Re: [PATCH] svm/avic: Fix invalidate logical APIC id entry

["Suthikulpanit, Suravee" <Suravee.Suthikulpanit@amd.com>]

On 3/22/19 8:04 PM, Vitaly Kuznetsov wrote:
> "Suthikulpanit, Suravee" <Suravee.Suthikulpanit@amd.com> writes:
> 
>> Only clear the valid bit when invalidate logical APIC id entry.
>> The current logic clear the valid bit, but also set the rest of
>> the bits (including reserved bits) to 1.
>>
>> Fixes: 98d90582be2e ('svm: Fix AVIC DFR and LDR handling')
>> Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
>> ---
>>   arch/x86/kvm/svm.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
>> index 7a4ce1a22ca0..f4fb766e474c 100644
>> --- a/arch/x86/kvm/svm.c
>> +++ b/arch/x86/kvm/svm.c
>> @@ -4640,7 +4640,7 @@ static void avic_invalidate_logical_id_entry(struct kvm_vcpu *vcpu)
>>   	u32 *entry = avic_get_logical_id_entry(vcpu, svm->ldr_reg, flat);
>>   
>>   	if (entry)
>> -		WRITE_ONCE(*entry, (u32) ~AVIC_LOGICAL_ID_ENTRY_VALID_MASK);
>> +		WRITE_ONCE(*entry, (u32)(*entry & ~AVIC_LOGICAL_ID_ENTRY_VALID_MASK));
> 
> I'm not sure how important this is, but this change creates a tiny time
> window between reading from *entry and writing there. Should we use
> atomic bitops instead?

Sure. That is also simpler. I'll send out V2 w/ clear_bit() instead.

Suravee

> 
>>   }
>>   
>>   static int avic_handle_ldr_update(struct kvm_vcpu *vcpu)
>