[PATCH] ppc64: SMU partitions & device-tree

Benjamin Herrenschmidt benh at kernel.crashing.org
Fri Sep 23 16:07:10 EST 2005

> > It does, but I felt like prom_add_property() should be made a bit safer.
> > It's definitely an abuse of the devtree_lock, but knowing how rarely
> > prom_add_property() is to be called, I don't think it's that much of a
> > problem, what do you think ?
> As long as you don't expect us to need to _remove_ properties from
> live nodes, I think it's ok.  Otherwise we'll need something like a
> per-node lock.

Yup, I don't expect to allow any property removal, nor to have this
property addition be more than an occasional matter. I just want to make
100% sure to guard against duplicate additions. In my current SMU
driver, this is protected by a semaphore though, but heh...


More information about the Linuxppc64-dev mailing list