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 20150618: p17 FreeBSD-EN-15:08.sendmail
15 Improvements to sendmail TLS/DH interoperability. [EN-15:08]
17 20150612: p16 FreeBSD-SA-15:10.openssl
18 Fix multiple vulnerabilities in OpenSSL. [SA-15:10]
20 20150609: p15 FreeBSD-EN-15:06.file
22 Updated base system file(1) to 5.22 to address multiple denial
25 20150513: p14 FreeBSD-EN-15:04.freebsd-update
27 Fix bug with freebsd-update(8) that does not ensure the previous
28 upgrade was completed. [EN-15:04]
30 20150407: p13 FreeBSD-SA-15:04.igmp [revised]
34 Improved patch for SA-15:04.igmp.
36 Fix multiple vulnerabilities of ntp. [SA-15:07]
38 Fix Denial of Service with IPv6 Router Advertisements. [SA-15:09]
41 Fix patch for SA-15:06.openssl.
43 20150319: p11 FreeBSD-SA-15:06.openssl
44 Fix multiple vulnerabilities in OpenSSL. [SA-15:06]
46 20150225: p10 FreeBSD-SA-15:04.igmp
49 FreeBSD-EN-15:02.openssl
50 FreeBSD-EN-15:03.freebsd-update
52 Fix integer overflow in IGMP protocol. [SA-15:04]
54 Fix BIND remote denial of service vulnerability. [SA-15:05]
56 Fix vt(4) crash with improper ioctl parameters. [EN-15:01]
58 Updated base system OpenSSL to 0.9.8zd. [EN-15:02]
60 Fix freebsd-update libraries update ordering issue. [EN-15:03]
62 20150127: p9 FreeBSD-SA-15:02.kmem
65 Fix SCTP SCTP_SS_VALUE kernel memory corruption and disclosure
66 vulnerability. [SA-15:02]
68 Fix SCTP stream reset vulnerability. [SA-15:03]
70 20150114: p8 FreeBSD-SA-15:01.openssl
71 Fix multiple vulnerabilities in OpenSSL. [SA-15:01]
73 20141223: p7 FreeBSD-SA-14:31.ntp
74 FreeBSD-EN-14:13.freebsd-update
76 Fix multiple vulnerabilities in NTP suite. [SA-14:31]
77 Fix directory deletion issue in freebsd-update. [EN-14:13]
79 20141210: p6 FreeBSD-SA-14:28.file
82 Fix multiple vulnerabilities in file(1) and libmagic(3).
85 Fix BIND remote denial of service vulnerability. [SA-14:29]
87 20141104: p5 FreeBSD-SA-14:25.setlogin
91 Fix kernel stack disclosure in setlogin(2) / getlogin(2).
94 Fix remote command execution in ftp(1). [SA-14:26]
96 Fix NFSv4 and ZFS cache consistency issue. [EN-14:12]
98 20141022: p4 FreeBSD-EN-14:10.tzdata
99 FreeBSD-EN-14:11.crypt
101 Time zone data file update. [EN-14:10]
103 Change crypt(3) default hashing algorithm back to DES. [EN-14:11]
105 20141021: p3 FreeBSD-SA-14:20.rtsold
106 FreeBSD-SA-14:21.routed
107 FreeBSD-SA-14:22.namei
108 FreeBSD-SA-14:23.openssl
110 Fix rtsold(8) remote buffer overflow vulnerability. [SA-14:20]
112 Fix routed(8) remote denial of service vulnerability. [SA-14:21]
114 Fix memory leak in sandboxed namei lookup. [SA-14:22]
116 Fix OpenSSL multiple vulnerabilities. [SA-14:23]
118 20140916: p2 FreeBSD-SA-14:19.tcp
119 Fix Denial of Service in TCP packet processing. [SA-14:19]
121 20140909: p1 FreeBSD-SA-14:18.openssl
122 Fix OpenSSL multiple vulnerabilities. [SA-14:18]
128 On i386 and amd64 systems, the onifconsole flag is now set by default
129 in /etc/ttys for ttyu0. This causes ttyu0 to be automatically enabled
130 as a login TTY if it is set in the bootloader as an active kernel
131 console. No changes in behavior should result otherwise. To revert to
132 the previous behavior, set ttyu0 to "off" in /etc/ttys.
135 Clang and llvm have been upgraded to 3.4.1 release.
138 Clang and llvm have been upgraded to 3.4 release.
141 The nve(4) driver for NVIDIA nForce MCP Ethernet adapters has
142 been deprecated and will not be part of FreeBSD 11.0 and later
143 releases. If you use this driver, please consider switching to
144 the nfe(4) driver instead.
147 The behavior of gss_pseudo_random() for the krb5 mechanism
148 has changed, for applications requesting a longer random string
149 than produced by the underlying enctype's pseudo-random() function.
150 In particular, the random string produced from a session key of
151 enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
152 be different at the 17th octet and later, after this change.
153 The counter used in the PRF+ construction is now encoded as a
154 big-endian integer in accordance with RFC 4402.
155 __FreeBSD_version is bumped to 902505.
161 Behavior of devfs rules path matching has been changed.
162 Pattern is now always matched against fully qualified devfs
163 path and slash characters must be explicitly matched by
164 slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
165 subdirectories must be reviewed.
168 hastctl(8)'s `status' command output changed to terse one-liner format.
169 Scripts using this should switch to `list' command or be rewritten.
172 Fix a bug that allowed a tracing process (e.g. gdb) to write
173 to a memory-mapped file in the traced process's address space
174 even if neither the traced process nor the tracing process had
175 write access to that file.
178 Added ZFS TRIM support which is enabled by default. To disable
179 ZFS TRIM support set vfs.zfs.trim.enabled=0 in loader.conf.
181 Creating new ZFS pools and adding new devices to existing pools
182 first performs a full device level TRIM which can take a significant
183 amount of time. The sysctl vfs.zfs.vdev.trim_on_init can be set to 0
184 to disable this behaviour.
186 ZFS TRIM requires the underlying device support BIO_DELETE which
187 is currently provided by methods such as ATA TRIM and SCSI UNMAP
188 via CAM, which are typically supported by SSD's.
190 Stats for ZFS TRIM can be monitored by looking at the sysctl's
191 under kstat.zfs.misc.zio_trim.
194 `list' command has been added to hastctl(8). For now, it is full
195 equivalent of `status' command.
196 WARNING: in the near future the output of hastctl's status command
197 will change to more terse format. If you use `hastctl status'
198 for parsing in your scripts, switch to `hastctl list'.
201 The mergemaster command now uses the default MAKEOBJDIRPREFIX
202 rather than creating it's own in the temporary directory in
203 order allow access to bootstrapped versions of tools such as
204 install and mtree. When upgrading from version of FreeBSD where
205 the install command does not support -l, you will need to
206 install a new mergemaster command if mergemaster -p is required.
207 This can be accomplished with the command (cd src/usr.sbin/mergemaster
210 Due to the use of the new -l option to install(1) during build
211 and install, you must take care not to directly set the INSTALL
212 make variable in your /etc/make.conf, /etc/src.conf, or on the
213 command line. If you wish to use the -C flag for all installs
214 you may be able to add INSTALL+=-C to /etc/make.conf or
218 Fix a bug that allows NFS clients to issue READDIR on files.
221 The install(1) option -M has changed meaning and now takes an
222 argument that is a file or path to append logs to. In the
223 unlikely event that -M was the last option on the command line
224 and the command line contained at least two files and a target
225 directory the first file will have logs appended to it. The -M
226 option served little practical purpose in the last decade so it's
227 used expected to be extremely rare.
230 A new compression method (lz4) has been merged to. Please refer to
231 zpool-features(7) for more information.
233 Please refer to the "ZFS notes" section of this file for information
234 on upgrading boot ZFS pools.
237 The VFS KBI was changed with the merge of several nullfs
238 optimizations and fixes. All filesystem modules must be
242 With the addition of auditdistd(8), a new auditdistd user is now
243 depended on during installworld. "mergemaster -p" can be used to add
244 the user prior to installworld, as documented in the handbook.
250 A new version of ZFS (pool version 5000) has been merged to 9-STABLE.
251 Starting with this version the old system of ZFS pool versioning
252 is superseded by "feature flags". This concept enables forward
253 compatibility against certain future changes in functionality of ZFS
254 pools. The first two read-only compatible "feature flags" for ZFS
255 pools are "com.delphix:async_destroy" and "com.delphix:empty_bpobj".
256 For more information read the new zpool-features(7) manual page.
257 Please refer to the "ZFS notes" section of this file for information
258 on upgrading boot ZFS pools.
261 The commit introducing bsd.compiler.mk breaks the traditional
262 building of kernels before this point. Add -m ${SRC}/share/mk
263 (for the right value of SRC) to your command lines to work
264 around; update your useland to a point after this; or use the
265 buildkernel/installkernel top-level targets. See also 20120829.
268 The IPFIREWALL_FORWARD kernel option has been removed. Its
269 functionality now turned on by default.
272 The random(4) support for the VIA hardware random number
273 generator (`PADLOCK') is no longer enabled unconditionally.
274 Add the PADLOCK_RNG option in the custom kernel config if
275 needed. The GENERIC kernels on i386 and amd64 do include the
276 option, so the change only affects the custom kernel
280 The amd64 kernel now uses xsetbv, xrstor instructions. To compile with
281 the traditional method, you must update your system with an installworld
282 before the kernel will build. The documented make buildkernel/installkernel
283 interfaces (coupled with fresh make kernel-toolchain) continue to work.
286 The sparc64 ZFS loader has been changed to no longer try to auto-
287 detect ZFS providers based on diskN aliases but now requires these
288 to be explicitly listed in the OFW boot-device environment variable.
291 Now unix domain sockets behave "as expected" on nullfs(5). Previously
292 nullfs(5) did not pass through all behaviours to the underlying layer,
293 as a result if we bound to a socket on the lower layer we could connect
294 only to the lower path; if we bound to the upper layer we could connect
295 only to the upper path. The new behavior is one can connect to both the
296 lower and the upper paths regardless what layer path one binds to.
299 The acpi_wmi(4) status device /dev/wmistat has been renamed to
303 A new VOP_ADVISE() was added to support posix_fadvise(2). All
304 filesystem modules must be recompiled.
307 The interface of the VOP_VPTOCNP(9) changed, now the returned
308 vnode shall be referenced, previously it was required to be
309 only held. All in-tree filesystems are converted.
315 The broken amd(4) driver has been replaced with esp(4) in the amd64,
316 i386 and pc98 GENERIC kernel configuration files.
319 This commit modifies vfs_register() so that it uses a hash
320 calculation to set vfc_typenum, which is enabled by default.
321 The first time a system is booted after this change, the
322 vfc_typenum values will change for all file systems. The
323 main effect of this is a change to the NFS server file handles
324 for file systems that use vfc_typenum in their fsid, such as ZFS.
325 It will, however, prevent vfc_typenum from changing when file
326 systems are loaded in a different order for subsequent reboots.
327 To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
328 until you are ready to remount all NFS clients after a reboot.
331 Bump the shared library version numbers for libraries that
332 do not use symbol versioning, have changed the ABI compared
333 to stable/8 and which shared library version was not bumped.
334 Done as part of 9.0-RELEASE cycle.
337 During the merge of Capsicum features, the fget(9) KPI was modified.
338 This may require the rebuilding of out-of-tree device drivers --
339 issues have been reported specifically with the nVidia device driver.
340 __FreeBSD_version is bumped to 900041.
342 Also, there is a period between 20110811 and 20110814 where the
343 special devices /dev/{stdin,stdout,stderr} did not work correctly.
344 Building world from a kernel during that window may not work.
347 The packet filter (pf) code has been updated to OpenBSD 4.5.
348 You need to update userland tools to be in sync with kernel.
349 This update breaks backward compatibility with earlier pfsync(4)
350 versions. Care must be taken when updating redundant firewall setups.
353 The following sysctls and tunables are retired on x86 platforms:
355 machdep.hlt_logical_cpus
356 The following sysctl is retired:
357 machdep.hyperthreading_allowed
358 The sysctls were supposed to provide a way to dynamically offline and
359 online selected CPUs on x86 platforms, but the implementation has not
360 been reliable especially with SCHED_ULE scheduler.
361 machdep.hyperthreading_allowed tunable is still available to ignore
362 hyperthreading CPUs at OS level.
363 Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
364 where X is an APIC ID of a CPU. Be advised, though, that disabling
365 CPUs in non-uniform fashion will result in non-uniform topology and
366 may lead to sub-optimal system performance with SCHED_ULE, which is
370 cpumask_t type is retired and cpuset_t is used in order to describe
374 Changes to ifconfig(8) for dynamic address family detection mandate
375 that you are running a kernel of 20110525 or later. Make sure to
376 follow the update procedure to boot a new kernel before installing
380 Support for sun4v architecture is officially dropped
383 Several KPI breaking changes have been committed to the mii(4) layer,
384 the PHY drivers and consequently some Ethernet drivers using mii(4).
385 This means that miibus.ko and the modules of the affected Ethernet
386 drivers need to be recompiled.
388 Note to kernel developers: Given that the OUI bit reversion problem
389 was fixed as part of these changes all mii(4) commits related to OUIs,
390 i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
391 handling, no longer can be merged verbatim to stable/8 and previous
395 Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
396 into their kernel configurations along with 'device pci'.
399 The default NFS client is now the new NFS client, so fstype "newnfs"
400 is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
401 Although mounts via fstype "nfs" will usually work without userland
402 changes, it is recommended that the mount(8) and mount_nfs(8)
403 commands be rebuilt from sources and that a link to mount_nfs called
404 mount_oldnfs be created. The new client is compiled into the
405 kernel with "options NFSCL" and this is needed for diskless root
406 file systems. The GENERIC kernel configs have been changed to use
407 NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
408 To use the regular/old client, you can "mount -t oldnfs ...". For
409 a diskless root file system, you must also include a line like:
411 vfs.root.mountfrom="oldnfs:"
413 in the boot/loader.conf on the root fs on the NFS server to make
414 a diskless root fs use the old client.
417 The GENERIC kernels for all architectures now default to the new
418 CAM-based ATA stack. It means that all legacy ATA drivers were
419 removed and replaced by respective CAM drivers. If you are using
420 ATA device names in /etc/fstab or other places, make sure to update
421 them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
422 where 'Y's are the sequential numbers starting from zero for each type
423 in order of detection, unless configured otherwise with tunables,
424 see cam(4)). There will be symbolic links created in /dev/ to map
425 old adX devices to the respective adaY. They should provide basic
426 compatibility for file systems mounting in most cases, but they do
427 not support old user-level APIs and do not have respective providers
428 in GEOM. Consider using updated management tools with new device names.
430 It is possible to load devices ahci, ata, siis and mvs as modules,
431 but option ATA_CAM should remain in kernel configuration to make ata
432 module work as CAM driver supporting legacy ATA controllers. Device ata
433 still can be used in modular fashion (atacore + ...). Modules atadisk
434 and atapi* are not used and won't affect operation in ATA_CAM mode.
435 Note that to use CAM-based ATA kernel should include CAM devices
436 scbus, pass, da (or explicitly ada), cd and optionally others. All of
437 them are parts of the cam module.
439 ataraid(4) functionality is now supported by the RAID GEOM class.
440 To use it you can load geom_raid kernel module and use graid(8) tool
441 for management. Instead of /dev/arX device names, use /dev/raid/rX.
443 No kernel config options or code have been removed, so if a problem
444 arises, please report it and optionally revert to the old ATA stack.
445 In order to do it you can remove from the kernel config:
450 , and instead add back:
451 device atadisk # ATA disk drives
452 device ataraid # ATA RAID drives
453 device atapicd # ATAPI CDROM drives
454 device atapifd # ATAPI floppy drives
455 device atapist # ATAPI tape drives
458 The default NFS server has been changed to the new server, which
459 was referred to as the experimental server. If you need to switch
460 back to the old NFS server, you must now put the "-o" option on
461 both the mountd and nfsd commands. This can be done using the
462 mountd_flags and nfs_server_flags rc.conf variables until an
463 update to the rc scripts is committed, which is coming soon.
466 The GNU Objective-C runtime library (libobjc), and other Objective-C
467 related components have been removed from the base system. If you
468 require an Objective-C library, please use one of the available ports.
471 ath(4) has been split into bus- and device- modules. if_ath contains
472 the HAL, the TX rate control and the network device code. if_ath_pci
473 contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
474 contains the AHB glue. Users need to load both if_ath_pci and if_ath
475 in order to use ath on everything else.
477 TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
478 need to load if_ath and if_ath_pci for ath(4) operation.
481 As part of the replacement of sysinstall, the process of building
482 release media has changed significantly. For details, please re-read
483 release(7), which has been updated to reflect the new build process.
486 GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
487 is the last available version under GPLv2. It brings a number of new
488 features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
489 4.1 and SSE 4.2), better support for powerpc64, a number of new
490 directives, and lots of other small improvements. See the ChangeLog
491 file in contrib/binutils for the full details.
494 IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
495 compliant, and will now use half of hash for authentication.
496 This will break interoperability with all stacks (including all
497 actual FreeBSD versions) who implement
498 draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
500 The only workaround with such peers is to use another HMAC
501 algorithm for IPsec ("phase 2") authentication.
504 Remove the uio_yield prototype and symbol. This function has
505 been misnamed since it was introduced and should not be
506 globally exposed with this name. The equivalent functionality
507 is now available using kern_yield(curthread->td_user_pri).
508 The function remains undocumented.
511 A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
512 symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
513 for scalar sysctls is defined but disabled. Code that needs
514 UQUAD to pass the type checking that must compile on older
515 systems where the define is not present can check against
516 __FreeBSD_version >= 900030.
518 The system dialog(1) has been replaced with a new version previously
519 in ports as devel/cdialog. dialog(1) is mostly command-line compatible
520 with the previous version, but the libdialog associated with it has
521 a largely incompatible API. As such, the original version of libdialog
522 will be kept temporarily as libodialog, until its base system consumers
523 are replaced or updated. Bump __FreeBSD_version to 900030.
526 If you are trying to run make universe on a -stable system, and you get
527 the following warning:
528 "Makefile", line 356: "Target architecture for i386/conf/GENERIC
529 unknown. config(8) likely too old."
530 or something similar to it, then you must upgrade your -stable system
531 to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
532 stable/8) or build the config from the latest stable/8 branch and
533 install it on your system.
535 Prior to this date, building a current universe on 8-stable system from
536 between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
537 error in the first kernel build phase. A new config on those old
538 systems will fix that problem for older versions of -current.
541 The TCP stack has been modified to allow Khelp modules to interact with
542 it via helper hook points and store per-connection data in the TCP
543 control block. Bump __FreeBSD_version to 900029. User space tools that
544 rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
548 Generic IEEE 802.3 annex 31B full duplex flow control support has been
549 added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
550 with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
551 to take advantage of it instead of using custom implementations. This
552 means that these drivers now no longer unconditionally advertise
553 support for flow control but only do so if flow control is a selected
554 media option. This was implemented in the generic support that way in
555 order to allow flow control to be switched on and off via ifconfig(8)
556 with the PHY specific default to typically off in order to protect
557 from unwanted effects. Consequently, if you used flow control with
558 one of the above mentioned drivers you now need to explicitly enable
560 ifconfig bge0 media auto mediaopt flowcontrol
562 Along with the above mentioned changes generic support for setting
563 1000baseT master mode also has been added and brgphy(4), ciphy(4),
564 e1000phy(4) as well as ip1000phy(4) have been converted to take
565 advantage of it. This means that these drivers now no longer take the
566 link0 parameter for selecting master mode but the master media option
567 has to be used instead, for example like in the following:
568 ifconfig bge0 media 1000baseT mediaopt full-duplex,master
570 Selection of master mode now is also available with all other PHY
571 drivers supporting 1000baseT.
574 The TCP stack has received a significant update to add support for
575 modularised congestion control and generally improve the clarity of
576 congestion control decisions. Bump __FreeBSD_version to 900025. User
577 space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
578 sockstat) need to be recompiled.
581 The man(1) utility has been replaced by a new version that no longer
582 uses /etc/manpath.config. Please consult man.conf(5) for how to
583 migrate local entries to the new format.
586 The copyright strings printed by login(1) and sshd(8) at the time of a
587 new connection have been removed to follow other operating systems and
591 A workaround for a fixed ld bug has been removed in kernel code,
592 so make sure that your system ld is built from sources after
593 revision 210245 from 2010-07-19 (r211583 if building head kernel
594 on stable/8, r211584 for stable/7; both from 2010-08-21).
595 A symptom of incorrect ld version is different addresses for
596 set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
599 The $ipv6_prefer variable in rc.conf(5) has been split into
600 $ip6addrctl_policy and $ipv6_activate_all_interfaces.
602 The $ip6addrctl_policy is a variable to choose a pre-defined
603 address selection policy set by ip6addrctl(8). A value
604 "ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
607 The $ipv6_activate_all_interfaces specifies whether IFDISABLED
608 flag (see an entry of 20090926) is set on an interface with no
609 corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
610 security reason. If you want IPv6 link-local address on all
611 interfaces by default, set this to "YES".
613 The old ipv6_prefer="YES" is equivalent to
614 ipv6_activate_all_interfaces="YES" and
615 ip6addrctl_policy="ipv6_prefer".
618 DTrace has grown support for userland tracing. Due to this, DTrace is
619 now i386 and amd64 only.
620 dtruss(1) is now installed by default on those systems and a new
621 kernel module is needed for userland tracing: fasttrap.
622 No changes to your kernel config file are necessary to enable
623 userland tracing, but you might consider adding 'STRIP=' and
624 'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
625 to have informative userland stack traces in DTrace (ustack).
628 The acpi_aiboost(4) driver has been removed in favor of the new
629 aibs(4) driver. You should update your kernel configuration file.
632 BSD grep has been imported to the base system and it is built by
633 default. It is completely BSD licensed, highly GNU-compatible, uses
634 less memory than its GNU counterpart and has a small codebase.
635 However, it is slower than its GNU counterpart, which is mostly
636 noticeable for larger searches, for smaller ones it is measurable
637 but not significant. The reason is complex, the most important factor
638 is that we lack a modern and efficient regex library and GNU
639 overcomes this by optimizing the searches internally. Future work
640 on improving the regex performance is planned, for the meantime,
641 users that need better performance, can build GNU grep instead by
642 setting the WITH_GNU_GREP knob.
645 Due to the import of powerpc64 support, all existing powerpc kernel
646 configuration files must be updated with a machine directive like this:
647 machine powerpc powerpc
649 In addition, an updated config(8) is required to build powerpc kernels
653 A new version of ZFS (version 15) has been merged to -HEAD.
654 This version uses a python library for the following subcommands:
655 zfs allow, zfs unallow, zfs groupspace, zfs userspace.
656 For full functionality of these commands the following port must
657 be installed: sysutils/py-zfs
660 'vm_page's are now hashed by physical address to an array of mutexes.
661 Currently this is only used to serialize access to hold_count. Over
662 time the page queue mutex will be peeled away. This changes the size
663 of pmap on every architecture. And requires all callers of vm_page_hold
664 and vm_page_unhold to be updated.
667 WITH_CTF can now be specified in src.conf (not recommended, there
668 are some problems with static executables), make.conf (would also
669 affect ports which do not use GNU make and do not override the
670 compile targets) or in the kernel config (via "makeoptions
672 When WITH_CTF was specified there before this was silently ignored,
673 so make sure that WITH_CTF is not used in places which could lead
674 to unwanted behavior.
677 The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
678 to allow 32-bit compatibility on non-x86 platforms. All kernel
679 configurations on amd64 and ia64 platforms using these options must
680 be modified accordingly.
683 The utmp user accounting database has been replaced with utmpx,
684 the user accounting interface standardized by POSIX.
685 Unfortunately the semantics of utmp and utmpx don't match,
686 making it practically impossible to support both interfaces.
687 The user accounting database is used by tools like finger(1),
688 last(1), talk(1), w(1) and ac(8).
690 All applications in the base system use utmpx. This means only
691 local binaries (e.g. from the ports tree) may still use these
692 utmp database files. These applications must be rebuilt to make
695 After the system has been upgraded, it is safe to remove the old
696 log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
697 assuming their contents is of no importance anymore. Old wtmp
698 databases can only be used by last(1) and ac(8) after they have
699 been converted to the new format using wtmpcvt(1).
702 Introduce the kernel thread "deadlock resolver" (which can be enabled
703 via the DEADLKRES option, see NOTES for more details) and the
704 sleepq_type() function for sleepqueues.
707 The rc.firewall and rc.firewall6 were unified, and
708 rc.firewall6 and rc.d/ip6fw were removed.
709 According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
710 variables are obsoleted. Instead, the following new rc
711 variables are added to rc.d/ipfw:
713 firewall_client_net_ipv6, firewall_simple_iif_ipv6,
714 firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
715 firewall_simple_onet_ipv6, firewall_trusted_ipv6
717 The meanings correspond to the relevant IPv4 variables.
723 The default terminal emulation for syscons(4) has been changed
724 from cons25 to xterm on all platforms except pc98. This means
725 that the /etc/ttys file needs to be updated to ensure correct
726 operation of applications on the console.
728 The terminal emulation style can be toggled per window by using
729 vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
730 options can be used to change the compile-time default back to
733 To prevent graphical artifacts, make sure the TERM environment
734 variable is set to match the terminal emulation that is being
735 performed by syscons(4).
738 The layout of the structure ieee80211req_scan_result has changed.
739 Applications that require wireless scan results (e.g. ifconfig(8))
740 from net80211 need to be recompiled.
742 Applications such as wpa_supplicant(8) may require a full world
743 build without using NO_CLEAN in order to get synchronized with the
747 The iwn(4) driver has been updated to support the 5000 and 5150 series.
748 There's one kernel module for each firmware. Adding "device iwnfw"
749 to the kernel configuration file means including all three firmware
750 images inside the kernel. If you want to include just the one for
751 your wireless card, use the devices iwn4965fw, iwn5000fw or
755 The rc.d/network_ipv6, IPv6 configuration script has been integrated
756 into rc.d/netif. The changes are the following:
758 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
759 for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
760 Note that both variables need the "inet6" keyword at the head.
762 Do not set $ipv6_network_interfaces manually if you do not
763 understand what you are doing. It is not needed in most cases.
765 $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
768 2. $ipv6_enable is obsolete. Use $ipv6_prefer and
769 "inet6 accept_rtadv" keyword in ifconfig(8) instead.
771 If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
772 all configured interfaces have "inet6 accept_rtadv" in the
773 $ifconfig_IF_ipv6. These are for backward compatibility.
775 3. A new variable $ipv6_prefer has been added. If NO, IPv6
776 functionality of interfaces with no corresponding
777 $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
778 and the default address selection policy of ip6addrctl(8)
779 is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
780 Note that if you want to configure IPv6 functionality on the
781 disabled interfaces after boot, first you need to clear the flag by
782 using ifconfig(8) like:
784 ifconfig em0 inet6 -ifdisabled
786 If YES, the default address selection policy is set as
789 The default value of $ipv6_prefer is NO.
791 4. If your system need to receive Router Advertisement messages,
792 define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
793 scripts automatically invoke rtsol(8) when the interface becomes
794 UP. The Router Advertisement messages are used for SLAAC
795 (State-Less Address AutoConfiguration).
798 802.11s D3.03 support was committed. This is incompatible with the
799 previous code, which was based on D3.0.
802 A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
803 of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
804 control whether accepting Router Advertisement messages or not.
805 Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
806 a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
807 The ifconfig(8) utility now supports these flags.
810 ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
811 mount(8) and -a option for df(1) to see them.
814 The old tunable hw.bus.devctl_disable has been superseded by
815 hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
816 replaced by hw.bus.devctl_queue=0. The default for this new tunable
820 Remove the option STOP_NMI. The default action is now to use NMI only
821 for KDB via the newly introduced function stop_cpus_hard() and
822 maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
825 The stable/8 branch created in subversion. This corresponds to the
826 RELENG_8 branch in CVS.
829 Bump the shared library version numbers for all libraries that do not
830 use symbol versioning as part of the 8.0-RELEASE cycle. Bump
831 __FreeBSD_version to 800105.
834 Due to changes in the implementation of virtual network stack support,
835 all network-related kernel modules must be recompiled. As this change
836 breaks the ABI, bump __FreeBSD_version to 800104.
839 The TOE interface to the TCP syncache has been modified to remove
840 struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
841 The cxgb driver is the only TOE consumer affected by this change, and
842 needs to be recompiled along with the kernel. As this change breaks
843 the ABI, bump __FreeBSD_version to 800103.
846 Padding has been added to struct tcpcb, sackhint and tcpstat in
847 <netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
848 maintaining the ABI. However, this change breaks the ABI, so bump
849 __FreeBSD_version to 800102. User space tools that rely on the size of
850 any of these structs (e.g. sockstat) need to be recompiled.
853 The NFS_LEGACYRPC option has been removed along with the old kernel
854 RPC implementation that this option selected. Kernel configurations
855 may need to be adjusted.
858 The network interface device nodes at /dev/net/<interface> have been
859 removed. All ioctl operations can be performed the normal way using
860 routing sockets. The kqueue functionality can generally be replaced
861 with routing sockets.
864 The documentation from the FreeBSD Documentation Project (Handbook,
865 FAQ, etc.) is now installed via packages by sysinstall(8) and under
866 the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
869 The ABI of various structures related to the SYSV IPC API have been
870 changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
871 options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
875 Layout of struct vnet has changed as routing related variables were
876 moved to their own Vimage module. Modules need to be recompiled. Bump
877 __FreeBSD_version to 800099.
880 NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
881 respectively. As long as no more than 16 groups per process are used,
882 no changes should be visible. When more than 16 groups are used, old
883 binaries may fail if they call getgroups() or getgrouplist() with
884 statically sized storage. Recompiling will work around this, but
885 applications should be modified to use dynamically allocated storage
886 for group arrays as POSIX.1-2008 does not cap an implementation's
887 number of supported groups at NGROUPS_MAX+1 as previous versions did.
889 NFS and portalfs mounts may also be affected as the list of groups is
890 truncated to 16. Users of NFS who use more than 16 groups, should
891 take care that negative group permissions are not used on the exported
892 file systems as they will not be reliable unless a GSSAPI based
893 authentication method is used.
896 The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
897 option compiles in the support for adaptive spinning for lockmgrs
898 which want to enable it. The lockinit() function now accepts the flag
899 LK_ADAPTIVE in order to make the lock object subject to adaptive
900 spinning when both held in write and read mode.
903 The layout of the structure returned by IEEE80211_IOC_STA_INFO has
904 changed. User applications that use this ioctl need to be rebuilt.
907 The layout of struct thread has changed. Kernel and modules need to
911 The layout of structs ifnet, domain, protosw and vnet_net has changed.
912 Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
915 window(1) has been removed from the base system. It can now be
916 installed from ports. The port is called misc/window.
919 The way we are storing and accessing `routing table' entries has
920 changed. Programs reading the FIB, like netstat, need to be
924 A new netisr implementation has been added for FreeBSD 8. Network
925 file system modules, such as igmp, ipdivert, and others, should be
927 Bump __FreeBSD_version to 800096.
930 Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
934 Add VOP_ACCESSX(9). File system modules need to be rebuilt.
935 Bump __FreeBSD_version to 800094.
938 Add mnt_xflag field to 'struct mount'. File system modules need to be
940 Bump __FreeBSD_version to 800093.
943 The compiling option ADAPTIVE_SX has been retired while it has been
944 introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
945 The KPI for sx_init_flags() changes as accepting flags:
946 SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
947 been introduced in order to handle the reversed logic.
948 Bump __FreeBSD_version to 800092.
951 Add support for hierarchical jails. Remove global securelevel.
952 Bump __FreeBSD_version to 800091.
955 The layout of struct vnet_net has changed, therefore modules
957 Bump __FreeBSD_version to 800090.
960 The newly imported zic(8) produces a new format in the output. Please
961 run tzsetup(8) to install the newly created data to /etc/localtime.
964 The sysctl tree for the usb stack has renamed from hw.usb2.* to
965 hw.usb.* and is now consistent again with previous releases.
968 802.11 monitor mode support was revised and driver api's were changed.
969 Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
970 of DLT_IEEE802_11. No user-visible data structures were changed but
971 applications that use DLT_IEEE802_11 may require changes.
972 Bump __FreeBSD_version to 800088.
975 The layout of the following structs has changed: sysctl_oid,
976 socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
977 vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
978 panics may be experienced. World rebuild is required for
979 correctly checking networking state from userland.
980 Bump __FreeBSD_version to 800085.
983 MLDv2 and Source-Specific Multicast (SSM) have been merged
984 to the IPv6 stack. VIMAGE hooks are in but not yet used.
985 The implementation of SSM within FreeBSD's IPv6 stack closely
986 follows the IPv4 implementation.
988 For kernel developers:
990 * The most important changes are that the ip6_output() and
991 ip6_input() paths no longer take the IN6_MULTI_LOCK,
992 and this lock has been downgraded to a non-recursive mutex.
994 * As with the changes to the IPv4 stack to support SSM, filtering
995 of inbound multicast traffic must now be performed by transport
996 protocols within the IPv6 stack. This does not apply to TCP and
997 SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
999 * The KPIs used by IPv6 multicast are similar to those used by
1000 the IPv4 stack, with the following differences:
1001 * im6o_mc_filter() is analogous to imo_multicast_filter().
1002 * The legacy KAME entry points in6_joingroup and in6_leavegroup()
1003 are shimmed to in6_mc_join() and in6_mc_leave() respectively.
1004 * IN6_LOOKUP_MULTI() has been deprecated and removed.
1005 * IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
1006 for MLDv1 have an additional 'timer' argument which is used to
1007 jitter the initial membership report for the solicited-node
1008 multicast membership on-link.
1009 * This is not strictly needed for MLDv2, which already jitters
1010 its report transmissions. However, the 'timer' argument is
1011 preserved in case MLDv1 is active on the interface.
1013 * The KAME linked-list based IPv6 membership implementation has
1014 been refactored to use a vector similar to that used by the IPv4
1016 Code which maintains a list of its own multicast memberships
1017 internally, e.g. carp, has been updated to reflect the new
1020 * There is a known Lock Order Reversal (LOR) due to in6_setscope()
1021 acquiring the IF_AFDATA_LOCK and being called within ip6_output().
1022 Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
1023 implementation constraint which needs to be addressed in HEAD.
1025 For application developers:
1027 * The changes are broadly similar to those made for the IPv4
1030 * The use of IPv4 and IPv6 multicast socket options on the same
1031 socket, using mapped addresses, HAS NOT been tested or supported.
1033 * There are a number of issues with the implementation of various
1034 IPv6 multicast APIs which need to be resolved in the API surface
1035 before the implementation is fully compatible with KAME userland
1036 use, and these are mostly to do with interface index treatment.
1038 * The literature available discusses the use of either the delta / ASM
1039 API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
1040 using setsourcefilter(3)/getsourcefilter(3). For more information
1041 please refer to RFC 3768, 'Socket Interface Extensions for
1042 Multicast Source Filters'.
1044 * Applications which use the published RFC 3678 APIs should be fine.
1046 For systems administrators:
1048 * The mtest(8) utility has been refactored to support IPv6, in
1049 addition to IPv4. Interface addresses are no longer accepted
1050 as arguments, their names must be used instead. The utility
1051 will map the interface name to its first IPv4 address as
1052 returned by getifaddrs(3).
1054 * The ifmcstat(8) utility has also been updated to print the MLDv2
1055 endpoint state and source filter lists via sysctl(3).
1057 * The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
1058 loopback of IPv6 multicast datagrams by default; it defaults to 1
1059 to preserve the existing behaviour. Disabling multicast loopback is
1060 recommended for optimal system performance.
1062 * The IPv6 MROUTING code has been changed to examine this sysctl
1063 instead of attempting to perform a group lookup before looping
1064 back forwarded datagrams.
1066 Bump __FreeBSD_version to 800084.
1069 Implement low-level Bluetooth HCI API.
1070 Bump __FreeBSD_version to 800083.
1073 The layout of struct malloc_type, used by modules to register new
1074 memory allocation types, has changed. Most modules will need to
1075 be rebuilt or panics may be experienced.
1076 Bump __FreeBSD_version to 800081.
1079 Anticipate overflowing inp_flags - add inp_flags2.
1080 This changes most offsets in inpcb, so checking v4 connection
1081 state will require a world rebuild.
1082 Bump __FreeBSD_version to 800080.
1085 Add an llentry to struct route and struct route_in6. Modules
1086 embedding a struct route will need to be recompiled.
1087 Bump __FreeBSD_version to 800079.
1090 The size of rt_metrics_lite and by extension rtentry has changed.
1091 Networking administration apps will need to be recompiled.
1092 The route command now supports show as an alias for get, weighting
1093 of routes, sticky and nostick flags to alter the behavior of stateful
1095 Bump __FreeBSD_version to 800078.
1098 Do not use Giant for kbdmux(4) locking. This is wrong and
1099 apparently causing more problems than it solves. This will
1100 re-open the issue where interrupt handlers may race with
1101 kbdmux(4) in polling mode. Typical symptoms include (but
1102 not limited to) duplicated and/or missing characters when
1103 low level console functions (such as gets) are used while
1104 interrupts are enabled (for example geli password prompt,
1105 mountroot prompt etc.). Disabling kbdmux(4) may help.
1108 The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
1109 kernel modules referencing any of the above need to be recompiled.
1110 Bump __FreeBSD_version to 800075.
1113 GEOM_PART has become the default partition slicer for storage devices,
1114 replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
1115 introduces some changes:
1117 MSDOS/EBR: the devices created from MSDOS extended partition entries
1118 (EBR) can be named differently than with GEOM_MBR and are now symlinks
1119 to devices with offset-based names. fstabs may need to be modified.
1121 BSD: the "geometry does not match label" warning is harmless in most
1122 cases but it points to problems in file system misalignment with
1123 disk geometry. The "c" partition is now implicit, covers the whole
1124 top-level drive and cannot be (mis)used by users.
1126 General: Kernel dumps are now not allowed to be written to devices
1127 whose partition types indicate they are meant to be used for file
1128 systems (or, in case of MSDOS partitions, as something else than
1131 Most of these changes date approximately from 200812.
1134 The uscanner(4) driver has been removed from the kernel. This follows
1135 Linux removing theirs in 2.6 and making libusb the default interface
1136 (supported by sane).
1139 The multicast forwarding code has been cleaned up. netstat(1)
1140 only relies on KVM now for printing bandwidth upcall meters.
1141 The IPv4 and IPv6 modules are split into ip_mroute_mod and
1142 ip6_mroute_mod respectively. The config(5) options for statically
1143 compiling this code remain the same, i.e. 'options MROUTING'.
1146 Support for the IFF_NEEDSGIANT network interface flag has been
1147 removed, which means that non-MPSAFE network device drivers are no
1148 longer supported. In particular, if_ar, if_sr, and network device
1149 drivers from the old (legacy) USB stack can no longer be built or
1153 POSIX.1 Native Language Support (NLS) has been enabled in libc and
1154 a bunch of new language catalog files have also been added.
1155 This means that some common libc messages are now localized and
1156 they depend on the LC_MESSAGES environmental variable.
1159 The k8temp(4) driver has been renamed to amdtemp(4) since
1160 support for Family 10 and Family 11 CPU families was added.
1163 IGMPv3 and Source-Specific Multicast (SSM) have been merged
1164 to the IPv4 stack. VIMAGE hooks are in but not yet used.
1166 For kernel developers, the most important changes are that the
1167 ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
1168 and this lock has been downgraded to a non-recursive mutex.
1170 Transport protocols (UDP, Raw IP) are now responsible for filtering
1171 inbound multicast traffic according to group membership and source
1172 filters. The imo_multicast_filter() KPI exists for this purpose.
1173 Transports which do not use multicast (SCTP, TCP) already reject
1174 multicast by default. Forwarding and receive performance may improve
1175 as a mutex acquisition is no longer needed in the ip_input()
1176 low-level input path. in_addmulti() and in_delmulti() are shimmed
1177 to new KPIs which exist to support SSM in-kernel.
1179 For application developers, it is recommended that loopback of
1180 multicast datagrams be disabled for best performance, as this
1181 will still cause the lock to be taken for each looped-back
1182 datagram transmission. The net.inet.ip.mcast.loop sysctl may
1183 be tuned to 0 to disable loopback by default; it defaults to 1
1184 to preserve the existing behaviour.
1186 For systems administrators, to obtain best performance with
1187 multicast reception and multiple groups, it is always recommended
1188 that a card with a suitably precise hash filter is used. Hash
1189 collisions will still result in the lock being taken within the
1190 transport protocol input path to check group membership.
1192 If deploying FreeBSD in an environment with IGMP snooping switches,
1193 it is recommended that the net.inet.igmp.sendlocal sysctl remain
1194 enabled; this forces 224.0.0.0/24 group membership to be announced
1197 The size of 'struct igmpstat' has changed; netstat needs to be
1198 recompiled to reflect this.
1199 Bump __FreeBSD_version to 800070.
1202 libusb20.so.1 is now installed as libusb.so.1 and the ports system
1203 updated to use it. This requires a buildworld/installworld in order to
1204 update the library and dependencies (usbconfig, etc). Its advisable to
1205 rebuild all ports which uses libusb. More specific directions are given
1206 in the ports collection UPDATING file. Any /etc/libmap.conf entries for
1207 libusb are no longer required and can be removed.
1210 A workaround is committed to allow the creation of System V shared
1211 memory segment of size > 2 GB on the 64-bit architectures.
1212 Due to a limitation of the existing ABI, the shm_segsz member
1213 of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
1214 wrong for large segments. Note that limits must be explicitly
1215 raised to allow such segments to be created.
1218 The layout of struct ifnet has changed, requiring a rebuild of all
1219 network device driver modules.
1222 The /dev handling for the new USB stack has changed, a
1223 buildworld/installworld is required for libusb20.
1226 The new USB2 stack has now been permanently moved in and all kernel and
1227 module names reverted to their previous values (eg, usb, ehci, ohci,
1228 ums, ...). The old usb stack can be compiled in by prefixing the name
1229 with the letter 'o', the old usb modules have been removed.
1230 Updating entry 20090216 for xorg and 20090215 for libmap may still
1234 The rc.conf(5) option if_up_delay has been renamed to
1235 defaultroute_delay to better reflect its purpose. If you have
1236 customized this setting in /etc/rc.conf you need to update it to
1240 xorg 7.4 wants to configure its input devices via hald which does not
1241 yet work with USB2. If the keyboard/mouse does not work in xorg then
1243 Option "AllowEmptyInput" "off"
1244 to your ServerLayout section. This will cause X to use the configured
1245 kbd and mouse sections from your xorg.conf.
1248 The GENERIC kernels for all architectures now default to the new USB2
1249 stack. No kernel config options or code have been removed so if a
1250 problem arises please report it and optionally revert to the old USB
1251 stack. If you are loading USB kernel modules or have a custom kernel
1252 that includes GENERIC then ensure that usb names are also changed over,
1253 eg uftdi -> usb2_serial_ftdi.
1255 Older programs linked against the ports libusb 0.1 need to be
1256 redirected to the new stack's libusb20. /etc/libmap.conf can
1258 # Map old usb library to new one for usb2 stack
1259 libusb-0.1.so.8 libusb20.so.1
1262 All USB ethernet devices now attach as interfaces under the name ueN
1263 (eg. ue0). This is to provide a predictable name as vendors often
1264 change usb chipsets in a product without notice.
1267 The ichsmb(4) driver has been changed to require SMBus slave
1268 addresses be left-justified (xxxxxxx0b) rather than right-justified.
1269 All of the other SMBus controller drivers require left-justified
1270 slave addresses, so this change makes all the drivers provide the
1274 INET6 statistics (struct ip6stat) was updated.
1275 netstat(1) needs to be recompiled.
1278 NTFS has been removed from GENERIC kernel on amd64 to match
1279 GENERIC on i386. Should not cause any issues since mount_ntfs(8)
1280 will load ntfs.ko module automatically when NTFS support is
1281 actually needed, unless ntfs.ko is not installed or security
1282 level prohibits loading kernel modules. If either is the case,
1283 "options NTFS" has to be added into kernel config.
1286 TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
1287 New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
1288 800061. User space tools that rely on the size of struct tcpcb in
1289 tcp_var.h (e.g. sockstat) need to be recompiled.
1292 ng_tty(4) module updated to match the new TTY subsystem.
1293 Due to API change, user-level applications must be updated.
1294 New API support added to mpd5 CVS and expected to be present
1295 in next mpd5.3 release.
1298 With __FreeBSD_version 800060 the makefs tool is part of
1299 the base system (it was a port).
1302 The afdata and ifnet locks have been changed from mutexes to
1303 rwlocks, network modules will need to be re-compiled.
1306 __FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
1307 RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
1308 The new code reduced struct rtentry{} by 16 bytes on 32-bit
1309 architecture and 40 bytes on 64-bit architecture. The userland
1310 applications "arp" and "ndp" have been updated accordingly.
1311 The output from "netstat -r" shows only routing entries and
1312 none of the L2 information.
1315 __FreeBSD_version 800057 marks the switchover from the
1316 binary ath hal to source code. Users must add the line:
1318 options AH_SUPPORT_AR5416
1320 to their kernel config files when specifying:
1324 The ath_hal module no longer exists; the code is now compiled
1325 together with the driver in the ath module. It is now
1326 possible to tailor chip support (i.e. reduce the set of chips
1327 and thereby the code size); consult ath_hal(4) for details.
1330 __FreeBSD_version 800054 adds memory barriers to
1331 <machine/atomic.h>, new interfaces to ifnet to facilitate
1332 multiple hardware transmit queues for cards that support
1333 them, and a lock-less ring-buffer implementation to
1334 enable drivers to more efficiently manage queueing of
1338 A new version of ZFS (version 13) has been merged to -HEAD.
1339 This version has zpool attribute "listsnapshots" off by
1340 default, which means "zfs list" does not show snapshots,
1341 and is the same as Solaris behavior.
1344 dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
1347 The uhci, ohci, ehci and slhci USB Host controller drivers have
1348 been put into separate modules. If you load the usb module
1349 separately through loader.conf you will need to load the
1350 appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
1351 controller add the following to loader.conf:
1357 The ABI used by the PMC toolset has changed. Please keep
1358 userland (libpmc(3)) and the kernel module (hwpmc(4)) in
1362 atapci kernel module now includes only generic PCI ATA
1363 driver. AHCI driver moved to ataahci kernel module.
1364 All vendor-specific code moved into separate kernel modules:
1365 ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
1366 atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
1367 atamarvell, atamicron, atanational, atanetcell, atanvidia,
1368 atapromise, ataserverworks, atasiliconimage, atasis, atavia
1371 The TTY subsystem of the kernel has been replaced by a new
1372 implementation, which provides better scalability and an
1373 improved driver model. Most common drivers have been migrated to
1374 the new TTY subsystem, while others have not. The following
1375 drivers have not yet been ported to the new TTY layer:
1378 cy, digi, rc, rp, sio
1384 ng_h4, ng_tty, ppp, sl, snp
1386 Adding these drivers to your kernel configuration file shall
1387 cause compilation to fail.
1390 ntpd has been upgraded to 4.2.4p5.
1393 OpenSSH has been upgraded to 5.1p1.
1395 For many years, FreeBSD's version of OpenSSH preferred DSA
1396 over RSA for host and user authentication keys. With this
1397 upgrade, we've switched to the vendor's default of RSA over
1398 DSA. This may cause upgraded clients to warn about unknown
1399 host keys even for previously known hosts. Users should
1400 follow the usual procedure for verifying host keys before
1401 accepting the RSA key.
1403 This can be circumvented by setting the "HostKeyAlgorithms"
1404 option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
1407 Please note that the sequence of keys offered for
1408 authentication has been changed as well. You may want to
1409 specify IdentityFile in a different order to revert this
1413 The sio(4) driver has been removed from the i386 and amd64
1414 kernel configuration files. This means uart(4) is now the
1415 default serial port driver on those platforms as well.
1417 To prevent collisions with the sio(4) driver, the uart(4) driver
1418 uses different names for its device nodes. This means the
1419 onboard serial port will now most likely be called "ttyu0"
1420 instead of "ttyd0". You may need to reconfigure applications to
1421 use the new device names.
1423 When using the serial port as a boot console, be sure to update
1424 /boot/device.hints and /etc/ttys before booting the new kernel.
1425 If you forget to do so, you can still manually specify the hints
1426 at the loader prompt:
1428 set hint.uart.0.at="isa"
1429 set hint.uart.0.port="0x3F8"
1430 set hint.uart.0.flags="0x10"
1431 set hint.uart.0.irq="4"
1435 The gpt(8) utility has been removed. Use gpart(8) to partition
1439 The version that Linuxulator emulates was changed from 2.4.2
1440 to 2.6.16. If you experience any problems with Linux binaries
1441 please try to set sysctl compat.linux.osrelease to 2.4.2 and
1442 if it fixes the problem contact emulation mailing list.
1445 ISDN4BSD (I4B) was removed from the src tree. You may need to
1446 update a your kernel configuration and remove relevant entries.
1449 I have checked in code to support multiple routing tables.
1450 See the man pages setfib(1) and setfib(2).
1451 This is a hopefully backwards compatible version,
1452 but to make use of it you need to compile your kernel
1453 with options ROUTETABLES=2 (or more up to 16).
1456 The 802.11 wireless support was redone to enable multi-bss
1457 operation on devices that are capable. The underlying device
1458 is no longer used directly but instead wlanX devices are
1459 cloned with ifconfig. This requires changes to rc.conf files.
1460 For example, change:
1461 ifconfig_ath0="WPA DHCP"
1464 ifconfig_wlan0="WPA DHCP"
1465 see rc.conf(5) for more details. In addition, mergemaster of
1466 /etc/rc.d is highly recommended. Simultaneous update of userland
1467 and kernel wouldn't hurt either.
1469 As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
1470 modules were merged into the base wlan module. All references
1471 to these modules (e.g. in kernel config files) must be removed.
1474 psm(4) has gained write(2) support in native operation level.
1475 Arbitrary commands can be written to /dev/psm%d and status can
1476 be read back from it. Therefore, an application is responsible
1477 for status validation and error recovery. It is a no-op in
1478 other operation levels.
1481 Support for KSE threading has been removed from the kernel. To
1482 run legacy applications linked against KSE libmap.conf may
1483 be used. The following libmap.conf may be used to ensure
1484 compatibility with any prior release:
1486 libpthread.so.1 libthr.so.1
1487 libpthread.so.2 libthr.so.2
1488 libkse.so.3 libthr.so.3
1491 The layout of struct vmspace has changed. This affects libkvm
1492 and any executables that link against libkvm and use the
1493 kvm_getprocs() function. In particular, but not exclusively,
1494 it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
1495 The effects are minimal, but it's advisable to upgrade world
1499 The latest em driver no longer has support in it for the
1500 82575 adapter, this is now moved to the igb driver. The
1501 split was done to make new features that are incompatible
1502 with older hardware easier to do.
1505 The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
1506 likewise the kernel option is now GEOM_LINUX_LVM.
1509 The default NFS mount mode has changed from UDP to TCP for
1510 increased reliability. If you rely on (insecurely) NFS
1511 mounting across a firewall you may need to update your
1515 Belatedly note the addition of m_collapse for compacting
1519 The fts(3) structures have been changed to use adequate
1520 integer types for their members and so to be able to cope
1521 with huge file trees. The old fts(3) ABI is preserved
1522 through symbol versioning in libc, so third-party binaries
1523 using fts(3) should still work, although they will not take
1524 advantage of the extended types. At the same time, some
1525 third-party software might fail to build after this change
1526 due to unportable assumptions made in its source code about
1527 fts(3) structure members. Such software should be fixed
1528 by its vendor or, in the worst case, in the ports tree.
1529 FreeBSD_version 800015 marks this change for the unlikely
1530 case that a portable fix is impossible.
1533 To upgrade to -current after this date, you must be running
1534 FreeBSD not older than 6.0-RELEASE. Upgrading to -current
1535 from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
1538 The ADAPTIVE_GIANT kernel option has been retired because its
1539 functionality is the default now.
1542 The AT keyboard emulation of sunkbd(4) has been turned on
1543 by default. In order to make the special symbols of the Sun
1544 keyboards driven by sunkbd(4) work under X these now have
1545 to be configured the same way as Sun USB keyboards driven
1546 by ukbd(4) (which also does AT keyboard emulation), f.e.:
1548 Option "XkbLayout" "us"
1549 Option "XkbRules" "xorg"
1550 Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
1553 It has been decided that it is desirable to provide ABI
1554 backwards compatibility to the FreeBSD 4/5/6 versions of the
1555 PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
1556 broken with the introduction of PCI domain support (see the
1557 20070930 entry). Unfortunately, this required the ABI of
1558 PCIOCGETCONF to be broken again in order to be able to
1559 provide backwards compatibility to the old version of that
1560 IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
1561 again. As for prominent ports this affects neither pciutils
1562 nor xorg-server this time, the hal port needs to be rebuilt
1566 The misnamed kthread_create() and friends have been renamed
1567 to kproc_create() etc. Many of the callers already
1568 used kproc_start()..
1569 I will return kthread_create() and friends in a while
1570 with implementations that actually create threads, not procs.
1571 Renaming corresponds with version 800002.
1580 Avoid using make -j when upgrading. While generally safe, there are
1581 sometimes problems using -j to upgrade. If your upgrade fails with
1582 -j, please try again without -j. From time to time in the past there
1583 have been problems using -j with buildworld and/or installworld. This
1584 is especially true when upgrading between "distant" versions (eg one
1585 that cross a major release boundary or several minor releases, or when
1586 several months have passed on the -current branch).
1588 Sometimes, obscure build problems are the result of environment
1589 poisoning. This can happen because the make utility reads its
1590 environment when searching for values for global variables. To run
1591 your build attempts in an "environmental clean room", prefix all make
1592 commands with 'env -i '. See the env(1) manual page for more details.
1594 When upgrading from one major version to another it is generally best
1595 to upgrade to the latest code in the currently installed branch first,
1596 then do an upgrade to the new branch. This is the best-tested upgrade
1597 path, and has the highest probability of being successful. Please try
1598 this approach before reporting problems with a major version upgrade.
1602 When upgrading the boot ZFS pool to a new version, always follow
1605 1.) recompile and reinstall the ZFS boot loader and boot block
1606 (this is part of "make buildworld" and "make installworld")
1608 2.) update the ZFS boot block on your boot drive
1610 The following example updates the ZFS boot block on the first
1611 partition (freebsd-boot) of a GPT partitioned drive ad0:
1612 "gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
1614 Non-boot pools do not need these updates.
1618 If you are updating from a prior version of FreeBSD (even one just
1619 a few days old), you should follow this procedure. It is the most
1620 failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
1622 make kernel-toolchain
1623 make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
1624 make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
1626 To test a kernel once
1627 ---------------------
1628 If you just want to boot a kernel once (because you are not sure
1629 if it works, or if you want to boot a known bad kernel to provide
1630 debugging information) run
1631 make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
1632 nextboot -k testkernel
1634 To just build a kernel when you know that it won't mess you up
1635 --------------------------------------------------------------
1636 This assumes you are already running a CURRENT system. Replace
1637 ${arch} with the architecture of your machine (e.g. "i386",
1638 "arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
1640 cd src/sys/${arch}/conf
1641 config KERNEL_NAME_HERE
1642 cd ../compile/KERNEL_NAME_HERE
1647 If this fails, go to the "To build a kernel" section.
1649 To rebuild everything and install it on the current system.
1650 -----------------------------------------------------------
1651 # Note: sometimes if you are running current you gotta do more than
1652 # is listed here if you are upgrading from a really old current.
1654 <make sure you have good level 0 dumps>
1656 make kernel KERNCONF=YOUR_KERNEL_HERE
1658 <reboot in single user> [3]
1666 To cross-install current onto a separate partition
1667 --------------------------------------------------
1668 # In this approach we use a separate partition to hold
1669 # current's root, 'usr', and 'var' directories. A partition
1670 # holding "/", "/usr" and "/var" should be about 2GB in
1673 <make sure you have good level 0 dumps>
1676 make buildkernel KERNCONF=YOUR_KERNEL_HERE
1677 <maybe newfs current's root partition>
1678 <mount current's root partition on directory ${CURRENT_ROOT}>
1679 make installworld DESTDIR=${CURRENT_ROOT}
1680 make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
1681 make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
1682 cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
1683 <edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
1684 <reboot into current>
1685 <do a "native" rebuild/install as described in the previous section>
1686 <maybe install compatibility libraries from ports/misc/compat*>
1690 To upgrade in-place from stable to current
1691 ----------------------------------------------
1692 <make sure you have good level 0 dumps>
1694 make kernel KERNCONF=YOUR_KERNEL_HERE [8]
1696 <reboot in single user> [3]
1703 Make sure that you've read the UPDATING file to understand the
1704 tweaks to various things you need. At this point in the life
1705 cycle of current, things change often and you are on your own
1706 to cope. The defaults can also change, so please read ALL of
1707 the UPDATING entries.
1709 Also, if you are tracking -current, you must be subscribed to
1710 freebsd-current@freebsd.org. Make sure that before you update
1711 your sources that you have read and understood all the recent
1712 messages there. If in doubt, please track -stable which has
1713 much fewer pitfalls.
1715 [1] If you have third party modules, such as vmware, you
1716 should disable them at this point so they don't crash your
1719 [3] From the bootblocks, boot -s, and then do
1724 adjkerntz -i # if CMOS is wall time
1725 Also, when doing a major release upgrade, it is required that
1726 you boot into single user mode to do the installworld.
1728 [4] Note: This step is non-optional. Failure to do this step
1729 can result in a significant reduction in the functionality of the
1730 system. Attempting to do it by hand is not recommended and those
1731 that pursue this avenue should read this file carefully, as well
1732 as the archives of freebsd-current and freebsd-hackers mailing lists
1733 for potential gotchas. The -U option is also useful to consider.
1734 See mergemaster(8) for more information.
1736 [5] Usually this step is a noop. However, from time to time
1737 you may need to do this if you get unknown user in the following
1738 step. It never hurts to do it all the time. You may need to
1739 install a new mergemaster (cd src/usr.sbin/mergemaster && make
1740 install) after the buildworld before this step if you last updated
1741 from [78]-stable or 9-stable before 20130430.
1743 [6] This only deletes old files and directories. Old libraries
1744 can be deleted by "make delete-old-libs", but you have to make
1745 sure that no program is using those libraries anymore.
1747 [8] In order to have a kernel that can run the 4.x binaries needed to
1748 do an installworld, you must include the COMPAT_FREEBSD4 option in
1749 your kernel. Failure to do so may leave you with a system that is
1750 hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
1751 required to run the 5.x binaries on more recent kernels. And so on
1752 for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
1754 Make sure that you merge any new devices from GENERIC since the
1755 last time you updated your kernel config file.
1757 [9] When checking out sources, you must include the -P flag to have
1758 cvs prune empty directories.
1760 If CPUTYPE is defined in your /etc/make.conf, make sure to use the
1761 "?=" instead of the "=" assignment operator, so that buildworld can
1762 override the CPUTYPE if it needs to.
1764 MAKEOBJDIRPREFIX must be defined in an environment variable, and
1765 not on the command line, or in /etc/make.conf. buildworld will
1766 warn if it is improperly defined.
1769 This file contains a list, in reverse chronological order, of major
1770 breakages in tracking -current. It is not guaranteed to be a complete
1771 list of such breakages, and only contains entries since October 10, 2007.
1772 If you need to see UPDATING entries from before that date, you will need
1773 to fetch an UPDATING file from an older FreeBSD release.
1775 Copyright information:
1777 Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
1779 Redistribution, publication, translation and use, with or without
1780 modification, in full or in part, in any form or format of this
1781 document are permitted without further permission from the author.
1783 THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
1784 IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
1785 WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
1786 DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
1787 INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
1788 (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
1789 SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
1790 HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
1791 STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
1792 IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
1793 POSSIBILITY OF SUCH DAMAGE.
1795 Contact Warner Losh if you have any questions about your use of