dmaengine: qcom: bam_dma: Handle the return value of bam_dma_resume

[PATCH] dmaengine: qcom: bam_dma: Handle the return value of bam_dma_resume

Posted by Chen Ni 1 year, 5 months ago

As pm_runtime_force_resume() can return error numbers, it should be
better to check the return value and deal with the exception.

Fixes: 0ac9c3dd0d6f ("dmaengine: qcom: bam_dma: fix runtime PM underflow")
Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
---
 drivers/dma/qcom/bam_dma.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/dma/qcom/bam_dma.c b/drivers/dma/qcom/bam_dma.c
index 5e7d332731e0..d2f5a77dfade 100644
--- a/drivers/dma/qcom/bam_dma.c
+++ b/drivers/dma/qcom/bam_dma.c
@@ -1460,9 +1460,7 @@ static int __maybe_unused bam_dma_resume(struct device *dev)
 	if (ret)
 		return ret;
 
-	pm_runtime_force_resume(dev);
-
-	return 0;
+	return pm_runtime_force_resume(dev);
 }
 
 static const struct dev_pm_ops bam_dma_pm_ops = {
-- 
2.25.1

Re: [PATCH] dmaengine: qcom: bam_dma: Handle the return value of bam_dma_resume

Posted by Markus Elfring 1 year, 5 months ago

> As pm_runtime_force_resume() can return error numbers, it should be
> better to check the return value and deal with the exception.

1. Please improve such a change description with an imperative wording.
   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.10#n94

2. Is there a need to complete the exception handling with a clk_unprepare() call?

3. How do you think about to use a summary phrase like “Return value from
   a pm_runtime_force_resume() call in bam_dma_resume()”?

4. Were any special source code analysis tools involved from your research organisation?

Regards,
Markus

Re: [PATCH] dmaengine: qcom: bam_dma: Handle the return value of bam_dma_resume

Posted by Dmitry Baryshkov 1 year, 5 months ago

On Wed, Jul 17, 2024 at 03:35:53PM GMT, Chen Ni wrote:
> As pm_runtime_force_resume() can return error numbers, it should be
> better to check the return value and deal with the exception.
> 
> Fixes: 0ac9c3dd0d6f ("dmaengine: qcom: bam_dma: fix runtime PM underflow")
> Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
> ---
>  drivers/dma/qcom/bam_dma.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/drivers/dma/qcom/bam_dma.c b/drivers/dma/qcom/bam_dma.c
> index 5e7d332731e0..d2f5a77dfade 100644
> --- a/drivers/dma/qcom/bam_dma.c
> +++ b/drivers/dma/qcom/bam_dma.c
> @@ -1460,9 +1460,7 @@ static int __maybe_unused bam_dma_resume(struct device *dev)
>  	if (ret)
>  		return ret;
>  
> -	pm_runtime_force_resume(dev);
> -
> -	return 0;
> +	return pm_runtime_force_resume(dev);

Which function will unroll the earlier clk_prepare() if we return an
error here?

>  }
>  
>  static const struct dev_pm_ops bam_dma_pm_ops = {
> -- 
> 2.25.1
> 

-- 
With best wishes
Dmitry

Re: [PATCH] dmaengine: qcom: bam_dma: Handle the return value of bam_dma_resume

Posted by Konrad Dybcio 1 year, 5 months ago

On 17.07.2024 9:35 AM, Chen Ni wrote:
> As pm_runtime_force_resume() can return error numbers, it should be
> better to check the return value and deal with the exception.
> 
> Fixes: 0ac9c3dd0d6f ("dmaengine: qcom: bam_dma: fix runtime PM underflow")
> Signed-off-by: Chen Ni <nichen@iscas.ac.cn>
> ---

Reviewed-by: Konrad Dybcio <konrad.dybcio@linaro.org>

Konrad