From caf1cbe3677576ef860f7bbba9074d0e6a0bcabb Mon Sep 17 00:00:00 2001 From: James Smart Date: Fri, 23 Oct 2020 15:26:04 -0700 Subject: [PATCH] nvme-fc: track error_recovery while connecting Whenever there are errors during CONNECTING, the driver recovers by aborting all outstanding ios and counts on the io completion to fail them and thus the connection/association they are on. However, the connection failure depends on a failure state from the core routines. Not all commands that are issued by the core routine are guaranteed to cause a failure of the core routine. They may be treated as a failure status and the status is then ignored. As such, whenever the transport enters error_recovery while CONNECTING, it will set a new flag indicating an association failed. The create_association routine which creates and initializes the controller, will monitor the state of the flag as well as the core routine error status and ensure the association fails if there was an error. Signed-off-by: James Smart Signed-off-by: Christoph Hellwig --- drivers/nvme/host/fc.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/drivers/nvme/host/fc.c b/drivers/nvme/host/fc.c index 3c002bdcace3..cfb6ef74c742 100644 --- a/drivers/nvme/host/fc.c +++ b/drivers/nvme/host/fc.c @@ -146,7 +146,8 @@ struct nvme_fc_rport { /* fc_ctrl flags values - specified as bit positions */ #define ASSOC_ACTIVE 0 -#define FCCTRL_TERMIO 1 +#define ASSOC_FAILED 1 +#define FCCTRL_TERMIO 2 struct nvme_fc_ctrl { spinlock_t lock; @@ -2988,6 +2989,8 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) ctrl->cnum, ctrl->lport->localport.port_name, ctrl->rport->remoteport.port_name, ctrl->ctrl.opts->subsysnqn); + clear_bit(ASSOC_FAILED, &ctrl->flags); + /* * Create the admin queue */ @@ -3016,7 +3019,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) */ ret = nvme_enable_ctrl(&ctrl->ctrl); - if (ret) + if (ret || test_bit(ASSOC_FAILED, &ctrl->flags)) goto out_disconnect_admin_queue; ctrl->ctrl.max_segments = ctrl->lport->ops->max_sgl_segments; @@ -3026,7 +3029,7 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) blk_mq_unquiesce_queue(ctrl->ctrl.admin_q); ret = nvme_init_identify(&ctrl->ctrl); - if (ret) + if (ret || test_bit(ASSOC_FAILED, &ctrl->flags)) goto out_disconnect_admin_queue; /* sanity checks */ @@ -3071,9 +3074,9 @@ nvme_fc_create_association(struct nvme_fc_ctrl *ctrl) ret = nvme_fc_create_io_queues(ctrl); else ret = nvme_fc_recreate_io_queues(ctrl); - if (ret) - goto out_term_aen_ops; } + if (ret || test_bit(ASSOC_FAILED, &ctrl->flags)) + goto out_term_aen_ops; changed = nvme_change_ctrl_state(&ctrl->ctrl, NVME_CTRL_LIVE); @@ -3301,6 +3304,7 @@ __nvme_fc_terminate_io(struct nvme_fc_ctrl *ctrl) */ if (ctrl->ctrl.state == NVME_CTRL_CONNECTING) { __nvme_fc_abort_outstanding_ios(ctrl, true); + set_bit(ASSOC_FAILED, &ctrl->flags); return; }