From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Andres Freund <andres(at)anarazel(dot)de> |
Cc: | Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, David Rowley <david(dot)rowley(at)2ndquadrant(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>, Amit Khandekar <amitdkhan(dot)pg(at)gmail(dot)com>, Ashutosh Bapat <ashutosh(dot)bapat(at)enterprisedb(dot)com>, Robert Haas <robertmhaas(at)gmail(dot)com> |
Subject: | Re: Atomics for heap_parallelscan_nextpage() |
Date: | 2017-08-16 17:40:09 |
Message-ID: | 30938.1502905209@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
I wrote:
> I can confirm that on dromedary, that regression test case is attempting
> to create a TOC with a not-well-aligned size: 93268 = 0x16c54 bytes.
... although, on closer look, it still seems like we have a fundamental
bit of schizophrenia here, because on this machine
$ grep ALIGN pg_config.h
#define ALIGNOF_DOUBLE 4
#define ALIGNOF_INT 4
#define ALIGNOF_LONG 4
#define ALIGNOF_LONG_LONG_INT 4
#define ALIGNOF_SHORT 2
#define MAXIMUM_ALIGNOF 4
Basically, therefore, ISTM that it is not a good thing that the atomics
code thinks it can rely on 8-byte-aligned data when the entire rest of
the system believes that 4-byte alignment is enough for anything.
I was wondering why the shm_toc code was using BUFFERALIGN and not
MAXALIGN, and I now suspect that the answer is "it's an entirely
undocumented kluge to make the atomics code not crash on 32-bit
machines, so long as nobody puts a pg_atomic_uint64 anywhere except
in a shm_toc".
I'm not sure that that's good enough, and I'm damn sure that it
shouldn't be undocumented.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Tom Lane | 2017-08-16 17:44:28 | Re: Atomics for heap_parallelscan_nextpage() |
Previous Message | Tom Lane | 2017-08-16 17:32:44 | Re: Atomics for heap_parallelscan_nextpage() |