2013-09-18 07:33:50

by Guennadi Liakhovetski

Subject: [PATCH] DMA: ste_dma40: use a power of 2 check

dst_addr_width and src_addr_width should be a power of 2. Currently the
driver checks, that they both lie between 1 and 8 and that they are eqal
to 1 or even. This however leaves an invalid value of 6 uncaught. Use an
explicit power of 2 check instead.

Signed-off-by: Guennadi Liakhovetski <[email protected]>
---

Compile tested only. I think 6 is an invalid value in this case, and since
you check for odd calues, it seems to me that any value can be supplied
here, so, looks like a power of 2 test is a better match for this.

diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
index 82d2b97..3d5e4ee 100644
--- a/drivers/dma/ste_dma40.c
+++ b/drivers/dma/ste_dma40.c
@@ -14,6 +14,7 @@
#include <linux/platform_device.h>
#include <linux/clk.h>
#include <linux/delay.h>
+#include <linux/log2.h>
#include <linux/pm.h>
#include <linux/pm_runtime.h>
#include <linux/err.h>
@@ -2796,8 +2797,8 @@ static int d40_set_runtime_config(struct dma_chan *chan,
src_addr_width > DMA_SLAVE_BUSWIDTH_8_BYTES ||
dst_addr_width <= DMA_SLAVE_BUSWIDTH_UNDEFINED ||
dst_addr_width > DMA_SLAVE_BUSWIDTH_8_BYTES ||
- ((src_addr_width > 1) && (src_addr_width & 1)) ||
- ((dst_addr_width > 1) && (dst_addr_width & 1)))
+ !is_power_of_2(src_addr_width) ||
+ !is_power_of_2(dst_addr_width))
return -EINVAL;

cfg->src_info.data_width = src_addr_width;

2013-09-23 05:12:07

by Vinod Koul

Subject: Re: [PATCH] DMA: ste_dma40: use a power of 2 check

On Wed, Sep 18, 2013 at 09:33:08AM +0200, Guennadi Liakhovetski wrote:
> dst_addr_width and src_addr_width should be a power of 2. Currently the
> driver checks, that they both lie between 1 and 8 and that they are eqal
typo ^^^^^
> to 1 or even. This however leaves an invalid value of 6 uncaught. Use an
> explicit power of 2 check instead.
>
Applied, with typo fix above..

~Vinod
> Signed-off-by: Guennadi Liakhovetski <[email protected]>
> ---
>
> Compile tested only. I think 6 is an invalid value in this case, and since
> you check for odd calues, it seems to me that any value can be supplied
> here, so, looks like a power of 2 test is a better match for this.
>
> diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c
> index 82d2b97..3d5e4ee 100644
> --- a/drivers/dma/ste_dma40.c
> +++ b/drivers/dma/ste_dma40.c
> @@ -14,6 +14,7 @@
> #include <linux/platform_device.h>
> #include <linux/clk.h>
> #include <linux/delay.h>
> +#include <linux/log2.h>
> #include <linux/pm.h>
> #include <linux/pm_runtime.h>
> #include <linux/err.h>
> @@ -2796,8 +2797,8 @@ static int d40_set_runtime_config(struct dma_chan *chan,
> src_addr_width > DMA_SLAVE_BUSWIDTH_8_BYTES ||
> dst_addr_width <= DMA_SLAVE_BUSWIDTH_UNDEFINED ||
> dst_addr_width > DMA_SLAVE_BUSWIDTH_8_BYTES ||
> - ((src_addr_width > 1) && (src_addr_width & 1)) ||
> - ((dst_addr_width > 1) && (dst_addr_width & 1)))
> + !is_power_of_2(src_addr_width) ||
> + !is_power_of_2(dst_addr_width))
> return -EINVAL;
>
> cfg->src_info.data_width = src_addr_width;

--