From: Sagi Grimberg <sagi@grimberg.me>
Date: Sun, 21 Mar 2021 07:08:48 +0000 (-0700)
Subject: nvme-tcp: block BH in sk state_change sk callback
X-Git-Url: http://git.maquefel.me/?a=commitdiff_plain;h=8b73b45d54a14588f86792869bfb23098ea254cb;p=linux.git

nvme-tcp: block BH in sk state_change sk callback

The TCP stack can run from process context for a long time
so we should disable BH here.

Fixes: 3f2304f8c6d6 ("nvme-tcp: add NVMe over TCP host driver")
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
---

diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 7de9bee1e5e96..b9e8ea3a7501f 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -870,7 +870,7 @@ static void nvme_tcp_state_change(struct sock *sk)
 {
 	struct nvme_tcp_queue *queue;
 
-	read_lock(&sk->sk_callback_lock);
+	read_lock_bh(&sk->sk_callback_lock);
 	queue = sk->sk_user_data;
 	if (!queue)
 		goto done;
@@ -891,7 +891,7 @@ static void nvme_tcp_state_change(struct sock *sk)
 
 	queue->state_change(sk);
 done:
-	read_unlock(&sk->sk_callback_lock);
+	read_unlock_bh(&sk->sk_callback_lock);
 }
 
 static inline bool nvme_tcp_queue_more(struct nvme_tcp_queue *queue)