Re: [Qemu-devel] [PATCH] virtio: zero vq->inuse in virtio_reset()

[Cornelia Huck]

On Wed,  7 Sep 2016 11:51:25 -0400
Stefan Hajnoczi  wrote:

> vq->inuse must be zeroed upon device reset like most other virtqueue
> fields.
> 
> In theory, virtio_reset() just needs assert(vq->inuse == 0) since
> devices must clean up in-flight requests during reset (requests cannot
> not be leaked!).
> 
> In practice, it is difficult to achieve vq->inuse == 0 across reset
> because balloon, blk, 9p, etc implement various different strategies for
> cleaning up requests.  Most devices call g_free(elem) directly without
> telling virtio.c that the VirtQueueElement is cleaned up.  Therefore
> vq->inuse is not decremented during reset.
> 
> This patch zeroes vq->inuse and trusts that devices are not leaking
> VirtQueueElements across reset.
> 
> I will send a follow-up series that refactors request life-cycle across
> all devices and converts vq->inuse = 0 into assert(vq->inuse == 0) but
> this more invasive approach is not appropriate for stable trees.
> 
> Signed-off-by: Stefan Hajnoczi 
> ---
>  hw/virtio/virtio.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
> index 74c085c..e8a13a5 100644
> --- a/hw/virtio/virtio.c
> +++ b/hw/virtio/virtio.c
> @@ -822,6 +822,7 @@ void virtio_reset(void *opaque)
>  vdev->vq[i].signalled_used_valid = false;
>  vdev->vq[i].notification = true;
>  vdev->vq[i].vring.num = vdev->vq[i].vring.num_default;
> +vdev->vq[i].inuse = 0;
>  }
>  }
> 

This looks sane as a minimal change, and survives my smoketests.

Reviewed-by: Cornelia Huck 

Re: [Qemu-devel] [PATCH] virtio: zero vq->inuse in virtio_reset()

[Ladi Prosek]

On Wed, Sep 7, 2016 at 5:54 PM, Stefan Hajnoczi  wrote:
> On Wed, Sep 7, 2016 at 11:51 AM, Stefan Hajnoczi  wrote:
>> vq->inuse must be zeroed upon device reset like most other virtqueue
>> fields.
>>
>> In theory, virtio_reset() just needs assert(vq->inuse == 0) since
>> devices must clean up in-flight requests during reset (requests cannot
>> not be leaked!).
>>
>> In practice, it is difficult to achieve vq->inuse == 0 across reset
>> because balloon, blk, 9p, etc implement various different strategies for
>> cleaning up requests.  Most devices call g_free(elem) directly without
>> telling virtio.c that the VirtQueueElement is cleaned up.  Therefore
>> vq->inuse is not decremented during reset.
>>
>> This patch zeroes vq->inuse and trusts that devices are not leaking
>> VirtQueueElements across reset.
>>
>> I will send a follow-up series that refactors request life-cycle across
>> all devices and converts vq->inuse = 0 into assert(vq->inuse == 0) but
>> this more invasive approach is not appropriate for stable trees.
>>
>> Signed-off-by: Stefan Hajnoczi 
>> ---
>>  hw/virtio/virtio.c | 1 +
>>  1 file changed, 1 insertion(+)
>>
>> diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
>> index 74c085c..e8a13a5 100644
>> --- a/hw/virtio/virtio.c
>> +++ b/hw/virtio/virtio.c
>> @@ -822,6 +822,7 @@ void virtio_reset(void *opaque)
>>  vdev->vq[i].signalled_used_valid = false;
>>  vdev->vq[i].notification = true;
>>  vdev->vq[i].vring.num = vdev->vq[i].vring.num_default;
>> +vdev->vq[i].inuse = 0;
>>  }
>>  }
>>
>> --
>> 2.7.4
>
> CCing qemu-stable

Reviewed-by: Ladi Prosek 

Re: [Qemu-devel] [PATCH] virtio: zero vq->inuse in virtio_reset()

[Stefan Hajnoczi]

On Wed, Sep 7, 2016 at 11:51 AM, Stefan Hajnoczi  wrote:
> vq->inuse must be zeroed upon device reset like most other virtqueue
> fields.
>
> In theory, virtio_reset() just needs assert(vq->inuse == 0) since
> devices must clean up in-flight requests during reset (requests cannot
> not be leaked!).
>
> In practice, it is difficult to achieve vq->inuse == 0 across reset
> because balloon, blk, 9p, etc implement various different strategies for
> cleaning up requests.  Most devices call g_free(elem) directly without
> telling virtio.c that the VirtQueueElement is cleaned up.  Therefore
> vq->inuse is not decremented during reset.
>
> This patch zeroes vq->inuse and trusts that devices are not leaking
> VirtQueueElements across reset.
>
> I will send a follow-up series that refactors request life-cycle across
> all devices and converts vq->inuse = 0 into assert(vq->inuse == 0) but
> this more invasive approach is not appropriate for stable trees.
>
> Signed-off-by: Stefan Hajnoczi 
> ---
>  hw/virtio/virtio.c | 1 +
>  1 file changed, 1 insertion(+)
>
> diff --git a/hw/virtio/virtio.c b/hw/virtio/virtio.c
> index 74c085c..e8a13a5 100644
> --- a/hw/virtio/virtio.c
> +++ b/hw/virtio/virtio.c
> @@ -822,6 +822,7 @@ void virtio_reset(void *opaque)
>  vdev->vq[i].signalled_used_valid = false;
>  vdev->vq[i].notification = true;
>  vdev->vq[i].vring.num = vdev->vq[i].vring.num_default;
> +vdev->vq[i].inuse = 0;
>  }
>  }
>
> --
> 2.7.4

CCing qemu-stable