John Levon wrote:
> On Thu, Mar 25, 2004 at 03:25:28PM -0500, Will Cohen wrote:
>>@@ -68,7 +74,7 @@
>> static int valid_kernel_stack(struct frame_head * head, struct pt_regs * regs)
>> unsigned long headaddr = (unsigned long)head;
>>- unsigned long stack = (unsigned long)regs;
>>+ unsigned long stack = (unsigned long)®s;
> What's this hunk about ?
I should have edited this out of the patch. However, I do wonder if the
original test is correct. How would examining what regs points at give
information about the stack? Is the struct regs points to pushed on the
stack? Looking at the code in arch/i386/kernel/traps.c and code
include/linux/sched.h for kstack_end the testing looked a lot different
than this valid_kernel_stack.