Skip to contentSkip to conversation

pinctrl: madera: Fix uninitialized variable issue in madera_mux_set_mux

PATCH Superseded Gustavo A. R. Silva 2 participants 3 messages Oct 10, 2018

Subsystems (3)
GPIO Pin control Cirrus logic madera codec drivers

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore

Gustavo A. R. Silva

author
#0 lore raw

PATCH 1/1 pinctrl: madera: Fix uninitialized variable issue in madera_mux_set_mux

There is a potential execution path in which variable *ret* is checked in an IF statement, and then its value is used to report an error at line 659 without being properly initialized previously:

659 if (ret)
660 dev_err(priv->dev, "Failed to write to 0x%x (%d)\n", reg, ret);

Fix this by initializing variable *ret* to -1.

Addresses-Coverity-ID: 1471969 ("Uninitialized scalar variable")

Fixes 218d72a77b0b ("pinctrl: madera: Add driver for Cirrus Logic Madera codecs")
Signed-off-by Gustavo A. R. Silva <[email protected]>
---
 drivers/pinctrl/cirrus/pinctrl-madera-core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pinctrl/cirrus/pinctrl-madera-core.c +1 -1
index 0266302..a99ace8 100644
@@ -608,7 +608,7 @@ static int madera_mux_set_mux(struct pinctrl_dev *pctldev,
 	unsigned int n_chip_groups = priv->chip->n_pin_groups;
 	const char *func_name = madera_mux_funcs[selector].name;
 	unsigned int reg;
-	int i, ret;
+	int i, ret = -1;
 
 	dev_dbg(priv->dev, "%s selecting %u (%s) for group %u (%s)\n",
 		__func__, selector, func_name, group,
Signature · 1 line
2.7.4
maintainer
#1 lore raw
Quoted 21 lines · Gustavo A. R. Silva

There is a potential execution path in which variable *ret* is checked in an IF statement, and then its value is used to report an error at line 659 without being properly initialized previously:

659 if (ret)
660 dev_err(priv->dev, "Failed to write to 0x%x (%d)\n", reg, ret);

Fix this by initializing variable *ret* to -1.

Addresses-Coverity-ID: 1471969 ("Uninitialized scalar variable")

Fixes 218d72a77b0b ("pinctrl: madera: Add driver for Cirrus Logic Madera codecs")
Signed-off-by Gustavo A. R. Silva <[email protected]>
---
 drivers/pinctrl/cirrus/pinctrl-madera-core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pinctrl/cirrus/pinctrl-madera-core.c +0 -0
index 0266302..a99ace8 100644

drivers/pinctrl/cirrus/pinctrl-madera-core.c static int madera_mux_set_mux(struct pinctrl_dev *pctldev, · line 612

Quoted 6 lines
drivers/pinctrl/cirrus/pinctrl-madera-core.c +1 -1
@@ -608,7 +608,7 @@ static int madera_mux_set_mux(struct pinctrl_dev *pctldev,
 	unsigned int n_chip_groups = priv->chip->n_pin_groups;
 	const char *func_name = madera_mux_funcs[selector].name;
 	unsigned int reg;
-	int i, ret;
+	int i, ret = -1;

I don't believe this would fully address the issue since if you initialise ret to -1 you will still drop into the if statement and you will use the reg variable which should also be uninitialised on that code path.

Feels like initialising to 0 would probably be better anyway since the code path looks like the case were you have n_pins == 0 and nothing really failed in that case just nothing needed to be done.

Thanks,
Charles

Quoted 4 lines
drivers/pinctrl/cirrus/pinctrl-madera-core.c +0 -0
 	dev_dbg(priv->dev, "%s selecting %u (%s) for group %u (%s)\n",
 		__func__, selector, func_name, group,
Signature · 1 line
2.7.4
author
#2 lore raw

Hi Charles,

Quoted 10 lines

I don't believe this would fully address the issue since if you initialise ret to -1 you will still drop into the if statement and you will use the reg variable which should also be uninitialised on that code path.

Feels like initialising to 0 would probably be better anyway since the code path looks like the case were you have n_pins == 0 and nothing really failed in that case just nothing needed to be done.

You're right. I got confused with another similar issue I fixed, recently.

I'll send v2 shortly.

Thanks for the feedback.
--
Gustavo