i965: Split sampler count variable to be per-stage.

Currently, we only have a single sampler state table shared among all
stages, so we just copy wm.sampler_count into vs.sampler_count.

In the future, each shader stage will have its own SAMPLER_STATE table,
at which point we'll need these separate sampler counts.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Paul Berry <stereotype441@gmail.com>

9 files changed, 27 insertions, 20 deletions

diff --git a/src/mesa/drivers/dri/i965/brw_context.h b/src/mesa/drivers/dri/i965/brw_context.h
index 419cf17..d7c3472 100644
--- a/src/mesa/drivers/dri/i965/brw_context.h
+++ b/src/mesa/drivers/dri/i965/brw_context.h
@@ -1066,7 +1066,6 @@ struct brw_context
 
    /** SAMPLER_STATE count and offset */
    struct {
-      GLuint count;
       uint32_t offset;
    } sampler;
 
@@ -1110,6 +1109,8 @@ struct brw_context
 
       uint32_t bind_bo_offset;
       uint32_t surf_offset[BRW_MAX_VS_SURFACES];
+
+      uint32_t sampler_count;
    } vs;
 
    struct {
@@ -1183,6 +1184,8 @@ struct brw_context
       uint32_t bind_bo_offset;
       uint32_t surf_offset[BRW_MAX_WM_SURFACES];
 
+      uint32_t sampler_count;
+
       struct {
          struct ra_regs *regs;
 
diff --git a/src/mesa/drivers/dri/i965/brw_vs_state.c b/src/mesa/drivers/dri/i965/brw_vs_state.c
index ddaf914..13aabac 100644
--- a/src/mesa/drivers/dri/i965/brw_vs_state.c
+++ b/src/mesa/drivers/dri/i965/brw_vs_state.c
@@ -142,7 +142,7 @@ brw_upload_vs_unit(struct brw_context *brw)
       vs->vs5.sampler_count = 0; /* hardware requirement */
    else {
       /* CACHE_NEW_SAMPLER */
-      vs->vs5.sampler_count = (brw->sampler.count + 3) / 4;
+      vs->vs5.sampler_count = (brw->vs.sampler_count + 3) / 4;
    }
 
 
@@ -155,7 +155,7 @@ brw_upload_vs_unit(struct brw_context *brw)
 
    /* Set the sampler state pointer, and its reloc
     */
-   if (brw->sampler.count) {
+   if (brw->vs.sampler_count) {
       vs->vs5.sampler_state_pointer =
          (brw->batch.bo->offset + brw->sampler.offset) >> 5;
       drm_intel_bo_emit_reloc(brw->batch.bo,
diff --git a/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c b/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
index 5457671..40a6d5b 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_sampler_state.c
@@ -377,17 +377,19 @@ brw_upload_samplers(struct brw_context *brw)
    /* ARB programs use the texture unit number as the sampler index, so we
     * need to find the highest unit used.  A bit-count will not work.
     */
-   brw->sampler.count = _mesa_fls(SamplersUsed);
+   brw->wm.sampler_count = _mesa_fls(SamplersUsed);
+   /* Currently we only use one sampler state table.  Mirror the count. */
+   brw->vs.sampler_count = brw->wm.sampler_count;
 
-   if (brw->sampler.count == 0)
+   if (brw->wm.sampler_count == 0)
       return;
 
    samplers = brw_state_batch(brw, AUB_TRACE_SAMPLER_STATE,
-			      brw->sampler.count * sizeof(*samplers),
+			      brw->wm.sampler_count * sizeof(*samplers),
 			      32, &brw->sampler.offset);
-   memset(samplers, 0, brw->sampler.count * sizeof(*samplers));
+   memset(samplers, 0, brw->wm.sampler_count * sizeof(*samplers));
 
-   for (unsigned s = 0; s < brw->sampler.count; s++) {
+   for (unsigned s = 0; s < brw->wm.sampler_count; s++) {
       if (SamplersUsed & (1 << s)) {
          const unsigned unit = (fs->SamplersUsed & (1 << s)) ?
             fs->SamplerUnits[s] : vs->SamplerUnits[s];
diff --git a/src/mesa/drivers/dri/i965/brw_wm_state.c b/src/mesa/drivers/dri/i965/brw_wm_state.c
index 631f351..106d628 100644
--- a/src/mesa/drivers/dri/i965/brw_wm_state.c
+++ b/src/mesa/drivers/dri/i965/brw_wm_state.c
@@ -144,10 +144,10 @@ brw_upload_wm_unit(struct brw_context *brw)
       wm->wm4.sampler_count = 0; /* hardware requirement */
    else {
       /* CACHE_NEW_SAMPLER */
-      wm->wm4.sampler_count = (brw->sampler.count + 1) / 4;
+      wm->wm4.sampler_count = (brw->wm.sampler_count + 1) / 4;
    }
 
-   if (brw->sampler.count) {
+   if (brw->wm.sampler_count) {
       /* reloc */
       wm->wm4.sampler_state_pointer = (brw->batch.bo->offset +
 				       brw->sampler.offset) >> 5;
@@ -225,7 +225,7 @@ brw_upload_wm_unit(struct brw_context *brw)
    }
 
    /* Emit sampler state relocation */
-   if (brw->sampler.count != 0) {
+   if (brw->wm.sampler_count != 0) {
       drm_intel_bo_emit_reloc(brw->batch.bo,
 			      brw->wm.state_offset +
 			      offsetof(struct brw_wm_unit_state, wm4),
diff --git a/src/mesa/drivers/dri/i965/gen6_vs_state.c b/src/mesa/drivers/dri/i965/gen6_vs_state.c
index da20713..4af7cda 100644
--- a/src/mesa/drivers/dri/i965/gen6_vs_state.c
+++ b/src/mesa/drivers/dri/i965/gen6_vs_state.c
@@ -149,7 +149,7 @@ upload_vs_state(struct brw_context *brw)
    OUT_BATCH(_3DSTATE_VS << 16 | (6 - 2));
    OUT_BATCH(brw->vs.prog_offset);
    OUT_BATCH(floating_point_mode |
-	     ((ALIGN(brw->sampler.count, 4)/4) << GEN6_VS_SAMPLER_COUNT_SHIFT));
+	     ((ALIGN(brw->vs.sampler_count, 4)/4) << GEN6_VS_SAMPLER_COUNT_SHIFT));
 
    if (brw->vs.prog_data->base.total_scratch) {
       OUT_RELOC(brw->vs.scratch_bo,
diff --git a/src/mesa/drivers/dri/i965/gen6_wm_state.c b/src/mesa/drivers/dri/i965/gen6_wm_state.c
index 24c96cc..e286785 100644
--- a/src/mesa/drivers/dri/i965/gen6_wm_state.c
+++ b/src/mesa/drivers/dri/i965/gen6_wm_state.c
@@ -140,7 +140,7 @@ upload_wm_state(struct brw_context *brw)
       dw2 |= GEN6_WM_FLOATING_POINT_MODE_ALT;
 
    /* CACHE_NEW_SAMPLER */
-   dw2 |= (ALIGN(brw->sampler.count, 4) / 4) << GEN6_WM_SAMPLER_COUNT_SHIFT;
+   dw2 |= (ALIGN(brw->wm.sampler_count, 4) / 4) << GEN6_WM_SAMPLER_COUNT_SHIFT;
    dw4 |= (brw->wm.prog_data->first_curbe_grf <<
 	   GEN6_WM_DISPATCH_START_GRF_SHIFT_0);
    dw4 |= (brw->wm.prog_data->first_curbe_grf_16 <<
diff --git a/src/mesa/drivers/dri/i965/gen7_sampler_state.c b/src/mesa/drivers/dri/i965/gen7_sampler_state.c
index 7ce58ce..f09c6b3 100644
--- a/src/mesa/drivers/dri/i965/gen7_sampler_state.c
+++ b/src/mesa/drivers/dri/i965/gen7_sampler_state.c
@@ -195,17 +195,19 @@ gen7_upload_samplers(struct brw_context *brw)
 
    GLbitfield SamplersUsed = vs->SamplersUsed | fs->SamplersUsed;
 
-   brw->sampler.count = _mesa_fls(SamplersUsed);
+   brw->wm.sampler_count = _mesa_fls(SamplersUsed);
+   /* Currently we only use one sampler state table.  Mirror the count. */
+   brw->vs.sampler_count = brw->wm.sampler_count;
 
-   if (brw->sampler.count == 0)
+   if (brw->wm.sampler_count == 0)
       return;
 
    samplers = brw_state_batch(brw, AUB_TRACE_SAMPLER_STATE,
-			      brw->sampler.count * sizeof(*samplers),
+			      brw->wm.sampler_count * sizeof(*samplers),
 			      32, &brw->sampler.offset);
-   memset(samplers, 0, brw->sampler.count * sizeof(*samplers));
+   memset(samplers, 0, brw->wm.sampler_count * sizeof(*samplers));
 
-   for (unsigned s = 0; s < brw->sampler.count; s++) {
+   for (unsigned s = 0; s < brw->wm.sampler_count; s++) {
       if (SamplersUsed & (1 << s)) {
          const unsigned unit = (fs->SamplersUsed & (1 << s)) ?
             fs->SamplerUnits[s] : vs->SamplerUnits[s];
diff --git a/src/mesa/drivers/dri/i965/gen7_vs_state.c b/src/mesa/drivers/dri/i965/gen7_vs_state.c
index 0340da4..634bd95 100644
--- a/src/mesa/drivers/dri/i965/gen7_vs_state.c
+++ b/src/mesa/drivers/dri/i965/gen7_vs_state.c
@@ -89,7 +89,7 @@ upload_vs_state(struct brw_context *brw)
    OUT_BATCH(_3DSTATE_VS << 16 | (6 - 2));
    OUT_BATCH(brw->vs.prog_offset);
    OUT_BATCH(floating_point_mode |
-	     ((ALIGN(brw->sampler.count, 4)/4) << GEN6_VS_SAMPLER_COUNT_SHIFT));
+	     ((ALIGN(brw->vs.sampler_count, 4)/4) << GEN6_VS_SAMPLER_COUNT_SHIFT));
 
    if (brw->vs.prog_data->base.total_scratch) {
       OUT_RELOC(brw->vs.scratch_bo,
diff --git a/src/mesa/drivers/dri/i965/gen7_wm_state.c b/src/mesa/drivers/dri/i965/gen7_wm_state.c
index 1bc6e2e..d079a52 100644
--- a/src/mesa/drivers/dri/i965/gen7_wm_state.c
+++ b/src/mesa/drivers/dri/i965/gen7_wm_state.c
@@ -162,7 +162,7 @@ upload_ps_state(struct brw_context *brw)
    dw2 = dw4 = dw5 = 0;
 
    /* CACHE_NEW_SAMPLER */
-   dw2 |= (ALIGN(brw->sampler.count, 4) / 4) << GEN7_PS_SAMPLER_COUNT_SHIFT;
+   dw2 |= (ALIGN(brw->wm.sampler_count, 4) / 4) << GEN7_PS_SAMPLER_COUNT_SHIFT;
 
    /* Use ALT floating point mode for ARB fragment programs, because they
     * require 0^0 == 1.  Even though _CurrentFragmentProgram is used for