Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-16 Thread Moritz Muehlenhoff
Juan Gallego wrote:
 is sarge affected by this vulnerability? or has sarge been archived and i 
 missed the announcement?

The main attack vector - pygrub/xen - doesn't exist in Sarge. The other attacks
are more or less theoretical and hardly justify modifications to an important
core package like this.

Cheers,
Moritz


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-14 Thread Nico Golde
Hi Steve,
* Steve Kemp [EMAIL PROTECTED] [2007-12-07 20:26]:
 On Fri Dec 07, 2007 at 18:41:35 +0100, Nico Golde wrote:
  What about those, are they unimportant?
  They are still present in the etch code. I stumbled
  upon them while preparing a testing-security upload.
 
   Uknown.  I used the patch provided by Theodore Tso, which he
  is/was planning on using for Sid/Ubuntu.
 
   If there are missing bits then we'll need to reissue the update,
  but right now I believed the patch was as complete as it needed
  to be.
[...] 
I asked Ted about this, I just quote what he wrote:
I don't consider that to be a high
priority issue, since it's not likely that an attacker would 
be able to trick an administrator to run resize2fs on some random 
filesystem image while running as root.

So decide on your own if this warrants an update of the DSA, 
he will include this in 1.40.4.
Kind regards
Nico

-- 
Nico Golde - http://www.ngolde.de - [EMAIL PROTECTED] - GPG: 0x73647CFF
For security reasons, all text in this mail is double-rot13 encrypted.


pgpccPMXo3kax.pgp
Description: PGP signature


Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-08 Thread Nico Golde
Hi Steve,
* Steve Kemp [EMAIL PROTECTED] [2007-12-07 20:26]:
 On Fri Dec 07, 2007 at 18:41:35 +0100, Nico Golde wrote:
 
  What about those, are they unimportant?
  They are still present in the etch code. I stumbled
  upon them while preparing a testing-security upload.
 
   Uknown.  I used the patch provided by Theodore Tso, which he
  is/was planning on using for Sid/Ubuntu.

Oh ok.

   If there are missing bits then we'll need to reissue the update,
  but right now I believed the patch was as complete as it needed
  to be.

Ok, I am waiting for his reply, I attached my patch to the 
bug report in unstable. From what I see every multiplication 
with fs-blocksize needs to be checked, all of these are 
coming from the file system. Let's see what he does :)

http://people.debian.org/~nion/nmu-diff/e2fsprogs-1.40.2-1_1.40.2-1+lenny1.patch
YFYI this is the patch I used for testing-security.
Cheers
Nico
-- 
Nico Golde - http://www.ngolde.de - [EMAIL PROTECTED] - GPG: 0x73647CFF
For security reasons, all text in this mail is double-rot13 encrypted.


pgp7nInBmG6BX.pgp
Description: PGP signature


Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-07 Thread Juan Gallego
On 2007-12-07 13:21-, Steve Kemp [EMAIL PROTECTED] wrote:

| For the stable distribution (etch), this problem has been fixed in version
| 1.39+1.40-WIP-2006.11.14+dfsg-2etch1.

| For the unstable distribution (sid), this problem will be fixed shortly.

hi Steve and others,

is sarge affected by this vulnerability? or has sarge been archived and i 
missed the announcement?

tia,
-- 
juan


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-07 Thread Steve Kemp
On Fri Dec 07, 2007 at 09:46:21 -0500, Juan Gallego wrote:

 | For the stable distribution (etch), this problem has been fixed in version
 | 1.39+1.40-WIP-2006.11.14+dfsg-2etch1.
 
 | For the unstable distribution (sid), this problem will be fixed shortly.

 is sarge affected by this vulnerability? or has sarge been archived and i 
 missed the announcement?

  Sarge is affected, but I don't yet have a working patch for that.

  There should be an update shortly, but this is pretty low-risk and
 it seemed sensible to release now, rather than waiting.

Steve
-- 


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-07 Thread Nico Golde
Hi,
* Nico Golde [EMAIL PROTECTED] [2007-12-07 18:32]:
[...] 
  Rafal Wojtczuk of McAfee AVERT Research discovered that e2fsprogs,
  ext2 file system utilities and libraries, contained multiple
  integer overflows in memory allocations, based on sizes taken directly
  from filesystem information.  These could result in heap-based
  overflows potentially allowing the execution of arbitrary code.
  
  For the stable distribution (etch), this problem has been fixed in version
  1.39+1.40-WIP-2006.11.14+dfsg-2etch1.
 [...] 
 e2fsck/swapfs.c:retval = ext2fs_get_mem(fs-blocksize * 
 fs-inode_blocks_per_group,
 resize/resize2fs.c: retval = ext2fs_get_mem(fs-blocksize * 
 fs-inode_blocks_per_group,
 resize/resize2fs.c: retval = ext2fs_get_mem(fs-blocksize *
 resize/resize2fs.c: retval = ext2fs_get_mem(rfs-old_fs-blocksize * 3, 
 block_buf);
 resize/extent.c:retval = ext2fs_get_mem(sizeof(struct 
 ext2_extent_entry) *
 
 What about those, are they unimportant? They are still present in the etch 
 code. I stumbled
 upon them while preparing a testing-security upload.

Sorry, this mail was originally only addressed to Steve but 
since I also got this mail through the debian-security list 
it ended up here now :)
Anyway, I looked again into these and from my point of view 
the released DSA is incomplete, I fixed those for 
testing-security by using get_mem_array as well.
Kind regards
Nico
-- 
Nico Golde - http://www.ngolde.de - [EMAIL PROTECTED] - GPG: 0x73647CFF
For security reasons, all text in this mail is double-rot13 encrypted.


pgpzvamUYiL2l.pgp
Description: PGP signature


Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-07 Thread Steve Kemp
On Fri Dec 07, 2007 at 18:41:35 +0100, Nico Golde wrote:

 What about those, are they unimportant?
 They are still present in the etch code. I stumbled
 upon them while preparing a testing-security upload.

  Uknown.  I used the patch provided by Theodore Tso, which he
 is/was planning on using for Sid/Ubuntu.

  If there are missing bits then we'll need to reissue the update,
 but right now I believed the patch was as complete as it needed
 to be.

 Sorry, this mail was originally only addressed to Steve but 
 since I also got this mail through the debian-security list 
 it ended up here now :)

  Fair enough.

Steve
-- 
# The Debian Security Audit Project.
http://www.debian.org/security/audit


-- 
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]



Re: [SECURITY] [DSA 1422-1] New e2fsprogs packages fix arbitrary code execution

2007-12-07 Thread Nico Golde
Hi Steve,
* Steve Kemp [EMAIL PROTECTED] [2007-12-07 14:32]:
 
 Debian Security Advisory DSA-1422[EMAIL PROTECTED]
 http://www.debian.org/security/   Steve Kemp
 December 07, 2007 http://www.debian.org/security/faq
 
 
 Package: e2fsprogs
 Vulnerability  : integer overfows
 Problem type   : local
 Debian-specific: no
 CVE Id(s)  : CVE-2007-5497
 
 Rafal Wojtczuk of McAfee AVERT Research discovered that e2fsprogs,
 ext2 file system utilities and libraries, contained multiple
 integer overflows in memory allocations, based on sizes taken directly
 from filesystem information.  These could result in heap-based
 overflows potentially allowing the execution of arbitrary code.
 
 For the stable distribution (etch), this problem has been fixed in version
 1.39+1.40-WIP-2006.11.14+dfsg-2etch1.
[...] 
e2fsck/swapfs.c:retval = ext2fs_get_mem(fs-blocksize * 
fs-inode_blocks_per_group,
resize/resize2fs.c: retval = ext2fs_get_mem(fs-blocksize * 
fs-inode_blocks_per_group,
resize/resize2fs.c: retval = ext2fs_get_mem(fs-blocksize *
resize/resize2fs.c: retval = ext2fs_get_mem(rfs-old_fs-blocksize * 3, 
block_buf);
resize/extent.c:retval = ext2fs_get_mem(sizeof(struct 
ext2_extent_entry) *

What about those, are they unimportant? They are still present in the etch 
code. I stumbled
upon them while preparing a testing-security upload.
Kind regards
Nico

-- 
Nico Golde - http://www.ngolde.de - [EMAIL PROTECTED] - GPG: 0x73647CFF
For security reasons, all text in this mail is double-rot13 encrypted.


pgp3qie05eItu.pgp
Description: PGP signature