netdev - Re: [PATCH] 9p/trans_virtio: fix hot-unplug

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Thu, 12 Mar 2015 11:54:10 +1030
From:	Rusty Russell <rusty@...tcorp.com.au>
To:	"Michael S. Tsirkin" <mst@...hat.com>, linux-kernel@...r.kernel.org
Cc:	virtualization@...ts.linux-foundation.org,
	Eric Van Hensbergen <ericvh@...il.com>,
	Ron Minnich <rminnich@...dia.gov>,
	Latchesar Ionkov <lucho@...kov.net>,
	"David S. Miller" <davem@...emloft.net>,
	v9fs-developer@...ts.sourceforge.net, netdev@...r.kernel.org
Subject: Re: [PATCH] 9p/trans_virtio: fix hot-unplug

"Michael S. Tsirkin" <mst@...hat.com> writes:
> On device hot-unplug, 9p/virtio currently will kfree channel while
> it might still be in use.
>
> Of course, it might stay used forever, so it's an extremely ugly hack,
> but it seems better than use-after-free that we have now.
>
> Signed-off-by: Michael S. Tsirkin <mst@...hat.com>

I'll apply it, but it looks like a bandaid.

The right answer would seem to be:
1) Use reference counting: 1 for client, 1 for transport.
2) When hot unplug, complete (ie. fail) all outstanding requests.
3) From then on, fail all incoming requests.
4) When refcount hits 0, free the structure.

Thanks,
Rusty.



> ---
>  net/9p/trans_virtio.c | 25 +++++++++++++++++++++----
>  1 file changed, 21 insertions(+), 4 deletions(-)
>
> diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c
> index d8e376a..d1b2f306 100644
> --- a/net/9p/trans_virtio.c
> +++ b/net/9p/trans_virtio.c
> @@ -658,14 +658,31 @@ p9_virtio_create(struct p9_client *client, const char *devname, char *args)
>  static void p9_virtio_remove(struct virtio_device *vdev)
>  {
>  	struct virtio_chan *chan = vdev->priv;
> -
> -	if (chan->inuse)
> -		p9_virtio_close(chan->client);
> -	vdev->config->del_vqs(vdev);
> +	unsigned long warning_time;
> +	bool inuse;
>  
>  	mutex_lock(&virtio_9p_lock);
> +
> +	/* Remove self from list so we don't get new users. */
>  	list_del(&chan->chan_list);
> +	warning_time = jiffies;
> +
> +	/* Wait for existing users to close. */
> +	while (chan->inuse) {
> +		mutex_unlock(&virtio_9p_lock);
> +		msleep(250);
> +               if (time_after(jiffies, warning_time + 10 * HZ)) {
> +			dev_emerg(&vdev->dev, "p9_virtio_remove: "
> +				  "waiting for device in use.\n");
> +			warning_time = jiffies;
> +               }
> +		mutex_lock(&virtio_9p_lock);
> +	}
> +
>  	mutex_unlock(&virtio_9p_lock);
> +
> +	vdev->config->del_vqs(vdev);
> +
>  	sysfs_remove_file(&(vdev->dev.kobj), &dev_attr_mount_tag.attr);
>  	kobject_uevent(&(vdev->dev.kobj), KOBJ_CHANGE);
>  	kfree(chan->tag);
> -- 
> MST
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html