TCP-AO got a bit more flaky
Subsystems (2)
Jakub Kicinski
maintainerHi Dmitry,
the changes in the timer subsystem Linus pulled for 6.9 made some of the TCP AO tests significantly more flaky.
We pulled them in on March 12th, the failures prior to that are legit:
https://netdev.bots.linux.dev/flakes.html?br-cnt=184&min-flip=0&pw-n=n&tn-needle=tcp-ao
PTAL whenever you have some spare cycles.
Hi Jakub,
Quoted 10 lines · Jakub Kicinski
Hi Dmitry,
the changes in the timer subsystem Linus pulled for 6.9 made some of the TCP AO tests significantly more flaky.
We pulled them in on March 12th, the failures prior to that are legit:
https://netdev.bots.linux.dev/flakes.html?br-cnt=184&min-flip=0&pw-n=n&tn-needle=tcp-aoPTAL whenever you have some spare cycles.
Certainly, will do this week, thanks for pinging!
Signature · 1 line
Dmitry
Jakub Kicinski
maintainerQuoted 9 lines · On Mon, 25 Mar 2024 14:16:04 +0000 Dmitry Safonov wrote:
Quoted 7 lines
the changes in the timer subsystem Linus pulled for 6.9 made some of the TCP AO tests significantly more flaky.
We pulled them in on March 12th, the failures prior to that are legit:
https://netdev.bots.linux.dev/flakes.html?br-cnt=184&min-flip=0&pw-n=n&tn-needle=tcp-aoPTAL whenever you have some spare cycles.
Certainly, will do this week, thanks for pinging!
Hi Dmitry! Do you have any spare cycles to spend on this? It's the main source of noise for us. It's not a huge deal but if you're busy I'd like to disable the rst-ipv* tests, at least.
Quoted 8 lines · Jakub Kicinski
Quoted 6 lines · On Mon, 25 Mar 2024 14:16:04 +0000 Dmitry Safonov wrote:
Quoted 6 lines
the changes in the timer subsystem Linus pulled for 6.9 made some of the TCP AO tests significantly more flaky.
We pulled them in on March 12th, the failures prior to that are legit:
https://netdev.bots.linux.dev/flakes.html?br-cnt=184&min-flip=0&pw-n=n&tn-needle=tcp-ao
Quoted 7 lines
Quoted 3 lines
Quoted 1 line
PTAL whenever you have some spare cycles.
Certainly, will do this week, thanks for pinging!
Hi Dmitry! Do you have any spare cycles to spend on this? It's the main source of noise for us. It's not a huge deal but if you're busy I'd like to disable the rst-ipv* tests, at least.
Hi Jakub, thanks for pinging.
I have a patch that addresses one of the issues I saw on these tests and currently working on another patch. I suggest we disable rst-ipv* tests on Monday if I don't finish the patch set by Monday.
Thanks and excuses for the delay: my work laptop was running linux and that was not fitting into corporate policy of having enterprise spy software. So, I had to urgently find a new machine for contributing and sending patches.
Signature · 1 line
Dmitry