> I also made the locking a bit better, so the normal operations will work > without locking. Locking is only done when then unlocked version fails.
I do not understand this. >From what i can see, the locking is only preventing two write operations to occur on "cache". But it is still possible for a read operation to occur while a write operation is in progress. This cannot be. From experience i know that it's not just unexpected results, and ConcurrentModificationExceptions, but a range of bizarre exceptions can occur that you'd never expect to have to catch. It's late and i've had a few drinks so maybe i'm misunderstanding something here...? ~mck -- "Physics is the universe's operating system." Steven R Garman | www.semb.wever.org | www.sesat.no | www.sesam.no | _______________________________________________ Kernel-development mailing list [email protected] http://sesat.no/mailman/listinfo/kernel-development
