Bug#790694: linux-image-3.16.0-4-powerpc64: nouveau driver and msi interrupts

Hi Ben,

I haven't applied the patch, I thought I would do a little more investigations before proceeding with disabling msi interrupts for all nv47gpu's.

I have recompiled a 4k pagesize kernel again but enabling msi support. I can use the module option to disable msi interrupt's of nouveau if need be.

But your right, the kernel does assign an msi address for the Quadro FX4500 GPU, however as mentioned before, it seems to work with the nouveau framebuffer console.

But when Xorg starts up with 2d acceleration, the gpu locks up. Going back to legacy interrupts and it seems to work fine again...

I have attached the detailed lspci -vvv output along with the output of /cat/proc/interrupts.

Please note, you can see one interrupt is triggered with msi interrupts and nouveau, but it just hangs the gpu and the nouveau interrupt count doesn't increment.

I can still ssh to it though, just the graphics is dead.

Any tips where to proceed? Should I recompile the kernel with pci debugging support to hopefully give more helpful feedback?

I think there are a few more issues with the PowerMac 11.2 support, but perhaps we can look at this issue.

Thanks Ben,

Best Regards,

Peter

On Mon, Jul 6, 2015 at 11:56 AM, Ben Hutchings <ben@decadent.org.uk> wrote:

On Mon, 2015-07-06 at 10:57 +1000, Peter Saisanas wrote:
> Hi Ben,
> Thanks for this, ill give it a go as soon as i get a chance.
> I'd say msi interrupts are fine on x86 with this class of gpu, but
> with powerpc or this particular config it may be a problem.

I asked upstream and was told MSIs should work on this PowerMac.

> I can give you the output of lspci -vvv when running with msi enabled
> on this particular GPU.
> It would just hang when starting xorg.
> When running /cat/proc/interrupts with msi enabled, it would show
> only one interrupt triggered on whatever cpu and hang the gpu if i
> recall.
> Strangely enough, msi interrupts on powerpc seem to work fine just
> with the nouveau console framebuffer...

The driver possibly doesn't ever need to wait for interrupts when writi
ng text to the console.

> If i recall, even with msi enabled, the msi address was still 0x0.
> But i will send you a log for your reference.
>
> Yes, and i am configuring kernel with 4kb kernel pagesize as there
> still seems to be an issue with 64kb kernel pagesizes and nouveau.

I saw that bug report as well. I'm not sure what to do about it -
other distributions were also using 64K pages for 64-bit PowerPC the
last time I looked, and there may be good reasons to do that.

Ben.

--
Ben Hutchings
If you seem to know what you are doing, you'll be given more to do.