Re: [PATCH] atmel_serial: fix spinlock lockup in RS485 code

From: Nicolas Ferre
Date: Thu Dec 15 2011 - 05:00:43 EST


On 12/15/2011 09:47 AM, Claudio Scordino :
> Hi all.
>
> We have encountered a spinlock lockup in the RS485 part of the
> atmel_serial driver.
>
> Here is the trace:
>
> [ 68.030000] BUG: spinlock lockup on CPU#0, dnp_master/244, c07e82c0
> [ 68.040000] [<c002a790>] (unwind_backtrace+0x0/0xf0) from
> [<c012f8f0>] (do_raw_spin_lock+0x10c/0x148)
> [ 68.050000] [<c012f8f0>] (do_raw_spin_lock+0x10c/0x148) from
> [<c015152c>] (atmel_tasklet_func+0x10/0x708)
> [ 68.060000] [<c015152c>] (atmel_tasklet_func+0x10/0x708) from
> [<c003bfd4>] (tasklet_action+0x8c/0xf4)
> [ 68.070000] [<c003bfd4>] (tasklet_action+0x8c/0xf4) from [<c003c630>]
> (__do_softirq+0xa4/0x180)
> [ 68.080000] [<c003c630>] (__do_softirq+0xa4/0x180) from [<c003c764>]
> (irq_exit+0x58/0xb0)
> [ 68.090000] [<c003c764>] (irq_exit+0x58/0xb0) from [<c0021084>]
> (asm_do_IRQ+0x84/0xa4)
> [ 68.100000] [<c0021084>] (asm_do_IRQ+0x84/0xa4) from [<c002622c>]
> (__irq_svc+0x4c/0x98)
> [ 68.110000] Exception stack(0xc341fdc8 to 0xc341fe10)
> [ 68.110000] fdc0: c07e82c0 00000000 00000200
> c4818000 c341fe10 c07e82c0
> [ 68.120000] fde0: 00000000 c07e82c0 c38d3318 c341e000 bebcee00
> 010a0b60 c07e83fc c341fe10
> [ 68.130000] fe00: c0151d54 c023d9c0 60000013 ffffffff
> [ 68.140000] [<c002622c>] (__irq_svc+0x4c/0x98) from [<c023d9c0>]
> (_raw_spin_unlock+0x0/0x48)
> [ 68.140000] [<c023d9c0>] (_raw_spin_unlock+0x0/0x48) from
> [<00000000>] ( (null))
>
> The soft lockup occurs later
>
> [ 126.750000] BUG: soft lockup - CPU#0 stuck for 61s! [dnp_master:244]
> [ 126.750000] irq event stamp: 0
> [ 126.750000] hardirqs last enabled at (0): [< (null)>] (null)
> [ 126.750000] hardirqs last disabled at (0): [<c0035e30>]
> copy_process+0x308/0xe78
> [ 126.750000] softirqs last enabled at (0): [<c0035e30>]
> copy_process+0x308/0xe78
> [ 126.750000] softirqs last disabled at (0): [< (null)>] (null)
>
>
> The following patch (already tested by Dave, who also reported the bug)
> fixes the problem.
>
> Regards,
>
> Claudio
>
>
>
>
> Subject: atmel_serial: fix spinlock lockup in RS485 code
>
> Patch to fix a spinlock lockup in the driver that sometimes happens when
> the
> tasklet starts.
>
> Signed-off-by: Claudio Scordino <claudio@xxxxxxxxxxxxxxx>
> Signed-off-by: Dave Bender <codehero@xxxxxxxxx>
> Tested-by: Dave Bender <codehero@xxxxxxxxx>

Acked-by: Nicolas Ferre <nicolas.ferre@xxxxxxxxx>

Claudio, can you please send the patch to Alan with my "Acked-by" and an
addition to Linux stable just after the signature lines:

Cc: stable <stable@xxxxxxxxxx>

Thanks, best regards,

> ---
> drivers/tty/serial/atmel_serial.c | 5 +++--
> 1 files changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/tty/serial/atmel_serial.c
> b/drivers/tty/serial/atmel_serial.c
> index 4c823f3..90c8e3a 100644
> --- a/drivers/tty/serial/atmel_serial.c
> +++ b/drivers/tty/serial/atmel_serial.c
> @@ -212,8 +212,9 @@ void atmel_config_rs485(struct uart_port *port,
> struct serial_rs485 *rs485conf)
> {
> struct atmel_uart_port *atmel_port = to_atmel_uart_port(port);
> unsigned int mode;
> + unsigned long flags;
>
> - spin_lock(&port->lock);
> + spin_lock_irqsave(&port->lock, flags);
>
> /* Disable interrupts */
> UART_PUT_IDR(port, atmel_port->tx_done_mask);
> @@ -244,7 +245,7 @@ void atmel_config_rs485(struct uart_port *port,
> struct serial_rs485 *rs485conf)
> /* Enable interrupts */
> UART_PUT_IER(port, atmel_port->tx_done_mask);
>
> - spin_unlock(&port->lock);
> + spin_unlock_irqrestore(&port->lock, flags);
>
> }
>


--
Nicolas Ferre
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/