All of lore.kernel.org
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: linux-kernel@vger.kernel.org, kvm@vger.kernel.org
Cc: vkuznets@redhat.com, Joerg Roedel <jroedel@suse.de>
Subject: [PATCH 24/24] KVM: nSVM: implement KVM_GET_NESTED_STATE and KVM_SET_NESTED_STATE
Date: Wed, 20 May 2020 13:21:45 -0400	[thread overview]
Message-ID: <20200520172145.23284-25-pbonzini@redhat.com> (raw)
In-Reply-To: <20200520172145.23284-1-pbonzini@redhat.com>

Similar to VMX, the state that is captured through the currently available
IOCTLs is a mix of L1 and L2 state, dependent on whether the L2 guest was
running at the moment when the process was interrupted to save its state.

In particular, the SVM-specific state for nested virtualization includes
the L1 saved state (including the interrupt flag), the cached L2 controls,
and the GIF.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 arch/x86/include/uapi/asm/kvm.h |  17 +++-
 arch/x86/kvm/cpuid.h            |   5 ++
 arch/x86/kvm/svm/nested.c       | 147 ++++++++++++++++++++++++++++++++
 arch/x86/kvm/svm/svm.c          |   1 +
 arch/x86/kvm/vmx/nested.c       |   5 --
 arch/x86/kvm/x86.c              |   3 +-
 6 files changed, 171 insertions(+), 7 deletions(-)

diff --git a/arch/x86/include/uapi/asm/kvm.h b/arch/x86/include/uapi/asm/kvm.h
index 3f3f780c8c65..12075a9de1c1 100644
--- a/arch/x86/include/uapi/asm/kvm.h
+++ b/arch/x86/include/uapi/asm/kvm.h
@@ -385,18 +385,22 @@ struct kvm_sync_regs {
 #define KVM_X86_QUIRK_MISC_ENABLE_NO_MWAIT (1 << 4)
 
 #define KVM_STATE_NESTED_FORMAT_VMX	0
-#define KVM_STATE_NESTED_FORMAT_SVM	1	/* unused */
+#define KVM_STATE_NESTED_FORMAT_SVM	1
 
 #define KVM_STATE_NESTED_GUEST_MODE	0x00000001
 #define KVM_STATE_NESTED_RUN_PENDING	0x00000002
 #define KVM_STATE_NESTED_EVMCS		0x00000004
 #define KVM_STATE_NESTED_MTF_PENDING	0x00000008
+#define KVM_STATE_NESTED_GIF_SET	0x00000100
 
 #define KVM_STATE_NESTED_SMM_GUEST_MODE	0x00000001
 #define KVM_STATE_NESTED_SMM_VMXON	0x00000002
 
 #define KVM_STATE_NESTED_VMX_VMCS_SIZE	0x1000
 
+#define KVM_STATE_NESTED_SVM_VMCB_SIZE	0x1000
+
+
 struct kvm_vmx_nested_state_data {
 	__u8 vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE];
 	__u8 shadow_vmcs12[KVM_STATE_NESTED_VMX_VMCS_SIZE];
@@ -411,6 +415,15 @@ struct kvm_vmx_nested_state_hdr {
 	} smm;
 };
 
+struct kvm_svm_nested_state_data {
+	/* Save area only used if KVM_STATE_NESTED_RUN_PENDING.  */
+	__u8 vmcb12[KVM_STATE_NESTED_SVM_VMCB_SIZE];
+};
+
+struct kvm_svm_nested_state_hdr {
+	__u64 vmcb_pa;
+};
+
 /* for KVM_CAP_NESTED_STATE */
 struct kvm_nested_state {
 	__u16 flags;
@@ -419,6 +432,7 @@ struct kvm_nested_state {
 
 	union {
 		struct kvm_vmx_nested_state_hdr vmx;
+		struct kvm_svm_nested_state_hdr svm;
 
 		/* Pad the header to 128 bytes.  */
 		__u8 pad[120];
@@ -431,6 +445,7 @@ struct kvm_nested_state {
 	 */
 	union {
 		struct kvm_vmx_nested_state_data vmx[0];
+		struct kvm_svm_nested_state_data svm[0];
 	} data;
 };
 
diff --git a/arch/x86/kvm/cpuid.h b/arch/x86/kvm/cpuid.h
index 63a70f6a3df3..05434cd9342f 100644
--- a/arch/x86/kvm/cpuid.h
+++ b/arch/x86/kvm/cpuid.h
@@ -303,4 +303,9 @@ static __always_inline void kvm_cpu_cap_check_and_set(unsigned int x86_feature)
 		kvm_cpu_cap_set(x86_feature);
 }
 
+static inline bool page_address_valid(struct kvm_vcpu *vcpu, gpa_t gpa)
+{
+	return PAGE_ALIGNED(gpa) && !(gpa >> cpuid_maxphyaddr(vcpu));
+}
+
 #endif
diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c
index 087a04ae74e4..001d1f830076 100644
--- a/arch/x86/kvm/svm/nested.c
+++ b/arch/x86/kvm/svm/nested.c
@@ -25,6 +25,7 @@
 #include "trace.h"
 #include "mmu.h"
 #include "x86.h"
+#include "cpuid.h"
 #include "lapic.h"
 #include "svm.h"
 
@@ -243,6 +244,8 @@ static void load_nested_vmcb_control(struct vcpu_svm *svm,
 {
 	copy_vmcb_control_area(&svm->nested.ctl, control);
 
+	/* Copy it here because nested_svm_check_controls will check it.  */
+	svm->nested.ctl.asid           = control->asid;
 	svm->nested.ctl.msrpm_base_pa &= ~0x0fffULL;
 	svm->nested.ctl.iopm_base_pa  &= ~0x0fffULL;
 }
@@ -870,6 +873,150 @@ int nested_svm_exit_special(struct vcpu_svm *svm)
 	return NESTED_EXIT_CONTINUE;
 }
 
+static int svm_get_nested_state(struct kvm_vcpu *vcpu,
+				struct kvm_nested_state __user *user_kvm_nested_state,
+				u32 user_data_size)
+{
+	struct vcpu_svm *svm;
+	struct kvm_nested_state kvm_state = {
+		.flags = 0,
+		.format = KVM_STATE_NESTED_FORMAT_SVM,
+		.size = sizeof(kvm_state),
+	};
+	struct vmcb __user *user_vmcb = (struct vmcb __user *)
+		&user_kvm_nested_state->data.svm[0];
+
+	if (!vcpu)
+		return kvm_state.size + KVM_STATE_NESTED_SVM_VMCB_SIZE;
+
+	svm = to_svm(vcpu);
+
+	if (user_data_size < kvm_state.size)
+		goto out;
+
+	/* First fill in the header and copy it out.  */
+	if (is_guest_mode(vcpu)) {
+		kvm_state.hdr.svm.vmcb_pa = svm->nested.vmcb;
+		kvm_state.size += KVM_STATE_NESTED_SVM_VMCB_SIZE;
+		kvm_state.flags |= KVM_STATE_NESTED_GUEST_MODE;
+
+		if (svm->nested.nested_run_pending)
+			kvm_state.flags |= KVM_STATE_NESTED_RUN_PENDING;
+	}
+
+	if (gif_set(svm))
+		kvm_state.flags |= KVM_STATE_NESTED_GIF_SET;
+
+	if (copy_to_user(user_kvm_nested_state, &kvm_state, sizeof(kvm_state)))
+		return -EFAULT;
+
+	if (!is_guest_mode(vcpu))
+		goto out;
+
+	/*
+	 * Copy over the full size of the VMCB rather than just the size
+	 * of the struct.
+	 */
+	if (memzero_user(user_vmcb, KVM_STATE_NESTED_SVM_VMCB_SIZE))
+		return -EFAULT;
+	if (copy_to_user(&user_vmcb->control, &svm->nested.ctl,
+			 sizeof(user_vmcb->control)))
+		return -EFAULT;
+	if (copy_to_user(&user_vmcb->save, &svm->nested.hsave->save,
+			 sizeof(user_vmcb->save)))
+		return -EFAULT;
+
+out:
+	return kvm_state.size;
+}
+
+static int svm_set_nested_state(struct kvm_vcpu *vcpu,
+				struct kvm_nested_state __user *user_kvm_nested_state,
+				struct kvm_nested_state *kvm_state)
+{
+	struct vcpu_svm *svm = to_svm(vcpu);
+	struct vmcb *hsave = svm->nested.hsave;
+	struct vmcb __user *user_vmcb = (struct vmcb __user *)
+		&user_kvm_nested_state->data.svm[0];
+	struct vmcb_control_area ctl;
+	struct vmcb_save_area save;
+	u32 cr0;
+
+	if (kvm_state->format != KVM_STATE_NESTED_FORMAT_SVM)
+		return -EINVAL;
+
+	if (kvm_state->flags & ~(KVM_STATE_NESTED_GUEST_MODE |
+				 KVM_STATE_NESTED_RUN_PENDING |
+				 KVM_STATE_NESTED_GIF_SET))
+		return -EINVAL;
+
+	/*
+	 * If in guest mode, vcpu->arch.efer actually refers to the L2 guest's
+	 * EFER.SVME, but EFER.SVME still has to be 1 for VMRUN to succeed.
+	 */
+	if (!(vcpu->arch.efer & EFER_SVME)) {
+		/* GIF=1 and no guest mode are required if SVME=0.  */
+		if (kvm_state->flags != KVM_STATE_NESTED_GIF_SET)
+			return -EINVAL;
+	}
+
+	/* SMM temporarily disables SVM, so we cannot be in guest mode.  */
+	if (is_smm(vcpu) && (kvm_state->flags & KVM_STATE_NESTED_GUEST_MODE))
+		return -EINVAL;
+
+	if (!(kvm_state->flags & KVM_STATE_NESTED_GUEST_MODE)) {
+		svm_leave_nested(svm);
+		goto out_set_gif;
+	}
+
+	if (!page_address_valid(vcpu, kvm_state->hdr.svm.vmcb_pa))
+		return -EINVAL;
+	if (kvm_state->size < sizeof(*kvm_state) + KVM_STATE_NESTED_SVM_VMCB_SIZE)
+		return -EINVAL;
+	if (copy_from_user(&ctl, &user_vmcb->control, sizeof(ctl)))
+		return -EFAULT;
+	if (copy_from_user(&save, &user_vmcb->save, sizeof(save)))
+		return -EFAULT;
+
+	if (!nested_vmcb_check_controls(&ctl))
+		return -EINVAL;
+
+	/*
+	 * Processor state contains L2 state.  Check that it is
+	 * valid for guest mode (see nested_vmcb_checks).
+	 */
+	cr0 = kvm_read_cr0(vcpu);
+        if (((cr0 & X86_CR0_CD) == 0) && (cr0 & X86_CR0_NW))
+                return -EINVAL;
+
+	/*
+	 * Validate host state saved from before VMRUN (see
+	 * nested_svm_check_permissions).
+	 * TODO: validate reserved bits for all saved state.
+	 */
+	if (!(save.cr0 & X86_CR0_PG))
+		return -EINVAL;
+
+	/*
+	 * All checks done, we can enter guest mode.  L1 control fields
+	 * come from the nested save state.  Guest state is already
+	 * in the registers, the save area of the nested state instead
+	 * contains saved L1 state.
+	 */
+	copy_vmcb_control_area(&hsave->control, &svm->vmcb->control);
+	hsave->save = save;
+
+	svm->nested.vmcb = kvm_state->hdr.svm.vmcb_pa;
+	load_nested_vmcb_control(svm, &ctl);
+	nested_prepare_vmcb_control(svm);
+
+out_set_gif:
+	svm_set_gif(svm, !!(kvm_state->flags & KVM_STATE_NESTED_GIF_SET));
+	return 0;
+}
+
 struct kvm_x86_nested_ops svm_nested_ops = {
 	.check_events = svm_check_nested_events,
+	.get_state = svm_get_nested_state,
+	.set_state = svm_set_nested_state,
 };
diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c
index 56be704ffe95..f64d071715d2 100644
--- a/arch/x86/kvm/svm/svm.c
+++ b/arch/x86/kvm/svm/svm.c
@@ -1191,6 +1191,7 @@ static int svm_create_vcpu(struct kvm_vcpu *vcpu)
 		svm->avic_is_running = true;
 
 	svm->nested.hsave = page_address(hsave_page);
+	clear_page(svm->nested.hsave);
 
 	svm->msrpm = page_address(msrpm_pages);
 	svm_vcpu_init_msrpm(svm->msrpm);
diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c
index 51ebb60e1533..106fc6fceb97 100644
--- a/arch/x86/kvm/vmx/nested.c
+++ b/arch/x86/kvm/vmx/nested.c
@@ -437,11 +437,6 @@ static void vmx_inject_page_fault_nested(struct kvm_vcpu *vcpu,
 	}
 }
 
-static bool page_address_valid(struct kvm_vcpu *vcpu, gpa_t gpa)
-{
-	return PAGE_ALIGNED(gpa) && !(gpa >> cpuid_maxphyaddr(vcpu));
-}
-
 static int nested_vmx_check_io_bitmap_controls(struct kvm_vcpu *vcpu,
 					       struct vmcs12 *vmcs12)
 {
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 0001b2addc66..3e9252737f9d 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -4626,7 +4626,8 @@ long kvm_arch_vcpu_ioctl(struct file *filp,
 
 		if (kvm_state.flags &
 		    ~(KVM_STATE_NESTED_RUN_PENDING | KVM_STATE_NESTED_GUEST_MODE
-		      | KVM_STATE_NESTED_EVMCS | KVM_STATE_NESTED_MTF_PENDING))
+		      | KVM_STATE_NESTED_EVMCS | KVM_STATE_NESTED_MTF_PENDING
+		      | KVM_STATE_NESTED_GIF_SET))
 			break;
 
 		/* nested_run_pending implies guest_mode.  */
-- 
2.18.2


  parent reply	other threads:[~2020-05-20 17:22 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-05-20 17:21 [PATCH 00/24] KVM: nSVM: event fixes and migration support Paolo Bonzini
2020-05-20 17:21 ` [PATCH 01/24] KVM: nSVM: fix condition for filtering async PF Paolo Bonzini
2020-05-26  0:23   ` Sasha Levin
2020-05-20 17:21 ` [PATCH 02/24] KVM: nSVM: leave ASID aside in copy_vmcb_control_area Paolo Bonzini
2020-05-26  0:23   ` Sasha Levin
2020-05-20 17:21 ` [PATCH 03/24] KVM: nSVM: inject exceptions via svm_check_nested_events Paolo Bonzini
2020-05-20 17:21 ` [PATCH 04/24] KVM: nSVM: remove exit_required Paolo Bonzini
2020-05-20 17:21 ` [PATCH 05/24] KVM: nSVM: correctly inject INIT vmexits Paolo Bonzini
2020-05-20 17:21 ` [PATCH 06/24] KVM: nSVM: move map argument out of enter_svm_guest_mode Paolo Bonzini
2020-05-20 17:21 ` [PATCH 07/24] KVM: nSVM: extract load_nested_vmcb_control Paolo Bonzini
2020-05-20 17:21 ` [PATCH 08/24] KVM: nSVM: extract preparation of VMCB for nested run Paolo Bonzini
2020-05-20 17:21 ` [PATCH 09/24] KVM: nSVM: clean up tsc_offset update Paolo Bonzini
2020-05-20 17:21 ` [PATCH 10/24] KVM: nSVM: pass vmcb_control_area to copy_vmcb_control_area Paolo Bonzini
2020-05-20 17:21 ` [PATCH 11/24] KVM: nSVM: remove trailing padding for struct vmcb_control_area Paolo Bonzini
2020-05-20 17:21 ` [PATCH 12/24] KVM: nSVM: save all control fields in svm->nested Paolo Bonzini
2020-05-20 17:21 ` [PATCH 13/24] KVM: nSVM: do not reload pause filter fields from VMCB Paolo Bonzini
2020-05-20 17:21 ` [PATCH 14/24] KVM: nSVM: remove HF_VINTR_MASK Paolo Bonzini
2020-05-20 17:21 ` [PATCH 15/24] KVM: nSVM: remove HF_HIF_MASK Paolo Bonzini
2020-05-20 17:21 ` [PATCH 16/24] KVM: nSVM: split nested_vmcb_check_controls Paolo Bonzini
2020-05-20 17:21 ` [PATCH 17/24] KVM: nSVM: do all MMU switch work in init/uninit functions Paolo Bonzini
2020-05-20 17:21 ` [PATCH 18/24] KVM: nSVM: leave guest mode when clearing EFER.SVME Paolo Bonzini
2020-05-20 17:21 ` [PATCH 19/24] KVM: nSVM: extract svm_set_gif Paolo Bonzini
2020-05-20 17:21 ` [PATCH 20/24] KVM: MMU: pass arbitrary CR0/CR4/EFER to kvm_init_shadow_mmu Paolo Bonzini
2020-05-20 17:21 ` [PATCH 21/24] KVM: x86: always update CR3 in VMCB Paolo Bonzini
2020-05-20 18:22   ` Sean Christopherson
2020-05-20 20:14     ` Paolo Bonzini
2020-05-22 22:47       ` Sean Christopherson
2020-05-23  7:07         ` Paolo Bonzini
2020-05-20 18:24   ` Sean Christopherson
2020-05-20 17:21 ` [PATCH 22/24] uaccess: add memzero_user Paolo Bonzini
2020-05-20 20:40   ` Christoph Hellwig
2020-05-20 20:50     ` Al Viro
2020-05-20 21:13     ` Paolo Bonzini
2020-05-20 21:28       ` Al Viro
2020-05-20 21:33         ` Paolo Bonzini
2020-05-20 17:21 ` [PATCH 23/24] selftests: kvm: add a SVM version of state-test Paolo Bonzini
2020-05-20 17:21 ` Paolo Bonzini [this message]
2020-05-20 19:24 ` [PATCH 00/24] KVM: nSVM: event fixes and migration support Maxim Levitsky
2020-05-20 20:42   ` Paolo Bonzini
2020-05-20 21:08     ` Maxim Levitsky
2020-05-20 21:15       ` 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=20200520172145.23284-25-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=jroedel@suse.de \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=vkuznets@redhat.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.