From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTP id 973C2DDEE9 for ; Thu, 14 Jun 2007 17:40:08 +1000 (EST) In-Reply-To: <46709A84.7030307@am.sony.com> References: <20070612181825.730300780@am.sony.com>> <466EEC01.6080807@am.sony.com> <46709A84.7030307@am.sony.com> Mime-Version: 1.0 (Apple Message framework v623) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: <6ef2a5debaf78e8f2fa7de8b72653203@kernel.crashing.org> From: Segher Boessenkool Subject: Re: [patch 29/30] PS3: Device tree source. Date: Thu, 14 Jun 2007 09:39:59 +0200 To: Geoff Levand Cc: linuxppc-dev@ozlabs.org, paulus@samba.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , > I will use model = "SonyPS3"; Okie. >>> + compatible = "PS3"; >> >> name = "sony,ps3" > > > As Ben H. said, the top level 'name' cannot be set, Bug in the flat device tree stuff. > and > I found dtc exits with an error if you include one. > > The compatible entry is generally used to test for machine > type, so I think that needs to remain, and I will use > compatible = "sony,ps3". "compatible" is a set of alternates for "name", so you can use it if you have to. >>> + * Need to keep linux,platform for a while, not used by kernel. >>> + */ >> >> Why that? > > Newer dtc seems to be OK with this removed, old ones exited with error. Ok. So you'll remove it? >>> + CBE,PPE { >> >> PowerPC,CBE@0 or cpu@0 -- the important thing is you need >> the @0 unit address. > > OK, I set it up as cpu@0, since the actual type seems to be probed for. PowerPC CPUs have a defined way of identifying themselves exactly, and this PVR value is supposed to be in the device tree too (although Linux probes it directly from the CPU). You could still put the name in the "compatible" property, but CPU nodes are a bit special, I don't think that would gain anything. >>> + d-cache-size = <8000>; >> >> I thought DL1 was 64kB, I might be wrong though? > > The BE handbook says 32. Ok, good. Segher