We're spending huge amounts of time on lock contention during
end_io processing because we unconditionally assume we are overwriting
an existing extent in the file for each IO.

This checks to see if we are outside i_size, and if so, it uses a
less expensive readonly search of the btree to look for existing
extents.
Signed-off-by: Chris Mason <chris.mason@oracle.com>

Btrfs has an optimization where it will preallocate dentries during
readdir to fill in enough information to open the inode without an extra
lookup.

But, we're calling d_alloc, which is doing GFP_KERNEL allocations, and
that leads to deadlocks because our readdir code has tree locks held.

For now, disable this optimization.  We'll fix the gfp mask in the next
merge window.
Signed-off-by: Chris Mason <chris.mason@oracle.com>

Fix out-of-space checking, addressing a warning and potential resource
leak when resizing the filesystem down while allocating blocks.
Signed-off-by: Daniel J Blueman <daniel@quora.org>
Reviewed-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

may_commit_transaction() calls
        spin_lock(&space_info->lock);
        spin_lock(&delayed_rsv->lock);
and update_global_block_rsv() calls
        spin_lock(&block_rsv->lock);
        spin_lock(&sinfo->lock);

Lockdep complains about this at run time.
Everywhere except in update_global_block_rsv(), the space_info lock is
the outer lock, therefore the locking order in update_global_block_rsv()
is changed.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

I was seeing root_list corruption on unmount during fs resize in 3.4-rc4; add
correct locking to address this.
Signed-off-by: Daniel J Blueman <daniel@quora.org>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

btrfs_map_block sets mirror_num, so that the repair code knows eventually
which device gave us the read error. For RAID10, mirror_num must be 1 or 2.
Before this fix mirror_num was incorrectly related to our stripe index.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

btrfs_start_delalloc_inodes will just walk the list of delalloc inodes and
start writing them out, but it doesn't splice the list or anything so as
long as somebody is doing work on the box you could end up in this section
_forever_.  So just remove it, it's not needed anyway since sync will start
writeback on all inodes anyway, all we need to do is wait for ordered
extents and then we can commit the transaction.  In my horrible torture test
sync goes from taking 4 minutes to about 1.5 minutes.  Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

The bitfield member mount_opt was too small by one bit to hold the mount
option that enabled to include data extents in the integrity checker.
Since the same issue happened when the BTRFS_MOUNT_PANIC_ON_FATAL_ERROR
option was added (git rebase silently merges so that the increase of the
size of the bitfield member is lost), the bit limit was removed entirely.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>

Each CRC or header error was counted twice, this is now fixed.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>

When a filesystem is mounted with the degraded option, it is
possible that some of the devices are not there.
btrfs_ioctl_dev_info() crashs in this case because the device
name is a NULL pointer. This ioctl was only used for scrub.
Signed-off-by: Stefan Behrens <sbehrens@giantdisaster.de>

It is basically a good thing if we are interruptible when waiting for
free space, but the generality in which it is implemented currently
leads to system calls being interruptible that are not documented this
way. For example git can't handle interrupted unlink(), leading to
corrupt repos under space pressure.
Instead we raise the bar to only be interruptible by SIGKILL.
Thanks to David Sterba for suggesting this.
Signed-off-by: Arne Jansen <sensille@gmx.net>

The caller expects this function to return with the lock held and
releases it immediately on error.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>

A user reported a panic where we were trying to fix a bad mirror but the
mirror number we were giving was 0, which is invalid.  This is because we
don't do the transid verification until after the read, so as far as the
read code is concerned the read was a success.  So instead store the mirror
we read from so that if there is some failure post read we know which mirror
to try next and which mirror needs to be fixed if we find a good copy of the
block.  Thanks,
Signed-off-by: Josef Bacik <josef@redhat.com>

Free fs_devices as done in the error-handling code just below.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>

Cc: Jeff Mahoney <jeffm@suse.com>
Cc: Chris Mason <chris.mason@oracle.com>
Cc: Josef Bacik <josef@redhat.com>
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>

Fix a bug, where in case we need to adjust stripe_size so that the
length of the resulting chunk is less than or equal to max_chunk_size,
DUP chunks turn out to be only half as big as they could be.

Cc: Arne Jansen <sensille@gmx.net>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>

iref_to_path and iterate_irefs both increment the eb's refcount to use it
after releasing the path. Both depend on consistent data remaining in the
extent buffer and need a read lock to protect it.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>

Avoid calling free_extent_buffer more than once when the iterator function
returns non-zero. The only code that uses this is scrub repair for corrupted
nodatasum blocks.
Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>

Make free_ipath() behave like most other freeing functions in the
kernel and gracefully do nothing when passed a NULL pointer.

Besides this making the bahaviour consistent with functions such as
kfree(), vfree(), btrfs_free_path() etc etc, it also fixes a real NULL
deref issue in fs/btrfs/ioctl.c::btrfs_ioctl_ino_to_path(). In that
function we have this code:

...
        ipath = init_ipath(size, root, path);
        if (IS_ERR(ipath)) {
                ret = PTR_ERR(ipath);
                ipath = NULL;
                goto out;
        }
...
out:
        btrfs_free_path(path);
        free_ipath(ipath);
...

If we ever take the true branch of that 'if' statement we'll end up
passing a NULL pointer to free_ipath() which will subsequently
dereference it and we'll go "Boom" :-(
This patch will avoid that.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>

clear_extent_bit()
{
    next_node = rb_next(&state->rb_node);
    ...
    clear_state_bit(state);  <-- this may free next_node
    if (next_node) {
        state = rb_entry(next_node);
        ...
    }
}

clear_state_bit() calls merge_state() which may free the next node
of the passing extent_state, so clear_extent_bit() may end up
referencing freed memory.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>

Currently it returns a set of bits that were cleared, but this return
value is not used at all.

Moreover it doesn't seem to be useful, because we may clear the bits
of a few extent_states, but only the cleared bits of last one is
returned.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>

Added in commit 49b25e05
("btrfs: enhance transaction abort infrastructure")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.cz>

Our code is not ready to cope with a sectorsize that's not equal to PAGE_SIZE.
It will lead to hanging-on while writing something.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>

Normally when there are 2 copies of a block, we add both to the
reada extent tree and prefetch only the one that is easier to reach.
This way we can better utilize multiple devices.
In case of DUP this makes no sense as both copies reside on the
same device.
Signed-off-by: Arne Jansen <sensille@gmx.net>

When inserting into the radix tree returns EEXIST, get the existing
entry without giving up the spinlock in between.
There was a race for both the zones trees and the extent tree.
Signed-off-by: Arne Jansen <sensille@gmx.net>

Follow those instructions, and you'll trigger a warning in the
beginning of d_set_d_op():

  # mkfs.btrfs /dev/loop3
  # mount /dev/loop3 /mnt
  # btrfs sub create /mnt/sub
  # btrfs sub snap /mnt /mnt/snap
  # touch /mnt/snap/sub
  touch: cannot touch `tmp': Permission denied

__d_alloc() set d_op to sb->s_d_op (btrfs_dentry_operations), and
then simple_lookup() reset it to simple_dentry_operations, which
triggered the warning.
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>

A user reported that booting his box up with btrfs root on 3.4 was way
slower than on 3.3 because I removed the ideal caching code.  It turns out
that we don't load the free space cache if we're in a commit for deadlock
reasons, but since we're reading the cache and it hasn't changed yet we are
safe reading the inode and free space item from the commit root, so do that
and remove all of the deadlock checks so we don't unnecessarily skip loading
the free space cache.  The user reported this fixed the slowness.  Thanks,
Tested-by: Calvin Walton <calvin.walton@kepstin.ca>
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

49b25e05 introduced a use-after-free bug
that caused spurious -EIO's to be returned.

Do the check before we free the transaction.

Cc: David Sterba <dsterba@suse.cz>
Cc: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: Dave Jones <davej@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

bio_alloc() has the possibility of returning NULL.
So, it is necessary to check the return value.
Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

This fixes a regression introduced by fc67c450.  spin_is_locked() always
returns 0 on UP kernels, which caused assert in get_restripe_target() to
be fired on every call from btrfs_reduce_alloc_profile() on UP systems.
Remove it completely for now, it's not clear if it's going to be needed
in future.
Reported-by: Bobby Powers <bobbypowers@gmail.com>
Reported-by: Mitch Harder <mitch.harder@sabayonlinux.org>
Tested-by: Mitch Harder <mitch.harder@sabayonlinux.org>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

We miscalculate the length of extents we're discarding, and it leads to
an eof of device.
Reported-by: Daniel Blueman <daniel@quora.org>
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

We'd have to be passing bogus extent buffers for this uninit variable to
actually be used, but set it to zero just in case.
Signed-off-by: Chris Mason <chris.mason@oracle.com>

This reverts commit 5500cdbe.

We've had a number of complaints of early enospc that bisect down
to this patch.  We'll hae to fix the reservations differently.

CC: stable@kernel.org
Signed-off-by: Chris Mason <chris.mason@oracle.com>

Dave Sterba had put in patches to look for mixed data/metadata groups
with metadata bigger than 4KB.  But these ended up in the wrong place
and it wasn't testing the feature flag correctly.

This updates the tests to make sure our sizes are matching
Signed-off-by: Chris Mason <chris.mason@oracle.com>

When we use autodefrag, we forget to update the index which indicates
the last page we've dirty.  And we'll set dirty flags on a same set of
pages again and again.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

$ mkfs.btrfs /dev/sdb7
$ mount /dev/sdb7 /mnt/btrfs/ -oautodefrag
$ dd if=/dev/zero of=/mnt/btrfs/foobar bs=4k count=10 oflag=direct 2>/dev/null
$ filefrag -v /mnt/btrfs/foobar
Filesystem type is: 9123683e
File size of /mnt/btrfs/foobar is 40960 (10 blocks, blocksize 4096)
 ext logical physical expected length flags
   0       0     3072              10 eof
/mnt/btrfs/foobar: 1 extent found

Now we have a big real extent [0, 40960), but autodefrag will still defrag it.

$ sync
$ filefrag -v /mnt/btrfs/foobar
Filesystem type is: 9123683e
File size of /mnt/btrfs/foobar is 40960 (10 blocks, blocksize 4096)
 ext logical physical expected length flags
   0       0     3082              10 eof
/mnt/btrfs/foobar: 1 extent found

So if we already find a big real extent, we're ok about that, just skip it.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

If our file's layout is as follows:
| hole | data1 | hole | data2 |

we do not need to defrag this file, because this file has holes and
cannot be merged into one extent.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

$ mkfs.btrfs disk
$ mount disk /mnt -o autodefrag
$ dd if=/dev/zero of=/mnt/foobar bs=4k count=10 2>/dev/null && sync
$ for i in `seq 9 -2 0`; do dd if=/dev/zero of=/mnt/foobar bs=4k count=1 \
  seek=$i conv=notrunc 2> /dev/null; done && sync

then we'll get to defrag "foobar" again and again.
So does option "-o autodefrag,compress".

Reasons:
When the cleaner kthread gets to fetch inodes from the defrag tree and defrag
them, it will dirty pages and submit them, this will comes to another DATA COW
where the processing inode will be inserted to the defrag tree again.

This patch sets a rule for COW code, i.e. insert an inode when we're really
going to make some defragments.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

commit 600a45e1
(Btrfs: fix deadlock on page lock when doing auto-defragment)
fixes the deadlock on page, but it also introduces another bug.

A page may have been truncated after unlock & lock.
So we need to find it again to get the right one.

And since we've held i_mutex lock, inode size remains unchanged and
we can drop isize overflow checks.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>

The bug is from running xfstests 209 with autodefrag.

The race is as follows:
       t1                       t2(autodefrag)
   direct IO
     invalidate pagecache
     dio(old data)             add_inode_defrag
     invalidate pagecache
   endio

   direct IO
     invalidate pagecache
                                run_defrag
                                  readpage(old data)
                                  set page dirty (old data)
     dio(new data, rewrite)
     invalidate pagecache (*)
     endio

t2(autodefrag) will get old data into pagecache via readpage and set
pagecache dirty.  Meanwhile, invalidate pagecache(*) will fail due to
dirty flags in pages.  So the old data may be flushed into disk by
flush thread, which will lead to data loss.

And so does the case of user defragment progs.

The patch fixes this race by holding i_mutex when we readpage and set page dirty.
Signed-off-by: Liu Bo <liubo2009@cn.fujitsu.com>
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>