Calling migrate_timer from context_switch() seems to introduce
unnecessary overhead.
Why did you choose to insert migrate_timer() to context_switch()
instead of inserting it ot the following position?
diff -r 8c6bb45901e7 xen/arch/ia64/xen/hypercall.c
--- a/xen/arch/ia64/xen/hypercall.c Wed Aug 16 14:28:57 2006 -0600
+++ b/xen/arch/ia64/xen/hypercall.c Mon Aug 21 13:46:05 2006 +0900
@@ -233,7 +233,10 @@ fw_hypercall (struct pt_regs *regs)
}
else {
perfc_incrc(pal_halt_light);
- do_sched_op_compat(SCHEDOP_yield, 0);
<<<<<<<<<<<<< migrate_timer() >>>>>>>>>>>>>>>>>>>>>>>>>>>
+ set_timer(&v->arch.hlt_timer,
+ vcpu_get_next_timer_ns(v));
+ do_sched_op_compat(SCHEDOP_block, 0);
+ stop_timer(&v->arch.hlt_timer);
}
regs->r8 = 0;
regs->r9 = 0;
On Wed, Aug 23, 2006 at 07:29:11PM +0900, Atsushi SAKAI wrote:
> Hi, Isaku
>
> Sorry for confusing.
> It should replace from "for context_switch" to "to context_switch"
> migrate_timer is in context_switch.
>
> Thanks
> Atsushi
>
> >Hi Atsushi.
> >
> >On Wed, Aug 23, 2006 at 05:48:15PM +0900, Atsushi SAKAI wrote:
> >
> >> 1)migrate_timer for hlt_timer_fn is added for context_switch
> >> This makes correct pCPU work for timer.
> >
> >Is it necessary to call migrate_timer() every context switch
> >instead of calling it right before set_timer(&hlt_timer)?
> >
> >--
> >yamahata
> >
>
>
>
>
>
>
>
> _______________________________________________
> Xen-ia64-devel mailing list
> Xen-ia64-devel@lists.xensource.com
> http://lists.xensource.com/xen-ia64-devel
--
yamahata
_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel
