diff mbox series

doc: add kernel-doc comment for u64_to_user_ptr

Message ID 20230102184349.10399-1-federico.vaga@vaga.pv.it
State New
Headers
Series doc: add kernel-doc comment for u64_to_user_ptr |

Commit Message

Federico Vaga Jan. 2, 2023, 6:43 p.m. UTC 
  Add a kernel-doc comment in kernel.h to document the macro
`u64_to_user_ptr`.

As of today, this macro is mentioned in the documentation in
'ioctl.rst' and 'botching-up-ioctls.rst'

Signed-off-by: Federico Vaga <federico.vaga@vaga.pv.it>
---
 include/linux/kernel.h | 4 ++++
 1 file changed, 4 insertions(+)

Comments

Andy Shevchenko Jan. 2, 2023, 7:03 p.m. UTC | #1 
On Mon, Jan 02, 2023 at 07:43:49PM +0100, Federico Vaga wrote:
> Add a kernel-doc comment in kernel.h to document the macro
> `u64_to_user_ptr`.

We refer functions as func(), so `u64_to_user_ptr` --> u64_to_user_ptr().

> As of today, this macro is mentioned in the documentation in
> 'ioctl.rst' and 'botching-up-ioctls.rst'

Missing period at the end.

...

> +/**
> + * u64_to_user_ptr - convert an unsigned 64bit number into a user pointer

64-bit

> + * @x: the number to convert

Isn't 'number' is a bit misleading here?
It decodes the user pointer, that is encoded into unsigned 64-bit value.
Unfortunately I am not a native speaker, I can't propose anything better.
It might be that the 'number' is quite good choice, dunno.

> + */
Federico Vaga Jan. 2, 2023, 7:15 p.m. UTC | #2 
On Mon, Jan 02, 2023 at 09:03:34PM +0200, Andy Shevchenko wrote:
>On Mon, Jan 02, 2023 at 07:43:49PM +0100, Federico Vaga wrote:
>> Add a kernel-doc comment in kernel.h to document the macro
>> `u64_to_user_ptr`.
>
>We refer functions as func(), so `u64_to_user_ptr` --> u64_to_user_ptr().
>
>> As of today, this macro is mentioned in the documentation in
>> 'ioctl.rst' and 'botching-up-ioctls.rst'
>
>Missing period at the end.
>
>...
>
>> +/**
>> + * u64_to_user_ptr - convert an unsigned 64bit number into a user pointer
>
>64-bit
>
>> + * @x: the number to convert
>
>Isn't 'number' is a bit misleading here?
>It decodes the user pointer, that is encoded into unsigned 64-bit value.
>Unfortunately I am not a native speaker, I can't propose anything better.
>It might be that the 'number' is quite good choice, dunno.

I agree about the macro's purpose. However, it does its job even if the 64-bit
value isn't a valid pointer (and this is what you will get). For this reason I
used a generic "number".

>> + */
>
>-- 
>With Best Regards,
>Andy Shevchenko
>
>
diff mbox series

Patch

diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index fe6efb24d151..8bcd126f64f2 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -56,6 +56,10 @@ 
 
 #define PTR_IF(cond, ptr)	((cond) ? (ptr) : NULL)
 
+/**
+ * u64_to_user_ptr - convert an unsigned 64bit number into a user pointer
+ * @x: the number to convert
+ */
 #define u64_to_user_ptr(x) (		\
 {					\
 	typecheck(u64, (x));		\