Re: Bug: XLogReader mishandles oversized multi-page xl_tot_len (potential memory corruption)

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Matthias van de Meent <boekewurm+postgres(at)gmail(dot)com>
Cc: David K <dkarapetyan(at)gmail(dot)com>, pgsql-hackers(at)lists(dot)postgresql(dot)org
Subject: Re: Bug: XLogReader mishandles oversized multi-page xl_tot_len (potential memory corruption)
Date: 2026-08-04 21:59:58
Message-ID: anJg3rHwUKDkwqSO@paquier.xyz
Views: Whole Thread | Raw Message | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Mon, Aug 03, 2026 at 11:23:47AM +0200, Matthias van de Meent wrote:
> On Mon, 3 Aug 2026 at 01:21, Michael Paquier <michael(at)paquier(dot)xyz> wrote:
>> Another thing to note: this basically breaks the recovery test
>> 039_end_of_wal.pl. Could you look at that please?
>
> Also fixed, including a new test case for >XLogRecordMaxSize.

Makes sense to me this way, so applied on HEAD.

I was also wondering for a bit why bae868caf222 has decided to stick
with a 2GB value for the record size in the written record headers,
but as far as I can see it's an arbitrary value that Thomas has picked
up when writing the tests, see the first flavor here:
https://www.postgresql.org/message-id/CA+hUKGLhFdd-G1DCk9Ze3KnQ_2jUkxyXvGwV_Ha=DWMrDnGHng@mail.gmail.com

And the resulting commit has inherited it.
--
Michael

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Melanie Plageman 2026-08-04 22:13:21 Re: Why clearing the VM doesn't require registering vm buffer in wal record
Previous Message Jeff Davis 2026-08-04 21:30:18 Re: CREATE SUBSCRIPTION ... SERVER vs. pg_dump, etc.