13:03:05 <jki> #startmeeting CIP IRC weekly meeting 13:03:05 <collab-meetbot> Meeting started Thu Jun 13 13:03:05 2024 UTC and is due to finish in 60 minutes. The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:03:05 <collab-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:03:05 <collab-meetbot> The meeting name has been set to 'cip_irc_weekly_meeting' 13:03:12 <jki> #topic AI review 13:03:17 <jki> - prepare blog entry on SLTS kernel state and challenges [Jan] 13:03:27 <jki> no news, unfortunately 13:03:55 <jki> I've nothing more recorded otherwise 13:04:09 <jki> 5 13:04:10 <jki> 4 13:04:11 <jki> 3 13:04:12 <jki> 2 13:04:14 <jki> 1 13:04:15 <jki> #topic Kernel maintenance updates 13:04:26 <pave1> I was reviewing 6.1.93. 13:04:31 <uli> i've been reviewing 6.1.93 13:04:32 <masami> This week reported 8 new CVEs and 4 updated CVEs. 13:04:46 <iwamatsu__> I reviewed 6.1.93. 13:06:17 <jki> anything else? 13:06:47 <jki> 8 CVEs sound like vacation phase 13:07:15 <pave1> Yeah, we'll just get 196 next week :-). 13:07:34 <jki> 5 13:07:36 <jki> 4 13:07:38 <jki> 3 13:07:40 <jki> 2 13:07:41 <jki> 1 13:07:44 <jki> #topic Kernel release status 13:07:52 <jki> 4.4 is late 13:08:01 <uli> there were concerns about the quality of a patch 13:08:08 <uli> so pave1 suggested to wait for the 4.19 release 13:08:14 <uli> but that release hasn't come so far 13:08:23 <uli> i'm inclined to drop the patch and move on 13:08:29 <pave1> Yeah. I don't believe we should be basing our releases on 4.19-rc. 13:08:47 <jki> which patch in particular? 13:08:52 <uli> x86/insn: Fix PUSH instruction in x86 instruction decoder opcode map 13:09:22 <pave1> 4.19.316-rc1 is out, so I expect 4.19.316 in few days. 13:09:22 <uli> stable doesn't have a schedule; we do 13:10:04 <pave1> I mean -- there's not neccessarily anything wrong with the named patch, 13:10:18 <pave1> but if it was not yet released in part of 4.19.x, we should not be releasing it. 13:11:02 <uli> that makes it impossible to relese anything on time 13:11:09 <pave1> No, why? 13:11:13 <uli> i've been basing almost every release so far on an rc 13:11:20 <pave1> Don't :-). 13:11:20 <uli> because the release doesn't happen when we need it 13:11:43 <pave1> Just take patches from latest version than happen, and ignore anything newer. 13:12:08 <uli> that's an option 13:12:18 <pave1> Patches that are not in in 4.19.x are straight from mainline, and untested. 13:12:18 <uli> but introduces quite some latency 13:12:41 <pave1> They may be completely bogus, breaking build, etc. 13:12:52 <pave1> Yes, I believe we should live with latency. 13:13:15 <pave1> Mainline stuff gets applied to queue-4.19. Then -rc1 comes 13:13:26 <pave1> people test that, and bad patches are dropped. Then we get 4.19.x 13:13:39 <pave1> We should ignore everything before it gets to 4.19.x 13:13:53 <uli> ok, i can live with that. 13:14:25 <jki> unless it looks very critical - but that requires special handling anyway 13:14:45 <pave1> Yep. 13:14:58 <pave1> and when things do get very critical, we do get -stable releases, anyway. 13:15:28 <uli> i can do that from now on. 13:15:34 <jki> good 13:15:35 <pave1> Thank you :-). 13:15:52 <jki> then 4.4 will come soon 13:15:57 <uli> yep. 13:15:58 <pave1> With this one, 4.19.316 should happen around Monday, so it might be easy to just wait. 13:16:16 <pave1> (And verify the patches are still there in the release). 13:16:32 <pave1> Or we can stash them and reapply after the release. 13:16:56 <uli> i can wait until monday night, and if there isn't anything yet, i'll release base on .315 13:17:02 <uli> *based 13:17:41 <jki> good, moving on... 13:17:43 <jki> 5 13:17:45 <jki> 4 13:17:47 <jki> 3 13:17:50 <jki> 2 13:17:50 <pave1> Works for me. "Sat, 15 Jun 2024 11:31:50 +0000." is the date Greg quoted for 4.19.316. 13:17:59 <jki> 1 13:18:02 <uli> ok 13:18:13 <jki> #topic Kernel testing 13:18:35 <arisut> no news afaik 13:18:39 <patersonc> jki: Are there any updates for the Siemens LAVA lab? 13:19:08 <arisut> looks still down from last time I checked 13:19:23 <jki> nope, but our team was on a Retreat (partial excuse) 13:19:37 <patersonc> Okay 13:21:06 <patersonc> I'm not sure I have any news this week 13:21:36 <jki> anything else on testing? 13:22:04 <jki> 5 13:22:06 <jki> 4 13:22:08 <jki> 3 13:22:09 <jki> 2 13:22:11 <jki> 1 13:22:14 <jki> #topic AOB 13:22:27 <pave1> I'll be travelling for two weeks starting Jun 24. 13:22:38 <masami> I will be absent from the next meeting and the one after that. 13:22:50 <jki> ok 13:23:18 <arisut> I will absent next meeting 13:23:39 <pave1> There's Linux RT meeting at Jun 26; not sure if I can make it there. 13:23:51 <jki> I should be there 13:24:14 <jki> can wear both hats this time 13:24:34 <pave1> Iwamatsu-san also has our hat :-). 13:25:10 <iwamatsu__> yes :) 13:25:15 <jki> does that time slot fit you? 13:26:02 <iwamatsu__> it is no problem. 13:26:09 <jki> ok, fine 13:26:56 <jki> something else: we discussed internally to look into the topic of mapping CVEs (via commits) on kernel configs 13:27:18 <jki> is anyone aware of preexisting public approaches? 13:29:02 <jki> seems not :) 13:29:12 <pave1> So in git@gitlab.com:cip-project/cip-kernel/lts-commit-list.git 13:29:19 <pave1> we have bin/commit.py. 13:29:38 <pave1> That computes "relevant" for a commit based on source file list. 13:29:42 <masami> https://events.linuxfoundation.org/archive/2022/open-source-summit-japan/program/schedule/ 13:29:58 <pave1> And we generate source file lists manually from our configs from time to time. 13:30:01 <masami> title "Config based CVE matching for 13:30:01 <masami> Linux kernel" 13:30:18 <pave1> We use that for reviews. 13:30:20 <jki> found it - thanks! 13:30:49 <masami> koverage command in kmax repo looks helpful. https://github.com/paulgazz/kmax 13:32:47 <jki> I think it would be good for us to execute automated mapping (as far as possible) for incoming CVEs as well as config changes 13:33:18 <jki> we may have some time to try out something and then discuss further 13:35:27 <jki> ok - anything else? 13:36:14 <jki> 5 13:36:16 <jki> 4 13:36:17 <jki> 3 13:36:19 <jki> 2 13:36:20 <jki> 1 13:36:22 <jki> #endmeeting