From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.genesi-usa.com (mithrandir.softwarenexus.net [66.98.186.96]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTP id 9CC2ADDE1B for ; Wed, 13 Jun 2007 21:28:13 +1000 (EST) Received: from 82-46-178-156.cable.ubr06.king.blueyonder.co.uk ([82.46.178.156] helo=[192.168.2.228]) by mail.genesi-usa.com with esmtpa (Exim 4.66 (FreeBSD)) (envelope-from ) id 1HyQmF-000IYA-0p for linuxppc-dev@ozlabs.org; Wed, 13 Jun 2007 11:12:39 +0000 Message-ID: <466FD4EC.80304@genesi-usa.com> Date: Wed, 13 Jun 2007 12:28:44 +0100 From: Matt Sealey MIME-Version: 1.0 To: ppc-dev Subject: Enforcing built-in driver init order? Content-Type: text/plain; charset=UTF-8 List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , I'm trying to work out how to decouple the BestComm SRAM driver from the BestComm DMA engine. SRAM is far more useful than just as a DMA buffer, especially as it can stay active over deep sleep and SDRAM refresh turned off, as a temporary place to put things. Even with a full set of tasks there are still some kilobytes free. In future processors the SRAM will only get larger meaning you can do more things with it, and it may be used for more than DMA. My simple problem is, if I split the two drivers apart, the SRAM platform driver is not guaranteed to load before the DMA platform driver. If it were a module I could have depmod handle it and load them in order but this isn't possible since DMA is required on boot, and SRAM is required for DMA. Any clues or hints on how it can be EASILY enforced that the SRAM platform driver is present (and can be checked for??) before the BestComm engine is turned on and tasks are arbitrarily loaded? -- Matt Sealey Genesi, Manager, Developer Relations