Horton

Re: Issue in GIN fast-insert: XLogBeginInsert + Read/LockBuffer ordering

Zhang Mingli <zmlpostgres@gmail.com>

View thread View on postgr.es
From: Zhang Mingli <zmlpostgres@gmail.com>
To: Michael Paquier <michael@paquier.xyz>
Cc: PostgreSQL Hackers <pgsql-hackers@postgresql.org>, Matthias van de Meent <boekewurm+postgres@gmail.com>
Date: 2022-10-12T13:39:11Z
Lists: pgsql-hackers

Attachments

HI,

On Oct 12, 2022, 10:09 +0800, Michael Paquier <michael@paquier.xyz>, wrote:
>
> Nice catches, both of you. Let's adjust everything spotted in one
> shot. Matthias' patch makes the ordering right, but the
> initialization path a bit harder to follow when using a separate list.
> The code is short so it does not strike me as a big deal, and it comes
> from the fact that we need to lock an existing buffer when merging two
> lists. For the branch where we insert into a tail page, the pages are
> already locked but it looks to be enough to move XLogBeginInsert()
> before the first XLogRegisterBuffer() call.
>
> Would any of you like to write a patch?
> --
> Michael
Patch added.

Regards,
Zhang Mingli

Commits

  1. Fix ordering issue with WAL operations in GIN fast insert path