dccp: avoid deadlock in dccp_v4_ctl_send_reset (95556a88) · Commits · 戴 / test

In the prep work I did before enabling BH while handling socket backlog, I missed two points in DCCP : 1) dccp_v4_ctl_send_reset() uses bh_lock_sock(), assuming BH were blocked. It is not anymore always true. 2) dccp_v4_route_skb() was using __IP_INC_STATS() instead of IP_INC_STATS() A similar fix was done for TCP, in commit ("ipv4: tcp: ip_send_unicast_reply() is not BH safe") Fixes: ("dccp: do not assume DCCP code is non preemptible") Fixes: ("net: do not block BH while processing socket backlog") Signed-off-by:

Eric Dumazet <edumazet@google.com> Reported-by:

Dmitry Vyukov <dvyukov@google.com> Signed-off-by:

David S. Miller <davem@davemloft.net>

net/dccp/ipv4.c

+6 −4

Original line number	Diff line number	Diff line
		@@ -462,7 +462,7 @@ static struct dst_entry* dccp_v4_route_skb(struct net net, struct sock sk,
		security_skb_classify_flow(skb, flowi4_to_flowi(&fl4));
		rt = ip_route_output_flow(net, &fl4, sk);
		if (IS_ERR(rt)) {
		__IP_INC_STATS(net, IPSTATS_MIB_OUTNOROUTES);
		IP_INC_STATS(net, IPSTATS_MIB_OUTNOROUTES);
		return NULL;
		}

		@@ -527,15 +527,17 @@ static void dccp_v4_ctl_send_reset(const struct sock sk, struct sk_buff rxskb)
		rxiph->daddr);
		skb_dst_set(skb, dst_clone(dst));

		local_bh_disable();
		bh_lock_sock(ctl_sk);
		err = ip_build_and_send_pkt(skb, ctl_sk,
		rxiph->daddr, rxiph->saddr, NULL);
		bh_unlock_sock(ctl_sk);

		if (net_xmit_eval(err) == 0) {
		DCCP_INC_STATS(DCCP_MIB_OUTSEGS);
		DCCP_INC_STATS(DCCP_MIB_OUTRSTS);
		__DCCP_INC_STATS(DCCP_MIB_OUTSEGS);
		__DCCP_INC_STATS(DCCP_MIB_OUTRSTS);
		}
		local_bh_enable();
		out:
		dst_release(dst);
		}

Admin message