For btrfs raid, while discarding a range of space, we'll need to know
the start offset and length to discard for each device, and it's done
in btrfs_map_block().

However the calculation is a bit complex for raid0 and raid10, so I
reimplement it based on a fact that:

        dev1          dev2           dev3    (raid0)
        -----------------------------------
        s0 s3 s6      s1 s4 s7       s2 s5

Each device has (total_stripes / nr_dev) stripes, or plus one.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

We pre-allocate a btrfs bio with fixed size, and then may re-allocate
memory if we find stripes are bigger than the fixed size. But this
pre-allocation is not necessary.

Also we don't have to calcuate the stripe number twice.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

Some functions never use the transaction handle passed to them.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

Check and reserve space for btrfs_update_inode().
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

We can recover from errors and return -errno to user space.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

It can return -ENOMEM.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

If we run into some failure path in io_ctl_prepare_pages(),
io_ctl->pages[] array may have some NULL pointers.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

I got this while running xfstests:

[24256.836098] block group 317849600 has an wrong amount of free space
[24256.836100] btrfs: failed to load free space cache for block group 317849600

We should clamp the extent returned by find_first_extent_bit(),
so the start of the extent won't smaller than the start of the
block group.
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

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

This closes races where btrfs is calling d_instantiate too soon during
inode creation.  All of the callers of btrfs_add_nondir are updated to
instantiate after the inode is fully setup in memory.
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Dan Carpenter noticed that we were doing a double unlock on the worker
lock, and sometimes picking a worker thread without the lock held.

This fixes both errors.
Signed-off-by: NChris Mason <chris.mason@oracle.com>
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>

The btrfs io submission threads can build up massive plug lists.  This
keeps things more reasonable so we don't hand over huge dumps of IO at
once.
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Merge branch 'for-chris' of http://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-work into integration

Conflicts:
	fs/btrfs/inode.c
Signed-off-by: NChris Mason <chris.mason@oracle.com>

btrfs_update_inode is sometimes called with a null reservation.
Signed-off-by: NChris Mason <chris.mason@oracle.com>

A user reported a problem booting into a new kernel with the old format inodes.
He was panicing in cow_file_range while writing out the inode cache.  This is
because if the block group is not cached we'll just skip writing out the cache,
however if it gets dirtied again in the same transaction and it finished caching
we'd go ahead and write it out, but since we set cache_generation to the transid
we think we've already truncated it and will just carry on, running into
cow_file_range and blowing up.  We need to make sure we only set
cache_generation if we've done the truncate.  The user tested this patch and
verified that the panic no longer occured.  Thanks,
Reported-and-Tested-by: NKlaus Bitto <klaus.bitto@gmail.com>
Signed-off-by: NJosef Bacik <josef@redhat.com>

I've been hitting this BUG_ON() in btrfs_orphan_add when running xfstest 269 in
a loop.  This is because we will add an orphan item, do the truncate, the
truncate will fail for whatever reason (*cough*ENOSPC*cough*) and then we're
left with an orphan item still in the fs.  Then we come back later to do another
truncate and it blows up because we already have an orphan item.  This is ok so
just fix the BUG_ON() to only BUG() if ret is not EEXIST.  Thanks,
Signed-off-by: NJosef Bacik <josef@redhat.com>

We were occasionaly leaking space when running xfstest 269.  This is because if
we failed to start the transaction in the truncate loop we'd just goto out, but
we need to break so that the inode is removed from the orphan list and the space
is properly freed.  Thanks,
Signed-off-by: NJosef Bacik <josef@redhat.com>

Running xfstests 269 with some tracing my scripts kept spitting out errors about
releasing bytes that we didn't actually have reserved.  This took me down a huge
rabbit hole and it turns out the way we deal with reserved_extents is wrong,
we need to only be setting it if the reservation succeeds, otherwise the free()
method will come in and unreserve space that isn't actually reserved yet, which
can lead to other warnings and such.  The math was all working out right in the
end, but it caused all sorts of other issues in addition to making my scripts
yell and scream and generally make it impossible for me to track down the
original issue I was looking for.  The other problem is with our error handling
in the reservation code.  There are two cases that we need to deal with

1) We raced with free.  In this case free won't free anything because csum_bytes
is modified before we dro the lock in our reservation path, so free rightly
doesn't release any space because the reservation code may be depending on that
reservation.  However if we fail, we need the reservation side to do the free at
that point since that space is no longer in use.  So as it stands the code was
doing this fine and it worked out, except in case #2

2) We don't race with free.  Nobody comes in and changes anything, and our
reservation fails.  In this case we didn't reserve anything anyway and we just
need to clean up csum_bytes but not free anything.  So we keep track of
csum_bytes before we drop the lock and if it hasn't changed we know we can just
decrement csum_bytes and carry on.

Because of the case where we can race with free()'s since we have to drop our
spin_lock to do the reservation, I'm going to serialize all reservations with
the i_mutex.  We already get this for free in the heavy use paths, truncate and
file write all hold the i_mutex, just needed to add it to page_mkwrite and
various ioctl/balance things.  With this patch my space leak scripts no longer
scream bloody murder.  Thanks,
Signed-off-by: NJosef Bacik <josef@redhat.com>

Now that we're properly keeping track of delayed inode space we've been getting
a lot of warnings out of btrfs_dirty_inode() when running xfstest 83.  This is
because a bunch of people call mark_inode_dirty, which is void so we can't
return ENOSPC.  This needs to be fixed in a few areas

1) file_update_time - this updates the mtime and such when writing to a file,
which will call mark_inode_dirty.  So copy file_update_time into btrfs so we can
call btrfs_dirty_inode directly and return an error if we get one appropriately.

2) fix symlinks to use btrfs_setattr for ->setattr.  For some reason we weren't
setting ->setattr for symlinks, even though we should have been.  This catches
one of the cases where we were getting errors in mark_inode_dirty.

3) Fix btrfs_setattr and btrfs_setsize to call btrfs_dirty_inode directly
instead of mark_inode_dirty.  This lets us return errors properly for truncate
and chown/anything related to setattr.

4) Add a new btrfs_fs_dirty_inode which will just call btrfs_dirty_inode and
print an error if we have one.  The only remaining user we can't control for
this is touch_atime(), but we don't really want to keep people from walking
down the tree if we don't have space to save the atime update, so just complain
but don't worry about it.

With this patch xfstests 83 complains a handful of times instead of hundreds of
times.  Thanks,
Signed-off-by: NJosef Bacik <josef@redhat.com>

Al pointed out we have some random problems with the way we account for
num_workers_starting in the async thread stuff.  First of all we need to make
sure to decrement num_workers_starting if we fail to start the worker, so make
__btrfs_start_workers do this.  Also fix __btrfs_start_workers so that it
doesn't call btrfs_stop_workers(), there is no point in stopping everybody if we
failed to create a worker.  Also check_pending_worker_creates needs to call
__btrfs_start_work in it's work function since it already increments
num_workers_starting.

People only start one worker at a time, so get rid of the num_workers argument
everywhere, and make btrfs_queue_worker a void since it will always succeed.
Thanks,
Signed-off-by: NJosef Bacik <josef@redhat.com>

The Smack LSM hook for security_d_instantiate checks
the inode's i_op->getxattr value to determine if the
containing filesystem supports extended attributes.
The BTRFS filesystem sets the inode's i_op value only
after it has instantiated the inode. This results in
Smack incorrectly giving new BTRFS inodes attributes
from the filesystem defaults on the assumption that
values can't be stored on the filesystem. This patch
moves the assignment of inode operation vectors ahead
of the calls to d_instantiate, letting Smack know that
the filesystem supports extended attributes. There
should be no impact on the performance or behavior of
BTRFS.
Signed-off-by: NCasey Schaufler <casey@schaufler-ca.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

If we have a constant stream of end_io completions or crc work,
we can hit softlockup messages from the async helper threads.  This
adds a cond_resched() into the loop to avoid them.
Signed-off-by: NChris Mason <chris.mason@oracle.com>

To reproduce the bug:

    # touch /mnt/tmp
    # stat /mnt/tmp | grep Change
    Change: 2011-12-09 09:32:23.412105981 +0800
    # chattr +i /mnt/tmp
    # stat /mnt/tmp | grep Change
    Change: 2011-12-09 09:32:43.198105295 +0800
    # umount /mnt
    # mount /dev/loop1 /mnt
    # stat /mnt/tmp | grep Change
    Change: 2011-12-09 09:32:23.412105981 +0800

We should update ctime of in-memory inode before calling
btrfs_update_inode().
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Since we have the free space caches, btrfs_orphan_cleanup also runs for
the tree_root. Unfortunately this also cleans up the orphans used to mark
subvol deletions in progress.

Currently if a subvol deletion gets interrupted twice by umount/mount, the
deletion will not be continued and the space permanently lost, though it
would be possible to write a tool to recover those lost subvol deletions.
This patch checks if the orphan belongs to a subvol (dead root) and skips
the deletion.
Signed-off-by: NArne Jansen <sensille@gmx.net>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

When we use raid0 as the data profile, df command may show us a very
inaccurate value of the available space, which may be much less than the
real one. It may make the users puzzled. Fix it by changing the calculation
of the available space, and making it be more similar to a fake chunk
allocation.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Btrfsck report errors after the 83th case of xfstests was run, The error
number is 400, it means the used disk space of the file is wrong.

The reason of this bug is that:
The file truncation may fail when the space of the file system is not enough,
and leave some file extents, whose offset are beyond the end of the files.
When we want to expand those files, we will drop those file extents, and
put in dummy file extents, and then we should update the i-node. But btrfs
forgets to do it.

This patch adds the forgotten i-node update.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Btrfsck report error 100 after the 83th case of xfstests was run, it means
the i_size of the file is wrong.

The reason of this bug is that:
Btrfs increased i_size of the file at the beginning, but it failed to expand
the file, and failed to update the i_size to the old size because there is no
enough space in the file system, so we found a wrong i_size.

This patch fixes this bug by updating the i_size just when we pass the file
expanding and get enough space to update i-node.
Signed-off-by: NMiao Xie <miaox@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

btrfs_end_bio checks the number of errors on a bio against the max
number of errors allowed before sending any EIOs up to the higher
levels.

If we got enough copies of the bio done for a given raid level, it is
supposed to clear the bio error flag and return success.

We have pointers to the original bio sent down by the higher layers and
pointers to any cloned bios we made for raid purposes.  If the original
bio happens to be the one that got an io error, but not the last one to
finish, it might not have the BIO_UPTODATE bit set.

Then, when the last bio does finish, we'll call bio_end_io on the
original bio.  It won't have the uptodate bit set and we'll end up
sending EIO to the higher layers.

We already had a check for this, it just was conditional on getting the
IO error on the very last bio.  Make the check unconditional so we eat
the EIOs properly.
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Drop spin lock in convert_extent_bit() when memory alloc fails,
otherwise, it will be a deadlock.
Signed-off-by: NLiu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

If we call ioctl(BTRFS_IOC_ADD_DEV) directly, we'll succeed in adding
a readonly device to a btrfs filesystem, and btrfs will write to
that device, emitting kernel errors:

[ 3109.833692] lost page write due to I/O error on loop2
[ 3109.833720] lost page write due to I/O error on loop2
...
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

When we find an existing cluster, we switch to its block group as the
current block group, possibly skipping multiple blocks in the process.
Furthermore, under heavy contention, multiple threads may fail to
allocate from a cluster and then release just-created clusters just to
proceed to create new ones in a different block group.

This patch tries to allocate from an existing cluster regardless of its
block group, and doesn't switch to that group, instead proceeding to
try to allocate a cluster from the group it was iterating before the
attempt.
Signed-off-by: NAlexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

If we reach LOOP_NO_EMPTY_SIZE, we won't even try to use a cluster that
others might have set up.  Odds are that there won't be one, but if
someone else succeeded in setting it up, we might as well use it, even
if we don't try to set up a cluster again.
Signed-off-by: NAlexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Commit 4a54c8c1 introduced raid-repair, killing the individual
readpage_io_failed_hook entries from inode.c and disk-io.c. Commit
4bb31e92 introduced new readahead code, adding a readpage_io_failed_hook to
disk-io.c.

The raid-repair commit had logic to disable raid-repair, if
readpage_io_failed_hook is set. Thus, the readahead commit effectively
disabled raid-repair for meta data.

This commit changes the logic to always attempt raid-repair when needed and
call the readpage_io_failed_hook in case raid-repair fails. This is much
more straight forward and should have been like that from the beginning.
Signed-off-by: NJan Schmidt <list.btrfs@jan-o-sch.net>
Reported-by: NStefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

If we don't have a cluster, don't bother trying to allocate from it,
jumping right away to the attempt to allocate a new cluster.
Signed-off-by: NAlexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

We test whether a block group has enough free space to hold the
requested block, but when we're doing clustered allocation, we can
save some cycles by testing whether it has enough room for the cluster
upfront, otherwise we end up attempting to set up a cluster and
failing.  Only in the NO_EMPTY_SIZE loop do we attempt an unclustered
allocation, and by then we'll have zeroed the cluster size, so this
patch won't stop us from using the block group as a last resort.
Signed-off-by: NAlexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

Instead of starting at zero (offset is always zero), request a cluster
starting at search_start, that denotes the beginning of the current
block group.
Signed-off-by: NAlexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

The field that indicates the size of the largest contiguous chunk of
free space in the cluster is not initialized when setting up bitmaps,
it's only increased when we find a larger contiguous chunk.  We end up
retaining a larger value than appropriate for highly-fragmented
clusters, which may cause pointless searches for large contiguous
groups, and even cause clusters that do not meet the density
requirements to be set up.
Signed-off-by: NAlexandre Oliva <oliva@lsd.ic.unicamp.br>
Signed-off-by: NChris Mason <chris.mason@oracle.com>

We're failing to create clusters with bitmaps because
setup_cluster_no_bitmap checks that the list is empty before inserting
the bitmap entry in the list for setup_cluster_bitmap, but the list
field is only initialized when it is restored from the on-disk free
space cache, or when it is written out to disk.

Besides a potential race condition due to the multiple use of the list
field, filesystem performance severely degrades over time: as we use
up all non-bitmap free extents, the try-to-set-up-cluster dance is
done at every metadata block allocation.  For every block group, we
fail to set up a cluster, and after failing on them all up to twice,
we fall back to the much slower unclustered allocation.

To make matters worse, before the unclustered allocation, we try to
create new block groups until we reach the 1% threshold, which
introduces additional bitmaps and thus block groups that we'll iterate
over at each metadata block request.

To reproduce this bug:

  # dd if=/dev/zero of=img bs=1M count=256
  # mkfs.btrfs img
  # losetup -r /dev/loop1 img
  # mount /dev/loop1 /mnt
  OOPS!!

It triggered BUG_ON(!nr_devices) in btrfs_calc_avail_data_space().

To fix this, instead of checking write-only devices, we check all open
deivces:

  # df -h /dev/loop1
  Filesystem            Size  Used Avail Use% Mounted on
  /dev/loop1            250M   28K  238M   1% /mnt
Signed-off-by: NLi Zefan <lizf@cn.fujitsu.com>

It seems overly harsh to fail a resize of a btrfs file system to the
same size when a shrink or grow would succeed.  User app GParted trips
over this error.  Allow it by bypassing the shrink or grow operation.
Signed-off-by: NMike Fleetwood <mike.fleetwood@googlemail.com>