Problem with clock setting

Joel Sherrill joel.sherrill at
Mon Jan 29 15:18:28 UTC 2007

Lars Törnkvist wrote:
>> -----Original Message-----
>> From: Joel Sherrill [mailto:joel.sherrill at] 
>> Do you have a test case I could play with?  It sounds like a 
>> possible bug and I want to look at it in more detail.
>> --joel
> We don't have a test case, just the complete application, so I will have to fix one. This can take some time. It would help if you tell me what you want to have. No hardware dependencies I suspect but what else do you, and don't you, want to have?
The smallest thing you can produce that will duplicate the problem.  :)

What time does RTEMS think it is? 
What time are you changing it to?
How many timers are active and how many fire during the time change?

"wake when" timers that are scheduled between old and new time will be 
executed when you
move the time forward so there is potential for a pause there.

I can conceive of a situation where you move the time forward 30 seconds 
and the timer
schedules itself 1 second ahead of its previous time (which is still now 
technically in the past).
It will fire again and again until it trickles itself into the present time.

If you can capture the state of the Watchdog chains when the time is 
changed, it should be enough to just
get a set of tasks and timers hacked together to create the same situation.
> Regards
> Lasse T.

More information about the users mailing list