We don't need it, rsv->size is set once and never changes throughout
its lifetime, so just use that for the reserve size.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
I ran into an issue where there was some reference being held on an
inode that I couldn't track. This assert wasn't triggered, but it at
least rules out we're doing something stupid.
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>
Allocating new chunks modifies both the extent and chunk tree, which can
trigger new chunk allocations. So instead of doing list_for_each_safe,
just do while (!list_empty()) so we make sure we don't exit with other
pending bg's still on our list.
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Liu Bo <bo.liu@linux.alibaba.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We need to clear the max_extent_size when we clear bits from a bitmap
since it could have been from the range that contains the
max_extent_size.
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Liu Bo <bo.liu@linux.alibaba.com>
Signed-off-by: Josef Bacik <jbacik@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
If we're allocating a new space cache inode it's likely going to be
under a transaction handle, so we need to use memalloc_nofs_save() in
order to avoid deadlocks, and more importantly lockdep messages that
make xfstests fail.
CC: stable@vger.kernel.org # 4.4+
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>
We want to release the unused reservation we have since it refills the
delayed refs reserve, which will make everything go smoother when
running the delayed refs if we're short on our reservation.
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Btrfs's btree locking has two modes, spinning mode and blocking mode,
while searching btree, locking is always acquired in spinning mode and
then converted to blocking mode if necessary, and in some hot paths we may
switch the locking back to spinning mode by btrfs_clear_path_blocking().
When acquiring locks, both of reader and writer need to wait for blocking
readers and writers to complete before doing read_lock()/write_lock().
The problem is that btrfs_clear_path_blocking() needs to switch nodes
in the path to blocking mode at first (by btrfs_set_path_blocking) to
make lockdep happy before doing its actual clearing blocking job.
When switching to blocking mode from spinning mode, it consists of
step 1) bumping up blocking readers counter and
step 2) read_unlock()/write_unlock(),
this has caused serious ping-pong effect if there're a great amount of
concurrent readers/writers, as waiters will be woken up and go to
sleep immediately.
1) Killing this kind of ping-pong results in a big improvement in my 1600k
files creation script,
MNT=/mnt/btrfs
mkfs.btrfs -f /dev/sdf
mount /dev/def $MNT
time fsmark -D 10000 -S0 -n 100000 -s 0 -L 1 -l /tmp/fs_log.txt \
-d $MNT/0 -d $MNT/1 \
-d $MNT/2 -d $MNT/3 \
-d $MNT/4 -d $MNT/5 \
-d $MNT/6 -d $MNT/7 \
-d $MNT/8 -d $MNT/9 \
-d $MNT/10 -d $MNT/11 \
-d $MNT/12 -d $MNT/13 \
-d $MNT/14 -d $MNT/15
w/o patch:
real 2m27.307s
user 0m12.839s
sys 13m42.831s
w/ patch:
real 1m2.273s
user 0m15.802s
sys 8m16.495s
1.1) latency histogram from funclatency[1]
Overall with the patch, there're ~50% less write lock acquisition and
the 95% max latency that write lock takes also reduces to ~100ms from
>500ms.
--------------------------------------------
w/o patch:
--------------------------------------------
Function = btrfs_tree_lock
msecs : count distribution
0 -> 1 : 2385222 |****************************************|
2 -> 3 : 37147 | |
4 -> 7 : 20452 | |
8 -> 15 : 13131 | |
16 -> 31 : 3877 | |
32 -> 63 : 3900 | |
64 -> 127 : 2612 | |
128 -> 255 : 974 | |
256 -> 511 : 165 | |
512 -> 1023 : 13 | |
Function = btrfs_tree_read_lock
msecs : count distribution
0 -> 1 : 6743860 |****************************************|
2 -> 3 : 2146 | |
4 -> 7 : 190 | |
8 -> 15 : 38 | |
16 -> 31 : 4 | |
--------------------------------------------
w/ patch:
--------------------------------------------
Function = btrfs_tree_lock
msecs : count distribution
0 -> 1 : 1318454 |****************************************|
2 -> 3 : 6800 | |
4 -> 7 : 3664 | |
8 -> 15 : 2145 | |
16 -> 31 : 809 | |
32 -> 63 : 219 | |
64 -> 127 : 10 | |
Function = btrfs_tree_read_lock
msecs : count distribution
0 -> 1 : 6854317 |****************************************|
2 -> 3 : 2383 | |
4 -> 7 : 601 | |
8 -> 15 : 92 | |
2) dbench also proves the improvement,
dbench -t 120 -D /mnt/btrfs 16
w/o patch:
Throughput 158.363 MB/sec
w/ patch:
Throughput 449.52 MB/sec
3) xfstests didn't show any additional failures.
One thing to note is that callers may set path->leave_spinning to have
all nodes in the path stay in spinning mode, which means callers are
ready to not sleep before releasing the path, but it won't cause
problems if they don't want to sleep in blocking mode.
[1]: https://github.com/iovisor/bcc/blob/master/tools/funclatency.py
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The value of blocking_readers is increased only when the lock is taken
for read, no way we can fail the condition with the write lock.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The replace_wait and bio_counter were mistakenly added to fs_info in
commit c404e0dc2c ("Btrfs: fix use-after-free in the finishing
procedure of the device replace"), but they logically belong to
fs_info::dev_replace. Besides, bio_counter is a very generic name and is
confusing in bare fs_info context.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The exit sequence in btrfs_dev_replace_start does not allow to simply
add a label to the right place so the error handling after starting
transaction failure jumps there. Currently there's a lock that pairs
with the unlock in the section, which is unnecessary and only raises
questions. Add a variable to track the locking status and avoid the
extra locking.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Too trivial, the purpose can be simply documented in a comment.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The wrapper is too trivial, open coding does not make it less readable.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There's a single caller and the function name does not say it's actually
taking the lock, so open coding makes it more explicit.
For now, btrfs_dev_replace_read_lock is used instead of read_lock so
it's paired with the unlocking wrapper in the same block.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This member seems to be copied from the extent_buffer locking scheme and
is at least used to assert that the read lock/unlock is properly nested.
In some way. While the _inc/_dec are called inside the read lock
section, the asserts are both inside and outside, so the ordering is not
guaranteed and we can see read/inc/dec ordered in any way
(theoretically).
A missing call of btrfs_dev_replace_clear_lock_blocking could cause
unexpected read_locks count, so this at least looks like a valid
assertion, but this will become unnecessary with later updates.
Reviewed-by: Anand Jain <anand.jain@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Although we have tree level check at tree read runtime, it's completely
based on its parent level.
We still need to do accurate level check to avoid invalid tree blocks
sneak into kernel space.
The check itself is simple, for leaf its level should always be 0.
For nodes its level should be in range [1, BTRFS_MAX_LEVEL - 1].
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Su Yue <suy.fnst@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For qgroup_trace_extent_swap(), if we find one leaf that needs to be
traced, we will also iterate all file extents and trace them.
This is OK if we're relocating data block groups, but if we're
relocating metadata block groups, balance code itself has ensured that
both subtree of file tree and reloc tree contain the same contents.
That's to say, if we're relocating metadata block groups, all file
extents in reloc and file tree should match, thus no need to trace them.
This should reduce the total number of dirty extents processed in metadata
block group balance.
[[Benchmark]] (with all previous enhancement)
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.
| v4.19-rc1 | w/ patchset | diff (*)
---------------------------------------------------------------
relocated extents | 22929 | 22851 | -0.3%
qgroup dirty extents | 227757 | 140886 | -38.1%
time (sys) | 65.253s | 37.464s | -42.6%
time (real) | 74.032s | 44.722s | -39.6%
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Reloc tree doesn't contribute to qgroup numbers, as we have accounted
them at balance time (see replace_path()).
Skipping the unneeded subtree tracing should reduce the overhead.
[[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.
| v4.19-rc1 | w/ patchset | diff (*)
---------------------------------------------------------------
relocated extents | 22929 | 22900 | -0.1%
qgroup dirty extents | 227757 | 167139 | -26.6%
time (sys) | 65.253s | 50.123s | -23.2%
time (real) | 74.032s | 52.551s | -29.0%
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Before this patch, with quota enabled during balance, we need to mark
the whole subtree dirty for quota.
E.g.
OO = Old tree blocks (from file tree)
NN = New tree blocks (from reloc tree)
File tree (src) Reloc tree (dst)
OO (a) NN (a)
/ \ / \
(b) OO OO (c) (b) NN NN (c)
/ \ / \ / \ / \
OO OO OO OO (d) OO OO OO NN (d)
For old balance + quota case, quota will mark the whole src and dst tree
dirty, including all the 3 old tree blocks in reloc tree.
It's doable for small file tree or new tree blocks are all located at
lower level.
But for large file tree or new tree blocks are all located at higher
level, this will lead to mark the whole tree dirty, and be unbelievably
slow.
This patch will change how we handle such balance with quota enabled
case.
Now we will search from (b) and (c) for any new tree blocks whose
generation is equal to @last_snapshot, and only mark them dirty.
In above case, we only need to trace tree blocks NN(b), NN(c) and NN(d).
(NN(a) will be traced when COW happens for nodeptr modification). And
also for tree blocks OO(b), OO(c), OO(d). (OO(a) will be traced when COW
happens for nodeptr modification.)
For above case, we could skip 3 tree blocks, but for larger tree, we can
skip tons of unmodified tree blocks, and hugely speed up balance.
This patch will introduce a new function,
btrfs_qgroup_trace_subtree_swap(), which will do the following main
work:
1) Read out real root eb
And setup basic dst_path for later calls
2) Call qgroup_trace_new_subtree_blocks()
To trace all new tree blocks in reloc tree and their counter
parts in the file tree.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Introduce new function, qgroup_trace_new_subtree_blocks(), to iterate
all new tree blocks in a reloc tree.
So that qgroup could skip unrelated tree blocks during balance, which
should hugely speedup balance speed when quota is enabled.
The function qgroup_trace_new_subtree_blocks() itself only cares about
new tree blocks in reloc tree.
All its main works are:
1) Read out tree blocks according to parent pointers
2) Do recursive depth-first search
Will call the same function on all its children tree blocks, with
search level set to current level -1.
And will also skip all children whose generation is smaller than
@last_snapshot.
3) Call qgroup_trace_extent_swap() to trace tree blocks
So although we have parameter list related to source file tree, it's not
used at all, but only passed to qgroup_trace_extent_swap().
Thus despite the tree read code, the core should be pretty short and all
about recursive depth-first search.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Introduce a new function, qgroup_trace_extent_swap(), which will be used
later for balance qgroup speedup.
The basis idea of balance is swapping tree blocks between reloc tree and
the real file tree.
The swap will happen in highest tree block, but there may be a lot of
tree blocks involved.
For example:
OO = Old tree blocks
NN = New tree blocks allocated during balance
File tree (257) Reloc tree for 257
L2 OO NN
/ \ / \
L1 OO OO (a) OO NN (a)
/ \ / \ / \ / \
L0 OO OO OO OO OO OO NN NN
(b) (c) (b) (c)
When calling qgroup_trace_extent_swap(), we will pass:
@src_eb = OO(a)
@dst_path = [ nodes[1] = NN(a), nodes[0] = NN(c) ]
@dst_level = 0
@root_level = 1
In that case, qgroup_trace_extent_swap() will search from OO(a) to
reach OO(c), then mark both OO(c) and NN(c) as qgroup dirty.
The main work of qgroup_trace_extent_swap() can be split into 3 parts:
1) Tree search from @src_eb
It should acts as a simplified btrfs_search_slot().
The key for search can be extracted from @dst_path->nodes[dst_level]
(first key).
2) Mark the final tree blocks in @src_path and @dst_path qgroup dirty
NOTE: In above case, OO(a) and NN(a) won't be marked qgroup dirty.
They should be marked during preivous (@dst_level = 1) iteration.
3) Mark file extents in leaves dirty
We don't have good way to pick out new file extents only.
So we still follow the old method by scanning all file extents in
the leave.
This function can free us from keeping two pathes, thus later we only need
to care about how to iterate all new tree blocks in reloc tree.
Signed-off-by: Qu Wenruo <wqu@suse.com>
[ copy changelog to function comment ]
Signed-off-by: David Sterba <dsterba@suse.com>
Number of qgroup dirty extents is directly linked to the performance
overhead, so add a new trace event, trace_qgroup_num_dirty_extents(), to
record how many dirty extents is processed in
btrfs_qgroup_account_extents().
This will be pretty handy to analyze later balance performance
improvement.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/relocation.c:build_backref_tree() is some code from 2009 era,
although it works pretty fine, it's not that easy to understand.
Especially combined with the complex btrfs backref format.
This patch adds some basic comment for the backref build part of the
code, making it less hard to read, at least for backref searching part.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The only aops we define for symlinks are identical to the aops for
regular files. This has been the case since symlink support was added in
commit 2b8d99a723 ("Btrfs: symlinks and hard links"). As far as I can
tell, there wasn't a good reason to have separate aops then, and there
isn't now, so let's just do what most other filesystems do and reuse the
same structure.
Signed-off-by: Omar Sandoval <osandov@fb.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
During buffered writes, we follow this basic series of steps:
again:
lock all the pages
wait for writeback on all the pages
Take the extent range lock
wait for ordered extents on the whole range
clean all the pages
if (copy_from_user_in_atomic() hits a fault) {
drop our locks
goto again;
}
dirty all the pages
release all the locks
The extra waiting, cleaning and locking are there to make sure we don't
modify pages in flight to the drive, after they've been crc'd.
If some of the pages in the range were already dirty when the write
began, and we need to goto again, we create a window where a dirty page
has been cleaned and unlocked. It may be reclaimed before we're able to
lock it again, which means we'll read the old contents off the drive and
lose any modifications that had been pending writeback.
We don't actually need to clean the pages. All of the other locking in
place makes sure we don't start IO on the pages, so we can just leave
them dirty for the duration of the write.
Fixes: 73d59314e6 (the original btrfs merge)
CC: stable@vger.kernel.org # v4.4+
Signed-off-by: Chris Mason <clm@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The helper does the same math and we take care about the special case
when flags is 0 too.
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Refactor the delayed refs loop by using the newly introduced
btrfs_run_delayed_refs_for_head function. This greatly simplifies
__btrfs_run_delayed_refs and makes it more obvious what is happening.
We now have 1 loop which iterates the existing delayed_heads and then
each selected ref head is processed by the new helper. All existing
semantics of the code are preserved so 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>
This patch introduces a new helper encompassing the implicit inner loop
in __btrfs_run_delayed_refs which processes all the refs for a given
head. The code is mostly copy/paste, the only difference is that if we
detect a newer reference then -EAGAIN is returned so that callers can
react correctly.
Also, at the end of the loop the head is relocked and
btrfs_merge_delayed_refs is run again to retain the pre-refactoring
semantics.
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is in preparation to refactor the giant loop in
__btrfs_run_delayed_refs. As a first step define a new function
which implements acquiring a reference to a btrfs_delayed_refs_head and
use it. 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>
Avoid the inline ifdefs and use two sections for self-tests enabled and
disabled.
Though there could be no ifdef and unconditional test_bit of
BTRFS_FS_STATE_DUMMY_FS_INFO, the static inline can help to optimize out
any code that would depend on conditions using btrfs_is_testing.
As this is only for the testing code, drop unlikely().
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The data used only for tests are better placed at the end of the
structure so that they don't change the structure layout. All new
members of btrfs_root should be placed before.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The helper find_lock_delalloc_range is now conditionally built static,
dpending on whether the self-tests are enabled or not. There's a macro
that is supposed to hide the export, used only once. To discourage
further use, drop it an add a public wrapper for the helper needed by
tests.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
rb_first_cached() trades an extra pointer "leftmost" for doing the same
job as rb_first() but in O(1).
While resolving indirect refs and missing refs, it always looks for the
first rb entry in a while loop, it's helpful to use rb_first_cached
instead.
For more details about the optimization see patch "Btrfs: delayed-refs:
use rb_first_cached for href_root".
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
rb_first_cached() trades an extra pointer "leftmost" for doing the
same job as rb_first() but in O(1).
As evict_inode_truncate_pages() removes all extent mapping by always
looking for the first rb entry, it's helpful to use rb_first_cached
instead.
For more details about the optimization see patch "Btrfs: delayed-refs:
use rb_first_cached for href_root".
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
rb_first_cached() trades an extra pointer "leftmost" for doing the same job as
rb_first() but in O(1).
Functions manipulating delayed_item need to get the first entry, this converts
it to use rb_first_cached().
For more details about the optimization see patch "Btrfs: delayed-refs:
use rb_first_cached for href_root".
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
rb_first_cached() trades an extra pointer "leftmost" for doing the same
job as rb_first() but in O(1).
Functions manipulating href->ref_tree need to get the first entry, this
converts href->ref_tree to use rb_first_cached().
For more details about the optimization see patch "Btrfs: delayed-refs:
use rb_first_cached for href_root".
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
rb_first_cached() trades an extra pointer "leftmost" for doing the same
job as rb_first() but in O(1).
Functions manipulating href_root need to get the first entry, this
converts href_root to use rb_first_cached().
This patch is first in the sequenct of similar updates to other rbtrees
and this is analysis of the expected behaviour and improvements.
There's a common pattern:
while (node = rb_first) {
entry = rb_entry(node)
next = rb_next(node)
rb_erase(node)
cleanup(entry)
}
rb_first needs to traverse the tree up to logN depth, rb_erase can
completely reshuffle the tree. With the caching we'll skip the traversal
in rb_first. That's a cached memory access vs looped pointer
dereference trade-off that IMHO has a clear winner.
Measurements show there's not much difference in a sample tree with
10000 nodes: 4.5s / rb_first and 4.8s / rb_first_cached. Real effects of
caching and pointer chasing are unpredictable though.
Further optimzations can be done to avoid the expensive rb_erase step.
In some cases it's ok to process the nodes in any order, so the tree can
be traversed in post-order, not rebalancing the children nodes and just
calling free. Care must be taken regarding the next node.
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ update changelog from mail discussions ]
Signed-off-by: David Sterba <dsterba@suse.com>
While testing my backport I noticed there was a panic if I ran
generic/416 generic/417 generic/418 all in a row. This just happened to
uncover a race where we had outstanding IO after we destroy all of our
workqueues, and then we'd go to queue the endio work on those free'd
workqueues.
This is because we aren't waiting for the caching threads to be done
before freeing everything up, so to fix this make sure we wait on any
outstanding caching that's being done before we free up the block group,
so we're sure to be done with all IO by the time we get to
btrfs_stop_all_workers(). This fixes the panic I was seeing
consistently in testing.
------------[ cut here ]------------
kernel BUG at fs/btrfs/volumes.c:6112!
SMP PTI
Modules linked in:
CPU: 1 PID: 27165 Comm: kworker/u4:7 Not tainted 4.16.0-02155-g3553e54a578d-dirty #875
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-2.el7 04/01/2014
Workqueue: btrfs-cache btrfs_cache_helper
RIP: 0010:btrfs_map_bio+0x346/0x370
RSP: 0000:ffffc900061e79d0 EFLAGS: 00010202
RAX: 0000000000000000 RBX: ffff880071542e00 RCX: 0000000000533000
RDX: ffff88006bb74380 RSI: 0000000000000008 RDI: ffff880078160000
RBP: 0000000000000001 R08: ffff8800781cd200 R09: 0000000000503000
R10: ffff88006cd21200 R11: 0000000000000000 R12: 0000000000000000
R13: 0000000000000000 R14: ffff8800781cd200 R15: ffff880071542e00
FS: 0000000000000000(0000) GS:ffff88007fd00000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000000000817ffc4 CR3: 0000000078314000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
btree_submit_bio_hook+0x8a/0xd0
submit_one_bio+0x5d/0x80
read_extent_buffer_pages+0x18a/0x320
btree_read_extent_buffer_pages+0xbc/0x200
? alloc_extent_buffer+0x359/0x3e0
read_tree_block+0x3d/0x60
read_block_for_search.isra.30+0x1a5/0x360
btrfs_search_slot+0x41b/0xa10
btrfs_next_old_leaf+0x212/0x470
caching_thread+0x323/0x490
normal_work_helper+0xc5/0x310
process_one_work+0x141/0x340
worker_thread+0x44/0x3c0
kthread+0xf8/0x130
? process_one_work+0x340/0x340
? kthread_bind+0x10/0x10
ret_from_fork+0x35/0x40
RIP: btrfs_map_bio+0x346/0x370 RSP: ffffc900061e79d0
---[ end trace 827eb13e50846033 ]---
Kernel panic - not syncing: Fatal exception
Kernel Offset: disabled
---[ end Kernel panic - not syncing: Fatal exception
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Commit 499f377f49 (btrfs: iterate over unused chunk space in FITRIM)
fixed free space trimming, but introduced latency when it was running.
This is due to it pinning the transaction using both a incremented
refcount and holding the commit root sem for the duration of a single
trim operation.
This was to ensure safety but it's unnecessary. We already hold the the
chunk mutex so we know that the chunk we're using can't be allocated
while we're trimming it.
In order to check against chunks allocated already in this transaction,
we need to check the pending chunks list. To to that safely without
joining the transaction (or attaching than then having to commit it) we
need to ensure that the dev root's commit root doesn't change underneath
us and the pending chunk lists stays around until we're done with it.
We can ensure the former by holding the commit root sem and the latter
by pinning the transaction. We do this now, but the critical section
covers the trim operation itself and we don't need to do that.
This patch moves the pinning and unpinning logic into helpers and unpins
the transaction after performing the search and check for pending
chunks.
Limiting the critical section of the transaction pinning improves the
latency substantially on slower storage (e.g. image files over NFS).
Fixes: 499f377f49 ("btrfs: iterate over unused chunk space in FITRIM")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We check whether any device the file system is using supports discard in
the ioctl call, but then we attempt to trim free extents on every device
regardless of whether discard is supported. Due to the way we mask off
EOPNOTSUPP, we can end up issuing the trim operations on each free range
on devices that don't support it, just wasting time.
Fixes: 499f377f49 ("btrfs: iterate over unused chunk space in FITRIM")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_trim_fs iterates over the fs_devices->alloc_list while holding the
device_list_mutex. The problem is that ->alloc_list is protected by the
chunk mutex. We don't want to hold the chunk mutex over the trim of the
entire file system. Fortunately, the ->dev_list list is protected by
the dev_list mutex and while it will give us all devices, including
read-only devices, we already just skip the read-only devices. Then we
can continue to take and release the chunk mutex while scanning each
device.
Fixes: 499f377f49 ("btrfs: iterate over unused chunk space in FITRIM")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
[BUG]
fstrim on some btrfs only trims the unallocated space, not trimming any
space in existing block groups.
[CAUSE]
Before fstrim_range passed to btrfs_trim_fs(), it gets truncated to
range [0, super->total_bytes). So later btrfs_trim_fs() will only be
able to trim block groups in range [0, super->total_bytes).
While for btrfs, any bytenr aligned to sectorsize is valid, since btrfs
uses its logical address space, there is nothing limiting the location
where we put block groups.
For filesystem with frequent balance, it's quite easy to relocate all
block groups and bytenr of block groups will start beyond
super->total_bytes.
In that case, btrfs will not trim existing block groups.
[FIX]
Just remove the truncation in btrfs_ioctl_fitrim(), so btrfs_trim_fs()
can get the unmodified range, which is normally set to [0, U64_MAX].
Reported-by: Chris Murphy <lists@colorremedies.com>
Fixes: f4c697e640 ("btrfs: return EINVAL if start > total_bytes in fitrim ioctl")
CC: <stable@vger.kernel.org> # v4.4+
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Function btrfs_trim_fs() doesn't handle errors in a consistent way. If
error happens when trimming existing block groups, it will skip the
remaining blocks and continue to trim unallocated space for each device.
The return value will only reflect the final error from device trimming.
This patch will fix such behavior by:
1) Recording the last error from block group or device trimming
The return value will also reflect the last error during trimming.
Make developer more aware of the problem.
2) Continuing trimming if possible
If we failed to trim one block group or device, we could still try
the next block group or device.
3) Report number of failures during block group and device trimming
It would be less noisy, but still gives user a brief summary of
what's going wrong.
Such behavior can avoid confusion for cases like failure to trim the
first block group and then only unallocated space is trimmed.
Reported-by: Chris Murphy <lists@colorremedies.com>
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ add bg_ret and dev_ret to the messages ]
Signed-off-by: David Sterba <dsterba@suse.com>
When we fail to start a transaction in btrfs_dev_replace_start, we leave
dev_replace->replace_start set to STARTED but clear ->srcdev and
->tgtdev. Later, that can result in an Oops in
btrfs_dev_replace_progress when having state set to STARTED or SUSPENDED
implies that ->srcdev is valid.
Also fix error handling when the state is already STARTED or SUSPENDED
while starting. That, too, will clear ->srcdev and ->tgtdev even though
it doesn't own them. This should be an impossible case to hit since we
should be protected by the BTRFS_FS_EXCL_OP bit being set. Let's add an
ASSERT there while we're at it.
Fixes: e93c89c1aa (Btrfs: add new sources for device replace code)
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
remove_extent_mapping uses the variable "ret" for return value, but it
is not modified after initialzation. Further, I find that any of the
callers do not handle the return value and the callees are only simple
functions so the return values does not need to be passed.
Signed-off-by: zhong jiang <zhongjiang@huawei.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In btrfs_search_old_slot get_old_root is always used with the assumption
it cannot fail. However, this is not true in rare circumstance it can
fail and return null. This will lead to null point dereference when the
header is read. Fix this by checking the return value and properly
handling NULL by setting ret to -EIO and returning gracefully.
Coverity-id: 1087503
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
In btrfs_orphan_cleanup the final 'if (ret) goto out' cannot ever be
executed. This is due to the last assignment to 'ret' depending on the
return value of btrfs_iget. If an error other than -ENOENT is returned
then the loop is prematurely terminated by 'goto out'. On the other
hand, if the error value is ENOENT then a subsequent if branch is
executed that always re-assigns 'ret' and in case it's an error just
terminates the loop. No functional changes.
Coverity-id: 1437392
Signed-off-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: Lu Fengqi <lufq.fnst@cn.fujitsu.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When we delete an inode,
btrfs_evict_inode() {
truncate_inode_pages_final()
truncate_inode_pages_range()
lock_page()
truncate_cleanup_page()
btrfs_invalidatepage()
wait_on_page_writeback
btrfs_lookup_ordered_range()
cancel_dirty_page()
unlock_page()
...
btrfs_wait_ordered_range()
...
As VFS has called ->invalidatepage() to get all ordered extents done (if
there are any) and truncated all page cache pages (no dirty pages to
writeback after this step), wait_ordered_range() is just a noop.
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
As long as @eb is marked with EXTENT_BUFFER_DIRTY, all of its pages
are dirty, so no need to set pages dirty again.
Ftrace showed that the loop took 10us on my dev box, so removing this
can save us at least 10us if eb is already dirty and otherwise avoid a
potentially expensive calls to set_page_dirty.
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Just in case that someone breaks the rule that pages are dirty as long
as eb is dirty. The next patch will dirty the pages conditionally.
Signed-off-by: Liu Bo <bo.liu@linux.alibaba.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
