3.16.63-rc1 review patch. If anyone has any objections, please let me know. ------------------ From: Parav Pandit <parav@xxxxxxxxxxxx> commit 33f93e1ebcf5acfaef06cda2d3e373730519e33e upstream. In case of LAP are used for RoCE, it can lead to a problem of sleeping a context while spin lock is held in below flow. cm_lap_handler ->spin_lock -> <..switch_case..> -> cm_init_av_for_response -> ib_init_ah_from_wc -> rdma_addr_find_l2_eth_by_grh wait_for_completion() Therefore ah attribute initialization is done for incoming lap requests outside of the lock context. Signed-off-by: Parav Pandit <parav@xxxxxxxxxxxx> Reviewed-by: Daniel Jurgens <danielj@xxxxxxxxxxxx> Signed-off-by: Leon Romanovsky <leon@xxxxxxxxxx> Signed-off-by: Jason Gunthorpe <jgg@xxxxxxxxxxxx> Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx> --- drivers/infiniband/core/cm.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) --- a/drivers/infiniband/core/cm.c +++ b/drivers/infiniband/core/cm.c @@ -2770,6 +2770,12 @@ static int cm_lap_handler(struct cm_work if (!cm_id_priv) return -EINVAL; + ret = cm_init_av_for_response(work->port, work->mad_recv_wc->wc, + work->mad_recv_wc->recv_buf.grh, + &cm_id_priv->av); + if (ret) + goto deref; + param = &work->cm_event.param.lap_rcvd; param->alternate_path = &work->path[0]; cm_format_path_from_lap(cm_id_priv, param->alternate_path, lap_msg); @@ -2809,11 +2815,6 @@ static int cm_lap_handler(struct cm_work cm_id_priv->id.lap_state = IB_CM_LAP_RCVD; cm_id_priv->tid = lap_msg->hdr.tid; - ret = cm_init_av_for_response(work->port, work->mad_recv_wc->wc, - work->mad_recv_wc->recv_buf.grh, - &cm_id_priv->av); - if (ret) - goto unlock; cm_init_av_by_path(param->alternate_path, &cm_id_priv->alt_av, cm_id_priv); ret = atomic_inc_and_test(&cm_id_priv->work_count);
