I'm not sure whether avoiding incrementing here is an ideal move, but
this diff definitely works toward a local optimum. Namely, that error
check is technically meaningless because signed overflow is undefined.
ok? Or would people prefer a solution that's robust to changing
*curpps's type?
Index: sys/kern/kern_time.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_time.c,v
retrieving revision 1.96
diff -u -p -r1.96 kern_time.c
--- sys/kern/kern_time.c 5 Dec 2015 10:11:53 -0000 1.96
+++ sys/kern/kern_time.c 10 Feb 2016 05:25:35 -0000
@@ -765,20 +765,8 @@ ppsratecheck(struct timeval *lasttime, i
else
rv = 0;
-#if 1 /*DIAGNOSTIC?*/
- /* be careful about wrap-around */
- if (*curpps + 1 > *curpps)
- *curpps = *curpps + 1;
-#else
- /*
- * assume that there's not too many calls to this function.
- * not sure if the assumption holds, as it depends on *caller's*
- * behavior, not the behavior of this function.
- * IMHO it is wrong to make assumption on the caller's behavior,
- * so the above #if is #if 1, not #ifdef DIAGNOSTIC.
- */
- *curpps = *curpps + 1;
-#endif
+ if (*curpps != INT_MAX)
+ (*curpps)++;
return (rv);
}
