Re: [PATCH 0/13] dw_mmc cleanup

[Shawn Lin <shawn.lin@rock-chips.com>]

在 2025/11/25 星期二 20:33, Ulf Hansson 写道:
> On Thu, 20 Nov 2025 at 11:29, Shawn Lin <shawn.lin@rock-chips.com> wrote:
>>
>> Hi Ulf & Jaehoon
>>
>> As you can see, dw_mmc is likely one of the most complex and difficult-to-read
>> host driver at present. It maintains various states and bottom-half scheduling,
>> containing a significant amount of redundant code—including a multi-slot mechanism
>> that has been unsupported for over a decade.
>>
>> Jaehoon attempted to remove cur_slot more than ten years ago, but the driver still
>> retains the slot structure and the associated queue mechanism designed to support
>> multiple slots. This has made the already complex code even harder to read and maintain.
>>
>> The first four patches aim to eliminate some of the redundant code, while the remaining
>> patches are intended to ultimately remove the dw_mci_slot variable. To facilitate review
>> and minimize the risk of regression, each patch is designed to accomplish a single,
>> clear objective.
>>
>> This series have been tested on the RK3588S EVB1 platform.
>>
>>
>>
>> Shawn Lin (13):
>>    mmc: dw_mmc: Remove unused struct dma_pdata
>>    mmc: dw_mmc: add dw_mci_prepare_desc() for both of 32bit and 64bit DMA
>>    mmc: dw_mmc: Remove vqmmc_enabled from struct dw_mci
>>    mmc: dw_mmc: Remove unused header files and keep alphabetical order
>>    mmc: dw_mmc: Move struct mmc_host from struct dw_mci_slot to struct
>>      dw_mci
>>    mmc: dw_mmc: Let glue drivers to use struct dw_mci as possible
>>    mmc: dw_mmc: Move flags from struct dw_mci_slot to struct dw_mci
>>    mmc: dw_mmc: Remove id from dw_mci_slot
>>    mmc: dw_mmc: Remove sdio_id from struct dw_mci_slot
>>    mmc: dw_mmc: Move clock rate stuff from struct dw_mci_slot to struct
>>      dw_mci
>>    mmc: dw_mmc: Remove mrq from struct dw_mci_slot
>>    mmc: dw_mmc: Remove queue from dw_mci
>>    mmc: dw_mmc: Remove struct dw_mci_slot
>>
>>   drivers/mmc/host/dw_mmc-exynos.c      |   9 +-
>>   drivers/mmc/host/dw_mmc-hi3798cv200.c |   6 +-
>>   drivers/mmc/host/dw_mmc-hi3798mv200.c |  17 +-
>>   drivers/mmc/host/dw_mmc-k3.c          |  21 +-
>>   drivers/mmc/host/dw_mmc-pltfm.c       |   2 +-
>>   drivers/mmc/host/dw_mmc-rockchip.c    |   9 +-
>>   drivers/mmc/host/dw_mmc-starfive.c    |   5 +-
>>   drivers/mmc/host/dw_mmc.c             | 606 ++++++++++++++--------------------
>>   drivers/mmc/host/dw_mmc.h             |  83 ++---
>>   9 files changed, 297 insertions(+), 461 deletions(-)
>>
>> --
>> 2.7.4
>>
> 
> I have walked through the series and overall this looks good to me,
> besides the comments on patch 3 and the obvious reported build
> problems.

Yes, I have fixed them locally and will look more into your comment on
patch 3.

> 
> Moreover, I would appreciate it if you could work a bit on the commit
> messages. No empty commit messages please - and try to describe a bit
> more on why/what each patch does.

Sure, will do.

> 
> Kind regards
> Uffe
>