[PATCH] UBUNTU: SAUCE: Don't register vga16fb framebuffer if other framebuffers are present

[Amit Kucheria]

On 10 Mar 25, Chase Douglas wrote:
> Using the vga16fb framebuffer is not safe when other framebuffers are
> present. This fixes the case where the vga16fb module is loaded after
> a better framebuffer is loaded (since vga16fb is by definition the
> worst-case framebuffer).
> 
> There does not appear to be any locking around the num_registered_fb, so
> this is a hack at best. However, in Lucid we build vga16fb as a module.
> Modules are loaded serially, so this should be ok for Lucid. In M, we
> will transition to efifb and drop vga16fb, so this is a one time hack.
> 
> This prevents sudo lshw from corrupting /dev/fb0 by writing to vga16fb
> through /dev/fb1.
> 
> BugLink: http://bugs.launchpad.net/bugs/527369
> 
> Signed-off-by: Chase Douglas <chase.douglas at canonical.com>
> ---
>  drivers/video/vga16fb.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/video/vga16fb.c b/drivers/video/vga16fb.c
> index efde41d..b1d62ef 100644
> --- a/drivers/video/vga16fb.c
> +++ b/drivers/video/vga16fb.c
> @@ -1355,7 +1355,7 @@ static int __init vga16fb_probe(struct platform_device *dev)
>  
>  	vga16fb_update_fix(info);
>  
> -	if (register_framebuffer(info) < 0) {
> +	if (num_registered_fb > 0 || register_framebuffer(info) < 0) {
>  		printk(KERN_ERR "vga16fb: unable to register framebuffer\n");
>  		ret = -EINVAL;
>  		goto err_check_var;
> -- 
> 1.7.0

We should differentiate the case when there was a real error from the one
where we bailed out on purpose. I think the extra cost to patch size would be
justified.

Regards,
Amit

-- 
----------------------------------------------------------------------
Amit Kucheria, Kernel Engineer || amit.kucheria at canonical.com
----------------------------------------------------------------------