Re: [PATCH v2] highmem: Make kunmap_{local,atomic}() take pointers to const void

[Fabio M. De Francesco]

On mercoledì 15 giugno 2022 07:12:56 CEST Fabio M. De Francesco wrote:
> kunmap_{local,atomic}() currently take pointers to void. However, this
> is semantically incorrect, since these functions do not change the memory
> their arguments point to.
> 
> Therefore, make this semantics explicit by modifying the
> kunmap_{local,atomic}() prototypes to take pointers to const void.
> 
> As side effects, compilers will likely produce more efficient code and
> they won't any longer need casts to pointers to void where these
> functions take arguments of type pointer to const void.
> 
> Suggested-by: David Sterba <dsterba@xxxxxxx>
> Suggested-by: Ira Weiny <ira.weiny@xxxxxxxxx>
> Signed-off-by: Fabio M. De Francesco <fmdefrancesco@xxxxxxxxx>
> ---
> 
> v1->v2: Change the commit message to clearly explain why these functions
> should require pointers to const void. The fundamental argument behind
> the commit message changes is semantic correctness. Other bonuses come as 
> side effects. Obviously there are no changes to the code. 
> 
> Many thanks to David Sterba and Ira Weiny for suggestions and reviews.
> 
>  include/linux/highmem-internal.h | 10 +++++-----
>  mm/highmem.c                     |  2 +-
>  2 files changed, 6 insertions(+), 6 deletions(-)
> 
As said in another email, this is the second version of "[PATCH] highmem: 
Make __kunmap_{local,atomic}() take "const void *"". You see that I changed 
subject (involuntarily, that was just a mistake when formatting).

This is why I'm writing: make Maintainers find easily which patch this is 
the second version of.

Sorry for that unwanted change to the subject.

Thanks,

Fabio