On Tue, Mar 26, 2019 at 10:26:21AM +0200, Heikki Krogerus wrote:
> On Tue, Mar 26, 2019 at 01:46:40PM +0900, Greg KH wrote:
> > On Tue, Mar 19, 2019 at 11:03:10AM +0200, Heikki Krogerus wrote:
> > > On Fri, Mar 15, 2019 at 03:43:21PM -0700, anil.kumar.k@xxxxxxxxx wrote:
> > > > From: Anil Kumar <anil.kumar.k@xxxxxxxxx>
> > > >
> > > > typec_register_port function used variable port->id without
> > > > initializing it. This patch fixes the issue by using the local
> > > > instance for the error case
> > > >
> > > > Issue is not observed in the mainline tree (5.x)
> >
> > Why is this not observed there? What patch fixed this? Why can we not
> > backport that fix?
>
> Commit 23481121c81d (accidentally) fixes the problem, but you can't
> backport that alone. You would need at least the following five
> patches with it (23481121c81d is the second last):
>
> 148b0aa78e4e ("platform: x86: intel_cht_int33fe: Remove the old connections for the muxes")
> 23481121c81d ("usb: typec: class: Don't use port parent for getting mux handles")
> 495965a1002a ("platform: x86: intel_cht_int33fe: Add connections for the USB Type-C port")
> 78d2b54b134e ("platform: x86: intel_cht_int33fe: Add connection for the DP alt mode")
> 140a4ec4addd ("platform: x86: intel_cht_int33fe: Register all connections at once")
> cd7753d37138 ("drivers: base: Helpers for adding device connection descriptions")
I've now tested these on top of v4.19.x, and I'm confident that these
are all the prerequisites that are needed. I'll send the original
patch to stable@xxxxxxxxxxxxxxx ml.
thanks,
--
heikki
