Skip to contentSkip to conversation

net: clear skb->tstamp in forwarding paths

PATCH In review Eric Dumazet 2 participants 2 messages Dec 14, 2018 – Dec 15, 2018

Subsystems (3)
Networking Network drivers IPv4/IPv6
lore
Share & tools
lore

Eric Dumazet

author
#0 lore raw

PATCH 1/1 net: clear skb->tstamp in forwarding paths

Sergey reported that forwarding was no longer working if fq packet scheduler was used.

This is caused by the recent switch to EDT model, since incoming packets might have been timestamped by __net_timestamp()

__net_timestamp() uses ktime_get_real(), while fq expects packets using CLOCK_MONOTONIC base.

The fix is to clear skb->tstamp in forwarding paths.

Fixes 80b14dee2bea ("net: Add a new socket option for a future transmit time.")
Fixes fb420d5d91c1 ("tcp/fq: move back to CLOCK_MONOTONIC")
Signed-off-by Eric Dumazet <[email protected]>
Reported-by Sergey Matyukevich <[email protected]>
Tested-by Sergey Matyukevich <[email protected]>
---
 net/ipv4/ip_forward.c | 1 +
 net/ipv6/ip6_output.c | 1 +
 2 files changed, 2 insertions(+)
net/ipv4/ip_forward.c +1 -0
index 32662e9e5d218868341169bba1dc3ab430952c58..d5984d31ab931487fdea844aa4a240df00d2b64e 100644
@@ -72,6 +72,7 @@ static int ip_forward_finish(struct net *net, struct sock *sk, struct sk_buff *s
 	if (unlikely(opt->optlen))
 		ip_forward_options(skb);
 
+	skb->tstamp = 0;
 	return dst_output(net, sk, skb);
 }
net/ipv6/ip6_output.c +1 -0
index fcd3c66ded1620d0d320fd7b1753f2734d1dfa46..4591ca4bdbe8831f8183c4df0d4237d023b5bb68 100644
@@ -378,6 +378,7 @@ static inline int ip6_forward_finish(struct net *net, struct sock *sk,
 	__IP6_INC_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTFORWDATAGRAMS);
 	__IP6_ADD_STATS(net, ip6_dst_idev(dst), IPSTATS_MIB_OUTOCTETS, skb->len);
 
+	skb->tstamp = 0;
 	return dst_output(net, sk, skb);
 }
Signature · 1 line
2.20.0.405.gbc1bbc6f85-goog
maintainer
#1 lore raw

From: Eric Dumazet <[email protected]>
Date: Fri, 14 Dec 2018 06:46:49 -0800

Quoted 9 lines

Sergey reported that forwarding was no longer working if fq packet scheduler was used.

This is caused by the recent switch to EDT model, since incoming packets might have been timestamped by __net_timestamp()

__net_timestamp() uses ktime_get_real(), while fq expects packets using CLOCK_MONOTONIC base.

Quoted 7 lines

The fix is to clear skb->tstamp in forwarding paths.

Fixes 80b14dee2bea ("net: Add a new socket option for a future transmit time.")
Fixes fb420d5d91c1 ("tcp/fq: move back to CLOCK_MONOTONIC")
Signed-off-by Eric Dumazet <[email protected]>
Reported-by Sergey Matyukevich <[email protected]>
Tested-by Sergey Matyukevich <[email protected]>

Applied and queued up for -stable, thanks Eric.