[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAFULd4YrDiu7zUdZXRkLNiN=6zC4hPWchymrbMA=ag3qG+OarA@mail.gmail.com>
Date: Fri, 22 Mar 2024 11:37:31 +0100
From: Uros Bizjak <ubizjak@...il.com>
To: Alexey Dobriyan <adobriyan@...il.com>
Cc: tglx@...utronix.de, mingo@...hat.com, bp@...en8.de,
dave.hansen@...ux.intel.com, x86@...nel.org, linux-kernel@...r.kernel.org,
hpa@...or.com
Subject: Re: [PATCH 1/1] x86_64: cleanup memset16(), memset32(), memset64()
On Fri, Mar 22, 2024 at 11:05 AM Alexey Dobriyan <adobriyan@...il.com> wrote:
>
> * use "+" constraint modifier,
> simplify inputs and output lists,
> delete dummy variables with meaningless names,
>
> "&" only makes sense in complex assembly creating constraints on
> intermediate registers. But 1 instruction assemblies don't have
> inner body so to speak.
>
> * write "rep stos*" on one line:
> Rep prefix is integral part of x86 instruction. I'm not sure why
> people separate "rep" with newline or semicolon pretending rep is
> separate instruction. It is not and never was.
The reason for this was that some archaic assemblers rejected "rep
insn" on one line. I have checked that the minimum required
binutils-2.25 assembles this without problems.
> * use __auto_type for maximum copy pasta experience,
>
> * reformat a bit to make everything looks nicer.
>
> Note that "memory" clobber is too much if "n" is known at compile time.
> However,
>
> "=m" (*(T(*)[n])s)
>
> doesn't work because -Wvla even if "n" is compile time constant:
>
> if (BCP(n)) {
> rep stos
> : "=m" (*(T(*)[n])s)
> } else {
> rep stosw
> : "memory"
> }
>
> The above doesn't work.
>
> Signed-off-by: Alexey Dobriyan <adobriyan@...il.com>
Reviewed-by: Uros Bizjak <ubizjak@...il.com>
> ---
> arch/x86/include/asm/string_64.h | 45 +++++++++++++++++---------------
> 1 file changed, 24 insertions(+), 21 deletions(-)
>
> diff --git a/arch/x86/include/asm/string_64.h b/arch/x86/include/asm/string_64.h
> index 857d364b9888..9d0b324eab21 100644
> --- a/arch/x86/include/asm/string_64.h
> +++ b/arch/x86/include/asm/string_64.h
> @@ -30,37 +30,40 @@ void *__memset(void *s, int c, size_t n);
> #define __HAVE_ARCH_MEMSET16
> static inline void *memset16(uint16_t *s, uint16_t v, size_t n)
> {
> - long d0, d1;
> - asm volatile("rep\n\t"
> - "stosw"
> - : "=&c" (d0), "=&D" (d1)
> - : "a" (v), "1" (s), "0" (n)
> - : "memory");
> - return s;
> + const __auto_type s0 = s;
> + asm volatile (
> + "rep stosw"
> + : "+D" (s), "+c" (n)
> + : "a" (v)
> + : "memory"
> + );
> + return s0;
> }
>
> #define __HAVE_ARCH_MEMSET32
> static inline void *memset32(uint32_t *s, uint32_t v, size_t n)
> {
> - long d0, d1;
> - asm volatile("rep\n\t"
> - "stosl"
> - : "=&c" (d0), "=&D" (d1)
> - : "a" (v), "1" (s), "0" (n)
> - : "memory");
> - return s;
> + const __auto_type s0 = s;
> + asm volatile (
> + "rep stosl"
> + : "+D" (s), "+c" (n)
> + : "a" (v)
> + : "memory"
> + );
> + return s0;
> }
>
> #define __HAVE_ARCH_MEMSET64
> static inline void *memset64(uint64_t *s, uint64_t v, size_t n)
> {
> - long d0, d1;
> - asm volatile("rep\n\t"
> - "stosq"
> - : "=&c" (d0), "=&D" (d1)
> - : "a" (v), "1" (s), "0" (n)
> - : "memory");
> - return s;
> + const __auto_type s0 = s;
> + asm volatile (
> + "rep stosq"
> + : "+D" (s), "+c" (n)
> + : "a" (v)
> + : "memory"
> + );
> + return s0;
> }
> #endif
>
> --
> 2.43.2
>
Powered by blists - more mailing lists