Radeon regression in 6.6 kernel

Phillip Susi phill at thesusis.net
Wed Nov 29 20:24:06 UTC 2023


Luben Tuikov <ltuikov89 at gmail.com> writes:

> I remember that the problem was really that amdgpu called drm_sched_entity_init(),
> in amdgpu_ttm_set_buffer_funcs_status() without actually having initialized the scheduler
> used therein. For instance, the code before commit b70438004a14f4, looked like this:
>

<snip>

> 		sched = &ring->sched;                             <-- LT: No one has initialized this scheduler
>	        r = drm_sched_entity_init(&adev->mman.entity,     <-- Oopses, now that sched->sched_rq is not

<snip>

> Before commit 56e449603f0ac5, amdgpu was getting away with this, because the sched->sched_rq
> was a static array.
>
> Ideally, amdgpu code would be fixed.

This sounds like an initilization problem that resulted in an OOPS at
boot time, but I don't remember seeing that.  I just get a failure on
system suspend.


More information about the amd-gfx mailing list