From: Colin Ian King <[email protected]>
Currently the for-loop using an unsigned int for the loop counter
which is problematic when comparing it to the signed int
gt->gpio_cs_count. This is an issue because if the signed int is
negative (for example, the call to of_gpio_count failed) then
the negative loop bound is implicitly cast to an unsigned int on
the comparison to loop counter i and will yield a very large value,
eventually causing an array bounds overflow on hw->gpio_cs.
Fix this by simply making the loop counter i a signed int;
Fixes: ca632f556697 ("spi: reorganize drivers")
Signed-off-by: Colin Ian King <[email protected]>
---
drivers/spi/spi-oc-tiny.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/spi/spi-oc-tiny.c b/drivers/spi/spi-oc-tiny.c
index 085f580be7ec..81f74b938dc9 100644
--- a/drivers/spi/spi-oc-tiny.c
+++ b/drivers/spi/spi-oc-tiny.c
@@ -206,7 +206,7 @@ static int tiny_spi_of_probe(struct platform_device *pdev)
{
struct tiny_spi *hw = platform_get_drvdata(pdev);
struct device_node *np = pdev->dev.of_node;
- unsigned int i;
+ int i;
u32 val;
if (!np)
--
2.20.1
On 3/31/2019 2:53 AM, Colin King wrote:
> From: Colin Ian King <[email protected]>
>
> Currently the for-loop using an unsigned int for the loop counter
> which is problematic when comparing it to the signed int
> gt->gpio_cs_count. This is an issue because if the signed int is
> negative (for example, the call to of_gpio_count failed) then
> the negative loop bound is implicitly cast to an unsigned int on
> the comparison to loop counter i and will yield a very large value,
> eventually causing an array bounds overflow on hw->gpio_cs.
>
> Fix this by simply making the loop counter i a signed int;
>
> Fixes: ca632f556697 ("spi: reorganize drivers")
> Signed-off-by: Colin Ian King <[email protected]>
> ---
> drivers/spi/spi-oc-tiny.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/spi/spi-oc-tiny.c b/drivers/spi/spi-oc-tiny.c
> index 085f580be7ec..81f74b938dc9 100644
> --- a/drivers/spi/spi-oc-tiny.c
> +++ b/drivers/spi/spi-oc-tiny.c
> @@ -206,7 +206,7 @@ static int tiny_spi_of_probe(struct platform_device *pdev)
> {
> struct tiny_spi *hw = platform_get_drvdata(pdev);
> struct device_node *np = pdev->dev.of_node;
> - unsigned int i;
> + int i;
same here as well you are putting wrapper over a original issue by not
checking the returned value.
Please fix that in both patches.
Thanks.
Mukesh
> u32 val;
>
> if (!np)
On 3/31/2019 2:53 AM, Colin King wrote:
> From: Colin Ian King <[email protected]>
>
> Currently the for-loop using an unsigned int for the loop counter
> which is problematic when comparing it to the signed int
> gt->gpio_cs_count. This is an issue because if the signed int is
> negative (for example, the call to of_gpio_count failed) then
> the negative loop bound is implicitly cast to an unsigned int on
> the comparison to loop counter i and will yield a very large value,
> eventually causing an array bounds overflow on hw->gpio_cs.
>
> Fix this by simply making the loop counter i a signed int;
>
> Fixes: ca632f556697 ("spi: reorganize drivers")
> Signed-off-by: Colin Ian King <[email protected]>
> ---
> drivers/spi/spi-oc-tiny.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/spi/spi-oc-tiny.c b/drivers/spi/spi-oc-tiny.c
> index 085f580be7ec..81f74b938dc9 100644
> --- a/drivers/spi/spi-oc-tiny.c
> +++ b/drivers/spi/spi-oc-tiny.c
> @@ -206,7 +206,7 @@ static int tiny_spi_of_probe(struct platform_device *pdev)
> {
> struct tiny_spi *hw = platform_get_drvdata(pdev);
> struct device_node *np = pdev->dev.of_node;
> - unsigned int i;
> + int i;
Same issue, here as well like your other patch. you are putting wrapper
over a original issue by not checking the returned value.
Please fix that in both patches.
Thanks.
Mukesh
> u32 val;
>
> if (!np)
On Sat, Mar 30, 2019 at 09:23:00PM +0000, Colin King wrote:
> Fixes: ca632f556697 ("spi: reorganize drivers")
Please think about the fixes tags you're applying, this is obviously not
the source of the issue.