[Mesa-dev] [RFC 04/24] nvc0: bind images on 3D shaders for Kepler

[Samuel Pitoiset]

On 04/15/2016 09:41 AM, Hans de Goede wrote:
> Hi,
>
> On 15-04-16 00:01, Samuel Pitoiset wrote:
>>
>>
>> On 04/14/2016 08:03 PM, Pierre Moreau wrote:
>>> On 01:56 AM - Apr 13 2016, Samuel Pitoiset wrote:
>>>> Similar to surfaces validation for compute shaders.
>>>>
>>>> Signed-off-by: Samuel Pitoiset <samuel.pitoiset at gmail.com>
>>>> ---
>>>>   src/gallium/drivers/nouveau/nvc0/nvc0_program.c |  4 +++-
>>>>   src/gallium/drivers/nouveau/nvc0/nvc0_tex.c     | 26
>>>> ++++++++++++++++++++++++-
>>>>   2 files changed, 28 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
>>>> b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
>>>> index ced8130..8e73227 100644
>>>> --- a/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
>>>> +++ b/src/gallium/drivers/nouveau/nvc0/nvc0_program.c
>>>> @@ -561,12 +561,14 @@ nvc0_program_translate(struct nvc0_program
>>>> *prog, uint16_t chipset,
>>>>      } else {
>>>>         if (chipset >= NVISA_GK104_CHIPSET) {
>>>>            info->io.texBindBase = NVC0_CB_AUX_TEX_INFO(0);
>>>> +         info->io.suInfoBase = NVC0_CB_AUX_SU_INFO(0);
>>>> +      } else {
>>>> +         info->io.suInfoBase = 0; /* TODO */
>>>>         }
>>>>         info->io.sampleInfoBase = NVC0_CB_AUX_SAMPLE_INFO;
>>>>         info->io.bufInfoBase = NVC0_CB_AUX_BUF_INFO(0);
>>>>         info->io.msInfoCBSlot = 15;
>>>>         info->io.msInfoBase = 0; /* TODO */
>>>> -      info->io.suInfoBase = 0; /* TODO */
>>>>      }
>>>>
>>>>      info->assignSlots = nvc0_program_assign_varying_slots;
>>>> diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_tex.c
>>>> b/src/gallium/drivers/nouveau/nvc0/nvc0_tex.c
>>>> index 585b1e5..7cac31d 100644
>>>> --- a/src/gallium/drivers/nouveau/nvc0/nvc0_tex.c
>>>> +++ b/src/gallium/drivers/nouveau/nvc0/nvc0_tex.c
>>>> @@ -875,7 +875,31 @@ nvc0_update_surface_bindings(struct
>>>> nvc0_context *nvc0)
>>>>   static inline void
>>>>   nve4_update_surface_bindings(struct nvc0_context *nvc0)
>>>>   {
>>>> -   /* TODO */
>>>> +   struct nouveau_pushbuf *push = nvc0->base.pushbuf;
>>>> +   struct nvc0_screen *screen = nvc0->screen;
>>>> +   int i, j, s;
>>>> +
>>>> +   for (s = 0; s < 5; s++) {
>>>
>>> Similar to my comment on patch 1, shouldn’t you have `s < 6` here?
>>> (Except if
>>> you follow Ilia’s suggestion.)
>>
>> No, because s = 5 is for compute shaders and they are bound in
>> nve4_compute_validate_surfaces().
>
> Maybe add a comment explaining this to the code ? So that we do not
> end up wondering later why things are as they are ?

Well, such a loop is used at many places in the driver. If you trace 
down where nve4_update_surface_bindings() is called, you will see that 
it's for 3D only. I don't think adding a comment will be really relevant 
here.

>
> Regards,
>
> Ha\ns
>
>
>>
>>>
>>> Pierre
>>>
>>>> +      BEGIN_NVC0(push, NVC0_3D(CB_SIZE), 3);
>>>> +      PUSH_DATA (push, 2048);
>>>> +      PUSH_DATAh(push, screen->uniform_bo->offset +
>>>> NVC0_CB_AUX_INFO(s));
>>>> +      PUSH_DATA (push, screen->uniform_bo->offset +
>>>> NVC0_CB_AUX_INFO(s));
>>>> +      BEGIN_1IC0(push, NVC0_3D(CB_POS), 1 + 16 * NVC0_MAX_IMAGES);
>>>> +      PUSH_DATA (push, NVC0_CB_AUX_SU_INFO(0));
>>>> +
>>>> +      for (i = 0; i < NVC0_MAX_IMAGES; ++i) {
>>>> +         struct pipe_image_view *view = &nvc0->images[s][i];
>>>> +         if (view->resource) {
>>>> +            struct nv04_resource *res = nv04_resource(view->resource);
>>>> +
>>>> +            nve4_set_surface_info(push, view, screen);
>>>> +            BCTX_REFN(nvc0->bufctx_3d, 3D_SUF, res, RDWR);
>>>> +         } else {
>>>> +            for (j = 0; j < 16; j++)
>>>> +               PUSH_DATA(push, 0);
>>>> +         }
>>>> +      }
>>>> +   }
>>>>   }
>>>>
>>>>   void
>>>> --
>>>> 2.8.0
>>>>
>>>> _______________________________________________
>>>> mesa-dev mailing list
>>>> mesa-dev at lists.freedesktop.org
>>>> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>> _______________________________________________
>> mesa-dev mailing list
>> mesa-dev at lists.freedesktop.org
>> https://lists.freedesktop.org/mailman/listinfo/mesa-dev

-- 
-Samuel