[PATCH V1] arm: tegra: dma: not required to move requestor when stopping.

[PATCH V1] arm: tegra: dma: not required to move requestor when stopping.

[Laxman Dewangan]

It is not require to move the requestor of dma to INVALID
option before stopping dma.

Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
---
 arch/arm/mach-tegra/dma.c |    8 --------
 1 files changed, 0 insertions(+), 8 deletions(-)

diff --git a/arch/arm/mach-tegra/dma.c b/arch/arm/mach-tegra/dma.c
index 998c55d..abea4f6 100644
--- a/arch/arm/mach-tegra/dma.c
+++ b/arch/arm/mach-tegra/dma.c
@@ -52,8 +52,6 @@
 #define CSR_ONCE				(1<<27)
 #define CSR_FLOW				(1<<21)
 #define CSR_REQ_SEL_SHIFT			16
-#define CSR_REQ_SEL_MASK			(0x1F<<CSR_REQ_SEL_SHIFT)
-#define CSR_REQ_SEL_INVALID			(31<<CSR_REQ_SEL_SHIFT)
 #define CSR_WCOUNT_SHIFT			2
 #define CSR_WCOUNT_MASK				0xFFFC
 
@@ -183,18 +181,12 @@ static void tegra_dma_stop(struct tegra_dma_channel *ch)
 
 static int tegra_dma_cancel(struct tegra_dma_channel *ch)
 {
-	u32 csr;
 	unsigned long irq_flags;
 
 	spin_lock_irqsave(&ch->lock, irq_flags);
 	while (!list_empty(&ch->list))
 		list_del(ch->list.next);
 
-	csr = readl(ch->addr + APB_DMA_CHAN_CSR);
-	csr &= ~CSR_REQ_SEL_MASK;
-	csr |= CSR_REQ_SEL_INVALID;
-	writel(csr, ch->addr + APB_DMA_CHAN_CSR);
-
 	tegra_dma_stop(ch);
 
 	spin_unlock_irqrestore(&ch->lock, irq_flags);
-- 
1.7.1.1

RE: [PATCH V1] arm: tegra: dma: not required to move requestor when stopping.

[Stephen Warren]

Laxman Dewangan wrote at Tuesday, January 24, 2012 1:11 AM:
> It is not require to move the requestor of dma to INVALID
> option before stopping dma.
> 
> Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>

Acked-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>

I know that Laxman has checked with the HW engineers, and they see no
reason to do this. I was originally worried about hung FIFOs preventing
an in-progress DMA from completely without the removed code, but according
to HW, that isn't a concern.

-- 
nvpublic

Re: [PATCH V1] arm: tegra: dma: not required to move requestor when stopping.

[Laxman Dewangan]

On Wednesday 25 January 2012 02:27 AM, Stephen Warren wrote:
> Laxman Dewangan wrote at Tuesday, January 24, 2012 1:11 AM:
>> It is not require to move the requestor of dma to INVALID
>> option before stopping dma.
>>
>> Signed-off-by: Laxman Dewangan<ldewangan@nvidia.com>
> Acked-by: Stephen Warren<swarren@nvidia.com>
> Tested-by: Stephen Warren<swarren@nvidia.com>
>
> I know that Laxman has checked with the HW engineers, and they see no
> reason to do this. I was originally worried about hung FIFOs preventing
> an in-progress DMA from completely without the removed code, but according
> to HW, that isn't a concern.
>
Olaf,
Can it be applied if it is OK to you?

Thanks,
Laxman

Re: [PATCH V1] arm: tegra: dma: not required to move requestor when stopping.

[Olof Johansson]

On Tue, Jan 31, 2012 at 04:32:48PM +0530, Laxman Dewangan wrote:
> On Wednesday 25 January 2012 02:27 AM, Stephen Warren wrote:
> >Laxman Dewangan wrote at Tuesday, January 24, 2012 1:11 AM:
> >>It is not require to move the requestor of dma to INVALID
> >>option before stopping dma.
> >>
> >>Signed-off-by: Laxman Dewangan<ldewangan@nvidia.com>
> >Acked-by: Stephen Warren<swarren@nvidia.com>
> >Tested-by: Stephen Warren<swarren@nvidia.com>
> >
> >I know that Laxman has checked with the HW engineers, and they see no
> >reason to do this. I was originally worried about hung FIFOs preventing
> >an in-progress DMA from completely without the removed code, but according
> >to HW, that isn't a concern.
> >
> Olaf,
> Can it be applied if it is OK to you?

Yep, applied.


-Olof