I've sent a different patch to fix the propagation of bfd remote state: http://patchwork.openvswitch.org/patch/2623/
As mentioned, this patch code was in an earlier version of a patchset, and although not strictly necessary, I think it cleans up the codebase a bit. I think the commit message should be reworded though. I'll plan to do this. On 24 December 2013 16:25, Joe Stringer <[email protected]> wrote: > Ah, I reported in the message that it fixes the issue but I am not > sure, still investigating. This somehow got dropped from earlier > versions of the patchset though, and I think should still be merged. > > On 24 December 2013 16:19, Joe Stringer <[email protected]> wrote: >> Commit 7f8f2757f3 ("bridge: Only store instant_stats on device changes") >> introduced a race condition where instant stats would sometimes not be >> updated to the database. In this case, monitor status reported by >> 'ovs-appctl bfd/show' would differ from 'ovs-vsctl list int'. This patch >> fixes the issue by ensuring that the main thread waits on >> connectivity_seqno in instant_stats_wait(). >> >> Bug #22136. >> >> Signed-off-by: Joe Stringer <[email protected]> >> --- >> vswitchd/bridge.c | 1 + >> 1 file changed, 1 insertion(+) >> >> diff --git a/vswitchd/bridge.c b/vswitchd/bridge.c >> index 6311ff3..f569699 100644 >> --- a/vswitchd/bridge.c >> +++ b/vswitchd/bridge.c >> @@ -2253,6 +2253,7 @@ instant_stats_run(void) >> static void >> instant_stats_wait(void) >> { >> + seq_wait(connectivity_seq_get(), connectivity_seqno); >> if (instant_txn) { >> ovsdb_idl_txn_wait(instant_txn); >> } else if (instant_stats_could_have_changed) { >> -- >> 1.7.9.5 >> _______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
