> 2. The Linux kernel is not reentrant and therefore a multi-processor system
> must have any kernel calls serialized and handled by one processor. Comments
> here please?

True only for 2.0.x

> 3. The Linux scheduler cannot preempt the kernel and therefore the
> scheduler will not context switch to a
> new process B  while running a process A and receiving an interrupt that
> unblocks B. Comments here please?

As with normal unix systems we don't pre-empt executing kernel space. It's
good practice for all but hard real time systems. Note the -executing-.
All kernel space is short thinks with sleeps between. Sleep time in kernel
is pre-emptable

Alan


-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/mentre/smp-faq/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]

Reply via email to