fedora_arm_and_aarch64_status_meeting
LOGS
15:00:11 <pwhalen> #startmeeting Fedora ARM and AArch64 Status Meeting
15:00:11 <zodbot> Meeting started Tue Oct 15 15:00:11 2019 UTC.
15:00:11 <zodbot> This meeting is logged and archived in a public location.
15:00:11 <zodbot> The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:11 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:11 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting'
15:00:11 <pwhalen> #chair pwhalen pbrobinson jlinton
15:00:11 <pwhalen> #topic roll call
15:00:11 <zodbot> Current chairs: jlinton pbrobinson pwhalen
15:00:22 <pwhalen> good morning folks, who's here today?
15:00:28 <jlinton> Good morning
15:00:39 * nirik is lurking
15:01:53 * pwhalen waits a couple minutes for others
15:03:36 * pbrobinson is here
15:04:28 <pwhalen> alright, lets get started
15:04:30 <pwhalen> #topic 1) ==== Userspace Status  ====
15:04:48 <pwhalen> any new user space issues?
15:05:05 * pwhalen has been on PTO the past week so not sure the status
15:05:22 <pbrobinson> there's the ongoing dnf shitshow ^W problem
15:05:31 <pbrobinson> I have it close to the top of my list
15:05:40 <pbrobinson> but there's been a bunch of escalations
15:05:44 <pwhalen> which one?
15:06:03 <pbrobinson> and damn, it's not hard to work out, just revert all the shit
15:06:17 <pwhalen> I think thats why nirik is here. the unknown arch bug?
15:06:27 <nirik> yeah.
15:06:40 <pbrobinson> the non existent sub architecture that someone who doesn't understand how Arm works someone invented
15:06:45 <nirik> even with workarounds I can get a vm installed, but it fails composes
15:07:06 <pbrobinson> nirik: yea, I basically want to seriously hurt people
15:07:19 <pbrobinson> it should never have been merged without someone that knows Arm reviewing it
15:07:19 <nirik> https://bugzilla.redhat.com/show_bug.cgi?id=1691430 for the record
15:07:32 <pwhalen> iirc there is one more revert needed?
15:07:35 <pbrobinson> thank nirik, my search fu was failing
15:08:01 <nirik> we can't bring our new hw on line for armv7 until this is fixed.
15:08:12 <pbrobinson> pwhalen: I need to dig back through the changes, I think the original stuff needs to be reverted plus anything that tries to "fix" it
15:08:39 <pbrobinson> nirik: can you give me access to that?
15:08:51 <nirik> are you not on it? huh, sure...
15:09:27 <pbrobinson> nirik: I'm sure I have access, no idea what the HW names are etc
15:09:34 <pbrobinson> not had time to look TBH
15:09:36 <nirik> you seem to be cc'ed...
15:09:54 <pbrobinson> ok, my inbox is a mess, will take a look
15:10:37 <pbrobinson> I pushed a fix to the kernel last night which I think fixes some of the UEFI on v7 issues I saw, I then need to revisit the cloud/container build stuff
15:13:17 <pwhalen> any other user space issues to discuss?
15:14:20 <pbrobinson> oh, yes, we should have pypy and pypy3 for aarch64 shortly
15:14:36 <pbrobinson> the new 7.2 release finally supports aarch64
15:15:13 <pwhalen> excellent
15:16:30 <pwhalen> #info pypy and pypy3 for aarch64 added in the new 7.2 release. Building now in F32.
15:16:50 <nirik> pbrobinson: let me know if I can help any with any of that...
15:16:57 <nirik> (likely not, but I thought I would offer)
15:17:18 <pbrobinson> nirik: will do
15:17:52 <pwhalen> #topic 2) ==== Kernel Status ====
15:18:01 <pwhalen> #info F31: kernel-5.3.6-300.fc31
15:18:01 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1400114
15:18:25 <pwhalen> just started testing this kernel today. No issues so far. Anyone else?
15:19:51 <jlinton> well 5.3.5 (or the other package that got upgraded) caused the tx2 not to shutdown
15:20:05 <jlinton> I need to track that down but I discovered it last fri right before I left for the long weekend
15:20:09 <pbrobinson> jlinton: by TX2 you mean Cavium?
15:20:13 <jlinton> yes,
15:20:50 <pbrobinson> OK, is it OK with 5.3.4 which is the current stable F-31 kernel (ie the one ATM destined for GA)
15:21:09 <pbrobinson> #info the current stable F-31 shipping kernel is 5.3.4
15:21:13 <jlinton> I think it was ok, but I suspect its not the kernel, its probably a service holding up systemd
15:21:35 <jlinton> I need to plug in a cable to see, since the 'esc' function to clear the fedora logo isn't working
15:22:32 <pbrobinson> OK
15:22:49 <pbrobinson> or drop "rhgb quiet" from the cmd line
15:23:29 <pwhalen> #info F32: kernel-5.4.0-0.rc3.git0.1.fc32
15:23:29 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1400203
15:24:00 <pwhalen> #info Please test and report any issues to the list or #fedora-arm.
15:24:18 <jlinton> Also, I ran the kernel stress test on the hikey, and realized that without the fan the stress test never passes, or my patience expired.
15:24:46 <pbrobinson> LOL
15:24:47 <pwhalen> did it work before?
15:24:51 <jlinton> I'm wondering how many of these other SBC's are in that boat
15:25:00 <pbrobinson> hikey has always had heat problems
15:25:05 <jlinton> I never tried running the stress test on that board before.
15:25:09 <pwhalen> seems like quite a few do
15:25:13 <pwhalen> ok
15:25:16 <jlinton> I've got a fairly large heatsink on it, but that isn't sufficient
15:25:33 <pwhalen> :( seeing thermal issues on other hardware too
15:26:53 <pwhalen> any other kernel issues?
15:27:35 <pwhalen> #topic 3) ==== Bootloader Status ====
15:27:37 <jlinton> I'm wondering if we should have a bug/bad thermal throtting page/etc for the boards that can't complete the stress test
15:27:42 <pwhalen> #undo
15:27:42 <zodbot> Removing item from minutes: <MeetBot.items.Topic object at 0x7f93a6eaab90>
15:28:13 <pwhalen> jlinton sounds like a plan, will make note to do that
15:28:48 <pbrobinson> jlinton: what's the stress test?
15:29:12 <jlinton> Its the "stress" option to the fedora kernel tests
15:29:24 * jlinton digging up the wiki page
15:30:09 <pwhalen> #link https://fedoraproject.org/wiki/QA:Testcase_kernel_regression
15:32:00 <jlinton> Yah, thats one of them, I thought it mentioned that there is a "stress" option to runtest.sh
15:32:20 <pwhalen> I thought so too, but might just be in the help
15:32:44 <jlinton> Either way its `./runtest.sh -t stress`
15:33:43 <pwhalen> thanks
15:33:48 <pwhalen> #topic 3) ==== Bootloader Status ====
15:33:59 <pwhalen> #info uboot-tools-2019.10-2.fc31
15:34:00 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1398293
15:34:15 <pwhalen> which is included in the most recent f31 composes.
15:34:37 <pwhalen> any other bootloader changes?
15:34:45 <pbrobinson> no
15:34:52 <pwhalen> good :)
15:34:58 <tdawson> This one allows the rock960 to boot graphically :)
15:34:59 <pwhalen> any issues so far?
15:35:03 <pwhalen> w00t!
15:35:05 <pbrobinson> I think that build is mostly good, I'm sure I missed some things but I fixed a whole bunch of stuff in it
15:35:06 <pwhalen> howdy tdawson
15:35:16 <tdawson> Howdy ... sorry I'm late.
15:35:26 <pbrobinson> tdawson: what do you mean by "boot graphically"?
15:35:57 <tdawson> Well, before that, everything I did had to be over the serial cable.
15:36:20 <pbrobinson> tdawson: I doubt you get U-Boot output on HDMI with that
15:36:46 <pbrobinson> you should get it later in the linux boot process but that's nothing to do with u-boot
15:36:48 <tdawson> Oh ... you are correct ... it does start halfway through
15:37:01 <pbrobinson> yea, that's unrelated to u-boot
15:37:38 <pbrobinson> I'm working with Rockchips to get it all from U-Boot but I'm not there
15:37:54 <tdawson> Ahh ... yep ... it's the the images from the uboot-images-armv8
15:38:28 <pwhalen> great progress so far
15:39:12 <pbrobinson> and the Rock64 and RockPro64 work with that build too
15:39:39 <pbrobinson> and I literally got a pre-prod PineBook Pro given to me earlier today
15:39:47 <tdawson> So yep ... what I said was incorrect, but those images are built at the time the package is built.  But you are correct.  If someone is expecting to see the u-boot prompt over HDMI ... that doesn't happen.
15:40:12 <tdawson> Cool
15:41:13 <pbrobinson> yea, I plan to fix u-boot over HDMI for those boards, or more work with rockchips to fix it
15:41:17 <pbrobinson> just need more time!
15:42:02 <pwhalen> anything else worth noting?
15:42:35 <pbrobinson> nothing from me
15:43:24 <pwhalen> #topic 4) ==== Fedora 31 ====
15:43:51 <pwhalen> #info Go/No Go meeting this Thursday - Oct 17th
15:44:08 <pbrobinson> so find all the breakage before then :-P
15:44:10 <pwhalen> #info List of blockers - https://qa.fedoraproject.org/blockerbugs/milestone/31/final/buglist
15:44:46 <pwhalen> seems like most are on their way to being fixed so an RC might happen today
15:45:22 <pwhalen> any problems not listed?
15:46:39 <pwhalen> excellent :)
15:49:11 <pwhalen> #info Final Release - 2019-10-22 (target)
15:49:18 <pwhalen> #topic 5)  == Open Floor ==
15:49:26 <pwhalen> Anything for open floor?
15:49:57 <pbrobinson> people should start thinking about Fedora 32 and what we need there
15:51:12 <tdawson> Nothing from me.
15:51:58 <pwhalen> Thanks for coming folks, please test test test!
15:52:02 <pwhalen> #endmeeting