From: Vitaly Kuznetsov <vkuznets@redhat.com>
To: kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>
Cc: Sean Christopherson <sean.j.christopherson@intel.com>,
Wanpeng Li <wanpengli@tencent.com>,
Jim Mattson <jmattson@google.com>,
Junaid Shahid <junaids@google.com>,
linux-kernel@vger.kernel.org
Subject: [PATCH v3 7/9] KVM: nSVM: implement nested_svm_load_cr3() and use it for host->guest switch
Date: Thu, 9 Jul 2020 16:53:56 +0200 [thread overview]
Message-ID: <20200709145358.1560330-8-vkuznets@redhat.com> (raw)
In-Reply-To: <20200709145358.1560330-1-vkuznets@redhat.com>
Undesired triple fault gets injected to L1 guest on SVM when L2 is
launched with certain CR3 values. #TF is raised by mmu_check_root()
check in fast_pgd_switch() and the root cause is that when
kvm_set_cr3() is called from nested_prepare_vmcb_save() with NPT
enabled CR3 points to a nGPA so we can't check it with
kvm_is_visible_gfn().
Using generic kvm_set_cr3() when switching to nested guest is not
a great idea as we'll have to distinguish between 'real' CR3s and
'nested' CR3s to e.g. not call kvm_mmu_new_pgd() with nGPA. Following
nVMX implement nested-specific nested_svm_load_cr3() doing the job.
Note: the current implementation is sub-optimal as we always do TLB
flush/MMU sync but this is still an improvement as we at least stop doing
kvm_mmu_reset_context().
Fixes: 7c390d350f8b ("kvm: x86: Add fast CR3 switch code path")
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
---
arch/x86/kvm/mmu/mmu.c | 2 ++
arch/x86/kvm/svm/nested.c | 33 +++++++++++++++++++++++++--------
2 files changed, 27 insertions(+), 8 deletions(-)
diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 69fa51af8cbf..1c3a231f825b 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -4986,6 +4986,8 @@ void kvm_init_shadow_npt_mmu(struct kvm_vcpu *vcpu, u32 cr0, u32 cr4, u32 efer,
union kvm_mmu_role new_role =
kvm_calc_shadow_mmu_root_page_role(vcpu, false);
+ __kvm_mmu_new_pgd(vcpu, nested_cr3, new_role.base, false, false);
+
if (new_role.as_u64 != context->mmu_role.as_u64)
shadow_mmu_init_context(vcpu, context, cr0, cr4, efer, new_role);
}
diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c
index 5f001d2c41d1..5ddf20941cf9 100644
--- a/arch/x86/kvm/svm/nested.c
+++ b/arch/x86/kvm/svm/nested.c
@@ -323,7 +323,28 @@ static inline bool nested_npt_enabled(struct vcpu_svm *svm)
static int nested_svm_load_cr3(struct kvm_vcpu *vcpu, unsigned long cr3,
bool nested_npt)
{
- return kvm_set_cr3(vcpu, cr3);
+ if (cr3 & rsvd_bits(cpuid_maxphyaddr(vcpu), 63))
+ return -EINVAL;
+
+ if (!nested_npt && is_pae_paging(vcpu) &&
+ (cr3 != kvm_read_cr3(vcpu) || pdptrs_changed(vcpu))) {
+ if (!load_pdptrs(vcpu, vcpu->arch.walk_mmu, cr3))
+ return -EINVAL;
+ }
+
+ /*
+ * TODO: optimize unconditional TLB flush/MMU sync here and in
+ * kvm_init_shadow_npt_mmu().
+ */
+ if (!nested_npt)
+ kvm_mmu_new_pgd(vcpu, cr3, false, false);
+
+ vcpu->arch.cr3 = cr3;
+ kvm_register_mark_available(vcpu, VCPU_EXREG_CR3);
+
+ kvm_init_mmu(vcpu, false);
+
+ return 0;
}
static int nested_prepare_vmcb_save(struct vcpu_svm *svm, struct vmcb *nested_vmcb)
@@ -342,6 +363,9 @@ static int nested_prepare_vmcb_save(struct vcpu_svm *svm, struct vmcb *nested_vm
svm_set_cr0(&svm->vcpu, nested_vmcb->save.cr0);
svm_set_cr4(&svm->vcpu, nested_vmcb->save.cr4);
+ if (nested_npt_enabled(svm))
+ nested_svm_init_mmu_context(&svm->vcpu);
+
ret = nested_svm_load_cr3(&svm->vcpu, nested_vmcb->save.cr3,
nested_npt_enabled(svm));
@@ -364,13 +388,6 @@ static int nested_prepare_vmcb_save(struct vcpu_svm *svm, struct vmcb *nested_vm
static void nested_prepare_vmcb_control(struct vcpu_svm *svm)
{
const u32 mask = V_INTR_MASKING_MASK | V_GIF_ENABLE_MASK | V_GIF_MASK;
- if (nested_npt_enabled(svm))
- nested_svm_init_mmu_context(&svm->vcpu);
-
- /* Guest paging mode is active - reset mmu */
- kvm_mmu_reset_context(&svm->vcpu);
-
- svm_flush_tlb(&svm->vcpu);
svm->vmcb->control.tsc_offset = svm->vcpu.arch.tsc_offset =
svm->vcpu.arch.l1_tsc_offset + svm->nested.ctl.tsc_offset;
--
2.25.4
next prev parent reply other threads:[~2020-07-09 14:54 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-09 14:53 [PATCH v3 0/9] KVM: nSVM: fixes for CR3/MMU switch upon nested guest entry/exit Vitaly Kuznetsov
2020-07-09 14:53 ` [PATCH v3 1/9] KVM: nSVM: split kvm_init_shadow_npt_mmu() from kvm_init_shadow_mmu() Vitaly Kuznetsov
2020-07-09 14:53 ` [PATCH v3 2/9] KVM: nSVM: stop dereferencing vcpu->arch.mmu to get the context in kvm_init_shadow{,_npt}_mmu() Vitaly Kuznetsov
2020-07-09 17:47 ` Paolo Bonzini
2020-07-09 14:53 ` [PATCH v3 3/9] KVM: nSVM: reset nested_run_pending upon nested_svm_vmrun_msrpm() failure Vitaly Kuznetsov
2020-07-09 17:48 ` Paolo Bonzini
2020-07-09 14:53 ` [PATCH v3 4/9] KVM: nSVM: prepare to handle errors from enter_svm_guest_mode() Vitaly Kuznetsov
2020-07-09 14:53 ` [PATCH v3 5/9] KVM: nSVM: introduce nested_svm_load_cr3() Vitaly Kuznetsov
2020-07-09 14:53 ` [PATCH v3 6/9] KVM: nSVM: move kvm_set_cr3() after nested_svm_uninit_mmu_context() Vitaly Kuznetsov
2020-07-09 17:51 ` Paolo Bonzini
2020-07-09 14:53 ` Vitaly Kuznetsov [this message]
2020-07-09 17:57 ` [PATCH v3 7/9] KVM: nSVM: implement nested_svm_load_cr3() and use it for host->guest switch Paolo Bonzini
2020-07-09 17:59 ` Paolo Bonzini
2020-07-10 11:40 ` Vitaly Kuznetsov
2020-07-10 12:08 ` Paolo Bonzini
2020-07-10 12:30 ` Vitaly Kuznetsov
2020-07-09 14:53 ` [PATCH v3 8/9] KVM: nSVM: use nested_svm_load_cr3() on guest->host switch Vitaly Kuznetsov
2020-07-09 14:53 ` [PATCH v3 9/9] KVM: x86: drop superfluous mmu_check_root() from fast_pgd_switch() Vitaly Kuznetsov
2020-07-09 18:01 ` [PATCH v3 0/9] KVM: nSVM: fixes for CR3/MMU switch upon nested guest entry/exit Paolo Bonzini
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=20200709145358.1560330-8-vkuznets@redhat.com \
--to=vkuznets@redhat.com \
--cc=jmattson@google.com \
--cc=junaids@google.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=sean.j.christopherson@intel.com \
--cc=wanpengli@tencent.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).