Commit d2527ad3 authored by Willem de Bruijn's avatar Willem de Bruijn Committed by Jakub Kicinski
Browse files

net: preserve MSG_ZEROCOPY with forwarding 



MSG_ZEROCOPY data must be copied before data is queued to local
sockets, to avoid indefinite timeout until memory release.

This test is performed by skb_orphan_frags_rx, which is called when
looping an egress skb to packet sockets, error queue or ingress path.

To preserve zerocopy for skbs that are looped to ingress but are then
forwarded to an egress device rather than delivered locally, defer
this last check until an skb enters the local IP receive path.

This is analogous to existing behavior of skb_clear_delivery_time.

Signed-off-by: default avatarWillem de Bruijn <willemb@google.com>
Link: https://patch.msgid.link/20250630194312.1571410-2-willemdebruijn.kernel@gmail.com


Signed-off-by: default avatarJakub Kicinski <kuba@kernel.org>
parent 04b1d18c

net/core/dev.c

+0 −2
Original line number Diff line number Diff line
@@ -5937,8 +5937,6 @@ static int __netif_receive_skb_core(struct sk_buff **pskb, bool pfmemalloc, 
	}



	if (pt_prev) {

		if (unlikely(skb_orphan_frags_rx(skb, GFP_ATOMIC)))

			goto drop;

		*ppt_prev = pt_prev;

	} else {

drop:

net/ipv4/ip_input.c

+6 −0
Original line number Diff line number Diff line
@@ -226,6 +226,12 @@ void ip_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int protocol) 


static int ip_local_deliver_finish(struct net *net, struct sock *sk, struct sk_buff *skb)

{

	if (unlikely(skb_orphan_frags_rx(skb, GFP_ATOMIC))) {

		__IP_INC_STATS(net, IPSTATS_MIB_INDISCARDS);

		kfree_skb_reason(skb, SKB_DROP_REASON_NOMEM);

		return 0;

	}



	skb_clear_delivery_time(skb);

	__skb_pull(skb, skb_network_header_len(skb));

net/ipv6/ip6_input.c

+7 −0
Original line number Diff line number Diff line
@@ -478,6 +478,13 @@ void ip6_protocol_deliver_rcu(struct net *net, struct sk_buff *skb, int nexthdr, 


static int ip6_input_finish(struct net *net, struct sock *sk, struct sk_buff *skb)

{

	if (unlikely(skb_orphan_frags_rx(skb, GFP_ATOMIC))) {

		__IP6_INC_STATS(net, ip6_dst_idev(skb_dst(skb)),

				IPSTATS_MIB_INDISCARDS);

		kfree_skb_reason(skb, SKB_DROP_REASON_NOMEM);

		return 0;

	}



	skb_clear_delivery_time(skb);

	ip6_protocol_deliver_rcu(net, skb, 0, false);