pinctrl: sx150x bug

pinctrl: sx150x bug

[Martin DEVERA]

Hello,

I encountered bug in SX1502 expander driver in 5.7.7. Here is relevant 
DTS part:

compatible = "semtech,sx1502q";
gpio4_cfg_pins: gpio2-cfg {
                         pins = "gpio5";
                         output-high;
     };

And part of OOPS:

[    0.673996] [<c023cfa6>] (gpiochip_get_data) from [<c023b235>] 
(sx150x_gpio_direction_output+0xd)
[    0.683259] [<c023b235>] (sx150x_gpio_direction_output) from 
[<c023b363>] (sx150x_pinconf_set+0x)
[    0.692796] [<c023b363>] (sx150x_pinconf_set) from [<c0238fef>] 
(pinconf_apply_setting+0x39/0x7e)
[    0.701635] [<c0238fef>] (pinconf_apply_setting) from [<c0236c77>] 
(pinctrl_commit_state+0xa5/0x)
[    0.710648] [<c0236c77>] (pinctrl_commit_state) from [<c0237e03>] 
(pinctrl_enable+0xff/0x1d4)
[    0.719139] [<c0237e03>] (pinctrl_enable) from [<c023b791>] 
(sx150x_probe+0x1a3/0x358)
[    0.727027] [<c023b791>] (sx150x_probe) from [<c02c38bf>] 
(i2c_device_probe+0x1bb/0x1dc)

The problem is that sx150x_pinconf_set uses sx150x_gpio_direction_output 
but gpio is not
setup yet. Patch below fixes it but I'm not sure whether is it correct:

diff --git a/drivers/pinctrl/pinctrl-sx150x.c 
b/drivers/pinctrl/pinctrl-sx150x.c
index 6e74bd87d959..3f5651edd336 100644
--- a/drivers/pinctrl/pinctrl-sx150x.c
+++ b/drivers/pinctrl/pinctrl-sx150x.c
@@ -1154,12 +1154,6 @@ static int sx150x_probe(struct i2c_client *client,
                 return ret;
         }

-       ret = pinctrl_enable(pctl->pctldev);
-       if (ret) {
-               dev_err(dev, "Failed to enable pinctrl device\n");
-               return ret;
-       }
-
         /* Register GPIO controller */
         pctl->gpio.base = -1;
         pctl->gpio.ngpio = pctl->data->npins;
@@ -1191,6 +1185,12 @@ static int sx150x_probe(struct i2c_client *client,
         if (ret)
                 return ret;

+       ret = pinctrl_enable(pctl->pctldev);
+       if (ret) {
+               dev_err(dev, "Failed to enable pinctrl device\n");
+               return ret;
+       }
+
         ret = gpiochip_add_pin_range(&pctl->gpio, dev_name(dev),
                                      0, 0, pctl->data->npins);
         if (ret)

Re: pinctrl: sx150x bug

[Linus Walleij]

Hi Martin,

thanks for your report, let's check with Peter, Neil and Andrey who also
use this expander if they also see this problem (CC).

On Wed, Aug 5, 2020 at 11:28 AM Martin DEVERA <devik@eaxlabs.cz> wrote:

> I encountered bug in SX1502 expander driver in 5.7.7. Here is relevant
> DTS part:
>
> compatible = "semtech,sx1502q";
> gpio4_cfg_pins: gpio2-cfg {
>                          pins = "gpio5";
>                          output-high;
>      };
>
> And part of OOPS:
>
> [    0.673996] [<c023cfa6>] (gpiochip_get_data) from [<c023b235>]
> (sx150x_gpio_direction_output+0xd)
> [    0.683259] [<c023b235>] (sx150x_gpio_direction_output) from
> [<c023b363>] (sx150x_pinconf_set+0x)
> [    0.692796] [<c023b363>] (sx150x_pinconf_set) from [<c0238fef>]
> (pinconf_apply_setting+0x39/0x7e)
> [    0.701635] [<c0238fef>] (pinconf_apply_setting) from [<c0236c77>]
> (pinctrl_commit_state+0xa5/0x)
> [    0.710648] [<c0236c77>] (pinctrl_commit_state) from [<c0237e03>]
> (pinctrl_enable+0xff/0x1d4)
> [    0.719139] [<c0237e03>] (pinctrl_enable) from [<c023b791>]
> (sx150x_probe+0x1a3/0x358)
> [    0.727027] [<c023b791>] (sx150x_probe) from [<c02c38bf>]
> (i2c_device_probe+0x1bb/0x1dc)
>
> The problem is that sx150x_pinconf_set uses sx150x_gpio_direction_output
> but gpio is not
> setup yet. Patch below fixes it but I'm not sure whether is it correct:
>
> diff --git a/drivers/pinctrl/pinctrl-sx150x.c
> b/drivers/pinctrl/pinctrl-sx150x.c
> index 6e74bd87d959..3f5651edd336 100644
> --- a/drivers/pinctrl/pinctrl-sx150x.c
> +++ b/drivers/pinctrl/pinctrl-sx150x.c
> @@ -1154,12 +1154,6 @@ static int sx150x_probe(struct i2c_client *client,
>                  return ret;
>          }
>
> -       ret = pinctrl_enable(pctl->pctldev);
> -       if (ret) {
> -               dev_err(dev, "Failed to enable pinctrl device\n");
> -               return ret;
> -       }
> -
>          /* Register GPIO controller */
>          pctl->gpio.base = -1;
>          pctl->gpio.ngpio = pctl->data->npins;
> @@ -1191,6 +1185,12 @@ static int sx150x_probe(struct i2c_client *client,
>          if (ret)
>                  return ret;
>
> +       ret = pinctrl_enable(pctl->pctldev);
> +       if (ret) {
> +               dev_err(dev, "Failed to enable pinctrl device\n");
> +               return ret;
> +       }
> +
>          ret = gpiochip_add_pin_range(&pctl->gpio, dev_name(dev),
>                                       0, 0, pctl->data->npins);
>          if (ret)

I don't see any problem with the patch, can you send a proper patch
with git-send-email so we can test it and apply it if it works for the
other users? Include the mentioned people on CC.

Yours,
Linus Walleij

Re: pinctrl: sx150x bug

[Andrey Smirnov]

On Thu, Aug 6, 2020 at 2:41 AM Linus Walleij <linus.walleij@linaro.org> wrote:
>
> Hi Martin,
>
> thanks for your report, let's check with Peter, Neil and Andrey who also
> use this expander if they also see this problem (CC).
>

Looks reasonable. I haven't used that HW in a while, so I added Chris
Healy to this thread, since he might be in a better position to
comment on this and maybe even provide a Tested-by

> On Wed, Aug 5, 2020 at 11:28 AM Martin DEVERA <devik@eaxlabs.cz> wrote:
>
> > I encountered bug in SX1502 expander driver in 5.7.7. Here is relevant
> > DTS part:
> >
> > compatible = "semtech,sx1502q";
> > gpio4_cfg_pins: gpio2-cfg {
> >                          pins = "gpio5";
> >                          output-high;
> >      };
> >
> > And part of OOPS:
> >
> > [    0.673996] [<c023cfa6>] (gpiochip_get_data) from [<c023b235>]
> > (sx150x_gpio_direction_output+0xd)
> > [    0.683259] [<c023b235>] (sx150x_gpio_direction_output) from
> > [<c023b363>] (sx150x_pinconf_set+0x)
> > [    0.692796] [<c023b363>] (sx150x_pinconf_set) from [<c0238fef>]
> > (pinconf_apply_setting+0x39/0x7e)
> > [    0.701635] [<c0238fef>] (pinconf_apply_setting) from [<c0236c77>]
> > (pinctrl_commit_state+0xa5/0x)
> > [    0.710648] [<c0236c77>] (pinctrl_commit_state) from [<c0237e03>]
> > (pinctrl_enable+0xff/0x1d4)
> > [    0.719139] [<c0237e03>] (pinctrl_enable) from [<c023b791>]
> > (sx150x_probe+0x1a3/0x358)
> > [    0.727027] [<c023b791>] (sx150x_probe) from [<c02c38bf>]
> > (i2c_device_probe+0x1bb/0x1dc)
> >
> > The problem is that sx150x_pinconf_set uses sx150x_gpio_direction_output
> > but gpio is not
> > setup yet. Patch below fixes it but I'm not sure whether is it correct:
> >
> > diff --git a/drivers/pinctrl/pinctrl-sx150x.c
> > b/drivers/pinctrl/pinctrl-sx150x.c
> > index 6e74bd87d959..3f5651edd336 100644
> > --- a/drivers/pinctrl/pinctrl-sx150x.c
> > +++ b/drivers/pinctrl/pinctrl-sx150x.c
> > @@ -1154,12 +1154,6 @@ static int sx150x_probe(struct i2c_client *client,
> >                  return ret;
> >          }
> >
> > -       ret = pinctrl_enable(pctl->pctldev);
> > -       if (ret) {
> > -               dev_err(dev, "Failed to enable pinctrl device\n");
> > -               return ret;
> > -       }
> > -
> >          /* Register GPIO controller */
> >          pctl->gpio.base = -1;
> >          pctl->gpio.ngpio = pctl->data->npins;
> > @@ -1191,6 +1185,12 @@ static int sx150x_probe(struct i2c_client *client,
> >          if (ret)
> >                  return ret;
> >
> > +       ret = pinctrl_enable(pctl->pctldev);
> > +       if (ret) {
> > +               dev_err(dev, "Failed to enable pinctrl device\n");
> > +               return ret;
> > +       }
> > +
> >          ret = gpiochip_add_pin_range(&pctl->gpio, dev_name(dev),
> >                                       0, 0, pctl->data->npins);
> >          if (ret)
>
> I don't see any problem with the patch, can you send a proper patch
> with git-send-email so we can test it and apply it if it works for the
> other users? Include the mentioned people on CC.
>
> Yours,
> Linus Walleij

Re: pinctrl: sx150x bug

[Linus Walleij]

On Wed, Aug 5, 2020 at 11:28 AM Martin DEVERA <devik@eaxlabs.cz> wrote:

> I encountered bug in SX1502 expander driver in 5.7.7. Here is relevant
> DTS part:
>
> compatible = "semtech,sx1502q";
> gpio4_cfg_pins: gpio2-cfg {
>                          pins = "gpio5";
>                          output-high;
>      };
>
> And part of OOPS:
>
> [    0.673996] [<c023cfa6>] (gpiochip_get_data) from [<c023b235>]
> (sx150x_gpio_direction_output+0xd)
> [    0.683259] [<c023b235>] (sx150x_gpio_direction_output) from
> [<c023b363>] (sx150x_pinconf_set+0x)
> [    0.692796] [<c023b363>] (sx150x_pinconf_set) from [<c0238fef>]
> (pinconf_apply_setting+0x39/0x7e)
> [    0.701635] [<c0238fef>] (pinconf_apply_setting) from [<c0236c77>]
> (pinctrl_commit_state+0xa5/0x)
> [    0.710648] [<c0236c77>] (pinctrl_commit_state) from [<c0237e03>]
> (pinctrl_enable+0xff/0x1d4)
> [    0.719139] [<c0237e03>] (pinctrl_enable) from [<c023b791>]
> (sx150x_probe+0x1a3/0x358)
> [    0.727027] [<c023b791>] (sx150x_probe) from [<c02c38bf>]
> (i2c_device_probe+0x1bb/0x1dc)
>
> The problem is that sx150x_pinconf_set uses sx150x_gpio_direction_output
> but gpio is not
> setup yet. Patch below fixes it but I'm not sure whether is it correct:

I simply rebased and applied this patch now.

Yours,
Linus Walleij