Re: Can we remove extra memset in BloomInitPage, GinInitPage and SpGistInitPage when we have it in PageInit? - Mailing list pgsql-hackers

From Pavel Borisov
Subject Re: Can we remove extra memset in BloomInitPage, GinInitPage and SpGistInitPage when we have it in PageInit?
Date
Msg-id CALT9ZEGMOGWQuGON1z-Xx9FTS4RU05Jgf+qbL2CoPrQpqXEx6w@mail.gmail.com
Whole thread Raw
In response to Re: Can we remove extra memset in BloomInitPage, GinInitPage and SpGistInitPage when we have it in PageInit?  (Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>)
List pgsql-hackers
ср, 7 апр. 2021 г. в 10:18, Bharath Rupireddy <bharath.rupireddyforpostgres@gmail.com>:
On Wed, Apr 7, 2021 at 11:44 AM Michael Paquier <michael@paquier.xyz> wrote:
>
> On Wed, Apr 07, 2021 at 06:31:19AM +0530, Bharath Rupireddy wrote:
> > Setting p->pd_flags = 0; is unnecessary and redundant after memsetting
> > the page to zeros. Also, see the existing code for pd_prune_xid,
> > similarly I've done that for pd_flags. I think it's okay with /*
> > p->pd_flags = 0;        done by above MemSet */.
>
> Sure, but this one does not hurt much either as-is, so I have left it
> out, and applied the rest.

Thanks for pushing the patch.

I wanted to comment out p->pd_flags = 0; in PageInit similar to the
pd_prune_xid just for consistency.
    /* p->pd_prune_xid = InvalidTransactionId;        done by above MemSet */

With Regards,
Bharath Rupireddy.
EnterpriseDB: http://www.enterprisedb.com



I've investigated the commit, and I think there is just one more thing that can make the page init more even. I propose my very small patch on this in another discussion branch: 
https://www.postgresql.org/message-id/CALT9ZEFFq2-n5Lmfg59L6Hm3ZrgCexyhR9eqme7v1jodtXGg1A@mail.gmail.com

If you want, feel free to discuss it and push if consider the change relevant.

--
Best regards,
Pavel Borisov

Postgres Professional: http://postgrespro.com

pgsql-hackers by date:

Previous
From: Pavel Borisov
Date:
Subject: [PATCH] Improve treatment of page special and page header alignment during page init.
Next
From: "Euler Taveira"
Date:
Subject: Re: Any objections to implementing LogicalDecodeMessageCB for pgoutput?