/**
* \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_blorp_exec(struct brw_context *brw,
const struct brw_blorp_params *params)
{
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;
/* Emit workaround flushes when we switch from drawing to blorping. */
brw_emit_post_sync_nonzero_flush(brw);
brw_upload_state_base_address(brw);
gen6_emit_3dstate_multisample(brw, params->dst.num_samples);
gen6_emit_3dstate_sample_mask(brw,
params->dst.num_samples > 1 ?
(1 << params->dst.num_samples) - 1 : 1);
gen6_blorp_emit_vertices(brw, params);
gen6_blorp_emit_urb_config(brw, params);
if (params->wm_prog_data) {
cc_blend_state_offset = gen6_blorp_emit_blend_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,
depthstencil_offset, cc_state_offset);
if (params->wm_prog_data) {
uint32_t wm_surf_offset_renderbuffer;
uint32_t wm_surf_offset_texture = 0;
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,
wm_surf_offset_renderbuffer,
wm_surf_offset_texture);
}
if (params->src.mt) {
const uint32_t 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);
gen6_blorp_emit_sf_config(brw, params);
if (params->wm_prog_data)
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);
if (params->wm_prog_data)
gen6_blorp_emit_binding_table_pointers(brw, 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);
}
/**
* \copydoc gen6_blorp_exec()
*/
void
gen7_blorp_exec(struct brw_context *brw,
const brw_blorp_params *params)
{
if (brw->gen >= 8)
return;
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 sampler_offset = 0;
uint32_t prog_offset = params->get_wm_prog(brw, &prog_data);
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);
gen7_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);
gen7_blorp_emit_blend_state_pointer(brw, params, cc_blend_state_offset);
gen7_blorp_emit_cc_state_pointer(brw, params, cc_state_offset);
}
depthstencil_offset = gen6_blorp_emit_depth_stencil_state(brw, params);
gen7_blorp_emit_depth_stencil_state_pointers(brw, params,
depthstencil_offset);
if (params->use_wm_prog) {
uint32_t wm_surf_offset_renderbuffer;
uint32_t wm_surf_offset_texture = 0;
wm_push_const_offset = gen6_blorp_emit_wm_constants(brw, params);
intel_miptree_used_for_rendering(params->dst.mt);
wm_surf_offset_renderbuffer =
gen7_blorp_emit_surface_state(brw, params, ¶ms->dst,
I915_GEM_DOMAIN_RENDER,
I915_GEM_DOMAIN_RENDER,
true /* is_render_target */);
if (params->src.mt) {
wm_surf_offset_texture =
gen7_blorp_emit_surface_state(brw, params, ¶ms->src,
I915_GEM_DOMAIN_SAMPLER, 0,
false /* is_render_target */);
}
wm_bind_bo_offset =
gen6_blorp_emit_binding_table(brw, params,
wm_surf_offset_renderbuffer,
wm_surf_offset_texture);
sampler_offset = gen7_blorp_emit_sampler_state(brw, params);
}
gen7_blorp_emit_vs_disable(brw, params);
gen7_blorp_emit_hs_disable(brw, params);
gen7_blorp_emit_te_disable(brw, params);
gen7_blorp_emit_ds_disable(brw, params);
gen7_blorp_emit_gs_disable(brw, params);
gen7_blorp_emit_streamout_disable(brw, params);
gen6_blorp_emit_clip_disable(brw, params);
gen7_blorp_emit_sf_config(brw, params);
gen7_blorp_emit_wm_config(brw, params, prog_data);
if (params->use_wm_prog) {
gen7_blorp_emit_binding_table_pointers_ps(brw, params,
wm_bind_bo_offset);
gen7_blorp_emit_sampler_state_pointers_ps(brw, params, sampler_offset);
gen7_blorp_emit_constant_ps(brw, params, wm_push_const_offset);
} else {
gen7_blorp_emit_constant_ps_disable(brw, params);
}
gen7_blorp_emit_ps_config(brw, params, prog_offset, prog_data);
gen7_blorp_emit_cc_viewport(brw, params);
if (params->depth.mt)
gen7_blorp_emit_depth_stencil_config(brw, params);
else
gen7_blorp_emit_depth_disable(brw, params);
gen7_blorp_emit_clear_params(brw, params);
gen6_blorp_emit_drawing_rectangle(brw, params);
gen7_blorp_emit_primitive(brw, params);
}
/**
* Sets up a surface state structure to point at the given region.
* While it is only used for the front/back buffer currently, it should be
* usable for further buffers when doing ARB_draw_buffer support.
*/
static void
gen6_update_renderbuffer_surface(struct brw_context *brw,
struct gl_renderbuffer *rb,
bool layered,
unsigned int unit)
{
struct gl_context *ctx = &brw->ctx;
struct intel_renderbuffer *irb = intel_renderbuffer(rb);
struct intel_mipmap_tree *mt = irb->mt;
uint32_t *surf;
uint32_t format = 0;
/* _NEW_BUFFERS */
mesa_format rb_format = _mesa_get_render_format(ctx, intel_rb_format(irb));
uint32_t surftype;
int depth = MAX2(irb->layer_count, 1);
const GLenum gl_target =
rb->TexImage ? rb->TexImage->TexObject->Target : GL_TEXTURE_2D;
uint32_t surf_index =
brw->wm.prog_data->binding_table.render_target_start + unit;
intel_miptree_used_for_rendering(irb->mt);
surf = brw_state_batch(brw, AUB_TRACE_SURFACE_STATE, 6 * 4, 32,
&brw->wm.base.surf_offset[surf_index]);
format = brw->render_target_format[rb_format];
if (unlikely(!brw->format_supported_as_render_target[rb_format])) {
_mesa_problem(ctx, "%s: renderbuffer format %s unsupported\n",
__func__, _mesa_get_format_name(rb_format));
}
switch (gl_target) {
case GL_TEXTURE_CUBE_MAP_ARRAY:
case GL_TEXTURE_CUBE_MAP:
surftype = BRW_SURFACE_2D;
depth *= 6;
break;
case GL_TEXTURE_3D:
depth = MAX2(irb->mt->logical_depth0, 1);
/* fallthrough */
default:
surftype = translate_tex_target(gl_target);
break;
}
const int min_array_element = layered ? 0 : irb->mt_layer;
surf[0] = SET_FIELD(surftype, BRW_SURFACE_TYPE) |
SET_FIELD(format, BRW_SURFACE_FORMAT);
/* reloc */
assert(mt->offset % mt->cpp == 0);
surf[1] = mt->bo->offset64 + mt->offset;
/* In the gen6 PRM Volume 1 Part 1: Graphics Core, Section 7.18.3.7.1
* (Surface Arrays For all surfaces other than separate stencil buffer):
*
* "[DevSNB] Errata: Sampler MSAA Qpitch will be 4 greater than the value
* calculated in the equation above , for every other odd Surface Height
* starting from 1 i.e. 1,5,9,13"
*
* Since this Qpitch errata only impacts the sampler, we have to adjust the
* input for the rendering surface to achieve the same qpitch. For the
* affected heights, we increment the height by 1 for the rendering
* surface.
*/
int height0 = irb->mt->logical_height0;
if (brw->gen == 6 && irb->mt->num_samples > 1 && (height0 % 4) == 1)
height0++;
surf[2] = SET_FIELD(mt->logical_width0 - 1, BRW_SURFACE_WIDTH) |
SET_FIELD(height0 - 1, BRW_SURFACE_HEIGHT) |
SET_FIELD(irb->mt_level - irb->mt->first_level, BRW_SURFACE_LOD);
surf[3] = brw_get_surface_tiling_bits(mt->tiling) |
SET_FIELD(depth - 1, BRW_SURFACE_DEPTH) |
SET_FIELD(mt->pitch - 1, BRW_SURFACE_PITCH);
surf[4] = brw_get_surface_num_multisamples(mt->num_samples) |
SET_FIELD(min_array_element, BRW_SURFACE_MIN_ARRAY_ELEMENT) |
SET_FIELD(depth - 1, BRW_SURFACE_RENDER_TARGET_VIEW_EXTENT);
surf[5] = (mt->align_h == 4 ? BRW_SURFACE_VERTICAL_ALIGN_ENABLE : 0);
drm_intel_bo_emit_reloc(brw->batch.bo,
brw->wm.base.surf_offset[surf_index] + 4,
mt->bo,
surf[1] - mt->bo->offset64,
I915_GEM_DOMAIN_RENDER,
I915_GEM_DOMAIN_RENDER);
}
static uint32_t
gen8_blorp_emit_surface_states(struct brw_context *brw,
const struct brw_blorp_params *params)
{
uint32_t wm_surf_offset_renderbuffer;
uint32_t wm_surf_offset_texture = 0;
intel_miptree_used_for_rendering(params->dst.mt);
wm_surf_offset_renderbuffer =
brw_blorp_emit_surface_state(brw, ¶ms->dst,
I915_GEM_DOMAIN_RENDER,
I915_GEM_DOMAIN_RENDER,
true /* is_render_target */);
if (params->src.mt) {
const struct brw_blorp_surface_info *surface = ¶ms->src;
struct intel_mipmap_tree *mt = surface->mt;
/* If src is a 2D multisample array texture on Gen7+ using
* INTEL_MSAA_LAYOUT_UMS or INTEL_MSAA_LAYOUT_CMS, src layer is the
* physical layer holding sample 0. So, for example, if mt->num_samples
* == 4, then logical layer n corresponds to layer == 4*n.
*
* Multisampled depth and stencil surfaces have the samples interleaved
* (INTEL_MSAA_LAYOUT_IMS) and therefore the layer doesn't need
* adjustment.
*/
const unsigned layer_divider =
(mt->msaa_layout == INTEL_MSAA_LAYOUT_UMS ||
mt->msaa_layout == INTEL_MSAA_LAYOUT_CMS) ?
MAX2(mt->num_samples, 1) : 1;
const bool is_cube = mt->target == GL_TEXTURE_CUBE_MAP_ARRAY ||
mt->target == GL_TEXTURE_CUBE_MAP;
const unsigned depth = (is_cube ? 6 : 1) * mt->logical_depth0;
const unsigned layer = mt->target != GL_TEXTURE_3D ?
surface->layer / layer_divider : 0;
struct isl_view view = {
.format = surface->brw_surfaceformat,
.base_level = surface->level,
.levels = mt->last_level - surface->level + 1,
.base_array_layer = layer,
.array_len = depth - layer,
.channel_select = {
swizzle_to_scs(GET_SWZ(surface->swizzle, 0)),
swizzle_to_scs(GET_SWZ(surface->swizzle, 1)),
swizzle_to_scs(GET_SWZ(surface->swizzle, 2)),
swizzle_to_scs(GET_SWZ(surface->swizzle, 3)),
},
.usage = ISL_SURF_USAGE_TEXTURE_BIT,
};
brw_emit_surface_state(brw, mt, &view,
brw->gen >= 9 ? SKL_MOCS_WB : BDW_MOCS_WB,
false, &wm_surf_offset_texture, -1,
I915_GEM_DOMAIN_SAMPLER, 0);
}
return gen6_blorp_emit_binding_table(brw,
wm_surf_offset_renderbuffer,
wm_surf_offset_texture);
}
/**
* \copydoc gen6_blorp_exec()
*/
void
gen8_blorp_exec(struct brw_context *brw, const struct brw_blorp_params *params)
{
uint32_t wm_bind_bo_offset = 0;
brw_upload_state_base_address(brw);
gen7_blorp_emit_cc_viewport(brw);
gen7_l3_state.emit(brw);
gen7_blorp_emit_urb_config(brw, params);
const uint32_t cc_blend_state_offset =
gen8_blorp_emit_blend_state(brw, params);
gen7_blorp_emit_blend_state_pointer(brw, cc_blend_state_offset);
const uint32_t cc_state_offset = gen6_blorp_emit_cc_state(brw);
gen7_blorp_emit_cc_state_pointer(brw, cc_state_offset);
gen8_blorp_disable_constant_state(brw, _3DSTATE_CONSTANT_VS);
gen8_blorp_disable_constant_state(brw, _3DSTATE_CONSTANT_HS);
gen8_blorp_disable_constant_state(brw, _3DSTATE_CONSTANT_DS);
gen8_blorp_disable_constant_state(brw, _3DSTATE_CONSTANT_GS);
gen8_blorp_disable_constant_state(brw, _3DSTATE_CONSTANT_PS);
wm_bind_bo_offset = gen8_blorp_emit_surface_states(brw, params);
gen7_blorp_emit_binding_table_pointers_ps(brw, wm_bind_bo_offset);
if (params->src.mt) {
const uint32_t sampler_offset =
gen6_blorp_emit_sampler_state(brw, BRW_MAPFILTER_LINEAR, 0, true);
gen7_blorp_emit_sampler_state_pointers_ps(brw, sampler_offset);
}
gen8_emit_3dstate_multisample(brw, params->dst.num_samples);
gen6_emit_3dstate_sample_mask(brw,
params->dst.num_samples > 1 ?
(1 << params->dst.num_samples) - 1 : 1);
gen8_disable_stages.emit(brw);
gen8_blorp_emit_vs_disable(brw);
gen8_blorp_emit_hs_disable(brw);
gen7_blorp_emit_te_disable(brw);
gen8_blorp_emit_ds_disable(brw);
gen8_blorp_emit_gs_disable(brw);
gen8_blorp_emit_streamout_disable(brw);
gen6_blorp_emit_clip_disable(brw);
gen8_blorp_emit_raster_state(brw);
gen8_blorp_emit_sbe_state(brw, params);
gen8_blorp_emit_sf_config(brw);
gen8_blorp_emit_ps_blend(brw);
gen8_blorp_emit_ps_extra(brw, params);
gen8_blorp_emit_ps_config(brw, params);
gen8_blorp_emit_depth_stencil_state(brw, params);
gen8_blorp_emit_wm_state(brw);
gen8_blorp_emit_depth_disable(brw);
gen7_blorp_emit_clear_params(brw, params);
gen6_blorp_emit_drawing_rectangle(brw, params);
gen8_blorp_emit_vf_topology(brw);
gen8_blorp_emit_vf_sys_gen_vals_state(brw);
gen6_blorp_emit_vertices(brw, params);
gen8_blorp_emit_vf_instancing_state(brw, params);
gen8_blorp_emit_vf_state(brw);
gen7_blorp_emit_primitive(brw, params);
if (brw->gen < 9)
gen8_write_pma_stall_bits(brw, 0);
}
static void
blorp_surf_for_miptree(struct brw_context *brw,
struct blorp_surf *surf,
struct intel_mipmap_tree *mt,
bool is_render_target,
uint32_t safe_aux_usage,
unsigned *level,
unsigned start_layer, unsigned num_layers,
struct isl_surf tmp_surfs[2])
{
if (mt->msaa_layout == INTEL_MSAA_LAYOUT_UMS ||
mt->msaa_layout == INTEL_MSAA_LAYOUT_CMS) {
const unsigned num_samples = MAX2(1, mt->num_samples);
for (unsigned i = 0; i < num_layers; i++) {
for (unsigned s = 0; s < num_samples; s++) {
const unsigned phys_layer = (start_layer + i) * num_samples + s;
intel_miptree_check_level_layer(mt, *level, phys_layer);
}
}
} else {
for (unsigned i = 0; i < num_layers; i++)
intel_miptree_check_level_layer(mt, *level, start_layer + i);
}
intel_miptree_get_isl_surf(brw, mt, &tmp_surfs[0]);
surf->surf = &tmp_surfs[0];
surf->addr = (struct blorp_address) {
.buffer = mt->bo,
.offset = mt->offset,
.read_domains = is_render_target ? I915_GEM_DOMAIN_RENDER :
I915_GEM_DOMAIN_SAMPLER,
.write_domain = is_render_target ? I915_GEM_DOMAIN_RENDER : 0,
};
if (brw->gen == 6 && mt->format == MESA_FORMAT_S_UINT8 &&
mt->array_layout == ALL_SLICES_AT_EACH_LOD) {
/* Sandy bridge stencil and HiZ use this ALL_SLICES_AT_EACH_LOD hack in
* order to allow for layered rendering. The hack makes each LOD of the
* stencil or HiZ buffer a single tightly packed array surface at some
* offset into the surface. Since ISL doesn't know how to deal with the
* crazy ALL_SLICES_AT_EACH_LOD layout and since we have to do a manual
* offset of it anyway, we might as well do the offset here and keep the
* hacks inside the i965 driver.
*
* See also gen6_depth_stencil_state.c
*/
uint32_t offset;
apply_gen6_stencil_hiz_offset(&tmp_surfs[0], mt, *level, &offset);
surf->addr.offset += offset;
*level = 0;
}
struct isl_surf *aux_surf = &tmp_surfs[1];
intel_miptree_get_aux_isl_surf(brw, mt, aux_surf, &surf->aux_usage);
if (surf->aux_usage != ISL_AUX_USAGE_NONE) {
if (surf->aux_usage == ISL_AUX_USAGE_HIZ) {
/* If we're not going to use it as a depth buffer, resolve HiZ */
if (!(safe_aux_usage & (1 << ISL_AUX_USAGE_HIZ))) {
for (unsigned i = 0; i < num_layers; i++) {
intel_miptree_slice_resolve_depth(brw, mt, *level,
start_layer + i);
/* If we're rendering to it then we'll need a HiZ resolve once
* we're done before we can use it with HiZ again.
*/
if (is_render_target)
intel_miptree_slice_set_needs_hiz_resolve(mt, *level,
start_layer + i);
}
surf->aux_usage = ISL_AUX_USAGE_NONE;
}
} else if (!(safe_aux_usage & (1 << surf->aux_usage))) {
uint32_t flags = 0;
if (safe_aux_usage & (1 << ISL_AUX_USAGE_CCS_E))
flags |= INTEL_MIPTREE_IGNORE_CCS_E;
intel_miptree_resolve_color(brw, mt,
*level, start_layer, num_layers, flags);
assert(!intel_miptree_has_color_unresolved(mt, *level, 1,
start_layer, num_layers));
surf->aux_usage = ISL_AUX_USAGE_NONE;
}
}
if (is_render_target)
intel_miptree_used_for_rendering(brw, mt, *level,
start_layer, num_layers);
if (surf->aux_usage != ISL_AUX_USAGE_NONE) {
/* We only really need a clear color if we also have an auxiliary
* surface. Without one, it does nothing.
*/
surf->clear_color = intel_miptree_get_isl_clear_color(brw, mt);
surf->aux_surf = aux_surf;
surf->aux_addr = (struct blorp_address) {
.read_domains = is_render_target ? I915_GEM_DOMAIN_RENDER :
I915_GEM_DOMAIN_SAMPLER,
.write_domain = is_render_target ? I915_GEM_DOMAIN_RENDER : 0,
};
if (mt->mcs_buf) {
surf->aux_addr.buffer = mt->mcs_buf->bo;
surf->aux_addr.offset = mt->mcs_buf->offset;
} else {
assert(surf->aux_usage == ISL_AUX_USAGE_HIZ);
struct intel_mipmap_tree *hiz_mt = mt->hiz_buf->mt;
if (hiz_mt) {
surf->aux_addr.buffer = hiz_mt->bo;
if (brw->gen == 6 &&
hiz_mt->array_layout == ALL_SLICES_AT_EACH_LOD) {
/* gen6 requires the HiZ buffer to be manually offset to the
* right location. We could fixup the surf but it doesn't
* matter since most of those fields don't matter.
*/
apply_gen6_stencil_hiz_offset(aux_surf, hiz_mt, *level,
&surf->aux_addr.offset);
} else {
surf->aux_addr.offset = 0;
}
assert(hiz_mt->pitch == aux_surf->row_pitch);
} else {
surf->aux_addr.buffer = mt->hiz_buf->aux_base.bo;
surf->aux_addr.offset = mt->hiz_buf->aux_base.offset;
}
}
} else {
surf->aux_addr = (struct blorp_address) {
.buffer = NULL,
};
memset(&surf->clear_color, 0, sizeof(surf->clear_color));
}
assert((surf->aux_usage == ISL_AUX_USAGE_NONE) ==
(surf->aux_addr.buffer == NULL));
}
