Re: [SECURITY] [DLA 1283-1] python-crypto security update

Hi Brian

This is what I think we should do.

1) Send a new DLA telling that the fix is only partial and not complete and in addtion that elgamal encryption is not supported by the library and should not be used.

2) Mark the CVE as no-dsa/ignored in the security database.

Suggested DLA text. Any opinion about this text? I'm not a native English speaker so my language may not be the best.

Package : python-crypto
Version : 2.6-4+deb7u8
CVE ID : CVE-2018-6594
Debian Bug : 889999

This is an update to DLA-1283-1. In DLA-1283-1 it is claimed that the issue

described in CVE-2018-6594 is fixed. It turns out that the fix is partial and

upstream has decided not to fix the issue as it would break compatibility and

that ElGamal encryption was not intended to work on its own.

The recommendation is still to upgrade python-crypto packages but in addition

to that consider that the fix is not complete. If you application using python-crypto

is implementing ElGamal encryption you should consider changing to some

other encryption method.

There will be no further update to python-crypto for this specific CVE. A fix would

break compatibility, the problem has been ignored by regular Debian Security team

due to its minor nature and in addition to that we are close to the end of life of the

Wheezy security support.

CVE-2018-6594:

python-crypto generated weak ElGamal key parameters, which allowed attackers to
obtain sensitive information by reading ciphertext data (i.e., it did not have
semantic security in face of a ciphertext-only attack).

For Debian 7 "Wheezy", the problem has been partially fixed in version
2.6-4+deb7u8.

We recommend that you upgrade your python-crypto packages.

Further information about Debian LTS security advisories, how to apply
these updates to your system and frequently asked questions can be
found at: https://wiki.debian.org/LTS

On 3 April 2018 at 00:02, Brian May <bam@debian.org> wrote:

Ola Lundqvist <ola@inguza.com> writes:

> Do we have a fix that solve the problem? If we do we can simply upload a
> new version with the fix and describe it accordingly.
> If it is fixed in some cases it may be considered fixed.
>
> I have not checked the details about this specific problem.

There are no known fixes.

Upstream argues that a fix is unnecessarily, because the key
vulnerability is only a problem when using the encryption, which is not
"is not meant to be used by itself". Furthermore they say fixing this
will break "backward compatibility with PGP".

For full details, read the upstream bug report:
https://github.com/Legrandin/pycryptodome/issues/90

Or, put another way, upstream renamed the "encrypt" function to
"_encrypt" a while back to indicate this should not be used. I think it
really depends on your point of view if this is sufficient to fix a
security vulnerability in key creation (another function) when used
directly with encryption.
--
Brian May <bam@debian.org>

--- Inguza Technology AB --- MSc in Information Technology ----

/ ola@inguza.com Folkebogatan 26 \

| opal@debian.org 654 68 KARLSTAD |

| http://inguza.com/ Mobile: +46 (0)70-332 1551 |

\ gpg/f.p.: 7090 A92B 18FE 7994 0C36 4FE4 18A1 B1CF 0FE5 3DD9 /

---------------------------------------------------------------