Andi Kleen <ak@xxxxxx> wrote:On Thursday 09 February 2006 19:04, Andrew Morton wrote:Ashok Raj <ashok.raj@xxxxxxxxx> wrote:What is this performance penalty exactly?The problem was with ACPI just simply looking at the namespace doesntWe need to fix this asap - the performance penalty for HOTPLUG_CPU=y,
exactly give us an idea of how many processors are possible in this platform.
NR_CPUS=lots will be appreciable.
All those for_each_cpu() loops will hit NR_CPUS cachelines instead of
hweight(cpu_possible_map) cachelines.