[patch] max17042_battery: fix a couple buffer overflows
Dan Carpenter
dan.carpenter at oracle.com
Thu Mar 15 22:37:32 EST 2012
There are a couple issues here caused by confusion between sizeof()
and ARRAY_SIZE(). "table_size" should be the number of elements, but we
should allocate it with kcalloc() so that we allocate the correct number
of bytes.
In max17042_init_model() we don't allocate enough space so we go past
the end of the array in max17042_read_model_data() and
max17042_model_data_compare().
In max17042_verify_model_lock() we allocate the right amount of space
but we call max17042_read_model_data() with the wrong number of elements
and also in the for loop we go past the end of the array.
Signed-off-by: Dan Carpenter <dan.carpenter at oracle.com>
---
I don't have this hardware. The original code is clearly buggy, but I
can't test that my fix is correct. Please review carefully.
diff --git a/drivers/power/max17042_battery.c b/drivers/power/max17042_battery.c
index e36763a..f8cd48c 100644
--- a/drivers/power/max17042_battery.c
+++ b/drivers/power/max17042_battery.c
@@ -328,11 +328,10 @@ static inline int max17042_model_data_compare(struct max17042_chip *chip,
static int max17042_init_model(struct max17042_chip *chip)
{
int ret;
- int table_size =
- sizeof(chip->pdata->config_data->cell_char_tbl)/sizeof(u16);
+ int table_size = ARRAY_SIZE(chip->pdata->config_data->cell_char_tbl);
u16 *temp_data;
- temp_data = kzalloc(table_size, GFP_KERNEL);
+ temp_data = kcalloc(table_size, sizeof(*temp_data), GFP_KERNEL);
if (!temp_data)
return -ENOMEM;
@@ -357,12 +356,11 @@ static int max17042_init_model(struct max17042_chip *chip)
static int max17042_verify_model_lock(struct max17042_chip *chip)
{
int i;
- int table_size =
- sizeof(chip->pdata->config_data->cell_char_tbl);
+ int table_size = ARRAY_SIZE(chip->pdata->config_data->cell_char_tbl);
u16 *temp_data;
int ret = 0;
- temp_data = kzalloc(table_size, GFP_KERNEL);
+ temp_data = kcalloc(table_size, sizeof(*temp_data), GFP_KERNEL);
if (!temp_data)
return -ENOMEM;
More information about the devicetree-discuss
mailing list