13:01:10 <jki> #startmeeting CIP IRC weekly meeting
13:01:10 <brlogger> Meeting started Thu Dec  2 13:01:10 2021 UTC and is due to finish in 60 minutes.  The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:01:10 <brlogger> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:01:10 <brlogger> The meeting name has been set to 'cip_irc_weekly_meeting'
13:01:14 <jki> Hi everyone!
13:01:19 <pave1> hi!
13:01:20 <uli> hello
13:01:23 <masami> hi
13:01:45 <patersonc[m]> ay up
13:01:50 <iwamatsu> hi
13:02:28 <alicefm> Hi
13:04:06 <jki> #topic AI review
13:04:14 <jki> 1. Combine root filesystem with kselftest binary (finishings) - iwamatsu & alicef
13:04:41 <alicef> kernelci is currently testing cip with kselftest
13:05:16 <alicef> I also replayed to pave1 mail about the current regressions
13:05:49 <alicef> we are currently adding the kernelci automatic bisection feature to the CIP kernel tree
13:06:30 <alicef> We should start to get cip kernel bisection mail in the future
13:06:43 <alicef> anything to add ? or question ?
13:06:53 <jki> what's missing to close this specific AI (and maybe start new ones)?
13:07:20 <jki> isar-cip-core merge at least, anything else?
13:07:53 <alicef> nothing other than isar-cip-core merge I think
13:08:31 <iwamatsu> alicef: can anyone receive the build result email? is there an I/F to reguster?
13:09:01 <alicef> iwamatsu: you mean bisection email ?
13:09:22 <iwamatsu> test result for CIP kernel
13:09:23 <alicef> all mail concerning cip kernel are already sended to cip-dev mailing list
13:10:00 <pave1> alicef: Thanks for the email reply.
13:10:12 <pave1> alicef: If we can get the automatic bisection to work, yes, that will be useful.
13:10:25 <alicef> from what I could see we are not missing any mail from cip-dev
13:10:57 <pave1> alicef: Best way would be to bisect between v4.19.217 and v4.19.217-cip62, if that can be arranged.
13:11:01 <iwamatsu> Can I receive it with other email address. e.g. personal email address.
13:11:38 <pave1> alicef: Plus, for this to be useful, we'd really need some way to submit patches/trees for testing.
13:11:49 <alicef> iwamatsu: is not something that as been done. usually kernelci report to kernelci ml and cip-dev ml
13:12:22 <patersonc[m]> We can set it up to report to other emails if we want, just need to submit an MR to kernelci
13:12:23 <iwamatsu> alicef: I see.
13:12:24 <alicef> pave1: no currently is not something that is supported by kernelci to do bisect between completely different tree
13:12:27 <iwamatsu> got it
13:12:41 <patersonc[m]> pave1: We could add another tree to Kernelci. Then you push to that tree to test etc.
13:13:25 <pave1> patersonc: That would be useful. ONe really needs a way to verify that patch fixes the issue, etc.
13:13:30 <alicef> pave1: you could have a testing branch on cip
13:13:45 <alicef> I don't think having another tree is needed
13:14:01 <alicef> just some branch like 4.19-cip-test
13:14:09 <alicef> is enough
13:14:24 <pave1> patersonc,alicef: Branch works, too. BUt it is going to be cluttered with test commits by different people.
13:14:41 <pave1> patersonc,alicef: How long will it take between push and getting results?
13:14:49 <alicef> sure. whatever works for you
13:15:02 <pave1> patersonc,alicef: Plus I guess that will cause quite a lot of noise on the lists :-(.
13:15:59 <alicef> pave1: it can take some time to get results. not sure how much time is not immediate afaik
13:16:34 <pave1> alicef: Is it more like hour, day or a week?
13:16:57 <alicef> pave1: depends we could add a job on jenkins only for your tree where you get your results from staging.
13:17:16 <alicef> pave1: need to be discussed with kernelci TSC
13:17:33 <alicef> depend from the resources TSC have for such thing
13:18:42 <alicef> but if you are just trying to reproduce tests you could also rebuild on your machine, push to a ftp server and send lava jobs with such artifacts
13:19:25 <alicef> depend where you want to test such commits
13:19:46 <alicef> if is just a board or if you want to test with a bunch of board
13:19:59 <pave1> alicef: Local builds don't really work too well.
13:20:27 <alicef> having a personal tree for pushing test to kernelci is possible
13:20:47 <pave1> alicef: yep, that would be nice.
13:20:48 <alicef> pave1: why not ? you are using the same enviroment anyway
13:21:58 <pave1> alicef: Maintaining all the cross-compilers (etc) is quite a pain, plus you need fast CPU and fast network for that.
13:22:31 <alicef> ok
13:23:42 <alicef> pave1: do you already have a tree for testing that you want tested ?
13:24:10 <pave1> alicef: I do have https://gitlab.com/cip-project/cip-kernel/linux-cip/tree/ci/pavel/linux-test .
13:24:29 <pave1> alicef: I push random kernels there (4.4, 4.19, 5.10, -realtime versions) when I need something tested.
13:24:41 <alicef> you can add it here https://github.com/kernelci/kernelci-core/blob/eae1f11042ae21383178595a675adfee0dac2412/config/core/build-configs-cip.yaml#L1
13:25:20 <pave1> Ok, perhaps we can discuss this further over email?
13:25:41 <patersonc[m]> https://github.com/kernelci/kernelci-core/issues/new?assignees=&labels=&template=new-kernel-branch.md&title=Add+branch+BRANCH+from+TREE
13:25:52 <alicef> sure
13:26:14 <alicef> patersonc[m]: that is not for cip branch, i suppose
13:26:40 <patersonc[m]> I think it could be. You specify the tree as cip
13:26:49 <alicef> ok
13:27:06 <alicef> and you can call your repo like pave1:
13:27:20 <alicef> and add some test to it
13:27:33 <alicef> some other kernelci dev is doing it
13:28:10 <alicef> we can also specify to send the email only to you from the jenkin task
13:29:09 <alicef> iwamatsu: we can send mail from specific people from changing the jenkin task, but not sure about the main cip tree. I can look into it
13:29:58 <iwamatsu> alicef: got it
13:30:17 <alicef> do you need to get cip tree mail to some personal mail ? if yes you can write it on private and I can look into adding it
13:31:48 <iwamatsu> answer is yes. company emails are limited...
13:31:59 <alicef> ok
13:33:16 <jki> ok, done with this topic?
13:33:41 <jki> alicef: please send a new patch series for isar-cip-core - or let me know if I can help
13:33:48 <alicef> iwamatsu: send me a mail and i will send a pull request to jenkins for adding your email
13:34:34 <alicef> jki: ok
13:35:01 <iwamatsu> alicef: thanks
13:35:20 <alicef> iwamatsu: np, sorry for the confusion
13:36:13 <jki> 2. Look into S3 artifact upload issues - patersonc
13:36:18 <jki> done, right?
13:36:33 <patersonc[m]> Done!
13:36:45 <jki> thanks!
13:36:47 <patersonc[m]> Issue is now fixed
13:36:52 <patersonc[m]> Although please let me know if you see any more issues
13:36:53 <iwamatsu> Yey
13:37:07 <pave1> Yep, will do.
13:37:36 <jki> 3. Perform initial comparison of KernelCI results 5.10 LTS vs. CIP - iwamatsu & alicef
13:38:43 <alicef> I did some comparison about 4.19 and sended by mail. not yet worked at 5.10 comparison
13:38:45 <iwamatsu> Not yet.
13:39:06 <jki> ah, sorry, maybe my mistake what was planned
13:39:21 <jki> or was it 5.10?
13:39:40 <iwamatsu> maybe all CIP kernel.
13:39:45 <alicef> I thought was a general comparison about lts-lts+1 and cip-cip+1
13:40:30 <alicef> we want this to be automated ?
13:40:49 <iwamatsu> I think so.
13:41:24 <alicef> here is comparison with 4.19.217 and 4.19.217-cip62 https://www.diffchecker.com/JjmXiP7b
13:41:31 <alicef> as example
13:41:44 <iwamatsu> about 5.10.y, we may need to proceed with the release of the 5.10.y CIP kernel..
13:41:51 <alicef> that i sended today to pavel request
13:42:18 <alicef> yes about 5.10 we still don't have a release.
13:42:55 <alicef> we should release before doing comparison or i can just compare with the last commit
13:43:14 <iwamatsu> is that diff email is automatic?
13:43:29 <iwamatsu> I see.
13:43:35 <alicef> not yet automatic but is possible to implement it to kernelci
13:43:42 <iwamatsu> nice
13:43:59 <iwamatsu> ok, lets discues email about this.
13:44:23 <alicef> but is not straight forward it still need some function for make it work and probably also need to be talked with kernelci tsc
13:44:37 <iwamatsu> I see
13:44:57 <alicef> I could start by opening a issue to kernelci and see others opinions
13:46:07 <alicef> currently I'm doing it manually by searching each email result
13:46:41 <alicef> and because the result format is standard they can be easily diffed
13:47:48 <alicef> I'm not sure if kernelci is keeping a list of previous sended email for comparison purpose
13:47:55 <alicef> I don't think so
13:48:05 <alicef> that could be a problem
13:48:50 <alicef> I need to investigate on how to automate this
13:49:23 <alicef> iwamatsu: any date on the 5.10 release ?
13:49:47 <jki> so, alicef will discuss this topic with kernelci, iwamatsu will first release 5.10 and wait for alicef's results?
13:50:26 <alicef> work for me
13:50:31 <iwamatsu> not decide yet > release
13:51:12 <iwamatsu> I do not wait for alice's work.
13:51:23 <jki> ok
13:51:44 <alicef> ok
13:52:04 <jki> maybe discuss 5.10 release separately - anything else for this comparison topic?
13:52:48 <alicef> something like the pasted diffchecker is enough as comparison format ?
13:53:01 <alicef> or have you in mind something different ?
13:53:43 <alicef> if such reports could be automated and sent by email would be enough ?
13:54:07 <alicef> for closing this AI
13:54:35 <iwamatsu> First of all, I think diff is fine.
13:55:56 <jki> okay
13:56:11 <jki> 4. Propose tweet on KernelCI-CIP collaboration progress - alicef
13:56:41 <jki> alicef: I think you asked for this :)
13:57:00 <alicef> my idea, as currently we have a CIP panel project on KernelCI, was to tweet about this event.
13:57:39 <alicef> https://github.com/orgs/kernelci/projects/11
13:57:56 <alicef> that show the status of project testing on kernelci
13:58:25 <alicef> also as we have enabled kselftest testing on kernelci, maybe tweet also about this
13:58:44 <jki> yes, do you have some text to report on that? we are happy to help tweaking it if needed
13:58:51 <alicef> cip kselftest testing on kernelci
13:59:22 <alicef> mmm
13:59:43 <jki> need not be now, of course
13:59:47 <alicef> we need to think about the text to report
14:00:00 <alicef> I don't have one now
14:00:52 <jki> send some draft to chris and me, and then we can continue and ask neal to tweet
14:00:59 <alicef> ok
14:01:10 <jki> #topic Kernel maintenance updates
14:01:12 <alicef> thanks
14:02:00 <pave1> I am reviewing 5.10.83... and did 4.19-rt release.
14:02:07 <uli> reviewing 5.10.80
14:02:15 <masami> There is 5 new CVEs, and two of them are not fixed in the mainline yet.
14:02:22 <iwamatsu> I am reviewing 5.10.83
14:02:23 <masami> Patch to fix CVE-2021-4001 for stable/5.10 which sent last week, was merged.
14:02:51 <pave1> I should do 4.4-rt release, soon, I'm waiting -rt team to release their kernel, and it would make sense to coordinate 4.4-cip with that.
14:03:41 <iwamatsu> And I am creating a review list creating/pushing tool
14:04:44 <iwamatsu> pave1: I can ask ahbout this to RT project.
14:05:17 <jki> ok, anything else?
14:05:26 <pave1> iwamatsu: That may make sense if they don't release it 'soon'.
14:05:41 <iwamatsu> s/ahbout this/4.4-rt release/g
14:06:06 <iwamatsu> pava1: ok
14:07:11 <jki> what is missing / what are conditions for a 5.10 release?
14:07:34 <pave1> I believe we are ready for 5.10 release.
14:08:03 <jki> head filled with latest or desired LTS version?
14:08:34 <iwamatsu> I am going to work this tomorrow.
14:08:36 <iwamatsu> yes.
14:09:12 <iwamatsu> I am updating to latset LTS.
14:09:33 <jki> cool!
14:10:01 <jki> anything else?
14:10:14 <alicefm> I can force a check from KernelCI on 5.10 current status if is needed
14:12:09 <jki> #topic Kernel testing
14:12:24 <jki> nothing more here, I suppose
14:12:32 <patersonc[m]> a bit
14:12:41 <jki> ok, go ahead
14:13:13 <patersonc[m]> I've started using an EU based S3 bucket for storing our CI artifacts. This speeds up our uploading/downloading which is good
14:13:26 <alicefm> Nice
14:13:37 <patersonc[m]> Currently I'm working on using the "latest" isar-cip-core image for our Kernel testing.
14:13:50 <patersonc[m]> jki: I've got a question about that, perhaps for later if you're available
14:15:23 <jki> yeah, we can do that
14:16:41 <jki> #topic AOB
14:16:45 <jki> anything?
14:18:31 <jki> 3
14:18:32 <jki> 2
14:18:34 <jki> 1
14:18:39 <jki> #endmeeting