1 Updating Information for FreeBSD current users
3 This file is maintained and copyrighted by M. Warner Losh <imp@freebsd.org>.
4 See end of file for further details. For commonly done items, please see the
5 COMMON ITEMS: section later in the file. These instructions assume that you
6 basically know what you are doing. If not, then please consult the FreeBSD
9 http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html
11 Items affecting the ports and packages system can be found in
12 /usr/ports/UPDATING. Please read that file before running portupgrade.
15 It is no longer necessary to include the dwarf version in your DEBUG
16 options in your kernel config file. The bug that required it to be
17 placed in the config file has bene fixed. DEBUG should now just
18 contain -g. The build system will automatically update things
19 to do the right thing.
25 On i386 and amd64 systems, the onifconsole flag is now set by default
26 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
27 as a login TTY if it is set in the bootloader as an active kernel
28 console. No changes in behavior should result otherwise. To revert to
29 the previous behavior, set ttyu0 to "off" in /etc/ttys.
32 Clang and llvm have been upgraded to 3.4.1 release.
35 Clang and llvm have been upgraded to 3.4 release.
38 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
39 been deprecated and will not be part of FreeBSD 11.0 and later
40 releases. If you use this driver, please consider switching to
41 the nfe(4) driver instead.
44 The behavior of gss_pseudo_random() for the krb5 mechanism
45 has changed, for applications requesting a longer random string
46 than produced by the underlying enctype's pseudo-random() function.
47 In particular, the random string produced from a session key of
48 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
49 be different at the 17th octet and later, after this change.
50 The counter used in the PRF+ construction is now encoded as a
51 big-endian integer in accordance with RFC 4402.
52 __FreeBSD_version is bumped to 902505.
58 Behavior of devfs rules path matching has been changed.
59 Pattern is now always matched against fully qualified devfs
60 path and slash characters must be explicitly matched by
61 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
62 subdirectories must be reviewed.
65 hastctl(8)'s `status' command output changed to terse one-liner format.
66 Scripts using this should switch to `list' command or be rewritten.
69 Fix a bug that allowed a tracing process (e.g. gdb) to write
70 to a memory-mapped file in the traced process's address space
71 even if neither the traced process nor the tracing process had
72 write access to that file.
75 Added ZFS TRIM support which is enabled by default. To disable
76 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
78 Creating new ZFS pools and adding new devices to existing pools
79 first performs a full device level TRIM which can take a significant
80 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
81 to disable this behaviour.
83 ZFS TRIM requires the underlying device support BIO_DELETE which
84 is currently provided by methods such as ATA TRIM and SCSI UNMAP
85 via CAM, which are typically supported by SSD's.
87 Stats for ZFS TRIM can be monitored by looking at the sysctl's
88 under kstat.zfs.misc.zio_trim.
91 `list' command has been added to hastctl(8). For now, it is full
92 equivalent of `status' command.
93 WARNING: in the near future the output of hastctl's status command
94 will change to more terse format. If you use `hastctl status'
95 for parsing in your scripts, switch to `hastctl list'.
98 The mergemaster command now uses the default MAKEOBJDIRPREFIX
99 rather than creating it's own in the temporary directory in
100 order allow access to bootstrapped versions of tools such as
101 install and mtree. When upgrading from version of FreeBSD where
102 the install command does not support -l, you will need to
103 install a new mergemaster command if mergemaster -p is required.
104 This can be accomplished with the command (cd src/usr.sbin/mergemaster
107 Due to the use of the new -l option to install(1) during build
108 and install, you must take care not to directly set the INSTALL
109 make variable in your /etc/make.conf, /etc/src.conf, or on the
110 command line. If you wish to use the -C flag for all installs
111 you may be able to add INSTALL+=-C to /etc/make.conf or
115 Fix a bug that allows NFS clients to issue READDIR on files.
118 The install(1) option -M has changed meaning and now takes an
119 argument that is a file or path to append logs to. In the
120 unlikely event that -M was the last option on the command line
121 and the command line contained at least two files and a target
122 directory the first file will have logs appended to it. The -M
123 option served little practical purpose in the last decade so it's
124 used expected to be extremely rare.
127 A new compression method (lz4) has been merged to. Please refer to
128 zpool-features(7) for more information.
130 Please refer to the "ZFS notes" section of this file for information
131 on upgrading boot ZFS pools.
134 The VFS KBI was changed with the merge of several nullfs
135 optimizations and fixes. All filesystem modules must be
139 With the addition of auditdistd(8), a new auditdistd user is now
140 depended on during installworld. "mergemaster -p" can be used to add
141 the user prior to installworld, as documented in the handbook.
147 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
148 Starting with this version the old system of ZFS pool versioning
149 is superseded by "feature flags". This concept enables forward
150 compatibility against certain future changes in functionality of ZFS
151 pools. The first two read-only compatible "feature flags" for ZFS
152 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
153 For more information read the new zpool-features(7) manual page.
154 Please refer to the "ZFS notes" section of this file for information
155 on upgrading boot ZFS pools.
158 The commit introducing bsd.compiler.mk breaks the traditional
159 building of kernels before this point. Add -m ${SRC}/share/mk
160 (for the right value of SRC) to your command lines to work
161 around; update your useland to a point after this; or use the
162 buildkernel/installkernel top-level targets. See also 20120829.
165 The IPFIREWALL_FORWARD kernel option has been removed. Its
166 functionality now turned on by default.
169 The random(4) support for the VIA hardware random number
170 generator (`PADLOCK') is no longer enabled unconditionally.
171 Add the PADLOCK_RNG option in the custom kernel config if
172 needed. The GENERIC kernels on i386 and amd64 do include the
173 option, so the change only affects the custom kernel
177 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
178 the traditional method, you must update your system with an installworld
179 before the kernel will build. The documented make buildkernel/installkernel
180 interfaces (coupled with fresh make kernel-toolchain) continue to work.
183 The sparc64 ZFS loader has been changed to no longer try to auto-
184 detect ZFS providers based on diskN aliases but now requires these
185 to be explicitly listed in the OFW boot-device environment variable.
188 Now unix domain sockets behave "as expected" on nullfs(5). Previously
189 nullfs(5) did not pass through all behaviours to the underlying layer,
190 as a result if we bound to a socket on the lower layer we could connect
191 only to the lower path; if we bound to the upper layer we could connect
192 only to the upper path. The new behavior is one can connect to both the
193 lower and the upper paths regardless what layer path one binds to.
196 The acpi_wmi(4) status device /dev/wmistat has been renamed to
200 A new VOP_ADVISE() was added to support posix_fadvise(2). All
201 filesystem modules must be recompiled.
204 The interface of the VOP_VPTOCNP(9) changed, now the returned
205 vnode shall be referenced, previously it was required to be
206 only held. All in-tree filesystems are converted.
212 The broken amd(4) driver has been replaced with esp(4) in the amd64,
213 i386 and pc98 GENERIC kernel configuration files.
216 This commit modifies vfs_register() so that it uses a hash
217 calculation to set vfc_typenum, which is enabled by default.
218 The first time a system is booted after this change, the
219 vfc_typenum values will change for all file systems. The
220 main effect of this is a change to the NFS server file handles
221 for file systems that use vfc_typenum in their fsid, such as ZFS.
222 It will, however, prevent vfc_typenum from changing when file
223 systems are loaded in a different order for subsequent reboots.
224 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
225 until you are ready to remount all NFS clients after a reboot.
228 Bump the shared library version numbers for libraries that
229 do not use symbol versioning, have changed the ABI compared
230 to stable/8 and which shared library version was not bumped.
231 Done as part of 9.0-RELEASE cycle.
234 During the merge of Capsicum features, the fget(9) KPI was modified.
235 This may require the rebuilding of out-of-tree device drivers --
236 issues have been reported specifically with the nVidia device driver.
237 __FreeBSD_version is bumped to 900041.
239 Also, there is a period between 20110811 and 20110814 where the
240 special devices /dev/{stdin,stdout,stderr} did not work correctly.
241 Building world from a kernel during that window may not work.
244 The packet filter (pf) code has been updated to OpenBSD 4.5.
245 You need to update userland tools to be in sync with kernel.
246 This update breaks backward compatibility with earlier pfsync(4)
247 versions. Care must be taken when updating redundant firewall setups.
250 The following sysctls and tunables are retired on x86 platforms:
252 machdep.hlt_logical_cpus
253 The following sysctl is retired:
254 machdep.hyperthreading_allowed
255 The sysctls were supposed to provide a way to dynamically offline and
256 online selected CPUs on x86 platforms, but the implementation has not
257 been reliable especially with SCHED_ULE scheduler.
258 machdep.hyperthreading_allowed tunable is still available to ignore
259 hyperthreading CPUs at OS level.
260 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
261 where X is an APIC ID of a CPU. Be advised, though, that disabling
262 CPUs in non-uniform fashion will result in non-uniform topology and
263 may lead to sub-optimal system performance with SCHED_ULE, which is
267 cpumask_t type is retired and cpuset_t is used in order to describe
271 Changes to ifconfig(8) for dynamic address family detection mandate
272 that you are running a kernel of 20110525 or later. Make sure to
273 follow the update procedure to boot a new kernel before installing
277 Support for sun4v architecture is officially dropped
280 Several KPI breaking changes have been committed to the mii(4) layer,
281 the PHY drivers and consequently some Ethernet drivers using mii(4).
282 This means that miibus.ko and the modules of the affected Ethernet
283 drivers need to be recompiled.
285 Note to kernel developers: Given that the OUI bit reversion problem
286 was fixed as part of these changes all mii(4) commits related to OUIs,
287 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
288 handling, no longer can be merged verbatim to stable/8 and previous
292 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
293 into their kernel configurations along with 'device pci'.
296 The default NFS client is now the new NFS client, so fstype "newnfs"
297 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
298 Although mounts via fstype "nfs" will usually work without userland
299 changes, it is recommended that the mount(8) and mount_nfs(8)
300 commands be rebuilt from sources and that a link to mount_nfs called
301 mount_oldnfs be created. The new client is compiled into the
302 kernel with "options NFSCL" and this is needed for diskless root
303 file systems. The GENERIC kernel configs have been changed to use
304 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
305 To use the regular/old client, you can "mount -t oldnfs ...". For
306 a diskless root file system, you must also include a line like:
308 vfs.root.mountfrom="oldnfs:"
310 in the boot/loader.conf on the root fs on the NFS server to make
311 a diskless root fs use the old client.
314 The GENERIC kernels for all architectures now default to the new
315 CAM-based ATA stack. It means that all legacy ATA drivers were
316 removed and replaced by respective CAM drivers. If you are using
317 ATA device names in /etc/fstab or other places, make sure to update
318 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
319 where 'Y's are the sequential numbers starting from zero for each type
320 in order of detection, unless configured otherwise with tunables,
321 see cam(4)). There will be symbolic links created in /dev/ to map
322 old adX devices to the respective adaY. They should provide basic
323 compatibility for file systems mounting in most cases, but they do
324 not support old user-level APIs and do not have respective providers
325 in GEOM. Consider using updated management tools with new device names.
327 It is possible to load devices ahci, ata, siis and mvs as modules,
328 but option ATA_CAM should remain in kernel configuration to make ata
329 module work as CAM driver supporting legacy ATA controllers. Device ata
330 still can be used in modular fashion (atacore + ...). Modules atadisk
331 and atapi* are not used and won't affect operation in ATA_CAM mode.
332 Note that to use CAM-based ATA kernel should include CAM devices
333 scbus, pass, da (or explicitly ada), cd and optionally others. All of
334 them are parts of the cam module.
336 ataraid(4) functionality is now supported by the RAID GEOM class.
337 To use it you can load geom_raid kernel module and use graid(8) tool
338 for management. Instead of /dev/arX device names, use /dev/raid/rX.
340 No kernel config options or code have been removed, so if a problem
341 arises, please report it and optionally revert to the old ATA stack.
342 In order to do it you can remove from the kernel config:
347 , and instead add back:
348 device atadisk # ATA disk drives
349 device ataraid # ATA RAID drives
350 device atapicd # ATAPI CDROM drives
351 device atapifd # ATAPI floppy drives
352 device atapist # ATAPI tape drives
355 The default NFS server has been changed to the new server, which
356 was referred to as the experimental server. If you need to switch
357 back to the old NFS server, you must now put the "-o" option on
358 both the mountd and nfsd commands. This can be done using the
359 mountd_flags and nfs_server_flags rc.conf variables until an
360 update to the rc scripts is committed, which is coming soon.
363 The GNU Objective-C runtime library (libobjc), and other Objective-C
364 related components have been removed from the base system. If you
365 require an Objective-C library, please use one of the available ports.
368 ath(4) has been split into bus- and device- modules. if_ath contains
369 the HAL, the TX rate control and the network device code. if_ath_pci
370 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
371 contains the AHB glue. Users need to load both if_ath_pci and if_ath
372 in order to use ath on everything else.
374 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
375 need to load if_ath and if_ath_pci for ath(4) operation.
378 As part of the replacement of sysinstall, the process of building
379 release media has changed significantly. For details, please re-read
380 release(7), which has been updated to reflect the new build process.
383 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
384 is the last available version under GPLv2. It brings a number of new
385 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
386 4.1 and SSE 4.2), better support for powerpc64, a number of new
387 directives, and lots of other small improvements. See the ChangeLog
388 file in contrib/binutils for the full details.
391 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
392 compliant, and will now use half of hash for authentication.
393 This will break interoperability with all stacks (including all
394 actual FreeBSD versions) who implement
395 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
397 The only workaround with such peers is to use another HMAC
398 algorithm for IPsec ("phase 2") authentication.
401 Remove the uio_yield prototype and symbol. This function has
402 been misnamed since it was introduced and should not be
403 globally exposed with this name. The equivalent functionality
404 is now available using kern_yield(curthread->td_user_pri).
405 The function remains undocumented.
408 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
409 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
410 for scalar sysctls is defined but disabled. Code that needs
411 UQUAD to pass the type checking that must compile on older
412 systems where the define is not present can check against
413 __FreeBSD_version >= 900030.
415 The system dialog(1) has been replaced with a new version previously
416 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
417 with the previous version, but the libdialog associated with it has
418 a largely incompatible API. As such, the original version of libdialog
419 will be kept temporarily as libodialog, until its base system consumers
420 are replaced or updated. Bump __FreeBSD_version to 900030.
423 If you are trying to run make universe on a -stable system, and you get
424 the following warning:
425 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
426 unknown. config(8) likely too old."
427 or something similar to it, then you must upgrade your -stable system
428 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
429 stable/8) or build the config from the latest stable/8 branch and
430 install it on your system.
432 Prior to this date, building a current universe on 8-stable system from
433 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
434 error in the first kernel build phase. A new config on those old
435 systems will fix that problem for older versions of -current.
438 The TCP stack has been modified to allow Khelp modules to interact with
439 it via helper hook points and store per-connection data in the TCP
440 control block. Bump __FreeBSD_version to 900029. User space tools that
441 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
445 Generic IEEE 802.3 annex 31B full duplex flow control support has been
446 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
447 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
448 to take advantage of it instead of using custom implementations. This
449 means that these drivers now no longer unconditionally advertise
450 support for flow control but only do so if flow control is a selected
451 media option. This was implemented in the generic support that way in
452 order to allow flow control to be switched on and off via ifconfig(8)
453 with the PHY specific default to typically off in order to protect
454 from unwanted effects. Consequently, if you used flow control with
455 one of the above mentioned drivers you now need to explicitly enable
457 ifconfig bge0 media auto mediaopt flowcontrol
459 Along with the above mentioned changes generic support for setting
460 1000baseT master mode also has been added and brgphy(4), ciphy(4),
461 e1000phy(4) as well as ip1000phy(4) have been converted to take
462 advantage of it. This means that these drivers now no longer take the
463 link0 parameter for selecting master mode but the master media option
464 has to be used instead, for example like in the following:
465 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
467 Selection of master mode now is also available with all other PHY
468 drivers supporting 1000baseT.
471 The TCP stack has received a significant update to add support for
472 modularised congestion control and generally improve the clarity of
473 congestion control decisions. Bump __FreeBSD_version to 900025. User
474 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
475 sockstat) need to be recompiled.
478 The man(1) utility has been replaced by a new version that no longer
479 uses /etc/manpath.config. Please consult man.conf(5) for how to
480 migrate local entries to the new format.
483 The copyright strings printed by login(1) and sshd(8) at the time of a
484 new connection have been removed to follow other operating systems and
488 A workaround for a fixed ld bug has been removed in kernel code,
489 so make sure that your system ld is built from sources after
490 revision 210245 from 2010-07-19 (r211583 if building head kernel
491 on stable/8, r211584 for stable/7; both from 2010-08-21).
492 A symptom of incorrect ld version is different addresses for
493 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
496 The $ipv6_prefer variable in rc.conf(5) has been split into
497 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
499 The $ip6addrctl_policy is a variable to choose a pre-defined
500 address selection policy set by ip6addrctl(8). A value
501 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
504 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
505 flag (see an entry of 20090926) is set on an interface with no
506 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
507 security reason. If you want IPv6 link-local address on all
508 interfaces by default, set this to "YES".
510 The old ipv6_prefer="YES" is equivalent to
511 ipv6_activate_all_interfaces="YES" and
512 ip6addrctl_policy="ipv6_prefer".
515 DTrace has grown support for userland tracing. Due to this, DTrace is
516 now i386 and amd64 only.
517 dtruss(1) is now installed by default on those systems and a new
518 kernel module is needed for userland tracing: fasttrap.
519 No changes to your kernel config file are necessary to enable
520 userland tracing, but you might consider adding 'STRIP=' and
521 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
522 to have informative userland stack traces in DTrace (ustack).
525 The acpi_aiboost(4) driver has been removed in favor of the new
526 aibs(4) driver. You should update your kernel configuration file.
529 BSD grep has been imported to the base system and it is built by
530 default. It is completely BSD licensed, highly GNU-compatible, uses
531 less memory than its GNU counterpart and has a small codebase.
532 However, it is slower than its GNU counterpart, which is mostly
533 noticeable for larger searches, for smaller ones it is measurable
534 but not significant. The reason is complex, the most important factor
535 is that we lack a modern and efficient regex library and GNU
536 overcomes this by optimizing the searches internally. Future work
537 on improving the regex performance is planned, for the meantime,
538 users that need better performance, can build GNU grep instead by
539 setting the WITH_GNU_GREP knob.
542 Due to the import of powerpc64 support, all existing powerpc kernel
543 configuration files must be updated with a machine directive like this:
544 machine powerpc powerpc
546 In addition, an updated config(8) is required to build powerpc kernels
550 A new version of ZFS (version 15) has been merged to -HEAD.
551 This version uses a python library for the following subcommands:
552 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
553 For full functionality of these commands the following port must
554 be installed: sysutils/py-zfs
557 'vm_page's are now hashed by physical address to an array of mutexes.
558 Currently this is only used to serialize access to hold_count. Over
559 time the page queue mutex will be peeled away. This changes the size
560 of pmap on every architecture. And requires all callers of vm_page_hold
561 and vm_page_unhold to be updated.
564 WITH_CTF can now be specified in src.conf (not recommended, there
565 are some problems with static executables), make.conf (would also
566 affect ports which do not use GNU make and do not override the
567 compile targets) or in the kernel config (via "makeoptions
569 When WITH_CTF was specified there before this was silently ignored,
570 so make sure that WITH_CTF is not used in places which could lead
571 to unwanted behavior.
574 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
575 to allow 32-bit compatibility on non-x86 platforms. All kernel
576 configurations on amd64 and ia64 platforms using these options must
577 be modified accordingly.
580 The utmp user accounting database has been replaced with utmpx,
581 the user accounting interface standardized by POSIX.
582 Unfortunately the semantics of utmp and utmpx don't match,
583 making it practically impossible to support both interfaces.
584 The user accounting database is used by tools like finger(1),
585 last(1), talk(1), w(1) and ac(8).
587 All applications in the base system use utmpx. This means only
588 local binaries (e.g. from the ports tree) may still use these
589 utmp database files. These applications must be rebuilt to make
592 After the system has been upgraded, it is safe to remove the old
593 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
594 assuming their contents is of no importance anymore. Old wtmp
595 databases can only be used by last(1) and ac(8) after they have
596 been converted to the new format using wtmpcvt(1).
599 Introduce the kernel thread "deadlock resolver" (which can be enabled
600 via the DEADLKRES option, see NOTES for more details) and the
601 sleepq_type() function for sleepqueues.
604 The rc.firewall and rc.firewall6 were unified, and
605 rc.firewall6 and rc.d/ip6fw were removed.
606 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
607 variables are obsoleted. Instead, the following new rc
608 variables are added to rc.d/ipfw:
610 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
611 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
612 firewall_simple_onet_ipv6, firewall_trusted_ipv6
614 The meanings correspond to the relevant IPv4 variables.
620 The default terminal emulation for syscons(4) has been changed
621 from cons25 to xterm on all platforms except pc98. This means
622 that the /etc/ttys file needs to be updated to ensure correct
623 operation of applications on the console.
625 The terminal emulation style can be toggled per window by using
626 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
627 options can be used to change the compile-time default back to
630 To prevent graphical artifacts, make sure the TERM environment
631 variable is set to match the terminal emulation that is being
632 performed by syscons(4).
635 The layout of the structure ieee80211req_scan_result has changed.
636 Applications that require wireless scan results (e.g. ifconfig(8))
637 from net80211 need to be recompiled.
639 Applications such as wpa_supplicant(8) may require a full world
640 build without using NO_CLEAN in order to get synchronized with the
644 The iwn(4) driver has been updated to support the 5000 and 5150 series.
645 There's one kernel module for each firmware. Adding "device iwnfw"
646 to the kernel configuration file means including all three firmware
647 images inside the kernel. If you want to include just the one for
648 your wireless card, use the devices iwn4965fw, iwn5000fw or
652 The rc.d/network_ipv6, IPv6 configuration script has been integrated
653 into rc.d/netif. The changes are the following:
655 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
656 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
657 Note that both variables need the "inet6" keyword at the head.
659 Do not set $ipv6_network_interfaces manually if you do not
660 understand what you are doing. It is not needed in most cases.
662 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
665 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
666 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
668 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
669 all configured interfaces have "inet6 accept_rtadv" in the
670 $ifconfig_IF_ipv6. These are for backward compatibility.
672 3. A new variable $ipv6_prefer has been added. If NO, IPv6
673 functionality of interfaces with no corresponding
674 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
675 and the default address selection policy of ip6addrctl(8)
676 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
677 Note that if you want to configure IPv6 functionality on the
678 disabled interfaces after boot, first you need to clear the flag by
679 using ifconfig(8) like:
681 ifconfig em0 inet6 -ifdisabled
683 If YES, the default address selection policy is set as
686 The default value of $ipv6_prefer is NO.
688 4. If your system need to receive Router Advertisement messages,
689 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
690 scripts automatically invoke rtsol(8) when the interface becomes
691 UP. The Router Advertisement messages are used for SLAAC
692 (State-Less Address AutoConfiguration).
695 802.11s D3.03 support was committed. This is incompatible with the
696 previous code, which was based on D3.0.
699 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
700 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
701 control whether accepting Router Advertisement messages or not.
702 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
703 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
704 The ifconfig(8) utility now supports these flags.
707 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
708 mount(8) and -a option for df(1) to see them.
711 The old tunable hw.bus.devctl_disable has been superseded by
712 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
713 replaced by hw.bus.devctl_queue=0. The default for this new tunable
717 Remove the option STOP_NMI. The default action is now to use NMI only
718 for KDB via the newly introduced function stop_cpus_hard() and
719 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
722 The stable/8 branch created in subversion. This corresponds to the
723 RELENG_8 branch in CVS.
726 Bump the shared library version numbers for all libraries that do not
727 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
728 __FreeBSD_version to 800105.
731 Due to changes in the implementation of virtual network stack support,
732 all network-related kernel modules must be recompiled. As this change
733 breaks the ABI, bump __FreeBSD_version to 800104.
736 The TOE interface to the TCP syncache has been modified to remove
737 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
738 The cxgb driver is the only TOE consumer affected by this change, and
739 needs to be recompiled along with the kernel. As this change breaks
740 the ABI, bump __FreeBSD_version to 800103.
743 Padding has been added to struct tcpcb, sackhint and tcpstat in
744 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
745 maintaining the ABI. However, this change breaks the ABI, so bump
746 __FreeBSD_version to 800102. User space tools that rely on the size of
747 any of these structs (e.g. sockstat) need to be recompiled.
750 The NFS_LEGACYRPC option has been removed along with the old kernel
751 RPC implementation that this option selected. Kernel configurations
752 may need to be adjusted.
755 The network interface device nodes at /dev/net/<interface> have been
756 removed. All ioctl operations can be performed the normal way using
757 routing sockets. The kqueue functionality can generally be replaced
758 with routing sockets.
761 The documentation from the FreeBSD Documentation Project (Handbook,
762 FAQ, etc.) is now installed via packages by sysinstall(8) and under
763 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
766 The ABI of various structures related to the SYSV IPC API have been
767 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
768 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
772 Layout of struct vnet has changed as routing related variables were
773 moved to their own Vimage module. Modules need to be recompiled. Bump
774 __FreeBSD_version to 800099.
777 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
778 respectively. As long as no more than 16 groups per process are used,
779 no changes should be visible. When more than 16 groups are used, old
780 binaries may fail if they call getgroups() or getgrouplist() with
781 statically sized storage. Recompiling will work around this, but
782 applications should be modified to use dynamically allocated storage
783 for group arrays as POSIX.1-2008 does not cap an implementation's
784 number of supported groups at NGROUPS_MAX+1 as previous versions did.
786 NFS and portalfs mounts may also be affected as the list of groups is
787 truncated to 16. Users of NFS who use more than 16 groups, should
788 take care that negative group permissions are not used on the exported
789 file systems as they will not be reliable unless a GSSAPI based
790 authentication method is used.
793 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
794 option compiles in the support for adaptive spinning for lockmgrs
795 which want to enable it. The lockinit() function now accepts the flag
796 LK_ADAPTIVE in order to make the lock object subject to adaptive
797 spinning when both held in write and read mode.
800 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
801 changed. User applications that use this ioctl need to be rebuilt.
804 The layout of struct thread has changed. Kernel and modules need to
808 The layout of structs ifnet, domain, protosw and vnet_net has changed.
809 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
812 window(1) has been removed from the base system. It can now be
813 installed from ports. The port is called misc/window.
816 The way we are storing and accessing `routing table' entries has
817 changed. Programs reading the FIB, like netstat, need to be
821 A new netisr implementation has been added for FreeBSD 8. Network
822 file system modules, such as igmp, ipdivert, and others, should be
824 Bump __FreeBSD_version to 800096.
827 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
831 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
832 Bump __FreeBSD_version to 800094.
835 Add mnt_xflag field to 'struct mount'. File system modules need to be
837 Bump __FreeBSD_version to 800093.
840 The compiling option ADAPTIVE_SX has been retired while it has been
841 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
842 The KPI for sx_init_flags() changes as accepting flags:
843 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
844 been introduced in order to handle the reversed logic.
845 Bump __FreeBSD_version to 800092.
848 Add support for hierarchical jails. Remove global securelevel.
849 Bump __FreeBSD_version to 800091.
852 The layout of struct vnet_net has changed, therefore modules
854 Bump __FreeBSD_version to 800090.
857 The newly imported zic(8) produces a new format in the output. Please
858 run tzsetup(8) to install the newly created data to /etc/localtime.
861 The sysctl tree for the usb stack has renamed from hw.usb2.* to
862 hw.usb.* and is now consistent again with previous releases.
865 802.11 monitor mode support was revised and driver api's were changed.
866 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
867 of DLT_IEEE802_11. No user-visible data structures were changed but
868 applications that use DLT_IEEE802_11 may require changes.
869 Bump __FreeBSD_version to 800088.
872 The layout of the following structs has changed: sysctl_oid,
873 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
874 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
875 panics may be experienced. World rebuild is required for
876 correctly checking networking state from userland.
877 Bump __FreeBSD_version to 800085.
880 MLDv2 and Source-Specific Multicast (SSM) have been merged
881 to the IPv6 stack. VIMAGE hooks are in but not yet used.
882 The implementation of SSM within FreeBSD's IPv6 stack closely
883 follows the IPv4 implementation.
885 For kernel developers:
887 * The most important changes are that the ip6_output() and
888 ip6_input() paths no longer take the IN6_MULTI_LOCK,
889 and this lock has been downgraded to a non-recursive mutex.
891 * As with the changes to the IPv4 stack to support SSM, filtering
892 of inbound multicast traffic must now be performed by transport
893 protocols within the IPv6 stack. This does not apply to TCP and
894 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
896 * The KPIs used by IPv6 multicast are similar to those used by
897 the IPv4 stack, with the following differences:
898 * im6o_mc_filter() is analogous to imo_multicast_filter().
899 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
900 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
901 * IN6_LOOKUP_MULTI() has been deprecated and removed.
902 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
903 for MLDv1 have an additional 'timer' argument which is used to
904 jitter the initial membership report for the solicited-node
905 multicast membership on-link.
906 * This is not strictly needed for MLDv2, which already jitters
907 its report transmissions. However, the 'timer' argument is
908 preserved in case MLDv1 is active on the interface.
910 * The KAME linked-list based IPv6 membership implementation has
911 been refactored to use a vector similar to that used by the IPv4
913 Code which maintains a list of its own multicast memberships
914 internally, e.g. carp, has been updated to reflect the new
917 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
918 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
919 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
920 implementation constraint which needs to be addressed in HEAD.
922 For application developers:
924 * The changes are broadly similar to those made for the IPv4
927 * The use of IPv4 and IPv6 multicast socket options on the same
928 socket, using mapped addresses, HAS NOT been tested or supported.
930 * There are a number of issues with the implementation of various
931 IPv6 multicast APIs which need to be resolved in the API surface
932 before the implementation is fully compatible with KAME userland
933 use, and these are mostly to do with interface index treatment.
935 * The literature available discusses the use of either the delta / ASM
936 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
937 using setsourcefilter(3)/getsourcefilter(3). For more information
938 please refer to RFC 3768, 'Socket Interface Extensions for
939 Multicast Source Filters'.
941 * Applications which use the published RFC 3678 APIs should be fine.
943 For systems administrators:
945 * The mtest(8) utility has been refactored to support IPv6, in
946 addition to IPv4. Interface addresses are no longer accepted
947 as arguments, their names must be used instead. The utility
948 will map the interface name to its first IPv4 address as
949 returned by getifaddrs(3).
951 * The ifmcstat(8) utility has also been updated to print the MLDv2
952 endpoint state and source filter lists via sysctl(3).
954 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
955 loopback of IPv6 multicast datagrams by default; it defaults to 1
956 to preserve the existing behaviour. Disabling multicast loopback is
957 recommended for optimal system performance.
959 * The IPv6 MROUTING code has been changed to examine this sysctl
960 instead of attempting to perform a group lookup before looping
961 back forwarded datagrams.
963 Bump __FreeBSD_version to 800084.
966 Implement low-level Bluetooth HCI API.
967 Bump __FreeBSD_version to 800083.
970 The layout of struct malloc_type, used by modules to register new
971 memory allocation types, has changed. Most modules will need to
972 be rebuilt or panics may be experienced.
973 Bump __FreeBSD_version to 800081.
976 Anticipate overflowing inp_flags - add inp_flags2.
977 This changes most offsets in inpcb, so checking v4 connection
978 state will require a world rebuild.
979 Bump __FreeBSD_version to 800080.
982 Add an llentry to struct route and struct route_in6. Modules
983 embedding a struct route will need to be recompiled.
984 Bump __FreeBSD_version to 800079.
987 The size of rt_metrics_lite and by extension rtentry has changed.
988 Networking administration apps will need to be recompiled.
989 The route command now supports show as an alias for get, weighting
990 of routes, sticky and nostick flags to alter the behavior of stateful
992 Bump __FreeBSD_version to 800078.
995 Do not use Giant for kbdmux(4) locking. This is wrong and
996 apparently causing more problems than it solves. This will
997 re-open the issue where interrupt handlers may race with
998 kbdmux(4) in polling mode. Typical symptoms include (but
999 not limited to) duplicated and/or missing characters when
1000 low level console functions (such as gets) are used while
1001 interrupts are enabled (for example geli password prompt,
1002 mountroot prompt etc.). Disabling kbdmux(4) may help.
1005 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1006 kernel modules referencing any of the above need to be recompiled.
1007 Bump __FreeBSD_version to 800075.
1010 GEOM_PART has become the default partition slicer for storage devices,
1011 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1012 introduces some changes:
1014 MSDOS/EBR: the devices created from MSDOS extended partition entries
1015 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1016 to devices with offset-based names. fstabs may need to be modified.
1018 BSD: the "geometry does not match label" warning is harmless in most
1019 cases but it points to problems in file system misalignment with
1020 disk geometry. The "c" partition is now implicit, covers the whole
1021 top-level drive and cannot be (mis)used by users.
1023 General: Kernel dumps are now not allowed to be written to devices
1024 whose partition types indicate they are meant to be used for file
1025 systems (or, in case of MSDOS partitions, as something else than
1028 Most of these changes date approximately from 200812.
1031 The uscanner(4) driver has been removed from the kernel. This follows
1032 Linux removing theirs in 2.6 and making libusb the default interface
1033 (supported by sane).
1036 The multicast forwarding code has been cleaned up. netstat(1)
1037 only relies on KVM now for printing bandwidth upcall meters.
1038 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1039 ip6_mroute_mod respectively. The config(5) options for statically
1040 compiling this code remain the same, i.e. 'options MROUTING'.
1043 Support for the IFF_NEEDSGIANT network interface flag has been
1044 removed, which means that non-MPSAFE network device drivers are no
1045 longer supported. In particular, if_ar, if_sr, and network device
1046 drivers from the old (legacy) USB stack can no longer be built or
1050 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1051 a bunch of new language catalog files have also been added.
1052 This means that some common libc messages are now localized and
1053 they depend on the LC_MESSAGES environmental variable.
1056 The k8temp(4) driver has been renamed to amdtemp(4) since
1057 support for Family 10 and Family 11 CPU families was added.
1060 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1061 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1063 For kernel developers, the most important changes are that the
1064 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1065 and this lock has been downgraded to a non-recursive mutex.
1067 Transport protocols (UDP, Raw IP) are now responsible for filtering
1068 inbound multicast traffic according to group membership and source
1069 filters. The imo_multicast_filter() KPI exists for this purpose.
1070 Transports which do not use multicast (SCTP, TCP) already reject
1071 multicast by default. Forwarding and receive performance may improve
1072 as a mutex acquisition is no longer needed in the ip_input()
1073 low-level input path. in_addmulti() and in_delmulti() are shimmed
1074 to new KPIs which exist to support SSM in-kernel.
1076 For application developers, it is recommended that loopback of
1077 multicast datagrams be disabled for best performance, as this
1078 will still cause the lock to be taken for each looped-back
1079 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1080 be tuned to 0 to disable loopback by default; it defaults to 1
1081 to preserve the existing behaviour.
1083 For systems administrators, to obtain best performance with
1084 multicast reception and multiple groups, it is always recommended
1085 that a card with a suitably precise hash filter is used. Hash
1086 collisions will still result in the lock being taken within the
1087 transport protocol input path to check group membership.
1089 If deploying FreeBSD in an environment with IGMP snooping switches,
1090 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1091 enabled; this forces 224.0.0.0/24 group membership to be announced
1094 The size of 'struct igmpstat' has changed; netstat needs to be
1095 recompiled to reflect this.
1096 Bump __FreeBSD_version to 800070.
1099 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1100 updated to use it. This requires a buildworld/installworld in order to
1101 update the library and dependencies (usbconfig, etc). Its advisable to
1102 rebuild all ports which uses libusb. More specific directions are given
1103 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1104 libusb are no longer required and can be removed.
1107 A workaround is committed to allow the creation of System V shared
1108 memory segment of size > 2 GB on the 64-bit architectures.
1109 Due to a limitation of the existing ABI, the shm_segsz member
1110 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1111 wrong for large segments. Note that limits must be explicitly
1112 raised to allow such segments to be created.
1115 The layout of struct ifnet has changed, requiring a rebuild of all
1116 network device driver modules.
1119 The /dev handling for the new USB stack has changed, a
1120 buildworld/installworld is required for libusb20.
1123 The new USB2 stack has now been permanently moved in and all kernel and
1124 module names reverted to their previous values (eg, usb, ehci, ohci,
1125 ums, ...). The old usb stack can be compiled in by prefixing the name
1126 with the letter 'o', the old usb modules have been removed.
1127 Updating entry 20090216 for xorg and 20090215 for libmap may still
1131 The rc.conf(5) option if_up_delay has been renamed to
1132 defaultroute_delay to better reflect its purpose. If you have
1133 customized this setting in /etc/rc.conf you need to update it to
1137 xorg 7.4 wants to configure its input devices via hald which does not
1138 yet work with USB2. If the keyboard/mouse does not work in xorg then
1140 Option "AllowEmptyInput" "off"
1141 to your ServerLayout section. This will cause X to use the configured
1142 kbd and mouse sections from your xorg.conf.
1145 The GENERIC kernels for all architectures now default to the new USB2
1146 stack. No kernel config options or code have been removed so if a
1147 problem arises please report it and optionally revert to the old USB
1148 stack. If you are loading USB kernel modules or have a custom kernel
1149 that includes GENERIC then ensure that usb names are also changed over,
1150 eg uftdi -> usb2_serial_ftdi.
1152 Older programs linked against the ports libusb 0.1 need to be
1153 redirected to the new stack's libusb20. /etc/libmap.conf can
1155 # Map old usb library to new one for usb2 stack
1156 libusb-0.1.so.8 libusb20.so.1
1159 All USB ethernet devices now attach as interfaces under the name ueN
1160 (eg. ue0). This is to provide a predictable name as vendors often
1161 change usb chipsets in a product without notice.
1164 The ichsmb(4) driver has been changed to require SMBus slave
1165 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1166 All of the other SMBus controller drivers require left-justified
1167 slave addresses, so this change makes all the drivers provide the
1171 INET6 statistics (struct ip6stat) was updated.
1172 netstat(1) needs to be recompiled.
1175 NTFS has been removed from GENERIC kernel on amd64 to match
1176 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1177 will load ntfs.ko module automatically when NTFS support is
1178 actually needed, unless ntfs.ko is not installed or security
1179 level prohibits loading kernel modules. If either is the case,
1180 "options NTFS" has to be added into kernel config.
1183 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1184 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1185 800061. User space tools that rely on the size of struct tcpcb in
1186 tcp_var.h (e.g. sockstat) need to be recompiled.
1189 ng_tty(4) module updated to match the new TTY subsystem.
1190 Due to API change, user-level applications must be updated.
1191 New API support added to mpd5 CVS and expected to be present
1192 in next mpd5.3 release.
1195 With __FreeBSD_version 800060 the makefs tool is part of
1196 the base system (it was a port).
1199 The afdata and ifnet locks have been changed from mutexes to
1200 rwlocks, network modules will need to be re-compiled.
1203 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1204 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1205 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1206 architecture and 40 bytes on 64-bit architecture. The userland
1207 applications "arp" and "ndp" have been updated accordingly.
1208 The output from "netstat -r" shows only routing entries and
1209 none of the L2 information.
1212 __FreeBSD_version 800057 marks the switchover from the
1213 binary ath hal to source code. Users must add the line:
1215 options AH_SUPPORT_AR5416
1217 to their kernel config files when specifying:
1221 The ath_hal module no longer exists; the code is now compiled
1222 together with the driver in the ath module. It is now
1223 possible to tailor chip support (i.e. reduce the set of chips
1224 and thereby the code size); consult ath_hal(4) for details.
1227 __FreeBSD_version 800054 adds memory barriers to
1228 <machine/atomic.h>, new interfaces to ifnet to facilitate
1229 multiple hardware transmit queues for cards that support
1230 them, and a lock-less ring-buffer implementation to
1231 enable drivers to more efficiently manage queueing of
1235 A new version of ZFS (version 13) has been merged to -HEAD.
1236 This version has zpool attribute "listsnapshots" off by
1237 default, which means "zfs list" does not show snapshots,
1238 and is the same as Solaris behavior.
1241 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1244 The uhci, ohci, ehci and slhci USB Host controller drivers have
1245 been put into separate modules. If you load the usb module
1246 separately through loader.conf you will need to load the
1247 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1248 controller add the following to loader.conf:
1254 The ABI used by the PMC toolset has changed. Please keep
1255 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1259 atapci kernel module now includes only generic PCI ATA
1260 driver. AHCI driver moved to ataahci kernel module.
1261 All vendor-specific code moved into separate kernel modules:
1262 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1263 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1264 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1265 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1268 The TTY subsystem of the kernel has been replaced by a new
1269 implementation, which provides better scalability and an
1270 improved driver model. Most common drivers have been migrated to
1271 the new TTY subsystem, while others have not. The following
1272 drivers have not yet been ported to the new TTY layer:
1275 cy, digi, rc, rp, sio
1281 ng_h4, ng_tty, ppp, sl, snp
1283 Adding these drivers to your kernel configuration file shall
1284 cause compilation to fail.
1287 ntpd has been upgraded to 4.2.4p5.
1290 OpenSSH has been upgraded to 5.1p1.
1292 For many years, FreeBSD's version of OpenSSH preferred DSA
1293 over RSA for host and user authentication keys. With this
1294 upgrade, we've switched to the vendor's default of RSA over
1295 DSA. This may cause upgraded clients to warn about unknown
1296 host keys even for previously known hosts. Users should
1297 follow the usual procedure for verifying host keys before
1298 accepting the RSA key.
1300 This can be circumvented by setting the "HostKeyAlgorithms"
1301 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1304 Please note that the sequence of keys offered for
1305 authentication has been changed as well. You may want to
1306 specify IdentityFile in a different order to revert this
1310 The sio(4) driver has been removed from the i386 and amd64
1311 kernel configuration files. This means uart(4) is now the
1312 default serial port driver on those platforms as well.
1314 To prevent collisions with the sio(4) driver, the uart(4) driver
1315 uses different names for its device nodes. This means the
1316 onboard serial port will now most likely be called "ttyu0"
1317 instead of "ttyd0". You may need to reconfigure applications to
1318 use the new device names.
1320 When using the serial port as a boot console, be sure to update
1321 /boot/device.hints and /etc/ttys before booting the new kernel.
1322 If you forget to do so, you can still manually specify the hints
1323 at the loader prompt:
1325 set hint.uart.0.at="isa"
1326 set hint.uart.0.port="0x3F8"
1327 set hint.uart.0.flags="0x10"
1328 set hint.uart.0.irq="4"
1332 The gpt(8) utility has been removed. Use gpart(8) to partition
1336 The version that Linuxulator emulates was changed from 2.4.2
1337 to 2.6.16. If you experience any problems with Linux binaries
1338 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1339 if it fixes the problem contact emulation mailing list.
1342 ISDN4BSD (I4B) was removed from the src tree. You may need to
1343 update a your kernel configuration and remove relevant entries.
1346 I have checked in code to support multiple routing tables.
1347 See the man pages setfib(1) and setfib(2).
1348 This is a hopefully backwards compatible version,
1349 but to make use of it you need to compile your kernel
1350 with options ROUTETABLES=2 (or more up to 16).
1353 The 802.11 wireless support was redone to enable multi-bss
1354 operation on devices that are capable. The underlying device
1355 is no longer used directly but instead wlanX devices are
1356 cloned with ifconfig. This requires changes to rc.conf files.
1357 For example, change:
1358 ifconfig_ath0="WPA DHCP"
1361 ifconfig_wlan0="WPA DHCP"
1362 see rc.conf(5) for more details. In addition, mergemaster of
1363 /etc/rc.d is highly recommended. Simultaneous update of userland
1364 and kernel wouldn't hurt either.
1366 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1367 modules were merged into the base wlan module. All references
1368 to these modules (e.g. in kernel config files) must be removed.
1371 psm(4) has gained write(2) support in native operation level.
1372 Arbitrary commands can be written to /dev/psm%d and status can
1373 be read back from it. Therefore, an application is responsible
1374 for status validation and error recovery. It is a no-op in
1375 other operation levels.
1378 Support for KSE threading has been removed from the kernel. To
1379 run legacy applications linked against KSE libmap.conf may
1380 be used. The following libmap.conf may be used to ensure
1381 compatibility with any prior release:
1383 libpthread.so.1 libthr.so.1
1384 libpthread.so.2 libthr.so.2
1385 libkse.so.3 libthr.so.3
1388 The layout of struct vmspace has changed. This affects libkvm
1389 and any executables that link against libkvm and use the
1390 kvm_getprocs() function. In particular, but not exclusively,
1391 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1392 The effects are minimal, but it's advisable to upgrade world
1396 The latest em driver no longer has support in it for the
1397 82575 adapter, this is now moved to the igb driver. The
1398 split was done to make new features that are incompatible
1399 with older hardware easier to do.
1402 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1403 likewise the kernel option is now GEOM_LINUX_LVM.
1406 The default NFS mount mode has changed from UDP to TCP for
1407 increased reliability. If you rely on (insecurely) NFS
1408 mounting across a firewall you may need to update your
1412 Belatedly note the addition of m_collapse for compacting
1416 The fts(3) structures have been changed to use adequate
1417 integer types for their members and so to be able to cope
1418 with huge file trees. The old fts(3) ABI is preserved
1419 through symbol versioning in libc, so third-party binaries
1420 using fts(3) should still work, although they will not take
1421 advantage of the extended types. At the same time, some
1422 third-party software might fail to build after this change
1423 due to unportable assumptions made in its source code about
1424 fts(3) structure members. Such software should be fixed
1425 by its vendor or, in the worst case, in the ports tree.
1426 FreeBSD_version 800015 marks this change for the unlikely
1427 case that a portable fix is impossible.
1430 To upgrade to -current after this date, you must be running
1431 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1432 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1435 The ADAPTIVE_GIANT kernel option has been retired because its
1436 functionality is the default now.
1439 The AT keyboard emulation of sunkbd(4) has been turned on
1440 by default. In order to make the special symbols of the Sun
1441 keyboards driven by sunkbd(4) work under X these now have
1442 to be configured the same way as Sun USB keyboards driven
1443 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1445 Option "XkbLayout" "us"
1446 Option "XkbRules" "xorg"
1447 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1450 It has been decided that it is desirable to provide ABI
1451 backwards compatibility to the FreeBSD 4/5/6 versions of the
1452 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1453 broken with the introduction of PCI domain support (see the
1454 20070930 entry). Unfortunately, this required the ABI of
1455 PCIOCGETCONF to be broken again in order to be able to
1456 provide backwards compatibility to the old version of that
1457 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1458 again. As for prominent ports this affects neither pciutils
1459 nor xorg-server this time, the hal port needs to be rebuilt
1463 The misnamed kthread_create() and friends have been renamed
1464 to kproc_create() etc. Many of the callers already
1465 used kproc_start()..
1466 I will return kthread_create() and friends in a while
1467 with implementations that actually create threads, not procs.
1468 Renaming corresponds with version 800002.
1477 Avoid using make -j when upgrading. While generally safe, there are
1478 sometimes problems using -j to upgrade. If your upgrade fails with
1479 -j, please try again without -j. From time to time in the past there
1480 have been problems using -j with buildworld and/or installworld. This
1481 is especially true when upgrading between "distant" versions (eg one
1482 that cross a major release boundary or several minor releases, or when
1483 several months have passed on the -current branch).
1485 Sometimes, obscure build problems are the result of environment
1486 poisoning. This can happen because the make utility reads its
1487 environment when searching for values for global variables. To run
1488 your build attempts in an "environmental clean room", prefix all make
1489 commands with 'env -i '. See the env(1) manual page for more details.
1491 When upgrading from one major version to another it is generally best
1492 to upgrade to the latest code in the currently installed branch first,
1493 then do an upgrade to the new branch. This is the best-tested upgrade
1494 path, and has the highest probability of being successful. Please try
1495 this approach before reporting problems with a major version upgrade.
1499 When upgrading the boot ZFS pool to a new version, always follow
1502 1.) recompile and reinstall the ZFS boot loader and boot block
1503 (this is part of "make buildworld" and "make installworld")
1505 2.) update the ZFS boot block on your boot drive
1507 The following example updates the ZFS boot block on the first
1508 partition (freebsd-boot) of a GPT partitioned drive ad0:
1509 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1511 Non-boot pools do not need these updates.
1515 If you are updating from a prior version of FreeBSD (even one just
1516 a few days old), you should follow this procedure. It is the most
1517 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1519 make kernel-toolchain
1520 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1521 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1523 To test a kernel once
1524 ---------------------
1525 If you just want to boot a kernel once (because you are not sure
1526 if it works, or if you want to boot a known bad kernel to provide
1527 debugging information) run
1528 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1529 nextboot -k testkernel
1531 To just build a kernel when you know that it won't mess you up
1532 --------------------------------------------------------------
1533 This assumes you are already running a CURRENT system. Replace
1534 ${arch} with the architecture of your machine (e.g. "i386",
1535 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1537 cd src/sys/${arch}/conf
1538 config KERNEL_NAME_HERE
1539 cd ../compile/KERNEL_NAME_HERE
1544 If this fails, go to the "To build a kernel" section.
1546 To rebuild everything and install it on the current system.
1547 -----------------------------------------------------------
1548 # Note: sometimes if you are running current you gotta do more than
1549 # is listed here if you are upgrading from a really old current.
1551 <make sure you have good level 0 dumps>
1553 make kernel KERNCONF=YOUR_KERNEL_HERE
1555 <reboot in single user> [3]
1563 To cross-install current onto a separate partition
1564 --------------------------------------------------
1565 # In this approach we use a separate partition to hold
1566 # current's root, 'usr', and 'var' directories. A partition
1567 # holding "/", "/usr" and "/var" should be about 2GB in
1570 <make sure you have good level 0 dumps>
1573 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1574 <maybe newfs current's root partition>
1575 <mount current's root partition on directory ${CURRENT_ROOT}>
1576 make installworld DESTDIR=${CURRENT_ROOT}
1577 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1578 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1579 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1580 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1581 <reboot into current>
1582 <do a "native" rebuild/install as described in the previous section>
1583 <maybe install compatibility libraries from ports/misc/compat*>
1587 To upgrade in-place from stable to current
1588 ----------------------------------------------
1589 <make sure you have good level 0 dumps>
1591 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1593 <reboot in single user> [3]
1600 Make sure that you've read the UPDATING file to understand the
1601 tweaks to various things you need. At this point in the life
1602 cycle of current, things change often and you are on your own
1603 to cope. The defaults can also change, so please read ALL of
1604 the UPDATING entries.
1606 Also, if you are tracking -current, you must be subscribed to
1607 freebsd-current@freebsd.org. Make sure that before you update
1608 your sources that you have read and understood all the recent
1609 messages there. If in doubt, please track -stable which has
1610 much fewer pitfalls.
1612 [1] If you have third party modules, such as vmware, you
1613 should disable them at this point so they don't crash your
1616 [3] From the bootblocks, boot -s, and then do
1621 adjkerntz -i # if CMOS is wall time
1622 Also, when doing a major release upgrade, it is required that
1623 you boot into single user mode to do the installworld.
1625 [4] Note: This step is non-optional. Failure to do this step
1626 can result in a significant reduction in the functionality of the
1627 system. Attempting to do it by hand is not recommended and those
1628 that pursue this avenue should read this file carefully, as well
1629 as the archives of freebsd-current and freebsd-hackers mailing lists
1630 for potential gotchas. The -U option is also useful to consider.
1631 See mergemaster(8) for more information.
1633 [5] Usually this step is a noop. However, from time to time
1634 you may need to do this if you get unknown user in the following
1635 step. It never hurts to do it all the time. You may need to
1636 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1637 install) after the buildworld before this step if you last updated
1638 from [78]-stable or 9-stable before 20130430.
1640 [6] This only deletes old files and directories. Old libraries
1641 can be deleted by "make delete-old-libs", but you have to make
1642 sure that no program is using those libraries anymore.
1644 [8] In order to have a kernel that can run the 4.x binaries needed to
1645 do an installworld, you must include the COMPAT_FREEBSD4 option in
1646 your kernel. Failure to do so may leave you with a system that is
1647 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1648 required to run the 5.x binaries on more recent kernels. And so on
1649 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1651 Make sure that you merge any new devices from GENERIC since the
1652 last time you updated your kernel config file.
1654 [9] When checking out sources, you must include the -P flag to have
1655 cvs prune empty directories.
1657 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1658 "?=" instead of the "=" assignment operator, so that buildworld can
1659 override the CPUTYPE if it needs to.
1661 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1662 not on the command line, or in /etc/make.conf. buildworld will
1663 warn if it is improperly defined.
1666 This file contains a list, in reverse chronological order, of major
1667 breakages in tracking -current. It is not guaranteed to be a complete
1668 list of such breakages, and only contains entries since October 10, 2007.
1669 If you need to see UPDATING entries from before that date, you will need
1670 to fetch an UPDATING file from an older FreeBSD release.
1672 Copyright information:
1674 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1676 Redistribution, publication, translation and use, with or without
1677 modification, in full or in part, in any form or format of this
1678 document are permitted without further permission from the author.
1680 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1681 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1682 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1683 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1684 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1685 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1686 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1687 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1688 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1689 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1690 POSSIBILITY OF SUCH DAMAGE.
1692 Contact Warner Losh if you have any questions about your use of