[Xen-ia64-devel][PATCH] found a small bugRE: [Xen-ia64-devel] [PATCH] pal_halt_light emulatefor domU TAKE3

["Xu, Anthony" <anthony.xu@xxxxxxxxx>]

Alex/Sakai,

I found a small bug in pal_halt_light emulator 
when I investigated VTI-domain hang issue.

When init_timer is called in alloc_vcpu_struct,
v->processor is not initialized.
     
Signed-off-by: Anthony Xu < anthony.xu@xxxxxxxxx >

--Anthony

>-----Original Message-----
>From: Atsushi SAKAI [mailto:sakaia@xxxxxxxxxxxxxx]
>Sent: 2006?8?24? 9:50
>To: Isaku Yamahata; Xu, Anthony
>Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
>Subject: Re: [Xen-ia64-devel] [PATCH] pal_halt_light emulatefor domU TAKE3
>
>Status
>
>I already started testing yamahata's suggested code from yesterday night.
>But it does not come to 16h (currently 13h30m running is passed at 10:30
>JST(+9:00))
>Test Configuration is 2xDomU Linux Kernel Compile Test on 4xpCPU w/ credit
>scheduler.
>
>1)migrate_timer for hlt_timer_fn is added at fw_hypercall(pal_halt_light)
>  This makes correct pCPU work for timer.
>
>2)To avoid soft lockup message(10sec timer stop) at Dom0/U,
>  I add the condition for checking in vcpu_get_next_timer_ns.
>  domain_itm > ia64_get_itc
>  If above condition is not satisfied
>  Just set the timer interval itm_delta from NOW().
>
>3)Remove vcpu_pend_timer from hlt_timer_fn.
>  It previously makes the state that vcpu is running but timer(VIRQ_ITC) is
>stopped.
>
>(References)
>about migrate_timer position
>http://lists.xensource.com/archives/html/xen-ia64-devel/2006-08/msg00324.ht
>ml
>Yamahata's for VTI migrate timer.
>http://lists.xensource.com/archives/html/xen-ia64-devel/2006-07/msg00375.ht
>ml
>About vcpu_get_next_timer_ns
>http://lists.xensource.com/archives/html/xen-ia64-devel/2006-08/msg00290.ht
>ml
>stop_timer
>http://lists.xensource.com/archives/html/xen-ia64-devel/2006-07/msg00171.ht
>ml
>
>
>Signed-off-by: Atsushi SAKAI <sakaia@xxxxxxxxxxxxxx>
>
>
>
>
>Thanks
>Atsushi SAKAI
>
>
>
>>On Wed, Aug 23, 2006 at 09:32:24PM +0800, Xu, Anthony wrote:
>>
>>> I agree with Yamahata partially,
>>> If I'm right, credit scheduler only schedules runnable vcpus,
>>> So when context_switch is called, the scheduled in vcpu's hlt_timer must be
>>> Stopped, it is unnecessary to call migrate_timer in context_switch.
>>
>>Agreed.
>>
>>--
>>yamahata
>>
>
>
>
>

Attachment: pal_halt_light_fix.patch
Description: pal_halt_light_fix.patch

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel