[PATCH] ARM: EXYNOS: Handle of_find_device_by_node and kstrdup failures

[PATCH] ARM: EXYNOS: Handle of_find_device_by_node and kstrdup failures

[Krzysztof Kozlowski]

Prevent possible NULL pointer dereference of pointer returned by
of_find_device_by_node(). Handle this by skipping such power domain.

Additionally fail the init on kstrdup() failure. Such case is actually
not fatal because the name for power domain allocated by kstrdup() is
used only in printk. Still as a precaution handle this as an error
condition.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
---
 arch/arm/mach-exynos/pm_domains.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/arch/arm/mach-exynos/pm_domains.c b/arch/arm/mach-exynos/pm_domains.c
index fd7640a6b503..4c6c29ba58ef 100644
--- a/arch/arm/mach-exynos/pm_domains.c
+++ b/arch/arm/mach-exynos/pm_domains.c
@@ -116,6 +116,12 @@ static __init int exynos4_pm_init_power_domain(void)
 		struct device *dev;
 
 		pdev = of_find_device_by_node(np);
+		if (!pdev) {
+			pr_err("%s: failed to find device for node %s\n",
+					__func__, np->name);
+			of_node_put(np);
+			continue;
+		}
 		dev = &pdev->dev;
 
 		pd = kzalloc(sizeof(*pd), GFP_KERNEL);
@@ -126,6 +132,12 @@ static __init int exynos4_pm_init_power_domain(void)
 		}
 
 		pd->pd.name = kstrdup(np->name, GFP_KERNEL);
+		if (!pd->pd.name) {
+			kfree(pd);
+			of_node_put(np);
+			return -ENOMEM;
+		}
+
 		pd->name = pd->pd.name;
 		pd->base = of_iomap(np, 0);
 		if (!pd->base) {
-- 
1.9.1

[PATCH] ARM: EXYNOS: Handle of_find_device_by_node and kstrdup failures

[Krzysztof Kozlowski]

On wto, 2015-02-03 at 18:28 +0100, Krzysztof Kozlowski wrote:
> Prevent possible NULL pointer dereference of pointer returned by
> of_find_device_by_node(). Handle this by skipping such power domain.
> 
> Additionally fail the init on kstrdup() failure. Such case is actually
> not fatal because the name for power domain allocated by kstrdup() is
> used only in printk. Still as a precaution handle this as an error
> condition.
> 
> Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>

Hi Kukjin,

What is the status of this patch and the "ARM: EXYNOS: Handle of
of_iomap() failure". The latter should be in your tree but I cannot find
it.

Best regards,
Krzysztof

> ---
>  arch/arm/mach-exynos/pm_domains.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/arch/arm/mach-exynos/pm_domains.c b/arch/arm/mach-exynos/pm_domains.c
> index fd7640a6b503..4c6c29ba58ef 100644
> --- a/arch/arm/mach-exynos/pm_domains.c
> +++ b/arch/arm/mach-exynos/pm_domains.c
> @@ -116,6 +116,12 @@ static __init int exynos4_pm_init_power_domain(void)
>  		struct device *dev;
>  
>  		pdev = of_find_device_by_node(np);
> +		if (!pdev) {
> +			pr_err("%s: failed to find device for node %s\n",
> +					__func__, np->name);
> +			of_node_put(np);
> +			continue;
> +		}
>  		dev = &pdev->dev;
>  
>  		pd = kzalloc(sizeof(*pd), GFP_KERNEL);
> @@ -126,6 +132,12 @@ static __init int exynos4_pm_init_power_domain(void)
>  		}
>  
>  		pd->pd.name = kstrdup(np->name, GFP_KERNEL);
> +		if (!pd->pd.name) {
> +			kfree(pd);
> +			of_node_put(np);
> +			return -ENOMEM;
> +		}
> +
>  		pd->name = pd->pd.name;
>  		pd->base = of_iomap(np, 0);
>  		if (!pd->base) {

[PATCH] ARM: EXYNOS: Handle of_find_device_by_node and kstrdup failures

[Krzysztof Kozlowski]

2015-02-24 16:42 GMT+01:00 Krzysztof Kozlowski <k.kozlowski@samsung.com>:
> On wto, 2015-02-03 at 18:28 +0100, Krzysztof Kozlowski wrote:
>> Prevent possible NULL pointer dereference of pointer returned by
>> of_find_device_by_node(). Handle this by skipping such power domain.
>>
>> Additionally fail the init on kstrdup() failure. Such case is actually
>> not fatal because the name for power domain allocated by kstrdup() is
>> used only in printk. Still as a precaution handle this as an error
>> condition.
>>
>> Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
>
> Hi Kukjin,
>
> What is the status of this patch and the "ARM: EXYNOS: Handle of
> of_iomap() failure". The latter should be in your tree but I cannot find
> it.
>
> Best regards,
> Krzysztof

Hi Kukjin,

You haven't applied this and the "ARM: EXYNOS: Handle of of_iomap()
failure" yet. Any comments?

Best regards,
Krzysztof