Re: [PATCH[[2.5][3-11] update dvb subsystem core

[Christoph Hellwig]

On Wed, May 07, 2003 at 11:19:30AM +0200, Michael Hunold wrote:
> The code does not behave differently. If DVB_DEVFS_ONLY is set, then the 
> old chardev register interface is omitted.

Which is a different behaviour, now you can't create a device node
manually anymor.  Also note that the feature you rely on here (devfs
presetting file->private_data) will go away in the next round of patches,
see Al Viro's patchit for a generic replacement that works with or
without devfs.

> > Just removing this makes the code even more incorrect.  You need to
> > add a ->owner member and call try_module_get on it before calling into
> > the module (and handle the return value..)
> 
> There is a functional dependency between the dvb-core and the actual dvb 
> driver. So there is no need to increase the module count of the dvb-core 
> if a new adapter is registered IMHO, because you cannot unload the 
> dvb-core before the driver anyway.

Okay, you're right I should have read more of the code to get the global
picture.  You still wan't an owner field for at least struct dvb_device
device, though - but the try_module_get must go into dvb_generic_open
and maybe in more other places where you use the "backend" modules.