/* * Copyright © 2016 Broadcom * * 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. */ #include #include "util/macros.h" #include "broadcom/common/v3d_device_info.h" #include "qpu_instr.h" #ifndef QPU_MASK #define QPU_MASK(high, low) ((((uint64_t)1<<((high)-(low)+1))-1)<<(low)) /* Using the GNU statement expression extension */ #define QPU_SET_FIELD(value, field) \ ({ \ uint64_t fieldval = (uint64_t)(value) << field ## _SHIFT; \ assert((fieldval & ~ field ## _MASK) == 0); \ fieldval & field ## _MASK; \ }) #define QPU_GET_FIELD(word, field) ((uint32_t)(((word) & field ## _MASK) >> field ## _SHIFT)) #define QPU_UPDATE_FIELD(inst, value, field) \ (((inst) & ~(field ## _MASK)) | QPU_SET_FIELD(value, field)) #endif /* QPU_MASK */ #define VC5_QPU_OP_MUL_SHIFT 58 #define VC5_QPU_OP_MUL_MASK QPU_MASK(63, 58) #define VC5_QPU_SIG_SHIFT 53 #define VC5_QPU_SIG_MASK QPU_MASK(57, 53) # define VC5_QPU_SIG_THRSW_BIT 0x1 # define VC5_QPU_SIG_LDUNIF_BIT 0x2 # define VC5_QPU_SIG_LDTMU_BIT 0x4 # define VC5_QPU_SIG_LDVARY_BIT 0x8 #define VC5_QPU_COND_SHIFT 46 #define VC5_QPU_COND_MASK QPU_MASK(52, 46) #define VC5_QPU_COND_IFA 0 #define VC5_QPU_COND_IFB 1 #define VC5_QPU_COND_IFNA 2 #define VC5_QPU_COND_IFNB 3 #define VC5_QPU_MM QPU_MASK(45, 45) #define VC5_QPU_MA QPU_MASK(44, 44) #define V3D_QPU_WADDR_M_SHIFT 38 #define V3D_QPU_WADDR_M_MASK QPU_MASK(43, 38) #define VC5_QPU_BRANCH_ADDR_LOW_SHIFT 35 #define VC5_QPU_BRANCH_ADDR_LOW_MASK QPU_MASK(55, 35) #define V3D_QPU_WADDR_A_SHIFT 32 #define V3D_QPU_WADDR_A_MASK QPU_MASK(37, 32) #define VC5_QPU_BRANCH_COND_SHIFT 32 #define VC5_QPU_BRANCH_COND_MASK QPU_MASK(34, 32) #define VC5_QPU_BRANCH_ADDR_HIGH_SHIFT 24 #define VC5_QPU_BRANCH_ADDR_HIGH_MASK QPU_MASK(31, 24) #define VC5_QPU_OP_ADD_SHIFT 24 #define VC5_QPU_OP_ADD_MASK QPU_MASK(31, 24) #define VC5_QPU_MUL_B_SHIFT 21 #define VC5_QPU_MUL_B_MASK QPU_MASK(23, 21) #define VC5_QPU_BRANCH_MSFIGN_SHIFT 21 #define VC5_QPU_BRANCH_MSFIGN_MASK QPU_MASK(22, 21) #define VC5_QPU_MUL_A_SHIFT 18 #define VC5_QPU_MUL_A_MASK QPU_MASK(20, 18) #define VC5_QPU_ADD_B_SHIFT 15 #define VC5_QPU_ADD_B_MASK QPU_MASK(17, 15) #define VC5_QPU_BRANCH_BDU_SHIFT 15 #define VC5_QPU_BRANCH_BDU_MASK QPU_MASK(17, 15) #define VC5_QPU_BRANCH_UB QPU_MASK(14, 14) #define VC5_QPU_ADD_A_SHIFT 12 #define VC5_QPU_ADD_A_MASK QPU_MASK(14, 12) #define VC5_QPU_BRANCH_BDI_SHIFT 12 #define VC5_QPU_BRANCH_BDI_MASK QPU_MASK(13, 12) #define VC5_QPU_RADDR_A_SHIFT 6 #define VC5_QPU_RADDR_A_MASK QPU_MASK(11, 6) #define VC5_QPU_RADDR_B_SHIFT 0 #define VC5_QPU_RADDR_B_MASK QPU_MASK(5, 0) #define THRSW .thrsw = true #define LDUNIF .ldunif = true #define LDTMU .ldtmu = true #define LDVARY .ldvary = true #define LDVPM .ldvpm = true #define SMIMM .small_imm = true #define LDTLB .ldtlb = true #define LDTLBU .ldtlbu = true #define UCB .ucb = true #define ROT .rotate = true #define WRTMUC .wrtmuc = true static const struct v3d_qpu_sig v33_sig_map[] = { /* MISC R3 R4 R5 */ [0] = { }, [1] = { THRSW, }, [2] = { LDUNIF }, [3] = { THRSW, LDUNIF }, [4] = { LDTMU, }, [5] = { THRSW, LDTMU, }, [6] = { LDTMU, LDUNIF }, [7] = { THRSW, LDTMU, LDUNIF }, [8] = { LDVARY, }, [9] = { THRSW, LDVARY, }, [10] = { LDVARY, LDUNIF }, [11] = { THRSW, LDVARY, LDUNIF }, [12] = { LDVARY, LDTMU, }, [13] = { THRSW, LDVARY, LDTMU, }, [14] = { SMIMM, LDVARY, }, [15] = { SMIMM, }, [16] = { LDTLB, }, [17] = { LDTLBU, }, /* 18-21 reserved */ [22] = { UCB, }, [23] = { ROT, }, [24] = { LDVPM, }, [25] = { THRSW, LDVPM, }, [26] = { LDVPM, LDUNIF }, [27] = { THRSW, LDVPM, LDUNIF }, [28] = { LDVPM, LDTMU, }, [29] = { THRSW, LDVPM, LDTMU, }, [30] = { SMIMM, LDVPM, }, [31] = { SMIMM, }, }; bool v3d_qpu_sig_unpack(const struct v3d_device_info *devinfo, uint32_t packed_sig, struct v3d_qpu_sig *sig) { if (packed_sig >= ARRAY_SIZE(v33_sig_map)) return false; *sig = v33_sig_map[packed_sig]; /* Signals with zeroed unpacked contents after element 0 are reserved. */ return (packed_sig == 0 || memcmp(sig, &v33_sig_map[0], sizeof(*sig) != 0)); } bool v3d_qpu_sig_pack(const struct v3d_device_info *devinfo, const struct v3d_qpu_sig *sig, uint32_t *packed_sig) { static const struct v3d_qpu_sig *map; map = v33_sig_map; for (int i = 0; i < ARRAY_SIZE(v33_sig_map); i++) { if (memcmp(&map[i], sig, sizeof(*sig)) == 0) { *packed_sig = i; return true; } } return false; } bool v3d_qpu_flags_unpack(const struct v3d_device_info *devinfo, uint32_t packed_cond, struct v3d_qpu_flags *cond) { static const enum v3d_qpu_cond cond_map[4] = { [0] = V3D_QPU_COND_IFA, [1] = V3D_QPU_COND_IFB, [2] = V3D_QPU_COND_IFNA, [3] = V3D_QPU_COND_IFNB, }; cond->ac = V3D_QPU_COND_NONE; cond->mc = V3D_QPU_COND_NONE; cond->apf = V3D_QPU_PF_NONE; cond->mpf = V3D_QPU_PF_NONE; cond->auf = V3D_QPU_UF_NONE; cond->muf = V3D_QPU_UF_NONE; if (packed_cond == 0) { return true; } else if (packed_cond >> 2 == 0) { cond->apf = packed_cond & 0x3; } else if (packed_cond >> 4 == 0) { cond->auf = (packed_cond & 0xf) - 4 + V3D_QPU_UF_ANDZ; } else if (packed_cond == 0x10) { return false; } else if (packed_cond >> 2 == 0x4) { cond->mpf = packed_cond & 0x3; } else if (packed_cond >> 4 == 0x1) { cond->muf = (packed_cond & 0xf) - 4 + V3D_QPU_UF_ANDZ; } else if (packed_cond >> 4 == 0x2) { cond->ac = ((packed_cond >> 2) & 0x3) + V3D_QPU_COND_IFA; cond->mpf = packed_cond & 0x3; } else if (packed_cond >> 4 == 0x3) { cond->mc = ((packed_cond >> 2) & 0x3) + V3D_QPU_COND_IFA; cond->apf = packed_cond & 0x3; } else if (packed_cond >> 6) { cond->mc = cond_map[(packed_cond >> 4) & 0x3]; if (((packed_cond >> 2) & 0x3) == 0) { cond->ac = cond_map[packed_cond & 0x3]; } else { cond->auf = (packed_cond & 0xf) - 4 + V3D_QPU_UF_ANDZ; } } return true; } bool v3d_qpu_flags_pack(const struct v3d_device_info *devinfo, const struct v3d_qpu_flags *cond, uint32_t *packed_cond) { #define AC (1 << 0) #define MC (1 << 1) #define APF (1 << 2) #define MPF (1 << 3) #define AUF (1 << 4) #define MUF (1 << 5) static const struct { uint8_t flags_present; uint8_t bits; } flags_table[] = { { 0, 0 }, { APF, 0 }, { AUF, 0 }, { MPF, (1 << 4) }, { MUF, (1 << 4) }, { AC, (1 << 5) }, { AC | MPF, (1 << 5) }, { MC, (1 << 5) | (1 << 4) }, { MC | APF, (1 << 5) | (1 << 4) }, { MC | AC, (1 << 6) }, { MC | AUF, (1 << 6) }, }; uint8_t flags_present = 0; if (cond->ac != V3D_QPU_COND_NONE) flags_present |= AC; if (cond->mc != V3D_QPU_COND_NONE) flags_present |= MC; if (cond->apf != V3D_QPU_PF_NONE) flags_present |= APF; if (cond->mpf != V3D_QPU_PF_NONE) flags_present |= MPF; if (cond->auf != V3D_QPU_UF_NONE) flags_present |= AUF; if (cond->muf != V3D_QPU_UF_NONE) flags_present |= MUF; for (int i = 0; i < ARRAY_SIZE(flags_table); i++) { if (flags_table[i].flags_present != flags_present) continue; *packed_cond = flags_table[i].bits; *packed_cond |= cond->apf; *packed_cond |= cond->mpf; if (flags_present & AUF) *packed_cond |= cond->auf - V3D_QPU_UF_ANDZ + 4; if (flags_present & MUF) *packed_cond |= cond->muf - V3D_QPU_UF_ANDZ + 4; if (flags_present & AC) *packed_cond |= (cond->ac - V3D_QPU_COND_IFA) << 2; if (flags_present & MC) { if (*packed_cond & (1 << 6)) *packed_cond |= (cond->mc - V3D_QPU_COND_IFA) << 4; else *packed_cond |= (cond->mc - V3D_QPU_COND_IFA) << 2; } return true; } return false; } /* Make a mapping of the table of opcodes in the spec. The opcode is * determined by a combination of the opcode field, and in the case of 0 or * 1-arg opcodes, the mux_b field as well. */ #define MUX_MASK(bot, top) (((1 << (top + 1)) - 1) - ((1 << (bot)) - 1)) #define ANYMUX MUX_MASK(0, 7) struct opcode_desc { uint8_t opcode_first; uint8_t opcode_last; uint8_t mux_b_mask; uint8_t mux_a_mask; uint8_t op; /* 0 if it's the same across V3D versions, or a specific V3D version. */ uint8_t ver; }; static const struct opcode_desc add_ops[] = { /* FADD is FADDNF depending on the order of the mux_a/mux_b. */ { 0, 47, ANYMUX, ANYMUX, V3D_QPU_A_FADD }, { 0, 47, ANYMUX, ANYMUX, V3D_QPU_A_FADDNF }, { 53, 55, ANYMUX, ANYMUX, V3D_QPU_A_VFPACK }, { 56, 56, ANYMUX, ANYMUX, V3D_QPU_A_ADD }, { 57, 59, ANYMUX, ANYMUX, V3D_QPU_A_VFPACK }, { 60, 60, ANYMUX, ANYMUX, V3D_QPU_A_SUB }, { 61, 63, ANYMUX, ANYMUX, V3D_QPU_A_VFPACK }, { 64, 111, ANYMUX, ANYMUX, V3D_QPU_A_FSUB }, { 120, 120, ANYMUX, ANYMUX, V3D_QPU_A_MIN }, { 121, 121, ANYMUX, ANYMUX, V3D_QPU_A_MAX }, { 122, 122, ANYMUX, ANYMUX, V3D_QPU_A_UMIN }, { 123, 123, ANYMUX, ANYMUX, V3D_QPU_A_UMAX }, { 124, 124, ANYMUX, ANYMUX, V3D_QPU_A_SHL }, { 125, 125, ANYMUX, ANYMUX, V3D_QPU_A_SHR }, { 126, 126, ANYMUX, ANYMUX, V3D_QPU_A_ASR }, { 127, 127, ANYMUX, ANYMUX, V3D_QPU_A_ROR }, /* FMIN is instead FMAX depending on the order of the mux_a/mux_b. */ { 128, 175, ANYMUX, ANYMUX, V3D_QPU_A_FMIN }, { 128, 175, ANYMUX, ANYMUX, V3D_QPU_A_FMAX }, { 176, 180, ANYMUX, ANYMUX, V3D_QPU_A_VFMIN }, { 181, 181, ANYMUX, ANYMUX, V3D_QPU_A_AND }, { 182, 182, ANYMUX, ANYMUX, V3D_QPU_A_OR }, { 183, 183, ANYMUX, ANYMUX, V3D_QPU_A_XOR }, { 184, 184, ANYMUX, ANYMUX, V3D_QPU_A_VADD }, { 185, 185, ANYMUX, ANYMUX, V3D_QPU_A_VSUB }, { 186, 186, 1 << 0, ANYMUX, V3D_QPU_A_NOT }, { 186, 186, 1 << 1, ANYMUX, V3D_QPU_A_NEG }, { 186, 186, 1 << 2, ANYMUX, V3D_QPU_A_FLAPUSH }, { 186, 186, 1 << 3, ANYMUX, V3D_QPU_A_FLBPUSH }, { 186, 186, 1 << 4, ANYMUX, V3D_QPU_A_FLBPOP }, { 186, 186, 1 << 6, ANYMUX, V3D_QPU_A_SETMSF }, { 186, 186, 1 << 7, ANYMUX, V3D_QPU_A_SETREVF }, { 187, 187, 1 << 0, 1 << 0, V3D_QPU_A_NOP, 0 }, { 187, 187, 1 << 0, 1 << 1, V3D_QPU_A_TIDX }, { 187, 187, 1 << 0, 1 << 2, V3D_QPU_A_EIDX }, { 187, 187, 1 << 0, 1 << 3, V3D_QPU_A_LR }, { 187, 187, 1 << 0, 1 << 4, V3D_QPU_A_VFLA }, { 187, 187, 1 << 0, 1 << 5, V3D_QPU_A_VFLNA }, { 187, 187, 1 << 0, 1 << 6, V3D_QPU_A_VFLB }, { 187, 187, 1 << 0, 1 << 7, V3D_QPU_A_VFLNB }, { 187, 187, 1 << 1, MUX_MASK(0, 2), V3D_QPU_A_FXCD }, { 187, 187, 1 << 1, 1 << 3, V3D_QPU_A_XCD }, { 187, 187, 1 << 1, MUX_MASK(4, 6), V3D_QPU_A_FYCD }, { 187, 187, 1 << 1, 1 << 7, V3D_QPU_A_YCD }, { 187, 187, 1 << 2, 1 << 0, V3D_QPU_A_MSF }, { 187, 187, 1 << 2, 1 << 1, V3D_QPU_A_REVF }, { 187, 187, 1 << 2, 1 << 2, V3D_QPU_A_VDWWT }, { 187, 187, 1 << 2, 1 << 5, V3D_QPU_A_TMUWT }, { 187, 187, 1 << 2, 1 << 6, V3D_QPU_A_VPMWT }, { 187, 187, 1 << 3, ANYMUX, V3D_QPU_A_VPMSETUP }, /* FIXME: MORE COMPLICATED */ /* { 190, 191, ANYMUX, ANYMUX, V3D_QPU_A_VFMOVABSNEGNAB }, */ { 192, 239, ANYMUX, ANYMUX, V3D_QPU_A_FCMP }, { 240, 244, ANYMUX, ANYMUX, V3D_QPU_A_VFMAX }, { 245, 245, MUX_MASK(0, 2), ANYMUX, V3D_QPU_A_FROUND }, { 245, 245, 1 << 3, ANYMUX, V3D_QPU_A_FTOIN }, { 245, 245, MUX_MASK(4, 6), ANYMUX, V3D_QPU_A_FTRUNC }, { 245, 245, 1 << 7, ANYMUX, V3D_QPU_A_FTOIZ }, { 246, 246, MUX_MASK(0, 2), ANYMUX, V3D_QPU_A_FFLOOR }, { 246, 246, 1 << 3, ANYMUX, V3D_QPU_A_FTOUZ }, { 246, 246, MUX_MASK(4, 6), ANYMUX, V3D_QPU_A_FCEIL }, { 246, 246, 1 << 7, ANYMUX, V3D_QPU_A_FTOC }, { 247, 247, MUX_MASK(0, 2), ANYMUX, V3D_QPU_A_FDX }, { 247, 247, MUX_MASK(4, 6), ANYMUX, V3D_QPU_A_FDY }, /* The stvpms are distinguished by the waddr field. */ { 248, 248, ANYMUX, ANYMUX, V3D_QPU_A_STVPMV }, { 248, 248, ANYMUX, ANYMUX, V3D_QPU_A_STVPMD }, { 248, 248, ANYMUX, ANYMUX, V3D_QPU_A_STVPMP }, { 252, 252, MUX_MASK(0, 2), ANYMUX, V3D_QPU_A_ITOF }, { 252, 252, 1 << 3, ANYMUX, V3D_QPU_A_CLZ }, { 252, 252, MUX_MASK(4, 6), ANYMUX, V3D_QPU_A_UTOF }, }; static const struct opcode_desc mul_ops[] = { { 1, 1, ANYMUX, ANYMUX, V3D_QPU_M_ADD }, { 2, 2, ANYMUX, ANYMUX, V3D_QPU_M_SUB }, { 3, 3, ANYMUX, ANYMUX, V3D_QPU_M_UMUL24 }, { 4, 8, ANYMUX, ANYMUX, V3D_QPU_M_VFMUL }, { 9, 9, ANYMUX, ANYMUX, V3D_QPU_M_SMUL24 }, { 10, 10, ANYMUX, ANYMUX, V3D_QPU_M_MULTOP }, { 14, 14, ANYMUX, ANYMUX, V3D_QPU_M_FMOV }, { 15, 15, MUX_MASK(0, 3), ANYMUX, V3D_QPU_M_FMOV }, { 15, 15, 1 << 4, 1 << 0, V3D_QPU_M_NOP, 0 }, { 15, 15, 1 << 7, ANYMUX, V3D_QPU_M_MOV }, { 16, 63, ANYMUX, ANYMUX, V3D_QPU_M_FMUL }, }; static const struct opcode_desc * lookup_opcode(const struct opcode_desc *opcodes, size_t num_opcodes, uint32_t opcode, uint32_t mux_a, uint32_t mux_b) { for (int i = 0; i < num_opcodes; i++) { const struct opcode_desc *op_desc = &opcodes[i]; if (opcode < op_desc->opcode_first || opcode > op_desc->opcode_last) continue; if (!(op_desc->mux_b_mask & (1 << mux_b))) continue; if (!(op_desc->mux_a_mask & (1 << mux_a))) continue; return op_desc; } return NULL; } static bool v3d_qpu_float32_unpack_unpack(uint32_t packed, enum v3d_qpu_input_unpack *unpacked) { switch (packed) { case 0: *unpacked = V3D_QPU_UNPACK_ABS; return true; case 1: *unpacked = V3D_QPU_UNPACK_NONE; return true; case 2: *unpacked = V3D_QPU_UNPACK_L; return true; case 3: *unpacked = V3D_QPU_UNPACK_H; return true; default: return false; } } static bool v3d_qpu_float32_unpack_pack(enum v3d_qpu_input_unpack unpacked, uint32_t *packed) { switch (unpacked) { case V3D_QPU_UNPACK_ABS: *packed = 0; return true; case V3D_QPU_UNPACK_NONE: *packed = 1; return true; case V3D_QPU_UNPACK_L: *packed = 2; return true; case V3D_QPU_UNPACK_H: *packed = 3; return true; default: return false; } } static bool v3d_qpu_float16_unpack_unpack(uint32_t packed, enum v3d_qpu_input_unpack *unpacked) { switch (packed) { case 0: *unpacked = V3D_QPU_UNPACK_NONE; return true; case 1: *unpacked = V3D_QPU_UNPACK_REPLICATE_32F_16; return true; case 2: *unpacked = V3D_QPU_UNPACK_REPLICATE_L_16; return true; case 3: *unpacked = V3D_QPU_UNPACK_REPLICATE_H_16; return true; case 4: *unpacked = V3D_QPU_UNPACK_SWAP_16; return true; default: return false; } } static bool v3d_qpu_float16_unpack_pack(enum v3d_qpu_input_unpack unpacked, uint32_t *packed) { switch (unpacked) { case V3D_QPU_UNPACK_NONE: *packed = 0; return true; case V3D_QPU_UNPACK_REPLICATE_32F_16: *packed = 1; return true; case V3D_QPU_UNPACK_REPLICATE_L_16: *packed = 2; return true; case V3D_QPU_UNPACK_REPLICATE_H_16: *packed = 3; return true; case V3D_QPU_UNPACK_SWAP_16: *packed = 4; return true; default: return false; } } static bool v3d_qpu_float32_pack_pack(enum v3d_qpu_input_unpack unpacked, uint32_t *packed) { switch (unpacked) { case V3D_QPU_PACK_NONE: *packed = 0; return true; case V3D_QPU_PACK_L: *packed = 1; return true; case V3D_QPU_PACK_H: *packed = 2; return true; default: return false; } } static bool v3d_qpu_add_unpack(const struct v3d_device_info *devinfo, uint64_t packed_inst, struct v3d_qpu_instr *instr) { uint32_t op = QPU_GET_FIELD(packed_inst, VC5_QPU_OP_ADD); uint32_t mux_a = QPU_GET_FIELD(packed_inst, VC5_QPU_ADD_A); uint32_t mux_b = QPU_GET_FIELD(packed_inst, VC5_QPU_ADD_B); uint32_t waddr = QPU_GET_FIELD(packed_inst, V3D_QPU_WADDR_A); uint32_t map_op = op; /* Some big clusters of opcodes are replicated with unpack * flags */ if (map_op >= 249 && map_op <= 251) map_op = (map_op - 249 + 245); if (map_op >= 253 && map_op <= 255) map_op = (map_op - 253 + 245); const struct opcode_desc *desc = lookup_opcode(add_ops, ARRAY_SIZE(add_ops), map_op, mux_a, mux_b); if (!desc) return false; instr->alu.add.op = desc->op; /* FADD/FADDNF and FMIN/FMAX are determined by the orders of the * operands. */ if (((op >> 2) & 3) * 8 + mux_a > (op & 3) * 8 + mux_b) { if (instr->alu.add.op == V3D_QPU_A_FMIN) instr->alu.add.op = V3D_QPU_A_FMAX; if (instr->alu.add.op == V3D_QPU_A_FADD) instr->alu.add.op = V3D_QPU_A_FADDNF; } /* Some QPU ops require a bit more than just basic opcode and mux a/b * comparisons to distinguish them. */ switch (instr->alu.add.op) { case V3D_QPU_A_STVPMV: case V3D_QPU_A_STVPMD: case V3D_QPU_A_STVPMP: switch (waddr) { case 0: instr->alu.add.op = V3D_QPU_A_STVPMV; break; case 1: instr->alu.add.op = V3D_QPU_A_STVPMD; break; case 2: instr->alu.add.op = V3D_QPU_A_STVPMP; break; default: return false; } break; default: break; } switch (instr->alu.add.op) { case V3D_QPU_A_FADD: case V3D_QPU_A_FADDNF: case V3D_QPU_A_FSUB: case V3D_QPU_A_FMIN: case V3D_QPU_A_FMAX: case V3D_QPU_A_FCMP: instr->alu.add.output_pack = (op >> 4) & 0x3; if (!v3d_qpu_float32_unpack_unpack((op >> 2) & 0x3, &instr->alu.add.a_unpack)) { return false; } if (!v3d_qpu_float32_unpack_unpack((op >> 0) & 0x3, &instr->alu.add.b_unpack)) { return false; } break; case V3D_QPU_A_FFLOOR: case V3D_QPU_A_FROUND: case V3D_QPU_A_FTRUNC: case V3D_QPU_A_FCEIL: case V3D_QPU_A_FDX: case V3D_QPU_A_FDY: instr->alu.add.output_pack = mux_b & 0x3; if (!v3d_qpu_float32_unpack_unpack((op >> 2) & 0x3, &instr->alu.add.a_unpack)) { return false; } break; case V3D_QPU_A_FTOIN: case V3D_QPU_A_FTOIZ: case V3D_QPU_A_FTOUZ: case V3D_QPU_A_FTOC: instr->alu.add.output_pack = V3D_QPU_PACK_NONE; if (!v3d_qpu_float32_unpack_unpack((op >> 2) & 0x3, &instr->alu.add.a_unpack)) { return false; } break; case V3D_QPU_A_VFMIN: case V3D_QPU_A_VFMAX: if (!v3d_qpu_float16_unpack_unpack(op & 0x7, &instr->alu.add.a_unpack)) { return false; } instr->alu.add.output_pack = V3D_QPU_PACK_NONE; instr->alu.add.b_unpack = V3D_QPU_UNPACK_NONE; break; default: instr->alu.add.output_pack = V3D_QPU_PACK_NONE; instr->alu.add.a_unpack = V3D_QPU_UNPACK_NONE; instr->alu.add.b_unpack = V3D_QPU_UNPACK_NONE; break; } instr->alu.add.a = mux_a; instr->alu.add.b = mux_b; instr->alu.add.waddr = QPU_GET_FIELD(packed_inst, V3D_QPU_WADDR_A); instr->alu.add.magic_write = packed_inst & VC5_QPU_MA; return true; } static bool v3d_qpu_mul_unpack(const struct v3d_device_info *devinfo, uint64_t packed_inst, struct v3d_qpu_instr *instr) { uint32_t op = QPU_GET_FIELD(packed_inst, VC5_QPU_OP_MUL); uint32_t mux_a = QPU_GET_FIELD(packed_inst, VC5_QPU_MUL_A); uint32_t mux_b = QPU_GET_FIELD(packed_inst, VC5_QPU_MUL_B); { const struct opcode_desc *desc = lookup_opcode(mul_ops, ARRAY_SIZE(mul_ops), op, mux_a, mux_b); if (!desc) return false; instr->alu.mul.op = desc->op; } switch (instr->alu.mul.op) { case V3D_QPU_M_FMUL: instr->alu.mul.output_pack = ((op >> 4) & 0x3) - 1; if (!v3d_qpu_float32_unpack_unpack((op >> 2) & 0x3, &instr->alu.mul.a_unpack)) { return false; } if (!v3d_qpu_float32_unpack_unpack((op >> 0) & 0x3, &instr->alu.mul.b_unpack)) { return false; } break; case V3D_QPU_M_FMOV: instr->alu.mul.output_pack = (((op & 1) << 1) + ((mux_b >> 2) & 1)); if (!v3d_qpu_float32_unpack_unpack(mux_b & 0x3, &instr->alu.mul.a_unpack)) { return false; } break; default: instr->alu.mul.output_pack = V3D_QPU_PACK_NONE; instr->alu.mul.a_unpack = V3D_QPU_UNPACK_NONE; instr->alu.mul.b_unpack = V3D_QPU_UNPACK_NONE; break; } instr->alu.mul.a = mux_a; instr->alu.mul.b = mux_b; instr->alu.mul.waddr = QPU_GET_FIELD(packed_inst, V3D_QPU_WADDR_M); instr->alu.mul.magic_write = packed_inst & VC5_QPU_MM; return true; } static bool v3d_qpu_add_pack(const struct v3d_device_info *devinfo, const struct v3d_qpu_instr *instr, uint64_t *packed_instr) { uint32_t waddr = instr->alu.add.waddr; uint32_t mux_a = instr->alu.add.a; uint32_t mux_b = instr->alu.add.b; int nsrc = v3d_qpu_add_op_num_src(instr->alu.add.op); const struct opcode_desc *desc; int opcode; for (desc = add_ops; desc != &add_ops[ARRAY_SIZE(add_ops)]; desc++) { if (desc->op == instr->alu.add.op) break; } if (desc == &add_ops[ARRAY_SIZE(add_ops)]) return false; opcode = desc->opcode_first; /* If an operation doesn't use an arg, its mux values may be used to * identify the operation type. */ if (nsrc < 2) mux_b = ffs(desc->mux_b_mask) - 1; if (nsrc < 1) mux_a = ffs(desc->mux_a_mask) - 1; switch (instr->alu.add.op) { case V3D_QPU_A_STVPMV: waddr = 0; break; case V3D_QPU_A_STVPMD: waddr = 1; break; case V3D_QPU_A_STVPMP: waddr = 2; break; default: break; } switch (instr->alu.add.op) { case V3D_QPU_A_FADD: case V3D_QPU_A_FADDNF: case V3D_QPU_A_FSUB: case V3D_QPU_A_FMIN: case V3D_QPU_A_FMAX: case V3D_QPU_A_FCMP: { uint32_t output_pack; uint32_t a_unpack; uint32_t b_unpack; if (!v3d_qpu_float32_pack_pack(instr->alu.add.output_pack, &output_pack)) { return false; } opcode |= output_pack << 4; if (!v3d_qpu_float32_unpack_pack(instr->alu.add.a_unpack, &a_unpack)) { return false; } if (!v3d_qpu_float32_unpack_pack(instr->alu.add.b_unpack, &b_unpack)) { return false; } /* These operations with commutative operands are * distinguished by which order their operands come in. */ bool ordering = a_unpack * 8 + mux_a > b_unpack * 8 + mux_b; if (((instr->alu.add.op == V3D_QPU_A_FMIN || instr->alu.add.op == V3D_QPU_A_FADD) && ordering) || ((instr->alu.add.op == V3D_QPU_A_FMAX || instr->alu.add.op == V3D_QPU_A_FADDNF) && !ordering)) { uint32_t temp; temp = a_unpack; a_unpack = b_unpack; b_unpack = temp; temp = mux_a; mux_a = mux_b; mux_b = temp; } opcode |= a_unpack << 2; opcode |= b_unpack << 0; break; } case V3D_QPU_A_FFLOOR: case V3D_QPU_A_FROUND: case V3D_QPU_A_FTRUNC: case V3D_QPU_A_FCEIL: case V3D_QPU_A_FDX: case V3D_QPU_A_FDY: { uint32_t packed; if (!v3d_qpu_float32_pack_pack(instr->alu.add.output_pack, &packed)) { return false; } mux_b |= packed; if (!v3d_qpu_float32_unpack_pack(instr->alu.add.a_unpack, &packed)) { return false; } if (packed == 0) return false; opcode |= packed << 2; break; } case V3D_QPU_A_FTOIN: case V3D_QPU_A_FTOIZ: case V3D_QPU_A_FTOUZ: case V3D_QPU_A_FTOC: if (instr->alu.add.output_pack != V3D_QPU_PACK_NONE) return false; uint32_t packed; if (!v3d_qpu_float32_unpack_pack(instr->alu.add.a_unpack, &packed)) { return false; } if (packed == 0) return false; opcode |= packed << 2; break; case V3D_QPU_A_VFMIN: case V3D_QPU_A_VFMAX: if (instr->alu.add.output_pack != V3D_QPU_PACK_NONE || instr->alu.add.b_unpack != V3D_QPU_UNPACK_NONE) { return false; } if (!v3d_qpu_float16_unpack_pack(instr->alu.add.a_unpack, &packed)) { return false; } opcode |= packed; break; default: if (instr->alu.add.op != V3D_QPU_A_NOP && (instr->alu.add.output_pack != V3D_QPU_PACK_NONE || instr->alu.add.a_unpack != V3D_QPU_UNPACK_NONE || instr->alu.add.b_unpack != V3D_QPU_UNPACK_NONE)) { return false; } break; } *packed_instr |= QPU_SET_FIELD(mux_a, VC5_QPU_ADD_A); *packed_instr |= QPU_SET_FIELD(mux_b, VC5_QPU_ADD_B); *packed_instr |= QPU_SET_FIELD(opcode, VC5_QPU_OP_ADD); *packed_instr |= QPU_SET_FIELD(waddr, V3D_QPU_WADDR_A); if (instr->alu.add.magic_write) *packed_instr |= VC5_QPU_MA; return true; } static bool v3d_qpu_mul_pack(const struct v3d_device_info *devinfo, const struct v3d_qpu_instr *instr, uint64_t *packed_instr) { uint32_t mux_a = instr->alu.mul.a; uint32_t mux_b = instr->alu.mul.b; int nsrc = v3d_qpu_mul_op_num_src(instr->alu.mul.op); const struct opcode_desc *desc; for (desc = mul_ops; desc != &mul_ops[ARRAY_SIZE(mul_ops)]; desc++) { if (desc->op == instr->alu.mul.op) break; } if (desc == &mul_ops[ARRAY_SIZE(mul_ops)]) return false; uint32_t opcode = desc->opcode_first; /* Some opcodes have a single valid value for their mux a/b, so set * that here. If mux a/b determine packing, it will be set below. */ if (nsrc < 2) mux_b = ffs(desc->mux_b_mask) - 1; if (nsrc < 1) mux_a = ffs(desc->mux_a_mask) - 1; switch (instr->alu.mul.op) { case V3D_QPU_M_FMUL: { uint32_t packed; if (!v3d_qpu_float32_pack_pack(instr->alu.mul.output_pack, &packed)) { return false; } /* No need for a +1 because desc->opcode_first has a 1 in this * field. */ opcode += packed << 4; if (!v3d_qpu_float32_unpack_pack(instr->alu.mul.a_unpack, &packed)) { return false; } opcode |= packed << 2; if (!v3d_qpu_float32_unpack_pack(instr->alu.mul.b_unpack, &packed)) { return false; } opcode |= packed << 0; break; } case V3D_QPU_M_FMOV: { uint32_t packed; if (!v3d_qpu_float32_pack_pack(instr->alu.mul.output_pack, &packed)) { return false; } opcode |= (packed >> 1) & 1; mux_b = (packed & 1) << 2; if (!v3d_qpu_float32_unpack_pack(instr->alu.mul.a_unpack, &packed)) { return false; } mux_b |= packed; break; } default: break; } *packed_instr |= QPU_SET_FIELD(mux_a, VC5_QPU_MUL_A); *packed_instr |= QPU_SET_FIELD(mux_b, VC5_QPU_MUL_B); *packed_instr |= QPU_SET_FIELD(opcode, VC5_QPU_OP_MUL); *packed_instr |= QPU_SET_FIELD(instr->alu.mul.waddr, V3D_QPU_WADDR_M); if (instr->alu.mul.magic_write) *packed_instr |= VC5_QPU_MM; return true; } static bool v3d_qpu_instr_unpack_alu(const struct v3d_device_info *devinfo, uint64_t packed_instr, struct v3d_qpu_instr *instr) { instr->type = V3D_QPU_INSTR_TYPE_ALU; if (!v3d_qpu_sig_unpack(devinfo, QPU_GET_FIELD(packed_instr, VC5_QPU_SIG), &instr->sig)) return false; if (!v3d_qpu_flags_unpack(devinfo, QPU_GET_FIELD(packed_instr, VC5_QPU_COND), &instr->flags)) return false; instr->raddr_a = QPU_GET_FIELD(packed_instr, VC5_QPU_RADDR_A); instr->raddr_b = QPU_GET_FIELD(packed_instr, VC5_QPU_RADDR_B); if (!v3d_qpu_add_unpack(devinfo, packed_instr, instr)) return false; if (!v3d_qpu_mul_unpack(devinfo, packed_instr, instr)) return false; return true; } static bool v3d_qpu_instr_unpack_branch(const struct v3d_device_info *devinfo, uint64_t packed_instr, struct v3d_qpu_instr *instr) { instr->type = V3D_QPU_INSTR_TYPE_BRANCH; uint32_t cond = QPU_GET_FIELD(packed_instr, VC5_QPU_BRANCH_COND); if (cond == 0) instr->branch.cond = V3D_QPU_BRANCH_COND_ALWAYS; else if (V3D_QPU_BRANCH_COND_A0 + (cond - 2) <= V3D_QPU_BRANCH_COND_ALLNA) instr->branch.cond = V3D_QPU_BRANCH_COND_A0 + (cond - 2); else return false; uint32_t msfign = QPU_GET_FIELD(packed_instr, VC5_QPU_BRANCH_MSFIGN); if (msfign == 3) return false; instr->branch.msfign = msfign; instr->branch.bdi = QPU_GET_FIELD(packed_instr, VC5_QPU_BRANCH_BDI); instr->branch.ub = packed_instr & VC5_QPU_BRANCH_UB; if (instr->branch.ub) { instr->branch.bdu = QPU_GET_FIELD(packed_instr, VC5_QPU_BRANCH_BDU); } instr->branch.raddr_a = QPU_GET_FIELD(packed_instr, VC5_QPU_RADDR_A); instr->branch.offset = 0; instr->branch.offset += QPU_GET_FIELD(packed_instr, VC5_QPU_BRANCH_ADDR_LOW) << 3; instr->branch.offset += QPU_GET_FIELD(packed_instr, VC5_QPU_BRANCH_ADDR_HIGH) << 24; return true; } bool v3d_qpu_instr_unpack(const struct v3d_device_info *devinfo, uint64_t packed_instr, struct v3d_qpu_instr *instr) { if (QPU_GET_FIELD(packed_instr, VC5_QPU_OP_MUL) != 0) { return v3d_qpu_instr_unpack_alu(devinfo, packed_instr, instr); } else { uint32_t sig = QPU_GET_FIELD(packed_instr, VC5_QPU_SIG); if ((sig & 24) == 16) { return v3d_qpu_instr_unpack_branch(devinfo, packed_instr, instr); } else { return false; } } } static bool v3d_qpu_instr_pack_alu(const struct v3d_device_info *devinfo, const struct v3d_qpu_instr *instr, uint64_t *packed_instr) { uint32_t sig; if (!v3d_qpu_sig_pack(devinfo, &instr->sig, &sig)) return false; *packed_instr |= QPU_SET_FIELD(sig, VC5_QPU_SIG); if (instr->type == V3D_QPU_INSTR_TYPE_ALU) { *packed_instr |= QPU_SET_FIELD(instr->raddr_a, VC5_QPU_RADDR_A); *packed_instr |= QPU_SET_FIELD(instr->raddr_b, VC5_QPU_RADDR_B); if (!v3d_qpu_add_pack(devinfo, instr, packed_instr)) return false; if (!v3d_qpu_mul_pack(devinfo, instr, packed_instr)) return false; uint32_t flags; if (!v3d_qpu_flags_pack(devinfo, &instr->flags, &flags)) return false; *packed_instr |= QPU_SET_FIELD(flags, VC5_QPU_COND); } return true; } static bool v3d_qpu_instr_pack_branch(const struct v3d_device_info *devinfo, const struct v3d_qpu_instr *instr, uint64_t *packed_instr) { *packed_instr |= QPU_SET_FIELD(16, VC5_QPU_SIG); if (instr->branch.cond != V3D_QPU_BRANCH_COND_ALWAYS) { *packed_instr |= QPU_SET_FIELD(2 + (instr->branch.cond - V3D_QPU_BRANCH_COND_A0), VC5_QPU_BRANCH_COND); } *packed_instr |= QPU_SET_FIELD(instr->branch.msfign, VC5_QPU_BRANCH_MSFIGN); *packed_instr |= QPU_SET_FIELD(instr->branch.bdi, VC5_QPU_BRANCH_BDI); if (instr->branch.ub) { *packed_instr |= VC5_QPU_BRANCH_UB; *packed_instr |= QPU_SET_FIELD(instr->branch.bdu, VC5_QPU_BRANCH_BDU); } switch (instr->branch.bdi) { case V3D_QPU_BRANCH_DEST_ABS: case V3D_QPU_BRANCH_DEST_REL: *packed_instr |= QPU_SET_FIELD(instr->branch.msfign, VC5_QPU_BRANCH_MSFIGN); *packed_instr |= QPU_SET_FIELD((instr->branch.offset & ~0xff000000) >> 3, VC5_QPU_BRANCH_ADDR_LOW); *packed_instr |= QPU_SET_FIELD(instr->branch.offset >> 24, VC5_QPU_BRANCH_ADDR_HIGH); case V3D_QPU_BRANCH_DEST_REGFILE: *packed_instr |= QPU_SET_FIELD(instr->branch.raddr_a, VC5_QPU_RADDR_A); break; default: break; } return true; } bool v3d_qpu_instr_pack(const struct v3d_device_info *devinfo, const struct v3d_qpu_instr *instr, uint64_t *packed_instr) { *packed_instr = 0; switch (instr->type) { case V3D_QPU_INSTR_TYPE_ALU: return v3d_qpu_instr_pack_alu(devinfo, instr, packed_instr); case V3D_QPU_INSTR_TYPE_BRANCH: return v3d_qpu_instr_pack_branch(devinfo, instr, packed_instr); default: return false; } }