Horton

Re: Allow non-superuser to cancel superuser tasks.

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Nathan Bossart <nathandbossart@gmail.com>
Cc: Alexander Kukushkin <cyberdemn@gmail.com>, "Andrey M. Borodin" <x4mmm@yandex-team.ru>, Michael Paquier <michael@paquier.xyz>, Kirill Reshke <reshkekirill@gmail.com>, "Leung, Anthony" <antholeu@amazon.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2024-11-26T17:27:33Z
Lists: pgsql-hackers 
Hi,

On 2024-11-22 20:44:34 -0600, Nathan Bossart wrote:
> On Fri, Nov 22, 2024 at 06:13:16PM -0500, Andres Freund wrote:
> >> -		if (procStatus && procStatus->st_backendType == B_AUTOVAC_WORKER)
> >> +		if (pgstat_get_backend_type(pid) == B_AUTOVAC_WORKER)
> > 
> > Because we already mapped the pid to a ProcNumber, it'd be cheaper to access
> > the backend status via procnumber.
> 
> D'oh, I missed that ProcNumber could be used as an index for the
> BackendStatusArray.  Is the attached more like what you are imagining?

Yes.

I'd probably add two function header comments:

1) explicit caution that this is fetching information not from the snapshot
   but "live" data
2) the return value might be out of date, that the procnumber needs to be
   valid and that the caller is responsible for permission checking

I'd also add a comment do the code saying that it's fine to bypass the
changecount mechanism, because we're reading a single 4 byte integer.

Greetings,

Andres Freund

Commits

  1. Look up backend type in pg_signal_backend() more cheaply.

  2. Add tap test for pg_signal_autovacuum role

  3. Introduce pg_signal_autovacuum_worker.

  4. Add a slot synchronization function.

  5. Ban role pg_signal_backend from more superuser backend types.