Received-SPF: pass (google.com: domain of linux-kernel+bounces-155951-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
Reply-To: Sean Christopherson <seanjc@google.com>
Date: Tue, 23 Apr 2024 15:15:21 -0700
In-Reply-To: <20240423221521.2923759-1-seanjc@google.com>
Precedence: bulk
Mime-Version: 1.0
References: <20240423221521.2923759-1-seanjc@google.com>
Message-ID: <20240423221521.2923759-5-seanjc@google.com>
Subject: [PATCH 4/4] KVM: x86: Move shadow_phys_bits into "kvm_host", as "maxphyaddr"
From: Sean Christopherson <seanjc@google.com>
To: Sean Christopherson <seanjc@google.com>, Paolo Bonzini <pbonzini@redhat.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"

Move shadow_phys_bits into "struct kvm_host_values", i.e. into KVM's
global "kvm_host" variable, so that it is automatically exported for use
in vendor modules.  Rename the variable/field to maxphyaddr to more
clearly capture what value it holds, now that it's used outside of the
MMU (and because the "shadow" part is more than a bit misleading as the
variable is not at all unique to shadow paging).

Recomputing the raw/true host.MAXPHYADDR on every use can be subtly
expensive, e.g. it will incur a VM-Exit on the CPUID if KVM is running as
a nested hypervisor.  Vendor code already has access to the information,
e.g. by directly doing CPUID or by invoking kvm_get_shadow_phys_bits(), so
there's no tangible benefit to making it MMU-only.

Signed-off-by: Sean Christopherson <seanjc@google.com>
---
 arch/x86/kvm/mmu.h      | 27 +--------------------------
 arch/x86/kvm/mmu/mmu.c  |  2 +-
 arch/x86/kvm/mmu/spte.c | 24 +++++++++++++++++++++---
 arch/x86/kvm/vmx/vmx.c  | 14 ++++++--------
 arch/x86/kvm/vmx/vmx.h  |  2 +-
 arch/x86/kvm/x86.h      |  7 +++++++
 6 files changed, 37 insertions(+), 39 deletions(-)

diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
index ef970aea26e7..0d63637f46d7 100644
--- a/arch/x86/kvm/mmu.h
+++ b/arch/x86/kvm/mmu.h
@@ -57,12 +57,6 @@ static __always_inline u64 rsvd_bits(int s, int e)
 	return ((2ULL << (e - s)) - 1) << s;
 }
 
-/*
- * The number of non-reserved physical address bits irrespective of features
- * that repurpose legal bits, e.g. MKTME.
- */
-extern u8 __ro_after_init shadow_phys_bits;
-
 static inline gfn_t kvm_mmu_max_gfn(void)
 {
 	/*
@@ -76,30 +70,11 @@ static inline gfn_t kvm_mmu_max_gfn(void)
 	 * than hardware's real MAXPHYADDR.  Using the host MAXPHYADDR
 	 * disallows such SPTEs entirely and simplifies the TDP MMU.
 	 */
-	int max_gpa_bits = likely(tdp_enabled) ? shadow_phys_bits : 52;
+	int max_gpa_bits = likely(tdp_enabled) ? kvm_host.maxphyaddr : 52;
 
 	return (1ULL << (max_gpa_bits - PAGE_SHIFT)) - 1;
 }
 
-static inline u8 kvm_get_shadow_phys_bits(void)
-{
-	/*
-	 * boot_cpu_data.x86_phys_bits is reduced when MKTME or SME are detected
-	 * in CPU detection code, but the processor treats those reduced bits as
-	 * 'keyID' thus they are not reserved bits. Therefore KVM needs to look at
-	 * the physical address bits reported by CPUID.
-	 */
-	if (likely(boot_cpu_data.extended_cpuid_level >= 0x80000008))
-		return cpuid_eax(0x80000008) & 0xff;
-
-	/*
-	 * Quite weird to have VMX or SVM but not MAXPHYADDR; probably a VM with
-	 * custom CPUID.  Proceed with whatever the kernel found since these features
-	 * aren't virtualizable (SME/SEV also require CPUIDs higher than 0x80000008).
-	 */
-	return boot_cpu_data.x86_phys_bits;
-}
-
 u8 kvm_mmu_get_max_tdp_level(void);
 
 void kvm_mmu_set_mmio_spte_mask(u64 mmio_value, u64 mmio_mask, u64 access_mask);
diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
index 12ad01929dce..c30bffa441cf 100644
--- a/arch/x86/kvm/mmu/mmu.c
+++ b/arch/x86/kvm/mmu/mmu.c
@@ -4933,7 +4933,7 @@ static void reset_rsvds_bits_mask_ept(struct kvm_vcpu *vcpu,
 
 static inline u64 reserved_hpa_bits(void)
 {
-	return rsvd_bits(shadow_phys_bits, 63);
+	return rsvd_bits(kvm_host.maxphyaddr, 63);
 }
 
 /*
diff --git a/arch/x86/kvm/mmu/spte.c b/arch/x86/kvm/mmu/spte.c
index 927f4abbe973..d49a3f928b0b 100644
--- a/arch/x86/kvm/mmu/spte.c
+++ b/arch/x86/kvm/mmu/spte.c
@@ -43,7 +43,25 @@ u64 __read_mostly shadow_acc_track_mask;
 u64 __read_mostly shadow_nonpresent_or_rsvd_mask;
 u64 __read_mostly shadow_nonpresent_or_rsvd_lower_gfn_mask;
 
-u8 __ro_after_init shadow_phys_bits;
+static u8 __init kvm_get_host_maxphyaddr(void)
+{
+	/*
+	 * boot_cpu_data.x86_phys_bits is reduced when MKTME or SME are detected
+	 * in CPU detection code, but the processor treats those reduced bits as
+	 * 'keyID' thus they are not reserved bits. Therefore KVM needs to look at
+	 * the physical address bits reported by CPUID, i.e. the raw MAXPHYADDR,
+	 * when reasoning about CPU behavior with respect to MAXPHYADDR.
+	 */
+	if (likely(boot_cpu_data.extended_cpuid_level >= 0x80000008))
+		return cpuid_eax(0x80000008) & 0xff;
+
+	/*
+	 * Quite weird to have VMX or SVM but not MAXPHYADDR; probably a VM with
+	 * custom CPUID.  Proceed with whatever the kernel found since these features
+	 * aren't virtualizable (SME/SEV also require CPUIDs higher than 0x80000008).
+	 */
+	return boot_cpu_data.x86_phys_bits;
+}
 
 void __init kvm_mmu_spte_module_init(void)
 {
@@ -56,7 +74,7 @@ void __init kvm_mmu_spte_module_init(void)
 	 */
 	allow_mmio_caching = enable_mmio_caching;
 
-	shadow_phys_bits = kvm_get_shadow_phys_bits();
+	kvm_host.maxphyaddr = kvm_get_host_maxphyaddr();
 }
 
 static u64 generation_mmio_spte_mask(u64 gen)
@@ -492,7 +510,7 @@ void kvm_mmu_reset_all_pte_masks(void)
 	 * 52-bit physical addresses then there are no reserved PA bits in the
 	 * PTEs and so the reserved PA approach must be disabled.
 	 */
-	if (shadow_phys_bits < 52)
+	if (kvm_host.maxphyaddr < 52)
 		mask = BIT_ULL(51) | PT_PRESENT_MASK;
 	else
 		mask = 0;
diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c
index cb1bd9aebac4..185b07bbbc16 100644
--- a/arch/x86/kvm/vmx/vmx.c
+++ b/arch/x86/kvm/vmx/vmx.c
@@ -8337,18 +8337,16 @@ static void __init vmx_setup_me_spte_mask(void)
 	u64 me_mask = 0;
 
 	/*
-	 * kvm_get_shadow_phys_bits() returns shadow_phys_bits.  Use
-	 * the former to avoid exposing shadow_phys_bits.
-	 *
 	 * On pre-MKTME system, boot_cpu_data.x86_phys_bits equals to
-	 * shadow_phys_bits.  On MKTME and/or TDX capable systems,
+	 * kvm_host.maxphyaddr.  On MKTME and/or TDX capable systems,
 	 * boot_cpu_data.x86_phys_bits holds the actual physical address
-	 * w/o the KeyID bits, and shadow_phys_bits equals to MAXPHYADDR
-	 * reported by CPUID.  Those bits between are KeyID bits.
+	 * w/o the KeyID bits, and kvm_host.maxphyaddr equals to
+	 * MAXPHYADDR reported by CPUID.  Those bits between are KeyID bits.
 	 */
-	if (boot_cpu_data.x86_phys_bits != kvm_get_shadow_phys_bits())
+	if (boot_cpu_data.x86_phys_bits != kvm_host.maxphyaddr)
 		me_mask = rsvd_bits(boot_cpu_data.x86_phys_bits,
-			kvm_get_shadow_phys_bits() - 1);
+				    kvm_host.maxphyaddr - 1);
+
 	/*
 	 * Unlike SME, host kernel doesn't support setting up any
 	 * MKTME KeyID on Intel platforms.  No memory encryption
diff --git a/arch/x86/kvm/vmx/vmx.h b/arch/x86/kvm/vmx/vmx.h
index 90f9e4434646..e7343023fbce 100644
--- a/arch/x86/kvm/vmx/vmx.h
+++ b/arch/x86/kvm/vmx/vmx.h
@@ -723,7 +723,7 @@ static inline bool vmx_need_pf_intercept(struct kvm_vcpu *vcpu)
 		return true;
 
 	return allow_smaller_maxphyaddr &&
-	       cpuid_maxphyaddr(vcpu) < kvm_get_shadow_phys_bits();
+	       cpuid_maxphyaddr(vcpu) < kvm_host.maxphyaddr;
 }
 
 static inline bool is_unrestricted_guest(struct kvm_vcpu *vcpu)
diff --git a/arch/x86/kvm/x86.h b/arch/x86/kvm/x86.h
index e69fff7d1f21..a88c65d3ea26 100644
--- a/arch/x86/kvm/x86.h
+++ b/arch/x86/kvm/x86.h
@@ -34,6 +34,13 @@ struct kvm_caps {
 };
 
 struct kvm_host_values {
+	/*
+	 * The host's raw MAXPHYADDR, i.e. the number of non-reserved physical
+	 * address bits irrespective of features that repurpose legal bits,
+	 * e.g. MKTME.
+	 */
+	u8 maxphyaddr;
+
 	u64 efer;
 	u64 xcr0;
 	u64 xss;
-- 
2.44.0.769.g3c40516874-goog