From mboxrd@z Thu Jan 1 00:00:00 1970 From: Benjamin Herrenschmidt Subject: Re: [patch 5/6] ps3: BD/DVD/CD-ROM Storage Driver Date: Tue, 17 Jul 2007 07:38:48 +1000 Message-ID: <1184621929.25235.79.camel@localhost.localdomain> References: <20070704132212.726923000@pademelon.sonytel.be> <1184335333.3402.20.camel@localhost.localdomain> <200707131619.54898.arnd@arndb.de> <1184340528.3402.32.camel@localhost.localdomain> <20070716121649.GI5195@kernel.dk> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from gate.crashing.org ([63.228.1.57]:50358 "EHLO gate.crashing.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755999AbXGPVjf (ORCPT ); Mon, 16 Jul 2007 17:39:35 -0400 In-Reply-To: Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Geert Uytterhoeven Cc: Jens Axboe , James Bottomley , Arnd Bergmann , linux-scsi@vger.kernel.org, Linux Kernel Development , Alessandro Rubini , Linux/PPC Development , Paul Mackerras > Upon closer look, while flush_kernel_dcache_page() is a no-op on ppc64, > flush_dcache_page() isn't. So I'd prefer to not call it if not really needed. > > And according to James, flush_kernel_dcache_page() should be sufficient... > > So I'm getting puzzled again... flush_dcache_page() handles icache vs. dcache coherency by clearing the PG_arch_1 bit in the struct page that indicates that the page is cache clean. You -must- call it if you're going to use any form of CPU access to write to the page (basically dirtying the data cache) and that page can be ever mapped into user space and executed from. I don't know what flush_kernel_dcache_page() does and if it needs a similar treatement, it's a new interface, so maybe Jens and or James can tell me more about it.. Cheers, Ben.