#include "brw_defines.h"
#include "brw_state.h"
-#include "gen6_hiz.h"
+#include "brw_blorp.h"
+#include "gen6_blorp.h"
/**
- * \name Constants for HiZ VBO
+ * \name Constants for BLORP VBO
* \{
- *
- * \see brw_context::hiz::vertex_bo
*/
-#define GEN6_HIZ_NUM_VERTICES 3
-#define GEN6_HIZ_NUM_VUE_ELEMS 8
-#define GEN6_HIZ_VBO_SIZE (GEN6_HIZ_NUM_VERTICES \
- * GEN6_HIZ_NUM_VUE_ELEMS \
- * sizeof(float))
+#define GEN6_BLORP_NUM_VERTICES 3
+#define GEN6_BLORP_NUM_VUE_ELEMS 8
+#define GEN6_BLORP_VBO_SIZE (GEN6_BLORP_NUM_VERTICES \
+ * GEN6_BLORP_NUM_VUE_ELEMS \
+ * sizeof(float))
/** \} */
void
-gen6_hiz_emit_batch_head(struct brw_context *brw)
+gen6_blorp_emit_batch_head(struct brw_context *brw,
+ const brw_blorp_params *params)
{
- struct gl_context *ctx = &brw->intel.ctx;
- struct intel_context *intel = &brw->intel;
-
/* To ensure that the batch contains only the resolve, flush the batch
* before beginning and after finishing emitting the resolve packets.
- *
- * Ideally, we would not need to flush for the resolve op. But, I suspect
- * that it's unsafe for CMD_PIPELINE_SELECT to occur multiple times in
- * a single batch, and there is no safe way to ensure that other than by
- * fencing the resolve with flushes. Ideally, we would just detect if
- * a batch is in progress and do the right thing, but that would require
- * the ability to *safely* access brw_context::state::dirty::brw
- * outside of the brw_upload_state() codepath.
- */
- intel_flush(ctx);
-
- /* CMD_PIPELINE_SELECT
- *
- * Select the 3D pipeline, as opposed to the media pipeline.
*/
- {
- BEGIN_BATCH(1);
- OUT_BATCH(brw->CMD_PIPELINE_SELECT << 16);
- ADVANCE_BATCH();
- }
-
- /* 3DSTATE_MULTISAMPLE */
- {
- int length = intel->gen == 7 ? 4 : 3;
+ intel_batchbuffer_flush(brw);
+}
- BEGIN_BATCH(length);
- OUT_BATCH(_3DSTATE_MULTISAMPLE << 16 | (length - 2));
- OUT_BATCH(MS_PIXEL_LOCATION_CENTER |
- MS_NUMSAMPLES_1);
- OUT_BATCH(0);
- if (length >= 4)
- OUT_BATCH(0);
- ADVANCE_BATCH();
- }
+/**
+ * CMD_STATE_BASE_ADDRESS
+ *
+ * From the Sandy Bridge PRM, Volume 1, Part 1, Table STATE_BASE_ADDRESS:
+ * The following commands must be reissued following any change to the
+ * base addresses:
+ * 3DSTATE_CC_POINTERS
+ * 3DSTATE_BINDING_TABLE_POINTERS
+ * 3DSTATE_SAMPLER_STATE_POINTERS
+ * 3DSTATE_VIEWPORT_STATE_POINTERS
+ * MEDIA_STATE_POINTERS
+ */
+void
+gen6_blorp_emit_state_base_address(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ uint8_t mocs = brw->gen == 7 ? GEN7_MOCS_L3 : 0;
- /* 3DSTATE_SAMPLE_MASK */
- {
- BEGIN_BATCH(2);
- OUT_BATCH(_3DSTATE_SAMPLE_MASK << 16 | (2 - 2));
- OUT_BATCH(1);
- ADVANCE_BATCH();
- }
+ BEGIN_BATCH(10);
+ OUT_BATCH(CMD_STATE_BASE_ADDRESS << 16 | (10 - 2));
+ OUT_BATCH(mocs << 8 | /* GeneralStateMemoryObjectControlState */
+ mocs << 4 | /* StatelessDataPortAccessMemoryObjectControlState */
+ 1); /* GeneralStateBaseAddressModifyEnable */
- /* CMD_STATE_BASE_ADDRESS
- *
- * From the Sandy Bridge PRM, Volume 1, Part 1, Table STATE_BASE_ADDRESS:
- * The following commands must be reissued following any change to the
- * base addresses:
- * 3DSTATE_CC_POINTERS
- * 3DSTATE_BINDING_TABLE_POINTERS
- * 3DSTATE_SAMPLER_STATE_POINTERS
- * 3DSTATE_VIEWPORT_STATE_POINTERS
- * MEDIA_STATE_POINTERS
- */
- {
- BEGIN_BATCH(10);
- OUT_BATCH(CMD_STATE_BASE_ADDRESS << 16 | (10 - 2));
- OUT_BATCH(1); /* GeneralStateBaseAddressModifyEnable */
- /* SurfaceStateBaseAddress */
- OUT_RELOC(intel->batch.bo, I915_GEM_DOMAIN_SAMPLER, 0, 1);
- /* DynamicStateBaseAddress */
- OUT_RELOC(intel->batch.bo, (I915_GEM_DOMAIN_RENDER |
- I915_GEM_DOMAIN_INSTRUCTION), 0, 1);
- OUT_BATCH(1); /* IndirectObjectBaseAddress */
+ /* SurfaceStateBaseAddress */
+ OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_SAMPLER, 0, 1);
+ /* DynamicStateBaseAddress */
+ OUT_RELOC(brw->batch.bo, (I915_GEM_DOMAIN_RENDER |
+ I915_GEM_DOMAIN_INSTRUCTION), 0, 1);
+ OUT_BATCH(1); /* IndirectObjectBaseAddress */
+ if (params->use_wm_prog) {
+ OUT_RELOC(brw->cache.bo, I915_GEM_DOMAIN_INSTRUCTION, 0,
+ 1); /* Instruction base address: shader kernels */
+ } else {
OUT_BATCH(1); /* InstructionBaseAddress */
- OUT_BATCH(1); /* GeneralStateUpperBound */
- OUT_BATCH(1); /* DynamicStateUpperBound */
- OUT_BATCH(1); /* IndirectObjectUpperBound*/
- OUT_BATCH(1); /* InstructionAccessUpperBound */
- ADVANCE_BATCH();
}
+ OUT_BATCH(1); /* GeneralStateUpperBound */
+ /* Dynamic state upper bound. Although the documentation says that
+ * programming it to zero will cause it to be ignored, that is a lie.
+ * If this isn't programmed to a real bound, the sampler border color
+ * pointer is rejected, causing border color to mysteriously fail.
+ */
+ OUT_BATCH(0xfffff001);
+ OUT_BATCH(1); /* IndirectObjectUpperBound*/
+ OUT_BATCH(1); /* InstructionAccessUpperBound */
+ ADVANCE_BATCH();
}
+
void
-gen6_hiz_emit_vertices(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- unsigned int level,
- unsigned int layer)
+gen6_blorp_emit_vertices(struct brw_context *brw,
+ const brw_blorp_params *params)
{
- struct intel_context *intel = &brw->intel;
uint32_t vertex_offset;
/* Setup VBO for the rectangle primitive..
* "Vertex URB Entry (VUE) Formats".
*/
{
- const int width = mt->level[level].width;
- const int height = mt->level[level].height;
float *vertex_data;
- const float vertices[GEN6_HIZ_VBO_SIZE] = {
- /* v0 */ 0, 0, 0, 0, 0, height, 0, 1,
- /* v1 */ 0, 0, 0, 0, width, height, 0, 1,
- /* v2 */ 0, 0, 0, 0, 0, 0, 0, 1,
+ const float vertices[GEN6_BLORP_VBO_SIZE] = {
+ /* v0 */ 0, 0, 0, 0, (float) params->x0, (float) params->y1, 0, 1,
+ /* v1 */ 0, 0, 0, 0, (float) params->x1, (float) params->y1, 0, 1,
+ /* v2 */ 0, 0, 0, 0, (float) params->x0, (float) params->y0, 0, 1,
};
- vertex_data = (float *) brw_state_batch(brw, AUB_TRACE_NO_TYPE,
- GEN6_HIZ_VBO_SIZE, 32,
+ vertex_data = (float *) brw_state_batch(brw, AUB_TRACE_VERTEX_BUFFER,
+ GEN6_BLORP_VBO_SIZE, 32,
&vertex_offset);
- memcpy(vertex_data, vertices, GEN6_HIZ_VBO_SIZE);
+ memcpy(vertex_data, vertices, GEN6_BLORP_VBO_SIZE);
}
/* 3DSTATE_VERTEX_BUFFERS */
const int batch_length = 1 + 4 * num_buffers;
uint32_t dw0 = GEN6_VB0_ACCESS_VERTEXDATA |
- (GEN6_HIZ_NUM_VUE_ELEMS * sizeof(float)) << BRW_VB0_PITCH_SHIFT;
+ (GEN6_BLORP_NUM_VUE_ELEMS * sizeof(float)) << BRW_VB0_PITCH_SHIFT;
- if (intel->gen >= 7)
+ 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(intel->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
+ OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
vertex_offset);
/* end address */
- OUT_RELOC(intel->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
- vertex_offset + GEN6_HIZ_VBO_SIZE - 1);
+ OUT_RELOC(brw->batch.bo, I915_GEM_DOMAIN_VERTEX, 0,
+ vertex_offset + GEN6_BLORP_VBO_SIZE - 1);
OUT_BATCH(0);
ADVANCE_BATCH();
}
/* 3DSTATE_VERTEX_ELEMENTS
*
* Fetch dwords 0 - 7 from each VUE. See the comments above where
- * hiz->vertex_bo is filled with data.
+ * the vertex_bo is filled with data.
*/
{
const int num_elements = 2;
}
}
-/**
- * \brief Execute a HiZ op on a miptree slice.
+
+/* 3DSTATE_URB
*
- * To execute the HiZ op, this function manually constructs and emits a batch
- * to "draw" the HiZ op's rectangle primitive. The batchbuffer is flushed
- * before constructing and after emitting the batch.
+ * Assign the entire URB to the VS. Even though the VS disabled, URB space
+ * is still needed because the clipper loads the VUE's from the URB. From
+ * the Sandybridge PRM, Volume 2, Part 1, Section 3DSTATE,
+ * Dword 1.15:0 "VS Number of URB Entries":
+ * This field is always used (even if VS Function Enable is DISABLED).
*
- * This function alters no GL state.
- *
- * For an overview of HiZ ops, see the following sections of the Sandy Bridge
- * PRM, Volume 1, Part 2:
- * - 7.5.3.1 Depth Buffer Clear
- * - 7.5.3.2 Depth Buffer Resolve
- * - 7.5.3.3 Hierarchical Depth Buffer Resolve
+ * The warning below appears in the PRM (Section 3DSTATE_URB), but we can
+ * safely ignore it because this batch contains only one draw call.
+ * Because of URB corruption caused by allocating a previous GS unit
+ * URB entry to the VS unit, software is required to send a “GS NULL
+ * Fence” (Send URB fence with VS URB size == 1 and GS URB size == 0)
+ * plus a dummy DRAW call before any case where VS will be taking over
+ * GS URB space.
*/
static void
-gen6_hiz_exec(struct intel_context *intel,
- struct intel_mipmap_tree *mt,
- unsigned int level,
- unsigned int layer,
- enum gen6_hiz_op op)
+gen6_blorp_emit_urb_config(struct brw_context *brw,
+ const brw_blorp_params *params)
{
- struct gl_context *ctx = &intel->ctx;
- struct brw_context *brw = brw_context(ctx);
- uint32_t draw_x, draw_y;
- uint32_t tile_mask_x, tile_mask_y;
+ BEGIN_BATCH(3);
+ OUT_BATCH(_3DSTATE_URB << 16 | (3 - 2));
+ OUT_BATCH(brw->urb.max_vs_entries << GEN6_URB_VS_ENTRIES_SHIFT);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
- assert(op != GEN6_HIZ_OP_DEPTH_CLEAR); /* Not implemented yet. */
- assert(mt->hiz_mt != NULL);
- intel_miptree_check_level_layer(mt, level, layer);
- {
- /* Construct a dummy renderbuffer just to extract tile offsets. */
- struct intel_renderbuffer rb;
- rb.mt = mt;
- rb.mt_level = level;
- rb.mt_layer = layer;
- intel_renderbuffer_set_draw_offset(&rb);
- draw_x = rb.draw_x;
- draw_y = rb.draw_y;
- }
+/* BLEND_STATE */
+uint32_t
+gen6_blorp_emit_blend_state(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ uint32_t cc_blend_state_offset;
+
+ struct gen6_blend_state *blend = (struct gen6_blend_state *)
+ brw_state_batch(brw, AUB_TRACE_BLEND_STATE,
+ sizeof(struct gen6_blend_state), 64,
+ &cc_blend_state_offset);
- /* Compute masks to determine how much of draw_x and draw_y should be
- * performed using the fine adjustment of "depth coordinate offset X/Y"
- * (dw5 of 3DSTATE_DEPTH_BUFFER). See the emit_depthbuffer() function for
- * details.
+ 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;
+
+ 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];
+
+ /* When blitting from an XRGB source to a ARGB destination, we need to
+ * interpret the missing channel as 1.0. Blending can do that for us:
+ * we simply use the RGB values from the fragment shader ("source RGB"),
+ * but smash the alpha channel to 1.
*/
- {
- uint32_t depth_mask_x, depth_mask_y, hiz_mask_x, hiz_mask_y;
- intel_region_get_tile_masks(mt->region, &depth_mask_x, &depth_mask_y);
- intel_region_get_tile_masks(mt->hiz_mt->region,
- &hiz_mask_x, &hiz_mask_y);
+ if (params->src.mt &&
+ _mesa_get_format_bits(params->dst.mt->format, GL_ALPHA_BITS) > 0 &&
+ _mesa_get_format_bits(params->src.mt->format, GL_ALPHA_BITS) == 0) {
+ blend->blend0.blend_enable = 1;
+ blend->blend0.ia_blend_enable = 1;
- /* Each HiZ row represents 2 rows of pixels */
- hiz_mask_y = hiz_mask_y << 1 | 1;
+ blend->blend0.blend_func = BRW_BLENDFUNCTION_ADD;
+ blend->blend0.ia_blend_func = BRW_BLENDFUNCTION_ADD;
- tile_mask_x = depth_mask_x | hiz_mask_x;
- tile_mask_y = depth_mask_y | hiz_mask_y;
+ blend->blend0.source_blend_factor = BRW_BLENDFACTOR_SRC_COLOR;
+ blend->blend0.dest_blend_factor = BRW_BLENDFACTOR_ZERO;
+ blend->blend0.ia_source_blend_factor = BRW_BLENDFACTOR_ONE;
+ blend->blend0.ia_dest_blend_factor = BRW_BLENDFACTOR_ZERO;
}
- gen6_hiz_emit_batch_head(brw);
- gen6_hiz_emit_vertices(brw, mt, level, layer);
+ return cc_blend_state_offset;
+}
+
- /* 3DSTATE_URB
- *
- * Assign the entire URB to the VS. Even though the VS disabled, URB space
- * is still needed because the clipper loads the VUE's from the URB. From
- * the Sandybridge PRM, Volume 2, Part 1, Section 3DSTATE,
- * Dword 1.15:0 "VS Number of URB Entries":
- * This field is always used (even if VS Function Enable is DISABLED).
- *
- * The warning below appears in the PRM (Section 3DSTATE_URB), but we can
- * safely ignore it because this batch contains only one draw call.
- * Because of URB corruption caused by allocating a previous GS unit
- * URB entry to the VS unit, software is required to send a “GS NULL
- * Fence” (Send URB fence with VS URB size == 1 and GS URB size == 0)
- * plus a dummy DRAW call before any case where VS will be taking over
- * GS URB space.
+/* CC_STATE */
+uint32_t
+gen6_blorp_emit_cc_state(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ uint32_t cc_state_offset;
+
+ struct gen6_color_calc_state *cc = (struct gen6_color_calc_state *)
+ brw_state_batch(brw, AUB_TRACE_CC_STATE,
+ sizeof(gen6_color_calc_state), 64,
+ &cc_state_offset);
+ memset(cc, 0, sizeof(*cc));
+
+ return cc_state_offset;
+}
+
+
+/**
+ * \param out_offset is relative to
+ * CMD_STATE_BASE_ADDRESS.DynamicStateBaseAddress.
+ */
+uint32_t
+gen6_blorp_emit_depth_stencil_state(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ uint32_t depthstencil_offset;
+
+ struct gen6_depth_stencil_state *state;
+ state = (struct gen6_depth_stencil_state *)
+ brw_state_batch(brw, AUB_TRACE_DEPTH_STENCIL_STATE,
+ sizeof(*state), 64,
+ &depthstencil_offset);
+ memset(state, 0, sizeof(*state));
+
+ /* See the following sections of the Sandy Bridge PRM, Volume 1, Part2:
+ * - 7.5.3.1 Depth Buffer Clear
+ * - 7.5.3.2 Depth Buffer Resolve
+ * - 7.5.3.3 Hierarchical Depth Buffer Resolve
*/
- {
- BEGIN_BATCH(3);
- OUT_BATCH(_3DSTATE_URB << 16 | (3 - 2));
- OUT_BATCH(brw->urb.max_vs_entries << GEN6_URB_VS_ENTRIES_SHIFT);
- OUT_BATCH(0);
- ADVANCE_BATCH();
+ state->ds2.depth_write_enable = 1;
+ if (params->hiz_op == GEN6_HIZ_OP_DEPTH_RESOLVE) {
+ state->ds2.depth_test_enable = 1;
+ state->ds2.depth_test_func = BRW_COMPAREFUNCTION_NEVER;
}
- /* 3DSTATE_CC_STATE_POINTERS
- *
- * The pointer offsets are relative to
- * CMD_STATE_BASE_ADDRESS.DynamicStateBaseAddress.
- *
- * The HiZ op doesn't use BLEND_STATE or COLOR_CALC_STATE.
- */
- {
- uint32_t depthstencil_offset;
- gen6_hiz_emit_depth_stencil_state(brw, op, &depthstencil_offset);
-
- BEGIN_BATCH(4);
- OUT_BATCH(_3DSTATE_CC_STATE_POINTERS << 16 | (4 - 2));
- OUT_BATCH(1); /* BLEND_STATE offset */
- OUT_BATCH(depthstencil_offset | 1); /* DEPTH_STENCIL_STATE offset */
- OUT_BATCH(1); /* COLOR_CALC_STATE offset */
- ADVANCE_BATCH();
+ return depthstencil_offset;
+}
+
+
+/* 3DSTATE_CC_STATE_POINTERS
+ *
+ * The pointer offsets are relative to
+ * CMD_STATE_BASE_ADDRESS.DynamicStateBaseAddress.
+ *
+ * The HiZ op doesn't use BLEND_STATE or COLOR_CALC_STATE.
+ */
+static void
+gen6_blorp_emit_cc_state_pointers(struct brw_context *brw,
+ const brw_blorp_params *params,
+ uint32_t cc_blend_state_offset,
+ uint32_t depthstencil_offset,
+ uint32_t cc_state_offset)
+{
+ BEGIN_BATCH(4);
+ OUT_BATCH(_3DSTATE_CC_STATE_POINTERS << 16 | (4 - 2));
+ OUT_BATCH(cc_blend_state_offset | 1); /* BLEND_STATE offset */
+ OUT_BATCH(depthstencil_offset | 1); /* DEPTH_STENCIL_STATE offset */
+ OUT_BATCH(cc_state_offset | 1); /* COLOR_CALC_STATE offset */
+ ADVANCE_BATCH();
+}
+
+
+/* WM push constants */
+uint32_t
+gen6_blorp_emit_wm_constants(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ uint32_t wm_push_const_offset;
+
+ void *constants = brw_state_batch(brw, AUB_TRACE_WM_CONSTANTS,
+ sizeof(params->wm_push_consts),
+ 32, &wm_push_const_offset);
+ memcpy(constants, ¶ms->wm_push_consts,
+ sizeof(params->wm_push_consts));
+
+ return wm_push_const_offset;
+}
+
+
+/* SURFACE_STATE for renderbuffer or texture surface (see
+ * brw_update_renderbuffer_surface and brw_update_texture_surface)
+ */
+static uint32_t
+gen6_blorp_emit_surface_state(struct brw_context *brw,
+ const brw_blorp_params *params,
+ const brw_blorp_surface_info *surface,
+ uint32_t read_domains, uint32_t write_domain)
+{
+ uint32_t wm_surf_offset;
+ uint32_t width = surface->width;
+ uint32_t height = surface->height;
+ if (surface->num_samples > 1) {
+ /* Since gen6 uses INTEL_MSAA_LAYOUT_IMS, width and height are measured
+ * in samples. But SURFACE_STATE wants them in pixels, so we need to
+ * divide them each by 2.
+ */
+ width /= 2;
+ height /= 2;
}
+ struct intel_region *region = surface->mt->region;
+ uint32_t tile_x, tile_y;
+
+ uint32_t *surf = (uint32_t *)
+ brw_state_batch(brw, AUB_TRACE_SURFACE_STATE, 6 * 4, 32,
+ &wm_surf_offset);
+
+ surf[0] = (BRW_SURFACE_2D << BRW_SURFACE_TYPE_SHIFT |
+ BRW_SURFACE_MIPMAPLAYOUT_BELOW << BRW_SURFACE_MIPLAYOUT_SHIFT |
+ BRW_SURFACE_CUBEFACE_ENABLES |
+ surface->brw_surfaceformat << BRW_SURFACE_FORMAT_SHIFT);
+
+ /* reloc */
+ surf[1] = (surface->compute_tile_offsets(&tile_x, &tile_y) +
+ region->bo->offset);
+
+ surf[2] = (0 << BRW_SURFACE_LOD_SHIFT |
+ (width - 1) << BRW_SURFACE_WIDTH_SHIFT |
+ (height - 1) << BRW_SURFACE_HEIGHT_SHIFT);
+
+ uint32_t tiling = surface->map_stencil_as_y_tiled
+ ? BRW_SURFACE_TILED | BRW_SURFACE_TILED_Y
+ : brw_get_surface_tiling_bits(region->tiling);
+ uint32_t pitch_bytes = region->pitch;
+ if (surface->map_stencil_as_y_tiled)
+ pitch_bytes *= 2;
+ surf[3] = (tiling |
+ 0 << BRW_SURFACE_DEPTH_SHIFT |
+ (pitch_bytes - 1) << BRW_SURFACE_PITCH_SHIFT);
+
+ surf[4] = brw_get_surface_num_multisamples(surface->num_samples);
+
+ /* Note that the low bits of these fields are missing, so
+ * there's the possibility of getting in trouble.
+ */
+ assert(tile_x % 4 == 0);
+ assert(tile_y % 2 == 0);
+ surf[5] = ((tile_x / 4) << BRW_SURFACE_X_OFFSET_SHIFT |
+ (tile_y / 2) << BRW_SURFACE_Y_OFFSET_SHIFT |
+ (surface->mt->align_h == 4 ?
+ BRW_SURFACE_VERTICAL_ALIGN_ENABLE : 0));
- /* 3DSTATE_VS
+ /* Emit relocation to surface contents */
+ drm_intel_bo_emit_reloc(brw->batch.bo,
+ wm_surf_offset + 4,
+ region->bo,
+ surf[1] - region->bo->offset,
+ read_domains, write_domain);
+
+ return wm_surf_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 wm_bind_bo_offset;
+ uint32_t *bind = (uint32_t *)
+ brw_state_batch(brw, AUB_TRACE_BINDING_TABLE,
+ sizeof(uint32_t) *
+ BRW_BLORP_NUM_BINDING_TABLE_ENTRIES,
+ 32, /* alignment */
+ &wm_bind_bo_offset);
+ bind[BRW_BLORP_RENDERBUFFER_BINDING_TABLE_INDEX] =
+ wm_surf_offset_renderbuffer;
+ bind[BRW_BLORP_TEXTURE_BINDING_TABLE_INDEX] = wm_surf_offset_texture;
+
+ return wm_bind_bo_offset;
+}
+
+
+/**
+ * SAMPLER_STATE. See brw_update_sampler_state().
+ */
+static uint32_t
+gen6_blorp_emit_sampler_state(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ uint32_t sampler_offset;
+
+ struct brw_sampler_state *sampler = (struct brw_sampler_state *)
+ brw_state_batch(brw, AUB_TRACE_SAMPLER_STATE,
+ sizeof(struct brw_sampler_state),
+ 32, &sampler_offset);
+ memset(sampler, 0, sizeof(*sampler));
+
+ sampler->ss0.min_filter = BRW_MAPFILTER_LINEAR;
+ sampler->ss0.mip_filter = BRW_MIPFILTER_NONE;
+ sampler->ss0.mag_filter = BRW_MAPFILTER_LINEAR;
+
+ sampler->ss1.r_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
+ sampler->ss1.s_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
+ sampler->ss1.t_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
+
+ sampler->ss0.min_mag_neq = 1;
+
+ /* Set LOD bias:
+ */
+ sampler->ss0.lod_bias = 0;
+
+ sampler->ss0.lod_preclamp = 1; /* OpenGL mode */
+ sampler->ss0.default_color_mode = 0; /* OpenGL/DX10 mode */
+
+ /* Set BaseMipLevel, MaxLOD, MinLOD:
*
- * Disable vertex shader.
+ * XXX: I don't think that using firstLevel, lastLevel works,
+ * because we always setup the surface state as if firstLevel ==
+ * level zero. Probably have to subtract firstLevel from each of
+ * these:
*/
- {
- /* From the BSpec, Volume 2a, Part 3 "Vertex Shader", Section
+ sampler->ss0.base_level = U_FIXED(0, 1);
+
+ sampler->ss1.max_lod = U_FIXED(0, 6);
+ sampler->ss1.min_lod = U_FIXED(0, 6);
+
+ sampler->ss3.non_normalized_coord = 1;
+
+ sampler->ss3.address_round |= BRW_ADDRESS_ROUNDING_ENABLE_U_MIN |
+ BRW_ADDRESS_ROUNDING_ENABLE_V_MIN |
+ BRW_ADDRESS_ROUNDING_ENABLE_R_MIN;
+ sampler->ss3.address_round |= BRW_ADDRESS_ROUNDING_ENABLE_U_MAG |
+ BRW_ADDRESS_ROUNDING_ENABLE_V_MAG |
+ BRW_ADDRESS_ROUNDING_ENABLE_R_MAG;
+
+ return sampler_offset;
+}
+
+
+/**
+ * 3DSTATE_SAMPLER_STATE_POINTERS. See upload_sampler_state_pointers().
+ */
+static void
+gen6_blorp_emit_sampler_state_pointers(struct brw_context *brw,
+ const brw_blorp_params *params,
+ uint32_t sampler_offset)
+{
+ BEGIN_BATCH(4);
+ OUT_BATCH(_3DSTATE_SAMPLER_STATE_POINTERS << 16 |
+ VS_SAMPLER_STATE_CHANGE |
+ GS_SAMPLER_STATE_CHANGE |
+ PS_SAMPLER_STATE_CHANGE |
+ (4 - 2));
+ OUT_BATCH(0); /* VS */
+ OUT_BATCH(0); /* GS */
+ OUT_BATCH(sampler_offset);
+ ADVANCE_BATCH();
+}
+
+
+/* 3DSTATE_VS
+ *
+ * Disable vertex shader.
+ */
+void
+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.
+ *
+ * [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(intel);
-
- BEGIN_BATCH(6);
- OUT_BATCH(_3DSTATE_VS << 16 | (6 - 2));
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
- ADVANCE_BATCH();
+ intel_emit_post_sync_nonzero_flush(brw);
}
- /* 3DSTATE_GS
- *
- * Disable the geometry shader.
- */
- {
- BEGIN_BATCH(7);
- OUT_BATCH(_3DSTATE_GS << 16 | (7 - 2));
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
+ /* Disable the push constant buffers. */
+ BEGIN_BATCH(5);
+ OUT_BATCH(_3DSTATE_CONSTANT_VS << 16 | (5 - 2));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+
+ BEGIN_BATCH(6);
+ OUT_BATCH(_3DSTATE_VS << 16 | (6 - 2));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
+
+
+/* 3DSTATE_GS
+ *
+ * Disable the geometry shader.
+ */
+void
+gen6_blorp_emit_gs_disable(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ /* Disable all the constant buffers. */
+ BEGIN_BATCH(5);
+ OUT_BATCH(_3DSTATE_CONSTANT_GS << 16 | (5 - 2));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+
+ BEGIN_BATCH(7);
+ OUT_BATCH(_3DSTATE_GS << 16 | (7 - 2));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
+
+
+/* 3DSTATE_CLIP
+ *
+ * Disable the clipper.
+ *
+ * The BLORP op emits a rectangle primitive, which requires clipping to
+ * be disabled. From page 10 of the Sandy Bridge PRM Volume 2 Part 1
+ * Section 1.3 "3D Primitives Overview":
+ * RECTLIST:
+ * Either the CLIP unit should be DISABLED, or the CLIP unit's Clip
+ * Mode should be set to a value other than CLIPMODE_NORMAL.
+ *
+ * Also disable perspective divide. This doesn't change the clipper's
+ * output, but does spare a few electrons.
+ */
+void
+gen6_blorp_emit_clip_disable(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ BEGIN_BATCH(4);
+ OUT_BATCH(_3DSTATE_CLIP << 16 | (4 - 2));
+ OUT_BATCH(0);
+ OUT_BATCH(GEN6_CLIP_PERSPECTIVE_DIVIDE_DISABLE);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
+
+
+/* 3DSTATE_SF
+ *
+ * Disable ViewportTransformEnable (dw2.1)
+ *
+ * From the SandyBridge PRM, Volume 2, Part 1, Section 1.3, "3D
+ * Primitives Overview":
+ * RECTLIST: Viewport Mapping must be DISABLED (as is typical with the
+ * use of screen- space coordinates).
+ *
+ * A solid rectangle must be rendered, so set FrontFaceFillMode (dw2.4:3)
+ * and BackFaceFillMode (dw2.5:6) to SOLID(0).
+ *
+ * From the Sandy Bridge PRM, Volume 2, Part 1, Section
+ * 6.4.1.1 3DSTATE_SF, Field FrontFaceFillMode:
+ * SOLID: Any triangle or rectangle object found to be front-facing
+ * is rendered as a solid object. This setting is required when
+ * (rendering rectangle (RECTLIST) objects.
+ */
+static void
+gen6_blorp_emit_sf_config(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ BEGIN_BATCH(20);
+ OUT_BATCH(_3DSTATE_SF << 16 | (20 - 2));
+ OUT_BATCH((1 - 1) << GEN6_SF_NUM_OUTPUTS_SHIFT | /* only position */
+ 1 << GEN6_SF_URB_ENTRY_READ_LENGTH_SHIFT |
+ 0 << GEN6_SF_URB_ENTRY_READ_OFFSET_SHIFT);
+ OUT_BATCH(0); /* dw2 */
+ OUT_BATCH(params->num_samples > 1 ? GEN6_SF_MSRAST_ON_PATTERN : 0);
+ for (int i = 0; i < 16; ++i)
OUT_BATCH(0);
- ADVANCE_BATCH();
- }
+ ADVANCE_BATCH();
+}
- /* 3DSTATE_CLIP
- *
- * Disable the clipper.
- *
- * The HiZ op emits a rectangle primitive, which requires clipping to
- * be disabled. From page 10 of the Sandy Bridge PRM Volume 2 Part 1
- * Section 1.3 "3D Primitives Overview":
- * RECTLIST:
- * Either the CLIP unit should be DISABLED, or the CLIP unit's Clip
- * Mode should be set to a value other than CLIPMODE_NORMAL.
+
+/**
+ * Enable or disable thread dispatch and set the HiZ op appropriately.
+ */
+static void
+gen6_blorp_emit_wm_config(struct brw_context *brw,
+ const brw_blorp_params *params,
+ uint32_t prog_offset,
+ brw_blorp_prog_data *prog_data)
+{
+ uint32_t dw2, dw4, dw5, dw6;
+
+ /* Even when thread dispatch is disabled, max threads (dw5.25:31) must be
+ * nonzero to prevent the GPU from hanging. While the documentation doesn't
+ * mention this explicitly, it notes that the valid range for the field is
+ * [1,39] = [2,40] threads, which excludes zero.
*
- * Also disable perspective divide. This doesn't change the clipper's
- * output, but does spare a few electrons.
+ * To be safe (and to minimize extraneous code) we go ahead and fully
+ * configure the WM state whether or not there is a WM program.
*/
- {
- BEGIN_BATCH(4);
- OUT_BATCH(_3DSTATE_CLIP << 16 | (4 - 2));
- OUT_BATCH(0);
- OUT_BATCH(GEN6_CLIP_PERSPECTIVE_DIVIDE_DISABLE);
- OUT_BATCH(0);
- ADVANCE_BATCH();
+
+ dw2 = dw4 = dw5 = dw6 = 0;
+ switch (params->hiz_op) {
+ case GEN6_HIZ_OP_DEPTH_CLEAR:
+ dw4 |= GEN6_WM_DEPTH_CLEAR;
+ break;
+ case GEN6_HIZ_OP_DEPTH_RESOLVE:
+ dw4 |= GEN6_WM_DEPTH_RESOLVE;
+ break;
+ case GEN6_HIZ_OP_HIZ_RESOLVE:
+ dw4 |= GEN6_WM_HIERARCHICAL_DEPTH_RESOLVE;
+ break;
+ case GEN6_HIZ_OP_NONE:
+ break;
+ default:
+ assert(0);
+ break;
+ }
+ dw5 |= GEN6_WM_LINE_AA_WIDTH_1_0;
+ dw5 |= GEN6_WM_LINE_END_CAP_AA_WIDTH_0_5;
+ dw5 |= (brw->max_wm_threads - 1) << GEN6_WM_MAX_THREADS_SHIFT;
+ dw6 |= 0 << GEN6_WM_BARYCENTRIC_INTERPOLATION_MODE_SHIFT; /* No interp */
+ dw6 |= 0 << GEN6_WM_NUM_SF_OUTPUTS_SHIFT; /* No inputs from SF */
+ if (params->use_wm_prog) {
+ dw2 |= 1 << GEN6_WM_SAMPLER_COUNT_SHIFT; /* Up to 4 samplers */
+ dw4 |= prog_data->first_curbe_grf << GEN6_WM_DISPATCH_START_GRF_SHIFT_0;
+ dw5 |= GEN6_WM_16_DISPATCH_ENABLE;
+ dw5 |= GEN6_WM_KILL_ENABLE; /* TODO: temporarily smash on */
+ dw5 |= GEN6_WM_DISPATCH_ENABLE; /* We are rendering */
}
- /* 3DSTATE_SF
- *
- * Disable ViewportTransformEnable (dw2.1)
- *
- * From the SandyBridge PRM, Volume 2, Part 1, Section 1.3, "3D
- * Primitives Overview":
- * RECTLIST: Viewport Mapping must be DISABLED (as is typical with the
- * use of screen- space coordinates).
- *
- * A solid rectangle must be rendered, so set FrontFaceFillMode (dw2.4:3)
- * and BackFaceFillMode (dw2.5:6) to SOLID(0).
- *
- * From the Sandy Bridge PRM, Volume 2, Part 1, Section
- * 6.4.1.1 3DSTATE_SF, Field FrontFaceFillMode:
- * SOLID: Any triangle or rectangle object found to be front-facing
- * is rendered as a solid object. This setting is required when
- * (rendering rectangle (RECTLIST) objects.
- */
- {
- BEGIN_BATCH(20);
- OUT_BATCH(_3DSTATE_SF << 16 | (20 - 2));
- OUT_BATCH((1 - 1) << GEN6_SF_NUM_OUTPUTS_SHIFT | /* only position */
- 1 << GEN6_SF_URB_ENTRY_READ_LENGTH_SHIFT |
- 0 << GEN6_SF_URB_ENTRY_READ_OFFSET_SHIFT);
- for (int i = 0; i < 18; ++i)
- OUT_BATCH(0);
- ADVANCE_BATCH();
+ if (params->num_samples > 1) {
+ dw6 |= GEN6_WM_MSRAST_ON_PATTERN;
+ if (prog_data && prog_data->persample_msaa_dispatch)
+ dw6 |= GEN6_WM_MSDISPMODE_PERSAMPLE;
+ else
+ dw6 |= GEN6_WM_MSDISPMODE_PERPIXEL;
+ } else {
+ dw6 |= GEN6_WM_MSRAST_OFF_PIXEL;
+ dw6 |= GEN6_WM_MSDISPMODE_PERSAMPLE;
}
- /* 3DSTATE_WM
- *
- * Disable thread dispatch (dw5.19) and enable the HiZ op.
- *
- * Even though thread dispatch is disabled, max threads (dw5.25:31) must be
- * nonzero to prevent the GPU from hanging. See the valid ranges in the
- * BSpec, Volume 2a.11 Windower, Section 3DSTATE_WM, Dword 5.25:31
- * "Maximum Number Of Threads".
+ BEGIN_BATCH(9);
+ OUT_BATCH(_3DSTATE_WM << 16 | (9 - 2));
+ OUT_BATCH(params->use_wm_prog ? prog_offset : 0);
+ OUT_BATCH(dw2);
+ OUT_BATCH(0); /* No scratch needed */
+ OUT_BATCH(dw4);
+ OUT_BATCH(dw5);
+ OUT_BATCH(dw6);
+ OUT_BATCH(0); /* No other programs */
+ OUT_BATCH(0); /* No other programs */
+ ADVANCE_BATCH();
+}
+
+
+static void
+gen6_blorp_emit_constant_ps(struct brw_context *brw,
+ const brw_blorp_params *params,
+ uint32_t wm_push_const_offset)
+{
+ /* Make sure the push constants fill an exact integer number of
+ * registers.
*/
- {
- uint32_t dw4 = 0;
-
- switch (op) {
- case GEN6_HIZ_OP_DEPTH_CLEAR:
- assert(!"not implemented");
- dw4 |= GEN6_WM_DEPTH_CLEAR;
- break;
- case GEN6_HIZ_OP_DEPTH_RESOLVE:
- dw4 |= GEN6_WM_DEPTH_RESOLVE;
- break;
- case GEN6_HIZ_OP_HIZ_RESOLVE:
- dw4 |= GEN6_WM_HIERARCHICAL_DEPTH_RESOLVE;
- break;
- default:
- assert(0);
- break;
- }
+ assert(sizeof(brw_blorp_wm_push_constants) % 32 == 0);
- BEGIN_BATCH(9);
- OUT_BATCH(_3DSTATE_WM << 16 | (9 - 2));
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(0);
- OUT_BATCH(dw4);
- OUT_BATCH((brw->max_wm_threads - 1) << GEN6_WM_MAX_THREADS_SHIFT);
- OUT_BATCH((1 - 1) << GEN6_WM_NUM_SF_OUTPUTS_SHIFT); /* only position */
- OUT_BATCH(0);
- OUT_BATCH(0);
- ADVANCE_BATCH();
- }
+ /* There must be at least one register worth of push constant data. */
+ assert(BRW_BLORP_NUM_PUSH_CONST_REGS > 0);
+
+ /* Enable push constant buffer 0. */
+ BEGIN_BATCH(5);
+ OUT_BATCH(_3DSTATE_CONSTANT_PS << 16 |
+ GEN6_CONSTANT_BUFFER_0_ENABLE |
+ (5 - 2));
+ OUT_BATCH(wm_push_const_offset + (BRW_BLORP_NUM_PUSH_CONST_REGS - 1));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
+
+static void
+gen6_blorp_emit_constant_ps_disable(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ /* Disable the push constant buffers. */
+ BEGIN_BATCH(5);
+ OUT_BATCH(_3DSTATE_CONSTANT_PS << 16 | (5 - 2));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
+
+/**
+ * 3DSTATE_BINDING_TABLE_POINTERS
+ */
+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);
+ OUT_BATCH(_3DSTATE_BINDING_TABLE_POINTERS << 16 |
+ GEN6_BINDING_TABLE_MODIFY_PS |
+ (4 - 2));
+ OUT_BATCH(0); /* vs -- ignored */
+ OUT_BATCH(0); /* gs -- ignored */
+ OUT_BATCH(wm_bind_bo_offset); /* wm/ps */
+ ADVANCE_BATCH();
+}
+
+
+static void
+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;
+
+ brw_get_depthstencil_tile_masks(params->depth.mt,
+ params->depth.level,
+ params->depth.layer,
+ NULL,
+ &tile_mask_x, &tile_mask_y);
/* 3DSTATE_DEPTH_BUFFER */
{
- uint32_t width = mt->level[level].width;
- uint32_t height = mt->level[level].height;
-
uint32_t tile_x = draw_x & tile_mask_x;
uint32_t tile_y = draw_y & tile_mask_y;
- uint32_t offset = intel_region_get_aligned_offset(mt->region,
- draw_x & ~tile_mask_x,
- draw_y & ~tile_mask_y);
+ uint32_t offset =
+ intel_region_get_aligned_offset(params->depth.mt->region,
+ 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
* tile_x and tile_y to 0. This is a temporary workaround until we come
* up with a better solution.
*/
+ 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;
- uint32_t format;
- switch (mt->format) {
- case MESA_FORMAT_Z16: format = BRW_DEPTHFORMAT_D16_UNORM; break;
- case MESA_FORMAT_Z32_FLOAT: format = BRW_DEPTHFORMAT_D32_FLOAT; break;
- case MESA_FORMAT_X8_Z24: format = BRW_DEPTHFORMAT_D24_UNORM_X8_UINT; break;
- default: assert(0); break;
- }
-
- intel_emit_post_sync_nonzero_flush(intel);
- intel_emit_depth_stall_flushes(intel);
+ intel_emit_post_sync_nonzero_flush(brw);
+ intel_emit_depth_stall_flushes(brw);
BEGIN_BATCH(7);
OUT_BATCH(_3DSTATE_DEPTH_BUFFER << 16 | (7 - 2));
- OUT_BATCH(((mt->region->pitch * mt->region->cpp) - 1) |
- format << 18 |
+ OUT_BATCH((params->depth.mt->region->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);
- OUT_RELOC(mt->region->bo,
+ OUT_RELOC(params->depth.mt->region->bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
offset);
OUT_BATCH(BRW_SURFACE_MIPMAPLAYOUT_BELOW << 1 |
- (width + tile_x - 1) << 6 |
- (height + tile_y - 1) << 19);
+ (params->depth.width + tile_x - 1) << 6 |
+ (params->depth.height + tile_y - 1) << 19);
OUT_BATCH(0);
OUT_BATCH(tile_x |
tile_y << 16);
/* 3DSTATE_HIER_DEPTH_BUFFER */
{
- struct intel_region *hiz_region = mt->hiz_mt->region;
+ struct intel_region *hiz_region = params->depth.mt->hiz_mt->region;
uint32_t hiz_offset =
intel_region_get_aligned_offset(hiz_region,
draw_x & ~tile_mask_x,
- (draw_y & ~tile_mask_y) / 2);
+ (draw_y & ~tile_mask_y) / 2, false);
BEGIN_BATCH(3);
OUT_BATCH((_3DSTATE_HIER_DEPTH_BUFFER << 16) | (3 - 2));
- OUT_BATCH(hiz_region->pitch * hiz_region->cpp - 1);
+ OUT_BATCH(hiz_region->pitch - 1);
OUT_RELOC(hiz_region->bo,
I915_GEM_DOMAIN_RENDER, I915_GEM_DOMAIN_RENDER,
hiz_offset);
OUT_BATCH(0);
ADVANCE_BATCH();
}
+}
- /* 3DSTATE_CLEAR_PARAMS
- *
- * From the Sandybridge PRM, Volume 2, Part 1, Section 3DSTATE_CLEAR_PARAMS:
- * [DevSNB] 3DSTATE_CLEAR_PARAMS packet must follow the DEPTH_BUFFER_STATE
- * packet when HiZ is enabled and the DEPTH_BUFFER_STATE changes.
- */
- {
- BEGIN_BATCH(2);
- OUT_BATCH(_3DSTATE_CLEAR_PARAMS << 16 | (2 - 2));
- OUT_BATCH(0);
- ADVANCE_BATCH();
- }
- /* 3DSTATE_DRAWING_RECTANGLE */
- {
- BEGIN_BATCH(4);
- OUT_BATCH(_3DSTATE_DRAWING_RECTANGLE << 16 | (4 - 2));
- OUT_BATCH(0);
- OUT_BATCH(((mt->level[level].width - 1) & 0xffff) |
- ((mt->level[level].height - 1) << 16));
- OUT_BATCH(0);
- ADVANCE_BATCH();
- }
+static void
+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);
- /* 3DPRIMITIVE */
- {
- BEGIN_BATCH(6);
- OUT_BATCH(CMD_3D_PRIM << 16 | (6 - 2) |
- _3DPRIM_RECTLIST << GEN4_3DPRIM_TOPOLOGY_TYPE_SHIFT |
- GEN4_3DPRIM_VERTEXBUFFER_ACCESS_SEQUENTIAL);
- OUT_BATCH(3); /* vertex count per instance */
- OUT_BATCH(0);
- OUT_BATCH(1); /* instance count */
- OUT_BATCH(0);
- OUT_BATCH(0);
- ADVANCE_BATCH();
- }
+ BEGIN_BATCH(7);
+ OUT_BATCH(_3DSTATE_DEPTH_BUFFER << 16 | (7 - 2));
+ OUT_BATCH((BRW_DEPTHFORMAT_D32_FLOAT << 18) |
+ (BRW_SURFACE_NULL << 29));
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
+}
- /* See comments above at first invocation of intel_flush() in
- * gen6_hiz_emit_batch_head().
- */
- intel_flush(ctx);
- /* Be safe. */
- brw->state.dirty.brw = ~0;
- brw->state.dirty.cache = ~0;
+/* 3DSTATE_CLEAR_PARAMS
+ *
+ * From the Sandybridge PRM, Volume 2, Part 1, Section 3DSTATE_CLEAR_PARAMS:
+ * [DevSNB] 3DSTATE_CLEAR_PARAMS packet must follow the DEPTH_BUFFER_STATE
+ * packet when HiZ is enabled and the DEPTH_BUFFER_STATE changes.
+ */
+static void
+gen6_blorp_emit_clear_params(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ BEGIN_BATCH(2);
+ OUT_BATCH(_3DSTATE_CLEAR_PARAMS << 16 |
+ GEN5_DEPTH_CLEAR_VALID |
+ (2 - 2));
+ OUT_BATCH(params->depth.mt ? params->depth.mt->depth_clear_value : 0);
+ ADVANCE_BATCH();
}
-/**
- * \param out_offset is relative to
- * CMD_STATE_BASE_ADDRESS.DynamicStateBaseAddress.
- */
+
+/* 3DSTATE_DRAWING_RECTANGLE */
void
-gen6_hiz_emit_depth_stencil_state(struct brw_context *brw,
- enum gen6_hiz_op op,
- uint32_t *out_offset)
+gen6_blorp_emit_drawing_rectangle(struct brw_context *brw,
+ const brw_blorp_params *params)
{
- struct gen6_depth_stencil_state *state;
- state = (struct gen6_depth_stencil_state *)
- brw_state_batch(brw, AUB_TRACE_DEPTH_STENCIL_STATE,
- sizeof(*state), 64,
- out_offset);
- memset(state, 0, sizeof(*state));
-
- /* See the following sections of the Sandy Bridge PRM, Volume 1, Part2:
- * - 7.5.3.1 Depth Buffer Clear
- * - 7.5.3.2 Depth Buffer Resolve
- * - 7.5.3.3 Hierarchical Depth Buffer Resolve
- */
- state->ds2.depth_write_enable = 1;
- if (op == GEN6_HIZ_OP_DEPTH_RESOLVE) {
- state->ds2.depth_test_enable = 1;
- state->ds2.depth_test_func = COMPAREFUNC_NEVER;
- }
+ 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(0);
+ ADVANCE_BATCH();
}
-/** \see intel_context::vtbl::resolve_hiz_slice */
+/* 3DSTATE_VIEWPORT_STATE_POINTERS */
void
-gen6_resolve_hiz_slice(struct intel_context *intel,
- struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer)
+gen6_blorp_emit_viewport_state(struct brw_context *brw,
+ const brw_blorp_params *params)
{
- gen6_hiz_exec(intel, mt, level, layer, GEN6_HIZ_OP_HIZ_RESOLVE);
+ struct brw_cc_viewport *ccv;
+ uint32_t cc_vp_offset;
+
+ ccv = (struct brw_cc_viewport *)brw_state_batch(brw, AUB_TRACE_CC_VP_STATE,
+ sizeof(*ccv), 32,
+ &cc_vp_offset);
+
+ ccv->min_depth = 0.0;
+ ccv->max_depth = 1.0;
+
+ BEGIN_BATCH(4);
+ OUT_BATCH(_3DSTATE_VIEWPORT_STATE_POINTERS << 16 | (4 - 2) |
+ GEN6_CC_VIEWPORT_MODIFY);
+ OUT_BATCH(0); /* clip VP */
+ OUT_BATCH(0); /* SF VP */
+ OUT_BATCH(cc_vp_offset);
+ ADVANCE_BATCH();
+}
+
+
+/* 3DPRIMITIVE */
+static void
+gen6_blorp_emit_primitive(struct brw_context *brw,
+ const brw_blorp_params *params)
+{
+ BEGIN_BATCH(6);
+ OUT_BATCH(CMD_3D_PRIM << 16 | (6 - 2) |
+ _3DPRIM_RECTLIST << GEN4_3DPRIM_TOPOLOGY_TYPE_SHIFT |
+ GEN4_3DPRIM_VERTEXBUFFER_ACCESS_SEQUENTIAL);
+ OUT_BATCH(3); /* vertex count per instance */
+ OUT_BATCH(0);
+ OUT_BATCH(1); /* instance count */
+ OUT_BATCH(0);
+ OUT_BATCH(0);
+ ADVANCE_BATCH();
}
-/** \see intel_context::vtbl::resolve_depth_slice */
+
+/**
+ * \brief Execute a blit or render pass operation.
+ *
+ * To execute the operation, this function manually constructs and emits a
+ * batch to draw a rectangle primitive. The batchbuffer is flushed before
+ * constructing and after emitting the batch.
+ *
+ * This function alters no GL state.
+ */
void
-gen6_resolve_depth_slice(struct intel_context *intel,
- struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer)
+gen6_blorp_exec(struct brw_context *brw,
+ const brw_blorp_params *params)
{
- gen6_hiz_exec(intel, mt, level, layer, GEN6_HIZ_OP_DEPTH_RESOLVE);
+ brw_blorp_prog_data *prog_data = NULL;
+ uint32_t cc_blend_state_offset = 0;
+ uint32_t cc_state_offset = 0;
+ uint32_t depthstencil_offset;
+ uint32_t wm_push_const_offset = 0;
+ uint32_t wm_bind_bo_offset = 0;
+
+ uint32_t prog_offset = params->get_wm_prog(brw, &prog_data);
+ gen6_blorp_emit_batch_head(brw, params);
+ gen6_emit_3dstate_multisample(brw, params->num_samples);
+ gen6_emit_3dstate_sample_mask(brw, params->num_samples, 1.0, false, ~0u);
+ gen6_blorp_emit_state_base_address(brw, params);
+ gen6_blorp_emit_vertices(brw, params);
+ 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);
+ }
+ depthstencil_offset = gen6_blorp_emit_depth_stencil_state(brw, params);
+ gen6_blorp_emit_cc_state_pointers(brw, params, cc_blend_state_offset,
+ depthstencil_offset, cc_state_offset);
+ if (params->use_wm_prog) {
+ uint32_t wm_surf_offset_renderbuffer;
+ uint32_t wm_surf_offset_texture = 0;
+ uint32_t sampler_offset;
+ wm_push_const_offset = gen6_blorp_emit_wm_constants(brw, params);
+ intel_miptree_used_for_rendering(params->dst.mt);
+ wm_surf_offset_renderbuffer =
+ gen6_blorp_emit_surface_state(brw, params, ¶ms->dst,
+ I915_GEM_DOMAIN_RENDER,
+ I915_GEM_DOMAIN_RENDER);
+ if (params->src.mt) {
+ wm_surf_offset_texture =
+ gen6_blorp_emit_surface_state(brw, params, ¶ms->src,
+ I915_GEM_DOMAIN_SAMPLER, 0);
+ }
+ wm_bind_bo_offset =
+ gen6_blorp_emit_binding_table(brw, params,
+ 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);
+ }
+ gen6_blorp_emit_vs_disable(brw, params);
+ gen6_blorp_emit_gs_disable(brw, params);
+ gen6_blorp_emit_clip_disable(brw, params);
+ gen6_blorp_emit_sf_config(brw, params);
+ if (params->use_wm_prog)
+ gen6_blorp_emit_constant_ps(brw, params, wm_push_const_offset);
+ else
+ 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_viewport_state(brw, params);
+
+ if (params->depth.mt)
+ gen6_blorp_emit_depth_stencil_config(brw, params);
+ else
+ gen6_blorp_emit_depth_disable(brw, params);
+ gen6_blorp_emit_clear_params(brw, params);
+ gen6_blorp_emit_drawing_rectangle(brw, params);
+ gen6_blorp_emit_primitive(brw, params);
}
+