Re: [RFC PATCH v3.1 00/27] Add support UHS-II for GL9755

[Adrian Hunter]

On 1/12/20 5:09 am, AKASHI Takahiro wrote:
> Adrian,
> 
> Thank you for your review comments.
> 
> On Thu, Nov 26, 2020 at 10:18:55AM +0200, Adrian Hunter wrote:
>> On 25/11/20 9:41 am, AKASHI Takahiro wrote:
>>> Gentle ping;
>>>
>>> On Fri, Nov 06, 2020 at 11:26:59AM +0900, AKASHI Takahiro wrote:
>>>> This is an interim snapshot of our next version, v4, for enabling
>>>> UHS-II on MMC/SD.
>>>>
>>>> It is focused on 'sdhci' side to address Adrian's comments regarding
>>>> "modularising" sdhci-uhs2.c.
>>>> The whole aim of this version is to get early feedback from Adrian (and
>>>> others) on this issue. Without any consensus about the code structure,
>>>
>>> Any comments so far?
>>>
>>
>> Overall, I like this approach of separating UHS2 from legacy sdhci as much
>> as possible.  The only major change, is to drop support for legacy quirks
>> and features that you do not need.  The reason for that, is that there may
>> be few drivers that end up with UHS-II support (opting instead for SD
>> Express), so there is no point going to a lot of trouble to support things
>> that never get used.
>>
>> From what I have seen that looks like it includes:
>> 	- any quirks
> 
> GLI driver (gl9755) needs
>   * SDHCI_QUIRK_NO_ENDATTR_IN_NOPDESC
>   * SDHCI_QUIRK2_BROKEN_DDR50
> but they are managed in sdhci code.
> 
>> 	- SDHCI LED support
>> 	- external DMA support
> 
> Should we add 'depends on !SDHCI_UHS2' to MMC_SDHCI_EXTERNAL_DMA?

No because SDHCI_UHS2 does not necessary mean UHS-II is actually being used
i.e. same kernel can support both even if no controller uses both