12:01:05 <jki> #startmeeting CIP IRC weekly meeting 12:01:05 <collab-meetbot`> Meeting started Thu Jul 16 12:01:05 2026 UTC and is due to finish in 60 minutes. The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot. 12:01:05 <collab-meetbot`> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 12:01:05 <collab-meetbot`> The meeting name has been set to 'cip_irc_weekly_meeting' 12:01:13 <jki> #topic AI review 12:01:20 <jki> - start blogpost draft [iwamatsu-san] 12:01:29 <iwamatsu> hello 12:01:31 <iwamatsu> no update 12:01:39 <iwamatsu> > blogpost draft 12:01:58 <jki> ok - I guess we will eventually be asked for a potential slot date 12:02:03 <iwamatsu> I am going to start tomorrow. 12:02:26 <jki> good, just let us know if we should help anywhere 12:02:36 <jki> no other AIs recorded 12:02:44 <jki> 5 12:02:46 <jki> 4 12:02:47 <jki> 3 12:02:49 <jki> 2 12:02:50 <jki> 1 12:02:53 <jki> #topic Kernel maintenance updates 12:03:12 <masami> This week reported 3 new CVEs and 0 updated CVEs. 12:03:18 <pave1> I was reviewing 6.12.95 nad various patch series on the mailing list. 12:03:25 <pave1> Bascially catching up after travels. 12:03:33 <uli> working on 4.19; 4.4 is still missing a few reviews 12:03:41 <iwamatsu> I am reviewing 6.12.94. 12:03:52 <pave1> uli: that batman-adv thingines? 12:03:57 <pave1> thingies? 12:04:06 <uli> that's what i mean 12:04:39 <pave1> Ok, I'll put them back on my todo list. 12:04:55 <uli> it's not strictly necessary; iiuc iwamatsu-san reviewed all of it 12:05:10 <pave1> ("What is batman and why are we reviewing that? Is there some actual user? Why is it enabled in our configs?") 12:05:19 <uli> excellent questions :) 12:05:49 <jki> which batman are we talking about? :) 12:05:59 <jki> and which config asks for him? 12:06:29 <pave1> -stable patches are full of batman... 12:06:40 <pave1> Like this: 270f518e2aab7 batman-adv: tt: fix negative tt_buff_len 12:06:56 <jki> CONFIG_BATMAN_ADV is fairly often set... 12:07:11 <uli> jki: moxa_* and siemens-i386-rt (the latter has basically everything enabled :( ) 12:07:40 <pave1> Yeah. And I'm tired of that particular superhero :-). 12:07:41 <jki> the latter is quite old 12:08:05 <jki> 4.19.y/x86/siemens_iot2000.config:# CONFIG_BATMAN_ADV is not set 12:08:14 <jki> we do not feel guilty here ;) 12:08:27 <pave1> It was nontrivial part of 4.4-st review. 12:08:49 <pave1> It is some kind of mesh networking IIRC. And we might want to take a second look if we really need it. 12:09:08 <jki> moxa config pull it in in 4.4. 12:09:57 <jki> siemens_i386-rt_defconfig do on 4.19 and 5.10, ok... 12:11:24 <jki> anyway, please raise that on the mailing list towards all affected config owners 12:11:34 <pave1> If someone could take a look and perhaps remove it from the config. Another one probably not needed is qla2xxx. 12:11:42 <pave1> Ok, will do. 12:12:05 <jki> please also provide reasons along that so that we can forward 12:12:22 <pave1> (There's a ton of unused stuff, but these have people actually patching them, so they cause some load.) 12:12:42 <jki> sure 12:13:19 <jki> some configs might have received too little feature reviews before we accepted them... 12:13:33 <jki> ok - more maintenance topics? 12:14:16 <jki> 5 12:14:18 <jki> 4 12:14:20 <jki> 3 12:14:22 <jki> 2 12:14:23 <jki> 1 12:14:26 <jki> #topic Kernel release status 12:14:55 <jki> 4.4 regular and 4.19-rt are due 12:15:12 <jki> 4.4, we already discussed - anything about 4.19-rt? 12:15:33 <pave1> I need to take a look. Should be ok. 12:15:38 <jki> oh, 6.1 is due as well 12:15:59 <jki> flipped after a 12:16:01 <iwamatsu> Yes, I am going to do it. 12:16:07 <jki> after I ran the script this morning 12:16:13 <jki> good 12:16:27 <jki> so, no issues in sight, just work :) 12:16:47 <jki> anything to add? 12:16:55 <jki> 5 12:16:57 <jki> 4 12:16:59 <jki> 3 12:17:01 <jki> 2 12:17:02 <jki> 1 12:17:04 <jki> #topic Kernel testing 12:17:35 <arisut> working on a new release of kci-dev with mcp features 12:17:43 <patersonc> Nice 12:18:19 <patersonc> I've no updates this week 12:19:45 <jki> anything else on testing? 12:19:59 <jki> 5 12:20:02 <jki> 4 12:20:03 <jki> 3 12:20:04 <jki> 2 12:20:06 <jki> 1 12:20:08 <jki> #topic AOB 12:20:34 <jki> news from cip-llm-gateway.ciplatform.org: it's basically running 12:20:55 <pave1> Nice :-). What is needed to play with it? (And what to expect?) 12:20:56 <jki> uli noticed some thermal... limitations, though, we are looking into that 12:21:01 <uli> i should be able to provide the scripts and instructions for use by next meeting 12:21:12 <jki> that sounds great! 12:21:49 <uli> the gpu runs hot, so it's throttled and performance is mediocre. 12:22:06 <jki> yeah, we need to understand that 12:22:24 <pave1> What should be my expectations, compared to gpt-5.4-nano (duck.ai)? 12:22:52 <uli> i haven't done any large-scale experiments with cloud models, but 12:23:25 <uli> from what limit experience i have, the regular free chatgpt models make the same mistakes :) 12:24:06 <pave1> Ok, that sounds good :-). My local experiments ... were not comparable to gpt-5.4-nano, not even remotely. 12:24:20 <jki> I have various more advanced models at hand and can also do cross checks where useful 12:25:02 <jki> we as CIP also have Codex access now 12:25:22 <jki> via me - drop me data and prompts 12:25:42 <jki> security insensitive ones for now, more later 12:26:17 <pave1> So what would be expectations from codex? "Write me a calculator application for Commodore 64, in Pascal"? 12:26:48 <jki> use Uli's setup eventually, just with a different backend 12:26:57 <pave1> Ok. 12:28:59 <jki> other businesses for today? 12:29:21 <jki> 5 12:29:23 <jki> 4 12:29:25 <jki> 3 12:29:27 <jki> 2 12:29:29 <jki> 1 12:29:31 <jki> #endmeeting