All of lore.kernel.org
 help / color / mirror / Atom feed
From: Maurizio Lombardi <mlombard@redhat.com>
To: kbusch@meta.com
Cc: hare@kernel.org, sagi@grimberg.me, emilne@redhat.com,
	jmeneghi@redhat.com, linux-nvme@lists.infradead.org
Subject: [PATCH 4/5] nvme-rdma: use the cancel command to perform an abort if target supports it
Date: Fri, 10 May 2024 18:30:25 +0200	[thread overview]
Message-ID: <20240510163026.786898-5-mlombard@redhat.com> (raw)
In-Reply-To: <20240510163026.786898-1-mlombard@redhat.com>

If available, use Cancel command to abort the command that timed
out instead of resetting the controller.

Signed-off-by: Maurizio Lombardi <mlombard@redhat.com>
---
 drivers/nvme/host/rdma.c | 28 ++++++++++++++++++++++++----
 1 file changed, 24 insertions(+), 4 deletions(-)

diff --git a/drivers/nvme/host/rdma.c b/drivers/nvme/host/rdma.c
index 366f0bb4ebfc..a45b45965b9e 100644
--- a/drivers/nvme/host/rdma.c
+++ b/drivers/nvme/host/rdma.c
@@ -74,6 +74,7 @@ struct nvme_rdma_request {
 	struct nvme_rdma_sgl	data_sgl;
 	struct nvme_rdma_sgl	*metadata_sgl;
 	bool			use_sig_mr;
+	bool			aborted;
 };
 
 enum nvme_rdma_queue_flags {
@@ -1956,16 +1957,19 @@ static enum blk_eh_timer_return nvme_rdma_timeout(struct request *rq)
 {
 	struct nvme_rdma_request *req = blk_mq_rq_to_pdu(rq);
 	struct nvme_rdma_queue *queue = req->queue;
-	struct nvme_rdma_ctrl *ctrl = queue->ctrl;
+	struct nvme_rdma_ctrl *rdma_ctrl = queue->ctrl;
+	struct nvme_ctrl *ctrl = &rdma_ctrl->ctrl;
 	struct nvme_command *cmd = req->req.cmd;
 	int qid = nvme_rdma_queue_idx(queue);
+	int error;
+	u32 effects;
 
-	dev_warn(ctrl->ctrl.device,
+	dev_warn(ctrl->device,
 		 "I/O tag %d (%04x) opcode %#x (%s) QID %d timeout\n",
 		 rq->tag, nvme_cid(rq), cmd->common.opcode,
 		 nvme_fabrics_opcode_str(qid, cmd), qid);
 
-	if (nvme_ctrl_state(&ctrl->ctrl) != NVME_CTRL_LIVE) {
+	if (nvme_ctrl_state(ctrl) != NVME_CTRL_LIVE) {
 		/*
 		 * If we are resetting, connecting or deleting we should
 		 * complete immediately because we may block controller
@@ -1983,11 +1987,26 @@ static enum blk_eh_timer_return nvme_rdma_timeout(struct request *rq)
 		return BLK_EH_DONE;
 	}
 
+	if (!req->aborted) {
+		effects = le32_to_cpu(ctrl->effects->iocs[nvme_cmd_cancel]);
+
+		if (!(effects & NVME_CMD_EFFECTS_CSUPP))
+			goto err_recovery;
+
+		error = nvme_submit_cancel_req(ctrl, rq, qid);
+		if (error)
+			goto err_recovery;
+
+		req->aborted = true;
+		return BLK_EH_RESET_TIMER;
+	}
+
 	/*
 	 * LIVE state should trigger the normal error recovery which will
 	 * handle completing this request.
 	 */
-	nvme_rdma_error_recovery(ctrl);
+err_recovery:
+	nvme_rdma_error_recovery(rdma_ctrl);
 	return BLK_EH_RESET_TIMER;
 }
 
@@ -2011,6 +2030,7 @@ static blk_status_t nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx,
 		return nvme_fail_nonready_command(&queue->ctrl->ctrl, rq);
 
 	dev = queue->device->dev;
+	req->aborted = false;
 
 	req->sqe.dma = ib_dma_map_single(dev, req->sqe.data,
 					 sizeof(struct nvme_command),
-- 
2.39.3



  parent reply	other threads:[~2024-05-10 16:30 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-05-10 16:30 [PATCH 0/5] Support for Cancel commands for host (TCP and RDMA) Maurizio Lombardi
2024-05-10 16:30 ` [PATCH 1/5] nvme: add definitions for cancel command Maurizio Lombardi
2024-05-10 16:30 ` [PATCH 2/5] nvme-core: add a function to submit a " Maurizio Lombardi
2024-05-12 14:06   ` Sagi Grimberg
2024-06-20  8:31     ` Maurizio Lombardi
2024-06-20 13:01       ` Maurizio Lombardi
2024-06-24 10:06         ` Sagi Grimberg
2024-06-26 18:38           ` [PATCH v2 7/7] nvme: add reserved ioq tags for cancel John Meneghini
2024-06-26 19:03             ` Keith Busch
2024-06-26 19:20               ` John Meneghini
2024-06-27  8:12                 ` Maurizio Lombardi
2024-06-26 19:10             ` John Meneghini
2024-05-10 16:30 ` [PATCH 3/5] nvme-tcp: use the cancel command to perform an abort if target supports it Maurizio Lombardi
2024-05-10 16:30 ` Maurizio Lombardi [this message]
2024-05-10 16:30 ` [donotmerge PATCH 5/5] nvmet: target support for cancel commands Maurizio Lombardi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20240510163026.786898-5-mlombard@redhat.com \
    --to=mlombard@redhat.com \
    --cc=emilne@redhat.com \
    --cc=hare@kernel.org \
    --cc=jmeneghi@redhat.com \
    --cc=kbusch@meta.com \
    --cc=linux-nvme@lists.infradead.org \
    --cc=sagi@grimberg.me \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.