[PATCH] Improve treatment of page special and page header alignment during page init.

From: Pavel Borisov <pashkin(dot)elfe(at)gmail(dot)com>
To: Postgres hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org>
Subject: [PATCH] Improve treatment of page special and page header alignment during page init.
Date: 2021-04-07 12:02:18
Message-ID: CALT9ZEFFq2-n5Lmfg59L6Hm3ZrgCexyhR9eqme7v1jodtXGg1A@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I was looking at changes in Sp-Gist by
commit 4c0239cb7a7775e3183cb575e62703d71bf3302d
(discussion
https://postgr.es/m/CALj2ACViOo2qyaPT7krWm4LRyRTw9kOXt+g6PfNmYuGA=YHj9A@mail.gmail.com
) and realized that during PageInit, both page header and page special are
expected to be maxaligned but in reality, their treatment is quite
different:
1. page special size is silently enforced to be maxaligned by PageInit()
even if caller-specified specialSize is not of a maxalign'ed size.
2. page header size alignment is not checked at all (but we expect it
maxalign'ed, yes).

I'd propose do both things in the same way: just Assert both sizes are
maxalign'ed during page init.

I dived further and it appears that the only caller, who provides not
properly aligned page special is fill_seq_with_data() and corrected it.

I am really convinced, that _callers_ should care about proper special
size. So now PageInit() just checks the right lengths of page special and
page header with assert, not enforcing size change silently. PFA my small
patch on this. I'd propose it to commit if in the HEAD only likewise the
commit 4c0239cb7a7775e3183cb575e62703d71bf3302d.

What do you think?

--
Best regards,
Pavel Borisov

Postgres Professional: http://postgrespro.com <http://www.postgrespro.com>

Attachment Content-Type Size
v1-0001-Ensure-same-treatment-of-page-header-and-page-spe.patch application/octet-stream 2.4 KB

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Pavel Borisov 2021-04-07 12:08:51 Re: Can we remove extra memset in BloomInitPage, GinInitPage and SpGistInitPage when we have it in PageInit?
Previous Message Magnus Hagander 2021-04-07 12:01:42 Re: hba.c:3160:18: warning: comparison of unsigned enum expression