2020-08-09 09:00:26

by Dinghao Liu

Subject: [PATCH] ASoC: intel: Fix memleak in sst_media_open

When power_up_sst() fails, stream needs to be freed
just like when try_module_get() fails. However, current
code is returning directly and ends up leaking memory.

Fixes: 0121327c1a68b ("ASoC: Intel: mfld-pcm: add control for powering up/down dsp")
Signed-off-by: Dinghao Liu <[email protected]>
---
sound/soc/intel/atom/sst-mfld-platform-pcm.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/sound/soc/intel/atom/sst-mfld-platform-pcm.c b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
index 49b9f18472bc..79fedf9e3da1 100644
--- a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
+++ b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
@@ -330,8 +330,10 @@ static int sst_media_open(struct snd_pcm_substream *substream,
runtime->private_data = stream;

ret_val = power_up_sst(stream);
- if (ret_val < 0)
+ if (ret_val < 0) {
+ kfree(stream);
return ret_val;
+ }

/* Make sure, that the period size is always even */
snd_pcm_hw_constraint_step(substream->runtime, 0,
--
2.17.1

2020-08-10 14:36:15

by Pierre-Louis Bossart

Subject: Re: [PATCH] ASoC: intel: Fix memleak in sst_media_open

On 8/9/20 3:56 AM, Dinghao Liu wrote:
> When power_up_sst() fails, stream needs to be freed
> just like when try_module_get() fails. However, current
> code is returning directly and ends up leaking memory.
>
> Fixes: 0121327c1a68b ("ASoC: Intel: mfld-pcm: add control for powering up/down dsp")
> Signed-off-by: Dinghao Liu <[email protected]>
> ---
> sound/soc/intel/atom/sst-mfld-platform-pcm.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/sound/soc/intel/atom/sst-mfld-platform-pcm.c b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> index 49b9f18472bc..79fedf9e3da1 100644
> --- a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> +++ b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> @@ -330,8 +330,10 @@ static int sst_media_open(struct snd_pcm_substream *substream,
> runtime->private_data = stream;
>
> ret_val = power_up_sst(stream);
> - if (ret_val < 0)
> + if (ret_val < 0) {
> + kfree(stream);
> return ret_val;
> + }
>
> /* Make sure, that the period size is always even */
> snd_pcm_hw_constraint_step(substream->runtime, 0,
>

Alternate suggestion to fix this, it's not great when we mix gotos and
returns.

diff --git a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
index 49b9f18472bc..b1cac7abdc0a 100644
--- a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
+++ b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
@@ -331,7 +331,7 @@ static int sst_media_open(struct snd_pcm_substream
*substream,

ret_val = power_up_sst(stream);
if (ret_val < 0)
- return ret_val;
+ goto out_power_up;

/* Make sure, that the period size is always even */
snd_pcm_hw_constraint_step(substream->runtime, 0,
@@ -340,8 +340,9 @@ static int sst_media_open(struct snd_pcm_substream
*substream,
return snd_pcm_hw_constraint_integer(runtime,
SNDRV_PCM_HW_PARAM_PERIODS);
out_ops:
- kfree(stream);
mutex_unlock(&sst_lock);
+out_power_up:
+ kfree(stream);
return ret_val;
}

2020-08-13 08:31:57

by Dinghao Liu

Subject: Re: Re: [PATCH] ASoC: intel: Fix memleak in sst_media_open

>
> On 8/9/20 3:56 AM, Dinghao Liu wrote:
> > When power_up_sst() fails, stream needs to be freed
> > just like when try_module_get() fails. However, current
> > code is returning directly and ends up leaking memory.
> >
> > Fixes: 0121327c1a68b ("ASoC: Intel: mfld-pcm: add control for powering up/down dsp")
> > Signed-off-by: Dinghao Liu <[email protected]>
> > ---
> > sound/soc/intel/atom/sst-mfld-platform-pcm.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/sound/soc/intel/atom/sst-mfld-platform-pcm.c b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> > index 49b9f18472bc..79fedf9e3da1 100644
> > --- a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> > +++ b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> > @@ -330,8 +330,10 @@ static int sst_media_open(struct snd_pcm_substream *substream,
> > runtime->private_data = stream;
> >
> > ret_val = power_up_sst(stream);
> > - if (ret_val < 0)
> > + if (ret_val < 0) {
> > + kfree(stream);
> > return ret_val;
> > + }
> >
> > /* Make sure, that the period size is always even */
> > snd_pcm_hw_constraint_step(substream->runtime, 0,
> >
>
> Alternate suggestion to fix this, it's not great when we mix gotos and
> returns.
>
> diff --git a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> index 49b9f18472bc..b1cac7abdc0a 100644
> --- a/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> +++ b/sound/soc/intel/atom/sst-mfld-platform-pcm.c
> @@ -331,7 +331,7 @@ static int sst_media_open(struct snd_pcm_substream
> *substream,
>
> ret_val = power_up_sst(stream);
> if (ret_val < 0)
> - return ret_val;
> + goto out_power_up;
>
> /* Make sure, that the period size is always even */
> snd_pcm_hw_constraint_step(substream->runtime, 0,
> @@ -340,8 +340,9 @@ static int sst_media_open(struct snd_pcm_substream
> *substream,
> return snd_pcm_hw_constraint_integer(runtime,
> SNDRV_PCM_HW_PARAM_PERIODS);
> out_ops:
> - kfree(stream);
> mutex_unlock(&sst_lock);
> +out_power_up:
> + kfree(stream);
> return ret_val;
> }

Thank you for your advice! This is better and I will send
a new patch soon.

Regards,
Dinghao