[PATCH] ntpd: don't stay at short polling interval

Miroslav Lichvar mlichvar at redhat.com
Fri Jan 9 14:29:32 UTC 2015


On Sun, Jan 04, 2015 at 06:12:41PM +0100, Denys Vlasenko wrote:
> > Great, thanks!
> >
> > The changes you have made to the patch seems to disable increasing the
> > polling interval when the -w option is used, was that intended?
> 
> Are you referring to this hunk being not applied?

I meant the following hunk and the other changes related to it.

        filter_datapoints(p);
        q = select_and_cluster();
-       rc = -1;
+       rc = 1;
        if (q) {
-               rc = 0;
                if (!(option_mask32 & OPT_w)) {
                        rc = update_local_clock(q);

In git there is now:

        filter_datapoints(p);
        q = select_and_cluster();
        rc = 0;
        if (q) {
                if (!(option_mask32 & OPT_w)) {
                        rc = update_local_clock(q);

which means with -w the polling interval will stay at the minimum
unless the selection fails and q is zero. I guess -w is mainly used
for debugging, so it's probably not a big deal, I'd just not expect it
to influence the poll control.

-- 
Miroslav Lichvar


More information about the busybox mailing list