| Message ID | 20160113070759.20248.86252.stgit@aravindap |
|---|---|
| State | Changes Requested |
| Headers | show |
On Wed, Jan 13, 2016 at 12:37:59PM +0530, Aravinda Prasad wrote: > This patch introduces a new KVM capability to control > how KVM behaves on machine check exception (MCE). > Without this capability, KVM redirects machine check > exceptions to guest's 0x200 vector if the address in > error belongs to the guest. With this capability KVM > causes a guest exit with NMI exit reason. > > This is required to avoid problems if a new kernel/KVM > is used with an old QEMU for guests that don't issue > "ibm,nmi-register". As old QEMU does not understand the > NMI exit type, it treats it as a fatal error. However, > the guest could have handled the machine check error > if the exception was delivered to guest's 0x200 interrupt > vector instead of NMI exit in case of old QEMU. > > QEMU part can be found at: > http://lists.nongnu.org/archive/html/qemu-ppc/2015-12/msg00199.html > > Change Log v3: > - Split the patch into 2. First patch introduces the > new capability while the second one enhances KVM to > redirect MCE. > - Fix access width bug > - Rebased to v4.4-rc7 > > Change Log v2: > - Added KVM capability > > Signed-off-by: Aravinda Prasad <aravinda@linux.vnet.ibm.com> > --- > arch/powerpc/include/asm/kvm_host.h | 1 + > arch/powerpc/kernel/asm-offsets.c | 1 + > arch/powerpc/kvm/powerpc.c | 7 +++++++ > include/uapi/linux/kvm.h | 1 + > 4 files changed, 10 insertions(+) > > diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h > index cfa758c..9ac2b84 100644 > --- a/arch/powerpc/include/asm/kvm_host.h > +++ b/arch/powerpc/include/asm/kvm_host.h > @@ -243,6 +243,7 @@ struct kvm_arch { > int hpt_cma_alloc; > struct dentry *debugfs_dir; > struct dentry *htab_dentry; > + u8 fwnmi_enabled; Um.. I don't see anything in this patch or 2/2 which actually tests this flag... > #endif /* CONFIG_KVM_BOOK3S_HV_POSSIBLE */ > #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE > struct mutex hpt_mutex; > diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c > index 221d584..6a4e81a 100644 > --- a/arch/powerpc/kernel/asm-offsets.c > +++ b/arch/powerpc/kernel/asm-offsets.c > @@ -506,6 +506,7 @@ int main(void) > DEFINE(KVM_ENABLED_HCALLS, offsetof(struct kvm, arch.enabled_hcalls)); > DEFINE(KVM_LPCR, offsetof(struct kvm, arch.lpcr)); > DEFINE(KVM_VRMA_SLB_V, offsetof(struct kvm, arch.vrma_slb_v)); > + DEFINE(KVM_FWNMI, offsetof(struct kvm, arch.fwnmi_enabled)); > DEFINE(VCPU_DSISR, offsetof(struct kvm_vcpu, arch.shregs.dsisr)); > DEFINE(VCPU_DAR, offsetof(struct kvm_vcpu, arch.shregs.dar)); > DEFINE(VCPU_VPA, offsetof(struct kvm_vcpu, arch.vpa.pinned_addr)); > diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c > index 6fd2405..a8399b5 100644 > --- a/arch/powerpc/kvm/powerpc.c > +++ b/arch/powerpc/kvm/powerpc.c > @@ -570,6 +570,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext) > r = 1; > break; > #endif > + case KVM_CAP_PPC_FWNMI: > + r = 1; > + break; > default: > r = 0; > break; > @@ -1132,6 +1135,10 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu, > break; > } > #endif /* CONFIG_KVM_XICS */ > + case KVM_CAP_PPC_FWNMI: > + r = 0; > + vcpu->kvm->arch.fwnmi_enabled = true; > + break; > default: > r = -EINVAL; > break; > diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h > index 03f3618..d8a07b5 100644 > --- a/include/uapi/linux/kvm.h > +++ b/include/uapi/linux/kvm.h > @@ -831,6 +831,7 @@ struct kvm_ppc_smmu_info { > #define KVM_CAP_GUEST_DEBUG_HW_WPS 120 > #define KVM_CAP_SPLIT_IRQCHIP 121 > #define KVM_CAP_IOEVENTFD_ANY_LENGTH 122 > +#define KVM_CAP_PPC_FWNMI 123 > > #ifdef KVM_CAP_IRQ_ROUTING > >
On Thu, Jan 14, 2016 at 11:02:39AM +1100, David Gibson wrote: > On Wed, Jan 13, 2016 at 12:37:59PM +0530, Aravinda Prasad wrote: > > This patch introduces a new KVM capability to control > > how KVM behaves on machine check exception (MCE). > > Without this capability, KVM redirects machine check > > exceptions to guest's 0x200 vector if the address in > > error belongs to the guest. With this capability KVM > > causes a guest exit with NMI exit reason. > > > > This is required to avoid problems if a new kernel/KVM > > is used with an old QEMU for guests that don't issue > > "ibm,nmi-register". As old QEMU does not understand the > > NMI exit type, it treats it as a fatal error. However, > > the guest could have handled the machine check error > > if the exception was delivered to guest's 0x200 interrupt > > vector instead of NMI exit in case of old QEMU. > > > > QEMU part can be found at: > > http://lists.nongnu.org/archive/html/qemu-ppc/2015-12/msg00199.html > > > > Change Log v3: > > - Split the patch into 2. First patch introduces the > > new capability while the second one enhances KVM to > > redirect MCE. > > - Fix access width bug > > - Rebased to v4.4-rc7 > > > > Change Log v2: > > - Added KVM capability > > > > Signed-off-by: Aravinda Prasad <aravinda@linux.vnet.ibm.com> > > --- > > arch/powerpc/include/asm/kvm_host.h | 1 + > > arch/powerpc/kernel/asm-offsets.c | 1 + > > arch/powerpc/kvm/powerpc.c | 7 +++++++ > > include/uapi/linux/kvm.h | 1 + > > 4 files changed, 10 insertions(+) > > > > diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h > > index cfa758c..9ac2b84 100644 > > --- a/arch/powerpc/include/asm/kvm_host.h > > +++ b/arch/powerpc/include/asm/kvm_host.h > > @@ -243,6 +243,7 @@ struct kvm_arch { > > int hpt_cma_alloc; > > struct dentry *debugfs_dir; > > struct dentry *htab_dentry; > > + u8 fwnmi_enabled; > > Um.. I don't see anything in this patch or 2/2 which actually tests > this flag... Sorry, I missed it in the asm, spotted it now. Reviewed-by: David Gibson <david@gibson.dropbear.id.au> > > > #endif /* CONFIG_KVM_BOOK3S_HV_POSSIBLE */ > > #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE > > struct mutex hpt_mutex; > > diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c > > index 221d584..6a4e81a 100644 > > --- a/arch/powerpc/kernel/asm-offsets.c > > +++ b/arch/powerpc/kernel/asm-offsets.c > > @@ -506,6 +506,7 @@ int main(void) > > DEFINE(KVM_ENABLED_HCALLS, offsetof(struct kvm, arch.enabled_hcalls)); > > DEFINE(KVM_LPCR, offsetof(struct kvm, arch.lpcr)); > > DEFINE(KVM_VRMA_SLB_V, offsetof(struct kvm, arch.vrma_slb_v)); > > + DEFINE(KVM_FWNMI, offsetof(struct kvm, arch.fwnmi_enabled)); > > DEFINE(VCPU_DSISR, offsetof(struct kvm_vcpu, arch.shregs.dsisr)); > > DEFINE(VCPU_DAR, offsetof(struct kvm_vcpu, arch.shregs.dar)); > > DEFINE(VCPU_VPA, offsetof(struct kvm_vcpu, arch.vpa.pinned_addr)); > > diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c > > index 6fd2405..a8399b5 100644 > > --- a/arch/powerpc/kvm/powerpc.c > > +++ b/arch/powerpc/kvm/powerpc.c > > @@ -570,6 +570,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext) > > r = 1; > > break; > > #endif > > + case KVM_CAP_PPC_FWNMI: > > + r = 1; > > + break; > > default: > > r = 0; > > break; > > @@ -1132,6 +1135,10 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu, > > break; > > } > > #endif /* CONFIG_KVM_XICS */ > > + case KVM_CAP_PPC_FWNMI: > > + r = 0; > > + vcpu->kvm->arch.fwnmi_enabled = true; > > + break; > > default: > > r = -EINVAL; > > break; > > diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h > > index 03f3618..d8a07b5 100644 > > --- a/include/uapi/linux/kvm.h > > +++ b/include/uapi/linux/kvm.h > > @@ -831,6 +831,7 @@ struct kvm_ppc_smmu_info { > > #define KVM_CAP_GUEST_DEBUG_HW_WPS 120 > > #define KVM_CAP_SPLIT_IRQCHIP 121 > > #define KVM_CAP_IOEVENTFD_ANY_LENGTH 122 > > +#define KVM_CAP_PPC_FWNMI 123 > > > > #ifdef KVM_CAP_IRQ_ROUTING > > > > >
On Wed, Jan 13, 2016 at 12:37:59PM +0530, Aravinda Prasad wrote: > This patch introduces a new KVM capability to control > how KVM behaves on machine check exception (MCE). > Without this capability, KVM redirects machine check > exceptions to guest's 0x200 vector if the address in > error belongs to the guest. With this capability KVM > causes a guest exit with NMI exit reason. > > This is required to avoid problems if a new kernel/KVM > is used with an old QEMU for guests that don't issue > "ibm,nmi-register". As old QEMU does not understand the > NMI exit type, it treats it as a fatal error. However, > the guest could have handled the machine check error > if the exception was delivered to guest's 0x200 interrupt > vector instead of NMI exit in case of old QEMU. [snip] > @@ -1132,6 +1135,10 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu, > break; > } > #endif /* CONFIG_KVM_XICS */ > + case KVM_CAP_PPC_FWNMI: > + r = 0; > + vcpu->kvm->arch.fwnmi_enabled = true; > + break; Might we ever want to set this flag back to false after setting it to true? If so perhaps we should do vcpu->kvm->arch.fwnmi_enabled = !!cap->args[0]. However, I admit I can't actually think of a situation where we would need to reset it. :) Paul. -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Saturday 23 January 2016 03:50 PM, Paul Mackerras wrote: > On Wed, Jan 13, 2016 at 12:37:59PM +0530, Aravinda Prasad wrote: >> This patch introduces a new KVM capability to control >> how KVM behaves on machine check exception (MCE). >> Without this capability, KVM redirects machine check >> exceptions to guest's 0x200 vector if the address in >> error belongs to the guest. With this capability KVM >> causes a guest exit with NMI exit reason. >> >> This is required to avoid problems if a new kernel/KVM >> is used with an old QEMU for guests that don't issue >> "ibm,nmi-register". As old QEMU does not understand the >> NMI exit type, it treats it as a fatal error. However, >> the guest could have handled the machine check error >> if the exception was delivered to guest's 0x200 interrupt >> vector instead of NMI exit in case of old QEMU. > > [snip] > >> @@ -1132,6 +1135,10 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu, >> break; >> } >> #endif /* CONFIG_KVM_XICS */ >> + case KVM_CAP_PPC_FWNMI: >> + r = 0; >> + vcpu->kvm->arch.fwnmi_enabled = true; >> + break; > > Might we ever want to set this flag back to false after setting it to > true? If so perhaps we should do vcpu->kvm->arch.fwnmi_enabled = > !!cap->args[0]. However, I admit I can't actually think of a > situation where we would need to reset it. :) Even I am not able to think of any situation where resetting is required. Regards, Aravinda > > Paul. >
diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h index cfa758c..9ac2b84 100644 --- a/arch/powerpc/include/asm/kvm_host.h +++ b/arch/powerpc/include/asm/kvm_host.h @@ -243,6 +243,7 @@ struct kvm_arch { int hpt_cma_alloc; struct dentry *debugfs_dir; struct dentry *htab_dentry; + u8 fwnmi_enabled; #endif /* CONFIG_KVM_BOOK3S_HV_POSSIBLE */ #ifdef CONFIG_KVM_BOOK3S_PR_POSSIBLE struct mutex hpt_mutex; diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c index 221d584..6a4e81a 100644 --- a/arch/powerpc/kernel/asm-offsets.c +++ b/arch/powerpc/kernel/asm-offsets.c @@ -506,6 +506,7 @@ int main(void) DEFINE(KVM_ENABLED_HCALLS, offsetof(struct kvm, arch.enabled_hcalls)); DEFINE(KVM_LPCR, offsetof(struct kvm, arch.lpcr)); DEFINE(KVM_VRMA_SLB_V, offsetof(struct kvm, arch.vrma_slb_v)); + DEFINE(KVM_FWNMI, offsetof(struct kvm, arch.fwnmi_enabled)); DEFINE(VCPU_DSISR, offsetof(struct kvm_vcpu, arch.shregs.dsisr)); DEFINE(VCPU_DAR, offsetof(struct kvm_vcpu, arch.shregs.dar)); DEFINE(VCPU_VPA, offsetof(struct kvm_vcpu, arch.vpa.pinned_addr)); diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c index 6fd2405..a8399b5 100644 --- a/arch/powerpc/kvm/powerpc.c +++ b/arch/powerpc/kvm/powerpc.c @@ -570,6 +570,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext) r = 1; break; #endif + case KVM_CAP_PPC_FWNMI: + r = 1; + break; default: r = 0; break; @@ -1132,6 +1135,10 @@ static int kvm_vcpu_ioctl_enable_cap(struct kvm_vcpu *vcpu, break; } #endif /* CONFIG_KVM_XICS */ + case KVM_CAP_PPC_FWNMI: + r = 0; + vcpu->kvm->arch.fwnmi_enabled = true; + break; default: r = -EINVAL; break; diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h index 03f3618..d8a07b5 100644 --- a/include/uapi/linux/kvm.h +++ b/include/uapi/linux/kvm.h @@ -831,6 +831,7 @@ struct kvm_ppc_smmu_info { #define KVM_CAP_GUEST_DEBUG_HW_WPS 120 #define KVM_CAP_SPLIT_IRQCHIP 121 #define KVM_CAP_IOEVENTFD_ANY_LENGTH 122 +#define KVM_CAP_PPC_FWNMI 123 #ifdef KVM_CAP_IRQ_ROUTING
This patch introduces a new KVM capability to control how KVM behaves on machine check exception (MCE). Without this capability, KVM redirects machine check exceptions to guest's 0x200 vector if the address in error belongs to the guest. With this capability KVM causes a guest exit with NMI exit reason. This is required to avoid problems if a new kernel/KVM is used with an old QEMU for guests that don't issue "ibm,nmi-register". As old QEMU does not understand the NMI exit type, it treats it as a fatal error. However, the guest could have handled the machine check error if the exception was delivered to guest's 0x200 interrupt vector instead of NMI exit in case of old QEMU. QEMU part can be found at: http://lists.nongnu.org/archive/html/qemu-ppc/2015-12/msg00199.html Change Log v3: - Split the patch into 2. First patch introduces the new capability while the second one enhances KVM to redirect MCE. - Fix access width bug - Rebased to v4.4-rc7 Change Log v2: - Added KVM capability Signed-off-by: Aravinda Prasad <aravinda@linux.vnet.ibm.com> --- arch/powerpc/include/asm/kvm_host.h | 1 + arch/powerpc/kernel/asm-offsets.c | 1 + arch/powerpc/kvm/powerpc.c | 7 +++++++ include/uapi/linux/kvm.h | 1 + 4 files changed, 10 insertions(+) -- To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html