From: Jason Gunthorpe jgg@nvidia.com
[ Upstream commit d54f23c09ec62670901f1a2a4712a5218522ca2b ]
cma_process_remove() triggers an unconditional rdma_destroy_id() for internal_id's and skips the event deliver and transition through RDMA_CM_DEVICE_REMOVAL.
This is confusing and unnecessary. internal_id always has cma_listen_handler() as the handler, have it catch the RDMA_CM_DEVICE_REMOVAL event and directly consume it and signal removal.
This way the FSM sequence never skips the DEVICE_REMOVAL case and the logic in this hard to test area is simplified.
Link: https://lore.kernel.org/r/20200723070707.1771101-2-leon@kernel.org Signed-off-by: Leon Romanovsky leonro@mellanox.com Signed-off-by: Jason Gunthorpe jgg@nvidia.com Signed-off-by: Sasha Levin sashal@kernel.org --- drivers/infiniband/core/cma.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index c30cf5307ce3e..537eeebde5f4d 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -2482,6 +2482,10 @@ static int cma_listen_handler(struct rdma_cm_id *id, { struct rdma_id_private *id_priv = id->context;
+ /* Listening IDs are always destroyed on removal */ + if (event->event == RDMA_CM_EVENT_DEVICE_REMOVAL) + return -1; + id->context = id_priv->id.context; id->event_handler = id_priv->id.event_handler; trace_cm_event_handler(id_priv, event); @@ -4829,7 +4833,7 @@ static void cma_process_remove(struct cma_device *cma_dev) cma_id_get(id_priv); mutex_unlock(&lock);
- ret = id_priv->internal_id ? 1 : cma_remove_id_dev(id_priv); + ret = cma_remove_id_dev(id_priv); cma_id_put(id_priv); if (ret) rdma_destroy_id(&id_priv->id);