On Thu, Mar 15, 2012 at 10:33 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Daniel Farina <daniel@heroku.com> writes:
>> The way MyCancelKey is checked now is backwards, in my mind. It seems
>> like it would be better checked by the receiving PID (one can use a
>> check/recheck also, if so inclined). Is there a large caveat to that?
>
> You mean, other than the fact that kill(2) can't transmit such a key?
I was planning on using an out-of-line mechanism. Bad idea?
> But actually I don't see what you hope to gain from such a change,
> even if it can be made to work. Anyone who can do kill(SIGINT) can
> do kill(SIGKILL), say --- so you have to be able to trust the signal
> sender. What's the point of not trusting it to verify the client
> identity?
No longer true with pg_cancel_backend not-by-superuser, no? Now there
are new people who can do kill(SIGINT) (modulus the already existing
cancel requests).
--
fdr
