13:07:24 <iwamatsu__> #startmeeting CIP IRC weekly meeting 13:07:24 <collab-meetbot> Meeting started Thu Feb 13 13:07:24 2025 UTC and is due to finish in 60 minutes. The chair is iwamatsu__. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:07:24 <collab-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:07:24 <collab-meetbot> The meeting name has been set to 'cip_irc_weekly_meeting' 13:07:35 <iwamatsu__> #topic AI review 13:08:21 <iwamatsu__> no AI on mail from jki. 13:08:29 <iwamatsu__> anything missed? 13:08:31 <iwamatsu__> 5 13:08:41 <iwamatsu__> 4 13:08:44 <iwamatsu__> 3 13:08:46 <iwamatsu__> 2 13:08:47 <iwamatsu__> 1 13:08:57 <iwamatsu__> #topic Kernel maintenance updates 13:08:59 <pave1> I was reviewing 6.1.129, now I'm working on autosel. 13:09:06 <uli_> i did 4.4 and 4.19 13:09:06 <masami> This week reported 20 new CVEs and 0 updated CVEs. 13:09:14 <uli_> thanks for the reviews! 13:09:28 <iwamatsu__> I am reviewing 6.1.129. 13:09:36 <pave1> thanks for releases! 13:09:43 <iwamatsu__> uli_: Thanks for release. 13:10:19 <jki> hi - and sorry for the delay 13:10:28 <jki> meeting already started? 13:10:36 <iwamatsu__> We already recieved some patch for 4.19.y 13:10:53 <iwamatsu__> jki: Yes, I takeover. 13:10:57 <jki> thanks! 13:11:03 <iwamatsu__> nop 13:12:29 <iwamatsu__> Do we need an announcement for 4.19.y? 13:12:40 <uli_> didn't i send one? 13:13:46 <iwamatsu__> https://lore.kernel.org/cip-dev/1733593115.596079.1739239912119@webmail.strato.de/T/#u 13:14:00 <iwamatsu__> > This is the first 4.19 release maintained exclusively by CIP. 13:14:23 <uli_> was that too subtle? :) 13:14:25 <iwamatsu__> uli_: I see, OK. 13:15:15 <iwamatsu__> sorry for this noise. 13:15:38 <iwamatsu__> anything else? 13:15:41 <iwamatsu__> 5 13:15:43 <iwamatsu__> 4 13:15:44 <iwamatsu__> 3 13:15:45 <iwamatsu__> 2 13:15:47 <iwamatsu__> 1 13:15:56 <iwamatsu__> #topic Kernel release status 13:16:34 <pave1> I'll do 6.1-cip and 6.1.128-rt. january was a little late so this catches it up. 13:16:55 <iwamatsu__> pave1: thank you. 13:17:06 <iwamatsu__> I am working for linux-5.10.y 13:17:25 <iwamatsu__> linux-4.4.y-cip-rt is late. 13:17:40 <uli_> 4.4 and 4.19 are done 13:17:54 <pave1> aha, ok, let me do 4.4-rt, too. 13:18:18 <iwamatsu__> ok, thank you. 13:18:38 <iwamatsu__> anything else? 13:18:40 <iwamatsu__> 5 13:18:42 <iwamatsu__> 4 13:18:44 <iwamatsu__> 3 13:18:45 <iwamatsu__> 2 13:18:46 <iwamatsu__> 1 13:18:56 <iwamatsu__> #topic Kernel testing 13:19:15 <patersonc> Howdy 13:19:30 <arisut> no updates from me. working on kci-dev bisections and KernelCI blog 13:19:30 <patersonc> I started to look at the device that wasn't booting with 6.6 last week 13:19:48 <patersonc> But then it seems that it's working with 6.6 again now - so I guess something got fixed? 13:20:01 <pave1> scary :-) 13:20:12 <iwamatsu__> same version? 13:20:28 <patersonc> Fixed between rc2 and the final release 13:20:32 <patersonc> I forget the exact version 13:20:52 <pave1> But I guess if it went away we don't need to debug that. 13:21:01 <patersonc> I certainly don't plan to :P 13:21:12 <iwamatsu__> if we want to take a reason, we need to run bisect..., but it is not important. 13:21:20 <patersonc> yep 13:21:38 <patersonc> iwamatsu__ did you get a chance to look into an LTP rootfs for the x86 / MCOM board? 13:22:32 <iwamatsu__> patersonc: OK, I will take a look. 13:22:43 <patersonc> Thanks 13:23:19 <patersonc> I guess nothing else from my side 13:23:24 <jki> I would have an RT testing topic 13:23:46 <jki> does anyone already know this: https://github.com/Linutronix/RTC-Testbench? 13:24:11 <patersonc> I didn't 13:24:32 <jki> I would have the proposal to build up end-to-end RT testing this way, first for x86 13:24:51 <jki> later for other archs/boards supporting XDP RT networking this way 13:25:00 <jki> (and a bit of TSN) 13:25:35 <patersonc> Fun 13:25:36 <jki> would stress both RT workloads (similar to cyclictest, though) AND RT I/O via network 13:25:59 <jki> setup is not totally straightforward, also because it will need two boxes obviously 13:26:08 <jki> but it could be rewarding for our RT testing 13:26:26 <patersonc> Sure 13:26:38 <pave1> If it uncovers somethin, it will be 'fun' to debugg. 13:26:40 <jki> I'm not sure if KernelCI is already prepared to handle something like this as well, so I would start small in our domain first 13:26:42 <iwamatsu__> I am interesting this. 13:27:00 <pave1> but yes, more testing is needed for rt. 13:27:09 <iwamatsu__> +1 13:27:27 <jki> regarding RT testing: have you seen this backport of my colleague? https://lore.kernel.org/stable/20250129153226.818485-1-florian.bezdeka@siemens.com/ 13:27:57 <patersonc> w.r.t kernelCI perhaps open an issue in their github about it? 13:27:58 <jki> I was able to trigger that (harmless) warning withing no time by booting a plain debian-rt 6.1 kernel in a VM 13:28:13 <jki> it is not just that our setup is improveable... 13:29:28 <jki> should we open the kernelci issue already now, or better try out first in our lab? 13:30:03 <arisut> https://lore.kernel.org/all/20250211090511.xeFJVnH_@linutronix.de/ 13:30:15 <arisut> PREEMPT_RT testing on kernelci 13:31:05 <jki> arisut: thanks for the pointer 13:31:08 <jki> let me read... 13:31:50 <iwamatsu__> anything missed? 13:32:02 <iwamatsu__> 5 13:32:04 <iwamatsu__> 4 13:32:06 <iwamatsu__> 3 13:32:08 <iwamatsu__> 2 13:32:09 <iwamatsu__> 1 13:32:31 <iwamatsu__> #topic AOB 13:32:42 <paulbarker> Hi folks, I have two questions - the first is about the cip-kernel-config repo, from the point-of-view of backporting support for Renesas SoCs 13:32:48 <paulbarker> When a CIP release is tagged in linux-cip, there is no corresponding tag in cip-kernel-config so it can be difficult to match up the two repositories when looking at an old CIP release. 13:32:54 <paulbarker> Is there anything we can do to improve that? My preference would be to move the CIP kernel configs in to the appropriate branches of linux-cip so they are always aligned with the rest of the tree. 13:33:08 <arisut> tldr KernelCI is starting to implement preempt_rt testing but still on the implementation phase 13:33:35 <pave1> we normally don't update configs when kernel changes 13:33:48 <pave1> I see renesas backports will need that. 13:34:13 <jki> we have some kind of lock-stepping in isar-cip-core 13:34:15 <paulbarker> pave1: We need to update the config when we backport a new driver which has a new Kconfig option 13:34:32 <jki> as we have both the configs and the kernel referenced there by versions 13:34:38 <pave1> latest config should work on the corresponding -cip branch. 13:35:01 <pave1> but yes, we should probably use same tags we do on kernel there. 13:36:00 <jki> but how often will that actually happen? 13:36:07 <pave1> (and if change in config breaks something on older -cip kernels, we should be extra careful. Does that happen?) 13:36:29 <jki> is this case a one-off or can it really repeat more often in the future? 13:36:33 <paulbarker> pave1: cip-kernel-config may have newer config symbols enabled which don't exist in Kconfig files in older releases 13:36:45 <pave1> suggestion would be only put tag into configs when it needs to change, after corresponding release. 13:37:19 <pave1> paul: yes, that's expected. But that won't cause any real problems, right? 13:37:38 <pave1> warning at oldconfig time or something... we can live with that. 13:38:02 <pave1> just warn us if you rename a symbol or something... 13:38:37 <paulbarker> Ok, let's think about it, may come back to it next week 13:38:44 <paulbarker> Second question, I have a CVE fix to backport to 4.4 & 4.19, should I be targeting the -cip or -st branches? 13:39:06 <pave1> -st. 13:39:09 <iwamatsu__> -st 13:39:33 <paulbarker> Thanks, I'll try to do the backport tomorrow 13:39:48 <pave1> thank you :-) 13:40:01 <jki> I have some CVE-related thing as well 13:40:04 <iwamatsu__> paulbarker: thank you 13:40:34 <iwamatsu__> jki: please 13:40:38 <jki> learned that my CERT colleagues are happily consuming our CVE yaml files :) 13:41:16 <jki> they just asked (so far) if we could easily also write down the cip-kernel versions that received fixes, not only the commit shas 13:41:34 <jki> no idea how much effort that would be, though 13:41:36 <pave1> git can do that. 13:41:44 <jki> yeah, we know 13:41:59 <jki> but that is another step that downstream has to script then 13:42:19 <jki> (i gave the same answer initially) 13:42:48 <pave1> yes and it may slow things down a bit. 13:42:48 <masami> Added fixed version in yaml files may be easy to users. 13:43:06 <pave1> but it seems like reasonable request. 13:44:11 <masami> I'll create sample yaml file which contains version informaion. 13:44:29 <jki> would that be something to automate in the end? 13:45:06 <masami> yes. it has to be automated. 13:45:18 <jki> masami: thanks in advance! 13:45:30 <masami> jki: np :) 13:45:50 <iwamatsu__> anything else? 13:46:05 <iwamatsu__> 5 13:46:07 <iwamatsu__> 4 13:46:08 <iwamatsu__> 3 13:46:10 <iwamatsu__> 2 13:46:11 <iwamatsu__> 1 13:46:23 <iwamatsu__> #endmeeting