#include "intel_batchbuffer.h"
#include "intel_mipmap_tree.h"
-#include "intel_resolve_map.h"
#include "intel_tex.h"
#include "intel_blit.h"
#include "intel_fbo.h"
#define FILE_DEBUG_FLAG DEBUG_MIPTREE
static void *intel_miptree_map_raw(struct brw_context *brw,
- struct intel_mipmap_tree *mt);
+ struct intel_mipmap_tree *mt,
+ GLbitfield mode);
static void intel_miptree_unmap_raw(struct intel_mipmap_tree *mt);
if (brw->gen >= 9) {
mesa_format linear_format = _mesa_get_srgb_format_linear(mt->format);
- const uint32_t brw_format = brw_isl_format_for_mesa_format(linear_format);
- return isl_format_supports_ccs_e(&brw->screen->devinfo, brw_format);
+ const enum isl_format isl_format =
+ brw_isl_format_for_mesa_format(linear_format);
+ return isl_format_supports_ccs_e(&brw->screen->devinfo, isl_format);
} else
return true;
}
}
}
+static bool
+create_mapping_table(GLenum target, unsigned first_level, unsigned last_level,
+ unsigned depth0, struct intel_mipmap_level *table)
+{
+ for (unsigned level = first_level; level <= last_level; level++) {
+ const unsigned d =
+ target == GL_TEXTURE_3D ? minify(depth0, level) : depth0;
+
+ table[level].slice = calloc(d, sizeof(*table[0].slice));
+ if (!table[level].slice)
+ goto unwind;
+ }
+
+ return true;
+
+unwind:
+ for (unsigned level = first_level; level <= last_level; level++)
+ free(table[level].slice);
+
+ return false;
+}
/**
* @param for_bo Indicates that the caller is
INTEL_AUX_DISABLE_ALL : INTEL_AUX_DISABLE_NONE;
mt->aux_disable |= INTEL_AUX_DISABLE_CCS;
mt->is_scanout = (layout_flags & MIPTREE_LAYOUT_FOR_SCANOUT) != 0;
- exec_list_make_empty(&mt->hiz_map);
- exec_list_make_empty(&mt->color_resolve_map);
+ mt->aux_state = NULL;
mt->cpp = _mesa_get_format_bytes(format);
mt->num_samples = num_samples;
mt->compressed = _mesa_is_format_compressed(format);
mt->msaa_layout = INTEL_MSAA_LAYOUT_NONE;
mt->refcount = 1;
+ if (brw->gen == 6 && format == MESA_FORMAT_S_UINT8)
+ layout_flags |= MIPTREE_LAYOUT_GEN6_HIZ_STENCIL;
+
int depth_multiply = 1;
if (num_samples > 1) {
/* Adjust width/height/depth for MSAA */
}
}
+ if (!create_mapping_table(target, first_level, last_level, depth0,
+ mt->level)) {
+ free(mt);
+ return NULL;
+ }
+
/* Set array_layout to ALL_SLICES_AT_EACH_LOD when array_spacing_lod0 can
* be used. array_spacing_lod0 is only used for non-IMS MSAA surfaces on
* Gen 7 and 8. On Gen 8 and 9 this layout is not available but it is still
intel_miptree_wants_hiz_buffer(brw, mt)))) {
uint32_t stencil_flags = MIPTREE_LAYOUT_ACCELERATED_UPLOAD;
if (brw->gen == 6) {
- stencil_flags |= MIPTREE_LAYOUT_FORCE_ALL_SLICE_AT_LOD |
- MIPTREE_LAYOUT_TILING_ANY;
+ stencil_flags |= MIPTREE_LAYOUT_TILING_ANY;
}
mt->stencil_mt = intel_miptree_create(brw,
}
}
- if (layout_flags & MIPTREE_LAYOUT_FORCE_ALL_SLICE_AT_LOD)
- mt->array_layout = ALL_SLICES_AT_EACH_LOD;
+ if (layout_flags & MIPTREE_LAYOUT_GEN6_HIZ_STENCIL)
+ mt->array_layout = GEN6_HIZ_STENCIL;
/*
* Obey HALIGN_16 constraints for Gen8 and Gen9 buffers which are
}
}
+/** \brief Assert that the level and layer are valid for the miptree. */
+void
+intel_miptree_check_level_layer(const struct intel_mipmap_tree *mt,
+ uint32_t level,
+ uint32_t layer)
+{
+ (void) mt;
+ (void) level;
+ (void) layer;
+
+ assert(level >= mt->first_level);
+ assert(level <= mt->last_level);
+
+ if (mt->surf.size > 0)
+ assert(layer < (mt->surf.dim == ISL_SURF_DIM_3D ?
+ minify(mt->surf.phys_level0_sa.depth, level) :
+ mt->surf.phys_level0_sa.array_len));
+ else
+ assert(layer < mt->level[level].depth);
+}
+
+static enum isl_aux_state **
+create_aux_state_map(struct intel_mipmap_tree *mt,
+ enum isl_aux_state initial)
+{
+ const uint32_t levels = mt->last_level + 1;
+
+ uint32_t total_slices = 0;
+ for (uint32_t level = 0; level < levels; level++)
+ total_slices += mt->level[level].depth;
+
+ const size_t per_level_array_size = levels * sizeof(enum isl_aux_state *);
+
+ /* We're going to allocate a single chunk of data for both the per-level
+ * reference array and the arrays of aux_state. This makes cleanup
+ * significantly easier.
+ */
+ const size_t total_size = per_level_array_size +
+ total_slices * sizeof(enum isl_aux_state);
+ void *data = malloc(total_size);
+ if (data == NULL)
+ return NULL;
+
+ enum isl_aux_state **per_level_arr = data;
+ enum isl_aux_state *s = data + per_level_array_size;
+ for (uint32_t level = 0; level < levels; level++) {
+ per_level_arr[level] = s;
+ for (uint32_t a = 0; a < mt->level[level].depth; a++)
+ *(s++) = initial;
+ }
+ assert((void *)s == data + total_size);
+
+ return per_level_arr;
+}
+
+static void
+free_aux_state_map(enum isl_aux_state **state)
+{
+ free(state);
+}
+
+static struct intel_mipmap_tree *
+make_surface(struct brw_context *brw, GLenum target, mesa_format format,
+ unsigned first_level, unsigned last_level,
+ unsigned width0, unsigned height0, unsigned depth0,
+ unsigned num_samples, enum isl_tiling isl_tiling,
+ isl_surf_usage_flags_t isl_usage_flags, uint32_t alloc_flags,
+ struct brw_bo *bo)
+{
+ struct intel_mipmap_tree *mt = calloc(sizeof(*mt), 1);
+ if (!mt)
+ return NULL;
+
+ if (!create_mapping_table(target, first_level, last_level, depth0,
+ mt->level)) {
+ free(mt);
+ return NULL;
+ }
+
+ if (target == GL_TEXTURE_CUBE_MAP ||
+ target == GL_TEXTURE_CUBE_MAP_ARRAY)
+ isl_usage_flags |= ISL_SURF_USAGE_CUBE_BIT;
+
+ DBG("%s: %s %s %ux %u:%u:%u %d..%d <-- %p\n",
+ __func__,
+ _mesa_enum_to_string(target),
+ _mesa_get_format_name(format),
+ num_samples, width0, height0, depth0,
+ first_level, last_level, mt);
+
+ struct isl_surf_init_info init_info = {
+ .dim = get_isl_surf_dim(target),
+ .format = translate_tex_format(brw, format, false),
+ .width = width0,
+ .height = height0,
+ .depth = target == GL_TEXTURE_3D ? depth0 : 1,
+ .levels = last_level - first_level + 1,
+ .array_len = target == GL_TEXTURE_3D ? 1 : depth0,
+ .samples = MAX2(num_samples, 1),
+ .usage = isl_usage_flags,
+ .tiling_flags = 1u << isl_tiling
+ };
+
+ if (!isl_surf_init_s(&brw->isl_dev, &mt->surf, &init_info))
+ goto fail;
+
+ assert(mt->surf.size % mt->surf.row_pitch == 0);
+
+ if (!bo) {
+ mt->bo = brw_bo_alloc_tiled(brw->bufmgr, "isl-miptree",
+ mt->surf.size,
+ isl_tiling_to_bufmgr_tiling(isl_tiling),
+ mt->surf.row_pitch, alloc_flags);
+ if (!mt->bo)
+ goto fail;
+ } else {
+ mt->bo = bo;
+ }
+
+ mt->first_level = first_level;
+ mt->last_level = last_level;
+ mt->target = target;
+ mt->format = format;
+ mt->refcount = 1;
+ mt->aux_state = NULL;
+
+ return mt;
+
+fail:
+ intel_miptree_release(&mt);
+ return NULL;
+}
+
static struct intel_mipmap_tree *
miptree_create(struct brw_context *brw,
GLenum target,
GLuint num_samples,
uint32_t layout_flags)
{
+ if (brw->gen == 6 && format == MESA_FORMAT_S_UINT8)
+ return make_surface(brw, target, format, first_level, last_level,
+ width0, height0, depth0, num_samples, ISL_TILING_W,
+ ISL_SURF_USAGE_STENCIL_BIT |
+ ISL_SURF_USAGE_TEXTURE_BIT,
+ BO_ALLOC_FOR_RENDER, NULL);
+
struct intel_mipmap_tree *mt;
mesa_format tex_format = format;
mesa_format etc_format = MESA_FORMAT_NONE;
if (layout_flags & MIPTREE_LAYOUT_ACCELERATED_UPLOAD)
alloc_flags |= BO_ALLOC_FOR_RENDER;
- unsigned long pitch;
mt->etc_format = etc_format;
if (format == MESA_FORMAT_S_UINT8) {
/* Align to size of W tile, 64x64. */
- mt->bo = drm_bacon_bo_alloc_tiled(brw->bufmgr, "miptree",
- ALIGN(mt->total_width, 64),
- ALIGN(mt->total_height, 64),
- mt->cpp, &mt->tiling, &pitch,
- alloc_flags);
+ mt->bo = brw_bo_alloc_tiled_2d(brw->bufmgr, "miptree",
+ ALIGN(mt->total_width, 64),
+ ALIGN(mt->total_height, 64),
+ mt->cpp, mt->tiling, &mt->pitch,
+ alloc_flags);
} else {
- mt->bo = drm_bacon_bo_alloc_tiled(brw->bufmgr, "miptree",
- mt->total_width, mt->total_height,
- mt->cpp, &mt->tiling, &pitch,
- alloc_flags);
+ mt->bo = brw_bo_alloc_tiled_2d(brw->bufmgr, "miptree",
+ mt->total_width, mt->total_height,
+ mt->cpp, mt->tiling, &mt->pitch,
+ alloc_flags);
}
- mt->pitch = pitch;
+ if (layout_flags & MIPTREE_LAYOUT_FOR_SCANOUT)
+ mt->bo->cache_coherent = false;
return mt;
}
*/
if (brw->gen < 6 && mt->bo->size >= brw->max_gtt_map_object_size &&
mt->tiling == I915_TILING_Y) {
- unsigned long pitch = mt->pitch;
const uint32_t alloc_flags =
(layout_flags & MIPTREE_LAYOUT_ACCELERATED_UPLOAD) ?
BO_ALLOC_FOR_RENDER : 0;
mt->total_width, mt->total_height);
mt->tiling = I915_TILING_X;
- drm_bacon_bo_unreference(mt->bo);
- mt->bo = drm_bacon_bo_alloc_tiled(brw->bufmgr, "miptree",
- mt->total_width, mt->total_height, mt->cpp,
- &mt->tiling, &pitch, alloc_flags);
- mt->pitch = pitch;
+ brw_bo_unreference(mt->bo);
+ mt->bo = brw_bo_alloc_tiled_2d(brw->bufmgr, "miptree",
+ mt->total_width, mt->total_height, mt->cpp,
+ mt->tiling, &mt->pitch, alloc_flags);
}
mt->offset = 0;
struct intel_mipmap_tree *
intel_miptree_create_for_bo(struct brw_context *brw,
- drm_bacon_bo *bo,
+ struct brw_bo *bo,
mesa_format format,
uint32_t offset,
uint32_t width,
uint32_t tiling, swizzle;
GLenum target;
- drm_bacon_bo_get_tiling(bo, &tiling, &swizzle);
+ brw_bo_get_tiling(bo, &tiling, &swizzle);
/* Nothing will be able to use this miptree with the BO if the offset isn't
* aligned.
if (!mt)
return NULL;
- drm_bacon_bo_reference(bo);
+ brw_bo_reference(bo);
mt->bo = bo;
mt->pitch = pitch;
mt->offset = offset;
void
intel_update_winsys_renderbuffer_miptree(struct brw_context *intel,
struct intel_renderbuffer *irb,
- drm_bacon_bo *bo,
+ struct brw_bo *bo,
uint32_t width, uint32_t height,
uint32_t pitch)
{
if (hiz_buf == NULL)
return;
- if (hiz_buf->mt)
- intel_miptree_release(&hiz_buf->mt);
- else
- drm_bacon_bo_unreference(hiz_buf->aux_base.bo);
+ brw_bo_unreference(hiz_buf->aux_base.bo);
free(hiz_buf);
}
DBG("%s deleting %p\n", __func__, *mt);
- drm_bacon_bo_unreference((*mt)->bo);
+ brw_bo_unreference((*mt)->bo);
intel_miptree_release(&(*mt)->stencil_mt);
intel_miptree_release(&(*mt)->r8stencil_mt);
intel_miptree_hiz_buffer_free((*mt)->hiz_buf);
if ((*mt)->mcs_buf) {
- drm_bacon_bo_unreference((*mt)->mcs_buf->bo);
+ brw_bo_unreference((*mt)->mcs_buf->bo);
free((*mt)->mcs_buf);
}
- intel_resolve_map_clear(&(*mt)->hiz_map);
- intel_resolve_map_clear(&(*mt)->color_resolve_map);
+ free_aux_state_map((*mt)->aux_state);
intel_miptree_release(&(*mt)->plane[0]);
intel_miptree_release(&(*mt)->plane[1]);
if (mt->target == GL_TEXTURE_CUBE_MAP)
depth = 6;
+ if (mt->surf.size > 0) {
+ if (level >= mt->surf.levels)
+ return false;
+
+ const unsigned level_depth =
+ mt->surf.dim == ISL_SURF_DIM_3D ?
+ minify(mt->surf.logical_level0_px.depth, level) :
+ mt->surf.logical_level0_px.array_len;
+
+ return width == minify(mt->surf.logical_level0_px.width, level) &&
+ height == minify(mt->surf.logical_level0_px.height, level) &&
+ depth == level_depth &&
+ MAX2(image->NumSamples, 1) == mt->surf.samples;
+ }
+
int level_depth = mt->level[level].depth;
if (mt->num_samples > 1) {
switch (mt->msaa_layout) {
DBG("%s level %d, depth %d, offset %d,%d\n", __func__,
level, d, x, y);
- assert(mt->level[level].slice == NULL);
+ assert(mt->level[level].slice);
- mt->level[level].slice = calloc(d, sizeof(*mt->level[0].slice));
mt->level[level].slice[0].x_offset = mt->level[level].level_x;
mt->level[level].slice[0].y_offset = mt->level[level].level_y;
}
GLuint level, GLuint slice,
GLuint *x, GLuint *y)
{
+ if (mt->surf.size > 0) {
+ uint32_t x_offset_sa, y_offset_sa;
+
+ /* Given level is relative to level zero while the miptree may be
+ * represent just a subset of all levels starting from 'first_level'.
+ */
+ assert(level >= mt->first_level);
+ level -= mt->first_level;
+
+ const unsigned z = mt->surf.dim == ISL_SURF_DIM_3D ? slice : 0;
+ slice = mt->surf.dim == ISL_SURF_DIM_3D ? 0 : slice;
+ isl_surf_get_image_offset_sa(&mt->surf, level, slice, z,
+ &x_offset_sa, &y_offset_sa);
+
+ *x = x_offset_sa;
+ *y = y_offset_sa;
+ return;
+ }
+
assert(slice < mt->level[level].depth);
*x = mt->level[level].slice[slice].x_offset;
static void
intel_miptree_copy_slice_sw(struct brw_context *brw,
- struct intel_mipmap_tree *dst_mt,
struct intel_mipmap_tree *src_mt,
- int level,
- int slice,
- int width,
- int height)
+ unsigned src_level, unsigned src_layer,
+ struct intel_mipmap_tree *dst_mt,
+ unsigned dst_level, unsigned dst_layer,
+ unsigned width, unsigned height)
{
void *src, *dst;
ptrdiff_t src_stride, dst_stride;
- int cpp = dst_mt->cpp;
+ const unsigned cpp = dst_mt->surf.size > 0 ?
+ (isl_format_get_layout(dst_mt->surf.format)->bpb / 8) : dst_mt->cpp;
intel_miptree_map(brw, src_mt,
- level, slice,
+ src_level, src_layer,
0, 0,
width, height,
GL_MAP_READ_BIT | BRW_MAP_DIRECT_BIT,
&src, &src_stride);
intel_miptree_map(brw, dst_mt,
- level, slice,
+ dst_level, dst_layer,
0, 0,
width, height,
GL_MAP_WRITE_BIT | GL_MAP_INVALIDATE_RANGE_BIT |
}
}
- intel_miptree_unmap(brw, dst_mt, level, slice);
- intel_miptree_unmap(brw, src_mt, level, slice);
+ intel_miptree_unmap(brw, dst_mt, dst_level, dst_layer);
+ intel_miptree_unmap(brw, src_mt, src_level, src_layer);
/* Don't forget to copy the stencil data over, too. We could have skipped
* passing BRW_MAP_DIRECT_BIT, but that would have meant intel_miptree_map
*/
if (dst_mt->stencil_mt) {
assert(src_mt->stencil_mt);
- intel_miptree_copy_slice_sw(brw, dst_mt->stencil_mt, src_mt->stencil_mt,
- level, slice, width, height);
+ intel_miptree_copy_slice_sw(brw,
+ src_mt->stencil_mt, src_level, src_layer,
+ dst_mt->stencil_mt, dst_level, dst_layer,
+ width, height);
}
}
-static void
+void
intel_miptree_copy_slice(struct brw_context *brw,
- struct intel_mipmap_tree *dst_mt,
- struct intel_mipmap_tree *src_mt,
- int level,
- int face,
- int depth)
+ struct intel_mipmap_tree *src_mt,
+ unsigned src_level, unsigned src_layer,
+ struct intel_mipmap_tree *dst_mt,
+ unsigned dst_level, unsigned dst_layer)
{
mesa_format format = src_mt->format;
- uint32_t width = minify(src_mt->physical_width0, level - src_mt->first_level);
- uint32_t height = minify(src_mt->physical_height0, level - src_mt->first_level);
- int slice;
-
- if (face > 0)
- slice = face;
- else
- slice = depth;
+ uint32_t width, height;
+
+ if (src_mt->surf.size > 0) {
+ width = minify(src_mt->surf.phys_level0_sa.width,
+ src_level - src_mt->first_level);
+ height = minify(src_mt->surf.phys_level0_sa.height,
+ src_level - src_mt->first_level);
+
+ if (src_mt->surf.dim == ISL_SURF_DIM_3D)
+ assert(src_layer < minify(src_mt->surf.phys_level0_sa.depth,
+ src_level - src_mt->first_level));
+ else
+ assert(src_layer < src_mt->surf.phys_level0_sa.array_len);
+ } else {
+ width = minify(src_mt->physical_width0,
+ src_level - src_mt->first_level);
+ height = minify(src_mt->physical_height0,
+ src_level - src_mt->first_level);
+ assert(src_layer < src_mt->level[src_level].depth);
+ }
- assert(depth < src_mt->level[level].depth);
assert(src_mt->format == dst_mt->format);
if (dst_mt->compressed) {
*/
if (src_mt->stencil_mt) {
intel_miptree_copy_slice_sw(brw,
- dst_mt, src_mt,
- level, slice,
+ src_mt, src_level, src_layer,
+ dst_mt, dst_level, dst_layer,
width, height);
return;
}
uint32_t dst_x, dst_y, src_x, src_y;
- intel_miptree_get_image_offset(dst_mt, level, slice, &dst_x, &dst_y);
- intel_miptree_get_image_offset(src_mt, level, slice, &src_x, &src_y);
+ intel_miptree_get_image_offset(dst_mt, dst_level, dst_layer,
+ &dst_x, &dst_y);
+ intel_miptree_get_image_offset(src_mt, src_level, src_layer,
+ &src_x, &src_y);
DBG("validate blit mt %s %p %d,%d/%d -> mt %s %p %d,%d/%d (%dx%d)\n",
_mesa_get_format_name(src_mt->format),
width, height);
if (!intel_miptree_blit(brw,
- src_mt, level, slice, 0, 0, false,
- dst_mt, level, slice, 0, 0, false,
+ src_mt, src_level, src_layer, 0, 0, false,
+ dst_mt, dst_level, dst_layer, 0, 0, false,
width, height, GL_COPY)) {
perf_debug("miptree validate blit for %s failed\n",
_mesa_get_format_name(format));
- intel_miptree_copy_slice_sw(brw, dst_mt, src_mt, level, slice,
+ intel_miptree_copy_slice_sw(brw,
+ src_mt, src_level, src_layer,
+ dst_mt, dst_level, dst_layer,
width, height);
}
}
struct intel_texture_object *intel_obj =
intel_texture_object(intelImage->base.Base.TexObject);
int level = intelImage->base.Base.Level;
- int face = intelImage->base.Base.Face;
-
- GLuint depth;
- if (intel_obj->base.Target == GL_TEXTURE_1D_ARRAY)
- depth = intelImage->base.Base.Height;
- else
- depth = intelImage->base.Base.Depth;
+ const unsigned face = intelImage->base.Base.Face;
+ unsigned start_layer, end_layer;
+
+ if (intel_obj->base.Target == GL_TEXTURE_1D_ARRAY) {
+ assert(face == 0);
+ assert(intelImage->base.Base.Height);
+ start_layer = 0;
+ end_layer = intelImage->base.Base.Height - 1;
+ } else if (face > 0) {
+ start_layer = face;
+ end_layer = face;
+ } else {
+ assert(intelImage->base.Base.Depth);
+ start_layer = 0;
+ end_layer = intelImage->base.Base.Depth - 1;
+ }
if (!invalidate) {
- for (int slice = 0; slice < depth; slice++) {
- intel_miptree_copy_slice(brw, dst_mt, src_mt, level, face, slice);
+ for (unsigned i = start_layer; i <= end_layer; i++) {
+ intel_miptree_copy_slice(brw,
+ src_mt, level, i,
+ dst_mt, level, i);
}
}
*
* Note: the clear value for MCS buffers is all 1's, so we memset to 0xff.
*/
- const int ret = brw_bo_map_gtt(brw, mt->mcs_buf->bo, "miptree");
- if (unlikely(ret)) {
+ void *map = brw_bo_map(brw, mt->mcs_buf->bo, MAP_WRITE);
+ if (unlikely(map == NULL)) {
fprintf(stderr, "Failed to map mcs buffer into GTT\n");
- drm_bacon_bo_unreference(mt->mcs_buf->bo);
+ brw_bo_unreference(mt->mcs_buf->bo);
free(mt->mcs_buf);
return;
}
- void *data = mt->mcs_buf->bo->virtual;
+ void *data = map;
memset(data, init_value, mt->mcs_buf->size);
- drm_bacon_bo_unmap(mt->mcs_buf->bo);
+ brw_bo_unmap(mt->mcs_buf->bo);
}
static struct intel_miptree_aux_buffer *
-intel_mcs_miptree_buf_create(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- mesa_format format,
- unsigned mcs_width,
- unsigned mcs_height,
- uint32_t layout_flags)
+intel_alloc_aux_buffer(struct brw_context *brw,
+ const char *name,
+ const struct isl_surf *aux_surf,
+ uint32_t alloc_flags,
+ struct intel_mipmap_tree *mt)
{
struct intel_miptree_aux_buffer *buf = calloc(sizeof(*buf), 1);
- struct intel_mipmap_tree *temp_mt;
-
if (!buf)
- return NULL;
+ return false;
- /* From the Ivy Bridge PRM, Vol4 Part1 p76, "MCS Base Address":
- *
- * "The MCS surface must be stored as Tile Y."
+ buf->size = aux_surf->size;
+ buf->pitch = aux_surf->row_pitch;
+ buf->qpitch = isl_surf_get_array_pitch_sa_rows(aux_surf);
+
+ /* ISL has stricter set of alignment rules then the drm allocator.
+ * Therefore one can pass the ISL dimensions in terms of bytes instead of
+ * trying to recalculate based on different format block sizes.
*/
- layout_flags |= MIPTREE_LAYOUT_TILING_Y;
- temp_mt = miptree_create(brw,
- mt->target,
- format,
- mt->first_level,
- mt->last_level,
- mcs_width,
- mcs_height,
- mt->logical_depth0,
- 0 /* num_samples */,
- layout_flags);
- if (!temp_mt) {
+ buf->bo = brw_bo_alloc_tiled(brw->bufmgr, name, buf->size,
+ I915_TILING_Y, buf->pitch, alloc_flags);
+ if (!buf->bo) {
free(buf);
return NULL;
}
- buf->bo = temp_mt->bo;
- buf->offset = temp_mt->offset;
- buf->size = temp_mt->total_height * temp_mt->pitch;
- buf->pitch = temp_mt->pitch;
- buf->qpitch = temp_mt->qpitch;
-
- /* Just hang on to the BO which backs the AUX buffer; the rest of the miptree
- * structure should go away. We use miptree create simply as a means to make
- * sure all the constraints for the buffer are satisfied.
- */
- drm_bacon_bo_reference(temp_mt->bo);
- intel_miptree_release(&temp_mt);
+ buf->surf = *aux_surf;
return buf;
}
assert(mt->mcs_buf == NULL);
assert((mt->aux_disable & INTEL_AUX_DISABLE_MCS) == 0);
- /* Choose the correct format for the MCS buffer. All that really matters
- * is that we allocate the right buffer size, since we'll always be
- * accessing this miptree using MCS-specific hardware mechanisms, which
- * infer the correct format based on num_samples.
- */
- mesa_format format;
- switch (num_samples) {
- case 2:
- case 4:
- /* 8 bits/pixel are required for MCS data when using 4x MSAA (2 bits for
- * each sample).
- */
- format = MESA_FORMAT_R_UNORM8;
- break;
- case 8:
- /* 32 bits/pixel are required for MCS data when using 8x MSAA (3 bits
- * for each sample, plus 8 padding bits).
- */
- format = MESA_FORMAT_R_UINT32;
- break;
- case 16:
- /* 64 bits/pixel are required for MCS data when using 16x MSAA (4 bits
- * for each sample).
- */
- format = MESA_FORMAT_RG_UINT32;
- break;
- default:
- unreachable("Unrecognized sample count in intel_miptree_alloc_mcs");
- };
+ /* Multisampled miptrees are only supported for single level. */
+ assert(mt->first_level == 0);
+ enum isl_aux_state **aux_state =
+ create_aux_state_map(mt, ISL_AUX_STATE_CLEAR);
+ if (!aux_state)
+ return false;
- mt->mcs_buf =
- intel_mcs_miptree_buf_create(brw, mt,
- format,
- mt->logical_width0,
- mt->logical_height0,
- MIPTREE_LAYOUT_ACCELERATED_UPLOAD);
- if (!mt->mcs_buf)
+ struct isl_surf temp_main_surf;
+ struct isl_surf temp_mcs_surf;
+
+ /* Create first an ISL presentation for the main color surface and let ISL
+ * calculate equivalent MCS surface against it.
+ */
+ intel_miptree_get_isl_surf(brw, mt, &temp_main_surf);
+ assert(isl_surf_get_mcs_surf(&brw->isl_dev, &temp_main_surf,
+ &temp_mcs_surf));
+
+ const uint32_t alloc_flags = BO_ALLOC_FOR_RENDER;
+ mt->mcs_buf = intel_alloc_aux_buffer(brw, "mcs-miptree",
+ &temp_mcs_surf, alloc_flags, mt);
+ if (!mt->mcs_buf) {
+ free(aux_state);
return false;
+ }
- intel_miptree_init_mcs(brw, mt, 0xFF);
+ mt->aux_state = aux_state;
- /* Multisampled miptrees are only supported for single level. */
- assert(mt->first_level == 0);
- intel_miptree_set_fast_clear_state(brw, mt, mt->first_level, 0,
- mt->logical_depth0,
- INTEL_FAST_CLEAR_STATE_CLEAR);
+ intel_miptree_init_mcs(brw, mt, 0xFF);
return true;
}
-
bool
intel_miptree_alloc_non_msrt_mcs(struct brw_context *brw,
struct intel_mipmap_tree *mt,
assert(temp_ccs_surf.size &&
(temp_ccs_surf.size % temp_ccs_surf.row_pitch == 0));
- struct intel_miptree_aux_buffer *buf = calloc(sizeof(*buf), 1);
- if (!buf)
+ enum isl_aux_state **aux_state =
+ create_aux_state_map(mt, ISL_AUX_STATE_PASS_THROUGH);
+ if (!aux_state)
return false;
- buf->size = temp_ccs_surf.size;
- buf->pitch = temp_ccs_surf.row_pitch;
- buf->qpitch = isl_surf_get_array_pitch_sa_rows(&temp_ccs_surf);
-
/* In case of compression mcs buffer needs to be initialised requiring the
* buffer to be immediately mapped to cpu space for writing. Therefore do
* not use the gpu access flag which can cause an unnecessary delay if the
*/
const uint32_t alloc_flags =
is_lossless_compressed ? 0 : BO_ALLOC_FOR_RENDER;
- uint32_t tiling = I915_TILING_Y;
- unsigned long pitch;
-
- /* ISL has stricter set of alignment rules then the drm allocator.
- * Therefore one can pass the ISL dimensions in terms of bytes instead of
- * trying to recalculate based on different format block sizes.
- */
- buf->bo = drm_bacon_bo_alloc_tiled(brw->bufmgr, "ccs-miptree",
- buf->pitch, buf->size / buf->pitch,
- 1, &tiling, &pitch, alloc_flags);
- if (buf->bo) {
- assert(pitch == buf->pitch);
- assert(tiling == I915_TILING_Y);
- } else {
- free(buf);
+ mt->mcs_buf = intel_alloc_aux_buffer(brw, "ccs-miptree",
+ &temp_ccs_surf, alloc_flags, mt);
+ if (!mt->mcs_buf) {
+ free(aux_state);
return false;
}
-
- mt->mcs_buf = buf;
+
+ mt->aux_state = aux_state;
/* From Gen9 onwards single-sampled (non-msrt) auxiliary buffers are
* used for lossless compression which requires similar initialisation
hz_height = DIV_ROUND_UP(hz_qpitch * Z0, 2 * 8) * 8;
}
- unsigned long pitch;
- uint32_t tiling = I915_TILING_Y;
- buf->aux_base.bo = drm_bacon_bo_alloc_tiled(brw->bufmgr, "hiz",
- hz_width, hz_height, 1,
- &tiling, &pitch,
- BO_ALLOC_FOR_RENDER);
+ buf->aux_base.bo = brw_bo_alloc_tiled_2d(brw->bufmgr, "hiz",
+ hz_width, hz_height, 1,
+ I915_TILING_Y, &buf->aux_base.pitch,
+ BO_ALLOC_FOR_RENDER);
if (!buf->aux_base.bo) {
free(buf);
return NULL;
- } else if (tiling != I915_TILING_Y) {
- drm_bacon_bo_unreference(buf->aux_base.bo);
- free(buf);
- return NULL;
}
buf->aux_base.size = hz_width * hz_height;
- buf->aux_base.pitch = pitch;
return buf;
}
hz_height = DIV_ROUND_UP(buf->aux_base.qpitch, 2 * 8) * 8 * Z0;
}
- unsigned long pitch;
- uint32_t tiling = I915_TILING_Y;
- buf->aux_base.bo = drm_bacon_bo_alloc_tiled(brw->bufmgr, "hiz",
- hz_width, hz_height, 1,
- &tiling, &pitch,
- BO_ALLOC_FOR_RENDER);
+ buf->aux_base.bo = brw_bo_alloc_tiled_2d(brw->bufmgr, "hiz",
+ hz_width, hz_height, 1,
+ I915_TILING_Y, &buf->aux_base.pitch,
+ BO_ALLOC_FOR_RENDER);
if (!buf->aux_base.bo) {
free(buf);
return NULL;
- } else if (tiling != I915_TILING_Y) {
- drm_bacon_bo_unreference(buf->aux_base.bo);
- free(buf);
- return NULL;
}
buf->aux_base.size = hz_width * hz_height;
- buf->aux_base.pitch = pitch;
return buf;
}
struct intel_mipmap_tree *mt)
{
struct intel_miptree_hiz_buffer *buf = calloc(sizeof(*buf), 1);
- uint32_t layout_flags = MIPTREE_LAYOUT_ACCELERATED_UPLOAD;
-
- if (brw->gen == 6)
- layout_flags |= MIPTREE_LAYOUT_FORCE_ALL_SLICE_AT_LOD;
-
if (!buf)
return NULL;
- layout_flags |= MIPTREE_LAYOUT_TILING_ANY;
- buf->mt = intel_miptree_create(brw,
- mt->target,
- mt->format,
- mt->first_level,
- mt->last_level,
- mt->logical_width0,
- mt->logical_height0,
- mt->logical_depth0,
- mt->num_samples,
- layout_flags);
- if (!buf->mt) {
+ struct isl_surf temp_main_surf;
+ intel_miptree_get_isl_surf(brw, mt, &temp_main_surf);
+
+ if (!isl_surf_get_hiz_surf(&brw->isl_dev, &temp_main_surf,
+ &buf->aux_base.surf)) {
free(buf);
return NULL;
}
- buf->aux_base.bo = buf->mt->bo;
- buf->aux_base.size = buf->mt->total_height * buf->mt->pitch;
- buf->aux_base.pitch = buf->mt->pitch;
-
- /* On gen6 hiz is unconditionally laid out packing all slices
- * at each level-of-detail (LOD). This means there is no valid qpitch
- * setting. In fact, this is ignored when hardware is setup - there is no
- * hardware qpitch setting of hiz on gen6.
- */
- buf->aux_base.qpitch = 0;
+ struct isl_surf *surf = &buf->aux_base.surf;
+ buf->aux_base.bo = brw_bo_alloc_tiled(brw->bufmgr, "hiz", surf->size,
+ I915_TILING_Y, surf->row_pitch,
+ BO_ALLOC_FOR_RENDER);
+ if (!buf->aux_base.bo) {
+ free(buf);
+ return NULL;
+ }
return buf;
}
assert(mt->hiz_buf == NULL);
assert((mt->aux_disable & INTEL_AUX_DISABLE_HIZ) == 0);
+ enum isl_aux_state **aux_state =
+ create_aux_state_map(mt, ISL_AUX_STATE_AUX_INVALID);
+ if (!aux_state)
+ return false;
+
if (brw->gen == 7) {
mt->hiz_buf = intel_gen7_hiz_buf_create(brw, mt);
} else if (brw->gen >= 8) {
mt->hiz_buf = intel_hiz_miptree_buf_create(brw, mt);
}
- if (!mt->hiz_buf)
+ if (!mt->hiz_buf) {
+ free(aux_state);
return false;
+ }
- /* Mark that all slices need a HiZ resolve. */
- for (unsigned level = mt->first_level; level <= mt->last_level; ++level) {
- if (!intel_miptree_level_enable_hiz(brw, mt, level))
- continue;
-
- for (unsigned layer = 0; layer < mt->level[level].depth; ++layer) {
- struct intel_resolve_map *m = malloc(sizeof(struct intel_resolve_map));
- exec_node_init(&m->link);
- m->level = level;
- m->layer = layer;
- m->need = BLORP_HIZ_OP_HIZ_RESOLVE;
+ for (unsigned level = mt->first_level; level <= mt->last_level; ++level)
+ intel_miptree_level_enable_hiz(brw, mt, level);
- exec_list_push_tail(&mt->hiz_map, &m->link);
- }
- }
+ mt->aux_state = aux_state;
return true;
}
* Does the miptree slice have hiz enabled?
*/
bool
-intel_miptree_level_has_hiz(struct intel_mipmap_tree *mt, uint32_t level)
+intel_miptree_level_has_hiz(const struct intel_mipmap_tree *mt, uint32_t level)
{
intel_miptree_check_level_layer(mt, level, 0);
return mt->level[level].has_hiz;
}
-void
-intel_miptree_slice_set_needs_hiz_resolve(struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer)
+bool
+intel_miptree_has_color_unresolved(const struct intel_mipmap_tree *mt,
+ unsigned start_level, unsigned num_levels,
+ unsigned start_layer, unsigned num_layers)
{
- if (!intel_miptree_level_has_hiz(mt, level))
- return;
+ assert(_mesa_is_format_color_format(mt->format));
- intel_resolve_map_set(&mt->hiz_map,
- level, layer, BLORP_HIZ_OP_HIZ_RESOLVE);
-}
+ if (!mt->mcs_buf)
+ return false;
+ /* Clamp the level range to fit the miptree */
+ assert(start_level + num_levels >= start_level);
+ const uint32_t last_level =
+ MIN2(mt->last_level, start_level + num_levels - 1);
+ start_level = MAX2(mt->first_level, start_level);
+ num_levels = last_level - start_level + 1;
+
+ for (uint32_t level = start_level; level <= last_level; level++) {
+ const uint32_t level_layers = MIN2(num_layers, mt->level[level].depth);
+ for (unsigned a = 0; a < level_layers; a++) {
+ enum isl_aux_state aux_state =
+ intel_miptree_get_aux_state(mt, level, start_layer + a);
+ assert(aux_state != ISL_AUX_STATE_AUX_INVALID);
+ if (aux_state != ISL_AUX_STATE_PASS_THROUGH)
+ return true;
+ }
+ }
-void
-intel_miptree_slice_set_needs_depth_resolve(struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer)
+ return false;
+}
+
+static void
+intel_miptree_check_color_resolve(const struct brw_context *brw,
+ const struct intel_mipmap_tree *mt,
+ unsigned level, unsigned layer)
{
- if (!intel_miptree_level_has_hiz(mt, level))
+
+ if ((mt->aux_disable & INTEL_AUX_DISABLE_CCS) || !mt->mcs_buf)
return;
- intel_resolve_map_set(&mt->hiz_map,
- level, layer, BLORP_HIZ_OP_DEPTH_RESOLVE);
+ /* Fast color clear is supported for mipmapped surfaces only on Gen8+. */
+ assert(brw->gen >= 8 ||
+ (level == 0 && mt->first_level == 0 && mt->last_level == 0));
+
+ /* Compression of arrayed msaa surfaces is supported. */
+ if (mt->num_samples > 1)
+ return;
+
+ /* Fast color clear is supported for non-msaa arrays only on Gen8+. */
+ assert(brw->gen >= 8 || (layer == 0 && mt->logical_depth0 == 1));
+
+ (void)level;
+ (void)layer;
}
-void
-intel_miptree_set_all_slices_need_depth_resolve(struct intel_mipmap_tree *mt,
- uint32_t level)
+static enum blorp_fast_clear_op
+get_ccs_d_resolve_op(enum isl_aux_state aux_state,
+ bool ccs_supported, bool fast_clear_supported)
{
- uint32_t layer;
- uint32_t end_layer = mt->level[level].depth;
+ assert(ccs_supported == fast_clear_supported);
- for (layer = 0; layer < end_layer; layer++) {
- intel_miptree_slice_set_needs_depth_resolve(mt, level, layer);
+ switch (aux_state) {
+ case ISL_AUX_STATE_CLEAR:
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ if (!ccs_supported)
+ return BLORP_FAST_CLEAR_OP_RESOLVE_FULL;
+ else
+ return BLORP_FAST_CLEAR_OP_NONE;
+
+ case ISL_AUX_STATE_PASS_THROUGH:
+ return BLORP_FAST_CLEAR_OP_NONE;
+
+ case ISL_AUX_STATE_RESOLVED:
+ case ISL_AUX_STATE_AUX_INVALID:
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ break;
}
+
+ unreachable("Invalid aux state for CCS_D");
}
-static bool
-intel_miptree_slice_resolve(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer,
- enum blorp_hiz_op need)
+static enum blorp_fast_clear_op
+get_ccs_e_resolve_op(enum isl_aux_state aux_state,
+ bool ccs_supported, bool fast_clear_supported)
{
- intel_miptree_check_level_layer(mt, level, layer);
+ switch (aux_state) {
+ case ISL_AUX_STATE_CLEAR:
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ if (!ccs_supported)
+ return BLORP_FAST_CLEAR_OP_RESOLVE_FULL;
+ else if (!fast_clear_supported)
+ return BLORP_FAST_CLEAR_OP_RESOLVE_PARTIAL;
+ else
+ return BLORP_FAST_CLEAR_OP_NONE;
- struct intel_resolve_map *item =
- intel_resolve_map_get(&mt->hiz_map, level, layer);
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ if (!ccs_supported)
+ return BLORP_FAST_CLEAR_OP_RESOLVE_FULL;
+ else
+ return BLORP_FAST_CLEAR_OP_NONE;
- if (!item || item->need != need)
- return false;
+ case ISL_AUX_STATE_PASS_THROUGH:
+ return BLORP_FAST_CLEAR_OP_NONE;
- intel_hiz_exec(brw, mt, level, layer, need);
- intel_resolve_map_remove(item);
- return true;
-}
+ case ISL_AUX_STATE_RESOLVED:
+ case ISL_AUX_STATE_AUX_INVALID:
+ break;
+ }
-bool
-intel_miptree_slice_resolve_hiz(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer)
-{
- return intel_miptree_slice_resolve(brw, mt, level, layer,
- BLORP_HIZ_OP_HIZ_RESOLVE);
+ unreachable("Invalid aux state for CCS_E");
}
-bool
-intel_miptree_slice_resolve_depth(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- uint32_t level,
- uint32_t layer)
+static void
+intel_miptree_prepare_ccs_access(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ uint32_t level, uint32_t layer,
+ bool aux_supported,
+ bool fast_clear_supported)
{
- return intel_miptree_slice_resolve(brw, mt, level, layer,
- BLORP_HIZ_OP_DEPTH_RESOLVE);
-}
+ enum isl_aux_state aux_state = intel_miptree_get_aux_state(mt, level, layer);
-static bool
-intel_miptree_all_slices_resolve(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- enum blorp_hiz_op need)
-{
- bool did_resolve = false;
+ enum blorp_fast_clear_op resolve_op;
+ if (intel_miptree_is_lossless_compressed(brw, mt)) {
+ resolve_op = get_ccs_e_resolve_op(aux_state, aux_supported,
+ fast_clear_supported);
+ } else {
+ resolve_op = get_ccs_d_resolve_op(aux_state, aux_supported,
+ fast_clear_supported);
+ }
- foreach_list_typed_safe(struct intel_resolve_map, map, link, &mt->hiz_map) {
- if (map->need != need)
- continue;
+ if (resolve_op != BLORP_FAST_CLEAR_OP_NONE) {
+ intel_miptree_check_color_resolve(brw, mt, level, layer);
+ brw_blorp_resolve_color(brw, mt, level, layer, resolve_op);
- intel_hiz_exec(brw, mt, map->level, map->layer, need);
- intel_resolve_map_remove(map);
- did_resolve = true;
- }
+ switch (resolve_op) {
+ case BLORP_FAST_CLEAR_OP_RESOLVE_FULL:
+ /* The CCS full resolve operation destroys the CCS and sets it to the
+ * pass-through state. (You can also think of this as being both a
+ * resolve and an ambiguate in one operation.)
+ */
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_PASS_THROUGH);
+ break;
+
+ case BLORP_FAST_CLEAR_OP_RESOLVE_PARTIAL:
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_NO_CLEAR);
+ break;
- return did_resolve;
+ default:
+ unreachable("Invalid resolve op");
+ }
+ }
}
-bool
-intel_miptree_all_slices_resolve_hiz(struct brw_context *brw,
- struct intel_mipmap_tree *mt)
+static void
+intel_miptree_finish_ccs_write(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ uint32_t level, uint32_t layer,
+ bool written_with_ccs)
{
- return intel_miptree_all_slices_resolve(brw, mt,
- BLORP_HIZ_OP_HIZ_RESOLVE);
+ enum isl_aux_state aux_state = intel_miptree_get_aux_state(mt, level, layer);
+
+ if (intel_miptree_is_lossless_compressed(brw, mt)) {
+ switch (aux_state) {
+ case ISL_AUX_STATE_CLEAR:
+ assert(written_with_ccs);
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_CLEAR);
+ break;
+
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ assert(written_with_ccs);
+ break; /* Nothing to do */
+
+ case ISL_AUX_STATE_PASS_THROUGH:
+ if (written_with_ccs) {
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_NO_CLEAR);
+ } else {
+ /* Nothing to do */
+ }
+ break;
+
+ case ISL_AUX_STATE_RESOLVED:
+ case ISL_AUX_STATE_AUX_INVALID:
+ unreachable("Invalid aux state for CCS_E");
+ }
+ } else {
+ /* CCS_D is a bit simpler */
+ switch (aux_state) {
+ case ISL_AUX_STATE_CLEAR:
+ assert(written_with_ccs);
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_CLEAR);
+ break;
+
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ assert(written_with_ccs);
+ break; /* Nothing to do */
+
+ case ISL_AUX_STATE_PASS_THROUGH:
+ /* Nothing to do */
+ break;
+
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ case ISL_AUX_STATE_RESOLVED:
+ case ISL_AUX_STATE_AUX_INVALID:
+ unreachable("Invalid aux state for CCS_D");
+ }
+ }
}
-bool
-intel_miptree_all_slices_resolve_depth(struct brw_context *brw,
- struct intel_mipmap_tree *mt)
-{
- return intel_miptree_all_slices_resolve(brw, mt,
- BLORP_HIZ_OP_DEPTH_RESOLVE);
+static void
+intel_miptree_finish_mcs_write(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ uint32_t level, uint32_t layer,
+ bool written_with_aux)
+{
+ switch (intel_miptree_get_aux_state(mt, level, layer)) {
+ case ISL_AUX_STATE_CLEAR:
+ assert(written_with_aux);
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_CLEAR);
+ break;
+
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ assert(written_with_aux);
+ break; /* Nothing to do */
+
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ case ISL_AUX_STATE_RESOLVED:
+ case ISL_AUX_STATE_PASS_THROUGH:
+ case ISL_AUX_STATE_AUX_INVALID:
+ unreachable("Invalid aux state for MCS");
+ }
}
-enum intel_fast_clear_state
-intel_miptree_get_fast_clear_state(const struct intel_mipmap_tree *mt,
- unsigned level, unsigned layer)
-{
- intel_miptree_check_level_layer(mt, level, layer);
+static void
+intel_miptree_prepare_hiz_access(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ uint32_t level, uint32_t layer,
+ bool hiz_supported, bool fast_clear_supported)
+{
+ enum blorp_hiz_op hiz_op = BLORP_HIZ_OP_NONE;
+ switch (intel_miptree_get_aux_state(mt, level, layer)) {
+ case ISL_AUX_STATE_CLEAR:
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ if (!hiz_supported || !fast_clear_supported)
+ hiz_op = BLORP_HIZ_OP_DEPTH_RESOLVE;
+ break;
+
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ if (!hiz_supported)
+ hiz_op = BLORP_HIZ_OP_DEPTH_RESOLVE;
+ break;
- const struct intel_resolve_map *item =
- intel_resolve_map_const_get(&mt->color_resolve_map, level, layer);
+ case ISL_AUX_STATE_PASS_THROUGH:
+ case ISL_AUX_STATE_RESOLVED:
+ break;
- if (!item)
- return INTEL_FAST_CLEAR_STATE_RESOLVED;
+ case ISL_AUX_STATE_AUX_INVALID:
+ if (hiz_supported)
+ hiz_op = BLORP_HIZ_OP_HIZ_RESOLVE;
+ break;
+ }
+
+ if (hiz_op != BLORP_HIZ_OP_NONE) {
+ intel_hiz_exec(brw, mt, level, layer, 1, hiz_op);
+
+ switch (hiz_op) {
+ case BLORP_HIZ_OP_DEPTH_RESOLVE:
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_RESOLVED);
+ break;
- return item->fast_clear_state;
+ case BLORP_HIZ_OP_HIZ_RESOLVE:
+ /* The HiZ resolve operation is actually an ambiguate */
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_PASS_THROUGH);
+ break;
+
+ default:
+ unreachable("Invalid HiZ op");
+ }
+ }
}
static void
-intel_miptree_check_color_resolve(const struct brw_context *brw,
- const struct intel_mipmap_tree *mt,
- unsigned level, unsigned layer)
-{
+intel_miptree_finish_hiz_write(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ uint32_t level, uint32_t layer,
+ bool written_with_hiz)
+{
+ switch (intel_miptree_get_aux_state(mt, level, layer)) {
+ case ISL_AUX_STATE_CLEAR:
+ assert(written_with_hiz);
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_CLEAR);
+ break;
- if ((mt->aux_disable & INTEL_AUX_DISABLE_CCS) || !mt->mcs_buf)
- return;
+ case ISL_AUX_STATE_COMPRESSED_NO_CLEAR:
+ case ISL_AUX_STATE_COMPRESSED_CLEAR:
+ assert(written_with_hiz);
+ break; /* Nothing to do */
- /* Fast color clear is supported for mipmapped surfaces only on Gen8+. */
- assert(brw->gen >= 8 ||
- (level == 0 && mt->first_level == 0 && mt->last_level == 0));
+ case ISL_AUX_STATE_RESOLVED:
+ if (written_with_hiz) {
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_NO_CLEAR);
+ } else {
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_AUX_INVALID);
+ }
+ break;
- /* Compression of arrayed msaa surfaces is supported. */
- if (mt->num_samples > 1)
- return;
+ case ISL_AUX_STATE_PASS_THROUGH:
+ if (written_with_hiz) {
+ intel_miptree_set_aux_state(brw, mt, level, layer, 1,
+ ISL_AUX_STATE_COMPRESSED_NO_CLEAR);
+ }
+ break;
- /* Fast color clear is supported for non-msaa arrays only on Gen8+. */
- assert(brw->gen >= 8 || (layer == 0 && mt->logical_depth0 == 1));
+ case ISL_AUX_STATE_AUX_INVALID:
+ assert(!written_with_hiz);
+ break;
+ }
+}
- (void)level;
- (void)layer;
+static inline uint32_t
+miptree_level_range_length(const struct intel_mipmap_tree *mt,
+ uint32_t start_level, uint32_t num_levels)
+{
+ assert(start_level >= mt->first_level);
+ assert(start_level <= mt->last_level);
+
+ if (num_levels == INTEL_REMAINING_LAYERS)
+ num_levels = mt->last_level - start_level + 1;
+ /* Check for overflow */
+ assert(start_level + num_levels >= start_level);
+ assert(start_level + num_levels <= mt->last_level + 1);
+
+ return num_levels;
}
-void
-intel_miptree_set_fast_clear_state(const struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- unsigned level,
- unsigned first_layer,
- unsigned num_layers,
- enum intel_fast_clear_state new_state)
+static inline uint32_t
+miptree_layer_range_length(const struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t num_layers)
{
- /* Setting the state to resolved means removing the item from the list
- * altogether.
- */
- assert(new_state != INTEL_FAST_CLEAR_STATE_RESOLVED);
+ assert(level <= mt->last_level);
+ uint32_t total_num_layers;
- intel_miptree_check_color_resolve(brw, mt, level, first_layer);
+ if (mt->surf.size > 0)
+ total_num_layers = mt->surf.dim == ISL_SURF_DIM_3D ?
+ minify(mt->surf.phys_level0_sa.depth, level) :
+ mt->surf.phys_level0_sa.array_len;
+ else
+ total_num_layers = mt->level[level].depth;
- assert(first_layer + num_layers <= mt->physical_depth0);
+ assert(start_layer < total_num_layers);
+ if (num_layers == INTEL_REMAINING_LAYERS)
+ num_layers = total_num_layers - start_layer;
+ /* Check for overflow */
+ assert(start_layer + num_layers >= start_layer);
+ assert(start_layer + num_layers <= total_num_layers);
- for (unsigned i = 0; i < num_layers; i++)
- intel_resolve_map_set(&mt->color_resolve_map, level,
- first_layer + i, new_state);
+ return num_layers;
}
-bool
-intel_miptree_has_color_unresolved(const struct intel_mipmap_tree *mt,
- unsigned start_level, unsigned num_levels,
- unsigned start_layer, unsigned num_layers)
+void
+intel_miptree_prepare_access(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ uint32_t start_level, uint32_t num_levels,
+ uint32_t start_layer, uint32_t num_layers,
+ bool aux_supported, bool fast_clear_supported)
{
- return intel_resolve_map_find_any(&mt->color_resolve_map,
- start_level, num_levels,
- start_layer, num_layers) != NULL;
+ num_levels = miptree_level_range_length(mt, start_level, num_levels);
+
+ if (_mesa_is_format_color_format(mt->format)) {
+ if (!mt->mcs_buf)
+ return;
+
+ if (mt->num_samples > 1) {
+ /* Nothing to do for MSAA */
+ assert(aux_supported && fast_clear_supported);
+ } else {
+ for (uint32_t l = 0; l < num_levels; l++) {
+ const uint32_t level = start_level + l;
+ const uint32_t level_layers =
+ miptree_layer_range_length(mt, level, start_layer, num_layers);
+ for (uint32_t a = 0; a < level_layers; a++) {
+ intel_miptree_prepare_ccs_access(brw, mt, level,
+ start_layer + a, aux_supported,
+ fast_clear_supported);
+ }
+ }
+ }
+ } else if (mt->format == MESA_FORMAT_S_UINT8) {
+ /* Nothing to do for stencil */
+ } else {
+ if (!mt->hiz_buf)
+ return;
+
+ for (uint32_t l = 0; l < num_levels; l++) {
+ const uint32_t level = start_level + l;
+ if (!intel_miptree_level_has_hiz(mt, level))
+ continue;
+
+ const uint32_t level_layers =
+ miptree_layer_range_length(mt, level, start_layer, num_layers);
+ for (uint32_t a = 0; a < level_layers; a++) {
+ intel_miptree_prepare_hiz_access(brw, mt, level, start_layer + a,
+ aux_supported,
+ fast_clear_supported);
+ }
+ }
+ }
}
void
-intel_miptree_used_for_rendering(const struct brw_context *brw,
- struct intel_mipmap_tree *mt, unsigned level,
- unsigned start_layer, unsigned num_layers)
+intel_miptree_finish_write(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t num_layers,
+ bool written_with_aux)
{
- const bool is_lossless_compressed =
- intel_miptree_is_lossless_compressed(brw, mt);
+ num_layers = miptree_layer_range_length(mt, level, start_layer, num_layers);
- for (unsigned i = 0; i < num_layers; ++i) {
- const enum intel_fast_clear_state fast_clear_state =
- intel_miptree_get_fast_clear_state(mt, level, start_layer + i);
+ if (_mesa_is_format_color_format(mt->format)) {
+ if (!mt->mcs_buf)
+ return;
- /* If the buffer was previously in fast clear state, change it to
- * unresolved state, since it won't be guaranteed to be clear after
- * rendering occurs.
- */
- if (is_lossless_compressed ||
- fast_clear_state == INTEL_FAST_CLEAR_STATE_CLEAR) {
- intel_miptree_set_fast_clear_state(
- brw, mt, level, start_layer + i, 1,
- INTEL_FAST_CLEAR_STATE_UNRESOLVED);
+ if (mt->num_samples > 1) {
+ for (uint32_t a = 0; a < num_layers; a++) {
+ intel_miptree_finish_mcs_write(brw, mt, level, start_layer + a,
+ written_with_aux);
+ }
+ } else {
+ for (uint32_t a = 0; a < num_layers; a++) {
+ intel_miptree_finish_ccs_write(brw, mt, level, start_layer + a,
+ written_with_aux);
+ }
+ }
+ } else if (mt->format == MESA_FORMAT_S_UINT8) {
+ /* Nothing to do for stencil */
+ } else {
+ if (!intel_miptree_level_has_hiz(mt, level))
+ return;
+
+ for (uint32_t a = 0; a < num_layers; a++) {
+ intel_miptree_finish_hiz_write(brw, mt, level, start_layer + a,
+ written_with_aux);
}
}
}
-static bool
-intel_miptree_needs_color_resolve(const struct brw_context *brw,
- const struct intel_mipmap_tree *mt,
- int flags)
+enum isl_aux_state
+intel_miptree_get_aux_state(const struct intel_mipmap_tree *mt,
+ uint32_t level, uint32_t layer)
{
- if (mt->aux_disable & INTEL_AUX_DISABLE_CCS)
- return false;
+ intel_miptree_check_level_layer(mt, level, layer);
- const bool is_lossless_compressed =
- intel_miptree_is_lossless_compressed(brw, mt);
+ if (_mesa_is_format_color_format(mt->format)) {
+ assert(mt->mcs_buf != NULL);
+ assert(mt->num_samples <= 1 || mt->msaa_layout == INTEL_MSAA_LAYOUT_CMS);
+ } else if (mt->format == MESA_FORMAT_S_UINT8) {
+ unreachable("Cannot get aux state for stencil");
+ } else {
+ assert(intel_miptree_level_has_hiz(mt, level));
+ }
- /* From gen9 onwards there is new compression scheme for single sampled
- * surfaces called "lossless compressed". These don't need to be always
- * resolved.
- */
- if ((flags & INTEL_MIPTREE_IGNORE_CCS_E) && is_lossless_compressed)
+ return mt->aux_state[level][layer];
+}
+
+void
+intel_miptree_set_aux_state(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t num_layers,
+ enum isl_aux_state aux_state)
+{
+ num_layers = miptree_layer_range_length(mt, level, start_layer, num_layers);
+
+ if (_mesa_is_format_color_format(mt->format)) {
+ assert(mt->mcs_buf != NULL);
+ assert(mt->num_samples <= 1 || mt->msaa_layout == INTEL_MSAA_LAYOUT_CMS);
+ } else if (mt->format == MESA_FORMAT_S_UINT8) {
+ unreachable("Cannot get aux state for stencil");
+ } else {
+ assert(intel_miptree_level_has_hiz(mt, level));
+ }
+
+ for (unsigned a = 0; a < num_layers; a++)
+ mt->aux_state[level][start_layer + a] = aux_state;
+}
+
+/* On Gen9 color buffers may be compressed by the hardware (lossless
+ * compression). There are, however, format restrictions and care needs to be
+ * taken that the sampler engine is capable for re-interpreting a buffer with
+ * format different the buffer was originally written with.
+ *
+ * For example, SRGB formats are not compressible and the sampler engine isn't
+ * capable of treating RGBA_UNORM as SRGB_ALPHA. In such a case the underlying
+ * color buffer needs to be resolved so that the sampling surface can be
+ * sampled as non-compressed (i.e., without the auxiliary MCS buffer being
+ * set).
+ */
+static bool
+can_texture_with_ccs(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ mesa_format view_format)
+{
+ if (!intel_miptree_is_lossless_compressed(brw, mt))
return false;
- /* Fast color clear resolves only make sense for non-MSAA buffers. */
- if (mt->msaa_layout != INTEL_MSAA_LAYOUT_NONE && !is_lossless_compressed)
+ enum isl_format isl_mt_format = brw_isl_format_for_mesa_format(mt->format);
+ enum isl_format isl_view_format = brw_isl_format_for_mesa_format(view_format);
+
+ if (!isl_formats_are_ccs_e_compatible(&brw->screen->devinfo,
+ isl_mt_format, isl_view_format)) {
+ perf_debug("Incompatible sampling format (%s) for rbc (%s)\n",
+ _mesa_get_format_name(view_format),
+ _mesa_get_format_name(mt->format));
return false;
+ }
return true;
}
-bool
-intel_miptree_resolve_color(struct brw_context *brw,
- struct intel_mipmap_tree *mt, unsigned level,
- unsigned start_layer, unsigned num_layers,
- int flags)
+static void
+intel_miptree_prepare_texture_slices(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ mesa_format view_format,
+ uint32_t start_level, uint32_t num_levels,
+ uint32_t start_layer, uint32_t num_layers,
+ bool *aux_supported_out)
+{
+ bool aux_supported, clear_supported;
+ if (_mesa_is_format_color_format(mt->format)) {
+ if (mt->num_samples > 1) {
+ aux_supported = clear_supported = true;
+ } else {
+ aux_supported = can_texture_with_ccs(brw, mt, view_format);
+
+ /* Clear color is specified as ints or floats and the conversion is
+ * done by the sampler. If we have a texture view, we would have to
+ * perform the clear color conversion manually. Just disable clear
+ * color.
+ */
+ clear_supported = aux_supported && (mt->format == view_format);
+ }
+ } else if (mt->format == MESA_FORMAT_S_UINT8) {
+ aux_supported = clear_supported = false;
+ } else {
+ aux_supported = clear_supported = intel_miptree_sample_with_hiz(brw, mt);
+ }
+
+ intel_miptree_prepare_access(brw, mt, start_level, num_levels,
+ start_layer, num_layers,
+ aux_supported, clear_supported);
+ if (aux_supported_out)
+ *aux_supported_out = aux_supported;
+}
+
+void
+intel_miptree_prepare_texture(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ mesa_format view_format,
+ bool *aux_supported_out)
{
- intel_miptree_check_color_resolve(brw, mt, level, start_layer);
+ intel_miptree_prepare_texture_slices(brw, mt, view_format,
+ 0, INTEL_REMAINING_LEVELS,
+ 0, INTEL_REMAINING_LAYERS,
+ aux_supported_out);
+}
- if (!intel_miptree_needs_color_resolve(brw, mt, flags))
- return false;
+void
+intel_miptree_prepare_image(struct brw_context *brw,
+ struct intel_mipmap_tree *mt)
+{
+ /* The data port doesn't understand any compression */
+ intel_miptree_prepare_access(brw, mt, 0, INTEL_REMAINING_LEVELS,
+ 0, INTEL_REMAINING_LAYERS, false, false);
+}
- /* Arrayed fast clear is only supported for gen8+. */
- assert(brw->gen >= 8 || num_layers == 1);
+void
+intel_miptree_prepare_fb_fetch(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t num_layers)
+{
+ intel_miptree_prepare_texture_slices(brw, mt, mt->format, level, 1,
+ start_layer, num_layers, NULL);
+}
- bool resolved = false;
- for (unsigned i = 0; i < num_layers; ++i) {
- intel_miptree_check_level_layer(mt, level, start_layer + i);
+void
+intel_miptree_prepare_render(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t layer_count,
+ bool srgb_enabled)
+{
+ /* If FRAMEBUFFER_SRGB is used on Gen9+ then we need to resolve any of
+ * the single-sampled color renderbuffers because the CCS buffer isn't
+ * supported for SRGB formats. This only matters if FRAMEBUFFER_SRGB is
+ * enabled because otherwise the surface state will be programmed with
+ * the linear equivalent format anyway.
+ */
+ if (brw->gen == 9 && srgb_enabled && mt->num_samples <= 1 &&
+ _mesa_get_srgb_format_linear(mt->format) != mt->format) {
- struct intel_resolve_map *item =
- intel_resolve_map_get(&mt->color_resolve_map, level,
- start_layer + i);
+ /* Lossless compression is not supported for SRGB formats, it
+ * should be impossible to get here with such surfaces.
+ */
+ assert(!intel_miptree_is_lossless_compressed(brw, mt));
+ intel_miptree_prepare_access(brw, mt, level, 1, start_layer, layer_count,
+ false, false);
+ }
- if (item) {
- assert(item->fast_clear_state != INTEL_FAST_CLEAR_STATE_RESOLVED);
+ /* For layered rendering non-compressed fast cleared buffers need to be
+ * resolved. Surface state can carry only one fast color clear value
+ * while each layer may have its own fast clear color value. For
+ * compressed buffers color value is available in the color buffer.
+ */
+ if (layer_count > 1 &&
+ !(mt->aux_disable & INTEL_AUX_DISABLE_CCS) &&
+ !intel_miptree_is_lossless_compressed(brw, mt)) {
+ assert(brw->gen >= 8);
- brw_blorp_resolve_color(brw, mt, level, start_layer);
- intel_resolve_map_remove(item);
- resolved = true;
- }
+ intel_miptree_prepare_access(brw, mt, level, 1, start_layer, layer_count,
+ false, false);
}
+}
- return resolved;
+void
+intel_miptree_finish_render(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t layer_count)
+{
+ assert(_mesa_is_format_color_format(mt->format));
+ intel_miptree_finish_write(brw, mt, level, start_layer, layer_count,
+ mt->mcs_buf != NULL);
}
void
-intel_miptree_all_slices_resolve_color(struct brw_context *brw,
- struct intel_mipmap_tree *mt,
- int flags)
+intel_miptree_prepare_depth(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t layer_count)
{
- if (!intel_miptree_needs_color_resolve(brw, mt, flags))
- return;
-
- foreach_list_typed_safe(struct intel_resolve_map, map, link,
- &mt->color_resolve_map) {
- assert(map->fast_clear_state != INTEL_FAST_CLEAR_STATE_RESOLVED);
+ intel_miptree_prepare_access(brw, mt, level, 1, start_layer, layer_count,
+ mt->hiz_buf != NULL, mt->hiz_buf != NULL);
+}
- brw_blorp_resolve_color(brw, mt, map->level, map->layer);
- intel_resolve_map_remove(map);
+void
+intel_miptree_finish_depth(struct brw_context *brw,
+ struct intel_mipmap_tree *mt, uint32_t level,
+ uint32_t start_layer, uint32_t layer_count,
+ bool depth_written)
+{
+ if (depth_written) {
+ intel_miptree_finish_write(brw, mt, level, start_layer, layer_count,
+ mt->hiz_buf != NULL);
}
}
*/
assert(mt->msaa_layout == INTEL_MSAA_LAYOUT_NONE || mt->num_samples <= 1);
+ intel_miptree_prepare_access(brw, mt, 0, INTEL_REMAINING_LEVELS,
+ 0, INTEL_REMAINING_LAYERS, false, false);
+
if (mt->mcs_buf) {
- intel_miptree_all_slices_resolve_color(brw, mt, 0);
mt->aux_disable |= (INTEL_AUX_DISABLE_CCS | INTEL_AUX_DISABLE_MCS);
- drm_bacon_bo_unreference(mt->mcs_buf->bo);
+ brw_bo_unreference(mt->mcs_buf->bo);
free(mt->mcs_buf);
mt->mcs_buf = NULL;
* execute any will likely crash due to the missing aux buffer. So let's
* delete all pending ops.
*/
- exec_list_make_empty(&mt->color_resolve_map);
+ free(mt->aux_state);
+ mt->aux_state = NULL;
}
if (mt->hiz_buf) {
mt->aux_disable |= INTEL_AUX_DISABLE_HIZ;
- intel_miptree_all_slices_resolve_depth(brw, mt);
intel_miptree_hiz_buffer_free(mt->hiz_buf);
mt->hiz_buf = NULL;
* any will likely crash due to the missing aux buffer. So let's delete
* all pending ops.
*/
- exec_list_make_empty(&mt->hiz_map);
+ free(mt->aux_state);
+ mt->aux_state = NULL;
}
}
struct intel_mipmap_tree *src,
struct intel_mipmap_tree *dst)
{
+ unsigned src_w, src_h, dst_w, dst_h;
+
+ if (src->surf.size > 0) {
+ src_w = src->surf.logical_level0_px.width;
+ src_h = src->surf.logical_level0_px.height;
+ } else {
+ src_w = src->logical_width0;
+ src_h = src->logical_height0;
+ }
+
+ if (dst->surf.size > 0) {
+ dst_w = dst->surf.logical_level0_px.width;
+ dst_h = dst->surf.logical_level0_px.height;
+ } else {
+ dst_w = dst->logical_width0;
+ dst_h = dst->logical_height0;
+ }
+
brw_blorp_blit_miptrees(brw,
src, 0 /* level */, 0 /* layer */,
src->format, SWIZZLE_XYZW,
dst, 0 /* level */, 0 /* layer */, dst->format,
- 0, 0,
- src->logical_width0, src->logical_height0,
- 0, 0,
- dst->logical_width0, dst->logical_height0,
+ 0, 0, src_w, src_h,
+ 0, 0, dst_w, dst_h,
GL_NEAREST, false, false /*mirror x, y*/,
false, false);
if (src->stencil_mt) {
+ if (src->stencil_mt->surf.size > 0) {
+ src_w = src->stencil_mt->surf.logical_level0_px.width;
+ src_h = src->stencil_mt->surf.logical_level0_px.height;
+ } else {
+ src_w = src->stencil_mt->logical_width0;
+ src_h = src->stencil_mt->logical_height0;
+ }
+
+ if (dst->stencil_mt->surf.size > 0) {
+ dst_w = dst->stencil_mt->surf.logical_level0_px.width;
+ dst_h = dst->stencil_mt->surf.logical_level0_px.height;
+ } else {
+ dst_w = dst->stencil_mt->logical_width0;
+ dst_h = dst->stencil_mt->logical_height0;
+ }
+
brw_blorp_blit_miptrees(brw,
src->stencil_mt, 0 /* level */, 0 /* layer */,
src->stencil_mt->format, SWIZZLE_XYZW,
dst->stencil_mt, 0 /* level */, 0 /* layer */,
dst->stencil_mt->format,
- 0, 0,
- src->logical_width0, src->logical_height0,
- 0, 0,
- dst->logical_width0, dst->logical_height0,
+ 0, 0, src_w, src_h,
+ 0, 0, dst_w, dst_h,
GL_NEAREST, false, false /*mirror x, y*/,
false, false /* decode/encode srgb */);
}
const uint32_t r8stencil_flags =
MIPTREE_LAYOUT_ACCELERATED_UPLOAD | MIPTREE_LAYOUT_TILING_Y |
MIPTREE_LAYOUT_DISABLE_AUX;
- assert(brw->gen > 6); /* Handle MIPTREE_LAYOUT_FORCE_ALL_SLICE_AT_LOD */
+ assert(brw->gen > 6); /* Handle MIPTREE_LAYOUT_GEN6_HIZ_STENCIL */
mt->r8stencil_mt = intel_miptree_create(brw,
src->target,
MESA_FORMAT_R_UINT8,
for (int level = src->first_level; level <= src->last_level; level++) {
const unsigned depth = src->level[level].depth;
- const int layers_per_blit =
- (dst->msaa_layout == INTEL_MSAA_LAYOUT_UMS ||
- dst->msaa_layout == INTEL_MSAA_LAYOUT_CMS) ?
- dst->num_samples : 1;
for (unsigned layer = 0; layer < depth; layer++) {
- brw_blorp_blit_miptrees(brw,
+ brw_blorp_copy_miptrees(brw,
src, level, layer,
- src->format, SWIZZLE_X,
- dst, level, layers_per_blit * layer,
- MESA_FORMAT_R_UNORM8,
- 0, 0,
+ dst, level, layer,
+ 0, 0, 0, 0,
minify(src->logical_width0, level),
- minify(src->logical_height0, level),
- 0, 0,
- minify(dst->logical_width0, level),
- minify(dst->logical_height0, level),
- GL_NEAREST, false, false /*mirror x, y*/,
- false, false /* decode/encode srgb */);
+ minify(src->logical_height0, level));
}
}
}
static void *
-intel_miptree_map_raw(struct brw_context *brw, struct intel_mipmap_tree *mt)
+intel_miptree_map_raw(struct brw_context *brw,
+ struct intel_mipmap_tree *mt,
+ GLbitfield mode)
{
- /* CPU accesses to color buffers don't understand fast color clears, so
- * resolve any pending fast color clears before we map.
- */
- intel_miptree_all_slices_resolve_color(brw, mt, 0);
-
- drm_bacon_bo *bo = mt->bo;
+ struct brw_bo *bo = mt->bo;
if (brw_batch_references(&brw->batch, bo))
intel_batchbuffer_flush(brw);
- /* brw_bo_map() uses a WB mmaping of the buffer's backing storage. It
- * will utilize the CPU cache even if the buffer is incoherent with the
- * GPU (i.e. any writes will be stored in the cache and not flushed to
- * memory and so will be invisible to the GPU or display engine). This
- * is the majority of buffers on a !llc machine, but even on a llc
- * almost all scanouts are incoherent with the CPU. A WB write into the
- * backing storage of the current scanout will not be immediately
- * visible on the screen. The transfer from cache to screen is slow and
- * indeterministic causing visible glitching on the screen. Never use
- * this WB mapping for writes to an active scanout (reads are fine, so
- * long as cache consistency is maintained).
- */
- if (mt->tiling != I915_TILING_NONE || mt->is_scanout)
- brw_bo_map_gtt(brw, bo, "miptree");
- else
- brw_bo_map(brw, bo, true, "miptree");
-
- return bo->virtual;
+ return brw_bo_map(brw, bo, mode);
}
static void
intel_miptree_unmap_raw(struct intel_mipmap_tree *mt)
{
- drm_bacon_bo_unmap(mt->bo);
+ brw_bo_unmap(mt->bo);
}
static void
y /= bh;
x /= bw;
- base = intel_miptree_map_raw(brw, mt) + mt->offset;
+ base = intel_miptree_map_raw(brw, mt, map->mode) + mt->offset;
if (base == NULL)
map->ptr = NULL;
}
}
- map->ptr = intel_miptree_map_raw(brw, map->linear_mt);
+ map->ptr = intel_miptree_map_raw(brw, map->linear_mt, map->mode);
DBG("%s: %d,%d %dx%d from mt %p (%s) %d,%d = %p/%d\n", __func__,
map->x, map->y, map->w, map->h,
image_x += map->x;
image_y += map->y;
- void *src = intel_miptree_map_raw(brw, mt);
+ void *src = intel_miptree_map_raw(brw, mt, map->mode);
if (!src)
return;
* temporary buffer back out.
*/
if (!(map->mode & GL_MAP_INVALIDATE_RANGE_BIT)) {
+ /* ISL uses a stencil pitch value that is expected by hardware whereas
+ * traditional miptree uses half of that. Below the value gets supplied
+ * to intel_offset_S8() which expects the legacy interpretation.
+ */
+ const unsigned pitch = mt->surf.size > 0 ?
+ mt->surf.row_pitch / 2 : mt->pitch;
uint8_t *untiled_s8_map = map->ptr;
- uint8_t *tiled_s8_map = intel_miptree_map_raw(brw, mt);
+ uint8_t *tiled_s8_map = intel_miptree_map_raw(brw, mt, GL_MAP_READ_BIT);
unsigned int image_x, image_y;
intel_miptree_get_image_offset(mt, level, slice, &image_x, &image_y);
for (uint32_t y = 0; y < map->h; y++) {
for (uint32_t x = 0; x < map->w; x++) {
- ptrdiff_t offset = intel_offset_S8(mt->pitch,
+ ptrdiff_t offset = intel_offset_S8(pitch,
x + image_x + map->x,
y + image_y + map->y,
brw->has_swizzling);
unsigned int slice)
{
if (map->mode & GL_MAP_WRITE_BIT) {
+ /* ISL uses a stencil pitch value that is expected by hardware whereas
+ * traditional miptree uses half of that. Below the value gets supplied
+ * to intel_offset_S8() which expects the legacy interpretation.
+ */
+ const unsigned pitch = mt->surf.size > 0 ?
+ mt->surf.row_pitch / 2: mt->pitch;
unsigned int image_x, image_y;
uint8_t *untiled_s8_map = map->ptr;
- uint8_t *tiled_s8_map = intel_miptree_map_raw(brw, mt);
+ uint8_t *tiled_s8_map = intel_miptree_map_raw(brw, mt, GL_MAP_WRITE_BIT);
intel_miptree_get_image_offset(mt, level, slice, &image_x, &image_y);
for (uint32_t y = 0; y < map->h; y++) {
for (uint32_t x = 0; x < map->w; x++) {
- ptrdiff_t offset = intel_offset_S8(mt->pitch,
+ ptrdiff_t offset = intel_offset_S8(pitch,
image_x + x + map->x,
image_y + y + map->y,
brw->has_swizzling);
image_x += map->x;
image_y += map->y;
- uint8_t *dst = intel_miptree_map_raw(brw, mt)
+ uint8_t *dst = intel_miptree_map_raw(brw, mt, GL_MAP_WRITE_BIT)
+ image_y * mt->pitch
+ image_x * mt->cpp;
* temporary buffer back out.
*/
if (!(map->mode & GL_MAP_INVALIDATE_RANGE_BIT)) {
+ /* ISL uses a stencil pitch value that is expected by hardware whereas
+ * traditional miptree uses half of that. Below the value gets supplied
+ * to intel_offset_S8() which expects the legacy interpretation.
+ */
+ const unsigned s_pitch = s_mt->surf.size > 0 ?
+ s_mt->surf.row_pitch / 2 : s_mt->pitch;
uint32_t *packed_map = map->ptr;
- uint8_t *s_map = intel_miptree_map_raw(brw, s_mt);
- uint32_t *z_map = intel_miptree_map_raw(brw, z_mt);
+ uint8_t *s_map = intel_miptree_map_raw(brw, s_mt, GL_MAP_READ_BIT);
+ uint32_t *z_map = intel_miptree_map_raw(brw, z_mt, GL_MAP_READ_BIT);
unsigned int s_image_x, s_image_y;
unsigned int z_image_x, z_image_y;
for (uint32_t y = 0; y < map->h; y++) {
for (uint32_t x = 0; x < map->w; x++) {
int map_x = map->x + x, map_y = map->y + y;
- ptrdiff_t s_offset = intel_offset_S8(s_mt->pitch,
+ ptrdiff_t s_offset = intel_offset_S8(s_pitch,
map_x + s_image_x,
map_y + s_image_y,
brw->has_swizzling);
bool map_z32f_x24s8 = mt->format == MESA_FORMAT_Z_FLOAT32;
if (map->mode & GL_MAP_WRITE_BIT) {
+ /* ISL uses a stencil pitch value that is expected by hardware whereas
+ * traditional miptree uses half of that. Below the value gets supplied
+ * to intel_offset_S8() which expects the legacy interpretation.
+ */
+ const unsigned s_pitch = s_mt->surf.size > 0 ?
+ s_mt->surf.row_pitch / 2 : s_mt->pitch;
uint32_t *packed_map = map->ptr;
- uint8_t *s_map = intel_miptree_map_raw(brw, s_mt);
- uint32_t *z_map = intel_miptree_map_raw(brw, z_mt);
+ uint8_t *s_map = intel_miptree_map_raw(brw, s_mt, GL_MAP_WRITE_BIT);
+ uint32_t *z_map = intel_miptree_map_raw(brw, z_mt, GL_MAP_WRITE_BIT);
unsigned int s_image_x, s_image_y;
unsigned int z_image_x, z_image_y;
for (uint32_t y = 0; y < map->h; y++) {
for (uint32_t x = 0; x < map->w; x++) {
- ptrdiff_t s_offset = intel_offset_S8(s_mt->pitch,
+ ptrdiff_t s_offset = intel_offset_S8(s_pitch,
x + s_image_x + map->x,
y + s_image_y + map->y,
brw->has_swizzling);
return;
}
- intel_miptree_slice_resolve_depth(brw, mt, level, slice);
- if (map->mode & GL_MAP_WRITE_BIT) {
- intel_miptree_slice_set_needs_hiz_resolve(mt, level, slice);
- }
+ intel_miptree_access_raw(brw, mt, level, slice,
+ map->mode & GL_MAP_WRITE_BIT);
if (mt->format == MESA_FORMAT_S_UINT8) {
intel_miptree_map_s8(brw, mt, map, level, slice);
enum isl_dim_layout
get_isl_dim_layout(const struct gen_device_info *devinfo, uint32_t tiling,
- GLenum target)
+ GLenum target, enum miptree_array_layout array_layout)
{
+ if (array_layout == GEN6_HIZ_STENCIL)
+ return ISL_DIM_LAYOUT_GEN6_STENCIL_HIZ;
+
switch (target) {
case GL_TEXTURE_1D:
case GL_TEXTURE_1D_ARRAY:
{
surf->dim = get_isl_surf_dim(mt->target);
surf->dim_layout = get_isl_dim_layout(&brw->screen->devinfo,
- mt->tiling, mt->target);
+ mt->tiling, mt->target,
+ mt->array_layout);
if (mt->num_samples > 1) {
switch (mt->msaa_layout) {
surf->phys_level0_sa.array_len = mt->physical_depth0;
}
- surf->levels = mt->last_level + 1;
+ surf->levels = mt->last_level - mt->first_level + 1;
surf->samples = MAX2(mt->num_samples, 1);
surf->size = 0; /* TODO */
switch (surf->dim_layout) {
case ISL_DIM_LAYOUT_GEN4_2D:
case ISL_DIM_LAYOUT_GEN4_3D:
+ case ISL_DIM_LAYOUT_GEN6_STENCIL_HIZ:
if (brw->gen >= 9) {
surf->array_pitch_el_rows = mt->qpitch;
} else {
surf->array_pitch_span = ISL_ARRAY_PITCH_SPAN_FULL;
break;
case ALL_SLICES_AT_EACH_LOD:
+ case GEN6_HIZ_STENCIL:
surf->array_pitch_span = ISL_ARRAY_PITCH_SPAN_COMPACT;
break;
default:
surf->usage |= ISL_SURF_USAGE_CUBE_BIT;
}
+enum isl_aux_usage
+intel_miptree_get_aux_isl_usage(const struct brw_context *brw,
+ const struct intel_mipmap_tree *mt)
+{
+ if (mt->hiz_buf)
+ return ISL_AUX_USAGE_HIZ;
+
+ if (!mt->mcs_buf)
+ return ISL_AUX_USAGE_NONE;
+
+ if (mt->num_samples > 1) {
+ assert(mt->msaa_layout == INTEL_MSAA_LAYOUT_CMS);
+ return ISL_AUX_USAGE_MCS;
+ }
+
+ if (intel_miptree_is_lossless_compressed(brw, mt)) {
+ assert(brw->gen >= 9);
+ return ISL_AUX_USAGE_CCS_E;
+ }
+
+ if ((mt->aux_disable & INTEL_AUX_DISABLE_CCS) == 0)
+ return ISL_AUX_USAGE_CCS_D;
+
+ unreachable("Invalid MCS miptree");
+}
+
/* WARNING: THE SURFACE CREATED BY THIS FUNCTION IS NOT COMPLETE AND CANNOT BE
* USED FOR ANY REAL CALCULATIONS. THE ONLY VALID USE OF SUCH A SURFACE IS TO
* PASS IT INTO isl_surf_fill_state.
void
intel_miptree_get_aux_isl_surf(struct brw_context *brw,
const struct intel_mipmap_tree *mt,
- struct isl_surf *surf,
- enum isl_aux_usage *usage)
+ enum isl_aux_usage usage,
+ struct isl_surf *surf)
{
uint32_t aux_pitch, aux_qpitch;
if (mt->mcs_buf) {
aux_pitch = mt->mcs_buf->pitch;
aux_qpitch = mt->mcs_buf->qpitch;
-
- if (mt->num_samples > 1) {
- assert(mt->msaa_layout == INTEL_MSAA_LAYOUT_CMS);
- *usage = ISL_AUX_USAGE_MCS;
- } else if (intel_miptree_is_lossless_compressed(brw, mt)) {
- assert(brw->gen >= 9);
- *usage = ISL_AUX_USAGE_CCS_E;
- } else if ((mt->aux_disable & INTEL_AUX_DISABLE_CCS) == 0) {
- *usage = ISL_AUX_USAGE_CCS_D;
- } else {
- unreachable("Invalid MCS miptree");
- }
} else if (mt->hiz_buf) {
aux_pitch = mt->hiz_buf->aux_base.pitch;
aux_qpitch = mt->hiz_buf->aux_base.qpitch;
-
- *usage = ISL_AUX_USAGE_HIZ;
} else {
- *usage = ISL_AUX_USAGE_NONE;
return;
}
intel_miptree_get_isl_surf(brw, mt, surf);
/* Figure out the format and tiling of the auxiliary surface */
- switch (*usage) {
+ switch (usage) {
case ISL_AUX_USAGE_NONE:
unreachable("Invalid auxiliary usage");
surf->array_pitch_el_rows =
aux_qpitch / isl_format_get_layout(surf->format)->bh;
}
-
-union isl_color_value
-intel_miptree_get_isl_clear_color(struct brw_context *brw,
- const struct intel_mipmap_tree *mt)
-{
- union isl_color_value clear_color;
-
- if (_mesa_get_format_base_format(mt->format) == GL_DEPTH_COMPONENT) {
- clear_color.i32[0] = mt->depth_clear_value;
- clear_color.i32[1] = 0;
- clear_color.i32[2] = 0;
- clear_color.i32[3] = 0;
- } else if (brw->gen >= 9) {
- clear_color.i32[0] = mt->gen9_fast_clear_color.i[0];
- clear_color.i32[1] = mt->gen9_fast_clear_color.i[1];
- clear_color.i32[2] = mt->gen9_fast_clear_color.i[2];
- clear_color.i32[3] = mt->gen9_fast_clear_color.i[3];
- } else if (_mesa_is_format_integer(mt->format)) {
- clear_color.i32[0] = (mt->fast_clear_color_value & (1u << 31)) != 0;
- clear_color.i32[1] = (mt->fast_clear_color_value & (1u << 30)) != 0;
- clear_color.i32[2] = (mt->fast_clear_color_value & (1u << 29)) != 0;
- clear_color.i32[3] = (mt->fast_clear_color_value & (1u << 28)) != 0;
- } else {
- clear_color.f32[0] = (mt->fast_clear_color_value & (1u << 31)) != 0;
- clear_color.f32[1] = (mt->fast_clear_color_value & (1u << 30)) != 0;
- clear_color.f32[2] = (mt->fast_clear_color_value & (1u << 29)) != 0;
- clear_color.f32[3] = (mt->fast_clear_color_value & (1u << 28)) != 0;
- }
-
- return clear_color;
-}