Ticket #42 (assigned task)
Reread all the scheduler synchronization, semaphore and sleep code
| Reported by: | dave | Owned by: | dave |
|---|---|---|---|
| Priority: | critical | Milestone: | NxOS 0.2 |
| Component: | marvin | Version: | |
| Keywords: | Cc: |
Description
With the addition of scheduling decisions to the scheduler unlock function, Marvin's scheduler is that much more responsive and solid. However, during the first test run of the sleep tests under those conditions, the sleeping task locked up. I haven't been able to reproduce the phenomenon, but I suspect that I may have introduced a race by changing the unlocking semantics. A thorough examination of all the scheduler code is in order, just to be on the safe side.
Change History
Note: See
TracTickets for help on using
tickets.
