Re: [PATCH v8 02/17] Add cmpxchg16b support for x86-64

["Jan Beulich" <JBeulich@suse.com>]

>>> On 12.10.15 at 10:54, <feng.wu@intel.com> wrote:
> --- a/xen/include/asm-x86/x86_64/system.h
> +++ b/xen/include/asm-x86/x86_64/system.h
> @@ -6,6 +6,39 @@
>                                     (unsigned long)(n),sizeof(*(ptr))))
>  
>  /*
> + * Atomic 16 bytes compare and exchange.  Compare OLD with MEM, if
> + * identical, store NEW in MEM.  Return the initial value in MEM.
> + * Success is indicated by comparing RETURN with OLD.
> + *
> + * This function can only be called when cpu_has_cx16 is true.
> + */
> +
> +static always_inline __uint128_t __cmpxchg16b(
> +    volatile void *ptr, const __uint128_t *old, const __uint128_t *new)
> +{
> +    __uint128_t prev;
> +    uint64_t new_high = *new >> 64;
> +    uint64_t new_low = *new;
> +
> +    ASSERT(cpu_has_cx16);
> +
> +    asm volatile ( "lock; cmpxchg16b %3"
> +                   : "=A" (prev)
> +                   : "c" (new_high), "b" (new_low),
> +                     "m" (*__xg(ptr)), "0" (*old) );

I was about to apply this when I spotted that this is still wrong:
The (requested) removal of the memory clobber requires that
the memory location (all 16 bytes of it) appear as output. With
__xg() covering 100 bytes, it should be possible to simply make
the operand an in- and output ("+m") instead of just an input.

Jan