Re: [PATCH V1 1/3] mmc: sdhci: Allow platform controlled voltage switching

[Adrian Hunter <adrian.hunter@intel.com>]

On 25/07/18 15:23, Vijay Viswanath wrote:
> Hi Adrian,
> 
> 
> On 7/25/2018 5:23 PM, Adrian Hunter wrote:
>> On 20/07/18 13:46, Vijay Viswanath wrote:
>>> Some controllers can have internal mechanism to inform the SW that it
>>> is ready for voltage switching. For such controllers, changing voltage
>>> before the HW is ready can result in various issues.
>>>
>>> During setup/cleanup of host, check whether regulator enable/disable
>>> was already done by platform driver.
>>>
>>> Signed-off-by: Vijay Viswanath <vviswana@codeaurora.org>
>>> ---
>>>   drivers/mmc/host/sdhci.c | 21 ++++++++++++++++-----
>>>   1 file changed, 16 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c
>>> index 1c828e0..494a1e2 100644
>>> --- a/drivers/mmc/host/sdhci.c
>>> +++ b/drivers/mmc/host/sdhci.c
>>> @@ -3472,6 +3472,7 @@ int sdhci_setup_host(struct sdhci_host *host)
>>>       unsigned int override_timeout_clk;
>>>       u32 max_clk;
>>>       int ret;
>>> +    bool enable_vqmmc = false;
>>>         WARN_ON(host == NULL);
>>>       if (host == NULL)
>>> @@ -3485,7 +3486,12 @@ int sdhci_setup_host(struct sdhci_host *host)
>>>        * the host can take the appropriate action if regulators are not
>>>        * available.
>>>        */
>>> -    ret = mmc_regulator_get_supply(mmc);
>>> +    if (!mmc->supply.vmmc) {
>>> +        ret = mmc_regulator_get_supply(mmc);
>>> +        enable_vqmmc  = true;
>>> +    } else {
>>> +        ret = 0;
>>> +    }
>>>       if (ret)
>>>           return ret;
>>
>> Why not
>>
>>     if (!mmc->supply.vmmc) {
>>         ret = mmc_regulator_get_supply(mmc);
>>           if (ret)
>>                return ret;
>>         enable_vqmmc  = true;
>>     }
>>
> 
> looks neater. Will do.
> 
>>>   @@ -3736,7 +3742,10 @@ int sdhci_setup_host(struct sdhci_host *host)
>>>         /* If vqmmc regulator and no 1.8V signalling, then there's no UHS */
>>>       if (!IS_ERR(mmc->supply.vqmmc)) {
>>> -        ret = regulator_enable(mmc->supply.vqmmc);
>>> +        if (enable_vqmmc)
>>> +            ret = regulator_enable(mmc->supply.vqmmc);
>>
>> Please introduce host->vqmmc_enabled = !ret;
>>
> 
> Any reason to introduce vqmmc_enabled variable in sdhci_host structure ? We
> can get around with a local variable and using regulator_is_enabled API.

regulator_enable() uses reference counting, so we have to use
regulator_disable() exactly the same number of times that we use
regulator_enable().

> 
>>> +        else
>>> +            ret = 0;
>>>           if (!regulator_is_supported_voltage(mmc->supply.vqmmc, 1700000,
>>>                               1950000))
>>>               host->caps1 &= ~(SDHCI_SUPPORT_SDR104 |
>>> @@ -3984,7 +3993,7 @@ int sdhci_setup_host(struct sdhci_host *host)
>>>       return 0;
>>>     unreg:
>>> -    if (!IS_ERR(mmc->supply.vqmmc))
>>> +    if (!IS_ERR(mmc->supply.vqmmc) && enable_vqmmc)
>>
>> And just make this
>>
>>     if (host->vqmmc_enabled)
>>
>>>           regulator_disable(mmc->supply.vqmmc);
>>>   undma:
>>>       if (host->align_buffer)
>>> @@ -4002,7 +4011,8 @@ void sdhci_cleanup_host(struct sdhci_host *host)
>>>   {
>>>       struct mmc_host *mmc = host->mmc;
>>>   -    if (!IS_ERR(mmc->supply.vqmmc))
>>> +    if (!IS_ERR(mmc->supply.vqmmc) &&
>>> +            (regulator_is_enabled(mmc->supply.vqmmc) > 0))
>>
>>     if (host->vqmmc_enabled)
>>
>>>           regulator_disable(mmc->supply.vqmmc);
>>>         if (host->align_buffer)
>>> @@ -4135,7 +4145,8 @@ void sdhci_remove_host(struct sdhci_host *host, int
>>> dead)
>>>         tasklet_kill(&host->finish_tasklet);
>>>   -    if (!IS_ERR(mmc->supply.vqmmc))
>>> +    if (!IS_ERR(mmc->supply.vqmmc) &&
>>> +            (regulator_is_enabled(mmc->supply.vqmmc) > 0))
>>
>>     if (host->vqmmc_enabled)
>>
>>>           regulator_disable(mmc->supply.vqmmc);
>>>         if (host->align_buffer)
>>>
>>
>> -- 
>> 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
>>
> 
> Thanks,
> Vijay
>