Re: SIGFPE handler is naive

On Tue, Aug 14, 2012 at 6:50 AM, Greg Stark <stark@mit.edu> wrote:
> It is possible to check if the signal was synchronous or was sent from
> an external process. You can check siginfo->si_pid to see who sent you
> the signal. I'm not sure checking that and handling it at
> check_for_interrupts if it's asynchronous is the best solution or not
> though.

If that's portable it might be an option, but I doubt that it is.

> I'm a bit confused. Didn't Tom do the laborious process of checking
> the whole source tree for situations where there's shared memory
> cleanup to be done in and arrange for it to happen? That was the
> blocking factor to get pg_cancel_backend() to work. Is the problem
> that the sigfpe handler doesn't invoke atexit() handlers?

No, the problem is that SIGFPE throws an error *from the signal
handler* rather than waiting for ProcessInterrupts().

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company