Re: [PATCH V2] fbcon -- fix race between open and removal of framebuffers

[Jack Stone <jwjstone@fastmail.fm>]

On 10/05/2011 22:06, Jack Stone wrote:
> Hi Tim,
> 
> One more quick question:
> 
> On 10/05/2011 13:47, Tim Gardner wrote:
> +static struct fb_info *get_framebuffer_info(int idx)
> +__acquires(&registered_lock)
> +__releases(&registered_lock)
> +{
> +	struct fb_info *fb_info;
> +
> +	spin_lock(&registered_lock);
> +	fb_info = registered_fb[idx];
> +	fb_info->ref_count++;
> +	spin_unlock(&registered_lock);
> +
> +	return fb_info;
> +}
> 
> [snip]
> 
>  static int
>  fb_open(struct inode *inode, struct file *file)
>  __acquires(&info->lock)
> @@ -1363,13 +1421,18 @@ __releases(&info->lock)
> 
>  	if (fbidx >= FB_MAX)
>  		return -ENODEV;
> -	info = registered_fb[fbidx];
> -	if (!info)
> +	info = get_framebuffer_info(fbidx);
> +	if (!info) {
>  		request_module("fb%d", fbidx);
> -	info = registered_fb[fbidx];
> +		info = get_framebuffer_info(fbidx);
> +	}
>  	if (!info)
>  		return -ENODEV;
> 
> This section of code implies that get_framebuffer_info can return NULL
> but in that case wouldn't the fb_info->ref_count++ have oopsed?
> 
> You could add the simple case of
> 
> if(fb_info)
> 	fb_info->ref_count++
> 
> to get_framebuffer_info. That should cover it.
> 

Just read your later patch. Sorry for the extra email.

Thanks,

Jack