mirror/gmsm
1
0
Fork 0
mirror of https://github.com/emmansun/gmsm.git synced 2025-04-26 20:26:19 +08:00
Code Issues Packages Projects Releases Wiki Activity

sm3: amd64, avx version, remove usage of RORX #157

Browse Source
This commit is contained in:
Sun Yimin 2023-09-19 17:32:53 +08:00 committed by GitHub
parent cee7547606
commit 077664e809
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 55 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
sm3/sm3block_amd64.go
View File

@ -6,7 +6,7 @@ package sm3
import "golang.org/x/sys/cpu" import "golang.org/x/sys/cpu"
var useAVX2 = cpu.X86.HasAVX2 && cpu.X86.HasBMI2 var useAVX2 = cpu.X86.HasAVX2 && cpu.X86.HasBMI2
var useAVX = cpu.X86.HasAVX && cpu.X86.HasBMI2 var useAVX = cpu.X86.HasAVX
//go:noescape //go:noescape
func blockAMD64(dig *digest, p []byte) func blockAMD64(dig *digest, p []byte)

81
sm3/sm3block_avx_amd64.s
View File

@ -60,7 +60,8 @@
// For rounds [0 - 16) // For rounds [0 - 16)
#define ROUND_AND_SCHED_N_0_0(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_0_0(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 0 ############################// ; \ // ############################# RND N + 0 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12, RORXL is BMI2 instr MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
VPALIGNR $12, XWORD0, XWORD1, XTMP0; \ // XTMP0 = W[-13] = {w6,w5,w4,w3} VPALIGNR $12, XWORD0, XWORD1, XTMP0; \ // XTMP0 = W[-13] = {w6,w5,w4,w3}
@ -86,16 +87,19 @@
VPXOR XTMP1, XTMP0, XTMP0; \ // XTMP0 = W[-6] ^ (W[-13] rol 7) VPXOR XTMP1, XTMP0, XTMP0; \ // XTMP0 = W[-6] ^ (W[-13] rol 7)
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
RORXL $23, y2, y0; \ MOVL y2, y0; \
ROLL $9, y0; \
VPALIGNR $12, XWORD1, XWORD2, XTMP1; \ // XTMP1 = W[-9] = {w10,w9,w8,w7} VPALIGNR $12, XWORD1, XWORD2, XTMP1; \ // XTMP1 = W[-9] = {w10,w9,w8,w7}
RORXL $15, y2, d; \ MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XWORD0, XTMP1, XTMP1; \ // XTMP1 = W[-9] ^ W[-16] VPXOR XWORD0, XTMP1, XTMP1; \ // XTMP1 = W[-9] ^ W[-16]
#define ROUND_AND_SCHED_N_0_1(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_0_1(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 1 ############################// ; \ // ############################# RND N + 1 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
VPSHUFD $0xA5, XWORD3, XTMP2; \ // XTMP2 = W[-3] {BBAA} {w14,w14,w13,w13} VPSHUFD $0xA5, XWORD3, XTMP2; \ // XTMP2 = W[-3] {BBAA} {w14,w14,w13,w13}
@ -121,15 +125,18 @@
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
VPSRLQ $9, XTMP2, XTMP4; \ // XTMP4 = XTMP2 rol 23 {xxxA} VPSRLQ $9, XTMP2, XTMP4; \ // XTMP4 = XTMP2 rol 23 {xxxA}
RORXL $23, y2, y0; \ MOVL y2, y0; \
RORXL $15, y2, d; \ ROLL $9, y0; \
MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XTMP2, XTMP4, XTMP4; \ // XTMP4 = XTMP2 ^ (XTMP2 rol 23 {xxxA}) VPXOR XTMP2, XTMP4, XTMP4; \ // XTMP4 = XTMP2 ^ (XTMP2 rol 23 {xxxA})
#define ROUND_AND_SCHED_N_0_2(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_0_2(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 2 ############################// ; \ // ############################# RND N + 2 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
ADDL y0, y2; \ // y2 = a <<< 12 + e + T ADDL y0, y2; \ // y2 = a <<< 12 + e + T
@ -154,16 +161,19 @@
ADDL y1, y2; \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 ADDL y1, y2; \ // y2 = GG(e, f, g) + h + SS1 + W = tt2
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
MOVL y2, y0; \
ROLL $9, y0; \
VPOR XTMP3, XTMP4, XTMP4; \ // XTMP4 = (W[-3] rol 15) {DCxx} VPOR XTMP3, XTMP4, XTMP4; \ // XTMP4 = (W[-3] rol 15) {DCxx}
RORXL $23, y2, y0; \ MOVL y2, d; \
RORXL $15, y2, d; \ ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XTMP1, XTMP4, XTMP4; \ // XTMP4 = W[-9] ^ W[-16] ^ (W[-3] rol 15) {DCxx} VPXOR XTMP1, XTMP4, XTMP4; \ // XTMP4 = W[-9] ^ W[-16] ^ (W[-3] rol 15) {DCxx}
#define ROUND_AND_SCHED_N_0_3(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_0_3(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 3 ############################// ; \ // ############################# RND N + 3 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
VPSLLD $15, XTMP4, XTMP5; \ VPSLLD $15, XTMP4, XTMP5; \
@ -188,9 +198,11 @@
ADDL y1, y2; \ // y2 = GG(e, f, g) + h + SS1 + W = tt2 ADDL y1, y2; \ // y2 = GG(e, f, g) + h + SS1 + W = tt2
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
MOVL y2, y0; \
ROLL $9, y0; \
VPXOR XTMP3, XTMP1, XTMP1; \ // XTMP1 = XTMP4 ^ (XTMP4 rol 15 {DCxx}) ^ (XTMP4 rol 23 {DCxx}) VPXOR XTMP3, XTMP1, XTMP1; \ // XTMP1 = XTMP4 ^ (XTMP4 rol 15 {DCxx}) ^ (XTMP4 rol 23 {DCxx})
RORXL $23, y2, y0; \ MOVL y2, d; \
RORXL $15, y2, d; \ ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XTMP1, XTMP0, XWORD0; \ // XWORD0 = {W[3], W[2], W[1], W[0]} VPXOR XTMP1, XTMP0, XWORD0; \ // XWORD0 = {W[3], W[2], W[1], W[0]}
@ -198,7 +210,8 @@
// For rounds [16 - 64) // For rounds [16 - 64)
#define ROUND_AND_SCHED_N_1_0(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_1_0(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 0 ############################// ; \ // ############################# RND N + 0 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
VPALIGNR $12, XWORD0, XWORD1, XTMP0; \ // XTMP0 = W[-13] = {w6,w5,w4,w3} VPALIGNR $12, XWORD0, XWORD1, XTMP0; \ // XTMP0 = W[-13] = {w6,w5,w4,w3}
@ -229,15 +242,18 @@
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
VPALIGNR $12, XWORD1, XWORD2, XTMP1; \ // XTMP1 = W[-9] = {w10,w9,w8,w7} VPALIGNR $12, XWORD1, XWORD2, XTMP1; \ // XTMP1 = W[-9] = {w10,w9,w8,w7}
RORXL $23, y2, y0; \ MOVL y2, y0; \
RORXL $15, y2, d; \ ROLL $9, y0; \
MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XWORD0, XTMP1, XTMP1; \ // XTMP1 = W[-9] ^ W[-16] VPXOR XWORD0, XTMP1, XTMP1; \ // XTMP1 = W[-9] ^ W[-16]
#define ROUND_AND_SCHED_N_1_1(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_1_1(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 1 ############################// ; \ // ############################# RND N + 1 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
ADDL y0, y2; \ // y2 = a <<< 12 + e + T ADDL y0, y2; \ // y2 = a <<< 12 + e + T
@ -267,15 +283,18 @@
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
VPSRLQ $9, XTMP2, XTMP4; \ // XTMP4 = XTMP2 rol 23 {xxxA} VPSRLQ $9, XTMP2, XTMP4; \ // XTMP4 = XTMP2 rol 23 {xxxA}
RORXL $23, y2, y0; \ MOVL y2, y0; \
RORXL $15, y2, d; \ ROLL $9, y0; \
MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XTMP2, XTMP4, XTMP4; \ // XTMP4 = XTMP2 XOR (XTMP2 rol 23 {xxxA}) VPXOR XTMP2, XTMP4, XTMP4; \ // XTMP4 = XTMP2 XOR (XTMP2 rol 23 {xxxA})
#define ROUND_AND_SCHED_N_1_2(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_1_2(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 2 ############################// ; \ // ############################# RND N + 2 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
ADDL y0, y2; \ // y2 = a <<< 12 + e + T ADDL y0, y2; \ // y2 = a <<< 12 + e + T
@ -305,15 +324,18 @@
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
VPOR XTMP3, XTMP4, XTMP4; \ // XTMP4 = (W[-3] rol 15) {DCBA} VPOR XTMP3, XTMP4, XTMP4; \ // XTMP4 = (W[-3] rol 15) {DCBA}
RORXL $23, y2, y0; \ MOVL y2, y0; \
RORXL $15, y2, d; \ ROLL $9, y0; \
MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XTMP1, XTMP4, XTMP4; \ // XTMP4 = W[-9] ^ W[-16] ^ (W[-3] rol 15) {DCBA} VPXOR XTMP1, XTMP4, XTMP4; \ // XTMP4 = W[-9] ^ W[-16] ^ (W[-3] rol 15) {DCBA}
#define ROUND_AND_SCHED_N_1_3(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \ #define ROUND_AND_SCHED_N_1_3(disp, const, a, b, c, d, e, f, g, h, XWORD0, XWORD1, XWORD2, XWORD3) \
; \ // ############################# RND N + 3 ############################// ; \ // ############################# RND N + 3 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
ADDL y0, y2; \ // y2 = a <<< 12 + e + T ADDL y0, y2; \ // y2 = a <<< 12 + e + T
@ -343,8 +365,10 @@
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
VPXOR XTMP3, XTMP1, XTMP1; \ // XTMP1 = XTMP4 ^ (XTMP4 rol 15 {DCBA}) ^ (XTMP4 rol 23 {DCBA}) VPXOR XTMP3, XTMP1, XTMP1; \ // XTMP1 = XTMP4 ^ (XTMP4 rol 15 {DCBA}) ^ (XTMP4 rol 23 {DCBA})
RORXL $23, y2, y0; \ MOVL y2, y0; \
RORXL $15, y2, d; \ ROLL $9, y0; \
MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)
VPXOR XTMP1, XTMP0, XWORD0; \ // XWORD0 = {W[3], W[2], W[1], W[0]} VPXOR XTMP1, XTMP0, XWORD0; \ // XWORD0 = {W[3], W[2], W[1], W[0]}
@ -352,7 +376,8 @@
// For rounds [16 - 64) // For rounds [16 - 64)
#define DO_ROUND_N_1(disp, idx, const, a, b, c, d, e, f, g, h) \ #define DO_ROUND_N_1(disp, idx, const, a, b, c, d, e, f, g, h) \
; \ // ############################# RND N + 0 ############################// ; \ // ############################# RND N + 0 ############################//
RORXL $20, a, y0; \ // y0 = a <<< 12 MOVL a, y0; \
ROLL $12, y0; \ // y0 = a <<< 12
MOVL e, y2; \ MOVL e, y2; \
ADDL $const, y2; \ ADDL $const, y2; \
ADDL y0, y2; \ // y2 = a <<< 12 + e + T ADDL y0, y2; \ // y2 = a <<< 12 + e + T
@ -380,8 +405,10 @@
ROLL $9, b; \ ROLL $9, b; \
ROLL $19, f; \ ROLL $19, f; \
; \ ; \
RORXL $23, y2, y0; \ MOVL y2, y0; \
RORXL $15, y2, d; \ ROLL $9, y0; \
MOVL y2, d; \
ROLL $17, d; \
XORL y0, d; \ XORL y0, d; \
XORL y2, d; \ // d = P(tt2) XORL y2, d; \ // d = P(tt2)