| Message ID | 20210421230929.434822-1-stephen@networkplumber.org (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Ajit Khaparde |
| Headers | show |
| Series | bnxt: skip getting statistics for queues not started | expand |
| Context | Check | Description |
|---|---|---|
| ci/iol-mellanox-Functional | success | Functional Testing PASS |
| ci/iol-mellanox-Performance | success | Performance Testing PASS |
| ci/iol-intel-Performance | success | Performance Testing PASS |
| ci/iol-abi-testing | success | Testing PASS |
| ci/iol-testing | success | Testing PASS |
| ci/github-robot | success | github build: passed |
| ci/intel-Testing | success | Testing PASS |
| ci/Intel-compilation | success | Compilation OK |
| ci/checkpatch | success | coding style OK |
On Wed, Apr 21, 2021 at 4:09 PM Stephen Hemminger <stephen@networkplumber.org> wrote: > > An application using rte_flow may define a large number of queues > but only use a small subset of them at any one time. > > Since querying the status of each queue requires a request/spin/reply > with the firmware, optimize by skipping the request for queues not running. > For those queues the statistics will be 0. > > This cuts the cost of single xstats query in half and has even > bigger gain for simple stats query. > > Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> Changed the commit headline to "net/bnxt: skip get statistics for stopped queues" Patch applied to dpdk-next-net-brcm. Thanks > --- > drivers/net/bnxt/bnxt_stats.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/net/bnxt/bnxt_stats.c b/drivers/net/bnxt/bnxt_stats.c > index bb4b2eee194e..11767e06d074 100644 > --- a/drivers/net/bnxt/bnxt_stats.c > +++ b/drivers/net/bnxt/bnxt_stats.c > @@ -528,6 +528,8 @@ int bnxt_stats_get_op(struct rte_eth_dev *eth_dev, > struct bnxt_rx_queue *rxq = bp->rx_queues[i]; > struct bnxt_cp_ring_info *cpr = rxq->cp_ring; > > + if (!rxq->rx_started) > + continue; > rc = bnxt_hwrm_ctx_qstats(bp, cpr->hw_stats_ctx_id, i, > bnxt_stats, 1); > if (unlikely(rc)) > @@ -543,6 +545,8 @@ int bnxt_stats_get_op(struct rte_eth_dev *eth_dev, > struct bnxt_tx_queue *txq = bp->tx_queues[i]; > struct bnxt_cp_ring_info *cpr = txq->cp_ring; > > + if (!txq->tx_started) > + continue; > rc = bnxt_hwrm_ctx_qstats(bp, cpr->hw_stats_ctx_id, i, > bnxt_stats, 0); > if (unlikely(rc)) > -- > 2.30.2 >
diff --git a/drivers/net/bnxt/bnxt_stats.c b/drivers/net/bnxt/bnxt_stats.c index bb4b2eee194e..11767e06d074 100644 --- a/drivers/net/bnxt/bnxt_stats.c +++ b/drivers/net/bnxt/bnxt_stats.c @@ -528,6 +528,8 @@ int bnxt_stats_get_op(struct rte_eth_dev *eth_dev, struct bnxt_rx_queue *rxq = bp->rx_queues[i]; struct bnxt_cp_ring_info *cpr = rxq->cp_ring; + if (!rxq->rx_started) + continue; rc = bnxt_hwrm_ctx_qstats(bp, cpr->hw_stats_ctx_id, i, bnxt_stats, 1); if (unlikely(rc)) @@ -543,6 +545,8 @@ int bnxt_stats_get_op(struct rte_eth_dev *eth_dev, struct bnxt_tx_queue *txq = bp->tx_queues[i]; struct bnxt_cp_ring_info *cpr = txq->cp_ring; + if (!txq->tx_started) + continue; rc = bnxt_hwrm_ctx_qstats(bp, cpr->hw_stats_ctx_id, i, bnxt_stats, 0); if (unlikely(rc))
An application using rte_flow may define a large number of queues but only use a small subset of them at any one time. Since querying the status of each queue requires a request/spin/reply with the firmware, optimize by skipping the request for queues not running. For those queues the statistics will be 0. This cuts the cost of single xstats query in half and has even bigger gain for simple stats query. Signed-off-by: Stephen Hemminger <stephen@networkplumber.org> --- drivers/net/bnxt/bnxt_stats.c | 4 ++++ 1 file changed, 4 insertions(+)