| 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: <201010151202.31629.hverkuil@xs4all.nl>
Date: Fri, 15 Oct 2010 12:02:31 +0200
From: Hans Verkuil <hverkuil@...all.nl>
To: Andrew Morton <akpm@...ux-foundation.org>
Cc: Valdis.Kletnieks@...edu,
Mauro Carvalho Chehab <mchehab@...hat.com>,
linux-kernel@...r.kernel.org, linux-media@...r.kernel.org
Subject: Re: mmotm 2010-10-13 - GSPCA SPCA561 webcam driver broken
On Friday, October 15, 2010 11:05:26 Andrew Morton wrote:
> On Fri, 15 Oct 2010 10:45:45 +0200 Hans Verkuil <hverkuil@...all.nl> wrote:
>
> > On Thursday, October 14, 2010 22:06:29 Valdis.Kletnieks@...edu wrote:
> > > On Wed, 13 Oct 2010 17:13:25 PDT, akpm@...ux-foundation.org said:
> > > > The mm-of-the-moment snapshot 2010-10-13-17-13 has been uploaded to
> > > >
> > > > http://userweb.kernel.org/~akpm/mmotm/
> > >
> > > This broke my webcam. I bisected it down to this commit, and things
> > > work again after reverting the 2 code lines of change.
> > >
> > > commit 9e4d79a98ebd857ec729f5fa8f432f35def4d0da
> > > Author: Hans Verkuil <hverkuil@...all.nl>
> > > Date: Sun Sep 26 08:16:56 2010 -0300
> > >
> > > V4L/DVB: v4l2-dev: after a disconnect any ioctl call will be blocked
> > >
> > > Until now all fops except release and (unlocked_)ioctl returned an error
> > > after the device node was unregistered. Extend this as well to the ioctl
> > > fops. There is nothing useful that an application can do here and it
> > > complicates the driver code unnecessarily.
> > >
> > > Signed-off-by: Hans Verkuil <hverkuil@...all.nl>
> > > Signed-off-by: Mauro Carvalho Chehab <mchehab@...hat.com>
> > >
> > >
> > > diff --git a/drivers/media/video/v4l2-dev.c b/drivers/media/video/v4l2-dev.c
> > > index d4a3532..f069c61 100644
> > > --- a/drivers/media/video/v4l2-dev.c
> > > +++ b/drivers/media/video/v4l2-dev.c
> > > @@ -221,8 +221,8 @@ static long v4l2_ioctl(struct file *filp, unsigned int cmd,
> > > struct video_device *vdev = video_devdata(filp);
> > > int ret;
> > >
> > > - /* Allow ioctl to continue even if the device was unregistered.
> > > - Things like dequeueing buffers might still be useful. */
> > > + if (!vdev->fops->ioctl)
> > > + return -ENOTTY;
> > > if (vdev->fops->unlocked_ioctl) {
> > > ret = vdev->fops->unlocked_ioctl(filp, cmd, arg);
> > > } else if (vdev->fops->ioctl) {
> > >
> > > I suspect this doesn't do what's intended if a driver is using ->unlocked_ioctl
> > > rather than ->ioctl, and it should be reverted - it only saves at most one
> > > if statement.
> > >
> > >
> >
> > I'm not sure what is going on here. It looks like this patch is mangled in your
> > tree since the same patch in the v4l-dvb repository looks like this:
> >
> > diff --git a/drivers/media/video/v4l2-dev.c b/drivers/media/video/v4l2-dev.c
> > index 32575a6..26d39c4 100644
> > --- a/drivers/media/video/v4l2-dev.c
> > +++ b/drivers/media/video/v4l2-dev.c
> > @@ -222,8 +222,8 @@ static int v4l2_ioctl(struct inode *inode, struct file *filp,
> >
> > if (!vdev->fops->ioctl)
> > return -ENOTTY;
> > - /* Allow ioctl to continue even if the device was unregistered.
> > - Things like dequeueing buffers might still be useful. */
> > + if (!video_is_registered(vdev))
> > + return -ENODEV;
> > return vdev->fops->ioctl(filp, cmd, arg);
> > }
> >
> > @@ -234,8 +234,8 @@ static long v4l2_unlocked_ioctl(struct file *filp,
> >
> > if (!vdev->fops->unlocked_ioctl)
> > return -ENOTTY;
> > - /* Allow ioctl to continue even if the device was unregistered.
> > - Things like dequeueing buffers might still be useful. */
> > + if (!video_is_registered(vdev))
> > + return -ENODEV;
> > return vdev->fops->unlocked_ioctl(filp, cmd, arg);
> > }
> >
> > In your diff there is a mismatch between ioctl and unlocked_ioctl which no doubt
> > is causing all the problems for you.
>
> The patch which Valdis quoted is what is in linux-next. I'm not
> at which stage the mangling happened?
>
OK, I see what happened. My original patch:
http://git.linuxtv.org/media_tree.git?a=commitdiff;h=30de1fa062e42a74c54e94c8977d7dcef9a5049f;hp=f39385558f50f0b5b2bc9b47c187b81a8188fb10
clashed with Arnd Bergmann's patch:
http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=commit;h=86a5ef7d777cdd61dfe82379d559dbea069aea3d
Someone tried to resolve the conflict but made a mistake.
The result is this:
static long v4l2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
struct video_device *vdev = video_devdata(filp);
int ret;
if (!vdev->fops->ioctl)
return -ENOTTY;
if (vdev->fops->unlocked_ioctl) {
if (vdev->lock)
mutex_lock(vdev->lock);
ret = vdev->fops->unlocked_ioctl(filp, cmd, arg);
if (vdev->lock)
mutex_unlock(vdev->lock);
} else if (vdev->fops->ioctl) {
/* TODO: convert all drivers to unlocked_ioctl */
lock_kernel();
ret = vdev->fops->ioctl(filp, cmd, arg);
unlock_kernel();
} else
ret = -ENOTTY;
return ret;
}
But this is wrong.
What it should be is this:
static long v4l2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
struct video_device *vdev = video_devdata(filp);
int ret = -ENODEV;
if (vdev->fops->unlocked_ioctl) {
if (vdev->lock)
mutex_lock(vdev->lock);
if (video_is_registered(vdev))
ret = vdev->fops->unlocked_ioctl(filp, cmd, arg);
if (vdev->lock)
mutex_unlock(vdev->lock);
} else if (vdev->fops->ioctl) {
/* TODO: convert all drivers to unlocked_ioctl */
lock_kernel();
if (video_is_registered(vdev))
ret = vdev->fops->ioctl(filp, cmd, arg);
unlock_kernel();
} else
ret = -ENOTTY;
return ret;
}
Mauro, is this something for you to fix?
Regards,
Hans
--
Hans Verkuil - video4linux developer - sponsored by TANDBERG, part of Cisco
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists