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 An additional fix for the issue described in the 20150615 sendmail
16 entry below has been been committed in revision 284788.
19 The latest version of openssl rejects TLS handshakes with DH
20 parameters below 768 bits. sendmail releases prior to 8.15.2
21 (not yet released), defaulted to a 512 bit DH parameter setting
22 for client connections. To improve interoperability, the
23 sendmail default for client connections has been raised to
27 It is no longer necessary to include the dwarf version in your DEBUG
28 options in your kernel config file. The bug that required it to be
29 placed in the config file has bene fixed. DEBUG should now just
30 contain -g. The build system will automatically update things
31 to do the right thing.
37 On i386 and amd64 systems, the onifconsole flag is now set by default
38 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
39 as a login TTY if it is set in the bootloader as an active kernel
40 console. No changes in behavior should result otherwise. To revert to
41 the previous behavior, set ttyu0 to "off" in /etc/ttys.
44 Clang and llvm have been upgraded to 3.4.1 release.
47 Clang and llvm have been upgraded to 3.4 release.
50 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
51 been deprecated and will not be part of FreeBSD 11.0 and later
52 releases. If you use this driver, please consider switching to
53 the nfe(4) driver instead.
56 The behavior of gss_pseudo_random() for the krb5 mechanism
57 has changed, for applications requesting a longer random string
58 than produced by the underlying enctype's pseudo-random() function.
59 In particular, the random string produced from a session key of
60 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
61 be different at the 17th octet and later, after this change.
62 The counter used in the PRF+ construction is now encoded as a
63 big-endian integer in accordance with RFC 4402.
64 __FreeBSD_version is bumped to 902505.
70 Behavior of devfs rules path matching has been changed.
71 Pattern is now always matched against fully qualified devfs
72 path and slash characters must be explicitly matched by
73 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
74 subdirectories must be reviewed.
77 hastctl(8)'s `status' command output changed to terse one-liner format.
78 Scripts using this should switch to `list' command or be rewritten.
81 Fix a bug that allowed a tracing process (e.g. gdb) to write
82 to a memory-mapped file in the traced process's address space
83 even if neither the traced process nor the tracing process had
84 write access to that file.
87 Added ZFS TRIM support which is enabled by default. To disable
88 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
90 Creating new ZFS pools and adding new devices to existing pools
91 first performs a full device level TRIM which can take a significant
92 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
93 to disable this behaviour.
95 ZFS TRIM requires the underlying device support BIO_DELETE which
96 is currently provided by methods such as ATA TRIM and SCSI UNMAP
97 via CAM, which are typically supported by SSD's.
99 Stats for ZFS TRIM can be monitored by looking at the sysctl's
100 under kstat.zfs.misc.zio_trim.
103 `list' command has been added to hastctl(8). For now, it is full
104 equivalent of `status' command.
105 WARNING: in the near future the output of hastctl's status command
106 will change to more terse format. If you use `hastctl status'
107 for parsing in your scripts, switch to `hastctl list'.
110 The mergemaster command now uses the default MAKEOBJDIRPREFIX
111 rather than creating it's own in the temporary directory in
112 order allow access to bootstrapped versions of tools such as
113 install and mtree. When upgrading from version of FreeBSD where
114 the install command does not support -l, you will need to
115 install a new mergemaster command if mergemaster -p is required.
116 This can be accomplished with the command (cd src/usr.sbin/mergemaster
119 Due to the use of the new -l option to install(1) during build
120 and install, you must take care not to directly set the INSTALL
121 make variable in your /etc/make.conf, /etc/src.conf, or on the
122 command line. If you wish to use the -C flag for all installs
123 you may be able to add INSTALL+=-C to /etc/make.conf or
127 Fix a bug that allows NFS clients to issue READDIR on files.
130 The install(1) option -M has changed meaning and now takes an
131 argument that is a file or path to append logs to. In the
132 unlikely event that -M was the last option on the command line
133 and the command line contained at least two files and a target
134 directory the first file will have logs appended to it. The -M
135 option served little practical purpose in the last decade so it's
136 used expected to be extremely rare.
139 A new compression method (lz4) has been merged to. Please refer to
140 zpool-features(7) for more information.
142 Please refer to the "ZFS notes" section of this file for information
143 on upgrading boot ZFS pools.
146 The VFS KBI was changed with the merge of several nullfs
147 optimizations and fixes. All filesystem modules must be
151 With the addition of auditdistd(8), a new auditdistd user is now
152 depended on during installworld. "mergemaster -p" can be used to add
153 the user prior to installworld, as documented in the handbook.
159 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
160 Starting with this version the old system of ZFS pool versioning
161 is superseded by "feature flags". This concept enables forward
162 compatibility against certain future changes in functionality of ZFS
163 pools. The first two read-only compatible "feature flags" for ZFS
164 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
165 For more information read the new zpool-features(7) manual page.
166 Please refer to the "ZFS notes" section of this file for information
167 on upgrading boot ZFS pools.
170 The commit introducing bsd.compiler.mk breaks the traditional
171 building of kernels before this point. Add -m ${SRC}/share/mk
172 (for the right value of SRC) to your command lines to work
173 around; update your useland to a point after this; or use the
174 buildkernel/installkernel top-level targets. See also 20120829.
177 The IPFIREWALL_FORWARD kernel option has been removed. Its
178 functionality now turned on by default.
181 The random(4) support for the VIA hardware random number
182 generator (`PADLOCK') is no longer enabled unconditionally.
183 Add the PADLOCK_RNG option in the custom kernel config if
184 needed. The GENERIC kernels on i386 and amd64 do include the
185 option, so the change only affects the custom kernel
189 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
190 the traditional method, you must update your system with an installworld
191 before the kernel will build. The documented make buildkernel/installkernel
192 interfaces (coupled with fresh make kernel-toolchain) continue to work.
195 The sparc64 ZFS loader has been changed to no longer try to auto-
196 detect ZFS providers based on diskN aliases but now requires these
197 to be explicitly listed in the OFW boot-device environment variable.
200 Now unix domain sockets behave "as expected" on nullfs(5). Previously
201 nullfs(5) did not pass through all behaviours to the underlying layer,
202 as a result if we bound to a socket on the lower layer we could connect
203 only to the lower path; if we bound to the upper layer we could connect
204 only to the upper path. The new behavior is one can connect to both the
205 lower and the upper paths regardless what layer path one binds to.
208 The acpi_wmi(4) status device /dev/wmistat has been renamed to
212 A new VOP_ADVISE() was added to support posix_fadvise(2). All
213 filesystem modules must be recompiled.
216 The interface of the VOP_VPTOCNP(9) changed, now the returned
217 vnode shall be referenced, previously it was required to be
218 only held. All in-tree filesystems are converted.
224 The broken amd(4) driver has been replaced with esp(4) in the amd64,
225 i386 and pc98 GENERIC kernel configuration files.
228 This commit modifies vfs_register() so that it uses a hash
229 calculation to set vfc_typenum, which is enabled by default.
230 The first time a system is booted after this change, the
231 vfc_typenum values will change for all file systems. The
232 main effect of this is a change to the NFS server file handles
233 for file systems that use vfc_typenum in their fsid, such as ZFS.
234 It will, however, prevent vfc_typenum from changing when file
235 systems are loaded in a different order for subsequent reboots.
236 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
237 until you are ready to remount all NFS clients after a reboot.
240 Bump the shared library version numbers for libraries that
241 do not use symbol versioning, have changed the ABI compared
242 to stable/8 and which shared library version was not bumped.
243 Done as part of 9.0-RELEASE cycle.
246 During the merge of Capsicum features, the fget(9) KPI was modified.
247 This may require the rebuilding of out-of-tree device drivers --
248 issues have been reported specifically with the nVidia device driver.
249 __FreeBSD_version is bumped to 900041.
251 Also, there is a period between 20110811 and 20110814 where the
252 special devices /dev/{stdin,stdout,stderr} did not work correctly.
253 Building world from a kernel during that window may not work.
256 The packet filter (pf) code has been updated to OpenBSD 4.5.
257 You need to update userland tools to be in sync with kernel.
258 This update breaks backward compatibility with earlier pfsync(4)
259 versions. Care must be taken when updating redundant firewall setups.
262 The following sysctls and tunables are retired on x86 platforms:
264 machdep.hlt_logical_cpus
265 The following sysctl is retired:
266 machdep.hyperthreading_allowed
267 The sysctls were supposed to provide a way to dynamically offline and
268 online selected CPUs on x86 platforms, but the implementation has not
269 been reliable especially with SCHED_ULE scheduler.
270 machdep.hyperthreading_allowed tunable is still available to ignore
271 hyperthreading CPUs at OS level.
272 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
273 where X is an APIC ID of a CPU. Be advised, though, that disabling
274 CPUs in non-uniform fashion will result in non-uniform topology and
275 may lead to sub-optimal system performance with SCHED_ULE, which is
279 cpumask_t type is retired and cpuset_t is used in order to describe
283 Changes to ifconfig(8) for dynamic address family detection mandate
284 that you are running a kernel of 20110525 or later. Make sure to
285 follow the update procedure to boot a new kernel before installing
289 Support for sun4v architecture is officially dropped
292 Several KPI breaking changes have been committed to the mii(4) layer,
293 the PHY drivers and consequently some Ethernet drivers using mii(4).
294 This means that miibus.ko and the modules of the affected Ethernet
295 drivers need to be recompiled.
297 Note to kernel developers: Given that the OUI bit reversion problem
298 was fixed as part of these changes all mii(4) commits related to OUIs,
299 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
300 handling, no longer can be merged verbatim to stable/8 and previous
304 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
305 into their kernel configurations along with 'device pci'.
308 The default NFS client is now the new NFS client, so fstype "newnfs"
309 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
310 Although mounts via fstype "nfs" will usually work without userland
311 changes, it is recommended that the mount(8) and mount_nfs(8)
312 commands be rebuilt from sources and that a link to mount_nfs called
313 mount_oldnfs be created. The new client is compiled into the
314 kernel with "options NFSCL" and this is needed for diskless root
315 file systems. The GENERIC kernel configs have been changed to use
316 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
317 To use the regular/old client, you can "mount -t oldnfs ...". For
318 a diskless root file system, you must also include a line like:
320 vfs.root.mountfrom="oldnfs:"
322 in the boot/loader.conf on the root fs on the NFS server to make
323 a diskless root fs use the old client.
326 The GENERIC kernels for all architectures now default to the new
327 CAM-based ATA stack. It means that all legacy ATA drivers were
328 removed and replaced by respective CAM drivers. If you are using
329 ATA device names in /etc/fstab or other places, make sure to update
330 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
331 where 'Y's are the sequential numbers starting from zero for each type
332 in order of detection, unless configured otherwise with tunables,
333 see cam(4)). There will be symbolic links created in /dev/ to map
334 old adX devices to the respective adaY. They should provide basic
335 compatibility for file systems mounting in most cases, but they do
336 not support old user-level APIs and do not have respective providers
337 in GEOM. Consider using updated management tools with new device names.
339 It is possible to load devices ahci, ata, siis and mvs as modules,
340 but option ATA_CAM should remain in kernel configuration to make ata
341 module work as CAM driver supporting legacy ATA controllers. Device ata
342 still can be used in modular fashion (atacore + ...). Modules atadisk
343 and atapi* are not used and won't affect operation in ATA_CAM mode.
344 Note that to use CAM-based ATA kernel should include CAM devices
345 scbus, pass, da (or explicitly ada), cd and optionally others. All of
346 them are parts of the cam module.
348 ataraid(4) functionality is now supported by the RAID GEOM class.
349 To use it you can load geom_raid kernel module and use graid(8) tool
350 for management. Instead of /dev/arX device names, use /dev/raid/rX.
352 No kernel config options or code have been removed, so if a problem
353 arises, please report it and optionally revert to the old ATA stack.
354 In order to do it you can remove from the kernel config:
359 , and instead add back:
360 device atadisk # ATA disk drives
361 device ataraid # ATA RAID drives
362 device atapicd # ATAPI CDROM drives
363 device atapifd # ATAPI floppy drives
364 device atapist # ATAPI tape drives
367 The default NFS server has been changed to the new server, which
368 was referred to as the experimental server. If you need to switch
369 back to the old NFS server, you must now put the "-o" option on
370 both the mountd and nfsd commands. This can be done using the
371 mountd_flags and nfs_server_flags rc.conf variables until an
372 update to the rc scripts is committed, which is coming soon.
375 The GNU Objective-C runtime library (libobjc), and other Objective-C
376 related components have been removed from the base system. If you
377 require an Objective-C library, please use one of the available ports.
380 ath(4) has been split into bus- and device- modules. if_ath contains
381 the HAL, the TX rate control and the network device code. if_ath_pci
382 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
383 contains the AHB glue. Users need to load both if_ath_pci and if_ath
384 in order to use ath on everything else.
386 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
387 need to load if_ath and if_ath_pci for ath(4) operation.
390 As part of the replacement of sysinstall, the process of building
391 release media has changed significantly. For details, please re-read
392 release(7), which has been updated to reflect the new build process.
395 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
396 is the last available version under GPLv2. It brings a number of new
397 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
398 4.1 and SSE 4.2), better support for powerpc64, a number of new
399 directives, and lots of other small improvements. See the ChangeLog
400 file in contrib/binutils for the full details.
403 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
404 compliant, and will now use half of hash for authentication.
405 This will break interoperability with all stacks (including all
406 actual FreeBSD versions) who implement
407 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
409 The only workaround with such peers is to use another HMAC
410 algorithm for IPsec ("phase 2") authentication.
413 Remove the uio_yield prototype and symbol. This function has
414 been misnamed since it was introduced and should not be
415 globally exposed with this name. The equivalent functionality
416 is now available using kern_yield(curthread->td_user_pri).
417 The function remains undocumented.
420 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
421 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
422 for scalar sysctls is defined but disabled. Code that needs
423 UQUAD to pass the type checking that must compile on older
424 systems where the define is not present can check against
425 __FreeBSD_version >= 900030.
427 The system dialog(1) has been replaced with a new version previously
428 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
429 with the previous version, but the libdialog associated with it has
430 a largely incompatible API. As such, the original version of libdialog
431 will be kept temporarily as libodialog, until its base system consumers
432 are replaced or updated. Bump __FreeBSD_version to 900030.
435 If you are trying to run make universe on a -stable system, and you get
436 the following warning:
437 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
438 unknown. config(8) likely too old."
439 or something similar to it, then you must upgrade your -stable system
440 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
441 stable/8) or build the config from the latest stable/8 branch and
442 install it on your system.
444 Prior to this date, building a current universe on 8-stable system from
445 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
446 error in the first kernel build phase. A new config on those old
447 systems will fix that problem for older versions of -current.
450 The TCP stack has been modified to allow Khelp modules to interact with
451 it via helper hook points and store per-connection data in the TCP
452 control block. Bump __FreeBSD_version to 900029. User space tools that
453 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
457 Generic IEEE 802.3 annex 31B full duplex flow control support has been
458 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
459 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
460 to take advantage of it instead of using custom implementations. This
461 means that these drivers now no longer unconditionally advertise
462 support for flow control but only do so if flow control is a selected
463 media option. This was implemented in the generic support that way in
464 order to allow flow control to be switched on and off via ifconfig(8)
465 with the PHY specific default to typically off in order to protect
466 from unwanted effects. Consequently, if you used flow control with
467 one of the above mentioned drivers you now need to explicitly enable
469 ifconfig bge0 media auto mediaopt flowcontrol
471 Along with the above mentioned changes generic support for setting
472 1000baseT master mode also has been added and brgphy(4), ciphy(4),
473 e1000phy(4) as well as ip1000phy(4) have been converted to take
474 advantage of it. This means that these drivers now no longer take the
475 link0 parameter for selecting master mode but the master media option
476 has to be used instead, for example like in the following:
477 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
479 Selection of master mode now is also available with all other PHY
480 drivers supporting 1000baseT.
483 The TCP stack has received a significant update to add support for
484 modularised congestion control and generally improve the clarity of
485 congestion control decisions. Bump __FreeBSD_version to 900025. User
486 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
487 sockstat) need to be recompiled.
490 The man(1) utility has been replaced by a new version that no longer
491 uses /etc/manpath.config. Please consult man.conf(5) for how to
492 migrate local entries to the new format.
495 The copyright strings printed by login(1) and sshd(8) at the time of a
496 new connection have been removed to follow other operating systems and
500 A workaround for a fixed ld bug has been removed in kernel code,
501 so make sure that your system ld is built from sources after
502 revision 210245 from 2010-07-19 (r211583 if building head kernel
503 on stable/8, r211584 for stable/7; both from 2010-08-21).
504 A symptom of incorrect ld version is different addresses for
505 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
508 The $ipv6_prefer variable in rc.conf(5) has been split into
509 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
511 The $ip6addrctl_policy is a variable to choose a pre-defined
512 address selection policy set by ip6addrctl(8). A value
513 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
516 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
517 flag (see an entry of 20090926) is set on an interface with no
518 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
519 security reason. If you want IPv6 link-local address on all
520 interfaces by default, set this to "YES".
522 The old ipv6_prefer="YES" is equivalent to
523 ipv6_activate_all_interfaces="YES" and
524 ip6addrctl_policy="ipv6_prefer".
527 DTrace has grown support for userland tracing. Due to this, DTrace is
528 now i386 and amd64 only.
529 dtruss(1) is now installed by default on those systems and a new
530 kernel module is needed for userland tracing: fasttrap.
531 No changes to your kernel config file are necessary to enable
532 userland tracing, but you might consider adding 'STRIP=' and
533 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
534 to have informative userland stack traces in DTrace (ustack).
537 The acpi_aiboost(4) driver has been removed in favor of the new
538 aibs(4) driver. You should update your kernel configuration file.
541 BSD grep has been imported to the base system and it is built by
542 default. It is completely BSD licensed, highly GNU-compatible, uses
543 less memory than its GNU counterpart and has a small codebase.
544 However, it is slower than its GNU counterpart, which is mostly
545 noticeable for larger searches, for smaller ones it is measurable
546 but not significant. The reason is complex, the most important factor
547 is that we lack a modern and efficient regex library and GNU
548 overcomes this by optimizing the searches internally. Future work
549 on improving the regex performance is planned, for the meantime,
550 users that need better performance, can build GNU grep instead by
551 setting the WITH_GNU_GREP knob.
554 Due to the import of powerpc64 support, all existing powerpc kernel
555 configuration files must be updated with a machine directive like this:
556 machine powerpc powerpc
558 In addition, an updated config(8) is required to build powerpc kernels
562 A new version of ZFS (version 15) has been merged to -HEAD.
563 This version uses a python library for the following subcommands:
564 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
565 For full functionality of these commands the following port must
566 be installed: sysutils/py-zfs
569 'vm_page's are now hashed by physical address to an array of mutexes.
570 Currently this is only used to serialize access to hold_count. Over
571 time the page queue mutex will be peeled away. This changes the size
572 of pmap on every architecture. And requires all callers of vm_page_hold
573 and vm_page_unhold to be updated.
576 WITH_CTF can now be specified in src.conf (not recommended, there
577 are some problems with static executables), make.conf (would also
578 affect ports which do not use GNU make and do not override the
579 compile targets) or in the kernel config (via "makeoptions
581 When WITH_CTF was specified there before this was silently ignored,
582 so make sure that WITH_CTF is not used in places which could lead
583 to unwanted behavior.
586 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
587 to allow 32-bit compatibility on non-x86 platforms. All kernel
588 configurations on amd64 and ia64 platforms using these options must
589 be modified accordingly.
592 The utmp user accounting database has been replaced with utmpx,
593 the user accounting interface standardized by POSIX.
594 Unfortunately the semantics of utmp and utmpx don't match,
595 making it practically impossible to support both interfaces.
596 The user accounting database is used by tools like finger(1),
597 last(1), talk(1), w(1) and ac(8).
599 All applications in the base system use utmpx. This means only
600 local binaries (e.g. from the ports tree) may still use these
601 utmp database files. These applications must be rebuilt to make
604 After the system has been upgraded, it is safe to remove the old
605 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
606 assuming their contents is of no importance anymore. Old wtmp
607 databases can only be used by last(1) and ac(8) after they have
608 been converted to the new format using wtmpcvt(1).
611 Introduce the kernel thread "deadlock resolver" (which can be enabled
612 via the DEADLKRES option, see NOTES for more details) and the
613 sleepq_type() function for sleepqueues.
616 The rc.firewall and rc.firewall6 were unified, and
617 rc.firewall6 and rc.d/ip6fw were removed.
618 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
619 variables are obsoleted. Instead, the following new rc
620 variables are added to rc.d/ipfw:
622 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
623 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
624 firewall_simple_onet_ipv6, firewall_trusted_ipv6
626 The meanings correspond to the relevant IPv4 variables.
632 The default terminal emulation for syscons(4) has been changed
633 from cons25 to xterm on all platforms except pc98. This means
634 that the /etc/ttys file needs to be updated to ensure correct
635 operation of applications on the console.
637 The terminal emulation style can be toggled per window by using
638 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
639 options can be used to change the compile-time default back to
642 To prevent graphical artifacts, make sure the TERM environment
643 variable is set to match the terminal emulation that is being
644 performed by syscons(4).
647 The layout of the structure ieee80211req_scan_result has changed.
648 Applications that require wireless scan results (e.g. ifconfig(8))
649 from net80211 need to be recompiled.
651 Applications such as wpa_supplicant(8) may require a full world
652 build without using NO_CLEAN in order to get synchronized with the
656 The iwn(4) driver has been updated to support the 5000 and 5150 series.
657 There's one kernel module for each firmware. Adding "device iwnfw"
658 to the kernel configuration file means including all three firmware
659 images inside the kernel. If you want to include just the one for
660 your wireless card, use the devices iwn4965fw, iwn5000fw or
664 The rc.d/network_ipv6, IPv6 configuration script has been integrated
665 into rc.d/netif. The changes are the following:
667 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
668 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
669 Note that both variables need the "inet6" keyword at the head.
671 Do not set $ipv6_network_interfaces manually if you do not
672 understand what you are doing. It is not needed in most cases.
674 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
677 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
678 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
680 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
681 all configured interfaces have "inet6 accept_rtadv" in the
682 $ifconfig_IF_ipv6. These are for backward compatibility.
684 3. A new variable $ipv6_prefer has been added. If NO, IPv6
685 functionality of interfaces with no corresponding
686 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
687 and the default address selection policy of ip6addrctl(8)
688 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
689 Note that if you want to configure IPv6 functionality on the
690 disabled interfaces after boot, first you need to clear the flag by
691 using ifconfig(8) like:
693 ifconfig em0 inet6 -ifdisabled
695 If YES, the default address selection policy is set as
698 The default value of $ipv6_prefer is NO.
700 4. If your system need to receive Router Advertisement messages,
701 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
702 scripts automatically invoke rtsol(8) when the interface becomes
703 UP. The Router Advertisement messages are used for SLAAC
704 (State-Less Address AutoConfiguration).
707 802.11s D3.03 support was committed. This is incompatible with the
708 previous code, which was based on D3.0.
711 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
712 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
713 control whether accepting Router Advertisement messages or not.
714 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
715 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
716 The ifconfig(8) utility now supports these flags.
719 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
720 mount(8) and -a option for df(1) to see them.
723 The old tunable hw.bus.devctl_disable has been superseded by
724 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
725 replaced by hw.bus.devctl_queue=0. The default for this new tunable
729 Remove the option STOP_NMI. The default action is now to use NMI only
730 for KDB via the newly introduced function stop_cpus_hard() and
731 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
734 The stable/8 branch created in subversion. This corresponds to the
735 RELENG_8 branch in CVS.
738 Bump the shared library version numbers for all libraries that do not
739 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
740 __FreeBSD_version to 800105.
743 Due to changes in the implementation of virtual network stack support,
744 all network-related kernel modules must be recompiled. As this change
745 breaks the ABI, bump __FreeBSD_version to 800104.
748 The TOE interface to the TCP syncache has been modified to remove
749 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
750 The cxgb driver is the only TOE consumer affected by this change, and
751 needs to be recompiled along with the kernel. As this change breaks
752 the ABI, bump __FreeBSD_version to 800103.
755 Padding has been added to struct tcpcb, sackhint and tcpstat in
756 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
757 maintaining the ABI. However, this change breaks the ABI, so bump
758 __FreeBSD_version to 800102. User space tools that rely on the size of
759 any of these structs (e.g. sockstat) need to be recompiled.
762 The NFS_LEGACYRPC option has been removed along with the old kernel
763 RPC implementation that this option selected. Kernel configurations
764 may need to be adjusted.
767 The network interface device nodes at /dev/net/<interface> have been
768 removed. All ioctl operations can be performed the normal way using
769 routing sockets. The kqueue functionality can generally be replaced
770 with routing sockets.
773 The documentation from the FreeBSD Documentation Project (Handbook,
774 FAQ, etc.) is now installed via packages by sysinstall(8) and under
775 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
778 The ABI of various structures related to the SYSV IPC API have been
779 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
780 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
784 Layout of struct vnet has changed as routing related variables were
785 moved to their own Vimage module. Modules need to be recompiled. Bump
786 __FreeBSD_version to 800099.
789 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
790 respectively. As long as no more than 16 groups per process are used,
791 no changes should be visible. When more than 16 groups are used, old
792 binaries may fail if they call getgroups() or getgrouplist() with
793 statically sized storage. Recompiling will work around this, but
794 applications should be modified to use dynamically allocated storage
795 for group arrays as POSIX.1-2008 does not cap an implementation's
796 number of supported groups at NGROUPS_MAX+1 as previous versions did.
798 NFS and portalfs mounts may also be affected as the list of groups is
799 truncated to 16. Users of NFS who use more than 16 groups, should
800 take care that negative group permissions are not used on the exported
801 file systems as they will not be reliable unless a GSSAPI based
802 authentication method is used.
805 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
806 option compiles in the support for adaptive spinning for lockmgrs
807 which want to enable it. The lockinit() function now accepts the flag
808 LK_ADAPTIVE in order to make the lock object subject to adaptive
809 spinning when both held in write and read mode.
812 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
813 changed. User applications that use this ioctl need to be rebuilt.
816 The layout of struct thread has changed. Kernel and modules need to
820 The layout of structs ifnet, domain, protosw and vnet_net has changed.
821 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
824 window(1) has been removed from the base system. It can now be
825 installed from ports. The port is called misc/window.
828 The way we are storing and accessing `routing table' entries has
829 changed. Programs reading the FIB, like netstat, need to be
833 A new netisr implementation has been added for FreeBSD 8. Network
834 file system modules, such as igmp, ipdivert, and others, should be
836 Bump __FreeBSD_version to 800096.
839 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
843 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
844 Bump __FreeBSD_version to 800094.
847 Add mnt_xflag field to 'struct mount'. File system modules need to be
849 Bump __FreeBSD_version to 800093.
852 The compiling option ADAPTIVE_SX has been retired while it has been
853 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
854 The KPI for sx_init_flags() changes as accepting flags:
855 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
856 been introduced in order to handle the reversed logic.
857 Bump __FreeBSD_version to 800092.
860 Add support for hierarchical jails. Remove global securelevel.
861 Bump __FreeBSD_version to 800091.
864 The layout of struct vnet_net has changed, therefore modules
866 Bump __FreeBSD_version to 800090.
869 The newly imported zic(8) produces a new format in the output. Please
870 run tzsetup(8) to install the newly created data to /etc/localtime.
873 The sysctl tree for the usb stack has renamed from hw.usb2.* to
874 hw.usb.* and is now consistent again with previous releases.
877 802.11 monitor mode support was revised and driver api's were changed.
878 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
879 of DLT_IEEE802_11. No user-visible data structures were changed but
880 applications that use DLT_IEEE802_11 may require changes.
881 Bump __FreeBSD_version to 800088.
884 The layout of the following structs has changed: sysctl_oid,
885 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
886 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
887 panics may be experienced. World rebuild is required for
888 correctly checking networking state from userland.
889 Bump __FreeBSD_version to 800085.
892 MLDv2 and Source-Specific Multicast (SSM) have been merged
893 to the IPv6 stack. VIMAGE hooks are in but not yet used.
894 The implementation of SSM within FreeBSD's IPv6 stack closely
895 follows the IPv4 implementation.
897 For kernel developers:
899 * The most important changes are that the ip6_output() and
900 ip6_input() paths no longer take the IN6_MULTI_LOCK,
901 and this lock has been downgraded to a non-recursive mutex.
903 * As with the changes to the IPv4 stack to support SSM, filtering
904 of inbound multicast traffic must now be performed by transport
905 protocols within the IPv6 stack. This does not apply to TCP and
906 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
908 * The KPIs used by IPv6 multicast are similar to those used by
909 the IPv4 stack, with the following differences:
910 * im6o_mc_filter() is analogous to imo_multicast_filter().
911 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
912 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
913 * IN6_LOOKUP_MULTI() has been deprecated and removed.
914 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
915 for MLDv1 have an additional 'timer' argument which is used to
916 jitter the initial membership report for the solicited-node
917 multicast membership on-link.
918 * This is not strictly needed for MLDv2, which already jitters
919 its report transmissions. However, the 'timer' argument is
920 preserved in case MLDv1 is active on the interface.
922 * The KAME linked-list based IPv6 membership implementation has
923 been refactored to use a vector similar to that used by the IPv4
925 Code which maintains a list of its own multicast memberships
926 internally, e.g. carp, has been updated to reflect the new
929 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
930 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
931 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
932 implementation constraint which needs to be addressed in HEAD.
934 For application developers:
936 * The changes are broadly similar to those made for the IPv4
939 * The use of IPv4 and IPv6 multicast socket options on the same
940 socket, using mapped addresses, HAS NOT been tested or supported.
942 * There are a number of issues with the implementation of various
943 IPv6 multicast APIs which need to be resolved in the API surface
944 before the implementation is fully compatible with KAME userland
945 use, and these are mostly to do with interface index treatment.
947 * The literature available discusses the use of either the delta / ASM
948 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
949 using setsourcefilter(3)/getsourcefilter(3). For more information
950 please refer to RFC 3768, 'Socket Interface Extensions for
951 Multicast Source Filters'.
953 * Applications which use the published RFC 3678 APIs should be fine.
955 For systems administrators:
957 * The mtest(8) utility has been refactored to support IPv6, in
958 addition to IPv4. Interface addresses are no longer accepted
959 as arguments, their names must be used instead. The utility
960 will map the interface name to its first IPv4 address as
961 returned by getifaddrs(3).
963 * The ifmcstat(8) utility has also been updated to print the MLDv2
964 endpoint state and source filter lists via sysctl(3).
966 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
967 loopback of IPv6 multicast datagrams by default; it defaults to 1
968 to preserve the existing behaviour. Disabling multicast loopback is
969 recommended for optimal system performance.
971 * The IPv6 MROUTING code has been changed to examine this sysctl
972 instead of attempting to perform a group lookup before looping
973 back forwarded datagrams.
975 Bump __FreeBSD_version to 800084.
978 Implement low-level Bluetooth HCI API.
979 Bump __FreeBSD_version to 800083.
982 The layout of struct malloc_type, used by modules to register new
983 memory allocation types, has changed. Most modules will need to
984 be rebuilt or panics may be experienced.
985 Bump __FreeBSD_version to 800081.
988 Anticipate overflowing inp_flags - add inp_flags2.
989 This changes most offsets in inpcb, so checking v4 connection
990 state will require a world rebuild.
991 Bump __FreeBSD_version to 800080.
994 Add an llentry to struct route and struct route_in6. Modules
995 embedding a struct route will need to be recompiled.
996 Bump __FreeBSD_version to 800079.
999 The size of rt_metrics_lite and by extension rtentry has changed.
1000 Networking administration apps will need to be recompiled.
1001 The route command now supports show as an alias for get, weighting
1002 of routes, sticky and nostick flags to alter the behavior of stateful
1004 Bump __FreeBSD_version to 800078.
1007 Do not use Giant for kbdmux(4) locking. This is wrong and
1008 apparently causing more problems than it solves. This will
1009 re-open the issue where interrupt handlers may race with
1010 kbdmux(4) in polling mode. Typical symptoms include (but
1011 not limited to) duplicated and/or missing characters when
1012 low level console functions (such as gets) are used while
1013 interrupts are enabled (for example geli password prompt,
1014 mountroot prompt etc.). Disabling kbdmux(4) may help.
1017 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1018 kernel modules referencing any of the above need to be recompiled.
1019 Bump __FreeBSD_version to 800075.
1022 GEOM_PART has become the default partition slicer for storage devices,
1023 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1024 introduces some changes:
1026 MSDOS/EBR: the devices created from MSDOS extended partition entries
1027 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1028 to devices with offset-based names. fstabs may need to be modified.
1030 BSD: the "geometry does not match label" warning is harmless in most
1031 cases but it points to problems in file system misalignment with
1032 disk geometry. The "c" partition is now implicit, covers the whole
1033 top-level drive and cannot be (mis)used by users.
1035 General: Kernel dumps are now not allowed to be written to devices
1036 whose partition types indicate they are meant to be used for file
1037 systems (or, in case of MSDOS partitions, as something else than
1040 Most of these changes date approximately from 200812.
1043 The uscanner(4) driver has been removed from the kernel. This follows
1044 Linux removing theirs in 2.6 and making libusb the default interface
1045 (supported by sane).
1048 The multicast forwarding code has been cleaned up. netstat(1)
1049 only relies on KVM now for printing bandwidth upcall meters.
1050 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1051 ip6_mroute_mod respectively. The config(5) options for statically
1052 compiling this code remain the same, i.e. 'options MROUTING'.
1055 Support for the IFF_NEEDSGIANT network interface flag has been
1056 removed, which means that non-MPSAFE network device drivers are no
1057 longer supported. In particular, if_ar, if_sr, and network device
1058 drivers from the old (legacy) USB stack can no longer be built or
1062 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1063 a bunch of new language catalog files have also been added.
1064 This means that some common libc messages are now localized and
1065 they depend on the LC_MESSAGES environmental variable.
1068 The k8temp(4) driver has been renamed to amdtemp(4) since
1069 support for Family 10 and Family 11 CPU families was added.
1072 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1073 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1075 For kernel developers, the most important changes are that the
1076 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1077 and this lock has been downgraded to a non-recursive mutex.
1079 Transport protocols (UDP, Raw IP) are now responsible for filtering
1080 inbound multicast traffic according to group membership and source
1081 filters. The imo_multicast_filter() KPI exists for this purpose.
1082 Transports which do not use multicast (SCTP, TCP) already reject
1083 multicast by default. Forwarding and receive performance may improve
1084 as a mutex acquisition is no longer needed in the ip_input()
1085 low-level input path. in_addmulti() and in_delmulti() are shimmed
1086 to new KPIs which exist to support SSM in-kernel.
1088 For application developers, it is recommended that loopback of
1089 multicast datagrams be disabled for best performance, as this
1090 will still cause the lock to be taken for each looped-back
1091 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1092 be tuned to 0 to disable loopback by default; it defaults to 1
1093 to preserve the existing behaviour.
1095 For systems administrators, to obtain best performance with
1096 multicast reception and multiple groups, it is always recommended
1097 that a card with a suitably precise hash filter is used. Hash
1098 collisions will still result in the lock being taken within the
1099 transport protocol input path to check group membership.
1101 If deploying FreeBSD in an environment with IGMP snooping switches,
1102 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1103 enabled; this forces 224.0.0.0/24 group membership to be announced
1106 The size of 'struct igmpstat' has changed; netstat needs to be
1107 recompiled to reflect this.
1108 Bump __FreeBSD_version to 800070.
1111 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1112 updated to use it. This requires a buildworld/installworld in order to
1113 update the library and dependencies (usbconfig, etc). Its advisable to
1114 rebuild all ports which uses libusb. More specific directions are given
1115 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1116 libusb are no longer required and can be removed.
1119 A workaround is committed to allow the creation of System V shared
1120 memory segment of size > 2 GB on the 64-bit architectures.
1121 Due to a limitation of the existing ABI, the shm_segsz member
1122 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1123 wrong for large segments. Note that limits must be explicitly
1124 raised to allow such segments to be created.
1127 The layout of struct ifnet has changed, requiring a rebuild of all
1128 network device driver modules.
1131 The /dev handling for the new USB stack has changed, a
1132 buildworld/installworld is required for libusb20.
1135 The new USB2 stack has now been permanently moved in and all kernel and
1136 module names reverted to their previous values (eg, usb, ehci, ohci,
1137 ums, ...). The old usb stack can be compiled in by prefixing the name
1138 with the letter 'o', the old usb modules have been removed.
1139 Updating entry 20090216 for xorg and 20090215 for libmap may still
1143 The rc.conf(5) option if_up_delay has been renamed to
1144 defaultroute_delay to better reflect its purpose. If you have
1145 customized this setting in /etc/rc.conf you need to update it to
1149 xorg 7.4 wants to configure its input devices via hald which does not
1150 yet work with USB2. If the keyboard/mouse does not work in xorg then
1152 Option "AllowEmptyInput" "off"
1153 to your ServerLayout section. This will cause X to use the configured
1154 kbd and mouse sections from your xorg.conf.
1157 The GENERIC kernels for all architectures now default to the new USB2
1158 stack. No kernel config options or code have been removed so if a
1159 problem arises please report it and optionally revert to the old USB
1160 stack. If you are loading USB kernel modules or have a custom kernel
1161 that includes GENERIC then ensure that usb names are also changed over,
1162 eg uftdi -> usb2_serial_ftdi.
1164 Older programs linked against the ports libusb 0.1 need to be
1165 redirected to the new stack's libusb20. /etc/libmap.conf can
1167 # Map old usb library to new one for usb2 stack
1168 libusb-0.1.so.8 libusb20.so.1
1171 All USB ethernet devices now attach as interfaces under the name ueN
1172 (eg. ue0). This is to provide a predictable name as vendors often
1173 change usb chipsets in a product without notice.
1176 The ichsmb(4) driver has been changed to require SMBus slave
1177 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1178 All of the other SMBus controller drivers require left-justified
1179 slave addresses, so this change makes all the drivers provide the
1183 INET6 statistics (struct ip6stat) was updated.
1184 netstat(1) needs to be recompiled.
1187 NTFS has been removed from GENERIC kernel on amd64 to match
1188 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1189 will load ntfs.ko module automatically when NTFS support is
1190 actually needed, unless ntfs.ko is not installed or security
1191 level prohibits loading kernel modules. If either is the case,
1192 "options NTFS" has to be added into kernel config.
1195 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1196 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1197 800061. User space tools that rely on the size of struct tcpcb in
1198 tcp_var.h (e.g. sockstat) need to be recompiled.
1201 ng_tty(4) module updated to match the new TTY subsystem.
1202 Due to API change, user-level applications must be updated.
1203 New API support added to mpd5 CVS and expected to be present
1204 in next mpd5.3 release.
1207 With __FreeBSD_version 800060 the makefs tool is part of
1208 the base system (it was a port).
1211 The afdata and ifnet locks have been changed from mutexes to
1212 rwlocks, network modules will need to be re-compiled.
1215 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1216 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1217 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1218 architecture and 40 bytes on 64-bit architecture. The userland
1219 applications "arp" and "ndp" have been updated accordingly.
1220 The output from "netstat -r" shows only routing entries and
1221 none of the L2 information.
1224 __FreeBSD_version 800057 marks the switchover from the
1225 binary ath hal to source code. Users must add the line:
1227 options AH_SUPPORT_AR5416
1229 to their kernel config files when specifying:
1233 The ath_hal module no longer exists; the code is now compiled
1234 together with the driver in the ath module. It is now
1235 possible to tailor chip support (i.e. reduce the set of chips
1236 and thereby the code size); consult ath_hal(4) for details.
1239 __FreeBSD_version 800054 adds memory barriers to
1240 <machine/atomic.h>, new interfaces to ifnet to facilitate
1241 multiple hardware transmit queues for cards that support
1242 them, and a lock-less ring-buffer implementation to
1243 enable drivers to more efficiently manage queueing of
1247 A new version of ZFS (version 13) has been merged to -HEAD.
1248 This version has zpool attribute "listsnapshots" off by
1249 default, which means "zfs list" does not show snapshots,
1250 and is the same as Solaris behavior.
1253 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1256 The uhci, ohci, ehci and slhci USB Host controller drivers have
1257 been put into separate modules. If you load the usb module
1258 separately through loader.conf you will need to load the
1259 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1260 controller add the following to loader.conf:
1266 The ABI used by the PMC toolset has changed. Please keep
1267 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1271 atapci kernel module now includes only generic PCI ATA
1272 driver. AHCI driver moved to ataahci kernel module.
1273 All vendor-specific code moved into separate kernel modules:
1274 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1275 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1276 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1277 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1280 The TTY subsystem of the kernel has been replaced by a new
1281 implementation, which provides better scalability and an
1282 improved driver model. Most common drivers have been migrated to
1283 the new TTY subsystem, while others have not. The following
1284 drivers have not yet been ported to the new TTY layer:
1287 cy, digi, rc, rp, sio
1293 ng_h4, ng_tty, ppp, sl, snp
1295 Adding these drivers to your kernel configuration file shall
1296 cause compilation to fail.
1299 ntpd has been upgraded to 4.2.4p5.
1302 OpenSSH has been upgraded to 5.1p1.
1304 For many years, FreeBSD's version of OpenSSH preferred DSA
1305 over RSA for host and user authentication keys. With this
1306 upgrade, we've switched to the vendor's default of RSA over
1307 DSA. This may cause upgraded clients to warn about unknown
1308 host keys even for previously known hosts. Users should
1309 follow the usual procedure for verifying host keys before
1310 accepting the RSA key.
1312 This can be circumvented by setting the "HostKeyAlgorithms"
1313 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1316 Please note that the sequence of keys offered for
1317 authentication has been changed as well. You may want to
1318 specify IdentityFile in a different order to revert this
1322 The sio(4) driver has been removed from the i386 and amd64
1323 kernel configuration files. This means uart(4) is now the
1324 default serial port driver on those platforms as well.
1326 To prevent collisions with the sio(4) driver, the uart(4) driver
1327 uses different names for its device nodes. This means the
1328 onboard serial port will now most likely be called "ttyu0"
1329 instead of "ttyd0". You may need to reconfigure applications to
1330 use the new device names.
1332 When using the serial port as a boot console, be sure to update
1333 /boot/device.hints and /etc/ttys before booting the new kernel.
1334 If you forget to do so, you can still manually specify the hints
1335 at the loader prompt:
1337 set hint.uart.0.at="isa"
1338 set hint.uart.0.port="0x3F8"
1339 set hint.uart.0.flags="0x10"
1340 set hint.uart.0.irq="4"
1344 The gpt(8) utility has been removed. Use gpart(8) to partition
1348 The version that Linuxulator emulates was changed from 2.4.2
1349 to 2.6.16. If you experience any problems with Linux binaries
1350 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1351 if it fixes the problem contact emulation mailing list.
1354 ISDN4BSD (I4B) was removed from the src tree. You may need to
1355 update a your kernel configuration and remove relevant entries.
1358 I have checked in code to support multiple routing tables.
1359 See the man pages setfib(1) and setfib(2).
1360 This is a hopefully backwards compatible version,
1361 but to make use of it you need to compile your kernel
1362 with options ROUTETABLES=2 (or more up to 16).
1365 The 802.11 wireless support was redone to enable multi-bss
1366 operation on devices that are capable. The underlying device
1367 is no longer used directly but instead wlanX devices are
1368 cloned with ifconfig. This requires changes to rc.conf files.
1369 For example, change:
1370 ifconfig_ath0="WPA DHCP"
1373 ifconfig_wlan0="WPA DHCP"
1374 see rc.conf(5) for more details. In addition, mergemaster of
1375 /etc/rc.d is highly recommended. Simultaneous update of userland
1376 and kernel wouldn't hurt either.
1378 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1379 modules were merged into the base wlan module. All references
1380 to these modules (e.g. in kernel config files) must be removed.
1383 psm(4) has gained write(2) support in native operation level.
1384 Arbitrary commands can be written to /dev/psm%d and status can
1385 be read back from it. Therefore, an application is responsible
1386 for status validation and error recovery. It is a no-op in
1387 other operation levels.
1390 Support for KSE threading has been removed from the kernel. To
1391 run legacy applications linked against KSE libmap.conf may
1392 be used. The following libmap.conf may be used to ensure
1393 compatibility with any prior release:
1395 libpthread.so.1 libthr.so.1
1396 libpthread.so.2 libthr.so.2
1397 libkse.so.3 libthr.so.3
1400 The layout of struct vmspace has changed. This affects libkvm
1401 and any executables that link against libkvm and use the
1402 kvm_getprocs() function. In particular, but not exclusively,
1403 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1404 The effects are minimal, but it's advisable to upgrade world
1408 The latest em driver no longer has support in it for the
1409 82575 adapter, this is now moved to the igb driver. The
1410 split was done to make new features that are incompatible
1411 with older hardware easier to do.
1414 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1415 likewise the kernel option is now GEOM_LINUX_LVM.
1418 The default NFS mount mode has changed from UDP to TCP for
1419 increased reliability. If you rely on (insecurely) NFS
1420 mounting across a firewall you may need to update your
1424 Belatedly note the addition of m_collapse for compacting
1428 The fts(3) structures have been changed to use adequate
1429 integer types for their members and so to be able to cope
1430 with huge file trees. The old fts(3) ABI is preserved
1431 through symbol versioning in libc, so third-party binaries
1432 using fts(3) should still work, although they will not take
1433 advantage of the extended types. At the same time, some
1434 third-party software might fail to build after this change
1435 due to unportable assumptions made in its source code about
1436 fts(3) structure members. Such software should be fixed
1437 by its vendor or, in the worst case, in the ports tree.
1438 FreeBSD_version 800015 marks this change for the unlikely
1439 case that a portable fix is impossible.
1442 To upgrade to -current after this date, you must be running
1443 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1444 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1447 The ADAPTIVE_GIANT kernel option has been retired because its
1448 functionality is the default now.
1451 The AT keyboard emulation of sunkbd(4) has been turned on
1452 by default. In order to make the special symbols of the Sun
1453 keyboards driven by sunkbd(4) work under X these now have
1454 to be configured the same way as Sun USB keyboards driven
1455 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1457 Option "XkbLayout" "us"
1458 Option "XkbRules" "xorg"
1459 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1462 It has been decided that it is desirable to provide ABI
1463 backwards compatibility to the FreeBSD 4/5/6 versions of the
1464 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1465 broken with the introduction of PCI domain support (see the
1466 20070930 entry). Unfortunately, this required the ABI of
1467 PCIOCGETCONF to be broken again in order to be able to
1468 provide backwards compatibility to the old version of that
1469 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1470 again. As for prominent ports this affects neither pciutils
1471 nor xorg-server this time, the hal port needs to be rebuilt
1475 The misnamed kthread_create() and friends have been renamed
1476 to kproc_create() etc. Many of the callers already
1477 used kproc_start()..
1478 I will return kthread_create() and friends in a while
1479 with implementations that actually create threads, not procs.
1480 Renaming corresponds with version 800002.
1489 Avoid using make -j when upgrading. While generally safe, there are
1490 sometimes problems using -j to upgrade. If your upgrade fails with
1491 -j, please try again without -j. From time to time in the past there
1492 have been problems using -j with buildworld and/or installworld. This
1493 is especially true when upgrading between "distant" versions (eg one
1494 that cross a major release boundary or several minor releases, or when
1495 several months have passed on the -current branch).
1497 Sometimes, obscure build problems are the result of environment
1498 poisoning. This can happen because the make utility reads its
1499 environment when searching for values for global variables. To run
1500 your build attempts in an "environmental clean room", prefix all make
1501 commands with 'env -i '. See the env(1) manual page for more details.
1503 When upgrading from one major version to another it is generally best
1504 to upgrade to the latest code in the currently installed branch first,
1505 then do an upgrade to the new branch. This is the best-tested upgrade
1506 path, and has the highest probability of being successful. Please try
1507 this approach before reporting problems with a major version upgrade.
1511 When upgrading the boot ZFS pool to a new version, always follow
1514 1.) recompile and reinstall the ZFS boot loader and boot block
1515 (this is part of "make buildworld" and "make installworld")
1517 2.) update the ZFS boot block on your boot drive
1519 The following example updates the ZFS boot block on the first
1520 partition (freebsd-boot) of a GPT partitioned drive ad0:
1521 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1523 Non-boot pools do not need these updates.
1527 If you are updating from a prior version of FreeBSD (even one just
1528 a few days old), you should follow this procedure. It is the most
1529 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1531 make kernel-toolchain
1532 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1533 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1535 To test a kernel once
1536 ---------------------
1537 If you just want to boot a kernel once (because you are not sure
1538 if it works, or if you want to boot a known bad kernel to provide
1539 debugging information) run
1540 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1541 nextboot -k testkernel
1543 To just build a kernel when you know that it won't mess you up
1544 --------------------------------------------------------------
1545 This assumes you are already running a CURRENT system. Replace
1546 ${arch} with the architecture of your machine (e.g. "i386",
1547 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1549 cd src/sys/${arch}/conf
1550 config KERNEL_NAME_HERE
1551 cd ../compile/KERNEL_NAME_HERE
1556 If this fails, go to the "To build a kernel" section.
1558 To rebuild everything and install it on the current system.
1559 -----------------------------------------------------------
1560 # Note: sometimes if you are running current you gotta do more than
1561 # is listed here if you are upgrading from a really old current.
1563 <make sure you have good level 0 dumps>
1565 make kernel KERNCONF=YOUR_KERNEL_HERE
1567 <reboot in single user> [3]
1575 To cross-install current onto a separate partition
1576 --------------------------------------------------
1577 # In this approach we use a separate partition to hold
1578 # current's root, 'usr', and 'var' directories. A partition
1579 # holding "/", "/usr" and "/var" should be about 2GB in
1582 <make sure you have good level 0 dumps>
1585 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1586 <maybe newfs current's root partition>
1587 <mount current's root partition on directory ${CURRENT_ROOT}>
1588 make installworld DESTDIR=${CURRENT_ROOT}
1589 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1590 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1591 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1592 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1593 <reboot into current>
1594 <do a "native" rebuild/install as described in the previous section>
1595 <maybe install compatibility libraries from ports/misc/compat*>
1599 To upgrade in-place from stable to current
1600 ----------------------------------------------
1601 <make sure you have good level 0 dumps>
1603 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1605 <reboot in single user> [3]
1612 Make sure that you've read the UPDATING file to understand the
1613 tweaks to various things you need. At this point in the life
1614 cycle of current, things change often and you are on your own
1615 to cope. The defaults can also change, so please read ALL of
1616 the UPDATING entries.
1618 Also, if you are tracking -current, you must be subscribed to
1619 freebsd-current@freebsd.org. Make sure that before you update
1620 your sources that you have read and understood all the recent
1621 messages there. If in doubt, please track -stable which has
1622 much fewer pitfalls.
1624 [1] If you have third party modules, such as vmware, you
1625 should disable them at this point so they don't crash your
1628 [3] From the bootblocks, boot -s, and then do
1633 adjkerntz -i # if CMOS is wall time
1634 Also, when doing a major release upgrade, it is required that
1635 you boot into single user mode to do the installworld.
1637 [4] Note: This step is non-optional. Failure to do this step
1638 can result in a significant reduction in the functionality of the
1639 system. Attempting to do it by hand is not recommended and those
1640 that pursue this avenue should read this file carefully, as well
1641 as the archives of freebsd-current and freebsd-hackers mailing lists
1642 for potential gotchas. The -U option is also useful to consider.
1643 See mergemaster(8) for more information.
1645 [5] Usually this step is a noop. However, from time to time
1646 you may need to do this if you get unknown user in the following
1647 step. It never hurts to do it all the time. You may need to
1648 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1649 install) after the buildworld before this step if you last updated
1650 from [78]-stable or 9-stable before 20130430.
1652 [6] This only deletes old files and directories. Old libraries
1653 can be deleted by "make delete-old-libs", but you have to make
1654 sure that no program is using those libraries anymore.
1656 [8] In order to have a kernel that can run the 4.x binaries needed to
1657 do an installworld, you must include the COMPAT_FREEBSD4 option in
1658 your kernel. Failure to do so may leave you with a system that is
1659 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1660 required to run the 5.x binaries on more recent kernels. And so on
1661 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1663 Make sure that you merge any new devices from GENERIC since the
1664 last time you updated your kernel config file.
1666 [9] When checking out sources, you must include the -P flag to have
1667 cvs prune empty directories.
1669 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1670 "?=" instead of the "=" assignment operator, so that buildworld can
1671 override the CPUTYPE if it needs to.
1673 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1674 not on the command line, or in /etc/make.conf. buildworld will
1675 warn if it is improperly defined.
1678 This file contains a list, in reverse chronological order, of major
1679 breakages in tracking -current. It is not guaranteed to be a complete
1680 list of such breakages, and only contains entries since October 10, 2007.
1681 If you need to see UPDATING entries from before that date, you will need
1682 to fetch an UPDATING file from an older FreeBSD release.
1684 Copyright information:
1686 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1688 Redistribution, publication, translation and use, with or without
1689 modification, in full or in part, in any form or format of this
1690 document are permitted without further permission from the author.
1692 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1693 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1694 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1695 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1696 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1697 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1698 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1699 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1700 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1701 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1702 POSSIBILITY OF SUCH DAMAGE.
1704 Contact Warner Losh if you have any questions about your use of