From mboxrd@z Thu Jan  1 00:00:00 1970
From: Vincent Bernat <vincent@bernat.im>
Subject: Re: [PATCH net-next v1] ipvs: add consistent source hashing scheduling
Date: Mon, 02 Apr 2018 19:57:12 +0200
Message-ID: <m3zi2lcwzr.fsf@luffy.cx>
References: <20180402172025.7380-1-vincent@bernat.im>
        <63c63fab-12ed-8b15-e00e-dc54e5c144eb@gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Cc: Wensong Zhang <wensong@linux-vs.org>,
        Simon Horman <horms@verge.net.au>,
        Julian Anastasov <ja@ssi.bg>,
        "David S. Miller" <davem@davemloft.net>, netdev@vger.kernel.org,
        lvs-devel@vger.kernel.org
To: Eric Dumazet <eric.dumazet@gmail.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from bart.luffy.cx ([78.47.78.131]:50532 "EHLO bart.luffy.cx"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1752920AbeDBR5P (ORCPT <rfc822;netdev@vger.kernel.org>);
        Mon, 2 Apr 2018 13:57:15 -0400
In-Reply-To: <63c63fab-12ed-8b15-e00e-dc54e5c144eb@gmail.com> (Eric Dumazet's
        message of "Mon, 2 Apr 2018 10:33:27 -0700")
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

 =E2=9D=A6  2 avril 2018 10:33 -0700, Eric Dumazet <eric.dumazet@gmail.com>=
=C2=A0:

>> +static inline u32
>> +ip_vs_csh_permutation(struct ip_vs_dest *d, int j)
>> +{
>> +	u32 offset, skip;
>> +	__be32 addr_fold =3D d->addr.ip;
>> +
>> +#ifdef CONFIG_IP_VS_IPV6
>> +	if (d->af =3D=3D AF_INET6)
>> +		addr_fold =3D d->addr.ip6[0]^d->addr.ip6[1]^
>> +			d->addr.ip6[2]^d->addr.ip6[3];
>> +#endif
>> +	addr_fold =3D ntohl(addr_fold) + ntohs(d->port);
>> +	offset =3D hash_32(addr_fold, 32) % IP_VS_CSH_TAB_SIZE;
>> +	skip =3D (hash_32(addr_fold + 1, 32) % (IP_VS_CSH_TAB_SIZE - 1)) + 1;
>> +	return (offset + j * skip) % IP_VS_CSH_TAB_SIZE;
>> +}
>> +
>
> This does not look very strong to me, particularly the IPv6 folding
>
> I would rather use __ipv6_addr_jhash() instead of ipv6_addr_hash(),
> even if it is hard coded ;)

I can switch to ipv6_addr_hash(). However, switching to
__ipv6_addr_jhash seems useless as I would need to hardcode the initial
value: people use source hashing to get the same result from one host to
another. Am I missing something?
--=20
Each module should do one thing well.
            - The Elements of Programming Style (Kernighan & Plauger)