DRV260XSW-LINUX: drv260x_resume()

Hello Yaita, 

I am consulting our firmware team on this and will respond to you as soon as possible. 

Best Regards, 

Justin Beigel

Hi Justin-san,

Thank you for your support.
I will wait for your feedback.

Best Regards,
Yaita

Yes, it is. I wonder which drv260x you mean, drv2603 or drv2604?

Hi Shenghao-san,

Thank you for your reply.
My customer uses DRV2604L and wonders if the code is correct.


The datasheet shows we should set EN=1 first then STANDBY=0 when SHUTDOWN to ACTIVE.
However, it seems the code shows the opposite sequence.
Is the code correct?

Best Regards,
Yaita

Hi Shenghao-san,

Have you looked into the inquiry?
If so, may I have your comments?

Best Regards,
Yaita

Hi Shenghao-san,

I gave notice of your comment to my customer however he didn't accept it.

Regarding drv260x_suspend(), it shows enable_gpio is set to 0 and transits to "Shutdown" state.
So drv260x_resume() should be resume from the "Shutdown" state.

He believes the following sequence is correct for drv260x_resume().
-------------------------------------------------
1) enable_gpio is set to 1: Shutdown -> Standby
2) STANDBY register is set to 0: Standby -> Active
-------------------------------------------------
He believes drv260x.c shows the opposite sequence.
Could you look into it again?

Best Regards,
Yaita

Hi Shenghao-san,

May I have your comments?
I have to reply to my customer.

Best Regards,
Yaita

Hi,

If EN set to 0, the ram bin will be lost. So during suspend, the code won't set EN to 0. Suspend is not shutdown.

Hi Shenghao-san,

Regarding drv260x_suspend(), it shows enable_gpio is set to 0 and transits to "Shutdown" state.
How should I think about that? Is the code incorrect?

Best Regards,
Yaita

during suspend, do not any handing on GPIO to the chip. There's no GPIO handling to chip in our sample code

Hi Shenghao-san,

It seems there is a difference between the image you attached and the image we saw.
Where is the code you attached?

We see the following sample code.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/input/misc/drv260x.c
The link above can be accessed at the following address.
https://www.ti.com/tool/DRV260XSW-LINUX

Best Regards,
Yaita

Hi Shenghao-san,

May I have your reply?

Best Regards,
Yaita

May I have your reply?

Best Regards,
Yaita

Hi Yaita-san

Do you have idea of who submit the code to kernel.org? If Ti guy did, I could have asked him the reason.  Usually, we ti guys only support the code from www.ti.com. I insist on the handling from www.ti.com, because, in theory, GPIO to chip won't be any change during suspending, or it will make chip reset.

Hi Shenghao-san,

I don't know who submit the code to kernel.org.
However, I believe TI members submit the code because the code can be accessed from the following ti.com.
https://www.ti.com/tool/DRV260XSW-LINUX

Could you tell me how to access the following code that you attached?



Best Regards,
Yaita

Hi Shenghao-san,

Please tell me where is the code you attached.
Can we access the code?



Best Regards,
Yaita

Hi Yaita-san

The link is as following:

CGit @ Texas Instruments - Open Source Git Repositories - git.TI.com/cgit

As to the drv260x code from Linux Kernel, the guy upload to the kernel branch has already left.

I will check further from other guys.

As you know, drv2604 and drv2605 are different on the ram firmware.

For drv2604, I will check whether the register setting and firmware will be lost if enable pin set to low.

For drv2605, I will check whether the register setting will be lost if enable pin set to low.

If so, the code can't set enable pin during suspend. Thanks.

Hi Shenghao-san,

After all, I don't know what the conclusion is.



The datasheet shows we should set EN=1 first then STANDBY=0 when SHUTDOWN to ACTIVE.
However, it seems the following code shows the opposite sequence.
Did you say the following code is incorrect?

----------------------------------------
static int __maybe_unused drv260x_resume(struct device *dev)
{
struct drv260x_data *haptics = dev_get_drvdata(dev);
int ret = 0;

mutex_lock(&haptics->input_dev->mutex);

if (input_device_enabled(haptics->input_dev)) {
ret = regulator_enable(haptics->regulator);
if (ret) {
dev_err(dev, "Failed to enable regulator\n");
goto out;
}

ret = regmap_update_bits(haptics->regmap,
DRV260X_MODE,
DRV260X_STANDBY_MASK, 0);
if (ret) {
dev_err(dev, "Failed to unset standby mode\n");
regulator_disable(haptics->regulator);
goto out;
}

gpiod_set_value(haptics->enable_gpio, 1);
}

out:
mutex_unlock(&haptics->input_dev->mutex);
return ret;
}
----------------------------------------
Best Regards,
Yaita

So sorry that such diagram made you confusion. In order to save power consumption, the code will disable the EN pin in suspend mode, just like shutdown mode.

Shenghao-san,

I understand we can transit Shutdown to Active when STANDBY=0 first then EN=1.
Is my understanding correct?

Best Regards,
Yaita