From: Vitaly Kuznetsov <vkuznets@redhat.com>
To: Sean Christopherson <seanjc@google.com>
Cc: kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>,
Wanpeng Li <wanpengli@tencent.com>,
Jim Mattson <jmattson@google.com>,
Michael Kelley <mikelley@microsoft.com>,
Siddharth Chandrasekaran <sidcha@amazon.de>,
Yuan Yao <yuan.yao@linux.intel.com>,
Maxim Levitsky <mlevitsk@redhat.com>,
linux-hyperv@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v10 02/39] KVM: x86: hyper-v: Resurrect dedicated KVM_REQ_HV_TLB_FLUSH flag
Date: Thu, 22 Sep 2022 17:37:05 +0200 [thread overview]
Message-ID: <87sfkjmndq.fsf@redhat.com> (raw)
In-Reply-To: <Yyx+AJLacVzOdtBr@google.com>
Sean Christopherson <seanjc@google.com> writes:
> On Thu, Sep 22, 2022, Vitaly Kuznetsov wrote:
>> Now let's get to VMX and the point of my confusion (and thanks in
>> advance for educating me!):
>> AFAIU, when EPT is in use:
>> KVM_REQ_TLB_FLUSH_CURRENT == invept
>> KVM_REQ_TLB_FLUSH_GUEST = invvpid
>>
>> For "normal" mappings (which are mapped on both stages) this is the same
>> thing as they're 'tagged' with both VPID and 'EPT root'. The question is
>> what's left. Given your comment, do I understand correctly that in case
>> of an invalid mapping in the guest (GVA doesn't resolve to a GPA), this
>> will only be tagged with VPID but not with 'EPT root' (as the CPU never
>> reached to the second translation stage)? We certainly can't ignore
>> these. Another (probably pure theoretical question) is what are the
>> mappings which are tagged with 'EPT root' but don't have a VPID tag?
>
> Intel puts mappings into three categories, which for non-root mode equates to:
>
> linear == GVA => GPA
> guest-physical == GPA => HPA
> combined == GVA => HPA
>
> and essentially the categories that consume the GVA are tagged with the VPID
> (linear and combined), and categories that consume the GPA are tagged with the
> EPTP address (guest-physical and combined).
>
>> Are these the mapping which happen when e.g. vCPU has paging disabled?
>
> No, these mappings can be created at all times. Even with CR0.PG=1, the guest
> can generate GPAs without going through a GVA=>GPA translation, e.g. the page tables
> themselves, RTIT (Intel PT) addresses, etc... And even for combined/full
> translations, the CPU can insert TLB entries for just the GPA=>HPA part.
>
> E.g. when a page is allocated by/for userspace, the kernel will zero the page using
> the kernel's direct map, but userspace will access the page via a different GVA.
> I.e. the guest effectively aliases GPA(x) with GVA(k) and GVA(u). By inserting
> the GPA(x) => HPA(y) into the TLB, when guest userspace access GVA(u), the CPU
> encounters a TLB miss on GVA(u) => GPA(x), but gets a TLB hit on GPA(x) => HPA(y).
>
> Separating EPT flushes from VPID (and PCID) flushes allows the CPU to retain
> the partial TLB entries, e.g. a host change in the EPT tables will result in the
> guest-physical and combined mappings being invalidated, but linear mappings can
> be kept.
>
Thanks a bunch! For some reason I though it's always the full thing (combined)
which is tagged with both VPID/PCID and EPTP and linear/guest-physical
are just 'corner' cases (but are still combined and tagged). Apparently,
it's not like that.
> I'm 99% certain AMD also caches partial entries, e.g. see the blurb on INVLPGA
> not affecting NPT translations, AMD just doesn't provide a way for the host to
> flush _only_ NPT translations. Maybe the performance benefits weren't significant
> enough to justify the extra complexity?
>
>> These are probably unrelated to Hyper-V TLB flushing.
>>
>> To preserve the 'small' optimization, we can probably move
>> kvm_clear_request(KVM_REQ_HV_TLB_FLUSH, vcpu);
>>
>> to nested_svm_transition_tlb_flush() or, in case this sounds too
>> hackish
>
> Move it to svm_flush_tlb_current(), because the justification is that on SVM,
> flushing "current" TLB entries also flushes "guest" TLB entries due to the more
> coarse-grained ASID-based TLB flush. E.g.
>
> diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
> index dd599afc85f5..a86b41503723 100644
> --- a/arch/x86/kvm/svm/svm.c
> +++ b/arch/x86/kvm/svm/svm.c
> @@ -3737,6 +3737,13 @@ static void svm_flush_tlb_current(struct kvm_vcpu *vcpu)
> {
> struct vcpu_svm *svm = to_svm(vcpu);
>
> + /*
> + * Unlike VMX, SVM doesn't provide a way to flush only NPT TLB entries.
> + * A TLB flush for the current ASID flushes both "host" and "guest" TLB
> + * entries, and thus is a superset of Hyper-V's fine grained flushing.
> + */
> + kvm_hv_vcpu_purge_flush_tlb(vcpu);
> +
> /*
> * Flush only the current ASID even if the TLB flush was invoked via
> * kvm_flush_remote_tlbs(). Although flushing remote TLBs requires all
>
>> we can drop it for now and add it to the (already overfull)
>> bucket of the "optimize nested_svm_transition_tlb_flush()".
>
> I think even long term, purging Hyper-V's FIFO in svm_flush_tlb_current() is the
> correct/desired behavior. This doesn't really have anything to do with nSVM,
> it's all about SVM not providing a way to flush only NPT entries.
True that, silly me forgot that even without any nesting, Hyper-V TLB
flush after svm_flush_tlb_current() makes no sense.
>
--
Vitaly
next prev parent reply other threads:[~2022-09-22 15:37 UTC|newest]
Thread overview: 64+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-21 15:23 [PATCH v10 00/39] KVM: x86: hyper-v: Fine-grained TLB flush + L2 TLB flush features Vitaly Kuznetsov
2022-09-21 15:23 ` [PATCH v10 01/39] KVM: x86: Rename 'enable_direct_tlbflush' to 'enable_l2_tlb_flush' Vitaly Kuznetsov
2022-09-21 15:23 ` [PATCH v10 02/39] KVM: x86: hyper-v: Resurrect dedicated KVM_REQ_HV_TLB_FLUSH flag Vitaly Kuznetsov
2022-09-21 16:23 ` Sean Christopherson
2022-09-21 16:45 ` Sean Christopherson
2022-09-22 9:35 ` Vitaly Kuznetsov
2022-09-22 9:31 ` Vitaly Kuznetsov
2022-09-22 15:23 ` Sean Christopherson
2022-09-22 15:37 ` Vitaly Kuznetsov [this message]
2022-09-21 15:24 ` [PATCH v10 03/39] KVM: x86: hyper-v: Introduce TLB flush fifo Vitaly Kuznetsov
2022-09-21 16:56 ` Sean Christopherson
2022-09-22 9:42 ` Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 04/39] KVM: x86: hyper-v: Add helper to read hypercall data for array Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 05/39] KVM: x86: hyper-v: Handle HVCALL_FLUSH_VIRTUAL_ADDRESS_LIST{,EX} calls gently Vitaly Kuznetsov
2022-09-21 17:00 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 06/39] KVM: x86: hyper-v: Expose support for extended gva ranges for flush hypercalls Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 07/39] KVM: x86: Prepare kvm_hv_flush_tlb() to handle L2's GPAs Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 08/39] x86/hyperv: Introduce HV_MAX_SPARSE_VCPU_BANKS/HV_VCPUS_PER_SPARSE_BANK constants Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 09/39] KVM: x86: hyper-v: Use HV_MAX_SPARSE_VCPU_BANKS/HV_VCPUS_PER_SPARSE_BANK instead of raw '64' Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 10/39] KVM: x86: hyper-v: Don't use sparse_set_to_vcpu_mask() in kvm_hv_send_ipi() Vitaly Kuznetsov
2022-09-21 20:54 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 11/39] KVM: x86: hyper-v: Create a separate fifo for L2 TLB flush Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 12/39] KVM: x86: hyper-v: Use preallocated buffer in 'struct kvm_vcpu_hv' instead of on-stack 'sparse_banks' Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 13/39] KVM: nVMX: Keep track of hv_vm_id/hv_vp_id when eVMCS is in use Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 14/39] KVM: nSVM: Keep track of Hyper-V hv_vm_id/hv_vp_id Vitaly Kuznetsov
2022-09-21 21:16 ` Sean Christopherson
2022-09-22 9:51 ` Vitaly Kuznetsov
2022-09-22 19:52 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 15/39] KVM: x86: Introduce .hv_inject_synthetic_vmexit_post_tlb_flush() nested hook Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 16/39] KVM: x86: hyper-v: Introduce kvm_hv_is_tlb_flush_hcall() Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 17/39] KVM: x86: hyper-v: L2 TLB flush Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 18/39] KVM: x86: hyper-v: Introduce fast guest_hv_cpuid_has_l2_tlb_flush() check Vitaly Kuznetsov
2022-09-21 21:19 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 19/39] KVM: nVMX: hyper-v: Cache VP assist page in 'struct kvm_vcpu_hv' Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 20/39] KVM: nVMX: hyper-v: Enable L2 TLB flush Vitaly Kuznetsov
2022-09-21 21:24 ` Sean Christopherson
2022-09-22 16:05 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 21/39] KVM: nSVM: " Vitaly Kuznetsov
2022-09-21 21:31 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 22/39] KVM: x86: Expose Hyper-V L2 TLB flush feature Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 23/39] KVM: selftests: Better XMM read/write helpers Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 24/39] KVM: selftests: Move HYPERV_LINUX_OS_ID definition to a common header Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 25/39] KVM: selftests: Move the function doing Hyper-V hypercall " Vitaly Kuznetsov
2022-09-21 21:51 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 26/39] KVM: selftests: Hyper-V PV IPI selftest Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 27/39] KVM: selftests: Fill in vm->vpages_mapped bitmap in virt_map() too Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 28/39] KVM: selftests: Export vm_vaddr_unused_gap() to make it possible to request unmapped ranges Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 29/39] KVM: selftests: Export _vm_get_page_table_entry() Vitaly Kuznetsov
2022-09-21 22:13 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 30/39] KVM: selftests: Hyper-V PV TLB flush selftest Vitaly Kuznetsov
2022-09-21 22:52 ` Sean Christopherson
2022-10-03 13:01 ` Vitaly Kuznetsov
2022-10-03 15:47 ` Sean Christopherson
2022-10-03 16:00 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 31/39] KVM: selftests: Sync 'struct hv_enlightened_vmcs' definition with hyperv-tlfs.h Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 32/39] KVM: selftests: Sync 'struct hv_vp_assist_page' " Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 33/39] KVM: selftests: Move Hyper-V VP assist page enablement out of evmcs.h Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 34/39] KVM: selftests: Split off load_evmcs() from load_vmcs() Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 35/39] KVM: selftests: Create a vendor independent helper to allocate Hyper-V specific test pages Vitaly Kuznetsov
2022-09-21 22:59 ` Sean Christopherson
2022-09-21 15:24 ` [PATCH v10 36/39] KVM: selftests: Allocate Hyper-V partition assist page Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 37/39] KVM: selftests: evmcs_test: Introduce L2 TLB flush test Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 38/39] KVM: selftests: hyperv_svm_test: " Vitaly Kuznetsov
2022-09-21 15:24 ` [PATCH v10 39/39] KVM: selftests: Rename 'evmcs_test' to 'hyperv_evmcs' Vitaly Kuznetsov
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=87sfkjmndq.fsf@redhat.com \
--to=vkuznets@redhat.com \
--cc=jmattson@google.com \
--cc=kvm@vger.kernel.org \
--cc=linux-hyperv@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mikelley@microsoft.com \
--cc=mlevitsk@redhat.com \
--cc=pbonzini@redhat.com \
--cc=seanjc@google.com \
--cc=sidcha@amazon.de \
--cc=wanpengli@tencent.com \
--cc=yuan.yao@linux.intel.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).