Originally, CVE-2005-3534: [1] http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2005-3534 has been assigned to NBD and addressed in nbd-v2.8.3 version: [2] http://sourceforge.net/project/shownotes.php?release_id=380202&group_id=13229 via changeset: [3] https://github.com/yoe/nbd/commit/4ed24fe0d64c7cc9963c57b52cad1555ad7c6b60 But nbd-v2.9.0: [4] http://sourceforge.net/projects/nbd/files/nbd/2.9.0/ contains the issue again. This flaw was fixed second time via upstream changeset: [5] https://github.com/yoe/nbd/commit/3ef52043861ab16352d49af89e048ba6339d6df8 References: [6] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=611187 https://bugzilla.redhat.com/show_bug.cgi?id=673562
i think nbd-2.9.20 is fixed, and it's in the tree now
(In reply to comment #1) > i think nbd-2.9.20 is fixed, and it's in the tree now > Great, thanks. Arches, please test and mark stable: =sys-block/nbd-2.9.20 Target keywords : "amd64 ppc ppc64 x86"
amd64 done
I tested on x86 but stopped "make check" after a few hours as it looks like an infinite loop!? However, the server & client seem to work nevertheless.
x86 done. Thanks Andreas. I saw the same behaviour. Could you file a bug for that?
ppc/ppc64 stable
Thank you everyone. GLSA request filed.
(In reply to comment #5) > x86 done. Thanks Andreas. I saw the same behaviour. Could you file a bug for > that? > I started it again and left it running in a screen session, 24 hours later it really finished successful! :-)
CVE-2011-0530 (http://nvd.nist.gov/nvd.cfm?cvename=CVE-2011-0530): Buffer overflow in the mainloop function in nbd-server.c in the server in Network Block Device (nbd) before 2.9.20 might allow remote attackers to execute arbitrary code via a long request. NOTE: this issue exists because of a CVE-2005-3534 regression.
This issue was resolved and addressed in GLSA 201206-35 at http://security.gentoo.org/glsa/glsa-201206-35.xml by GLSA coordinator Stefan Behte (craig).