Regression with mainline kernel on rpi4

Sudip Mukherjee sudipm.mukherjee at gmail.com
Wed Sep 22 17:18:05 UTC 2021


On Wed, Sep 22, 2021 at 12:28 PM Maxime Ripard <maxime at cerno.tech> wrote:
>
> On Wed, Sep 22, 2021 at 11:10:34AM +0100, Sudip Mukherjee wrote:
> > On Wed, Sep 22, 2021 at 10:57 AM Maxime Ripard <maxime at cerno.tech> wrote:
> > >

<snip>

>
> Still works fine (and it required some mangling of the kernel command line).
>
> If we summarize:
>
>   - You initially just dumped a panic and a link to your QA, without any
>     more context:

The SHA was also given, and I didn't know what else you would need.
The openQA link was given to show the dmesg.

>
>   - Then stating that you're not doing any test, really;

Yes, and I still say that, its just a boot test.

>
>   - Well, except for booting Ubuntu, but no other modification
>
>   - But you're not booting the standard image
>
>   - And with a custom initrd

yes, something which has always worked in boot-testing LTS kernel or
mainline kernel.

>
>   - And that QA link states that you're booting from QEMU, but you're
>     not.

I only found that the "WORKER_CLASS" has the name "qemu_rpi4", that is
a name which I choose to give as that worker laptop is connected to
rpi4 and also running qemu tests. If you want I can change the name of
the WORKER_CLASS. :)
iiuc, dmesg shows if its booting in a qemu or on a real hardware.

>
> Please provide a full documentation on what you're doing to generate
> that image, from scratch, in order to get that panic you reported
> previously.

I have now ordered another rpi4 board and will create the image from
scratch and give you the steps.

>
> I've spent an entire day trying to make sense of what you're doing
> exactly to get into that situation. I have other things to work on and I
> don't plan on figuring out any random CI system.

I am not really sure why you are trying to figure out a random CI
system. I can reproduce the problem in our setup everytime I test with
that reverted commit and I have already said I am happy to test with a
debug patch or anything else.


-- 
Regards
Sudip


More information about the dri-devel mailing list