Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Johannes Thumshirn <jthumshirn@suse.de>
To:     Christoph Hellwig <hch@lst.de>, Sagi Grimberg <sagi@grimberg.me>,
        Keith Busch <keith.busch@intel.com>
Cc:     James Smart <james.smart@broadcom.com>,
        Hannes Reinecke <hare@suse.de>, Ewan Milne <emilne@redhat.com>,
        Max Gurtovoy <maxg@mellanox.com>,
        Linux NVMe Mailinglist <linux-nvme@lists.infradead.org>,
        Linux Kernel Mailinglist <linux-kernel@vger.kernel.org>,
        Johannes Thumshirn <jthumshirn@suse.de>
Subject: [PATCH 3/4] nvmet: loop: abort commands before resetting controller
Date:   Thu, 19 Jul 2018 15:28:37 +0200
Message-Id: <20180719132838.15556-4-jthumshirn@suse.de>
In-Reply-To: <20180719132838.15556-1-jthumshirn@suse.de>
References: <20180719132838.15556-1-jthumshirn@suse.de>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Currently if a timeout on loop happens we just go ahead and reset the
underlying transport.

Instead try to abort the timeout out command and if this fails as well
continue the error path escalation and tear down the transport.

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
---
 drivers/nvme/target/loop.c | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/drivers/nvme/target/loop.c b/drivers/nvme/target/loop.c
index af7fbf4132b0..b6c355125cb0 100644
--- a/drivers/nvme/target/loop.c
+++ b/drivers/nvme/target/loop.c
@@ -141,9 +141,24 @@ static enum blk_eh_timer_return
 nvme_loop_timeout(struct request *rq, bool reserved)
 {
 	struct nvme_loop_iod *iod = blk_mq_rq_to_pdu(rq);
+	struct nvme_ctrl *ctrl = &iod->queue->ctrl->ctrl;
+	int qid = nvme_loop_queue_idx(iod->queue);
+	int ret;
+
+	dev_warn(ctrl->device,
+		 "I/O %d QID %d timeout, aborting\n",
+		 rq->tag, qid);
+
+	ret = nvme_abort_cmd(ctrl, rq, cpu_to_le16(qid));
+	if (!ret)
+		return BLK_EH_RESET_TIMER;
+
+	dev_warn(ctrl->device,
+		 "I/O %d QID %d abort failed %d, reset controller\n",
+		 rq->tag, qid, ret);
 
 	/* queue error recovery */
-	nvme_reset_ctrl(&iod->queue->ctrl->ctrl);
+	nvme_reset_ctrl(ctrl);
 
 	/* fail with DNR on admin cmd timeout */
 	nvme_req(rq)->status = NVME_SC_ABORT_REQ | NVME_SC_DNR;
-- 
2.16.4