From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ben Hutchings <bhutchings@solarflare.com>
Subject: Re: [PATCH net-next-2.6] tg3: 64bits stats
Date: Mon, 05 Jul 2010 19:41:03 +0100
Message-ID: <1278355263.2087.116.camel@achroite.uk.solarflarecom.com>
References: <1278321242.2877.18.camel@edumazet-laptop>
	 <1278345780.2877.376.camel@edumazet-laptop>
	 <1278351070.2087.70.camel@achroite.uk.solarflarecom.com>
	 <1278353230.2877.601.camel@edumazet-laptop>
Mime-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE
Cc: Matt Carlson <mcarlson@broadcom.com>,
	Michael Chan <mchan@broadcom.com>,
	netdev <netdev@vger.kernel.org>,
	David Miller <davem@davemloft.net>
To: Eric Dumazet <eric.dumazet@gmail.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from exchange.solarflare.com ([216.237.3.220]:2319 "EHLO
	exchange.solarflare.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1751557Ab0GESlI convert rfc822-to-8bit (ORCPT
	<rfc822;netdev@vger.kernel.org>); Mon, 5 Jul 2010 14:41:08 -0400
In-Reply-To: <1278353230.2877.601.camel@edumazet-laptop>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>

On Mon, 2010-07-05 at 20:07 +0200, Eric Dumazet wrote:
> Le lundi 05 juillet 2010 =C3=A0 18:31 +0100, Ben Hutchings a =C3=A9cr=
it :
> > On Mon, 2010-07-05 at 18:03 +0200, Eric Dumazet wrote:
> > > Le lundi 05 juillet 2010 =C3=A0 11:14 +0200, Eric Dumazet a =C3=A9=
crit :
> > > > After commit be1f3c2c027c (net: Enable 64-bit net device statis=
tics on
> > > > 32-bit architectures), we can now provide 64bit stats, even on =
32bit
> > > > arches.
> > > >=20
> > > > Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
> > > > ---
> > >=20
> > > Please disregard this patch.
> > >=20
> > > There is small possibility a reader might read a 64bit value whil=
e
> > > another writer makes a change to it, changing high 32bit value.
> > >=20
> > > A change in core network would be needed to make this 100% safe,
> > > possibly using a seqlock to protect dev->stats64
> >=20
> > I really didn't want to add this overhead and complication to reade=
rs
> > when only some drivers need it.
> >=20
>=20
> Overhead would be minimal, only in get_stats() method and only for
> drivers that want to provide 64 bit stats...
>=20
> Clearly, rx/tx path must not have any overhead.
>=20
> Right now, even RTNL 64bit stats are not safe.

Ugh, I was assuming that callers of dev_get_stats() would hold
dev_base_lock.  However only netstat_show() seems to do so currently.

Ben.

> Should we revert prior patches or try to make progress ?
>=20
>=20
>=20
--=20
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.