2018-11-12 12:25:13

by Suthikulpanit, Suravee

Subject: [PATCH] svm: Add mutex_lock to protect apic_access_page_done on AMD systems

From: Wei Wang <[email protected]>

There is a race condition when accessing kvm->arch.apic_access_page_done.
Due to it, x86_set_memory_region will fail when creating the second vcpu
for a svm guest.

Add a mutex_lock to serialize the accesses to apic_access_page_done.
This lock is also used by vmx for the same purpose.

Signed-off-by: Wei Wang <[email protected]>
Signed-off-by: Amadeusz Juskowiak <[email protected]>
Signed-off-by: Julian Stecklina <[email protected]>
Signed-off-by: Suravee Suthikulpanit <[email protected]>
---
arch/x86/kvm/svm.c | 19 +++++++++++--------
1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index d96092b35936..8be1718fed1d 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -1658,20 +1658,23 @@ static u64 *avic_get_physical_id_entry(struct kvm_vcpu *vcpu,
static int avic_init_access_page(struct kvm_vcpu *vcpu)
{
struct kvm *kvm = vcpu->kvm;
- int ret;
+ int ret = 0;

+ mutex_lock(&kvm->slots_lock);
if (kvm->arch.apic_access_page_done)
- return 0;
+ goto out;

- ret = x86_set_memory_region(kvm,
- APIC_ACCESS_PAGE_PRIVATE_MEMSLOT,
- APIC_DEFAULT_PHYS_BASE,
- PAGE_SIZE);
+ ret = __x86_set_memory_region(kvm,
+ APIC_ACCESS_PAGE_PRIVATE_MEMSLOT,
+ APIC_DEFAULT_PHYS_BASE,
+ PAGE_SIZE);
if (ret)
- return ret;
+ goto out;

kvm->arch.apic_access_page_done = true;
- return 0;
+out:
+ mutex_unlock(&kvm->slots_lock);
+ return ret;
}

static int avic_init_backing_page(struct kvm_vcpu *vcpu)
--
2.17.1

2018-11-12 13:38:37

by Joerg Roedel

Subject: Re: [PATCH] svm: Add mutex_lock to protect apic_access_page_done on AMD systems

On Mon, Nov 12, 2018 at 12:23:14PM +0000, Suthikulpanit, Suravee wrote:
> From: Wei Wang <[email protected]>
>
> There is a race condition when accessing kvm->arch.apic_access_page_done.
> Due to it, x86_set_memory_region will fail when creating the second vcpu
> for a svm guest.
>
> Add a mutex_lock to serialize the accesses to apic_access_page_done.
> This lock is also used by vmx for the same purpose.
>
> Signed-off-by: Wei Wang <[email protected]>
> Signed-off-by: Amadeusz Juskowiak <[email protected]>
> Signed-off-by: Julian Stecklina <[email protected]>
> Signed-off-by: Suravee Suthikulpanit <[email protected]>

Reviewed-by: Joerg Roedel <[email protected]>

2018-11-25 17:35:37

by Paolo Bonzini

Subject: Re: [PATCH] svm: Add mutex_lock to protect apic_access_page_done on AMD systems

On 12/11/18 14:37, [email protected] wrote:
> On Mon, Nov 12, 2018 at 12:23:14PM +0000, Suthikulpanit, Suravee wrote:
>> From: Wei Wang <[email protected]>
>>
>> There is a race condition when accessing kvm->arch.apic_access_page_done.
>> Due to it, x86_set_memory_region will fail when creating the second vcpu
>> for a svm guest.
>>
>> Add a mutex_lock to serialize the accesses to apic_access_page_done.
>> This lock is also used by vmx for the same purpose.
>>
>> Signed-off-by: Wei Wang <[email protected]>
>> Signed-off-by: Amadeusz Juskowiak <[email protected]>
>> Signed-off-by: Julian Stecklina <[email protected]>
>> Signed-off-by: Suravee Suthikulpanit <[email protected]>
> Reviewed-by: Joerg Roedel <[email protected]>

Queued, thanks.

Paolo