Re: [PATCH] kconfig: Remove last trace of GENERIC_IO

[Robin Murphy <robin.murphy@arm.com>]

On 07/02/18 17:15, Rob Herring wrote:
> On Feb 6, 2018 8:14 AM, "Robin Murphy" <robin.murphy@arm.com> wrote:
>>
>> On 06/02/18 16:06, Rob Herring wrote:
>>>
>>> On Tue, Feb 6, 2018 at 7:11 AM, Robin Murphy <robin.murphy@arm.com> wrote:
>>>>
>>>> Since commit 9de8da47742b ("kconfig: kill off GENERIC_IO option"), this
>>>> select has done nothing, and is now just confusing. Clean it up.
>>>>
>>>> Signed-off-by: Robin Murphy <robin.murphy@arm.com>
>>>> ---
>>>>    lib/Kconfig | 1 -
>>>>    1 file changed, 1 deletion(-)
>>>
>>>
>>> We already have:
>>>
>>> https://lkml.org/lkml/2018/1/17/924
>>>
>>> and this series:
>>>
>>> https://lkml.org/lkml/2018/2/4/208
>>
>>
>> Agh, Ulf's patch is even sat in my inbox...
>>
>> This was just a drive-by from looking into HAS_IOMEM, sorry for the noise :)
> 
> What are you looking into there? IMO, we should get rid of HAS_IOMEM.
> It's basically a !UML option to turn off a bunch of drivers in UML.

I was just having a quick investigate out of curiosity after Geert 
raised it in the context of awkward COMPILE_TEST dependencies here:

https://lists.linuxfoundation.org/pipermail/iommu/2018-February/025963.html

I don't have the bandwidth myself to actually work on burying it, but it 
does seem like that might be feasible.

Robin.