Mark Fasheh [Thu, 5 Nov 2015 22:38:00 +0000 (14:38 -0800)]
btrfs: qgroup: account shared subtree during snapshot delete
Commit 0ed4792 ('btrfs: qgroup: Switch to new extent-oriented qgroup
mechanism.') removed our qgroup accounting during
btrfs_drop_snapshot(). Predictably, this results in qgroup numbers
going bad shortly after a snapshot is removed.
Fix this by adding a dirty extent record when we encounter extents during
our shared subtree walk. This effectively restores the functionality we had
with the original shared subtree walking code in 1152651 (btrfs: qgroup:
account shared subtrees during snapshot delete).
The idea with the original patch (and this one) is that shared subtrees can
get skipped during drop_snapshot. The shared subtree walk then allows us a
chance to visit those extents and add them to the qgroup work for later
processing. This ultimately makes the accounting for drop snapshot work.
The new qgroup code nicely handles all the other extents during the tree
walk via the ref dec/inc functions so we don't have to add actions beyond
what we had originally.
Signed-off-by: Mark Fasheh <mfasheh@suse.de> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Thu, 5 Nov 2015 22:37:58 +0000 (14:37 -0800)]
Btrfs: use btrfs_get_fs_root in resolve_indirect_ref
The backref code will look up the fs_root we're trying to resolve our indirect
refs for, unfortunately we use btrfs_read_fs_root_no_name, which returns -ENOENT
if the ref is 0. This isn't helpful for the qgroup stuff with snapshot delete
as it won't be able to search down the snapshot we are deleting, which will
cause us to miss roots. So use btrfs_get_fs_root and send false for check_ref
so we can always get the root we're looking for. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Mark Fasheh <mfasheh@suse.de> Signed-off-by: Chris Mason <clm@fb.com>
Justin Maggard [Fri, 6 Nov 2015 18:36:42 +0000 (10:36 -0800)]
btrfs: qgroup: fix quota disable during rescan
There's a race condition that leads to a NULL pointer dereference if you
disable quotas while a quota rescan is running. To fix this, we just need
to wait for the quota rescan worker to actually exit before tearing down
the quota structures.
Signed-off-by: Justin Maggard <jmaggard@netgear.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Mon, 23 Nov 2015 15:25:16 +0000 (15:25 +0000)]
Btrfs: fix race between cleaner kthread and space cache writeout
When a block group becomes unused and the cleaner kthread is currently
running, we can end up getting the current transaction aborted with error
-ENOENT when we try to commit the transaction, leading to the following
trace:
The following diagram illustrates a sequence of steps leading to this
problem:
CPU 1 CPU 2
<at transaction N>
adds bg A to list
fs_info->unused_bgs
adds bg B to list
fs_info->unused_bgs
<transaction kthread
commits transaction N
and wakes up the
cleaner kthread>
cleaner kthread
delete_unused_bgs()
sees bg A in list
fs_info->unused_bgs
btrfs_start_transaction()
<transaction N + 1 starts>
deletes bg A
update_block_group(bg C)
--> adds bg C to list
fs_info->unused_bgs
deletes bg B
sees bg C in the list
fs_info->unused_bgs
btrfs_remove_chunk(bg C)
btrfs_remove_block_group(bg C)
--> checks if the block group
is in a dirty list, and
because it isn't now, it
does nothing
--> the block group item
is deleted from the
extent tree
--> adds bg C to list
transaction->dirty_bgs
some task calls
btrfs_commit_transaction(t N + 1)
commit_cowonly_roots()
btrfs_write_dirty_block_groups()
--> sees bg C in cur_trans->dirty_bgs
--> calls write_one_cache_group()
which returns -ENOENT because
it did not find the block group
item in the extent tree
--> transaction aborte with -ENOENT
because write_one_cache_group()
returned that error
So fix this by adding a block group to the list of dirty block groups
before adding it to the list of unused block groups.
This happened on a stress test using fsstress plus concurrent calls to
fallocate 20G and truncate (releasing part of the space allocated with
fallocate).
Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Thu, 19 Nov 2015 11:45:48 +0000 (11:45 +0000)]
Btrfs: fix scrub preventing unused block groups from being deleted
Currently scrub can race with the cleaner kthread when the later attempts
to delete an unused block group, and the result is preventing the cleaner
kthread from ever deleting later the block group - unless the block group
becomes used and unused again. The following diagram illustrates that
race:
CPU 1 CPU 2
cleaner kthread
btrfs_delete_unused_bgs()
gets block group X from
fs_info->unused_bgs and
removes it from that list
scrub_enumerate_chunks()
searches device tree using
its commit root
finds device extent for
block group X
gets block group X from the tree
fs_info->block_group_cache_tree
(via btrfs_lookup_block_group())
sets bg X to RO
sees the block group is
already RO and therefore
doesn't delete it nor adds
it back to unused list
So fix this by making scrub add the block group again to the list of
unused block groups if the block group is still unused when it finished
scrubbing it and it hasn't been removed already.
Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Thu, 19 Nov 2015 10:57:20 +0000 (10:57 +0000)]
Btrfs: fix race between scrub and block group deletion
Scrub can race with the cleaner kthread deleting block groups that are
unused (and with relocation too) leading to a failure with error -EINVAL
that gets returned to user space.
The following diagram illustrates how it happens:
CPU 1 CPU 2
cleaner kthread
btrfs_delete_unused_bgs()
gets block group X from
fs_info->unused_bgs
sets block group to RO
btrfs_remove_chunk(bg X)
deletes device extents
scrub_enumerate_chunks()
searches device tree using
its commit root
finds device extent for
block group X
gets block group X from the tree
fs_info->block_group_cache_tree
(via btrfs_lookup_block_group())
sets bg X to RO (again)
btrfs_remove_block_group(bg X)
deletes block group from
fs_info->block_group_cache_tree
removes extent map from
fs_info->mapping_tree
scrub_chunk(offset X)
searches fs_info->mapping_tree
for extent map starting at
offset X
--> doesn't find any such
extent map
--> returns -EINVAL and scrub
errors out to userspace
with -EINVAL
Fix this by dealing with an extent map lookup failure as an indicator of
block group deletion.
Issue reproduced with fstest btrfs/071.
Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
This is because of unprotected use of rcu_dereference in
btrfs_scratch_superblocks. We can't add rcu locks around the whole
function because we read the superblock.
The fix will use the rcu string buffer directly without the rcu locking.
Thi is safe as the device will not go away in the meantime. We're
holding the device list mutexes.
Restructuring the code to narrow down the rcu section turned out to be
impossible, we need to call filp_open (through update_dev_time) on the
buffer and this could call kmalloc/__might_sleep. We could call kstrdup
with GFP_ATOMIC but it's not absolutely necessary.
Fixes: 12b1c2637b6e (Btrfs: enhance btrfs_scratch_superblock to scratch all superblocks) Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Reason:
When big data writen to fs, the whole free space will be allocated
for data chunk.
And operation as scrub need to set_block_ro(), and when there is
only one metadata chunk in system(or other metadata chunks
are all full), the function will try to allocate a new chunk,
and failed because no space in device.
Fix:
When set_block_ro failed for metadata chunk, it is not a problem
because scrub_lock paused commit_trancaction in same time, and
metadata are always cowed, so the on-the-fly writepages will not
write data into same place with scrub/replace.
Let replace continue in this case is no problem.
Tested by above script, and xfstests/011, plus 100 times xfstests/070.
Changelog v1->v2:
1: Add detail comments in source and commit-message.
2: Add dmesg detail into commit-message.
3: Limit return value of -ENOSPC to be passed.
All suggested by: Filipe Manana <fdmanana@gmail.com>
Suggested-by: Filipe Manana <fdmanana@gmail.com> Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
David Sterba [Mon, 16 Nov 2015 15:50:13 +0000 (16:50 +0100)]
btrfs: fix clashing number of the enhanced balance usage filter
I've accidentally picked an already used number for the enhanced usage
filter represented by BTRFS_BALANCE_ARGS_USAGE_RANGE, clashing with
BTRFS_BALANCE_ARGS_CONVERT. Introduced during the development phase,
no backward compatibility issues.
Reported-by: Holger Hoffstätte <holger.hoffstaette@googlemail.com> Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Fixes: bc3094673f22 ("btrfs: extend balance filter usage to take minimum and maximum") Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Fri, 13 Nov 2015 23:57:17 +0000 (23:57 +0000)]
Btrfs: fix the number of transaction units needed to remove a block group
We were using only 1 transaction unit when attempting to delete an unused
block group but in reality we need 3 + N units, where N corresponds to the
number of stripes. We were accounting only for the addition of the orphan
item (for the block group's free space cache inode) but we were not
accounting that we need to delete one block group item from the extent
tree, one free space item from the tree of tree roots and N device extent
items from the device tree.
While one unit is not enough, it worked most of the time because for each
single unit we are too pessimistic and assume an entire tree path, with
the highest possible heigth (8), needs to be COWed with eventual node
splits at every possible level in the tree, so there was usually enough
reserved space for removing all the items and adding the orphan item.
However after adding the orphan item, writepages() can by called by the VM
subsystem against the btree inode when we are under memory pressure, which
causes writeback to start for the nodes we COWed before, this forces the
operation to remove the free space item to COW again some (or all of) the
same nodes (in the tree of tree roots). Even without writepages() being
called, we could fail with ENOSPC because these items are located in
multiple trees and one of them might have a higher heigth and require
node/leaf splits at many levels, exhausting all the reserved space before
removing all the items and adding the orphan.
In the kernel 4.0 release, commit 3d84be799194 ("Btrfs: fix BUG_ON in
btrfs_orphan_add() when delete unused block group"), we attempted to fix
a BUG_ON due to ENOSPC when trying to add the orphan item by making the
cleaner kthread reserve one transaction unit before attempting to remove
the block group, but this was not enough. We had a couple user reports
still hitting the same BUG_ON after 4.0, like Stefan Priebe's report on
a 4.2-rc6 kernel for example:
So fix this by reserving all the necessary units of metadata.
Reported-by: Stefan Priebe <s.priebe@profihost.ag> Fixes: 3d84be799194 ("Btrfs: fix BUG_ON in btrfs_orphan_add() when delete unused block group") Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Fri, 13 Nov 2015 23:57:16 +0000 (23:57 +0000)]
Btrfs: use global reserve when deleting unused block group after ENOSPC
It's possible to reach a state where the cleaner kthread isn't able to
start a transaction to delete an unused block group due to lack of enough
free metadata space and due to lack of unallocated device space to allocate
a new metadata block group as well. If this happens try to use space from
the global block group reserve just like we do for unlink operations, so
that we don't reach a permanent state where starting a transaction for
filesystem operations (file creation, renames, etc) keeps failing with
-ENOSPC. Such an unfortunate state was observed on a machine where over
a dozen unused data block groups existed and the cleaner kthread was
failing to delete them due to ENOSPC error when attempting to start a
transaction, and even running balance with a -dusage=0 filter failed with
ENOSPC as well. Also unmounting and mounting again the filesystem didn't
help. Allowing the cleaner kthread to use the global block reserve to
delete the unused data block groups fixed the problem.
Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Jeff Mahoney <jeffm@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
The fsync call passes 0 and LLONG_MAX, the range length does not fit to
loff_t and overflows, but the value is converted to u64 so it silently
works as expected.
The minimal fix is a typecast to u64, switching functions to take
(start, end) instead of (start, len) would be more intrusive.
Coccinelle script found that there's one more opencoded calculation of
the length.
mount -o nospace_cache "$TEST_DEV" "$TEST_DIR"
umount "$TEST_DEV"
mount -o nospace_cache "$TEST_DEV" "$TEST_DIR"
btrfs filesystem usage $TEST_DIR
We can see the data chunk changed from raid1 to single:
# btrfs filesystem usage $TEST_DIR
Data,single: Size:8.00MiB, Used:0.00B
/dev/vdg 8.00MiB
#
Reason:
When a empty filesystem mount with -o nospace_cache, the last
data blockgroup will be auto-removed in umount.
Then if we mount it again, there is no data chunk in the
filesystem, so the only available data profile is 0x0, result
is all new chunks are created as single type.
Fix:
Don't auto-delete last blockgroup for a raid type.
Test:
Test by above script, and confirmed the logic by debug output.
Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Zhao Lei [Wed, 19 Aug 2015 14:39:18 +0000 (22:39 +0800)]
btrfs: scrub: setup all fields for sblock_to_check
scrub_setup_recheck_block() isn't setup all necessary fields for
sblock_to_check because history reason.
So current code need more arguments in severial functions,
and more local variables, just to passing these lacked values to
necessary place.
This patch setup above fields to sblock_to_check in
scrub_setup_recheck_block(), for:
1: more cleanup for function arg, local variable
2: to make sblock_to_check complete, then we can use sblock_to_check
without concern about some uninitialized member.
Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Zhao Lei [Tue, 25 Aug 2015 13:31:40 +0000 (21:31 +0800)]
btrfs: scrub: set error stats when tree block spanning stripes
It is better to show error stats to user when we found tree block
spanning stripes.
On a btrfs created by old version of btrfs-convert:
Before patch:
# btrfs scrub start -B /dev/vdh
scrub done for 8b342d35-2904-41ab-b3cb-2f929709cf47
scrub started at Tue Aug 25 21:19:09 2015 and finished after 00:00:00
total bytes scrubbed: 53.54MiB with 0 errors
# dmesg
...
[ 128.711434] BTRFS error (device vdh): scrub: tree block 27054080 spanning stripes, ignored. logical=27000832
[ 128.712744] BTRFS error (device vdh): scrub: tree block 27054080 spanning stripes, ignored. logical=27066368
...
After patch:
# btrfs scrub start -B /dev/vdh
scrub done for ff7f844b-7a4e-4b1a-88a9-8252ab25be1b
scrub started at Tue Aug 25 21:42:29 2015 and finished after 00:00:00
total bytes scrubbed: 53.60MiB with 2 errors
error details:
corrected errors: 0, uncorrectable errors: 2, unverified errors: 0
ERROR: There are uncorrectable errors.
# dmesg
...omit...
#
Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Mon, 9 Nov 2015 18:06:38 +0000 (18:06 +0000)]
Btrfs: fix race when listing an inode's xattrs
When listing a inode's xattrs we have a time window where we race against
a concurrent operation for adding a new hard link for our inode that makes
us not return any xattr to user space. In order for this to happen, the
first xattr of our inode needs to be at slot 0 of a leaf and the previous
leaf must still have room for an inode ref (or extref) item, and this can
happen because an inode's listxattrs callback does not lock the inode's
i_mutex (nor does the VFS does it for us), but adding a hard link to an
inode makes the VFS lock the inode's i_mutex before calling the inode's
link callback.
The race illustrated by the following sequence diagram is possible:
CPU 1 CPU 2
btrfs_listxattr()
searches for key (257 XATTR_ITEM 0)
gets path with path->nodes[0] == leaf X
and path->slots[0] == N
because path->slots[0] is >=
btrfs_header_nritems(leaf X), it calls
btrfs_next_leaf()
btrfs_next_leaf()
releases the path
adds key (257 INODE_REF 666)
to the end of leaf X (slot N),
and leaf X now has N + 1 items
searches for the key (257 INODE_REF 256),
with path->keep_locks == 1, because that
is the last key it saw in leaf X before
releasing the path
ends up at leaf X again and it verifies
that the key (257 INODE_REF 256) is no
longer the last key in leaf X, so it
returns with path->nodes[0] == leaf X
and path->slots[0] == N, pointing to
the new item with key (257 INODE_REF 666)
btrfs_listxattr's loop iteration sees that
the type of the key pointed by the path is
different from the type BTRFS_XATTR_ITEM_KEY
and so it breaks the loop and stops looking
for more xattr items
--> the application doesn't get any xattr
listed for our inode
So fix this by breaking the loop only if the key's type is greater than
BTRFS_XATTR_ITEM_KEY and skip the current key if its type is smaller.
Cc: stable@vger.kernel.org Signed-off-by: Filipe Manana <fdmanana@suse.com>
Filipe Manana [Mon, 9 Nov 2015 00:33:58 +0000 (00:33 +0000)]
Btrfs: fix race leading to BUG_ON when running delalloc for nodatacow
If we are using the NO_HOLES feature, we have a tiny time window when
running delalloc for a nodatacow inode where we can race with a concurrent
link or xattr add operation leading to a BUG_ON.
This happens because at run_delalloc_nocow() we end up casting a leaf item
of type BTRFS_INODE_[REF|EXTREF]_KEY or of type BTRFS_XATTR_ITEM_KEY to a
file extent item (struct btrfs_file_extent_item) and then analyse its
extent type field, which won't match any of the expected extent types
(values BTRFS_FILE_EXTENT_[REG|PREALLOC|INLINE]) and therefore trigger an
explicit BUG_ON(1).
The following sequence diagram shows how the race happens when running a
no-cow dellaloc range [4K, 8K[ for inode 257 and we have the following
neighbour leafs:
(Note the implicit hole for inode 257 regarding the [0, 8K[ range)
CPU 1 CPU 2
run_dealloc_nocow()
btrfs_lookup_file_extent()
--> searches for a key with value
(257 EXTENT_DATA 4096) in the
fs/subvol tree
--> returns us a path with
path->nodes[0] == leaf X and
path->slots[0] == N
because path->slots[0] is >=
btrfs_header_nritems(leaf X), it
calls btrfs_next_leaf()
btrfs_next_leaf()
--> releases the path
hard link added to our inode,
with key (257 INODE_REF 500)
added to the end of leaf X,
so leaf X now has N + 1 keys
--> searches for the key
(257 INODE_REF 256), because
it was the last key in leaf X
before it released the path,
with path->keep_locks set to 1
--> ends up at leaf X again and
it verifies that the key
(257 INODE_REF 256) is no longer
the last key in the leaf, so it
returns with path->nodes[0] ==
leaf X and path->slots[0] == N,
pointing to the new item with
key (257 INODE_REF 500)
the loop iteration of run_dealloc_nocow()
does not break out the loop and continues
because the key referenced in the path
at path->nodes[0] and path->slots[0] is
for inode 257, its type is < BTRFS_EXTENT_DATA_KEY
and its offset (500) is less then our delalloc
range's end (8192)
the item pointed by the path, an inode reference item,
is (incorrectly) interpreted as a file extent item and
we get an invalid extent type, leading to the BUG_ON(1):
This happened because the item we were processing did not match a file
extent item (its key type != BTRFS_EXTENT_DATA_KEY), and even on this
case we cast the item to a struct btrfs_file_extent_item pointer and
then find a type field value that does not match any of the expected
values (BTRFS_FILE_EXTENT_[REG|PREALLOC|INLINE]). This scenario happens
due to a tiny time window where a race can happen as exemplified below.
For example, consider the following scenario where we're using the
NO_HOLES feature and we have the following two neighbour leafs:
Our inode 257 has an implicit hole in the range [0, 8K[ (implicit rather
than explicit because NO_HOLES is enabled). Now if our inode has an
ordered extent for the range [4K, 8K[ that is finishing, the following
can happen:
CPU 1 CPU 2
btrfs_finish_ordered_io()
insert_reserved_file_extent()
__btrfs_drop_extents()
Searches for the key
(257 EXTENT_DATA 4096) through
btrfs_lookup_file_extent()
Key not found and we get a path where
path->nodes[0] == leaf X and
path->slots[0] == N
Because path->slots[0] is >=
btrfs_header_nritems(leaf X), we call
btrfs_next_leaf()
btrfs_next_leaf() releases the path
inserts key
(257 INODE_REF 4096)
at the end of leaf X,
leaf X now has N + 1 keys,
and the new key is at
slot N
btrfs_next_leaf() searches for
key (257 INODE_REF 256), with
path->keep_locks set to 1,
because it was the last key it
saw in leaf X
finds it in leaf X again and
notices it's no longer the last
key of the leaf, so it returns 0
with path->nodes[0] == leaf X and
path->slots[0] == N (which is now
< btrfs_header_nritems(leaf X)),
pointing to the new key
(257 INODE_REF 4096)
__btrfs_drop_extents() casts the
item at path->nodes[0], slot
path->slots[0], to a struct
btrfs_file_extent_item - it does
not skip keys for the target
inode with a type less than
BTRFS_EXTENT_DATA_KEY
(BTRFS_INODE_REF_KEY < BTRFS_EXTENT_DATA_KEY)
sees a bogus value for the type
field triggering the WARN_ON in
the trace shown above, and sets
extent_end = search_start (4096)
does the if-then-else logic to
fixup 0 length extent items created
by a past bug from hole punching:
if (extent_end == key.offset &&
extent_end >= search_start)
goto delete_extent_item;
that evaluates to true and it ends
up deleting the key pointed to by
path->slots[0], (257 INODE_REF 4096),
from leaf X
The same could happen for example for a xattr that ends up having a key
with an offset value that matches search_start (very unlikely but not
impossible).
So fix this by ensuring that keys smaller than BTRFS_EXTENT_DATA_KEY are
skipped, never casted to struct btrfs_file_extent_item and never deleted
by accident. Also protect against the unexpected case of getting a key
for a lower inode number by skipping that key and issuing a warning.
Cc: stable@vger.kernel.org Signed-off-by: Filipe Manana <fdmanana@suse.com>
Filipe Manana [Thu, 5 Nov 2015 10:51:37 +0000 (10:51 +0000)]
Btrfs: fix sleeping inside atomic context in qgroup rescan worker
We are holding a btree path with spinning locks and then we attempt to
clone an extent buffer, which calls kmem_cache_alloc() and this function
can sleep, causing the following trace to be reported on a debug kernel:
[107118.218536] BUG: sleeping function called from invalid context at mm/slab.c:2871
[107118.224110] in_atomic(): 1, irqs_disabled(): 0, pid: 19148, name: kworker/u32:3
[107118.226120] INFO: lockdep is turned off.
[107118.226843] Preemption disabled at:[<ffffffffa05ffa22>] btrfs_clear_lock_blocking_rw+0x96/0xea [btrfs]
Filipe Manana [Thu, 5 Nov 2015 10:06:23 +0000 (10:06 +0000)]
Btrfs: fix race waiting for qgroup rescan worker
We were initializing the completion (fs_info->qgroup_rescan_completion)
object after releasing the qgroup rescan lock, which gives a small time
window for a rescan waiter to not actually wait for the rescan worker
to finish. Example:
CPU 1 CPU 2
fs_info->qgroup_rescan_completion->done is 0
btrfs_qgroup_rescan_worker()
complete_all(&fs_info->qgroup_rescan_completion)
sets fs_info->qgroup_rescan_completion->done
to UINT_MAX / 2
... do some other stuff ....
qgroup_rescan_init()
mutex_lock(&fs_info->qgroup_rescan_lock)
set flag BTRFS_QGROUP_STATUS_FLAG_RESCAN
in fs_info->qgroup_flags
mutex_unlock(&fs_info->qgroup_rescan_lock)
btrfs_qgroup_wait_for_completion()
mutex_lock(&fs_info->qgroup_rescan_lock)
sees flag BTRFS_QGROUP_STATUS_FLAG_RESCAN
in fs_info->qgroup_flags
mutex_unlock(&fs_info->qgroup_rescan_lock)
Justin Maggard [Wed, 4 Nov 2015 23:56:16 +0000 (15:56 -0800)]
btrfs: qgroup: exit the rescan worker during umount
I was hitting a consistent NULL pointer dereference during shutdown that
showed the trace running through end_workqueue_bio(). I traced it back to
the endio_meta_workers workqueue being poked after it had already been
destroyed.
Eventually I found that the root cause was a qgroup rescan that was still
in progress while we were stopping all the btrfs workers.
Currently we explicitly pause balance and scrub operations in
close_ctree(), but we do nothing to stop the qgroup rescan. We should
probably be doing the same for qgroup rescan, but that's a much larger
change. This small change is good enough to allow me to unmount without
crashing.
Signed-off-by: Justin Maggard <jmaggard@netgear.com> Reviewed-by: Filipe Manana <fdmanana@suse.com>
Filipe Manana [Wed, 4 Nov 2015 09:52:04 +0000 (09:52 +0000)]
Btrfs: fix extent accounting for partial direct IO writes
When doing a write using direct IO we can end up not doing the whole write
operation using the direct IO path, in that case we fallback to a buffered
write to do the remaining IO. This happens for example if the range we are
writing to contains a compressed extent.
When we do a partial write and fallback to buffered IO, due to the
existence of a compressed extent for example, we end up not adjusting the
outstanding extents counter of our inode which ends up getting decremented
twice, once by the DIO ordered extent for the partial write and once again
by btrfs_direct_IO(), resulting in an arithmetic underflow at
extent-tree.c:drop_outstanding_extent(). For example if we have:
extents [ prealloc extent ] [ compressed extent ]
offsets A B C D E
and at the moment our inode's outstanding extents counter is 0, if we do a
direct IO write against the range [B, D[ (which has a length smaller than
128Mb), we end up bumping our inode's outstanding extents counter to 1, we
create a DIO ordered extent for the range [B, C[ and then fallback to a
buffered write for the range [C, D[. The direct IO handler
(inode.c:btrfs_direct_IO()) decrements the outstanding extents counter by
1, leaving it with a value of 0, through a call to
btrfs_delalloc_release_space() and then shortly after the DIO ordered
extent finishes and calls btrfs_delalloc_release_metadata() which ends
up to attempt to decrement the inode's outstanding extents counter by 1,
resulting in an assertion failure at drop_outstanding_extent() because
the operation would result in an arithmetic underflow (0 - 1). This
produces the following trace:
So fix this by ensuring we adjust the outstanding extents counter when we
do the fallback just like we do for the case where the whole write can be
done through the direct IO path.
We were also adjusting the outstanding extents counter by a constant value
of 1, which is incorrect because we were ignorning that we account extents
in BTRFS_MAX_EXTENT_SIZE units, o fix that as well.
The following test case for fstests reproduces this issue:
seq=`basename $0`
seqres=$RESULT_DIR/$seq
echo "QA output created by $seq"
tmp=/tmp/$$
status=1 # failure is the default!
trap "_cleanup; exit \$status" 0 1 2 3 15
_cleanup()
{
rm -f $tmp.*
}
# get standard environment, filters and checks
. ./common/rc
. ./common/filter
# real QA test starts here
_need_to_be_root
_supported_fs btrfs
_supported_os Linux
_require_scratch
_require_xfs_io_command "falloc"
# Create a compressed extent covering the range [700K, 800K[.
$XFS_IO_PROG -f -s -c "pwrite -S 0xaa -b 100K 700K 100K" \
$SCRATCH_MNT/foo | _filter_xfs_io
# Create prealloc extent covering the range [600K, 700K[.
$XFS_IO_PROG -c "falloc 600K 100K" $SCRATCH_MNT/foo
# Write 80K of data to the range [640K, 720K[ using direct IO. This
# range covers both the prealloc extent and the compressed extent.
# Because there's a compressed extent in the range we are writing to,
# the DIO write code path ends up only writing the first 60k of data,
# which goes to the prealloc extent, and then falls back to buffered IO
# for writing the remaining 20K of data - because that remaining data
# maps to a file range containing a compressed extent.
# When falling back to buffered IO, we used to trigger an assertion when
# releasing reserved space due to bad accounting of the inode's
# outstanding extents counter, which was set to 1 but we ended up
# decrementing it by 1 twice, once through the ordered extent for the
# 60K of data we wrote using direct IO, and once through the main direct
# IO handler (inode.cbtrfs_direct_IO()) because the direct IO write
# wrote less than 80K of data (60K).
$XFS_IO_PROG -d -c "pwrite -S 0xbb -b 80K 640K 80K" \
$SCRATCH_MNT/foo | _filter_xfs_io
# Now similar test as above but for very large write operations. This
# triggers special cases for an inode's outstanding extents accounting,
# as internally btrfs logically splits extents into 128Mb units.
$XFS_IO_PROG -f -s \
-c "pwrite -S 0xaa -b 128M 258M 128M" \
-c "falloc 0 258M" \
$SCRATCH_MNT/bar | _filter_xfs_io
$XFS_IO_PROG -d -c "pwrite -S 0xbb -b 256M 3M 256M" $SCRATCH_MNT/bar \
| _filter_xfs_io
# Now verify the file contents are correct and that they are the same
# even after unmounting and mounting the fs again (or evicting the page
# cache).
#
# For file foo, all bytes in the range [0, 640K[ must have a value of
# 0x00, all bytes in the range [640K, 720K[ must have a value of 0xbb
# and all bytes in the range [720K, 800K[ must have a value of 0xaa.
#
# For file bar, all bytes in the range [0, 3M[ must havea value of 0x00,
# all bytes in the range [3M, 259M[ must have a value of 0xbb and all
# bytes in the range [259M, 386M[ must have a value of 0xaa.
#
echo "File digests before remounting the file system:"
md5sum $SCRATCH_MNT/foo | _filter_scratch
md5sum $SCRATCH_MNT/bar | _filter_scratch
_scratch_remount
echo "File digests after remounting the file system:"
md5sum $SCRATCH_MNT/foo | _filter_scratch
md5sum $SCRATCH_MNT/bar | _filter_scratch
status=0
exit
Fixes: e1cbbfa5f5aa ("Btrfs: fix outstanding_extents accounting in DIO") Fixes: 3e05bde8c3c2 ("Btrfs: only adjust outstanding_extents when we do a short write") Signed-off-by: Filipe Manana <fdmanana@suse.com>
Filipe Manana [Mon, 2 Nov 2015 12:32:44 +0000 (12:32 +0000)]
Btrfs: fix hole punching when using the no-holes feature
When we are using the no-holes feature, if we punch a hole into a file
range that already contains a hole which overlaps the range we are passing
to fallocate(), we end up removing the extent map that represents the
existing hole without adding a new one. This happens because with the
no-holes feature we do not have explicit extent items to represent holes
and therefore the call to __btrfs_drop_extents(), made from
btrfs_punch_hole(), returns an end offset to the variable drop_end that
is smaller than the end of the range passed to fallocate(), while it
drops all existing extent maps in that range.
Normally having a missing extent map is not a problem, for example for
a readpages() operation we just end up building the extent map by
looking at the fs/subvol tree for a matching extent item (or a lack of
one for implicit holes). However for an fsync that uses the fast path,
which needs to look at the list of modified extent maps, this means
the fsync will not record information about the complete hole we had
before the fallocate() call into the log tree, resulting in a file with
content/layout that does not match what we had neither before nor after
the hole punch operation.
The following test case for fstests reproduces the issue. It fails without
this change because we get a file with a different digest after the fsync
log replay and also with a different extent/hole layout.
seq=`basename $0`
seqres=$RESULT_DIR/$seq
echo "QA output created by $seq"
tmp=/tmp/$$
status=1 # failure is the default!
trap "_cleanup; exit \$status" 0 1 2 3 15
_cleanup()
{
_cleanup_flakey
rm -f $tmp.*
}
# get standard environment, filters and checks
. ./common/rc
. ./common/filter
. ./common/punch
. ./common/dmflakey
# real QA test starts here
_need_to_be_root
_supported_fs generic
_supported_os Linux
_require_scratch
_require_xfs_io_command "fpunch"
_require_xfs_io_command "fiemap"
_require_dm_target flakey
_require_metadata_journaling $SCRATCH_DEV
# This test was motivated by an issue found in btrfs when the btrfs
# no-holes feature is enabled (introduced in kernel 3.14). So enable
# the feature if the fs being tested is btrfs.
if [ $FSTYP == "btrfs" ]; then
_require_btrfs_fs_feature "no_holes"
_require_btrfs_mkfs_feature "no-holes"
MKFS_OPTIONS="$MKFS_OPTIONS -O no-holes"
fi
# Create out test file with some data and then fsync it.
# We do the fsync only to make sure the last fsync we do in this test
# triggers the fast code path of btrfs' fsync implementation, a
# condition necessary to trigger the bug btrfs had.
$XFS_IO_PROG -f -c "pwrite -S 0xaa 0K 128K" \
-c "fsync" \
$SCRATCH_MNT/foobar | _filter_xfs_io
# Now punch a hole against the range [96K, 128K[.
$XFS_IO_PROG -c "fpunch 96K 32K" $SCRATCH_MNT/foobar
# Punch another hole against a range that overlaps the previous range
# and ends beyond eof.
$XFS_IO_PROG -c "fpunch 64K 128K" $SCRATCH_MNT/foobar
# Punch another hole against a range that overlaps the first range
# ([96K, 128K[) and ends at eof.
$XFS_IO_PROG -c "fpunch 32K 96K" $SCRATCH_MNT/foobar
# Fsync our file. We want to verify that, after a power failure and
# mounting the filesystem again, the file content reflects all the hole
# punch operations.
$XFS_IO_PROG -c "fsync" $SCRATCH_MNT/foobar
echo "File digest before power failure:"
md5sum $SCRATCH_MNT/foobar | _filter_scratch
echo "Fiemap before power failure:"
$XFS_IO_PROG -c "fiemap -v" $SCRATCH_MNT/foobar | _filter_fiemap
# Silently drop all writes and umount to simulate a crash/power failure.
_load_flakey_table $FLAKEY_DROP_WRITES
_unmount_flakey
# Allow writes again, mount to trigger log replay and validate file
# contents.
_load_flakey_table $FLAKEY_ALLOW_WRITES
_mount_flakey
echo "File digest after log replay:"
# Must match the same digest we got before the power failure.
md5sum $SCRATCH_MNT/foobar | _filter_scratch
echo "Fiemap after log replay:"
# Must match the same extent listing we got before the power failure.
$XFS_IO_PROG -c "fiemap -v" $SCRATCH_MNT/foobar | _filter_fiemap
_unmount_flakey
status=0
exit
Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
The above call trace is seen even on x86_64; albeit very rarely and that too
with nodesize set to 64k and with nospace_cache mount option being used.
The reason for the above call trace is,
btrfs_remove_chunk
check_system_chunk
Allocate chunk if required
For each physical stripe on underlying device,
btrfs_free_dev_extent
...
Take lock on Device tree's root node
btrfs_cow_block("dev tree's root node");
btrfs_reserve_extent
find_free_extent
index = BTRFS_RAID_DUP;
have_caching_bg = false;
When in LOOP_CACHING_NOWAIT state, Assume we find a block group
which is being cached; Hence have_caching_bg is set to true
When repeating the search for the next RAID index, we set
have_caching_bg to false.
Hence right after completing the LOOP_CACHING_NOWAIT state, we incorrectly
skip LOOP_CACHING_WAIT state and move to LOOP_ALLOC_CHUNK state where we
allocate a chunk and try to add entries corresponding to the chunk's physical
stripe into the device tree. When doing so the task deadlocks itself waiting
for the blocking lock on the root node of the device tree.
This commit fixes the issue by introducing a new local variable to help
indicate as to whether a block group of any RAID type is being cached.
Signed-off-by: Chandan Rajendra <chandan@linux.vnet.ibm.com> Reviewed-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Thu, 29 Oct 2015 09:28:46 +0000 (17:28 +0800)]
btrfs: Fix a data space underflow warning
Even with quota disabled, generic/127 will trigger a kernel warning by
underflow data space info.
The bug is caused by buffered write, which in case of short copy, the
start parameter for btrfs_delalloc_release_space() is wrong, and
round_up/down() in btrfs_delalloc_release() extents the range to page
aligned, decreasing one more page than expected.
This patch will fix it by passing correct start.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Mon, 26 Oct 2015 06:11:18 +0000 (14:11 +0800)]
btrfs: qgroup: Fix a race in delayed_ref which leads to abort trans
Between btrfs_allocerved_file_extent() and
btrfs_add_delayed_qgroup_reserve(), there is a window that delayed_refs
are run and delayed ref head maybe freed before
btrfs_add_delayed_qgroup_reserve().
This will cause btrfs_dad_delayed_qgroup_reserve() to return -ENOENT,
and cause transaction to be aborted.
This patch will record qgroup reserve space info into delayed_ref_head
at btrfs_add_delayed_ref(), to eliminate the race window.
Reported-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Jiri Kosina [Mon, 26 Oct 2015 06:06:19 +0000 (15:06 +0900)]
btrfs: clear PF_NOFREEZE in cleaner_kthread()
cleaner_kthread() kthread calls try_to_freeze() at the beginning of every
cleanup attempt. This operation can't ever succeed though, as the kthread
hasn't marked itself as freezable.
Before (hopefully eventually) kthread freezing gets converted to fileystem
freezing, we'd rather mark cleaner_kthread() freezable (as my
understanding is that it can generate filesystem I/O during suspend).
Signed-off-by: Jiri Kosina <jkosina@suse.cz> Signed-off-by: Chris Mason <clm@fb.com>
David Sterba [Fri, 16 Oct 2015 17:15:50 +0000 (19:15 +0200)]
btrfs: add balance filters limits, stripes and usage to supported mask
Enable the extended 'limit' syntax (a range), the new 'stripes' and
extended 'usage' syntax (a range) filters in the filters mask. The patch
comes separate and not within the series that introduced the new filters
because the patch adding the mask was merged in a late rc. The
integration branch was based on an older rc and could not merge the
patch due to the missing changes.
Prerequisities:
* btrfs: check unsupported filters in balance arguments
* btrfs: extend balance filter limit to take minimum and maximum
* btrfs: add balance filter for stripes
* btrfs: extend balance filter usage to take minimum and maximum
Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
David Sterba [Tue, 20 Oct 2015 16:22:13 +0000 (18:22 +0200)]
btrfs: extend balance filter usage to take minimum and maximum
Similar to the 'limit' filter, we can enhance the 'usage' filter to
accept a range. The change is backward compatible, the range is applied
only in connection with the BTRFS_BALANCE_ARGS_USAGE_RANGE flag.
We don't have a usecase yet, the current syntax has been sufficient. The
enhancement should provide parity with other range-like filters.
Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Balance block groups which have the given number of stripes, defined by
a range min..max. This is useful to selectively rebalance only chunks
that do not span enough devices, applies to RAID0/10/5/6.
Signed-off-by: Gabríel Arthúr Pétursson <gabriel@system.is>
[ renamed bargs members, added to the UAPI, wrote the changelog ] Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
David Sterba [Sat, 10 Oct 2015 15:16:50 +0000 (17:16 +0200)]
btrfs: extend balance filter limit to take minimum and maximum
The 'limit' filter is underdesigned, it should have been a range for
[min,max], with some relaxed semantics when one of the bounds is
missing. Besides that, using a full u64 for a single value is a waste of
bytes.
Let's fix both by extending the use of the u64 bytes for the [min,max]
range. This can be done in a backward compatible way, the range will be
interpreted only if the appropriate flag is set
(BTRFS_BALANCE_ARGS_LIMIT_RANGE).
Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Chris Mason [Tue, 13 Oct 2015 18:06:48 +0000 (14:06 -0400)]
btrfs: fix use after free iterating extrefs
The code for btrfs inode-resolve has never worked properly for
files with enough hard links to trigger extrefs. It was trying to
get the leaf out of a path after freeing the path:
The fix here is to use the extent buffer we cloned just a little higher
up to avoid deadlocks caused by using the leaf in the path.
Signed-off-by: Chris Mason <clm@fb.com>
cc: stable@vger.kernel.org # v3.7+
cc: Mark Fasheh <mfasheh@suse.de> Reviewed-by: Filipe Manana <fdmanana@suse.com> Reviewed-by: Mark Fasheh <mfasheh@suse.de> Signed-off-by: Chris Mason <clm@fb.com>
David Sterba [Mon, 12 Oct 2015 14:55:54 +0000 (16:55 +0200)]
btrfs: check unsupported filters in balance arguments
We don't verify that all the balance filter arguments supplemented by
the flags are actually known to the kernel. Thus we let it silently pass
and do nothing.
At the moment this means only the 'limit' filter, but we're going to add
a few more soon so it's better to have that fixed. Also in older stable
kernels so that it works with newer userspace tools.
Cc: stable@vger.kernel.org # 3.16+ Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Chris Mason <clm@fb.com>
Filipe Manana [Fri, 23 Oct 2015 06:52:54 +0000 (07:52 +0100)]
Btrfs: fix regression running delayed references when using qgroups
In the kernel 4.2 merge window we had a big changes to the implementation
of delayed references and qgroups which made the no_quota field of delayed
references not used anymore. More specifically the no_quota field is not
used anymore as of:
commit 0ed4792af0e8 ("btrfs: qgroup: Switch to new extent-oriented qgroup mechanism.")
Leaving the no_quota field actually prevents delayed references from
getting merged, which in turn cause the following BUG_ON(), at
fs/btrfs/extent-tree.c, to be hit when qgroups are enabled:
static int run_delayed_tree_ref(...)
{
(...)
BUG_ON(node->ref_mod != 1);
(...)
}
2) Ref2 bytenr X, action = BTRFS_DROP_DELAYED_REF, no_quota = 0, added.
It's not merged with Ref1 because Ref1->no_quota != Ref2->no_quota.
3) Ref3 bytenr X, action = BTRFS_ADD_DELAYED_REF, no_quota = 1, added.
It's not merged with the reference at the tail of the list of refs
for bytenr X because the reference at the tail, Ref2 is incompatible
due to Ref2->no_quota != Ref3->no_quota.
4) Ref4 bytenr X, action = BTRFS_DROP_DELAYED_REF, no_quota = 0, added.
It's not merged with the reference at the tail of the list of refs
for bytenr X because the reference at the tail, Ref3 is incompatible
due to Ref3->no_quota != Ref4->no_quota.
5) We run delayed references, trigger merging of delayed references,
through __btrfs_run_delayed_refs() -> btrfs_merge_delayed_refs().
6) Ref1 and Ref3 are merged as Ref1->no_quota = Ref3->no_quota and
all other conditions are satisfied too. So Ref1 gets a ref_mod
value of 2.
7) Ref2 and Ref4 are merged as Ref2->no_quota = Ref4->no_quota and
all other conditions are satisfied too. So Ref2 gets a ref_mod
value of 2.
8) Ref1 and Ref2 aren't merged, because they have different values
for their no_quota field.
9) Delayed reference Ref1 is picked for running (select_delayed_ref()
always prefers references with an action == BTRFS_ADD_DELAYED_REF).
So run_delayed_tree_ref() is called for Ref1 which triggers the
BUG_ON because Ref1->red_mod != 1 (equals 2).
So fix this by removing the no_quota field, as it's not used anymore as
of commit 0ed4792af0e8 ("btrfs: qgroup: Switch to new extent-oriented
qgroup mechanism.").
The use of no_quota was also buggy in at least two places:
1) At delayed-refs.c:btrfs_add_delayed_tree_ref() - we were setting
no_quota to 0 instead of 1 when the following condition was true:
is_fstree(ref_root) || !fs_info->quota_enabled
2) At extent-tree.c:__btrfs_inc_extent_ref() - we were attempting to
reset a node's no_quota when the condition "!is_fstree(root_objectid)
|| !root->fs_info->quota_enabled" was true but we did it only in
an unused local stack variable, that is, we never reset the no_quota
value in the node itself.
This fixes the remainder of problems several people have been having when
running delayed references, mostly while a balance is running in parallel,
on a 4.2+ kernel.
Very special thanks to Stéphane Lesimple for helping debugging this issue
and testing this fix on his multi terabyte filesystem (which took more
than one day to balance alone, plus fsck, etc).
Also, this fixes deadlock issue when using the clone ioctl with qgroups
enabled, as reported by Elias Probst in the mailing list. The deadlock
happens because after calling btrfs_insert_empty_item we have our path
holding a write lock on a leaf of the fs/subvol tree and then before
releasing the path we called check_ref() which did backref walking, when
qgroups are enabled, and tried to read lock the same leaf. The trace for
this case is the following:
The problem goes away by eleminating check_ref(), which no longer is
needed as its purpose was to get a value for the no_quota field of
a delayed reference (this patch removes the no_quota field as mentioned
earlier).
Reported-by: Stéphane Lesimple <stephane_btrfs@lesimple.fr> Tested-by: Stéphane Lesimple <stephane_btrfs@lesimple.fr> Reported-by: Elias Probst <mail@eliasprobst.eu> Reported-by: Peter Becker <floyd.net@gmail.com> Reported-by: Malte Schröder <malte@tnxip.de> Reported-by: Derek Dongray <derek@valedon.co.uk> Reported-by: Erkki Seppala <flux-btrfs@inside.org> Cc: stable@vger.kernel.org # 4.2+ Signed-off-by: Filipe Manana <fdmanana@suse.com> Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Filipe Manana [Thu, 22 Oct 2015 08:47:34 +0000 (09:47 +0100)]
Btrfs: fix regression when running delayed references
In the kernel 4.2 merge window we had a refactoring/rework of the delayed
references implementation in order to fix certain problems with qgroups.
However that rework introduced one more regression that leads to the
following trace when running delayed references for metadata:
This happens because the new delayed references code no longer merges
delayed references that have different sequence values. The following
steps are an example sequence leading to this issue:
1) Transaction N starts, fs_info->tree_mod_seq has value 0;
2) Extent buffer (btree node) A is allocated, delayed reference Ref1 for
bytenr A is created, with a value of 1 and a seq value of 0;
3) fs_info->tree_mod_seq is incremented to 1;
4) Extent buffer A is deleted through btrfs_del_items(), which calls
btrfs_del_leaf(), which in turn calls btrfs_free_tree_block(). The
later returns the metadata extent associated to extent buffer A to
the free space cache (the range is not pinned), because the extent
buffer was created in the current transaction (N) and writeback never
happened for the extent buffer (flag BTRFS_HEADER_FLAG_WRITTEN not set
in the extent buffer).
This creates the delayed reference Ref2 for bytenr A, with a value
of -1 and a seq value of 1;
5) Delayed reference Ref2 is not merged with Ref1 when we create it,
because they have different sequence numbers (decided at
add_delayed_ref_tail_merge());
6) fs_info->tree_mod_seq is incremented to 2;
7) Some task attempts to allocate a new extent buffer (done at
extent-tree.c:find_free_extent()), but due to heavy fragmentation
and running low on metadata space the clustered allocation fails
and we fall back to unclustered allocation, which finds the
extent at offset A, so a new extent buffer at offset A is allocated.
This creates delayed reference Ref3 for bytenr A, with a value of 1
and a seq value of 2;
8) Ref3 is not merged neither with Ref2 nor Ref1, again because they
all have different seq values;
9) We start running the delayed references (__btrfs_run_delayed_refs());
10) The delayed Ref1 is the first one being applied, which ends up
creating an inline extent backref in the extent tree;
10) Next the delayed reference Ref3 is selected for execution, and not
Ref2, because select_delayed_ref() always gives a preference for
positive references (that have an action of BTRFS_ADD_DELAYED_REF);
11) When running Ref3 we encounter alreay the inline extent backref
in the extent tree at insert_inline_extent_backref(), which makes
us hit the following BUG_ON:
BUG_ON(owner < BTRFS_FIRST_FREE_OBJECTID);
This is always true because owner corresponds to the level of the
extent buffer/btree node in the btree.
For the scenario described above we hit the BUG_ON because we never merge
references that have different seq values.
We used to do the merging before the 4.2 kernel, more specifically, before
the commmits:
c6fc24549960 ("btrfs: delayed-ref: Use list to replace the ref_root in ref_head.") c43d160fcd5e ("btrfs: delayed-ref: Cleanup the unneeded functions.")
This issue became more exposed after the following change that was added
to 4.2 as well:
cffc3374e567 ("Btrfs: fix order by which delayed references are run")
Which in turn fixed another regression by the two commits previously
mentioned.
So fix this by bringing back the delayed reference merge code, with the
proper adaptations so that it operates against the new data structure
(linked list vs old red black tree implementation).
This issue was hit running fstest btrfs/063 in a loop. Several people have
reported this issue in the mailing list when running on kernels 4.2+.
Very special thanks to Stéphane Lesimple for helping debugging this issue
and testing this fix on his multi terabyte filesystem (which took more
than one day to balance alone, plus fsck, etc).
Fixes: c6fc24549960 ("btrfs: delayed-ref: Use list to replace the ref_root in ref_head.") Reported-by: Peter Becker <floyd.net@gmail.com> Reported-by: Stéphane Lesimple <stephane_btrfs@lesimple.fr> Tested-by: Stéphane Lesimple <stephane_btrfs@lesimple.fr> Reported-by: Malte Schröder <malte@tnxip.de> Reported-by: Derek Dongray <derek@valedon.co.uk> Reported-by: Erkki Seppala <flux-btrfs@inside.org> Cc: stable@vger.kernel.org # 4.2+ Signed-off-by: Filipe Manana <fdmanana@suse.com> Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Josef Bacik [Fri, 2 Oct 2015 20:12:23 +0000 (16:12 -0400)]
Btrfs: don't do extra bitmap search in one bit case
When we make ctl->unit allocations from a bitmap there is no point in searching
for the next 0 in the bitmap. If we've found a bit we're done and can just exit
the loop. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Fri, 2 Oct 2015 20:09:42 +0000 (16:09 -0400)]
Btrfs: keep track of largest extent in bitmaps
We can waste a lot of time searching through bitmaps when we are heavily
fragmented trying to find large contiguous areas that don't exist in the bitmap.
So keep track of the max extent size when we do a full search of a bitmap so
that next time around we can just skip the expensive searching if our max size
is less than what we are looking for. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Fri, 2 Oct 2015 19:25:10 +0000 (15:25 -0400)]
Btrfs: don't keep trying to build clusters if we are fragmented
If we are extremely fragmented then we won't be able to create a free_cluster.
So if this happens set last_ptr->fragmented so that all future allcations will
give up trying to create a cluster. When we unpin extents we will unset
->fragmented if we free up a sufficient amount of space in a block group.
Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Thu, 1 Oct 2015 18:54:10 +0000 (14:54 -0400)]
Btrfs: cut down on loops through the allocator
We try really really hard to make allocations, but sometimes it is just not
going to happen, especially when free space is extremely fragmented. So add a
few short cuts through the looping states. For example if we couldn't allocate
a chunk, just go straight to the NO_EMPTY_SIZE loop. If there are no uncached
block groups and we've done a full search, go straight to the ALLOC_CHUNK stage.
And finally if we already have empty_size and empty_cluster set to 0 go ahead
and return -ENOSPC. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Thu, 1 Oct 2015 16:55:18 +0000 (12:55 -0400)]
Btrfs: don't continue setting up space cache when enospc
If we hit ENOSPC when setting up a space cache don't bother setting up any of
the other space cache's in this transaction, it'll just induce unnecessary
latency. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Tue, 29 Sep 2015 15:40:47 +0000 (11:40 -0400)]
Btrfs: keep track of max_extent_size per space_info
When we are heavily fragmented we can induce a lot of latency trying to make an
allocation happen that is simply not going to happen. Thankfully we keep track
of our max_extent_size when going through the allocator, so if we get to the
point where we are exiting find_free_extent with ENOSPC then set our
space_info->max_extent_size so we can keep future allocations from having to pay
this cost. We reset the max_extent_size whenever we release pinned bytes back
into this space info so we can redo all the work. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Fri, 25 Sep 2015 20:13:11 +0000 (16:13 -0400)]
Btrfs: don't loop in allocator for space cache
The space cache needs to have contiguous allocations, and the allocator tries to
make allocations by reducing the amount of bytes requested and re-searching.
But this just makes us waste time when we are very fragmented, so if we can't
find our space just exit, don't bother trying to search again. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Thu, 24 Sep 2015 14:46:10 +0000 (10:46 -0400)]
Btrfs: add a flags field to btrfs_transaction
I want to set some per transaction flags, so instead of adding yet another int
lets just convert the current two int indicators to flags and add a flags field
for future use. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Wed, 23 Sep 2015 21:11:16 +0000 (17:11 -0400)]
Btrfs: fix prealloc under heavy fragmentation conditions
If we are heavily fragmented we will continually try to prealloc the largest
extent size we can every time we call btrfs_reserve_extent. This can be very
expensive when we are heavily fragmented, burning lots of CPU cycles and loops
through the allocator. So instead notice when we get a smaller chunk from the
allocator than what we specified and use this as the new maximum size we try to
allocate. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Wed, 23 Sep 2015 18:54:14 +0000 (14:54 -0400)]
Btrfs: add fragment=* debug mount option
In tracking down these weird bitmap problems it was helpful to artificially
create an extremely fragmented file system. These mount options let us either
fragment data or metadata or both. With these options I could reproduce all
sorts of weird latencies and hangs that occur under extreme fragmentation and
get them fixed. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Mon, 5 Oct 2015 14:35:29 +0000 (10:35 -0400)]
Btrfs: fix qgroup sanity tests
With my changes to allow us to find old roots when resolving indirect refs I
introduced a regression to the sanity tests. Since we don't really care to go
down into the fs roots we just need to have the old behavior of returning ENOENT
for dummy roots for the sanity tests. In the future if we want to get fancy we
can populate the test fs trees with the references as well. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Josef Bacik [Thu, 24 Sep 2015 20:17:39 +0000 (16:17 -0400)]
Btrfs: change how we wait for pending ordered extents
We have a mechanism to make sure we don't lose updates for ordered extents that
were logged in the transaction that is currently running. We add the ordered
extent to a transaction list and then the transaction waits on all the ordered
extents in that list. However are substantially large file systems this list
can be extremely large, and can give us soft lockups, since the ordered extents
don't remove themselves from the list when they do complete.
To fix this we simply add a counter to the transaction that is incremented any
time we have a logged extent that needs to be completed in the current
transaction. Then when the ordered extent finally completes it decrements the
per transaction counter and wakes up the transaction if we are the last ones.
This will eliminate the softlockup. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Thu, 8 Oct 2015 10:19:37 +0000 (18:19 +0800)]
btrfs: qgroup: Avoid calling btrfs_free_reserved_data_space in clear_bit_hook
In clear_bit_hook, qgroup reserved data is already handled quite well,
either released by finish_ordered_io or invalidatepage.
So calling btrfs_qgroup_free_data() here is completely meaningless, and
since btrfs_qgroup_free_data() will lock io_tree, so it can't be called
with io_tree lock hold.
This patch will add a new function
btrfs_free_reserved_data_space_noquota() for clear_bit_hook() to cease
the lockdep warning.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
btrfs: fallocate: Add support to accurate qgroup reserve
Now fallocate will do accurate qgroup reserve space check, unlike old
method, which will always reserve the whole length of the range.
With this patch, fallocate will:
1) Iterate the desired range and mark in data rsv map
Only range which is going to be allocated will be recorded in data
rsv map and reserve the space.
For already allocated range (normal/prealloc extent) they will be
skipped.
Also, record the marked range into a new list for later use.
2) If 1) succeeded, do real file extent allocate.
And at file extent allocation time, corresponding range will be
removed from the range in data rsv map.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
For btrfs_invalidatepage() and its variant evict_inode_truncate_page(),
there will be pages don't reach disk.
In that case, their reserved space won't be release nor freed by
finish_ordered_io() nor delayed_ref handler.
So we must free their qgroup reserved space, or we will leaking reserved
space again.
So this will patch will call btrfs_qgroup_free_data() for
invalidatepage() and its variant evict_inode_truncate_page().
And due to the nature of new btrfs_qgroup_reserve/free_data() reserved
space will only be reserved or freed once, so for pages which are
already flushed to disk, their reserved space will be released and freed
by delayed_ref handler.
Double free won't be a problem.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
For NOCOW and inline case, there will be no delayed_ref created for
them, so we should free their reserved data space at proper
time(finish_ordered_io for NOCOW and cow_file_inline for inline).
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Cleanup the old facilities which use old btrfs_qgroup_reserve() function
call, replace them with the newer version, and remove the "__" prefix in
them.
Also, make btrfs_qgroup_reserve/free() functions private, as they are
now only used inside qgroup codes.
Now, the whole btrfs qgroup is swithed to use the new reserve facilities.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
btrfs: extent-tree: Switch to new check_data_free_space and free_reserved_data_space
Use new reserve/free for buffered write and inode cache.
For buffered write case, as nodatacow write won't increase quota account,
so unlike old behavior which does reserve before check nocow, now we
check nocow first and then only reserve data if we can't do nocow write.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
btrfs: extent-tree: Add new version of btrfs_check_data_free_space and btrfs_free_reserved_data_space.
Add new functions __btrfs_check_data_free_space() and
__btrfs_free_reserved_data_space() to work with new accurate qgroup
reserved space framework.
The new function will replace old btrfs_check_data_free_space() and
btrfs_free_reserved_data_space() respectively, but until all the change
is done, let's just use the new name.
Also, export internal use function btrfs_alloc_data_chunk_ondemand(), as
now qgroup reserve requires precious bytes, some operation can't get the
accurate number in advance(like fallocate).
But data space info check and data chunk allocate doesn't need to be
that accurate, and can be called at the beginning.
So export it for later operations.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
btrfs: delayed_ref: release and free qgroup reserved at proper timing
Qgroup reserved space needs to be released from inode dirty map and get
freed at different timing:
1) Release when the metadata is written into tree
After corresponding metadata is written into tree, any newer write will
be COWed(don't include NOCOW case yet).
So we must release its range from inode dirty range map, or we will
forget to reserve needed range, causing accounting exceeding the limit.
2) Free reserved bytes when delayed ref is run
When delayed refs are run, qgroup accounting will follow soon and turn
the reserved bytes into rfer/excl numbers.
As run_delayed_refs and qgroup accounting are all done at
commit_transaction() time, we are safe to free reserved space in
run_delayed_ref time().
With these timing to release/free reserved space, we should be able to
resolve the long existing qgroup reserve space leak problem.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Mon, 12 Oct 2015 08:28:06 +0000 (16:28 +0800)]
btrfs: qgroup: Introduce functions to release/free qgroup reserve data
space
Introduce functions btrfs_qgroup_release/free_data() to release/free
reserved data range.
Release means, just remove the data range from io_tree, but doesn't
free the reserved space.
This is for normal buffered write case, when data is written into disc
and its metadata is added into tree, its reserved space should still be
kept until commit_trans().
So in that case, we only release dirty range, but keep the reserved
space recorded some other place until commit_tran().
Free means not only remove data range, but also free reserved space.
This is used for case for cleanup and invalidate page.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Mon, 12 Oct 2015 07:35:38 +0000 (15:35 +0800)]
btrfs: extent_io: Introduce new function clear_record_extent_bits()
Introduce new function clear_record_extent_bits(), which will clear bits
for given range and record the details about which ranges are cleared
and how many bytes in total it changes.
This provides the basis for later qgroup reserve codes.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Mon, 12 Oct 2015 06:53:37 +0000 (14:53 +0800)]
btrfs: extent_io: Introduce new function set_record_extent_bits
Introduce new function set_record_extent_bits(), which will not only set
given bits, but also record how many bytes are changed, and detailed
range info.
This is quite important for later qgroup reserve framework.
The number of bytes will be used to do qgroup reserve, and detailed
range info will be used to cleanup for EQUOT case.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
Qu Wenruo [Mon, 12 Oct 2015 04:08:16 +0000 (12:08 +0800)]
btrfs: extent_io: Introduce needed structure for recoding set/clear bits
Add a new structure, extent_change_set, to record how many bytes are
changed in one set/clear_extent_bits() operation, with detailed changed
ranges info.
This provides the needed facilities for later qgroup reserve framework.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com> Signed-off-by: Chris Mason <clm@fb.com>
reada is using -1 instead of the -ENOMEM defined macro to specify that
a buffer allocation failed. Since the error number is propagated, the
caller will get a -EPERM which is the wrong error condition.
Also, updating the caller to return the exact value from
reada_add_block.
Smatch tool warning:
reada_add_block() warn: returning -1 instead of -ENOMEM is sloppy
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com> Signed-off-by: David Sterba <dsterba@suse.com>
check-integrity is using -1 instead of the -ENOMEM defined macro to
specify that a buffer allocation failed. Since the error number is
propagated, the caller will get a -EPERM which is the wrong error
condition.
Also, the smatch tool complains with the following warnings:
btrfsic_process_superblock() warn: returning -1 instead of -ENOMEM is sloppy
btrfsic_read_block() warn: returning -1 instead of -ENOMEM is sloppy
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Luis de Bethencourt <luisbg@osg.samsung.com> Signed-off-by: David Sterba <dsterba@suse.com>
Byongho Lee [Wed, 14 Oct 2015 05:05:24 +0000 (14:05 +0900)]
btrfs: compress: put variables defined per compress type in struct to make cache friendly
Below variables are defined per compress type.
- struct list_head comp_idle_workspace[BTRFS_COMPRESS_TYPES]
- spinlock_t comp_workspace_lock[BTRFS_COMPRESS_TYPES]
- int comp_num_workspace[BTRFS_COMPRESS_TYPES]
- atomic_t comp_alloc_workspace[BTRFS_COMPRESS_TYPES]
- wait_queue_head_t comp_workspace_wait[BTRFS_COMPRESS_TYPES]
BTW, while accessing one compress type of these variables, the next or
before address is other compress types of it.
So this patch puts these variables in a struct to make cache friendly.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Byongho Lee <bhlee.kernel@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Byongho Lee [Thu, 8 Oct 2015 11:49:34 +0000 (20:49 +0900)]
btrfs: cleanup iterating over prop_handlers array
This patch eliminates the last item of prop_handlers array which is used
to check end of array and instead uses ARRAY_SIZE macro.
Though this is a very tiny optimization, using ARRAY_SIZE macro is a
good practice to iterate array.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Byongho Lee <bhlee.kernel@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Tue, 22 Sep 2015 21:00:07 +0000 (21:00 +0000)]
btrfs: declare rsv_count as unsigned int instead of int
rsv_count ultimately gets passed to start_transaction() which
now takes an unsigned int as its num_items parameter.
The value of rsv_count should always be positive so declare it
as being unsigned.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Tue, 22 Sep 2015 20:59:15 +0000 (20:59 +0000)]
btrfs: change num_items type from u64 to unsigned int
The value of num_items that start_transaction() ultimately
always takes is a small one, so a 64 bit integer is overkill.
Also change num_items for btrfs_start_transaction() and
btrfs_start_transaction_lflush() as well.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Improve readability by generalizing the profile validity checks.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Zhao Lei [Tue, 15 Sep 2015 13:08:07 +0000 (21:08 +0800)]
btrfs: use btrfs_raid_array for btrfs_get_num_tolerated_disk_barrier_failures()
btrfs_raid_array[] is used to define all raid attributes, use it
to get tolerated_failures in btrfs_get_num_tolerated_disk_barrier_failures(),
instead of complex condition in function.
It can make code simple and auto-support other possible raid-type in
future.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Signed-off-by: David Sterba <dsterba@suse.com>
Zhao Lei [Tue, 15 Sep 2015 13:08:06 +0000 (21:08 +0800)]
btrfs: Move btrfs_raid_array to public
This array is used to record attributes of each raid type,
make it public, and many functions will benifit with this array.
For example, num_tolerated_disk_barrier_failures(), we can
avoid complex conditions in this function, and get raid attribute
simply by accessing above array.
It can also make code logic simple, reduce duplication code, and
increase maintainability.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Zhao Lei <zhaolei@cn.fujitsu.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Wed, 9 Sep 2015 00:18:50 +0000 (00:18 +0000)]
btrfs: use a single if() statement for one outcome in get_block_rsv()
Rather than have three separate if() statements for the same outcome
we should just OR them together in the same if() statement.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Mon, 7 Sep 2015 14:24:37 +0000 (17:24 +0300)]
btrfs: memset cur_trans->delayed_refs to zero
Use memset() to null out the btrfs_delayed_ref_root of
btrfs_transaction instead of setting all the members to 0 by hand.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Thu, 27 Aug 2015 23:53:45 +0000 (23:53 +0000)]
btrfs: trimming some start_transaction() code away
Just call kmem_cache_zalloc() instead of calling kmem_cache_alloc().
We're just initializing most fields to 0, false and NULL later on
_anyway_, so to make the code mode readable and potentially gain
a bit of performance (completely untested claim), we should fill our
btrfs_trans_handle with zeros on allocation then just initialize
those five remaining fields (not counting the list_heads) as normal.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Mon, 24 Aug 2015 21:56:28 +0000 (21:56 +0000)]
btrfs: Fixed declaration of old_len
old_len is used to store the return value of btrfs_item_size_nr().
The return value of btrfs_item_size_nr() is of type u32.
To improve code correctness and avoid mixing signed and unsigned
integers I've changed old_len to be of type u32 as well.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Alexandru Moise [Sun, 23 Aug 2015 16:01:42 +0000 (16:01 +0000)]
btrfs: Fixed dsize and last_off declarations
The return values of btrfs_item_offset_nr and btrfs_item_size_nr are of
type u32. To avoid mixing signed and unsigned integers we should also
declare dsize and last_off to be of type u32.
Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: Alexandru Moise <00moses.alexander00@gmail.com> Signed-off-by: David Sterba <dsterba@suse.com>
Filipe Manana [Fri, 16 Oct 2015 11:34:25 +0000 (12:34 +0100)]
Btrfs: fix truncation of compressed and inlined extents
When truncating a file to a smaller size which consists of an inline
extent that is compressed, we did not discard (or made unusable) the
data between the new file size and the old file size, wasting metadata
space and allowing for the truncated data to be leaked and the data
corruption/loss mentioned below.
We were also not correctly decrementing the number of bytes used by the
inode, we were setting it to zero, giving a wrong report for callers of
the stat(2) syscall. The fsck tool also reported an error about a mismatch
between the nbytes of the file versus the real space used by the file.
Now because we weren't discarding the truncated region of the file, it
was possible for a caller of the clone ioctl to actually read the data
that was truncated, allowing for a security breach without requiring root
access to the system, using only standard filesystem operations. The
scenario is the following:
1) User A creates a file which consists of an inline and compressed
extent with a size of 2000 bytes - the file is not accessible to
any other users (no read, write or execution permission for anyone
else);
2) The user truncates the file to a size of 1000 bytes;
3) User A makes the file world readable;
4) User B creates a file consisting of an inline extent of 2000 bytes;
5) User B issues a clone operation from user A's file into its own
file (using a length argument of 0, clone the whole range);
6) User B now gets to see the 1000 bytes that user A truncated from
its file before it made its file world readbale. User B also lost
the bytes in the range [1000, 2000[ bytes from its own file, but
that might be ok if his/her intention was reading stale data from
user A that was never supposed to be public.
Note that this contrasts with the case where we truncate a file from 2000
bytes to 1000 bytes and then truncate it back from 1000 to 2000 bytes. In
this case reading any byte from the range [1000, 2000[ will return a value
of 0x00, instead of the original data.
This problem exists since the clone ioctl was added and happens both with
and without my recent data loss and file corruption fixes for the clone
ioctl (patch "Btrfs: fix file corruption and data loss after cloning
inline extents").
So fix this by truncating the compressed inline extents as we do for the
non-compressed case, which involves decompressing, if the data isn't already
in the page cache, compressing the truncated version of the extent, writing
the compressed content into the inline extent and then truncate it.
The following test case for fstests reproduces the problem. In order for
the test to pass both this fix and my previous fix for the clone ioctl
that forbids cloning a smaller inline extent into a larger one,
which is titled "Btrfs: fix file corruption and data loss after cloning
inline extents", are needed. Without that other fix the test fails in a
different way that does not leak the truncated data, instead part of
destination file gets replaced with zeroes (because the destination file
has a larger inline extent than the source).
seq=`basename $0`
seqres=$RESULT_DIR/$seq
echo "QA output created by $seq"
tmp=/tmp/$$
status=1 # failure is the default!
trap "_cleanup; exit \$status" 0 1 2 3 15
_cleanup()
{
rm -f $tmp.*
}
# get standard environment, filters and checks
. ./common/rc
. ./common/filter
# real QA test starts here
_need_to_be_root
_supported_fs btrfs
_supported_os Linux
_require_scratch
_require_cloner
# Create our test files. File foo is going to be the source of a clone operation
# and consists of a single inline extent with an uncompressed size of 512 bytes,
# while file bar consists of a single inline extent with an uncompressed size of
# 256 bytes. For our test's purpose, it's important that file bar has an inline
# extent with a size smaller than foo's inline extent.
$XFS_IO_PROG -f -c "pwrite -S 0xa1 0 128" \
-c "pwrite -S 0x2a 128 384" \
$SCRATCH_MNT/foo | _filter_xfs_io
$XFS_IO_PROG -f -c "pwrite -S 0xbb 0 256" $SCRATCH_MNT/bar | _filter_xfs_io
# Now durably persist all metadata and data. We do this to make sure that we get
# on disk an inline extent with a size of 512 bytes for file foo.
sync
# Now truncate our file foo to a smaller size. Because it consists of a
# compressed and inline extent, btrfs did not shrink the inline extent to the
# new size (if the extent was not compressed, btrfs would shrink it to 128
# bytes), it only updates the inode's i_size to 128 bytes.
$XFS_IO_PROG -c "truncate 128" $SCRATCH_MNT/foo
# Now clone foo's inline extent into bar.
# This clone operation should fail with errno EOPNOTSUPP because the source
# file consists only of an inline extent and the file's size is smaller than
# the inline extent of the destination (128 bytes < 256 bytes). However the
# clone ioctl was not prepared to deal with a file that has a size smaller
# than the size of its inline extent (something that happens only for compressed
# inline extents), resulting in copying the full inline extent from the source
# file into the destination file.
#
# Note that btrfs' clone operation for inline extents consists of removing the
# inline extent from the destination inode and copy the inline extent from the
# source inode into the destination inode, meaning that if the destination
# inode's inline extent is larger (N bytes) than the source inode's inline
# extent (M bytes), some bytes (N - M bytes) will be lost from the destination
# file. Btrfs could copy the source inline extent's data into the destination's
# inline extent so that we would not lose any data, but that's currently not
# done due to the complexity that would be needed to deal with such cases
# (specially when one or both extents are compressed), returning EOPNOTSUPP, as
# it's normally not a very common case to clone very small files (only case
# where we get inline extents) and copying inline extents does not save any
# space (unlike for normal, non-inlined extents).
$CLONER_PROG -s 0 -d 0 -l 0 $SCRATCH_MNT/foo $SCRATCH_MNT/bar
# Now because the above clone operation used to succeed, and due to foo's inline
# extent not being shinked by the truncate operation, our file bar got the whole
# inline extent copied from foo, making us lose the last 128 bytes from bar
# which got replaced by the bytes in range [128, 256[ from foo before foo was
# truncated - in other words, data loss from bar and being able to read old and
# stale data from foo that should not be possible to read anymore through normal
# filesystem operations. Contrast with the case where we truncate a file from a
# size N to a smaller size M, truncate it back to size N and then read the range
# [M, N[, we should always get the value 0x00 for all the bytes in that range.
# We expected the clone operation to fail with errno EOPNOTSUPP and therefore
# not modify our file's bar data/metadata. So its content should be 256 bytes
# long with all bytes having the value 0xbb.
#
# Without the btrfs bug fix, the clone operation succeeded and resulted in
# leaking truncated data from foo, the bytes that belonged to its range
# [128, 256[, and losing data from bar in that same range. So reading the
# file gave us the following content:
#
# 0000000 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1
# *
# 0000200 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a 2a
# *
# 0000400
echo "File bar's content after the clone operation:"
od -t x1 $SCRATCH_MNT/bar
# Also because the foo's inline extent was not shrunk by the truncate
# operation, btrfs' fsck, which is run by the fstests framework everytime a
# test completes, failed reporting the following error:
#
# root 5 inode 257 errors 400, nbytes wrong
status=0
exit
Cc: stable@vger.kernel.org Signed-off-by: Filipe Manana <fdmanana@suse.com>