Granularity in getitimer() and gettimeofday()

[tim at NLM-VAX.ARPA]

From:  Tim Davis <tim at NLM-VAX.ARPA>

We are attempting to do some real-time I/O statistics on our VAX 11/780
with 4.2 BSD.  The getitimer(2) man page says the resolution of the system
clock is rounded to 10 microseconds; however, the timevals we get back from
getitimer() and gettimeofday() have the .tv_usec field rounded to 10
MILLIseconds.  We haven't been able to locate this rounding action in the
source.  Does any one know what's going on here?

Tim Davis
LHNCBC/National Library of Medicine
(301) 496-1936

tim at nlm-vax ARPA