From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from mx1.redhat.com ([209.132.183.28]:24736 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753765Ab2BGIWh (ORCPT ); Tue, 7 Feb 2012 03:22:37 -0500 Date: Tue, 7 Feb 2012 09:22:33 +0100 From: Stanislaw Gruszka To: wwguy Cc: Intel Linux Wireless , linux-wireless@vger.kernel.org Subject: Re: [RFC 4/4] iwlwifi: use writeb,writel,readl directly Message-ID: <20120207082232.GD2246@redhat.com> (sfid-20120207_092240_998257_3FF8630D) References: <1328544564-8696-1-git-send-email-sgruszka@redhat.com> <1328544564-8696-4-git-send-email-sgruszka@redhat.com> <1328545972.14547.15.camel@wwguy-ubuntu> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <1328545972.14547.15.camel@wwguy-ubuntu> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, Feb 06, 2012 at 08:32:52AM -0800, wwguy wrote: > I believe iwl driver were use readl/writel before switch to > ioread32/iowrite32 (long time ago) > > http://www.gossamer-threads.com/lists/linux/kernel/960145 > > also, AFAIK, the readl/writel IO functions still works but their use in > new code is discouraged. I do not think so, we have tons of drivers that use readl/writel. PCI express devices can not use port-based IO, they will be incredible slow then. I think confusion come from a fact that iwlwifi use pci_ioremp() for mapping registers, which can also maps port-based IO areas. Simply changing that into pci_ioremap_bar(), and make driver depend on CONFIG_HAS_IOMEM will remove that inconsistency. Stanislaw