freebsd-dev/UPDATING

1791 lines
74 KiB
Plaintext
Raw Normal View History

2018-07-27 21:36:29 +00:00
Updating Information for FreeBSD current users.
This file is maintained and copyrighted by M. Warner Losh <imp@freebsd.org>.
See end of file for further details. For commonly done items, please see the
COMMON ITEMS: section later in the file. These instructions assume that you
basically know what you are doing. If not, then please consult the FreeBSD
handbook:
https://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html
Items affecting the ports and packages system can be found in
/usr/ports/UPDATING. Please read that file before running portupgrade.
NOTE: FreeBSD has switched from gcc to clang. If you have trouble bootstrapping
from older versions of FreeBSD, try WITHOUT_CLANG and WITH_GCC to bootstrap to
the tip of head, and then rebuild without this option. The bootstrap process
from older version of current across the gcc/clang cutover is a bit fragile.
NOTE TO PEOPLE WHO THINK THAT FreeBSD 12.x IS SLOW:
FreeBSD 12.x has many debugging features turned on, in both the kernel
and userland. These features attempt to detect incorrect use of
system primitives, and encourage loud failure through extra sanity
checking and fail stop semantics. They also substantially impact
system performance. If you want to do performance measurement,
benchmarking, and optimization, you'll want to turn them off. This
includes various WITNESS- related kernel options, INVARIANTS, malloc
debugging flags in userland, and various verbose features in the
kernel. Many developers choose to disable these features on build
machines to maximize performance. (To completely disable malloc
debugging, define MALLOC_PRODUCTION in /etc/make.conf, or to merely
disable the most expensive debugging functionality run
"ln -s 'abort:false,junk:false' /etc/malloc.conf".)
20180808:
The default pager for most commands has been changed to "less". To
restore the old behavior, set PAGER="more" and MANPAGER="more -s" in
your environment.
20180731:
The jedec_ts(4) driver has been removed. A superset of its functionality
is available in the jedec_dimm(4) driver, and the manpage for that
driver includes migration instructions. If you have "device jedec_ts"
in your kernel configuration file, it must be removed.
20180730:
amd64/GENERIC now has EFI runtime services, EFIRT, enabled by default.
This should have no effect if the kernel is booted via BIOS/legacy boot.
EFIRT may be disabled via a loader tunable, efi.rt.disabled, if a system
has a buggy firmware that prevents a successful boot due to use of
runtime services.
20180727:
Atmel AT91RM9200 and AT91SAM9, Cavium CNS 11xx and XScale
support has been removed from the tree. These ports were
obsolete and/or known to be broken for many years.
Finalize the boot manager protocol support for next-stage boot loading. If we are booting in a conforming UEFI Boot Manager Environment, then use the BootCurrent variable to find the BootXXXX we're using. Once we find that, then if it contains more than one EFI_DEVICE_PATH in its what to boot section, try to use the last one as the kernel to load. This will also set the default root partition as well. If there's only one path, or if there's an error along the way, assume that nothing specific was specified and revert to the old algorithm. If something was specified, but not found, then fail the boot. Otherwise you that, specific thing. On FreeBSD, this can be set using efibootmgr -l <loader> -k <kernel>. We try a few variations of kernel to cope with the fact that UEFI comes from a DOS world where paths might be upper case and/or contain back-slashes. Note: In an ideal world, we'd work out where we are in chain loading by looking at the passed-in image handle and doing name matching. However, that's unreliable since at least boot1.efi booted images don't have that, hence the assumption that loader.efi needs to load the last thing on the list, if possible. The reason we fail for something specific is so that we can fully participate in the UEFI Boot Manager Protocol and fail over to the next item in the list of BootOrder choices when something goes wrong at this stage. This implements was was talked about in freebsd-arch@ last year https://docs.freebsd.org/cgi/getmsg.cgi?fetch=3576+0+archive/2017/freebsd-arch/20171022.freebsd-arch and documented in full (after changed resulting from the discussion) in https://docs.google.com/document/d/1aK9IqF-60JPEbUeSAUAkYjF2W_8EnmczFs6RqCT90Jg/edit# although one or two minor details may have been modified in this implementation to make it work, and the ZFS MEDIA PATH extension isn't implemented. This does not yet move things to ESP:\efi\freebsd\loader.efi. RelNotes: Yes Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D16403
2018-07-23 20:36:59 +00:00
20180723:
loader.efi has been augmented to participate more fully in the
UEFI boot manager protocol. loader.efi will now look at the
BootXXXX environment variable to determine if a specific kernel
or root partition was specified. XXXX is derived from BootCurrent.
efibootmgr(8) manages these standard UEFI variables.
Finalize the boot manager protocol support for next-stage boot loading. If we are booting in a conforming UEFI Boot Manager Environment, then use the BootCurrent variable to find the BootXXXX we're using. Once we find that, then if it contains more than one EFI_DEVICE_PATH in its what to boot section, try to use the last one as the kernel to load. This will also set the default root partition as well. If there's only one path, or if there's an error along the way, assume that nothing specific was specified and revert to the old algorithm. If something was specified, but not found, then fail the boot. Otherwise you that, specific thing. On FreeBSD, this can be set using efibootmgr -l <loader> -k <kernel>. We try a few variations of kernel to cope with the fact that UEFI comes from a DOS world where paths might be upper case and/or contain back-slashes. Note: In an ideal world, we'd work out where we are in chain loading by looking at the passed-in image handle and doing name matching. However, that's unreliable since at least boot1.efi booted images don't have that, hence the assumption that loader.efi needs to load the last thing on the list, if possible. The reason we fail for something specific is so that we can fully participate in the UEFI Boot Manager Protocol and fail over to the next item in the list of BootOrder choices when something goes wrong at this stage. This implements was was talked about in freebsd-arch@ last year https://docs.freebsd.org/cgi/getmsg.cgi?fetch=3576+0+archive/2017/freebsd-arch/20171022.freebsd-arch and documented in full (after changed resulting from the discussion) in https://docs.google.com/document/d/1aK9IqF-60JPEbUeSAUAkYjF2W_8EnmczFs6RqCT90Jg/edit# although one or two minor details may have been modified in this implementation to make it work, and the ZFS MEDIA PATH extension isn't implemented. This does not yet move things to ESP:\efi\freebsd\loader.efi. RelNotes: Yes Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D16403
2018-07-23 20:36:59 +00:00
20180720:
zfsloader's functionality has now been folded into loader.
zfsloader is no longer necessary once you've updated your
boot blocks. For a transition period, we will install a
hardlink for zfsloader to loader to allow a smooth transition
until the boot blocks can be updated (hard link because old
zfs boot blocks don't understand symlinks).
20180719:
ARM64 now have efifb support, if you want to have serial console
on your arm64 board when an screen is connected and the bootloader
setup a frambuffer for us to use, just add :
boot_serial=YES
boot_multicons=YES
in /boot/loader.conf
For Raspberry Pi 3 (RPI) users, this is needed even if you don't have
an screen connected as the firmware will setup a framebuffer are that
u-boot will expose as an EFI framebuffer.
20180719:
New uid:gid added, ntpd:ntpd (123:123). Be sure to run mergemaster
or take steps to update /etc/passwd before doing installworld on
existing systems. Also, rc.d/ntpd now starts ntpd(8) as user ntpd
if the new mac_ntpd(4) policy is available, unless ntpd_flags or
the ntp config file contain options that change file/dir locations.
When such options (e.g., "statsdir" or "crypto") are used, ntpd can
still be run as non-root by setting ntpd_user=ntpd in rc.conf, after
taking steps to ensure that all required files/dirs are accessible
by the ntpd user.
2018-07-19 23:54:18 +00:00
20180717:
Big endian arm support has been removed.
20180711:
The static environment setup in kernel configs is no longer mutually
exclusive with the loader(8) environment by default. In order to
restore the previous default behavior of disabling the loader(8)
environment if a static environment is present, you must specify
loader_env.disabled=1 in the static environment.
20180705:
The ABI of syscalls used by management tools like sockstat and
netstat has been broken to allow 32-bit binaries to work on
64-bit kernels without modification. These programs will need
to match the kernel in order to function. External programs may
require minor modifications to accommodate a change of type in
structures from pointers to 64-bit virtual addresses.
20180702:
On i386 and amd64 atomics are now inlined. Out of tree modules using
atomics will need to be rebuilt.
20180701:
The '%I' format in the kern.corefile sysctl limits the number of
core files that a process can generate to the number stored in the
debug.ncores sysctl. The '%I' format is replaced by the single digit
index. Previously, if all indexes were taken the kernel would overwrite
only a core file with the highest index in a filename.
Currently the system will create a new core file if there is a free
index or if all slots are taken it will overwrite the oldest one.
20180630:
Clang, llvm, lld, lldb, compiler-rt and libc++ have been upgraded to
6.0.1. Please see the 20141231 entry below for information about
prerequisites and upgrading, if you are not already using clang 3.5.0
or higher.
20180628:
r335753 introduced a new quoting method. However, etc/devd/devmatch.conf
needed to be changed to work with it. This change was made with r335763
and requires a mergemaster / etcupdate / etc to update the installed file.
20180612:
r334930 changed the interface between the NFS modules, so they all
need to be rebuilt. r335018 did a __FreeBSD_version bump for this.
20180530:
The kernel / userland interface for devinfo changed, so you'll
need a new kernel and userland as a pair for it to work (rebuilding
lib/libdevinfo is all that's required). devinfo and devmatch will
not work, but everything else will when there's a mismatch.
20180523:
The on-disk format for hwpmc callchain records has changed to include
threadid corresponding to a given record. This changes the field offsets
and thus requires that libpmcstat be rebuilt before using a kernel
later than r334108.
20180517:
The vxge(4) driver has been removed. This driver was introduced into
HEAD one week before the Exar left the Ethernet market and is not
known to be used. If you have device vxge in your kernel config file
it must be removed.
2018-05-16 21:07:12 +00:00
20180510:
The amd64 kernel now requires a ld that supports ifunc to produce a
working kernel, either lld or a newer binutils. lld is built by default
on amd64, and the 'buildkernel' target uses it automatically. However,
it is not the default linker, so building the kernel the traditional
way requires LD=ld.lld on the command line (or LD=/usr/local/bin/ld for
binutils port/package). lld will soon be default, and this requirement
will go away.
20180508:
The nxge(4) driver has been removed. This driver was for PCI-X 10g
cards made by s2io/Neterion. The company was aquired by Exar and
no longer sells or supports Ethernet products. If you have device
nxge in your kernel config file it must be removed.
20180504:
The tz database (tzdb) has been updated to 2018e. This version more
correctly models time stamps in time zones with negative DST such as
Europe/Dublin (from 1971 on), Europe/Prague (1946/7), and
Africa/Windhoek (1994/2017). This does not affect the UT offsets, only
time zone abbreviations and the tm_isdst flag.
20180502:
The ixgb(4) driver has been removed. This driver was for an early and
uncommon legacy PCI 10GbE for a single ASIC, Intel 82597EX. Intel
quickly shifted to the long lived ixgbe family. If you have device
ixgb in your kernel config file it must be removed.
20180501:
The lmc(4) driver has been removed. This was a WAN interface
card that was already reportedly rare in 2003, and had an ambiguous
license. If you have device lmc in your kernel config file it must
be removed.
20180413:
Support for Arcnet networks has been removed. If you have device
arcnet or device cm in your kernel config file they must be
removed.
20180411:
Support for FDDI networks has been removed. If you have device
fddi or device fpa in your kernel config file they must be
removed.
20180406:
In addition to supporting RFC 3164 formatted messages, the
syslogd(8) service is now capable of parsing RFC 5424 formatted
log messages. The main benefit of using RFC 5424 is that clients
may now send log messages with timestamps containing year numbers,
microseconds and time zone offsets.
Similarly, the syslog(3) C library function has been altered to
send RFC 5424 formatted messages to the local system logging
daemon. On systems using syslogd(8), this change should have no
negative impact, as long as syslogd(8) and the C library are
updated at the same time. On systems using a different system
logging daemon, it may be necessary to make configuration
adjustments, depending on the software used.
When using syslog-ng, add the 'syslog-protocol' flag to local
input sources to enable parsing of RFC 5424 formatted messages:
source src {
unix-dgram("/var/run/log" flags(syslog-protocol));
}
When using rsyslog, disable the 'SysSock.UseSpecialParser' option
of the 'imuxsock' module to let messages be processed by the
regular RFC 3164/5424 parsing pipeline:
module(load="imuxsock" SysSock.UseSpecialParser="off")
Do note that these changes only affect communication between local
applications and syslogd(8). The format that syslogd(8) uses to
store messages on disk or forward messages to other systems
remains unchanged. syslogd(8) still uses RFC 3164 for these
purposes. Options to customize this behaviour will be added in the
future. Utilities that process log files stored in /var/log are
thus expected to continue to function as before.
__FreeBSD_version has been incremented to 1200061 to denote this
change.
2018-03-29 17:44:55 +00:00
20180328:
Support for token ring networks has been removed. If you
have "device token" in your kernel config you should remove
it. No device drivers supported token ring.
20180323:
makefs was modified to be able to tag ISO9660 El Torito boot catalog
entries as EFI instead of overloading the i386 tag as done previously.
The amd64 mkisoimages.sh script used to build amd64 ISO images for
release was updated to use this. This may mean that makefs must be
updated before "make cdrom" can be run in the release directory. This
should be as simple as:
$ cd $SRCDIR/usr.sbin/makefs
$ make depend all install
Add Lua as a scripting langauge to /boot/loader liblua glues the lua run time into the boot loader. It implements all the runtime routines that lua expects. In addition, it has a few standard 'C' headers that nueter various aspects of the LUA build that are too specific to lua to be in libsa. Many refinements from the original code to improve implementation and the number of included lua libraries. Use int64_t for lua_Number. Have "/boot/lua" be the default module path. Numerous cleanups from the original GSoC project, including hacking libsa to allow lua to be built with only one change outside luaconf.h. Add the final bit of lua glue to bring in liblua and plug into the multiple interpreter framework, previously committed. Add LOADER_LUA option, currently off by default. Presently, this is an experimental option. One must opt-in to using this by defining WITH_LOADER_LUA and WITHOUT_FORTH. It's been lightly tested, so keep a backup copy of your old loader handy. The menu code, coming in the next commit, hasn't been exhaustively tested. A LUA boot loader is 60k larger than a FORTH one, which is 80k larger than a no-interpreter one. Subtle changes in size may tip things past some subtle limit (the binary is ~430k now when built with LUA). A future version may offer coexistance. Bump FreeBSD version to 1200058 to mark the milestone. Pedro Souza's 2014 Summer of Code project. Rui Paulo, Pedro Arthur, Zakary Nafziger and Wojciech A. Koszek also contributed. Warner Losh reworked it extensively into its current form. Obtained from: https://wiki.freebsd.org/SummerOfCode2014/LuaLoader Sponsored by: Google Summer of Code Relnotes: Yes MFC After: 1 month Differential Review: https://reviews.freebsd.org/D14295
2018-02-12 15:31:53 +00:00
20180212:
FreeBSD boot loader enhanced with Lua scripting. It's purely opt-in for
now by building WITH_LOADER_LUA and WITHOUT_FORTH in /etc/src.conf.
Co-existance for the transition period will come shortly. Booting is a
complex environment and test coverage for Lua-enabled loaders has been
thin, so it would be prudent to assume it might not work and make
provisions for backup boot methods.
20180211:
devmatch functionality has been turned on in devd. It will automatically
load drivers for unattached devices. This may cause unexpected drivers to
be loaded. Please report any problems to current@ and imp@freebsd.org.
20180114:
Clang, llvm, lld, lldb, compiler-rt and libc++ have been upgraded to
6.0.0. Please see the 20141231 entry below for information about
prerequisites and upgrading, if you are not already using clang 3.5.0
or higher.
20180110:
LLVM's lld linker is now used as the FreeBSD/amd64 bootstrap linker.
This means it is used to link the kernel and userland libraries and
executables, but is not yet installed as /usr/bin/ld by default.
To revert to ld.bfd as the bootstrap linker, in /etc/src.conf set
WITHOUT_LLD_BOOTSTRAP=yes
2018-01-10 16:52:00 +00:00
20180110:
2018-01-12 17:33:20 +00:00
On i386, pmtimer has been removed. Its functionality has been folded
into apm. It was a no-op on ACPI in current for a while now (but was still
needed on i386 in FreeBSD 11 and earlier). Users may need to remove it
from kernel config files.
2018-01-10 16:52:00 +00:00
20180104:
The use of RSS hash from the network card aka flowid has been
disabled by default for lagg(4) as it's currently incompatible with
the lacp and loadbalance protocols.
This can be re-enabled by setting the following in loader.conf:
net.link.lagg.default_use_flowid="1"
20180102:
The SW_WATCHDOG option is no longer necessary to enable the
hardclock-based software watchdog if no hardware watchdog is
configured. As before, SW_WATCHDOG will cause the software
watchdog to be enabled even if a hardware watchdog is configured.
20171215:
r326887 fixes the issue described in the 20171214 UPDATING entry.
r326888 flips the switch back to building GELI support always.
20171214:
r362593 broke ZFS + GELI support for reasons unknown. However,
it also broke ZFS support generally, so GELI has been turned off
by default as the lesser evil in r326857. If you boot off ZFS and/or
GELI, it might not be a good time to update.
20171125:
PowerPC users must update loader(8) by rebuilding world before
installing a new kernel, as the protocol connecting them has
changed. Without the update, loader metadata will not be passed
successfully to the kernel and users will have to enter their
root partition at the kernel mountroot prompt to continue booting.
Newer versions of loader can boot old kernels without issue.
2017-11-26 04:55:23 +00:00
20171110:
The LOADER_FIREWIRE_SUPPORT build variable as been renamed to
WITH/OUT_LOADER_FIREWIRE. LOADER_{NO_,}GELI_SUPPORT has been renamed
to WITH/OUT_LOADER_GELI.
20171106:
The naive and non-compliant support of posix_fallocate(2) in ZFS
has been removed as of r325320. The system call now returns EINVAL
when used on a ZFS file. Although the new behavior complies with the
standard, some consumers are not prepared to cope with it.
One known victim is lld prior to r325420.
20171102:
Building in a FreeBSD src checkout will automatically create object
directories now rather than store files in the current directory if
2017-11-17 15:46:19 +00:00
'make obj' was not ran. Calling 'make obj' is no longer necessary.
This feature can be disabled by setting WITHOUT_AUTO_OBJ=yes in
/etc/src-env.conf (not /etc/src.conf), or passing the option in the
environment.
20171101:
The default MAKEOBJDIR has changed from /usr/obj/<srcdir> for native
builds, and /usr/obj/<arch>/<srcdir> for cross-builds, to a unified
/usr/obj/<srcdir>/<arch>. This behavior can be changed to the old
format by setting WITHOUT_UNIFIED_OBJDIR=yes in /etc/src-env.conf,
the environment, or with -DWITHOUT_UNIFIED_OBJDIR when building.
The UNIFIED_OBJDIR option is a transitional feature that will be
removed for 12.0 release; please migrate to the new format for any
tools by looking up the OBJDIR used by 'make -V .OBJDIR' means rather
than hardcoding paths.
20171028:
The native-xtools target no longer installs the files by default to the
OBJDIR. Use the native-xtools-install target with a DESTDIR to install
to ${DESTDIR}/${NXTP} where NXTP defaults to /nxb-bin.
20171021:
As part of the boot loader infrastructure cleanup, LOADER_*_SUPPORT
options are changing from controlling the build if defined / undefined
to controlling the build with explicit 'yes' or 'no' values. They will
shift to WITH/WITHOUT options to match other options in the system.
20171010:
libstand has turned into a private library for sys/boot use only.
It is no longer supported as a public interface outside of sys/boot.
20171005:
The arm port has split armv6 into armv6 and armv7. armv7 is now
a valid TARGET_ARCH/MACHINE_ARCH setting. If you have an armv7 system
and are running a kernel from before r324363, you will need to add
MACHINE_ARCH=armv7 to 'make buildworld' to do a native build.
2017-10-04 09:54:55 +00:00
20171003:
When building multiple kernels using KERNCONF, non-existent KERNCONF
files will produce an error and buildkernel will fail. Previously
missing KERNCONF files silently failed giving no indication as to
why, only to subsequently discover during installkernel that the
desired kernel was never built in the first place.
20170912:
The default serial number format for CTL LUNs has changed. This will
affect users who use /dev/diskid/* device nodes, or whose FibreChannel
or iSCSI clients care about their LUNs' serial numbers. Users who
require serial number stability should hardcode serial numbers in
/etc/ctl.conf .
20170912:
For 32-bit arm compiled for hard-float support, soft-floating point
binaries now always get their shared libraries from
LD_SOFT_LIBRARY_PATH (in the past, this was only used if
/usr/libsoft also existed). Only users with a hard-float ld.so, but
soft-float everything else should be affected.
20170826:
The geli password typed at boot is now hidden. To restore the previous
behavior, see geli(8) for configuration options.
20170825:
Move PMTUD blackhole counters to TCPSTATS and remove them from bare
sysctl values. Minor nit, but requires a rebuild of both world/kernel
to complete.
20170814:
"make check" behavior (made in ^/head@r295380) has been changed to
execute from a limited sandbox, as opposed to executing from
${TESTSDIR}.
Behavioral changes:
- The "beforecheck" and "aftercheck" targets are now specified.
- ${CHECKDIR} (added in commit noted above) has been removed.
- Legacy behavior can be enabled by setting
WITHOUT_MAKE_CHECK_USE_SANDBOX in src.conf(5) or the environment.
If the limited sandbox mode is enabled, "make check" will execute
"make distribution", then install, execute the tests, and clean up the
sandbox if successful.
The "make distribution" and "make install" targets are typically run as
root to set appropriate permissions and ownership at installation time.
The end-user should set "WITH_INSTALL_AS_USER" in src.conf(5) or the
environment if executing "make check" with limited sandbox mode using
an unprivileged user.
20170808:
Since the switch to GPT disk labels, fsck for UFS/FFS has been
unable to automatically find alternate superblocks. As of r322297,
the information needed to find alternate superblocks has been
moved to the end of the area reserved for the boot block.
Filesystems created with a newfs of this vintage or later
will create the recovery information. If you have a filesystem
created prior to this change and wish to have a recovery block
created for your filesystem, you can do so by running fsck in
2018-01-12 17:33:20 +00:00
foreground mode (i.e., do not use the -p or -y options). As it
starts, fsck will ask ``SAVE DATA TO FIND ALTERNATE SUPERBLOCKS''
to which you should answer yes.
2017-07-28 21:14:28 +00:00
20170728:
As of r321665, an NFSv4 server configuration that services
Kerberos mounts or clients that do not support the uid/gid in
owner/owner_group string capability, must explicitly enable
the nfsuserd daemon by adding nfsuserd_enable="YES" to the
machine's /etc/rc.conf file.
20170722:
Clang, llvm, lldb, compiler-rt and libc++ have been upgraded to 5.0.0.
Please see the 20141231 entry below for information about prerequisites
and upgrading, if you are not already using clang 3.5.0 or higher.
20170701:
WITHOUT_RCMDS is now the default. Set WITH_RCMDS if you need the
r-commands (rlogin, rsh, etc.) to be built with the base system.
20170625:
The FreeBSD/powerpc platform now uses a 64-bit type for time_t. This is
a very major ABI incompatible change, so users of FreeBSD/powerpc must
be careful when performing source upgrades. It is best to run
'make installworld' from an alternate root system, either a live
CD/memory stick, or a temporary root partition. Additionally, all ports
must be recompiled. powerpc64 is largely unaffected, except in the case
of 32-bit compatibility. All 32-bit binaries will be affected.
20170623:
2017-06-23 18:06:20 +00:00
Forward compatibility for the "ino64" project have been committed. This
will allow most new binaries to run on older kernels in a limited
fashion. This prevents many of the common foot-shooting actions in the
upgrade as well as the limited ability to roll back the kernel across
the ino64 upgrade. Complicated use cases may not work properly, though
enough simpler ones work to allow recovery in most situations.
20170620:
Switch back to the BSDL dtc (Device Tree Compiler). Set WITH_GPL_DTC
if you require the GPL compiler.
2017-06-23 18:06:20 +00:00
20170618:
The internal ABI used for communication between the NFS kernel modules
was changed by r320085, so __FreeBSD_version was bumped to
ensure all the NFS related modules are updated together.
20170617:
The ABI of struct event was changed by extending the data
member to 64bit and adding ext fields. For upgrade, same
precautions as for the entry 20170523 "ino64" must be
followed.
20170531:
The GNU roff toolchain has been removed from base. To render manpages
which are not supported by mandoc(1), man(1) can fallback on GNU roff
from ports (and recommends to install it).
To render roff(7) documents, consider using GNU roff from ports or the
heirloom doctools roff toolchain from ports via pkg install groff or
via pkg install heirloom-doctools.
20170524:
The ath(4) and ath_hal(4) modules now build piecemeal to allow for
smaller runtime footprint builds. This is useful for embedded systems
which only require one chipset support.
If you load it as a module, make sure this is in /boot/loader.conf:
if_ath_load="YES"
This will load the HAL, all chip/RF backends and if_ath_pci.
If you have if_ath_pci in /boot/loader.conf, ensure it is after
if_ath or it will not load any HAL chipset support.
If you want to selectively load things (eg on ye cheape ARM/MIPS
platforms where RAM is at a premium) you should:
* load ath_hal
* load the chip modules in question
* load ath_rate, ath_dfs
* load ath_main
* load if_ath_pci and/or if_ath_ahb depending upon your particular
bus bind type - this is where probe/attach is done.
For further comments/feedback, poke adrian@ .
20170523:
The "ino64" 64-bit inode project has been committed, which extends
a number of types to 64 bits. Upgrading in place requires care and
adherence to the documented upgrade procedure.
If using a custom kernel configuration ensure that the
COMPAT_FREEBSD11 option is included (as during the upgrade the
system will be running the ino64 kernel with the existing world).
For the safest in-place upgrade begin by removing previous build
artifacts via "rm -rf /usr/obj/*". Then, carefully follow the
full procedure documented below under the heading "To rebuild
everything and install it on the current system." Specifically,
a reboot is required after installing the new kernel before
installing world.
20170424:
The NATM framework including the en(4), fatm(4), hatm(4), and
patm(4) devices has been removed. Consumers should plan a
migration before the end-of-life date for FreeBSD 11.
20170420:
GNU diff has been replaced by a BSD licensed diff. Some features of GNU
diff has not been implemented, if those are needed a newer version of
GNU diff is available via the diffutils package under the gdiff name.
20170413:
As of r316810 for ipfilter, keep frags is no longer assumed when
keep state is specified in a rule. r316810 aligns ipfilter with
documentation in man pages separating keep frags from keep state.
This allows keep state to be specified without forcing keep frags
and allows keep frags to be specified independently of keep state.
To maintain previous behaviour, also specify keep frags with
2017-04-14 05:32:03 +00:00
keep state (as documented in ipf.conf.5).
20170407:
arm64 builds now use the base system LLD 4.0.0 linker by default,
instead of requiring that the aarch64-binutils port or package be
installed. To continue using aarch64-binutils, set
CROSS_BINUTILS_PREFIX=/usr/local/aarch64-freebsd/bin .
20170405:
The UDP optimization in entry 20160818 that added the sysctl
net.inet.udp.require_l2_bcast has been reverted. L2 broadcast
packets will no longer be treated as L3 broadcast packets.
20170331:
Binds and sends to the loopback addresses, IPv6 and IPv4, will now
use any explicitly assigned loopback address available in the jail
instead of using the first assigned address of the jail.
20170329:
The ctl.ko module no longer implements the iSCSI target frontend:
cfiscsi.ko does instead.
If building cfiscsi.ko as a kernel module, the module can be loaded
via one of the following methods:
- `cfiscsi_load="YES"` in loader.conf(5).
- Add `cfiscsi` to `$kld_list` in rc.conf(5).
- ctladm(8)/ctld(8), when compiled with iSCSI support
(`WITH_ISCSI=yes` in src.conf(5))
Please see cfiscsi(4) for more details.
- Add support for eMMC "partitions". Besides the user data area, i. e. the default partition, eMMC v4.41 and later devices can additionally provide up to: 1 enhanced user data area partition 2 boot partitions 1 RPMB (Replay Protected Memory Block) partition 4 general purpose partitions (optionally with a enhanced or extended attribute) Of these "partitions", only the enhanced user data area one actually slices the user data area partition and, thus, gets handled with the help of geom_flashmap(4). The other types of partitions have address space independent from the default partition and need to be switched to via CMD6 (SWITCH), i. e. constitute a set of additional "disks". The second kind of these "partitions" doesn't fit that well into the design of mmc(4) and mmcsd(4). I've decided to let mmcsd(4) hook all of these "partitions" up as disk(9)'s (except for the RPMB partition as it didn't seem to make much sense to be able to put a file-system there and may require authentication; therefore, RPMB partitions are solely accessible via the newly added IOCTL interface currently; see also below). This approach for one resulted in cleaner code. Second, it retains the notion of mmcsd(4) children corresponding to a single physical device each. With the addition of some layering violations, it also would have been possible for mmc(4) to add separate mmcsd(4) instances with one disk each for all of these "partitions", however. Still, both mmc(4) and mmcsd(4) share some common code now e. g. for issuing CMD6, which has been factored out into mmc_subr.c. Besides simply subdividing eMMC devices, some Intel NUCs having UEFI code in the boot partitions etc., another use case for the partition support is the activation of pseudo-SLC mode, which manufacturers of eMMC chips typically associate with the enhanced user data area and/ or the enhanced attribute of general purpose partitions. CAVEAT EMPTOR: Partitioning eMMC devices is a one-time operation. - Now that properly issuing CMD6 is crucial (so data isn't written to the wrong partition for example), make a step into the direction of correctly handling the timeout for these commands in the MMC layer. Also, do a SEND_STATUS when CMD6 is invoked with an R1B response as recommended by relevant specifications. However, quite some work is left to be done in this regard; all other R1B-type commands done by the MMC layer also should be followed by a SEND_STATUS (CMD13), the erase timeout calculations/handling as documented in specifications are entirely ignored so far, the MMC layer doesn't provide timeouts applicable up to the bridge drivers and at least sdhci(4) currently is hardcoding 1 s as timeout for all command types unconditionally. Let alone already available return codes often not being checked in the MMC layer ... - Add an IOCTL interface to mmcsd(4); this is sufficiently compatible with Linux so that the GNU mmc-utils can be ported to and used with FreeBSD (note that due to the remaining deficiencies outlined above SANITIZE operations issued by/with `mmc` currently most likely will fail). These latter will be added to ports as sysutils/mmc-utils in a bit. Among others, the `mmc` tool of the GNU mmc-utils allows for partitioning eMMC devices (tested working). - For devices following the eMMC specification v4.41 or later, year 0 is 2013 rather than 1997; so correct this for assembling the device ID string properly. - Let mmcsd.ko depend on mmc.ko. Additionally, bump MMC_VERSION as at least for some of the above a matching pair is required. - In the ACPI front-end of sdhci(4) describe the Intel eMMC and SDXC controllers as such in order to match the PCI one. Additionally, in the entry for the 80860F14 SDXC controller remove the eMMC-only SDHCI_QUIRK_INTEL_POWER_UP_RESET. OKed by: imp Submitted by: ian (mmc_switch_status() implementation)
2017-03-16 22:23:04 +00:00
20170316:
The mmcsd.ko module now additionally depends on geom_flashmap.ko.
Also, mmc.ko and mmcsd.ko need to be a matching pair built from the
same source (previously, the dependency of mmcsd.ko on mmc.ko was
missing, but mmcsd.ko now will refuse to load if it is incompatible
with mmc.ko).
20170315:
The syntax of ipfw(8) named states was changed to avoid ambiguity.
If you have used named states in the firewall rules, you need to modify
them after installworld and before rebooting. Now named states must
be prefixed with colon.
20170311:
The old drm (sys/dev/drm/) drivers for i915 and radeon have been
removed as the userland we provide cannot use them. The KMS version
(sys/dev/drm2) supports the same hardware.
20170302:
Clang, llvm, lldb, compiler-rt and libc++ have been upgraded to 4.0.0.
Please see the 20141231 entry below for information about prerequisites
and upgrading, if you are not already using clang 3.5.0 or higher.
20170221:
The code that provides support for ZFS .zfs/ directory functionality
has been reimplemented. It's not possible now to create a snapshot
by mkdir under .zfs/snapshot/. That should be the only user visible
change.
2017-02-17 06:22:00 +00:00
20170216:
EISA bus support has been removed. The WITH_EISA option is no longer
valid.
20170215:
MCA bus support has been removed.
20170127:
The WITH_LLD_AS_LD / WITHOUT_LLD_AS_LD build knobs have been renamed
WITH_LLD_IS_LD / WITHOUT_LLD_IS_LD, for consistency with CLANG_IS_CC.
20170112:
The EM_MULTIQUEUE kernel configuration option is deprecated now that
the em(4) driver conforms to iflib specifications.
20170109:
The igb(4), em(4) and lem(4) ethernet drivers are now implemented via
IFLIB. If you have a custom kernel configuration that excludes em(4)
but you use igb(4), you need to re-add em(4) to your custom configuration.
20161217:
Clang, llvm, lldb, compiler-rt and libc++ have been upgraded to 3.9.1.
Please see the 20141231 entry below for information about prerequisites
and upgrading, if you are not already using clang 3.5.0 or higher.
20161124:
Clang, llvm, lldb, compiler-rt and libc++ have been upgraded to 3.9.0.
Please see the 20141231 entry below for information about prerequisites
and upgrading, if you are not already using clang 3.5.0 or higher.
20161119:
The layout of the pmap structure has changed for powerpc to put the pmap
statistics at the front for all CPU variations. libkvm(3) and all tools
that link against it need to be recompiled.
add iic interface to ig4 driver, move isl and cyapa to iicbus Summary: The hardware does not expose a classic SMBus interface. Instead it has a lower level interface that can express a far richer I2C protocol than what smbus offers. However, the interface does not provide a way to explicitly generate the I2C stop and start conditions. It's only possible to request that the stop condition is generated after transferring the next byte in either direction. So, at least one data byte must always be transferred. Thus, some I2C sequences are impossible to generate, e.g., an equivalent of smbus quick command (<start>-<slave addr>-<r/w bit>-<stop>). At the same time isl(4) and cyapa(4) are moved to iicbus and now they use iicbus_transfer for communication. Previously they used smbus_trans() interface that is not defined by the SMBus protocol and was implemented only by ig4(4). In fact, that interface was impossible to implement for the typical SMBus controllers like intpm(4) or ichsmb(4) where a type of the SMBus command must be programmed. The plan is to remove smbus_trans() and all its uses. As an aside, the smbus_trans() method deviates from the standard, but perhaps backwards, FreeBSD convention of using 8-bit slave addresses (shifted by 1 bit to the left). The method expects 7-bit addresses. There is a user facing consequence of this change. A user must now provide device hints for isl and cyapa that specify an iicbus to use and a slave address on it. On Chromebook hardware where isl and cyapa devices are commonly found it is also possible to use a new chromebook_platform(4) driver that automatically configures isl and cyapa devices. There is no need to provide the device hints in that case, Right now smbus(4) driver tries to discover all slaves on the bus. That is very dangerous. Fortunately, the probing code uses smbus_trans() to do its job, so it is really enabled for ig4 only. The plan is to remove that auto-probing code and smbus_trans(). Tested by: grembo, Matthias Apitz <guru@unixarea.de> (w/o chromebook_platform) Discussed with: grembo, imp Reviewed by: wblock (docs) MFC after: 1 month Relnotes: yes Differential Revision: https://reviews.freebsd.org/D8172
2016-10-30 12:15:33 +00:00
20161030:
isl(4) and cyapa(4) drivers now require a new driver,
chromebook_platform(4), to work properly on Chromebook-class hardware.
On other types of hardware the drivers may need to be configured using
device hints. Please see the corresponding manual pages for details.
2016-10-17 21:35:13 +00:00
20161017:
The urtwn(4) driver was merged into rtwn(4) and now consists of
rtwn(4) main module + rtwn_usb(4) and rtwn_pci(4) bus-specific
parts.
Also, firmware for RTL8188CE was renamed due to possible name
conflict (rtwnrtl8192cU(B) -> rtwnrtl8192cE(B))
20161015:
GNU rcs has been removed from base. It is available as packages:
- rcs: Latest GPLv3 GNU rcs version.
2016-10-15 15:55:04 +00:00
- rcs57: Copy of the latest version of GNU rcs (GPLv2) before it was
removed from base.
20161008:
Use of the cc_cdg, cc_chd, cc_hd, or cc_vegas congestion control
modules now requires that the kernel configuration contain the
TCP_HHOOK option. (This option is included in the GENERIC kernel.)
20161003:
The WITHOUT_ELFCOPY_AS_OBJCOPY src.conf(5) knob has been retired.
ELF Tool Chain's elfcopy is always installed as /usr/bin/objcopy.
20160924:
Relocatable object files with the extension of .So have been renamed
to use an extension of .pico instead. The purpose of this change is
to avoid a name clash with shared libraries on case-insensitive file
systems. On those file systems, foo.So is the same file as foo.so.
20160918:
2016-09-18 15:55:57 +00:00
GNU rcs has been turned off by default. It can (temporarily) be built
again by adding WITH_RCS knob in src.conf.
Otherwise, GNU rcs is available from packages:
- rcs: Latest GPLv3 GNU rcs version.
- rcs57: Copy of the latest version of GNU rcs (GPLv2) from base.
20160918:
The backup_uses_rcs functionality has been removed from rc.subr.
20160908:
The queue(3) debugging macro, QUEUE_MACRO_DEBUG, has been split into
two separate components, QUEUE_MACRO_DEBUG_TRACE and
QUEUE_MACRO_DEBUG_TRASH. Define both for the original
QUEUE_MACRO_DEBUG behavior.
20160824:
r304787 changed some ioctl interfaces between the iSCSI userspace
programs and the kernel. ctladm, ctld, iscsictl, and iscsid must be
rebuilt to work with new kernels. __FreeBSD_version has been bumped
to 1200005.
20160818:
The UDP receive code has been updated to only treat incoming UDP
packets that were addressed to an L2 broadcast address as L3
broadcast packets. It is not expected that this will affect any
standards-conforming UDP application. The new behaviour can be
disabled by setting the sysctl net.inet.udp.require_l2_bcast to
0.
20160818:
Remove the openbsd_poll system call.
__FreeBSD_version has been bumped because of this.
20160708:
The stable/11 branch has been created from head@r302406.
20160622:
The libc stub for the pipe(2) system call has been replaced with
a wrapper that calls the pipe2(2) system call and the pipe(2)
system call is now only implemented by the kernels that include
"options COMPAT_FREEBSD10" in their config file (this is the
default). Users should ensure that this option is enabled in
their kernel or upgrade userspace to r302092 before upgrading their
kernel.
20160527:
CAM will now strip leading spaces from SCSI disks' serial numbers.
2016-11-17 00:21:55 +00:00
This will affect users who create UFS filesystems on SCSI disks using
those disk's diskid device nodes. For example, if /etc/fstab
previously contained a line like
"/dev/diskid/DISK-%20%20%20%20%20%20%20ABCDEFG0123456", you should
change it to "/dev/diskid/DISK-ABCDEFG0123456". Users of geom
transforms like gmirror may also be affected. ZFS users should
generally be fine.
20160523:
The bitstring(3) API has been updated with new functionality and
improved performance. But it is binary-incompatible with the old API.
Objects built with the new headers may not be linked against objects
built with the old headers.
20160520:
The brk and sbrk functions have been removed from libc on arm64.
Binutils from ports has been updated to not link to these
functions and should be updated to the latest version before
installing a new libc.
20160517:
The armv6 port now defaults to hard float ABI. Limited support
for running both hardfloat and soft float on the same system
is available using the libraries installed with -DWITH_LIBSOFT.
This has only been tested as an upgrade path for installworld
and packages may fail or need manual intervention to run. New
packages will be needed.
To update an existing self-hosted armv6hf system, you must add
TARGET_ARCH=armv6 on the make command line for both the build
and the install steps.
20160510:
Kernel modules compiled outside of a kernel build now default to
installing to /boot/modules instead of /boot/kernel. Many kernel
modules built this way (such as those in ports) already overrode
KMODDIR explicitly to install into /boot/modules. However,
manually building and installing a module from /sys/modules will
now install to /boot/modules instead of /boot/kernel.
2016-04-14 22:38:13 +00:00
20160414:
2016-04-15 04:45:37 +00:00
The CAM I/O scheduler has been committed to the kernel. There should be
no user visible impact. This does enable NCQ Trim on ada SSDs. While the
list of known rogues that claim support for this but actually corrupt
data is believed to be complete, be on the lookout for data
corruption. The known rogue list is believed to be complete:
o Crucial MX100, M550 drives with MU01 firmware.
o Micron M510 and M550 drives with MU01 firmware.
o Micron M500 prior to MU07 firmware
o Samsung 830, 840, and 850 all firmwares
o FCCT M500 all firmwares
Crucial has firmware http://www.crucial.com/usa/en/support-ssd-firmware
with working NCQ TRIM. For Micron branded drives, see your sales rep for
updated firmware. Black listed drives will work correctly because these
drives work correctly so long as no NCQ TRIMs are sent to them. Given
this list is the same as found in Linux, it's believed there are no
other rogues in the market place. All other models from the above
vendors work.
To be safe, if you are at all concerned, you can quirk each of your
drives to prevent NCQ from being sent by setting:
kern.cam.ada.X.quirks="0x2"
in loader.conf. If the drive requires the 4k sector quirk, set the
quirks entry to 0x3.
2016-04-14 22:38:13 +00:00
20160330:
The FAST_DEPEND build option has been removed and its functionality is
now the one true way. The old mkdep(1) style of 'make depend' has
been removed. See 20160311 for further details.
Use uintmax_t (typedef'd to rman_res_t type) for rman ranges. On some architectures, u_long isn't large enough for resource definitions. Particularly, powerpc and arm allow 36-bit (or larger) physical addresses, but type `long' is only 32-bit. This extends rman's resources to uintmax_t. With this change, any resource can feasibly be placed anywhere in physical memory (within the constraints of the driver). Why uintmax_t and not something machine dependent, or uint64_t? Though it's possible for uintmax_t to grow, it's highly unlikely it will become 128-bit on 32-bit architectures. 64-bit architectures should have plenty of RAM to absorb the increase on resource sizes if and when this occurs, and the number of resources on memory-constrained systems should be sufficiently small as to not pose a drastic overhead. That being said, uintmax_t was chosen for source clarity. If it's specified as uint64_t, all printf()-like calls would either need casts to uintmax_t, or be littered with PRI*64 macros. Casts to uintmax_t aren't horrible, but it would also bake into the API for resource_list_print_type() either a hidden assumption that entries get cast to uintmax_t for printing, or these calls would need the PRI*64 macros. Since source code is meant to be read more often than written, I chose the clearest path of simply using uintmax_t. Tested on a PowerPC p5020-based board, which places all device resources in 0xfxxxxxxxx, and has 8GB RAM. Regression tested on qemu-system-i386 Regression tested on qemu-system-mips (malta profile) Tested PAE and devinfo on virtualbox (live CD) Special thanks to bz for his testing on ARM. Reviewed By: bz, jhb (previous) Relnotes: Yes Sponsored by: Alex Perez/Inertial Computing Differential Revision: https://reviews.freebsd.org/D4544
2016-03-18 01:28:41 +00:00
20160317:
Resource range types have grown from unsigned long to uintmax_t. All
drivers, and anything using libdevinfo, need to be recompiled.
20160311:
WITH_FAST_DEPEND is now enabled by default for in-tree and out-of-tree
builds. It no longer runs mkdep(1) during 'make depend', and the
'make depend' stage can safely be skipped now as it is auto ran
when building 'make all' and will generate all SRCS and DPSRCS before
building anything else. Dependencies are gathered at compile time with
-MF flags kept in separate .depend files per object file. Users should
run 'make cleandepend' once if using -DNO_CLEAN to clean out older
stale .depend files.
20160306:
On amd64, clang 3.8.0 can now insert sections of type AMD64_UNWIND into
kernel modules. Therefore, if you load any kernel modules at boot time,
please install the boot loaders after you install the kernel, but before
rebooting, e.g.:
make buildworld
make buildkernel KERNCONF=YOUR_KERNEL_HERE
make installkernel KERNCONF=YOUR_KERNEL_HERE
make -C sys/boot install
<reboot in single user>
Then follow the usual steps, described in the General Notes section,
below.
20160305:
Clang, llvm, lldb and compiler-rt have been upgraded to 3.8.0. Please
see the 20141231 entry below for information about prerequisites and
upgrading, if you are not already using clang 3.5.0 or higher.
20160301:
The AIO subsystem is now a standard part of the kernel. The
VFS_AIO kernel option and aio.ko kernel module have been removed.
Due to stability concerns, asynchronous I/O requests are only
permitted on sockets and raw disks by default. To enable
asynchronous I/O requests on all file types, set the
vfs.aio.enable_unsafe sysctl to a non-zero value.
20160226:
The ELF object manipulation tool objcopy is now provided by the
ELF Tool Chain project rather than by GNU binutils. It should be a
drop-in replacement, with the addition of arm64 support. The
(temporary) src.conf knob WITHOUT_ELFCOPY_AS_OBJCOPY knob may be set
to obtain the GNU version if necessary.
20160129:
Building ZFS pools on top of zvols is prohibited by default. That
feature has never worked safely; it's always been prone to deadlocks.
Using a zvol as the backing store for a VM guest's virtual disk will
still work, even if the guest is using ZFS. Legacy behavior can be
restored by setting vfs.zfs.vol.recursive=1.
20160119:
The NONE and HPN patches has been removed from OpenSSH. They are
still available in the security/openssh-portable port.
20160113:
With the addition of ypldap(8), a new _ypldap user is now required
during installworld. "mergemaster -p" can be used to add the user
prior to installworld, as documented in the handbook.
20151216:
The tftp loader (pxeboot) now uses the option root-path directive. As a
consequence it no longer looks for a pxeboot.4th file on the tftp
server. Instead it uses the regular /boot infrastructure as with the
other loaders.
20151211:
The code to start recording plug and play data into the modules has
been committed. While the old tools will properly build a new kernel,
a number of warnings about "unknown metadata record 4" will be produced
for an older kldxref. To avoid such warnings, make sure to rebuild
the kernel toolchain (or world). Make sure that you have r292078 or
later when trying to build 292077 or later before rebuilding.
20151207:
Debug data files are now built by default with 'make buildworld' and
installed with 'make installworld'. This facilitates debugging but
requires more disk space both during the build and for the installed
world. Debug files may be disabled by setting WITHOUT_DEBUG_FILES=yes
in src.conf(5).
20151130:
r291527 changed the internal interface between the nfsd.ko and
nfscommon.ko modules. As such, they must both be upgraded to-gether.
__FreeBSD_version has been bumped because of this.
20151108:
Add support for unicode collation strings leads to a change of
order of files listed by ls(1) for example. To get back to the old
behaviour, set LC_COLLATE environment variable to "C".
Databases administrators will need to reindex their databases given
collation results will be different.
2015-11-17 19:40:48 +00:00
Due to a bug in install(1) it is recommended to remove the ancient
locales before running make installworld.
rm -rf /usr/share/locale/*
2015-10-30 20:51:33 +00:00
20151030:
The OpenSSL has been upgraded to 1.0.2d. Any binaries requiring
libcrypto.so.7 or libssl.so.7 must be recompiled.
20151020:
Qlogic 24xx/25xx firmware images were updated from 5.5.0 to 7.3.0.
Kernel modules isp_2400_multi and isp_2500_multi were removed and
should be replaced with isp_2400 and isp_2500 modules respectively.
20151017:
The build previously allowed using 'make -n' to not recurse into
sub-directories while showing what commands would be executed, and
'make -n -n' to recursively show commands. Now 'make -n' will recurse
and 'make -N' will not.
20151012:
If you specify SENDMAIL_MC or SENDMAIL_CF in make.conf, mergemaster
and etcupdate will now use this file. A custom sendmail.cf is now
updated via this mechanism rather than via installworld. If you had
excluded sendmail.cf in mergemaster.rc or etcupdate.conf, you may
want to remove the exclusion or change it to "always install".
/etc/mail/sendmail.cf is now managed the same way regardless of
whether SENDMAIL_MC/SENDMAIL_CF is used. If you are not using
SENDMAIL_MC/SENDMAIL_CF there should be no change in behavior.
20151011:
Compatibility shims for legacy ATA device names have been removed.
It includes ATA_STATIC_ID kernel option, kern.cam.ada.legacy_aliases
and kern.geom.raid.legacy_aliases loader tunables, kern.devalias.*
environment variables, /dev/ad* and /dev/ar* symbolic links.
20151006:
2015-10-05 19:48:12 +00:00
Clang, llvm, lldb, compiler-rt and libc++ have been upgraded to 3.7.0.
Please see the 20141231 entry below for information about prerequisites
and upgrading, if you are not already using clang 3.5.0 or higher.
2015-09-24 16:56:44 +00:00
20150924:
Kernel debug files have been moved to /usr/lib/debug/boot/kernel/,
and renamed from .symbols to .debug. This reduces the size requirements
on the boot partition or file system and provides consistency with
userland debug files.
When using the supported kernel installation method the
/usr/lib/debug/boot/kernel directory will be renamed (to kernel.old)
as is done with /boot/kernel.
Developers wishing to maintain the historical behavior of installing
debug files in /boot/kernel/ can set KERN_DEBUGDIR="" in src.conf(5).
20150827:
The wireless drivers had undergone changes that remove the 'parent
interface' from the ifconfig -l output. The rc.d network scripts
used to check presence of a parent interface in the list, so old
scripts would fail to start wireless networking. Thus, etcupdate(3)
or mergemaster(8) run is required after kernel update, to update your
rc.d scripts in /etc.
20150827:
pf no longer supports 'scrub fragment crop' or 'scrub fragment drop-ovl'
These configurations are now automatically interpreted as
'scrub fragment reassemble'.
20150817:
Kernel-loadable modules for the random(4) device are back. To use
them, the kernel must have
device random
options RANDOM_LOADABLE
kldload(8) can then be used to load random_fortuna.ko
or random_yarrow.ko. Please note that due to the indirect
function calls that the loadable modules need to provide,
the build-in variants will be slightly more efficient.
The random(4) kernel option RANDOM_DUMMY has been retired due to
unpopularity. It was not all that useful anyway.
20150813:
The WITHOUT_ELFTOOLCHAIN_TOOLS src.conf(5) knob has been retired.
Control over building the ELF Tool Chain tools is now provided by
the WITHOUT_TOOLCHAIN knob.
20150810:
The polarity of Pulse Per Second (PPS) capture events with the
uart(4) driver has been corrected. Prior to this change the PPS
"assert" event corresponded to the trailing edge of a positive PPS
pulse and the "clear" event was the leading edge of the next pulse.
As the width of a PPS pulse in a typical GPS receiver is on the
order of 1 millisecond, most users will not notice any significant
difference with this change.
Anyone who has compensated for the historical polarity reversal by
configuring a negative offset equal to the pulse width will need to
remove that workaround.
20150809:
The default group assigned to /dev/dri entries has been changed
from 'wheel' to 'video' with the id of '44'. If you want to have
access to the dri devices please add yourself to the video group
with:
# pw groupmod video -m $USER
20150806:
The menu.rc and loader.rc files will now be replaced during
upgrades. Please migrate local changes to menu.rc.local and
loader.rc.local instead.
20150805:
GNU Binutils versions of addr2line, c++filt, nm, readelf, size,
strings and strip have been removed. The src.conf(5) knob
WITHOUT_ELFTOOLCHAIN_TOOLS no longer provides the binutils tools.
20150728:
As ZFS requires more kernel stack pages than is the default on some
architectures e.g. i386, it now warns if KSTACK_PAGES is less than
ZFS_MIN_KSTACK_PAGES (which is 4 at the time of writing).
Please consider using 'options KSTACK_PAGES=X' where X is greater
than or equal to ZFS_MIN_KSTACK_PAGES i.e. 4 in such configurations.
20150706:
sendmail has been updated to 8.15.2. Starting with FreeBSD 11.0
and sendmail 8.15, sendmail uses uncompressed IPv6 addresses by
default, i.e., they will not contain "::". For example, instead
of ::1, it will be 0:0:0:0:0:0:0:1. This permits a zero subnet
to have a more specific match, such as different map entries for
IPv6:0:0 vs IPv6:0. This change requires that configuration
data (including maps, files, classes, custom ruleset, etc.) must
use the same format, so make certain such configuration data is
upgrading. As a very simple check search for patterns like
'IPv6:[0-9a-fA-F:]*::' and 'IPv6::'. To return to the old
behavior, set the m4 option confUSE_COMPRESSED_IPV6_ADDRESSES or
the cf option UseCompressedIPv6Addresses.
Huge cleanup of random(4) code. * GENERAL - Update copyright. - Make kernel options for RANDOM_YARROW and RANDOM_DUMMY. Set neither to ON, which means we want Fortuna - If there is no 'device random' in the kernel, there will be NO random(4) device in the kernel, and the KERN_ARND sysctl will return nothing. With RANDOM_DUMMY there will be a random(4) that always blocks. - Repair kern.arandom (KERN_ARND sysctl). The old version went through arc4random(9) and was a bit weird. - Adjust arc4random stirring a bit - the existing code looks a little suspect. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Redo read_random(9) so as to duplicate random(4)'s read internals. This makes it a first-class citizen rather than a hack. - Move stuff out of locked regions when it does not need to be there. - Trim RANDOM_DEBUG printfs. Some are excess to requirement, some behind boot verbose. - Use SYSINIT to sequence the startup. - Fix init/deinit sysctl stuff. - Make relevant sysctls also tunables. - Add different harvesting "styles" to allow for different requirements (direct, queue, fast). - Add harvesting of FFS atime events. This needs to be checked for weighing down the FS code. - Add harvesting of slab allocator events. This needs to be checked for weighing down the allocator code. - Fix the random(9) manpage. - Loadable modules are not present for now. These will be re-engineered when the dust settles. - Use macros for locks. - Fix comments. * src/share/man/... - Update the man pages. * src/etc/... - The startup/shutdown work is done in D2924. * src/UPDATING - Add UPDATING announcement. * src/sys/dev/random/build.sh - Add copyright. - Add libz for unit tests. * src/sys/dev/random/dummy.c - Remove; no longer needed. Functionality incorporated into randomdev.*. * live_entropy_sources.c live_entropy_sources.h - Remove; content moved. - move content to randomdev.[ch] and optimise. * src/sys/dev/random/random_adaptors.c src/sys/dev/random/random_adaptors.h - Remove; plugability is no longer used. Compile-time algorithm selection is the way to go. * src/sys/dev/random/random_harvestq.c src/sys/dev/random/random_harvestq.h - Add early (re)boot-time randomness caching. * src/sys/dev/random/randomdev_soft.c src/sys/dev/random/randomdev_soft.h - Remove; no longer needed. * src/sys/dev/random/uint128.h - Provide a fake uint128_t; if a real one ever arrived, we can use that instead. All that is needed here is N=0, N++, N==0, and some localised trickery is used to manufacture a 128-bit 0ULLL. * src/sys/dev/random/unit_test.c src/sys/dev/random/unit_test.h - Improve unit tests; previously the testing human needed clairvoyance; now the test will do a basic check of compressibility. Clairvoyant talent is still a good idea. - This is still a long way off a proper unit test. * src/sys/dev/random/fortuna.c src/sys/dev/random/fortuna.h - Improve messy union to just uint128_t. - Remove unneeded 'static struct fortuna_start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) * src/sys/dev/random/yarrow.c src/sys/dev/random/yarrow.h - Improve messy union to just uint128_t. - Remove unneeded 'staic struct start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) - Fix some magic numbers elsewhere used as FAST and SLOW. Differential Revision: https://reviews.freebsd.org/D2025 Reviewed by: vsevolod,delphij,rwatson,trasz,jmg Approved by: so (delphij)
2015-06-30 17:00:45 +00:00
20150630:
The default kernel entropy-processing algorithm is now
Fortuna, replacing Yarrow.
Huge cleanup of random(4) code. * GENERAL - Update copyright. - Make kernel options for RANDOM_YARROW and RANDOM_DUMMY. Set neither to ON, which means we want Fortuna - If there is no 'device random' in the kernel, there will be NO random(4) device in the kernel, and the KERN_ARND sysctl will return nothing. With RANDOM_DUMMY there will be a random(4) that always blocks. - Repair kern.arandom (KERN_ARND sysctl). The old version went through arc4random(9) and was a bit weird. - Adjust arc4random stirring a bit - the existing code looks a little suspect. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Redo read_random(9) so as to duplicate random(4)'s read internals. This makes it a first-class citizen rather than a hack. - Move stuff out of locked regions when it does not need to be there. - Trim RANDOM_DEBUG printfs. Some are excess to requirement, some behind boot verbose. - Use SYSINIT to sequence the startup. - Fix init/deinit sysctl stuff. - Make relevant sysctls also tunables. - Add different harvesting "styles" to allow for different requirements (direct, queue, fast). - Add harvesting of FFS atime events. This needs to be checked for weighing down the FS code. - Add harvesting of slab allocator events. This needs to be checked for weighing down the allocator code. - Fix the random(9) manpage. - Loadable modules are not present for now. These will be re-engineered when the dust settles. - Use macros for locks. - Fix comments. * src/share/man/... - Update the man pages. * src/etc/... - The startup/shutdown work is done in D2924. * src/UPDATING - Add UPDATING announcement. * src/sys/dev/random/build.sh - Add copyright. - Add libz for unit tests. * src/sys/dev/random/dummy.c - Remove; no longer needed. Functionality incorporated into randomdev.*. * live_entropy_sources.c live_entropy_sources.h - Remove; content moved. - move content to randomdev.[ch] and optimise. * src/sys/dev/random/random_adaptors.c src/sys/dev/random/random_adaptors.h - Remove; plugability is no longer used. Compile-time algorithm selection is the way to go. * src/sys/dev/random/random_harvestq.c src/sys/dev/random/random_harvestq.h - Add early (re)boot-time randomness caching. * src/sys/dev/random/randomdev_soft.c src/sys/dev/random/randomdev_soft.h - Remove; no longer needed. * src/sys/dev/random/uint128.h - Provide a fake uint128_t; if a real one ever arrived, we can use that instead. All that is needed here is N=0, N++, N==0, and some localised trickery is used to manufacture a 128-bit 0ULLL. * src/sys/dev/random/unit_test.c src/sys/dev/random/unit_test.h - Improve unit tests; previously the testing human needed clairvoyance; now the test will do a basic check of compressibility. Clairvoyant talent is still a good idea. - This is still a long way off a proper unit test. * src/sys/dev/random/fortuna.c src/sys/dev/random/fortuna.h - Improve messy union to just uint128_t. - Remove unneeded 'static struct fortuna_start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) * src/sys/dev/random/yarrow.c src/sys/dev/random/yarrow.h - Improve messy union to just uint128_t. - Remove unneeded 'staic struct start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) - Fix some magic numbers elsewhere used as FAST and SLOW. Differential Revision: https://reviews.freebsd.org/D2025 Reviewed by: vsevolod,delphij,rwatson,trasz,jmg Approved by: so (delphij)
2015-06-30 17:00:45 +00:00
Assuming you have 'device random' in your kernel config
file, the configurations allow a kernel option to override
this default. You may choose *ONE* of:
Huge cleanup of random(4) code. * GENERAL - Update copyright. - Make kernel options for RANDOM_YARROW and RANDOM_DUMMY. Set neither to ON, which means we want Fortuna - If there is no 'device random' in the kernel, there will be NO random(4) device in the kernel, and the KERN_ARND sysctl will return nothing. With RANDOM_DUMMY there will be a random(4) that always blocks. - Repair kern.arandom (KERN_ARND sysctl). The old version went through arc4random(9) and was a bit weird. - Adjust arc4random stirring a bit - the existing code looks a little suspect. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Redo read_random(9) so as to duplicate random(4)'s read internals. This makes it a first-class citizen rather than a hack. - Move stuff out of locked regions when it does not need to be there. - Trim RANDOM_DEBUG printfs. Some are excess to requirement, some behind boot verbose. - Use SYSINIT to sequence the startup. - Fix init/deinit sysctl stuff. - Make relevant sysctls also tunables. - Add different harvesting "styles" to allow for different requirements (direct, queue, fast). - Add harvesting of FFS atime events. This needs to be checked for weighing down the FS code. - Add harvesting of slab allocator events. This needs to be checked for weighing down the allocator code. - Fix the random(9) manpage. - Loadable modules are not present for now. These will be re-engineered when the dust settles. - Use macros for locks. - Fix comments. * src/share/man/... - Update the man pages. * src/etc/... - The startup/shutdown work is done in D2924. * src/UPDATING - Add UPDATING announcement. * src/sys/dev/random/build.sh - Add copyright. - Add libz for unit tests. * src/sys/dev/random/dummy.c - Remove; no longer needed. Functionality incorporated into randomdev.*. * live_entropy_sources.c live_entropy_sources.h - Remove; content moved. - move content to randomdev.[ch] and optimise. * src/sys/dev/random/random_adaptors.c src/sys/dev/random/random_adaptors.h - Remove; plugability is no longer used. Compile-time algorithm selection is the way to go. * src/sys/dev/random/random_harvestq.c src/sys/dev/random/random_harvestq.h - Add early (re)boot-time randomness caching. * src/sys/dev/random/randomdev_soft.c src/sys/dev/random/randomdev_soft.h - Remove; no longer needed. * src/sys/dev/random/uint128.h - Provide a fake uint128_t; if a real one ever arrived, we can use that instead. All that is needed here is N=0, N++, N==0, and some localised trickery is used to manufacture a 128-bit 0ULLL. * src/sys/dev/random/unit_test.c src/sys/dev/random/unit_test.h - Improve unit tests; previously the testing human needed clairvoyance; now the test will do a basic check of compressibility. Clairvoyant talent is still a good idea. - This is still a long way off a proper unit test. * src/sys/dev/random/fortuna.c src/sys/dev/random/fortuna.h - Improve messy union to just uint128_t. - Remove unneeded 'static struct fortuna_start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) * src/sys/dev/random/yarrow.c src/sys/dev/random/yarrow.h - Improve messy union to just uint128_t. - Remove unneeded 'staic struct start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) - Fix some magic numbers elsewhere used as FAST and SLOW. Differential Revision: https://reviews.freebsd.org/D2025 Reviewed by: vsevolod,delphij,rwatson,trasz,jmg Approved by: so (delphij)
2015-06-30 17:00:45 +00:00
options RANDOM_YARROW # Legacy /dev/random algorithm.
options RANDOM_DUMMY # Blocking-only driver.
Huge cleanup of random(4) code. * GENERAL - Update copyright. - Make kernel options for RANDOM_YARROW and RANDOM_DUMMY. Set neither to ON, which means we want Fortuna - If there is no 'device random' in the kernel, there will be NO random(4) device in the kernel, and the KERN_ARND sysctl will return nothing. With RANDOM_DUMMY there will be a random(4) that always blocks. - Repair kern.arandom (KERN_ARND sysctl). The old version went through arc4random(9) and was a bit weird. - Adjust arc4random stirring a bit - the existing code looks a little suspect. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Redo read_random(9) so as to duplicate random(4)'s read internals. This makes it a first-class citizen rather than a hack. - Move stuff out of locked regions when it does not need to be there. - Trim RANDOM_DEBUG printfs. Some are excess to requirement, some behind boot verbose. - Use SYSINIT to sequence the startup. - Fix init/deinit sysctl stuff. - Make relevant sysctls also tunables. - Add different harvesting "styles" to allow for different requirements (direct, queue, fast). - Add harvesting of FFS atime events. This needs to be checked for weighing down the FS code. - Add harvesting of slab allocator events. This needs to be checked for weighing down the allocator code. - Fix the random(9) manpage. - Loadable modules are not present for now. These will be re-engineered when the dust settles. - Use macros for locks. - Fix comments. * src/share/man/... - Update the man pages. * src/etc/... - The startup/shutdown work is done in D2924. * src/UPDATING - Add UPDATING announcement. * src/sys/dev/random/build.sh - Add copyright. - Add libz for unit tests. * src/sys/dev/random/dummy.c - Remove; no longer needed. Functionality incorporated into randomdev.*. * live_entropy_sources.c live_entropy_sources.h - Remove; content moved. - move content to randomdev.[ch] and optimise. * src/sys/dev/random/random_adaptors.c src/sys/dev/random/random_adaptors.h - Remove; plugability is no longer used. Compile-time algorithm selection is the way to go. * src/sys/dev/random/random_harvestq.c src/sys/dev/random/random_harvestq.h - Add early (re)boot-time randomness caching. * src/sys/dev/random/randomdev_soft.c src/sys/dev/random/randomdev_soft.h - Remove; no longer needed. * src/sys/dev/random/uint128.h - Provide a fake uint128_t; if a real one ever arrived, we can use that instead. All that is needed here is N=0, N++, N==0, and some localised trickery is used to manufacture a 128-bit 0ULLL. * src/sys/dev/random/unit_test.c src/sys/dev/random/unit_test.h - Improve unit tests; previously the testing human needed clairvoyance; now the test will do a basic check of compressibility. Clairvoyant talent is still a good idea. - This is still a long way off a proper unit test. * src/sys/dev/random/fortuna.c src/sys/dev/random/fortuna.h - Improve messy union to just uint128_t. - Remove unneeded 'static struct fortuna_start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) * src/sys/dev/random/yarrow.c src/sys/dev/random/yarrow.h - Improve messy union to just uint128_t. - Remove unneeded 'staic struct start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) - Fix some magic numbers elsewhere used as FAST and SLOW. Differential Revision: https://reviews.freebsd.org/D2025 Reviewed by: vsevolod,delphij,rwatson,trasz,jmg Approved by: so (delphij)
2015-06-30 17:00:45 +00:00
If you have neither, you get Fortuna. For most people,
read no further, Fortuna will give a /dev/random that works
like it always used to, and the difference will be irrelevant.
If you remove 'device random', you get *NO* kernel-processed
2015-11-17 19:43:40 +00:00
entropy at all. This may be acceptable to folks building
Huge cleanup of random(4) code. * GENERAL - Update copyright. - Make kernel options for RANDOM_YARROW and RANDOM_DUMMY. Set neither to ON, which means we want Fortuna - If there is no 'device random' in the kernel, there will be NO random(4) device in the kernel, and the KERN_ARND sysctl will return nothing. With RANDOM_DUMMY there will be a random(4) that always blocks. - Repair kern.arandom (KERN_ARND sysctl). The old version went through arc4random(9) and was a bit weird. - Adjust arc4random stirring a bit - the existing code looks a little suspect. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Redo read_random(9) so as to duplicate random(4)'s read internals. This makes it a first-class citizen rather than a hack. - Move stuff out of locked regions when it does not need to be there. - Trim RANDOM_DEBUG printfs. Some are excess to requirement, some behind boot verbose. - Use SYSINIT to sequence the startup. - Fix init/deinit sysctl stuff. - Make relevant sysctls also tunables. - Add different harvesting "styles" to allow for different requirements (direct, queue, fast). - Add harvesting of FFS atime events. This needs to be checked for weighing down the FS code. - Add harvesting of slab allocator events. This needs to be checked for weighing down the allocator code. - Fix the random(9) manpage. - Loadable modules are not present for now. These will be re-engineered when the dust settles. - Use macros for locks. - Fix comments. * src/share/man/... - Update the man pages. * src/etc/... - The startup/shutdown work is done in D2924. * src/UPDATING - Add UPDATING announcement. * src/sys/dev/random/build.sh - Add copyright. - Add libz for unit tests. * src/sys/dev/random/dummy.c - Remove; no longer needed. Functionality incorporated into randomdev.*. * live_entropy_sources.c live_entropy_sources.h - Remove; content moved. - move content to randomdev.[ch] and optimise. * src/sys/dev/random/random_adaptors.c src/sys/dev/random/random_adaptors.h - Remove; plugability is no longer used. Compile-time algorithm selection is the way to go. * src/sys/dev/random/random_harvestq.c src/sys/dev/random/random_harvestq.h - Add early (re)boot-time randomness caching. * src/sys/dev/random/randomdev_soft.c src/sys/dev/random/randomdev_soft.h - Remove; no longer needed. * src/sys/dev/random/uint128.h - Provide a fake uint128_t; if a real one ever arrived, we can use that instead. All that is needed here is N=0, N++, N==0, and some localised trickery is used to manufacture a 128-bit 0ULLL. * src/sys/dev/random/unit_test.c src/sys/dev/random/unit_test.h - Improve unit tests; previously the testing human needed clairvoyance; now the test will do a basic check of compressibility. Clairvoyant talent is still a good idea. - This is still a long way off a proper unit test. * src/sys/dev/random/fortuna.c src/sys/dev/random/fortuna.h - Improve messy union to just uint128_t. - Remove unneeded 'static struct fortuna_start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) * src/sys/dev/random/yarrow.c src/sys/dev/random/yarrow.h - Improve messy union to just uint128_t. - Remove unneeded 'staic struct start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) - Fix some magic numbers elsewhere used as FAST and SLOW. Differential Revision: https://reviews.freebsd.org/D2025 Reviewed by: vsevolod,delphij,rwatson,trasz,jmg Approved by: so (delphij)
2015-06-30 17:00:45 +00:00
embedded systems, but has complications. Carry on reading,
and it is assumed you know what you need.
*PLEASE* read random(4) and random(9) if you are in the
2015-11-17 19:43:40 +00:00
habit of tweaking kernel configs, and/or if you are a member
Huge cleanup of random(4) code. * GENERAL - Update copyright. - Make kernel options for RANDOM_YARROW and RANDOM_DUMMY. Set neither to ON, which means we want Fortuna - If there is no 'device random' in the kernel, there will be NO random(4) device in the kernel, and the KERN_ARND sysctl will return nothing. With RANDOM_DUMMY there will be a random(4) that always blocks. - Repair kern.arandom (KERN_ARND sysctl). The old version went through arc4random(9) and was a bit weird. - Adjust arc4random stirring a bit - the existing code looks a little suspect. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Redo read_random(9) so as to duplicate random(4)'s read internals. This makes it a first-class citizen rather than a hack. - Move stuff out of locked regions when it does not need to be there. - Trim RANDOM_DEBUG printfs. Some are excess to requirement, some behind boot verbose. - Use SYSINIT to sequence the startup. - Fix init/deinit sysctl stuff. - Make relevant sysctls also tunables. - Add different harvesting "styles" to allow for different requirements (direct, queue, fast). - Add harvesting of FFS atime events. This needs to be checked for weighing down the FS code. - Add harvesting of slab allocator events. This needs to be checked for weighing down the allocator code. - Fix the random(9) manpage. - Loadable modules are not present for now. These will be re-engineered when the dust settles. - Use macros for locks. - Fix comments. * src/share/man/... - Update the man pages. * src/etc/... - The startup/shutdown work is done in D2924. * src/UPDATING - Add UPDATING announcement. * src/sys/dev/random/build.sh - Add copyright. - Add libz for unit tests. * src/sys/dev/random/dummy.c - Remove; no longer needed. Functionality incorporated into randomdev.*. * live_entropy_sources.c live_entropy_sources.h - Remove; content moved. - move content to randomdev.[ch] and optimise. * src/sys/dev/random/random_adaptors.c src/sys/dev/random/random_adaptors.h - Remove; plugability is no longer used. Compile-time algorithm selection is the way to go. * src/sys/dev/random/random_harvestq.c src/sys/dev/random/random_harvestq.h - Add early (re)boot-time randomness caching. * src/sys/dev/random/randomdev_soft.c src/sys/dev/random/randomdev_soft.h - Remove; no longer needed. * src/sys/dev/random/uint128.h - Provide a fake uint128_t; if a real one ever arrived, we can use that instead. All that is needed here is N=0, N++, N==0, and some localised trickery is used to manufacture a 128-bit 0ULLL. * src/sys/dev/random/unit_test.c src/sys/dev/random/unit_test.h - Improve unit tests; previously the testing human needed clairvoyance; now the test will do a basic check of compressibility. Clairvoyant talent is still a good idea. - This is still a long way off a proper unit test. * src/sys/dev/random/fortuna.c src/sys/dev/random/fortuna.h - Improve messy union to just uint128_t. - Remove unneeded 'static struct fortuna_start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) * src/sys/dev/random/yarrow.c src/sys/dev/random/yarrow.h - Improve messy union to just uint128_t. - Remove unneeded 'staic struct start_cache'. - Tighten up up arithmetic. - Provide a method to allow eternal junk to be introduced; harden it against blatant by compress/hashing. - Assert that locks are held correctly. - Fix the nasty pre- and post-read overloading by providing explictit functions to do these tasks. - Turn into self-sufficient module (no longer requires randomdev_soft.[ch]) - Fix some magic numbers elsewhere used as FAST and SLOW. Differential Revision: https://reviews.freebsd.org/D2025 Reviewed by: vsevolod,delphij,rwatson,trasz,jmg Approved by: so (delphij)
2015-06-30 17:00:45 +00:00
of the embedded community, wanting specific and not-usual
behaviour from your security subsystems.
NOTE!! If you use RANDOM_DUMMY and/or have no 'device
random', you will NOT have a functioning /dev/random, and
many cryptographic features will not work, including SSH.
You may also find strange behaviour from the random(3) set
of library functions, in particular sranddev(3), srandomdev(3)
and arc4random(3). The reason for this is that the KERN_ARND
sysctl only returns entropy if it thinks it has some to
share, and with RANDOM_DUMMY or no 'device random' this
will never happen.
20150623:
An additional fix for the issue described in the 20150614 sendmail
entry below has been committed in revision 284717.
20150616:
FreeBSD's old make (fmake) has been removed from the system. It is
available as the devel/fmake port or via pkg install fmake.
2015-07-19 14:34:35 +00:00
20150615:
The fix for the issue described in the 20150614 sendmail entry
below has been committed in revision 284436. The work
around described in that entry is no longer needed unless the
default setting is overridden by a confDH_PARAMETERS configuration
setting of '5' or pointing to a 512 bit DH parameter file.
20150614:
ALLOW_DEPRECATED_ATF_TOOLS/ATFFILE support has been removed from
atf.test.mk (included from bsd.test.mk). Please upgrade devel/atf
and devel/kyua to version 0.20+ and adjust any calling code to work
with Kyuafile and kyua.
20150614:
The import of openssl to address the FreeBSD-SA-15:10.openssl
security advisory includes a change which rejects handshakes
with DH parameters below 768 bits. sendmail releases prior
to 8.15.2 (not yet released), defaulted to a 512 bit
DH parameter setting for client connections. To work around
this interoperability, sendmail can be configured to use a
2048 bit DH parameter by:
2015-07-19 14:34:35 +00:00
1. Edit /etc/mail/`hostname`.mc
2. If a setting for confDH_PARAMETERS does not exist or
exists and is set to a string beginning with '5',
replace it with '2'.
3. If a setting for confDH_PARAMETERS exists and is set to
a file path, create a new file with:
openssl dhparam -out /path/to/file 2048
4. Rebuild the .cf file:
cd /etc/mail/; make; make install
5. Restart sendmail:
cd /etc/mail/; make restart
A sendmail patch is coming, at which time this file will be
updated.
20150604:
Generation of legacy formatted entries have been disabled by default
in pwd_mkdb(8), as all base system consumers of the legacy formatted
entries were converted to use the new format by default when the new,
machine independent format have been added and supported since FreeBSD
5.x.
Please see the pwd_mkdb(8) manual page for further details.
20150525:
Clang and llvm have been upgraded to 3.6.1 release. Please see the
20141231 entry below for information about prerequisites and upgrading,
if you are not already using 3.5.0 or higher.
20150521:
TI platform code switched to using vendor DTS files and this update
may break existing systems running on Beaglebone, Beaglebone Black,
and Pandaboard:
- dtb files should be regenerated/reinstalled. Filenames are the
same but content is different now
- GPIO addressing was changed, now each GPIO bank (32 pins per bank)
has its own /dev/gpiocX device, e.g. pin 121 on /dev/gpioc0 in old
addressing scheme is now pin 25 on /dev/gpioc3.
- Pandaboard: /etc/ttys should be updated, serial console device is
now /dev/ttyu2, not /dev/ttyu0
20150501:
soelim(1) from gnu/usr.bin/groff has been replaced by usr.bin/soelim.
If you need the GNU extension from groff soelim(1), install groff
from package: pkg install groff, or via ports: textproc/groff.
20150423:
chmod, chflags, chown and chgrp now affect symlinks in -R mode as
defined in symlink(7); previously symlinks were silently ignored.
20150415:
The const qualifier has been removed from iconv(3) to comply with
POSIX. The ports tree is aware of this from r384038 onwards.
20150416:
Libraries specified by LIBADD in Makefiles must have a corresponding
DPADD_<lib> variable to ensure correct dependencies. This is now
enforced in src.libnames.mk.
20150324:
From legacy ata(4) driver was removed support for SATA controllers
supported by more functional drivers ahci(4), siis(4) and mvs(4).
Kernel modules ataahci and ataadaptec were removed completely,
replaced by ahci and mvs modules respectively.
20150315:
Clang, llvm and lldb have been upgraded to 3.6.0 release. Please see
the 20141231 entry below for information about prerequisites and
upgrading, if you are not already using 3.5.0 or higher.
20150307:
The 32-bit PowerPC kernel has been changed to a position-independent
executable. This can only be booted with a version of loader(8)
newer than January 31, 2015, so make sure to update both world and
kernel before rebooting.
20150217:
If you are running a -CURRENT kernel since r273872 (Oct 30th, 2014),
but before r278950, the RNG was not seeded properly. Immediately
upgrade the kernel to r278950 or later and regenerate any keys (e.g.
ssh keys or openssl keys) that were generated w/ a kernel from that
2015-02-21 13:00:52 +00:00
range. This does not affect programs that directly used /dev/random
or /dev/urandom. All userland uses of arc4random(3) are affected.
20150210:
The autofs(4) ABI was changed in order to restore binary compatibility
with 10.1-RELEASE. The automountd(8) daemon needs to be rebuilt to work
with the new kernel.
20150131:
The powerpc64 kernel has been changed to a position-independent
executable. This can only be booted with a new version of loader(8),
so make sure to update both world and kernel before rebooting.
20150118:
Clang and llvm have been upgraded to 3.5.1 release. This is a bugfix
only release, no new features have been added. Please see the 20141231
entry below for information about prerequisites and upgrading, if you
are not already using 3.5.0.
20150107:
ELF tools addr2line, elfcopy (strip), nm, size, and strings are now
taken from the ELF Tool Chain project rather than GNU binutils. They
should be drop-in replacements, with the addition of arm64 support.
The WITHOUT_ELFTOOLCHAIN_TOOLS= knob may be used to obtain the
binutils tools, if necessary. See 20150805 for updated information.
20150105:
The default Unbound configuration now enables remote control
using a local socket. Users who have already enabled the
local_unbound service should regenerate their configuration
by running "service local_unbound setup" as root.
2015-07-19 14:34:35 +00:00
20150102:
The GNU texinfo and GNU info pages have been removed.
To be able to view GNU info pages please install texinfo from ports.
20141231:
Clang, llvm and lldb have been upgraded to 3.5.0 release.
2014-12-31 18:11:09 +00:00
As of this release, a prerequisite for building clang, llvm and lldb is
a C++11 capable compiler and C++11 standard library. This means that to
be able to successfully build the cross-tools stage of buildworld, with
clang as the bootstrap compiler, your system compiler or cross compiler
should either be clang 3.3 or later, or gcc 4.8 or later, and your
system C++ library should be libc++, or libdstdc++ from gcc 4.8 or
later.
2014-12-31 18:11:09 +00:00
On any standard FreeBSD 10.x or 11.x installation, where clang and
libc++ are on by default (that is, on x86 or arm), this should work out
of the box.
On 9.x installations where clang is enabled by default, e.g. on x86 and
powerpc, libc++ will not be enabled by default, so libc++ should be
built (with clang) and installed first. If both clang and libc++ are
missing, build clang first, then use it to build libc++.
On 8.x and earlier installations, upgrade to 9.x first, and then follow
the instructions for 9.x above.
2014-12-31 18:11:09 +00:00
Sparc64 and mips users are unaffected, as they still use gcc 4.2.1 by
default, and do not build clang.
Many embedded systems are resource constrained, and will not be able to
build clang in a reasonable time, or in some cases at all. In those
cases, cross building bootable systems on amd64 is a workaround.
This new version of clang introduces a number of new warnings, of which
the following are most likely to appear:
-Wabsolute-value
This warns in two cases, for both C and C++:
* When the code is trying to take the absolute value of an unsigned
quantity, which is effectively a no-op, and almost never what was
intended. The code should be fixed, if at all possible. If you are
sure that the unsigned quantity can be safely cast to signed, without
loss of information or undefined behavior, you can add an explicit
cast, or disable the warning.
* When the code is trying to take an absolute value, but the called
abs() variant is for the wrong type, which can lead to truncation.
If you want to disable the warning instead of fixing the code, please
make sure that truncation will not occur, or it might lead to unwanted
side-effects.
-Wtautological-undefined-compare and
-Wundefined-bool-conversion
These warn when C++ code is trying to compare 'this' against NULL, while
'this' should never be NULL in well-defined C++ code. However, there is
some legacy (pre C++11) code out there, which actively abuses this
feature, which was less strictly defined in previous C++ versions.
Squid and openjdk do this, for example. The warning can be turned off
for C++98 and earlier, but compiling the code in C++11 mode might result
in unexpected behavior; for example, the parts of the program that are
unreachable could be optimized away.
20141222:
The old NFS client and server (kernel options NFSCLIENT, NFSSERVER)
kernel sources have been removed. The .h files remain, since some
utilities include them. This will need to be fixed later.
If "mount -t oldnfs ..." is attempted, it will fail.
If the "-o" option on mountd(8), nfsd(8) or nfsstat(1) is used,
the utilities will report errors.
20141121:
The handling of LOCAL_LIB_DIRS has been altered to skip addition of
directories to top level SUBDIR variable when their parent
directory is included in LOCAL_DIRS. Users with build systems with
such hierarchies and without SUBDIR entries in the parent
directory Makefiles should add them or add the directories to
LOCAL_DIRS.
20141109:
2015-01-05 13:46:37 +00:00
faith(4) and faithd(8) have been removed from the base system. Faith
has been obsolete for a very long time.
20141104:
vt(4), the new console driver, is enabled by default. It brings
support for Unicode and double-width characters, as well as
support for UEFI and integration with the KMS kernel video
drivers.
You may need to update your console settings in /etc/rc.conf,
most probably the keymap. During boot, /etc/rc.d/syscons will
indicate what you need to do.
vt(4) still has issues and lacks some features compared to
syscons(4). See the wiki for up-to-date information:
https://wiki.freebsd.org/Newcons
If you want to keep using syscons(4), you can do so by adding
the following line to /boot/loader.conf:
kern.vty=sc
20141102:
pjdfstest has been integrated into kyua as an opt-in test suite.
2015-01-05 13:46:37 +00:00
Please see share/doc/pjdfstest/README for more details on how to
execute it.
20141009:
gperf has been removed from the base system for architectures
that use clang. Ports that require gperf will obtain it from the
devel/gperf port.
20140923:
pjdfstest has been moved from tools/regression/pjdfstest to
contrib/pjdfstest .
20140922:
At svn r271982, The default linux compat kernel ABI has been adjusted
to 2.6.18 in support of the linux-c6 compat ports infrastructure
update. If you wish to continue using the linux-f10 compat ports,
add compat.linux.osrelease=2.6.16 to your local sysctl.conf. Users are
encouraged to update their linux-compat packages to linux-c6 during
their next update cycle.
20140729:
The ofwfb driver, used to provide a graphics console on PowerPC when
2015-01-05 13:46:37 +00:00
using vt(4), no longer allows mmap() of all physical memory. This
will prevent Xorg on PowerPC with some ATI graphics cards from
initializing properly unless x11-servers/xorg-server is updated to
1.12.4_8 or newer.
20140723:
The xdev targets have been converted to using TARGET and
TARGET_ARCH instead of XDEV and XDEV_ARCH.
20140719:
The default unbound configuration has been modified to address
issues with reverse lookups on networks that use private
address ranges. If you use the local_unbound service, run
"service local_unbound setup" as root to regenerate your
configuration, then "service local_unbound reload" to load the
new configuration.
20140709:
The GNU texinfo and GNU info pages are not built and installed
anymore, WITH_INFO knob has been added to allow to built and install
them again.
UPDATE: see 20150102 entry on texinfo's removal
20140708:
The GNU readline library is now an INTERNALLIB - that is, it is
statically linked into consumers (GDB and variants) in the base
system, and the shared library is no longer installed. The
devel/readline port is available for third party software that
requires readline.
20140702:
The Itanium architecture (ia64) has been removed from the list of
known architectures. This is the first step in the removal of the
architecture.
2014-07-01 22:32:11 +00:00
20140701:
Commit r268115 has added NFSv4.1 server support, merged from
projects/nfsv4.1-server. Since this includes changes to the
internal interfaces between the NFS related modules, a full
build of the kernel and modules will be necessary.
__FreeBSD_version has been bumped.
20140629:
The WITHOUT_VT_SUPPORT kernel config knob has been renamed
WITHOUT_VT. (The other _SUPPORT knobs have a consistent meaning
which differs from the behaviour controlled by this knob.)
2014-06-24 18:56:45 +00:00
20140619:
Maximal length of the serial number in CTL was increased from 16 to
64 chars, that breaks ABI. All CTL-related tools, such as ctladm
and ctld, need to be rebuilt to work with a new kernel.
20140606:
The libatf-c and libatf-c++ major versions were downgraded to 0 and
1 respectively to match the upstream numbers. They were out of
sync because, when they were originally added to FreeBSD, the
upstream versions were not respected. These libraries are private
and not yet built by default, so renumbering them should be a
non-issue. However, unclean source trees will yield broken test
programs once the operator executes "make delete-old-libs" after a
"make installworld".
Additionally, the atf-sh binary was made private by moving it into
/usr/libexec/. Already-built shell test programs will keep the
path to the old binary so they will break after "make delete-old"
is run.
If you are using WITH_TESTS=yes (not the default), wipe the object
tree and rebuild from scratch to prevent spurious test failures.
This is only needed once: the misnumbered libraries and misplaced
binaries have been added to OptionalObsoleteFiles.inc so they will
be removed during a clean upgrade.
20140512:
Clang and llvm have been upgraded to 3.4.1 release.
20140508:
We bogusly installed src.opts.mk in /usr/share/mk. This file should
be removed to avoid issues in the future (and has been added to
ObsoleteFiles.inc).
20140505:
/etc/src.conf now affects only builds of the FreeBSD src tree. In the
past, it affected all builds that used the bsd.*.mk files. The old
behavior was a bug, but people may have relied upon it. To get this
behavior back, you can .include /etc/src.conf from /etc/make.conf
(which is still global and isn't changed). This also changes the
behavior of incremental builds inside the tree of individual
directories. Set MAKESYSPATH to ".../share/mk" to do that.
Although this has survived make universe and some upgrade scenarios,
other upgrade scenarios may have broken. At least one form of
temporary breakage was fixed with MAKESYSPATH settings for buildworld
as well... In cases where MAKESYSPATH isn't working with this
setting, you'll need to set it to the full path to your tree.
One side effect of all this cleaning up is that bsd.compiler.mk
is no longer implicitly included by bsd.own.mk. If you wish to
use COMPILER_TYPE, you must now explicitly include bsd.compiler.mk
as well.
20140430:
The lindev device has been removed since /dev/full has been made a
standard device. __FreeBSD_version has been bumped.
20140424:
The knob WITHOUT_VI was added to the base system, which controls
building ex(1), vi(1), etc. Older releases of FreeBSD required ex(1)
in order to reorder files share/termcap and didn't build ex(1) as a
build tool, so building/installing with WITH_VI is highly advised for
build hosts for older releases.
This issue has been fixed in stable/9 and stable/10 in r277022 and
r276991, respectively.
2014-04-18 17:06:18 +00:00
20140418:
The YES_HESIOD knob has been removed. It has been obsolete for
a decade. Please move to using WITH_HESIOD instead or your builds
will silently lack HESIOD.
20140405:
The uart(4) driver has been changed with respect to its handling
of the low-level console. Previously the uart(4) driver prevented
any process from changing the baudrate or the CLOCAL and HUPCL
control flags. By removing the restrictions, operators can make
changes to the serial console port without having to reboot.
However, when getty(8) is started on the serial device that is
associated with the low-level console, a misconfigured terminal
line in /etc/ttys will now have a real impact.
Before upgrading the kernel, make sure that /etc/ttys has the
serial console device configured as 3wire without baudrate to
preserve the previous behaviour. E.g:
ttyu0 "/usr/libexec/getty 3wire" vt100 on secure
20140306:
Support for libwrap (TCP wrappers) in rpcbind was disabled by default
to improve performance. To re-enable it, if needed, run rpcbind
with command line option -W.
2014-03-03 16:34:42 +00:00
20140226:
Switched back to the GPL dtc compiler due to updates in the upstream
dts files not being supported by the BSDL dtc compiler. You will need
to rebuild your kernel toolchain to pick up the new compiler. Core dumps
may result while building dtb files during a kernel build if you fail
to do so. Set WITHOUT_GPL_DTC if you require the BSDL compiler.
20140216:
Clang and llvm have been upgraded to 3.4 release.
20140216:
The nve(4) driver has been removed. Please use the nfe(4) driver
for NVIDIA nForce MCP Ethernet adapters instead.
20140212:
An ABI incompatibility crept into the libc++ 3.4 import in r261283.
This could cause certain C++ applications using shared libraries built
against the previous version of libc++ to crash. The incompatibility
has now been fixed, but any C++ applications or shared libraries built
between r261283 and r261801 should be recompiled.
20140204:
OpenSSH will now ignore errors caused by kernel lacking of Capsicum
capability mode support. Please note that enabling the feature in
kernel is still highly recommended.
20140131:
OpenSSH is now built with sandbox support, and will use sandbox as
the default privilege separation method. This requires Capsicum
capability mode support in kernel.
2014-01-28 19:07:28 +00:00
20140128:
The libelf and libdwarf libraries have been updated to newer
versions from upstream. Shared library version numbers for
these two libraries were bumped. Any ports or binaries
requiring these two libraries should be recompiled.
__FreeBSD_version is bumped to 1100006.
20140110:
If a Makefile in a tests/ directory was auto-generating a Kyuafile
instead of providing an explicit one, this would prevent such
Makefile from providing its own Kyuafile in the future during
NO_CLEAN builds. This has been fixed in the Makefiles but manual
intervention is needed to clean an objdir if you use NO_CLEAN:
# find /usr/obj -name Kyuafile | xargs rm -f
20131213:
The behavior of gss_pseudo_random() for the krb5 mechanism
has changed, for applications requesting a longer random string
than produced by the underlying enctype's pseudo-random() function.
In particular, the random string produced from a session key of
enctype aes256-cts-hmac-sha1-96 or aes256-cts-hmac-sha1-96 will
be different at the 17th octet and later, after this change.
The counter used in the PRF+ construction is now encoded as a
big-endian integer in accordance with RFC 4402.
__FreeBSD_version is bumped to 1100004.
20131108:
The WITHOUT_ATF build knob has been removed and its functionality
has been subsumed into the more generic WITHOUT_TESTS. If you were
using the former to disable the build of the ATF libraries, you
should change your settings to use the latter.
20131025:
The default version of mtree is nmtree which is obtained from
NetBSD. The output is generally the same, but may vary
slightly. If you found you need identical output adding
"-F freebsd9" to the command line should do the trick. For the
time being, the old mtree is available as fmtree.
20131014:
libbsdyml has been renamed to libyaml and moved to /usr/lib/private.
This will break ports-mgmt/pkg. Rebuild the port, or upgrade to pkg
1.1.4_8 and verify bsdyml not linked in, before running "make
delete-old-libs":
# make -C /usr/ports/ports-mgmt/pkg build deinstall install clean
or
# pkg install pkg; ldd /usr/local/sbin/pkg | grep bsdyml
20131010:
The stable/10 branch has been created in subversion from head
revision r256279.
COMMON ITEMS:
General Notes
-------------
Avoid using make -j when upgrading. While generally safe, there are
sometimes problems using -j to upgrade. If your upgrade fails with
-j, please try again without -j. From time to time in the past there
have been problems using -j with buildworld and/or installworld. This
is especially true when upgrading between "distant" versions (eg one
that cross a major release boundary or several minor releases, or when
several months have passed on the -current branch).
Sometimes, obscure build problems are the result of environment
poisoning. This can happen because the make utility reads its
environment when searching for values for global variables. To run
your build attempts in an "environmental clean room", prefix all make
commands with 'env -i '. See the env(1) manual page for more details.
When upgrading from one major version to another it is generally best to
upgrade to the latest code in the currently installed branch first, then
do an upgrade to the new branch. This is the best-tested upgrade path,
and has the highest probability of being successful. Please try this
approach if you encounter problems with a major version upgrade. Since
2018-01-12 17:33:20 +00:00
the stable 4.x branch point, one has generally been able to upgrade from
anywhere in the most recent stable branch to head / current (or even the
last couple of stable branches). See the top of this file when there's
an exception.
When upgrading a live system, having a root shell around before
installing anything can help undo problems. Not having a root shell
around can lead to problems if pam has changed too much from your
starting point to allow continued authentication after the upgrade.
This file should be read as a log of events. When a later event changes
information of a prior event, the prior event should not be deleted.
Instead, a pointer to the entry with the new information should be
placed in the old entry. Readers of this file should also sanity check
older entries before relying on them blindly. Authors of new entries
should write them with this in mind.
Merge ZFS version 15 and almost all OpenSolaris bugfixes referenced in Solaris 10 updates 141445-09 and 142901-14. Detailed information: (OpenSolaris revisions and Bug IDs, Solaris 10 patch numbers) 7844:effed23820ae 6755435 zfs_open() and zfs_close() needs to use ZFS_ENTER/ZFS_VERIFY_ZP (141445-01) 7897:e520d8258820 6748436 inconsistent zpool.cache in boot_archive could panic a zfs root filesystem upon boot-up (141445-01) 7965:b795da521357 6740164 zpool attach can create an illegal root pool (141909-02) 8084:b811cc60d650 6769612 zpool_import() will continue to write to cachefile even if altroot is set (N/A) 8121:7fd09d4ebd9c 6757430 want an option for zdb to disable space map loading and leak tracking (141445-01) 8129:e4f45a0bfbb0 6542860 ASSERT: reason != VDEV_LABEL_REMOVE||vdev_inuse(vd, crtxg, reason, 0) (141445-01) 8188:fd00c0a81e80 6761100 want zdb option to select older uberblocks (141445-01) 8190:6eeea43ced42 6774886 zfs_setattr() won't allow ndmp to restore SUNWattr_rw (141445-01) 8225:59a9961c2aeb 6737463 panic while trying to write out config file if root pool import fails (141445-01) 8227:f7d7be9b1f56 6765294 Refactor replay (141445-01) 8228:51e9ca9ee3a5 6572357 libzfs should do more to avoid mnttab lookups (141909-01) 6572376 zfs_iter_filesystems and zfs_iter_snapshots get objset stats twice (141909-01) 8241:5a60f16123ba 6328632 zpool offline is a bit too conservative (141445-01) 6739487 ASSERT: txg <= spa_final_txg due to scrub/export race (141445-01) 6767129 ASSERT: cvd->vdev_isspare, in spa_vdev_detach() (141445-01) 6747698 checksum failures after offline -t / export / import / scrub (141445-01) 6745863 ZFS writes to disk after it has been offlined (141445-01) 6722540 50% slowdown on scrub/resilver with certain vdev configurations (141445-01) 6759999 resilver logic rewrites ditto blocks on both source and destination (141445-01) 6758107 I/O should never suspend during spa_load() (141445-01) 6776548 codereview(1) runs off the page when faced with multi-line comments (N/A) 6761406 AMD errata 91 workaround doesn't work on 64-bit systems (141445-01) 8242:e46e4b2f0a03 6770866 GRUB/ZFS should require physical path or devid, but not both (141445-01) 8269:03a7e9050cfd 6674216 "zfs share" doesn't work, but "zfs set sharenfs=on" does (141445-01) 6621164 $SRC/cmd/zfs/zfs_main.c seems to have a syntax error in the translation note (141445-01) 6635482 i18n problems in libzfs_dataset.c and zfs_main.c (141445-01) 6595194 "zfs get" VALUE column is as wide as NAME (141445-01) 6722991 vdev_disk.c: error checking for ddi_pathname_to_dev_t() must test for NODEV (141445-01) 6396518 ASSERT strings shouldn't be pre-processed (141445-01) 8274:846b39508aff 6713916 scrub/resilver needlessly decompress data (141445-01) 8343:655db2375fed 6739553 libzfs_status msgid table is out of sync (141445-01) 6784104 libzfs unfairly rejects numerical values greater than 2^63 (141445-01) 6784108 zfs_realloc() should not free original memory on failure (141445-01) 8525:e0e0e525d0f8 6788830 set large value to reservation cause core dump (141445-01) 6791064 want sysevents for ZFS scrub (141445-01) 6791066 need to be able to set cachefile on faulted pools (141445-01) 6791071 zpool_do_import() should not enable datasets on faulted pools (141445-01) 6792134 getting multiple properties on a faulted pool leads to confusion (141445-01) 8547:bcc7b46e5ff7 6792884 Vista clients cannot access .zfs (141445-01) 8632:36ef517870a3 6798384 It can take a village to raise a zio (141445-01) 8636:7e4ce9158df3 6551866 deadlock between zfs_write(), zfs_freesp(), and zfs_putapage() (141909-01) 6504953 zfs_getpage() misunderstands VOP_GETPAGE() interface (141909-01) 6702206 ZFS read/writer lock contention throttles sendfile() benchmark (141445-01) 6780491 Zone on a ZFS filesystem has poor fork/exec performance (141445-01) 6747596 assertion failed: DVA_EQUAL(BP_IDENTITY(&zio->io_bp_orig), BP_IDENTITY(zio->io_bp))); (141445-01) 8692:692d4668b40d 6801507 ZFS read aggregation should not mind the gap (141445-01) 8697:e62d2612c14d 6633095 creating a filesystem with many properties set is slow (141445-01) 8768:dfecfdbb27ed 6775697 oracle crashes when overwriting after hitting quota on zfs (141909-01) 8811:f8deccf701cf 6790687 libzfs mnttab caching ignores external changes (141445-01) 6791101 memory leak from libzfs_mnttab_init (141445-01) 8845:91af0d9c0790 6800942 smb_session_create() incorrectly stores IP addresses (N/A) 6582163 Access Control List (ACL) for shares (141445-01) 6804954 smb_search - shortname field should be space padded following the NULL terminator (N/A) 6800184 Panic at smb_oplock_conflict+0x35() (N/A) 8876:59d2e67b4b65 6803822 Reboot after replacement of system disk in a ZFS mirror drops to grub> prompt (141445-01) 8924:5af812f84759 6789318 coredump when issue zdb -uuuu poolname/ (141445-01) 6790345 zdb -dddd -e poolname coredump (141445-01) 6797109 zdb: 'zdb -dddddd pool_name/fs_name inode' coredump if the file with inode was deleted (141445-01) 6797118 zdb: 'zdb -dddddd poolname inum' coredump if I miss the fs name (141445-01) 6803343 shareiscsi=on failed, iscsitgtd failed request to share (141445-01) 9030:243fd360d81f 6815893 hang mounting a dataset after booting into a new boot environment (141445-01) 9056:826e1858a846 6809691 'zpool create -f' no longer overwrites ufs infomation (141445-01) 9179:d8fbd96b79b3 6790064 zfs needs to determine uid and gid earlier in create process (141445-01) 9214:8d350e5d04aa 6604992 forced unmount + being in .zfs/snapshot/<snap1> = not happy (141909-01) 6810367 assertion failed: dvp->v_flag & VROOT, file: ../../common/fs/gfs.c, line: 426 (141909-01) 9229:e3f8b41e5db4 6807765 ztest_dsl_dataset_promote_busy needs to clean up after ENOSPC (141445-01) 9230:e4561e3eb1ef 6821169 offlining a device results in checksum errors (141445-01) 6821170 ZFS should not increment error stats for unavailable devices (141445-01) 6824006 need to increase issue and interrupt taskqs threads in zfs (141445-01) 9234:bffdc4fc05c4 6792139 recovering from a suspended pool needs some work (141445-01) 6794830 reboot command hangs on a failed zfs pool (141445-01) 9246:67c03c93c071 6824062 System panicked in zfs_mount due to NULL pointer dereference when running btts and svvs tests (141909-01) 9276:a8a7fc849933 6816124 System crash running zpool destroy on broken zpool (141445-03) 9355:09928982c591 6818183 zfs snapshot -r is slow due to set_snap_props() doing txg_wait_synced() for each new snapshot (141445-03) 9391:413d0661ef33 6710376 log device can show incorrect status when other parts of pool are degraded (141445-03) 9396:f41cf682d0d3 (part already merged) 6501037 want user/group quotas on ZFS (141445-03) 6827260 assertion failed in arc_read(): hdr == pbuf->b_hdr (141445-03) 6815592 panic: No such hold X on refcount Y from zfs_znode_move (141445-03) 6759986 zfs list shows temporary %clone when doing online zfs recv (141445-03) 9404:319573cd93f8 6774713 zfs ignores canmount=noauto when sharenfs property != off (141445-03) 9412:4aefd8704ce0 6717022 ZFS DMU needs zero-copy support (141445-03) 9425:e7ffacaec3a8 6799895 spa_add_spares() needs to be protected by config lock (141445-03) 6826466 want to post sysevents on hot spare activation (141445-03) 6826468 spa 'allowfaulted' needs some work (141445-03) 6826469 kernel support for storing vdev FRU information (141445-03) 6826470 skip posting checksum errors from DTL regions of leaf vdevs (141445-03) 6826471 I/O errors after device remove probe can confuse FMA (141445-03) 6826472 spares should enjoy some of the benefits of cache devices (141445-03) 9443:2a96d8478e95 6833711 gang leaders shouldn't have to be logical (141445-03) 9463:d0bd231c7518 6764124 want zdb to be able to checksum metadata blocks only (141445-03) 9465:8372081b8019 6830237 zfs panic in zfs_groupmember() (141445-03) 9466:1fdfd1fed9c4 6833162 phantom log device in zpool status (141445-03) 9469:4f68f041ddcd 6824968 add ZFS userquota support to rquotad (141445-03) 9470:6d827468d7b5 6834217 godfather I/O should reexecute (141445-03) 9480:fcff33da767f 6596237 Stop looking and start ganging (141909-02) 9493:9933d599bc93 6623978 lwb->lwb_buf != NULL, file ../../../uts/common/fs/zfs/zil.c, line 787, function zil_lwb_commit (141445-06) 9512:64cafcbcc337 6801810 Commit of aligned streaming rewrites to ZIL device causes unwanted disk reads (N/A) 9515:d3b739d9d043 6586537 async zio taskqs can block out userland commands (142901-09) 9554:787363635b6a 6836768 zfs_userspace() callback has no way to indicate failure (N/A) 9574:1eb6a6ab2c57 6838062 zfs panics when an error is encountered in space_map_load() (141909-02) 9583:b0696cd037cc 6794136 Panic BAD TRAP: type=e when importing degraded zraid pool. (141909-03) 9630:e25a03f552e0 6776104 "zfs import" deadlock between spa_unload() and spa_async_thread() (141445-06) 9653:a70048a304d1 6664765 Unable to remove files when using fat-zap and quota exceeded on ZFS filesystem (141445-06) 9688:127be1845343 6841321 zfs userspace / zfs get userused@ doesn't work on mounted snapshot (N/A) 6843069 zfs get userused@S-1-... doesn't work (N/A) 9873:8ddc892eca6e 6847229 assertion failed: refcount_count(&tx->tx_space_written) + delta <= tx->tx_space_towrite in dmu_tx.c (141445-06) 9904:d260bd3fd47c 6838344 kernel heap corruption detected on zil while stress testing (141445-06) 9951:a4895b3dd543 6844900 zfs_ioc_userspace_upgrade leaks (N/A) 10040:38b25aeeaf7a 6857012 zfs panics on zpool import (141445-06) 10000:241a51d8720c 6848242 zdb -e no longer works as expected (N/A) 10100:4a6965f6bef8 6856634 snv_117 not booting: zfs_parse_bootfs: error2 (141445-07) 10160:a45b03783d44 6861983 zfs should use new name <-> SID interfaces (N/A) 6862984 userquota commands can hang (141445-06) 10299:80845694147f 6696858 zfs receive of incremental replication stream can dereference NULL pointer and crash (N/A) 10302:a9e3d1987706 6696858 zfs receive of incremental replication stream can dereference NULL pointer and crash (fix lint) (N/A) 10575:2a8816c5173b (partial merge) 6882227 spa_async_remove() shouldn't do a full clear (142901-14) 10800:469478b180d9 6880764 fsync on zfs is broken if writes are greater than 32kb on a hard crash and no log attached (142901-09) 6793430 zdb -ivvvv assertion failure: bp->blk_cksum.zc_word[2] == dmu_objset_id(zilog->zl_os) (N/A) 10801:e0bf032e8673 (partial merge) 6822816 assertion failed: zap_remove_int(ds_next_clones_obj) returns ENOENT (142901-09) 10810:b6b161a6ae4a 6892298 buf->b_hdr->b_state != arc_anon, file: ../../common/fs/zfs/arc.c, line: 2849 (142901-09) 10890:499786962772 6807339 spurious checksum errors when replacing a vdev (142901-13) 11249:6c30f7dfc97b 6906110 bad trap panic in zil_replay_log_record (142901-13) 6906946 zfs replay isn't handling uid/gid correctly (142901-13) 11454:6e69bacc1a5a 6898245 suspended zpool should not cause rest of the zfs/zpool commands to hang (142901-10) 11546:42ea6be8961b (partial merge) 6833999 3-way deadlock in dsl_dataset_hold_ref() and dsl_sync_task_group_sync() (142901-09) Discussed with: pjd Approved by: delphij (mentor) Obtained from: OpenSolaris (multiple Bug IDs) MFC after: 2 months
2010-07-12 23:49:04 +00:00
ZFS notes
---------
When upgrading the boot ZFS pool to a new version, always follow
these two steps:
1.) recompile and reinstall the ZFS boot loader and boot block
(this is part of "make buildworld" and "make installworld")
2.) update the ZFS boot block on your boot drive
The following example updates the ZFS boot block on the first
partition (freebsd-boot) of a GPT partitioned drive ada0:
"gpart bootcode -p /boot/gptzfsboot -i 1 ada0"
Merge ZFS version 15 and almost all OpenSolaris bugfixes referenced in Solaris 10 updates 141445-09 and 142901-14. Detailed information: (OpenSolaris revisions and Bug IDs, Solaris 10 patch numbers) 7844:effed23820ae 6755435 zfs_open() and zfs_close() needs to use ZFS_ENTER/ZFS_VERIFY_ZP (141445-01) 7897:e520d8258820 6748436 inconsistent zpool.cache in boot_archive could panic a zfs root filesystem upon boot-up (141445-01) 7965:b795da521357 6740164 zpool attach can create an illegal root pool (141909-02) 8084:b811cc60d650 6769612 zpool_import() will continue to write to cachefile even if altroot is set (N/A) 8121:7fd09d4ebd9c 6757430 want an option for zdb to disable space map loading and leak tracking (141445-01) 8129:e4f45a0bfbb0 6542860 ASSERT: reason != VDEV_LABEL_REMOVE||vdev_inuse(vd, crtxg, reason, 0) (141445-01) 8188:fd00c0a81e80 6761100 want zdb option to select older uberblocks (141445-01) 8190:6eeea43ced42 6774886 zfs_setattr() won't allow ndmp to restore SUNWattr_rw (141445-01) 8225:59a9961c2aeb 6737463 panic while trying to write out config file if root pool import fails (141445-01) 8227:f7d7be9b1f56 6765294 Refactor replay (141445-01) 8228:51e9ca9ee3a5 6572357 libzfs should do more to avoid mnttab lookups (141909-01) 6572376 zfs_iter_filesystems and zfs_iter_snapshots get objset stats twice (141909-01) 8241:5a60f16123ba 6328632 zpool offline is a bit too conservative (141445-01) 6739487 ASSERT: txg <= spa_final_txg due to scrub/export race (141445-01) 6767129 ASSERT: cvd->vdev_isspare, in spa_vdev_detach() (141445-01) 6747698 checksum failures after offline -t / export / import / scrub (141445-01) 6745863 ZFS writes to disk after it has been offlined (141445-01) 6722540 50% slowdown on scrub/resilver with certain vdev configurations (141445-01) 6759999 resilver logic rewrites ditto blocks on both source and destination (141445-01) 6758107 I/O should never suspend during spa_load() (141445-01) 6776548 codereview(1) runs off the page when faced with multi-line comments (N/A) 6761406 AMD errata 91 workaround doesn't work on 64-bit systems (141445-01) 8242:e46e4b2f0a03 6770866 GRUB/ZFS should require physical path or devid, but not both (141445-01) 8269:03a7e9050cfd 6674216 "zfs share" doesn't work, but "zfs set sharenfs=on" does (141445-01) 6621164 $SRC/cmd/zfs/zfs_main.c seems to have a syntax error in the translation note (141445-01) 6635482 i18n problems in libzfs_dataset.c and zfs_main.c (141445-01) 6595194 "zfs get" VALUE column is as wide as NAME (141445-01) 6722991 vdev_disk.c: error checking for ddi_pathname_to_dev_t() must test for NODEV (141445-01) 6396518 ASSERT strings shouldn't be pre-processed (141445-01) 8274:846b39508aff 6713916 scrub/resilver needlessly decompress data (141445-01) 8343:655db2375fed 6739553 libzfs_status msgid table is out of sync (141445-01) 6784104 libzfs unfairly rejects numerical values greater than 2^63 (141445-01) 6784108 zfs_realloc() should not free original memory on failure (141445-01) 8525:e0e0e525d0f8 6788830 set large value to reservation cause core dump (141445-01) 6791064 want sysevents for ZFS scrub (141445-01) 6791066 need to be able to set cachefile on faulted pools (141445-01) 6791071 zpool_do_import() should not enable datasets on faulted pools (141445-01) 6792134 getting multiple properties on a faulted pool leads to confusion (141445-01) 8547:bcc7b46e5ff7 6792884 Vista clients cannot access .zfs (141445-01) 8632:36ef517870a3 6798384 It can take a village to raise a zio (141445-01) 8636:7e4ce9158df3 6551866 deadlock between zfs_write(), zfs_freesp(), and zfs_putapage() (141909-01) 6504953 zfs_getpage() misunderstands VOP_GETPAGE() interface (141909-01) 6702206 ZFS read/writer lock contention throttles sendfile() benchmark (141445-01) 6780491 Zone on a ZFS filesystem has poor fork/exec performance (141445-01) 6747596 assertion failed: DVA_EQUAL(BP_IDENTITY(&zio->io_bp_orig), BP_IDENTITY(zio->io_bp))); (141445-01) 8692:692d4668b40d 6801507 ZFS read aggregation should not mind the gap (141445-01) 8697:e62d2612c14d 6633095 creating a filesystem with many properties set is slow (141445-01) 8768:dfecfdbb27ed 6775697 oracle crashes when overwriting after hitting quota on zfs (141909-01) 8811:f8deccf701cf 6790687 libzfs mnttab caching ignores external changes (141445-01) 6791101 memory leak from libzfs_mnttab_init (141445-01) 8845:91af0d9c0790 6800942 smb_session_create() incorrectly stores IP addresses (N/A) 6582163 Access Control List (ACL) for shares (141445-01) 6804954 smb_search - shortname field should be space padded following the NULL terminator (N/A) 6800184 Panic at smb_oplock_conflict+0x35() (N/A) 8876:59d2e67b4b65 6803822 Reboot after replacement of system disk in a ZFS mirror drops to grub> prompt (141445-01) 8924:5af812f84759 6789318 coredump when issue zdb -uuuu poolname/ (141445-01) 6790345 zdb -dddd -e poolname coredump (141445-01) 6797109 zdb: 'zdb -dddddd pool_name/fs_name inode' coredump if the file with inode was deleted (141445-01) 6797118 zdb: 'zdb -dddddd poolname inum' coredump if I miss the fs name (141445-01) 6803343 shareiscsi=on failed, iscsitgtd failed request to share (141445-01) 9030:243fd360d81f 6815893 hang mounting a dataset after booting into a new boot environment (141445-01) 9056:826e1858a846 6809691 'zpool create -f' no longer overwrites ufs infomation (141445-01) 9179:d8fbd96b79b3 6790064 zfs needs to determine uid and gid earlier in create process (141445-01) 9214:8d350e5d04aa 6604992 forced unmount + being in .zfs/snapshot/<snap1> = not happy (141909-01) 6810367 assertion failed: dvp->v_flag & VROOT, file: ../../common/fs/gfs.c, line: 426 (141909-01) 9229:e3f8b41e5db4 6807765 ztest_dsl_dataset_promote_busy needs to clean up after ENOSPC (141445-01) 9230:e4561e3eb1ef 6821169 offlining a device results in checksum errors (141445-01) 6821170 ZFS should not increment error stats for unavailable devices (141445-01) 6824006 need to increase issue and interrupt taskqs threads in zfs (141445-01) 9234:bffdc4fc05c4 6792139 recovering from a suspended pool needs some work (141445-01) 6794830 reboot command hangs on a failed zfs pool (141445-01) 9246:67c03c93c071 6824062 System panicked in zfs_mount due to NULL pointer dereference when running btts and svvs tests (141909-01) 9276:a8a7fc849933 6816124 System crash running zpool destroy on broken zpool (141445-03) 9355:09928982c591 6818183 zfs snapshot -r is slow due to set_snap_props() doing txg_wait_synced() for each new snapshot (141445-03) 9391:413d0661ef33 6710376 log device can show incorrect status when other parts of pool are degraded (141445-03) 9396:f41cf682d0d3 (part already merged) 6501037 want user/group quotas on ZFS (141445-03) 6827260 assertion failed in arc_read(): hdr == pbuf->b_hdr (141445-03) 6815592 panic: No such hold X on refcount Y from zfs_znode_move (141445-03) 6759986 zfs list shows temporary %clone when doing online zfs recv (141445-03) 9404:319573cd93f8 6774713 zfs ignores canmount=noauto when sharenfs property != off (141445-03) 9412:4aefd8704ce0 6717022 ZFS DMU needs zero-copy support (141445-03) 9425:e7ffacaec3a8 6799895 spa_add_spares() needs to be protected by config lock (141445-03) 6826466 want to post sysevents on hot spare activation (141445-03) 6826468 spa 'allowfaulted' needs some work (141445-03) 6826469 kernel support for storing vdev FRU information (141445-03) 6826470 skip posting checksum errors from DTL regions of leaf vdevs (141445-03) 6826471 I/O errors after device remove probe can confuse FMA (141445-03) 6826472 spares should enjoy some of the benefits of cache devices (141445-03) 9443:2a96d8478e95 6833711 gang leaders shouldn't have to be logical (141445-03) 9463:d0bd231c7518 6764124 want zdb to be able to checksum metadata blocks only (141445-03) 9465:8372081b8019 6830237 zfs panic in zfs_groupmember() (141445-03) 9466:1fdfd1fed9c4 6833162 phantom log device in zpool status (141445-03) 9469:4f68f041ddcd 6824968 add ZFS userquota support to rquotad (141445-03) 9470:6d827468d7b5 6834217 godfather I/O should reexecute (141445-03) 9480:fcff33da767f 6596237 Stop looking and start ganging (141909-02) 9493:9933d599bc93 6623978 lwb->lwb_buf != NULL, file ../../../uts/common/fs/zfs/zil.c, line 787, function zil_lwb_commit (141445-06) 9512:64cafcbcc337 6801810 Commit of aligned streaming rewrites to ZIL device causes unwanted disk reads (N/A) 9515:d3b739d9d043 6586537 async zio taskqs can block out userland commands (142901-09) 9554:787363635b6a 6836768 zfs_userspace() callback has no way to indicate failure (N/A) 9574:1eb6a6ab2c57 6838062 zfs panics when an error is encountered in space_map_load() (141909-02) 9583:b0696cd037cc 6794136 Panic BAD TRAP: type=e when importing degraded zraid pool. (141909-03) 9630:e25a03f552e0 6776104 "zfs import" deadlock between spa_unload() and spa_async_thread() (141445-06) 9653:a70048a304d1 6664765 Unable to remove files when using fat-zap and quota exceeded on ZFS filesystem (141445-06) 9688:127be1845343 6841321 zfs userspace / zfs get userused@ doesn't work on mounted snapshot (N/A) 6843069 zfs get userused@S-1-... doesn't work (N/A) 9873:8ddc892eca6e 6847229 assertion failed: refcount_count(&tx->tx_space_written) + delta <= tx->tx_space_towrite in dmu_tx.c (141445-06) 9904:d260bd3fd47c 6838344 kernel heap corruption detected on zil while stress testing (141445-06) 9951:a4895b3dd543 6844900 zfs_ioc_userspace_upgrade leaks (N/A) 10040:38b25aeeaf7a 6857012 zfs panics on zpool import (141445-06) 10000:241a51d8720c 6848242 zdb -e no longer works as expected (N/A) 10100:4a6965f6bef8 6856634 snv_117 not booting: zfs_parse_bootfs: error2 (141445-07) 10160:a45b03783d44 6861983 zfs should use new name <-> SID interfaces (N/A) 6862984 userquota commands can hang (141445-06) 10299:80845694147f 6696858 zfs receive of incremental replication stream can dereference NULL pointer and crash (N/A) 10302:a9e3d1987706 6696858 zfs receive of incremental replication stream can dereference NULL pointer and crash (fix lint) (N/A) 10575:2a8816c5173b (partial merge) 6882227 spa_async_remove() shouldn't do a full clear (142901-14) 10800:469478b180d9 6880764 fsync on zfs is broken if writes are greater than 32kb on a hard crash and no log attached (142901-09) 6793430 zdb -ivvvv assertion failure: bp->blk_cksum.zc_word[2] == dmu_objset_id(zilog->zl_os) (N/A) 10801:e0bf032e8673 (partial merge) 6822816 assertion failed: zap_remove_int(ds_next_clones_obj) returns ENOENT (142901-09) 10810:b6b161a6ae4a 6892298 buf->b_hdr->b_state != arc_anon, file: ../../common/fs/zfs/arc.c, line: 2849 (142901-09) 10890:499786962772 6807339 spurious checksum errors when replacing a vdev (142901-13) 11249:6c30f7dfc97b 6906110 bad trap panic in zil_replay_log_record (142901-13) 6906946 zfs replay isn't handling uid/gid correctly (142901-13) 11454:6e69bacc1a5a 6898245 suspended zpool should not cause rest of the zfs/zpool commands to hang (142901-10) 11546:42ea6be8961b (partial merge) 6833999 3-way deadlock in dsl_dataset_hold_ref() and dsl_sync_task_group_sync() (142901-09) Discussed with: pjd Approved by: delphij (mentor) Obtained from: OpenSolaris (multiple Bug IDs) MFC after: 2 months
2010-07-12 23:49:04 +00:00
Non-boot pools do not need these updates.
To build a kernel
-----------------
If you are updating from a prior version of FreeBSD (even one just
a few days old), you should follow this procedure. It is the most
failsafe as it uses a /usr/obj tree with a fresh mini-buildworld,
make kernel-toolchain
make -DALWAYS_CHECK_MAKE buildkernel KERNCONF=YOUR_KERNEL_HERE
make -DALWAYS_CHECK_MAKE installkernel KERNCONF=YOUR_KERNEL_HERE
To test a kernel once
---------------------
If you just want to boot a kernel once (because you are not sure
if it works, or if you want to boot a known bad kernel to provide
debugging information) run
make installkernel KERNCONF=YOUR_KERNEL_HERE KODIR=/boot/testkernel
nextboot -k testkernel
To rebuild everything and install it on the current system.
-----------------------------------------------------------
# Note: sometimes if you are running current you gotta do more than
# is listed here if you are upgrading from a really old current.
<make sure you have good level 0 dumps>
make buildworld
make buildkernel KERNCONF=YOUR_KERNEL_HERE
make installkernel KERNCONF=YOUR_KERNEL_HERE
[1]
<reboot in single user> [3]
mergemaster -Fp [5]
make installworld
mergemaster -Fi [4]
make delete-old [6]
<reboot>
To cross-install current onto a separate partition
--------------------------------------------------
# In this approach we use a separate partition to hold
# current's root, 'usr', and 'var' directories. A partition
# holding "/", "/usr" and "/var" should be about 2GB in
# size.
<make sure you have good level 0 dumps>
<boot into -stable>
make buildworld
make buildkernel KERNCONF=YOUR_KERNEL_HERE
<maybe newfs current's root partition>
<mount current's root partition on directory ${CURRENT_ROOT}>
make installworld DESTDIR=${CURRENT_ROOT} -DDB_FROM_SRC
make distribution DESTDIR=${CURRENT_ROOT} # if newfs'd
make installkernel KERNCONF=YOUR_KERNEL_HERE DESTDIR=${CURRENT_ROOT}
cp /etc/fstab ${CURRENT_ROOT}/etc/fstab # if newfs'd
<edit ${CURRENT_ROOT}/etc/fstab to mount "/" from the correct partition>
<reboot into current>
<do a "native" rebuild/install as described in the previous section>
<maybe install compatibility libraries from ports/misc/compat*>
<reboot>
To upgrade in-place from stable to current
----------------------------------------------
<make sure you have good level 0 dumps>
make buildworld [9]
make buildkernel KERNCONF=YOUR_KERNEL_HERE [8]
make installkernel KERNCONF=YOUR_KERNEL_HERE
[1]
<reboot in single user> [3]
mergemaster -Fp [5]
make installworld
mergemaster -Fi [4]
make delete-old [6]
<reboot>
2000-08-28 03:54:51 +00:00
Make sure that you've read the UPDATING file to understand the
tweaks to various things you need. At this point in the life
cycle of current, things change often and you are on your own
to cope. The defaults can also change, so please read ALL of
the UPDATING entries.
Also, if you are tracking -current, you must be subscribed to
freebsd-current@freebsd.org. Make sure that before you update
your sources that you have read and understood all the recent
messages there. If in doubt, please track -stable which has
much fewer pitfalls.
[1] If you have third party modules, such as vmware, you
should disable them at this point so they don't crash your
system on reboot.
[3] From the bootblocks, boot -s, and then do
fsck -p
mount -u /
mount -a
cd src
adjkerntz -i # if CMOS is wall time
2002-04-25 02:14:48 +00:00
Also, when doing a major release upgrade, it is required that
you boot into single user mode to do the installworld.
[4] Note: This step is non-optional. Failure to do this step
can result in a significant reduction in the functionality of the
system. Attempting to do it by hand is not recommended and those
that pursue this avenue should read this file carefully, as well
as the archives of freebsd-current and freebsd-hackers mailing lists
for potential gotchas. The -U option is also useful to consider.
See mergemaster(8) for more information.
2018-01-12 17:33:20 +00:00
[5] Usually this step is a no-op. However, from time to time
you may need to do this if you get unknown user in the following
step. It never hurts to do it all the time. You may need to
install a new mergemaster (cd src/usr.sbin/mergemaster && make
install) after the buildworld before this step if you last updated
from current before 20130425 or from -stable before 20130430.
[6] This only deletes old files and directories. Old libraries
can be deleted by "make delete-old-libs", but you have to make
sure that no program is using those libraries anymore.
[8] In order to have a kernel that can run the 4.x binaries needed to
do an installworld, you must include the COMPAT_FREEBSD4 option in
your kernel. Failure to do so may leave you with a system that is
hard to boot to recover. A similar kernel option COMPAT_FREEBSD5 is
required to run the 5.x binaries on more recent kernels. And so on
for COMPAT_FREEBSD6 and COMPAT_FREEBSD7.
Make sure that you merge any new devices from GENERIC since the
last time you updated your kernel config file.
[9] If CPUTYPE is defined in your /etc/make.conf, make sure to use the
"?=" instead of the "=" assignment operator, so that buildworld can
override the CPUTYPE if it needs to.
MAKEOBJDIRPREFIX must be defined in an environment variable, and
not on the command line, or in /etc/make.conf. buildworld will
warn if it is improperly defined.
FORMAT:
This file contains a list, in reverse chronological order, of major
breakages in tracking -current. It is not guaranteed to be a complete
list of such breakages, and only contains entries since September 23, 2011.
If you need to see UPDATING entries from before that date, you will need
to fetch an UPDATING file from an older FreeBSD release.
Copyright information:
Copyright 1998-2009 M. Warner Losh. All Rights Reserved.
Redistribution, publication, translation and use, with or without
modification, in full or in part, in any form or format of this
document are permitted without further permission from the author.
THIS DOCUMENT IS PROVIDED BY WARNER LOSH ``AS IS'' AND ANY EXPRESS OR
IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL WARNER LOSH BE LIABLE FOR ANY DIRECT,
INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.
Contact Warner Losh if you have any questions about your use of
this document.
1999-08-28 01:35:59 +00:00
$FreeBSD$