When dwc3_resume_common() returns an error, runtime pm is left in
disabled state in dwc3_resume(). The next dwc3_suspend_common()
should be skipped as the device is already in suspended state but
it's not because power.disable_depth is non-zero.
Ensures runtime PM is always re-enabled even after failed resume
attempts.

Fixes: 68c26fe58182 ("usb: dwc3: set pm runtime active before resume common")
Cc: stable@vger.kernel.org
Signed-off-by: Roy Luo <royluo@google.com>
---
 drivers/usb/dwc3/core.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index ccc3895dbd7f..1928b074b2df 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -2537,7 +2537,7 @@ static int dwc3_suspend(struct device *dev)
 static int dwc3_resume(struct device *dev)
 {
 	struct dwc3	*dwc = dev_get_drvdata(dev);
-	int		ret;
+	int		ret = 0;
 
 	pinctrl_pm_select_default_state(dev);
 
@@ -2547,12 +2547,11 @@ static int dwc3_resume(struct device *dev)
 	ret = dwc3_resume_common(dwc, PMSG_RESUME);
 	if (ret) {
 		pm_runtime_set_suspended(dev);
-		return ret;
 	}
 
 	pm_runtime_enable(dev);
 
-	return 0;
+	return ret;
 }
 
 static void dwc3_complete(struct device *dev)

base-commit: ad618736883b8970f66af799e34007475fe33a68
-- 
2.46.0.469.g59c65b2a67-goog

On Fri, Sep 06, 2024, Roy Luo wrote:
> When dwc3_resume_common() returns an error, runtime pm is left in
> disabled state in dwc3_resume(). The next dwc3_suspend_common()

What issue did you see when dwc3_suspend_common is not skipped?

BR,
Thinh

> should be skipped as the device is already in suspended state but
> it's not because power.disable_depth is non-zero.
> Ensures runtime PM is always re-enabled even after failed resume
> attempts.
> 
> Fixes: 68c26fe58182 ("usb: dwc3: set pm runtime active before resume common")
> Cc: stable@vger.kernel.org
> Signed-off-by: Roy Luo <royluo@google.com>
> ---
>  drivers/usb/dwc3/core.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
> index ccc3895dbd7f..1928b074b2df 100644
> --- a/drivers/usb/dwc3/core.c
> +++ b/drivers/usb/dwc3/core.c
> @@ -2537,7 +2537,7 @@ static int dwc3_suspend(struct device *dev)
>  static int dwc3_resume(struct device *dev)
>  {
>  	struct dwc3	*dwc = dev_get_drvdata(dev);
> -	int		ret;
> +	int		ret = 0;
>  
>  	pinctrl_pm_select_default_state(dev);
>  
> @@ -2547,12 +2547,11 @@ static int dwc3_resume(struct device *dev)
>  	ret = dwc3_resume_common(dwc, PMSG_RESUME);
>  	if (ret) {
>  		pm_runtime_set_suspended(dev);
> -		return ret;
>  	}
>  
>  	pm_runtime_enable(dev);
>  
> -	return 0;
> +	return ret;
>  }
>  
>  static void dwc3_complete(struct device *dev)
> 
> base-commit: ad618736883b8970f66af799e34007475fe33a68
> -- 
> 2.46.0.469.g59c65b2a67-goog
> 

On Fri, Sep 6, 2024 at 5:58 PM Thinh Nguyen <Thinh.Nguyen@synopsys.com> wrote:
>
> On Fri, Sep 06, 2024, Roy Luo wrote:
> > When dwc3_resume_common() returns an error, runtime pm is left in
> > disabled state in dwc3_resume(). The next dwc3_suspend_common()
>
> What issue did you see when dwc3_suspend_common is not skipped?

Apologies for the delayed response.

To answer your question, if dwc3_suspend_common() isn't skipped, it
can lead to issues because dwc->dev is already in a suspended state.
This could mean its parent devices (like the power domain or glue
driver) are also suspended and may have released resources that dwc
requires.
Consequently, calling dwc3_suspend_common() in this situation could
result in attempts to access unclocked or unpowered registers.

Regards,
Roy Luo