#include <stdio.h>
#include "brw_device_info.h"
#include "main/mtypes.h"
+#include "main/macros.h"
#ifdef __cplusplus
extern "C" {
* appear in *classes.
*/
int aligned_pairs_class;
- } fs_reg_sets[2];
+ } fs_reg_sets[3];
void (*shader_debug_log)(void *, const char *str, ...) PRINTFLIKE(2, 3);
void (*shader_perf_log)(void *, const char *str, ...) PRINTFLIKE(2, 3);
bool scalar_stage[MESA_SHADER_STAGES];
struct gl_shader_compiler_options glsl_compiler_options[MESA_SHADER_STAGES];
+
+ /**
+ * Apply workarounds for SIN and COS output range problems.
+ * This can negatively impact performance.
+ */
+ bool precise_trig;
};
* For Sandybridge, which shader w/a we need for gather quirks.
*/
enum gen6_gather_sampler_wa gen6_gather_wa[MAX_SAMPLERS];
+
+ /**
+ * Texture units that have a YUV image bound.
+ */
+ uint32_t y_u_v_image_mask;
+ uint32_t y_uv_image_mask;
+ uint32_t yx_xuxv_image_mask;
};
struct brw_sampler_prog_key_data tex;
};
+/** The program key for Tessellation Control Shaders. */
+struct brw_tcs_prog_key
+{
+ unsigned program_string_id;
+
+ GLenum tes_primitive_mode;
+
+ unsigned input_vertices;
+
+ /** A bitfield of per-patch outputs written. */
+ uint32_t patch_outputs_written;
+
+ /** A bitfield of per-vertex outputs written. */
+ uint64_t outputs_written;
+
+ struct brw_sampler_prog_key_data tex;
+};
+
+/** The program key for Tessellation Evaluation Shaders. */
+struct brw_tes_prog_key
+{
+ unsigned program_string_id;
+
+ /** A bitfield of per-patch inputs read. */
+ uint32_t patch_inputs_read;
+
+ /** A bitfield of per-vertex inputs read. */
+ uint64_t inputs_read;
+
+ struct brw_sampler_prog_key_data tex;
+};
+
/** The program key for Geometry Shaders. */
struct brw_gs_prog_key
{
uint8_t iz_lookup;
bool stats_wm:1;
bool flat_shade:1;
- bool persample_shading:1;
- bool persample_2x:1;
unsigned nr_color_regions:5;
bool replicate_alpha:1;
- bool render_to_fbo:1;
bool clamp_fragment_color:1;
- bool compute_pos_offset:1;
- bool compute_sample_id:1;
+ bool persample_interp:1;
+ bool multisample_fbo:1;
unsigned line_aa:2;
bool high_quality_derivatives:1;
+ bool force_dual_color_blend:1;
uint16_t drawable_height;
uint64_t input_slots_valid;
uint32_t abo_start;
uint32_t image_start;
uint32_t shader_time_start;
+ uint32_t plane_start[3];
/** @} */
} binding_table;
unsigned curb_read_length;
unsigned total_scratch;
+ unsigned total_shared;
/**
* Register where the thread expects to find input data from the URB
GLuint num_varying_inputs;
- GLuint dispatch_grf_start_reg_16;
- GLuint reg_blocks;
- GLuint reg_blocks_16;
+ uint8_t reg_blocks_0;
+ uint8_t reg_blocks_2;
+
+ uint8_t dispatch_grf_start_reg_2;
+ uint32_t prog_offset_2;
struct {
/** @{
bool computed_stencil;
bool early_fragment_tests;
- bool no_8;
+ bool dispatch_8;
+ bool dispatch_16;
bool dual_src_blend;
+ bool persample_dispatch;
bool uses_pos_offset;
bool uses_omask;
bool uses_kill;
+ bool uses_src_depth;
+ bool uses_src_w;
+ bool uses_sample_mask;
+ bool has_side_effects;
bool pulls_bary;
- uint32_t prog_offset_16;
/**
* Mask of which interpolation modes are required by the fragment shader.
*/
uint32_t barycentric_interp_modes;
+ /**
+ * Mask of which FS inputs are marked flat by the shader source. This is
+ * needed for setting up 3DSTATE_SF/SBE.
+ */
+ uint32_t flat_inputs;
+
/**
* Map from gl_varying_slot to the position within the FS setup data
* payload where the varying's attribute vertex deltas should be delivered.
int urb_setup[VARYING_SLOT_MAX];
};
+struct brw_push_const_block {
+ unsigned dwords; /* Dword count, not reg aligned */
+ unsigned regs;
+ unsigned size; /* Bytes, register aligned */
+};
+
struct brw_cs_prog_data {
struct brw_stage_prog_data base;
GLuint dispatch_grf_start_reg_16;
unsigned local_size[3];
unsigned simd_size;
+ unsigned threads;
bool uses_barrier;
bool uses_num_work_groups;
- unsigned local_invocation_id_regs;
+ int thread_local_id_index;
+
+ struct {
+ struct brw_push_const_block cross_thread;
+ struct brw_push_const_block per_thread;
+ struct brw_push_const_block total;
+ } push;
struct {
/** @{
* additional processing is applied before storing them in the VUE), the
* value is -1.
*/
- signed char varying_to_slot[BRW_VARYING_SLOT_COUNT];
+ signed char varying_to_slot[VARYING_SLOT_TESS_MAX];
/**
* Map from VUE slot to gl_varying_slot value. For slots that do not
*
* For slots that are not in use, the value is BRW_VARYING_SLOT_PAD.
*/
- signed char slot_to_varying[BRW_VARYING_SLOT_COUNT];
+ signed char slot_to_varying[VARYING_SLOT_TESS_MAX];
/**
* Total number of VUE slots in use
*/
int num_slots;
+
+ /**
+ * Number of per-patch VUE slots. Only valid for tessellation control
+ * shader outputs and tessellation evaluation shader inputs.
+ */
+ int num_per_patch_slots;
+
+ /**
+ * Number of per-vertex VUE slots. Only valid for tessellation control
+ * shader outputs and tessellation evaluation shader inputs.
+ */
+ int num_per_vertex_slots;
};
void brw_print_vue_map(FILE *fp, const struct brw_vue_map *vue_map);
GLbitfield64 slots_valid,
bool separate_shader);
+void brw_compute_tess_vue_map(struct brw_vue_map *const vue_map,
+ const GLbitfield64 slots_valid,
+ const GLbitfield is_patch);
+
enum shader_dispatch_mode {
DISPATCH_MODE_4X1_SINGLE = 0,
DISPATCH_MODE_4X2_DUAL_INSTANCE = 1,
GLuint urb_read_length;
GLuint total_grf;
+ uint32_t cull_distance_mask;
+
/* Used for calculating urb partitions. In the VS, this is the size of the
* URB entry used for both input and output to the thread. In the GS, this
* is the size of the URB entry used for output.
GLbitfield64 inputs_read;
unsigned nr_attributes;
+ unsigned nr_attribute_slots;
bool uses_vertexid;
bool uses_instanceid;
+ bool uses_basevertex;
+ bool uses_baseinstance;
+ bool uses_drawid;
+};
+
+struct brw_tcs_prog_data
+{
+ struct brw_vue_prog_data base;
+
+ /** Number vertices in output patch */
+ int instances;
+};
+
+
+struct brw_tes_prog_data
+{
+ struct brw_vue_prog_data base;
+
+ enum brw_tess_partitioning partitioning;
+ enum brw_tess_output_topology output_topology;
+ enum brw_tess_domain domain;
};
struct brw_gs_prog_data
{
struct brw_vue_prog_data base;
+ unsigned vertices_in;
+
/**
* Size of an output vertex, measured in HWORDS (32 bytes).
*/
/** @} */
+struct brw_compiler *
+brw_compiler_create(void *mem_ctx, const struct brw_device_info *devinfo);
+
/**
* Compile a vertex shader.
*
unsigned *final_assembly_size,
char **error_str);
+/**
+ * Compile a tessellation control shader.
+ *
+ * Returns the final assembly and the program's size.
+ */
+const unsigned *
+brw_compile_tcs(const struct brw_compiler *compiler,
+ void *log_data,
+ void *mem_ctx,
+ const struct brw_tcs_prog_key *key,
+ struct brw_tcs_prog_data *prog_data,
+ const struct nir_shader *nir,
+ int shader_time_index,
+ unsigned *final_assembly_size,
+ char **error_str);
+
+/**
+ * Compile a tessellation evaluation shader.
+ *
+ * Returns the final assembly and the program's size.
+ */
+const unsigned *
+brw_compile_tes(const struct brw_compiler *compiler, void *log_data,
+ void *mem_ctx,
+ const struct brw_tes_prog_key *key,
+ struct brw_tes_prog_data *prog_data,
+ const struct nir_shader *shader,
+ struct gl_shader_program *shader_prog,
+ int shader_time_index,
+ unsigned *final_assembly_size,
+ char **error_str);
+
/**
* Compile a vertex shader.
*
struct gl_program *prog,
int shader_time_index8,
int shader_time_index16,
+ bool allow_spilling,
bool use_rep_send,
unsigned *final_assembly_size,
char **error_str);
unsigned *final_assembly_size,
char **error_str);
+static inline uint32_t
+encode_slm_size(const struct brw_device_info *devinfo, uint32_t bytes)
+{
+ uint32_t slm_size = 0;
+
+ /* Shared Local Memory is specified as powers of two, and encoded in
+ * INTERFACE_DESCRIPTOR_DATA with the following representations:
+ *
+ * Size | 0 kB | 1 kB | 2 kB | 4 kB | 8 kB | 16 kB | 32 kB | 64 kB |
+ * -------------------------------------------------------------------
+ * Gen7-8 | 0 | none | none | 1 | 2 | 4 | 8 | 16 |
+ * -------------------------------------------------------------------
+ * Gen9+ | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
+ */
+ assert(bytes <= 64 * 1024);
+
+ if (bytes > 0) {
+ /* Shared Local Memory Size is specified as powers of two. */
+ slm_size = util_next_power_of_two(bytes);
+
+ if (devinfo->gen >= 9) {
+ /* Use a minimum of 1kB; turn an exponent of 10 (1024 kB) into 1. */
+ slm_size = ffs(MAX2(slm_size, 1024)) - 10;
+ } else {
+ /* Use a minimum of 4kB; convert to the pre-Gen9 representation. */
+ slm_size = MAX2(slm_size, 4096) / 4096;
+ }
+ }
+
+ return slm_size;
+}
+
#ifdef __cplusplus
} /* extern "C" */
#endif