Re: [PATCH 1/4] Input: fix poller start order on uninhibit

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Initial events from `dev->open()` are dropped during device uninhibit
--

commit f4fc68b8f579c449b7f9056431e51975c324c92e
Author: Dmitry Torokhov <dmitry.torokhov@gmail.com>

Input: fix poller start order on uninhibit

This commit resolves an issue where initial events generated by the poller
are dropped during a device uninhibit operation. It defers starting the
poller until after the device is fully marked as uninhibited.

> diff --git a/drivers/input/input.c b/drivers/input/input.c
> index cf6fecea79b82..d99dfe69e12d8 100644
> --- a/drivers/input/input.c
> +++ b/drivers/input/input.c
> @@ -1793,8 +1793,6 @@ static int input_uninhibit_device(struct input_dev *dev)
>  			if (error)
>  				return error;
>  		}

[Severity: High]
This isn't a bug introduced by this patch, but it appears the exact same
logic flaw exists for dev->open() which is called just above this snippet
in input_uninhibit_device():

    if (dev->open) {
        error = dev->open(dev);
        if (error)
            return error;
    }

Since dev->inhibited is still true when dev->open() is called, if a driver
generates initial synchronous events during its open routine (for example,
gpio_keys_report_state() called by gpio_keys_open()), won't those events
also be silently filtered out by input_get_disposition() returning
INPUT_IGNORE_EVENT?

Since this patch addresses this state mismatch for the poller, should the
dev->open() call also be deferred until after dev->inhibited is set to false?

> -		if (dev->poller)
> -			input_dev_poller_start(dev->poller);
>  	}
>  
>  	dev->inhibited = false;
> @@ -1802,6 +1800,9 @@ static int input_uninhibit_device(struct input_dev *dev)
>  	scoped_guard(spinlock_irq, &dev->event_lock)
>  		input_dev_toggle(dev, true);
>  
> +	if (dev->users && dev->poller)
> +		input_dev_poller_start(dev->poller);
> +
>  	return 0;
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260624055008.2494980-1-dmitry.torokhov@gmail.com?part=1