[PATCH v1.1] Randomize memcpy benchmark addresses.

[Siddhesh Poyarekar]

On Thu, Sep 05, 2013 at 08:41:02AM +0100, Will Newton wrote:
> >
> > -  if (memcmp (dst, src, len) != 0)
> > -    {
> > -      error (0, 0, "Wrong result in function %s dst \"%s\" src \"%s\"",
> > -            impl->name, dst, src);
> > -      ret = 1;
> > -      return;
> > -    }
> 
> I am not sure how useful this sanity check is. Obviously the glibc
> memcpy implementation is tested elsewhere but the simple
> implementation isn't.
> 

It is here because the string benchmarks were dual purpose before (in
string/test-*) and I had just copied them here in benchtests as a
starting point.  The sanity check could be dropped (since it should be
there in string/test-*) but it should be a separate patch since it is
not related in any way to the randomization patch.

Siddhesh