[Devel] [PATCH hci-8.1 v2 0/6] dm/push_backup: event fd interface and fixes

[Andrey Zhadchenko]

The push_backup target currently makes the backup tool wait for the next
cluster range inside the "push_backup_read" target message. dm runs
target_message() under dm_get_live_table(), i.e. with md->io_barrier SRCU
read-locked, so while the tool sleeps there anything that needs
synchronize_srcu(&md->io_barrier) on that device is stuck: suspend, table
reload/remove and dm_internal_suspend_fast() from "@stats_create". An
unrelated dm-stats user hangs, and everything that later queues behind
md->suspend_lock hangs with it.

Patch 6 adds a file descriptor based interface instead. The tool gets an
anon inode fd via the "push_backup_get_fd" message and then uses plain
read()/write() of struct push_backup_range to fetch the next range and to
acknowledge it, without going through dm messages on the data path. The
fd pins struct push_backup with a kref so it survives target removal and
simply returns -ENODEV afterwards. The legacy messages are kept for
compatibility.

Patches 1-5 are fixes found while working on this:

  1. init the lock, waitqueue and timer before anything that can fail so
     pb_destroy() is safe from any error path;
  2. clear pb->alive under pb->lock like the other users do;
  3. drop the redundant mod_timer() in push_backup_write(). The timer
     re-arms itself from pb->deadline_jiffies, so updating the deadline
     under the lock is enough. This was also the only place that armed
     the timer outside the window push_backup_stop() waits for, which
     matters once writes may arrive via the fd without pb->ctl_rwsem;
  4. pass absolute expiry to mod_timer(). Both remaining callers passed
     a relative value, so after ~5 min of uptime the timer fired and
     re-armed on every tick while a write was delayed;
  5. arm the timer under pb->lock. pb_timer_func() and
     postpone_if_required_for_backup() both decided under the lock and
     armed after dropping it, so a stale S64_MAX could overwrite a fresh
     deadline and park the timer ~12 days out with delayed writes still
     pending. With arming done under the lock the RCU dance between
     postpone and stop is no longer needed and is removed.

v2:
  - 2/6: reworded the commit message, dropped the KCSAN claim.
  - 5/6: new patch: arm the deadline timer under pb->lock and drop the
    rcu_read_lock()/synchronize_rcu() between
    postpone_if_required_for_backup() and push_backup_stop().
  - 6/6: resubmit_pbios() now frees the pbio before
    dm_requeue_original_rq(). The requeue drops the md reference held by
    the in-flight request, after which __dm_destroy() may run .dtr and
    destroy pbio_pool under the fd writer's feet.

https://virtuozzo.atlassian.net/browse/VSTOR-152912

Andrey Zhadchenko (6):
  drivers/md/push_backup: correct initialization
  drivers/md/push_backup: clear pb->alive under pb->lock
  drivers/md/push_backup: do not re-arm deadline timer in
    push_backup_write()
  drivers/md/push_backup: fix mod_timer() usage
  drivers/md/push_backup: arm deadline timer under pb->lock
  dm/push_backup: add event file descriptor interface

 drivers/md/dm-push-backup.c         | 252 +++++++++++++++++++++-------
 include/uapi/linux/dm-push-backup.h |  36 ++++
 2 files changed, 229 insertions(+), 59 deletions(-)
 create mode 100644 include/uapi/linux/dm-push-backup.h

-- 
2.43.5