Horton

Re: POC: make mxidoff 64 bits

wenhui qiu <qiuwenhuifx@gmail.com>

View thread
From: wenhui qiu <qiuwenhuifx@gmail.com>
To: Maxim Orlov <orlovmg@gmail.com>
Cc: Alexander Korotkov <aekorotkov@gmail.com>, Ashutosh Bapat <ashutosh.bapat.oss@gmail.com>, Heikki Linnakangas <hlinnaka@iki.fi>, Postgres hackers <pgsql-hackers@lists.postgresql.org>
Date: 2025-09-16T12:12:17Z
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.

Hi Maxim
  Thanks for your continued efforts to get XID64 implemented.
>   32kB page may contain then 2^13-2 offsets, each is maxed by 2^18+1.
> Therefore, offset from base will never overflow 2^31 and will always
> fit uint32.

> It appears logical to me.
Agree +1 , but I have a question: I remember the XID64 patch got split into
a few threads. How are these threads related? The original one was seen as
too big a change, so it was broken up after people raised concerns.

Thanks

On Mon, Sep 15, 2025 at 11:42 PM Maxim Orlov <orlovmg@gmail.com> wrote:

>
>
> On Sat, 13 Sept 2025 at 16:34, Alexander Korotkov <aekorotkov@gmail.com>
> wrote:
>
>>
>> Therefore, we can change from each 8 of 32-bit multixact offsets
>> (takes 32-bytes) to one 64-bit offset + 7 of 24-bit offset increments
>> (takes 29-bytes).  The actual multixact offsets can be calculated at
>> the fly, overhead shouldn't be significant.  What do you think?
>>
>>
> Thank you for your review; I'm pleased to hear from you again.
>
> Yes, because the maximum number of mxoff is limited by the number of
> running transactions, we may do it that way.
> However, it is a bit wired to have offsets with the 7-byte "base".
>
> I believe we may take advantage of the 64XID patch's notion of putting a
> 8 byte base followed by 4 byte offsets for particular page.
>
> 32kB page may contain then 2^13-2 offsets, each is maxed by 2^18+1.
> Therefore, offset from base will never overflow 2^31 and will always
> fit uint32.
>
> It appears logical to me.
>
>
> --
> Best regards,
> Maxim Orlov.
>