git: 28ed159f2669 - main - pps: Round to closest integer in pps_event()

The branch main has been updated by imp:

URL: https://cgit.FreeBSD.org/src/commit/?id=28ed159f2669b8d401ca95bf5c6e9ff06a997ef9

commit 28ed159f2669b8d401ca95bf5c6e9ff06a997ef9
Author:     Sebastian Huber <sebastian.huber@embedded-brains.de>
AuthorDate: 2023-02-27 21:49:10 +0000
Commit:     Warner Losh <imp@FreeBSD.org>
CommitDate: 2023-02-27 22:10:55 +0000

    pps: Round to closest integer in pps_event()
    
    The comment above bintime2timespec() says:
    
      When converting between timestamps on parallel timescales of differing
      resolutions it is historical and scientific practice to round down.
    
    However, the delta_nsec value is a time difference and not a timestamp.  Also
    the rounding errors accumulate in the frequency accumulator, see hardpps().
    So, rounding to the closest integer is probably slightly better.
    
    Reviewed by: imp
    Pull Request: https://github.com/freebsd/freebsd-src/pull/604
---
 sys/kern/kern_tc.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/sys/kern/kern_tc.c b/sys/kern/kern_tc.c
index 0dc233896baa..e4b3d5e80529 100644
--- a/sys/kern/kern_tc.c
+++ b/sys/kern/kern_tc.c
@@ -1882,6 +1882,7 @@ pps_event(struct pps_state *pps, int event)
 #ifdef PPS_SYNC
 	if (fhard) {
 		uint64_t delta_nsec;
+		uint64_t freq;
 
 		/*
 		 * Feed the NTP PLL/FLL.
@@ -1893,7 +1894,8 @@ pps_event(struct pps_state *pps, int event)
 		tcount &= captc->tc_counter_mask;
 		delta_nsec = 1000000000;
 		delta_nsec *= tcount;
-		delta_nsec /= captc->tc_frequency;
+		freq = captc->tc_frequency;
+		delta_nsec = (delta_nsec + freq / 2) / freq;
 		hardpps(tsp, (long)delta_nsec);
 	}
 #endif