/* * Copyright (C) 2019 Connor Abbott * Copyright (C) 2019 Lyude Paul * Copyright (C) 2019 Ryan Houdek * * Permission is hereby granted, free of charge, to any person obtaining a * copy of this software and associated documentation files (the "Software"), * to deal in the Software without restriction, including without limitation * the rights to use, copy, modify, merge, publish, distribute, sublicense, * and/or sell copies of the Software, and to permit persons to whom the * Software is furnished to do so, subject to the following conditions: * * The above copyright notice and this permission notice (including the next * paragraph) shall be included in all copies or substantial portions of the * Software. * * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, 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. */ #ifndef __bifrost_h__ #define __bifrost_h__ #include #include struct bifrost_header { unsigned unk0 : 7; // If true, convert any infinite result of any floating-point operation to // the biggest representable number. unsigned suppress_inf: 1; // Convert any NaN results to 0. unsigned suppress_nan : 1; unsigned unk1 : 2; // true if the execution mask of the next clause is the same as the mask of // the current clause. unsigned back_to_back : 1; unsigned no_end_of_shader: 1; unsigned unk2 : 2; // Set to true for fragment shaders, to implement this bit of spec text // from section 7.1.5 of the GLSL ES spec: // // "Stores to image and buffer variables performed by helper invocations // have no effect on the underlying image or buffer memory." // // Helper invocations are threads (invocations) corresponding to pixels in // a quad that aren't actually part of the triangle, but are included to // make derivatives work correctly. They're usually turned on, but they // need to be masked off for GLSL-level stores. This bit seems to be the // only bit that's actually different between fragment shaders and other // shaders, so this is probably what it's doing. unsigned elide_writes : 1; // If backToBack is off: // - true for conditional branches and fallthrough // - false for unconditional branches // The blob seems to always set it to true if back-to-back is on. unsigned branch_cond : 1; // This bit is set when the next clause writes to the data register of some // previous clause. unsigned datareg_writebarrier: 1; unsigned datareg : 6; unsigned scoreboard_deps: 8; unsigned scoreboard_index: 3; unsigned clause_type: 4; unsigned unk3 : 1; // part of clauseType? unsigned next_clause_type: 4; unsigned unk4 : 1; // part of nextClauseType? } __attribute__((packed)); enum bifrost_packed_src { BIFROST_SRC_PORT0 = 0, BIFROST_SRC_PORT1 = 1, BIFROST_SRC_PORT3 = 2, BIFROST_SRC_STAGE = 3, BIFROST_SRC_CONST_LO = 4, BIFROST_SRC_CONST_HI = 5, BIFROST_SRC_PASS_FMA = 6, BIFROST_SRC_PASS_ADD = 7, }; struct bifrost_fma_inst { unsigned src0 : 3; unsigned op : 20; } __attribute__((packed)); struct bifrost_add_inst { unsigned src0 : 3; unsigned op : 17; } __attribute__((packed)); enum bifrost_outmod { BIFROST_NONE = 0x0, BIFROST_POS = 0x1, BIFROST_SAT_SIGNED = 0x2, BIFROST_SAT = 0x3, }; enum bifrost_roundmode { BIFROST_RTE = 0x0, /* round to even */ BIFROST_RTP = 0x1, /* round to positive */ BIFROST_RTN = 0x2, /* round to negative */ BIFROST_RTZ = 0x3 /* round to zero */ }; /* NONE: Same as fmax() and fmin() -- return the other * number if any number is NaN. Also always return +0 if * one argument is +0 and the other is -0. * * NAN_WINS: Instead of never returning a NaN, always return * one. The "greater"/"lesser" NaN is always returned, first * by checking the sign and then the mantissa bits. * * SRC1_WINS: For max, implement src0 > src1 ? src0 : src1. * For min, implement src0 < src1 ? src0 : src1. This * includes handling NaN's and signedness of 0 differently * from above, since +0 and -0 compare equal and comparisons * always return false for NaN's. As a result, this mode is * *not* commutative. * * SRC0_WINS: For max, implement src0 < src1 ? src1 : src0 * For min, implement src0 > src1 ? src1 : src0 */ enum bifrost_minmax_mode { BIFROST_MINMAX_NONE = 0x0, BIFROST_NAN_WINS = 0x1, BIFROST_SRC1_WINS = 0x2, BIFROST_SRC0_WINS = 0x3, }; #define BIFROST_FMA_OP_FADD32 (0x58 >> 2) struct bifrost_fma_add { unsigned src0 : 3; unsigned src1 : 3; unsigned src1_abs : 1; unsigned src0_neg : 1; unsigned src1_neg : 1; unsigned unk : 3; unsigned src0_abs : 1; enum bifrost_outmod outmod : 2; enum bifrost_roundmode roundmode : 2; unsigned op : 6; } __attribute__((packed)); #define BIFROST_FMA_OP_FMA (0x00) struct bifrost_fma_fma { unsigned src0 : 3; unsigned src1 : 3; unsigned src2 : 3; unsigned src_expand : 3; unsigned src0_abs : 1; unsigned unk : 4; unsigned src0_neg : 1; unsigned src2_neg : 1; unsigned src1_abs : 1; unsigned src2_abs : 1; unsigned op : 5; } __attribute__((packed)); enum bifrost_csel_cond { BIFROST_FEQ_F = 0x0, BIFROST_FGT_F = 0x1, BIFROST_FGE_F = 0x2, BIFROST_IEQ_F = 0x3, BIFROST_IGT_I = 0x4, BIFROST_IGE_I = 0x5, BIFROST_UGT_I = 0x6, BIFROST_UGE_I = 0x7 }; struct bifrost_csel4 { unsigned src0 : 3; unsigned src1 : 3; unsigned src2 : 3; unsigned src3 : 3; enum bifrost_csel_cond cond : 3; unsigned op : 8; } __attribute__((packed)); struct bifrost_shift_fma { unsigned src0 : 3; unsigned src1 : 3; unsigned src2 : 3; unsigned half : 3; /* 000 for i32, 100 for i8, 111 for v2i16 */ unsigned unk : 1; /* always set? */ unsigned invert_1 : 1; /* Inverts sources to combining op */ /* For XOR, switches RSHIFT to LSHIFT since only one invert needed */ unsigned invert_2 : 1; unsigned op : 8; } __attribute__((packed)); struct bifrost_shift_add { unsigned src0 : 3; unsigned src1 : 3; unsigned src2 : 3; unsigned zero : 2; unsigned invert_1 : 1; unsigned invert_2 : 1; unsigned op : 7; } __attribute__((packed)); enum bifrost_ldst_type { BIFROST_LDST_F16 = 0, BIFROST_LDST_F32 = 1, BIFROST_LDST_I32 = 2, BIFROST_LDST_U32 = 3 }; struct bifrost_ld_var_addr { unsigned src0 : 3; unsigned src1 : 3; unsigned location : 5; enum bifrost_ldst_type type : 2; unsigned op : 7; } __attribute__((packed)); struct bifrost_ld_attr { unsigned src0 : 3; unsigned src1 : 3; unsigned location : 5; unsigned channels : 2; /* MALI_POSITIVE */ enum bifrost_ldst_type type : 2; unsigned op : 5; } __attribute__((packed)); enum bifrost_interp_mode { BIFROST_INTERP_PER_FRAG = 0x0, BIFROST_INTERP_CENTROID = 0x1, BIFROST_INTERP_DEFAULT = 0x2, BIFROST_INTERP_EXPLICIT = 0x3 }; #define BIFROST_ADD_OP_LD_VAR_16 (0x1a << 1) #define BIFROST_ADD_OP_LD_VAR_32 (0x0a << 1) struct bifrost_ld_var { unsigned src0 : 3; /* If top two bits set, indirect with src in bottom three */ unsigned addr : 5; unsigned channels : 2; /* MALI_POSITIVE */ enum bifrost_interp_mode interp_mode : 2; unsigned reuse : 1; unsigned flat : 1; unsigned op : 6; } __attribute__((packed)); struct bifrost_tex_ctrl { unsigned sampler_index : 4; // also used to signal indirects unsigned tex_index : 7; bool no_merge_index : 1; // whether to merge (direct) sampler & texture indices bool filter : 1; // use the usual filtering pipeline (0 for texelFetch & textureGather) unsigned unk0 : 2; bool texel_offset : 1; // *Offset() bool is_shadow : 1; bool is_array : 1; unsigned tex_type : 2; // 2D, 3D, Cube, Buffer bool compute_lod : 1; // 0 for *Lod() bool not_supply_lod : 1; // 0 for *Lod() or when a bias is applied bool calc_gradients : 1; // 0 for *Grad() unsigned unk1 : 1; unsigned result_type : 4; // integer, unsigned, float TODO: why is this 4 bits? unsigned unk2 : 4; } __attribute__((packed)); struct bifrost_dual_tex_ctrl { unsigned sampler_index0 : 2; unsigned unk0 : 2; unsigned tex_index0 : 2; unsigned sampler_index1 : 2; unsigned tex_index1 : 2; unsigned unk1 : 22; } __attribute__((packed)); enum branch_bit_size { BR_SIZE_32 = 0, BR_SIZE_16XX = 1, BR_SIZE_16YY = 2, // For the above combinations of bitsize and location, an extra bit is // encoded via comparing the sources. The only possible source of ambiguity // would be if the sources were the same, but then the branch condition // would be always true or always false anyways, so we can ignore it. But // this no longer works when comparing the y component to the x component, // since it's valid to compare the y component of a source against its own // x component. Instead, the extra bit is encoded via an extra bitsize. BR_SIZE_16YX0 = 3, BR_SIZE_16YX1 = 4, BR_SIZE_32_AND_16X = 5, BR_SIZE_32_AND_16Y = 6, // Used for comparisons with zero and always-true, see below. I think this // only works for integer comparisons. BR_SIZE_ZERO = 7, }; enum bifrost_reg_write_unit { REG_WRITE_NONE = 0, // don't write REG_WRITE_TWO, // write using reg2 REG_WRITE_THREE, // write using reg3 }; struct bifrost_regs { unsigned uniform_const : 8; unsigned reg2 : 6; unsigned reg3 : 6; unsigned reg0 : 5; unsigned reg1 : 6; unsigned ctrl : 4; } __attribute__((packed)); enum bifrost_branch_cond { BR_COND_LT = 0, BR_COND_LE = 1, BR_COND_GE = 2, BR_COND_GT = 3, // Equal vs. not-equal determined by src0/src1 comparison BR_COND_EQ = 4, // floating-point comparisons // Becomes UNE when you flip the arguments BR_COND_OEQ = 5, // TODO what happens when you flip the arguments? BR_COND_OGT = 6, BR_COND_OLT = 7, }; enum bifrost_branch_code { BR_ALWAYS = 63, }; struct bifrost_branch { unsigned src0 : 3; /* For BR_SIZE_ZERO, upper two bits become ctrl */ unsigned src1 : 3; /* Offset source -- always uniform/const but * theoretically could support indirect jumps? */ unsigned src2 : 3; enum bifrost_branch_cond cond : 3; enum branch_bit_size size : 3; unsigned op : 5; }; /* Clause packing */ #define BIFROST_FMA_NOP (0x701960) #define BIFROST_ADD_NOP (0x3D960) struct bifrost_fmt1 { unsigned ins_0 : 3; unsigned tag : 5; uint64_t ins_1 : 64; unsigned ins_2 : 11; uint64_t header : 45; } __attribute__((packed)); #define BIFROST_FMT1_INSTRUCTIONS 0b00101 #define BIFROST_FMT1_FINAL 0b01001 #define BIFROST_FMT1_CONSTANTS 0b00001 enum bifrost_reg_control { BIFROST_WRITE_FMA_P2 = 1, BIFROST_WRITE_FMA_P2_READ_P3 = 2, BIFROST_WRITE_FMA_P2_READ_P3_ALT = 3, BIFROST_READ_P3 = 4, BIFROST_WRITE_ADD_P2 = 5, BIFROST_WRITE_ADD_P2_READ_P3 = 6, BIFROST_WRITE_ADD_P2_FMA_P3 = 7, BIFROST_FIRST_NONE = 8, BIFROST_FIRST_WRITE_FMA_P2 = 9, BIFROST_REG_NONE = 11, BIFROST_FIRST_READ_P3 = 12, BIFROST_FIRST_WRITE_ADD_P2 = 13, BIFROST_FIRST_WRITE_ADD_P2_READ_P3 = 14, BIFROST_FIRST_WRITE_ADD_P2_FMA_P3 = 15 }; #endif