Re: [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Roman Kagan]

On Fri, Oct 09, 2015 at 04:41:15PM +0200, Paolo Bonzini wrote:
> On 09/10/2015 15:39, Denis V. Lunev wrote:
> > A new vcpu exit is introduced to notify the userspace of the
> > changes in Hyper-V synic configuraion triggered by guest writing to the
> > corresponding MSRs.
> 
> Why is this exit necessary?

The guest writes to synic-related MSRs and that should take "immediate"
effect.

E.g. it may decide to disable or relocate the message page by writing to
SIMP MSR.  The host is then supposed to stop accessing the old message
page before the vCPU proceeds to the next instruction.  Hence the exit,
to allow the userspace to react accordingly before reentering the guest.

Roman.
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Eric Blake]

On 10/12/2015 07:42 AM, Eric Blake wrote:
> On 10/09/2015 07:39 AM, Denis V. Lunev wrote:
>> From: Andrey Smetanin 
>>
>> A new vcpu exit is introduced to notify the userspace of the
>> changes in Hyper-V synic configuraion triggered by guest writing to the
> 
> s/configuraion/configuration/
> Is 'synic' intended?  Is it short for something (if so, spelling it out
> may help)?

Ah, I see it now from patch 1/2: SYNthetic Interrupt Controller.

>> +related with Hyper-V emulation. Currently used to synchronize modified
>> +Hyper-V synic state with userspace.
> 
> Again, is 'synic' intended?  Hmm, I see it throughout the patch, so it
> looks intentional, but I keep trying to read it as a typo for 'sync'.

Don't know if SynIC would make it any easier to read?

-- 
Eric Blake   eblake redhat com+1-919-301-3266
Libvirt virtualization library http://libvirt.org



signature.asc
Description: OpenPGP digital signature
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Eric Blake]

On 10/09/2015 07:39 AM, Denis V. Lunev wrote:
> From: Andrey Smetanin 
> 
> A new vcpu exit is introduced to notify the userspace of the
> changes in Hyper-V synic configuraion triggered by guest writing to the

s/configuraion/configuration/
Is 'synic' intended?  Is it short for something (if so, spelling it out
may help)?


> +++ b/Documentation/virtual/kvm/api.txt
> @@ -3331,6 +3331,12 @@ the userspace IOAPIC should process the EOI and 
> retrigger the interrupt if
>  it is still asserted.  Vector is the LAPIC interrupt vector for which the
>  EOI was received.
>  
> + /* KVM_EXIT_HYPERV */
> +struct kvm_hyperv_exit hyperv;
> +Indicates that the VCPU's exits into userspace to process some tasks

s/VCPU's/VCPU/

> +related with Hyper-V emulation. Currently used to synchronize modified
> +Hyper-V synic state with userspace.

Again, is 'synic' intended?  Hmm, I see it throughout the patch, so it
looks intentional, but I keep trying to read it as a typo for 'sync'.

-- 
Eric Blake   eblake redhat com+1-919-301-3266
Libvirt virtualization library http://libvirt.org



signature.asc
Description: OpenPGP digital signature
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Denis V. Lunev]

On 10/12/2015 04:42 PM, Eric Blake wrote:

On 10/09/2015 07:39 AM, Denis V. Lunev wrote:

From: Andrey Smetanin 

A new vcpu exit is introduced to notify the userspace of the
changes in Hyper-V synic configuraion triggered by guest writing to the

s/configuraion/configuration/
Is 'synic' intended?  Is it short for something (if so, spelling it out
may help)?



+++ b/Documentation/virtual/kvm/api.txt
@@ -3331,6 +3331,12 @@ the userspace IOAPIC should process the EOI and 
retrigger the interrupt if
  it is still asserted.  Vector is the LAPIC interrupt vector for which the
  EOI was received.
  
+		/* KVM_EXIT_HYPERV */

+struct kvm_hyperv_exit hyperv;
+Indicates that the VCPU's exits into userspace to process some tasks

s/VCPU's/VCPU/


+related with Hyper-V emulation. Currently used to synchronize modified
+Hyper-V synic state with userspace.

Again, is 'synic' intended?  Hmm, I see it throughout the patch, so it
looks intentional, but I keep trying to read it as a typo for 'sync'.


this is not a typo :)

this is an abbreviation for synthetic interrupt controller,
pls compare with this

./arch/x86/include/uapi/asm/hyperv.h: * Basic SynIC MSRs 
(HV_X64_MSR_SCONTROL through HV_X64_MSR_EOM


Though there is some sense in the question itself. I think
that it would be better to keep naming it SynIC as
in the original kernel code.

Den
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Roman Kagan]

On Mon, Oct 12, 2015 at 07:42:42AM -0600, Eric Blake wrote:
> On 10/09/2015 07:39 AM, Denis V. Lunev wrote:
> > From: Andrey Smetanin 
> > 
> > A new vcpu exit is introduced to notify the userspace of the
> > changes in Hyper-V synic configuraion triggered by guest writing to the
> Again, is 'synic' intended?  Hmm, I see it throughout the patch, so it
> looks intentional, but I keep trying to read it as a typo for 'sync'.

I tend to mistype it as 'cynic' as better matching what it is ;)

Note taken, we'll address that in the next round, thanks.

Roman.
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Paolo Bonzini]

On 12/10/2015 15:42, Eric Blake wrote:
> > +related with Hyper-V emulation. Currently used to synchronize modified
> > +Hyper-V synic state with userspace.
>
> Again, is 'synic' intended?  Hmm, I see it throughout the patch, so it
> looks intentional, but I keep trying to read it as a typo for 'sync'.

It's synthetic interrupt controller, but indeed it's a bit obscure.
Perhaps we can change it in the docs to "Hyper-V synthetic interrupt state".

Paolo
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

RE: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[KY Srinivasan]

> -Original Message-
> From: Denis V. Lunev [mailto:d...@openvz.org]
> Sent: Monday, October 12, 2015 6:46 AM
> To: Eric Blake <ebl...@redhat.com>
> Cc: Gleb Natapov <g...@kernel.org>; qemu-de...@nongnu.org;
> virtualization@lists.linux-foundation.org; rka...@virtuozzo.com; Paolo
> Bonzini <pbonz...@redhat.com>; Andrey Smetanin
> <asmeta...@virtuozzo.com>; Vitaly Kuznetsov <vkuzn...@redhat.com>; KY
> Srinivasan <k...@microsoft.com>
> Subject: Re: [Qemu-devel] [PATCH 2/2] kvm/x86: Hyper-V kvm exit
> 
> On 10/12/2015 04:42 PM, Eric Blake wrote:
> > On 10/09/2015 07:39 AM, Denis V. Lunev wrote:
> >> From: Andrey Smetanin <asmeta...@virtuozzo.com>
> >>
> >> A new vcpu exit is introduced to notify the userspace of the
> >> changes in Hyper-V synic configuraion triggered by guest writing to the
> > s/configuraion/configuration/
> > Is 'synic' intended?  Is it short for something (if so, spelling it out
> > may help)?
> >
> >
> >> +++ b/Documentation/virtual/kvm/api.txt
> >> @@ -3331,6 +3331,12 @@ the userspace IOAPIC should process the EOI
> and retrigger the interrupt if
> >>   it is still asserted.  Vector is the LAPIC interrupt vector for which the
> >>   EOI was received.
> >>
> >> +  /* KVM_EXIT_HYPERV */
> >> +struct kvm_hyperv_exit hyperv;
> >> +Indicates that the VCPU's exits into userspace to process some tasks
> > s/VCPU's/VCPU/
> >
> >> +related with Hyper-V emulation. Currently used to synchronize modified
> >> +Hyper-V synic state with userspace.
> > Again, is 'synic' intended?  Hmm, I see it throughout the patch, so it
> > looks intentional, but I keep trying to read it as a typo for 'sync'.
> >
> this is not a typo :)

Yes; the Hyper-V public functional spec has chosen this name;
it stands for Synthetic Interrupt Controller.

K. Y

___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

[PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Denis V. Lunev]

From: Andrey Smetanin 

A new vcpu exit is introduced to notify the userspace of the
changes in Hyper-V synic configuraion triggered by guest writing to the
corresponding MSRs.

Signed-off-by: Andrey Smetanin 
Reviewed-by: Roman Kagan 
Signed-off-by: Denis V. Lunev 
CC: Vitaly Kuznetsov 
CC: "K. Y. Srinivasan" 
CC: Gleb Natapov 
CC: Paolo Bonzini 
---
 Documentation/virtual/kvm/api.txt |  6 ++
 arch/x86/include/asm/kvm_host.h   |  1 +
 arch/x86/kvm/hyperv.c | 17 +
 arch/x86/kvm/x86.c|  6 ++
 include/linux/kvm_host.h  |  1 +
 include/uapi/linux/kvm.h  | 17 +
 6 files changed, 48 insertions(+)

diff --git a/Documentation/virtual/kvm/api.txt 
b/Documentation/virtual/kvm/api.txt
index 34cc068..cffe670 100644
--- a/Documentation/virtual/kvm/api.txt
+++ b/Documentation/virtual/kvm/api.txt
@@ -3331,6 +3331,12 @@ the userspace IOAPIC should process the EOI and 
retrigger the interrupt if
 it is still asserted.  Vector is the LAPIC interrupt vector for which the
 EOI was received.
 
+   /* KVM_EXIT_HYPERV */
+struct kvm_hyperv_exit hyperv;
+Indicates that the VCPU's exits into userspace to process some tasks
+related with Hyper-V emulation. Currently used to synchronize modified
+Hyper-V synic state with userspace.
+
/* Fix the size of the union. */
char padding[256];
};
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index e614a543..f515e01 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -392,6 +392,7 @@ struct kvm_vcpu_hv {
u64 hv_vapic;
s64 runtime_offset;
struct kvm_vcpu_hv_synic synic;
+   struct kvm_hyperv_exit exit;
 };
 
 struct kvm_vcpu_arch {
diff --git a/arch/x86/kvm/hyperv.c b/arch/x86/kvm/hyperv.c
index 15c3c02..174ce041 100644
--- a/arch/x86/kvm/hyperv.c
+++ b/arch/x86/kvm/hyperv.c
@@ -91,6 +91,20 @@ static int synic_set_sint(struct kvm_vcpu_hv_synic *synic, 
int sint, u64 data)
return 0;
 }
 
+static void synic_exit(struct kvm_vcpu_hv_synic *synic, u32 msr)
+{
+   struct kvm_vcpu *vcpu = synic_to_vcpu(synic);
+   struct kvm_vcpu_hv *hv_vcpu = >arch.hyperv;
+
+   hv_vcpu->exit.type = KVM_EXIT_HYPERV_SYNIC;
+   hv_vcpu->exit.u.synic.msr = msr;
+   hv_vcpu->exit.u.synic.control = synic->control;
+   hv_vcpu->exit.u.synic.evt_page = synic->evt_page;
+   hv_vcpu->exit.u.synic.msg_page = synic->msg_page;
+
+   kvm_make_request(KVM_REQ_HV_EXIT, vcpu);
+}
+
 static int synic_set_msr(struct kvm_vcpu_hv_synic *synic,
 u32 msr, u64 data, bool host)
 {
@@ -103,6 +117,7 @@ static int synic_set_msr(struct kvm_vcpu_hv_synic *synic,
switch (msr) {
case HV_X64_MSR_SCONTROL:
synic->control = data;
+   synic_exit(synic, msr);
break;
case HV_X64_MSR_SVERSION:
if (!host) {
@@ -119,6 +134,7 @@ static int synic_set_msr(struct kvm_vcpu_hv_synic *synic,
break;
}
synic->evt_page = data;
+   synic_exit(synic, msr);
break;
case HV_X64_MSR_SIMP:
if (data & HV_SYNIC_SIMP_ENABLE)
@@ -128,6 +144,7 @@ static int synic_set_msr(struct kvm_vcpu_hv_synic *synic,
break;
}
synic->msg_page = data;
+   synic_exit(synic, msr);
break;
case HV_X64_MSR_EOM: {
int i;
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 7580e9c..4c80d18 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -6335,6 +6335,12 @@ static int vcpu_enter_guest(struct kvm_vcpu *vcpu)
r = 0;
goto out;
}
+   if (kvm_check_request(KVM_REQ_HV_EXIT, vcpu)) {
+   vcpu->run->exit_reason = KVM_EXIT_HYPERV;
+   vcpu->run->hyperv = vcpu->arch.hyperv.exit;
+   r = 0;
+   goto out;
+   }
}
 
/*
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 30fac73..d80b031 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -143,6 +143,7 @@ static inline bool is_error_page(struct page *page)
 #define KVM_REQ_HV_CRASH  27
 #define KVM_REQ_IOAPIC_EOI_EXIT   28
 #define KVM_REQ_HV_RESET  29
+#define KVM_REQ_HV_EXIT   30
 
 #define KVM_USERSPACE_IRQ_SOURCE_ID0
 #define KVM_IRQFD_RESAMPLE_IRQ_SOURCE_ID   1
diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h
index 27ce460..6e32f75 100644
--- a/include/uapi/linux/kvm.h

Re: [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Paolo Bonzini]

On 09/10/2015 15:39, Denis V. Lunev wrote:
> From: Andrey Smetanin 
> 
> A new vcpu exit is introduced to notify the userspace of the
> changes in Hyper-V synic configuraion triggered by guest writing to the
> corresponding MSRs.
> 
> Signed-off-by: Andrey Smetanin 
> Reviewed-by: Roman Kagan 
> Signed-off-by: Denis V. Lunev 
> CC: Vitaly Kuznetsov 
> CC: "K. Y. Srinivasan" 
> CC: Gleb Natapov 
> CC: Paolo Bonzini 

Why is this exit necessary?

Paolo
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

Re: [PATCH 2/2] kvm/x86: Hyper-V kvm exit

[Paolo Bonzini]

On 09/10/2015 16:53, Roman Kagan wrote:
>> > Why is this exit necessary?
> The guest writes to synic-related MSRs and that should take "immediate"
> effect.
> 
> E.g. it may decide to disable or relocate the message page by writing to
> SIMP MSR.  The host is then supposed to stop accessing the old message
> page before the vCPU proceeds to the next instruction.  Hence the exit,
> to allow the userspace to react accordingly before reentering the guest.

Ok, thanks!

Paolo
___
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization