diff options
| author | Simon Rettberg | 2020-07-17 18:19:58 +0200 |
|---|---|---|
| committer | Simon Rettberg | 2020-07-17 18:19:58 +0200 |
| commit | a231e6283e1c79a33d6f3bdf786832f10e355d0c (patch) | |
| tree | a5be40fa9fe93230d06de56ccac3c378f80c3833 /src/kernel/net-txrx.c | |
| parent | [KERNEL] After nullptr check, MAYBE RETURN!? (diff) | |
| download | dnbd3-ng-a231e6283e1c79a33d6f3bdf786832f10e355d0c.tar.gz dnbd3-ng-a231e6283e1c79a33d6f3bdf786832f10e355d0c.tar.xz dnbd3-ng-a231e6283e1c79a33d6f3bdf786832f10e355d0c.zip | |
[KERNEL] Trying to get locking under control
Current code doesn't pay much attention to properly locking when
accessing structures that are being used in multiple workers/threads.
Try to fix this, slowly.
Diffstat (limited to 'src/kernel/net-txrx.c')
| -rw-r--r-- | src/kernel/net-txrx.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/src/kernel/net-txrx.c b/src/kernel/net-txrx.c index c839561..51c684d 100644 --- a/src/kernel/net-txrx.c +++ b/src/kernel/net-txrx.c @@ -345,8 +345,8 @@ int dnbd3_receive_cmd_get_block_mq(struct dnbd3_sock *sock, error_sock(sock, "unexpected reply (%d) %p", tag, req); if (req) { cmd = blk_mq_rq_to_pdu(req); - mutex_lock(&cmd->lock); debug_sock(sock, "requeue request"); + mutex_lock(&cmd->lock); dnbd3_requeue_cmd(cmd); mutex_unlock(&cmd->lock); } |