DPDK patches and discussions
 help / color / mirror / Atom feed
From: Ajit Khaparde <ajit.khaparde@broadcom.com>
To: dev@dpdk.org
Cc: ferruh.yigit@intel.com, Rahul Gupta <rahul.gupta@broadcom.com>,
	Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>,
	Kalesh Anakkur Purayil <kalesh-anakkur.purayil@broadcom.com>
Subject: [dpdk-dev] [PATCH 13/17] net/bnxt: handle flow flush handling
Date: Mon,  2 Sep 2019 19:18:57 -0700	[thread overview]
Message-ID: <20190903021901.25895-14-ajit.khaparde@broadcom.com> (raw)
In-Reply-To: <20190903021901.25895-1-ajit.khaparde@broadcom.com>

We are not freeing all the flows when a flow_flush is called.
Iterate through all the flows belonging to all the VNICs in use and
free the filters.

Signed-off-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Reviewed-by: Rahul Gupta <rahul.gupta@broadcom.com>
Reviewed-by: Venkat Duvvuru <venkatkumar.duvvuru@broadcom.com>
Reviewed-by: Kalesh Anakkur Purayil <kalesh-anakkur.purayil@broadcom.com>
---
 drivers/net/bnxt/bnxt_flow.c | 56 ++++++++++++++++++------------------
 1 file changed, 28 insertions(+), 28 deletions(-)

diff --git a/drivers/net/bnxt/bnxt_flow.c b/drivers/net/bnxt/bnxt_flow.c
index 71a5266ce..9d943cd14 100644
--- a/drivers/net/bnxt/bnxt_flow.c
+++ b/drivers/net/bnxt/bnxt_flow.c
@@ -907,7 +907,6 @@ bnxt_validate_and_parse_flow(struct rte_eth_dev *dev,
 	const struct rte_flow_action *act =
 		bnxt_flow_non_void_action(actions);
 	struct bnxt *bp = dev->data->dev_private;
-	struct rte_eth_conf *dev_conf = &bp->eth_dev->data->dev_conf;
 	const struct rte_flow_action_queue *act_q;
 	const struct rte_flow_action_rss *rss;
 	const struct rte_flow_action_vf *act_vf;
@@ -982,9 +981,6 @@ bnxt_validate_and_parse_flow(struct rte_eth_dev *dev,
 
 		rxq = bp->rx_queues[act_q->index];
 
-		if (!(dev_conf->rxmode.mq_mode & ETH_MQ_RX_RSS) && rxq)
-			goto use_vnic;
-
 		//if (!rxq ||
 		    //bp->vnic_info[0].fw_grp_ids[act_q->index] !=
 		    //INVALID_HW_RING_ID ||
@@ -1567,25 +1563,6 @@ bnxt_flow_create(struct rte_eth_dev *dev,
 
 	vnic = find_matching_vnic(bp, filter);
 done:
-	if (!ret) {
-		flow->filter = filter;
-		flow->vnic = vnic;
-		/* VNIC is set only in case of queue or RSS action */
-		if (vnic) {
-			/*
-			 * RxQ0 is not used for flow filters.
-			 */
-
-			if (update_flow) {
-				ret = -EXDEV;
-				goto free_flow;
-			}
-			STAILQ_INSERT_TAIL(&vnic->filter, filter, next);
-		}
-		PMD_DRV_LOG(ERR, "Successfully created flow.\n");
-		STAILQ_INSERT_TAIL(&vnic->flow_list, flow, next);
-		return flow;
-	}
 	if (!ret) {
 		flow->filter = filter;
 		flow->vnic = vnic;
@@ -1593,10 +1570,13 @@ bnxt_flow_create(struct rte_eth_dev *dev,
 			ret = -EXDEV;
 			goto free_flow;
 		}
+
+		STAILQ_INSERT_TAIL(&vnic->filter, filter, next);
 		PMD_DRV_LOG(ERR, "Successfully created flow.\n");
 		STAILQ_INSERT_TAIL(&vnic->flow_list, flow, next);
 		return flow;
 	}
+
 free_filter:
 	bnxt_free_filter(bp, filter);
 free_flow:
@@ -1727,6 +1707,7 @@ static int
 bnxt_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error)
 {
 	struct bnxt *bp = dev->data->dev_private;
+	struct bnxt_filter_info *filter = NULL;
 	struct bnxt_vnic_info *vnic;
 	struct rte_flow *flow;
 	unsigned int i;
@@ -1734,11 +1715,12 @@ bnxt_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error)
 
 	for (i = 0; i < bp->max_vnics; i++) {
 		vnic = &bp->vnic_info[i];
-		if (vnic->fw_vnic_id == INVALID_VNIC_ID)
+		if (vnic && vnic->fw_vnic_id == INVALID_VNIC_ID)
 			continue;
 
-		STAILQ_FOREACH(flow, &vnic->flow_list, next) {
-			struct bnxt_filter_info *filter = flow->filter;
+		while (!STAILQ_EMPTY(&vnic->flow_list)) {
+			flow = STAILQ_FIRST(&vnic->flow_list);
+			filter = flow->filter;
 
 			if (filter->filter_type ==
 			    HWRM_CFA_TUNNEL_REDIRECT_FILTER &&
@@ -1757,7 +1739,7 @@ bnxt_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error)
 				ret = bnxt_hwrm_clear_em_filter(bp, filter);
 			if (filter->filter_type == HWRM_CFA_NTUPLE_FILTER)
 				ret = bnxt_hwrm_clear_ntuple_filter(bp, filter);
-			else if (!i)
+			else if (i)
 				ret = bnxt_hwrm_clear_l2_filter(bp, filter);
 
 			if (ret) {
@@ -1770,10 +1752,28 @@ bnxt_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error)
 				return -rte_errno;
 			}
 done:
-			bnxt_free_filter(bp, filter);
 			STAILQ_REMOVE(&vnic->flow_list, flow,
 				      rte_flow, next);
+
+			STAILQ_REMOVE(&vnic->filter,
+				      filter,
+				      bnxt_filter_info,
+				      next);
+			bnxt_free_filter(bp, filter);
+
 			rte_free(flow);
+
+			/* If this was the last flow associated with this vnic,
+			 * switch the queue back to RSS pool.
+			 */
+			if (STAILQ_EMPTY(&vnic->flow_list)) {
+				rte_free(vnic->fw_grp_ids);
+				if (vnic->rx_queue_cnt > 1)
+					bnxt_hwrm_vnic_ctx_free(bp, vnic);
+				bnxt_hwrm_vnic_free(bp, vnic);
+				vnic->rx_queue_cnt = 0;
+				bp->nr_vnics--;
+			}
 		}
 	}
 
-- 
2.20.1 (Apple Git-117)


  parent reply	other threads:[~2019-09-03  2:21 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-03  2:18 [dpdk-dev] [PATCH 00/17] bnxt patchset to improve rte flow support Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 01/17] net/bnxt: return standard error codes for HWRM command Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 02/17] net/bnxt: refactor code to allow dynamic creation of VNIC Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 03/17] net/bnxt: allow flow creation when RSS is enabled Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 04/17] net/bnxt: add support to create SMAC and inner DMAC filters Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 05/17] net/bnxt: add support for RSS action Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 06/17] net/bnxt: parse priority attribute for flow creation Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 07/17] net/bnxt: delete and flush L2 filters cleanly Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 08/17] net/bnxt: cleanup vnic after flow validate Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 09/17] net/bnxt: allow only unicast MAC address filter creation Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 10/17] net/bnxt: properly handle ring cleanup in case of error Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 11/17] net/bnxt: check device is started before flow creation Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 12/17] net/bnxt: check for invalid VNIC ID in vnic tpa cfg Ajit Khaparde
2019-09-03  2:18 ` Ajit Khaparde [this message]
2019-09-03  2:18 ` [dpdk-dev] [PATCH 14/17] net/bnxt: validate RSS hash key length Ajit Khaparde
2019-09-03  2:18 ` [dpdk-dev] [PATCH 15/17] net/bnxt: handle cleanup if flow creation fails Ajit Khaparde
2019-09-03  2:19 ` [dpdk-dev] [PATCH 16/17] net/bnxt: synchronize between flow related functions Ajit Khaparde
2019-09-03  2:19 ` [dpdk-dev] [PATCH 17/17] net/bnxt: drop untagged frames when specified Ajit Khaparde
2019-10-01 17:02 ` [dpdk-dev] [PATCH 00/17] bnxt patchset to improve rte flow support Ferruh Yigit
2019-10-01 17:32   ` Ajit Khaparde

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20190903021901.25895-14-ajit.khaparde@broadcom.com \
    --to=ajit.khaparde@broadcom.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=kalesh-anakkur.purayil@broadcom.com \
    --cc=rahul.gupta@broadcom.com \
    --cc=venkatkumar.duvvuru@broadcom.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).