Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1636672 > unrolled thread

[PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization

Started byBandan Das <bsd@redhat.com>
First post2017-05-05 21:30 +0200
Last post2017-05-11 07:10 +0200
Articles 9 — 3 participants

Back to article view | Back to linux.kernel

Contents

  [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization Bandan Das <bsd@redhat.com> - 2017-05-05 21:30 +0200
    [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation Bandan Das <bsd@redhat.com> - 2017-05-05 21:30 +0200
      Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty  logging emulation "Huang, Kai" <kai.huang@linux.intel.com> - 2017-05-10 13:00 +0200
        Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation Bandan Das <bsd@redhat.com> - 2017-05-10 18:00 +0200
          Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty  logging emulation "Huang, Kai" <kai.huang@linux.intel.com> - 2017-05-11 01:30 +0200
            Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation Bandan Das <bsd@redhat.com> - 2017-05-11 20:40 +0200
    Re: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for  Nested Virtualization Paolo Bonzini <pbonzini@redhat.com> - 2017-05-09 17:30 +0200
      Re: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization Bandan Das <bsd@redhat.com> - 2017-05-09 18:10 +0200
        Re: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for  Nested Virtualization Paolo Bonzini <pbonzini@redhat.com> - 2017-05-11 07:10 +0200

#1636672 — [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization

FromBandan Das <bsd@redhat.com>
Date2017-05-05 21:30 +0200
Subject[PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization
Message-ID<tDR5o-5Ye-5@gated-at.bofh.it>
v2:
2/3: Clear out all bits except bit 12
3/3: Slightly modify an existing comment, honor L0's
PML setting when clearing it for L1

v1:
http://www.spinics.net/lists/kvm/msg149247.html

These patches implement PML on top of EPT A/D emulation
(ae1e2d1082ae).

When dirty bit is being set, we write the gpa to the
buffer provided by L1. If the index overflows, we just
change the exit reason before running L1.

Bandan Das (3):
  kvm: x86: Add a hook for arch specific dirty logging emulation
  nVMX: Implement emulated Page Modification Logging
  nVMX: Advertise PML to L1 hypervisor

 arch/x86/include/asm/kvm_host.h |  2 +
 arch/x86/kvm/mmu.c              | 15 +++++++
 arch/x86/kvm/mmu.h              |  1 +
 arch/x86/kvm/paging_tmpl.h      |  4 ++
 arch/x86/kvm/vmx.c              | 97 ++++++++++++++++++++++++++++++++++++++---
 5 files changed, 112 insertions(+), 7 deletions(-)

-- 
2.9.3

[toc] | [next] | [standalone]

#1636674 — [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation

FromBandan Das <bsd@redhat.com>
Date2017-05-05 21:30 +0200
Subject[PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation
Message-ID<tDR5o-5Ye-17@gated-at.bofh.it>
In reply to#1636672
When KVM updates accessed/dirty bits, this hook can be used
to invoke an arch specific function that implements/emulates
dirty logging such as PML.

Signed-off-by: Bandan Das <bsd@redhat.com>
---
 arch/x86/include/asm/kvm_host.h |  2 ++
 arch/x86/kvm/mmu.c              | 15 +++++++++++++++
 arch/x86/kvm/mmu.h              |  1 +
 arch/x86/kvm/paging_tmpl.h      |  4 ++++
 4 files changed, 22 insertions(+)

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index f5bddf92..9c761fe 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -1020,6 +1020,8 @@ struct kvm_x86_ops {
 	void (*enable_log_dirty_pt_masked)(struct kvm *kvm,
 					   struct kvm_memory_slot *slot,
 					   gfn_t offset, unsigned long mask);
+	int (*write_log_dirty)(struct kvm_vcpu *vcpu);
+
 	/* pmu operations of sub-arch */
 	const struct kvm_pmu_ops *pmu_ops;
 
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 5586765..5d3376f 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1498,6 +1498,21 @@ void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm,
 		kvm_mmu_write_protect_pt_masked(kvm, slot, gfn_offset, mask);
 }
 
+/**
+ * kvm_arch_write_log_dirty - emulate dirty page logging
+ * @vcpu: Guest mode vcpu
+ *
+ * Emulate arch specific page modification logging for the
+ * nested hypervisor
+ */
+int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu)
+{
+	if (kvm_x86_ops->write_log_dirty)
+		return kvm_x86_ops->write_log_dirty(vcpu);
+
+	return 0;
+}
+
 bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
 				    struct kvm_memory_slot *slot, u64 gfn)
 {
diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
index d8ccb32..2797580 100644
--- a/arch/x86/kvm/mmu.h
+++ b/arch/x86/kvm/mmu.h
@@ -202,4 +202,5 @@ void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
 void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
 bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
 				    struct kvm_memory_slot *slot, u64 gfn);
+int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
 #endif
diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index 314d207..5624174 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -226,6 +226,10 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu,
 		if (level == walker->level && write_fault &&
 				!(pte & PT_GUEST_DIRTY_MASK)) {
 			trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte));
+#if PTTYPE == PTTYPE_EPT
+			if (kvm_arch_write_log_dirty(vcpu))
+				return -EINVAL;
+#endif
 			pte |= PT_GUEST_DIRTY_MASK;
 		}
 		if (pte == orig_pte)
-- 
2.9.3

[toc] | [prev] | [next] | [standalone]

#1638747 — Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation

From"Huang, Kai" <kai.huang@linux.intel.com>
Date2017-05-10 13:00 +0200
SubjectRe: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation
Message-ID<tFxvz-85D-7@gated-at.bofh.it>
In reply to#1636674

On 5/6/2017 7:25 AM, Bandan Das wrote:
> When KVM updates accessed/dirty bits, this hook can be used
> to invoke an arch specific function that implements/emulates
> dirty logging such as PML.
>
> Signed-off-by: Bandan Das <bsd@redhat.com>
> ---
>  arch/x86/include/asm/kvm_host.h |  2 ++
>  arch/x86/kvm/mmu.c              | 15 +++++++++++++++
>  arch/x86/kvm/mmu.h              |  1 +
>  arch/x86/kvm/paging_tmpl.h      |  4 ++++
>  4 files changed, 22 insertions(+)
>
> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> index f5bddf92..9c761fe 100644
> --- a/arch/x86/include/asm/kvm_host.h
> +++ b/arch/x86/include/asm/kvm_host.h
> @@ -1020,6 +1020,8 @@ struct kvm_x86_ops {
>  	void (*enable_log_dirty_pt_masked)(struct kvm *kvm,
>  					   struct kvm_memory_slot *slot,
>  					   gfn_t offset, unsigned long mask);
> +	int (*write_log_dirty)(struct kvm_vcpu *vcpu);

Hi,

Thanks for adding PML to nested support!

IMHO this callback is only used for write L2's dirty gpa to L1's PML 
buffer, so probably it's better to change the name to something like: 
nested_write_log_dirty.

Thanks,
-Kai

> +
>  	/* pmu operations of sub-arch */
>  	const struct kvm_pmu_ops *pmu_ops;
>
> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
> index 5586765..5d3376f 100644
> --- a/arch/x86/kvm/mmu.c
> +++ b/arch/x86/kvm/mmu.c
> @@ -1498,6 +1498,21 @@ void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm,
>  		kvm_mmu_write_protect_pt_masked(kvm, slot, gfn_offset, mask);
>  }
>
> +/**
> + * kvm_arch_write_log_dirty - emulate dirty page logging
> + * @vcpu: Guest mode vcpu
> + *
> + * Emulate arch specific page modification logging for the
> + * nested hypervisor
> + */
> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu)
> +{
> +	if (kvm_x86_ops->write_log_dirty)
> +		return kvm_x86_ops->write_log_dirty(vcpu);
> +
> +	return 0;
> +}

kvm_nested_arch_write_log_dirty?

> +
>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>  				    struct kvm_memory_slot *slot, u64 gfn)
>  {
> diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
> index d8ccb32..2797580 100644
> --- a/arch/x86/kvm/mmu.h
> +++ b/arch/x86/kvm/mmu.h
> @@ -202,4 +202,5 @@ void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>  void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>  				    struct kvm_memory_slot *slot, u64 gfn);
> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
>  #endif
> diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
> index 314d207..5624174 100644
> --- a/arch/x86/kvm/paging_tmpl.h
> +++ b/arch/x86/kvm/paging_tmpl.h
> @@ -226,6 +226,10 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu,
>  		if (level == walker->level && write_fault &&
>  				!(pte & PT_GUEST_DIRTY_MASK)) {
>  			trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte));
> +#if PTTYPE == PTTYPE_EPT
> +			if (kvm_arch_write_log_dirty(vcpu))
> +				return -EINVAL;
> +#endif
>  			pte |= PT_GUEST_DIRTY_MASK;
>  		}
>  		if (pte == orig_pte)
>

[toc] | [prev] | [next] | [standalone]

#1638929 — Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation

FromBandan Das <bsd@redhat.com>
Date2017-05-10 18:00 +0200
SubjectRe: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation
Message-ID<tFCbU-2yr-19@gated-at.bofh.it>
In reply to#1638747
Hi Kai,

"Huang, Kai" <kai.huang@linux.intel.com> writes:

> On 5/6/2017 7:25 AM, Bandan Das wrote:
>> When KVM updates accessed/dirty bits, this hook can be used
>> to invoke an arch specific function that implements/emulates
>> dirty logging such as PML.
>>
>> Signed-off-by: Bandan Das <bsd@redhat.com>
>> ---
>>  arch/x86/include/asm/kvm_host.h |  2 ++
>>  arch/x86/kvm/mmu.c              | 15 +++++++++++++++
>>  arch/x86/kvm/mmu.h              |  1 +
>>  arch/x86/kvm/paging_tmpl.h      |  4 ++++
>>  4 files changed, 22 insertions(+)
>>
>> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
>> index f5bddf92..9c761fe 100644
>> --- a/arch/x86/include/asm/kvm_host.h
>> +++ b/arch/x86/include/asm/kvm_host.h
>> @@ -1020,6 +1020,8 @@ struct kvm_x86_ops {
>>  	void (*enable_log_dirty_pt_masked)(struct kvm *kvm,
>>  					   struct kvm_memory_slot *slot,
>>  					   gfn_t offset, unsigned long mask);
>> +	int (*write_log_dirty)(struct kvm_vcpu *vcpu);
>
> Hi,
>
> Thanks for adding PML to nested support!
>
> IMHO this callback is only used for write L2's dirty gpa to L1's PML
> buffer, so probably it's better to change the name to something like:
> nested_write_log_dirty.

The name was meant more to signify what it does: i.e. write dirty log rather
than where in the hierarchy it's being used :) But I guess, a nested_ prefix
doesn't hurt either.

Bandan

> Thanks,
> -Kai
>
>> +
>>  	/* pmu operations of sub-arch */
>>  	const struct kvm_pmu_ops *pmu_ops;
>>
>> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
>> index 5586765..5d3376f 100644
>> --- a/arch/x86/kvm/mmu.c
>> +++ b/arch/x86/kvm/mmu.c
>> @@ -1498,6 +1498,21 @@ void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm,
>>  		kvm_mmu_write_protect_pt_masked(kvm, slot, gfn_offset, mask);
>>  }
>>
>> +/**
>> + * kvm_arch_write_log_dirty - emulate dirty page logging
>> + * @vcpu: Guest mode vcpu
>> + *
>> + * Emulate arch specific page modification logging for the
>> + * nested hypervisor
>> + */
>> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu)
>> +{
>> +	if (kvm_x86_ops->write_log_dirty)
>> +		return kvm_x86_ops->write_log_dirty(vcpu);
>> +
>> +	return 0;
>> +}
>
> kvm_nested_arch_write_log_dirty?
>
>> +
>>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>>  				    struct kvm_memory_slot *slot, u64 gfn)
>>  {
>> diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
>> index d8ccb32..2797580 100644
>> --- a/arch/x86/kvm/mmu.h
>> +++ b/arch/x86/kvm/mmu.h
>> @@ -202,4 +202,5 @@ void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>>  void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>>  				    struct kvm_memory_slot *slot, u64 gfn);
>> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
>>  #endif
>> diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
>> index 314d207..5624174 100644
>> --- a/arch/x86/kvm/paging_tmpl.h
>> +++ b/arch/x86/kvm/paging_tmpl.h
>> @@ -226,6 +226,10 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu,
>>  		if (level == walker->level && write_fault &&
>>  				!(pte & PT_GUEST_DIRTY_MASK)) {
>>  			trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte));
>> +#if PTTYPE == PTTYPE_EPT
>> +			if (kvm_arch_write_log_dirty(vcpu))
>> +				return -EINVAL;
>> +#endif
>>  			pte |= PT_GUEST_DIRTY_MASK;
>>  		}
>>  		if (pte == orig_pte)
>>

[toc] | [prev] | [next] | [standalone]

#1639108 — Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation

From"Huang, Kai" <kai.huang@linux.intel.com>
Date2017-05-11 01:30 +0200
SubjectRe: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation
Message-ID<tFJdn-70m-3@gated-at.bofh.it>
In reply to#1638929

On 5/11/2017 3:53 AM, Bandan Das wrote:
> Hi Kai,
>
> "Huang, Kai" <kai.huang@linux.intel.com> writes:
>
>> On 5/6/2017 7:25 AM, Bandan Das wrote:
>>> When KVM updates accessed/dirty bits, this hook can be used
>>> to invoke an arch specific function that implements/emulates
>>> dirty logging such as PML.
>>>
>>> Signed-off-by: Bandan Das <bsd@redhat.com>
>>> ---
>>>  arch/x86/include/asm/kvm_host.h |  2 ++
>>>  arch/x86/kvm/mmu.c              | 15 +++++++++++++++
>>>  arch/x86/kvm/mmu.h              |  1 +
>>>  arch/x86/kvm/paging_tmpl.h      |  4 ++++
>>>  4 files changed, 22 insertions(+)
>>>
>>> diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
>>> index f5bddf92..9c761fe 100644
>>> --- a/arch/x86/include/asm/kvm_host.h
>>> +++ b/arch/x86/include/asm/kvm_host.h
>>> @@ -1020,6 +1020,8 @@ struct kvm_x86_ops {
>>>  	void (*enable_log_dirty_pt_masked)(struct kvm *kvm,
>>>  					   struct kvm_memory_slot *slot,
>>>  					   gfn_t offset, unsigned long mask);
>>> +	int (*write_log_dirty)(struct kvm_vcpu *vcpu);
>>
>> Hi,
>>
>> Thanks for adding PML to nested support!
>>
>> IMHO this callback is only used for write L2's dirty gpa to L1's PML
>> buffer, so probably it's better to change the name to something like:
>> nested_write_log_dirty.
>
> The name was meant more to signify what it does: i.e. write dirty log rather
> than where in the hierarchy it's being used :) But I guess, a nested_ prefix
> doesn't hurt either.

Hi Bandan,

I was just suggesting. You and Paolo still make the decision :)

Thanks,
-Kai
>
> Bandan
>
>> Thanks,
>> -Kai
>>
>>> +
>>>  	/* pmu operations of sub-arch */
>>>  	const struct kvm_pmu_ops *pmu_ops;
>>>
>>> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
>>> index 5586765..5d3376f 100644
>>> --- a/arch/x86/kvm/mmu.c
>>> +++ b/arch/x86/kvm/mmu.c
>>> @@ -1498,6 +1498,21 @@ void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm,
>>>  		kvm_mmu_write_protect_pt_masked(kvm, slot, gfn_offset, mask);
>>>  }
>>>
>>> +/**
>>> + * kvm_arch_write_log_dirty - emulate dirty page logging
>>> + * @vcpu: Guest mode vcpu
>>> + *
>>> + * Emulate arch specific page modification logging for the
>>> + * nested hypervisor
>>> + */
>>> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu)
>>> +{
>>> +	if (kvm_x86_ops->write_log_dirty)
>>> +		return kvm_x86_ops->write_log_dirty(vcpu);
>>> +
>>> +	return 0;
>>> +}
>>
>> kvm_nested_arch_write_log_dirty?
>>
>>> +
>>>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>>>  				    struct kvm_memory_slot *slot, u64 gfn)
>>>  {
>>> diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
>>> index d8ccb32..2797580 100644
>>> --- a/arch/x86/kvm/mmu.h
>>> +++ b/arch/x86/kvm/mmu.h
>>> @@ -202,4 +202,5 @@ void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>>>  void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>>>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>>>  				    struct kvm_memory_slot *slot, u64 gfn);
>>> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
>>>  #endif
>>> diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
>>> index 314d207..5624174 100644
>>> --- a/arch/x86/kvm/paging_tmpl.h
>>> +++ b/arch/x86/kvm/paging_tmpl.h
>>> @@ -226,6 +226,10 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu,
>>>  		if (level == walker->level && write_fault &&
>>>  				!(pte & PT_GUEST_DIRTY_MASK)) {
>>>  			trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte));
>>> +#if PTTYPE == PTTYPE_EPT
>>> +			if (kvm_arch_write_log_dirty(vcpu))
>>> +				return -EINVAL;
>>> +#endif
>>>  			pte |= PT_GUEST_DIRTY_MASK;
>>>  		}
>>>  		if (pte == orig_pte)
>>>
>

[toc] | [prev] | [next] | [standalone]

#1639951 — Re: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation

FromBandan Das <bsd@redhat.com>
Date2017-05-11 20:40 +0200
SubjectRe: [PATCH v2 1/3] kvm: x86: Add a hook for arch specific dirty logging emulation
Message-ID<tG1ah-1ut-11@gated-at.bofh.it>
In reply to#1639108
"Huang, Kai" <kai.huang@linux.intel.com> writes:

...
> Hi Bandan,
>
> I was just suggesting. You and Paolo still make the decision :)

Sure Kai, I don't mind the name change at all.
The maintainer has already picked this up and I don't think
the function name change is worth submitting a follow up.
Thank you very much for the review! :)

Bandan

> Thanks,
> -Kai
>>
>> Bandan
>>
>>> Thanks,
>>> -Kai
>>>
>>>> +
>>>>  	/* pmu operations of sub-arch */
>>>>  	const struct kvm_pmu_ops *pmu_ops;
>>>>
>>>> diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
>>>> index 5586765..5d3376f 100644
>>>> --- a/arch/x86/kvm/mmu.c
>>>> +++ b/arch/x86/kvm/mmu.c
>>>> @@ -1498,6 +1498,21 @@ void kvm_arch_mmu_enable_log_dirty_pt_masked(struct kvm *kvm,
>>>>  		kvm_mmu_write_protect_pt_masked(kvm, slot, gfn_offset, mask);
>>>>  }
>>>>
>>>> +/**
>>>> + * kvm_arch_write_log_dirty - emulate dirty page logging
>>>> + * @vcpu: Guest mode vcpu
>>>> + *
>>>> + * Emulate arch specific page modification logging for the
>>>> + * nested hypervisor
>>>> + */
>>>> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu)
>>>> +{
>>>> +	if (kvm_x86_ops->write_log_dirty)
>>>> +		return kvm_x86_ops->write_log_dirty(vcpu);
>>>> +
>>>> +	return 0;
>>>> +}
>>>
>>> kvm_nested_arch_write_log_dirty?
>>>
>>>> +
>>>>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>>>>  				    struct kvm_memory_slot *slot, u64 gfn)
>>>>  {
>>>> diff --git a/arch/x86/kvm/mmu.h b/arch/x86/kvm/mmu.h
>>>> index d8ccb32..2797580 100644
>>>> --- a/arch/x86/kvm/mmu.h
>>>> +++ b/arch/x86/kvm/mmu.h
>>>> @@ -202,4 +202,5 @@ void kvm_mmu_gfn_disallow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>>>>  void kvm_mmu_gfn_allow_lpage(struct kvm_memory_slot *slot, gfn_t gfn);
>>>>  bool kvm_mmu_slot_gfn_write_protect(struct kvm *kvm,
>>>>  				    struct kvm_memory_slot *slot, u64 gfn);
>>>> +int kvm_arch_write_log_dirty(struct kvm_vcpu *vcpu);
>>>>  #endif
>>>> diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
>>>> index 314d207..5624174 100644
>>>> --- a/arch/x86/kvm/paging_tmpl.h
>>>> +++ b/arch/x86/kvm/paging_tmpl.h
>>>> @@ -226,6 +226,10 @@ static int FNAME(update_accessed_dirty_bits)(struct kvm_vcpu *vcpu,
>>>>  		if (level == walker->level && write_fault &&
>>>>  				!(pte & PT_GUEST_DIRTY_MASK)) {
>>>>  			trace_kvm_mmu_set_dirty_bit(table_gfn, index, sizeof(pte));
>>>> +#if PTTYPE == PTTYPE_EPT
>>>> +			if (kvm_arch_write_log_dirty(vcpu))
>>>> +				return -EINVAL;
>>>> +#endif
>>>>  			pte |= PT_GUEST_DIRTY_MASK;
>>>>  		}
>>>>  		if (pte == orig_pte)
>>>>
>>

[toc] | [prev] | [next] | [standalone]

#1638192 — Re: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization

FromPaolo Bonzini <pbonzini@redhat.com>
Date2017-05-09 17:30 +0200
SubjectRe: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization
Message-ID<tFffj-3xV-9@gated-at.bofh.it>
In reply to#1636672

On 05/05/2017 21:25, Bandan Das wrote:
> v2:
> 2/3: Clear out all bits except bit 12
> 3/3: Slightly modify an existing comment, honor L0's
> PML setting when clearing it for L1
> 
> v1:
> http://www.spinics.net/lists/kvm/msg149247.html
> 
> These patches implement PML on top of EPT A/D emulation
> (ae1e2d1082ae).
> 
> When dirty bit is being set, we write the gpa to the
> buffer provided by L1. If the index overflows, we just
> change the exit reason before running L1.

I tested this with api/dirty-log-perf, and nested PML is more than 3
times faster than pml=0.  I want to do a few more tests because I don't
see any PML full exits in the L1 trace, but it seems to be a nice
improvement!

Paolo

> Bandan Das (3):
>   kvm: x86: Add a hook for arch specific dirty logging emulation
>   nVMX: Implement emulated Page Modification Logging
>   nVMX: Advertise PML to L1 hypervisor
> 
>  arch/x86/include/asm/kvm_host.h |  2 +
>  arch/x86/kvm/mmu.c              | 15 +++++++
>  arch/x86/kvm/mmu.h              |  1 +
>  arch/x86/kvm/paging_tmpl.h      |  4 ++
>  arch/x86/kvm/vmx.c              | 97 ++++++++++++++++++++++++++++++++++++++---
>  5 files changed, 112 insertions(+), 7 deletions(-)
>

[toc] | [prev] | [next] | [standalone]

#1638245

FromBandan Das <bsd@redhat.com>
Date2017-05-09 18:10 +0200
Message-ID<tFfS1-41K-15@gated-at.bofh.it>
In reply to#1638192
Paolo Bonzini <pbonzini@redhat.com> writes:

> On 05/05/2017 21:25, Bandan Das wrote:
>> v2:
>> 2/3: Clear out all bits except bit 12
>> 3/3: Slightly modify an existing comment, honor L0's
>> PML setting when clearing it for L1
>> 
>> v1:
>> http://www.spinics.net/lists/kvm/msg149247.html
>> 
>> These patches implement PML on top of EPT A/D emulation
>> (ae1e2d1082ae).
>> 
>> When dirty bit is being set, we write the gpa to the
>> buffer provided by L1. If the index overflows, we just
>> change the exit reason before running L1.
>
> I tested this with api/dirty-log-perf, and nested PML is more than 3
> times faster than pml=0.  I want to do a few more tests because I don't
> see any PML full exits in the L1 trace, but it seems to be a nice
> improvement!

Thanks for testing! Regarding the PML full exits, I did notice their
absence. I induced it artifically in my testing with a lower index
and it seemed to work fine.

> Paolo
>
>> Bandan Das (3):
>>   kvm: x86: Add a hook for arch specific dirty logging emulation
>>   nVMX: Implement emulated Page Modification Logging
>>   nVMX: Advertise PML to L1 hypervisor
>> 
>>  arch/x86/include/asm/kvm_host.h |  2 +
>>  arch/x86/kvm/mmu.c              | 15 +++++++
>>  arch/x86/kvm/mmu.h              |  1 +
>>  arch/x86/kvm/paging_tmpl.h      |  4 ++
>>  arch/x86/kvm/vmx.c              | 97 ++++++++++++++++++++++++++++++++++++++---
>>  5 files changed, 112 insertions(+), 7 deletions(-)
>>

[toc] | [prev] | [next] | [standalone]

#1639195 — Re: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization

FromPaolo Bonzini <pbonzini@redhat.com>
Date2017-05-11 07:10 +0200
SubjectRe: [PATCH v2 0/3] nVMX: Emulated Page Modification Logging for Nested Virtualization
Message-ID<tFOwp-22k-1@gated-at.bofh.it>
In reply to#1638245

On 09/05/2017 18:03, Bandan Das wrote:
>> I tested this with api/dirty-log-perf, and nested PML is more than 3
>> times faster than pml=0.  I want to do a few more tests because I don't
>> see any PML full exits in the L1 trace, but it seems to be a nice
>> improvement!
>
> Thanks for testing! Regarding the PML full exits, I did notice their
> absence. I induced it artifically in my testing with a lower index
> and it seemed to work fine.

Yes, tracing showed that it's because the guest's EXTERNAL_INTERRUPT
exits are too frequent.

Paolo

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web