Re: [PATCH v13 5/6] usb: dwc3: qcom: Keep power domain on to retain controller status

[Matthias Kaehlcke <mka@chromium.org>]

On Tue, Apr 12, 2022 at 12:46:53AM +0530, Sandeep Maheswaram wrote:
> Keep the power domain always on during runtime suspend or if the
> controller supports wakeup in order to retain controller status
> and to support wakeup from devices.
> 
> Signed-off-by: Sandeep Maheswaram <quic_c_sanm@quicinc.com>
> ---
>  drivers/usb/dwc3/dwc3-qcom.c | 11 ++++++++++-
>  1 file changed, 10 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/dwc3/dwc3-qcom.c b/drivers/usb/dwc3/dwc3-qcom.c
> index 9804a19..9747be6 100644
> --- a/drivers/usb/dwc3/dwc3-qcom.c
> +++ b/drivers/usb/dwc3/dwc3-qcom.c
> @@ -17,6 +17,7 @@
>  #include <linux/of_platform.h>
>  #include <linux/platform_device.h>
>  #include <linux/phy/phy.h>
> +#include <linux/pm_domain.h>
>  #include <linux/usb/of.h>
>  #include <linux/reset.h>
>  #include <linux/iopoll.h>
> @@ -724,6 +725,7 @@ static int dwc3_qcom_probe(struct platform_device *pdev)
>  	struct resource		*res, *parent_res = NULL;
>  	int			ret, i;
>  	bool			ignore_pipe_clk;
> +	struct generic_pm_domain *genpd;

nit: I'm not really a fan of gazillions of whitespaces between the type
and the variable name, but if they are kept all variable names above
should move a tab to the right. In any case the style in this file isn't
consistent, so an alternative would be to just get rid of the alignment
completely.

>  	qcom = devm_kzalloc(&pdev->dev, sizeof(*qcom), GFP_KERNEL);
>  	if (!qcom)
> @@ -732,6 +734,8 @@ static int dwc3_qcom_probe(struct platform_device *pdev)
>  	platform_set_drvdata(pdev, qcom);
>  	qcom->dev = &pdev->dev;
>  
> +	genpd = pd_to_genpd(qcom->dev->pm_domain);
> +
>  	if (has_acpi_companion(dev)) {
>  		qcom->acpi_pdata = acpi_device_get_match_data(dev);
>  		if (!qcom->acpi_pdata) {
> @@ -839,7 +843,12 @@ static int dwc3_qcom_probe(struct platform_device *pdev)
>  	if (ret)
>  		goto interconnect_exit;
>  
> -	device_init_wakeup(&pdev->dev, 1);
> +	genpd->flags |= GENPD_FLAG_RPM_ALWAYS_ON;
> +
> +	if (device_may_wakeup(&qcom->dwc3->dev))
> +		genpd->flags |= GENPD_FLAG_ALWAYS_ON;

In v12 only GENPD_FLAG_ALWAYS_ON was set, not GENPD_FLAG_RPM_ALWAYS_ON.
I asked a few times for a change log, in this instance it would be
useful to explain why GENPD_FLAG_RPM_ALWAYS_ON is now set.

> +
> +	device_init_wakeup(&pdev->dev, device_may_wakeup(&qcom->dwc3->dev));

device_may_wakeup() isn't an expensive operation, but it's still not nice to
call it twice in three lines of code. Instead you could do this:

	if (device_may_wakeup(&qcom->dwc3->dev)) {
		genpd->flags |= GENPD_FLAG_ALWAYS_ON;
		device_init_wakeup(&pdev->dev, true);
	}