mirror of
https://github.com/netwide-assembler/nasm.git
synced 2024-11-27 08:10:07 +08:00
d59b2602b3
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
182 lines
5.7 KiB
Perl
Executable File
182 lines
5.7 KiB
Perl
Executable File
#!/usr/bin/perl
|
|
## --------------------------------------------------------------------------
|
|
##
|
|
## Copyright 1996-2009 The NASM Authors - All Rights Reserved
|
|
## See the file AUTHORS included with the NASM distribution for
|
|
## the specific copyright holders.
|
|
##
|
|
## Redistribution and use in source and binary forms, with or without
|
|
## modification, are permitted provided that the following
|
|
## conditions are met:
|
|
##
|
|
## * Redistributions of source code must retain the above copyright
|
|
## notice, this list of conditions and the following disclaimer.
|
|
## * Redistributions in binary form must reproduce the above
|
|
## copyright notice, this list of conditions and the following
|
|
## disclaimer in the documentation and/or other materials provided
|
|
## with the distribution.
|
|
##
|
|
## THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
|
|
## CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
|
|
## INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
|
|
## MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
|
|
## DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
|
|
## CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
|
|
## SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
|
|
## NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
|
|
## LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
## HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
|
|
## CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
|
|
## OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
|
|
## EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
|
|
##
|
|
## --------------------------------------------------------------------------
|
|
|
|
#
|
|
# Generate a perfect hash for directive parsing
|
|
#
|
|
# Usage: directiv.pl directiv.dat directiv.c directiv.h
|
|
#
|
|
|
|
require 'phash.ph';
|
|
|
|
my($output, $directives_dat, $outfile) = @ARGV;
|
|
|
|
@directives = ();
|
|
@specials = ('none', 'unknown');
|
|
|
|
open(DD, "< ${directives_dat}\0")
|
|
or die "$0: cannot open: ${directives_dat}: $!\n";
|
|
while (defined($line = <DD>)) {
|
|
chomp $line;
|
|
if ($line =~ /^\s*([[:alnum:]]+)\s*(|[\;\#].*)$/) {
|
|
push(@directives, $1);
|
|
}
|
|
}
|
|
close(DD);
|
|
|
|
if ($output eq 'h') {
|
|
open(H, "> ${outfile}\0")
|
|
or die "$0: cannot create: ${outfile}: $!\n";
|
|
|
|
print H "/*\n";
|
|
print H " * This file is generated from directiv.dat\n";
|
|
print H " * by directiv.pl; do not edit.\n";
|
|
print H " */\n";
|
|
print H "\n";
|
|
|
|
print H "#ifndef NASM_DIRECTIVES_H\n";
|
|
print H "#define NASM_DIRECTIVES_H\n";
|
|
print H "\n";
|
|
|
|
$c = '{';
|
|
print H "enum directives ";
|
|
foreach $d (@specials) {
|
|
print H "$c\n D_$d";
|
|
$c = ',';
|
|
}
|
|
foreach $d (@directives) {
|
|
print H "$c\n D_\U$d";
|
|
$c = ',';
|
|
}
|
|
print H "\n};\n\n";
|
|
printf H "extern const char * const directives[%d];\n",
|
|
scalar(@directives)+scalar(@specials);
|
|
print H "enum directives find_directive(const char *token);\n\n";
|
|
print H "#endif /* NASM_DIRECTIVES_H */\n";
|
|
} elsif ($output eq 'c') {
|
|
%directive = ();
|
|
$n = 0;
|
|
foreach $d (@directives) {
|
|
if (exists($directive{$d})) {
|
|
die "$0: $directives_dat: duplicate directive: $d\n";
|
|
}
|
|
$directive{$d} = $n++; # This is zero-based, unlike the enum!
|
|
}
|
|
|
|
@hashinfo = gen_perfect_hash(\%directive);
|
|
if (!@hashinfo) {
|
|
die "$0: no hash found\n";
|
|
}
|
|
|
|
# Paranoia...
|
|
verify_hash_table(\%directive, \@hashinfo);
|
|
|
|
($n, $sv, $g) = @hashinfo;
|
|
|
|
die if ($n & ($n-1));
|
|
|
|
open(C, "> ${outfile}\0")
|
|
or die "$0: cannot create: ${directives_c}: $!\n";
|
|
|
|
print C "/*\n";
|
|
print C " * This file is generated from directiv.dat\n";
|
|
print C " * by directiv.pl; do not edit.\n";
|
|
print C " */\n";
|
|
print C "\n";
|
|
|
|
print C "#include \"compiler.h\"\n";
|
|
print C "#include <string.h>\n";
|
|
print C "#include \"nasm.h\"\n";
|
|
print C "#include \"hashtbl.h\"\n";
|
|
print C "#include \"directiv.h\"\n";
|
|
print C "\n";
|
|
|
|
printf C "const char * const directives[%d] =\n",
|
|
scalar(@directives)+scalar(@specials);
|
|
$c = '{';
|
|
foreach $d (@specials) {
|
|
print C "$c\n NULL";
|
|
$c = ',';
|
|
}
|
|
foreach $d (@directives) {
|
|
print C "$c\n \"$d\"";
|
|
$c = ',';
|
|
}
|
|
print C "\n};\n\n";
|
|
|
|
print C "enum directives find_directive(const char *token)\n";
|
|
print C "{\n";
|
|
|
|
# Put a large value in unused slots. This makes it extremely unlikely
|
|
# that any combination that involves unused slot will pass the range test.
|
|
# This speeds up rejection of unrecognized tokens, i.e. identifiers.
|
|
print C "#define UNUSED 16383\n";
|
|
|
|
print C " static const int16_t hash1[$n] = {\n";
|
|
for ($i = 0; $i < $n; $i++) {
|
|
my $h = ${$g}[$i*2+0];
|
|
print C " ", defined($h) ? $h : 'UNUSED', ",\n";
|
|
}
|
|
print C " };\n";
|
|
|
|
print C " static const int16_t hash2[$n] = {\n";
|
|
for ($i = 0; $i < $n; $i++) {
|
|
my $h = ${$g}[$i*2+1];
|
|
print C " ", defined($h) ? $h : 'UNUSED', ",\n";
|
|
}
|
|
print C " };\n";
|
|
|
|
print C " uint32_t k1, k2;\n";
|
|
print C " uint64_t crc;\n";
|
|
# For correct overflow behavior, "ix" should be unsigned of the same
|
|
# width as the hash arrays.
|
|
print C " uint16_t ix;\n";
|
|
print C "\n";
|
|
printf C " crc = crc64i(UINT64_C(0x%08x%08x), token);\n",
|
|
$$sv[0], $$sv[1];
|
|
print C " k1 = (uint32_t)crc;\n";
|
|
print C " k2 = (uint32_t)(crc >> 32);\n";
|
|
print C "\n";
|
|
printf C " ix = hash1[k1 & 0x%x] + hash2[k2 & 0x%x];\n", $n-1, $n-1;
|
|
printf C " if (ix >= %d)\n", scalar(@directives);
|
|
print C " return D_unknown;\n";
|
|
print C "\n";
|
|
printf C " ix += %d;\n", scalar(@specials);
|
|
print C " if (nasm_stricmp(token, directives[ix]))\n";
|
|
print C " return D_unknown;\n";
|
|
print C "\n";
|
|
print C " return ix;\n";
|
|
print C "}\n";
|
|
}
|