Bug#1056170: libhsa-runtime64-1: ROCr must assume xnack is disabled

[Christian Kastner <ckk@debian.org>]

Hi Cory,

thanks for clarifying, I indeed misunderstood a few things.

On 2023-11-24 16:42, Cordell Bloor wrote:
>> However, unlikely as it may seem, I'd still like to ask: is there any
>> risk of negatively affecting the graphics side of this? Can this change
>> somehow break a regular user's video output?
>>
>> This is far-fetched, but it's not entirely inconceivable that some
>> external stack might rely on the current behavior.
> 
> Yes, there is always a risk when enabling a new feature that it will
> introduce bugs. I see there's an issue on the amdgpu bug tracker with a
> user who has both an AMD GPU and an NVIDIA GPU on their system. It seems
> that HSA_AMD_SVM is causing issues with switching the NVIDIA card back
> and forth between the host driver and vfio-pci [2].

In that case, I'd like to postpone asking the Kernel Team for now, at
least until this known issue has been addressed. (I'm not concerned
about this particular use case, but rather that the issue may be a
symptom of an underlying cause with broader reach).

That doesn't mean we can't do our own experiments, in fact I like the
idea of "forking" unstable with a customized kernel more and more, call
it "unstable-amdsvm" or whatever.

The advantage of having our own APT repo is that it's pretty easy to do
things like that.

Best,
Christian