[U-Boot] [PATCH 2/5] driver: usb: add EHCI driver for hi3787cv200 SoC

[Marek Vasut]

On 05/08/2017 11:15 AM, Jorge Ramirez wrote:
> On 05/07/2017 02:34 PM, Marek Vasut wrote:
>> On 05/07/2017 12:56 PM, Jorge Ramirez wrote:
>>> On 05/06/2017 01:05 AM, Marek Vasut wrote:
>>>> On 05/05/2017 07:55 PM, Jorge Ramirez wrote:
>>>>> On 05/05/2017 05:34 PM, Marek Vasut wrote:
>>>>>> On 05/05/2017 04:32 PM, Tom Rini wrote:
>>>>>>> On Thu, May 04, 2017 at 03:47:07PM +0200, Jorge Ramirez-Ortiz wrote:
>>>>>>>
>>>>>>> CC'ing Marek...
>>>>>>>
>>>>>>>> Signed-off-by: Jorge Ramirez-Ortiz <jorge.ramirez-ortiz at linaro.org>
>>>>>>>> ---
>>>>>>>>     .../arm/include/asm/arch-hi3798cv200/hi3798cv200.h |  93
>>>>>>>> ++++++++++
>>>>>>>>     drivers/usb/host/Kconfig                           |   6 +
>>>>>>>>     drivers/usb/host/Makefile                          |   1 +
>>>>>>>>     drivers/usb/host/ehci-hi3798cv200.c                | 196
>>>>>>>> +++++++++++++++++++++
>>>>>>>>     4 files changed, 296 insertions(+)
>>>>>>>>     create mode 100644
>>>>>>>> arch/arm/include/asm/arch-hi3798cv200/hi3798cv200.h
>>>>>>>>     create mode 100644 drivers/usb/host/ehci-hi3798cv200.c
>>>>>>>>
>>>>>>>> diff --git a/arch/arm/include/asm/arch-hi3798cv200/hi3798cv200.h
>>>>>>>> b/arch/arm/include/asm/arch-hi3798cv200/hi3798cv200.h
>>>>>>>> new file mode 100644
>>>>>>>> index 0000000..c67fda1
>>>>>>>> --- /dev/null
>>>>>>>> +++ b/arch/arm/include/asm/arch-hi3798cv200/hi3798cv200.h
>>>>>>>> @@ -0,0 +1,93 @@
>>>>>>>> +/*
>>>>>>>> + * (C) Copyright 2017 Linaro
>>>>>>>> + * Jorge Ramirez-Ortiz <jorge.ramirez-ortiz at linaro.org>
>>>>>>>> + *
>>>>>>>> + * SPDX-License-Identifier:    GPL-2.0+
>>>>>>>> + */
>>>>>>>> +
>>>>>>>> +#ifndef __HI3798cv200_H__
>>>>>>>> +#define __HI3798cv200_H__
>>>>>> This should be a separate patch ..
>>>>> hi Marek
>>>>>
>>>>> to be clear, are you asking a single patch containing hi3798cv200.h?
>>>> Yes, this is arch stuff and is not in any way part of the USB driver
>>>> IMO.
>>> sure.
>>> I suppose that what are really saying is that the usb driver has to be
>>> upstreamed after the architecture and not before if such a dependency
>>> exists (ie not using the device tree)
>> Yes, you can be asked to change stuff in the arch code , at which point
>> you might have to modify the (dead, but accepted) usb code.
>>
>> If you can move arch stuff to arch code, even better. Then you might
>> even be able to use ehci-generic.
> 
> yes, that is what I ended up doing. thanks for the suggestion.
> 
>>
>>>> [...]
>>>>
>>>>>>>> +};
>>>>>>>> +
>>>>>>>> +#endif
>>>>>>>> diff --git a/drivers/usb/host/Kconfig b/drivers/usb/host/Kconfig
>>>>>>>> index 0bf8274..a749d0e 100644
>>>>>>>> --- a/drivers/usb/host/Kconfig
>>>>>>>> +++ b/drivers/usb/host/Kconfig
>>>>>>>> @@ -128,6 +128,12 @@ config USB_EHCI_ZYNQ
>>>>>>>>         ---help---
>>>>>>>>           Enable support for Zynq on-chip EHCI USB controller
>>>>>>>>     +config USB_EHCI_POPLAR
>>>>>>>> +    bool "Support for HI3798cv200 EHCI USB controller"
>>>>>> Does HI stand for Hitachi ?
>>>>> Hisilicon
>>>> Don't we already have some hisilicon stuff in the tree ? Maybe some
>>>> 96board ?
>>>>
>>>> In fact, could this driver be replaced by ehci-generic ?
>>> yes I think so. the driver is basically platform initialization code and
>>> the generic driver.
>>> the problem I have is that there is not usb host support in the kernel
>>> for it so we dont have a device node in the tree.
>>>
>>> should I commit the linux kernel device tree (first patch in the series)
>>> and then add the node when I push the DM usb driver?
>> Ideally check the bindings with Rob Herring, otherwise I don't care.
>> If you contribute to kernel too, awesome.
> 
> in v2 I pushed the bindings to allow me to use the ehci-generic driver.
> 
> the kernel team working on this platform will push a more detailed node
> definition to korg - usb host is still under development - at which
> point I'll merge it back.
> I hope that this is acceptable?

Works for me if you don't invent weird stuff in the bindings and keep
them reasonably in sync. Thanks

-- 
Best regards,
Marek Vasut