From f67971e683e81d7ba4739728511ae6e52a1b6321 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 2 Nov 2017 18:10:03 -0700 Subject: tcp: tcp_fragment() should not assume rtx skbs While stress testing MTU probing, we had crashes in list_del() that we root-caused to the fact that tcp_fragment() is unconditionally inserting the freshly allocated skb into tsorted_sent_queue list. But this list is supposed to contain skbs that were sent. This was mostly harmless until MTU probing was enabled. Fortunately we can use the tcp_queue enum added later (but in same linux version) for rtx-rb-tree to fix the bug. Fixes: e2080072ed2d ("tcp: new list for sent but unacked skbs for RACK recovery") Signed-off-by: Eric Dumazet Cc: Yuchung Cheng Cc: Neal Cardwell Cc: Soheil Hassas Yeganeh Cc: Alexei Starovoitov Cc: Priyaranjan Jha Acked-by: Neal Cardwell Acked-by: Soheil Hassas Yeganeh Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net/ipv4') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 06a0c89ffe40..822962ece284 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1395,7 +1395,8 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue, /* Link BUFF into the send queue. */ __skb_header_release(buff); tcp_insert_write_queue_after(skb, buff, sk, tcp_queue); - list_add(&buff->tcp_tsorted_anchor, &skb->tcp_tsorted_anchor); + if (tcp_queue == TCP_FRAG_IN_RTX_QUEUE) + list_add(&buff->tcp_tsorted_anchor, &skb->tcp_tsorted_anchor); return 0; } -- cgit v1.2.3