From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from kirsty.vergenet.net ([202.4.237.240]:42605 "EHLO kirsty.vergenet.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751537AbdEIJ2Q (ORCPT ); Tue, 9 May 2017 05:28:16 -0400 Date: Tue, 9 May 2017 11:28:01 +0200 From: Simon Horman To: Arnd Bergmann Cc: Wolfram Sang , Ulf Hansson , Magnus Damm , linux-mmc@vger.kernel.org, Linux-Renesas Subject: Re: [PATCH/RFC 0/5] mmc: renesas-sdhi: refactor DMA support Message-ID: <20170509092801.GA8417@verge.net.au> References: <1493723743-22821-1-git-send-email-horms+renesas@verge.net.au> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: linux-renesas-soc-owner@vger.kernel.org List-ID: On Tue, May 02, 2017 at 02:03:06PM +0200, Arnd Bergmann wrote: > On Tue, May 2, 2017 at 1:15 PM, Simon Horman wrote: > > Hi Wolfram, Hi Arnd, Hi all, > > > > the intention of this patch-set is to refactor the DMA support in > > the Renesas SDHI driver in order to make it easier to add support > > for using the SDHI hardware with different DMA implementations. > > > > This is based on earlier work, posted as "[PATCH/RFC v3 0/6] mmc: > > renesas_sdhi: add R-Car Gen-3 DMA support". It attempts to implement > > the reworking of the driver proposed by Arnd[1] in his review of that > > patch-set. > > > > [1] http://www.spinics.net/lists/linux-mmc/msg38004.html > > > > Unlike that patch-set this patch-set does not add support for > > R-Car Gen-3 DMA. Rather it focuses on refactoring the code. > > Looks fine overall. Remind me why we are not using the dma-engine > framework though. IIRC this is because at least one of the possible > DMA implementations is tightly integrated with the TMIO register > set and not easily split out into another driver, right? In a nutshell, yes, that is my understanding. The DMA implementation that is currently present in mainline uses the SYS-DMAC device and the DMA Engine framework is used. However, we expect to support devices that make use of built-in bus master engine for SDHI DMA. In this case it does not seem to make much sense to break out DMA support into a separate driver as it can only be used as part of SDHI which it is a part of. This is as opposed to SYS-DMAC where the SYS-DMAC device may be used by several different devices, f.e. SDHI and SCIF (serial).