skein/amd64/skein_block_asm.s

307 F_O         =   -FRAME_OFFS
326     movq    %rdi, ctxPtr+F_O(%rbp)  #save caller's parameters on the stack
327     movq    %rsi, blkPtr+F_O(%rbp)
328     movq    %rdx, blkCnt+F_O(%rbp)
329     movq    %rcx, bitAdd+F_O(%rbp)
371     movq    ctxPtr+F_O(%rbp),%rsi   #h (pointer)
373     movq    blkPtr+F_O(%rbp),%rcx   #blkPtr
374     leaq    Wcopy +F_O(%rbp),%r8    #wPtr
375     leaq    ksKey +F_O(%rbp),%r9    #key pointer
376     leaq    ksTwk +F_O(%rbp),%rax   #tweak pointer
479     addq    bitAdd+F_O(%rbp)  ,%r13  #computed updated tweak value T0
491     movq    blkPtr +F_O(%rbp) ,%rsi  #esi --> input block
500     movq    %rax,Wcopy+ 0+F_O(%rbp)  #save copy of input block
501     movq    %rbx,Wcopy+ 8+F_O(%rbp)
502     movq    %rcx,Wcopy+16+F_O(%rbp)
503     movq    %rdx,Wcopy+24+F_O(%rbp)
514     movq    %r8 ,ksKey+ 0+F_O(%rbp)  #save key schedule on stack for Skein_Debug_Block
515     movq    %r9 ,ksKey+ 8+F_O(%rbp)
516     movq    %r10,ksKey+16+F_O(%rbp)
517     movq    %r11,ksKey+24+F_O(%rbp)
518     movq    %r12,ksKey+32+F_O(%rbp)
520     movq    %r13,ksTwk+ 0+F_O(%rbp)
521     movq    %r14,ksTwk+ 8+F_O(%rbp)
522     movq    %r15,ksTwk+16+F_O(%rbp)
534     movq    %r8 ,ksKey+40+F_O(%rbp)  #save key schedule on stack for looping code
535     movq    %r9 ,ksKey+ 8+F_O(%rbp)
536     movq    %r10,ksKey+16+F_O(%rbp)
537     movq    %r11,ksKey+24+F_O(%rbp)
538     movq    %r12,ksKey+32+F_O(%rbp)
540     movq    %r13,ksTwk+24+F_O(%rbp)
541     movq    %r14,ksTwk+ 8+F_O(%rbp)
542     movq    %r15,ksTwk+16+F_O(%rbp)
545     movq    %rsi,blkPtr  +F_O(%rbp)  #update block pointer
567                     movq ksKey+8*1+F_O(%rbp,%rdi,8),%r8
580                     movq ksTwk+8*1+F_O(%rbp,%rdi,8),%r13
589                     movq ksKey+8*2+F_O(%rbp,%rdi,8),%r9
595                     movq ksKey+8*4+F_O(%rbp,%rdi,8),%r11
610                     movq ksKey+8*3+F_O(%rbp,%rdi,8),%r10
616                     movq %r8,ksKey+8*6+F_O(%rbp,%rdi,8)  #"rotate" the key
625                     addq      ksTwk+8*2+F_O(%rbp,%rdi,8),%r10  #precompute key + tweak
626                     movq %r13,ksTwk+8*4+F_O(%rbp,%rdi,8)       #"rotate" the tweak
656     movq    ctxPtr +F_O(%rbp),%rdi           #restore rdi --> context
661     xorq    Wcopy + 0+F_O (%rbp),%rax
662     xorq    Wcopy + 8+F_O (%rbp),%rbx
663     xorq    Wcopy +16+F_O (%rbp),%rcx
664     xorq    Wcopy +24+F_O (%rbp),%rdx
674     decq    blkCnt+F_O(%rbp)
686     movq    %rax,X_stk+ 0+F_O(%rbp)  #save X[] state on stack so debug routines can access it
687     movq    %rbx,X_stk+ 8+F_O(%rbp)  #(use FP_ since rsp has changed!)
688     movq    %rcx,X_stk+16+F_O(%rbp)
689     movq    %rdi,X_stk+24+F_O(%rbp)
691     movq    ctxPtr+F_O(%rbp),%rsi    #ctx_hdr_ptr
752     R_512_OneRound  8, 9,10,11,12,13,14,15,%((_RR_)+0),<movq ksKey+8*(((_II_)+3) % 9)+F_O(%rbp),%rax>,,<movq ksKey+8*(((_II_)+4) % 9)+F_O(%rbp),%rbx>
753     R_512_OneRound 10, 9,12,15,14,13, 8,11,%((_RR_)+1),<movq ksKey+8*(((_II_)+5) % 9)+F_O(%rbp),%rcx>,,<movq ksKey+8*(((_II_)+6) % 9)+F_O(%rbp),%rdx>
754     R_512_OneRound 12, 9,14,11, 8,13,10,15,%((_RR_)+2),<movq ksKey+8*(((_II_)+7) % 9)+F_O(%rbp),%rsi>,,<addq ksTwk+8*(((_II_)+0) % 3)+F_O(%rbp),%rcx>
755     R_512_OneRound 14, 9, 8,15,10,13,12,11,%((_RR_)+3),<addq ksTwk+8*(((_II_)+1) % 3)+F_O(%rbp),%rdx>,
757     addq    ksKey+8*(((_II_)+0)%9)+F_O(%rbp),%r8
759     addq    ksKey+8*(((_II_)+1)%9)+F_O(%rbp),%r9
761     addq    ksKey+8*(((_II_)+2)%9)+F_O(%rbp),%r10
768     R_512_OneRound  8, 9,10,11,12,13,14,15,%((_RR_)+0),<movq ksKey+8*6+F_O(%rbp,%rdi,8),%rdx>,<movq      ksTwk-8*1+F_O(%rbp,%rdi,8),%rax>,<movq      ksKey-8*1+F_O(%rbp,%rdi,8),%rsi>
769     R_512_OneRound 10, 9,12,15,14,13, 8,11,%((_RR_)+1),<movq ksKey+8*5+F_O(%rbp,%rdi,8),%rcx>,<movq %rax,ksTwk+8*2+F_O(%rbp,%rdi,8)     >,<movq %rsi,ksKey+8*8+F_O(%rbp,%rdi,8)>
770     R_512_OneRound 12, 9,14,11, 8,13,10,15,%((_RR_)+2),<movq ksKey+8*4+F_O(%rbp,%rdi,8),%rbx>,<addq      ksTwk+8*1+F_O(%rbp,%rdi,8),%rdx>,<movq      ksKey+8*7+F_O(%rbp,%rdi,8),%rsi>
771     R_512_OneRound 14, 9, 8,15,10,13,12,11,%((_RR_)+3),<movq ksKey+8*3+F_O(%rbp,%rdi,8),%rax>,<addq      ksTwk+8*0+F_O(%rbp,%rdi,8),%rcx>
773     addq    ksKey+8*0+F_O(%rbp,%rdi,8),%r8
776     addq    ksKey+8*1+F_O(%rbp,%rdi,8),%r9
779     addq    ksKey+8*2+F_O(%rbp,%rdi,8),%r10
806     addq    bitAdd+F_O(%rbp),%rax     #computed updated tweak value T0
810     movq    %rax,ksTwk+ 0+F_O(%rbp)
812     movq    blkPtr +F_O(%rbp),%rsi    #%rsi --> input block
813     movq    %rbx,ksTwk+ 8+F_O(%rbp)
814     movq    %rcx,ksTwk+16+F_O(%rbp)
818       movq  %r\_Rn_,ksKey+8*(_Rn_-8)+F_O(%rbp)
820       movq  %rdx,ksKey+8*(8)+F_O(%rbp)#save key schedule parity
833     movq    %rax,Wcopy+ 0+F_O(%rbp)   #keep local copy for feedforward
834     movq    %rbx,Wcopy+ 8+F_O(%rbp)
837     movq    %rcx,Wcopy+16+F_O(%rbp)
838     movq    %rdx,Wcopy+24+F_O(%rbp)
848     movq    %rax,Wcopy+32+F_O(%rbp)
849     movq    %rbx,Wcopy+40+F_O(%rbp)
850     movq    %rcx,Wcopy+48+F_O(%rbp)
851     movq    %rdx,Wcopy+56+F_O(%rbp)
862     movq    %rsi,blkPtr+F_O(%rbp)     #update block pointer
887     movq    ctxPtr +F_O(%rbp),%rdi           #restore rdi --> context
896       xorq  Wcopy+8*(_Rn_-8)+F_O(%rbp),%r\_Rn_  #feedforward XOR
905     decq    blkCnt+F_O(%rbp)
919     movq    %r\_Rn_,X_stk+8*(_Rn_-8)+F_O(%rbp)
921     movq    ctxPtr+F_O(%rbp),%rsi    #ctx_hdr_ptr
1103     addq         bitAdd+  F_O(%rbp),%r8    #computed updated tweak value T0
1107     movq    %r8 ,ksTwk+ 0+F_O(%rbp)
1108     movq    %r9 ,ksTwk+ 8+F_O(%rbp)        #keep values in %r8 ,%r9  for initial tweak injection below
1109     movq    %r10,ksTwk+16+F_O(%rbp)
1113     movq         blkPtr +F_O(%rbp),%rsi    # rsi --> input block
1121       movq  %r14,ksKey +8*_rN_+F_O(%rbp)   #save key schedule word on stack
1122       movq  %r15,Wcopy +8*_rN_+F_O(%rbp)   #save local msg Wcopy
1134       movq  %rcx,Wcopy+8*_oo_+F_O(%rbp)    #save copy of msg word for feedforward
1142     movq    %rax,ksKey+8*WCNT+F_O(%rbp)    #save key schedule parity
1147     movq     %rsi,blkPtr+F_O(%rbp)         #save bumped msg ptr