[PATCH v3 0/2] imx-sdma: memory barrier fix

[PATCH v3 0/2] imx-sdma: memory barrier fix

[Richard Zhao]

Changes in v3:
 - use writel instead of explicitly call wmb.

Changes in v2:
 - add "let sdma_run_channel call sdma_enable_channel"

[PATCH v3 1/2] dma/imx-sdma: let sdma_run_channel call
[PATCH v3 2/2] dma/imx-sdma: use writel to write register when

Thanks
Richard

[PATCH v3 1/2] dma/imx-sdma: let sdma_run_channel call sdma_enable_channel

[Richard Zhao]

Let all enable channel code call sdma_enable_channel.

Signed-off-by: Richard Zhao <richard.zhao@linaro.org>
---
 drivers/dma/imx-sdma.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index f59fd8f..c2bc4f1 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -394,6 +394,11 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
 	return 0;
 }
 
+static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
+{
+	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
+}
+
 /*
  * sdma_run_channel - run a channel and wait till it's done
  */
@@ -405,7 +410,7 @@ static int sdma_run_channel(struct sdma_channel *sdmac)
 
 	init_completion(&sdmac->done);
 
-	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
+	sdma_enable_channel(sdma, channel);
 
 	ret = wait_for_completion_timeout(&sdmac->done, HZ);
 
@@ -811,11 +816,6 @@ out:
 	return ret;
 }
 
-static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
-{
-	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
-}
-
 static dma_cookie_t sdma_assign_cookie(struct sdma_channel *sdmac)
 {
 	dma_cookie_t cookie = sdmac->chan.cookie;
-- 
1.7.5.4

Re: [PATCH v3 1/2] dma/imx-sdma: let sdma_run_channel call sdma_enable_channel

[Shawn Guo]

On Thu, Jan 05, 2012 at 03:59:21PM +0800, Richard Zhao wrote:
> Let all enable channel code call sdma_enable_channel.
> 
> Signed-off-by: Richard Zhao <richard.zhao@linaro.org>

Acked-by: Shawn Guo <shawn.guo@linaro.org>

-- 
Regards,
Shawn

[PATCH v3 2/2] dma/imx-sdma: use writel to write register when enable a channel

[Richard Zhao]

dma_alloc_coherent memory may be bufferable when set
CONFIG_ARM_DMA_MEM_BUFFERABLE. We need to add nececcary
memory barrier. writel implicitly call wmb in such case.

Signed-off-by: Richard Zhao <richard.zhao@linaro.org>
---
 drivers/dma/imx-sdma.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index c2bc4f1..e987468 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -396,7 +396,7 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
 
 static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
 {
-	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
+	writel(1 << channel, sdma->regs + SDMA_H_START);
 }
 
 /*
-- 
1.7.5.4

Re: [PATCH v3 2/2] dma/imx-sdma: use writel to write register when enable a channel

[Shawn Guo]

On Thu, Jan 05, 2012 at 03:59:22PM +0800, Richard Zhao wrote:
> dma_alloc_coherent memory may be bufferable when set
> CONFIG_ARM_DMA_MEM_BUFFERABLE. We need to add nececcary
> memory barrier. writel implicitly call wmb in such case.
> 
> Signed-off-by: Richard Zhao <richard.zhao@linaro.org>
> ---
>  drivers/dma/imx-sdma.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
> index c2bc4f1..e987468 100644
> --- a/drivers/dma/imx-sdma.c
> +++ b/drivers/dma/imx-sdma.c
> @@ -396,7 +396,7 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
>  
>  static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
>  {
> -	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
> +	writel(1 << channel, sdma->regs + SDMA_H_START);
>  }
>  

As educated by Arnd, generally it's safer to use pair of readl/writel
than __raw_readl/__raw_writel in driver.  I'm wondering if it's a good
opportunity for us to change the pair all over this driver.

-- 
Regards,
Shawn

Re: [PATCH v3 2/2] dma/imx-sdma: use writel to write register when enable a channel

[Richard Zhao]

Hi Shawn,

On Thu, Jan 05, 2012 at 05:23:17PM +0800, Shawn Guo wrote:
> On Thu, Jan 05, 2012 at 03:59:22PM +0800, Richard Zhao wrote:
> > dma_alloc_coherent memory may be bufferable when set
> > CONFIG_ARM_DMA_MEM_BUFFERABLE. We need to add nececcary
> > memory barrier. writel implicitly call wmb in such case.
> > 
> > Signed-off-by: Richard Zhao <richard.zhao@linaro.org>
> > ---
> >  drivers/dma/imx-sdma.c |    2 +-
> >  1 files changed, 1 insertions(+), 1 deletions(-)
> > 
> > diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
> > index c2bc4f1..e987468 100644
> > --- a/drivers/dma/imx-sdma.c
> > +++ b/drivers/dma/imx-sdma.c
> > @@ -396,7 +396,7 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
> >  
> >  static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
> >  {
> > -	__raw_writel(1 << channel, sdma->regs + SDMA_H_START);
> > +	writel(1 << channel, sdma->regs + SDMA_H_START);
> >  }
> >  
> 
> As educated by Arnd, generally it's safer to use pair of readl/writel
> than __raw_readl/__raw_writel in driver.  I'm wondering if it's a good
> opportunity for us to change the pair all over this driver.
Russel mentioned that too. But I cannot understand.
If CONFIG_ARM_DMA_MEM_BUFFERABLE, readl/writel always call rmb/wmb. Do
we realy need it? It might affect performance.

Du to the concern, I didn't convert all the readl/writel.

Thanks
Richard
> 
> -- 
> Regards,
> Shawn
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
> 

Re: [PATCH v3 2/2] dma/imx-sdma: use writel to write register when enable a channel

[Eric Miao]

On Thu, Jan 5, 2012 at 5:19 PM, Richard Zhao <richard.zhao@freescale.com> wrote:
> Hi Shawn,
>
> On Thu, Jan 05, 2012 at 05:23:17PM +0800, Shawn Guo wrote:
>> On Thu, Jan 05, 2012 at 03:59:22PM +0800, Richard Zhao wrote:
>> > dma_alloc_coherent memory may be bufferable when set
>> > CONFIG_ARM_DMA_MEM_BUFFERABLE. We need to add nececcary
>> > memory barrier. writel implicitly call wmb in such case.
>> >
>> > Signed-off-by: Richard Zhao <richard.zhao@linaro.org>
>> > ---
>> >  drivers/dma/imx-sdma.c |    2 +-
>> >  1 files changed, 1 insertions(+), 1 deletions(-)
>> >
>> > diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
>> > index c2bc4f1..e987468 100644
>> > --- a/drivers/dma/imx-sdma.c
>> > +++ b/drivers/dma/imx-sdma.c
>> > @@ -396,7 +396,7 @@ static int sdma_config_ownership(struct sdma_channel *sdmac,
>> >
>> >  static void sdma_enable_channel(struct sdma_engine *sdma, int channel)
>> >  {
>> > -   __raw_writel(1 << channel, sdma->regs + SDMA_H_START);
>> > +   writel(1 << channel, sdma->regs + SDMA_H_START);
>> >  }
>> >
>>
>> As educated by Arnd, generally it's safer to use pair of readl/writel
>> than __raw_readl/__raw_writel in driver.  I'm wondering if it's a good
>> opportunity for us to change the pair all over this driver.
> Russel mentioned that too. But I cannot understand.
> If CONFIG_ARM_DMA_MEM_BUFFERABLE, readl/writel always call rmb/wmb. Do
> we realy need it? It might affect performance.

It's generally a good idea to use readl() and writel() unless it's performance
critical, e.g. where endian conversion is not necessary and IO read/write
order doesn't matter.

As MT_DEVICE is strongly ordered, I think it'll be OK for __raw_{read,write}l()
to be used for those ioremap() register accesses, not necessarily to those
ioremap_*() version though.

Re: [PATCH v3 2/2] dma/imx-sdma: use writel to write register when enable a channel

[Shawn Guo]

On Thu, Jan 05, 2012 at 05:19:05PM +0800, Richard Zhao wrote:
> Russel mentioned that too. But I cannot understand.
> If CONFIG_ARM_DMA_MEM_BUFFERABLE, readl/writel always call rmb/wmb. Do
> we realy need it? It might affect performance.
> 
For safety, I think we need it.  The operations should be really fast.

-- 
Regards,
Shawn