Re: [RFC][PATCH] timekeeping: Limit system time to prevent 32-bit time_t overflow

[Miroslav Lichvar <mlichvar@redhat.com>]

On Wed, Apr 15, 2015 at 12:17:36PM -0400, Justin Keller wrote:
> Is there a reason for "step = leap"?

It's there to not change the behavior when a leap second occurs, the
clock still needs to be stepped. I guess it could be optimized a bit,
if it used "if (unlikely(leap || tk->xtime_sec >= time_max_sec))", the
64-bit step variable wouldn't have to be used in normal operation.

> >                 /* Figure out if its a leap sec and apply if needed */
> >                 leap = second_overflow(tk->xtime_sec);
> > -               if (unlikely(leap)) {
> > +               step = leap;
> > +
> > +               /* If the system time reached the maximum, step it back */
> > +               if (unlikely(tk->xtime_sec >= time_max_sec)) {
> > +                       step = time_max_sec - tk->xtime_sec - SEC_PER_WEEK;
> > +                       printk(KERN_NOTICE
> > +                               "Clock: maximum time reached, stepping back\n");
> > +               }
> > +
> > +               if (unlikely(step)) {
> >                         struct timespec64 ts;
> >
> > -                       tk->xtime_sec += leap;
> > +                       tk->xtime_sec += step;

-- 
Miroslav Lichvar