drbd: fix for possible deadlock on IO error during resync

Browse source

Scenario:

Something (say, flush-147:0) is in drbd_al_begin_io,
holding a local_cnt, waiting for the resync to make progress.

Disk fails, worker in after_state_ch does drbd_rs_cancel_all,
then waits for local_cnt to drop to zero.

flush-147:0 is woken by drbd_rs_cancel_all, needs to write an AL
transaction, and queues that on the worker.

Deadlock.

Fix: do not wait in the worker, have put_ldev() trigger the
state change D_FAILED -> D_DISKLESS when necessary.
put_ldev() cannot do the state change directly, as it may or may not
already hold various spinlocks. We queue a short work instead.

Signed-off-by: Philipp Reisner <philipp.reisner@linbit.com>
Signed-off-by: Lars Ellenberg <lars.ellenberg@linbit.com>

...

This commit is contained in:

Lars Ellenberg 2010-09-14 20:26:27 +02:00 • committed by Philipp Reisner

parent 22cc37a943

commit e9e6f3ec53

2 changed files with 54 additions and 22 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

8

drivers/block/drbd/drbd_int.h

View file

72

drivers/block/drbd/drbd_main.c

View file