[PATCH v5 1/5] phy: add a driver for the Rockchip SoC internal USB2.0 PHY

Kishon Vijay Abraham I kishon at ti.com
Thu Dec 11 00:36:21 PST 2014


Hi,

On Thursday 11 December 2014 01:14 PM, Yunzhi Li wrote:
> Hi Kishon:
> 
> On 2014/12/11 14:02, Kishon Vijay Abraham I wrote:
>> Hi,
>>
>> On Wednesday 10 December 2014 04:16 PM, Yunzhi Li wrote:
>>> This patch to add a generic PHY driver for ROCKCHIP usb PHYs,
>>> currently this driver can support RK3288. The RK3288 SoC have
>>> three independent USB PHY IPs which are all configured through a
>>> set of registers located in the GRF (general register files)
>>> module.
>>>
>>> Signed-off-by: Yunzhi Li <lyz at rock-chips.com>
>>>
>>> +
>>> +#define ROCKCHIP_RK3288_UOC(n)    (0x320 + n * 0x14)
>>> +
>>> +/*
>>> + * The higher 16-bit of this register is used for write protection
>>> + * only if BIT(13 + 16) set to 1 the BIT(13) can be written.
>>> + */
>>> +#define SIDDQ_MSK        BIT(13 + 16)
>> I think here the "MSK" is misleading. it should be something that refers write
>> protection?
> So, #define SIDDQ_WRITE_ENA  BIT(29) , could be ok ?

yeah. should be fine.
>>> +#define SIDDQ_ON        BIT(13)
>>> +#define SIDDQ_OFF        (0 << 13)
>>> +
>>> +struct rockchip_usb_phy {
>>> +    struct regmap    *reg_base;
>>> +    unsigned int    reg_offset;
>>> +    struct clk    *clk;
>>> +    struct phy    *phy;
>>> +    unsigned    index;
>>> +};
>>> +
>>> +struct rockchip_usb_phy_priv {
>>> +    struct rockchip_usb_phy    *phys;
>>> +    unsigned        nphys;
>>> +};
>>> +
>>> +static int rockchip_usb_phy_power(struct rockchip_usb_phy *phy,
>>> +                       bool siddq)
>>> +{
>>> +    return regmap_write(phy->reg_base, phy->reg_offset,
>>> +                SIDDQ_MSK | (siddq ? SIDDQ_ON : SIDDQ_OFF));
>> Shouldn't we actually reset the bit for power off?
> Sorry, which bit you refer to here  and why should it be reset? could you give
> more infomation.

Never mind. Initially was thinking it might affect bits other than siddq but
anyway you have write lock. So should be fine.

Thanks
Kishon



More information about the Linux-rockchip mailing list