Input: force feedback - potential integer wrap in input_ff_create()
The problem here is that max_effects can wrap on 32 bits systems.
We'd allocate a smaller amount of data than sizeof(struct ff_device).
The call to kcalloc() on the next line would fail but it would write
the NULL return outside of the memory we just allocated causing data
corruption.
The call path is that uinput_setup_device() get ->ff_effects_max from
the user and sets the value in the ->private_data struct. From there
it is:
-> uinput_ioctl_handler()
-> uinput_create_device()
-> input_ff_create(dev, udev->ff_effects_max);
I've also changed ff_effects_max so it's an unsigned int instead of
a signed int as a cleanup.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
diff --git a/drivers/input/ff-core.c b/drivers/input/ff-core.c
index 3367f76..480eb9d 100644
--- a/drivers/input/ff-core.c
+++ b/drivers/input/ff-core.c
@@ -309,9 +309,10 @@
* Once ff device is created you need to setup its upload, erase,
* playback and other handlers before registering input device
*/
-int input_ff_create(struct input_dev *dev, int max_effects)
+int input_ff_create(struct input_dev *dev, unsigned int max_effects)
{
struct ff_device *ff;
+ size_t ff_dev_size;
int i;
if (!max_effects) {
@@ -319,8 +320,12 @@
return -EINVAL;
}
- ff = kzalloc(sizeof(struct ff_device) +
- max_effects * sizeof(struct file *), GFP_KERNEL);
+ ff_dev_size = sizeof(struct ff_device) +
+ max_effects * sizeof(struct file *);
+ if (ff_dev_size < max_effects) /* overflow */
+ return -EINVAL;
+
+ ff = kzalloc(ff_dev_size, GFP_KERNEL);
if (!ff)
return -ENOMEM;