ADVANCE_BATCH();
}
+static void
+gen6_blorp_emit_vertex_buffer_state(struct brw_context *brw,
+ unsigned num_elems,
+ unsigned vbo_size,
+ uint32_t vertex_offset)
+{
+ /* 3DSTATE_VERTEX_BUFFERS */
+ const int num_buffers = 1;
+ const int batch_length = 1 + 4 * num_buffers;
+
+ uint32_t dw0 = GEN6_VB0_ACCESS_VERTEXDATA |
+ (num_elems * sizeof(float)) << BRW_VB0_PITCH_SHIFT;
+
+ if (brw->gen >= 7)
+ dw0 |= GEN7_VB0_ADDRESS_MODIFYENABLE;
+
+ if (brw->gen == 7)
+ dw0 |= GEN7_MOCS_L3 << 16;
+
+ BEGIN_BATCH(batch_length);
+ OUT_BATCH((_3DSTATE_VERTEX_BUFFERS << 16) | (batch_length - 2));
+ OUT_BATCH(dw0);
+ /* start address */
+ OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
+ vertex_offset);
+ /* end address */
+ OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
+ vertex_offset + vbo_size - 1);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
void
gen6_blorp_emit_vertices(struct brw_context *brw,
memcpy(vertex_data, vertices, GEN6_BLORP_VBO_SIZE);
}
- /* 3DSTATE_VERTEX_BUFFERS */
- {
- const int num_buffers = 1;
- const int batch_length = 1 + 4 * num_buffers;
-
- uint32_t dw0 = GEN6_VB0_ACCESS_VERTEXDATA |
- (GEN6_BLORP_NUM_VUE_ELEMS * sizeof(float)) << BRW_VB0_PITCH_SHIFT;
-
- if (brw->gen >= 7)
- dw0 |= GEN7_VB0_ADDRESS_MODIFYENABLE;
-
- if (brw->gen == 7)
- dw0 |= GEN7_MOCS_L3 << 16;
-
- BEGIN_BATCH(batch_length);
- OUT_BATCH((_3DSTATE_VERTEX_BUFFERS << 16) | (batch_length - 2));
- OUT_BATCH(dw0);
- /* start address */
- OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
- vertex_offset);
- /* end address */
- OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
- vertex_offset + GEN6_BLORP_VBO_SIZE - 1);
- OUT_BATCH(0);
- ADVANCE_BATCH();
- }
+ gen6_blorp_emit_vertex_buffer_state(brw, GEN6_BLORP_NUM_VUE_ELEMS,
+ GEN6_BLORP_VBO_SIZE,
+ vertex_offset);
/* 3DSTATE_VERTEX_ELEMENTS
*
{
uint32_t cc_blend_state_offset;
+ assume(params->num_draw_buffers);
+
+ const unsigned size = params->num_draw_buffers *
+ sizeof(struct gen6_blend_state);
struct gen6_blend_state *blend = (struct gen6_blend_state *)
- brw_state_batch(brw, AUB_TRACE_BLEND_STATE,
- sizeof(struct gen6_blend_state), 64,
+ brw_state_batch(brw, AUB_TRACE_BLEND_STATE, size, 64,
&cc_blend_state_offset);
- memset(blend, 0, sizeof(*blend));
-
- blend->blend1.pre_blend_clamp_enable = 1;
- blend->blend1.post_blend_clamp_enable = 1;
- blend->blend1.clamp_range = BRW_RENDERTARGET_CLAMPRANGE_FORMAT;
+ memset(blend, 0, size);
- blend->blend1.write_disable_r = params->color_write_disable[0];
- blend->blend1.write_disable_g = params->color_write_disable[1];
- blend->blend1.write_disable_b = params->color_write_disable[2];
- blend->blend1.write_disable_a = params->color_write_disable[3];
+ for (unsigned i = 0; i < params->num_draw_buffers; ++i) {
+ blend[i].blend1.pre_blend_clamp_enable = 1;
+ blend[i].blend1.post_blend_clamp_enable = 1;
+ blend[i].blend1.clamp_range = BRW_RENDERTARGET_CLAMPRANGE_FORMAT;
+ }
return cc_blend_state_offset;
}
/* CC_STATE */
uint32_t
-gen6_blorp_emit_cc_state(struct brw_context *brw,
- const brw_blorp_params *params)
+gen6_blorp_emit_cc_state(struct brw_context *brw)
{
uint32_t cc_state_offset;
/* BINDING_TABLE. See brw_wm_binding_table(). */
uint32_t
gen6_blorp_emit_binding_table(struct brw_context *brw,
- const brw_blorp_params *params,
uint32_t wm_surf_offset_renderbuffer,
uint32_t wm_surf_offset_texture)
{
*/
uint32_t
gen6_blorp_emit_sampler_state(struct brw_context *brw,
- const brw_blorp_params *params)
+ unsigned tex_filter, unsigned max_lod,
+ bool non_normalized_coords)
{
uint32_t sampler_offset;
uint32_t *sampler_state = (uint32_t *)
brw_emit_sampler_state(brw,
sampler_state,
sampler_offset,
- BRW_MAPFILTER_LINEAR, /* min filter */
- BRW_MAPFILTER_LINEAR, /* mag filter */
+ tex_filter, /* min filter */
+ tex_filter, /* mag filter */
BRW_MIPFILTER_NONE,
BRW_ANISORATIO_2,
address_rounding,
BRW_TEXCOORDMODE_CLAMP,
BRW_TEXCOORDMODE_CLAMP,
0, /* min LOD */
- 0, /* max LOD */
+ max_lod,
0, /* LOD bias */
0, /* base miplevel */
0, /* shadow function */
- true, /* non-normalized coordinates */
+ non_normalized_coords,
0); /* border color offset - unused */
return sampler_offset;
*/
static void
gen6_blorp_emit_sampler_state_pointers(struct brw_context *brw,
- const brw_blorp_params *params,
uint32_t sampler_offset)
{
BEGIN_BATCH(4);
gen6_blorp_emit_vs_disable(struct brw_context *brw,
const brw_blorp_params *params)
{
- if (brw->gen == 6) {
- /* From the BSpec, 3D Pipeline > Geometry > Vertex Shader > State,
- * 3DSTATE_VS, Dword 5.0 "VS Function Enable":
- *
- * [DevSNB] A pipeline flush must be programmed prior to a
- * 3DSTATE_VS command that causes the VS Function Enable to
- * toggle. Pipeline flush can be executed by sending a PIPE_CONTROL
- * command with CS stall bit set and a post sync operation.
- */
- intel_emit_post_sync_nonzero_flush(brw);
- }
+ /* From the BSpec, 3D Pipeline > Geometry > Vertex Shader > State,
+ * 3DSTATE_VS, Dword 5.0 "VS Function Enable":
+ *
+ * [DevSNB] A pipeline flush must be programmed prior to a
+ * 3DSTATE_VS command that causes the VS Function Enable to
+ * toggle. Pipeline flush can be executed by sending a PIPE_CONTROL
+ * command with CS stall bit set and a post sync operation.
+ *
+ * We've already done one at the start of the BLORP operation.
+ */
/* Disable the push constant buffers. */
BEGIN_BATCH(5);
OUT_BATCH(0);
OUT_BATCH(0);
ADVANCE_BATCH();
+ brw->gs.enabled = false;
}
* output, but does spare a few electrons.
*/
void
-gen6_blorp_emit_clip_disable(struct brw_context *brw,
- const brw_blorp_params *params)
+gen6_blorp_emit_clip_disable(struct brw_context *brw)
{
BEGIN_BATCH(4);
OUT_BATCH(_3DSTATE_CLIP << 16 | (4 - 2));
{
BEGIN_BATCH(20);
OUT_BATCH(_3DSTATE_SF << 16 | (20 - 2));
- OUT_BATCH((1 - 1) << GEN6_SF_NUM_OUTPUTS_SHIFT | /* only position */
+ OUT_BATCH(params->num_varyings << GEN6_SF_NUM_OUTPUTS_SHIFT |
1 << GEN6_SF_URB_ENTRY_READ_LENGTH_SHIFT |
- 0 << GEN6_SF_URB_ENTRY_READ_OFFSET_SHIFT);
+ BRW_SF_URB_ENTRY_READ_OFFSET <<
+ GEN6_SF_URB_ENTRY_READ_OFFSET_SHIFT);
OUT_BATCH(0); /* dw2 */
OUT_BATCH(params->dst.num_samples > 1 ? GEN6_SF_MSRAST_ON_PATTERN : 0);
for (int i = 0; i < 16; ++i)
*/
static void
gen6_blorp_emit_binding_table_pointers(struct brw_context *brw,
- const brw_blorp_params *params,
uint32_t wm_bind_bo_offset)
{
BEGIN_BATCH(4);
gen6_blorp_emit_depth_stencil_config(struct brw_context *brw,
const brw_blorp_params *params)
{
- struct gl_context *ctx = &brw->ctx;
- uint32_t draw_x = params->depth.x_offset;
- uint32_t draw_y = params->depth.y_offset;
- uint32_t tile_mask_x, tile_mask_y;
+ uint32_t surfwidth, surfheight;
uint32_t surftype;
unsigned int depth = MAX2(params->depth.mt->logical_depth0, 1);
GLenum gl_target = params->depth.mt->target;
+ unsigned int lod;
switch (gl_target) {
case GL_TEXTURE_CUBE_MAP_ARRAY:
break;
}
- brw_get_depthstencil_tile_masks(params->depth.mt,
- params->depth.level,
- params->depth.layer,
- NULL,
- &tile_mask_x, &tile_mask_y);
+ const unsigned min_array_element = params->depth.layer;
- /* 3DSTATE_DEPTH_BUFFER */
- {
- uint32_t tile_x = draw_x & tile_mask_x;
- uint32_t tile_y = draw_y & tile_mask_y;
- uint32_t offset =
- intel_miptree_get_aligned_offset(params->depth.mt,
- draw_x & ~tile_mask_x,
- draw_y & ~tile_mask_y, false);
-
- /* According to the Sandy Bridge PRM, volume 2 part 1, pp326-327
- * (3DSTATE_DEPTH_BUFFER dw5), in the documentation for "Depth
- * Coordinate Offset X/Y":
- *
- * "The 3 LSBs of both offsets must be zero to ensure correct
- * alignment"
- *
- * We have no guarantee that tile_x and tile_y are correctly aligned,
- * since they are determined by the mipmap layout, which is only aligned
- * to multiples of 4.
- *
- * So, to avoid hanging the GPU, just smash the low order 3 bits of
- * tile_x and tile_y to 0. This is a temporary workaround until we come
- * up with a better solution.
+ lod = params->depth.level - params->depth.mt->first_level;
+
+ if (params->hiz_op != GEN6_HIZ_OP_NONE && lod == 0) {
+ /* HIZ ops for lod 0 may set the width & height a little
+ * larger to allow the fast depth clear to fit the hardware
+ * alignment requirements. (8x4)
*/
- WARN_ONCE((tile_x & 7) || (tile_y & 7),
- "Depth/stencil buffer needs alignment to 8-pixel boundaries.\n"
- "Truncating offset, bad rendering may occur.\n");
- tile_x &= ~7;
- tile_y &= ~7;
+ surfwidth = params->depth.width;
+ surfheight = params->depth.height;
+ } else {
+ surfwidth = params->depth.mt->logical_width0;
+ surfheight = params->depth.mt->logical_height0;
+ }
- intel_emit_post_sync_nonzero_flush(brw);
- intel_emit_depth_stall_flushes(brw);
+ /* 3DSTATE_DEPTH_BUFFER */
+ {
+ brw_emit_depth_stall_flushes(brw);
BEGIN_BATCH(7);
+ /* 3DSTATE_DEPTH_BUFFER dw0 */
OUT_BATCH(_3DSTATE_DEPTH_BUFFER << 16 | (7 - 2));
+
+ /* 3DSTATE_DEPTH_BUFFER dw1 */
OUT_BATCH((params->depth.mt->pitch - 1) |
params->depth_format << 18 |
1 << 21 | /* separate stencil enable */
1 << 22 | /* hiz enable */
BRW_TILEWALK_YMAJOR << 26 |
1 << 27 | /* y-tiled */
- BRW_SURFACE_2D << 29);
+ surftype << 29);
+
+ /* 3DSTATE_DEPTH_BUFFER dw2 */
OUT_RELOC(params->depth.mt->bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
- offset);
+ 0);
+
+ /* 3DSTATE_DEPTH_BUFFER dw3 */
OUT_BATCH(BRW_SURFACE_MIPMAPLAYOUT_BELOW << 1 |
- (params->depth.width + tile_x - 1) << 6 |
- (params->depth.height + tile_y - 1) << 19);
+ (surfwidth - 1) << 6 |
+ (surfheight - 1) << 19 |
+ lod << 2);
+
+ /* 3DSTATE_DEPTH_BUFFER dw4 */
+ OUT_BATCH((depth - 1) << 21 |
+ min_array_element << 10 |
+ (depth - 1) << 1);
+
+ /* 3DSTATE_DEPTH_BUFFER dw5 */
OUT_BATCH(0);
- OUT_BATCH(tile_x |
- tile_y << 16);
+
+ /* 3DSTATE_DEPTH_BUFFER dw6 */
OUT_BATCH(0);
ADVANCE_BATCH();
}
/* 3DSTATE_HIER_DEPTH_BUFFER */
{
- struct intel_mipmap_tree *hiz_mt = params->depth.mt->hiz_mt;
- uint32_t hiz_offset =
- intel_miptree_get_aligned_offset(hiz_mt,
- draw_x & ~tile_mask_x,
- (draw_y & ~tile_mask_y) / 2, false);
+ struct intel_mipmap_tree *hiz_mt = params->depth.mt->hiz_buf->mt;
+ uint32_t offset = 0;
+
+ if (hiz_mt->array_layout == ALL_SLICES_AT_EACH_LOD) {
+ offset = intel_miptree_get_aligned_offset(hiz_mt,
+ hiz_mt->level[lod].level_x,
+ hiz_mt->level[lod].level_y,
+ false);
+ }
BEGIN_BATCH(3);
OUT_BATCH((_3DSTATE_HIER_DEPTH_BUFFER << 16) | (3 - 2));
OUT_BATCH(hiz_mt->pitch - 1);
OUT_RELOC(hiz_mt->bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
- hiz_offset);
+ offset);
ADVANCE_BATCH();
}
gen6_blorp_emit_depth_disable(struct brw_context *brw,
const brw_blorp_params *params)
{
- intel_emit_post_sync_nonzero_flush(brw);
- intel_emit_depth_stall_flushes(brw);
+ brw_emit_depth_stall_flushes(brw);
BEGIN_BATCH(7);
OUT_BATCH(_3DSTATE_DEPTH_BUFFER << 16 | (7 - 2));
gen6_blorp_emit_drawing_rectangle(struct brw_context *brw,
const brw_blorp_params *params)
{
- if (brw->gen == 6)
- intel_emit_post_sync_nonzero_flush(brw);
-
BEGIN_BATCH(4);
OUT_BATCH(_3DSTATE_DRAWING_RECTANGLE << 16 | (4 - 2));
OUT_BATCH(0);
- OUT_BATCH(((params->x1 - 1) & 0xffff) |
- ((params->y1 - 1) << 16));
+ OUT_BATCH(((MAX2(params->x1, params->x0) - 1) & 0xffff) |
+ ((MAX2(params->y1, params->y0) - 1) << 16));
OUT_BATCH(0);
ADVANCE_BATCH();
}
GEN4_3DPRIM_VERTEXBUFFER_ACCESS_SEQUENTIAL);
OUT_BATCH(3); /* vertex count per instance */
OUT_BATCH(0);
- OUT_BATCH(1); /* instance count */
+ OUT_BATCH(params->num_layers); /* instance count */
OUT_BATCH(0);
OUT_BATCH(0);
ADVANCE_BATCH();
-
- /* Only used on Sandybridge; harmless to set elsewhere. */
- brw->batch.need_workaround_flush = true;
}
/**
uint32_t prog_offset = params->get_wm_prog(brw, &prog_data);
/* Emit workaround flushes when we switch from drawing to blorping. */
- brw->batch.need_workaround_flush = true;
+ brw_emit_post_sync_nonzero_flush(brw);
gen6_emit_3dstate_multisample(brw, params->dst.num_samples);
gen6_emit_3dstate_sample_mask(brw,
gen6_blorp_emit_urb_config(brw, params);
if (params->use_wm_prog) {
cc_blend_state_offset = gen6_blorp_emit_blend_state(brw, params);
- cc_state_offset = gen6_blorp_emit_cc_state(brw, params);
+ cc_state_offset = gen6_blorp_emit_cc_state(brw);
}
depthstencil_offset = gen6_blorp_emit_depth_stencil_state(brw, params);
gen6_blorp_emit_cc_state_pointers(brw, params, cc_blend_state_offset,
I915_GEM_DOMAIN_SAMPLER, 0);
}
wm_bind_bo_offset =
- gen6_blorp_emit_binding_table(brw, params,
+ gen6_blorp_emit_binding_table(brw,
wm_surf_offset_renderbuffer,
wm_surf_offset_texture);
- sampler_offset = gen6_blorp_emit_sampler_state(brw, params);
- gen6_blorp_emit_sampler_state_pointers(brw, params, sampler_offset);
+ sampler_offset =
+ gen6_blorp_emit_sampler_state(brw, BRW_MAPFILTER_LINEAR, 0, true);
+ gen6_blorp_emit_sampler_state_pointers(brw, sampler_offset);
}
gen6_blorp_emit_vs_disable(brw, params);
gen6_blorp_emit_gs_disable(brw, params);
- gen6_blorp_emit_clip_disable(brw, params);
+ gen6_blorp_emit_clip_disable(brw);
gen6_blorp_emit_sf_config(brw, params);
if (params->use_wm_prog)
gen6_blorp_emit_constant_ps(brw, params, wm_push_const_offset);
gen6_blorp_emit_constant_ps_disable(brw, params);
gen6_blorp_emit_wm_config(brw, params, prog_offset, prog_data);
if (params->use_wm_prog)
- gen6_blorp_emit_binding_table_pointers(brw, params, wm_bind_bo_offset);
+ gen6_blorp_emit_binding_table_pointers(brw, wm_bind_bo_offset);
gen6_blorp_emit_viewport_state(brw, params);
if (params->depth.mt)