* Re: [PATCH] linux/container_of.h: Remove redundant type cast in container_of_const()
2024-08-09 11:27 ` Andy Shevchenko
@ 2024-08-09 12:04 ` Zijun Hu
2024-08-12 14:17 ` Sakari Ailus
2024-08-13 9:17 ` Greg Kroah-Hartman
2 siblings, 0 replies; 5+ messages in thread
From: Zijun Hu @ 2024-08-09 12:04 UTC (permalink / raw)
To: Andy Shevchenko
Cc: Greg Kroah-Hartman, Jason Gunthorpe, Sakari Ailus,
Matthew Wilcox (Oracle),
Rafael J. Wysocki, linux-kernel, Zijun Hu
On 2024/8/9 19:27, Andy Shevchenko wrote:
> On Fri, Aug 02, 2024 at 11:15:15PM +0800, Zijun Hu wrote:
>> From: Zijun Hu <quic_zijuhu@quicinc.com>
>>
>> Remove redundant (type *) cast for default branch in container_of_const()
>> since the cast has been done by container_of().
>
> While it might have same effect, the below is explicitly clear about both
> cases. With your patch it will become inconsistent.
>
my change is more obvious to say that container_of_const() is same as
container_of() for none const pointer parameter @ptr. it is simpler and
easier to understand.
> ...
>
>> #define container_of_const(ptr, type, member) \
>> _Generic(ptr, \
>> const typeof(*(ptr)) *: ((const type *)container_of(ptr, type, member)),\
>
> (see, in the above line the cast is still present / required)
>
yes, for above case, (const type *) is suitable and required.
>> - default: ((type *)container_of(ptr, type, member)) \
>> + default: container_of(ptr, type, member) \
>> )
>
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] linux/container_of.h: Remove redundant type cast in container_of_const()
2024-08-09 11:27 ` Andy Shevchenko
2024-08-09 12:04 ` Zijun Hu
@ 2024-08-12 14:17 ` Sakari Ailus
2024-08-13 9:17 ` Greg Kroah-Hartman
2 siblings, 0 replies; 5+ messages in thread
From: Sakari Ailus @ 2024-08-12 14:17 UTC (permalink / raw)
To: Andy Shevchenko
Cc: Zijun Hu, Greg Kroah-Hartman, Jason Gunthorpe,
Matthew Wilcox (Oracle),
Rafael J. Wysocki, linux-kernel, Zijun Hu
Hi Andy, Zijun,
On Fri, Aug 09, 2024 at 02:27:23PM +0300, Andy Shevchenko wrote:
> On Fri, Aug 02, 2024 at 11:15:15PM +0800, Zijun Hu wrote:
> > From: Zijun Hu <quic_zijuhu@quicinc.com>
> >
> > Remove redundant (type *) cast for default branch in container_of_const()
> > since the cast has been done by container_of().
>
> While it might have same effect, the below is explicitly clear about both
> cases. With your patch it will become inconsistent.
In the const case it's actually required whereas container_of() already
does exactly the same cast, rendering the cast done below redundant.
There's nothing inconsistent in removing a needless cast.
Removing it is my preference as well.
>
> ...
>
> > #define container_of_const(ptr, type, member) \
> > _Generic(ptr, \
> > const typeof(*(ptr)) *: ((const type *)container_of(ptr, type, member)),\
>
> (see, in the above line the cast is still present / required)
>
> > - default: ((type *)container_of(ptr, type, member)) \
> > + default: container_of(ptr, type, member) \
> > )
>
--
Kind regards,
Sakari Ailus
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH] linux/container_of.h: Remove redundant type cast in container_of_const()
2024-08-09 11:27 ` Andy Shevchenko
2024-08-09 12:04 ` Zijun Hu
2024-08-12 14:17 ` Sakari Ailus
@ 2024-08-13 9:17 ` Greg Kroah-Hartman
2 siblings, 0 replies; 5+ messages in thread
From: Greg Kroah-Hartman @ 2024-08-13 9:17 UTC (permalink / raw)
To: Andy Shevchenko
Cc: Zijun Hu, Jason Gunthorpe, Sakari Ailus, Matthew Wilcox (Oracle),
Rafael J. Wysocki, linux-kernel, Zijun Hu
On Fri, Aug 09, 2024 at 02:27:23PM +0300, Andy Shevchenko wrote:
> On Fri, Aug 02, 2024 at 11:15:15PM +0800, Zijun Hu wrote:
> > From: Zijun Hu <quic_zijuhu@quicinc.com>
> >
> > Remove redundant (type *) cast for default branch in container_of_const()
> > since the cast has been done by container_of().
>
> While it might have same effect, the below is explicitly clear about both
> cases. With your patch it will become inconsistent.
Which is why I did it this way. It's safe to keep as-is, so let's do so
please.
Remember, we write code for people first, compilers second. There's
nothing wrong with the code as-is and it makes it more obvious what is
happening when someone reads it.
thanks,
greg k-h
^ permalink raw reply [flat|nested] 5+ messages in thread