From: Harlan Stenn <stenn@ntp.org>
Date: Tue, 12 Oct 2010 05:28:24 +0000 (-0400)
Subject: [Bug 750]: Non-existing device causes coredump with RIPE-NCC driver
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=43839409418a1cd6da9505d72a69b6b749a01b22;p=thirdparty%2Fntp.git

[Bug 750]: Non-existing device causes coredump with RIPE-NCC driver

bk: 4cb3f1f8R7gFSNL509ywMU0oxh_IRw
---

diff --git a/ChangeLog b/ChangeLog
index fb24902a61..bf01e325ea 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,12 +1,12 @@
 ---
 
-* [Bug 750]: Non-existing device causes coredump with RIPE-NCC driver.
-* [Bug 1567]: Support Arbiter 1093C Satellite Clock on Windows.
-* [Bug 1643]: Range-check the decoding of the RIPE-NCC status codes.
-* [Bug 1644]: cvo.sh should use lsb_release to identify linux distros.
-* [Bug 1659]: Support Truetime Satellite Clocks on Windows.
-* [Bug 1660]: On some systems, test is in /usr/bin, not /bin.
-* [Bug 1661]: Re-indent refclock_ripencc.c.
+* [Bug 750] Non-existing device causes coredump with RIPE-NCC driver.
+* [Bug 1567] Support Arbiter 1093C Satellite Clock on Windows.
+* [Bug 1643] Range-check the decoding of the RIPE-NCC status codes.
+* [Bug 1644] cvo.sh should use lsb_release to identify linux distros.
+* [Bug 1659] Support Truetime Satellite Clocks on Windows.
+* [Bug 1660] On some systems, test is in /usr/bin, not /bin.
+* [Bug 1661] Re-indent refclock_ripencc.c.
 
 ---
 (4.2.6p3-RC2) 2010/09/25 Released by Harlan Stenn <stenn@ntp.org>
diff --git a/ntpd/refclock_ripencc.c b/ntpd/refclock_ripencc.c
index 02a1dc4dc4..3ce3a73750 100644
--- a/ntpd/refclock_ripencc.c
+++ b/ntpd/refclock_ripencc.c
@@ -514,7 +514,7 @@ ripencc_start(int unit, struct peer *peer)
 	pp->io.srcclock = (caddr_t)peer;
 	pp->io.datalen = 0;
 	if (!io_addclock(&pp->io)) {
-		/* XXX HMS: Do we need to set pp->io.fd to -1 here? */
+		pp->io.fd = -1;
 		(void) close(fd);
 		free(up);
 		return (0);