[PATCH v5 5/5] drm/amdgpu: move PD/PT bos on LRU again

[Michel Dänzer]

On 2018-08-28 7:03 p.m., Michel Dänzer wrote:
> On 2018-08-28 11:14 a.m., Michel Dänzer wrote:
>> On 2018-08-22 9:52 a.m., Huang Rui wrote:
>>> The new bulk moving functionality is ready, the overhead of moving PD/PT bos to
>>> LRU is fixed. So move them on LRU again.
>>>
>>> Signed-off-by: Huang Rui <ray.huang at amd.com>
>>> Tested-by: Mike Lothian <mike at fireburn.co.uk>
>>> Tested-by: Dieter Nützel <Dieter at nuetzel-hh.de>
>>> Acked-by: Chunming Zhou <david1.zhou at amd.com>
>>> Reviewed-by: Junwei Zhang <Jerry.Zhang at amd.com>
>>> ---
>>>  drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> index db1f28a..d195a3d 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
>>> @@ -1107,7 +1107,7 @@ int amdgpu_vm_update_directories(struct amdgpu_device *adev,
>>>  					   struct amdgpu_vm_bo_base,
>>>  					   vm_status);
>>>  		bo_base->moved = false;
>>> -		list_del_init(&bo_base->vm_status);
>>> +		list_move(&bo_base->vm_status, &vm->idle);
>>>  
>>>  		bo = bo_base->bo->parent;
>>>  		if (!bo)
>>>
>>
>> Since this change, I'm getting various badness when running piglit using
>> radeonsi on Bonaire, see the attached dmesg excerpt.
>>
>> Reverting just this change on top of current amd-staging-drm-next avoids
>> the problem.
>>
>> Looks like some list manipulation isn't sufficiently protected against
>> concurrent execution?
> 
> KASAN pointed me to one issue:
> https://patchwork.freedesktop.org/patch/246212/
> 
> However, this doesn't fully fix the problem.

Ray, any ideas yet for solving this? If not, let's revert this change
for now.


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer