[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAC_iWjJqZA4itFNyf1Czur8ACZ8S+9JhSsgjstC2Pho8V4xZDA@mail.gmail.com>
Date: Fri, 1 Mar 2024 12:40:15 +0200
From: Ilias Apalodimas <ilias.apalodimas@...aro.org>
To: Jesper Dangaard Brouer <hawk@...nel.org>
Cc: Lorenzo Bianconi <lorenzo@...nel.org>, Jakub Kicinski <kuba@...nel.org>, netdev@...r.kernel.org,
lorenzo.bianconi@...hat.com, davem@...emloft.net, edumazet@...gle.com,
pabeni@...hat.com
Subject: Re: [PATCH net-next] net: page_pool: make page_pool_create inline
Hi Jesper,
On Fri, 1 Mar 2024 at 12:33, Jesper Dangaard Brouer <hawk@...nel.org> wrote:
>
>
>
> On 01/03/2024 11.27, Ilias Apalodimas wrote:
> >
> > On Wed, 14 Feb 2024 at 21:14, Lorenzo Bianconi <lorenzo@...nel.org> wrote:
> >>
> >>> On Wed, 14 Feb 2024 19:01:28 +0100 Lorenzo Bianconi wrote:
> >>>> Make page_pool_create utility routine inline a remove exported symbol.
> >>>
> >>> But why? If you add the kdoc back the LoC saved will be 1.
> >>
> >> I would remove the symbol exported since it is just a wrapper for
> >> page_pool_create_percpu()
> >
> > I don't mind either. But the explanation above must be part of the
> > commit message.
>
> I hope my benchmark module will still work...
>
> https://github.com/netoptimizer/prototype-kernel/blob/master/kernel/lib/bench_page_pool_simple.c#L181
Since we rely on this so much, perhaps we should make an effort to
merge it, even in a version that only works for x86.
I know it's fairly easy to compile and run locally, but having it as
part of the CI is better -- more people will be able to look at the
results quickly and determine if we have a performance degradation
Cheers
/Ilias
>
>
> --Jesper
Powered by blists - more mailing lists