15:00:35 <pwhalen> #startmeeting Fedora ARM and AArch64 Status Meeting
15:00:35 <pwhalen> #chair pwhalen pbrobinson jlinton coremodule
15:00:35 <pwhalen> #topic roll call
15:00:35 <zodbot> Meeting started Tue Apr 27 15:00:35 2021 UTC.
15:00:35 <zodbot> This meeting is logged and archived in a public location.
15:00:35 <zodbot> The chair is pwhalen. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:35 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:35 <zodbot> The meeting name has been set to 'fedora_arm_and_aarch64_status_meeting'
15:00:35 <zodbot> Current chairs: coremodule jlinton pbrobinson pwhalen
15:00:54 <pwhalen> good morning folks, who's here today?
15:00:59 * jlinton waves
15:01:21 * pwhalen high fives jlinton
15:01:25 * pbrobinson o/
15:01:43 * pwhalen \o
15:01:50 <pwhalen> Ok, lets get started
15:02:00 <pwhalen> #topic 0) ==== Fedora Linux 34 Released today! (April 27th)
15:02:00 <pwhalen> #link https://fedoramagazine.org/announcing-fedora-34/
15:02:42 <pwhalen> Thanks for all your help testing jlinton, greatly appreciated. Especially with such a quick turn around from compose to sign off.
15:03:17 * coremodule is here.
15:03:21 <pwhalen> Has anyone hit any issues we should document for common bugs?
15:03:45 <pbrobinson> thanks pwhalen, we had a couple of right blockers at the end but we got it through the gate without too much pain
15:03:46 <pwhalen> Armhfp disk images missing browsers comes to mind
15:04:05 <pbrobinson> meh
15:04:35 <jlinton> What did you spin up the workstation image on? and did it have a fbdev?
15:04:39 <pwhalen> yea, just hit it testing.. installed midori.. I guess people know what to do
15:05:58 <pwhalen> I tested rpi4 for workstation... pine64.. Others tested the nano
15:06:30 <jlinton> I didn't revisit the problem I was seeing with the kde image to see if it was happening with the workstation image, but it was getting stuck before displaying the first boot config.
15:06:56 <jlinton> it was working fine on the honeycomb with a gpu, but my rpi4 with fbdev was just hanging (as did anything that appeared to be using wayland)
15:07:37 <pbrobinson> I didn't test KDE as the KDE SIG had agreed in the acceptance of their feature to do the testing
15:07:57 <jlinton> the install media didn't have the same issues, so if one is using fbdev, it works when installed from the install image.
15:08:13 <pwhalen> wayland works on the pi4, xorg is a bit more responsive
15:08:14 <jlinton> well it didn't look like a KDE issue in the final spin.
15:08:33 <jlinton> thats with the rpi gpu driver?
15:08:48 <jlinton> aka in DT mode?
15:09:50 <pwhalen> yes.. but now you have me worried
15:10:00 <jlinton> Hence why I was concerned about the workstation image, I probably should just spin it up and see what happens.
15:10:24 <pwhalen> Did anyone else have issues with the rpi4?
15:10:59 <coremodule> I tested workstation, minimal, and server on rpi4/00 and didn't have any issues
15:11:34 <jlinton> which was definitly in DT mode, I'm mostly just worried about dumb fbdev, more than the rpi itself.
15:12:01 <coremodule> I am having trouble getting my jetson nano to output video at all. Do anyone else have issues with jetson nano video?
15:12:08 <coremodule> *does
15:12:25 <defolos> I'm having trouble getting that thing to boot at all…
15:12:41 <pbrobinson> coremodule: do you get the nvidia logo at the beginning?
15:12:47 <defolos> but I think I've gotta update the SPI flash or something for that
15:13:21 <pbrobinson> defolos: https://nullr0ute.com/2020/11/installing-fedora-on-the-nvidia-jetson-nano/ although I need to add a note about L4T versions and py2
15:13:36 <coremodule> pbrobinson, yes, I always get the logo, but depending on the monitor/image combo, sometimes I just never get video out. I have two different hdmi monitors that I try it with.
15:13:45 <pwhalen> I tested rc1 workstation and minimal on the nano, but using the f33 GA uboot
15:14:03 <pbrobinson> I tested with the F-34 U-Boot's
15:14:10 <pbrobinson> and it was fine, sometimes a bit slow
15:14:18 <defolos> <pbrobinson "defolos: https://nullr0ute.com/2"> tried that, but it didn't want to switch into recovery mode ☹️
15:14:24 <pbrobinson> I also have a serial console so I can also see what's going on there
15:15:13 <pbrobinson> defolos: I put a pair of micro switches from an old PC case on reset/recovery which made it easier
15:15:20 <coremodule> if I use the serial console, I can see it boot, but I never get video on the hdmi. or at least not in the 30 minutes I waited
15:15:26 <pbrobinson> but I've also used just a jumper cable without issues
15:15:55 <pbrobinson> jumperered the recovery, powered it on, removed the jumper cable
15:18:13 <defolos> yeah, I guess I have to try it again
15:18:25 <pwhalen> thunderbirdtr tested all images on the nano, only KDE failed
15:18:28 <defolos> tried to short circuit the recovery against GND but it did nothing
15:19:10 <pwhalen> If people could make note, ping add any issues they find in common bugs, that would be appreciated.
15:19:31 <pbrobinson> defolos: the pinouts are documented in the manual
15:19:42 <pbrobinson> pwhalen++ will do
15:20:19 <pwhalen> Anything else for f34?
15:20:47 <pwhalen> #topic 1) ==== Userspace Status  ====
15:21:10 <pwhalen> Any user space issues? In f34 or otherwise?
15:21:39 <pbrobinson> just the known firefox on arm32 here, I did another scratch build a few days ago but not got as far as looking at the logs
15:22:29 <jlinton> That wpa_supplicant bug is someone annoying and its been open for three releases now.
15:23:02 <pwhalen> jlinton: agreed. I can't recall why we don't just add it
15:23:36 <jlinton> basically wpa_supplicant is a weak dependency and doesn't get installed if the first install is a server config. Which then in turn kills NetworkManager when it can't find wpa_supplicant if there is a wifi adapter.
15:24:07 <pbrobinson> because it's a "where do we add it thing", it's because it's a soft dep and it also could use iwd, which is I suspect what gets pulled in
15:24:50 <pwhalen> oh right, I recall iwd now
15:26:02 <pwhalen> Any other user space issues?
15:26:19 <pwhalen> #topic 2) ==== Kernel Status ====
15:26:27 <pwhalen> #info kernel-5.11.16-300.fc34
15:26:27 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1738747
15:26:50 <pwhalen> #info kernel-5.12.0-198.fc35
15:26:50 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1740788
15:26:56 <pwhalen> 5.12 GA
15:27:16 <pwhalen> I've done a bit of testing, but need to do more this week.
15:27:41 <pwhalen> Be great if others could as well, I expect that will land in F34 in the next couple of weeks
15:28:00 <jlinton> As a F34 userpace note, BTI is working enough to get around the systemd WX deny for services. But services marked as such will not have BTI enabled on their main executable for the time being.
15:28:57 <pwhalen> jlinton: thanks for the update there
15:29:18 <pwhalen> #info Please test and report any issues to the list or #fedora-arm.
15:30:31 <pwhalen> #topic 3) ==== Bootloader Status ====
15:30:58 <pwhalen> #info uboot-tools-2021.04-2.fc34
15:30:58 <pwhalen> #link https://koji.fedoraproject.org/koji/buildinfo?buildID=1738800
15:31:24 <pwhalen> Many thanks to pbrobinson for the quick fix for uboot last week
15:31:31 <pwhalen> pbrobinson++
15:31:45 <pbrobinson> also upstream 2021.07-rc1 is out today, I will probably not look at that this week
15:32:04 <pwhalen> understandable :)
15:32:13 <pbrobinson> I will probably push a new build of 2021.04 to F-34 with the upstream fix for the 32 bit issue and a few other bits
15:32:13 <pwhalen> #topic 4) ==== Fedora 35 ====
15:32:26 <pwhalen> #link OpenQA Testing - https://openqa.fedoraproject.org/tests/overview.html?distri=fedora&version=Rawhide&build=Fedora-Rawhide-20210425.n.0&groupid=5
15:32:41 <pwhalen> Looks busted.
15:33:03 <pwhalen> Early, but did anyone have anything they wanted to mention for F35?
15:33:18 <jlinton> MTE heap checking?
15:33:45 <jlinton> Looks like that should get enabled RSN in glibc, so it should be in F35
15:34:12 <pbrobinson> cool, does it need toolchain tweaks or anything like that we'd need a feature for?
15:35:04 <jlinton> No, the kernel bit is already there, its just a matter of building glibc with it enabled, and then putting it in the enviroment for machines that have the capability
15:35:22 <jlinton> that will likely cause a lot of new bugs in various applications doing unusual things with pointers.
15:35:36 <pwhalen> fun! :)
15:35:42 <pbrobinson> I'm guessing it needs HW that actually supports it
15:35:54 <jlinton> but since it requires user intervention to fully enable, shouldn't be visible in normal circumstances.
15:36:04 <jlinton> HW is recommended :)
15:36:34 <pwhalen> is it 8.5 hw?
15:36:57 <jlinton> Yes, that is the one.
15:37:22 <pwhalen> thanks. Ok.
15:37:28 <pwhalen> Anything else for f35?
15:37:46 <pbrobinson> nothing for this week
15:37:48 <pwhalen> #topic 5)  == Open Floor ==
15:38:13 <jlinton> https://www.anandtech.com/show/16073/arm-announces-neoverse-v1-n2
15:38:24 <jlinton> also its RH summit today
15:38:56 <jlinton> ah that was the old link,
15:39:30 <jlinton> https://www.anandtech.com/show/16640/arm-announces-neoverse-v1-n2-platforms-cpus-cmn700-mesh
15:39:39 <jlinton> that one has perf/feature projections.
15:40:40 <pbrobinson> jlinton: yes, had seen it, what's the arm rev it's based on?
15:41:24 <jlinton> "The architectural features of the Neoverse V1 are probably the most complicated in terms of describing – essentially, it’s a v8.4 baseline architecture which also pulls v8.5 and v8.6 features in for the HPC oriented workloads the design is aimed for."
15:42:00 <pwhalen> coremodule: no video on the nano, was that with the f34 uboot?
15:42:16 <pbrobinson> jlinton: glorious as usual arm, thanks
15:42:36 <jlinton> Well that V1, the N2 is v9
15:42:48 <jlinton> Lol
15:43:24 <jlinton> V1 has SVE of course, which will make it a lot more visible to fedora.
15:44:40 <pbrobinson> why's that?
15:44:51 <pbrobinson> I thought all the HPCers were CentOS
15:45:19 <pbrobinson> and SVE1 or 2?
15:45:21 <jlinton> Yes, but while its HPC targeted I think it should be more generally available.
15:45:25 * pbrobinson runs
15:46:07 <jlinton> Same answer as before? Lol
15:47:17 <jlinton> I think re availablity the bit about "cloud native HPC" tells the story.
15:47:53 <pwhalen> heh
15:48:16 <pwhalen> Anything else for today?
15:49:06 <pwhalen> #endmeeting