Re: Boundary value check in lazy_tid_reaped() - Mailing list pgsql-hackers

From Thomas Munro
Subject Re: Boundary value check in lazy_tid_reaped()
Date
Msg-id CA+hUKG++PXpa1xqAQJCc0x5Z6L3qjdyTr8B1Rp3Q2cp=NYishA@mail.gmail.com
Whole thread Raw
In response to Boundary value check in lazy_tid_reaped()  (Masahiko Sawada <masahiko.sawada@2ndquadrant.com>)
Responses Re: Boundary value check in lazy_tid_reaped()  (Thomas Munro <thomas.munro@gmail.com>)
Re: Boundary value check in lazy_tid_reaped()  (Peter Geoghegan <pg@bowt.ie>)
List pgsql-hackers
On Sun, Aug 30, 2020 at 11:08 PM Masahiko Sawada
<masahiko.sawada@2ndquadrant.com> wrote:
> So my proposal is to add boundary value check in lazy_tid_reaped()
> before executing bsearch(3). This will help when index vacuum happens
> multiple times or when garbage tuples are concentrated to a narrow
> range.

Makes sense if it's often out of range.

> I thought that we can have a generic function wrapping bsearch(3) that
> does boundary value checks and then does bsearch(3) so that we can use
> it in other similar places as well. But the attached patch doesn't do
> that as I'd like to hear opinions on the proposal first.

I wonder if you would also see a speed-up with a bsearch() replacement
that is inlineable, so it can inline the comparator (instead of
calling it through a function pointer).  I wonder if something more
like (lblk << 32 | loff) - (rblk << 32 | roff) would go faster than
the branchy comparator.



pgsql-hackers by date:

Previous
From: Andres Freund
Date:
Subject: Re: Clang UndefinedBehaviorSanitize (Postgres14) Detected undefined-behavior
Next
From: Victor Spirin
Date:
Subject: Re: Sometimes the output to the stdout in Windows disappears