Re: [RFC PATCH 1/1] mmc: dw_mmc: Add runtime pm to dw_mmc

[Karol Wrona <k.wrona@samsung.com>]

On 01/22/2015 10:02 AM, Ulf Hansson wrote:
> On 21 January 2015 at 17:43, Karol Wrona <k.wrona@samsung.com> wrote:
>> This patch adds runtime pm handling to dw_mmc and enables it for dw_mmc-exynos.
>> It mainly uses mci_request/mci_request_end for mmc host state information.
>>
>> Signed-off-by: Karol Wrona <k.wrona@samsung.com>
>> ---
>>  drivers/mmc/host/dw_mmc-exynos.c |   69 ++++++++++++++++++++++++++++++++++++--
>>  drivers/mmc/host/dw_mmc.c        |   65 +++++++++++++++++++++++++++++++++--
>>  2 files changed, 130 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/mmc/host/dw_mmc-exynos.c b/drivers/mmc/host/dw_mmc-exynos.c
>> index 12a5eaa..7281c6f 100644
>> --- a/drivers/mmc/host/dw_mmc-exynos.c
>> +++ b/drivers/mmc/host/dw_mmc-exynos.c
>> @@ -17,12 +17,15 @@
[...]
>>
>> +#define DWMMC_AUTOSUSPEND_DELAY 200
> 
> Normally we use 50 as default. Any reason to why you can't use that?
> 
Thanks for looking at that.
No special reason. I will check lower delay value.

> Hmm, maybe we should have such a default defined in a common mmc host
> header file!?
Will do.
> 
>> +
>>  /* Variations in Exynos specific dw-mshc controller */
>>  enum dw_mci_exynos_type {
>>         DW_MCI_TYPE_EXYNOS4210,
>> @@ -97,6 +100,30 @@ static int dw_mci_exynos_setup_clock(struct dw_mci *host)
>>         return 0;
>>  }
>>
>> +#ifdef CONFIG_PM
>> +static int dw_mci_exynos_runtime_suspend(struct device *dev)
>> +{
>> +       struct dw_mci *host = dev_get_drvdata(dev);
>> +
>> +       /* empty for now */
>> +
>> +       return 0;
>> +}
>> +
>> +static int dw_mci_exynos_runtime_resume(struct device *dev)
>> +{
>> +       struct dw_mci *host = dev_get_drvdata(dev);
>> +
>> +       /* empty for now */
>> +
>> +       return 0;
>> +}
>> +#else
>> +#define dw_mci_exynos_runtime_suspend          NULL
>> +#define dw_mci_exynos_runtime_resume           NULL
>> +
>> +#endif /* CONFIG_PM */
> 
> I would suggest you to remove all the above code from this patch. If
> you want to add the callbacks, let's anyway do that from a separate
> patch.
Will do.

Thanks,
Karol