linux-kernel - Re: [PATCH 7/7] serial: 8250_port: Remove calls to runtime PM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <YaYBNIGwOkHts6wn@hovoldconsulting.com>
Date:   Tue, 30 Nov 2021 11:47:16 +0100
From:   Johan Hovold <johan@...nel.org>
To:     Tony Lindgren <tony@...mide.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Andy Shevchenko <andriy.shevchenko@...el.com>,
        Jiri Slaby <jirislaby@...nel.org>,
        Sebastian Andrzej Siewior <bigeasy@...utronix.de>,
        Vignesh Raghavendra <vigneshr@...com>,
        linux-serial@...r.kernel.org, linux-omap@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
Subject: Re: [PATCH 7/7] serial: 8250_port: Remove calls to runtime PM

On Mon, Nov 15, 2021 at 10:42:03AM +0200, Tony Lindgren wrote:
> From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
> 
> Since we now have runtime PM calls in serial_core.c the individual
> drivers do not need them anymore for the struct uart_ops related
> functions.
> 
> Remove runtime PM calls in 8250 driver. This still leaves the flag for
> UART_CAP_RPM for serial8250_rpm_get_tx(), serial8250_rpm_put_tx() and
> serial8250_wakeup() to manage the reference count for serial TX.
> 
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
> [tony@...mide.com: updated to remove the exported functions too]
> Signed-off-by: Tony Lindgren <tony@...mide.com>

> @@ -1477,8 +1459,11 @@ static enum hrtimer_restart serial8250_em485_handle_stop_tx(struct hrtimer *t)
>  			stop_tx_timer);
>  	struct uart_8250_port *p = em485->port;
>  	unsigned long flags;
> +	int err;
>  
> -	serial8250_rpm_get(p);
> +	err = pm_runtime_resume_and_get(p->port.dev);

This also won't work as this function is called on timer expiry; you
cannot do a synchronous resume here.

> +	if (err < 0)
> +		goto out_rpm_err;
>  	spin_lock_irqsave(&p->port.lock, flags);
>  	if (em485->active_timer == &em485->stop_tx_timer) {
>  		p->rs485_stop_tx(p);
> @@ -1486,8 +1471,9 @@ static enum hrtimer_restart serial8250_em485_handle_stop_tx(struct hrtimer *t)
>  		em485->tx_stopped = true;
>  	}
>  	spin_unlock_irqrestore(&p->port.lock, flags);
> -	serial8250_rpm_put(p);
> -
> +	pm_runtime_mark_last_busy(p->port.dev);
> +	pm_runtime_put_autosuspend(p->port.dev);
> +out_rpm_err:
>  	return HRTIMER_NORESTART;
>  }
  
>  void serial8250_read_char(struct uart_8250_port *up, unsigned char lsr)
> @@ -1984,15 +1966,11 @@ static unsigned int serial8250_tx_empty(struct uart_port *port)
>  	unsigned long flags;
>  	unsigned int lsr;
>  
> -	serial8250_rpm_get(up);
> -
>  	spin_lock_irqsave(&port->lock, flags);
>  	lsr = serial_port_in(port, UART_LSR);
>  	up->lsr_saved_flags |= lsr & LSR_SAVE_FLAGS;
>  	spin_unlock_irqrestore(&port->lock, flags);
>  
> -	serial8250_rpm_put(up);
> -
>  	return (lsr & BOTH_EMPTY) == BOTH_EMPTY ? TIOCSER_TEMT : 0;
>  }

As I mentioned elsewhere, the corresponding get+put is missing in serial
core now.

>  static void serial8250_put_poll_char(struct uart_port *port,
>  			 unsigned char c)
>  {
>  	unsigned int ier;
>  	struct uart_8250_port *up = up_to_u8250p(port);
>  
> -	serial8250_rpm_get(up);
>  	/*
>  	 *	First save the IER then disable the interrupts
>  	 */
> @@ -2155,7 +2117,6 @@ static void serial8250_put_poll_char(struct uart_port *port,
>  	 */
>  	wait_for_xmitr(up, BOTH_EMPTY);
>  	serial_port_out(port, UART_IER, ier);
> -	serial8250_rpm_put(up);
>  }

And this is a console callback; where is it guaranteed that the console
is never suspended?

Johan