Render YUV400 image on Ivybridge

Signed-off-by: Xiang, Haihao <haihao.xiang@intel.com>

diff --git a/src/i965_render.c b/src/i965_render.c
index 0c2efb8..626b1e6 100644 (file)
--- a/src/i965_render.c
+++ b/src/i965_render.c
@@ -987,10 +987,16 @@ i965_render_upload_constants(VADriverContextP ctx,
     assert(render_state->curbe.bo->virtual);
     constant_buffer = render_state->curbe.bo->virtual;
 
-    if (obj_surface->fourcc == VA_FOURCC('N','V','1','2'))
-        *constant_buffer = 1;
-    else
-        *constant_buffer = 0;
+    if (obj_surface->subsampling == SUBSAMPLE_YUV400) {
+        assert(obj_surface->fourcc == VA_FOURCC('I', 'M', 'C', '1') ||
+               obj_surface->fourcc == VA_FOURCC('I', 'M', 'C', '3'));
+        *constant_buffer = 2;
+    } else {
+        if (obj_surface->fourcc == VA_FOURCC('N', 'V', '1', '2'))
+            *constant_buffer = 1;
+        else
+            *constant_buffer = 0;
+    }
 
     dri_bo_unmap(render_state->curbe.bo);
 }

diff --git a/src/shaders/render/exa_wm_src_sample_planar.g7a b/src/shaders/render/exa_wm_src_sample_planar.g7a
index 23d880f..8a47fb5 100644 (file)
--- a/src/shaders/render/exa_wm_src_sample_planar.g7a
+++ b/src/shaders/render/exa_wm_src_sample_planar.g7a
@@ -33,15 +33,18 @@ include(`exa_wm.g4i')
 define(`src_msg_gen7',   `g65')
 define(`src_msg_ind_gen7',`65')
 /* UV flag */
-define(`nv12', `g6.0<0,1,0>UW')
+define(`uv_flag', `g6.0<0,1,0>UW')
 
 /* prepare sampler read back gX register, which would be written back to output */
 
 /* use simd16 sampler, param 0 is u, param 1 is v. */
 /* 'payload' loading, assuming tex coord start from g4 */
-cmp.g.f0.0 (1) null nv12 0x0UW {align1};
+cmp.e.f0.0 (1) null uv_flag 0x1UW {align1};
 (f0.0) jmpi INTERLEAVED_UV;
 
+cmp.e.f0.0 (1) null uv_flag 0x2UW {align1};
+(f0.0) jmpi CONSTANT_UV;
+
 /* load r */
 mov (1) g0.8<1>UD               0x0000e000UD { align1 mask_disable };
 mov (8) src_msg_gen7<1>UD       g0<8,8,1>UD  { align1 mask_disable };
@@ -69,6 +72,12 @@ send (16) src_msg_ind_gen7           /* msg reg index */
 
 jmpi SAMPLE_Y;
 
+CONSTANT_UV:
+mov (16)        src_sample_g<1>f        0.5f { compr align1 mask_disable };
+mov (16)        src_sample_b<1>f        0.5f { compr align1 mask_disable };
+
+jmpi SAMPLE_Y;
+
 INTERLEAVED_UV:
 mov (1) g0.8<1>UD       0x0000c000UD { align1 mask_disable };
 mov (8) src_msg_gen7<1>UD    g0<8,8,1>UD  { align1 mask_disable };
@@ -81,6 +90,7 @@ send (16) src_msg_ind_gen7           /* msg reg index */
                                 /* here(src->dst) we should use src_sampler and src_surface */
         mlen 5 rlen 4 { align1 };   /* required message len 5, readback len 8 */
 
+
 SAMPLE_Y:
 mov (1) g0.8<1>UD      0x0000e000UD { align1 mask_disable };
 mov (8) src_msg_gen7<1>UD      g0<8,8,1>UD  { align1 mask_disable };

diff --git a/src/shaders/render/exa_wm_src_sample_planar.g7b b/src/shaders/render/exa_wm_src_sample_planar.g7b
index 617ecf2..598bd66 100644 (file)
--- a/src/shaders/render/exa_wm_src_sample_planar.g7b
+++ b/src/shaders/render/exa_wm_src_sample_planar.g7b
@@ -1,4 +1,6 @@
-   { 0x03000010, 0x20002d3c, 0x000000c0, 0x00000000 },
+   { 0x01000010, 0x20002d3c, 0x000000c0, 0x00010001 },
+   { 0x00010020, 0x34001c00, 0x00001400, 0x00000018 },
+   { 0x01000010, 0x20002d3c, 0x000000c0, 0x00020002 },
    { 0x00010020, 0x34001c00, 0x00001400, 0x0000000e },
    { 0x00000201, 0x20080061, 0x00000000, 0x0000e000 },
    { 0x00600201, 0x28200021, 0x008d0000, 0x00000000 },
@@ -6,6 +8,9 @@
    { 0x00000201, 0x20080061, 0x00000000, 0x0000e000 },
    { 0x00600201, 0x28200021, 0x008d0000, 0x00000000 },
    { 0x02800031, 0x22401ca9, 0x00000820, 0x0a2c0405 },
+   { 0x00000020, 0x34001c00, 0x00001400, 0x0000000c },
+   { 0x00800201, 0x220003fd, 0x00000000, 0x3f000000 },
+   { 0x00800201, 0x224003fd, 0x00000000, 0x3f000000 },
    { 0x00000020, 0x34001c00, 0x00001400, 0x00000006 },
    { 0x00000201, 0x20080061, 0x00000000, 0x0000c000 },
    { 0x00600201, 0x28200021, 0x008d0000, 0x00000000 },