From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Message-ID: <9e4733910710240823v2022661ftb1b754a86cab88f3@mail.gmail.com> Date: Wed, 24 Oct 2007 11:23:40 -0400 From: "Jon Smirl" To: "Grant Likely" Subject: Re: Audio codec device tree entries In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 References: <9e4733910710221859q6ea54810nba58907d5ddd966d@mail.gmail.com> <471E12C7.8020509@freescale.com> <8416ea754e013a67441aec778c81ad73@kernel.crashing.org> <9e4733910710231529h1089eacdy888306f20af92555@mail.gmail.com> <471F52ED.10007@freescale.com> <9e4733910710240800y24952e70g8c318e35e2e45e2e@mail.gmail.com> Cc: PowerPC dev list , Timur Tabi List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 10/24/07, Grant Likely wrote: > > The DTC experts need to tell us which way to make the pointers between > > i2s and i2c for the codec. Here's a another way it could be done that > > looks more like the ac97 model. > > I *really* don't think this is a good idea. Put the node on the bus > that the device is addressed from. I2S is the *data* path, not the > *control* path, but you cannot control the codec from there. > > Your suggestion only looks more like the AC97 model if you're looking > at the data path. If you're looking at the control path it looks > completely different. The device tree convention is to orient around > the control path. It doesn't make any difference to me which one we pick. I'm just listing all of the possible combinations. I just want to pick a model so that we can write the code. I see your point about putting the child node onto the control bus. ac97 is both a control and data bus. For the i2s case the child should go onto the i2c bus. -- Jon Smirl jonsmirl@gmail.com