mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: John Ogness <john.ogness@linutronix.de>
To: Yunhui Cui <cuiyunhui@bytedance.com>,
	arnd@arndb.de, andriy.shevchenko@linux.intel.com,
	benjamin.larsson@genexis.eu, cuiyunhui@bytedance.com,
	gregkh@linuxfoundation.org, heikki.krogerus@linux.intel.com,
	ilpo.jarvinen@linux.intel.com, jirislaby@kernel.org,
	jkeeping@inmusicbrands.com, linux-kernel@vger.kernel.org,
	linux-serial@vger.kernel.org, markus.mayer@linaro.org,
	matt.porter@linaro.org, namcao@linutronix.de, paulmck@kernel.org,
	pmladek@suse.com, schnelle@linux.ibm.com,
	sunilvl@ventanamicro.com, tim.kryger@linaro.org
Subject: Re: [PATCH v6 2/4] serial: 8250: avoid potential PSLVERR issue
Date: Wed, 21 May 2025 10:03:43 +0206	[thread overview]
Message-ID: <84y0uq9z0o.fsf@jogness.linutronix.de> (raw)
In-Reply-To: <20250513024212.74658-2-cuiyunhui@bytedance.com>

On 2025-05-13, Yunhui Cui <cuiyunhui@bytedance.com> wrote:
> diff --git a/drivers/tty/serial/8250/8250_port.c b/drivers/tty/serial/8250/8250_port.c
> index 07fe818dffa34..9a04f24b0c762 100644
> --- a/drivers/tty/serial/8250/8250_port.c
> +++ b/drivers/tty/serial/8250/8250_port.c
> @@ -2133,25 +2132,22 @@ static void wait_for_xmitr(struct uart_8250_port *up, int bits)
>  static int serial8250_get_poll_char(struct uart_port *port)
>  {
>  	struct uart_8250_port *up = up_to_u8250p(port);
> -	int status;
> +	int status = NO_POLL_CHAR;
>  	u16 lsr;
> +	unsigned long flags;
>  
>  	serial8250_rpm_get(up);
>  
> +	uart_port_lock_irqsave(port, &flags);
>  	lsr = serial_port_in(port, UART_LSR);
> +	if (lsr & UART_LSR_DR)
> +		status = serial_port_in(port, UART_RX);
> +	uart_port_unlock_irqrestore(port, flags);

I realize I previously made a comment saying it was OK to add the spin
locking here. But I have changed my mind. Please remove this spin
locking. It is not necessary because with kgdb all the other CPUs are
quiesced, so there is no need to synchronize with the console. Also, it
will deadlock if kgdb took over while the port was locked.

> @@ -2513,7 +2514,6 @@ void serial8250_do_shutdown(struct uart_port *port)
>  		port->mctrl &= ~TIOCM_OUT2;
>  
>  	serial8250_set_mctrl(port, port->mctrl);
> -	uart_port_unlock_irqrestore(port, flags);
>  
>  	/*
>  	 * Disable break condition and FIFOs
> @@ -2521,6 +2521,14 @@ void serial8250_do_shutdown(struct uart_port *port)
>  	serial_port_out(port, UART_LCR,
>  			serial_port_in(port, UART_LCR) & ~UART_LCR_SBC);
>  	serial8250_clear_fifos(up);
> +	/*
> +	 * Read data port to reset things, and then unlink from
> +	 * the IRQ chain.
> +	 * Since reading UART_RX clears interrupts, doing so with
> +	 * FIFO disabled won't trigger PSLVERR.
> +	 */
> +	serial_port_in(port, UART_RX);
> +	uart_port_unlock_irqrestore(port, flags);
>  
>  #ifdef CONFIG_SERIAL_8250_RSA
>  	/*
> @@ -2529,11 +2537,6 @@ void serial8250_do_shutdown(struct uart_port *port)
>  	disable_rsa(up);
>  #endif
>  
> -	/*
> -	 * Read data port to reset things, and then unlink from
> -	 * the IRQ chain.
> -	 */
> -	serial_port_in(port, UART_RX);

I am thinking you should keep the read here and instead move the unlock
below the read. This would mean the lock/unlock in disable_rsa() need to
be removed. (The function comments for disable_rsa() aready say that the
caller needs to hold the port lock.)

I am thinking something like the below (untested) diff instead of the
above 2 hunks.

John Ogness

diff --git a/drivers/tty/serial/8250/8250_port.c b/drivers/tty/serial/8250/8250_port.c
index 22b3f7a193070..51467383aaf5a 100644
--- a/drivers/tty/serial/8250/8250_port.c
+++ b/drivers/tty/serial/8250/8250_port.c
@@ -781,8 +781,6 @@ static void disable_rsa(struct uart_8250_port *up)
 
 	if (up->port.type == PORT_RSA &&
 	    up->port.uartclk == SERIAL_RSA_BAUD_BASE * 16) {
-		uart_port_lock_irq(&up->port);
-
 		mode = serial_in(up, UART_RSA_MSR);
 		result = !(mode & UART_RSA_MSR_FIFO);
 
@@ -794,7 +792,6 @@ static void disable_rsa(struct uart_8250_port *up)
 
 		if (result)
 			up->port.uartclk = SERIAL_RSA_BAUD_BASE_LO * 16;
-		uart_port_unlock_irq(&up->port);
 	}
 }
 #endif /* CONFIG_SERIAL_8250_RSA */
@@ -2536,7 +2533,6 @@ void serial8250_do_shutdown(struct uart_port *port)
 		port->mctrl &= ~TIOCM_OUT2;
 
 	serial8250_set_mctrl(port, port->mctrl);
-	uart_port_unlock_irqrestore(port, flags);
 
 	/*
 	 * Disable break condition and FIFOs
@@ -2555,8 +2551,12 @@ void serial8250_do_shutdown(struct uart_port *port)
 	/*
 	 * Read data port to reset things, and then unlink from
 	 * the IRQ chain.
+	 *
+	 * Since reading UART_RX clears interrupts, doing so with
+	 * FIFO disabled won't trigger PSLVERR.
 	 */
 	serial_port_in(port, UART_RX);
+	uart_port_unlock_irqrestore(port, flags);
 	serial8250_rpm_put(up);
 
 	up->ops->release_irq(up);

  reply	other threads:[~2025-05-21  7:57 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-05-13  2:42 [PATCH v6 1/4] serial: 8250: fix panic due to PSLVERR Yunhui Cui
2025-05-13  2:42 ` [PATCH v6 2/4] serial: 8250: avoid potential PSLVERR issue Yunhui Cui
2025-05-21  7:57   ` John Ogness [this message]
2025-05-28  3:29     ` [External] " yunhui cui
2025-05-13  2:42 ` [PATCH v6 3/4] serial: 8250_dw: assert port->lock is held in dw8250_force_idle() Yunhui Cui
2025-05-13  2:42 ` [PATCH v6 4/4] serial: 8250_dw: fix PSLVERR on RX_TIMEOUT Yunhui Cui
2025-05-21  6:21 ` [PATCH v6 1/4] serial: 8250: fix panic due to PSLVERR yunhui cui

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=84y0uq9z0o.fsf@jogness.linutronix.de \
    --to=john.ogness@linutronix.de \
    --cc=andriy.shevchenko@linux.intel.com \
    --cc=arnd@arndb.de \
    --cc=benjamin.larsson@genexis.eu \
    --cc=cuiyunhui@bytedance.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=heikki.krogerus@linux.intel.com \
    --cc=ilpo.jarvinen@linux.intel.com \
    --cc=jirislaby@kernel.org \
    --cc=jkeeping@inmusicbrands.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-serial@vger.kernel.org \
    --cc=markus.mayer@linaro.org \
    --cc=matt.porter@linaro.org \
    --cc=namcao@linutronix.de \
    --cc=paulmck@kernel.org \
    --cc=pmladek@suse.com \
    --cc=schnelle@linux.ibm.com \
    --cc=sunilvl@ventanamicro.com \
    --cc=tim.kryger@linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®