On Tue, Mar 12, 2013 at 09:00:53PM +0800, Zhi-zhou Zhang wrote: > Here should return 64-bit types rather than 32-bit types. Or we > may get wrong return value if high 32-bit isn't equal to zero. David Daney's patch which I've applied not two hours before you posted this patch solves the same issue though a different way. So I'm just going to drop this patch. Thanks anyway! Ralf
