15:00:42 <pwhalen> #startmeeting Fedora ARM and AArch64 Status Meeting
15:00:42 <zodbot> Meeting started Tue Jun 16 15:00:42 2020 UTC.
15:00:42 <zodbot> This meeting is logged and archived in a public location.
15:00:42 <zodbot> The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:42 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:42 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting'
15:00:42 <pwhalen> #chair pwhalen pbrobinson jlinton
15:00:42 <zodbot> Current chairs: jlinton pbrobinson pwhalen
15:00:59 <pwhalen> #topic roll call
15:00:59 <pwhalen> good morning folks, who's here today?
15:01:05 * pbrobinson is here
15:01:48 * pwhalen gives it a moment for others to join
15:03:34 * jlinton waves
15:04:00 <pbrobinson> hey jlinton
15:04:12 <pwhalen> howdy jlinton
15:04:18 <pwhalen> lets get started
15:04:25 <pwhalen> #topic 1) ==== Userspace Status  ====
15:04:42 <pwhalen> Any new user space issues?
15:06:04 <pwhalen> I'm not aware of anything, but was out last week so just getting caught up.
15:06:07 <jlinton> Not here, my basic graphics+compiler stack seems to be working well.
15:06:14 <pbrobinson> the glibc one?
15:06:14 <pwhalen> #info No issues reported.
15:06:24 <pwhalen> er, was waiting for F33
15:06:40 <pbrobinson> ¯\_(ツ)_/¯ it's userspace ;-)
15:06:46 <pwhalen> sure..
15:06:56 <pwhalen> #info systemd crashes when booting rawhide on armhfp
15:06:56 <pwhalen> #link https://bugzilla.redhat.com/show_bug.cgi?id=1831239
15:06:58 <jlinton> Yah, I've been juggling it with a _DMA() problem I can't seem to find
15:07:33 <jlinton> I need to verify with a clean install and an rd.break that I can force a core file <sigh>
15:08:13 <jlinton> Just out of curriousity I was meaning to try it with selinux off
15:08:51 <pwhalen> I tried that, didnt help. Think there are logs there with it off
15:09:33 <pwhalen> jlinton: with some guidance I am happy to help
15:10:03 <pwhalen> I've been testing kernels for armhfp, but userspace status is completely unknown in F33
15:10:13 <jlinton> Well, I've just applied the fedora ./configure settings to a git tree, then I started bisecting the branch as I mentioned in the defect.
15:10:58 <jlinton> that should get you to the failing commit, and from there stepping the early init should get you to the crashing command
15:11:03 <jlinton> Its the latter I haven't done.
15:13:08 <pwhalen> jlinton: do you think you will have time for the latter?
15:15:14 <pwhalen> or do we have someone else that can take a look?
15:16:34 <jlinton> No, let me drop this _DMA() so I can just focus on the glibc and the coresight patches.
15:16:48 <pbrobinson> thanks jlinton
15:17:21 <pwhalen> big thanks jlinton
15:17:43 <pwhalen> #topic 2) ==== Kernel Status ====
15:17:55 <pwhalen> #info kernel-5.8.0-0.rc1.1.fc33
15:17:55 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1523678
15:18:16 <pwhalen> started to test this week, no issues so far on either armhfp or aarch64
15:18:19 <jlinton> Ok, yah there is a acpi _DMA() limit/mask problem, seemingly only with the xhci
15:19:01 <jlinton> keeps machines with ZONE_DMA requirements from working (rpi4 being one)
15:20:08 <pbrobinson> there's another regression with the rpi too, I think rpi3, but the maintainer is looking at that
15:20:23 <pbrobinson> I'm slowly catching up on kernel stuff
15:20:31 <jlinton> I thought it would be straightforward but the adapter takes a generic fault, so I'm sorting through piles of buffers trying to figure out which one caused it.
15:21:24 <pbrobinson> surely being rpi4 people are falling over themselves to fix it?
15:22:22 <jlinton> Right, i'm not even sure why its not hitting the DT case, except luck? or the fact that there is an init difference between the xhci-pci and xhci-plat drivers.
15:22:54 <jlinton> I don't think its an rpi bug, so much as one that gets triggered there easier due to the DMA situation
15:23:16 <pbrobinson> OK, I figured as much
15:23:47 <pwhalen> Any other kernel issues?
15:24:14 <pwhalen> #topic 3) ==== Bootloader Status ====
15:24:30 <pbrobinson> I did get a new rc4 build done end of last week
15:24:30 <pwhalen> #info New U-Boot available for testing -  uboot-tools-2020.07-0.3.rc4.fc33
15:24:30 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1522690
15:24:41 <pbrobinson> mostly untested
15:25:00 <pwhalen> Just started testing it here, will report back if I find any issues
15:25:34 <pwhalen> #topic 4) ==== Fedora 33 ====
15:25:34 <pwhalen> #link https://fedoraproject.org/wiki/Test_Results:Current_Installation_Test
15:26:23 <jlinton> https://pagure.io/fesco/issue/2403
15:26:35 <pwhalen> Blank currently for arm/aarch64. OpenQA is down right now due to the DC move, I'll get some testing done today
15:27:06 <pwhalen> jlinton: nice!
15:27:58 <pwhalen> #info Aarch64 Pointer Authentication Change for Fedora 33 approved.
15:27:58 <pwhalen> #link https://fedoraproject.org/wiki/Changes/Aarch64_PointerAuthentication
15:28:58 <pwhalen> Anything else for F33?
15:29:05 <pbrobinson> jlinton: let me know if you need me to do changes for the rpm flags for the PAC stuff
15:29:25 <jlinton> Yah, I will put a patch on the defect.
15:29:47 <jlinton> I'm trying to decide if I should create an aarch64 specific hardening section, or just put it in the optimization flags
15:32:05 <pwhalen> #topic 5)  == Open Floor ==
15:32:14 <pwhalen> Anything else for today?
15:33:04 <pbrobinson> nothing from me
15:34:28 <pwhalen> #endmeeting