From mboxrd@z Thu Jan  1 00:00:00 1970
From: peter.griffin@linaro.org (Peter Griffin)
Date: Mon, 9 May 2016 18:30:43 +0100
Subject: [PATCH 03/18] dmaengine: st_fdma: Add STMicroelectronics FDMA
 engine driver support
In-Reply-To: <20160502093015.GD2274@localhost>
References: <1461236675-10176-1-git-send-email-peter.griffin@linaro.org>
 <1461236675-10176-4-git-send-email-peter.griffin@linaro.org>
 <20160426165622.GR2274@localhost>
 <20160427125923.GA12756@griffinp-ThinkPad-X1-Carbon-2nd>
 <20160502093015.GD2274@localhost>
Message-ID: <20160509173043.GA25540@griffinp-ThinkPad-X1-Carbon-2nd>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Hi Vinod,

On Mon, 02 May 2016, Vinod Koul wrote:

> On Wed, Apr 27, 2016 at 01:59:23PM +0100, Peter Griffin wrote:
> > > bunch of this seems similar to cyclic case, can you create common setup
> > > routine for these, anyway cyclic is a special cases of slave_sg
> > 
> > In v4 I've made a st_fdma_prep_common() which abstracts out all the common
> > checks at the beginning of the *_prep*() functions.
> > 
> > In v3 fill_fw_node() is already (from one of your previous reviews)
> > abstracting out all the common parts from these loops. So everything
> > that is now left is actually differences between the two setups.
> > 
> > Is that Ok?
> 
> Sounds better
> 
> > > and you irq is still enabled and tasklets can be scheduled!!
> > > 
> > Eeek! Very good point. Also looking at some other drivers we
> > should be doing a of_dma_controller_free() and
> > dma_async_device_unregister().
> > 
> > So something like this: -
> > 
> >   st_fdma_disable(); /*disables irqs*/
> >   of_dma_controller_free(pdev->dev.of_node);
> >   dma_async_device_unregister(&priv->slave);
> >   st_fdma_clk_disable(fdev);
> 
> you can call devm_free_irq() explictyly, that will syncronize irq and
> free it up.

Ok will do/

> Also you should kill the tasklet afterwards.
> 
We aren't using a tasklet in this driver.

Peter.