git@vger.kernel.org list mirror (unofficial, one of many)
 help / color / mirror / code / Atom feed
From: Jeff King <peff@peff.net>
To: git@vger.kernel.org
Subject: pack corruption post-mortem
Date: Wed, 16 Oct 2013 04:34:01 -0400	[thread overview]
Message-ID: <20131016083400.GA31266@sigill.intra.peff.net> (raw)

I was recently presented with a repository with a corrupted packfile,
and was asked if the data was recoverable. This post-mortem describes
the steps I took to investigate and fix the problem. I thought others
might find the process interesting, and it might help somebody in the
same situation.

I started with an fsck, which found a problem with exactly one object
(I've used $pack and $obj below to keep the output readable, and also
because I'll refer to them later):

    $ git fsck
    error: $pack SHA1 checksum mismatch
    error: index CRC mismatch for object $obj from $pack at offset 51653873
    error: inflate: data stream error (incorrect data check)
    error: cannot unpack $obj from $pack at offset 51653873

The pack checksum failing means a byte is munged somewhere, and it is
presumably in the object mentioned (since both the index checksum and
zlib were failing).

Reading the zlib source code, I found that "incorrect data check" means
that the adler-32 checksum at the end of the zlib data did not match the
inflated data. So stepping the data through zlib would not help, as it
did not fail until the very end, when we realize the crc does not match.
The problematic bytes could be anywhere in the object data.

The first thing I did was pull the broken data out of the packfile. I
needed to know how big the object was, which I found out with:

  $ git show-index <$idx | cut -d' ' -f1 | sort -n | grep -A1 51653873
  51653873
  51664736

Show-index gives us the list of objects and their offsets. We throw away
everything but the offsets, and then sort them so that our interesting
offset (which we got from the fsck output above) is followed immediately
by the offset of the next object. Now we know that the object data is
10863 bytes long, and we can grab it with:

  dd if=$pack of=object bs=1 skip=51653873 count=10863

I inspected a hexdump of the data, looking for any obvious bogosity
(e.g., a 4K run of zeroes would be a good sign of filesystem
corruption). But everything looked pretty reasonable.

Note that the "object" file isn't fit for feeding straight to zlib; it
has the git packed object header, which is variable-length. We want to
strip that off so we can start playing with the zlib data directly. You
can either work your way through it manually (the format is described in
Documentation/technical/pack-format.txt), or you can walk through it in
a debugger. I did the latter, creating a valid pack like:

  # pack magic and version
  printf 'PACK\0\0\0\2' >tmp.pack
  # pack has one object
  printf '\0\0\0\1' >>tmp.pack
  # now add our object data
  cat object >>tmp.pack
  # and then append the pack trailer
  /path/to/git.git/test-sha1 -b <tmp.pack >trailer
  cat trailer >>tmp.pack

and then running "git index-pack tmp.pack" in the debugger (stop at
unpack_raw_entry). Doing this, I found that there were 3 bytes of header
(and the header itself had a sane type and size). So I stripped those
off with:

  dd if=object of=zlib bs=1 skip=3

I ran the result through zlib's inflate using a custom C program. And
while it did report the error, I did get the right number of output
bytes (i.e., it matched git's size header that we decoded above). But
feeding the result back to "git hash-object" didn't produce the same
sha1. So there were some wrong bytes, but I didn't know which. The file
happened to be C source code, so I hoped I could notice something
obviously wrong with it, but I didn't. I even got it to compile!

I also tried comparing it to other versions of the same path in the
repository, hoping that there would be some part of the diff that didn't
make sense. Unfortunately, this happened to be the only revision of this
particular file in the repository, so I had nothing to compare against.

So I took a different approach. Working under the guess that the
corruption was limited to a single byte, I wrote a program to munge each
byte individually, and try inflating the result. Since the object was
only 10K compressed, that worked out to about 2.5M attempts, which took
a few minutes.

The program I used is here:

-- >8 --
#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <signal.h>
#include <zlib.h>

static int try_zlib(unsigned char *buf, int len)
{
  /* make this absurdly large so we don't have to loop */
  static unsigned char out[1024*1024];
  z_stream z;
  int ret;

  memset(&z, 0, sizeof(z));
  inflateInit(&z);

  z.next_in = buf;
  z.avail_in = len;
  z.next_out = out;
  z.avail_out = sizeof(out);

  ret = inflate(&z, 0);
  inflateEnd(&z);
  return ret >= 0;
}

/* eye candy */
static int counter = 0;
static void progress(int sig)
{
  fprintf(stderr, "\r%d", counter);
  alarm(1);
}

int main(void)
{
  /* oversized so we can read the whole buffer in */
  unsigned char buf[1024*1024];
  int len;
  unsigned i, j;

  signal(SIGALRM, progress);
  alarm(1);

  len = read(0, buf, sizeof(buf));
  for (i = 0; i < len; i++) {
    unsigned char c = buf[i];
    for (j = 0; j <= 0xff; j++) {
      buf[i] = j;

      counter++;
      if (try_zlib(buf, len))
        printf("i=%d, j=%x\n", i, j);
    }
    buf[i] = c;
  }

  alarm(0);
  fprintf(stderr, "\n");
  return 0;
}
-- >8 --

I compiled and ran with:

  gcc -Wall -Werror -O3 munge.c -o munge -lz
  ./munge <zlib

There were a few false positives early on (if you write "no data" in the
zlib header, zlib thinks it's just fine :) ). But I got a hit about
halfway through:

  i=5642, j=c7

I let it run to completion, and got a few more hits at the end (where it
was munging the crc to match our broken data). So there was a good
chance this middle hit was the source of the problem.

I confirmed by tweaking the byte in a hex editor, zlib inflating the
result (no errors!), and then piping the output into "git hash-object",
which reported the sha1 of the broken object. Success!

I fixed the packfile itself with:

  chmod +w $pack
  printf '\xc7' | dd of=$pack bs=1 seek=51659518 conv=notrunc
  chmod -w $pack

The '\xc7' comes from the replacement byte our "munge" program found.
The offset 51659518 is derived by taking the original object offset
(51653873), adding the replacement offset found by "munge" (5642), and
then adding back in the 3 bytes of git header we stripped.

After that, "git fsck" ran clean.

As for the corruption itself, I was lucky that it was indeed a single
byte. In fact, it turned out to be a single bit. The byte 0xc7 was
corrupted to 0xc5. So presumably it was caused by faulty hardware, or a
cosmic ray.

And the aborted attempt to look at the inflated output to see what was
wrong? I could have looked forever and never found it. Here's the diff
between what the corrupted data inflates to, versus the real data:

  -       cp = strtok (arg, "+");
  +       cp = strtok (arg, ".");

It tweaked one byte and still ended up as valid, readable C that just
happened to do something totally different! One takeaway is that on a
less unlucky day, looking at the zlib output might have actually been
helpful, as most random changes would actually break the C code.

But more importantly, git's hashing and checksumming noticed a problem
that easily could have gone undetected in another system. The result
still compiled, but would have caused an interesting bug (that would
have been blamed on some random commit).

-Peff

             reply	other threads:[~2013-10-16  8:34 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-10-16  8:34 Jeff King [this message]
2013-10-16  8:59 ` Duy Nguyen
2013-10-16 15:41 ` Martin Fick
2013-10-17  0:35   ` Jeff King
2013-10-17 15:47     ` Junio C Hamano
2013-10-25  7:55       ` Jeff King
2013-10-17  1:06   ` Duy Nguyen
2013-10-19 10:32 ` Duy Nguyen
2013-10-19 14:41   ` Nicolas Pitre
2013-10-19 19:17     ` Shawn Pearce
2013-10-20 20:56       ` Nicolas Pitre
2013-10-20  4:44     ` Duy Nguyen
2013-10-20 21:08       ` Nicolas Pitre
2015-04-01 21:08 ` [PATCH] howto: document more tools for recovery corruption Jeff King
2015-04-01 22:21   ` Junio C Hamano
2015-04-02  0:49     ` Jeff King

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: http://vger.kernel.org/majordomo-info.html

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20131016083400.GA31266@sigill.intra.peff.net \
    --to=peff@peff.net \
    --cc=git@vger.kernel.org \
    --subject='Re: pack corruption post-mortem' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

git@vger.kernel.org list mirror (unofficial, one of many)

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://public-inbox.org/git
	git clone --mirror http://ou63pmih66umazou.onion/git
	git clone --mirror http://czquwvybam4bgbro.onion/git
	git clone --mirror http://hjrcffqmbrq6wope.onion/git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V1 git git/ https://public-inbox.org/git \
		git@vger.kernel.org
	public-inbox-index git

Example config snippet for mirrors.
Newsgroups are available over NNTP:
	nntp://news.public-inbox.org/inbox.comp.version-control.git
	nntp://7fh6tueqddpjyxjmgtdiueylzoqt6pt7hec3pukyptlmohoowvhde4yd.onion/inbox.comp.version-control.git
	nntp://ie5yzdi7fg72h7s4sdcztq5evakq23rdt33mfyfcddc5u3ndnw24ogqd.onion/inbox.comp.version-control.git
	nntp://4uok3hntl7oi7b4uf4rtfwefqeexfzil2w6kgk2jn5z2f764irre7byd.onion/inbox.comp.version-control.git
	nntp://news.gmane.io/gmane.comp.version-control.git
 note: .onion URLs require Tor: https://www.torproject.org/

code repositories for project(s) associated with this inbox:

	https://80x24.org/mirrors/git.git

AGPL code for this site: git clone https://public-inbox.org/public-inbox.git