_GNU_GREP knob.
502 Due to the import of powerpc64 support, all existing powerpc kernel
503 configuration files must be updated with a machine directive like this:
504 machine powerpc powerpc
506 In addition, an updated config(8) is required to build powerpc kernels
510 A new version of ZFS (version 15) has been merged to -HEAD.
511 This version uses a python library for the following subcommands:
512 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
513 For full functionality of these commands the following port must
514 be installed: sysutils/py-zfs
517 'vm_page's are now hashed by physical address to an array of mutexes.
518 Currently this is only used to serialize access to hold_count. Over
519 time the page queue mutex will be peeled away. This changes the size
520 of pmap on every architecture. And requires all callers of vm_page_hold
521 and vm_page_unhold to be updated.
524 WITH_CTF can now be specified in src.conf (not recommended, there
525 are some problems with static executables), make.conf (would also
526 affect ports which do not use GNU make and do not override the
527 compile targets) or in the kernel config (via "makeoptions
529 When WITH_CTF was specified there before this was silently ignored,
530 so make sure that WITH_CTF is not used in places which could lead
531 to unwanted behavior.
534 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
535 to allow 32-bit compatibility on non-x86 platforms. All kernel
536 configurations on amd64 and ia64 platforms using these options must
537 be modified accordingly.
540 The utmp user accounting database has been replaced with utmpx,
541 the user accounting interface standardized by POSIX.
542 Unfortunately the semantics of utmp and utmpx don't match,
543 making it practically impossible to support both interfaces.
544 The user accounting database is used by tools like finger(1),
545 last(1), talk(1), w(1) and ac(8).
547 All applications in the base system use utmpx. This means only
548 local binaries (e.g. from the ports tree) may still use these
549 utmp database files. These applications must be rebuilt to make
552 After the system has been upgraded, it is safe to remove the old
553 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
554 assuming their contents is of no importance anymore. Old wtmp
555 databases can only be used by last(1) and ac(8) after they have
556 been converted to the new format using wtmpcvt(1).
559 Introduce the kernel thread "deadlock resolver" (which can be enabled
560 via the DEADLKRES option, see NOTES for more details) and the
561 sleepq_type() function for sleepqueues.
564 The rc.firewall and rc.firewall6 were unified, and
565 rc.firewall6 and rc.d/ip6fw were removed.
566 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
567 variables are obsoleted. Instead, the following new rc
568 variables are added to rc.d/ipfw:
570 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
571 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
572 firewall_simple_onet_ipv6, firewall_trusted_ipv6
574 The meanings correspond to the relevant IPv4 variables.
580 The default terminal emulation for syscons(4) has been changed
581 from cons25 to xterm on all platforms except pc98. This means
582 that the /etc/ttys file needs to be updated to ensure correct
583 operation of applications on the console.
585 The terminal emulation style can be toggled per window by using
586 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
587 options can be used to change the compile-time default back to
590 To prevent graphical artifacts, make sure the TERM environment
591 variable is set to match the terminal emulation that is being
592 performed by syscons(4).
595 The layout of the structure ieee80211req_scan_result has changed.
596 Applications that require wireless scan results (e.g. ifconfig(8))
597 from net80211 need to be recompiled.
599 Applications such as wpa_supplicant(8) may require a full world
600 build without using NO_CLEAN in order to get synchronized with the
604 The iwn(4) driver has been updated to support the 5000 and 5150 series.
605 There's one kernel module for each firmware. Adding "device iwnfw"
606 to the kernel configuration file means including all three firmware
607 images inside the kernel. If you want to include just the one for
608 your wireless card, use the devices iwn4965fw, iwn5000fw or
612 The rc.d/network_ipv6, IPv6 configuration script has been integrated
613 into rc.d/netif. The changes are the following:
615 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
616 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
617 Note that both variables need the "inet6" keyword at the head.
619 Do not set $ipv6_network_interfaces manually if you do not
620 understand what you are doing. It is not needed in most cases.
622 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
625 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
626 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
628 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
629 all configured interfaces have "inet6 accept_rtadv" in the
630 $ifconfig_IF_ipv6. These are for backward compatibility.
632 3. A new variable $ipv6_prefer has been added. If NO, IPv6
633 functionality of interfaces with no corresponding
634 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
635 and the default address selection policy of ip6addrctl(8)
636 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
637 Note that if you want to configure IPv6 functionality on the
638 disabled interfaces after boot, first you need to clear the flag by
639 using ifconfig(8) like:
641 ifconfig em0 inet6 -ifdisabled
643 If YES, the default address selection policy is set as
646 The default value of $ipv6_prefer is NO.
648 4. If your system need to receive Router Advertisement messages,
649 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
650 scripts automatically invoke rtsol(8) when the interface becomes
651 UP. The Router Advertisement messages are used for SLAAC
652 (State-Less Address AutoConfiguration).
655 802.11s D3.03 support was committed. This is incompatible with the
656 previous code, which was based on D3.0.
659 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
660 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
661 control whether accepting Router Advertisement messages or not.
662 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
663 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
664 The ifconfig(8) utility now supports these flags.
667 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
668 mount(8) and -a option for df(1) to see them.
671 The old tunable hw.bus.devctl_disable has been superseded by
672 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
673 replaced by hw.bus.devctl_queue=0. The default for this new tunable
677 Remove the option STOP_NMI. The default action is now to use NMI only
678 for KDB via the newly introduced function stop_cpus_hard() and
679 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
682 The stable/8 branch created in subversion. This corresponds to the
683 RELENG_8 branch in CVS.
686 Bump the shared library version numbers for all libraries that do not
687 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
688 __FreeBSD_version to 800105.
691 Due to changes in the implementation of virtual network stack support,
692 all network-related kernel modules must be recompiled. As this change
693 breaks the ABI, bump __FreeBSD_version to 800104.
696 The TOE interface to the TCP syncache has been modified to remove
697 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
698 The cxgb driver is the only TOE consumer affected by this change, and
699 needs to be recompiled along with the kernel. As this change breaks
700 the ABI, bump __FreeBSD_version to 800103.
703 Padding has been added to struct tcpcb, sackhint and tcpstat in
704 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
705 maintaining the ABI. However, this change breaks the ABI, so bump
706 __FreeBSD_version to 800102. User space tools that rely on the size of
707 any of these structs (e.g. sockstat) need to be recompiled.
710 The NFS_LEGACYRPC option has been removed along with the old kernel
711 RPC implementation that this option selected. Kernel configurations
712 may need to be adjusted.
715 The network interface device nodes at /dev/net/<interface> have been
716 removed. All ioctl operations can be performed the normal way using
717 routing sockets. The kqueue functionality can generally be replaced
718 with routing sockets.
721 The documentation from the FreeBSD Documentation Project (Handbook,
722 FAQ, etc.) is now installed via packages by sysinstall(8) and under
723 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
726 The ABI of various structures related to the SYSV IPC API have been
727 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
728 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
732 Layout of struct vnet has changed as routing related variables were
733 moved to their own Vimage module. Modules need to be recompiled. Bump
734 __FreeBSD_version to 800099.
737 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
738 respectively. As long as no more than 16 groups per process are used,
739 no changes should be visible. When more than 16 groups are used, old
740 binaries may fail if they call getgroups() or getgrouplist() with
741 statically sized storage. Recompiling will work around this, but
742 applications should be modified to use dynamically allocated storage
743 for group arrays as POSIX.1-2008 does not cap an implementation's
744 number of supported groups at NGROUPS_MAX+1 as previous versions did.
746 NFS and portalfs mounts may also be affected as the list of groups is
747 truncated to 16. Users of NFS who use more than 16 groups, should
748 take care that negative group permissions are not used on the exported
749 file systems as they will not be reliable unless a GSSAPI based
750 authentication method is used.
753 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
754 option compiles in the support for adaptive spinning for lockmgrs
755 which want to enable it. The lockinit() function now accepts the flag
756 LK_ADAPTIVE in order to make the lock object subject to adaptive
757 spinning when both held in write and read mode.
760 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
761 changed. User applications that use this ioctl need to be rebuilt.
764 The layout of struct thread has changed. Kernel and modules need to
768 The layout of structs ifnet, domain, protosw and vnet_net has changed.
769 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
772 window(1) has been removed from the base system. It can now be
773 installed from ports. The port is called misc/window.
776 The way we are storing and accessing `routing table' entries has
777 changed. Programs reading the FIB, like netstat, need to be
781 A new netisr implementation has been added for FreeBSD 8. Network
782 file system modules, such as igmp, ipdivert, and others, should be
784 Bump __FreeBSD_version to 800096.
787 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
791 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
792 Bump __FreeBSD_version to 800094.
795 Add mnt_xflag field to 'struct mount'. File system modules need to be
797 Bump __FreeBSD_version to 800093.
800 The compiling option ADAPTIVE_SX has been retired while it has been
801 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
802 The KPI for sx_init_flags() changes as accepting flags:
803 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
804 been introduced in order to handle the reversed logic.
805 Bump __FreeBSD_version to 800092.
808 Add support for hierarchical jails. Remove global securelevel.
809 Bump __FreeBSD_version to 800091.
812 The layout of struct vnet_net has changed, therefore modules
814 Bump __FreeBSD_version to 800090.
817 The newly imported zic(8) produces a new format in the output. Please
818 run tzsetup(8) to install the newly created data to /etc/localtime.
821 The sysctl tree for the usb stack has renamed from hw.usb2.* to
822 hw.usb.* and is now consistent again with previous releases.
825 802.11 monitor mode support was revised and driver api's were changed.
826 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
827 of DLT_IEEE802_11. No user-visible data structures were changed but
828 applications that use DLT_IEEE802_11 may require changes.
829 Bump __FreeBSD_version to 800088.
832 The layout of the following structs has changed: sysctl_oid,
833 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
834 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
835 panics may be experienced. World rebuild is required for
836 correctly checking networking state from userland.
837 Bump __FreeBSD_version to 800085.
840 MLDv2 and Source-Specific Multicast (SSM) have been merged
841 to the IPv6 stack. VIMAGE hooks are in but not yet used.
842 The implementation of SSM within FreeBSD's IPv6 stack closely
843 follows the IPv4 implementation.
845 For kernel developers:
847 * The most important changes are that the ip6_output() and
848 ip6_input() paths no longer take the IN6_MULTI_LOCK,
849 and this lock has been downgraded to a non-recursive mutex.
851 * As with the changes to the IPv4 stack to support SSM, filtering
852 of inbound multicast traffic must now be performed by transport
853 protocols within the IPv6 stack. This does not apply to TCP and
854 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
856 * The KPIs used by IPv6 multicast are similar to those used by
857 the IPv4 stack, with the following differences:
858 * im6o_mc_filter() is analogous to imo_multicast_filter().
859 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
860 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
861 * IN6_LOOKUP_MULTI() has been deprecated and removed.
862 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
863 for MLDv1 have an additional 'timer' argument which is used to
864 jitter the initial membership report for the solicited-node
865 multicast membership on-link.
866 * This is not strictly needed for MLDv2, which already jitters
867 its report transmissions. However, the 'timer' argument is
868 preserved in case MLDv1 is active on the interface.
870 * The KAME linked-list based IPv6 membership implementation has
871 been refactored to use a vector similar to that used by the IPv4
873 Code which maintains a list of its own multicast memberships
874 internally, e.g. carp, has been updated to reflect the new
877 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
878 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
879 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
880 implementation constraint which needs to be addressed in HEAD.
882 For application developers:
884 * The changes are broadly similar to those made for the IPv4
887 * The use of IPv4 and IPv6 multicast socket options on the same
888 socket, using mapped addresses, HAS NOT been tested or supported.
890 * There are a number of issues with the implementation of various
891 IPv6 multicast APIs which need to be resolved in the API surface
892 before the implementation is fully compatible with KAME userland
893 use, and these are mostly to do with interface index treatment.
895 * The literature available discusses the use of either the delta / ASM
896 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
897 using setsourcefilter(3)/getsourcefilter(3). For more information
898 please refer to RFC 3768, 'Socket Interface Extensions for
899 Multicast Source Filters'.
901 * Applications which use the published RFC 3678 APIs should be fine.
903 For systems administrators:
905 * The mtest(8) utility has been refactored to support IPv6, in
906 addition to IPv4. Interface addresses are no longer accepted
907 as arguments, their names must be used instead. The utility
908 will map the interface name to its first IPv4 address as
909 returned by getifaddrs(3).
911 * The ifmcstat(8) utility has also been updated to print the MLDv2
912 endpoint state and source filter lists via sysctl(3).
914 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
915 loopback of IPv6 multicast datagrams by default; it defaults to 1
916 to preserve the existing behaviour. Disabling multicast loopback is
917 recommended for optimal system performance.
919 * The IPv6 MROUTING code has been changed to examine this sysctl
920 instead of attempting to perform a group lookup before looping
921 back forwarded datagrams.
923 Bump __FreeBSD_version to 800084.
926 Implement low-level Bluetooth HCI API.
927 Bump __FreeBSD_version to 800083.
930 The layout of struct malloc_type, used by modules to register new
931 memory allocation types, has changed. Most modules will need to
932 be rebuilt or panics may be experienced.
933 Bump __FreeBSD_version to 800081.
936 Anticipate overflowing inp_flags - add inp_flags2.
937 This changes most offsets in inpcb, so checking v4 connection
938 state will require a world rebuild.
939 Bump __FreeBSD_version to 800080.
942 Add an llentry to struct route and struct route_in6. Modules
943 embedding a struct route will need to be recompiled.
944 Bump __FreeBSD_version to 800079.
947 The size of rt_metrics_lite and by extension rtentry has changed.
948 Networking administration apps will need to be recompiled.
949 The route command now supports show as an alias for get, weighting
950 of routes, sticky and nostick flags to alter the behavior of stateful
952 Bump __FreeBSD_version to 800078.
955 Do not use Giant for kbdmux(4) locking. This is wrong and
956 apparently causing more problems than it solves. This will
957 re-open the issue where interrupt handlers may race with
958 kbdmux(4) in polling mode. Typical symptoms include (but
959 not limited to) duplicated and/or missing characters when
960 low level console functions (such as gets) are used while
961 interrupts are enabled (for example geli password prompt,
962 mountroot prompt etc.). Disabling kbdmux(4) may help.
965 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
966 kernel modules referencing any of the above need to be recompiled.
967 Bump __FreeBSD_version to 800075.
970 GEOM_PART has become the default partition slicer for storage devices,
971 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
972 introduces some changes:
974 MSDOS/EBR: the devices created from MSDOS extended partition entries
975 (EBR) can be named differently than with GEOM_MBR and are now symlinks
976 to devices with offset-based names. fstabs may need to be modified.
978 BSD: the "geometry does not match label" warning is harmless in most
979 cases but it points to problems in file system misalignment with
980 disk geometry. The "c" partition is now implicit, covers the whole
981 top-level drive and cannot be (mis)used by users.
983 General: Kernel dumps are now not allowed to be written to devices
984 whose partition types indicate they are meant to be used for file
985 systems (or, in case of MSDOS partitions, as something else than
988 Most of these changes date approximately from 200812.
991 The uscanner(4) driver has been removed from the kernel. This follows
992 Linux removing theirs in 2.6 and making libusb the default interface
996 The multicast forwarding code has been cleaned up. netstat(1)
997 only relies on KVM now for printing bandwidth upcall meters.
998 The IPv4 and IPv6 modules are split into ip_mroute_mod and
999 ip6_mroute_mod respectively. The config(5) options for statically
1000 compiling this code remain the same, i.e. 'options MROUTING'.
1003 Support for the IFF_NEEDSGIANT network interface flag has been
1004 removed, which means that non-MPSAFE network device drivers are no
1005 longer supported. In particular, if_ar, if_sr, and network device
1006 drivers from the old (legacy) USB stack can no longer be built or
1010 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1011 a bunch of new language catalog files have also been added.
1012 This means that some common libc messages are now localized and
1013 they depend on the LC_MESSAGES environmental variable.
1016 The k8temp(4) driver has been renamed to amdtemp(4) since
1017 support for Family 10 and Family 11 CPU families was added.
1020 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1021 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1023 For kernel developers, the most important changes are that the
1024 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1025 and this lock has been downgraded to a non-recursive mutex.
1027 Transport protocols (UDP, Raw IP) are now responsible for filtering
1028 inbound multicast traffic according to group membership and source
1029 filters. The imo_multicast_filter() KPI exists for this purpose.
1030 Transports which do not use multicast (SCTP, TCP) already reject
1031 multicast by default. Forwarding and receive performance may improve
1032 as a mutex acquisition is no longer needed in the ip_input()
1033 low-level input path. in_addmulti() and in_delmulti() are shimmed
1034 to new KPIs which exist to support SSM in-kernel.
1036 For application developers, it is recommended that loopback of
1037 multicast datagrams be disabled for best performance, as this
1038 will still cause the lock to be taken for each looped-back
1039 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1040 be tuned to 0 to disable loopback by default; it defaults to 1
1041 to preserve the existing behaviour.
1043 For systems administrators, to obtain best performance with
1044 multicast reception and multiple groups, it is always recommended
1045 that a card with a suitably precise hash filter is used. Hash
1046 collisions will still result in the lock being taken within the
1047 transport protocol input path to check group membership.
1049 If deploying FreeBSD in an environment with IGMP snooping switches,
1050 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1051 enabled; this forces 224.0.0.0/24 group membership to be announced
1054 The size of 'struct igmpstat' has changed; netstat needs to be
1055 recompiled to reflect this.
1056 Bump __FreeBSD_version to 800070.
1059 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1060 updated to use it. This requires a buildworld/installworld in order to
1061 update the library and dependencies (usbconfig, etc). Its advisable to
1062 rebuild all ports which uses libusb. More specific directions are given
1063 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1064 libusb are no longer required and can be removed.
1067 A workaround is committed to allow the creation of System V shared
1068 memory segment of size > 2 GB on the 64-bit architectures.
1069 Due to a limitation of the existing ABI, the shm_segsz member
1070 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1071 wrong for large segments. Note that limits must be explicitly
1072 raised to allow such segments to be created.
1075 The layout of struct ifnet has changed, requiring a rebuild of all
1076 network device driver modules.
1079 The /dev handling for the new USB stack has changed, a
1080 buildworld/installworld is required for libusb20.
1083 The new USB2 stack has now been permanently moved in and all kernel and
1084 module names reverted to their previous values (eg, usb, ehci, ohci,
1085 ums, ...). The old usb stack can be compiled in by prefixing the name
1086 with the letter 'o', the old usb modules have been removed.
1087 Updating entry 20090216 for xorg and 20090215 for libmap may still
1091 The rc.conf(5) option if_up_delay has been renamed to
1092 defaultroute_delay to better reflect its purpose. If you have
1093 customized this setting in /etc/rc.conf you need to update it to
1097 xorg 7.4 wants to configure its input devices via hald which does not
1098 yet work with USB2. If the keyboard/mouse does not work in xorg then
1100 Option "AllowEmptyInput" "off"
1101 to your ServerLayout section. This will cause X to use the configured
1102 kbd and mouse sections from your xorg.conf.
1105 The GENERIC kernels for all architectures now default to the new USB2
1106 stack. No kernel config options or code have been removed so if a
1107 problem arises please report it and optionally revert to the old USB
1108 stack. If you are loading USB kernel modules or have a custom kernel
1109 that includes GENERIC then ensure that usb names are also changed over,
1110 eg uftdi -> usb2_serial_ftdi.
1112 Older programs linked against the ports libusb 0.1 need to be
1113 redirected to the new stack's libusb20. /etc/libmap.conf can
1115 # Map old usb library to new one for usb2 stack
1116 libusb-0.1.so.8 libusb20.so.1
1119 All USB ethernet devices now attach as interfaces under the name ueN
1120 (eg. ue0). This is to provide a predictable name as vendors often
1121 change usb chipsets in a product without notice.
1124 The ichsmb(4) driver has been changed to require SMBus slave
1125 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1126 All of the other SMBus controller drivers require left-justified
1127 slave addresses, so this change makes all the drivers provide the
1131 INET6 statistics (struct ip6stat) was updated.
1132 netstat(1) needs to be recompiled.
1135 NTFS has been removed from GENERIC kernel on amd64 to match
1136 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1137 will load ntfs.ko module automatically when NTFS support is
1138 actually needed, unless ntfs.ko is not installed or security
1139 level prohibits loading kernel modules. If either is the case,
1140 "options NTFS" has to be added into kernel config.
1143 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1144 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1145 800061. User space tools that rely on the size of struct tcpcb in
1146 tcp_var.h (e.g. sockstat) need to be recompiled.
1149 ng_tty(4) module updated to match the new TTY subsystem.
1150 Due to API change, user-level applications must be updated.
1151 New API support added to mpd5 CVS and expected to be present
1152 in next mpd5.3 release.
1155 With __FreeBSD_version 800060 the makefs tool is part of
1156 the base system (it was a port).
1159 The afdata and ifnet locks have been changed from mutexes to
1160 rwlocks, network modules will need to be re-compiled.
1163 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1164 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1165 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1166 architecture and 40 bytes on 64-bit architecture. The userland
1167 applications "arp" and "ndp" have been updated accordingly.
1168 The output from "netstat -r" shows only routing entries and
1169 none of the L2 information.
1172 __FreeBSD_version 800057 marks the switchover from the
1173 binary ath hal to source code. Users must add the line:
1175 options AH_SUPPORT_AR5416
1177 to their kernel config files when specifying:
1181 The ath_hal module no longer exists; the code is now compiled
1182 together with the driver in the ath module. It is now
1183 possible to tailor chip support (i.e. reduce the set of chips
1184 and thereby the code size); consult ath_hal(4) for details.
1187 __FreeBSD_version 800054 adds memory barriers to
1188 <machine/atomic.h>, new interfaces to ifnet to facilitate
1189 multiple hardware transmit queues for cards that support
1190 them, and a lock-less ring-buffer implementation to
1191 enable drivers to more efficiently manage queueing of
1195 A new version of ZFS (version 13) has been merged to -HEAD.
1196 This version has zpool attribute "listsnapshots" off by
1197 default, which means "zfs list" does not show snapshots,
1198 and is the same as Solaris behavior.
1201 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1204 The uhci, ohci, ehci and slhci USB Host controller drivers have
1205 been put into separate modules. If you load the usb module
1206 separately through loader.conf you will need to load the
1207 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1208 controller add the following to loader.conf:
1214 The ABI used by the PMC toolset has changed. Please keep
1215 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1219 atapci kernel module now includes only generic PCI ATA
1220 driver. AHCI driver moved to ataahci kernel module.
1221 All vendor-specific code moved into separate kernel modules:
1222 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1223 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1224 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1225 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1228 The TTY subsystem of the kernel has been replaced by a new
1229 implementation, which provides better scalability and an
1230 improved driver model. Most common drivers have been migrated to
1231 the new TTY subsystem, while others have not. The following
1232 drivers have not yet been ported to the new TTY layer:
1235 cy, digi, rc, rp, sio
1241 ng_h4, ng_tty, ppp, sl, snp
1243 Adding these drivers to your kernel configuration file shall
1244 cause compilation to fail.
1247 ntpd has been upgraded to 4.2.4p5.
1250 OpenSSH has been upgraded to 5.1p1.
1252 For many years, FreeBSD's version of OpenSSH preferred DSA
1253 over RSA for host and user authentication keys. With this
1254 upgrade, we've switched to the vendor's default of RSA over
1255 DSA. This may cause upgraded clients to warn about unknown
1256 host keys even for previously known hosts. Users should
1257 follow the usual procedure for verifying host keys before
1258 accepting the RSA key.
1260 This can be circumvented by setting the "HostKeyAlgorithms"
1261 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1264 Please note that the sequence of keys offered for
1265 authentication has been changed as well. You may want to
1266 specify IdentityFile in a different order to revert this
1270 The sio(4) driver has been removed from the i386 and amd64
1271 kernel configuration files. This means uart(4) is now the
1272 default serial port driver on those platforms as well.
1274 To prevent collisions with the sio(4) driver, the uart(4) driver
1275 uses different names for its device nodes. This means the
1276 onboard serial port will now most likely be called "ttyu0"
1277 instead of "ttyd0". You may need to reconfigure applications to
1278 use the new device names.
1280 When using the serial port as a boot console, be sure to update
1281 /boot/device.hints and /etc/ttys before booting the new kernel.
1282 If you forget to do so, you can still manually specify the hints
1283 at the loader prompt:
1285 set hint.uart.0.at="isa"
1286 set hint.uart.0.port="0x3F8"
1287 set hint.uart.0.flags="0x10"
1288 set hint.uart.0.irq="4"
1292 The gpt(8) utility has been removed. Use gpart(8) to partition
1296 The version that Linuxulator emulates was changed from 2.4.2
1297 to 2.6.16. If you experience any problems with Linux binaries
1298 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1299 if it fixes the problem contact emulation mailing list.
1302 ISDN4BSD (I4B) was removed from the src tree. You may need to
1303 update a your kernel configuration and remove relevant entries.
1306 I have checked in code to support multiple routing tables.
1307 See the man pages setfib(1) and setfib(2).
1308 This is a hopefully backwards compatible version,
1309 but to make use of it you need to compile your kernel
1310 with options ROUTETABLES=2 (or more up to 16).
1313 The 802.11 wireless support was redone to enable multi-bss
1314 operation on devices that are capable. The underlying device
1315 is no longer used directly but instead wlanX devices are
1316 cloned with ifconfig. This requires changes to rc.conf files.
1317 For example, change:
1318 ifconfig_ath0="WPA DHCP"
1321 ifconfig_wlan0="WPA DHCP"
1322 see rc.conf(5) for more details. In addition, mergemaster of
1323 /etc/rc.d is highly recommended. Simultaneous update of userland
1324 and kernel wouldn't hurt either.
1326 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1327 modules were merged into the base wlan module. All references
1328 to these modules (e.g. in kernel config files) must be removed.
1331 psm(4) has gained write(2) support in native operation level.
1332 Arbitrary commands can be written to /dev/psm%d and status can
1333 be read back from it. Therefore, an application is responsible
1334 for status validation and error recovery. It is a no-op in
1335 other operation levels.
1338 Support for KSE threading has been removed from the kernel. To
1339 run legacy applications linked against KSE libmap.conf may
1340 be used. The following libmap.conf may be used to ensure
1341 compatibility with any prior release:
1343 libpthread.so.1 libthr.so.1
1344 libpthread.so.2 libthr.so.2
1345 libkse.so.3 libthr.so.3
1348 The layout of struct vmspace has changed. This affects libkvm
1349 and any executables that link against libkvm and use the
1350 kvm_getprocs() function. In particular, but not exclusively,
1351 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1352 The effects are minimal, but it's advisable to upgrade world
1356 The latest em driver no longer has support in it for the
1357 82575 adapter, this is now moved to the igb driver. The
1358 split was done to make new features that are incompatible
1359 with older hardware easier to do.
1362 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1363 likewise the kernel option is now GEOM_LINUX_LVM.
1366 The default NFS mount mode has changed from UDP to TCP for
1367 increased reliability. If you rely on (insecurely) NFS
1368 mounting across a firewall you may need to update your
1372 Belatedly note the addition of m_collapse for compacting
1376 The fts(3) structures have been changed to use adequate
1377 integer types for their members and so to be able to cope
1378 with huge file trees. The old fts(3) ABI is preserved
1379 through symbol versioning in libc, so third-party binaries
1380 using fts(3) should still work, although they will not take
1381 advantage of the extended types. At the same time, some
1382 third-party software might fail to build after this change
1383 due to unportable assumptions made in its source code about
1384 fts(3) structure members. Such software should be fixed
1385 by its vendor or, in the worst case, in the ports tree.
1386 FreeBSD_version 800015 marks this change for the unlikely
1387 case that a portable fix is impossible.
1390 To upgrade to -current after this date, you must be running
1391 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1392 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1395 The ADAPTIVE_GIANT kernel option has been retired because its
1396 functionality is the default now.
1399 The AT keyboard emulation of sunkbd(4) has been turned on
1400 by default. In order to make the special symbols of the Sun
1401 keyboards driven by sunkbd(4) work under X these now have
1402 to be configured the same way as Sun USB keyboards driven
1403 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1405 Option "XkbLayout" "us"
1406 Option "XkbRules" "xorg"
1407 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1410 It has been decided that it is desirable to provide ABI
1411 backwards compatibility to the FreeBSD 4/5/6 versions of the
1412 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1413 broken with the introduction of PCI domain support (see the
1414 20070930 entry). Unfortunately, this required the ABI of
1415 PCIOCGETCONF to be broken again in order to be able to
1416 provide backwards compatibility to the old version of that
1417 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1418 again. As for prominent ports this affects neither pciutils
1419 nor xorg-server this time, the hal port needs to be rebuilt
1423 The misnamed kthread_create() and friends have been renamed
1424 to kproc_create() etc. Many of the callers already
1425 used kproc_start()..
1426 I will return kthread_create() and friends in a while
1427 with implementations that actually create threads, not procs.
1428 Renaming corresponds with version 800002.
1437 Avoid using make -j when upgrading. While generally safe, there are
1438 sometimes problems using -j to upgrade. If your upgrade fails with
1439 -j, please try again without -j. From time to time in the past there
1440 have been problems using -j with buildworld and/or installworld. This
1441 is especially true when upgrading between "distant" versions (eg one
1442 that cross a major release boundary or several minor releases, or when
1443 several months have passed on the -current branch).
1445 Sometimes, obscure build problems are the result of environment
1446 poisoning. This can happen because the make utility reads its
1447 environment when searching for values for global variables. To run
1448 your build attempts in an "environmental clean room", prefix all make
1449 commands with 'env -i '. See the env(1) manual page for more details.
1451 When upgrading from one major version to another it is generally best
1452 to upgrade to the latest code in the currently installed branch first,
1453 then do an upgrade to the new branch. This is the best-tested upgrade
1454 path, and has the highest probability of being successful. Please try
1455 this approach before reporting problems with a major version upgrade.
1459 When upgrading the boot ZFS pool to a new version, always follow
1462 1.) recompile and reinstall the ZFS boot loader and boot block
1463 (this is part of "make buildworld" and "make installworld")
1465 2.) update the ZFS boot block on your boot drive
1467 The following example updates the ZFS boot block on the first
1468 partition (freebsd-boot) of a GPT partitioned drive ad0:
1469 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1471 Non-boot pools do not need these updates.
1475 If you are updating from a prior version of FreeBSD (even one just
1476 a few days old), you should follow this procedure. It is the most
1477 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1479 make kernel-toolchain
1480 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1481 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1483 To test a kernel once
1484 ---------------------
1485 If you just want to boot a kernel once (because you are not sure
1486 if it works, or if you want to boot a known bad kernel to provide
1487 debugging information) run
1488 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1489 nextboot -k testkernel
1491 To just build a kernel when you know that it won't mess you up
1492 --------------------------------------------------------------
1493 This assumes you are already running a CURRENT system. Replace
1494 ${arch} with the architecture of your machine (e.g. "i386",
1495 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1497 cd src/sys/${arch}/conf
1498 config KERNEL_NAME_HERE
1499 cd ../compile/KERNEL_NAME_HERE
1504 If this fails, go to the "To build a kernel" section.
1506 To rebuild everything and install it on the current system.
1507 -----------------------------------------------------------
1508 # Note: sometimes if you are running current you gotta do more than
1509 # is listed here if you are upgrading from a really old current.
1511 <make sure you have good level 0 dumps>
1513 make kernel KERNCONF=YOUR_KERNEL_HERE
1515 <reboot in single user> [3]
1523 To cross-install current onto a separate partition
1524 --------------------------------------------------
1525 # In this approach we use a separate partition to hold
1526 # current's root, 'usr', and 'var' directories. A partition
1527 # holding "/", "/usr" and "/var" should be about 2GB in
1530 <make sure you have good level 0 dumps>
1533 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1534 <maybe newfs current's root partition>
1535 <mount current's root partition on directory ${CURRENT_ROOT}>
1536 make installworld DESTDIR=${CURRENT_ROOT}
1537 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1538 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1539 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1540 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1541 <reboot into current>
1542 <do a "native" rebuild/install as described in the previous section>
1543 <maybe install compatibility libraries from ports/misc/compat*>
1547 To upgrade in-place from stable to current
1548 ----------------------------------------------
1549 <make sure you have good level 0 dumps>
1551 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1553 <reboot in single user> [3]
1560 Make sure that you've read the UPDATING file to understand the
1561 tweaks to various things you need. At this point in the life
1562 cycle of current, things change often and you are on your own
1563 to cope. The defaults can also change, so please read ALL of
1564 the UPDATING entries.
1566 Also, if you are tracking -current, you must be subscribed to
1567 freebsd-current@freebsd.org. Make sure that before you update
1568 your sources that you have read and understood all the recent
1569 messages there. If in doubt, please track -stable which has
1570 much fewer pitfalls.
1572 [1] If you have third party modules, such as vmware, you
1573 should disable them at this point so they don't crash your
1576 [3] From the bootblocks, boot -s, and then do
1581 adjkerntz -i # if CMOS is wall time
1582 Also, when doing a major release upgrade, it is required that
1583 you boot into single user mode to do the installworld.
1585 [4] Note: This step is non-optional. Failure to do this step
1586 can result in a significant reduction in the functionality of the
1587 system. Attempting to do it by hand is not recommended and those
1588 that pursue this avenue should read this file carefully, as well
1589 as the archives of freebsd-current and freebsd-hackers mailing lists
1590 for potential gotchas. The -U option is also useful to consider.
1591 See mergemaster(8) for more information.
1593 [5] Usually this step is a noop. However, from time to time
1594 you may need to do this if you get unknown user in the following
1595 step. It never hurts to do it all the time. You may need to
1596 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1597 install) after the buildworld before this step if you last updated
1598 from [78]-stable or 9-stable before 20130430.
1600 [6] This only deletes old files and directories. Old libraries
1601 can be deleted by "make delete-old-libs", but you have to make
1602 sure that no program is using those libraries anymore.
1604 [8] In order to have a kernel that can run the 4.x binaries needed to
1605 do an installworld, you must include the COMPAT_FREEBSD4 option in
1606 your kernel. Failure to do so may leave you with a system that is
1607 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1608 required to run the 5.x binaries on more recent kernels. And so on
1609 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1611 Make sure that you merge any new devices from GENERIC since the
1612 last time you updated your kernel config file.
1614 [9] When checking out sources, you must include the -P flag to have
1615 cvs prune empty directories.
1617 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1618 "?=" instead of the "=" assignment operator, so that buildworld can
1619 override the CPUTYPE if it needs to.
1621 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1622 not on the command line, or in /etc/make.conf. buildworld will
1623 warn if it is improperly defined.
1626 This file contains a list, in reverse chronological order, of major
1627 breakages in tracking -current. It is not guaranteed to be a complete
1628 list of such breakages, and only contains entries since October 10, 2007.
1629 If you need to see UPDATING entries from before that date, you will need
1630 to fetch an UPDATING file from an older FreeBSD release.
1632 Copyright information:
1634 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1636 Redistribution, publication, translation and use, with or without
1637 modification, in full or in part, in any form or format of this
1638 document are permitted without further permission from the author.
1640 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1641 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1642 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1643 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1644 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1645 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1646 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1647 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1648 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1649 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1650 POSSIBILITY OF SUCH DAMAGE.
1652 Contact Warner Losh if you have any questions about your use of