[PATCH v2] misc: atmel-ssc: Handle return value of clk_prepare_enable and clk_prepare

clk_prepare_enable() and clk_prepare() can fail here and
we must check its return value.

Signed-off-by: Arvind Yadav <redacted>
---
 drivers/misc/atmel-ssc.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/misc/atmel-ssc.c b/drivers/misc/atmel-ssc.c
index b2a0340..df34b81 100644
--- a/drivers/misc/atmel-ssc.c
+++ b/drivers/misc/atmel-ssc.c

@@ -30,6 +30,7 @@ struct ssc_device *ssc_request(unsigned int ssc_num)
 {
 	int ssc_valid = 0;
 	struct ssc_device *ssc;
+	int ret;
 
 	spin_lock(&user_lock);
 	list_for_each_entry(ssc, &ssc_list, list) {

@@ -60,7 +61,11 @@ struct ssc_device *ssc_request(unsigned int ssc_num)
 	ssc->user++;
 	spin_unlock(&user_lock);
 
-	clk_prepare(ssc->clk);
+	ret = clk_prepare(ssc->clk);
+	if (ret) {
+		pr_err("Failed to prepare clock\n");
+		return ERR_PTR(ret);
+	}
 
 	return ssc;
 }

@@ -195,6 +200,7 @@ static int ssc_probe(struct platform_device *pdev)
 	struct resource *regs;
 	struct ssc_device *ssc;
 	const struct atmel_ssc_platform_data *plat_dat;
+	int ret;
 
 	ssc = devm_kzalloc(&pdev->dev, sizeof(struct ssc_device), GFP_KERNEL);
 	if (!ssc) {

@@ -229,7 +235,9 @@ static int ssc_probe(struct platform_device *pdev)
 	}
 
 	/* disable all interrupts */
-	clk_prepare_enable(ssc->clk);
+	ret = clk_prepare_enable(ssc->clk);
+	if (ret)
+		return ret;
 	ssc_writel(ssc->regs, IDR, -1);
 	ssc_readl(ssc->regs, SR);
 	clk_disable_unprepare(ssc->clk);

-- 
1.9.1

On Fri, Jun 02, 2017 at 11:09:02AM +0530, Arvind Yadav wrote:

clk_prepare_enable() and clk_prepare() can fail here and
we must check its return value.

Signed-off-by: Arvind Yadav <redacted>
---
 drivers/misc/atmel-ssc.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

What changed from v1?  Always put that information below the --- line.

v3 please?

Hi,

It is getting tiring to get patches that are nor even compile tested
resulting from whatever static analysis tool you used.

This patch has almost no value and v1 was clearly wrong.

Do you realize clk_prepare and clk_prepare_enable will never fail for
the SSC?

On 02/06/2017 at 11:09:02 +0530, Arvind Yadav wrote:

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

Hi,

Yes, Patch v1 was wrong that's why i have push v2.
clk_prepare and clk_prepare_enable can fail. There
is not harm to check it's return value. It'll not impact present
functionality.

-arvind

On Sunday 04 June 2017 03:47 AM, Alexandre Belloni wrote:

Hi,

It is getting tiring to get patches that are nor even compile tested
resulting from whatever static analysis tool you used.

This patch has almost no value and v1 was clearly wrong.

Do you realize clk_prepare and clk_prepare_enable will never fail for
the SSC?

On 02/06/2017 at 11:09:02 +0530, Arvind Yadav wrote:

quoted

clk_prepare_enable() and clk_prepare() can fail here and
we must check its return value.

Signed-off-by: Arvind Yadav <redacted>
---
  drivers/misc/atmel-ssc.c | 12 ++++++++++--
  1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/misc/atmel-ssc.c b/drivers/misc/atmel-ssc.c
index b2a0340..df34b81 100644
--- a/drivers/misc/atmel-ssc.c
+++ b/drivers/misc/atmel-ssc.c

@@ -30,6 +30,7 @@ struct ssc_device *ssc_request(unsigned int ssc_num)
  {
  	int ssc_valid = 0;
  	struct ssc_device *ssc;
+	int ret;
  
  	spin_lock(&user_lock);
  	list_for_each_entry(ssc, &ssc_list, list) {

@@ -60,7 +61,11 @@ struct ssc_device *ssc_request(unsigned int ssc_num)
  	ssc->user++;
  	spin_unlock(&user_lock);
  
-	clk_prepare(ssc->clk);
+	ret = clk_prepare(ssc->clk);
+	if (ret) {
+		pr_err("Failed to prepare clock\n");
+		return ERR_PTR(ret);
+	}
  
  	return ssc;
  }

@@ -195,6 +200,7 @@ static int ssc_probe(struct platform_device *pdev)
  	struct resource *regs;
  	struct ssc_device *ssc;
  	const struct atmel_ssc_platform_data *plat_dat;
+	int ret;
  
  	ssc = devm_kzalloc(&pdev->dev, sizeof(struct ssc_device), GFP_KERNEL);
  	if (!ssc) {

@@ -229,7 +235,9 @@ static int ssc_probe(struct platform_device *pdev)
  	}
  
  	/* disable all interrupts */
-	clk_prepare_enable(ssc->clk);
+	ret = clk_prepare_enable(ssc->clk);
+	if (ret)
+		return ret;
  	ssc_writel(ssc->regs, IDR, -1);
  	ssc_readl(ssc->regs, SR);
  	clk_disable_unprepare(ssc->clk);

-- 
1.9.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

On 05/06/2017 at 14:53:30 +0530, Arvind Yadav wrote:

Hi,

Yes, Patch v1 was wrong that's why i have push v2.
clk_prepare and clk_prepare_enable can fail. There

No this is not true, they will never fail for the SSC.

is not harm to check it's return value. It'll not impact present
functionality.

It does impact boot time, this patch adds 112 bytes to a compressed
kernel for no reason.

-arvind

On Sunday 04 June 2017 03:47 AM, Alexandre Belloni wrote:

quoted

Hi,

It is getting tiring to get patches that are nor even compile tested
resulting from whatever static analysis tool you used.

This patch has almost no value and v1 was clearly wrong.

Do you realize clk_prepare and clk_prepare_enable will never fail for
the SSC?

On 02/06/2017 at 11:09:02 +0530, Arvind Yadav wrote:

quoted

clk_prepare_enable() and clk_prepare() can fail here and
we must check its return value.

Signed-off-by: Arvind Yadav <redacted>
---
  drivers/misc/atmel-ssc.c | 12 ++++++++++--
  1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/drivers/misc/atmel-ssc.c b/drivers/misc/atmel-ssc.c
index b2a0340..df34b81 100644
--- a/drivers/misc/atmel-ssc.c
+++ b/drivers/misc/atmel-ssc.c

@@ -30,6 +30,7 @@ struct ssc_device *ssc_request(unsigned int ssc_num)
  {
  	int ssc_valid = 0;
  	struct ssc_device *ssc;
+	int ret;
  	spin_lock(&user_lock);
  	list_for_each_entry(ssc, &ssc_list, list) {

@@ -60,7 +61,11 @@ struct ssc_device *ssc_request(unsigned int ssc_num)
  	ssc->user++;
  	spin_unlock(&user_lock);
-	clk_prepare(ssc->clk);
+	ret = clk_prepare(ssc->clk);
+	if (ret) {
+		pr_err("Failed to prepare clock\n");
+		return ERR_PTR(ret);
+	}
  	return ssc;
  }

@@ -195,6 +200,7 @@ static int ssc_probe(struct platform_device *pdev)
  	struct resource *regs;
  	struct ssc_device *ssc;
  	const struct atmel_ssc_platform_data *plat_dat;
+	int ret;
  	ssc = devm_kzalloc(&pdev->dev, sizeof(struct ssc_device), GFP_KERNEL);
  	if (!ssc) {

@@ -229,7 +235,9 @@ static int ssc_probe(struct platform_device *pdev)
  	}
  	/* disable all interrupts */
-	clk_prepare_enable(ssc->clk);
+	ret = clk_prepare_enable(ssc->clk);
+	if (ret)
+		return ret;
  	ssc_writel(ssc->regs, IDR, -1);
  	ssc_readl(ssc->regs, SR);
  	clk_disable_unprepare(ssc->clk);

-- 
1.9.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

On Mon, Jun 05, 2017 at 11:36:52AM +0200, Alexandre Belloni wrote:

On 05/06/2017 at 14:53:30 +0530, Arvind Yadav wrote:

quoted

Hi,

Yes, Patch v1 was wrong that's why i have push v2.
clk_prepare and clk_prepare_enable can fail. There

No this is not true, they will never fail for the SSC.

How is anyone supposed to know this?  Just check the functions correctly
and move on, if they can never fail, wonderful, that's a code path that
is not going to be run.

quoted

is not harm to check it's return value. It'll not impact present
functionality.

It does impact boot time, this patch adds 112 bytes to a compressed
kernel for no reason.

It solves the issue that if someone copy/paste from this code, they will
implement something incorrectly.  You can spare the 112 bytes (which
really, seems very large for just 2 error cases, are you sure that's
right?  Just drop the string if you care about size here.

thanks,

greg k-h

On 05/06/2017 at 14:05:15 +0200, Greg KH wrote:

On Mon, Jun 05, 2017 at 11:36:52AM +0200, Alexandre Belloni wrote:

quoted

On 05/06/2017 at 14:53:30 +0530, Arvind Yadav wrote:

quoted

Hi,

Yes, Patch v1 was wrong that's why i have push v2.
clk_prepare and clk_prepare_enable can fail. There

No this is not true, they will never fail for the SSC.

How is anyone supposed to know this?  Just check the functions correctly
and move on, if they can never fail, wonderful, that's a code path that
is not going to be run.

Anyone able (and taking the time) to read code can know that.

I'm really not okay with having useless code paths. Many people are
already complaining (rightfully) that the kernel is bloated, let's try
to not make the situation worse and keep in mind that Linux can run on
tiny SoCs.

quoted

quoted

is not harm to check it's return value. It'll not impact present
functionality.

It does impact boot time, this patch adds 112 bytes to a compressed
kernel for no reason.

It solves the issue that if someone copy/paste from this code, they will
implement something incorrectly.  You can spare the 112 bytes (which
really, seems very large for just 2 error cases, are you sure that's
right?  Just drop the string if you care about size here.

I'm sure about the size, I've measured it on top of v4.12-rc1 before
replying.

I don't think anyone will copy/paste from this code but if that happen
and is missed by the maintainer, I'm pretty sure someone will be quick
to run his preferred static analysis tool and send a patch (hopefully,
testing it really compiles first).

thanks,

greg k-h

-- 
Alexandre Belloni, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com