On 04/01/2017 21:41, Cao, Lei wrote:
> Introduce new memory tracking ioctls to support performant
> checkpointing solutions.
>
> See patch 4 for details on the new ioctls.
>
> Signed-off-by: Lei Cao <lei.cao@xxxxxxxxxxx>
> ---
> arch/x86/include/asm/kvm_host.h | 2 ++
> arch/x86/kvm/x86.c | 3 +++
> include/uapi/linux/kvm.h | 8 +++++++
> virt/kvm/kvm_main.c | 47 ++++++++++++++++++++++++++++++++++++++++-
> 4 files changed, 59 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index 2e25038..6dfb14a 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -45,6 +45,8 @@
>
> #define KVM_PIO_PAGE_OFFSET 1
> #define KVM_COALESCED_MMIO_PAGE_OFFSET 2
> +#define KVM_DIRTY_LOG_PAGE_OFFSET 3
Because this part of the mmap area is variable-sized, let's move it up
(e.g. to 1024).
> +
> #define KVM_HALT_POLL_NS_DEFAULT 400000
>
> #define KVM_IRQCHIP_NUM_PINS KVM_IOAPIC_NUM_PINS
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 0b387d6..5707129 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -2688,6 +2688,9 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
> case KVM_CAP_X2APIC_API:
> r = KVM_X2APIC_API_VALID_FLAGS;
> break;
> + case KVM_CAP_DIRTY_LOG_LIST:
> + r = KVM_DIRTY_LOG_PAGE_OFFSET;
> + break;
This can be in virt/kvm/kvm_main.c kvm_vm_ioctl_check_extension_generic
(predicated by #ifdef KVM_DIRTY_LOG_PAGE_OFFSET).
> default:
> r = 0;
> break;
> diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
> index cac48ed..05332de 100644
> --- a/include/uapi/linux/kvm.h
> +++ b/include/uapi/linux/kvm.h
> @@ -871,6 +871,7 @@ struct kvm_ppc_smmu_info {
> #define KVM_CAP_S390_USER_INSTR0 130
> #define KVM_CAP_MSI_DEVID 131
> #define KVM_CAP_PPC_HTM 132
> +#define KVM_CAP_DIRTY_LOG_LIST 133
>
> #ifdef KVM_CAP_IRQ_ROUTING
>
> @@ -1282,6 +1283,13 @@ struct kvm_s390_ucas_mapping {
> /* Available with KVM_CAP_X86_SMM */
> #define KVM_SMI _IO(KVMIO, 0xb7)
>
> +/* Set the per-vcpu dirty log size */
> +#define KVM_SET_DIRTY_LOG_SIZE _IOW(KVMIO, 0xc0, __u32)
> +/* Rearm traps for all dirty pages */
> +#define KVM_RESET_DIRTY_PAGES _IO(KVMIO, 0xc1)
> +/* Get the current dirty count */
> +#define KVM_GET_DIRTY_COUNT _IOR(KVMIO, 0xc2, __u32)
> +
> #define KVM_DEV_ASSIGN_ENABLE_IOMMU (1 << 0)
> #define KVM_DEV_ASSIGN_PCI_2_3 (1 << 1)
> #define KVM_DEV_ASSIGN_MASK_INTX (1 << 2)
> diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
> index 823544c..8e6f2b1 100644
> --- a/virt/kvm/kvm_main.c
> +++ b/virt/kvm/kvm_main.c
> @@ -2945,6 +2945,23 @@ static long kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
> return kvm_vm_ioctl_check_extension(kvm, arg);
> }
>
> +#ifdef KVM_DIRTY_LOG_PAGE_OFFSET
> +static int kvm_mt_set_dirty_log_size(struct kvm *kvm, u32 size)
> +{
> + return -EINVAL;
> +}
Please rename to kvm_vm_ioctl_enable_dirty_log_list.
> +static int kvm_mt_reset_all_gfns(struct kvm *kvm)
> +{
> + return -EINVAL;
> +}
Please rename to kvm_vm_ioctl_reset_dirty_pages.
> +static int kvm_mt_get_dirty_count(struct kvm *kvm, u32 *count)
> +{
> + return -EINVAL;
> +}
This should not be necessary, as discussed in the cover letter.
> +#endif /* KVM_DIRTY_LOG_PAGE_OFFSET*/
> +
> static long kvm_vm_ioctl(struct file *filp,
> unsigned int ioctl, unsigned long arg)
> {
> @@ -3101,9 +3118,37 @@ static long kvm_vm_ioctl(struct file *filp,
> r = 0;
> break;
> }
> - case KVM_CHECK_EXTENSION:
> + case KVM_CHECK_EXTENSION: {
> r = kvm_vm_ioctl_check_extension_generic(kvm, arg);
> break;
> + }
> +#ifdef KVM_DIRTY_LOG_PAGE_OFFSET
> + case KVM_SET_DIRTY_LOG_SIZE: {
> + u32 size;
> +
> + r = -EFAULT;
> + if (copy_from_user(&size, argp, sizeof(u32)))
> + goto out;
> + r = kvm_mt_set_dirty_log_size(kvm, size);
> + break;
> + }
Please use KVM_ENABLE_CAP for this functionality. I'm attaching a patch
to introduce a new function kvm_vcpu_ioctl_enable_cap_generic, you can
extend it with support for enabling KVM_CAP_DIRTY_LOG_LIST.
Thanks,
Paolo
> + case KVM_RESET_DIRTY_PAGES: {
> + r = kvm_mt_reset_all_gfns(kvm);
> + break;
> + }
> + case KVM_GET_DIRTY_COUNT: {
> + u32 count;
> +
> + r = kvm_mt_get_dirty_count(kvm, &count);
> + if (r)
> + goto out;
> + r = -EFAULT;
> + if (copy_to_user(argp, &count, sizeof(count)))
> + goto out;
> + r = 0;
> + break;
> + }
> +#endif /* KVM_DIRTY_LOG_PAGE_OFFSET */
> default:
> r = kvm_arch_vm_ioctl(filp, ioctl, arg);
> }
>
From: Paolo Bonzini <pbonzini@xxxxxxxxxx>
Subject: [PATCH] kvm: move KVM_CAP_ENABLE_CAP_VM handling to virt/kvm/
We will shortly add a capability that can be enabled in a generic
way. Add a new function kvm_vcpu_ioctl_enable_cap_generic to hold
the KVM_ENABLE_CAP support for that kind of capability.
Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
diff --git a/Documentation/virtual/kvm/api.txt b/Documentation/virtual/kvm/api.txt
index 03145b7cafaa..3a2721d2470e 100644
--- a/Documentation/virtual/kvm/api.txt
+++ b/Documentation/virtual/kvm/api.txt
@@ -1006,10 +1006,15 @@ documentation when it pops into existence).
4.37 KVM_ENABLE_CAP
-Capability: KVM_CAP_ENABLE_CAP, KVM_CAP_ENABLE_CAP_VM
-Architectures: x86 (only KVM_CAP_ENABLE_CAP_VM),
- mips (only KVM_CAP_ENABLE_CAP), ppc, s390
-Type: vcpu ioctl, vm ioctl (with KVM_CAP_ENABLE_CAP_VM)
+Capability: KVM_CAP_ENABLE_CAP
+Architectures: mips, ppc, s390
+Type: vcpu ioctl
+Parameters: struct kvm_enable_cap (in)
+Returns: 0 on success; -1 on error
+
+Capability: KVM_CAP_ENABLE_CAP_VM
+Architectures: all
+Type: vcpu ioctl
Parameters: struct kvm_enable_cap (in)
Returns: 0 on success; -1 on error
diff --git a/arch/powerpc/kvm/powerpc.c b/arch/powerpc/kvm/powerpc.c
index efd1183a6b16..10dd88a4eee0 100644
--- a/arch/powerpc/kvm/powerpc.c
+++ b/arch/powerpc/kvm/powerpc.c
@@ -507,7 +507,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
case KVM_CAP_PPC_UNSET_IRQ:
case KVM_CAP_PPC_IRQ_LEVEL:
case KVM_CAP_ENABLE_CAP:
- case KVM_CAP_ENABLE_CAP_VM:
case KVM_CAP_ONE_REG:
case KVM_CAP_IOEVENTFD:
case KVM_CAP_DEVICE_CTRL:
@@ -1358,8 +1357,8 @@ int kvm_vm_ioctl_irq_line(struct kvm *kvm, struct kvm_irq_level *irq_event,
}
-static int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
- struct kvm_enable_cap *cap)
+int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
+ struct kvm_enable_cap *cap)
{
int r;
@@ -1412,15 +1411,6 @@ long kvm_arch_vm_ioctl(struct file *filp,
break;
}
- case KVM_ENABLE_CAP:
- {
- struct kvm_enable_cap cap;
- r = -EFAULT;
- if (copy_from_user(&cap, argp, sizeof(cap)))
- goto out;
- r = kvm_vm_ioctl_enable_cap(kvm, &cap);
- break;
- }
#ifdef CONFIG_PPC_BOOK3S_64
case KVM_CREATE_SPAPR_TCE_64: {
struct kvm_create_spapr_tce_64 create_tce_64;
diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index bec71e902be3..cc231075b783 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -366,7 +366,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
case KVM_CAP_S390_CSS_SUPPORT:
case KVM_CAP_IOEVENTFD:
case KVM_CAP_DEVICE_CTRL:
- case KVM_CAP_ENABLE_CAP_VM:
case KVM_CAP_S390_IRQCHIP:
case KVM_CAP_VM_ATTRIBUTES:
case KVM_CAP_MP_STATE:
@@ -480,7 +479,7 @@ static void icpt_operexc_on_all_vcpus(struct kvm *kvm)
}
}
-static int kvm_vm_ioctl_enable_cap(struct kvm *kvm, struct kvm_enable_cap *cap)
+int kvm_vm_ioctl_enable_cap(struct kvm *kvm, struct kvm_enable_cap *cap)
{
int r;
@@ -1232,14 +1231,6 @@ long kvm_arch_vm_ioctl(struct file *filp,
r = kvm_s390_inject_vm(kvm, &s390int);
break;
}
- case KVM_ENABLE_CAP: {
- struct kvm_enable_cap cap;
- r = -EFAULT;
- if (copy_from_user(&cap, argp, sizeof(cap)))
- break;
- r = kvm_vm_ioctl_enable_cap(kvm, &cap);
- break;
- }
case KVM_CREATE_IRQCHIP: {
struct kvm_irq_routing_entry routing;
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 2de017e98b13..0aea15aee272 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2628,7 +2628,6 @@ int kvm_vm_ioctl_check_extension(struct kvm *kvm, long ext)
case KVM_CAP_HYPERV_TIME:
case KVM_CAP_IOAPIC_POLARITY_IGNORED:
case KVM_CAP_TSC_DEADLINE_TIMER:
- case KVM_CAP_ENABLE_CAP_VM:
case KVM_CAP_DISABLE_QUIRKS:
case KVM_CAP_SET_BOOT_CPU_ID:
case KVM_CAP_SPLIT_IRQCHIP:
@@ -3829,8 +3828,8 @@ int kvm_vm_ioctl_irq_line(struct kvm *kvm, struct kvm_irq_level *irq_event,
return 0;
}
-static int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
- struct kvm_enable_cap *cap)
+int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
+ struct kvm_enable_cap *cap)
{
int r;
@@ -4138,15 +4137,6 @@ long kvm_arch_vm_ioctl(struct file *filp,
r = 0;
break;
}
- case KVM_ENABLE_CAP: {
- struct kvm_enable_cap cap;
-
- r = -EFAULT;
- if (copy_from_user(&cap, argp, sizeof(cap)))
- goto out;
- r = kvm_vm_ioctl_enable_cap(kvm, &cap);
- break;
- }
default:
r = kvm_vm_ioctl_assigned_device(kvm, ioctl, arg);
}
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 1c5190dab2c1..33d9974f9125 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -718,6 +718,8 @@ int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
int kvm_vm_ioctl_irq_line(struct kvm *kvm, struct kvm_irq_level *irq_level,
bool line_status);
+int kvm_vm_ioctl_enable_cap(struct kvm *kvm,
+ struct kvm_enable_cap *cap);
long kvm_arch_vm_ioctl(struct file *filp,
unsigned int ioctl, unsigned long arg);
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 823544c166be..b051f1cfc5da 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -2928,6 +2928,7 @@ static long kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
#endif
case KVM_CAP_IOEVENTFD_ANY_LENGTH:
case KVM_CAP_CHECK_EXTENSION_VM:
+ case KVM_CAP_ENABLE_CAP_VM:
return 1;
#ifdef CONFIG_HAVE_KVM_IRQ_ROUTING
case KVM_CAP_IRQ_ROUTING:
@@ -2945,6 +2946,18 @@ static long kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg)
return kvm_vm_ioctl_check_extension(kvm, arg);
}
+int __attribute__((weak)) kvm_vm_ioctl_enable_cap(struct kvm *kvm,
+ struct kvm_enable_cap *cap)
+{
+ return -EINVAL;
+}
+
+static int kvm_vm_ioctl_enable_cap_generic(struct kvm *kvm,
+ struct kvm_enable_cap *cap)
+{
+ return kvm_vm_ioctl_enable_cap(kvm, cap);
+}
+
static long kvm_vm_ioctl(struct file *filp,
unsigned int ioctl, unsigned long arg)
{
@@ -2958,6 +2971,15 @@ static long kvm_vm_ioctl(struct file *filp,
case KVM_CREATE_VCPU:
r = kvm_vm_ioctl_create_vcpu(kvm, arg);
break;
+ case KVM_ENABLE_CAP: {
+ struct kvm_enable_cap cap;
+
+ r = -EFAULT;
+ if (copy_from_user(&cap, argp, sizeof(cap)))
+ goto out;
+ r = kvm_vm_ioctl_enable_cap_generic(kvm, &cap);
+ break;
+ }
case KVM_SET_USER_MEMORY_REGION: {
struct kvm_userspace_memory_region kvm_userspace_mem;
