Re: [(resend)patch 3/7] Notify page fault call chain for ia64

[Keshavamurthy Anil S <anil.s.keshavamurthy@intel.com>]

On Thu, Apr 20, 2006 at 12:18:25PM +1000, Keith Owens wrote:
> Why is register_page_fault_notifier defined in traps.c?  Surely it
> should be in mm/fault.c, which is the only place that uses the chain.
Ah..that is because I blindly followed {register/unregister}_die_notifier()
implementation. It can moved to mm/faults.c and I will rework my patch.

> 
> > trap_init (void)
> > {
> >Index: linux-2.6.17-rc1-mm3/arch/ia64/mm/fault.c
> >===================================================================
> >--- linux-2.6.17-rc1-mm3.orig/arch/ia64/mm/fault.c
> >+++ linux-2.6.17-rc1-mm3/arch/ia64/mm/fault.c
> >@@ -84,7 +84,7 @@ ia64_do_page_fault (unsigned long addres
> > 	/*
> > 	 * This is to handle the kprobes on user space access instructions
> > 	 */
> >-	if (notify_die(DIE_PAGE_FAULT, "page fault", regs, code, TRAP_BRKPT,
> >+	if (notify_page_fault(DIE_PAGE_FAULT, "page fault", regs, code, TRAP_BRKPT,
> > 					SIGSEGV) == NOTIFY_STOP)
> > 		return;
> 
> Since this is a critical path, please remove all references to
> notify_page_fault() and its register functions when CONFIG_KPROBES=n.
Yes, I agree with you and thanks for your feedback.

-Anil