[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <018346fd-3a7c-4cb2-8456-61a03ee435b3@linux.dev>
Date: Thu, 19 Jun 2025 14:57:59 +0100
From: Vadim Fedorenko <vadim.fedorenko@...ux.dev>
To: Vikas Gupta <vikas.gupta@...adcom.com>, davem@...emloft.net,
edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
andrew+netdev@...n.ch, horms@...nel.org
Cc: netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
michael.chan@...adcom.com, pavan.chebbi@...adcom.com,
vsrama-krishna.nemani@...adcom.com,
Bhargava Chenna Marreddy <bhargava.marreddy@...adcom.com>,
Rajashekar Hudumula <rajashekar.hudumula@...adcom.com>
Subject: Re: [net-next, 09/10] bng_en: Initialize default configuration
On 18/06/2025 15:47, Vikas Gupta wrote:
> Query resources from the firmware and, based on the
> availability of resources, initialize the default
> settings. The default settings include:
> 1. Rings and other resource reservations with the
> firmware. This ensures that resources are reserved
> before network and auxiliary devices are created.
> 2. Mapping the BAR, which helps access doorbells since
> its size is known after querying the firmware.
> 3. Retrieving the TCs and hardware CoS queue mappings.
>
> Signed-off-by: Vikas Gupta <vikas.gupta@...adcom.com>
> Reviewed-by: Bhargava Chenna Marreddy <bhargava.marreddy@...adcom.com>
> Reviewed-by: Rajashekar Hudumula <rajashekar.hudumula@...adcom.com>
> ---
[...]
> diff --git a/drivers/net/ethernet/broadcom/bnge/bnge_hwrm_lib.c b/drivers/net/ethernet/broadcom/bnge/bnge_hwrm_lib.c
> index 577ac5dcdd9b..ffdec285948e 100644
> --- a/drivers/net/ethernet/broadcom/bnge/bnge_hwrm_lib.c
> +++ b/drivers/net/ethernet/broadcom/bnge/bnge_hwrm_lib.c
> @@ -647,3 +647,57 @@ int bnge_hwrm_vnic_qcaps(struct bnge_dev *bd)
>
> return rc;
> }
> +
> +#define BNGE_CNPQ(q_profile) \
> + ((q_profile) == \
> + QUEUE_QPORTCFG_RESP_QUEUE_ID0_SERVICE_PROFILE_LOSSY_ROCE_CNP)
> +
> +int bnge_hwrm_queue_qportcfg(struct bnge_dev *bd)
> +{
> + struct hwrm_queue_qportcfg_output *resp;
> + struct hwrm_queue_qportcfg_input *req;
> + u8 i, j, *qptr;
> + bool no_rdma;
> + int rc = 0;
The initialization is not needed. The very next line re-assigns the
value of rc.
> +
> + rc = hwrm_req_init(bd, req, HWRM_QUEUE_QPORTCFG);
> + if (rc)
> + return rc;
> +
> + resp = hwrm_req_hold(bd, req);
> + rc = hwrm_req_send(bd, req);
> + if (rc)
> + goto qportcfg_exit;
> +
> + if (!resp->max_configurable_queues) {
> + rc = -EINVAL;
> + goto qportcfg_exit;
> + }
> + bd->max_tc = resp->max_configurable_queues;
> + bd->max_lltc = resp->max_configurable_lossless_queues;
> + if (bd->max_tc > BNGE_MAX_QUEUE)
> + bd->max_tc = BNGE_MAX_QUEUE;
> +
> + no_rdma = !bnge_is_roce_en(bd);
> + qptr = &resp->queue_id0;
> + for (i = 0, j = 0; i < bd->max_tc; i++) {
> + bd->q_info[j].queue_id = *qptr;
> + bd->q_ids[i] = *qptr++;
> + bd->q_info[j].queue_profile = *qptr++;
> + bd->tc_to_qidx[j] = j;
> + if (!BNGE_CNPQ(bd->q_info[j].queue_profile) || no_rdma)
> + j++;
> + }
> + bd->max_q = bd->max_tc;
> + bd->max_tc = max_t(u8, j, 1);
> +
> + if (resp->queue_cfg_info & QUEUE_QPORTCFG_RESP_QUEUE_CFG_INFO_ASYM_CFG)
> + bd->max_tc = 1;
> +
> + if (bd->max_lltc > bd->max_tc)
> + bd->max_lltc = bd->max_tc;
> +
> +qportcfg_exit:
> + hwrm_req_drop(bd, req);
> + return rc;
> +}
[...]
Powered by blists - more mailing lists