Re: [PATCH v3] mmc: mtk-sd: reduce CIT for better performance

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 06/06/23 10:17, Wenbin Mei (梅文彬) ha scritto:
> On Mon, 2023-06-05 at 10:48 +0200, AngeloGioacchino Del Regno wrote:
>>   	
>> External email : Please do not click links or open attachments until
>> you have verified the sender or the content.
>>   Il 05/06/23 08:01, Wenbin Mei ha scritto:
>>> CQHCI_SSC1 indicates to CQE the polling period to use when using
>> periodic
>>> SEND_QUEUE_STATUS(CMD13) polling.
>>> Since MSDC CQE uses msdc_hclk as ITCFVAL, so driver should use hclk
>>> frequency to get the actual time.
>>> The default value 0x1000 that corresponds to 150us for MediaTek
>> SoCs, let's
>>> decrease it to 0x40 that corresponds to 2.35us, which can improve
>> the
>>> performance of some eMMC devices.
>>>
>>> Signed-off-by: Wenbin Mei <wenbin.mei@mediatek.com>
>>
>> OK! That's almost good now. There's only one consideration here: if
>> MediaTek
>> SoCs *require* msdc_hclk to calculate the CIT time, this means that
>> this clock
>> is critical for CQHCI functionality.
>>
>> If msdc_hclk is not present, CQHCI cannot work correctly... so you
>> don't have
>> to cover the case in which there's no msdc_hclk clock: if that's not
>> present,
>> either fail probing, or disable CQHCI.
>>
>>> ---
>>>    drivers/mmc/host/cqhci.h  |  1 +
>>>    drivers/mmc/host/mtk-sd.c | 47
>> +++++++++++++++++++++++++++++++++++++++
>>>    2 files changed, 48 insertions(+)
>>>
>>> diff --git a/drivers/mmc/host/cqhci.h b/drivers/mmc/host/cqhci.h
>>> index ba9387ed90eb..292b89ebd978 100644
>>> --- a/drivers/mmc/host/cqhci.h
>>> +++ b/drivers/mmc/host/cqhci.h
>>> @@ -23,6 +23,7 @@
>>>    /* capabilities */
>>>    #define CQHCI_CAP0x04
>>>    #define CQHCI_CAP_CS0x10000000 /* Crypto Support */
>>> +#define CQHCI_CAP_ITCFMUL(x)(((x) & GENMASK(15, 12)) >> 12)
>>>    
>>>    /* configuration */
>>>    #define CQHCI_CFG0x08
>>> diff --git a/drivers/mmc/host/mtk-sd.c b/drivers/mmc/host/mtk-sd.c
>>> index edade0e54a0c..c221ef8a6992 100644
>>> --- a/drivers/mmc/host/mtk-sd.c
>>> +++ b/drivers/mmc/host/mtk-sd.c
>>> @@ -473,6 +473,7 @@ struct msdc_host {
>>>    struct msdc_tune_para def_tune_para; /* default tune setting */
>>>    struct msdc_tune_para saved_tune_para; /* tune result of
>> CMD21/CMD19 */
>>>    struct cqhci_host *cq_host;
>>> +u32 cq_ssc1_time;
>>>    };
>>>    
>>>    static const struct mtk_mmc_compatible mt2701_compat = {
>>> @@ -2450,9 +2451,50 @@ static void
>> msdc_hs400_enhanced_strobe(struct mmc_host *mmc,
>>>    }
>>>    }
>>>    
>>> +static void msdc_cqe_cit_cal(struct msdc_host *host, u64 timer_ns)
>>
>> static int msdc_cqe_cit_cal(....)
>>
> Sorry, I missed this comment.
> I think there is no need to return a value.
> Becuase msdc_hclk is exist, and if not present, it will return earily.
> Even if it goes to the default case in the switch flow, we will assign
> a default value.
> So I think it's better to return null, do you think it is okay?
> 

Yeah ignore this comment; I've noticed that HCLK is already mandatory, otherwise
the probe function will fail earlier anyway.

Thanks,
Angelo