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 The install(1) option -M has changed meaning and now takes an
16 argument that is a file or path to append logs to. In the
17 unlikely event that -M was the last option on the command line
18 and the command line contained at least two files and a target
19 directory the first file will have logs appended to it. The -M
20 option served little practical purpose in the last decade so it's
21 used expected to be extremely rare.
24 A new compression method (lz4) has been merged to. Please refer to
25 zpool-features(7) for more information.
27 Please refer to the "ZFS notes" section of this file for information
28 on upgrading boot ZFS pools.
31 The VFS KBI was changed with the merge of several nullfs
32 optimizations and fixes. All filesystem modules must be
36 With the addition of auditdistd(8), a new auditdistd user is now
37 depended on during installworld. "mergemaster -p" can be used to add
38 the user prior to installworld, as documented in the handbook.
44 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
45 Starting with this version the old system of ZFS pool versioning
46 is superseded by "feature flags". This concept enables forward
47 compatibility against certain future changes in functionality of ZFS
48 pools. The first two read-only compatible "feature flags" for ZFS
49 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
50 For more information read the new zpool-features(7) manual page.
51 Please refer to the "ZFS notes" section of this file for information
52 on upgrading boot ZFS pools.
55 The commit introducing bsd.compiler.mk breaks the traditional
56 building of kernels before this point. Add -m ${SRC}/share/mk
57 (for the right value of SRC) to your command lines to work
58 around; update your useland to a point after this; or use the
59 buildkernel/installkernel top-level targets. See also 20120829.
62 The IPFIREWALL_FORWARD kernel option has been removed. Its
63 functionality now turned on by default.
66 The random(4) support for the VIA hardware random number
67 generator (`PADLOCK') is no longer enabled unconditionally.
68 Add the PADLOCK_RNG option in the custom kernel config if
69 needed. The GENERIC kernels on i386 and amd64 do include the
70 option, so the change only affects the custom kernel
74 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
75 the traditional method, you must update your system with an installworld
76 before the kernel will build. The documented make buildkernel/installkernel
77 interfaces (coupled with fresh make kernel-toolchain) continue to work.
80 The sparc64 ZFS loader has been changed to no longer try to auto-
81 detect ZFS providers based on diskN aliases but now requires these
82 to be explicitly listed in the OFW boot-device environment variable.
85 Now unix domain sockets behave "as expected" on nullfs(5). Previously
86 nullfs(5) did not pass through all behaviours to the underlying layer,
87 as a result if we bound to a socket on the lower layer we could connect
88 only to the lower path; if we bound to the upper layer we could connect
89 only to the upper path. The new behavior is one can connect to both the
90 lower and the upper paths regardless what layer path one binds to.
93 The acpi_wmi(4) status device /dev/wmistat has been renamed to
97 A new VOP_ADVISE() was added to support posix_fadvise(2). All
98 filesystem modules must be recompiled.
101 The interface of the VOP_VPTOCNP(9) changed, now the returned
102 vnode shall be referenced, previously it was required to be
103 only held. All in-tree filesystems are converted.
109 The broken amd(4) driver has been replaced with esp(4) in the amd64,
110 i386 and pc98 GENERIC kernel configuration files.
113 This commit modifies vfs_register() so that it uses a hash
114 calculation to set vfc_typenum, which is enabled by default.
115 The first time a system is booted after this change, the
116 vfc_typenum values will change for all file systems. The
117 main effect of this is a change to the NFS server file handles
118 for file systems that use vfc_typenum in their fsid, such as ZFS.
119 It will, however, prevent vfc_typenum from changing when file
120 systems are loaded in a different order for subsequent reboots.
121 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
122 until you are ready to remount all NFS clients after a reboot.
125 Bump the shared library version numbers for libraries that
126 do not use symbol versioning, have changed the ABI compared
127 to stable/8 and which shared library version was not bumped.
128 Done as part of 9.0-RELEASE cycle.
131 During the merge of Capsicum features, the fget(9) KPI was modified.
132 This may require the rebuilding of out-of-tree device drivers --
133 issues have been reported specifically with the nVidia device driver.
134 __FreeBSD_version is bumped to 900041.
136 Also, there is a period between 20110811 and 20110814 where the
137 special devices /dev/{stdin,stdout,stderr} did not work correctly.
138 Building world from a kernel during that window may not work.
141 The packet filter (pf) code has been updated to OpenBSD 4.5.
142 You need to update userland tools to be in sync with kernel.
143 This update breaks backward compatibility with earlier pfsync(4)
144 versions. Care must be taken when updating redundant firewall setups.
147 The following sysctls and tunables are retired on x86 platforms:
149 machdep.hlt_logical_cpus
150 The following sysctl is retired:
151 machdep.hyperthreading_allowed
152 The sysctls were supposed to provide a way to dynamically offline and
153 online selected CPUs on x86 platforms, but the implementation has not
154 been reliable especially with SCHED_ULE scheduler.
155 machdep.hyperthreading_allowed tunable is still available to ignore
156 hyperthreading CPUs at OS level.
157 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
158 where X is an APIC ID of a CPU. Be advised, though, that disabling
159 CPUs in non-uniform fashion will result in non-uniform topology and
160 may lead to sub-optimal system performance with SCHED_ULE, which is
164 cpumask_t type is retired and cpuset_t is used in order to describe
168 Changes to ifconfig(8) for dynamic address family detection mandate
169 that you are running a kernel of 20110525 or later. Make sure to
170 follow the update procedure to boot a new kernel before installing
174 Support for sun4v architecture is officially dropped
177 Several KPI breaking changes have been committed to the mii(4) layer,
178 the PHY drivers and consequently some Ethernet drivers using mii(4).
179 This means that miibus.ko and the modules of the affected Ethernet
180 drivers need to be recompiled.
182 Note to kernel developers: Given that the OUI bit reversion problem
183 was fixed as part of these changes all mii(4) commits related to OUIs,
184 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
185 handling, no longer can be merged verbatim to stable/8 and previous
189 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
190 into their kernel configurations along with 'device pci'.
193 The default NFS client is now the new NFS client, so fstype "newnfs"
194 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
195 Although mounts via fstype "nfs" will usually work without userland
196 changes, it is recommended that the mount(8) and mount_nfs(8)
197 commands be rebuilt from sources and that a link to mount_nfs called
198 mount_oldnfs be created. The new client is compiled into the
199 kernel with "options NFSCL" and this is needed for diskless root
200 file systems. The GENERIC kernel configs have been changed to use
201 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
202 To use the regular/old client, you can "mount -t oldnfs ...". For
203 a diskless root file system, you must also include a line like:
205 vfs.root.mountfrom="oldnfs:"
207 in the boot/loader.conf on the root fs on the NFS server to make
208 a diskless root fs use the old client.
211 The GENERIC kernels for all architectures now default to the new
212 CAM-based ATA stack. It means that all legacy ATA drivers were
213 removed and replaced by respective CAM drivers. If you are using
214 ATA device names in /etc/fstab or other places, make sure to update
215 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
216 where 'Y's are the sequential numbers starting from zero for each type
217 in order of detection, unless configured otherwise with tunables,
218 see cam(4)). There will be symbolic links created in /dev/ to map
219 old adX devices to the respective adaY. They should provide basic
220 compatibility for file systems mounting in most cases, but they do
221 not support old user-level APIs and do not have respective providers
222 in GEOM. Consider using updated management tools with new device names.
224 It is possible to load devices ahci, ata, siis and mvs as modules,
225 but option ATA_CAM should remain in kernel configuration to make ata
226 module work as CAM driver supporting legacy ATA controllers. Device ata
227 still can be used in modular fashion (atacore + ...). Modules atadisk
228 and atapi* are not used and won't affect operation in ATA_CAM mode.
229 Note that to use CAM-based ATA kernel should include CAM devices
230 scbus, pass, da (or explicitly ada), cd and optionally others. All of
231 them are parts of the cam module.
233 ataraid(4) functionality is now supported by the RAID GEOM class.
234 To use it you can load geom_raid kernel module and use graid(8) tool
235 for management. Instead of /dev/arX device names, use /dev/raid/rX.
237 No kernel config options or code have been removed, so if a problem
238 arises, please report it and optionally revert to the old ATA stack.
239 In order to do it you can remove from the kernel config:
244 , and instead add back:
245 device atadisk # ATA disk drives
246 device ataraid # ATA RAID drives
247 device atapicd # ATAPI CDROM drives
248 device atapifd # ATAPI floppy drives
249 device atapist # ATAPI tape drives
252 The default NFS server has been changed to the new server, which
253 was referred to as the experimental server. If you need to switch
254 back to the old NFS server, you must now put the "-o" option on
255 both the mountd and nfsd commands. This can be done using the
256 mountd_flags and nfs_server_flags rc.conf variables until an
257 update to the rc scripts is committed, which is coming soon.
260 The GNU Objective-C runtime library (libobjc), and other Objective-C
261 related components have been removed from the base system. If you
262 require an Objective-C library, please use one of the available ports.
265 ath(4) has been split into bus- and device- modules. if_ath contains
266 the HAL, the TX rate control and the network device code. if_ath_pci
267 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
268 contains the AHB glue. Users need to load both if_ath_pci and if_ath
269 in order to use ath on everything else.
271 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
272 need to load if_ath and if_ath_pci for ath(4) operation.
275 As part of the replacement of sysinstall, the process of building
276 release media has changed significantly. For details, please re-read
277 release(7), which has been updated to reflect the new build process.
280 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
281 is the last available version under GPLv2. It brings a number of new
282 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
283 4.1 and SSE 4.2), better support for powerpc64, a number of new
284 directives, and lots of other small improvements. See the ChangeLog
285 file in contrib/binutils for the full details.
288 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
289 compliant, and will now use half of hash for authentication.
290 This will break interoperability with all stacks (including all
291 actual FreeBSD versions) who implement
292 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
294 The only workaround with such peers is to use another HMAC
295 algorithm for IPsec ("phase 2") authentication.
298 Remove the uio_yield prototype and symbol. This function has
299 been misnamed since it was introduced and should not be
300 globally exposed with this name. The equivalent functionality
301 is now available using kern_yield(curthread->td_user_pri).
302 The function remains undocumented.
305 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
306 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
307 for scalar sysctls is defined but disabled. Code that needs
308 UQUAD to pass the type checking that must compile on older
309 systems where the define is not present can check against
310 __FreeBSD_version >= 900030.
312 The system dialog(1) has been replaced with a new version previously
313 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
314 with the previous version, but the libdialog associated with it has
315 a largely incompatible API. As such, the original version of libdialog
316 will be kept temporarily as libodialog, until its base system consumers
317 are replaced or updated. Bump __FreeBSD_version to 900030.
320 If you are trying to run make universe on a -stable system, and you get
321 the following warning:
322 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
323 unknown. config(8) likely too old."
324 or something similar to it, then you must upgrade your -stable system
325 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
326 stable/8) or build the config from the latest stable/8 branch and
327 install it on your system.
329 Prior to this date, building a current universe on 8-stable system from
330 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
331 error in the first kernel build phase. A new config on those old
332 systems will fix that problem for older versions of -current.
335 The TCP stack has been modified to allow Khelp modules to interact with
336 it via helper hook points and store per-connection data in the TCP
337 control block. Bump __FreeBSD_version to 900029. User space tools that
338 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
342 Generic IEEE 802.3 annex 31B full duplex flow control support has been
343 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
344 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
345 to take advantage of it instead of using custom implementations. This
346 means that these drivers now no longer unconditionally advertise
347 support for flow control but only do so if flow control is a selected
348 media option. This was implemented in the generic support that way in
349 order to allow flow control to be switched on and off via ifconfig(8)
350 with the PHY specific default to typically off in order to protect
351 from unwanted effects. Consequently, if you used flow control with
352 one of the above mentioned drivers you now need to explicitly enable
354 ifconfig bge0 media auto mediaopt flowcontrol
356 Along with the above mentioned changes generic support for setting
357 1000baseT master mode also has been added and brgphy(4), ciphy(4),
358 e1000phy(4) as well as ip1000phy(4) have been converted to take
359 advantage of it. This means that these drivers now no longer take the
360 link0 parameter for selecting master mode but the master media option
361 has to be used instead, for example like in the following:
362 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
364 Selection of master mode now is also available with all other PHY
365 drivers supporting 1000baseT.
368 The TCP stack has received a significant update to add support for
369 modularised congestion control and generally improve the clarity of
370 congestion control decisions. Bump __FreeBSD_version to 900025. User
371 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
372 sockstat) need to be recompiled.
375 The man(1) utility has been replaced by a new version that no longer
376 uses /etc/manpath.config. Please consult man.conf(5) for how to
377 migrate local entries to the new format.
380 The copyright strings printed by login(1) and sshd(8) at the time of a
381 new connection have been removed to follow other operating systems and
385 A workaround for a fixed ld bug has been removed in kernel code,
386 so make sure that your system ld is built from sources after
387 revision 210245 from 2010-07-19 (r211583 if building head kernel
388 on stable/8, r211584 for stable/7; both from 2010-08-21).
389 A symptom of incorrect ld version is different addresses for
390 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
393 The $ipv6_prefer variable in rc.conf(5) has been split into
394 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
396 The $ip6addrctl_policy is a variable to choose a pre-defined
397 address selection policy set by ip6addrctl(8). A value
398 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
401 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
402 flag (see an entry of 20090926) is set on an interface with no
403 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
404 security reason. If you want IPv6 link-local address on all
405 interfaces by default, set this to "YES".
407 The old ipv6_prefer="YES" is equivalent to
408 ipv6_activate_all_interfaces="YES" and
409 ip6addrctl_policy="ipv6_prefer".
412 DTrace has grown support for userland tracing. Due to this, DTrace is
413 now i386 and amd64 only.
414 dtruss(1) is now installed by default on those systems and a new
415 kernel module is needed for userland tracing: fasttrap.
416 No changes to your kernel config file are necessary to enable
417 userland tracing, but you might consider adding 'STRIP=' and
418 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
419 to have informative userland stack traces in DTrace (ustack).
422 The acpi_aiboost(4) driver has been removed in favor of the new
423 aibs(4) driver. You should update your kernel configuration file.
426 BSD grep has been imported to the base system and it is built by
427 default. It is completely BSD licensed, highly GNU-compatible, uses
428 less memory than its GNU counterpart and has a small codebase.
429 However, it is slower than its GNU counterpart, which is mostly
430 noticeable for larger searches, for smaller ones it is measurable
431 but not significant. The reason is complex, the most important factor
432 is that we lack a modern and efficient regex library and GNU
433 overcomes this by optimizing the searches internally. Future work
434 on improving the regex performance is planned, for the meantime,
435 users that need better performance, can build GNU grep instead by
436 setting the WITH_GNU_GREP knob.
439 Due to the import of powerpc64 support, all existing powerpc kernel
440 configuration files must be updated with a machine directive like this:
441 machine powerpc powerpc
443 In addition, an updated config(8) is required to build powerpc kernels
447 A new version of ZFS (version 15) has been merged to -HEAD.
448 This version uses a python library for the following subcommands:
449 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
450 For full functionality of these commands the following port must
451 be installed: sysutils/py-zfs
454 'vm_page's are now hashed by physical address to an array of mutexes.
455 Currently this is only used to serialize access to hold_count. Over
456 time the page queue mutex will be peeled away. This changes the size
457 of pmap on every architecture. And requires all callers of vm_page_hold
458 and vm_page_unhold to be updated.
461 WITH_CTF can now be specified in src.conf (not recommended, there
462 are some problems with static executables), make.conf (would also
463 affect ports which do not use GNU make and do not override the
464 compile targets) or in the kernel config (via "makeoptions
466 When WITH_CTF was specified there before this was silently ignored,
467 so make sure that WITH_CTF is not used in places which could lead
468 to unwanted behavior.
471 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
472 to allow 32-bit compatibility on non-x86 platforms. All kernel
473 configurations on amd64 and ia64 platforms using these options must
474 be modified accordingly.
477 The utmp user accounting database has been replaced with utmpx,
478 the user accounting interface standardized by POSIX.
479 Unfortunately the semantics of utmp and utmpx don't match,
480 making it practically impossible to support both interfaces.
481 The user accounting database is used by tools like finger(1),
482 last(1), talk(1), w(1) and ac(8).
484 All applications in the base system use utmpx. This means only
485 local binaries (e.g. from the ports tree) may still use these
486 utmp database files. These applications must be rebuilt to make
489 After the system has been upgraded, it is safe to remove the old
490 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
491 assuming their contents is of no importance anymore. Old wtmp
492 databases can only be used by last(1) and ac(8) after they have
493 been converted to the new format using wtmpcvt(1).
496 Introduce the kernel thread "deadlock resolver" (which can be enabled
497 via the DEADLKRES option, see NOTES for more details) and the
498 sleepq_type() function for sleepqueues.
501 The rc.firewall and rc.firewall6 were unified, and
502 rc.firewall6 and rc.d/ip6fw were removed.
503 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
504 variables are obsoleted. Instead, the following new rc
505 variables are added to rc.d/ipfw:
507 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
508 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
509 firewall_simple_onet_ipv6, firewall_trusted_ipv6
511 The meanings correspond to the relevant IPv4 variables.
517 The default terminal emulation for syscons(4) has been changed
518 from cons25 to xterm on all platforms except pc98. This means
519 that the /etc/ttys file needs to be updated to ensure correct
520 operation of applications on the console.
522 The terminal emulation style can be toggled per window by using
523 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
524 options can be used to change the compile-time default back to
527 To prevent graphical artifacts, make sure the TERM environment
528 variable is set to match the terminal emulation that is being
529 performed by syscons(4).
532 The layout of the structure ieee80211req_scan_result has changed.
533 Applications that require wireless scan results (e.g. ifconfig(8))
534 from net80211 need to be recompiled.
536 Applications such as wpa_supplicant(8) may require a full world
537 build without using NO_CLEAN in order to get synchronized with the
541 The iwn(4) driver has been updated to support the 5000 and 5150 series.
542 There's one kernel module for each firmware. Adding "device iwnfw"
543 to the kernel configuration file means including all three firmware
544 images inside the kernel. If you want to include just the one for
545 your wireless card, use the devices iwn4965fw, iwn5000fw or
549 The rc.d/network_ipv6, IPv6 configuration script has been integrated
550 into rc.d/netif. The changes are the following:
552 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
553 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
554 Note that both variables need the "inet6" keyword at the head.
556 Do not set $ipv6_network_interfaces manually if you do not
557 understand what you are doing. It is not needed in most cases.
559 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
562 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
563 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
565 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
566 all configured interfaces have "inet6 accept_rtadv" in the
567 $ifconfig_IF_ipv6. These are for backward compatibility.
569 3. A new variable $ipv6_prefer has been added. If NO, IPv6
570 functionality of interfaces with no corresponding
571 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
572 and the default address selection policy of ip6addrctl(8)
573 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
574 Note that if you want to configure IPv6 functionality on the
575 disabled interfaces after boot, first you need to clear the flag by
576 using ifconfig(8) like:
578 ifconfig em0 inet6 -ifdisabled
580 If YES, the default address selection policy is set as
583 The default value of $ipv6_prefer is NO.
585 4. If your system need to receive Router Advertisement messages,
586 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
587 scripts automatically invoke rtsol(8) when the interface becomes
588 UP. The Router Advertisement messages are used for SLAAC
589 (State-Less Address AutoConfiguration).
592 802.11s D3.03 support was committed. This is incompatible with the
593 previous code, which was based on D3.0.
596 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
597 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
598 control whether accepting Router Advertisement messages or not.
599 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
600 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
601 The ifconfig(8) utility now supports these flags.
604 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
605 mount(8) and -a option for df(1) to see them.
608 The old tunable hw.bus.devctl_disable has been superseded by
609 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
610 replaced by hw.bus.devctl_queue=0. The default for this new tunable
614 Remove the option STOP_NMI. The default action is now to use NMI only
615 for KDB via the newly introduced function stop_cpus_hard() and
616 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
619 The stable/8 branch created in subversion. This corresponds to the
620 RELENG_8 branch in CVS.
623 Bump the shared library version numbers for all libraries that do not
624 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
625 __FreeBSD_version to 800105.
628 Due to changes in the implementation of virtual network stack support,
629 all network-related kernel modules must be recompiled. As this change
630 breaks the ABI, bump __FreeBSD_version to 800104.
633 The TOE interface to the TCP syncache has been modified to remove
634 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
635 The cxgb driver is the only TOE consumer affected by this change, and
636 needs to be recompiled along with the kernel. As this change breaks
637 the ABI, bump __FreeBSD_version to 800103.
640 Padding has been added to struct tcpcb, sackhint and tcpstat in
641 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
642 maintaining the ABI. However, this change breaks the ABI, so bump
643 __FreeBSD_version to 800102. User space tools that rely on the size of
644 any of these structs (e.g. sockstat) need to be recompiled.
647 The NFS_LEGACYRPC option has been removed along with the old kernel
648 RPC implementation that this option selected. Kernel configurations
649 may need to be adjusted.
652 The network interface device nodes at /dev/net/<interface> have been
653 removed. All ioctl operations can be performed the normal way using
654 routing sockets. The kqueue functionality can generally be replaced
655 with routing sockets.
658 The documentation from the FreeBSD Documentation Project (Handbook,
659 FAQ, etc.) is now installed via packages by sysinstall(8) and under
660 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
663 The ABI of various structures related to the SYSV IPC API have been
664 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
665 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
669 Layout of struct vnet has changed as routing related variables were
670 moved to their own Vimage module. Modules need to be recompiled. Bump
671 __FreeBSD_version to 800099.
674 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
675 respectively. As long as no more than 16 groups per process are used,
676 no changes should be visible. When more than 16 groups are used, old
677 binaries may fail if they call getgroups() or getgrouplist() with
678 statically sized storage. Recompiling will work around this, but
679 applications should be modified to use dynamically allocated storage
680 for group arrays as POSIX.1-2008 does not cap an implementation's
681 number of supported groups at NGROUPS_MAX+1 as previous versions did.
683 NFS and portalfs mounts may also be affected as the list of groups is
684 truncated to 16. Users of NFS who use more than 16 groups, should
685 take care that negative group permissions are not used on the exported
686 file systems as they will not be reliable unless a GSSAPI based
687 authentication method is used.
690 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
691 option compiles in the support for adaptive spinning for lockmgrs
692 which want to enable it. The lockinit() function now accepts the flag
693 LK_ADAPTIVE in order to make the lock object subject to adaptive
694 spinning when both held in write and read mode.
697 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
698 changed. User applications that use this ioctl need to be rebuilt.
701 The layout of struct thread has changed. Kernel and modules need to
705 The layout of structs ifnet, domain, protosw and vnet_net has changed.
706 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
709 window(1) has been removed from the base system. It can now be
710 installed from ports. The port is called misc/window.
713 The way we are storing and accessing `routing table' entries has
714 changed. Programs reading the FIB, like netstat, need to be
718 A new netisr implementation has been added for FreeBSD 8. Network
719 file system modules, such as igmp, ipdivert, and others, should be
721 Bump __FreeBSD_version to 800096.
724 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
728 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
729 Bump __FreeBSD_version to 800094.
732 Add mnt_xflag field to 'struct mount'. File system modules need to be
734 Bump __FreeBSD_version to 800093.
737 The compiling option ADAPTIVE_SX has been retired while it has been
738 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
739 The KPI for sx_init_flags() changes as accepting flags:
740 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
741 been introduced in order to handle the reversed logic.
742 Bump __FreeBSD_version to 800092.
745 Add support for hierarchical jails. Remove global securelevel.
746 Bump __FreeBSD_version to 800091.
749 The layout of struct vnet_net has changed, therefore modules
751 Bump __FreeBSD_version to 800090.
754 The newly imported zic(8) produces a new format in the output. Please
755 run tzsetup(8) to install the newly created data to /etc/localtime.
758 The sysctl tree for the usb stack has renamed from hw.usb2.* to
759 hw.usb.* and is now consistent again with previous releases.
762 802.11 monitor mode support was revised and driver api's were changed.
763 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
764 of DLT_IEEE802_11. No user-visible data structures were changed but
765 applications that use DLT_IEEE802_11 may require changes.
766 Bump __FreeBSD_version to 800088.
769 The layout of the following structs has changed: sysctl_oid,
770 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
771 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
772 panics may be experienced. World rebuild is required for
773 correctly checking networking state from userland.
774 Bump __FreeBSD_version to 800085.
777 MLDv2 and Source-Specific Multicast (SSM) have been merged
778 to the IPv6 stack. VIMAGE hooks are in but not yet used.
779 The implementation of SSM within FreeBSD's IPv6 stack closely
780 follows the IPv4 implementation.
782 For kernel developers:
784 * The most important changes are that the ip6_output() and
785 ip6_input() paths no longer take the IN6_MULTI_LOCK,
786 and this lock has been downgraded to a non-recursive mutex.
788 * As with the changes to the IPv4 stack to support SSM, filtering
789 of inbound multicast traffic must now be performed by transport
790 protocols within the IPv6 stack. This does not apply to TCP and
791 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
793 * The KPIs used by IPv6 multicast are similar to those used by
794 the IPv4 stack, with the following differences:
795 * im6o_mc_filter() is analogous to imo_multicast_filter().
796 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
797 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
798 * IN6_LOOKUP_MULTI() has been deprecated and removed.
799 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
800 for MLDv1 have an additional 'timer' argument which is used to
801 jitter the initial membership report for the solicited-node
802 multicast membership on-link.
803 * This is not strictly needed for MLDv2, which already jitters
804 its report transmissions. However, the 'timer' argument is
805 preserved in case MLDv1 is active on the interface.
807 * The KAME linked-list based IPv6 membership implementation has
808 been refactored to use a vector similar to that used by the IPv4
810 Code which maintains a list of its own multicast memberships
811 internally, e.g. carp, has been updated to reflect the new
814 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
815 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
816 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
817 implementation constraint which needs to be addressed in HEAD.
819 For application developers:
821 * The changes are broadly similar to those made for the IPv4
824 * The use of IPv4 and IPv6 multicast socket options on the same
825 socket, using mapped addresses, HAS NOT been tested or supported.
827 * There are a number of issues with the implementation of various
828 IPv6 multicast APIs which need to be resolved in the API surface
829 before the implementation is fully compatible with KAME userland
830 use, and these are mostly to do with interface index treatment.
832 * The literature available discusses the use of either the delta / ASM
833 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
834 using setsourcefilter(3)/getsourcefilter(3). For more information
835 please refer to RFC 3768, 'Socket Interface Extensions for
836 Multicast Source Filters'.
838 * Applications which use the published RFC 3678 APIs should be fine.
840 For systems administrators:
842 * The mtest(8) utility has been refactored to support IPv6, in
843 addition to IPv4. Interface addresses are no longer accepted
844 as arguments, their names must be used instead. The utility
845 will map the interface name to its first IPv4 address as
846 returned by getifaddrs(3).
848 * The ifmcstat(8) utility has also been updated to print the MLDv2
849 endpoint state and source filter lists via sysctl(3).
851 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
852 loopback of IPv6 multicast datagrams by default; it defaults to 1
853 to preserve the existing behaviour. Disabling multicast loopback is
854 recommended for optimal system performance.
856 * The IPv6 MROUTING code has been changed to examine this sysctl
857 instead of attempting to perform a group lookup before looping
858 back forwarded datagrams.
860 Bump __FreeBSD_version to 800084.
863 Implement low-level Bluetooth HCI API.
864 Bump __FreeBSD_version to 800083.
867 The layout of struct malloc_type, used by modules to register new
868 memory allocation types, has changed. Most modules will need to
869 be rebuilt or panics may be experienced.
870 Bump __FreeBSD_version to 800081.
873 Anticipate overflowing inp_flags - add inp_flags2.
874 This changes most offsets in inpcb, so checking v4 connection
875 state will require a world rebuild.
876 Bump __FreeBSD_version to 800080.
879 Add an llentry to struct route and struct route_in6. Modules
880 embedding a struct route will need to be recompiled.
881 Bump __FreeBSD_version to 800079.
884 The size of rt_metrics_lite and by extension rtentry has changed.
885 Networking administration apps will need to be recompiled.
886 The route command now supports show as an alias for get, weighting
887 of routes, sticky and nostick flags to alter the behavior of stateful
889 Bump __FreeBSD_version to 800078.
892 Do not use Giant for kbdmux(4) locking. This is wrong and
893 apparently causing more problems than it solves. This will
894 re-open the issue where interrupt handlers may race with
895 kbdmux(4) in polling mode. Typical symptoms include (but
896 not limited to) duplicated and/or missing characters when
897 low level console functions (such as gets) are used while
898 interrupts are enabled (for example geli password prompt,
899 mountroot prompt etc.). Disabling kbdmux(4) may help.
902 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
903 kernel modules referencing any of the above need to be recompiled.
904 Bump __FreeBSD_version to 800075.
907 GEOM_PART has become the default partition slicer for storage devices,
908 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
909 introduces some changes:
911 MSDOS/EBR: the devices created from MSDOS extended partition entries
912 (EBR) can be named differently than with GEOM_MBR and are now symlinks
913 to devices with offset-based names. fstabs may need to be modified.
915 BSD: the "geometry does not match label" warning is harmless in most
916 cases but it points to problems in file system misalignment with
917 disk geometry. The "c" partition is now implicit, covers the whole
918 top-level drive and cannot be (mis)used by users.
920 General: Kernel dumps are now not allowed to be written to devices
921 whose partition types indicate they are meant to be used for file
922 systems (or, in case of MSDOS partitions, as something else than
925 Most of these changes date approximately from 200812.
928 The uscanner(4) driver has been removed from the kernel. This follows
929 Linux removing theirs in 2.6 and making libusb the default interface
933 The multicast forwarding code has been cleaned up. netstat(1)
934 only relies on KVM now for printing bandwidth upcall meters.
935 The IPv4 and IPv6 modules are split into ip_mroute_mod and
936 ip6_mroute_mod respectively. The config(5) options for statically
937 compiling this code remain the same, i.e. 'options MROUTING'.
940 Support for the IFF_NEEDSGIANT network interface flag has been
941 removed, which means that non-MPSAFE network device drivers are no
942 longer supported. In particular, if_ar, if_sr, and network device
943 drivers from the old (legacy) USB stack can no longer be built or
947 POSIX.1 Native Language Support (NLS) has been enabled in libc and
948 a bunch of new language catalog files have also been added.
949 This means that some common libc messages are now localized and
950 they depend on the LC_MESSAGES environmental variable.
953 The k8temp(4) driver has been renamed to amdtemp(4) since
954 support for Family 10 and Family 11 CPU families was added.
957 IGMPv3 and Source-Specific Multicast (SSM) have been merged
958 to the IPv4 stack. VIMAGE hooks are in but not yet used.
960 For kernel developers, the most important changes are that the
961 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
962 and this lock has been downgraded to a non-recursive mutex.
964 Transport protocols (UDP, Raw IP) are now responsible for filtering
965 inbound multicast traffic according to group membership and source
966 filters. The imo_multicast_filter() KPI exists for this purpose.
967 Transports which do not use multicast (SCTP, TCP) already reject
968 multicast by default. Forwarding and receive performance may improve
969 as a mutex acquisition is no longer needed in the ip_input()
970 low-level input path. in_addmulti() and in_delmulti() are shimmed
971 to new KPIs which exist to support SSM in-kernel.
973 For application developers, it is recommended that loopback of
974 multicast datagrams be disabled for best performance, as this
975 will still cause the lock to be taken for each looped-back
976 datagram transmission. The net.inet.ip.mcast.loop sysctl may
977 be tuned to 0 to disable loopback by default; it defaults to 1
978 to preserve the existing behaviour.
980 For systems administrators, to obtain best performance with
981 multicast reception and multiple groups, it is always recommended
982 that a card with a suitably precise hash filter is used. Hash
983 collisions will still result in the lock being taken within the
984 transport protocol input path to check group membership.
986 If deploying FreeBSD in an environment with IGMP snooping switches,
987 it is recommended that the net.inet.igmp.sendlocal sysctl remain
988 enabled; this forces 224.0.0.0/24 group membership to be announced
991 The size of 'struct igmpstat' has changed; netstat needs to be
992 recompiled to reflect this.
993 Bump __FreeBSD_version to 800070.
996 libusb20.so.1 is now installed as libusb.so.1 and the ports system
997 updated to use it. This requires a buildworld/installworld in order to
998 update the library and dependencies (usbconfig, etc). Its advisable to
999 rebuild all ports which uses libusb. More specific directions are given
1000 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1001 libusb are no longer required and can be removed.
1004 A workaround is committed to allow the creation of System V shared
1005 memory segment of size > 2 GB on the 64-bit architectures.
1006 Due to a limitation of the existing ABI, the shm_segsz member
1007 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1008 wrong for large segments. Note that limits must be explicitly
1009 raised to allow such segments to be created.
1012 The layout of struct ifnet has changed, requiring a rebuild of all
1013 network device driver modules.
1016 The /dev handling for the new USB stack has changed, a
1017 buildworld/installworld is required for libusb20.
1020 The new USB2 stack has now been permanently moved in and all kernel and
1021 module names reverted to their previous values (eg, usb, ehci, ohci,
1022 ums, ...). The old usb stack can be compiled in by prefixing the name
1023 with the letter 'o', the old usb modules have been removed.
1024 Updating entry 20090216 for xorg and 20090215 for libmap may still
1028 The rc.conf(5) option if_up_delay has been renamed to
1029 defaultroute_delay to better reflect its purpose. If you have
1030 customized this setting in /etc/rc.conf you need to update it to
1034 xorg 7.4 wants to configure its input devices via hald which does not
1035 yet work with USB2. If the keyboard/mouse does not work in xorg then
1037 Option "AllowEmptyInput" "off"
1038 to your ServerLayout section. This will cause X to use the configured
1039 kbd and mouse sections from your xorg.conf.
1042 The GENERIC kernels for all architectures now default to the new USB2
1043 stack. No kernel config options or code have been removed so if a
1044 problem arises please report it and optionally revert to the old USB
1045 stack. If you are loading USB kernel modules or have a custom kernel
1046 that includes GENERIC then ensure that usb names are also changed over,
1047 eg uftdi -> usb2_serial_ftdi.
1049 Older programs linked against the ports libusb 0.1 need to be
1050 redirected to the new stack's libusb20. /etc/libmap.conf can
1052 # Map old usb library to new one for usb2 stack
1053 libusb-0.1.so.8 libusb20.so.1
1056 All USB ethernet devices now attach as interfaces under the name ueN
1057 (eg. ue0). This is to provide a predictable name as vendors often
1058 change usb chipsets in a product without notice.
1061 The ichsmb(4) driver has been changed to require SMBus slave
1062 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1063 All of the other SMBus controller drivers require left-justified
1064 slave addresses, so this change makes all the drivers provide the
1068 INET6 statistics (struct ip6stat) was updated.
1069 netstat(1) needs to be recompiled.
1072 NTFS has been removed from GENERIC kernel on amd64 to match
1073 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1074 will load ntfs.ko module automatically when NTFS support is
1075 actually needed, unless ntfs.ko is not installed or security
1076 level prohibits loading kernel modules. If either is the case,
1077 "options NTFS" has to be added into kernel config.
1080 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1081 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1082 800061. User space tools that rely on the size of struct tcpcb in
1083 tcp_var.h (e.g. sockstat) need to be recompiled.
1086 ng_tty(4) module updated to match the new TTY subsystem.
1087 Due to API change, user-level applications must be updated.
1088 New API support added to mpd5 CVS and expected to be present
1089 in next mpd5.3 release.
1092 With __FreeBSD_version 800060 the makefs tool is part of
1093 the base system (it was a port).
1096 The afdata and ifnet locks have been changed from mutexes to
1097 rwlocks, network modules will need to be re-compiled.
1100 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1101 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1102 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1103 architecture and 40 bytes on 64-bit architecture. The userland
1104 applications "arp" and "ndp" have been updated accordingly.
1105 The output from "netstat -r" shows only routing entries and
1106 none of the L2 information.
1109 __FreeBSD_version 800057 marks the switchover from the
1110 binary ath hal to source code. Users must add the line:
1112 options AH_SUPPORT_AR5416
1114 to their kernel config files when specifying:
1118 The ath_hal module no longer exists; the code is now compiled
1119 together with the driver in the ath module. It is now
1120 possible to tailor chip support (i.e. reduce the set of chips
1121 and thereby the code size); consult ath_hal(4) for details.
1124 __FreeBSD_version 800054 adds memory barriers to
1125 <machine/atomic.h>, new interfaces to ifnet to facilitate
1126 multiple hardware transmit queues for cards that support
1127 them, and a lock-less ring-buffer implementation to
1128 enable drivers to more efficiently manage queueing of
1132 A new version of ZFS (version 13) has been merged to -HEAD.
1133 This version has zpool attribute "listsnapshots" off by
1134 default, which means "zfs list" does not show snapshots,
1135 and is the same as Solaris behavior.
1138 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1141 The uhci, ohci, ehci and slhci USB Host controller drivers have
1142 been put into separate modules. If you load the usb module
1143 separately through loader.conf you will need to load the
1144 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1145 controller add the following to loader.conf:
1151 The ABI used by the PMC toolset has changed. Please keep
1152 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1156 atapci kernel module now includes only generic PCI ATA
1157 driver. AHCI driver moved to ataahci kernel module.
1158 All vendor-specific code moved into separate kernel modules:
1159 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1160 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1161 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1162 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1165 The TTY subsystem of the kernel has been replaced by a new
1166 implementation, which provides better scalability and an
1167 improved driver model. Most common drivers have been migrated to
1168 the new TTY subsystem, while others have not. The following
1169 drivers have not yet been ported to the new TTY layer:
1172 cy, digi, rc, rp, sio
1178 ng_h4, ng_tty, ppp, sl, snp
1180 Adding these drivers to your kernel configuration file shall
1181 cause compilation to fail.
1184 ntpd has been upgraded to 4.2.4p5.
1187 OpenSSH has been upgraded to 5.1p1.
1189 For many years, FreeBSD's version of OpenSSH preferred DSA
1190 over RSA for host and user authentication keys. With this
1191 upgrade, we've switched to the vendor's default of RSA over
1192 DSA. This may cause upgraded clients to warn about unknown
1193 host keys even for previously known hosts. Users should
1194 follow the usual procedure for verifying host keys before
1195 accepting the RSA key.
1197 This can be circumvented by setting the "HostKeyAlgorithms"
1198 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1201 Please note that the sequence of keys offered for
1202 authentication has been changed as well. You may want to
1203 specify IdentityFile in a different order to revert this
1207 The sio(4) driver has been removed from the i386 and amd64
1208 kernel configuration files. This means uart(4) is now the
1209 default serial port driver on those platforms as well.
1211 To prevent collisions with the sio(4) driver, the uart(4) driver
1212 uses different names for its device nodes. This means the
1213 onboard serial port will now most likely be called "ttyu0"
1214 instead of "ttyd0". You may need to reconfigure applications to
1215 use the new device names.
1217 When using the serial port as a boot console, be sure to update
1218 /boot/device.hints and /etc/ttys before booting the new kernel.
1219 If you forget to do so, you can still manually specify the hints
1220 at the loader prompt:
1222 set hint.uart.0.at="isa"
1223 set hint.uart.0.port="0x3F8"
1224 set hint.uart.0.flags="0x10"
1225 set hint.uart.0.irq="4"
1229 The gpt(8) utility has been removed. Use gpart(8) to partition
1233 The version that Linuxulator emulates was changed from 2.4.2
1234 to 2.6.16. If you experience any problems with Linux binaries
1235 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1236 if it fixes the problem contact emulation mailing list.
1239 ISDN4BSD (I4B) was removed from the src tree. You may need to
1240 update a your kernel configuration and remove relevant entries.
1243 I have checked in code to support multiple routing tables.
1244 See the man pages setfib(1) and setfib(2).
1245 This is a hopefully backwards compatible version,
1246 but to make use of it you need to compile your kernel
1247 with options ROUTETABLES=2 (or more up to 16).
1250 The 802.11 wireless support was redone to enable multi-bss
1251 operation on devices that are capable. The underlying device
1252 is no longer used directly but instead wlanX devices are
1253 cloned with ifconfig. This requires changes to rc.conf files.
1254 For example, change:
1255 ifconfig_ath0="WPA DHCP"
1258 ifconfig_wlan0="WPA DHCP"
1259 see rc.conf(5) for more details. In addition, mergemaster of
1260 /etc/rc.d is highly recommended. Simultaneous update of userland
1261 and kernel wouldn't hurt either.
1263 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1264 modules were merged into the base wlan module. All references
1265 to these modules (e.g. in kernel config files) must be removed.
1268 psm(4) has gained write(2) support in native operation level.
1269 Arbitrary commands can be written to /dev/psm%d and status can
1270 be read back from it. Therefore, an application is responsible
1271 for status validation and error recovery. It is a no-op in
1272 other operation levels.
1275 Support for KSE threading has been removed from the kernel. To
1276 run legacy applications linked against KSE libmap.conf may
1277 be used. The following libmap.conf may be used to ensure
1278 compatibility with any prior release:
1280 libpthread.so.1 libthr.so.1
1281 libpthread.so.2 libthr.so.2
1282 libkse.so.3 libthr.so.3
1285 The layout of struct vmspace has changed. This affects libkvm
1286 and any executables that link against libkvm and use the
1287 kvm_getprocs() function. In particular, but not exclusively,
1288 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1289 The effects are minimal, but it's advisable to upgrade world
1293 The latest em driver no longer has support in it for the
1294 82575 adapter, this is now moved to the igb driver. The
1295 split was done to make new features that are incompatible
1296 with older hardware easier to do.
1299 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1300 likewise the kernel option is now GEOM_LINUX_LVM.
1303 The default NFS mount mode has changed from UDP to TCP for
1304 increased reliability. If you rely on (insecurely) NFS
1305 mounting across a firewall you may need to update your
1309 Belatedly note the addition of m_collapse for compacting
1313 The fts(3) structures have been changed to use adequate
1314 integer types for their members and so to be able to cope
1315 with huge file trees. The old fts(3) ABI is preserved
1316 through symbol versioning in libc, so third-party binaries
1317 using fts(3) should still work, although they will not take
1318 advantage of the extended types. At the same time, some
1319 third-party software might fail to build after this change
1320 due to unportable assumptions made in its source code about
1321 fts(3) structure members. Such software should be fixed
1322 by its vendor or, in the worst case, in the ports tree.
1323 FreeBSD_version 800015 marks this change for the unlikely
1324 case that a portable fix is impossible.
1327 To upgrade to -current after this date, you must be running
1328 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1329 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1332 The ADAPTIVE_GIANT kernel option has been retired because its
1333 functionality is the default now.
1336 The AT keyboard emulation of sunkbd(4) has been turned on
1337 by default. In order to make the special symbols of the Sun
1338 keyboards driven by sunkbd(4) work under X these now have
1339 to be configured the same way as Sun USB keyboards driven
1340 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1342 Option "XkbLayout" "us"
1343 Option "XkbRules" "xorg"
1344 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1347 It has been decided that it is desirable to provide ABI
1348 backwards compatibility to the FreeBSD 4/5/6 versions of the
1349 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1350 broken with the introduction of PCI domain support (see the
1351 20070930 entry). Unfortunately, this required the ABI of
1352 PCIOCGETCONF to be broken again in order to be able to
1353 provide backwards compatibility to the old version of that
1354 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1355 again. As for prominent ports this affects neither pciutils
1356 nor xorg-server this time, the hal port needs to be rebuilt
1360 The misnamed kthread_create() and friends have been renamed
1361 to kproc_create() etc. Many of the callers already
1362 used kproc_start()..
1363 I will return kthread_create() and friends in a while
1364 with implementations that actually create threads, not procs.
1365 Renaming corresponds with version 800002.
1374 Avoid using make -j when upgrading. While generally safe, there are
1375 sometimes problems using -j to upgrade. If your upgrade fails with
1376 -j, please try again without -j. From time to time in the past there
1377 have been problems using -j with buildworld and/or installworld. This
1378 is especially true when upgrading between "distant" versions (eg one
1379 that cross a major release boundary or several minor releases, or when
1380 several months have passed on the -current branch).
1382 Sometimes, obscure build problems are the result of environment
1383 poisoning. This can happen because the make utility reads its
1384 environment when searching for values for global variables. To run
1385 your build attempts in an "environmental clean room", prefix all make
1386 commands with 'env -i '. See the env(1) manual page for more details.
1388 When upgrading from one major version to another it is generally best
1389 to upgrade to the latest code in the currently installed branch first,
1390 then do an upgrade to the new branch. This is the best-tested upgrade
1391 path, and has the highest probability of being successful. Please try
1392 this approach before reporting problems with a major version upgrade.
1396 When upgrading the boot ZFS pool to a new version, always follow
1399 1.) recompile and reinstall the ZFS boot loader and boot block
1400 (this is part of "make buildworld" and "make installworld")
1402 2.) update the ZFS boot block on your boot drive
1404 The following example updates the ZFS boot block on the first
1405 partition (freebsd-boot) of a GPT partitioned drive ad0:
1406 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1408 Non-boot pools do not need these updates.
1412 If you are updating from a prior version of FreeBSD (even one just
1413 a few days old), you should follow this procedure. It is the most
1414 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1416 make kernel-toolchain
1417 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1418 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1420 To test a kernel once
1421 ---------------------
1422 If you just want to boot a kernel once (because you are not sure
1423 if it works, or if you want to boot a known bad kernel to provide
1424 debugging information) run
1425 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1426 nextboot -k testkernel
1428 To just build a kernel when you know that it won't mess you up
1429 --------------------------------------------------------------
1430 This assumes you are already running a CURRENT system. Replace
1431 ${arch} with the architecture of your machine (e.g. "i386",
1432 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1434 cd src/sys/${arch}/conf
1435 config KERNEL_NAME_HERE
1436 cd ../compile/KERNEL_NAME_HERE
1441 If this fails, go to the "To build a kernel" section.
1443 To rebuild everything and install it on the current system.
1444 -----------------------------------------------------------
1445 # Note: sometimes if you are running current you gotta do more than
1446 # is listed here if you are upgrading from a really old current.
1448 <make sure you have good level 0 dumps>
1450 make kernel KERNCONF=YOUR_KERNEL_HERE
1452 <reboot in single user> [3]
1460 To cross-install current onto a separate partition
1461 --------------------------------------------------
1462 # In this approach we use a separate partition to hold
1463 # current's root, 'usr', and 'var' directories. A partition
1464 # holding "/", "/usr" and "/var" should be about 2GB in
1467 <make sure you have good level 0 dumps>
1470 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1471 <maybe newfs current's root partition>
1472 <mount current's root partition on directory ${CURRENT_ROOT}>
1473 make installworld DESTDIR=${CURRENT_ROOT}
1474 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1475 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1476 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1477 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1478 <reboot into current>
1479 <do a "native" rebuild/install as described in the previous section>
1480 <maybe install compatibility libraries from ports/misc/compat*>
1484 To upgrade in-place from stable to current
1485 ----------------------------------------------
1486 <make sure you have good level 0 dumps>
1488 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1490 <reboot in single user> [3]
1497 Make sure that you've read the UPDATING file to understand the
1498 tweaks to various things you need. At this point in the life
1499 cycle of current, things change often and you are on your own
1500 to cope. The defaults can also change, so please read ALL of
1501 the UPDATING entries.
1503 Also, if you are tracking -current, you must be subscribed to
1504 freebsd-current@freebsd.org. Make sure that before you update
1505 your sources that you have read and understood all the recent
1506 messages there. If in doubt, please track -stable which has
1507 much fewer pitfalls.
1509 [1] If you have third party modules, such as vmware, you
1510 should disable them at this point so they don't crash your
1513 [3] From the bootblocks, boot -s, and then do
1518 adjkerntz -i # if CMOS is wall time
1519 Also, when doing a major release upgrade, it is required that
1520 you boot into single user mode to do the installworld.
1522 [4] Note: This step is non-optional. Failure to do this step
1523 can result in a significant reduction in the functionality of the
1524 system. Attempting to do it by hand is not recommended and those
1525 that pursue this avenue should read this file carefully, as well
1526 as the archives of freebsd-current and freebsd-hackers mailing lists
1527 for potential gotchas. The -U option is also useful to consider.
1528 See mergemaster(8) for more information.
1530 [5] Usually this step is a noop. However, from time to time
1531 you may need to do this if you get unknown user in the following
1532 step. It never hurts to do it all the time. You may need to
1533 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1534 install) after the buildworld before this step if you last updated
1535 from current before 20020224 or from -stable before 20020408.
1537 [6] This only deletes old files and directories. Old libraries
1538 can be deleted by "make delete-old-libs", but you have to make
1539 sure that no program is using those libraries anymore.
1541 [8] In order to have a kernel that can run the 4.x binaries needed to
1542 do an installworld, you must include the COMPAT_FREEBSD4 option in
1543 your kernel. Failure to do so may leave you with a system that is
1544 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1545 required to run the 5.x binaries on more recent kernels. And so on
1546 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1548 Make sure that you merge any new devices from GENERIC since the
1549 last time you updated your kernel config file.
1551 [9] When checking out sources, you must include the -P flag to have
1552 cvs prune empty directories.
1554 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1555 "?=" instead of the "=" assignment operator, so that buildworld can
1556 override the CPUTYPE if it needs to.
1558 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1559 not on the command line, or in /etc/make.conf. buildworld will
1560 warn if it is improperly defined.
1563 This file contains a list, in reverse chronological order, of major
1564 breakages in tracking -current. It is not guaranteed to be a complete
1565 list of such breakages, and only contains entries since October 10, 2007.
1566 If you need to see UPDATING entries from before that date, you will need
1567 to fetch an UPDATING file from an older FreeBSD release.
1569 Copyright information:
1571 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1573 Redistribution, publication, translation and use, with or without
1574 modification, in full or in part, in any form or format of this
1575 document are permitted without further permission from the author.
1577 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1578 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1579 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1580 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1581 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1582 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1583 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1584 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1585 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1586 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1587 POSSIBILITY OF SUCH DAMAGE.
1589 Contact Warner Losh if you have any questions about your use of