Samuel Thibault, le Fri 28 Mar 2014 16:33:39 +0100, a écrit :
> see a disassemble of memcpy, it ends up being a rep movsb
> %ds:(%esi),%es:(%edi), which is just exactly what we need to tell the
> processor).

Oops, sorry, wrong example, that one is coming from libc. Anyway,
relying on the host libc is not a good thing on the long run.  I'd
rather rely on the gcc __built_mem* etc.

Samuel

Reply via email to