When ptlrpc_start_thread fails to create a new thread, it will
finalize and free a struct ptlrpc_thread created and used here.
Considering this, it can be a problem when ptlrpc_svcpt_stop_thread
is driven and handles the struct ptlrpc_thread right after or right
before failure of cfs_create_thread. Because this situation let
the both of ptlrpc_start_thread and ptlrpc_svcpt_stop_threads
access the freed ptlrpc_thread and cause OS panic. Or, it may
happen that ptlrpc_svcpt_stop_threads waits forever holding an
already-freed waitq.
This patch adds an error handling into ptlrpc_start_thread to fix
this problem.
Signed-off-by: Hiroya Nozaki <nozaki.hiroya@jp.fujitsu.com>
Change-Id: Ic25f40f8650c65e21abe4df127fc20f9c7d0fcd1
Reviewed-on: http://review.whamcloud.com/5552
Tested-by: Hudson
Tested-by: Maloo <whamcloud.maloo@gmail.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Nikitas Angelinas <nikitas_angelinas@xyratex.com>
Reviewed-by: Keith Mannthey <keith.mannthey@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
CERROR("cannot start thread '%s': rc %d\n",
thread->t_name, rc);
spin_lock(&svcpt->scp_lock);
CERROR("cannot start thread '%s': rc %d\n",
thread->t_name, rc);
spin_lock(&svcpt->scp_lock);
- cfs_list_del(&thread->t_link);
--svcpt->scp_nthrs_starting;
--svcpt->scp_nthrs_starting;
- spin_unlock(&svcpt->scp_lock);
-
- OBD_FREE(thread, sizeof(*thread));
+ if (thread_is_stopping(thread)) {
+ /* this ptlrpc_thread is being hanled
+ * by ptlrpc_svcpt_stop_threads now
+ */
+ thread_add_flags(thread, SVC_STOPPED);
+ cfs_waitq_signal(&thread->t_ctl_waitq);
+ spin_unlock(&svcpt->scp_lock);
+ } else {
+ cfs_list_del(&thread->t_link);
+ spin_unlock(&svcpt->scp_lock);
+ OBD_FREE_PTR(thread);
+ }