Re: [PATCH] Fix check for valid stack pointer during backtrace

[tsbogend@xxxxxxxxxxxxxxxx (Thomas Bogendoerfer)]

On Mon, May 12, 2008 at 11:24:21PM +0900, Atsushi Nemoto wrote:
> > -	unsigned long *sp = (unsigned long *)(reg29 & ~3);
> > +	unsigned long __user *sp = (unsigned long __user *)(reg29 & ~3);
> 
> Please drop this change, while kstack_end expect non __user pointer.

arg, of course. Thanks.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessary a
good idea.                                                [ RFC1925, 2.3 ]