Re: [PATCH v2] ALSA: hda: cs35l41: Fix ACPI device reference leak on probe failure

[Takashi Iwai]

On Sun, 11 Oct 2026 13:42:51 +0200,
Ma Ke wrote:
> 
> cs35l41_hda_probe() takes an acpi_device through
> cs35l41_hda_read_acpi() and stores it in cs35l41->dacpi. It is
> released in the err: label, which
> is reached from every later failure.
> 
> One failure is not routed there:
> 
>         ret = cs35l41_hda_read_acpi(cs35l41, device_name, id);
>         if (ret)
>                 return dev_err_probe(cs35l41->dev, ret, "...");
> 
> cs35l41->acpi_subsystem_id is allocated inside cs35l41_hda_read_acpi()
> before that point, so it is leaked whenever the function fails.
> 
> Route the failure through a new err_acpi label that releases it.
> 
> That label must not release cs35l41->dacpi a second time.  Every
> failure path in cs35l41_hda_read_acpi() that can be taken after the
> assignment already drops that reference: cs35l41_hda_parse_acpi() does
> so in its own err: label, and the !physdev path does it directly.
> Clear cs35l41->dacpi wherever it is dropped, so the pointer never
> outlives the reference and the put in err_acpi is a no-op on those
> paths.
> 
> Found by code review.
> 
> Fixes: 447106e92a0c ("ALSA: hda: cs35l41: Support mute notifications for CS35L41 HDA")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Ma Ke <make_ruc2021@xxxxxxx>

Now Sashiko complains other things; please verify whether they are
valid or not.

Through those developments, I'm thinking whether it'd be cleaner
rather to move the unref and kfree into the caller side, i.e. doing it
only in cs35l41_hda_probe() error path (but conditionally), instead,
so something like below.


thanks,

Takashi

diff --git a/sound/hda/codecs/side-codecs/cs35l41_hda.c b/sound/hda/codecs/side-codecs/cs35l41_hda.c
index 62296be45c55..e6f0ae39876d 100644
--- a/sound/hda/codecs/side-codecs/cs35l41_hda.c
+++ b/sound/hda/codecs/side-codecs/cs35l41_hda.c
@@ -1907,7 +1907,6 @@ int cs35l41_hda_parse_acpi(struct cs35l41_hda *cs35l41, struct device *physdev,
 	hw_cfg->valid = false;
 	hw_cfg->gpio1.valid = false;
 	hw_cfg->gpio2.valid = false;
-	acpi_dev_put(cs35l41->dacpi);
 
 	return ret;
 }
@@ -1928,10 +1927,8 @@ static int cs35l41_hda_read_acpi(struct cs35l41_hda *cs35l41, const char *hid, i
 	cs35l41->dacpi = adev;
 	struct device *physdev __free(put_device) =
 		get_device(acpi_get_first_physical_node(adev));
-	if (!physdev) {
-		acpi_dev_put(adev);
+	if (!physdev)
 		return -ENODEV;
-	}
 
 	sub = acpi_get_subsystem_id(ACPI_HANDLE(physdev));
 	if (IS_ERR(sub))
@@ -1986,8 +1983,10 @@ int cs35l41_hda_probe(struct device *dev, const char *device_name, int id, int i
 	dev_set_drvdata(dev, cs35l41);
 
 	ret = cs35l41_hda_read_acpi(cs35l41, device_name, id);
-	if (ret)
-		return dev_err_probe(cs35l41->dev, ret, "Platform not supported\n");
+	if (ret) {
+		dev_err_probe(cs35l41->dev, ret, "Platform not supported\n");
+		goto err_acpi;
+	}
 
 	if (IS_ERR(cs35l41->reset_gpio)) {
 		ret = PTR_ERR(cs35l41->reset_gpio);
@@ -2076,6 +2075,7 @@ int cs35l41_hda_probe(struct device *dev, const char *device_name, int id, int i
 		gpiod_set_value_cansleep(cs35l41->reset_gpio, 0);
 	gpiod_put(cs35l41->reset_gpio);
 	gpiod_put(cs35l41->cs_gpio);
+err_acpi:
 	acpi_dev_put(cs35l41->dacpi);
 	kfree(cs35l41->acpi_subsystem_id);