ftimer - report realtime itimer status
ftimer provides the cpu number of the processor handling the fast clock
used by the real time itimer facility. It also reports any outstanding
real time itimer timeouts.
The fast clock is inactive until it is first used, then remains active
from that time onward. The fast clock typically becomes active when a
realtime process (i.e., those running with a non-degrading priority, see
npri(1)) executes setitimer(2), or less frequently when some special
kernel driver needs the fast clock enabled for a high resolution interval
timer and executes an internal kernel routine.
A user can control which processor handles the fast clock interrupts by
using the mpadmin(1) command or the sysmp(2) programmatic syscall.
Note, however, that ftimer(1) has no real relevance for Challenge/Onyx
platforms, as "fast clock" interrupts are handled by all processors, not
just by a single processor. For Challenge/Onyx the distinction between
"fast clock" itimers and "normal" itimers is made on the basis of what
time resolution is available, not on the basis of which physical hardware
clock ("fast" or "normal") implements the itimer. As with the other
platforms, realtime processes can get access to high resolution itimers,
and non-realtime processes can only use "normal" resolution itimers (see
setitimer(2) and timers(5)).
npri(1), getitimer(2), setitimer(2), realtime(5), timers(5), mpadmin(1).
PPPPaaaaggggeeee 1111 [ Back ]