diff options
Diffstat (limited to 'crypto/sha/asm/sha512-ppc.pl')
| -rwxr-xr-x | crypto/sha/asm/sha512-ppc.pl | 460 |
1 files changed, 460 insertions, 0 deletions
diff --git a/crypto/sha/asm/sha512-ppc.pl b/crypto/sha/asm/sha512-ppc.pl new file mode 100755 index 000000000000..6b44a68e599e --- /dev/null +++ b/crypto/sha/asm/sha512-ppc.pl @@ -0,0 +1,460 @@ +#!/usr/bin/env perl + +# ==================================================================== +# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL +# project. The module is, however, dual licensed under OpenSSL and +# CRYPTOGAMS licenses depending on where you obtain it. For further +# details see http://www.openssl.org/~appro/cryptogams/. +# ==================================================================== + +# I let hardware handle unaligned input, except on page boundaries +# (see below for details). Otherwise straightforward implementation +# with X vector in register bank. The module is big-endian [which is +# not big deal as there're no little-endian targets left around]. + +# sha256 | sha512 +# -m64 -m32 | -m64 -m32 +# --------------------------------------+----------------------- +# PPC970,gcc-4.0.0 +50% +38% | +40% +410%(*) +# Power6,xlc-7 +150% +90% | +100% +430%(*) +# +# (*) 64-bit code in 32-bit application context, which actually is +# on TODO list. It should be noted that for safe deployment in +# 32-bit *mutli-threaded* context asyncronous signals should be +# blocked upon entry to SHA512 block routine. This is because +# 32-bit signaling procedure invalidates upper halves of GPRs. +# Context switch procedure preserves them, but not signaling:-( + +# Second version is true multi-thread safe. Trouble with the original +# version was that it was using thread local storage pointer register. +# Well, it scrupulously preserved it, but the problem would arise the +# moment asynchronous signal was delivered and signal handler would +# dereference the TLS pointer. While it's never the case in openssl +# application or test suite, we have to respect this scenario and not +# use TLS pointer register. Alternative would be to require caller to +# block signals prior calling this routine. For the record, in 32-bit +# context R2 serves as TLS pointer, while in 64-bit context - R13. + +$flavour=shift; +$output =shift; + +if ($flavour =~ /64/) { + $SIZE_T=8; + $LRSAVE=2*$SIZE_T; + $STU="stdu"; + $UCMP="cmpld"; + $SHL="sldi"; + $POP="ld"; + $PUSH="std"; +} elsif ($flavour =~ /32/) { + $SIZE_T=4; + $LRSAVE=$SIZE_T; + $STU="stwu"; + $UCMP="cmplw"; + $SHL="slwi"; + $POP="lwz"; + $PUSH="stw"; +} else { die "nonsense $flavour"; } + +$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1; +( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or +( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or +die "can't locate ppc-xlate.pl"; + +open STDOUT,"| $^X $xlate $flavour $output" || die "can't call $xlate: $!"; + +if ($output =~ /512/) { + $func="sha512_block_data_order"; + $SZ=8; + @Sigma0=(28,34,39); + @Sigma1=(14,18,41); + @sigma0=(1, 8, 7); + @sigma1=(19,61, 6); + $rounds=80; + $LD="ld"; + $ST="std"; + $ROR="rotrdi"; + $SHR="srdi"; +} else { + $func="sha256_block_data_order"; + $SZ=4; + @Sigma0=( 2,13,22); + @Sigma1=( 6,11,25); + @sigma0=( 7,18, 3); + @sigma1=(17,19,10); + $rounds=64; + $LD="lwz"; + $ST="stw"; + $ROR="rotrwi"; + $SHR="srwi"; +} + +$FRAME=32*$SIZE_T+16*$SZ; +$LOCALS=6*$SIZE_T; + +$sp ="r1"; +$toc="r2"; +$ctx="r3"; # zapped by $a0 +$inp="r4"; # zapped by $a1 +$num="r5"; # zapped by $t0 + +$T ="r0"; +$a0 ="r3"; +$a1 ="r4"; +$t0 ="r5"; +$t1 ="r6"; +$Tbl="r7"; + +$A ="r8"; +$B ="r9"; +$C ="r10"; +$D ="r11"; +$E ="r12"; +$F ="r13"; $F="r2" if ($SIZE_T==8);# reassigned to exempt TLS pointer +$G ="r14"; +$H ="r15"; + +@V=($A,$B,$C,$D,$E,$F,$G,$H); +@X=("r16","r17","r18","r19","r20","r21","r22","r23", + "r24","r25","r26","r27","r28","r29","r30","r31"); + +$inp="r31"; # reassigned $inp! aliases with @X[15] + +sub ROUND_00_15 { +my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_; +$code.=<<___; + $LD $T,`$i*$SZ`($Tbl) + $ROR $a0,$e,$Sigma1[0] + $ROR $a1,$e,$Sigma1[1] + and $t0,$f,$e + andc $t1,$g,$e + add $T,$T,$h + xor $a0,$a0,$a1 + $ROR $a1,$a1,`$Sigma1[2]-$Sigma1[1]` + or $t0,$t0,$t1 ; Ch(e,f,g) + add $T,$T,@X[$i] + xor $a0,$a0,$a1 ; Sigma1(e) + add $T,$T,$t0 + add $T,$T,$a0 + + $ROR $a0,$a,$Sigma0[0] + $ROR $a1,$a,$Sigma0[1] + and $t0,$a,$b + and $t1,$a,$c + xor $a0,$a0,$a1 + $ROR $a1,$a1,`$Sigma0[2]-$Sigma0[1]` + xor $t0,$t0,$t1 + and $t1,$b,$c + xor $a0,$a0,$a1 ; Sigma0(a) + add $d,$d,$T + xor $t0,$t0,$t1 ; Maj(a,b,c) + add $h,$T,$a0 + add $h,$h,$t0 + +___ +} + +sub ROUND_16_xx { +my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_; +$i-=16; +$code.=<<___; + $ROR $a0,@X[($i+1)%16],$sigma0[0] + $ROR $a1,@X[($i+1)%16],$sigma0[1] + $ROR $t0,@X[($i+14)%16],$sigma1[0] + $ROR $t1,@X[($i+14)%16],$sigma1[1] + xor $a0,$a0,$a1 + $SHR $a1,@X[($i+1)%16],$sigma0[2] + xor $t0,$t0,$t1 + $SHR $t1,@X[($i+14)%16],$sigma1[2] + add @X[$i],@X[$i],@X[($i+9)%16] + xor $a0,$a0,$a1 ; sigma0(X[(i+1)&0x0f]) + xor $t0,$t0,$t1 ; sigma1(X[(i+14)&0x0f]) + add @X[$i],@X[$i],$a0 + add @X[$i],@X[$i],$t0 +___ +&ROUND_00_15($i,$a,$b,$c,$d,$e,$f,$g,$h); +} + +$code=<<___; +.machine "any" +.text + +.globl $func +.align 6 +$func: + $STU $sp,-$FRAME($sp) + mflr r0 + $SHL $num,$num,`log(16*$SZ)/log(2)` + + $PUSH $ctx,`$FRAME-$SIZE_T*22`($sp) + + $PUSH $toc,`$FRAME-$SIZE_T*20`($sp) + $PUSH r13,`$FRAME-$SIZE_T*19`($sp) + $PUSH r14,`$FRAME-$SIZE_T*18`($sp) + $PUSH r15,`$FRAME-$SIZE_T*17`($sp) + $PUSH r16,`$FRAME-$SIZE_T*16`($sp) + $PUSH r17,`$FRAME-$SIZE_T*15`($sp) + $PUSH r18,`$FRAME-$SIZE_T*14`($sp) + $PUSH r19,`$FRAME-$SIZE_T*13`($sp) + $PUSH r20,`$FRAME-$SIZE_T*12`($sp) + $PUSH r21,`$FRAME-$SIZE_T*11`($sp) + $PUSH r22,`$FRAME-$SIZE_T*10`($sp) + $PUSH r23,`$FRAME-$SIZE_T*9`($sp) + $PUSH r24,`$FRAME-$SIZE_T*8`($sp) + $PUSH r25,`$FRAME-$SIZE_T*7`($sp) + $PUSH r26,`$FRAME-$SIZE_T*6`($sp) + $PUSH r27,`$FRAME-$SIZE_T*5`($sp) + $PUSH r28,`$FRAME-$SIZE_T*4`($sp) + $PUSH r29,`$FRAME-$SIZE_T*3`($sp) + $PUSH r30,`$FRAME-$SIZE_T*2`($sp) + $PUSH r31,`$FRAME-$SIZE_T*1`($sp) + $PUSH r0,`$FRAME+$LRSAVE`($sp) + + $LD $A,`0*$SZ`($ctx) + mr $inp,r4 ; incarnate $inp + $LD $B,`1*$SZ`($ctx) + $LD $C,`2*$SZ`($ctx) + $LD $D,`3*$SZ`($ctx) + $LD $E,`4*$SZ`($ctx) + $LD $F,`5*$SZ`($ctx) + $LD $G,`6*$SZ`($ctx) + $LD $H,`7*$SZ`($ctx) + + bl LPICmeup +LPICedup: + andi. r0,$inp,3 + bne Lunaligned +Laligned: + add $num,$inp,$num + $PUSH $num,`$FRAME-$SIZE_T*24`($sp) ; end pointer + $PUSH $inp,`$FRAME-$SIZE_T*23`($sp) ; inp pointer + bl Lsha2_block_private + b Ldone + +; PowerPC specification allows an implementation to be ill-behaved +; upon unaligned access which crosses page boundary. "Better safe +; than sorry" principle makes me treat it specially. But I don't +; look for particular offending word, but rather for the input +; block which crosses the boundary. Once found that block is aligned +; and hashed separately... +.align 4 +Lunaligned: + subfic $t1,$inp,4096 + andi. $t1,$t1,`4096-16*$SZ` ; distance to closest page boundary + beq Lcross_page + $UCMP $num,$t1 + ble- Laligned ; didn't cross the page boundary + subfc $num,$t1,$num + add $t1,$inp,$t1 + $PUSH $num,`$FRAME-$SIZE_T*25`($sp) ; save real remaining num + $PUSH $t1,`$FRAME-$SIZE_T*24`($sp) ; intermediate end pointer + $PUSH $inp,`$FRAME-$SIZE_T*23`($sp) ; inp pointer + bl Lsha2_block_private + ; $inp equals to the intermediate end pointer here + $POP $num,`$FRAME-$SIZE_T*25`($sp) ; restore real remaining num +Lcross_page: + li $t1,`16*$SZ/4` + mtctr $t1 + addi r20,$sp,$LOCALS ; aligned spot below the frame +Lmemcpy: + lbz r16,0($inp) + lbz r17,1($inp) + lbz r18,2($inp) + lbz r19,3($inp) + addi $inp,$inp,4 + stb r16,0(r20) + stb r17,1(r20) + stb r18,2(r20) + stb r19,3(r20) + addi r20,r20,4 + bdnz Lmemcpy + + $PUSH $inp,`$FRAME-$SIZE_T*26`($sp) ; save real inp + addi $t1,$sp,`$LOCALS+16*$SZ` ; fictitious end pointer + addi $inp,$sp,$LOCALS ; fictitious inp pointer + $PUSH $num,`$FRAME-$SIZE_T*25`($sp) ; save real num + $PUSH $t1,`$FRAME-$SIZE_T*24`($sp) ; end pointer + $PUSH $inp,`$FRAME-$SIZE_T*23`($sp) ; inp pointer + bl Lsha2_block_private + $POP $inp,`$FRAME-$SIZE_T*26`($sp) ; restore real inp + $POP $num,`$FRAME-$SIZE_T*25`($sp) ; restore real num + addic. $num,$num,`-16*$SZ` ; num-- + bne- Lunaligned + +Ldone: + $POP r0,`$FRAME+$LRSAVE`($sp) + $POP $toc,`$FRAME-$SIZE_T*20`($sp) + $POP r13,`$FRAME-$SIZE_T*19`($sp) + $POP r14,`$FRAME-$SIZE_T*18`($sp) + $POP r15,`$FRAME-$SIZE_T*17`($sp) + $POP r16,`$FRAME-$SIZE_T*16`($sp) + $POP r17,`$FRAME-$SIZE_T*15`($sp) + $POP r18,`$FRAME-$SIZE_T*14`($sp) + $POP r19,`$FRAME-$SIZE_T*13`($sp) + $POP r20,`$FRAME-$SIZE_T*12`($sp) + $POP r21,`$FRAME-$SIZE_T*11`($sp) + $POP r22,`$FRAME-$SIZE_T*10`($sp) + $POP r23,`$FRAME-$SIZE_T*9`($sp) + $POP r24,`$FRAME-$SIZE_T*8`($sp) + $POP r25,`$FRAME-$SIZE_T*7`($sp) + $POP r26,`$FRAME-$SIZE_T*6`($sp) + $POP r27,`$FRAME-$SIZE_T*5`($sp) + $POP r28,`$FRAME-$SIZE_T*4`($sp) + $POP r29,`$FRAME-$SIZE_T*3`($sp) + $POP r30,`$FRAME-$SIZE_T*2`($sp) + $POP r31,`$FRAME-$SIZE_T*1`($sp) + mtlr r0 + addi $sp,$sp,$FRAME + blr + .long 0 + .byte 0,12,4,1,0x80,18,3,0 + .long 0 + +.align 4 +Lsha2_block_private: +___ +for($i=0;$i<16;$i++) { +$code.=<<___ if ($SZ==4); + lwz @X[$i],`$i*$SZ`($inp) +___ +# 64-bit loads are split to 2x32-bit ones, as CPU can't handle +# unaligned 64-bit loads, only 32-bit ones... +$code.=<<___ if ($SZ==8); + lwz $t0,`$i*$SZ`($inp) + lwz @X[$i],`$i*$SZ+4`($inp) + insrdi @X[$i],$t0,32,0 +___ + &ROUND_00_15($i,@V); + unshift(@V,pop(@V)); +} +$code.=<<___; + li $T,`$rounds/16-1` + mtctr $T +.align 4 +Lrounds: + addi $Tbl,$Tbl,`16*$SZ` +___ +for(;$i<32;$i++) { + &ROUND_16_xx($i,@V); + unshift(@V,pop(@V)); +} +$code.=<<___; + bdnz- Lrounds + + $POP $ctx,`$FRAME-$SIZE_T*22`($sp) + $POP $inp,`$FRAME-$SIZE_T*23`($sp) ; inp pointer + $POP $num,`$FRAME-$SIZE_T*24`($sp) ; end pointer + subi $Tbl,$Tbl,`($rounds-16)*$SZ` ; rewind Tbl + + $LD r16,`0*$SZ`($ctx) + $LD r17,`1*$SZ`($ctx) + $LD r18,`2*$SZ`($ctx) + $LD r19,`3*$SZ`($ctx) + $LD r20,`4*$SZ`($ctx) + $LD r21,`5*$SZ`($ctx) + $LD r22,`6*$SZ`($ctx) + addi $inp,$inp,`16*$SZ` ; advance inp + $LD r23,`7*$SZ`($ctx) + add $A,$A,r16 + add $B,$B,r17 + $PUSH $inp,`$FRAME-$SIZE_T*23`($sp) + add $C,$C,r18 + $ST $A,`0*$SZ`($ctx) + add $D,$D,r19 + $ST $B,`1*$SZ`($ctx) + add $E,$E,r20 + $ST $C,`2*$SZ`($ctx) + add $F,$F,r21 + $ST $D,`3*$SZ`($ctx) + add $G,$G,r22 + $ST $E,`4*$SZ`($ctx) + add $H,$H,r23 + $ST $F,`5*$SZ`($ctx) + $ST $G,`6*$SZ`($ctx) + $UCMP $inp,$num + $ST $H,`7*$SZ`($ctx) + bne Lsha2_block_private + blr + .long 0 + .byte 0,12,0x14,0,0,0,0,0 +___ + +# Ugly hack here, because PPC assembler syntax seem to vary too +# much from platforms to platform... +$code.=<<___; +.align 6 +LPICmeup: + mflr r0 + bcl 20,31,\$+4 + mflr $Tbl ; vvvvvv "distance" between . and 1st data entry + addi $Tbl,$Tbl,`64-8` + mtlr r0 + blr + .long 0 + .byte 0,12,0x14,0,0,0,0,0 + .space `64-9*4` +___ +$code.=<<___ if ($SZ==8); + .long 0x428a2f98,0xd728ae22,0x71374491,0x23ef65cd + .long 0xb5c0fbcf,0xec4d3b2f,0xe9b5dba5,0x8189dbbc + .long 0x3956c25b,0xf348b538,0x59f111f1,0xb605d019 + .long 0x923f82a4,0xaf194f9b,0xab1c5ed5,0xda6d8118 + .long 0xd807aa98,0xa3030242,0x12835b01,0x45706fbe + .long 0x243185be,0x4ee4b28c,0x550c7dc3,0xd5ffb4e2 + .long 0x72be5d74,0xf27b896f,0x80deb1fe,0x3b1696b1 + .long 0x9bdc06a7,0x25c71235,0xc19bf174,0xcf692694 + .long 0xe49b69c1,0x9ef14ad2,0xefbe4786,0x384f25e3 + .long 0x0fc19dc6,0x8b8cd5b5,0x240ca1cc,0x77ac9c65 + .long 0x2de92c6f,0x592b0275,0x4a7484aa,0x6ea6e483 + .long 0x5cb0a9dc,0xbd41fbd4,0x76f988da,0x831153b5 + .long 0x983e5152,0xee66dfab,0xa831c66d,0x2db43210 + .long 0xb00327c8,0x98fb213f,0xbf597fc7,0xbeef0ee4 + .long 0xc6e00bf3,0x3da88fc2,0xd5a79147,0x930aa725 + .long 0x06ca6351,0xe003826f,0x14292967,0x0a0e6e70 + .long 0x27b70a85,0x46d22ffc,0x2e1b2138,0x5c26c926 + .long 0x4d2c6dfc,0x5ac42aed,0x53380d13,0x9d95b3df + .long 0x650a7354,0x8baf63de,0x766a0abb,0x3c77b2a8 + .long 0x81c2c92e,0x47edaee6,0x92722c85,0x1482353b + .long 0xa2bfe8a1,0x4cf10364,0xa81a664b,0xbc423001 + .long 0xc24b8b70,0xd0f89791,0xc76c51a3,0x0654be30 + .long 0xd192e819,0xd6ef5218,0xd6990624,0x5565a910 + .long 0xf40e3585,0x5771202a,0x106aa070,0x32bbd1b8 + .long 0x19a4c116,0xb8d2d0c8,0x1e376c08,0x5141ab53 + .long 0x2748774c,0xdf8eeb99,0x34b0bcb5,0xe19b48a8 + .long 0x391c0cb3,0xc5c95a63,0x4ed8aa4a,0xe3418acb + .long 0x5b9cca4f,0x7763e373,0x682e6ff3,0xd6b2b8a3 + .long 0x748f82ee,0x5defb2fc,0x78a5636f,0x43172f60 + .long 0x84c87814,0xa1f0ab72,0x8cc70208,0x1a6439ec + .long 0x90befffa,0x23631e28,0xa4506ceb,0xde82bde9 + .long 0xbef9a3f7,0xb2c67915,0xc67178f2,0xe372532b + .long 0xca273ece,0xea26619c,0xd186b8c7,0x21c0c207 + .long 0xeada7dd6,0xcde0eb1e,0xf57d4f7f,0xee6ed178 + .long 0x06f067aa,0x72176fba,0x0a637dc5,0xa2c898a6 + .long 0x113f9804,0xbef90dae,0x1b710b35,0x131c471b + .long 0x28db77f5,0x23047d84,0x32caab7b,0x40c72493 + .long 0x3c9ebe0a,0x15c9bebc,0x431d67c4,0x9c100d4c + .long 0x4cc5d4be,0xcb3e42b6,0x597f299c,0xfc657e2a + .long 0x5fcb6fab,0x3ad6faec,0x6c44198c,0x4a475817 +___ +$code.=<<___ if ($SZ==4); + .long 0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5 + .long 0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5 + .long 0xd807aa98,0x12835b01,0x243185be,0x550c7dc3 + .long 0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174 + .long 0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc + .long 0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da + .long 0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7 + .long 0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967 + .long 0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13 + .long 0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85 + .long 0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3 + .long 0xd192e819,0xd6990624,0xf40e3585,0x106aa070 + .long 0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5 + .long 0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3 + .long 0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208 + .long 0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2 +___ + +$code =~ s/\`([^\`]*)\`/eval $1/gem; +print $code; +close STDOUT; |