On Friday, 27 November 1998,
     DENIS TONN wrote to IVAN TODOROSKI about Poking inside the kernel

IT>> 1)
IT>>   PRIORITY=ABSOLUTE
IT>>   PMSHELL thread priorities: default
IT>>   OS/2 kernel: unmodified
IT>>
IT>>   Dragging windows around freezes other applications. It seems there is
IT>>   a boost applied, even though priority is set to ABSOLUTE.

 DT>  I doubt it. I suspect what you are seeing here is related to the
 DT> non-preemtable portions of the driver.
 DT>  I have looked at the internal priorities when PRIORITY=ABSOLUTE, and
 DT> they never change except under application control. The scheduler will
 DT> do a pure "look for the highest priority, ready to run" thread, an
 DT> dispatch it.
 DT>  Keep in mind that the PM "drawing thread" will be
involved here, and
 DT> it may have a higher priority.

  Well I tried to do my best to isolate and take into consideration all
  these effects.

  Just consider the following results:

  In all cases the priority of the PMSHELL threads was set to Regular,
  delta 0 (checked, re-checked, triple-checked), so this source is
  eliminated.

  Non-preemptivity of driver operations has no observable effect here
  because if you raise the priority of the background process to Time
  Critical it won't freeze, it WON'T EVEN SLOW DOWN, not even a little,
  when you drag other windows around. So some portions of the redrawing
  of the window surely are non-preemptive, but they're so finegrained
  that the effect is negligible. If the non-preemptivity of driver
  operations was the major reason for this freezing, then changing the
  priority of the background process would not have changed a thing, but
  it did.


  So, what did I do?

  Remember, all threads of PMSHELL.EXE are set to Regular, delta 0 in
  the following experiments.

1)  Default kernel, PRIORITY=ABSOLUTE
       Applications freeze when you drag windows around.

2)  Default kernel, PRIORITY=DYNAMIC
       Applications freeze when you drag windows around.

3)  "Flat" kernel, PRIORITY=ABSOLUTE
       Applications freeze when you drag windows around.

4)  "Flat" kernel, PRIORITY=DYNAMIC
       Applications *DON'T* freeze when you drag windows around.


    What conclusions would you draw from this?

IT>>   The conclusion is that PM drawing operations get a priority boost,
IT>>   even when PRIORITY=ABSOLUTE, and that's why they run at higher
IT>>   priority than the calling program.

 DT>  No, the PM drawing threads do NOT get "boosted". They
may be setting
 DT> themselves to a higher level (temporarily) or passing off the work to
 DT> high priority threads (more likely), but they are NOT boosted with
 DT> PRIORITY=ABSOLUTE.

  Maybe so, maybe it's not an actual "boost", but all other things being
  equal, drawing operations do have a higher priority in case 1) then in
  case 4) above. So cases 1) and 4) are not totally equivalent, which
  means that setting PRIORITY=ABSOLUTE is NOT the same as "flattening"
  the kernel priority table. How else could these 4 experiments be
  explained?

 DT>  Your experiments in this area are doing more to expose the behaviour
 DT> of PM and the graphics driver than they are in exposing the behaviour
 DT> of priority adjustment. Keep this FIRMLY in mind with the above "test
 DT> scenerio".
 DT>  If you wish to actually play with testing the priority boosting of
 DT> the kernel, you need to eliminate these variables from your
 DT> experiments.

  Well ALL things are equal, only things that are varied is the PRIORITY
  setting and the kernel priority table.

  And I did some of those tests with applications and threads you
  proposed (I was using OREXX, not C, but it souldn't matter, since
  OREXX has a RexxUtil function to set the priority), and everything
  works absolutely as expected, there seem to be no anomalies in the
  priority scheduler, which puzzles me even more!

  I would be extremely interested in your educated guess about the
  reason for case 4) above.

P.S. I checked out the IBM Developers site. Now there is one cool site!
     I downloaded something called "OS/2 Debugging Handbook" last night.
     Hope it's the right one... Gonna go read it later tonight, thanks
     for the tip!

                                                            - Ivan -

.!. As I said before, I never repeat myself.
--- Terminate 5.00/Pro [OS/2]
 þ TerMail/QWK þ