On 4/19/23 11:47, Daniel Wagner wrote:
We need to free the resources in the opposite order as we allocate them. Signed-off-by: Daniel Wagner <dwagner@xxxxxxx> --- tests/nvme/rc | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tests/nvme/rc b/tests/nvme/rc index b44239446dcf..ec0cc2d8d8cc 100644 --- a/tests/nvme/rc +++ b/tests/nvme/rc @@ -204,10 +204,10 @@ _cleanup_fcloop() { local remote_wwnn="${3:-$def_remote_wwnn}" local remote_wwpn="${4:-$def_remote_wwpn}"- _nvme_fcloop_del_rport "${local_wwnn}" "${local_wwpn}" \- "${remote_wwnn}" "${remote_wwpn}" _nvme_fcloop_del_tport "${remote_wwnn}" "${remote_wwpn}" _nvme_fcloop_del_lport "${local_wwnn}" "${local_wwpn}" + _nvme_fcloop_del_rport "${local_wwnn}" "${local_wwpn}" \ + "${remote_wwnn}" "${remote_wwpn}" }_cleanup_nvmet() {
Does this fix a bug? if it does, than it should probably be documented that there is a driver bug because userspace teardown ordering should not trigger a driver bug.
