The iowait_sdma_drained() callback lacked locking to
protect the qp s_flags field.
This causes the s_flags to be out of sync
on multiple CPUs, potentially corrupting the s_flags.
Fixes: a545f5308b6c ("staging/rdma/hfi: fix CQ completion order issue")
Reviewed-by: Sebastian Sanchez <sebastian.sanchez@intel.com>
Signed-off-by: Mike Marciniszyn <mike.marciniszyn@intel.com>
Reviewed-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
         * do the flush work until that QP's
         * sdma work has finished.
         */
+       spin_lock(&qp->s_lock);
        if (qp->s_flags & RVT_S_WAIT_DMA) {
                qp->s_flags &= ~RVT_S_WAIT_DMA;
                hfi1_schedule_send(qp);
        }
+       spin_unlock(&qp->s_lock);
 }
 
 /**