Even though a separate root is passed in, we're still operating on the
extent root. Let's use that for the trace point.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_init_new_device only uses the root passed in via the ioctl to
start the transaction. Nothing else that happens is related to whatever
root the user used to initiate the ioctl. We can drop the root requirement
and just use fs_info->dev_root instead.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There are many functions that are always called with the same root
argument. Rather than passing the same root every time, we can
pass an fs_info pointer instead and have the function get the root
pointer itself.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There are 11 functions that accept a root parameter and immediately
overwrite it. We can pass those an fs_info pointer instead.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This issue was found when I tried to delete a heavily reflinked file,
when deleting such files, other transaction operation will not have a
chance to make progress, for example, start_transaction() will blocked
in wait_current_trans(root) for long time, sometimes it even triggers
soft lockups, and the time taken to delete such heavily reflinked file
is also very large, often hundreds of seconds. Using perf top, it reports
that:
PerfTop: 7416 irqs/sec kernel:99.8% exact: 0.0% [4000Hz cpu-clock], (all, 4 CPUs)
---------------------------------------------------------------------------------------
84.37% [btrfs] [k] __btrfs_run_delayed_refs.constprop.80
11.02% [kernel] [k] delay_tsc
0.79% [kernel] [k] _raw_spin_unlock_irq
0.78% [kernel] [k] _raw_spin_unlock_irqrestore
0.45% [kernel] [k] do_raw_spin_lock
0.18% [kernel] [k] __slab_alloc
It seems __btrfs_run_delayed_refs() took most cpu time, after some debug
work, I found it's select_delayed_ref() causing this issue, for a delayed
head, in our case, it'll be full of BTRFS_DROP_DELAYED_REF nodes, but
select_delayed_ref() will firstly try to iterate node list to find
BTRFS_ADD_DELAYED_REF nodes, obviously it's a disaster in this case, and
waste much time.
To fix this issue, we introduce a new ref_add_list in struct btrfs_delayed_ref_head,
then in select_delayed_ref(), if this list is not empty, we can directly use
nodes in this list. With this patch, it just took about 10~15 seconds to
delte the same file. Now using perf top, it reports that:
PerfTop: 2734 irqs/sec kernel:99.5% exact: 0.0% [4000Hz cpu-clock], (all, 4 CPUs)
----------------------------------------------------------------------------------------
20.74% [kernel] [k] _raw_spin_unlock_irqrestore
16.33% [kernel] [k] __slab_alloc
5.41% [kernel] [k] lock_acquired
4.42% [kernel] [k] lock_acquire
4.05% [kernel] [k] lock_release
3.37% [kernel] [k] _raw_spin_unlock_irq
For normal files, this patch also gives help, at least we do not need to
iterate whole list to found BTRFS_ADD_DELAYED_REF nodes.
Signed-off-by: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit 62b99540a1 (btrfs: relocation: Fix leaking qgroups numbers
on data extents) only fixes the problem partly.
The previous fix is to trace all new data extents at transaction commit
time when balance finishes.
However balance is not done in a large transaction, every path
replacement can happen in its own transaction.
This makes the fix useless if transaction commits during relocation.
For example:
relocate_block_group()
|-merge_reloc_roots()
| |- merge_reloc_root()
| |- btrfs_start_transaction() <- Trans X
| |- replace_path() <- Cause leak
| |- btrfs_end_transaction_throttle() <- Trans X commits here
| | Leak not fixed
| |
| |- btrfs_start_transaction() <- Trans Y
| |- replace_path() <- Cause leak
| |- btrfs_end_transaction_throttle() <- Trans Y ends
| but not committed
|-btrfs_join_transaction() <- Still trans Y
|-qgroup_fix() <- Only fixes data leak
| in trans Y
|-btrfs_commit_transaction() <- Trans Y commits
In that case, qgroup fixup can only fix data leak in trans Y, data leak
in trans X is out of fix.
So the correct fix should happen in the same transaction of
replace_path().
This patch fixes it by tracing both subtrees of tree block swap, so it
can fix the problem and ensure all leaking and fix are in the same
transaction, so no leak again.
Reported-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-and-Tested-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Move account_shared_subtree() to qgroup.c and rename it to
btrfs_qgroup_trace_subtree().
Do the same thing for account_leaf_items() and rename it to
btrfs_qgroup_trace_leaf_items().
Since all these functions are only for qgroup, move them to qgroup.c and
export them is more appropriate.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-and-Tested-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Rename btrfs_qgroup_insert_dirty_extent(_nolock) to
btrfs_qgroup_trace_extent(_nolock), according to the new
reserve/trace/account naming schema.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-and-Tested-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add explaination how btrfs qgroups work.
Qgroup is split into 3 main phrases:
1) Reserve
To ensure qgroup doesn't exceed its limit
2) Trace
To info qgroup to trace which extent
3) Account
Calculate qgroup number change for each traced extent.
This should save quite some time for new developers.
Signed-off-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Reviewed-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
And remove the bogus check for a NULL return value from kmap, which
can't happen. While we're at it: I don't think that kmapping up to 256
will work without deadlocks on highmem machines, a better idea would
be to use vm_map_ram to map all of them into a single virtual address
range. Incidentally that would also simplify the code a lot.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Rework the loop a little bit to use the generic bio_for_each_segment_all
helper for iterating over the bio.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Use the bvec offset and len members to prepare for multipage bvecs.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: David Sterba <dsterba@suse.com>
Instead of using bi_vcnt to calculate it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Use bio_for_each_segment_all to iterate over the segments instead.
This requires a bit of reshuffling so that we only lookup up the ordered
item once inside the bio_for_each_segment_all loop.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Just use bio_for_each_segment_all to iterate over all segments.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Just use bio_for_each_segment_all to iterate over all segments.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Pass the full bio to the decompression routines and use bio iterators
to iterate over the data in the bio.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: David Sterba <dsterba@suse.com>
This fixes the WARN_ON on BTRFS_I(inode)->reserved_extents in
btrfs_destroy_inode and the WARN_ON on nonzero delalloc bytes on umount
with qgroups enabled.
I was able to reproduce this by setting up a small (~500kb) quota limit
and writing a file one byte at a time until I hit the limit. The warnings
would all hit on umount.
The root cause is that we would reserve a block-sized range in both
the reservation and the quota in btrfs_check_data_free_space, but if we
encountered a problem (like e.g. EDQUOT), we would only release the single
byte in the qgroup reservation. That caused an iotree state split, which
increased the number of outstanding extents, in turn disallowing releasing
the metadata reservation.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: Qu Wenruo <quwenruo@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
At this point we will have dropped extent entries from the file, so if we fail
to insert the new hole entries then we are leaving the fs in a corrupt state
(albeit an easily fixed one). Abort the transaciton if this happens so we can
avoid corrupting the fs. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In order to do hole punching we have a block reserve to hold the reservation we
need to drop the extents in our range. Since we could end up dropping a lot of
extents we set rsv->failfast so we can just loop around again and drop the
remaining of the range. Unfortunately we unconditionally fill the hole extents
in and start from the last extent we encountered, which we may or may not have
dropped. So this can result in overlapping file extent entries, which can be
tripped over in a variety of ways, either by hitting BUG_ON(!ret) in
fill_holes() after the search, or in btrfs_set_item_key_safe() in
btrfs_drop_extent() at a later time by an unrelated task. Fix this by only
setting drop_end to the last extent we did actually drop. This way our holes
are filled in properly for the range that we did drop, and the rest of the range
that remains to be dropped is actually dropped. Thanks,
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If we process the last item in the leaf and hit an I/O error while
reading the next leaf, we return -EIO without having adjusted the
position. Since we have emitted dirents, getdents() will return
the byte count to the user instead of the error. Subsequent callers
will emit the last successful dirent again, and return -EIO again,
with the same result. Callers loop forever.
Instead, if we always increment ctx->pos after emitting or skipping
the dirent, we'll be sure that we won't hit the same one again. When
we go to process the next leaf, we won't have emitted any dirents
and the -EIO will be returned to the user properly. We also don't
need to track if we've emitted a dirent already or if we've changed
the position yet.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit 3de4586c52 (Btrfs: Allow subvolumes and snapshots anywhere
in the directory tree) introduced the current system of placing
snapshots in the directory tree. It also introduced the behavior of
creating the snapshot and then creating the directory entries for it.
We've kept this code around for compatibility reasons, but it turns
out that no file systems with the old tree_root based snapshots can
be mounted on newer (>= 2009) kernels anyway. About a month after the
above commit, commit 2a7108ad89 (Btrfs: rev the disk format for the
inode compat and csum selection changes) landed, changing the superblock
magic number.
As a result, we know that we'll never encounter tree_root-based dirents
or have to deal with skipping our own snapshot dirents. Since that
also means that we're now only iterating over DIR_INDEX items, which only
contain one directory entry per leaf item, we don't need to loop over
the leaf item contents anymore either.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If zlib_inflateInit2 fails, the input page is never unmapped.
Add a call to kunmap when it fails.
Signed-off-by: Nick Terrell <nickrterrell@gmail.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The balance status item contains currently known filter values, but the
stripes filter was unintentionally not among them. This would mean, that
interrupted and automatically restarted balance does not apply the
stripe filters.
Fixes: dee32d0ac3
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: David Sterba <dsterba@suse.com>
'btrfs_iget()' can not return NULL, so this test can be removed.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
csum member of struct btrfs_super_block has array type of u8. It makes
sense that function btrfs_csum_final should be also declared to accept
u8 *. I changed the declaration of method void btrfs_csum_final(u32 crc,
char *result); to void btrfs_csum_final(u32 crc, u8 *result);
Signed-off-by: Domagoj Tršan <domagoj.trsan@gmail.com>
[ changed cast to u8 at several call sites ]
Signed-off-by: David Sterba <dsterba@suse.com>
If we have
|0--hole--4095||4096--preallocate--12287|
instead of using preallocated space, a 8K direct write will just
create a new 8K extent and it'll end up with
|0--new extent--8191||8192--preallocate--12287|
It's because we find a hole em and then go to create a new 8K
extent directly without adjusting @len.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: Chris Mason <clm@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There is no need to call kfree() if memdup_user() fails, as no memory
was allocated and the error in the error-valued pointer should be returned.
Signed-off-by: Shailendra Verma <shailendra.v@samsung.com>
[ edit subject ]
Signed-off-by: David Sterba <dsterba@suse.com>
Using copy_extent_buffer is suitable for copying betwenn buffers from an
arbitrary offset and deals with page boundaries. This is not necessary
when doing a full extent_buffer-to-extent_buffer copy. We can utilize
the copy_page helper as well.
Signed-off-by: David Sterba <dsterba@suse.com>
The only memset we do is to 0, so sink the parameter to the function and
simplify all calls. Rename the function to reflect the behaviour.
Signed-off-by: David Sterba <dsterba@suse.com>
The fsid and chunk tree uuid are always located in the first page,
we don't need the to use write_extent_buffer.
Signed-off-by: David Sterba <dsterba@suse.com>
During the time, the function has been shrunk to the point that it just
calls find_extent_buffer, just passing the parameters.
Signed-off-by: David Sterba <dsterba@suse.com>
We dereference fs_info several times, besides that post-mount functions
should never see a NULL fs_info.
Signed-off-by: David Sterba <dsterba@suse.com>
The lock is held, we make the same lookup that previously failed with
EEXIST and we don't insert NULL pointers.
Signed-off-by: David Sterba <dsterba@suse.com>
Originally, the eb and start were passed separately in case eb is NULL.
Since the readahead has been refactored in 4.6, this is not true anymore
and we can get rid of the parameter.
Signed-off-by: David Sterba <dsterba@suse.com>
'start' is not used since "btrfs: reada: Pass reada_extent into
__readahead_hook directly" (6e39dbe8b9).
Signed-off-by: David Sterba <dsterba@suse.com>
We can't touch the eb directly in case the function is called with a
non-zero error, so we can read the eb level when needed.
Signed-off-by: David Sterba <dsterba@suse.com>
The helpers are not meant to be generic, the name is misleading. Convert
them to static inlines for type checking.
Signed-off-by: David Sterba <dsterba@suse.com>
They're not even documented anywhere, letting users with no recourse but
to RTFS. It's no big burden to output the bitfield as words.
Also, display unknown flags as hex.
Signed-off-by: Adam Borowski <kilobyte@angband.pl>
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: David Sterba <dsterba@suse.com>
My QEMU VM was seeing inexplicable I/O errors that I tracked down to
errors coming from the qcow2 virtual drive in the host system. The qcow2
file is a nocow file on my Btrfs drive, which QEMU opens with O_DIRECT.
Every once in awhile, pread() or pwrite() would return EEXIST, which
makes no sense. This turned out to be a bug in btrfs_get_extent().
Commit 8dff9c8534 ("Btrfs: deal with duplciates during extent_map
insertion in btrfs_get_extent") fixed a case in btrfs_get_extent() where
two threads race on adding the same extent map to an inode's extent map
tree. However, if the added em is merged with an adjacent em in the
extent tree, then we'll end up with an existing extent that is not
identical to but instead encompasses the extent we tried to add. When we
call merge_extent_mapping() to find the nonoverlapping part of the new
em, the arithmetic overflows because there is no such thing. We then end
up trying to add a bogus em to the em_tree, which results in a EEXIST
that can bubble all the way up to userspace.
Fix it by extending the identical extent map special case.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Tickets_id's name may result in some misunderstandings, it just indicates
the next ticket will be handled and is not stored per ticket.
Fixes: ce12965 ("btrfs: introduce tickets_id to determine whether
asynchronous metadata reclaim work makes progress")
Signed-off-by: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_map_block supports different types of mappings, which to a large
extent resemble block layer operations. But they don't always do, and
currently btrfs dangerously overlays it's own flag over the block layer
flags. This is just asking for a conflict, so introduce a different
map flags enum inside of btrfs instead.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
