xf86-video-ati: Branch 'r6xx-r7xx-support' - 2 commits
Alex Deucher
agd5f at kemper.freedesktop.org
Wed Feb 11 12:32:04 PST 2009
src/r600_exa.c | 554 +++++++++++++++++++++++++----------------
src/r600_textured_videofuncs.c | 15 -
2 files changed, 352 insertions(+), 217 deletions(-)
New commits:
commit bd141aa73a77f68301715fb3b5664e2082327a80
Author: Alex Deucher <alexdeucher at gmail.com>
Date: Wed Feb 11 15:31:47 2009 -0500
R6xx/R7xx Xv: switch packed over to Yang's new shader code
diff --git a/src/r600_exa.c b/src/r600_exa.c
index 34e67d8..542d42d 100644
--- a/src/r600_exa.c
+++ b/src/r600_exa.c
@@ -2841,7 +2841,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
// 2
ps[i++] = CF_ALLOC_IMP_EXP_DWORD0(ARRAY_BASE(CF_PIXEL_MRT0),
TYPE(SQ_EXPORT_PIXEL),
- RW_GPR(3),
+ RW_GPR(2),
RW_REL(ABSOLUTE),
INDEX_GPR(0),
ELEM_SIZE(3));
@@ -2856,96 +2856,88 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
CF_INST(SQ_CF_INST_EXPORT_DONE),
WHOLE_QUAD_MODE(0),
BARRIER(1));
+ /* Undo scaling of Y'CbCr values
+ * Y' is scaled from 16:235
+ * Cb/Cr are scaled from 16:240
+ */
// 3 - alu 0
- // DP4 gpr[2].x gpr[1].x c[0].x
+ // MULADD gpr[1].x gpr[1].x c[3].x c[3].y
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(259),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
LAST(0));
- ps[i++] = ALU_DWORD1_OP2(info->ChipFamily,
- SRC0_ABS(0),
- SRC1_ABS(0),
- UPDATE_EXECUTE_MASK(0),
- UPDATE_PRED(0),
- WRITE_MASK(1),
- FOG_MERGE(0),
- OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_102),
- DST_GPR(2),
+ ps[i++] = ALU_DWORD1_OP3(SRC2_SEL(259),
+ SRC2_REL(ABSOLUTE),
+ SRC2_ELEM(ELEM_Y),
+ SRC2_NEG(0),
+ ALU_INST(SQ_OP3_INST_MULADD),
+ BANK_SWIZZLE(SQ_ALU_VEC_012),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_X),
CLAMP(1));
// 4 - alu 1
- // DP4 gpr[2].y gpr[1].y c[0].y
+ // MULADD gpr[1].y gpr[1].y c[3].z c[3].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(259),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_Y),
+ SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
LAST(0));
- ps[i++] = ALU_DWORD1_OP2(info->ChipFamily,
- SRC0_ABS(0),
- SRC1_ABS(0),
- UPDATE_EXECUTE_MASK(0),
- UPDATE_PRED(0),
- WRITE_MASK(0),
- FOG_MERGE(0),
- OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_102),
- DST_GPR(2),
+ ps[i++] = ALU_DWORD1_OP3(SRC2_SEL(259),
+ SRC2_REL(ABSOLUTE),
+ SRC2_ELEM(ELEM_W),
+ SRC2_NEG(0),
+ ALU_INST(SQ_OP3_INST_MULADD),
+ BANK_SWIZZLE(SQ_ALU_VEC_012),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Y),
- CLAMP(1));
+ CLAMP(0));
// 5 - alu 2
- // DP4 gpr[2].z gpr[1].z c[0].z
+ // MULADD gpr[1].z gpr[1].z c[3].z c[3].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(259),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
LAST(0));
- ps[i++] = ALU_DWORD1_OP2(info->ChipFamily,
- SRC0_ABS(0),
- SRC1_ABS(0),
- UPDATE_EXECUTE_MASK(0),
- UPDATE_PRED(0),
- WRITE_MASK(0),
- FOG_MERGE(0),
- OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_102),
- DST_GPR(2),
+ ps[i++] = ALU_DWORD1_OP3(SRC2_SEL(259),
+ SRC2_REL(ABSOLUTE),
+ SRC2_ELEM(ELEM_W),
+ SRC2_NEG(0),
+ ALU_INST(SQ_OP3_INST_MULADD),
+ BANK_SWIZZLE(SQ_ALU_VEC_012),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Z),
- CLAMP(1));
+ CLAMP(0));
// 6 - alu 3
- // DP4 gpr[2].w gpr[1].w c[0].w
- ps[i++] = ALU_DWORD0(SRC0_SEL(1),
+ // MOV gpr[1].w 0.0
+ ps[i++] = ALU_DWORD0(SRC0_SEL(SQ_ALU_SRC_0),
SRC0_REL(ABSOLUTE),
- SRC0_ELEM(ELEM_W),
+ SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(SQ_ALU_SRC_0),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_W),
+ SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -2955,22 +2947,22 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(0),
+ WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_021),
- DST_GPR(2),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_W),
- CLAMP(1));
+ CLAMP(0));
// 7 - alu 4
- // DP4 gpr[2].x gpr[1].x c[1].x
+ // DP4 gpr[2].x gpr[1].x c[0].x
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
@@ -2982,7 +2974,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(0),
+ WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -2992,12 +2984,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_X),
CLAMP(1));
// 8 - alu 5
- // DP4 gpr[2].y gpr[1].y c[1].y
+ // DP4 gpr[2].y gpr[1].y c[0].y
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Y),
SRC1_NEG(0),
@@ -3009,7 +3001,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3019,12 +3011,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Y),
CLAMP(1));
// 9 - alu 6
- // DP4 gpr[2].z gpr[1].z c[1].z
+ // DP4 gpr[2].z gpr[1].z c[0].z
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
@@ -3046,12 +3038,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Z),
CLAMP(1));
// 10 - alu 7
- // DP4 gpr[2].w gpr[1].w c[1].w
+ // DP4 gpr[2].w gpr[1].w c[0].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_W),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_W),
SRC1_NEG(0),
@@ -3073,12 +3065,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_W),
CLAMP(1));
// 11 - alu 8
- // DP4 gpr[2].x gpr[1].x c[2].x
+ // DP4 gpr[2].x gpr[1].x c[1].x
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
@@ -3100,12 +3092,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_X),
CLAMP(1));
// 12 - alu 9
- // DP4 gpr[2].y gpr[1].y c[2].y
+ // DP4 gpr[2].y gpr[1].y c[1].y
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Y),
SRC1_NEG(0),
@@ -3117,7 +3109,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(0),
+ WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3127,12 +3119,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Y),
CLAMP(1));
// 13 - alu 10
- // DP4 gpr[2].z gpr[1].z c[2].z
+ // DP4 gpr[2].z gpr[1].z c[1].z
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
@@ -3144,7 +3136,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3154,12 +3146,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Z),
CLAMP(1));
// 14 - alu 11
- // DP4 gpr[2].w gpr[1].w c[2].w
+ // DP4 gpr[2].w gpr[1].w c[1].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_W),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_W),
SRC1_NEG(0),
@@ -3181,12 +3173,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_W),
CLAMP(1));
// 15 - alu 12
- // MOV gpr[3].x gpr[2].x
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].x gpr[1].x c[2].x
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
@@ -3198,24 +3190,24 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_210),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_X),
- CLAMP(0));
+ CLAMP(1));
// 16 - alu 13
- // MOV gpr[3].y gpr[2].y
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].y gpr[1].y c[2].y
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_X),
+ SRC1_ELEM(ELEM_Y),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3225,24 +3217,24 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_210),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Y),
- CLAMP(0));
+ CLAMP(1));
// 17 - alu 14
- // MOV gpr[3].z gpr[2].z
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].z gpr[1].z c[2].z
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_X),
+ SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3255,21 +3247,21 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_210),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Z),
- CLAMP(0));
+ CLAMP(1));
// 18 - alu 15
- // MOV gpr[3].w gpr[2].w
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].w gpr[1].w c[2].w
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_W),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_X),
+ SRC1_ELEM(ELEM_W),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3279,15 +3271,15 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_012),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_021),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_W),
- CLAMP(0));
+ CLAMP(1));
// 19 - alignment
ps[i++] = 0x00000000;
ps[i++] = 0x00000000;
diff --git a/src/r600_textured_videofuncs.c b/src/r600_textured_videofuncs.c
index 4a7391c..c06512a 100644
--- a/src/r600_textured_videofuncs.c
+++ b/src/r600_textured_videofuncs.c
@@ -202,7 +202,7 @@ R600DisplayTexturedVideo(ScrnInfoPtr pScrn, RADEONPortPrivPtr pPriv)
accel_state->ps_size, accel_state->ps_mc_addr);
ps_conf.shader_addr = accel_state->ps_mc_addr;
- ps_conf.num_gprs = 4;
+ ps_conf.num_gprs = 3;
ps_conf.stack_size = 0;
ps_conf.uncached_first_inst = 1;
ps_conf.clamp_consts = 0;
commit cb4a1ceaa32d6847b146a31291772e1183972ee7
Author: Yang Zhao <yang at yangman.ca>
Date: Wed Feb 11 15:18:00 2009 -0500
R6xx/R7xx Xv: Planar - Properly scale Y'CbCr values before converting to RGB
According to MPEG-2 spec, Y' and Cb/Cr values are scaled to [16, 235]
and [16, 240], respectively, when packed into bytes. Properly take care
of the reverse scaling before translating to RGB.
Conversion matrix has been simplified to remove 3rd column, as the fitting
to [-0.5, 0.5] can be done with scaling.
Redundant MOV instructions were also removed, and now only 3 GPRs are required.
diff --git a/src/r600_exa.c b/src/r600_exa.c
index a0f227b..34e67d8 100644
--- a/src/r600_exa.c
+++ b/src/r600_exa.c
@@ -3349,7 +3349,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
ps[i++] = TEX_DWORD_PAD;
// xv ps planar ----------------------------------
- i = accel_state->xv_ps_offset_planar / 4;
+ i = accel_state->xv_ps_offset_planar / 4;
// 0
ps[i++] = CF_DWORD0(ADDR(20));
ps[i++] = CF_DWORD1(POP_COUNT(0),
@@ -3378,7 +3378,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
// 2
ps[i++] = CF_ALLOC_IMP_EXP_DWORD0(ARRAY_BASE(CF_PIXEL_MRT0),
TYPE(SQ_EXPORT_PIXEL),
- RW_GPR(3),
+ RW_GPR(2),
RW_REL(ABSOLUTE),
INDEX_GPR(0),
ELEM_SIZE(3));
@@ -3393,96 +3393,88 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
CF_INST(SQ_CF_INST_EXPORT_DONE),
WHOLE_QUAD_MODE(0),
BARRIER(1));
+ /* Undo scaling of Y'CbCr values
+ * Y' is scaled from 16:235
+ * Cb/Cr are scaled from 16:240
+ */
// 3 - alu 0
- // DP4 gpr[2].x gpr[1].x c[0].x
+ // MULADD gpr[1].x gpr[1].x c[3].x c[3].y
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(259),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
LAST(0));
- ps[i++] = ALU_DWORD1_OP2(info->ChipFamily,
- SRC0_ABS(0),
- SRC1_ABS(0),
- UPDATE_EXECUTE_MASK(0),
- UPDATE_PRED(0),
- WRITE_MASK(1),
- FOG_MERGE(0),
- OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_102),
- DST_GPR(2),
+ ps[i++] = ALU_DWORD1_OP3(SRC2_SEL(259),
+ SRC2_REL(ABSOLUTE),
+ SRC2_ELEM(ELEM_Y),
+ SRC2_NEG(0),
+ ALU_INST(SQ_OP3_INST_MULADD),
+ BANK_SWIZZLE(SQ_ALU_VEC_012),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_X),
CLAMP(1));
// 4 - alu 1
- // DP4 gpr[2].y gpr[1].y c[0].y
+ // MULADD gpr[1].y gpr[1].y c[3].z c[3].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(259),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_Y),
+ SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
LAST(0));
- ps[i++] = ALU_DWORD1_OP2(info->ChipFamily,
- SRC0_ABS(0),
- SRC1_ABS(0),
- UPDATE_EXECUTE_MASK(0),
- UPDATE_PRED(0),
- WRITE_MASK(0),
- FOG_MERGE(0),
- OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_102),
- DST_GPR(2),
+ ps[i++] = ALU_DWORD1_OP3(SRC2_SEL(259),
+ SRC2_REL(ABSOLUTE),
+ SRC2_ELEM(ELEM_W),
+ SRC2_NEG(0),
+ ALU_INST(SQ_OP3_INST_MULADD),
+ BANK_SWIZZLE(SQ_ALU_VEC_012),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Y),
- CLAMP(1));
+ CLAMP(0));
// 5 - alu 2
- // DP4 gpr[2].z gpr[1].z c[0].z
+ // MULADD gpr[1].z gpr[1].z c[3].z c[3].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(259),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
LAST(0));
- ps[i++] = ALU_DWORD1_OP2(info->ChipFamily,
- SRC0_ABS(0),
- SRC1_ABS(0),
- UPDATE_EXECUTE_MASK(0),
- UPDATE_PRED(0),
- WRITE_MASK(0),
- FOG_MERGE(0),
- OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_102),
- DST_GPR(2),
+ ps[i++] = ALU_DWORD1_OP3(SRC2_SEL(259),
+ SRC2_REL(ABSOLUTE),
+ SRC2_ELEM(ELEM_W),
+ SRC2_NEG(0),
+ ALU_INST(SQ_OP3_INST_MULADD),
+ BANK_SWIZZLE(SQ_ALU_VEC_012),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Z),
- CLAMP(1));
+ CLAMP(0));
// 6 - alu 3
- // DP4 gpr[2].w gpr[1].w c[0].w
- ps[i++] = ALU_DWORD0(SRC0_SEL(1),
+ // MOV gpr[1].w 0.0
+ ps[i++] = ALU_DWORD0(SRC0_SEL(SQ_ALU_SRC_0),
SRC0_REL(ABSOLUTE),
- SRC0_ELEM(ELEM_W),
+ SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(256),
+ SRC1_SEL(SQ_ALU_SRC_0),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_W),
+ SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3492,22 +3484,22 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(0),
+ WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
- BANK_SWIZZLE(SQ_ALU_VEC_021),
- DST_GPR(2),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(1),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_W),
- CLAMP(1));
+ CLAMP(0));
// 7 - alu 4
- // DP4 gpr[2].x gpr[1].x c[1].x
+ // DP4 gpr[2].x gpr[1].x c[0].x
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
@@ -3519,7 +3511,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(0),
+ WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3529,12 +3521,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_X),
CLAMP(1));
// 8 - alu 5
- // DP4 gpr[2].y gpr[1].y c[1].y
+ // DP4 gpr[2].y gpr[1].y c[0].y
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Y),
SRC1_NEG(0),
@@ -3546,7 +3538,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3556,12 +3548,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Y),
CLAMP(1));
// 9 - alu 6
- // DP4 gpr[2].z gpr[1].z c[1].z
+ // DP4 gpr[2].z gpr[1].z c[0].z
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
@@ -3583,12 +3575,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Z),
CLAMP(1));
// 10 - alu 7
- // DP4 gpr[2].w gpr[1].w c[1].w
+ // DP4 gpr[2].w gpr[1].w c[0].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_W),
SRC0_NEG(0),
- SRC1_SEL(257),
+ SRC1_SEL(256),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_W),
SRC1_NEG(0),
@@ -3610,12 +3602,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_W),
CLAMP(1));
// 11 - alu 8
- // DP4 gpr[2].x gpr[1].x c[2].x
+ // DP4 gpr[2].x gpr[1].x c[1].x
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
@@ -3637,12 +3629,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_X),
CLAMP(1));
// 12 - alu 9
- // DP4 gpr[2].y gpr[1].y c[2].y
+ // DP4 gpr[2].y gpr[1].y c[1].y
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Y),
SRC1_NEG(0),
@@ -3654,7 +3646,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(0),
+ WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3664,12 +3656,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Y),
CLAMP(1));
// 13 - alu 10
- // DP4 gpr[2].z gpr[1].z c[2].z
+ // DP4 gpr[2].z gpr[1].z c[1].z
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
@@ -3681,7 +3673,7 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
ALU_INST(SQ_OP2_INST_DOT4),
@@ -3691,12 +3683,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_Z),
CLAMP(1));
// 14 - alu 11
- // DP4 gpr[2].w gpr[1].w c[2].w
+ // DP4 gpr[2].w gpr[1].w c[1].w
ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_W),
SRC0_NEG(0),
- SRC1_SEL(258),
+ SRC1_SEL(257),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_W),
SRC1_NEG(0),
@@ -3718,12 +3710,12 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
DST_ELEM(ELEM_W),
CLAMP(1));
// 15 - alu 12
- // MOV gpr[3].x gpr[2].x
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].x gpr[1].x c[2].x
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_X),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
SRC1_ELEM(ELEM_X),
SRC1_NEG(0),
@@ -3735,24 +3727,24 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_210),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_X),
- CLAMP(0));
+ CLAMP(1));
// 16 - alu 13
- // MOV gpr[3].y gpr[2].y
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].y gpr[1].y c[2].y
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Y),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_X),
+ SRC1_ELEM(ELEM_Y),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3762,24 +3754,24 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_210),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Y),
- CLAMP(0));
+ CLAMP(1));
// 17 - alu 14
- // MOV gpr[3].z gpr[2].z
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].z gpr[1].z c[2].z
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_Z),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_X),
+ SRC1_ELEM(ELEM_Z),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3792,21 +3784,21 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
WRITE_MASK(1),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_210),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_102),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_Z),
- CLAMP(0));
+ CLAMP(1));
// 18 - alu 15
- // MOV gpr[3].w gpr[2].w
- ps[i++] = ALU_DWORD0(SRC0_SEL(2),
+ // DP4 gpr[2].w gpr[1].w c[2].w
+ ps[i++] = ALU_DWORD0(SRC0_SEL(1),
SRC0_REL(ABSOLUTE),
SRC0_ELEM(ELEM_W),
SRC0_NEG(0),
- SRC1_SEL(0),
+ SRC1_SEL(258),
SRC1_REL(ABSOLUTE),
- SRC1_ELEM(ELEM_X),
+ SRC1_ELEM(ELEM_W),
SRC1_NEG(0),
INDEX_MODE(SQ_INDEX_LOOP),
PRED_SEL(SQ_PRED_SEL_OFF),
@@ -3816,15 +3808,15 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
SRC1_ABS(0),
UPDATE_EXECUTE_MASK(0),
UPDATE_PRED(0),
- WRITE_MASK(1),
+ WRITE_MASK(0),
FOG_MERGE(0),
OMOD(SQ_ALU_OMOD_OFF),
- ALU_INST(SQ_OP2_INST_MOV),
- BANK_SWIZZLE(SQ_ALU_VEC_012),
- DST_GPR(3),
+ ALU_INST(SQ_OP2_INST_DOT4),
+ BANK_SWIZZLE(SQ_ALU_VEC_021),
+ DST_GPR(2),
DST_REL(ABSOLUTE),
DST_ELEM(ELEM_W),
- CLAMP(0));
+ CLAMP(1));
// 19 - alignment
ps[i++] = 0x00000000;
ps[i++] = 0x00000000;
@@ -4057,6 +4049,150 @@ R600LoadShaders(ScrnInfoPtr pScrn, ScreenPtr pScreen)
MEGA_FETCH(0));
vs[i++] = VTX_DWORD_PAD;
+ // comp mask vs ---------------------------------------
+ i = accel_state->comp_mask_vs_offset / 4;
+ //0
+ vs[i++] = CF_DWORD0(ADDR(4));
+ vs[i++] = CF_DWORD1(POP_COUNT(0),
+ CF_CONST(0),
+ COND(SQ_CF_COND_ACTIVE),
+ I_COUNT(3),
+ CALL_COUNT(0),
+ END_OF_PROGRAM(0),
+ VALID_PIXEL_MODE(0),
+ CF_INST(SQ_CF_INST_VTX),
+ WHOLE_QUAD_MODE(0),
+ BARRIER(1));
+ //1 - dst
+ vs[i++] = CF_ALLOC_IMP_EXP_DWORD0(ARRAY_BASE(CF_POS0),
+ TYPE(SQ_EXPORT_POS),
+ RW_GPR(2),
+ RW_REL(ABSOLUTE),
+ INDEX_GPR(0),
+ ELEM_SIZE(0));
+ vs[i++] = CF_ALLOC_IMP_EXP_DWORD1_SWIZ(SRC_SEL_X(SQ_SEL_X),
+ SRC_SEL_Y(SQ_SEL_Y),
+ SRC_SEL_Z(SQ_SEL_Z),
+ SRC_SEL_W(SQ_SEL_W),
+ R6xx_ELEM_LOOP(0),
+ BURST_COUNT(1),
+ END_OF_PROGRAM(0),
+ VALID_PIXEL_MODE(0),
+ CF_INST(SQ_CF_INST_EXPORT_DONE),
+ WHOLE_QUAD_MODE(0),
+ BARRIER(1));
+ //2 - src
+ vs[i++] = CF_ALLOC_IMP_EXP_DWORD0(ARRAY_BASE(0),
+ TYPE(SQ_EXPORT_PARAM),
+ RW_GPR(1),
+ RW_REL(ABSOLUTE),
+ INDEX_GPR(0),
+ ELEM_SIZE(0));
+ vs[i++] = CF_ALLOC_IMP_EXP_DWORD1_SWIZ(SRC_SEL_X(SQ_SEL_X),
+ SRC_SEL_Y(SQ_SEL_Y),
+ SRC_SEL_Z(SQ_SEL_Z),
+ SRC_SEL_W(SQ_SEL_W),
+ R6xx_ELEM_LOOP(0),
+ BURST_COUNT(1),
+ END_OF_PROGRAM(0),
+ VALID_PIXEL_MODE(0),
+ CF_INST(SQ_CF_INST_EXPORT),
+ WHOLE_QUAD_MODE(0),
+ BARRIER(0));
+ //3 - mask
+ vs[i++] = CF_ALLOC_IMP_EXP_DWORD0(ARRAY_BASE(1),
+ TYPE(SQ_EXPORT_PARAM),
+ RW_GPR(0),
+ RW_REL(ABSOLUTE),
+ INDEX_GPR(0),
+ ELEM_SIZE(0));
+ vs[i++] = CF_ALLOC_IMP_EXP_DWORD1_SWIZ(SRC_SEL_X(SQ_SEL_X),
+ SRC_SEL_Y(SQ_SEL_Y),
+ SRC_SEL_Z(SQ_SEL_Z),
+ SRC_SEL_W(SQ_SEL_W),
+ R6xx_ELEM_LOOP(0),
+ BURST_COUNT(1),
+ END_OF_PROGRAM(1),
+ VALID_PIXEL_MODE(0),
+ CF_INST(SQ_CF_INST_EXPORT_DONE),
+ WHOLE_QUAD_MODE(0),
+ BARRIER(0));
+ //4/5 - dst
+ vs[i++] = VTX_DWORD0(VTX_INST(SQ_VTX_INST_FETCH),
+ FETCH_TYPE(SQ_VTX_FETCH_VERTEX_DATA),
+ FETCH_WHOLE_QUAD(0),
+ BUFFER_ID(0),
+ SRC_GPR(0),
+ SRC_REL(ABSOLUTE),
+ SRC_SEL_X(SQ_SEL_X),
+ MEGA_FETCH_COUNT(24));
+ vs[i++] = VTX_DWORD1_GPR(DST_GPR(2),
+ DST_REL(0),
+ DST_SEL_X(SQ_SEL_X),
+ DST_SEL_Y(SQ_SEL_Y),
+ DST_SEL_Z(SQ_SEL_0),
+ DST_SEL_W(SQ_SEL_1),
+ USE_CONST_FIELDS(0),
+ DATA_FORMAT(FMT_32_32_FLOAT), //xxx
+ NUM_FORMAT_ALL(SQ_NUM_FORMAT_NORM), //xxx
+ FORMAT_COMP_ALL(SQ_FORMAT_COMP_SIGNED), //xxx
+ SRF_MODE_ALL(SRF_MODE_ZERO_CLAMP_MINUS_ONE));
+ vs[i++] = VTX_DWORD2(OFFSET(0),
+ ENDIAN_SWAP(ENDIAN_NONE),
+ CONST_BUF_NO_STRIDE(0),
+ MEGA_FETCH(1));
+ vs[i++] = VTX_DWORD_PAD;
+ //6/7 - src
+ vs[i++] = VTX_DWORD0(VTX_INST(SQ_VTX_INST_FETCH),
+ FETCH_TYPE(SQ_VTX_FETCH_VERTEX_DATA),
+ FETCH_WHOLE_QUAD(0),
+ BUFFER_ID(0),
+ SRC_GPR(0),
+ SRC_REL(ABSOLUTE),
+ SRC_SEL_X(SQ_SEL_X),
+ MEGA_FETCH_COUNT(8));
+ vs[i++] = VTX_DWORD1_GPR(DST_GPR(1),
+ DST_REL(0),
+ DST_SEL_X(SQ_SEL_X),
+ DST_SEL_Y(SQ_SEL_Y),
+ DST_SEL_Z(SQ_SEL_0),
+ DST_SEL_W(SQ_SEL_1),
+ USE_CONST_FIELDS(0),
+ DATA_FORMAT(FMT_32_32_FLOAT), //xxx
+ NUM_FORMAT_ALL(SQ_NUM_FORMAT_NORM), //xxx
+ FORMAT_COMP_ALL(SQ_FORMAT_COMP_SIGNED), //xxx
+ SRF_MODE_ALL(SRF_MODE_ZERO_CLAMP_MINUS_ONE));
+ vs[i++] = VTX_DWORD2(OFFSET(8),
+ ENDIAN_SWAP(ENDIAN_NONE),
+ CONST_BUF_NO_STRIDE(0),
+ MEGA_FETCH(0));
+ vs[i++] = VTX_DWORD_PAD;
+ //8/9 - mask
+ vs[i++] = VTX_DWORD0(VTX_INST(SQ_VTX_INST_FETCH),
+ FETCH_TYPE(SQ_VTX_FETCH_VERTEX_DATA),
+ FETCH_WHOLE_QUAD(0),
+ BUFFER_ID(0),
+ SRC_GPR(0),
+ SRC_REL(ABSOLUTE),
+ SRC_SEL_X(SQ_SEL_X),
+ MEGA_FETCH_COUNT(8));
+ vs[i++] = VTX_DWORD1_GPR(DST_GPR(0),
+ DST_REL(0),
+ DST_SEL_X(SQ_SEL_X),
+ DST_SEL_Y(SQ_SEL_Y),
+ DST_SEL_Z(SQ_SEL_0),
+ DST_SEL_W(SQ_SEL_1),
+ USE_CONST_FIELDS(0),
+ DATA_FORMAT(FMT_32_32_FLOAT), //xxx
+ NUM_FORMAT_ALL(SQ_NUM_FORMAT_NORM), //xxx
+ FORMAT_COMP_ALL(SQ_FORMAT_COMP_SIGNED), //xxx
+ SRF_MODE_ALL(SRF_MODE_ZERO_CLAMP_MINUS_ONE));
+ vs[i++] = VTX_DWORD2(OFFSET(16),
+ ENDIAN_SWAP(ENDIAN_NONE),
+ CONST_BUF_NO_STRIDE(0),
+ MEGA_FETCH(0));
+ vs[i++] = VTX_DWORD_PAD;
+
// comp mask ps ---------------------------------------
// not yet
diff --git a/src/r600_textured_videofuncs.c b/src/r600_textured_videofuncs.c
index d5add19..4a7391c 100644
--- a/src/r600_textured_videofuncs.c
+++ b/src/r600_textured_videofuncs.c
@@ -118,9 +118,16 @@ R600DisplayTexturedVideo(ScrnInfoPtr pScrn, RADEONPortPrivPtr pPriv)
int uv_offset;
static float ps_alu_consts[] = {
- 1.0, 0.0, 1.13983, -1.13983/2, // r - c[0]
- 1.0, -0.39465, -0.5806, (0.39465+0.5806)/2, // g - c[1]
- 1.0, 2.03211, 0.0, -2.03211/2, // b - c[2]
+ 1.0, 0.0, 1.4020, 0, // r - c[0]
+ 1.0, -0.34414, -0.71414, 0, // g - c[1]
+ 1.0, 1.7720, 0.0, 0, // b - c[2]
+ /* Constants for undoing Y'CbCr scaling
+ * - Y' is scaled from 16:235
+ * - Cb/Cr are scaled from 16:240
+ * Unscaled value N' = N * N_mul + N_shift (N' in range [-0.5, 0.5])
+ * Vector is [Y_mul, Y_shfit, C_mul, C_shift]
+ */
+ 256.0/219.0, -16.0/219.0, 256.0/224.0, -128.0/224.0,
};
CLEAR (cb_conf);
More information about the xorg-commit
mailing list