Re: [PATCH 4/4] hwmon: (nzxt-kraken3) Stop device IO before calling hid_hw_stop

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Unconditional call to `hid_device_io_stop()` in the probe error path triggers a spurious kernel warning and improperly unwinds initialization.
--

--- Patch [4]: [PATCH 4/4] hwmon: (nzxt-kraken3) Stop device IO before calling hid_hw_stop ---
commit 1a6a1a210c8f1b5e86d4cd6021a1f5a8d033e172
Author: Guenter Roeck <linux@roeck-us.net>

hwmon: (nzxt-kraken3) Stop device IO before calling hid_hw_stop

Calling hid_hw_stop() does not stop the device IO, causing a race condition
if probe fails. This commit fixes the problem by calling
hid_device_io_stop() before calling hid_hw_stop().

> diff --git a/drivers/hwmon/nzxt-kraken3.c b/drivers/hwmon/nzxt-kraken3.c
> index d00409bcab93a..cca53252b343a 100644
> --- a/drivers/hwmon/nzxt-kraken3.c
> +++ b/drivers/hwmon/nzxt-kraken3.c
> @@ -970,6 +970,7 @@ static int kraken3_probe(struct hid_device *hdev, const struct hid_device_id *id
>  fail_and_close:
>  	hid_hw_close(hdev);
>  fail_and_stop:
> +	hid_device_io_stop(hdev);

[Severity: High]
Does this code unconditionally call hid_device_io_stop() on error paths where
hid_device_io_start() has not yet been called?

Looking at kraken3_probe() in drivers/hwmon/nzxt-kraken3.c, if we hit the
default case in the product switch statement or if memory allocation for
priv->buffer fails, we jump to fail_and_close:

        default:
                ret = -ENODEV;
                goto fail_and_close;

At this point, hid_device_io_start() hasn't executed.

Because hid_device_io_stop() checks if io_started is false and emits a
dev_warn() in that case, this will trigger a spurious kernel warning on
these benign error paths. On systems with panic_on_warn enabled, this could
lead to a panic just by plugging in an unsupported device.

Additionally, when failing after hid_device_io_start() has been called
(for example, if kraken3_init_device() fails), the error handling jumps to
fail_and_close. This means hid_hw_close() is called before
hid_device_io_stop().

Should a new fail_and_stop_io label be added to ensure the teardown
happens in the correct LIFO order and to avoid warnings on early errors?

>  	hid_hw_stop(hdev);
>  	return ret;
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260708015806.213790-1-linux@roeck-us.net?part=4