Re: [PATCH usb-next] usb: dwc3: Use devm_of_platform_populate

[Jisheng Zhang]

On Mon, 9 Nov 2020 10:34:14 + Russell King - ARM Linux admin  wrote:

> 
> 
> On Mon, Nov 09, 2020 at 09:59:53AM +0800, Jisheng Zhang wrote:
> > diff --git a/drivers/usb/dwc3/dwc3-meson-g12a.c 
> > b/drivers/usb/dwc3/dwc3-meson-g12a.c
> > index 417e05381b5d..83015bb7b926 100644
> > --- a/drivers/usb/dwc3/dwc3-meson-g12a.c
> > +++ b/drivers/usb/dwc3/dwc3-meson-g12a.c
> > @@ -702,7 +702,6 @@ static int dwc3_meson_g12a_probe(struct platform_device 
> > *pdev)
> >  {
> >   struct dwc3_meson_g12a  *priv;
> >   struct device   *dev = >dev;
> > - struct device_node  *np = dev->of_node;
> >   void __iomem *base;
> >   int ret, i;
> >
> > @@ -794,7 +793,7 @@ static int dwc3_meson_g12a_probe(struct platform_device 
> > *pdev)
> >   goto err_phys_power;
> >   }
> >
> > - ret = of_platform_populate(np, NULL, NULL, dev);
> > + ret = devm_of_platform_populate(dev);
> >   if (ret)
> >   goto err_phys_power;
> >
> > @@ -832,8 +831,6 @@ static int dwc3_meson_g12a_remove(struct 
> > platform_device *pdev)
> >   if (priv->drvdata->otg_switch_supported)
> >   usb_role_switch_unregister(priv->role_switch);
> >
> > - of_platform_depopulate(dev);
> > -
> >   for (i = 0 ; i < PHY_COUNT ; ++i) {
> >   phy_power_off(priv->phys[i]);
> >   phy_exit(priv->phys[i]);  
> 
> Does it matter that the order that things happen in
> dwc3_meson_g12a_remove() is changed as a result of your patch? Was
> the code relying on the platform devices being depopulated before
> powering off the PHYs?

...

> 
> Same here... and for anywhere else in this patch that you're deleting
> a of_platform_depopulate().

oops, good question! I believe it's not safe to move the depopulate after
the phy_* APIs.

So please ignore this patch.

> 
> You effectively are moving the call to of_platform_depopulate() *after*
> the driver's .remove function has been called.
> 
> --
>

Re: [PATCH usb-next] usb: dwc3: Use devm_of_platform_populate

[Russell King - ARM Linux admin]

On Mon, Nov 09, 2020 at 09:59:53AM +0800, Jisheng Zhang wrote:
> diff --git a/drivers/usb/dwc3/dwc3-meson-g12a.c 
> b/drivers/usb/dwc3/dwc3-meson-g12a.c
> index 417e05381b5d..83015bb7b926 100644
> --- a/drivers/usb/dwc3/dwc3-meson-g12a.c
> +++ b/drivers/usb/dwc3/dwc3-meson-g12a.c
> @@ -702,7 +702,6 @@ static int dwc3_meson_g12a_probe(struct platform_device 
> *pdev)
>  {
>   struct dwc3_meson_g12a  *priv;
>   struct device   *dev = >dev;
> - struct device_node  *np = dev->of_node;
>   void __iomem *base;
>   int ret, i;
>  
> @@ -794,7 +793,7 @@ static int dwc3_meson_g12a_probe(struct platform_device 
> *pdev)
>   goto err_phys_power;
>   }
>  
> - ret = of_platform_populate(np, NULL, NULL, dev);
> + ret = devm_of_platform_populate(dev);
>   if (ret)
>   goto err_phys_power;
>  
> @@ -832,8 +831,6 @@ static int dwc3_meson_g12a_remove(struct platform_device 
> *pdev)
>   if (priv->drvdata->otg_switch_supported)
>   usb_role_switch_unregister(priv->role_switch);
>  
> - of_platform_depopulate(dev);
> -
>   for (i = 0 ; i < PHY_COUNT ; ++i) {
>   phy_power_off(priv->phys[i]);
>   phy_exit(priv->phys[i]);

Does it matter that the order that things happen in
dwc3_meson_g12a_remove() is changed as a result of your patch? Was
the code relying on the platform devices being depopulated before
powering off the PHYs?

> diff --git a/drivers/usb/dwc3/dwc3-of-simple.c 
> b/drivers/usb/dwc3/dwc3-of-simple.c
> index e62ecd22b3ed..f1c267e39d62 100644
> --- a/drivers/usb/dwc3/dwc3-of-simple.c
> +++ b/drivers/usb/dwc3/dwc3-of-simple.c
> @@ -73,7 +73,7 @@ static int dwc3_of_simple_probe(struct platform_device 
> *pdev)
>   if (ret)
>   goto err_resetc_assert;
>  
> - ret = of_platform_populate(np, NULL, NULL, dev);
> + ret = devm_of_platform_populate(dev);
>   if (ret)
>   goto err_clk_put;
>  
> @@ -97,8 +97,6 @@ static int dwc3_of_simple_probe(struct platform_device 
> *pdev)
>  
>  static void __dwc3_of_simple_teardown(struct dwc3_of_simple *simple)
>  {
> - of_platform_depopulate(simple->dev);
> -
>   clk_bulk_disable_unprepare(simple->num_clocks, simple->clks);
>   clk_bulk_put_all(simple->num_clocks, simple->clks);
>   simple->num_clocks = 0;

Same here... and for anywhere else in this patch that you're deleting
a of_platform_depopulate().

You effectively are moving the call to of_platform_depopulate() *after*
the driver's .remove function has been called.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!