libwin-hivex-perl 1.3.3-1 / usr / bin / hivexregedit

#!/usr/bin/perl -w
# Copyright (C) 2010-2011 Red Hat Inc.
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.

use warnings;
use strict;

use Win::Hivex;
use Win::Hivex::Regedit qw(reg_import reg_export);

use Pod::Usage;
use Getopt::Long;

=encoding utf8

=head1 NAME

hivexregedit - Merge and export Registry changes from regedit-format files.

=head1 SYNOPSIS

 hivexregedit --merge [--prefix prefix] [--encoding enc] \
         hivefile [regfile]

 hivexregedit --export [--prefix prefix] hivefile key > regfile

=head1 DESCRIPTION

Please note hivexregedit is a low-level tool for manipulating hive
files directly.  To merge or export registry changes to Windows
virtual machines it's better to use L<virt-win-reg(1)>.

Given a local binary ("hive") file, there are two modes.  C<--merge>
imports (merges) changes from a regedit-format file into the hive.  It
is similar to using the C</s> switch in Windows regedit.exe.

C<--export> exports a Registry key (recursively) into the regedit format.

=head2 ENCODING

C<hivexregedit> expects that regedit files have already been re-encoded
in the local encoding.  Usually on Linux hosts, this means UTF-8 with
Unix-style line endings.  Since Windows regedit files are often in
UTF-16LE with Windows-style line endings, you may need to re-encode the
whole file before or after processing.

To re-encode a file from Windows format to Linux (before processing it
with the C<--merge> option), you would do something like this:

 iconv -f utf-16le -t utf-8 < win.reg | dos2unix > linux.reg

To go in the opposite direction, after using C<--export> and before
sending the file to a Windows user, do something like this:

 unix2dos linux.reg | iconv -f utf-8 -t utf-16le > win.reg

For more information about encoding, see L<Win::Hivex::Regedit(3)>.

If you are unsure about the current encoding, use the L<file(1)>
command.  Recent versions of Windows regedit.exe produce a UTF-16LE
file with Windows-style (CRLF) line endings, like this:

 $ file software.reg
 software.reg: Little-endian UTF-16 Unicode text, with very long lines,
 with CRLF line terminators

This file would need conversion before you could C<--merge> it.

=head2 SHELL QUOTING

Be careful when passing parameters containing C<\> (backslash) in the
shell.  Usually you will have to use 'single quotes' or double
backslashes (but not both) to protect them from the shell.

=head2 CurrentControlSet etc.

Registry keys like C<CurrentControlSet> don't really exist in the
Windows Registry at the level of the hive file, and therefore you
cannot modify these.

C<CurrentControlSet> is usually an alias for C<ControlSet001>.  In
some circumstances it might refer to another control set.  The way
to find out is to look at the C<HKLM\SYSTEM\Select> key:

 $ hivexregedit --export SYSTEM '\Select'
 [\Select]
 "Current"=dword:00000001
 "Default"=dword:00000001
 "Failed"=dword:00000000
 "LastKnownGood"=dword:00000002

"Current" is the one which Windows will choose when it boots.

Similarly, other C<Current...> keys in the path may need to
be replaced.

=head1 EXAMPLE

 $ virt-cat WindowsGuest /Windows/System32/config/software > software.hive
 $ hivexregedit --export \
     --prefix 'HKEY_LOCAL_MACHINE\SOFTWARE' \
     software.hive '\Microsoft' > ms-keys.reg

 $ hivexregedit --merge system.hive \
     --prefix 'HKEY_LOCAL_MACHINE\SYSTEM' additions.reg

=head1 OPTIONS

=over 4

=cut

my $help;

=item B<--help>

Display help.

=cut

my $debug;

=item B<--debug>

Enable debugging in the hivex library.  This is useful for diagnosing
bugs and also malformed hive files.

=cut

my $merge;

=item B<--merge>

 hivexregedit --merge [--prefix prefix] [--encoding enc] \
         hivefile [regfile]

Merge C<regfile> (a regedit-format text file) into the hive
C<hivefile>.  If C<regfile> is omitted, then the program reads from
standard input.  (Also you can give multiple input files).

C<--prefix> specifies the Windows Registry prefix.  It is almost
always necessary to use this when dealing with real hive files.

C<--encoding> specifies the encoding for unmarked strings in the
input.  It defaults to C<UTF-16LE> which should work for recent
versions of Windows.  Another possibility is to use C<ASCII>.

=cut

my $export;

=item B<--export>

 hivexregedit --export [--prefix prefix] hivefile key > regfile

C<key> is a path within the hive C<hivefile>.  (The key should not
contain any prefix and should be quoted to defend backslashes from the
shell).  The key is exported, recursively, to standard output in the
textual regedit format.

C<--prefix> specifies the Windows Registry prefix.  It is almost
always necessary to use this when dealing with real hive files.

=cut

my $prefix;

=item B<--prefix> prefix

Hive files and Windows Registry key names are indirectly related.  For
example, inside the software hive, all keys are stored relative to
C<HKEY_LOCAL_MACHINE\SOFTWARE>.  Thus
C<HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft> appears in the hive file as
C<\Microsoft>.

The hive format itself does not store this prefix, so you have to
supply it based on outside knowledge.  (L<virt-win-reg(1)>, amongst
other things, already knows about this).

Usually it is sufficient to pass the parameter
C<--prefix 'HKEY_LOCAL_MACHINE\SOFTWARE'> or similar when doing
merges and exports.

=cut

my $encoding;

=item B<--encoding> UTF-16LE|ASCII

When merging (only), you may need to specify the encoding for strings
to be used in the hive file.  This is explained in detail in
L<Win::Hivex::Regedit(3)/ENCODING STRINGS>.

The default is to use UTF-16LE, which should work with recent versions
of Windows.

=cut

my $unsafe_printable_strings;

=item B<--unsafe-printable-strings>

When exporting (only), assume strings are UTF-16LE and print them as
strings instead of hex sequences.  Remove the final zero codepoint
from strings if present.

This is unsafe and does not preserve the fidelity of strings in the
original hive for various reasons:

=over 4

=item *

Assumes the original encoding is UTF-16LE.  ASCII strings and strings
in other encodings will be corrupted by this transformation.

=item *

Assumes that everything which has type 1 or 2 is really a string
and that everything else is not a string, but the type field in
real hives is not reliable.

=item *

Loses information about whether a zero codepoint followed the string
in the hive or not.

=back

This all happens because the hive itself contains no information about
how strings are encoded (see
L<Win::Hivex::Regedit(3)/ENCODING STRINGS>).

You should only use this option for quick hacking and debugging of the
hive contents, and I<never> use it if the output is going to be passed
into another program or stored in another hive.

=back

=cut

GetOptions ("help|?" => \$help,
            "debug" => \$debug,
            "merge|import" => \$merge,
            "export" => \$export,
            "prefix=s" => \$prefix,
            "encoding=s" => \$encoding,
            "unsafe-printable-strings" => \$unsafe_printable_strings,
    ) or pod2usage (2);
pod2usage (1) if $help;

if ($merge && $export) {
    die "hivexregedit: cannot use --merge and --export at the same time\n"
}

unless ($merge || $export) {
    die "hivexregedit: use --merge or --export, see the manpage for help\n"
}

if ($export && defined $encoding) {
    die "hivexregedit: --encoding has no effect when used with --export\n"
}

if ($merge) {                   # --merge (reg_import)
    if (@ARGV < 1) {
        die "hivexregedit --merge hivefile [input.reg ...]\n"
    }

    my $hivefile = shift @ARGV;

    my $h = Win::Hivex->open ($hivefile, write => 1, debug => $debug);

    # Read from stdin unless other files have been specified.
    unshift (@ARGV, '-') unless @ARGV;

    foreach (@ARGV) {
        open FILE, $_ or die "$_: $!";
        reg_import (\*FILE, sub {
            local $_ = shift;
            # Remove prefix from the start of the path, matching
            # case insensitively.
            if (defined $prefix) {
                my $len = length $prefix;
                if (length $_ >= $len &&
                    lc (substr ($_, 0, $len)) eq lc ($prefix)) {
                    $_ = substr ($_, $len);
                }
            }
            ($h, $_)
        });
    }

    $h->commit (undef);
} else {                        # --export (reg_export)
    if (@ARGV != 2) {
        die "hivexregedit --export hivefile key\n"
    }

    my $hivefile = shift @ARGV;
    my $key = shift @ARGV;

    my $h = Win::Hivex->open ($hivefile, debug => $debug);

    print "Windows Registry Editor Version 5.00\n\n";

    reg_export ($h, $key, \*STDOUT,
                prefix => $prefix,
                unsafe_printable_strings => $unsafe_printable_strings);
}

=head1 SEE ALSO

L<virt-win-reg(1)>,
L<Win::Hivex::Regedit(3)>,
L<Win::Hivex(3)>,
L<hivexsh(1)>,
L<dos2unix(1)>,
L<unix2dos(1)>,
L<iconv(1)>,
L<http://libguestfs.org/>.

=head1 AUTHOR

Richard W.M. Jones L<http://people.redhat.com/~rjones/>

=head1 COPYRIGHT

Copyright (C) 2010 Red Hat Inc.

This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.