#include "brw_wm.h"
#include "brw_shader.h"
}
+#include "gen8_generator.h"
#include "glsl/glsl_types.h"
#include "glsl/ir.h"
+#define MAX_SAMPLER_MESSAGE_SIZE 11
+
class bblock_t;
namespace {
struct acp_entry;
bool is_one() const;
bool is_null() const;
bool is_valid_3src() const;
+ bool is_contiguous() const;
fs_reg retype(uint32_t type);
+ fs_reg &apply_stride(unsigned stride);
/** Register file: GRF, MRF, IMM. */
enum register_file file;
int type;
bool negate;
bool abs;
- bool sechalf;
struct brw_reg fixed_hw_reg;
- int smear; /* -1, or a channel of the reg to smear to all channels. */
+
+ /** Smear a channel of the reg to all channels. */
+ fs_reg &set_smear(unsigned subreg);
/** Value for file == IMM */
union {
float f;
} imm;
+ /**
+ * Offset in bytes from the start of the register. Values up to a
+ * backend_reg::reg_offset unit are valid.
+ */
+ int subreg_offset;
+
+ /** Register region horizontal stride */
+ int stride;
+
fs_reg *reladdr;
};
+static inline fs_reg
+offset(fs_reg reg, unsigned delta)
+{
+ assert(delta == 0 || (reg.file != HW_REG && reg.file != IMM));
+ reg.reg_offset += delta;
+ return reg;
+}
+
+static inline fs_reg
+byte_offset(fs_reg reg, unsigned delta)
+{
+ assert(delta == 0 || (reg.file != HW_REG && reg.file != IMM));
+ reg.subreg_offset += delta;
+ return reg;
+}
+
+/**
+ * Get either of the 8-component halves of a 16-component register.
+ *
+ * Note: this also works if \c reg represents a SIMD16 pair of registers.
+ */
+static inline fs_reg
+half(const fs_reg ®, unsigned idx)
+{
+ assert(idx < 2);
+ assert(idx == 0 || (reg.file != HW_REG && reg.file != IMM));
+ return byte_offset(reg, 8 * idx * reg.stride * type_sz(reg.type));
+}
+
static const fs_reg reg_undef;
static const fs_reg reg_null_f(retype(brw_null_reg(), BRW_REGISTER_TYPE_F));
static const fs_reg reg_null_d(retype(brw_null_reg(), BRW_REGISTER_TYPE_D));
fs_inst *MAD(fs_reg dst, fs_reg c, fs_reg b, fs_reg a);
fs_inst *ADDC(fs_reg dst, fs_reg src0, fs_reg src1);
fs_inst *SUBB(fs_reg dst, fs_reg src0, fs_reg src1);
+ fs_inst *SEL(fs_reg dst, fs_reg src0, fs_reg src1);
int type_size(const struct glsl_type *type);
fs_inst *get_instruction_generating_reg(fs_inst *start,
void setup_pull_constants();
void invalidate_live_intervals();
void calculate_live_intervals();
+ void calculate_register_pressure();
bool opt_algebraic();
bool opt_cse();
bool opt_cse_local(bblock_t *block, exec_list *aeb);
bool try_constant_propagate(fs_inst *inst, acp_entry *entry);
bool opt_copy_propagate_local(void *mem_ctx, bblock_t *block,
exec_list *acp);
- bool register_coalesce_2();
+ bool register_coalesce();
bool compute_to_mrf();
bool dead_code_eliminate();
bool dead_code_eliminate_local();
fs_reg *emit_fragcoord_interpolation(ir_variable *ir);
fs_inst *emit_linterp(const fs_reg &attr, const fs_reg &interp,
glsl_interp_qualifier interpolation_mode,
- bool is_centroid);
+ bool is_centroid, bool is_sample);
fs_reg *emit_frontfacing_interpolation(ir_variable *ir);
fs_reg *emit_samplepos_setup(ir_variable *ir);
fs_reg *emit_sampleid_setup(ir_variable *ir);
+ fs_reg *emit_samplemaskin_setup(ir_variable *ir);
fs_reg *emit_general_interpolation(ir_variable *ir);
void emit_interpolation_setup_gen4();
void emit_interpolation_setup_gen6();
fs_reg sample_index);
fs_inst *emit_texture_gen7(ir_texture *ir, fs_reg dst, fs_reg coordinate,
fs_reg shadow_comp, fs_reg lod, fs_reg lod2,
- fs_reg sample_index);
+ fs_reg sample_index, fs_reg mcs, int sampler);
+ fs_reg emit_mcs_fetch(ir_texture *ir, fs_reg coordinate, int sampler);
+ void emit_gen6_gather_wa(uint8_t wa, fs_reg dst);
fs_reg fix_math_operand(fs_reg src);
fs_inst *emit_math(enum opcode op, fs_reg dst, fs_reg src0);
fs_inst *emit_math(enum opcode op, fs_reg dst, fs_reg src0, fs_reg src1);
bool try_emit_saturate(ir_expression *ir);
bool try_emit_mad(ir_expression *ir, int mul_arg);
void try_replace_with_sel();
+ bool opt_peephole_sel();
+ bool opt_peephole_predicated_break();
+ bool opt_saturate_propagation();
void emit_bool_to_cond_code(ir_rvalue *condition);
void emit_if_gen6(ir_if *ir);
void emit_unspill(fs_inst *inst, fs_reg reg, uint32_t spill_offset,
void setup_builtin_uniform_values(ir_variable *ir);
int implied_mrf_writes(fs_inst *inst);
+ virtual void dump_instructions();
void dump_instruction(backend_instruction *inst);
void visit_atomic_counter_intrinsic(ir_call *ir);
int *virtual_grf_end;
brw::fs_live_variables *live_intervals;
+ int *regs_live_at_ip;
+
/* This is the map from UNIFORM hw_reg + reg_offset as generated by
* the visitor to the packed uniform number after
* remove_dead_constants() that represents the actual uploaded
const unsigned *generate_assembly(exec_list *simd8_instructions,
exec_list *simd16_instructions,
- unsigned *assembly_size);
+ unsigned *assembly_size,
+ FILE *dump_file = NULL);
private:
- void generate_code(exec_list *instructions);
+ void generate_code(exec_list *instructions, FILE *dump_file);
void generate_fb_write(fs_inst *inst);
+ void generate_blorp_fb_write(fs_inst *inst);
void generate_pixel_xy(struct brw_reg dst, bool is_x);
void generate_linterp(fs_inst *inst, struct brw_reg dst,
struct brw_reg *src);
struct brw_reg dst,
struct brw_reg surf_index);
- void mark_surface_used(unsigned surf_index);
-
void patch_discard_jumps_to_fb_writes();
struct brw_context *brw;
struct brw_wm_compile *c;
struct gl_shader_program *prog;
- struct gl_shader *shader;
const struct gl_fragment_program *fp;
unsigned dispatch_width; /**< 8 or 16 */
void *mem_ctx;
};
+/**
+ * The fragment shader code generator.
+ *
+ * Translates FS IR to actual i965 assembly code.
+ */
+class gen8_fs_generator : public gen8_generator
+{
+public:
+ gen8_fs_generator(struct brw_context *brw,
+ struct brw_wm_compile *c,
+ struct gl_shader_program *prog,
+ struct gl_fragment_program *fp,
+ bool dual_source_output);
+ ~gen8_fs_generator();
+
+ const unsigned *generate_assembly(exec_list *simd8_instructions,
+ exec_list *simd16_instructions,
+ unsigned *assembly_size);
+
+private:
+ void generate_code(exec_list *instructions);
+ void generate_fb_write(fs_inst *inst);
+ void generate_linterp(fs_inst *inst, struct brw_reg dst,
+ struct brw_reg *src);
+ void generate_tex(fs_inst *inst, struct brw_reg dst, struct brw_reg src);
+ void generate_math1(fs_inst *inst, struct brw_reg dst, struct brw_reg src);
+ void generate_math2(fs_inst *inst, struct brw_reg dst,
+ struct brw_reg src0, struct brw_reg src1);
+ void generate_ddx(fs_inst *inst, struct brw_reg dst, struct brw_reg src);
+ void generate_ddy(fs_inst *inst, struct brw_reg dst, struct brw_reg src,
+ bool negate_value);
+ void generate_scratch_write(fs_inst *inst, struct brw_reg src);
+ void generate_scratch_read(fs_inst *inst, struct brw_reg dst);
+ void generate_scratch_read_gen7(fs_inst *inst, struct brw_reg dst);
+ void generate_uniform_pull_constant_load(fs_inst *inst,
+ struct brw_reg dst,
+ struct brw_reg index,
+ struct brw_reg offset);
+ void generate_varying_pull_constant_load(fs_inst *inst,
+ struct brw_reg dst,
+ struct brw_reg index,
+ struct brw_reg offset);
+ void generate_mov_dispatch_to_flags(fs_inst *ir);
+ void generate_set_simd4x2_offset(fs_inst *ir,
+ struct brw_reg dst,
+ struct brw_reg offset);
+ void generate_discard_jump(fs_inst *ir);
+
+ void patch_discard_jumps_to_fb_writes();
+
+ void mark_surface_used(unsigned surf_index);
+
+ struct brw_wm_compile *c;
+ const struct gl_fragment_program *fp;
+
+ unsigned dispatch_width; /** 8 or 16 */
+
+ bool dual_source_output;
+
+ exec_list discard_halt_patches;
+};
+
bool brw_do_channel_expressions(struct exec_list *instructions);
bool brw_do_vector_splitting(struct exec_list *instructions);
bool brw_fs_precompile(struct gl_context *ctx, struct gl_shader_program *prog);