cannam@154: #!/usr/bin/perl
cannam@154: # Copyright (C) 2002-2013 Xiph.org Foundation
cannam@154: #
cannam@154: # Redistribution and use in source and binary forms, with or without
cannam@154: # modification, are permitted provided that the following conditions
cannam@154: # are met:
cannam@154: #
cannam@154: # - Redistributions of source code must retain the above copyright
cannam@154: # notice, this list of conditions and the following disclaimer.
cannam@154: #
cannam@154: # - Redistributions in binary form must reproduce the above copyright
cannam@154: # notice, this list of conditions and the following disclaimer in the
cannam@154: # documentation and/or other materials provided with the distribution.
cannam@154: #
cannam@154: # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
cannam@154: # ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
cannam@154: # LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
cannam@154: # A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
cannam@154: # OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
cannam@154: # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
cannam@154: # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
cannam@154: # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
cannam@154: # LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
cannam@154: # NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
cannam@154: # SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
cannam@154: 
cannam@154: my $bigend;  # little/big endian
cannam@154: my $nxstack;
cannam@154: my $apple = 0;
cannam@154: my $symprefix = "";
cannam@154: 
cannam@154: $nxstack = 0;
cannam@154: 
cannam@154: eval 'exec /usr/local/bin/perl -S $0 ${1+"$@"}'
cannam@154:     if $running_under_some_shell;
cannam@154: 
cannam@154: while ($ARGV[0] =~ /^-/) {
cannam@154:     $_ = shift;
cannam@154:   last if /^--$/;
cannam@154:     if (/^-n$/) {
cannam@154:     $nflag++;
cannam@154:     next;
cannam@154:     }
cannam@154:     if (/^--apple$/) {
cannam@154:         $apple = 1;
cannam@154:         $symprefix = "_";
cannam@154:         next;
cannam@154:     }
cannam@154:     die "I don't recognize this switch: $_\\n";
cannam@154: }
cannam@154: $printit++ unless $nflag;
cannam@154: 
cannam@154: $\ = "\n";      # automatically add newline on print
cannam@154: $n=0;
cannam@154: 
cannam@154: $thumb = 0;     # ARM mode by default, not Thumb.
cannam@154: @proc_stack = ();
cannam@154: 
cannam@154: printf ("    .syntax unified\n");
cannam@154: 
cannam@154: LINE:
cannam@154: while (<>) {
cannam@154: 
cannam@154:     # For ADRLs we need to add a new line after the substituted one.
cannam@154:     $addPadding = 0;
cannam@154: 
cannam@154:     # First, we do not dare to touch *anything* inside double quotes, do we?
cannam@154:     # Second, if you want a dollar character in the string,
cannam@154:     # insert two of them -- that's how ARM C and assembler treat strings.
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+DCB[ \t]*\"/$1:   .ascii \"/   && do { s/\$\$/\$/g; next };
cannam@154:     s/\bDCB\b[ \t]*\"/.ascii \"/                          && do { s/\$\$/\$/g; next };
cannam@154:     s/^(\S+)\s+RN\s+(\S+)/$1 .req r$2/                    && do { s/\$\$/\$/g; next };
cannam@154:     # If there's nothing on a line but a comment, don't try to apply any further
cannam@154:     #  substitutions (this is a cheap hack to avoid mucking up the license header)
cannam@154:     s/^([ \t]*);/$1@/                                     && do { s/\$\$/\$/g; next };
cannam@154:     # If substituted -- leave immediately !
cannam@154: 
cannam@154:     s/@/,:/;
cannam@154:     s/;/@/;
cannam@154:     while ( /@.*'/ ) {
cannam@154:       s/(@.*)'/$1/g;
cannam@154:     }
cannam@154:     s/\{FALSE\}/0/g;
cannam@154:     s/\{TRUE\}/1/g;
cannam@154:     s/\{(\w\w\w\w+)\}/$1/g;
cannam@154:     s/\bINCLUDE[ \t]*([^ \t\n]+)/.include \"$1\"/;
cannam@154:     s/\bGET[ \t]*([^ \t\n]+)/.include \"${ my $x=$1; $x =~ s|\.s|-gnu.S|; \$x }\"/;
cannam@154:     s/\bIMPORT\b/.extern/;
cannam@154:     s/\bEXPORT\b\s*/.global $symprefix/;
cannam@154:     s/^(\s+)\[/$1IF/;
cannam@154:     s/^(\s+)\|/$1ELSE/;
cannam@154:     s/^(\s+)\]/$1ENDIF/;
cannam@154:     s/IF *:DEF:/ .ifdef/;
cannam@154:     s/IF *:LNOT: *:DEF:/ .ifndef/;
cannam@154:     s/ELSE/ .else/;
cannam@154:     s/ENDIF/ .endif/;
cannam@154: 
cannam@154:     if( /\bIF\b/ ) {
cannam@154:       s/\bIF\b/ .if/;
cannam@154:       s/=/==/;
cannam@154:     }
cannam@154:     if ( $n == 2) {
cannam@154:         s/\$/\\/g;
cannam@154:     }
cannam@154:     if ($n == 1) {
cannam@154:         s/\$//g;
cannam@154:         s/label//g;
cannam@154:     $n = 2;
cannam@154:       }
cannam@154:     if ( /MACRO/ ) {
cannam@154:       s/MACRO *\n/.macro/;
cannam@154:       $n=1;
cannam@154:     }
cannam@154:     if ( /\bMEND\b/ ) {
cannam@154:       s/\bMEND\b/.endm/;
cannam@154:       $n=0;
cannam@154:     }
cannam@154: 
cannam@154:     # ".rdata" doesn't work in 'as' version 2.13.2, as it is ".rodata" there.
cannam@154:     #
cannam@154:     if ( /\bAREA\b/ ) {
cannam@154:         my $align;
cannam@154:         $align = "2";
cannam@154:         if ( /ALIGN=(\d+)/ ) {
cannam@154:             $align = $1;
cannam@154:         }
cannam@154:         if ( /CODE/ ) {
cannam@154:             $nxstack = 1;
cannam@154:         }
cannam@154:         s/^(.+)CODE(.+)READONLY(.*)/    .text/;
cannam@154:         s/^(.+)DATA(.+)READONLY(.*)/    .section .rdata/;
cannam@154:         s/^(.+)\|\|\.data\|\|(.+)/    .data/;
cannam@154:         s/^(.+)\|\|\.bss\|\|(.+)/    .bss/;
cannam@154:         s/$/;   .p2align $align/;
cannam@154:         # Enable NEON instructions but don't produce a binary that requires
cannam@154:         # ARMv7. RVCT does not have equivalent directives, so we just do this
cannam@154:         # for all CODE areas.
cannam@154:         if ( /.text/ ) {
cannam@154:             # Separating .arch, .fpu, etc., by semicolons does not work (gas
cannam@154:             # thinks the semicolon is part of the arch name, even when there's
cannam@154:             # whitespace separating them). Sadly this means our line numbers
cannam@154:             # won't match the original source file (we could use the .line
cannam@154:             # directive, which is documented to be obsolete, but then gdb will
cannam@154:             # show the wrong line in the translated source file).
cannam@154:             s/$/;   .arch armv7-a\n   .fpu neon\n   .object_arch armv4t/ unless ($apple);
cannam@154:         }
cannam@154:     }
cannam@154: 
cannam@154:     s/\|\|\.constdata\$(\d+)\|\|/.L_CONST$1/;       # ||.constdata$3||
cannam@154:     s/\|\|\.bss\$(\d+)\|\|/.L_BSS$1/;               # ||.bss$2||
cannam@154:     s/\|\|\.data\$(\d+)\|\|/.L_DATA$1/;             # ||.data$2||
cannam@154:     s/\|\|([a-zA-Z0-9_]+)\@([a-zA-Z0-9_]+)\|\|/@ $&/;
cannam@154:     s/^(\s+)\%(\s)/    .space $1/;
cannam@154: 
cannam@154:     s/\|(.+)\.(\d+)\|/\.$1_$2/;                     # |L80.123| -> .L80_123
cannam@154:     s/\bCODE32\b/.code 32/ && do {$thumb = 0};
cannam@154:     s/\bCODE16\b/.code 16/ && do {$thumb = 1};
cannam@154:     if (/\bPROC\b/)
cannam@154:     {
cannam@154:         my $prefix;
cannam@154:         my $proc;
cannam@154:         /^([A-Za-z_\.]\w+)\b/;
cannam@154:         $proc = $1;
cannam@154:         $prefix = "";
cannam@154:         if ($proc)
cannam@154:         {
cannam@154:             $prefix = $prefix.sprintf("\t.type\t%s, %%function", $proc) unless ($apple);
cannam@154:             # Make sure we $prefix isn't empty here (for the $apple case).
cannam@154:             # We handle mangling the label here, make sure it doesn't match
cannam@154:             # the label handling below (if $prefix would be empty).
cannam@154:             $prefix = $prefix."; ";
cannam@154:             push(@proc_stack, $proc);
cannam@154:             s/^[A-Za-z_\.]\w+/$symprefix$&:/;
cannam@154:         }
cannam@154:         $prefix = $prefix."\t.thumb_func; " if ($thumb);
cannam@154:         s/\bPROC\b/@ $&/;
cannam@154:         $_ = $prefix.$_;
cannam@154:     }
cannam@154:     s/^(\s*)(S|Q|SH|U|UQ|UH)ASX\b/$1$2ADDSUBX/;
cannam@154:     s/^(\s*)(S|Q|SH|U|UQ|UH)SAX\b/$1$2SUBADDX/;
cannam@154:     if (/\bENDP\b/)
cannam@154:     {
cannam@154:         my $proc;
cannam@154:         s/\bENDP\b/@ $&/;
cannam@154:         $proc = pop(@proc_stack);
cannam@154:         $_ = "\t.size $proc, .-$proc".$_ if ($proc && !$apple);
cannam@154:     }
cannam@154:     s/\bSUBT\b/@ $&/;
cannam@154:     s/\bDATA\b/@ $&/;   # DATA directive is deprecated -- Asm guide, p.7-25
cannam@154:     s/\bKEEP\b/@ $&/;
cannam@154:     s/\bEXPORTAS\b/@ $&/;
cannam@154:     s/\|\|(.)+\bEQU\b/@ $&/;
cannam@154:     s/\|\|([\w\$]+)\|\|/$1/;
cannam@154:     s/\bENTRY\b/@ $&/;
cannam@154:     s/\bASSERT\b/@ $&/;
cannam@154:     s/\bGBLL\b/@ $&/;
cannam@154:     s/\bGBLA\b/@ $&/;
cannam@154:     s/^\W+OPT\b/@ $&/;
cannam@154:     s/:OR:/|/g;
cannam@154:     s/:SHL:/<</g;
cannam@154:     s/:SHR:/>>/g;
cannam@154:     s/:AND:/&/g;
cannam@154:     s/:LAND:/&&/g;
cannam@154:     s/CPSR/cpsr/;
cannam@154:     s/SPSR/spsr/;
cannam@154:     s/ALIGN$/.balign 4/;
cannam@154:     s/ALIGN\s+([0-9x]+)$/.balign $1/;
cannam@154:     s/psr_cxsf/psr_all/;
cannam@154:     s/LTORG/.ltorg/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+EQU/ .set $1,/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+SETL/ .set $1,/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+SETA/ .set $1,/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+\*/ .set $1,/;
cannam@154: 
cannam@154:     #  {PC} + 0xdeadfeed  -->  . + 0xdeadfeed
cannam@154:     s/\{PC\} \+/ \. +/;
cannam@154: 
cannam@154:     # Single hex constant on the line !
cannam@154:     #
cannam@154:     # >>> NOTE <<<
cannam@154:     #   Double-precision floats in gcc are always mixed-endian, which means
cannam@154:     #   bytes in two words are little-endian, but words are big-endian.
cannam@154:     #   So, 0x0000deadfeed0000 would be stored as 0x0000dead at low address
cannam@154:     #   and 0xfeed0000 at high address.
cannam@154:     #
cannam@154:     s/\bDCFD\b[ \t]+0x([a-fA-F0-9]{8})([a-fA-F0-9]{8})/.long 0x$1, 0x$2/;
cannam@154:     # Only decimal constants on the line, no hex !
cannam@154:     s/\bDCFD\b[ \t]+([0-9\.\-]+)/.double $1/;
cannam@154: 
cannam@154:     # Single hex constant on the line !
cannam@154: #    s/\bDCFS\b[ \t]+0x([a-f0-9]{8})([a-f0-9]{8})/.long 0x$1, 0x$2/;
cannam@154:     # Only decimal constants on the line, no hex !
cannam@154: #    s/\bDCFS\b[ \t]+([0-9\.\-]+)/.double $1/;
cannam@154:     s/\bDCFS[ \t]+0x/.word 0x/;
cannam@154:     s/\bDCFS\b/.float/;
cannam@154: 
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+DCD/$1 .word/;
cannam@154:     s/\bDCD\b/.word/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+DCW/$1 .short/;
cannam@154:     s/\bDCW\b/.short/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+DCB/$1 .byte/;
cannam@154:     s/\bDCB\b/.byte/;
cannam@154:     s/^([A-Za-z_]\w*)[ \t]+\%/.comm $1,/;
cannam@154:     s/^[A-Za-z_\.]\w+/$&:/;
cannam@154:     s/^(\d+)/$1:/;
cannam@154:     s/\%(\d+)/$1b_or_f/;
cannam@154:     s/\%[Bb](\d+)/$1b/;
cannam@154:     s/\%[Ff](\d+)/$1f/;
cannam@154:     s/\%[Ff][Tt](\d+)/$1f/;
cannam@154:     s/&([\dA-Fa-f]+)/0x$1/;
cannam@154:     if ( /\b2_[01]+\b/ ) {
cannam@154:       s/\b2_([01]+)\b/conv$1&&&&/g;
cannam@154:       while ( /[01][01][01][01]&&&&/ ) {
cannam@154:         s/0000&&&&/&&&&0/g;
cannam@154:         s/0001&&&&/&&&&1/g;
cannam@154:         s/0010&&&&/&&&&2/g;
cannam@154:         s/0011&&&&/&&&&3/g;
cannam@154:         s/0100&&&&/&&&&4/g;
cannam@154:         s/0101&&&&/&&&&5/g;
cannam@154:         s/0110&&&&/&&&&6/g;
cannam@154:         s/0111&&&&/&&&&7/g;
cannam@154:         s/1000&&&&/&&&&8/g;
cannam@154:         s/1001&&&&/&&&&9/g;
cannam@154:         s/1010&&&&/&&&&A/g;
cannam@154:         s/1011&&&&/&&&&B/g;
cannam@154:         s/1100&&&&/&&&&C/g;
cannam@154:         s/1101&&&&/&&&&D/g;
cannam@154:         s/1110&&&&/&&&&E/g;
cannam@154:         s/1111&&&&/&&&&F/g;
cannam@154:       }
cannam@154:       s/000&&&&/&&&&0/g;
cannam@154:       s/001&&&&/&&&&1/g;
cannam@154:       s/010&&&&/&&&&2/g;
cannam@154:       s/011&&&&/&&&&3/g;
cannam@154:       s/100&&&&/&&&&4/g;
cannam@154:       s/101&&&&/&&&&5/g;
cannam@154:       s/110&&&&/&&&&6/g;
cannam@154:       s/111&&&&/&&&&7/g;
cannam@154:       s/00&&&&/&&&&0/g;
cannam@154:       s/01&&&&/&&&&1/g;
cannam@154:       s/10&&&&/&&&&2/g;
cannam@154:       s/11&&&&/&&&&3/g;
cannam@154:       s/0&&&&/&&&&0/g;
cannam@154:       s/1&&&&/&&&&1/g;
cannam@154:       s/conv&&&&/0x/g;
cannam@154:     }
cannam@154: 
cannam@154:     if ( /commandline/)
cannam@154:     {
cannam@154:         if( /-bigend/)
cannam@154:         {
cannam@154:             $bigend=1;
cannam@154:         }
cannam@154:     }
cannam@154: 
cannam@154:     if ( /\bDCDU\b/ )
cannam@154:     {
cannam@154:         my $cmd=$_;
cannam@154:         my $value;
cannam@154:         my $prefix;
cannam@154:         my $w1;
cannam@154:         my $w2;
cannam@154:         my $w3;
cannam@154:         my $w4;
cannam@154: 
cannam@154:         s/\s+DCDU\b/@ $&/;
cannam@154: 
cannam@154:         $cmd =~ /\bDCDU\b\s+0x(\d+)/;
cannam@154:         $value = $1;
cannam@154:         $value =~ /(\w\w)(\w\w)(\w\w)(\w\w)/;
cannam@154:         $w1 = $1;
cannam@154:         $w2 = $2;
cannam@154:         $w3 = $3;
cannam@154:         $w4 = $4;
cannam@154: 
cannam@154:         if( $bigend ne "")
cannam@154:         {
cannam@154:             # big endian
cannam@154:             $prefix = "\t.byte\t0x".$w1.";".
cannam@154:                       "\t.byte\t0x".$w2.";".
cannam@154:                       "\t.byte\t0x".$w3.";".
cannam@154:                       "\t.byte\t0x".$w4."; ";
cannam@154:         }
cannam@154:         else
cannam@154:         {
cannam@154:             # little endian
cannam@154:             $prefix = "\t.byte\t0x".$w4.";".
cannam@154:                       "\t.byte\t0x".$w3.";".
cannam@154:                       "\t.byte\t0x".$w2.";".
cannam@154:                       "\t.byte\t0x".$w1."; ";
cannam@154:         }
cannam@154:         $_=$prefix.$_;
cannam@154:     }
cannam@154: 
cannam@154:     if ( /\badrl\b/i )
cannam@154:     {
cannam@154:         s/\badrl\s+(\w+)\s*,\s*(\w+)/ldr $1,=$2/i;
cannam@154:         $addPadding = 1;
cannam@154:     }
cannam@154:     s/\bEND\b/@ END/;
cannam@154: } continue {
cannam@154:     printf ("%s", $_) if $printit;
cannam@154:     if ($addPadding != 0)
cannam@154:     {
cannam@154:         printf ("   mov r0,r0\n");
cannam@154:         $addPadding = 0;
cannam@154:     }
cannam@154: }
cannam@154: #If we had a code section, mark that this object doesn't need an executable
cannam@154: # stack.
cannam@154: if ($nxstack && !$apple) {
cannam@154:     printf ("    .section\t.note.GNU-stack,\"\",\%\%progbits\n");
cannam@154: }