From 4a269818a7eb8577d32d8b2879099c689ddbd856 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 11 Oct 2017 13:27:29 -0700 Subject: [PATCH] tcp: fix tcp_unlink_write_queue() Yury reported crash with this signature : [ 554.034021] [] 0xffff80003ccd5a58 [ 554.034156] [] skb_release_all+0x14/0x30 [ 554.034288] [] __kfree_skb+0x14/0x28 [ 554.034409] [] tcp_sendmsg_locked+0x4dc/0xcc8 [ 554.034541] [] tcp_sendmsg+0x34/0x58 [ 554.034659] [] inet_sendmsg+0x2c/0xf8 [ 554.034783] [] sock_sendmsg+0x18/0x30 [ 554.034928] [] SyS_sendto+0x84/0xf8 Problem is that skb->destructor contains garbage, and this is because I accidentally removed tcp_skb_tsorted_anchor_cleanup() from tcp_unlink_write_queue() This would trigger with a write(fd, , len) attempt, and we will add to packetdrill this capability to avoid future regressions. Fixes: 75c119afe14f ("tcp: implement rb-tree based retransmit queue") Reported-by: Yury Norov Tested-by: Yury Norov Signed-off-by: Eric Dumazet Signed-off-by: David S. Miller --- include/net/tcp.h | 1 + 1 file changed, 1 insertion(+) diff --git a/include/net/tcp.h b/include/net/tcp.h index 5a95e5886b55..15163454174b 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -1712,6 +1712,7 @@ static inline void tcp_insert_write_queue_before(struct sk_buff *new, static inline void tcp_unlink_write_queue(struct sk_buff *skb, struct sock *sk) { + tcp_skb_tsorted_anchor_cleanup(skb); __skb_unlink(skb, &sk->sk_write_queue); }