Re: [PART2 PATCH v4 10/11] svm: Introduce AMD IOMMU avic_ga_log_notifier

From: Radim KrÄmÃÅ
Date: Wed Jul 13 2016 - 10:30:38 EST


2016-07-13 08:20-0500, Suravee Suthikulpanit:
> From: Suravee Suthikulpanit <Suravee.Suthikulpanit@xxxxxxx>
>
> This patch introduces avic_ga_log_notifier, which will be called
> by IOMMU driver whenever it handles the Guest vAPIC (GA) log entry.
>
> Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@xxxxxxx>
> ---
> arch/x86/include/asm/kvm_host.h | 2 ++
> arch/x86/kvm/svm.c | 68 +++++++++++++++++++++++++++++++++++++++--
> 2 files changed, 68 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index 69e62862..a9466ad 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -776,9 +776,11 @@ struct kvm_arch {
> bool disabled_lapic_found;
>
> /* Struct members for AVIC */
> + u32 avic_vm_id;
> u32 ldr_mode;
> struct page *avic_logical_id_table_page;
> struct page *avic_physical_id_table_page;
> + struct hlist_node hnode;
> };
>
> struct kvm_vm_stat {
> diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
> index 16ef31b..1d9f2f6 100644
> --- a/arch/x86/kvm/svm.c
> +++ b/arch/x86/kvm/svm.c
> @@ -34,6 +34,8 @@
> #include <linux/sched.h>
> #include <linux/trace_events.h>
> #include <linux/slab.h>
> +#include <linux/amd-iommu.h>
> +#include <linux/hashtable.h>
>
> #include <asm/apic.h>
> #include <asm/perf_event.h>
> @@ -928,6 +930,53 @@ static void svm_disable_lbrv(struct vcpu_svm *svm)
> set_msr_interception(msrpm, MSR_IA32_LASTINTTOIP, 0, 0);
> }
>
> +/* Note:
> + * This hash table is used to map VM_ID to a struct kvm_arch,
> + * when handling AMD IOMMU GALOG notification to schedule in
> + * a particular vCPU.
> + */
> +#define SVM_VM_DATA_HASH_BITS 8
> +DECLARE_HASHTABLE(svm_vm_data_hash, SVM_VM_DATA_HASH_BITS);
> +static spinlock_t svm_vm_data_hash_lock;
> +
> +/* Note:
> + * This function is called from IOMMU driver to notify
> + * SVM to schedule in a particular vCPU of a particular VM.
> + */
> +static int avic_ga_log_notifier(int vm_id, int vcpu_id)
> +{
> + unsigned long flags;
> + struct kvm_arch *ka = NULL;
> + struct kvm_vcpu *vcpu = NULL;
> + struct vcpu_svm *svm = NULL;
> +
> + pr_debug("SVM: %s: vm_id=%#x, vcpu_id=%#x\n", __func__, vm_id, vcpu_id);
> +
> + spin_lock_irqsave(&svm_vm_data_hash_lock, flags);
> + hash_for_each_possible(svm_vm_data_hash, ka, hnode, vm_id) {
> + struct kvm *kvm = container_of(ka, struct kvm, arch);
> +
> + vcpu = kvm_get_vcpu_by_id(kvm, vcpu_id);

The first result is not neccessarily the correct one.

With more than active 256 VMs, there is a guaranteed collision that
cannot be disambiguated, so VCPUs in both VMs need to be woken up.

Having a 24 bit vm_id and checking that
kvm->*.avic_id & 0xfffff == vm_id
would help a bit to avoid useless wakeups, but the collision cannot be
avoided.

> + break;
> + }
> + spin_unlock_irqrestore(&svm_vm_data_hash_lock, flags);
> +
> + if (!vcpu)
> + return 0;
> +
> + svm = to_svm(vcpu);
> +
> + /* Note:
> + * At this point, the IOMMU should have already set the pending
> + * bit in the vAPIC backing page. So, we just need to schedule
> + * in the vcpu.
> + */
> + if (vcpu->mode == OUTSIDE_GUEST_MODE)
> + kvm_vcpu_wake_up(vcpu);
> +
> + return 0;
> +}