On 26/11/18 22:05, Jim Mattson wrote:
> On Wed, May 30, 2018 at 9:33 AM Konrad Rzeszutek Wilk
> <konrad.wilk@xxxxxxxxxx> wrote:
>>
>> On Tue, May 22, 2018 at 09:54:20AM -0700, Jim Mattson wrote:
>>> Previously, we only called indirect_branch_prediction_barrier on the
>>> logical CPU that freed a vmcb. This function should be called on all
>>> logical CPUs that last loaded the vmcb in question.
>>>
>>> Fixes: 15d45071523d ("KVM/x86: Add IBPB support")
>>> Reported-by: Neel Natu <neelnatu@xxxxxxxxxx>
>>> Signed-off-by: Jim Mattson <jmattson@xxxxxxxxxx>
>> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
>>
>> Thank you!
>>> ---
>>> arch/x86/kvm/svm.c | 20 +++++++++++++++-----
>>> 1 file changed, 15 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
>>> index 220e5a89465a..ffa27f75e323 100644
>>> --- a/arch/x86/kvm/svm.c
>>> +++ b/arch/x86/kvm/svm.c
>>> @@ -2172,21 +2172,31 @@ static struct kvm_vcpu *svm_create_vcpu(struct kvm *kvm, unsigned int id)
>>> return ERR_PTR(err);
>>> }
>>>
>>> +static void svm_clear_current_vmcb(struct vmcb *vmcb)
>>> +{
>>> + int i;
>>> +
>>> + for_each_online_cpu(i)
>>> + cmpxchg(&per_cpu(svm_data, i)->current_vmcb, vmcb, NULL);
>>> +}
>>> +
>>> static void svm_free_vcpu(struct kvm_vcpu *vcpu)
>>> {
>>> struct vcpu_svm *svm = to_svm(vcpu);
>>>
>>> + /*
>>> + * The vmcb page can be recycled, causing a false negative in
>>> + * svm_vcpu_load(). So, ensure that no logical CPU has this
>>> + * vmcb page recorded as its current vmcb.
>>> + */
>>> + svm_clear_current_vmcb(svm->vmcb);
>>> +
>>> __free_page(pfn_to_page(__sme_clr(svm->vmcb_pa) >> PAGE_SHIFT));
>>> __free_pages(virt_to_page(svm->msrpm), MSRPM_ALLOC_ORDER);
>>> __free_page(virt_to_page(svm->nested.hsave));
>>> __free_pages(virt_to_page(svm->nested.msrpm), MSRPM_ALLOC_ORDER);
>>> kvm_vcpu_uninit(vcpu);
>>> kmem_cache_free(kvm_vcpu_cache, svm);
>>> - /*
>>> - * The vmcb page can be recycled, causing a false negative in
>>> - * svm_vcpu_load(). So do a full IBPB now.
>>> - */
>>> - indirect_branch_prediction_barrier();
>>> }
>>>
>>> static void svm_vcpu_load(struct kvm_vcpu *vcpu, int cpu)
>>> --
>>> 2.17.0.441.gb46fe60e1d-goog
>>>
>
> Ping.
>
Queued, thanks.
Paolo
