From: Eric Dumazet edumazet@google.com
Two upstream commits squashed together for v4.14 stable :
commit 88f8598d0a302a08380eadefd09b9f5cb1c4c428 upstream.
Previously TCP only warns if its RTO timer fires and the retransmission queue is empty, but it'll cause null pointer reference later on. It's better to avoid such catastrophic failure and simply exit with a warning.
Squashed with "tcp: refactor tcp_retransmit_timer()" :
commit 0d580fbd2db084a5c96ee9c00492236a279d5e0f upstream.
It appears linux-4.14 stable needs a backport of commit 88f8598d0a30 ("tcp: exit if nothing to retransmit on RTO timeout")
Since tcp_rtx_queue_empty() is not in pre 4.15 kernels, let's refactor tcp_retransmit_timer() to only use tcp_rtx_queue_head()
Signed-off-by: Yuchung Cheng ycheng@google.com Signed-off-by: Eric Dumazet edumazet@google.com Reviewed-by: Neal Cardwell ncardwell@google.com Reviewed-by: Soheil Hassas Yeganeh soheil@google.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- net/ipv4/tcp_timer.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
--- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -413,6 +413,7 @@ void tcp_retransmit_timer(struct sock *s struct tcp_sock *tp = tcp_sk(sk); struct net *net = sock_net(sk); struct inet_connection_sock *icsk = inet_csk(sk); + struct sk_buff *skb;
if (tp->fastopen_rsk) { WARN_ON_ONCE(sk->sk_state != TCP_SYN_RECV && @@ -423,10 +424,13 @@ void tcp_retransmit_timer(struct sock *s */ return; } + if (!tp->packets_out) - goto out; + return;
- WARN_ON(tcp_write_queue_empty(sk)); + skb = tcp_rtx_queue_head(sk); + if (WARN_ON_ONCE(!skb)) + return;
tp->tlp_high_seq = 0;
@@ -459,7 +463,7 @@ void tcp_retransmit_timer(struct sock *s goto out; } tcp_enter_loss(sk); - tcp_retransmit_skb(sk, tcp_write_queue_head(sk), 1); + tcp_retransmit_skb(sk, skb, 1); __sk_dst_reset(sk); goto out_reset_timer; }