crypto/intel/sha1edp.s

77 	2. The registers are used in a circular buffering mode. For example, we start with W28,W24,...,W0 (with W0 indicating the most recent 16-byte)
78 		i=0, W28,W24,...,W0
79 		i=4, W24,W20,...,W28
80 		i=8, W20,W16,...,W24
112 #define stack_size	(12+16*2+16*11+16*4)			// 12-bytes (alignment) + extra 2 + 3 (W24/W28/XMM_SHUFB_BSWAP) + 8 (xmm0-xmm7) + 4 (WK(t))
132 #define W24 	%xmm8
136 #define W24     12*16(sp)
377 	// this is a variant of W_PRECALC_32_79_0_ssse3 for i386 (as W24/W28 are stored in memory, not in registers)
386 	// this is a variant of W_PRECALC_32_79_0_nossse3 for i386 (as W24/W28 are stored in memory, not in registers)
410 	// this is a variant of W_PRECALC_32_79_2 for i386 (as W24/W28 are stored in memory, not in registers)
411 	// this should be used when the input is either W24 or W28 on i386 architecture
530 	// i=0 	: W28,W24,W20,W16,W12,W8,W4,W0
536 	// i=4	: W24,W20,W16,W12,W8,W4,W0,W28
542 	// i=8  : W20,W16,W12,W8,W4,W0,W28,W24
544 	W_PRECALC_00_15_1	W24					// convert W_TMP to big-endian, and save W24 = W_TMP
545 	W_PRECALC_00_15_2						// W_TMP = W24 + K
546 	W_PRECALC_00_15_3	11					// 32(sp) = W_TMP = W24 + K
548 	// i=12 : W16,W12,W8,W4,W0,W28,W24,W20
559 	// i=16 : W12,W8,W4,W0,W28,W24,W20,W16
560 	W_PRECALC_16_31_0	W0,W28,W24,W20,W16
569 	// i=20 : W8,W4,W0,W28,W24,W20,W16,W12
570 	W_PRECALC_16_31_0	W28,W24,W20,W16,W12
579 	// i=24 : W4,W0,W28,W24,W20,W16,W12,W8
580 	W_PRECALC_16_31_0	W24,W20,W16,W12,W8
582 	W_PRECALC_16_31_1	W24,W8
589 	// i=28 : W0,W28,W24,W20,W16,W12,W8,W4
599 	// i=32 : W28,W24,W20,W16,W12,W8,W4,W0
611 	// i=36 : W24,W20,W16,W12,W8,W4,W0,W28
613 	W_PRECALC_32_79_0	W24,W4,W0,W28
615 	W_PRECALC_32_79_0_i386	W24,W4,W0,W28
629 	// i=40 : W20,W16,W12,W8,W4,W0,W28,W24
633 	W_PRECALC_32_79_0	W20,W0,W28,W24
635 	W_PRECALC_32_79_0_i386	W20,W0,W28,W24
638 	W_PRECALC_32_79_1	W8,W24
641 	W_PRECALC_32_79_2	W24
643 	W_PRECALC_32_79_2_i386	W24
646 	W_PRECALC_32_79_3	W24,26,K_XMM
649 	// i=44 : W16,W12,W8,W4,W0,W28,W24,W20
650 	W_PRECALC_32_79_0	W16,W28,W24,W20
659 	// i=48 : W12,W8,W4,W0,W28,W24,W20,W16
660 	W_PRECALC_32_79_0	W12,W24,W20,W16
669 	// i=52 : W8,W4,W0,W28,W24,W20,W16,W12
681 	// i=56 : W4,W0,W28,W24,W20,W16,W12,W8
684 	W_PRECALC_32_79_1	W24,W8
691 	// i=60 : W0,W28,W24,W20,W16,W12,W8,W4
703 	// i=64 : W28,W24,W20,W16,W12,W8,W4,W0
713 	// i=68 : W24,W20,W16,W12,W8,W4,W0,W28
715 	W_PRECALC_32_79_0	W24,W4,W0,W28
717 	W_PRECALC_32_79_0_i386	W24,W4,W0,W28
731 	// i=72 : W20,W16,W12,W8,W4,W0,W28,W24
733 	W_PRECALC_32_79_0	W20,W0,W28,W24
735 	W_PRECALC_32_79_0_i386	W20,W0,W28,W24
738 	W_PRECALC_32_79_1	W8,W24
741 	W_PRECALC_32_79_2	W24
743 	W_PRECALC_32_79_2_i386	W24
746 	W_PRECALC_32_79_3	W24,58,K_XMM
751 	// i=76 : W16,W12,W8,W4,W0,W28,W24,W20
752 	W_PRECALC_32_79_0	W16,W28,W24,W20
764 	// i=0  : W28,W24,W20,W16,W12,W8,W4,W0
774 	// i=4  : W24,W20,W16,W12,W8,W4,W0,W28
784 	// i=8  : W20,W16,W12,W8,W4,W0,W28,W24
787 	W_PRECALC_00_15_1	W24					// convert W_TMP to big-endian, and save W24 = W_TMP
789 	W_PRECALC_00_15_2						// W_TMP = W24 + K
791 	W_PRECALC_00_15_3	11					// 32(sp) = W_TMP = W24 + K
794 	// i=12 : W16,W12,W8,W4,W0,W28,W24,W20
835 	// i=0 	: W28,W24,W20,W16,W12,W8,W4,W0
841 	// i=4	: W24,W20,W16,W12,W8,W4,W0,W28
847 	// i=8  : W20,W16,W12,W8,W4,W0,W28,W24
849 	W_PRECALC_00_15_1	W24					// convert W_TMP to big-endian, and save W24 = W_TMP
850 	W_PRECALC_00_15_2						// W_TMP = W24 + K
851 	W_PRECALC_00_15_3	11					// 32(sp) = W_TMP = W24 + K
853 	// i=12 : W16,W12,W8,W4,W0,W28,W24,W20
864 	// circular buffer : W12,W8,W4,W0,W28,W24,W20,W16
865 	W_PRECALC_16_31_0	W0,W28,W24,W20,W16
875 	// W8,W4,W0,W28,W24,W20,W16,W12
876 	W_PRECALC_16_31_0	W28,W24,W20,W16,W12
887 	// W4,W0,W28,W24,W20,W16,W12,W8
888 	W_PRECALC_16_31_0	W24,W20,W16,W12,W8
890 	W_PRECALC_16_31_1	W24,W8
899 	// W0,W28,W24,W20,W16,W12,W8,W4
911 	// W28,W24,W20,W16,W12,W8,W4,W0
922 	// W24,W20,W16,W12,W8,W4,W0,W28
924 	W_PRECALC_32_79_0	W24,W4,W0,W28
926 	W_PRECALC_32_79_0_i386	W24,W4,W0,W28
943 	// W20,W16,W12,W8,W4,W0,W28,W24
945 	W_PRECALC_32_79_0	W20,W0,W28,W24
947 	W_PRECALC_32_79_0_i386	W20,W0,W28,W24
950 	W_PRECALC_32_79_1	W8,W24
953 	W_PRECALC_32_79_2	W24
955 	W_PRECALC_32_79_2_i386	W24
958 	W_PRECALC_32_79_3	W24,26,K_XMM
962 	// W16,W12,W8,W4,W0,W28,W24,W20
963 	W_PRECALC_32_79_0	W16,W28,W24,W20
973 	// W12,W8,W4,W0,W28,W24,W20,W16
974 	W_PRECALC_32_79_0	W12,W24,W20,W16
984 	// W8,W4,W0,W28,W24,W20,W16,W12
995 	// W4,W0,W28,W24,W20,W16,W12,W8
998 	W_PRECALC_32_79_1	W24,W8
1008 	// W0,W28,W24,W20,W16,W12,W8,W4
1019 	// W28,W24,W20,W16,W12,W8,W4,W0
1030 	// W24,W20,W16,W12,W8,W4,W0,W28
1032 	W_PRECALC_32_79_0	W24,W4,W0,W28
1034 	W_PRECALC_32_79_0_i386	W24,W4,W0,W28
1049 	// W20,W16,W12,W8,W4,W0,W28,W24
1051 	W_PRECALC_32_79_0	W20,W0,W28,W24
1053 	W_PRECALC_32_79_0_i386	W20,W0,W28,W24
1056 	W_PRECALC_32_79_1	W8,W24
1059 	W_PRECALC_32_79_2	W24
1061 	W_PRECALC_32_79_2_i386	W24
1064 	W_PRECALC_32_79_3	W24,58,K_XMM
1070 	// W16,W12,W8,W4,W0,W28,W24,W20
1071 	W_PRECALC_32_79_0	W16,W28,W24,W20
1083 	// i=0  : W28,W24,W20,W16,W12,W8,W4,W0
1093 	// i=4  : W24,W20,W16,W12,W8,W4,W0,W28
1103 	// i=8  : W20,W16,W12,W8,W4,W0,W28,W24
1106 	W_PRECALC_00_15_1	W24					// convert W_TMP to big-endian, and save W24 = W_TMP
1108 	W_PRECALC_00_15_2						// W_TMP = W24 + K
1110 	W_PRECALC_00_15_3	11					// 32(sp) = W_TMP = W24 + K
1113 	// i=12 : W16,W12,W8,W4,W0,W28,W24,W20