peridot/vendor/github.com/cloudflare/circl/math/fp448/fp_amd64.h

592 lines
22 KiB
C
Raw Normal View History

// This code was imported from https://github.com/armfazh/rfc7748_precomputed
// CHECK_BMI2ADX triggers bmi2adx if supported,
// otherwise it fallbacks to legacy code.
#define CHECK_BMI2ADX(label, legacy, bmi2adx) \
CMPB ·hasBmi2Adx(SB), $0 \
JE label \
bmi2adx \
RET \
label: \
legacy \
RET
// cselect is a conditional move
// if b=1: it copies y into x;
// if b=0: x remains with the same value;
// if b<> 0,1: undefined.
// Uses: AX, DX, FLAGS
// Instr: x86_64, cmov
#define cselect(x,y,b) \
TESTQ b, b \
MOVQ 0+x, AX; MOVQ 0+y, DX; CMOVQNE DX, AX; MOVQ AX, 0+x; \
MOVQ 8+x, AX; MOVQ 8+y, DX; CMOVQNE DX, AX; MOVQ AX, 8+x; \
MOVQ 16+x, AX; MOVQ 16+y, DX; CMOVQNE DX, AX; MOVQ AX, 16+x; \
MOVQ 24+x, AX; MOVQ 24+y, DX; CMOVQNE DX, AX; MOVQ AX, 24+x; \
MOVQ 32+x, AX; MOVQ 32+y, DX; CMOVQNE DX, AX; MOVQ AX, 32+x; \
MOVQ 40+x, AX; MOVQ 40+y, DX; CMOVQNE DX, AX; MOVQ AX, 40+x; \
MOVQ 48+x, AX; MOVQ 48+y, DX; CMOVQNE DX, AX; MOVQ AX, 48+x;
// cswap is a conditional swap
// if b=1: x,y <- y,x;
// if b=0: x,y remain with the same values;
// if b<> 0,1: undefined.
// Uses: AX, DX, R8, FLAGS
// Instr: x86_64, cmov
#define cswap(x,y,b) \
TESTQ b, b \
MOVQ 0+x, AX; MOVQ AX, R8; MOVQ 0+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 0+x; MOVQ DX, 0+y; \
MOVQ 8+x, AX; MOVQ AX, R8; MOVQ 8+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 8+x; MOVQ DX, 8+y; \
MOVQ 16+x, AX; MOVQ AX, R8; MOVQ 16+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 16+x; MOVQ DX, 16+y; \
MOVQ 24+x, AX; MOVQ AX, R8; MOVQ 24+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 24+x; MOVQ DX, 24+y; \
MOVQ 32+x, AX; MOVQ AX, R8; MOVQ 32+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 32+x; MOVQ DX, 32+y; \
MOVQ 40+x, AX; MOVQ AX, R8; MOVQ 40+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 40+x; MOVQ DX, 40+y; \
MOVQ 48+x, AX; MOVQ AX, R8; MOVQ 48+y, DX; CMOVQNE DX, AX; CMOVQNE R8, DX; MOVQ AX, 48+x; MOVQ DX, 48+y;
// additionLeg adds x and y and stores in z
// Uses: AX, DX, R8-R14, FLAGS
// Instr: x86_64
#define additionLeg(z,x,y) \
MOVQ 0+x, R8; ADDQ 0+y, R8; \
MOVQ 8+x, R9; ADCQ 8+y, R9; \
MOVQ 16+x, R10; ADCQ 16+y, R10; \
MOVQ 24+x, R11; ADCQ 24+y, R11; \
MOVQ 32+x, R12; ADCQ 32+y, R12; \
MOVQ 40+x, R13; ADCQ 40+y, R13; \
MOVQ 48+x, R14; ADCQ 48+y, R14; \
MOVQ $0, AX; ADCQ $0, AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
ADDQ AX, R8; MOVQ $0, AX; \
ADCQ $0, R9; \
ADCQ $0, R10; \
ADCQ DX, R11; \
ADCQ $0, R12; \
ADCQ $0, R13; \
ADCQ $0, R14; \
ADCQ $0, AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
ADDQ AX, R8; MOVQ R8, 0+z; \
ADCQ $0, R9; MOVQ R9, 8+z; \
ADCQ $0, R10; MOVQ R10, 16+z; \
ADCQ DX, R11; MOVQ R11, 24+z; \
ADCQ $0, R12; MOVQ R12, 32+z; \
ADCQ $0, R13; MOVQ R13, 40+z; \
ADCQ $0, R14; MOVQ R14, 48+z;
// additionAdx adds x and y and stores in z
// Uses: AX, DX, R8-R15, FLAGS
// Instr: x86_64, adx
#define additionAdx(z,x,y) \
MOVL $32, R15; \
XORL DX, DX; \
MOVQ 0+x, R8; ADCXQ 0+y, R8; \
MOVQ 8+x, R9; ADCXQ 8+y, R9; \
MOVQ 16+x, R10; ADCXQ 16+y, R10; \
MOVQ 24+x, R11; ADCXQ 24+y, R11; \
MOVQ 32+x, R12; ADCXQ 32+y, R12; \
MOVQ 40+x, R13; ADCXQ 40+y, R13; \
MOVQ 48+x, R14; ADCXQ 48+y, R14; \
;;;;;;;;;;;;;;; ADCXQ DX, DX; \
XORL AX, AX; \
ADCXQ DX, R8; SHLXQ R15, DX, DX; \
ADCXQ AX, R9; \
ADCXQ AX, R10; \
ADCXQ DX, R11; \
ADCXQ AX, R12; \
ADCXQ AX, R13; \
ADCXQ AX, R14; \
ADCXQ AX, AX; \
XORL DX, DX; \
ADCXQ AX, R8; MOVQ R8, 0+z; SHLXQ R15, AX, AX; \
ADCXQ DX, R9; MOVQ R9, 8+z; \
ADCXQ DX, R10; MOVQ R10, 16+z; \
ADCXQ AX, R11; MOVQ R11, 24+z; \
ADCXQ DX, R12; MOVQ R12, 32+z; \
ADCXQ DX, R13; MOVQ R13, 40+z; \
ADCXQ DX, R14; MOVQ R14, 48+z;
// subtraction subtracts y from x and stores in z
// Uses: AX, DX, R8-R14, FLAGS
// Instr: x86_64
#define subtraction(z,x,y) \
MOVQ 0+x, R8; SUBQ 0+y, R8; \
MOVQ 8+x, R9; SBBQ 8+y, R9; \
MOVQ 16+x, R10; SBBQ 16+y, R10; \
MOVQ 24+x, R11; SBBQ 24+y, R11; \
MOVQ 32+x, R12; SBBQ 32+y, R12; \
MOVQ 40+x, R13; SBBQ 40+y, R13; \
MOVQ 48+x, R14; SBBQ 48+y, R14; \
MOVQ $0, AX; SETCS AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
SUBQ AX, R8; MOVQ $0, AX; \
SBBQ $0, R9; \
SBBQ $0, R10; \
SBBQ DX, R11; \
SBBQ $0, R12; \
SBBQ $0, R13; \
SBBQ $0, R14; \
SETCS AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
SUBQ AX, R8; MOVQ R8, 0+z; \
SBBQ $0, R9; MOVQ R9, 8+z; \
SBBQ $0, R10; MOVQ R10, 16+z; \
SBBQ DX, R11; MOVQ R11, 24+z; \
SBBQ $0, R12; MOVQ R12, 32+z; \
SBBQ $0, R13; MOVQ R13, 40+z; \
SBBQ $0, R14; MOVQ R14, 48+z;
// maddBmi2Adx multiplies x and y and accumulates in z
// Uses: AX, DX, R15, FLAGS
// Instr: x86_64, bmi2, adx
#define maddBmi2Adx(z,x,y,i,r0,r1,r2,r3,r4,r5,r6) \
MOVQ i+y, DX; XORL AX, AX; \
MULXQ 0+x, AX, R8; ADOXQ AX, r0; ADCXQ R8, r1; MOVQ r0,i+z; \
MULXQ 8+x, AX, r0; ADOXQ AX, r1; ADCXQ r0, r2; MOVQ $0, R8; \
MULXQ 16+x, AX, r0; ADOXQ AX, r2; ADCXQ r0, r3; \
MULXQ 24+x, AX, r0; ADOXQ AX, r3; ADCXQ r0, r4; \
MULXQ 32+x, AX, r0; ADOXQ AX, r4; ADCXQ r0, r5; \
MULXQ 40+x, AX, r0; ADOXQ AX, r5; ADCXQ r0, r6; \
MULXQ 48+x, AX, r0; ADOXQ AX, r6; ADCXQ R8, r0; \
;;;;;;;;;;;;;;;;;;; ADOXQ R8, r0;
// integerMulAdx multiplies x and y and stores in z
// Uses: AX, DX, R8-R15, FLAGS
// Instr: x86_64, bmi2, adx
#define integerMulAdx(z,x,y) \
MOVL $0,R15; \
MOVQ 0+y, DX; XORL AX, AX; MOVQ $0, R8; \
MULXQ 0+x, AX, R9; MOVQ AX, 0+z; \
MULXQ 8+x, AX, R10; ADCXQ AX, R9; \
MULXQ 16+x, AX, R11; ADCXQ AX, R10; \
MULXQ 24+x, AX, R12; ADCXQ AX, R11; \
MULXQ 32+x, AX, R13; ADCXQ AX, R12; \
MULXQ 40+x, AX, R14; ADCXQ AX, R13; \
MULXQ 48+x, AX, R15; ADCXQ AX, R14; \
;;;;;;;;;;;;;;;;;;;; ADCXQ R8, R15; \
maddBmi2Adx(z,x,y, 8, R9,R10,R11,R12,R13,R14,R15) \
maddBmi2Adx(z,x,y,16,R10,R11,R12,R13,R14,R15, R9) \
maddBmi2Adx(z,x,y,24,R11,R12,R13,R14,R15, R9,R10) \
maddBmi2Adx(z,x,y,32,R12,R13,R14,R15, R9,R10,R11) \
maddBmi2Adx(z,x,y,40,R13,R14,R15, R9,R10,R11,R12) \
maddBmi2Adx(z,x,y,48,R14,R15, R9,R10,R11,R12,R13) \
MOVQ R15, 56+z; \
MOVQ R9, 64+z; \
MOVQ R10, 72+z; \
MOVQ R11, 80+z; \
MOVQ R12, 88+z; \
MOVQ R13, 96+z; \
MOVQ R14, 104+z;
// maddLegacy multiplies x and y and accumulates in z
// Uses: AX, DX, R15, FLAGS
// Instr: x86_64
#define maddLegacy(z,x,y,i) \
MOVQ i+y, R15; \
MOVQ 0+x, AX; MULQ R15; MOVQ AX, R8; ;;;;;;;;;;;; MOVQ DX, R9; \
MOVQ 8+x, AX; MULQ R15; ADDQ AX, R9; ADCQ $0, DX; MOVQ DX, R10; \
MOVQ 16+x, AX; MULQ R15; ADDQ AX, R10; ADCQ $0, DX; MOVQ DX, R11; \
MOVQ 24+x, AX; MULQ R15; ADDQ AX, R11; ADCQ $0, DX; MOVQ DX, R12; \
MOVQ 32+x, AX; MULQ R15; ADDQ AX, R12; ADCQ $0, DX; MOVQ DX, R13; \
MOVQ 40+x, AX; MULQ R15; ADDQ AX, R13; ADCQ $0, DX; MOVQ DX, R14; \
MOVQ 48+x, AX; MULQ R15; ADDQ AX, R14; ADCQ $0, DX; \
ADDQ 0+i+z, R8; MOVQ R8, 0+i+z; \
ADCQ 8+i+z, R9; MOVQ R9, 8+i+z; \
ADCQ 16+i+z, R10; MOVQ R10, 16+i+z; \
ADCQ 24+i+z, R11; MOVQ R11, 24+i+z; \
ADCQ 32+i+z, R12; MOVQ R12, 32+i+z; \
ADCQ 40+i+z, R13; MOVQ R13, 40+i+z; \
ADCQ 48+i+z, R14; MOVQ R14, 48+i+z; \
ADCQ $0, DX; MOVQ DX, 56+i+z;
// integerMulLeg multiplies x and y and stores in z
// Uses: AX, DX, R8-R15, FLAGS
// Instr: x86_64
#define integerMulLeg(z,x,y) \
MOVQ 0+y, R15; \
MOVQ 0+x, AX; MULQ R15; MOVQ AX, 0+z; ;;;;;;;;;;;; MOVQ DX, R8; \
MOVQ 8+x, AX; MULQ R15; ADDQ AX, R8; ADCQ $0, DX; MOVQ DX, R9; MOVQ R8, 8+z; \
MOVQ 16+x, AX; MULQ R15; ADDQ AX, R9; ADCQ $0, DX; MOVQ DX, R10; MOVQ R9, 16+z; \
MOVQ 24+x, AX; MULQ R15; ADDQ AX, R10; ADCQ $0, DX; MOVQ DX, R11; MOVQ R10, 24+z; \
MOVQ 32+x, AX; MULQ R15; ADDQ AX, R11; ADCQ $0, DX; MOVQ DX, R12; MOVQ R11, 32+z; \
MOVQ 40+x, AX; MULQ R15; ADDQ AX, R12; ADCQ $0, DX; MOVQ DX, R13; MOVQ R12, 40+z; \
MOVQ 48+x, AX; MULQ R15; ADDQ AX, R13; ADCQ $0, DX; MOVQ DX,56+z; MOVQ R13, 48+z; \
maddLegacy(z,x,y, 8) \
maddLegacy(z,x,y,16) \
maddLegacy(z,x,y,24) \
maddLegacy(z,x,y,32) \
maddLegacy(z,x,y,40) \
maddLegacy(z,x,y,48)
// integerSqrLeg squares x and stores in z
// Uses: AX, CX, DX, R8-R15, FLAGS
// Instr: x86_64
#define integerSqrLeg(z,x) \
XORL R15, R15; \
MOVQ 0+x, CX; \
MOVQ CX, AX; MULQ CX; MOVQ AX, 0+z; MOVQ DX, R8; \
ADDQ CX, CX; ADCQ $0, R15; \
MOVQ 8+x, AX; MULQ CX; ADDQ AX, R8; ADCQ $0, DX; MOVQ DX, R9; MOVQ R8, 8+z; \
MOVQ 16+x, AX; MULQ CX; ADDQ AX, R9; ADCQ $0, DX; MOVQ DX, R10; \
MOVQ 24+x, AX; MULQ CX; ADDQ AX, R10; ADCQ $0, DX; MOVQ DX, R11; \
MOVQ 32+x, AX; MULQ CX; ADDQ AX, R11; ADCQ $0, DX; MOVQ DX, R12; \
MOVQ 40+x, AX; MULQ CX; ADDQ AX, R12; ADCQ $0, DX; MOVQ DX, R13; \
MOVQ 48+x, AX; MULQ CX; ADDQ AX, R13; ADCQ $0, DX; MOVQ DX, R14; \
\
MOVQ 8+x, CX; \
MOVQ CX, AX; ADDQ R15, CX; MOVQ $0, R15; ADCQ $0, R15; \
;;;;;;;;;;;;;; MULQ CX; ADDQ AX, R9; ADCQ $0, DX; MOVQ R9,16+z; \
MOVQ R15, AX; NEGQ AX; ANDQ 8+x, AX; ADDQ AX, DX; ADCQ $0, R11; MOVQ DX, R8; \
ADDQ 8+x, CX; ADCQ $0, R15; \
MOVQ 16+x, AX; MULQ CX; ADDQ AX, R10; ADCQ $0, DX; ADDQ R8, R10; ADCQ $0, DX; MOVQ DX, R8; MOVQ R10, 24+z; \
MOVQ 24+x, AX; MULQ CX; ADDQ AX, R11; ADCQ $0, DX; ADDQ R8, R11; ADCQ $0, DX; MOVQ DX, R8; \
MOVQ 32+x, AX; MULQ CX; ADDQ AX, R12; ADCQ $0, DX; ADDQ R8, R12; ADCQ $0, DX; MOVQ DX, R8; \
MOVQ 40+x, AX; MULQ CX; ADDQ AX, R13; ADCQ $0, DX; ADDQ R8, R13; ADCQ $0, DX; MOVQ DX, R8; \
MOVQ 48+x, AX; MULQ CX; ADDQ AX, R14; ADCQ $0, DX; ADDQ R8, R14; ADCQ $0, DX; MOVQ DX, R9; \
\
MOVQ 16+x, CX; \
MOVQ CX, AX; ADDQ R15, CX; MOVQ $0, R15; ADCQ $0, R15; \
;;;;;;;;;;;;;; MULQ CX; ADDQ AX, R11; ADCQ $0, DX; MOVQ R11, 32+z; \
MOVQ R15, AX; NEGQ AX; ANDQ 16+x,AX; ADDQ AX, DX; ADCQ $0, R13; MOVQ DX, R8; \
ADDQ 16+x, CX; ADCQ $0, R15; \
MOVQ 24+x, AX; MULQ CX; ADDQ AX, R12; ADCQ $0, DX; ADDQ R8, R12; ADCQ $0, DX; MOVQ DX, R8; MOVQ R12, 40+z; \
MOVQ 32+x, AX; MULQ CX; ADDQ AX, R13; ADCQ $0, DX; ADDQ R8, R13; ADCQ $0, DX; MOVQ DX, R8; \
MOVQ 40+x, AX; MULQ CX; ADDQ AX, R14; ADCQ $0, DX; ADDQ R8, R14; ADCQ $0, DX; MOVQ DX, R8; \
MOVQ 48+x, AX; MULQ CX; ADDQ AX, R9; ADCQ $0, DX; ADDQ R8, R9; ADCQ $0, DX; MOVQ DX,R10; \
\
MOVQ 24+x, CX; \
MOVQ CX, AX; ADDQ R15, CX; MOVQ $0, R15; ADCQ $0, R15; \
;;;;;;;;;;;;;; MULQ CX; ADDQ AX, R13; ADCQ $0, DX; MOVQ R13, 48+z; \
MOVQ R15, AX; NEGQ AX; ANDQ 24+x,AX; ADDQ AX, DX; ADCQ $0, R9; MOVQ DX, R8; \
ADDQ 24+x, CX; ADCQ $0, R15; \
MOVQ 32+x, AX; MULQ CX; ADDQ AX, R14; ADCQ $0, DX; ADDQ R8, R14; ADCQ $0, DX; MOVQ DX, R8; MOVQ R14, 56+z; \
MOVQ 40+x, AX; MULQ CX; ADDQ AX, R9; ADCQ $0, DX; ADDQ R8, R9; ADCQ $0, DX; MOVQ DX, R8; \
MOVQ 48+x, AX; MULQ CX; ADDQ AX, R10; ADCQ $0, DX; ADDQ R8, R10; ADCQ $0, DX; MOVQ DX,R11; \
\
MOVQ 32+x, CX; \
MOVQ CX, AX; ADDQ R15, CX; MOVQ $0, R15; ADCQ $0, R15; \
;;;;;;;;;;;;;; MULQ CX; ADDQ AX, R9; ADCQ $0, DX; MOVQ R9, 64+z; \
MOVQ R15, AX; NEGQ AX; ANDQ 32+x,AX; ADDQ AX, DX; ADCQ $0, R11; MOVQ DX, R8; \
ADDQ 32+x, CX; ADCQ $0, R15; \
MOVQ 40+x, AX; MULQ CX; ADDQ AX, R10; ADCQ $0, DX; ADDQ R8, R10; ADCQ $0, DX; MOVQ DX, R8; MOVQ R10, 72+z; \
MOVQ 48+x, AX; MULQ CX; ADDQ AX, R11; ADCQ $0, DX; ADDQ R8, R11; ADCQ $0, DX; MOVQ DX,R12; \
\
XORL R13, R13; \
XORL R14, R14; \
MOVQ 40+x, CX; \
MOVQ CX, AX; ADDQ R15, CX; MOVQ $0, R15; ADCQ $0, R15; \
;;;;;;;;;;;;;; MULQ CX; ADDQ AX, R11; ADCQ $0, DX; MOVQ R11, 80+z; \
MOVQ R15, AX; NEGQ AX; ANDQ 40+x,AX; ADDQ AX, DX; ADCQ $0, R13; MOVQ DX, R8; \
ADDQ 40+x, CX; ADCQ $0, R15; \
MOVQ 48+x, AX; MULQ CX; ADDQ AX, R12; ADCQ $0, DX; ADDQ R8, R12; ADCQ $0, DX; MOVQ DX, R8; MOVQ R12, 88+z; \
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ADDQ R8, R13; ADCQ $0,R14; \
\
XORL R9, R9; \
MOVQ 48+x, CX; \
MOVQ CX, AX; ADDQ R15, CX; MOVQ $0, R15; ADCQ $0, R15; \
;;;;;;;;;;;;;; MULQ CX; ADDQ AX, R13; ADCQ $0, DX; MOVQ R13, 96+z; \
MOVQ R15, AX; NEGQ AX; ANDQ 48+x,AX; ADDQ AX, DX; ADCQ $0, R9; MOVQ DX, R8; \
;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; ADDQ R8,R14; ADCQ $0, R9; MOVQ R14, 104+z;
// integerSqrAdx squares x and stores in z
// Uses: AX, CX, DX, R8-R15, FLAGS
// Instr: x86_64, bmi2, adx
#define integerSqrAdx(z,x) \
XORL R15, R15; \
MOVQ 0+x, DX; \
;;;;;;;;;;;;;; MULXQ DX, AX, R8; MOVQ AX, 0+z; \
ADDQ DX, DX; ADCQ $0, R15; CLC; \
MULXQ 8+x, AX, R9; ADCXQ AX, R8; MOVQ R8, 8+z; \
MULXQ 16+x, AX, R10; ADCXQ AX, R9; MOVQ $0, R8;\
MULXQ 24+x, AX, R11; ADCXQ AX, R10; \
MULXQ 32+x, AX, R12; ADCXQ AX, R11; \
MULXQ 40+x, AX, R13; ADCXQ AX, R12; \
MULXQ 48+x, AX, R14; ADCXQ AX, R13; \
;;;;;;;;;;;;;;;;;;;; ADCXQ R8, R14; \
\
MOVQ 8+x, DX; \
MOVQ DX, AX; ADDQ R15, DX; MOVQ $0, R15; ADCQ $0, R15; \
MULXQ AX, AX, CX; \
MOVQ R15, R8; NEGQ R8; ANDQ 8+x, R8; \
ADDQ AX, R9; MOVQ R9, 16+z; \
ADCQ CX, R8; \
ADCQ $0, R11; \
ADDQ 8+x, DX; \
ADCQ $0, R15; \
XORL R9, R9; ;;;;;;;;;;;;;;;;;;;;; ADOXQ R8, R10; \
MULXQ 16+x, AX, CX; ADCXQ AX, R10; ADOXQ CX, R11; MOVQ R10, 24+z; \
MULXQ 24+x, AX, CX; ADCXQ AX, R11; ADOXQ CX, R12; MOVQ $0, R10; \
MULXQ 32+x, AX, CX; ADCXQ AX, R12; ADOXQ CX, R13; \
MULXQ 40+x, AX, CX; ADCXQ AX, R13; ADOXQ CX, R14; \
MULXQ 48+x, AX, CX; ADCXQ AX, R14; ADOXQ CX, R9; \
;;;;;;;;;;;;;;;;;;; ADCXQ R10, R9; \
\
MOVQ 16+x, DX; \
MOVQ DX, AX; ADDQ R15, DX; MOVQ $0, R15; ADCQ $0, R15; \
MULXQ AX, AX, CX; \
MOVQ R15, R8; NEGQ R8; ANDQ 16+x, R8; \
ADDQ AX, R11; MOVQ R11, 32+z; \
ADCQ CX, R8; \
ADCQ $0, R13; \
ADDQ 16+x, DX; \
ADCQ $0, R15; \
XORL R11, R11; ;;;;;;;;;;;;;;;;;;; ADOXQ R8, R12; \
MULXQ 24+x, AX, CX; ADCXQ AX, R12; ADOXQ CX, R13; MOVQ R12, 40+z; \
MULXQ 32+x, AX, CX; ADCXQ AX, R13; ADOXQ CX, R14; MOVQ $0, R12; \
MULXQ 40+x, AX, CX; ADCXQ AX, R14; ADOXQ CX, R9; \
MULXQ 48+x, AX, CX; ADCXQ AX, R9; ADOXQ CX, R10; \
;;;;;;;;;;;;;;;;;;; ADCXQ R11,R10; \
\
MOVQ 24+x, DX; \
MOVQ DX, AX; ADDQ R15, DX; MOVQ $0, R15; ADCQ $0, R15; \
MULXQ AX, AX, CX; \
MOVQ R15, R8; NEGQ R8; ANDQ 24+x, R8; \
ADDQ AX, R13; MOVQ R13, 48+z; \
ADCQ CX, R8; \
ADCQ $0, R9; \
ADDQ 24+x, DX; \
ADCQ $0, R15; \
XORL R13, R13; ;;;;;;;;;;;;;;;;;;; ADOXQ R8, R14; \
MULXQ 32+x, AX, CX; ADCXQ AX, R14; ADOXQ CX, R9; MOVQ R14, 56+z; \
MULXQ 40+x, AX, CX; ADCXQ AX, R9; ADOXQ CX, R10; MOVQ $0, R14; \
MULXQ 48+x, AX, CX; ADCXQ AX, R10; ADOXQ CX, R11; \
;;;;;;;;;;;;;;;;;;; ADCXQ R12,R11; \
\
MOVQ 32+x, DX; \
MOVQ DX, AX; ADDQ R15, DX; MOVQ $0, R15; ADCQ $0, R15; \
MULXQ AX, AX, CX; \
MOVQ R15, R8; NEGQ R8; ANDQ 32+x, R8; \
ADDQ AX, R9; MOVQ R9, 64+z; \
ADCQ CX, R8; \
ADCQ $0, R11; \
ADDQ 32+x, DX; \
ADCQ $0, R15; \
XORL R9, R9; ;;;;;;;;;;;;;;;;;;;;; ADOXQ R8, R10; \
MULXQ 40+x, AX, CX; ADCXQ AX, R10; ADOXQ CX, R11; MOVQ R10, 72+z; \
MULXQ 48+x, AX, CX; ADCXQ AX, R11; ADOXQ CX, R12; \
;;;;;;;;;;;;;;;;;;; ADCXQ R13,R12; \
\
MOVQ 40+x, DX; \
MOVQ DX, AX; ADDQ R15, DX; MOVQ $0, R15; ADCQ $0, R15; \
MULXQ AX, AX, CX; \
MOVQ R15, R8; NEGQ R8; ANDQ 40+x, R8; \
ADDQ AX, R11; MOVQ R11, 80+z; \
ADCQ CX, R8; \
ADCQ $0, R13; \
ADDQ 40+x, DX; \
ADCQ $0, R15; \
XORL R11, R11; ;;;;;;;;;;;;;;;;;;; ADOXQ R8, R12; \
MULXQ 48+x, AX, CX; ADCXQ AX, R12; ADOXQ CX, R13; MOVQ R12, 88+z; \
;;;;;;;;;;;;;;;;;;; ADCXQ R14,R13; \
\
MOVQ 48+x, DX; \
MOVQ DX, AX; ADDQ R15, DX; MOVQ $0, R15; ADCQ $0, R15; \
MULXQ AX, AX, CX; \
MOVQ R15, R8; NEGQ R8; ANDQ 48+x, R8; \
XORL R10, R10; ;;;;;;;;;;;;;; ADOXQ CX, R14; \
;;;;;;;;;;;;;; ADCXQ AX, R13; ;;;;;;;;;;;;;; MOVQ R13, 96+z; \
;;;;;;;;;;;;;; ADCXQ R8, R14; MOVQ R14, 104+z;
// reduceFromDoubleLeg finds a z=x modulo p such that z<2^448 and stores in z
// Uses: AX, R8-R15, FLAGS
// Instr: x86_64
#define reduceFromDoubleLeg(z,x) \
/* ( ,2C13,2C12,2C11,2C10|C10,C9,C8, C7) + (C6,...,C0) */ \
/* (r14, r13, r12, r11, r10,r9,r8,r15) */ \
MOVQ 80+x,AX; MOVQ AX,R10; \
MOVQ $0xFFFFFFFF00000000, R8; \
ANDQ R8,R10; \
\
MOVQ $0,R14; \
MOVQ 104+x,R13; SHLQ $1,R13,R14; \
MOVQ 96+x,R12; SHLQ $1,R12,R13; \
MOVQ 88+x,R11; SHLQ $1,R11,R12; \
MOVQ 72+x, R9; SHLQ $1,R10,R11; \
MOVQ 64+x, R8; SHLQ $1,R10; \
MOVQ $0xFFFFFFFF,R15; ANDQ R15,AX; ORQ AX,R10; \
MOVQ 56+x,R15; \
\
ADDQ 0+x,R15; MOVQ R15, 0+z; MOVQ 56+x,R15; \
ADCQ 8+x, R8; MOVQ R8, 8+z; MOVQ 64+x, R8; \
ADCQ 16+x, R9; MOVQ R9,16+z; MOVQ 72+x, R9; \
ADCQ 24+x,R10; MOVQ R10,24+z; MOVQ 80+x,R10; \
ADCQ 32+x,R11; MOVQ R11,32+z; MOVQ 88+x,R11; \
ADCQ 40+x,R12; MOVQ R12,40+z; MOVQ 96+x,R12; \
ADCQ 48+x,R13; MOVQ R13,48+z; MOVQ 104+x,R13; \
ADCQ $0,R14; \
/* (c10c9,c9c8,c8c7,c7c13,c13c12,c12c11,c11c10) + (c6,...,c0) */ \
/* ( r9, r8, r15, r13, r12, r11, r10) */ \
MOVQ R10, AX; \
SHRQ $32,R11,R10; \
SHRQ $32,R12,R11; \
SHRQ $32,R13,R12; \
SHRQ $32,R15,R13; \
SHRQ $32, R8,R15; \
SHRQ $32, R9, R8; \
SHRQ $32, AX, R9; \
\
ADDQ 0+z,R10; \
ADCQ 8+z,R11; \
ADCQ 16+z,R12; \
ADCQ 24+z,R13; \
ADCQ 32+z,R15; \
ADCQ 40+z, R8; \
ADCQ 48+z, R9; \
ADCQ $0,R14; \
/* ( c7) + (c6,...,c0) */ \
/* (r14) */ \
MOVQ R14, AX; SHLQ $32, AX; \
ADDQ R14,R10; MOVQ $0,R14; \
ADCQ $0,R11; \
ADCQ $0,R12; \
ADCQ AX,R13; \
ADCQ $0,R15; \
ADCQ $0, R8; \
ADCQ $0, R9; \
ADCQ $0,R14; \
/* ( c7) + (c6,...,c0) */ \
/* (r14) */ \
MOVQ R14, AX; SHLQ $32,AX; \
ADDQ R14,R10; MOVQ R10, 0+z; \
ADCQ $0,R11; MOVQ R11, 8+z; \
ADCQ $0,R12; MOVQ R12,16+z; \
ADCQ AX,R13; MOVQ R13,24+z; \
ADCQ $0,R15; MOVQ R15,32+z; \
ADCQ $0, R8; MOVQ R8,40+z; \
ADCQ $0, R9; MOVQ R9,48+z;
// reduceFromDoubleAdx finds a z=x modulo p such that z<2^448 and stores in z
// Uses: AX, R8-R15, FLAGS
// Instr: x86_64, adx
#define reduceFromDoubleAdx(z,x) \
/* ( ,2C13,2C12,2C11,2C10|C10,C9,C8, C7) + (C6,...,C0) */ \
/* (r14, r13, r12, r11, r10,r9,r8,r15) */ \
MOVQ 80+x,AX; MOVQ AX,R10; \
MOVQ $0xFFFFFFFF00000000, R8; \
ANDQ R8,R10; \
\
MOVQ $0,R14; \
MOVQ 104+x,R13; SHLQ $1,R13,R14; \
MOVQ 96+x,R12; SHLQ $1,R12,R13; \
MOVQ 88+x,R11; SHLQ $1,R11,R12; \
MOVQ 72+x, R9; SHLQ $1,R10,R11; \
MOVQ 64+x, R8; SHLQ $1,R10; \
MOVQ $0xFFFFFFFF,R15; ANDQ R15,AX; ORQ AX,R10; \
MOVQ 56+x,R15; \
\
XORL AX,AX; \
ADCXQ 0+x,R15; MOVQ R15, 0+z; MOVQ 56+x,R15; \
ADCXQ 8+x, R8; MOVQ R8, 8+z; MOVQ 64+x, R8; \
ADCXQ 16+x, R9; MOVQ R9,16+z; MOVQ 72+x, R9; \
ADCXQ 24+x,R10; MOVQ R10,24+z; MOVQ 80+x,R10; \
ADCXQ 32+x,R11; MOVQ R11,32+z; MOVQ 88+x,R11; \
ADCXQ 40+x,R12; MOVQ R12,40+z; MOVQ 96+x,R12; \
ADCXQ 48+x,R13; MOVQ R13,48+z; MOVQ 104+x,R13; \
ADCXQ AX,R14; \
/* (c10c9,c9c8,c8c7,c7c13,c13c12,c12c11,c11c10) + (c6,...,c0) */ \
/* ( r9, r8, r15, r13, r12, r11, r10) */ \
MOVQ R10, AX; \
SHRQ $32,R11,R10; \
SHRQ $32,R12,R11; \
SHRQ $32,R13,R12; \
SHRQ $32,R15,R13; \
SHRQ $32, R8,R15; \
SHRQ $32, R9, R8; \
SHRQ $32, AX, R9; \
\
XORL AX,AX; \
ADCXQ 0+z,R10; \
ADCXQ 8+z,R11; \
ADCXQ 16+z,R12; \
ADCXQ 24+z,R13; \
ADCXQ 32+z,R15; \
ADCXQ 40+z, R8; \
ADCXQ 48+z, R9; \
ADCXQ AX,R14; \
/* ( c7) + (c6,...,c0) */ \
/* (r14) */ \
MOVQ R14, AX; SHLQ $32, AX; \
CLC; \
ADCXQ R14,R10; MOVQ $0,R14; \
ADCXQ R14,R11; \
ADCXQ R14,R12; \
ADCXQ AX,R13; \
ADCXQ R14,R15; \
ADCXQ R14, R8; \
ADCXQ R14, R9; \
ADCXQ R14,R14; \
/* ( c7) + (c6,...,c0) */ \
/* (r14) */ \
MOVQ R14, AX; SHLQ $32, AX; \
CLC; \
ADCXQ R14,R10; MOVQ R10, 0+z; MOVQ $0,R14; \
ADCXQ R14,R11; MOVQ R11, 8+z; \
ADCXQ R14,R12; MOVQ R12,16+z; \
ADCXQ AX,R13; MOVQ R13,24+z; \
ADCXQ R14,R15; MOVQ R15,32+z; \
ADCXQ R14, R8; MOVQ R8,40+z; \
ADCXQ R14, R9; MOVQ R9,48+z;
// addSub calculates two operations: x,y = x+y,x-y
// Uses: AX, DX, R8-R15, FLAGS
#define addSub(x,y) \
MOVQ 0+x, R8; ADDQ 0+y, R8; \
MOVQ 8+x, R9; ADCQ 8+y, R9; \
MOVQ 16+x, R10; ADCQ 16+y, R10; \
MOVQ 24+x, R11; ADCQ 24+y, R11; \
MOVQ 32+x, R12; ADCQ 32+y, R12; \
MOVQ 40+x, R13; ADCQ 40+y, R13; \
MOVQ 48+x, R14; ADCQ 48+y, R14; \
MOVQ $0, AX; ADCQ $0, AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
ADDQ AX, R8; MOVQ $0, AX; \
ADCQ $0, R9; \
ADCQ $0, R10; \
ADCQ DX, R11; \
ADCQ $0, R12; \
ADCQ $0, R13; \
ADCQ $0, R14; \
ADCQ $0, AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
ADDQ AX, R8; MOVQ 0+x,AX; MOVQ R8, 0+x; MOVQ AX, R8; \
ADCQ $0, R9; MOVQ 8+x,AX; MOVQ R9, 8+x; MOVQ AX, R9; \
ADCQ $0, R10; MOVQ 16+x,AX; MOVQ R10, 16+x; MOVQ AX, R10; \
ADCQ DX, R11; MOVQ 24+x,AX; MOVQ R11, 24+x; MOVQ AX, R11; \
ADCQ $0, R12; MOVQ 32+x,AX; MOVQ R12, 32+x; MOVQ AX, R12; \
ADCQ $0, R13; MOVQ 40+x,AX; MOVQ R13, 40+x; MOVQ AX, R13; \
ADCQ $0, R14; MOVQ 48+x,AX; MOVQ R14, 48+x; MOVQ AX, R14; \
SUBQ 0+y, R8; \
SBBQ 8+y, R9; \
SBBQ 16+y, R10; \
SBBQ 24+y, R11; \
SBBQ 32+y, R12; \
SBBQ 40+y, R13; \
SBBQ 48+y, R14; \
MOVQ $0, AX; SETCS AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
SUBQ AX, R8; MOVQ $0, AX; \
SBBQ $0, R9; \
SBBQ $0, R10; \
SBBQ DX, R11; \
SBBQ $0, R12; \
SBBQ $0, R13; \
SBBQ $0, R14; \
SETCS AX; \
MOVQ AX, DX; \
SHLQ $32, DX; \
SUBQ AX, R8; MOVQ R8, 0+y; \
SBBQ $0, R9; MOVQ R9, 8+y; \
SBBQ $0, R10; MOVQ R10, 16+y; \
SBBQ DX, R11; MOVQ R11, 24+y; \
SBBQ $0, R12; MOVQ R12, 32+y; \
SBBQ $0, R13; MOVQ R13, 40+y; \
SBBQ $0, R14; MOVQ R14, 48+y;