[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <1283870626-9472-1-git-send-email-mlichvar@redhat.com>
Date: Tue, 7 Sep 2010 16:43:46 +0200
From: Miroslav Lichvar <mlichvar@...hat.com>
To: linux-kernel@...r.kernel.org
Cc: john stultz <johnstul@...ibm.com>,
Miroslav Lichvar <mlichvar@...hat.com>
Subject: [PATCH] ntp: Clamp PLL update interval
Clamp update interval to reduce PLL gain with low sampling rate (e.g.
intermittent network connection) to avoid instability.
The clamp roughly corresponds to the loop time constant, it's 8 * poll
interval for SHIFT_PLL 2 and 32 * poll interval for SHIFT_PLL 4. This
gives good results without affecting the gain in normal conditions where
ntpd skips only up to seven consecutive samples.
Signed-off-by: Miroslav Lichvar <mlichvar@...hat.com>
---
kernel/time/ntp.c | 14 +++++++++++---
1 files changed, 11 insertions(+), 3 deletions(-)
diff --git a/kernel/time/ntp.c b/kernel/time/ntp.c
index c631168..d232189 100644
--- a/kernel/time/ntp.c
+++ b/kernel/time/ntp.c
@@ -149,10 +149,18 @@ static void ntp_update_offset(long offset)
time_reftime = get_seconds();
offset64 = offset;
- freq_adj = (offset64 * secs) <<
- (NTP_SCALE_SHIFT - 2 * (SHIFT_PLL + 2 + time_constant));
+ freq_adj = ntp_update_offset_fll(offset64, secs);
- freq_adj += ntp_update_offset_fll(offset64, secs);
+ /*
+ * Clamp update interval to reduce PLL gain with low
+ * sampling rate (e.g. intermittent network connection)
+ * to avoid instability.
+ */
+ if (unlikely(secs > 1 << (SHIFT_PLL + 1 + time_constant)))
+ secs = 1 << (SHIFT_PLL + 1 + time_constant);
+
+ freq_adj += (offset64 * secs) <<
+ (NTP_SCALE_SHIFT - 2 * (SHIFT_PLL + 2 + time_constant));
freq_adj = min(freq_adj + time_freq, MAXFREQ_SCALED);
--
1.7.2
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists