When user space do memory recovery, it will check whether KVM and
guest support the error recovery, only when both of them support,
user space will do the error recovery. This patch exports this
capability of KVM to user space.
Cc: Peter Maydell <[email protected]>
Signed-off-by: Dongjiu Geng <[email protected]>
---
v1->v2:
1. check whether host support memory failure instead of RAS capability
https://patchwork.kernel.org/patch/10730827/
v1:
1. User space needs to check this capability of host is suggested by Peter[1],
this patch as RFC tag because user space patches are still under review,
so this kernel patch is firstly sent out for review.
[1]: https://patchwork.codeaurora.org/patch/652261/
---
Documentation/virtual/kvm/api.txt | 9 +++++++++
arch/arm64/kvm/reset.c | 3 +++
include/uapi/linux/kvm.h | 1 +
3 files changed, 13 insertions(+)
diff --git a/Documentation/virtual/kvm/api.txt b/Documentation/virtual/kvm/api.txt
index cd209f7..822a57b 100644
--- a/Documentation/virtual/kvm/api.txt
+++ b/Documentation/virtual/kvm/api.txt
@@ -4895,3 +4895,12 @@ Architectures: x86
This capability indicates that KVM supports paravirtualized Hyper-V IPI send
hypercalls:
HvCallSendSyntheticClusterIpi, HvCallSendSyntheticClusterIpiEx.
+
+8.21 KVM_CAP_ARM_MEMORY_ERROR_RECOVERY
+
+Architectures: arm, arm64
+
+This capability indicates that guest memory error can be detected by the host which
+supports the error recovery. When user space do recovery, such as QEMU, it will
+check whether host and guest all support memory error recovery, only when both of them
+support, user space will do the error recovery.
diff --git a/arch/arm64/kvm/reset.c b/arch/arm64/kvm/reset.c
index b72a3dd..b6e3986 100644
--- a/arch/arm64/kvm/reset.c
+++ b/arch/arm64/kvm/reset.c
@@ -84,6 +84,9 @@ int kvm_arch_vm_ioctl_check_extension(struct kvm *kvm, long ext)
case KVM_CAP_ARM_INJECT_SERROR_ESR:
r = cpus_have_const_cap(ARM64_HAS_RAS_EXTN);
break;
+ case KVM_CAP_ARM_MEMORY_ERROR_RECOVERY:
+ r= IS_ENABLED(CONFIG_MEMORY_FAILURE);
+ break;
case KVM_CAP_SET_GUEST_DEBUG:
case KVM_CAP_VCPU_ATTRIBUTES:
r = 1;
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 2b7a652..3b19580 100644
--- a/include/uapi/linux/kvm.h
+++ b/include/uapi/linux/kvm.h
@@ -975,6 +975,7 @@ struct kvm_ppc_resize_hpt {
#define KVM_CAP_HYPERV_ENLIGHTENED_VMCS 163
#define KVM_CAP_EXCEPTION_PAYLOAD 164
#define KVM_CAP_ARM_VM_IPA_SIZE 165
+#define KVM_CAP_ARM_MEMORY_ERROR_RECOVERY 166
#ifdef KVM_CAP_IRQ_ROUTING
--
2.7.4
On Mon, 13 May 2019 at 07:32, Dongjiu Geng <[email protected]> wrote:
>
> When user space do memory recovery, it will check whether KVM and
> guest support the error recovery, only when both of them support,
> user space will do the error recovery. This patch exports this
> capability of KVM to user space.
>
> Cc: Peter Maydell <[email protected]>
> Signed-off-by: Dongjiu Geng <[email protected]>
> ---
> v1->v2:
> 1. check whether host support memory failure instead of RAS capability
> https://patchwork.kernel.org/patch/10730827/
>
> v1:
> 1. User space needs to check this capability of host is suggested by Peter[1],
> this patch as RFC tag because user space patches are still under review,
> so this kernel patch is firstly sent out for review.
>
> [1]: https://patchwork.codeaurora.org/patch/652261/
> ---
I thought the conclusion of the thread on the v1 patch was that
userspace doesn't need to specifically ask the host kernel if
it has support for this -- if it does not, then the host kernel
will just never deliver userspace any SIGBUS with MCEERR code,
which is fine. Or am I still confused?
thanks
-- PMM
On 2019/5/13 17:44, Peter Maydell wrote:
> On Mon, 13 May 2019 at 07:32, Dongjiu Geng <[email protected]> wrote:
>>
>> When user space do memory recovery, it will check whether KVM and
>> guest support the error recovery, only when both of them support,
>> user space will do the error recovery. This patch exports this
>> capability of KVM to user space.
>>
>> Cc: Peter Maydell <[email protected]>
>> Signed-off-by: Dongjiu Geng <[email protected]>
>> ---
>> v1->v2:
>> 1. check whether host support memory failure instead of RAS capability
>> https://patchwork.kernel.org/patch/10730827/
>>
>> v1:
>> 1. User space needs to check this capability of host is suggested by Peter[1],
>> this patch as RFC tag because user space patches are still under review,
>> so this kernel patch is firstly sent out for review.
>>
>> [1]: https://patchwork.codeaurora.org/patch/652261/
>> ---
>
> I thought the conclusion of the thread on the v1 patch was that
> userspace doesn't need to specifically ask the host kernel if
> it has support for this -- if it does not, then the host kernel
> will just never deliver userspace any SIGBUS with MCEERR code,
> which is fine. Or am I still confused?
thanks Peter's quick reply.
yes, I think so, if it does not support, then the host kernel
will just never deliver userspace any SIGBUS with MCEERR code.
so maybe we do not need this patch.
>
> thanks
> -- PMM
> .
>