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 Items affecting the ports and packages system can be found in
10 /usr/ports/UPDATING. Please read that file before running portupgrade.
12 20130402: p7 FreeBSD-SA-13:03.openssl FreeBSD-SA-13:04.bind
13 Fix multiple vulnerabilities in OpenSSL.
15 Fix Denial of Service vulnerability in named(8).
17 20130218: p6 FreeBSD-SA-13:01.bind FreeBSD-SA-13:02.libc
18 Fix Denial of Service vulnerability in named(8) with DNS64.
20 Fix Denial of Service vulnerability in libc's glob(3) functionality.
22 20121122: p5 FreeBSD-SA-12:06.bind FreeBSD-SA-12:07.hostapd
23 FreeBSD-SA-12:08.linux
24 Fix multiple Denial of Service vulnerabilities with named(8).
26 Fix insufficient message length validation for EAP-TLS messages.
28 Fix Linux compatibility layer input validation error.
30 20120806: p4 FreeBSD-SA-12:05.bind
31 Fix named(8) DNSSEC validation Denial of Service.
33 20120612: p3 FreeBSD-SA-12:03.bind
34 FreeBSD-SA-12:04.sysret
35 FreeBSD-EN-12:02.ipv6refcount
36 Fix a problem where zero-length RDATA fields can cause named to crash.
39 Correct a privilege escalation when returning from kernel if
40 running FreeBSD/amd64 on non-AMD processors. [12:04]
42 Fix reference count errors in IPv6 code. [EN-12:02]
44 20120530: p2 FreeBSD-SA-12:01.openssl (revised),
45 FreeBSD-SA-12:02.crypt
46 Update the previous openssl fix. [12:01]
48 Fix a bug in crypt(3) ignoring characters of a passphrase. [12:02]
50 20120503: p1 FreeBSD-SA-12:01.openssl
51 Fix multiple OpenSSL vulnerabilities.
57 The broken amd(4) driver has been replaced with esp(4) in the amd64,
58 i386 and pc98 GENERIC kernel configuration files.
61 This commit modifies vfs_register() so that it uses a hash
62 calculation to set vfc_typenum, which is enabled by default.
63 The first time a system is booted after this change, the
64 vfc_typenum values will change for all file systems. The
65 main effect of this is a change to the NFS server file handles
66 for file systems that use vfc_typenum in their fsid, such as ZFS.
67 It will, however, prevent vfc_typenum from changing when file
68 systems are loaded in a different order for subsequent reboots.
69 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
70 until you are ready to remount all NFS clients after a reboot.
73 Bump the shared library version numbers for libraries that
74 do not use symbol versioning, have changed the ABI compared
75 to stable/8 and which shared library version was not bumped.
76 Done as part of 9.0-RELEASE cycle.
79 During the merge of Capsicum features, the fget(9) KPI was modified.
80 This may require the rebuilding of out-of-tree device drivers --
81 issues have been reported specifically with the nVidia device driver.
82 __FreeBSD_version is bumped to 900041.
84 Also, there is a period between 20110811 and 20110814 where the
85 special devices /dev/{stdin,stdout,stderr} did not work correctly.
86 Building world from a kernel during that window may not work.
89 The packet filter (pf) code has been updated to OpenBSD 4.5.
90 You need to update userland tools to be in sync with kernel.
91 This update breaks backward compatibility with earlier pfsync(4)
92 versions. Care must be taken when updating redundant firewall setups.
95 The following sysctls and tunables are retired on x86 platforms:
97 machdep.hlt_logical_cpus
98 The following sysctl is retired:
99 machdep.hyperthreading_allowed
100 The sysctls were supposed to provide a way to dynamically offline and
101 online selected CPUs on x86 platforms, but the implementation has not
102 been reliable especially with SCHED_ULE scheduler.
103 machdep.hyperthreading_allowed tunable is still available to ignore
104 hyperthreading CPUs at OS level.
105 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
106 where X is an APIC ID of a CPU. Be advised, though, that disabling
107 CPUs in non-uniform fashion will result in non-uniform topology and
108 may lead to sub-optimal system performance with SCHED_ULE, which is
112 cpumask_t type is retired and cpuset_t is used in order to describe
116 Changes to ifconfig(8) for dynamic address family detection mandate
117 that you are running a kernel of 20110525 or later. Make sure to
118 follow the update procedure to boot a new kernel before installing
122 Support for sun4v architecture is officially dropped
125 Several KPI breaking changes have been committed to the mii(4) layer,
126 the PHY drivers and consequently some Ethernet drivers using mii(4).
127 This means that miibus.ko and the modules of the affected Ethernet
128 drivers need to be recompiled.
130 Note to kernel developers: Given that the OUI bit reversion problem
131 was fixed as part of these changes all mii(4) commits related to OUIs,
132 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
133 handling, no longer can be merged verbatim to stable/8 and previous
137 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
138 into their kernel configurations along with 'device pci'.
141 The default NFS client is now the new NFS client, so fstype "newnfs"
142 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
143 Although mounts via fstype "nfs" will usually work without userland
144 changes, it is recommended that the mount(8) and mount_nfs(8)
145 commands be rebuilt from sources and that a link to mount_nfs called
146 mount_oldnfs be created. The new client is compiled into the
147 kernel with "options NFSCL" and this is needed for diskless root
148 file systems. The GENERIC kernel configs have been changed to use
149 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
150 To use the regular/old client, you can "mount -t oldnfs ...". For
151 a diskless root file system, you must also include a line like:
153 vfs.root.mountfrom="oldnfs:"
155 in the boot/loader.conf on the root fs on the NFS server to make
156 a diskless root fs use the old client.
159 The GENERIC kernels for all architectures now default to the new
160 CAM-based ATA stack. It means that all legacy ATA drivers were
161 removed and replaced by respective CAM drivers. If you are using
162 ATA device names in /etc/fstab or other places, make sure to update
163 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
164 where 'Y's are the sequential numbers starting from zero for each type
165 in order of detection, unless configured otherwise with tunables,
166 see cam(4)). There will be symbolic links created in /dev/ to map
167 old adX devices to the respective adaY. They should provide basic
168 compatibility for file systems mounting in most cases, but they do
169 not support old user-level APIs and do not have respective providers
170 in GEOM. Consider using updated management tools with new device names.
172 It is possible to load devices ahci, ata, siis and mvs as modules,
173 but option ATA_CAM should remain in kernel configuration to make ata
174 module work as CAM driver supporting legacy ATA controllers. Device ata
175 still can be used in modular fashion (atacore + ...). Modules atadisk
176 and atapi* are not used and won't affect operation in ATA_CAM mode.
177 Note that to use CAM-based ATA kernel should include CAM devices
178 scbus, pass, da (or explicitly ada), cd and optionally others. All of
179 them are parts of the cam module.
181 ataraid(4) functionality is now supported by the RAID GEOM class.
182 To use it you can load geom_raid kernel module and use graid(8) tool
183 for management. Instead of /dev/arX device names, use /dev/raid/rX.
185 No kernel config options or code have been removed, so if a problem
186 arises, please report it and optionally revert to the old ATA stack.
187 In order to do it you can remove from the kernel config:
192 , and instead add back:
193 device atadisk # ATA disk drives
194 device ataraid # ATA RAID drives
195 device atapicd # ATAPI CDROM drives
196 device atapifd # ATAPI floppy drives
197 device atapist # ATAPI tape drives
200 The default NFS server has been changed to the new server, which
201 was referred to as the experimental server. If you need to switch
202 back to the old NFS server, you must now put the "-o" option on
203 both the mountd and nfsd commands. This can be done using the
204 mountd_flags and nfs_server_flags rc.conf variables until an
205 update to the rc scripts is committed, which is coming soon.
208 The GNU Objective-C runtime library (libobjc), and other Objective-C
209 related components have been removed from the base system. If you
210 require an Objective-C library, please use one of the available ports.
213 ath(4) has been split into bus- and device- modules. if_ath contains
214 the HAL, the TX rate control and the network device code. if_ath_pci
215 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
216 contains the AHB glue. Users need to load both if_ath_pci and if_ath
217 in order to use ath on everything else.
219 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
220 need to load if_ath and if_ath_pci for ath(4) operation.
223 As part of the replacement of sysinstall, the process of building
224 release media has changed significantly. For details, please re-read
225 release(7), which has been updated to reflect the new build process.
228 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
229 is the last available version under GPLv2. It brings a number of new
230 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
231 4.1 and SSE 4.2), better support for powerpc64, a number of new
232 directives, and lots of other small improvements. See the ChangeLog
233 file in contrib/binutils for the full details.
236 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
237 compliant, and will now use half of hash for authentication.
238 This will break interoperability with all stacks (including all
239 actual FreeBSD versions) who implement
240 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
242 The only workaround with such peers is to use another HMAC
243 algorithm for IPsec ("phase 2") authentication.
246 Remove the uio_yield prototype and symbol. This function has
247 been misnamed since it was introduced and should not be
248 globally exposed with this name. The equivalent functionality
249 is now available using kern_yield(curthread->td_user_pri).
250 The function remains undocumented.
253 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
254 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
255 for scalar sysctls is defined but disabled. Code that needs
256 UQUAD to pass the type checking that must compile on older
257 systems where the define is not present can check against
258 __FreeBSD_version >= 900030.
260 The system dialog(1) has been replaced with a new version previously
261 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
262 with the previous version, but the libdialog associated with it has
263 a largely incompatible API. As such, the original version of libdialog
264 will be kept temporarily as libodialog, until its base system consumers
265 are replaced or updated. Bump __FreeBSD_version to 900030.
268 If you are trying to run make universe on a -stable system, and you get
269 the following warning:
270 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
271 unknown. config(8) likely too old."
272 or something similar to it, then you must upgrade your -stable system
273 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
274 stable/8) or build the config from the latest stable/8 branch and
275 install it on your system.
277 Prior to this date, building a current universe on 8-stable system from
278 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
279 error in the first kernel build phase. A new config on those old
280 systems will fix that problem for older versions of -current.
283 The TCP stack has been modified to allow Khelp modules to interact with
284 it via helper hook points and store per-connection data in the TCP
285 control block. Bump __FreeBSD_version to 900029. User space tools that
286 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
290 Generic IEEE 802.3 annex 31B full duplex flow control support has been
291 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
292 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
293 to take advantage of it instead of using custom implementations. This
294 means that these drivers now no longer unconditionally advertise
295 support for flow control but only do so if flow control is a selected
296 media option. This was implemented in the generic support that way in
297 order to allow flow control to be switched on and off via ifconfig(8)
298 with the PHY specific default to typically off in order to protect
299 from unwanted effects. Consequently, if you used flow control with
300 one of the above mentioned drivers you now need to explicitly enable
302 ifconfig bge0 media auto mediaopt flowcontrol
304 Along with the above mentioned changes generic support for setting
305 1000baseT master mode also has been added and brgphy(4), ciphy(4),
306 e1000phy(4) as well as ip1000phy(4) have been converted to take
307 advantage of it. This means that these drivers now no longer take the
308 link0 parameter for selecting master mode but the master media option
309 has to be used instead, for example like in the following:
310 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
312 Selection of master mode now is also available with all other PHY
313 drivers supporting 1000baseT.
316 The TCP stack has received a significant update to add support for
317 modularised congestion control and generally improve the clarity of
318 congestion control decisions. Bump __FreeBSD_version to 900025. User
319 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
320 sockstat) need to be recompiled.
323 The man(1) utility has been replaced by a new version that no longer
324 uses /etc/manpath.config. Please consult man.conf(5) for how to
325 migrate local entries to the new format.
328 The copyright strings printed by login(1) and sshd(8) at the time of a
329 new connection have been removed to follow other operating systems and
333 A workaround for a fixed ld bug has been removed in kernel code,
334 so make sure that your system ld is built from sources after
335 revision 210245 from 2010-07-19 (r211583 if building head kernel
336 on stable/8, r211584 for stable/7; both from 2010-08-21).
337 A symptom of incorrect ld version is different addresses for
338 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
341 The $ipv6_prefer variable in rc.conf(5) has been split into
342 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
344 The $ip6addrctl_policy is a variable to choose a pre-defined
345 address selection policy set by ip6addrctl(8). A value
346 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
349 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
350 flag (see an entry of 20090926) is set on an interface with no
351 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
352 security reason. If you want IPv6 link-local address on all
353 interfaces by default, set this to "YES".
355 The old ipv6_prefer="YES" is equivalent to
356 ipv6_activate_all_interfaces="YES" and
357 ip6addrctl_policy="ipv6_prefer".
360 DTrace has grown support for userland tracing. Due to this, DTrace is
361 now i386 and amd64 only.
362 dtruss(1) is now installed by default on those systems and a new
363 kernel module is needed for userland tracing: fasttrap.
364 No changes to your kernel config file are necessary to enable
365 userland tracing, but you might consider adding 'STRIP=' and
366 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
367 to have informative userland stack traces in DTrace (ustack).
370 The acpi_aiboost(4) driver has been removed in favor of the new
371 aibs(4) driver. You should update your kernel configuration file.
374 BSD grep has been imported to the base system and it is built by
375 default. It is completely BSD licensed, highly GNU-compatible, uses
376 less memory than its GNU counterpart and has a small codebase.
377 However, it is slower than its GNU counterpart, which is mostly
378 noticeable for larger searches, for smaller ones it is measurable
379 but not significant. The reason is complex, the most important factor
380 is that we lack a modern and efficient regex library and GNU
381 overcomes this by optimizing the searches internally. Future work
382 on improving the regex performance is planned, for the meantime,
383 users that need better performance, can build GNU grep instead by
384 setting the WITH_GNU_GREP knob.
387 Due to the import of powerpc64 support, all existing powerpc kernel
388 configuration files must be updated with a machine directive like this:
389 machine powerpc powerpc
391 In addition, an updated config(8) is required to build powerpc kernels
395 A new version of ZFS (version 15) has been merged to -HEAD.
396 This version uses a python library for the following subcommands:
397 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
398 For full functionality of these commands the following port must
399 be installed: sysutils/py-zfs
402 'vm_page's are now hashed by physical address to an array of mutexes.
403 Currently this is only used to serialize access to hold_count. Over
404 time the page queue mutex will be peeled away. This changes the size
405 of pmap on every architecture. And requires all callers of vm_page_hold
406 and vm_page_unhold to be updated.
409 WITH_CTF can now be specified in src.conf (not recommended, there
410 are some problems with static executables), make.conf (would also
411 affect ports which do not use GNU make and do not override the
412 compile targets) or in the kernel config (via "makeoptions
414 When WITH_CTF was specified there before this was silently ignored,
415 so make sure that WITH_CTF is not used in places which could lead
416 to unwanted behavior.
419 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
420 to allow 32-bit compatibility on non-x86 platforms. All kernel
421 configurations on amd64 and ia64 platforms using these options must
422 be modified accordingly.
425 The utmp user accounting database has been replaced with utmpx,
426 the user accounting interface standardized by POSIX.
427 Unfortunately the semantics of utmp and utmpx don't match,
428 making it practically impossible to support both interfaces.
429 The user accounting database is used by tools like finger(1),
430 last(1), talk(1), w(1) and ac(8).
432 All applications in the base system use utmpx. This means only
433 local binaries (e.g. from the ports tree) may still use these
434 utmp database files. These applications must be rebuilt to make
437 After the system has been upgraded, it is safe to remove the old
438 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
439 assuming their contents is of no importance anymore. Old wtmp
440 databases can only be used by last(1) and ac(8) after they have
441 been converted to the new format using wtmpcvt(1).
444 Introduce the kernel thread "deadlock resolver" (which can be enabled
445 via the DEADLKRES option, see NOTES for more details) and the
446 sleepq_type() function for sleepqueues.
449 The rc.firewall and rc.firewall6 were unified, and
450 rc.firewall6 and rc.d/ip6fw were removed.
451 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
452 variables are obsoleted. Instead, the following new rc
453 variables are added to rc.d/ipfw:
455 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
456 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
457 firewall_simple_onet_ipv6, firewall_trusted_ipv6
459 The meanings correspond to the relevant IPv4 variables.
465 The default terminal emulation for syscons(4) has been changed
466 from cons25 to xterm on all platforms except pc98. This means
467 that the /etc/ttys file needs to be updated to ensure correct
468 operation of applications on the console.
470 The terminal emulation style can be toggled per window by using
471 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
472 options can be used to change the compile-time default back to
475 To prevent graphical artifacts, make sure the TERM environment
476 variable is set to match the terminal emulation that is being
477 performed by syscons(4).
480 The layout of the structure ieee80211req_scan_result has changed.
481 Applications that require wireless scan results (e.g. ifconfig(8))
482 from net80211 need to be recompiled.
484 Applications such as wpa_supplicant(8) may require a full world
485 build without using NO_CLEAN in order to get synchronized with the
489 The iwn(4) driver has been updated to support the 5000 and 5150 series.
490 There's one kernel module for each firmware. Adding "device iwnfw"
491 to the kernel configuration file means including all three firmware
492 images inside the kernel. If you want to include just the one for
493 your wireless card, use the the devices iwn4965fw, iwn5000fw or
497 The rc.d/network_ipv6, IPv6 configuration script has been integrated
498 into rc.d/netif. The changes are the following:
500 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
501 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
502 Note that both variables need the "inet6" keyword at the head.
504 Do not set $ipv6_network_interfaces manually if you do not
505 understand what you are doing. It is not needed in most cases.
507 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
510 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
511 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
513 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
514 all configured interfaces have "inet6 accept_rtadv" in the
515 $ifconfig_IF_ipv6. These are for backward compatibility.
517 3. A new variable $ipv6_prefer has been added. If NO, IPv6
518 functionality of interfaces with no corresponding
519 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
520 and the default address selection policy of ip6addrctl(8)
521 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
522 Note that if you want to configure IPv6 functionality on the
523 disabled interfaces after boot, first you need to clear the flag by
524 using ifconfig(8) like:
526 ifconfig em0 inet6 -ifdisabled
528 If YES, the default address selection policy is set as
531 The default value of $ipv6_prefer is NO.
533 4. If your system need to receive Router Advertisement messages,
534 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
535 scripts automatically invoke rtsol(8) when the interface becomes
536 UP. The Router Advertisement messages are used for SLAAC
537 (State-Less Address AutoConfiguration).
540 802.11s D3.03 support was committed. This is incompatible with the
541 previous code, which was based on D3.0.
544 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
545 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
546 control whether accepting Router Advertisement messages or not.
547 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
548 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
549 The ifconfig(8) utility now supports these flags.
552 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
553 mount(8) and -a option for df(1) to see them.
556 The old tunable hw.bus.devctl_disable has been superseded by
557 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
558 replaced by hw.bus.devctl_queue=0. The default for this new tunable
562 Remove the option STOP_NMI. The default action is now to use NMI only
563 for KDB via the newly introduced function stop_cpus_hard() and
564 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
567 The stable/8 branch created in subversion. This corresponds to the
568 RELENG_8 branch in CVS.
571 Bump the shared library version numbers for all libraries that do not
572 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
573 __FreeBSD_version to 800105.
576 Due to changes in the implementation of virtual network stack support,
577 all network-related kernel modules must be recompiled. As this change
578 breaks the ABI, bump __FreeBSD_version to 800104.
581 The TOE interface to the TCP syncache has been modified to remove
582 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
583 The cxgb driver is the only TOE consumer affected by this change, and
584 needs to be recompiled along with the kernel. As this change breaks
585 the ABI, bump __FreeBSD_version to 800103.
588 Padding has been added to struct tcpcb, sackhint and tcpstat in
589 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
590 maintaining the ABI. However, this change breaks the ABI, so bump
591 __FreeBSD_version to 800102. User space tools that rely on the size of
592 any of these structs (e.g. sockstat) need to be recompiled.
595 The NFS_LEGACYRPC option has been removed along with the old kernel
596 RPC implementation that this option selected. Kernel configurations
597 may need to be adjusted.
600 The network interface device nodes at /dev/net/<interface> have been
601 removed. All ioctl operations can be performed the normal way using
602 routing sockets. The kqueue functionality can generally be replaced
603 with routing sockets.
606 The documentation from the FreeBSD Documentation Project (Handbook,
607 FAQ, etc.) is now installed via packages by sysinstall(8) and under
608 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
611 The ABI of various structures related to the SYSV IPC API have been
612 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
613 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
617 Layout of struct vnet has changed as routing related variables were
618 moved to their own Vimage module. Modules need to be recompiled. Bump
619 __FreeBSD_version to 800099.
622 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
623 respectively. As long as no more than 16 groups per process are used,
624 no changes should be visible. When more than 16 groups are used, old
625 binaries may fail if they call getgroups() or getgrouplist() with
626 statically sized storage. Recompiling will work around this, but
627 applications should be modified to use dynamically allocated storage
628 for group arrays as POSIX.1-2008 does not cap an implementation's
629 number of supported groups at NGROUPS_MAX+1 as previous versions did.
631 NFS and portalfs mounts may also be affected as the list of groups is
632 truncated to 16. Users of NFS who use more than 16 groups, should
633 take care that negative group permissions are not used on the exported
634 file systems as they will not be reliable unless a GSSAPI based
635 authentication method is used.
638 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
639 option compiles in the support for adaptive spinning for lockmgrs
640 which want to enable it. The lockinit() function now accepts the flag
641 LK_ADAPTIVE in order to make the lock object subject to adaptive
642 spinning when both held in write and read mode.
645 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
646 changed. User applications that use this ioctl need to be rebuilt.
649 The layout of struct thread has changed. Kernel and modules need to
653 The layout of structs ifnet, domain, protosw and vnet_net has changed.
654 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
657 window(1) has been removed from the base system. It can now be
658 installed from ports. The port is called misc/window.
661 The way we are storing and accessing `routing table' entries has
662 changed. Programs reading the FIB, like netstat, need to be
666 A new netisr implementation has been added for FreeBSD 8. Network
667 file system modules, such as igmp, ipdivert, and others, should be
669 Bump __FreeBSD_version to 800096.
672 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
676 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
677 Bump __FreeBSD_version to 800094.
680 Add mnt_xflag field to 'struct mount'. File system modules need to be
682 Bump __FreeBSD_version to 800093.
685 The compiling option ADAPTIVE_SX has been retired while it has been
686 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
687 The KPI for sx_init_flags() changes as accepting flags:
688 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
689 been introduced in order to handle the reversed logic.
690 Bump __FreeBSD_version to 800092.
693 Add support for hierarchical jails. Remove global securelevel.
694 Bump __FreeBSD_version to 800091.
697 The layout of struct vnet_net has changed, therefore modules
699 Bump __FreeBSD_version to 800090.
702 The newly imported zic(8) produces a new format in the output. Please
703 run tzsetup(8) to install the newly created data to /etc/localtime.
706 The sysctl tree for the usb stack has renamed from hw.usb2.* to
707 hw.usb.* and is now consistent again with previous releases.
710 802.11 monitor mode support was revised and driver api's were changed.
711 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
712 of DLT_IEEE802_11. No user-visible data structures were changed but
713 applications that use DLT_IEEE802_11 may require changes.
714 Bump __FreeBSD_version to 800088.
717 The layout of the following structs has changed: sysctl_oid,
718 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
719 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
720 panics may be experienced. World rebuild is required for
721 correctly checking networking state from userland.
722 Bump __FreeBSD_version to 800085.
725 MLDv2 and Source-Specific Multicast (SSM) have been merged
726 to the IPv6 stack. VIMAGE hooks are in but not yet used.
727 The implementation of SSM within FreeBSD's IPv6 stack closely
728 follows the IPv4 implementation.
730 For kernel developers:
732 * The most important changes are that the ip6_output() and
733 ip6_input() paths no longer take the IN6_MULTI_LOCK,
734 and this lock has been downgraded to a non-recursive mutex.
736 * As with the changes to the IPv4 stack to support SSM, filtering
737 of inbound multicast traffic must now be performed by transport
738 protocols within the IPv6 stack. This does not apply to TCP and
739 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
741 * The KPIs used by IPv6 multicast are similar to those used by
742 the IPv4 stack, with the following differences:
743 * im6o_mc_filter() is analogous to imo_multicast_filter().
744 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
745 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
746 * IN6_LOOKUP_MULTI() has been deprecated and removed.
747 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
748 for MLDv1 have an additional 'timer' argument which is used to
749 jitter the initial membership report for the solicited-node
750 multicast membership on-link.
751 * This is not strictly needed for MLDv2, which already jitters
752 its report transmissions. However, the 'timer' argument is
753 preserved in case MLDv1 is active on the interface.
755 * The KAME linked-list based IPv6 membership implementation has
756 been refactored to use a vector similar to that used by the IPv4
758 Code which maintains a list of its own multicast memberships
759 internally, e.g. carp, has been updated to reflect the new
762 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
763 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
764 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
765 implementation constraint which needs to be addressed in HEAD.
767 For application developers:
769 * The changes are broadly similar to those made for the IPv4
772 * The use of IPv4 and IPv6 multicast socket options on the same
773 socket, using mapped addresses, HAS NOT been tested or supported.
775 * There are a number of issues with the implementation of various
776 IPv6 multicast APIs which need to be resolved in the API surface
777 before the implementation is fully compatible with KAME userland
778 use, and these are mostly to do with interface index treatment.
780 * The literature available discusses the use of either the delta / ASM
781 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
782 using setsourcefilter(3)/getsourcefilter(3). For more information
783 please refer to RFC 3768, 'Socket Interface Extensions for
784 Multicast Source Filters'.
786 * Applications which use the published RFC 3678 APIs should be fine.
788 For systems administrators:
790 * The mtest(8) utility has been refactored to support IPv6, in
791 addition to IPv4. Interface addresses are no longer accepted
792 as arguments, their names must be used instead. The utility
793 will map the interface name to its first IPv4 address as
794 returned by getifaddrs(3).
796 * The ifmcstat(8) utility has also been updated to print the MLDv2
797 endpoint state and source filter lists via sysctl(3).
799 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
800 loopback of IPv6 multicast datagrams by default; it defaults to 1
801 to preserve the existing behaviour. Disabling multicast loopback is
802 recommended for optimal system performance.
804 * The IPv6 MROUTING code has been changed to examine this sysctl
805 instead of attempting to perform a group lookup before looping
806 back forwarded datagrams.
808 Bump __FreeBSD_version to 800084.
811 Implement low-level Bluetooth HCI API.
812 Bump __FreeBSD_version to 800083.
815 The layout of struct malloc_type, used by modules to register new
816 memory allocation types, has changed. Most modules will need to
817 be rebuilt or panics may be experienced.
818 Bump __FreeBSD_version to 800081.
821 Anticipate overflowing inp_flags - add inp_flags2.
822 This changes most offsets in inpcb, so checking v4 connection
823 state will require a world rebuild.
824 Bump __FreeBSD_version to 800080.
827 Add an llentry to struct route and struct route_in6. Modules
828 embedding a struct route will need to be recompiled.
829 Bump __FreeBSD_version to 800079.
832 The size of rt_metrics_lite and by extension rtentry has changed.
833 Networking administration apps will need to be recompiled.
834 The route command now supports show as an alias for get, weighting
835 of routes, sticky and nostick flags to alter the behavior of stateful
837 Bump __FreeBSD_version to 800078.
840 Do not use Giant for kbdmux(4) locking. This is wrong and
841 apparently causing more problems than it solves. This will
842 re-open the issue where interrupt handlers may race with
843 kbdmux(4) in polling mode. Typical symptoms include (but
844 not limited to) duplicated and/or missing characters when
845 low level console functions (such as gets) are used while
846 interrupts are enabled (for example geli password prompt,
847 mountroot prompt etc.). Disabling kbdmux(4) may help.
850 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
851 kernel modules referencing any of the above need to be recompiled.
852 Bump __FreeBSD_version to 800075.
855 GEOM_PART has become the default partition slicer for storage devices,
856 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
857 introduces some changes:
859 MSDOS/EBR: the devices created from MSDOS extended partition entries
860 (EBR) can be named differently than with GEOM_MBR and are now symlinks
861 to devices with offset-based names. fstabs may need to be modified.
863 BSD: the "geometry does not match label" warning is harmless in most
864 cases but it points to problems in file system misalignment with
865 disk geometry. The "c" partition is now implicit, covers the whole
866 top-level drive and cannot be (mis)used by users.
868 General: Kernel dumps are now not allowed to be written to devices
869 whose partition types indicate they are meant to be used for file
870 systems (or, in case of MSDOS partitions, as something else than
873 Most of these changes date approximately from 200812.
876 The uscanner(4) driver has been removed from the kernel. This follows
877 Linux removing theirs in 2.6 and making libusb the default interface
881 The multicast forwarding code has been cleaned up. netstat(1)
882 only relies on KVM now for printing bandwidth upcall meters.
883 The IPv4 and IPv6 modules are split into ip_mroute_mod and
884 ip6_mroute_mod respectively. The config(5) options for statically
885 compiling this code remain the same, i.e. 'options MROUTING'.
888 Support for the IFF_NEEDSGIANT network interface flag has been
889 removed, which means that non-MPSAFE network device drivers are no
890 longer supported. In particular, if_ar, if_sr, and network device
891 drivers from the old (legacy) USB stack can no longer be built or
895 POSIX.1 Native Language Support (NLS) has been enabled in libc and
896 a bunch of new language catalog files have also been added.
897 This means that some common libc messages are now localized and
898 they depend on the LC_MESSAGES environmental variable.
901 The k8temp(4) driver has been renamed to amdtemp(4) since
902 support for Family 10 and Family 11 CPU families was added.
905 IGMPv3 and Source-Specific Multicast (SSM) have been merged
906 to the IPv4 stack. VIMAGE hooks are in but not yet used.
908 For kernel developers, the most important changes are that the
909 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
910 and this lock has been downgraded to a non-recursive mutex.
912 Transport protocols (UDP, Raw IP) are now responsible for filtering
913 inbound multicast traffic according to group membership and source
914 filters. The imo_multicast_filter() KPI exists for this purpose.
915 Transports which do not use multicast (SCTP, TCP) already reject
916 multicast by default. Forwarding and receive performance may improve
917 as a mutex acquisition is no longer needed in the ip_input()
918 low-level input path. in_addmulti() and in_delmulti() are shimmed
919 to new KPIs which exist to support SSM in-kernel.
921 For application developers, it is recommended that loopback of
922 multicast datagrams be disabled for best performance, as this
923 will still cause the lock to be taken for each looped-back
924 datagram transmission. The net.inet.ip.mcast.loop sysctl may
925 be tuned to 0 to disable loopback by default; it defaults to 1
926 to preserve the existing behaviour.
928 For systems administrators, to obtain best performance with
929 multicast reception and multiple groups, it is always recommended
930 that a card with a suitably precise hash filter is used. Hash
931 collisions will still result in the lock being taken within the
932 transport protocol input path to check group membership.
934 If deploying FreeBSD in an environment with IGMP snooping switches,
935 it is recommended that the net.inet.igmp.sendlocal sysctl remain
936 enabled; this forces 224.0.0.0/24 group membership to be announced
939 The size of 'struct igmpstat' has changed; netstat needs to be
940 recompiled to reflect this.
941 Bump __FreeBSD_version to 800070.
944 libusb20.so.1 is now installed as libusb.so.1 and the ports system
945 updated to use it. This requires a buildworld/installworld in order to
946 update the library and dependencies (usbconfig, etc). Its advisable to
947 rebuild all ports which uses libusb. More specific directions are given
948 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
949 libusb are no longer required and can be removed.
952 A workaround is committed to allow the creation of System V shared
953 memory segment of size > 2 GB on the 64-bit architectures.
954 Due to a limitation of the existing ABI, the shm_segsz member
955 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
956 wrong for large segments. Note that limits must be explicitly
957 raised to allow such segments to be created.
960 The layout of struct ifnet has changed, requiring a rebuild of all
961 network device driver modules.
964 The /dev handling for the new USB stack has changed, a
965 buildworld/installworld is required for libusb20.
968 The new USB2 stack has now been permanently moved in and all kernel and
969 module names reverted to their previous values (eg, usb, ehci, ohci,
970 ums, ...). The old usb stack can be compiled in by prefixing the name
971 with the letter 'o', the old usb modules have been removed.
972 Updating entry 20090216 for xorg and 20090215 for libmap may still
976 The rc.conf(5) option if_up_delay has been renamed to
977 defaultroute_delay to better reflect its purpose. If you have
978 customized this setting in /etc/rc.conf you need to update it to
982 xorg 7.4 wants to configure its input devices via hald which does not
983 yet work with USB2. If the keyboard/mouse does not work in xorg then
985 Option "AllowEmptyInput" "off"
986 to your ServerLayout section. This will cause X to use the configured
987 kbd and mouse sections from your xorg.conf.
990 The GENERIC kernels for all architectures now default to the new USB2
991 stack. No kernel config options or code have been removed so if a
992 problem arises please report it and optionally revert to the old USB
993 stack. If you are loading USB kernel modules or have a custom kernel
994 that includes GENERIC then ensure that usb names are also changed over,
995 eg uftdi -> usb2_serial_ftdi.
997 Older programs linked against the ports libusb 0.1 need to be
998 redirected to the new stack's libusb20. /etc/libmap.conf can
1000 # Map old usb library to new one for usb2 stack
1001 libusb-0.1.so.8 libusb20.so.1
1004 All USB ethernet devices now attach as interfaces under the name ueN
1005 (eg. ue0). This is to provide a predictable name as vendors often
1006 change usb chipsets in a product without notice.
1009 The ichsmb(4) driver has been changed to require SMBus slave
1010 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1011 All of the other SMBus controller drivers require left-justified
1012 slave addresses, so this change makes all the drivers provide the
1016 INET6 statistics (struct ip6stat) was updated.
1017 netstat(1) needs to be recompiled.
1020 NTFS has been removed from GENERIC kernel on amd64 to match
1021 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1022 will load ntfs.ko module automatically when NTFS support is
1023 actually needed, unless ntfs.ko is not installed or security
1024 level prohibits loading kernel modules. If either is the case,
1025 "options NTFS" has to be added into kernel config.
1028 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1029 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1030 800061. User space tools that rely on the size of struct tcpcb in
1031 tcp_var.h (e.g. sockstat) need to be recompiled.
1034 ng_tty(4) module updated to match the new TTY subsystem.
1035 Due to API change, user-level applications must be updated.
1036 New API support added to mpd5 CVS and expected to be present
1037 in next mpd5.3 release.
1040 With __FreeBSD_version 800060 the makefs tool is part of
1041 the base system (it was a port).
1044 The afdata and ifnet locks have been changed from mutexes to
1045 rwlocks, network modules will need to be re-compiled.
1048 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1049 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1050 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1051 architecture and 40 bytes on 64-bit architecture. The userland
1052 applications "arp" and "ndp" have been updated accordingly.
1053 The output from "netstat -r" shows only routing entries and
1054 none of the L2 information.
1057 __FreeBSD_version 800057 marks the switchover from the
1058 binary ath hal to source code. Users must add the line:
1060 options AH_SUPPORT_AR5416
1062 to their kernel config files when specifying:
1066 The ath_hal module no longer exists; the code is now compiled
1067 together with the driver in the ath module. It is now
1068 possible to tailor chip support (i.e. reduce the set of chips
1069 and thereby the code size); consult ath_hal(4) for details.
1072 __FreeBSD_version 800054 adds memory barriers to
1073 <machine/atomic.h>, new interfaces to ifnet to facilitate
1074 multiple hardware transmit queues for cards that support
1075 them, and a lock-less ring-buffer implementation to
1076 enable drivers to more efficiently manage queueing of
1080 A new version of ZFS (version 13) has been merged to -HEAD.
1081 This version has zpool attribute "listsnapshots" off by
1082 default, which means "zfs list" does not show snapshots,
1083 and is the same as Solaris behavior.
1086 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1089 The uhci, ohci, ehci and slhci USB Host controller drivers have
1090 been put into separate modules. If you load the usb module
1091 separately through loader.conf you will need to load the
1092 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1093 controller add the following to loader.conf:
1099 The ABI used by the PMC toolset has changed. Please keep
1100 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1104 atapci kernel module now includes only generic PCI ATA
1105 driver. AHCI driver moved to ataahci kernel module.
1106 All vendor-specific code moved into separate kernel modules:
1107 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1108 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1109 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1110 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1113 The TTY subsystem of the kernel has been replaced by a new
1114 implementation, which provides better scalability and an
1115 improved driver model. Most common drivers have been migrated to
1116 the new TTY subsystem, while others have not. The following
1117 drivers have not yet been ported to the new TTY layer:
1120 cy, digi, rc, rp, sio
1126 ng_h4, ng_tty, ppp, sl, snp
1128 Adding these drivers to your kernel configuration file shall
1129 cause compilation to fail.
1132 ntpd has been upgraded to 4.2.4p5.
1135 OpenSSH has been upgraded to 5.1p1.
1137 For many years, FreeBSD's version of OpenSSH preferred DSA
1138 over RSA for host and user authentication keys. With this
1139 upgrade, we've switched to the vendor's default of RSA over
1140 DSA. This may cause upgraded clients to warn about unknown
1141 host keys even for previously known hosts. Users should
1142 follow the usual procedure for verifying host keys before
1143 accepting the RSA key.
1145 This can be circumvented by setting the "HostKeyAlgorithms"
1146 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1149 Please note that the sequence of keys offered for
1150 authentication has been changed as well. You may want to
1151 specify IdentityFile in a different order to revert this
1155 The sio(4) driver has been removed from the i386 and amd64
1156 kernel configuration files. This means uart(4) is now the
1157 default serial port driver on those platforms as well.
1159 To prevent collisions with the sio(4) driver, the uart(4) driver
1160 uses different names for its device nodes. This means the
1161 onboard serial port will now most likely be called "ttyu0"
1162 instead of "ttyd0". You may need to reconfigure applications to
1163 use the new device names.
1165 When using the serial port as a boot console, be sure to update
1166 /boot/device.hints and /etc/ttys before booting the new kernel.
1167 If you forget to do so, you can still manually specify the hints
1168 at the loader prompt:
1170 set hint.uart.0.at="isa"
1171 set hint.uart.0.port="0x3F8"
1172 set hint.uart.0.flags="0x10"
1173 set hint.uart.0.irq="4"
1177 The gpt(8) utility has been removed. Use gpart(8) to partition
1181 The version that Linuxulator emulates was changed from 2.4.2
1182 to 2.6.16. If you experience any problems with Linux binaries
1183 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1184 if it fixes the problem contact emulation mailing list.
1187 ISDN4BSD (I4B) was removed from the src tree. You may need to
1188 update a your kernel configuration and remove relevant entries.
1191 I have checked in code to support multiple routing tables.
1192 See the man pages setfib(1) and setfib(2).
1193 This is a hopefully backwards compatible version,
1194 but to make use of it you need to compile your kernel
1195 with options ROUTETABLES=2 (or more up to 16).
1198 The 802.11 wireless support was redone to enable multi-bss
1199 operation on devices that are capable. The underlying device
1200 is no longer used directly but instead wlanX devices are
1201 cloned with ifconfig. This requires changes to rc.conf files.
1202 For example, change:
1203 ifconfig_ath0="WPA DHCP"
1206 ifconfig_wlan0="WPA DHCP"
1207 see rc.conf(5) for more details. In addition, mergemaster of
1208 /etc/rc.d is highly recommended. Simultaneous update of userland
1209 and kernel wouldn't hurt either.
1211 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1212 modules were merged into the base wlan module. All references
1213 to these modules (e.g. in kernel config files) must be removed.
1216 psm(4) has gained write(2) support in native operation level.
1217 Arbitrary commands can be written to /dev/psm%d and status can
1218 be read back from it. Therefore, an application is responsible
1219 for status validation and error recovery. It is a no-op in
1220 other operation levels.
1223 Support for KSE threading has been removed from the kernel. To
1224 run legacy applications linked against KSE libmap.conf may
1225 be used. The following libmap.conf may be used to ensure
1226 compatibility with any prior release:
1228 libpthread.so.1 libthr.so.1
1229 libpthread.so.2 libthr.so.2
1230 libkse.so.3 libthr.so.3
1233 The layout of struct vmspace has changed. This affects libkvm
1234 and any executables that link against libkvm and use the
1235 kvm_getprocs() function. In particular, but not exclusively,
1236 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1237 The effects are minimal, but it's advisable to upgrade world
1241 The latest em driver no longer has support in it for the
1242 82575 adapter, this is now moved to the igb driver. The
1243 split was done to make new features that are incompatible
1244 with older hardware easier to do.
1247 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1248 likewise the kernel option is now GEOM_LINUX_LVM.
1251 The default NFS mount mode has changed from UDP to TCP for
1252 increased reliability. If you rely on (insecurely) NFS
1253 mounting across a firewall you may need to update your
1257 Belatedly note the addition of m_collapse for compacting
1261 The fts(3) structures have been changed to use adequate
1262 integer types for their members and so to be able to cope
1263 with huge file trees. The old fts(3) ABI is preserved
1264 through symbol versioning in libc, so third-party binaries
1265 using fts(3) should still work, although they will not take
1266 advantage of the extended types. At the same time, some
1267 third-party software might fail to build after this change
1268 due to unportable assumptions made in its source code about
1269 fts(3) structure members. Such software should be fixed
1270 by its vendor or, in the worst case, in the ports tree.
1271 FreeBSD_version 800015 marks this change for the unlikely
1272 case that a portable fix is impossible.
1275 To upgrade to -current after this date, you must be running
1276 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1277 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1280 The ADAPTIVE_GIANT kernel option has been retired because its
1281 functionality is the default now.
1284 The AT keyboard emulation of sunkbd(4) has been turned on
1285 by default. In order to make the special symbols of the Sun
1286 keyboards driven by sunkbd(4) work under X these now have
1287 to be configured the same way as Sun USB keyboards driven
1288 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1290 Option "XkbLayout" "us"
1291 Option "XkbRules" "xorg"
1292 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1295 It has been decided that it is desirable to provide ABI
1296 backwards compatibility to the FreeBSD 4/5/6 versions of the
1297 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1298 broken with the introduction of PCI domain support (see the
1299 20070930 entry). Unfortunately, this required the ABI of
1300 PCIOCGETCONF to be broken again in order to be able to
1301 provide backwards compatibility to the old version of that
1302 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1303 again. As for prominent ports this affects neither pciutils
1304 nor xorg-server this time, the hal port needs to be rebuilt
1308 The misnamed kthread_create() and friends have been renamed
1309 to kproc_create() etc. Many of the callers already
1310 used kproc_start()..
1311 I will return kthread_create() and friends in a while
1312 with implementations that actually create threads, not procs.
1313 Renaming corresponds with version 800002.
1322 Avoid using make -j when upgrading. While generally safe, there are
1323 sometimes problems using -j to upgrade. If your upgrade fails with
1324 -j, please try again without -j. From time to time in the past there
1325 have been problems using -j with buildworld and/or installworld. This
1326 is especially true when upgrading between "distant" versions (eg one
1327 that cross a major release boundary or several minor releases, or when
1328 several months have passed on the -current branch).
1330 Sometimes, obscure build problems are the result of environment
1331 poisoning. This can happen because the make utility reads its
1332 environment when searching for values for global variables. To run
1333 your build attempts in an "environmental clean room", prefix all make
1334 commands with 'env -i '. See the env(1) manual page for more details.
1336 When upgrading from one major version to another it is generally best
1337 to upgrade to the latest code in the currently installed branch first,
1338 then do an upgrade to the new branch. This is the best-tested upgrade
1339 path, and has the highest probability of being successful. Please try
1340 this approach before reporting problems with a major version upgrade.
1344 When upgrading the boot ZFS pool to a new version, always follow
1347 1.) recompile and reinstall the ZFS boot loader and boot block
1348 (this is part of "make buildworld" and "make installworld")
1350 2.) update the ZFS boot block on your boot drive
1352 The following example updates the ZFS boot block on the first
1353 partition (freebsd-boot) of a GPT partitioned drive ad0:
1354 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1356 Non-boot pools do not need these updates.
1360 If you are updating from a prior version of FreeBSD (even one just
1361 a few days old), you should follow this procedure. It is the most
1362 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1364 make kernel-toolchain
1365 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1366 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1368 To test a kernel once
1369 ---------------------
1370 If you just want to boot a kernel once (because you are not sure
1371 if it works, or if you want to boot a known bad kernel to provide
1372 debugging information) run
1373 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1374 nextboot -k testkernel
1376 To just build a kernel when you know that it won't mess you up
1377 --------------------------------------------------------------
1378 This assumes you are already running a CURRENT system. Replace
1379 ${arch} with the architecture of your machine (e.g. "i386",
1380 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1382 cd src/sys/${arch}/conf
1383 config KERNEL_NAME_HERE
1384 cd ../compile/KERNEL_NAME_HERE
1389 If this fails, go to the "To build a kernel" section.
1391 To rebuild everything and install it on the current system.
1392 -----------------------------------------------------------
1393 # Note: sometimes if you are running current you gotta do more than
1394 # is listed here if you are upgrading from a really old current.
1396 <make sure you have good level 0 dumps>
1398 make kernel KERNCONF=YOUR_KERNEL_HERE
1400 <reboot in single user> [3]
1408 To cross-install current onto a separate partition
1409 --------------------------------------------------
1410 # In this approach we use a separate partition to hold
1411 # current's root, 'usr', and 'var' directories. A partition
1412 # holding "/", "/usr" and "/var" should be about 2GB in
1415 <make sure you have good level 0 dumps>
1418 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1419 <maybe newfs current's root partition>
1420 <mount current's root partition on directory ${CURRENT_ROOT}>
1421 make installworld DESTDIR=${CURRENT_ROOT}
1422 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1423 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1424 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1425 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1426 <reboot into current>
1427 <do a "native" rebuild/install as described in the previous section>
1428 <maybe install compatibility libraries from ports/misc/compat*>
1432 To upgrade in-place from 8.x-stable to current
1433 ----------------------------------------------
1434 <make sure you have good level 0 dumps>
1436 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1438 <reboot in single user> [3]
1445 Make sure that you've read the UPDATING file to understand the
1446 tweaks to various things you need. At this point in the life
1447 cycle of current, things change often and you are on your own
1448 to cope. The defaults can also change, so please read ALL of
1449 the UPDATING entries.
1451 Also, if you are tracking -current, you must be subscribed to
1452 freebsd-current@freebsd.org. Make sure that before you update
1453 your sources that you have read and understood all the recent
1454 messages there. If in doubt, please track -stable which has
1455 much fewer pitfalls.
1457 [1] If you have third party modules, such as vmware, you
1458 should disable them at this point so they don't crash your
1461 [3] From the bootblocks, boot -s, and then do
1466 adjkerntz -i # if CMOS is wall time
1467 Also, when doing a major release upgrade, it is required that
1468 you boot into single user mode to do the installworld.
1470 [4] Note: This step is non-optional. Failure to do this step
1471 can result in a significant reduction in the functionality of the
1472 system. Attempting to do it by hand is not recommended and those
1473 that pursue this avenue should read this file carefully, as well
1474 as the archives of freebsd-current and freebsd-hackers mailing lists
1475 for potential gotchas. The -U option is also useful to consider.
1476 See mergemaster(8) for more information.
1478 [5] Usually this step is a noop. However, from time to time
1479 you may need to do this if you get unknown user in the following
1480 step. It never hurts to do it all the time. You may need to
1481 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1482 install) after the buildworld before this step if you last updated
1483 from current before 20020224 or from -stable before 20020408.
1485 [6] This only deletes old files and directories. Old libraries
1486 can be deleted by "make delete-old-libs", but you have to make
1487 sure that no program is using those libraries anymore.
1489 [8] In order to have a kernel that can run the 4.x binaries needed to
1490 do an installworld, you must include the COMPAT_FREEBSD4 option in
1491 your kernel. Failure to do so may leave you with a system that is
1492 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1493 required to run the 5.x binaries on more recent kernels. And so on
1494 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1496 Make sure that you merge any new devices from GENERIC since the
1497 last time you updated your kernel config file.
1499 [9] When checking out sources, you must include the -P flag to have
1500 cvs prune empty directories.
1502 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1503 "?=" instead of the "=" assignment operator, so that buildworld can
1504 override the CPUTYPE if it needs to.
1506 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1507 not on the command line, or in /etc/make.conf. buildworld will
1508 warn if it is improperly defined.
1511 This file contains a list, in reverse chronological order, of major
1512 breakages in tracking -current. It is not guaranteed to be a complete
1513 list of such breakages, and only contains entries since October 10, 2007.
1514 If you need to see UPDATING entries from before that date, you will need
1515 to fetch an UPDATING file from an older FreeBSD release.
1517 Copyright information:
1519 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1521 Redistribution, publication, translation and use, with or without
1522 modification, in full or in part, in any form or format of this
1523 document are permitted without further permission from the author.
1525 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1526 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1527 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1528 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1529 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1530 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1531 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1532 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1533 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1534 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1535 POSSIBILITY OF SUCH DAMAGE.
1537 Contact Warner Losh if you have any questions about your use of