[PATCH] serial: ifx6x60: add_timer is not safe in the mrdy_assert function

[PATCH] serial: ifx6x60: add_timer is not safe in the mrdy_assert function

[Jun Chen]

This patch make use of mod_timer instead of add_timer in the mrdy_assert function.
Because the srdy interrupter can go high when we are running function mrdy_assert and mrdy_assert
can be called by multi-entry. In our medfield platform, spi stress test can encounter this
error logs triggered by the BUG_ON of add_timer function.This patch had been tested on
our medfield platform.

the scenario:
      CPU0							CPU1
mrdy_assert
set_bit(IFX_SPI_STATE_TIMER_PENDING)
								ifx_spi_handle_srdy
								...
								clear_bit(IFX_SPI_STATE_TIMER_PENDING)
								...
								mrdy_assert
								set_bit(IFX_SPI_STATE_TIMER_PENDING)
								...
								add_timer
...
add_timer

cc:liu chuansheng <chuansheng.liu@intel.com>
cc:Bi Chao <chao.bi@intel.com>
Signed-off-by: Chen Jun <jun.d.chen@intel.com>
---
 drivers/tty/serial/ifx6x60.c |    4 +---
 1 files changed, 1 insertions(+), 3 deletions(-)

diff --git a/drivers/tty/serial/ifx6x60.c b/drivers/tty/serial/ifx6x60.c
index 769396a..e56ed33 100644
--- a/drivers/tty/serial/ifx6x60.c
+++ b/drivers/tty/serial/ifx6x60.c
@@ -190,9 +190,7 @@ static void mrdy_assert(struct ifx_spi_device *ifx_dev)
 	if (!val) {
 		if (!test_and_set_bit(IFX_SPI_STATE_TIMER_PENDING,
 				      &ifx_dev->flags)) {
-			ifx_dev->spi_timer.expires =
-				jiffies + IFX_SPI_TIMEOUT_SEC*HZ;
-			add_timer(&ifx_dev->spi_timer);
+			mod_timer(&ifx_dev->spi_timer,jiffies + IFX_SPI_TIMEOUT_SEC*HZ);
 
 		}
 	}
-- 
1.7.4.1

Re: [PATCH] serial: ifx6x60: add_timer is not safe in the mrdy_assert function

[Alan Cox]

On Mon, 22 Oct 2012 10:23:07 -0400
Jun Chen <jun.d.chen@intel.com> wrote:

> 
> This patch make use of mod_timer instead of add_timer in the mrdy_assert function.
> Because the srdy interrupter can go high when we are running function mrdy_assert and mrdy_assert
> can be called by multi-entry. In our medfield platform, spi stress test can encounter this
> error logs triggered by the BUG_ON of add_timer function.This patch had been tested on
> our medfield platform.
> 
> the scenario:
>       CPU0							CPU1
> mrdy_assert
> set_bit(IFX_SPI_STATE_TIMER_PENDING)
> 								ifx_spi_handle_srdy
> 								...
> 								clear_bit(IFX_SPI_STATE_TIMER_PENDING)
> 								...
> 								mrdy_assert
> 								set_bit(IFX_SPI_STATE_TIMER_PENDING)
> 								...
> 								add_timer
> ...
> add_timer
> 
> cc:liu chuansheng <chuansheng.liu@intel.com>
> cc:Bi Chao <chao.bi@intel.com>
> Signed-off-by: Chen Jun <jun.d.chen@intel.com>

Acked-by: Alan Cox <alan@linux.intel.com>