[PATCH 4/7] mmc: dw_mmc: add samsung exynos5250 specific extentions

Thomas Abraham thomas.abraham at linaro.org
Thu May 10 06:55:15 EDT 2012


On 2 May 2012 13:19, Jaehoon Chung <jh80.chung at samsung.com> wrote:
> On 05/02/2012 04:01 PM, Kyungmin Park wrote:
>
>> Hi,
>>
>> On 5/2/12, Thomas Abraham <thomas.abraham at linaro.org> wrote:
>>> The instantiation of the Synopsis Designware controller on Exynos5250
>>> include extension for SDR and DDR specific tx/rx phase shift timing
>>> and CIU internal divider. In addition to that, the option to skip the
>>> command hold stage is also introduced. Add support for these Exynos5250
>>> specfic extenstions.

[...]

>>> @@ -265,6 +266,10 @@ static u32 dw_mci_prepare_command(struct mmc_host *mmc,
>>> struct mmc_command *cmd)
>>>                      cmdr |= SDMMC_CMD_DAT_WR;
>>>      }
>>>
>>> +    if (slot->host->drv_data->ctrl_type == DW_MCI_TYPE_EXYNOS5250)
>>> +            if (SDMMC_CLKSEL_GET_SELCLK_DRV(mci_readl(slot->host, CLKSEL)))
>>> +                    cmdr |= SDMMC_USE_HOLD_REG;
>> Some other board, custom SOC also can use this HOLD register. So it's
>> not EXYNOS5250 specific one. I think we introduce the more generic
>> quirks for this instead of SOC specific.
>
> One more, I think that also need to check the IMPLEMENT_HOLD_REG bit in HCON register.
> It has dependency with that.

The above code is specific to Exynos5250 and hence it is not required
to check the IMPLEMENT_HOLD_REG bit in HCON register. On Exynos5250,
the hold register is implemented and available.


> As Mr.Park is mentioned, this register is clock phasing.
> In spec, card is enumerated in SDR12 or SDR25 mode, the application must program the use_hold_reg.

Exynos5250 hardware manual specifies additional restrictions on the
use of hold register. The above code checks for those restrictions and
programs the USE_HOLD_REG accordingly. Please let me know if there is
any condition that is not handled by the above code.

Thanks,
Thomas.

>
> Best Regards,
> Jaehoon Chung
>



More information about the linux-arm-kernel mailing list