When doing a device replace, while at scrub.c:scrub_enumerate_chunks(), we
set the block group to RO mode and then wait for any ongoing writes into
extents of the block group to complete. While doing that wait we overwrite
the value of the variable 'ret' and can break out of the loop if an error
happens without turning the block group back into RW mode. So what happens
is the following:
1) btrfs_inc_block_group_ro() returns 0, meaning it set the block group
to RO mode (its ->ro field set to 1 or incremented to some value > 1);
2) Then btrfs_wait_ordered_roots() returns a value > 0;
3) Then if either joining or committing the transaction fails, we break
out of the loop wihtout calling btrfs_dec_block_group_ro(), leaving
the block group in RO mode forever.
To fix this, just remove the code that waits for ongoing writes to extents
of the block group, since it's not needed because in the initial setup
phase of a device replace operation, before starting to find all chunks
and their extents, we set the target device for replace while holding
fs_info->dev_replace->rwsem, which ensures that after releasing that
semaphore, any writes into the source device are made to the target device
as well (__btrfs_map_block() guarantees that). So while at
scrub_enumerate_chunks() we only need to worry about finding and copying
extents (from the source device to the target device) that were written
before we started the device replace operation.
Fixes: f0e9b7d640 ("Btrfs: fix race setting block group readonly during device replace")
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit 9e0af23764 ("Btrfs: fix task hang under heavy compressed
write") worked around the issue that a recycled work item could get a
false dependency on the original work item due to how the workqueue code
guarantees non-reentrancy. It did so by giving different work functions
to different types of work.
However, the fixes in the previous few patches are more complete, as
they prevent a work item from being recycled at all (except for a tiny
window that the kernel workqueue code handles for us). This obsoletes
the previous fix, so we don't need the unique helpers for correctness.
The only other reason to keep them would be so they show up in stack
traces, but they always seem to be optimized to a tail call, so they
don't show up anyways. So, let's just get rid of the extra indirection.
While we're here, rename normal_work_helper() to the more informative
btrfs_work_helper().
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The file ctree.h serves as a header for everything and has become quite
bloated. Split some helpers that are generic and create a new file that
should be the catch-all for code that's not btrfs-specific.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_lock_and_flush_ordered_range() loads given "*cached_state" into
cachedp, which, in general, is NULL. Then, lock_extent_bits() updates
"cachedp", but it never goes backs to the caller. Thus the caller still
see its "cached_state" to be NULL and never free the state allocated
under btrfs_lock_and_flush_ordered_range(). As a result, we will
see massive state leak with e.g. fstests btrfs/005. Fix this bug by
properly handling the pointers.
Fixes: bd80d94efb ("btrfs: Always use a cached extent_state in btrfs_lock_and_flush_ordered_range")
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have code for data and metadata reservations for delalloc. There's
quite a bit of code here, and it's used in a lot of places so I've
separated it out to it's own file. inode.c and file.c are already
pretty large, and this code is complicated enough to live in its own
space.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
BTRFS has the implicit assumption that a checksum in btrfs_orderd_sums
is 4 bytes. While this is true for CRC32C, it is not for any other
checksum.
Change the data type to be a byte array and adjust loop index
calculation accordingly.
This includes moving the adjustment of 'index' by 'ins_size' in
btrfs_csum_file_blocks() before dividing 'ins_size' by the checksum
size, because before this patch the 'sums' member of 'struct
btrfs_ordered_sum' was 4 Bytes in size and afterwards it is only one
byte.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In case no cached_state argument is passed to
btrfs_lock_and_flush_ordered_range use one locally in the function. This
optimises the case when an ordered extent is found since the unlock
function will be able to unlock that state directly without searching
for it again.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There is a certain idiom used in multiple places in btrfs' codebase,
dealing with flushing an ordered range. Factor this in a separate
function that can be reused. Future patches will replace the existing
code with that function.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When diagnosing a slowdown of generic/224 I noticed we were not doing
anything when calling into shrink_delalloc(). This is because all
writes in 224 are O_DIRECT, not delalloc, and thus our delalloc_bytes
counter is 0, which short circuits most of the work inside of
shrink_delalloc(). However O_DIRECT writes still consume metadata
resources and generate ordered extents, which we can still wait on.
Fix this by tracking outstanding DIO write bytes, and use this as well
as the delalloc bytes counter to decide if we need to lookup and wait on
any ordered extents. If we have more DIO writes than delalloc bytes
we'll go ahead and wait on any ordered extents regardless of our flush
state as flushing delalloc is likely to not gain us anything.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
[ use dio instead of odirect in identifiers ]
Signed-off-by: David Sterba <dsterba@suse.com>
Ordered csums are keyed off of a btrfs_ordered_extent, which already has
a reference to the inode. This implies that an explicit inode argument
is redundant. So remove it.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Recent multi-page biovec rework allowed creation of bios that can span
large regions - up to 128 megabytes in the case of btrfs. OTOH btrfs'
submission path currently allocates a contiguous array to store the
checksums for every bio submitted. This means we can request up to
(128mb / BTRFS_SECTOR_SIZE) * 4 bytes + 32bytes of memory from kmalloc.
On busy systems with possibly fragmented memory said kmalloc can fail
which will trigger BUG_ON due to improper error handling IO submission
context in btrfs.
Until error handling is improved or bios in btrfs limited to a more
manageable size (e.g. 1m) let's use kvmalloc to fallback to vmalloc for
such large allocations. There is no hard requirement that the memory
allocated for checksums during IO submission has to be contiguous, but
this is a simple fix that does not require several non-contiguous
allocations.
For small writes this is unlikely to have any visible effect since
kmalloc will still satisfy allocation requests as usual. For larger
requests the code will just fallback to vmalloc.
We've performed evaluation on several workload types and there was no
significant difference kmalloc vs kvmalloc.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Tracking pending ordered extents per transaction was introduced in commit
50d9aa99bd ("Btrfs: make sure logged extents complete in the current
transaction V3") and later updated in commit 161c3549b4 ("Btrfs: change
how we wait for pending ordered extents").
However now that on fsync we always wait for ordered extents to complete
before logging, done in commit 5636cf7d6d ("btrfs: remove the logged
extents infrastructure"), we no longer need the stuff to track for pending
ordered extents, which was not completely removed in the mentioned commit.
So remove the remaining of the pending ordered extents infrastructure.
Reviewed-by: Liu Bo <bo.liu@linux.alibaba.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The logic to check if the inode is already in the log can now be
simplified since we always wait for the ordered extents to complete
before deciding whether the inode needs to be logged. The big comment
about it can go away too.
CC: Filipe Manana <fdmanana@suse.com>
Suggested-by: Filipe Manana <fdmanana@suse.com>
[ code and changelog copied from mail discussion ]
Signed-off-by: David Sterba <dsterba@suse.com>
This is no longer used anywhere, remove all of it.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Use the wrappers and reduce the amount of low-level details about the
waitqueue management.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Remove GPL boilerplate text (long, short, one-line) and keep the rest,
ie. personal, company or original source copyright statements. Add the
SPDX header.
Signed-off-by: David Sterba <dsterba@suse.com>
Before this patch, btrfs qgroup is mixing per-transcation meta rsv with
preallocated meta rsv, making it quite easy to underflow qgroup meta
reservation.
Since we have the new qgroup meta rsv types, apply it to delalloc
reservation.
Now for delalloc, most of its reserved space will use META_PREALLOC qgroup
rsv type.
And for callers reducing outstanding extent like btrfs_finish_ordered_io(),
they will convert corresponding META_PREALLOC reservation to
META_PERTRANS.
This is mainly due to the fact that current qgroup numbers will only be
updated in btrfs_commit_transaction(), that's to say if we don't keep
such placeholder reservation, we can exceed qgroup limitation.
And for callers freeing outstanding extent in error handler, we will
just free META_PREALLOC bytes.
This behavior makes callers of btrfs_qgroup_release_meta() or
btrfs_qgroup_convert_meta() to be aware of which type they are.
So in this patch, btrfs_delalloc_release_metadata() and its callers get
an extra parameter to info qgroup to do correct meta convert/release.
The good news is, even we use the wrong type (convert or free), it won't
cause obvious bug, as prealloc type is always in good shape, and the
type only affects how per-trans meta is increased or not.
So the worst case will be at most metadata limitation can be sometimes
exceeded (no convert at all) or metadata limitation is reached too soon
(no free at all).
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The __cold functions are placed to a special section, as they're
expected to be called rarely. This could help i-cache prefetches or help
compiler to decide which branches are more/less likely to be taken
without any other annotations needed.
Though we can't add more __exit annotations, it's still possible to add
__cold (that's also added with __exit). That way the following function
categories are tagged:
- printf wrappers, error messages
- exit helpers
Signed-off-by: David Sterba <dsterba@suse.com>
Right now we do a lot of weird hoops around outstanding_extents in order
to keep the extent count consistent. This is because we logically
transfer the outstanding_extent count from the initial reservation
through the set_delalloc_bits. This makes it pretty difficult to get a
handle on how and when we need to mess with outstanding_extents.
Fix this by revamping the rules of how we deal with outstanding_extents.
Now instead everybody that is holding on to a delalloc extent is
required to increase the outstanding extents count for itself. This
means we'll have something like this
btrfs_delalloc_reserve_metadata - outstanding_extents = 1
btrfs_set_extent_delalloc - outstanding_extents = 2
btrfs_release_delalloc_extents - outstanding_extents = 1
for an initial file write. Now take the append write where we extend an
existing delalloc range but still under the maximum extent size
btrfs_delalloc_reserve_metadata - outstanding_extents = 2
btrfs_set_extent_delalloc
btrfs_set_bit_hook - outstanding_extents = 3
btrfs_merge_extent_hook - outstanding_extents = 2
btrfs_delalloc_release_extents - outstanding_extnets = 1
In order to make the ordered extent transition we of course must now
make ordered extents carry their own outstanding_extent reservation, so
for cow_file_range we end up with
btrfs_add_ordered_extent - outstanding_extents = 2
clear_extent_bit - outstanding_extents = 1
btrfs_remove_ordered_extent - outstanding_extents = 0
This makes all manipulations of outstanding_extents much more explicit.
Every successful call to btrfs_delalloc_reserve_metadata _must_ now be
combined with btrfs_release_delalloc_extents, even in the error case, as
that is the only function that actually modifies the
outstanding_extents counter.
The drawback to this is now we are much more likely to have transient
cases where outstanding_extents is much larger than it actually should
be. This could happen before as we manipulated the delalloc bits, but
now it happens basically at every write. This may put more pressure on
the ENOSPC flushing code, but I think making this code simpler is worth
the cost. I have another change coming to mitigate this side-effect
somewhat.
I also added trace points for the counter manipulation. These were used
by a bpf script I wrote to help track down leak issues.
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Dave Jones hit a WARN_ON(nr < 0) in btrfs_wait_ordered_roots() with
v4.12-rc6. This was because commit 70e7af244 made it possible for
calc_reclaim_items_nr() to return a negative number. It's not really a
bug in that commit, it just didn't go far enough down the stack to find
all the possible 64->32 bit overflows.
This switches calc_reclaim_items_nr() to return a u64 and changes everyone
that uses the results of that math to u64 as well.
Reported-by: Dave Jones <davej@codemonkey.org.uk>
Fixes: 70e7af2 ("Btrfs: fix delalloc accounting leak caused by u32 overflow")
Signed-off-by: Chris Mason <clm@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
refcount_t type and corresponding API should be
used instead of atomic_t when the variable is used as
a reference counter. This allows to avoid accidental
refcounter overflows that might lead to use-after-free
situations.
Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
Signed-off-by: Hans Liljestrand <ishkamiel@gmail.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David Windsor <dwindsor@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
refcount_t type and corresponding API should be
used instead of atomic_t when the variable is used as
a reference counter. This allows to avoid accidental
refcounter overflows that might lead to use-after-free
situations.
Signed-off-by: Elena Reshetova <elena.reshetova@intel.com>
Signed-off-by: Hans Liljestrand <ishkamiel@gmail.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: David Windsor <dwindsor@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Since we have a good helper entry_end, use it for ordered extent.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ whitespace reformatting ]
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_ordered_update_i_size can be called by truncate and endio, but
only endio takes ordered_extent which contains the completed IO.
while truncating down a file, if there are some in-flight IOs,
btrfs_ordered_update_i_size in endio will set disk_i_size to
@orig_offset that is zero. If truncating-down fails somehow, we try to
recover in memory isize with this zero'd disk_i_size.
Fix it by only updating disk_i_size with @orig_offset when
btrfs_ordered_update_i_size is not called from endio while truncating
down and waiting for in-flight IOs completing their work before recover
in-memory size.
Besides fixing the above issue, add an assertion for last_size to double
check we truncate down to the desired size.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In routines where someptr->fs_info is referenced multiple times, we
introduce a convenience variable. This makes the code considerably
more readable.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We track the node sizes per-root, but they never vary from the values
in the superblock. This patch messes with the 80-column style a bit,
but subsequent patches to factor out root->fs_info into a convenience
variable fix it up again.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
CodingStyle chapter 2:
"[...] never break user-visible strings such as printk messages,
because that breaks the ability to grep for them."
This patch unsplits user-visible strings.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
BTRFS is using a variety of slab caches to satisfy internal needs.
Those slab caches are always allocated with the SLAB_RECLAIM_ACCOUNT,
meaning allocations from the caches are going to be accounted as
SReclaimable. At the same time btrfs is not registering any shrinkers
whatsoever, thus preventing memory from the slabs to be shrunk. This
means those caches are not in fact reclaimable.
To fix this remove the SLAB_RECLAIM_ACCOUNT on all caches apart from the
inode cache, since this one is being freed by the generic VFS super_block
shrinker. Also set the transaction related caches as SLAB_TEMPORARY,
to better document the lifetime of the objects (it just translates
to SLAB_RECLAIM_ACCOUNT).
Signed-off-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When doing truncate operation, btrfs_setsize() will first call
truncate_setsize() to set new inode->i_size, but if later
btrfs_truncate() fails, btrfs_setsize() will call
"i_size_write(inode, BTRFS_I(inode)->disk_i_size)" to reset the
inmemory inode size, now bug occurs. It's because for truncate
case btrfs_ordered_update_i_size() directly uses inode->i_size
to update BTRFS_I(inode)->disk_i_size, indeed we should use the
"offset" argument to update disk_i_size. Here is the call graph:
==>btrfs_truncate()
====>btrfs_truncate_inode_items()
======>btrfs_ordered_update_i_size(inode, last_size, NULL);
Here btrfs_ordered_update_i_size()'s offset argument is last_size.
And below test case can reveal this bug:
dd if=/dev/zero of=fs.img bs=$((1024*1024)) count=100
dev=$(losetup --show -f fs.img)
mkdir -p /mnt/mntpoint
mkfs.btrfs -f $dev
mount $dev /mnt/mntpoint
cd /mnt/mntpoint
echo "workdir is: /mnt/mntpoint"
blocksize=$((128 * 1024))
dd if=/dev/zero of=testfile bs=$blocksize count=1
sync
count=$((17*1024*1024*1024/blocksize))
echo "file size is:" $((count*blocksize))
for ((i = 1; i <= $count; i++)); do
i=$((i + 1))
dst_offset=$((blocksize * i))
xfs_io -f -c "reflink testfile 0 $dst_offset $blocksize"\
testfile > /dev/null
done
sync
truncate --size 0 testfile
ls -l testfile
du -sh testfile
exit
In this case, truncate operation will fail for enospc reason and
"du -sh testfile" returns value greater than 0, but testfile's
size is 0, we need to reflect correct inode->i_size.
Signed-off-by: Wang Xiaoguang <wangxg.fnst@cn.fujitsu.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
When we do a device replace, for each device extent we find from the
source device, we set the corresponding block group to readonly mode to
prevent writes into it from happening while we are copying the device
extent from the source to the target device. However just before we set
the block group to readonly mode some concurrent task might have already
allocated an extent from it or decided it could perform a nocow write
into one of its extents, which can make the device replace process to
miss copying an extent since it uses the extent tree's commit root to
search for extents and only once it finishes searching for all extents
belonging to the block group it does set the left cursor to the logical
end address of the block group - this is a problem if the respective
ordered extents finish while we are searching for extents using the
extent tree's commit root and no transaction commit happens while we
are iterating the tree, since it's the delayed references created by the
ordered extents (when they complete) that insert the extent items into
the extent tree (using the non-commit root of course).
Example:
CPU 1 CPU 2
btrfs_dev_replace_start()
btrfs_scrub_dev()
scrub_enumerate_chunks()
--> finds device extent belonging
to block group X
<transaction N starts>
starts buffered write
against some inode
writepages is run against
that inode forcing dellaloc
to run
btrfs_writepages()
extent_writepages()
extent_write_cache_pages()
__extent_writepage()
writepage_delalloc()
run_delalloc_range()
cow_file_range()
btrfs_reserve_extent()
--> allocates an extent
from block group X
(which is not yet
in RO mode)
btrfs_add_ordered_extent()
--> creates ordered extent Y
flush_epd_write_bio()
--> bio against the extent from
block group X is submitted
btrfs_inc_block_group_ro(bg X)
--> sets block group X to readonly
scrub_chunk(bg X)
scrub_stripe(device extent from srcdev)
--> keeps searching for extent items
belonging to the block group using
the extent tree's commit root
--> it never blocks due to
fs_info->scrub_pause_req as no
one tries to commit transaction N
--> copies all extents found from the
source device into the target device
--> finishes search loop
bio completes
ordered extent Y completes
and creates delayed data
reference which will add an
extent item to the extent
tree when run (typically
at transaction commit time)
--> so the task doing the
scrub/device replace
at CPU 1 misses this
and does not copy this
extent into the new/target
device
btrfs_dec_block_group_ro(bg X)
--> turns block group X back to RW mode
dev_replace->cursor_left is set to the
logical end offset of block group X
So fix this by waiting for all cow and nocow writes after setting a block
group to readonly mode.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Josef Bacik <jbacik@fb.com>
Before the relocation process of a block group starts, it sets the block
group to readonly mode, then flushes all delalloc writes and then finally
it waits for all ordered extents to complete. This last step includes
waiting for ordered extents destinated at extents allocated in other block
groups, making us waste unecessary time.
So improve this by waiting only for ordered extents that fall into the
block group's range.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: Liu Bo <bo.li.liu@oracle.com>
So that its better organized.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Cleanup.
kmem_cache_destroy has support NULL argument checking,
so drop the double null testing before calling it.
Signed-off-by: Kinglong Mee <kinglongmee@gmail.com>
Signed-off-by: David Sterba <dsterba@suse.com>
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>
If we fail to submit a bio for a direct IO request, we were grabbing the
corresponding ordered extent and decrementing its reference count twice,
once for our lookup reference and once for the ordered tree reference.
This was a problem because it caused the ordered extent to be freed
without removing it from the ordered tree and any lists it might be
attached to, leaving dangling pointers to the ordered extent around.
Example trace with CONFIG_DEBUG_PAGEALLOC=y:
[161779.858707] BUG: unable to handle kernel paging request at 0000000087654330
[161779.859983] IP: [<ffffffff8124ca68>] rb_prev+0x22/0x3b
[161779.860636] PGD 34d818067 PUD 0
[161779.860636] Oops: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC
(...)
[161779.860636] Call Trace:
[161779.860636] [<ffffffffa06b36a6>] __tree_search+0xd9/0xf9 [btrfs]
[161779.860636] [<ffffffffa06b3708>] tree_search+0x42/0x63 [btrfs]
[161779.860636] [<ffffffffa06b4868>] ? btrfs_lookup_ordered_range+0x2d/0xa5 [btrfs]
[161779.860636] [<ffffffffa06b4873>] btrfs_lookup_ordered_range+0x38/0xa5 [btrfs]
[161779.860636] [<ffffffffa06aab8e>] btrfs_get_blocks_direct+0x11b/0x615 [btrfs]
[161779.860636] [<ffffffff8119727f>] do_blockdev_direct_IO+0x5ff/0xb43
[161779.860636] [<ffffffffa06aaa73>] ? btrfs_page_exists_in_range+0x1ad/0x1ad [btrfs]
[161779.860636] [<ffffffffa06a2c9a>] ? btrfs_get_extent_fiemap+0x1bc/0x1bc [btrfs]
[161779.860636] [<ffffffff811977f5>] __blockdev_direct_IO+0x32/0x34
[161779.860636] [<ffffffffa06a2c9a>] ? btrfs_get_extent_fiemap+0x1bc/0x1bc [btrfs]
[161779.860636] [<ffffffffa06a10ae>] btrfs_direct_IO+0x198/0x21f [btrfs]
[161779.860636] [<ffffffffa06a2c9a>] ? btrfs_get_extent_fiemap+0x1bc/0x1bc [btrfs]
[161779.860636] [<ffffffff81112ca1>] generic_file_direct_write+0xb3/0x128
[161779.860636] [<ffffffffa06affaa>] ? btrfs_file_write_iter+0x15f/0x3e0 [btrfs]
[161779.860636] [<ffffffffa06b004c>] btrfs_file_write_iter+0x201/0x3e0 [btrfs]
(...)
We were also not freeing the btrfs_dio_private we allocated previously,
which kmemleak reported with the following trace in its sysfs file:
unreferenced object 0xffff8803f553bf80 (size 96):
comm "xfs_io", pid 4501, jiffies 4295039588 (age 173.936s)
hex dump (first 32 bytes):
88 6c 9b f5 02 88 ff ff 00 00 00 00 00 00 00 00 .l..............
00 00 00 00 00 00 00 00 00 00 c4 00 00 00 00 00 ................
backtrace:
[<ffffffff81161ffe>] create_object+0x172/0x29a
[<ffffffff8145870f>] kmemleak_alloc+0x25/0x41
[<ffffffff81154e64>] kmemleak_alloc_recursive.constprop.40+0x16/0x18
[<ffffffff811579ed>] kmem_cache_alloc_trace+0xfb/0x148
[<ffffffffa03d8cff>] btrfs_submit_direct+0x65/0x16a [btrfs]
[<ffffffff811968dc>] dio_bio_submit+0x62/0x8f
[<ffffffff811975fe>] do_blockdev_direct_IO+0x97e/0xb43
[<ffffffff811977f5>] __blockdev_direct_IO+0x32/0x34
[<ffffffffa03d70ae>] btrfs_direct_IO+0x198/0x21f [btrfs]
[<ffffffff81112ca1>] generic_file_direct_write+0xb3/0x128
[<ffffffffa03e604d>] btrfs_file_write_iter+0x201/0x3e0 [btrfs]
[<ffffffff8116586a>] __vfs_write+0x7c/0xa5
[<ffffffff81165da9>] vfs_write+0xa0/0xe4
[<ffffffff81166675>] SyS_pwrite64+0x64/0x82
[<ffffffff81464fd7>] system_call_fastpath+0x12/0x6f
[<ffffffffffffffff>] 0xffffffffffffffff
For read requests we weren't doing any cleanup either (none of the work
done by btrfs_endio_direct_read()), so a failure submitting a bio for a
read request would leave a range in the inode's io_tree locked forever,
blocking any future operations (both reads and writes) against that range.
So fix this by making sure we do the same cleanup that we do for the case
where the bio submission succeeds.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
We don't need to attach ordered extents that have completed to the current
transaction. Doing so only makes us hold memory for longer than necessary
and delaying the iput of the inode until the transaction is committed (for
each created ordered extent we do an igrab and then schedule an asynchronous
iput when the ordered extent's reference count drops to 0), preventing the
inode from being evictable until the transaction commits.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Commit 3a8b36f378 ("Btrfs: fix data loss in the fast fsync path") added
a performance regression for that causes an unnecessary sync of the log
trees (fs/subvol and root log trees) when 2 consecutive fsyncs are done
against a file, without no writes or any metadata updates to the inode in
between them and if a transaction is committed before the second fsync is
called.
Huang Ying reported this to lkml (https://lkml.org/lkml/2015/3/18/99)
after a test sysbench test that measured a -62% decrease of file io
requests per second for that tests' workload.
The test is:
echo performance > /sys/devices/system/cpu/cpu0/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu1/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu2/cpufreq/scaling_governor
echo performance > /sys/devices/system/cpu/cpu3/cpufreq/scaling_governor
mkfs -t btrfs /dev/sda2
mount -t btrfs /dev/sda2 /fs/sda2
cd /fs/sda2
for ((i = 0; i < 1024; i++)); do fallocate -l 67108864 testfile.$i; done
sysbench --test=fileio --max-requests=0 --num-threads=4 --max-time=600 \
--file-test-mode=rndwr --file-total-size=68719476736 --file-io-mode=sync \
--file-num=1024 run
A test on kvm guest, running a debug kernel gave me the following results:
Without 3a8b36f378: 16.01 reqs/sec
With 3a8b36f378: 3.39 reqs/sec
With 3a8b36f378 and this patch: 16.04 reqs/sec
Reported-by: Huang Ying <ying.huang@intel.com>
Tested-by: Huang, Ying <ying.huang@intel.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
After commit 8407f55326
("Btrfs: fix data corruption after fast fsync and writeback error"),
during wait_ordered_extents(), we wait for ordered extent setting
BTRFS_ORDERED_IO_DONE or BTRFS_ORDERED_IOERR, at which point we've
already got checksum information, so we don't need to check
(csum_bytes_left == 0) in the whole logging path.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Chris Mason <clm@fb.com>
When waiting for the writeback of block group cache we returned
immediately if there was an error during writeback without waiting
for the ordered extent to complete. This left a short time window
where if some other task attempts to start the writeout for the same
block group cache it can attempt to add a new ordered extent, starting
at the same offset (0) before the previous one is removed from the
ordered tree, causing an ordered tree panic (calls BUG()).
This normally doesn't happen in other write paths, such as buffered
writes or direct IO writes for regular files, since before marking
page ranges dirty we lock the ranges and wait for any ordered extents
within the range to complete first.
Fix this by making btrfs_wait_ordered_range() not return immediately
if it gets an error from the writeback, waiting for all ordered extents
to complete first.
This issue happened often when running the fstest btrfs/088 and it's
easy to trigger it by running in a loop until the panic happens:
for ((i = 1; i <= 10000; i++)) do ./check btrfs/088 ; done
[17156.862573] BTRFS critical (device sdc): panic in ordered_data_tree_panic:70: Inconsistency in ordered tree at offset 0 (errno=-17 Object already exists)
[17156.864052] ------------[ cut here ]------------
[17156.864052] kernel BUG at fs/btrfs/ordered-data.c:70!
(...)
[17156.864052] Call Trace:
[17156.864052] [<ffffffffa03876e3>] btrfs_add_ordered_extent+0x12/0x14 [btrfs]
[17156.864052] [<ffffffffa03787e2>] run_delalloc_nocow+0x5bf/0x747 [btrfs]
[17156.864052] [<ffffffffa03789ff>] run_delalloc_range+0x95/0x353 [btrfs]
[17156.864052] [<ffffffffa038b7fe>] writepage_delalloc.isra.16+0xb9/0x13f [btrfs]
[17156.864052] [<ffffffffa038d75b>] __extent_writepage+0x129/0x1f7 [btrfs]
[17156.864052] [<ffffffffa038da5a>] extent_write_cache_pages.isra.15.constprop.28+0x231/0x2f4 [btrfs]
[17156.864052] [<ffffffff810ad2af>] ? __module_text_address+0x12/0x59
[17156.864052] [<ffffffff8107d33d>] ? trace_hardirqs_on+0xd/0xf
[17156.864052] [<ffffffffa038df76>] extent_writepages+0x4b/0x5c [btrfs]
[17156.864052] [<ffffffff81144431>] ? kmem_cache_free+0x9b/0xce
[17156.864052] [<ffffffffa0376a46>] ? btrfs_submit_direct+0x3fc/0x3fc [btrfs]
[17156.864052] [<ffffffffa0389cd6>] ? free_extent_state+0x8c/0xc1 [btrfs]
[17156.864052] [<ffffffffa0374871>] btrfs_writepages+0x28/0x2a [btrfs]
[17156.864052] [<ffffffff8110c4c8>] do_writepages+0x23/0x2c
[17156.864052] [<ffffffff81102f36>] __filemap_fdatawrite_range+0x5a/0x61
[17156.864052] [<ffffffff81102f6e>] filemap_fdatawrite_range+0x13/0x15
[17156.864052] [<ffffffffa0383ef7>] btrfs_fdatawrite_range+0x21/0x48 [btrfs]
[17156.864052] [<ffffffffa03ab89e>] __btrfs_write_out_cache.isra.14+0x2d9/0x3a7 [btrfs]
[17156.864052] [<ffffffffa03ac1ab>] ? btrfs_write_out_cache+0x41/0xdc [btrfs]
[17156.864052] [<ffffffffa03ac1fd>] btrfs_write_out_cache+0x93/0xdc [btrfs]
[17156.864052] [<ffffffffa0363847>] ? btrfs_start_dirty_block_groups+0x13a/0x2b2 [btrfs]
[17156.864052] [<ffffffffa03638e6>] btrfs_start_dirty_block_groups+0x1d9/0x2b2 [btrfs]
[17156.864052] [<ffffffff8107d33d>] ? trace_hardirqs_on+0xd/0xf
[17156.864052] [<ffffffffa037209e>] btrfs_commit_transaction+0x130/0x9c9 [btrfs]
[17156.864052] [<ffffffffa034c748>] btrfs_sync_fs+0xe1/0x12d [btrfs]
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
We can have multiple fsync operations against the same file during the
same transaction and they can collect the same ordered extents while they
don't complete (still accessible from the inode's ordered tree). If this
happens, those ordered extents will never get their reference counts
decremented to 0, leading to memory leaks and inode leaks (an iput for an
ordered extent's inode is scheduled only when the ordered extent's refcount
drops to 0). The following sequence diagram explains this race:
CPU 1 CPU 2
btrfs_sync_file()
btrfs_sync_file()
mutex_lock(inode->i_mutex)
btrfs_log_inode()
btrfs_get_logged_extents()
--> collects ordered extent X
--> increments ordered
extent X's refcount
btrfs_submit_logged_extents()
mutex_unlock(inode->i_mutex)
mutex_lock(inode->i_mutex)
btrfs_sync_log()
btrfs_wait_logged_extents()
--> list_del_init(&ordered->log_list)
btrfs_log_inode()
btrfs_get_logged_extents()
--> Adds ordered extent X
to logged_list because
at this point:
list_empty(&ordered->log_list)
&& test_bit(BTRFS_ORDERED_LOGGED,
&ordered->flags) == 0
--> Increments ordered extent
X's refcount
--> check if ordered extent's io is
finished or not, start it if
necessary and wait for it to finish
--> sets bit BTRFS_ORDERED_LOGGED
on ordered extent X's flags
and adds it to trans->ordered
btrfs_sync_log() finishes
btrfs_submit_logged_extents()
btrfs_log_inode() finishes
mutex_unlock(inode->i_mutex)
btrfs_sync_file() finishes
btrfs_sync_log()
btrfs_wait_logged_extents()
--> Sees ordered extent X has the
bit BTRFS_ORDERED_LOGGED set in
its flags
--> X's refcount is untouched
btrfs_sync_log() finishes
btrfs_sync_file() finishes
btrfs_commit_transaction()
--> called by transaction kthread for e.g.
btrfs_wait_pending_ordered()
--> waits for ordered extent X to
complete
--> decrements ordered extent X's
refcount by 1 only, corresponding
to the increment done by the fsync
task ran by CPU 1
In the scenario of the above diagram, after the transaction commit,
the ordered extent will remain with a refcount of 1 forever, leaking
the ordered extent structure and preventing the i_count of its inode
from ever decreasing to 0, since the delayed iput is scheduled only
when the ordered extent's refcount drops to 0, preventing the inode
from ever being evicted by the VFS.
Fix this by using the flag BTRFS_ORDERED_LOGGED differently. Use it to
mean that an ordered extent is already being processed by an fsync call,
which will attach it to the current transaction, preventing it from being
collected by subsequent fsync operations against the same inode.
This race was introduced with the following change (added in 3.19 and
backported to stable 3.18 and 3.17):
Btrfs: make sure logged extents complete in the current transaction V3
commit 50d9aa99bd
I ran into this issue while running xfstests/generic/113 in a loop, which
failed about 1 out of 10 runs with the following warning in dmesg:
[ 2612.440038] WARNING: CPU: 4 PID: 22057 at fs/btrfs/disk-io.c:3558 free_fs_root+0x36/0x133 [btrfs]()
[ 2612.442810] Modules linked in: btrfs crc32c_generic xor raid6_pq nfsd auth_rpcgss oid_registry nfs_acl nfs lockd grace fscache sunrpc loop processor parport_pc parport psmouse therma
l_sys i2c_piix4 serio_raw pcspkr evdev microcode button i2c_core ext4 crc16 jbd2 mbcache sd_mod sg sr_mod cdrom virtio_scsi ata_generic virtio_pci ata_piix virtio_ring libata virtio flo
ppy e1000 scsi_mod [last unloaded: btrfs]
[ 2612.452711] CPU: 4 PID: 22057 Comm: umount Tainted: G W 3.19.0-rc5-btrfs-next-4+ #1
[ 2612.454921] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
[ 2612.457709] 0000000000000009 ffff8801342c3c78 ffffffff8142425e ffff88023ec8f2d8
[ 2612.459829] 0000000000000000 ffff8801342c3cb8 ffffffff81045308 ffff880046460000
[ 2612.461564] ffffffffa036da56 ffff88003d07b000 ffff880046460000 ffff880046460068
[ 2612.463163] Call Trace:
[ 2612.463719] [<ffffffff8142425e>] dump_stack+0x4c/0x65
[ 2612.464789] [<ffffffff81045308>] warn_slowpath_common+0xa1/0xbb
[ 2612.466026] [<ffffffffa036da56>] ? free_fs_root+0x36/0x133 [btrfs]
[ 2612.467247] [<ffffffff810453c5>] warn_slowpath_null+0x1a/0x1c
[ 2612.468416] [<ffffffffa036da56>] free_fs_root+0x36/0x133 [btrfs]
[ 2612.469625] [<ffffffffa036f2a7>] btrfs_drop_and_free_fs_root+0x93/0x9b [btrfs]
[ 2612.471251] [<ffffffffa036f353>] btrfs_free_fs_roots+0xa4/0xd6 [btrfs]
[ 2612.472536] [<ffffffff8142612e>] ? wait_for_completion+0x24/0x26
[ 2612.473742] [<ffffffffa0370bbc>] close_ctree+0x1f3/0x33c [btrfs]
[ 2612.475477] [<ffffffff81059d1d>] ? destroy_workqueue+0x148/0x1ba
[ 2612.476695] [<ffffffffa034e3da>] btrfs_put_super+0x19/0x1b [btrfs]
[ 2612.477911] [<ffffffff81153e53>] generic_shutdown_super+0x73/0xef
[ 2612.479106] [<ffffffff811540e2>] kill_anon_super+0x13/0x1e
[ 2612.480226] [<ffffffffa034e1e3>] btrfs_kill_super+0x17/0x23 [btrfs]
[ 2612.481471] [<ffffffff81154307>] deactivate_locked_super+0x3b/0x50
[ 2612.482686] [<ffffffff811547a7>] deactivate_super+0x3f/0x43
[ 2612.483791] [<ffffffff8116b3ed>] cleanup_mnt+0x59/0x78
[ 2612.484842] [<ffffffff8116b44c>] __cleanup_mnt+0x12/0x14
[ 2612.485900] [<ffffffff8105d019>] task_work_run+0x8f/0xbc
[ 2612.486960] [<ffffffff810028d8>] do_notify_resume+0x5a/0x6b
[ 2612.488083] [<ffffffff81236e5b>] ? trace_hardirqs_on_thunk+0x3a/0x3f
[ 2612.489333] [<ffffffff8142a17f>] int_signal+0x12/0x17
[ 2612.490353] ---[ end trace 54a960a6bdcb8d93 ]---
[ 2612.557253] VFS: Busy inodes after unmount of sdb. Self-destruct in 5 seconds. Have a nice day...
Kmemleak confirmed the ordered extent leak (and btrfs inode specific
structures such as delayed nodes):
$ cat /sys/kernel/debug/kmemleak
unreferenced object 0xffff880154290db0 (size 576):
comm "btrfsck", pid 21980, jiffies 4295542503 (age 1273.412s)
hex dump (first 32 bytes):
01 40 00 00 01 00 00 00 b0 1d f1 4e 01 88 ff ff .@.........N....
00 00 00 00 00 00 00 00 c8 0d 29 54 01 88 ff ff ..........)T....
backtrace:
[<ffffffff8141d74d>] kmemleak_update_trace+0x4c/0x6a
[<ffffffff8122f2c0>] radix_tree_node_alloc+0x6d/0x83
[<ffffffff8122fb26>] __radix_tree_create+0x109/0x190
[<ffffffff8122fbdd>] radix_tree_insert+0x30/0xac
[<ffffffffa03b9bde>] btrfs_get_or_create_delayed_node+0x130/0x187 [btrfs]
[<ffffffffa03bb82d>] btrfs_delayed_delete_inode_ref+0x32/0xac [btrfs]
[<ffffffffa0379dae>] __btrfs_unlink_inode+0xee/0x288 [btrfs]
[<ffffffffa037c715>] btrfs_unlink_inode+0x1e/0x40 [btrfs]
[<ffffffffa037c797>] btrfs_unlink+0x60/0x9b [btrfs]
[<ffffffff8115d7f0>] vfs_unlink+0x9c/0xed
[<ffffffff8115f5de>] do_unlinkat+0x12c/0x1fa
[<ffffffff811601a7>] SyS_unlinkat+0x29/0x2b
[<ffffffff81429e92>] system_call_fastpath+0x12/0x17
[<ffffffffffffffff>] 0xffffffffffffffff
unreferenced object 0xffff88014ef11db0 (size 576):
comm "rm", pid 22009, jiffies 4295542593 (age 1273.052s)
hex dump (first 32 bytes):
02 00 00 00 01 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 c8 1d f1 4e 01 88 ff ff ...........N....
backtrace:
[<ffffffff8141d74d>] kmemleak_update_trace+0x4c/0x6a
[<ffffffff8122f2c0>] radix_tree_node_alloc+0x6d/0x83
[<ffffffff8122fb26>] __radix_tree_create+0x109/0x190
[<ffffffff8122fbdd>] radix_tree_insert+0x30/0xac
[<ffffffffa03b9bde>] btrfs_get_or_create_delayed_node+0x130/0x187 [btrfs]
[<ffffffffa03bb82d>] btrfs_delayed_delete_inode_ref+0x32/0xac [btrfs]
[<ffffffffa0379dae>] __btrfs_unlink_inode+0xee/0x288 [btrfs]
[<ffffffffa037c715>] btrfs_unlink_inode+0x1e/0x40 [btrfs]
[<ffffffffa037c797>] btrfs_unlink+0x60/0x9b [btrfs]
[<ffffffff8115d7f0>] vfs_unlink+0x9c/0xed
[<ffffffff8115f5de>] do_unlinkat+0x12c/0x1fa
[<ffffffff811601a7>] SyS_unlinkat+0x29/0x2b
[<ffffffff81429e92>] system_call_fastpath+0x12/0x17
[<ffffffffffffffff>] 0xffffffffffffffff
unreferenced object 0xffff8800336feda8 (size 584):
comm "aio-stress", pid 22031, jiffies 4295543006 (age 1271.400s)
hex dump (first 32 bytes):
00 40 3e 00 00 00 00 00 00 00 8f 42 00 00 00 00 .@>........B....
00 00 01 00 00 00 00 00 00 00 01 00 00 00 00 00 ................
backtrace:
[<ffffffff8114eb34>] create_object+0x172/0x29a
[<ffffffff8141d790>] kmemleak_alloc+0x25/0x41
[<ffffffff81141ae6>] kmemleak_alloc_recursive.constprop.52+0x16/0x18
[<ffffffff81145288>] kmem_cache_alloc+0xf7/0x198
[<ffffffffa0389243>] __btrfs_add_ordered_extent+0x43/0x309 [btrfs]
[<ffffffffa038968b>] btrfs_add_ordered_extent_dio+0x12/0x14 [btrfs]
[<ffffffffa03810e2>] btrfs_get_blocks_direct+0x3ef/0x571 [btrfs]
[<ffffffff81181349>] do_blockdev_direct_IO+0x62a/0xb47
[<ffffffff8118189a>] __blockdev_direct_IO+0x34/0x36
[<ffffffffa03776e5>] btrfs_direct_IO+0x16a/0x1e8 [btrfs]
[<ffffffff81100373>] generic_file_direct_write+0xb8/0x12d
[<ffffffffa038615c>] btrfs_file_write_iter+0x24b/0x42f [btrfs]
[<ffffffff8118bb0d>] aio_run_iocb+0x2b7/0x32e
[<ffffffff8118c99a>] do_io_submit+0x26e/0x2ff
[<ffffffff8118ca3b>] SyS_io_submit+0x10/0x12
[<ffffffff81429e92>] system_call_fastpath+0x12/0x17
CC: <stable@vger.kernel.org> # 3.19, 3.18 and 3.17
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Instead of collecting all ordered extents from the inode's ordered tree
and then wait for all of them to complete, just collect the ones that
overlap the fsync range.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
Liu Bo pointed out that my previous fix would lose the generation update in the
scenario I described. It is actually much worse than that, we could lose the
entire extent if we lose power right after the transaction commits. Consider
the following
write extent 0-4k
log extent in log tree
commit transaction
< power fail happens here
ordered extent completes
We would lose the 0-4k extent because it hasn't updated the actual fs tree, and
the transaction commit will reset the log so it isn't replayed. If we lose
power before the transaction commit we are save, otherwise we are not.
Fix this by keeping track of all extents we logged in this transaction. Then
when we go to commit the transaction make sure we wait for all of those ordered
extents to complete before proceeding. This will make sure that if we lose
power after the transaction commit we still have our data. This also fixes the
problem of the improperly updated extent generation. Thanks,
cc: stable@vger.kernel.org
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: Chris Mason <clm@fb.com>
To avoid duplicating this double filemap_fdatawrite_range() call for
inodes with async extents (compressed writes) so often.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Chris Mason <clm@fb.com>
This has been reported and discussed for a long time, and this hang occurs in
both 3.15 and 3.16.
Btrfs now migrates to use kernel workqueue, but it introduces this hang problem.
Btrfs has a kind of work queued as an ordered way, which means that its
ordered_func() must be processed in the way of FIFO, so it usually looks like --
normal_work_helper(arg)
work = container_of(arg, struct btrfs_work, normal_work);
work->func() <---- (we name it work X)
for ordered_work in wq->ordered_list
ordered_work->ordered_func()
ordered_work->ordered_free()
The hang is a rare case, first when we find free space, we get an uncached block
group, then we go to read its free space cache inode for free space information,
so it will
file a readahead request
btrfs_readpages()
for page that is not in page cache
__do_readpage()
submit_extent_page()
btrfs_submit_bio_hook()
btrfs_bio_wq_end_io()
submit_bio()
end_workqueue_bio() <--(ret by the 1st endio)
queue a work(named work Y) for the 2nd
also the real endio()
So the hang occurs when work Y's work_struct and work X's work_struct happens
to share the same address.
A bit more explanation,
A,B,C -- struct btrfs_work
arg -- struct work_struct
kthread:
worker_thread()
pick up a work_struct from @worklist
process_one_work(arg)
worker->current_work = arg; <-- arg is A->normal_work
worker->current_func(arg)
normal_work_helper(arg)
A = container_of(arg, struct btrfs_work, normal_work);
A->func()
A->ordered_func()
A->ordered_free() <-- A gets freed
B->ordered_func()
submit_compressed_extents()
find_free_extent()
load_free_space_inode()
... <-- (the above readhead stack)
end_workqueue_bio()
btrfs_queue_work(work C)
B->ordered_free()
As if work A has a high priority in wq->ordered_list and there are more ordered
works queued after it, such as B->ordered_func(), its memory could have been
freed before normal_work_helper() returns, which means that kernel workqueue
code worker_thread() still has worker->current_work pointer to be work
A->normal_work's, ie. arg's address.
Meanwhile, work C is allocated after work A is freed, work C->normal_work
and work A->normal_work are likely to share the same address(I confirmed this
with ftrace output, so I'm not just guessing, it's rare though).
When another kthread picks up work C->normal_work to process, and finds our
kthread is processing it(see find_worker_executing_work()), it'll think
work C as a collision and skip then, which ends up nobody processing work C.
So the situation is that our kthread is waiting forever on work C.
Besides, there're other cases that can lead to deadlock, but the real problem
is that all btrfs workqueue shares one work->func, -- normal_work_helper,
so this makes each workqueue to have its own helper function, but only a
wraper pf normal_work_helper.
With this patch, I no long hit the above hang.
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: Chris Mason <clm@fb.com>
Truncates and renames are often used to replace old versions of a file
with new versions. Applications often expect this to be an atomic
replacement, even if they haven't done anything to make sure the new
version is fully on disk.
Btrfs has strict flushing in place to make sure that renaming over an
old file with a new file will fully flush out the new file before
allowing the transaction commit with the rename to complete.
This ordering means the commit code needs to be able to lock file pages,
and there are a few paths in the filesystem where we will try to end a
transaction with the page lock held. It's rare, but these things can
deadlock.
This patch removes the ordered flushes and switches to a best effort
filemap_flush like ext4 uses. It's not perfect, but it should fix the
deadlocks.
Signed-off-by: Chris Mason <clm@fb.com>
