Horton

Re: ResourceOwner refactoring

Andres Freund <andres@anarazel.de>

View thread View on postgr.es
From: Andres Freund <andres@anarazel.de>
To: Heikki Linnakangas <hlinnaka@iki.fi>
Cc: Aleksander Alekseev <aleksander@timescale.com>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>, Julien Rouhaud <rjuju123@gmail.com>, Robert Haas <robertmhaas@gmail.com>
Date: 2023-07-10T19:40:00Z
Lists: pgsql-hackers 
Hi,

On 2023-07-10 12:14:46 -0700, Andres Freund wrote:
> >  /*
> > - * Initially allocated size of a ResourceArray.  Must be power of two since
> > - * we'll use (arraysize - 1) as mask for hashing.
> > + * Size of the fixed-size array to hold most-recently remembered resources.
> >   */
> > -#define RESARRAY_INIT_SIZE 16
> > +#define RESOWNER_ARRAY_SIZE 32
>
> That's 512 bytes, pretty large to be searched sequentially. It's somewhat sad
> that the array needs to include 8 byte of ResourceOwnerFuncs...

On that note:

It's perhaps worth noting that this change actually *increases* the size of
ResourceOwnerData. Previously:

        /* size: 576, cachelines: 9, members: 19 */
        /* sum members: 572, holes: 1, sum holes: 4 */

now:
        /* size: 696, cachelines: 11, members: 13 */
        /* sum members: 693, holes: 1, sum holes: 3 */

It won't make a difference memory-usage wise, given aset.c's rounding
behaviour. But it does mean that more memory needs to be zeroed, as
ResourceOwnerCreate() uses MemoryContextAllocZero.

As there's really no need to initialize the long ->arr, ->locks, it seems
worth to just initialize explicitly instead of using MemoryContextAllocZero().


With the new representation, is there any point in having ->locks? We still
need ->nlocks to be able to switch to the lossy behaviour, but there doesn't
seem to be much point in having the separate array.

Greetings,

Andres Freund

Commits

  1. Make RelationFlushRelation() work without ResourceOwner during abort

  2. Fix bug in bulk extending temp relation after failure

  3. Add missing PGDLLIMPORT markings

  4. Add test_dsa module.

  5. Clear CurrentResourceOwner earlier in CommitTransaction.

  6. Fix dsa.c with different resource owners.

  7. Fix bug in the new ResourceOwner implementation.

  8. Change pgcrypto to use the new ResourceOwner mechanism.

  9. Use a faster hash function in resource owners.

  10. Make ResourceOwners more easily extensible.

  11. Move a few ResourceOwnerEnlarge() calls for safety and clarity.