4.19-stable review patch. If anyone has any objections, please let me know.
------------------
From: Guenter Roeck <linux@xxxxxxxxxxxx>
[ Upstream commit cbf7467828cd4ec7ceac7a8b5b5ddb2f69f07b0e ]
Using DIV_ROUND_CLOSEST() on an unbound value can result in underflows.
Indeed, module test scripts report:
temp1_max: Suspected underflow: [min=0, read 255000, written -9223372036854775808]
temp1_crit: Suspected underflow: [min=0, read 255000, written -9223372036854775808]
Fix by introducing an extra set of clamping.
Fixes: 5372d2d71c46 ("hwmon: Driver for Maxim MAX6697 and compatibles")
Reviewed-by: Tzung-Bi Shih <tzungbi@xxxxxxxxxx>
Signed-off-by: Guenter Roeck <linux@xxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/hwmon/max6697.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/hwmon/max6697.c b/drivers/hwmon/max6697.c
index 6df28fe0577da..7e7f59c68ce6e 100644
--- a/drivers/hwmon/max6697.c
+++ b/drivers/hwmon/max6697.c
@@ -314,6 +314,7 @@ static ssize_t set_temp(struct device *dev,
return ret;
mutex_lock(&data->update_lock);
+ temp = clamp_val(temp, -1000000, 1000000); /* prevent underflow */
temp = DIV_ROUND_CLOSEST(temp, 1000) + data->temp_offset;
temp = clamp_val(temp, 0, data->type == max6581 ? 255 : 127);
data->temp[nr][index] = temp;
--
2.43.0
