When input_mt_init_slots() fails, input should be
freed to prevent memleak.
Fixes: 9a6a4193d65b8 ("HID: Add driver for USB ELAN Touchpad")
Signed-off-by: Dinghao Liu <[email protected]>
---
drivers/hid/hid-elan.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/hid/hid-elan.c b/drivers/hid/hid-elan.c
index 45c4f888b7c4..858442004258 100644
--- a/drivers/hid/hid-elan.c
+++ b/drivers/hid/hid-elan.c
@@ -188,6 +188,7 @@ static int elan_input_configured(struct hid_device *hdev, struct hid_input *hi)
ret = input_mt_init_slots(input, ELAN_MAX_FINGERS, INPUT_MT_POINTER);
if (ret) {
hid_err(hdev, "Failed to init elan MT slots: %d\n", ret);
+ input_free_device(input);
return ret;
}
--
2.17.1
On Mon, 24 Aug 2020, Dinghao Liu wrote:
> When input_mt_init_slots() fails, input should be
> freed to prevent memleak.
>
> Fixes: 9a6a4193d65b8 ("HID: Add driver for USB ELAN Touchpad")
> Signed-off-by: Dinghao Liu <[email protected]>
> ---
> drivers/hid/hid-elan.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/drivers/hid/hid-elan.c b/drivers/hid/hid-elan.c
> index 45c4f888b7c4..858442004258 100644
> --- a/drivers/hid/hid-elan.c
> +++ b/drivers/hid/hid-elan.c
> @@ -188,6 +188,7 @@ static int elan_input_configured(struct hid_device *hdev, struct hid_input *hi)
> ret = input_mt_init_slots(input, ELAN_MAX_FINGERS, INPUT_MT_POINTER);
> if (ret) {
> hid_err(hdev, "Failed to init elan MT slots: %d\n", ret);
> + input_free_device(input);
> return ret;
Good catch, but apparently it's not the only memleak there --
input_mt_init_slots() allocates the input_mt slots and friends, so we need
input_mt_destroy_slots() there as well.
Could you please add this to your patch too, while you are at fixing this
error codepath anyway, and resubmit?
Thanks!
--
Jiri Kosina
SUSE Labs
> On Mon, 24 Aug 2020, Dinghao Liu wrote:
>
> > When input_mt_init_slots() fails, input should be
> > freed to prevent memleak.
> >
> > Fixes: 9a6a4193d65b8 ("HID: Add driver for USB ELAN Touchpad")
> > Signed-off-by: Dinghao Liu <[email protected]>
> > ---
> > drivers/hid/hid-elan.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/hid/hid-elan.c b/drivers/hid/hid-elan.c
> > index 45c4f888b7c4..858442004258 100644
> > --- a/drivers/hid/hid-elan.c
> > +++ b/drivers/hid/hid-elan.c
> > @@ -188,6 +188,7 @@ static int elan_input_configured(struct hid_device *hdev, struct hid_input *hi)
> > ret = input_mt_init_slots(input, ELAN_MAX_FINGERS, INPUT_MT_POINTER);
> > if (ret) {
> > hid_err(hdev, "Failed to init elan MT slots: %d\n", ret);
> > + input_free_device(input);
> > return ret;
>
> Good catch, but apparently it's not the only memleak there --
> input_mt_init_slots() allocates the input_mt slots and friends, so we need
> input_mt_destroy_slots() there as well.
>
> Could you please add this to your patch too, while you are at fixing this
> error codepath anyway, and resubmit?
>
Thank you for your advice! I will fix this and resend the patch soon.
Regards,
Dinghao