Re: [RFC PATCH v1] usb/phy add sprd ums512 usbphy

[Baolin Wang]

On 3/18/2023 11:37 PM, Cixi Geng wrote:
> On Mon, 2023-03-13 at 17:11 +0800, Baolin Wang wrote:
> >

...

> > > +
> > > +MODULE_DESCRIPTION("UNISOC USB PHY driver");
> > > +MODULE_LICENSE("GPL");
> > > diff --git a/drivers/usb/phy/phy-sprd-ums512.h
> > > b/drivers/usb/phy/phy-sprd-ums512.h
> > > new file mode 100644
> > > index 000000000000..903da0573eae
> > > --- /dev/null
> > > +++ b/drivers/usb/phy/phy-sprd-ums512.h
> > > @@ -0,0 +1,39 @@
> > > +/* SPDX-License-Identifier: GPL-2.0+ OR MIT */
> > > +/*
> > > + * Spreadtrum UMS512 SOC USB registers file
> > > + *
> > > + * Copyright C 2022, Spreadtrum Communications Inc.
> > > + */
> > > +
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_DATABUS16_8
> > >      0x10000000
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_DMPULLDOWN                 0x8
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_DPPULLDOWN                 0x10
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_ISO_SW_EN                  0x1
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_PS_PD_L                    0x8
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_PS_PD_S                    0x10
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_RESERVED                   0xff
> > > ff
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_TFREGRES                   0x1f
> > > 80000
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_TUNEHSAMP                  0x60
> > > 00000
> > > +#define
> > > MASK_ANLG_PHY_G2_ANALOG_USB20_USB20_VBUSVLDEXT                 0x10
> > > 000
> > > +#define
> > > MASK_ANLG_PHY_G2_DBG_SEL_ANALOG_USB20_USB20_DMPULLDOWN         0x2
> > > +#define
> > > MASK_ANLG_PHY_G2_DBG_SEL_ANALOG_USB20_USB20_DPPULLDOWN         0x4
> > > +#define
> > > MASK_AON_APB_ANA_EB                                            0x10
> > > 00
> > > +#define
> > > MASK_AON_APB_CGM_DPHY_REF_EN                                   0x40
> > > 0
> > > +#define
> > > MASK_AON_APB_CGM_OTG_REF_EN                                    0x10
> > > 00
> > > +#define
> > > MASK_AON_APB_OTG_PHY_SOFT_RST                                  0x20
> > > 0
> > > +#define
> > > MASK_AON_APB_OTG_UTMI_EB                                       0x10
> > > 0
> > > +#define
> > > MASK_AON_APB_OTG_UTMI_SOFT_RST                                 0x10
> > > 0
> > > +#define
> > > MASK_AON_APB_OTG_VBUS_VALID_PHYREG                             0x10
> > > 00000
> > > +#define
> > > MASK_AON_APB_USB2_PHY_IDDIG                                    0x8
> > > +#define
> > > MASK_AON_APB_UTMI_WIDTH_SEL                                    0x40
> > > 000000
> > > +#define
> > > REG_ANLG_PHY_G2_ANALOG_USB20_USB20_BATTER_PLL                  0x00
> > > 5c
> > > +#define
> > > REG_ANLG_PHY_G2_ANALOG_USB20_USB20_ISO_SW                      0x00
> > > 70
> > > +#define
> > > REG_ANLG_PHY_G2_ANALOG_USB20_USB20_TRIMMING                    0x00
> > > 64
> > > +#define
> > > REG_ANLG_PHY_G2_ANALOG_USB20_USB20_UTMI_CTL1                   0x00
> > > 58
> > > +#define
> > > REG_ANLG_PHY_G2_ANALOG_USB20_USB20_UTMI_CTL2                   0x00
> > > 60
> > > +#define
> > > REG_ANLG_PHY_G2_ANALOG_USB20_REG_SEL_CFG_0                     0x00
> > > 74
> > > +#define
> > > REG_AON_APB_APB_EB1                                            0x00
> > > 04
> > > +#define
> > > REG_AON_APB_APB_RST1                                           0x00
> > > 10
> > > +#define
> > > REG_AON_APB_CGM_REG1                                           0x01
> > > 38
> > > +#define
> > > REG_AON_APB_OTG_PHY_CTRL                                       0x02
> > > 08
> > > +#define
> > > REG_AON_APB_OTG_PHY_TEST                                       0x02
> > > 04
> >
> > Move them to the driver file and please rename the ugly macro names,
> > too
> > long :(
> >
> > And why not move the usb phy driver to be a generic phy driver? I
> > mean
> > move it to the drivers/phy.
> Do you mean all the usb-phy need move to driver/phy? or just this
> driver?

If I remember correctly, new phy drivers should go to the generic phy 
framework unless there are some reasons to prevent.

Greg, how do you think?