linux-serial.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/2] serial: mrst_max3110: Check the irq number before enable/disabe irq in PM hooks
@ 2013-10-09  2:39 Feng Tang
  2013-10-09  2:39 ` [PATCH 2/2] serial: mrst_max3110: Fix race condition between spi transfers Feng Tang
  0 siblings, 1 reply; 2+ messages in thread
From: Feng Tang @ 2013-10-09  2:39 UTC (permalink / raw)
  To: Greg KH, linux-serial; +Cc: Feng Tang

We should check the validity of the irq number before calling
disable_irq() and enable_irq() in the suspend/resume function,
as "max->irq == 0" means the irq is not enabled for max3110
device, otherwise it will hurt device whose irq number is really 0.

Signed-off-by: Feng Tang <feng.tang@intel.com>
---
 drivers/tty/serial/mrst_max3110.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/tty/serial/mrst_max3110.c b/drivers/tty/serial/mrst_max3110.c
index a67e708..ee77e73 100644
--- a/drivers/tty/serial/mrst_max3110.c
+++ b/drivers/tty/serial/mrst_max3110.c
@@ -749,7 +749,8 @@ static int serial_m3110_suspend(struct device *dev)
 	struct spi_device *spi = to_spi_device(dev);
 	struct uart_max3110 *max = spi_get_drvdata(spi);
 
-	disable_irq(max->irq);
+	if (max->irq > 0)
+		disable_irq(max->irq);
 	uart_suspend_port(&serial_m3110_reg, &max->port);
 	max3110_out(max, max->cur_conf | WC_SW_SHDI);
 	return 0;
@@ -762,7 +763,8 @@ static int serial_m3110_resume(struct device *dev)
 
 	max3110_out(max, max->cur_conf);
 	uart_resume_port(&serial_m3110_reg, &max->port);
-	enable_irq(max->irq);
+	if (max->irq > 0)
+		enable_irq(max->irq);
 	return 0;
 }
 
-- 
1.7.0.4


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [PATCH 2/2] serial: mrst_max3110: Fix race condition between spi transfers
  2013-10-09  2:39 [PATCH 1/2] serial: mrst_max3110: Check the irq number before enable/disabe irq in PM hooks Feng Tang
@ 2013-10-09  2:39 ` Feng Tang
  0 siblings, 0 replies; 2+ messages in thread
From: Feng Tang @ 2013-10-09  2:39 UTC (permalink / raw)
  To: Greg KH, linux-serial; +Cc: Bin Gao, Feng Tang

From: Bin Gao <bin.gao@intel.com>

There is a race between termios configuration and xmit that can cause the
intel_mid_ssp_spi driver to stall.

Serializing spi transactions fixes the problem.

Signed-off-by: Bin Gao <bin.gao@intel.com>
Signed-off-by: Feng Tang <feng.tang@intel.com>
---
 drivers/tty/serial/mrst_max3110.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/drivers/tty/serial/mrst_max3110.c b/drivers/tty/serial/mrst_max3110.c
index ee77e73..565779d 100644
--- a/drivers/tty/serial/mrst_max3110.c
+++ b/drivers/tty/serial/mrst_max3110.c
@@ -61,6 +61,7 @@ struct uart_max3110 {
 	struct task_struct *main_thread;
 	struct task_struct *read_thread;
 	struct mutex thread_mutex;
+	struct mutex io_mutex;
 
 	u32 baud;
 	u16 cur_conf;
@@ -90,6 +91,7 @@ static int max3110_write_then_read(struct uart_max3110 *max,
 	struct spi_transfer	x;
 	int ret;
 
+	mutex_lock(&max->io_mutex);
 	spi_message_init(&message);
 	memset(&x, 0, sizeof x);
 	x.len = len;
@@ -104,6 +106,7 @@ static int max3110_write_then_read(struct uart_max3110 *max,
 
 	/* Do the i/o */
 	ret = spi_sync(spi, &message);
+	mutex_unlock(&max->io_mutex);
 	return ret;
 }
 
@@ -805,6 +808,7 @@ static int serial_m3110_probe(struct spi_device *spi)
 	max->irq = (u16)spi->irq;
 
 	mutex_init(&max->thread_mutex);
+	mutex_init(&max->io_mutex);
 
 	max->word_7bits = 0;
 	max->parity = 0;
-- 
1.7.0.4


^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2013-10-09  2:47 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2013-10-09  2:39 [PATCH 1/2] serial: mrst_max3110: Check the irq number before enable/disabe irq in PM hooks Feng Tang
2013-10-09  2:39 ` [PATCH 2/2] serial: mrst_max3110: Fix race condition between spi transfers Feng Tang

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).