This patch applies the suggestions made by Con Kolivas for improving the smpnice code. The non cosmetic part of the patch addresses the fact the mapping from nice values to task load weights for negative nice values does not match the implied CPU allocations in the function task_timeslice(). As suggested by Con the mapping function now uses the time slice information directly (via a slightly modified interface). Signed-off-by: Peter Williams Peter -- Peter Williams pwil3058@bigpond.net.au "Learning, n. The kind of ignorance distinguishing the studious." -- Ambrose Bierce