#include "util/u_framebuffer.h"
#include "util/u_dual_blend.h"
#include "evergreen_compute.h"
+#include "util/u_math.h"
+
+static INLINE unsigned evergreen_array_mode(unsigned mode)
+{
+ switch (mode) {
+ case RADEON_SURF_MODE_LINEAR_ALIGNED: return V_028C70_ARRAY_LINEAR_ALIGNED;
+ break;
+ case RADEON_SURF_MODE_1D: return V_028C70_ARRAY_1D_TILED_THIN1;
+ break;
+ case RADEON_SURF_MODE_2D: return V_028C70_ARRAY_2D_TILED_THIN1;
+ default:
+ case RADEON_SURF_MODE_LINEAR: return V_028C70_ARRAY_LINEAR_GENERAL;
+ }
+}
static uint32_t eg_num_banks(uint32_t nbanks)
{
case PIPE_TEXTURE_3D:
return V_030000_SQ_TEX_DIM_3D;
case PIPE_TEXTURE_CUBE:
+ case PIPE_TEXTURE_CUBE_ARRAY:
return V_030000_SQ_TEX_DIM_CUBEMAP;
}
}
return V_028040_Z_16;
case PIPE_FORMAT_Z24X8_UNORM:
case PIPE_FORMAT_Z24_UNORM_S8_UINT:
+ case PIPE_FORMAT_X8Z24_UNORM:
+ case PIPE_FORMAT_S8_UINT_Z24_UNORM:
return V_028040_Z_24;
case PIPE_FORMAT_Z32_FLOAT:
case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
case PIPE_FORMAT_L32A32_UINT:
case PIPE_FORMAT_L32A32_SINT:
case PIPE_FORMAT_L32A32_FLOAT:
+ case PIPE_FORMAT_R8A8_UNORM:
+ case PIPE_FORMAT_R8A8_SNORM:
+ case PIPE_FORMAT_R8A8_UINT:
+ case PIPE_FORMAT_R8A8_SINT:
+ case PIPE_FORMAT_R16A16_UNORM:
+ case PIPE_FORMAT_R16A16_SNORM:
+ case PIPE_FORMAT_R16A16_UINT:
+ case PIPE_FORMAT_R16A16_SINT:
+ case PIPE_FORMAT_R16A16_FLOAT:
+ case PIPE_FORMAT_R32A32_UINT:
+ case PIPE_FORMAT_R32A32_SINT:
+ case PIPE_FORMAT_R32A32_FLOAT:
return V_028C70_SWAP_ALT;
case PIPE_FORMAT_R8G8_UNORM:
case PIPE_FORMAT_R8G8_SNORM:
case PIPE_FORMAT_R8G8B8A8_SINT:
case PIPE_FORMAT_R8G8B8A8_UINT:
case PIPE_FORMAT_R8G8B8X8_UNORM:
+ case PIPE_FORMAT_R8G8B8X8_SNORM:
+ case PIPE_FORMAT_R8G8B8X8_SRGB:
+ case PIPE_FORMAT_R8G8B8X8_UINT:
+ case PIPE_FORMAT_R8G8B8X8_SINT:
return V_028C70_SWAP_STD;
case PIPE_FORMAT_A8B8G8R8_UNORM:
case PIPE_FORMAT_X8Z24_UNORM:
case PIPE_FORMAT_S8_UINT_Z24_UNORM:
- return V_028C70_SWAP_STD;
+ return V_028C70_SWAP_STD_REV;
case PIPE_FORMAT_R10G10B10A2_UNORM:
case PIPE_FORMAT_R10G10B10X2_SNORM:
case PIPE_FORMAT_B10G10R10A2_UNORM:
case PIPE_FORMAT_B10G10R10A2_UINT:
+ case PIPE_FORMAT_B10G10R10X2_UNORM:
return V_028C70_SWAP_ALT;
case PIPE_FORMAT_R11G11B10_FLOAT:
case PIPE_FORMAT_R16G16B16A16_UINT:
case PIPE_FORMAT_R16G16B16A16_SINT:
case PIPE_FORMAT_R16G16B16A16_FLOAT:
+ case PIPE_FORMAT_R16G16B16X16_UNORM:
+ case PIPE_FORMAT_R16G16B16X16_SNORM:
+ case PIPE_FORMAT_R16G16B16X16_FLOAT:
+ case PIPE_FORMAT_R16G16B16X16_UINT:
+ case PIPE_FORMAT_R16G16B16X16_SINT:
case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
/* 128-bit buffers. */
case PIPE_FORMAT_R32G32B32A32_UNORM:
case PIPE_FORMAT_R32G32B32A32_SINT:
case PIPE_FORMAT_R32G32B32A32_UINT:
+ case PIPE_FORMAT_R32G32B32X32_FLOAT:
+ case PIPE_FORMAT_R32G32B32X32_UINT:
+ case PIPE_FORMAT_R32G32B32X32_SINT:
return V_028C70_SWAP_STD;
default:
R600_ERR("unsupported colorswap format %d\n", format);
case PIPE_FORMAT_R8G8_SNORM:
case PIPE_FORMAT_R8G8_UINT:
case PIPE_FORMAT_R8G8_SINT:
+ case PIPE_FORMAT_R8A8_UNORM:
+ case PIPE_FORMAT_R8A8_SNORM:
+ case PIPE_FORMAT_R8A8_UINT:
+ case PIPE_FORMAT_R8A8_SINT:
return V_028C70_COLOR_8_8;
case PIPE_FORMAT_R16_UNORM:
case PIPE_FORMAT_R8G8B8A8_SNORM:
case PIPE_FORMAT_R8G8B8A8_UNORM:
case PIPE_FORMAT_R8G8B8X8_UNORM:
+ case PIPE_FORMAT_R8G8B8X8_SNORM:
+ case PIPE_FORMAT_R8G8B8X8_SRGB:
+ case PIPE_FORMAT_R8G8B8X8_UINT:
+ case PIPE_FORMAT_R8G8B8X8_SINT:
case PIPE_FORMAT_R8SG8SB8UX8U_NORM:
case PIPE_FORMAT_X8B8G8R8_UNORM:
case PIPE_FORMAT_X8R8G8B8_UNORM:
case PIPE_FORMAT_R10G10B10X2_SNORM:
case PIPE_FORMAT_B10G10R10A2_UNORM:
case PIPE_FORMAT_B10G10R10A2_UINT:
+ case PIPE_FORMAT_B10G10R10X2_UNORM:
case PIPE_FORMAT_R10SG10SB10SA2U_NORM:
return V_028C70_COLOR_2_10_10_10;
case PIPE_FORMAT_R16G16_FLOAT:
case PIPE_FORMAT_L16A16_FLOAT:
+ case PIPE_FORMAT_R16A16_FLOAT:
return V_028C70_COLOR_16_16_FLOAT;
case PIPE_FORMAT_R16G16_UNORM:
case PIPE_FORMAT_L16A16_SNORM:
case PIPE_FORMAT_L16A16_UINT:
case PIPE_FORMAT_L16A16_SINT:
+ case PIPE_FORMAT_R16A16_UNORM:
+ case PIPE_FORMAT_R16A16_SNORM:
+ case PIPE_FORMAT_R16A16_UINT:
+ case PIPE_FORMAT_R16A16_SINT:
return V_028C70_COLOR_16_16;
case PIPE_FORMAT_R11G11B10_FLOAT:
case PIPE_FORMAT_R16G16B16A16_SINT:
case PIPE_FORMAT_R16G16B16A16_UNORM:
case PIPE_FORMAT_R16G16B16A16_SNORM:
+ case PIPE_FORMAT_R16G16B16X16_UNORM:
+ case PIPE_FORMAT_R16G16B16X16_SNORM:
+ case PIPE_FORMAT_R16G16B16X16_UINT:
+ case PIPE_FORMAT_R16G16B16X16_SINT:
return V_028C70_COLOR_16_16_16_16;
case PIPE_FORMAT_R16G16B16A16_FLOAT:
+ case PIPE_FORMAT_R16G16B16X16_FLOAT:
return V_028C70_COLOR_16_16_16_16_FLOAT;
case PIPE_FORMAT_R32G32_FLOAT:
case PIPE_FORMAT_L32A32_FLOAT:
+ case PIPE_FORMAT_R32A32_FLOAT:
return V_028C70_COLOR_32_32_FLOAT;
case PIPE_FORMAT_R32G32_SINT:
case PIPE_FORMAT_R32G32B32A32_UNORM:
case PIPE_FORMAT_R32G32B32A32_SINT:
case PIPE_FORMAT_R32G32B32A32_UINT:
+ case PIPE_FORMAT_R32G32B32X32_UINT:
+ case PIPE_FORMAT_R32G32B32X32_SINT:
return V_028C70_COLOR_32_32_32_32;
case PIPE_FORMAT_R32G32B32A32_FLOAT:
+ case PIPE_FORMAT_R32G32B32X32_FLOAT:
return V_028C70_COLOR_32_32_32_32_FLOAT;
/* YUV buffers. */
return ss;
}
+static struct pipe_sampler_view *
+texture_buffer_sampler_view(struct r600_pipe_sampler_view *view,
+ unsigned width0, unsigned height0)
+
+{
+ struct pipe_context *ctx = view->base.context;
+ struct r600_texture *tmp = (struct r600_texture*)view->base.texture;
+ uint64_t va;
+ int stride = util_format_get_blocksize(view->base.format);
+ unsigned format, num_format, format_comp, endian;
+ unsigned swizzle_res;
+ unsigned char swizzle[4];
+ const struct util_format_description *desc;
+
+ 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,
+ &format, &num_format, &format_comp,
+ &endian);
+
+ desc = util_format_description(view->base.format);
+
+ swizzle_res = r600_get_swizzle_combined(desc->swizzle, swizzle, TRUE);
+
+ va = r600_resource_va(ctx->screen, view->base.texture);
+ view->tex_resource = &tmp->resource;
+
+ view->skip_mip_address_reloc = true;
+ view->tex_resource_words[0] = va;
+ view->tex_resource_words[1] = width0 - 1;
+ view->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_SRF_MODE_ALL(1) |
+ S_030008_ENDIAN_SWAP(endian);
+ view->tex_resource_words[3] = swizzle_res;
+ /*
+ * 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
+ * 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);
+ return &view->base;
+}
+
struct pipe_sampler_view *
evergreen_create_sampler_view_custom(struct pipe_context *ctx,
struct pipe_resource *texture,
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, tile_type = 0;
+ unsigned char swizzle[4], array_mode = 0, non_disp_tiling = 0;
unsigned height, depth, width;
unsigned macro_aspect, tile_split, bankh, bankw, nbanks;
+ enum pipe_format pipe_format = state->format;
+ struct radeon_surface_level *surflevel;
if (view == NULL)
return NULL;
view->base.reference.count = 1;
view->base.context = ctx;
+ if (texture->target == PIPE_BUFFER)
+ return texture_buffer_sampler_view(view, width0, height0);
+
swizzle[0] = state->swizzle_r;
swizzle[1] = state->swizzle_g;
swizzle[2] = state->swizzle_b;
swizzle[3] = state->swizzle_a;
- format = r600_translate_texformat(ctx->screen, state->format,
+ tile_split = tmp->surface.tile_split;
+ surflevel = tmp->surface.level;
+
+ /* Texturing with separate depth and stencil. */
+ if (tmp->is_depth && !tmp->is_flushing_texture) {
+ switch (pipe_format) {
+ case PIPE_FORMAT_Z32_FLOAT_S8X24_UINT:
+ 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;
+ 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;
+ break;
+ default:;
+ }
+ }
+
+ format = r600_translate_texformat(ctx->screen, pipe_format,
swizzle,
&word4, &yuv_format);
assert(format != ~0);
return NULL;
}
- if (tmp->is_depth && !tmp->is_flushing_texture) {
- if (!r600_init_flushed_depth_texture(ctx, texture, NULL)) {
- FREE(view);
- return NULL;
- }
- tmp = tmp->flushed_depth_texture;
- }
-
endian = r600_colorformat_endian_swap(format);
width = width0;
height = height0;
- depth = tmp->surface.level[0].npix_z;
- pitch = tmp->surface.level[0].nblk_x * util_format_get_blockwidth(state->format);
- tile_type = tmp->tile_type;
+ depth = texture->depth0;
+ pitch = surflevel[0].nblk_x * util_format_get_blockwidth(pipe_format);
+ non_disp_tiling = tmp->non_disp_tiling;
- switch (tmp->surface.level[0].mode) {
+ switch (surflevel[0].mode) {
case RADEON_SURF_MODE_LINEAR_ALIGNED:
array_mode = V_028C70_ARRAY_LINEAR_ALIGNED;
break;
array_mode = V_028C70_ARRAY_LINEAR_GENERAL;
break;
}
- tile_split = tmp->surface.tile_split;
macro_aspect = tmp->surface.mtilea;
bankw = tmp->surface.bankw;
bankh = tmp->surface.bankh;
/* 128 bit formats require tile type = 1 */
if (rscreen->chip_class == CAYMAN) {
- if (util_format_get_blocksize(state->format) >= 16)
- tile_type = 1;
+ if (util_format_get_blocksize(pipe_format) >= 16)
+ non_disp_tiling = 1;
}
nbanks = eg_num_banks(rscreen->tiling_info.num_banks);
depth = texture->array_size;
} else if (texture->target == PIPE_TEXTURE_2D_ARRAY) {
depth = texture->array_size;
- }
+ } else if (texture->target == PIPE_TEXTURE_CUBE_ARRAY)
+ depth = texture->array_size / 6;
view->tex_resource = &tmp->resource;
view->tex_resource_words[0] = (S_030000_DIM(r600_tex_dim(texture->target, texture->nr_samples)) |
S_030000_PITCH((pitch / 8) - 1) |
S_030000_TEX_WIDTH(width - 1));
if (rscreen->chip_class == CAYMAN)
- view->tex_resource_words[0] |= CM_S_030000_NON_DISP_TILING_ORDER(tile_type);
+ view->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(tile_type);
+ 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) |
S_030004_TEX_DEPTH(depth - 1) |
S_030004_ARRAY_MODE(array_mode));
- view->tex_resource_words[2] = (tmp->surface.level[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ view->tex_resource_words[2] = (surflevel[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
/* TEX_RESOURCE_WORD3.MIP_ADDRESS */
if (texture->nr_samples > 1 && rscreen->msaa_texture_support == MSAA_TEXTURE_COMPRESSED) {
view->tex_resource_words[3] = (tmp->fmask_offset + r600_resource_va(ctx->screen, texture)) >> 8;
}
} else if (state->u.tex.last_level && texture->nr_samples <= 1) {
- view->tex_resource_words[3] = (tmp->surface.level[1].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ view->tex_resource_words[3] = (surflevel[1].offset + r600_resource_va(ctx->screen, texture)) >> 8;
} else {
- view->tex_resource_words[3] = (tmp->surface.level[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
+ view->tex_resource_words[3] = (surflevel[0].offset + r600_resource_va(ctx->screen, texture)) >> 8;
}
view->tex_resource_words[4] = (word4 |
S_03001C_BANK_WIDTH(bankw) |
S_03001C_BANK_HEIGHT(bankh) |
S_03001C_MACRO_TILE_ASPECT(macro_aspect) |
- S_03001C_NUM_BANKS(nbanks);
+ S_03001C_NUM_BANKS(nbanks) |
+ S_03001C_DEPTH_SAMPLE_ORDER(tmp->is_depth && !tmp->is_flushing_texture);
return &view->base;
}
static void evergreen_emit_clip_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
struct pipe_clip_state *state = &rctx->clip_state.state;
r600_write_context_reg_seq(cs, R_0285BC_PA_CL_UCP0_X, 6*4);
static void evergreen_emit_scissor_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
struct pipe_scissor_state *state = &rctx->scissor.scissor;
uint32_t tl, br;
unsigned color_info, color_attrib, color_dim = 0;
unsigned format, swap, ntype, endian;
uint64_t offset, base_offset;
- unsigned tile_type, macro_aspect, tile_split, bankh, bankw, fmask_bankh, nbanks;
+ unsigned non_disp_tiling, macro_aspect, tile_split, bankh, bankw, fmask_bankh, nbanks;
const struct util_format_description *desc;
int i;
bool blend_clamp = 0, blend_bypass = 0;
- if (rtex->is_depth && !rtex->is_flushing_texture) {
- r600_init_flushed_depth_texture(&rctx->context, pipe_tex, NULL);
- rtex = rtex->flushed_depth_texture;
- assert(rtex);
- }
-
offset = rtex->surface.level[level].offset;
if (rtex->surface.level[level].mode < RADEON_SURF_MODE_1D) {
offset += rtex->surface.level[level].slice_size *
switch (rtex->surface.level[level].mode) {
case RADEON_SURF_MODE_LINEAR_ALIGNED:
color_info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_LINEAR_ALIGNED);
- tile_type = 1;
+ non_disp_tiling = 1;
break;
case RADEON_SURF_MODE_1D:
color_info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_1D_TILED_THIN1);
- tile_type = rtex->tile_type;
+ 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);
- tile_type = rtex->tile_type;
+ non_disp_tiling = rtex->non_disp_tiling;
break;
case RADEON_SURF_MODE_LINEAR:
default:
color_info = S_028C70_ARRAY_MODE(V_028C70_ARRAY_LINEAR_GENERAL);
- tile_type = 1;
+ non_disp_tiling = 1;
break;
}
tile_split = rtex->surface.tile_split;
/* 128 bit formats require tile type = 1 */
if (rscreen->chip_class == CAYMAN) {
if (util_format_get_blocksize(surf->base.format) >= 16)
- tile_type = 1;
+ non_disp_tiling = 1;
}
nbanks = eg_num_banks(rscreen->tiling_info.num_banks);
desc = util_format_description(surf->base.format);
S_028C74_BANK_WIDTH(bankw) |
S_028C74_BANK_HEIGHT(bankh) |
S_028C74_MACRO_TILE_ASPECT(macro_aspect) |
- S_028C74_NON_DISP_TILING_ORDER(tile_type) |
+ S_028C74_NON_DISP_TILING_ORDER(non_disp_tiling) |
S_028C74_FMASK_BANK_HEIGHT(fmask_bankh);
if (rctx->chip_class == CAYMAN && rtex->resource.b.b.nr_samples > 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;
}
if (rtex->surface.flags & RADEON_SURF_SBUFFER) {
- uint64_t stencil_offset = rtex->surface.stencil_offset;
- unsigned i, stile_split = rtex->surface.stencil_tile_split;
+ uint64_t stencil_offset;
+ unsigned stile_split = rtex->surface.stencil_tile_split;
stile_split = eg_tile_split(stile_split);
+
+ stencil_offset = rtex->surface.stencil_level[level].offset;
stencil_offset += r600_resource_va(screen, surf->base.texture);
- stencil_offset += rtex->surface.level[level].offset / 4;
- stencil_offset >>= 8;
-
- /* We're guessing the stencil offset from the depth offset.
- * Make sure each mipmap level has a unique offset. */
- for (i = 1; i <= level; i++) {
- /* If two levels have the same address, add 256
- * to the offset of the smaller level. */
- if ((rtex->surface.level[i-1].offset / 4) >> 8 ==
- (rtex->surface.level[i].offset / 4) >> 8) {
- stencil_offset++;
- }
- }
- surf->db_stencil_base = stencil_offset;
+ surf->db_stencil_base = stencil_offset >> 8;
surf->db_stencil_info = S_028044_FORMAT(V_028044_STENCIL_8) |
S_028044_TILE_SPLIT(stile_split);
} else {
S_028044_FORMAT(V_028044_STENCIL_8);
}
+ surf->htile_enabled = 0;
+ /* use htile only for first level */
+ if (rtex->htile && !level) {
+ uint64_t va = r600_resource_va(&rctx->screen->screen, &rtex->htile->b.b);
+ surf->htile_enabled = 1;
+ surf->db_htile_data_base = va >> 8;
+ surf->db_htile_surface = S_028ABC_HTILE_WIDTH(1) |
+ S_028ABC_HTILE_HEIGHT(1) |
+ S_028ABC_LINEAR(1);
+ surf->db_depth_info |= S_028040_TILE_SURFACE_ENABLE(1);
+ surf->db_preload_control = 0;
+ }
+
surf->depth_initialized = true;
}
uint32_t i, log_samples;
if (rctx->framebuffer.state.nr_cbufs) {
- rctx->flags |= R600_CONTEXT_CB_FLUSH;
+ rctx->flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
if (rctx->framebuffer.state.cbufs[0]->texture->nr_samples > 1) {
rctx->flags |= R600_CONTEXT_FLUSH_AND_INV_CB_META;
}
}
if (rctx->framebuffer.state.zsbuf) {
- rctx->flags |= R600_CONTEXT_DB_FLUSH;
+ rctx->flags |= R600_CONTEXT_WAIT_3D_IDLE | R600_CONTEXT_FLUSH_AND_INV;
}
util_copy_framebuffer_state(&rctx->framebuffer.state, state);
surf = (struct r600_surface*)state->cbufs[i];
rtex = (struct r600_texture*)surf->base.texture;
+ r600_context_add_resource_size(ctx, state->cbufs[i]->texture);
+
if (!surf->color_initialized) {
evergreen_init_color_surface(rctx, surf);
}
if (state->zsbuf) {
surf = (struct r600_surface*)state->zsbuf;
+ r600_context_add_resource_size(ctx, state->zsbuf->texture);
+
if (!surf->depth_initialized) {
evergreen_init_depth_surface(rctx, surf);
}
rctx->poly_offset_state.zs_format = state->zsbuf->format;
rctx->poly_offset_state.atom.dirty = true;
}
+
+ if (rctx->db_state.rsurf != surf) {
+ rctx->db_state.rsurf = surf;
+ rctx->db_state.atom.dirty = true;
+ rctx->db_misc_state.atom.dirty = true;
+ }
+ } else if (rctx->db_state.rsurf) {
+ rctx->db_state.rsurf = NULL;
+ rctx->db_state.atom.dirty = true;
+ rctx->db_misc_state.atom.dirty = true;
}
if (rctx->cb_misc_state.nr_cbufs != state->nr_cbufs) {
};
static unsigned max_dist_8x = 7;
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
unsigned max_dist = 0;
switch (nr_samples) {
};
static unsigned max_dist_16x = 8;
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
unsigned max_dist = 0;
switch (nr_samples) {
static void evergreen_emit_framebuffer_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
struct pipe_framebuffer_state *state = &rctx->framebuffer.state;
unsigned nr_cbufs = state->nr_cbufs;
unsigned i, tl, br;
/* Colorbuffers. */
for (i = 0; i < nr_cbufs; i++) {
struct r600_surface *cb = (struct r600_surface*)state->cbufs[i];
- unsigned reloc = r600_context_bo_reloc(rctx, (struct r600_resource*)cb->base.texture,
+ unsigned reloc = r600_context_bo_reloc(rctx,
+ &rctx->rings.gfx,
+ (struct r600_resource*)cb->base.texture,
RADEON_USAGE_READWRITE);
r600_write_context_reg_seq(cs, R_028C60_CB_COLOR0_BASE + i * 0x3C, 11);
((struct r600_surface*)state->cbufs[0])->cb_color_info);
if (!rctx->keep_tiling_flags) {
- unsigned reloc = r600_context_bo_reloc(rctx, (struct r600_resource*)state->cbufs[0]->texture,
+ unsigned reloc = r600_context_bo_reloc(rctx,
+ &rctx->rings.gfx,
+ (struct r600_resource*)state->cbufs[0]->texture,
RADEON_USAGE_READWRITE);
r600_write_value(cs, PKT3(PKT3_NOP, 0, 0)); /* R_028C70_CB_COLOR0_INFO */
/* ZS buffer. */
if (state->zsbuf) {
struct r600_surface *zb = (struct r600_surface*)state->zsbuf;
- unsigned reloc = r600_context_bo_reloc(rctx, (struct r600_resource*)state->zsbuf->texture,
+ unsigned reloc = r600_context_bo_reloc(rctx,
+ &rctx->rings.gfx,
+ (struct r600_resource*)state->zsbuf->texture,
RADEON_USAGE_READWRITE);
r600_write_context_reg(cs, R_028B78_PA_SU_POLY_OFFSET_DB_FMT_CNTL,
static void evergreen_emit_polygon_offset(struct r600_context *rctx, struct r600_atom *a)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.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;
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:
static void evergreen_emit_cb_misc_state(struct r600_context *rctx, struct r600_atom *atom)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.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;
r600_write_value(cs, 0xf | (a->dual_src_blend ? ps_colormask : 0) | fb_colormask); /* R_02823C_CB_SHADER_MASK */
}
+static void evergreen_emit_db_state(struct r600_context *rctx, struct r600_atom *atom)
+{
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
+ struct r600_db_state *a = (struct r600_db_state*)atom;
+
+ if (a->rsurf && a->rsurf->htile_enabled) {
+ struct r600_texture *rtex = (struct r600_texture *)a->rsurf->base.texture;
+ unsigned reloc_idx;
+
+ r600_write_context_reg(cs, R_02802C_DB_DEPTH_CLEAR, fui(rtex->depth_clear));
+ r600_write_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, a->rsurf->db_htile_surface);
+ r600_write_context_reg(cs, R_028AC8_DB_PRELOAD_CONTROL, a->rsurf->db_preload_control);
+ r600_write_context_reg(cs, R_028014_DB_HTILE_DATA_BASE, a->rsurf->db_htile_data_base);
+ reloc_idx = r600_context_bo_reloc(rctx, &rctx->rings.gfx, rtex->htile, RADEON_USAGE_READWRITE);
+ cs->buf[cs->cdw++] = PKT3(PKT3_NOP, 0, 0);
+ cs->buf[cs->cdw++] = reloc_idx;
+ } else {
+ r600_write_context_reg(cs, R_028ABC_DB_HTILE_SURFACE, 0);
+ r600_write_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->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.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 db_render_override =
- S_02800C_FORCE_HIZ_ENABLE(V_02800C_FORCE_DISABLE) |
S_02800C_FORCE_HIS_ENABLE0(V_02800C_FORCE_DISABLE) |
S_02800C_FORCE_HIS_ENABLE1(V_02800C_FORCE_DISABLE);
}
db_render_override |= S_02800C_NOOP_CULL_DISABLE(1);
}
-
+ if (rctx->db_state.rsurf && rctx->db_state.rsurf->htile_enabled) {
+ /* FORCE_OFF means HiZ/HiS are determined by DB_SHADER_CONTROL */
+ db_render_override |= S_02800C_FORCE_HIZ_ENABLE(V_02800C_FORCE_OFF);
+ /* This is to fix a lockup when hyperz and alpha test are enabled at
+ * the same time somehow GPU get confuse on which order to pick for
+ * z test
+ */
+ if (rctx->alphatest_state.sx_alpha_test_control) {
+ db_render_override |= S_02800C_FORCE_SHADER_Z_ORDER(1);
+ }
+ } else {
+ db_render_override |= S_02800C_FORCE_HIZ_ENABLE(V_02800C_FORCE_DISABLE);
+ }
if (a->flush_depthstencil_through_cb) {
assert(a->copy_depth || a->copy_stencil);
S_028000_STENCIL_COPY_ENABLE(a->copy_stencil) |
S_028000_COPY_CENTROID(1) |
S_028000_COPY_SAMPLE(a->copy_sample);
+ } else if (a->flush_depthstencil_in_place) {
+ db_render_control |= S_028000_DEPTH_COMPRESS_DISABLE(1) |
+ S_028000_STENCIL_COMPRESS_DISABLE(1);
+ db_render_override |= S_02800C_DISABLE_PIXEL_RATE_TILES(1);
+ }
+ if (a->htile_clear) {
+ /* FIXME we might want to disable cliprect here */
+ db_render_control |= S_028000_DEPTH_CLEAR_ENABLE(1);
}
r600_write_context_reg_seq(cs, R_028000_DB_RENDER_CONTROL, 2);
unsigned resource_offset,
unsigned pkt_flags)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
r600_write_value(cs, 0xc0000000); /* RESOURCEi_WORD7 */
r600_write_value(cs, PKT3(PKT3_NOP, 0, 0) | pkt_flags);
- r600_write_value(cs, r600_context_bo_reloc(rctx, rbuffer, RADEON_USAGE_READ));
+ r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, rbuffer, RADEON_USAGE_READ));
}
state->dirty_mask = 0;
}
unsigned reg_alu_constbuf_size,
unsigned reg_alu_const_cache)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
r600_write_context_reg(cs, reg_alu_const_cache + buffer_index * 4, va >> 8);
r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, r600_context_bo_reloc(rctx, rbuffer, RADEON_USAGE_READ));
+ r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, rbuffer, RADEON_USAGE_READ));
r600_write_value(cs, PKT3(PKT3_SET_RESOURCE, 8, 0));
r600_write_value(cs, (buffer_id_base + buffer_index) * 8);
r600_write_value(cs, 0xc0000000); /* RESOURCEi_WORD7 */
r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, r600_context_bo_reloc(rctx, rbuffer, RADEON_USAGE_READ));
+ r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, rbuffer, RADEON_USAGE_READ));
dirty_mask &= ~(1 << buffer_index);
}
struct r600_samplerview_state *state,
unsigned resource_id_base)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
uint32_t dirty_mask = state->dirty_mask;
while (dirty_mask) {
r600_write_value(cs, (resource_id_base + resource_index) * 8);
r600_write_array(cs, 8, rview->tex_resource_words);
- reloc = r600_context_bo_reloc(rctx, rview->tex_resource,
+ reloc = r600_context_bo_reloc(rctx, &rctx->rings.gfx, rview->tex_resource,
RADEON_USAGE_READ);
r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
r600_write_value(cs, reloc);
unsigned resource_id_base,
unsigned border_index_reg)
{
- struct radeon_winsys_cs *cs = rctx->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
uint32_t dirty_mask = texinfo->states.dirty_mask;
while (dirty_mask) {
struct r600_sample_mask *s = (struct r600_sample_mask*)a;
uint8_t mask = s->sample_mask;
- r600_write_context_reg(rctx->cs, R_028C3C_PA_SC_AA_MASK,
+ r600_write_context_reg(rctx->rings.gfx.cs, R_028C3C_PA_SC_AA_MASK,
mask | (mask << 8) | (mask << 16) | (mask << 24));
}
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->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
uint16_t mask = s->sample_mask;
r600_write_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->cs;
+ struct radeon_winsys_cs *cs = rctx->rings.gfx.cs;
struct r600_cso_state *state = (struct r600_cso_state*)a;
- struct r600_resource *shader = (struct r600_resource*)state->cso;
+ struct r600_fetch_shader *shader = (struct r600_fetch_shader*)state->cso;
r600_write_context_reg(cs, R_0288A4_SQ_PGM_START_FS,
- r600_resource_va(rctx->context.screen, &shader->b.b) >> 8);
+ (r600_resource_va(rctx->context.screen, &shader->buffer->b.b) + shader->offset) >> 8);
r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
- r600_write_value(cs, r600_context_bo_reloc(rctx, shader, RADEON_USAGE_READ));
+ r600_write_value(cs, r600_context_bo_reloc(rctx, &rctx->rings.gfx, shader->buffer, RADEON_USAGE_READ));
}
void evergreen_init_state_functions(struct r600_context *rctx)
r600_init_atom(rctx, &rctx->clip_misc_state.atom, id++, r600_emit_clip_misc_state, 6);
r600_init_atom(rctx, &rctx->clip_state.atom, id++, evergreen_emit_clip_state, 26);
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->rasterizer_state.atom, id++, r600_emit_cso_state, 0);
r600_store_value(cb, 0); /* CM_R_0288E8_SQ_LDS_ALLOC */
r600_store_value(cb, 0); /* R_0288EC_SQ_LDS_ALLOC_PS */
- r600_store_context_reg_seq(cb, R_028380_SQ_VTX_SEMANTIC_0, 34);
- r600_store_value(cb, 0); /* R_028380_SQ_VTX_SEMANTIC_0 */
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0); /* R_0283FC_SQ_VTX_SEMANTIC_31 */
+ r600_store_context_reg(cb, R_0288F0_SQ_VTX_SEMANTIC_CLEAR, ~0);
+
+ r600_store_context_reg_seq(cb, R_028400_VGT_MAX_VTX_INDX, 2);
r600_store_value(cb, ~0); /* R_028400_VGT_MAX_VTX_INDX */
r600_store_value(cb, 0); /* R_028404_VGT_MIN_VTX_INDX */
r600_store_ctl_const(cb, R_03CFF0_SQ_VTX_BASE_VTX_LOC, 0);
- r600_store_context_reg_seq(cb, R_028028_DB_STENCIL_CLEAR, 2);
- r600_store_value(cb, 0); /* R_028028_DB_STENCIL_CLEAR */
- r600_store_value(cb, 0x3F800000); /* R_02802C_DB_DEPTH_CLEAR */
+ r600_store_context_reg(cb, R_028028_DB_STENCIL_CLEAR, 0);
r600_store_context_reg(cb, R_0286DC_SPI_FOG_CNTL, 0);
r600_store_config_reg(cb, R_008A14_PA_CL_ENHANCE, (3 << 1) | 1);
- r600_store_context_reg_seq(cb, R_028380_SQ_VTX_SEMANTIC_0, 34);
- r600_store_value(cb, 0); /* R_028380_SQ_VTX_SEMANTIC_0 */
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0);
- r600_store_value(cb, 0); /* R_0283FC_SQ_VTX_SEMANTIC_31 */
+ r600_store_context_reg(cb, R_0288F0_SQ_VTX_SEMANTIC_CLEAR, ~0);
+
+ r600_store_context_reg_seq(cb, R_028400_VGT_MAX_VTX_INDX, 2);
r600_store_value(cb, ~0); /* R_028400_VGT_MAX_VTX_INDX */
r600_store_value(cb, 0); /* R_028404_VGT_MIN_VTX_INDX */
r600_store_ctl_const(cb, R_03CFF0_SQ_VTX_BASE_VTX_LOC, 0);
- r600_store_context_reg_seq(cb, R_028028_DB_STENCIL_CLEAR, 2);
- r600_store_value(cb, 0); /* R_028028_DB_STENCIL_CLEAR */
- r600_store_value(cb, 0x3F800000); /* R_02802C_DB_DEPTH_CLEAR */
+ r600_store_context_reg(cb, R_028028_DB_STENCIL_CLEAR, 0);
r600_store_context_reg(cb, R_028200_PA_SC_WINDOW_OFFSET, 0);
r600_store_context_reg(cb, R_02820C_PA_SC_CLIPRECT_RULE, 0xFFFF);
struct r600_context *rctx = (struct r600_context *)ctx;
struct r600_pipe_state *rstate = &shader->rstate;
struct r600_shader *rshader = &shader->shader;
- unsigned i, exports_ps, num_cout, spi_ps_in_control_0, spi_input_z, spi_ps_in_control_1, db_shader_control;
+ unsigned i, exports_ps, num_cout, spi_ps_in_control_0, spi_input_z, spi_ps_in_control_1, db_shader_control = 0;
int pos_index = -1, face_index = -1;
int ninterp = 0;
boolean have_linear = FALSE, have_centroid = FALSE, have_perspective = FALSE;
rstate->nregs = 0;
- db_shader_control = S_02880C_Z_ORDER(V_02880C_EARLY_Z_THEN_LATE_Z);
for (i = 0; i < rshader->ninput; i++) {
/* evergreen NUM_INTERP only contains values interpolated into the LDS,
POSITION goes via GPRs from the SC so isn't counted */
V_02880C_EXPORT_DB_FULL) |
S_02880C_ALPHA_TO_MASK_DISABLE(rctx->framebuffer.cb0_is_integer);
+ /* When alpha test is enabled we can't trust the hw to make the proper
+ * decision on the order in which ztest should be run related to fragment
+ * shader execution.
+ *
+ * If alpha test is enabled perform early z rejection (RE_Z) but don't early
+ * write to the zbuffer. Write to zbuffer is delayed after fragment shader
+ * execution and thus after alpha test so if discarded by the alpha test
+ * the z value is not written.
+ */
+ if (rctx->alphatest_state.sx_alpha_test_control) {
+ db_shader_control |= S_02880C_Z_ORDER(V_02880C_RE_Z);
+ } else {
+ db_shader_control |= S_02880C_Z_ORDER(V_02880C_EARLY_Z_THEN_LATE_Z);
+ }
+
if (db_shader_control != rctx->db_misc_state.db_shader_control) {
rctx->db_misc_state.db_shader_control = db_shader_control;
rctx->db_misc_state.atom.dirty = true;
}
}
+
+static void evergreen_dma_copy_tile(struct r600_context *rctx,
+ struct pipe_resource *dst,
+ unsigned dst_level,
+ unsigned dst_x,
+ unsigned dst_y,
+ unsigned dst_z,
+ struct pipe_resource *src,
+ unsigned src_level,
+ unsigned src_x,
+ unsigned src_y,
+ unsigned src_z,
+ unsigned copy_height,
+ unsigned pitch,
+ unsigned bpp)
+{
+ struct radeon_winsys_cs *cs = rctx->rings.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 ncopy, height, cheight, detile, i, x, y, z, src_mode, dst_mode;
+ unsigned sub_cmd, bank_h, bank_w, mt_aspect, nbanks, tile_split;
+ uint64_t base, addr;
+
+ /* make sure that the dma ring is only one active */
+ rctx->rings.gfx.flush(rctx, RADEON_FLUSH_ASYNC);
+
+ dst_mode = rdst->surface.level[dst_level].mode;
+ src_mode = rsrc->surface.level[src_level].mode;
+ /* downcast linear aligned to linear to simplify test */
+ src_mode = src_mode == RADEON_SURF_MODE_LINEAR_ALIGNED ? RADEON_SURF_MODE_LINEAR : src_mode;
+ dst_mode = dst_mode == RADEON_SURF_MODE_LINEAR_ALIGNED ? RADEON_SURF_MODE_LINEAR : dst_mode;
+ assert(dst_mode != src_mode);
+
+ y = 0;
+ sub_cmd = 0x8;
+ lbpp = util_logbase2(bpp);
+ pitch_tile_max = ((pitch / bpp) >> 3) - 1;
+ nbanks = eg_num_banks(rctx->screen->tiling_info.num_banks);
+
+ if (dst_mode == RADEON_SURF_MODE_LINEAR) {
+ /* 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) >> 6;
+ 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;
+ 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;
+ 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);
+ base += r600_resource_va(&rctx->screen->screen, src);
+ addr += r600_resource_va(&rctx->screen->screen, dst);
+ } 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) >> 6;
+ 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;
+ 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;
+ 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);
+ base += r600_resource_va(&rctx->screen->screen, dst);
+ addr += r600_resource_va(&rctx->screen->screen, src);
+ }
+
+ size = (copy_height * pitch) >> 2;
+ ncopy = (size / 0x000fffff) + !!(size % 0x000fffff);
+ r600_need_dma_space(rctx, ncopy * 9);
+
+ for (i = 0; i < ncopy; i++) {
+ cheight = copy_height;
+ if (((cheight * pitch) >> 2) > 0x000fffff) {
+ cheight = (0x000fffff << 2) / pitch;
+ }
+ size = (cheight * pitch) >> 2;
+ /* emit reloc before writting cs so that cs is always in consistent state */
+ r600_context_bo_reloc(rctx, &rctx->rings.dma, &rsrc->resource, RADEON_USAGE_READ);
+ r600_context_bo_reloc(rctx, &rctx->rings.dma, &rdst->resource, RADEON_USAGE_WRITE);
+ 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);
+ cs->buf[cs->cdw++] = addr & 0xfffffffc;
+ cs->buf[cs->cdw++] = (addr >> 32UL) & 0xff;
+ copy_height -= cheight;
+ addr += cheight * pitch;
+ y += cheight;
+ }
+}
+
+boolean evergreen_dma_blit(struct pipe_context *ctx,
+ struct pipe_resource *dst,
+ unsigned dst_level,
+ unsigned dst_x, unsigned dst_y, unsigned dst_z,
+ struct pipe_resource *src,
+ unsigned src_level,
+ const struct pipe_box *src_box)
+{
+ struct r600_context *rctx = (struct r600_context *)ctx;
+ struct r600_texture *rsrc = (struct r600_texture*)src;
+ struct r600_texture *rdst = (struct r600_texture*)dst;
+ unsigned dst_pitch, src_pitch, bpp, dst_mode, src_mode, copy_height;
+ unsigned src_w, dst_w;
+
+ if (rctx->rings.dma.cs == NULL) {
+ return FALSE;
+ }
+ if (src->format != dst->format) {
+ return FALSE;
+ }
+
+ 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;
+ 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;
+ /* downcast linear aligned to linear to simplify test */
+ src_mode = src_mode == RADEON_SURF_MODE_LINEAR_ALIGNED ? RADEON_SURF_MODE_LINEAR : src_mode;
+ dst_mode = dst_mode == RADEON_SURF_MODE_LINEAR_ALIGNED ? RADEON_SURF_MODE_LINEAR : dst_mode;
+
+ if (src_pitch != dst_pitch || src_box->x || dst_x || src_w != dst_w) {
+ /* FIXME evergreen can do partial blit */
+ return FALSE;
+ }
+ /* the x test here are currently useless (because we don't support partial blit)
+ * but keep them around so we don't forget about those
+ */
+ if ((src_pitch & 0x7) || (src_box->x & 0x7) || (dst_x & 0x7) || (src_box->y & 0x7) || (dst_y & 0x7)) {
+ return FALSE;
+ }
+
+ if (src_mode == dst_mode) {
+ uint64_t dst_offset, src_offset;
+ /* simple dma blit would do NOTE code here assume :
+ * src_box.x/y == 0
+ * 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 += src_box->y * src_pitch + src_box->x * bpp;
+ dst_offset = rdst->surface.level[dst_level].offset;
+ dst_offset += rdst->surface.level[dst_level].slice_size * dst_z;
+ dst_offset += dst_y * dst_pitch + dst_x * bpp;
+ evergreen_dma_copy(rctx, dst, src, dst_offset, src_offset,
+ src_box->height * src_pitch);
+ } else {
+ evergreen_dma_copy_tile(rctx, dst, dst_level, dst_x, dst_y, dst_z,
+ src, src_level, src_box->x, src_box->y, src_box->z,
+ copy_height, dst_pitch, bpp);
+ }
+ return TRUE;
+}
projects / mesa.git / blobdiff