| Message ID | 20230817145941.1091418-1-bmasney@redhat.com (mailing list archive) |
|---|---|
| State | Not Applicable |
| Headers | show |
| Series | pinctrl: qcom-pmic-gpio: silence -EPROBE_DEFER message on probe | expand |
On Thu, Aug 17, 2023 at 05:01:19PM +0200, Konrad Dybcio wrote: > On 17.08.2023 16:59, Brian Masney wrote: > > The following message shows up one or more times when booting a Qualcomm > > SA8775 Development board: > > > > qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: can't add gpio chip > > > > Convert this over to use dev_err_probe() to silence this message. > > > > Signed-off-by: Brian Masney <bmasney@redhat.com> > > --- > That looks odd, why would it ever defer? > > SPMI should be up by the time it gets a chance to probe. You replied within the same minute of me posting that patch, which is the fastest review I've had to date on an upstream kernel list. Before we continue, please verify: [ ] I am not a robot :) So SPMI is up and probes normally the first time, and is up by time this driver probes. I think the probe deferral is happening somewhere in pinctrl, however I am not sure exactly where. I added some tracers to the kernel command line and here's some relevant log messages: device: 'c440000.spmi:pmic@2:gpio@8800': device_add bus: 'platform': add device c440000.spmi:pmic@2:gpio@8800 PM: Adding info for platform:c440000.spmi:pmic@2:gpio@8800 bus: 'platform': __driver_probe_device: matched device c440000.spmi:pmic@2:gpio@8800 with driver qcom-spmi-gpio bus: 'platform': really_probe: probing driver qcom-spmi-gpio with device c440000.spmi:pmic@2:gpio@8800 qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: no pinctrl handle qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: try to register 12 pins ... pinctrl core: registered pin 0 (gpio1) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 1 (gpio2) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 2 (gpio3) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 3 (gpio4) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 4 (gpio5) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 5 (gpio6) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 6 (gpio7) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 7 (gpio8) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 8 (gpio9) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 9 (gpio10) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 10 (gpio11) on c440000.spmi:pmic@2:gpio@8800 pinctrl core: registered pin 11 (gpio12) on c440000.spmi:pmic@2:gpio@8800 qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: no hogs found qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: error -EPROBE_DEFER: can't add gpio chip qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: Driver qcom-spmi-gpio requests probe deferral platform c440000.spmi:pmic@2:gpio@8800: Added to deferred list The second time it probes the device is successfully added. Brian
On 17/08/2023 18:26, Brian Masney wrote: > On Thu, Aug 17, 2023 at 05:01:19PM +0200, Konrad Dybcio wrote: >> On 17.08.2023 16:59, Brian Masney wrote: >>> The following message shows up one or more times when booting a Qualcomm >>> SA8775 Development board: >>> >>> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: can't add gpio chip >>> >>> Convert this over to use dev_err_probe() to silence this message. >>> >>> Signed-off-by: Brian Masney <bmasney@redhat.com> >>> --- >> That looks odd, why would it ever defer? >> >> SPMI should be up by the time it gets a chance to probe. > > You replied within the same minute of me posting that patch, which is > the fastest review I've had to date on an upstream kernel list. Before > we continue, please verify: > > [ ] I am not a robot > > :) > > So SPMI is up and probes normally the first time, and is up by time this > driver probes. I think the probe deferral is happening somewhere in > pinctrl, however I am not sure exactly where. I added some tracers to > the kernel command line and here's some relevant log messages: > > device: 'c440000.spmi:pmic@2:gpio@8800': device_add > bus: 'platform': add device c440000.spmi:pmic@2:gpio@8800 > PM: Adding info for platform:c440000.spmi:pmic@2:gpio@8800 > bus: 'platform': __driver_probe_device: matched device c440000.spmi:pmic@2:gpio@8800 with driver qcom-spmi-gpio > bus: 'platform': really_probe: probing driver qcom-spmi-gpio with device c440000.spmi:pmic@2:gpio@8800 > qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: no pinctrl handle > qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: try to register 12 pins ... > pinctrl core: registered pin 0 (gpio1) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 1 (gpio2) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 2 (gpio3) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 3 (gpio4) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 4 (gpio5) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 5 (gpio6) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 6 (gpio7) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 7 (gpio8) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 8 (gpio9) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 9 (gpio10) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 10 (gpio11) on c440000.spmi:pmic@2:gpio@8800 > pinctrl core: registered pin 11 (gpio12) on c440000.spmi:pmic@2:gpio@8800 > qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: no hogs found > qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: error -EPROBE_DEFER: can't add gpio chip > qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: Driver qcom-spmi-gpio requests probe deferral > platform c440000.spmi:pmic@2:gpio@8800: Added to deferred list > > The second time it probes the device is successfully added. There is a bug in DTS. I'll send a patch. Best regards, Krzysztof
On 18/08/2023 15:51, Krzysztof Kozlowski wrote: >> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: no hogs found >> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: error -EPROBE_DEFER: can't add gpio chip >> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: Driver qcom-spmi-gpio requests probe deferral >> platform c440000.spmi:pmic@2:gpio@8800: Added to deferred list >> >> The second time it probes the device is successfully added. > > There is a bug in DTS. I'll send a patch. https://lore.kernel.org/linux-arm-msm/20230818135538.47481-1-krzysztof.kozlowski@linaro.org/T/#u Best regards, Krzysztof
On Fri, Aug 18, 2023 at 03:59:18PM +0200, Krzysztof Kozlowski wrote: > On 18/08/2023 15:51, Krzysztof Kozlowski wrote: > >> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: no hogs found > >> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: error -EPROBE_DEFER: can't add gpio chip > >> qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: Driver qcom-spmi-gpio requests probe deferral > >> platform c440000.spmi:pmic@2:gpio@8800: Added to deferred list > >> > >> The second time it probes the device is successfully added. > > > > There is a bug in DTS. I'll send a patch. > > https://lore.kernel.org/linux-arm-msm/20230818135538.47481-1-krzysztof.kozlowski@linaro.org/T/#u Thanks for sending that. I didn't look at the DTS. Let's just drop the patch that I posted here since that message may help someone in the future identify this same type of issue for another platform. Brian
diff --git a/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c b/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c index deded9c6fd7d..d7576fa14935 100644 --- a/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c +++ b/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c @@ -1154,10 +1154,8 @@ static int pmic_gpio_probe(struct platform_device *pdev) girq->child_irq_domain_ops.translate = pmic_gpio_domain_translate; ret = gpiochip_add_data(&state->chip, state); - if (ret) { - dev_err(state->dev, "can't add gpio chip\n"); - return ret; - } + if (ret) + return dev_err_probe(state->dev, ret, "can't add gpio chip\n"); /* * For DeviceTree-supported systems, the gpio core checks the
The following message shows up one or more times when booting a Qualcomm SA8775 Development board: qcom-spmi-gpio c440000.spmi:pmic@2:gpio@8800: can't add gpio chip Convert this over to use dev_err_probe() to silence this message. Signed-off-by: Brian Masney <bmasney@redhat.com> --- drivers/pinctrl/qcom/pinctrl-spmi-gpio.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)