tcp: Revert "tcp: tcp_probe: use spin_lock_bh()"
authorEric Dumazet <edumazet@google.com>
Tue, 21 Feb 2017 14:21:47 +0000 (06:21 -0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 15 Sep 2018 07:43:01 +0000 (09:43 +0200)
commit 29869d66870a715177bfb505f66a7e0e8bcc89c3 upstream.

This reverts commit e70ac171658679ecf6bea4bbd9e9325cd6079d2b.

jtcp_rcv_established() is in fact called with hard irq being disabled.

Initial bug report from Ricardo Nabinger Sanchez [1] still needs
to be investigated, but does not look like a TCP bug.

[1] https://www.spinics.net/lists/netdev/msg420960.html

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: kernel test robot <xiaolong.ye@intel.com>
Cc: Ricardo Nabinger Sanchez <rnsanchez@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/ipv4/tcp_probe.c

index 3d063eb..f6c50af 100644 (file)
@@ -117,7 +117,7 @@ static void jtcp_rcv_established(struct sock *sk, struct sk_buff *skb,
             (fwmark > 0 && skb->mark == fwmark)) &&
            (full || tp->snd_cwnd != tcp_probe.lastcwnd)) {
 
-               spin_lock_bh(&tcp_probe.lock);
+               spin_lock(&tcp_probe.lock);
                /* If log fills, just silently drop */
                if (tcp_probe_avail() > 1) {
                        struct tcp_log *p = tcp_probe.log + tcp_probe.head;
@@ -157,7 +157,7 @@ static void jtcp_rcv_established(struct sock *sk, struct sk_buff *skb,
                        tcp_probe.head = (tcp_probe.head + 1) & (bufsize - 1);
                }
                tcp_probe.lastcwnd = tp->snd_cwnd;
-               spin_unlock_bh(&tcp_probe.lock);
+               spin_unlock(&tcp_probe.lock);
 
                wake_up(&tcp_probe.wait);
        }