*/
#include "r600_formats.h"
#include "r600_shader.h"
+#include "r600_query.h"
#include "evergreend.h"
#include "pipe/p_shader_tokens.h"
return 0;
}
-static unsigned r600_tex_dim(unsigned dim, unsigned nr_samples)
+static unsigned r600_tex_dim(struct r600_texture *rtex,
+ unsigned view_target, unsigned nr_samples)
{
- switch (dim) {
+ unsigned res_target = rtex->resource.b.b.target;
+
+ if (view_target == PIPE_TEXTURE_CUBE ||
+ view_target == PIPE_TEXTURE_CUBE_ARRAY)
+ res_target = view_target;
+ /* If interpreting cubemaps as something else, set 2D_ARRAY. */
+ else if (res_target == PIPE_TEXTURE_CUBE ||
+ res_target == PIPE_TEXTURE_CUBE_ARRAY)
+ res_target = PIPE_TEXTURE_2D_ARRAY;
+
+ switch (res_target) {
default:
case PIPE_TEXTURE_1D:
return V_030000_SQ_TEX_DIM_1D;
return r600_translate_dbformat(format) != ~0U;
}
-boolean evergreen_is_format_supported(struct pipe_screen *screen,
- enum pipe_format format,
- enum pipe_texture_target target,
- unsigned sample_count,
- unsigned usage)
+bool evergreen_is_format_supported(struct pipe_screen *screen,
+ enum pipe_format format,
+ enum pipe_texture_target target,
+ unsigned sample_count,
+ unsigned storage_sample_count,
+ unsigned usage)
{
struct r600_screen *rscreen = (struct r600_screen*)screen;
unsigned retval = 0;
if (target >= PIPE_MAX_TEXTURE_TYPES) {
R600_ERR("r600: unsupported texture type %d\n", target);
- return FALSE;
+ return false;
}
- if (!util_format_is_supported(format, usage))
- return FALSE;
+ if (MAX2(1, sample_count) != MAX2(1, storage_sample_count))
+ return false;
if (sample_count > 1) {
if (!rscreen->has_msaa)
- return FALSE;
+ return false;
switch (sample_count) {
case 2:
case 8:
break;
default:
- return FALSE;
+ return false;
}
}
retval |= PIPE_BIND_VERTEX_BUFFER;
}
- if (usage & PIPE_BIND_TRANSFER_READ)
- retval |= PIPE_BIND_TRANSFER_READ;
- if (usage & PIPE_BIND_TRANSFER_WRITE)
- retval |= PIPE_BIND_TRANSFER_WRITE;
-
if ((usage & PIPE_BIND_LINEAR) &&
!util_format_is_compressed(format) &&
!(usage & PIPE_BIND_DEPTH_STENCIL))
r600_init_command_buffer(&rs->buffer, 30);
rs->scissor_enable = state->scissor;
+ rs->clip_halfz = state->clip_halfz;
rs->flatshade = state->flatshade;
rs->sprite_coord_enable = state->sprite_coord_enable;
+ rs->rasterizer_discard = state->rasterizer_discard;
rs->two_side = state->light_twoside;
rs->clip_plane_enable = state->clip_plane_enable;
rs->pa_sc_line_stipple = state->line_stipple_enable ?
S_028A0C_LINE_PATTERN(state->line_stipple_pattern) |
S_028A0C_REPEAT_COUNT(state->line_stipple_factor) : 0;
rs->pa_cl_clip_cntl =
- S_028810_PS_UCP_MODE(3) |
S_028810_DX_CLIP_SPACE_DEF(state->clip_halfz) |
- S_028810_ZCLIP_NEAR_DISABLE(!state->depth_clip) |
- S_028810_ZCLIP_FAR_DISABLE(!state->depth_clip) |
+ S_028810_ZCLIP_NEAR_DISABLE(!state->depth_clip_near) |
+ S_028810_ZCLIP_FAR_DISABLE(!state->depth_clip_far) |
S_028810_DX_LINEAR_ATTR_CLIP_ENA(1) |
S_028810_DX_RASTERIZATION_KILL(state->rasterizer_discard);
rs->multisample_enable = state->multisample;
rs->offset_units = state->offset_units;
rs->offset_scale = state->offset_scale * 16.0f;
rs->offset_enable = state->offset_point || state->offset_line || state->offset_tri;
+ rs->offset_units_unscaled = state->offset_units_unscaled;
if (state->point_size_per_vertex) {
psize_min = util_get_min_point_size(state);
unsigned max_aniso = rscreen->force_aniso >= 0 ? rscreen->force_aniso
: state->max_anisotropy;
unsigned max_aniso_ratio = r600_tex_aniso_filter(max_aniso);
+ float max_lod = state->max_lod;
if (!ss) {
return NULL;
}
+ /* If the min_mip_filter is NONE, then the texture has no mipmapping and
+ * MIP_FILTER will also be set to NONE. However, if more then one LOD is
+ * configured, then the texture lookup seems to fail for some specific texture
+ * formats. Forcing the number of LODs to one in this case fixes it. */
+ if (state->min_mip_filter == PIPE_TEX_MIPFILTER_NONE)
+ max_lod = state->min_lod;
+
ss->border_color_use = sampler_state_needs_border_color(state);
/* R_03C000_SQ_TEX_SAMPLER_WORD0_0 */
/* R_03C004_SQ_TEX_SAMPLER_WORD1_0 */
ss->tex_sampler_words[1] =
S_03C004_MIN_LOD(S_FIXED(CLAMP(state->min_lod, 0, 15), 8)) |
- S_03C004_MAX_LOD(S_FIXED(CLAMP(state->max_lod, 0, 15), 8));
+ S_03C004_MAX_LOD(S_FIXED(CLAMP(max_lod, 0, 15), 8));
/* R_03C008_SQ_TEX_SAMPLER_WORD2_0 */
ss->tex_sampler_words[2] =
S_03C008_LOD_BIAS(S_FIXED(CLAMP(state->lod_bias, -16, 16), 8)) |
return ss;
}
-static struct pipe_sampler_view *
-texture_buffer_sampler_view(struct r600_context *rctx,
- struct r600_pipe_sampler_view *view,
- unsigned width0, unsigned height0)
-
+struct eg_buf_res_params {
+ enum pipe_format pipe_format;
+ unsigned offset;
+ unsigned size;
+ unsigned char swizzle[4];
+ bool uncached;
+ bool force_swizzle;
+ bool size_in_bytes;
+};
+
+static void evergreen_fill_buffer_resource_words(struct r600_context *rctx,
+ struct pipe_resource *buffer,
+ struct eg_buf_res_params *params,
+ bool *skip_mip_address_reloc,
+ unsigned tex_resource_words[8])
{
- struct r600_texture *tmp = (struct r600_texture*)view->base.texture;
+ struct r600_texture *tmp = (struct r600_texture*)buffer;
uint64_t va;
- int stride = util_format_get_blocksize(view->base.format);
+ int stride = util_format_get_blocksize(params->pipe_format);
unsigned format, num_format, format_comp, endian;
unsigned swizzle_res;
- unsigned char swizzle[4];
const struct util_format_description *desc;
- unsigned offset = view->base.u.buf.first_element * stride;
- unsigned size = (view->base.u.buf.last_element - view->base.u.buf.first_element + 1) * stride;
-
- swizzle[0] = view->base.swizzle_r;
- swizzle[1] = view->base.swizzle_g;
- swizzle[2] = view->base.swizzle_b;
- swizzle[3] = view->base.swizzle_a;
- r600_vertex_data_type(view->base.format,
+ r600_vertex_data_type(params->pipe_format,
&format, &num_format, &format_comp,
&endian);
- desc = util_format_description(view->base.format);
+ desc = util_format_description(params->pipe_format);
- swizzle_res = r600_get_swizzle_combined(desc->swizzle, swizzle, TRUE);
-
- va = tmp->resource.gpu_address + offset;
- view->tex_resource = &tmp->resource;
+ if (params->force_swizzle)
+ swizzle_res = r600_get_swizzle_combined(params->swizzle, NULL, TRUE);
+ else
+ swizzle_res = r600_get_swizzle_combined(desc->swizzle, params->swizzle, TRUE);
- view->skip_mip_address_reloc = true;
- view->tex_resource_words[0] = va;
- view->tex_resource_words[1] = size - 1;
- view->tex_resource_words[2] = S_030008_BASE_ADDRESS_HI(va >> 32UL) |
+ va = tmp->resource.gpu_address + params->offset;
+ *skip_mip_address_reloc = true;
+ tex_resource_words[0] = va;
+ tex_resource_words[1] = params->size - 1;
+ tex_resource_words[2] = S_030008_BASE_ADDRESS_HI(va >> 32UL) |
S_030008_STRIDE(stride) |
S_030008_DATA_FORMAT(format) |
S_030008_NUM_FORMAT_ALL(num_format) |
S_030008_FORMAT_COMP_ALL(format_comp) |
S_030008_ENDIAN_SWAP(endian);
- view->tex_resource_words[3] = swizzle_res;
+ tex_resource_words[3] = swizzle_res | S_03000C_UNCACHED(params->uncached);
/*
- * in theory dword 4 is for number of elements, for use with resinfo,
- * but it seems to utterly fail to work, the amd gpu shader analyser
+ * dword 4 is for number of elements, for use with resinfo,
+ * albeit the amd gpu shader analyser
* uses a const buffer to store the element sizes for buffer txq
*/
- view->tex_resource_words[4] = 0;
- view->tex_resource_words[5] = view->tex_resource_words[6] = 0;
- view->tex_resource_words[7] = S_03001C_TYPE(V_03001C_SQ_TEX_VTX_VALID_BUFFER);
+ tex_resource_words[4] = params->size_in_bytes ? params->size : (params->size / stride);
+
+ tex_resource_words[5] = tex_resource_words[6] = 0;
+ tex_resource_words[7] = S_03001C_TYPE(V_03001C_SQ_TEX_VTX_VALID_BUFFER);
+}
+
+static struct pipe_sampler_view *
+texture_buffer_sampler_view(struct r600_context *rctx,
+ struct r600_pipe_sampler_view *view,
+ unsigned width0, unsigned height0)
+{
+ struct r600_texture *tmp = (struct r600_texture*)view->base.texture;
+ struct eg_buf_res_params params;
+
+ memset(¶ms, 0, sizeof(params));
+
+ params.pipe_format = view->base.format;
+ params.offset = view->base.u.buf.offset;
+ params.size = view->base.u.buf.size;
+ params.swizzle[0] = view->base.swizzle_r;
+ params.swizzle[1] = view->base.swizzle_g;
+ params.swizzle[2] = view->base.swizzle_b;
+ params.swizzle[3] = view->base.swizzle_a;
+
+ evergreen_fill_buffer_resource_words(rctx, view->base.texture,
+ ¶ms, &view->skip_mip_address_reloc,
+ view->tex_resource_words);
+ view->tex_resource = &tmp->resource;
if (tmp->resource.gpu_address)
- LIST_ADDTAIL(&view->list, &rctx->b.texture_buffers);
+ list_addtail(&view->list, &rctx->texture_buffers);
return &view->base;
}
-struct pipe_sampler_view *
-evergreen_create_sampler_view_custom(struct pipe_context *ctx,
- struct pipe_resource *texture,
- const struct pipe_sampler_view *state,
- unsigned width0, unsigned height0,
- unsigned force_level)
+struct eg_tex_res_params {
+ enum pipe_format pipe_format;
+ int force_level;
+ unsigned width0;
+ unsigned height0;
+ unsigned first_level;
+ unsigned last_level;
+ unsigned first_layer;
+ unsigned last_layer;
+ unsigned target;
+ unsigned char swizzle[4];
+};
+
+static int evergreen_fill_tex_resource_words(struct r600_context *rctx,
+ struct pipe_resource *texture,
+ struct eg_tex_res_params *params,
+ bool *skip_mip_address_reloc,
+ unsigned tex_resource_words[8])
{
- struct r600_context *rctx = (struct r600_context*)ctx;
- struct r600_screen *rscreen = (struct r600_screen*)ctx->screen;
- struct r600_pipe_sampler_view *view = CALLOC_STRUCT(r600_pipe_sampler_view);
+ struct r600_screen *rscreen = (struct r600_screen*)rctx->b.b.screen;
struct r600_texture *tmp = (struct r600_texture*)texture;
unsigned format, endian;
uint32_t word4 = 0, yuv_format = 0, pitch = 0;
- unsigned char swizzle[4], array_mode = 0, non_disp_tiling = 0;
+ unsigned char array_mode = 0, non_disp_tiling = 0;
unsigned height, depth, width;
unsigned macro_aspect, tile_split, bankh, bankw, nbanks, fmask_bankh;
- enum pipe_format pipe_format = state->format;
- struct radeon_surf_level *surflevel;
+ struct legacy_surf_level *surflevel;
unsigned base_level, first_level, last_level;
unsigned dim, last_layer;
uint64_t va;
bool do_endian_swap = FALSE;
- if (!view)
- return NULL;
-
- /* initialize base object */
- view->base = *state;
- view->base.texture = NULL;
- pipe_reference(NULL, &texture->reference);
- view->base.texture = texture;
- view->base.reference.count = 1;
- view->base.context = ctx;
-
- if (state->target == PIPE_BUFFER)
- return texture_buffer_sampler_view(rctx, view, width0, height0);
-
- swizzle[0] = state->swizzle_r;
- swizzle[1] = state->swizzle_g;
- swizzle[2] = state->swizzle_b;
- swizzle[3] = state->swizzle_a;
-
- tile_split = tmp->surface.tile_split;
- surflevel = tmp->surface.level;
+ tile_split = tmp->surface.u.legacy.tile_split;
+ surflevel = tmp->surface.u.legacy.level;
/* Texturing with separate depth and stencil. */
- if (tmp->is_depth && !tmp->is_flushing_texture) {
- switch (pipe_format) {
+ if (tmp->db_compatible) {
+ switch (params->pipe_format) {
case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
- pipe_format = PIPE_FORMAT_Z32_FLOAT;
+ params->pipe_format = PIPE_FORMAT_Z32_FLOAT;
break;
case PIPE_FORMAT_X8Z24_UNORM:
case PIPE_FORMAT_S8_UINT_Z24_UNORM:
- /* Z24 is always stored like this. */
- pipe_format = PIPE_FORMAT_Z24X8_UNORM;
+ /* Z24 is always stored like this for DB
+ * compatibility.
+ */
+ params->pipe_format = PIPE_FORMAT_Z24X8_UNORM;
break;
case PIPE_FORMAT_X24S8_UINT:
case PIPE_FORMAT_S8X24_UINT:
case PIPE_FORMAT_X32_S8X24_UINT:
- pipe_format = PIPE_FORMAT_S8_UINT;
- tile_split = tmp->surface.stencil_tile_split;
- surflevel = tmp->surface.stencil_level;
+ params->pipe_format = PIPE_FORMAT_S8_UINT;
+ tile_split = tmp->surface.u.legacy.stencil_tile_split;
+ surflevel = tmp->surface.u.legacy.stencil_level;
break;
default:;
}
}
if (R600_BIG_ENDIAN)
- do_endian_swap = !(tmp->is_depth && !tmp->is_flushing_texture);
+ do_endian_swap = !tmp->db_compatible;
- format = r600_translate_texformat(ctx->screen, pipe_format,
- swizzle,
+ format = r600_translate_texformat(rctx->b.b.screen, params->pipe_format,
+ params->swizzle,
&word4, &yuv_format, do_endian_swap);
assert(format != ~0);
if (format == ~0) {
- FREE(view);
- return NULL;
+ return -1;
}
endian = r600_colorformat_endian_swap(format, do_endian_swap);
base_level = 0;
- first_level = state->u.tex.first_level;
- last_level = state->u.tex.last_level;
- width = width0;
- height = height0;
+ first_level = params->first_level;
+ last_level = params->last_level;
+ width = params->width0;
+ height = params->height0;
depth = texture->depth0;
- if (force_level) {
- base_level = force_level;
+ if (params->force_level) {
+ base_level = params->force_level;
first_level = 0;
last_level = 0;
- width = u_minify(width, force_level);
- height = u_minify(height, force_level);
- depth = u_minify(depth, force_level);
+ width = u_minify(width, params->force_level);
+ height = u_minify(height, params->force_level);
+ depth = u_minify(depth, params->force_level);
}
- pitch = surflevel[base_level].nblk_x * util_format_get_blockwidth(pipe_format);
+ pitch = surflevel[base_level].nblk_x * util_format_get_blockwidth(params->pipe_format);
non_disp_tiling = tmp->non_disp_tiling;
switch (surflevel[base_level].mode) {
array_mode = V_028C70_ARRAY_1D_TILED_THIN1;
break;
}
- macro_aspect = tmp->surface.mtilea;
- bankw = tmp->surface.bankw;
- bankh = tmp->surface.bankh;
+ macro_aspect = tmp->surface.u.legacy.mtilea;
+ bankw = tmp->surface.u.legacy.bankw;
+ bankh = tmp->surface.u.legacy.bankh;
tile_split = eg_tile_split(tile_split);
macro_aspect = eg_macro_tile_aspect(macro_aspect);
bankw = eg_bank_wh(bankw);
/* 128 bit formats require tile type = 1 */
if (rscreen->b.chip_class == CAYMAN) {
- if (util_format_get_blocksize(pipe_format) >= 16)
+ if (util_format_get_blocksize(params->pipe_format) >= 16)
non_disp_tiling = 1;
}
nbanks = eg_num_banks(rscreen->b.info.r600_num_banks);
- if (state->target == PIPE_TEXTURE_1D_ARRAY) {
- height = 1;
- depth = texture->array_size;
- } else if (state->target == PIPE_TEXTURE_2D_ARRAY) {
- depth = texture->array_size;
- } else if (state->target == PIPE_TEXTURE_CUBE_ARRAY)
- depth = texture->array_size / 6;
va = tmp->resource.gpu_address;
- if (state->format == PIPE_FORMAT_X24S8_UINT ||
- state->format == PIPE_FORMAT_S8X24_UINT ||
- state->format == PIPE_FORMAT_X32_S8X24_UINT ||
- state->format == PIPE_FORMAT_S8_UINT)
- view->is_stencil_sampler = true;
-
- view->tex_resource = &tmp->resource;
-
/* array type views and views into array types need to use layer offset */
- dim = state->target;
- if (state->target != PIPE_TEXTURE_CUBE)
- dim = MAX2(state->target, texture->target);
+ dim = r600_tex_dim(tmp, params->target, texture->nr_samples);
+
+ if (dim == V_030000_SQ_TEX_DIM_1D_ARRAY) {
+ height = 1;
+ depth = texture->array_size;
+ } else if (dim == V_030000_SQ_TEX_DIM_2D_ARRAY ||
+ dim == V_030000_SQ_TEX_DIM_2D_ARRAY_MSAA) {
+ depth = texture->array_size;
+ } else if (dim == V_030000_SQ_TEX_DIM_CUBEMAP)
+ depth = texture->array_size / 6;
- view->tex_resource_words[0] = (S_030000_DIM(r600_tex_dim(dim, texture->nr_samples)) |
- S_030000_PITCH((pitch / 8) - 1) |
- S_030000_TEX_WIDTH(width - 1));
+ tex_resource_words[0] = (S_030000_DIM(dim) |
+ S_030000_PITCH((pitch / 8) - 1) |
+ S_030000_TEX_WIDTH(width - 1));
if (rscreen->b.chip_class == CAYMAN)
- view->tex_resource_words[0] |= CM_S_030000_NON_DISP_TILING_ORDER(non_disp_tiling);
+ tex_resource_words[0] |= CM_S_030000_NON_DISP_TILING_ORDER(non_disp_tiling);
else
- view->tex_resource_words[0] |= S_030000_NON_DISP_TILING_ORDER(non_disp_tiling);
- view->tex_resource_words[1] = (S_030004_TEX_HEIGHT(height - 1) |
+ tex_resource_words[0] |= S_030000_NON_DISP_TILING_ORDER(non_disp_tiling);
+ tex_resource_words[1] = (S_030004_TEX_HEIGHT(height - 1) |
S_030004_TEX_DEPTH(depth - 1) |
S_030004_ARRAY_MODE(array_mode));
- view->tex_resource_words[2] = (surflevel[base_level].offset + va) >> 8;
+ tex_resource_words[2] = (surflevel[base_level].offset + va) >> 8;
+ *skip_mip_address_reloc = false;
/* TEX_RESOURCE_WORD3.MIP_ADDRESS */
if (texture->nr_samples > 1 && rscreen->has_compressed_msaa_texturing) {
if (tmp->is_depth) {
/* disable FMASK (0 = disabled) */
- view->tex_resource_words[3] = 0;
- view->skip_mip_address_reloc = true;
+ tex_resource_words[3] = 0;
+ *skip_mip_address_reloc = true;
} else {
/* FMASK should be in MIP_ADDRESS for multisample textures */
- view->tex_resource_words[3] = (tmp->fmask.offset + va) >> 8;
+ tex_resource_words[3] = (tmp->fmask.offset + va) >> 8;
}
} else if (last_level && texture->nr_samples <= 1) {
- view->tex_resource_words[3] = (surflevel[1].offset + va) >> 8;
+ tex_resource_words[3] = (surflevel[1].offset + va) >> 8;
} else {
- view->tex_resource_words[3] = (surflevel[base_level].offset + va) >> 8;
+ tex_resource_words[3] = (surflevel[base_level].offset + va) >> 8;
}
- last_layer = state->u.tex.last_layer;
- if (state->target != texture->target && depth == 1) {
- last_layer = state->u.tex.first_layer;
+ last_layer = params->last_layer;
+ if (params->target != texture->target && depth == 1) {
+ last_layer = params->first_layer;
}
- view->tex_resource_words[4] = (word4 |
- S_030010_ENDIAN_SWAP(endian));
- view->tex_resource_words[5] = S_030014_BASE_ARRAY(state->u.tex.first_layer) |
- S_030014_LAST_ARRAY(last_layer);
- view->tex_resource_words[6] = S_030018_TILE_SPLIT(tile_split);
+ tex_resource_words[4] = (word4 |
+ S_030010_ENDIAN_SWAP(endian));
+ tex_resource_words[5] = S_030014_BASE_ARRAY(params->first_layer) |
+ S_030014_LAST_ARRAY(last_layer);
+ tex_resource_words[6] = S_030018_TILE_SPLIT(tile_split);
if (texture->nr_samples > 1) {
unsigned log_samples = util_logbase2(texture->nr_samples);
if (rscreen->b.chip_class == CAYMAN) {
- view->tex_resource_words[4] |= S_030010_LOG2_NUM_FRAGMENTS(log_samples);
+ tex_resource_words[4] |= S_030010_LOG2_NUM_FRAGMENTS(log_samples);
}
/* LAST_LEVEL holds log2(nr_samples) for multisample textures */
- view->tex_resource_words[5] |= S_030014_LAST_LEVEL(log_samples);
- view->tex_resource_words[6] |= S_030018_FMASK_BANK_HEIGHT(fmask_bankh);
+ tex_resource_words[5] |= S_030014_LAST_LEVEL(log_samples);
+ tex_resource_words[6] |= S_030018_FMASK_BANK_HEIGHT(fmask_bankh);
} else {
bool no_mip = first_level == last_level;
- view->tex_resource_words[4] |= S_030010_BASE_LEVEL(first_level);
- view->tex_resource_words[5] |= S_030014_LAST_LEVEL(last_level);
+ tex_resource_words[4] |= S_030010_BASE_LEVEL(first_level);
+ tex_resource_words[5] |= S_030014_LAST_LEVEL(last_level);
/* aniso max 16 samples */
- view->tex_resource_words[6] |= S_030018_MAX_ANISO_RATIO(no_mip ? 0 : 4);
+ tex_resource_words[6] |= S_030018_MAX_ANISO_RATIO(no_mip ? 0 : 4);
}
- view->tex_resource_words[7] = S_03001C_DATA_FORMAT(format) |
+ tex_resource_words[7] = S_03001C_DATA_FORMAT(format) |
S_03001C_TYPE(V_03001C_SQ_TEX_VTX_VALID_TEXTURE) |
S_03001C_BANK_WIDTH(bankw) |
S_03001C_BANK_HEIGHT(bankh) |
S_03001C_MACRO_TILE_ASPECT(macro_aspect) |
S_03001C_NUM_BANKS(nbanks) |
- S_03001C_DEPTH_SAMPLE_ORDER(tmp->is_depth && !tmp->is_flushing_texture);
+ S_03001C_DEPTH_SAMPLE_ORDER(tmp->db_compatible);
+ return 0;
+}
+
+struct pipe_sampler_view *
+evergreen_create_sampler_view_custom(struct pipe_context *ctx,
+ struct pipe_resource *texture,
+ const struct pipe_sampler_view *state,
+ unsigned width0, unsigned height0,
+ unsigned force_level)
+{
+ struct r600_context *rctx = (struct r600_context*)ctx;
+ struct r600_pipe_sampler_view *view = CALLOC_STRUCT(r600_pipe_sampler_view);
+ struct r600_texture *tmp = (struct r600_texture*)texture;
+ struct eg_tex_res_params params;
+ int ret;
+
+ if (!view)
+ return NULL;
+
+ /* initialize base object */
+ view->base = *state;
+ view->base.texture = NULL;
+ pipe_reference(NULL, &texture->reference);
+ view->base.texture = texture;
+ view->base.reference.count = 1;
+ view->base.context = ctx;
+
+ if (state->target == PIPE_BUFFER)
+ return texture_buffer_sampler_view(rctx, view, width0, height0);
+
+ memset(¶ms, 0, sizeof(params));
+ params.pipe_format = state->format;
+ params.force_level = force_level;
+ params.width0 = width0;
+ params.height0 = height0;
+ params.first_level = state->u.tex.first_level;
+ params.last_level = state->u.tex.last_level;
+ params.first_layer = state->u.tex.first_layer;
+ params.last_layer = state->u.tex.last_layer;
+ params.target = state->target;
+ params.swizzle[0] = state->swizzle_r;
+ params.swizzle[1] = state->swizzle_g;
+ params.swizzle[2] = state->swizzle_b;
+ params.swizzle[3] = state->swizzle_a;
+
+ ret = evergreen_fill_tex_resource_words(rctx, texture, ¶ms,
+ &view->skip_mip_address_reloc,
+ view->tex_resource_words);
+ if (ret != 0) {
+ FREE(view);
+ return NULL;
+ }
+
+ if (state->format == PIPE_FORMAT_X24S8_UINT ||
+ state->format == PIPE_FORMAT_S8X24_UINT ||
+ state->format == PIPE_FORMAT_X32_S8X24_UINT ||
+ state->format == PIPE_FORMAT_S8_UINT)
+ view->is_stencil_sampler = true;
+
+ view->tex_resource = &tmp->resource;
+
return &view->base;
}
static void evergreen_emit_config_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_config_state *a = (struct r600_config_state*)atom;
radeon_set_config_reg_seq(cs, R_008C04_SQ_GPR_RESOURCE_MGMT_1, 3);
static void evergreen_emit_clip_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct pipe_clip_state *state = &rctx->clip_state.state;
radeon_set_context_reg_seq(cs, R_0285BC_PA_CL_UCP0_X, 6*4);
*br = S_028244_BR_X(scissor.maxx) | S_028244_BR_Y(scissor.maxy);
}
-/**
- * This function intializes the CB* register values for RATs. It is meant
- * to be used for 1D aligned buffers that do not have an associated
- * radeon_surf.
- */
-void evergreen_init_color_surface_rat(struct r600_context *rctx,
- struct r600_surface *surf)
+struct r600_tex_color_info {
+ unsigned info;
+ unsigned view;
+ unsigned dim;
+ unsigned pitch;
+ unsigned slice;
+ unsigned attrib;
+ unsigned ntype;
+ unsigned fmask;
+ unsigned fmask_slice;
+ uint64_t offset;
+ boolean export_16bpc;
+};
+
+static void evergreen_set_color_surface_buffer(struct r600_context *rctx,
+ struct r600_resource *res,
+ enum pipe_format pformat,
+ unsigned first_element,
+ unsigned last_element,
+ struct r600_tex_color_info *color)
{
- struct pipe_resource *pipe_buffer = surf->base.texture;
- unsigned format = r600_translate_colorformat(rctx->b.chip_class,
- surf->base.format, FALSE);
- unsigned endian = r600_colorformat_endian_swap(format, FALSE);
- unsigned swap = r600_translate_colorswap(surf->base.format, FALSE);
- unsigned block_size =
- align(util_format_get_blocksize(pipe_buffer->format), 4);
+ unsigned format, swap, ntype, endian;
+ const struct util_format_description *desc;
+ unsigned block_size = util_format_get_blocksize(res->b.b.format);
unsigned pitch_alignment =
MAX2(64, rctx->screen->b.info.pipe_interleave_bytes / block_size);
- unsigned pitch = align(pipe_buffer->width0, pitch_alignment);
-
- surf->cb_color_base = r600_resource(pipe_buffer)->gpu_address >> 8;
-
- surf->cb_color_pitch = (pitch / 8) - 1;
-
- surf->cb_color_slice = 0;
-
- surf->cb_color_view = 0;
-
- surf->cb_color_info =
- S_028C70_ENDIAN(endian)
- | S_028C70_FORMAT(format)
- | S_028C70_ARRAY_MODE(V_028C70_ARRAY_LINEAR_ALIGNED)
- | S_028C70_NUMBER_TYPE(V_028C70_NUMBER_UINT)
- | S_028C70_COMP_SWAP(swap)
- | S_028C70_BLEND_BYPASS(1) /* We must set this bit because we
- * are using NUMBER_UINT */
- | S_028C70_RAT(1)
- ;
+ unsigned pitch = align(res->b.b.width0, pitch_alignment);
+ int i;
+ unsigned width_elements;
- surf->cb_color_attrib = S_028C74_NON_DISP_TILING_ORDER(1);
+ width_elements = last_element - first_element + 1;
- /* For buffers, CB_COLOR0_DIM needs to be set to the number of
- * elements. */
- surf->cb_color_dim = pipe_buffer->width0;
+ format = r600_translate_colorformat(rctx->b.chip_class, pformat, FALSE);
+ swap = r600_translate_colorswap(pformat, FALSE);
- /* Set the buffer range the GPU will have access to: */
- util_range_add(&r600_resource(pipe_buffer)->valid_buffer_range,
- 0, pipe_buffer->width0);
+ endian = r600_colorformat_endian_swap(format, FALSE);
- surf->cb_color_fmask = surf->cb_color_base;
- surf->cb_color_fmask_slice = 0;
+ desc = util_format_description(pformat);
+ for (i = 0; i < 4; i++) {
+ if (desc->channel[i].type != UTIL_FORMAT_TYPE_VOID) {
+ break;
+ }
+ }
+ ntype = V_028C70_NUMBER_UNORM;
+ if (desc->colorspace == UTIL_FORMAT_COLORSPACE_SRGB)
+ ntype = V_028C70_NUMBER_SRGB;
+ else if (desc->channel[i].type == UTIL_FORMAT_TYPE_SIGNED) {
+ if (desc->channel[i].normalized)
+ ntype = V_028C70_NUMBER_SNORM;
+ else if (desc->channel[i].pure_integer)
+ ntype = V_028C70_NUMBER_SINT;
+ } else if (desc->channel[i].type == UTIL_FORMAT_TYPE_UNSIGNED) {
+ if (desc->channel[i].normalized)
+ ntype = V_028C70_NUMBER_UNORM;
+ else if (desc->channel[i].pure_integer)
+ ntype = V_028C70_NUMBER_UINT;
+ } else if (desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT) {
+ ntype = V_028C70_NUMBER_FLOAT;
+ }
+
+ pitch = (pitch / 8) - 1;
+ color->pitch = S_028C64_PITCH_TILE_MAX(pitch);
+
+ color->info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_LINEAR_ALIGNED);
+ color->info |= S_028C70_FORMAT(format) |
+ S_028C70_COMP_SWAP(swap) |
+ S_028C70_BLEND_CLAMP(0) |
+ S_028C70_BLEND_BYPASS(1) |
+ S_028C70_NUMBER_TYPE(ntype) |
+ S_028C70_ENDIAN(endian);
+ color->attrib = S_028C74_NON_DISP_TILING_ORDER(1);
+ color->ntype = ntype;
+ color->export_16bpc = false;
+ color->dim = width_elements - 1;
+ color->slice = 0; /* (width_elements / 64) - 1;*/
+ color->view = 0;
+ color->offset = (res->gpu_address + first_element) >> 8;
+
+ color->fmask = color->offset;
+ color->fmask_slice = 0;
}
-void evergreen_init_color_surface(struct r600_context *rctx,
- struct r600_surface *surf)
+static void evergreen_set_color_surface_common(struct r600_context *rctx,
+ struct r600_texture *rtex,
+ unsigned level,
+ unsigned first_layer,
+ unsigned last_layer,
+ enum pipe_format pformat,
+ struct r600_tex_color_info *color)
{
struct r600_screen *rscreen = rctx->screen;
- struct r600_texture *rtex = (struct r600_texture*)surf->base.texture;
- unsigned level = surf->base.u.tex.level;
unsigned pitch, slice;
- unsigned color_info, color_attrib, color_dim = 0, color_view;
- unsigned format, swap, ntype, endian;
- uint64_t offset, base_offset;
unsigned non_disp_tiling, macro_aspect, tile_split, bankh, bankw, fmask_bankh, nbanks;
+ unsigned format, swap, ntype, endian;
const struct util_format_description *desc;
- int i;
bool blend_clamp = 0, blend_bypass = 0, do_endian_swap = FALSE;
+ int i;
+
+ color->offset = rtex->surface.u.legacy.level[level].offset;
+ color->view = S_028C6C_SLICE_START(first_layer) |
+ S_028C6C_SLICE_MAX(last_layer);
- offset = rtex->surface.level[level].offset;
- color_view = S_028C6C_SLICE_START(surf->base.u.tex.first_layer) |
- S_028C6C_SLICE_MAX(surf->base.u.tex.last_layer);
+ color->offset += rtex->resource.gpu_address;
+ color->offset >>= 8;
- pitch = (rtex->surface.level[level].nblk_x) / 8 - 1;
- slice = (rtex->surface.level[level].nblk_x * rtex->surface.level[level].nblk_y) / 64;
+ color->dim = 0;
+ pitch = (rtex->surface.u.legacy.level[level].nblk_x) / 8 - 1;
+ slice = (rtex->surface.u.legacy.level[level].nblk_x * rtex->surface.u.legacy.level[level].nblk_y) / 64;
if (slice) {
slice = slice - 1;
}
- color_info = 0;
- switch (rtex->surface.level[level].mode) {
+
+ color->info = 0;
+ switch (rtex->surface.u.legacy.level[level].mode) {
default:
case RADEON_SURF_MODE_LINEAR_ALIGNED:
- color_info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_LINEAR_ALIGNED);
+ color->info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_LINEAR_ALIGNED);
non_disp_tiling = 1;
break;
case RADEON_SURF_MODE_1D:
- color_info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_1D_TILED_THIN1);
+ color->info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_1D_TILED_THIN1);
non_disp_tiling = rtex->non_disp_tiling;
break;
case RADEON_SURF_MODE_2D:
- color_info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_2D_TILED_THIN1);
+ color->info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_2D_TILED_THIN1);
non_disp_tiling = rtex->non_disp_tiling;
break;
}
- tile_split = rtex->surface.tile_split;
- macro_aspect = rtex->surface.mtilea;
- bankw = rtex->surface.bankw;
- bankh = rtex->surface.bankh;
+ tile_split = rtex->surface.u.legacy.tile_split;
+ macro_aspect = rtex->surface.u.legacy.mtilea;
+ bankw = rtex->surface.u.legacy.bankw;
+ bankh = rtex->surface.u.legacy.bankh;
if (rtex->fmask.size)
fmask_bankh = rtex->fmask.bank_height;
else
- fmask_bankh = rtex->surface.bankh;
+ fmask_bankh = rtex->surface.u.legacy.bankh;
tile_split = eg_tile_split(tile_split);
macro_aspect = eg_macro_tile_aspect(macro_aspect);
bankw = eg_bank_wh(bankw);
bankh = eg_bank_wh(bankh);
fmask_bankh = eg_bank_wh(fmask_bankh);
- /* 128 bit formats require tile type = 1 */
if (rscreen->b.chip_class == CAYMAN) {
- if (util_format_get_blocksize(surf->base.format) >= 16)
+ if (util_format_get_blocksize(pformat) >= 16)
non_disp_tiling = 1;
}
nbanks = eg_num_banks(rscreen->b.info.r600_num_banks);
- desc = util_format_description(surf->base.format);
+ desc = util_format_description(pformat);
for (i = 0; i < 4; i++) {
if (desc->channel[i].type != UTIL_FORMAT_TYPE_VOID) {
break;
}
}
-
- color_attrib = S_028C74_TILE_SPLIT(tile_split)|
- S_028C74_NUM_BANKS(nbanks) |
- S_028C74_BANK_WIDTH(bankw) |
- S_028C74_BANK_HEIGHT(bankh) |
- S_028C74_MACRO_TILE_ASPECT(macro_aspect) |
- S_028C74_NON_DISP_TILING_ORDER(non_disp_tiling) |
- S_028C74_FMASK_BANK_HEIGHT(fmask_bankh);
+ color->attrib = S_028C74_TILE_SPLIT(tile_split)|
+ S_028C74_NUM_BANKS(nbanks) |
+ S_028C74_BANK_WIDTH(bankw) |
+ S_028C74_BANK_HEIGHT(bankh) |
+ S_028C74_MACRO_TILE_ASPECT(macro_aspect) |
+ S_028C74_NON_DISP_TILING_ORDER(non_disp_tiling) |
+ S_028C74_FMASK_BANK_HEIGHT(fmask_bankh);
if (rctx->b.chip_class == CAYMAN) {
- color_attrib |= S_028C74_FORCE_DST_ALPHA_1(desc->swizzle[3] ==
+ color->attrib |= S_028C74_FORCE_DST_ALPHA_1(desc->swizzle[3] ==
PIPE_SWIZZLE_1);
if (rtex->resource.b.b.nr_samples > 1) {
unsigned log_samples = util_logbase2(rtex->resource.b.b.nr_samples);
- color_attrib |= S_028C74_NUM_SAMPLES(log_samples) |
+ color->attrib |= S_028C74_NUM_SAMPLES(log_samples) |
S_028C74_NUM_FRAGMENTS(log_samples);
}
}
ntype = V_028C70_NUMBER_UNORM;
else if (desc->channel[i].pure_integer)
ntype = V_028C70_NUMBER_UINT;
+ } else if (desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT) {
+ ntype = V_028C70_NUMBER_FLOAT;
}
if (R600_BIG_ENDIAN)
- do_endian_swap = !(rtex->is_depth && !rtex->is_flushing_texture);
+ do_endian_swap = !rtex->db_compatible;
- format = r600_translate_colorformat(rctx->b.chip_class, surf->base.format,
- do_endian_swap);
+ format = r600_translate_colorformat(rctx->b.chip_class, pformat, do_endian_swap);
assert(format != ~0);
-
- swap = r600_translate_colorswap(surf->base.format, do_endian_swap);
+ swap = r600_translate_colorswap(pformat, do_endian_swap);
assert(swap != ~0);
endian = r600_colorformat_endian_swap(format, do_endian_swap);
blend_bypass = 1;
}
- surf->alphatest_bypass = ntype == V_028C70_NUMBER_UINT || ntype == V_028C70_NUMBER_SINT;
-
- color_info |= S_028C70_FORMAT(format) |
+ color->ntype = ntype;
+ color->info |= S_028C70_FORMAT(format) |
S_028C70_COMP_SWAP(swap) |
S_028C70_BLEND_CLAMP(blend_clamp) |
S_028C70_BLEND_BYPASS(blend_bypass) |
+ S_028C70_SIMPLE_FLOAT(1) |
S_028C70_NUMBER_TYPE(ntype) |
S_028C70_ENDIAN(endian);
+ if (rtex->fmask.size) {
+ color->info |= S_028C70_COMPRESSION(1);
+ }
+
/* EXPORT_NORM is an optimzation that can be enabled for better
* performance in certain cases.
* EXPORT_NORM can be enabled if:
* - 11-bit or smaller UNORM/SNORM/SRGB
* - 16-bit or smaller FLOAT
*/
+ color->export_16bpc = false;
if (desc->colorspace != UTIL_FORMAT_COLORSPACE_ZS &&
((desc->channel[i].size < 12 &&
desc->channel[i].type != UTIL_FORMAT_TYPE_FLOAT &&
ntype != V_028C70_NUMBER_UINT && ntype != V_028C70_NUMBER_SINT) ||
(desc->channel[i].size < 17 &&
desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT))) {
- color_info |= S_028C70_SOURCE_FORMAT(V_028C70_EXPORT_4C_16BPC);
- surf->export_16bpc = true;
+ color->info |= S_028C70_SOURCE_FORMAT(V_028C70_EXPORT_4C_16BPC);
+ color->export_16bpc = true;
}
+ color->pitch = S_028C64_PITCH_TILE_MAX(pitch);
+ color->slice = S_028C68_SLICE_TILE_MAX(slice);
+
if (rtex->fmask.size) {
- color_info |= S_028C70_COMPRESSION(1);
+ color->fmask = (rtex->resource.gpu_address + rtex->fmask.offset) >> 8;
+ color->fmask_slice = S_028C88_TILE_MAX(rtex->fmask.slice_tile_max);
+ } else {
+ color->fmask = color->offset;
+ color->fmask_slice = S_028C88_TILE_MAX(slice);
}
+}
+
+/**
+ * This function intializes the CB* register values for RATs. It is meant
+ * to be used for 1D aligned buffers that do not have an associated
+ * radeon_surf.
+ */
+void evergreen_init_color_surface_rat(struct r600_context *rctx,
+ struct r600_surface *surf)
+{
+ struct pipe_resource *pipe_buffer = surf->base.texture;
+ struct r600_tex_color_info color;
+
+ evergreen_set_color_surface_buffer(rctx, (struct r600_resource *)surf->base.texture,
+ surf->base.format, 0, pipe_buffer->width0,
+ &color);
+
+ surf->cb_color_base = color.offset;
+ surf->cb_color_dim = color.dim;
+ surf->cb_color_info = color.info | S_028C70_RAT(1);
+ surf->cb_color_pitch = color.pitch;
+ surf->cb_color_slice = color.slice;
+ surf->cb_color_view = color.view;
+ surf->cb_color_attrib = color.attrib;
+ surf->cb_color_fmask = color.fmask;
+ surf->cb_color_fmask_slice = color.fmask_slice;
+
+ surf->cb_color_view = 0;
- base_offset = rtex->resource.gpu_address;
+ /* Set the buffer range the GPU will have access to: */
+ util_range_add(pipe_buffer, &r600_resource(pipe_buffer)->valid_buffer_range,
+ 0, pipe_buffer->width0);
+}
+
+
+void evergreen_init_color_surface(struct r600_context *rctx,
+ struct r600_surface *surf)
+{
+ struct r600_texture *rtex = (struct r600_texture*)surf->base.texture;
+ unsigned level = surf->base.u.tex.level;
+ struct r600_tex_color_info color;
+
+ evergreen_set_color_surface_common(rctx, rtex, level,
+ surf->base.u.tex.first_layer,
+ surf->base.u.tex.last_layer,
+ surf->base.format,
+ &color);
+
+ surf->alphatest_bypass = color.ntype == V_028C70_NUMBER_UINT ||
+ color.ntype == V_028C70_NUMBER_SINT;
+ surf->export_16bpc = color.export_16bpc;
/* XXX handle enabling of CB beyond BASE8 which has different offset */
- surf->cb_color_base = (base_offset + offset) >> 8;
- surf->cb_color_dim = color_dim;
- surf->cb_color_info = color_info;
- surf->cb_color_pitch = S_028C64_PITCH_TILE_MAX(pitch);
- surf->cb_color_slice = S_028C68_SLICE_TILE_MAX(slice);
- surf->cb_color_view = color_view;
- surf->cb_color_attrib = color_attrib;
- if (rtex->fmask.size) {
- surf->cb_color_fmask = (base_offset + rtex->fmask.offset) >> 8;
- surf->cb_color_fmask_slice = S_028C88_TILE_MAX(rtex->fmask.slice_tile_max);
- } else {
- surf->cb_color_fmask = surf->cb_color_base;
- surf->cb_color_fmask_slice = S_028C88_TILE_MAX(slice);
- }
+ surf->cb_color_base = color.offset;
+ surf->cb_color_dim = color.dim;
+ surf->cb_color_info = color.info;
+ surf->cb_color_pitch = color.pitch;
+ surf->cb_color_slice = color.slice;
+ surf->cb_color_view = color.view;
+ surf->cb_color_attrib = color.attrib;
+ surf->cb_color_fmask = color.fmask;
+ surf->cb_color_fmask_slice = color.fmask_slice;
surf->color_initialized = true;
}
struct r600_screen *rscreen = rctx->screen;
struct r600_texture *rtex = (struct r600_texture*)surf->base.texture;
unsigned level = surf->base.u.tex.level;
- struct radeon_surf_level *levelinfo = &rtex->surface.level[level];
+ struct legacy_surf_level *levelinfo = &rtex->surface.u.legacy.level[level];
uint64_t offset;
unsigned format, array_mode;
unsigned macro_aspect, tile_split, bankh, bankw, nbanks;
assert(format != ~0);
offset = rtex->resource.gpu_address;
- offset += rtex->surface.level[level].offset;
+ offset += rtex->surface.u.legacy.level[level].offset;
- switch (rtex->surface.level[level].mode) {
+ switch (rtex->surface.u.legacy.level[level].mode) {
case RADEON_SURF_MODE_2D:
array_mode = V_028C70_ARRAY_2D_TILED_THIN1;
break;
array_mode = V_028C70_ARRAY_1D_TILED_THIN1;
break;
}
- tile_split = rtex->surface.tile_split;
- macro_aspect = rtex->surface.mtilea;
- bankw = rtex->surface.bankw;
- bankh = rtex->surface.bankh;
+ tile_split = rtex->surface.u.legacy.tile_split;
+ macro_aspect = rtex->surface.u.legacy.mtilea;
+ bankw = rtex->surface.u.legacy.bankw;
+ bankh = rtex->surface.u.legacy.bankh;
tile_split = eg_tile_split(tile_split);
macro_aspect = eg_macro_tile_aspect(macro_aspect);
bankw = eg_bank_wh(bankw);
surf->db_depth_slice = S_02805C_SLICE_TILE_MAX(levelinfo->nblk_x *
levelinfo->nblk_y / 64 - 1);
- switch (surf->base.format) {
- case PIPE_FORMAT_Z24X8_UNORM:
- case PIPE_FORMAT_Z24_UNORM_S8_UINT:
- case PIPE_FORMAT_X8Z24_UNORM:
- case PIPE_FORMAT_S8_UINT_Z24_UNORM:
- surf->pa_su_poly_offset_db_fmt_cntl =
- S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS((char)-24);
- break;
- case PIPE_FORMAT_Z32_FLOAT:
- case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
- surf->pa_su_poly_offset_db_fmt_cntl =
- S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS((char)-23) |
- S_028B78_POLY_OFFSET_DB_IS_FLOAT_FMT(1);
- break;
- case PIPE_FORMAT_Z16_UNORM:
- surf->pa_su_poly_offset_db_fmt_cntl =
- S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS((char)-16);
- break;
- default:;
- }
-
- if (rtex->surface.flags & RADEON_SURF_SBUFFER) {
+ if (rtex->surface.has_stencil) {
uint64_t stencil_offset;
- unsigned stile_split = rtex->surface.stencil_tile_split;
+ unsigned stile_split = rtex->surface.u.legacy.stencil_tile_split;
stile_split = eg_tile_split(stile_split);
- stencil_offset = rtex->surface.stencil_level[level].offset;
+ stencil_offset = rtex->surface.u.legacy.stencil_level[level].offset;
stencil_offset += rtex->resource.gpu_address;
surf->db_stencil_base = stencil_offset >> 8;
S_028044_FORMAT(V_028044_STENCIL_8);
}
- /* use htile only for first level */
- if (rtex->htile_buffer && !level) {
- uint64_t va = rtex->htile_buffer->gpu_address;
+ if (r600_htile_enabled(rtex, level)) {
+ uint64_t va = rtex->resource.gpu_address + rtex->htile_offset;
surf->db_htile_data_base = va >> 8;
surf->db_htile_surface = S_028ABC_HTILE_WIDTH(1) |
S_028ABC_HTILE_HEIGHT(1) |
struct r600_surface *surf;
struct r600_texture *rtex;
uint32_t i, log_samples;
-
- if (rctx->framebuffer.state.nr_cbufs) {
- rctx->b.flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
- rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_CB |
- R600_CONTEXT_FLUSH_AND_INV_CB_META;
- }
- if (rctx->framebuffer.state.zsbuf) {
- rctx->b.flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
- rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_DB;
-
- rtex = (struct r600_texture*)rctx->framebuffer.state.zsbuf->texture;
- if (rtex->htile_buffer) {
- rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_DB_META;
- }
- }
+ uint32_t target_mask = 0;
+ /* Flush TC when changing the framebuffer state, because the only
+ * client not using TC that can change textures is the framebuffer.
+ * Other places don't typically have to flush TC.
+ */
+ rctx->b.flags |= R600_CONTEXT_WAIT_3D_IDLE |
+ R600_CONTEXT_FLUSH_AND_INV |
+ R600_CONTEXT_FLUSH_AND_INV_CB |
+ R600_CONTEXT_FLUSH_AND_INV_CB_META |
+ R600_CONTEXT_FLUSH_AND_INV_DB |
+ R600_CONTEXT_FLUSH_AND_INV_DB_META |
+ R600_CONTEXT_INV_TEX_CACHE;
util_copy_framebuffer_state(&rctx->framebuffer.state, state);
if (!surf)
continue;
+ target_mask |= (0xf << (i * 4));
+
rtex = (struct r600_texture*)surf->base.texture;
r600_context_add_resource_size(ctx, state->cbufs[i]->texture);
rctx->framebuffer.export_16bpc = false;
}
- if (rtex->fmask.size && rtex->cmask.size) {
+ if (rtex->fmask.size) {
rctx->framebuffer.compressed_cb_mask |= 1 << i;
}
}
r600_mark_atom_dirty(rctx, &rctx->db_misc_state.atom);
}
- if (rctx->cb_misc_state.nr_cbufs != state->nr_cbufs) {
+ if (rctx->cb_misc_state.nr_cbufs != state->nr_cbufs ||
+ rctx->cb_misc_state.bound_cbufs_target_mask != target_mask) {
+ rctx->cb_misc_state.bound_cbufs_target_mask = target_mask;
rctx->cb_misc_state.nr_cbufs = state->nr_cbufs;
r600_mark_atom_dirty(rctx, &rctx->cb_misc_state.atom);
}
r600_mark_atom_dirty(rctx, &rctx->framebuffer.atom);
r600_set_sample_locations_constant_buffer(rctx);
+ rctx->framebuffer.do_update_surf_dirtiness = true;
}
static void evergreen_set_min_samples(struct pipe_context *ctx, unsigned min_samples)
}
/* 8xMSAA */
-static uint32_t sample_locs_8x[] = {
+static const uint32_t sample_locs_8x[] = {
FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
FILL_SREG(-7, -1, -3, -7, 7, -3, -5, 7),
FILL_SREG(-1, 1, 1, 5, 3, -5, 5, 3),
static void evergreen_emit_msaa_state(struct r600_context *rctx, int nr_samples, int ps_iter_samples)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
unsigned max_dist = 0;
switch (nr_samples) {
S_028C00_EXPAND_LINE_WIDTH(1)); /* R_028C00_PA_SC_LINE_CNTL */
radeon_emit(cs, S_028C04_MSAA_NUM_SAMPLES(util_logbase2(nr_samples)) |
S_028C04_MAX_SAMPLE_DIST(max_dist)); /* R_028C04_PA_SC_AA_CONFIG */
- radeon_set_context_reg(cs, EG_R_028A4C_PA_SC_MODE_CNTL_1,
+ radeon_set_context_reg(cs, R_028A4C_PA_SC_MODE_CNTL_1,
EG_S_028A4C_PS_ITER_SAMPLE(ps_iter_samples > 1) |
EG_S_028A4C_FORCE_EOV_CNTDWN_ENABLE(1) |
EG_S_028A4C_FORCE_EOV_REZ_ENABLE(1));
radeon_set_context_reg_seq(cs, R_028C00_PA_SC_LINE_CNTL, 2);
radeon_emit(cs, S_028C00_LAST_PIXEL(1)); /* R_028C00_PA_SC_LINE_CNTL */
radeon_emit(cs, 0); /* R_028C04_PA_SC_AA_CONFIG */
- radeon_set_context_reg(cs, EG_R_028A4C_PA_SC_MODE_CNTL_1,
+ radeon_set_context_reg(cs, R_028A4C_PA_SC_MODE_CNTL_1,
EG_S_028A4C_FORCE_EOV_CNTDWN_ENABLE(1) |
EG_S_028A4C_FORCE_EOV_REZ_ENABLE(1));
}
}
+static void evergreen_emit_image_state(struct r600_context *rctx, struct r600_atom *atom,
+ int immed_id_base, int res_id_base, int offset, uint32_t pkt_flags)
+{
+ struct r600_image_state *state = (struct r600_image_state *)atom;
+ struct pipe_framebuffer_state *fb_state = &rctx->framebuffer.state;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ struct r600_texture *rtex;
+ struct r600_resource *resource;
+ int i;
+
+ for (i = 0; i < R600_MAX_IMAGES; i++) {
+ struct r600_image_view *image = &state->views[i];
+ unsigned reloc, immed_reloc;
+ int idx = i + offset;
+
+ if (!pkt_flags)
+ idx += fb_state->nr_cbufs + (rctx->dual_src_blend ? 1 : 0);
+ if (!image->base.resource)
+ continue;
+
+ resource = (struct r600_resource *)image->base.resource;
+ if (resource->b.b.target != PIPE_BUFFER)
+ rtex = (struct r600_texture *)image->base.resource;
+ else
+ rtex = NULL;
+
+ reloc = radeon_add_to_buffer_list(&rctx->b,
+ &rctx->b.gfx,
+ resource,
+ RADEON_USAGE_READWRITE,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+
+ immed_reloc = radeon_add_to_buffer_list(&rctx->b,
+ &rctx->b.gfx,
+ resource->immed_buffer,
+ RADEON_USAGE_READWRITE,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+
+ if (pkt_flags)
+ radeon_compute_set_context_reg_seq(cs, R_028C60_CB_COLOR0_BASE + idx * 0x3C, 13);
+ else
+ radeon_set_context_reg_seq(cs, R_028C60_CB_COLOR0_BASE + idx * 0x3C, 13);
+
+ radeon_emit(cs, image->cb_color_base); /* R_028C60_CB_COLOR0_BASE */
+ radeon_emit(cs, image->cb_color_pitch); /* R_028C64_CB_COLOR0_PITCH */
+ radeon_emit(cs, image->cb_color_slice); /* R_028C68_CB_COLOR0_SLICE */
+ radeon_emit(cs, image->cb_color_view); /* R_028C6C_CB_COLOR0_VIEW */
+ radeon_emit(cs, image->cb_color_info); /* R_028C70_CB_COLOR0_INFO */
+ radeon_emit(cs, image->cb_color_attrib); /* R_028C74_CB_COLOR0_ATTRIB */
+ radeon_emit(cs, image->cb_color_dim); /* R_028C78_CB_COLOR0_DIM */
+ radeon_emit(cs, rtex ? rtex->cmask.base_address_reg : image->cb_color_base); /* R_028C7C_CB_COLOR0_CMASK */
+ radeon_emit(cs, rtex ? rtex->cmask.slice_tile_max : 0); /* R_028C80_CB_COLOR0_CMASK_SLICE */
+ radeon_emit(cs, image->cb_color_fmask); /* R_028C84_CB_COLOR0_FMASK */
+ radeon_emit(cs, image->cb_color_fmask_slice); /* R_028C88_CB_COLOR0_FMASK_SLICE */
+ radeon_emit(cs, rtex ? rtex->color_clear_value[0] : 0); /* R_028C8C_CB_COLOR0_CLEAR_WORD0 */
+ radeon_emit(cs, rtex ? rtex->color_clear_value[1] : 0); /* R_028C90_CB_COLOR0_CLEAR_WORD1 */
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C60_CB_COLOR0_BASE */
+ radeon_emit(cs, reloc);
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C74_CB_COLOR0_ATTRIB */
+ radeon_emit(cs, reloc);
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C7C_CB_COLOR0_CMASK */
+ radeon_emit(cs, reloc);
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C84_CB_COLOR0_FMASK */
+ radeon_emit(cs, reloc);
+
+ if (pkt_flags)
+ radeon_compute_set_context_reg(cs, R_028B9C_CB_IMMED0_BASE + (idx * 4), resource->immed_buffer->gpu_address >> 8);
+ else
+ radeon_set_context_reg(cs, R_028B9C_CB_IMMED0_BASE + (idx * 4), resource->immed_buffer->gpu_address >> 8);
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0)); /**/
+ radeon_emit(cs, immed_reloc);
+
+ radeon_emit(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
+ radeon_emit(cs, (immed_id_base + i + offset) * 8);
+ radeon_emit_array(cs, image->immed_resource_words, 8);
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, immed_reloc);
+
+ radeon_emit(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
+ radeon_emit(cs, (res_id_base + i + offset) * 8);
+ radeon_emit_array(cs, image->resource_words, 8);
+
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, reloc);
+
+ if (!image->skip_mip_address_reloc) {
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, reloc);
+ }
+ }
+}
+
+static void evergreen_emit_fragment_image_state(struct r600_context *rctx, struct r600_atom *atom)
+{
+ evergreen_emit_image_state(rctx, atom,
+ R600_IMAGE_IMMED_RESOURCE_OFFSET,
+ R600_IMAGE_REAL_RESOURCE_OFFSET, 0, 0);
+}
+
+static void evergreen_emit_compute_image_state(struct r600_context *rctx, struct r600_atom *atom)
+{
+ evergreen_emit_image_state(rctx, atom,
+ EG_FETCH_CONSTANTS_OFFSET_CS + R600_IMAGE_IMMED_RESOURCE_OFFSET,
+ EG_FETCH_CONSTANTS_OFFSET_CS + R600_IMAGE_REAL_RESOURCE_OFFSET,
+ 0, RADEON_CP_PACKET3_COMPUTE_MODE);
+}
+
+static void evergreen_emit_fragment_buffer_state(struct r600_context *rctx, struct r600_atom *atom)
+{
+ int offset = util_bitcount(rctx->fragment_images.enabled_mask);
+ evergreen_emit_image_state(rctx, atom,
+ R600_IMAGE_IMMED_RESOURCE_OFFSET,
+ R600_IMAGE_REAL_RESOURCE_OFFSET, offset, 0);
+}
+
+static void evergreen_emit_compute_buffer_state(struct r600_context *rctx, struct r600_atom *atom)
+{
+ int offset = util_bitcount(rctx->compute_images.enabled_mask);
+ evergreen_emit_image_state(rctx, atom,
+ EG_FETCH_CONSTANTS_OFFSET_CS + R600_IMAGE_IMMED_RESOURCE_OFFSET,
+ EG_FETCH_CONSTANTS_OFFSET_CS + R600_IMAGE_REAL_RESOURCE_OFFSET,
+ offset, RADEON_CP_PACKET3_COMPUTE_MODE);
+}
+
static void evergreen_emit_framebuffer_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct pipe_framebuffer_state *state = &rctx->framebuffer.state;
unsigned nr_cbufs = state->nr_cbufs;
unsigned i, tl, br;
&rctx->b.gfx,
(struct r600_resource*)cb->base.texture,
RADEON_USAGE_READWRITE,
- tex->surface.nsamples > 1 ?
+ tex->resource.b.b.nr_samples > 1 ?
RADEON_PRIO_COLOR_BUFFER_MSAA :
RADEON_PRIO_COLOR_BUFFER);
if (tex->cmask_buffer && tex->cmask_buffer != &tex->resource) {
cmask_reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
tex->cmask_buffer, RADEON_USAGE_READWRITE,
- RADEON_PRIO_CMASK);
+ RADEON_PRIO_SEPARATE_META);
} else {
cmask_reloc = reloc;
}
radeon_emit(cs, reloc);
}
/* set CB_COLOR1_INFO for possible dual-src blending */
- if (i == 1 && state->cbufs[0]) {
+ if (rctx->framebuffer.dual_src_blend && i == 1 && state->cbufs[0]) {
radeon_set_context_reg(cs, R_028C70_CB_COLOR0_INFO + 1 * 0x3C,
cb->cb_color_info | tex->cb_color_info);
i++;
}
+ i += util_bitcount(rctx->fragment_images.enabled_mask);
+ i += util_bitcount(rctx->fragment_buffers.enabled_mask);
for (; i < 8 ; i++)
radeon_set_context_reg(cs, R_028C70_CB_COLOR0_INFO + i * 0x3C, 0);
for (; i < 12; i++)
RADEON_PRIO_DEPTH_BUFFER_MSAA :
RADEON_PRIO_DEPTH_BUFFER);
- radeon_set_context_reg(cs, R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL,
- zb->pa_su_poly_offset_db_fmt_cntl);
radeon_set_context_reg(cs, R_028008_DB_DEPTH_VIEW, zb->db_depth_view);
radeon_set_context_reg_seq(cs, R_028040_DB_Z_INFO, 8);
if (rctx->b.chip_class == EVERGREEN) {
evergreen_emit_msaa_state(rctx, rctx->framebuffer.nr_samples, rctx->ps_iter_samples);
} else {
- cayman_emit_msaa_sample_locs(cs, rctx->framebuffer.nr_samples);
- cayman_emit_msaa_config(cs, rctx->framebuffer.nr_samples, rctx->ps_iter_samples, 0);
+ cayman_emit_msaa_state(cs, rctx->framebuffer.nr_samples,
+ rctx->ps_iter_samples, 0);
}
}
static void evergreen_emit_polygon_offset(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_poly_offset_state *state = (struct r600_poly_offset_state*)a;
float offset_units = state->offset_units;
float offset_scale = state->offset_scale;
+ uint32_t pa_su_poly_offset_db_fmt_cntl = 0;
- switch (state->zs_format) {
- case PIPE_FORMAT_Z24X8_UNORM:
- case PIPE_FORMAT_Z24_UNORM_S8_UINT:
- case PIPE_FORMAT_X8Z24_UNORM:
- case PIPE_FORMAT_S8_UINT_Z24_UNORM:
- offset_units *= 2.0f;
- break;
- case PIPE_FORMAT_Z16_UNORM:
- offset_units *= 4.0f;
- break;
- default:;
+ if (!state->offset_units_unscaled) {
+ switch (state->zs_format) {
+ case PIPE_FORMAT_Z24X8_UNORM:
+ case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ case PIPE_FORMAT_X8Z24_UNORM:
+ case PIPE_FORMAT_S8_UINT_Z24_UNORM:
+ offset_units *= 2.0f;
+ pa_su_poly_offset_db_fmt_cntl =
+ S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS((char)-24);
+ break;
+ case PIPE_FORMAT_Z16_UNORM:
+ offset_units *= 4.0f;
+ pa_su_poly_offset_db_fmt_cntl =
+ S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS((char)-16);
+ break;
+ default:
+ pa_su_poly_offset_db_fmt_cntl =
+ S_028B78_POLY_OFFSET_NEG_NUM_DB_BITS((char)-23) |
+ S_028B78_POLY_OFFSET_DB_IS_FLOAT_FMT(1);
+ }
}
radeon_set_context_reg_seq(cs, R_028B80_PA_SU_POLY_OFFSET_FRONT_SCALE, 4);
radeon_emit(cs, fui(offset_units));
radeon_emit(cs, fui(offset_scale));
radeon_emit(cs, fui(offset_units));
+
+ radeon_set_context_reg(cs, R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL,
+ pa_su_poly_offset_db_fmt_cntl);
+}
+
+uint32_t evergreen_construct_rat_mask(struct r600_context *rctx, struct r600_cb_misc_state *a,
+ unsigned nr_cbufs)
+{
+ unsigned base_mask = 0;
+ unsigned dirty_mask = a->image_rat_enabled_mask;
+ while (dirty_mask) {
+ unsigned idx = u_bit_scan(&dirty_mask);
+ base_mask |= (0xf << (idx * 4));
+ }
+ unsigned offset = util_last_bit(a->image_rat_enabled_mask);
+ dirty_mask = a->buffer_rat_enabled_mask;
+ while (dirty_mask) {
+ unsigned idx = u_bit_scan(&dirty_mask);
+ base_mask |= (0xf << (idx + offset) * 4);
+ }
+ return base_mask << (nr_cbufs * 4);
}
static void evergreen_emit_cb_misc_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_cb_misc_state *a = (struct r600_cb_misc_state*)atom;
- unsigned fb_colormask = (1ULL << ((unsigned)a->nr_cbufs * 4)) - 1;
- unsigned ps_colormask = (1ULL << ((unsigned)a->nr_ps_color_outputs * 4)) - 1;
-
+ unsigned fb_colormask = a->bound_cbufs_target_mask;
+ unsigned ps_colormask = a->ps_color_export_mask;
+ unsigned rat_colormask = evergreen_construct_rat_mask(rctx, a, a->nr_cbufs);
radeon_set_context_reg_seq(cs, R_028238_CB_TARGET_MASK, 2);
- radeon_emit(cs, a->blend_colormask & fb_colormask); /* R_028238_CB_TARGET_MASK */
+ radeon_emit(cs, (a->blend_colormask & fb_colormask) | rat_colormask); /* R_028238_CB_TARGET_MASK */
/* This must match the used export instructions exactly.
* Other values may lead to undefined behavior and hangs.
*/
static void evergreen_emit_db_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_db_state *a = (struct r600_db_state*)atom;
if (a->rsurf && a->rsurf->db_htile_surface) {
radeon_set_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, a->rsurf->db_htile_surface);
radeon_set_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, a->rsurf->db_preload_control);
radeon_set_context_reg(cs, R_028014_DB_HTILE_DATA_BASE, a->rsurf->db_htile_data_base);
- reloc_idx = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, rtex->htile_buffer,
- RADEON_USAGE_READWRITE, RADEON_PRIO_HTILE);
- cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
- cs->buf[cs->cdw++] = reloc_idx;
+ reloc_idx = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, &rtex->resource,
+ RADEON_USAGE_READWRITE, RADEON_PRIO_SEPARATE_META);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc_idx);
} else {
radeon_set_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, 0);
radeon_set_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, 0);
static void evergreen_emit_db_misc_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_db_misc_state *a = (struct r600_db_misc_state*)atom;
unsigned db_render_control = 0;
unsigned db_count_control = 0;
unsigned resource_offset,
unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
unsigned buffer_index = u_bit_scan(&dirty_mask);
vb = &state->vb[buffer_index];
- rbuffer = (struct r600_resource*)vb->buffer;
+ rbuffer = (struct r600_resource*)vb->buffer.resource;
assert(rbuffer);
va = rbuffer->gpu_address + vb->buffer_offset;
unsigned reg_alu_const_cache,
unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
va = rbuffer->gpu_address + cb->buffer_offset;
- if (!gs_ring_buffer) {
+ if (buffer_index < R600_MAX_HW_CONST_BUFFERS) {
radeon_set_context_reg_flag(cs, reg_alu_constbuf_size + buffer_index * 4,
DIV_ROUND_UP(cb->buffer_size, 256), pkt_flags);
radeon_set_context_reg_flag(cs, reg_alu_const_cache + buffer_index * 4, va >> 8,
pkt_flags);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
+ radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, rbuffer,
+ RADEON_USAGE_READ, RADEON_PRIO_CONST_BUFFER));
}
- radeon_emit(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
- radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, rbuffer,
- RADEON_USAGE_READ, RADEON_PRIO_CONST_BUFFER));
-
radeon_emit(cs, PKT3(PKT3_SET_RESOURCE, 8, 0) | pkt_flags);
radeon_emit(cs, (buffer_id_base + buffer_index) * 8);
radeon_emit(cs, va); /* RESOURCEi_WORD0 */
- radeon_emit(cs, rbuffer->b.b.width0 - cb->buffer_offset - 1); /* RESOURCEi_WORD1 */
+ radeon_emit(cs, cb->buffer_size -1); /* RESOURCEi_WORD1 */
radeon_emit(cs, /* RESOURCEi_WORD2 */
S_030008_ENDIAN_SWAP(gs_ring_buffer ? ENDIAN_NONE : r600_endian_swap(32)) |
S_030008_STRIDE(gs_ring_buffer ? 4 : 16) |
0);
}
+void evergreen_setup_scratch_buffers(struct r600_context *rctx) {
+ static const struct {
+ unsigned ring_base;
+ unsigned item_size;
+ unsigned ring_size;
+ } regs[EG_NUM_HW_STAGES] = {
+ [R600_HW_STAGE_PS] = { R_008C68_SQ_PSTMP_RING_BASE, R_028914_SQ_PSTMP_RING_ITEMSIZE, R_008C6C_SQ_PSTMP_RING_SIZE },
+ [R600_HW_STAGE_VS] = { R_008C60_SQ_VSTMP_RING_BASE, R_028910_SQ_VSTMP_RING_ITEMSIZE, R_008C64_SQ_VSTMP_RING_SIZE },
+ [R600_HW_STAGE_GS] = { R_008C58_SQ_GSTMP_RING_BASE, R_02890C_SQ_GSTMP_RING_ITEMSIZE, R_008C5C_SQ_GSTMP_RING_SIZE },
+ [R600_HW_STAGE_ES] = { R_008C50_SQ_ESTMP_RING_BASE, R_028908_SQ_ESTMP_RING_ITEMSIZE, R_008C54_SQ_ESTMP_RING_SIZE },
+ [EG_HW_STAGE_LS] = { R_008E10_SQ_LSTMP_RING_BASE, R_028830_SQ_LSTMP_RING_ITEMSIZE, R_008E14_SQ_LSTMP_RING_SIZE },
+ [EG_HW_STAGE_HS] = { R_008E18_SQ_HSTMP_RING_BASE, R_028834_SQ_HSTMP_RING_ITEMSIZE, R_008E1C_SQ_HSTMP_RING_SIZE }
+ };
+
+ for (unsigned i = 0; i < EG_NUM_HW_STAGES; i++) {
+ struct r600_pipe_shader *stage = rctx->hw_shader_stages[i].shader;
+
+ if (stage && unlikely(stage->scratch_space_needed)) {
+ r600_setup_scratch_area_for_shader(rctx, stage,
+ &rctx->scratch_buffers[i], regs[i].ring_base, regs[i].item_size, regs[i].ring_size);
+ }
+ }
+}
+
static void evergreen_emit_sampler_views(struct r600_context *rctx,
struct r600_samplerview_state *state,
unsigned resource_id_base, unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
static void evergreen_emit_tes_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
{
+ if (!rctx->tes_shader)
+ return;
evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_TESS_EVAL].views,
EG_FETCH_CONSTANTS_OFFSET_VS + R600_MAX_CONST_BUFFERS, 0);
}
static void evergreen_emit_cs_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
{
evergreen_emit_sampler_views(rctx, &rctx->samplers[PIPE_SHADER_COMPUTE].views,
- EG_FETCH_CONSTANTS_OFFSET_CS + 2, RADEON_CP_PACKET3_COMPUTE_MODE);
+ EG_FETCH_CONSTANTS_OFFSET_CS + R600_MAX_CONST_BUFFERS, RADEON_CP_PACKET3_COMPUTE_MODE);
+}
+
+static void evergreen_convert_border_color(union pipe_color_union *in,
+ union pipe_color_union *out,
+ enum pipe_format format)
+{
+ if (util_format_is_pure_integer(format) &&
+ !util_format_is_depth_or_stencil(format)) {
+ const struct util_format_description *d = util_format_description(format);
+
+ for (int i = 0; i < d->nr_channels; ++i) {
+ int cs = d->channel[i].size;
+ if (d->channel[i].type == UTIL_FORMAT_TYPE_SIGNED)
+ out->f[i] = (double)(in->i[i]) / ((1ul << (cs - 1)) - 1 );
+ else if (d->channel[i].type == UTIL_FORMAT_TYPE_UNSIGNED)
+ out->f[i] = (double)(in->ui[i]) / ((1ul << cs) - 1 );
+ else
+ out->f[i] = 0;
+ }
+
+ } else {
+ switch (format) {
+ case PIPE_FORMAT_X24S8_UINT:
+ case PIPE_FORMAT_X32_S8X24_UINT:
+ out->f[0] = (double)(in->ui[0]) / 255.0;
+ out->f[1] = out->f[2] = out->f[3] = 0.0f;
+ break;
+ default:
+ memcpy(out->f, in->f, 4 * sizeof(float));
+ }
+ }
}
static void evergreen_emit_sampler_states(struct r600_context *rctx,
unsigned border_index_reg,
unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
uint32_t dirty_mask = texinfo->states.dirty_mask;
+ union pipe_color_union border_color = {{0,0,0,1}};
+ union pipe_color_union *border_color_ptr = &border_color;
while (dirty_mask) {
struct r600_pipe_sampler_state *rstate;
rstate = texinfo->states.states[i];
assert(rstate);
+ if (rstate->border_color_use) {
+ struct r600_pipe_sampler_view *rview = texinfo->views.views[i];
+ if (rview) {
+ evergreen_convert_border_color(&rstate->border_color,
+ &border_color, rview->base.format);
+ } else {
+ border_color_ptr = &rstate->border_color;
+ }
+ }
+
radeon_emit(cs, PKT3(PKT3_SET_SAMPLER, 3, 0) | pkt_flags);
radeon_emit(cs, (resource_id_base + i) * 3);
radeon_emit_array(cs, rstate->tex_sampler_words, 3);
if (rstate->border_color_use) {
radeon_set_config_reg_seq(cs, border_index_reg, 5);
radeon_emit(cs, i);
- radeon_emit_array(cs, rstate->border_color.ui, 4);
+ radeon_emit_array(cs, border_color_ptr->ui, 4);
}
}
texinfo->states.dirty_mask = 0;
static void evergreen_emit_tes_sampler_states(struct r600_context *rctx, struct r600_atom *atom)
{
+ if (!rctx->tes_shader)
+ return;
evergreen_emit_sampler_states(rctx, &rctx->samplers[PIPE_SHADER_TESS_EVAL], 18,
R_00A414_TD_VS_SAMPLER0_BORDER_INDEX, 0);
}
static void cayman_emit_sample_mask(struct r600_context *rctx, struct r600_atom *a)
{
struct r600_sample_mask *s = (struct r600_sample_mask*)a;
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
uint16_t mask = s->sample_mask;
radeon_set_context_reg_seq(cs, CM_R_028C38_PA_SC_AA_MASK_X0Y0_X1Y0, 2);
static void evergreen_emit_vertex_fetch_shader(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_cso_state *state = (struct r600_cso_state*)a;
struct r600_fetch_shader *shader = (struct r600_fetch_shader*)state->cso;
+ if (!shader)
+ return;
+
radeon_set_context_reg(cs, R_0288A4_SQ_PGM_START_FS,
(shader->buffer->gpu_address + shader->offset) >> 8);
radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, shader->buffer,
RADEON_USAGE_READ,
- RADEON_PRIO_INTERNAL_SHADER));
+ RADEON_PRIO_SHADER_BINARY));
}
static void evergreen_emit_shader_stages(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_shader_stages_state *state = (struct r600_shader_stages_state*)a;
uint32_t v = 0, v2 = 0, primid = 0, tf_param = 0;
static void evergreen_emit_gs_rings(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->b.gfx.cs;
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
struct r600_gs_rings_state *state = (struct r600_gs_rings_state*)a;
struct r600_resource *rbuffer;
radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, rbuffer,
RADEON_USAGE_READWRITE,
- RADEON_PRIO_RINGS_STREAMOUT));
+ RADEON_PRIO_SHADER_RINGS));
radeon_set_config_reg(cs, R_008C44_SQ_ESGS_RING_SIZE,
state->esgs_ring.buffer_size >> 8);
radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
radeon_emit(cs, radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, rbuffer,
RADEON_USAGE_READWRITE,
- RADEON_PRIO_RINGS_STREAMOUT));
+ RADEON_PRIO_SHADER_RINGS));
radeon_set_config_reg(cs, R_008C4C_SQ_GSVS_RING_SIZE,
state->gsvs_ring.buffer_size >> 8);
} else {
static void cayman_init_atom_start_cs(struct r600_context *rctx)
{
struct r600_command_buffer *cb = &rctx->start_cs_cmd;
- int tmp, i;
+ int i;
r600_init_command_buffer(cb, 338);
r600_store_value(cb, 0x76543210); /* CM_R_028BD4_PA_SC_CENTROID_PRIORITY_0 */
r600_store_value(cb, 0xfedcba98); /* CM_R_028BD8_PA_SC_CENTROID_PRIORITY_1 */
+ r600_store_context_reg(cb, R_028724_GDS_ADDR_SIZE, 0x3fff);
r600_store_context_reg_seq(cb, R_0288E8_SQ_LDS_ALLOC, 2);
r600_store_value(cb, 0); /* R_0288E8_SQ_LDS_ALLOC */
r600_store_value(cb, 0); /* R_0288EC_SQ_LDS_ALLOC_PS */
r600_store_context_reg(cb, R_028200_PA_SC_WINDOW_OFFSET, 0);
r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
- r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * R600_MAX_VIEWPORTS);
- for (tmp = 0; tmp < R600_MAX_VIEWPORTS; tmp++) {
- r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
- r600_store_value(cb, fui(1.0)); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
- }
-
r600_store_context_reg(cb, R_028230_PA_SC_EDGERULE, 0xAAAAAAAA);
r600_store_context_reg(cb, R_028820_PA_CL_NANINF_CNTL, 0);
r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
r600_store_context_reg(cb, R_028230_PA_SC_EDGERULE, 0xAAAAAAAA);
- r600_store_context_reg_seq(cb, R_0282D0_PA_SC_VPORT_ZMIN_0, 2 * R600_MAX_VIEWPORTS);
- for (tmp = 0; tmp < R600_MAX_VIEWPORTS; tmp++) {
- r600_store_value(cb, 0); /* R_0282D0_PA_SC_VPORT_ZMIN_0 */
- r600_store_value(cb, fui(1.0)); /* R_0282D4_PA_SC_VPORT_ZMAX_0 */
- }
-
r600_store_context_reg(cb, R_0286DC_SPI_FOG_CNTL, 0);
r600_store_context_reg(cb, R_028820_PA_CL_NANINF_CNTL, 0);
spi_baryc_cntl |= spi_baryc_enable_bit[k];
have_perspective |= k < 3;
have_linear |= !(k < 3);
+ if (rshader->input[i].uses_interpolate_at_centroid) {
+ k = eg_get_interpolator_index(
+ rshader->input[i].interpolate,
+ TGSI_INTERPOLATE_LOC_CENTROID);
+ spi_baryc_cntl |= spi_baryc_enable_bit[k];
+ }
}
}
if (sid) {
tmp = S_028644_SEMANTIC(sid);
+ /* D3D 9 behaviour. GL is undefined */
+ if (rshader->input[i].name == TGSI_SEMANTIC_COLOR && rshader->input[i].sid == 0)
+ tmp |= S_028644_DEFAULT_VAL(3);
+
if (rshader->input[i].name == TGSI_SEMANTIC_POSITION ||
rshader->input[i].interpolate == TGSI_INTERPOLATE_CONSTANT ||
(rshader->input[i].interpolate == TGSI_INTERPOLATE_COLOR &&
db_shader_control |= S_02880C_STENCIL_EXPORT_ENABLE(stencil_export);
db_shader_control |= S_02880C_MASK_EXPORT_ENABLE(mask_export);
+ if (shader->selector->info.properties[TGSI_PROPERTY_FS_EARLY_DEPTH_STENCIL]) {
+ db_shader_control |= S_02880C_DEPTH_BEFORE_SHADER(1) |
+ S_02880C_EXEC_ON_NOOP(shader->selector->info.writes_memory);
+ } else if (shader->selector->info.writes_memory) {
+ db_shader_control |= S_02880C_EXEC_ON_HIER_FAIL(1);
+ }
+
switch (rshader->ps_conservative_z) {
default: /* fall through */
case TGSI_FS_DEPTH_LAYOUT_ANY:
exports_ps |= 1;
}
- num_cout = rshader->nr_ps_color_exports;
+ num_cout = rshader->ps_export_highest + 1;
exports_ps |= S_02884C_EXPORT_COLORS(num_cout);
if (!exports_ps) {
exports_ps = 2;
}
shader->nr_ps_color_outputs = num_cout;
+ shader->ps_color_export_mask = rshader->ps_color_export_mask;
if (ninterp == 0) {
ninterp = 1;
have_perspective = TRUE;
r600_store_value(cb, /* R_028844_SQ_PGM_RESOURCES_PS */
S_028844_NUM_GPRS(rshader->bc.ngpr) |
S_028844_PRIME_CACHE_ON_DRAW(1) |
+ S_028844_DX10_CLAMP(1) |
S_028844_STACK_SIZE(rshader->bc.nstack));
/* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
r600_store_context_reg(cb, R_028890_SQ_PGM_RESOURCES_ES,
S_028890_NUM_GPRS(rshader->bc.ngpr) |
+ S_028890_DX10_CLAMP(1) |
S_028890_STACK_SIZE(rshader->bc.nstack));
r600_store_context_reg(cb, R_02888C_SQ_PGM_START_ES,
shader->bo->gpu_address >> 8);
r600_store_context_reg(cb, R_028878_SQ_PGM_RESOURCES_GS,
S_028878_NUM_GPRS(rshader->bc.ngpr) |
+ S_028878_DX10_CLAMP(1) |
S_028878_STACK_SIZE(rshader->bc.nstack));
r600_store_context_reg(cb, R_028874_SQ_PGM_START_GS,
shader->bo->gpu_address >> 8);
S_0286C4_VS_EXPORT_COUNT(nparams - 1));
r600_store_context_reg(cb, R_028860_SQ_PGM_RESOURCES_VS,
S_028860_NUM_GPRS(rshader->bc.ngpr) |
+ S_028860_DX10_CLAMP(1) |
S_028860_STACK_SIZE(rshader->bc.nstack));
if (rshader->vs_position_window_space) {
r600_store_context_reg(cb, R_028818_PA_CL_VTE_CNTL,
/* After that, the NOP relocation packet must be emitted (shader->bo, RADEON_USAGE_READ). */
shader->pa_cl_vs_out_cntl =
- S_02881C_VS_OUT_CCDIST0_VEC_ENA((rshader->clip_dist_write & 0x0F) != 0) |
- S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->clip_dist_write & 0xF0) != 0) |
+ S_02881C_VS_OUT_CCDIST0_VEC_ENA((rshader->cc_dist_mask & 0x0F) != 0) |
+ S_02881C_VS_OUT_CCDIST1_VEC_ENA((rshader->cc_dist_mask & 0xF0) != 0) |
S_02881C_VS_OUT_MISC_VEC_ENA(rshader->vs_out_misc_write) |
S_02881C_USE_VTX_POINT_SIZE(rshader->vs_out_point_size) |
S_02881C_USE_VTX_EDGE_FLAG(rshader->vs_out_edgeflag) |
r600_init_command_buffer(cb, 32);
r600_store_context_reg(cb, R_0288BC_SQ_PGM_RESOURCES_HS,
S_0288BC_NUM_GPRS(rshader->bc.ngpr) |
+ S_0288BC_DX10_CLAMP(1) |
S_0288BC_STACK_SIZE(rshader->bc.nstack));
r600_store_context_reg(cb, R_0288B8_SQ_PGM_START_HS,
shader->bo->gpu_address >> 8);
r600_init_command_buffer(cb, 32);
r600_store_context_reg(cb, R_0288D4_SQ_PGM_RESOURCES_LS,
S_0288D4_NUM_GPRS(rshader->bc.ngpr) |
+ S_0288D4_DX10_CLAMP(1) |
S_0288D4_STACK_SIZE(rshader->bc.nstack));
r600_store_context_reg(cb, R_0288D0_SQ_PGM_START_LS,
shader->bo->gpu_address >> 8);
* get a hang unless you flush the DB in between. For now just use
* LATE_Z.
*/
- if (rctx->alphatest_state.sx_alpha_test_control) {
+ if (rctx->alphatest_state.sx_alpha_test_control || rctx->ps_shader->info.writes_memory) {
db_shader_control |= S_02880C_Z_ORDER(V_02880C_LATE_Z);
} else {
db_shader_control |= S_02880C_Z_ORDER(V_02880C_EARLY_Z_THEN_LATE_Z);
unsigned pitch,
unsigned bpp)
{
- struct radeon_winsys_cs *cs = rctx->b.dma.cs;
+ struct radeon_cmdbuf *cs = rctx->b.dma.cs;
struct r600_texture *rsrc = (struct r600_texture*)src;
struct r600_texture *rdst = (struct r600_texture*)dst;
unsigned array_mode, lbpp, pitch_tile_max, slice_tile_max, size;
unsigned sub_cmd, bank_h, bank_w, mt_aspect, nbanks, tile_split, non_disp_tiling = 0;
uint64_t base, addr;
- dst_mode = rdst->surface.level[dst_level].mode;
- src_mode = rsrc->surface.level[src_level].mode;
+ dst_mode = rdst->surface.u.legacy.level[dst_level].mode;
+ src_mode = rsrc->surface.u.legacy.level[src_level].mode;
assert(dst_mode != src_mode);
/* non_disp_tiling bit needs to be set for depth, stencil, and fmask surfaces */
if (dst_mode == RADEON_SURF_MODE_LINEAR_ALIGNED) {
/* T2L */
array_mode = evergreen_array_mode(src_mode);
- slice_tile_max = (rsrc->surface.level[src_level].nblk_x * rsrc->surface.level[src_level].nblk_y) / (8*8);
+ slice_tile_max = (rsrc->surface.u.legacy.level[src_level].nblk_x * rsrc->surface.u.legacy.level[src_level].nblk_y) / (8*8);
slice_tile_max = slice_tile_max ? slice_tile_max - 1 : 0;
/* linear height must be the same as the slice tile max height, it's ok even
* if the linear destination/source have smaller heigh as the size of the
* dma packet will be using the copy_height which is always smaller or equal
* to the linear height
*/
- height = rsrc->surface.level[src_level].npix_y;
+ height = u_minify(rsrc->resource.b.b.height0, src_level);
detile = 1;
x = src_x;
y = src_y;
z = src_z;
- base = rsrc->surface.level[src_level].offset;
- addr = rdst->surface.level[dst_level].offset;
- addr += rdst->surface.level[dst_level].slice_size * dst_z;
+ base = rsrc->surface.u.legacy.level[src_level].offset;
+ addr = rdst->surface.u.legacy.level[dst_level].offset;
+ addr += (uint64_t)rdst->surface.u.legacy.level[dst_level].slice_size_dw * 4 * dst_z;
addr += dst_y * pitch + dst_x * bpp;
- bank_h = eg_bank_wh(rsrc->surface.bankh);
- bank_w = eg_bank_wh(rsrc->surface.bankw);
- mt_aspect = eg_macro_tile_aspect(rsrc->surface.mtilea);
- tile_split = eg_tile_split(rsrc->surface.tile_split);
+ bank_h = eg_bank_wh(rsrc->surface.u.legacy.bankh);
+ bank_w = eg_bank_wh(rsrc->surface.u.legacy.bankw);
+ mt_aspect = eg_macro_tile_aspect(rsrc->surface.u.legacy.mtilea);
+ tile_split = eg_tile_split(rsrc->surface.u.legacy.tile_split);
base += rsrc->resource.gpu_address;
addr += rdst->resource.gpu_address;
} else {
/* L2T */
array_mode = evergreen_array_mode(dst_mode);
- slice_tile_max = (rdst->surface.level[dst_level].nblk_x * rdst->surface.level[dst_level].nblk_y) / (8*8);
+ slice_tile_max = (rdst->surface.u.legacy.level[dst_level].nblk_x * rdst->surface.u.legacy.level[dst_level].nblk_y) / (8*8);
slice_tile_max = slice_tile_max ? slice_tile_max - 1 : 0;
/* linear height must be the same as the slice tile max height, it's ok even
* if the linear destination/source have smaller heigh as the size of the
* dma packet will be using the copy_height which is always smaller or equal
* to the linear height
*/
- height = rdst->surface.level[dst_level].npix_y;
+ height = u_minify(rdst->resource.b.b.height0, dst_level);
detile = 0;
x = dst_x;
y = dst_y;
z = dst_z;
- base = rdst->surface.level[dst_level].offset;
- addr = rsrc->surface.level[src_level].offset;
- addr += rsrc->surface.level[src_level].slice_size * src_z;
+ base = rdst->surface.u.legacy.level[dst_level].offset;
+ addr = rsrc->surface.u.legacy.level[src_level].offset;
+ addr += (uint64_t)rsrc->surface.u.legacy.level[src_level].slice_size_dw * 4 * src_z;
addr += src_y * pitch + src_x * bpp;
- bank_h = eg_bank_wh(rdst->surface.bankh);
- bank_w = eg_bank_wh(rdst->surface.bankw);
- mt_aspect = eg_macro_tile_aspect(rdst->surface.mtilea);
- tile_split = eg_tile_split(rdst->surface.tile_split);
+ bank_h = eg_bank_wh(rdst->surface.u.legacy.bankh);
+ bank_w = eg_bank_wh(rdst->surface.u.legacy.bankw);
+ mt_aspect = eg_macro_tile_aspect(rdst->surface.u.legacy.mtilea);
+ tile_split = eg_tile_split(rdst->surface.u.legacy.tile_split);
base += rdst->resource.gpu_address;
addr += rsrc->resource.gpu_address;
}
size = (cheight * pitch) / 4;
/* emit reloc before writing cs so that cs is always in consistent state */
radeon_add_to_buffer_list(&rctx->b, &rctx->b.dma, &rsrc->resource,
- RADEON_USAGE_READ, RADEON_PRIO_SDMA_TEXTURE);
+ RADEON_USAGE_READ, 0);
radeon_add_to_buffer_list(&rctx->b, &rctx->b.dma, &rdst->resource,
- RADEON_USAGE_WRITE, RADEON_PRIO_SDMA_TEXTURE);
- cs->buf[cs->cdw++] = DMA_PACKET(DMA_PACKET_COPY, sub_cmd, size);
- cs->buf[cs->cdw++] = base >> 8;
- cs->buf[cs->cdw++] = (detile << 31) | (array_mode << 27) |
- (lbpp << 24) | (bank_h << 21) |
- (bank_w << 18) | (mt_aspect << 16);
- cs->buf[cs->cdw++] = (pitch_tile_max << 0) | ((height - 1) << 16);
- cs->buf[cs->cdw++] = (slice_tile_max << 0);
- cs->buf[cs->cdw++] = (x << 0) | (z << 18);
- cs->buf[cs->cdw++] = (y << 0) | (tile_split << 21) | (nbanks << 25) | (non_disp_tiling << 28);
- cs->buf[cs->cdw++] = addr & 0xfffffffc;
- cs->buf[cs->cdw++] = (addr >> 32UL) & 0xff;
+ RADEON_USAGE_WRITE, 0);
+ radeon_emit(cs, DMA_PACKET(DMA_PACKET_COPY, sub_cmd, size));
+ radeon_emit(cs, base >> 8);
+ radeon_emit(cs, (detile << 31) | (array_mode << 27) |
+ (lbpp << 24) | (bank_h << 21) |
+ (bank_w << 18) | (mt_aspect << 16));
+ radeon_emit(cs, (pitch_tile_max << 0) | ((height - 1) << 16));
+ radeon_emit(cs, (slice_tile_max << 0));
+ radeon_emit(cs, (x << 0) | (z << 18));
+ radeon_emit(cs, (y << 0) | (tile_split << 21) | (nbanks << 25) | (non_disp_tiling << 28));
+ radeon_emit(cs, addr & 0xfffffffc);
+ radeon_emit(cs, (addr >> 32UL) & 0xff);
copy_height -= cheight;
addr += cheight * pitch;
y += cheight;
}
- r600_dma_emit_wait_idle(&rctx->b);
}
static void evergreen_dma_copy(struct pipe_context *ctx,
goto fallback;
}
+ if (rctx->cmd_buf_is_compute) {
+ rctx->b.gfx.flush(rctx, PIPE_FLUSH_ASYNC, NULL);
+ rctx->cmd_buf_is_compute = false;
+ }
+
if (dst->target == PIPE_BUFFER && src->target == PIPE_BUFFER) {
evergreen_dma_copy_buffer(rctx, dst, src, dst_x, src_box->x, src_box->width);
return;
dst_y = util_format_get_nblocksy(src->format, dst_y);
bpp = rdst->surface.bpe;
- dst_pitch = rdst->surface.level[dst_level].pitch_bytes;
- src_pitch = rsrc->surface.level[src_level].pitch_bytes;
- src_w = rsrc->surface.level[src_level].npix_x;
- dst_w = rdst->surface.level[dst_level].npix_x;
+ dst_pitch = rdst->surface.u.legacy.level[dst_level].nblk_x * rdst->surface.bpe;
+ src_pitch = rsrc->surface.u.legacy.level[src_level].nblk_x * rsrc->surface.bpe;
+ src_w = u_minify(rsrc->resource.b.b.width0, src_level);
+ dst_w = u_minify(rdst->resource.b.b.width0, dst_level);
copy_height = src_box->height / rsrc->surface.blk_h;
- dst_mode = rdst->surface.level[dst_level].mode;
- src_mode = rsrc->surface.level[src_level].mode;
+ dst_mode = rdst->surface.u.legacy.level[dst_level].mode;
+ src_mode = rsrc->surface.u.legacy.level[src_level].mode;
if (src_pitch != dst_pitch || src_box->x || dst_x || src_w != dst_w) {
/* FIXME evergreen can do partial blit */
* dst_x/y == 0
* dst_pitch == src_pitch
*/
- src_offset= rsrc->surface.level[src_level].offset;
- src_offset += rsrc->surface.level[src_level].slice_size * src_box->z;
+ src_offset= rsrc->surface.u.legacy.level[src_level].offset;
+ src_offset += (uint64_t)rsrc->surface.u.legacy.level[src_level].slice_size_dw * 4 * src_box->z;
src_offset += src_y * src_pitch + src_x * bpp;
- dst_offset = rdst->surface.level[dst_level].offset;
- dst_offset += rdst->surface.level[dst_level].slice_size * dst_z;
+ dst_offset = rdst->surface.u.legacy.level[dst_level].offset;
+ dst_offset += (uint64_t)rdst->surface.u.legacy.level[dst_level].slice_size_dw * 4 * dst_z;
dst_offset += dst_y * dst_pitch + dst_x * bpp;
evergreen_dma_copy_buffer(rctx, dst, src, dst_offset, src_offset,
src_box->height * src_pitch);
memcpy(rctx->tess_state, default_outer_level, sizeof(float) * 4);
memcpy(rctx->tess_state+4, default_inner_level, sizeof(float) * 2);
- rctx->tess_state_dirty = true;
+ rctx->driver_consts[PIPE_SHADER_TESS_CTRL].tcs_default_levels_dirty = true;
+}
+
+static void evergreen_setup_immed_buffer(struct r600_context *rctx,
+ struct r600_image_view *rview,
+ enum pipe_format pformat)
+{
+ struct r600_screen *rscreen = (struct r600_screen *)rctx->b.b.screen;
+ uint32_t immed_size = rscreen->b.info.max_se * 256 * 64 * util_format_get_blocksize(pformat);
+ struct eg_buf_res_params buf_params;
+ bool skip_reloc = false;
+ struct r600_resource *resource = (struct r600_resource *)rview->base.resource;
+ if (!resource->immed_buffer) {
+ eg_resource_alloc_immed(&rscreen->b, resource, immed_size);
+ }
+
+ memset(&buf_params, 0, sizeof(buf_params));
+ buf_params.pipe_format = pformat;
+ buf_params.size = resource->immed_buffer->b.b.width0;
+ buf_params.swizzle[0] = PIPE_SWIZZLE_X;
+ buf_params.swizzle[1] = PIPE_SWIZZLE_Y;
+ buf_params.swizzle[2] = PIPE_SWIZZLE_Z;
+ buf_params.swizzle[3] = PIPE_SWIZZLE_W;
+ buf_params.uncached = 1;
+ evergreen_fill_buffer_resource_words(rctx, &resource->immed_buffer->b.b,
+ &buf_params, &skip_reloc,
+ rview->immed_resource_words);
+}
+
+static void evergreen_set_hw_atomic_buffers(struct pipe_context *ctx,
+ unsigned start_slot,
+ unsigned count,
+ const struct pipe_shader_buffer *buffers)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_atomic_buffer_state *astate;
+ unsigned i, idx;
+
+ astate = &rctx->atomic_buffer_state;
+
+ /* we'd probably like to expand this to 8 later so put the logic in */
+ for (i = start_slot, idx = 0; i < start_slot + count; i++, idx++) {
+ const struct pipe_shader_buffer *buf;
+ struct pipe_shader_buffer *abuf;
+
+ abuf = &astate->buffer[i];
+
+ if (!buffers || !buffers[idx].buffer) {
+ pipe_resource_reference(&abuf->buffer, NULL);
+ continue;
+ }
+ buf = &buffers[idx];
+
+ pipe_resource_reference(&abuf->buffer, buf->buffer);
+ abuf->buffer_offset = buf->buffer_offset;
+ abuf->buffer_size = buf->buffer_size;
+ }
+}
+
+static void evergreen_set_shader_buffers(struct pipe_context *ctx,
+ enum pipe_shader_type shader, unsigned start_slot,
+ unsigned count,
+ const struct pipe_shader_buffer *buffers,
+ unsigned writable_bitmask)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_image_state *istate = NULL;
+ struct r600_image_view *rview;
+ struct r600_tex_color_info color;
+ struct eg_buf_res_params buf_params;
+ struct r600_resource *resource;
+ unsigned i, idx;
+ unsigned old_mask;
+
+ if (shader != PIPE_SHADER_FRAGMENT &&
+ shader != PIPE_SHADER_COMPUTE && count == 0)
+ return;
+
+ if (shader == PIPE_SHADER_FRAGMENT)
+ istate = &rctx->fragment_buffers;
+ else if (shader == PIPE_SHADER_COMPUTE)
+ istate = &rctx->compute_buffers;
+
+ old_mask = istate->enabled_mask;
+ for (i = start_slot, idx = 0; i < start_slot + count; i++, idx++) {
+ const struct pipe_shader_buffer *buf;
+ unsigned res_type;
+
+ rview = &istate->views[i];
+
+ if (!buffers || !buffers[idx].buffer) {
+ pipe_resource_reference((struct pipe_resource **)&rview->base.resource, NULL);
+ istate->enabled_mask &= ~(1 << i);
+ continue;
+ }
+
+ buf = &buffers[idx];
+ pipe_resource_reference((struct pipe_resource **)&rview->base.resource, buf->buffer);
+
+ resource = (struct r600_resource *)rview->base.resource;
+
+ evergreen_setup_immed_buffer(rctx, rview, PIPE_FORMAT_R32_UINT);
+
+ color.offset = 0;
+ color.view = 0;
+ evergreen_set_color_surface_buffer(rctx, resource,
+ PIPE_FORMAT_R32_UINT,
+ buf->buffer_offset,
+ buf->buffer_offset + buf->buffer_size,
+ &color);
+
+ res_type = V_028C70_BUFFER;
+
+ rview->cb_color_base = color.offset;
+ rview->cb_color_dim = color.dim;
+ rview->cb_color_info = color.info |
+ S_028C70_RAT(1) |
+ S_028C70_RESOURCE_TYPE(res_type);
+ rview->cb_color_pitch = color.pitch;
+ rview->cb_color_slice = color.slice;
+ rview->cb_color_view = color.view;
+ rview->cb_color_attrib = color.attrib;
+ rview->cb_color_fmask = color.fmask;
+ rview->cb_color_fmask_slice = color.fmask_slice;
+
+ memset(&buf_params, 0, sizeof(buf_params));
+ buf_params.pipe_format = PIPE_FORMAT_R32_UINT;
+ buf_params.offset = buf->buffer_offset;
+ buf_params.size = buf->buffer_size;
+ buf_params.swizzle[0] = PIPE_SWIZZLE_X;
+ buf_params.swizzle[1] = PIPE_SWIZZLE_Y;
+ buf_params.swizzle[2] = PIPE_SWIZZLE_Z;
+ buf_params.swizzle[3] = PIPE_SWIZZLE_W;
+ buf_params.force_swizzle = true;
+ buf_params.uncached = 1;
+ buf_params.size_in_bytes = true;
+ evergreen_fill_buffer_resource_words(rctx, &resource->b.b,
+ &buf_params,
+ &rview->skip_mip_address_reloc,
+ rview->resource_words);
+
+ istate->enabled_mask |= (1 << i);
+ }
+
+ istate->atom.num_dw = util_bitcount(istate->enabled_mask) * 46;
+
+ if (old_mask != istate->enabled_mask)
+ r600_mark_atom_dirty(rctx, &rctx->framebuffer.atom);
+
+ /* construct the target mask */
+ if (rctx->cb_misc_state.buffer_rat_enabled_mask != istate->enabled_mask) {
+ rctx->cb_misc_state.buffer_rat_enabled_mask = istate->enabled_mask;
+ r600_mark_atom_dirty(rctx, &rctx->cb_misc_state.atom);
+ }
+
+ if (shader == PIPE_SHADER_FRAGMENT)
+ r600_mark_atom_dirty(rctx, &istate->atom);
+}
+
+static void evergreen_set_shader_images(struct pipe_context *ctx,
+ enum pipe_shader_type shader, unsigned start_slot,
+ unsigned count,
+ const struct pipe_image_view *images)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ unsigned i;
+ struct r600_image_view *rview;
+ struct pipe_resource *image;
+ struct r600_resource *resource;
+ struct r600_tex_color_info color;
+ struct eg_buf_res_params buf_params;
+ struct eg_tex_res_params tex_params;
+ unsigned old_mask;
+ struct r600_image_state *istate = NULL;
+ int idx;
+ if (shader != PIPE_SHADER_FRAGMENT && shader != PIPE_SHADER_COMPUTE && count == 0)
+ return;
+
+ if (shader == PIPE_SHADER_FRAGMENT)
+ istate = &rctx->fragment_images;
+ else if (shader == PIPE_SHADER_COMPUTE)
+ istate = &rctx->compute_images;
+
+ assert (shader == PIPE_SHADER_FRAGMENT || shader == PIPE_SHADER_COMPUTE);
+
+ old_mask = istate->enabled_mask;
+ for (i = start_slot, idx = 0; i < start_slot + count; i++, idx++) {
+ unsigned res_type;
+ const struct pipe_image_view *iview;
+ rview = &istate->views[i];
+
+ if (!images || !images[idx].resource) {
+ pipe_resource_reference((struct pipe_resource **)&rview->base.resource, NULL);
+ istate->enabled_mask &= ~(1 << i);
+ istate->compressed_colortex_mask &= ~(1 << i);
+ istate->compressed_depthtex_mask &= ~(1 << i);
+ continue;
+ }
+
+ iview = &images[idx];
+ image = iview->resource;
+ resource = (struct r600_resource *)image;
+
+ r600_context_add_resource_size(ctx, image);
+
+ rview->base = *iview;
+ rview->base.resource = NULL;
+ pipe_resource_reference((struct pipe_resource **)&rview->base.resource, image);
+
+ evergreen_setup_immed_buffer(rctx, rview, iview->format);
+
+ bool is_buffer = image->target == PIPE_BUFFER;
+ struct r600_texture *rtex = (struct r600_texture *)image;
+ if (!is_buffer & rtex->db_compatible)
+ istate->compressed_depthtex_mask |= 1 << i;
+ else
+ istate->compressed_depthtex_mask &= ~(1 << i);
+
+ if (!is_buffer && rtex->cmask.size)
+ istate->compressed_colortex_mask |= 1 << i;
+ else
+ istate->compressed_colortex_mask &= ~(1 << i);
+ if (!is_buffer) {
+
+ evergreen_set_color_surface_common(rctx, rtex,
+ iview->u.tex.level,
+ iview->u.tex.first_layer,
+ iview->u.tex.last_layer,
+ iview->format,
+ &color);
+ color.dim = S_028C78_WIDTH_MAX(u_minify(image->width0, iview->u.tex.level) - 1) |
+ S_028C78_HEIGHT_MAX(u_minify(image->height0, iview->u.tex.level) - 1);
+ } else {
+ color.offset = 0;
+ color.view = 0;
+ evergreen_set_color_surface_buffer(rctx, resource,
+ iview->format,
+ iview->u.buf.offset,
+ iview->u.buf.size,
+ &color);
+ }
+
+ switch (image->target) {
+ case PIPE_BUFFER:
+ res_type = V_028C70_BUFFER;
+ break;
+ case PIPE_TEXTURE_1D:
+ res_type = V_028C70_TEXTURE1D;
+ break;
+ case PIPE_TEXTURE_1D_ARRAY:
+ res_type = V_028C70_TEXTURE1DARRAY;
+ break;
+ case PIPE_TEXTURE_2D:
+ case PIPE_TEXTURE_RECT:
+ res_type = V_028C70_TEXTURE2D;
+ break;
+ case PIPE_TEXTURE_3D:
+ res_type = V_028C70_TEXTURE3D;
+ break;
+ case PIPE_TEXTURE_2D_ARRAY:
+ case PIPE_TEXTURE_CUBE:
+ case PIPE_TEXTURE_CUBE_ARRAY:
+ res_type = V_028C70_TEXTURE2DARRAY;
+ break;
+ default:
+ assert(0);
+ res_type = 0;
+ break;
+ }
+
+ rview->cb_color_base = color.offset;
+ rview->cb_color_dim = color.dim;
+ rview->cb_color_info = color.info |
+ S_028C70_RAT(1) |
+ S_028C70_RESOURCE_TYPE(res_type);
+ rview->cb_color_pitch = color.pitch;
+ rview->cb_color_slice = color.slice;
+ rview->cb_color_view = color.view;
+ rview->cb_color_attrib = color.attrib;
+ rview->cb_color_fmask = color.fmask;
+ rview->cb_color_fmask_slice = color.fmask_slice;
+
+ if (image->target != PIPE_BUFFER) {
+ memset(&tex_params, 0, sizeof(tex_params));
+ tex_params.pipe_format = iview->format;
+ tex_params.force_level = 0;
+ tex_params.width0 = image->width0;
+ tex_params.height0 = image->height0;
+ tex_params.first_level = iview->u.tex.level;
+ tex_params.last_level = iview->u.tex.level;
+ tex_params.first_layer = iview->u.tex.first_layer;
+ tex_params.last_layer = iview->u.tex.last_layer;
+ tex_params.target = image->target;
+ tex_params.swizzle[0] = PIPE_SWIZZLE_X;
+ tex_params.swizzle[1] = PIPE_SWIZZLE_Y;
+ tex_params.swizzle[2] = PIPE_SWIZZLE_Z;
+ tex_params.swizzle[3] = PIPE_SWIZZLE_W;
+ evergreen_fill_tex_resource_words(rctx, &resource->b.b, &tex_params,
+ &rview->skip_mip_address_reloc,
+ rview->resource_words);
+
+ } else {
+ memset(&buf_params, 0, sizeof(buf_params));
+ buf_params.pipe_format = iview->format;
+ buf_params.size = iview->u.buf.size;
+ buf_params.offset = iview->u.buf.offset;
+ buf_params.swizzle[0] = PIPE_SWIZZLE_X;
+ buf_params.swizzle[1] = PIPE_SWIZZLE_Y;
+ buf_params.swizzle[2] = PIPE_SWIZZLE_Z;
+ buf_params.swizzle[3] = PIPE_SWIZZLE_W;
+ evergreen_fill_buffer_resource_words(rctx, &resource->b.b,
+ &buf_params,
+ &rview->skip_mip_address_reloc,
+ rview->resource_words);
+ }
+ istate->enabled_mask |= (1 << i);
+ }
+
+ istate->atom.num_dw = util_bitcount(istate->enabled_mask) * 46;
+ istate->dirty_buffer_constants = TRUE;
+ rctx->b.flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
+ rctx->b.flags |= R600_CONTEXT_FLUSH_AND_INV_CB |
+ R600_CONTEXT_FLUSH_AND_INV_CB_META;
+
+ if (old_mask != istate->enabled_mask)
+ r600_mark_atom_dirty(rctx, &rctx->framebuffer.atom);
+
+ if (rctx->cb_misc_state.image_rat_enabled_mask != istate->enabled_mask) {
+ rctx->cb_misc_state.image_rat_enabled_mask = istate->enabled_mask;
+ r600_mark_atom_dirty(rctx, &rctx->cb_misc_state.atom);
+ }
+
+ if (shader == PIPE_SHADER_FRAGMENT)
+ r600_mark_atom_dirty(rctx, &istate->atom);
+}
+
+static void evergreen_get_pipe_constant_buffer(struct r600_context *rctx,
+ enum pipe_shader_type shader, uint slot,
+ struct pipe_constant_buffer *cbuf)
+{
+ struct r600_constbuf_state *state = &rctx->constbuf_state[shader];
+ struct pipe_constant_buffer *cb;
+ cbuf->user_buffer = NULL;
+
+ cb = &state->cb[slot];
+
+ cbuf->buffer_size = cb->buffer_size;
+ pipe_resource_reference(&cbuf->buffer, cb->buffer);
+}
+
+static void evergreen_get_shader_buffers(struct r600_context *rctx,
+ enum pipe_shader_type shader,
+ uint start_slot, uint count,
+ struct pipe_shader_buffer *sbuf)
+{
+ assert(shader == PIPE_SHADER_COMPUTE);
+ int idx, i;
+ struct r600_image_state *istate = &rctx->compute_buffers;
+ struct r600_image_view *rview;
+
+ for (i = start_slot, idx = 0; i < start_slot + count; i++, idx++) {
+
+ rview = &istate->views[i];
+
+ pipe_resource_reference(&sbuf[idx].buffer, rview->base.resource);
+ if (rview->base.resource) {
+ uint64_t rview_va = ((struct r600_resource *)rview->base.resource)->gpu_address;
+
+ uint64_t prog_va = rview->resource_words[0];
+
+ prog_va += ((uint64_t)G_030008_BASE_ADDRESS_HI(rview->resource_words[2])) << 32;
+ prog_va -= rview_va;
+
+ sbuf[idx].buffer_offset = prog_va & 0xffffffff;
+ sbuf[idx].buffer_size = rview->resource_words[1] + 1;;
+ } else {
+ sbuf[idx].buffer_offset = 0;
+ sbuf[idx].buffer_size = 0;
+ }
+ }
+}
+
+static void evergreen_save_qbo_state(struct pipe_context *ctx, struct r600_qbo_state *st)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ st->saved_compute = rctx->cs_shader_state.shader;
+
+ /* save constant buffer 0 */
+ evergreen_get_pipe_constant_buffer(rctx, PIPE_SHADER_COMPUTE, 0, &st->saved_const0);
+ /* save ssbo 0 */
+ evergreen_get_shader_buffers(rctx, PIPE_SHADER_COMPUTE, 0, 3, st->saved_ssbo);
}
+
void evergreen_init_state_functions(struct r600_context *rctx)
{
unsigned id = 1;
rctx->config_state.dyn_gpr_enabled = true;
}
r600_init_atom(rctx, &rctx->framebuffer.atom, id++, evergreen_emit_framebuffer_state, 0);
+ r600_init_atom(rctx, &rctx->fragment_images.atom, id++, evergreen_emit_fragment_image_state, 0);
+ r600_init_atom(rctx, &rctx->compute_images.atom, id++, evergreen_emit_compute_image_state, 0);
+ r600_init_atom(rctx, &rctx->fragment_buffers.atom, id++, evergreen_emit_fragment_buffer_state, 0);
+ r600_init_atom(rctx, &rctx->compute_buffers.atom, id++, evergreen_emit_compute_buffer_state, 0);
/* shader const */
r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_VERTEX].atom, id++, evergreen_emit_vs_constant_buffers, 0);
r600_init_atom(rctx, &rctx->constbuf_state[PIPE_SHADER_GEOMETRY].atom, id++, evergreen_emit_gs_constant_buffers, 0);
r600_init_atom(rctx, &rctx->db_misc_state.atom, id++, evergreen_emit_db_misc_state, 10);
r600_init_atom(rctx, &rctx->db_state.atom, id++, evergreen_emit_db_state, 14);
r600_init_atom(rctx, &rctx->dsa_state.atom, id++, r600_emit_cso_state, 0);
- r600_init_atom(rctx, &rctx->poly_offset_state.atom, id++, evergreen_emit_polygon_offset, 6);
+ r600_init_atom(rctx, &rctx->poly_offset_state.atom, id++, evergreen_emit_polygon_offset, 9);
r600_init_atom(rctx, &rctx->rasterizer_state.atom, id++, r600_emit_cso_state, 0);
r600_add_atom(rctx, &rctx->b.scissors.atom, id++);
r600_add_atom(rctx, &rctx->b.viewports.atom, id++);
rctx->b.b.set_polygon_stipple = evergreen_set_polygon_stipple;
rctx->b.b.set_min_samples = evergreen_set_min_samples;
rctx->b.b.set_tess_state = evergreen_set_tess_state;
+ rctx->b.b.set_hw_atomic_buffers = evergreen_set_hw_atomic_buffers;
+ rctx->b.b.set_shader_images = evergreen_set_shader_images;
+ rctx->b.b.set_shader_buffers = evergreen_set_shader_buffers;
if (rctx->b.chip_class == EVERGREEN)
rctx->b.b.get_sample_position = evergreen_get_sample_position;
else
rctx->b.b.get_sample_position = cayman_get_sample_position;
rctx->b.dma_copy = evergreen_dma_copy;
+ rctx->b.save_qbo_state = evergreen_save_qbo_state;
evergreen_init_compute_state_functions(rctx);
}
unsigned input_vertex_size, output_vertex_size;
unsigned input_patch_size, pervertex_output_patch_size, output_patch_size;
unsigned output_patch0_offset, perpatch_output_offset, lds_size;
- uint32_t values[16];
+ uint32_t values[8];
unsigned num_waves;
unsigned num_pipes = rctx->screen->b.info.r600_max_quad_pipes;
unsigned wave_divisor = (16 * num_pipes);
if (rctx->lds_alloc != 0 &&
rctx->last_ls == ls &&
- !rctx->tess_state_dirty &&
rctx->last_num_tcs_input_cp == num_tcs_input_cp &&
rctx->last_tcs == tcs)
return;
rctx->lds_alloc = (lds_size | (num_waves << 14));
- memcpy(&values[8], rctx->tess_state, 6 * sizeof(float));
- values[14] = 0;
- values[15] = 0;
-
- rctx->tess_state_dirty = false;
rctx->last_ls = ls;
rctx->last_tcs = tcs;
rctx->last_num_tcs_input_cp = num_tcs_input_cp;
constbuf.user_buffer = values;
- constbuf.buffer_size = 16 * 4;
+ constbuf.buffer_size = 8 * 4;
rctx->b.b.set_constant_buffer(&rctx->b.b, PIPE_SHADER_VERTEX,
R600_LDS_INFO_CONST_BUFFER, &constbuf);
}
void evergreen_set_ls_hs_config(struct r600_context *rctx,
- struct radeon_winsys_cs *cs,
+ struct radeon_cmdbuf *cs,
uint32_t ls_hs_config)
{
radeon_set_context_reg(cs, R_028B58_VGT_LS_HS_CONFIG, ls_hs_config);
}
void evergreen_set_lds_alloc(struct r600_context *rctx,
- struct radeon_winsys_cs *cs,
+ struct radeon_cmdbuf *cs,
uint32_t lds_alloc)
{
radeon_set_context_reg(cs, R_0288E8_SQ_LDS_ALLOC, lds_alloc);
}
return true;
}
+
+#define AC_ENCODE_TRACE_POINT(id) (0xcafe0000 | ((id) & 0xffff))
+
+void eg_trace_emit(struct r600_context *rctx)
+{
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ unsigned reloc;
+
+ if (rctx->b.chip_class < EVERGREEN)
+ return;
+
+ /* This must be done after r600_need_cs_space. */
+ reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
+ (struct r600_resource*)rctx->trace_buf, RADEON_USAGE_WRITE,
+ RADEON_PRIO_CP_DMA);
+
+ rctx->trace_id++;
+ radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx, rctx->trace_buf,
+ RADEON_USAGE_READWRITE, RADEON_PRIO_TRACE);
+ radeon_emit(cs, PKT3(PKT3_MEM_WRITE, 3, 0));
+ radeon_emit(cs, rctx->trace_buf->gpu_address);
+ radeon_emit(cs, rctx->trace_buf->gpu_address >> 32 | MEM_WRITE_32_BITS | MEM_WRITE_CONFIRM);
+ radeon_emit(cs, rctx->trace_id);
+ radeon_emit(cs, 0);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, AC_ENCODE_TRACE_POINT(rctx->trace_id));
+}
+
+static void evergreen_emit_set_append_cnt(struct r600_context *rctx,
+ struct r600_shader_atomic *atomic,
+ struct r600_resource *resource,
+ uint32_t pkt_flags)
+{
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ unsigned reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
+ resource,
+ RADEON_USAGE_READ,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+ uint64_t dst_offset = resource->gpu_address + (atomic->start * 4);
+ uint32_t base_reg_0 = R_02872C_GDS_APPEND_COUNT_0;
+
+ uint32_t reg_val = (base_reg_0 + atomic->hw_idx * 4 - EVERGREEN_CONTEXT_REG_OFFSET) >> 2;
+
+ radeon_emit(cs, PKT3(PKT3_SET_APPEND_CNT, 2, 0) | pkt_flags);
+ radeon_emit(cs, (reg_val << 16) | 0x3);
+ radeon_emit(cs, dst_offset & 0xfffffffc);
+ radeon_emit(cs, (dst_offset >> 32) & 0xff);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+}
+
+static void evergreen_emit_event_write_eos(struct r600_context *rctx,
+ struct r600_shader_atomic *atomic,
+ struct r600_resource *resource,
+ uint32_t pkt_flags)
+{
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ uint32_t event = EVENT_TYPE_PS_DONE;
+ uint32_t base_reg_0 = R_02872C_GDS_APPEND_COUNT_0;
+ uint32_t reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
+ resource,
+ RADEON_USAGE_WRITE,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+ uint64_t dst_offset = resource->gpu_address + (atomic->start * 4);
+ uint32_t reg_val = (base_reg_0 + atomic->hw_idx * 4) >> 2;
+
+ if (pkt_flags == RADEON_CP_PACKET3_COMPUTE_MODE)
+ event = EVENT_TYPE_CS_DONE;
+
+ radeon_emit(cs, PKT3(PKT3_EVENT_WRITE_EOS, 3, 0) | pkt_flags);
+ radeon_emit(cs, EVENT_TYPE(event) | EVENT_INDEX(6));
+ radeon_emit(cs, (dst_offset) & 0xffffffff);
+ radeon_emit(cs, (0 << 29) | ((dst_offset >> 32) & 0xff));
+ radeon_emit(cs, reg_val);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+}
+
+static void cayman_emit_event_write_eos(struct r600_context *rctx,
+ struct r600_shader_atomic *atomic,
+ struct r600_resource *resource,
+ uint32_t pkt_flags)
+{
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ uint32_t event = EVENT_TYPE_PS_DONE;
+ uint32_t reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
+ resource,
+ RADEON_USAGE_WRITE,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+ uint64_t dst_offset = resource->gpu_address + (atomic->start * 4);
+
+ if (pkt_flags == RADEON_CP_PACKET3_COMPUTE_MODE)
+ event = EVENT_TYPE_CS_DONE;
+
+ radeon_emit(cs, PKT3(PKT3_EVENT_WRITE_EOS, 3, 0) | pkt_flags);
+ radeon_emit(cs, EVENT_TYPE(event) | EVENT_INDEX(6));
+ radeon_emit(cs, (dst_offset) & 0xffffffff);
+ radeon_emit(cs, (1 << 29) | ((dst_offset >> 32) & 0xff));
+ radeon_emit(cs, (atomic->hw_idx) | (1 << 16));
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+}
+
+/* writes count from a buffer into GDS */
+static void cayman_write_count_to_gds(struct r600_context *rctx,
+ struct r600_shader_atomic *atomic,
+ struct r600_resource *resource,
+ uint32_t pkt_flags)
+{
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ unsigned reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
+ resource,
+ RADEON_USAGE_READ,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+ uint64_t dst_offset = resource->gpu_address + (atomic->start * 4);
+
+ radeon_emit(cs, PKT3(PKT3_CP_DMA, 4, 0) | pkt_flags);
+ radeon_emit(cs, dst_offset & 0xffffffff);
+ radeon_emit(cs, PKT3_CP_DMA_CP_SYNC | PKT3_CP_DMA_DST_SEL(1) | ((dst_offset >> 32) & 0xff));// GDS
+ radeon_emit(cs, atomic->hw_idx * 4);
+ radeon_emit(cs, 0);
+ radeon_emit(cs, PKT3_CP_DMA_CMD_DAS | 4);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+}
+
+void evergreen_emit_atomic_buffer_setup_count(struct r600_context *rctx,
+ struct r600_pipe_shader *cs_shader,
+ struct r600_shader_atomic *combined_atomics,
+ uint8_t *atomic_used_mask_p)
+{
+ uint8_t atomic_used_mask = 0;
+ int i, j, k;
+ bool is_compute = cs_shader ? true : false;
+
+ for (i = 0; i < (is_compute ? 1 : EG_NUM_HW_STAGES); i++) {
+ uint8_t num_atomic_stage;
+ struct r600_pipe_shader *pshader;
+
+ if (is_compute)
+ pshader = cs_shader;
+ else
+ pshader = rctx->hw_shader_stages[i].shader;
+ if (!pshader)
+ continue;
+
+ num_atomic_stage = pshader->shader.nhwatomic_ranges;
+ if (!num_atomic_stage)
+ continue;
+
+ for (j = 0; j < num_atomic_stage; j++) {
+ struct r600_shader_atomic *atomic = &pshader->shader.atomics[j];
+ int natomics = atomic->end - atomic->start + 1;
+
+ for (k = 0; k < natomics; k++) {
+ /* seen this in a previous stage */
+ if (atomic_used_mask & (1u << (atomic->hw_idx + k)))
+ continue;
+
+ combined_atomics[atomic->hw_idx + k].hw_idx = atomic->hw_idx + k;
+ combined_atomics[atomic->hw_idx + k].buffer_id = atomic->buffer_id;
+ combined_atomics[atomic->hw_idx + k].start = atomic->start + k;
+ combined_atomics[atomic->hw_idx + k].end = combined_atomics[atomic->hw_idx + k].start + 1;
+ atomic_used_mask |= (1u << (atomic->hw_idx + k));
+ }
+ }
+ }
+ *atomic_used_mask_p = atomic_used_mask;
+}
+
+void evergreen_emit_atomic_buffer_setup(struct r600_context *rctx,
+ bool is_compute,
+ struct r600_shader_atomic *combined_atomics,
+ uint8_t atomic_used_mask)
+{
+ struct r600_atomic_buffer_state *astate = &rctx->atomic_buffer_state;
+ unsigned pkt_flags = 0;
+ uint32_t mask;
+
+ if (is_compute)
+ pkt_flags = RADEON_CP_PACKET3_COMPUTE_MODE;
+
+ mask = atomic_used_mask;
+ if (!mask)
+ return;
+
+ while (mask) {
+ unsigned atomic_index = u_bit_scan(&mask);
+ struct r600_shader_atomic *atomic = &combined_atomics[atomic_index];
+ struct r600_resource *resource = r600_resource(astate->buffer[atomic->buffer_id].buffer);
+ assert(resource);
+
+ if (rctx->b.chip_class == CAYMAN)
+ cayman_write_count_to_gds(rctx, atomic, resource, pkt_flags);
+ else
+ evergreen_emit_set_append_cnt(rctx, atomic, resource, pkt_flags);
+ }
+}
+
+void evergreen_emit_atomic_buffer_save(struct r600_context *rctx,
+ bool is_compute,
+ struct r600_shader_atomic *combined_atomics,
+ uint8_t *atomic_used_mask_p)
+{
+ struct radeon_cmdbuf *cs = rctx->b.gfx.cs;
+ struct r600_atomic_buffer_state *astate = &rctx->atomic_buffer_state;
+ uint32_t pkt_flags = 0;
+ uint32_t event = EVENT_TYPE_PS_DONE;
+ uint32_t mask;
+ uint64_t dst_offset;
+ unsigned reloc;
+
+ if (is_compute)
+ pkt_flags = RADEON_CP_PACKET3_COMPUTE_MODE;
+
+ mask = *atomic_used_mask_p;
+ if (!mask)
+ return;
+
+ while (mask) {
+ unsigned atomic_index = u_bit_scan(&mask);
+ struct r600_shader_atomic *atomic = &combined_atomics[atomic_index];
+ struct r600_resource *resource = r600_resource(astate->buffer[atomic->buffer_id].buffer);
+ assert(resource);
+
+ if (rctx->b.chip_class == CAYMAN)
+ cayman_emit_event_write_eos(rctx, atomic, resource, pkt_flags);
+ else
+ evergreen_emit_event_write_eos(rctx, atomic, resource, pkt_flags);
+ }
+
+ if (pkt_flags == RADEON_CP_PACKET3_COMPUTE_MODE)
+ event = EVENT_TYPE_CS_DONE;
+
+ ++rctx->append_fence_id;
+ reloc = radeon_add_to_buffer_list(&rctx->b, &rctx->b.gfx,
+ r600_resource(rctx->append_fence),
+ RADEON_USAGE_READWRITE,
+ RADEON_PRIO_SHADER_RW_BUFFER);
+ dst_offset = r600_resource(rctx->append_fence)->gpu_address;
+ radeon_emit(cs, PKT3(PKT3_EVENT_WRITE_EOS, 3, 0) | pkt_flags);
+ radeon_emit(cs, EVENT_TYPE(event) | EVENT_INDEX(6));
+ radeon_emit(cs, dst_offset & 0xffffffff);
+ radeon_emit(cs, (2 << 29) | ((dst_offset >> 32) & 0xff));
+ radeon_emit(cs, rctx->append_fence_id);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+
+ radeon_emit(cs, PKT3(PKT3_WAIT_REG_MEM, 5, 0) | pkt_flags);
+ radeon_emit(cs, WAIT_REG_MEM_GEQUAL | WAIT_REG_MEM_MEMORY | (1 << 8));
+ radeon_emit(cs, dst_offset & 0xffffffff);
+ radeon_emit(cs, ((dst_offset >> 32) & 0xff));
+ radeon_emit(cs, rctx->append_fence_id);
+ radeon_emit(cs, 0xffffffff);
+ radeon_emit(cs, 0xa);
+ radeon_emit(cs, PKT3(PKT3_NOP, 0, 0));
+ radeon_emit(cs, reloc);
+}