[PATCH] dma/imx-sdma: add wmb before enable/run a channel

[Uwe Kleine-König]

Hello Richard,

On Sun, Jan 01, 2012 at 11:27:40AM +0800, Richard Zhao wrote:
> dma_alloc_coherent memory may be bufferable. We need to add
> nececcary memory barrier.
necessary

> 
> Signed-off-by: Richard Zhao <richard.zhao at linaro.org>
> ---
>  drivers/dma/imx-sdma.c |    2 ++
>  1 files changed, 2 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
> index 2ebb2bc..8cd9492 100644
> --- a/drivers/dma/imx-sdma.c
> +++ b/drivers/dma/imx-sdma.c
> @@ -403,6 +403,7 @@ static int sdma_run_channel(struct sdma_channel *sdmac)
>  
>  	init_completion(&sdmac->done);
>  
> +	wmb();
>  	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
>  
>  	ret = wait_for_completion_timeout(&sdmac->done, HZ);
> @@ -809,6 +810,7 @@ out:
>  
>  static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
>  {
> +	wmb();
>  	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
>  }
I don't know the sdma code very well, but wouldn't it be nice to call
sdma_enable_channel() in sdma_run_channel() instead of open coding it?

As you probably need some rearrangements to have sdma_enable_channel()
declared in sdma_run_channel() this might be better done with a seperate
patch.

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |