nvme-rdma: fix crash casue by destroy id while resolving addr

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA
Link: https://gitee.com/openeuler/kernel/issues/I1WGZE

-------------------------------------------------

when nvme_rdma_wait_for_cm timeout happened before resolve addr
timeout, we will destroy id while still resolving addr, this will
cause a crash.
so change wait for cm timeout to double amount of resolve addr timeout.
Signed-off-by: NRuozhu Li <liruozhu@huawei.com>
Reviewed-by: NChao Leng <lengchao@huawei.com>
Reviewed-by: NJike Cheng <chengjike.cheng@huawei.com>
Signed-off-by: NLijie <lijie34@huawei.com>
Acked-by: NHanjun Guo <guohanjun@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

drivers/nvme/host/rdma.c

@@ -240,7 +240,7 @@ static void nvme_rdma_qp_event(struct ib_event *event, void *context)
 static int nvme_rdma_wait_for_cm(struct nvme_rdma_queue *queue)
 {
 	wait_for_completion_interruptible_timeout(&queue->cm_done,
-			msecs_to_jiffies(NVME_RDMA_CONNECT_TIMEOUT_MS) + 1);
+			msecs_to_jiffies(2 * NVME_RDMA_CONNECT_TIMEOUT_MS));
 	return queue->cm_error;
 }