15:01:17 <pwhalen> #startmeeting Fedora ARM and AArch64 Status Meeting
15:01:17 <zodbot> Meeting started Tue Feb  2 15:01:17 2021 UTC.
15:01:17 <zodbot> This meeting is logged and archived in a public location.
15:01:17 <zodbot> The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:17 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:17 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting'
15:01:17 <pwhalen> #chair pwhalen pbrobinson jlinton coremodule
15:01:17 <zodbot> Current chairs: coremodule jlinton pbrobinson pwhalen
15:01:17 <pwhalen> #topic roll call
15:01:24 * jlinton waves
15:01:29 * coremodule is here!
15:03:03 * pwhalen gives it another minute
15:05:21 <pwhalen> #topic 1) ==== Userspace Status  ====
15:05:31 <pwhalen> Any new issues?
15:06:06 <coremodule> nothing from me
15:06:48 <pwhalen> #info No new issues reported.
15:06:57 <pwhalen> #topic 2) ==== Kernel Status ====
15:07:08 <pwhalen> #info kernel-5.11.0-0.rc6.141.fc34
15:07:09 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1697758
15:07:49 <pwhalen> So far, not hit anything new. Anyone else?
15:09:10 <jlinton> No, 5.11 is working ok (post rc2 or so) on the couple devices I've been running it on
15:09:24 <pwhalen> jlinton: that coresight issue needs to be fixed in firmware?
15:09:53 <jlinton> Yes, there isn't anything we can do about bad tables, unless we simply quirk the machine to disable coresight.
15:10:05 <jlinton> or we create our own table and load it manually.
15:10:51 <pwhalen> ok :(
15:11:02 <jlinton> I guess we should have an offline discussion about shaking the OEM bush some more.
15:11:20 <pwhalen> yea
15:11:39 <pwhalen> Our biggest kernel issue is armhfp
15:12:03 <jlinton> The oz build, or something else?
15:13:00 <pwhalen> Standard kernel gives no output, lpae boots but the panics because it cant unpack the initramfs
15:13:35 <pwhalen> "Initramfs unpacking failed: no cpio magic"
15:14:10 <jlinton> qemu/kvm or something else?
15:14:12 <pwhalen> It was suggested the kernel is overlapping with the initramfs
15:14:37 <pwhalen> on hardware it actually seems to work.. in qemu it doesnt
15:14:44 <jlinton> this is with grub right?
15:15:32 <pwhalen> right. Grub works on hw, fails in qemu
15:15:49 <pwhalen> syslinux fails all over, but lpae gives some output
15:16:33 <jlinton> Let, me see if I can find some time this week to spin that up.
15:16:54 <jlinton> Although, I'm wondering if we should consider getting rid of LPAE soon too.
15:16:54 <pwhalen> that would be fantastic, I'll get a bug filed with some details as well
15:17:36 <jlinton> I think the general conversation WRT the 32-bit kernel is that people with more than a GB or so of ram will tend to be running 64-bit capable machines now, and should move to them.
15:17:57 <jlinton> "high-end" 32-bit machines aren't being produced anymore for the most part.
15:18:53 <pwhalen> we need the lpae for our builders
15:19:08 <pwhalen> which is the other issue we have for our kernel
15:19:39 <pwhalen> https://bugzilla.redhat.com/show_bug.cgi?id=1920183
15:19:52 <pwhalen> "Overeager OOM kills on 5.10 kernels on 32bit arm with lpae"
15:20:03 <jlinton> Well, maybe tie that discussion together with building in a container rather than a full VM?
15:20:39 <jlinton> since the trend with Cortex-* cores is 32-bit EL0 only.
15:21:12 <pwhalen> I dont know all that would be required for the buildsystem to switch over to it
15:21:33 <pwhalen> sounds liek a good future change
15:24:12 <pwhalen> #info Please test and report any issues to the list or #fedora-arm.
15:24:28 <jlinton> I wonder how much of that OOM bug is just gcc11+other things getting fatter
15:24:42 <jlinton> and more packages hitting RAM limits
15:27:20 <pwhalen> Quite possible.
15:27:48 <pwhalen> #topic 3) ==== Bootloader Status ====
15:27:59 <pwhalen> #info uboot-tools-2021.04-0.1.rc1.fc34
15:27:59 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1699614
15:28:26 <pwhalen> Not tested this yet, but will take a look today
15:28:59 <pwhalen> Any issues with the previous build? Or in general?
15:30:58 <pwhalen> #topic 4) ==== Fedora 34 ====
15:31:07 <pwhalen> #info Latest nominated compose
15:31:07 <pwhalen> #link https://fedoraproject.org/wiki/Test_Results:Current_Installation_Test
15:31:54 <pwhalen> openqa results are looking better in the last compose - https://openqa.fedoraproject.org/tests/overview?distri=fedora&version=Rawhide&build=Fedora-Rawhide-20210201.n.1&groupid=5
15:32:28 <pbrobinson> pwhalen: on u-boot, testing for that would be good as 2021.04 will be the GA release for F34
15:33:02 <pwhalen> will do, thanks
15:33:13 <pwhalen> this is concerning - https://openqa.fedoraproject.org/tests/767675#step/uefi_postinstall/8
15:33:39 <pwhalen> looks like graphical issues again? Maybe debug kernel?
15:34:41 <pwhalen> #info Please test and report any issues to the list or #fedora-arm.
15:34:53 <coremodule> huh, weird
15:35:19 <pwhalen> coremodule: its happened before, I cant recall how it was fixed or if it just went away
15:35:27 <jlinton> I think its been doing that for a while, its been a bit hit/miss. I've assumed it has something to do with the arm memory mapping mismatches/etc with the VM framebuffer
15:36:05 <pwhalen> jlinton: yea I've seen it before as well, but looks to be more consistent
15:37:10 <pwhalen> #info Fedora 34 Schedule - https://fedorapeople.org/groups/schedule/f-34/f-34-key-tasks.html
15:38:24 <pwhalen> Any other concerns for F34?
15:39:19 <pbrobinson> still awaiting the approval of the uefi on armv7 change
15:39:28 <pbrobinson> but other than that nothing of note
15:40:56 <pwhalen> #info Fedora 34 Changeset
15:40:56 <pwhalen> #link https://fedoraproject.org/wiki/Releases/34/ChangeSet
15:41:01 <pwhalen> #topic 5)  == Open Floor ==
15:41:06 <pwhalen> Anything else for today?
15:41:57 <coremodule> i don't have anything
15:43:17 <pwhalen> #endmeeting