2014-04-17 15:54:01

by Sam Protsenko

Subject: [PATCH] acpi: respect const qualifier

This patch fixes next warning:

<------------------------------- cut here --------------------------------->
drivers/acpi/property.c: In function ‘acpi_free_properties’:
drivers/acpi/property.c:35:2: warning: passing argument 1 of ‘acpi_os_free’
discards ‘const’ qualifier from pointer
target type [enabled by default]

ACPI_FREE(adev->properties);
<------------------------------- cut here --------------------------------->

The problem is next:

- acpi_free_properties() calls ACPI_FREE(), passing "const *"
parameter to it
- ACPI_FREE() is macro for acpi_os_free()
- acpi_os_free() takes "void *" parameter
- acpi_os_free() calls kfree()
- but kfree() requires "const void *" parameter

So acpi_os_free() should require "const void *" parameter as well.

Signed-off-by: Semen Protsenko <[email protected]>
---
include/acpi/platform/aclinux.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/acpi/platform/aclinux.h b/include/acpi/platform/aclinux.h
index 93c55ed..c6fa763 100644
--- a/include/acpi/platform/aclinux.h
+++ b/include/acpi/platform/aclinux.h
@@ -174,7 +174,7 @@ static inline void *acpi_os_allocate_zeroed(acpi_size size)
#define ACPI_USE_ALTERNATE_PROTOTYPE_acpi_os_allocate_zeroed
#define USE_NATIVE_ALLOCATE_ZEROED

-static inline void acpi_os_free(void *memory)
+static inline void acpi_os_free(const void *memory)
{
kfree(memory);
}
--
1.9.1

2014-04-21 20:29:08

by Rafael J. Wysocki

Subject: Re: [PATCH] acpi: respect const qualifier

On Thursday, April 17, 2014 06:53:26 PM Semen Protsenko wrote:
> This patch fixes next warning:
>
> <------------------------------- cut here --------------------------------->
> drivers/acpi/property.c: In function ‘acpi_free_properties’:
> drivers/acpi/property.c:35:2: warning: passing argument 1 of ‘acpi_os_free’
> discards ‘const’ qualifier from pointer
> target type [enabled by default]
>
> ACPI_FREE(adev->properties);
> <------------------------------- cut here --------------------------------->

This is not a mainline kernel problem, surely?

>
> The problem is next:
>
> - acpi_free_properties() calls ACPI_FREE(), passing "const *"
> parameter to it
> - ACPI_FREE() is macro for acpi_os_free()
> - acpi_os_free() takes "void *" parameter
> - acpi_os_free() calls kfree()
> - but kfree() requires "const void *" parameter
>
> So acpi_os_free() should require "const void *" parameter as well.
>
> Signed-off-by: Semen Protsenko <[email protected]>
> ---
> include/acpi/platform/aclinux.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/include/acpi/platform/aclinux.h b/include/acpi/platform/aclinux.h
> index 93c55ed..c6fa763 100644
> --- a/include/acpi/platform/aclinux.h
> +++ b/include/acpi/platform/aclinux.h
> @@ -174,7 +174,7 @@ static inline void *acpi_os_allocate_zeroed(acpi_size size)
> #define ACPI_USE_ALTERNATE_PROTOTYPE_acpi_os_allocate_zeroed
> #define USE_NATIVE_ALLOCATE_ZEROED
>
> -static inline void acpi_os_free(void *memory)
> +static inline void acpi_os_free(const void *memory)
> {
> kfree(memory);
> }
>

--
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

2014-04-25 15:02:24

by Sam Protsenko

Subject: Re: [PATCH] acpi: respect const qualifier

> This is not a mainline kernel problem, surely?

Correct, this warning is just example of possible issue that can be
occurred because of missing "const" qualifier. Actually I saw this
warning in Linaro ACPI kernel (intended for ARMv8 AArch64
architecture).

But I believe that this issue should be fixed in mainline kernel,
because it can cause the same warning in the future.

On 21 April 2014 23:44, Rafael J. Wysocki <[email protected]> wrote:
> On Thursday, April 17, 2014 06:53:26 PM Semen Protsenko wrote:
>> This patch fixes next warning:
>>
>> <------------------------------- cut here --------------------------------->
>> drivers/acpi/property.c: In function ‘acpi_free_properties’:
>> drivers/acpi/property.c:35:2: warning: passing argument 1 of ‘acpi_os_free’
>> discards ‘const’ qualifier from pointer
>> target type [enabled by default]
>>
>> ACPI_FREE(adev->properties);
>> <------------------------------- cut here --------------------------------->
>
> This is not a mainline kernel problem, surely?
>
>>
>> The problem is next:
>>
>> - acpi_free_properties() calls ACPI_FREE(), passing "const *"
>> parameter to it
>> - ACPI_FREE() is macro for acpi_os_free()
>> - acpi_os_free() takes "void *" parameter
>> - acpi_os_free() calls kfree()
>> - but kfree() requires "const void *" parameter
>>
>> So acpi_os_free() should require "const void *" parameter as well.
>>
>> Signed-off-by: Semen Protsenko <[email protected]>
>> ---
>> include/acpi/platform/aclinux.h | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/include/acpi/platform/aclinux.h b/include/acpi/platform/aclinux.h
>> index 93c55ed..c6fa763 100644
>> --- a/include/acpi/platform/aclinux.h
>> +++ b/include/acpi/platform/aclinux.h
>> @@ -174,7 +174,7 @@ static inline void *acpi_os_allocate_zeroed(acpi_size size)
>> #define ACPI_USE_ALTERNATE_PROTOTYPE_acpi_os_allocate_zeroed
>> #define USE_NATIVE_ALLOCATE_ZEROED
>>
>> -static inline void acpi_os_free(void *memory)
>> +static inline void acpi_os_free(const void *memory)
>> {
>> kfree(memory);
>> }
>>
>
> --
> I speak only for myself.
> Rafael J. Wysocki, Intel Open Source Technology Center.