linux-kernel - Re: [PATCH v3] media: usb: uvc: use vb2_ops_wait

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 for Android: free password hash cracker in your pocket

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

Message-ID: <1532665.63kW7uD9fo@avalon>
Date:	Sat, 29 Nov 2014 20:11:41 +0200
From:	Laurent Pinchart <laurent.pinchart@...asonboard.com>
To:	Prabhakar Lad <prabhakar.csengg@...il.com>
Cc:	LMML <linux-media@...r.kernel.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Hans Verkuil <hverkuil@...all.nl>
Subject: Re: [PATCH v3] media: usb: uvc: use vb2_ops_wait_prepare/finish helper

Hi Prabhakar,

On Friday 28 November 2014 17:06:52 Prabhakar Lad wrote:
> On Thu, Nov 27, 2014 at 9:32 PM, Laurent Pinchart wrote:
> > Hi Prabhakar,
> 
> [Snip]
> 
> >> +     queue->queue.lock = &queue->mutex;
> > 
> > I'm a bit concerned that this would introduce future breakages. Setting
> > the queue.lock pointer enables locking in all vb2_fop_* and vb2_ops_wait_*
> > functions. The uvcvideo driver isn't ready for that, but doesn't use the
> > vb2_fop_* functions yet, so that's not an issue. However, in the future,
> > videobuf2 might use the lock in more places, including functions used by
> > the uvcvideo driver. This could then cause breakages.
> 
> Even if in future if videobuf2 uses this lock it would be in helpers mostly,
> so any way it doesn’t harm :)

My concern is about vb2 starting using the lock in existing helpers used by 
the uvcvideo driver. I suppose we can deal with it later.

> > It would be better to completely convert the uvcvideo driver to the
> > vb2_fop_* functions if we want to use vb2_ops_*. I'm not sure how complex
> > that would be though, and whether it would be possible while still
> > keeping the fine-grained locking implemented by the uvcvideo driver. Do
> > you think it should be attempted ?
> 
> mmap & poll should be fairly simple, looks like open & release cannot be
> dropped as it does some usb_autopm_get/put_interface() calls which I am not
> aware of.

I've looked at that, and there's a race condition in vb2_fop_poll() (for which 
I've already sent a patch) and possible in vb2_mmap() (raised the issue on 
#v4l today) as well that need to be fixed first.

Anyway, for this patch

Acked-by: Laurent Pinchart <laurent.pinchart@...asonboard.com>

Have you tested it by the way ?

Should I take it in my tree or will you send a pull request for the whole 
series ?

> >>       ret = vb2_queue_init(&queue->queue);
> >>       if (ret)
> >>               return ret;

-- 
Regards,

Laurent Pinchart

--
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/