13:02:03 <jki> #startmeeting CIP IRC weekly meeting 13:02:03 <collab-meetbot`> Meeting started Thu Jul 10 13:02:03 2025 UTC and is due to finish in 60 minutes. The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:02:03 <collab-meetbot`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:02:03 <collab-meetbot`> The meeting name has been set to 'cip_irc_weekly_meeting' 13:02:12 <jki> #topic AI review 13:02:13 <iwamatsu__> hello 13:02:19 <jki> none on my list 13:02:27 <jki> 5 13:02:29 <jki> 4 13:02:31 <jki> 3 13:02:32 <jki> 2 13:02:34 <jki> 1 13:02:36 <jki> #topic Kernel maintenance updates 13:02:51 <uli_> still working on 4.19, it's a big one 13:02:51 <masami> This week reported 170 new CVEs and 9 updated CVEs. 13:03:10 <iwamatsu__> I reviewed 6.12.35 and 36. 13:03:41 <pave1> I'm reviewing 6.12.35 and .36. 13:04:50 <jki> anything to add? 13:05:44 <jki> 5 13:05:46 <jki> 4 13:05:48 <jki> 3 13:05:49 <jki> 2 13:05:50 <jki> 1 13:05:53 <jki> #topic Kernel release status 13:06:09 <jki> 6.1-rt is late - waiting for upstream? 13:06:38 <iwamatsu__> I am preparing now. I am going to release tomorrow. 13:06:38 <pave1> Not really, more confusion on my part. 13:06:58 <iwamatsu__> I am preparing now. I am going to release tomorrow. > 6.12.y-cip 13:07:07 <jki> ok, assumed so :) 13:07:20 <jki> pavel: you just missed the deadline? 13:07:36 <jki> do we have a suitable base for 6.1-rt? 13:07:45 <pave1> I'll need to take a look. 13:07:50 <jki> ok 13:08:02 <pave1> As cadence changed from 1 to two months, I simply added 2 months to the date. 13:08:07 <pave1> Which ... was not correct thing to do :-) 13:08:17 <jki> we have a common script :) 13:08:44 <pave1> and that's good because it catches errors. I'm keeping whiteboard in local file. 13:09:15 <jki> anything else on releases? 13:09:27 <jki> 5 13:09:29 <jki> 4 13:09:31 <jki> 3 13:09:32 <jki> 2 13:09:34 <jki> 1 13:09:36 <jki> #topic Kernel testing 13:10:09 <arisut> I finally got cip configurations on KernelCI 13:10:10 <patersonc> Nothing from me 13:10:47 <arisut> we had a problem with identical jobs names that was blocking the cip configurations introduction but as been fixed 13:11:09 <patersonc> Good 13:11:13 <arisut> currently I'm adding lab-cip boards to KernelCI 13:11:23 <patersonc> Thanks! 13:11:28 <jki> nice 13:12:13 <jki> btw, the configs we have in x.y-cip vs. those in regular branches in cip-kernel-configs are not in sync 13:12:18 <jki> is that a problem? 13:12:53 <arisut> what do you mean by x.y-cip configuration ? 13:13:05 <arisut> the one inside the cip kernel repo? 13:13:23 <arisut> we are using cip-kernel-configs 13:13:28 <jki> in the cip-kernel-config repo 13:13:37 <jki> we have branches for every kernel there 13:13:45 <arisut> yes 13:13:54 <jki> but feature maintance only happens against the -cip branches 13:14:37 <jki> at least in most cases 13:15:27 <arisut> you mean dose not happen in -cip-rt branches for example ? 13:15:35 <arisut> does 13:15:51 <jki> also 13:15:53 <arisut> or in .y version 13:16:10 <jki> for focus is on -cip for features 13:16:13 <arisut> currently I'm using only the -cip branches 13:16:19 <arisut> afair 13:16:19 <jki> -rt is normally added via snippet 13:16:36 <jki> okay, that would be up-to-date 13:16:37 <arisut> maybe -cip-rt also 13:17:11 <jki> I never touched a cip-rt for feature extensions, but that is because we use the rt snippet for enabling RT 13:17:14 <arisut> yes also cip-rt 13:17:20 <arisut> ok 13:17:31 <arisut> is that a problem? 13:18:15 <jki> I was just wondering if it affects our testing efforst negatively if we add feature X to .y-cip but not to .y+1 or so 13:19:00 <jki> configs for regular stable kernels are taken as snapshots from the -cip kernels when we add the stable ones, right? 13:19:15 <jki> no more updates afterwards 13:19:21 <pave1> Yeah, making sure features we use are tested in newer stable kernels would make some sense. 13:19:27 <arisut> maybe we could loose regression bugs on such features ? 13:20:00 <jki> we currently only spot regressions for recently added features if related patches land in -cip 13:20:28 <jki> if we want to change that, we likely need to automate forward-porting of feature additions 13:20:41 <jki> similar to the cip_merged config generation 13:22:24 <jki> well, will need someone to implement that... 13:22:48 <pave1> Or we can ask people submitting new features to do that. 13:22:52 <pave1> :-). 13:23:04 <jki> manual checks don't work 13:23:20 <pave1> They may be good enough. 13:23:24 <jki> process is already complicated enough, just went through this 13:23:51 <pave1> There will be manual component as submitter knows which kernel to enable feature on. 13:24:12 <jki> ...which CIP kernel 13:24:23 <jki> that is what people care about in CIP 13:24:37 <patersonc> Having different config versions for every kernel kernel version is a pain. But I guess a single defconfig wouldn't work either because different kernel versions have new features we may want to test? 13:24:45 <pave1> Aha, I thought that idea would be: 13:24:57 <pave1> when "shiny cat" feature is ported from 6.16 to 6.1-cip 13:25:15 <pave1> we also need to enable 'shiny cat" in 6.12-cip and 6.16-stable test configurations. 13:25:43 <pave1> (and 6.17-stable and 6.18-stable) so that we see regressions early. 13:25:43 <jki> and that should be auto-checked at least 13:25:54 <jki> or it will constantly be forgotten 13:26:05 <jki> better would be auto-generation of the stable configs 13:26:19 <jki> so that -cip ones remain the only base 13:26:47 <jki> just like no one is supposed to touch the merged configs manually anymore 13:28:03 <jki> ok - more on testing? or configs? :) 13:28:22 <jki> 4 13:28:25 <jki> 4 13:28:27 <jki> 3 13:28:29 <jki> 2 13:28:31 <jki> 1 13:28:34 <jki> #topic AOB 13:28:47 <jki> FYI: my erofs problem was a kernel bug 13:29:12 <pave1> Good. so -- solved? 13:29:13 <jki> arm32-only, patch is on the way, also for stable - kudos to the erofs maintainer 13:30:19 <jki> https://lore.kernel.org/linux-erofs/da52a0d3-9a3b-4465-bf17-cf2ad8044330@linux.alibaba.com/T/#u 13:30:46 <jki> anything else for today? 13:31:27 <jki> 5 13:31:29 <jki> 4 13:31:31 <jki> 3 13:31:33 <jki> 2 13:31:35 <jki> 1 13:31:38 <jki> #endmeeting