tq 2440

[gregkh@suse.de (Greg KH)]

On Thu, Jul 15, 2010 at 06:33:22PM +0100, Russell King - ARM Linux wrote:
> On Thu, Jul 15, 2010 at 10:22:18AM -0700, Greg KH wrote:
> > On Thu, Jul 15, 2010 at 06:10:29PM +0100, Russell King - ARM Linux wrote:
> > > On Thu, Jul 15, 2010 at 09:51:41AM -0700, Greg KH wrote:
> > > > On Thu, Jul 15, 2010 at 05:35:26PM +0100, Russell King - ARM Linux wrote:
> > > > > On Thu, Jul 15, 2010 at 11:16:30PM +0800, Conke Hu wrote:
> > > > > > ever notice to the following kernel log?
> > > > > > "Device 's3c2440-nand' does not have a release() function, it is
> > > > > > broken and must."
> > > > > > release() function should be implemented  in the platform_device.
> > > > > 
> > > > > That's not telling you to provide a release function.  The warning is
> > > > > telling you that a device is being unregistered which doesn't have a
> > > > > release function.
> > > > > 
> > > > > Consider this point - maybe it doesn't have a release function because
> > > > > it's not supposed to be unregistered?
> > > > 
> > > > Heh, well, it is being unregistered, and at that point in time, the
> > > > kernel complains.
> > > 
> > > The unregistration occurs because platform_register_devices() (which is
> > > used by arch code to register a block of platform devices) undoes its
> > > work if one of the devices fails to register.
> > > 
> > > I've long since thought, since I created that function, that this probably
> > > isn't desirable behaviour, and it should continue to register as many
> > > devices as it possibly can.
> > > 
> > > > All kobjects need a release function that actually frees it.  If not,
> > > > that is a logic bug.  Please see the Documentation/kobject.txt file for
> > > > details.
> > > 
> > > How do you free a statically declared platform device?
> > 
> > You never unregister it :)
> 
> Indeed, and one way to do that is to fix the double-registration of
> dm9000.0.

I agree.

> Another way to avoid the other complaints is to remove
> the unregistration in platform_register_devices().

No, because you could have created a platform device with a call to
platform_device_alloc() and then called platform_device_register() and
then later, platform_device_unregister(), right?

thanks,

greg k-h