Re: [PATCH] mmc: dw_mmc: modify DATA register offset

[James Hogan <james.hogan@imgtec.com>]

On 10/17/2011 11:04 AM, Jaehoon Chung wrote:
> On 10/17/2011 06:00 PM, James Hogan wrote:
> 
>> Hi
>>
>> On 10/17/2011 09:46 AM, Jaehoon Chung wrote:
>>> Hi James.
>>>
>>> On 10/17/2011 05:27 PM, James Hogan wrote:
>>>
>>>> Hi,
>>>>
>>>> On 10/17/2011 08:05 AM, Jaehoon Chung wrote:
>>>>> In dw_mmc 2.40a spec, Data register's offset is changed.
>>>>> Now we used Data register offset is 0x100. but if somebody use 2.40a
>>>>> controller, must use 0x200 for Data register.
>>>>>
>>>>> This patch is added version-id checking point and using SDMMC_DATA(x)
>>>>> instead of SDMMC_DATA. (assume 2.40a is the latest version)
>>>>>
>>>>> Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
>>>>> Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
>>>>> ---
>>>>>  drivers/mmc/host/dw_mmc.c  |   66 ++++++++++++++++++++++++++++++--------------
>>>>>  drivers/mmc/host/dw_mmc.h  |   10 ++++++-
>>>>>  include/linux/mmc/dw_mmc.h |    2 +
>>>>>  3 files changed, 56 insertions(+), 22 deletions(-)
>>>>>
>>>>> diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
>>>>> index 701f14e..3aaeb08 100644
>>>>> --- a/drivers/mmc/host/dw_mmc.c
>>>>> +++ b/drivers/mmc/host/dw_mmc.c
>>>>> @@ -1043,7 +1043,8 @@ static void dw_mci_push_data16(struct dw_mci *host, void *buf, int cnt)
>>>>>  		buf += len;
>>>>>  		cnt -= len;
>>>>>  		if (!sg_next(host->sg) || host->part_buf_count == 2) {
>>>>> -			mci_writew(host, DATA, host->part_buf16);
>>>>> +			mci_writew(host, DATA(host->data_offset),
>>>>> +					host->part_buf16);
>>>>>  			host->part_buf_count = 0;
>>>>>  		}
>>>>>  	}
>>>>
>>>> I really think it would be more concise to just have something like this:
>>>> mci_writew(host, host->data_offset, host->part_buf16);
>>>> ...
>>>>
>>>>> +#define DATA_OFFSET		0
>>>>> +#define DATA_240A_OFFSET	0x100
>>>>
>>>> and then have these as register positions like the other #defines, e.g.
>>>> #define SDMMC_DATA      0x100
>>>> #define SDMMC_DATA_240A 0x200
>>>>
>>>
>>>
>>> Sorry, if change your suggestion, how do you control SDMMC_##reg?
>>> mci_readl(dev, reg) __raw_readl(dev->regs + SDMMC_##reg)
>>
>> Ah ok, sorry. I see what you mean now. I'd forgotton the mci_readl macro
>> did that!
>>
>> I suppose there's a couple of ways that you could avoid the offset from
>> 0x100.
>>
>> 1) could define a register macro which takes a raw offset:
>> #define SDMMC_RAW(x)		(x)
>> mci_writew(host, RAW(host->data_offset), host->part_buf16);
>>
>> 2) could define the DATA register macro which takes a struct dw_mci* as
>> an argument:
>> #define SDMMC_DATA(HOST)        ((HOST)->data_offset)
>> mci_writew(host, DATA(host), host->part_buf16);
>>
>> I don't have a strong preference between these.
> 
> My suggestion is also similar to your suggestions.
> But your suggestions is used data->offset assigned 0x100 or 0x200. right?

yes, that's right. Using (host->regs + 0x100 + host->data_offset) is 2
additions, which is one reason why I suggest having just (host->regs +
host->data_offset).

> My suggestions is used (DATA + (x)). 
> All of them must use the macro like DATA(x). right?

yes, unfortunately, unless the mci_write* macros are altered, something
like what you did before, but I think since DATA is still being treated
as a single register, changing the SDMMC_DATA macro would be better.

Cheers
James

> 
> My suggestion and yours are difference which offset used.
> I will resend this patch after modify...
> 
> Best regards,
> Jaehoon Chung
> 
>>
>> Thanks
>> James
>>
>>>
>>>>> @@ -1952,6 +1964,18 @@ static int dw_mci_probe(struct platform_device *pdev)
>>>>>  	}
>>>>>  
>>>>>  	/*
>>>>> +	 * In 2.40a spec, Data offset is changed.
>>>>> +	 * Need to check the version-id and set data-offset for DATA register.
>>>>> +	 */
>>>>> +	host->verid = SDMMC_GET_VERID(mci_readl(host, VERID));
>>>>> +	dev_info(&pdev->dev, "Version ID is %04x\n", host->verid);
>>>>> +
>>>>> +	if (host->verid < DW_MMC_240A)
>>>>> +		host->data_offset = DATA_OFFSET;
>>>>> +	else
>>>>> +		host->data_offset = DATA_240A_OFFSET;
>>>>> +
>>>>> +	/*
>>>>>  	 * Enable interrupts for command done, data over, data empty, card det,
>>>>>  	 * receive ready and error such as transmit, receive timeout, crc error
>>>>>  	 */
>>>>> diff --git a/drivers/mmc/host/dw_mmc.h b/drivers/mmc/host/dw_mmc.h
>>>>> index bfa3c1c..965fd19 100644
>>>>> --- a/drivers/mmc/host/dw_mmc.h
>>>>> +++ b/drivers/mmc/host/dw_mmc.h
>>>>> @@ -14,6 +14,8 @@
>>>>>  #ifndef _DW_MMC_H_
>>>>>  #define _DW_MMC_H_
>>>>>  
>>>>> +#define DW_MMC_240A		0x240a
>>>>> +
>>>>>  #define SDMMC_CTRL		0x000
>>>>>  #define SDMMC_PWREN		0x004
>>>>>  #define SDMMC_CLKDIV		0x008
>>>>> @@ -51,7 +53,11 @@
>>>>>  #define SDMMC_IDINTEN		0x090
>>>>>  #define SDMMC_DSCADDR		0x094
>>>>>  #define SDMMC_BUFADDR		0x098
>>>>> -#define SDMMC_DATA		0x100
>>>>> +#define SDMMC_DATA(x)		(0x100 + (x))
>>>>> +
>>>>> +/* Data offset is difference according to Verision */
>>>>
>>>> should that be "version"?
>>>
>>> Typo..should fix that.
>>>
>>>>
>>>>> +#define DATA_OFFSET		0
>>>>> +#define DATA_240A_OFFSET	0x100
>>>>>  
>>>>>  /* shift bit field */
>>>>>  #define _SBF(f, v)		((v) << (f))
>>>>> @@ -130,6 +136,8 @@
>>>>>  #define SDMMC_IDMAC_ENABLE		BIT(7)
>>>>>  #define SDMMC_IDMAC_FB			BIT(1)
>>>>>  #define SDMMC_IDMAC_SWRESET		BIT(0)
>>>>> +/* Version ID register define */
>>>>> +#define SDMMC_GET_VERID(x)		((x) & 0xFFFF)
>>>>>  
>>>>>  /* Register access macros */
>>>>>  #define mci_readl(dev, reg)			\
>>>>> diff --git a/include/linux/mmc/dw_mmc.h b/include/linux/mmc/dw_mmc.h
>>>>> index 6b46819..6928e29 100644
>>>>> --- a/include/linux/mmc/dw_mmc.h
>>>>> +++ b/include/linux/mmc/dw_mmc.h
>>>>> @@ -147,6 +147,8 @@ struct dw_mci {
>>>>>  	u32			current_speed;
>>>>>  	u32			num_slots;
>>>>>  	u32			fifoth_val;
>>>>> +	u16			verid;
>>>>> +	u16			data_offset;
>>>>>  	struct platform_device	*pdev;
>>>>>  	struct dw_mci_board	*pdata;
>>>>>  	struct dw_mci_slot	*slot[MAX_MCI_SLOTS];
>>>>
>>>> The kerneldoc comment above struct dw_mci should be updated to describe
>>>> the new fields.
>>>
>>> I will add the comment for new fields,
>>>
>>> Best Regards,
>>> Jaehoon Chung
>>>
>>>>
>>>
>>>> Other than that it looks good to me.
>>>>
>>>> Thanks
>>>> James
>>>>
>>>> --
>>>> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
>>>> the body of a message to majordomo@vger.kernel.org
>>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>>
>>>
>>>
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>
> 
>