RDMA/core: let rdma_connect_locked() call lockdep_assert_held(&id_priv->handler_mutex)

rdma_accept() also has this, so this is now more consistent and may
prevent bugs in future.

Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Leon Romanovsky <leon@kernel.org>
Cc: linux-rdma@vger.kernel.org
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Link: https://patch.msgid.link/20251008165913.444276-1-metze@samba.org
Signed-off-by: Leon Romanovsky <leon@kernel.org>

diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c
index 5b2d3ae3f9fce85baa71214242658d12f3b87460..95e89f5c147c2c605e99a3fe3c0cddc224e91bf4 100644 (file)
--- a/drivers/infiniband/core/cma.c
+++ b/drivers/infiniband/core/cma.c
@@ -4475,6 +4475,8 @@ int rdma_connect_locked(struct rdma_cm_id *id,
                container_of(id, struct rdma_id_private, id);
        int ret;
 
+       lockdep_assert_held(&id_priv->handler_mutex);
+
        if (!cma_comp_exch(id_priv, RDMA_CM_ROUTE_RESOLVED, RDMA_CM_CONNECT))
                return -EINVAL;