Re: [PATCH v2] kvm: fix waitqueue_active without memory barrier in virt/kvm/async_pf.c

[William Dauchy <william@gandi.net>]

[-- Attachment #1: Type: text/plain, Size: 3043 bytes --]

Hi Paolo,

Looking at the history of this function, is it reasonable to say it
fixes the following commit?
af585b9 KVM: Halt vcpu if page it tries to access is swapped out

Does it make it a good candidate for -stable?

Thanks,

On Oct09 12:21, Kosuke Tatsukawa wrote:
> async_pf_execute() seems to be missing a memory barrier which might
> cause the waker to not notice the waiter and miss sending a wake_up as
> in the following figure.
> 
>         async_pf_execute                    kvm_vcpu_block
> ------------------------------------------------------------------------
> spin_lock(&vcpu->async_pf.lock);
> if (waitqueue_active(&vcpu->wq))
> /* The CPU might reorder the test for
>    the waitqueue up here, before
>    prior writes complete */
>                                     prepare_to_wait(&vcpu->wq, &wait,
>                                       TASK_INTERRUPTIBLE);
>                                     /*if (kvm_vcpu_check_block(vcpu) < 0) */
>                                      /*if (kvm_arch_vcpu_runnable(vcpu)) { */
>                                       ...
>                                       return (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE &&
>                                         !vcpu->arch.apf.halted)
>                                         || !list_empty_careful(&vcpu->async_pf.done)
>                                      ...
>                                      return 0;
> list_add_tail(&apf->link,
>   &vcpu->async_pf.done);
> spin_unlock(&vcpu->async_pf.lock);
>                                     waited = true;
>                                     schedule();
> ------------------------------------------------------------------------
> 
> The attached patch adds the missing memory barrier.
> 
> I found this issue when I was looking through the linux source code
> for places calling waitqueue_active() before wake_up*(), but without
> preceding memory barriers, after sending a patch to fix a similar
> issue in drivers/tty/n_tty.c  (Details about the original issue can be
> found here: https://lkml.org/lkml/2015/9/28/849).
> 
> Signed-off-by: Kosuke Tatsukawa <tatsu@ab.jp.nec.com>
> ---
> v2:
>   - Fixed comment based on feedback from Paolo
> v1:
>   - https://lkml.org/lkml/2015/10/8/994
> ---
>  virt/kvm/async_pf.c |    6 ++++++
>  1 files changed, 6 insertions(+), 0 deletions(-)
> 
> diff --git a/virt/kvm/async_pf.c b/virt/kvm/async_pf.c
> index 44660ae..a0999d7 100644
> --- a/virt/kvm/async_pf.c
> +++ b/virt/kvm/async_pf.c
> @@ -94,6 +94,12 @@ static void async_pf_execute(struct work_struct *work)
>  
>  	trace_kvm_async_pf_completed(addr, gva);
>  
> +	/*
> +	 * Memory barrier is required here to make sure change to
> +	 * vcpu->async_pf.done is visible from other CPUs.  This memory
> +	 * barrier pairs with prepare_to_wait's set_current_state()
> +	 */
> +	smp_mb();
>  	if (waitqueue_active(&vcpu->wq))
>  		wake_up_interruptible(&vcpu->wq);
>  
> -- 
> 1.7.1

-- 
William

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 181 bytes --]