| Message ID | 20161124172104.GA17619@potion (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
----- Original Message ----- > From: "Radim Krčmář" <rkrcmar@redhat.com> > To: "Paolo Bonzini" <pbonzini@redhat.com> > Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org > Sent: Thursday, November 24, 2016 6:21:04 PM > Subject: Re: [PATCH 2/6] KVM: x86: decouple irqchip_in_kernel() and pic_irqchip() > > 2016-11-24 11:55-0500, Paolo Bonzini: > >> irqchip_in_kernel() tried to save a bit by reusing pic_irqchip(), but it > >> just complicated the code. > >> Add kvm->arch.irqchip_kvm that matches kvm->arch.irqchip_split. > >> (Turning them into an exclusive type would be nicer.) > > > > Then do it. ;) > > It is hard to name! :) > > I would squash something like this if the names were ok: > > diff --git a/arch/x86/include/asm/kvm_host.h > b/arch/x86/include/asm/kvm_host.h > index 929228ec2839..726235f0e3f3 100644 > --- a/arch/x86/include/asm/kvm_host.h > +++ b/arch/x86/include/asm/kvm_host.h > @@ -706,6 +706,12 @@ struct kvm_hv { > HV_REFERENCE_TSC_PAGE tsc_ref; > }; > > +enum kvm_kernel_irqchip { kvm_kernel_irqchip_mode? > + KVM_IRQCHIP_NONE, > + KVM_IRQCHIP_KVM, /* created in KVM_CREATE_IRQCHIP */ > + KVM_IRQCHIP_SPLIT, /* created in KVM_CAP_SPLIT_IRQCHIP */ Since you pretty much asked to nitpick, KVM_IRQCHIP_KERNEL would match irqchip_in_kernel better. Also, s/in/with/? :) > +}; > + > struct kvm_arch { > unsigned int n_used_mmu_pages; > unsigned int n_requested_mmu_pages; > @@ -778,8 +784,7 @@ struct kvm_arch { > > u64 disabled_quirks; > > - bool irqchip_kvm; > - bool irqchip_split; > + enum kvm_kernel_irqchip irqchip; irqchip_mode? > u8 nr_reserved_ioapic_pins; > > bool disabled_lapic_found; > diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h > index a80515e38645..f90ca9e0affc 100644 > --- a/arch/x86/kvm/irq.h > +++ b/arch/x86/kvm/irq.h > @@ -93,12 +93,12 @@ static inline int pic_in_kernel(struct kvm *kvm) > > static inline int irqchip_split(struct kvm *kvm) > { > - return kvm->arch.irqchip_split; > + return kvm->arch.irqchip == KVM_IRQCHIP_SPLIT; > } > > static inline int irqchip_kvm(struct kvm *kvm) > { > - return kvm->arch.irqchip_kvm; > + return kvm->arch.irqchip == KVM_IRQCHIP_KVM; > } > > static inline int irqchip_in_kernel(struct kvm *kvm) > diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c > index 4c364a13b17c..99c63b73dd35 100644 > --- a/arch/x86/kvm/x86.c > +++ b/arch/x86/kvm/x86.c > @@ -3834,7 +3834,7 @@ static int kvm_vm_ioctl_enable_cap(struct kvm *kvm, > goto split_irqchip_unlock; > /* Pairs with irqchip_in_kernel. */ > smp_wmb(); > - kvm->arch.irqchip_split = true; > + kvm->arch.irqchip = KVM_IRQCHIP_SPLIT; > kvm->arch.nr_reserved_ioapic_pins = cap->args[0]; > r = 0; > split_irqchip_unlock: > @@ -3931,7 +3931,7 @@ long kvm_arch_vm_ioctl(struct file *filp, > } > /* Write kvm->irq_routing before enabling irqchip_in_kernel. */ > smp_wmb(); > - kvm->arch.irqchip_kvm = true; > + kvm->arch.irqchip = KVM_IRQCHIP_KVM; > create_irqchip_unlock: > mutex_unlock(&kvm->lock); > break; > -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
2016-11-25 03:59-0500, Paolo Bonzini: > ----- Original Message ----- >> From: "Radim Krčmář" <rkrcmar@redhat.com> >> To: "Paolo Bonzini" <pbonzini@redhat.com> >> Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org >> Sent: Thursday, November 24, 2016 6:21:04 PM >> Subject: Re: [PATCH 2/6] KVM: x86: decouple irqchip_in_kernel() and pic_irqchip() >> >> 2016-11-24 11:55-0500, Paolo Bonzini: >> >> irqchip_in_kernel() tried to save a bit by reusing pic_irqchip(), but it >> >> just complicated the code. >> >> Add kvm->arch.irqchip_kvm that matches kvm->arch.irqchip_split. >> >> (Turning them into an exclusive type would be nicer.) >> > >> > Then do it. ;) >> >> It is hard to name! :) >> >> I would squash something like this if the names were ok: >> >> diff --git a/arch/x86/include/asm/kvm_host.h >> b/arch/x86/include/asm/kvm_host.h >> index 929228ec2839..726235f0e3f3 100644 >> --- a/arch/x86/include/asm/kvm_host.h >> +++ b/arch/x86/include/asm/kvm_host.h >> @@ -706,6 +706,12 @@ struct kvm_hv { >> HV_REFERENCE_TSC_PAGE tsc_ref; >> }; >> >> +enum kvm_kernel_irqchip { > > kvm_kernel_irqchip_mode? If we append the mode, what about just "kvm_irqchip_mode"? irqchip_in_kernel() tells which one is in the kernel. >> + KVM_IRQCHIP_NONE, >> + KVM_IRQCHIP_KVM, /* created in KVM_CREATE_IRQCHIP */ >> + KVM_IRQCHIP_SPLIT, /* created in KVM_CAP_SPLIT_IRQCHIP */ > > Since you pretty much asked to nitpick, I am always interested in nitpicks! > KVM_IRQCHIP_KERNEL would > match irqchip_in_kernel better. Matching the enum name prefix would be nice, so I'd rename it to enum kvm_irqchip_kernel_mode then. I'd keep them this way if we go with enum kvm_irqchip_mode. > Also, s/in/with/? :) Ok. >> +}; >> + >> struct kvm_arch { >> unsigned int n_used_mmu_pages; >> unsigned int n_requested_mmu_pages; >> @@ -778,8 +784,7 @@ struct kvm_arch { >> >> u64 disabled_quirks; >> >> - bool irqchip_kvm; >> - bool irqchip_split; >> + enum kvm_kernel_irqchip irqchip; > > irqchip_mode? Yes, thanks. -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 25/11/2016 18:11, Radim Krčmář wrote: > 2016-11-25 03:59-0500, Paolo Bonzini: >> ----- Original Message ----- >>> From: "Radim Krčmář" <rkrcmar@redhat.com> >>> To: "Paolo Bonzini" <pbonzini@redhat.com> >>> Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org >>> Sent: Thursday, November 24, 2016 6:21:04 PM >>> Subject: Re: [PATCH 2/6] KVM: x86: decouple irqchip_in_kernel() and pic_irqchip() >>> >>> 2016-11-24 11:55-0500, Paolo Bonzini: >>>>> irqchip_in_kernel() tried to save a bit by reusing pic_irqchip(), but it >>>>> just complicated the code. >>>>> Add kvm->arch.irqchip_kvm that matches kvm->arch.irqchip_split. >>>>> (Turning them into an exclusive type would be nicer.) >>>> >>>> Then do it. ;) >>> >>> It is hard to name! :) >>> >>> I would squash something like this if the names were ok: >>> >>> diff --git a/arch/x86/include/asm/kvm_host.h >>> b/arch/x86/include/asm/kvm_host.h >>> index 929228ec2839..726235f0e3f3 100644 >>> --- a/arch/x86/include/asm/kvm_host.h >>> +++ b/arch/x86/include/asm/kvm_host.h >>> @@ -706,6 +706,12 @@ struct kvm_hv { >>> HV_REFERENCE_TSC_PAGE tsc_ref; >>> }; >>> >>> +enum kvm_kernel_irqchip { >> >> kvm_kernel_irqchip_mode? > > If we append the mode, what about just "kvm_irqchip_mode"? > > irqchip_in_kernel() tells which one is in the kernel. > >>> + KVM_IRQCHIP_NONE, >>> + KVM_IRQCHIP_KVM, /* created in KVM_CREATE_IRQCHIP */ >>> + KVM_IRQCHIP_SPLIT, /* created in KVM_CAP_SPLIT_IRQCHIP */ >> >> Since you pretty much asked to nitpick, > > I am always interested in nitpicks! > >> KVM_IRQCHIP_KERNEL would >> match irqchip_in_kernel better. > > Matching the enum name prefix would be nice, so I'd rename it to > enum kvm_irqchip_kernel_mode then. I'd keep them this way if we go with > enum kvm_irqchip_mode. kvm_irqchip_mode is best I think (NONE/KERNEL/SPLIT). Paolo >> Also, s/in/with/? :) > > Ok. > >>> +}; >>> + >>> struct kvm_arch { >>> unsigned int n_used_mmu_pages; >>> unsigned int n_requested_mmu_pages; >>> @@ -778,8 +784,7 @@ struct kvm_arch { >>> >>> u64 disabled_quirks; >>> >>> - bool irqchip_kvm; >>> - bool irqchip_split; >>> + enum kvm_kernel_irqchip irqchip; >> >> irqchip_mode? > > Yes, thanks. > -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h index 929228ec2839..726235f0e3f3 100644 --- a/arch/x86/include/asm/kvm_host.h +++ b/arch/x86/include/asm/kvm_host.h @@ -706,6 +706,12 @@ struct kvm_hv { HV_REFERENCE_TSC_PAGE tsc_ref; }; +enum kvm_kernel_irqchip { + KVM_IRQCHIP_NONE, + KVM_IRQCHIP_KVM, /* created in KVM_CREATE_IRQCHIP */ + KVM_IRQCHIP_SPLIT, /* created in KVM_CAP_SPLIT_IRQCHIP */ +}; + struct kvm_arch { unsigned int n_used_mmu_pages; unsigned int n_requested_mmu_pages; @@ -778,8 +784,7 @@ struct kvm_arch { u64 disabled_quirks; - bool irqchip_kvm; - bool irqchip_split; + enum kvm_kernel_irqchip irqchip; u8 nr_reserved_ioapic_pins; bool disabled_lapic_found; diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h index a80515e38645..f90ca9e0affc 100644 --- a/arch/x86/kvm/irq.h +++ b/arch/x86/kvm/irq.h @@ -93,12 +93,12 @@ static inline int pic_in_kernel(struct kvm *kvm) static inline int irqchip_split(struct kvm *kvm) { - return kvm->arch.irqchip_split; + return kvm->arch.irqchip == KVM_IRQCHIP_SPLIT; } static inline int irqchip_kvm(struct kvm *kvm) { - return kvm->arch.irqchip_kvm; + return kvm->arch.irqchip == KVM_IRQCHIP_KVM; } static inline int irqchip_in_kernel(struct kvm *kvm) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 4c364a13b17c..99c63b73dd35 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -3834,7 +3834,7 @@ static int kvm_vm_ioctl_enable_cap(struct kvm *kvm, goto split_irqchip_unlock; /* Pairs with irqchip_in_kernel. */ smp_wmb(); - kvm->arch.irqchip_split = true; + kvm->arch.irqchip = KVM_IRQCHIP_SPLIT; kvm->arch.nr_reserved_ioapic_pins = cap->args[0]; r = 0; split_irqchip_unlock: @@ -3931,7 +3931,7 @@ long kvm_arch_vm_ioctl(struct file *filp, } /* Write kvm->irq_routing before enabling irqchip_in_kernel. */ smp_wmb(); - kvm->arch.irqchip_kvm = true; + kvm->arch.irqchip = KVM_IRQCHIP_KVM; create_irqchip_unlock: mutex_unlock(&kvm->lock); break;