From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wolfram Sang Subject: Re: [RFC] i2c: Providing hooks for i2c multimaster bus arbitration. Date: Tue, 12 Feb 2013 11:03:35 +0100 Message-ID: <20130212100335.GA5490@nekote.pengutronix.de> References: <1359968595-21291-1-git-send-email-yuvaraj.cd@samsung.com> <20130210181710.GB12760@nekote.pengutronix.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-i2c-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Simon Glass Cc: Doug Anderson , Wolfram Sang , Yuvaraj Kumar C D , Yuvaraj Kumar C D , linux-samsung-soc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Grant Grundler , Mark Brown , "linux-i2c-u79uwXL29TY76Z2rM5mHXA@public.gmane.org" , naveen krishna , "linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org" List-Id: linux-i2c@vger.kernel.org Hi, > > At the moment perhaps the strongest argument for why this particular > > arbitration scheme is needed is that's what the EC (embedded > > controller) on the ARM Chromebook uses. There have been several > > arguments in-house about whether this was the most ideal way to > > structure things, but that's what we shipped with. Thus, if we want > > to be able to talk to i2c devices on this bus (contains the keyboard, > > battery, and a bunch of power switches) we need something in the > > system that implements this arbitration scheme. > > > > There was some discussion about using the built-in I2C multi-master > support but no one had had a good experience getting it to work > reliably in a real system so we ended up deciding not to rely on this. So, were there no experiences at all or bad experiences when trying? I am interested in hearing what I2C multi master restrictions might exist in practice. Thanks, Wolfram