Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <47772b9ae10be6dbe5b0cfcd0bc14efc8fb22c0c.camel@linux.ibm.com>
Subject: Re: [PATCH v2 0/7] Implement inline static calls on PPC32 - v2
From:   Benjamin Gray <bgray@linux.ibm.com>
To:     Christophe Leroy <christophe.leroy@csgroup.eu>,
        Ard Biesheuvel <ardb@kernel.org>
Cc:     X86 ML <x86@kernel.org>, Peter Zijlstra <peterz@infradead.org>,
        Chen Zhongjin <chenzhongjin@huawei.com>,
        Dave Hansen <dave.hansen@linux.intel.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        Nicholas Piggin <npiggin@gmail.com>,
        Jason Baron <jbaron@akamai.com>,
        Ingo Molnar <mingo@redhat.com>,
        "sv@linux.ibm.com" <sv@linux.ibm.com>,
        "Steven Rostedt (VMware)" <rostedt@goodmis.org>,
        "H. Peter Anvin" <hpa@zytor.com>, Borislav Petkov <bp@alien8.de>,
        Thomas Gleixner <tglx@linutronix.de>,
        "agust@denx.de" <agust@denx.de>,
        "open list:LINUX FOR POWERPC (32-BIT AND 64-BIT)" 
        <linuxppc-dev@lists.ozlabs.org>,
        Josh Poimboeuf <jpoimboe@kernel.org>
Date:   Thu, 08 Sep 2022 10:13:57 +1000
In-Reply-To: <d35a2039-1755-b0be-6733-bb7ec19b2ea8@csgroup.eu>
References: <cover.1657301423.git.christophe.leroy@csgroup.eu>
         <CAMj1kXFqs=YAbTDJOgzpse9ZkggSxPNNJJphEA=J94FQzF55qg@mail.gmail.com>
         <d35a2039-1755-b0be-6733-bb7ec19b2ea8@csgroup.eu>
Content-Type: multipart/signed; micalg="pgp-sha512";
        protocol="application/pgp-signature"; boundary="=-WWbX6z7r0y1yL9TfB4uB"
User-Agent: Evolution 3.44.4 (3.44.4-1.fc36) 
MIME-Version: 1.0
Precedence: bulk


--=-WWbX6z7r0y1yL9TfB4uB
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, 2022-09-01 at 16:46 +0000, Christophe Leroy wrote:
> Surprisingly, I get worst performance with inline static call than
> with=20
> out of line static call:

I'm not sure what hackbench is doing, but when microbenchmarking 64 bit
out-of-line calls in a loop I saw a similar thing where adding more
indirection improved the performance despite doing more work. The cause
seemed to be a combination of using older hardware and the target being
too short (just an integer increment). Moving to a newer machine and
adding a lot of NOPs to the target made the performance make sense.

--=-WWbX6z7r0y1yL9TfB4uB
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: This is a digitally signed message part

-----BEGIN PGP SIGNATURE-----

iHUEABYKAB0WIQQZBU/uXBOXyOcthtB2V5JPzcRbYwUCYxkzxQAKCRB2V5JPzcRb
Y5ANAQCtq5gqxsuqsnj4PNcdhDxdIwjJ1xWdzadoauMrpaeTkAEA+PFvFtRbb4pe
33Eb8TrRhen+BbD9crA5+UmCQX2ZKAM=
=TVgb
-----END PGP SIGNATURE-----

--=-WWbX6z7r0y1yL9TfB4uB--