From mboxrd@z Thu Jan 1 00:00:00 1970 From: Sergei Shtylyov Subject: Re: [PATCH] Palmchip BK3710 IDE driver Date: Wed, 23 Jan 2008 16:38:06 +0300 Message-ID: <4797433E.4040707@ru.mvista.com> References: <200801212144.29511.asalnikov@ru.mvista.com> <4796526F.1090103@ru.mvista.com> <20080122222207.18221939@lxorguk.ukuu.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from h155.mvista.com ([63.81.120.155]:8109 "EHLO imap.sh.mvista.com" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1751346AbYAWNhQ (ORCPT ); Wed, 23 Jan 2008 08:37:16 -0500 In-Reply-To: <20080122222207.18221939@lxorguk.ukuu.org.uk> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Anton Salnikov Cc: Alan Cox , linux-ide@vger.kernel.org, bzolnier@gmail.com Hello. Alan Cox wrote: >> Why you chose to use ioread32() and iowrite32() if your device is strictly >>memory mapped? Those functions add some overhead, and boil down to readl() and > There are distinct portability advantages but you shouldn't mix > ioread32/iowrite32 with ioremap as that isn't guaranteed to work. Hm, pci_iomap() boils down to ioremap(), so > readl/writel does fine and fixes up the driver. Which brings up another issue with the latest patch: Anton, where do you call ioremap() or something alike to get the virtual address of the registers? MBR, Sergei