lkml.org 
[lkml]   [2022]   [Jul]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH] KVM: x86: Do not block APIC write for non ICR registers
From
Paolo / Sean,

Please let me know if you want me to send v2 with changes proposed by Sean.

Regards,
Suravee

On 7/19/22 10:24 AM, Suthikulpanit, Suravee wrote:
> Sean,
>
> On 7/19/2022 12:28 AM, Sean Christopherson wrote:
>> On Mon, Jul 18, 2022, Suravee Suthikulpanit wrote:
>>> The commit 5413bcba7ed5 ("KVM: x86: Add support for vICR APIC-write
>>> VM-Exits in x2APIC mode") introduces logic to prevent APIC write
>>> for offset other than ICR. This breaks x2AVIC support, which requires
>>> KVM to trap and emulate x2APIC MSR writes.
>>>
>>> Therefore, removes the warning and modify to logic to allow MSR write.
>>>
>>> Fixes: 5413bcba7ed5 ("KVM: x86: Add support for vICR APIC-write VM-Exits in x2APIC mode")
>>
>> This tag is wrong, I believe it should be:
>>
>>    Fixes: 4d1d7942e36a ("KVM: SVM: Introduce logic to (de)activate x2AVIC mode")
>>
>> And that absolutely matters because this should not be backported to older
>> kernels that don't support x2avic.
>
> The commit 5413bcba7ed5 is the one that modifies the logic in the kvm_apic_write_nodecode().
> I understand your point that the 5413bcba7ed is committed later than 4d1d7942e36a and being
> affected by the change. However, if there is a case that only x2AVIC stuff is being backported
> w/o the virtualize IPI stuff, then this fix is not needed. Hence, I would say the fix is for
> the 5413bcba7ed5 as specified in the original patch.
>
>>> .....
>>>            */
>>> -        if (WARN_ON_ONCE(offset != APIC_ICR))
>>> +        if (offset == APIC_ICR) {
>>> +            kvm_apic_send_ipi(apic, (u32)val, (u32)(val >> 32));
>>> +            trace_kvm_apic_write(APIC_ICR, val);
>>>               return;
>>> -
>>> -        kvm_lapic_msr_read(apic, offset, &val);
>>> -        kvm_apic_send_ipi(apic, (u32)val, (u32)(val >> 32));
>>> -        trace_kvm_apic_write(APIC_ICR, val);
>>> +        }
>>> +        kvm_lapic_msr_write(apic, offset, val);
>>
>> Because this lacks the TODO below, what about tweaking this so that there's a
>> single call to kvm_lapic_msr_write()?  gcc-11 even generates more efficient code
>> for this.  Alternatively, the ICR path can be an early return inside a single
>> x2APIC check, but gcc generate identical code and I like making x2APIC+ICR stand
>> out as being truly special.
>
> That sounds good.
>
>> Compile tested only.
>>
>> ---
>> From: Sean Christopherson <seanjc@google.com>
>> Date: Mon, 18 Jul 2022 10:16:02 -0700
>> Subject: [PATCH] KVM: x86: Handle trap-like x2APIC accesses for any APIC
>>   register
>>
>> Handle trap-like VM-Exits for all APIC registers when the guest is in
>> x2APIC mode and drop the now-stale WARN that KVM encounters trap-like
>> exits only for ICR.  On Intel, only writes to ICR can be trap-like when
>> APICv and x2APIC are enabled, but AMD's x2AVIC can trap more registers,
>> e.g. LDR and DFR.
>>
>> Fixes: 4d1d7942e36a ("KVM: SVM: Introduce logic to (de)activate x2AVIC mode")
>> Reported-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
>> Cc: Zeng Guang <guang.zeng@intel.com>
>> Cc: Maxim Levitsky <mlevitsk@redhat.com>
>> Signed-off-by: Sean Christopherson <seanjc@google.com>
>> ---
>>   arch/x86/kvm/lapic.c | 21 ++++++++++-----------
>>   1 file changed, 10 insertions(+), 11 deletions(-)
>>
>> diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
>> index 9d4f73c4dc02..95bb1ef37a12 100644
>> --- a/arch/x86/kvm/lapic.c
>> +++ b/arch/x86/kvm/lapic.c
>> @@ -2283,21 +2283,20 @@ void kvm_apic_write_nodecode(struct kvm_vcpu *vcpu, u32 offset)
>>       struct kvm_lapic *apic = vcpu->arch.apic;
>>       u64 val;
>>
>> -    if (apic_x2apic_mode(apic)) {
>> -        /*
>> -         * When guest APIC is in x2APIC mode and IPI virtualization
>> -         * is enabled, accessing APIC_ICR may cause trap-like VM-exit
>> -         * on Intel hardware. Other offsets are not possible.
>> -         */
>> -        if (WARN_ON_ONCE(offset != APIC_ICR))
>> -            return;
>> -
>> +    if (apic_x2apic_mode(apic))
>>           kvm_lapic_msr_read(apic, offset, &val);
>> +    else
>> +        val = kvm_lapic_get_reg(apic, offset);
>> +
>> +    /*
>> +     * ICR is a single 64-bit register when x2APIC is enabled.  For legacy
>> +     * xAPIC, ICR writes need to go down the common (slightly slower) path
>> +     * to get the upper half from ICR2.
>> +     */
>> +    if (apic_x2apic_mode(apic) && offset == APIC_ICR) {
>>           kvm_apic_send_ipi(apic, (u32)val, (u32)(val >> 32));
>>           trace_kvm_apic_write(APIC_ICR, val);
>>       } else {
>> -        val = kvm_lapic_get_reg(apic, offset);
>> -
>>           /* TODO: optimize to just emulate side effect w/o one more write */
>>           kvm_lapic_reg_write(apic, offset, (u32)val);
>>       }
>>
>> base-commit: 8031d87aa9953ddeb047a5356ebd0b240c30f233
>> --
>
> Tested-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
>
> Suravee

\
 
 \ /
  Last update: 2022-07-20 11:52    [W:0.185 / U:0.028 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site