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.
14 20150612: p16 FreeBSD-SA-15:10.openssl
15 Fix multiple vulnerabilities in OpenSSL. [SA-15:10]
17 20150609: p15 FreeBSD-EN-15:06.file
19 Updated base system file(1) to 5.22 to address multiple denial
22 20150513: p14 FreeBSD-EN-15:04.freebsd-update
24 Fix bug with freebsd-update(8) that does not ensure the previous
25 upgrade was completed. [EN-15:04]
27 20150407: p13 FreeBSD-SA-15:04.igmp [revised]
31 Improved patch for SA-15:04.igmp.
33 Fix multiple vulnerabilities of ntp. [SA-15:07]
35 Fix Denial of Service with IPv6 Router Advertisements. [SA-15:09]
38 Fix patch for SA-15:06.openssl.
40 20150319: p11 FreeBSD-SA-15:06.openssl
41 Fix multiple vulnerabilities in OpenSSL. [SA-15:06]
43 20150225: p10 FreeBSD-SA-15:04.igmp
46 FreeBSD-EN-15:02.openssl
47 FreeBSD-EN-15:03.freebsd-update
49 Fix integer overflow in IGMP protocol. [SA-15:04]
51 Fix BIND remote denial of service vulnerability. [SA-15:05]
53 Fix vt(4) crash with improper ioctl parameters. [EN-15:01]
55 Updated base system OpenSSL to 0.9.8zd. [EN-15:02]
57 Fix freebsd-update libraries update ordering issue. [EN-15:03]
59 20150127: p9 FreeBSD-SA-15:02.kmem
62 Fix SCTP SCTP_SS_VALUE kernel memory corruption and disclosure
63 vulnerability. [SA-15:02]
65 Fix SCTP stream reset vulnerability. [SA-15:03]
67 20150114: p8 FreeBSD-SA-15:01.openssl
68 Fix multiple vulnerabilities in OpenSSL. [SA-15:01]
70 20141223: p7 FreeBSD-SA-14:31.ntp
71 FreeBSD-EN-14:13.freebsd-update
73 Fix multiple vulnerabilities in NTP suite. [SA-14:31]
74 Fix directory deletion issue in freebsd-update. [EN-14:13]
76 20141210: p6 FreeBSD-SA-14:28.file
79 Fix multiple vulnerabilities in file(1) and libmagic(3).
82 Fix BIND remote denial of service vulnerability. [SA-14:29]
84 20141104: p5 FreeBSD-SA-14:25.setlogin
88 Fix kernel stack disclosure in setlogin(2) / getlogin(2).
91 Fix remote command execution in ftp(1). [SA-14:26]
93 Fix NFSv4 and ZFS cache consistency issue. [EN-14:12]
95 20141022: p4 FreeBSD-EN-14:10.tzdata
96 FreeBSD-EN-14:11.crypt
98 Time zone data file update. [EN-14:10]
100 Change crypt(3) default hashing algorithm back to DES. [EN-14:11]
102 20141021: p3 FreeBSD-SA-14:20.rtsold
103 FreeBSD-SA-14:21.routed
104 FreeBSD-SA-14:22.namei
105 FreeBSD-SA-14:23.openssl
107 Fix rtsold(8) remote buffer overflow vulnerability. [SA-14:20]
109 Fix routed(8) remote denial of service vulnerability. [SA-14:21]
111 Fix memory leak in sandboxed namei lookup. [SA-14:22]
113 Fix OpenSSL multiple vulnerabilities. [SA-14:23]
115 20140916: p2 FreeBSD-SA-14:19.tcp
116 Fix Denial of Service in TCP packet processing. [SA-14:19]
118 20140909: p1 FreeBSD-SA-14:18.openssl
119 Fix OpenSSL multiple vulnerabilities. [SA-14:18]
125 On i386 and amd64 systems, the onifconsole flag is now set by default
126 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
127 as a login TTY if it is set in the bootloader as an active kernel
128 console. No changes in behavior should result otherwise. To revert to
129 the previous behavior, set ttyu0 to "off" in /etc/ttys.
132 Clang and llvm have been upgraded to 3.4.1 release.
135 Clang and llvm have been upgraded to 3.4 release.
138 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
139 been deprecated and will not be part of FreeBSD 11.0 and later
140 releases. If you use this driver, please consider switching to
141 the nfe(4) driver instead.
144 The behavior of gss_pseudo_random() for the krb5 mechanism
145 has changed, for applications requesting a longer random string
146 than produced by the underlying enctype's pseudo-random() function.
147 In particular, the random string produced from a session key of
148 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
149 be different at the 17th octet and later, after this change.
150 The counter used in the PRF+ construction is now encoded as a
151 big-endian integer in accordance with RFC 4402.
152 __FreeBSD_version is bumped to 902505.
158 Behavior of devfs rules path matching has been changed.
159 Pattern is now always matched against fully qualified devfs
160 path and slash characters must be explicitly matched by
161 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
162 subdirectories must be reviewed.
165 hastctl(8)'s `status' command output changed to terse one-liner format.
166 Scripts using this should switch to `list' command or be rewritten.
169 Fix a bug that allowed a tracing process (e.g. gdb) to write
170 to a memory-mapped file in the traced process's address space
171 even if neither the traced process nor the tracing process had
172 write access to that file.
175 Added ZFS TRIM support which is enabled by default. To disable
176 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
178 Creating new ZFS pools and adding new devices to existing pools
179 first performs a full device level TRIM which can take a significant
180 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
181 to disable this behaviour.
183 ZFS TRIM requires the underlying device support BIO_DELETE which
184 is currently provided by methods such as ATA TRIM and SCSI UNMAP
185 via CAM, which are typically supported by SSD's.
187 Stats for ZFS TRIM can be monitored by looking at the sysctl's
188 under kstat.zfs.misc.zio_trim.
191 `list' command has been added to hastctl(8). For now, it is full
192 equivalent of `status' command.
193 WARNING: in the near future the output of hastctl's status command
194 will change to more terse format. If you use `hastctl status'
195 for parsing in your scripts, switch to `hastctl list'.
198 The mergemaster command now uses the default MAKEOBJDIRPREFIX
199 rather than creating it's own in the temporary directory in
200 order allow access to bootstrapped versions of tools such as
201 install and mtree. When upgrading from version of FreeBSD where
202 the install command does not support -l, you will need to
203 install a new mergemaster command if mergemaster -p is required.
204 This can be accomplished with the command (cd src/usr.sbin/mergemaster
207 Due to the use of the new -l option to install(1) during build
208 and install, you must take care not to directly set the INSTALL
209 make variable in your /etc/make.conf, /etc/src.conf, or on the
210 command line. If you wish to use the -C flag for all installs
211 you may be able to add INSTALL+=-C to /etc/make.conf or
215 Fix a bug that allows NFS clients to issue READDIR on files.
218 The install(1) option -M has changed meaning and now takes an
219 argument that is a file or path to append logs to. In the
220 unlikely event that -M was the last option on the command line
221 and the command line contained at least two files and a target
222 directory the first file will have logs appended to it. The -M
223 option served little practical purpose in the last decade so it's
224 used expected to be extremely rare.
227 A new compression method (lz4) has been merged to. Please refer to
228 zpool-features(7) for more information.
230 Please refer to the "ZFS notes" section of this file for information
231 on upgrading boot ZFS pools.
234 The VFS KBI was changed with the merge of several nullfs
235 optimizations and fixes. All filesystem modules must be
239 With the addition of auditdistd(8), a new auditdistd user is now
240 depended on during installworld. "mergemaster -p" can be used to add
241 the user prior to installworld, as documented in the handbook.
247 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
248 Starting with this version the old system of ZFS pool versioning
249 is superseded by "feature flags". This concept enables forward
250 compatibility against certain future changes in functionality of ZFS
251 pools. The first two read-only compatible "feature flags" for ZFS
252 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
253 For more information read the new zpool-features(7) manual page.
254 Please refer to the "ZFS notes" section of this file for information
255 on upgrading boot ZFS pools.
258 The commit introducing bsd.compiler.mk breaks the traditional
259 building of kernels before this point. Add -m ${SRC}/share/mk
260 (for the right value of SRC) to your command lines to work
261 around; update your useland to a point after this; or use the
262 buildkernel/installkernel top-level targets. See also 20120829.
265 The IPFIREWALL_FORWARD kernel option has been removed. Its
266 functionality now turned on by default.
269 The random(4) support for the VIA hardware random number
270 generator (`PADLOCK') is no longer enabled unconditionally.
271 Add the PADLOCK_RNG option in the custom kernel config if
272 needed. The GENERIC kernels on i386 and amd64 do include the
273 option, so the change only affects the custom kernel
277 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
278 the traditional method, you must update your system with an installworld
279 before the kernel will build. The documented make buildkernel/installkernel
280 interfaces (coupled with fresh make kernel-toolchain) continue to work.
283 The sparc64 ZFS loader has been changed to no longer try to auto-
284 detect ZFS providers based on diskN aliases but now requires these
285 to be explicitly listed in the OFW boot-device environment variable.
288 Now unix domain sockets behave "as expected" on nullfs(5). Previously
289 nullfs(5) did not pass through all behaviours to the underlying layer,
290 as a result if we bound to a socket on the lower layer we could connect
291 only to the lower path; if we bound to the upper layer we could connect
292 only to the upper path. The new behavior is one can connect to both the
293 lower and the upper paths regardless what layer path one binds to.
296 The acpi_wmi(4) status device /dev/wmistat has been renamed to
300 A new VOP_ADVISE() was added to support posix_fadvise(2). All
301 filesystem modules must be recompiled.
304 The interface of the VOP_VPTOCNP(9) changed, now the returned
305 vnode shall be referenced, previously it was required to be
306 only held. All in-tree filesystems are converted.
312 The broken amd(4) driver has been replaced with esp(4) in the amd64,
313 i386 and pc98 GENERIC kernel configuration files.
316 This commit modifies vfs_register() so that it uses a hash
317 calculation to set vfc_typenum, which is enabled by default.
318 The first time a system is booted after this change, the
319 vfc_typenum values will change for all file systems. The
320 main effect of this is a change to the NFS server file handles
321 for file systems that use vfc_typenum in their fsid, such as ZFS.
322 It will, however, prevent vfc_typenum from changing when file
323 systems are loaded in a different order for subsequent reboots.
324 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
325 until you are ready to remount all NFS clients after a reboot.
328 Bump the shared library version numbers for libraries that
329 do not use symbol versioning, have changed the ABI compared
330 to stable/8 and which shared library version was not bumped.
331 Done as part of 9.0-RELEASE cycle.
334 During the merge of Capsicum features, the fget(9) KPI was modified.
335 This may require the rebuilding of out-of-tree device drivers --
336 issues have been reported specifically with the nVidia device driver.
337 __FreeBSD_version is bumped to 900041.
339 Also, there is a period between 20110811 and 20110814 where the
340 special devices /dev/{stdin,stdout,stderr} did not work correctly.
341 Building world from a kernel during that window may not work.
344 The packet filter (pf) code has been updated to OpenBSD 4.5.
345 You need to update userland tools to be in sync with kernel.
346 This update breaks backward compatibility with earlier pfsync(4)
347 versions. Care must be taken when updating redundant firewall setups.
350 The following sysctls and tunables are retired on x86 platforms:
352 machdep.hlt_logical_cpus
353 The following sysctl is retired:
354 machdep.hyperthreading_allowed
355 The sysctls were supposed to provide a way to dynamically offline and
356 online selected CPUs on x86 platforms, but the implementation has not
357 been reliable especially with SCHED_ULE scheduler.
358 machdep.hyperthreading_allowed tunable is still available to ignore
359 hyperthreading CPUs at OS level.
360 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
361 where X is an APIC ID of a CPU. Be advised, though, that disabling
362 CPUs in non-uniform fashion will result in non-uniform topology and
363 may lead to sub-optimal system performance with SCHED_ULE, which is
367 cpumask_t type is retired and cpuset_t is used in order to describe
371 Changes to ifconfig(8) for dynamic address family detection mandate
372 that you are running a kernel of 20110525 or later. Make sure to
373 follow the update procedure to boot a new kernel before installing
377 Support for sun4v architecture is officially dropped
380 Several KPI breaking changes have been committed to the mii(4) layer,
381 the PHY drivers and consequently some Ethernet drivers using mii(4).
382 This means that miibus.ko and the modules of the affected Ethernet
383 drivers need to be recompiled.
385 Note to kernel developers: Given that the OUI bit reversion problem
386 was fixed as part of these changes all mii(4) commits related to OUIs,
387 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
388 handling, no longer can be merged verbatim to stable/8 and previous
392 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
393 into their kernel configurations along with 'device pci'.
396 The default NFS client is now the new NFS client, so fstype "newnfs"
397 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
398 Although mounts via fstype "nfs" will usually work without userland
399 changes, it is recommended that the mount(8) and mount_nfs(8)
400 commands be rebuilt from sources and that a link to mount_nfs called
401 mount_oldnfs be created. The new client is compiled into the
402 kernel with "options NFSCL" and this is needed for diskless root
403 file systems. The GENERIC kernel configs have been changed to use
404 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
405 To use the regular/old client, you can "mount -t oldnfs ...". For
406 a diskless root file system, you must also include a line like:
408 vfs.root.mountfrom="oldnfs:"
410 in the boot/loader.conf on the root fs on the NFS server to make
411 a diskless root fs use the old client.
414 The GENERIC kernels for all architectures now default to the new
415 CAM-based ATA stack. It means that all legacy ATA drivers were
416 removed and replaced by respective CAM drivers. If you are using
417 ATA device names in /etc/fstab or other places, make sure to update
418 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
419 where 'Y's are the sequential numbers starting from zero for each type
420 in order of detection, unless configured otherwise with tunables,
421 see cam(4)). There will be symbolic links created in /dev/ to map
422 old adX devices to the respective adaY. They should provide basic
423 compatibility for file systems mounting in most cases, but they do
424 not support old user-level APIs and do not have respective providers
425 in GEOM. Consider using updated management tools with new device names.
427 It is possible to load devices ahci, ata, siis and mvs as modules,
428 but option ATA_CAM should remain in kernel configuration to make ata
429 module work as CAM driver supporting legacy ATA controllers. Device ata
430 still can be used in modular fashion (atacore + ...). Modules atadisk
431 and atapi* are not used and won't affect operation in ATA_CAM mode.
432 Note that to use CAM-based ATA kernel should include CAM devices
433 scbus, pass, da (or explicitly ada), cd and optionally others. All of
434 them are parts of the cam module.
436 ataraid(4) functionality is now supported by the RAID GEOM class.
437 To use it you can load geom_raid kernel module and use graid(8) tool
438 for management. Instead of /dev/arX device names, use /dev/raid/rX.
440 No kernel config options or code have been removed, so if a problem
441 arises, please report it and optionally revert to the old ATA stack.
442 In order to do it you can remove from the kernel config:
447 , and instead add back:
448 device atadisk # ATA disk drives
449 device ataraid # ATA RAID drives
450 device atapicd # ATAPI CDROM drives
451 device atapifd # ATAPI floppy drives
452 device atapist # ATAPI tape drives
455 The default NFS server has been changed to the new server, which
456 was referred to as the experimental server. If you need to switch
457 back to the old NFS server, you must now put the "-o" option on
458 both the mountd and nfsd commands. This can be done using the
459 mountd_flags and nfs_server_flags rc.conf variables until an
460 update to the rc scripts is committed, which is coming soon.
463 The GNU Objective-C runtime library (libobjc), and other Objective-C
464 related components have been removed from the base system. If you
465 require an Objective-C library, please use one of the available ports.
468 ath(4) has been split into bus- and device- modules. if_ath contains
469 the HAL, the TX rate control and the network device code. if_ath_pci
470 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
471 contains the AHB glue. Users need to load both if_ath_pci and if_ath
472 in order to use ath on everything else.
474 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
475 need to load if_ath and if_ath_pci for ath(4) operation.
478 As part of the replacement of sysinstall, the process of building
479 release media has changed significantly. For details, please re-read
480 release(7), which has been updated to reflect the new build process.
483 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
484 is the last available version under GPLv2. It brings a number of new
485 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
486 4.1 and SSE 4.2), better support for powerpc64, a number of new
487 directives, and lots of other small improvements. See the ChangeLog
488 file in contrib/binutils for the full details.
491 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
492 compliant, and will now use half of hash for authentication.
493 This will break interoperability with all stacks (including all
494 actual FreeBSD versions) who implement
495 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
497 The only workaround with such peers is to use another HMAC
498 algorithm for IPsec ("phase 2") authentication.
501 Remove the uio_yield prototype and symbol. This function has
502 been misnamed since it was introduced and should not be
503 globally exposed with this name. The equivalent functionality
504 is now available using kern_yield(curthread->td_user_pri).
505 The function remains undocumented.
508 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
509 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
510 for scalar sysctls is defined but disabled. Code that needs
511 UQUAD to pass the type checking that must compile on older
512 systems where the define is not present can check against
513 __FreeBSD_version >= 900030.
515 The system dialog(1) has been replaced with a new version previously
516 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
517 with the previous version, but the libdialog associated with it has
518 a largely incompatible API. As such, the original version of libdialog
519 will be kept temporarily as libodialog, until its base system consumers
520 are replaced or updated. Bump __FreeBSD_version to 900030.
523 If you are trying to run make universe on a -stable system, and you get
524 the following warning:
525 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
526 unknown. config(8) likely too old."
527 or something similar to it, then you must upgrade your -stable system
528 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
529 stable/8) or build the config from the latest stable/8 branch and
530 install it on your system.
532 Prior to this date, building a current universe on 8-stable system from
533 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
534 error in the first kernel build phase. A new config on those old
535 systems will fix that problem for older versions of -current.
538 The TCP stack has been modified to allow Khelp modules to interact with
539 it via helper hook points and store per-connection data in the TCP
540 control block. Bump __FreeBSD_version to 900029. User space tools that
541 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
545 Generic IEEE 802.3 annex 31B full duplex flow control support has been
546 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
547 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
548 to take advantage of it instead of using custom implementations. This
549 means that these drivers now no longer unconditionally advertise
550 support for flow control but only do so if flow control is a selected
551 media option. This was implemented in the generic support that way in
552 order to allow flow control to be switched on and off via ifconfig(8)
553 with the PHY specific default to typically off in order to protect
554 from unwanted effects. Consequently, if you used flow control with
555 one of the above mentioned drivers you now need to explicitly enable
557 ifconfig bge0 media auto mediaopt flowcontrol
559 Along with the above mentioned changes generic support for setting
560 1000baseT master mode also has been added and brgphy(4), ciphy(4),
561 e1000phy(4) as well as ip1000phy(4) have been converted to take
562 advantage of it. This means that these drivers now no longer take the
563 link0 parameter for selecting master mode but the master media option
564 has to be used instead, for example like in the following:
565 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
567 Selection of master mode now is also available with all other PHY
568 drivers supporting 1000baseT.
571 The TCP stack has received a significant update to add support for
572 modularised congestion control and generally improve the clarity of
573 congestion control decisions. Bump __FreeBSD_version to 900025. User
574 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
575 sockstat) need to be recompiled.
578 The man(1) utility has been replaced by a new version that no longer
579 uses /etc/manpath.config. Please consult man.conf(5) for how to
580 migrate local entries to the new format.
583 The copyright strings printed by login(1) and sshd(8) at the time of a
584 new connection have been removed to follow other operating systems and
588 A workaround for a fixed ld bug has been removed in kernel code,
589 so make sure that your system ld is built from sources after
590 revision 210245 from 2010-07-19 (r211583 if building head kernel
591 on stable/8, r211584 for stable/7; both from 2010-08-21).
592 A symptom of incorrect ld version is different addresses for
593 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
596 The $ipv6_prefer variable in rc.conf(5) has been split into
597 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
599 The $ip6addrctl_policy is a variable to choose a pre-defined
600 address selection policy set by ip6addrctl(8). A value
601 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
604 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
605 flag (see an entry of 20090926) is set on an interface with no
606 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
607 security reason. If you want IPv6 link-local address on all
608 interfaces by default, set this to "YES".
610 The old ipv6_prefer="YES" is equivalent to
611 ipv6_activate_all_interfaces="YES" and
612 ip6addrctl_policy="ipv6_prefer".
615 DTrace has grown support for userland tracing. Due to this, DTrace is
616 now i386 and amd64 only.
617 dtruss(1) is now installed by default on those systems and a new
618 kernel module is needed for userland tracing: fasttrap.
619 No changes to your kernel config file are necessary to enable
620 userland tracing, but you might consider adding 'STRIP=' and
621 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
622 to have informative userland stack traces in DTrace (ustack).
625 The acpi_aiboost(4) driver has been removed in favor of the new
626 aibs(4) driver. You should update your kernel configuration file.
629 BSD grep has been imported to the base system and it is built by
630 default. It is completely BSD licensed, highly GNU-compatible, uses
631 less memory than its GNU counterpart and has a small codebase.
632 However, it is slower than its GNU counterpart, which is mostly
633 noticeable for larger searches, for smaller ones it is measurable
634 but not significant. The reason is complex, the most important factor
635 is that we lack a modern and efficient regex library and GNU
636 overcomes this by optimizing the searches internally. Future work
637 on improving the regex performance is planned, for the meantime,
638 users that need better performance, can build GNU grep instead by
639 setting the WITH_GNU_GREP knob.
642 Due to the import of powerpc64 support, all existing powerpc kernel
643 configuration files must be updated with a machine directive like this:
644 machine powerpc powerpc
646 In addition, an updated config(8) is required to build powerpc kernels
650 A new version of ZFS (version 15) has been merged to -HEAD.
651 This version uses a python library for the following subcommands:
652 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
653 For full functionality of these commands the following port must
654 be installed: sysutils/py-zfs
657 'vm_page's are now hashed by physical address to an array of mutexes.
658 Currently this is only used to serialize access to hold_count. Over
659 time the page queue mutex will be peeled away. This changes the size
660 of pmap on every architecture. And requires all callers of vm_page_hold
661 and vm_page_unhold to be updated.
664 WITH_CTF can now be specified in src.conf (not recommended, there
665 are some problems with static executables), make.conf (would also
666 affect ports which do not use GNU make and do not override the
667 compile targets) or in the kernel config (via "makeoptions
669 When WITH_CTF was specified there before this was silently ignored,
670 so make sure that WITH_CTF is not used in places which could lead
671 to unwanted behavior.
674 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
675 to allow 32-bit compatibility on non-x86 platforms. All kernel
676 configurations on amd64 and ia64 platforms using these options must
677 be modified accordingly.
680 The utmp user accounting database has been replaced with utmpx,
681 the user accounting interface standardized by POSIX.
682 Unfortunately the semantics of utmp and utmpx don't match,
683 making it practically impossible to support both interfaces.
684 The user accounting database is used by tools like finger(1),
685 last(1), talk(1), w(1) and ac(8).
687 All applications in the base system use utmpx. This means only
688 local binaries (e.g. from the ports tree) may still use these
689 utmp database files. These applications must be rebuilt to make
692 After the system has been upgraded, it is safe to remove the old
693 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
694 assuming their contents is of no importance anymore. Old wtmp
695 databases can only be used by last(1) and ac(8) after they have
696 been converted to the new format using wtmpcvt(1).
699 Introduce the kernel thread "deadlock resolver" (which can be enabled
700 via the DEADLKRES option, see NOTES for more details) and the
701 sleepq_type() function for sleepqueues.
704 The rc.firewall and rc.firewall6 were unified, and
705 rc.firewall6 and rc.d/ip6fw were removed.
706 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
707 variables are obsoleted. Instead, the following new rc
708 variables are added to rc.d/ipfw:
710 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
711 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
712 firewall_simple_onet_ipv6, firewall_trusted_ipv6
714 The meanings correspond to the relevant IPv4 variables.
720 The default terminal emulation for syscons(4) has been changed
721 from cons25 to xterm on all platforms except pc98. This means
722 that the /etc/ttys file needs to be updated to ensure correct
723 operation of applications on the console.
725 The terminal emulation style can be toggled per window by using
726 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
727 options can be used to change the compile-time default back to
730 To prevent graphical artifacts, make sure the TERM environment
731 variable is set to match the terminal emulation that is being
732 performed by syscons(4).
735 The layout of the structure ieee80211req_scan_result has changed.
736 Applications that require wireless scan results (e.g. ifconfig(8))
737 from net80211 need to be recompiled.
739 Applications such as wpa_supplicant(8) may require a full world
740 build without using NO_CLEAN in order to get synchronized with the
744 The iwn(4) driver has been updated to support the 5000 and 5150 series.
745 There's one kernel module for each firmware. Adding "device iwnfw"
746 to the kernel configuration file means including all three firmware
747 images inside the kernel. If you want to include just the one for
748 your wireless card, use the devices iwn4965fw, iwn5000fw or
752 The rc.d/network_ipv6, IPv6 configuration script has been integrated
753 into rc.d/netif. The changes are the following:
755 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
756 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
757 Note that both variables need the "inet6" keyword at the head.
759 Do not set $ipv6_network_interfaces manually if you do not
760 understand what you are doing. It is not needed in most cases.
762 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
765 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
766 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
768 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
769 all configured interfaces have "inet6 accept_rtadv" in the
770 $ifconfig_IF_ipv6. These are for backward compatibility.
772 3. A new variable $ipv6_prefer has been added. If NO, IPv6
773 functionality of interfaces with no corresponding
774 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
775 and the default address selection policy of ip6addrctl(8)
776 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
777 Note that if you want to configure IPv6 functionality on the
778 disabled interfaces after boot, first you need to clear the flag by
779 using ifconfig(8) like:
781 ifconfig em0 inet6 -ifdisabled
783 If YES, the default address selection policy is set as
786 The default value of $ipv6_prefer is NO.
788 4. If your system need to receive Router Advertisement messages,
789 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
790 scripts automatically invoke rtsol(8) when the interface becomes
791 UP. The Router Advertisement messages are used for SLAAC
792 (State-Less Address AutoConfiguration).
795 802.11s D3.03 support was committed. This is incompatible with the
796 previous code, which was based on D3.0.
799 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
800 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
801 control whether accepting Router Advertisement messages or not.
802 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
803 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
804 The ifconfig(8) utility now supports these flags.
807 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
808 mount(8) and -a option for df(1) to see them.
811 The old tunable hw.bus.devctl_disable has been superseded by
812 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
813 replaced by hw.bus.devctl_queue=0. The default for this new tunable
817 Remove the option STOP_NMI. The default action is now to use NMI only
818 for KDB via the newly introduced function stop_cpus_hard() and
819 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
822 The stable/8 branch created in subversion. This corresponds to the
823 RELENG_8 branch in CVS.
826 Bump the shared library version numbers for all libraries that do not
827 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
828 __FreeBSD_version to 800105.
831 Due to changes in the implementation of virtual network stack support,
832 all network-related kernel modules must be recompiled. As this change
833 breaks the ABI, bump __FreeBSD_version to 800104.
836 The TOE interface to the TCP syncache has been modified to remove
837 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
838 The cxgb driver is the only TOE consumer affected by this change, and
839 needs to be recompiled along with the kernel. As this change breaks
840 the ABI, bump __FreeBSD_version to 800103.
843 Padding has been added to struct tcpcb, sackhint and tcpstat in
844 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
845 maintaining the ABI. However, this change breaks the ABI, so bump
846 __FreeBSD_version to 800102. User space tools that rely on the size of
847 any of these structs (e.g. sockstat) need to be recompiled.
850 The NFS_LEGACYRPC option has been removed along with the old kernel
851 RPC implementation that this option selected. Kernel configurations
852 may need to be adjusted.
855 The network interface device nodes at /dev/net/<interface> have been
856 removed. All ioctl operations can be performed the normal way using
857 routing sockets. The kqueue functionality can generally be replaced
858 with routing sockets.
861 The documentation from the FreeBSD Documentation Project (Handbook,
862 FAQ, etc.) is now installed via packages by sysinstall(8) and under
863 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
866 The ABI of various structures related to the SYSV IPC API have been
867 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
868 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
872 Layout of struct vnet has changed as routing related variables were
873 moved to their own Vimage module. Modules need to be recompiled. Bump
874 __FreeBSD_version to 800099.
877 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
878 respectively. As long as no more than 16 groups per process are used,
879 no changes should be visible. When more than 16 groups are used, old
880 binaries may fail if they call getgroups() or getgrouplist() with
881 statically sized storage. Recompiling will work around this, but
882 applications should be modified to use dynamically allocated storage
883 for group arrays as POSIX.1-2008 does not cap an implementation's
884 number of supported groups at NGROUPS_MAX+1 as previous versions did.
886 NFS and portalfs mounts may also be affected as the list of groups is
887 truncated to 16. Users of NFS who use more than 16 groups, should
888 take care that negative group permissions are not used on the exported
889 file systems as they will not be reliable unless a GSSAPI based
890 authentication method is used.
893 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
894 option compiles in the support for adaptive spinning for lockmgrs
895 which want to enable it. The lockinit() function now accepts the flag
896 LK_ADAPTIVE in order to make the lock object subject to adaptive
897 spinning when both held in write and read mode.
900 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
901 changed. User applications that use this ioctl need to be rebuilt.
904 The layout of struct thread has changed. Kernel and modules need to
908 The layout of structs ifnet, domain, protosw and vnet_net has changed.
909 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
912 window(1) has been removed from the base system. It can now be
913 installed from ports. The port is called misc/window.
916 The way we are storing and accessing `routing table' entries has
917 changed. Programs reading the FIB, like netstat, need to be
921 A new netisr implementation has been added for FreeBSD 8. Network
922 file system modules, such as igmp, ipdivert, and others, should be
924 Bump __FreeBSD_version to 800096.
927 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
931 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
932 Bump __FreeBSD_version to 800094.
935 Add mnt_xflag field to 'struct mount'. File system modules need to be
937 Bump __FreeBSD_version to 800093.
940 The compiling option ADAPTIVE_SX has been retired while it has been
941 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
942 The KPI for sx_init_flags() changes as accepting flags:
943 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
944 been introduced in order to handle the reversed logic.
945 Bump __FreeBSD_version to 800092.
948 Add support for hierarchical jails. Remove global securelevel.
949 Bump __FreeBSD_version to 800091.
952 The layout of struct vnet_net has changed, therefore modules
954 Bump __FreeBSD_version to 800090.
957 The newly imported zic(8) produces a new format in the output. Please
958 run tzsetup(8) to install the newly created data to /etc/localtime.
961 The sysctl tree for the usb stack has renamed from hw.usb2.* to
962 hw.usb.* and is now consistent again with previous releases.
965 802.11 monitor mode support was revised and driver api's were changed.
966 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
967 of DLT_IEEE802_11. No user-visible data structures were changed but
968 applications that use DLT_IEEE802_11 may require changes.
969 Bump __FreeBSD_version to 800088.
972 The layout of the following structs has changed: sysctl_oid,
973 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
974 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
975 panics may be experienced. World rebuild is required for
976 correctly checking networking state from userland.
977 Bump __FreeBSD_version to 800085.
980 MLDv2 and Source-Specific Multicast (SSM) have been merged
981 to the IPv6 stack. VIMAGE hooks are in but not yet used.
982 The implementation of SSM within FreeBSD's IPv6 stack closely
983 follows the IPv4 implementation.
985 For kernel developers:
987 * The most important changes are that the ip6_output() and
988 ip6_input() paths no longer take the IN6_MULTI_LOCK,
989 and this lock has been downgraded to a non-recursive mutex.
991 * As with the changes to the IPv4 stack to support SSM, filtering
992 of inbound multicast traffic must now be performed by transport
993 protocols within the IPv6 stack. This does not apply to TCP and
994 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
996 * The KPIs used by IPv6 multicast are similar to those used by
997 the IPv4 stack, with the following differences:
998 * im6o_mc_filter() is analogous to imo_multicast_filter().
999 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
1000 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
1001 * IN6_LOOKUP_MULTI() has been deprecated and removed.
1002 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
1003 for MLDv1 have an additional 'timer' argument which is used to
1004 jitter the initial membership report for the solicited-node
1005 multicast membership on-link.
1006 * This is not strictly needed for MLDv2, which already jitters
1007 its report transmissions. However, the 'timer' argument is
1008 preserved in case MLDv1 is active on the interface.
1010 * The KAME linked-list based IPv6 membership implementation has
1011 been refactored to use a vector similar to that used by the IPv4
1013 Code which maintains a list of its own multicast memberships
1014 internally, e.g. carp, has been updated to reflect the new
1017 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
1018 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
1019 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
1020 implementation constraint which needs to be addressed in HEAD.
1022 For application developers:
1024 * The changes are broadly similar to those made for the IPv4
1027 * The use of IPv4 and IPv6 multicast socket options on the same
1028 socket, using mapped addresses, HAS NOT been tested or supported.
1030 * There are a number of issues with the implementation of various
1031 IPv6 multicast APIs which need to be resolved in the API surface
1032 before the implementation is fully compatible with KAME userland
1033 use, and these are mostly to do with interface index treatment.
1035 * The literature available discusses the use of either the delta / ASM
1036 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
1037 using setsourcefilter(3)/getsourcefilter(3). For more information
1038 please refer to RFC 3768, 'Socket Interface Extensions for
1039 Multicast Source Filters'.
1041 * Applications which use the published RFC 3678 APIs should be fine.
1043 For systems administrators:
1045 * The mtest(8) utility has been refactored to support IPv6, in
1046 addition to IPv4. Interface addresses are no longer accepted
1047 as arguments, their names must be used instead. The utility
1048 will map the interface name to its first IPv4 address as
1049 returned by getifaddrs(3).
1051 * The ifmcstat(8) utility has also been updated to print the MLDv2
1052 endpoint state and source filter lists via sysctl(3).
1054 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
1055 loopback of IPv6 multicast datagrams by default; it defaults to 1
1056 to preserve the existing behaviour. Disabling multicast loopback is
1057 recommended for optimal system performance.
1059 * The IPv6 MROUTING code has been changed to examine this sysctl
1060 instead of attempting to perform a group lookup before looping
1061 back forwarded datagrams.
1063 Bump __FreeBSD_version to 800084.
1066 Implement low-level Bluetooth HCI API.
1067 Bump __FreeBSD_version to 800083.
1070 The layout of struct malloc_type, used by modules to register new
1071 memory allocation types, has changed. Most modules will need to
1072 be rebuilt or panics may be experienced.
1073 Bump __FreeBSD_version to 800081.
1076 Anticipate overflowing inp_flags - add inp_flags2.
1077 This changes most offsets in inpcb, so checking v4 connection
1078 state will require a world rebuild.
1079 Bump __FreeBSD_version to 800080.
1082 Add an llentry to struct route and struct route_in6. Modules
1083 embedding a struct route will need to be recompiled.
1084 Bump __FreeBSD_version to 800079.
1087 The size of rt_metrics_lite and by extension rtentry has changed.
1088 Networking administration apps will need to be recompiled.
1089 The route command now supports show as an alias for get, weighting
1090 of routes, sticky and nostick flags to alter the behavior of stateful
1092 Bump __FreeBSD_version to 800078.
1095 Do not use Giant for kbdmux(4) locking. This is wrong and
1096 apparently causing more problems than it solves. This will
1097 re-open the issue where interrupt handlers may race with
1098 kbdmux(4) in polling mode. Typical symptoms include (but
1099 not limited to) duplicated and/or missing characters when
1100 low level console functions (such as gets) are used while
1101 interrupts are enabled (for example geli password prompt,
1102 mountroot prompt etc.). Disabling kbdmux(4) may help.
1105 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1106 kernel modules referencing any of the above need to be recompiled.
1107 Bump __FreeBSD_version to 800075.
1110 GEOM_PART has become the default partition slicer for storage devices,
1111 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1112 introduces some changes:
1114 MSDOS/EBR: the devices created from MSDOS extended partition entries
1115 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1116 to devices with offset-based names. fstabs may need to be modified.
1118 BSD: the "geometry does not match label" warning is harmless in most
1119 cases but it points to problems in file system misalignment with
1120 disk geometry. The "c" partition is now implicit, covers the whole
1121 top-level drive and cannot be (mis)used by users.
1123 General: Kernel dumps are now not allowed to be written to devices
1124 whose partition types indicate they are meant to be used for file
1125 systems (or, in case of MSDOS partitions, as something else than
1128 Most of these changes date approximately from 200812.
1131 The uscanner(4) driver has been removed from the kernel. This follows
1132 Linux removing theirs in 2.6 and making libusb the default interface
1133 (supported by sane).
1136 The multicast forwarding code has been cleaned up. netstat(1)
1137 only relies on KVM now for printing bandwidth upcall meters.
1138 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1139 ip6_mroute_mod respectively. The config(5) options for statically
1140 compiling this code remain the same, i.e. 'options MROUTING'.
1143 Support for the IFF_NEEDSGIANT network interface flag has been
1144 removed, which means that non-MPSAFE network device drivers are no
1145 longer supported. In particular, if_ar, if_sr, and network device
1146 drivers from the old (legacy) USB stack can no longer be built or
1150 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1151 a bunch of new language catalog files have also been added.
1152 This means that some common libc messages are now localized and
1153 they depend on the LC_MESSAGES environmental variable.
1156 The k8temp(4) driver has been renamed to amdtemp(4) since
1157 support for Family 10 and Family 11 CPU families was added.
1160 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1161 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1163 For kernel developers, the most important changes are that the
1164 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1165 and this lock has been downgraded to a non-recursive mutex.
1167 Transport protocols (UDP, Raw IP) are now responsible for filtering
1168 inbound multicast traffic according to group membership and source
1169 filters. The imo_multicast_filter() KPI exists for this purpose.
1170 Transports which do not use multicast (SCTP, TCP) already reject
1171 multicast by default. Forwarding and receive performance may improve
1172 as a mutex acquisition is no longer needed in the ip_input()
1173 low-level input path. in_addmulti() and in_delmulti() are shimmed
1174 to new KPIs which exist to support SSM in-kernel.
1176 For application developers, it is recommended that loopback of
1177 multicast datagrams be disabled for best performance, as this
1178 will still cause the lock to be taken for each looped-back
1179 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1180 be tuned to 0 to disable loopback by default; it defaults to 1
1181 to preserve the existing behaviour.
1183 For systems administrators, to obtain best performance with
1184 multicast reception and multiple groups, it is always recommended
1185 that a card with a suitably precise hash filter is used. Hash
1186 collisions will still result in the lock being taken within the
1187 transport protocol input path to check group membership.
1189 If deploying FreeBSD in an environment with IGMP snooping switches,
1190 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1191 enabled; this forces 224.0.0.0/24 group membership to be announced
1194 The size of 'struct igmpstat' has changed; netstat needs to be
1195 recompiled to reflect this.
1196 Bump __FreeBSD_version to 800070.
1199 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1200 updated to use it. This requires a buildworld/installworld in order to
1201 update the library and dependencies (usbconfig, etc). Its advisable to
1202 rebuild all ports which uses libusb. More specific directions are given
1203 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1204 libusb are no longer required and can be removed.
1207 A workaround is committed to allow the creation of System V shared
1208 memory segment of size > 2 GB on the 64-bit architectures.
1209 Due to a limitation of the existing ABI, the shm_segsz member
1210 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1211 wrong for large segments. Note that limits must be explicitly
1212 raised to allow such segments to be created.
1215 The layout of struct ifnet has changed, requiring a rebuild of all
1216 network device driver modules.
1219 The /dev handling for the new USB stack has changed, a
1220 buildworld/installworld is required for libusb20.
1223 The new USB2 stack has now been permanently moved in and all kernel and
1224 module names reverted to their previous values (eg, usb, ehci, ohci,
1225 ums, ...). The old usb stack can be compiled in by prefixing the name
1226 with the letter 'o', the old usb modules have been removed.
1227 Updating entry 20090216 for xorg and 20090215 for libmap may still
1231 The rc.conf(5) option if_up_delay has been renamed to
1232 defaultroute_delay to better reflect its purpose. If you have
1233 customized this setting in /etc/rc.conf you need to update it to
1237 xorg 7.4 wants to configure its input devices via hald which does not
1238 yet work with USB2. If the keyboard/mouse does not work in xorg then
1240 Option "AllowEmptyInput" "off"
1241 to your ServerLayout section. This will cause X to use the configured
1242 kbd and mouse sections from your xorg.conf.
1245 The GENERIC kernels for all architectures now default to the new USB2
1246 stack. No kernel config options or code have been removed so if a
1247 problem arises please report it and optionally revert to the old USB
1248 stack. If you are loading USB kernel modules or have a custom kernel
1249 that includes GENERIC then ensure that usb names are also changed over,
1250 eg uftdi -> usb2_serial_ftdi.
1252 Older programs linked against the ports libusb 0.1 need to be
1253 redirected to the new stack's libusb20. /etc/libmap.conf can
1255 # Map old usb library to new one for usb2 stack
1256 libusb-0.1.so.8 libusb20.so.1
1259 All USB ethernet devices now attach as interfaces under the name ueN
1260 (eg. ue0). This is to provide a predictable name as vendors often
1261 change usb chipsets in a product without notice.
1264 The ichsmb(4) driver has been changed to require SMBus slave
1265 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1266 All of the other SMBus controller drivers require left-justified
1267 slave addresses, so this change makes all the drivers provide the
1271 INET6 statistics (struct ip6stat) was updated.
1272 netstat(1) needs to be recompiled.
1275 NTFS has been removed from GENERIC kernel on amd64 to match
1276 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1277 will load ntfs.ko module automatically when NTFS support is
1278 actually needed, unless ntfs.ko is not installed or security
1279 level prohibits loading kernel modules. If either is the case,
1280 "options NTFS" has to be added into kernel config.
1283 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1284 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1285 800061. User space tools that rely on the size of struct tcpcb in
1286 tcp_var.h (e.g. sockstat) need to be recompiled.
1289 ng_tty(4) module updated to match the new TTY subsystem.
1290 Due to API change, user-level applications must be updated.
1291 New API support added to mpd5 CVS and expected to be present
1292 in next mpd5.3 release.
1295 With __FreeBSD_version 800060 the makefs tool is part of
1296 the base system (it was a port).
1299 The afdata and ifnet locks have been changed from mutexes to
1300 rwlocks, network modules will need to be re-compiled.
1303 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1304 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1305 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1306 architecture and 40 bytes on 64-bit architecture. The userland
1307 applications "arp" and "ndp" have been updated accordingly.
1308 The output from "netstat -r" shows only routing entries and
1309 none of the L2 information.
1312 __FreeBSD_version 800057 marks the switchover from the
1313 binary ath hal to source code. Users must add the line:
1315 options AH_SUPPORT_AR5416
1317 to their kernel config files when specifying:
1321 The ath_hal module no longer exists; the code is now compiled
1322 together with the driver in the ath module. It is now
1323 possible to tailor chip support (i.e. reduce the set of chips
1324 and thereby the code size); consult ath_hal(4) for details.
1327 __FreeBSD_version 800054 adds memory barriers to
1328 <machine/atomic.h>, new interfaces to ifnet to facilitate
1329 multiple hardware transmit queues for cards that support
1330 them, and a lock-less ring-buffer implementation to
1331 enable drivers to more efficiently manage queueing of
1335 A new version of ZFS (version 13) has been merged to -HEAD.
1336 This version has zpool attribute "listsnapshots" off by
1337 default, which means "zfs list" does not show snapshots,
1338 and is the same as Solaris behavior.
1341 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1344 The uhci, ohci, ehci and slhci USB Host controller drivers have
1345 been put into separate modules. If you load the usb module
1346 separately through loader.conf you will need to load the
1347 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1348 controller add the following to loader.conf:
1354 The ABI used by the PMC toolset has changed. Please keep
1355 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1359 atapci kernel module now includes only generic PCI ATA
1360 driver. AHCI driver moved to ataahci kernel module.
1361 All vendor-specific code moved into separate kernel modules:
1362 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1363 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1364 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1365 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1368 The TTY subsystem of the kernel has been replaced by a new
1369 implementation, which provides better scalability and an
1370 improved driver model. Most common drivers have been migrated to
1371 the new TTY subsystem, while others have not. The following
1372 drivers have not yet been ported to the new TTY layer:
1375 cy, digi, rc, rp, sio
1381 ng_h4, ng_tty, ppp, sl, snp
1383 Adding these drivers to your kernel configuration file shall
1384 cause compilation to fail.
1387 ntpd has been upgraded to 4.2.4p5.
1390 OpenSSH has been upgraded to 5.1p1.
1392 For many years, FreeBSD's version of OpenSSH preferred DSA
1393 over RSA for host and user authentication keys. With this
1394 upgrade, we've switched to the vendor's default of RSA over
1395 DSA. This may cause upgraded clients to warn about unknown
1396 host keys even for previously known hosts. Users should
1397 follow the usual procedure for verifying host keys before
1398 accepting the RSA key.
1400 This can be circumvented by setting the "HostKeyAlgorithms"
1401 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1404 Please note that the sequence of keys offered for
1405 authentication has been changed as well. You may want to
1406 specify IdentityFile in a different order to revert this
1410 The sio(4) driver has been removed from the i386 and amd64
1411 kernel configuration files. This means uart(4) is now the
1412 default serial port driver on those platforms as well.
1414 To prevent collisions with the sio(4) driver, the uart(4) driver
1415 uses different names for its device nodes. This means the
1416 onboard serial port will now most likely be called "ttyu0"
1417 instead of "ttyd0". You may need to reconfigure applications to
1418 use the new device names.
1420 When using the serial port as a boot console, be sure to update
1421 /boot/device.hints and /etc/ttys before booting the new kernel.
1422 If you forget to do so, you can still manually specify the hints
1423 at the loader prompt:
1425 set hint.uart.0.at="isa"
1426 set hint.uart.0.port="0x3F8"
1427 set hint.uart.0.flags="0x10"
1428 set hint.uart.0.irq="4"
1432 The gpt(8) utility has been removed. Use gpart(8) to partition
1436 The version that Linuxulator emulates was changed from 2.4.2
1437 to 2.6.16. If you experience any problems with Linux binaries
1438 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1439 if it fixes the problem contact emulation mailing list.
1442 ISDN4BSD (I4B) was removed from the src tree. You may need to
1443 update a your kernel configuration and remove relevant entries.
1446 I have checked in code to support multiple routing tables.
1447 See the man pages setfib(1) and setfib(2).
1448 This is a hopefully backwards compatible version,
1449 but to make use of it you need to compile your kernel
1450 with options ROUTETABLES=2 (or more up to 16).
1453 The 802.11 wireless support was redone to enable multi-bss
1454 operation on devices that are capable. The underlying device
1455 is no longer used directly but instead wlanX devices are
1456 cloned with ifconfig. This requires changes to rc.conf files.
1457 For example, change:
1458 ifconfig_ath0="WPA DHCP"
1461 ifconfig_wlan0="WPA DHCP"
1462 see rc.conf(5) for more details. In addition, mergemaster of
1463 /etc/rc.d is highly recommended. Simultaneous update of userland
1464 and kernel wouldn't hurt either.
1466 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1467 modules were merged into the base wlan module. All references
1468 to these modules (e.g. in kernel config files) must be removed.
1471 psm(4) has gained write(2) support in native operation level.
1472 Arbitrary commands can be written to /dev/psm%d and status can
1473 be read back from it. Therefore, an application is responsible
1474 for status validation and error recovery. It is a no-op in
1475 other operation levels.
1478 Support for KSE threading has been removed from the kernel. To
1479 run legacy applications linked against KSE libmap.conf may
1480 be used. The following libmap.conf may be used to ensure
1481 compatibility with any prior release:
1483 libpthread.so.1 libthr.so.1
1484 libpthread.so.2 libthr.so.2
1485 libkse.so.3 libthr.so.3
1488 The layout of struct vmspace has changed. This affects libkvm
1489 and any executables that link against libkvm and use the
1490 kvm_getprocs() function. In particular, but not exclusively,
1491 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1492 The effects are minimal, but it's advisable to upgrade world
1496 The latest em driver no longer has support in it for the
1497 82575 adapter, this is now moved to the igb driver. The
1498 split was done to make new features that are incompatible
1499 with older hardware easier to do.
1502 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1503 likewise the kernel option is now GEOM_LINUX_LVM.
1506 The default NFS mount mode has changed from UDP to TCP for
1507 increased reliability. If you rely on (insecurely) NFS
1508 mounting across a firewall you may need to update your
1512 Belatedly note the addition of m_collapse for compacting
1516 The fts(3) structures have been changed to use adequate
1517 integer types for their members and so to be able to cope
1518 with huge file trees. The old fts(3) ABI is preserved
1519 through symbol versioning in libc, so third-party binaries
1520 using fts(3) should still work, although they will not take
1521 advantage of the extended types. At the same time, some
1522 third-party software might fail to build after this change
1523 due to unportable assumptions made in its source code about
1524 fts(3) structure members. Such software should be fixed
1525 by its vendor or, in the worst case, in the ports tree.
1526 FreeBSD_version 800015 marks this change for the unlikely
1527 case that a portable fix is impossible.
1530 To upgrade to -current after this date, you must be running
1531 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1532 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1535 The ADAPTIVE_GIANT kernel option has been retired because its
1536 functionality is the default now.
1539 The AT keyboard emulation of sunkbd(4) has been turned on
1540 by default. In order to make the special symbols of the Sun
1541 keyboards driven by sunkbd(4) work under X these now have
1542 to be configured the same way as Sun USB keyboards driven
1543 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1545 Option "XkbLayout" "us"
1546 Option "XkbRules" "xorg"
1547 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1550 It has been decided that it is desirable to provide ABI
1551 backwards compatibility to the FreeBSD 4/5/6 versions of the
1552 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1553 broken with the introduction of PCI domain support (see the
1554 20070930 entry). Unfortunately, this required the ABI of
1555 PCIOCGETCONF to be broken again in order to be able to
1556 provide backwards compatibility to the old version of that
1557 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1558 again. As for prominent ports this affects neither pciutils
1559 nor xorg-server this time, the hal port needs to be rebuilt
1563 The misnamed kthread_create() and friends have been renamed
1564 to kproc_create() etc. Many of the callers already
1565 used kproc_start()..
1566 I will return kthread_create() and friends in a while
1567 with implementations that actually create threads, not procs.
1568 Renaming corresponds with version 800002.
1577 Avoid using make -j when upgrading. While generally safe, there are
1578 sometimes problems using -j to upgrade. If your upgrade fails with
1579 -j, please try again without -j. From time to time in the past there
1580 have been problems using -j with buildworld and/or installworld. This
1581 is especially true when upgrading between "distant" versions (eg one
1582 that cross a major release boundary or several minor releases, or when
1583 several months have passed on the -current branch).
1585 Sometimes, obscure build problems are the result of environment
1586 poisoning. This can happen because the make utility reads its
1587 environment when searching for values for global variables. To run
1588 your build attempts in an "environmental clean room", prefix all make
1589 commands with 'env -i '. See the env(1) manual page for more details.
1591 When upgrading from one major version to another it is generally best
1592 to upgrade to the latest code in the currently installed branch first,
1593 then do an upgrade to the new branch. This is the best-tested upgrade
1594 path, and has the highest probability of being successful. Please try
1595 this approach before reporting problems with a major version upgrade.
1599 When upgrading the boot ZFS pool to a new version, always follow
1602 1.) recompile and reinstall the ZFS boot loader and boot block
1603 (this is part of "make buildworld" and "make installworld")
1605 2.) update the ZFS boot block on your boot drive
1607 The following example updates the ZFS boot block on the first
1608 partition (freebsd-boot) of a GPT partitioned drive ad0:
1609 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1611 Non-boot pools do not need these updates.
1615 If you are updating from a prior version of FreeBSD (even one just
1616 a few days old), you should follow this procedure. It is the most
1617 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1619 make kernel-toolchain
1620 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1621 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1623 To test a kernel once
1624 ---------------------
1625 If you just want to boot a kernel once (because you are not sure
1626 if it works, or if you want to boot a known bad kernel to provide
1627 debugging information) run
1628 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1629 nextboot -k testkernel
1631 To just build a kernel when you know that it won't mess you up
1632 --------------------------------------------------------------
1633 This assumes you are already running a CURRENT system. Replace
1634 ${arch} with the architecture of your machine (e.g. "i386",
1635 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1637 cd src/sys/${arch}/conf
1638 config KERNEL_NAME_HERE
1639 cd ../compile/KERNEL_NAME_HERE
1644 If this fails, go to the "To build a kernel" section.
1646 To rebuild everything and install it on the current system.
1647 -----------------------------------------------------------
1648 # Note: sometimes if you are running current you gotta do more than
1649 # is listed here if you are upgrading from a really old current.
1651 <make sure you have good level 0 dumps>
1653 make kernel KERNCONF=YOUR_KERNEL_HERE
1655 <reboot in single user> [3]
1663 To cross-install current onto a separate partition
1664 --------------------------------------------------
1665 # In this approach we use a separate partition to hold
1666 # current's root, 'usr', and 'var' directories. A partition
1667 # holding "/", "/usr" and "/var" should be about 2GB in
1670 <make sure you have good level 0 dumps>
1673 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1674 <maybe newfs current's root partition>
1675 <mount current's root partition on directory ${CURRENT_ROOT}>
1676 make installworld DESTDIR=${CURRENT_ROOT}
1677 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1678 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1679 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1680 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1681 <reboot into current>
1682 <do a "native" rebuild/install as described in the previous section>
1683 <maybe install compatibility libraries from ports/misc/compat*>
1687 To upgrade in-place from stable to current
1688 ----------------------------------------------
1689 <make sure you have good level 0 dumps>
1691 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1693 <reboot in single user> [3]
1700 Make sure that you've read the UPDATING file to understand the
1701 tweaks to various things you need. At this point in the life
1702 cycle of current, things change often and you are on your own
1703 to cope. The defaults can also change, so please read ALL of
1704 the UPDATING entries.
1706 Also, if you are tracking -current, you must be subscribed to
1707 freebsd-current@freebsd.org. Make sure that before you update
1708 your sources that you have read and understood all the recent
1709 messages there. If in doubt, please track -stable which has
1710 much fewer pitfalls.
1712 [1] If you have third party modules, such as vmware, you
1713 should disable them at this point so they don't crash your
1716 [3] From the bootblocks, boot -s, and then do
1721 adjkerntz -i # if CMOS is wall time
1722 Also, when doing a major release upgrade, it is required that
1723 you boot into single user mode to do the installworld.
1725 [4] Note: This step is non-optional. Failure to do this step
1726 can result in a significant reduction in the functionality of the
1727 system. Attempting to do it by hand is not recommended and those
1728 that pursue this avenue should read this file carefully, as well
1729 as the archives of freebsd-current and freebsd-hackers mailing lists
1730 for potential gotchas. The -U option is also useful to consider.
1731 See mergemaster(8) for more information.
1733 [5] Usually this step is a noop. However, from time to time
1734 you may need to do this if you get unknown user in the following
1735 step. It never hurts to do it all the time. You may need to
1736 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1737 install) after the buildworld before this step if you last updated
1738 from [78]-stable or 9-stable before 20130430.
1740 [6] This only deletes old files and directories. Old libraries
1741 can be deleted by "make delete-old-libs", but you have to make
1742 sure that no program is using those libraries anymore.
1744 [8] In order to have a kernel that can run the 4.x binaries needed to
1745 do an installworld, you must include the COMPAT_FREEBSD4 option in
1746 your kernel. Failure to do so may leave you with a system that is
1747 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1748 required to run the 5.x binaries on more recent kernels. And so on
1749 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1751 Make sure that you merge any new devices from GENERIC since the
1752 last time you updated your kernel config file.
1754 [9] When checking out sources, you must include the -P flag to have
1755 cvs prune empty directories.
1757 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1758 "?=" instead of the "=" assignment operator, so that buildworld can
1759 override the CPUTYPE if it needs to.
1761 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1762 not on the command line, or in /etc/make.conf. buildworld will
1763 warn if it is improperly defined.
1766 This file contains a list, in reverse chronological order, of major
1767 breakages in tracking -current. It is not guaranteed to be a complete
1768 list of such breakages, and only contains entries since October 10, 2007.
1769 If you need to see UPDATING entries from before that date, you will need
1770 to fetch an UPDATING file from an older FreeBSD release.
1772 Copyright information:
1774 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1776 Redistribution, publication, translation and use, with or without
1777 modification, in full or in part, in any form or format of this
1778 document are permitted without further permission from the author.
1780 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1781 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1782 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1783 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1784 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1785 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1786 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1787 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1788 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1789 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1790 POSSIBILITY OF SUCH DAMAGE.
1792 Contact Warner Losh if you have any questions about your use of
projects / FreeBSD / releng / 9.3.git / blob