Re: [PATCH RFC 2/2] mfd: add Renesas RPC-IF driver

[Sergei Shtylyov <sergei.shtylyov@xxxxxxxxxxxxxxxxxx>]

On 06/19/2019 09:18 AM, Lee Jones wrote:

>>>>>>> So is an RPC-IF a real hardware device.  Can you share the datasheet?
>>>>>>
>>>>>> Unfortunately the datasheet for the R-Car Gen3 and RZ/G2 SoCs is
>>>>>> not yet public.
>>>>>
>>>>> When will it be public?
>>>>
>>>> Dunno. RZ/G1 documentation became public a few months after the SoC
>>>> release.
>>>>
>>>>> Do you have access to it?
>>>>
>>>> Yes I do.
>>>
>>> Great.  Maybe you can help Sergei with his 'undocumented bits' issue.
>>
>>    You're an optimist. :-)
>>    I think I have the same gen3 manual v1.50 as Geert. It won't help as the
>> bits that constitute the magic numbers in the driver are not described (not
>> even the default values are listed for the most of them).
> 
> Then how did you get hold of them?

   Mason did, not me. And they were copied from the bootloader, IIRC.

>>>>>> However, a very similar hardware block is present in the RZ/A2M SoC.
>>>>>> Please see Chapter 20 ("SPI Multi I/O Bus Controller") of the "RZ/A2M Group
>>>>>> User’s Manual: Hardware", which you can download from
>>>>>> https://www.renesas.com/eu/en/products/microcontrollers-microprocessors/rz/rza/rza2m.html#documents
>>>>>
>>>>>   "The SPI multi I/O bus controller enables the direct connection of
>>>>>    serial flash, OctaFlashTM, XccelaTM flash, or HyperFlashTM memory
>>>>>    devices to this LSI chip.
>>>>>
>>>>>    This module allows the connected serial flash, OctaFlashTM, XccelaTM
>>>>>    flash, or HyperFlashTM memory devices to be accessed by reading the
>>>>>    external address space, or using Manual mode to transmit and receive
>>>>>    data."
>>>>>
>>>>> Looks like a flash device to me.
>>>>
>>>> The external address space is a small window.
>>
>>    Yeah, it only provides 64 MiB window into the flash chip.
>>
>>>>> Can the SPI portion be used to connect generic SPI devices?
>>>>
>>>> I'll defer that to the people who worked on the driver...
>>
>>    Yes. Or at least it should, looking at the driver code...
> 
> Judging by that response, I'm guessing that this is unused/untested.

   I certainly haven't tested it, perhaps Mason did?

MBR, Sergei