Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:8 as permitted sender) client-ip=2620:137:e000::3:8;
Date:   Sat, 16 Sep 2023 09:50:36 +0100
From:   Conor Dooley <conor@kernel.org>
To:     Charlie Jenkins <charlie@rivosinc.com>
Cc:     Palmer Dabbelt <palmer@dabbelt.com>,
        Samuel Holland <samuel.holland@sifive.com>,
        David Laight <David.Laight@aculab.com>,
        linux-riscv@lists.infradead.org, linux-kernel@vger.kernel.org,
        linux-arch@vger.kernel.org,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Albert Ou <aou@eecs.berkeley.edu>,
        Arnd Bergmann <arnd@arndb.de>
Subject: Re: [PATCH v6 1/4] asm-generic: Improve csum_fold
Message-ID: <20230916-jackpot-guy-01c2024c6a63@spud>
References: <20230915-optimize_checksum-v6-0-14a6cf61c618@rivosinc.com>
 <20230915-optimize_checksum-v6-1-14a6cf61c618@rivosinc.com>
MIME-Version: 1.0
Content-Type: multipart/signed; micalg=pgp-sha256;
        protocol="application/pgp-signature"; boundary="60Aaxr3gdQT472cE"
Content-Disposition: inline
In-Reply-To: <20230915-optimize_checksum-v6-1-14a6cf61c618@rivosinc.com>
Precedence: bulk


--60Aaxr3gdQT472cE
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Sep 15, 2023 at 10:01:17AM -0700, Charlie Jenkins wrote:
> This csum_fold implementation introduced into arch/arc by Vineet Gupta
> is better than the default implementation on at least arc, x86, and
> riscv. Using GCC trunk and compiling non-inlined version, this
> implementation has 41.6667%, 25% fewer instructions on riscv64, x86-64
> respectively with -O3 optimization. Most implmentations override this
> default in asm, but this should be more performant than all of those
> other implementations except for arm which has barrel shifting and
> sparc32 which has a carry flag.
>=20
> Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
> Reviewed-by: David Laight <david.laight@aculab.com>
> ---
>  include/asm-generic/checksum.h | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>=20
> diff --git a/include/asm-generic/checksum.h b/include/asm-generic/checksu=
m.h
> index 43e18db89c14..37f5ec70ac93 100644
> --- a/include/asm-generic/checksum.h
> +++ b/include/asm-generic/checksum.h
> @@ -31,9 +31,7 @@ extern __sum16 ip_fast_csum(const void *iph, unsigned i=
nt ihl);
>  static inline __sum16 csum_fold(__wsum csum)
>  {
>  	u32 sum =3D (__force u32)csum;
> -	sum =3D (sum & 0xffff) + (sum >> 16);
> -	sum =3D (sum & 0xffff) + (sum >> 16);
> -	return (__force __sum16)~sum;
> +	return (__force __sum16)((~sum - ror32(sum, 16)) >> 16);

Breaks the build on RISC-V in a way that is repaired by later patches in
the series, so you likely did not notice:

=2E/include/asm-generic/checksum.h:34:35: error: call to undeclared functio=
n 'ror32'; ISO C99 and later do not support implicit function declarations =
[-Wimplicit-function-declaration]
=2E./include/linux/bitops.h:134:21: error: static declaration of 'ror32' fo=
llows non-static declaration

Cheers,
Conor.

--60Aaxr3gdQT472cE
Content-Type: application/pgp-signature; name="signature.asc"

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

iHUEABYIAB0WIQRh246EGq/8RLhDjO14tDGHoIJi0gUCZQVsXAAKCRB4tDGHoIJi
0mnEAQCTLnl3iMEBVo3Tzp6o42hU6amG6PduNyUQd8smnnQAVAD/RhsXtPuthkcC
tk+v8xXjpz227mU8sGL65Usl2rfDRgk=
=GX46
-----END PGP SIGNATURE-----

--60Aaxr3gdQT472cE--