RE: [PATCH 1/1] ACPI: NULL pointer checking

[Moore, Robert]

> -----Original Message-----
> From: Zhenhua [mailto:lizhenhuajiyang@xxxxxxx]
> Sent: Wednesday, September 6, 2017 6:35 AM
> To: devel@xxxxxxxxxx; Moore, Robert <robert.moore@xxxxxxxxx>; Zheng, Lv
> <lv.zheng@xxxxxxxxx>; Wysocki, Rafael J <rafael.j.wysocki@xxxxxxxxx>;
> lenb@xxxxxxxxxx; linux-acpi@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx
> Cc: Zhenhua <lizhenhuajiyang@xxxxxxx>
> Subject: [PATCH 1/1] ACPI: NULL pointer checking
> 
> These two pointers should be checked, for some broken devices they may
> cause system crash.
> 
> Signed-off-by: Zhenhua <lizhenhuajiyang@xxxxxxx>
> ---
>  drivers/acpi/acpica/nsaccess.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/acpi/acpica/nsaccess.c
> b/drivers/acpi/acpica/nsaccess.c index f2733f51ca8d..151abb0ef4d5 100644
> --- a/drivers/acpi/acpica/nsaccess.c
> +++ b/drivers/acpi/acpica/nsaccess.c
> @@ -338,7 +338,7 @@ acpi_ns_lookup(union acpi_generic_state *scope_info,
>  			return_ACPI_STATUS(AE_AML_INTERNAL);
>  		}
> 
> -		if (!(flags & ACPI_NS_PREFIX_IS_SCOPE)) {
> +		if ((!prefix_node) && (!(flags & ACPI_NS_PREFIX_IS_SCOPE))) {
[Moore, Robert] 

Is there an actual case where the original code failed?



>  			/*
>  			 * This node might not be a actual "scope" node (such as
> a
>  			 * Device/Method, etc.)  It could be a Package or other
> object @@ -369,7 +369,7 @@ acpi_ns_lookup(union acpi_generic_state
> *scope_info,
>  		ACPI_DEBUG_PRINT((ACPI_DB_NAMES,
>  				  "Null Pathname (Zero segments), Flags=%X\n",
>  				  flags));
> -	} else {
> +	} else if (path) {
>  		/*
>  		 * Name pointer is valid (and must be in internal name
> format)
>  		 *
> --
> 2.14.1
>