All of lore.kernel.org
 help / color / mirror / Atom feed
From: Erwan Le Ray <erwan.leray@foss.st.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Jiri Slaby <jslaby@suse.com>,
	Maxime Coquelin <mcoquelin.stm32@gmail.com>,
	Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: <linux-serial@vger.kernel.org>,
	<linux-stm32@st-md-mailman.stormreply.com>,
	<linux-arm-kernel@lists.infradead.org>,
	<linux-kernel@vger.kernel.org>,
	Erwan Le Ray <erwan.leray@foss.st.com>,
	Fabrice Gasnier <fabrice.gasnier@foss.st.com>,
	Valentin Caron <valentin.caron@foss.st.com>
Subject: [PATCH 05/13] serial: stm32: fix a deadlock condition with wakeup event
Date: Fri, 19 Feb 2021 18:47:28 +0100	[thread overview]
Message-ID: <20210219174736.1022-6-erwan.leray@foss.st.com> (raw)
In-Reply-To: <20210219174736.1022-1-erwan.leray@foss.st.com>

Deadlock issue is seen when enabling CONFIG_PROVE_LOCKING=Y, and uart
console as wakeup source. Deadlock occurs when resuming from low power
mode if system is waked up via usart console.
The deadlock is triggered 100% when also disabling console suspend prior
to go to suspend.

Simplified call stack, deadlock condition:
- stm32_console_write <-- spin_lock already held
- print_circular_bug
- pm_wakeup_dev_event <-- triggers lockdep as seen above
- stm32_receive_chars
- stm32_interrupt <-- wakeup via uart console, takes the lock

So, revisit spin_lock in stm32-usart driver:
- there is no need to hold the lock to access ICR (atomic clear of status
  flags)
- only hold the lock inside stm32_receive_chars() routine (no need to
  call pm_wakeup_dev_event with lock held)
- keep stm32_transmit_chars() routine called with lock held

Fixes: 48a6092fb41f ("serial: stm32-usart: Add STM32 USART Driver")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>

diff --git a/drivers/tty/serial/stm32-usart.c b/drivers/tty/serial/stm32-usart.c
index d409a23806b1..524c55751023 100644
--- a/drivers/tty/serial/stm32-usart.c
+++ b/drivers/tty/serial/stm32-usart.c
@@ -214,13 +214,18 @@ static void stm32_usart_receive_chars(struct uart_port *port, bool threaded)
 	struct tty_port *tport = &port->state->port;
 	struct stm32_port *stm32_port = to_stm32_port(port);
 	struct stm32_usart_offsets *ofs = &stm32_port->info->ofs;
-	unsigned long c;
+	unsigned long c, flags;
 	u32 sr;
 	char flag;
 
 	if (irqd_is_wakeup_set(irq_get_irq_data(port->irq)))
 		pm_wakeup_event(tport->tty->dev, 0);
 
+	if (threaded)
+		spin_lock_irqsave(&port->lock, flags);
+	else
+		spin_lock(&port->lock);
+
 	while (stm32_usart_pending_rx(port, &sr, &stm32_port->last_res,
 				      threaded)) {
 		sr |= USART_SR_DUMMY_RX;
@@ -276,9 +281,12 @@ static void stm32_usart_receive_chars(struct uart_port *port, bool threaded)
 		uart_insert_char(port, sr, USART_SR_ORE, c, flag);
 	}
 
-	spin_unlock(&port->lock);
+	if (threaded)
+		spin_unlock_irqrestore(&port->lock, flags);
+	else
+		spin_unlock(&port->lock);
+
 	tty_flip_buffer_push(tport);
-	spin_lock(&port->lock);
 }
 
 static void stm32_usart_tx_dma_complete(void *arg)
@@ -459,8 +467,6 @@ static irqreturn_t stm32_usart_interrupt(int irq, void *ptr)
 	struct stm32_usart_offsets *ofs = &stm32_port->info->ofs;
 	u32 sr;
 
-	spin_lock(&port->lock);
-
 	sr = readl_relaxed(port->membase + ofs->isr);
 
 	if ((sr & USART_SR_RTOF) && ofs->icr != UNDEF_REG)
@@ -474,10 +480,11 @@ static irqreturn_t stm32_usart_interrupt(int irq, void *ptr)
 	if ((sr & USART_SR_RXNE) && !(stm32_port->rx_ch))
 		stm32_usart_receive_chars(port, false);
 
-	if ((sr & USART_SR_TXE) && !(stm32_port->tx_ch))
+	if ((sr & USART_SR_TXE) && !(stm32_port->tx_ch)) {
+		spin_lock(&port->lock);
 		stm32_usart_transmit_chars(port);
-
-	spin_unlock(&port->lock);
+		spin_unlock(&port->lock);
+	}
 
 	if (stm32_port->rx_ch)
 		return IRQ_WAKE_THREAD;
@@ -490,13 +497,9 @@ static irqreturn_t stm32_usart_threaded_interrupt(int irq, void *ptr)
 	struct uart_port *port = ptr;
 	struct stm32_port *stm32_port = to_stm32_port(port);
 
-	spin_lock(&port->lock);
-
 	if (stm32_port->rx_ch)
 		stm32_usart_receive_chars(port, true);
 
-	spin_unlock(&port->lock);
-
 	return IRQ_HANDLED;
 }
 
-- 
2.17.1


WARNING: multiple messages have this Message-ID (diff)
From: Erwan Le Ray <erwan.leray@foss.st.com>
To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Jiri Slaby <jslaby@suse.com>,
	Maxime Coquelin <mcoquelin.stm32@gmail.com>,
	Alexandre Torgue <alexandre.torgue@foss.st.com>
Cc: Valentin Caron <valentin.caron@foss.st.com>,
	Erwan Le Ray <erwan.leray@foss.st.com>,
	linux-kernel@vger.kernel.org,
	Fabrice Gasnier <fabrice.gasnier@foss.st.com>,
	linux-serial@vger.kernel.org,
	linux-stm32@st-md-mailman.stormreply.com,
	linux-arm-kernel@lists.infradead.org
Subject: [PATCH 05/13] serial: stm32: fix a deadlock condition with wakeup event
Date: Fri, 19 Feb 2021 18:47:28 +0100	[thread overview]
Message-ID: <20210219174736.1022-6-erwan.leray@foss.st.com> (raw)
In-Reply-To: <20210219174736.1022-1-erwan.leray@foss.st.com>

Deadlock issue is seen when enabling CONFIG_PROVE_LOCKING=Y, and uart
console as wakeup source. Deadlock occurs when resuming from low power
mode if system is waked up via usart console.
The deadlock is triggered 100% when also disabling console suspend prior
to go to suspend.

Simplified call stack, deadlock condition:
- stm32_console_write <-- spin_lock already held
- print_circular_bug
- pm_wakeup_dev_event <-- triggers lockdep as seen above
- stm32_receive_chars
- stm32_interrupt <-- wakeup via uart console, takes the lock

So, revisit spin_lock in stm32-usart driver:
- there is no need to hold the lock to access ICR (atomic clear of status
  flags)
- only hold the lock inside stm32_receive_chars() routine (no need to
  call pm_wakeup_dev_event with lock held)
- keep stm32_transmit_chars() routine called with lock held

Fixes: 48a6092fb41f ("serial: stm32-usart: Add STM32 USART Driver")
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>

diff --git a/drivers/tty/serial/stm32-usart.c b/drivers/tty/serial/stm32-usart.c
index d409a23806b1..524c55751023 100644
--- a/drivers/tty/serial/stm32-usart.c
+++ b/drivers/tty/serial/stm32-usart.c
@@ -214,13 +214,18 @@ static void stm32_usart_receive_chars(struct uart_port *port, bool threaded)
 	struct tty_port *tport = &port->state->port;
 	struct stm32_port *stm32_port = to_stm32_port(port);
 	struct stm32_usart_offsets *ofs = &stm32_port->info->ofs;
-	unsigned long c;
+	unsigned long c, flags;
 	u32 sr;
 	char flag;
 
 	if (irqd_is_wakeup_set(irq_get_irq_data(port->irq)))
 		pm_wakeup_event(tport->tty->dev, 0);
 
+	if (threaded)
+		spin_lock_irqsave(&port->lock, flags);
+	else
+		spin_lock(&port->lock);
+
 	while (stm32_usart_pending_rx(port, &sr, &stm32_port->last_res,
 				      threaded)) {
 		sr |= USART_SR_DUMMY_RX;
@@ -276,9 +281,12 @@ static void stm32_usart_receive_chars(struct uart_port *port, bool threaded)
 		uart_insert_char(port, sr, USART_SR_ORE, c, flag);
 	}
 
-	spin_unlock(&port->lock);
+	if (threaded)
+		spin_unlock_irqrestore(&port->lock, flags);
+	else
+		spin_unlock(&port->lock);
+
 	tty_flip_buffer_push(tport);
-	spin_lock(&port->lock);
 }
 
 static void stm32_usart_tx_dma_complete(void *arg)
@@ -459,8 +467,6 @@ static irqreturn_t stm32_usart_interrupt(int irq, void *ptr)
 	struct stm32_usart_offsets *ofs = &stm32_port->info->ofs;
 	u32 sr;
 
-	spin_lock(&port->lock);
-
 	sr = readl_relaxed(port->membase + ofs->isr);
 
 	if ((sr & USART_SR_RTOF) && ofs->icr != UNDEF_REG)
@@ -474,10 +480,11 @@ static irqreturn_t stm32_usart_interrupt(int irq, void *ptr)
 	if ((sr & USART_SR_RXNE) && !(stm32_port->rx_ch))
 		stm32_usart_receive_chars(port, false);
 
-	if ((sr & USART_SR_TXE) && !(stm32_port->tx_ch))
+	if ((sr & USART_SR_TXE) && !(stm32_port->tx_ch)) {
+		spin_lock(&port->lock);
 		stm32_usart_transmit_chars(port);
-
-	spin_unlock(&port->lock);
+		spin_unlock(&port->lock);
+	}
 
 	if (stm32_port->rx_ch)
 		return IRQ_WAKE_THREAD;
@@ -490,13 +497,9 @@ static irqreturn_t stm32_usart_threaded_interrupt(int irq, void *ptr)
 	struct uart_port *port = ptr;
 	struct stm32_port *stm32_port = to_stm32_port(port);
 
-	spin_lock(&port->lock);
-
 	if (stm32_port->rx_ch)
 		stm32_usart_receive_chars(port, true);
 
-	spin_unlock(&port->lock);
-
 	return IRQ_HANDLED;
 }
 
-- 
2.17.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

  parent reply	other threads:[~2021-02-19 17:49 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-19 17:47 [PATCH 00/13] stm32 usart various fixes Erwan Le Ray
2021-02-19 17:47 ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 01/13] serial: stm32: fix probe and remove order for dma Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 02/13] serial: stm32: fix startup by enabling usart for reception Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 03/13] serial: stm32: fix incorrect characters on console Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 04/13] serial: stm32: fix TX and RX FIFO thresholds Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` Erwan Le Ray [this message]
2021-02-19 17:47   ` [PATCH 05/13] serial: stm32: fix a deadlock condition with wakeup event Erwan Le Ray
2021-02-19 17:47 ` [PATCH 06/13] serial: stm32: fix wake-up flag handling Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 07/13] serial: stm32: fix a deadlock in set_termios Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 08/13] serial: stm32: fix tx dma completion, release channel Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 09/13] serial: stm32: call stm32_transmit_chars locked Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 10/13] serial: stm32: fix FIFO flush in startup and set_termios Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 11/13] serial: stm32: add FIFO flush when port is closed Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 12/13] serial: stm32: fix tx_empty condition Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-02-19 17:47 ` [PATCH 13/13] serial: stm32: add support for "flush_buffer" ops Erwan Le Ray
2021-02-19 17:47   ` Erwan Le Ray
2021-03-03 19:31 ` [PATCH 00/13] stm32 usart various fixes Greg Kroah-Hartman
2021-03-03 19:31   ` Greg Kroah-Hartman

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=20210219174736.1022-6-erwan.leray@foss.st.com \
    --to=erwan.leray@foss.st.com \
    --cc=alexandre.torgue@foss.st.com \
    --cc=fabrice.gasnier@foss.st.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=jslaby@suse.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-serial@vger.kernel.org \
    --cc=linux-stm32@st-md-mailman.stormreply.com \
    --cc=mcoquelin.stm32@gmail.com \
    --cc=valentin.caron@foss.st.com \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.