[PATCH net] selftests: netfilter: nft_queue.sh: increase udp_gro_ct subtest timeout

[PATCH net] selftests: netfilter: nft_queue.sh: increase udp_gro_ct subtest timeout

[Florian Westphal]

Jakub reports test flakes on debug kernels:
 FAIL: test_udp_gro_ct: Expected software segmentation to occur, had 23 and 17

The nf_queue listener exits when it doesn't receive a new packet after
2s, seems this is too low. Use a 5 second timeout instead, same as the
duration of the packet generation side.

Fixes: 59ecffa3995e ("selftests: netfilter: nft_queue.sh: add udp fraglist gro test case")
Reported-by: Jakub Kicinski <kuba@kernel.org>
Closes: https://lore.kernel.org/netdev/20260218184114.0b405b72@kernel.org/
Signed-off-by: Florian Westphal <fw@strlen.de>
---
 Hello, unfortunately I can't repro the flakes here, but 2s is rather
 short and a nf_queue listener that exits while packets are still queued
 does explain the error message (more packets counted as incoming than
 outgoing).

 tools/testing/selftests/net/netfilter/nft_queue.sh | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/tools/testing/selftests/net/netfilter/nft_queue.sh b/tools/testing/selftests/net/netfilter/nft_queue.sh
index 139bc1211878..5d22a4dc8721 100755
--- a/tools/testing/selftests/net/netfilter/nft_queue.sh
+++ b/tools/testing/selftests/net/netfilter/nft_queue.sh
@@ -591,6 +591,7 @@ EOF
 test_udp_gro_ct()
 {
 	local errprefix="FAIL: test_udp_gro_ct:"
+	local timeout=5
 
 	ip netns exec "$nsrouter" conntrack -F 2>/dev/null
 
@@ -630,10 +631,10 @@ table inet udpq {
 	}
 }
 EOF
-	timeout 10 ip netns exec "$ns2" socat UDP-LISTEN:12346,fork,pf=ipv4 OPEN:"$TMPFILE1",trunc &
+	timeout "$timeout" ip netns exec "$ns2" socat UDP-LISTEN:12346,fork,pf=ipv4 OPEN:"$TMPFILE1",trunc &
 	local rpid=$!
 
-	ip netns exec "$nsrouter" ./nf_queue -G -c -q 1 -t 2 > "$TMPFILE2" &
+	ip netns exec "$nsrouter" ./nf_queue -G -c -q 1 -t "$timeout" > "$TMPFILE2" &
 	local nfqpid=$!
 
 	ip netns exec "$nsrouter" ethtool -K "veth0" rx-udp-gro-forwarding on rx-gro-list on generic-receive-offload on
@@ -644,11 +645,11 @@ EOF
 	local bs=512
 	local count=$(((32 * 1024 * 1024) / bs))
 	dd if=/dev/zero bs="$bs" count="$count" 2>/dev/null | for i in $(seq 1 16); do
-		timeout 5 ip netns exec "$ns1" \
+		timeout "$timeout" ip netns exec "$ns1" \
 			socat -u -b 512 STDIN UDP-DATAGRAM:10.0.2.99:12346,reuseport,bind=0.0.0.0:55221 &
 	done
 
-	busywait 10000 test -s "$TMPFILE1"
+	busywait $((timeout * 1000)) test -s "$TMPFILE1"
 
 	kill "$rpid"
 
-- 
2.52.0

Re: [PATCH net] selftests: netfilter: nft_queue.sh: increase udp_gro_ct subtest timeout

[Jakub Kicinski]

On Mon, 23 Feb 2026 20:00:58 +0100 Florian Westphal wrote:
> Jakub reports test flakes on debug kernels:
>  FAIL: test_udp_gro_ct: Expected software segmentation to occur, had 23 and 17
> 
> The nf_queue listener exits when it doesn't receive a new packet after
> 2s, seems this is too low. Use a 5 second timeout instead, same as the
> duration of the packet generation side.
> 
> Fixes: 59ecffa3995e ("selftests: netfilter: nft_queue.sh: add udp fraglist gro test case")
> Reported-by: Jakub Kicinski <kuba@kernel.org>
> Closes: https://lore.kernel.org/netdev/20260218184114.0b405b72@kernel.org/
> Signed-off-by: Florian Westphal <fw@strlen.de>
> ---
>  Hello, unfortunately I can't repro the flakes here, but 2s is rather
>  short and a nf_queue listener that exits while packets are still queued
>  does explain the error message (more packets counted as incoming than
>  outgoing).

Thanks for looking into it! Unfortunately we saw the problem again with
this patch applied:
https://netdev-ctrl.bots.linux.dev/logs/vmksft/nf-dbg/results/531881/3-nft-queue-sh/stdout

Looks like it repros once every 25 runs for us:
https://netdev.bots.linux.dev/contest.html?executor=vmksft-nf-dbg&test=nft-queue-sh

Re: [PATCH net] selftests: netfilter: nft_queue.sh: increase udp_gro_ct subtest timeout

[Florian Westphal]

Jakub Kicinski <kuba@kernel.org> wrote:
> Thanks for looking into it! Unfortunately we saw the problem again with
> this patch applied:
> https://netdev-ctrl.bots.linux.dev/logs/vmksft/nf-dbg/results/531881/3-nft-queue-sh/stdout
> 
> Looks like it repros once every 25 runs for us:
> https://netdev.bots.linux.dev/contest.html?executor=vmksft-nf-dbg&test=nft-queue-sh

Yes, I already saw that it did not help, I can't explain this so far.  I
restarted a long test run (1000 runs) with '--cpus 4' on my laptop, I
hope it triggers this time :-/

You can mark this patch as deferred or changes requested, I will
hopefully be able to send a v2 that gets rid of this problem.