linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chris Clayton <chris2553@googlemail.com>
To: Brandon Philips <brandon@ifup.org>
Cc: Torsten Kaiser <just.for.lkml@googlemail.com>,
	Dave Young <hidave.darkstar@gmail.com>,
	linux-media@vger.kernel.org, linux-kernel@vger.kernel.org,
	Mauro Carvalho Chehab <mchehab@infradead.org>,
	Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Subject: Re: [PATCH] bttv: fix mutex use before init
Date: Wed, 15 Dec 2010 18:44:04 +0000	[thread overview]
Message-ID: <201012151844.04105.chris2553@googlemail.com> (raw)
In-Reply-To: <20101214003024.GA3575@hanuman.home.ifup.org>

On Tuesday 14 December 2010, Brandon Philips wrote:
> On 17:13 Sun 12 Dec 2010, Torsten Kaiser wrote:
> >  * change &fh->cap.vb_lock in bttv_open() AND radio_open() to
> > &btv->init.cap.vb_lock
> >  * add a mutex_init(&btv->init.cap.vb_lock) to the setup of init in
> > bttv_probe()
>
> That seems like a reasonable suggestion. An openSUSE user submitted this
> bug to our tracker too. Here is the patch I am having him test.
>
> Would you mind testing it?
>
> From 456dc0ce36db523c4c0c8a269f4eec43a72de1dc Mon Sep 17 00:00:00 2001
> From: Brandon Philips <bphilips@suse.de>
> Date: Mon, 13 Dec 2010 16:21:55 -0800
> Subject: [PATCH] bttv: fix locking for btv->init
>
> Fix locking for the btv->init by using btv->init.cap.vb_lock and in the
> process fix uninitialized deref introduced in c37db91fd0d.
>
> Signed-off-by: Brandon Philips <bphilips@suse.de>
> ---
>  drivers/media/video/bt8xx/bttv-driver.c |   24 +++++++++++++-----------
>  1 files changed, 13 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/media/video/bt8xx/bttv-driver.c
> b/drivers/media/video/bt8xx/bttv-driver.c index a529619..e656424 100644
> --- a/drivers/media/video/bt8xx/bttv-driver.c
> +++ b/drivers/media/video/bt8xx/bttv-driver.c
> @@ -2391,16 +2391,11 @@ static int setup_window_lock(struct bttv_fh *fh,
> struct bttv *btv, fh->ov.field    = win->field;
>  	fh->ov.setup_ok = 1;
>
> -	/*
> -	 * FIXME: btv is protected by btv->lock mutex, while btv->init
> -	 *	  is protected by fh->cap.vb_lock. This seems to open the
> -	 *	  possibility for some race situations. Maybe the better would
> -	 *	  be to unify those locks or to use another way to store the
> -	 *	  init values that will be consumed by videobuf callbacks
> -	 */
> +	mutex_lock(&btv->init.cap.vb_lock);
>  	btv->init.ov.w.width   = win->w.width;
>  	btv->init.ov.w.height  = win->w.height;
>  	btv->init.ov.field     = win->field;
> +	mutex_unlock(&btv->init.cap.vb_lock);
>
>  	/* update overlay if needed */
>  	retval = 0;
> @@ -2620,9 +2615,11 @@ static int bttv_s_fmt_vid_cap(struct file *file,
> void *priv, fh->cap.last         = V4L2_FIELD_NONE;
>  	fh->width            = f->fmt.pix.width;
>  	fh->height           = f->fmt.pix.height;
> +	mutex_lock(&btv->init.cap.vb_lock);
>  	btv->init.fmt        = fmt;
>  	btv->init.width      = f->fmt.pix.width;
>  	btv->init.height     = f->fmt.pix.height;
> +	mutex_unlock(&btv->init.cap.vb_lock);
>  	mutex_unlock(&fh->cap.vb_lock);
>
>  	return 0;
> @@ -2855,6 +2852,7 @@ static int bttv_s_fbuf(struct file *file, void *f,
>
>  	retval = 0;
>  	fh->ovfmt = fmt;
> +	mutex_lock(&btv->init.cap.vb_lock);
>  	btv->init.ovfmt = fmt;
>  	if (fb->flags & V4L2_FBUF_FLAG_OVERLAY) {
>  		fh->ov.w.left   = 0;
> @@ -2876,6 +2874,7 @@ static int bttv_s_fbuf(struct file *file, void *f,
>  			retval = bttv_switch_overlay(btv, fh, new);
>  		}
>  	}
> +	mutex_unlock(&btv->init.cap.vb_lock);
>  	mutex_unlock(&fh->cap.vb_lock);
>  	return retval;
>  }
> @@ -3141,6 +3140,7 @@ static int bttv_s_crop(struct file *file, void *f,
> struct v4l2_crop *crop) fh->do_crop = 1;
>
>  	mutex_lock(&fh->cap.vb_lock);
> +	mutex_lock(&btv->init.cap.vb_lock);
>
>  	if (fh->width < c.min_scaled_width) {
>  		fh->width = c.min_scaled_width;
> @@ -3158,6 +3158,7 @@ static int bttv_s_crop(struct file *file, void *f,
> struct v4l2_crop *crop) btv->init.height = c.max_scaled_height;
>  	}
>
> +	mutex_unlock(&btv->init.cap.vb_lock);
>  	mutex_unlock(&fh->cap.vb_lock);
>
>  	return 0;
> @@ -3302,9 +3303,9 @@ static int bttv_open(struct file *file)
>  	 * Let's first copy btv->init at fh, holding cap.vb_lock, and then work
>  	 * with the rest of init, holding btv->lock.
>  	 */
> -	mutex_lock(&fh->cap.vb_lock);
> +	mutex_lock(&btv->init.cap.vb_lock);
>  	*fh = btv->init;
> -	mutex_unlock(&fh->cap.vb_lock);
> +	mutex_unlock(&btv->init.cap.vb_lock);
>
>  	fh->type = type;
>  	fh->ov.setup_ok = 0;
> @@ -3502,9 +3503,9 @@ static int radio_open(struct file *file)
>  	if (unlikely(!fh))
>  		return -ENOMEM;
>  	file->private_data = fh;
> -	mutex_lock(&fh->cap.vb_lock);
> +	mutex_lock(&btv->init.cap.vb_lock);
>  	*fh = btv->init;
> -	mutex_unlock(&fh->cap.vb_lock);
> +	mutex_unlock(&btv->init.cap.vb_lock);
>
>  	mutex_lock(&btv->lock);
>  	v4l2_prio_open(&btv->prio, &fh->prio);
> @@ -4489,6 +4490,7 @@ static int __devinit bttv_probe(struct pci_dev *dev,
>  	btv->opt_coring     = coring;
>
>  	/* fill struct bttv with some useful defaults */
> +	mutex_init(&btv->init.cap.vb_lock);
>  	btv->init.btv         = btv;
>  	btv->init.ov.w.width  = 320;
>  	btv->init.ov.w.height = 240;

The patch is good here too. Thanks.

Tested-by: Chris Clayton <chris2553@googlemail.com>



-- 
The more I see, the more I know. The more I know, the less I understand. 
Changing Man - Paul Weller

  parent reply	other threads:[~2010-12-15 18:44 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-12-12 13:15 [PATCH] bttv: fix mutex use before init Dave Young
2010-12-12 16:13 ` Torsten Kaiser
2010-12-13 14:04   ` Dave Young
2010-12-13 19:06     ` Torsten Kaiser
2010-12-14  0:30   ` Brandon Philips
2010-12-14 12:05     ` Dave Young
2010-12-14 20:56     ` Torsten Kaiser
2010-12-14 21:13       ` Torsten Kaiser
2010-12-14 21:48         ` Brandon Philips
2010-12-14 21:43       ` Brandon Philips
2010-12-15  2:42         ` Dave Young
2010-12-15  6:47         ` Torsten Kaiser
2010-12-15 18:44     ` Chris Clayton [this message]
2010-12-15 21:45       ` Mauro Carvalho Chehab
2010-12-16 17:26         ` Chris Clayton
2010-12-17 14:05         ` Torsten Kaiser
2010-12-17 16:07         ` Brandon Philips
2010-12-17 20:11           ` Mauro Carvalho Chehab

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=201012151844.04105.chris2553@googlemail.com \
    --to=chris2553@googlemail.com \
    --cc=brandon@ifup.org \
    --cc=g.liakhovetski@gmx.de \
    --cc=hidave.darkstar@gmail.com \
    --cc=just.for.lkml@googlemail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab@infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).