Re: kobject_register failed for intelfb (-EACCES) (Re: 2.6.12-rc4-mm1)

Re: kobject_register failed for intelfb (-EACCES) (Re: 2.6.12-rc4-mm1)

[Andrew Morton]

Greg KH <greg@kroah.com> wrote:
>
> 
>  On Thu, May 12, 2005 at 04:58:01PM +0400, Alexey Dobriyan wrote:
>  > kobject Intel(R) 830M/845G/852GM/855GM/865G/915G Framebuffer Driver:
>  > registering. parent: <NULL>, set: drivers
>  > kobject_register failed for Intel(R) 830M/845G/852GM/855GM/865G/915G
> 
>  Someone tried to put a "/" in a kobject name, which is not allowed.
>  Actually the name seems to be set to:
>  	"Intel(R) 830M/845G/852GM/855GM/865G/915G Framebuffer Driver"
>  which is a bit verbous if you want to create a directory name :)

I don't think that part of the driver has changed in some time.  Is there
something new in your trees which would trigger this?

Seems like a fix such as this will be needed:

--- 25/drivers/video/intelfb/intelfbdrv.c~intelfbdrv-naming-fix	2005-05-12 08:54:46.000000000 -0700
+++ 25-akpm/drivers/video/intelfb/intelfbdrv.c	2005-05-12 08:55:03.000000000 -0700
@@ -214,7 +214,7 @@ static struct fb_ops intel_fb_ops = {
 
 /* PCI driver module table */
 static struct pci_driver intelfb_driver = {
-	.name =		"Intel(R) " SUPPORTED_CHIPSETS " Framebuffer Driver",
+	.name =		"intelfb",
 	.id_table =	intelfb_pci_table,
 	.probe =	intelfb_pci_register,
 	.remove =	__devexit_p(intelfb_pci_unregister)
_

Re: kobject_register failed for intelfb (-EACCES) (Re: 2.6.12-rc4-mm1)

[Greg KH]

On Thu, May 12, 2005 at 08:59:33AM -0700, Andrew Morton wrote:
> Greg KH <greg@kroah.com> wrote:
> >
> > 
> >  On Thu, May 12, 2005 at 04:58:01PM +0400, Alexey Dobriyan wrote:
> >  > kobject Intel(R) 830M/845G/852GM/855GM/865G/915G Framebuffer Driver:
> >  > registering. parent: <NULL>, set: drivers
> >  > kobject_register failed for Intel(R) 830M/845G/852GM/855GM/865G/915G
> > 
> >  Someone tried to put a "/" in a kobject name, which is not allowed.
> >  Actually the name seems to be set to:
> >  	"Intel(R) 830M/845G/852GM/855GM/865G/915G Framebuffer Driver"
> >  which is a bit verbous if you want to create a directory name :)
> 
> I don't think that part of the driver has changed in some time.  Is there
> something new in your trees which would trigger this?

No, not that I know of.

> Seems like a fix such as this will be needed:
> 
> --- 25/drivers/video/intelfb/intelfbdrv.c~intelfbdrv-naming-fix	2005-05-12 08:54:46.000000000 -0700
> +++ 25-akpm/drivers/video/intelfb/intelfbdrv.c	2005-05-12 08:55:03.000000000 -0700
> @@ -214,7 +214,7 @@ static struct fb_ops intel_fb_ops = {
>  
>  /* PCI driver module table */
>  static struct pci_driver intelfb_driver = {
> -	.name =		"Intel(R) " SUPPORTED_CHIPSETS " Framebuffer Driver",
> +	.name =		"intelfb",

Did the SUPPORTED_CHIPSETS macro change somehow?

Anyway, the patch looks correct to me.

thanks,

greg k-h

Re: kobject_register failed for intelfb (-EACCES) (Re: 2.6.12-rc4-mm1)

[Alexey Dobriyan]

On Thursday 12 May 2005 19:59, Andrew Morton wrote:
> Greg KH <greg@kroah.com> wrote:
> >  On Thu, May 12, 2005 at 04:58:01PM +0400, Alexey Dobriyan wrote:
> >  > kobject Intel(R) 830M/845G/852GM/855GM/865G/915G Framebuffer Driver:
> >  > registering. parent: <NULL>, set: drivers
> >  > kobject_register failed for Intel(R) 830M/845G/852GM/855GM/865G/915G
> > 
> >  Someone tried to put a "/" in a kobject name, which is not allowed.
> >  Actually the name seems to be set to:
> >  	"Intel(R) 830M/845G/852GM/855GM/865G/915G Framebuffer Driver"
> >  which is a bit verbous if you want to create a directory name :)

> Seems like a fix such as this will be needed:

> -	.name =		"Intel(R) " SUPPORTED_CHIPSETS " Framebuffer Driver",
> +	.name =		"intelfb",

It works.