cpufreq: exynos: Fix for memory leak in case SoC name does not match

PATCH In review Kukjin Kim 4 participants 6 messages Aug 4, 2015 – Aug 14, 2015

Subsystems (2)

Arm Samsung Exynos CPU frequency scaling framework

lore

Share & tools

lore

Kukjin Kim

author

2015-08-04 07:45 UTC #0 lore raw

PATCH 1/1 cpufreq: exynos: Fix for memory leak in case SoC name does not match

From: Shailendra Verma <[email protected]>

During probe free the memory allocated to "exynos_info" in case of unknown SoC type.

Signed-off-by Shailendra Verma <[email protected]>

Acked-by Viresh Kumar <[email protected]>

Acked-by Lukasz Majewski <[email protected]>

[k.kozlowski: Rebased the patch around if(of_machine_is_compatible)]

Signed-off-by Krzysztof Kozlowski <[email protected]>

Signed-off-by Kukjin Kim <[email protected]>

---
Hi Rafael and Viresh,

Since this is a fix to prevent memory leak, it would be better if this could be sent to linus during -rc via cpufreq tree.

Thanks,
Kukjin

 drivers/cpufreq/exynos-cpufreq.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

drivers/cpufreq/exynos-cpufreq.c +4 -2

index ae5b2bd..fa3dd84 100644
@@ -180,7 +180,7 @@ static int exynos_cpufreq_probe(struct platform_device *pdev)
 		ret = exynos5250_cpufreq_init(exynos_info);
 	} else {
 		pr_err("%s: Unknown SoC type\n", __func__);
-		return -ENODEV;
+		ret = -ENODEV;
 	}
 
 	if (ret)

@@ -188,12 +188,14 @@ static int exynos_cpufreq_probe(struct platform_device *pdev)
 
 	if (exynos_info->set_freq == NULL) {
 		dev_err(&pdev->dev, "No set_freq function (ERR)\n");
+		ret = -EINVAL;
 		goto err_vdd_arm;
 	}
 
 	arm_regulator = regulator_get(NULL, "vdd_arm");
 	if (IS_ERR(arm_regulator)) {
 		dev_err(&pdev->dev, "failed to get resource vdd_arm\n");
+		ret = -EINVAL;
 		goto err_vdd_arm;
 	}

@@ -225,7 +227,7 @@ err_cpufreq_reg:
 	regulator_put(arm_regulator);
 err_vdd_arm:
 	kfree(exynos_info);
-	return -EINVAL;
+	return ret;
 }
 
 static struct platform_driver exynos_cpufreq_platdrv = {

Signature · 1 line

2.0.0

Rafael J. Wysocki

2015-08-05 23:56 UTC #1 lore raw

Quoted 11 lines · Kukjin Kim

From: Shailendra Verma <[email protected]>
During probe free the memory allocated to "exynos_info" in case of unknown SoC type.
Signed-off-by Shailendra Verma <[email protected]>
Acked-by Viresh Kumar <[email protected]>
Acked-by Lukasz Majewski <[email protected]>
[k.kozlowski: Rebased the patch around if(of_machine_is_compatible)]
Signed-off-by Krzysztof Kozlowski <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>

Quoted 5 lines

---
Hi Rafael and Viresh,
Since this is a fix to prevent memory leak, it would be better if this could be sent to linus during -rc via cpufreq tree.

I'm not planning more pull requests for 4.2 and this one doesn't seem to be a good enough reason for a pull request of its own, though.

Of course, if there are more urgent fixes, I can push it along with them, but there are none so far.

Thanks,
Rafael

Krzysztof Kozlowski

2015-08-08 07:36 UTC #2 lore raw

W dniu 06.08.2015 o 09:23, Rafael J. Wysocki pisze:

Quoted 15 lines · Kukjin Kim

Quoted 14 lines
From: Shailendra Verma <[email protected]>
During probe free the memory allocated to "exynos_info" in case of unknown SoC type.
Signed-off-by Shailendra Verma <[email protected]>
Acked-by Viresh Kumar <[email protected]>
Acked-by Lukasz Majewski <[email protected]>
[k.kozlowski: Rebased the patch around if(of_machine_is_compatible)]
Signed-off-by Krzysztof Kozlowski <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>
---
Hi Rafael and Viresh,

Quoted 8 lines

Quoted 2 lines
Since this is a fix to prevent memory leak, it would be better if this could be sent to linus during -rc via cpufreq tree.
I'm not planning more pull requests for 4.2 and this one doesn't seem to be a good enough reason for a pull request of its own, though.
Of course, if there are more urgent fixes, I can push it along with them, but there are none so far.

The cpufreq driver will be removed completely in v4.3 or v4.4 with patchset adding cpufreq-dt support for Exynos 4x12. This means that this patch makes sense only for 4.2 and as a stable backport (but it was not marked as such).

Anyone thinks that the patch will be worth stable back-porting?

Best regards,
Krzysztof

Viresh Kumar

maintainer

2015-08-08 13:29 UTC #3 lore raw

Quoted 6 lines · Krzysztof Kozlowski

The cpufreq driver will be removed completely in v4.3 or v4.4 with patchset adding cpufreq-dt support for Exynos 4x12. This means that this patch makes sense only for 4.2 and as a stable backport (but it was not marked as such).
Anyone thinks that the patch will be worth stable back-porting?

Doesn't look worth enough to me..

Signature · 1 line

viresh

Rafael J. Wysocki

2015-08-14 09:14 UTC #4 lore raw

In reply to Krzysztof Kozlowski

Quoted 22 lines · Krzysztof Kozlowski

W dniu 06.08.2015 o 09:23, Rafael J. Wysocki pisze:
Quoted 21 lines · Kukjin Kim
Quoted 11 lines
From: Shailendra Verma <[email protected]>
During probe free the memory allocated to "exynos_info" in case of unknown SoC type.
Signed-off-by Shailendra Verma <[email protected]>
Acked-by Viresh Kumar <[email protected]>
Acked-by Lukasz Majewski <[email protected]>
[k.kozlowski: Rebased the patch around if(of_machine_is_compatible)]
Signed-off-by Krzysztof Kozlowski <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>
Quoted 5 lines
---
Hi Rafael and Viresh,
Since this is a fix to prevent memory leak, it would be better if this could be sent to linus during -rc via cpufreq tree.
I'm not planning more pull requests for 4.2 and this one doesn't seem to be a good enough reason for a pull request of its own, though.

Quoted 7 lines

Quoted 2 lines
Of course, if there are more urgent fixes, I can push it along with them, but there are none so far.
The cpufreq driver will be removed completely in v4.3 or v4.4 with patchset adding cpufreq-dt support for Exynos 4x12. This means that this patch makes sense only for 4.2 and as a stable backport (but it was not marked as such).

Since I have one more ACPI fix for 4.2, I've applied this one and will push it to Linus next week.

Thanks,
Rafael

Kukjin Kim

author

2015-08-14 09:34 UTC #5 lore raw

Quoted 27 lines · Rafael J. Wysocki

Quoted 25 lines · Krzysztof Kozlowski
W dniu 06.08.2015 o 09:23, Rafael J. Wysocki pisze:
Quoted 23 lines · Kukjin Kim
Quoted 11 lines
From: Shailendra Verma <[email protected]>
During probe free the memory allocated to "exynos_info" in case of unknown SoC type.
Signed-off-by Shailendra Verma <[email protected]>
Acked-by Viresh Kumar <[email protected]>
Acked-by Lukasz Majewski <[email protected]>
[k.kozlowski: Rebased the patch around if(of_machine_is_compatible)]
Signed-off-by Krzysztof Kozlowski <[email protected]>
Signed-off-by Kukjin Kim <[email protected]>
Quoted 5 lines
---
Hi Rafael and Viresh,
Since this is a fix to prevent memory leak, it would be better if this could be sent to linus during -rc via cpufreq tree.
I'm not planning more pull requests for 4.2 and this one doesn't seem to be a good enough reason for a pull request of its own, though.
Of course, if there are more urgent fixes, I can push it along with them, but there are none so far.

Quoted 8 lines

Quoted 4 lines
The cpufreq driver will be removed completely in v4.3 or v4.4 with patchset adding cpufreq-dt support for Exynos 4x12. This means that this patch makes sense only for 4.2 and as a stable backport (but it was not marked as such).
Since I have one more ACPI fix for 4.2, I've applied this one and will push it to Linus next week.

Rafael, thanks.

- Kukjin