fedora_arm_and_aarch64_status_meeting
LOGS
15:00:23 <pwhalen> #startmeeting Fedora ARM and AArch64 Status Meeting
15:00:23 <zodbot> Meeting started Tue Dec  3 15:00:23 2019 UTC.
15:00:23 <zodbot> This meeting is logged and archived in a public location.
15:00:23 <zodbot> The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:23 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:23 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting'
15:00:23 <pwhalen> #chair pwhalen pbrobinson jlinton
15:00:23 <zodbot> Current chairs: jlinton pbrobinson pwhalen
15:00:23 <pwhalen> #topic roll call
15:00:35 <pwhalen> good morning folks, who's here today?
15:01:39 * pbrobinson is here
15:01:54 * tdawson is here
15:02:14 * nirik is lurking
15:02:32 <tdawson> Though, all my arm work has been iot  ... so not too much to contribute today.
15:02:56 <pwhalen> ok, lets get started
15:03:05 <pwhalen> #topic 1) ==== Userspace Status  ====
15:03:55 <pwhalen> does anyone have any issues to discuss?
15:04:20 <pbrobinson> nothing I'm aware of
15:04:47 * nirik had 2 items, but can wait for the open floor or whatever...
15:04:49 <pwhalen> nirik, you here for the lpae/xfs issue? I'm digging in now but not yet reproduced. will ping the bz when I do
15:05:00 <pwhalen> whats the other? :)
15:05:08 <nirik> yeah, so thats one... sort of. I am not sure it's really xfs...
15:05:26 <pwhalen> seen anything weird since going to ext4?
15:05:28 <nirik> it seems armv7 vm's behave badly with larger amounts of memory
15:05:42 <nirik> up to about 24gb is fine...
15:06:00 <nirik> then from 25-40 they boot ok, but i/o gets really slow...
15:06:12 <nirik> over 40ish they don't boot, or panic on boot
15:06:25 <nirik> I haven't filed this yet, I haven't had time to sit down and duplicate it.
15:06:27 <pwhalen> huh, could that have been the xfs issue, or it completely separate?
15:06:45 <nirik> it could have been the xfs issue... that was one with lots of memory
15:07:08 <nirik> I'll try and sit down and document it and file a bug this week if I can get time.
15:07:11 <pwhalen> ok, might also be why I havent been able to reproduce. I didnt give it anywhere near that
15:07:15 <pbrobinson> nirik: I'm not sure more than ~24gb will be very useful, it's using LPAE
15:07:28 <nirik> yeah, I don't know what the limits are there...
15:07:41 <nirik> for now I have been giving new armv7 buildvm's 24.
15:08:09 <pbrobinson> and I'm not sure upstream would necessarily care for larger amounts for lpae although I could ask
15:08:41 <nirik> it wasnt clear where the limit was... 24, 32?
15:09:44 <nirik> anyhow, just wanted to mention that as a headsup.
15:09:59 <nirik> (or in case you all had seen it and it was not worth looking more)
15:09:59 <pwhalen> sorry, was googling and not seeing it on a quick search
15:10:50 <pbrobinson> for the time being I suggest leaving it at 24
15:11:05 <pwhalen> nirik, thanks. Will continue to try to reproduce
15:11:20 <nirik> pbrobinson: will do.
15:11:54 <nirik> pwhalen: thanks
15:12:07 <pwhalen> any other user space issues?
15:12:43 <pwhalen> #topic 2) ==== Kernel Status ====
15:12:59 <pwhalen> #info F32: kernel-5.4.0-2.fc32
15:12:59 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1416758
15:13:38 <pwhalen> I've tested a variety of boards, not seeing any issues so far
15:14:54 <pwhalen> looks like 5.5 rc's are going to land soon, recent build attempts have failed
15:15:06 <pwhalen> #info Please test and report any issues to the list or #fedora-arm.
15:16:14 <pwhalen> any other kernel news, issues?
15:16:33 <pbrobinson> nothing of note from me but I'm some what behind
15:16:55 <pwhalen> #topic 3) ==== Bootloader Status ====
15:17:11 <pwhalen> #info uboot-tools-2020.01-0.3.rc4.fc32
15:17:11 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1419083
15:17:32 <pwhalen> building now
15:17:43 <pbrobinson> yes, nothing of note there
15:18:30 <pwhalen> #info Please test and report any issues to the list or #fedora-arm.
15:19:00 <pwhalen> #topic 4) ==== Fedora 32 ====
15:19:06 <pwhalen> #info Latest nominated nightly
15:19:07 <pwhalen> #link https://fedoraproject.org/wiki/Test_Results:Current_Installation_Test
15:19:32 <pwhalen> so we've got testable nightlies again as of last Friday
15:20:07 <pwhalen> there were some selinux bugs, and other glibc/seccomp issues on armv7 that are now fixed
15:21:07 <pwhalen> might be some more selinux bugs
15:21:31 <pbrobinson> oh yay
15:21:59 <pwhalen> #info Fedora 32 Schedule - https://fedorapeople.org/groups/schedule/f-32/f-32-key-tasks.html
15:22:44 <pwhalen> does anyone plan any changes for F32? Anything we should start talking about?
15:23:28 <pwhalen> deadlines over the holidays :/
15:23:33 <pbrobinson> probably a few I should start thinking about
15:24:45 <pwhalen> we'll discuss more next week perhaps.
15:24:49 <pwhalen> anything else for f32?
15:26:20 <pwhalen> #topic 5)  == Open Floor ==
15:26:35 <pwhalen> anything else for today?
15:26:50 <nirik> so, I had one thing more... not sure if it's meeting worthy, or should just be discussed later in channel...
15:27:55 <pwhalen> go for it :)
15:28:05 <nirik> we have 3 aarch64 mustangs pbrobinson set up to do appliance/container builds for armv7... they don't seem to be working lately, and I am not sure how to debug them. No screenshot on the failed tasks... if I try virsh console I don't get anything either... I guess they are using the efi boot setup?
15:29:42 <pwhalen> nirik, yes, they worked then something got changed and it all broke again :(
15:30:05 <pwhalen> but it is efi boot
15:30:39 <nirik> ok. when is that better than normal boot? :)
15:31:13 <nirik> anyhow, we should try and fix those composes if we can. I'm happy to help, just not sure how to debug
15:31:21 <pbrobinson> nirik: yes, I need to get back to debugging that issue, it's just purely been a time and other escalation issue :(
15:31:22 <pwhalen> when it works? :P
15:31:49 <nirik> should we track it somewhere? (I am not sure where...)
15:32:15 <pbrobinson> I had started working on it pre F-31 GA when the dnf/rpm issue came along and consumed all "close to free time I didn't have" that I had allocated to it
15:32:43 <pbrobinson> and I've not yet had cycles to get back, it's got about 6 tabs in my browser dedicated to it
15:34:09 <nirik> ah fun
15:34:16 <nirik> well, if there's anything I can do to help, let me know.
15:34:22 <nirik> thats all I had. :)
15:35:17 <pwhalen> that's all I had as well, I'll leave it open for a couple of minutes
15:35:23 <pbrobinson> nirik: thanks and will do, I suspect we just need a fix or two for it, there's a couple of bugs that sort of track it
15:35:55 <nirik> yeah, I hope it's something easy.
15:36:12 <pbrobinson> these sort of track the issues I suspect are at fault https://bugzilla.redhat.com/show_bug.cgi?id=1705676 https://bugzilla.redhat.com/show_bug.cgi?id=1661288
15:36:13 <nirik> is it worth trying to not use efi? or thats needed?
15:36:36 <pbrobinson> It was working for a very short period of time
15:36:46 <nirik> also, I guess we can't do this nested on aarch64 vm's right?
15:37:14 <nirik> I guess I could set 2 of the new ampere boxes to be buildhw's and do it on them... ?
15:39:01 <pwhalen> right, nested wont work unfortunately
15:40:32 <nirik> or just keep the mustangs around. ;) anyhow...
15:43:15 <pwhalen> lets continue on the arm channel
15:43:51 <pwhalen> #endmeeting