Re: [PATCH v3] USB: PHY: Palmas USB Transceiver Driver

[Graeme Gregory <gg@slimlogic.co.uk>]

On 26/03/13 06:03, Kishon Vijay Abraham I wrote:
> Hi,
>
> On Monday 25 March 2013 03:16 PM, Laxman Dewangan wrote:
>> On Monday 25 March 2013 03:02 PM, Kishon Vijay Abraham I wrote:
>>> From: Graeme Gregory <gg@slimlogic.co.uk>
>>>
>>> This is the driver for the OTG transceiver built into the Palmas
>>> chip. It
>>> handles the various USB OTG events that can be generated by cable
>>> insertion/removal.
>>>
>>> Signed-off-by: Graeme Gregory <gg@slimlogic.co.uk>
>>> Signed-off-by: Moiz Sonasath <m-sonasath@ti.com>
>>> Signed-off-by: Ruchika Kharwar <ruchika@ti.com>
>>> Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
>>> Signed-off-by: Sebastien Guiriec <s-guiriec@ti.com>
>>> ---
>>
>> I think this driver is more over the cable connection like vbus
>> detetcion or ID pin detection.
>> Then why not it is implemented based on extcon framework?
>
> extcon framework uses notification mechanism and Felipe dint like
> using notification here. right Felipe?
>>
>> That way, generic usb driver (like tegra_usb driver) will get
>> notification through extcon.
>>
>> We need this cable detection through extcon on our tegra solution
>> through the Palmas.
>>
>>
>> +#include <linux/of.h>
>> +#include <linux/of_platform.h>
>> +
>> +static int palmas_usb_read(struct palmas *palmas, unsigned int reg,
>> +               unsigned int *dest)
>> +{
>> +       unsigned int addr;
>> +       int slave;
>> +
>> +       slave = PALMAS_BASE_TO_SLAVE(PALMAS_USB_OTG_BASE);
>> +       addr = PALMAS_BASE_TO_REG(PALMAS_USB_OTG_BASE, reg);
>> +
>> +       return regmap_read(palmas->regmap[slave], addr, dest);
>>
>>
>> Please use the generic api for palmas_read()/palmas_write(0 as it will
>> be ease on debugging on register access.
>> Direct regmap_read() does not help much on this.
>
> Graeme,
> Any reason why you dint use palmas_read()/palmas_write here?
> Btw palmas_read()/palmas_write() internally uses regmap APIs.
Because I was not a fan of tightly coupling the child devices to the
parent MFD. palmas_read/write were added by Laxman.

>>
>>> +}
>>> +
>>> +static int palmas_usb_write(struct palmas *palmas, unsigned int reg,
>>> +               unsigned int data)
>>> +{
>>> +       unsigned int addr;
>>> +       int slave;
>>> +
>>> +       slave = PALMAS_BASE_TO_SLAVE(PALMAS_USB_OTG_BASE);
>>> +       addr = PALMAS_BASE_TO_REG(PALMAS_USB_OTG_BASE, reg);
>>> +
>>> +       return regmap_write(palmas->regmap[slave], addr, data);
>>
>> Same as above.
>>
>>
>>
>>> +
>>> +       if (status != OMAP_DWC3_UNKNOWN) {
>>> +               palmas_usb->linkstat = status;
>>> +               palmas_usb->mailboxstat = dwc3_omap_mailbox(status);
>> Omap specific call, why? This is generic palma driver.
>
> hmm.. I think we should either fall-back to the notification mechanism
> or have the client drivers pass function pointer here. Felipe?
>>
>>
>>> +
>>
>>> +       palmas_usb->dev         = &pdev->dev;
>>> +
>>> +       palmas_usb->irq1 = regmap_irq_get_virq(palmas->irq_data,
>>> +                                               PALMAS_ID_OTG_IRQ);
>>> +       palmas_usb->irq2 = regmap_irq_get_virq(palmas->irq_data,
>>> +                                               PALMAS_ID_IRQ);
>>> +       palmas_usb->irq3 = regmap_irq_get_virq(palmas->irq_data,
>>> +                                               PALMAS_VBUS_OTG_IRQ);
>>> +       palmas_usb->irq4 = regmap_irq_get_virq(palmas->irq_data,
>>> +                                               PALMAS_VBUS_IRQ);
>>
>> Should be come from platform_get_irq() through platform driver.
>
> No. It can be obtained from regmap too.
>
> Thanks
> Kishon