Messages in this thread | | | Date | Fri, 3 Jul 2009 20:43:53 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH -tip] x86: atomic64: inline atomic64_read() |
| |
* Arjan van de Ven <arjan@infradead.org> wrote:
> On Fri, 3 Jul 2009 20:18:15 +0200 > Ingo Molnar <mingo@elte.hu> wrote: > > > > > * Arjan van de Ven <arjan@infradead.org> wrote: > > > > > On Fri, 3 Jul 2009 20:04:50 +0200 > > > Ingo Molnar <mingo@elte.hu> wrote: > > > > > > > > +static inline u64 atomic64_read(atomic64_t *ptr) > > > > > +{ > > > > > + u64 res; > > > > > + > > > > > + asm volatile( > > > > > + "mov %%ebx, %%eax\n\t" > > > > > + "mov %%ecx, %%edx\n\t" > > > > > + LOCK_PREFIX "cmpxchg8b %1\n" > > > > > > also afaik cmpxchg implies the lock prefix, so no need to make an > > > explicit one... it just adds code size and alternatives-work > > > > good point! > > > > Is the same true of XADD as well? If yes then we can remove the lock > > prefix from those as well - there's a few uses of that on 32-bit and > > 64-bit as well. > > > > ok never mind I was wrong, > it seems cmpxchg does NOT imply a lock prefix, nor does xadd.
Indeed, i just checked the SDM and the lock prefix is not implicit.
Ignore the patch i sent out ;-)
Ingo
| |