nvme-rdma: rework rdma connection establishment error path
authorSagi Grimberg <sagi@grimberg.me>
Thu, 4 May 2017 10:33:10 +0000 (13:33 +0300)
committerChristoph Hellwig <hch@lst.de>
Thu, 15 Jun 2017 12:28:20 +0000 (14:28 +0200)
Instead of introducing a flag for if the queue is allocated,
simply free the rdma resources when we get the error.

We allocate the queue rdma resources when we have an address
resolution, their we allocate (or take a reference on) our device
so we should free it when we have error after the address resolution
namely:
1. route resolution error
2. connect reject
3. connect error
4. peer unreachable error

Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
drivers/nvme/host/rdma.c

index 2d4a74045d4486fa6e0ca64ee5397d96e6b65a8a..aea78d67f8beba6c221559c91c0412507e047ef5 100644 (file)
@@ -81,9 +81,8 @@ struct nvme_rdma_request {
 
 enum nvme_rdma_queue_flags {
        NVME_RDMA_Q_CONNECTED           = 0,
-       NVME_RDMA_IB_QUEUE_ALLOCATED    = 1,
-       NVME_RDMA_Q_DELETING            = 2,
-       NVME_RDMA_Q_LIVE                = 3,
+       NVME_RDMA_Q_DELETING            = 1,
+       NVME_RDMA_Q_LIVE                = 2,
 };
 
 struct nvme_rdma_queue {
@@ -466,9 +465,6 @@ static void nvme_rdma_destroy_queue_ib(struct nvme_rdma_queue *queue)
        struct nvme_rdma_device *dev;
        struct ib_device *ibdev;
 
-       if (!test_and_clear_bit(NVME_RDMA_IB_QUEUE_ALLOCATED, &queue->flags))
-               return;
-
        dev = queue->device;
        ibdev = dev->dev;
        rdma_destroy_qp(queue->cm_id);
@@ -525,7 +521,6 @@ static int nvme_rdma_create_queue_ib(struct nvme_rdma_queue *queue)
                ret = -ENOMEM;
                goto out_destroy_qp;
        }
-       set_bit(NVME_RDMA_IB_QUEUE_ALLOCATED, &queue->flags);
 
        return 0;
 
@@ -590,7 +585,6 @@ static int nvme_rdma_init_queue(struct nvme_rdma_ctrl *ctrl,
        return 0;
 
 out_destroy_cm_id:
-       nvme_rdma_destroy_queue_ib(queue);
        rdma_destroy_id(queue->cm_id);
        return ret;
 }
@@ -1374,12 +1368,14 @@ static int nvme_rdma_cm_handler(struct rdma_cm_id *cm_id,
                complete(&queue->cm_done);
                return 0;
        case RDMA_CM_EVENT_REJECTED:
+               nvme_rdma_destroy_queue_ib(queue);
                cm_error = nvme_rdma_conn_rejected(queue, ev);
                break;
-       case RDMA_CM_EVENT_ADDR_ERROR:
        case RDMA_CM_EVENT_ROUTE_ERROR:
        case RDMA_CM_EVENT_CONNECT_ERROR:
        case RDMA_CM_EVENT_UNREACHABLE:
+               nvme_rdma_destroy_queue_ib(queue);
+       case RDMA_CM_EVENT_ADDR_ERROR:
                dev_dbg(queue->ctrl->ctrl.device,
                        "CM error event %d\n", ev->event);
                cm_error = -ECONNRESET;