Re: [PATCH 3/4] x86/asm: delete dummy variables in movdir64b()

["H. Peter Anvin" <hpa@zytor.com>]

On March 7, 2025 8:15:42 AM PST, Alexey Dobriyan <adobriyan@gmail.com> wrote:
>On Fri, Mar 07, 2025 at 12:49:26PM +0100, Ingo Molnar wrote:
>> 
>> * Alexey Dobriyan <adobriyan@gmail.com> wrote:
>> 
>> > Cast to pointer-to-array instead.
>> > 
>> > Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
>> > ---
>> >  arch/x86/include/asm/special_insns.h | 9 +++------
>> >  1 file changed, 3 insertions(+), 6 deletions(-)
>> > 
>> > diff --git a/arch/x86/include/asm/special_insns.h b/arch/x86/include/asm/special_insns.h
>> > index d349aa0f0a83..b24c6c945c38 100644
>> > --- a/arch/x86/include/asm/special_insns.h
>> > +++ b/arch/x86/include/asm/special_insns.h
>> > @@ -215,13 +215,10 @@ static __always_inline void serialize(void)
>> >  /* The dst parameter must be 64-bytes aligned */
>> >  static inline void movdir64b(void *dst, const void *src)
>> >  {
>> > -	const struct { char _[64]; } *__src = src;
>> > -	struct { char _[64]; } *__dst = dst;
>> > -
>> >  	/*
>> >  	 * MOVDIR64B %(rdx), rax.
>> >  	 *
>> > -	 * Both __src and __dst must be memory constraints in order to tell the
>> > +	 * Both src and dst must be memory constraints in order to tell the
>> >  	 * compiler that no other memory accesses should be reordered around
>> >  	 * this one.
>> >  	 *
>> > @@ -230,8 +227,8 @@ static inline void movdir64b(void *dst, const void *src)
>> >  	 * I.e., not the pointers but what they point to, thus the deref'ing '*'.
>> >  	 */
>> >  	asm volatile(".byte 0x66, 0x0f, 0x38, 0xf8, 0x02"
>> > -		     : "+m" (*__dst)
>> > -		     :  "m" (*__src), "a" (__dst), "d" (__src));
>> > +		     : "+m" (*(char(*)[64])dst)
>> > +		     :  "m" (*(const char(*)[64])src), "a" (dst), "d" (src));
>> 
>> In what world is putting type casts inside asm() statements an 
>> improvement to the code?
>
>In the same world where creating distracting variable whose only purpose
>is to make a cast is considered not good.
>
>Notice the cast is shorter, there is not "struct", so it is positive in
>both vertical and horizontal directions.

Stuffing more into a single statement – and asm statements are invariably complex – is not good for clarity. Please, no.