From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann To: Daniel Mack Subject: Re: [PATCH 00/20] ARM: pxa: move core and drivers to dmaengine Date: Wed, 7 Aug 2013 18:08:51 +0200 References: <1375889649-14638-1-git-send-email-zonque@gmail.com> In-Reply-To: <1375889649-14638-1-git-send-email-zonque@gmail.com> MIME-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Message-Id: <201308071808.51400.arnd@arndb.de> Cc: mark.rutland@arm.com, s.neumann@raumfeld.com, linux-mtd@lists.infradead.org, haojian.zhuang@linaro.org, cxie4@marvell.com, lars@metafoo.de, nico@linaro.org, vinod.koul@intel.com, marek.vasut@gmail.com, ezequiel.garcia@free-electrons.com, rmk+kernel@arm.linux.org.uk, devicetree@vger.kernel.org, samuel@sortiz.org, broonie@kernel.org, mika.westerberg@linux.intel.com, linux-arm-kernel@lists.infradead.org, thomas.petazzoni@free-electrons.com, eric.y.miao@gmail.com, gregkh@linuxfoundation.org, g.liakhovetski@gmx.de, sachin.kamat@linaro.org, kernel@pengutronix.de, djbw@fb.com, davem@davemloft.net List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Wednesday 07 August 2013, Daniel Mack wrote: > I've been working on teaching the mmp-pdma driver more functions and > porting tree-wide scattered pxa specific drivers over to dmaengine > implementations. I posted the first round of patches for mmp-pdma here: > > http://marc.info/?l=linux-arm-kernel&m=137587082530228 > > > With the following patches applied, I can boot a PXA3xx board boot with > all DMA runtime information determined from DT, using pxa3xx-nand, > pxamci and audio components (cyclic DMA). > > However, the transition is quite intrusive and spans across several > subsystem, and due to the nature of the current pxa DMA implementation, > it cannot be gap-less. All drivers currently request an exclusive > channel via code in arch/arm/plat-pxa/dma.c and then do direct register > modifications regarding their obtained channel. We can't allow this > with the mmp-pdma driver of course, and I also have no idea how to > provide sane stubs for the existing hooks which end up in dmaengine > calls. > > Hence, all drivers have to be ported over in one series, and all these > changes should be merged by one pull request eventually in order to > prevent both build and functional breakage. Haojian's repository seems > most suitable for that, as he's the PXA maintainer. Very nice series! I just looked over these and only found a few minor details that could be improved. Arnd From mboxrd@z Thu Jan 1 00:00:00 1970 From: arnd@arndb.de (Arnd Bergmann) Date: Wed, 7 Aug 2013 18:08:51 +0200 Subject: [PATCH 00/20] ARM: pxa: move core and drivers to dmaengine In-Reply-To: <1375889649-14638-1-git-send-email-zonque@gmail.com> References: <1375889649-14638-1-git-send-email-zonque@gmail.com> Message-ID: <201308071808.51400.arnd@arndb.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wednesday 07 August 2013, Daniel Mack wrote: > I've been working on teaching the mmp-pdma driver more functions and > porting tree-wide scattered pxa specific drivers over to dmaengine > implementations. I posted the first round of patches for mmp-pdma here: > > http://marc.info/?l=linux-arm-kernel&m=137587082530228 > > > With the following patches applied, I can boot a PXA3xx board boot with > all DMA runtime information determined from DT, using pxa3xx-nand, > pxamci and audio components (cyclic DMA). > > However, the transition is quite intrusive and spans across several > subsystem, and due to the nature of the current pxa DMA implementation, > it cannot be gap-less. All drivers currently request an exclusive > channel via code in arch/arm/plat-pxa/dma.c and then do direct register > modifications regarding their obtained channel. We can't allow this > with the mmp-pdma driver of course, and I also have no idea how to > provide sane stubs for the existing hooks which end up in dmaengine > calls. > > Hence, all drivers have to be ported over in one series, and all these > changes should be merged by one pull request eventually in order to > prevent both build and functional breakage. Haojian's repository seems > most suitable for that, as he's the PXA maintainer. Very nice series! I just looked over these and only found a few minor details that could be improved. Arnd