[PATCH v7 2/5] phy: add a driver for the Rockchip SoC internal USB2.0 PHY

[Doug Anderson]

Hi,

On Fri, Dec 12, 2014 at 11:24 PM, Kishon Vijay Abraham I <kishon at ti.com> wrote:
> hi,
>
> On Saturday 13 December 2014 05:49 AM, Doug Anderson wrote:
>> Yunzhi,
>>
>> On Fri, Dec 12, 2014 at 7:07 AM, Yunzhi Li <lyz at rock-chips.com> 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>
>>>
>>> ---
>>>
>>> Changes in v7:
>>> - Accept Kishon's comments to use phandle args to find a phy
>>>   struct directly and get rid of using a custom of_xlate
>>>   function.
>>
>> I'm going to assume you didn't test this version, since it doesn't
>> work for me.  At suspend time power is high and my printouts in the
>> powerup/powerdown code aren't called...
>>
>>
>>> +       for_each_available_child_of_node(dev->of_node, child) {
>>> +               rk_phy = devm_kzalloc(dev, sizeof(*rk_phy), GFP_KERNEL);
>>> +               if (!rk_phy)
>>> +                       return -ENOMEM;
>>> +
>>> +               if (of_property_read_u32(child, "reg", &reg_offset)) {
>>> +                       dev_err(dev, "missing reg property in node %s\n",
>>> +                               child->name);
>>> +                       return -EINVAL;
>>> +               }
>>> +
>>> +               rk_phy->reg_offset = reg_offset;
>>> +               rk_phy->reg_base = grf;
>>> +
>>> +               rk_phy->clk = of_clk_get_by_name(child, "phyclk");
>>> +               if (IS_ERR(rk_phy->clk))
>>> +                       rk_phy->clk = NULL;
>>> +
>>> +               rk_phy->phy = devm_phy_create(dev, child, &ops);
>>> +               if (IS_ERR(rk_phy->phy)) {
>>> +                       dev_err(dev, "failed to create PHY\n");
>>> +                       return PTR_ERR(rk_phy->phy);
>>> +               }
>>> +               phy_set_drvdata(rk_phy->phy, rk_phy);
>>> +       }
>>> +
>>> +       phy_provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate);
>>
>> I think your bug is here.  I think you now need to register 3 phy
>> providers, not just one.
>
> No there should be only one phy provider. It means the bug is elsewhere.

Ah.  That's what I get for testing on a backported kernel.  I bet it's
because I'm missing:

2a4c370 phy: core: Fix of_phy_provider_lookup to return PHY provider
for sub node

Hrm, that made things better, but I still only got one printout when I
expected 3 (one for each user of the PHY).  I bet there are more picks
I need then...  :-/  Ah, yup.  When I pick the whole load of PHY
related stuff then I get all 3.  :)

I'll do more testing when I have more time and post up a Tested-by, then...

-Doug