This patch against 2.4.23 adds an x86 sysctl,
/proc/sys/kernel/idle_poll, which allows idle behavior to be toggled
between polling and using HLT instructions.

Since this is the first time I've ever messed with kernel code, I'd
appreciate it if those more knowledgeable can give it a quick glance and
let me know if I've made any particularly obvious errors.  For example,
should I be doing something to ensure that the value of pm_idle_poll is
actually read from memory in default_idle?  (sysctl.c doesn't like it if
I just make pm_idle_poll volatile.)

Thanks!
-- 
========================================================================
Ian Pilcher                                        i.pilcher@comcast.net
========================================================================