Skip to contentSkip to conversation

Error code handling by GPIO subsystem for PIN_CONFIG_PERSIST_STATE config

DISCUSSION LakshmiPraveen Kopparthi 2 participants 3 messages May 3, 2021 GPIO

lore
Share & tools
lore
#0 lore raw

Hello GPIO Maintainers,

I am writing a GPIO device driver for one of our PCIe based devices.While addressing the warnings and errors reported by checkpatch script, I ran into an problem while fixing this.

WARNING: ENOTSUPP is not a SUSV4 error code, prefer EOPNOTSUPP#197:
FILE: kernel_src/drivers/gpio/gpio-pci1xxxx.c:159:+ ret = -ENOTSUPP;

This warning is reported against the gpio_chip -> set_config() callback's implementation because for the not supported configurations, -ENOTSUPP is returned by this callback.If -ENOTSUPP is replaced by --
-EOPNOTSUPP , the GPIOs are not accessible from both the device node
(char device in /dev) and from sysfs interface as well.

On further analysis it is understood that while trying to access a gpio from user space,

The gpio sub system calls gpiod_set_transitory() API to issue a set configuration with PIN_CONFIG_PERSIST_STATE as a parameter.In this context, set_config() call back in the device's source code gets called. Since the PIN_CONFIG_PERSIST_STATE configuration is not supported by our device, It returns an error code.

If the error code is -EOPNOTSUPP, then the gpio access request is terminated by returning the error code as EOPNOTSUPP.If set_config() returns ENOTSUPP then the API- gpio_set_config_with_argument_optional() internally updates the return value to true thus making the gpio access request to proceed further.

If gpio_set_config_with_argument_optional() API updated to handle the EOPNOTSUPP along with ENOTSUPP, then the issue seemed to be resolved.

I am planning to submit a patch to add support for our device. Please let me know If my understanding is correct or not and also the way forword on this issue.

Thanks,
Lakshmi Praveen.

#1 lore raw
Quoted 6 lines · LakshmiPraveen Kopparthi

Hello GPIO Maintainers,

I am writing a GPIO device driver for one of our PCIe based devices.While addressing the warnings and errors reported by checkpatch script, I ran into an problem while fixing this.

The problem is that you started fixing the symptoms, not a cause. The cause is... checkpatch itself. Please, fix it instead (The warning should be rewritten to be not so frightening).

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#2 lore raw
Quoted 5 lines · Andy Shevchenko
Quoted 3 lines · LakshmiPraveen Kopparthi

Hello GPIO Maintainers,

Quoted 7 lines
Quoted 3 lines

I am writing a GPIO device driver for one of our PCIe based devices.While addressing the warnings and errors reported by checkpatch script, I ran into an problem while fixing this.

The problem is that you started fixing the symptoms, not a cause. The cause is... checkpatch itself. Please, fix it instead (The warning should be rewritten to be not so frightening).

Read this [1] discussion to understand how to proceed.

[1]: https://lore.kernel.org/lkml/[email protected]/

Signature · 2 lines
With Best Regards,
Andy Shevchenko