[PATCH] gpio: rcar: Improve clock error handling and reporting

[PATCH] gpio: rcar: Improve clock error handling and reporting

[Geert Uytterhoeven]

If the Renesas R-Car GPIO driver cannot find a functional clock, it
prints a warning, .e.g.

	gpio_rcar ffc40000.gpio: unable to get clock

and continues, as the clock is optional, depending on the SoC type.
This warning may confuse users.

To fix this, add a flag to indicate that the clock is mandatory or
optional:
  - If the clock is mandatory (on R-Car Gen2), a missing clock is now
    treated as a fatal error,
  - If the clock is optional (on R-Car Gen1), the warning is no longer
    printed.

The new flag is merged with the existing has_both_edge_trigger boolean
into a bitfield to save space.

Suggested-by: Magnus Damm <magnus.damm@gmail.com>
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
This depends on "gpio: rcar: Remove obsolete platform data support".

Tested on r8a7779/marzen (no functional clock) and r8a7791/koelsch (with
and without mandatory functional clock).

 drivers/gpio/gpio-rcar.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/drivers/gpio/gpio-rcar.c b/drivers/gpio/gpio-rcar.c
index bb6e363d5178634c..c890ac14b23a65ea 100644
--- a/drivers/gpio/gpio-rcar.c
+++ b/drivers/gpio/gpio-rcar.c
@@ -39,6 +39,7 @@ struct gpio_rcar_priv {
 	struct clk *clk;
 	unsigned int irq_parent;
 	unsigned has_both_edge_trigger:1;
+	unsigned needs_clk:1;
 };
 
 #define IOINTSEL 0x00	/* General IO/Interrupt Switching Register */
@@ -316,15 +317,18 @@ static int gpio_rcar_direction_output(struct gpio_chip *chip, unsigned offset,
 }
 
 struct gpio_rcar_info {
-	bool has_both_edge_trigger;
+	unsigned has_both_edge_trigger:1;
+	unsigned needs_clk:1;
 };
 
 static const struct gpio_rcar_info gpio_rcar_info_gen1 = {
-	.has_both_edge_trigger = false,
+	.has_both_edge_trigger = 0,
+	.needs_clk = 0,
 };
 
 static const struct gpio_rcar_info gpio_rcar_info_gen2 = {
-	.has_both_edge_trigger = true,
+	.has_both_edge_trigger = 1,
+	.needs_clk = 1,
 };
 
 static const struct of_device_id gpio_rcar_of_table[] = {
@@ -371,6 +375,7 @@ static int gpio_rcar_parse_dt(struct gpio_rcar_priv *p, unsigned int *npins)
 	ret = of_parse_phandle_with_fixed_args(np, "gpio-ranges", 3, 0, &args);
 	*npins = ret == 0 ? args.args[2] : RCAR_MAX_GPIO_PER_BANK;
 	p->has_both_edge_trigger = info->has_both_edge_trigger;
+	p->needs_clk = info->needs_clk;
 
 	if (*npins == 0 || *npins > RCAR_MAX_GPIO_PER_BANK) {
 		dev_warn(&p->pdev->dev,
@@ -409,7 +414,11 @@ static int gpio_rcar_probe(struct platform_device *pdev)
 
 	p->clk = devm_clk_get(dev, NULL);
 	if (IS_ERR(p->clk)) {
-		dev_warn(dev, "unable to get clock\n");
+		if (p->needs_clk) {
+			dev_err(dev, "unable to get clock\n");
+			ret = PTR_ERR(p->clk);
+			goto err0;
+		}
 		p->clk = NULL;
 	}
 
-- 
1.9.1

Re: [PATCH] gpio: rcar: Improve clock error handling and reporting

[Linus Walleij]

On Tue, Nov 24, 2015 at 4:05 PM, Geert Uytterhoeven
<geert+renesas@glider.be> wrote:

> If the Renesas R-Car GPIO driver cannot find a functional clock, it
> prints a warning, .e.g.
>
>         gpio_rcar ffc40000.gpio: unable to get clock
>
> and continues, as the clock is optional, depending on the SoC type.
> This warning may confuse users.
>
> To fix this, add a flag to indicate that the clock is mandatory or
> optional:
>   - If the clock is mandatory (on R-Car Gen2), a missing clock is now
>     treated as a fatal error,
>   - If the clock is optional (on R-Car Gen1), the warning is no longer
>     printed.
>
> The new flag is merged with the existing has_both_edge_trigger boolean
> into a bitfield to save space.

*DON'T* try to save bits in a megabyte kernel like this.
The better readability is worth the extra bits or bytes it takes.
Besides, did you really try to compile it before/after that change
to verify that it actually saved that space?

> Suggested-by: Magnus Damm <magnus.damm@gmail.com>
> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
> ---
> This depends on "gpio: rcar: Remove obsolete platform data support".

Where is that patch? This does not apply to my tree,
probably because of this.

Please sent the patches in a series.

>         unsigned has_both_edge_trigger:1;
> +       unsigned needs_clk:1;

both should be bool and assigned true/false.

Yours,
Linus Walleij

Re: [PATCH] gpio: rcar: Improve clock error handling and reporting

[Geert Uytterhoeven]

Hi Linus,

On Tue, Dec 1, 2015 at 10:54 AM, Linus Walleij <linus.walleij@linaro.org> wrote:
> On Tue, Nov 24, 2015 at 4:05 PM, Geert Uytterhoeven
> <geert+renesas@glider.be> wrote:
>
>> If the Renesas R-Car GPIO driver cannot find a functional clock, it
>> prints a warning, .e.g.
>>
>>         gpio_rcar ffc40000.gpio: unable to get clock
>>
>> and continues, as the clock is optional, depending on the SoC type.
>> This warning may confuse users.
>>
>> To fix this, add a flag to indicate that the clock is mandatory or
>> optional:
>>   - If the clock is mandatory (on R-Car Gen2), a missing clock is now
>>     treated as a fatal error,
>>   - If the clock is optional (on R-Car Gen1), the warning is no longer
>>     printed.
>>
>> The new flag is merged with the existing has_both_edge_trigger boolean
>> into a bitfield to save space.
>
> *DON'T* try to save bits in a megabyte kernel like this.
> The better readability is worth the extra bits or bytes it takes.
> Besides, did you really try to compile it before/after that change
> to verify that it actually saved that space?

Is it better readable?

Size also depends on how many more flags we'll add later.

>> Suggested-by: Magnus Damm <magnus.damm@gmail.com>
>> Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
>> ---
>> This depends on "gpio: rcar: Remove obsolete platform data support".
>
> Where is that patch? This does not apply to my tree,
> probably because of this.

IIRC, you said "include it in your pull request" for that series.

> Please sent the patches in a series.

I can send both to you, as they're not pinctrl but gpio.

>>         unsigned has_both_edge_trigger:1;
>> +       unsigned needs_clk:1;
>
> both should be bool and assigned true/false.

OK, if you prefer it that way.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

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

Re: [PATCH] gpio: rcar: Improve clock error handling and reporting

[Linus Walleij]

On Tue, Dec 1, 2015 at 11:09 AM, Geert Uytterhoeven
<geert@linux-m68k.org> wrote:
> On Tue, Dec 1, 2015 at 10:54 AM, Linus Walleij <linus.walleij@linaro.org> wrote:
>> On Tue, Nov 24, 2015 at 4:05 PM, Geert Uytterhoeven
>> <geert+renesas@glider.be> wrote:

>>> The new flag is merged with the existing has_both_edge_trigger boolean
>>> into a bitfield to save space.
>>
>> *DON'T* try to save bits in a megabyte kernel like this.
>> The better readability is worth the extra bits or bytes it takes.
>> Besides, did you really try to compile it before/after that change
>> to verify that it actually saved that space?
>
> Is it better readable?

OK sorry for being such an ass, I'm a bit stressed out.

> Size also depends on how many more flags we'll add later.

Yeah. But I really like the bools, I think the static checkers can also
make better use of them, but that's just me.

>> Where is that patch? This does not apply to my tree,
>> probably because of this.
>
> IIRC, you said "include it in your pull request" for that series.

OK so this needs to go in with the pin control stuff? Then it's cool...

>> Please sent the patches in a series.
>
> I can send both to you, as they're not pinctrl but gpio.
>
>>>         unsigned has_both_edge_trigger:1;
>>> +       unsigned needs_clk:1;
>>
>> both should be bool and assigned true/false.
>
> OK, if you prefer it that way.

Thanks!

Yours,
Linus Walleij