From: Alexander Aring aahringo@redhat.com
[ Upstream commit eec054b5a7cfe6d1f1598a323b05771ee99857b5 ]
This patch fixes the flushing of send work before shutdown. The function cancel_work_sync() is not the right workqueue functionality to use here as it would cancel the work if the work queues itself. In cases of EAGAIN in send() for dlm message we need to be sure that everything is send out before. The function flush_work() will ensure that every send work is be done inclusive in EAGAIN cases.
Signed-off-by: Alexander Aring aahringo@redhat.com Signed-off-by: David Teigland teigland@redhat.com Signed-off-by: Sasha Levin sashal@kernel.org --- fs/dlm/lowcomms.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c index 79f56f16bc2c..44e2716ac158 100644 --- a/fs/dlm/lowcomms.c +++ b/fs/dlm/lowcomms.c @@ -612,10 +612,7 @@ static void shutdown_connection(struct connection *con) { int ret;
- if (cancel_work_sync(&con->swork)) { - log_print("canceled swork for node %d", con->nodeid); - clear_bit(CF_WRITE_PENDING, &con->flags); - } + flush_work(&con->swork);
mutex_lock(&con->sock_mutex); /* nothing to shutdown */