From mboxrd@z Thu Jan 1 00:00:00 1970 From: Francois Romieu Subject: Re: [PATCH 1/1] More ethtool support for sis900 Date: Sun, 6 Mar 2005 01:15:46 +0100 Message-ID: <20050306001546.GE25116@electric-eye.fr.zoreil.com> References: <20050305134011.23638.68926@localhost.localdomain> <4229FA32.4000401@pobox.com> <02a49476862ae18433e5b80aafa616fd@libero.it> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Jeff Garzik , netdev@oss.sgi.com To: Daniele Venzano Content-Disposition: inline In-Reply-To: <02a49476862ae18433e5b80aafa616fd@libero.it> Sender: netdev-bounce@oss.sgi.com Errors-to: netdev-bounce@oss.sgi.com List-Id: netdev.vger.kernel.org Daniele Venzano : [...] > I saw the locking, but I couldn't come up with a reason for it. Is it > needed because of kernel wide preemption ? Usually you do not want simultaneous accesses to the mii interface (link events, Tx timeout recovery or so). >>From a quick glance at the sis900 driver, I would expect the lock to protect against sis900_timer() (assuming you add a simple spinlock to it as well). -- Ueimor