gpio: Add missing open drain/source handling to gpiod_set_value_cansleep()

Commit Message

Linus Walleij Jan. 10, 2018, 1:17 p.m. UTC

On Tue, Jan 9, 2018 at 7:08 PM, Geert Uytterhoeven
<[e-mail]> wrote:

> Since commit f11a04464ae57e8d ("i2c: gpio: Enable working over slow
> can_sleep GPIOs"), probing the i2c RTC connected to an i2c-gpio bus on
> r8a7740/armadillo fails with:
>
>     rtc-s35390a 0-0030: error resetting chip
>     rtc-s35390a: probe of 0-0030 failed with error -5
>
> More debug code reveals:
>
>     i2c i2c-0: master_xfer[0] R, addr=0x30, len=1
>     i2c i2c-0: NAK from device addr 0x30 msg #0
>     s35390a_get_reg: ret = -6
>
> Commit 02e479808b5d62f8 ("gpio: Alter semantics of *raw* operations to
> actually be raw") moved open drain/source handling from
> gpiod_set_raw_value_commit() to gpiod_set_value(), but forgot to take
> into account that gpiod_set_value_cansleep() also needs this handling.
> The i2c protocol mandates that i2c signals are open drain, hence i2c
> communication fails.
>
> Fix this by adding the missing handling to gpiod_set_value_cansleep(),
> using a new common helper __gpiod_set_value_nocheck().
>
> Fixes: 02e479808b5d62f8 ("gpio: Alter semantics of *raw* operations to actually be raw")
> Signed-off-by: Geert Uytterhoeven <[e-mail]>
> ---
> The bad gpio commit is in v4.15-rc1 and later.
> The i2c commit exposing this is not yet upstream, but in i2c/for-next.

Whoops!

It appears Meltdown and Spectre saves my ass because now I have
time to get this upstream before v4.15...

Applied for fixes with the following tweaks: I removed the __prefix (I don't
like them) and added kerneldoc.

                value = !value;
@@ -2919,7 +2928,7 @@ void gpiod_set_value(struct gpio_desc *desc, int value)
 {
        VALIDATE_DESC_VOID(desc);
        WARN_ON(desc->gdev->chip->can_sleep);
-       __gpiod_set_value_nocheck(desc, value);
+       gpiod_set_value_nocheck(desc, value);
 }
 EXPORT_SYMBOL_GPL(gpiod_set_value);

@@ -3247,7 +3256,7 @@ void gpiod_set_value_cansleep(struct gpio_desc
*desc, int value)
 {
        might_sleep_if(extra_checks);
        VALIDATE_DESC_VOID(desc);
-       __gpiod_set_value_nocheck(desc, value);
+       gpiod_set_value_nocheck(desc, value);
 }
 EXPORT_SYMBOL_GPL(gpiod_set_value_cansleep);


Yours,
Linus Walleij

Comments

Geert Uytterhoeven Jan. 10, 2018, 1:29 p.m. UTC | #1

Hi Linus,

On Wed, Jan 10, 2018 at 2:17 PM, Linus Walleij <[e-mail]> wrote:
> On Tue, Jan 9, 2018 at 7:08 PM, Geert Uytterhoeven
> <[e-mail]> wrote:
>
>> Since commit f11a04464ae57e8d ("i2c: gpio: Enable working over slow
>> can_sleep GPIOs"), probing the i2c RTC connected to an i2c-gpio bus on
>> r8a7740/armadillo fails with:
>>
>>     rtc-s35390a 0-0030: error resetting chip
>>     rtc-s35390a: probe of 0-0030 failed with error -5
>>
>> More debug code reveals:
>>
>>     i2c i2c-0: master_xfer[0] R, addr=0x30, len=1
>>     i2c i2c-0: NAK from device addr 0x30 msg #0
>>     s35390a_get_reg: ret = -6
>>
>> Commit 02e479808b5d62f8 ("gpio: Alter semantics of *raw* operations to
>> actually be raw") moved open drain/source handling from
>> gpiod_set_raw_value_commit() to gpiod_set_value(), but forgot to take
>> into account that gpiod_set_value_cansleep() also needs this handling.
>> The i2c protocol mandates that i2c signals are open drain, hence i2c
>> communication fails.
>>
>> Fix this by adding the missing handling to gpiod_set_value_cansleep(),
>> using a new common helper __gpiod_set_value_nocheck().
>>
>> Fixes: 02e479808b5d62f8 ("gpio: Alter semantics of *raw* operations to actually be raw")
>> Signed-off-by: Geert Uytterhoeven <[e-mail]>
>> ---
>> The bad gpio commit is in v4.15-rc1 and later.
>> The i2c commit exposing this is not yet upstream, but in i2c/for-next.
>
> Whoops!
>
> It appears Meltdown and Spectre saves my ass because now I have
> time to get this upstream before v4.15...
>
> Applied for fixes with the following tweaks: I removed the __prefix (I don't
> like them) and added kerneldoc.

Looks good to me (also in git). Thanks!

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [e-mail]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Patch

diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index 1693bb9ab72c..14532d9576e4 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -2892,7 +2892,16 @@  void gpiod_set_raw_value(struct gpio_desc
*desc, int value)
 }
 EXPORT_SYMBOL_GPL(gpiod_set_raw_value);

-static void __gpiod_set_value_nocheck(struct gpio_desc *desc, int value)
+/**
+ * gpiod_set_value_nocheck() - set a GPIO line value without checking
+ * @desc: the descriptor to set the value on
+ * @value: value to set
+ *
+ * This sets the value of a GPIO line backing a descriptor, applying
+ * different semantic quirks like active low and open drain/source
+ * handling.
+ */
+static void gpiod_set_value_nocheck(struct gpio_desc *desc, int value)
 {
        if (test_bit(FLAG_ACTIVE_LOW, &desc->flags))