_Thread_Dispatch_disable_level gone negative.
Leon Pollak
leonp at plris.com
Thu Apr 29 10:23:34 UTC 2010
On Thursday April 29 2010, Nick Thomas wrote:
> Hi,
>
> I have had problems in the past with _ISR_Nest_Level and
> _Thread_Dispatch_disable_level.
> These sometimes get stuck at 1, so RTEMS thinks it is in a ISR, which
> prevents any tasks from running.
>
> Still trying to track down the cause of this issue, but it looks like
> something is going wrong in one of the ISR routines, either it's not using
> a conventional return path or something.
>
> But, some new info now: I have just seen the same problem occur, but
> _Thread_Dispatch_disable_level is 0xffffffff , i.e. -1;
> This is having the same effect, that no tasks get to run.
>
> But, what could possibly cause this RTEMS variable to go negative.
>
>
>
> I am using RTEMS 4.7.1 on a PowerPC. I know it's an older version of RTEMS,
> but it's too late in the project to change that now.
>
>
> Any comments/suggestions are appreciated regarding this issue.
Nick,
I think I had the similar problem, when 0xffffffff was the value of comtrol
variable - do not remember exactly its name, but it was something in
_Watchdog_.....
This was the problem in old version of RTEMS which is solved in upper
versions.
Joel pointed me to the corresponding PR number, but I do not remember it
too...:-)
Anyway, you need to upgrade...:-((
This is what I finally did, although it cost me a lot of efforts...
--
Leon
More information about the users
mailing list