[Intel-gfx] [PATCH v3 06/15] vfio/mtty: Use the new device life cycle helpers

[Alex Williamson]

On Fri,  9 Sep 2022 18:22:38 +0800
Kevin Tian <kevin.tian at intel.com> wrote:

> From: Yi Liu <yi.l.liu at intel.com>
> 
> and manage available ports inside @init/@release.
> 
> Signed-off-by: Yi Liu <yi.l.liu at intel.com>
> Signed-off-by: Kevin Tian <kevin.tian at intel.com>
> Reviewed-by: Jason Gunthorpe <jgg at nvidia.com>
> ---
>  samples/vfio-mdev/mtty.c | 67 +++++++++++++++++++++++-----------------
>  1 file changed, 39 insertions(+), 28 deletions(-)
> 
> diff --git a/samples/vfio-mdev/mtty.c b/samples/vfio-mdev/mtty.c
> index f42a59ed2e3f..41301d50b247 100644
> --- a/samples/vfio-mdev/mtty.c
> +++ b/samples/vfio-mdev/mtty.c
...
> +static int mtty_probe(struct mdev_device *mdev)
> +{
> +	struct mdev_state *mdev_state;
> +	int ret;
> +
> +	mdev_state = vfio_alloc_device(mdev_state, vdev, &mdev->dev,
> +				       &mtty_dev_ops);
> +	if (IS_ERR(mdev_state))
> +		return PTR_ERR(mdev_state);
>  
>  	ret = vfio_register_emulated_iommu_dev(&mdev_state->vdev);
>  	if (ret)
> -		goto err_vconfig;
> +		goto err_put_vdev;
>  	dev_set_drvdata(&mdev->dev, mdev_state);
>  	return 0;
>  
> -err_vconfig:
> -	kfree(mdev_state->vconfig);
> -err_state:
> -	vfio_uninit_group_dev(&mdev_state->vdev);
> -	kfree(mdev_state);
> -err_nr_ports:
> -	atomic_add(nr_ports, &mdev_avail_ports);
> +err_put_vdev:
> +	vfio_put_device(&mdev_state->vdev);
>  	return ret;
>  }
>  
> +static void mtty_release_dev(struct vfio_device *vdev)
> +{
> +	struct mdev_state *mdev_state =
> +		container_of(vdev, struct mdev_state, vdev);
> +
> +	kfree(mdev_state->vconfig);
> +	vfio_free_device(vdev);
> +	atomic_add(mdev_state->nr_ports, &mdev_avail_ports);

I must be missing something, isn't this a use-after-free?

mdev_state is allocated via vfio_alloc_device(), where vdev is the
first entry in that structure, so this is equivalent to
kvfree(mdev_state).  mbochs has the same issue.  mdpy and vfio-ap
adjust global counters after vfio_free_device(), which I think muddies
the situation.  Shouldn't we look suspiciously at any .release callback
where vfio_free_device() isn't the last thing executed?  Thanks,

Alex