When running the 083th case of xfstests on the filesystem with
"compress-force=lzo", the following WARNINGs were triggered.
  WARNING: at fs/btrfs/inode.c:7908
  WARNING: at fs/btrfs/inode.c:7909
  WARNING: at fs/btrfs/inode.c:7911
  WARNING: at fs/btrfs/extent-tree.c:4510
  WARNING: at fs/btrfs/extent-tree.c:4511

This problem was introduced by the patch "Btrfs: fix deadlock due
to unsubmitted". In this patch, there are two bugs which caused
the above problem.

The 1st one is a off-by-one bug, if the DIO write return 0, it is
also a short write, we need release the reserved space for it. But
we didn't do it in that patch. Fix it by change "ret > 0" to
"ret >= 0".

The 2nd one is ->outstanding_extents was increased twice when
a short write happened. As we know, ->outstanding_extents is
a counter to keep track of the number of extent items we may
use duo to delalloc, when we reserve the free space for a
delalloc write, we assume that the write will introduce just
one extent item, so we increase ->outstanding_extents by 1 at
that time. And then we will increase it every time we split the
write, it is done at the beginning of btrfs_get_blocks_direct().
So when a short write happens, we needn't increase
->outstanding_extents again. But this patch done.

In order to fix the 2nd problem, I re-write the logic for
->outstanding_extents operation. We don't increase it at the
beginning of btrfs_get_blocks_direct(), instead, we just
increase it when the split actually happens.
Reported-by: NMitch Harder <mitch.harder@sabayonlinux.org>
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

This comes from one of btrfs's project ideas,
As we defragment files, we break any sharing from other snapshots.
The balancing code will preserve the sharing, and defrag needs to grow this
as well.

Now we're able to fill the blank with this patch, in which we make full use of
backref walking stuff.

Here is the basic idea,
o  set the writeback ranges started by defragment with flag EXTENT_DEFRAG
o  at endio, after we finish updating fs tree, we use backref walking to find
   all parents of the ranges and re-link them with the new COWed file layout by
   adding corresponding backrefs.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Very large fallocate requests are cpu bound and result in extents with a
repeating pattern of ever decreasing size:

$ time fallocate -l 1T file
real	0m13.039s

( an excerpt of the extents from btrfs-debug-tree: )
  prealloc data disk byte 1536292564992 nr 397312
  prealloc data disk byte 1536292962304 nr 196608
  prealloc data disk byte 1536293158912 nr 98304
  prealloc data disk byte 1536293257216 nr 49152
  prealloc data disk byte 1536293306368 nr 24576
  prealloc data disk byte 1536293330944 nr 12288
  prealloc data disk byte 1536293343232 nr 8192
  prealloc data disk byte 1536293351424 nr 4096
  prealloc data disk byte 1536293355520 nr 4096
  prealloc data disk byte 1536293359616 nr 4096

The excessive cpu use comes from __btrfs_prealloc_file_range() trying to
allocate the entire remaining size after each extent is allocated.
btrfs_reserve_extent() repeatedly cuts this requested size in half until
it gets down to the size that the allocators can return.  We limit the
problem for now by capping each reservation at 256 meg.

The small extents come from a masking bug when decreasing the requested
reservation size.  The high 32bits are cleared and the remaining low
bits might happen to reserve a small size.   Fix this by using
round_down() which properly casts the mask.

After these fixes huge fallocate requests are fast and result in nice
large extents:

$ time fallocate -l 1T file
real	0m0.082s

  prealloc data disk byte 1112425889792 nr 268435456
  prealloc data disk byte 1112694325248 nr 268435456
  prealloc data disk byte 1112962760704 nr 268435456
Reported-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: NZach Brown <zab@redhat.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Right now inode cache inode is treated as the same as space cache
inode, ie. keep inode in memory till putting super.

But this leads to an awkward situation.

If we're going to delete a snapshot/subvolume, btrfs will not
actually delete it and return free space, but will add it to dead
roots list until the last inode on this snap/subvol being destroyed.
Then we'll fetch deleted roots and cleanup them via cleaner thread.

So here is the problem, if we enable inode cache option, each
snap/subvol has a cached inode which is used to store inode allcation
information.  And this cache inode will be kept in memory, as the above
said.  So with inode cache, snap/subvol can only be added into
dead roots list during freeing roots stage in umount, so that we can
ONLY get space back after another remount(we cleanup dead roots on mount).

But the real thing is we'll no more use the snap/subvol if we mark it
deleted, so we can safely iput its cache inode when we delete snap/subvol.

Another thing is that we need to change the rules of droping inode, we
don't keep snap/subvol's cache inode in memory till end so that we can
add snap/subvol into dead roots list in time.
Reported-by: NMitch Harder <mitch.harder@sabayonlinux.org>
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

This idea is from ext4. By this patch, we can make the dio write parallel,
and improve the performance. But because we can not update isize without
i_mutex, the unlocked dio write just can be done in front of the EOF.

We needn't worry about the race between dio write and truncate, because the
truncate need wait untill all the dio write end.

And we also needn't worry about the race between dio write and punch hole,
because we have extent lock to protect our operation.

I ran fio to test the performance of this feature.

== Hardware ==
CPU: Intel(R) Core(TM)2 Duo CPU     E7500  @ 2.93GHz
Mem: 2GB
SSD: Intel X25-M 120GB (Test Partition: 60GB)

== config file ==
[global]
ioengine=psync
direct=1
bs=4k
size=32G
runtime=60
directory=/mnt/btrfs/
filename=testfile
group_reporting
thread

[file1]
numjobs=1 # 2 4
rw=randwrite

== result (KBps) ==
write	1	2	4
lock	24936	24738	24726
nolock	24962	30866	32101

== result (iops) ==
write	1	2	4
lock	6234	6184	6181
nolock	6240	7716	8025
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Currently, we can do unlocked dio reads, but the following race
is possible:

dio_read_task			truncate_task
				->btrfs_setattr()
->btrfs_direct_IO
    ->__blockdev_direct_IO
      ->btrfs_get_block
				  ->btrfs_truncate()
				 #alloc truncated blocks
				 #to other inode
      ->submit_io()
     #INFORMATION LEAK

In order to avoid this problem, we must serialize unlocked dio reads with
truncate. There are two approaches:
- use extent lock to protect the extent that we truncate
- use inode_dio_wait() to make sure the truncating task will wait for
  the read DIO.

If we use the 1st one, we will meet the endless truncation problem due to
the nonlocked read DIO after we implement the nonlocked write DIO. It is
because we still need invoke inode_dio_wait() avoid the race between write
DIO and truncation. By that time, we have to introduce

  btrfs_inode_{block, resume}_nolock_dio()

again. That is we have to implement this patch again, so I choose the 2nd
way to fix the problem.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The deadlock problem happened when running fsstress(a test program in LTP).

Steps to reproduce:
 # mkfs.btrfs -b 100M <partition>
 # mount <partition> <mnt>
 # <Path>/fsstress -p 3 -n 10000000 -d <mnt>

The reason is:
btrfs_direct_IO()
 |->do_direct_IO()
     |->get_page()
     |->get_blocks()
     |	 |->btrfs_delalloc_resereve_space()
     |	 |->btrfs_add_ordered_extent() -------	Add a new ordered extent
     |->dio_send_cur_page(page0) --------------	We didn't submit bio here
     |->get_page()
     |->get_blocks()
	 |->btrfs_delalloc_resereve_space()
	     |->flush_space()
		 |->btrfs_start_ordered_extent()
		     |->wait_event() ----------	Wait the completion of
						the ordered extent that is
						mentioned above

But because we didn't submit the bio that is mentioned above, the ordered
extent can not complete, we would wait for its completion forever.

There are two methods which can fix this deadlock problem:
1. submit the bio before we invoke get_blocks()
2. reserve the space before we do dio

Though the 1st is the simplest way, we need modify the code of VFS, and it
is likely to break contiguous requests, and introduce performance regression
for the other filesystems.

So we have to choose the 2nd way.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Cc: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

I noticed we were getting lots of warnings with xfstest 83 because we have
reservations outstanding.  This is because we moved the orphan add outside
of the truncate, but we don't actually cleanup our reservation if something
fails.  This fixes the problem and I no longer see warnings.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Sometimes xfstest 83 will fail to remount the scratch device because we've
gotten ourselves so full that we cannot cleanup the orphan items.  In this
case check to see if we're doing the orphan cleanup and if we are allow us
to steal our reservation from the global block rsv.  With this patch I've
not been able to reproduce the failed mount problem.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

I noticed we would deadlock if we aborted a transaction while doing
compressed io.  This is because we don't unlock our pages if something goes
horribly wrong.  To fix this we need to make sure that we call
extent_clear_unlock_delalloc in order to unlock all the pages.  If we have
to cow in the async submission thread we need to make sure to unlock our
locked_page as the cow error path will not unlock the locked page as it
depends on the caller to unlock that page.  With this patch we no longer
deadlock on the page lock when we have an aborted transaction.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Dave sent me a panic where we were doing the orphan cleanup and panic'ed
trying to release our reservation from the orphan block rsv.  The reason for
this is because our orphan block rsv had been free'd out from underneath us
because the transaction commit found that there were no orphan inodes
according to its count and decided to free it.  This is incorrect so make
sure we inc the orphan inodes count so the accounting is all done properly.
This would also cause the warning in the orphan commit code normally if you
had any orphans to cleanup as they would only decrement the orphan count so
you'd get a negative orphan count which could cause problems during runtime.
Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

When a transaction aborts or there's an EIO on an ordered extent or any
error really we will not free up the space we reserved for this ordered
extent.  This results in warnings from the block group cache cleanup in the
case of a transaction abort, or leaking space in the case of EIO on an
ordered extent.  Fix this up by free'ing the reserved space if we have an
error at all trying to complete an ordered extent.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

We need not use a global lock to protect the delalloc_bytes of the
inode, just use its own lock. In this way, we can reduce the lock
contention and ->delalloc_lock will just protect delalloc inode
list.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

fs_info->delalloc_bytes is accessed very frequently, so use percpu
counter instead of the u64 variant for it to reduce the lock
contention.

This patch also fixed the problem that we access the variant
without the lock protection.At worst, we would not flush the
delalloc inodes, and just return ENOSPC error when we still have
some free space in the fs.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The header file will then be installed under /usr/include/linux so that
userspace applications can refer to Btrfs ioctls by name and use the same
structs used internally in the kernel.
Signed-off-by: NFilipe Brandenburger <filbranden@google.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

This reverts commit 2794ed01.

Wasn't supposed to get used in btrfs_mknod, it was supposed to be in
btrfs_create, which was done in commit
9185aa58.
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

btrfs_start_delalloc_inodes() needn't traverse and flush the delalloc inodes
repeatedly. It is because we can regard the data that the users write after
we start delalloc inodes flush as the one which is after the delalloc inodes
flush is done, and we can flush it next time.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The API in tree log code has done sort of changes, and it proves that
we can benifit from using token, so do the same thing here.

function_graph tracer's timer shows that it costs nearly half time
of before(39.788us -> 22.391us).
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Since we don't actually copy the extent information from the source tree in
the fast case we don't need to wait for ordered io to be completed in order
to fsync, we just need to wait for the io to be completed.  So when we're
logging our file just attach all of the ordered extents to the log, and then
when the log syncs just wait for IO_DONE on the ordered extents and then
write the super.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Use wrapper page_offset to get byte-offset into filesystem object for page.
Signed-off-by: NLiu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

We're running into having 50-100 orphans left over with xfstests 83
because of ENOSPC when trying to start the transaction for the inode update.
But in fact, it makes no sense in updating the inode for the new size while
we're deleting the stupid thing. This patch fixes this problem.
Reported-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

This builds on David Woodhouse's original Btrfs raid5/6 implementation.
The code has changed quite a bit, blame Chris Mason for any bugs.

Read/modify/write is done after the higher levels of the filesystem have
prepared a given bio.  This means the higher layers are not responsible
for building full stripes, and they don't need to query for the topology
of the extents that may get allocated during delayed allocation runs.
It also means different files can easily share the same stripe.

But, it does expose us to incorrect parity if we crash or lose power
while doing a read/modify/write cycle.  This will be addressed in a
later commit.

Scrub is unable to repair crc errors on raid5/6 chunks.

Discard does not work on raid5/6 (yet)

The stripe size is fixed at 64KiB per disk.  This will be tunable
in a later commit.
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We'll want to merge writes so they can fill a full RAID[56] stripe, but
not necessarily reads.
Signed-off-by: NDavid Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

btrfs_start_delalloc_inodes() locks the delalloc_inodes list, fetches the
first inode, unlocks the list, triggers btrfs_alloc_delalloc_work/
btrfs_queue_worker for this inode, and then it locks the list, checks the
head of the list again. But because we don't delete the first inode that it
deals with before, it will fetch the same inode. As a result, this function
allocates a huge amount of btrfs_delalloc_work structures, and OOM happens.

Fix this problem by splice this delalloc list.
Reported-by: NAlex Lyakas <alex.btrfs@zadarastorage.com>
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

truncate() vs. ftruncate() differ in the VFS; truncate()
doesn't set (ATTR_CTIME | ATTR_MTIME), and it's up to the
fs to do the timestamp updates if the size changes.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: NJosef Bacik <josef@toxicpanda.com>

btrfs_cont_expand() tries to free an IS_ERR em as it gets an error from
btrfs_get_extent() and breaks out of its loop.

An instance of -EEXIST was reported in the wild:

  https://bugzilla.redhat.com/show_bug.cgi?id=874407

I have no idea if that -EEXIST is surprising, or not.  Regardless, this
error handling should be cleaned up to handle other reasonable errors
(ENOMEM, EIO; whatever).

This seemed to be the only buggy freeing of the relatively rare IS_ERR
em so I opted to fix the caller rather than teach free_extent_map() to
use IS_ERR_OR_NULL().
Signed-off-by: NZach Brown <zab@redhat.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: NJosef Bacik <josef@toxicpanda.com>

xfstests case 285 complains.

It it because btrfs did not try to find unwritten delalloc
bytes(only dirty pages, not yet writeback) behind prealloc
extents, it ends up finding nothing while we're with SEEK_DATA.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Running xfstests 83 in a loop would sometimes fail the fsck.  This happens
because if we invalidate a page that already has an ordered extent setup for
it we will complete the ordered extent ourselves, assuming that the truncate
will clean everything up.  The problem with this is there is plenty of time
for the truncate to fail after we've done this work.  So to fix this we need
to add the orphan item first to make sure the cleanup gets done properly,
and then we can truncate the pagecache and all that stuff and be safe.  This
fixes the btrfsck failures I was seeing while running 83 in a loop.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The code that relied on that flag was ripped out of btrfs quite some
time ago, and never added back. Josef indicated that he was going to
take a different approach to the problem in btrfs, and that we
could just eliminate this flag.

Cc: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: NJeff Layton <jlayton@redhat.com>
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

Users report a bug, the reproducer is:
$ mkfs.btrfs /dev/loop0
$ mount /dev/loop0 /mnt/btrfs/
$ mkdir /mnt/btrfs/dir
$ chattr +C /mnt/btrfs/dir/
$ dd if=/dev/zero of=/mnt/btrfs/dir/foo bs=4K count=10;
$ lsattr /mnt/btrfs/dir/foo
---------------C- /mnt/btrfs/dir/foo
$ filefrag /mnt/btrfs/dir/foo
/mnt/btrfs/dir/foo: 1 extent found    ---> an extent
$ dd if=/dev/zero of=/mnt/btrfs/dir/foo bs=4K count=1 seek=5 conv=notrunc,nocreat; sync
$ filefrag /mnt/btrfs/dir/foo
/mnt/btrfs/dir/foo: 3 extents found   ---> with nocow, btrfs breaks the extent into three parts

The new created file should not only inherit the NODATACOW flag, but also
honor NODATASUM flag, because we must do COW on a file extent with checksum.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

The handling for directory crc hash overflows was fairly obscure,
split_leaf returns EOVERFLOW when we try to extend the item and that is
supposed to bubble up to userland.  For a while it did so, but along the
way we added better handling of errors and forced the FS readonly if we
hit IO errors during the directory insertion.

Along the way, we started testing only for EEXIST and the EOVERFLOW case
was dropped.  The end result is that we may force the FS readonly if we
catch a directory hash bucket overflow.

This fixes a few problem spots.  First I add tests for EOVERFLOW in the
places where we can safely just return the error up the chain.

btrfs_rename is harder though, because it tries to insert the new
directory item only after it has already unlinked anything the rename
was going to overwrite.  Rather than adding very complex logic, I added
a helper to test for the hash overflow case early while it is still safe
to bail out.

Snapshot and subvolume creation had a similar problem, so they are using
the new helper now too.
Signed-off-by: NChris Mason <chris.mason@fusionio.com>
Reported-by: NPascal Junod <pascal@junod.info>

When a new file is created with btrfs_create(), the inode will initially be
created with permissions 0666 and later on in btrfs_init_acl() it will be
adapted to mask out the umask bits. The problem is that this change won't make
it into the btrfs_inode unless there's another change to the inode (e.g. writing
content changing the size or touching the file changing the mtime.)

This fix adds a call to btrfs_update_inode() to btrfs_create() to make sure that
the change will not get lost if the in-memory inode is flushed before other
changes are made to the file.
Signed-off-by: NFilipe Brandenburger <filbranden@google.com>
Reviewed-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

This starts a transaction and dirties the inode everytime we call it, which
is super expensive if you have a write heavy workload.  We will be updating
the inode when the IO completes and we reserve the space for the inode
update when we reserve space for the write, so there is no chance of loss of
information or enospc issues.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We don't really need to copy extents from the source tree since we have all
of the information already available to us in the extent_map tree.  So
instead just write the extents straight to the log tree and don't bother to
copy the extent items from the source tree.
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

We are going to use EM's to log extents in the future, so we need to not
mark them as prealloc if they aren't actually prealloc extents.  Instead
mark them with FILLING so we know to ammend mod_start/mod_len and that way
we don't confuse the extent logging code.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

If we've written to a prealloc extent we need to know the original block len
for the extent.  We can't figure this out currently since ->block_len is
just set to the extent length.  So introduce ->orig_block_len so that we
know how many bytes were in the original extent for proper extent logging
that future patches will need.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

The tree logging stuff needs the csums to be on the ordered extents in order
to log them properly, so mark that we're sync and inline the csum creation
so we don't have to wait on the csumming to be done when logging extents
that are still in flight.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

Currently we copy all the file information into the log, inode item, the
refs, xattrs etc.  Except most of this doesn't change from fsync to fsync,
just the inode item changes.  So set a flag if an xattr changes or a link is
added, and otherwise only log the inode item.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

ret variant may be set to 0 if we read page successfully, but it might be
released before we lock it again. On this case, if we fail to allocate a
new page, we will return 0, it is wrong, fix it.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>

If we runt the direct IO, we should not run auto defrag, because it may
introduce buffered IO vs direcIO problem, and make direct IO slow down.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@fusionio.com>