15:00:12 <coremodule> #startmeeting Fedora ARM Working Group Meeting
15:00:12 <zodbot> Meeting started Tue May 10 15:00:12 2022 UTC.
15:00:12 <zodbot> This meeting is logged and archived in a public location.
15:00:12 <zodbot> The chair is coremodule. Information about MeetBot at https://fedoraproject.org/wiki/Zodbot#Meeting_Functions.
15:00:12 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:12 <zodbot> The meeting name has been set to 'fedora_arm_working_group_meeting'
15:00:12 <coremodule> #chair pwhalen pbrobinson jlinton coremodule
15:00:12 <coremodule> #topic roll call
15:00:12 <zodbot> Current chairs: coremodule jlinton pbrobinson pwhalen
15:00:21 <coremodule> Hi everyone, who is around for an ARM meeting today?
15:02:27 <jforbes> I am around
15:02:32 <pbrobinson> I'm here
15:03:22 <coremodule> hi jforbes
15:03:24 <coremodule> hi pbrobinson
15:03:33 <coremodule> #topic 1) ==== Stable Releases ====
15:03:43 <coremodule> #info Fedora 36 officially launches today!
15:05:04 <coremodule> #topic 2) ==== Kernel Status ====
15:05:09 <coremodule> #info kernel-5.18.0-0.rc6.47.fc37
15:05:14 <coremodule> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1964165
15:05:27 <coremodule> has anyone tried the new kernel?
15:05:42 <coremodule> I haven't yet
15:06:38 <coremodule> #info Please test and report in #fedora-arm any issues.
15:06:42 <jlinton> I wonder if we should have the CLUSTER scheduling turned on. It causes problems on the ampere, and possibly others
15:07:50 <jlinton> (it seems the default was selected a back in the 5.16 timeframe, and I didn't notice errors on the little machines I was mostly using during that timeframe)
15:09:32 <coremodule> maybe so, jforbes could we do that?
15:10:36 <jforbes> I haven't set the configs for 5.18 yet, but I can give that a run
15:10:55 <jforbes> Configs are on my list for this week
15:11:59 <coremodule> cool, thanks
15:11:59 <jlinton> I've got a patch i've been contemplating to get rid of the borken messages, but until that gets clear I think we should probably have it off, or accept that the scheduler domains are broken, at least on ampere.
15:12:29 <jforbes> jlinton: I can turn it off just for aarch64
15:14:04 <jlinton> Yah, I saw its on x86, but didn't pay really close attention to how its being triggered there, if its tied to the PPTT the same way it is on aarch64 then it probably never triggers since I've not seen an x86 with a PPTT yet.
15:14:29 <jforbes> Will not pull that back to 5.17 though, so there will be time to test 5.18
15:15:17 <jlinton> well the machine still boots, but it tosses errors, and the first MC domain has 2 cores each, which probably kills benchmarks that run a lot of threads
15:15:57 <jforbes> jlinton: should show up in builds on rawhide starting tomorrow
15:16:43 <coremodule> thanks jforbes
15:17:12 <coremodule> #topic 3) ==== Bootloader Status ====
15:17:12 <coremodule> #info grub2-2.06-40.fc37
15:17:13 <coremodule> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1963244
15:17:13 <coremodule> #info uboot-tools-2022.07-0.1.rc1.fc37
15:17:13 <coremodule> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1956320
15:19:46 <coremodule> I haven't tested either, anyone else?
15:21:59 <coremodule> #topic 4) ==== Fedora 37 status ====
15:22:33 <coremodule> no more armhfp! for better or for worse
15:23:00 <coremodule> (I say better, less to test, hah)
15:23:15 * pwhalen shows up late
15:23:23 <coremodule> hi pwhalen
15:23:31 <pwhalen> good mornin'!
15:23:34 <coremodule> #chair pwhalen jforbes
15:23:34 <zodbot> Current chairs: coremodule jforbes jlinton pbrobinson pwhalen
15:24:30 <jforbes> coremodule: Unfortunately I still have to build kernels or we may be in for a bit of an issue when doing stable rebases
15:24:41 <pwhalen> Looks reasonable in openqa. Some of the failures look like needle issues
15:25:54 <jlinton> How long could we keep building a few core things for F36/arm?
15:26:07 <jlinton> aka could we extend its life out more than the usual two years?
15:27:24 <jlinton> The extended support version, that zombies on until people lose interest?
15:27:29 <pwhalen> Nothing stops people from picking it up, but I dont think in the official koji instance
15:28:11 <jforbes> I would honestly rather not build a single 32bit arm kernel that doesn't have to be built since it is the longest build by a good bitr
15:30:06 <jlinton> Yah, i'm all on board with the lets kill it, because I don't think a full featured distro belongs  on a 32-bit platform at this point. I'm just tossing out random ideas if someone really yells loudly.
15:31:16 <pwhalen> #topic 5) ==== Open Floor ====
15:31:27 <pwhalen> Anything else for today?
15:31:45 <coremodule> nothing from me! thanks for picking it up
15:33:37 <pwhalen> curious why containers are failing in f35 and not f36 for armhfp
15:34:22 <pwhalen> nothing else from me :)
15:34:41 <coremodule> #endmeeting