On Fri, Jul 26, 2019 at 05:14:36PM +0800, Jia-Ju Bai wrote:
> In wm97xx_ts_input_open(), there is an if statement on line 507 to check
> whether wm->mach_ops is NULL:
> if (wm->mach_ops && wm->mach_ops->acc_enabled)
>
> When wm->mach_ops is NULL, it is used on line 521:
> wm97xx_init_pen_irq(wm);
> BUG_ON(!wm->mach_ops->irq_enable);
> BUG_ON(!wm->mach_ops->irq_gpio);
> wm97xx_reg_write(..., reg & ~(wm->mach_ops->irq_gpio))
>
> Thus, possible null-pointer dereferences may occur.
>
> To fix these bugs, wm->mach_ops is checked at the beginning of
> wm97xx_init_pen_irq().
>
> These bugs found by a static analysis tool STCheck written by us.
>
> Signed-off-by: Jia-Ju Bai <baijiaju1990@xxxxxxxxx>
> ---
> v2:
> * Add a new check of wm->mach_ops in wm97xx_init_pen_irq().
> Thank Charles for helpful advice.
>
> ---
> drivers/input/touchscreen/wm97xx-core.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/input/touchscreen/wm97xx-core.c b/drivers/input/touchscreen/wm97xx-core.c
> index 0a174bd82915..50b016abf492 100644
> --- a/drivers/input/touchscreen/wm97xx-core.c
> +++ b/drivers/input/touchscreen/wm97xx-core.c
> @@ -374,6 +374,9 @@ static int wm97xx_init_pen_irq(struct wm97xx *wm)
> {
> u16 reg;
>
> + if (!wm->mach_ops)
> + return -EINVAL;
> +
Probably worth adding an printk in here too, the calling function
doesn't check the return value of this function so otherwise
there will be no indication this failed.
Thanks,
Charles
> /* If an interrupt is supplied an IRQ enable operation must also be
> * provided. */
> BUG_ON(!wm->mach_ops->irq_enable);
> --
> 2.17.0
>
