Need a second set of eyeballs for a possible startup race condition in vc04_services/vchiq.

Greg KH greg at
Tue Jan 31 00:25:16 PST 2017

On Mon, Jan 30, 2017 at 01:17:55PM -0800, Michael Zoran wrote:
> On Tue, 2017-01-31 at 00:01 +0300, Dan Carpenter wrote:
> > It's hard to review this because there are no callers and the hash
> > you're talking about is an RPI hash...  You have no idea how lazy I
> > am.
> > 
> > You're right, that code looks racy but it's almost certainly harmless
> > depending on how it's called.  It has to race at the very very
> > beginning
> > otherwise it's fine.
> I wonder if out of caution it would make sense to start stripping out
> some of these entry points that don't have an obvious caller in the
> kernel tree?

If you aren't going to need to add them back soon for another one of
these drivers, yes.  If they are going to be needed, I suggest adding
the driver now so we know what should, and should not, be kept.


greg k-h

More information about the linux-rpi-kernel mailing list