Re: Locally exploitable races? OpenBSD

[Gregory Steuck <greg@nest.cx>]

>>>>> "Todd" == Todd C Miller <Todd.Miller@courtesan.com> writes:

    Todd> What he's talking about is completely irrelevant because it
    Todd> requires things use rfork() and nothing in the system does...

    Todd> Should the races get fixed?  Yes.  Is it a real problem?
    Todd> Nope.  Should developers respond to obvious flame bait?  No
    Todd> again.

If I am not mistaken it's a stability, not security issue. The problem
is real in sense that it gives a local user a way to destabilize
(crash?)  system. Then again, having seen those nasty free() exploits I
won't be surprised if it can be turned into something dangerous.

Is there an issue with bringing in FreeBSD's fix for sys_dup2?