Re: [PATCH 2/8] KVM: x86: extend struct kvm_vcpu_pv_apf_data with token info

From: Vitaly Kuznetsov
Date: Sat May 23 2020 - 12:34:33 EST


Vivek Goyal <vgoyal@xxxxxxxxxx> writes:

> On Mon, May 11, 2020 at 06:47:46PM +0200, Vitaly Kuznetsov wrote:
>> Currently, APF mechanism relies on the #PF abuse where the token is being
>> passed through CR2. If we switch to using interrupts to deliver page-ready
>> notifications we need a different way to pass the data. Extent the existing
>> 'struct kvm_vcpu_pv_apf_data' with token information for page-ready
>> notifications.
>>
>> The newly introduced apf_put_user_ready() temporary puts both reason
>> and token information, this will be changed to put token only when we
>> switch to interrupt based notifications.
>>
>> Signed-off-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
>> ---
>> arch/x86/include/uapi/asm/kvm_para.h | 3 ++-
>> arch/x86/kvm/x86.c | 17 +++++++++++++----
>> 2 files changed, 15 insertions(+), 5 deletions(-)
>>
>> diff --git a/arch/x86/include/uapi/asm/kvm_para.h b/arch/x86/include/uapi/asm/kvm_para.h
>> index 2a8e0b6b9805..e3602a1de136 100644
>> --- a/arch/x86/include/uapi/asm/kvm_para.h
>> +++ b/arch/x86/include/uapi/asm/kvm_para.h
>> @@ -113,7 +113,8 @@ struct kvm_mmu_op_release_pt {
>>
>> struct kvm_vcpu_pv_apf_data {
>> __u32 reason;
>
> Hi Vitaly,
>
> Given we are redoing it, can we convert "reason" into a flag instead
> and use bit 0 for signalling "page not present" Then rest of the 31
> bits can be used for other purposes. I potentially want to use one bit to
> signal error (if it is known at the time of injecting #PF).

Yes, I think we can do that. The existing KVM_PV_REASON_PAGE_READY and
KVM_PV_REASON_PAGE_NOT_PRESENT are mutually exclusive and can be
converted to flags (we'll only have KVM_PV_REASON_PAGE_NOT_PRESENT in
use when this series is merged).

>
>> - __u8 pad[60];
>> + __u32 pageready_token;
>> + __u8 pad[56];
>
> Given token is 32 bit, for returning error in "page ready" type messages,
> I will probably use padding bytes and create pagready_flag and use one
> of the bits to signal error.

In case we're intended to pass more data in synchronous notifications,
shall we leave some blank space after 'flags' ('reason' previously) and
before 'token'?

--
Vitaly