On Tue, Aug 05, 2014 at 02:33:25PM +0200, Sebastian Andrzej Siewior wrote: > On 08/04/2014 12:13 PM, Markus Pargmann wrote: > > >> In the long term I would suggest to move the DT probe over to the musb > >> core code and we wouldn't need the node assignment anymore… > >> > >> [0] http://www.spinics.net/lists/linux-usb/msg94701.html > > > > Really old thread but I just noticed that this solution does not work > > for automatically parsed properties like 'default' pinctrl. I added > > 'default' pinctrl settings for the musb device node. They are > > automatically setup by the driver core infrastracture as expected. > > > > The problem is that the 'default' pincontrol settings will be applied > > again for the musb-hdrc because they are extracted from the of_ndoe > > data. This fails because they are already claimed by the 'musb-dsps' > > driver. musb-hdrc continues probing, but there is a big error message > > that it failed: > > > > [ 1.181912] pinctrl-single 44e10800.pinmux: pin 44e10a20.0 already requested by 47401c00.usb; cannot clai > > m for musb-hdrc.1.auto > > [ 1.194053] pinctrl-single 44e10800.pinmux: pin-136 (musb-hdrc.1.auto) status -22 > > [ 1.201930] pinctrl-single 44e10800.pinmux: could not request pin 136 (44e10a20.0) from group pinmux_usb1 > > _pins on device pinctrl-single > > [ 1.214790] musb-hdrc musb-hdrc.1.auto: Error applying setting, reverse things back > > > > Any ideas how to solve it for mainline? > > Just what I said last time and I hope Felipe has the same opinion: > Teach musb-hdrc DT. Which means "ti,musb-am33xx" would be listed for > for musb-hdrc itself and we would lose that child device. The > additional ops that we have would be passed via the data field next to > the id. sounds good to me. -- balbi
Attachment:
signature.asc
Description: Digital signature
