[PATCH] nvme-tcp: add basic support for the C2HTermReq PDU

[Sagi Grimberg]

On 14/02/2025 21:00, Maurizio Lombardi wrote:
> Previously, the NVMe/TCP host driver did not handle the C2HTermReq PDU,
> instead printing "unsupported pdu type (3)" when received. This patch adds
> support for processing the C2HTermReq PDU, allowing the driver
> to print the Fatal Error Status field.
>
> Example of output:
> nvme nvme4: Received C2HTermReq (FES = Invalid PDU Header Field)
>
> Signed-off-by: Maurizio Lombardi <mlombard at redhat.com>
> ---
>   drivers/nvme/host/tcp.c  | 37 +++++++++++++++++++++++++++++++++++++
>   include/linux/nvme-tcp.h |  2 ++
>   2 files changed, 39 insertions(+)
>
> diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c
> index 841238f38fdd..8f783185575d 100644
> --- a/drivers/nvme/host/tcp.c
> +++ b/drivers/nvme/host/tcp.c
> @@ -763,6 +763,40 @@ static int nvme_tcp_handle_r2t(struct nvme_tcp_queue *queue,
>   	return 0;
>   }
>   
> +static void nvme_tcp_handle_c2h_term(struct nvme_tcp_queue *queue,
> +		struct nvme_tcp_term_pdu *pdu)
> +{
> +	u16 fes;
> +	const char *msg;
> +	u32 plen = le32_to_cpu(pdu->hdr.plen);
> +
> +	static const char * const msg_table[] = {
> +		[NVME_TCP_FES_INVALID_PDU_HDR] = "Invalid PDU Header Field",
> +		[NVME_TCP_FES_PDU_SEQ_ERR] = "PDU Sequence Error",
> +		[NVME_TCP_FES_HDR_DIGEST_ERR] = "Header Digest Error",
> +		[NVME_TCP_FES_DATA_OUT_OF_RANGE] = "Data Transfer Out Of Range",
> +		[NVME_TCP_FES_R2T_LIMIT_EXCEEDED] = "R2T Limit Exceeded",
> +		[NVME_TCP_FES_UNSUPPORTED_PARAM] = "Unsupported Parameter",
> +	};
> +
> +	if (plen < NVME_TCP_MIN_C2HTERM_PLEN ||
> +	    plen > NVME_TCP_MAX_C2HTERM_PLEN) {
> +		dev_err(queue->ctrl->ctrl.device,
> +			"Received a malformed C2HTermReq PDU (plen = %u)\n",
> +			plen);
> +		return;
> +	}
> +
> +	fes = le16_to_cpu(pdu->fes);
> +	if (fes && fes < ARRAY_SIZE(msg_table))
> +		msg = msg_table[fes];
> +	else
> +		msg = "N/A";

msg = "unknown" ? we shouldn't suggest N/A because it should be very 
much applicable.

> +
> +	dev_err(queue->ctrl->ctrl.device,
> +		"Received C2HTermReq (FES = %s)\n", msg);
> +}
> +

AFAIR, the goal with C2HTerm was to also log the offending PDU itself. 
However I don't think
its really necessary looking back.

Can you test this also with header digest enabled?

Conditioned that the above passes just fine, plus one minor nit,
Reviewed-by: Sagi Grimberg <sagi at grimberg.me>