Re: XMAX_LOCK_ONLY and XMAX_COMMITTED (fk/multixact code) - Mailing list pgsql-hackers

From Bossart, Nathan
Subject Re: XMAX_LOCK_ONLY and XMAX_COMMITTED (fk/multixact code)
Date
Msg-id 239E70B4-8896-4807-9463-ECD1FF71A560@amazon.com
Whole thread Raw
In response to Re: XMAX_LOCK_ONLY and XMAX_COMMITTED (fk/multixact code)  ("Bossart, Nathan" <bossartn@amazon.com>)
List pgsql-hackers
On 11/30/21, 4:54 PM, "Bossart, Nathan" <bossartn@amazon.com> wrote:
> v2 attached.

I accidentally left a redundant check in v2, so here is a v3 without
it.

My proposed patch adds a few checks for the unsupported bit patterns
in the visibility code, but it is far from exhaustive.  I'm wondering
if it might be better just to add a function or macro that everything
exported from heapam_visibility.c is expected to call.  My guess is
the main argument against that would be the possible performance
impact.

Nathan


Attachment

pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: Lots of memory allocated when reassigning Large Objects
Next
From: "Bossart, Nathan"
Date:
Subject: Re: Correct error message for end-of-recovery record TLI