On Mon, Jan 20, 2003 at 05:42:16PM -0800, john stultz wrote:
> get_cycles() is a poor method for determining "real time".
> Please use do_gettimeofday().
Does do_gettimeofday() exist on all platforms? Does it indeed
give actual wall clock time, instead of the inaccurate time jiffies can
give?
> I believe you mean "udelay for 60 micro-seconds"
No, I mean 60 *seconds*. There have been drivers and such that
do pause the entire box that long.
> Wait, so 180 seconds is the margin of error?
Yup, for the default. I'm not beholden to these specific
defaults. They're just the defaults we use in our environment.
> > + if (tsc_diff > hangcheck_tsc_margin) {
>
> but now we're using it to compare cycles! 180sec != 180 cycles
Look at the calculations. I'm comparing cycles to cycles,
calculated from the original seconds.
> Additionally, this code doesn't take systems that have unsync'ed TSCs,
> or systems that change cpu frequency into account. Again, please use
> do_gettimeofday(). Then you can then talk about the values returned in
> secs and usecs, and I believe things will be much more clear.
I'll look into it, but it must absolutely be in terms of wall
clock time as measured from outside the system.
Joel
--Life's Little Instruction Book #99
"Think big thoughts, but relish small pleasures."
Joel Becker Senior Member of Technical Staff Oracle Corporation E-mail: joel.becker@oracle.com Phone: (650) 506-8127 - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Jan 23 2003 - 22:00:25 EST