In the search loop of the binary search function, we are doing a division
by 2 of the sum of the high and low slots. Because the slots are integers,
the generated assembly code for it is the following on x86_64:
0x00000000000141f1 <+145>: mov %eax,%ebx
0x00000000000141f3 <+147>: shr $0x1f,%ebx
0x00000000000141f6 <+150>: add %eax,%ebx
0x00000000000141f8 <+152>: sar %ebx
It's a few more instructions than a simple right shift, because signed
integer division needs to round towards zero. However we know that slots
can never be negative (btrfs_header_nritems() returns an u32), so we
can instead use unsigned types for the low and high slots and therefore
use unsigned integer division, which results in a single instruction on
x86_64:
0x00000000000141f0 <+144>: shr %ebx
So use unsigned types for the slots and therefore unsigned division.
This is part of a small patchset comprised of the following two patches:
btrfs: eliminate extra call when doing binary search on extent buffer
btrfs: do unsigned integer division in the extent buffer binary search loop
The following fs_mark test was run on a non-debug kernel (Debian's default
kernel config) before and after applying the patchset:
$ cat test.sh
#!/bin/bash
DEV=/dev/sdi
MNT=/mnt/sdi
MOUNT_OPTIONS="-o ssd"
MKFS_OPTIONS="-O no-holes -R free-space-tree"
FILES=100000
THREADS=$(nproc --all)
FILE_SIZE=0
umount $DEV &> /dev/null
mkfs.btrfs -f $MKFS_OPTIONS $DEV
mount $MOUNT_OPTIONS $DEV $MNT
OPTS="-S 0 -L 6 -n $FILES -s $FILE_SIZE -t $THREADS -k"
for ((i = 1; i <= $THREADS; i++)); do
OPTS="$OPTS -d $MNT/d$i"
done
fs_mark $OPTS
umount $MNT
Results before applying patchset:
FSUse% Count Size Files/sec App Overhead
2 1200000 0 174472.0 11549868
4 2400000 0 253503.0 11694618
4 3600000 0 257833.1 11611508
6 4800000 0 247089.5 11665983
6 6000000 0 211296.1 12121244
10 7200000 0 187330.6 12548565
Results after applying patchset:
FSUse% Count Size Files/sec App Overhead
2 1200000 0 207556.0 11393252
4 2400000 0 266751.1 11347909
4 3600000 0 274397.5 11270058
6 4800000 0 259608.4 11442250
6 6000000 0 238895.8 11635921
8 7200000 0 211942.2 11873825
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The function btrfs_bin_search() is just a wrapper around the function
generic_bin_search(), which passes the same arguments plus a default
low slot with a value of 0. This adds an unnecessary extra function
call, since btrfs_bin_search() is not static. So improve on this by
making btrfs_bin_search() an inline function that calls
generic_bin_search(), renaming the later to btrfs_generic_bin_search()
and exporting it.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
btrfs_clean_tree_block is a misnomer, it's just
clear_extent_buffer_dirty with some extra accounting around it. Rename
this to btrfs_clear_buffer_dirty to make it more clear it belongs with
it's setter, btrfs_mark_buffer_dirty.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We check the header generation in the extent buffer against the current
running transaction id to see if it's safe to clear DIRTY on this
buffer. Generally speaking if we're clearing the buffer dirty we're
holding the transaction open, but in the case of cleaning up an aborted
transaction we don't, so we have extra checks in that path to check the
transid. To allow for a future cleanup go ahead and pass in the trans
handle so we don't have to rely on ->running_transaction being set.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add annotations to functions that might sleep due to allocations or IO
and could be called from various contexts. In case of btrfs_search_slot
it's not obvious why it would sleep:
btrfs_search_slot
setup_nodes_for_search
reada_for_balance
btrfs_readahead_node_child
btrfs_readahead_tree_block
btrfs_find_create_tree_block
alloc_extent_buffer
kmem_cache_zalloc
/* allocate memory non-atomically, might sleep */
kmem_cache_alloc(GFP_NOFS|__GFP_NOFAIL|__GFP_ZERO)
read_extent_buffer_pages
submit_extent_page
/* disk IO, might sleep */
submit_one_bio
Other examples where the sleeping could happen is in 3 places might
sleep in update_qgroup_limit_item(), as shown below:
update_qgroup_limit_item
btrfs_alloc_path
/* allocate memory non-atomically, might sleep */
kmem_cache_zalloc(btrfs_path_cachep, GFP_NOFS)
Signed-off-by: ChenXiaoSong <chenxiaosong2@huawei.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is simply the same thing as btrfs_item_nr_offset(leaf, 0), so
remove this helper and replace it's usage with the above statement.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have some gnarly memmove and copy_extent_buffer calls for leaf
manipulation. This is because our item offsets aren't absolute, they're
based on 0 being where the items start in the leaf, which is after the
btrfs_header. This means any manipulation of the data requires adding
sizeof(struct btrfs_header) to the offsets we pull from the items.
Moving the items themselves is easier as the helpers are absolute
offsets, however we of course have to call the helpers to get the
offsets for the item numbers. This makes for
copy_extent_buffer/memmove_extent_buffer calls that are kind of hard to
reason about what's happening.
Fix this by pushing this logic into helpers. For data we'll only use
the item provided offsets, and the helpers will use the
BTRFS_LEAF_DATA_OFFSET addition for the offsets. Additionally for the
item manipulation simply pass in the item numbers, and then the helpers
will call the offset helper to get the actual offset into the leaf.
The diffstat makes this look like more code, but that's simply because I
added comments for the helpers, it's net negative for the amount of
code, and is easier to reason.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is a change needed for extent tree v2, as we will be growing the
header size. This exists in btrfs-progs currently, and not having it
makes syncing accessors.[ch] more problematic. So make this change to
set us up for extent tree v2 and match what btrfs-progs does to make
syncing easier.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is actually a change for extent tree v2, but it exists in
btrfs-progs but not in the kernel. This makes it annoying to sync
accessors.h with btrfs-progs, and since this is the way I need it for
extent-tree v2 simply update these helpers to take the extent buffer in
order to make syncing possible now, and make the extent tree v2 stuff
easier moving forward.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
These helpers use functions that are in multiple places, which makes it
tricky to sync them into btrfs-progs. Move them to file-item.h and then
include file-item.h in places that use these helpers.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is only used in ctree.c, with the exception of zero'ing out extent
buffers we're getting ready to write out. In theory we shouldn't have
an extent buffer with 0 items that we're writing out, however I'd rather
be safe than sorry so open code it in extent_io.c, and then copy the
helper into ctree.c. This will make it easier to sync accessors.[ch]
into btrfs-progs, as this requires a helper that isn't defined in
accessors.h.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There are several different tree block parentness check parameters used
across several helpers:
- level
Mandatory
- transid
Under most cases it's mandatory, but there are several backref cases
which skips this check.
- owner_root
- first_key
Utilized by most top-down tree search routine. Otherwise can be
skipped.
Those four members are not always mandatory checks, and some of them are
the same u64, which means if some arguments got swapped compiler will
not catch it.
Furthermore if we're going to further expand the parentness check, we
need to modify quite some helpers just to add one more parameter.
This patch will concentrate all these members into a structure called
btrfs_tree_parent_check, and pass that structure for the following
helpers:
- btrfs_read_extent_buffer()
- read_tree_block()
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Move these out of ctree.h into relocation.h to cut down on code in
ctree.h
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Update, reformat or reword function comments. This also removes the kdoc
marker so we don't get reports when the function name is missing.
Changes made:
- remove kdoc markers
- reformat the brief description to be a proper sentence
- reword to imperative voice
- align parameter list
- fix typos
Signed-off-by: David Sterba <dsterba@suse.com>
Move all the extent tree related prototypes to extent-tree.h out of
ctree.h, and then go include it everywhere needed so everything
compiles.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This is specific to the item-accessor code, move it out of ctree.h into
accessor.h/.c and then update the users to include the new header file.
This un-inlines btrfs_init_map_token, however this is only called once
per function so it's not critical to be inlined. This also saves 904
bytes of code on a release build.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Anand Jain <anand.jain@oracle.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're going to use fs.h to hold fs wide related helpers and definitions,
move the FS_STATE enum and related helpers to fs.h, and then update all
files that need these definitions to include fs.h.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Anand Jain <anand.jain@oracle.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 have a bunch of printk helpers that are in ctree.h. These have
nothing to do with ctree.c, so move them into their own header.
Subsequent patches will cleanup the printk helpers.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
All callers of btrfs_tree_mod_log_insert_key() are now passing a GFP_NOFS
flag to it, so remove the flag from it and from alloc_tree_mod_elem() and
use it directly within alloc_tree_mod_elem().
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When fixing up the first key of each node above the current level, at
fixup_low_keys(), we are doing a GFP_ATOMIC allocation for inserting an
operation record for the tree mod log. However we can do just fine with
GFP_NOFS nowadays. The need for GFP_ATOMIC was for the old days when we
had custom locks with spinning behaviour for extent buffers and we were
in spinning mode while at fixup_low_keys(). Now we use rw semaphores for
extent buffer locks, so we can safely use GFP_NOFS.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
This uses btrfs_header_nritems, which I will be moving out of ctree.h.
In order to avoid needing to include the relevant header in ctree.h,
simply move this helper function into ctree.c.
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
[ rename parameters ]
Signed-off-by: David Sterba <dsterba@suse.com>
This is local to the ctree code, remove it from ctree.h and inode.c,
create new init/exit functions for the cachep, and move it locally to
ctree.c.
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When doing a nowait buffered write we can trigger the following assertion:
[11138.437027] assertion failed: !path->nowait, in fs/btrfs/ctree.c:4658
[11138.438251] ------------[ cut here ]------------
[11138.438254] kernel BUG at fs/btrfs/messages.c:259!
[11138.438762] invalid opcode: 0000 [#1] PREEMPT SMP DEBUG_PAGEALLOC PTI
[11138.439450] CPU: 4 PID: 1091021 Comm: fsstress Not tainted 6.1.0-rc4-btrfs-next-128 #1
[11138.440611] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.14.0-0-g155821a1990b-prebuilt.qemu.org 04/01/2014
[11138.442553] RIP: 0010:btrfs_assertfail+0x19/0x1b [btrfs]
[11138.443583] Code: 5b 41 5a 41 (...)
[11138.446437] RSP: 0018:ffffbaf0cf05b840 EFLAGS: 00010246
[11138.447235] RAX: 0000000000000039 RBX: ffffbaf0cf05b938 RCX: 0000000000000000
[11138.448303] RDX: 0000000000000000 RSI: ffffffffb2ef59f6 RDI: 00000000ffffffff
[11138.449370] RBP: ffff9165f581eb68 R08: 00000000ffffffff R09: 0000000000000001
[11138.450493] R10: ffff9167a88421f8 R11: 0000000000000000 R12: ffff9164981b1000
[11138.451661] R13: 000000008c8f1000 R14: ffff9164991d4000 R15: ffff9164981b1000
[11138.452225] FS: 00007f1438a66440(0000) GS:ffff9167ad600000(0000) knlGS:0000000000000000
[11138.452949] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[11138.453394] CR2: 00007f1438a64000 CR3: 0000000100c36002 CR4: 0000000000370ee0
[11138.454057] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[11138.454879] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[11138.455779] Call Trace:
[11138.456211] <TASK>
[11138.456598] btrfs_next_old_leaf.cold+0x18/0x1d [btrfs]
[11138.457827] ? kmem_cache_alloc+0x18d/0x2a0
[11138.458516] btrfs_lookup_csums_range+0x149/0x4d0 [btrfs]
[11138.459407] csum_exist_in_range+0x56/0x110 [btrfs]
[11138.460271] can_nocow_file_extent+0x27c/0x310 [btrfs]
[11138.461155] can_nocow_extent+0x1ec/0x2e0 [btrfs]
[11138.461672] btrfs_check_nocow_lock+0x114/0x1c0 [btrfs]
[11138.462951] btrfs_buffered_write+0x44c/0x8e0 [btrfs]
[11138.463482] btrfs_do_write_iter+0x42b/0x5f0 [btrfs]
[11138.463982] ? lock_release+0x153/0x4a0
[11138.464347] io_write+0x11b/0x570
[11138.464660] ? lock_release+0x153/0x4a0
[11138.465213] ? lock_is_held_type+0xe8/0x140
[11138.466003] io_issue_sqe+0x63/0x4a0
[11138.466339] io_submit_sqes+0x238/0x770
[11138.466741] __do_sys_io_uring_enter+0x37b/0xb10
[11138.467206] ? lock_is_held_type+0xe8/0x140
[11138.467879] ? syscall_enter_from_user_mode+0x1d/0x50
[11138.468688] do_syscall_64+0x38/0x90
[11138.469265] entry_SYSCALL_64_after_hwframe+0x63/0xcd
[11138.470017] RIP: 0033:0x7f1438c539e6
This is because to check if we can NOCOW, we check that if we can NOCOW
into an extent (it's prealloc extent or the inode has NOCOW attribute),
and then check if there are csums for the extent's range in the csum tree.
The search may leave us beyond the last slot of a leaf, and then when
we call btrfs_next_leaf() we end up at btrfs_next_old_leaf() with a
time_seq of 0.
This triggers a failure of the first assertion at btrfs_next_old_leaf(),
since we have a nowait path. With assertions disabled, we simply don't
respect the NOWAIT semantics, allowing the write to block on locks or
blocking on IO for reading an extent buffer from disk.
Fix this by:
1) Triggering the assertion only if time_seq is not 0, which means that
search is being done by a tree mod log user, and in the buffered and
direct IO write paths we don't use the tree mod log;
2) Implementing NOWAIT semantics at btrfs_next_old_leaf(). Any failure to
lock an extent buffer should return immediately and not retry the
search, as well as if we need to do IO to read an extent buffer from
disk.
Fixes: c922b016f3 ("btrfs: assert nowait mode is not used for some btree search functions")
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Add ENOMEM among the error codes that don't print stack trace on
transaction abort. We've got several reports from syzbot that detects
stacks as errors but caused by limiting memory. As this is an artificial
condition we don't need to know where exactly the error happens, the
abort and error cleanup will continue like e.g. for EIO.
As the transaction aborts code needs to be inline in a lot of code, the
implementation cases about minimal bloat. The error codes are in a
separate function and the WARN uses the condition directly. This
increases the code size by 571 bytes on release build.
Alternatives considered: add -ENOMEM among the errors, this increases
size by 2340 bytes, various attempts to combine the WARN and helper
calls, increase by 700 or more bytes.
Example syzbot reports (error -12):
- https://syzkaller.appspot.com/bug?extid=5244d35be7f589cf093e
- https://syzkaller.appspot.com/bug?extid=9c37714c07194d816417
Signed-off-by: David Sterba <dsterba@suse.com>
Adds nowait asserts to btree search functions which are not used by
buffered IO and direct IO paths.
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Stefan Roesch <shr@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
For NOWAIT IOCBs we'll need a way to tell search to not wait on locks
or anything. Accomplish this by adding a path->nowait flag that will
use trylocks and skip reading of metadata, returning -EAGAIN in either
of these cases. For now we only need this for reads, so only the read
side is handled. Add an ASSERT() to catch anybody trying to use this
for writes so they know they'll have to implement the write side.
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Stefan Roesch <shr@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We have been hitting the following lockdep splat with btrfs/187 recently
WARNING: possible circular locking dependency detected
5.19.0-rc8+ #775 Not tainted
------------------------------------------------------
btrfs/752500 is trying to acquire lock:
ffff97e1875a97b8 (btrfs-treloc-02#2){+.+.}-{3:3}, at: __btrfs_tree_lock+0x24/0x110
but task is already holding lock:
ffff97e1875a9278 (btrfs-tree-01/1){+.+.}-{3:3}, at: __btrfs_tree_lock+0x24/0x110
which lock already depends on the new lock.
the existing dependency chain (in reverse order) is:
-> #2 (btrfs-tree-01/1){+.+.}-{3:3}:
down_write_nested+0x41/0x80
__btrfs_tree_lock+0x24/0x110
btrfs_init_new_buffer+0x7d/0x2c0
btrfs_alloc_tree_block+0x120/0x3b0
__btrfs_cow_block+0x136/0x600
btrfs_cow_block+0x10b/0x230
btrfs_search_slot+0x53b/0xb70
btrfs_lookup_inode+0x2a/0xa0
__btrfs_update_delayed_inode+0x5f/0x280
btrfs_async_run_delayed_root+0x24c/0x290
btrfs_work_helper+0xf2/0x3e0
process_one_work+0x271/0x590
worker_thread+0x52/0x3b0
kthread+0xf0/0x120
ret_from_fork+0x1f/0x30
-> #1 (btrfs-tree-01){++++}-{3:3}:
down_write_nested+0x41/0x80
__btrfs_tree_lock+0x24/0x110
btrfs_search_slot+0x3c3/0xb70
do_relocation+0x10c/0x6b0
relocate_tree_blocks+0x317/0x6d0
relocate_block_group+0x1f1/0x560
btrfs_relocate_block_group+0x23e/0x400
btrfs_relocate_chunk+0x4c/0x140
btrfs_balance+0x755/0xe40
btrfs_ioctl+0x1ea2/0x2c90
__x64_sys_ioctl+0x88/0xc0
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x63/0xcd
-> #0 (btrfs-treloc-02#2){+.+.}-{3:3}:
__lock_acquire+0x1122/0x1e10
lock_acquire+0xc2/0x2d0
down_write_nested+0x41/0x80
__btrfs_tree_lock+0x24/0x110
btrfs_lock_root_node+0x31/0x50
btrfs_search_slot+0x1cb/0xb70
replace_path+0x541/0x9f0
merge_reloc_root+0x1d6/0x610
merge_reloc_roots+0xe2/0x260
relocate_block_group+0x2c8/0x560
btrfs_relocate_block_group+0x23e/0x400
btrfs_relocate_chunk+0x4c/0x140
btrfs_balance+0x755/0xe40
btrfs_ioctl+0x1ea2/0x2c90
__x64_sys_ioctl+0x88/0xc0
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x63/0xcd
other info that might help us debug this:
Chain exists of:
btrfs-treloc-02#2 --> btrfs-tree-01 --> btrfs-tree-01/1
Possible unsafe locking scenario:
CPU0 CPU1
---- ----
lock(btrfs-tree-01/1);
lock(btrfs-tree-01);
lock(btrfs-tree-01/1);
lock(btrfs-treloc-02#2);
*** DEADLOCK ***
7 locks held by btrfs/752500:
#0: ffff97e292fdf460 (sb_writers#12){.+.+}-{0:0}, at: btrfs_ioctl+0x208/0x2c90
#1: ffff97e284c02050 (&fs_info->reclaim_bgs_lock){+.+.}-{3:3}, at: btrfs_balance+0x55f/0xe40
#2: ffff97e284c00878 (&fs_info->cleaner_mutex){+.+.}-{3:3}, at: btrfs_relocate_block_group+0x236/0x400
#3: ffff97e292fdf650 (sb_internal#2){.+.+}-{0:0}, at: merge_reloc_root+0xef/0x610
#4: ffff97e284c02378 (btrfs_trans_num_writers){++++}-{0:0}, at: join_transaction+0x1a8/0x5a0
#5: ffff97e284c023a0 (btrfs_trans_num_extwriters){++++}-{0:0}, at: join_transaction+0x1a8/0x5a0
#6: ffff97e1875a9278 (btrfs-tree-01/1){+.+.}-{3:3}, at: __btrfs_tree_lock+0x24/0x110
stack backtrace:
CPU: 1 PID: 752500 Comm: btrfs Not tainted 5.19.0-rc8+ #775
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-2.fc32 04/01/2014
Call Trace:
dump_stack_lvl+0x56/0x73
check_noncircular+0xd6/0x100
? lock_is_held_type+0xe2/0x140
__lock_acquire+0x1122/0x1e10
lock_acquire+0xc2/0x2d0
? __btrfs_tree_lock+0x24/0x110
down_write_nested+0x41/0x80
? __btrfs_tree_lock+0x24/0x110
__btrfs_tree_lock+0x24/0x110
btrfs_lock_root_node+0x31/0x50
btrfs_search_slot+0x1cb/0xb70
? lock_release+0x137/0x2d0
? _raw_spin_unlock+0x29/0x50
? release_extent_buffer+0x128/0x180
replace_path+0x541/0x9f0
merge_reloc_root+0x1d6/0x610
merge_reloc_roots+0xe2/0x260
relocate_block_group+0x2c8/0x560
btrfs_relocate_block_group+0x23e/0x400
btrfs_relocate_chunk+0x4c/0x140
btrfs_balance+0x755/0xe40
btrfs_ioctl+0x1ea2/0x2c90
? lock_is_held_type+0xe2/0x140
? lock_is_held_type+0xe2/0x140
? __x64_sys_ioctl+0x88/0xc0
__x64_sys_ioctl+0x88/0xc0
do_syscall_64+0x38/0x90
entry_SYSCALL_64_after_hwframe+0x63/0xcd
This isn't necessarily new, it's just tricky to hit in practice. There
are two competing things going on here. With relocation we create a
snapshot of every fs tree with a reloc tree. Any extent buffers that
get initialized here are initialized with the reloc root lockdep key.
However since it is a snapshot, any blocks that are currently in cache
that originally belonged to the fs tree will have the normal tree
lockdep key set. This creates the lock dependency of
reloc tree -> normal tree
for the extent buffer locking during the first phase of the relocation
as we walk down the reloc root to relocate blocks.
However this is problematic because the final phase of the relocation is
merging the reloc root into the original fs root. This involves
searching down to any keys that exist in the original fs root and then
swapping the relocated block and the original fs root block. We have to
search down to the fs root first, and then go search the reloc root for
the block we need to replace. This creates the dependency of
normal tree -> reloc tree
which is why lockdep complains.
Additionally even if we were to fix this particular mismatch with a
different nesting for the merge case, we're still slotting in a block
that has a owner of the reloc root objectid into a normal tree, so that
block will have its lockdep key set to the tree reloc root, and create a
lockdep splat later on when we wander into that block from the fs root.
Unfortunately the only solution here is to make sure we do not set the
lockdep key to the reloc tree lockdep key normally, and then reset any
blocks we wander into from the reloc root when we're doing the merged.
This solves the problem of having mixed tree reloc keys intermixed with
normal tree keys, and then allows us to make sure in the merge case we
maintain the lock order of
normal tree -> reloc tree
We handle this by setting a bit on the reloc root when we do the search
for the block we want to relocate, and any block we search into or COW
at that point gets set to the reloc tree key. This works correctly
because we only ever COW down to the parent node, so we aren't resetting
the key for the block we're linking into the fs root.
With this patch we no longer have the lockdep splat in btrfs/187.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The parameter has been added in 2009 in the infamous monster commit
5d4f98a28c ("Btrfs: Mixed back reference (FORWARD ROLLING FORMAT
CHANGE)") but not used ever since. We can sink it and allow further
simplifications.
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Btrfs doesn't check whether the tree block respects the root owner.
This means, if a tree block referred by a parent in extent tree, but has
owner of 5, btrfs can still continue reading the tree block, as long as
it doesn't trigger other sanity checks.
Normally this is fine, but combined with the empty tree check in
check_leaf(), if we hit an empty extent tree, but the root node has
csum tree owner, we can let such extent buffer to sneak in.
Shrink the hole by:
- Do extra eb owner check at tree read time
- Make sure the root owner extent buffer exactly matches the root id.
Unfortunately we can't yet completely patch the hole, there are several
call sites can't pass all info we need:
- For reloc/log trees
Their owner is key::offset, not key::objectid.
We need the full root key to do that accurate check.
For now, we just skip the ownership check for those trees.
- For add_data_references() of relocation
That call site doesn't have any parent/ownership info, as all the
bytenrs are all from btrfs_find_all_leafs().
- For direct backref items walk
Direct backref items records the parent bytenr directly, thus unlike
indirect backref item, we don't do a full tree search.
Thus in that case, we don't have full parent owner to check.
For the later two cases, they all pass 0 as @owner_root, thus we can
skip those cases if @owner_root is 0.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There is a common pattern when searching for a key in btrfs:
* Call btrfs_search_slot to find the slot for the key
* Enter an endless loop:
* If the found slot is larger than the no. of items in the current
leaf, check the next leaf
* If it's still not found in the next leaf, terminate the loop
* Otherwise do something with the found key
* Increment the current slot and continue
To reduce code duplication, we can replace this code pattern with an
iterator macro, similar to the existing for_each_X macros found
elsewhere in the kernel. This also makes the code easier to understand
for newcomers by putting a name to the encapsulated functionality.
Signed-off-by: Marcos Paulo de Souza <mpdesouza@suse.com>
Signed-off-by: Gabriel Niebler <gniebler@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The function btrfs_read_buffer() is useless, it just calls
btree_read_extent_buffer_pages() with exactly the same arguments.
So remove it and rename btree_read_extent_buffer_pages() to
btrfs_read_extent_buffer(), which is a shorter name, has the "btrfs_"
prefix (since it's used outside disk-io.c) and the name is clear enough
about what it does.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The comment at the top of read_block_for_search() is very outdated, as it
refers to the blocking versus spinning path locking modes. We no longer
have these two locking modes after we switched the btree locks from custom
code to rw semaphores. So update the comment to stop referring to the
blocking mode and put it more up to date.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When reading a btree node (or leaf), at read_block_for_search(), if we
can't find its extent buffer in the cache (the fs_info->buffer_radix
radix tree), then we unlock all upper level nodes before reading the
btree node/leaf from disk, to prevent blocking other tasks for too long.
However if we find that the extent buffer is in the cache but it is not
up to date, we don't unlock upper level nodes before reading it from disk,
potentially blocking other tasks on upper level nodes for too long.
Fix this inconsistent behaviour by unlocking upper level nodes if we need
to read a node/leaf from disk because its in-memory extent buffer is not
up to date. If we unlocked upper level nodes then we must return -EAGAIN
to the caller, just like the case where the extent buffer is not cached in
memory. And like that case, we determine if upper level nodes are locked
by checking only if the parent node is locked - if it isn't, then no other
upper level nodes are locked.
This is actually a rare case, as if we have an extent buffer in memory,
it typically has the uptodate flag set and passes all the checks done by
btrfs_buffer_uptodate().
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When reading a btree node, at read_block_for_search(), if we don't find
the node's (or leaf) extent buffer in the cache, we will read it from
disk. Since that requires waiting on IO, we release all upper level nodes
from our path before reading the target node/leaf, and then return -EAGAIN
to the caller, which will make the caller restart the while btree search.
However we are causing the restart of btree search even for cases where
it is not necessary:
1) We have a path with ->skip_locking set to true, typically when doing
a search on a commit root, so we are never holding locks on any node;
2) We are doing a read search (the "ins_len" argument passed to
btrfs_search_slot() is 0), or we are doing a search to modify an
existing key (the "cow" argument passed to btrfs_search_slot() has
a value of 1 and "ins_len" is 0), in which case we never hold locks
for upper level nodes;
3) We are doing a search to insert or delete a key, in which case we may
or may not have upper level nodes locked. That depends on the current
minimum write lock levels at btrfs_search_slot(), if we had to split
or merge parent nodes, if we had to COW upper level nodes and if
we ever visited slot 0 of an upper level node. It's still common to
not have upper level nodes locked, but our current node must be at
least at level 1, for insertions, or at least at level 2 for deletions.
In these cases when we have locks on upper level nodes, they are always
write locks.
These cases where we are not holding locks on upper level nodes far
outweigh the cases where we are holding locks, so it's completely wasteful
to retry the whole search when we have no upper nodes locked.
So change the logic to not return -EAGAIN, and make the caller retry the
search, when we don't have the parent node locked - when it's not locked
it means no other upper level nodes are locked as well.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There is one oddball error handling of btrfs_read_buffer():
ret = btrfs_read_buffer(tmp, gen, parent_level - 1, &first_key);
if (!ret) {
*eb_ret = tmp;
return 0;
}
free_extent_buffer(tmp);
btrfs_release_path(p);
return -EIO;
While all other call sites check the error first. Unify the behavior.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We had an error handling pattern for read_tree_block() like this:
eb = read_tree_block();
if (IS_ERR(eb)) {
/*
* Handling error here
* Normally ended up with return or goto out.
*/
} else if (!extent_buffer_uptodate(eb)) {
/*
* Different error handling here
* Normally also ended up with return or goto out;
*/
}
This is fine, but if we want to add extra check for each
read_tree_block(), the existing if-else-if is not that expandable and
will take reader some seconds to figure out there is no extra branch.
Here we change it to a more common way, without the extra else:
eb = read_tree_block();
if (IS_ERR(eb)) {
/*
* Handling error here
*/
return eb or goto out;
}
if (!extent_buffer_uptodate(eb)) {
/*
* Different error handling here
*/
return eb or goto out;
}
This also removes some oddball call sites which uses some creative way
to check error.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When deleting items from a leaf, we always compute the sum of the data
sizes of the items that are going to be deleted. However we only use
that sum when the last item to delete is behind the last item in the
leaf. This unnecessarily wastes CPU time when we are deleting either
the whole leaf or from some slot > 0 up to the last item in the leaf,
and both of these cases are common (e.g. truncation operation, either
as a result of truncate(2) or when logging inodes, deleting checksums
after removing a large enough extent, etc).
So compute only the sum of the data sizes if the last item to be
deleted does not match the last item in the leaf.
This change if part of a patchset that is comprised of the following
patches:
1/6 btrfs: remove unnecessary leaf free space checks when pushing items
2/6 btrfs: avoid unnecessary COW of leaves when deleting items from a leaf
3/6 btrfs: avoid unnecessary computation when deleting items from a leaf
4/6 btrfs: remove constraint on number of visited leaves when replacing extents
5/6 btrfs: remove useless path release in the fast fsync path
6/6 btrfs: prepare extents to be logged before locking a log tree path
The last patch in the series has some performance test result in its
changelog.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When we delete items from a leaf, if we end up with more than two thirds
of unused leaf space, we try to delete the leaf by moving all its items
into its left and right neighbour leaves. Sometimes that is not possible
because there is not enough free space in the left and right leaves, and
in that case we end up not deleting our leaf.
The way we are doing this is not ideal and can be improved in the
following ways:
1) When we call push_leaf_left(), we pass a value of 1 byte to the data
size parameter of push_leaf_left(). This is not realistic value because
no item can have a size less than 25 bytes, which is the size of struct
btrfs_item. This means that means that if the left leaf has not enough
free space to push any item, we end up COWing it even if we end up not
changing its content at all.
COWing that leaf means allocating a new metadata extent, marking it
dirty and doing more IO when committing a transaction or when syncing a
log tree. For a log tree case, it's particularly more important to
avoid the useless COW operation, as more IO can imply a higher latency
for an fsync operation.
So instead of passing 1 as the minimum data size for push_leaf_left(),
pass the size of the first item in our leaf, as we don't want to COW
the left leaf if we can't at least push the first item of our leaf;
2) When we call push_leaf_right(), we also pass a value of 1 byte as the
data size parameter of push_leaf_right(). Like the previous case, it
will also result in COWing the right leaf even if we are not able to
move any items into it, since there can't be any item with a size
smaller than 25 bytes (the size of struct btrfs_item).
So instead of passing 1 as the minimum data size to push_leaf_right(),
pass a size that corresponds to the sum of the size of all the
remaining items in our leaf. We are not interested in moving less than
that, because if we do, we are not able to delete our leaf and we have
COWed the right leaf for nothing. Plus, moving only some of the items
of our leaf, it means an even less balanced tree.
Just like the previous case, we want to avoid the useless COW of the
right leaf, this way we don't have to spend time allocating one new
metadata extent, and doing more IO when committing a transaction or
syncing a log tree. For the log tree case it's specially more important
because more IO can result in a higher latency for a fsync operation.
So adjust the minimum data size passed to push_leaf_left() and
push_leaf_right() as mentioned above.
This change if part of a patchset that is comprised of the following
patches:
1/6 btrfs: remove unnecessary leaf free space checks when pushing items
2/6 btrfs: avoid unnecessary COW of leaves when deleting items from a leaf
3/6 btrfs: avoid unnecessary computation when deleting items from a leaf
4/6 btrfs: remove constraint on number of visited leaves when replacing extents
5/6 btrfs: remove useless path release in the fast fsync path
6/6 btrfs: prepare extents to be logged before locking a log tree path
Not being able to delete a leaf that became less than 1/3 full after
deleting items from it is actually common. For example, for the fio test
mentioned in the changelog of patch 6/6, we are only able to delete a
leaf at btrfs_del_items() about 5.3% of the time, due to its left and
right neighbour leaves not having enough free space to push all the
remaining items into them.
The last patch in the series has some performance test result in its
changelog.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When trying to push items from a leaf into its left and right neighbours,
we lock the left or right leaf, check if it has the required minimum free
space, COW the leaf and then check again if it has the minimum required
free space. This second check is pointless:
1) Most and foremost because it's not needed. We have a write lock on the
leaf and on its parent node, so no one can come in and change either
the pre-COW or post-COW version of the leaf for the whole duration of
the push_leaf_left() and push_leaf_right() calls;
2) The call to btrfs_leaf_free_space() is not trivial, it has a fair
amount of arithmetic operations and access to fields in the leaf's
header and items, so it's not very cheap.
So remove the duplicated free space checks.
This change if part of a patchset that is comprised of the following
patches:
1/6 btrfs: remove unnecessary leaf free space checks when pushing items
2/6 btrfs: avoid unnecessary COW of leaves when deleting items from a leaf
3/6 btrfs: avoid unnecessary computation when deleting items from a leaf
4/6 btrfs: remove constraint on number of visited leaves when replacing extents
5/6 btrfs: remove useless path release in the fast fsync path
6/6 btrfs: prepare extents to be logged before locking a log tree path
The last patch in the series has some performance test result in its
changelog.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The purpose of this function is to unlock all nodes in a btrfs path
which are above 'lowest_unlock' and whose slot used is different than 0.
As such it used slightly awkward structure of 'if' as well as somewhat
cryptic "no_skip" control variable which denotes whether we should
check the current level of skipability or no.
This patch does the following (cosmetic) refactorings:
* Renames 'no_skip' to 'check_skip' and makes it a boolean. This
variable controls whether we are below the lowest_unlock/skip_level
levels.
* Consolidates the 2 conditions which warrant checking whether the
current level should be skipped under 1 common if (check_skip) branch,
this increase indentation level but is not critical.
* Consolidates the 'skip_level < i && i >= lowest_unlock' and
'i >= lowest_unlock && i > skip_level' condition into a common branch
since those are identical.
* Eliminates the local extent_buffer variable as in this case it doesn't
bring anything to function readability.
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>
The comment refers to the old extent buffer locking code, where we used to
have custom locks that had blocking and spinning behaviour modes. That is
not the case anymore, since we have transitioned to rw semaphores, so the
comment does not offer any value anymore. Remove it.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
After calling split_leaf() we BUG_ON() if the returned value is greater
than zero. However split_leaf() only returns 0, in case of success, or a
negative value in case of an error.
The reason for the BUG_ON() is that if we ever get a positive return
value from split_leaf(), we can not simply propagate it to the callers
of btrfs_search_slot(), as that would be interpreted as "key not found"
and not as an error. That means it could result in callers ending up
causing some potential silent corruption.
So change the BUG_ON() to an ASSERT(), and in case assertions are
disabled, produce a warning and set the return value to an error, to make
it not possible to get into a silent corruption and having the error not
noticed.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
There's quite a significant amount of code for doing the key search for a
leaf at btrfs_search_slot(), with a couple labels and gotos in it, plus
btrfs_search_slot() is already big enough.
So move the logic that does the key search on a leaf into a new helper
function. This makes it better organized, removing the need for the labels
and the gotos, as well as reducing the indentation level and the size of
btrfs_search_slot().
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When inserting a key, we check if the write_lock_level is less than 1,
and if so we set it to 1, release the path and retry the tree traversal.
However that is unnecessary, because when ins_len is greater than 0, we
know that write_lock_level can never be less than 1.
The logic to retry is also buggy, because in case ins_len was decremented,
due to an exact key match and the search is not meant for item extension
(path->search_for_extension is 0), we retry without incrementing ins_len,
which would make the next retry decrement it again by the same amount.
So remove the check for write_lock_level being less than 1 and add an
assertion to assert it's always >= 1.
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
When inserting a new key, we release the write lock on the leaf's parent
only after doing the binary search on the leaf. This is because if the
key ends up at slot 0, we will have to update the key at slot 0 of the
parent node. The same reasoning applies to any other upper level nodes
when their slot is 0. We also need to keep the parent locked in case the
leaf does not have enough free space to insert the new key/item, because
in that case we will split the leaf and we will need to add a new key to
the parent due to a new leaf resulting from the split operation.
However if the leaf has enough space for the new key and the key does not
end up at slot 0 of the leaf we could release our write lock on the parent
before doing the binary search on the leaf to figure out the destination
slot. That leads to reducing the amount of time other tasks are blocked
waiting to lock the parent, therefore increasing parallelism when there
are other tasks that are trying to access other leaves accessible through
the same parent. This also applies to other upper nodes besides the
immediate parent, when their slot is 0, since we keep locks on them until
we figure out if the leaf slot is slot 0 or not.
In fact, having the key ending at up slot 0 when is rare. Typically it
only happens when the key is less than or equals to the smallest, the
"left most", key of the entire btree, during a split attempt when we try
to push to the right sibling leaf or when the caller just wants to update
the item of an existing key. It's also very common that a leaf has enough
space to insert a new key, since after a split we move about half of the
keys from one into the new leaf.
So unlock the parent, and any other upper level nodes, when during a key
insertion we notice the key is greater then the first key in the leaf and
the leaf has enough free space. After unlocking the upper level nodes, do
the binary search using a low boundary of slot 1 and not slot 0, to figure
out the slot where the key will be inserted (or where the key already is
in case it exists and the caller wants to modify its item data).
This extra comparison, with the first key, is cheap and the key is very
likely already in a cache line because it immediately follows the header
of the extent buffer and we have recently read the level field of the
header (which in fact is the last field of the header).
The following fs_mark test was run on a non-debug kernel (debian's default
kernel config), with a 12 cores intel CPU, and using a NVMe device:
$ cat run-fsmark.sh
#!/bin/bash
DEV=/dev/nvme0n1
MNT=/mnt/nvme0n1
MOUNT_OPTIONS="-o ssd"
MKFS_OPTIONS="-O no-holes -R free-space-tree"
FILES=100000
THREADS=$(nproc --all)
FILE_SIZE=0
echo "performance" | \
tee /sys/devices/system/cpu/cpu*/cpufreq/scaling_governor
mkfs.btrfs -f $MKFS_OPTIONS $DEV
mount $MOUNT_OPTIONS $DEV $MNT
OPTS="-S 0 -L 10 -n $FILES -s $FILE_SIZE -t $THREADS -k"
for ((i = 1; i <= $THREADS; i++)); do
OPTS="$OPTS -d $MNT/d$i"
done
fs_mark $OPTS
umount $MNT
Before this change:
FSUse% Count Size Files/sec App Overhead
0 1200000 0 165273.6 5958381
0 2400000 0 190938.3 6284477
0 3600000 0 181429.1 6044059
0 4800000 0 173979.2 6223418
0 6000000 0 139288.0 6384560
0 7200000 0 163000.4 6520083
1 8400000 0 57799.2 5388544
1 9600000 0 66461.6 5552969
2 10800000 0 49593.5 5163675
2 12000000 0 57672.1 4889398
After this change:
FSUse% Count Size Files/sec App Overhead
0 1200000 0 167987.3 (+1.6%) 6272730
0 2400000 0 198563.9 (+4.0%) 6048847
0 3600000 0 197436.6 (+8.8%) 6163637
0 4800000 0 202880.7 (+16.6%) 6371771
1 6000000 0 167275.9 (+20.1%) 6556733
1 7200000 0 204051.2 (+25.2%) 6817091
1 8400000 0 69622.8 (+20.5%) 5525675
1 9600000 0 69384.5 (+4.4%) 5700723
1 10800000 0 61454.1 (+23.9%) 5363754
3 12000000 0 61908.7 (+7.3%) 5370196
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Right now generic_bin_search() always uses a low boundary slot of 0, but
in the next patch we'll want to often skip slot 0 when searching for a
key. So make generic_bin_search() have the low boundary slot specified
as an argument, and move the check for the extent buffer level from
btrfs_bin_search() to generic_bin_search() to avoid adding another
wrapper around generic_bin_search().
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Now that we clear the extent buffer uptodate if we fail to write it out
we need to check to see if our root node is uptodate before we search
down it. Otherwise we could return stale data (or potentially corrupt
data that was caught by the write verification step) and think that the
path is OK to search down.
CC: stable@vger.kernel.org # 5.4+
Reviewed-by: Nikolay Borisov <nborisov@suse.com>
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Signed-off-by: David Sterba <dsterba@suse.com>
We don't allow send and balance/relocation to run in parallel in order
to prevent send failing or silently producing some bad stream. This is
because while send is using an extent (specially metadata) or about to
read a metadata extent and expecting it belongs to a specific parent
node, relocation can run, the transaction used for the relocation is
committed and the extent gets reallocated while send is still using the
extent, so it ends up with a different content than expected. This can
result in just failing to read a metadata extent due to failure of the
validation checks (parent transid, level, etc), failure to find a
backreference for a data extent, and other unexpected failures. Besides
reallocation, there's also a similar problem of an extent getting
discarded when it's unpinned after the transaction used for block group
relocation is committed.
The restriction between balance and send was added in commit 9e967495e0
("Btrfs: prevent send failures and crashes due to concurrent relocation"),
kernel 5.3, while the more general restriction between send and relocation
was added in commit 1cea5cf0e6 ("btrfs: ensure relocation never runs
while we have send operations running"), kernel 5.14.
Both send and relocation can be very long running operations. Relocation
because it has to do a lot of IO and expensive backreference lookups in
case there are many snapshots, and send due to read IO when operating on
very large trees. This makes it inconvenient for users and tools to deal
with scheduling both operations.
For zoned filesystem we also have automatic block group relocation, so
send can fail with -EAGAIN when users least expect it or send can end up
delaying the block group relocation for too long. In the future we might
also get the automatic block group relocation for non zoned filesystems.
This change makes it possible for send and relocation to run in parallel.
This is achieved the following way:
1) For all tree searches, send acquires a read lock on the commit root
semaphore;
2) After each tree search, and before releasing the commit root semaphore,
the leaf is cloned and placed in the search path (struct btrfs_path);
3) After releasing the commit root semaphore, the changed_cb() callback
is invoked, which operates on the leaf and writes commands to the pipe
(or file in case send/receive is not used with a pipe). It's important
here to not hold a lock on the commit root semaphore, because if we did
we could deadlock when sending and receiving to the same filesystem
using a pipe - the send task blocks on the pipe because it's full, the
receive task, which is the only consumer of the pipe, triggers a
transaction commit when attempting to create a subvolume or reserve
space for a write operation for example, but the transaction commit
blocks trying to write lock the commit root semaphore, resulting in a
deadlock;
4) Before moving to the next key, or advancing to the next change in case
of an incremental send, check if a transaction used for relocation was
committed (or is about to finish its commit). If so, release the search
path(s) and restart the search, to where we were before, so that we
don't operate on stale extent buffers. The search restarts are always
possible because both the send and parent roots are RO, and no one can
add, remove of update keys (change their offset) in RO trees - the
only exception is deduplication, but that is still not allowed to run
in parallel with send;
5) Periodically check if there is contention on the commit root semaphore,
which means there is a transaction commit trying to write lock it, and
release the semaphore and reschedule if there is contention, so as to
avoid causing any significant delays to transaction commits.
This leaves some room for optimizations for send to have less path
releases and re searching the trees when there's relocation running, but
for now it's kept simple as it performs quite well (on very large trees
with resulting send streams in the order of a few hundred gigabytes).
Test case btrfs/187, from fstests, stresses relocation, send and
deduplication attempting to run in parallel, but without verifying if send
succeeds and if it produces correct streams. A new test case will be added
that exercises relocation happening in parallel with send and then checks
that send succeeds and the resulting streams are correct.
A final note is that for now this still leaves the mutual exclusion
between send operations and deduplication on files belonging to a root
used by send operations. A solution for that will be slightly more complex
but it will eventually be built on top of this change.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
The name btrfs_item_end_nr() is a bit of a misnomer, as it's actually
the offset of the end of the data the item points to. In fact all of
the helpers that we use btrfs_item_end_nr() use data in their name, like
BTRFS_LEAF_DATA_SIZE() and leaf_data(). Rename to btrfs_item_data_end()
to make it clear what this helper is giving us.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Now that all call sites are using the slot number to modify item values,
rename the SETGET helpers to raw_item_*(), and then rework the _nr()
helpers to be the btrfs_item_*() btrfs_set_item_*() helpers, and then
rename all of the callers to the new helpers.
Signed-off-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Filipe Manana