From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <507FB495.7050104@freescale.com> Date: Thu, 18 Oct 2012 15:49:41 +0800 From: Huang Shijie MIME-Version: 1.0 To: Marek Vasut Subject: Re: [PATCH] dma: add new DMA control commands References: <1350538335-29026-1-git-send-email-b32955@freescale.com> <1350541111.5263.3.camel@vkoul-udesk3> <507FA595.4020507@freescale.com> <201210180914.58527.marex@denx.de> In-Reply-To: <201210180914.58527.marex@denx.de> Content-Type: text/plain; charset="UTF-8"; format=flowed Content-Transfer-Encoding: quoted-printable Cc: alsa-devel@alsa-project.org, tiwai@suse.de, artem.bityutskiy@linux.intel.com, perex@perex.cz, linux-mtd@lists.infradead.org, linux-i2c@vger.kernel.org, Huang Shijie , linux@arm.linux.org.uk, cjb@laptop.org, lrg@ti.com, ben-linux@fluff.org, linux-arm-kernel@lists.infradead.org, Fabio Estevam , Vinod Koul , broonie@opensource.wolfsonmicro.com, linux-mmc@vger.kernel.org, w.sang@pengutronix.de, linux-kernel@vger.kernel.org, djbw@fb.com, khali@linux-fr.org, shawn.guo@linaro.org, dwmw2@infradead.org List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , =E4=BA=8E 2012=E5=B9=B410=E6=9C=8818=E6=97=A5 15:14, Marek Vasut =E5=86=99= =E9=81=93: > Dear Huang Shijie, > > Why such massive CC ? > >> =E4=BA=8E 2012=E5=B9=B410=E6=9C=8818=E6=97=A5 14:18, Vinod Koul =E5=86= =99=E9=81=93: >>> Why cant you do start (prepare clock etc) when you submit the descrip= tor >>> to dmaengine. Can be done in tx_submit callback. >>> Similarly remove the clock when dma transaction gets completed. >> I ever thought this method too. >> >> But it will become low efficient in the following case: >> >> Assuming the gpmi-nand driver has to read out 1024 pages in one >> _SINGLE_ read operation. >> The gpmi-nand will submit the descriptor to dmaengine per page. > It will? Then GPMI NAND is flat our broken ... again. yes. Please read the NAND chip spec about the comand READ PAGE(00h-30h) and=20 the code nand_do_read_ops(). The nand chip limits us only read one page out one ti= me. So the driver will submit the descriptor to dmaengine per page. >> So with >> your method, >> the system will repeat the enable/disable dma clock 1024 time. > Yes, it is the driver that's wrong. not the driver. >> At every >> enable/disable dma clock, >> the system has to enable the clock chain and it's parents ... >> >> But with this patch, we only need to enable/disable dma clock one time= , >> just at we select the nand chip. > You are fixing a driver problem at a framework level, wrong. > > So, check how the MXS SPI driver handles descriptor chaining. Indeed, t= he > Sigmatel screwed the DMA stuff good. But if you analyze the SPI driver,= you'll > notice a few important points that might come handy when you fix the GP= MI NAND > driver properly. > > The direction to take here is: > 1) Implement DMA chaining into the GPMI NAND driver How can i implement the DMA chain if the nand chip READ-PAGE command=20 gives us the one page limit? thanks Huang Shijie > 2) You might need to do one PIO transfer to reconfigure the IP register= s between > each segment of the DMA chain (just as MXS SPI does it) > 3) You might run out of DMA descriptors when doing too long chains -- s= o you > might need to fix that part of the mxs DMA driver.