From 8c3088f895a0cfdf630973a9cd32f5e085eb973c Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 6 Aug 2019 17:09:14 +0200 Subject: [PATCH] tcp: be more careful in tcp_fragment() commit b617158dc096709d8600c53b6052144d12b89fab upstream. Some applications set tiny SO_SNDBUF values and expect TCP to just work. Recent patches to address CVE-2019-11478 broke them in case of losses, since retransmits might be prevented. We should allow these flows to make progress. This patch allows the first and last skb in retransmit queue to be split even if memory limits are hit. It also adds the some room due to the fact that tcp_sendmsg() and tcp_sendpage() might overshoot sk_wmem_queued by about one full TSO skb (64KB size). Note this allowance was already present in stable backports for kernels < 4.15 Note for < 4.15 backports : tcp_rtx_queue_tail() will probably look like : static inline struct sk_buff *tcp_rtx_queue_tail(const struct sock *sk) { struct sk_buff *skb = tcp_send_head(sk); return skb ? tcp_write_queue_prev(sk, skb) : tcp_write_queue_tail(sk); } Fixes: f070ef2ac667 ("tcp: tcp_fragment() should apply sane memory limits") Signed-off-by: Eric Dumazet Reported-by: Andrew Prout Tested-by: Andrew Prout Tested-by: Jonathan Lemon Tested-by: Michal Kubecek Acked-by: Neal Cardwell Acked-by: Yuchung Cheng Acked-by: Christoph Paasch Cc: Jonathan Looney Signed-off-by: David S. Miller Signed-off-by: Matthieu Baerts Signed-off-by: Sasha Levin --- include/net/tcp.h | 17 +++++++++++++++++ net/ipv4/tcp_output.c | 11 ++++++++++- 2 files changed, 27 insertions(+), 1 deletion(-) diff --git a/include/net/tcp.h b/include/net/tcp.h index 0b477a1e1177..7994e569644e 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -1688,6 +1688,23 @@ static inline void tcp_check_send_head(struct sock *sk, struct sk_buff *skb_unli tcp_sk(sk)->highest_sack = NULL; } +static inline struct sk_buff *tcp_rtx_queue_head(const struct sock *sk) +{ + struct sk_buff *skb = tcp_write_queue_head(sk); + + if (skb == tcp_send_head(sk)) + skb = NULL; + + return skb; +} + +static inline struct sk_buff *tcp_rtx_queue_tail(const struct sock *sk) +{ + struct sk_buff *skb = tcp_send_head(sk); + + return skb ? tcp_write_queue_prev(sk, skb) : tcp_write_queue_tail(sk); +} + static inline void __tcp_add_write_queue_tail(struct sock *sk, struct sk_buff *skb) { __skb_queue_tail(&sk->sk_write_queue, skb); diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index a5960b9b6741..a99086bf26ea 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1264,6 +1264,7 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *buff; int nsize, old_factor; + long limit; int nlen; u8 flags; @@ -1274,7 +1275,15 @@ int tcp_fragment(struct sock *sk, struct sk_buff *skb, u32 len, if (nsize < 0) nsize = 0; - if (unlikely((sk->sk_wmem_queued >> 1) > sk->sk_sndbuf + 0x20000)) { + /* tcp_sendmsg() can overshoot sk_wmem_queued by one full size skb. + * We need some allowance to not penalize applications setting small + * SO_SNDBUF values. + * Also allow first and last skb in retransmit queue to be split. + */ + limit = sk->sk_sndbuf + 2 * SKB_TRUESIZE(GSO_MAX_SIZE); + if (unlikely((sk->sk_wmem_queued >> 1) > limit && + skb != tcp_rtx_queue_head(sk) && + skb != tcp_rtx_queue_tail(sk))) { NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPWQUEUETOOBIG); return -ENOMEM; } -- 2.20.1