Re: Watchdog timers & Cooperative OS'es

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View

Quoted text here. Click to load it

That can be quite hard.

In the end, it all depends on what _exactly_ you want the dog to watch
out for: complete breakdown of multi-tasking (i.e. one task hangs in a
loop and no others get any CPU time) only, or partial breakdown too,
as in: task-switching is still being done, but at least one task is
getting executed too infrequently to get its work done.

It also depends rather strongly on the exact type of scheduling
algorithm you're using, esp. on whether the scheduler tries to
guarantee a maximum interval between invocations of a given thread, at
a given priority.

And it depends on what your watchdog time-out handler will *do*, once
it's triggered.

Quoted text here. Click to load it

Putting it in the idle thread is not reliable at all --- it won't be
updated at all if the other threads are so busy that the idle thread
hardly ever gets executed at all.  So the WD might well misfire in the
exact situation where it hurts most.

Putting it in the highest-priority thread will only detect complete
break-down of task-switching.  It won't notice if some of the
medium-priority tasks are overloaded, but the top-priority one still
gets executed.

You may have to put the WD update into *every* thread instead.  Or,
since you appear to be writing your own scheduler, right into the
scheduler itself.  E.g. at each task switch for a task that has to finish
its next invocation before some known time interval, it should update
the watchdog to fire at that time if it's currently set to fire later.
--
Hans-Bernhard Broeker ( snipped-for-privacy@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Re: Watchdog timers & Cooperative OS'es
Quoted text here. Click to load it


Ralph,

Here is an article from embedded.com:

http://www.embedded.com/2000/0011/0011feat4.htm

near the end of the article it talks about the high priority vs. low
priority watchdog timer threads.

Are you planning on haveing each thread report on its health? If not
you could get in a condition where the lower priority threads are
starving but the watchdog at the high priority is still kicking the
watchdog.

David

Re: Watchdog timers & Cooperative OS'es

Quoted text here. Click to load it

Thanks for the link, a good read.

Ralph



Re: Watchdog timers & Cooperative OS'es
Quoted text here. Click to load it

I've seen the high priority vs. low priority choice develop into a holy war.
The high-priority option will cause the WDT to trigger reset when the system
goes to pot, or the scheduler stops working properly.  But it won't detect
an instance where one of the threads locks up.  The low priority option will
trigger the reset in the same instances as the high priority type will, as
well as when a thread above idle blocks processing on lower priority threads.

Both options are appealing, but often a combination of the two is the best.
Set up a flag that the idle thread can use to indicate that it has executed.
Have the high priority thread check for this flag and reset it.  If the
flag hasn't been set, the thread can access state information so you can
determine why the idle thread didn't execute.  Then, depending on your
desired system behavior, you can toggle the watchdog or not.

This idea can be extended so that each of the threads has a bit that must
be set so the watchdog thread can verify that they've executed.  By using
both a low and high priority thread for WDT handling, you gain a lot of
flexibility in the way error situations are handled.

Re: Watchdog timers & Cooperative OS'es
Quoted text here. Click to load it
reliable
priority in
task
war.
system
will
threads.
best.
executed.


Thanks Dingo,

This is a very good idea, I can have my idle thread set a global to some
count, and the high priority watchdog thread count that down, if it ever
reaches 0 then it stops kicking the watch dog.

Seems that this really is the best of both worlds. Easy for tasks to use.

I am also going to check for free ram in the watchdog, the system runs with
a nice buffer of free ram, if it drops too low it can reset (which it should
never do), meaning all my tasks don't need to bother checking for low ram
situations that shouldn't happen anyway.

Ralph



Site Timeline