Re: [PATCH] Re: time() glitch on 2.4.18: solved

[Jim Paris]

> > > +		if (count > LATCH) {
> > 
> > may be (count >= LATCH) would be even better ?
> 
> Some PIT clones seem to hold the LATCH value momentarily judging by
> other things that were triggered wrongly by >=

If so, then that's a separate problem: the later code

	count = ((LATCH-1) - count) * TICK_SIZE;
	delay_at_last_interrupt = (count + LATCH/2) / LATCH;

will cause delay_at_last_interrupt to be negative, so you'll see
backwards jumps in time and occasional wraparound of usecs as I did.
Perhaps a 

	if (count == LATCH) 
	   count--;

after the reset code?

-jim