When we migrate a kvm guest that uses pvclock between two hosts, we may
suffer a large skew. This is because there can be significant differences
between the monotonic clock of the hosts involved. When a new host with
a much larger monotonic time starts running the guest, the view of time
will be significantly impacted.
Situation is much worse when we do the opposite, and migrate to a host with
a smaller monotonic clock.
This proposed ioctl will allow userspace to inform us what is the monotonic
clock value in the source host, so we can keep the time skew short, and
more importantly, never goes backwards. Userspace may also need to trigger
the current data, since from the first migration onwards, it won't be
reflected by a simple call to clock_gettime() anymore.
[ v2: uses a struct with a padding ]
[ v3: provide an ioctl to get clock data too ]
[ v4: used fixed-width signed type for delta ]
[ v5: nitpick fixes, documentation update, and capability bit ]
Signed-off-by: Glauber Costa <[email protected]>
---
Documentation/kvm/api.txt | 34 ++++++++++++++++++++++++++++++++++
arch/x86/include/asm/kvm_host.h | 1 +
arch/x86/kvm/x86.c | 36 +++++++++++++++++++++++++++++++++++-
include/linux/kvm.h | 8 ++++++++
4 files changed, 78 insertions(+), 1 deletions(-)
diff --git a/Documentation/kvm/api.txt b/Documentation/kvm/api.txt
index 5a4bc8c..2353c4b 100644
--- a/Documentation/kvm/api.txt
+++ b/Documentation/kvm/api.txt
@@ -593,6 +593,40 @@ struct kvm_irqchip {
} chip;
};
+4.27 KVM_GET_CLOCK
+
+Capability: KVM_CAP_ADJUST_CLOCK
+Architectures: x86
+Type: vm ioctl
+Parameters: struct kvm_clock_data (out)
+Returns: 0 on success, -1 on error
+
+Gets the current timestamp of kvmclock as seen by the current guest. In
+conjunction with KVM_SET_CLOCK, it is used to ensure monotonicity on scenarios
+such as migration.
+
+struct kvm_clock_data {
+ __u64 clock; /* kvmclock current value */
+ __u64 pad[2];
+};
+
+4.28 KVM_SET_CLOCK
+
+Capability: KVM_CAP_ADJUST_CLOCK
+Architectures: x86
+Type: vm ioctl
+Parameters: struct kvm_clock_data (in)
+Returns: 0 on success, -1 on error
+
+Sets the current timestamp of kvmclock to the valued specific in its parameter.
+In conjunction with KVM_GET_CLOCK, it is used to ensure monotonicity on scenarios
+such as migration.
+
+struct kvm_clock_data {
+ __u64 clock; /* kvmclock current value */
+ __u64 pad[2];
+};
+
5. The kvm_run structure
Application code obtains a pointer to the kvm_run structure by
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index d838922..d759a1f 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -412,6 +412,7 @@ struct kvm_arch{
unsigned long irq_sources_bitmap;
unsigned long irq_states[KVM_IOAPIC_NUM_PINS];
u64 vm_init_tsc;
+ s64 kvmclock_offset;
};
struct kvm_vm_stat {
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 9b96953..a6fc13a 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -677,7 +677,8 @@ static void kvm_write_guest_time(struct kvm_vcpu *v)
/* With all the info we got, fill in the values */
vcpu->hv_clock.system_time = ts.tv_nsec +
- (NSEC_PER_SEC * (u64)ts.tv_sec);
+ (NSEC_PER_SEC * (u64)ts.tv_sec) + v->kvm->arch.kvmclock_offset;
+
/*
* The interface expects us to write an even number signaling that the
* update is finished. Since the guest won't see the intermediate
@@ -1224,6 +1225,7 @@ int kvm_dev_ioctl_check_extension(long ext)
case KVM_CAP_PIT2:
case KVM_CAP_PIT_STATE2:
case KVM_CAP_SET_IDENTITY_MAP_ADDR:
+ case KVM_CAP_ADJUST_CLOCK:
r = 1;
break;
case KVM_CAP_COALESCED_MMIO:
@@ -2421,6 +2423,38 @@ long kvm_arch_vm_ioctl(struct file *filp,
r = 0;
break;
}
+ case KVM_SET_CLOCK: {
+ struct timespec now;
+ struct kvm_clock_data user_ns;
+ u64 now_ns;
+ s64 delta;
+
+ r = -EFAULT;
+ if (copy_from_user(&user_ns, argp, sizeof(user_ns)))
+ goto out;
+
+ r = 0;
+ ktime_get_ts(&now);
+ now_ns = timespec_to_ns(&now);
+ delta = user_ns.clock - now_ns;
+ kvm->arch.kvmclock_offset = delta;
+ break;
+ }
+ case KVM_GET_CLOCK: {
+ struct timespec now;
+ struct kvm_clock_data user_ns;
+ u64 now_ns;
+
+ ktime_get_ts(&now);
+ now_ns = timespec_to_ns(&now);
+ user_ns.clock = kvm->arch.kvmclock_offset + now_ns;
+
+ if (copy_to_user(argp, &user_ns, sizeof(user_ns)))
+ r = -EFAULT;
+
+ break;
+ }
+
default:
;
}
diff --git a/include/linux/kvm.h b/include/linux/kvm.h
index f8f8900..7fb04a9 100644
--- a/include/linux/kvm.h
+++ b/include/linux/kvm.h
@@ -436,6 +436,7 @@ struct kvm_ioeventfd {
#endif
#define KVM_CAP_IOEVENTFD 36
#define KVM_CAP_SET_IDENTITY_MAP_ADDR 37
+#define KVM_CAP_ADJUST_CLOCK 38
#ifdef KVM_CAP_IRQ_ROUTING
@@ -497,6 +498,11 @@ struct kvm_irqfd {
__u8 pad[20];
};
+struct kvm_clock_data {
+ __u64 clock;
+ __u64 pad[2];
+};
+
/*
* ioctls for VM fds
*/
@@ -546,6 +552,8 @@ struct kvm_irqfd {
#define KVM_CREATE_PIT2 _IOW(KVMIO, 0x77, struct kvm_pit_config)
#define KVM_SET_BOOT_CPU_ID _IO(KVMIO, 0x78)
#define KVM_IOEVENTFD _IOW(KVMIO, 0x79, struct kvm_ioeventfd)
+#define KVM_SET_CLOCK _IOW(KVMIO, 0x7a, struct kvm_clock_data)
+#define KVM_GET_CLOCK _IOR(KVMIO, 0x7b, struct kvm_clock_data)
/*
* ioctls for vcpu fds
--
1.6.2.2
On Fri, Oct 16, 2009 at 03:28:36PM -0400, Glauber Costa wrote:
> When we migrate a kvm guest that uses pvclock between two hosts, we may
> suffer a large skew. This is because there can be significant differences
> between the monotonic clock of the hosts involved. When a new host with
> a much larger monotonic time starts running the guest, the view of time
> will be significantly impacted.
>
> Situation is much worse when we do the opposite, and migrate to a host with
> a smaller monotonic clock.
>
> This proposed ioctl will allow userspace to inform us what is the monotonic
> clock value in the source host, so we can keep the time skew short, and
> more importantly, never goes backwards. Userspace may also need to trigger
> the current data, since from the first migration onwards, it won't be
> reflected by a simple call to clock_gettime() anymore.
>
> [ v2: uses a struct with a padding ]
> [ v3: provide an ioctl to get clock data too ]
> [ v4: used fixed-width signed type for delta ]
> [ v5: nitpick fixes, documentation update, and capability bit ]
>
> Signed-off-by: Glauber Costa <[email protected]>
Applied, thanks (the ioctl number changed though).
On 10/17/2009 04:28 AM, Glauber Costa wrote:
> When we migrate a kvm guest that uses pvclock between two hosts, we may
> suffer a large skew. This is because there can be significant differences
> between the monotonic clock of the hosts involved. When a new host with
> a much larger monotonic time starts running the guest, the view of time
> will be significantly impacted.
>
> Situation is much worse when we do the opposite, and migrate to a host with
> a smaller monotonic clock.
>
> This proposed ioctl will allow userspace to inform us what is the monotonic
> clock value in the source host, so we can keep the time skew short, and
> more importantly, never goes backwards. Userspace may also need to trigger
> the current data, since from the first migration onwards, it won't be
> reflected by a simple call to clock_gettime() anymore.
>
> +struct kvm_clock_data {
> + __u64 clock;
> + __u64 pad[2];
> +};
> +
>
The padding is not reusable without a flags word that we can set bits in
for extended functionality. Please add one (and a check that it is
zero, or -EINVAL).
Since it was already applied, please do it as an incremental patch.
--
I have a truly marvellous patch that fixes the bug which this
signature is too narrow to contain.