From: Sagi Grimberg <sagi@grimberg.me>
To: Keith Busch <kbusch@kernel.org>
Cc: linux-nvme@lists.infradead.org, hch@lst.de
Subject: Re: nvme tcp receive errors
Date: Wed, 31 Mar 2021 15:45:42 -0700 [thread overview]
Message-ID: <71bf89c6-4e45-e9d2-0c73-65428712dceb@grimberg.me> (raw)
In-Reply-To: <20210331222644.GA28381@redsun51.ssa.fujisawa.hgst.com>
>> What is the workload you are running? have an fio job file?
>> Is this I/O to a raw block device? or with fs or iosched?
>
> It's O_DIRECT to raw block device using libaio engine. No fs, page
> cache, or io scheduler are used.
I see.
>
> The fio job is generated by a script that cycles through various sizes,
> rw mixes, and io depth. It is not always consistent on which paricular
> set of parameters are running when the error message is observed,
> though. I can get more details if this will be helpful.
Try out a debug patch [1], and when this happens we can get some more
info on the request itself.
>> Also, I'm assuming that you are using Linux nvmet as the target
>> device?
>
> Not this time. The target is implemented in a hardware device.
Ha, cool...
[1]:
---
diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
index 69f59d2c5799..b218a41ac088 100644
--- a/drivers/nvme/host/tcp.c
+++ b/drivers/nvme/host/tcp.c
@@ -37,6 +37,14 @@ enum nvme_tcp_send_state {
NVME_TCP_SEND_DDGST,
};
+enum nvme_tcp_cmd_state {
+ NVME_TCP_CMD_QUEUED = 0,
+ NVME_TCP_CMD_SEND_PDU,
+ NVME_TCP_CMD_PENDING_DATA,
+ NVME_TCP_CMD_DATA_DONE,
+ NVME_TCP_CMD_DONE,
+};
+
struct nvme_tcp_request {
struct nvme_request req;
void *pdu;
@@ -56,6 +64,7 @@ struct nvme_tcp_request {
size_t offset;
size_t data_sent;
enum nvme_tcp_send_state state;
+ enum nvme_tcp_cmd_state cmd_state;
};
enum nvme_tcp_queue_flags {
@@ -482,6 +491,7 @@ static void nvme_tcp_error_recovery(struct nvme_ctrl
*ctrl)
static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
struct nvme_completion *cqe)
{
+ struct nvme_tcp_request *req;
struct request *rq;
rq = blk_mq_tag_to_rq(nvme_tcp_tagset(queue), cqe->command_id);
@@ -493,6 +503,8 @@ static int nvme_tcp_process_nvme_cqe(struct
nvme_tcp_queue *queue,
return -EINVAL;
}
+ req = blk_mq_rq_to_pdu(rq);
+ req->cmd_state = NVME_TCP_CMD_DONE;
if (!nvme_try_complete_req(rq, cqe->status, cqe->result))
nvme_complete_rq(rq);
queue->nr_cqe++;
@@ -503,6 +515,7 @@ static int nvme_tcp_process_nvme_cqe(struct
nvme_tcp_queue *queue,
static int nvme_tcp_handle_c2h_data(struct nvme_tcp_queue *queue,
struct nvme_tcp_data_pdu *pdu)
{
+ struct nvme_tcp_request *req;
struct request *rq;
rq = blk_mq_tag_to_rq(nvme_tcp_tagset(queue), pdu->command_id);
@@ -512,11 +525,12 @@ static int nvme_tcp_handle_c2h_data(struct
nvme_tcp_queue *queue,
nvme_tcp_queue_id(queue), pdu->command_id);
return -ENOENT;
}
+ req = blk_mq_rq_to_pdu(rq);
if (!blk_rq_payload_bytes(rq)) {
dev_err(queue->ctrl->ctrl.device,
- "queue %d tag %#x unexpected data\n",
- nvme_tcp_queue_id(queue), rq->tag);
+ "queue %d tag %#x unexpected data cmd_state %d\n",
+ nvme_tcp_queue_id(queue), rq->tag, req->cmd_state);
return -EIO;
}
@@ -755,7 +769,9 @@ static int nvme_tcp_recv_data(struct nvme_tcp_queue
*queue, struct sk_buff *skb,
nvme_tcp_ddgst_final(queue->rcv_hash,
&queue->exp_ddgst);
queue->ddgst_remaining = NVME_TCP_DIGEST_LENGTH;
} else {
+ req->cmd_state = NVME_TCP_CMD_DATA_DONE;
if (pdu->hdr.flags & NVME_TCP_F_DATA_SUCCESS) {
+ req->cmd_state = NVME_TCP_CMD_DONE;
nvme_tcp_end_request(rq, NVME_SC_SUCCESS);
queue->nr_cqe++;
}
@@ -796,7 +812,10 @@ static int nvme_tcp_recv_ddgst(struct
nvme_tcp_queue *queue,
if (pdu->hdr.flags & NVME_TCP_F_DATA_SUCCESS) {
struct request *rq =
blk_mq_tag_to_rq(nvme_tcp_tagset(queue),
pdu->command_id);
+ struct nvme_tcp_request *req;
+ req = blk_mq_rq_to_pdu(rq);
+ req->cmd_state = NVME_TCP_CMD_DONE;
nvme_tcp_end_request(rq, NVME_SC_SUCCESS);
queue->nr_cqe++;
}
@@ -944,6 +963,7 @@ static int nvme_tcp_try_send_data(struct
nvme_tcp_request *req)
nvme_tcp_ddgst_final(queue->snd_hash,
&req->ddgst);
req->state = NVME_TCP_SEND_DDGST;
+ req->cmd_state = NVME_TCP_CMD_DATA_DONE;
req->offset = 0;
} else {
nvme_tcp_done_send_req(queue);
@@ -979,6 +999,7 @@ static int nvme_tcp_try_send_cmd_pdu(struct
nvme_tcp_request *req)
len -= ret;
if (!len) {
+ req->cmd_state = req->data_len ?
NVME_TCP_CMD_PENDING_DATA : NVME_TCP_CMD_DATA_DONE;
if (inline_data) {
req->state = NVME_TCP_SEND_DATA;
if (queue->data_digest)
@@ -2329,6 +2350,7 @@ static blk_status_t nvme_tcp_queue_rq(struct
blk_mq_hw_ctx *hctx,
blk_mq_start_request(rq);
+ req->cmd_state = NVME_TCP_CMD_QUEUED;
nvme_tcp_queue_request(req, true, bd->last);
return BLK_STS_OK;
---
_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
next prev parent reply other threads:[~2021-03-31 22:46 UTC|newest]
Thread overview: 50+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-31 16:18 nvme tcp receive errors Keith Busch
2021-03-31 19:10 ` Sagi Grimberg
2021-03-31 20:49 ` Keith Busch
2021-03-31 22:16 ` Sagi Grimberg
2021-03-31 22:26 ` Keith Busch
2021-03-31 22:45 ` Sagi Grimberg [this message]
2021-04-02 17:11 ` Keith Busch
2021-04-02 17:27 ` Sagi Grimberg
2021-04-05 14:37 ` Keith Busch
2021-04-07 19:53 ` Keith Busch
2021-04-09 21:38 ` Sagi Grimberg
2021-04-27 23:39 ` Keith Busch
2021-04-27 23:55 ` Sagi Grimberg
2021-04-28 15:58 ` Keith Busch
2021-04-28 17:42 ` Sagi Grimberg
2021-04-28 18:01 ` Keith Busch
2021-04-28 23:06 ` Sagi Grimberg
2021-04-29 3:33 ` Keith Busch
2021-04-29 4:52 ` Sagi Grimberg
2021-05-03 18:51 ` Keith Busch
2021-05-03 19:58 ` Sagi Grimberg
2021-05-03 20:25 ` Keith Busch
2021-05-04 19:29 ` Sagi Grimberg
2021-04-09 18:04 ` Sagi Grimberg
2021-04-14 0:29 ` Keith Busch
2021-04-21 5:33 ` Sagi Grimberg
2021-04-21 14:28 ` Keith Busch
2021-04-21 16:59 ` Sagi Grimberg
2021-04-26 15:31 ` Keith Busch
2021-04-27 3:10 ` Sagi Grimberg
2021-04-27 18:12 ` Keith Busch
2021-04-27 23:58 ` Sagi Grimberg
2021-04-30 23:42 ` Sagi Grimberg
2021-05-03 14:28 ` Keith Busch
2021-05-03 19:36 ` Sagi Grimberg
2021-05-03 19:38 ` Sagi Grimberg
2021-05-03 19:44 ` Keith Busch
2021-05-03 20:00 ` Sagi Grimberg
2021-05-04 14:36 ` Keith Busch
2021-05-04 18:15 ` Sagi Grimberg
2021-05-04 19:14 ` Keith Busch
2021-05-10 18:06 ` Keith Busch
2021-05-10 18:18 ` Sagi Grimberg
2021-05-10 18:30 ` Keith Busch
2021-05-10 21:07 ` Sagi Grimberg
2021-05-11 3:00 ` Keith Busch
2021-05-11 17:17 ` Sagi Grimberg
2021-05-13 15:48 ` Keith Busch
2021-05-13 19:53 ` Sagi Grimberg
2021-05-17 20:48 ` Keith Busch
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=71bf89c6-4e45-e9d2-0c73-65428712dceb@grimberg.me \
--to=sagi@grimberg.me \
--cc=hch@lst.de \
--cc=kbusch@kernel.org \
--cc=linux-nvme@lists.infradead.org \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).