Message-ID: <1512561090.17323.32.camel@gmail.com>
Date:   Wed, 06 Dec 2017 17:21:30 +0530
From:   kaiwan.billimoria@...il.com
To:     "Tobin C. Harding" <me@...in.cc>
Cc:     Alexander Kapshuk <alexander.kapshuk@...il.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        "kernel-hardening@...ts.openwall.com" 
        <kernel-hardening@...ts.openwall.com>
Subject: Re: [PATCH v3] scripts: leaking_addresses: add support for 32-bit
 kernel addresses

On Wed, 2017-12-06 at 15:04 +1100, Tobin C. Harding wrote:
> On Tue, Dec 05, 2017 at 11:56:44AM +0530, kaiwan.billimoria@...il.com wrote:
> > Currently, leaking_addresses.pl only supports scanning 64 bit
> > architectures. This is due to how the regular expressions are formed. We
> > can do better than this. 32 architectures can be supported if we take
> > into consideration the kernel virtual address split (via the PAGE_OFFSET
> > kernel configurable).
> > 
> > Add support for ix86 32 bit architectures.
> >  - Add command line option for page offset.
> >  - Add command line option for kernel configuration file.
> >  - Parse kernel config file for page offset (CONFIG_PAGE_OFFSET).
> >  - Use page offset when checking for kernel virtual addresses.
> > 
> > 
> > Signed-off-by: Kaiwan N Billimoria <kaiwan.billimoria@...il.com>
> > ---
> 
> Right, this is starting to look awesome.

Great!

> > Note- This patch represents co development by Tobin and Kaiwan (plus suggestions from 
> > Alexander Kapshuk). Applies on Tobin's tree 'leaks' branch on top of commit 680db1ef560f
> > (leaking_addresses: fix typo function not called).
> > 
> > 
> >  scripts/leaking_addresses.pl | 169 +++++++++++++++++++++++++++++++++++++------
> >  1 file changed, 148 insertions(+), 21 deletions(-)
> > 
> > diff --git a/scripts/leaking_addresses.pl b/scripts/leaking_addresses.pl
> > index 2d5336b3e1ea..6b015980d117 100755
> > --- a/scripts/leaking_addresses.pl
> > +++ b/scripts/leaking_addresses.pl
> > @@ -24,6 +24,7 @@ use Cwd 'abs_path';
> >  use Term::ANSIColor qw(:constants);
> >  use Getopt::Long qw(:config no_auto_abbrev);
> >  use Config;
> > +use feature 'state';
> >  
> >  my $P = $0;
> >  my $V = '0.01';
> > @@ -37,18 +38,20 @@ my $TIMEOUT = 10;
> >  # Script can only grep for kernel addresses on the following architectures. If
> >  # your architecture is not listed here and has a grep'able kernel address please
> >  # consider submitting a patch.
> > -my @SUPPORTED_ARCHITECTURES = ('x86_64', 'ppc64');
> > +my @SUPPORTED_ARCHITECTURES = ('x86_64', 'ppc64',  'i[3456]86');
> >  
> >  # Command line options.
> >  my $help = 0;
> >  my $debug = 0;
> > -my $raw = 0;
> > -my $output_raw = "";	# Write raw results to file.
> > -my $input_raw = "";	# Read raw results from file instead of scanning.
> > +my $raw = 0;                   # Show raw output.
> > +my $output_raw = "";           # Write raw results to file.
> > +my $input_raw = "";            # Read raw results from file instead of scanning.
> > +my $suppress_dmesg = 0;	       # Don't show dmesg in output.
> > +my $squash_by_path = 0;	       # Summary report grouped by absolute path.
> > +my $squash_by_filename = 0;    # Summary report grouped by filename.
> >  
> > -my $suppress_dmesg = 0;		# Don't show dmesg in output.
> > -my $squash_by_path = 0;		# Summary report grouped by absolute path.
> > -my $squash_by_filename = 0;	# Summary report grouped by filename.
> > +my $page_offset_32bit = 0;     # 32-bit: value of CONFIG_PAGE_OFFSET
> > +my $kernel_config_file = "";   # Kernel configuration file.
> >  
> >  # Do not parse these files (absolute path).
> >  my @skip_parse_files_abs = ('/proc/kmsg',
> > @@ -97,14 +100,16 @@ Version: $V
> >  
> >  Options:
> >  
> > -	-o, --output-raw=<file>  Save results for future processing.
> > -	-i, --input-raw=<file>   Read results from file instead of scanning.
> > -	    --raw                Show raw results (default).
> > -	    --suppress-dmesg     Do not show dmesg results.
> > -	    --squash-by-path     Show one result per unique path.
> > -	    --squash-by-filename Show one result per unique filename.
> > -	-d, --debug              Display debugging output.
> > -	-h, --help, --version    Display this help and exit.
> > +	-o, --output-raw=<file>         Save results for future processing.
> > +	-i, --input-raw=<file>          Read results from file instead of scanning.
> > +		--raw                       Show raw results (default).
> > +		--suppress-dmesg            Do not show dmesg results.
> > +		--squash-by-path            Show one result per unique path.
> > +		--squash-by-filename        Show one result per unique filename.
> > +	--page-offset-32bit=<hex>       PAGE_OFFSET value (for 32-bit kernels).
> > +	--kernel-config-file=<file>     Kernel configuration file (e.g /boot/config)
> > +	-d, --debug                     Display debugging output.
> > +	-h, --help, --version           Display this help and exit.
> >  
> >  Examples:
> >  
> > @@ -117,7 +122,10 @@ Examples:
> >  	# View summary report.
> >  	$0 --input-raw scan.out --squash-by-filename
> >  
> > -Scans the running (64 bit) kernel for potential leaking addresses.
> > +	# Scan kernel on a 32-bit system with a 2GB:2GB virtual address split.
> > +	$0 --page-offset-32bit=0x80000000
> > +
> > +Scans the running kernel for potential leaking addresses.
> >  
> >  EOM
> >  	exit($exitcode);
> > @@ -133,6 +141,8 @@ GetOptions(
> >  	'squash-by-path'        => \$squash_by_path,
> >  	'squash-by-filename'    => \$squash_by_filename,
> >  	'raw'                   => \$raw,
> > +	'page-offset-32bit=o'   => \$page_offset_32bit,
> > +	'kernel-config-file=s'  => \$kernel_config_file,
> >  ) or help(1);
> >  
> >  help(0) if ($help);
> > @@ -148,6 +158,7 @@ if (!$input_raw and ($squash_by_path or $squash_by_filename)) {
> >  	exit(128);
> >  }
> >  
> > +show_detected_architecture() if $debug;
> >  if (!is_supported_architecture()) {
> >  	printf "\nScript does not support your architecture, sorry.\n";
> >  	printf "\nCurrently we support: \n\n";
> > @@ -179,7 +190,7 @@ sub dprint
> >  
> >  sub is_supported_architecture
> >  {
> > -	return (is_x86_64() or is_ppc64());
> > +	return (is_x86_64() or is_ppc64() or is_ix86_32());
> >  }
> >  
> >  sub is_x86_64
> > @@ -202,10 +213,40 @@ sub is_ppc64
> >  	return 0;
> >  }
> >  
> > +sub is_ix86_32
> > +{
> > +	my $archname = $Config{archname};
> > +
> > +	if ($archname =~ m/i[3456]86-linux/) {
> > +		return 1;
> > +	}
> > +	return 0;
> > +}
> > +
> > +sub show_detected_architecture
> > +{
> > +	printf "Detected architecture: ";
> > +	if (is_ix86_32()) {
> > +		printf "32 bit x86\n";
> > +	} elsif (is_x86_64()) {
> > +		printf "x86_64\n";
> > +	} elsif (is_ppc64()) {
> > +		printf "ppc64\n";
> > +	} else {
> > +		printf "failed to detect architecture\n"
> > +	}
> > +}
> > +
> >  sub is_false_positive
> >  {
> >  	my ($match) = @_;
> >  
> > +	if (is_ix86_32()) {
> > +		return is_false_positive_ix86_32($match);
> > +	}
> > +
> > +	# 64 bit architectures
> > +
> >  	if ($match =~ '\b(0x)?(f|F){16}\b' or
> >  	    $match =~ '\b(0x)?0{16}\b') {
> >  		return 1;
> > @@ -222,6 +263,89 @@ sub is_false_positive
> >  	return 0;
> >  }
> >  
> > +sub is_false_positive_ix86_32
> > +{
> > +	my ($match) = @_;
> > +	state $page_offset = get_page_offset(); # only gets called once
> 
> nit: new line here

Will do
> > +	if ($match =~ '\b(0x)?(f|F){8}\b') {
> > +		return 1;
> > +	}
> > +
> > +	my $addr32 = eval hex($match);
> > +	if ($addr32 < $page_offset) {
> > +		return 1;
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> > +sub get_page_offset
> > +{
> > +	my $page_offset;
> > +	my $default_offset = hex("0xc0000000");
> > +	my @config_files;
> 
> 	my $tmp_file = "";
> 
> See comments below for reasoning.
> 
> > +	# Allow --page-offset-32bit to override.
> > +	if ($page_offset_32bit != 0) {
> > +		return $page_offset_32bit;
> > +	}
> > +
> > +	# Allow --kernel-config-file to override.
> > +	if ($kernel_config_file ne "") {
> > +		@config_files = ($kernel_config_file);
> > +	} else {
> > +		my $config_file = '/boot/config-' . `uname -r`;
> > +		@config_files = ($config_file, '/boot/config');
> > +	}
> > +
> > +	if (-R "/proc/config.gz") {
> > +		my $tmp_file = "/tmp/tmpkconf";
> 
> 		$tmp_file = "/tmp/tmpkconf";
> 
> > +		if (system("gunzip < /proc/config.gz > $tmp_file")) {
> > +			dprint " parse_kernel_config: system(gunzip...) failed\n";
> > +		} else {
> > +			$page_offset = parse_kernel_config_file($tmp_file);
> > +			if ($page_offset ne "") {
> > +				return hex($page_offset);
> > +			}
> > +		}
> > +		system("rm -f $tmp_file");
> > +	}
> 
> The logic is a bit broken here. sub returns without rm'ing tmp file. 

Caught! :-)  Thanks..

> I
> believe we discussed using
> 
> 	@config_files = ($tmp_file);
> 
> Then continuing to iterate @config_files as done. 

I thought, why not just do this:

	if (-R "/proc/config.gz") {
		my $tmp_file = "/tmp/tmpkconf";
		if (system("gunzip < /proc/config.gz > $tmp_file")) {
			dprint " parse_kernel_config: system(gunzip...) failed\n";
		} else {
			$page_offset = parse_kernel_config_file($tmp_file);
			if ($page_offset ne "") {
				system("rm -f $tmp_file");
				return hex($page_offset);
			}
			system("rm -f $tmp_file");
		}
	}

Also, this way, the '$tmp_file' var remains localized to the handling of
the /proc/config.gz file 'if' statement scope.

> > +
> > +	foreach my $config_file (@config_files) {
> > +		chomp $config_file;
> > +		$page_offset = parse_kernel_config_file($config_file);
> > +		if ($page_offset ne "") {
> > +			return hex($page_offset);
> > +		}
> > +	}
> 
> We may need to use 'last' instead of returning so we can check for
> 
> 	if ($tmp_file ne "") {
> 		system("rm -f $tmp_file");
> 	}

Not required, if we use the manner I propose above..
> 
> And one final (particularly trivial) nitpick
> 
> Can you use the brief commit log with prefix
> 
> 	leaking_addresses:
>  
> please. That prefix is what is currently used. Using 'scripts:' makes it
> hard to fit a descriptive message within 52 characters.
> 
Understood, sorry for the current patch series not using this style.

> I know we have changed it already, but perhaps it should mention x86 not
> just 32 bit (since it is not 32 bit generic).
> 
> I realized while reviewing your code that there is no reason for this to
> be x86 specific, if we can get a config file with CONFIG_PAGE_OFFSET
> then we can scan the kernel like this irrespective of architecture. Perl
> doesn't manage to correctly identify the RaspberryPi I tried it on as 32
> bit so we may not be able to do it how we currently are.
Interesting..
> 
> I'm mentioning this because I don't want you to go to all this work and
> then remove a bunch of your code immediately while making it 32 bit
> generic. If you want to work on a generic version then I'm happy to work
> with you on it.
Sure, lets try for a generic ver! Thanks for your help on this..
As your experience woth the R Pi shows, we may have to just resort to building a
generic framework of sorts, letting folks "plugin" appropriate "truth values"
for their particular platform; this way, we support as much as we can for now
and, going forward, it's generic.
As of right now though, am unsure what this "generic framework" is..

>  If you would prefer to just get this done and merged
> then we can do that too.
> 
> As I've said before I'm new to the maintainer role so still learning how
> best to approach things. Thanks for your patience.
IMO, you're doing just great (me, am not so sure :) ).

Thanks,
Kaiwan.
> Hope this helps,
> Tobin.

Powered by blists - more mailing lists