On Mon, May 06, 2019 at 10:05:07AM +, Koenig, Christian wrote:
> Am 06.05.19 um 10:04 schrieb Daniel Vetter:
> > [SNIP]
> + /* pin buffer into GTT */
> + return amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
> >>> This is kinda what I mean with "shouldn't we pin the attachment" - afaiui
>
Am 06.05.19 um 10:04 schrieb Daniel Vetter:
> [SNIP]
+ /* pin buffer into GTT */
+ return amdgpu_bo_pin(bo, AMDGPU_GEM_DOMAIN_GTT);
>>> This is kinda what I mean with "shouldn't we pin the attachment" - afaiui
>>> this can fail is someone already pinned the buffer into vram. And that
>>>
On Fri, May 03, 2019 at 02:35:02PM +0200, Christian König wrote:
> Am 30.04.19 um 16:16 schrieb Daniel Vetter:
> > [SNIP]
> > > /**
> > > - * amdgpu_gem_map_attach - _buf_ops.attach implementation
> > > - * @dma_buf: Shared DMA buffer
> > > + * amdgpu_gem_pin_dma_buf - _buf_ops.pin_dma_buf
Am 30.04.19 um 16:16 schrieb Daniel Vetter:
[SNIP]
/**
- * amdgpu_gem_map_attach - _buf_ops.attach implementation
- * @dma_buf: Shared DMA buffer
+ * amdgpu_gem_pin_dma_buf - _buf_ops.pin_dma_buf implementation
+ *
+ * @dma_buf: DMA-buf to pin in memory
+ *
+ * Pin the BO which is backing the
On Fri, Apr 26, 2019 at 02:36:36PM +0200, Christian König wrote:
> The caching of SGT's is actually quite harmful and should probably removed
> altogether when all drivers are audited.
>
> Start by providing a separate DMA-buf export implementation in amdgpu. This is
> also a prerequisite of
The caching of SGT's is actually quite harmful and should probably removed
altogether when all drivers are audited.
Start by providing a separate DMA-buf export implementation in amdgpu. This is
also a prerequisite of unpinned DMA-buf handling.
v2: fix unintended recursion, remove debugging