On Wed, 2023-10-25 at 17:23 +0200, Vitaly Kuznetsov wrote:
> Hyper-V partition assist page is used when KVM runs on top of Hyper-V and
> is not used for Windows/Hyper-V guests on KVM, this means that 'hv_pa_pg'
> placement in 'struct kvm_hv' is unfortunate. As a preparation to making
> Hyper-V emulation optional, move 'hv_pa_pg' to 'struct kvm_arch' and put it
> under CONFIG_HYPERV.
>
> While on it, introduce hv_get_partition_assist_page() helper to allocate
> partition assist page. Move the comment explaining why we use a single page
> for all vCPUs from VMX and expand it a bit.
>
> No functional change intended.
>
> Signed-off-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
> ---
> arch/x86/include/asm/kvm_host.h | 2 +-
> arch/x86/kvm/kvm_onhyperv.h | 20 ++++++++++++++++++++
> arch/x86/kvm/svm/svm_onhyperv.c | 10 +++-------
> arch/x86/kvm/vmx/vmx.c | 14 +++-----------
> arch/x86/kvm/x86.c | 4 +++-
> 5 files changed, 30 insertions(+), 20 deletions(-)
>
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index d107516b4591..7fb2810f4573 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1125,7 +1125,6 @@ struct kvm_hv {
> */
> unsigned int synic_auto_eoi_used;
>
> - struct hv_partition_assist_pg *hv_pa_pg;
> struct kvm_hv_syndbg hv_syndbg;
> };
>
> @@ -1447,6 +1446,7 @@ struct kvm_arch {
> #if IS_ENABLED(CONFIG_HYPERV)
> hpa_t hv_root_tdp;
> spinlock_t hv_root_tdp_lock;
> + struct hv_partition_assist_pg *hv_pa_pg;
> #endif
> /*
> * VM-scope maximum vCPU ID. Used to determine the size of structures
> diff --git a/arch/x86/kvm/kvm_onhyperv.h b/arch/x86/kvm/kvm_onhyperv.h
> index f9ca3e7432b2..eefab3dc8498 100644
> --- a/arch/x86/kvm/kvm_onhyperv.h
> +++ b/arch/x86/kvm/kvm_onhyperv.h
> @@ -10,6 +10,26 @@
> int hv_flush_remote_tlbs_range(struct kvm *kvm, gfn_t gfn, gfn_t nr_pages);
> int hv_flush_remote_tlbs(struct kvm *kvm);
> void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp);
> +static inline hpa_t hv_get_partition_assist_page(struct kvm_vcpu *vcpu)
> +{
> + /*
> + * Partition assist page is something which Hyper-V running in L0
> + * requires from KVM running in L1 before direct TLB flush for L2
> + * guests can be enabled. KVM doesn't currently use the page but to
> + * comply with TLFS it still needs to be allocated. For now, this
> + * is a single page shared among all vCPUs.
Perfect!
> + */
> + struct hv_partition_assist_pg **p_hv_pa_pg =
> + &vcpu->kvm->arch.hv_pa_pg;
> +
> + if (!*p_hv_pa_pg)
> + *p_hv_pa_pg = kzalloc(PAGE_SIZE, GFP_KERNEL_ACCOUNT);
> +
> + if (!*p_hv_pa_pg)
> + return INVALID_PAGE;
> +
> + return __pa(*p_hv_pa_pg);
> +}
> #else /* !CONFIG_HYPERV */
> static inline int hv_flush_remote_tlbs(struct kvm *kvm)
> {
> diff --git a/arch/x86/kvm/svm/svm_onhyperv.c b/arch/x86/kvm/svm/svm_onhyperv.c
> index 7af8422d3382..3971b3ea5d04 100644
> --- a/arch/x86/kvm/svm/svm_onhyperv.c
> +++ b/arch/x86/kvm/svm/svm_onhyperv.c
> @@ -18,18 +18,14 @@
> int svm_hv_enable_l2_tlb_flush(struct kvm_vcpu *vcpu)
> {
> struct hv_vmcb_enlightenments *hve;
> - struct hv_partition_assist_pg **p_hv_pa_pg =
> - &to_kvm_hv(vcpu->kvm)->hv_pa_pg;
> + hpa_t partition_assist_page = hv_get_partition_assist_page(vcpu);
>
> - if (!*p_hv_pa_pg)
> - *p_hv_pa_pg = kzalloc(PAGE_SIZE, GFP_KERNEL);
> -
> - if (!*p_hv_pa_pg)
> + if (partition_assist_page == INVALID_PAGE)
> return -ENOMEM;
>
> hve = &to_svm(vcpu)->vmcb->control.hv_enlightenments;
>
> - hve->partition_assist_page = __pa(*p_hv_pa_pg);
> + hve->partition_assist_page = partition_assist_page;
> hve->hv_vm_id = (unsigned long)vcpu->kvm;
> if (!hve->hv_enlightenments_control.nested_flush_hypercall) {
> hve->hv_enlightenments_control.nested_flush_hypercall = 1;
> diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
> index be20a60047b1..cb4591405f14 100644
> --- a/arch/x86/kvm/vmx/vmx.c
> +++ b/arch/x86/kvm/vmx/vmx.c
> @@ -523,22 +523,14 @@ module_param(enlightened_vmcs, bool, 0444);
> static int hv_enable_l2_tlb_flush(struct kvm_vcpu *vcpu)
> {
> struct hv_enlightened_vmcs *evmcs;
> - struct hv_partition_assist_pg **p_hv_pa_pg =
> - &to_kvm_hv(vcpu->kvm)->hv_pa_pg;
> - /*
> - * Synthetic VM-Exit is not enabled in current code and so All
> - * evmcs in singe VM shares same assist page.
> - */
> - if (!*p_hv_pa_pg)
> - *p_hv_pa_pg = kzalloc(PAGE_SIZE, GFP_KERNEL_ACCOUNT);
> + hpa_t partition_assist_page = hv_get_partition_assist_page(vcpu);
>
> - if (!*p_hv_pa_pg)
> + if (partition_assist_page == INVALID_PAGE)
> return -ENOMEM;
>
> evmcs = (struct hv_enlightened_vmcs *)to_vmx(vcpu)->loaded_vmcs->vmcs;
>
> - evmcs->partition_assist_page =
> - __pa(*p_hv_pa_pg);
> + evmcs->partition_assist_page = partition_assist_page;
> evmcs->hv_vm_id = (unsigned long)vcpu->kvm;
> evmcs->hv_enlightenments_control.nested_flush_hypercall = 1;
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index d632931fa545..cc2524598368 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -12425,7 +12425,9 @@ void kvm_arch_sched_in(struct kvm_vcpu *vcpu, int cpu)
>
> void kvm_arch_free_vm(struct kvm *kvm)
> {
> - kfree(to_kvm_hv(kvm)->hv_pa_pg);
> +#if IS_ENABLED(CONFIG_HYPERV)
> + kfree(kvm->arch.hv_pa_pg);
> +#endif
> __kvm_arch_free_vm(kvm);
> }
>
Reviewed-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
Best regards,
Maxim Levitsky
