Lines Matching refs:W_TMP
121 // symbolizing registers or stack memory with algorithmic variables W0,W4,...,W28 + W_TMP, W_TMP2, and XMM_SHUFB_BSWAP for code with ssse3 support
123 #define W_TMP %xmm0
194 1. W_TMP = new 16 bytes from MESSAGE[]
195 2. W_TMP = pshufb(W_TMP, XMM_SHUFB_BSWAP); save to W circular buffer for updating W
197 4. save quadruple W[i]+K[i] = W_TMP in the stack memory;
206 xmovu $0*4(BUFFER_PTR), W_TMP // read 16-bytes into W_TMP, BUFFER_PTR possibly not 16-byte aligned
209 xmovu $0*4(T1), W_TMP // read 16-bytes into W_TMP, BUFFER_PTR possibly not 16-byte aligned
214 pshufb XMM_SHUFB_BSWAP, W_TMP // convert W_TMP from little-endian into big-endian
215 xmov W_TMP, $0 // save W_TMP in the circular buffer
220 paddd (K_BASE), W_TMP // W_TMP += {K,K,K,K};
223 paddd (T1), W_TMP // W_TMP += {K,K,K,K};
228 xmov W_TMP, WK($0&~3) // save quadruple W[i]+K in the stack memory, which would be used later for updating the hashes A/B/C/D/E
234 2. load the 16-bytes from the aligned stack memory into W_TMP
268 xmov 14*16(sp), W_TMP // load the bswapped 16-bytes from the aligned stack memory
269 xmov W_TMP, $0 // save W = W_TMP in the circular buffer
284 1. W = W3 ^ W8 ^ W14 ^ W16; W_TMP = W; W_TMP2 = (W[i] 0 0 0);
285 2. W_TMP = (W3 ^ W8 ^ W14 ^ W16) rol 1; split (W[i] 0 0 0) rol 2 in W_TMP2 and W
286 3. W = W_TMP = W_TMP ^ W_TMP2 ^ W = (W3 ^ W8 ^ W14 ^ W16) rol 1 ^ (W[i] 0 0 0) rol 2; WK = W _TMP+K;
293 xmov $3, W_TMP // W_TMP = W4
294 psrldq $$4, W_TMP // W_TMP = W3
299 pxor $0, W_TMP // W_TMP = W3 ^ W16
300 pxor W_TMP, $1 // W = W3 ^ W16 ^ W8 ^ W14
302 xmov $1, W_TMP // W_TMP = W3 ^ W16 ^ W8 ^ W14
308 pslld $$1, W_TMP // (W3 ^ W16 ^ W8 ^ W14)<<1
309 por $0, W_TMP // W_TMP = (W3 ^ W16 ^ W8 ^ W14) rol 1
319 pxor $0, W_TMP
320 pxor W_TMP2, W_TMP // W_TMP = (W3 ^ W16 ^ W8 ^ W14) rol 1 ^ (W[i] 0 0 0) rol 2
321 xmov W_TMP, $0 // save W = W_TMP in the W circular buffer
323 paddd $2(K_BASE), W_TMP // W+K
325 paddd $2(T1), W_TMP // W+K
327 xmov W_TMP, WK($1&~3) // save WK = W+K for later update of the hashes A/B/C/D/E
336 xmov $0, W_TMP // W16 = (w13 w14 w15 w16)
338 psrldq $$8, W_TMP // shift right to make (0 0 w13 w14)
339 por W_TMP, $4 // W = W14 = (w11 w12 w13 w14)
341 xmov $3, W_TMP // W_TMP = W4 = (w1 w2 w3 w4)
342 psrldq $$4, W_TMP // W_TMP = W3 = (0 w1 w2 w3)
362 xmov $2, W_TMP // (w1 w2 w3 w4)
364 palignr $$8, $1, W_TMP // W_tmp = (w3 w4 w5 w6) = W6;
369 xmov $2, W_TMP // (w1 w2 w3 w4)
372 pslldq $$8, W_TMP // (w3 w4 0 0)
374 por W_TMP2, W_TMP // W_tmp = (w3 w4 w5 w6) = W6
379 xmov $3, W_TMP // W32
380 pxor $0, W_TMP // W28 ^ W32
381 xmov W_TMP, $3 // W = W28 ^ W32;
382 xmov $2, W_TMP // W4
383 palignr $$8, $1, W_TMP // W_tmp = (w3 w4 w5 w6) = W6;
388 xmov $3, W_TMP // W32
389 pxor $0, W_TMP // W28 ^ W32
390 xmov W_TMP, $3 // W = W28 ^ W32
391 xmov $2, W_TMP // W4 = (w1 w2 w3 w4)
393 pslldq $$8, W_TMP // (w3 w4 0 0)
395 por W_TMP2, W_TMP // W_tmp = (w3 w4 w5 w6) = W6
399 pxor $0, W_TMP // W_tmp = W6 ^ W16
400 pxor $1, W_TMP // W_tmp = W6 ^ W16 ^ W28 ^ W32
401 xmov W_TMP, $1 // W = W_tmp = W6 ^ W16 ^ W28 ^ W32
406 pslld $$2, W_TMP // W << 2
407 por $0, W_TMP // W_tmp = (W6 ^ W16 ^ W28 ^ W32) rol 2
416 pslld $$2, W_TMP // W_tmp << 2
417 por $0, W_TMP // W_tmp = (W6 ^ W16 ^ W28 ^ W32) rol 2
422 xmov W_TMP, $0 // W = (W6 ^ W16 ^ W28 ^ W32) rol 2
423 paddd $2(K_BASE), W_TMP // W + K
424 xmov W_TMP, WK($1&~3) // write W+K
427 xmov W_TMP, $0 // W = (W6 ^ W16 ^ W28 ^ W32) rol 2
428 paddd $2(T1), W_TMP // W_tmp = W + K
429 xmov W_TMP, WK($1&~3) // write WK
531 W_PRECALC_00_15_0 0 // W_TMP = (BUFFER_PTR)
532 W_PRECALC_00_15_1 W0 // convert W_TMP to big-endian, and save W0 = W_TMP
533 W_PRECALC_00_15_2 // W_TMP = W0 + K
534 W_PRECALC_00_15_3 3 // (sp) = W_TMP = W0 + K
537 W_PRECALC_00_15_0 4 // W_TMP = 16(BUFFER_PTR)
538 W_PRECALC_00_15_1 W28 // convert W_TMP to big-endian, and save W28 = W_TMP
539 W_PRECALC_00_15_2 // W_TMP = W28 + K
540 W_PRECALC_00_15_3 7 // 16(sp) = W_TMP = W28 + K
543 W_PRECALC_00_15_0 8 // W_TMP = 32(BUFFER_PTR)
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
549 W_PRECALC_00_15_0 12 // W_TMP = 48(BUFFER_PTR)
550 W_PRECALC_00_15_1 W20 // convert W_TMP to big-endian, and save W20 = W_TMP
551 W_PRECALC_00_15_2 // W_TMP = W20 + K
552 W_PRECALC_00_15_3 15 // 48(sp) = W_TMP = W20 + K
765 W_PRECALC_00_15_0 0 // W_TMP = (BUFFER_PTR)
767 W_PRECALC_00_15_1 W0 // convert W_TMP to big-endian, and save W0 = W_TMP
769 W_PRECALC_00_15_2 // W_TMP = W0 + K
771 W_PRECALC_00_15_3 3 // (sp) = W_TMP = W0 + K
775 W_PRECALC_00_15_0 4 // W_TMP = 16(BUFFER_PTR)
777 W_PRECALC_00_15_1 W28 // convert W_TMP to big-endian, and save W28 = W_TMP
779 W_PRECALC_00_15_2 // W_TMP = W28 + K
781 W_PRECALC_00_15_3 7 // 16(sp) = W_TMP = W28 + K[0]
785 W_PRECALC_00_15_0 8 // W_TMP = 32(BUFFER_PTR)
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
795 W_PRECALC_00_15_0 12 // W_TMP = 48(BUFFER_PTR)
797 W_PRECALC_00_15_1 W20 // convert W_TMP to big-endian, and save W20 = W_TMP
799 W_PRECALC_00_15_2 // W_TMP = W20 + K
801 W_PRECALC_00_15_3 15 // 48(sp) = W_TMP = W20 + K
836 W_PRECALC_00_15_0 0 // W_TMP = (BUFFER_PTR)
837 W_PRECALC_00_15_1 W0 // convert W_TMP to big-endian, and save W0 = W_TMP
838 W_PRECALC_00_15_2 // W_TMP = W0 + K
839 W_PRECALC_00_15_3 3 // (sp) = W_TMP = W0 + K
842 W_PRECALC_00_15_0 4 // W_TMP = 16(BUFFER_PTR)
843 W_PRECALC_00_15_1 W28 // convert W_TMP to big-endian, and save W28 = W_TMP
844 W_PRECALC_00_15_2 // W_TMP = W28 + K
845 W_PRECALC_00_15_3 7 // 16(sp) = W_TMP = W28 + K
848 W_PRECALC_00_15_0 8 // W_TMP = 32(BUFFER_PTR)
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
854 W_PRECALC_00_15_0 12 // W_TMP = 48(BUFFER_PTR)
855 W_PRECALC_00_15_1 W20 // convert W_TMP to big-endian, and save W20 = W_TMP
856 W_PRECALC_00_15_2 // W_TMP = W20 + K
857 W_PRECALC_00_15_3 15 // 48(sp) = W_TMP = W20 + K
1084 W_PRECALC_00_15_0 0 // W_TMP = (BUFFER_PTR)
1086 W_PRECALC_00_15_1 W0 // convert W_TMP to big-endian, and save W0 = W_TMP
1088 W_PRECALC_00_15_2 // W_TMP = W0 + K
1090 W_PRECALC_00_15_3 3 // (sp) = W_TMP = W0 + K
1094 W_PRECALC_00_15_0 4 // W_TMP = 16(BUFFER_PTR)
1096 W_PRECALC_00_15_1 W28 // convert W_TMP to big-endian, and save W28 = W_TMP
1098 W_PRECALC_00_15_2 // W_TMP = W28 + K
1100 W_PRECALC_00_15_3 7 // 16(sp) = W_TMP = W28 + K[0]
1104 W_PRECALC_00_15_0 8 // W_TMP = 32(BUFFER_PTR)
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
1114 W_PRECALC_00_15_0 12 // W_TMP = 48(BUFFER_PTR)
1116 W_PRECALC_00_15_1 W20 // convert W_TMP to big-endian, and save W20 = W_TMP
1118 W_PRECALC_00_15_2 // W_TMP = W20 + K
1120 W_PRECALC_00_15_3 15 // 48(sp) = W_TMP = W20 + K