| 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
| ||
|
Message-ID: <082ba39a-aa00-ac49-ac2a-eceac0bca5ff@xs4all.nl>
Date: Tue, 6 Sep 2022 15:25:58 +0200
From: Hans Verkuil <hverkuil@...all.nl>
To: Soumya Negi <soumya.negi97@...il.com>,
Mauro Carvalho Chehab <mchehab@...nel.org>,
Shuah Khan <skhan@...uxfoundation.org>
Cc: linux-kernel-mentees@...ts.linuxfoundation.org,
linux-media@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [RFT PATCH] media: em28xx: Enable v4l2 file ops at the end of
em28xx_v4l2_init()
Hi Soumya,
On 29/08/2022 15:25, Soumya Negi wrote:
> Fix syzbot bug KASAN: use-after-free Read in v4l2_fh_open
> Syzbot link:
> https://syzkaller.appspot.com/bug?id=0e3c97f1c4112e102c9988afd5eff079350eab7f
> Repro: https://syzkaller.appspot.com/text?tag=ReproC&x=12663ebdd00000
>
> The bug arises because the em28xx driver registers a v4l2 video
> device(struct video_device) with the V4L2 interface in
> em28xx_v4l2_init() but the v4l2 extension initialization eventually
> fails and the video device is unregistered. v4l2_open() in the V4L2
> interface views the device as registered and allows calls to
> em28xx_v4l2_open(). Due to race between video_unregister_device() and
> v4l2_open(), em28xx_v4l2_open() is accessing video device after it has
> been freed(by the release callback) and is passing it on to
> v4l2_fh_open().
>
> In em28xx_v4l2_init(), temporarily disable v4l2 file operations on
> struct video_device devices(video, vbi, radio) before registering them
> and enable the file ops at the end when v4l2 extension initialization
> succeeds.
That's not the right approach. The problem is the roll-your-own v4l2->ref
refcount. Instead this should use the refcount that is built into struct
v4l2_device.
Specifically v4l2_dev->release should be set to the release function, and
v4l2_device_put(&v4l2->v4l2_dev); should be called in remove() or in the
error path of em28xx_v4l2_init().
Using this the release callback of v4l2_device will be called when the last
open file handle is closed, i.e. it keeps track of the open()s.
The roll-your-own approach in the em28xx driver (written before this v4l2_device
refcounting existed), is not taking that into account, so that causes these
problems.
What is also bad is that em28xx_vb2_setup() is called after the video devices
are registered. That should happen before.
Regards,
Hans
>
> Reported-by: syzbot+b2391895514ed9ef4a8e@...kaller.appspotmail.com
> Signed-off-by: Soumya Negi <soumya.negi97@...il.com>
> ---
> drivers/media/usb/em28xx/em28xx-video.c | 45 +++++++++++++++++++++----
> 1 file changed, 39 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/media/usb/em28xx/em28xx-video.c b/drivers/media/usb/em28xx/em28xx-video.c
> index 8181c0e6a25b..900a1eacb1c7 100644
> --- a/drivers/media/usb/em28xx/em28xx-video.c
> +++ b/drivers/media/usb/em28xx/em28xx-video.c
> @@ -2320,6 +2320,19 @@ static int em28xx_v4l2_close(struct file *filp)
> return 0;
> }
>
> +/* Used to temporarily disable file operations on video_device until successful
> + * initialization in em28xx_v4l2_init().
> + */
> +static const struct v4l2_file_operations em28xx_v4l_fops_empty = {
> + .owner = THIS_MODULE,
> + .open = NULL,
> + .release = NULL,
> + .read = NULL,
> + .poll = NULL,
> + .mmap = NULL,
> + .unlocked_ioctl = NULL,
> +};
> +
> static const struct v4l2_file_operations em28xx_v4l_fops = {
> .owner = THIS_MODULE,
> .open = em28xx_v4l2_open,
> @@ -2375,12 +2388,22 @@ static const struct v4l2_ioctl_ops video_ioctl_ops = {
> };
>
> static const struct video_device em28xx_video_template = {
> - .fops = &em28xx_v4l_fops,
> + .fops = &em28xx_v4l_fops_empty,
> .ioctl_ops = &video_ioctl_ops,
> .release = video_device_release_empty,
> .tvnorms = V4L2_STD_ALL,
> };
>
> +/* Used to temporarily disable file operations on video_device until successful
> + * initialization in em28xx_v4l2_init().
> + */
> +static const struct v4l2_file_operations radio_fops_empty = {
> + .owner = THIS_MODULE,
> + .open = NULL,
> + .release = NULL,
> + .unlocked_ioctl = NULL,
> +};
> +
> static const struct v4l2_file_operations radio_fops = {
> .owner = THIS_MODULE,
> .open = em28xx_v4l2_open,
> @@ -2404,7 +2427,7 @@ static const struct v4l2_ioctl_ops radio_ioctl_ops = {
> };
>
> static struct video_device em28xx_radio_template = {
> - .fops = &radio_fops,
> + .fops = &radio_fops_empty,
> .ioctl_ops = &radio_ioctl_ops,
> .release = video_device_release_empty,
> };
> @@ -2833,9 +2856,6 @@ static int em28xx_v4l2_init(struct em28xx *dev)
> "can't register radio device\n");
> goto unregister_dev;
> }
> - dev_info(&dev->intf->dev,
> - "Registered radio device as %s\n",
> - video_device_node_name(&v4l2->radio_dev));
> }
>
> /* Init entities at the Media Controller */
> @@ -2851,14 +2871,27 @@ static int em28xx_v4l2_init(struct em28xx *dev)
> }
> #endif
>
> + /* Enable v4l2 file operations for v4l2 video video_device */
> + v4l2->vdev.fops = &em28xx_v4l_fops;
> dev_info(&dev->intf->dev,
> "V4L2 video device registered as %s\n",
> video_device_node_name(&v4l2->vdev));
>
> - if (video_is_registered(&v4l2->vbi_dev))
> + if (video_is_registered(&v4l2->vbi_dev)) {
> + /* Enable v4l2 file operations for v4l2 vbi video_device */
> + v4l2->vbi_dev.fops = &em28xx_v4l_fops;
> dev_info(&dev->intf->dev,
> "V4L2 VBI device registered as %s\n",
> video_device_node_name(&v4l2->vbi_dev));
> + }
> +
> + if (video_is_registered(&v4l2->radio_dev)) {
> + /* Enable v4l2 file operations for v4l2 radio video_device */
> + v4l2->radio_dev.fops = &radio_fops;
> + dev_info(&dev->intf->dev,
> + "Registered radio device as %s\n",
> + video_device_node_name(&v4l2->radio_dev));
> + }
>
> /* Save some power by putting tuner to sleep */
> v4l2_device_call_all(&v4l2->v4l2_dev, 0, tuner, standby);
Powered by blists - more mailing lists