* [TEST] gre_multipath_nh_res takes forever
@ 2024-02-15 17:50 Jakub Kicinski
2024-02-19 16:50 ` Ido Schimmel
0 siblings, 1 reply; 3+ messages in thread
From: Jakub Kicinski @ 2024-02-15 17:50 UTC (permalink / raw)
To: Ido Schimmel, Petr Machata; +Cc: netdev@vger.kernel.org
Hi!
I bumped the timeouts for forwarding tests with kernel debug enabled
yesterday, and gre_multipath_nh_res still doesn't finish even tho it
runs for close to 3 hours. On a non-debug kernel it takes around 30min.
You probably have a better idea how to address this than me, but it
seems to time out on the IPv6 tests - I wonder if using mausezahn
instead of ping for IPv6 would help a bit?
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [TEST] gre_multipath_nh_res takes forever
2024-02-15 17:50 [TEST] gre_multipath_nh_res takes forever Jakub Kicinski
@ 2024-02-19 16:50 ` Ido Schimmel
2024-02-19 19:17 ` Jakub Kicinski
0 siblings, 1 reply; 3+ messages in thread
From: Ido Schimmel @ 2024-02-19 16:50 UTC (permalink / raw)
To: Jakub Kicinski; +Cc: Petr Machata, netdev@vger.kernel.org
Hi,
On Thu, Feb 15, 2024 at 09:50:05AM -0800, Jakub Kicinski wrote:
> Hi!
>
> I bumped the timeouts for forwarding tests with kernel debug enabled
> yesterday, and gre_multipath_nh_res still doesn't finish even tho it
> runs for close to 3 hours. On a non-debug kernel it takes around 30min.
>
> You probably have a better idea how to address this than me, but it
> seems to time out on the IPv6 tests - I wonder if using mausezahn
> instead of ping for IPv6 would help a bit?
Converted from ping6 to mausezahn, but the test was still relatively
slow because of the 1msec delay we pass the mausezahn. After converting
to mausezahn and removing the delay the runtime dropped from ~350
seconds to ~15 seconds on my system.
I think I will parametrize the delay for platforms that might need it,
but default it to 0 as it's not needed with veth. The same change is
needed in other tests. Will submit it together with Petr's selftest
changes.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [TEST] gre_multipath_nh_res takes forever
2024-02-19 16:50 ` Ido Schimmel
@ 2024-02-19 19:17 ` Jakub Kicinski
0 siblings, 0 replies; 3+ messages in thread
From: Jakub Kicinski @ 2024-02-19 19:17 UTC (permalink / raw)
To: Ido Schimmel; +Cc: Petr Machata, netdev@vger.kernel.org
On Mon, 19 Feb 2024 18:50:58 +0200 Ido Schimmel wrote:
> > I bumped the timeouts for forwarding tests with kernel debug enabled
> > yesterday, and gre_multipath_nh_res still doesn't finish even tho it
> > runs for close to 3 hours. On a non-debug kernel it takes around 30min.
> >
> > You probably have a better idea how to address this than me, but it
> > seems to time out on the IPv6 tests - I wonder if using mausezahn
> > instead of ping for IPv6 would help a bit?
>
> Converted from ping6 to mausezahn, but the test was still relatively
> slow because of the 1msec delay we pass the mausezahn. After converting
> to mausezahn and removing the delay the runtime dropped from ~350
> seconds to ~15 seconds on my system.
Great!
> I think I will parametrize the delay for platforms that might need it,
> but default it to 0 as it's not needed with veth. The same change is
> needed in other tests. Will submit it together with Petr's selftest
> changes.
SG, thanks!
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2024-02-19 19:17 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-02-15 17:50 [TEST] gre_multipath_nh_res takes forever Jakub Kicinski
2024-02-19 16:50 ` Ido Schimmel
2024-02-19 19:17 ` Jakub Kicinski
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).