From: Sage Weil <sage@redhat.com>
Date: Fri, 31 Oct 2014 23:25:09 +0000 (-0700)
Subject: msg/Pipe: inject delay in stop_and_wait
X-Git-Tag: v0.87.1~31^2~2
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=c3335c7aa6c1e6c3f1879c0cd3cd2f13091221be;p=ceph.git

msg/Pipe: inject delay in stop_and_wait

Inject a delay in stop_and_wait.  This will mostly affect the connection
race Pipe takeover code which currently calls stop_and_wait while holding
the msgr->lock.  This should make it easier for a racing fast_dispatch
method to get stuck on a call that (indirectly) needs the msgr lock.
See #9921.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 2fe5c4c305218fdb1771857e4e0ef7c98a8d0fb6)
---

diff --git a/src/msg/Pipe.cc b/src/msg/Pipe.cc
index e73e2b1c8326..32e20db3f156 100644
--- a/src/msg/Pipe.cc
+++ b/src/msg/Pipe.cc
@@ -1437,6 +1437,15 @@ void Pipe::stop_and_wait()
 {
   if (state != STATE_CLOSED)
     stop();
+
+  if (msgr->cct->_conf->ms_inject_internal_delays) {
+    ldout(msgr->cct, 10) << __func__ << " sleep for "
+			 << msgr->cct->_conf->ms_inject_internal_delays
+			 << dendl;
+    utime_t t;
+    t.set_from_double(msgr->cct->_conf->ms_inject_internal_delays);
+    t.sleep();
+  }
   
   // HACK: we work around an annoying deadlock here.  If the fast
   // dispatch method calls mark_down() on itself, it can block here