Re: [PATCH] HID: magicmouse: do not keep a stale msc->input if no input is claimed

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Deterministic NULL pointer dereference in event handlers after a failed input registration.
--

--- Patch [1]: [PATCH] HID: magicmouse: do not keep a stale msc->input if no input is claimed ---
commit 6fea999082f28bca408072a6788f3616b0ebbd9f
Author: Jose Villaseñor Montfort <[email protected]>

HID: magicmouse: do not keep a stale msc->input if no input is claimed

This patch clears msc->input when the HID core does not claim an input
device. This is intended to ensure existing NULL checks safely handle the
failure path where hidinput_connect() frees the input devices.

Link: https://lore.kernel.org/linux-input/[email protected]/

> diff --git a/drivers/hid/hid-magicmouse.c b/drivers/hid/hid-magicmouse.c
> index 802a3479e24b9..2f14094a63b1e 100644
> --- a/drivers/hid/hid-magicmouse.c
> +++ b/drivers/hid/hid-magicmouse.c
> @@ -900,6 +900,16 @@ static int magicmouse_probe(struct hid_device *hdev,
>  		return ret;
>  	}
>  
> +	/*
> +	 * When hidinput_connect() fails it frees every input device it
> +	 * created, but that does not fail hid_hw_start(): the core simply
> +	 * does not claim an input. msc->input, cached in ->input_mapping
> +	 * while the report descriptor was parsed, would then be a dangling
> +	 * pointer that passes every NULL check. Trust the core's claim.
> +	 */
> +	if (!(hdev->claimed & HID_CLAIMED_INPUT))
> +		msc->input = NULL;
> +

[Severity: High]
If msc->input is set to NULL here due to a failure, could this cause a NULL
pointer dereference for certain USB devices?

Looking further down in magicmouse_probe(), there is an early return on
success for specific devices that bypasses the failure check:

drivers/hid/hid-magicmouse.c:magicmouse_probe() {
    ...
    if (!(hdev->claimed & HID_CLAIMED_INPUT))
        msc->input = NULL;
    ...
    if (is_usb_magicmouse2(id->vendor, id->product) ||
        (is_usb_magictrackpad2(id->vendor, id->product) &&
         hdev->type != HID_TYPE_USBMOUSE))
        return 0;

    if (!msc->input) {
        hid_err(hdev, "magicmouse input not registered\n");
        ...
}

If the probe succeeds by returning 0, the driver remains bound.

When an input report is later received, will magicmouse_raw_event() and
magicmouse_event() unconditionally dereference the NULL msc->input pointer?

For example, looking at magicmouse_raw_event():

drivers/hid/hid-magicmouse.c:magicmouse_raw_event() {
    ...
    struct input_dev *input = msc->input;
    ...
    if (input->id.product == USB_DEVICE_ID_APPLE_MAGICMOUSE ||
    ...
}

And similarly in magicmouse_event():

drivers/hid/hid-magicmouse.c:magicmouse_event() {
    ...
    struct magicmouse_sc *msc = hid_get_drvdata(hdev);
    if ((msc->input->id.product == USB_DEVICE_ID_APPLE_MAGICMOUSE2 ||
    ...
}

>  	if (is_usb_magicmouse2(id->vendor, id->product) ||
>  	    is_usb_magictrackpad2(id->vendor, id->product)) {
>  		timer_setup(&msc->battery_timer, magicmouse_battery_timer_tick, 0);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1