| From: | Andres Freund <andres(at)2ndquadrant(dot)com> |
|---|---|
| To: | Florian Pflug <fgp(at)phlo(dot)org> |
| Cc: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, PostgreSQL-development Hackers <pgsql-hackers(at)postgresql(dot)org>, Heikki Linnakangas <hlinnakangas(at)vmware(dot)com> |
| Subject: | Re: Memory ordering issue in LWLockRelease, WakeupWaiters, WALInsertSlotRelease |
| Date: | 2014-02-15 11:11:03 |
| Message-ID: | 20140215111103.GB20973@alap3.anarazel.de |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On 2014-02-15 04:20:17 +0100, Florian Pflug wrote:
> Another idea would be to do as you suggest and only mark the PGPROC pointers
> volatile, but to additionally add a check for queue corruption somewhere. We should
> be able to detect that - if we ever hit this issue, LWLockRelease should find a
> PGPROC while traversing the queue whose lwWaitLink is NULL but which isn't equal to
> lock->tail. If that ever happens, we'd simply PANIC.
My current conclusion is that backporting barriers.h is by far the most
reasonable way to go. The compiler problems have been ironed out by
now...
Arguments against?
Greetings,
Andres Freund
--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andres Freund | 2014-02-15 11:21:14 | Re: Patch: show xid and xmin in pg_stat_activity and pg_stat_replication |
| Previous Message | Hannu Krosing | 2014-02-15 11:02:41 | Re: Ctrl+C from sh can shut down daemonized PostgreSQL cluster |