From: Vitaly Kuznetsov <vkuznets@redhat.com>
To: Maxim Levitsky <mlevitsk@redhat.com>
Cc: Sean Christopherson <seanjc@google.com>,
Wanpeng Li <wanpengli@tencent.com>,
Jim Mattson <jmattson@google.com>,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH v3 0/4] KVM: x86: hyper-v: Conditionally allow SynIC with APICv/AVIC
Date: Mon, 14 Jun 2021 09:40:18 +0200 [thread overview]
Message-ID: <87zgvsx5b1.fsf@vitty.brq.redhat.com> (raw)
In-Reply-To: <f294faba4e5d25aba8773f36170d1309236edd3b.camel@redhat.com>
Maxim Levitsky <mlevitsk@redhat.com> writes:
> On Wed, 2021-06-09 at 17:09 +0200, Vitaly Kuznetsov wrote:
>> Changes since v2:
>> - First two patches got merged, rebase.
>> - Use 'enable_apicv = avic = ...' in PATCH1 [Paolo]
>> - Collect R-b tags for PATCH2 [Sean, Max]
>> - Use hv_apicv_update_work() to get out of SRCU lock [Max]
>> - "KVM: x86: Check for pending interrupts when APICv is getting disabled"
>> added.
>>
>> Original description:
>>
>> APICV_INHIBIT_REASON_HYPERV is currently unconditionally forced upon
>> SynIC activation as SynIC's AutoEOI is incompatible with APICv/AVIC. It is,
>> however, possible to track whether the feature was actually used by the
>> guest and only inhibit APICv/AVIC when needed.
>>
>> The series can be tested with the followin hack:
>>
>> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
>> index 9a48f138832d..65a9974f80d9 100644
>> --- a/arch/x86/kvm/cpuid.c
>> +++ b/arch/x86/kvm/cpuid.c
>> @@ -147,6 +147,13 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu)
>> vcpu->arch.ia32_misc_enable_msr &
>> MSR_IA32_MISC_ENABLE_MWAIT);
>> }
>> +
>> + /* Dirty hack: force HV_DEPRECATING_AEOI_RECOMMENDED. Not to be merged! */
>> + best = kvm_find_cpuid_entry(vcpu, HYPERV_CPUID_ENLIGHTMENT_INFO, 0);
>> + if (best) {
>> + best->eax &= ~HV_X64_APIC_ACCESS_RECOMMENDED;
>> + best->eax |= HV_DEPRECATING_AEOI_RECOMMENDED;
>> + }
>> }
>> EXPORT_SYMBOL_GPL(kvm_update_cpuid_runtime);
>>
>> Vitaly Kuznetsov (4):
>> KVM: x86: Use common 'enable_apicv' variable for both APICv and AVIC
>> KVM: x86: Drop vendor specific functions for APICv/AVIC enablement
>> KVM: x86: Check for pending interrupts when APICv is getting disabled
>> KVM: x86: hyper-v: Deactivate APICv only when AutoEOI feature is in
>> use
>>
>> arch/x86/include/asm/kvm_host.h | 9 +++++-
>> arch/x86/kvm/hyperv.c | 51 +++++++++++++++++++++++++++++----
>> arch/x86/kvm/svm/avic.c | 14 ++++-----
>> arch/x86/kvm/svm/svm.c | 22 ++++++++------
>> arch/x86/kvm/svm/svm.h | 2 --
>> arch/x86/kvm/vmx/capabilities.h | 1 -
>> arch/x86/kvm/vmx/vmx.c | 2 --
>> arch/x86/kvm/x86.c | 18 ++++++++++--
>> 8 files changed, 86 insertions(+), 33 deletions(-)
>>
>
> Hi!
>
> I hate to say it, but at least one of my VMs doesn't boot amymore
> with avic=1, after the recent updates. I'll bisect this soon,
> but this is likely related to this series.
>
> I will also review this series very soon.
>
> When the VM fails, it hangs on the OVMF screen and I see this
> in qemu logs:
>
> KVM: injection failed, MSI lost (Operation not permitted)
> KVM: injection failed, MSI lost (Operation not permitted)
> KVM: injection failed, MSI lost (Operation not permitted)
> KVM: injection failed, MSI lost (Operation not permitted)
> KVM: injection failed, MSI lost (Operation not permitted)
> KVM: injection failed, MSI lost (Operation not permitted)
>
-EPERM?? Interesting... strace(1) may come handy...
$ git grep EPERM kvm/queue arch/x86/kvm/
kvm/queue:arch/x86/kvm/x86.c: ret = -KVM_EPERM;
(just this one)
kvm_emulate_hypercall():
...
b3646477d458f arch/x86/kvm/x86.c (Jason Baron 2021-01-14 22:27:56 -0500 8433) if (static_call(kvm_x86_get_cpl)(vcpu) != 0) {
07708c4af1346 arch/x86/kvm/x86.c (Jan Kiszka 2009-08-03 18:43:28 +0200 8434) ret = -KVM_EPERM;
696ca779a928d arch/x86/kvm/x86.c (Radim Krčmář 2018-05-24 17:50:56 +0200 8435) goto out;
07708c4af1346 arch/x86/kvm/x86.c (Jan Kiszka 2009-08-03 18:43:28 +0200 8436) }
...
Doesn't seem we have any updates here, curious what your bisection will
point us to.
--
Vitaly
next prev parent reply other threads:[~2021-06-14 7:40 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-09 15:09 [PATCH v3 0/4] KVM: x86: hyper-v: Conditionally allow SynIC with APICv/AVIC Vitaly Kuznetsov
2021-06-09 15:09 ` [PATCH v3 1/4] KVM: x86: Use common 'enable_apicv' variable for both APICv and AVIC Vitaly Kuznetsov
2021-06-09 15:09 ` [PATCH v3 2/4] KVM: x86: Drop vendor specific functions for APICv/AVIC enablement Vitaly Kuznetsov
2021-06-09 15:09 ` [PATCH v3 3/4] KVM: x86: Check for pending interrupts when APICv is getting disabled Vitaly Kuznetsov
2021-06-09 15:09 ` [PATCH v3 4/4] KVM: x86: hyper-v: Deactivate APICv only when AutoEOI feature is in use Vitaly Kuznetsov
2021-06-10 13:17 ` [PATCH v3 0/4] KVM: x86: hyper-v: Conditionally allow SynIC with APICv/AVIC Paolo Bonzini
2021-06-11 18:50 ` Maxim Levitsky
2021-06-14 7:40 ` Vitaly Kuznetsov [this message]
2021-06-14 9:51 ` Maxim Levitsky
2021-06-14 13:08 ` Paolo Bonzini
2021-06-14 18:21 ` Maxim Levitsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87zgvsx5b1.fsf@vitty.brq.redhat.com \
--to=vkuznets@redhat.com \
--cc=jmattson@google.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mlevitsk@redhat.com \
--cc=pbonzini@redhat.com \
--cc=seanjc@google.com \
--cc=wanpengli@tencent.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).