Re: [PATCH v2 10/15] kvm: x86: hyperv: Use APICv deactivate request interface

[Alexander Graf]

On 15.08.19 18:25, Suthikulpanit, Suravee wrote:
Since disabling APICv has to be done for all vcpus on AMD-based system,
adopt the newly introduced kvm_make_apicv_deactivate_request() intereface.

typo


Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpa...@amd.com>
---
  arch/x86/kvm/hyperv.c | 12 ++++++++++--
  1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/arch/x86/kvm/hyperv.c b/arch/x86/kvm/hyperv.c
index a39e38f..4f71a39 100644
--- a/arch/x86/kvm/hyperv.c
+++ b/arch/x86/kvm/hyperv.c
@@ -772,9 +772,17 @@ int kvm_hv_activate_synic(struct kvm_vcpu *vcpu, bool 
dont_zero_synic_pages)
  
  	/*
         * Hyper-V SynIC auto EOI SINT's are
-        * not compatible with APICV, so deactivate APICV
+        * not compatible with APICV, so  request

double space

+        * to deactivate APICV permanently.
+        *
+        * Since this requires updating
+        * APIC_ACCESS_PAGE_PRIVATE_MEMSLOT,
+        * also take srcu lock.
         */
-       kvm_vcpu_deactivate_apicv(vcpu);
+       vcpu->srcu_idx = srcu_read_lock(&vcpu->kvm->srcu);
+       kvm_make_apicv_deactivate_request(vcpu, true);
+       srcu_read_unlock(&vcpu->kvm->srcu, vcpu->srcu_idx);

Overall, I'm not terribly happy with the srcu locks. Can't we handle the 
memslot changes outside of the lock region, inside the respective 
request handlers somehow?


Alex

+
        synic->active = true;
        synic->dont_zero_synic_pages = dont_zero_synic_pages;
        return 0;