Josef Bacik 77d20c685b btrfs: do not block starts waiting on previous transaction commit ... Internally I got a report of very long stalls on normal operations like creating a new file when auto relocation was running. The reporter used the 'bpf offcputime' tracer to show that we would get stuck in start_transaction for 5 to 30 seconds, and were always being woken up by the transaction commit. Using my timing-everything script, which times how long a function takes and what percentage of that total time is taken up by its children, I saw several traces like this 1083 took 32812902424 ns 29929002926 ns 91.2110% wait_for_commit_duration 25568 ns 7.7920e-05% commit_fs_roots_duration 1007751 ns 0.00307% commit_cowonly_roots_duration 446855602 ns 1.36182% btrfs_run_delayed_refs_duration 271980 ns 0.00082% btrfs_run_delayed_items_duration 2008 ns 6.1195e-06% btrfs_apply_pending_changes_duration 9656 ns 2.9427e-05% switch_commit_roots_duration 1598 ns 4.8700e-06% btrfs_commit_device_sizes_duration 4314 ns 1.3147e-05% btrfs_free_log_root_tree_duration Here I was only tracing functions that happen where we are between START_COMMIT and UNBLOCKED in order to see what would be keeping us blocked for so long. The wait_for_commit() we do is where we wait for a previous transaction that hasn't completed it's commit. This can include all of the unpin work and other cleanups, which tends to be the longest part of our transaction commit. There is no reason we should be blocking new things from entering the transaction at this point, it just adds to random latency spikes for no reason. Fix this by adding a PREP stage. This allows us to properly deal with multiple committers coming in at the same time, we retain the behavior that the winner waits on the previous transaction and the losers all wait for this transaction commit to occur. Nothing else is blocked during the PREP stage, and then once the wait is complete we switch to COMMIT_START and all of the same behavior as before is maintained. Reviewed-by: Filipe Manana <fdmanana@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>		2023-09-08 14:10:49 +02:00
..
9p	fs/9p: Remove unused extern declaration	2023-07-20 19:21:48 +00:00
adfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
affs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
afs	afs: Fix accidental truncation when storing data	2023-07-04 12:24:32 -07:00
autofs	arch/*/configs/*defconfig: Replace AUTOFS4_FS by AUTOFS_FS	2023-07-29 14:08:22 -07:00
befs	befs: Replace all non-returning strlcpy with strscpy	2023-05-30 16:42:00 -07:00
bfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
btrfs	btrfs: do not block starts waiting on previous transaction commit	2023-09-08 14:10:49 +02:00
cachefiles	v6.5/vfs.file	2023-06-26 10:14:36 -07:00
ceph	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
coda	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
configfs	fs: consolidate duplicate dt_type helpers	2023-04-03 09:23:54 +02:00
cramfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
crypto	fscrypt: Replace 1-element array with flexible array	2023-05-23 19:46:09 -07:00
debugfs	debugfs: Correct the 'debugfs_create_str' docs	2023-05-31 19:02:14 +01:00
devpts	devpts: simplify two-level sysctl registration for pty_kern_table	2023-03-13 12:36:34 +01:00
dlm	dlm for 6.5	2023-06-29 13:27:50 -07:00
ecryptfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
efivarfs	efivarfs: expose used and total size	2023-05-17 18:21:34 +02:00
efs
erofs	erofs: drop unnecessary WARN_ON() in erofs_kill_sb()	2023-08-01 16:12:24 +08:00
exfat	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
exportfs	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
ext2	\n	2023-06-29 13:39:51 -07:00
ext4	ext4: fix rbtree traversal bug in ext4_mb_use_preallocated	2023-07-23 08:21:14 -04:00
f2fs	f2fs update for 6.5-rc1	2023-07-05 14:14:37 -07:00
fat	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
freevxfs	There is no particular theme here - mainly quick hits all over the tree.	2023-02-23 17:55:40 -08:00
fscache	fscache: Use clear_and_wake_up_bit() in fscache_create_volume_work()	2023-01-30 12:51:54 +00:00
fuse	fuse update for 6.5	2023-07-19 11:00:27 -07:00
gfs2	gfs2: Don't use filemap_splice_read	2023-08-07 18:42:04 +02:00
hfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
hfsplus	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
hostfs	Landlock updates for v6.5-rc1	2023-06-27 17:10:27 -07:00
hpfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
hugetlbfs	hugetlb: revert use of page_cache_next_miss()	2023-06-23 16:59:32 -07:00
iomap	iomap: micro optimize the ki_pos assignment in iomap_file_buffered_write	2023-07-17 08:49:57 -07:00
isofs
jbd2	jbd2: remove __journal_try_to_free_buffer()	2023-07-10 23:09:21 -04:00
jffs2	for-6.5/splice-2023-06-23	2023-06-26 11:52:12 -07:00
jfs	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
kernfs	driver core changes for 6.5-rc1	2023-07-03 12:56:23 -07:00
lockd	NFS client updates for Linux 6.5	2023-07-01 14:38:25 -07:00
minix	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
netfs	Move netfs_extract_iter_to_sg() to lib/scatterlist.c	2023-06-08 13:42:33 +02:00
nfs	NFS client updates for Linux 6.5	2023-07-01 14:38:25 -07:00
nfs_common	NFSv4.2: remove MODULE_LICENSE in non-modules	2023-04-13 13:13:52 -07:00
nfsd	nfsd-6.5 fixes:	2023-08-03 09:26:34 -07:00
nilfs2	nilfs2: fix use-after-free of nilfs_root in dirtying inodes via iput	2023-08-04 13:03:43 -07:00
nls	fs/nls: make load_nls() take a const parameter	2023-07-25 00:30:02 -05:00
notify	fanotify: disallow mount/sb marks on kernel internal pseudo fs	2023-07-04 13:29:29 +02:00
ntfs	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
ntfs3	driver ntfs3 for linux 6.5	2023-07-07 14:59:38 -07:00
ocfs2	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
omfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
openpromfs
orangefs	orangefs: Provide a splice-read wrapper	2023-05-24 08:42:16 -06:00
overlayfs	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
proc	14 hotfixes. 11 of these are cc:stable and the remainder address post-6.4	2023-08-11 14:19:20 -07:00
pstore	pstore updates for v6.5-rc1	2023-06-27 21:21:32 -07:00
qnx4	qnx4: credit contributors in CREDITS	2023-03-14 12:56:30 -06:00
qnx6	qnx6: credit contributor and mark filesystem orphan	2023-03-14 12:56:30 -06:00
quota	quota: fix warning in dqgrab()	2023-06-05 16:50:30 +02:00
ramfs	- Yosry Ahmed brought back some cgroup v1 stats in OOM logs.	2023-06-28 10:28:11 -07:00
reiserfs	- Yosry Ahmed brought back some cgroup v1 stats in OOM logs.	2023-06-28 10:28:11 -07:00
romfs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
smb	smb: client: fix null auth	2023-08-16 00:26:07 -05:00
squashfs	squashfs: fix cache race with migration	2023-07-08 09:29:30 -07:00
sysfs	sysfs: Skip empty folders creation	2023-06-15 13:37:53 +02:00
sysv	for-6.5/splice-2023-06-23	2023-06-26 11:52:12 -07:00
tracefs	fs: port ->mkdir() to pass mnt_idmap	2023-01-19 09:24:26 +01:00
ubifs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
udf	\n	2023-06-29 13:39:51 -07:00
ufs	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
unicode	unicode: remove MODULE_LICENSE in non-modules	2023-04-13 13:13:54 -07:00
vboxsf	hardening fixes for v6.5-rc6	2023-08-08 14:59:49 -07:00
verity	fsverity: improve documentation for builtin signature support	2023-06-20 22:47:55 -07:00
xfs	xfs: convert flex-array declarations in xfs attr shortform objects	2023-07-17 08:48:56 -07:00
zonefs	zonefs: fix synchronous direct writes to sequential files	2023-08-10 12:59:47 +09:00
aio.c	fs/aio: Stop allocating aio rings from HIGHMEM	2023-06-15 09:22:23 +02:00
anon_inodes.c
attr.c	nfs: use vfs setgid helper	2023-03-30 08:51:48 +02:00
bad_inode.c	fs: port ->permission() to pass mnt_idmap	2023-01-19 09:24:28 +01:00
binfmt_elf_fdpic.c	binfmt: Slightly simplify elf_fdpic_map_file()	2023-05-30 15:49:46 -07:00
binfmt_elf_test.c
binfmt_elf.c	Merge branch 'expand-stack'	2023-06-28 20:35:21 -07:00
binfmt_flat.c
binfmt_misc.c	binfmt_misc: fix shift-out-of-bounds in check_special_flags	2022-12-02 13:57:04 -08:00
binfmt_script.c
buffer.c	\n	2023-06-29 13:39:51 -07:00
char_dev.c	vfs: Replace all non-returning strlcpy with strscpy	2023-05-15 09:42:01 +02:00
compat_binfmt_elf.c
coredump.c	v6.5/vfs.misc	2023-06-26 09:50:21 -07:00
d_path.c	fs: d_path: include internal.h	2023-05-17 09:16:59 +02:00
dax.c	dax: enable dax fault handler to report VM_FAULT_HWPOISON	2023-06-26 07:54:23 -06:00
dcache.c
direct-io.c	- Yosry Ahmed brought back some cgroup v1 stats in OOM logs.	2023-06-28 10:28:11 -07:00
drop_caches.c
eventfd.c	eventfd: show the EFD_SEMAPHORE flag in fdinfo	2023-06-15 09:22:23 +02:00
eventpoll.c	v6.5/vfs.misc	2023-06-26 09:50:21 -07:00
exec.c	\n	2023-06-29 13:31:44 -07:00
fcntl.c	fs.idmapped.v6.3	2023-02-20 11:53:11 -08:00
fhandle.c	fsnotify: move fsnotify_open() hook into do_dentry_open()	2023-06-12 10:43:45 +02:00
file_table.c	fs: move cleanup from init_file() into its callers	2023-07-02 13:15:49 +02:00
file.c	fs: rely on ->iterate_shared to determine f_pos locking	2023-08-06 15:08:36 +02:00
filesystems.c
fs_context.c	fs: avoid empty option when generating legacy mount string	2023-06-07 21:49:55 +02:00
fs_parser.c	ext4: journal_path mount options should follow links	2022-12-01 10:46:54 -05:00
fs_pin.c
fs_struct.c
fs_types.c
fs-writeback.c	writeback: move wb_over_bg_thresh() call outside lock section	2023-06-09 16:25:14 -07:00
fsopen.c
init.c	fs: port ->permission() to pass mnt_idmap	2023-01-19 09:24:28 +01:00
inode.c	locking: remove spin_lock_prefetch	2023-08-12 09:18:47 -07:00
internal.h	v6.5/vfs.file	2023-06-26 10:14:36 -07:00
ioctl.c	fs: port inode_owner_or_capable() to mnt_idmap	2023-01-19 09:24:29 +01:00
Kconfig	smb: move client and server files to common directory fs/smb	2023-05-24 16:29:21 -05:00
Kconfig.binfmt
kernel_read_file.c
libfs.c	fs: factor out a direct_write_fallback helper	2023-06-09 16:25:53 -07:00
locks.c	filelocks: use mount idmapping for setlease permission check	2023-03-09 22:36:12 +01:00
Makefile	for-6.5/block-2023-06-23	2023-06-26 12:47:20 -07:00
mbcache.c	ext4: fix deadlock due to mbcache entry corruption	2022-12-08 21:49:25 -05:00
mnt_idmapping.c	fs: move mnt_idmap	2023-01-19 09:24:30 +01:00
mount.h
mpage.c	mpage: use folios in bio end_io handler	2023-04-18 16:30:02 -07:00
namei.c	fs: no need to check source	2023-07-04 10:20:29 +02:00
namespace.c	v6.5/vfs.mount	2023-06-26 10:27:04 -07:00
nsfs.c	kill the last remaining user of proc_ns_fget()	2023-04-20 22:55:35 -04:00
open.c	open: make RESOLVE_CACHED correctly test for O_TMPFILE	2023-08-06 15:08:35 +02:00
pipe.c	pipe: check for IOCB_NOWAIT alongside O_NONBLOCK	2023-05-12 17:17:27 +02:00
pnode.c	fs: allow to mount beneath top mount	2023-05-19 04:30:22 +02:00
pnode.h	fs: allow to mount beneath top mount	2023-05-19 04:30:22 +02:00
posix_acl.c	acl: don't depend on IOP_XATTR	2023-03-06 09:59:20 +01:00
proc_namespace.c	tty, proc, kernfs, random: Use copy_splice_read()	2023-05-24 08:42:16 -06:00
read_write.c	splice: Use filemap_splice_read() instead of generic_file_splice_read()	2023-05-24 08:42:17 -06:00
readdir.c	vfs: get rid of old '->iterate' directory operation	2023-08-06 15:08:35 +02:00
remap_range.c	fs: use UB-safe check for signed addition overflow in remap_verify_area	2023-05-24 11:03:59 +02:00
select.c
seq_file.c	use less confusing names for iov_iter direction initializers	2022-11-25 13:01:55 -05:00
signalfd.c
splice.c	splice, net: Fix splice_to_socket() for O_NONBLOCK socket	2023-07-26 21:56:06 -07:00
stack.c
stat.c	fs.idmapped.v6.3	2023-02-20 11:53:11 -08:00
statfs.c	statfs: enforce statfs[64] structure initialization	2023-05-17 15:20:17 +02:00
super.c	\n	2023-06-29 13:39:51 -07:00
sync.c
sysctls.c	sysctl: Refactor base paths registrations	2023-05-23 21:43:26 -07:00
timerfd.c
userfaultfd.c	Merge mm-hotfixes-stable into mm-stable to pick up depended-upon changes.	2023-06-23 16:58:19 -07:00
utimes.c	fs.idmapped.v6.3	2023-02-20 11:53:11 -08:00
xattr.c	fs: don't call posix_acl_listxattr in generic_listxattr	2023-05-17 15:25:20 +02:00