linux.git/fs/fs-writeback.c, branch v6.12 Linux kernel source tree inode: port __I_SYNC to var event 2024-08-30T06:22:39+00:00 Christian Brauner brauner@kernel.org 2024-08-23T12:47:37+00:00 532980cb1bff9b942c23fe94324dee560f5f57a5 Port the __I_SYNC mechanism to use the new var event mechanism. Link: https://lore.kernel.org/r/20240823-work-i_state-v3-3-5cd5fd207a57@kernel.org Reviewed-by: Josef Bacik <josef@toxicpanda.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
Port the __I_SYNC mechanism to use the new var event mechanism.

Link: https://lore.kernel.org/r/20240823-work-i_state-v3-3-5cd5fd207a57@kernel.org
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
vfs: drop one lock trip in evict() 2024-08-30T06:22:38+00:00 Mateusz Guzik mjguzik@gmail.com 2024-08-13T14:36:26+00:00 57510c58b5832c8cb36516cdba48543133f3ab85 Most commonly neither I_LRU_ISOLATING nor I_SYNC are set, but the stock kernel takes a back-to-back relock trip to check for them. It probably can be avoided altogether, but for now massage things back to just one lock acquire. Signed-off-by: Mateusz Guzik <mjguzik@gmail.com> Link: https://lore.kernel.org/r/20240813143626.1573445-1-mjguzik@gmail.com Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
Most commonly neither I_LRU_ISOLATING nor I_SYNC are set, but the stock
kernel takes a back-to-back relock trip to check for them.

It probably can be avoided altogether, but for now massage things back
to just one lock acquire.

Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
Link: https://lore.kernel.org/r/20240813143626.1573445-1-mjguzik@gmail.com
Reviewed-by: Zhihao Cheng <chengzhihao1@huawei.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs: don't flush in-flight wb switches for superblocks without cgroup writeback 2024-08-19T11:45:02+00:00 Haifeng Xu haifeng.xu@shopee.com 2024-07-26T03:05:25+00:00 c393eaa85349941badf3ce5f087400dbaf3bbe02 When deactivating any type of superblock, it had to wait for the in-flight wb switches to be completed. wb switches are executed in inode_switch_wbs_work_fn() which needs to acquire the wb_switch_rwsem and races against sync_inodes_sb(). If there are too much dirty data in the superblock, the waiting time may increase significantly. For superblocks without cgroup writeback such as tmpfs, they have nothing to do with the wb swithes, so the flushing can be avoided. Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com> Link: https://lore.kernel.org/r/20240726030525.180330-1-haifeng.xu@shopee.com Reviewed-by: Jan Kara <jack@suse.cz> Suggested-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
When deactivating any type of superblock, it had to wait for the in-flight
wb switches to be completed. wb switches are executed in inode_switch_wbs_work_fn()
which needs to acquire the wb_switch_rwsem and races against sync_inodes_sb().
If there are too much dirty data in the superblock, the waiting time may increase
significantly.

For superblocks without cgroup writeback such as tmpfs, they have nothing to
do with the wb swithes, so the flushing can be avoided.

Signed-off-by: Haifeng Xu <haifeng.xu@shopee.com>
Link: https://lore.kernel.org/r/20240726030525.180330-1-haifeng.xu@shopee.com
Reviewed-by: Jan Kara <jack@suse.cz>
Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
sysctl: treewide: constify the ctl_table argument of proc_handlers 2024-07-24T18:59:29+00:00 Joel Granados j.granados@samsung.com 2024-07-24T18:59:29+00:00 78eb4ea25cd5fdbdae7eb9fdf87b99195ff67508 const qualify the struct ctl_table argument in the proc_handler function signatures. This is a prerequisite to moving the static ctl_table structs into .rodata data which will ensure that proc_handler function pointers cannot be modified. This patch has been generated by the following coccinelle script: ``` virtual patch @r1@ identifier ctl, write, buffer, lenp, ppos; identifier func !~ "appldata_(timer|interval)_handler|sched_(rt|rr)_handler|rds_tcp_skbuf_handler|proc_sctp_do_(hmac_alg|rto_min|rto_max|udp_port|alpha_beta|auth|probe_interval)"; @@ int func( - struct ctl_table *ctl + const struct ctl_table *ctl ,int write, void *buffer, size_t *lenp, loff_t *ppos); @r2@ identifier func, ctl, write, buffer, lenp, ppos; @@ int func( - struct ctl_table *ctl + const struct ctl_table *ctl ,int write, void *buffer, size_t *lenp, loff_t *ppos) { ... } @r3@ identifier func; @@ int func( - struct ctl_table * + const struct ctl_table * ,int , void *, size_t *, loff_t *); @r4@ identifier func, ctl; @@ int func( - struct ctl_table *ctl + const struct ctl_table *ctl ,int , void *, size_t *, loff_t *); @r5@ identifier func, write, buffer, lenp, ppos; @@ int func( - struct ctl_table * + const struct ctl_table * ,int write, void *buffer, size_t *lenp, loff_t *ppos); ``` * Code formatting was adjusted in xfs_sysctl.c to comply with code conventions. The xfs_stats_clear_proc_handler, xfs_panic_mask_proc_handler and xfs_deprecated_dointvec_minmax where adjusted. * The ctl_table argument in proc_watchdog_common was const qualified. This is called from a proc_handler itself and is calling back into another proc_handler, making it necessary to change it as part of the proc_handler migration. Co-developed-by: Thomas Weißschuh <linux@weissschuh.net> Signed-off-by: Thomas Weißschuh <linux@weissschuh.net> Co-developed-by: Joel Granados <j.granados@samsung.com> Signed-off-by: Joel Granados <j.granados@samsung.com> 
const qualify the struct ctl_table argument in the proc_handler function
signatures. This is a prerequisite to moving the static ctl_table
structs into .rodata data which will ensure that proc_handler function
pointers cannot be modified.

This patch has been generated by the following coccinelle script:

```
  virtual patch

  @r1@
  identifier ctl, write, buffer, lenp, ppos;
  identifier func !~ "appldata_(timer|interval)_handler|sched_(rt|rr)_handler|rds_tcp_skbuf_handler|proc_sctp_do_(hmac_alg|rto_min|rto_max|udp_port|alpha_beta|auth|probe_interval)";
  @@

  int func(
  - struct ctl_table *ctl
  + const struct ctl_table *ctl
    ,int write, void *buffer, size_t *lenp, loff_t *ppos);

  @r2@
  identifier func, ctl, write, buffer, lenp, ppos;
  @@

  int func(
  - struct ctl_table *ctl
  + const struct ctl_table *ctl
    ,int write, void *buffer, size_t *lenp, loff_t *ppos)
  { ... }

  @r3@
  identifier func;
  @@

  int func(
  - struct ctl_table *
  + const struct ctl_table *
    ,int , void *, size_t *, loff_t *);

  @r4@
  identifier func, ctl;
  @@

  int func(
  - struct ctl_table *ctl
  + const struct ctl_table *ctl
    ,int , void *, size_t *, loff_t *);

  @r5@
  identifier func, write, buffer, lenp, ppos;
  @@

  int func(
  - struct ctl_table *
  + const struct ctl_table *
    ,int write, void *buffer, size_t *lenp, loff_t *ppos);

```

* Code formatting was adjusted in xfs_sysctl.c to comply with code
  conventions. The xfs_stats_clear_proc_handler,
  xfs_panic_mask_proc_handler and xfs_deprecated_dointvec_minmax where
  adjusted.

* The ctl_table argument in proc_watchdog_common was const qualified.
  This is called from a proc_handler itself and is calling back into
  another proc_handler, making it necessary to change it as part of the
  proc_handler migration.

Co-developed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Co-developed-by: Joel Granados <j.granados@samsung.com>
Signed-off-by: Joel Granados <j.granados@samsung.com>
fs/writeback: remove unnecessary return in writeback_inodes_sb 2024-04-05T13:53:45+00:00 Kemeng Shi shikemeng@huaweicloud.com 2024-02-28T09:19:58+00:00 6a1ee87176ffed1d9e749bc66a2ad85be2d7dbb7 writeback_inodes_sb doesn't have return value, just remove unnecessary return in it. Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> Link: https://lore.kernel.org/r/20240228091958.288260-7-shikemeng@huaweicloud.com Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
writeback_inodes_sb doesn't have return value, just remove unnecessary
return in it.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Link: https://lore.kernel.org/r/20240228091958.288260-7-shikemeng@huaweicloud.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/writeback: correct comment of __wakeup_flusher_threads_bdi 2024-04-05T13:53:45+00:00 Kemeng Shi shikemeng@huaweicloud.com 2024-02-28T09:19:57+00:00 ba679de9c3fc511f457ea0ad8f5a22e9152fa07b Commit e8e8a0c6c9bfc ("writeback: move nr_pages == 0 logic to one location") removed parameter nr_pages of __wakeup_flusher_threads_bdi and we try to writeback all dirty pages in __wakeup_flusher_threads_bdi now. Just correct stale comment. Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> Link: https://lore.kernel.org/r/20240228091958.288260-6-shikemeng@huaweicloud.com Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
Commit e8e8a0c6c9bfc ("writeback: move nr_pages == 0 logic to one
location") removed parameter nr_pages of __wakeup_flusher_threads_bdi
and we try to writeback all dirty pages in __wakeup_flusher_threads_bdi
now. Just correct stale comment.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Link: https://lore.kernel.org/r/20240228091958.288260-6-shikemeng@huaweicloud.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/writeback: only calculate dirtied_before when b_io is empty 2024-04-05T13:53:45+00:00 Kemeng Shi shikemeng@huaweicloud.com 2024-02-28T09:19:56+00:00 639924abc1ae28eb05893a402081e8e6cff23b8a The dirtied_before is only used when b_io is not empty, so only calculate when b_io is not empty. Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> Link: https://lore.kernel.org/r/20240228091958.288260-5-shikemeng@huaweicloud.com Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
The dirtied_before is only used when b_io is not empty, so only calculate
when b_io is not empty.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Link: https://lore.kernel.org/r/20240228091958.288260-5-shikemeng@huaweicloud.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/writeback: remove unused parameter wb of finish_writeback_work 2024-04-05T13:53:45+00:00 Kemeng Shi shikemeng@huaweicloud.com 2024-02-28T09:19:55+00:00 2ddc93461214507b8e50ba7218d6260be8d623d1 Remove unused parameter wb of finish_writeback_work. Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> Link: https://lore.kernel.org/r/20240228091958.288260-4-shikemeng@huaweicloud.com Reviewed-by: Tim Chen <tim.c.chen@linux.intel.com> Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
Remove unused parameter wb of finish_writeback_work.

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Link: https://lore.kernel.org/r/20240228091958.288260-4-shikemeng@huaweicloud.com
Reviewed-by: Tim Chen <tim.c.chen@linux.intel.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/writeback: bail out if there is no more inodes for IO and queued once 2024-04-05T13:52:17+00:00 Kemeng Shi shikemeng@huaweicloud.com 2024-02-28T09:19:54+00:00 d92109891f21cf367caa2cc6dff11a4411d917f4 For case there is no more inodes for IO in io list from last wb_writeback, We may bail out early even there is inode in dirty list should be written back. Only bail out when we queued once to avoid missing dirtied inode. This is from code reading... Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> Link: https://lore.kernel.org/r/20240228091958.288260-3-shikemeng@huaweicloud.com Reviewed-by: Jan Kara <jack@suse.cz> [brauner@kernel.org: fold in memory corruption fix from Jan in [1]] Link: https://lore.kernel.org/r/20240405132346.bid7gibby3lxxhez@quack3 [1] Signed-off-by: Christian Brauner <brauner@kernel.org> 
For case there is no more inodes for IO in io list from last wb_writeback,
We may bail out early even there is inode in dirty list should be written
back. Only bail out when we queued once to avoid missing dirtied inode.

This is from code reading...

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Link: https://lore.kernel.org/r/20240228091958.288260-3-shikemeng@huaweicloud.com
Reviewed-by: Jan Kara <jack@suse.cz>
[brauner@kernel.org: fold in memory corruption fix from Jan in [1]]
Link: https://lore.kernel.org/r/20240405132346.bid7gibby3lxxhez@quack3 [1]
Signed-off-by: Christian Brauner <brauner@kernel.org>
fs/writeback: avoid to writeback non-expired inode in kupdate writeback 2024-03-26T08:01:18+00:00 Kemeng Shi shikemeng@huaweicloud.com 2024-02-28T09:19:53+00:00 ac0c18f2c693f0e7a44dbbb36b14d5141e5d20e5 In kupdate writeback, only expired inode (have been dirty for longer than dirty_expire_interval) is supposed to be written back. However, kupdate writeback will writeback non-expired inode left in b_io or b_more_io from last wb_writeback. As a result, writeback will keep being triggered unexpected when we keep dirtying pages even dirty memory is under threshold and inode is not expired. To be more specific: Assume dirty background threshold is > 1G and dirty_expire_centisecs is > 60s. When we running fio -size=1G -invalidate=0 -ioengine=libaio --time_based -runtime=60... (keep dirtying), the writeback will keep being triggered as following: wb_workfn wb_do_writeback wb_check_background_flush /* * Wb dirty background threshold starts at 0 if device was idle and * grows up when bandwidth of wb is updated. So a background * writeback is triggered. */ wb_over_bg_thresh /* * Dirtied inode will be written back and added to b_more_io list * after slice used up (because we keep dirtying the inode). */ wb_writeback Writeback is triggered per dirty_writeback_centisecs as following: wb_workfn wb_do_writeback wb_check_old_data_flush /* * Write back inode left in b_io and b_more_io from last wb_writeback * even the inode is non-expired and it will be added to b_more_io * again as slice will be used up (because we keep dirtying the * inode) */ wb_writeback Fix this by moving non-expired inode to dirty list instead of more io list for kupdate writeback in requeue_inode. Test as following: /* make it more easier to observe the issue */ echo 300000 > /proc/sys/vm/dirty_expire_centisecs echo 100 > /proc/sys/vm/dirty_writeback_centisecs /* create a idle device */ mkfs.ext4 -F /dev/vdb mount /dev/vdb /bdi1/ /* run buffer write with fio */ fio -name test -filename=/bdi1/file -size=800M -ioengine=libaio -bs=4K \ -iodepth=1 -rw=write -direct=0 --time_based -runtime=60 -invalidate=0 Fio result before fix (run three tests): 1360MB/s 1329MB/s 1455MB/s Fio result after fix (run three tests): 1737MB/s 1729MB/s 1789MB/s Writeback for non-expired inode is gone as expeted. Observe this with trace writeback_start and writeback_written as following: echo 1 > /sys/kernel/debug/tracing/events/writeback/writeback_start/enab echo 1 > /sys/kernel/debug/tracing/events/writeback/writeback_written/enable cat /sys/kernel/tracing/trace_pipe Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com> Link: https://lore.kernel.org/r/20240228091958.288260-2-shikemeng@huaweicloud.com Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> 
In kupdate writeback, only expired inode (have been dirty for longer than
dirty_expire_interval) is supposed to be written back. However, kupdate
writeback will writeback non-expired inode left in b_io or b_more_io from
last wb_writeback. As a result, writeback will keep being triggered
unexpected when we keep dirtying pages even dirty memory is under
threshold and inode is not expired. To be more specific:
Assume dirty background threshold is > 1G and dirty_expire_centisecs is
> 60s. When we running fio -size=1G -invalidate=0 -ioengine=libaio
--time_based -runtime=60... (keep dirtying), the writeback will keep
being triggered as following:
wb_workfn
  wb_do_writeback
    wb_check_background_flush
      /*
       * Wb dirty background threshold starts at 0 if device was idle and
       * grows up when bandwidth of wb is updated. So a background
       * writeback is triggered.
       */
      wb_over_bg_thresh
      /*
       * Dirtied inode will be written back and added to b_more_io list
       * after slice used up (because we keep dirtying the inode).
       */
      wb_writeback

Writeback is triggered per dirty_writeback_centisecs as following:
wb_workfn
  wb_do_writeback
    wb_check_old_data_flush
      /*
       * Write back inode left in b_io and b_more_io from last wb_writeback
       * even the inode is non-expired and it will be added to b_more_io
       * again as slice will be used up (because we keep dirtying the
       * inode)
       */
      wb_writeback

Fix this by moving non-expired inode to dirty list instead of more io
list for kupdate writeback in requeue_inode.

Test as following:
/* make it more easier to observe the issue */
echo 300000 > /proc/sys/vm/dirty_expire_centisecs
echo 100 > /proc/sys/vm/dirty_writeback_centisecs
/* create a idle device */
mkfs.ext4 -F /dev/vdb
mount /dev/vdb /bdi1/
/* run buffer write with fio */
fio -name test -filename=/bdi1/file -size=800M -ioengine=libaio -bs=4K \
-iodepth=1 -rw=write -direct=0 --time_based -runtime=60 -invalidate=0

Fio result before fix (run three tests):
1360MB/s
1329MB/s
1455MB/s

Fio result after fix (run three tests):
1737MB/s
1729MB/s
1789MB/s

Writeback for non-expired inode is gone as expeted. Observe this with trace
writeback_start and writeback_written as following:
echo 1 > /sys/kernel/debug/tracing/events/writeback/writeback_start/enab
echo 1 > /sys/kernel/debug/tracing/events/writeback/writeback_written/enable
cat /sys/kernel/tracing/trace_pipe

Signed-off-by: Kemeng Shi <shikemeng@huaweicloud.com>
Link: https://lore.kernel.org/r/20240228091958.288260-2-shikemeng@huaweicloud.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>