Messages in this thread | | | Date | Mon, 21 Apr 2014 12:26:06 -0500 | From | Josh Cartwright <> | Subject | Re: [PATCH 2/4] soc: qcom: Add GSBI driver |
| |
On Mon, Apr 21, 2014 at 12:11:18PM -0500, Andy Gross wrote: > On Mon, Apr 21, 2014 at 11:54:00AM -0500, Josh Cartwright wrote: > > > + if (of_property_read_u32(node, "qcom,mode", &mode)) { > > > + dev_err(gsbi->dev, "missing mode configuration\n"); > > > + return -EINVAL; > > > + } > > > > I'm wondering if you should really be a (very simple) pinctrl driver > > proper. > > Perhaps. But how would i reconcile more than one device node that uses the same > GSBI? One could still trounce the other unless I only allow one setting of the > GSBI. >
I don't understand, as long as the pins/functions have been specified properly to the pinctrl core, I would expect a conflicting configuration to be rejected.
Anyway, I wouldn't expect the subnodes to be consuming the GSBI pin configuration anyway (although that could probably be done), instead, I would expect the GSBI node to consume it's own pin configuration.
-- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
| |