bnxt patchset with bug fixes
cover bnxt patchset with bug fixes
This patch against dpdk-next-net contains bug fixes for issues found during testing.
This patch is dependent on bnxt patchset
https://patchwork.dpdk.org/project/dpdk/list/?series=6686
Please apply. Thanks
Ajit Khaparde (2):
net/bnxt: update trusted VF status only when it changes net/bnxt: get the default HWRM command timeout from firmware
Kalesh AP (5):
net/bnxt: change return value of few routines net/bnxt: return error if setting link up fail net/bnxt: remove redundant header file inclusion net/bnxt: fix to handle if change status in port start only net/bnxt: remove a useless check in validate flow routine
Rahul Gupta (2):
net/bnxt: reduce cleanup time during reset recovery net/bnxt: change msix vector to queue mapping
Santoshkumar Karanappa Rastapur (1):
net/bnxt: free default completion ring before VF configuration
drivers/net/bnxt/bnxt.h | 8 ++++++ drivers/net/bnxt/bnxt_cpr.c | 1 - drivers/net/bnxt/bnxt_ethdev.c | 45 ++++++++++++++++++++++------------ drivers/net/bnxt/bnxt_flow.c | 4 +-- drivers/net/bnxt/bnxt_hwrm.c | 22 ++++++++++++----- drivers/net/bnxt/bnxt_irq.c | 1 - drivers/net/bnxt/bnxt_irq.h | 3 --- drivers/net/bnxt/bnxt_ring.c | 7 ++---- drivers/net/bnxt/bnxt_ring.h | 2 +- drivers/net/bnxt/bnxt_rxq.c | 1 - drivers/net/bnxt/bnxt_rxr.c | 1 - drivers/net/bnxt/bnxt_txq.c | 1 - drivers/net/bnxt/bnxt_txr.c | 1 - 13 files changed, 57 insertions(+), 40 deletions(-)
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 1/10 net/bnxt: change return value of few routines
From: Kalesh AP <[email protected]>
Change bnxt_init_nic() and bnxt_init_ring_grps() to void functions as these can never fail.
--- drivers/net/bnxt/bnxt_ethdev.c | 13 +++---------- drivers/net/bnxt/bnxt_ring.c | 4 +--- drivers/net/bnxt/bnxt_ring.h | 2 +- 3 files changed, 5 insertions(+), 14 deletions(-)
drivers/net/bnxt/bnxt_ethdev.c +3 -10
@@ -518,20 +518,13 @@ static int bnxt_shutdown_nic(struct bnxt *bp)return 0 ; } -static int bnxt_init_nic (struct bnxt *bp) +static void bnxt_init_nic (struct bnxt *bp) { -int rc; - -if (BNXT_HAS_RING_GRPS (bp)) { - rc =bnxt_init_ring_grps (bp); -if (rc) -return rc; - } +if (BNXT_HAS_RING_GRPS (bp)) +bnxt_init_ring_grps (bp);bnxt_init_vnics (bp);bnxt_init_filters (bp); - -return 0 ; }/*
drivers/net/bnxt/bnxt_ring.c +1 -3
@@ -39,15 +39,13 @@ void bnxt_free_ring(struct bnxt_ring *ring)* Ring groups */ -int bnxt_init_ring_grps (struct bnxt *bp) +void bnxt_init_ring_grps (struct bnxt *bp) {unsigned int i;for (i =0 ; i < bp->max_ring_grps; i++)memset (&bp->grp_info[i], (uint8_t )HWRM_NA_SIGNATURE ,sizeof (struct bnxt_ring_grp_info )); - -return 0 ; }int bnxt_alloc_ring_grps (struct bnxt *bp)
drivers/net/bnxt/bnxt_ring.h +1 -1
@@ -65,7 +65,7 @@ struct bnxt_tx_ring_info;struct bnxt_rx_ring_info ;struct bnxt_cp_ring_info ;void bnxt_free_ring (struct bnxt_ring *ring); -int bnxt_init_ring_grps (struct bnxt *bp); +void bnxt_init_ring_grps (struct bnxt *bp);int bnxt_alloc_ring_grps (struct bnxt *bp); int bnxt_alloc_rings(struct bnxt *bp, uint16_t qidx, struct bnxt_tx_queue *txq,
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 2/10 net/bnxt: free default completion ring before VF configuration
In reply to Ajit KhapardeFrom: Santoshkumar Karanappa Rastapur <[email protected]>
If the VF resources need to be reconfigured using the bnxt_hwrm_func_reserve_vf_resc, make sure that the default completion ring is freed first before the resources are reserved.
Reallocate the ring once the VF resources are configured.
--- drivers/net/bnxt/bnxt_ethdev.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
drivers/net/bnxt/bnxt_ethdev.c +17 -0
@@ -662,11 +662,28 @@ static int bnxt_dev_configure_op(struct rte_eth_dev *eth_dev)return -ENOSPC ; } +/* If a resource has already been allocated - in this case +* it is the async completion ring, free it. Reallocate it after +* resource reservation. This will ensure the resource counts +* are calculated correctly. +*/ +if (!BNXT_HAS_NQ (bp) && bp->async_cp_ring) { +bnxt_disable_int (bp); +bnxt_free_cp_ring (bp, bp->async_cp_ring); + } + rc =bnxt_hwrm_func_reserve_vf_resc (bp, false);if (rc) {PMD_DRV_LOG (ERR ,"HWRM resource alloc fail:%x\n" , rc);return -ENOSPC ; } + +if (!BNXT_HAS_NQ (bp) && bp->async_cp_ring) { + rc =bnxt_alloc_async_cp_ring (bp); +if (rc) +return rc; +bnxt_enable_int (bp); + } }else {/* legacy driver needs to get updated values */ rc =bnxt_hwrm_func_qcaps (bp);
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 3/10 net/bnxt: return error if setting link up fail
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
Currently bnxt driver does not return error in case setting link up fails.
--- drivers/net/bnxt/bnxt_ethdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/bnxt/bnxt_ethdev.c +1 -1
@@ -923,7 +923,7 @@ static int bnxt_dev_set_link_up_op(struct rte_eth_dev *eth_dev) eth_dev->data->dev_link.link_status =1 ;bnxt_print_link_info (eth_dev); -return 0 ; +return rc; }static int bnxt_dev_set_link_down_op (struct rte_eth_dev *eth_dev)
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 4/10 net/bnxt: remove redundant header file inclusion
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
bnxt.h header includes bnxt_cpr.h and bnxt_util.h. There is no need to include these headers file explicitly.
This commit does not cause any functional change.
--- drivers/net/bnxt/bnxt_cpr.c | 1 - drivers/net/bnxt/bnxt_ethdev.c | 2 -- drivers/net/bnxt/bnxt_flow.c | 1 - drivers/net/bnxt/bnxt_hwrm.c | 1 - drivers/net/bnxt/bnxt_irq.c | 1 - drivers/net/bnxt/bnxt_ring.c | 1 - drivers/net/bnxt/bnxt_rxq.c | 1 - drivers/net/bnxt/bnxt_rxr.c | 1 - drivers/net/bnxt/bnxt_txq.c | 1 - drivers/net/bnxt/bnxt_txr.c | 1 - 10 files changed, 11 deletions(-)
drivers/net/bnxt/bnxt_cpr.c +0 -1
@@ -8,7 +8,6 @@#include <rte_cycles.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_hwrm.h" #include "bnxt_ring.h" #include "hsi_struct_def_dpdk.h"
drivers/net/bnxt/bnxt_ethdev.c +0 -2
@@ -14,7 +14,6 @@#include <rte_alarm.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_filter.h" #include "bnxt_hwrm.h" #include "bnxt_irq.h"
@@ -27,7 +26,6 @@#include "bnxt_vnic.h" #include "hsi_struct_def_dpdk.h" #include "bnxt_nvm_defs.h" -#include "bnxt_util.h" #define DRV_MODULE_NAME "bnxt" static const char bnxt_version[] =
drivers/net/bnxt/bnxt_flow.c +0 -1
@@ -17,7 +17,6 @@#include "bnxt_ring.h" #include "bnxt_rxq.h" #include "bnxt_vnic.h" -#include "bnxt_util.h" #include "hsi_struct_def_dpdk.h" static int
drivers/net/bnxt/bnxt_hwrm.c +0 -1
@@ -13,7 +13,6 @@#include <rte_version.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_filter.h" #include "bnxt_hwrm.h" #include "bnxt_rxq.h"
drivers/net/bnxt/bnxt_irq.c +0 -1
@@ -9,7 +9,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_irq.h" #include "bnxt_ring.h" #include "hsi_struct_def_dpdk.h"
drivers/net/bnxt/bnxt_ring.c +0 -1
@@ -9,7 +9,6 @@#include <unistd.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_hwrm.h" #include "bnxt_ring.h" #include "bnxt_rxq.h"
drivers/net/bnxt/bnxt_rxq.c +0 -1
@@ -8,7 +8,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_filter.h" #include "bnxt_hwrm.h" #include "bnxt_ring.h"
drivers/net/bnxt/bnxt_rxr.c +0 -1
@@ -12,7 +12,6 @@#include <rte_memory.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_ring.h" #include "bnxt_rxr.h" #include "bnxt_rxq.h"
drivers/net/bnxt/bnxt_txq.c +0 -1
@@ -8,7 +8,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_ring.h" #include "bnxt_txq.h" #include "bnxt_txr.h"
drivers/net/bnxt/bnxt_txr.c +0 -1
@@ -9,7 +9,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_ring.h" #include "bnxt_txq.h" #include "bnxt_txr.h"
Signature · 1 line
2.20.1 (Apple Git-117)
Re: [dpdk-dev] [PATCH 04/10] net/bnxt: remove redundant header file inclusion
Quoted 9 lines · Ajit Khaparde
From: Kalesh AP <[email protected]>
bnxt.hheader includesbnxt_cpr.handbnxt_util.h. There is no need to include these headers file explicitly.This commit does not cause any functional change.
Signed-off-by Kalesh AP <[email protected]>Reviewed-by Ajit Khaparde <[email protected]>
don't you want this patch to be backported? I think it may be good to backport these kind of small fixes too to reduce conflicts in the driver in the future.
PATCH 5/10 net/bnxt: update trusted VF status only when it changes
In reply to Ajit KhapardeWe are currently logging trusted VF information everytime a VF config change is indicated by the async message. Update it only when new setting is different from the current setting.
--- drivers/net/bnxt/bnxt_hwrm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/net/bnxt/bnxt_hwrm.c +4 -1
@@ -2873,10 +2873,13 @@ int bnxt_hwrm_func_qcfg(struct bnxt *bp, uint16_t *mtu)if (BNXT_PF (bp) && (flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_MULTI_HOST )) bp->flags |=BNXT_FLAG_MULTI_HOST ; -if (BNXT_VF (bp) && (flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_TRUSTED_VF )) { +if (BNXT_VF (bp) && + !BNXT_VF_IS_TRUSTED (bp) && + (flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_TRUSTED_VF )) { bp->flags |=BNXT_FLAG_TRUSTED_VF_EN ;PMD_DRV_LOG (INFO ,"Trusted VF cap enabled\n" ); }else if (BNXT_VF (bp) && +BNXT_VF_IS_TRUSTED (bp) && !(flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_TRUSTED_VF )) { bp->flags &= ~BNXT_FLAG_TRUSTED_VF_EN ;PMD_DRV_LOG (INFO ,"Trusted VF cap disabled\n" );
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 6/10 net/bnxt: get the default HWRM command timeout from firmware
In reply to Ajit KhapardeThe HWRM command timeout is set to a very high value. VER_GET command response returns the default request timeout value. Use this value for waiting for HWRM commands to complete. Poll for the valid bit every 1us instead of 600us.
--- drivers/net/bnxt/bnxt.h | 5 +++++ drivers/net/bnxt/bnxt_hwrm.c | 13 +++++++++---- 2 files changed, 14 insertions(+), 4 deletions(-)
drivers/net/bnxt/bnxt.h +5 -0
@@ -525,6 +525,11 @@ struct bnxt {uint16_t max_resp_len;uint16_t hwrm_max_ext_req_len; +/* default command timeout value of 50ms */ +#define HWRM_CMD_TIMEOUT 50000 +/* default HWRM request timeout value */ +uint32_t hwrm_cmd_timeout; +struct bnxt_link_info link_info;struct bnxt_cos_queue_info rx_cos_queue[BNXT_COS_QUEUE_COUNT ];struct bnxt_cos_queue_info tx_cos_queue[BNXT_COS_QUEUE_COUNT ];
drivers/net/bnxt/bnxt_hwrm.c +9 -4
@@ -25,8 +25,6 @@#include <rte_io.h> -#define HWRM_CMD_TIMEOUT 6000000 -#define HWRM_SHORT_CMD_TIMEOUT 50000#define HWRM_SPEC_CODE_1_8_3 0x10803#define HWRM_VERSION_1_9_1 0x10901#define HWRM_VERSION_1_9_2 0x10903
@@ -105,9 +103,9 @@ static int bnxt_hwrm_send_message(struct bnxt *bp, void *msg,/* For VER_GET command, set timeout as 50ms */ if (rte_cpu_to_le_16 (req->req_type) ==HWRM_VER_GET ) - timeout =HWRM_SHORT_CMD_TIMEOUT ; -else timeout =HWRM_CMD_TIMEOUT ; +else + timeout = bp->hwrm_cmd_timeout;if (bp->flags &BNXT_FLAG_SHORT_CMD || msg_len > bp->max_req_len) {
@@ -969,6 +967,13 @@ int bnxt_hwrm_ver_get(struct bnxt *bp) fw_version |= resp->hwrm_intf_upd_8b; bp->hwrm_spec_code = fw_version; +/* def_req_timeout value is in milliseconds */ + bp->hwrm_cmd_timeout =rte_le_to_cpu_16 (resp->def_req_timeout); +/* convert timeout to usec */ + bp->hwrm_cmd_timeout *=1000 ; +if (!bp->hwrm_cmd_timeout) + bp->hwrm_cmd_timeout =HWRM_CMD_TIMEOUT ; +if (resp->hwrm_intf_maj_8b !=HWRM_VERSION_MAJOR ) {PMD_DRV_LOG (ERR ,"Unsupported firmware API version\n" ); rc = -EINVAL ;
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 7/10 net/bnxt: reduce cleanup time during reset recovery
In reply to Ajit KhapardeFrom: Rahul Gupta <[email protected]>
In stop_op(), don't wait for link down event's ASYNC response from FW.
--- drivers/net/bnxt/bnxt_ethdev.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/net/bnxt/bnxt_ethdev.c +6 -2
@@ -960,8 +960,12 @@ static void bnxt_dev_stop_op(struct rte_eth_dev *eth_dev) eth_dev->data->dev_link.link_status =0 ; }bnxt_dev_set_link_down_op (eth_dev); -/* Wait for link to be reset and the async notification to process. */ -rte_delay_ms (BNXT_LINK_WAIT_INTERVAL *2 ); + +/* Wait for link to be reset and the async notification to process. +* During reset recovery, there is no need to wait +*/ +if (!is_bnxt_in_error (bp)) +rte_delay_ms (BNXT_LINK_WAIT_INTERVAL *2 );/* Clean queue intr-vector mapping */ rte_intr_efd_disable (intr_handle);
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 8/10 net/bnxt: change msix vector to queue mapping
In reply to Ajit KhapardeFrom: Rahul Gupta <[email protected]>
DPDK PCIe-VFIO framework configures base MSIX vector for interrupts which is supported by other h/w. In case of bnxt, base MSIX vector starts with the RX completion queue 0. To comply with the DPDK framework We need to increase the map index by 1 so that RXTX completion queues events can be delivered to appropriate event listeners by kernel VFIO.
--- drivers/net/bnxt/bnxt.h | 3 +++ drivers/net/bnxt/bnxt_irq.h | 3 --- drivers/net/bnxt/bnxt_ring.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-)
drivers/net/bnxt/bnxt.h +3 -0
@@ -58,6 +58,9 @@#define BNXT_NUM_ASYNC_CPR (bp) 1#endif +#define BNXT_MISC_VEC_ID RTE_INTR_VEC_ZERO_OFFSET +#define BNXT_RX_VEC_START RTE_INTR_VEC_RXTX_OFFSET +/* Chimp Communication Channel */ #define GRCPF_REG_CHIMP_CHANNEL_OFFSET 0x0#define GRCPF_REG_CHIMP_COMM_TRIGGER 0x100
drivers/net/bnxt/bnxt_irq.h +0 -3
@@ -6,9 +6,6 @@ #ifndef _BNXT_IRQ_H_ #define _BNXT_IRQ_H_ -#define BNXT_MISC_VEC_ID RTE_INTR_VEC_ZERO_OFFSET -#define BNXT_RX_VEC_START RTE_INTR_VEC_RXTX_OFFSET - struct bnxt_irq { rte_intr_callback_fn handler; unsigned int vector;
drivers/net/bnxt/bnxt_ring.c +1 -1
@@ -404,7 +404,7 @@ static int bnxt_alloc_cmpl_ring(struct bnxt *bp, int queue_index, {struct bnxt_ring *cp_ring = cpr->cp_ring_struct;uint32_t nq_ring_id =HWRM_NA_SIGNATURE ; -int cp_ring_index = queue_index +BNXT_NUM_ASYNC_CPR (bp); +int cp_ring_index = queue_index +BNXT_RX_VEC_START ;struct bnxt_cp_ring_info *nqr = bp->rxtx_nq_ring;uint8_t ring_type;int rc =0 ;
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 9/10 net/bnxt: fix to handle if change status in port start only
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
Driver should handle if change status during port start only. Also fixed to enable interrupts after issuing hwrm_if_change.
--- drivers/net/bnxt/bnxt_ethdev.c | 3 ++- drivers/net/bnxt/bnxt_hwrm.c | 3 +++ 2 files changed, 5 insertions(+), 1 deletion(-)
drivers/net/bnxt/bnxt_ethdev.c +2 -1
@@ -867,7 +867,6 @@ static int bnxt_dev_start_op(struct rte_eth_dev *eth_dev) bp->rx_cp_nr_rings, RTE_ETHDEV_QUEUE_STAT_CNTRS); } -bnxt_enable_int (bp); rc =bnxt_hwrm_if_change (bp,1 );if (!rc) {if (bp->flags &BNXT_FLAG_IF_CHANGE_HOT_FW_RESET_DONE ) {
@@ -877,6 +876,8 @@ static int bnxt_dev_start_op(struct rte_eth_dev *eth_dev) } } +bnxt_enable_int (bp); + rc =bnxt_init_chip (bp);if (rc) goto error;
drivers/net/bnxt/bnxt_hwrm.c +3 -0
@@ -4858,6 +4858,9 @@ int bnxt_hwrm_if_change(struct bnxt *bp, bool up) flags =rte_le_to_cpu_32 (resp->flags);HWRM_UNLOCK (); +if (!up) +return 0 ; +if (flags &HWRM_FUNC_DRV_IF_CHANGE_OUTPUT_FLAGS_HOT_FW_RESET_DONE ) {PMD_DRV_LOG (INFO ,"FW reset happened while port was down\n" ); bp->flags |=BNXT_FLAG_IF_CHANGE_HOT_FW_RESET_DONE ;
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH 10/10 net/bnxt: remove a useless check in validate flow routine
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
Since vnic->rx_queue_cnt is already checked against non-zero, removed the useless check from the inside condition check.
--- drivers/net/bnxt/bnxt_flow.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/net/bnxt/bnxt_flow.c +1 -2
@@ -1067,8 +1067,7 @@ bnxt_validate_and_parse_flow(struct rte_eth_dev *dev, goto ret; }if (vnic->rx_queue_cnt) { -if (vnic->rx_queue_cnt >1 || - vnic->start_grp_id != act_q->index) { +if (vnic->start_grp_id != act_q->index) {PMD_DRV_LOG (ERR ,"VNIC already in use\n" ); rte_flow_error_set(error,
Signature · 1 line
2.20.1 (Apple Git-117)
Re: [dpdk-dev] [PATCH 10/10] net/bnxt: remove a useless check in validate flow routine
Quoted 9 lines · Ajit Khaparde
From: Kalesh AP <[email protected]>
Since
vnic->rx_queue_cntis already checked against non-zero, removed the useless check from the inside condition check.Fixes 4be31562a4ba ("net/bnxt: allow flow creation when RSS is enabled")Signed-off-by Kalesh AP <[email protected]>Reviewed-by Ajit Khaparde <[email protected]>
Squashed into relevant commit in next-net, thanks.
(please send new version of the patchset on top of latest next-net)
v2 cover bnxt patchset with bug fixes
This patch against dpdk-next-net mostly contains bug fixes for issues found during testing.
v1->v2:
-Split patch 09/10 into two and reworded the commit messages. -Rebased against latest dpdk-next-net
-Some of the patches were missing the Fixes tag. Added it to them.
Ajit Khaparde (2):
net/bnxt: update trusted VF status only when it changes net/bnxt: get the default HWRM command timeout from firmware
Kalesh AP (7):
net/bnxt: fix return checks and return values
net/bnxt: return error if setting link up fail
net/bnxt: remove redundant header file inclusion net/bnxt: enable interrupts after checking interface status from FW net/bnxt: fix to handle interface change status in port start only net/bnxt: fix coding style issues
net/bnxt: remove unnecessary variable assignment
Rahul Gupta (2):
net/bnxt: reduce cleanup time during reset recovery net/bnxt: change msix vector to queue mapping
Santoshkumar Karanappa Rastapur (1):
net/bnxt: free default completion ring before VF configuration
drivers/net/bnxt/bnxt.h | 8 ++++++++ drivers/net/bnxt/bnxt_cpr.c | 1 - drivers/net/bnxt/bnxt_ethdev.c | 32 +++++++++++++++++++++++------- drivers/net/bnxt/bnxt_flow.c | 1 - drivers/net/bnxt/bnxt_hwrm.c | 33 ++++++++++++++++++------------- drivers/net/bnxt/bnxt_irq.c | 1 - drivers/net/bnxt/bnxt_irq.h | 3 --- drivers/net/bnxt/bnxt_ring.c | 3 +-- drivers/net/bnxt/bnxt_rxq.c | 15 ++++++-------- drivers/net/bnxt/bnxt_rxr.c | 1 - drivers/net/bnxt/bnxt_stats.c | 36 +++++++++++----------------------- drivers/net/bnxt/bnxt_txq.c | 1 - drivers/net/bnxt/bnxt_txr.c | 1 - 13 files changed, 70 insertions(+), 66 deletions(-)
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 1/12 net/bnxt: fix return checks and return values
From: Kalesh AP <[email protected]>
This patch fixes few checks and few return values while getting and clearing device statistics.
1. Fixed to return standard error code.
2. Clubbed few error checks
3. Removed an unnecessary return check
--- drivers/net/bnxt/bnxt_stats.c | 36 +++++++++++------------------------ 1 file changed, 11 insertions(+), 25 deletions(-)
drivers/net/bnxt/bnxt_stats.c +11 -25
@@ -360,7 +360,7 @@ int bnxt_stats_get_op(struct rte_eth_dev *eth_dev,memset (bnxt_stats,0 ,sizeof (*bnxt_stats));if (!(bp->flags &BNXT_FLAG_INIT_DONE )) {PMD_DRV_LOG (ERR ,"Device Initialization not complete!\n" ); -return -1 ; +return -EIO ; } num_q_stats = RTE_MIN(bp->rx_cp_nr_rings,
@@ -390,9 +390,8 @@ int bnxt_stats_get_op(struct rte_eth_dev *eth_dev,if (unlikely (rc))return rc; } + rc =bnxt_hwrm_func_qstats (bp,0xffff , bnxt_stats); -if (unlikely (rc)) -return rc;return rc; }
@@ -573,30 +572,17 @@ int bnxt_dev_xstats_reset_op(struct rte_eth_dev *eth_dev)if (ret)return ret; -if (bp->flags &BNXT_FLAG_PORT_STATS &&BNXT_SINGLE_PF (bp)) { - ret =bnxt_hwrm_port_clr_stats (bp); -if (ret !=0 ) { -PMD_DRV_LOG (ERR ,"Operation failed: %s\n" , -strerror (-ret)); -return ret; - } - } - - ret =0 ; - -if (BNXT_VF (bp)) { -PMD_DRV_LOG (ERR ,"Operation not supported on a VF device\n" ); - ret = -ENOTSUP ; - } -if (!BNXT_SINGLE_PF (bp)) { -PMD_DRV_LOG (ERR ,"Operation not supported on a MF device\n" ); - ret = -ENOTSUP ; - } -if (!(bp->flags &BNXT_FLAG_PORT_STATS )) { +if (BNXT_VF (bp) || !BNXT_SINGLE_PF (bp) || + !(bp->flags &BNXT_FLAG_PORT_STATS )) {PMD_DRV_LOG (ERR ,"Operation not supported\n" ); ret = -ENOTSUP ; } + ret =bnxt_hwrm_port_clr_stats (bp); +if (ret !=0 ) +PMD_DRV_LOG (ERR ,"Failed to reset xstats: %s\n" , +strerror (-ret)); +return ret; }
@@ -625,7 +611,7 @@ int bnxt_dev_xstats_get_by_id_op(struct rte_eth_dev *dev, const uint64_t *ids,for (i =0 ; i < limit; i++) {if (ids[i] >= stat_cnt) {PMD_DRV_LOG (ERR ,"id value isn't valid" ); -return -1 ; +return -EINVAL ; } values[i] = values_copy[ids[i]]; }
@@ -659,7 +645,7 @@ int bnxt_dev_xstats_get_names_by_id_op(struct rte_eth_dev *dev,for (i =0 ; i < limit; i++) {if (ids[i] >= stat_cnt) {PMD_DRV_LOG (ERR ,"id value isn't valid" ); -return -1 ; +return -EINVAL ; }strcpy (xstats_names[i].name, xstats_names_copy[ids[i]].name);
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 2/12 net/bnxt: free default completion ring before VF configuration
In reply to Ajit KhapardeFrom: Santoshkumar Karanappa Rastapur <[email protected]>
If the VF resources need to be reconfigured using the bnxt_hwrm_func_reserve_vf_resc, make sure that the default completion ring is freed first before the resources are reserved.
Reallocate the ring once the VF resources are configured.
--- drivers/net/bnxt/bnxt_ethdev.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+)
drivers/net/bnxt/bnxt_ethdev.c +17 -0
@@ -669,11 +669,28 @@ static int bnxt_dev_configure_op(struct rte_eth_dev *eth_dev)return -ENOSPC ; } +/* If a resource has already been allocated - in this case +* it is the async completion ring, free it. Reallocate it after +* resource reservation. This will ensure the resource counts +* are calculated correctly. +*/ +if (!BNXT_HAS_NQ (bp) && bp->async_cp_ring) { +bnxt_disable_int (bp); +bnxt_free_cp_ring (bp, bp->async_cp_ring); + } + rc =bnxt_hwrm_func_reserve_vf_resc (bp, false);if (rc) {PMD_DRV_LOG (ERR ,"HWRM resource alloc fail:%x\n" , rc);return -ENOSPC ; } + +if (!BNXT_HAS_NQ (bp) && bp->async_cp_ring) { + rc =bnxt_alloc_async_cp_ring (bp); +if (rc) +return rc; +bnxt_enable_int (bp); + } }else {/* legacy driver needs to get updated values */ rc =bnxt_hwrm_func_qcaps (bp);
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 3/12 net/bnxt: return error if setting link up fail
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
Currently bnxt driver does not return error in case setting link up fails.
--- drivers/net/bnxt/bnxt_ethdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/bnxt/bnxt_ethdev.c +1 -1
@@ -930,7 +930,7 @@ static int bnxt_dev_set_link_up_op(struct rte_eth_dev *eth_dev) eth_dev->data->dev_link.link_status =1 ;bnxt_print_link_info (eth_dev); -return 0 ; +return rc; }static int bnxt_dev_set_link_down_op (struct rte_eth_dev *eth_dev)
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 4/12 net/bnxt: remove redundant header file inclusion
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
bnxt.h header includes bnxt_cpr.h and bnxt_util.h. There is no need to include these headers file explicitly.
This commit does not cause any functional change.
--- drivers/net/bnxt/bnxt_cpr.c | 1 - drivers/net/bnxt/bnxt_ethdev.c | 2 -- drivers/net/bnxt/bnxt_flow.c | 1 - drivers/net/bnxt/bnxt_hwrm.c | 1 - drivers/net/bnxt/bnxt_irq.c | 1 - drivers/net/bnxt/bnxt_ring.c | 1 - drivers/net/bnxt/bnxt_rxq.c | 1 - drivers/net/bnxt/bnxt_rxr.c | 1 - drivers/net/bnxt/bnxt_txq.c | 1 - drivers/net/bnxt/bnxt_txr.c | 1 - 10 files changed, 11 deletions(-)
drivers/net/bnxt/bnxt_cpr.c +0 -1
@@ -8,7 +8,6 @@#include <rte_cycles.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_hwrm.h" #include "bnxt_ring.h" #include "hsi_struct_def_dpdk.h"
drivers/net/bnxt/bnxt_ethdev.c +0 -2
@@ -14,7 +14,6 @@#include <rte_alarm.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_filter.h" #include "bnxt_hwrm.h" #include "bnxt_irq.h"
@@ -27,7 +26,6 @@#include "bnxt_vnic.h" #include "hsi_struct_def_dpdk.h" #include "bnxt_nvm_defs.h" -#include "bnxt_util.h" #define DRV_MODULE_NAME "bnxt" static const char bnxt_version[] =
drivers/net/bnxt/bnxt_flow.c +0 -1
@@ -17,7 +17,6 @@#include "bnxt_ring.h" #include "bnxt_rxq.h" #include "bnxt_vnic.h" -#include "bnxt_util.h" #include "hsi_struct_def_dpdk.h" static int
drivers/net/bnxt/bnxt_hwrm.c +0 -1
@@ -13,7 +13,6 @@#include <rte_version.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_filter.h" #include "bnxt_hwrm.h" #include "bnxt_rxq.h"
drivers/net/bnxt/bnxt_irq.c +0 -1
@@ -9,7 +9,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_irq.h" #include "bnxt_ring.h" #include "hsi_struct_def_dpdk.h"
drivers/net/bnxt/bnxt_ring.c +0 -1
@@ -9,7 +9,6 @@#include <unistd.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_hwrm.h" #include "bnxt_ring.h" #include "bnxt_rxq.h"
drivers/net/bnxt/bnxt_rxq.c +0 -1
@@ -8,7 +8,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_filter.h" #include "bnxt_hwrm.h" #include "bnxt_ring.h"
drivers/net/bnxt/bnxt_rxr.c +0 -1
@@ -12,7 +12,6 @@#include <rte_memory.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_ring.h" #include "bnxt_rxr.h" #include "bnxt_rxq.h"
drivers/net/bnxt/bnxt_txq.c +0 -1
@@ -8,7 +8,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_ring.h" #include "bnxt_txq.h" #include "bnxt_txr.h"
drivers/net/bnxt/bnxt_txr.c +0 -1
@@ -9,7 +9,6 @@#include <rte_malloc.h> #include "bnxt.h" -#include "bnxt_cpr.h" #include "bnxt_ring.h" #include "bnxt_txq.h" #include "bnxt_txr.h"
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 5/12 net/bnxt: update trusted VF status only when it changes
In reply to Ajit KhapardeWe are currently logging trusted VF information everytime a VF config change is indicated by the async message. Update it only when new setting is different from the current setting.
--- drivers/net/bnxt/bnxt_hwrm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/net/bnxt/bnxt_hwrm.c +4 -1
@@ -2873,10 +2873,13 @@ int bnxt_hwrm_func_qcfg(struct bnxt *bp, uint16_t *mtu)if (BNXT_PF (bp) && (flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_MULTI_HOST )) bp->flags |=BNXT_FLAG_MULTI_HOST ; -if (BNXT_VF (bp) && (flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_TRUSTED_VF )) { +if (BNXT_VF (bp) && + !BNXT_VF_IS_TRUSTED (bp) && + (flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_TRUSTED_VF )) { bp->flags |=BNXT_FLAG_TRUSTED_VF_EN ;PMD_DRV_LOG (INFO ,"Trusted VF cap enabled\n" ); }else if (BNXT_VF (bp) && +BNXT_VF_IS_TRUSTED (bp) && !(flags &HWRM_FUNC_QCFG_OUTPUT_FLAGS_TRUSTED_VF )) { bp->flags &= ~BNXT_FLAG_TRUSTED_VF_EN ;PMD_DRV_LOG (INFO ,"Trusted VF cap disabled\n" );
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 6/12 net/bnxt: get the default HWRM command timeout from firmware
In reply to Ajit KhapardeThe HWRM command timeout is set to a very high value. VER_GET command response returns the default request timeout value. Use this value for waiting for HWRM commands to complete. Poll for the valid bit every 1us instead of 600us.
--- drivers/net/bnxt/bnxt.h | 5 +++++ drivers/net/bnxt/bnxt_hwrm.c | 13 +++++++++---- 2 files changed, 14 insertions(+), 4 deletions(-)
drivers/net/bnxt/bnxt.h +5 -0
@@ -525,6 +525,11 @@ struct bnxt {uint16_t max_resp_len;uint16_t hwrm_max_ext_req_len; +/* default command timeout value of 50ms */ +#define HWRM_CMD_TIMEOUT 50000 +/* default HWRM request timeout value */ +uint32_t hwrm_cmd_timeout; +struct bnxt_link_info link_info;struct bnxt_cos_queue_info rx_cos_queue[BNXT_COS_QUEUE_COUNT ];struct bnxt_cos_queue_info tx_cos_queue[BNXT_COS_QUEUE_COUNT ];
drivers/net/bnxt/bnxt_hwrm.c +9 -4
@@ -25,8 +25,6 @@#include <rte_io.h> -#define HWRM_CMD_TIMEOUT 6000000 -#define HWRM_SHORT_CMD_TIMEOUT 50000#define HWRM_SPEC_CODE_1_8_3 0x10803#define HWRM_VERSION_1_9_1 0x10901#define HWRM_VERSION_1_9_2 0x10903
@@ -105,9 +103,9 @@ static int bnxt_hwrm_send_message(struct bnxt *bp, void *msg,/* For VER_GET command, set timeout as 50ms */ if (rte_cpu_to_le_16 (req->req_type) ==HWRM_VER_GET ) - timeout =HWRM_SHORT_CMD_TIMEOUT ; -else timeout =HWRM_CMD_TIMEOUT ; +else + timeout = bp->hwrm_cmd_timeout;if (bp->flags &BNXT_FLAG_SHORT_CMD || msg_len > bp->max_req_len) {
@@ -969,6 +967,13 @@ int bnxt_hwrm_ver_get(struct bnxt *bp) fw_version |= resp->hwrm_intf_upd_8b; bp->hwrm_spec_code = fw_version; +/* def_req_timeout value is in milliseconds */ + bp->hwrm_cmd_timeout =rte_le_to_cpu_16 (resp->def_req_timeout); +/* convert timeout to usec */ + bp->hwrm_cmd_timeout *=1000 ; +if (!bp->hwrm_cmd_timeout) + bp->hwrm_cmd_timeout =HWRM_CMD_TIMEOUT ; +if (resp->hwrm_intf_maj_8b !=HWRM_VERSION_MAJOR ) {PMD_DRV_LOG (ERR ,"Unsupported firmware API version\n" ); rc = -EINVAL ;
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 7/12 net/bnxt: reduce cleanup time during reset recovery
In reply to Ajit KhapardeFrom: Rahul Gupta <[email protected]>
In stop_op(), don't wait for link down event's ASYNC response from FW.
--- drivers/net/bnxt/bnxt_ethdev.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/net/bnxt/bnxt_ethdev.c +6 -2
@@ -967,8 +967,12 @@ static void bnxt_dev_stop_op(struct rte_eth_dev *eth_dev) eth_dev->data->dev_link.link_status =0 ; }bnxt_dev_set_link_down_op (eth_dev); -/* Wait for link to be reset and the async notification to process. */ -rte_delay_ms (BNXT_LINK_WAIT_INTERVAL *2 ); + +/* Wait for link to be reset and the async notification to process. +* During reset recovery, there is no need to wait +*/ +if (!is_bnxt_in_error (bp)) +rte_delay_ms (BNXT_LINK_WAIT_INTERVAL *2 );/* Clean queue intr-vector mapping */ rte_intr_efd_disable (intr_handle);
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 8/12 net/bnxt: change msix vector to queue mapping
In reply to Ajit KhapardeFrom: Rahul Gupta <[email protected]>
DPDK PCIe-VFIO framework configures base MSIX vector for interrupts which is supported by other h/w. In case of bnxt, base MSIX vector starts with the RX completion queue 0. To comply with the DPDK framework We need to increase the map index by 1 so that RXTX completion queues events can be delivered to appropriate event listeners by kernel VFIO.
--- drivers/net/bnxt/bnxt.h | 3 +++ drivers/net/bnxt/bnxt_irq.h | 3 --- drivers/net/bnxt/bnxt_ring.c | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-)
drivers/net/bnxt/bnxt.h +3 -0
@@ -58,6 +58,9 @@#define BNXT_NUM_ASYNC_CPR (bp) 1#endif +#define BNXT_MISC_VEC_ID RTE_INTR_VEC_ZERO_OFFSET +#define BNXT_RX_VEC_START RTE_INTR_VEC_RXTX_OFFSET +/* Chimp Communication Channel */ #define GRCPF_REG_CHIMP_CHANNEL_OFFSET 0x0#define GRCPF_REG_CHIMP_COMM_TRIGGER 0x100
drivers/net/bnxt/bnxt_irq.h +0 -3
@@ -6,9 +6,6 @@ #ifndef _BNXT_IRQ_H_ #define _BNXT_IRQ_H_ -#define BNXT_MISC_VEC_ID RTE_INTR_VEC_ZERO_OFFSET -#define BNXT_RX_VEC_START RTE_INTR_VEC_RXTX_OFFSET - struct bnxt_irq { rte_intr_callback_fn handler; unsigned int vector;
drivers/net/bnxt/bnxt_ring.c +1 -1
@@ -406,7 +406,7 @@ static int bnxt_alloc_cmpl_ring(struct bnxt *bp, int queue_index, {struct bnxt_ring *cp_ring = cpr->cp_ring_struct;uint32_t nq_ring_id =HWRM_NA_SIGNATURE ; -int cp_ring_index = queue_index +BNXT_NUM_ASYNC_CPR (bp); +int cp_ring_index = queue_index +BNXT_RX_VEC_START ;struct bnxt_cp_ring_info *nqr = bp->rxtx_nq_ring;uint8_t ring_type;int rc =0 ;
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 9/12 net/bnxt: enable interrupts after checking interface status from FW
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
Driver issues hwrm_if_change when the port is started. This command returns whether the fw has undergone a reset while port was down and driver does re-init of resources if fw has undergone a reset. Fix to enable interrupts only after this check.
--- drivers/net/bnxt/bnxt_ethdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/bnxt/bnxt_ethdev.c +1 -1
@@ -874,7 +874,6 @@ static int bnxt_dev_start_op(struct rte_eth_dev *eth_dev) bp->rx_cp_nr_rings, RTE_ETHDEV_QUEUE_STAT_CNTRS); } -bnxt_enable_int (bp); rc =bnxt_hwrm_if_change (bp,1 );if (!rc) {if (bp->flags &BNXT_FLAG_IF_CHANGE_HOT_FW_RESET_DONE ) {
@@ -883,6 +882,7 @@ static int bnxt_dev_start_op(struct rte_eth_dev *eth_dev)return rc; } } +bnxt_enable_int (bp); rc =bnxt_init_chip (bp);if (rc)
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 10/12 net/bnxt: fix to handle interface change status in port start only
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
Driver issues hwrm_if_change when the port is started and stopped. This command returns whether the fw has undergone a reset while port was down. Fix to handle hwrm_if_change status during port start only.
--- drivers/net/bnxt/bnxt_hwrm.c | 3 +++ 1 file changed, 3 insertions(+)
drivers/net/bnxt/bnxt_hwrm.c +3 -0
@@ -4858,6 +4858,9 @@ int bnxt_hwrm_if_change(struct bnxt *bp, bool up) flags =rte_le_to_cpu_32 (resp->flags);HWRM_UNLOCK (); +if (!up) +return 0 ; +if (flags &HWRM_FUNC_DRV_IF_CHANGE_OUTPUT_FLAGS_HOT_FW_RESET_DONE ) {PMD_DRV_LOG (INFO ,"FW reset happened while port was down\n" ); bp->flags |=BNXT_FLAG_IF_CHANGE_HOT_FW_RESET_DONE ;
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 11/12 net/bnxt: fix coding style issues
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
- Remove unnecessary new line
- Remove unnecessary blank lines
- Align rte_* header file inclusion at one place
--- drivers/net/bnxt/bnxt_ethdev.c | 1 - drivers/net/bnxt/bnxt_hwrm.c | 11 +++-------- 2 files changed, 3 insertions(+), 9 deletions(-)
drivers/net/bnxt/bnxt_ethdev.c +0 -1
@@ -3002,7 +3002,6 @@ bnxt_parse_fdir_filter(struct bnxt *bp,return -EINVAL ; } -if (fdir_mode ==RTE_FDIR_MODE_PERFECT_MAC_VLAN ) {rte_memcpy (filter->dst_macaddr, fdir->input.flow.mac_vlan_flow.mac_addr.addr_bytes,6 );
drivers/net/bnxt/bnxt_hwrm.c +3 -8
@@ -11,6 +11,7 @@#include <rte_malloc.h> #include <rte_memzone.h> #include <rte_version.h> +#include <rte_io.h> #include "bnxt.h" #include "bnxt_filter.h"
@@ -23,8 +24,6 @@#include "bnxt_vnic.h" #include "hsi_struct_def_dpdk.h" -#include <rte_io.h> -#define HWRM_SPEC_CODE_1_8_3 0x10803#define HWRM_VERSION_1_9_1 0x10901#define HWRM_VERSION_1_9_2 0x10903
@@ -1483,8 +1482,7 @@ int bnxt_hwrm_ring_grp_alloc(struct bnxt *bp, unsigned int idx)HWRM_CHECK_RESULT (); - bp->grp_info[idx].fw_grp_id = -rte_le_to_cpu_16 (resp->ring_group_id); + bp->grp_info[idx].fw_grp_id =rte_le_to_cpu_16 (resp->ring_group_id);HWRM_UNLOCK ();
@@ -1542,8 +1540,7 @@ int bnxt_hwrm_stat_ctx_alloc(struct bnxt *bp, struct bnxt_cp_ring_info *cpr, req.update_period_ms =rte_cpu_to_le_32 (0 ); - req.stats_dma_addr = -rte_cpu_to_le_64 (cpr->hw_stats_map); + req.stats_dma_addr =rte_cpu_to_le_64 (cpr->hw_stats_map); rc =bnxt_hwrm_send_message (bp, &req,sizeof (req),BNXT_USE_CHIMP_MB );
@@ -3633,7 +3630,6 @@ int bnxt_hwrm_ctx_qstats(struct bnxt *bp, uint32_t cid, int idx, stats->q_obytes[idx] +=rte_le_to_cpu_64 (resp->tx_bcast_bytes); } -HWRM_UNLOCK ();return rc;
@@ -4218,7 +4214,6 @@ int bnxt_hwrm_set_ntuple_filter(struct bnxt *bp,HWRM_CFA_NTUPLE_FILTER_ALLOC_INPUT_ENABLES_DST_ID ; req.dst_id =rte_cpu_to_le_16 (dst_id); -if (filter->ip_addr_type) { req.ip_addr_type = filter->ip_addr_type; enables |=
Signature · 1 line
2.20.1 (Apple Git-117)
PATCH v2 12/12 net/bnxt: remove unnecessary variable assignment
In reply to Ajit KhapardeFrom: Kalesh AP <[email protected]>
There is no need to assign return value to a temporary variable. Instead return error directly in case of failure.
--- drivers/net/bnxt/bnxt_rxq.c | 14 ++++++-------- 1 file changed, 6 insertions(+), 8 deletions(-)
drivers/net/bnxt/bnxt_rxq.c +6 -8
@@ -384,10 +384,9 @@ bnxt_rx_queue_intr_enable_op(struct rte_eth_dev *eth_dev, uint16_t queue_id)if (eth_dev->data->rx_queues) { rxq = eth_dev->data->rx_queues[queue_id]; -if (!rxq) { - rc = -EINVAL ; -return rc; - } +if (!rxq) +return -EINVAL ; + cpr = rxq->cp_ring;B_CP_DB_REARM (cpr, cpr->cp_raw_cons); }
@@ -408,10 +407,9 @@ bnxt_rx_queue_intr_disable_op(struct rte_eth_dev *eth_dev, uint16_t queue_id)if (eth_dev->data->rx_queues) { rxq = eth_dev->data->rx_queues[queue_id]; -if (!rxq) { - rc = -EINVAL ; -return rc; - } +if (!rxq) +return -EINVAL ; + cpr = rxq->cp_ring;B_CP_DB_DISARM (cpr); }
Signature · 1 line
2.20.1 (Apple Git-117)
Quoted 21 lines · Ajit Khaparde
This patch against dpdk-next-net mostly contains bug fixes for issues found during testing.
v1->v2:
-Split patch 09/10 into two and reworded the commit messages. -Rebased against latest dpdk-next-net
-Some of the patches were missing the Fixes tag. Added it to them.Ajit Khaparde (2):
net/bnxt: update trusted VF status only when it changes net/bnxt: get the default HWRM command timeout from firmwareKalesh AP (7):
net/bnxt: fix return checks and return values
net/bnxt: return error if setting link up fail
net/bnxt: remove redundant header file inclusion net/bnxt: enable interrupts after checking interface status from FW net/bnxt: fix to handle interface change status in port start only net/bnxt: fix coding style issues
net/bnxt: remove unnecessary variable assignment
Quoted 6 lines
Rahul Gupta (2):
net/bnxt: reduce cleanup time during reset recovery net/bnxt: change msix vector to queue mappingSantoshkumar Karanappa Rastapur (1):
net/bnxt: free default completion ring before VF configuration
Series applied to dpdk-next-net/master, thanks.