summaryrefslogtreecommitdiff
path: root/src/lib/libcrypto/perlasm
diff options
context:
space:
mode:
Diffstat (limited to 'src/lib/libcrypto/perlasm')
-rwxr-xr-xsrc/lib/libcrypto/perlasm/x86_64-xlate.pl554
1 files changed, 554 insertions, 0 deletions
diff --git a/src/lib/libcrypto/perlasm/x86_64-xlate.pl b/src/lib/libcrypto/perlasm/x86_64-xlate.pl
new file mode 100755
index 0000000000..a4af769b4a
--- /dev/null
+++ b/src/lib/libcrypto/perlasm/x86_64-xlate.pl
@@ -0,0 +1,554 @@
1#!/usr/bin/env perl
2
3# Ascetic x86_64 AT&T to MASM assembler translator by <appro>.
4#
5# Why AT&T to MASM and not vice versa? Several reasons. Because AT&T
6# format is way easier to parse. Because it's simpler to "gear" from
7# Unix ABI to Windows one [see cross-reference "card" at the end of
8# file]. Because Linux targets were available first...
9#
10# In addition the script also "distills" code suitable for GNU
11# assembler, so that it can be compiled with more rigid assemblers,
12# such as Solaris /usr/ccs/bin/as.
13#
14# This translator is not designed to convert *arbitrary* assembler
15# code from AT&T format to MASM one. It's designed to convert just
16# enough to provide for dual-ABI OpenSSL modules development...
17# There *are* limitations and you might have to modify your assembler
18# code or this script to achieve the desired result...
19#
20# Currently recognized limitations:
21#
22# - can't use multiple ops per line;
23# - indirect calls and jumps are not supported;
24#
25# Dual-ABI styling rules.
26#
27# 1. Adhere to Unix register and stack layout [see the end for
28# explanation].
29# 2. Forget about "red zone," stick to more traditional blended
30# stack frame allocation. If volatile storage is actually required
31# that is. If not, just leave the stack as is.
32# 3. Functions tagged with ".type name,@function" get crafted with
33# unified Win64 prologue and epilogue automatically. If you want
34# to take care of ABI differences yourself, tag functions as
35# ".type name,@abi-omnipotent" instead.
36# 4. To optimize the Win64 prologue you can specify number of input
37# arguments as ".type name,@function,N." Keep in mind that if N is
38# larger than 6, then you *have to* write "abi-omnipotent" code,
39# because >6 cases can't be addressed with unified prologue.
40# 5. Name local labels as .L*, do *not* use dynamic labels such as 1:
41# (sorry about latter).
42# 6. Don't use [or hand-code with .byte] "rep ret." "ret" mnemonic is
43# required to identify the spots, where to inject Win64 epilogue!
44# But on the pros, it's then prefixed with rep automatically:-)
45# 7. Due to MASM limitations [and certain general counter-intuitivity
46# of ip-relative addressing] generation of position-independent
47# code is assisted by synthetic directive, .picmeup, which puts
48# address of the *next* instruction into target register.
49#
50# Example 1:
51# .picmeup %rax
52# lea .Label-.(%rax),%rax
53# Example 2:
54# .picmeup %rcx
55# .Lpic_point:
56# ...
57# lea .Label-.Lpic_point(%rcx),%rbp
58
59my $output = shift;
60
61{ my ($stddev,$stdino,@junk)=stat(STDOUT);
62 my ($outdev,$outino,@junk)=stat($output);
63
64 open STDOUT,">$output" || die "can't open $output: $!"
65 if ($stddev!=$outdev || $stdino!=$outino);
66}
67
68my $masmref=8 + 50727*2**-32; # 8.00.50727 shipped with VS2005
69my $masm=$masmref if ($output =~ /\.asm/);
70if ($masm && `ml64 2>&1` =~ m/Version ([0-9]+)\.([0-9]+)(\.([0-9]+))?/)
71{ $masm=$1 + $2*2**-16 + $4*2**-32; }
72
73my $current_segment;
74my $current_function;
75
76{ package opcode; # pick up opcodes
77 sub re {
78 my $self = shift; # single instance in enough...
79 local *line = shift;
80 undef $ret;
81
82 if ($line =~ /^([a-z][a-z0-9]*)/i) {
83 $self->{op} = $1;
84 $ret = $self;
85 $line = substr($line,@+[0]); $line =~ s/^\s+//;
86
87 undef $self->{sz};
88 if ($self->{op} =~ /^(movz)b.*/) { # movz is pain...
89 $self->{op} = $1;
90 $self->{sz} = "b";
91 } elsif ($self->{op} =~ /call/) {
92 $self->{sz} = ""
93 } elsif ($self->{op} =~ /([a-z]{3,})([qlwb])$/) {
94 $self->{op} = $1;
95 $self->{sz} = $2;
96 }
97 }
98 $ret;
99 }
100 sub size {
101 my $self = shift;
102 my $sz = shift;
103 $self->{sz} = $sz if (defined($sz) && !defined($self->{sz}));
104 $self->{sz};
105 }
106 sub out {
107 my $self = shift;
108 if (!$masm) {
109 if ($self->{op} eq "movz") { # movz is pain...
110 sprintf "%s%s%s",$self->{op},$self->{sz},shift;
111 } elsif ($self->{op} =~ /^set/) {
112 "$self->{op}";
113 } elsif ($self->{op} eq "ret") {
114 ".byte 0xf3,0xc3";
115 } else {
116 "$self->{op}$self->{sz}";
117 }
118 } else {
119 $self->{op} =~ s/^movz/movzx/;
120 if ($self->{op} eq "ret") {
121 $self->{op} = "";
122 if ($current_function->{abi} eq "svr4") {
123 $self->{op} = "mov rdi,QWORD PTR 8[rsp]\t;WIN64 epilogue\n\t".
124 "mov rsi,QWORD PTR 16[rsp]\n\t";
125 }
126 $self->{op} .= "DB\t0F3h,0C3h\t\t;repret";
127 }
128 $self->{op};
129 }
130 }
131}
132{ package const; # pick up constants, which start with $
133 sub re {
134 my $self = shift; # single instance in enough...
135 local *line = shift;
136 undef $ret;
137
138 if ($line =~ /^\$([^,]+)/) {
139 $self->{value} = $1;
140 $ret = $self;
141 $line = substr($line,@+[0]); $line =~ s/^\s+//;
142 }
143 $ret;
144 }
145 sub out {
146 my $self = shift;
147
148 if (!$masm) {
149 # Solaris /usr/ccs/bin/as can't handle multiplications
150 # in $self->{value}
151 $self->{value} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/egi;
152 $self->{value} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
153 sprintf "\$%s",$self->{value};
154 } else {
155 $self->{value} =~ s/0x([0-9a-f]+)/0$1h/ig;
156 sprintf "%s",$self->{value};
157 }
158 }
159}
160{ package ea; # pick up effective addresses: expr(%reg,%reg,scale)
161 sub re {
162 my $self = shift; # single instance in enough...
163 local *line = shift;
164 undef $ret;
165
166 if ($line =~ /^([^\(,]*)\(([%\w,]+)\)/) {
167 $self->{label} = $1;
168 ($self->{base},$self->{index},$self->{scale})=split(/,/,$2);
169 $self->{scale} = 1 if (!defined($self->{scale}));
170 $ret = $self;
171 $line = substr($line,@+[0]); $line =~ s/^\s+//;
172
173 $self->{base} =~ s/^%//;
174 $self->{index} =~ s/^%// if (defined($self->{index}));
175 }
176 $ret;
177 }
178 sub size {}
179 sub out {
180 my $self = shift;
181 my $sz = shift;
182
183 # Silently convert all EAs to 64-bit. This is required for
184 # elder GNU assembler and results in more compact code,
185 # *but* most importantly AES module depends on this feature!
186 $self->{index} =~ s/^[er](.?[0-9xpi])[d]?$/r\1/;
187 $self->{base} =~ s/^[er](.?[0-9xpi])[d]?$/r\1/;
188
189 if (!$masm) {
190 # Solaris /usr/ccs/bin/as can't handle multiplications
191 # in $self->{label}
192 $self->{label} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/egi;
193 $self->{label} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
194
195 if (defined($self->{index})) {
196 sprintf "%s(%%%s,%%%s,%d)",
197 $self->{label},$self->{base},
198 $self->{index},$self->{scale};
199 } else {
200 sprintf "%s(%%%s)", $self->{label},$self->{base};
201 }
202 } else {
203 %szmap = ( b=>"BYTE", w=>"WORD", l=>"DWORD", q=>"QWORD" );
204
205 $self->{label} =~ s/\./\$/g;
206 $self->{label} =~ s/0x([0-9a-f]+)/0$1h/ig;
207 $self->{label} = "($self->{label})" if ($self->{label} =~ /[\*\+\-\/]/);
208
209 if (defined($self->{index})) {
210 sprintf "%s PTR %s[%s*%d+%s]",$szmap{$sz},
211 $self->{label},
212 $self->{index},$self->{scale},
213 $self->{base};
214 } elsif ($self->{base} eq "rip") {
215 sprintf "%s PTR %s",$szmap{$sz},$self->{label};
216 } else {
217 sprintf "%s PTR %s[%s]",$szmap{$sz},
218 $self->{label},$self->{base};
219 }
220 }
221 }
222}
223{ package register; # pick up registers, which start with %.
224 sub re {
225 my $class = shift; # muliple instances...
226 my $self = {};
227 local *line = shift;
228 undef $ret;
229
230 if ($line =~ /^%(\w+)/) {
231 bless $self,$class;
232 $self->{value} = $1;
233 $ret = $self;
234 $line = substr($line,@+[0]); $line =~ s/^\s+//;
235 }
236 $ret;
237 }
238 sub size {
239 my $self = shift;
240 undef $ret;
241
242 if ($self->{value} =~ /^r[\d]+b$/i) { $ret="b"; }
243 elsif ($self->{value} =~ /^r[\d]+w$/i) { $ret="w"; }
244 elsif ($self->{value} =~ /^r[\d]+d$/i) { $ret="l"; }
245 elsif ($self->{value} =~ /^r[\w]+$/i) { $ret="q"; }
246 elsif ($self->{value} =~ /^[a-d][hl]$/i){ $ret="b"; }
247 elsif ($self->{value} =~ /^[\w]{2}l$/i) { $ret="b"; }
248 elsif ($self->{value} =~ /^[\w]{2}$/i) { $ret="w"; }
249 elsif ($self->{value} =~ /^e[a-z]{2}$/i){ $ret="l"; }
250
251 $ret;
252 }
253 sub out {
254 my $self = shift;
255 sprintf $masm?"%s":"%%%s",$self->{value};
256 }
257}
258{ package label; # pick up labels, which end with :
259 sub re {
260 my $self = shift; # single instance is enough...
261 local *line = shift;
262 undef $ret;
263
264 if ($line =~ /(^[\.\w]+\:)/) {
265 $self->{value} = $1;
266 $ret = $self;
267 $line = substr($line,@+[0]); $line =~ s/^\s+//;
268
269 $self->{value} =~ s/\.L/\$L/ if ($masm);
270 }
271 $ret;
272 }
273 sub out {
274 my $self = shift;
275
276 if (!$masm) {
277 $self->{value};
278 } elsif ($self->{value} ne "$current_function->{name}:") {
279 $self->{value};
280 } elsif ($current_function->{abi} eq "svr4") {
281 my $func = "$current_function->{name} PROC\n".
282 " mov QWORD PTR 8[rsp],rdi\t;WIN64 prologue\n".
283 " mov QWORD PTR 16[rsp],rsi\n";
284 my $narg = $current_function->{narg};
285 $narg=6 if (!defined($narg));
286 $func .= " mov rdi,rcx\n" if ($narg>0);
287 $func .= " mov rsi,rdx\n" if ($narg>1);
288 $func .= " mov rdx,r8\n" if ($narg>2);
289 $func .= " mov rcx,r9\n" if ($narg>3);
290 $func .= " mov r8,QWORD PTR 40[rsp]\n" if ($narg>4);
291 $func .= " mov r9,QWORD PTR 48[rsp]\n" if ($narg>5);
292 $func .= "\n";
293 } else {
294 "$current_function->{name} PROC";
295 }
296 }
297}
298{ package expr; # pick up expressioins
299 sub re {
300 my $self = shift; # single instance is enough...
301 local *line = shift;
302 undef $ret;
303
304 if ($line =~ /(^[^,]+)/) {
305 $self->{value} = $1;
306 $ret = $self;
307 $line = substr($line,@+[0]); $line =~ s/^\s+//;
308
309 $self->{value} =~ s/\.L/\$L/g if ($masm);
310 }
311 $ret;
312 }
313 sub out {
314 my $self = shift;
315 $self->{value};
316 }
317}
318{ package directive; # pick up directives, which start with .
319 sub re {
320 my $self = shift; # single instance is enough...
321 local *line = shift;
322 undef $ret;
323 my $dir;
324 my %opcode = # lea 2f-1f(%rip),%dst; 1: nop; 2:
325 ( "%rax"=>0x01058d48, "%rcx"=>0x010d8d48,
326 "%rdx"=>0x01158d48, "%rbx"=>0x011d8d48,
327 "%rsp"=>0x01258d48, "%rbp"=>0x012d8d48,
328 "%rsi"=>0x01358d48, "%rdi"=>0x013d8d48,
329 "%r8" =>0x01058d4c, "%r9" =>0x010d8d4c,
330 "%r10"=>0x01158d4c, "%r11"=>0x011d8d4c,
331 "%r12"=>0x01258d4c, "%r13"=>0x012d8d4c,
332 "%r14"=>0x01358d4c, "%r15"=>0x013d8d4c );
333
334 if ($line =~ /^\s*(\.\w+)/) {
335 if (!$masm) {
336 $self->{value} = $1;
337 $line =~ s/\@abi\-omnipotent/\@function/;
338 $line =~ s/\@function.*/\@function/;
339 if ($line =~ /\.picmeup\s+(%r[\w]+)/i) {
340 $self->{value} = sprintf "\t.long\t0x%x,0x90000000",$opcode{$1};
341 } elsif ($line =~ /\.asciz\s+"(.*)"$/) {
342 $self->{value} = ".byte\t".join(",",unpack("C*",$1),0);
343 } elsif ($line =~ /\.extern/) {
344 $self->{value} = ""; # swallow extern
345 } else {
346 $self->{value} = $line;
347 }
348 $line = "";
349 return $self;
350 }
351
352 $dir = $1;
353 $ret = $self;
354 undef $self->{value};
355 $line = substr($line,@+[0]); $line =~ s/^\s+//;
356 SWITCH: for ($dir) {
357 /\.(text)/
358 && do { my $v=undef;
359 $v="$current_segment\tENDS\n" if ($current_segment);
360 $current_segment = "_$1\$";
361 $current_segment =~ tr/[a-z]/[A-Z]/;
362 $v.="$current_segment\tSEGMENT ";
363 $v.=$masm>=$masmref ? "ALIGN(64)" : "PAGE";
364 $v.=" 'CODE'";
365 $self->{value} = $v;
366 last;
367 };
368 /\.extern/ && do { $self->{value} = "EXTRN\t".$line.":BYTE"; last; };
369 /\.globl/ && do { $self->{value} = "PUBLIC\t".$line; last; };
370 /\.type/ && do { ($sym,$type,$narg) = split(',',$line);
371 if ($type eq "\@function") {
372 undef $current_function;
373 $current_function->{name} = $sym;
374 $current_function->{abi} = "svr4";
375 $current_function->{narg} = $narg;
376 } elsif ($type eq "\@abi-omnipotent") {
377 undef $current_function;
378 $current_function->{name} = $sym;
379 }
380 last;
381 };
382 /\.size/ && do { if (defined($current_function)) {
383 $self->{value}="$current_function->{name}\tENDP";
384 undef $current_function;
385 }
386 last;
387 };
388 /\.align/ && do { $self->{value} = "ALIGN\t".$line; last; };
389 /\.(byte|value|long|quad)/
390 && do { my @arr = split(',',$line);
391 my $sz = substr($1,0,1);
392 my $last = pop(@arr);
393 my $conv = sub { my $var=shift;
394 if ($var=~s/0x([0-9a-f]+)/0$1h/i) { $var; }
395 else { sprintf"0%Xh",$var; }
396 };
397
398 $sz =~ tr/bvlq/BWDQ/;
399 $self->{value} = "\tD$sz\t";
400 for (@arr) { $self->{value} .= &$conv($_).","; }
401 $self->{value} .= &$conv($last);
402 last;
403 };
404 /\.picmeup/ && do { $self->{value} = sprintf"\tDD\t 0%Xh,090000000h",$opcode{$line};
405 last;
406 };
407 /\.asciz/ && do { if ($line =~ /^"(.*)"$/) {
408 my @str=unpack("C*",$1);
409 push @str,0;
410 while ($#str>15) {
411 $self->{value}.="DB\t"
412 .join(",",@str[0..15])."\n";
413 foreach (0..15) { shift @str; }
414 }
415 $self->{value}.="DB\t"
416 .join(",",@str) if (@str);
417 }
418 last;
419 };
420 }
421 $line = "";
422 }
423
424 $ret;
425 }
426 sub out {
427 my $self = shift;
428 $self->{value};
429 }
430}
431
432while($line=<>) {
433
434 chomp($line);
435
436 $line =~ s|[#!].*$||; # get rid of asm-style comments...
437 $line =~ s|/\*.*\*/||; # ... and C-style comments...
438 $line =~ s|^\s+||; # ... and skip white spaces in beginning
439
440 undef $label;
441 undef $opcode;
442 undef $dst;
443 undef $src;
444 undef $sz;
445
446 if ($label=label->re(\$line)) { print $label->out(); }
447
448 if (directive->re(\$line)) {
449 printf "%s",directive->out();
450 } elsif ($opcode=opcode->re(\$line)) { ARGUMENT: {
451
452 if ($src=register->re(\$line)) { opcode->size($src->size()); }
453 elsif ($src=const->re(\$line)) { }
454 elsif ($src=ea->re(\$line)) { }
455 elsif ($src=expr->re(\$line)) { }
456
457 last ARGUMENT if ($line !~ /^,/);
458
459 $line = substr($line,1); $line =~ s/^\s+//;
460
461 if ($dst=register->re(\$line)) { opcode->size($dst->size()); }
462 elsif ($dst=const->re(\$line)) { }
463 elsif ($dst=ea->re(\$line)) { }
464
465 } # ARGUMENT:
466
467 $sz=opcode->size();
468
469 if (defined($dst)) {
470 if (!$masm) {
471 printf "\t%s\t%s,%s", $opcode->out($dst->size()),
472 $src->out($sz),$dst->out($sz);
473 } else {
474 printf "\t%s\t%s,%s", $opcode->out(),
475 $dst->out($sz),$src->out($sz);
476 }
477 } elsif (defined($src)) {
478 printf "\t%s\t%s",$opcode->out(),$src->out($sz);
479 } else {
480 printf "\t%s",$opcode->out();
481 }
482 }
483
484 print $line,"\n";
485}
486
487print "\n$current_segment\tENDS\nEND\n" if ($masm);
488
489close STDOUT;
490
491#################################################
492# Cross-reference x86_64 ABI "card"
493#
494# Unix Win64
495# %rax * *
496# %rbx - -
497# %rcx #4 #1
498# %rdx #3 #2
499# %rsi #2 -
500# %rdi #1 -
501# %rbp - -
502# %rsp - -
503# %r8 #5 #3
504# %r9 #6 #4
505# %r10 * *
506# %r11 * *
507# %r12 - -
508# %r13 - -
509# %r14 - -
510# %r15 - -
511#
512# (*) volatile register
513# (-) preserved by callee
514# (#) Nth argument, volatile
515#
516# In Unix terms top of stack is argument transfer area for arguments
517# which could not be accomodated in registers. Or in other words 7th
518# [integer] argument resides at 8(%rsp) upon function entry point.
519# 128 bytes above %rsp constitute a "red zone" which is not touched
520# by signal handlers and can be used as temporal storage without
521# allocating a frame.
522#
523# In Win64 terms N*8 bytes on top of stack is argument transfer area,
524# which belongs to/can be overwritten by callee. N is the number of
525# arguments passed to callee, *but* not less than 4! This means that
526# upon function entry point 5th argument resides at 40(%rsp), as well
527# as that 32 bytes from 8(%rsp) can always be used as temporal
528# storage [without allocating a frame]. One can actually argue that
529# one can assume a "red zone" above stack pointer under Win64 as well.
530# Point is that at apparently no occasion Windows kernel would alter
531# the area above user stack pointer in true asynchronous manner...
532#
533# All the above means that if assembler programmer adheres to Unix
534# register and stack layout, but disregards the "red zone" existense,
535# it's possible to use following prologue and epilogue to "gear" from
536# Unix to Win64 ABI in leaf functions with not more than 6 arguments.
537#
538# omnipotent_function:
539# ifdef WIN64
540# movq %rdi,8(%rsp)
541# movq %rsi,16(%rsp)
542# movq %rcx,%rdi ; if 1st argument is actually present
543# movq %rdx,%rsi ; if 2nd argument is actually ...
544# movq %r8,%rdx ; if 3rd argument is ...
545# movq %r9,%rcx ; if 4th argument ...
546# movq 40(%rsp),%r8 ; if 5th ...
547# movq 48(%rsp),%r9 ; if 6th ...
548# endif
549# ...
550# ifdef WIN64
551# movq 8(%rsp),%rdi
552# movq 16(%rsp),%rsi
553# endif
554# ret
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-04-27 13:50:47 +0000