RE: [PATCH] phy: rcar-gen3-usb2: Add R-Car Gen3 USB2 PHY driver

[Yoshihiro Shimoda]

Hi Geert-san,

> Sent: Monday, August 24, 2015 9:13 PM
> 
> Hi Shimoda-san,
> 
> On Thu, Aug 20, 2015 at 8:28 AM, Yoshihiro Shimoda
> <yoshihiro.shimoda.uh@xxxxxxxxxxx> wrote:
> > This patch adds support for R-Car generation 3 USB2 PHY driver.
> > This SoC has 3 EHCI/OHCI channels, and the channel 0 is shared
> > with the HSUSB (USB2.0 peripheral) device.
> >
> > So, the purpose of this driver is:
> >  1) initializes some registers of SoC specific to use the
> >     {ehci,ohci}-platform driver.
> >
> >  2) detects id pin to select host or peripheral on the channel 0.
> >
> > For now, this driver only supports 1) above.
> 
> Thanks for your patch!

Thank you very much for your review!

> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/phy/rcar-gen3-phy-usb2.txt
> > @@ -0,0 +1,37 @@
> > +* Renesas R-Car generation 3 USB 2.0 PHY
> > +
> > +This file provides information on what the device node for the R-Car generation
> > +3 USB 2.0 PHY contains.
> > +
> > +Required properties:
> > +- compatible: "renesas,usb2-phy-r8a7795" if the device is a part of R8A7795 SoC.
> 
> "an R8A7795 SoC".

I will fix this in v2.
(I will submit v2 patch late today.)

> > +- reg: offset and length of the USB2.0 host register block.
> > +- reg-names: must be "usb2".
> > +- clocks: clock phandle and specifier pair.
> 
> "pair(s)."

I will fix this.

> > +- clock-names: string, clock input name, must be "usb2", and optional "hsusb".
> 
> names

Thank you for the point.
As you mentioned about runtime PM below, I will remove this property.

> > +- #phy-cells: see phy-bindings.txt in the same directory, must be <0>.
> > +
> > +Optional proparies:
> 
> properties

Oops, I will fix it.
I should have done aspell command before I submitted.
If I did that command, I found other typo in this patch :)

> > +To use a USB channel which EHCI/OHCI and HSUSB are combined, the device tree
> 
> s/which/where/
> 
> > +node should set HSUSB proparies to reg and reg-names proparies:
> 
> properties (x2)

I will fix it.

> > +- reg: offset and length of the HSUSB register block.
> > +- reg-names: must be "hsusb".
> > +
> > +Example (R-Car H3):
> > +
> > +       usb-phy@ee080200 {
> > +               compatible = "renesas,usb2-phy-r8a7795";
> > +               reg = <0 0xee080200 0 0x6ff>, <0 0xe6590100 0 0x100>;
> > +               reg-names = "usb2", "hsusb";
> > +               clocks = <&mstp7_clks R8A7795_CLK_EHCI0>,
> > +                        <&mstp7_clks R8A7795_CLK_HSUSB>;
> 
> If you ever want to enable Runtime PM, do you want the first MSTP clock to be
> managed automatically?
> What about the second optional clock?

Thank you very much for the point.
I didn't take care about Runtime PM. However, I want to enable Runtime PM
because usb host and function drivers support it.

> > +               clock-names = "usb2", "hsusb";
> > +       };
> > +
> > +       usb-phy@ee0a0200 {
> > +               compatible = "renesas,usb2-phy-r8a7795";
> > +               reg = <0 0xee0a0200 0 0x6ff>;
> > +               reg-names = "usb2";
> > +               clocks = <&mstp7_clks R8A7795_CLK_EHCI0>;
> > +               clock-names = "usb2";
> > +       };
> 
> > --- /dev/null
> > +++ b/drivers/phy/phy-rcar-gen3-usb2.c
> 
> > +static int rcar_gen3_phy_usb2_init(struct phy *p)
> > +{
> > +       struct rcar_gen3_phy_usb2_channel *channel = phy_get_drvdata(p);
> > +       unsigned long flags;
> > +       void __iomem *usb2_base = channel->usb2.base;
> > +       void __iomem *hsusb_base = channel->hsusb.base;
> > +       u32 tmp;
> > +
> > +       /* Since ops->init() is called once, this driver enables both clocks */
> > +       clk_prepare_enable(channel->usb2.clk);
> > +       clk_prepare_enable(channel->hsusb.clk);
> 
> So this driver manages its clock(s) itself, and doesn't support Runtime PM.

So, I will use remove clk APIs in this driver.

Best regards,
Yoshihiro Shimoda

> > +static int rcar_gen3_phy_usb2_exit(struct phy *p)
> > +{
> > +       struct rcar_gen3_phy_usb2_channel *channel = phy_get_drvdata(p);
> > +
> > +       writel(0, channel->usb2.base + USB2_INT_ENABLE);
> > +
> > +       clk_disable_unprepare(channel->hsusb.clk);
> > +       clk_disable_unprepare(channel->usb2.clk);
> > +
> > +       return 0;
> > +}
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
> 
> 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
N‹§²æ¸›yú²X¬¶ÇvØ–)Þ{.nlj·¥Š{±‘êX§¶›¡Ü}©ž²ÆzÚj:+v‰¨¾«‘êZ+€Êzf£¢·hšˆ§~†­†Ûÿû®w¥¢¸?™¨è&¢)ßf”ùy§m…á«a¶Úÿ0¶ìå