Re: [PATCH v2] HID: wacom: fix NULL pointer dereference in wacom_intuos_pad()

[Dmitry Torokhov]

On Sun, May 24, 2026 at 12:06:19AM +0900, Jinmo Yang wrote:
> wacom_intuos_pad() dereferences wacom->pad_input without a NULL check.
> When a Wacom tablet is created via UHID with parameters that route
> through wacom_bpt_irq() -> wacom_intuos_irq() -> wacom_intuos_pad(),
> but probe did not allocate pad_input, the call to
> wacom_report_numbered_buttons() passes a NULL input_dev, causing a
> general protection fault in input_get_drvdata().
> 
> Add a NULL check for pad_input at the top of wacom_intuos_pad() to
> bail out early when the pad input device was not set up.
> 
> The bug was found by syzkaller on linux-next 7.1.0-rc4 (next-20260522):
> 
>   Oops: general protection fault, probably for non-canonical address 0xdffffc0000000052
>   KASAN: null-ptr-deref in range [0x0000000000000290-0x0000000000000297]
>   RIP: 0010:input_get_drvdata include/linux/input.h:390 [inline]
>   RIP: 0010:wacom_report_numbered_buttons+0x37/0x210 drivers/hid/wacom_wac.c:4210
>   Call Trace:
>    wacom_intuos_pad drivers/hid/wacom_wac.c:643 [inline]
>    wacom_intuos_irq+0x29a/0x32b0 drivers/hid/wacom_wac.c:1042
>    wacom_bpt_irq drivers/hid/wacom_wac.c:3275 [inline]
>    wacom_wac_irq+0x12ed/0xaef0 drivers/hid/wacom_wac.c:3545
>    wacom_raw_event+0x677/0xa90 drivers/hid/wacom_sys.c:184
>    __hid_input_report.constprop.0+0x39a/0x4d0 drivers/hid/hid-core.c:2161
>    uhid_dev_input2 drivers/hid/uhid.c:618 [inline]
>    uhid_char_write+0xa8a/0xfa0 drivers/hid/uhid.c:776
> 
> Also confirmed on a Pixel 9 Pro (Android 16, kernel 6.1.124) where
> it causes an immediate kernel panic and reboot via /dev/uhid without
> requiring root privileges:
> 
>   KP: Oops: Fatal exception: comm:wacom_27qhdt
>   Reboot reason: 0xbaba - Kernel PANIC
> 
> Reproducer (unprivileged):
>   open("/dev/uhid", O_RDWR)
>   write(fd, UHID_CREATE2{vendor=0x056a, product=0x0020})
>   write(fd, UHID_INPUT2{report_id=0x0c, size=10})
> 
> Fixes: c7f0522a1ad1 ("HID: wacom: Slim down wacom_intuos_pad processing")
> Signed-off-by: Jinmo Yang <jinmo44.yang@xxxxxxxxx>
> ---
>  drivers/hid/wacom_wac.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/hid/wacom_wac.c b/drivers/hid/wacom_wac.c
> index da1f0ea85..251ddda3e 100644
> --- a/drivers/hid/wacom_wac.c
> +++ b/drivers/hid/wacom_wac.c
> @@ -515,7 +515,6 @@ static int wacom_intuos_pad(struct wacom_wac *wacom)
>  	struct wacom_features *features = &wacom->features;
>  	unsigned char *data = wacom->data;
>  	struct input_dev *input = wacom->pad_input;
> -	int i;
>  	int buttons = 0, nbuttons = features->numbered_buttons;
>  	int keys = 0, nkeys = 0;
>  	int ring1 = 0, ring2 = 0;
> @@ -523,6 +522,10 @@ static int wacom_intuos_pad(struct wacom_wac *wacom)
>  	bool prox = false;
>  	bool wrench = false, keyboard = false, mute_touch = false, menu = false,
>  	     info = false;
> +	int i;
> +
> +	if (!input)
> +		return 0;
>  
>  	/* pad packets. Works as a second tool and is always in prox */
>  	if (!(data[0] == WACOM_REPORT_INTUOSPAD || data[0] == WACOM_REPORT_INTUOS5PAD ||

I think there are many more places in the driver where it used
wacom->pad_input without verifying that it exists.

Thanks.

-- 
Dmitry