[PATCH -next] ASoC: stm: Fix PM disable depth imbalance in stm32_i2s_probe
The pm_runtime_enable will increase power disable depth. Thus a pairing decrement is needed on the error handling path to keep it balanced according to context.
Fixes:efc162cbd480f ("ASoC: stm: Use dev_err_probe() helper") Signed-off-by: Zhang Qilong zhangqilong3@huawei.com --- sound/soc/stm/stm32_i2s.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/sound/soc/stm/stm32_i2s.c b/sound/soc/stm/stm32_i2s.c index 6aafe793eec4..5a717443b105 100644 --- a/sound/soc/stm/stm32_i2s.c +++ b/sound/soc/stm/stm32_i2s.c @@ -1139,12 +1139,15 @@ static int stm32_i2s_probe(struct platform_device *pdev) pm_runtime_enable(&pdev->dev);
ret = snd_dmaengine_pcm_register(&pdev->dev, &stm32_i2s_pcm_config, 0); - if (ret) + if (ret) { + pm_runtime_disable(&pdev->dev); return dev_err_probe(&pdev->dev, ret, "PCM DMA register error\n"); + }
ret = snd_soc_register_component(&pdev->dev, &stm32_i2s_component, i2s->dai_drv, 1); if (ret) { + pm_runtime_disable(&pdev->dev); snd_dmaengine_pcm_unregister(&pdev->dev); return ret; }
Hello Zhang,
On 9/26/22 11:14, Zhang Qilong wrote:
The pm_runtime_enable will increase power disable depth. Thus a pairing decrement is needed on the error handling path to keep it balanced according to context.
Fixes:efc162cbd480f ("ASoC: stm: Use dev_err_probe() helper") Signed-off-by: Zhang Qilong zhangqilong3@huawei.com
sound/soc/stm/stm32_i2s.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/sound/soc/stm/stm32_i2s.c b/sound/soc/stm/stm32_i2s.c index 6aafe793eec4..5a717443b105 100644 --- a/sound/soc/stm/stm32_i2s.c +++ b/sound/soc/stm/stm32_i2s.c @@ -1139,12 +1139,15 @@ static int stm32_i2s_probe(struct platform_device *pdev) pm_runtime_enable(&pdev->dev);
ret = snd_dmaengine_pcm_register(&pdev->dev, &stm32_i2s_pcm_config, 0);
- if (ret)
if (ret) {
pm_runtime_disable(&pdev->dev);
return dev_err_probe(&pdev->dev, ret, "PCM DMA register error\n");
}
ret = snd_soc_register_component(&pdev->dev, &stm32_i2s_component, i2s->dai_drv, 1); if (ret) {
pm_runtime_disable(&pdev->dev);
snd_dmaengine_pcm_unregister(&pdev->dev); return ret; }
Seems that there the error has been introduced in this commit: 32a956a1fadf ("ASoC: stm32: i2s: add pm_runtime support") The pm_runtime_enable should be at the end of the stm32_i2s_probe as done here: https://github.com/STMicroelectronics/linux/blob/v5.15-stm32mp/sound/soc/stm...
Please, could you update your patch in this way? With also an update of the "Fixes:" reference in the commit message
Thanks, Arnaud
Hello Zhang,
On 9/26/22 11:14, Zhang Qilong wrote:
The pm_runtime_enable will increase power disable depth. Thus a pairing decrement is needed on the error handling path to keep it balanced according to context.
Fixes:efc162cbd480f ("ASoC: stm: Use dev_err_probe() helper") Signed-off-by: Zhang Qilong zhangqilong3@huawei.com
sound/soc/stm/stm32_i2s.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/sound/soc/stm/stm32_i2s.c b/sound/soc/stm/stm32_i2s.c index 6aafe793eec4..5a717443b105 100644 --- a/sound/soc/stm/stm32_i2s.c +++ b/sound/soc/stm/stm32_i2s.c @@ -1139,12 +1139,15 @@ static int stm32_i2s_probe(struct
platform_device *pdev)
pm_runtime_enable(&pdev->dev);
ret = snd_dmaengine_pcm_register(&pdev->dev,
&stm32_i2s_pcm_config, 0);
- if (ret)
- if (ret) {
return dev_err_probe(&pdev->dev, ret, "PCM DMA registerpm_runtime_disable(&pdev->dev);
error\n");
}
ret = snd_soc_register_component(&pdev->dev,
&stm32_i2s_component,
i2s->dai_drv, 1);
if (ret) {
snd_dmaengine_pcm_unregister(&pdev->dev); return ret; }pm_runtime_disable(&pdev->dev);
Seems that there the error has been introduced in this commit: 32a956a1fadf ("ASoC: stm32: i2s: add pm_runtime support") The pm_runtime_enable should be at the end of the stm32_i2s_probe as done here: https://github.com/STMicroelectronics/linux/blob/v5.15-stm32mp/sound/ soc/stm/stm32_i2s.c#L1200
Please, could you update your patch in this way? With also an update of the "Fixes:" reference in the commit message
Hi,
Very useful guide, and I have send v2 version. In addition, I found several similar problems at the same time and fixed it through a patch set.
Thanks, Zhang
Thanks, Arnaud
participants (3)
-
Arnaud POULIQUEN
-
Zhang Qilong
-
zhangqilong