On 5/26/26 12:11, Timur Kristóf wrote:
> On 2026. május 26., kedd 10:51:38 közép-európai nyári idő Christian König 
> wrote:
>> On 5/26/26 01:29, Timur Kristóf wrote:
>>> On 2026. április 27., hétfő 20:27:23 közép-európai nyári idő Christian
>>> König> 
>>> wrote:
>>>> Instead of coming up with more sophisticated names for states a VM BO
>>>> can be in, group them by the type of BO first and then by the state.
>>>>
>>>> So we end with BO type kernel, always_valid and individual and then
>>>> states
>>>> evicted, moved and idle.
>>>>
>>>> Not much functional change, except that evicted_user is moved back
>>>> together with the other BOs again which makes the handling in
>>>> amdgpu_vm_validate() a bit more complex.
>>>>
>>>> Also fixes a problem with user queues and amdgpu_vm_ready(). We didn't
>>>> considered the VM ready when user BOs were not ideally placed, harmless
>>>> performance impact for kernel queues but a complete show stopper for
>>>> userqueues.
>>>>
>>>> v2: fix a few typos in comments, rename the BO types to make them more
>>>>
>>>>     descriptive, fix a couple of bugs found during testing
>>>>
>>>> v3: squashed together with revert to old status lock handling, looks
>>>>
>>>>     like the first patch still had some bug which this one here should
>>>>     fix.
>>>>     Fix a missing lock around debugfs printing.
>>>>
>>>> Signed-off-by: Christian König <[email protected]>
>>>
>>> Hi Christian,
>>>
>>> This patch regresses GPU recovery.
>>> (Tested on SI, haven't had time to test other GPUs yet.)
>>>
>>> It hits the following warning:
>>> WARNING: drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c:635 at
>>> amdgpu_vm_pde_update And then this error:
>>> [drm:amdgpu_gem_va_update_vm [amdgpu]] *ERROR* Couldn't update BO_VA (-22)
>>>
>>> Please fix.
>>
>> Yeah that's a known issue. Fix is already pushed to amd-staging-drm-next.
> 
> Hi Christian,
> 
> Where is this fix? As far as I see "restructure VM state machine v2" is the 
> top 
> commit on amd-staging-drm-next right now.

The patch is called "drm/amdgpu: fix amdgpu_vm_bo_reset_state_machine" and 
should be on the mailing list.

Regards,
Christian.

> 
> Timur
> 
> 
> 

Reply via email to