The kernel lock validator
Posted May 31, 2006 6:54 UTC (Wed) by smurf
In reply to: The kernel lock validator
Parent article: The kernel lock validator
Umm, the validator doesn't detect all *possible* deadlock conditions -- just those that, given current and past locking history, might occur. It won't find tomorrow's deadlock if the code that could possibly cause it didn't run today.
Oh, and the scheduler doesn't run in constant time. It just delegates the job to small trivial chunks that run when each process starts+stops using the CPU, so technically it's still O(n), not O(1).
to post comments)