Re: [PATCH] [drivers/video/fb] Eliminate recursive mutex in compatfb ioctl path

[Geert Uytterhoeven]

On Sun, 2 Nov 2008, Keith Packard wrote:
> fb_compat_ioctl grabbed the fb mutex and then called fb_ioctl, which also
> grabs the same mutex. I traced through all existing fb_compat_ioctl
> functions in the various fb drivers and found that they also transform their
> arguments and call the native fb_ioctl function. So, it is correct to just
> remove the mutex_lock/mutex_unlock pair in the top-level compat path.

Last week I submitted an alternative fix: http://lkml.org/lkml/2008/10/31/162

Probably Rafael wants to keep track of this regression, too?

> Signed-off-by: Keith Packard <keithp@xxxxxxxxxx>
> ---
>  drivers/video/fbmem.c |    2 --
>  1 files changed, 0 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
> index cd5f20d..548cf26 100644
> --- a/drivers/video/fbmem.c
> +++ b/drivers/video/fbmem.c
> @@ -1253,7 +1253,6 @@ fb_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
>  	struct fb_ops *fb = info->fbops;
>  	long ret = -ENOIOCTLCMD;
>  
> -	mutex_lock(&info->lock);
>  	switch(cmd) {
>  	case FBIOGET_VSCREENINFO:
>  	case FBIOPUT_VSCREENINFO:
> @@ -1279,7 +1278,6 @@ fb_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
>  			ret = fb->fb_compat_ioctl(info, cmd, arg);
>  		break;
>  	}
> -	mutex_unlock(&info->lock);
>  	return ret;
>  }
>  #endif

With kind regards,

Geert Uytterhoeven
Software Architect

Sony Techsoft Centre Europe
The Corporate Village  Da Vincilaan 7-D1  B-1935 Zaventem  Belgium

Phone:    +32 (0)2 700 8453
Fax:      +32 (0)2 700 8622
E-mail:   Geert.Uytterhoeven@xxxxxxxxxxx
Internet: http://www.sony-europe.com/

A division of Sony Europe (Belgium) N.V.
VAT BE 0413.825.160 Â RPR Brussels
Fortis  BIC GEBABEBB  IBAN BE41293037680010