If you mount -o space_cache, the option will be persistent across mounts, but to
make sure the user knows that they did this, emit a message telling them if they
didn't mount with -o space_cache but the feature is still used.
Signed-off-by: Josef Bacik <josef@redhat.com>
If something goes wrong with the free space cache we need a way to make sure
it's not loaded on mount and that it's cleared for everybody. When you pass the
clear_cache option it will make it so all block groups are setup to be cleared,
which keeps them from being loaded and then they will be truncated when the
transaction is committed. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
There are just a few things that need to be fixed in the kernel to support mixed
data+metadata block groups. Mostly we just need to make sure that if we are
using mixed block groups that we continue to allocate mixed block groups as we
need them. Also we need to make sure __find_space_info will find our space info
if we search for DATA or METADATA only. Tested this with xfstests and it works
nicely. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
With the free space disk caching we can mark the block group as started with the
caching, but we don't have a caching ctl. This can race with anybody else who
tries to get the caching ctl before we cache (this is very hard to do btw). So
instead check to see if cache->caching_ctl is set, and if not return NULL.
Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
This patch actually loads the free space cache if it exists. The only thing
that really changes here is that we need to cache the block group if we're going
to remove an extent from it. Previously we did not do this since the caching
kthread would pick it up. With the on disk cache we don't have this luxury so
we need to make sure we read the on disk cache in first, and then remove the
extent, that way when the extent is unpinned the free space is added to the
block group. This has been tested with all sorts of things.
Signed-off-by: Josef Bacik <josef@redhat.com>
This is a simple bit, just dump the free space cache out to our preallocated
inode when we're writing out dirty block groups. There are a bunch of changes
in inode.c in order to account for special cases. Mostly when we're doing the
writeout we're holding trans_mutex, so we need to use the nolock transacation
functions. Also we can't do asynchronous completions since the async thread
could be blocked on already completed IO waiting for the transaction lock. This
has been tested with xfstests and btrfs filesystem balance, as well as my ENOSPC
tests. Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
flush_scheduled_work() is deprecated and scheduled to be removed.
sync_stop() currently cancels cpu_buffer works inside buffer_mutex and
flushes the system workqueue outside. Instead, split end_cpu_work()
into two parts - stopping further work enqueues and flushing works -
and do the former inside buffer_mutex and latter outside.
For stable kernels v2.6.35.y and v2.6.36.y.
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: stable@kernel.org
Signed-off-by: Robert Richter <robert.richter@amd.com>
The kernel build with CONFIG_OPROFILE and CPU_HOTPLUG enabled.
The oprofile is initialised using system timer in absence of hardware
counters supports. Oprofile isn't started from userland.
In this setup while doing a CPU offline the kernel hangs in infinite
for loop inside lock_hrtimer_base() function
This happens because as part of oprofile_cpu_notify(, it tries to
stop an hrtimer which was never started. These per-cpu hrtimers
are started when the oprfile is started.
echo 1 > /dev/oprofile/enable
This problem also existwhen the cpu is booted with maxcpus parameter
set. When bringing the remaining cpus online the timers are started
even if oprofile is not yet enabled.
This patch fix this issue by adding a state variable so that
these hrtimer start/stop is only attempted when oprofile is
started
For stable kernels v2.6.35.y and v2.6.36.y.
Reported-by: Jan Sebastien <s-jan@ti.com>
Tested-by: sricharan <r.sricharan@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Cc: stable@kernel.org
Signed-off-by: Robert Richter <robert.richter@amd.com>
a) switch ->put_device() to logfs_super *
b) actually call it on early failures in logfs_get_sb_device()
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
take setting s_bdev/s_mtd/s_devops to callers of logfs_get_sb_device(),
don't bother passing them separately
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
eventual replacement for ->get_sb() - does *not* get vfsmount,
return ERR_PTR(error) or root of subtree to be mounted.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
nameidata_to_filp() drops nd->path or transfers it to opened
file. In the former case it's a Bad Idea(tm) to do mnt_drop_write()
on nd->path.mnt, since we might race with umount and vfsmount in
question might be gone already.
Fix: don't drop it, then... IOW, have nameidata_to_filp() grab nd->path
in case it transfers it to file and do path_drop() in callers. After
they are through with accessing nd->path...
Reported-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Commit 22d4cd4c4d ("Allocate irq stacks seperate from percpu
area") removed NUMA affinity of IRQ stacks as side-effect of
the fix.
Using alloc_pages_node() instead of __get_free_pages() is safe,
even if the target node has no available LOWMEM pages :
alloc_pages_node() fallbacks to another node.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Acked-by: Brian Gerst <brgerst@gmail.com>
Cc: tj@kernel.org
Cc: torvalds@linux-foundation.org
Cc: Peter Zijlstra <peterz@infradead.org>
LKML-Reference: <1288276854.2649.607.camel@edumazet-laptop>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
The code for searching through the device list to read-balance in
raid1 is rather clumsy and hard to follow. Try to simplify it a bit.
No important functionality change here.
Signed-off-by: NeilBrown <neilb@suse.de>
When writing to an 'external' bitmap we don't currently unplug the
device before waiting, so we can get a 3msec delay each time;
So use REQ_UNPLUG to force and unplug.
Signed-off-by: NeilBrown <neilb@suse.de>
Enable Westmere support for all APIC modes on SGI UV.
Signed-off-by: Russ Anderson <rja@sgi.com>
LKML-Reference: <20101028224132.GB15804@sgi.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Considering following scenario - the touch is present on the screen
at the beginning of the last conversion sequence, but by the time
the last sequence is finished, the finger is lift off. The AD7879 data
available interrupt signals (DAV) completion, however some X,Y values
are not valid because the screen inputs were floating during the
acquisition.
The AD7877 acts differently here, since it only asserts DAV if the
touch is still present when the conversion sequence finished.
Based on the fact that this can only happen in the last sample of the
repeated conversion sequence, we simply skip the last (short glitches
are filtered by the AD7879 internal median and average filters).
This doesn't cause noticeable side effects, since the minimum conversion
interval is 9.44ms. We receive ~100 waypoint samples per second, so we
simply delay the result by 9.44ms.
We also reject samples where pressure is greater than pressure_max.
Signed-off-by: Michael Hennerich <michael.hennerich@analog.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Some menuconfigs in the Kconfig files have prompts and dependencies.
Currently, streamline_config misses these, and this can cause
streamline_config to keep modules enabled that should not be, and
even worse, not enable those that should.
This patch makes streamline_config process menuconfigs just like it
would process a config.
Reported-by: member graysky <graysky@archlinux.us>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
I noticed that some Kconfig files have multi line dependencies
that continue with a backslash. Those dependencies on the next
line will be missed by streamline_config.
For example:
config CS89x0
tristate "CS89x0 support"
depends on NET_ETHERNET && (ISA || EISA || MACH_IXDP2351 \
|| ARCH_IXDP2X01 || MACH_MX31ADS)
The "|| ARCH_IXDP2X01 || MACH_MX31ADS)" will not be processed.
This patch adds code to handle this case.
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
On Fri, Aug 17, 2010 at 01:43PM +0800, Américo Wang wrote:
> Acked-by: WANG Cong <xiyou.wangcong@gmail.com>
>
> BTW, I think we should add "use strict;" too.
Then I added "use strict;" to streamline_config.pl, I saw another
warning.
> Global symbol "$dir" requires explicit package name at
scripts/kconfig/streamline_config.pl line 286.
> Global symbol "$dir" requires explicit package name at
scripts/kconfig/streamline_config.pl line 287.
> Global symbol "$dir" requires explicit package name at
scripts/kconfig/streamline_config.pl line 288.
Then I added "my $dir;" to line 285.
Cc: Américo Wang <xiyou.wangcong@gmail.com>
Cc: Toralf Foerster <toralf.foerster@gmx.de>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: Hiromu Yakura <hiromu1996@gmail.com>
LKML-Reference: <1282042158.7160.9.camel@hiromu-Macbook>
[ changed to just add my in front of $dir instead of new line ]
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
When I added "use strict;" to streamline_config.pl, I saw the following
warnings:
> Global symbol "%prompt" requires explicit package name at
scripts/kconfig/streamline_config.pl line 183.
> Global symbol "%prompt" requires explicit package name at
scripts/kconfig/streamline_config.pl line 368.
The declaration of %prompt was incorrect, and should have been %prompts.
Cc: Toralf Foerster <toralf.foerster@gmx.de>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: Hiromu Yakura <hiromu1996@gmail.com>
LKML-Reference: <1281845597.11566.5.camel@camp10-laptop>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Chris Mason