linux-toradex.git/fs, branch v3.16-rc2

Merge tag 'locks-v3.16-2' of git://git.samba.org/jlayton/linux

2014-06-22T02:40:30+00:00

Pull file locking fixes from Jeff Layton:
 "File locking related bugfixes

  Nothing too earth-shattering here.  A fix for a potential regression
  due to a patch in pile #1, and the addition of a memory barrier to
  prevent a race condition between break_deleg and generic_add_lease"

* tag 'locks-v3.16-2' of git://git.samba.org/jlayton/linux:
  locks: set fl_owner for leases back to current->files
  locks: add missing memory barrier in break_deleg

Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs

2014-06-22T00:21:43+00:00

Pull btrfs fixes from Chris Mason:
 "This fixes some lockups in btrfs reported with rc1.  It probably has
  some performance impact because it is backing off our spinning locks
  more often and switching to a blocking lock.  I'll be able to nail
  that down next week, but for now I want to get the lockups taken care
  of.

  Otherwise some more stack reduction and assorted fixes"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
  Btrfs: fix wrong error handle when the device is missing or is not writeable
  Btrfs: fix deadlock when mounting a degraded fs
  Btrfs: use bio_endio_nodec instead of open code
  Btrfs: fix NULL pointer crash when running balance and scrub concurrently
  btrfs: Skip scrubbing removed chunks to avoid -ENOENT.
  Btrfs: fix broken free space cache after the system crashed
  Btrfs: make free space cache write out functions more readable
  Btrfs: remove unused wait queue in struct extent_buffer
  Btrfs: fix deadlocks with trylock on tree nodes

Merge branch 'for-3.16' of git://linux-nfs.org/~bfields/linux

2014-06-22T00:20:38+00:00

Pull nfsd bugfixes from Bruce Fields:
 "Fixes for a new regression from the xdr encoding rewrite, and a
  delegation problem we've had for a while (made somewhat more annoying
  by the vfs delegation support added in 3.13)"

* 'for-3.16' of git://linux-nfs.org/~bfields/linux:
  NFSD: fix bug for readdir of pseudofs
  NFSD: Don't hand out delegations for 30 seconds after recalling them.

Btrfs: fix wrong error handle when the device is missing or is not writeable

2014-06-19T21:20:56+00:00

The original bio might be submitted, so we shoud increase bi_remaining to
account for it when we deal with the error that the device is missing or
is not writeable, or we would skip the endio handle.

Signed-off-by: Miao Xie 
Signed-off-by: Chris Mason

Btrfs: fix deadlock when mounting a degraded fs

2014-06-19T21:20:56+00:00

The deadlock happened when we mount degraded filesystem, the reproduced
steps are following:
 # mkfs.btrfs -f -m raid1 -d raid1  
 # echo 1 > /sys/block/`basename `/device/delete
 # mount -o degraded  

The reason was that the counter -- bi_remaining was wrong. If the missing
or unwriteable device was the last device in the mapping array, we would
not submit the original bio, so we shouldn't increase bi_remaining of it
in btrfs_end_bio(), or we would skip the final endio handle.

Fix this problem by adding a flag into btrfs bio structure. If we submit
the original bio, we will set the flag, and we increase bi_remaining counter,
or we don't.

Though there is another way to fix it -- decrease bi_remaining counter of the
original bio when we make sure the original bio is not submitted, this method
need add more check and is easy to make mistake.

Signed-off-by: Miao Xie 
Reviewed-by: Liu Bo 
Signed-off-by: Chris Mason

Btrfs: use bio_endio_nodec instead of open code

2014-06-19T21:20:55+00:00

Signed-off-by: Miao Xie 
Signed-off-by: Chris Mason

Btrfs: fix NULL pointer crash when running balance and scrub concurrently

2014-06-19T21:20:55+00:00

While running balance, scrub, fsstress concurrently we hit the
following kernel crash:

[56561.448845] BTRFS info (device sde): relocating block group 11005853696 flags 132
[56561.524077] BUG: unable to handle kernel NULL pointer dereference at 0000000000000078
[56561.524237] IP: [] scrub_chunk.isra.12+0xdd/0x130 [btrfs]
[56561.524297] PGD 9be28067 PUD 7f3dd067 PMD 0
[56561.524325] Oops: 0000 [#1] SMP
[....]
[56561.527237] Call Trace:
[56561.527309]  [] scrub_enumerate_chunks+0x24e/0x490 [btrfs]
[56561.527392]  [] ? abort_exclusive_wait+0x50/0xb0
[56561.527476]  [] btrfs_scrub_dev+0x1a4/0x530 [btrfs]
[56561.527561]  [] btrfs_ioctl+0x13f7/0x2a90 [btrfs]
[56561.527639]  [] do_vfs_ioctl+0x2e0/0x4c0
[56561.527712]  [] ? vtime_account_user+0x54/0x60
[56561.527788]  [] ? __audit_syscall_entry+0x9c/0xf0
[56561.527870]  [] SyS_ioctl+0x81/0xa0
[56561.527941]  [] tracesys+0xdd/0xe2
[...]
[56561.528304] RIP  [] scrub_chunk.isra.12+0xdd/0x130 [btrfs]
[56561.528395]  RSP 
[56561.528454] CR2: 0000000000000078

This is because in btrfs_relocate_chunk(), we will free @bdev directly while
scrub may still hold extent mapping, and may access freed memory.

Fix this problem by wrapping freeing @bdev work into free_extent_map() which
is based on reference count.

Reported-by: Qu Wenruo 
Signed-off-by: Wang Shilong 
Signed-off-by: Miao Xie 
Signed-off-by: Chris Mason

btrfs: Skip scrubbing removed chunks to avoid -ENOENT.

2014-06-19T21:20:54+00:00

When run scrub with balance, sometimes -ENOENT will be returned, since
in scrub_enumerate_chunks() will search dev_extent in *COMMIT_ROOT*, but
btrfs_lookup_block_group() will search block group in *MEMORY*, so if a
chunk is removed but not committed, -ENOENT will be returned.

However, there is no need to stop scrubbing since other chunks may be
scrubbed without problem.

So this patch changes the behavior to skip removed chunks and continue
to scrub the rest.

Signed-off-by: Qu Wenruo 
Signed-off-by: Miao Xie 
Signed-off-by: Chris Mason

Btrfs: fix broken free space cache after the system crashed

2014-06-19T21:20:54+00:00

When we mounted the filesystem after the crash, we got the following
message:
  BTRFS error (device xxx): block group xxxx has wrong amount of free space
  BTRFS error (device xxx): failed to load free space cache for block group xxx

It is because we didn't update the metadata of the allocated space (in extent
tree) until the file data was written into the disk. During this time, there was
no information about the allocated spaces in either the extent tree nor the
free space cache. when we wrote out the free space cache at this time (commit
transaction), those spaces were lost. In fact, only the free space that is
used to store the file data had this problem, the others didn't because
the metadata of them is updated in the same transaction context.

There are many methods which can fix the above problem
- track the allocated space, and write it out when we write out the free
  space cache
- account the size of the allocated space that is used to store the file
  data, if the size is not zero, don't write out the free space cache.

The first one is complex and may make the performance drop down.
This patch chose the second method, we use a per-block-group variant to
account the size of that allocated space. Besides that, we also introduce
a per-block-group read-write semaphore to avoid the race between
the allocation and the free space cache write out.

Signed-off-by: Miao Xie 
Signed-off-by: Chris Mason

Btrfs: make free space cache write out functions more readable

2014-06-19T21:20:54+00:00

This patch makes the free space cache write out functions more readable,
and beisdes that, it also reduces the stack space that the function --
__btrfs_write_out_cache uses from 194bytes to 144bytes.

Signed-off-by: Miao Xie 
Signed-off-by: Chris Mason