Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.32 as permitted sender) client-ip=23.128.96.32;
Date:   Tue, 24 Oct 2023 14:48:53 +0200
From:   Johan Hovold <johan@kernel.org>
To:     Tony Lindgren <tony@atomide.com>
Cc:     Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jiri Slaby <jirislaby@kernel.org>,
        Rob Herring <robh@kernel.org>,
        Andy Shevchenko <andriy.shevchenko@intel.com>,
        Dhruva Gole <d-gole@ti.com>,
        Ilpo =?utf-8?B?SsOkcnZpbmVu?= <ilpo.jarvinen@linux.intel.com>,
        John Ogness <john.ogness@linutronix.de>,
        Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
        Vignesh Raghavendra <vigneshr@ti.com>,
        linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org,
        Maximilian Luz <luzmaximilian@gmail.com>
Subject: Re: [RFC PATCH 1/2] serial: core: Move tty and serdev to be children
 of serial core port device
Message-ID: <ZTe9NdS160EU1RJJ@hovoldconsulting.com>
References: <20231024113624.54364-1-tony@atomide.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20231024113624.54364-1-tony@atomide.com>
Precedence: bulk

On Tue, Oct 24, 2023 at 02:36:18PM +0300, Tony Lindgren wrote:
> Let's move tty and serdev controller to be children of the serial core port
> device. This way the runtime PM usage count of a child device propagates
> to the serial hardware device.

> The tty device moves happily with just a change of the parent device.

> @@ -3153,7 +3153,8 @@ static int serial_core_add_one_port(struct uart_driver *drv, struct uart_port *u
>  	 * setserial to be used to alter this port's parameters.
>  	 */
>  	tty_dev = tty_port_register_device_attr_serdev(port, drv->tty_driver,
> -			uport->line, uport->dev, port, uport->tty_groups);
> +			uport->line, uport->dev, &uport->port_dev->dev, port,
> +			uport->tty_groups);
>  	if (!IS_ERR(tty_dev)) {
>  		device_set_wakeup_capable(tty_dev, 1);
>  	} else {

>  struct device *tty_port_register_device_attr_serdev(struct tty_port *port,
>                 struct tty_driver *driver, unsigned index,
> -               struct device *device, void *drvdata,
> +               struct device *host, struct device *parent, void *drvdata,
>                 const struct attribute_group **attr_grp)
>  {
>         struct device *dev;
> 
>         tty_port_link_device(port, driver, index);
> 
> -       dev = serdev_tty_port_register(port, device, driver, index);
> +       dev = serdev_tty_port_register(port, host, parent, driver, index);
>         if (PTR_ERR(dev) != -ENODEV) {
>                 /* Skip creating cdev if we registered a serdev device */
>                 return dev;
>         }
> 
> -       return tty_register_device_attr(driver, index, device, drvdata,
> +       return tty_register_device_attr(driver, index, parent, drvdata,
>                         attr_grp);
>  }

Looks like this patch breaks the wakeup-irq hack in uart_suspend_port():

	tty_dev = device_find_child(uport->dev, &match, serial_match_port);
	if (tty_dev && device_may_wakeup(tty_dev)) {
		enable_irq_wake(uport->irq);
		put_device(tty_dev);
		mutex_unlock(&port->mutex);
		return 0;
	}

There may be more of these hard-coded assumptions, this one I happened
to be aware of.

Johan