Lockdep complains about btrfs's async commit:

[ 2372.462171] [ BUG: bad unlock balance detected! ]
[ 2372.462191] 3.12.0+ #32 Tainted: G        W
[ 2372.462209] -------------------------------------
[ 2372.462228] ceph-osd/14048 is trying to release lock (sb_internal) at:
[ 2372.462275] [<ffffffffa022cb10>] btrfs_commit_transaction_async+0x1b0/0x2a0 [btrfs]
[ 2372.462305] but there are no more locks to release!
[ 2372.462324]
[ 2372.462324] other info that might help us debug this:
[ 2372.462349] no locks held by ceph-osd/14048.
[ 2372.462367]
[ 2372.462367] stack backtrace:
[ 2372.462386] CPU: 2 PID: 14048 Comm: ceph-osd Tainted: G        W    3.12.0+ #32
[ 2372.462414] Hardware name: To Be Filled By O.E.M. To Be Filled By O.E.M./To be filled by O.E.M., BIOS 080015  11/09/2011
[ 2372.462455]  ffffffffa022cb10 ffff88007490fd28 ffffffff816f094a ffff8800378aa320
[ 2372.462491]  ffff88007490fd50 ffffffff810adf4c ffff8800378aa320 ffff88009af97650
[ 2372.462526]  ffffffffa022cb10 ffff88007490fd88 ffffffff810b01ee ffff8800898c0000
[ 2372.462562] Call Trace:
[ 2372.462584]  [<ffffffffa022cb10>] ? btrfs_commit_transaction_async+0x1b0/0x2a0 [btrfs]
[ 2372.462619]  [<ffffffff816f094a>] dump_stack+0x45/0x56
[ 2372.462642]  [<ffffffff810adf4c>] print_unlock_imbalance_bug+0xec/0x100
[ 2372.462677]  [<ffffffffa022cb10>] ? btrfs_commit_transaction_async+0x1b0/0x2a0 [btrfs]
[ 2372.462710]  [<ffffffff810b01ee>] lock_release+0x18e/0x210
[ 2372.462742]  [<ffffffffa022cb36>] btrfs_commit_transaction_async+0x1d6/0x2a0 [btrfs]
[ 2372.462783]  [<ffffffffa025a7ce>] btrfs_ioctl_start_sync+0x3e/0xc0 [btrfs]
[ 2372.462822]  [<ffffffffa025f1d3>] btrfs_ioctl+0x4c3/0x1f70 [btrfs]
[ 2372.462849]  [<ffffffff812c0321>] ? avc_has_perm+0x121/0x1b0
[ 2372.462873]  [<ffffffff812c0224>] ? avc_has_perm+0x24/0x1b0
[ 2372.462897]  [<ffffffff8107ecc8>] ? sched_clock_cpu+0xa8/0x100
[ 2372.462922]  [<ffffffff8117b145>] do_vfs_ioctl+0x2e5/0x4e0
[ 2372.462946]  [<ffffffff812c19e6>] ? file_has_perm+0x86/0xa0
[ 2372.462969]  [<ffffffff8117b3c1>] SyS_ioctl+0x81/0xa0
[ 2372.462991]  [<ffffffff817045a4>] tracesys+0xdd/0xe2

====================================================

It's because that we don't do the right thing when checking if it's ok to
tell lockdep that we're trying to release the rwsem.

If the trans handle's type is TRANS_ATTACH, we won't acquire the freeze rwsem, but
as TRANS_ATTACH fits the check (trans < TRANS_JOIN_NOLOCK), we'll release the freeze
rwsem, which makes lockdep complains a lot.
Reported-by: NMa Jianpeng <majianpeng@gmail.com>
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

The 'git blame' history shows that, the old transaction commit code has to do
twice to ensure roots are updated and we have to flush metadata and super block
manually, however, right now all of these can be handled well inside
the transaction commit code without extra efforts.

And the error handling part remains same with the current code, -- 'return to
caller once we get error'.

This saves us a transaction commit and a flush of super block, which are both
heavy operations according to ftrace output analysis.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

__btrfs_start_workers returns 0 in case it raced with
btrfs_stop_workers and lost the race.  This is wrong because worker in
this case is not allowed to start and is in fact destroyed.  Return
-EINVAL instead.
Signed-off-by: NIlya Dryomov <idryomov@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

This disables the "if needed, write the good copy back before the read
is completed" part of the read sequence for read-only mounts.

Cc: Jan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: NIlya Dryomov <idryomov@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Currently if we discover an error when scrubbing in ro mode we a)
blindly increment the uncorrectable_errors counter, and b) spam the
dmesg with the 'unable to fixup (regular) error at ...' message, even
though a) we haven't tried to determine if the error is correctable or
not, and b) we haven't tried to fixup anything.  Fix this.

Cc: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: NIlya Dryomov <idryomov@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

If we fsync, seek and write, rename and then fsync again we will lose the
modified hole extent because the rename will drop all of the modified extents
since we didn't do the fast search.  We need to only drop the modified extents
if we didn't do the fast search and we were logging the entire inode as we don't
need them anymore, otherwise this is being premature.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

If we rename a file that is already in the log and we fsync again we will lose
the new name.  This is because we just log the inode update and not the new ref.
To fix this we just need to check if we are logging the new name of the inode
and copy all the metadata instead of just updating the inode itself.  With this
patch my testcase now passes.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We can just return false for this so we stop doing the snapshot aware defrag
stuff.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

3 of 4 callers actually want file_inode()...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Heiko Carstens noticed that btrfs was using empty_zero_page
incorrectly.  He explained:

	The definition of empty_zero_page is architecture specific.  It
	is (currently) either a character array, an unsigned long
	containing the address of the empty_zero_page, or even worse
	only the address of the struct page belonging to the
	empty_zero_page.

This commit changes btrfs to use a for-loop instead.  On x86
the resulting .ko is smaller, and we're no longer worrying about
how each arch builds its zeros.
Reported-by: NHeiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

rename the function -- btrfs_start_all_delalloc_inodes(), and make its
name be compatible to btrfs_wait_ordered_roots(), since they are always
used at the same place.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

It is very likely that there are lots of ordered extents in the filesytem,
if we wait for the completion of all of them when we want to reclaim some
space for the metadata space reservation, we would be blocked for a long
time. The performance would drop down suddenly for a long time.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

It was very likely that there were lots of async delalloc pages in the
filesystem, if we waited until all the pages were flushed, we would be
blocked for a long time, and the performance would also drop down.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

In shrink_delalloc(), what we need reclaim is the metadata space, so
flushing pages by to_reclaim is not reasonable, it is very likely that
the pages we flush are not enough. And then we had to invoke the flush
function for several times, at the worst, we need call flush_space for
several times. It wasted time.

We improve this problem by converting the metadata space size we need
reserve to the delalloc bytes, By this way, we can flush the pages
by a reasonable number.

(Now we use a fixed number to do conversion, it is not flexible, maybe
 we can find a good way to improve it in the future.)
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

This patch picked up the code that was used to calculate the number of
the items for which we need reserve space, and we will use it in the next
patch.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We only allocate scrub workers if we pass all the necessary
checks, for example, there are no operation in progress.

Besides, move mutex lock protection outside of scrub_workers_get()
/scrub_workers_put().
Signed-off-by: NWang Shilong <wangsl.fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

I hit this problem with my no holes patch and it made me realize what the
problem was for bz 60834.  If the first item in the leaf is an inline extent and
we try to read anything starting from disk_bytenr onward we will read off the
end of the leaf.  So we need to check to see what it's type is, and if it's not
REG we can just break out.  This should fix this problem.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

The function write_ctree_super() in disk-io.c uses variable ret to return
the result of function write_all_supers(). Since, this variable serves
no purpose, hence the patch removes it and returns the call of the
called function.
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NRashika Kheria <rashika.kheria@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Fix spacing issues detected via checkpatch.pl in accordance with the
kernel style guidelines.
Signed-off-by: NDulshani Gunawardhana <dulshani.gunawardhana89@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Replace kmalloc(size * nr, ) with kmalloc_array(nr, size), thus making
it easier to check is that the calculation doesn't wrap or return a smaller allocation
Signed-off-by: NDulshani Gunawardhana <dulshani.gunawardhana89@gmail.com>
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Enclose macros with complex values within parenthesis in accordance to
checkpatch.pl.
Signed-off-by: NDulshani Gunawardhana <dulshani.gunawardhana89@gmail.com>
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Use WARN_ON()'s return value in place of WARN_ON(1) for cleaner source
code that outputs a more descriptive warnings. Also fix the styling
warning of redundant braces that came up as a result of this fix.
Signed-off-by: NDulshani Gunawardhana <dulshani.gunawardhana89@gmail.com>
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Remove redundant local zero structure, replacing it by the kernel's
global ZERO_PAGE.
Signed-off-by: NDulshani Gunawardhana <dulshani.gunawardhana89@gmail.com>
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Pack the structure btrfs_device in volumes.h to eliminate holes detected
by pahole, thus reducing binary memory footprint.
Signed-off-by: NDulshani Gunawardhana <dulshani.gunawardhana89@gmail.com>
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

This patch replaces multiple atomic_inc() with atomic_add() in
delayed-inode.c to reduce source code and have few instructions
for compilation.
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NRashika Kheria <rashika.kheria@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

The function free_root_pointers() in disk-io.h contains redundant code.
Therefore, this patch adds a helper function free_root_extent_buffers()
to free_root_pointers() to eliminate redundancy.
Reviewed-by: NZach Brown <zab@redhat.com>
Signed-off-by: NRashika Kheria <rashika.kheria@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Running balance and defrag concurrently can end up with a crash:

kernel BUG at fs/btrfs/relocation.c:4528!
RIP: 0010:[<ffffffffa01ac33b>]  [<ffffffffa01ac33b>] btrfs_reloc_cow_block+ 0x1eb/0x230 [btrfs]
Call Trace:
  [<ffffffffa01398c1>] ? update_ref_for_cow+0x241/0x380 [btrfs]
  [<ffffffffa0180bad>] ? copy_extent_buffer+0xad/0x110 [btrfs]
  [<ffffffffa0139da1>] __btrfs_cow_block+0x3a1/0x520 [btrfs]
  [<ffffffffa013a0b6>] btrfs_cow_block+0x116/0x1b0 [btrfs]
  [<ffffffffa013ddad>] btrfs_search_slot+0x43d/0x970 [btrfs]
  [<ffffffffa0153c57>] btrfs_lookup_file_extent+0x37/0x40 [btrfs]
  [<ffffffffa0172a5e>] __btrfs_drop_extents+0x11e/0xae0 [btrfs]
  [<ffffffffa013b3fd>] ? generic_bin_search.constprop.39+0x8d/0x1a0 [btrfs]
  [<ffffffff8117d14a>] ? kmem_cache_alloc+0x1da/0x200
  [<ffffffffa0138e7a>] ? btrfs_alloc_path+0x1a/0x20 [btrfs]
  [<ffffffffa0173ef0>] btrfs_drop_extents+0x60/0x90 [btrfs]
  [<ffffffffa016b24d>] relink_extent_backref+0x2ed/0x780 [btrfs]
  [<ffffffffa0162fe0>] ? btrfs_submit_bio_hook+0x1e0/0x1e0 [btrfs]
  [<ffffffffa01b8ed7>] ? iterate_inodes_from_logical+0x87/0xa0 [btrfs]
  [<ffffffffa016b909>] btrfs_finish_ordered_io+0x229/0xac0 [btrfs]
  [<ffffffffa016c3b5>] finish_ordered_fn+0x15/0x20 [btrfs]
  [<ffffffffa018cbe5>] worker_loop+0x125/0x4e0 [btrfs]
  [<ffffffffa018cac0>] ? btrfs_queue_worker+0x300/0x300 [btrfs]
  [<ffffffff81075ea0>] kthread+0xc0/0xd0
  [<ffffffff81075de0>] ? insert_kthread_work+0x40/0x40
  [<ffffffff8164796c>] ret_from_fork+0x7c/0xb0
  [<ffffffff81075de0>] ? insert_kthread_work+0x40/0x40
----------------------------------------------------------------------

It turns out to be that balance operation will bump root's @last_snapshot,
which enables snapshot-aware defrag path, and backref walking stuff will
find data reloc tree as refs' parent, and hit the BUG_ON() during COW.

As data reloc tree's data is just for relocation purpose, and will be deleted right
after relocation is done, it's unnecessary to walk those refs belonged to data reloc
tree, it'd be better to skip them.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

If something wrong happens in write endio, running snapshot-aware defragment
can end up with undefined results, maybe a crash, so we should avoid it.

In order to share similar code, this also adds a helper to free the struct for
snapshot-aware defrag.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

If we get any error while doing a dir index/item lookup in the
log tree, we were always unlinking the corresponding inode in
the subvolume. It makes sense to unlink only if the lookup failed
to find the dir index/item, which corresponds to NULL or -ENOENT,
and not when other errors happen (like a transient -ENOMEM or -EIO).
Signed-off-by: NFilipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We were setting the csums search offset and length to the right values if
the extent is compressed, but later on right before doing the csums lookup
we were overriding these two parameters regardless of compression being
set or not for the extent.
Signed-off-by: NFilipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We were ignoring the name component of the dir_item. Both the
name and data must fit within BTRFS_MAX_XATTR_SIZE(root).
Signed-off-by: NFilipe David Borba Manana <fdmanana@gmail.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Originally, we introduced scrub_super_lock to synchronize
tree log code with scrubbing super.

However we can replace scrub_super_lock with device_list_mutex,
because writing super will hold this mutex, this will reduce an extra
lock holding when writing supers in sync log code.
Signed-off-by: NWang Shilong <wangsl.fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We define a 'int' to get extent's generation by mistake,fix it.
Signed-off-by: NWang Shilong <wangsl.fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

After running space balance on a new fs, the fs check program outputed the
following warning message:
 free space inode generation (0) did not match free space cache generation (20)

Steps to reproduce:
 # mkfs.btrfs -f <dev>
 # mount <dev> <mnt>
 # btrfs balance start <mnt>
 # umount <mnt>
 # btrfs check <dev>

It was because there was no data space after the space balance, and the free
space write out task didn't try to allocate a new data chunk for the free space
inode when doing the reservation. So the data space reservation failed, and in
order to tell the free space loader that this free space inode could not be
trusted, the generation of the free space inode wasn't updated. Then the check
program found this problem and outputed the above message.

But in fact, it is safe that we try to allocate a new data chunk when we find
the data space is not enough. The patch fixes the above problem by this way.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

I noticed with my horrible snapshot excercisor that we were taking forever to
relocate the larger the file system got.  This appeared to be because we were
committing the transaction _constantly_.  There were a few places where we do
braindead things with metadata reservation, like start a transaction and then
try to refill the block rsv, which not only keeps us from committing a
transaction during the enospc stuff, but keeps us from doing some of the harder
flushing work which will make us more likely to need to commit the transaction.
We also were checking the block rsv and committing the transaction if the block
rsv was below a certain threshold, but we were doing this in a place where we
don't actually keep anything in the block rsv so this was always ending up false
so we always committed the transaction in this case.  I tested this to make sure
it didn't break anything, but it takes about 10 hours to get the box to this
state so I don't know how much of an impact it will really make.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

When using delalloc workers in a non-waiting way (like for enospc handling) we
can end up not actually waiting for the dirty pages to be started if we have
compression.  We need to add an extra filemap flush to make sure any async
extents that have started are actually moved along before returning.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

A user reported a list corruption warning from btrfs_remove_ordered_extent, it
is because we aren't taking the ordered_root_lock when we remove the inode from
the ordered operations list.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

This is just the write path, the only reason we start a transaction is so we can
check cross references, we don't make any actual changes, so there is no reason
to abort the transaction if we fail.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>