There's a cve request pending for a buffer overflow in the ARC2 key handling, it's described in this test case:
A patch is available here:
Mike Wiacek <email@example.com> is credited with finding this bug. No further detail is available and i'm afraid there's no packaged release yet.
herd, can you include this patch in our distribution?
I'm quite unsure about the status here. If that's exploitable, it seems a user can pass a overly long key to ARC2 and can write arbitrary memory with its content.
As pycrypto may be pulled as a PDEPEND of portage, i set this to "A1". If you think this is wrong, please correct me.
Buffer overflow in the PyCrypto ARC2 module 2.0.1 allows remote
attackers to cause a denial of service and possibly execute arbitrary
code via a large ARC2 key length.
ping, python herd. upstream committed a patch 4 weeks ago. Is there anything holding this back from being fixed in our tree?
dev-python/pycrypto-2.0.1-r8 in CVS now with suggested patch. I'm adding arches to this bug so they are aware of this and act accordingly. I'm also keeping this bug open.
Created attachment 183837 [details]
Unittest for ARC2 Buffer Overflow in CVE-2009-0544
This test case is a modified version of the one at securityfocus.com, so it runs on all python versions available in the tree.
Sparc stable for pycrypto-2.0.1-r8. All tests run fine.
Stable for HPPA.