12:03:01 <jki> #startmeeting CIP IRC weekly meeting
12:03:01 <collab-meetbot> Meeting started Thu Sep  1 12:03:01 2022 UTC and is due to finish in 60 minutes.  The chair is jki. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:03:01 <collab-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
12:03:01 <collab-meetbot> The meeting name has been set to 'cip_irc_weekly_meeting'
12:03:19 <jki> #topic AI review
12:03:26 <jki> 1. Resolve/ignore failures of KernelCI on 4.4-cip - alicefm
12:03:33 <alicef> hi
12:03:39 <alicef> no updates
12:04:16 <jki> but the ball regarding this is correctly in your field ATM?
12:05:28 <alicef> yes, we was discussing it with fbezdeka
12:05:57 <jki> ok
12:06:02 <jki> 2. Check cip devices on kernelci old pull request - patersonc
12:06:17 <patersonc[m]> Still not done
12:06:42 <jki> any other AIs from last weeks?
12:07:15 <jki> 3
12:07:17 <jki> 2
12:07:17 <pave1> No, IIRC.
12:07:22 <jki> 1
12:07:24 <jki> #topic Kernel maintenance updates
12:07:45 <uli> finished reviewing 5.10.137
12:07:57 <pave1> I'm travelling this week. Some reviews done on 5.10.138 and 5.10.140. 5.10.139 contains single patch.
12:08:03 <masami> This week reported 11 new CVEs and 7 updated CVEs.
12:08:04 <iwamatsu> I am reviewing 5.10.140
12:08:18 <pave1> I'm not sure if I'll be able to make it to the meeting tommorow (sorry).
12:09:19 <jki> which meeting?
12:09:48 <pave1> I may be mistaken. I thought there's one on Zoom.
12:10:01 <pave1> CIP TSC meeting.
12:10:04 <jki> eTSC was yesterday ;)
12:10:15 <pave1> Aha. Sorry :-).
12:10:49 <pave1> That makes things simpler.
12:12:06 <jki> there are some follow-ups from the meeting, more later
12:13:15 <jki> maybe for reference and discussion here: I did some stats on releases since April
12:13:19 <jki> 5.10:        10
12:13:25 <jki> 5.10-rt:    2
12:13:29 <jki> 4.19:        9
12:13:33 <jki> 4.19-rt:    2
12:13:37 <jki> 4.4:        2
12:13:43 <jki> 4.4-rt:     2
12:14:26 <jki> we were on track with 5.10 and 4.19 and with 4.4-rt IIRC
12:14:58 <jki> 5.10-rt, 4.19-rt and 4.4 are a bit behind the targets
12:15:09 <jki> OTOH, no one complained so far
12:16:32 <pave1> AFAICT last 5.10-rt was in July, so we should do one ASAP.
12:16:48 <pave1> 4.19-rt was also in July, so next should be sometime in September.
12:17:22 <pave1> 4.4-rt was in August, so next one should be in October.
12:18:12 <jki> 4.19-rt is indeed fine as well - 0.5 per month
12:18:23 <jki> 5.10-rt is one per month
12:18:27 <jki> so is 4.4
12:19:18 <jki> https://wiki.linuxfoundation.org/civilinfrastructureplatform/start#kernel_maintainership
12:20:23 <pave1> Agreed. We should do 5.10-cip-rt when next 5.10-rt is released.
12:20:50 <jki> would be good
12:21:34 <iwamatsu> But it depends on the RT release....
12:22:12 <jki> well, if they are releasing at lower rates than what we promise, we need to account for that
12:22:44 <jki> which likely mean lowering our promises and communicating that
12:23:52 <pave1> Let me check. Trouble with -stable-rt is that 4.9-rt, 4.19-rt and 5.10-rt each have different maintainer with different release rates.
12:25:21 <pave1> 5.10 rt has one release a month for
12:25:59 <jki> that would still fit
12:26:23 <jki> worst case if we miss one: one month without a release, but next one with 2 then
12:26:36 <pave1> 2022 (or more).
12:26:44 <pave1> But did not have one in August, AFAICT.
12:27:14 <pave1> So we should get "soon" when he releases.
12:27:26 <jki> good
12:27:46 <jki> any other maintenance topics?
12:28:08 <patersonc[m]> The other option is to fund the RT project and ask for more frequent releases
12:28:21 <jki> well, we do that already
12:28:37 <jki> but not so much for maintenance
12:28:39 <patersonc[m]> Then maybe we should ask them what is possible
12:28:53 <jki> I don't think that is directly paid out of the LF project
12:29:31 <jki> we can still ask for clarification and can try to prioritize this once the mainline merge is "done"
12:29:49 <jki> I think the next RT meeting is end of Sept
12:30:28 <jki> move on?
12:30:56 <jki> 3
12:30:58 <jki> 2
12:31:00 <jki> 1
12:31:03 <jki> #topic Kernel testing
12:32:10 <alicef> still no updates on CIP pr progress
12:34:17 <jki> anything is?
12:34:47 <alicef> KernelCI have some updates and still working on CIP pr but no updates
12:38:58 <jki> ok, anything else?
12:39:11 <jki> 3
12:39:13 <jki> 2
12:39:15 <jki> 1
12:39:20 <jki> #topic AOB
12:39:35 <jki> RISC-V and 5.10-cip
12:40:12 <jki> in order to ramp up support for RISC-V and then add the RZ-Five, I suggested to first officially declare support for that arch in our latest kernel
12:40:39 <jki> that would become a proposal to the TSC to vote on
12:40:48 <pave1> Ok. I guess we can do that for kernel.
12:41:02 <jki> right, only for the kernel so far
12:41:03 <iwamatsu> I think so
12:41:21 <iwamatsu> QEMU too?
12:41:22 <pave1> I was not able to get working userspace so far, and it looks like userspace is a bit of "work in progress" at the moment.
12:41:46 <pave1> Yeah, I guess we should do QEMU if we claim RISC-V support.
12:41:53 <jki> userspace depends on where you pick it up from, and when
12:42:11 <jki> right qemu-riscv64 would become the first target
12:42:20 <iwamatsu> Okay
12:42:24 <jki> for 5.10-cip then
12:42:24 <pave1> Do we want to claim risc-v-realtime?
12:42:31 <jki> good question...
12:42:46 <jki> I doubt that this is already seriously looked it, is it?
12:43:04 <pave1> I have not checked really, but -rt only supports listed architectures.
12:43:15 <jki> and is RISC-V listed at all?
12:43:23 <jki> if not, that would be easy for us :)
12:44:06 <pave1> I'd have to check. Lets add it to my todo list.
12:45:00 <jki> please check soon, then I can also send out the proposal sooner
12:45:19 <jki> so, what do we do for userspace for now?
12:45:47 <jki> buildroot? some frozen and known-to-be-good(TM) debian sid-version?
12:45:55 <patersonc[m]> Can we just use the Renesas/Yocto userspace for now?
12:46:05 <patersonc[m]> Does it need to be Debian?
12:46:09 <jki> does it have all features we need to kernel testing?
12:46:26 <jki> eventually, it has to be Debian, sure, but we can start differently
12:46:36 <pave1> So... I tried various debian versions and one Ubuntu version.
12:46:46 <patersonc[m]> May need some additions for LTP etc.
12:46:49 <pave1> On each version I tried, ldconfig segfaults and gcc segfaults.
12:47:01 <jki> against QEMU or RZ/Five?
12:47:22 <pave1> On RZ/Five. Other software worked quite well.. so this is quite suspect.
12:47:41 <jki> let's focus on QEMU first, I would say
12:47:57 <pave1> That makes sense, I guess.
12:48:08 <jki> meanwhile, chris & Co. can double-check your findings
12:48:17 <patersonc[m]> Ah yes, can't use the Renesas BSP for qemu - sorry :P
12:48:35 <iwamatsu> If we don't need a big application (GNOME etc), I think there's no problem with Debian/unstable.
12:48:59 <jki> Debian unstable is only stable when you freeze it via snapshot.d.o
12:49:14 <jki> but, yes, you can use that for even bigger applications
12:49:19 <jki> we have ROS2 running on it
12:49:42 <jki> with hacks...
12:50:33 <iwamatsu> No problem, Debian does not support ROS2 all packages yet... :-P
12:50:37 <jki> anyway, we should take what is easiest to enable for kernel testing, works on QEMU and should be easy to run on RZ/Five as well
12:52:08 <patersonc[m]> +1
12:52:20 <iwamatsu> BTW, I can't find RT-Linux Project and RISC-V support information. Probably not in the list.
12:52:26 <iwamatsu> +1
12:53:32 <jki> ok, then I will draft a proposal over adding RISC-V as architecture to 5.10-cip and adding qemu-riscv64 as new reference target
12:53:55 <jki> mentioning that -rt is not in scope as upstream does not support it yet
12:54:01 <patersonc[m]> Thank you
12:54:03 <jki> likely not even in 6.0-rt
12:54:06 <pave1> iwamatsu: Yes, I suspect so. AFAICT x86-64 is highest priority for -rt, arm64 being second one.
12:54:40 <iwamatsu> OK
12:54:47 <jki> perfect
12:55:46 <jki> any other business?
12:56:41 <patersonc[m]> Just apologies for not being here for the testing bit. I've nothing to add regardless
12:58:02 <jki> np
12:58:10 <jki> then let's close...
12:58:16 <jki> 3
12:58:18 <jki> 2
12:58:21 <jki> 1
12:58:23 <jki> #endmeeting