[Artful] (upstream) cherry-pick mmu_notifier fix to prevent Windows BSODs

Message ID 20171219085213.3qlssbsnz4l276wb@nora.maurer-it.com
State New
Headers show
Series
  • [Artful] (upstream) cherry-pick mmu_notifier fix to prevent Windows BSODs
Related show

Commit Message

Fabian Grünbichler Dec. 19, 2017, 8:52 a.m.
BugLink: https://bugs.launchpad.net/bugs/1738972

Impact: Windows 2016 and 10 VMs running in Qemu VMs using KVM trigger
BSODs under certain memory conditions

Fix: Single follow-up upstream cherry pick which fixes the problem.

Regression Potential: Merged in 4.15 and 4.14.6, tested by multiple
affected users.

see https://marc.info/?l=linux-kernel&m=151203441531472&w=2 for the long
writeup

clean cherry-pick from 4.14.6

Comments

Stefan Bader Jan. 23, 2018, 1:45 p.m. | #1
On 19.12.2017 09:52, Fabian Grünbichler wrote:
> BugLink: https://bugs.launchpad.net/bugs/1738972
> 
> Impact: Windows 2016 and 10 VMs running in Qemu VMs using KVM trigger
> BSODs under certain memory conditions
> 
> Fix: Single follow-up upstream cherry pick which fixes the problem.
> 
> Regression Potential: Merged in 4.15 and 4.14.6, tested by multiple
> affected users.
> 
> see https://marc.info/?l=linux-kernel&m=151203441531472&w=2 for the long
> writeup
> 
> clean cherry-pick from 4.14.6
> 
> 
> 

Clean cherry-pick from upstream stable, tested.

Acked-by: Stefan Bader <stefan.bader@canonical.com>
Colin King Jan. 23, 2018, 5:48 p.m. | #2
On 19/12/17 08:52, Fabian Grünbichler wrote:
> BugLink: https://bugs.launchpad.net/bugs/1738972
> 
> Impact: Windows 2016 and 10 VMs running in Qemu VMs using KVM trigger
> BSODs under certain memory conditions
> 
> Fix: Single follow-up upstream cherry pick which fixes the problem.
> 
> Regression Potential: Merged in 4.15 and 4.14.6, tested by multiple
> affected users.
> 
> see https://marc.info/?l=linux-kernel&m=151203441531472&w=2 for the long
> writeup
> 
> clean cherry-pick from 4.14.6
> 

Clean cherry pick with testing, so looks OK.

Acked-by: Colin Ian King <colin.king@canonical.com>
Stefan Bader Feb. 20, 2018, 9:16 a.m. | #3
On 19.12.2017 09:52, Fabian Grünbichler wrote:
> BugLink: https://bugs.launchpad.net/bugs/1738972
> 
> Impact: Windows 2016 and 10 VMs running in Qemu VMs using KVM trigger
> BSODs under certain memory conditions
> 
> Fix: Single follow-up upstream cherry pick which fixes the problem.
> 
> Regression Potential: Merged in 4.15 and 4.14.6, tested by multiple
> affected users.
> 
> see https://marc.info/?l=linux-kernel&m=151203441531472&w=2 for the long
> writeup
> 
> clean cherry-pick from 4.14.6
> 
> 
> 
Applied to artful/master-next-backlog

Patch

From 95faa25b783f41f86f7393efa1cf149b614503c2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Radim=20Kr=C4=8Dm=C3=A1=C5=99?= <rkrcmar@redhat.com>
Date: Thu, 30 Nov 2017 19:05:45 +0100
Subject: [PATCH][Artful] KVM: x86: fix APIC page invalidation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

BugLink: https://bugs.launchpad.net/bugs/1738972

commit b1394e745b9453dcb5b0671c205b770e87dedb87 upstream.

Implementation of the unpinned APIC page didn't update the VMCS address
cache when invalidation was done through range mmu notifiers.
This became a problem when the page notifier was removed.

Re-introduce the arch-specific helper and call it from ...range_start.

Reported-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Fixes: 38b9917350cb ("kvm: vmx: Implement set_apic_access_page_addr")
Fixes: 369ea8242c0f ("mm/rmap: update to new mmu_notifier semantic v2")
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Andrea Arcangeli <aarcange@redhat.com>
Tested-by: Wanpeng Li <wanpeng.li@hotmail.com>
Tested-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 58582f04bc87b9d8d848d9163ce3355dd6f00602)
Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
---
 arch/x86/include/asm/kvm_host.h |  3 +++
 arch/x86/kvm/x86.c              | 14 ++++++++++++++
 virt/kvm/kvm_main.c             |  8 ++++++++
 3 files changed, 25 insertions(+)

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index 92c9032502d8..b69af3df978a 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1437,4 +1437,7 @@  static inline int kvm_cpu_get_apicid(int mps_cpu)
 #endif
 }
 
+void kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
+		unsigned long start, unsigned long end);
+
 #endif /* _ASM_X86_KVM_HOST_H */
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 7351cdc46cc7..703cd4171921 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -6711,6 +6711,20 @@  static void kvm_vcpu_flush_tlb(struct kvm_vcpu *vcpu)
 	kvm_x86_ops->tlb_flush(vcpu);
 }
 
+void kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
+		unsigned long start, unsigned long end)
+{
+	unsigned long apic_address;
+
+	/*
+	 * The physical address of apic access page is stored in the VMCS.
+	 * Update it when it becomes invalid.
+	 */
+	apic_address = gfn_to_hva(kvm, APIC_DEFAULT_PHYS_BASE >> PAGE_SHIFT);
+	if (start <= apic_address && apic_address < end)
+		kvm_make_all_cpus_request(kvm, KVM_REQ_APIC_PAGE_RELOAD);
+}
+
 void kvm_vcpu_reload_apic_access_page(struct kvm_vcpu *vcpu)
 {
 	struct page *page = NULL;
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 4d81f6ded88e..3b3e54742263 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -136,6 +136,11 @@  static void kvm_uevent_notify_change(unsigned int type, struct kvm *kvm);
 static unsigned long long kvm_createvm_count;
 static unsigned long long kvm_active_vms;
 
+__weak void kvm_arch_mmu_notifier_invalidate_range(struct kvm *kvm,
+		unsigned long start, unsigned long end)
+{
+}
+
 bool kvm_is_reserved_pfn(kvm_pfn_t pfn)
 {
 	if (pfn_valid(pfn))
@@ -361,6 +366,9 @@  static void kvm_mmu_notifier_invalidate_range_start(struct mmu_notifier *mn,
 		kvm_flush_remote_tlbs(kvm);
 
 	spin_unlock(&kvm->mmu_lock);
+
+	kvm_arch_mmu_notifier_invalidate_range(kvm, start, end);
+
 	srcu_read_unlock(&kvm->srcu, idx);
 }
 
-- 
2.14.2