From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753494AbaIBK1R (ORCPT ); Tue, 2 Sep 2014 06:27:17 -0400 Received: from mxout5.netvision.net.il ([194.90.6.65]:40534 "EHLO mxout5.netvision.net.il" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752341AbaIBK1Q (ORCPT ); Tue, 2 Sep 2014 06:27:16 -0400 MIME-version: 1.0 Content-transfer-encoding: 7BIT Content-type: text/plain; CHARSET=US-ASCII; format=flowed Message-id: <54059AAD.2060905@gmail.com> Date: Tue, 02 Sep 2014 13:23:41 +0300 From: Eli Billauer User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.12) Gecko/20100907 Fedora/3.0.7-1.fc12 Thunderbird/3.0.7 To: Arend van Spriel Cc: Dan Carpenter , gregkh@linuxfoundation.org, arnd@arndb.de, devel@driverdev.osuosl.org, linux-kernel@vger.kernel.org Subject: Re: [PATCH] staging: xillybus: Move out of staging References: <1409471821-11809-1-git-send-email-eli.billauer@gmail.com> <20140901121318.GG5122@mwanda> <54046FDA.8060905@broadcom.com> In-reply-to: <54046FDA.8060905@broadcom.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hello, Arend. On 01/09/14 16:08, Arend van Spriel wrote: > Maybe it would be better to use the DMA-API instead of the PCI wrappers. I just found out that the PCI wrappers were indeed removed from the DMA-API.txt documentation back in 2010, saying that "driver writers are always able to use the DMA API with any bus": https://lkml.org/lkml/2010/3/8/470 On the other hand, I was under the impression that the PCI API is for (future?) platforms which may need a distinction between PCI and non-PCI. That is why the Xillybus driver does some acrobatics, so that the PCI API is used for PCI devices, and the general DMA API for non-PCI devices. Given that the distinction is already there, is it wise to remove it, and use the plain DMA API for all? That will simplify the code, of course, but if there's any esoteric platform that needs a different treatment for PCI devices, it's better to stay as is, I suppose. Thanks, Eli