I have a short question about locking/preemption. On a Uni-proc, I understand we have to implement locking for global data structures, when kernel preemption is enabled (as in 2.6) and its not neccessary when kernel preemption is
disabled.And in SMP locking always has to be implemented.
But what happens in the case of hyper-threaded (as intel p4-HT) machines ? Does the scheduler provide the behavior of a UP or SMP (assuming I have CONFIG_SMP enabled)?
rgds
Arun
--
"There are 10 people in the world - those who understand binary and those who dont !"
