Horton

Re: POC: make mxidoff 64 bits

Maxim Orlov <orlovmg@gmail.com>

View thread
From: Maxim Orlov <orlovmg@gmail.com>
To: Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>
Cc: wenhui qiu <qiuwenhuifx@gmail.com>, Heikki Linnakangas <hlinnaka@iki.fi>, Alexander Korotkov <aekorotkov@gmail.com>, Postgres hackers <pgsql-hackers@lists.postgresql.org>
Date: 2025-05-27T15:18:26Z
Lists: pgsql-hackers

Commits

Same data as JSON: GET /api/v1/messages/:b64id/commits the thread's linked commits as JSON, with link sources. API reference →

  1. Fix partial read handling in pg_upgrade's multixact conversion

  2. Increase timeout in multixid_conversion upgrade test

  3. Improve sanity checks on multixid members length

  4. Clarify comment on multixid offset wraparound check

  5. Never store 0 as the nextMXact

  6. Add runtime checks for bogus multixact offsets

  7. Widen MultiXactOffset to 64 bits

  8. Move pg_multixact SLRU page format definitions to a separate header

  9. Convert confusing macros in multixact.c to static inline functions

  10. Index SLRUs by 64-bit integers rather than by 32-bit integers

  11. Cope with possible failure of the oldest MultiXact to exist.

Attachments

On Tue, 29 Apr 2025 at 15:01, Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>
wrote:

>
> I notice that transam/README does not mention multixact except one place
> in section "Transaction Emulation during Recovery". I expected it to
> document how pg_multixact/members and pg_multixact/offset are used and what
> is their layout. It's not the responsibility of this patchset to document
> it, but it will be good if we add a section about multixacts in
> transam/README. It will make reviews easier.
>

Yeah, I agree, this is a big overlook, I think. Anyone who tries to
understand how pg_multixact works has to deal with the code.
Certainly, we need to address this issue.

But for now, here is a new rebase @ 70a13c528b6e382a381f.
The only change is that following commits 15a79c7 and a0ed19e, we must also
switch to PRIu64 format.

-- 
Best regards,
Maxim Orlov.