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  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:	Sun, 23 Dec 2012 18:37:43 -0200
From:	Mauro Carvalho Chehab <mchehab@...hat.com>
To:	Jörg Otte <jrg.otte@...il.com>
Cc:	linux-kernel@...r.kernel.org, linux-media@...r.kernel.org,
	Laurent Pinchart <laurent.pinchart@...asonboard.com>,
	Linus Torvalds <torvalds@...ux-foundation.org>
Subject: Re: [v3.8-rc1] Multimedia regression, ioctl(17,..)-API changed ?

Hi Jörg,

Em Sun, 23 Dec 2012 17:46:07 +0100
Jörg Otte <jrg.otte@...il.com> escreveu:

> With kernel v3.8 all multimedia programs under KDE4 don't work (Kubuntu 12.04).
> They alltogether ( at least Dragonplayer (Mediaplayer), Knotify4
> (system-sound),
> System-Settings-Multimedia,..) are looping forever producing 100% CPU-usage
> and must be killed.
> 
> With kernel 3.7 there are no problems.

Do you have any other non-uvc device to test?

> I compared an strace of Dragonplayer under 3.7 and 3.8 kernels. The
> main difference
> of both traces are the following corresponding outputs just before
> looping in v3.8
> begins:
> 
> v3.7:
> ioctl(17, VIDIOC_ENUMSTD, 0x7fff6cce66a0) = -1 EINVAL (Invalid argument)

This ioctl returns -ENOTTY already with other media drivers on v3.7.

> ioctl(17, VIDIOC_QUERYCTRL, 0x7fff6cce66f0) = -1 EINVAL (Invalid argument)

This change is new, and, AFAIKT, only UVC returns -ENOENT on 3.8-rc1.

This is likely the source of the troubles.

> 
> v3.8:
> ioctl(17, VIDIOC_ENUMSTD, 0x7fffc3be6990) = -1 ENOTTY (Inappropriate
> ioctl for device)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> 
> So error number EINVAL was changed to ENOTTY/ENOENT
> 
> When Dragonplayer under v3.8 comes to ioctl(17, VIDIOC_QUERYCTRL,...)
> and sees error
> number ENOENT instead of EINVAL it loops forever producing 100% CPU
> usage like so:
> 
>   .
>   .
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)
> ioctl(17, VIDIOC_QUERYCTRL, 0x7fffc3be69e0) = -1 ENOENT (No such file
> or directory)

Yeah, it started an endless loop here, likely because kde4 seems to be
expecting either 0 or -EINVAL error code for VIDIOC_QUERYCTRL.

It should be noticed that there are other valid error codes here. For
example, if this ioctl is not implemented, -ENOTTY may also be returned.
Fortunately, almost all drivers do implement this ioctl.

The expected return error codes for this ioctl are described at:
	http://linuxtv.org/downloads/v4l-dvb-apis/vidioc-queryctrl.html

In practice, except for the uvc driver, the current return codes are
EINVAL/EACCES/ENOTTY.

>   .
>   and so on
>   .
> 
> For me it looks like that KDE4 multimedia is not aware of the new error numbers.
> 
> Looking through the commits I found driver uvcvideo producing the changed
> error numbers.
> 
> commit f0ed2ce840b3a59b587e8aa398538141a86e9588
> Author: Laurent Pinchart <laurent.pinchart@...asonboard.com>
> [media] uvcvideo: Set error_idx properly for extended controls API failures
> 
> To verify this I built a v3.8-kernel without uvcvideo (USB_VIDEO_CLASS=n)
> and the problem disappeared!
> 
> Simply reverting the commit is not an option for me because then I am left
> with merge conflicts and I don't know how to resolve.
> 
> Unfortunately without uvcvideo I lost my usb-camera support.

Rafael made a patch fixing it, and Linus should be applying it.

Regards,
Mauro
--
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