package Verilog::VCD;
use warnings;
use strict;
use Carp qw(croak);
require Exporter;
our @ISA = qw(Exporter);
our @EXPORT_OK = qw(parse_vcd list_sigs get_timescale get_endtime);
our %EXPORT_TAGS = (all => \@EXPORT_OK);
our $VERSION = '0.03';
my $timescale;
my $endtime;
sub list_sigs {
# Parse input VCD file into data structure,
# then return just a list of the signal names.
my $file = shift;
unless (defined $file) {
croak('Error: list_sigs requires a filename. It seems like no ',
'filename was provided or filename was undefined');
}
my $vcd = parse_vcd($file, {only_sigs => 1});
my @sigs;
for my $code (keys %{ $vcd }) {
my @nets = @{ $vcd->{$code}->{nets} };
push @sigs, map { "$_->{hier}.$_->{name}" } @nets;
}
return @sigs;
}
sub parse_vcd {
# Parse input VCD file into data structure.
# Also, print t-v pairs to STDOUT, if requested.
my ($file, $opt) = @_;
unless (defined $file) {
croak('Error: parse_vcd requires a filename. It seems like no ',
'filename was provided or filename was undefined');
}
if ($opt) {
unless (ref($opt) eq 'HASH') {
croak('Error: If options are passed to parse_vcd, they must be ',
'passed as a hash reference.');
}
}
my $only_sigs = (exists $opt->{only_sigs}) ? 1 : 0;
my $all_sigs;
my %usigs;
if (exists $opt->{siglist}) {
%usigs = map { $_ => 1 } @{ $opt->{siglist} };
unless (%usigs) {
croak('Error: The signal list passed using siglist was empty.');
}
$all_sigs = 0;
}
else {
$all_sigs = 1;
}
my $use_stdout = (exists $opt->{use_stdout}) ? 1 : 0;
my %data;
my $mult;
my $num_sigs;
my @hier;
my $time = 0;
open my $fh, '<', $file or croak("Error: Can not open VCD file $file: $!");
while (<$fh>) {
chomp;
s/ ^ \s+ //x;
if (/ \$enddefinitions \b /x) {
$num_sigs = scalar keys %data;
unless ($num_sigs) {
if ($all_sigs) {
croak("Error: No signals were found in the VCD file $file.",
'Check the VCD file for proper $var syntax.');
}
else {
croak("Error: No matching signals were found in the VCD file $file.",
' Use list_sigs to view all signals in the VCD file.');
}
}
if (($num_sigs>1) and $use_stdout) {
croak("Error: There are too many signals ($num_sigs) for output ",
'to STDOUT. Use list_sigs to select a single signal.');
}
last if $only_sigs;
}
elsif (/ \$timescale \b /x) {
my $statement = $_;
my $line;
unless (/ \$end \b /x) {
do {
$line = <$fh>;
$statement .= " $line";
} until ($line =~ / \$end \b /x);
}
$mult = calc_mult($statement, $opt);
}
elsif (/ \$scope \b /x) {
# assumes all on one line
# $scope module dff $end
push @hier, (split)[2]; # just keep scope name
}
elsif (/ \$upscope \b /x) {
pop @hier;
}
elsif (/ \$var \b /x) {
# assumes all on one line:
# $var reg 1 *@ data $end
# $var wire 4 ) addr [3:0] $end
my (undef, $type, $size, $code, $name) = split /\s+/, $_, 5;
$name =~ s/ \s+ \$end .* //x;
$name =~ s/ \s //xg;
my $path = join '.', @hier;
my $full_name = "$path.$name";
push @{ $data{$code}{nets} }, {
type => $type,
name => $name,
size => $size,
hier => $path,
} if exists $usigs{$full_name} or $all_sigs;
}
elsif (/ ^ [#] (\d+) /x) {
$time = $mult * $1;
$endtime = $time;
}
elsif (/ ^ ([01zx]) (.+) /xi or / ^ [br] (\S+) \s+ (.+) /xi) {
my $value = lc $1;
my $code = $2;
if (exists $data{$code}) {
if ($use_stdout) {
print "$time $value\n";
}
else {
push @{ $data{$code}{tv} }, [$time, $value];
}
}
}
}
close $fh;
return \%data;
}
sub calc_mult {
# Calculate a new multiplier for time values.
# Input statement is complete timescale, for example:
# $timescale 10ns $end
# Input new_units is one of s|ms|us|ns|ps|fs.
# Return numeric multiplier.
# Also sets the package $timescale variable.
my ($statement, $opt) = @_;
my @fields = split /\s+/, $statement;
pop @fields; # delete $end from array
shift @fields; # delete $timescale from array
my $tscale = join '', @fields;
my $new_units;
if (exists $opt->{timescale}) {
$new_units = lc $opt->{timescale};
$new_units =~ s/\s//g;
$timescale = "1$new_units";
}
else {
$timescale = $tscale;
return 1;
}
my $mult;
my $units;
if ($tscale =~ / (\d+) ([a-z]+) /xi) {
$mult = $1;
$units = lc $2;
}
else {
croak("Error: Unsupported timescale found in VCD file: $tscale. ",
'Refer to the Verilog LRM.');
}
my %mults = (
'fs' => 1e-15,
'ps' => 1e-12,
'ns' => 1e-09,
'us' => 1e-06,
'ms' => 1e-03,
's' => 1e-00,
);
my $usage = join '|', sort { $mults{$a} <=> $mults{$b} } keys %mults;
my $scale;
if (exists $mults{$units}) {
$scale = $mults{$units};
}
else {
croak("Error: Unsupported timescale units found in VCD file: $units. ",
"Supported values are: $usage");
}
my $new_scale;
if (exists $mults{$new_units}) {
$new_scale = $mults{$new_units};
}
else {
croak("Error: Illegal user-supplied timescale: $new_units. ",
"Legal values are: $usage");
}
return (($mult * $scale) / $new_scale);
}
sub get_timescale {
return $timescale;
}
sub get_endtime {
return $endtime;
}
=head1 NAME
Verilog::VCD - Parse a Verilog VCD text file
=head1 VERSION
This document refers to Verilog::VCD version 0.03.
=head1 SYNOPSIS
use Verilog::VCD qw(parse_vcd);
my $vcd = parse_vcd('/path/to/some.vcd');
=head1 DESCRIPTION
Verilog is a Hardware Description Language (HDL) used to model digital logic.
While simulating logic circuits, the values of signals can be written out to
a Value Change Dump (VCD) file. This module can be used to parse a VCD file
so that further analysis can be performed on the simulation data. The entire
VCD file can be stored in a Perl data structure and manipulated using
standard hash and array operations.
=head2 Input File Syntax
The syntax of the VCD text file is described in the documentation of
the IEEE standard for Verilog. Only the four-state VCD format is supported.
The extended VCD format (with strength information) is not supported.
Since the input file is assumed to be legal VCD syntax, only minimal
validation is performed.
=head1 SUBROUTINES
=head2 parse_vcd($file, $opt_ref)
Parse a VCD file and return a reference to a data structure which
includes hierarchical signal definitions and time-value data for all
the specified signals. A file name is required. By default, all
signals in the VCD file are included, and times are in units
specified by the C<$timescale> VCD keyword.
my $vcd = parse_vcd('/path/to/some.vcd');
It returns a reference to a nested data structure. The top of the
structure is a Hash-of-Hashes. The keys to the top hash are the VCD
identifier codes for each signal. The following is an example
representation of a very simple VCD file. It shows one signal named
C<chip.cpu.alu.clk>, whose VCD code is C<+>. The time-value pairs
are stored as an Array-of-Arrays, referenced by the C<tv> key. The
time is always the first number in the pair, and the times are stored in
increasing order in the array.
{
'+' => {
'tv' => [
[
'0',
'1'
],
[
'12',
'0'
],
],
'nets' => [
{
'hier' => 'chip.cpu.alu.',
'name' => 'clk',
'type' => 'reg',
'size' => '1'
}
]
}
};
Since each code could have multiple hierarchical signal names, the names are
stored as an Array-of-Hashes, referenced by the C<nets> key. The example above
only shows one signal name for the code.
=head3 OPTIONS
Options to C<parse_vcd> should be passed as a hash reference.
=over 4
=item timescale
It is possible to scale all times in the VCD file to a desired timescale.
To specify a certain timescale, such as nanoseconds:
my $vcd = parse_vcd($file, {timescale => 'ns'});
Valid timescales are:
s ms us ns ps fs
=item siglist
If only a subset of the signals included in the VCD file are needed,
they can be specified by a signal list passed as an array reference.
The signals should be full hierarchical paths separated by the dot
character. For example:
my @signals = qw(
top.chip.clk
top.chip.cpu.alu.status
top.chip.cpu.alu.sum[15:0]
);
my $vcd = parse_vcd($file, {siglist => \@signals});
Limiting the number of signals can substantially reduce memory usage of the
returned data structure because only the time-value data for the selected
signals is loaded into the data structure.
=item use_stdout
It is possible to print time-value pairs directly to STDOUT for a
single signal using the C<use_stdout> option. If the VCD file has
more than one signal, the C<siglist> option must also be used, and there
must only be one signal specified. For example:
my $vcd = parse_vcd($file, {
use_stdout => 1,
siglist => [(top.clk)]
});
The time-value pairs are output as space-separated tokens, one per line.
For example:
0 x
15 0
277 1
500 0
Times are listed in the first column.
Times units can be controlled by the C<timescale> option.
=item only_sigs
Parse a VCD file and return a reference to a data structure which
includes only the hierarchical signal definitions. Parsing stops once
all signals have been found. Therefore, no time-value data are
included in the returned data structure. This is useful for
analyzing signals and hierarchies.
my $vcd = parse_vcd($file, {only_sigs => 1});
=back
=head2 list_sigs($file)
Parse a VCD file and return a list of all signals in the VCD file.
Parsing stops once all signals have been found. This is
helpful for deciding how to limit what signals are parsed.
Here is an example:
my @signals = list_sigs('input.vcd');
The signals are full hierarchical paths separated by the dot character
top.chip.cpu.alu.status
top.chip.cpu.alu.sum[15:0]
=head2 get_timescale( )
This returns a string corresponding to the timescale as specified
by the C<$timescale> VCD keyword. It returns the timescale for
the last VCD file parsed. If called before a file is parsed, it
returns an undefined value. If the C<parse_vcd> C<timescale> option
was used to specify a timescale, the specified value will be returned
instead of what is in the VCD file.
my $vcd = parse_vcd($file); # Parse a file first
my $ts = get_timescale(); # Then query the timescale
=head2 get_endtime( )
This returns the last time found in the VCD file, scaled
appropriately. It returns the last time for the last VCD file parsed.
If called before a file is parsed, it returns an undefined value.
my $vcd = parse_vcd($file); # Parse a file first
my $et = get_endtime(); # Then query the endtime
=head1 EXPORT
Nothing is exported by default. Functions may be exported individually, or
all functions may be exported at once, using the special tag C<:all>.
=head1 DIAGNOSTICS
Error conditions cause the program to die using C<croak> from the
L<Carp|Carp> Core module.
=head1 LIMITATIONS
Only the following VCD keywords are parsed:
$end $scope
$enddefinitions $upscope
$timescale $var
The extended VCD format (with strength information) is not supported.
The default mode of C<parse_vcd> is to load the entire VCD file into the
data structure. This could be a problem for huge VCD files. The best solution
to any memory problem is to plan ahead and keep VCD files as small as possible.
When simulating, dump fewer signals and scopes, and use shorter dumping
time ranges. Another technique is to parse only a small list of signals
using the C<siglist> option; this method only loads the desired signals into
the data structure. Finally, the C<use_stdout> option will parse the input VCD
file line-by-line, instead of loading it into the data structure, and directly
prints time-value data to STDOUT. The drawback is that this only applies to
one signal.
=head1 BUGS
There are no known bugs in this module.
=head1 SEE ALSO
Refer to the following Verilog documentation:
IEEE Standard for Verilog (c) Hardware Description Language
IEEE Std 1364-2005
Section 18.2, "Format of four-state VCD file"
=head1 AUTHOR
Gene Sullivan (gsullivan@cpan.org)
=head1 COPYRIGHT AND LICENSE
Copyright (c) 2012 Gene Sullivan. All rights reserved.
This module is free software; you can redistribute it and/or modify
it under the same terms as Perl itself. See L<perlartistic|perlartistic>.
=cut
1;