thr3ads.net - Nouveau - [Nouveau] [PATCH] nv50: fix setting of texture ms info to be per-stage [May 2014]

If this information is useful, please help other people find it:
Share via:

Ilia Mirkin

2014-May-10 03:19 UTC

[Nouveau] [PATCH] nv50: fix setting of texture ms info to be per-stage

Different textures may be bound to each slot for each stage. So we need
to be able to upload ms parameters for each one without stages
overwriting each other.

Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu>
Cc: "10.1 10.2" <mesa-stable at lists.freedesktop.org>
---
 src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp |  4 ++++
 src/gallium/drivers/nouveau/nv50/nv50_context.h               | 10 +++++-----
 src/gallium/drivers/nouveau/nv50/nv50_tex.c                   |  2 +-
 3 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
index eafc0a7..63db1d7 100644
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
@@ -591,6 +591,10 @@ void NV50LoweringPreSSA::loadTexMsInfo(uint32_t off, Value
**ms,
    Value *tmp = new_LValue(func, FILE_GPR);
    uint8_t b = prog->driver->io.resInfoCBSlot;
    off += prog->driver->io.suInfoBase;
+   if (prog->getType() > Program::TYPE_VERTEX)
+      off += 16 * 2 * 4;
+   if (prog->getType() > Program::TYPE_GEOMETRY)
+      off += 16 * 2 * 4;
    *ms_x = bld.mkLoadv(TYPE_U32, bld.mkSymbol(
                              FILE_MEMORY_CONST, b, TYPE_U32, off + 0), NULL);
    *ms_y = bld.mkLoadv(TYPE_U32, bld.mkSymbol(
diff --git a/src/gallium/drivers/nouveau/nv50/nv50_context.h
b/src/gallium/drivers/nouveau/nv50/nv50_context.h
index b776dee..3b7cb18 100644
--- a/src/gallium/drivers/nouveau/nv50/nv50_context.h
+++ b/src/gallium/drivers/nouveau/nv50/nv50_context.h
@@ -78,16 +78,16 @@
 /* 8 user clip planes, at 4 32-bit floats each */
 #define NV50_CB_AUX_UCP_OFFSET    0x0000
 #define NV50_CB_AUX_UCP_SIZE      (8 * 4 * 4)
-/* 256 textures, each with ms_x, ms_y u32 pairs */
+/* 16 textures * 3 shaders, each with ms_x, ms_y u32 pairs */
 #define NV50_CB_AUX_TEX_MS_OFFSET 0x0080
-#define NV50_CB_AUX_TEX_MS_SIZE   (256 * 2 * 4)
+#define NV50_CB_AUX_TEX_MS_SIZE   (16 * 3 * 2 * 4)
 /* For each MS level (4), 8 sets of 32-bit integer pairs sample offsets */
-#define NV50_CB_AUX_MS_OFFSET     0x880
+#define NV50_CB_AUX_MS_OFFSET     0x200
 #define NV50_CB_AUX_MS_SIZE       (4 * 8 * 4 * 2)
 /* Sample position pairs for the current output MS level */
-#define NV50_CB_AUX_SAMPLE_OFFSET 0x980
+#define NV50_CB_AUX_SAMPLE_OFFSET 0x300
 #define NV50_CB_AUX_SAMPLE_OFFSET_SIZE (4 * 8 * 2)
-/* next spot: 0x9c0 */
+/* next spot: 0x340 */
 /* 4 32-bit floats for the vertex runout, put at the end */
 #define NV50_CB_AUX_RUNOUT_OFFSET (NV50_CB_AUX_SIZE - 0x10)
 
diff --git a/src/gallium/drivers/nouveau/nv50/nv50_tex.c
b/src/gallium/drivers/nouveau/nv50/nv50_tex.c
index 5cfce3a..0d2604e 100644
--- a/src/gallium/drivers/nouveau/nv50/nv50_tex.c
+++ b/src/gallium/drivers/nouveau/nv50/nv50_tex.c
@@ -286,7 +286,7 @@ nv50_validate_tic(struct nv50_context *nv50, int s)
    }
    if (nv50->num_textures[s]) {
       BEGIN_NV04(push, NV50_3D(CB_ADDR), 1);
-      PUSH_DATA (push, (NV50_CB_AUX_TEX_MS_OFFSET << (8 - 2)) |
NV50_CB_AUX);
+      PUSH_DATA (push, ((NV50_CB_AUX_TEX_MS_OFFSET + 16 * s * 2 * 4) <<
(8 - 2)) | NV50_CB_AUX);
       BEGIN_NI04(push, NV50_3D(CB_DATA(0)), nv50->num_textures[s] * 2);
       for (i = 0; i < nv50->num_textures[s]; i++) {
          struct nv50_tic_entry *tic = nv50_tic_entry(nv50->textures[s][i]);
-- 
1.8.5.5

Ben Skeggs

2014-May-11 23:21 UTC

head link

[Nouveau] [PATCH] nv50: fix setting of texture ms info to be per-stage

On Sat, May 10, 2014 at 1:19 PM, Ilia Mirkin <imirkin at alum.mit.edu>
wrote:> Different textures may be bound to each slot for each stage. So we need
> to be able to upload ms parameters for each one without stages
> overwriting each other.
>
> Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu>
> Cc: "10.1 10.2" <mesa-stable at lists.freedesktop.org>Reviewed-by: Ben Skeggs <bskeggs at redhat.com>
> ---
>  src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp |  4 ++++
>  src/gallium/drivers/nouveau/nv50/nv50_context.h               | 10
+++++-----
>  src/gallium/drivers/nouveau/nv50/nv50_tex.c                   |  2 +-
>  3 files changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
> index eafc0a7..63db1d7 100644
> --- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
> +++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nv50.cpp
> @@ -591,6 +591,10 @@ void NV50LoweringPreSSA::loadTexMsInfo(uint32_t off,
Value **ms,
>     Value *tmp = new_LValue(func, FILE_GPR);
>     uint8_t b = prog->driver->io.resInfoCBSlot;
>     off += prog->driver->io.suInfoBase;
> +   if (prog->getType() > Program::TYPE_VERTEX)
> +      off += 16 * 2 * 4;
> +   if (prog->getType() > Program::TYPE_GEOMETRY)
> +      off += 16 * 2 * 4;
>     *ms_x = bld.mkLoadv(TYPE_U32, bld.mkSymbol(
>                               FILE_MEMORY_CONST, b, TYPE_U32, off + 0),
NULL);
>     *ms_y = bld.mkLoadv(TYPE_U32, bld.mkSymbol(
> diff --git a/src/gallium/drivers/nouveau/nv50/nv50_context.h
b/src/gallium/drivers/nouveau/nv50/nv50_context.h
> index b776dee..3b7cb18 100644
> --- a/src/gallium/drivers/nouveau/nv50/nv50_context.h
> +++ b/src/gallium/drivers/nouveau/nv50/nv50_context.h
> @@ -78,16 +78,16 @@
>  /* 8 user clip planes, at 4 32-bit floats each */
>  #define NV50_CB_AUX_UCP_OFFSET    0x0000
>  #define NV50_CB_AUX_UCP_SIZE      (8 * 4 * 4)
> -/* 256 textures, each with ms_x, ms_y u32 pairs */
> +/* 16 textures * 3 shaders, each with ms_x, ms_y u32 pairs */
>  #define NV50_CB_AUX_TEX_MS_OFFSET 0x0080
> -#define NV50_CB_AUX_TEX_MS_SIZE   (256 * 2 * 4)
> +#define NV50_CB_AUX_TEX_MS_SIZE   (16 * 3 * 2 * 4)
>  /* For each MS level (4), 8 sets of 32-bit integer pairs sample offsets */
> -#define NV50_CB_AUX_MS_OFFSET     0x880
> +#define NV50_CB_AUX_MS_OFFSET     0x200
>  #define NV50_CB_AUX_MS_SIZE       (4 * 8 * 4 * 2)
>  /* Sample position pairs for the current output MS level */
> -#define NV50_CB_AUX_SAMPLE_OFFSET 0x980
> +#define NV50_CB_AUX_SAMPLE_OFFSET 0x300
>  #define NV50_CB_AUX_SAMPLE_OFFSET_SIZE (4 * 8 * 2)
> -/* next spot: 0x9c0 */
> +/* next spot: 0x340 */
>  /* 4 32-bit floats for the vertex runout, put at the end */
>  #define NV50_CB_AUX_RUNOUT_OFFSET (NV50_CB_AUX_SIZE - 0x10)
>
> diff --git a/src/gallium/drivers/nouveau/nv50/nv50_tex.c
b/src/gallium/drivers/nouveau/nv50/nv50_tex.c
> index 5cfce3a..0d2604e 100644
> --- a/src/gallium/drivers/nouveau/nv50/nv50_tex.c
> +++ b/src/gallium/drivers/nouveau/nv50/nv50_tex.c
> @@ -286,7 +286,7 @@ nv50_validate_tic(struct nv50_context *nv50, int s)
>     }
>     if (nv50->num_textures[s]) {
>        BEGIN_NV04(push, NV50_3D(CB_ADDR), 1);
> -      PUSH_DATA (push, (NV50_CB_AUX_TEX_MS_OFFSET << (8 - 2)) |
NV50_CB_AUX);
> +      PUSH_DATA (push, ((NV50_CB_AUX_TEX_MS_OFFSET + 16 * s * 2 * 4)
<< (8 - 2)) | NV50_CB_AUX);
>        BEGIN_NI04(push, NV50_3D(CB_DATA(0)), nv50->num_textures[s] * 2);
>        for (i = 0; i < nv50->num_textures[s]; i++) {
>           struct nv50_tic_entry *tic =
nv50_tic_entry(nv50->textures[s][i]);
> --
> 1.8.5.5
>
> _______________________________________________
> Nouveau mailing list
> Nouveau at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/nouveau

Carl Worth

2014-May-14 00:44 UTC

head link

[Nouveau] [Mesa-stable] [PATCH] nv50: fix setting of texture ms info to be per-stage

Ilia Mirkin <imirkin at alum.mit.edu> writes:> Cc: "10.1 10.2" <mesa-stable at lists.freedesktop.org>
...>  /* Sample position pairs for the current output MS level */
> -#define NV50_CB_AUX_SAMPLE_OFFSET 0x980
> +#define NV50_CB_AUX_SAMPLE_OFFSET 0x300
>  #define NV50_CB_AUX_SAMPLE_OFFSET_SIZE (4 * 8 * 2)
> -/* next spot: 0x9c0 */
> +/* next spot: 0x340 */
>  /* 4 32-bit floats for the vertex runout, put at the end */
>  #define NV50_CB_AUX_RUNOUT_OFFSET (NV50_CB_AUX_SIZE - 0x10)
This hunk of the patch isn't applying cleanly to 10.1.

I think the conflict resolution should be obvious to an author or
competent reviewer, but I didn't trust myself to get it right.

If one of you can follow up with a patch that applies cleanly to the
10.1 branch, I'll be happy to apply it.

-Carl

-- 
carl.d.worth at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 818 bytes
Desc: not available
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20140513/5c1ccfe0/attachment.sig>

Maybe Matching Threads

Search for more seemingly similar threads

Nouveau - May 2014 - [PATCH] nv50: fix setting of texture ms info to be per-stage

[Nouveau] [PATCH] nv50: fix setting of texture ms info to be per-stage

[Nouveau] [PATCH] nv50: fix setting of texture ms info to be per-stage

[Nouveau] [Mesa-stable] [PATCH] nv50: fix setting of texture ms info to be per-stage

Maybe Matching Threads