After some thought, my patch was huge overkill. The correct solution (with
same result) should be as in following patch.
I don't see a point in forcing CLOCK_REALTIME being master and completely
ignoring current portState. This patch is supposed to fix that.
--- a/phc2sys.c 2022-10-14 09:24:53.643279501 +0200
+++ b/phc2sys.c 2022-10-14 09:17:27.256854202 +0200
@@ -558,11 +558,11 @@
rt->state = PS_SLAVE;
} else if (rt) {
if (rt->state != PS_MASTER) {
- rt->state = PS_MASTER;
- clock_reinit(priv, rt, rt->state);
+ pr_info("portState is not Master, not synchronizing");
+ } else {
+ LIST_INSERT_HEAD(&priv->dst_clocks, rt, dst_list);
+ pr_info("selecting %s for synchronization", rt->device);
}
- LIST_INSERT_HEAD(&priv->dst_clocks, rt, dst_list);
- pr_info("selecting %s for synchronization", rt->device);
}
priv->master = src;
pr_info("selecting %s as the master clock", src->device); _______________________________________________
Linuxptp-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
