Excerpts from Greg Smith's message of vie dic 16 11:19:52 -0300 2011:
> On 12/16/2011 08:42 AM, Robert Haas wrote:
> > the proposed patch would potentially result - in the extremely
> > unlikely event of a
> > super-fast PID wraparound - in someone cancelling a query they
> > otherwise wouldn't have been able to cancel.
> >
>
> So how might this get exploited?
>
> -Attach a debugger and put a breakpoint between the check and the kill
If you can attach a debugger to a backend, you already have the
credentials to the user running postmaster, which we assume to be a
"game over" condition.
I guess a more interesting exploitation would be to do this until the
killing backend randomly wins the race condition against a dying backend
and a new one starting up; this would require very rapid reuse of the
PID, plus very rapid startup of the new proces. I think we already
assume that this is not the case in other code paths, even on systems
that randomly (instead of sequentially) assign PIDs.
--
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
