On Mon, Apr 14, 2008 at 08:04:09PM +0200, Peter Zijlstra wrote:
quoted text
> On Mon, 2008-04-14 at 09:19 -0700, mark gross wrote:
> > I am working with one of the iwl4965 developers to add pm_qos based
> > power management to its power states.  We relalized that the block
> > notifier I used in the PM_QOS has no way of passing in any driver device
> > instance data.  In this case the developer expected the notification
> > call backs to call the iwl4965 notification function with an instance
> > pointer to the correct *dev.
> > 
> > Poking around I've noticed a handful of drivers using notifications that
> > seem to keep a list of instance pointers around so it can plug into the
> > notification infrastructure.  including : ipmi_msghandler.c adb_hid.c
> > md.c ips.c
> > 
> > As having a registered call back called with a private data pointer set
> > up at registration time  is such a common idiom I thought it might be a
> > good thing to add a private_data pointer to the struct notifier_block
> > and add the interfaces needed to pass the private data as the
> > notification chain is processed.
> 
> How is:
> 
> struct my_device_thingy {
> 	...
> 	struct notifier_block my_notifier;
> 	...
> };
> 
> int my_device_notifier_call(struct notifier_block *nb, unsigned long val
> 			    void *data)
> {
> 	struct my_device_thingy *mdt = container_of(nb, struct
> 				 my_device_thingy, my_notifier);
> 
> 	...
> }
> 
> not sufficient?

Because ... um....a.... that would work too.  

Please ignore my last patch.

Thanks!

--mgross

--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/