13:02:06 <jki> #startmeeting CIP IRC weekly meeting
13:02:06 <collab-meetbot`> Meeting started Thu Jan 19 13:02:06 2023 UTC and is due to finish in 60 minutes.  The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:02:06 <collab-meetbot`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:02:06 <collab-meetbot`> The meeting name has been set to 'cip_irc_weekly_meeting'
13:02:14 <jki> #topic AI review
13:02:28 <jki> I don't have any on the list
13:02:48 <jki> anything missed?
13:03:15 <pave1> I guess empty list is a good thing :-).
13:03:33 <jki> possibly... :)
13:03:44 <jki> moving on in
13:03:45 <jki> 3
13:03:48 <jki> 2
13:03:49 <jki> 1
13:03:52 <jki> #topic Kernel maintenance updates
13:04:12 <uli> i backported the remaining 4.9 patches to 4.4
13:04:26 <masami> This week reported 7 new CVEs and 8 updated CVEs.
13:04:27 <pave1> I did reviews on 5.10.163 and 164.
13:04:39 <iwamatsu> I reviewed 5.10.193 and 194.
13:04:48 <pave1> 4.19-rt is released, 5.10-rt hit bug in upstream.
13:05:50 <jki> 5.10.193??
13:06:13 <pave1> That looks like typo -- 6 vs. 9.
13:06:15 <jki> 163, I suppose :)
13:07:09 <jki> anything else?
13:07:19 <iwamatsu> sorry, 163 and 164.
13:07:56 <jki> 3
13:07:57 <jki> 2
13:08:00 <jki> 1
13:08:03 <jki> #topic Kernel release status
13:08:12 <jki> -4.4
13:08:28 <uli> pavel did the reviews, everything seems ok. i think it's ready to release.
13:08:46 <pave1> Yep, two more patches to look at, but lets not hold a release for that.
13:09:27 <jki> good. -rt would be able to follow?
13:09:52 <pave1> I guess the question is "who does the rt"? :-)
13:10:45 <jki> uli: would you feel ready for that?
13:11:25 <uli> i haven't looked at it at all yet...
13:11:33 <uli> next time, maybe?
13:11:49 <jki> pavel: would you do it again, giving uli some intro?
13:13:30 <pave1> jki: Makes sense. There is still series to be backported from 4.9-rt and we'll actually have to start watching 4.14-rt.
13:14:32 <jki> yep
13:15:07 <jki> were those trees rather silent rt-wise by now, only integration with stable?
13:15:52 <pave1> There were silent for a year, but there are some changes now (hence the backport). It seems to be "just nice to have" and not critical.
13:16:05 <jki> ok
13:16:12 <jki> -4.19
13:16:26 <jki> I think we are up-to-date here, right?
13:16:38 <pave1> I think so.
13:16:42 <jki> -5.10
13:17:00 <jki> -rt release in sight?
13:17:11 <pave1> I hit a bug in mainline -rt.
13:17:19 <jki> uh
13:17:34 <pave1> It is currently being discussed, I'm not only one hitting it.
13:17:56 <jki> ok, then we have a good reason to wait
13:18:10 <jki> anything else?
13:18:18 <jki> 3
13:18:20 <jki> 2
13:18:21 <pave1> It is assembly on arm64, not too evil but could not figure it out in 5 minutes, either.
13:18:34 <alicef> 4.19 we are missing the cip release as yesterday we got update from kernel upstream. I think iwamatsu san is currently working on it.
13:19:36 <jki> indeed!
13:19:38 <jki> ok
13:19:51 <iwamatsu> Yes, I am working about it. I will relelase 4.19.y-cip tomorrow.
13:20:01 <alicef> iwamatsu: thanks
13:21:07 <jki> good - then move on
13:21:10 <jki> #topic Kernel testing
13:21:35 <alicef> sent PR to kernelCI for updating the isar-cip-core
13:22:49 <patersonc[m]> just a couple of things from me
13:22:49 <patersonc[m]> 1) As I reported in the TSC, 5.10 gitlab CI builds should now include the riscv and qemu-riscv defconfigs
13:22:49 <patersonc[m]> 2) It sounds like we still need a discussion on testing _all_ of the stable kernel branches
13:23:11 <alicef> and writing PR for using more cip configs on KernelCI other than 4.19
13:23:25 <pave1> patersonc: Yep. I believe we should start testing 6.1.X -- that's no brainer.
13:23:31 <patersonc[m]> Thank you alicef
13:24:16 <pave1> Testing all the stables would not hurt, either, but in that case I'd like you to take over writing all the "no bugs detected" here to Greg ;-).
13:24:34 <pave1> I still plan to watch the results and debug failures, etc.
13:25:39 <patersonc[m]> I'm for it
13:25:47 <pave1> Ok, job is yours :-).
13:25:51 <jki> I also believe that running them all is not the issue, processing them is
13:26:15 <pave1> But maybe more importantly...
13:26:47 <pave1> It looks like upstream -rt trees are not tested adequately.
13:27:25 <pave1> Could we start testing them, too? In -stable testing, we won't make much difference but here we would
13:27:59 <jki> just throwing the same tests as for vanilla on them?
13:28:45 <pave1> jki: That would be a very good start.
13:28:56 <patersonc[m]> We can start testing -rt stable
13:29:29 <pave1> That would be great.
13:29:55 <jki> then let's go - but if we are multiplying CI times, just make sure to warn the TSC about the costs ;)
13:29:56 <patersonc[m]> Okay
13:30:28 <jki> or can we also leverage kernelci for that?
13:31:12 <patersonc[m]> Potentially
13:31:20 <patersonc[m]> Maybe they are already testing -rt I need to check
13:31:37 <patersonc[m]> If they are, then it's a question of whether we also want to test it with our set of configs
13:31:44 <jki> would be good to know, to foxus on the real gaps
13:32:01 <pave1> I'd suggest testing 6.1.X and -rt on our infrastructure.
13:32:25 <pave1> We already know there's a gap in 5.10-rt testing: it did not build on arm64 at all.
13:32:35 <patersonc[m]> https://linux.kernelci.org/job/rt-stable/
13:32:41 <patersonc[m]> So yes, they already test -rt stable
13:32:52 <pave1> ...and it took 3 days for person from TI to notice.
13:33:18 <patersonc[m]> But whether the RT project monitor the build/test reports is another thing
13:33:27 <jki> evaluating reports is the bottleneck again...
13:33:42 <patersonc[m]> 5.10 is reported to fail for Arm64: https://linux.kernelci.org/build/rt-stable/branch/v5.10-rt/kernel/v5.10.162-rt78/
13:34:02 <patersonc[m]> So usual story, lots of people building & testing things, not so many looking at the results :P
13:34:06 <pave1> jki: Yep, kernelci got that, but nobody noticed. It is the same fail: https://linux.kernelci.org/build/id/63c57fa09c4841bd961d39cd/logs/
13:35:08 <jki> at least the respective -rt stable maintainer should cross-check, I would say
13:35:35 <jki> or should be subscribed
13:35:40 <pave1> That's not the way it works, I'm afraid.
13:36:04 <patersonc[m]> Does the RT project have some sort of RC release?
13:37:01 <pave1> patersonc: I'm sure they did _some_ kind of -rc at least occassionaly. I'm not sure if they do it for every release.
13:37:15 <patersonc[m]> Okay
13:37:32 <patersonc[m]> Obviously that would be the best thing to test, but it's not so useful if it's not always done
13:38:14 <pave1> There are differnent people maintaining different stable-rt branches, so their workflows may differ.
13:38:59 <jki> then we may have to help with reporting - once again
13:39:35 <pave1> Yes, that would be a good thing to do.
13:42:31 <jki> I'll check internally what our group may contribute here - was on my mid-term plan anyway
13:43:38 <jki> anything else on testing?
13:44:00 <patersonc[m]> Not from me
13:44:24 <jki> who will take the AI for enbling more stable trees in our lab now?
13:45:19 <patersonc[m]> I will
13:45:29 <jki> perfect, thanks!
13:45:37 <jki> then moving on....
13:45:38 <jki> 3
13:45:40 <pave1> I believe we'll also need configs.
13:45:40 <jki> 2
13:46:05 <pave1> ...but using configs for older version we are already testing should work well enough for a start.
13:46:09 <patersonc[m]> pave1: ah good point
13:46:19 <jki> likely
13:46:55 <jki> but having a look at the resulting diff of the .configs would not be bad
13:48:06 <jki> now moving on in...
13:48:09 <jki> 3
13:48:11 <jki> 2
13:48:13 <jki> 1
13:48:15 <jki> #topic AOB
13:48:44 <jki> I'm on vacation next Thursday
13:48:54 <jki> would need a substitute
13:49:15 <pave1> I can do it I guess.
13:49:25 <jki> thanks!
13:50:08 <jki> anything else?
13:50:53 <jki> 3
13:50:55 <jki> 2
13:50:58 <jki> 1
13:51:00 <jki> #endmeeting