Commit d9abbf1c changed tree mod log locking around ROOT_REPLACE operations.
When a tree root is split, however, we were logging removal of all elements
from the root node before logging removal of half of the elements for the
split operation. This leads to a BUG_ON when rewinding.

This commit removes the erroneous logging of removal of all elements.
Signed-off-by: NJan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The following case will make the incompat/compat flag of the super block
be recovered.
 Task1					|Task2
 flags = btrfs_super_incompat_flags();	|
					|flags = btrfs_super_incompat_flags();
 flags |= new_flag1;			|
					|flags |= new_flag2;
 btrfs_set_super_incompat_flags(flags);	|
					|btrfs_set_super_incompat_flags(flags);
the new_flag1 is recovered.

In order to avoid this problem, we introduce a lock named super_lock into
the btrfs_fs_info structure. If we want to update incompat/compat flags
of the super block, we must hold it.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The new mount option is set after parsing the remount arguments,
so it is wrong that checking the autodefrag is close or not at
btrfs_remount_prepare(). Fix it.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Walking backref tree and btrfs quota rely on ulist very much.
This patch tries to use rb_tree to speed up search time.

The original code always checks whether an element
exists before adding a new element, however it costs O(n).

I try to add a rb_tree in the ulist,this is only used to speed up
search. I also do some measurements with quota enabled.

fsstress -p 4 -n 10000

Without this path:
real    0m51.058s       2m4.745s        1m28.222s       1m5.137s
user    0m0.035s        0m0.041s        0m0.105s        0m0.100s
sys     0m12.009s       0m11.246s       0m10.901s       0m10.999s       0m11.287s

With this path:
real    0m55.295s       0m50.960s       1m2.214s        0m48.273s
user    0m0.053s        0m0.095s        0m0.135s        0m0.107s
sys     0m7.766s        0m6.013s        0m6.319s        0m6.030s        0m6.532s

After applying the patch,the execute time is down by ~42%.(11.287s->6.532s)
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Reviewed-by: NJan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Two new flags are added to allow omitting the stream header and the
end command for btrfs send streams. This is used in cases where you
send multiple snapshots back-to-back in one stream.

This used to be encoded like this (with 2 snapshots in this example):
<stream header> + <sequence of commands> + <end cmd> +
<stream header> + <sequence of commands> + <end cmd> + EOF

The new format (if the two new flags are used) is this one:
<stream header> + <sequence of commands> +
                  <sequence of commands> + <end cmd>

Note that the currently existing receivers treat <end cmd> only as
an indication that a new <stream header> is following. This means,
you can just skip the sequence <end cmd> <stream header> without
loosing compatibility. As long as an EOF is following, the currently
existing receivers handle the new format (if the two new flags are
used) exactly as the old one.

So what is the benefit of this change? The goal is to be able to use
a single stream (one TCP connection) to multiplex a request/response
handshake plus Btrfs send streams, all in the same stream. In this
case you cannot evaluate an EOF condition as an end of the Btrfs send
stream. You need something else, and the <end cmd> is just perfect
for this purpose.

The summary is:
The format change is driven by the need to send several Btrfs send
streams over a single TCP connections, with the ability for a repeated
request/response handshake in the middle. And this format change does
not break any existing tool, it is completely compatible.

You could compare the old behaviour of the Btrfs send stream to the
one of ftp where you need a seperate request/response channel and
newly opened data transfer channels for each file, while the new
behaviour is more like http using a single stream for everything.
Signed-off-by: NStefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

__merge_refs() always return 0, it is unnecessary
for the caller to check the return value.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The only error return value of __add_prelim_ref() is -ENOMEM,
just return errors rather than trigger BUG_ON().
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Steps to reproduce:
	mkfs.btrfs <disk>
	mount <disk> <mnt>
	btrfs quota enable <mnt>
	btrfs sub create <mnt>/subv
	btrfs qgroup limit 10K <mnt>/subv
	btrfs quota disable <mnt>/subv

It is wrong for qgroup to reserve when disabling quota,
so just use tree_root to avoid edquot when disabling quota.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Step to reproduce:
	mkfs.btrfs <disk>
	mount <disk> <mnt>
	btrfs quota enable <mnt>
	btrfs qgroup limit 0/1 <mnt>
	dmesg

If the relative qgroup dosen't exist, flag 'BTRFS_QGROUP_STATUS_
FLAG_INCONSISTENT' will be set, and print the noise message.
This is wrong, we can just move find_qgroup_rb() before
update_qgroup_limit_item().this dosen't change the logic of the
function. But it can avoid unnecessary noise message and wrong set of flag.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The original code forgot to check 'inherit', we should
gurantee that all the qgroups in the struct 'inherit' exist.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Step to reproduce:
		mkfs.btrfs <disk>
		mount <disk> <mnt>
		btrfs quota enable <mnt>
		btrfs qgroup assign 0/1 1/1 <mnt>
		umount <mnt>
		btrfs-debug-tree <disk> | grep QGROUP
If we want to add a qgroup relation, we should gurantee that
'src' and 'dst' exist, otherwise, such qgroup relation should
not be allowed to create.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

We use mutex lock to protect all the user change operations.
So when we are calling find_qgroup_rb() to check whether qgroup
exists, we don't have to hold spin_lock.

Besides, when enabling/disabling quota, it must be single thread
when operations come here. spin lock must be firstly used to
clear quota_root when disabling quota, while enabling quota, spin
lock must be used to complete the last assign work.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The original code has one spin_lock 'qgroup_lock' to protect quota
configurations in memory. If we want to add a BTRFS_QGROUP_INFO_KEY,
it will be added to Btree firstly, and then update configurations in
memory,however, a race condition may happen between these operations.
For example:
	->add_qgroup_info_item()
		->add_qgroup_rb()

For the above case, del_qgroup_info_item() may happen just before
add_qgroup_rb().

What's worse, when we want to add a qgroup relation:
	->add_qgroup_relation_item()
		->add_qgroup_relations()

We don't have any checks whether 'src' and 'dst' exist before
add_qgroup_relation_item(), a race condition can also happen for
the above case.

To avoid race condition and have all the necessary checks, we introduce
a mutex lock 'qgroup_ioctl_lock', and we make all the user change operations
protected by the mutex lock.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Creating the subvolume/snapshots(including root subvolume) qgroup
auotomatically when enabling quota.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

__btrfs_unlink_inode() aborts its transaction when it sees errors after
it removes the directory item.  But it missed the case where
btrfs_del_dir_entries_in_log() returns an error.  If this happens then
the unlink appears to fail but the items have been removed without
updating the directory size.  The directory then has leaked bytes in
i_size and can never be removed.

Adding the missing transaction abort at least makes this failure
consistent with the other failure cases.

I noticed this while reading the code after someone on irc reported
having a directory with i_size but no entries.  I tested it by forcing
btrfs_del_dir_entries_in_log() to return -ENOMEM.
Signed-off-by: NZach Brown <zab@redhat.com>
Reviewed-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

This:

   # mkfs.btrfs /dev/sdb{1,2} ; wipefs -a /dev/sdb1; mount /dev/sdb2 /mnt/test

would lead to a blkdev open/close mismatch when the mount fails, and
a permanently busy (opened O_EXCL) sdb2:

   # wipefs -a /dev/sdb2
   wipefs: error: /dev/sdb2: probing initialization failed: Device or resource busy

It's because btrfs_open_devices() may open some devices, fail on
the last one, and return that failure stored in "ret."   The mount
then fails, but the caller then does not clean up the open devices.

Chris assures me that:

"btrfs_open_devices just means: go off and open every bdev you can from
this uuid.  It should return success if we opened any of them at all."

So change the logic to ignore any open failures; just skip processing
of that device.  Later on it's decided whether we have enough devices
to continue.
Reported-by: NJan Safranek <jsafrane@redhat.com>
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

It is very likely that there are several blocks in bio, it is very
inefficient if we get their csums one by one. This patch improves
this problem by getting the csums in batch.

According to the result of the following test, the execute time of
__btrfs_lookup_bio_sums() is down by ~28%(300us -> 217us).

 # dd if=<mnt>/file of=/dev/null bs=1M count=1024
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

A user sent me a btrfs-image of a file system that was panicing on mount during
the log recovery.  I had originally thought these problems were from a bug in
the free space cache code, but that was just a symptom of the problem.  The
problem is if your application does something like this

[prealloc][prealloc][prealloc]

the internal extent maps will merge those all together into one extent map, even
though on disk they are 3 separate extents.  So if you go to write into one of
these ranges the extent map will be right since we use the physical extent when
doing the write, but when we log the extents they will use the wrong sizes for
the remainder prealloc space.  If this doesn't happen to trip up the free space
cache (which it won't in a lot of cases) then you will get bogus entries in your
extent tree which will screw stuff up later.  The data and such will still work,
but everything else is broken.  This patch fixes this by not allowing extents
that are on the modified list to be merged.  This has the side effect that we
are no longer adding everything to the modified list all the time, which means
we now have to call btrfs_drop_extents every time we log an extent into the
tree.  So this allows me to drop all this speciality code I was using to get
around calling btrfs_drop_extents.  With this patch the testcase I've created no
longer creates a bogus file system after replaying the log.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

When logging changed extents I was logging ram_bytes as the current length,
which isn't correct, it's supposed to be the ram bytes of the original extent.
This is for compression where even if we split the extent we need to know the
ram bytes so when we uncompress the extent we know how big it will be.  This was
still working out right with compression for some reason but I think we were
getting lucky.  It was definitely off for prealloc which is why I noticed it,
btrfsck was complaining about it.  With this patch btrfsck no longer complains
after a log replay.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Dave was hitting a lockdep warning because we're now properly taking the ordered
operations mutex in the ordered wait stuff.  This is because some cases we will
have a trans handle when we are flushing delalloc space, but we can't wait on
ordered extents because we could potentially deadlock, so fix this by not doing
the wait if we have a trans handle.  Thanks
Reported-and-tested-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

I noticed that we will add a block group to the space info before we add it to
the block group cache rb tree, so we could potentially allocate from the block
group before it's able to be searched for.  I don't think this is too much of
a problem, the race window is microscopic, but just in case move the tree
insertion to above the space info linking.  This makes it easier to adjust the
error handling as well, so we can remove a couple of BUG_ON(ret)'s and have real
error handling setup for these scenarios.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

If btrfs_find_all_roots() fails, 'roots' has been freed or 'roots'
fails to allocate. We don't need to free it outside btrfs_find_all_roots()
again.Fix it.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The reason that BUG_ON() happens in these places is just
because of ENOMEM.

We try ro return ENOMEM rather than trigger BUG_ON(), the
caller will abort the transaction thus avoiding the kernel panic.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Reviewed-by: NMiao Xie <miaox@cn.fujitsu.com>
Reviewed-by: NJan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

A user reported a panic while running a balance.  What was happening was he was
relocating a block, which added the reference to the relocation tree.  Then
relocation would walk through the relocation tree and drop that reference and
free that block, and then it would walk down a snapshot which referenced the
same block and add another ref to the block.  The problem is this was all
happening in the same transaction, so the parent block was free'ed up when we
drop our reference which was immediately available for allocation, and then it
was used _again_ to add a reference for the same block from a different
snapshot.  This resulted in something like this in the delayed ref tree

add ref to 90234880, parent=2067398656, ref_root 1766, level 1
del ref to 90234880, parent=2067398656, ref_root 18446744073709551608, level 1
add ref to 90234880, parent=2067398656, ref_root 1767, level 1

as you can see the ref_root's don't match, because when we inc the ref we use
the header owner, which is the original tree the block belonged to, instead of
the data reloc tree.  Then when we remove the extent we use the reloc tree
objectid.  But none of this matters, since it is a shared reference which means
only the parent matters.  When the delayed ref stuff runs it adds all the
increments first, and then does all the drops, to make sure that we don't delete
the ref if we net a positive ref count.  But tree blocks aren't allowed to have
multiple refs from the same block, so this panics when it tries to add the
second ref.  We need the add and the drop to cancel each other out in memory so
we only do the final add.

So to fix this we need to adjust how the delayed refs are added to the tree.
Only the ref_root matters when it is a normal backref, and only the parent
matters when it is a shared backref.  So make our decision based on what ref
type we have.  This allows us to keep the ref_root in memory in case anybody
wants to use it for something else, and it allows the delayed refs to be merged
properly so we don't end up with this panic.

With this patch the users image no longer panics on mount, and it has a clean
fsck after a normal mount/umount cycle.  Thanks,

Cc: stable@vger.kernel.org
Reported-by: NRoman Mamedov <rm@romanrm.ru>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Testing my enospc log code I managed to abort a transaction during mount, which
put me into an infinite loop.  This is because of two things, first we don't
reset trans_no_join if we abort during transaction commit, which will force
anybody trying to start a transaction to just loop endlessly waiting for it to
be set to 0.  But this is still just a symptom, the second issue is we don't set
the fs state to error during errors on mount.  This is because we don't want to
do the flip read only thing during mount, but we still really want to set the fs
state to an error to keep us from even getting to the trans_no_join check.  So
fix both of these things, make sure to reset trans_no_join if we abort during a
commit, and make sure we set the fs state to error no matter if we're mounting
or not.  This should keep us from getting into this infinite loop again.
Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Steps to reproduce:
	mkfs.btrfs <disk>
	mount <disk> <mnt>
	btrfs quota enable <mnt>
	btrfs sub create <mnt>/subv

	i=1
	while [ $i -le 10000 ]
	do
		dd if=/dev/zero of=<mnt>/subv/data_$i bs=1K count=1
		i=$(($i+1))
		if [ $i -eq 500 ]
		then
			btrfs quota disable $mnt
		fi
	done
	dmesg
Obviously, this warn_on() is unnecessary, and it will be easily triggered.
Just remove it.
Signed-off-by: NWang Shilong <wangsl-fnst@cn.fujitsu.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

set_extent_bit()'s (u64 *failed_start) expects NULL not 0.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Document all current btrfs mount options.
Signed-off-by: NEric Sandeen <sandeen@redhat.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The subvolume ioctls block on the parent directory mutex that can be
held by other concurrent snapshot activity for a long time. Give the
user at least some chance to get out of this situation by allowing
to send a kill signal.
Signed-off-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Signed-off-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The messages

  btrfs: unlinked 123 orphans
  btrfs: truncated 456 orphans

are not useful to regular users and raise questions whether there are
problems with the filesystem.
Signed-off-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

This mount option was a workaround when subvol= assumed path relative
to the default subvolume, not the toplevel one. This was fixed long time
ago and subvolrootid has no effect.
Signed-off-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

With more than one btrfs volume mounted, it can be very difficult to find
out which volume is hitting an error. btrfs_error() will print this, but
it is currently rigged as more of a fatal error handler, while many of
the printk()s are currently for debugging and yet-unhandled cases.

This patch just changes the functions where the device information is
already available. Some cases remain where the root or fs_info is not
passed to the function emitting the error.

This may introduce some confusion with volumes backed by multiple devices
emitting errors referring to the primary device in the set instead of the
one on which the error occurred.

Use btrfs_printk(fs_info, format, ...) rather than writing the device
string every time, and introduce macro wrappers ala XFS for brevity.
Since the function already cannot be used for continuations, print a
newline as part of the btrfs_printk() message rather than at each caller.
Signed-off-by: NSimon Kirby <sim@hostway.ca>
Reviewed-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

The Kconfig title does not make much sense after the cleanup of
CONFIG_EXPERIMENTAL option, align the wording with other filesystems.
Signed-off-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Each time pick one dead root from the list and let the caller know if
it's needed to continue. This should improve responsiveness during
umount and balance which at some point waits for cleaning all currently
queued dead roots.

A new dead root is added to the end of the list, so the snapshots
disappear in the order of deletion.

The snapshot cleaning work is now done only from the cleaner thread and the
others wake it if needed.
Signed-off-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Signed-off-by: NZhi Yong Wu <wuzhy@linux.vnet.ibm.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Signed-off-by: NZhi Yong Wu <wuzhy@linux.vnet.ibm.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

Share the exactly same code of stopping workers.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

We currently store the first key of the tree block inside the reference for the
tree block in the extent tree.  This takes up quite a bit of space.  Make a new
key type for metadata which holds the level as the offset and completely removes
storing the btrfs_tree_block_info inside the extent ref.  This reduces the size
from 51 bytes to 33 bytes per extent reference for each tree block.  In practice
this results in a 30-35% decrease in the size of our extent tree, which means we
COW less and can keep more of the extent tree in memory which makes our heavy
metadata operations go much faster.  This is not an automatic format change, you
must enable it at mkfs time or with btrfstune.  This patch deals with having
metadata stored as either the old format or the new format so it is easy to
convert.  Thanks,
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>

free_root_pointers() has been introduced to cleanup all of tree roots,
so just use it instead.
Signed-off-by: NLiu Bo <bo.li.liu@oracle.com>
Reviewed-by: NDavid Sterba <dsterba@suse.cz>
Signed-off-by: NJosef Bacik <jbacik@fusionio.com>