julian [Fri, 21 Aug 2009 11:23:29 +0000 (11:23 +0000)]
MFC r196423
Fix ipfw's initialization functions to get the correct order of evaluation
to allow vnet and non vnet operation. Move some functions from ip_fw_pfil.c
to ip_fw2.c and mode to mostly using the SYSINIT and VNET_SYSINIT handlers
instead of the modevent handler. Correct some spelling errors in comments
in the affected code. Note this bug fixes a crash in NON VIMAGE kernels when
ipfw is unloaded.
This patch is a minimal patch for 8.0
I have a much larger patch that actually fixes the underlying problems
that will be applied after 8.0
Reviewed by: zec@, rwatson@, bz@(earlier version)
Approved by: re (rwatson)
zec [Fri, 21 Aug 2009 11:17:25 +0000 (11:17 +0000)]
MFC r196421:
Bugfix: all requests for creating vnets via vimage -c were always
reported as failures, even if the actual library / system call
would succeed, because error message would be reported if the return
value from jail_setv() call was >= 0, and if not, then if that same
value was < 0, i.e. always. The correct behavior is to abort (only)
if jail_setv() returns < 0.
jhb [Fri, 21 Aug 2009 03:14:39 +0000 (03:14 +0000)]
MFC 196417:
This patch fixes two bugs in sglist(9) and improves robustness of the API via
better semantics if a request to append an address range to an existing list
fails.
- When cloning an sglist, properly set the length in the new sglist instead of
leaving the new list empty.
- Properly compute the amount of data added to an sglist via
_sglist_append_buf(). This allows sglist_consume_uio() to properly update
uio_resid.
- When a request to append an address range to a scatter/gather list fails,
restore the sglist to the state it had at the start of the function call
instead of resetting it to an empty list.
kensmith [Fri, 21 Aug 2009 01:12:06 +0000 (01:12 +0000)]
MFC r196415:
Fix a boot hang for hptrr(4) caused by changes introduced in r195534.
It is necessary to make sure cpi->transport is set for xpt_scan_bus() to
work properly.
Submitted by: Bernhard Schmidt (scb+freebsd-current <at> techwires
<dot> net)
Reviewed by: scottl
Approved by: re (kib)
jkim [Thu, 20 Aug 2009 23:04:21 +0000 (23:04 +0000)]
MFC: r196412
Check whether the SMBIOS reports reasonable amount of memory. If it is
less than "avail memory", fall back to Maxmem to avoid user confusion.
We use SMBIOS information to display "real memory" since r190599 but
some broken SMBIOS implementation reported only half of actual memory.
zec [Thu, 20 Aug 2009 22:56:29 +0000 (22:56 +0000)]
MFC r196409:
vimage(8) is a legacy CLI interface for managing jails associated with
network stack instances, which is provided for compatibility with
older applications. This change brings it back to life in a followup
to the initial conversion of vimage to use the new jail(4)
userland-kernel API:
- when creating vimages via "vimage -c", by default turn on a few
options expected by legacy applications, such as allow operations on
raw sockets, FS mounts etc, and allow jail-related parameters to be
optionally configured.
- introduce the "-m" modifier which allows for configuring jail
parameters of existing vimages / vnet-jails.
- make "vimage name command ..." actually work.
- when reassigning ifnets to vnets using "vimage -i", attempt to rename
the ifnet as "ethXXX" on arrival in the target vnet. Several legacy
applications are known to depend heavily on such behavior.
- vimage -l lists only jails associated with vnets. The output is
sorted using vimage / jail names as keys.
- vimage -l by default searches only the current level in the jail
hierarchy. Recursive listing can be requested via -r switch.
- vimage -l by default prints only jail names on each line, making
such output suitable for pipelining to other commands. More verbose
output can be obtained via -v switch, and even more jail specific
information will be displayed if -j switch is turned on.
- there's no need to build vimage as statically linked, so update the
Makefile accordingly.
rwatson [Thu, 20 Aug 2009 21:29:49 +0000 (21:29 +0000)]
Merge r196267 from head to stable/8:
Rather than fix questionable ifnet list locking in the implementation of
the kern.polling.enable sysctl, remove the sysctl. It has been deprecated
since FreeBSD 6 in favour of per-ifnet polling flags.
jhb [Thu, 20 Aug 2009 20:53:36 +0000 (20:53 +0000)]
MFC 196404:
Change the 'resid' parameter to sglist_consume_uio() from an int to a
size_t to match the recent type change of the uio_resid member of struct
uio.
avg [Thu, 20 Aug 2009 11:04:31 +0000 (11:04 +0000)]
MFC 196399: fstat: fix fsid comparison when executed on systems with 64-bit long
This affects only fstat on zfs and devfs, only on 64-bit systems
and only when fsid is greater than 2^31 - 1.
When fstat examines a file via stat(2) it takes uint32_t st_dev
and assigns to (signed) (64-bit) long fsid, this results in
a positive value.
When fstat examines opened files it takes int32_t f_fsid.val[0]
and assigns to (signed) (64-bit) long fsid, this results in
a negative value.
So, while initially st_dev and f_fsid.val[0] have the same bit
values they get promoted to different 64-bit values because
of the signed-vs-unsigned difference.
A fix is to use "more natural" positive numbers by introducing
intermediate unsigned cast for f_fsid.val[0].
will [Thu, 20 Aug 2009 02:49:43 +0000 (02:49 +0000)]
MFC r196397 from head:
Fix CARP memory leaks on carp_if's malloc'd using M_CARP. This occurs when
CARP tries to free them using M_IFADDR after the last address for a virtual
host is removed and when detaching from the parent interface.
pjd [Thu, 20 Aug 2009 00:08:58 +0000 (00:08 +0000)]
MFC r196395:
Our libc doesn't implement control method for XDR (only kernel does) and it
will always return failure. Fix this by bringing userland implementation of
xdrmem_control() back. This allow 'zpool import' to work again.
Reported by: Thomas Backman <serenity@exscape.org>
Reviewed by: kmacy
Approved by: re (kib)
simon [Wed, 19 Aug 2009 21:01:32 +0000 (21:01 +0000)]
MFC 196392:
Add support for backing up the old kernel when installing a new kernel
using freebsd-update. This applies to using freebsd-update in "upgrade
mode" and normal freebsd-update on a security branch.
The backup kernel will be written to /boot/kernel.old, if the directory
does not exist, or the directory was created by freebsd-update in a
previous backup. Otherwise freebsd-update will generate a new directory
name for use by the backup. By default symbol files are not backed up
to save diskspace and avoid filling up the root partition.
This feature is fully configurable in the freebsd-update config file,
but defaults to enabled.
kmacy [Wed, 19 Aug 2009 20:17:36 +0000 (20:17 +0000)]
This change fixes a comment and addresses a complaint by kib@ by
moving a frequently executed flowtable syslog statement from being
conditional on bootverbose to conditional on a per-vnet flowtable
sysctl.
delphij [Wed, 19 Aug 2009 18:08:50 +0000 (18:08 +0000)]
MFC r196386:
Temporarily enhance em(4) and igb(4) hack to take account for IFF_NOARP.
Without this changeset there will be no way to prevent these NICs from
sending ARP, which is harmful in server farms that is configured as
"Direct Server Return" behind a load balancer.
A better fix would remove the whole hack completely but it would be
later than 8.0-RELEASE.
jhb [Wed, 19 Aug 2009 17:45:58 +0000 (17:45 +0000)]
MFC 196382:
Explicitly line up the CPU state labels with the calculated starting column
that takes into account the width of the largest CPU ID. On systems with
> 10 CPUs the labels for the first 10 CPUs were not lined up properly
otherwise.
marcel [Wed, 19 Aug 2009 16:40:45 +0000 (16:40 +0000)]
MFC change 196383:
Remove the dependency on the kernel -- in particular the gctl request to
the GEOM_BSD class -- to translate the absolute offsets in the label to
relative ones. This makes bslabel(8) work correctly with GEOM_PART and
also when the BSD label is nested under arbitrary partitioning schemes.
Inspired by: Eygene Ryabinkin <rea-fbsd@codelabs.ru>
Approved by: re (kib)
raj [Wed, 19 Aug 2009 14:48:59 +0000 (14:48 +0000)]
MFC r196380
Fix USB cache sync operations for platforms with non-coherent DMA.
- usb_pc_cpu_invalidate() is called between [consecutive] reads from a device,
so a sequence of BUS_DMASYNC_POSTREAD and _PREREAD should be used. Note we
cannot use or'ed shorthand ( _POSTREAD | _PREREAD) for BUS_DMASYNC flags, as
the low level bus dma sync operation is implementation dependent and we
cannot assume the required order of operations to be guaranteed.
- usb_pc_cpu_flush() is called before writing to a device, so
BUS_DMASYNC_PREWRITE should be used.
Submitted by: Grzegorz Bernacki
Reviewed by: HPS, arm@, usb@ ML
Tested by: HPS, Mike Tancsa
Approved by: re (kib)
Obtained from: Semihalf
ed [Wed, 19 Aug 2009 14:38:43 +0000 (14:38 +0000)]
MFC r196378:
Small changes to the warning message generated by pty(4):
- Only print the warning once, instead of filling up the screen.
- Use the word "legacy" for the pty_warningcnt description, to prevent
confusion.
- Use log() instead of printf().
Discussed with: rwatson, jhb
Approved by: re (kib)
edwin [Wed, 19 Aug 2009 03:58:12 +0000 (03:58 +0000)]
- Make note of the update of tzdata from 2008h to 2009j
- Make note of the update of tzcode from 2004a to 2009h
Add an extra alert that people who update via source or via
freebsd-update will have to run the tzsetup(8) utility.
mlaier [Wed, 19 Aug 2009 00:17:00 +0000 (00:17 +0000)]
MFC r196372:
If we cannot immediately get the pf_consistency_lock in the purge thread,
restart the scan after acquiring the lock the hard way. Otherwise we
might end up with a dead reference.
stas [Tue, 18 Aug 2009 21:13:00 +0000 (21:13 +0000)]
- MFC r196370.
Do not try to reevaluate current RX production index on each
loop iteration as it can be updated by the card while we
process the RX ring forcing us to process RX descriptors
for which DMA synchronisation operation has not been
performed. This fixes the bug when bge(4) drops packets
under high load.
Discussed with: yongari, marius
Approved by: re (kib)
kmacy [Tue, 18 Aug 2009 20:39:35 +0000 (20:39 +0000)]
MFC 196368
- change the interface to flowtable_lookup so that we don't rely on
the mbuf for obtaining the fib index
- check that a cached flow corresponds to the same fib index as the
packet for which we are doing the lookup
- at interface detach time flush any flows referencing stale rtentrys
associated with the interface that is going away (fixes reported
panics)
- reduce the time between cleans in case the cleaner is running at
the time the eventhandler is called and the wakeup is missed less
time will elapse before the eventhandler returns
- separate per-vnet initialization from global initialization
(pointed out by jeli@)
yongari [Tue, 18 Aug 2009 20:25:02 +0000 (20:25 +0000)]
MFC r196366:
Backout r193289. r193289 restored page select bits to previous
value instead of blindly resetting it to 0. However, it seems page
select bits of some 88E1116 PHY is initialized to invalid one such
that restoring page select bits after programming broke MII
register access. The correct solution would be reset page select
bits to 0 in PHY attach stage but it would require more testing.
Since we're in BETA stage such a change would be dangerous so just
back it out.
This change should fix nfe(4) breakage on NVIDIA MCP55.
Reported by: Ryan Rogers < webmaster <> doghouserepair dot com >
Sam Fourman Jr. < sfourman <> gmail dot com >
Tested by: Ryan Rogers < webmaster <> doghouserepair dot com >
Sam Fourman Jr. < sfourman <> gmail dot com >
Approved by: re (kib)
mav [Tue, 18 Aug 2009 09:36:25 +0000 (09:36 +0000)]
Fix copy/paste bug, that requests data read during ATA device probe sequence
for ATA_SETFEATURES/ATA_SF_SETXFER command which by definition transfers no
data. Most of controllers are irrelevant to this bug, but some nVidia's
doesn't.
attilio [Mon, 17 Aug 2009 16:33:53 +0000 (16:33 +0000)]
MFC r196334:
* Change the scope of the ASSERT_ATOMIC_LOAD() from a generic check to
a pointer-fetching specific operation check. Consequently, rename the
operation ASSERT_ATOMIC_LOAD_PTR().
* Fix the implementation of ASSERT_ATOMIC_LOAD_PTR() by checking
directly alignment on the word boundry, for all the given specific
architectures. That's a bit too strict for some common case, but it
assures safety.
* Add a comment explaining the scope of the macro
* Add a new stub in the lockmgr specific implementation
Tested by: marcel (initial version), marius
Reviewed by: rwatson, jhb (comment specific review)
Approved by: re (kib)
marcel [Mon, 17 Aug 2009 16:24:50 +0000 (16:24 +0000)]
MFC rev 196333:
The start of the EFI GPT partition in the PMBR can always be represented
by CHS addressing. Don't define these fields as 0xff, but rather define
them correctly. This prevents boot problems on PCs where GPT is being
used.
PR: 115406
Submitted by: Kent Hauser <kent@khauser.net>
Approved by: re (kib)
jhb [Mon, 17 Aug 2009 15:42:55 +0000 (15:42 +0000)]
- Remove self-referential mergeinfo from xen/netfront and xen/xenpci that
claims those directories were merged into themselves.
- Remove mergeinfo on xen/xenpci that claims the stable/7 xenpci was merged
into head.
jhay [Mon, 17 Aug 2009 15:39:47 +0000 (15:39 +0000)]
MFC: 196326
Fix parse() so that the partition to boot (load /boot/loader) from can
be set. The syntax as printed in main() is used: 0:ad(0p3)/boot/loader
pjd [Mon, 17 Aug 2009 10:02:31 +0000 (10:02 +0000)]
MFC r196309:
getcwd() (when __getcwd() fails) works by stating current directory, going up
(..), calling readdir and looking for previous directory inode. In case of
.zfs/ directory this doesn't work, because .zfs/ is hidden by default, so it
won't be visible in readdir output.
Fix this by implementing VPTOCNP for snapshot directories, so __getcwd()
doesn't fail and getcwd() doesn't have to use readdir method.
This fixes /bin/pwd from within .zfs/snapshot/<name>/.
pjd [Mon, 17 Aug 2009 09:03:47 +0000 (09:03 +0000)]
MFC r196295:
Remove OpenSolaris taskq port (it performs very poorly in our kernel) and
replace it with wrappers around our taskqueue(9).
To make it possible implement taskqueue_member() function which returns 1
if the given thread was created by the given taskqueue.
pjd [Mon, 17 Aug 2009 08:46:47 +0000 (08:46 +0000)]
MFC r196293:
Because taskqueue_run() can drop tq_mutex, we need to check if the
TQ_FLAGS_ACTIVE flag wasn't removed in the meantime, which means we missed a
wakeup.
pjd [Mon, 17 Aug 2009 08:38:41 +0000 (08:38 +0000)]
MFC r196291:
- Fix a race where /dev/zfs control device is created before ZFS is fully
initialized. Also destroy /dev/zfs before doing other deinitializations.
- Initialization through taskq is no longer needed and there is a race
where one of the zpool/zfs command loads zfs.ko and tries to do some work
immediately, but /dev/zfs is not there yet.
marcel [Sun, 16 Aug 2009 21:32:12 +0000 (21:32 +0000)]
MFC change 196278:
Emit a proper error message instead of dumping core when 1)
GEOM_PART does not exist in the kernel, and 2) the GEOM in
question does not exist.
Additionally abort in case of programming errors that result
in neither the class nor geom not being present in the gctl
request.
rink [Sun, 16 Aug 2009 10:25:58 +0000 (10:25 +0000)]
MFC r196272
Prevent sysinstall from needlessly waiting for confirmation when using an USB
device in non-interactive mode.
If there are no USB devices, sysinstall gives an error messages, and if there
is >1, it'll ask which one is to be used. This change allows a non-interactive
install from USB media to succeed without any user interaction if there is
exactly one USB disk device in the system it can use.
Submitted by: Daniel O'Connor < doconnorat gsoft dot com dot au >
Reviewed by: randi
Approved by: re (rwatson)
tuexen [Sat, 15 Aug 2009 21:37:16 +0000 (21:37 +0000)]
MFC r196260.
* Fix a bug where PR-SCTP settings are ignore when using implicit
association setup.
* Fix a bug where message with illegal stream ids are not deleted.
* Fix a crash when reporting back unsent messages from the send_queue.
* Fix a bug related to INIT retransmission when the socket is already
closed.
* Fix a bug where associations were stalled when partial delivery API
was enabled.
* Fix a bug where the receive buffer size was smaller than the
partial_delivery_point.
keramida [Sat, 15 Aug 2009 18:03:34 +0000 (18:03 +0000)]
MFC 196254 - iostat: add a bit of space between tty in/out columns
The columns for tty input and output may bump against each other
if the tty output needs more than 5 columns. Add a bit of space
that pushes everything 1 column to the right, but also avoids the
problem.
stas [Sat, 15 Aug 2009 15:18:29 +0000 (15:18 +0000)]
- Merge r196246:
Proprely intialize UART parameters at probe stage, so uart(4)
will initialize the FIFO memory correctly on attach. Before
that this values was intialized in only in at91_usart_bus_attach
which is called after the uart(4) memory allocation happens.
stas [Sat, 15 Aug 2009 15:12:46 +0000 (15:12 +0000)]
- Merge r196244:
Avoid overflowing the swap size counters in human-readable mode
by introducing the new CONVERT_BLOCKS macro which operates on
sizes already converted to number of blocks. With this macro
it is not longer needed to perform needless multiplica
qingli [Sat, 15 Aug 2009 00:04:12 +0000 (00:04 +0000)]
MFC 196234
In function ip_output(), the cached route is flushed when there is a
mismatch between the cached entry and the intended destination. The
cached rtentry{} is flushed but the associated llentry{} is not. This
causes the wrong destination MAC address being used in the output
packets. The fix is to flush the llentry{} when rtentry{} is cleared.
zec [Fri, 14 Aug 2009 23:05:10 +0000 (23:05 +0000)]
MFC r196230:
Appease VNET_DEBUG - in if_vmove we temporarily switch i.e.
recurse from one vnet to another which is OK, so no need
to flood the console with warnings here.
bz [Fri, 14 Aug 2009 21:50:47 +0000 (21:50 +0000)]
MFC r196226:
Add a new macro to test that a variable could be loaded atomically.
Check that the given variable is at most uintptr_t in size and that
it is aligned.
Note: ASSERT_ATOMIC_LOAD() uses ALIGN() to check for adequate
alignment -- however, the function of ALIGN() is to guarantee
alignment, and therefore may lead to stronger alignment
enforcement than necessary for types that are smaller than
sizeof(uintptr_t).
Add checks to mtx, rw and sx locks init functions to detect possible
breakage. This was used during debugging of the problem fixed with
r196118 where a pointer was on an un-aligned address in the dpcpu area.
In collaboration with: rwatson
Reviewed by: rwatson
jhb [Fri, 14 Aug 2009 20:57:21 +0000 (20:57 +0000)]
Adjust the handling of the local APIC PMC interrupt vector:
- Provide lapic_disable_pmc(), lapic_enable_pmc(), and lapic_reenable_pmc()
routines in the local APIC code that the hwpmc(4) driver can use to
manage the local APIC PMC interrupt vector.
- Do not enable the local APIC PMC interrupt vector by default when
HWPMC_HOOKS is enabled. Instead, the hwpmc(4) driver explicitly
enables the interrupt when it is succesfully initialized and disables
the interrupt when it is unloaded. This avoids enabling the interrupt
on unsupported CPUs which may result in spurious NMIs.
jhb [Fri, 14 Aug 2009 20:42:40 +0000 (20:42 +0000)]
MFC 196221:
Add the ability to build a release from an SVN checkout instead of a CVS
checkout. If SVNROOT is specified, then the source tree will be checked
out from that SVN repository instead of using CVS. ports and docs still
use CVS. If SVNROOT is not specified, then the source tree will be checked
out using CVS. An explicit SVN branch can be specified using SVNBRANCH
(e.g. SVNBRANCH=stable/8). If SVNBRANCH is not set but RELEASETAG is set
to a CVS branch (such as RELENG_8) the appropriate SVN branch will be
inferred from the CVS branch using svnbranch.awk.
Note that there are still several open questions about using SVN instead
of CVS in the release process. However, this does enable one to build a
release from an SVN repository if needed.
kib [Fri, 14 Aug 2009 11:17:34 +0000 (11:17 +0000)]
MFC r196205:
In nfs_upgrade_vnlock(), assert that the vnode is locked.
When downgrading, pass LK_RETRY to the vn_lock(), since otherwise
vn_lock() unlocks the doomed vnode, causing extra unlock.
julian [Fri, 14 Aug 2009 10:25:14 +0000 (10:25 +0000)]
MFC of r196201
URL: http://svn.freebsd.org/changeset/base/196201
Fix ipfw crash on uid or gid check.
Receiving any ip packet for which there is no existing socket will
crash if ipfw has a uid or gid test rule, as the uid/gid
of the non existent owner of said non existent socket is tested.
Brooks introduced this error as part of his >16 gids patch.
It appears to be a cut-n-paste error from similar code a few lines
before. The old code used the 'pcb' variable here, but in the
new code that switched the 'inp' variable, which is often NULL
and what is tested in the code further up. The rest of the multi-gid
patch for ipfw seems solid (and cleaner than previous code).
p.s. What's up with all the properties changing? It is a fresh checkout.
attilio [Thu, 13 Aug 2009 17:54:11 +0000 (17:54 +0000)]
MFC r196196:
* Completely remove the option STOP_NMI from the kernel. This option
has proven to have a good effect when entering KDB by using a NMI,
but it completely violates all the good rules about interrupts
disabled while holding a spinlock in other occasions. This can be the
cause of deadlocks on events where a normal IPI_STOP is expected.
* Add an new IPI called IPI_STOP_HARD on all the supported architectures.
This IPI is responsible for sending a stop message among CPUs using a
privileged channel when disponible. In other cases it just does match a
normal IPI_STOP.
Right now the IPI_STOP_HARD functionality uses a NMI on ia32 and amd64
architectures, while on the other has a normal IPI_STOP effect. It is
responsibility of maintainers to eventually implement an hard stop
when necessary and possible.
* Use the new IPI facility in order to implement a new userend SMP kernel
function called stop_cpus_hard(). That is specular to stop_cpu() but
it does use the privileged channel for the stopping facility.
* Let KDB use the newly introduced function stop_cpus_hard() and leave
stop_cpus() for all the other cases
* Disable interrupts on CPU0 when starting the process of APs suspension.
* Style cleanup and comments adding
This patch should fix the reboot/shutdown deadlocks many users are
constantly reporting on mailing lists.
Please don't forget to update your config file with the STOP_NMI
option removal
bz [Thu, 13 Aug 2009 15:02:02 +0000 (15:02 +0000)]
MFC r196185:
Rather than replicating the maths from the kernel, use the
value the kernel calculated directly as we already read it
with struct vnet. This will make kvm_vnet.c more resilent
in case of possible kernel changes.
projects / FreeBSD / stable / 8.git / log