[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20220311070346.45023-6-jiangshanlai@gmail.com>
Date: Fri, 11 Mar 2022 15:03:45 +0800
From: Lai Jiangshan <jiangshanlai@...il.com>
To: linux-kernel@...r.kernel.org, kvm@...r.kernel.org,
Paolo Bonzini <pbonzini@...hat.com>,
Sean Christopherson <seanjc@...gle.com>
Cc: Lai Jiangshan <jiangshan.ljs@...group.com>,
Vitaly Kuznetsov <vkuznets@...hat.com>,
Wanpeng Li <wanpengli@...cent.com>,
Jim Mattson <jmattson@...gle.com>,
Joerg Roedel <joro@...tes.org>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
"H. Peter Anvin" <hpa@...or.com>
Subject: [PATCH V2 5/5] KVM: X86: Only get rflags when needed in permission_fault()
From: Lai Jiangshan <jiangshan.ljs@...group.com>
The SMAP checking and rflags are only needed in permission_fault()
when it is supervisor access and SMAP is enabled. These information is
already encoded in the combination of mmu->permissions[] and the index.
So we can use the encoded information to see if we need the SMAP checking
instead of getting the rflags unconditionally.
Signed-off-by: Lai Jiangshan <jiangshan.ljs@...group.com>
---
arch/x86/kvm/mmu.h | 28 +++++++++++++++++++++++-----
1 file changed, 23 insertions(+), 5 deletions(-)
diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
index 4cb7a39ecd51..ceac1e9e21e9 100644
--- a/arch/x86/kvm/mmu.h
+++ b/arch/x86/kvm/mmu.h
@@ -218,13 +218,12 @@ static inline u8 permission_fault(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
{
/* strip nested paging fault error codes */
unsigned int pfec = access;
- unsigned long rflags = static_call(kvm_x86_get_rflags)(vcpu);
/*
* For explicit supervisor accesses, SMAP is disabled if EFLAGS.AC = 1.
* For implicit supervisor accesses, SMAP cannot be overridden.
*
- * SMAP works on supervisor accesses only, and not_smap can
+ * SMAP works on supervisor accesses only, and the SMAP checking bit can
* be set or not set when user access with neither has any bearing
* on the result.
*
@@ -233,11 +232,30 @@ static inline u8 permission_fault(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
* if SMAP checks are being disabled.
*/
bool explicit_access = !(access & PFERR_IMPLICIT_ACCESS);
- bool not_smap = (rflags & X86_EFLAGS_AC) && explicit_access;
- int index = (pfec + (!!not_smap << PFERR_RSVD_BIT)) >> 1;
- bool fault = (mmu->permissions[index] >> pte_access) & 1;
+ bool fault = (mmu->permissions[pfec >> 1] >> pte_access) & 1;
+ int index = (pfec + PFERR_RSVD_MASK) >> 1;
+ bool fault_not_smap = (mmu->permissions[index] >> pte_access) & 1;
u32 errcode = PFERR_PRESENT_MASK;
+ /*
+ * The value of fault has included SMAP checking if it is supervisor
+ * access and SMAP is enabled and encoded in mmu->permissions.
+ *
+ * fault fault_not_smap
+ * 0 0 not fault due to UWX nor SMAP
+ * 0 1 impossible combination
+ * 1 1 fault due to UWX
+ * 1 0 fault due to SMAP, need to check if
+ * SMAP is prevented
+ *
+ * SMAP is prevented only when X86_EFLAGS_AC is set on explicit
+ * supervisor access.
+ */
+ if (unlikely(fault && !fault_not_smap && explicit_access)) {
+ unsigned long rflags = static_call(kvm_x86_get_rflags)(vcpu);
+ fault = !(rflags & X86_EFLAGS_AC);
+ }
+
WARN_ON(pfec & (PFERR_PK_MASK | PFERR_RSVD_MASK));
if (unlikely(mmu->pkru_mask)) {
u32 pkru_bits, offset;
--
2.19.1.6.gb485710b
Powered by blists - more mailing lists