[PATCH v1] nvmet-rdma: Invoke fatal error on error completion

[PATCH v1] nvmet-rdma: Invoke fatal error on error completion

[Sagi Grimberg]

In case we got an error completion the rdma queue pair
is in error state, teardown the entire controller. Note
that in recv or read error completion we might not have
a controller yet, so check for the controller existence.

Signed-off-by: Sagi Grimberg <sagi at grimberg.me>
---
Changes from v0:
- Added curly braces to nvmet_rdma_error_comp conditionals
- aligned multi-line if condition indentation

 drivers/nvme/target/rdma.c | 46 ++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 38 insertions(+), 8 deletions(-)

diff --git a/drivers/nvme/target/rdma.c b/drivers/nvme/target/rdma.c
index b1c6e5bb0b70..5005cda9fdcc 100644
--- a/drivers/nvme/target/rdma.c
+++ b/drivers/nvme/target/rdma.c
@@ -134,6 +134,7 @@ static void nvmet_rdma_send_done(struct ib_cq *cq, struct ib_wc *wc);
 static void nvmet_rdma_recv_done(struct ib_cq *cq, struct ib_wc *wc);
 static void nvmet_rdma_read_data_done(struct ib_cq *cq, struct ib_wc *wc);
 static void nvmet_rdma_qp_event(struct ib_event *event, void *priv);
+static void nvmet_rdma_queue_disconnect(struct nvmet_rdma_queue *queue);
 
 static struct nvmet_fabrics_ops nvmet_rdma_ops;
 
@@ -486,12 +487,33 @@ static void nvmet_rdma_release_rsp(struct nvmet_rdma_rsp *rsp)
 	nvmet_rdma_put_rsp(rsp);
 }
 
+static void nvmet_rdma_error_comp(struct nvmet_rdma_queue *queue)
+{
+	if (queue->nvme_sq.ctrl) {
+		nvmet_ctrl_fatal_error(queue->nvme_sq.ctrl);
+	} else {
+		/*
+		 * we didn't setup the controller yet in case
+		 * of admin connect error, just disconnect and
+		 * cleanup the queue
+		 */
+		nvmet_rdma_queue_disconnect(queue);
+	}
+}
+
 static void nvmet_rdma_send_done(struct ib_cq *cq, struct ib_wc *wc)
 {
 	struct nvmet_rdma_rsp *rsp =
 		container_of(wc->wr_cqe, struct nvmet_rdma_rsp, send_cqe);
 
 	nvmet_rdma_release_rsp(rsp);
+
+	if (unlikely(wc->status != IB_WC_SUCCESS &&
+		     wc->status != IB_WC_WR_FLUSH_ERR)) {
+		pr_err("SEND for CQE 0x%p failed with status %s (%d).\n",
+			wc->wr_cqe, ib_wc_status_msg(wc->status), wc->status);
+		nvmet_rdma_error_comp(rsp->queue);
+	}
 }
 
 static void nvmet_rdma_queue_response(struct nvmet_req *req)
@@ -534,11 +556,13 @@ static void nvmet_rdma_read_data_done(struct ib_cq *cq, struct ib_wc *wc)
 			rsp->req.sg_cnt, nvmet_data_dir(&rsp->req));
 	rsp->n_rdma = 0;
 
-	if (unlikely(wc->status != IB_WC_SUCCESS &&
-		wc->status != IB_WC_WR_FLUSH_ERR)) {
-		pr_info("RDMA READ for CQE 0x%p failed with status %s (%d).\n",
-			wc->wr_cqe, ib_wc_status_msg(wc->status), wc->status);
-		nvmet_req_complete(&rsp->req, NVME_SC_DATA_XFER_ERROR);
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		nvmet_rdma_release_rsp(rsp);
+		if (wc->status != IB_WC_WR_FLUSH_ERR) {
+			pr_info("RDMA READ for CQE 0x%p failed with status %s (%d).\n",
+				wc->wr_cqe, ib_wc_status_msg(wc->status), wc->status);
+			nvmet_rdma_error_comp(queue);
+		}
 		return;
 	}
 
@@ -705,13 +729,19 @@ static void nvmet_rdma_recv_done(struct ib_cq *cq, struct ib_wc *wc)
 	struct nvmet_rdma_queue *queue = cq->cq_context;
 	struct nvmet_rdma_rsp *rsp;
 
-	if (unlikely(wc->status != IB_WC_SUCCESS))
+	if (unlikely(wc->status != IB_WC_SUCCESS)) {
+		if (wc->status != IB_WC_WR_FLUSH_ERR) {
+			pr_err("RECV for CQE 0x%p failed with status %s (%d)\n",
+				wc->wr_cqe, ib_wc_status_msg(wc->status),
+				wc->status);
+			nvmet_rdma_error_comp(queue);
+		}
 		return;
+	}
 
 	if (unlikely(wc->byte_len < sizeof(struct nvme_command))) {
 		pr_err("Ctrl Fatal Error: capsule size less than 64 bytes\n");
-		if (queue->nvme_sq.ctrl)
-			nvmet_ctrl_fatal_error(queue->nvme_sq.ctrl);
+		nvmet_rdma_error_comp(queue);
 		return;
 	}
 
-- 
1.9.1

[PATCH v1] nvmet-rdma: Invoke fatal error on error completion

[Christoph Hellwig]

Thanks,

applied to nvmet-all.3.