From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753432Ab1HJPNX (ORCPT ); Wed, 10 Aug 2011 11:13:23 -0400 Received: from cam-admin0.cambridge.arm.com ([217.140.96.50]:55167 "EHLO cam-admin0.cambridge.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753168Ab1HJPNV (ORCPT ); Wed, 10 Aug 2011 11:13:21 -0400 Date: Wed, 10 Aug 2011 16:12:56 +0100 From: Will Deacon To: Rob Herring Cc: Arnd Bergmann , Mark Rutland , "linux-arm-kernel@lists.infradead.org" , "devicetree-discuss@lists.ozlabs.org" , "linux-kernel@vger.kernel.org" , "linux@arm.linux.org.uk" , "tglx@linutronix.de" , "weizeng.he@csr.com" , "workgroup.linux@csr.com" , "'Barry Song'" <21cnbao@gmail.com>, "'Grant Likely'" , "'Olof Johansson'" Subject: Re: Subject: L2x0 OF properties do not include interrupt # Message-ID: <20110810151256.GN10121@e102144-lin.cambridge.arm.com> References: <000201cc575b$c1229010$4367b030$@rutland@arm.com> <201108101609.56995.arnd@arndb.de> <4E42965A.9010500@calxeda.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <4E42965A.9010500@calxeda.com> User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Aug 10, 2011 at 03:31:54PM +0100, Rob Herring wrote: > Here's all the interrupts on the PL310: > > DECERRINTR Decode error received on master ports from L3 > SLVERRINTR Slave error received on master ports from L3 > ERRRDINTR Error on L2 data RAM read > ERRRTINTR Error on L2 tag RAM read > ERRWDINTR Error on L2 data RAM write > ERRWTINTR Error on L2 tag RAM write > PARRDINTR Parity error on L2 data RAM read > PARRTINTR Parity error on L2 tag RAM read > ECNTRINTR Event Counter Overflow/Increment > L2CCINTR L2CC Combined Interrupt Output > > It's likely that you would want to split these to different drivers. > EDAC for parity/bus errors and hw-events for event counters, for example. I completely agree with you here. In reality, I expect we'll be lucky if even the combined interrupt is wired up so maybe the individual property names should only be added on an as-needed basis? For the time being we can happily get by with something like combined-interrupt = <...>. Or would you rather specify everything up front? Will