On 05.05.2020 18:12, Wolfram Sang wrote:
> On Wed, Apr 15, 2020 at 10:06:43AM +0300, Codrin Ciubotariu wrote:
>> devm_gpiod_get() usually calls gpio_request_enable() for non-strict pinmux
>> drivers. These puts the pins in GPIO mode, whithout notifying the pinctrl
>> driver. At this point, the I2C bus no longer owns the pins. To mux the
>> pins back to the I2C bus, we use the pinctrl driver to change the state
>> of the pins to GPIO, before using devm_gpiod_get(). After the pins are
>> received as GPIOs, we switch theer pinctrl state back to the default
>> one,
>>
>> Fixes: d3d3fdcc4c90 ("i2c: at91: implement i2c bus recovery")
>> Signed-off-by: Codrin Ciubotariu <codrin.ciubotariu@xxxxxxxxxxxxx>
>
> Applied to for-current, thanks!
Just looked at this patch and noticed that I should change the pinctrl
state back to default if we can't get the gpio pins. I will submit a
patch to fix this.
>
> This will do for 5.7. For 5.8 or 5.9, I can imagine to take the two
> pinctrl_state pointers into bus_recovery_info and handle all this in the
> core. I will try this later this week if noone is super-eager to try it
> out before.
>
By 'all this' you mean to move the entire function in the core, right?
Having just these two pointers bus_recinovery_info won't help much. I
can try it, if you haven't already started...
Best regards,
Codrin
