start_time patch for fhandler_process.cc
Wed Jul 9 10:07:00 GMT 2003
> On Tue, Jul 08, 2003 at 07:09:12PM +0100, Chris January wrote:
> >Try this Chris and see if it solves the start time problem.
> >2003-07-28 Chris January <kseitz@firstname.lastname@example.org>
> > * fhandler_process.cc (format_process_stat): Changed the
> calculation for
> Sorry, no.
> Unknown HZ value! (250) Assume 100.
> USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
> cgf 3452 0.0 1.0 2544 2680 ? R Aug08 0:00 procps auwx
> Now that I've read the description of what the field is supposed to
> contain, I'm wondering if the culprit is the "Unknown HZ value! (250)
> Assume 100."
> Could that be it?
Almost certainly. That would give you dates 250/100 = 2.5 times into the
future I should think.
Could the problem be you have an SMP machine? procps calibrates HZ using
uptime and total cpu jiffies. The /proc implementation asks NT how many
ticks it spent in kernel, user mode, etc. However maybe NT counts ticks more
than once if you have more than one processor?
More information about the Cygwin-patches