Hi Andi,

> > I check this. Maybe we have more of them. What do you propose as max
> > size value for kmalloc? 2*PAGE_SIZE or 4*PAGE_SIZE?
> 
> What better fits the intended use case. I don't know how many objects are expected
> here. Smaller is better probably.

I now looked carefully through your patch and changed and added some
parts to better fit into. I also fixed another RFCOMM refcount bug.
Please review it, before I send it to Dave.

> If you want to handle more objects this way you should use seq_file instead.

The general plan is to move over to sysfs.

Regards

Marcel