[PATCH] pch_dma: Fix suspend issue

[PATCH] pch_dma: Fix suspend issue

[Tomoya MORINAGA]

Currently, executing suspend/hibernation,
memory access violation occurs.

This patch fixes the issue
 - Modify array size (MAX_CHAN_NR)

Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.lapis-semi.com>
---
 drivers/dma/pch_dma.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/dma/pch_dma.c b/drivers/dma/pch_dma.c
index 1ac8d4b..42bcc80 100644
--- a/drivers/dma/pch_dma.c
+++ b/drivers/dma/pch_dma.c
@@ -60,7 +60,7 @@
 #define DMA_DESC_FOLLOW_WITHOUT_IRQ	0x2
 #define DMA_DESC_FOLLOW_WITH_IRQ	0x3
 
-#define MAX_CHAN_NR			8
+#define MAX_CHAN_NR			12
 
 #define DMA_MASK_CTL0_MODE	0x33333333
 #define DMA_MASK_CTL2_MODE	0x00003333
-- 
1.7.4.4

Re: [PATCH] pch_dma: Fix suspend issue

[Vinod Koul]

On Tue, 2011-10-04 at 16:27 +0900, Tomoya MORINAGA wrote:
> Currently, executing suspend/hibernation,
> memory access violation occurs.
> 
> This patch fixes the issue
>  - Modify array size (MAX_CHAN_NR)
> 
> Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.lapis-semi.com>
> ---
>  drivers/dma/pch_dma.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/dma/pch_dma.c b/drivers/dma/pch_dma.c
> index 1ac8d4b..42bcc80 100644
> --- a/drivers/dma/pch_dma.c
> +++ b/drivers/dma/pch_dma.c
> @@ -60,7 +60,7 @@
>  #define DMA_DESC_FOLLOW_WITHOUT_IRQ	0x2
>  #define DMA_DESC_FOLLOW_WITH_IRQ	0x3
>  
> -#define MAX_CHAN_NR			8
> +#define MAX_CHAN_NR			12
>  
>  #define DMA_MASK_CTL0_MODE	0x33333333
>  #define DMA_MASK_CTL2_MODE	0x00003333

How would changing the max number of descriptors fix the memory
violation?

-- 
~Vinod

Re: [PATCH] pch_dma: Fix suspend issue

[Tomoya MORINAGA]

(2011/10/07 14:00), Vinod Koul wrote:
> How would changing the max number of descriptors fix the memory
> violation?
>

In probe of pch_dma,  you can see the following code.

	for (i = 0; i < nr_channels; i++) {

snip...

		pd_chan->membase = &regs->desc[i];

snip...

	}

Max value of nr_channels is 12 defined at pci_table.
So, this caused memory access violation.

-- 
tomoya
ROHM Co., Ltd.

Re: [PATCH] pch_dma: Fix suspend issue

[Vinod Koul]

On Fri, 2011-10-07 at 14:54 +0900, Tomoya MORINAGA wrote:
> (2011/10/07 14:00), Vinod Koul wrote:
> > How would changing the max number of descriptors fix the memory
> > violation?
> >
> 
> In probe of pch_dma,  you can see the following code.
> 
> 	for (i = 0; i < nr_channels; i++) {
> 
> snip...
> 
> 		pd_chan->membase = &regs->desc[i];
> 
> snip...
> 
> 	}
> 
> Max value of nr_channels is 12 defined at pci_table.
> So, this caused memory access violation.
This is all is due to not saving nr_channels in probe and using that
instead.

Looking deeper, struct pch_dma defines array of length MAX_CHAN_NR for
channels. Then why do you allocate memory in probe as 
         pd = kzalloc(sizeof(struct pch_dma)+
                 sizeof(struct pch_dma_chan) * nr_channels, GFP_KERNEL);
what is the point in allocating additional memory for each channel?

Given this, why should there be predefined channel array in pcm_dma?

It would be great if we could fix this by dynamically allocating memory
for channels based on nr_channels and not wasting due to a static array.

-- 
~Vinod

Re: [PATCH] pch_dma: Fix suspend issue

[Tomoya MORINAGA]

(2011/10/07 15:14), Vinod Koul wrote:
> This is all is due to not saving nr_channels in probe and using that
> instead.
>
> Looking deeper, struct pch_dma defines array of length MAX_CHAN_NR for
> channels. Then why do you allocate memory in probe as
>           pd = kzalloc(sizeof(struct pch_dma)+
>                   sizeof(struct pch_dma_chan) * nr_channels, GFP_KERNEL);
> what is the point in allocating additional memory for each channel?
>
> Given this, why should there be predefined channel array in pcm_dma?
>
> It would be great if we could fix this by dynamically allocating memory
> for channels based on nr_channels and not wasting due to a static array.

You are right.
Current pch_dma driver looks like wasting memory.

We must review pch_dma driver again.

Thanks,
-- 
tomoya
ROHM Co., Ltd.

Re: [PATCH] pch_dma: Fix suspend issue

[Vinod Koul]

On Fri, 2011-10-07 at 19:33 +0900, Tomoya MORINAGA wrote:
> (2011/10/07 15:14), Vinod Koul wrote:
> > This is all is due to not saving nr_channels in probe and using that
> > instead.
> >
> > Looking deeper, struct pch_dma defines array of length MAX_CHAN_NR for
> > channels. Then why do you allocate memory in probe as
> >           pd = kzalloc(sizeof(struct pch_dma)+
> >                   sizeof(struct pch_dma_chan) * nr_channels, GFP_KERNEL);
> > what is the point in allocating additional memory for each channel?
> >
> > Given this, why should there be predefined channel array in pcm_dma?
> >
> > It would be great if we could fix this by dynamically allocating memory
> > for channels based on nr_channels and not wasting due to a static array.
> 
> You are right.
> Current pch_dma driver looks like wasting memory.
> 
> We must review pch_dma driver again.
Okay, can you update the changelog to indicate how it is fixing the
suspend issue, original change log doesn't help


-- 
~Vinod