All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH] usb: dwc3: fix runtime PM in error path
@ 2016-06-10  8:54 Roger Quadros
  2016-06-10  9:16 ` Felipe Balbi
                   ` (2 more replies)
  0 siblings, 3 replies; 6+ messages in thread
From: Roger Quadros @ 2016-06-10  8:54 UTC (permalink / raw)
  To: balbi, grygorii.strashko; +Cc: linux-usb, linux-kernel, Roger Quadros

If there is a failure after pm_runtime_enable/get_sync()
we need to call pm_runtime_disable/put_sync().

Otherwise it will lead to an unbalanced pm_runtime_enable() on the
subsequent probe if the earlier probe bailed out due to -EPROBE_DEFER.

Signed-off-by: Roger Quadros <rogerq@ti.com>
---
 drivers/usb/dwc3/core.c | 19 ++++++++++++-------
 1 file changed, 12 insertions(+), 7 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 5bf38f8..222d2c9 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -1419,7 +1419,7 @@ static int dwc3_probe(struct platform_device *pdev)
 	if (ret) {
 		dev_err(dwc->dev, "failed to allocate event buffers\n");
 		ret = -ENOMEM;
-		goto err0;
+		goto err1;
 	}
 
 	if (IS_ENABLED(CONFIG_USB_DWC3_HOST) &&
@@ -1436,12 +1436,12 @@ static int dwc3_probe(struct platform_device *pdev)
 
 	ret = dwc3_alloc_scratch_buffers(dwc);
 	if (ret)
-		goto err1;
+		goto err2;
 
 	ret = dwc3_core_init(dwc);
 	if (ret) {
 		dev_err(dev, "failed to initialize core\n");
-		goto err2;
+		goto err3;
 	}
 
 	/* Check the maximum_speed parameter */
@@ -1473,23 +1473,28 @@ static int dwc3_probe(struct platform_device *pdev)
 
 	ret = dwc3_core_init_mode(dwc);
 	if (ret)
-		goto err3;
+		goto err4;
 
 	dwc3_debugfs_init(dwc);
 	pm_runtime_put(dev);
 
 	return 0;
 
-err3:
+err4:
 	dwc3_event_buffers_cleanup(dwc);
 
-err2:
+err3:
 	dwc3_free_scratch_buffers(dwc);
 
-err1:
+err2:
 	dwc3_free_event_buffers(dwc);
 	dwc3_ulpi_exit(dwc);
 
+err1:
+	pm_runtime_allow(&pdev->dev);
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
 err0:
 	/*
 	 * restore res->start back to its original value so that, in case the
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 6+ messages in thread
* Re: [PATCH] usb: dwc3: fix runtime PM in error path
  2016-06-10  8:54 [PATCH] usb: dwc3: fix runtime PM in error path Roger Quadros
@ 2016-06-10  9:16 ` Felipe Balbi
  2016-06-10  9:52   ` Roger Quadros
  2016-06-10 11:37 ` [PATCH v2] " Roger Quadros
  2016-06-10 11:38 ` Roger Quadros
  2 siblings, 1 reply; 6+ messages in thread
From: Felipe Balbi @ 2016-06-10  9:16 UTC (permalink / raw)
  To: Roger Quadros, grygorii.strashko; +Cc: linux-usb, linux-kernel, Roger Quadros

[-- Attachment #1: Type: text/plain, Size: 467 bytes --]


Hi,

Roger Quadros <rogerq@ti.com> writes:
> If there is a failure after pm_runtime_enable/get_sync()
> we need to call pm_runtime_disable/put_sync().
>
> Otherwise it will lead to an unbalanced pm_runtime_enable() on the
> subsequent probe if the earlier probe bailed out due to -EPROBE_DEFER.
>
> Signed-off-by: Roger Quadros <rogerq@ti.com>

where is this supposed to be applied? You're not checking if
pm_runtime_get() failed at all.

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: [PATCH] usb: dwc3: fix runtime PM in error path
  2016-06-10  9:16 ` Felipe Balbi
@ 2016-06-10  9:52   ` Roger Quadros
  2016-06-10 10:05     ` Felipe Balbi
  0 siblings, 1 reply; 6+ messages in thread
From: Roger Quadros @ 2016-06-10  9:52 UTC (permalink / raw)
  To: Felipe Balbi, grygorii.strashko; +Cc: linux-usb, linux-kernel


[-- Attachment #1.1: Type: text/plain, Size: 655 bytes --]

On 10/06/16 12:16, Felipe Balbi wrote:
> 
> Hi,
> 
> Roger Quadros <rogerq@ti.com> writes:
>> If there is a failure after pm_runtime_enable/get_sync()
>> we need to call pm_runtime_disable/put_sync().
>>
>> Otherwise it will lead to an unbalanced pm_runtime_enable() on the
>> subsequent probe if the earlier probe bailed out due to -EPROBE_DEFER.
>>
>> Signed-off-by: Roger Quadros <rogerq@ti.com>
> 
> where is this supposed to be applied? You're not checking if

On your testing/next

> pm_runtime_get() failed at all.
> 

I'm not adding pm_runtime_get in my patch.
Just fixing up the failure path in core.c.

cheers,
-roger


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread
* Re: [PATCH] usb: dwc3: fix runtime PM in error path
  2016-06-10  9:52   ` Roger Quadros
@ 2016-06-10 10:05     ` Felipe Balbi
  0 siblings, 0 replies; 6+ messages in thread
From: Felipe Balbi @ 2016-06-10 10:05 UTC (permalink / raw)
  To: Roger Quadros, grygorii.strashko; +Cc: linux-usb, linux-kernel

[-- Attachment #1: Type: text/plain, Size: 960 bytes --]

Roger Quadros <rogerq@ti.com> writes:

> [ Unknown signature status ]
> On 10/06/16 12:16, Felipe Balbi wrote:
>> 
>> Hi,
>> 
>> Roger Quadros <rogerq@ti.com> writes:
>>> If there is a failure after pm_runtime_enable/get_sync()
>>> we need to call pm_runtime_disable/put_sync().
>>>
>>> Otherwise it will lead to an unbalanced pm_runtime_enable() on the
>>> subsequent probe if the earlier probe bailed out due to -EPROBE_DEFER.
>>>
>>> Signed-off-by: Roger Quadros <rogerq@ti.com>
>> 
>> where is this supposed to be applied? You're not checking if
>
> On your testing/next
>
>> pm_runtime_get() failed at all.
>> 
>
> I'm not adding pm_runtime_get in my patch.
> Just fixing up the failure path in core.c.

that's fine, but you're missing this hunk:

-       pm_runtime_get_sync(dev);
+
+       ret = pm_runtime_get_sync(dev);
+       if (ret < 0)
+               goto err1;
+
        pm_runtime_forbid(dev);

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 818 bytes --]

^ permalink raw reply	[flat|nested] 6+ messages in thread
* [PATCH v2] usb: dwc3: fix runtime PM in error path
  2016-06-10  8:54 [PATCH] usb: dwc3: fix runtime PM in error path Roger Quadros
  2016-06-10  9:16 ` Felipe Balbi
@ 2016-06-10 11:37 ` Roger Quadros
  2016-06-10 11:38 ` Roger Quadros
  2 siblings, 0 replies; 6+ messages in thread
From: Roger Quadros @ 2016-06-10 11:37 UTC (permalink / raw)
  To: balbi, grygorii.strashko; +Cc: linux-usb, linux-kernel

If there is a failure after pm_runtime_enable/get_sync()
we need to call pm_runtime_disable/put_sync().

Otherwise it will lead to an unbalanced pm_runtime_enable() on the
subsequent probe if the earlier probe bailed out due to -EPROBE_DEFER.

pm_runtime_get_sync() can fail as well so deal with that case too.

Signed-off-by: Roger Quadros <rogerq@ti.com>
---
v2:
- deal with pm_runtime_get_sync() failure

 drivers/usb/dwc3/core.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 131e7eb..d51c9a9 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -959,14 +959,17 @@ static int dwc3_probe(struct platform_device *pdev)
 	pm_runtime_use_autosuspend(dev);
 	pm_runtime_set_autosuspend_delay(dev, DWC3_DEFAULT_AUTOSUSPEND_DELAY);
 	pm_runtime_enable(dev);
-	pm_runtime_get_sync(dev);
+	ret = pm_runtime_get_sync(dev);
+	if (ret < 0)
+		goto err1;
+
 	pm_runtime_forbid(dev);
 
 	ret = dwc3_alloc_event_buffers(dwc, DWC3_EVENT_BUFFERS_SIZE);
 	if (ret) {
 		dev_err(dwc->dev, "failed to allocate event buffers\n");
 		ret = -ENOMEM;
-		goto err0;
+		goto err2;
 	}
 
 	if (IS_ENABLED(CONFIG_USB_DWC3_HOST) &&
@@ -983,12 +986,12 @@ static int dwc3_probe(struct platform_device *pdev)
 
 	ret = dwc3_alloc_scratch_buffers(dwc);
 	if (ret)
-		goto err1;
+		goto err3;
 
 	ret = dwc3_core_init(dwc);
 	if (ret) {
 		dev_err(dev, "failed to initialize core\n");
-		goto err2;
+		goto err4;
 	}
 
 	/* Check the maximum_speed parameter */
@@ -1020,23 +1023,29 @@ static int dwc3_probe(struct platform_device *pdev)
 
 	ret = dwc3_core_init_mode(dwc);
 	if (ret)
-		goto err3;
+		goto err5;
 
 	dwc3_debugfs_init(dwc);
 	pm_runtime_put(dev);
 
 	return 0;
 
-err3:
+err5:
 	dwc3_event_buffers_cleanup(dwc);
 
-err2:
+err4:
 	dwc3_free_scratch_buffers(dwc);
 
-err1:
+err3:
 	dwc3_free_event_buffers(dwc);
 	dwc3_ulpi_exit(dwc);
 
+err2:
+	pm_runtime_allow(&pdev->dev);
+err1:
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
 err0:
 	/*
 	 * restore res->start back to its original value so that, in case the
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 6+ messages in thread
* [PATCH v2] usb: dwc3: fix runtime PM in error path
  2016-06-10  8:54 [PATCH] usb: dwc3: fix runtime PM in error path Roger Quadros
  2016-06-10  9:16 ` Felipe Balbi
  2016-06-10 11:37 ` [PATCH v2] " Roger Quadros
@ 2016-06-10 11:38 ` Roger Quadros
  2 siblings, 0 replies; 6+ messages in thread
From: Roger Quadros @ 2016-06-10 11:38 UTC (permalink / raw)
  To: balbi, grygorii.strashko; +Cc: linux-usb, linux-kernel, rogerq

If there is a failure after pm_runtime_enable/get_sync()
we need to call pm_runtime_disable/put_sync().

Otherwise it will lead to an unbalanced pm_runtime_enable() on the
subsequent probe if the earlier probe bailed out due to -EPROBE_DEFER.

pm_runtime_get_sync() can fail as well so deal with that case too.

Signed-off-by: Roger Quadros <rogerq@ti.com>
---
v2:
- deal with pm_runtime_get_sync() failure

 drivers/usb/dwc3/core.c | 25 +++++++++++++++++--------
 1 file changed, 17 insertions(+), 8 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index 131e7eb..d51c9a9 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -959,14 +959,17 @@ static int dwc3_probe(struct platform_device *pdev)
 	pm_runtime_use_autosuspend(dev);
 	pm_runtime_set_autosuspend_delay(dev, DWC3_DEFAULT_AUTOSUSPEND_DELAY);
 	pm_runtime_enable(dev);
-	pm_runtime_get_sync(dev);
+	ret = pm_runtime_get_sync(dev);
+	if (ret < 0)
+		goto err1;
+
 	pm_runtime_forbid(dev);
 
 	ret = dwc3_alloc_event_buffers(dwc, DWC3_EVENT_BUFFERS_SIZE);
 	if (ret) {
 		dev_err(dwc->dev, "failed to allocate event buffers\n");
 		ret = -ENOMEM;
-		goto err0;
+		goto err2;
 	}
 
 	if (IS_ENABLED(CONFIG_USB_DWC3_HOST) &&
@@ -983,12 +986,12 @@ static int dwc3_probe(struct platform_device *pdev)
 
 	ret = dwc3_alloc_scratch_buffers(dwc);
 	if (ret)
-		goto err1;
+		goto err3;
 
 	ret = dwc3_core_init(dwc);
 	if (ret) {
 		dev_err(dev, "failed to initialize core\n");
-		goto err2;
+		goto err4;
 	}
 
 	/* Check the maximum_speed parameter */
@@ -1020,23 +1023,29 @@ static int dwc3_probe(struct platform_device *pdev)
 
 	ret = dwc3_core_init_mode(dwc);
 	if (ret)
-		goto err3;
+		goto err5;
 
 	dwc3_debugfs_init(dwc);
 	pm_runtime_put(dev);
 
 	return 0;
 
-err3:
+err5:
 	dwc3_event_buffers_cleanup(dwc);
 
-err2:
+err4:
 	dwc3_free_scratch_buffers(dwc);
 
-err1:
+err3:
 	dwc3_free_event_buffers(dwc);
 	dwc3_ulpi_exit(dwc);
 
+err2:
+	pm_runtime_allow(&pdev->dev);
+err1:
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
 err0:
 	/*
 	 * restore res->start back to its original value so that, in case the
-- 
2.7.4

^ permalink raw reply related	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2016-06-10 11:38 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-06-10  8:54 [PATCH] usb: dwc3: fix runtime PM in error path Roger Quadros
2016-06-10  9:16 ` Felipe Balbi
2016-06-10  9:52   ` Roger Quadros
2016-06-10 10:05     ` Felipe Balbi
2016-06-10 11:37 ` [PATCH v2] " Roger Quadros
2016-06-10 11:38 ` Roger Quadros

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.