On 07/06/18 16:44, Fabio Estevam wrote:
On Thu, Jun 7, 2018 at 12:23 PM, Richard Fitzgerald
<rf@xxxxxxxxxxxxxxxxxxxxx> wrote:
Most likely the bug that my patch fixed was masking a different
problem with the iMX system.
In imx53-qsb-common.dtsi the iomuxc pinctrl driver node has some hogs
and also has many other pinctrl configurations that are not hogs and
are used by other drivers. What I think is happening is that because of
the bug in the original dt_to_map_one_config() all the child nodes under
iomux would have been treated as hogs, including the ones that are not
hogs. So when the iomuxc driver is probed all the possible pinctrl
configurations would be applied immediately and you would see correctly
configured pin muxes even though the way they got configured was wrong.
I assume your problem is that those settings which are not hogs are not
getting set up now? That indicates that for some reason when the
No, only pinctrl_hog group is not being set up now.
individual drivers call pinctrl_get() the mappings do not get applied.
So the interesting case to debug is what happens when
dt_to_map_one_config() is called with pctldev==NULL, why are the
pinctrl mappings requested by all those drivers not mapped to the
iomuxc driver?
Which of the pinctrl configuration are not applied? Are all missing?
Or only some?
Only pinctrl_hog configuration is not taking effect.
Interesting, I'll take another look with that information. Your log
shows that it matches.
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
