Re: DSA failed to allocate memory

Thomas Munro <thomas.munro@gmail.com>

From: Thomas Munro <thomas.munro@gmail.com>

To: Tom Lane <tgl@sss.pgh.pa.us>

Cc: Dongming Liu <ldming101@gmail.com>, Robert Haas <robertmhaas@gmail.com>, pgsql-hackers <pgsql-hackers@postgresql.org>

Date: 2023-06-14T00:29:54Z

Lists: pgsql-hackers

Attachments

v4-0001-Re-bin-segment-when-memory-pages-are-freed.patch (text/x-patch) patch v4-0001
v4-0002-Add-a-test-module-to-exercise-dsa.c.patch (text/x-patch) patch v4-0002

On Mon, Feb 20, 2023 at 5:52 PM Thomas Munro <thomas.munro@gmail.com> wrote:
> I'm wondering about this bit in rebin_segment():
>
> +       if (segment_map->header == NULL)
> +               return;
>
> Why would we be rebinning an uninitialised/unused segment?

Answering my own question: because destroy_superblock() can do that.
So I think destroy_superblock() should test for that case, not
rebin_segment().  See attached.

Commits

Re-bin segment when memory pages are freed.
- 1605623ec686 11.21 landed
- 74ad9b0d1e7c 12.16 landed
- a0003572f214 13.12 landed
- b7ec66731dd8 14.9 landed
- 9ffb10f183f1 15.4 landed
- af8f9ec66bd1 16.0 landed
- 03f80daac8cf 17.0 landed
De-support floating-point timestamps.
- b6aa17e0ae36 10.0 cited