On Sat, Jun 30, 2007 at 08:09:47PM +0900, YAMAMOTO Takashi wrote:
> > Index: i386/locore.S
> > ===================================================================
> > RCS file: /cvsroot/src/sys/arch/xen/i386/locore.S,v
> > retrieving revision 1.25
> > diff -u -r1.25 locore.S
> > --- i386/locore.S	17 May 2007 14:51:35 -0000	1.25
> > +++ i386/locore.S	28 Jun 2007 13:27:49 -0000
> > @@ -663,6 +663,7 @@
> >  	 * Switch to newlwp's stack.
> >  	 */
> >  
> > +	CLI(%ebx)
> >  	movl	L_ADDR(%edi),%ebx
> >  	movl	PCB_EBP(%ebx),%ebp
> >  	movl	PCB_ESP(%ebx),%esp
> 
> can you explain why it's necessary?

I didn't find why an interrupt coming in there would cause problem,
but I can confirm that this prevent the KASSERT panic that was being reported.
Also, I notice that the code in netbsd-4 does the process stack swicthing
with interrupts disabled.

> 
> anyway, please don't call ras_lookup with interrupt disabled.

OK. I just wanted to make sure the new context was fully installed before
re-taking interrupts. I didn't notice a C function was being called.

> 
> > @@ -780,9 +789,29 @@
> >  	call	_C_LABEL(trap)
> >  	addl	$4,%esp
> >  	jmp	.Lsyscall_checkast
> > -1:	STI(%eax)
> > -	CHECK_DEFERRED_SWITCH(%eax)
> > +1:	CHECK_DEFERRED_SWITCH(%eax)
> >  	jnz	9f
> > +	STIC(%eax)
> > +	jz	14f
> > +	call	_C_LABEL(stipending)
> > +	testl	%eax,%eax
> > +	jz	14f
> > +	/* process pending interrupts */
> > +	CLI(%eax)
> > +	movl	CPUVAR(ILEVEL), %ebx
> > +	movl	$.Lsyscall_resume, %esi # address to resume loop at
> > +.Lsyscall_resume:
> > +	movl	%ebx,%eax		# get cpl
> > +	movl	CPUVAR(IUNMASK)(,%eax,4),%eax
> > +	andl	CPUVAR(IPENDING),%eax	# any non-masked bits left?
> > +	jz	17f
> > +	bsrl	%eax,%eax
> > +	btrl	%eax,CPUVAR(IPENDING)
> > +	movl	CPUVAR(ISOURCES)(,%eax,4),%eax
> > +	jmp	*IS_RESUME(%eax)
> > +17:	movl	%ebx, CPUVAR(ILEVEL)	#restore cpl
> > +	jmp	.Lsyscall_checkast
> > +14:
> >  #ifndef DIAGNOSTIC
> >  	INTRFASTEXIT
> >  #else /* DIAGNOSTIC */
> 
> can't these duplicated code be shared?

it could probably, but this would make some more call/return, or jump.
I'm not sure it's worth it.

> 
> i'm not sure why stipending() was needed in the first place,
> in addition to __sti() and do_hypervisor_callback().

__sti() is intended for C code, but it takes an extrat trap and 4 context
switches to process pending interrupts. processing interrupts in the current
context it more efficient.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--