[Mesa-dev] [PATCH RFC 8/8] nvc0: disable MS images on GM107+

[Ilia Mirkin]

On Tue, Jul 19, 2016 at 12:26 PM, Samuel Pitoiset
<samuel.pitoiset at gmail.com> wrote:
>
>
> On 07/19/2016 06:11 PM, Ilia Mirkin wrote:
>>
>> On Mon, Jul 18, 2016 at 4:56 PM, Samuel Pitoiset
>> <samuel.pitoiset at gmail.com> wrote:
>>>
>>> MS images have to be handled explicitly and I don't plan to implement
>>> them for now.
>>>
>>> Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
>>> ---
>>>  src/gallium/drivers/nouveau/nvc0/nvc0_screen.c | 7 +++++++
>>>  1 file changed, 7 insertions(+)
>>>
>>> diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
>>> b/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
>>> index f681631..a0d3495 100644
>>> --- a/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
>>> +++ b/src/gallium/drivers/nouveau/nvc0/nvc0_screen.c
>>> @@ -90,6 +90,13 @@ nvc0_screen_is_format_supported(struct pipe_screen
>>> *pscreen,
>>>                   PIPE_BIND_LINEAR |
>>>                   PIPE_BIND_SHARED);
>>>
>>> +   if (bindings & PIPE_BIND_SHADER_IMAGE &&
>>> +       nouveau_screen(pscreen)->class_3d >= GM107_3D_CLASS) {
>>> +      /* MS images are currently unsupported on Maxwell because they
>>> have to
>>> +       * be handled explicitly. */
>>> +      return false;
>>> +   }
>>
>>
>> Doesn't this disable *all* shader images, not just MS ones? I think
>> you want to check for samples > 1.
>
>
>
> No, because PIPE_BIND_SHADER_IMAGE is for MS images only (and PBO
> downloads), but you are right that checking samples > 1 is better.

No, PIPE_BIND_SHADER_IMAGE is for everything. Just because st/mesa
doesn't always check it doesn't mean another state tracker might not.