On Tue, 2015-05-26 at 13:08 +0200, Bart Van Assche wrote: > fc_invoke_resp() tests resp_task before setting it. Hence initialize > resp_task in fc_exch_em_alloc(). > > Signed-off-by: Bart Van Assche <[email protected]> > Cc: stable <[email protected]> > --- > drivers/scsi/libfc/fc_exch.c | 1 + > 1 file changed, 1 insertion(+) > > diff --git a/drivers/scsi/libfc/fc_exch.c b/drivers/scsi/libfc/fc_exch.c > index 1b3a094..fb4cc5e 100644 > --- a/drivers/scsi/libfc/fc_exch.c > +++ b/drivers/scsi/libfc/fc_exch.c > @@ -868,6 +868,7 @@ hit: > ep->rxid = FC_XID_UNKNOWN; > ep->class = mp->class; > ep->resp_active = 0; > + ep->resp_task = NULL;
This shouldn't be issue since entire ep struct is initialized first, so not sure what really fixed which should go to stable kernel also. Added resp_active should be also removed then have only non zero init stuff. //Vasu > init_waitqueue_head(&ep->resp_wq); > INIT_DELAYED_WORK(&ep->timeout_work, fc_exch_timeout); > out: _______________________________________________ fcoe-devel mailing list [email protected] http://lists.open-fcoe.org/mailman/listinfo/fcoe-devel
