case PIPE_FORMAT_R16G16_FLOAT:
case PIPE_FORMAT_R16G16_UINT:
case PIPE_FORMAT_R16G16_SINT:
+ case PIPE_FORMAT_R16G16B16_FLOAT:
+ case PIPE_FORMAT_R32G32B32_FLOAT:
case PIPE_FORMAT_R32_UINT:
case PIPE_FORMAT_R32_SINT:
case PIPE_FORMAT_R32_FLOAT:
float offset_units = rctx->rasterizer->offset_units;
unsigned offset_db_fmt_cntl = 0, depth;
- switch (rctx->framebuffer.zsbuf->texture->format) {
+ switch (rctx->framebuffer.zsbuf->format) {
case PIPE_FORMAT_Z24X8_UNORM:
case PIPE_FORMAT_Z24_UNORM_S8_UINT:
depth = -24;
target_mask |= (state->rt[0].colormask << (4 * i));
}
}
+
+ if (target_mask)
+ color_control |= S_028808_SPECIAL_OP(V_028808_NORMAL);
+ else
+ color_control |= S_028808_SPECIAL_OP(V_028808_DISABLE);
+
blend->cb_target_mask = target_mask;
blend->cb_color_control = color_control;
/* only MRT0 has dual src blend */
{
struct r600_screen *rscreen = (struct r600_screen*)ctx->screen;
struct r600_pipe_sampler_view *view = CALLOC_STRUCT(r600_pipe_sampler_view);
- struct r600_pipe_resource_state *rstate;
struct r600_resource_texture *tmp = (struct r600_resource_texture*)texture;
unsigned format, endian;
uint32_t word4 = 0, yuv_format = 0, pitch = 0;
if (view == NULL)
return NULL;
- rstate = &view->state;
/* initialize base object */
view->base = *state;
format = r600_translate_texformat(ctx->screen, state->format,
swizzle,
&word4, &yuv_format);
+ assert(format != ~0);
if (format == ~0) {
- format = 0;
+ FREE(view);
+ return NULL;
}
if (tmp->is_depth && !tmp->is_flushing_texture) {
- r600_texture_depth_flush(ctx, texture, TRUE);
+ if (!r600_init_flushed_depth_texture(ctx, texture, NULL)) {
+ FREE(view);
+ return NULL;
+ }
tmp = tmp->flushed_depth_texture;
}
depth = texture->array_size;
}
- rstate->bo[0] = &tmp->resource;
- rstate->bo[1] = &tmp->resource;
- rstate->bo_usage[0] = RADEON_USAGE_READ;
- rstate->bo_usage[1] = RADEON_USAGE_READ;
-
- rstate->val[0] = (S_038000_DIM(r600_tex_dim(texture->target)) |
- S_038000_TILE_MODE(array_mode) |
- S_038000_TILE_TYPE(tile_type) |
- S_038000_PITCH((pitch / 8) - 1) |
- S_038000_TEX_WIDTH(width - 1));
- rstate->val[1] = (S_038004_TEX_HEIGHT(height - 1) |
- S_038004_TEX_DEPTH(depth - 1) |
- S_038004_DATA_FORMAT(format));
- rstate->val[2] = tmp->offset[offset_level] >> 8;
- rstate->val[3] = tmp->offset[offset_level+1] >> 8;
- rstate->val[4] = (word4 |
- S_038010_SRF_MODE_ALL(V_038010_SRF_MODE_ZERO_CLAMP_MINUS_ONE) |
- S_038010_REQUEST_SIZE(1) |
- S_038010_ENDIAN_SWAP(endian) |
- S_038010_BASE_LEVEL(0));
- rstate->val[5] = (S_038014_LAST_LEVEL(last_level) |
- S_038014_BASE_ARRAY(state->u.tex.first_layer) |
- S_038014_LAST_ARRAY(state->u.tex.last_layer));
- rstate->val[6] = (S_038018_TYPE(V_038010_SQ_TEX_VTX_VALID_TEXTURE) |
- S_038018_MAX_ANISO(4 /* max 16 samples */));
+ view->tex_resource = &tmp->resource;
+ view->tex_resource_words[0] = (S_038000_DIM(r600_tex_dim(texture->target)) |
+ S_038000_TILE_MODE(array_mode) |
+ S_038000_TILE_TYPE(tile_type) |
+ S_038000_PITCH((pitch / 8) - 1) |
+ S_038000_TEX_WIDTH(width - 1));
+ view->tex_resource_words[1] = (S_038004_TEX_HEIGHT(height - 1) |
+ S_038004_TEX_DEPTH(depth - 1) |
+ S_038004_DATA_FORMAT(format));
+ view->tex_resource_words[2] = tmp->offset[offset_level] >> 8;
+ view->tex_resource_words[3] = tmp->offset[offset_level+1] >> 8;
+ view->tex_resource_words[4] = (word4 |
+ S_038010_SRF_MODE_ALL(V_038010_SRF_MODE_ZERO_CLAMP_MINUS_ONE) |
+ S_038010_REQUEST_SIZE(1) |
+ S_038010_ENDIAN_SWAP(endian) |
+ S_038010_BASE_LEVEL(0));
+ view->tex_resource_words[5] = (S_038014_LAST_LEVEL(last_level) |
+ S_038014_BASE_ARRAY(state->u.tex.first_layer) |
+ S_038014_LAST_ARRAY(state->u.tex.last_layer));
+ view->tex_resource_words[6] = (S_038018_TYPE(V_038010_SQ_TEX_VTX_VALID_TEXTURE) |
+ S_038018_MAX_ANISO(4 /* max 16 samples */));
} else {
width = tmp->surface.level[offset_level].npix_x;
height = tmp->surface.level[offset_level].npix_y;
break;
}
- rstate->bo[0] = &tmp->resource;
- rstate->bo[1] = &tmp->resource;
- rstate->bo_usage[0] = RADEON_USAGE_READ;
- rstate->bo_usage[1] = RADEON_USAGE_READ;
-
- rstate->val[0] = (S_038000_DIM(r600_tex_dim(texture->target)) |
- S_038000_TILE_MODE(array_mode) |
- S_038000_TILE_TYPE(tile_type) |
- S_038000_PITCH((pitch / 8) - 1) |
- S_038000_TEX_WIDTH(width - 1));
- rstate->val[1] = (S_038004_TEX_HEIGHT(height - 1) |
- S_038004_TEX_DEPTH(depth - 1) |
- S_038004_DATA_FORMAT(format));
- rstate->val[2] = tmp->surface.level[offset_level].offset >> 8;
+ view->tex_resource = &tmp->resource;
+ view->tex_resource_words[0] = (S_038000_DIM(r600_tex_dim(texture->target)) |
+ S_038000_TILE_MODE(array_mode) |
+ S_038000_TILE_TYPE(tile_type) |
+ S_038000_PITCH((pitch / 8) - 1) |
+ S_038000_TEX_WIDTH(width - 1));
+ view->tex_resource_words[1] = (S_038004_TEX_HEIGHT(height - 1) |
+ S_038004_TEX_DEPTH(depth - 1) |
+ S_038004_DATA_FORMAT(format));
+ view->tex_resource_words[2] = tmp->surface.level[offset_level].offset >> 8;
if (offset_level >= tmp->surface.last_level) {
- rstate->val[3] = tmp->surface.level[offset_level].offset >> 8;
+ view->tex_resource_words[3] = tmp->surface.level[offset_level].offset >> 8;
} else {
- rstate->val[3] = tmp->surface.level[offset_level + 1].offset >> 8;
+ view->tex_resource_words[3] = tmp->surface.level[offset_level + 1].offset >> 8;
}
- rstate->val[4] = (word4 |
- S_038010_SRF_MODE_ALL(V_038010_SRF_MODE_ZERO_CLAMP_MINUS_ONE) |
- S_038010_REQUEST_SIZE(1) |
- S_038010_ENDIAN_SWAP(endian) |
- S_038010_BASE_LEVEL(0));
- rstate->val[5] = (S_038014_LAST_LEVEL(last_level) |
- S_038014_BASE_ARRAY(state->u.tex.first_layer) |
- S_038014_LAST_ARRAY(state->u.tex.last_layer));
- rstate->val[6] = (S_038018_TYPE(V_038010_SQ_TEX_VTX_VALID_TEXTURE) |
- S_038018_MAX_ANISO(4 /* max 16 samples */));
+ view->tex_resource_words[4] = (word4 |
+ S_038010_SRF_MODE_ALL(V_038010_SRF_MODE_ZERO_CLAMP_MINUS_ONE) |
+ S_038010_REQUEST_SIZE(1) |
+ S_038010_ENDIAN_SWAP(endian) |
+ S_038010_BASE_LEVEL(0));
+ view->tex_resource_words[5] = (S_038014_LAST_LEVEL(last_level) |
+ S_038014_BASE_ARRAY(state->u.tex.first_layer) |
+ S_038014_LAST_ARRAY(state->u.tex.last_layer));
+ view->tex_resource_words[6] = (S_038018_TYPE(V_038010_SQ_TEX_VTX_VALID_TEXTURE) |
+ S_038018_MAX_ANISO(4 /* max 16 samples */));
}
return &view->base;
}
-static void r600_set_sampler_views(struct r600_context *rctx,
- struct r600_textures_info *dst,
- unsigned count,
- struct pipe_sampler_view **views,
- void (*set_resource)(struct r600_context*, struct r600_pipe_resource_state*, unsigned))
-{
- struct r600_pipe_sampler_view **rviews = (struct r600_pipe_sampler_view **)views;
- unsigned i;
-
- if (count)
- r600_inval_texture_cache(rctx);
-
- for (i = 0; i < count; i++) {
- if (rviews[i]) {
- if (((struct r600_resource_texture *)rviews[i]->base.texture)->is_depth)
- rctx->have_depth_texture = true;
-
- /* Changing from array to non-arrays textures and vice versa requires updating TEX_ARRAY_OVERRIDE. */
- if ((rviews[i]->base.texture->target == PIPE_TEXTURE_1D_ARRAY ||
- rviews[i]->base.texture->target == PIPE_TEXTURE_2D_ARRAY) != dst->is_array_sampler[i])
- dst->samplers_dirty = true;
-
- set_resource(rctx, &rviews[i]->state, i + R600_MAX_CONST_BUFFERS);
- } else {
- set_resource(rctx, NULL, i + R600_MAX_CONST_BUFFERS);
- }
-
- pipe_sampler_view_reference(
- (struct pipe_sampler_view **)&dst->views[i],
- views[i]);
- }
-
- for (i = count; i < dst->n_views; i++) {
- if (dst->views[i]) {
- set_resource(rctx, NULL, i + R600_MAX_CONST_BUFFERS);
- pipe_sampler_view_reference((struct pipe_sampler_view **)&dst->views[i], NULL);
- }
- }
-
- dst->n_views = count;
-}
-
static void r600_set_vs_sampler_views(struct pipe_context *ctx, unsigned count,
struct pipe_sampler_view **views)
{
struct r600_context *rctx = (struct r600_context *)ctx;
- r600_set_sampler_views(rctx, &rctx->vs_samplers, count, views,
- r600_context_pipe_state_set_vs_resource);
+ r600_set_sampler_views(rctx, &rctx->vs_samplers, count, views);
}
static void r600_set_ps_sampler_views(struct pipe_context *ctx, unsigned count,
struct pipe_sampler_view **views)
{
struct r600_context *rctx = (struct r600_context *)ctx;
- r600_set_sampler_views(rctx, &rctx->ps_samplers, count, views,
- r600_context_pipe_state_set_ps_resource);
+ r600_set_sampler_views(rctx, &rctx->ps_samplers, count, views);
}
static void r600_set_seamless_cubemap(struct r600_context *rctx, boolean enable)
/* TEX_ARRAY_OVERRIDE must be set for array textures to disable
* filtering between layers.
* Don't update TEX_ARRAY_OVERRIDE if we don't have the sampler view. */
- if (tex->views[i]) {
- if (tex->views[i]->base.texture->target == PIPE_TEXTURE_1D_ARRAY ||
- tex->views[i]->base.texture->target == PIPE_TEXTURE_2D_ARRAY) {
+ if (tex->views.views[i]) {
+ if (tex->views.views[i]->base.texture->target == PIPE_TEXTURE_1D_ARRAY ||
+ tex->views.views[i]->base.texture->target == PIPE_TEXTURE_2D_ARRAY) {
tex->samplers[i]->rstate.regs[0].value |= S_03C000_TEX_ARRAY_OVERRIDE(1);
tex->is_array_sampler[i] = true;
} else {
unsigned offset;
const struct util_format_description *desc;
int i;
- unsigned blend_bypass = 0, blend_clamp = 1;
+ bool blend_bypass = 0, blend_clamp = 1, alphatest_bypass;
surf = (struct r600_surface *)state->cbufs[cb];
rtex = (struct r600_resource_texture*)state->cbufs[cb]->texture;
- if (rtex->is_depth)
- rctx->have_depth_fb = TRUE;
-
if (rtex->is_depth && !rtex->is_flushing_texture) {
rtex = rtex->flushed_depth_texture;
}
}
format = r600_translate_colorformat(surf->base.format);
+ assert(format != ~0);
+
swap = r600_translate_colorswap(surf->base.format);
- if(rtex->resource.b.b.usage == PIPE_USAGE_STAGING) {
+ assert(swap != ~0);
+
+ if (rtex->resource.b.b.usage == PIPE_USAGE_STAGING) {
endian = ENDIAN_NONE;
} else {
endian = r600_colorformat_endian_swap(format);
blend_bypass = 1;
}
- if (ntype == V_0280A0_NUMBER_UINT || ntype == V_0280A0_NUMBER_SINT)
- rctx->sx_alpha_test_control |= S_028410_ALPHA_TEST_BYPASS(1);
- else
- rctx->sx_alpha_test_control &= C_028410_ALPHA_TEST_BYPASS;
+ /* Alpha-test is done on the first colorbuffer only. */
+ if (cb == 0) {
+ alphatest_bypass = ntype == V_0280A0_NUMBER_UINT || ntype == V_0280A0_NUMBER_SINT;
+ if (rctx->alphatest_state.bypass != alphatest_bypass) {
+ rctx->alphatest_state.bypass = alphatest_bypass;
+ r600_atom_dirty(rctx, &rctx->alphatest_state.atom);
+ }
+ }
color_info |= S_0280A0_FORMAT(format) |
S_0280A0_COMP_SWAP(swap) |
ntype != V_0280A0_NUMBER_UINT &&
ntype != V_0280A0_NUMBER_SINT) &&
G_0280A0_BLEND_CLAMP(color_info) &&
- !G_0280A0_BLEND_FLOAT32(color_info))
+ !G_0280A0_BLEND_FLOAT32(color_info)) {
color_info |= S_0280A0_SOURCE_FORMAT(V_0280A0_EXPORT_NORM);
+ } else {
+ rctx->export_16bpc = false;
+ }
} else {
/* EXPORT_NORM can be enabled if:
* - 11-bit or smaller UNORM/SNORM/SRGB
desc->channel[i].type != UTIL_FORMAT_TYPE_FLOAT &&
ntype != V_0280A0_NUMBER_UINT && ntype != V_0280A0_NUMBER_SINT) ||
(desc->channel[i].size < 17 &&
- desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT)))
+ desc->channel[i].type == UTIL_FORMAT_TYPE_FLOAT))) {
color_info |= S_0280A0_SOURCE_FORMAT(V_0280A0_EXPORT_NORM);
+ } else {
+ rctx->export_16bpc = false;
+ }
}
- if (cb == 0)
- rctx->color0_format = color_info;
+ /* for possible dual-src MRT write color info 1 */
+ if (cb == 0 && rctx->framebuffer.nr_cbufs == 1) {
+ r600_pipe_state_add_reg_bo(rstate,
+ R_0280A0_CB_COLOR0_INFO + 1 * 4,
+ color_info, &rtex->resource, RADEON_USAGE_READWRITE);
+ }
r600_pipe_state_add_reg_bo(rstate,
R_028040_CB_COLOR0_BASE + cb * 4,
}
}
- format = r600_translate_dbformat(state->zsbuf->texture->format);
+ format = r600_translate_dbformat(state->zsbuf->format);
+ assert(format != ~0);
r600_pipe_state_add_reg_bo(rstate, R_02800C_DB_DEPTH_BASE,
offset >> 8, &rtex->resource, RADEON_USAGE_READWRITE);
{
struct r600_context *rctx = (struct r600_context *)ctx;
struct r600_pipe_state *rstate = CALLOC_STRUCT(r600_pipe_state);
- uint32_t tl, br, shader_control;
+ uint32_t tl, br;
if (rstate == NULL)
return;
util_copy_framebuffer_state(&rctx->framebuffer, state);
/* build states */
- rctx->have_depth_fb = 0;
+ rctx->export_16bpc = true;
+ rctx->nr_cbufs = state->nr_cbufs;
+
for (int i = 0; i < state->nr_cbufs; i++) {
r600_cb(rctx, rstate, state, i);
}
r600_db(rctx, rstate, state);
}
- shader_control = 0;
- rctx->fb_cb_shader_mask = 0;
- for (int i = 0; i < state->nr_cbufs; i++) {
- shader_control |= 1 << i;
- rctx->fb_cb_shader_mask |= 0xf << (i * 4);
- }
tl = S_028240_TL_X(0) | S_028240_TL_Y(0) | S_028240_WINDOW_OFFSET_DISABLE(1);
br = S_028244_BR_X(state->width) | S_028244_BR_Y(state->height);
r600_pipe_state_add_reg(rstate,
R_028208_PA_SC_WINDOW_SCISSOR_BR, br);
+ /* Always enable the first colorbuffer in CB_SHADER_CONTROL. This
+ * will assure that the alpha-test will work even if there is
+ * no colorbuffer bound. */
r600_pipe_state_add_reg(rstate, R_0287A0_CB_SHADER_CONTROL,
- shader_control);
+ (1ull << MAX2(state->nr_cbufs, 1)) - 1);
+
+ if (state->nr_cbufs == 0 && rctx->alphatest_state.bypass) {
+ rctx->alphatest_state.bypass = false;
+ r600_atom_dirty(rctx, &rctx->alphatest_state.atom);
+ }
free(rctx->states[R600_PIPE_STATE_FRAMEBUFFER]);
rctx->states[R600_PIPE_STATE_FRAMEBUFFER] = rstate;
if (state->zsbuf) {
r600_polygon_offset_update(rctx);
}
+
+ if (rctx->cb_misc_state.nr_cbufs != state->nr_cbufs) {
+ rctx->cb_misc_state.nr_cbufs = state->nr_cbufs;
+ r600_atom_dirty(rctx, &rctx->cb_misc_state.atom);
+ }
+}
+
+static void r600_emit_cb_misc_state(struct r600_context *rctx, struct r600_atom *atom)
+{
+ struct radeon_winsys_cs *cs = rctx->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 multiwrite = a->multiwrite && a->nr_cbufs > 1;
+
+ r600_write_context_reg_seq(cs, R_028238_CB_TARGET_MASK, 2);
+ r600_write_value(cs, a->blend_colormask & fb_colormask); /* R_028238_CB_TARGET_MASK */
+ r600_write_value(cs, (a->dual_src_blend ? ps_colormask : 0) | fb_colormask); /* R_02823C_CB_SHADER_MASK */
+ r600_write_context_reg(cs, R_028808_CB_COLOR_CONTROL,
+ a->cb_color_control |
+ S_028808_MULTIWRITE_ENABLE(multiwrite));
}
static void r600_emit_db_misc_state(struct r600_context *rctx, struct r600_atom *atom)
}
db_render_override |= S_028D10_NOOP_CULL_DISABLE(1);
}
- if (a->flush_depthstencil_enabled) {
+ if (a->flush_depthstencil_through_cb) {
db_render_control |= S_028D0C_DEPTH_COPY_ENABLE(1) |
S_028D0C_STENCIL_COPY_ENABLE(1) |
S_028D0C_COPY_CENTROID(1);
static void r600_emit_vertex_buffers(struct r600_context *rctx, struct r600_atom *atom)
{
struct radeon_winsys_cs *cs = rctx->cs;
- struct pipe_vertex_buffer *vb = rctx->vertex_buffer;
- unsigned count = rctx->nr_vertex_buffers;
- unsigned i, offset;
+ uint32_t dirty_mask = rctx->vertex_buffer_state.dirty_mask;
- for (i = 0; i < count; i++) {
- struct r600_resource *rbuffer = (struct r600_resource*)vb[i].buffer;
+ while (dirty_mask) {
+ struct pipe_vertex_buffer *vb;
+ struct r600_resource *rbuffer;
+ unsigned offset;
+ unsigned buffer_index = u_bit_scan(&dirty_mask);
- if (!rbuffer) {
- continue;
- }
+ vb = &rctx->vertex_buffer_state.vb[buffer_index];
+ rbuffer = (struct r600_resource*)vb->buffer;
+ assert(rbuffer);
- offset = vb[i].buffer_offset;
+ offset = vb->buffer_offset;
/* fetch resources start at index 320 */
r600_write_value(cs, PKT3(PKT3_SET_RESOURCE, 7, 0));
- r600_write_value(cs, (320 + i) * 7);
+ r600_write_value(cs, (320 + buffer_index) * 7);
r600_write_value(cs, offset); /* RESOURCEi_WORD0 */
r600_write_value(cs, rbuffer->buf->size - offset - 1); /* RESOURCEi_WORD1 */
r600_write_value(cs, /* RESOURCEi_WORD2 */
S_038008_ENDIAN_SWAP(r600_endian_swap(32)) |
- S_038008_STRIDE(vb[i].stride));
+ S_038008_STRIDE(vb->stride));
r600_write_value(cs, 0); /* RESOURCEi_WORD3 */
r600_write_value(cs, 0); /* RESOURCEi_WORD4 */
r600_write_value(cs, 0); /* RESOURCEi_WORD5 */
state->dirty_mask = 0;
}
-static void r600_emit_vs_constant_buffer(struct r600_context *rctx, struct r600_atom *atom)
+static void r600_emit_vs_constant_buffers(struct r600_context *rctx, struct r600_atom *atom)
{
r600_emit_constant_buffers(rctx, &rctx->vs_constbuf_state, 160,
R_028180_ALU_CONST_BUFFER_SIZE_VS_0,
R_028980_ALU_CONST_CACHE_VS_0);
}
-static void r600_emit_ps_constant_buffer(struct r600_context *rctx, struct r600_atom *atom)
+static void r600_emit_ps_constant_buffers(struct r600_context *rctx, struct r600_atom *atom)
{
r600_emit_constant_buffers(rctx, &rctx->ps_constbuf_state, 0,
R_028140_ALU_CONST_BUFFER_SIZE_PS_0,
R_028940_ALU_CONST_CACHE_PS_0);
}
+static void r600_emit_sampler_views(struct r600_context *rctx,
+ struct r600_samplerview_state *state,
+ unsigned resource_id_base)
+{
+ struct radeon_winsys_cs *cs = rctx->cs;
+ uint32_t dirty_mask = state->dirty_mask;
+
+ while (dirty_mask) {
+ struct r600_pipe_sampler_view *rview;
+ unsigned resource_index = u_bit_scan(&dirty_mask);
+ unsigned reloc;
+
+ rview = state->views[resource_index];
+ assert(rview);
+
+ r600_write_value(cs, PKT3(PKT3_SET_RESOURCE, 7, 0));
+ r600_write_value(cs, (resource_id_base + resource_index) * 7);
+ r600_write_array(cs, 7, rview->tex_resource_words);
+
+ /* XXX The kernel needs two relocations. This is stupid. */
+ reloc = r600_context_bo_reloc(rctx, rview->tex_resource,
+ RADEON_USAGE_READ);
+ r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
+ r600_write_value(cs, reloc);
+ r600_write_value(cs, PKT3(PKT3_NOP, 0, 0));
+ r600_write_value(cs, reloc);
+ }
+ state->dirty_mask = 0;
+}
+
+static void r600_emit_vs_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
+{
+ r600_emit_sampler_views(rctx, &rctx->vs_samplers.views, 160 + R600_MAX_CONST_BUFFERS);
+}
+
+static void r600_emit_ps_sampler_views(struct r600_context *rctx, struct r600_atom *atom)
+{
+ r600_emit_sampler_views(rctx, &rctx->ps_samplers.views, R600_MAX_CONST_BUFFERS);
+}
+
void r600_init_state_functions(struct r600_context *rctx)
{
+ r600_init_atom(&rctx->cb_misc_state.atom, r600_emit_cb_misc_state, 0, 0);
+ r600_atom_dirty(rctx, &rctx->cb_misc_state.atom);
r600_init_atom(&rctx->db_misc_state.atom, r600_emit_db_misc_state, 4, 0);
r600_atom_dirty(rctx, &rctx->db_misc_state.atom);
- r600_init_atom(&rctx->vertex_buffer_state, r600_emit_vertex_buffers, 0, 0);
- r600_init_atom(&rctx->vs_constbuf_state.atom, r600_emit_vs_constant_buffer, 0, 0);
- r600_init_atom(&rctx->ps_constbuf_state.atom, r600_emit_ps_constant_buffer, 0, 0);
+ r600_init_atom(&rctx->vertex_buffer_state.atom, r600_emit_vertex_buffers, 0, 0);
+ r600_init_atom(&rctx->vs_constbuf_state.atom, r600_emit_vs_constant_buffers, 0, 0);
+ r600_init_atom(&rctx->ps_constbuf_state.atom, r600_emit_ps_constant_buffers, 0, 0);
+ r600_init_atom(&rctx->vs_samplers.views.atom, r600_emit_vs_sampler_views, 0, 0);
+ r600_init_atom(&rctx->ps_samplers.views.atom, r600_emit_ps_sampler_views, 0, 0);
rctx->context.create_blend_state = r600_create_blend_state;
rctx->context.create_depth_stencil_alpha_state = r600_create_dsa_state;
- rctx->context.create_fs_state = r600_create_shader_state;
+ rctx->context.create_fs_state = r600_create_shader_state_ps;
rctx->context.create_rasterizer_state = r600_create_rs_state;
rctx->context.create_sampler_state = r600_create_sampler_state;
rctx->context.create_sampler_view = r600_create_sampler_view;
rctx->context.create_vertex_elements_state = r600_create_vertex_elements;
- rctx->context.create_vs_state = r600_create_shader_state;
+ rctx->context.create_vs_state = r600_create_shader_state_vs;
rctx->context.bind_blend_state = r600_bind_blend_state;
rctx->context.bind_depth_stencil_alpha_state = r600_bind_dsa_state;
rctx->context.bind_fragment_sampler_states = r600_bind_ps_samplers;
rctx->context.set_stream_output_targets = r600_set_so_targets;
}
+/* Adjust GPR allocation on R6xx/R7xx */
void r600_adjust_gprs(struct r600_context *rctx)
{
struct r600_pipe_state rstate;
unsigned tmp;
int diff;
- if (rctx->chip_class >= EVERGREEN)
- return;
+ /* XXX: Following call moved from r600_bind_[ps|vs]_shader,
+ * it seems eg+ doesn't need it, r6xx/7xx probably need it only for
+ * adjusting the GPR allocation?
+ * Do we need this if we aren't really changing config below? */
+ r600_inval_shader_cache(rctx);
- if (!rctx->ps_shader || !rctx->vs_shader)
- return;
-
- if (rctx->ps_shader->shader.bc.ngpr > rctx->default_ps_gprs)
+ if (rctx->ps_shader->current->shader.bc.ngpr > rctx->default_ps_gprs)
{
- diff = rctx->ps_shader->shader.bc.ngpr - rctx->default_ps_gprs;
+ diff = rctx->ps_shader->current->shader.bc.ngpr - rctx->default_ps_gprs;
num_vs_gprs -= diff;
num_ps_gprs += diff;
}
- if (rctx->vs_shader->shader.bc.ngpr > rctx->default_vs_gprs)
+ if (rctx->vs_shader->current->shader.bc.ngpr > rctx->default_vs_gprs)
{
- diff = rctx->vs_shader->shader.bc.ngpr - rctx->default_vs_gprs;
+ diff = rctx->vs_shader->current->shader.bc.ngpr - rctx->default_vs_gprs;
num_ps_gprs -= diff;
num_vs_gprs += diff;
}
r600_store_context_reg(cb, R_0288A4_SQ_PGM_RESOURCES_FS, 0);
r600_store_context_reg(cb, R_0288DC_SQ_PGM_CF_OFFSET_FS, 0);
- if (rctx->chip_class == R700 && rctx->screen->info.r600_has_streamout)
+ if (rctx->chip_class == R700 && rctx->screen->has_streamout)
r600_store_context_reg(cb, R_028354_SX_SURFACE_SYNC, S_028354_SURFACE_SYNC_MASK(0xf));
r600_store_context_reg(cb, R_028800_DB_DEPTH_CONTROL, 0);
+ if (rctx->screen->has_streamout) {
+ r600_store_context_reg(cb, R_028B28_VGT_STRMOUT_DRAW_OPAQUE_OFFSET, 0);
+ }
r600_store_loop_const(cb, R_03E200_SQ_LOOP_CONST_0, 0x1000FFF);
r600_store_loop_const(cb, R_03E200_SQ_LOOP_CONST_0 + (32 * 4), 0x1000FFF);
int pos_index = -1, face_index = -1;
unsigned tmp, sid, ufi = 0;
int need_linear = 0;
+ unsigned z_export = 0, stencil_export = 0;
rstate->nregs = 0;
db_shader_control = S_02880C_Z_ORDER(V_02880C_EARLY_Z_THEN_LATE_Z);
for (i = 0; i < rshader->noutput; i++) {
if (rshader->output[i].name == TGSI_SEMANTIC_POSITION)
- db_shader_control |= S_02880C_Z_EXPORT_ENABLE(1);
+ z_export = 1;
if (rshader->output[i].name == TGSI_SEMANTIC_STENCIL)
- db_shader_control |= S_02880C_STENCIL_REF_EXPORT_ENABLE(1);
+ stencil_export = 1;
}
+ db_shader_control |= S_02880C_Z_EXPORT_ENABLE(z_export);
+ db_shader_control |= S_02880C_STENCIL_REF_EXPORT_ENABLE(stencil_export);
if (rshader->uses_kill)
db_shader_control |= S_02880C_KILL_ENABLE(1);
exports_ps = 0;
- num_cout = 0;
for (i = 0; i < rshader->noutput; i++) {
if (rshader->output[i].name == TGSI_SEMANTIC_POSITION ||
- rshader->output[i].name == TGSI_SEMANTIC_STENCIL)
+ rshader->output[i].name == TGSI_SEMANTIC_STENCIL) {
exports_ps |= 1;
- else if (rshader->output[i].name == TGSI_SEMANTIC_COLOR) {
- num_cout++;
}
}
+ num_cout = rshader->nr_ps_color_exports;
exports_ps |= S_028854_EXPORT_COLORS(num_cout);
if (!exports_ps) {
/* always at least export 1 component per pixel */
exports_ps = 2;
}
- shader->ps_cb_shader_mask = (1ULL << ((unsigned)num_cout * 4)) - 1;
+ shader->nr_ps_color_outputs = num_cout;
spi_ps_in_control_0 = S_0286CC_NUM_INTERP(rshader->ninput) |
S_0286CC_PERSP_GRADIENT_ENA(1)|
R_028854_SQ_PGM_EXPORTS_PS,
exports_ps);
/* only set some bits here, the other bits are set in the dsa state */
- r600_pipe_state_add_reg(rstate, R_02880C_DB_SHADER_CONTROL,
- db_shader_control);
+ shader->db_shader_control = db_shader_control;
+ shader->ps_depth_export = z_export | stencil_export;
shader->sprite_coord_enable = rctx->sprite_coord_enable;
if (rctx->rasterizer)
void *r600_create_db_flush_dsa(struct r600_context *rctx)
{
struct pipe_depth_stencil_alpha_state dsa;
- struct r600_pipe_state *rstate;
- struct r600_pipe_dsa *dsa_state;
boolean quirk = false;
if (rctx->family == CHIP_RV610 || rctx->family == CHIP_RV630 ||
dsa.stencil[0].writemask = 0xff;
}
- rstate = rctx->context.create_depth_stencil_alpha_state(&rctx->context, &dsa);
- dsa_state = (struct r600_pipe_dsa*)rstate;
- dsa_state->is_flush = true;
- return rstate;
+ return rctx->context.create_depth_stencil_alpha_state(&rctx->context, &dsa);
+}
+
+void r600_update_dual_export_state(struct r600_context * rctx)
+{
+ unsigned dual_export = rctx->export_16bpc && rctx->nr_cbufs &&
+ !rctx->ps_shader->current->ps_depth_export;
+ unsigned db_shader_control = rctx->ps_shader->current->db_shader_control |
+ S_02880C_DUAL_EXPORT_ENABLE(dual_export);
+
+ if (db_shader_control != rctx->db_shader_control) {
+ struct r600_pipe_state rstate;
+
+ rctx->db_shader_control = db_shader_control;
+ rstate.nregs = 0;
+ r600_pipe_state_add_reg(&rstate, R_02880C_DB_SHADER_CONTROL, db_shader_control);
+ r600_context_pipe_state_set(rctx, &rstate);
+ }
}