Messages in this thread | | | From | Tim Hockin <> | Subject | Bug or feature: schedule + 1 task | Date | Sat, 19 Jun 1999 21:45:10 -0500 (CDT) |
| |
I'm not sure if this should be classified as a bug or as expected behavior.
If there is exactly 1 task runnable on a system (probably true so long as # tasks < # cpus), that task will be chosen to run, even if it's goodness is -1000 or lower. Seeing as it is the only task available, this makes sense, but on the other hand, by getting a -1000 (or lower) goodness, it is basically being told NOT to run.
I ran across this while trying to simplify pset (processor binding) for late 2.2 and 2.3 kernels. If I have one task, and it is bound to CPU 1, but started running on CPU 2 (pre-bind), it will be rescheduled on CPU 2, regardless of the fact that goodness() detected this and returned -1000.
Should this be fixed, or should it be left as is?
Tim
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |