From: Sean Christopherson <seanjc@google.com>
To: Paolo Bonzini <pbonzini@redhat.com>,
Sean Christopherson <seanjc@google.com>,
Vitaly Kuznetsov <vkuznets@redhat.com>,
Jarkko Sakkinen <jarkko@kernel.org>
Cc: kvm@vger.kernel.org, linux-sgx@vger.kernel.org,
linux-kernel@vger.kernel.org,
Maxim Levitsky <mlevitsk@redhat.com>,
Hou Wenlong <houwenlong.hwl@antgroup.com>,
Xiaoyao Li <xiaoyao.li@intel.com>,
Kechen Lu <kechenl@nvidia.com>,
Oliver Upton <oliver.upton@linux.dev>,
Binbin Wu <binbin.wu@linux.intel.com>,
Yang Weijiang <weijiang.yang@intel.com>,
Robert Hoo <robert.hoo.linux@gmail.com>
Subject: [PATCH v3 33/57] KVM: x86: Remove unnecessary caching of KVM's PV CPUID base
Date: Wed, 27 Nov 2024 17:34:00 -0800 [thread overview]
Message-ID: <20241128013424.4096668-34-seanjc@google.com> (raw)
In-Reply-To: <20241128013424.4096668-1-seanjc@google.com>
Now that KVM only searches for KVM's PV CPUID base when userspace sets
guest CPUID, drop the cache and simply do the search every time.
Practically speaking, this is a nop except for situations where userspace
sets CPUID _after_ running the vCPU, which is anything but a hot path,
e.g. QEMU does so only when hotplugging a vCPU. And on the flip side,
caching guest CPUID information, especially information that is used to
query/modify _other_ CPUID state, is inherently dangerous as it's all too
easy to use stale information, i.e. KVM should only cache CPUID state when
the performance and/or programming benefits justify it.
Signed-off-by: Sean Christopherson <seanjc@google.com>
---
arch/x86/include/asm/kvm_host.h | 1 -
arch/x86/kvm/cpuid.c | 34 ++++++++-------------------------
2 files changed, 8 insertions(+), 27 deletions(-)
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index e159e44a6a1b..f076df9f18be 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -854,7 +854,6 @@ struct kvm_vcpu_arch {
int cpuid_nent;
struct kvm_cpuid_entry2 *cpuid_entries;
- struct kvm_hypervisor_cpuid kvm_cpuid;
bool is_amd_compatible;
/*
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 3ba0e6a67823..b402b9f59cbb 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -168,12 +168,7 @@ static int kvm_cpuid_check_equal(struct kvm_vcpu *vcpu, struct kvm_cpuid_entry2
/*
* Apply runtime CPUID updates to the incoming CPUID entries to avoid
- * false positives due mismatches on KVM-owned feature flags. Note,
- * runtime CPUID updates may consume other CPUID-driven vCPU state,
- * e.g. KVM or Xen CPUID bases. Updating runtime state before full
- * CPUID processing is functionally correct only because any change in
- * CPUID is disallowed, i.e. using stale data is ok because the below
- * checks will reject the change.
+ * false positives due mismatches on KVM-owned feature flags.
*
* Note! @e2 and @nent track the _old_ CPUID entries!
*/
@@ -231,28 +226,16 @@ static struct kvm_hypervisor_cpuid kvm_get_hypervisor_cpuid(struct kvm_vcpu *vcp
vcpu->arch.cpuid_nent, sig);
}
-static struct kvm_cpuid_entry2 *__kvm_find_kvm_cpuid_features(struct kvm_cpuid_entry2 *entries,
- int nent, u32 kvm_cpuid_base)
-{
- return cpuid_entry2_find(entries, nent, kvm_cpuid_base | KVM_CPUID_FEATURES,
- KVM_CPUID_INDEX_NOT_SIGNIFICANT);
-}
-
-static struct kvm_cpuid_entry2 *kvm_find_kvm_cpuid_features(struct kvm_vcpu *vcpu)
-{
- u32 base = vcpu->arch.kvm_cpuid.base;
-
- if (!base)
- return NULL;
-
- return __kvm_find_kvm_cpuid_features(vcpu->arch.cpuid_entries,
- vcpu->arch.cpuid_nent, base);
-}
-
static u32 kvm_apply_cpuid_pv_features_quirk(struct kvm_vcpu *vcpu)
{
- struct kvm_cpuid_entry2 *best = kvm_find_kvm_cpuid_features(vcpu);
+ struct kvm_hypervisor_cpuid kvm_cpuid;
+ struct kvm_cpuid_entry2 *best;
+ kvm_cpuid = kvm_get_hypervisor_cpuid(vcpu, KVM_SIGNATURE);
+ if (!kvm_cpuid.base)
+ return 0;
+
+ best = kvm_find_cpuid_entry(vcpu, kvm_cpuid.base | KVM_CPUID_FEATURES);
if (!best)
return 0;
@@ -483,7 +466,6 @@ static int kvm_set_cpuid(struct kvm_vcpu *vcpu, struct kvm_cpuid_entry2 *e2,
if (r)
goto err;
- vcpu->arch.kvm_cpuid = kvm_get_hypervisor_cpuid(vcpu, KVM_SIGNATURE);
#ifdef CONFIG_KVM_XEN
vcpu->arch.xen.cpuid = kvm_get_hypervisor_cpuid(vcpu, XEN_SIGNATURE);
#endif
--
2.47.0.338.g60cca15819-goog
next prev parent reply other threads:[~2024-11-28 1:35 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-28 1:33 [PATCH v3 00/57] KVM: x86: CPUID overhaul, fixes, and caching Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 01/57] KVM: x86: Use feature_bit() to clear CONSTANT_TSC when emulating CPUID Sean Christopherson
2024-12-13 10:53 ` Vitaly Kuznetsov
2024-11-28 1:33 ` [PATCH v3 02/57] KVM: x86: Limit use of F() and SF() to kvm_cpu_cap_{mask,init_kvm_defined}() Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 03/57] KVM: x86: Do all post-set CPUID processing during vCPU creation Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 04/57] KVM: x86: Explicitly do runtime CPUID updates "after" initial setup Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 05/57] KVM: x86: Account for KVM-reserved CR4 bits when passing through CR4 on VMX Sean Christopherson
2024-12-13 1:30 ` Chao Gao
2024-11-28 1:33 ` [PATCH v3 06/57] KVM: selftests: Update x86's set_sregs_test to match KVM's CPUID enforcement Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 07/57] KVM: selftests: Assert that vcpu->cpuid is non-NULL when getting CPUID entries Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 08/57] KVM: selftests: Refresh vCPU CPUID cache in __vcpu_get_cpuid_entry() Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 09/57] KVM: selftests: Verify KVM stuffs runtime CPUID OS bits on CR4 writes Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 10/57] KVM: x86: Move __kvm_is_valid_cr4() definition to x86.h Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 11/57] KVM: x86/pmu: Drop now-redundant refresh() during init() Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 12/57] KVM: x86: Drop now-redundant MAXPHYADDR and GPA rsvd bits from vCPU creation Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 13/57] KVM: x86: Disallow KVM_CAP_X86_DISABLE_EXITS after " Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 14/57] KVM: x86: Reject disabling of MWAIT/HLT interception when not allowed Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 15/57] KVM: x86: Drop the now unused KVM_X86_DISABLE_VALID_EXITS Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 16/57] KVM: selftests: Fix a bad TEST_REQUIRE() in x86's KVM PV test Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 17/57] KVM: selftests: Update x86's KVM PV test to match KVM's disabling exits behavior Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 18/57] KVM: x86: Zero out PV features cache when the CPUID leaf is not present Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 19/57] KVM: x86: Don't update PV features caches when enabling enforcement capability Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 20/57] KVM: x86: Do reverse CPUID sanity checks in __feature_leaf() Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 21/57] KVM: x86: Account for max supported CPUID leaf when getting raw host CPUID Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 22/57] KVM: x86: Unpack F() CPUID feature flag macros to one flag per line of code Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 23/57] KVM: x86: Rename kvm_cpu_cap_mask() to kvm_cpu_cap_init() Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 24/57] KVM: x86: Add a macro to init CPUID features that are 64-bit only Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 25/57] KVM: x86: Add a macro to precisely handle aliased 0x1.EDX CPUID features Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 26/57] KVM: x86: Handle kernel- and KVM-defined CPUID words in a single helper Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 27/57] KVM: x86: #undef SPEC_CTRL_SSBD in cpuid.c to avoid macro collisions Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 28/57] KVM: x86: Harden CPU capabilities processing against out-of-scope features Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 29/57] KVM: x86: Add a macro to init CPUID features that ignore host kernel support Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 30/57] KVM: x86: Add a macro to init CPUID features that KVM emulates in software Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 31/57] KVM: x86: Swap incoming guest CPUID into vCPU before massaging in KVM_SET_CPUID2 Sean Christopherson
2024-11-28 1:33 ` [PATCH v3 32/57] KVM: x86: Clear PV_UNHALT for !HLT-exiting only when userspace sets CPUID Sean Christopherson
2024-11-28 1:34 ` Sean Christopherson [this message]
2024-11-28 1:34 ` [PATCH v3 34/57] KVM: x86: Always operate on kvm_vcpu data in cpuid_entry2_find() Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 35/57] KVM: x86: Move kvm_find_cpuid_entry{,_index}() up near cpuid_entry2_find() Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 36/57] KVM: x86: Remove all direct usage of cpuid_entry2_find() Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 37/57] KVM: x86: Advertise TSC_DEADLINE_TIMER in KVM_GET_SUPPORTED_CPUID Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 38/57] KVM: x86: Advertise HYPERVISOR " Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 39/57] KVM: x86: Rename "governed features" helpers to use "guest_cpu_cap" Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 40/57] KVM: x86: Replace guts of "governed" features with comprehensive cpu_caps Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 41/57] KVM: x86: Initialize guest cpu_caps based on guest CPUID Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 42/57] KVM: x86: Extract code for generating per-entry emulated CPUID information Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 43/57] KVM: x86: Treat MONTIOR/MWAIT as a "partially emulated" feature Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 44/57] KVM: x86: Initialize guest cpu_caps based on KVM support Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 45/57] KVM: x86: Avoid double CPUID lookup when updating MWAIT at runtime Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 46/57] KVM: x86: Drop unnecessary check that cpuid_entry2_find() returns right leaf Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 47/57] KVM: x86: Update OS{XSAVE,PKE} bits in guest CPUID irrespective of host support Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 48/57] KVM: x86: Update guest cpu_caps at runtime for dynamic CPUID-based features Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 49/57] KVM: x86: Shuffle code to prepare for dropping guest_cpuid_has() Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 50/57] KVM: x86: Replace (almost) all guest CPUID feature queries with cpu_caps Sean Christopherson
2024-12-13 2:14 ` Chao Gao
2024-12-17 0:05 ` Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 51/57] KVM: x86: Drop superfluous host XSAVE check when adjusting guest XSAVES caps Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 52/57] KVM: x86: Add a macro for features that are synthesized into boot_cpu_data Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 53/57] KVM: x86: Pull CPUID capabilities from boot_cpu_data only as needed Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 54/57] KVM: x86: Rename "SF" macro to "SCATTERED_F" Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 55/57] KVM: x86: Explicitly track feature flags that require vendor enabling Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 56/57] KVM: x86: Explicitly track feature flags that are enabled at runtime Sean Christopherson
2024-11-28 1:34 ` [PATCH v3 57/57] KVM: x86: Use only local variables (no bitmask) to init kvm_cpu_caps Sean Christopherson
2024-12-18 1:15 ` Maxim Levitsky
2024-12-18 1:13 ` [PATCH v3 00/57] KVM: x86: CPUID overhaul, fixes, and caching Maxim Levitsky
2024-12-19 2:40 ` Sean Christopherson
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=20241128013424.4096668-34-seanjc@google.com \
--to=seanjc@google.com \
--cc=binbin.wu@linux.intel.com \
--cc=houwenlong.hwl@antgroup.com \
--cc=jarkko@kernel.org \
--cc=kechenl@nvidia.com \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-sgx@vger.kernel.org \
--cc=mlevitsk@redhat.com \
--cc=oliver.upton@linux.dev \
--cc=pbonzini@redhat.com \
--cc=robert.hoo.linux@gmail.com \
--cc=vkuznets@redhat.com \
--cc=weijiang.yang@intel.com \
--cc=xiaoyao.li@intel.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