Return-Path: Received: from mail-m2837.qiye.163.com (mail-m2837.qiye.163.com [103.74.28.37]) by mail19.linbit.com (LINBIT Mail Daemon) with ESMTP id B1F0D4201DC for ; Tue, 26 Apr 2022 10:09:08 +0200 (CEST) From: Rui Xu To: philipp.reisner@linbit.com, drbd-dev@lists.linbit.com, joel.colledge@linbit.com Date: Tue, 26 Apr 2022 16:09:03 +0800 Message-Id: <20220426080903.1831211-1-rui.xu@easystack.cn> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Cc: Rui Xu , dongsheng.yang@easystack.cn Subject: [Drbd-dev] [PATCH v2] drbd: fix a race of drbd_free_peer_req List-Id: "*Coordination* of development, patches, contributions -- *Questions* \(even to developers\) go to drbd-user, please." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Commit e061feb8 and 9a11ebeb5 introduce a bug to race with drbd_free_peer_req() in got_peer_ack. drbd_free_page_chain in drbd_finish_peer_reqs and got_peer_ack may happen concurrently, the sequence is as follows: drbd_finish_peer_reqs got_peer_ack drbd_free_page_chain drbd_free_peer_req spin_lock() page_chain_add spin_unlock() spin_lock() page_chain_add spin_unlock() We can see that page_chain_add will be called twice which will lead to crash. Swap drbd_may_finish_epoch() and drbd_free_page_chain() in e_end_block() to ensure that got_peer_ack will not be called while still processing the request in drbd_finish_peer_reqs(). Signed-off-by: Rui Xu --- drbd/drbd_receiver.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drbd/drbd_receiver.c b/drbd/drbd_receiver.c index f077f2b..d0f42d3 100644 --- a/drbd/drbd_receiver.c +++ b/drbd/drbd_receiver.c @@ -2560,9 +2560,10 @@ static int e_end_block(struct drbd_work *w, int cancel) } else D_ASSERT(device, drbd_interval_empty(&peer_req->i)); + drbd_free_page_chain(&peer_device->connection->transport, &peer_req->page_chain, 0); + drbd_may_finish_epoch(peer_device->connection, peer_req->epoch, EV_PUT + (cancel ? EV_CLEANUP : 0)); - drbd_free_page_chain(&peer_device->connection->transport, &peer_req->page_chain, 0); return err; } -- 1.8.3.1