Re: [PATCH v2] pinctrl: queue GPIO operations instead of defering

Re: [PATCH v2] pinctrl: queue GPIO operations instead of defering

[Stephen Warren]

On 08/17/2013 08:56 AM, Linus Walleij wrote:
> We currently defer probing of the caller if a pinctrl GPIO
> request or direction setting comes in before the range mapping
> a certain GPIO to a certain pin controller is available.
> 
> This can end up with a circular dependency: the GPIO driver
> needs the pin controller to be ready and the pin controller
> need the GPIO driver to be ready. This also happens if
> pin controllers and GPIO controllers compiled as modules
> are inserted in a certain order.
> 
> To break this circular dependence, queue any GPIO
> operations coming to the framework until the range is ready,
> instead of deferring the probe of the caller.

I just noticed that next-20130819 is failing on the Tegra114 Dalmore
board because of this patch. The kernel log is below. Reverting this
patch solves the problem

> [   12.427751] tegra114-pinctrl 70000868.pinmux: pin USB_VBUS_EN0 PN4 already requested by Tegra GPIOs:108; cannot claim for Tegra GPIOs:108
> [   12.440320] tegra114-pinctrl 70000868.pinmux: pin-108 (Tegra GPIOs:108) status -22
> [   12.447965] usb1_vbus: Failed to request enable GPIO108: -22
> [   12.454306] reg-fixed-voltage 3.regulator: Failed to register regulator: -22
> [   12.461508] reg-fixed-voltage: probe of 3.regulator failed with error -22
> [   12.469659] tegra114-pinctrl 70000868.pinmux: pin SPDIF_IN PK6 already requested by Tegra GPIOs:86; cannot claim for Tegra GPIOs:86
> [   12.481606] tegra114-pinctrl 70000868.pinmux: pin-86 (Tegra GPIOs:86) status -22
> [   12.489052] usb2_vbus: Failed to request enable GPIO86: -22
> [   12.495226] reg-fixed-voltage 4.regulator: Failed to register regulator: -22
> [   12.502449] reg-fixed-voltage: probe of 4.regulator failed with error -22
> [   12.510542] tegra114-pinctrl 70000868.pinmux: pin GMI_CLK PK1 already requested by Tegra GPIOs:81; cannot claim for Tegra GPIOs:81
> [   12.522403] tegra114-pinctrl 70000868.pinmux: pin-81 (Tegra GPIOs:81) status -22
> [   12.529896] vdd_hdmi_5v0: Failed to request enable GPIO81: -22
> [   12.536281] reg-fixed-voltage 5.regulator: Failed to register regulator: -22
> [   12.543465] reg-fixed-voltage: probe of 5.regulator failed with error -22
> [   12.552509] vdd_cam_1v8_reg: 1800 mV 
> [   12.558205] platform 7d008000.usb: Driver tegra-ehci requests probe deferral
> [   12.566600] platform 7d008000.usb-phy: Driver tegra-phy requests probe deferral
> [   12.577475] input: gpio-keys.2 as /devices/soc0/gpio-keys.2/input/input1
> [   12.588329] platform 7d008000.usb: Driver tegra-ehci requests probe deferral
> [   12.596967] platform 7d008000.usb-phy: Driver tegra-phy requests probe deferral

Re: [PATCH v2] pinctrl: queue GPIO operations instead of defering

[Stephen Warren]

On 08/19/2013 01:21 PM, Stephen Warren wrote:
> On 08/17/2013 08:56 AM, Linus Walleij wrote:
>> We currently defer probing of the caller if a pinctrl GPIO
>> request or direction setting comes in before the range mapping
>> a certain GPIO to a certain pin controller is available.
>>
>> This can end up with a circular dependency: the GPIO driver
>> needs the pin controller to be ready and the pin controller
>> need the GPIO driver to be ready. This also happens if
>> pin controllers and GPIO controllers compiled as modules
>> are inserted in a certain order.
>>
>> To break this circular dependence, queue any GPIO
>> operations coming to the framework until the range is ready,
>> instead of deferring the probe of the caller.
> 
> I just noticed that next-20130819 is failing on the Tegra114 Dalmore
> board because of this patch. The kernel log is below. Reverting this
> patch solves the problem

Linus, any thoughts here?

>> [   12.427751] tegra114-pinctrl 70000868.pinmux: pin USB_VBUS_EN0 PN4 already requested by Tegra GPIOs:108; cannot claim for Tegra GPIOs:108
>> [   12.440320] tegra114-pinctrl 70000868.pinmux: pin-108 (Tegra GPIOs:108) status -22
>> [   12.447965] usb1_vbus: Failed to request enable GPIO108: -22
>> [   12.454306] reg-fixed-voltage 3.regulator: Failed to register regulator: -22
>> [   12.461508] reg-fixed-voltage: probe of 3.regulator failed with error -22
>> [   12.469659] tegra114-pinctrl 70000868.pinmux: pin SPDIF_IN PK6 already requested by Tegra GPIOs:86; cannot claim for Tegra GPIOs:86
>> [   12.481606] tegra114-pinctrl 70000868.pinmux: pin-86 (Tegra GPIOs:86) status -22
>> [   12.489052] usb2_vbus: Failed to request enable GPIO86: -22
>> [   12.495226] reg-fixed-voltage 4.regulator: Failed to register regulator: -22
>> [   12.502449] reg-fixed-voltage: probe of 4.regulator failed with error -22
>> [   12.510542] tegra114-pinctrl 70000868.pinmux: pin GMI_CLK PK1 already requested by Tegra GPIOs:81; cannot claim for Tegra GPIOs:81
>> [   12.522403] tegra114-pinctrl 70000868.pinmux: pin-81 (Tegra GPIOs:81) status -22
>> [   12.529896] vdd_hdmi_5v0: Failed to request enable GPIO81: -22
>> [   12.536281] reg-fixed-voltage 5.regulator: Failed to register regulator: -22
>> [   12.543465] reg-fixed-voltage: probe of 5.regulator failed with error -22
>> [   12.552509] vdd_cam_1v8_reg: 1800 mV 
>> [   12.558205] platform 7d008000.usb: Driver tegra-ehci requests probe deferral
>> [   12.566600] platform 7d008000.usb-phy: Driver tegra-phy requests probe deferral
>> [   12.577475] input: gpio-keys.2 as /devices/soc0/gpio-keys.2/input/input1
>> [   12.588329] platform 7d008000.usb: Driver tegra-ehci requests probe deferral
>> [   12.596967] platform 7d008000.usb-phy: Driver tegra-phy requests probe deferral

Re: [PATCH v2] pinctrl: queue GPIO operations instead of defering

[Linus Walleij]

On Wed, Aug 21, 2013 at 7:45 PM, Stephen Warren <swarren@wwwdotorg.org> wrote:
> On 08/19/2013 01:21 PM, Stephen Warren wrote:
>> On 08/17/2013 08:56 AM, Linus Walleij wrote:
>>> We currently defer probing of the caller if a pinctrl GPIO
>>> request or direction setting comes in before the range mapping
>>> a certain GPIO to a certain pin controller is available.
>>>
>>> This can end up with a circular dependency: the GPIO driver
>>> needs the pin controller to be ready and the pin controller
>>> need the GPIO driver to be ready. This also happens if
>>> pin controllers and GPIO controllers compiled as modules
>>> are inserted in a certain order.
>>>
>>> To break this circular dependence, queue any GPIO
>>> operations coming to the framework until the range is ready,
>>> instead of deferring the probe of the caller.
>>
>> I just noticed that next-20130819 is failing on the Tegra114 Dalmore
>> board because of this patch. The kernel log is below. Reverting this
>> patch solves the problem
>
> Linus, any thoughts here?

None whatsoever, but logically it should be that your pinctrl+GPIO
driver pair depends on the GPIO driver coming first and having it's probing
deferred.

I'll rip out that patch for the time being...

Yours,
Linus Walleij