Skip to contentSkip to conversation

iio: imu: bno055: uninitialized variable bug in bno055_trigger_handler()

PATCH Applied Dan Carpenter 4 participants 5 messages Oct 14, 2022 – Nov 7, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 iio: imu: bno055: uninitialized variable bug in bno055_trigger_handler()

This bug is basically harmless, although it will trigger a runtime warning if you use KMSan. On the first iteration through the loop, the "best_delta" variable is uninitialized so re-order the condition to prevent reading uninitialized memory.

Fixes 4aefe1c2bd0c ("iio: imu: add Bosch Sensortec BNO055 core driver")
---
 drivers/iio/imu/bno055/bno055.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iio/imu/bno055/bno055.c +1 -1
index 307557a609e3..52744dd98e65 100644
@@ -632,7 +632,7 @@ static int bno055_set_regmask(struct bno055_priv *priv, int val, int val2,
 			return -EINVAL;
 		}
 		delta = abs(tbl_val - req_val);
-		if (delta < best_delta || first) {
+		if (first || delta < best_delta) {
 			best_delta = delta;
 			hwval = i;
 			first = false;
Signature · 1 line
2.35.1
reviewer Acked-by on v1 1/1
#1 lore raw
Quoted 14 lines

-----Original Message-----
From: Dan Carpenter <[email protected]>
Sent: Friday, October 14, 2022 11:40 AM
To: Jonathan Cameron <[email protected]>; Andrea Merello <[email protected]>
Cc: Lars-Peter Clausen <[email protected]>; [email protected]; [email protected]
Subject: [PATCH] iio: imu: bno055: uninitialized variable bug in bno055_trigger_handler()

[External]

Quoted 8 lines

This bug is basically harmless, although it will trigger a runtime warning if you use KMSan. On the first iteration through the loop, the "best_delta" variable is uninitialized so re-order the condition to prevent reading uninitialized memory.

Fixes 4aefe1c2bd0c ("iio: imu: add Bosch Sensortec BNO055 core driver")

---

maintainer
#2 lore raw
In reply to Dan Carpenter
Quoted 7 lines · On Fri, 14 Oct 2022 12:39:52 +0300 Dan Carpenter <[email protected]>…

This bug is basically harmless, although it will trigger a runtime warning if you use KMSan. On the first iteration through the loop, the "best_delta" variable is uninitialized so re-order the condition to prevent reading uninitialized memory.

Fixes 4aefe1c2bd0c ("iio: imu: add Bosch Sensortec BNO055 core driver")

You reported this a while back along with a second issue (false positive) with hwval.

I posted a patch fixing both
https://lore.kernel.org/linux-iio/[email protected]/

I don't really care which patch goes in, but curious to reasoning to not also deal with the hwval warning here?

Jonathan

Quoted 17 lines
---
 drivers/iio/imu/bno055/bno055.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iio/imu/bno055/bno055.c +1 -1
index 307557a609e3..52744dd98e65 100644
@@ -632,7 +632,7 @@ static int bno055_set_regmask(struct bno055_priv *priv, int val, int val2,
 			return -EINVAL;
 		}
 		delta = abs(tbl_val - req_val);
-		if (delta < best_delta || first) {
+		if (first || delta < best_delta) {
 			best_delta = delta;
 			hwval = i;
 			first = false;
maintainer Applied
#3 lore raw
Quoted 11 lines · On Sat, 15 Oct 2022 17:33:59 +0100 Jonathan Cameron <[email protected]> wrote:
Quoted 7 lines · On Fri, 14 Oct 2022 12:39:52 +0300 Dan Carpenter <[email protected]>…

This bug is basically harmless, although it will trigger a runtime warning if you use KMSan. On the first iteration through the loop, the "best_delta" variable is uninitialized so re-order the condition to prevent reading uninitialized memory.

Fixes 4aefe1c2bd0c ("iio: imu: add Bosch Sensortec BNO055 core driver")
Quoted 9 lines

You reported this a while back along with a second issue (false positive) with hwval.

I posted a patch fixing both
https://lore.kernel.org/linux-iio/[email protected]/

I don't really care which patch goes in, but curious to reasoning to not also deal with the hwval warning here?

Meh. Rather than not applying either patch, I'll pick this one up. We can deal with the hwval warning at a later date if necessary.

Applied to the fixes-togreg branch of iio.git.

Jonathan

Quoted 20 lines

Jonathan

Quoted 17 lines
---
 drivers/iio/imu/bno055/bno055.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iio/imu/bno055/bno055.c +1 -1
index 307557a609e3..52744dd98e65 100644
@@ -632,7 +632,7 @@ static int bno055_set_regmask(struct bno055_priv *priv, int val, int val2,
 			return -EINVAL;
 		}
 		delta = abs(tbl_val - req_val);
-		if (delta < best_delta || first) {
+		if (first || delta < best_delta) {
 			best_delta = delta;
 			hwval = i;
 			first = false;  
#4 lore raw
Quoted 17 lines · Jonathan Cameron
Quoted 14 lines · On Sat, 15 Oct 2022 17:33:59 +0100 Jonathan Cameron <[email protected]> wrote:
Quoted 7 lines · On Fri, 14 Oct 2022 12:39:52 +0300 Dan Carpenter <[email protected]>…

This bug is basically harmless, although it will trigger a runtime warning if you use KMSan. On the first iteration through the loop, the "best_delta" variable is uninitialized so re-order the condition to prevent reading uninitialized memory.

Fixes 4aefe1c2bd0c ("iio: imu: add Bosch Sensortec BNO055 core driver")

You reported this a while back along with a second issue (false positive) with hwval.

Quoted 9 lines
Quoted 6 lines

I posted a patch fixing both
https://lore.kernel.org/linux-iio/[email protected]/

I don't really care which patch goes in, but curious to reasoning to not also deal with the hwval warning here?

Meh. Rather than not applying either patch, I'll pick this one up. We can deal with the hwval warning at a later date if necessary.

Sorry, I am behind in my email.

There was a couple weeks in between sending the bug report and the patch so I forgot about it. Also I send those bug reports from a different system so I don't check that as part of the QC process. I'm moving to checking vger instead of grepping my outbox but I haven't automated that yet.

I don't get a "hwval" warning on my system so using the cross function DB must have silenced the false positive.

regards,
dan carpenter