From: Samik Gupta <samik.gupta@broadcom.com> This commit introduces a limit on the number of RX rings included in the RSS redirection table to a value no larger than the size supported by Thor as defined by BNXT_RSS_TBL_SIZE_THOR. Fixes: d819382543f3 ("net/bnxt: add RSS redirection table operations") Cc: stable@dpdk.org Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Samik Gupta <samik.gupta@broadcom.com> --- drivers/net/bnxt/bnxt.h | 2 +- drivers/net/bnxt/bnxt_ethdev.c | 15 +++++++++++++-- 2 files changed, 14 insertions(+), 3 deletions(-) diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h index 90ced972c..9bd4f1da9 100644 --- a/drivers/net/bnxt/bnxt.h +++ b/drivers/net/bnxt/bnxt.h @@ -389,7 +389,7 @@ struct bnxt_coal { #define DBR_TYPE_NQ (0xaULL << 60) #define DBR_TYPE_NQ_ARM (0xbULL << 60) -#define BNXT_RSS_TBL_SIZE_THOR 512 +#define BNXT_RSS_TBL_SIZE_THOR 512U #define BNXT_RSS_ENTRIES_PER_CTX_THOR 64 #define BNXT_MAX_RSS_CTXTS_THOR \ (BNXT_RSS_TBL_SIZE_THOR / BNXT_RSS_ENTRIES_PER_CTX_THOR) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 81c8f8d79..c363c8427 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -207,12 +207,15 @@ int is_bnxt_in_error(struct bnxt *bp) static uint16_t bnxt_rss_ctxts(const struct bnxt *bp) { + unsigned int num_rss_rings = RTE_MIN(bp->rx_nr_rings, + BNXT_RSS_TBL_SIZE_THOR); + if (!BNXT_CHIP_THOR(bp)) return 1; - return RTE_ALIGN_MUL_CEIL(bp->rx_nr_rings, + return RTE_ALIGN_MUL_CEIL(num_rss_rings, BNXT_RSS_ENTRIES_PER_CTX_THOR) / - BNXT_RSS_ENTRIES_PER_CTX_THOR; + BNXT_RSS_ENTRIES_PER_CTX_THOR; } uint16_t bnxt_rss_hash_tbl_size(const struct bnxt *bp) @@ -424,6 +427,14 @@ static int bnxt_setup_one_vnic(struct bnxt *bp, uint16_t vnic_id) if (dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS) { int j, nr_ctxs = bnxt_rss_ctxts(bp); + if (bp->rx_nr_rings > BNXT_RSS_TBL_SIZE_THOR) { + PMD_DRV_LOG(ERR, "RxQ cnt %d > reta_size %d\n", + bp->rx_nr_rings, BNXT_RSS_TBL_SIZE_THOR); + PMD_DRV_LOG(ERR, + "Only queues 0-%d will be in RSS table\n", + BNXT_RSS_TBL_SIZE_THOR - 1); + } + rc = 0; for (j = 0; j < nr_ctxs; j++) { rc = bnxt_hwrm_vnic_ctx_alloc(bp, vnic, j); -- 2.21.1 (Apple Git-122.3)
From: Samik Gupta <samik.gupta@broadcom.com> This commit reconfigures a vnic's default ring if the current default ring is stopped by the application. It picks the lowest numbered ring that is currently active to be the new default, and issues the hwrm_vnic_cfg command to update the configuration. Applies to adapters that are not Thor-based. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Signed-off-by: Samik Gupta <samik.gupta@broadcom.com> --- drivers/net/bnxt/bnxt_rxq.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/drivers/net/bnxt/bnxt_rxq.c b/drivers/net/bnxt/bnxt_rxq.c index e0ec34216..61196eba9 100644 --- a/drivers/net/bnxt/bnxt_rxq.c +++ b/drivers/net/bnxt/bnxt_rxq.c @@ -557,12 +557,12 @@ int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id) rc = bnxt_vnic_rss_configure(bp, vnic); } - if (BNXT_CHIP_THOR(bp)) { - /* Compute current number of active receive queues. */ - for (i = vnic->start_grp_id; i < vnic->end_grp_id; i++) - if (bp->rx_queues[i]->rx_started) - active_queue_cnt++; + /* Compute current number of active receive queues. */ + for (i = vnic->start_grp_id; i < vnic->end_grp_id; i++) + if (bp->rx_queues[i]->rx_started) + active_queue_cnt++; + if (BNXT_CHIP_THOR(bp)) { /* * For Thor, we need to ensure that the VNIC default receive * ring corresponds to an active receive queue. When no queue @@ -582,6 +582,22 @@ int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id) /* Reconfigure default receive ring. */ bnxt_hwrm_vnic_cfg(bp, vnic); } + } else if (active_queue_cnt) { + /* + * If the queue being stopped is the current default queue and + * there are other active queues, pick one of them as the + * default and reconfigure the vnic. + */ + if (vnic->dflt_ring_grp == bp->grp_info[rx_queue_id].fw_grp_id) { + for (i = vnic->start_grp_id; i < vnic->end_grp_id; i++) { + if (bp->rx_queues[i]->rx_started) { + vnic->dflt_ring_grp = + bp->grp_info[i].fw_grp_id; + bnxt_hwrm_vnic_cfg(bp, vnic); + break; + } + } + } } if (rc == 0) -- 2.21.1 (Apple Git-122.3)
From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> In __bnxt_hwrm_func_qcaps, when memory allocations fails driver is not releasing the hwrm lock. This patch fixes it by calling hwrm_unlock in that error case. Fixes: b7778e8a1c00 ("net/bnxt: refactor to properly allocate resources for PF/VF") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_hwrm.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c index ebbf504c0..784e9778a 100644 --- a/drivers/net/bnxt/bnxt_hwrm.c +++ b/drivers/net/bnxt/bnxt_hwrm.c @@ -718,6 +718,7 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp) sizeof(bp->pf->vf_info[0]) * new_max_vfs, 0); if (bp->pf->vf_info == NULL) { PMD_DRV_LOG(ERR, "Alloc vf info fail\n"); + HWRM_UNLOCK(); return -ENOMEM; } bp->pf->max_vfs = new_max_vfs; -- 2.21.1 (Apple Git-122.3)
From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> In bnxt_reta_update_op() and bnxt_rss_hash_update_op(), driver does not propagate the error back to the application when the fw command fails. Fixes: 378ab645bb0b ("net/bnxt: fix RSS RETA indirection table ops") Fixes: fcc0aa1edc10 ("net/bnxt: add RSS hash configuration") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_ethdev.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 8047b0b5d..33358779b 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -1843,8 +1843,8 @@ static int bnxt_reta_update_op(struct rte_eth_dev *eth_dev, } } - bnxt_hwrm_vnic_rss_cfg(bp, vnic); - return 0; + rc = bnxt_hwrm_vnic_rss_cfg(bp, vnic); + return rc; } static int bnxt_reta_query_op(struct rte_eth_dev *eth_dev, @@ -1949,8 +1949,8 @@ static int bnxt_rss_hash_update_op(struct rte_eth_dev *eth_dev, memcpy(vnic->rss_hash_key, rss_conf->rss_key, rss_conf->rss_key_len); rss_config: - bnxt_hwrm_vnic_rss_cfg(bp, vnic); - return 0; + rc = bnxt_hwrm_vnic_rss_cfg(bp, vnic); + return rc; } static int bnxt_rss_hash_conf_get_op(struct rte_eth_dev *eth_dev, -- 2.21.1 (Apple Git-122.3)
In case mutex init fails during initialization, start cleanup and fail the initialization process. Fixes: a73b8e939f10 ("net/bnxt: fix race between start and interrupt handler") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_ethdev.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 33358779b..0b14ca234 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -4739,8 +4739,10 @@ bnxt_init_locks(struct bnxt *bp) } err = pthread_mutex_init(&bp->def_cp_lock, NULL); - if (err) + if (err) { PMD_DRV_LOG(ERR, "Unable to initialize def_cp_lock\n"); + return err; + } err = pthread_mutex_init(&bp->health_check_lock, NULL); if (err) -- 2.21.1 (Apple Git-122.3)
&device requires the %u format specifer not the %d specifier, as &device is unsigned. Fixes: a46bbb57605b ("net/bnxt: update multi device design") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/tf_core/tf_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/bnxt/tf_core/tf_core.c b/drivers/net/bnxt/tf_core/tf_core.c index 24d49096a..3409cbbce 100644 --- a/drivers/net/bnxt/tf_core/tf_core.c +++ b/drivers/net/bnxt/tf_core/tf_core.c @@ -82,7 +82,7 @@ tf_open_session(struct tf *tfp, return rc; TFP_DRV_LOG(INFO, - "domain:%d, bus:%d, device:%d\n", + "domain:%d, bus:%d, device:%u\n", parms->session_id.internal.domain, parms->session_id.internal.bus, parms->session_id.internal.device); -- 2.21.1 (Apple Git-122.3)
From: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com> The current max_rings computation does not take into account the case when max_nq_rings is <= num_async_cpr. This results in a wrong value like 0, when max_nq_rings is 1. Fix this by subtracting num_async_cpr only when max_cp_rings > num_async_cpr. Apart from this, the entire logic is currently spread across a few macros, making it hard to read and debug this code. Move this code into an inline function. max_msix is not used in the max_rings calculation. Apparently the max_msix field returned in HWRM_RESC_QCAPS is only valid for Thor and newer chips. On Wh+ it will be equal to min_compl_rings. Also, when a function reset is performed on an application quit, FW will not reset the VF resource pool as per design. This can lead to a strange condition wherein the max_msix field on Wh+ keeps changing on each application re-load thereby throwing throwing off the max_rings computation. Fixes: f03e66cb64ce ("net/bnxt: limit queue count for NS3/Stingray devices") Cc: stable@dpdk.org Signed-off-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com> Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt.h | 41 +++++++++++++++++++++++----------- drivers/net/bnxt/bnxt_ethdev.c | 2 +- drivers/net/bnxt/bnxt_rxq.c | 7 +++--- drivers/net/bnxt/bnxt_txq.c | 2 +- 4 files changed, 34 insertions(+), 18 deletions(-) diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h index 9c1c87489..bc0935272 100644 --- a/drivers/net/bnxt/bnxt.h +++ b/drivers/net/bnxt/bnxt.h @@ -752,19 +752,6 @@ struct bnxt { uint16_t max_tx_rings; uint16_t max_rx_rings; #define MAX_STINGRAY_RINGS 128U -/* For sake of symmetry, max Tx rings == max Rx rings, one stat ctx for each */ -#define BNXT_MAX_RX_RINGS(bp) \ - (BNXT_STINGRAY(bp) ? RTE_MIN(RTE_MIN(bp->max_rx_rings / 2U, \ - MAX_STINGRAY_RINGS), \ - bp->max_stat_ctx / 2U) : \ - RTE_MIN(bp->max_rx_rings / 2U, \ - bp->max_stat_ctx / 2U)) -#define BNXT_MAX_TX_RINGS(bp) \ - (RTE_MIN((bp)->max_tx_rings, BNXT_MAX_RX_RINGS(bp))) - -#define BNXT_MAX_RINGS(bp) \ - (RTE_MIN((((bp)->max_cp_rings - BNXT_NUM_ASYNC_CPR(bp)) / 2U), \ - BNXT_MAX_TX_RINGS(bp))) #define BNXT_MAX_VF_REP_RINGS 8 @@ -822,6 +809,34 @@ struct bnxt { uint16_t tx_cfa_action; }; +static +inline uint16_t bnxt_max_rings(struct bnxt *bp) +{ + uint16_t max_tx_rings = bp->max_tx_rings; + uint16_t max_rx_rings = bp->max_rx_rings; + uint16_t max_cp_rings = bp->max_cp_rings; + uint16_t max_rings; + + /* For the sake of symmetry: + * max Tx rings == max Rx rings, one stat ctx for each. + */ + if (BNXT_STINGRAY(bp)) { + max_rx_rings = RTE_MIN(RTE_MIN(max_rx_rings / 2U, + MAX_STINGRAY_RINGS), + bp->max_stat_ctx / 2U); + } else { + max_rx_rings = RTE_MIN(max_rx_rings / 2U, + bp->max_stat_ctx / 2U); + } + + max_tx_rings = RTE_MIN(max_tx_rings, max_rx_rings); + if (max_cp_rings > BNXT_NUM_ASYNC_CPR(bp)) + max_cp_rings -= BNXT_NUM_ASYNC_CPR(bp); + max_rings = RTE_MIN(max_cp_rings / 2U, max_tx_rings); + + return max_rings; +} + #define BNXT_FC_TIMER 1 /* Timer freq in Sec Flow Counters */ /** diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 0b14ca234..bf8963577 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -920,7 +920,7 @@ static int bnxt_dev_info_get_op(struct rte_eth_dev *eth_dev, if (BNXT_PF(bp)) dev_info->max_vfs = pdev->max_vfs; - max_rx_rings = BNXT_MAX_RINGS(bp); + max_rx_rings = bnxt_max_rings(bp); /* For the sake of symmetry, max_rx_queues = max_tx_queues */ dev_info->max_rx_queues = max_rx_rings; dev_info->max_tx_queues = max_rx_rings; diff --git a/drivers/net/bnxt/bnxt_rxq.c b/drivers/net/bnxt/bnxt_rxq.c index 328cc994d..19e11e47b 100644 --- a/drivers/net/bnxt/bnxt_rxq.c +++ b/drivers/net/bnxt/bnxt_rxq.c @@ -311,7 +311,7 @@ int bnxt_rx_queue_setup_op(struct rte_eth_dev *eth_dev, if (rc) return rc; - if (queue_idx >= BNXT_MAX_RINGS(bp)) { + if (queue_idx >= bnxt_max_rings(bp)) { PMD_DRV_LOG(ERR, "Cannot create Rx ring %d. Only %d rings available\n", queue_idx, bp->max_rx_rings); @@ -364,8 +364,9 @@ int bnxt_rx_queue_setup_op(struct rte_eth_dev *eth_dev, eth_dev->data->rx_queues[queue_idx] = rxq; /* Allocate RX ring hardware descriptors */ - if (bnxt_alloc_rings(bp, queue_idx, NULL, rxq, rxq->cp_ring, NULL, - "rxr")) { + rc = bnxt_alloc_rings(bp, queue_idx, NULL, rxq, rxq->cp_ring, NULL, + "rxr"); + if (rc) { PMD_DRV_LOG(ERR, "ring_dma_zone_reserve for rx_ring failed!\n"); goto err; diff --git a/drivers/net/bnxt/bnxt_txq.c b/drivers/net/bnxt/bnxt_txq.c index c9792a2af..99a31cef2 100644 --- a/drivers/net/bnxt/bnxt_txq.c +++ b/drivers/net/bnxt/bnxt_txq.c @@ -98,7 +98,7 @@ int bnxt_tx_queue_setup_op(struct rte_eth_dev *eth_dev, if (rc) return rc; - if (queue_idx >= BNXT_MAX_RINGS(bp)) { + if (queue_idx >= bnxt_max_rings(bp)) { PMD_DRV_LOG(ERR, "Cannot create Tx ring %d. Only %d rings available\n", queue_idx, bp->max_tx_rings); -- 2.21.1 (Apple Git-122.3)
From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> 1. moved invalid VNIC id check to the beginning of the function. 2. removed a duplicate check which avoids unnecessary code indentation. Fixes: 49d0709b257fc ("net/bnxt: delete and flush L2 filters cleanly") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_hwrm.c | 42 +++++++++++++++++------------------- 1 file changed, 20 insertions(+), 22 deletions(-) diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c index 6f5402070..cee2656c1 100644 --- a/drivers/net/bnxt/bnxt_hwrm.c +++ b/drivers/net/bnxt/bnxt_hwrm.c @@ -4896,37 +4896,35 @@ int bnxt_vnic_rss_configure(struct bnxt *bp, struct bnxt_vnic_info *vnic) { unsigned int rss_idx, fw_idx, i; + if (vnic->fw_vnic_id == INVALID_HW_RING_ID) + return 0; + if (!(vnic->rss_table && vnic->hash_type)) return 0; if (BNXT_CHIP_P5(bp)) return bnxt_vnic_rss_configure_p5(bp, vnic); - if (vnic->fw_vnic_id == INVALID_HW_RING_ID) - return 0; - - if (vnic->rss_table && vnic->hash_type) { - /* - * Fill the RSS hash & redirection table with - * ring group ids for all VNICs - */ - for (rss_idx = 0, fw_idx = 0; rss_idx < HW_HASH_INDEX_SIZE; - rss_idx++, fw_idx++) { - for (i = 0; i < bp->rx_cp_nr_rings; i++) { - fw_idx %= bp->rx_cp_nr_rings; - if (vnic->fw_grp_ids[fw_idx] != - INVALID_HW_RING_ID) - break; - fw_idx++; - } - if (i == bp->rx_cp_nr_rings) - return 0; - vnic->rss_table[rss_idx] = vnic->fw_grp_ids[fw_idx]; + /* + * Fill the RSS hash & redirection table with + * ring group ids for all VNICs + */ + for (rss_idx = 0, fw_idx = 0; rss_idx < HW_HASH_INDEX_SIZE; + rss_idx++, fw_idx++) { + for (i = 0; i < bp->rx_cp_nr_rings; i++) { + fw_idx %= bp->rx_cp_nr_rings; + if (vnic->fw_grp_ids[fw_idx] != INVALID_HW_RING_ID) + break; + fw_idx++; } - return bnxt_hwrm_vnic_rss_cfg(bp, vnic); + + if (i == bp->rx_cp_nr_rings) + return 0; + + vnic->rss_table[rss_idx] = vnic->fw_grp_ids[fw_idx]; } - return 0; + return bnxt_hwrm_vnic_rss_cfg(bp, vnic); } static void bnxt_hwrm_set_coal_params(struct bnxt_coal *hw_coal, -- 2.21.1 (Apple Git-122.3)
From: Somnath Kotur <somnath.kotur@broadcom.com> This cmd should be called by every driver after 'hwrm_func_cfg' to get the actual number of resources allocated by the HWRM. The values returned in the cmd are the max values for that PF. Also, now that the max values for the PF are computed in probe itself, no need to invoke FUNC_QCAPs or any other cmd in dev_configure_op() as that would just override the actual max values obtained above. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_ethdev.c | 7 ------- drivers/net/bnxt/bnxt_hwrm.c | 36 +++++++++++++++++++++++++++++++++- 2 files changed, 35 insertions(+), 8 deletions(-) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index bf8963577..21d9e26ca 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -1071,13 +1071,6 @@ static int bnxt_dev_configure_op(struct rte_eth_dev *eth_dev) } pthread_mutex_unlock(&bp->def_cp_lock); - } else { - /* legacy driver needs to get updated values */ - rc = bnxt_hwrm_func_qcaps(bp); - if (rc) { - PMD_DRV_LOG(ERR, "hwrm func qcaps fail:%d\n", rc); - return rc; - } } /* Inherit new configurations */ diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c index cee2656c1..a3eb133c0 100644 --- a/drivers/net/bnxt/bnxt_hwrm.c +++ b/drivers/net/bnxt/bnxt_hwrm.c @@ -3456,6 +3456,35 @@ static int bnxt_update_max_resources(struct bnxt *bp, return 0; } +/* Update the PF resource values based on how many resources + * got allocated to it. + */ +static int bnxt_update_max_resources_pf_only(struct bnxt *bp) +{ + struct hwrm_func_qcfg_input req = {0}; + struct hwrm_func_qcfg_output *resp = bp->hwrm_cmd_resp_addr; + int rc; + + /* Get the actual allocated values now */ + HWRM_PREP(&req, HWRM_FUNC_QCFG, BNXT_USE_CHIMP_MB); + req.fid = rte_cpu_to_le_16(0xffff); + rc = bnxt_hwrm_send_message(bp, &req, sizeof(req), BNXT_USE_CHIMP_MB); + HWRM_CHECK_RESULT(); + + bp->max_rsscos_ctx = rte_le_to_cpu_16(resp->alloc_rsscos_ctx); + bp->max_stat_ctx = rte_le_to_cpu_16(resp->alloc_stat_ctx); + bp->max_cp_rings = rte_le_to_cpu_16(resp->alloc_cmpl_rings); + bp->max_tx_rings = rte_le_to_cpu_16(resp->alloc_tx_rings); + bp->max_rx_rings = rte_le_to_cpu_16(resp->alloc_rx_rings); + bp->max_l2_ctx = rte_le_to_cpu_16(resp->alloc_l2_ctx); + bp->max_ring_grps = rte_le_to_cpu_16(resp->alloc_hw_ring_grps); + bp->max_vnics = rte_le_to_cpu_16(resp->alloc_vnics); + + HWRM_UNLOCK(); + + return 0; +} + int bnxt_hwrm_func_qcfg_current_vf_vlan(struct bnxt *bp, int vf) { struct hwrm_func_qcfg_input req = {0}; @@ -3555,8 +3584,13 @@ int bnxt_hwrm_allocate_pf_only(struct bnxt *bp) HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE); bp->pf->func_cfg_flags |= HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE; + rc = bnxt_hwrm_pf_func_cfg(bp, &pf_resc); - rc = __bnxt_hwrm_func_qcaps(bp); + if (rc) + return rc; + + rc = bnxt_update_max_resources_pf_only(bp); + return rc; } -- 2.21.1 (Apple Git-122.3)
From: Samik Gupta <samik.gupta@broadcom.com> This commit introduces a limit on the number of RX rings included in the RSS redirection table to a value no larger than the size supported by Thor as defined by BNXT_RSS_TBL_SIZE_THOR. Fixes: d819382543f3 ("net/bnxt: add RSS redirection table operations") Cc: stable@dpdk.org Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Signed-off-by: Samik Gupta <samik.gupta@broadcom.com> --- drivers/net/bnxt/bnxt.h | 2 +- drivers/net/bnxt/bnxt_ethdev.c | 15 +++++++++++++-- 2 files changed, 14 insertions(+), 3 deletions(-) diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h index 90ced972c..9bd4f1da9 100644 --- a/drivers/net/bnxt/bnxt.h +++ b/drivers/net/bnxt/bnxt.h @@ -389,7 +389,7 @@ struct bnxt_coal { #define DBR_TYPE_NQ (0xaULL << 60) #define DBR_TYPE_NQ_ARM (0xbULL << 60) -#define BNXT_RSS_TBL_SIZE_THOR 512 +#define BNXT_RSS_TBL_SIZE_THOR 512U #define BNXT_RSS_ENTRIES_PER_CTX_THOR 64 #define BNXT_MAX_RSS_CTXTS_THOR \ (BNXT_RSS_TBL_SIZE_THOR / BNXT_RSS_ENTRIES_PER_CTX_THOR) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 81c8f8d79..c363c8427 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -207,12 +207,15 @@ int is_bnxt_in_error(struct bnxt *bp) static uint16_t bnxt_rss_ctxts(const struct bnxt *bp) { + unsigned int num_rss_rings = RTE_MIN(bp->rx_nr_rings, + BNXT_RSS_TBL_SIZE_THOR); + if (!BNXT_CHIP_THOR(bp)) return 1; - return RTE_ALIGN_MUL_CEIL(bp->rx_nr_rings, + return RTE_ALIGN_MUL_CEIL(num_rss_rings, BNXT_RSS_ENTRIES_PER_CTX_THOR) / - BNXT_RSS_ENTRIES_PER_CTX_THOR; + BNXT_RSS_ENTRIES_PER_CTX_THOR; } uint16_t bnxt_rss_hash_tbl_size(const struct bnxt *bp) @@ -424,6 +427,14 @@ static int bnxt_setup_one_vnic(struct bnxt *bp, uint16_t vnic_id) if (dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS) { int j, nr_ctxs = bnxt_rss_ctxts(bp); + if (bp->rx_nr_rings > BNXT_RSS_TBL_SIZE_THOR) { + PMD_DRV_LOG(ERR, "RxQ cnt %d > reta_size %d\n", + bp->rx_nr_rings, BNXT_RSS_TBL_SIZE_THOR); + PMD_DRV_LOG(ERR, + "Only queues 0-%d will be in RSS table\n", + BNXT_RSS_TBL_SIZE_THOR - 1); + } + rc = 0; for (j = 0; j < nr_ctxs; j++) { rc = bnxt_hwrm_vnic_ctx_alloc(bp, vnic, j); -- 2.21.1 (Apple Git-122.3)
From: Samik Gupta <samik.gupta@broadcom.com> This commit reconfigures a vnic's default ring if the current default ring is stopped by the application. It picks the lowest numbered ring that is currently active to be the new default, and issues the hwrm_vnic_cfg command to update the configuration. Applies to adapters that are not Thor-based. Fixes: 9b63c6fd70e3 ("net/bnxt: support Rx/Tx queue start/stop") Cc: stable@dpdk.org Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> Reviewed-by: Lance Richardson <lance.richardson@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Signed-off-by: Samik Gupta <samik.gupta@broadcom.com> --- drivers/net/bnxt/bnxt_rxq.c | 26 +++++++++++++++++++++----- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/drivers/net/bnxt/bnxt_rxq.c b/drivers/net/bnxt/bnxt_rxq.c index e0ec34216..61196eba9 100644 --- a/drivers/net/bnxt/bnxt_rxq.c +++ b/drivers/net/bnxt/bnxt_rxq.c @@ -557,12 +557,12 @@ int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id) rc = bnxt_vnic_rss_configure(bp, vnic); } - if (BNXT_CHIP_THOR(bp)) { - /* Compute current number of active receive queues. */ - for (i = vnic->start_grp_id; i < vnic->end_grp_id; i++) - if (bp->rx_queues[i]->rx_started) - active_queue_cnt++; + /* Compute current number of active receive queues. */ + for (i = vnic->start_grp_id; i < vnic->end_grp_id; i++) + if (bp->rx_queues[i]->rx_started) + active_queue_cnt++; + if (BNXT_CHIP_THOR(bp)) { /* * For Thor, we need to ensure that the VNIC default receive * ring corresponds to an active receive queue. When no queue @@ -582,6 +582,22 @@ int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id) /* Reconfigure default receive ring. */ bnxt_hwrm_vnic_cfg(bp, vnic); } + } else if (active_queue_cnt) { + /* + * If the queue being stopped is the current default queue and + * there are other active queues, pick one of them as the + * default and reconfigure the vnic. + */ + if (vnic->dflt_ring_grp == bp->grp_info[rx_queue_id].fw_grp_id) { + for (i = vnic->start_grp_id; i < vnic->end_grp_id; i++) { + if (bp->rx_queues[i]->rx_started) { + vnic->dflt_ring_grp = + bp->grp_info[i].fw_grp_id; + bnxt_hwrm_vnic_cfg(bp, vnic); + break; + } + } + } } if (rc == 0) -- 2.21.1 (Apple Git-122.3)
From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> In __bnxt_hwrm_func_qcaps, when memory allocations fails driver is not releasing the hwrm lock. This patch fixes it by calling hwrm_unlock in that error case. Fixes: b7778e8a1c00 ("net/bnxt: refactor to properly allocate resources for PF/VF") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_hwrm.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c index ebbf504c0..784e9778a 100644 --- a/drivers/net/bnxt/bnxt_hwrm.c +++ b/drivers/net/bnxt/bnxt_hwrm.c @@ -718,6 +718,7 @@ static int __bnxt_hwrm_func_qcaps(struct bnxt *bp) sizeof(bp->pf->vf_info[0]) * new_max_vfs, 0); if (bp->pf->vf_info == NULL) { PMD_DRV_LOG(ERR, "Alloc vf info fail\n"); + HWRM_UNLOCK(); return -ENOMEM; } bp->pf->max_vfs = new_max_vfs; -- 2.21.1 (Apple Git-122.3)
From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> In bnxt_reta_update_op() and bnxt_rss_hash_update_op(), driver does not propagate the error back to the application when the fw command fails. Fixes: 378ab645bb0b ("net/bnxt: fix RSS RETA indirection table ops") Fixes: fcc0aa1edc10 ("net/bnxt: add RSS hash configuration") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_ethdev.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 8047b0b5d..33358779b 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -1843,8 +1843,8 @@ static int bnxt_reta_update_op(struct rte_eth_dev *eth_dev, } } - bnxt_hwrm_vnic_rss_cfg(bp, vnic); - return 0; + rc = bnxt_hwrm_vnic_rss_cfg(bp, vnic); + return rc; } static int bnxt_reta_query_op(struct rte_eth_dev *eth_dev, @@ -1949,8 +1949,8 @@ static int bnxt_rss_hash_update_op(struct rte_eth_dev *eth_dev, memcpy(vnic->rss_hash_key, rss_conf->rss_key, rss_conf->rss_key_len); rss_config: - bnxt_hwrm_vnic_rss_cfg(bp, vnic); - return 0; + rc = bnxt_hwrm_vnic_rss_cfg(bp, vnic); + return rc; } static int bnxt_rss_hash_conf_get_op(struct rte_eth_dev *eth_dev, -- 2.21.1 (Apple Git-122.3)
In case mutex init fails during initialization, start cleanup and fail the initialization process. Fixes: a73b8e939f10 ("net/bnxt: fix race between start and interrupt handler") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_ethdev.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 33358779b..0b14ca234 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -4739,8 +4739,10 @@ bnxt_init_locks(struct bnxt *bp) } err = pthread_mutex_init(&bp->def_cp_lock, NULL); - if (err) + if (err) { PMD_DRV_LOG(ERR, "Unable to initialize def_cp_lock\n"); + return err; + } err = pthread_mutex_init(&bp->health_check_lock, NULL); if (err) -- 2.21.1 (Apple Git-122.3)
&device requires the %u format specifer not the %d specifier, as &device is unsigned. Fixes: a46bbb57605b ("net/bnxt: update multi device design") Cc: stable@dpdk.org Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/tf_core/tf_core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/bnxt/tf_core/tf_core.c b/drivers/net/bnxt/tf_core/tf_core.c index 24d49096a..3409cbbce 100644 --- a/drivers/net/bnxt/tf_core/tf_core.c +++ b/drivers/net/bnxt/tf_core/tf_core.c @@ -82,7 +82,7 @@ tf_open_session(struct tf *tfp, return rc; TFP_DRV_LOG(INFO, - "domain:%d, bus:%d, device:%d\n", + "domain:%d, bus:%d, device:%u\n", parms->session_id.internal.domain, parms->session_id.internal.bus, parms->session_id.internal.device); -- 2.21.1 (Apple Git-122.3)
From: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com> The current max_rings computation does not take into account the case when max_nq_rings is <= num_async_cpr. This results in a wrong value like 0, when max_nq_rings is 1. Fix this by subtracting num_async_cpr only when max_cp_rings > num_async_cpr. Apart from this, the entire logic is currently spread across a few macros, making it hard to read and debug this code. Move this code into an inline function. max_msix is not used in the max_rings calculation. Apparently the max_msix field returned in HWRM_RESC_QCAPS is only valid for Thor and newer chips. On Wh+ it will be equal to min_compl_rings. Also, when a function reset is performed on an application quit, FW will not reset the VF resource pool as per design. This can lead to a strange condition wherein the max_msix field on Wh+ keeps changing on each application re-load thereby throwing throwing off the max_rings computation. Fixes: f03e66cb64ce ("net/bnxt: limit queue count for NS3/Stingray devices") Cc: stable@dpdk.org Signed-off-by: Sriharsha Basavapatna <sriharsha.basavapatna@broadcom.com> Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt.h | 41 +++++++++++++++++++++++----------- drivers/net/bnxt/bnxt_ethdev.c | 2 +- drivers/net/bnxt/bnxt_rxq.c | 7 +++--- drivers/net/bnxt/bnxt_txq.c | 2 +- 4 files changed, 34 insertions(+), 18 deletions(-) diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h index 9c1c87489..bc0935272 100644 --- a/drivers/net/bnxt/bnxt.h +++ b/drivers/net/bnxt/bnxt.h @@ -752,19 +752,6 @@ struct bnxt { uint16_t max_tx_rings; uint16_t max_rx_rings; #define MAX_STINGRAY_RINGS 128U -/* For sake of symmetry, max Tx rings == max Rx rings, one stat ctx for each */ -#define BNXT_MAX_RX_RINGS(bp) \ - (BNXT_STINGRAY(bp) ? RTE_MIN(RTE_MIN(bp->max_rx_rings / 2U, \ - MAX_STINGRAY_RINGS), \ - bp->max_stat_ctx / 2U) : \ - RTE_MIN(bp->max_rx_rings / 2U, \ - bp->max_stat_ctx / 2U)) -#define BNXT_MAX_TX_RINGS(bp) \ - (RTE_MIN((bp)->max_tx_rings, BNXT_MAX_RX_RINGS(bp))) - -#define BNXT_MAX_RINGS(bp) \ - (RTE_MIN((((bp)->max_cp_rings - BNXT_NUM_ASYNC_CPR(bp)) / 2U), \ - BNXT_MAX_TX_RINGS(bp))) #define BNXT_MAX_VF_REP_RINGS 8 @@ -822,6 +809,34 @@ struct bnxt { uint16_t tx_cfa_action; }; +static +inline uint16_t bnxt_max_rings(struct bnxt *bp) +{ + uint16_t max_tx_rings = bp->max_tx_rings; + uint16_t max_rx_rings = bp->max_rx_rings; + uint16_t max_cp_rings = bp->max_cp_rings; + uint16_t max_rings; + + /* For the sake of symmetry: + * max Tx rings == max Rx rings, one stat ctx for each. + */ + if (BNXT_STINGRAY(bp)) { + max_rx_rings = RTE_MIN(RTE_MIN(max_rx_rings / 2U, + MAX_STINGRAY_RINGS), + bp->max_stat_ctx / 2U); + } else { + max_rx_rings = RTE_MIN(max_rx_rings / 2U, + bp->max_stat_ctx / 2U); + } + + max_tx_rings = RTE_MIN(max_tx_rings, max_rx_rings); + if (max_cp_rings > BNXT_NUM_ASYNC_CPR(bp)) + max_cp_rings -= BNXT_NUM_ASYNC_CPR(bp); + max_rings = RTE_MIN(max_cp_rings / 2U, max_tx_rings); + + return max_rings; +} + #define BNXT_FC_TIMER 1 /* Timer freq in Sec Flow Counters */ /** diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index 0b14ca234..bf8963577 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -920,7 +920,7 @@ static int bnxt_dev_info_get_op(struct rte_eth_dev *eth_dev, if (BNXT_PF(bp)) dev_info->max_vfs = pdev->max_vfs; - max_rx_rings = BNXT_MAX_RINGS(bp); + max_rx_rings = bnxt_max_rings(bp); /* For the sake of symmetry, max_rx_queues = max_tx_queues */ dev_info->max_rx_queues = max_rx_rings; dev_info->max_tx_queues = max_rx_rings; diff --git a/drivers/net/bnxt/bnxt_rxq.c b/drivers/net/bnxt/bnxt_rxq.c index 328cc994d..19e11e47b 100644 --- a/drivers/net/bnxt/bnxt_rxq.c +++ b/drivers/net/bnxt/bnxt_rxq.c @@ -311,7 +311,7 @@ int bnxt_rx_queue_setup_op(struct rte_eth_dev *eth_dev, if (rc) return rc; - if (queue_idx >= BNXT_MAX_RINGS(bp)) { + if (queue_idx >= bnxt_max_rings(bp)) { PMD_DRV_LOG(ERR, "Cannot create Rx ring %d. Only %d rings available\n", queue_idx, bp->max_rx_rings); @@ -364,8 +364,9 @@ int bnxt_rx_queue_setup_op(struct rte_eth_dev *eth_dev, eth_dev->data->rx_queues[queue_idx] = rxq; /* Allocate RX ring hardware descriptors */ - if (bnxt_alloc_rings(bp, queue_idx, NULL, rxq, rxq->cp_ring, NULL, - "rxr")) { + rc = bnxt_alloc_rings(bp, queue_idx, NULL, rxq, rxq->cp_ring, NULL, + "rxr"); + if (rc) { PMD_DRV_LOG(ERR, "ring_dma_zone_reserve for rx_ring failed!\n"); goto err; diff --git a/drivers/net/bnxt/bnxt_txq.c b/drivers/net/bnxt/bnxt_txq.c index c9792a2af..99a31cef2 100644 --- a/drivers/net/bnxt/bnxt_txq.c +++ b/drivers/net/bnxt/bnxt_txq.c @@ -98,7 +98,7 @@ int bnxt_tx_queue_setup_op(struct rte_eth_dev *eth_dev, if (rc) return rc; - if (queue_idx >= BNXT_MAX_RINGS(bp)) { + if (queue_idx >= bnxt_max_rings(bp)) { PMD_DRV_LOG(ERR, "Cannot create Tx ring %d. Only %d rings available\n", queue_idx, bp->max_tx_rings); -- 2.21.1 (Apple Git-122.3)
From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> 1. moved invalid VNIC id check to the beginning of the function. 2. removed a duplicate check which avoids unnecessary code indentation. Fixes: 49d0709b257fc ("net/bnxt: delete and flush L2 filters cleanly") Cc: stable@dpdk.org Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_hwrm.c | 42 +++++++++++++++++------------------- 1 file changed, 20 insertions(+), 22 deletions(-) diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c index 6f5402070..cee2656c1 100644 --- a/drivers/net/bnxt/bnxt_hwrm.c +++ b/drivers/net/bnxt/bnxt_hwrm.c @@ -4896,37 +4896,35 @@ int bnxt_vnic_rss_configure(struct bnxt *bp, struct bnxt_vnic_info *vnic) { unsigned int rss_idx, fw_idx, i; + if (vnic->fw_vnic_id == INVALID_HW_RING_ID) + return 0; + if (!(vnic->rss_table && vnic->hash_type)) return 0; if (BNXT_CHIP_P5(bp)) return bnxt_vnic_rss_configure_p5(bp, vnic); - if (vnic->fw_vnic_id == INVALID_HW_RING_ID) - return 0; - - if (vnic->rss_table && vnic->hash_type) { - /* - * Fill the RSS hash & redirection table with - * ring group ids for all VNICs - */ - for (rss_idx = 0, fw_idx = 0; rss_idx < HW_HASH_INDEX_SIZE; - rss_idx++, fw_idx++) { - for (i = 0; i < bp->rx_cp_nr_rings; i++) { - fw_idx %= bp->rx_cp_nr_rings; - if (vnic->fw_grp_ids[fw_idx] != - INVALID_HW_RING_ID) - break; - fw_idx++; - } - if (i == bp->rx_cp_nr_rings) - return 0; - vnic->rss_table[rss_idx] = vnic->fw_grp_ids[fw_idx]; + /* + * Fill the RSS hash & redirection table with + * ring group ids for all VNICs + */ + for (rss_idx = 0, fw_idx = 0; rss_idx < HW_HASH_INDEX_SIZE; + rss_idx++, fw_idx++) { + for (i = 0; i < bp->rx_cp_nr_rings; i++) { + fw_idx %= bp->rx_cp_nr_rings; + if (vnic->fw_grp_ids[fw_idx] != INVALID_HW_RING_ID) + break; + fw_idx++; } - return bnxt_hwrm_vnic_rss_cfg(bp, vnic); + + if (i == bp->rx_cp_nr_rings) + return 0; + + vnic->rss_table[rss_idx] = vnic->fw_grp_ids[fw_idx]; } - return 0; + return bnxt_hwrm_vnic_rss_cfg(bp, vnic); } static void bnxt_hwrm_set_coal_params(struct bnxt_coal *hw_coal, -- 2.21.1 (Apple Git-122.3)
From: Somnath Kotur <somnath.kotur@broadcom.com> This cmd should be called by every driver after 'hwrm_func_cfg' to get the actual number of resources allocated by the HWRM. The values returned in the cmd are the max values for that PF. Also, now that the max values for the PF are computed in probe itself, no need to invoke FUNC_QCAPs or any other cmd in dev_configure_op() as that would just override the actual max values obtained above. Fixes: f8168ca0e690 ("net/bnxt: support thor controller") Cc: stable@dpdk.org Signed-off-by: Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Ajit Khaparde <ajit.khaparde@broadcom.com> --- drivers/net/bnxt/bnxt_ethdev.c | 7 ------- drivers/net/bnxt/bnxt_hwrm.c | 36 +++++++++++++++++++++++++++++++++- 2 files changed, 35 insertions(+), 8 deletions(-) diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c index bf8963577..21d9e26ca 100644 --- a/drivers/net/bnxt/bnxt_ethdev.c +++ b/drivers/net/bnxt/bnxt_ethdev.c @@ -1071,13 +1071,6 @@ static int bnxt_dev_configure_op(struct rte_eth_dev *eth_dev) } pthread_mutex_unlock(&bp->def_cp_lock); - } else { - /* legacy driver needs to get updated values */ - rc = bnxt_hwrm_func_qcaps(bp); - if (rc) { - PMD_DRV_LOG(ERR, "hwrm func qcaps fail:%d\n", rc); - return rc; - } } /* Inherit new configurations */ diff --git a/drivers/net/bnxt/bnxt_hwrm.c b/drivers/net/bnxt/bnxt_hwrm.c index cee2656c1..a3eb133c0 100644 --- a/drivers/net/bnxt/bnxt_hwrm.c +++ b/drivers/net/bnxt/bnxt_hwrm.c @@ -3456,6 +3456,35 @@ static int bnxt_update_max_resources(struct bnxt *bp, return 0; } +/* Update the PF resource values based on how many resources + * got allocated to it. + */ +static int bnxt_update_max_resources_pf_only(struct bnxt *bp) +{ + struct hwrm_func_qcfg_input req = {0}; + struct hwrm_func_qcfg_output *resp = bp->hwrm_cmd_resp_addr; + int rc; + + /* Get the actual allocated values now */ + HWRM_PREP(&req, HWRM_FUNC_QCFG, BNXT_USE_CHIMP_MB); + req.fid = rte_cpu_to_le_16(0xffff); + rc = bnxt_hwrm_send_message(bp, &req, sizeof(req), BNXT_USE_CHIMP_MB); + HWRM_CHECK_RESULT(); + + bp->max_rsscos_ctx = rte_le_to_cpu_16(resp->alloc_rsscos_ctx); + bp->max_stat_ctx = rte_le_to_cpu_16(resp->alloc_stat_ctx); + bp->max_cp_rings = rte_le_to_cpu_16(resp->alloc_cmpl_rings); + bp->max_tx_rings = rte_le_to_cpu_16(resp->alloc_tx_rings); + bp->max_rx_rings = rte_le_to_cpu_16(resp->alloc_rx_rings); + bp->max_l2_ctx = rte_le_to_cpu_16(resp->alloc_l2_ctx); + bp->max_ring_grps = rte_le_to_cpu_16(resp->alloc_hw_ring_grps); + bp->max_vnics = rte_le_to_cpu_16(resp->alloc_vnics); + + HWRM_UNLOCK(); + + return 0; +} + int bnxt_hwrm_func_qcfg_current_vf_vlan(struct bnxt *bp, int vf) { struct hwrm_func_qcfg_input req = {0}; @@ -3555,8 +3584,13 @@ int bnxt_hwrm_allocate_pf_only(struct bnxt *bp) HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE); bp->pf->func_cfg_flags |= HWRM_FUNC_CFG_INPUT_FLAGS_STD_TX_RING_MODE_DISABLE; + rc = bnxt_hwrm_pf_func_cfg(bp, &pf_resc); - rc = __bnxt_hwrm_func_qcaps(bp); + if (rc) + return rc; + + rc = bnxt_update_max_resources_pf_only(bp); + return rc; } -- 2.21.1 (Apple Git-122.3)
On Wed, Dec 9, 2020 at 6:56 PM Ajit Khaparde <ajitkhaparde@gmail.com> wrote:
>
> In case mutex init fails during initialization, start cleanup and
> fail the initialization process.
>
> Fixes: a73b8e939f10 ("net/bnxt: fix race between start and interrupt handler")
> Cc: stable@dpdk.org
>
> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
> ---
> drivers/net/bnxt/bnxt_ethdev.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/bnxt/bnxt_ethdev.c b/drivers/net/bnxt/bnxt_ethdev.c
> index 33358779b..0b14ca234 100644
> --- a/drivers/net/bnxt/bnxt_ethdev.c
> +++ b/drivers/net/bnxt/bnxt_ethdev.c
> @@ -4739,8 +4739,10 @@ bnxt_init_locks(struct bnxt *bp)
> }
>
> err = pthread_mutex_init(&bp->def_cp_lock, NULL);
> - if (err)
> + if (err) {
> PMD_DRV_LOG(ERR, "Unable to initialize def_cp_lock\n");
> + return err;
> + }
>
> err = pthread_mutex_init(&bp->health_check_lock, NULL);
> if (err)
> --
> 2.21.1 (Apple Git-122.3)
>
Acked-by: Lance Richardson <lance.richardson@broadcom.com>
On Wed, Dec 9, 2020 at 6:56 PM Ajit Khaparde <ajitkhaparde@gmail.com> wrote:
>
> &device requires the %u format specifer not the %d specifier, as
> &device is unsigned.
>
> Fixes: a46bbb57605b ("net/bnxt: update multi device design")
> Cc: stable@dpdk.org
>
> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
> ---
> drivers/net/bnxt/tf_core/tf_core.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/net/bnxt/tf_core/tf_core.c b/drivers/net/bnxt/tf_core/tf_core.c
> index 24d49096a..3409cbbce 100644
> --- a/drivers/net/bnxt/tf_core/tf_core.c
> +++ b/drivers/net/bnxt/tf_core/tf_core.c
> @@ -82,7 +82,7 @@ tf_open_session(struct tf *tfp,
> return rc;
>
> TFP_DRV_LOG(INFO,
> - "domain:%d, bus:%d, device:%d\n",
> + "domain:%d, bus:%d, device:%u\n",
> parms->session_id.internal.domain,
> parms->session_id.internal.bus,
> parms->session_id.internal.device);
> --
> 2.21.1 (Apple Git-122.3)
>
Acked-by: Lance Richardson <lance.richardson@broadcom.com>