[PATCH v2] i2c: hix5hd2: Make sure clk is disabled in remove

[PATCH v2] i2c: hix5hd2: Make sure clk is disabled in remove

[Andi Shyti]

From: Alexey Khoroshilov <khoroshilov@ispras.ru>

pm_runtime_set_suspended() does not lead to call of suspend callback,
so clk may be left undisabled in hix5hd2_i2c_remove().

By the way, the patch adds error handling for clk_prepare_enable().

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Signed-off-by: Andi Shyti <andi.shyti@kernel.org>
---
Hi,

Look what I fished from the far December 2017 :)

It looked better to respin it rather than replying to such an old
mail.

I haven't made any modification to the patch exept for a little
rebase conflict. Here's a full changelog, anyway.

Changelog
=========
v1 -> v2:
 - Fished this out from the muddy pond.
 - Added my SoB
 - Fixed rebase conflict

Andi

 drivers/i2c/busses/i2c-hix5hd2.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/drivers/i2c/busses/i2c-hix5hd2.c b/drivers/i2c/busses/i2c-hix5hd2.c
index 0e34cbaca22dc..ec775ffefa9fc 100644
--- a/drivers/i2c/busses/i2c-hix5hd2.c
+++ b/drivers/i2c/busses/i2c-hix5hd2.c
@@ -421,7 +421,11 @@ static int hix5hd2_i2c_probe(struct platform_device *pdev)
 		dev_err(&pdev->dev, "cannot get clock\n");
 		return PTR_ERR(priv->clk);
 	}
-	clk_prepare_enable(priv->clk);
+	ret = clk_prepare_enable(priv->clk);
+	if (ret) {
+		dev_err(&pdev->dev, "cannot enable clock\n");
+		return ret;
+	}
 
 	strscpy(priv->adap.name, "hix5hd2-i2c", sizeof(priv->adap.name));
 	priv->dev = &pdev->dev;
@@ -469,8 +473,10 @@ static int hix5hd2_i2c_remove(struct platform_device *pdev)
 	struct hix5hd2_i2c_priv *priv = platform_get_drvdata(pdev);
 
 	i2c_del_adapter(&priv->adap);
-	pm_runtime_disable(priv->dev);
-	pm_runtime_set_suspended(priv->dev);
+
+	/* Make sure priv->clk is disabled */
+	pm_runtime_force_suspend(priv->dev);
+
 	clk_disable_unprepare(priv->clk);
 
 	return 0;
-- 
2.40.1

Re: [PATCH v2] i2c: hix5hd2: Make sure clk is disabled in remove

[Andi Shyti]

Hi Alexey,

are you still around? Your patch looks good, just two things...

On Fri, Jun 09, 2023 at 12:55:13AM +0200, Andi Shyti wrote:
> From: Alexey Khoroshilov <khoroshilov@ispras.ru>
> 
> pm_runtime_set_suspended() does not lead to call of suspend callback,
> so clk may be left undisabled in hix5hd2_i2c_remove().
> 
> By the way, the patch adds error handling for clk_prepare_enable().

could you please put this in a separate patch?

> Found by Linux Driver Verification project (linuxtesting.org).
> 
> Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
> Signed-off-by: Andi Shyti <andi.shyti@kernel.org>

if you're taking this, feel free to drop my SoB...

> ---
> Hi,
> 
> Look what I fished from the far December 2017 :)
> 
> It looked better to respin it rather than replying to such an old
> mail.
> 
> I haven't made any modification to the patch exept for a little
> rebase conflict. Here's a full changelog, anyway.
> 
> Changelog
> =========
> v1 -> v2:
>  - Fished this out from the muddy pond.
>  - Added my SoB
>  - Fixed rebase conflict

... but please keep the changelog.

> Andi
> 
>  drivers/i2c/busses/i2c-hix5hd2.c | 12 +++++++++---
>  1 file changed, 9 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-hix5hd2.c b/drivers/i2c/busses/i2c-hix5hd2.c
> index 0e34cbaca22dc..ec775ffefa9fc 100644
> --- a/drivers/i2c/busses/i2c-hix5hd2.c
> +++ b/drivers/i2c/busses/i2c-hix5hd2.c
> @@ -421,7 +421,11 @@ static int hix5hd2_i2c_probe(struct platform_device *pdev)
>  		dev_err(&pdev->dev, "cannot get clock\n");
>  		return PTR_ERR(priv->clk);
>  	}
> -	clk_prepare_enable(priv->clk);
> +	ret = clk_prepare_enable(priv->clk);

please use devm_clk_get_enabled();

Thanks,
Andi

> +	if (ret) {
> +		dev_err(&pdev->dev, "cannot enable clock\n");
> +		return ret;
> +	}
>  
>  	strscpy(priv->adap.name, "hix5hd2-i2c", sizeof(priv->adap.name));
>  	priv->dev = &pdev->dev;
> @@ -469,8 +473,10 @@ static int hix5hd2_i2c_remove(struct platform_device *pdev)
>  	struct hix5hd2_i2c_priv *priv = platform_get_drvdata(pdev);
>  
>  	i2c_del_adapter(&priv->adap);
> -	pm_runtime_disable(priv->dev);
> -	pm_runtime_set_suspended(priv->dev);
> +
> +	/* Make sure priv->clk is disabled */
> +	pm_runtime_force_suspend(priv->dev);
> +
>  	clk_disable_unprepare(priv->clk);
>  
>  	return 0;
> -- 
> 2.40.1
> 

Re: [PATCH v2] i2c: hix5hd2: Make sure clk is disabled in remove

[Alexey Khoroshilov]

Hi Andi,

On 09.06.2023 02:10, Andi Shyti wrote:
> Hi Alexey,
> 
> are you still around? Your patch looks good, just two things...
> 
> On Fri, Jun 09, 2023 at 12:55:13AM +0200, Andi Shyti wrote:
>> From: Alexey Khoroshilov <khoroshilov@ispras.ru>
>>
>> pm_runtime_set_suspended() does not lead to call of suspend callback,
>> so clk may be left undisabled in hix5hd2_i2c_remove().
>>
>> By the way, the patch adds error handling for clk_prepare_enable().
> 
> could you please put this in a separate patch?

Sure I can do that, thank you for catching it up!

--
Alexey

> 
>> Found by Linux Driver Verification project (linuxtesting.org).
>>
>> Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
>> Signed-off-by: Andi Shyti <andi.shyti@kernel.org>
> 
> if you're taking this, feel free to drop my SoB...
> 
>> ---
>> Hi,
>>
>> Look what I fished from the far December 2017 :)
>>
>> It looked better to respin it rather than replying to such an old
>> mail.
>>
>> I haven't made any modification to the patch exept for a little
>> rebase conflict. Here's a full changelog, anyway.
>>
>> Changelog
>> =========
>> v1 -> v2:
>>  - Fished this out from the muddy pond.
>>  - Added my SoB
>>  - Fixed rebase conflict
> 
> ... but please keep the changelog.
> 
>> Andi
>>
>>  drivers/i2c/busses/i2c-hix5hd2.c | 12 +++++++++---
>>  1 file changed, 9 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/i2c/busses/i2c-hix5hd2.c b/drivers/i2c/busses/i2c-hix5hd2.c
>> index 0e34cbaca22dc..ec775ffefa9fc 100644
>> --- a/drivers/i2c/busses/i2c-hix5hd2.c
>> +++ b/drivers/i2c/busses/i2c-hix5hd2.c
>> @@ -421,7 +421,11 @@ static int hix5hd2_i2c_probe(struct platform_device *pdev)
>>  		dev_err(&pdev->dev, "cannot get clock\n");
>>  		return PTR_ERR(priv->clk);
>>  	}
>> -	clk_prepare_enable(priv->clk);
>> +	ret = clk_prepare_enable(priv->clk);
> 
> please use devm_clk_get_enabled();
> 
> Thanks,
> Andi
> 
>> +	if (ret) {
>> +		dev_err(&pdev->dev, "cannot enable clock\n");
>> +		return ret;
>> +	}
>>  
>>  	strscpy(priv->adap.name, "hix5hd2-i2c", sizeof(priv->adap.name));
>>  	priv->dev = &pdev->dev;
>> @@ -469,8 +473,10 @@ static int hix5hd2_i2c_remove(struct platform_device *pdev)
>>  	struct hix5hd2_i2c_priv *priv = platform_get_drvdata(pdev);
>>  
>>  	i2c_del_adapter(&priv->adap);
>> -	pm_runtime_disable(priv->dev);
>> -	pm_runtime_set_suspended(priv->dev);
>> +
>> +	/* Make sure priv->clk is disabled */
>> +	pm_runtime_force_suspend(priv->dev);
>> +
>>  	clk_disable_unprepare(priv->clk);
>>  
>>  	return 0;
>> -- 
>> 2.40.1
>>

Re: [PATCH v2] i2c: hix5hd2: Make sure clk is disabled in remove

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 643 bytes --]

On Fri, Jun 09, 2023 at 12:55:13AM +0200, Andi Shyti wrote:
> From: Alexey Khoroshilov <khoroshilov@ispras.ru>
> 
> pm_runtime_set_suspended() does not lead to call of suspend callback,
> so clk may be left undisabled in hix5hd2_i2c_remove().
> 
> By the way, the patch adds error handling for clk_prepare_enable().
> 
> Found by Linux Driver Verification project (linuxtesting.org).
> 
> Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
> Signed-off-by: Andi Shyti <andi.shyti@kernel.org>

I have now applied "[PATCH 05/15] i2c: busses: hix5hd2: Use
devm_clk_get_enabled()". Can this patch then be dropped entirely?


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2] i2c: hix5hd2: Make sure clk is disabled in remove

[Andi Shyti]

Hi Wolfram,

On Fri, Jun 23, 2023 at 12:10:07PM +0200, Wolfram Sang wrote:
> On Fri, Jun 09, 2023 at 12:55:13AM +0200, Andi Shyti wrote:
> > From: Alexey Khoroshilov <khoroshilov@ispras.ru>
> > 
> > pm_runtime_set_suspended() does not lead to call of suspend callback,
> > so clk may be left undisabled in hix5hd2_i2c_remove().
> > 
> > By the way, the patch adds error handling for clk_prepare_enable().
> > 
> > Found by Linux Driver Verification project (linuxtesting.org).
> > 
> > Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
> > Signed-off-by: Andi Shyti <andi.shyti@kernel.org>
> 
> I have now applied "[PATCH 05/15] i2c: busses: hix5hd2: Use
> devm_clk_get_enabled()". Can this patch then be dropped entirely?

yes, this patch can be dropped. When I resent this I didn't plan
that refactoring.

Thanks,
Andi