Message ID | 1599734031-28746-1-git-send-email-chenhc@lemote.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | KVM: MIPS: Change the definition of kvm type | expand |
On 10/09/2020 12.33, Huacai Chen wrote: > MIPS defines two kvm types: > > #define KVM_VM_MIPS_TE 0 > #define KVM_VM_MIPS_VZ 1 > > In Documentation/virt/kvm/api.rst it is said that "You probably want to > use 0 as machine type", which implies that type 0 be the "automatic" or > "default" type. And, in user-space libvirt use the null-machine (with > type 0) to detect the kvm capability, which returns "KVM not supported" > on a VZ platform. > > I try to fix it in QEMU but it is ugly: > https://lists.nongnu.org/archive/html/qemu-devel/2020-08/msg05629.html > > And Thomas Huth suggests me to change the definition of kvm type: > https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03281.html > > So I define like this: > > #define KVM_VM_MIPS_AUTO 0 > #define KVM_VM_MIPS_VZ 1 > #define KVM_VM_MIPS_TE 2 > > Since VZ and TE cannot co-exists, using type 0 on a TE platform will > still return success (so old user-space tools have no problems on new > kernels); the advantage is that using type 0 on a VZ platform will not > return failure. So, the only problem is "new user-space tools use type > 2 on old kernels", but if we treat this as a kernel bug, we can backport > this patch to old stable kernels. > > Cc: stable@vger.kernel.org > Signed-off-by: Huacai Chen <chenhc@lemote.com> > --- > arch/mips/kvm/mips.c | 2 ++ > include/uapi/linux/kvm.h | 5 +++-- > 2 files changed, 5 insertions(+), 2 deletions(-) > > diff --git a/arch/mips/kvm/mips.c b/arch/mips/kvm/mips.c > index d7ba3f9..9efeb67 100644 > --- a/arch/mips/kvm/mips.c > +++ b/arch/mips/kvm/mips.c > @@ -138,6 +138,8 @@ extern void kvm_init_loongson_ipi(struct kvm *kvm); > int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) > { > switch (type) { > + case KVM_VM_MIPS_AUTO: > + break; > #ifdef CONFIG_KVM_MIPS_VZ > case KVM_VM_MIPS_VZ: > #else > diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h > index 29ba8e8..cfc1ae2 100644 > --- a/include/uapi/linux/kvm.h > +++ b/include/uapi/linux/kvm.h > @@ -790,9 +790,10 @@ struct kvm_ppc_resize_hpt { > #define KVM_VM_PPC_HV 1 > #define KVM_VM_PPC_PR 2 > > -/* on MIPS, 0 forces trap & emulate, 1 forces VZ ASE */ > -#define KVM_VM_MIPS_TE 0 > +/* on MIPS, 0 indicates auto, 1 forces VZ ASE, 2 forces trap & emulate */ > +#define KVM_VM_MIPS_AUTO 0 > #define KVM_VM_MIPS_VZ 1 > +#define KVM_VM_MIPS_TE 2 > > #define KVM_S390_SIE_PAGE_OFFSET 1 Thanks, I think that's the right way to go if we don't want mips to behave completely different compared to the other architectures (e.g. powerpc is also using 0 as "automatic" type in arch/powerpc/kvm/powerpc.c). Reviewed-by: Thomas Huth <thuth@redhat.com>
Hi [This is an automated email] This commit has been processed because it contains a -stable tag. The stable tag indicates that it's relevant for the following trees: all The bot has tested the following trees: v5.8.7, v5.4.63, v4.19.143, v4.14.196, v4.9.235, v4.4.235. v5.8.7: Build OK! v5.4.63: Build OK! v4.19.143: Build OK! v4.14.196: Build OK! v4.9.235: Failed to apply! Possible dependencies: 06c158c96ed8 ("KVM: MIPS/MMU: Convert guest physical map to page table") 1534b3964901 ("KVM: MIPS/MMU: Simplify ASID restoration") 1581ff3dbf69 ("KVM: MIPS/MMU: Move preempt/ASID handling to implementation") 91cdee5710d5 ("KVM: MIPS/T&E: Restore host asid on return to host") a2c046e40ff1 ("KVM: MIPS: Add vcpu_run() & vcpu_reenter() callbacks") a31b50d741bd ("KVM: MIPS/MMU: Invalidate GVA PTs on ASID changes") a60b8438bdba ("KVM: MIPS: Convert get/set_regs -> vcpu_load/put") a7ebb2e410f8 ("KVM: MIPS/T&E: active_mm = init_mm in guest context") a8a3c426772e ("KVM: MIPS: Add VZ & TE capabilities") c550d53934d8 ("KVM: MIPS: Remove duplicated ASIDs from vcpu") c92701322711 ("KVM: PPC: Book3S HV: Add userspace interfaces for POWER9 MMU") v4.4.235: Failed to apply! Possible dependencies: 107d44a2c5bf ("KVM: document KVM_REINJECT_CONTROL ioctl") 366baf28ee3f ("KVM: PPC: Use RCU for arch.spapr_tce_tables") 462ee11e58c9 ("KVM: PPC: Replace SPAPR_TCE_SHIFT with IOMMU_PAGE_SHIFT_4K") 58ded4201ff0 ("KVM: PPC: Add support for 64bit TCE windows") 5ee7af18642c ("KVM: PPC: Move reusable bits of H_PUT_TCE handler to helpers") a8a3c426772e ("KVM: MIPS: Add VZ & TE capabilities") c92701322711 ("KVM: PPC: Book3S HV: Add userspace interfaces for POWER9 MMU") d3695aa4f452 ("KVM: PPC: Add support for multiple-TCE hcalls") f8626985c7c2 ("KVM: PPC: Account TCE-containing pages in locked_vm") fcbb2ce67284 ("KVM: PPC: Rework H_PUT_TCE/H_GET_TCE handlers") fe26e52712cc ("KVM: PPC: Add @page_shift to kvmppc_spapr_tce_table") NOTE: The patch will not be queued to stable trees until it is upstream. How should we proceed with this patch?
On Thu, Sep 10, 2020 at 12:34 PM Huacai Chen <chenhc@lemote.com> wrote: > > MIPS defines two kvm types: > > #define KVM_VM_MIPS_TE 0 > #define KVM_VM_MIPS_VZ 1 > > In Documentation/virt/kvm/api.rst it is said that "You probably want to > use 0 as machine type", which implies that type 0 be the "automatic" or > "default" type. And, in user-space libvirt use the null-machine (with > type 0) to detect the kvm capability, which returns "KVM not supported" > on a VZ platform. > > I try to fix it in QEMU but it is ugly: > https://lists.nongnu.org/archive/html/qemu-devel/2020-08/msg05629.html I'm not sure this is helpful information to keep in the commit message. Otherwise: Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org> > > And Thomas Huth suggests me to change the definition of kvm type: > https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03281.html Suggested-by: Thomas Huth <thuth@redhat.com> > > So I define like this: > > #define KVM_VM_MIPS_AUTO 0 > #define KVM_VM_MIPS_VZ 1 > #define KVM_VM_MIPS_TE 2 > > Since VZ and TE cannot co-exists, using type 0 on a TE platform will > still return success (so old user-space tools have no problems on new > kernels); the advantage is that using type 0 on a VZ platform will not > return failure. So, the only problem is "new user-space tools use type > 2 on old kernels", but if we treat this as a kernel bug, we can backport > this patch to old stable kernels. > > Cc: stable@vger.kernel.org > Signed-off-by: Huacai Chen <chenhc@lemote.com> > --- > arch/mips/kvm/mips.c | 2 ++ > include/uapi/linux/kvm.h | 5 +++-- > 2 files changed, 5 insertions(+), 2 deletions(-) > > diff --git a/arch/mips/kvm/mips.c b/arch/mips/kvm/mips.c > index d7ba3f9..9efeb67 100644 > --- a/arch/mips/kvm/mips.c > +++ b/arch/mips/kvm/mips.c > @@ -138,6 +138,8 @@ extern void kvm_init_loongson_ipi(struct kvm *kvm); > int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) > { > switch (type) { > + case KVM_VM_MIPS_AUTO: > + break; > #ifdef CONFIG_KVM_MIPS_VZ > case KVM_VM_MIPS_VZ: > #else > diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h > index 29ba8e8..cfc1ae2 100644 > --- a/include/uapi/linux/kvm.h > +++ b/include/uapi/linux/kvm.h > @@ -790,9 +790,10 @@ struct kvm_ppc_resize_hpt { > #define KVM_VM_PPC_HV 1 > #define KVM_VM_PPC_PR 2 > > -/* on MIPS, 0 forces trap & emulate, 1 forces VZ ASE */ > -#define KVM_VM_MIPS_TE 0 > +/* on MIPS, 0 indicates auto, 1 forces VZ ASE, 2 forces trap & emulate */ > +#define KVM_VM_MIPS_AUTO 0 > #define KVM_VM_MIPS_VZ 1 > +#define KVM_VM_MIPS_TE 2 > > #define KVM_S390_SIE_PAGE_OFFSET 1 > > -- > 2.7.0 >
Hi, Sasha, On Fri, Sep 11, 2020 at 1:18 AM Sasha Levin <sashal@kernel.org> wrote: > > Hi > > [This is an automated email] > > This commit has been processed because it contains a -stable tag. > The stable tag indicates that it's relevant for the following trees: all > > The bot has tested the following trees: v5.8.7, v5.4.63, v4.19.143, v4.14.196, v4.9.235, v4.4.235. > > v5.8.7: Build OK! > v5.4.63: Build OK! > v4.19.143: Build OK! > v4.14.196: Build OK! > v4.9.235: Failed to apply! Possible dependencies: > 06c158c96ed8 ("KVM: MIPS/MMU: Convert guest physical map to page table") > 1534b3964901 ("KVM: MIPS/MMU: Simplify ASID restoration") > 1581ff3dbf69 ("KVM: MIPS/MMU: Move preempt/ASID handling to implementation") > 91cdee5710d5 ("KVM: MIPS/T&E: Restore host asid on return to host") > a2c046e40ff1 ("KVM: MIPS: Add vcpu_run() & vcpu_reenter() callbacks") > a31b50d741bd ("KVM: MIPS/MMU: Invalidate GVA PTs on ASID changes") > a60b8438bdba ("KVM: MIPS: Convert get/set_regs -> vcpu_load/put") > a7ebb2e410f8 ("KVM: MIPS/T&E: active_mm = init_mm in guest context") > a8a3c426772e ("KVM: MIPS: Add VZ & TE capabilities") > c550d53934d8 ("KVM: MIPS: Remove duplicated ASIDs from vcpu") > c92701322711 ("KVM: PPC: Book3S HV: Add userspace interfaces for POWER9 MMU") > > v4.4.235: Failed to apply! Possible dependencies: > 107d44a2c5bf ("KVM: document KVM_REINJECT_CONTROL ioctl") > 366baf28ee3f ("KVM: PPC: Use RCU for arch.spapr_tce_tables") > 462ee11e58c9 ("KVM: PPC: Replace SPAPR_TCE_SHIFT with IOMMU_PAGE_SHIFT_4K") > 58ded4201ff0 ("KVM: PPC: Add support for 64bit TCE windows") > 5ee7af18642c ("KVM: PPC: Move reusable bits of H_PUT_TCE handler to helpers") > a8a3c426772e ("KVM: MIPS: Add VZ & TE capabilities") > c92701322711 ("KVM: PPC: Book3S HV: Add userspace interfaces for POWER9 MMU") > d3695aa4f452 ("KVM: PPC: Add support for multiple-TCE hcalls") > f8626985c7c2 ("KVM: PPC: Account TCE-containing pages in locked_vm") > fcbb2ce67284 ("KVM: PPC: Rework H_PUT_TCE/H_GET_TCE handlers") > fe26e52712cc ("KVM: PPC: Add @page_shift to kvmppc_spapr_tce_table") > > > NOTE: The patch will not be queued to stable trees until it is upstream. > > How should we proceed with this patch? Backport to 4.14+ is enough. Huacai > > -- > Thanks > Sasha
On 10/09/20 12:33, Huacai Chen wrote: > MIPS defines two kvm types: > > #define KVM_VM_MIPS_TE 0 > #define KVM_VM_MIPS_VZ 1 > > In Documentation/virt/kvm/api.rst it is said that "You probably want to > use 0 as machine type", which implies that type 0 be the "automatic" or > "default" type. And, in user-space libvirt use the null-machine (with > type 0) to detect the kvm capability, which returns "KVM not supported" > on a VZ platform. > > I try to fix it in QEMU but it is ugly: > https://lists.nongnu.org/archive/html/qemu-devel/2020-08/msg05629.html > > And Thomas Huth suggests me to change the definition of kvm type: > https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03281.html > > So I define like this: > > #define KVM_VM_MIPS_AUTO 0 > #define KVM_VM_MIPS_VZ 1 > #define KVM_VM_MIPS_TE 2 > > Since VZ and TE cannot co-exists, using type 0 on a TE platform will > still return success (so old user-space tools have no problems on new > kernels); the advantage is that using type 0 on a VZ platform will not > return failure. So, the only problem is "new user-space tools use type > 2 on old kernels", but if we treat this as a kernel bug, we can backport > this patch to old stable kernels. I'm a bit wary to do that. However it's not an issue for QEMU since it generally updates the kernel headers. Paolo > Cc: stable@vger.kernel.org > Signed-off-by: Huacai Chen <chenhc@lemote.com> > --- > arch/mips/kvm/mips.c | 2 ++ > include/uapi/linux/kvm.h | 5 +++-- > 2 files changed, 5 insertions(+), 2 deletions(-) > > diff --git a/arch/mips/kvm/mips.c b/arch/mips/kvm/mips.c > index d7ba3f9..9efeb67 100644 > --- a/arch/mips/kvm/mips.c > +++ b/arch/mips/kvm/mips.c > @@ -138,6 +138,8 @@ extern void kvm_init_loongson_ipi(struct kvm *kvm); > int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) > { > switch (type) { > + case KVM_VM_MIPS_AUTO: > + break; > #ifdef CONFIG_KVM_MIPS_VZ > case KVM_VM_MIPS_VZ: > #else > diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h > index 29ba8e8..cfc1ae2 100644 > --- a/include/uapi/linux/kvm.h > +++ b/include/uapi/linux/kvm.h > @@ -790,9 +790,10 @@ struct kvm_ppc_resize_hpt { > #define KVM_VM_PPC_HV 1 > #define KVM_VM_PPC_PR 2 > > -/* on MIPS, 0 forces trap & emulate, 1 forces VZ ASE */ > -#define KVM_VM_MIPS_TE 0 > +/* on MIPS, 0 indicates auto, 1 forces VZ ASE, 2 forces trap & emulate */ > +#define KVM_VM_MIPS_AUTO 0 > #define KVM_VM_MIPS_VZ 1 > +#define KVM_VM_MIPS_TE 2 > > #define KVM_S390_SIE_PAGE_OFFSET 1 > >
On 11/09/2020 19.22, Paolo Bonzini wrote: > On 10/09/20 12:33, Huacai Chen wrote: >> MIPS defines two kvm types: >> >> #define KVM_VM_MIPS_TE 0 >> #define KVM_VM_MIPS_VZ 1 >> >> In Documentation/virt/kvm/api.rst it is said that "You probably want to >> use 0 as machine type", which implies that type 0 be the "automatic" or >> "default" type. And, in user-space libvirt use the null-machine (with >> type 0) to detect the kvm capability, which returns "KVM not supported" >> on a VZ platform. >> >> I try to fix it in QEMU but it is ugly: >> https://lists.nongnu.org/archive/html/qemu-devel/2020-08/msg05629.html >> >> And Thomas Huth suggests me to change the definition of kvm type: >> https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03281.html >> >> So I define like this: >> >> #define KVM_VM_MIPS_AUTO 0 >> #define KVM_VM_MIPS_VZ 1 >> #define KVM_VM_MIPS_TE 2 >> >> Since VZ and TE cannot co-exists, using type 0 on a TE platform will >> still return success (so old user-space tools have no problems on new >> kernels); the advantage is that using type 0 on a VZ platform will not >> return failure. So, the only problem is "new user-space tools use type >> 2 on old kernels", but if we treat this as a kernel bug, we can backport >> this patch to old stable kernels. > > I'm a bit wary to do that. However it's not an issue for QEMU since it > generally updates the kernel headers. Are there any other userspace programs beside QEMU that use KVM on MIPS? If there aren't any other serious userspace programs, I think we should go ahead with this patch here. Otherwise, what are the other programs that could be affected? Thomas
On 19/09/20 11:18, Thomas Huth wrote: >>> >>> So I define like this: >>> >>> #define KVM_VM_MIPS_AUTO 0 >>> #define KVM_VM_MIPS_VZ 1 >>> #define KVM_VM_MIPS_TE 2 >>> >>> Since VZ and TE cannot co-exists, using type 0 on a TE platform will >>> still return success (so old user-space tools have no problems on new >>> kernels); the advantage is that using type 0 on a VZ platform will not >>> return failure. So, the only problem is "new user-space tools use type >>> 2 on old kernels", but if we treat this as a kernel bug, we can backport >>> this patch to old stable kernels. >> I'm a bit wary to do that. However it's not an issue for QEMU since it >> generally updates the kernel headers. > Are there any other userspace programs beside QEMU that use KVM on MIPS? > If there aren't any other serious userspace programs, I think we should > go ahead with this patch here. Otherwise, what are the other programs > that could be affected? kvmtool (aka lkvm) I guess. I don't know if it is affected but I wouldn't be worried. Paolo
diff --git a/arch/mips/kvm/mips.c b/arch/mips/kvm/mips.c index d7ba3f9..9efeb67 100644 --- a/arch/mips/kvm/mips.c +++ b/arch/mips/kvm/mips.c @@ -138,6 +138,8 @@ extern void kvm_init_loongson_ipi(struct kvm *kvm); int kvm_arch_init_vm(struct kvm *kvm, unsigned long type) { switch (type) { + case KVM_VM_MIPS_AUTO: + break; #ifdef CONFIG_KVM_MIPS_VZ case KVM_VM_MIPS_VZ: #else diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h index 29ba8e8..cfc1ae2 100644 --- a/include/uapi/linux/kvm.h +++ b/include/uapi/linux/kvm.h @@ -790,9 +790,10 @@ struct kvm_ppc_resize_hpt { #define KVM_VM_PPC_HV 1 #define KVM_VM_PPC_PR 2 -/* on MIPS, 0 forces trap & emulate, 1 forces VZ ASE */ -#define KVM_VM_MIPS_TE 0 +/* on MIPS, 0 indicates auto, 1 forces VZ ASE, 2 forces trap & emulate */ +#define KVM_VM_MIPS_AUTO 0 #define KVM_VM_MIPS_VZ 1 +#define KVM_VM_MIPS_TE 2 #define KVM_S390_SIE_PAGE_OFFSET 1
MIPS defines two kvm types: #define KVM_VM_MIPS_TE 0 #define KVM_VM_MIPS_VZ 1 In Documentation/virt/kvm/api.rst it is said that "You probably want to use 0 as machine type", which implies that type 0 be the "automatic" or "default" type. And, in user-space libvirt use the null-machine (with type 0) to detect the kvm capability, which returns "KVM not supported" on a VZ platform. I try to fix it in QEMU but it is ugly: https://lists.nongnu.org/archive/html/qemu-devel/2020-08/msg05629.html And Thomas Huth suggests me to change the definition of kvm type: https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03281.html So I define like this: #define KVM_VM_MIPS_AUTO 0 #define KVM_VM_MIPS_VZ 1 #define KVM_VM_MIPS_TE 2 Since VZ and TE cannot co-exists, using type 0 on a TE platform will still return success (so old user-space tools have no problems on new kernels); the advantage is that using type 0 on a VZ platform will not return failure. So, the only problem is "new user-space tools use type 2 on old kernels", but if we treat this as a kernel bug, we can backport this patch to old stable kernels. Cc: stable@vger.kernel.org Signed-off-by: Huacai Chen <chenhc@lemote.com> --- arch/mips/kvm/mips.c | 2 ++ include/uapi/linux/kvm.h | 5 +++-- 2 files changed, 5 insertions(+), 2 deletions(-)