freebsd-skq/UPDATING

2117 lines
86 KiB
Plaintext
Raw Normal View History

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:
http://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 to bootstrap to the tip of
head, and then rebuild without this option. The bootstrap process from
older version of current is a bit fragile.
NOTE TO PEOPLE WHO THINK THAT FreeBSD 11.x IS SLOW:
FreeBSD 11.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".)
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 rc.d/jail script has been updated to support jail(8)
configuration file. The "jail_<jname>_*" rc.conf(5) variables
for per-jail configuration are automatically converted to
/var/run/jail.<jname>.conf before the jail(8) utility is invoked.
This is transparently backward compatible. See below about some
incompatibilities and rc.conf(5) manual page for more details.
These variables are now deprecated in favor of jail(8) configuration
file. One can use "rc.d/jail config <jname>" command to generate
a jail(8) configuration file in /var/run/jail.<jname>.conf without
running the jail(8) utility. The default pathname of the
configuration file is /etc/jail.conf and can be specified by
using $jail_conf or $jail_<jname>_conf variables.
Please note that jail_devfs_ruleset accepts an integer at
this moment. Please consider to rewrite the ruleset name
with an integer.
20130930:
BIND has been removed from the base system. If all you need
is a local resolver, simply enable and start the local_unbound
service instead. Otherwise, several versions of BIND are
available in the ports tree. The dns/bind99 port is one example.
With this change, nslookup(1) and dig(1) are no longer in the base
system. Users should instead use host(1) and drill(1) which are
in the base system. Alternatively, nslookup and dig can
be obtained by installing the dns/bind-tools port.
20130916:
With the addition of unbound(8), a new unbound user is now
required during installworld. "mergemaster -p" can be used to
add the user prior to installworld, as documented in the handbook.
20130911:
OpenSSH is now built with DNSSEC support, and will by default
silently trust signed SSHFP records. This can be controlled with
the VerifyHostKeyDNS client configuration setting. DNSSEC support
can be disabled entirely with the WITHOUT_LDNS option in src.conf.
20130906:
The GNU Compiler Collection and C++ standard library (libstdc++)
are no longer built by default on platforms where clang is the system
compiler. You can enable them with the WITH_GCC and WITH_GNUCXX
options in src.conf.
20130905:
The PROCDESC kernel option is now part of the GENERIC kernel
configuration and is required for the rwhod(8) to work.
If you are using custom kernel configuration, you should include
'options PROCDESC'.
20130905:
The API and ABI related to the Capsicum framework was modified
in backward incompatible way. The userland libraries and programs
have to be recompiled to work with the new kernel. This includes the
following libraries and programs, but the whole buildworld is
advised: libc, libprocstat, dhclient, tcpdump, hastd, hastctl,
kdump, procstat, rwho, rwhod, uniq.
20130903:
AES-NI intrinsic support has been added to gcc. The AES-NI module
has been updated to use this support. A new gcc is required to build
the aesni module on both i386 and amd64.
20130821:
The PADLOCK_RNG and RDRAND_RNG kernel options are now devices.
Thus "device padlock_rng" and "device rdrand_rng" should be
used instead of "options PADLOCK_RNG" & "options RDRAND_RNG".
20130813:
WITH_ICONV has been split into two feature sets. WITH_ICONV now
enables just the iconv* functionality and is now on by default.
WITH_LIBICONV_COMPAT enables the libiconv api and link time
compatability. Set WITHOUT_ICONV to build the old way.
If you have been using WITH_ICONV before, you will very likely
need to turn on WITH_LIBICONV_COMPAT.
20130806:
INVARIANTS option now enables DEBUG for code with OpenSolaris and
Illumos origin, including ZFS. If you have INVARIANTS in your
kernel configuration, then there is no need to set DEBUG or ZFS_DEBUG
explicitly.
DEBUG used to enable witness(9) tracking of OpenSolaris (mostly ZFS)
locks if WITNESS option was set. Because that generated a lot of
witness(9) reports and all of them were believed to be false
positives, this is no longer done. New option OPENSOLARIS_WITNESS
can be used to achieve the previous behavior.
20130806:
Timer values in IPv6 data structures now use time_uptime instead
of time_second. Although this is not a user-visible functional
change, userland utilities which directly use them---ndp(8),
rtadvd(8), and rtsold(8) in the base system---need to be updated
to r253970 or later.
20130802:
find -delete can now delete the pathnames given as arguments,
instead of only files found below them or if the pathname did
not contain any slashes. Formerly, the following error message
would result:
find: -delete: <path>: relative path potentially not safe
Deleting the pathnames given as arguments can be prevented
without error messages using -mindepth 1 or by changing
directory and passing "." as argument to find. This works in the
old as well as the new version of find.
20130726:
Behavior of devfs rules path matching has been changed.
Pattern is now always matched against fully qualified devfs
path and slash characters must be explicitly matched by
slashes in pattern (FNM_PATHNAME). Rulesets involving devfs
subdirectories must be reviewed.
20130716:
The default ARM ABI has changed to the ARM EABI. The old ABI is
incompatible with the ARM EABI and all programs and modules will
need to be rebuilt to work with a new kernel.
To keep using the old ABI ensure the WITHOUT_ARM_EABI knob is set.
NOTE: Support for the old ABI will be removed in the future and
users are advised to upgrade.
20130709:
pkg_install has been disconnected from the build if you really need it
you should add WITH_PKGTOOLS in your src.conf(5).
20130709:
Most of network statistics structures were changed to be able
keep 64-bits counters. Thus all tools, that work with networking
2013-07-14 20:22:16 +00:00
statistics, must be rebuilt (netstat(1), bsnmpd(1), etc.)
20130629:
Fix targets that run multiple make's to use && rather than ;
so that subsequent steps depend on success of previous.
NOTE: if building 'universe' with -j* on stable/8 or stable/9
it would be better to start the build using bmake, to avoid
overloading the machine.
20130618:
Fix a bug that allowed a tracing process (e.g. gdb) to write
to a memory-mapped file in the traced process's address space
even if neither the traced process nor the tracing process had
write access to that file.
20130615:
CVS has been removed from the base system. An exact copy
of the code is available from the devel/cvs port.
20130613:
Some people report the following error after the switch to bmake:
make: illegal option -- J
usage: make [-BPSXeiknpqrstv] [-C directory] [-D variable]
...
*** [buildworld] Error code 2
this likely due to an old instance of make in
${MAKEPATH} (${MAKEOBJDIRPREFIX}${.CURDIR}/make.${MACHINE})
which src/Makefile will use that blindly, if it exists, so if
you see the above error:
rm -rf `make -V MAKEPATH`
should resolve it.
20130516:
Use bmake by default.
Whereas before one could choose to build with bmake via
-DWITH_BMAKE one must now use -DWITHOUT_BMAKE to use the old
make. The goal is to remove these knobs for 10-RELEASE.
It is worth noting that bmake (like gmake) treats the command
line as the unit of failure, rather than statements within the
command line. Thus '(cd some/where && dosomething)' is safer
than 'cd some/where; dosomething'. The '()' allows consistent
behavior in parallel build.
20130429:
Fix a bug that allows NFS clients to issue READDIR on files.
20130426:
The WITHOUT_IDEA option has been removed because
the IDEA patent expired.
20130426:
The sysctl which controls TRIM support under ZFS has been renamed
from vfs.zfs.trim_disable -> vfs.zfs.trim.enabled and has been
enabled by default.
20130425:
The mergemaster command now uses the default MAKEOBJDIRPREFIX
rather than creating it's own in the temporary directory in
order allow access to bootstrapped versions of tools such as
install and mtree. When upgrading from version of FreeBSD where
the install command does not support -l, you will need to
install a new mergemaster command if mergemaster -p is required.
This can be accomplished with the command (cd src/usr.sbin/mergemaster
&& make install).
2013-04-06 13:47:44 +00:00
20130404:
Legacy ATA stack, disabled and replaced by new CAM-based one since
FreeBSD 9.0, completely removed from the sources. Kernel modules
atadisk and atapi*, user-level tools atacontrol and burncd are
removed. Kernel option `options ATA_CAM` is now permanently enabled
and removed.
20130319:
SOCK_CLOEXEC and SOCK_NONBLOCK flags have been added to socket(2)
and socketpair(2). Software, in particular Kerberos, may
automatically detect and use these during building. The resulting
binaries will not work on older kernels.
20130308:
CTL_DISABLE has also been added to the sparc64 GENERIC (for further
information, see the respective 20130304 entry).
20130304:
Recent commits to callout(9) changed the size of struct callout,
so the KBI is probably heavily disturbed. Also, some functions
in callout(9)/sleep(9)/sleepqueue(9)/condvar(9) KPIs were replaced
by macros. Every kernel module using it won't load, so rebuild
is requested.
The ctl device has been re-enabled in GENERIC for i386 and amd64,
but does not initialize by default (because of the new CTL_DISABLE
option) to save memory. To re-enable it, remove the CTL_DISABLE
option from the kernel config file or set kern.cam.ctl.disable=0
in /boot/loader.conf.
20130301:
The ctl device has been disabled in GENERIC for i386 and amd64.
This was done due to the extra memory being allocated at system
initialisation time by the ctl driver which was only used if
a CAM target device was created. This makes a FreeBSD system
unusable on 128MB or less of RAM.
20130208:
A new compression method (lz4) has been merged to -HEAD. Please
refer to zpool-features(7) for more information.
Please refer to the "ZFS notes" section of this file for information
on upgrading boot ZFS pools.
20130129:
A BSD-licensed patch(1) variant has been added and is installed
as bsdpatch, being the GNU version the default patch.
To inverse the logic and use the BSD-licensed one as default,
while having the GNU version installed as gnupatch, rebuild
2013-01-30 10:23:38 +00:00
and install world with the WITH_BSD_PATCH knob set.
20130121:
Due to the use of the new -l option to install(1) during build
and install, you must take care not to directly set the INSTALL
make variable in your /etc/make.conf, /etc/src.conf, or on the
command line. If you wish to use the -C flag for all installs
you may be able to add INSTALL+=-C to /etc/make.conf or
/etc/src.conf.
20130118:
The install(1) option -M has changed meaning and now takes an
argument that is a file or path to append logs to. In the
unlikely event that -M was the last option on the command line
and the command line contained at least two files and a target
directory the first file will have logs appended to it. The -M
option served little practical purpose in the last decade so its
use is expected to be extremely rare.
20121223:
After switching to Clang as the default compiler some users of ZFS
on i386 systems started to experience stack overflow kernel panics.
Please consider using 'options KSTACK_PAGES=4' in such configurations.
20121222:
GEOM_LABEL now mangles label names read from file system metadata.
Mangling affect labels containing spaces, non-printable characters,
'%' or '"'. Device names in /etc/fstab and other places may need to
be updated.
20121217:
By default, only the 10 most recent kernel dumps will be saved. To
restore the previous behaviour (no limit on the number of kernel dumps
stored in the dump directory) add the following line to /etc/rc.conf:
savecore_flags=""
20121201:
With the addition of auditdistd(8), a new auditdistd user is now
required during installworld. "mergemaster -p" can be used to
add the user prior to installworld, as documented in the handbook.
20121117:
The sin6_scope_id member variable in struct sockaddr_in6 is now
filled by the kernel before passing the structure to the userland via
sysctl or routing socket. This means the KAME-specific embedded scope
id in sin6_addr.s6_addr[2] is always cleared in userland application.
This behavior can be controlled by net.inet6.ip6.deembed_scopeid.
__FreeBSD_version is bumped to 1000025.
20121105:
On i386 and amd64 systems WITH_CLANG_IS_CC is now the default.
This means that the world and kernel will be compiled with clang
and that clang will be installed as /usr/bin/cc, /usr/bin/c++,
and /usr/bin/cpp. To disable this behavior and revert to building
with gcc, compile with WITHOUT_CLANG_IS_CC. Really old versions
of current may need to bootstrap WITHOUT_CLANG first if the clang
build fails (its compatibility window doesn't extend to the 9 stable
branch point).
20121102:
The IPFIREWALL_FORWARD kernel option has been removed. Its
functionality now turned on by default.
20121023:
The ZERO_COPY_SOCKET kernel option has been removed and
split into SOCKET_SEND_COW and SOCKET_RECV_PFLIP.
NB: SOCKET_SEND_COW uses the VM page based copy-on-write
mechanism which is not safe and may result in kernel crashes.
NB: The SOCKET_RECV_PFLIP mechanism is useless as no current
driver supports disposeable external page sized mbuf storage.
Proper replacements for both zero-copy mechanisms are under
consideration and will eventually lead to complete removal
of the two kernel options.
20121023:
The IPv4 network stack has been converted to network byte
order. The following modules need to be recompiled together
with kernel: carp(4), divert(4), gif(4), siftr(4), gre(4),
pf(4), ipfw(4), ng_ipfw(4), stf(4).
20121022:
Support for non-MPSAFE filesystems was removed from VFS. The
VFS_VERSION was bumped, all filesystem modules shall be
recompiled.
20121018:
All the non-MPSAFE filesystems have been disconnected from
the build. The full list includes: codafs, hpfs, ntfs, nwfs,
portalfs, smbfs, xfs.
20121016:
The interface cloning API and ABI has changed. The following
modules need to be recompiled together with kernel:
ipfw(4), pfsync(4), pflog(4), usb(4), wlan(4), stf(4),
vlan(4), disc(4), edsc(4), if_bridge(4), gif(4), tap(4),
faith(4), epair(4), enc(4), tun(4), if_lagg(4), gre(4).
20121015:
The sdhci driver was split in two parts: sdhci (generic SD Host
Controller logic) and sdhci_pci (actual hardware driver).
No kernel config modifications are required, but if you
load sdhc as a module you must switch to sdhci_pci instead.
20121014:
Import the FUSE kernel and userland support into base system.
20121013:
The GNU sort(1) program has been removed since the BSD-licensed
sort(1) has been the default for quite some time and no serious
problems have been reported. The corresponding WITH_GNU_SORT
2012-10-13 19:37:58 +00:00
knob has also gone.
20121006:
The pfil(9) API/ABI for AF_INET family has been changed. Packet
filtering modules: pf(4), ipfw(4), ipfilter(4) need to be recompiled
with new kernel.
20121001:
The net80211(4) ABI has been changed to allow for improved driver
PS-POLL and power-save support. All wireless drivers need to be
recompiled to work with the new kernel.
20120913:
The random(4) support for the VIA hardware random number
generator (`PADLOCK') is no longer enabled unconditionally.
Add the padlock_rng device in the custom kernel config if
needed. The GENERIC kernels on i386 and amd64 do include the
device, so the change only affects the custom kernel
configurations.
Merge the projects/pf/head branch, that was worked on for last six months, into head. The most significant achievements in the new code: o Fine grained locking, thus much better performance. o Fixes to many problems in pf, that were specific to FreeBSD port. New code doesn't have that many ifdefs and much less OpenBSDisms, thus is more attractive to our developers. Those interested in details, can browse through SVN log of the projects/pf/head branch. And for reference, here is exact list of revisions merged: r232043, r232044, r232062, r232148, r232149, r232150, r232298, r232330, r232332, r232340, r232386, r232390, r232391, r232605, r232655, r232656, r232661, r232662, r232663, r232664, r232673, r232691, r233309, r233782, r233829, r233830, r233834, r233835, r233836, r233865, r233866, r233868, r233873, r234056, r234096, r234100, r234108, r234175, r234187, r234223, r234271, r234272, r234282, r234307, r234309, r234382, r234384, r234456, r234486, r234606, r234640, r234641, r234642, r234644, r234651, r235505, r235506, r235535, r235605, r235606, r235826, r235991, r235993, r236168, r236173, r236179, r236180, r236181, r236186, r236223, r236227, r236230, r236252, r236254, r236298, r236299, r236300, r236301, r236397, r236398, r236399, r236499, r236512, r236513, r236525, r236526, r236545, r236548, r236553, r236554, r236556, r236557, r236561, r236570, r236630, r236672, r236673, r236679, r236706, r236710, r236718, r237154, r237155, r237169, r237314, r237363, r237364, r237368, r237369, r237376, r237440, r237442, r237751, r237783, r237784, r237785, r237788, r237791, r238421, r238522, r238523, r238524, r238525, r239173, r239186, r239644, r239652, r239661, r239773, r240125, r240130, r240131, r240136, r240186, r240196, r240212. I'd like to thank people who participated in early testing: Tested by: Florian Smeets <flo freebsd.org> Tested by: Chekaluk Vitaly <artemrts ukr.net> Tested by: Ben Wilber <ben desync.com> Tested by: Ian FREISLICH <ianf cloudseed.co.za>
2012-09-08 06:41:54 +00:00
20120908:
The pf(4) packet filter ABI has been changed. pfctl(8) and
snmp_pf module need to be recompiled to work with new kernel.
20120828:
A new ZFS feature flag "com.delphix:empty_bpobj" has been merged
to -HEAD. Pools that have empty_bpobj in active state can not be
imported read-write with ZFS implementations that do not support
this feature. For more information read the zpool-features(5)
manual page.
20120727:
The sparc64 ZFS loader has been changed to no longer try to auto-
detect ZFS providers based on diskN aliases but now requires these
to be explicitly listed in the OFW boot-device environment variable.
20120712:
The OpenSSL has been upgraded to 1.0.1c. Any binaries requiring
libcrypto.so.6 or libssl.so.6 must be recompiled. Also, there are
configuration changes. Make sure to merge /etc/ssl/openssl.cnf.
20120712:
The following sysctls and tunables have been renamed for consistency
with other variables:
kern.cam.da.da_send_ordered -> kern.cam.da.send_ordered
kern.cam.ada.ada_send_ordered -> kern.cam.ada.send_ordered
2012-06-28 08:22:00 +00:00
20120628:
The sort utility has been replaced with BSD sort. For now, GNU sort
is also available as "gnusort" or the default can be set back to
GNU sort by setting WITH_GNU_SORT. In this case, BSD sort will be
installed as "bsdsort".
20120611:
A new version of ZFS (pool version 5000) has been merged to -HEAD.
Starting with this version the old system of ZFS pool versioning
is superseded by "feature flags". This concept enables forward
compatibility against certain future changes in functionality of ZFS
pools. The first read-only compatible "feature flag" for ZFS pools
is named "com.delphix:async_destroy". For more information
read the new zpool-features(5) manual page.
Please refer to the "ZFS notes" section of this file for information
on upgrading boot ZFS pools.
20120417:
The malloc(3) implementation embedded in libc now uses sources imported
as contrib/jemalloc. The most disruptive API change is to
/etc/malloc.conf. If your system has an old-style /etc/malloc.conf,
delete it prior to installworld, and optionally re-create it using the
new format after rebooting. See malloc.conf(5) for details
(specifically the TUNING section and the "opt.*" entries in the MALLCTL
NAMESPACE section).
20120328:
Big-endian MIPS TARGET_ARCH values no longer end in "eb". mips64eb
is now spelled mips64. mipsn32eb is now spelled mipsn32. mipseb is
now spelled mips. This is to aid compatibility with third-party
software that expects this naming scheme in uname(3). Little-endian
2013-05-01 07:13:36 +00:00
settings are unchanged. If you are updating a big-endian mips64 machine
from before this change, you may need to set MACHINE_ARCH=mips64 in
your environment before the new build system will recognize your machine.
20120306:
Disable by default the option VFS_ALLOW_NONMPSAFE for all supported
platforms.
20120229:
Now unix domain sockets behave "as expected" on nullfs(5). Previously
nullfs(5) did not pass through all behaviours to the underlying layer,
as a result if we bound to a socket on the lower layer we could connect
only to the lower path; if we bound to the upper layer we could connect
only to the upper path. The new behavior is one can connect to both the
lower and the upper paths regardless what layer path one binds to.
20120211:
The getifaddrs upgrade path broken with 20111215 has been restored.
If you have upgraded in between 20111215 and 20120209 you need to
recompile libc again with your kernel. You still need to recompile
world to be able to configure CARP but this restriction already
comes from 20111215.
20120114:
The set_rcvar() function has been removed from /etc/rc.subr. All
base and ports rc.d scripts have been updated, so if you have a
port installed with a script in /usr/local/etc/rc.d you can either
hand-edit the rcvar= line, or reinstall the port.
An easy way to handle the mass-update of /etc/rc.d:
rm /etc/rc.d/* && mergemaster -i
20120109:
panic(9) now stops other CPUs in the SMP systems, disables interrupts
on the current CPU and prevents other threads from running.
This behavior can be reverted using the kern.stop_scheduler_on_panic
tunable/sysctl.
The new behavior can be incompatible with kern.sync_on_panic.
20111215:
The carp(4) facility has been changed significantly. Configuration
of the CARP protocol via ifconfig(8) has changed, as well as format
of CARP events submitted to devd(8) has changed. See manual pages
for more information. The arpbalance feature of carp(4) is currently
not supported anymore.
Size of struct in_aliasreq, struct in6_aliasreq has changed. User
utilities using SIOCAIFADDR, SIOCAIFADDR_IN6, e.g. ifconfig(8),
need to be recompiled.
2011-11-30 20:08:30 +00:00
20111122:
The acpi_wmi(4) status device /dev/wmistat has been renamed to
/dev/wmistat0.
20111108:
The option VFS_ALLOW_NONMPSAFE option has been added in order to
explicitely support non-MPSAFE filesystems.
It is on by default for all supported platform at this present
time.
20111101:
The broken amd(4) driver has been replaced with esp(4) in the amd64,
i386 and pc98 GENERIC kernel configuration files.
20110930:
sysinstall has been removed
20110923:
The stable/9 branch created in subversion. This corresponds to the
RELENG_9 branch in CVS.
20110913:
This commit modifies vfs_register() so that it uses a hash
calculation to set vfc_typenum, which is enabled by default.
The first time a system is booted after this change, the
vfc_typenum values will change for all file systems. The
main effect of this is a change to the NFS server file handles
for file systems that use vfc_typenum in their fsid, such as ZFS.
It will, however, prevent vfc_typenum from changing when file
systems are loaded in a different order for subsequent reboots.
To disable this, you can set vfs.typenumhash=0 in /boot/loader.conf
until you are ready to remount all NFS clients after a reboot.
20110828:
Bump the shared library version numbers for libraries that
do not use symbol versioning, have changed the ABI compared
to stable/8 and which shared library version was not bumped.
Done as part of 9.0-RELEASE cycle.
20110815:
During the merge of Capsicum features, the fget(9) KPI was modified.
This may require the rebuilding of out-of-tree device drivers --
issues have been reported specifically with the nVidia device driver.
__FreeBSD_version is bumped to 900041.
Also, there is a period between 20110811 and 20110814 where the
special devices /dev/{stdin,stdout,stderr} did not work correctly.
Building world from a kernel during that window may not work.
20110628:
The packet filter (pf) code has been updated to OpenBSD 4.5.
You need to update userland tools to be in sync with kernel.
This update breaks backward compatibility with earlier pfsync(4)
versions. Care must be taken when updating redundant firewall setups.
20110608:
The following sysctls and tunables are retired on x86 platforms:
machdep.hlt_cpus
machdep.hlt_logical_cpus
The following sysctl is retired:
machdep.hyperthreading_allowed
The sysctls were supposed to provide a way to dynamically offline and
online selected CPUs on x86 platforms, but the implementation has not
been reliable especially with SCHED_ULE scheduler.
machdep.hyperthreading_allowed tunable is still available to ignore
hyperthreading CPUs at OS level.
Individual CPUs can be disabled using hint.lapic.X.disabled tunable,
where X is an APIC ID of a CPU. Be advised, though, that disabling
CPUs in non-uniform fashion will result in non-uniform topology and
may lead to sub-optimal system performance with SCHED_ULE, which is
a default scheduler.
2011-06-07 08:07:41 +00:00
20110607:
cpumask_t type is retired and cpuset_t is used in order to describe
a mask of CPUs.
20110531:
Changes to ifconfig(8) for dynamic address family detection mandate
that you are running a kernel of 20110525 or later. Make sure to
follow the update procedure to boot a new kernel before installing
world.
20110513:
Support for sun4v architecture is officially dropped
20110503:
Several KPI breaking changes have been committed to the mii(4) layer,
the PHY drivers and consequently some Ethernet drivers using mii(4).
This means that miibus.ko and the modules of the affected Ethernet
drivers need to be recompiled.
Note to kernel developers: Given that the OUI bit reversion problem
was fixed as part of these changes all mii(4) commits related to OUIs,
i.e. to sys/dev/mii/miidevs, PHY driver probing and vendor specific
handling, no longer can be merged verbatim to stable/8 and previous
branches.
20110430:
Users of the Atheros AR71xx SoC code now need to add 'device ar71xx_pci'
into their kernel configurations along with 'device pci'.
20110427:
The default NFS client is now the new NFS client, so fstype "newnfs"
is now "nfs" and the regular/old NFS client is now fstype "oldnfs".
Although mounts via fstype "nfs" will usually work without userland
changes, it is recommended that the mount(8) and mount_nfs(8)
commands be rebuilt from sources and that a link to mount_nfs called
mount_oldnfs be created. The new client is compiled into the
kernel with "options NFSCL" and this is needed for diskless root
file systems. The GENERIC kernel configs have been changed to use
NFSCL and NFSD (the new server) instead of NFSCLIENT and NFSSERVER.
To use the regular/old client, you can "mount -t oldnfs ...". For
a diskless root file system, you must also include a line like:
vfs.root.mountfrom="oldnfs:"
in the boot/loader.conf on the root fs on the NFS server to make
a diskless root fs use the old client.
20110424:
The GENERIC kernels for all architectures now default to the new
CAM-based ATA stack. It means that all legacy ATA drivers were
removed and replaced by respective CAM drivers. If you are using
ATA device names in /etc/fstab or other places, make sure to update
them respectively (adX -> adaY, acdX -> cdY, afdX -> daY, astX -> saY,
where 'Y's are the sequential numbers starting from zero for each type
in order of detection, unless configured otherwise with tunables,
see cam(4)). There will be symbolic links created in /dev/ to map
old adX devices to the respective adaY. They should provide basic
compatibility for file systems mounting in most cases, but they do
not support old user-level APIs and do not have respective providers
in GEOM. Consider using updated management tools with new device names.
It is possible to load devices ahci, ata, siis and mvs as modules,
but option ATA_CAM should remain in kernel configuration to make ata
module work as CAM driver supporting legacy ATA controllers. Device ata
still can be used in modular fashion (atacore + ...). Modules atadisk
and atapi* are not used and won't affect operation in ATA_CAM mode.
Note that to use CAM-based ATA kernel should include CAM devices
scbus, pass, da (or explicitly ada), cd and optionally others. All of
them are parts of the cam module.
ataraid(4) functionality is now supported by the RAID GEOM class.
To use it you can load geom_raid kernel module and use graid(8) tool
for management. Instead of /dev/arX device names, use /dev/raid/rX.
No kernel config options or code have been removed, so if a problem
arises, please report it and optionally revert to the old ATA stack.
In order to do it you can remove from the kernel config:
options ATA_CAM
device ahci
device mvs
device siis
, and instead add back:
device atadisk # ATA disk drives
device ataraid # ATA RAID drives
device atapicd # ATAPI CDROM drives
device atapifd # ATAPI floppy drives
device atapist # ATAPI tape drives
20110423:
The default NFS server has been changed to the new server, which
was referred to as the experimental server. If you need to switch
back to the old NFS server, you must now put the "-o" option on
both the mountd and nfsd commands. This can be done using the
mountd_flags and nfs_server_flags rc.conf variables until an
update to the rc scripts is committed, which is coming soon.
20110418:
The GNU Objective-C runtime library (libobjc), and other Objective-C
related components have been removed from the base system. If you
require an Objective-C library, please use one of the available ports.
2011-03-31 08:32:53 +00:00
20110331:
ath(4) has been split into bus- and device- modules. if_ath contains
the HAL, the TX rate control and the network device code. if_ath_pci
contains the PCI bus glue. For Atheros MIPS embedded systems, if_ath_ahb
contains the AHB glue. Users need to load both if_ath_pci and if_ath
in order to use ath on everything else.
TO REPEAT: if_ath_ahb is not needed for normal users. Normal users only
need to load if_ath and if_ath_pci for ath(4) operation.
20110314:
As part of the replacement of sysinstall, the process of building
release media has changed significantly. For details, please re-read
release(7), which has been updated to reflect the new build process.
20110218:
GNU binutils 2.17.50 (as of 2007-07-03) has been merged to -HEAD. This
is the last available version under GPLv2. It brings a number of new
features, such as support for newer x86 CPU's (with SSE-3, SSSE-3, SSE
4.1 and SSE 4.2), better support for powerpc64, a number of new
directives, and lots of other small improvements. See the ChangeLog
file in contrib/binutils for the full details.
20110218:
IPsec's HMAC_SHA256-512 support has been fixed to be RFC4868
compliant, and will now use half of hash for authentication.
This will break interoperability with all stacks (including all
actual FreeBSD versions) who implement
draft-ietf-ipsec-ciph-sha-256-00 (they use 96 bits of hash for
authentication).
The only workaround with such peers is to use another HMAC
algorithm for IPsec ("phase 2") authentication.
20110207:
Remove the uio_yield prototype and symbol. This function has
been misnamed since it was introduced and should not be
globally exposed with this name. The equivalent functionality
is now available using kern_yield(curthread->td_user_pri).
The function remains undocumented.
20110112:
A SYSCTL_[ADD_]UQUAD was added for unsigned uint64_t pointers,
symmetric with the existing SYSCTL_[ADD_]QUAD. Type checking
for scalar sysctls is defined but disabled. Code that needs
UQUAD to pass the type checking that must compile on older
systems where the define is not present can check against
__FreeBSD_version >= 900030.
The system dialog(1) has been replaced with a new version previously
in ports as devel/cdialog. dialog(1) is mostly command-line compatible
with the previous version, but the libdialog associated with it has
a largely incompatible API. As such, the original version of libdialog
will be kept temporarily as libodialog, until its base system consumers
are replaced or updated. Bump __FreeBSD_version to 900030.
20110103:
If you are trying to run make universe on a -stable system, and you get
the following warning:
2011-01-12 19:28:52 +00:00
"Makefile", line 356: "Target architecture for i386/conf/GENERIC
unknown. config(8) likely too old."
or something similar to it, then you must upgrade your -stable system
to 8.2-Release or newer (really, any time after r210146 7/15/2010 in
stable/8) or build the config from the latest stable/8 branch and
install it on your system.
Prior to this date, building a current universe on 8-stable system from
between 7/15/2010 and 1/2/2011 would result in a weird shell parsing
2011-01-12 19:28:52 +00:00
error in the first kernel build phase. A new config on those old
systems will fix that problem for older versions of -current.
20101228:
The TCP stack has been modified to allow Khelp modules to interact with
it via helper hook points and store per-connection data in the TCP
control block. Bump __FreeBSD_version to 900029. User space tools that
rely on the size of struct tcpcb in tcp_var.h (e.g. sockstat) need to
be recompiled.
o Flesh out the generic IEEE 802.3 annex 31B full duplex flow control support in mii(4): - Merge generic flow control advertisement (which can be enabled by passing by MIIF_DOPAUSE to mii_attach(9)) and parsing support from NetBSD into mii_physubr.c and ukphy_subr.c. Unlike as in NetBSD, IFM_FLOW isn't implemented as a global option via the "don't care mask" but instead as a media specific option this. This has the following advantages: o allows flow control advertisement with autonegotiation to be turned on and off via ifconfig(8) with the default typically being off (though MIIF_FORCEPAUSE has been added causing flow control to be always advertised, allowing to easily MFC this changes for drivers that previously used home-grown support for flow control that behaved that way without breaking POLA) o allows to deal with PHY drivers where flow control advertisement with manual selection doesn't work or at least isn't implemented, like it's the case with brgphy(4), e1000phy(4) and ip1000phy(4), by setting MIIF_NOMANPAUSE o the available combinations of media options are readily available from the `ifconfig -m` output - Add IFM_FLOW to IFM_SHARED_OPTION_DESCRIPTIONS and IFM_ETH_RXPAUSE and IFM_ETH_TXPAUSE to IFM_SUBTYPE_ETHERNET_OPTION_DESCRIPTIONS so these are understood by ifconfig(8). o Make the master/slave support in mii(4) actually usable: - Change IFM_ETH_MASTER from being implemented as a global option via the "don't care mask" to a media specific one as it actually is only applicable to IFM_1000_T to date. - Let mii_phy_setmedia() set GTCR_MAN_MS in IFM_1000_T slave mode to actually configure manually selected slave mode (like we also do in the PHY specific implementations). - Add IFM_ETH_MASTER to IFM_SUBTYPE_ETHERNET_OPTION_DESCRIPTIONS so it is understood by ifconfig(8). o Switch bge(4), bce(4), msk(4), nfe(4) and stge(4) along with brgphy(4), e1000phy(4) and ip1000phy(4) to use the generic flow control support instead of home-grown solutions via IFM_FLAGs. This includes changing these PHY drivers and smcphy(4) to no longer unconditionally advertise support for flow control but only if the selected media has IFM_FLOW set (or MIIF_FORCEPAUSE is set) and implemented for these media variants, i.e. typically only for copper. o Switch brgphy(4), ciphy(4), e1000phy(4) and ip1000phy(4) to report and set IFM_1000_T master mode via IFM_ETH_MASTER instead of via IFF_LINK0 and some IFM_FLAGn. o Switch brgphy(4) to add at least the the supported copper media based on the contents of the BMSR via mii_phy_add_media() instead of hardcoding them. The latter approach seems to have developed historically, besides causing unnecessary code duplication it was also undesirable because brgphy_mii_phy_auto() already based the capability advertisement on the contents of the BMSR though. o Let brgphy(4) set IFM_1000_T master mode on all supported PHY and not just BCM5701. Apparently this was a misinterpretation of a workaround in the Linux tg3 driver; BCM5701 seem to require RGPHY_1000CTL_MSE and BRGPHY_1000CTL_MSC to be set when configuring autonegotiation but this doesn't mean we can't set these as well on other PHYs for manual media selection. o Let ukphy_status() report IFM_1000_T master mode via IFM_ETH_MASTER so IFM_1000_T master mode support now is generally available with all PHY drivers. o Don't let e1000phy(4) set master/slave bits for IFM_1000_SX as it's not applicable there. Reviewed by: yongari (plus additional testing) Obtained from: NetBSD (partially), OpenBSD (partially) MFC after: 2 weeks
2010-11-14 13:26:10 +00:00
20101114:
Generic IEEE 802.3 annex 31B full duplex flow control support has been
added to mii(4) and bge(4), bce(4), msk(4), nfe(4) and stge(4) along
with brgphy(4), e1000phy(4) as well as ip1000phy() have been converted
to take advantage of it instead of using custom implementations. This
means that these drivers now no longer unconditionally advertise
support for flow control but only do so if flow control is a selected
media option. This was implemented in the generic support that way in
order to allow flow control to be switched on and off via ifconfig(8)
with the PHY specific default to typically off in order to protect
from unwanted effects. Consequently, if you used flow control with
one of the above mentioned drivers you now need to explicitly enable
it, for example via:
ifconfig bge0 media auto mediaopt flowcontrol
Along with the above mentioned changes generic support for setting
1000baseT master mode also has been added and brgphy(4), ciphy(4),
e1000phy(4) as well as ip1000phy(4) have been converted to take
advantage of it. This means that these drivers now no longer take the
link0 parameter for selecting master mode but the master media option
has to be used instead, for example like in the following:
ifconfig bge0 media 1000baseT mediaopt full-duplex,master
Selection of master mode now is also available with all other PHY
drivers supporting 1000baseT.
This commit marks the first formal contribution of the "Five New TCP Congestion Control Algorithms for FreeBSD" FreeBSD Foundation funded project. More details about the project are available at: http://caia.swin.edu.au/freebsd/5cc/ - Add a KPI and supporting infrastructure to allow modular congestion control algorithms to be used in the net stack. Algorithms can maintain per-connection state if required, and connections maintain their own algorithm pointer, which allows different connections to concurrently use different algorithms. The TCP_CONGESTION socket option can be used with getsockopt()/setsockopt() to programmatically query or change the congestion control algorithm respectively from within an application at runtime. - Integrate the framework with the TCP stack in as least intrusive a manner as possible. Care was also taken to develop the framework in a way that should allow integration with other congestion aware transport protocols (e.g. SCTP) in the future. The hope is that we will one day be able to share a single set of congestion control algorithm modules between all congestion aware transport protocols. - Introduce a new congestion recovery (TF_CONGRECOVERY) state into the TCP stack and use it to decouple the meaning of recovery from a congestion event and recovery from packet loss (TF_FASTRECOVERY) a la RFC2581. ECN and delay based congestion control protocols don't generally need to recover from packet loss and need a different way to note a congestion recovery episode within the stack. - Remove the net.inet.tcp.newreno sysctl, which simplifies some portions of code and ensures the stack always uses the appropriate mechanisms for recovering from packet loss during a congestion recovery episode. - Extract the NewReno congestion control algorithm from the TCP stack and massage it into module form. NewReno is always built into the kernel and will remain the default algorithm for the forseeable future. Implementations of additional different algorithms will become available in the near future. - Bump __FreeBSD_version to 900025 and note in UPDATING that rebuilding code that relies on the size of "struct tcpcb" is required. Many thanks go to the Cisco University Research Program Fund at Community Foundation Silicon Valley and the FreeBSD Foundation. Their support of our work at the Centre for Advanced Internet Architectures, Swinburne University of Technology is greatly appreciated. In collaboration with: David Hayes <dahayes at swin edu au> and Grenville Armitage <garmitage at swin edu au> Sponsored by: Cisco URP, FreeBSD Foundation Reviewed by: rpaulo Tested by: David Hayes (and many others over the years) MFC after: 3 months
2010-11-12 06:41:55 +00:00
20101111:
The TCP stack has received a significant update to add support for
modularised congestion control and generally improve the clarity of
congestion control decisions. Bump __FreeBSD_version to 900025. User
space tools that rely on the size of struct tcpcb in tcp_var.h (e.g.
sockstat) need to be recompiled.
20101002:
The man(1) utility has been replaced by a new version that no longer
uses /etc/manpath.config. Please consult man.conf(5) for how to
migrate local entries to the new format.
20100928:
The copyright strings printed by login(1) and sshd(8) at the time of a
new connection have been removed to follow other operating systems and
upstream sshd.
2010-09-15 10:04:41 +00:00
20100915:
A workaround for a fixed ld bug has been removed in kernel code,
so make sure that your system ld is built from sources after
revision 210245 from 2010-07-19 (r211583 if building head kernel
on stable/8, r211584 for stable/7; both from 2010-08-21).
A symptom of incorrect ld version is different addresses for
set_pcpu section and __start_set_pcpu symbol in kernel and/or modules.
2010-09-15 10:04:41 +00:00
20100913:
The $ipv6_prefer variable in rc.conf(5) has been split into
$ip6addrctl_policy and $ipv6_activate_all_interfaces.
The $ip6addrctl_policy is a variable to choose a pre-defined
address selection policy set by ip6addrctl(8). A value
"ipv4_prefer", "ipv6_prefer" or "AUTO" can be specified. The
default is "AUTO".
The $ipv6_activate_all_interfaces specifies whether IFDISABLED
flag (see an entry of 20090926) is set on an interface with no
corresponding $ifconfig_IF_ipv6 line. The default is "NO" for
security reason. If you want IPv6 link-local address on all
interfaces by default, set this to "YES".
The old ipv6_prefer="YES" is equivalent to
ipv6_activate_all_interfaces="YES" and
ip6addrctl_policy="ipv6_prefer".
20100913:
DTrace has grown support for userland tracing. Due to this, DTrace is
now i386 and amd64 only.
dtruss(1) is now installed by default on those systems and a new
kernel module is needed for userland tracing: fasttrap.
No changes to your kernel config file are necessary to enable
userland tracing, but you might consider adding 'STRIP=' and
'CFLAGS+=-fno-omit-frame-pointer' to your make.conf if you want
to have informative userland stack traces in DTrace (ustack).
2010-07-25 18:32:59 +00:00
20100725:
The acpi_aiboost(4) driver has been removed in favor of the new
aibs(4) driver. You should update your kernel configuration file.
20100722:
BSD grep has been imported to the base system and it is built by
default. It is completely BSD licensed, highly GNU-compatible, uses
less memory than its GNU counterpart and has a small codebase.
However, it is slower than its GNU counterpart, which is mostly
noticeable for larger searches, for smaller ones it is measurable
but not significant. The reason is complex, the most important factor
is that we lack a modern and efficient regex library and GNU
overcomes this by optimizing the searches internally. Future work
on improving the regex performance is planned, for the meantime,
users that need better performance, can build GNU grep instead by
setting the WITH_GNU_GREP knob.
20100713:
Due to the import of powerpc64 support, all existing powerpc kernel
configuration files must be updated with a machine directive like this:
machine powerpc powerpc
In addition, an updated config(8) is required to build powerpc kernels
after this change.
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
20100713:
A new version of ZFS (version 15) has been merged to -HEAD.
This version uses a python library for the following subcommands:
zfs allow, zfs unallow, zfs groupspace, zfs userspace.
For full functionality of these commands the following port must
be installed: sysutils/py-zfs
20100429:
'vm_page's are now hashed by physical address to an array of mutexes.
Currently this is only used to serialize access to hold_count. Over
time the page queue mutex will be peeled away. This changes the size
of pmap on every architecture. And requires all callers of vm_page_hold
and vm_page_unhold to be updated.
20100402:
WITH_CTF can now be specified in src.conf (not recommended, there
are some problems with static executables), make.conf (would also
affect ports which do not use GNU make and do not override the
compile targets) or in the kernel config (via "makeoptions
WITH_CTF=yes").
When WITH_CTF was specified there before this was silently ignored,
so make sure that WITH_CTF is not used in places which could lead
to unwanted behavior.
20100311:
The kernel option COMPAT_IA32 has been replaced with COMPAT_FREEBSD32
to allow 32-bit compatibility on non-x86 platforms. All kernel
configurations on amd64 and ia64 platforms using these options must
be modified accordingly.
20100113:
The utmp user accounting database has been replaced with utmpx,
the user accounting interface standardized by POSIX.
Unfortunately the semantics of utmp and utmpx don't match,
making it practically impossible to support both interfaces.
The user accounting database is used by tools like finger(1),
last(1), talk(1), w(1) and ac(8).
All applications in the base system use utmpx. This means only
local binaries (e.g. from the ports tree) may still use these
utmp database files. These applications must be rebuilt to make
use of utmpx.
After the system has been upgraded, it is safe to remove the old
log files (/var/run/utmp, /var/log/lastlog and /var/log/wtmp*),
assuming their contents is of no importance anymore. Old wtmp
databases can only be used by last(1) and ac(8) after they have
been converted to the new format using wtmpcvt(1).
20100108:
Introduce the kernel thread "deadlock resolver" (which can be enabled
via the DEADLKRES option, see NOTES for more details) and the
sleepq_type() function for sleepqueues.
20091202:
The rc.firewall and rc.firewall6 were unified, and
rc.firewall6 and rc.d/ip6fw were removed.
According to the removal of rc.d/ip6fw, ipv6_firewall_* rc
variables are obsoleted. Instead, the following new rc
variables are added to rc.d/ipfw:
firewall_client_net_ipv6, firewall_simple_iif_ipv6,
firewall_simple_inet_ipv6, firewall_simple_oif_ipv6,
firewall_simple_onet_ipv6, firewall_trusted_ipv6
The meanings correspond to the relevant IPv4 variables.
20091125:
8.0-RELEASE.
20091113:
The default terminal emulation for syscons(4) has been changed
from cons25 to xterm on all platforms except pc98. This means
that the /etc/ttys file needs to be updated to ensure correct
operation of applications on the console.
The terminal emulation style can be toggled per window by using
vidcontrol(1)'s -T flag. The TEKEN_CONS25 kernel configuration
options can be used to change the compile-time default back to
cons25.
To prevent graphical artifacts, make sure the TERM environment
variable is set to match the terminal emulation that is being
performed by syscons(4).
20091109:
The layout of the structure ieee80211req_scan_result has changed.
Applications that require wireless scan results (e.g. ifconfig(8))
from net80211 need to be recompiled.
Applications such as wpa_supplicant(8) may require a full world
build without using NO_CLEAN in order to get synchronized with the
new structure.
20091025:
The iwn(4) driver has been updated to support the 5000 and 5150 series.
There's one kernel module for each firmware. Adding "device iwnfw"
to the kernel configuration file means including all three firmware
images inside the kernel. If you want to include just the one for
your wireless card, use the devices iwn4965fw, iwn5000fw or
iwn5150fw.
Fix several logic bugs in the previous IPv6 variable change and re-add $ipv6_enable support for backward compatibility. From UPDATING: 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF for IPv4. For aliases, $ifconfig_IF_aliasN should be used. Note that both variables need the "inet6" keyword at the head. Do not set $ipv6_network_interfaces manually if you do not understand what you are doing. It is not needed in most cases. $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but they are obsolete. 2. $ipv6_enable is obsolete. Use $ipv6_prefer and/or "inet6 accept_rtadv" keyword in ifconfig(8) instead. If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and all configured interfaces have "inet6 accept_rtadv" in the $ifconfig_IF_ipv6. These are for backward compatibility. 3. A new variable $ipv6_prefer has been added. If NO, IPv6 functionality of interfaces with no corresponding $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag, and the default address selection policy of ip6addrctl(8) is the IPv4-preferred one (see rc.d/ip6addrctl for more details). Note that if you want to configure IPv6 functionality on the disabled interfaces after boot, first you need to clear the flag by using ifconfig(8) like: ifconfig em0 inet6 -ifdisabled If YES, the default address selection policy is set as IPv6-preferred. The default value of $ipv6_prefer is NO. 4. If your system need to receive Router Advertisement messages, define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8) scripts automatically invoke rtsol(8) when the interface becomes UP. The Router Advertisement messages are used for SLAAC (State-Less Address AutoConfiguration).
2009-09-26 18:59:00 +00:00
20090926:
The rc.d/network_ipv6, IPv6 configuration script has been integrated
into rc.d/netif. The changes are the following:
1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF
for IPv4. For aliases, $ifconfig_IF_aliasN should be used.
Note that both variables need the "inet6" keyword at the head.
Do not set $ipv6_network_interfaces manually if you do not
understand what you are doing. It is not needed in most cases.
$ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but
they are obsolete.
2. $ipv6_enable is obsolete. Use $ipv6_prefer and
"inet6 accept_rtadv" keyword in ifconfig(8) instead.
If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and
all configured interfaces have "inet6 accept_rtadv" in the
$ifconfig_IF_ipv6. These are for backward compatibility.
3. A new variable $ipv6_prefer has been added. If NO, IPv6
functionality of interfaces with no corresponding
$ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag,
and the default address selection policy of ip6addrctl(8)
is the IPv4-preferred one (see rc.d/ip6addrctl for more details).
Note that if you want to configure IPv6 functionality on the
disabled interfaces after boot, first you need to clear the flag by
using ifconfig(8) like:
ifconfig em0 inet6 -ifdisabled
If YES, the default address selection policy is set as
IPv6-preferred.
The default value of $ipv6_prefer is NO.
4. If your system need to receive Router Advertisement messages,
define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8)
scripts automatically invoke rtsol(8) when the interface becomes
UP. The Router Advertisement messages are used for SLAAC
(State-Less Address AutoConfiguration).
20090922:
802.11s D3.03 support was committed. This is incompatible with the
previous code, which was based on D3.0.
Fix several logic bugs in the previous IPv6 variable change and re-add $ipv6_enable support for backward compatibility. From UPDATING: 1. To use IPv6, simply define $ifconfig_IF_ipv6 like $ifconfig_IF for IPv4. For aliases, $ifconfig_IF_aliasN should be used. Note that both variables need the "inet6" keyword at the head. Do not set $ipv6_network_interfaces manually if you do not understand what you are doing. It is not needed in most cases. $ipv6_ifconfig_IF and $ipv6_ifconfig_IF_aliasN still work, but they are obsolete. 2. $ipv6_enable is obsolete. Use $ipv6_prefer and/or "inet6 accept_rtadv" keyword in ifconfig(8) instead. If you define $ipv6_enable=YES, it means $ipv6_prefer=YES and all configured interfaces have "inet6 accept_rtadv" in the $ifconfig_IF_ipv6. These are for backward compatibility. 3. A new variable $ipv6_prefer has been added. If NO, IPv6 functionality of interfaces with no corresponding $ifconfig_IF_ipv6 is disabled by using "inet6 ifdisabled" flag, and the default address selection policy of ip6addrctl(8) is the IPv4-preferred one (see rc.d/ip6addrctl for more details). Note that if you want to configure IPv6 functionality on the disabled interfaces after boot, first you need to clear the flag by using ifconfig(8) like: ifconfig em0 inet6 -ifdisabled If YES, the default address selection policy is set as IPv6-preferred. The default value of $ipv6_prefer is NO. 4. If your system need to receive Router Advertisement messages, define "inet6 accept_rtadv" in $ifconfig_IF_ipv6. The rc(8) scripts automatically invoke rtsol(8) when the interface becomes UP. The Router Advertisement messages are used for SLAAC (State-Less Address AutoConfiguration).
2009-09-26 18:59:00 +00:00
20090912:
A sysctl variable net.inet6.ip6.accept_rtadv now sets the default value
of a per-interface flag ND6_IFF_ACCEPT_RTADV, not a global knob to
control whether accepting Router Advertisement messages or not.
Also, a per-interface flag ND6_IFF_AUTO_LINKLOCAL has been added and
a sysctl variable net.inet6.ip6.auto_linklocal is its default value.
The ifconfig(8) utility now supports these flags.
20090910:
ZFS snapshots are now mounted with MNT_IGNORE flag. Use -v option for
mount(8) and -a option for df(1) to see them.
20090825:
The old tunable hw.bus.devctl_disable has been superseded by
hw.bus.devctl_queue. hw.bus.devctl_disable=1 in loader.conf should be
replaced by hw.bus.devctl_queue=0. The default for this new tunable
is 1000.
2009-08-13 17:09:45 +00:00
20090813:
Remove the option STOP_NMI. The default action is now to use NMI only
for KDB via the newly introduced function stop_cpus_hard() and
maintain stop_cpus() to just use a normal IPI_STOP on ia32 and amd64.
20090803:
The stable/8 branch created in subversion. This corresponds to the
RELENG_8 branch in CVS.
2009-08-13 17:09:45 +00:00
20090719:
Bump the shared library version numbers for all libraries that do not
use symbol versioning as part of the 8.0-RELEASE cycle. Bump
__FreeBSD_version to 800105.
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
20090714:
Due to changes in the implementation of virtual network stack support,
all network-related kernel modules must be recompiled. As this change
breaks the ABI, bump __FreeBSD_version to 800104.
Build on Jeff Roberson's linker-set based dynamic per-CPU allocator (DPCPU), as suggested by Peter Wemm, and implement a new per-virtual network stack memory allocator. Modify vnet to use the allocator instead of monolithic global container structures (vinet, ...). This change solves many binary compatibility problems associated with VIMAGE, and restores ELF symbols for virtualized global variables. Each virtualized global variable exists as a "reference copy", and also once per virtual network stack. Virtualized global variables are tagged at compile-time, placing the in a special linker set, which is loaded into a contiguous region of kernel memory. Virtualized global variables in the base kernel are linked as normal, but those in modules are copied and relocated to a reserved portion of the kernel's vnet region with the help of a the kernel linker. Virtualized global variables exist in per-vnet memory set up when the network stack instance is created, and are initialized statically from the reference copy. Run-time access occurs via an accessor macro, which converts from the current vnet and requested symbol to a per-vnet address. When "options VIMAGE" is not compiled into the kernel, normal global ELF symbols will be used instead and indirection is avoided. This change restores static initialization for network stack global variables, restores support for non-global symbols and types, eliminates the need for many subsystem constructors, eliminates large per-subsystem structures that caused many binary compatibility issues both for monitoring applications (netstat) and kernel modules, removes the per-function INIT_VNET_*() macros throughout the stack, eliminates the need for vnet_symmap ksym(2) munging, and eliminates duplicate definitions of virtualized globals under VIMAGE_GLOBALS. Bump __FreeBSD_version and update UPDATING. Portions submitted by: bz Reviewed by: bz, zec Discussed with: gnn, jamie, jeff, jhb, julian, sam Suggested by: peter Approved by: re (kensmith)
2009-07-14 22:48:30 +00:00
20090713:
The TOE interface to the TCP syncache has been modified to remove
struct tcpopt (<netinet/tcp_var.h>) from the ABI of the network stack.
The cxgb driver is the only TOE consumer affected by this change, and
needs to be recompiled along with the kernel. As this change breaks
the ABI, bump __FreeBSD_version to 800103.
20090712:
Padding has been added to struct tcpcb, sackhint and tcpstat in
<netinet/tcp_var.h> to facilitate future MFCs and bug fixes whilst
maintaining the ABI. However, this change breaks the ABI, so bump
__FreeBSD_version to 800102. User space tools that rely on the size of
any of these structs (e.g. sockstat) need to be recompiled.
20090630:
The NFS_LEGACYRPC option has been removed along with the old kernel
RPC implementation that this option selected. Kernel configurations
may need to be adjusted.
20090629:
The network interface device nodes at /dev/net/<interface> have been
removed. All ioctl operations can be performed the normal way using
routing sockets. The kqueue functionality can generally be replaced
with routing sockets.
20090628:
The documentation from the FreeBSD Documentation Project (Handbook,
FAQ, etc.) is now installed via packages by sysinstall(8) and under
the /usr/local/share/doc/freebsd directory instead of /usr/share/doc.
20090624:
The ABI of various structures related to the SYSV IPC API have been
changed. As a result, the COMPAT_FREEBSD[456] and COMPAT_43 kernel
options now all require COMPAT_FREEBSD7. Bump __FreeBSD_version to
800100.
20090622:
Layout of struct vnet has changed as routing related variables were
moved to their own Vimage module. Modules need to be recompiled. Bump
__FreeBSD_version to 800099.
Rework the credential code to support larger values of NGROUPS and NGROUPS_MAX, eliminate ABI dependencies on them, and raise the to 1024 and 1023 respectively. (Previously they were equal, but under a close reading of POSIX, NGROUPS_MAX was defined to be too large by 1 since it is the number of supplemental groups, not total number of groups.) The bulk of the change consists of converting the struct ucred member cr_groups from a static array to a pointer. Do the equivalent in kinfo_proc. Introduce new interfaces crcopysafe() and crsetgroups() for duplicating a process credential before modifying it and for setting group lists respectively. Both interfaces take care for the details of allocating groups array. crsetgroups() takes care of truncating the group list to the current maximum (NGROUPS) if necessary. In the future, crsetgroups() may be responsible for insuring invariants such as sorting the supplemental groups to allow groupmember() to be implemented as a binary search. Because we can not change struct xucred without breaking application ABIs, we leave it alone and introduce a new XU_NGROUPS value which is always 16 and is to be used or NGRPS as appropriate for things such as NFS which need to use no more than 16 groups. When feasible, truncate the group list rather than generating an error. Minor changes: - Reduce the number of hand rolled versions of groupmember(). - Do not assign to both cr_gid and cr_groups[0]. - Modify ipfw to cache ucreds instead of part of their contents since they are immutable once referenced by more than one entity. Submitted by: Isilon Systems (initial implementation) X-MFC after: never PR: bin/113398 kern/133867
2009-06-19 17:10:35 +00:00
20090619:
NGROUPS_MAX and NGROUPS have been increased from 16 to 1023 and 1024
respectively. As long as no more than 16 groups per process are used,
no changes should be visible. When more than 16 groups are used, old
binaries may fail if they call getgroups() or getgrouplist() with
statically sized storage. Recompiling will work around this, but
applications should be modified to use dynamically allocated storage
for group arrays as POSIX.1-2008 does not cap an implementation's
number of supported groups at NGROUPS_MAX+1 as previous versions did.
NFS and portalfs mounts may also be affected as the list of groups is
truncated to 16. Users of NFS who use more than 16 groups, should
take care that negative group permissions are not used on the exported
file systems as they will not be reliable unless a GSSAPI based
authentication method is used.
20090616:
The compiling option ADAPTIVE_LOCKMGRS has been introduced. This
option compiles in the support for adaptive spinning for lockmgrs
which want to enable it. The lockinit() function now accepts the flag
LK_ADAPTIVE in order to make the lock object subject to adaptive
spinning when both held in write and read mode.
20090613:
The layout of the structure returned by IEEE80211_IOC_STA_INFO has
changed. User applications that use this ioctl need to be rebuilt.
20090611:
The layout of struct thread has changed. Kernel and modules need to
be rebuilt.
20090608:
The layout of structs ifnet, domain, protosw and vnet_net has changed.
Kernel modules need to be rebuilt. Bump __FreeBSD_version to 800097.
20090602:
window(1) has been removed from the base system. It can now be
installed from ports. The port is called misc/window.
20090601:
The way we are storing and accessing `routing table' entries has
changed. Programs reading the FIB, like netstat, need to be
re-compiled.
20090601:
A new netisr implementation has been added for FreeBSD 8. Network
file system modules, such as igmp, ipdivert, and others, should be
rebuilt.
Bump __FreeBSD_version to 800096.
20090530:
Remove the tunable/sysctl debug.mpsafevfs as its initial purpose is no
more valid.
20090530:
Add VOP_ACCESSX(9). File system modules need to be rebuilt.
Bump __FreeBSD_version to 800094.
20090529:
Add mnt_xflag field to 'struct mount'. File system modules need to be
rebuilt.
Bump __FreeBSD_version to 800093.
20090528:
The compiling option ADAPTIVE_SX has been retired while it has been
introduced the option NO_ADAPTIVE_SX which handles the reversed logic.
The KPI for sx_init_flags() changes as accepting flags:
SX_ADAPTIVESPIN flag has been retired while the SX_NOADAPTIVE flag has
been introduced in order to handle the reversed logic.
Bump __FreeBSD_version to 800092.
20090527:
Add support for hierarchical jails. Remove global securelevel.
Bump __FreeBSD_version to 800091.
20090523:
The layout of struct vnet_net has changed, therefore modules
need to be rebuilt.
Bump __FreeBSD_version to 800090.
20090523:
The newly imported zic(8) produces a new format in the output. Please
run tzsetup(8) to install the newly created data to /etc/localtime.
20090520:
The sysctl tree for the usb stack has renamed from hw.usb2.* to
hw.usb.* and is now consistent again with previous releases.
2009-05-20 20:05:56 +00:00
20090520:
802.11 monitor mode support was revised and driver api's were changed.
Drivers dependent on net80211 now support DLT_IEEE802_11_RADIO instead
of DLT_IEEE802_11. No user-visible data structures were changed but
applications that use DLT_IEEE802_11 may require changes.
2009-05-20 20:05:56 +00:00
Bump __FreeBSD_version to 800088.
Permit buiding kernels with options VIMAGE, restricted to only a single active network stack instance. Turning on options VIMAGE at compile time yields the following changes relative to default kernel build: 1) V_ accessor macros for virtualized variables resolve to structure fields via base pointers, instead of being resolved as fields in global structs or plain global variables. As an example, V_ifnet becomes: options VIMAGE: ((struct vnet_net *) vnet_net)->_ifnet default build: vnet_net_0._ifnet options VIMAGE_GLOBALS: ifnet 2) INIT_VNET_* macros will declare and set up base pointers to be used by V_ accessor macros, instead of resolving to whitespace: INIT_VNET_NET(ifp->if_vnet); becomes struct vnet_net *vnet_net = (ifp->if_vnet)->mod_data[VNET_MOD_NET]; 3) Memory for vnet modules registered via vnet_mod_register() is now allocated at run time in sys/kern/kern_vimage.c, instead of per vnet module structs being declared as globals. If required, vnet modules can now request the framework to provide them with allocated bzeroed memory by filling in the vmi_size field in their vmi_modinfo structures. 4) structs socket, ifnet, inpcbinfo, tcpcb and syncache_head are extended to hold a pointer to the parent vnet. options VIMAGE builds will fill in those fields as required. 5) curvnet is introduced as a new global variable in options VIMAGE builds, always pointing to the default and only struct vnet. 6) struct sysctl_oid has been extended with additional two fields to store major and minor virtualization module identifiers, oid_v_subs and oid_v_mod. SYSCTL_V_* family of macros will fill in those fields accordingly, and store the offset in the appropriate vnet container struct in oid_arg1. In sysctl handlers dealing with virtualized sysctls, the SYSCTL_RESOLVE_V_ARG1() macro will compute the address of the target variable and make it available in arg1 variable for further processing. Unused fields in structs vnet_inet, vnet_inet6 and vnet_ipfw have been deleted. Reviewed by: bz, rwatson Approved by: julian (mentor)
2009-04-30 13:36:26 +00:00
20090430:
The layout of the following structs has changed: sysctl_oid,
socket, ifnet, inpcbinfo, tcpcb, syncache_head, vnet_inet,
vnet_inet6 and vnet_ipfw. Most modules need to be rebuild or
panics may be experienced. World rebuild is required for
correctly checking networking state from userland.
Bump __FreeBSD_version to 800085.
Bite the bullet, and make the IPv6 SSM and MLDv2 mega-commit: import from p4 bms_netdev. Summary of changes: * Connect netinet6/in6_mcast.c to build. The legacy KAME KPIs are mostly preserved. * Eliminate now dead code from ip6_output.c. Don't do mbuf bingo, we are not going to do RFC 2292 style CMSG tricks for multicast options as they are not required by any current IPv6 normative reference. * Refactor transports (UDP, raw_ip6) to do own mcast filtering. SCTP, TCP unaffected by this change. * Add ip6_msource, in6_msource structs to in6_var.h. * Hookup mld_ifinfo state to in6_ifextra, allocate from domifattach path. * Eliminate IN6_LOOKUP_MULTI(), it is no longer referenced. Kernel consumers which need this should use in6m_lookup(). * Refactor IPv6 socket group memberships to use a vector (like IPv4). * Update ifmcstat(8) for IPv6 SSM. * Add witness lock order for IN6_MULTI_LOCK. * Move IN6_MULTI_LOCK out of lower ip6_output()/ip6_input() paths. * Introduce IP6STAT_ADD/SUB/INC/DEC as per rwatson's IPv4 cleanup. * Update carp(4) for new IPv6 SSM KPIs. * Virtualize ip6_mrouter socket. Changes mostly localized to IPv6 MROUTING. * Don't do a local group lookup in MROUTING. * Kill unused KAME prototypes in6_purgemkludge(), in6_restoremkludge(). * Preserve KAME DAD timer jitter behaviour in MLDv1 compatibility mode. * Bump __FreeBSD_version to 800084. * Update UPDATING. NOTE WELL: * This code hasn't been tested against real MLDv2 queriers (yet), although the on-wire protocol has been verified in Wireshark. * There are a few unresolved issues in the socket layer APIs to do with scope ID propagation. * There is a LOR present in ip6_output()'s use of in6_setscope() which needs to be resolved. See comments in mld6.c. This is believed to be benign and can't be avoided for the moment without re-introducing an indirect netisr. This work was mostly derived from the IGMPv3 implementation, and has been sponsored by a third party.
2009-04-29 19:19:13 +00:00
20090429:
MLDv2 and Source-Specific Multicast (SSM) have been merged
to the IPv6 stack. VIMAGE hooks are in but not yet used.
The implementation of SSM within FreeBSD's IPv6 stack closely
follows the IPv4 implementation.
For kernel developers:
* The most important changes are that the ip6_output() and
ip6_input() paths no longer take the IN6_MULTI_LOCK,
and this lock has been downgraded to a non-recursive mutex.
* As with the changes to the IPv4 stack to support SSM, filtering
of inbound multicast traffic must now be performed by transport
protocols within the IPv6 stack. This does not apply to TCP and
SCTP, however, it does apply to UDP in IPv6 and raw IPv6.
* The KPIs used by IPv6 multicast are similar to those used by
the IPv4 stack, with the following differences:
* im6o_mc_filter() is analogous to imo_multicast_filter().
* The legacy KAME entry points in6_joingroup and in6_leavegroup()
are shimmed to in6_mc_join() and in6_mc_leave() respectively.
* IN6_LOOKUP_MULTI() has been deprecated and removed.
* IPv6 relies on MLD for the DAD mechanism. KAME's internal KPIs
for MLDv1 have an additional 'timer' argument which is used to
jitter the initial membership report for the solicited-node
multicast membership on-link.
* This is not strictly needed for MLDv2, which already jitters
its report transmissions. However, the 'timer' argument is
preserved in case MLDv1 is active on the interface.
* The KAME linked-list based IPv6 membership implementation has
been refactored to use a vector similar to that used by the IPv4
stack.
Code which maintains a list of its own multicast memberships
internally, e.g. carp, has been updated to reflect the new
semantics.
* There is a known Lock Order Reversal (LOR) due to in6_setscope()
acquiring the IF_AFDATA_LOCK and being called within ip6_output().
Whilst MLDv2 tries to avoid this otherwise benign LOR, it is an
implementation constraint which needs to be addressed in HEAD.
For application developers:
* The changes are broadly similar to those made for the IPv4
stack.
* The use of IPv4 and IPv6 multicast socket options on the same
socket, using mapped addresses, HAS NOT been tested or supported.
* There are a number of issues with the implementation of various
IPv6 multicast APIs which need to be resolved in the API surface
before the implementation is fully compatible with KAME userland
use, and these are mostly to do with interface index treatment.
* The literature available discusses the use of either the delta / ASM
API with setsockopt(2)/getsockopt(2), or the full-state / ASM API
using setsourcefilter(3)/getsourcefilter(3). For more information
please refer to RFC 3768, 'Socket Interface Extensions for
Multicast Source Filters'.
* Applications which use the published RFC 3678 APIs should be fine.
For systems administrators:
* The mtest(8) utility has been refactored to support IPv6, in
addition to IPv4. Interface addresses are no longer accepted
as arguments, their names must be used instead. The utility
will map the interface name to its first IPv4 address as
returned by getifaddrs(3).
* The ifmcstat(8) utility has also been updated to print the MLDv2
endpoint state and source filter lists via sysctl(3).
* The net.inet6.ip6.mcast.loop sysctl may be tuned to 0 to disable
loopback of IPv6 multicast datagrams by default; it defaults to 1
to preserve the existing behaviour. Disabling multicast loopback is
recommended for optimal system performance.
* The IPv6 MROUTING code has been changed to examine this sysctl
instead of attempting to perform a group lookup before looping
back forwarded datagrams.
Bump __FreeBSD_version to 800084.
20090422:
Implement low-level Bluetooth HCI API.
Bump __FreeBSD_version to 800083.
20090419:
The layout of struct malloc_type, used by modules to register new
memory allocation types, has changed. Most modules will need to
be rebuilt or panics may be experienced.
Bump __FreeBSD_version to 800081.
20090415:
Anticipate overflowing inp_flags - add inp_flags2.
This changes most offsets in inpcb, so checking v4 connection
state will require a world rebuild.
Bump __FreeBSD_version to 800080.
20090415:
Add an llentry to struct route and struct route_in6. Modules
embedding a struct route will need to be recompiled.
Bump __FreeBSD_version to 800079.
20090414:
The size of rt_metrics_lite and by extension rtentry has changed.
Networking administration apps will need to be recompiled.
The route command now supports show as an alias for get, weighting
of routes, sticky and nostick flags to alter the behavior of stateful
load balancing.
Bump __FreeBSD_version to 800078.
2009-04-12 19:42:25 +00:00
20090408:
Do not use Giant for kbdmux(4) locking. This is wrong and
apparently causing more problems than it solves. This will
re-open the issue where interrupt handlers may race with
kbdmux(4) in polling mode. Typical symptoms include (but
not limited to) duplicated and/or missing characters when
low level console functions (such as gets) are used while
interrupts are enabled (for example geli password prompt,
mountroot prompt etc.). Disabling kbdmux(4) may help.
20090407:
The size of structs vnet_net, vnet_inet and vnet_ipfw has changed;
kernel modules referencing any of the above need to be recompiled.
Bump __FreeBSD_version to 800075.
20090320:
GEOM_PART has become the default partition slicer for storage devices,
replacing GEOM_MBR, GEOM_BSD, GEOM_PC98 and GEOM_GPT slicers. It
introduces some changes:
MSDOS/EBR: the devices created from MSDOS extended partition entries
(EBR) can be named differently than with GEOM_MBR and are now symlinks
to devices with offset-based names. fstabs may need to be modified.
BSD: the "geometry does not match label" warning is harmless in most
cases but it points to problems in file system misalignment with
disk geometry. The "c" partition is now implicit, covers the whole
top-level drive and cannot be (mis)used by users.
General: Kernel dumps are now not allowed to be written to devices
whose partition types indicate they are meant to be used for file
systems (or, in case of MSDOS partitions, as something else than
the "386BSD" type).
Most of these changes date approximately from 200812.
20090319:
The uscanner(4) driver has been removed from the kernel. This follows
Linux removing theirs in 2.6 and making libusb the default interface
(supported by sane).
Introduce a number of changes to the MROUTING code. This is purely a forwarding plane cleanup; no control plane code is involved. Summary: * Split IPv4 and IPv6 MROUTING support. The static compile-time kernel option remains the same, however, the modules may now be built for IPv4 and IPv6 separately as ip_mroute_mod and ip6_mroute_mod. * Clean up the IPv4 multicast forwarding code to use BSD queue and hash table constructs. Don't build our own timer abstractions when ratecheck() and timevalclear() etc will do. * Expose the multicast forwarding cache (MFC) and virtual interface table (VIF) as sysctls, to reduce netstat's dependence on libkvm for this information for running kernels. * bandwidth meters however still require libkvm. * Make the MFC hash table size a boot/load-time tunable ULONG, net.inet.ip.mfchashsize (defaults to 256). * Remove unused members from struct vif and struct mfc. * Kill RSVP support, as no current RSVP implementation uses it. These stubs could be moved to raw_ip.c. * Don't share locks or initialization between IPv4 and IPv6. * Don't use a static struct route_in6 in ip6_mroute.c. The v6 code is still using a cached struct route_in6, this is moved to mif6 for the time being. * More cleanup remains to be merged from ip_mroute.c to ip6_mroute.c. v4 path tested using ports/net/mcast-tools. v6 changes are mostly mechanical locking and *have not* been tested. As these changes partially break some kernel ABIs, they will not be MFCed. There is a lot more work to be done here. Reviewed by: Pavlin Radoslavov
2009-03-19 01:43:03 +00:00
20090319:
The multicast forwarding code has been cleaned up. netstat(1)
only relies on KVM now for printing bandwidth upcall meters.
The IPv4 and IPv6 modules are split into ip_mroute_mod and
ip6_mroute_mod respectively. The config(5) options for statically
compiling this code remain the same, i.e. 'options MROUTING'.
20090315:
Support for the IFF_NEEDSGIANT network interface flag has been
removed, which means that non-MPSAFE network device drivers are no
longer supported. In particular, if_ar, if_sr, and network device
drivers from the old (legacy) USB stack can no longer be built or
used.
20090313:
POSIX.1 Native Language Support (NLS) has been enabled in libc and
a bunch of new language catalog files have also been added.
This means that some common libc messages are now localized and
they depend on the LC_MESSAGES environmental variable.
20090313:
The k8temp(4) driver has been renamed to amdtemp(4) since
support for Family 10 and Family 11 CPU families was added.
20090309:
IGMPv3 and Source-Specific Multicast (SSM) have been merged
to the IPv4 stack. VIMAGE hooks are in but not yet used.
For kernel developers, the most important changes are that the
ip_output() and ip_input() paths no longer take the IN_MULTI_LOCK(),
and this lock has been downgraded to a non-recursive mutex.
Transport protocols (UDP, Raw IP) are now responsible for filtering
inbound multicast traffic according to group membership and source
filters. The imo_multicast_filter() KPI exists for this purpose.
Transports which do not use multicast (SCTP, TCP) already reject
multicast by default. Forwarding and receive performance may improve
as a mutex acquisition is no longer needed in the ip_input()
low-level input path. in_addmulti() and in_delmulti() are shimmed
to new KPIs which exist to support SSM in-kernel.
For application developers, it is recommended that loopback of
multicast datagrams be disabled for best performance, as this
will still cause the lock to be taken for each looped-back
datagram transmission. The net.inet.ip.mcast.loop sysctl may
be tuned to 0 to disable loopback by default; it defaults to 1
to preserve the existing behaviour.
For systems administrators, to obtain best performance with
multicast reception and multiple groups, it is always recommended
that a card with a suitably precise hash filter is used. Hash
collisions will still result in the lock being taken within the
transport protocol input path to check group membership.
If deploying FreeBSD in an environment with IGMP snooping switches,
it is recommended that the net.inet.igmp.sendlocal sysctl remain
enabled; this forces 224.0.0.0/24 group membership to be announced
via IGMP.
The size of 'struct igmpstat' has changed; netstat needs to be
recompiled to reflect this.
Bump __FreeBSD_version to 800070.
20090309:
libusb20.so.1 is now installed as libusb.so.1 and the ports system
updated to use it. This requires a buildworld/installworld in order to
update the library and dependencies (usbconfig, etc). Its advisable to
rebuild all ports which uses libusb. More specific directions are given
in the ports collection UPDATING file. Any /etc/libmap.conf entries for
2009-03-09 22:43:00 +00:00
libusb are no longer required and can be removed.
20090302:
A workaround is committed to allow the creation of System V shared
memory segment of size > 2 GB on the 64-bit architectures.
Due to a limitation of the existing ABI, the shm_segsz member
of the struct shmid_ds, returned by shmctl(IPC_STAT) call is
2009-05-23 09:24:07 +00:00
wrong for large segments. Note that limits must be explicitly
raised to allow such segments to be created.
20090301:
The layout of struct ifnet has changed, requiring a rebuild of all
network device driver modules.
2009-02-27 17:32:49 +00:00
20090227:
The /dev handling for the new USB stack has changed, a
buildworld/installworld is required for libusb20.
20090223:
The new USB2 stack has now been permanently moved in and all kernel and
module names reverted to their previous values (eg, usb, ehci, ohci,
2009-02-23 19:28:29 +00:00
ums, ...). The old usb stack can be compiled in by prefixing the name
with the letter 'o', the old usb modules have been removed.
Updating entry 20090216 for xorg and 20090215 for libmap may still
apply.
20090217:
The rc.conf(5) option if_up_delay has been renamed to
defaultroute_delay to better reflect its purpose. If you have
customized this setting in /etc/rc.conf you need to update it to
use the new name.
20090216:
xorg 7.4 wants to configure its input devices via hald which does not
yet work with USB2. If the keyboard/mouse does not work in xorg then
add
Option "AllowEmptyInput" "off"
to your ServerLayout section. This will cause X to use the configured
2009-02-17 10:49:36 +00:00
kbd and mouse sections from your xorg.conf.
20090215:
The GENERIC kernels for all architectures now default to the new USB2
stack. No kernel config options or code have been removed so if a
problem arises please report it and optionally revert to the old USB
stack. If you are loading USB kernel modules or have a custom kernel
that includes GENERIC then ensure that usb names are also changed over,
2009-02-17 10:49:36 +00:00
eg uftdi -> usb2_serial_ftdi.
Older programs linked against the ports libusb 0.1 need to be
redirected to the new stack's libusb20. /etc/libmap.conf can
be used for this:
# Map old usb library to new one for usb2 stack
libusb-0.1.so.8 libusb20.so.1
20090209:
All USB ethernet devices now attach as interfaces under the name ueN
(eg. ue0). This is to provide a predictable name as vendors often
change usb chipsets in a product without notice.
20090203:
The ichsmb(4) driver has been changed to require SMBus slave
addresses be left-justified (xxxxxxx0b) rather than right-justified.
All of the other SMBus controller drivers require left-justified
slave addresses, so this change makes all the drivers provide the
same interface.
20090201:
INET6 statistics (struct ip6stat) was updated.
netstat(1) needs to be recompiled.
20090119:
NTFS has been removed from GENERIC kernel on amd64 to match
GENERIC on i386. Should not cause any issues since mount_ntfs(8)
will load ntfs.ko module automatically when NTFS support is
actually needed, unless ntfs.ko is not installed or security
level prohibits loading kernel modules. If either is the case,
"options NTFS" has to be added into kernel config.
20090115:
TCP Appropriate Byte Counting (RFC 3465) support added to kernel.
New field in struct tcpcb breaks ABI, so bump __FreeBSD_version to
800061. User space tools that rely on the size of struct tcpcb in
tcp_var.h (e.g. sockstat) need to be recompiled.
20081225:
ng_tty(4) module updated to match the new TTY subsystem.
Due to API change, user-level applications must be updated.
New API support added to mpd5 CVS and expected to be present
in next mpd5.3 release.
20081219:
2008-12-19 23:12:14 +00:00
With __FreeBSD_version 800060 the makefs tool is part of
the base system (it was a port).
20081216:
2009-02-17 10:50:18 +00:00
The afdata and ifnet locks have been changed from mutexes to
rwlocks, network modules will need to be re-compiled.
20081214:
__FreeBSD_version 800059 incorporates the new arp-v2 rewrite.
RTF_CLONING, RTF_LLINFO and RTF_WASCLONED flags are eliminated.
2009-02-17 10:50:18 +00:00
The new code reduced struct rtentry{} by 16 bytes on 32-bit
architecture and 40 bytes on 64-bit architecture. The userland
applications "arp" and "ndp" have been updated accordingly.
The output from "netstat -r" shows only routing entries and
none of the L2 information.
20081130:
__FreeBSD_version 800057 marks the switchover from the
binary ath hal to source code. Users must add the line:
2008-12-01 23:09:58 +00:00
options AH_SUPPORT_AR5416
to their kernel config files when specifying:
device ath_hal
The ath_hal module no longer exists; the code is now compiled
together with the driver in the ath module. It is now
possible to tailor chip support (i.e. reduce the set of chips
and thereby the code size); consult ath_hal(4) for details.
20081121:
__FreeBSD_version 800054 adds memory barriers to
<machine/atomic.h>, new interfaces to ifnet to facilitate
multiple hardware transmit queues for cards that support
them, and a lock-less ring-buffer implementation to
enable drivers to more efficiently manage queueing of
packets.
20081117:
A new version of ZFS (version 13) has been merged to -HEAD.
This version has zpool attribute "listsnapshots" off by
default, which means "zfs list" does not show snapshots,
and is the same as Solaris behavior.
20081028:
dummynet(4) ABI has changed. ipfw(8) needs to be recompiled.
20081009:
The uhci, ohci, ehci and slhci USB Host controller drivers have
been put into separate modules. If you load the usb module
separately through loader.conf you will need to load the
appropriate *hci module as well. E.g. for a UHCI-based USB 2.0
controller add the following to loader.conf:
uhci_load="YES"
ehci_load="YES"
20081009:
The ABI used by the PMC toolset has changed. Please keep
userland (libpmc(3)) and the kernel module (hwpmc(4)) in
sync.
20081009:
atapci kernel module now includes only generic PCI ATA
driver. AHCI driver moved to ataahci kernel module.
All vendor-specific code moved into separate kernel modules:
ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek,
atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron,
atamarvell, atamicron, atanational, atanetcell, atanvidia,
atapromise, ataserverworks, atasiliconimage, atasis, atavia
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
20080820:
The TTY subsystem of the kernel has been replaced by a new
implementation, which provides better scalability and an
improved driver model. Most common drivers have been migrated to
the new TTY subsystem, while others have not. The following
drivers have not yet been ported to the new TTY layer:
PCI/ISA:
cy, digi, rc, rp, sio
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
USB:
ubser, ucycom
Integrate the new MPSAFE TTY layer to the FreeBSD operating system. The last half year I've been working on a replacement TTY layer for the FreeBSD kernel. The new TTY layer was designed to improve the following: - Improved driver model: The old TTY layer has a driver model that is not abstract enough to make it friendly to use. A good example is the output path, where the device drivers directly access the output buffers. This means that an in-kernel PPP implementation must always convert network buffers into TTY buffers. If a PPP implementation would be built on top of the new TTY layer (still needs a hooks layer, though), it would allow the PPP implementation to directly hand the data to the TTY driver. - Improved hotplugging: With the old TTY layer, it isn't entirely safe to destroy TTY's from the system. This implementation has a two-step destructing design, where the driver first abandons the TTY. After all threads have left the TTY, the TTY layer calls a routine in the driver, which can be used to free resources (unit numbers, etc). The pts(4) driver also implements this feature, which means posix_openpt() will now return PTY's that are created on the fly. - Improved performance: One of the major improvements is the per-TTY mutex, which is expected to improve scalability when compared to the old Giant locking. Another change is the unbuffered copying to userspace, which is both used on TTY device nodes and PTY masters. Upgrading should be quite straightforward. Unlike previous versions, existing kernel configuration files do not need to be changed, except when they reference device drivers that are listed in UPDATING. Obtained from: //depot/projects/mpsafetty/... Approved by: philip (ex-mentor) Discussed: on the lists, at BSDCan, at the DevSummit Sponsored by: Snow B.V., the Netherlands dcons(4) fixed by: kan
2008-08-20 08:31:58 +00:00
Line disciplines:
ng_h4, ng_tty, ppp, sl, snp
Adding these drivers to your kernel configuration file shall
cause compilation to fail.
2008-09-03 08:30:17 +00:00
20080818:
ntpd has been upgraded to 4.2.4p5.
20080801:
OpenSSH has been upgraded to 5.1p1.
For many years, FreeBSD's version of OpenSSH preferred DSA
over RSA for host and user authentication keys. With this
upgrade, we've switched to the vendor's default of RSA over
DSA. This may cause upgraded clients to warn about unknown
host keys even for previously known hosts. Users should
follow the usual procedure for verifying host keys before
accepting the RSA key.
This can be circumvented by setting the "HostKeyAlgorithms"
option to "ssh-dss,ssh-rsa" in ~/.ssh/config or on the ssh
command line.
Please note that the sequence of keys offered for
authentication has been changed as well. You may want to
specify IdentityFile in a different order to revert this
behavior.
20080713:
The sio(4) driver has been removed from the i386 and amd64
kernel configuration files. This means uart(4) is now the
default serial port driver on those platforms as well.
2008-07-25 09:30:53 +00:00
To prevent collisions with the sio(4) driver, the uart(4) driver
uses different names for its device nodes. This means the
onboard serial port will now most likely be called "ttyu0"
instead of "ttyd0". You may need to reconfigure applications to
use the new device names.
When using the serial port as a boot console, be sure to update
/boot/device.hints and /etc/ttys before booting the new kernel.
If you forget to do so, you can still manually specify the hints
at the loader prompt:
set hint.uart.0.at="isa"
set hint.uart.0.port="0x3F8"
set hint.uart.0.flags="0x10"
set hint.uart.0.irq="4"
boot -s
2008-06-09 21:33:57 +00:00
20080609:
The gpt(8) utility has been removed. Use gpart(8) to partition
disks instead.
20080603:
The version that Linuxulator emulates was changed from 2.4.2
to 2.6.16. If you experience any problems with Linux binaries
2008-07-25 09:30:53 +00:00
please try to set sysctl compat.linux.osrelease to 2.4.2 and
if it fixes the problem contact emulation mailing list.
20080525:
ISDN4BSD (I4B) was removed from the src tree. You may need to
update a your kernel configuration and remove relevant entries.
20080509:
I have checked in code to support multiple routing tables.
2008-07-25 09:13:18 +00:00
See the man pages setfib(1) and setfib(2).
This is a hopefully backwards compatible version,
but to make use of it you need to compile your kernel
with options ROUTETABLES=2 (or more up to 16).
2008-04-20 21:25:37 +00:00
20080420:
The 802.11 wireless support was redone to enable multi-bss
operation on devices that are capable. The underlying device
is no longer used directly but instead wlanX devices are
cloned with ifconfig. This requires changes to rc.conf files.
For example, change:
ifconfig_ath0="WPA DHCP"
to
wlans_ath0=wlan0
ifconfig_wlan0="WPA DHCP"
2008-04-29 19:55:18 +00:00
see rc.conf(5) for more details. In addition, mergemaster of
/etc/rc.d is highly recommended. Simultaneous update of userland
and kernel wouldn't hurt either.
2008-04-20 21:25:37 +00:00
As part of the multi-bss changes the wlan_scan_ap and wlan_scan_sta
modules were merged into the base wlan module. All references
to these modules (e.g. in kernel config files) must be removed.
20080408:
psm(4) has gained write(2) support in native operation level.
Arbitrary commands can be written to /dev/psm%d and status can
be read back from it. Therefore, an application is responsible
for status validation and error recovery. It is a no-op in
other operation levels.
20080312:
Support for KSE threading has been removed from the kernel. To
run legacy applications linked against KSE libmap.conf may
be used. The following libmap.conf may be used to ensure
compatibility with any prior release:
libpthread.so.1 libthr.so.1
libpthread.so.2 libthr.so.2
libkse.so.3 libthr.so.3
20080301:
The layout of struct vmspace has changed. This affects libkvm
and any executables that link against libkvm and use the
kvm_getprocs() function. In particular, but not exclusively,
it affects ps(1), fstat(1), pkill(1), systat(1), top(1) and w(1).
The effects are minimal, but it's advisable to upgrade world
nonetheless.
20080229:
The latest em driver no longer has support in it for the
82575 adapter, this is now moved to the igb driver. The
split was done to make new features that are incompatible
with older hardware easier to do.
20080220:
The new geom_lvm(4) geom class has been renamed to geom_linux_lvm(4),
likewise the kernel option is now GEOM_LINUX_LVM.
20080211:
The default NFS mount mode has changed from UDP to TCP for
increased reliability. If you rely on (insecurely) NFS
mounting across a firewall you may need to update your
firewall rules.
2008-02-08 21:24:58 +00:00
20080208:
Belatedly note the addition of m_collapse for compacting
mbuf chains.
Our fts(3) API, as inherited from 4.4BSD, suffers from integer fields in FTS and FTSENT structs being too narrow. In addition, the narrow types creep from there into fts.c. As a result, fts(3) consumers, e.g., find(1) or rm(1), can't handle file trees an ordinary user can create, which can have security implications. To fix the historic implementation of fts(3), OpenBSD and NetBSD have already changed <fts.h> in somewhat incompatible ways, so we are free to do so, too. This change is a superset of changes from the other BSDs with a few more improvements. It doesn't touch fts(3) functionality; it just extends integer types used by it to match modern reality and the C standard. Here are its points: o For C object sizes, use size_t unless it's 100% certain that the object will be really small. (Note that fts(3) can construct pathnames _much_ longer than PATH_MAX for its consumers.) o Avoid the short types because on modern platforms using them results in larger and slower code. Change shorts to ints as follows: - For variables than count simple, limited things like states, use plain vanilla `int' as it's the type of choice in C. - For a limited number of bit flags use `unsigned' because signed bit-wise operations are implementation-defined, i.e., unportable, in C. o For things that should be at least 64 bits wide, use long long and not int64_t, as the latter is an optional type. See FTSENT.fts_number aka FTS.fts_bignum. Extending fts_number `to satisfy future needs' is pointless because there is fts_pointer, which can be used to link to arbitrary data from an FTSENT. However, there already are fts(3) consumers that require fts_number, or fts_bignum, have at least 64 bits in it, so we must allow for them. o For the tree depth, use `long'. This is a trade-off between making this field too wide and allowing for 64-bit inode numbers and/or chain-mounted filesystems. On the one hand, `long' is almost enough for 32-bit filesystems on a 32-bit platform (our ino_t is uint32_t now). On the other hand, platforms with a 64-bit (or wider) `long' will be ready for 64-bit inode numbers, as well as for several 32-bit filesystems mounted one under another. Note that fts_level has to be signed because -1 is a magic value for it, FTS_ROOTPARENTLEVEL. o For the `nlinks' local var in fts_build(), use `long'. The logic in fts_build() requires that `nlinks' be signed, but our nlink_t currently is uint16_t. Therefore let's make the signed var wide enough to be able to represent 2^16-1 in pure C99, and even 2^32-1 on a 64-bit platform. Perhaps the logic should be changed just to use nlink_t, but it can be done later w/o breaking fts(3) ABI any more because `nlinks' is just a local var. This commit also inludes supporting stuff for the fts change: o Preserve the old versions of fts(3) functions through libc symbol versioning because the old versions appeared in all our former releases. o Bump __FreeBSD_version just in case. There is a small chance that some ill-written 3-rd party apps may fail to build or work correctly if compiled after this change. o Update the fts(3) manpage accordingly. In particular, remove references to fts_bignum, which was a FreeBSD-specific hack to work around the too narrow types of FTSENT members. Now fts_number is at least 64 bits wide (long long) and fts_bignum is an undocumented alias for fts_number kept around for compatibility reasons. According to Google Code Search, the only big consumers of fts_bignum are in our own source tree, so they can be fixed easily to use fts_number. o Mention the change in src/UPDATING. PR: bin/104458 Approved by: re (quite a while ago) Discussed with: deischen (the symbol versioning part) Reviewed by: -arch (mostly silence); das (generally OK, but we didn't agree on some types used; assuming that no objections on -arch let me to stick to my opinion)
2008-01-26 17:09:40 +00:00
20080126:
The fts(3) structures have been changed to use adequate
integer types for their members and so to be able to cope
with huge file trees. The old fts(3) ABI is preserved
through symbol versioning in libc, so third-party binaries
using fts(3) should still work, although they will not take
advantage of the extended types. At the same time, some
third-party software might fail to build after this change
due to unportable assumptions made in its source code about
fts(3) structure members. Such software should be fixed
by its vendor or, in the worst case, in the ports tree.
FreeBSD_version 800015 marks this change for the unlikely
case that a portable fix is impossible.
20080123:
To upgrade to -current after this date, you must be running
FreeBSD not older than 6.0-RELEASE. Upgrading to -current
from 5.x now requires a stop over at RELENG_6 or RELENG_7 systems.
20071128:
The ADAPTIVE_GIANT kernel option has been retired because its
functionality is the default now.
20071118:
The AT keyboard emulation of sunkbd(4) has been turned on
by default. In order to make the special symbols of the Sun
keyboards driven by sunkbd(4) work under X these now have
to be configured the same way as Sun USB keyboards driven
by ukbd(4) (which also does AT keyboard emulation), f.e.:
Option "XkbLayout" "us"
Option "XkbRules" "xorg"
Option "XkbSymbols" "pc(pc105)+sun_vndr/usb(sun_usb)+us"
20071024:
It has been decided that it is desirable to provide ABI
backwards compatibility to the FreeBSD 4/5/6 versions of the
PCIOCGETCONF, PCIOCREAD and PCIOCWRITE IOCTLs, which was
broken with the introduction of PCI domain support (see the
20070930 entry). Unfortunately, this required the ABI of
PCIOCGETCONF to be broken again in order to be able to
provide backwards compatibility to the old version of that
IOCTL. Thus consumers of PCIOCGETCONF have to be recompiled
again. As for prominent ports this affects neither pciutils
nor xorg-server this time, the hal port needs to be rebuilt
however.
20071020:
The misnamed kthread_create() and friends have been renamed
to kproc_create() etc. Many of the callers already
used kproc_start()..
I will return kthread_create() and friends in a while
with implementations that actually create threads, not procs.
Renaming corresponds with version 800002.
20071010:
RELENG_7 branched.
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 before reporting problems with a major version upgrade.
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.
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 ad0:
"gpart bootcode -p /boot/gptzfsboot -i 1 ad0"
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 just build a kernel when you know that it won't mess you up
--------------------------------------------------------------
This assumes you are already running a CURRENT system. Replace
${arch} with the architecture of your machine (e.g. "i386",
"arm", "amd64", "ia64", "pc98", "sparc64", "powerpc", "mips", etc).
cd src/sys/${arch}/conf
config KERNEL_NAME_HERE
cd ../compile/KERNEL_NAME_HERE
make depend
make
2004-06-22 10:33:58 +00:00
make install
If this fails, go to the "To build a kernel" section.
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 kernel 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 kernel KERNCONF=YOUR_KERNEL_HERE [8]
[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.
[5] Usually this step is a noop. 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] When checking out sources, you must include the -P flag to have
cvs prune empty directories.
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 October 10, 2007.
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
$FreeBSD$