testing the actual cpu limits, and latencies for application timers.
it looks like it just adds 60ns of latency on my system.
invoke: $ ./cpu-stress sec usec example: $ cpu-stress 1 10 398781499 1.000010,1.000015,0.000005 Asked delta: 1.000010 seconds Got delta: 1.000015 seconds Latency: 0.000005 seconds First line is the count, a measure of amount of work done by the cpu-stress program in while(1) count++; a,b,c represent the asked timer delta, actual timer delta and the latency respectively. You ask for a 1s+10us, get an alarm at 1s+15us and the latency (mostly due to the extra gettimeofday call). Last 3 lines are output to stderr, so it will not interfere with logging.
but it defaults to no colour if you don't have the package. so just do: easy_install termcolor to get (hopefully) awesome colours.
invoke as root: scripts/of.sh eth0,eth1,eth2.. to bridge eth0,eth1,.. it's just for knowing how to invoke.. not intended for real use ;)
.. that udev/tmpfs/devtmpfs were supposed to create. *sigh*!
... to solve the problem of "Begin: waiting for root fs" this wasn't a sata problem.
some packet forwarding problem.. will check it out sometime
tring dryrun of userspace isolated switch