* [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset @ 2022-01-17 8:26 Like Xu 2022-01-17 17:32 ` Paolo Bonzini 2022-01-18 18:30 ` [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Sean Christopherson 0 siblings, 2 replies; 9+ messages in thread From: Like Xu @ 2022-01-17 8:26 UTC (permalink / raw) To: Paolo Bonzini Cc: Wanpeng Li, Sean Christopherson, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel From: Like Xu <likexu@tencent.com> XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by both RESET and INIT. In both cases, the size in bytes of the XSAVE area containing all states enabled by XCR0 or (XCRO | IA32_XSS) needs to be updated. Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") Signed-off-by: Like Xu <likexu@tencent.com> --- arch/x86/kvm/x86.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 76b4803dd3bd..5748a57e1cb7 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -11134,6 +11134,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) struct kvm_cpuid_entry2 *cpuid_0x1; unsigned long old_cr0 = kvm_read_cr0(vcpu); unsigned long new_cr0; + bool need_update_cpuid = false; /* * Several of the "set" flows, e.g. ->set_cr0(), read other registers @@ -11199,6 +11200,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) vcpu->arch.msr_misc_features_enables = 0; + if (vcpu->arch.xcr0 != XFEATURE_MASK_FP) + need_update_cpuid = true; vcpu->arch.xcr0 = XFEATURE_MASK_FP; } @@ -11216,6 +11219,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); + if (vcpu->arch.ia32_xss) + need_update_cpuid = true; vcpu->arch.ia32_xss = 0; static_call(kvm_x86_vcpu_reset)(vcpu, init_event); @@ -11264,6 +11269,9 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) */ if (init_event) kvm_make_request(KVM_REQ_TLB_FLUSH_GUEST, vcpu); + + if (need_update_cpuid) + kvm_update_cpuid_runtime(vcpu); } EXPORT_SYMBOL_GPL(kvm_vcpu_reset); -- 2.33.1 ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset 2022-01-17 8:26 [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Like Xu @ 2022-01-17 17:32 ` Paolo Bonzini 2022-01-21 4:09 ` [PATCH v2] KVM: x86: Sync the states size with the XCR0/IA32_XSS at any time Like Xu 2022-01-18 18:30 ` [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Sean Christopherson 1 sibling, 1 reply; 9+ messages in thread From: Paolo Bonzini @ 2022-01-17 17:32 UTC (permalink / raw) To: Like Xu Cc: Wanpeng Li, Sean Christopherson, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel On 1/17/22 09:26, Like Xu wrote: > From: Like Xu <likexu@tencent.com> > > XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by > both RESET and INIT. In both cases, the size in bytes of the XSAVE > area containing all states enabled by XCR0 or (XCRO | IA32_XSS) > needs to be updated. > > Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") > Signed-off-by: Like Xu <likexu@tencent.com> Can you write a test case please? Thanks, Paolo > --- > arch/x86/kvm/x86.c | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 76b4803dd3bd..5748a57e1cb7 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -11134,6 +11134,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > struct kvm_cpuid_entry2 *cpuid_0x1; > unsigned long old_cr0 = kvm_read_cr0(vcpu); > unsigned long new_cr0; > + bool need_update_cpuid = false; > > /* > * Several of the "set" flows, e.g. ->set_cr0(), read other registers > @@ -11199,6 +11200,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > > vcpu->arch.msr_misc_features_enables = 0; > > + if (vcpu->arch.xcr0 != XFEATURE_MASK_FP) > + need_update_cpuid = true; > vcpu->arch.xcr0 = XFEATURE_MASK_FP; > } > > @@ -11216,6 +11219,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); > kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); > > + if (vcpu->arch.ia32_xss) > + need_update_cpuid = true; > vcpu->arch.ia32_xss = 0; > > static_call(kvm_x86_vcpu_reset)(vcpu, init_event); > @@ -11264,6 +11269,9 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > */ > if (init_event) > kvm_make_request(KVM_REQ_TLB_FLUSH_GUEST, vcpu); > + > + if (need_update_cpuid) > + kvm_update_cpuid_runtime(vcpu); > } > EXPORT_SYMBOL_GPL(kvm_vcpu_reset); > ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH v2] KVM: x86: Sync the states size with the XCR0/IA32_XSS at any time 2022-01-17 17:32 ` Paolo Bonzini @ 2022-01-21 4:09 ` Like Xu 2022-01-21 15:41 ` Sean Christopherson 0 siblings, 1 reply; 9+ messages in thread From: Like Xu @ 2022-01-21 4:09 UTC (permalink / raw) To: Paolo Bonzini Cc: Wanpeng Li, Sean Christopherson, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel From: Like Xu <likexu@tencent.com> XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by both RESET and INIT. The kvm_set_msr_common()'s handling of MSR_IA32_XSS also needs to update kvm_update_cpuid_runtime(). In the above cases, the size in bytes of the XSAVE area containing all states enabled by XCR0 or (XCRO | IA32_XSS) needs to be updated. For simplicity and consistency, legacy helpers are used to write values and call kvm_update_cpuid_runtime(), and it's not exactly a fast path. Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") Signed-off-by: Like Xu <likexu@tencent.com> --- v1 -> v2 Changelog: - Strongly prefer that use the helpers to write the values; (Sean) - Postpone IA32_XSS test cases once non-zero values are supported; (Paolo) - User space may call SET_CPUID2 after kvm_vcpu_reset(init_event=false); (Paolo) arch/x86/kvm/x86.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 55518b7d3b96..22d4b1d15e94 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -11256,7 +11256,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) vcpu->arch.msr_misc_features_enables = 0; - vcpu->arch.xcr0 = XFEATURE_MASK_FP; + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); } /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ @@ -11273,7 +11273,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); - vcpu->arch.ia32_xss = 0; + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); static_call(kvm_x86_vcpu_reset)(vcpu, init_event); -- 2.33.1 ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH v2] KVM: x86: Sync the states size with the XCR0/IA32_XSS at any time 2022-01-21 4:09 ` [PATCH v2] KVM: x86: Sync the states size with the XCR0/IA32_XSS at any time Like Xu @ 2022-01-21 15:41 ` Sean Christopherson 2022-01-23 4:37 ` [PATCH v3] KVM: x86: Sync the states size with the XCR0/IA32_XSS at, " Like Xu 0 siblings, 1 reply; 9+ messages in thread From: Sean Christopherson @ 2022-01-21 15:41 UTC (permalink / raw) To: Like Xu Cc: Paolo Bonzini, Wanpeng Li, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel On Fri, Jan 21, 2022, Like Xu wrote: > From: Like Xu <likexu@tencent.com> > > XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by > both RESET and INIT. The kvm_set_msr_common()'s handling of MSR_IA32_XSS > also needs to update kvm_update_cpuid_runtime(). In the above cases, the > size in bytes of the XSAVE area containing all states enabled by XCR0 or > (XCRO | IA32_XSS) needs to be updated. > > For simplicity and consistency, legacy helpers are used to write values s/legacy/existing "legacy" refers to something that is outdated/deprecated, which isn't what you intend here. > and call kvm_update_cpuid_runtime(), and it's not exactly a fast path. > > Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") > Signed-off-by: Like Xu <likexu@tencent.com> > --- > v1 -> v2 Changelog: > - Strongly prefer that use the helpers to write the values; (Sean) > - Postpone IA32_XSS test cases once non-zero values are supported; (Paolo) > - User space may call SET_CPUID2 after kvm_vcpu_reset(init_event=false); (Paolo) > > arch/x86/kvm/x86.c | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 55518b7d3b96..22d4b1d15e94 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -11256,7 +11256,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > > vcpu->arch.msr_misc_features_enables = 0; > > - vcpu->arch.xcr0 = XFEATURE_MASK_FP; > + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); > } > > /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ > @@ -11273,7 +11273,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); > kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); > > - vcpu->arch.ia32_xss = 0; > + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); This relies on a prep patch to invoke kvm_update_cpuid_runtime() in kvm_set_msr_common() for MSR_IA32_XSS, no? > static_call(kvm_x86_vcpu_reset)(vcpu, init_event); > > -- > 2.33.1 > > ^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH v3] KVM: x86: Sync the states size with the XCR0/IA32_XSS at, any time 2022-01-21 15:41 ` Sean Christopherson @ 2022-01-23 4:37 ` Like Xu 2022-01-26 16:37 ` Sean Christopherson 0 siblings, 1 reply; 9+ messages in thread From: Like Xu @ 2022-01-23 4:37 UTC (permalink / raw) To: Paolo Bonzini Cc: Wanpeng Li, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel, Sean Christopherson From: Like Xu <likexu@tencent.com> XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by both RESET and INIT. The kvm_set_msr_common()'s handling of MSR_IA32_XSS also needs to update kvm_update_cpuid_runtime(). In the above cases, the size in bytes of the XSAVE area containing all states enabled by XCR0 or (XCRO | IA32_XSS) needs to be updated. For simplicity and consistency, existing helpers are used to write values and call kvm_update_cpuid_runtime(), and it's not exactly a fast path. Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") Signed-off-by: Like Xu <likexu@tencent.com> --- v2 -> v3 Changelog: - Apply s/legacy/existing in the commit message; (Sean) - Invoke kvm_update_cpuid_runtime() for MSR_IA32_XSS; (Sean) arch/x86/kvm/x86.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 55518b7d3b96..4b509b26d9ab 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3535,6 +3535,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info) if (data & ~supported_xss) return 1; vcpu->arch.ia32_xss = data; + kvm_update_cpuid_runtime(vcpu); break; case MSR_SMI_COUNT: if (!msr_info->host_initiated) @@ -11256,7 +11257,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) vcpu->arch.msr_misc_features_enables = 0; - vcpu->arch.xcr0 = XFEATURE_MASK_FP; + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); } /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ @@ -11273,7 +11274,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); - vcpu->arch.ia32_xss = 0; + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); static_call(kvm_x86_vcpu_reset)(vcpu, init_event); -- 2.33.1 ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH v3] KVM: x86: Sync the states size with the XCR0/IA32_XSS at, any time 2022-01-23 4:37 ` [PATCH v3] KVM: x86: Sync the states size with the XCR0/IA32_XSS at, " Like Xu @ 2022-01-26 16:37 ` Sean Christopherson 2022-01-26 17:08 ` Paolo Bonzini 0 siblings, 1 reply; 9+ messages in thread From: Sean Christopherson @ 2022-01-26 16:37 UTC (permalink / raw) To: Like Xu Cc: Paolo Bonzini, Wanpeng Li, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel On Sun, Jan 23, 2022, Like Xu wrote: > From: Like Xu <likexu@tencent.com> > > XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by > both RESET and INIT. The kvm_set_msr_common()'s handling of MSR_IA32_XSS > also needs to update kvm_update_cpuid_runtime(). In the above cases, the > size in bytes of the XSAVE area containing all states enabled by XCR0 or > (XCRO | IA32_XSS) needs to be updated. > > For simplicity and consistency, existing helpers are used to write values > and call kvm_update_cpuid_runtime(), and it's not exactly a fast path. > > Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") > Signed-off-by: Like Xu <likexu@tencent.com> > --- > v2 -> v3 Changelog: > - Apply s/legacy/existing in the commit message; (Sean) > - Invoke kvm_update_cpuid_runtime() for MSR_IA32_XSS; (Sean) > > arch/x86/kvm/x86.c | 5 +++-- > 1 file changed, 3 insertions(+), 2 deletions(-) > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 55518b7d3b96..4b509b26d9ab 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -3535,6 +3535,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct > msr_data *msr_info) > if (data & ~supported_xss) > return 1; > vcpu->arch.ia32_xss = data; > + kvm_update_cpuid_runtime(vcpu); > break; > case MSR_SMI_COUNT: > if (!msr_info->host_initiated) > @@ -11256,7 +11257,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > > vcpu->arch.msr_misc_features_enables = 0; > > - vcpu->arch.xcr0 = XFEATURE_MASK_FP; > + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); > } > > /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ > @@ -11273,7 +11274,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); > kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); > > - vcpu->arch.ia32_xss = 0; > + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); Heh, this now conflicts with a patch Xiaoyao just posted, turns out the SDM was wrong. I think there's also some whitespace change or something that prevents this from applying cleanly. For convenience, I'll post a miniseries with this and Xiaoyao's patch. [*] https://lore.kernel.org/all/20220126034750.2495371-1-xiaoyao.li@intel.com > > static_call(kvm_x86_vcpu_reset)(vcpu, init_event); > > -- > 2.33.1 > > ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH v3] KVM: x86: Sync the states size with the XCR0/IA32_XSS at, any time 2022-01-26 16:37 ` Sean Christopherson @ 2022-01-26 17:08 ` Paolo Bonzini 0 siblings, 0 replies; 9+ messages in thread From: Paolo Bonzini @ 2022-01-26 17:08 UTC (permalink / raw) To: Sean Christopherson, Like Xu Cc: Wanpeng Li, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel On 1/26/22 17:37, Sean Christopherson wrote: > On Sun, Jan 23, 2022, Like Xu wrote: >> From: Like Xu <likexu@tencent.com> >> >> XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by >> both RESET and INIT. The kvm_set_msr_common()'s handling of MSR_IA32_XSS >> also needs to update kvm_update_cpuid_runtime(). In the above cases, the >> size in bytes of the XSAVE area containing all states enabled by XCR0 or >> (XCRO | IA32_XSS) needs to be updated. >> >> For simplicity and consistency, existing helpers are used to write values >> and call kvm_update_cpuid_runtime(), and it's not exactly a fast path. >> >> Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") >> Signed-off-by: Like Xu <likexu@tencent.com> >> --- >> v2 -> v3 Changelog: >> - Apply s/legacy/existing in the commit message; (Sean) >> - Invoke kvm_update_cpuid_runtime() for MSR_IA32_XSS; (Sean) >> >> arch/x86/kvm/x86.c | 5 +++-- >> 1 file changed, 3 insertions(+), 2 deletions(-) >> >> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c >> index 55518b7d3b96..4b509b26d9ab 100644 >> --- a/arch/x86/kvm/x86.c >> +++ b/arch/x86/kvm/x86.c >> @@ -3535,6 +3535,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct >> msr_data *msr_info) >> if (data & ~supported_xss) >> return 1; >> vcpu->arch.ia32_xss = data; >> + kvm_update_cpuid_runtime(vcpu); >> break; >> case MSR_SMI_COUNT: >> if (!msr_info->host_initiated) >> @@ -11256,7 +11257,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) >> >> vcpu->arch.msr_misc_features_enables = 0; >> >> - vcpu->arch.xcr0 = XFEATURE_MASK_FP; >> + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); >> } >> >> /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ >> @@ -11273,7 +11274,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) >> cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); >> kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); >> >> - vcpu->arch.ia32_xss = 0; >> + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); > > Heh, this now conflicts with a patch Xiaoyao just posted, turns out the SDM was > wrong. I think there's also some whitespace change or something that prevents > this from applying cleanly. For convenience, I'll post a miniseries with this > and Xiaoyao's patch. > > [*] https://lore.kernel.org/all/20220126034750.2495371-1-xiaoyao.li@intel.com It's okay, fixing the conflict in Xiaoyao's patch is trivial and---anyway---KVM does not support XSS != 0 for now so his patch doesn't really have any practical effect. Like's patch queued, thanks. Paolo ^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset 2022-01-17 8:26 [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Like Xu 2022-01-17 17:32 ` Paolo Bonzini @ 2022-01-18 18:30 ` Sean Christopherson 2022-01-19 7:25 ` Like Xu 1 sibling, 1 reply; 9+ messages in thread From: Sean Christopherson @ 2022-01-18 18:30 UTC (permalink / raw) To: Like Xu Cc: Paolo Bonzini, Wanpeng Li, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel On Mon, Jan 17, 2022, Like Xu wrote: > From: Like Xu <likexu@tencent.com> > > XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by > both RESET and INIT. In both cases, the size in bytes of the XSAVE > area containing all states enabled by XCR0 or (XCRO | IA32_XSS) > needs to be updated. > > Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") > Signed-off-by: Like Xu <likexu@tencent.com> > --- > arch/x86/kvm/x86.c | 8 ++++++++ > 1 file changed, 8 insertions(+) > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 76b4803dd3bd..5748a57e1cb7 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -11134,6 +11134,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > struct kvm_cpuid_entry2 *cpuid_0x1; > unsigned long old_cr0 = kvm_read_cr0(vcpu); > unsigned long new_cr0; > + bool need_update_cpuid = false; > > /* > * Several of the "set" flows, e.g. ->set_cr0(), read other registers > @@ -11199,6 +11200,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > > vcpu->arch.msr_misc_features_enables = 0; > > + if (vcpu->arch.xcr0 != XFEATURE_MASK_FP) > + need_update_cpuid = true; > vcpu->arch.xcr0 = XFEATURE_MASK_FP; > } > > @@ -11216,6 +11219,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); > kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); > > + if (vcpu->arch.ia32_xss) > + need_update_cpuid = true; This means that kvm_set_msr_common()'s handling of MSR_IA32_XSS also needs to update kvm_update_cpuid_runtime(). And then for bnoth XCR0 and XSS, I would very strongly prefer that use the helpers to write the values and let the helpers call kvm_update_cpuid_runtime(). Yes, that will mean kvm_update_cpuid_runtime() may be called multiple times during INIT, but that's already true (CR4), and this isn't exactly a fast path. diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 55518b7d3b96..22d4b1d15e94 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -11256,7 +11256,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) vcpu->arch.msr_misc_features_enables = 0; - vcpu->arch.xcr0 = XFEATURE_MASK_FP; + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); } /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ @@ -11273,7 +11273,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); - vcpu->arch.ia32_xss = 0; + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); static_call(kvm_x86_vcpu_reset)(vcpu, init_event); ^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset 2022-01-18 18:30 ` [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Sean Christopherson @ 2022-01-19 7:25 ` Like Xu 0 siblings, 0 replies; 9+ messages in thread From: Like Xu @ 2022-01-19 7:25 UTC (permalink / raw) To: Sean Christopherson Cc: Paolo Bonzini, Wanpeng Li, Jim Mattson, Vitaly Kuznetsov, Joerg Roedel, kvm, linux-kernel On 19/1/2022 2:30 am, Sean Christopherson wrote: > On Mon, Jan 17, 2022, Like Xu wrote: >> From: Like Xu <likexu@tencent.com> >> >> XCR0 is reset to 1 by RESET but not INIT and IA32_XSS is zeroed by >> both RESET and INIT. In both cases, the size in bytes of the XSAVE >> area containing all states enabled by XCR0 or (XCRO | IA32_XSS) >> needs to be updated. >> >> Fixes: a554d207dc46 ("KVM: X86: Processor States following Reset or INIT") >> Signed-off-by: Like Xu <likexu@tencent.com> >> --- >> arch/x86/kvm/x86.c | 8 ++++++++ >> 1 file changed, 8 insertions(+) >> >> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c >> index 76b4803dd3bd..5748a57e1cb7 100644 >> --- a/arch/x86/kvm/x86.c >> +++ b/arch/x86/kvm/x86.c >> @@ -11134,6 +11134,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) >> struct kvm_cpuid_entry2 *cpuid_0x1; >> unsigned long old_cr0 = kvm_read_cr0(vcpu); >> unsigned long new_cr0; >> + bool need_update_cpuid = false; >> >> /* >> * Several of the "set" flows, e.g. ->set_cr0(), read other registers >> @@ -11199,6 +11200,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) >> >> vcpu->arch.msr_misc_features_enables = 0; >> >> + if (vcpu->arch.xcr0 != XFEATURE_MASK_FP) >> + need_update_cpuid = true; >> vcpu->arch.xcr0 = XFEATURE_MASK_FP; >> } >> >> @@ -11216,6 +11219,8 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) >> cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); >> kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); >> >> + if (vcpu->arch.ia32_xss) >> + need_update_cpuid = true; > > This means that kvm_set_msr_common()'s handling of MSR_IA32_XSS also needs to > update kvm_update_cpuid_runtime(). And then for bnoth XCR0 and XSS, I would very > strongly prefer that use the helpers to write the values and let the helpers call Looks good to me and let me apply it in the next version. > kvm_update_cpuid_runtime(). Yes, that will mean kvm_update_cpuid_runtime() may be > called multiple times during INIT, but that's already true (CR4), and this isn't > exactly a fast path. An undisclosed lazy mechanism is under analyzed for performance gains. > > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 55518b7d3b96..22d4b1d15e94 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -11256,7 +11256,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > > vcpu->arch.msr_misc_features_enables = 0; > > - vcpu->arch.xcr0 = XFEATURE_MASK_FP; > + __kvm_set_xcr(vcpu, 0, XFEATURE_MASK_FP); > } > > /* All GPRs except RDX (handled below) are zeroed on RESET/INIT. */ > @@ -11273,7 +11273,7 @@ void kvm_vcpu_reset(struct kvm_vcpu *vcpu, bool init_event) > cpuid_0x1 = kvm_find_cpuid_entry(vcpu, 1, 0); > kvm_rdx_write(vcpu, cpuid_0x1 ? cpuid_0x1->eax : 0x600); > > - vcpu->arch.ia32_xss = 0; > + __kvm_set_msr(vcpu, MSR_IA32_XSS, 0, true); > > static_call(kvm_x86_vcpu_reset)(vcpu, init_event); > > ^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2022-01-26 17:08 UTC | newest] Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2022-01-17 8:26 [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Like Xu 2022-01-17 17:32 ` Paolo Bonzini 2022-01-21 4:09 ` [PATCH v2] KVM: x86: Sync the states size with the XCR0/IA32_XSS at any time Like Xu 2022-01-21 15:41 ` Sean Christopherson 2022-01-23 4:37 ` [PATCH v3] KVM: x86: Sync the states size with the XCR0/IA32_XSS at, " Like Xu 2022-01-26 16:37 ` Sean Christopherson 2022-01-26 17:08 ` Paolo Bonzini 2022-01-18 18:30 ` [PATCH] KVM: x86: Update the states size cpuid even if XCR0/IA32_XSS is reset Sean Christopherson 2022-01-19 7:25 ` Like Xu
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.