| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Andres Freund <andres(at)anarazel(dot)de> |
| Cc: | pgsql-hackers(at)postgresql(dot)org, Petr Jelinek <petr(at)2ndquadrant(dot)com>, Peter Eisentraut <peter_e(at)gmx(dot)net> |
| Subject: | Re: libpqrcv_PQexec() seems to violate latch protocol |
| Date: | 2017-06-06 21:14:59 |
| Message-ID: | 1236.1496783699@sss.pgh.pa.us |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Andres Freund <andres(at)anarazel(dot)de> writes:
> The function in $subject does:
> ResetLatch(&MyProc->procLatch);
> rc = WaitLatchOrSocket(&MyProc->procLatch,
> WL_POSTMASTER_DEATH | WL_SOCKET_READABLE |
> WL_LATCH_SET,
> PQsocket(streamConn),
> 0,
> WAIT_EVENT_LIBPQWALRECEIVER);
Yeah, this is certainly broken.
> Afaict, the ResetLatch() really should just instead be in the if (rc & WL_LATCH_SET) block.
And, to be specific, it should be before the CHECK_FOR_INTERRUPTS call,
since that is the useful work that we want to be sure occurs after
any latch-setting event.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Andres Freund | 2017-06-06 21:17:00 | Re: libpqrcv_PQexec() seems to violate latch protocol |
| Previous Message | Andres Freund | 2017-06-06 21:04:05 | libpqrcv_PQexec() seems to violate latch protocol |