cifs: on replayable errors back-off before replay, not after

On replayable errors, we call smb2_should_replays that does these
things today:
1. decide if we need to replay the command again
2. sleep to back-off the failed request
3. update the next sleep value

We will not be able to use this for async requests, when this is
processed in callbacks (as this will be called in cifsd threads that
should not sleep in response processing).

Modify the behaviour by taking the sleep out of smb2_should_replay
and performing the sleep for back-off just before actually
performing the replay.

Signed-off-by: Shyam Prasad N <sprasad@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
This commit is contained in:
Shyam Prasad N 2026-01-31 14:03:03 +05:30 • committed by Steve French
commit 16d480ed49
4 changed files with 129 additions and 42 deletions