[PATCH] xen/x86: Don't BUG on CPU0 offlining

[PATCH] xen/x86: Don't BUG on CPU0 offlining

[Vitaly Kuznetsov]

CONFIG_BOOTPARAM_HOTPLUG_CPU0 allows to offline CPU0 but Xen HVM guests
BUG() in xen_teardown_timer(). Remove the BUG_ON(), this is probably a
leftover from ancient times when CPU0 hotplug was impossible, it works
just fine for HVM.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
- CPU0 hotplug is currently broken on x86, see
  https://lkml.org/lkml/2017/6/26/529
---
 arch/x86/xen/time.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
index a1895a8..1ecb05d 100644
--- a/arch/x86/xen/time.c
+++ b/arch/x86/xen/time.c
@@ -309,7 +309,6 @@ static irqreturn_t xen_timer_interrupt(int irq, void *dev_id)
 void xen_teardown_timer(int cpu)
 {
 	struct clock_event_device *evt;
-	BUG_ON(cpu == 0);
 	evt = &per_cpu(xen_clock_events, cpu).evt;
 
 	if (evt->irq >= 0) {
-- 
2.9.4

Re: [Xen-devel] [PATCH] xen/x86: Don't BUG on CPU0 offlining

[Konrad Rzeszutek Wilk]

On Mon, Jun 26, 2017 at 06:39:30PM +0200, Vitaly Kuznetsov wrote:
> CONFIG_BOOTPARAM_HOTPLUG_CPU0 allows to offline CPU0 but Xen HVM guests
> BUG() in xen_teardown_timer(). Remove the BUG_ON(), this is probably a
> leftover from ancient times when CPU0 hotplug was impossible, it works
> just fine for HVM.
> 
> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>


Perhaps:

BUG_ON(cpu == 0 && xen_pv_domain()); 

?

> ---
> - CPU0 hotplug is currently broken on x86, see
>   https://lkml.org/lkml/2017/6/26/529
> ---
>  arch/x86/xen/time.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/arch/x86/xen/time.c b/arch/x86/xen/time.c
> index a1895a8..1ecb05d 100644
> --- a/arch/x86/xen/time.c
> +++ b/arch/x86/xen/time.c
> @@ -309,7 +309,6 @@ static irqreturn_t xen_timer_interrupt(int irq, void *dev_id)
>  void xen_teardown_timer(int cpu)
>  {
>  	struct clock_event_device *evt;
> -	BUG_ON(cpu == 0);
>  	evt = &per_cpu(xen_clock_events, cpu).evt;
>  
>  	if (evt->irq >= 0) {
> -- 
> 2.9.4
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> https://lists.xen.org/xen-devel

Re: [Xen-devel] [PATCH] xen/x86: Don't BUG on CPU0 offlining

[Vitaly Kuznetsov]

Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> writes:

> On Mon, Jun 26, 2017 at 06:39:30PM +0200, Vitaly Kuznetsov wrote:
>> CONFIG_BOOTPARAM_HOTPLUG_CPU0 allows to offline CPU0 but Xen HVM guests
>> BUG() in xen_teardown_timer(). Remove the BUG_ON(), this is probably a
>> leftover from ancient times when CPU0 hotplug was impossible, it works
>> just fine for HVM.
>> 
>> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>
> Perhaps:
>
> BUG_ON(cpu == 0 && xen_pv_domain()); 
>

Linus says no BUG_ONs :-)

xen_pv_cpu_disable() has the following:

    if (cpu == 0)
            return -EBUSY;

as a protection so we won't get to xen_teardown_timer() but if you think
additional BUG_ON() protection is justified I'm definitely not against
adding it.

-- 
  Vitaly

Re: [Xen-devel] [PATCH] xen/x86: Don't BUG on CPU0 offlining

[Juergen Groß]

On 06/28/2017 06:31 PM, Vitaly Kuznetsov wrote:
> Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> writes:
> 
>> On Mon, Jun 26, 2017 at 06:39:30PM +0200, Vitaly Kuznetsov wrote:
>>> CONFIG_BOOTPARAM_HOTPLUG_CPU0 allows to offline CPU0 but Xen HVM guests
>>> BUG() in xen_teardown_timer(). Remove the BUG_ON(), this is probably a
>>> leftover from ancient times when CPU0 hotplug was impossible, it works
>>> just fine for HVM.
>>>
>>> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
>>
>> Perhaps:
>>
>> BUG_ON(cpu == 0 && xen_pv_domain());
>>
> 
> Linus says no BUG_ONs :-)
> 
> xen_pv_cpu_disable() has the following:
> 
>      if (cpu == 0)
>              return -EBUSY;
> 
> as a protection so we won't get to xen_teardown_timer() but if you think
> additional BUG_ON() protection is justified I'm definitely not against
> adding it.

No, I don't think its needed. This should be handled at the appropriate
level, not down in the timer handling.

So:

Acked-by: Juergen Gross <jgross@suse.com>


Juergen

Re: [PATCH] xen/x86: Don't BUG on CPU0 offlining

[Juergen Gross]

On 26/06/17 18:39, Vitaly Kuznetsov wrote:
> CONFIG_BOOTPARAM_HOTPLUG_CPU0 allows to offline CPU0 but Xen HVM guests
> BUG() in xen_teardown_timer(). Remove the BUG_ON(), this is probably a
> leftover from ancient times when CPU0 hotplug was impossible, it works
> just fine for HVM.
> 
> Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>

Committed to xen/tip for-linus-4.13


Thanks,

Juergen