Skip to contentSkip to conversation

dme1737: Keep index within pwm_config[]

PATCH In review Roel Kluin 3 participants 5 messages Jul 23, 2009 – Jul 25, 2009

Subsystems (2)
lore
Share & tools
lore

Roel Kluin

author
#0 lore raw

PATCH 1/1 dme1737: Keep index within pwm_config[]

pwm_config is only 3 bytes - pwm_config[3] is out of range.

---
Credits to Parfait and Nathan Keynes,

drivers/hwmon/dme1737.c +1 -1
index 3df202a..57e8ed4 100644
@@ -1134,7 +1134,7 @@ static ssize_t show_pwm(struct device *dev, struct device_attribute *attr,
 		res = PWM_FREQ_FROM_REG(data->pwm_freq[ix]);
 		break;
 	case SYS_PWM_ENABLE:
-		if (ix > 3) {
+		if (ix >= 3) {
 			res = 1; /* pwm[5-6] hard-wired to manual mode */
 		} else {
 			res = PWM_EN_FROM_REG(data->pwm_config[ix]);

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

#1 lore raw
Quoted 11 lines · Roel Kluin

pwm_config is only 3 bytes - pwm_config[3] is out of range.

---
Credits to Parfait and Nathan Keynes,

drivers/hwmon/dme1737.c +0 -0
index 3df202a..57e8ed4 100644

drivers/hwmon/dme1737.c static ssize_t show_pwm(struct device *dev, struct device_attribute *attr, · line 1140

Quoted 8 lines
drivers/hwmon/dme1737.c +1 -1
@@ -1134,7 +1134,7 @@ static ssize_t show_pwm(struct device *dev, struct device_attribute *attr,
 		res = PWM_FREQ_FROM_REG(data->pwm_freq[ix]);
 		break;
 	case SYS_PWM_ENABLE:
-		if (ix > 3) {
+		if (ix >= 3) {
 			res = 1; /* pwm[5-6] hard-wired to manual mode */
 		} else {
 			res = PWM_EN_FROM_REG(data->pwm_config[ix]);

This code path is never called with ix = 3 (the device has no PWM4 output) so it doesn't make any difference in practice.

Signature · 6 lines
Jean Delvare

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#2 lore raw
In reply to Roel Kluin
Quoted 7 lines · On Fri, 24 Jul 2009 09:12:12 +0200 Jean Delvare <[email protected]> wrote:
Quoted 6 lines · Roel Kluin

pwm_config is only 3 bytes - pwm_config[3] is out of range.

---
Credits to Parfait and Nathan Keynes,

Quoted 17 lines

drivers/hwmon/dme1737.c static ssize_t show_pwm(struct device *dev, struct device_attribute *attr, · line 1140

Quoted 13 lines
drivers/hwmon/dme1737.c +1 -1
index 3df202a..57e8ed4 100644
@@ -1134,7 +1134,7 @@ static ssize_t show_pwm(struct device *dev, struct device_attribute *attr,
 		res = PWM_FREQ_FROM_REG(data->pwm_freq[ix]);
 		break;
 	case SYS_PWM_ENABLE:
-		if (ix > 3) {
+		if (ix >= 3) {
 			res = 1; /* pwm[5-6] hard-wired to manual mode */
 		} else {
 			res = PWM_EN_FROM_REG(data->pwm_config[ix]);

This code path is never called with ix = 3 (the device has no PWM4 output) so it doesn't make any difference in practice.

This possible bug was, I believe, detected by a static code scanner called "Parfait".

Hopefully people will continue to run Parfait against the kernel for years to come. It would be unfortunate if code sites such as this were to pop up in their reports again and again, so there is value in making the parfait warning go away, even if it is a cant-happen thing.

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

#3 lore raw
In reply to Roel Kluin
Quoted 28 lines · Andrew Morton
Quoted 7 lines · On Fri, 24 Jul 2009 09:12:12 +0200 Jean Delvare <[email protected]> wrote:
Quoted 6 lines · Roel Kluin

pwm_config is only 3 bytes - pwm_config[3] is out of range.

---
Credits to Parfait and Nathan Keynes,

Quoted 17 lines

drivers/hwmon/dme1737.c static ssize_t show_pwm(struct device *dev, struct device_attribute *attr, · line 1140

Quoted 13 lines
drivers/hwmon/dme1737.c +1 -1
index 3df202a..57e8ed4 100644
@@ -1134,7 +1134,7 @@ static ssize_t show_pwm(struct device *dev, struct device_attribute *attr,
 		res = PWM_FREQ_FROM_REG(data->pwm_freq[ix]);
 		break;
 	case SYS_PWM_ENABLE:
-		if (ix > 3) {
+		if (ix >= 3) {
 			res = 1; /* pwm[5-6] hard-wired to manual mode */
 		} else {
 			res = PWM_EN_FROM_REG(data->pwm_config[ix]);

This code path is never called with ix = 3 (the device has no PWM4 output) so it doesn't make any difference in practice.

Quoted 7 lines

This possible bug was, I believe, detected by a static code scanner called "Parfait".

Hopefully people will continue to run Parfait against the kernel for years to come. It would be unfortunate if code sites such as this were to pop up in their reports again and again, so there is value in making the parfait warning go away, even if it is a cant-happen thing.

Feel free to push the patch upstream yourself then. But please make sure the description is clear enough that it doesn't fix any actual bug.

Signature · 6 lines
Jean Delvare

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
author
#4 lore raw
In reply to Roel Kluin

The static code scanner "Parfait" reported this because pwm_config is only 3 bytes - pwm_config[3] is out of range. Since this code path is never called with ix = 3 (the device has no PWM4 output) this doesn't change anything in practice. But to encourage testing with Parfait, lets make the warning go away...

---

drivers/hwmon/dme1737.c +1 -1
index 3df202a..57e8ed4 100644
@@ -1134,7 +1134,7 @@ static ssize_t show_pwm(struct device *dev, struct device_attribute *attr,
 		res = PWM_FREQ_FROM_REG(data->pwm_freq[ix]);
 		break;
 	case SYS_PWM_ENABLE:
-		if (ix > 3) {
+		if (ix >= 3) {
 			res = 1; /* pwm[5-6] hard-wired to manual mode */
 		} else {
 			res = PWM_EN_FROM_REG(data->pwm_config[ix]);

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors