From: Sage Weil <sage@redhat.com>
Date: Wed, 13 Apr 2016 19:22:30 +0000 (-0400)
Subject: osd: fix watch reconnect race
X-Git-Tag: v10.2.0~24^2~1
X-Git-Url: http://git-server-git.apps.pok.os.sepia.ceph.com/?a=commitdiff_plain;h=c04fd4210f660c6458d5ecd9259429aff9fa4881;p=ceph.git

osd: fix watch reconnect race

It's possible to process a watch reconnect op, setting Watch::conn, before
the ms_handle_reset path tries to disconnect the old one.  Since reset()
was blindly disconnecting the "current" connection, we could race and
disconnect the new con instead of the old one.

Fix this by specifying which con to disconnect.

Fixes: http://tracker.ceph.com/issues/15441
Signed-off-by: Sage Weil <sage@redhat.com>
---

diff --git a/src/osd/OSD.cc b/src/osd/OSD.cc
index 7c260ff39aea..d37bc7d70bd3 100644
--- a/src/osd/OSD.cc
+++ b/src/osd/OSD.cc
@@ -4662,7 +4662,7 @@ bool OSD::ms_handle_reset(Connection *con)
   dout(1) << "ms_handle_reset con " << con << " session " << session << dendl;
   if (!session)
     return false;
-  session->wstate.reset();
+  session->wstate.reset(con);
   session->con.reset(NULL);  // break con <-> session ref cycle
   session_handle_reset(session);
   session->put();
diff --git a/src/osd/Watch.cc b/src/osd/Watch.cc
index f4411baf7fee..43846c03a16a 100644
--- a/src/osd/Watch.cc
+++ b/src/osd/Watch.cc
@@ -513,7 +513,7 @@ void WatchConState::removeWatch(WatchRef watch)
   watches.erase(watch);
 }
 
-void WatchConState::reset()
+void WatchConState::reset(Connection *con)
 {
   set<WatchRef> _watches;
   {
@@ -526,7 +526,11 @@ void WatchConState::reset()
     boost::intrusive_ptr<ReplicatedPG> pg((*i)->get_pg());
     pg->lock();
     if (!(*i)->is_discarded()) {
-      (*i)->disconnect();
+      if ((*i)->is_connected(con)) {
+	(*i)->disconnect();
+      } else {
+	generic_derr << __func__ << " not still connected to " << (*i) << dendl;
+      }
     }
     pg->unlock();
   }
diff --git a/src/osd/Watch.h b/src/osd/Watch.h
index 6e4ec37a6b51..c6843ee785de 100644
--- a/src/osd/Watch.h
+++ b/src/osd/Watch.h
@@ -202,6 +202,9 @@ public:
   bool is_connected() {
     return conn.get() != NULL;
   }
+  bool is_connected(Connection *con) {
+    return conn.get() == con;
+  }
 
   /// NOTE: must be called with pg lock held
   ~Watch();
@@ -290,7 +293,7 @@ public:
     );
 
   /// Called on session reset, disconnects watchers
-  void reset();
+  void reset(Connection *con);
 };
 
 #endif