After a succession of renames operations of different files and unlinking
one of them, if we fsync one of the renamed files we can end up with a
log that will either fail to replay at mount time or result in a filesystem
that is in an inconsistent state. One example scenario:
$ mkfs.btrfs -f /dev/sdb
$ mount /dev/sdb /mnt
$ mkdir /mnt/testdir
$ touch /mnt/testdir/fname1
$ touch /mnt/testdir/fname2
$ sync
$ mv /mnt/testdir/fname1 /mnt/testdir/fname3
$ rm -f /mnt/testdir/fname2
$ ln /mnt/testdir/fname3 /mnt/testdir/fname2
$ touch /mnt/testdir/fname1
$ xfs_io -c "fsync" /mnt/testdir/fname1
<power failure>
$ mount /dev/sdb /mnt
$ umount /mnt
$ btrfs check /dev/sdb
[1/7] checking root items
[2/7] checking extents
[3/7] checking free space cache
[4/7] checking fs roots
root 5 inode 259 errors 2, no orphan item
ERROR: errors found in fs roots
Opening filesystem to check...
Checking filesystem on /dev/sdc
UUID: 20e4abb8-5a19-4492-8bb4-6084125c2d0d
found 393216 bytes used, error(s) found
total csum bytes: 0
total tree bytes: 131072
total fs tree bytes: 32768
total extent tree bytes: 16384
btree space waste bytes: 122986
file data blocks allocated: 262144
referenced 262144
On a kernel without the first patch in this series, titled
"[PATCH] Btrfs: fix fsync after succession of renames of different files",
we get instead an error when mounting the filesystem due to failure of
replaying the log:
$ mount /dev/sdb /mnt
mount: mount /dev/sdb on /mnt failed: File exists
Fix this by logging the parent directory of an inode whenever we find an
inode that no longer exists (was unlinked in the current transaction),
during the procedure which finds inodes that have old names that collide
with new names of other inodes.
A test case for fstests follows soon.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
After a succession of rename operations of different files and fsyncing
one of them, such that each file gets a new name that corresponds to an
old name of another file, we can end up with a log that will cause a
failure when attempted to replay at mount time (an EEXIST error).
We currently have correct behaviour when such succession of renames
involves only two files, but if there are more files involved, we end up
not logging all the inodes that are needed, therefore resulting in a
failure when attempting to replay the log.
Example:
$ mkfs.btrfs -f /dev/sdb
$ mount /dev/sdb /mnt
$ mkdir /mnt/testdir
$ touch /mnt/testdir/fname1
$ touch /mnt/testdir/fname2
$ sync
$ mv /mnt/testdir/fname1 /mnt/testdir/fname3
$ mv /mnt/testdir/fname2 /mnt/testdir/fname4
$ ln /mnt/testdir/fname3 /mnt/testdir/fname2
$ touch /mnt/testdir/fname1
$ xfs_io -c "fsync" /mnt/testdir/fname1
<power failure>
$ mount /dev/sdb /mnt
mount: mount /dev/sdb on /mnt failed: File exists
So fix this by checking all inode dependencies when logging an inode. That
is, if one logged inode A has a new name that matches the old name of some
other inode B, check if inode B has a new name that matches the old name
of some other inode C, and so on. This fix is implemented not by doing any
recursive function calls but by using an iterative method using a linked
list that is used in a first-in-first-out fashion.
A test case for fstests follows soon.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Qgroups will do the old roots lookup at delayed ref time, which could be
while walking down the extent root while running a delayed ref. This
should be fine, except we specifically lock eb's in the backref walking
code irrespective of path->skip_locking, which deadlocks the system.
Fix up the backref code to honor path->skip_locking, nobody will be
modifying the commit_root when we're searching so it's completely safe
to do.
This happens since fb235dc06f ("btrfs: qgroup: Move half of the qgroup
accounting time out of commit trans"), kernel may lockup with quota
enabled.
There is one backref trace triggered by snapshot dropping along with
write operation in the source subvolume. The example can be reliably
reproduced:
btrfs-cleaner D 0 4062 2 0x80000000
Call Trace:
schedule+0x32/0x90
btrfs_tree_read_lock+0x93/0x130 [btrfs]
find_parent_nodes+0x29b/0x1170 [btrfs]
btrfs_find_all_roots_safe+0xa8/0x120 [btrfs]
btrfs_find_all_roots+0x57/0x70 [btrfs]
btrfs_qgroup_trace_extent_post+0x37/0x70 [btrfs]
btrfs_qgroup_trace_leaf_items+0x10b/0x140 [btrfs]
btrfs_qgroup_trace_subtree+0xc8/0xe0 [btrfs]
do_walk_down+0x541/0x5e3 [btrfs]
walk_down_tree+0xab/0xe7 [btrfs]
btrfs_drop_snapshot+0x356/0x71a [btrfs]
btrfs_clean_one_deleted_snapshot+0xb8/0xf0 [btrfs]
cleaner_kthread+0x12b/0x160 [btrfs]
kthread+0x112/0x130
ret_from_fork+0x27/0x50
When dropping snapshots with qgroup enabled, we will trigger backref
walk.
However such backref walk at that timing is pretty dangerous, as if one
of the parent nodes get WRITE locked by other thread, we could cause a
dead lock.
For example:
FS 260 FS 261 (Dropped)
node A node B
/ \ / \
node C node D node E
/ \ / \ / \
leaf F|leaf G|leaf H|leaf I|leaf J|leaf K
The lock sequence would be:
Thread A (cleaner) | Thread B (other writer)
-----------------------------------------------------------------------
write_lock(B) |
write_lock(D) |
^^^ called by walk_down_tree() |
| write_lock(A)
| write_lock(D) << Stall
read_lock(H) << for backref walk |
read_lock(D) << lock owner is |
the same thread A |
so read lock is OK |
read_lock(A) << Stall |
So thread A hold write lock D, and needs read lock A to unlock.
While thread B holds write lock A, while needs lock D to unlock.
This will cause a deadlock.
This is not only limited to snapshot dropping case. As the backref
walk, even only happens on commit trees, is breaking the normal top-down
locking order, makes it deadlock prone.
Fixes: fb235dc06f ("btrfs: qgroup: Move half of the qgroup accounting time out of commit trans")
CC: stable@vger.kernel.org # 4.14+
Reported-and-tested-by: David Sterba <dsterba@suse.com>
Reported-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
[ rebase to latest branch and fix lock assert bug in btrfs/007 ]
Signed-off-by: Qu Wenruo <wqu@suse.com>
[ copy logs and deadlock analysis from Qu's patch ]
Signed-off-by: David Sterba <dsterba@suse.com>
[BUG]
Btrfs qgroup will still hit EDQUOT under the following case:
$ dev=/dev/test/test
$ mnt=/mnt/btrfs
$ umount $mnt &> /dev/null
$ umount $dev &> /dev/null
$ mkfs.btrfs -f $dev
$ mount $dev $mnt -o nospace_cache
$ btrfs subv create $mnt/subv
$ btrfs quota enable $mnt
$ btrfs quota rescan -w $mnt
$ btrfs qgroup limit -e 1G $mnt/subv
$ fallocate -l 900M $mnt/subv/padding
$ sync
$ rm $mnt/subv/padding
# Hit EDQUOT
$ xfs_io -f -c "pwrite 0 512M" $mnt/subv/real_file
[CAUSE]
Since commit a514d63882 ("btrfs: qgroup: Commit transaction in advance
to reduce early EDQUOT"), btrfs is not forced to commit transaction to
reclaim more quota space.
Instead, we just check pertrans metadata reservation against some
threshold and try to do asynchronously transaction commit.
However in above case, the pertrans metadata reservation is pretty small
thus it will never trigger asynchronous transaction commit.
[FIX]
Instead of only accounting pertrans metadata reservation, we calculate
how much free space we have, and if there isn't much free space left,
commit transaction asynchronously to try to free some space.
This may slow down the fs when we have less than 32M free qgroup space,
but should reduce a lot of false EDQUOT, so the cost should be
acceptable.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
[BUG]
Btrfs/139 will fail with a high probability if the testing machine (VM)
has only 2G RAM.
Resulting the final write success while it should fail due to EDQUOT,
and the fs will have quota exceeding the limit by 16K.
The simplified reproducer will be: (needs a 2G ram VM)
$ mkfs.btrfs -f $dev
$ mount $dev $mnt
$ btrfs subv create $mnt/subv
$ btrfs quota enable $mnt
$ btrfs quota rescan -w $mnt
$ btrfs qgroup limit -e 1G $mnt/subv
$ for i in $(seq -w 1 8); do
xfs_io -f -c "pwrite 0 128M" $mnt/subv/file_$i > /dev/null
echo "file $i written" > /dev/kmsg
done
$ sync
$ btrfs qgroup show -pcre --raw $mnt
The last pwrite will not trigger EDQUOT and final 'qgroup show' will
show something like:
qgroupid rfer excl max_rfer max_excl parent child
-------- ---- ---- -------- -------- ------ -----
0/5 16384 16384 none none --- ---
0/256 1073758208 1073758208 none 1073741824 --- ---
And 1073758208 is larger than
> 1073741824.
[CAUSE]
It's a bug in btrfs qgroup data reserved space management.
For quota limit, we must ensure that:
reserved (data + metadata) + rfer/excl <= limit
Since rfer/excl is only updated at transaction commmit time, reserved
space needs to be taken special care.
One important part of reserved space is data, and for a new data extent
written to disk, we still need to take the reserved space until
rfer/excl numbers get updated.
Originally when an ordered extent finishes, we migrate the reserved
qgroup data space from extent_io tree to delayed ref head of the data
extent, expecting delayed ref will only be cleaned up at commit
transaction time.
However for small RAM machine, due to memory pressure dirty pages can be
flushed back to disk without committing a transaction.
The related events will be something like:
file 1 written
btrfs_finish_ordered_io: ino=258 ordered offset=0 len=54947840
btrfs_finish_ordered_io: ino=258 ordered offset=54947840 len=5636096
btrfs_finish_ordered_io: ino=258 ordered offset=61153280 len=57344
btrfs_finish_ordered_io: ino=258 ordered offset=61210624 len=8192
btrfs_finish_ordered_io: ino=258 ordered offset=60583936 len=569344
cleanup_ref_head: num_bytes=54947840
cleanup_ref_head: num_bytes=5636096
cleanup_ref_head: num_bytes=569344
cleanup_ref_head: num_bytes=57344
cleanup_ref_head: num_bytes=8192
^^^^^^^^^^^^^^^^ This will free qgroup data reserved space
file 2 written
...
file 8 written
cleanup_ref_head: num_bytes=8192
...
btrfs_commit_transaction <<< the only transaction committed during
the test
When file 2 is written, we have already freed 128M reserved qgroup data
space for ino 258. Thus later write won't trigger EDQUOT.
This allows us to write more data beyond qgroup limit.
In my 2G ram VM, it could reach about 1.2G before hitting EDQUOT.
[FIX]
By moving reserved qgroup data space from btrfs_delayed_ref_head to
btrfs_qgroup_extent_record, we can ensure that reserved qgroup data
space won't be freed half way before commit transaction, thus fix the
problem.
Fixes: f64d5ca868 ("btrfs: delayed_ref: Add new function to record reserved space into delayed ref")
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The member btrfs_fs_info::scrub_nocow_workers is unused since the nocow
optimization was removed from scrub in 9bebe665c3 ("btrfs: scrub:
Remove unused copy_nocow_pages and its callchain").
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The scrub worker pointers are not NULL iff the scrub is running, so
reset them back once the last reference is dropped. Add assertions to
the initial phase of scrub to verify that.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Use the refcount_t for fs_info::scrub_workers_refcnt instead of int so
we get the extra checks. All reference changes are still done under
scrub_lock.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
scrub_workers_refcnt is protected by scrub_lock, add lockdep_assert_held()
in scrub_workers_get().
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Suggested-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have killed volume mutex (commit: dccdb07bc9
btrfs: kill btrfs_fs_info::volume_mutex). This a trival one seems to have
escaped.
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There is no need to forward declare flush_write_bio(), as it only
depends on submit_one_bio(). Both of them are pretty small, just move
them to kill the forward declaration.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The variables and function parameters of __etree_search which pertain to
prev/next are grossly misnamed. Namely, prev_ret holds the next state
and not the previous. Similarly, next_ret actually holds the previous
extent state relating to the offset we are interested in. Fix this by
renaming the variables as well as switching the arguments order. No
functional changes.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
With the refactoring introduced in 8b62f87bad ("Btrfs: reworki
outstanding_extents") this flag became unused. Remove it and renumber
the following flags accordingly. No functional changes.
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>
There is no point in using a construct like 'if (!condition)
WARN_ON(1)'. Use WARN_ON(!condition) directly. No functional changes.
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>
We could generate a lot of delayed refs in evict but never have any left
over space from our block rsv to make up for that fact. So reserve some
extra space and give it to the transaction so it can be used to refill
the delayed refs rsv every loop through the truncate path.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For FLUSH_LIMIT flushers we really can only allocate chunks and flush
delayed inode items, everything else is problematic. I added a bunch of
new states and it lead to weirdness in the FLUSH_LIMIT case because I
forgot about how it worked. So instead explicitly declare the states
that are ok for flushing with FLUSH_LIMIT and use that for our state
machine. Then as we add new things that are safe we can just add them
to this list.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
With severe fragmentation we can end up with our inode rsv size being
huge during writeout, which would cause us to need to make very large
metadata reservations.
However we may not actually need that much once writeout is complete,
because of the over-reservation for the worst case.
So instead try to make our reservation, and if we couldn't make it
re-calculate our new reservation size and try again. If our reservation
size doesn't change between tries then we know we are actually out of
space and can error. Flushing that could have been running in parallel
did not make any space.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
[ rename to calc_refill_bytes, update comment and changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
With the introduction of the per-inode block_rsv it became possible to
have really really large reservation requests made because of data
fragmentation. Since the ticket stuff assumed that we'd always have
relatively small reservation requests it just killed all tickets if we
were unable to satisfy the current request.
However, this is generally not the case anymore. So fix this logic to
instead see if we had a ticket that we were able to give some
reservation to, and if we were continue the flushing loop again.
Likewise we make the tickets use the space_info_add_old_bytes() method
of returning what reservation they did receive in hopes that it could
satisfy reservations down the line.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We've done this forever because of the voodoo around knowing how much
space we have. However, we have better ways of doing this now, and on
normal file systems we'll easily have a global reserve of 512MiB, and
since metadata chunks are usually 1GiB that means we'll allocate
metadata chunks more readily. Instead use the actual used amount when
determining if we need to allocate a chunk or not.
This has a side effect for mixed block group fs'es where we are no
longer allocating enough chunks for the data/metadata requirements. To
deal with this add a ALLOC_CHUNK_FORCE step to the flushing state
machine. This will only get used if we've already made a full loop
through the flushing machinery and tried committing the transaction.
If we have then we can try and force a chunk allocation since we likely
need it to make progress. This resolves issues I was seeing with
the mixed bg tests in xfstests without the new flushing state.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
[ merged with patch "add ALLOC_CHUNK_FORCE to the flushing code" ]
Signed-off-by: David Sterba <dsterba@suse.com>
For enospc_debug having the block rsvs is super helpful to see if we've
done something wrong.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
may_commit_transaction will skip committing the transaction if we don't
have enough pinned space or if we're trying to find space for a SYSTEM
chunk. However, if we have pending free block groups in this transaction
we still want to commit as we may be able to allocate a chunk to make
our reservation. So instead of just returning ENOSPC, check if we have
free block groups pending, and if so commit the transaction to allow us
to use that free space.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Zstd compression requires different amounts of memory for each level of
compression. The prior patches implemented indirection to allow for each
compression type to manage their workspaces independently. This patch
uses this indirection to implement compression level support for zstd.
To manage the additional memory require, each compression level has its
own queue of workspaces. A global LRU is used to help with reclaim.
Reclaim is done via a timer which provides a mechanism to decrease
memory utilization by keeping only workspaces around that are sized
appropriately. Forward progress is guaranteed by a preallocated max
workspace hidden from the LRU.
When getting a workspace, it uses a bitmap to identify the levels that
are populated and scans up. If it finds a workspace that is greater than
it, it uses it, but does not update the last_used time and the
corresponding place in the LRU. If we hit memory pressure, we sleep on
the max level workspace. We continue to rescan in case we can use a
smaller workspace, but eventually should be able to obtain the max level
workspace or allocate one again should memory pressure subside.
The memory requirement for decompression is the same as level 1, and
therefore can use any of available workspace.
The number of workspaces is bound by an upper limit of the workqueue's
limit which currently is 2 (percpu limit). The reclaim timer is used to
free inactive/improperly sized workspaces and is set to 307s to avoid
colliding with transaction commit (every 30s).
Repeating the experiment from v2 [1], the Silesia corpus was copied to a
btrfs filesystem 10 times and then read back after dropping the caches.
The btrfs filesystem was on an SSD.
Level Ratio Compression (MB/s) Decompression (MB/s) Memory (KB)
1 2.658 438.47 910.51 780
2 2.744 364.86 886.55 1004
3 2.801 336.33 828.41 1260
4 2.858 286.71 886.55 1260
5 2.916 212.77 556.84 1388
6 2.363 119.82 990.85 1516
7 3.000 154.06 849.30 1516
8 3.011 159.54 875.03 1772
9 3.025 100.51 940.15 1772
10 3.033 118.97 616.26 1772
11 3.036 94.19 802.11 1772
12 3.037 73.45 931.49 1772
13 3.041 55.17 835.26 2284
14 3.087 44.70 716.78 2547
15 3.126 37.30 878.84 2547
[1] https://lore.kernel.org/linux-btrfs/20181031181108.289340-1-terrelln@fb.com/
Cc: Nick Terrell <terrelln@fb.com>
Cc: Omar Sandoval <osandov@osandov.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
It is possible based on the level configurations that a higher level
workspace uses less memory than a lower level workspace. In order to
reuse workspaces, this must be made a monotonic relationship. This
precomputes the required memory for each level and enforces the
monotonicity between level and memory required. This is also done
in upstream zstd in [1].
[1] a68b76afef
Cc: Nick Terrell <terrelln@fb.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Zstd currently only supports the default level of compression. This
patch switches to using the level passed in for btrfs zstd
configuration.
Zstd workspaces now keep track of the requested level as this can differ
from the size of the workspace.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently, the only user of set_level() is zlib which sets an internal
workspace parameter. As level is now plumbed into get_workspace(), this
can be handled there rather than separately.
This repurposes set_level() to bound the level passed in so it can be
used when setting the mounts compression level and as well as verifying
the level before getting a workspace. The other benefit is this divides
the meaning of compress(0) and get_workspace(0). The former means we
want to use the default compression level of the compression type. The
latter means we can use any workspace available.
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Zlib compression supports multiple levels, but doesn't require changing
in how a workspace itself is created and managed. Zstd introduces a
different memory requirement such that higher levels of compression
require more memory.
This requires changes in how the alloc()/get() methods work for zstd.
This pach plumbs compression level through the interface as a parameter
in preparation for zstd compression levels. This gives the compression
types opportunity to create/manage based on the compression level.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The previous patch added generic helpers for get_workspace() and
put_workspace(). Now, we can migrate ownership of the workspace_manager
to be in the compression type code as the compression code itself
doesn't care beyond being able to get a workspace. The init/cleanup and
get/put methods are abstracted so each compression algorithm can decide
how they want to manage their workspaces.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There are two levels of workspace management. First, alloc()/free()
which are responsible for actually creating and destroy workspaces.
Second, at a higher level, get()/put() which is the compression code
asking for a workspace from a workspace_manager.
The compression code shouldn't really care how it gets a workspace, but
that it got a workspace. This adds get_workspace() and put_workspace()
to be the higher level interface which is responsible for indexing into
the appropriate compression type. It also introduces
btrfs_put_workspace() and btrfs_get_workspace() to be the generic
implementations of the higher interface.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Workspace manager init and cleanup code is open coded inside a for loop
over the compression types. This forces each compression type to rely on
the same workspace manager implementation. This patch creates helper
methods that will be the generic implementation for btrfs workspace
management.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Make the workspace_manager own the interface operations rather than
managing index-paired arrays for the workspace_manager and compression
operations.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
While the heuristic workspaces aren't really compression workspaces,
they use the same interface for managing them. So rather than branching,
let's just handle them once again as the index 0 compression type.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is in preparation for zstd compression levels. As each level will
require different size of workspace, workspaces_list is no longer a
really fitting name.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
It is very easy to miss places that rely on a certain bitshifting for
decoding the type_level overloading. Add helpers to do this instead.
Cc: Omar Sandoval <osandov@osandov.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Support for a new command that can be used eg. as a command
$ btrfs device scan --forget [dev]'
(the final name may change though)
to undo the effects of 'btrfs device scan [dev]'. For this purpose
this patch proposes to use ioctl #5 as it was empty and is next to the
SCAN ioctl.
The new ioctl BTRFS_IOC_FORGET_DEV works only on the control device
(/dev/btrfs-control) to unregister one or all devices, devices that are
not mounted.
The argument is struct btrfs_ioctl_vol_args, ::name specifies the device
path. To unregister all device, the path is an empty string.
Again, the devices are removed only if they aren't part of a mounte
filesystem.
This new ioctl provides:
- release of unwanted btrfs_fs_devices and btrfs_devices structures
from memory if the device is not going to be mounted
- ability to mount filesystem in degraded mode, when one devices is
corrupted like in split brain raid1
- running test cases which would require reloading the kernel module
but this is not possible eg. due to mounted filesystem or built-in
Signed-off-by: Anand Jain <anand.jain@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
The throttle path doesn't take cleaner_delayed_iput_mutex, which means
we could think we're done flushing iputs in the data space reservation
path when we could have a throttler doing an iput. There's no real
reason to serialize the delayed iput flushing, so instead of taking the
cleaner_delayed_iput_mutex whenever we flush the delayed iputs just
replace it with an atomic counter and a waitqueue. This removes the
short (or long depending on how big the inode is) window where we think
there are no more pending iputs when there really are some.
The waiting is killable as it could be indirectly called from user
operations like fallocate or zero-range. Such call sites should handle
the error but otherwise it's not necessary. Eg. flush_space just needs
to attempt to make space by waiting on iputs.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
[ add killable comment and changelog parts ]
Signed-off-by: David Sterba <dsterba@suse.com>
Since inc_block_group_ro() would return -ENOSPC, outputting debug info
for enospc_debug mount option would be helpful to debug some balance
false ENOSPC report.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Inside qgroup_rsv_add/release(), we have trace events
trace_qgroup_update_reserve() to catch reserved space update.
However we still have two manual trace_qgroup_update_reserve() calls
just outside these functions. Remove these duplicated calls.
Fixes: 64ee4e751a ("btrfs: qgroup: Update trace events to use new separate rsv types")
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
A compiler warning (in a patch in development) pointed to a variable
that was used only inside and ASSERT:
u64 root_objectid = root->root_key.objectid;
ASSERT(root_objectid == ...);
fs/btrfs/relocation.c: In function ‘insert_dirty_subv’:
fs/btrfs/relocation.c:2138:6: warning: unused variable ‘root_objectid’ [-Wunused-variable]
u64 root_objectid = root->root_key.objectid;
^~~~~~~~~~~~~
When CONFIG_BRTFS_ASSERT isn't enabled, variable root_objectid isn't used.
Rework the assertion helper by adding a runtime check instead of the
'#ifdef CONFIG_BTRFS_ASSERT #else ...", so the compiler sees the
condition being passed into an inline function after preprocessing.
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
Reviewed-by: David Sterba <dsterba@suse.com>
[ update changelog ]
Signed-off-by: David Sterba <dsterba@suse.com>
The last caller that does not have a fixed value of lock is
btrfs_set_path_blocking, that actually does the same conditional swtich
by the lock type so we can merge the branches together and remove the
helper.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
Currently, the number of readers and writers is checked and in case
there are any, wait and redo the locks. There's some duplication
before the branches go back to again label, eg. calling wait_event on
blocking_readers twice.
The sequence is transformed
loop:
* wait for readers
* wait for writers
* write_lock
* check readers, unlock and wait for readers, loop
* check writers, unlock and wait for writers, loop
The new sequence is not exactly the same due to the simplification, for
readers it's slightly faster. For the writers, original code does
* wait for writers
* (loop) wait for readers
* wait for writers -- again
while the new goes directly to the reader check. This should behave the
same on a contended lock with multiple writers and readers, but can
reduce number of times we're waiting on something.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_set_lock_blocking is now only a simple wrapper around
btrfs_set_lock_blocking_write. The name does not bring any semantic
value that could not be inferred from the new function so there's no
point keeping it.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
We can use the right helper where the lock type is a fixed parameter.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
There are many callers that hardcode the desired lock type so we can
avoid the switch and call them directly. Split the current function to
two. There are no remaining users of btrfs_clear_lock_blocking_rw so
it's removed. The call sites will be converted in followup patches.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
There are many callers that hardcode the desired lock type so we can
avoid the switch and call them directly. Split the current function to
two but leave a helper that still takes the variable lock type to make
current code compile. The call sites will be converted in followup
patches.
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: David Sterba <dsterba@suse.com>
Since it's replaced by new delayed subtree swap code, remove the
original code.
The cleanup is small since most of its core function is still used by
delayed subtree swap trace.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Before this patch, qgroup code traces the whole subtree of subvolume and
reloc trees unconditionally.
This makes qgroup numbers consistent, but it could cause tons of
unnecessary extent tracing, which causes a lot of overhead.
However for subtree swap of balance, just swap both subtrees because
they contain the same contents and tree structure, so qgroup numbers
won't change.
It's the race window between subtree swap and transaction commit could
cause qgroup number change.
This patch will delay the qgroup subtree scan until COW happens for the
subtree root.
So if there is no other operations for the fs, balance won't cause extra
qgroup overhead. (best case scenario)
Depending on the workload, most of the subtree scan can still be
avoided.
Only for worst case scenario, it will fall back to old subtree swap
overhead. (scan all swapped subtrees)
[[Benchmark]]
Hardware:
VM 4G vRAM, 8 vCPUs,
disk is using 'unsafe' cache mode,
backing device is SAMSUNG 850 evo SSD.
Host has 16G ram.
Mkfs parameter:
--nodesize 4K (To bump up tree size)
Initial subvolume contents:
4G data copied from /usr and /lib.
(With enough regular small files)
Snapshots:
16 snapshots of the original subvolume.
each snapshot has 3 random files modified.
balance parameter:
-m
So the content should be pretty similar to a real world root fs layout.
And after file system population, there is no other activity, so it
should be the best case scenario.
| v4.20-rc1 | w/ patchset | diff
-----------------------------------------------------------------------
relocated extents | 22615 | 22457 | -0.1%
qgroup dirty extents | 163457 | 121606 | -25.6%
time (sys) | 22.884s | 18.842s | -17.6%
time (real) | 27.724s | 22.884s | -17.5%
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
To allow delayed subtree swap rescan, btrfs needs to record per-root
information about which tree blocks get swapped. This patch introduces
the required infrastructure.
The designed workflow will be:
1) Record the subtree root block that gets swapped.
During subtree swap:
O = Old tree blocks
N = New tree blocks
reloc tree subvolume tree X
Root Root
/ \ / \
NA OB OA OB
/ | | \ / | | \
NC ND OE OF OC OD OE OF
In this case, NA and OA are going to be swapped, record (NA, OA) into
subvolume tree X.
2) After subtree swap.
reloc tree subvolume tree X
Root Root
/ \ / \
OA OB NA OB
/ | | \ / | | \
OC OD OE OF NC ND OE OF
3a) COW happens for OB
If we are going to COW tree block OB, we check OB's bytenr against
tree X's swapped_blocks structure.
If it doesn't fit any, nothing will happen.
3b) COW happens for NA
Check NA's bytenr against tree X's swapped_blocks, and get a hit.
Then we do subtree scan on both subtrees OA and NA.
Resulting 6 tree blocks to be scanned (OA, OC, OD, NA, NC, ND).
Then no matter what we do to subvolume tree X, qgroup numbers will
still be correct.
Then NA's record gets removed from X's swapped_blocks.
4) Transaction commit
Any record in X's swapped_blocks gets removed, since there is no
modification to swapped subtrees, no need to trigger heavy qgroup
subtree rescan for them.
This will introduce 128 bytes overhead for each btrfs_root even qgroup
is not enabled. This is to reduce memory allocations and potential
failures.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Refactor btrfs_qgroup_trace_subtree_swap() into
qgroup_trace_subtree_swap(), which only needs two extent buffer and some
other bool to control the behavior.
This provides the basis for later delayed subtree scan work.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Relocation code will drop btrfs_root::reloc_root as soon as
merge_reloc_root() finishes.
However later qgroup code will need to access btrfs_root::reloc_root
after merge_reloc_root() for delayed subtree rescan.
So alter the timming of resetting btrfs_root:::reloc_root, make it
happens after transaction commit.
With this patch, we will introduce a new btrfs_root::state,
BTRFS_ROOT_DEAD_RELOC_TREE, to info part of btrfs_root::reloc_tree user
that although btrfs_root::reloc_tree is still non-NULL, but still it's
not used any more.
The lifespan of btrfs_root::reloc tree will become:
Old behavior | New
------------------------------------------------------------------------
btrfs_init_reloc_root() --- | btrfs_init_reloc_root() ---
set reloc_root | | set reloc_root |
| | |
| | |
merge_reloc_root() | | merge_reloc_root() |
|- btrfs_update_reloc_root() --- | |- btrfs_update_reloc_root() -+-
clear btrfs_root::reloc_root | set ROOT_DEAD_RELOC_TREE |
| record root into dirty |
| roots rbtree |
| |
| reloc_block_group() Or |
| btrfs_recover_relocation() |
| | After transaction commit |
| |- clean_dirty_subvols() ---
| clear btrfs_root::reloc_root
During ROOT_DEAD_RELOC_TREE set lifespan, the only user of
btrfs_root::reloc_tree should be qgroup.
Since reloc root needs a longer life-span, this patch will also delay
btrfs_drop_snapshot() call.
Now btrfs_drop_snapshot() is called in clean_dirty_subvols().
This patch will increase the size of btrfs_root by 16 bytes.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Filipe Manana