Linux Serial subsystem development
 help / color / mirror / Atom feed
From: Stefan Agner <stefan@agner.ch>
To: gregkh@linuxfoundation.org
Cc: jslaby@suse.com, bhuvanchandra.dv@toradex.com,
	linux-serial@vger.kernel.org, linux-kernel@vger.kernel.org,
	Stefan Agner <stefan@agner.ch>
Subject: [PATCH] tty: serial: fsl_lpuart: lock port on console write
Date: Fri, 24 Mar 2017 11:33:46 -0700	[thread overview]
Message-ID: <20170324183346.12178-1-stefan@agner.ch> (raw)

The console write code is not entirely race free (e.g. the operations
to disabling the UART interrupts are not atomic) hence locking is
required. This has been become apparent with the PREEMPT RT patchset
applied: With the fully preemptible kernel configuration the system
often ended up in a freeze already at startup.

Disable interrupts and lock using read_lock_irqsave. Try to lock in
the sysrq/oops case, but don't bother if locking fails.

Signed-off-by: Stefan Agner <stefan@agner.ch>
---
 drivers/tty/serial/fsl_lpuart.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/tty/serial/fsl_lpuart.c b/drivers/tty/serial/fsl_lpuart.c
index f02934ffb329..15df1ba78095 100644
--- a/drivers/tty/serial/fsl_lpuart.c
+++ b/drivers/tty/serial/fsl_lpuart.c
@@ -1705,6 +1705,13 @@ lpuart_console_write(struct console *co, const char *s, unsigned int count)
 {
 	struct lpuart_port *sport = lpuart_ports[co->index];
 	unsigned char  old_cr2, cr2;
+	unsigned long flags;
+	int locked = 1;
+
+	if (sport->port.sysrq || oops_in_progress)
+		locked = spin_trylock_irqsave(&sport->port.lock, flags);
+	else
+		spin_lock_irqsave(&sport->port.lock, flags);
 
 	/* first save CR2 and then disable interrupts */
 	cr2 = old_cr2 = readb(sport->port.membase + UARTCR2);
@@ -1719,6 +1726,9 @@ lpuart_console_write(struct console *co, const char *s, unsigned int count)
 		barrier();
 
 	writeb(old_cr2, sport->port.membase + UARTCR2);
+
+	if (locked)
+		spin_unlock_irqrestore(&sport->port.lock, flags);
 }
 
 static void
@@ -1726,6 +1736,13 @@ lpuart32_console_write(struct console *co, const char *s, unsigned int count)
 {
 	struct lpuart_port *sport = lpuart_ports[co->index];
 	unsigned long  old_cr, cr;
+	unsigned long flags;
+	int locked = 1;
+
+	if (sport->port.sysrq || oops_in_progress)
+		locked = spin_trylock_irqsave(&sport->port.lock, flags);
+	else
+		spin_lock_irqsave(&sport->port.lock, flags);
 
 	/* first save CR2 and then disable interrupts */
 	cr = old_cr = lpuart32_read(sport->port.membase + UARTCTRL);
@@ -1740,6 +1757,9 @@ lpuart32_console_write(struct console *co, const char *s, unsigned int count)
 		barrier();
 
 	lpuart32_write(old_cr, sport->port.membase + UARTCTRL);
+
+	if (locked)
+		spin_unlock_irqrestore(&sport->port.lock, flags);
 }
 
 /*
-- 
2.12.0

                 reply	other threads:[~2017-03-24 18:33 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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=20170324183346.12178-1-stefan@agner.ch \
    --to=stefan@agner.ch \
    --cc=bhuvanchandra.dv@toradex.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=jslaby@suse.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-serial@vger.kernel.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