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

Groups > linux.kernel > #1642369 > unrolled thread

[PATCH v3 0/2] kvm: arm/arm64: Fixes for race conditions

Back to article view | Back to linux.kernel

Contents

  [PATCH v3 0/2] kvm: arm/arm64: Fixes for race conditions Suzuki K Poulose <suzuki.poulose@arm.com> - 2017-05-16 11:40 +0200
    Re: [PATCH v3 0/2] kvm: arm/arm64: Fixes for race conditions Christoffer Dall <cdall@linaro.org> - 2017-05-16 12:00 +0200
      Re: [PATCH v3 0/2] kvm: arm/arm64: Fixes for race conditions Suzuki K Poulose <Suzuki.Poulose@arm.com> - 2017-05-16 12:10 +0200
        Re: [PATCH v3 0/2] kvm: arm/arm64: Fixes for race conditions Christoffer Dall <cdall@linaro.org> - 2017-05-16 15:30 +0200

#1642369 — [PATCH v3 0/2] kvm: arm/arm64: Fixes for race conditions

The patches fixes race conditions in stage2 pgd accesses.

Patch 1 is a fix up for the patch which has already been pushed to
kvmarm/master.

Patch 2 fixes a case where stage2 PGD could be free'd when we release
the kvm->mmu_lock to prevent VCPU starvation.

Applies on kvmarm/master tree

Changes since V2:
http://lists.infradead.org/pipermail/linux-arm-kernel/2017-May/504401.html

 - Patch 1 in V2 has been pushed to kvmarm/master. Added a fixup to the
   original patch to make sure we don't use cached stage2 pgd.
 - Patch 2: Handle cases where unmap_stage2_range may be called without
   checking the validity of stage2 pgd.


Changes since v1:
 http://lists.infradead.org/pipermail/linux-arm-kernel/2017-April/502867.html

 - Dropped patch for fixing mmu_notifier race condition, which couldn't be
   reproduced.
 - Added reviewed-by from Christoffer
 - Added new patch to fix another race condition

Suzuki K Poulose (2):
  kvm: arm/arm64: Force reading uncached stage2 PGD
  kvm: arm/arm64: Fix use after free of stage2 page table

 virt/kvm/arm/mmu.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

-- 
2.7.4

[toc] | [next] | [standalone]

#1642375

On Tue, May 16, 2017 at 10:34:53AM +0100, Suzuki K Poulose wrote:
> The patches fixes race conditions in stage2 pgd accesses.
> 
> Patch 1 is a fix up for the patch which has already been pushed to
> kvmarm/master.
> 
> Patch 2 fixes a case where stage2 PGD could be free'd when we release
> the kvm->mmu_lock to prevent VCPU starvation.
> 
> Applies on kvmarm/master tree

Thanks for the rebase and fix.

Applied to master, and for both patches::

Reviewed-by: Christoffer Dall <cdall@linaro.org>

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

#1642381

On 16/05/17 10:53, Christoffer Dall wrote:
> On Tue, May 16, 2017 at 10:34:53AM +0100, Suzuki K Poulose wrote:
>> The patches fixes race conditions in stage2 pgd accesses.
>>
>> Patch 1 is a fix up for the patch which has already been pushed to
>> kvmarm/master.
>>
>> Patch 2 fixes a case where stage2 PGD could be free'd when we release
>> the kvm->mmu_lock to prevent VCPU starvation.
>>
>> Applies on kvmarm/master tree
>
> Thanks for the rebase and fix.
>
> Applied to master, and for both patches::
>
> Reviewed-by: Christoffer Dall <cdall@linaro.org>
>

Thanks, I think these fixes needs to go to stable as well. I will wait for
these to hit mainline and then send a back port.

Suzuki

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

#1642561

On Tue, May 16, 2017 at 11:01:54AM +0100, Suzuki K Poulose wrote:
> On 16/05/17 10:53, Christoffer Dall wrote:
> >On Tue, May 16, 2017 at 10:34:53AM +0100, Suzuki K Poulose wrote:
> >>The patches fixes race conditions in stage2 pgd accesses.
> >>
> >>Patch 1 is a fix up for the patch which has already been pushed to
> >>kvmarm/master.
> >>
> >>Patch 2 fixes a case where stage2 PGD could be free'd when we release
> >>the kvm->mmu_lock to prevent VCPU starvation.
> >>
> >>Applies on kvmarm/master tree
> >
> >Thanks for the rebase and fix.
> >
> >Applied to master, and for both patches::
> >
> >Reviewed-by: Christoffer Dall <cdall@linaro.org>
> >
> 
> Thanks, I think these fixes needs to go to stable as well. I will wait for
> these to hit mainline and then send a back port.

Agreed, I already added the cc to stable when applying them, so I think
you'll see the report from the stable folks for the parts that don't
apply and can do the backport then.

Thanks,
-Christoffer

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web