13:01:18 <jki> #startmeeting CIP IRC weekly meeting
13:01:18 <brlogger> Meeting started Thu Oct 21 13:01:18 2021 UTC and is due to finish in 60 minutes.  The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:01:18 <brlogger> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:01:18 <brlogger> The meeting name has been set to 'cip_irc_weekly_meeting'
13:01:31 <jki> Hi everyone!
13:01:42 <jki> please let us know if you are around
13:01:48 <patersonc[m]> Hello
13:01:49 <iwamatsu> hi
13:01:50 <josiah|2> Hi
13:01:53 <masami> hi
13:01:54 <uli> hello
13:02:00 <pave1> hi
13:02:11 <alicef> o/
13:02:53 <jki> great, let's get started
13:03:00 <jki> #topic AI review
13:03:06 <jki> 1. Combine root filesystem with kselftest binary - iwamatsu & alicef
13:03:25 <alicef> I went on on the CIP core work
13:03:57 <alicef> we merged the pull request for using gz compressed rootfs on KernelCI and started testing the isar-cip-core images.
13:04:00 <alicef> isar-cip-core looks booting up from KernelCI but currently we cannot login as there is a root password
13:04:02 <alicef> https://lava.ciplatform.org/scheduler/job/482078#L809
13:04:04 <alicef> Would be useful for KernelCI to have the isar-cip-core password removed as the other rootfs are doing.
13:04:43 <alicef> jki: any opinion on this ?
13:04:46 <jki> that should be doable, though I don't have a pattern at hand
13:05:10 <alicef> is ok to remove the password from the isar-cip-core?
13:05:33 <jki> we could do that at least for the testing images being built
13:05:57 <alicef> ok nice. I will try to work on that
13:06:02 <jki> the "how" is not at hand for me, but we will find a solution
13:06:13 <alicef> I also updated the storage fileserver script for making a latest directory (even if probably will not used on KernelCI)
13:06:16 <alicef> and updated the isar-cip-core images  https://storage.staging.kernelci.org/images/rootfs/cip/
13:06:21 <jki> would be an Isar topic (if not documented there) -> isar-users ML
13:07:50 <jki> alicef: great progress, thanks a lot!
13:08:05 <alicef> lastly not related to cip core, I also started adding missing devices from https://lava.ciplatform.org/scheduler/device_types that are still not on KernelCI (currently testing it)
13:09:25 <alicef> that's all from me
13:09:40 <jki> 2. Look into S3 artifact upload issues - patersonc
13:09:51 <patersonc[m]> No update
13:10:21 <jki> ok - any new AIs?
13:10:46 <jki> 3
13:10:49 <jki> 2
13:10:51 <jki> 1
13:10:54 <jki> #topic Kernel maintenance updates
13:11:19 <pave1> I have reviewed patches for 5.10.74, still reviewing 5.10.75.
13:11:29 <uli> pushed 5.10.70/71 reviews, now reviewing 5.10.75
13:11:42 <masami> I sent CVE entry report. There is 7 new CVEs are reported in this week.
13:11:51 <iwamatsu> I reviewed 5.10.74.
13:11:51 <masami> I'll send backport patch of CVE-2021-20321 for 4.4.
13:14:08 <jki> regarding maintenance efforts: are we still on track with past estimations, roughly?
13:14:53 <jki> there were some concerns in the board, though I communicated that esitmations cannot be "exact", specifically here
13:15:50 <pave1> I see maybe +10% effort compared to previous year.
13:16:31 <pave1> It may decrease as 5.10 is getting older (and less active).
13:16:40 <jki> are there specific reasons for that, like +X% patch rate, release rate etc.?
13:17:07 <jki> right, so that natural curve of stable kernels "calmind down"
13:17:08 <pave1> I expect increase when we enter self-maintainence mode for 4.4
13:17:37 <pave1> Yes. And last year I was reviewing "calm" 4.19, and now I'm reviewing "hot" 5.10 (and 4.19 too).
13:19:05 <jki> understood
13:19:49 <patersonc[m]> o/
13:20:08 <patersonc[m]> When does CIP intend to start cip specific tags of v5.10?
13:20:24 <patersonc[m]> When some non-LTS patches are added?
13:20:51 <iwamatsu> maybe it will be decide next long TSC meeting.
13:21:42 <patersonc[m]> okay
13:21:59 <jki> yes, that would be good point to discuss
13:22:26 <patersonc[m]> I guess part of it is whether the board thinks we can afford it
13:22:34 <pave1> I believe we are ready to do that.
13:22:41 <jki> an important aspect is that we as WG group can make it clear that we can handle the additional effort with available resource
13:22:46 <pave1> We are doing reviews anyway, which are the expensive part.
13:22:55 <jki> that is an important point
13:23:11 <pave1> I believe we even checked that all patches in 4.19 tree are in 5.10 tree, too.
13:23:56 <jki> in addition, there was and likely still is the expectation of having 3 maintainers working on the kernels
13:24:28 <jki> at least being ready to do so if needed (longer absence, additional unexpected work etc.)
13:25:11 <patersonc[m]> Okay :)
13:25:18 <pave1> OTOH... I believe we have zero CIP specific changes in 5.10 tree, so ... there's no advantage of using -cip tree.
13:25:43 <jki> that is my technical argument as well when people ask
13:25:49 <jki> may change, though
13:26:12 <jki> and then there is also the psychological aspect of not yet having tagged...
13:26:37 <jki> stronger commitment
13:26:44 <patersonc[m]> pave1: Although as you say, we did add the missing patches from v4.19?
13:27:17 <jki> at we bit-identical with LTS or are we not?
13:27:46 <pave1> jki, patersonc: Not sure really. We may have backported patches from 5.11 to 4.19, and those would need to go to 5.10.
13:28:27 <pave1> so there probably is reason for 5.10-cip already.
13:28:41 <jki> then we should stress that
13:28:52 <patersonc[m]> I've got a feeling we already did that
13:28:52 <jki> but also update linux-5.10.y-cip...
13:29:07 <patersonc[m]> At least, we looked at the Renesas side of things for that
13:29:14 <jki> it's 4 months old on git.kernel.org
13:30:03 <pave1> I checked, and 5.10-cip has cip specific changes, such as 98eb71578 dt-bindings: pci: rcar-pci-ep: Document missing interrupts property
13:30:12 <pave1> So yes, we should probably update and release that.
13:30:38 <jki> I'll add that as new AI
13:30:49 <jki> and also adding that try to CI, I suppose
13:30:55 <iwamatsu> There was no story that 5.10-CIP was delayed four months ago....
13:31:01 <jki> s/try/tree
13:32:56 <alicef> 5.10-cip is already on kernelci
13:33:24 <jki> ah, perfect
13:33:37 <iwamatsu> And merge testing with the latest 5.10 tree is always done.
13:33:54 <jki> then it's just about official updating that branch
13:34:19 <iwamatsu> https://gitlab.com/cip-project/cip-kernel/linux-cip/-/commits/ci/iwamatsu/linux-5.10.y-cip-rc
13:34:35 <alicef> https://linux.kernelci.org/job/cip/branch/linux-5.10.y-cip/
13:36:32 <jki> ok - anything else under this topic?
13:36:55 <jki> 3
13:36:56 <jki> 2
13:36:59 <jki> 1
13:37:03 <jki> #topic Kernel testing
13:38:18 <jki> all covered before already?
13:38:23 <patersonc[m]> Nothing from me, I'll had over to alicef :)
13:38:33 <alicef> i already exhausted my topics on kernel testing
13:38:56 <jki> then move on in
13:38:58 <jki> 3
13:39:00 <jki> 2
13:39:03 <jki> 1
13:39:23 <jki> #topic AOB
13:41:40 <jki> uli: you are working in the group for a while now - do things work for you? any feedback/impressions?
13:42:23 <uli> works well for me so far
13:42:36 <uli> we are still looking for a better way to distribute the work, i guess
13:43:33 <pave1> Yes. I looked, and while etherpad does what we need, it is probably easier to just place it into git repository somewhere.
13:44:01 <pave1> Alternatively we could distribute work based on first letter of patch subject.
13:44:33 <pave1> We should probably create new repository for this, and agree that commit messages will not be used there.
13:45:06 <jki> maybe better on "commit number % developers" ;)
13:45:18 <pave1> jki: No.
13:45:34 <pave1> jki: You want series of related patches to go to the same person.
13:45:49 <jki> right
13:46:44 <uli> git repo sounds good to me
13:46:51 <iwamatsu> +1
13:47:30 <pave1> And I guess we all agree on "no real commit messages there"?
13:47:43 <pave1> Can we create repo ourselves?
13:48:14 <jki> formally, we need an approval for any new "project" under cip-project/
13:48:21 <iwamatsu> or we can use gitlab snippets
13:48:37 <jki> we can start on the playground and migrate later
13:49:02 <iwamatsu> https://gitlab.com/cip-project/cip-kernel/linux-cip/-/snippets
13:49:07 <jki> or use snippets if they are fine
13:50:20 <pave1> I have no experience with snippets.
13:51:01 <pave1> I guess repository could go under cip-project/cip-kernel/, but we'd want new repo there.
13:51:33 <uli> is it actually necessary to put it in any "official" place? it's just for internal coordination after all.
13:51:44 <jki> iwamatsu: could you propose a workflow based on snippets?
13:52:07 <jki> and if you want a git repo instead, again my proposal to create something in cip-playground first
13:52:13 <jki> that is "non-official"
13:52:32 <jki> https://gitlab.com/cip-playground
13:52:43 <iwamatsu> jki: OK, I will investigate and propose.
13:55:35 <alicef> thanks
13:56:41 <jki> thanks!
13:56:56 <jki> any other topics for today?
13:57:23 <jki> 3
13:57:28 <jki> 2
13:57:31 <jki> 1
13:57:35 <jki> #endmeeting