Re: [PATCH] mmc: sdhci-of-arasan: Disable clk_xin clock in the remove

From: Flavio Ceolin
Date: Tue Dec 12 2017 - 12:15:45 EST


Hi Michal,

> On 12.12.2017 09:19, Flavio Ceolin wrote:
>> clk_xin is properly prepared/enabled on sdhci_arasan_probe(), and
>> unprepared/disabled in the error path, but it is not being
>> unprepared/disabled on sdhci_arasan_remove().
>>
>> Found by Linux Driver Verification project (linuxtesting.org).
>>
>> Signed-off-by: Flavio Ceolin <flavio.ceolin@xxxxxxxxx>
>> ---
>> drivers/mmc/host/sdhci-of-arasan.c | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/drivers/mmc/host/sdhci-of-arasan.c b/drivers/mmc/host/sdhci-of-arasan.c
>> index 0720ea7..69bd260 100644
>> --- a/drivers/mmc/host/sdhci-of-arasan.c
>> +++ b/drivers/mmc/host/sdhci-of-arasan.c
>> @@ -692,6 +692,7 @@ static int sdhci_arasan_remove(struct platform_device *pdev)
>> struct sdhci_pltfm_host *pltfm_host = sdhci_priv(host);
>> struct sdhci_arasan_data *sdhci_arasan = sdhci_pltfm_priv(pltfm_host);
>> struct clk *clk_ahb = sdhci_arasan->clk_ahb;
>> + struct clk *clk_xin = devm_clk_get(&pdev->dev, "clk_xin");
>
> I don't think this is right. You have already asked for this clock in
> probe. It means you should reuse pltfm_host->clk = clk_xin;
>
> And if you look at sdhci_pltfm_unregister you will find out that
> clk_disable_unprepare(pltfm_host->clk);
> is called there.

Yep, I confirm that. You're right, the patch is wrong, sorry for that :)
>
>>
>> if (!IS_ERR(sdhci_arasan->phy)) {
>> if (sdhci_arasan->is_phy_on)
>> @@ -705,6 +706,9 @@ static int sdhci_arasan_remove(struct platform_device *pdev)
>>
>> clk_disable_unprepare(clk_ahb);
>>
>> + if (!IS_ERR(clk_xin))
>
> And clk_xin is required property.
>
>> + clk_disable_unprepare(clk_xin);
>> +
>> return ret;
>> }
>>
>>
>
> It means NACK from me.
>
> Thanks,
> Michal

Regards,
Flavio Ceolin