Re: [RFC][PATCH - 4/13] NTP cleanup: Breakup ntp_adjtimex()

[john stultz <johnstul@us.ibm.com>]

On Wed, 2005-08-10 at 18:27 -0700, john stultz wrote:
> All,
> 	This patch breaks up the complex nesting of code in ntp_adjtimex() by
> creating a ntp_hardupdate() function and simplifying some of the logic.
> This also mimics the documented NTP spec somewhat better.
> 
> Any comments or feedback would be greatly appreciated.

Ugh. I just caught a bug where I misplaced the parens. 

> -			} /* STA_PLL */
> +				else if (ntp_hardupdate(txc->offset, xtime))
> +					result = TIME_ERROR;
> +			}
>  		} /* txc->modes & ADJ_OFFSET */
 
That's wrong. The following patch fixes it. 

thanks
-john


diff --git a/kernel/ntp.c b/kernel/ntp.c
--- a/kernel/ntp.c
+++ b/kernel/ntp.c
@@ -388,9 +388,8 @@ int ntp_adjtimex(struct timex *txc)
 				/* adjtime() is independent from ntp_adjtime() */
 				if ((time_next_adjust = txc->offset) == 0)
 					time_adjust = 0;
-				else if (ntp_hardupdate(txc->offset, xtime))
-					result = TIME_ERROR;
-			}
+			} else if (ntp_hardupdate(txc->offset, xtime))
+				result = TIME_ERROR;
 		} /* txc->modes & ADJ_OFFSET */
 
 		if (txc->modes & ADJ_TICK) {