From: phil <phil>
Date: Mon, 13 Oct 2003 22:31:02 +0000 (+0000)
Subject: The pinger was deadlocking on itself, in some rare and improbably but
X-Git-Tag: v1_7_0_51~2^9~248
X-Git-Url: https://git.whamcloud.com/?a=commitdiff_plain;h=1217bde96ab9949d28a854c7600f13967f0cb44d;p=fs%2Flustre-release.git

The pinger was deadlocking on itself, in some rare and improbably but
very real cases, so I reworked it to be simpler.

It's started once from ptlrpc_init now, and stopped once from ptlrpc_exit

Adding and deleting imports no longer causes the thread to be created
and go away.  Most importantly, we no longer hold the sem in
stop_pinger across the wake_up and subsequent l_wait_event; if the
pinger thread was already doing something, blocked on the pinger_sem,
it would never wake up and deadlock would ensue.
---

diff --git a/lustre/ptlrpc/ptlrpc_internal.h b/lustre/ptlrpc/ptlrpc_internal.h
index cfb2b55..cdf6f69 100644
--- a/lustre/ptlrpc/ptlrpc_internal.h
+++ b/lustre/ptlrpc/ptlrpc_internal.h
@@ -98,5 +98,9 @@ enum {
 
 int ptlrpc_expire_one_request(struct ptlrpc_request *req);
 
+/* pinger.c */
+int ptlrpc_start_pinger(void);
+int ptlrpc_stop_pinger(void);
 void ptlrpc_pinger_sending_on_import(struct obd_import *imp);
+
 #endif /* PTLRPC_INTERNAL_H */
diff --git a/lustre/ptlrpc/ptlrpc_module.c b/lustre/ptlrpc/ptlrpc_module.c
index 8486b76..d0a8da5 100644
--- a/lustre/ptlrpc/ptlrpc_module.c
+++ b/lustre/ptlrpc/ptlrpc_module.c
@@ -113,11 +113,14 @@ __init int ptlrpc_init(void)
 
         ptlrpc_put_connection_superhack = ptlrpc_put_connection;
         ptlrpc_abort_inflight_superhack = ptlrpc_abort_inflight;
+
+        ptlrpc_start_pinger();
         RETURN(0);
 }
 
 static void __exit ptlrpc_exit(void)
 {
+        ptlrpc_stop_pinger();
         ptlrpc_exit_portals();
         ptlrpc_cleanup_connection();
         llog_cleanup_commit_master(0);