Re: [PATCH v2 2/2] Documentation: fpga: dfl: Add description for DFL UIO support

From: Tom Rix
Date: Mon Dec 14 2020 - 23:45:16 EST



On 12/14/20 6:22 PM, Xu Yilun wrote:
> On Mon, Dec 14, 2020 at 02:14:56PM -0800, Tom Rix wrote:
>> On 12/13/20 7:36 PM, Xu Yilun wrote:
>>> This patch adds description for UIO support for dfl devices on DFL
>>> bus.
>>>
>>> Signed-off-by: Xu Yilun <yilun.xu@xxxxxxxxx>
>>> ---
>>> v2: no doc in v1, add it for v2.
>>> ---
>>
>>> +components. They could instantiate a new private feature in the DFL, and then
>>> +get a DFL device in their system. In some cases users may need a userspace
>>> +driver for the DFL device:
>>> +
>>> +* Users may need to run some diagnostic test for their hardwares.
>> * Users may prototype the kernel driver in user space.
> Could we just add the line rather than replacing the previous line? I think this
> comment is describing a different usecase.

Yes, this is what i ment, please use your original.

I am offering another usecase, one I will use.

Add mine as well, if you want.

>
>>> +* Some hardware is designed for specific purposes and does not fit into one of
>>> + the standard kernel subsystems.
>>> +
>>> +This requires the direct access to the MMIO space and interrupt handling in
>>> +userspace. We implemented a dfl-uio-pdev module which exposes the UIO device
>> The dfl-uio-pdev module exposes
> Will change it.
>
>>> +interfaces. It adds the uio_pdrv_genirq platform device with the resources of
>>> +the DFL device, and let the generic UIO platform device driver provide UIO
>> the DLF device, and lets
> Will change it.
>
>>> +support to userspace.
>> Use FPGA_DFL_UIO_PDEV to enable this feature.
> I didn't get your idea for this.

I wanted the user to know which kconfig controls this feature.

Leave it out if you don't think it fits.

>
>>> +
>>> +The DFL UIO driver has a special matching algorithem. It will match any DFL
>>> +device which could not be handled by other DFL drivers. In this way, it will
>>> +not impact the functionality of the features which are already supported by the
>>> +system.
>> (not sure if this section is needed)
> I think we may keep it.

Ok.

Tom

>
> Thanks,
> Yilun
>