[strongSwan-dev] Issue in strongswan v4.4.1

[Martin Willi]

Hi,

> We have observed one issue in strong swan that two different SA have
> been assigned with same reqid.

Hm, while it very unlikely, it seems that we currently access that
counter variable non-atomic. This may, in theory, result in two
identical reqids if two threads create CHILD_SAs simultaneously.

Hard to say if this is actually the issue that lead to these non-unqiue
reqids. Anyway, we should fix that, and I've done so with the three
patches at [1].

Regards
Martin

[1]http://git.strongswan.org/?p=strongswan.git;a=shortlog;h=refs/heads/unique-sas