[LU-8765] dead loop in ptlrpc_replay_next()

• Type: Bug
• Resolution: Fixed
• Priority: Minor
• Fix Version/s: Lustre 2.10.0
• Severity: 3
• Assignee: Niu Yawei (Inactive)
• Reporter: Niu Yawei (Inactive)

Description

I see there is a defect about imp_replay_cursor which can lead to ptlrpc_replay_next() run into a dead loop:

• During replay, imp_replay_cursor moves to an open request A;
• Client close the file, so that rq_replay of the A open request is cleared;
• ptlrpc_replay_next() is called to continue replay, it calls ptlrpc_free_committed() to remove committed/closed request from replay/committed list, request A is removed from committed list; (The open request is still being held by the pending close request, so it's not freed);
• ptlrpc_replay_next() then try to move imp_replay_cursor to next, but the next is itself now, dead loop;

Comments

[Peter Jones]

Landed for 2.10

[Gerrit Updater]

Oleg Drokin ([e-mail]) merged in patch https://review.whamcloud.com/23418/
Subject: LU-8765 ptlrpc: update replay cursor when close during replay
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 1e8dfacb6f58d875d7840eb89a3af3e780659367

[Gerrit Updater]

Niu Yawei ([e-mail]) uploaded a new patch: http://review.whamcloud.com/23418
Subject: LU-8765 ptlrpc: update replay cursor when close during replay
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: f09f17661323a0f134d9ef02044b863693ee0a9c