768a6a6fad5d0147943d70ee98255f668fd594c0
[openssl.git] / crypto / sha / asm / sha512-ppc.pl
1 #!/usr/bin/env perl
2
3 # ====================================================================
4 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
5 # project. The module is, however, dual licensed under OpenSSL and
6 # CRYPTOGAMS licenses depending on where you obtain it. For further
7 # details see http://www.openssl.org/~appro/cryptogams/.
8 # ====================================================================
9
10 # I let hardware handle unaligned input, except on page boundaries
11 # (see below for details). Otherwise straightforward implementation
12 # with X vector in register bank. The module is big-endian [which is
13 # not big deal as there're no little-endian targets left around].
14
15 #                       sha256          |       sha512
16 #                       -m64    -m32    |       -m64    -m32
17 # --------------------------------------+-----------------------
18 # PPC970,gcc-4.0.0      +50%    +38%    |       +40%    +410%(*)
19 # Power6,xlc-7          +150%   +90%    |       +100%   +430%(*)
20 #
21 # (*)   64-bit code in 32-bit application context, which actually is
22 #       on TODO list. It should be noted that for safe deployment in
23 #       32-bit *mutli-threaded* context asyncronous signals should be
24 #       blocked upon entry to SHA512 block routine. This is because
25 #       32-bit signaling procedure invalidates upper halves of GPRs.
26 #       Context switch procedure preserves them, but not signaling:-(
27
28 # Second version is true multi-thread safe. Trouble with the original
29 # version was that it was using thread local storage pointer register.
30 # Well, it scrupulously preserved it, but the problem would arise the
31 # moment asynchronous signal was delivered and signal handler would
32 # dereference the TLS pointer. While it's never the case in openssl
33 # application or test suite, we have to respect this scenario and not
34 # use TLS pointer register. Alternative would be to require caller to
35 # block signals prior calling this routine. For the record, in 32-bit
36 # context R2 serves as TLS pointer, while in 64-bit context - R13.
37
38 $flavour=shift;
39 $output =shift;
40
41 if ($flavour =~ /64/) {
42         $SIZE_T=8;
43         $STU="stdu";
44         $UCMP="cmpld";
45         $SHL="sldi";
46         $POP="ld";
47         $PUSH="std";
48 } elsif ($flavour =~ /32/) {
49         $SIZE_T=4;
50         $STU="stwu";
51         $UCMP="cmplw";
52         $SHL="slwi";
53         $POP="lwz";
54         $PUSH="stw";
55 } else { die "nonsense $flavour"; }
56
57 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
58 ( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
59 ( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
60 die "can't locate ppc-xlate.pl";
61
62 open STDOUT,"| $^X $xlate $flavour $output" || die "can't call $xlate: $!";
63
64 if ($output =~ /512/) {
65         $func="sha512_block_data_order";
66         $SZ=8;
67         @Sigma0=(28,34,39);
68         @Sigma1=(14,18,41);
69         @sigma0=(1,  8, 7);
70         @sigma1=(19,61, 6);
71         $rounds=80;
72         $LD="ld";
73         $ST="std";
74         $ROR="rotrdi";
75         $SHR="srdi";
76 } else {
77         $func="sha256_block_data_order";
78         $SZ=4;
79         @Sigma0=( 2,13,22);
80         @Sigma1=( 6,11,25);
81         @sigma0=( 7,18, 3);
82         @sigma1=(17,19,10);
83         $rounds=64;
84         $LD="lwz";
85         $ST="stw";
86         $ROR="rotrwi";
87         $SHR="srwi";
88 }
89
90 $FRAME=32*$SIZE_T;
91
92 $sp ="r1";
93 $toc="r2";
94 $ctx="r3";      # zapped by $a0
95 $inp="r4";      # zapped by $a1
96 $num="r5";      # zapped by $t0
97
98 $T  ="r0";
99 $a0 ="r3";
100 $a1 ="r4";
101 $t0 ="r5";
102 $t1 ="r6";
103 $Tbl="r7";
104
105 $A  ="r8";
106 $B  ="r9";
107 $C  ="r10";
108 $D  ="r11";
109 $E  ="r12";
110 $F  ="r13";     $F="r2" if ($SIZE_T==8);# reassigned to exempt TLS pointer
111 $G  ="r14";
112 $H  ="r15";
113
114 @V=($A,$B,$C,$D,$E,$F,$G,$H);
115 @X=("r16","r17","r18","r19","r20","r21","r22","r23",
116     "r24","r25","r26","r27","r28","r29","r30","r31");
117
118 $inp="r31";     # reassigned $inp! aliases with @X[15]
119
120 sub ROUND_00_15 {
121 my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
122 $code.=<<___;
123         $LD     $T,`$i*$SZ`($Tbl)
124         $ROR    $a0,$e,$Sigma1[0]
125         $ROR    $a1,$e,$Sigma1[1]
126         and     $t0,$f,$e
127         andc    $t1,$g,$e
128         add     $T,$T,$h
129         xor     $a0,$a0,$a1
130         $ROR    $a1,$a1,`$Sigma1[2]-$Sigma1[1]`
131         or      $t0,$t0,$t1             ; Ch(e,f,g)
132         add     $T,$T,@X[$i]
133         xor     $a0,$a0,$a1             ; Sigma1(e)
134         add     $T,$T,$t0
135         add     $T,$T,$a0
136
137         $ROR    $a0,$a,$Sigma0[0]
138         $ROR    $a1,$a,$Sigma0[1]
139         and     $t0,$a,$b
140         and     $t1,$a,$c
141         xor     $a0,$a0,$a1
142         $ROR    $a1,$a1,`$Sigma0[2]-$Sigma0[1]`
143         xor     $t0,$t0,$t1
144         and     $t1,$b,$c
145         xor     $a0,$a0,$a1             ; Sigma0(a)
146         add     $d,$d,$T
147         xor     $t0,$t0,$t1             ; Maj(a,b,c)
148         add     $h,$T,$a0
149         add     $h,$h,$t0
150
151 ___
152 }
153
154 sub ROUND_16_xx {
155 my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
156 $i-=16;
157 $code.=<<___;
158         $ROR    $a0,@X[($i+1)%16],$sigma0[0]
159         $ROR    $a1,@X[($i+1)%16],$sigma0[1]
160         $ROR    $t0,@X[($i+14)%16],$sigma1[0]
161         $ROR    $t1,@X[($i+14)%16],$sigma1[1]
162         xor     $a0,$a0,$a1
163         $SHR    $a1,@X[($i+1)%16],$sigma0[2]
164         xor     $t0,$t0,$t1
165         $SHR    $t1,@X[($i+14)%16],$sigma1[2]
166         add     @X[$i],@X[$i],@X[($i+9)%16]
167         xor     $a0,$a0,$a1             ; sigma0(X[(i+1)&0x0f])
168         xor     $t0,$t0,$t1             ; sigma1(X[(i+14)&0x0f])
169         add     @X[$i],@X[$i],$a0
170         add     @X[$i],@X[$i],$t0
171 ___
172 &ROUND_00_15($i,$a,$b,$c,$d,$e,$f,$g,$h);
173 }
174
175 $code=<<___;
176 .machine        "any"
177 .text
178
179 .globl  $func
180 .align  6
181 $func:
182         mflr    r0
183         $STU    $sp,`-($FRAME+16*$SZ)`($sp)
184         $SHL    $num,$num,`log(16*$SZ)/log(2)`
185
186         $PUSH   $ctx,`$FRAME-$SIZE_T*22`($sp)
187
188         $PUSH   r0,`$FRAME-$SIZE_T*21`($sp)
189         $PUSH   $toc,`$FRAME-$SIZE_T*20`($sp)
190         $PUSH   r13,`$FRAME-$SIZE_T*19`($sp)
191         $PUSH   r14,`$FRAME-$SIZE_T*18`($sp)
192         $PUSH   r15,`$FRAME-$SIZE_T*17`($sp)
193         $PUSH   r16,`$FRAME-$SIZE_T*16`($sp)
194         $PUSH   r17,`$FRAME-$SIZE_T*15`($sp)
195         $PUSH   r18,`$FRAME-$SIZE_T*14`($sp)
196         $PUSH   r19,`$FRAME-$SIZE_T*13`($sp)
197         $PUSH   r20,`$FRAME-$SIZE_T*12`($sp)
198         $PUSH   r21,`$FRAME-$SIZE_T*11`($sp)
199         $PUSH   r22,`$FRAME-$SIZE_T*10`($sp)
200         $PUSH   r23,`$FRAME-$SIZE_T*9`($sp)
201         $PUSH   r24,`$FRAME-$SIZE_T*8`($sp)
202         $PUSH   r25,`$FRAME-$SIZE_T*7`($sp)
203         $PUSH   r26,`$FRAME-$SIZE_T*6`($sp)
204         $PUSH   r27,`$FRAME-$SIZE_T*5`($sp)
205         $PUSH   r28,`$FRAME-$SIZE_T*4`($sp)
206         $PUSH   r29,`$FRAME-$SIZE_T*3`($sp)
207         $PUSH   r30,`$FRAME-$SIZE_T*2`($sp)
208         $PUSH   r31,`$FRAME-$SIZE_T*1`($sp)
209
210         $LD     $A,`0*$SZ`($ctx)
211         mr      $inp,r4                         ; incarnate $inp
212         $LD     $B,`1*$SZ`($ctx)
213         $LD     $C,`2*$SZ`($ctx)
214         $LD     $D,`3*$SZ`($ctx)
215         $LD     $E,`4*$SZ`($ctx)
216         $LD     $F,`5*$SZ`($ctx)
217         $LD     $G,`6*$SZ`($ctx)
218         $LD     $H,`7*$SZ`($ctx)
219
220         b       LPICmeup
221 LPICedup:
222         andi.   r0,$inp,3
223         bne     Lunaligned
224 Laligned:
225         add     $num,$inp,$num
226         $PUSH   $num,`$FRAME-$SIZE_T*24`($sp)   ; end pointer
227         $PUSH   $inp,`$FRAME-$SIZE_T*23`($sp)   ; inp pointer
228         bl      Lsha2_block_private
229 Ldone:
230         $POP    r0,`$FRAME-$SIZE_T*21`($sp)
231         $POP    $toc,`$FRAME-$SIZE_T*20`($sp)
232         $POP    r13,`$FRAME-$SIZE_T*19`($sp)
233         $POP    r14,`$FRAME-$SIZE_T*18`($sp)
234         $POP    r15,`$FRAME-$SIZE_T*17`($sp)
235         $POP    r16,`$FRAME-$SIZE_T*16`($sp)
236         $POP    r17,`$FRAME-$SIZE_T*15`($sp)
237         $POP    r18,`$FRAME-$SIZE_T*14`($sp)
238         $POP    r19,`$FRAME-$SIZE_T*13`($sp)
239         $POP    r20,`$FRAME-$SIZE_T*12`($sp)
240         $POP    r21,`$FRAME-$SIZE_T*11`($sp)
241         $POP    r22,`$FRAME-$SIZE_T*10`($sp)
242         $POP    r23,`$FRAME-$SIZE_T*9`($sp)
243         $POP    r24,`$FRAME-$SIZE_T*8`($sp)
244         $POP    r25,`$FRAME-$SIZE_T*7`($sp)
245         $POP    r26,`$FRAME-$SIZE_T*6`($sp)
246         $POP    r27,`$FRAME-$SIZE_T*5`($sp)
247         $POP    r28,`$FRAME-$SIZE_T*4`($sp)
248         $POP    r29,`$FRAME-$SIZE_T*3`($sp)
249         $POP    r30,`$FRAME-$SIZE_T*2`($sp)
250         $POP    r31,`$FRAME-$SIZE_T*1`($sp)
251         mtlr    r0
252         addi    $sp,$sp,`$FRAME+16*$SZ`
253         blr
254 ___
255
256 # PowerPC specification allows an implementation to be ill-behaved
257 # upon unaligned access which crosses page boundary. "Better safe
258 # than sorry" principle makes me treat it specially. But I don't
259 # look for particular offending word, but rather for the input
260 # block which crosses the boundary. Once found that block is aligned
261 # and hashed separately...
262 $code.=<<___;
263 .align  4
264 Lunaligned:
265         subfic  $t1,$inp,4096
266         andi.   $t1,$t1,`4096-16*$SZ`   ; distance to closest page boundary
267         beq     Lcross_page
268         $UCMP   $num,$t1
269         ble-    Laligned                ; didn't cross the page boundary
270         subfc   $num,$t1,$num
271         add     $t1,$inp,$t1
272         $PUSH   $num,`$FRAME-$SIZE_T*25`($sp)   ; save real remaining num
273         $PUSH   $t1,`$FRAME-$SIZE_T*24`($sp)    ; intermediate end pointer
274         $PUSH   $inp,`$FRAME-$SIZE_T*23`($sp)   ; inp pointer
275         bl      Lsha2_block_private
276         ; $inp equals to the intermediate end pointer here
277         $POP    $num,`$FRAME-$SIZE_T*25`($sp)   ; restore real remaining num
278 Lcross_page:
279         li      $t1,`16*$SZ/4`
280         mtctr   $t1
281         addi    r20,$sp,$FRAME                  ; aligned spot below the frame
282 Lmemcpy:
283         lbz     r16,0($inp)
284         lbz     r17,1($inp)
285         lbz     r18,2($inp)
286         lbz     r19,3($inp)
287         addi    $inp,$inp,4
288         stb     r16,0(r20)
289         stb     r17,1(r20)
290         stb     r18,2(r20)
291         stb     r19,3(r20)
292         addi    r20,r20,4
293         bdnz    Lmemcpy
294
295         $PUSH   $inp,`$FRAME-$SIZE_T*26`($sp)   ; save real inp
296         addi    $t1,$sp,`$FRAME+16*$SZ`         ; fictitious end pointer
297         addi    $inp,$sp,$FRAME                 ; fictitious inp pointer
298         $PUSH   $num,`$FRAME-$SIZE_T*25`($sp)   ; save real num
299         $PUSH   $t1,`$FRAME-$SIZE_T*24`($sp)    ; end pointer
300         $PUSH   $inp,`$FRAME-$SIZE_T*23`($sp)   ; inp pointer
301         bl      Lsha2_block_private
302         $POP    $inp,`$FRAME-$SIZE_T*26`($sp)   ; restore real inp
303         $POP    $num,`$FRAME-$SIZE_T*25`($sp)   ; restore real num
304         addic.  $num,$num,`-16*$SZ`             ; num--
305         bne-    Lunaligned
306         b       Ldone
307 ___
308
309 $code.=<<___;
310 .align  4
311 Lsha2_block_private:
312 ___
313 for($i=0;$i<16;$i++) {
314 $code.=<<___ if ($SZ==4);
315         lwz     @X[$i],`$i*$SZ`($inp)
316 ___
317 # 64-bit loads are split to 2x32-bit ones, as CPU can't handle
318 # unaligned 64-bit loads, only 32-bit ones...
319 $code.=<<___ if ($SZ==8);
320         lwz     $t0,`$i*$SZ`($inp)
321         lwz     @X[$i],`$i*$SZ+4`($inp)
322         insrdi  @X[$i],$t0,32,0
323 ___
324         &ROUND_00_15($i,@V);
325         unshift(@V,pop(@V));
326 }
327 $code.=<<___;
328         li      $T,`$rounds/16-1`
329         mtctr   $T
330 .align  4
331 Lrounds:
332         addi    $Tbl,$Tbl,`16*$SZ`
333 ___
334 for(;$i<32;$i++) {
335         &ROUND_16_xx($i,@V);
336         unshift(@V,pop(@V));
337 }
338 $code.=<<___;
339         bdnz-   Lrounds
340
341         $POP    $ctx,`$FRAME-$SIZE_T*22`($sp)
342         $POP    $inp,`$FRAME-$SIZE_T*23`($sp)   ; inp pointer
343         $POP    $num,`$FRAME-$SIZE_T*24`($sp)   ; end pointer
344         subi    $Tbl,$Tbl,`($rounds-16)*$SZ`    ; rewind Tbl
345
346         $LD     r16,`0*$SZ`($ctx)
347         $LD     r17,`1*$SZ`($ctx)
348         $LD     r18,`2*$SZ`($ctx)
349         $LD     r19,`3*$SZ`($ctx)
350         $LD     r20,`4*$SZ`($ctx)
351         $LD     r21,`5*$SZ`($ctx)
352         $LD     r22,`6*$SZ`($ctx)
353         addi    $inp,$inp,`16*$SZ`              ; advance inp
354         $LD     r23,`7*$SZ`($ctx)
355         add     $A,$A,r16
356         add     $B,$B,r17
357         $PUSH   $inp,`$FRAME-$SIZE_T*23`($sp)
358         add     $C,$C,r18
359         $ST     $A,`0*$SZ`($ctx)
360         add     $D,$D,r19
361         $ST     $B,`1*$SZ`($ctx)
362         add     $E,$E,r20
363         $ST     $C,`2*$SZ`($ctx)
364         add     $F,$F,r21
365         $ST     $D,`3*$SZ`($ctx)
366         add     $G,$G,r22
367         $ST     $E,`4*$SZ`($ctx)
368         add     $H,$H,r23
369         $ST     $F,`5*$SZ`($ctx)
370         $ST     $G,`6*$SZ`($ctx)
371         $UCMP   $inp,$num
372         $ST     $H,`7*$SZ`($ctx)
373         bne     Lsha2_block_private
374         blr
375 ___
376
377 # Ugly hack here, because PPC assembler syntax seem to vary too
378 # much from platforms to platform...
379 $code.=<<___;
380 .align  6
381 LPICmeup:
382         bl      LPIC
383         addi    $Tbl,$Tbl,`64-4`        ; "distance" between . and last nop
384         b       LPICedup
385         nop
386         nop
387         nop
388         nop
389         nop
390 LPIC:   mflr    $Tbl
391         blr
392         nop
393         nop
394         nop
395         nop
396         nop
397         nop
398 ___
399 $code.=<<___ if ($SZ==8);
400         .long   0x428a2f98,0xd728ae22,0x71374491,0x23ef65cd
401         .long   0xb5c0fbcf,0xec4d3b2f,0xe9b5dba5,0x8189dbbc
402         .long   0x3956c25b,0xf348b538,0x59f111f1,0xb605d019
403         .long   0x923f82a4,0xaf194f9b,0xab1c5ed5,0xda6d8118
404         .long   0xd807aa98,0xa3030242,0x12835b01,0x45706fbe
405         .long   0x243185be,0x4ee4b28c,0x550c7dc3,0xd5ffb4e2
406         .long   0x72be5d74,0xf27b896f,0x80deb1fe,0x3b1696b1
407         .long   0x9bdc06a7,0x25c71235,0xc19bf174,0xcf692694
408         .long   0xe49b69c1,0x9ef14ad2,0xefbe4786,0x384f25e3
409         .long   0x0fc19dc6,0x8b8cd5b5,0x240ca1cc,0x77ac9c65
410         .long   0x2de92c6f,0x592b0275,0x4a7484aa,0x6ea6e483
411         .long   0x5cb0a9dc,0xbd41fbd4,0x76f988da,0x831153b5
412         .long   0x983e5152,0xee66dfab,0xa831c66d,0x2db43210
413         .long   0xb00327c8,0x98fb213f,0xbf597fc7,0xbeef0ee4
414         .long   0xc6e00bf3,0x3da88fc2,0xd5a79147,0x930aa725
415         .long   0x06ca6351,0xe003826f,0x14292967,0x0a0e6e70
416         .long   0x27b70a85,0x46d22ffc,0x2e1b2138,0x5c26c926
417         .long   0x4d2c6dfc,0x5ac42aed,0x53380d13,0x9d95b3df
418         .long   0x650a7354,0x8baf63de,0x766a0abb,0x3c77b2a8
419         .long   0x81c2c92e,0x47edaee6,0x92722c85,0x1482353b
420         .long   0xa2bfe8a1,0x4cf10364,0xa81a664b,0xbc423001
421         .long   0xc24b8b70,0xd0f89791,0xc76c51a3,0x0654be30
422         .long   0xd192e819,0xd6ef5218,0xd6990624,0x5565a910
423         .long   0xf40e3585,0x5771202a,0x106aa070,0x32bbd1b8
424         .long   0x19a4c116,0xb8d2d0c8,0x1e376c08,0x5141ab53
425         .long   0x2748774c,0xdf8eeb99,0x34b0bcb5,0xe19b48a8
426         .long   0x391c0cb3,0xc5c95a63,0x4ed8aa4a,0xe3418acb
427         .long   0x5b9cca4f,0x7763e373,0x682e6ff3,0xd6b2b8a3
428         .long   0x748f82ee,0x5defb2fc,0x78a5636f,0x43172f60
429         .long   0x84c87814,0xa1f0ab72,0x8cc70208,0x1a6439ec
430         .long   0x90befffa,0x23631e28,0xa4506ceb,0xde82bde9
431         .long   0xbef9a3f7,0xb2c67915,0xc67178f2,0xe372532b
432         .long   0xca273ece,0xea26619c,0xd186b8c7,0x21c0c207
433         .long   0xeada7dd6,0xcde0eb1e,0xf57d4f7f,0xee6ed178
434         .long   0x06f067aa,0x72176fba,0x0a637dc5,0xa2c898a6
435         .long   0x113f9804,0xbef90dae,0x1b710b35,0x131c471b
436         .long   0x28db77f5,0x23047d84,0x32caab7b,0x40c72493
437         .long   0x3c9ebe0a,0x15c9bebc,0x431d67c4,0x9c100d4c
438         .long   0x4cc5d4be,0xcb3e42b6,0x597f299c,0xfc657e2a
439         .long   0x5fcb6fab,0x3ad6faec,0x6c44198c,0x4a475817
440 ___
441 $code.=<<___ if ($SZ==4);
442         .long   0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
443         .long   0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
444         .long   0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
445         .long   0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
446         .long   0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
447         .long   0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
448         .long   0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
449         .long   0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
450         .long   0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
451         .long   0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
452         .long   0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
453         .long   0xd192e819,0xd6990624,0xf40e3585,0x106aa070
454         .long   0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
455         .long   0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
456         .long   0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
457         .long   0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
458 ___
459
460 $code =~ s/\`([^\`]*)\`/eval $1/gem;
461 print $code;
462 close STDOUT;