X-Git-Url: https://git.libre-soc.org/?a=blobdiff_plain;ds=sidebyside;f=src%2Fgallium%2Fdrivers%2Fradeonsi%2Fsi_state.h;h=5233be72462203ca88aa25c5102df2bd4c60a800;hb=19f3141e6ab34dc7389b5e1fa9f3dca5e23b2191;hp=8de86753f6d66138957edc86a1ba418046e92722;hpb=7be6186e0cecccd7a64e9ba242f17ddbe588d84a;p=mesa.git diff --git a/src/gallium/drivers/radeonsi/si_state.h b/src/gallium/drivers/radeonsi/si_state.h index 8de86753f6d..5233be72462 100644 --- a/src/gallium/drivers/radeonsi/si_state.h +++ b/src/gallium/drivers/radeonsi/si_state.h @@ -19,9 +19,6 @@ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE * USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * Authors: - * Christian König */ #ifndef SI_STATE_H @@ -30,6 +27,8 @@ #include "si_pm4.h" #include "radeon/r600_pipe_common.h" +#include "pipebuffer/pb_slab.h" + #define SI_NUM_GRAPHICS_SHADERS (PIPE_SHADER_TESS_EVAL+1) #define SI_NUM_SHADERS (PIPE_SHADER_COMPUTE+1) @@ -47,14 +46,17 @@ struct si_shader_selector; struct si_state_blend { struct si_pm4_state pm4; uint32_t cb_target_mask; - bool alpha_to_coverage; - bool alpha_to_one; - bool dual_src_blend; /* Set 0xf or 0x0 (4 bits) per render target if the following is * true. ANDed with spi_shader_col_format. */ + unsigned cb_target_enabled_4bit; unsigned blend_enable_4bit; unsigned need_src_alpha_4bit; + unsigned commutative_4bit; + bool alpha_to_coverage:1; + bool alpha_to_one:1; + bool dual_src_blend:1; + bool logicop_enable:1; }; struct si_state_rasterizer { @@ -63,6 +65,8 @@ struct si_state_rasterizer { struct si_pm4_state *pm4_poly_offset; unsigned pa_sc_line_stipple; unsigned pa_cl_clip_cntl; + float line_width; + float max_point_size; unsigned sprite_coord_enable:8; unsigned clip_plane_enable:8; unsigned flatshade:1; @@ -86,10 +90,36 @@ struct si_dsa_stencil_ref_part { uint8_t writemask[2]; }; +struct si_dsa_order_invariance { + /** Whether the final result in Z/S buffers is guaranteed to be + * invariant under changes to the order in which fragments arrive. */ + bool zs:1; + + /** Whether the set of fragments that pass the combined Z/S test is + * guaranteed to be invariant under changes to the order in which + * fragments arrive. */ + bool pass_set:1; + + /** Whether the last fragment that passes the combined Z/S test at each + * sample is guaranteed to be invariant under changes to the order in + * which fragments arrive. */ + bool pass_last:1; +}; + struct si_state_dsa { struct si_pm4_state pm4; - unsigned alpha_func; struct si_dsa_stencil_ref_part stencil_ref; + + /* 0 = without stencil buffer, 1 = when both Z and S buffers are present */ + struct si_dsa_order_invariance order_invariance[2]; + + ubyte alpha_func:3; + bool depth_enabled:1; + bool depth_write_enabled:1; + bool stencil_enabled:1; + bool stencil_write_enabled:1; + bool db_can_write:1; + }; struct si_stencil_ref { @@ -98,18 +128,23 @@ struct si_stencil_ref { struct si_dsa_stencil_ref_part dsa_part; }; -struct si_vertex_element +struct si_vertex_elements { - unsigned count; - unsigned first_vb_use_mask; - /* Vertex buffer descriptor list size aligned for optimal prefetch. */ - unsigned desc_list_byte_size; - - uint8_t fix_fetch[SI_MAX_ATTRIBS]; + uint32_t instance_divisors[SI_MAX_ATTRIBS]; uint32_t rsrc_word3[SI_MAX_ATTRIBS]; - uint32_t format_size[SI_MAX_ATTRIBS]; - struct pipe_vertex_element elements[SI_MAX_ATTRIBS]; + uint16_t src_offset[SI_MAX_ATTRIBS]; + uint8_t fix_fetch[SI_MAX_ATTRIBS]; + uint8_t format_size[SI_MAX_ATTRIBS]; + uint8_t vertex_buffer_index[SI_MAX_ATTRIBS]; + + uint8_t count; bool uses_instance_divisors; + + uint16_t first_vb_use_mask; + /* Vertex buffer descriptor list size aligned for optimal prefetch. */ + uint16_t desc_list_byte_size; + uint16_t instance_divisor_is_one; /* bitmask of inputs */ + uint16_t instance_divisor_is_fetched; /* bitmask of inputs */ }; union si_state { @@ -134,20 +169,20 @@ union si_state { union si_state_atoms { struct { /* The order matters. */ - struct r600_atom *prefetch_L2; struct r600_atom *render_cond; struct r600_atom *streamout_begin; struct r600_atom *streamout_enable; /* must be after streamout_begin */ struct r600_atom *framebuffer; struct r600_atom *msaa_sample_locs; struct r600_atom *db_render_state; + struct r600_atom *dpbb_state; struct r600_atom *msaa_config; struct r600_atom *sample_mask; struct r600_atom *cb_render_state; struct r600_atom *blend_color; struct r600_atom *clip_regs; struct r600_atom *clip_state; - struct r600_atom *shader_userdata; + struct r600_atom *shader_pointers; struct r600_atom *scissors; struct r600_atom *viewports; struct r600_atom *stencil_ref; @@ -177,6 +212,7 @@ enum { SI_VS_STREAMOUT_BUF3, SI_HS_CONST_DEFAULT_TESS_LEVELS, + SI_VS_CONST_INSTANCE_DIVISORS, SI_VS_CONST_CLIP_PLANES, SI_PS_CONST_POLY_STIPPLE, SI_PS_CONST_SAMPLE_POSITIONS, @@ -188,13 +224,12 @@ enum { * are contiguous: * * 0 - rw buffers - * 1 - vertex const buffers - * 2 - vertex shader buffers - * ... - * 5 - fragment const buffers - * ... - * 21 - compute const buffers + * 1 - vertex const and shader buffers + * 2 - vertex samplers and images + * 3 - fragment const and shader buffer * ... + * 11 - compute const and shader buffers + * 12 - compute samplers and images */ enum { SI_SHADER_DESCS_CONST_AND_SHADER_BUFFERS, @@ -209,6 +244,11 @@ enum { #define SI_NUM_DESCS (SI_DESCS_FIRST_SHADER + \ SI_NUM_SHADERS * SI_NUM_SHADER_DESCS) +#define SI_DESCS_SHADER_MASK(name) \ + u_bit_consecutive(SI_DESCS_FIRST_SHADER + \ + PIPE_SHADER_##name * SI_NUM_SHADER_DESCS, \ + SI_NUM_SHADER_DESCS) + /* This represents descriptors in memory, such as buffer resources, * image resources, and sampler states. */ @@ -217,59 +257,38 @@ struct si_descriptors { uint32_t *list; /* The list in mapped GPU memory. */ uint32_t *gpu_list; - /* The size of one descriptor. */ - unsigned element_dw_size; - /* The maximum number of descriptors. */ - unsigned num_elements; /* The buffer where the descriptors have been uploaded. */ struct r600_resource *buffer; - int buffer_offset; /* can be negative if not using lower slots */ - - /* Offset in CE RAM */ - unsigned ce_offset; + uint64_t gpu_address; - /* Slots allocated in CE RAM. If we get active slots outside of this - * range, direct uploads to memory will be used instead. This basically - * governs switching between onchip (CE) and offchip (upload) modes. - */ - unsigned first_ce_slot; - unsigned num_ce_slots; + /* The maximum number of descriptors. */ + uint32_t num_elements; /* Slots that are used by currently-bound shaders. - * With CE: It determines which slots are dumped to L2. - * It doesn't skip uploads to CE RAM. - * Without CE: It determines which slots are uploaded. + * It determines which slots are uploaded. */ - unsigned first_active_slot; - unsigned num_active_slots; - - /* Slots that have been changed and need to be uploaded. */ - uint64_t dirty_mask; - - /* Whether CE is used to upload this descriptor array. */ - bool uses_ce; - - /* The shader userdata offset within a shader where the 64-bit pointer to the descriptor - * array will be stored. */ - unsigned shader_userdata_offset; -}; + uint32_t first_active_slot; + uint32_t num_active_slots; -struct si_sampler_views { - struct pipe_sampler_view *views[SI_NUM_SAMPLERS]; - struct si_sampler_state *sampler_states[SI_NUM_SAMPLERS]; - - /* The i-th bit is set if that element is enabled (non-NULL resource). */ - unsigned enabled_mask; + /* The SGPR index where the 64-bit pointer to the descriptor array will + * be stored. */ + ubyte shader_userdata_offset; + /* The size of one descriptor. */ + ubyte element_dw_size; + /* If there is only one slot enabled, bind it directly instead of + * uploading descriptors. -1 if disabled. */ + signed char slot_index_to_bind_directly; }; struct si_buffer_resources { - enum radeon_bo_usage shader_usage; /* READ, WRITE, or READWRITE */ - enum radeon_bo_usage shader_usage_constbuf; - enum radeon_bo_priority priority; - enum radeon_bo_priority priority_constbuf; struct pipe_resource **buffers; /* this has num_buffers elements */ + enum radeon_bo_usage shader_usage:4; /* READ, WRITE, or READWRITE */ + enum radeon_bo_usage shader_usage_constbuf:4; + enum radeon_bo_priority priority:6; + enum radeon_bo_priority priority_constbuf:6; + /* The i-th bit is set if that element is enabled (non-NULL resource). */ unsigned enabled_mask; }; @@ -280,6 +299,9 @@ struct si_buffer_resources { #define si_pm4_state_changed(sctx, member) \ ((sctx)->queued.named.member != (sctx)->emitted.named.member) +#define si_pm4_state_enabled_and_changed(sctx, member) \ + ((sctx)->queued.named.member && si_pm4_state_changed(sctx, member)) + #define si_pm4_bind_state(sctx, member, value) \ do { \ (sctx)->queued.named.member = (value); \ @@ -296,9 +318,6 @@ struct si_buffer_resources { } while(0) /* si_descriptors.c */ -void si_ce_save_all_descriptors_at_ib_end(struct si_context* sctx); -void si_ce_restore_all_descriptors_at_ib_start(struct si_context *sctx); -void si_ce_enable_loads(struct radeon_winsys_cs *ib); void si_set_mutable_tex_desc_fields(struct si_screen *sscreen, struct r600_texture *tex, const struct legacy_surf_level *base_level_info, @@ -322,20 +341,27 @@ bool si_upload_graphics_shader_descriptors(struct si_context *sctx); bool si_upload_compute_shader_descriptors(struct si_context *sctx); void si_release_all_descriptors(struct si_context *sctx); void si_all_descriptors_begin_new_cs(struct si_context *sctx); +void si_all_resident_buffers_begin_new_cs(struct si_context *sctx); void si_upload_const_buffer(struct si_context *sctx, struct r600_resource **rbuffer, const uint8_t *ptr, unsigned size, uint32_t *const_offset); void si_update_all_texture_descriptors(struct si_context *sctx); void si_shader_change_notify(struct si_context *sctx); void si_update_needs_color_decompress_masks(struct si_context *sctx); -void si_emit_graphics_shader_userdata(struct si_context *sctx, +void si_emit_graphics_shader_pointers(struct si_context *sctx, struct r600_atom *atom); -void si_emit_compute_shader_userdata(struct si_context *sctx); +void si_emit_compute_shader_pointers(struct si_context *sctx); void si_set_rw_buffer(struct si_context *sctx, uint slot, const struct pipe_constant_buffer *input); void si_set_active_descriptors(struct si_context *sctx, unsigned desc_idx, uint64_t new_active_mask); void si_set_active_descriptors_for_shader(struct si_context *sctx, struct si_shader_selector *sel); +bool si_bindless_descriptor_can_reclaim_slab(void *priv, + struct pb_slab_entry *entry); +struct pb_slab *si_bindless_descriptor_slab_alloc(void *priv, unsigned heap, + unsigned entry_size, + unsigned group_index); +void si_bindless_descriptor_slab_free(void *priv, struct pb_slab *pslab); /* si_state.c */ struct si_shader_selector; @@ -368,25 +394,46 @@ si_create_sampler_view_custom(struct pipe_context *ctx, const struct pipe_sampler_view *state, unsigned width0, unsigned height0, unsigned force_level); +void si_update_fb_dirtiness_after_rendering(struct si_context *sctx); + +/* si_state_binning.c */ +void si_emit_dpbb_state(struct si_context *sctx, struct r600_atom *state); -/* si_state_shader.c */ +/* si_state_shaders.c */ bool si_update_shaders(struct si_context *sctx); void si_init_shader_functions(struct si_context *sctx); bool si_init_shader_cache(struct si_screen *sscreen); void si_destroy_shader_cache(struct si_screen *sscreen); -void si_init_shader_selector_async(void *job, int thread_index); void si_get_active_slot_masks(const struct tgsi_shader_info *info, uint32_t *const_and_shader_buffers, uint64_t *samplers_and_images); +void *si_get_blit_vs(struct si_context *sctx, enum blitter_attrib_type type, + unsigned num_layers); /* si_state_draw.c */ void si_init_ia_multi_vgt_param_table(struct si_context *sctx); void si_emit_cache_flush(struct si_context *sctx); -void si_ce_pre_draw_synchronization(struct si_context *sctx); -void si_ce_post_draw_synchronization(struct si_context *sctx); void si_draw_vbo(struct pipe_context *ctx, const struct pipe_draw_info *dinfo); +void si_draw_rectangle(struct blitter_context *blitter, + void *vertex_elements_cso, + blitter_get_vs_func get_vs, + int x1, int y1, int x2, int y2, + float depth, unsigned num_instances, + enum blitter_attrib_type type, + const union blitter_attrib *attrib); void si_trace_emit(struct si_context *sctx); +/* si_state_msaa.c */ +void si_init_msaa_functions(struct si_context *sctx); +void si_emit_sample_locations(struct radeon_winsys_cs *cs, int nr_samples); + +/* si_state_streamout.c */ +void si_streamout_buffers_dirty(struct si_context *sctx); +void si_emit_streamout_end(struct si_context *sctx); +void si_update_prims_generated_query_state(struct si_context *sctx, + unsigned type, int diff); +void si_init_streamout_functions(struct si_context *sctx); + static inline unsigned si_tile_mode_index(struct r600_texture *rtex, unsigned level, bool stencil)