Re: [PATCH v4 07/14] mfd: Add driver for Maxim 77802 Power Management IC

From: Doug Anderson
Date: Thu Jun 26 2014 - 12:29:59 EST


Javier,

On Thu, Jun 26, 2014 at 9:18 AM, Javier Martinez Canillas
<javier.martinez@xxxxxxxxxxxxxxx> wrote:
> Hello Doug,
>
> On 06/26/2014 06:12 PM, Doug Anderson wrote:
>> Javier,
>>
>> On Thu, Jun 26, 2014 at 4:13 AM, Javier Martinez Canillas
>> <javier.martinez@xxxxxxxxxxxxxxx> wrote:
>>>>> +
>>>>> +#ifdef CONFIG_PM_SLEEP
>>>>> +static int max77802_suspend(struct device *dev)
>>>>> +{
>>>>> + struct i2c_client *i2c = container_of(dev, struct i2c_client, dev);
>>>>> + struct max77802_dev *max77802 = i2c_get_clientdata(i2c);
>>>>> +
>>>>> + if (device_may_wakeup(dev))
>>>>> + enable_irq_wake(max77802->irq);
>>>>> +
>>>>> + disable_irq(max77802->irq);
>>>>
>>>> Can you add short comment why this is needed? I know why but just for
>>>> future generations which will wonder: "why do we need to disable the IRQ
>>>> while suspending?" :). Especially that this is rather a workaround for
>>>> issue in other driver (I2C bus).
>>>>
>>>
>>> Good idea, I'll add a comment here on next version so code archaeologists can
>>> figure out what what's going on here.
>>
>> Is the disable_irq() needed if you have
>> <https://patchwork.kernel.org/patch/4421891/>?
>>
>
> Probably not but I added the following comment:
>
> /*
> * The IRQ must be disabled during suspend since due wakeup
> * ordering issues it may be possible that the I2C controller
> * is still suspended when the interrupt happens so the IRQ
> * handler will fail to read the I2C bus.
> */
> disable_irq(max77802->irq);
>
> since in theory this PMIC can be used in other SoCs besides
> Exynos5420/Exynos5800 and it may be possible that the I2C controller driver for
> these other SoCs may not resume at noirq time.
>
> But on a second thought, this PMIC seems to be designed specially for these two
> Exynos SoCs so I guess it's safe to assume that it is not needed?

Right, there's a close coupling between PMICs and SoCs. The PMIC has
special sequencing and default voltage levels that were tuned exactly
for this SoC.

Note: Wolfram has not actually responded to my patch much less
accepted it. It's entirely possible that in another month or two
we'll hear back a big fat NAK. In that case your solution will be the
best one I can think of.


-Doug
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/