1#!/usr/bin/perl
2# Copyright (C) 2002-2013 Xiph.org Foundation
3#
4# Redistribution and use in source and binary forms, with or without
5# modification, are permitted provided that the following conditions
6# are met:
7#
8# - Redistributions of source code must retain the above copyright
9# notice, this list of conditions and the following disclaimer.
10#
11# - Redistributions in binary form must reproduce the above copyright
12# notice, this list of conditions and the following disclaimer in the
13# documentation and/or other materials provided with the distribution.
14#
15# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
16# ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
17# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
18# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
19# OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
20# EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
21# PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
22# PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
23# LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
24# NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
25# SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26
27my $bigend;  # little/big endian
28my $nxstack;
29my $apple = 0;
30my $symprefix = "";
31
32$nxstack = 0;
33
34eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
35    if $running_under_some_shell;
36
37while ($ARGV[0] =~ /^-/) {
38    $_ = shift;
39  last if /^--$/;
40    if (/^-n$/) {
41    $nflag++;
42    next;
43    }
44    if (/^--apple$/) {
45        $apple = 1;
46        $symprefix = "_";
47        next;
48    }
49    die "I don't recognize this switch: $_\\n";
50}
51$printit++ unless $nflag;
52
53$\ = "\n";      # automatically add newline on print
54$n=0;
55
56$thumb = 0;     # ARM mode by default, not Thumb.
57@proc_stack = ();
58
59printf ("    .syntax unified\n");
60
61LINE:
62while (<>) {
63
64    # For ADRLs we need to add a new line after the substituted one.
65    $addPadding = 0;
66
67    # First, we do not dare to touch *anything* inside double quotes, do we?
68    # Second, if you want a dollar character in the string,
69    # insert two of them -- that's how ARM C and assembler treat strings.
70    s/^([A-Za-z_]\w*)[ \t]+DCB[ \t]*\"/$1:   .ascii \"/   && do { s/\$\$/\$/g; next };
71    s/\bDCB\b[ \t]*\"/.ascii \"/                          && do { s/\$\$/\$/g; next };
72    s/^(\S+)\s+RN\s+(\S+)/$1 .req r$2/                    && do { s/\$\$/\$/g; next };
73    # If there's nothing on a line but a comment, don't try to apply any further
74    #  substitutions (this is a cheap hack to avoid mucking up the license header)
75    s/^([ \t]*);/$1@/                                     && do { s/\$\$/\$/g; next };
76    # If substituted -- leave immediately !
77
78    s/@/,:/;
79    s/;/@/;
80    while ( /@.*'/ ) {
81      s/(@.*)'/$1/g;
82    }
83    s/\{FALSE\}/0/g;
84    s/\{TRUE\}/1/g;
85    s/\{(\w\w\w\w+)\}/$1/g;
86    s/\bINCLUDE[ \t]*([^ \t\n]+)/.include \"$1\"/;
87    s/\bGET[ \t]*([^ \t\n]+)/.include \"${ my $x=$1; $x =~ s|\.s|-gnu.S|; \$x }\"/;
88    s/\bIMPORT\b/.extern/;
89    s/\bEXPORT\b\s*/.global $symprefix/;
90    s/^(\s+)\[/$1IF/;
91    s/^(\s+)\|/$1ELSE/;
92    s/^(\s+)\]/$1ENDIF/;
93    s/IF *:DEF:/ .ifdef/;
94    s/IF *:LNOT: *:DEF:/ .ifndef/;
95    s/ELSE/ .else/;
96    s/ENDIF/ .endif/;
97
98    if( /\bIF\b/ ) {
99      s/\bIF\b/ .if/;
100      s/=/==/;
101    }
102    if ( $n == 2) {
103        s/\$/\\/g;
104    }
105    if ($n == 1) {
106        s/\$//g;
107        s/label//g;
108    $n = 2;
109      }
110    if ( /MACRO/ ) {
111      s/MACRO *\n/.macro/;
112      $n=1;
113    }
114    if ( /\bMEND\b/ ) {
115      s/\bMEND\b/.endm/;
116      $n=0;
117    }
118
119    # ".rdata" doesn't work in 'as' version 2.13.2, as it is ".rodata" there.
120    #
121    if ( /\bAREA\b/ ) {
122        my $align;
123        $align = "2";
124        if ( /ALIGN=(\d+)/ ) {
125            $align = $1;
126        }
127        if ( /CODE/ ) {
128            $nxstack = 1;
129        }
130        s/^(.+)CODE(.+)READONLY(.*)/    .text/;
131        s/^(.+)DATA(.+)READONLY(.*)/    .section .rdata/;
132        s/^(.+)\|\|\.data\|\|(.+)/    .data/;
133        s/^(.+)\|\|\.bss\|\|(.+)/    .bss/;
134        s/$/;   .p2align $align/;
135        # Enable NEON instructions but don't produce a binary that requires
136        # ARMv7. RVCT does not have equivalent directives, so we just do this
137        # for all CODE areas.
138        if ( /.text/ ) {
139            # Separating .arch, .fpu, etc., by semicolons does not work (gas
140            # thinks the semicolon is part of the arch name, even when there's
141            # whitespace separating them). Sadly this means our line numbers
142            # won't match the original source file (we could use the .line
143            # directive, which is documented to be obsolete, but then gdb will
144            # show the wrong line in the translated source file).
145            s/$/;   .arch armv7-a\n   .fpu neon\n   .object_arch armv4t/ unless ($apple);
146        }
147    }
148
149    s/\|\|\.constdata\$(\d+)\|\|/.L_CONST$1/;       # ||.constdata$3||
150    s/\|\|\.bss\$(\d+)\|\|/.L_BSS$1/;               # ||.bss$2||
151    s/\|\|\.data\$(\d+)\|\|/.L_DATA$1/;             # ||.data$2||
152    s/\|\|([a-zA-Z0-9_]+)\@([a-zA-Z0-9_]+)\|\|/@ $&/;
153    s/^(\s+)\%(\s)/    .space $1/;
154
155    s/\|(.+)\.(\d+)\|/\.$1_$2/;                     # |L80.123| -> .L80_123
156    s/\bCODE32\b/.code 32/ && do {$thumb = 0};
157    s/\bCODE16\b/.code 16/ && do {$thumb = 1};
158    if (/\bPROC\b/)
159    {
160        my $prefix;
161        my $proc;
162        /^([A-Za-z_\.]\w+)\b/;
163        $proc = $1;
164        $prefix = "";
165        if ($proc)
166        {
167            $prefix = $prefix.sprintf("\t.type\t%s, %%function", $proc) unless ($apple);
168            # Make sure we $prefix isn't empty here (for the $apple case).
169            # We handle mangling the label here, make sure it doesn't match
170            # the label handling below (if $prefix would be empty).
171            $prefix = $prefix."; ";
172            push(@proc_stack, $proc);
173            s/^[A-Za-z_\.]\w+/$symprefix$&:/;
174        }
175        $prefix = $prefix."\t.thumb_func; " if ($thumb);
176        s/\bPROC\b/@ $&/;
177        $_ = $prefix.$_;
178    }
179    s/^(\s*)(S|Q|SH|U|UQ|UH)ASX\b/$1$2ADDSUBX/;
180    s/^(\s*)(S|Q|SH|U|UQ|UH)SAX\b/$1$2SUBADDX/;
181    if (/\bENDP\b/)
182    {
183        my $proc;
184        s/\bENDP\b/@ $&/;
185        $proc = pop(@proc_stack);
186        $_ = "\t.size $proc, .-$proc".$_ if ($proc && !$apple);
187    }
188    s/\bSUBT\b/@ $&/;
189    s/\bDATA\b/@ $&/;   # DATA directive is deprecated -- Asm guide, p.7-25
190    s/\bKEEP\b/@ $&/;
191    s/\bEXPORTAS\b/@ $&/;
192    s/\|\|(.)+\bEQU\b/@ $&/;
193    s/\|\|([\w\$]+)\|\|/$1/;
194    s/\bENTRY\b/@ $&/;
195    s/\bASSERT\b/@ $&/;
196    s/\bGBLL\b/@ $&/;
197    s/\bGBLA\b/@ $&/;
198    s/^\W+OPT\b/@ $&/;
199    s/:OR:/|/g;
200    s/:SHL:/<</g;
201    s/:SHR:/>>/g;
202    s/:AND:/&/g;
203    s/:LAND:/&&/g;
204    s/CPSR/cpsr/;
205    s/SPSR/spsr/;
206    s/ALIGN$/.balign 4/;
207    s/ALIGN\s+([0-9x]+)$/.balign $1/;
208    s/psr_cxsf/psr_all/;
209    s/LTORG/.ltorg/;
210    s/^([A-Za-z_]\w*)[ \t]+EQU/ .set $1,/;
211    s/^([A-Za-z_]\w*)[ \t]+SETL/ .set $1,/;
212    s/^([A-Za-z_]\w*)[ \t]+SETA/ .set $1,/;
213    s/^([A-Za-z_]\w*)[ \t]+\*/ .set $1,/;
214
215    #  {PC} + 0xdeadfeed  -->  . + 0xdeadfeed
216    s/\{PC\} \+/ \. +/;
217
218    # Single hex constant on the line !
219    #
220    # >>> NOTE <<<
221    #   Double-precision floats in gcc are always mixed-endian, which means
222    #   bytes in two words are little-endian, but words are big-endian.
223    #   So, 0x0000deadfeed0000 would be stored as 0x0000dead at low address
224    #   and 0xfeed0000 at high address.
225    #
226    s/\bDCFD\b[ \t]+0x([a-fA-F0-9]{8})([a-fA-F0-9]{8})/.long 0x$1, 0x$2/;
227    # Only decimal constants on the line, no hex !
228    s/\bDCFD\b[ \t]+([0-9\.\-]+)/.double $1/;
229
230    # Single hex constant on the line !
231#    s/\bDCFS\b[ \t]+0x([a-f0-9]{8})([a-f0-9]{8})/.long 0x$1, 0x$2/;
232    # Only decimal constants on the line, no hex !
233#    s/\bDCFS\b[ \t]+([0-9\.\-]+)/.double $1/;
234    s/\bDCFS[ \t]+0x/.word 0x/;
235    s/\bDCFS\b/.float/;
236
237    s/^([A-Za-z_]\w*)[ \t]+DCD/$1 .word/;
238    s/\bDCD\b/.word/;
239    s/^([A-Za-z_]\w*)[ \t]+DCW/$1 .short/;
240    s/\bDCW\b/.short/;
241    s/^([A-Za-z_]\w*)[ \t]+DCB/$1 .byte/;
242    s/\bDCB\b/.byte/;
243    s/^([A-Za-z_]\w*)[ \t]+\%/.comm $1,/;
244    s/^[A-Za-z_\.]\w+/$&:/;
245    s/^(\d+)/$1:/;
246    s/\%(\d+)/$1b_or_f/;
247    s/\%[Bb](\d+)/$1b/;
248    s/\%[Ff](\d+)/$1f/;
249    s/\%[Ff][Tt](\d+)/$1f/;
250    s/&([\dA-Fa-f]+)/0x$1/;
251    if ( /\b2_[01]+\b/ ) {
252      s/\b2_([01]+)\b/conv$1&&&&/g;
253      while ( /[01][01][01][01]&&&&/ ) {
254        s/0000&&&&/&&&&0/g;
255        s/0001&&&&/&&&&1/g;
256        s/0010&&&&/&&&&2/g;
257        s/0011&&&&/&&&&3/g;
258        s/0100&&&&/&&&&4/g;
259        s/0101&&&&/&&&&5/g;
260        s/0110&&&&/&&&&6/g;
261        s/0111&&&&/&&&&7/g;
262        s/1000&&&&/&&&&8/g;
263        s/1001&&&&/&&&&9/g;
264        s/1010&&&&/&&&&A/g;
265        s/1011&&&&/&&&&B/g;
266        s/1100&&&&/&&&&C/g;
267        s/1101&&&&/&&&&D/g;
268        s/1110&&&&/&&&&E/g;
269        s/1111&&&&/&&&&F/g;
270      }
271      s/000&&&&/&&&&0/g;
272      s/001&&&&/&&&&1/g;
273      s/010&&&&/&&&&2/g;
274      s/011&&&&/&&&&3/g;
275      s/100&&&&/&&&&4/g;
276      s/101&&&&/&&&&5/g;
277      s/110&&&&/&&&&6/g;
278      s/111&&&&/&&&&7/g;
279      s/00&&&&/&&&&0/g;
280      s/01&&&&/&&&&1/g;
281      s/10&&&&/&&&&2/g;
282      s/11&&&&/&&&&3/g;
283      s/0&&&&/&&&&0/g;
284      s/1&&&&/&&&&1/g;
285      s/conv&&&&/0x/g;
286    }
287
288    if ( /commandline/)
289    {
290        if( /-bigend/)
291        {
292            $bigend=1;
293        }
294    }
295
296    if ( /\bDCDU\b/ )
297    {
298        my $cmd=$_;
299        my $value;
300        my $prefix;
301        my $w1;
302        my $w2;
303        my $w3;
304        my $w4;
305
306        s/\s+DCDU\b/@ $&/;
307
308        $cmd =~ /\bDCDU\b\s+0x(\d+)/;
309        $value = $1;
310        $value =~ /(\w\w)(\w\w)(\w\w)(\w\w)/;
311        $w1 = $1;
312        $w2 = $2;
313        $w3 = $3;
314        $w4 = $4;
315
316        if( $bigend ne "")
317        {
318            # big endian
319            $prefix = "\t.byte\t0x".$w1.";".
320                      "\t.byte\t0x".$w2.";".
321                      "\t.byte\t0x".$w3.";".
322                      "\t.byte\t0x".$w4."; ";
323        }
324        else
325        {
326            # little endian
327            $prefix = "\t.byte\t0x".$w4.";".
328                      "\t.byte\t0x".$w3.";".
329                      "\t.byte\t0x".$w2.";".
330                      "\t.byte\t0x".$w1."; ";
331        }
332        $_=$prefix.$_;
333    }
334
335    if ( /\badrl\b/i )
336    {
337        s/\badrl\s+(\w+)\s*,\s*(\w+)/ldr $1,=$2/i;
338        $addPadding = 1;
339    }
340    s/\bEND\b/@ END/;
341} continue {
342    printf ("%s", $_) if $printit;
343    if ($addPadding != 0)
344    {
345        printf ("   mov r0,r0\n");
346        $addPadding = 0;
347    }
348}
349#If we had a code section, mark that this object doesn't need an executable
350# stack.
351if ($nxstack && !$apple) {
352    printf ("    .section\t.note.GNU-stack,\"\",\%\%progbits\n");
353}
354