| |
I am reworking parts of the task scheduler in tm.c, tm.h, et al.
I notice, reading the code, that there is an idle process which runs all
of the time. There is a note in tm.c; "execi needs pd_idle". It
appears to use it to locate the head of the process list.
If that is the extent of the need for the idle process, I don't think
that the mechanism that I have in mind is going to miss it if I remove
it. I am pretty much tearing up the linked list stuff and redoing it.
All of the other stuff (stack manipulation, _tm_switcher, the process
state model, yield(), exit(), etc. is remaining the same.
What I hope to gain by doing this is better performance when several
processes (tasks, whatever) have the same priority.
Does anybody see anything that I might be overlooking? Does anything
else depend on the idle process? All it does is an endless loop of
sleeps.
--
No honest business is promoted by spam
with the possible exception of Hormel.
|
|
Message has 1 Reply:
 | | Re: Idle process
|
| Hi Lou, the sleep instruction effectively shuts down the CPU until the next interrupt occurs. Thus, the idle process conserves battery power if no task wishes to run. I consider this a worthwile feature. If you can do this without a dedicated idle (...) (26 years ago, 8-Mar-99, to lugnet.robotics.rcx.legos)
|
10 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
