On Fri, 27 Oct 2006 14:31:41 -0700 Jeremy Fitzhardinge <jeremy at goop.org> wrote: > Pavel Machek wrote: > > Indentation is b0rken here. > > > > Oops. How strange. > > > And... is get_user right primitive for accessing area that may not be > > there? > > I'm pretty sure there's precedent for using __get_user in this way > (get_user is a different matter, since it cares about whether the > address is within the user part of the address space). Certainly in > arch/i386 code there shouldn't be a problem. Is there some other way to > achieve the same effect (without manually setting up an exception/fixup > block)? It'd be better to use include/linux/uaccess.h:probe_kernel_address() for this operation.