* [PATCH] net/bnxt: fix ring calculation for representors
@ 2022-02-09 5:58 Ajit Khaparde
2022-02-09 15:37 ` Ajit Khaparde
0 siblings, 1 reply; 5+ messages in thread
From: Ajit Khaparde @ 2022-02-09 5:58 UTC (permalink / raw)
To: dev; +Cc: stable, Somnath Kotur, Kalesh AP
[-- Attachment #1: Type: text/plain, Size: 3504 bytes --]
Currently the Tx and Rx ring count for representors is fixed.
It does not consider the number of rings created for the parent
function. And that can cause issues not only during initialization
but while running traffic.
Instead check the number of rings created for the parent function
while configuring the ring resources for representors.
Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
Cc: stable@dpdk.org
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
---
drivers/net/bnxt/bnxt.h | 2 +-
drivers/net/bnxt/bnxt_reps.c | 19 ++++++++++---------
2 files changed, 11 insertions(+), 10 deletions(-)
diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index 433f1c80be..0cbb58b2cf 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -831,7 +831,7 @@ struct bnxt {
uint16_t max_tx_rings;
uint16_t max_rx_rings;
#define MAX_STINGRAY_RINGS 236U
-#define BNXT_MAX_VF_REP_RINGS 8
+#define BNXT_MAX_VF_REP_RINGS 8U
uint16_t max_nq_rings;
uint16_t max_l2_ctx;
diff --git a/drivers/net/bnxt/bnxt_reps.c b/drivers/net/bnxt/bnxt_reps.c
index 5e140f0cdb..9ae8848382 100644
--- a/drivers/net/bnxt/bnxt_reps.c
+++ b/drivers/net/bnxt/bnxt_reps.c
@@ -548,7 +548,8 @@ int bnxt_rep_dev_info_get_op(struct rte_eth_dev *eth_dev,
dev_info->max_mac_addrs = parent_bp->max_l2_ctx;
dev_info->max_hash_mac_addrs = 0;
- max_rx_rings = BNXT_MAX_VF_REP_RINGS;
+ max_rx_rings = RTE_MIN(parent_bp->rx_nr_rings, BNXT_MAX_VF_REP_RINGS);
+
/* 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;
@@ -629,10 +630,10 @@ int bnxt_rep_rx_queue_setup_op(struct rte_eth_dev *eth_dev,
struct rte_mbuf **buf_ring;
int rc = 0;
- if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
+ if (queue_idx >= rep_bp->rx_nr_rings) {
PMD_DRV_LOG(ERR,
"Cannot create Rx ring %d. %d rings available\n",
- queue_idx, BNXT_MAX_VF_REP_RINGS);
+ queue_idx, rep_bp->rx_nr_rings);
return -EINVAL;
}
@@ -729,10 +730,10 @@ int bnxt_rep_tx_queue_setup_op(struct rte_eth_dev *eth_dev,
struct bnxt_tx_queue *parent_txq, *txq;
struct bnxt_vf_rep_tx_queue *vfr_txq;
- if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
+ if (queue_idx >= rep_bp->rx_nr_rings) {
PMD_DRV_LOG(ERR,
"Cannot create Tx rings %d. %d rings available\n",
- queue_idx, BNXT_MAX_VF_REP_RINGS);
+ queue_idx, rep_bp->rx_nr_rings);
return -EINVAL;
}
@@ -805,10 +806,10 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
struct rte_eth_stats *stats)
{
struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
- int i;
+ unsigned int i;
memset(stats, 0, sizeof(*stats));
- for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
+ for (i = 0; i < rep_bp->rx_nr_rings; i++) {
stats->obytes += rep_bp->tx_bytes[i];
stats->opackets += rep_bp->tx_pkts[i];
stats->ibytes += rep_bp->rx_bytes[i];
@@ -828,9 +829,9 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
int bnxt_rep_stats_reset_op(struct rte_eth_dev *eth_dev)
{
struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
- int i;
+ unsigned int i;
- for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
+ for (i = 0; i < rep_bp->rx_nr_rings; i++) {
rep_bp->tx_pkts[i] = 0;
rep_bp->tx_bytes[i] = 0;
rep_bp->rx_pkts[i] = 0;
--
2.32.0 (Apple Git-132)
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] net/bnxt: fix ring calculation for representors
2022-02-09 5:58 [PATCH] net/bnxt: fix ring calculation for representors Ajit Khaparde
@ 2022-02-09 15:37 ` Ajit Khaparde
2022-02-09 17:34 ` Ajit Khaparde
0 siblings, 1 reply; 5+ messages in thread
From: Ajit Khaparde @ 2022-02-09 15:37 UTC (permalink / raw)
To: dpdk-dev; +Cc: dpdk stable, Somnath Kotur, Kalesh AP, Ferruh Yigit
[-- Attachment #1: Type: text/plain, Size: 4333 bytes --]
On Tue, Feb 8, 2022 at 9:58 PM Ajit Khaparde <ajit.khaparde@broadcom.com> wrote:
>
> Currently the Tx and Rx ring count for representors is fixed.
> It does not consider the number of rings created for the parent
> function. And that can cause issues not only during initialization
> but while running traffic.
> Instead check the number of rings created for the parent function
> while configuring the ring resources for representors.
>
> Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
> Cc: stable@dpdk.org
>
> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Patch applied to dpdk-next-net-brcm.
> ---
> drivers/net/bnxt/bnxt.h | 2 +-
> drivers/net/bnxt/bnxt_reps.c | 19 ++++++++++---------
> 2 files changed, 11 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
> index 433f1c80be..0cbb58b2cf 100644
> --- a/drivers/net/bnxt/bnxt.h
> +++ b/drivers/net/bnxt/bnxt.h
> @@ -831,7 +831,7 @@ struct bnxt {
> uint16_t max_tx_rings;
> uint16_t max_rx_rings;
> #define MAX_STINGRAY_RINGS 236U
> -#define BNXT_MAX_VF_REP_RINGS 8
> +#define BNXT_MAX_VF_REP_RINGS 8U
>
> uint16_t max_nq_rings;
> uint16_t max_l2_ctx;
> diff --git a/drivers/net/bnxt/bnxt_reps.c b/drivers/net/bnxt/bnxt_reps.c
> index 5e140f0cdb..9ae8848382 100644
> --- a/drivers/net/bnxt/bnxt_reps.c
> +++ b/drivers/net/bnxt/bnxt_reps.c
> @@ -548,7 +548,8 @@ int bnxt_rep_dev_info_get_op(struct rte_eth_dev *eth_dev,
> dev_info->max_mac_addrs = parent_bp->max_l2_ctx;
> dev_info->max_hash_mac_addrs = 0;
>
> - max_rx_rings = BNXT_MAX_VF_REP_RINGS;
> + max_rx_rings = RTE_MIN(parent_bp->rx_nr_rings, BNXT_MAX_VF_REP_RINGS);
> +
> /* 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;
> @@ -629,10 +630,10 @@ int bnxt_rep_rx_queue_setup_op(struct rte_eth_dev *eth_dev,
> struct rte_mbuf **buf_ring;
> int rc = 0;
>
> - if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
> + if (queue_idx >= rep_bp->rx_nr_rings) {
> PMD_DRV_LOG(ERR,
> "Cannot create Rx ring %d. %d rings available\n",
> - queue_idx, BNXT_MAX_VF_REP_RINGS);
> + queue_idx, rep_bp->rx_nr_rings);
> return -EINVAL;
> }
>
> @@ -729,10 +730,10 @@ int bnxt_rep_tx_queue_setup_op(struct rte_eth_dev *eth_dev,
> struct bnxt_tx_queue *parent_txq, *txq;
> struct bnxt_vf_rep_tx_queue *vfr_txq;
>
> - if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
> + if (queue_idx >= rep_bp->rx_nr_rings) {
> PMD_DRV_LOG(ERR,
> "Cannot create Tx rings %d. %d rings available\n",
> - queue_idx, BNXT_MAX_VF_REP_RINGS);
> + queue_idx, rep_bp->rx_nr_rings);
> return -EINVAL;
> }
>
> @@ -805,10 +806,10 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
> struct rte_eth_stats *stats)
> {
> struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
> - int i;
> + unsigned int i;
>
> memset(stats, 0, sizeof(*stats));
> - for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
> + for (i = 0; i < rep_bp->rx_nr_rings; i++) {
> stats->obytes += rep_bp->tx_bytes[i];
> stats->opackets += rep_bp->tx_pkts[i];
> stats->ibytes += rep_bp->rx_bytes[i];
> @@ -828,9 +829,9 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
> int bnxt_rep_stats_reset_op(struct rte_eth_dev *eth_dev)
> {
> struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
> - int i;
> + unsigned int i;
>
> - for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
> + for (i = 0; i < rep_bp->rx_nr_rings; i++) {
> rep_bp->tx_pkts[i] = 0;
> rep_bp->tx_bytes[i] = 0;
> rep_bp->rx_pkts[i] = 0;
> --
> 2.32.0 (Apple Git-132)
>
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH] net/bnxt: fix ring calculation for representors
2022-02-09 15:37 ` Ajit Khaparde
@ 2022-02-09 17:34 ` Ajit Khaparde
2022-02-09 17:42 ` [PATCH v2] " Ajit Khaparde
0 siblings, 1 reply; 5+ messages in thread
From: Ajit Khaparde @ 2022-02-09 17:34 UTC (permalink / raw)
To: dpdk-dev; +Cc: dpdk stable, Somnath Kotur, Kalesh AP, Ferruh Yigit
[-- Attachment #1: Type: text/plain, Size: 4661 bytes --]
On Wed, Feb 9, 2022 at 7:37 AM Ajit Khaparde <ajit.khaparde@broadcom.com> wrote:
>
> On Tue, Feb 8, 2022 at 9:58 PM Ajit Khaparde <ajit.khaparde@broadcom.com> wrote:
> >
> > Currently the Tx and Rx ring count for representors is fixed.
> > It does not consider the number of rings created for the parent
> > function. And that can cause issues not only during initialization
> > but while running traffic.
> > Instead check the number of rings created for the parent function
> > while configuring the ring resources for representors.
> >
> > Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
> > Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
> > Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
> Patch applied to dpdk-next-net-brcm.
Reverting and sending a fresh version.
>
> > ---
> > drivers/net/bnxt/bnxt.h | 2 +-
> > drivers/net/bnxt/bnxt_reps.c | 19 ++++++++++---------
> > 2 files changed, 11 insertions(+), 10 deletions(-)
> >
> > diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
> > index 433f1c80be..0cbb58b2cf 100644
> > --- a/drivers/net/bnxt/bnxt.h
> > +++ b/drivers/net/bnxt/bnxt.h
> > @@ -831,7 +831,7 @@ struct bnxt {
> > uint16_t max_tx_rings;
> > uint16_t max_rx_rings;
> > #define MAX_STINGRAY_RINGS 236U
> > -#define BNXT_MAX_VF_REP_RINGS 8
> > +#define BNXT_MAX_VF_REP_RINGS 8U
> >
> > uint16_t max_nq_rings;
> > uint16_t max_l2_ctx;
> > diff --git a/drivers/net/bnxt/bnxt_reps.c b/drivers/net/bnxt/bnxt_reps.c
> > index 5e140f0cdb..9ae8848382 100644
> > --- a/drivers/net/bnxt/bnxt_reps.c
> > +++ b/drivers/net/bnxt/bnxt_reps.c
> > @@ -548,7 +548,8 @@ int bnxt_rep_dev_info_get_op(struct rte_eth_dev *eth_dev,
> > dev_info->max_mac_addrs = parent_bp->max_l2_ctx;
> > dev_info->max_hash_mac_addrs = 0;
> >
> > - max_rx_rings = BNXT_MAX_VF_REP_RINGS;
> > + max_rx_rings = RTE_MIN(parent_bp->rx_nr_rings, BNXT_MAX_VF_REP_RINGS);
> > +
> > /* 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;
> > @@ -629,10 +630,10 @@ int bnxt_rep_rx_queue_setup_op(struct rte_eth_dev *eth_dev,
> > struct rte_mbuf **buf_ring;
> > int rc = 0;
> >
> > - if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
> > + if (queue_idx >= rep_bp->rx_nr_rings) {
> > PMD_DRV_LOG(ERR,
> > "Cannot create Rx ring %d. %d rings available\n",
> > - queue_idx, BNXT_MAX_VF_REP_RINGS);
> > + queue_idx, rep_bp->rx_nr_rings);
> > return -EINVAL;
> > }
> >
> > @@ -729,10 +730,10 @@ int bnxt_rep_tx_queue_setup_op(struct rte_eth_dev *eth_dev,
> > struct bnxt_tx_queue *parent_txq, *txq;
> > struct bnxt_vf_rep_tx_queue *vfr_txq;
> >
> > - if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
> > + if (queue_idx >= rep_bp->rx_nr_rings) {
> > PMD_DRV_LOG(ERR,
> > "Cannot create Tx rings %d. %d rings available\n",
> > - queue_idx, BNXT_MAX_VF_REP_RINGS);
> > + queue_idx, rep_bp->rx_nr_rings);
> > return -EINVAL;
> > }
> >
> > @@ -805,10 +806,10 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
> > struct rte_eth_stats *stats)
> > {
> > struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
> > - int i;
> > + unsigned int i;
> >
> > memset(stats, 0, sizeof(*stats));
> > - for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
> > + for (i = 0; i < rep_bp->rx_nr_rings; i++) {
> > stats->obytes += rep_bp->tx_bytes[i];
> > stats->opackets += rep_bp->tx_pkts[i];
> > stats->ibytes += rep_bp->rx_bytes[i];
> > @@ -828,9 +829,9 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
> > int bnxt_rep_stats_reset_op(struct rte_eth_dev *eth_dev)
> > {
> > struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
> > - int i;
> > + unsigned int i;
> >
> > - for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
> > + for (i = 0; i < rep_bp->rx_nr_rings; i++) {
> > rep_bp->tx_pkts[i] = 0;
> > rep_bp->tx_bytes[i] = 0;
> > rep_bp->rx_pkts[i] = 0;
> > --
> > 2.32.0 (Apple Git-132)
> >
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread* [PATCH v2] net/bnxt: fix ring calculation for representors
2022-02-09 17:34 ` Ajit Khaparde
@ 2022-02-09 17:42 ` Ajit Khaparde
2022-02-09 18:19 ` Ajit Khaparde
0 siblings, 1 reply; 5+ messages in thread
From: Ajit Khaparde @ 2022-02-09 17:42 UTC (permalink / raw)
To: dev; +Cc: stable, Somnath Kotur, Kalesh AP
[-- Attachment #1: Type: text/plain, Size: 3783 bytes --]
Currently the Tx and Rx ring count for representors is fixed.
It does not consider the number of rings created for the parent
function. And that can cause issues not only during initialization
but while running traffic.
Instead check the number of rings created for the parent function
while configuring the ring resources for representors.
In some cases VF rep ring init may happen before the parent function's
rings have been setup. And this can cause representor ring count to be
configured as 0. In such cases, initialize the VF representor
ring count to 8.
Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
Cc: stable@dpdk.org
Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
---
drivers/net/bnxt/bnxt.h | 2 +-
drivers/net/bnxt/bnxt_reps.c | 21 ++++++++++++---------
2 files changed, 13 insertions(+), 10 deletions(-)
diff --git a/drivers/net/bnxt/bnxt.h b/drivers/net/bnxt/bnxt.h
index 433f1c80be..0cbb58b2cf 100644
--- a/drivers/net/bnxt/bnxt.h
+++ b/drivers/net/bnxt/bnxt.h
@@ -831,7 +831,7 @@ struct bnxt {
uint16_t max_tx_rings;
uint16_t max_rx_rings;
#define MAX_STINGRAY_RINGS 236U
-#define BNXT_MAX_VF_REP_RINGS 8
+#define BNXT_MAX_VF_REP_RINGS 8U
uint16_t max_nq_rings;
uint16_t max_l2_ctx;
diff --git a/drivers/net/bnxt/bnxt_reps.c b/drivers/net/bnxt/bnxt_reps.c
index 5e140f0cdb..e773932681 100644
--- a/drivers/net/bnxt/bnxt_reps.c
+++ b/drivers/net/bnxt/bnxt_reps.c
@@ -548,7 +548,10 @@ int bnxt_rep_dev_info_get_op(struct rte_eth_dev *eth_dev,
dev_info->max_mac_addrs = parent_bp->max_l2_ctx;
dev_info->max_hash_mac_addrs = 0;
- max_rx_rings = BNXT_MAX_VF_REP_RINGS;
+ max_rx_rings = parent_bp->rx_nr_rings ?
+ RTE_MIN(parent_bp->rx_nr_rings, BNXT_MAX_VF_REP_RINGS) :
+ BNXT_MAX_VF_REP_RINGS;
+
/* 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;
@@ -629,10 +632,10 @@ int bnxt_rep_rx_queue_setup_op(struct rte_eth_dev *eth_dev,
struct rte_mbuf **buf_ring;
int rc = 0;
- if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
+ if (queue_idx >= rep_bp->rx_nr_rings) {
PMD_DRV_LOG(ERR,
"Cannot create Rx ring %d. %d rings available\n",
- queue_idx, BNXT_MAX_VF_REP_RINGS);
+ queue_idx, rep_bp->rx_nr_rings);
return -EINVAL;
}
@@ -729,10 +732,10 @@ int bnxt_rep_tx_queue_setup_op(struct rte_eth_dev *eth_dev,
struct bnxt_tx_queue *parent_txq, *txq;
struct bnxt_vf_rep_tx_queue *vfr_txq;
- if (queue_idx >= BNXT_MAX_VF_REP_RINGS) {
+ if (queue_idx >= rep_bp->rx_nr_rings) {
PMD_DRV_LOG(ERR,
"Cannot create Tx rings %d. %d rings available\n",
- queue_idx, BNXT_MAX_VF_REP_RINGS);
+ queue_idx, rep_bp->rx_nr_rings);
return -EINVAL;
}
@@ -805,10 +808,10 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
struct rte_eth_stats *stats)
{
struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
- int i;
+ unsigned int i;
memset(stats, 0, sizeof(*stats));
- for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
+ for (i = 0; i < rep_bp->rx_nr_rings; i++) {
stats->obytes += rep_bp->tx_bytes[i];
stats->opackets += rep_bp->tx_pkts[i];
stats->ibytes += rep_bp->rx_bytes[i];
@@ -828,9 +831,9 @@ int bnxt_rep_stats_get_op(struct rte_eth_dev *eth_dev,
int bnxt_rep_stats_reset_op(struct rte_eth_dev *eth_dev)
{
struct bnxt_representor *rep_bp = eth_dev->data->dev_private;
- int i;
+ unsigned int i;
- for (i = 0; i < BNXT_MAX_VF_REP_RINGS; i++) {
+ for (i = 0; i < rep_bp->rx_nr_rings; i++) {
rep_bp->tx_pkts[i] = 0;
rep_bp->tx_bytes[i] = 0;
rep_bp->rx_pkts[i] = 0;
--
2.32.0 (Apple Git-132)
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread* Re: [PATCH v2] net/bnxt: fix ring calculation for representors
2022-02-09 17:42 ` [PATCH v2] " Ajit Khaparde
@ 2022-02-09 18:19 ` Ajit Khaparde
0 siblings, 0 replies; 5+ messages in thread
From: Ajit Khaparde @ 2022-02-09 18:19 UTC (permalink / raw)
To: dpdk-dev; +Cc: dpdk stable, Somnath Kotur, Kalesh AP
[-- Attachment #1: Type: text/plain, Size: 1054 bytes --]
On Wed, Feb 9, 2022 at 9:42 AM Ajit Khaparde <ajit.khaparde@broadcom.com> wrote:
>
> Currently the Tx and Rx ring count for representors is fixed.
> It does not consider the number of rings created for the parent
> function. And that can cause issues not only during initialization
> but while running traffic.
> Instead check the number of rings created for the parent function
> while configuring the ring resources for representors.
> In some cases VF rep ring init may happen before the parent function's
> rings have been setup. And this can cause representor ring count to be
> configured as 0. In such cases, initialize the VF representor
> ring count to 8.
>
> Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
> Cc: stable@dpdk.org
>
> Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
> Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
v1->v2: sets a default queue count if under some cases we find to be 0.
Patch applied to dpdk-next-net-brcm.
[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4218 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2022-02-09 18:19 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-02-09 5:58 [PATCH] net/bnxt: fix ring calculation for representors Ajit Khaparde
2022-02-09 15:37 ` Ajit Khaparde
2022-02-09 17:34 ` Ajit Khaparde
2022-02-09 17:42 ` [PATCH v2] " Ajit Khaparde
2022-02-09 18:19 ` Ajit Khaparde
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).