Re: [PATCH 05/16] drm/amdgpu/ttm: Use GART helper to map VRAM pages (v2)

["Timur Kristóf" <timur.kristof@gmail.com>]

On Tue, 2025-11-04 at 14:33 +0100, Christian König wrote:
> On 11/3/25 23:23, Timur Kristóf wrote:
> > Use the GART helper function introduced in the previous commit
> > to map the VRAM pages of the transfer window to GART.
> > No functional changes, just code cleanup.
> > 
> > Split this into a separate commit to make it easier to bisect,
> > in case there are problems in the future.
> > 
> > Signed-off-by: Timur Kristóf <timur.kristof@gmail.com>
> 
> Reviewed-by: Christian König <christian.koenig@amd.com>
> 
> Could also be squashed into commit #4. Otherwise commit #5 just adds
> unused functionality.
> 
> But not a must have from my side.
> 
> Regards,
> Christian.

I prefer to keep the two commits separate to make it easier to bisect
and troubleshoot. In case there are any problems with it, commit #5 can
be reverted without regressing the VCE1 functionality.

Thanks,
Timur

> 
> > ---
> >  drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 12 ++----------
> >  1 file changed, 2 insertions(+), 10 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> > b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> > index e226c3aff7d7..84f9d5a57d03 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
> > @@ -188,7 +188,6 @@ static int amdgpu_ttm_map_buffer(struct
> > ttm_buffer_object *bo,
> >  	struct amdgpu_job *job;
> >  	void *cpu_addr;
> >  	uint64_t flags;
> > -	unsigned int i;
> >  	int r;
> >  
> >  	BUG_ON(adev->mman.buffer_funcs->copy_max_bytes <
> > @@ -254,16 +253,9 @@ static int amdgpu_ttm_map_buffer(struct
> > ttm_buffer_object *bo,
> >  		dma_addr = &bo->ttm->dma_address[mm_cur->start >>
> > PAGE_SHIFT];
> >  		amdgpu_gart_map(adev, 0, num_pages, dma_addr,
> > flags, cpu_addr);
> >  	} else {
> > -		dma_addr_t dma_address;
> > -
> > -		dma_address = mm_cur->start;
> > -		dma_address += adev->vm_manager.vram_base_offset;
> > +		u64 pa = mm_cur->start + adev-
> > >vm_manager.vram_base_offset;
> >  
> > -		for (i = 0; i < num_pages; ++i) {
> > -			amdgpu_gart_map(adev, i << PAGE_SHIFT, 1,
> > &dma_address,
> > -					flags, cpu_addr);
> > -			dma_address += PAGE_SIZE;
> > -		}
> > +		amdgpu_gart_map_vram_range(adev, pa, 0, num_pages,
> > flags, cpu_addr);
> >  	}
> >  
> >  	dma_fence_put(amdgpu_job_submit(job));