Re: 2.6.35-rc1 regression with pvclock and smp guests

[Zachary Amsden <zamsden@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1875 bytes --]

On 09/30/2010 01:07 PM, Michael Tokarev wrote:
> 01.10.2010 03:02, Michael Tokarev wrote:
>    
>> 30.09.2010 23:05, Marcelo Tosatti wrote:
>> []
>>      
>>> Arjan, Michael, can you try the following:
>>>
>>>  From 3823c018162dc708b543cbdc680a4c7d63533fee Mon Sep 17 00:00:00 2001
>>> From: Zachary Amsden<zamsden@redhat.com>
>>> Date: Sat, 29 May 2010 17:52:46 -1000
>>> Subject: [KVM V2 04/25] Fix SVM VMCB reset
>>> Cc: Avi Kivity<avi@redhat.com>,
>>>      Marcelo Tosatti<mtosatti@redhat.com>,
>>>      Glauber Costa<glommer@redhat.com>,
>>>      linux-kernel@vger.kernel.org
>>>
>>> On reset, VMCB TSC should be set to zero.  Instead, code was setting
>>> tsc_offset to zero, which passes through the underlying TSC.
>>>
>>> Signed-off-by: Zachary Amsden<zamsden@redhat.com>
>>> ---
>>>   arch/x86/kvm/svm.c |    2 +-
>>>   1 files changed, 1 insertions(+), 1 deletions(-)
>>>
>>> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
>>> index 760c86e..46856d2 100644
>>> --- a/arch/x86/kvm/svm.c
>>> +++ b/arch/x86/kvm/svm.c
>>> @@ -781,7 +781,7 @@ static void init_vmcb(struct vcpu_svm *svm)
>>>
>>>   	control->iopm_base_pa = iopm_base;
>>>   	control->msrpm_base_pa = __pa(svm->msrpm);
>>> -	control->tsc_offset = 0;
>>> +	guest_write_tsc(&svm->vcpu, 0);
>>>   	control->int_ctl = V_INTR_MASKING_MASK;
>>>        
>> This fails to compile on 2.6.35.5:
>>
>> arch/x86/kvm/svm.c: In function ‘init_vmcb’:
>> arch/x86/kvm/svm.c:769: error: implicit declaration of function ‘guest_write_tsc’
>>
>> I'll take a look tomorrow where that comes from.. hopefully ;)
>>      
> Ok, that routine is static, defined in arch/x86/kvm/vmx.c
> (not svm.c).  I'm not sure it's ok to use it in svm.c
> directly, as it appears to be vmx-specific.
>
> Thanks!
>
> /mjt
>    


Can you try this patch to see if it helps?  I believe it is also safe 
for Xen, but cc'ing to double check.

[-- Attachment #2: kvmclock-fix-hack-1.patch --]
[-- Type: text/plain, Size: 807 bytes --]

Try to fix setup_percpu_clockdev by moving it before interrupts
are enabled.

Signed-off-by: Zachary Amsden <zamsden@redhat.com>

diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 8b3bfc4..40a383b 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -351,6 +351,8 @@ notrace static void __cpuinit start_secondary(void *unused)
 	unlock_vector_lock();
 	ipi_call_unlock();
 	per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;
+	x86_cpuinit.setup_percpu_clockev();
+
 	x86_platform.nmi_init();
 
 	/* enable local interrupts */
@@ -359,8 +361,6 @@ notrace static void __cpuinit start_secondary(void *unused)
 	/* to prevent fake stack check failure in clock setup */
 	boot_init_stack_canary();
 
-	x86_cpuinit.setup_percpu_clockev();
-
 	wmb();
 	cpu_idle();
 }