Monday, June 24, 2024

[fedora-arm] Re: Raspberry Pi 4 issues with Fedora Rawhide

>> Sorry for the delay in reply, it got lost in the mess of my inbox.
>>
>> > I am writing this email to raise awareness well ahead of the F41 branching
>> > on August 6 about the bugs I found while running Fedora Rawhide on the RPi 4.
>> > I have proposed these bugs as blockers for the F41 release:
>> >
>> > * gnome-initial-setup: Choosing avatar results in SetIconFile call failed for
>> > unknown reason [0]
>> > I proposed this as F41 Final blocker, but this is fairly minor bug and I can
>> > imagine that we would waive it if not fixed and create a common issue.
>>
>> TBH that doesn't look RPi specific, does it work if you set SELinux to
>> permissive? It seems like a perms/access problem to me.
>
>
> Yep, you're right. Setting up an avatar in gnome-initial-setup works on image booted up with selinux=0 on kernel cmdline.
>
>>
>> > * gsk: vulkan renderer breaks gtk4 apps on Raspberry Pi 4 and 400 [1]
>> > GSK now defaults to vulkan and it causes problems on RPi, I initially
>> > encountered crashing gnome-initial-setup (and later that all GTK4 apps are
>> > crashing upon startup). Thus I proposed it as a F41 Beta blocker. This
>> > crashing on app startup will be resolved in the next GTK release, the fix is
>> > already merged.
>> > However, there are still issues with GTK4 apps [2], all related to vulkan
>> > renderer [3] [4] [5]. It is not yet clear if the problems are bugs in GTK or
>> > mesa. In the end, if proven difficult to fix, we could always switch back to
>> > older renderer as suggested by Adam during F40 cycle [6].
>>
>> What is GSK? Was there an official Fedora change for the switch to Vulkan?
>>
>> > * Raspberry Pi 4 won't wake up from suspend [7]
>> > Although we don't have a criterion for suspend, I proposed this as a blocker
>> > bug. I cannot login every time I keep Raspberry Pi idling for 15 minutes and
>> > I have to restart it, this could also lead to loss of data. This is something
>> > different compared to the x86_64 situation. On x86_64 we probably wouldn't
>> > block on suspend on some particular hardware configuration, but I think we
>> > should block here since we support only a handful of ARM boards.
>> > Also, suspend on Raspberry Pi OS is disabled, so I'm not sure if suspend on
>> > Raspberry Pi is something we even want in Fedora.
>>
>> I am absolutely against blocking on suspend for RPi or any specific
>> platform, this is completely out of our control and it's not supported
>> even on downstream kernel forks, it's dependent on issues with closed
>> FW which is completely out of our control. There's reasons we don't
>> block on suspend on x86 so I am unsure why we'd explicitly decide to
>> torture arm maintainers with that.
>>
>> Peter
>
>
> Yes, I get that. I don't think that we should block on a specific platform, especially if it is out of our control. But it is a pretty horrible default that RPi with Fedora Workstation is unusable after 15 minutes. I don't know the internals or what it would take to fix it, but if it is just a switch to disable suspend for arm images... well then we should disable it. The bug has been re-proposed as blocking, there is a discussion issue for the blocker bug (https://pagure.io/fedora-qa/blocker-review/issue/1607). Please, please voice your concerns there, you are probably the most qualified among us.

I have
--
_______________________________________________
arm mailing list -- arm@lists.fedoraproject.org
To unsubscribe send an email to arm-leave@lists.fedoraproject.org
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/arm@lists.fedoraproject.org
Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue

No comments:

Post a Comment