From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Lindgren Subject: Re: [rft/rfc/patch-2.6.28-rc3+ 00/59] namespace cleanup patches Date: Tue, 11 Nov 2008 15:58:14 -0800 Message-ID: <20081111235811.GP3984@atomide.com> References: <1226444238-14614-1-git-send-email-me@felipebalbi.com> <20081111232141.GD23454@frodo> <20081111234711.GO3984@atomide.com> <20081111235039.GF23454@frodo> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mho-01-bos.mailhop.org ([63.208.196.178]:63270 "EHLO mho-01-bos.mailhop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753205AbYKKX6R (ORCPT ); Tue, 11 Nov 2008 18:58:17 -0500 Content-Disposition: inline In-Reply-To: <20081111235039.GF23454@frodo> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Felipe Balbi Cc: linux-omap@vger.kernel.org, Felipe Balbi * Felipe Balbi [081111 15:50]: > On Tue, Nov 11, 2008 at 03:47:13PM -0800, Tony Lindgren wrote: > > * Felipe Balbi [081111 15:22]: > > > On Wed, Nov 12, 2008 at 12:56:19AM +0200, Felipe Balbi wrote: > > > > From: Felipe Balbi > > > > > > > > Get rid of omap_{read,write}[bsl] defines. > > > > Virtual and physical addresses usage was fixed by Russel King and > > > > there's no need anymore for keeping a define to a standard kernel > > > > function. Get rid of those to keep the namespace cleaner. > > > > > > > > Build tested with all omap defconfigs. > > > > Depends on my previous "make clk fwk build again on omap1" patch. > > > > > > for convinience, these patches plus the clock fix for omap1 and the mmc > > > fix are available at [1] > > > > Felipe, have you tested any of these? To me looks like you're missing > > the associated ioremap before you can use __raw_read/write instead of > > omap_read/write. > > You're right, I should at least use __raw_read/write(IO_ADDRESS(...)); Adding the IO_ADDRESS there is the wrong fix though, let's rather keep what we have. > I'll fix it by tomorrow and test on omap3. How about we convert them one at a time to use ioremap + __raw_read/write instead? Pretty much the only place where ioremap cannot be used is in sram.c, and that has a comment about it. And clock fwk stuff we should not mess with right now. Cheers, Tony