Netdev List
 help / color / mirror / Atom feed
* [PATCH net] tcp: ignore locked zero CWND route metric
@ 2026-10-03 15:32 bestswngs
  2026-10-03 16:10 ` Eric Dumazet
  0 siblings, 1 reply; 4+ messages in thread
From: bestswngs @ 2026-10-03 15:32 UTC (permalink / raw)
  To: edumazet, ncardwell, netdev
  Cc: kuniyu, davem, horms, xmei5, co+71b6509cb791e531, Weiming Shi

From: Weiming Shi <bestswngs@gmail.com>

A locked RTAX_CWND route metric of zero is copied into the TCP metrics
cache and then into snd_cwnd_clamp. This sets the initial cwnd to zero.
With a cached SSTHRESH metric, a later Reno ACK can reach a divide by
zero in tcp_cong_avoid_ai(). A user with CAP_NET_ADMIN in a user-created
network namespace can trigger the oops.

Ignore a zero cached CWND metric when applying the locked clamp, leaving
the socket's initialized nonzero clamp in place.

  Oops: divide error: 0000 [#1] SMP KASAN NOPTI
  RIP: 0010:tcp_cong_avoid_ai (net/ipv4/tcp_cong.c:482)
  Call Trace:
  tcp_ack (net/ipv4/tcp_input.c:4440)
  tcp_rcv_established (net/ipv4/tcp_input.c:6678)
  tcp_v4_do_rcv (net/ipv4/tcp_ipv4.c:1854)
  tcp_v4_rcv (net/ipv4/tcp_ipv4.c:2248)
  Kernel panic - not syncing: Fatal exception in interrupt

Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local cache.")
Reported-by: <co+71b6509cb791e531@bugs.sh>
Assisted-by: LLM
Signed-off-by: Weiming Shi <bestswngs@gmail.com>
---
 net/ipv4/tcp_metrics.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c
index dc0c081fc1f3..ff41804ad086 100644
--- a/net/ipv4/tcp_metrics.c
+++ b/net/ipv4/tcp_metrics.c
@@ -484,8 +484,11 @@ void tcp_init_metrics(struct sock *sk)
 		goto reset;
 	}
 
-	if (tcp_metric_locked(tm, TCP_METRIC_CWND))
-		tp->snd_cwnd_clamp = tcp_metric_get(tm, TCP_METRIC_CWND);
+	if (tcp_metric_locked(tm, TCP_METRIC_CWND)) {
+		val = tcp_metric_get(tm, TCP_METRIC_CWND);
+		if (val)
+			tp->snd_cwnd_clamp = val;
+	}
 
 	val = READ_ONCE(net->ipv4.sysctl_tcp_no_ssthresh_metrics_save) ?
 	      0 : tcp_metric_get(tm, TCP_METRIC_SSTHRESH);
-- 
2.55.0


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH net] tcp: ignore locked zero CWND route metric
  2026-10-03 15:32 [PATCH net] tcp: ignore locked zero CWND route metric bestswngs
@ 2026-10-03 16:10 ` Eric Dumazet
  2026-10-03 17:09   ` Eric Dumazet
  0 siblings, 1 reply; 4+ messages in thread
From: Eric Dumazet @ 2026-10-03 16:10 UTC (permalink / raw)
  To: bestswngs, ncardwell, netdev
  Cc: kuniyu, davem, horms, xmei5, co+71b6509cb791e531



On 10/3/26 17:32, bestswngs@gmail.com wrote:
> From: Weiming Shi <bestswngs@gmail.com>
> 
> A locked RTAX_CWND route metric of zero is copied into the TCP metrics
> cache and then into snd_cwnd_clamp. This sets the initial cwnd to zero.
> With a cached SSTHRESH metric, a later Reno ACK can reach a divide by
> zero in tcp_cong_avoid_ai(). A user with CAP_NET_ADMIN in a user-created
> network namespace can trigger the oops.
> 
> Ignore a zero cached CWND metric when applying the locked clamp, leaving
> the socket's initialized nonzero clamp in place.
> 
>    Oops: divide error: 0000 [#1] SMP KASAN NOPTI
>    RIP: 0010:tcp_cong_avoid_ai (net/ipv4/tcp_cong.c:482)
>    Call Trace:
>    tcp_ack (net/ipv4/tcp_input.c:4440)
>    tcp_rcv_established (net/ipv4/tcp_input.c:6678)
>    tcp_v4_do_rcv (net/ipv4/tcp_ipv4.c:1854)
>    tcp_v4_rcv (net/ipv4/tcp_ipv4.c:2248)
>    Kernel panic - not syncing: Fatal exception in interrupt
> 
 > Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local 
cache.")> Reported-by: <co+71b6509cb791e531@bugs.sh>
> Assisted-by: LLM
> Signed-off-by: Weiming Shi <bestswngs@gmail.com>
> ---
>   net/ipv4/tcp_metrics.c | 7 +++++--
>   1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c
> index dc0c081fc1f3..ff41804ad086 100644
> --- a/net/ipv4/tcp_metrics.c
> +++ b/net/ipv4/tcp_metrics.c
> @@ -484,8 +484,11 @@ void tcp_init_metrics(struct sock *sk)
>   		goto reset;
>   	}
>   
> -	if (tcp_metric_locked(tm, TCP_METRIC_CWND))
> -		tp->snd_cwnd_clamp = tcp_metric_get(tm, TCP_METRIC_CWND);
> +	if (tcp_metric_locked(tm, TCP_METRIC_CWND)) {
> +		val = tcp_metric_get(tm, TCP_METRIC_CWND);

Please explain how we can come to this situation?

cwnd can never be zero, so how a save metric could contain zero?


> +		if (val)
> +			tp->snd_cwnd_clamp = val;
> +	}
>   
>   	val = READ_ONCE(net->ipv4.sysctl_tcp_no_ssthresh_metrics_save) ?
>   	      0 : tcp_metric_get(tm, TCP_METRIC_SSTHRESH);


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH net] tcp: ignore locked zero CWND route metric
  2026-10-03 16:10 ` Eric Dumazet
@ 2026-10-03 17:09   ` Eric Dumazet
  2026-10-04 16:51     ` Weiming Shi
  0 siblings, 1 reply; 4+ messages in thread
From: Eric Dumazet @ 2026-10-03 17:09 UTC (permalink / raw)
  To: bestswngs, ncardwell, netdev
  Cc: kuniyu, davem, horms, xmei5, co+71b6509cb791e531



On 10/3/26 18:10, Eric Dumazet wrote:
> 
> 
> On 10/3/26 17:32, bestswngs@gmail.com wrote:
>> From: Weiming Shi <bestswngs@gmail.com>
>>
>> A locked RTAX_CWND route metric of zero is copied into the TCP metrics
>> cache and then into snd_cwnd_clamp. This sets the initial cwnd to zero.
>> With a cached SSTHRESH metric, a later Reno ACK can reach a divide by
>> zero in tcp_cong_avoid_ai(). A user with CAP_NET_ADMIN in a user-created
>> network namespace can trigger the oops.
>>
>> Ignore a zero cached CWND metric when applying the locked clamp, leaving
>> the socket's initialized nonzero clamp in place.
>>
>>    Oops: divide error: 0000 [#1] SMP KASAN NOPTI
>>    RIP: 0010:tcp_cong_avoid_ai (net/ipv4/tcp_cong.c:482)
>>    Call Trace:
>>    tcp_ack (net/ipv4/tcp_input.c:4440)
>>    tcp_rcv_established (net/ipv4/tcp_input.c:6678)
>>    tcp_v4_do_rcv (net/ipv4/tcp_ipv4.c:1854)
>>    tcp_v4_rcv (net/ipv4/tcp_ipv4.c:2248)
>>    Kernel panic - not syncing: Fatal exception in interrupt
>>
>  > Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local 
> cache.")> Reported-by: <co+71b6509cb791e531@bugs.sh>
>> Assisted-by: LLM
>> Signed-off-by: Weiming Shi <bestswngs@gmail.com>
>> ---
>>   net/ipv4/tcp_metrics.c | 7 +++++--
>>   1 file changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c
>> index dc0c081fc1f3..ff41804ad086 100644
>> --- a/net/ipv4/tcp_metrics.c
>> +++ b/net/ipv4/tcp_metrics.c
>> @@ -484,8 +484,11 @@ void tcp_init_metrics(struct sock *sk)
>>           goto reset;
>>       }
>> -    if (tcp_metric_locked(tm, TCP_METRIC_CWND))
>> -        tp->snd_cwnd_clamp = tcp_metric_get(tm, TCP_METRIC_CWND);
>> +    if (tcp_metric_locked(tm, TCP_METRIC_CWND)) {
>> +        val = tcp_metric_get(tm, TCP_METRIC_CWND);
> 
> Please explain how we can come to this situation?
> 
> cwnd can never be zero, so how a save metric could contain zero?

That is to say :

The right fix is in ip_metrics_convert(), one time instead of at each 
TCP session init.

diff --git a/net/ipv4/metrics.c b/net/ipv4/metrics.c
index 
ad40762a8b38379e4ccdc505e3885adb8dfff763..50a08e592b207e85ccc2bb6f917379de61af4579 
100644
--- a/net/ipv4/metrics.c
+++ b/net/ipv4/metrics.c
@@ -54,6 +54,10 @@ static int ip_metrics_convert(struct nlattr *fc_mx,
                         NL_SET_ERR_MSG(extack, "Unknown flag set in 
feature mask in metrics attribute");
                         return -EINVAL;
                 }
+               if ((type == RTAX_CWND || type == RTAX_INITCWND) && !val) {
+                       NL_SET_ERR_MSG(extack, "CWND metric must be 
greater than zero");
+                       return -EINVAL;
+               }
                 metrics[type - 1] = val;
         }

pw-bot: cr


^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH net] tcp: ignore locked zero CWND route metric
  2026-10-03 17:09   ` Eric Dumazet
@ 2026-10-04 16:51     ` Weiming Shi
  0 siblings, 0 replies; 4+ messages in thread
From: Weiming Shi @ 2026-10-04 16:51 UTC (permalink / raw)
  To: Eric Dumazet
  Cc: ncardwell, netdev, kuniyu, davem, horms, xmei5,
	co+71b6509cb791e531

Eric Dumazet <edumazet@kernel.org> 于2026年10月4日周日 01:09写道:
>
>
>
> On 10/3/26 18:10, Eric Dumazet wrote:
> >
> >
> > On 10/3/26 17:32, bestswngs@gmail.com wrote:
> >> From: Weiming Shi <bestswngs@gmail.com>
> >>
> >> A locked RTAX_CWND route metric of zero is copied into the TCP metrics
> >> cache and then into snd_cwnd_clamp. This sets the initial cwnd to zero.
> >> With a cached SSTHRESH metric, a later Reno ACK can reach a divide by
> >> zero in tcp_cong_avoid_ai(). A user with CAP_NET_ADMIN in a user-created
> >> network namespace can trigger the oops.
> >>
> >> Ignore a zero cached CWND metric when applying the locked clamp, leaving
> >> the socket's initialized nonzero clamp in place.
> >>
> >>    Oops: divide error: 0000 [#1] SMP KASAN NOPTI
> >>    RIP: 0010:tcp_cong_avoid_ai (net/ipv4/tcp_cong.c:482)
> >>    Call Trace:
> >>    tcp_ack (net/ipv4/tcp_input.c:4440)
> >>    tcp_rcv_established (net/ipv4/tcp_input.c:6678)
> >>    tcp_v4_do_rcv (net/ipv4/tcp_ipv4.c:1854)
> >>    tcp_v4_rcv (net/ipv4/tcp_ipv4.c:2248)
> >>    Kernel panic - not syncing: Fatal exception in interrupt
> >>
> >  > Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local
> > cache.")> Reported-by: <co+71b6509cb791e531@bugs.sh>
> >> Assisted-by: LLM
> >> Signed-off-by: Weiming Shi <bestswngs@gmail.com>
> >> ---
> >>   net/ipv4/tcp_metrics.c | 7 +++++--
> >>   1 file changed, 5 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/net/ipv4/tcp_metrics.c b/net/ipv4/tcp_metrics.c
> >> index dc0c081fc1f3..ff41804ad086 100644
> >> --- a/net/ipv4/tcp_metrics.c
> >> +++ b/net/ipv4/tcp_metrics.c
> >> @@ -484,8 +484,11 @@ void tcp_init_metrics(struct sock *sk)
> >>           goto reset;
> >>       }
> >> -    if (tcp_metric_locked(tm, TCP_METRIC_CWND))
> >> -        tp->snd_cwnd_clamp = tcp_metric_get(tm, TCP_METRIC_CWND);
> >> +    if (tcp_metric_locked(tm, TCP_METRIC_CWND)) {
> >> +        val = tcp_metric_get(tm, TCP_METRIC_CWND);
> >
> > Please explain how we can come to this situation?
> >
> > cwnd can never be zero, so how a save metric could contain zero?
>
> That is to say :
>
> The right fix is in ip_metrics_convert(), one time instead of at each
> TCP session init.
>
> diff --git a/net/ipv4/metrics.c b/net/ipv4/metrics.c
> index
> ad40762a8b38379e4ccdc505e3885adb8dfff763..50a08e592b207e85ccc2bb6f917379de61af4579
> 100644
> --- a/net/ipv4/metrics.c
> +++ b/net/ipv4/metrics.c
> @@ -54,6 +54,10 @@ static int ip_metrics_convert(struct nlattr *fc_mx,
>                          NL_SET_ERR_MSG(extack, "Unknown flag set in
> feature mask in metrics attribute");
>                          return -EINVAL;
>                  }
> +               if ((type == RTAX_CWND || type == RTAX_INITCWND) && !val) {
> +                       NL_SET_ERR_MSG(extack, "CWND metric must be
> greater than zero");
> +                       return -EINVAL;
> +               }
>                  metrics[type - 1] = val;
>          }
>
> pw-bot: cr
>

Thanks for the review. v2 sent.

Best regards,
Weiming Shi

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2026-10-04 16:51 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-10-03 15:32 [PATCH net] tcp: ignore locked zero CWND route metric bestswngs
2026-10-03 16:10 ` Eric Dumazet
2026-10-03 17:09   ` Eric Dumazet
2026-10-04 16:51     ` Weiming Shi

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox