Skip to contentSkip to conversation

media: mali-c55: add missing of_reserved_mem_device_release()

PATCH Superseded David Carlier 2 participants 10 messages Mar 27 – Mar 28

Subsystems (2)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/3 media: mali-c55: add missing of_reserved_mem_device_release()

mali_c55_probe() calls of_reserved_mem_device_init() to associate reserved memory regions with the device. This function allocates a struct rmem_assigned_device and adds it to a global linked list, which must be explicitly released via of_reserved_mem_device_release() — there is no devm variant of this API.

However, neither the probe error paths nor mali_c55_remove() called of_reserved_mem_device_release(). Any probe failure after the of_reserved_mem_device_init() call, as well as every normal device removal, leaked the reserved memory association on the global list.

Fix this by adding an err_release_mem label at the end of the probe error chain and calling of_reserved_mem_device_release() in mali_c55_remove(). The remove teardown order is also corrected to call mali_c55_media_frameworks_deinit() before kfree(), mirroring the probe init order in reverse.

Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +8 -3
index c1a562cd214e..5cb59c70ffc9 100644
@@ -806,8 +806,10 @@ static int mali_c55_probe(struct platform_device *pdev)
 	vb2_dma_contig_set_max_seg_size(dev, UINT_MAX);
 
 	ret = __mali_c55_power_on(mali_c55);
-	if (ret)
-		return dev_err_probe(dev, ret, "failed to power on\n");
+	if (ret) {
+		dev_err_probe(dev, ret, "failed to power on\n");
+		goto err_release_mem;
+	}
 
 	ret = mali_c55_check_hwcfg(mali_c55);
 	if (ret)
@@ -846,6 +848,8 @@ static int mali_c55_probe(struct platform_device *pdev)
 	kfree(mali_c55->context.registers);
 err_power_off:
 	__mali_c55_power_off(mali_c55);
+err_release_mem:
+	of_reserved_mem_device_release(dev);
 
 	return ret;
 }
@@ -854,8 +858,9 @@ static void mali_c55_remove(struct platform_device *pdev)
 {
 	struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
-	kfree(mali_c55->context.registers);
 	mali_c55_media_frameworks_deinit(mali_c55);
+	kfree(mali_c55->context.registers);
+	of_reserved_mem_device_release(&pdev->dev);
 }
 
 static const struct of_device_id mali_c55_of_match[] = {
Signature · 1 line
2.53.0
#1 lore raw

PATCH v3 2/3 media: mali-c55: add missing pm_runtime_disable() in remove

pm_runtime_enable() is called during probe but mali_c55_remove() never calls pm_runtime_disable(), leaving the device's runtime PM state enabled after the driver is unbound.

Add the missing pm_runtime_disable() call to the remove path.

Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -0
index 5cb59c70ffc9..38b11d5ba168 100644
@@ -859,6 +859,7 @@ static void mali_c55_remove(struct platform_device *pdev)
 	struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
 	mali_c55_media_frameworks_deinit(mali_c55);
+	pm_runtime_disable(&pdev->dev);
 	kfree(mali_c55->context.registers);
 	of_reserved_mem_device_release(&pdev->dev);
 }
Signature · 1 line
2.53.0
#5 lore raw

PATCH v3 3/3 media: mali-c55: fix probe error path skipping pm_runtime_disable()

In reply to David Carlier

When mali_c55_media_frameworks_init() fails, the goto target jumps to err_free_context_registers, skipping pm_runtime_disable() despite pm_runtime having already been enabled earlier in the function.

Fix this by adding an err_pm_runtime_disable label and redirecting the frameworks init failure to it, so pm_runtime is properly unwound on that error path.

Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +2 -2
index 38b11d5ba168..f998f914b355 100644
@@ -828,7 +828,7 @@ static int mali_c55_probe(struct platform_device *pdev)
 
 	ret = mali_c55_media_frameworks_init(mali_c55);
 	if (ret)
-		goto err_free_context_registers;
+		goto err_pm_runtime_disable;
 
 	pm_runtime_idle(&pdev->dev);
 
@@ -843,8 +843,8 @@ static int mali_c55_probe(struct platform_device *pdev)
 
 err_deinit_media_frameworks:
 	mali_c55_media_frameworks_deinit(mali_c55);
+err_pm_runtime_disable:
 	pm_runtime_disable(&pdev->dev);
-err_free_context_registers:
 	kfree(mali_c55->context.registers);
 err_power_off:
 	__mali_c55_power_off(mali_c55);
Signature · 1 line
2.53.0
maintainer Reviewed-by on v3 1/3
#8 lore raw
In reply to David Carlier

Hi David

Quoted 11 lines · David Carlier

mali_c55_probe() calls of_reserved_mem_device_init() to associate reserved memory regions with the device. This function allocates a struct rmem_assigned_device and adds it to a global linked list, which must be explicitly released via of_reserved_mem_device_release() — there is no devm variant of this API.

However, neither the probe error paths nor mali_c55_remove() called of_reserved_mem_device_release(). Any probe failure after the of_reserved_mem_device_init() call, as well as every normal device removal, leaked the reserved memory association on the global list.

Quoted 8 lines

Fix this by adding an err_release_mem label at the end of the probe error chain and calling of_reserved_mem_device_release() in mali_c55_remove(). The remove teardown order is also corrected to call mali_c55_media_frameworks_deinit() before kfree(), mirroring the probe init order in reverse.

Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")

With your permission, I'll take these three patches in a follow-up series that fixes the module loading/unloading.

I could also specify it as pre-requisite to b4, I'll see what's easier

Thanks
j

Quoted 45 lines
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +8 -4
index c1a562cd214e..5cb59c70ffc9 100644
@@ -806,8 +806,10 @@ static int mali_c55_probe(struct platform_device *pdev)
 	vb2_dma_contig_set_max_seg_size(dev, UINT_MAX);
 
 	ret = __mali_c55_power_on(mali_c55);
-	if (ret)
-		return dev_err_probe(dev, ret, "failed to power on\n");
+	if (ret) {
+		dev_err_probe(dev, ret, "failed to power on\n");
+		goto err_release_mem;
+	}
 
 	ret = mali_c55_check_hwcfg(mali_c55);
 	if (ret)
@@ -846,6 +848,8 @@ static int mali_c55_probe(struct platform_device *pdev)
 	kfree(mali_c55->context.registers);
 err_power_off:
 	__mali_c55_power_off(mali_c55);
+err_release_mem:
+	of_reserved_mem_device_release(dev);
 
 	return ret;
 }
@@ -854,8 +858,9 @@ static void mali_c55_remove(struct platform_device *pdev)
 {
 	struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
-	kfree(mali_c55->context.registers);
 	mali_c55_media_frameworks_deinit(mali_c55);
+	kfree(mali_c55->context.registers);
+	of_reserved_mem_device_release(&pdev->dev);
 }
 
 static const struct of_device_id mali_c55_of_match[] = {
--

2.53.0

maintainer Reviewed-by on v3 2/3
#2 lore raw

Re: [PATCH v3 2/3] media: mali-c55: add missing pm_runtime_disable() in remove

In reply to David Carlier

Hi David

Quoted 5 lines · David Carlier

pm_runtime_enable() is called during probe but mali_c55_remove() never calls pm_runtime_disable(), leaving the device's runtime PM state enabled after the driver is unbound.

Add the missing pm_runtime_disable() call to the remove path.

The driver doesn't depend on PM, so we need to explicitly power-off the peripheral in remove() (and set the pm_runtime status to suspended)

Quoted 3 lines
Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")

This patch is however correct, I'll add the power-off on top.

Quoted 20 lines
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -1
index 5cb59c70ffc9..38b11d5ba168 100644
@@ -859,6 +859,7 @@ static void mali_c55_remove(struct platform_device *pdev)
 	struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
 	mali_c55_media_frameworks_deinit(mali_c55);
+	pm_runtime_disable(&pdev->dev);
 	kfree(mali_c55->context.registers);
 	of_reserved_mem_device_release(&pdev->dev);
 }
--

2.53.0

maintainer
#6 lore raw

Re: [PATCH v3 3/3] media: mali-c55: fix probe error path skipping pm_runtime_disable()

In reply to David Carlier

Hi David

Quoted 28 lines · David Carlier

When mali_c55_media_frameworks_init() fails, the goto target jumps to err_free_context_registers, skipping pm_runtime_disable() despite pm_runtime having already been enabled earlier in the function.

Fix this by adding an err_pm_runtime_disable label and redirecting the frameworks init failure to it, so pm_runtime is properly unwound on that error path.

Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -1
index 38b11d5ba168..f998f914b355 100644
@@ -828,7 +828,7 @@ static int mali_c55_probe(struct platform_device *pdev)
 
 	ret = mali_c55_media_frameworks_init(mali_c55);
 	if (ret)
-		goto err_free_context_registers;
+		goto err_pm_runtime_disable;
 
 	pm_runtime_idle(&pdev->dev);
 

drivers/media/platform/arm/mali-c55/mali-c55-core.c static int mali_c55_probe(struct platform_device *pdev) · line 846

Quoted 4 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -0
@@ -843,8 +843,8 @@ static int mali_c55_probe(struct platform_device *pdev)
 
 err_deinit_media_frameworks:
 	mali_c55_media_frameworks_deinit(mali_c55);
+err_pm_runtime_disable:

Is it necessary to

        pm_runtime_set_suspended(&pdev->dev);

?

Quoted 9 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +0 -2
 	pm_runtime_disable(&pdev->dev);
-err_free_context_registers:
 	kfree(mali_c55->context.registers);
 err_power_off:
 	__mali_c55_power_off(mali_c55);
--

2.53.0

#9 lore raw
In reply to Jacopo Mondi
Quoted 21 lines · Jacopo Mondi

Hi David

Quoted 17 lines · David Carlier

mali_c55_probe() calls of_reserved_mem_device_init() to associate reserved memory regions with the device. This function allocates a struct rmem_assigned_device and adds it to a global linked list, which must be explicitly released via of_reserved_mem_device_release() — there is no devm variant of this API.

However, neither the probe error paths nor mali_c55_remove() called of_reserved_mem_device_release(). Any probe failure after the of_reserved_mem_device_init() call, as well as every normal device removal, leaked the reserved memory association on the global list.

Fix this by adding an err_release_mem label at the end of the probe error chain and calling of_reserved_mem_device_release() in mali_c55_remove(). The remove teardown order is also corrected to call mali_c55_media_frameworks_deinit() before kfree(), mirroring the probe init order in reverse.

Quoted 7 lines
Quoted 2 lines
Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")

With your permission, I'll take these three patches in a follow-up series that fixes the module loading/unloading.

Yes please go ahead ! Cheers,

Quoted 51 lines

I could also specify it as pre-requisite to b4, I'll see what's easier

Thanks
j

Quoted 45 lines
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +8 -4
index c1a562cd214e..5cb59c70ffc9 100644
@@ -806,8 +806,10 @@ static int mali_c55_probe(struct platform_device *pdev)
      vb2_dma_contig_set_max_seg_size(dev, UINT_MAX);
 
      ret = __mali_c55_power_on(mali_c55);
-     if (ret)
-             return dev_err_probe(dev, ret, "failed to power on\n");
+     if (ret) {
+             dev_err_probe(dev, ret, "failed to power on\n");
+             goto err_release_mem;
+     }
 
      ret = mali_c55_check_hwcfg(mali_c55);
      if (ret)
@@ -846,6 +848,8 @@ static int mali_c55_probe(struct platform_device *pdev)
      kfree(mali_c55->context.registers);
 err_power_off:
      __mali_c55_power_off(mali_c55);
+err_release_mem:
+     of_reserved_mem_device_release(dev);
 
      return ret;
 }
@@ -854,8 +858,9 @@ static void mali_c55_remove(struct platform_device *pdev)
 {
      struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
-     kfree(mali_c55->context.registers);
      mali_c55_media_frameworks_deinit(mali_c55);
+     kfree(mali_c55->context.registers);
+     of_reserved_mem_device_release(&pdev->dev);
 }
 
 static const struct of_device_id mali_c55_of_match[] = {
--

2.53.0

#7 lore raw

Re: [PATCH v3 3/3] media: mali-c55: fix probe error path skipping pm_runtime_disable()

In reply to Jacopo Mondi
Quoted 39 lines · Jacopo Mondi

Hi David

Quoted 28 lines · David Carlier

When mali_c55_media_frameworks_init() fails, the goto target jumps to err_free_context_registers, skipping pm_runtime_disable() despite pm_runtime having already been enabled earlier in the function.

Fix this by adding an err_pm_runtime_disable label and redirecting the frameworks init failure to it, so pm_runtime is properly unwound on that error path.

Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -1
index 38b11d5ba168..f998f914b355 100644
@@ -828,7 +828,7 @@ static int mali_c55_probe(struct platform_device *pdev)
 
      ret = mali_c55_media_frameworks_init(mali_c55);
      if (ret)
-             goto err_free_context_registers;
+             goto err_pm_runtime_disable;
 
      pm_runtime_idle(&pdev->dev);
 

drivers/media/platform/arm/mali-c55/mali-c55-core.c static int mali_c55_probe(struct platform_device *pdev) · line 846

Quoted 4 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -0
@@ -843,8 +843,8 @@ static int mali_c55_probe(struct platform_device *pdev)
 
 err_deinit_media_frameworks:
      mali_c55_media_frameworks_deinit(mali_c55);
+err_pm_runtime_disable:

Is it necessary to

Quoted 10 lines
        pm_runtime_set_suspended(&pdev->dev);

?

Quoted 6 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +0 -2
      pm_runtime_disable(&pdev->dev);
-err_free_context_registers:
      kfree(mali_c55->context.registers);
 err_power_off:
      __mali_c55_power_off(mali_c55);
--

Hi,

Good point — yes, pm_runtime_set_active() is called earlier in
probe, so the error unwind should also call pm_runtime_set_suspended() to undo it and decrement the
parent's usage count. Same applies to the remove path.

I'll send a v4 with that added to both patches 2 and 3.

Thanks for the review,
David

Quoted 3 lines
Quoted 3 lines

2.53.0

maintainer
#3 lore raw

Re: [PATCH v3 2/3] media: mali-c55: add missing pm_runtime_disable() in remove

In reply to Jacopo Mondi

Hi David
since you have to send a v4 anyway, would you care about also
powering off the device duing remove ?

I have a patch, but since you're already looking at this it doesn't make much sense to pile another one on top.

Could you squash in the below suggestions ?

Quoted 6 lines · Jacopo Mondi

Hi David

Quoted 3 lines · David Carlier

pm_runtime_enable() is called during probe but mali_c55_remove() never calls pm_runtime_disable(), leaving the device's runtime PM state enabled after the driver is unbound.

I would use these as the commit message and the commit title:

media: mali-c55: Power-off the peripheral in remove()

The Mali C55 driver doesn't depend on PM. For this reason, if pm_runtime is not compiled in it is required to manually power-off the peripheral during the driver's remove() handler.

Also pm_runtime_enable() is called during probe but mali_c55_remove() never calls pm_runtime_disable(), leaving the device's runtime PM state enabled after the driver is unbound.

Manually power-off the peripheral in remove() if the peripheral has not been suspended using runtime_pm and disable runtime pm.

Quoted 19 lines
Quoted 2 lines

Add the missing pm_runtime_disable() call to the remove path.

The driver doesn't depend on PM, so we need to explicitly power-off the peripheral in remove() (and set the pm_runtime status to suspended)

Quoted 3 lines
Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")

This patch is however correct, I'll add the power-off on top.

Quoted 4 lines
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 1 +
 1 file changed, 1 insertion(+)
Quoted 8 lines

drivers/media/platform/arm/mali-c55/mali-c55-core.c static void mali_c55_remove(struct platform_device *pdev) · line 861

Quoted 8 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +0 -0
index 5cb59c70ffc9..38b11d5ba168 100644
@@ -859,6 +859,7 @@ static void mali_c55_remove(struct platform_device *pdev)
 	struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
 	mali_c55_media_frameworks_deinit(mali_c55);

And this should become:

+ if (!pm_runtime_suspended(&pdev->dev)) {
+ __mali_c55_power_off(mali_c55);
+ pm_runtime_set_suspended(&pdev->dev);
+ }

Quoted 1 line

drivers/media/platform/arm/mali-c55/mali-c55-core.c static void mali_c55_remove(struct platform_device *pdev) · line 862

Quoted 1 line
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -0
+	pm_runtime_disable(&pdev->dev);

Otherwise, if you want to keep your patch simpler, I'll add a patch for power-off on top.

Thanks
j

Quoted 7 lines
Quoted 7 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +0 -1
 	kfree(mali_c55->context.registers);
 	of_reserved_mem_device_release(&pdev->dev);
 }
--

2.53.0

#4 lore raw

Re: [PATCH v3 2/3] media: mali-c55: add missing pm_runtime_disable() in remove

Hi sorry I missed your previous messages ok looking into it. Cheers.

Quoted 82 lines · Jacopo Mondi

Hi David
since you have to send a v4 anyway, would you care about also
powering off the device duing remove ?

I have a patch, but since you're already looking at this it doesn't make much sense to pile another one on top.

Could you squash in the below suggestions ?

Quoted 6 lines · Jacopo Mondi

Hi David

Quoted 3 lines · David Carlier

pm_runtime_enable() is called during probe but mali_c55_remove() never calls pm_runtime_disable(), leaving the device's runtime PM state enabled after the driver is unbound.

I would use these as the commit message and the commit title:

media: mali-c55: Power-off the peripheral in remove()

The Mali C55 driver doesn't depend on PM. For this reason, if pm_runtime is not compiled in it is required to manually power-off the peripheral during the driver's remove() handler.

Also pm_runtime_enable() is called during probe but mali_c55_remove() never calls pm_runtime_disable(), leaving the device's runtime PM state enabled after the driver is unbound.

Manually power-off the peripheral in remove() if the peripheral has not been suspended using runtime_pm and disable runtime pm.

Quoted 19 lines
Quoted 2 lines

Add the missing pm_runtime_disable() call to the remove path.

The driver doesn't depend on PM, so we need to explicitly power-off the peripheral in remove() (and set the pm_runtime status to suspended)

Quoted 3 lines
Fixes d5f281f3dd29 ("media: mali-c55: Add Mali-C55 ISP driver")

This patch is however correct, I'll add the power-off on top.

Quoted 4 lines
---
 drivers/media/platform/arm/mali-c55/mali-c55-core.c | 1 +
 1 file changed, 1 insertion(+)
Quoted 8 lines

drivers/media/platform/arm/mali-c55/mali-c55-core.c static void mali_c55_remove(struct platform_device *pdev) · line 861

Quoted 8 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +0 -0
index 5cb59c70ffc9..38b11d5ba168 100644
@@ -859,6 +859,7 @@ static void mali_c55_remove(struct platform_device *pdev)
    struct mali_c55 *mali_c55 = platform_get_drvdata(pdev);
 
    mali_c55_media_frameworks_deinit(mali_c55);

And this should become:

(quoted diff) +4 -0
+       if (!pm_runtime_suspended(&pdev->dev)) {
+               __mali_c55_power_off(mali_c55);
+               pm_runtime_set_suspended(&pdev->dev);
+       }
Quoted 1 line

drivers/media/platform/arm/mali-c55/mali-c55-core.c static void mali_c55_remove(struct platform_device *pdev) · line 862

Quoted 1 line
drivers/media/platform/arm/mali-c55/mali-c55-core.c +1 -0
+   pm_runtime_disable(&pdev->dev);

Otherwise, if you want to keep your patch simpler, I'll add a patch for power-off on top.

Thanks
j

Quoted 7 lines
Quoted 7 lines
drivers/media/platform/arm/mali-c55/mali-c55-core.c +0 -1
    kfree(mali_c55->context.registers);
    of_reserved_mem_device_release(&pdev->dev);
 }
--

2.53.0