[dpdk-dev] [PATCH 2/4] net/ice: rework for generic flow enabling

[Wang, Ying A]

Hi, Xiaolong

> -----Original Message-----
> From: Ye, Xiaolong
> Sent: Saturday, September 7, 2019 12:13 AM
> To: Wang, Ying A <ying.a.wang at intel.com>
> Cc: Zhang, Qi Z <qi.z.zhang at intel.com>; Yang, Qiming
> <qiming.yang at intel.com>; dev at dpdk.org; Zhao1, Wei <wei.zhao1 at intel.com>
> Subject: Re: [PATCH 2/4] net/ice: rework for generic flow enabling
> 
> On 09/04, Ying Wang wrote:
> >The patch reworks the generic flow API (rte_flow) implementation.
> >It introduces an abstract layer which provides a unified interface for
> >low-level filter engine (switch, fdir, hash) to register supported
> >patterns and actions and implement flow validate/create/destroy/flush/
> >query activities.
> >
> >The patch also removes the existing switch filter implementation to
> >avoid compile error. Switch filter implementation for the new framework
> >will be added in the following patch.
> >
> >Signed-off-by: Ying Wang <ying.a.wang at intel.com>
> >---
> > drivers/net/ice/ice_ethdev.c        |  22 +-
> > drivers/net/ice/ice_ethdev.h        |  15 +-
> > drivers/net/ice/ice_generic_flow.c  | 768
> >+++++++++++++++--------------------
> > drivers/net/ice/ice_generic_flow.h  | 782
> >++++++++----------------------------
> > drivers/net/ice/ice_switch_filter.c | 511 -----------------------
> >drivers/net/ice/ice_switch_filter.h |  18 -
> > 6 files changed, 525 insertions(+), 1591 deletions(-)
> >
> >diff --git a/drivers/net/ice/ice_ethdev.c
> >b/drivers/net/ice/ice_ethdev.c index 4e0645db1..647aca3ed 100644
> >--- a/drivers/net/ice/ice_ethdev.c
> >+++ b/drivers/net/ice/ice_ethdev.c
> >@@ -15,7 +15,7 @@
> > #include "base/ice_dcb.h"
> > #include "ice_ethdev.h"
> > #include "ice_rxtx.h"
> >-#include "ice_switch_filter.h"
> >+#include "ice_generic_flow.h"
> >
> > /* devargs */
> > #define ICE_SAFE_MODE_SUPPORT_ARG "safe-mode-support"
> >@@ -1677,7 +1677,11 @@ ice_dev_init(struct rte_eth_dev *dev)
> > 	/* get base queue pairs index  in the device */
> > 	ice_base_queue_get(pf);
> >
> >-	TAILQ_INIT(&pf->flow_list);
> >+	ret = ice_flow_init(ad);
> >+	if (ret) {
> >+		PMD_INIT_LOG(ERR, "Failed to initialize flow");
> >+		return ret;
> >+	}
> >
> > 	return 0;
> >
> >@@ -1796,6 +1800,8 @@ ice_dev_close(struct rte_eth_dev *dev)  {
> > 	struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
> > 	struct ice_hw *hw = ICE_DEV_PRIVATE_TO_HW(dev->data-
> >dev_private);
> >+	struct ice_adapter *ad =
> >+		ICE_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
> >
> > 	/* Since stop will make link down, then the link event will be
> > 	 * triggered, disable the irq firstly to avoid the port_infoe etc @@
> >-1806,6 +1812,8 @@ ice_dev_close(struct rte_eth_dev *dev)
> >
> > 	ice_dev_stop(dev);
> >
> >+	ice_flow_uninit(ad);
> >+
> > 	/* release all queue resource */
> > 	ice_free_queues(dev);
> >
> >@@ -1822,8 +1830,6 @@ ice_dev_uninit(struct rte_eth_dev *dev)  {
> > 	struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev);
> > 	struct rte_intr_handle *intr_handle = &pci_dev->intr_handle;
> >-	struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
> >-	struct rte_flow *p_flow;
> >
> > 	ice_dev_close(dev);
> >
> >@@ -1840,14 +1846,6 @@ ice_dev_uninit(struct rte_eth_dev *dev)
> > 	/* unregister callback func from eal lib */
> > 	rte_intr_callback_unregister(intr_handle,
> > 				     ice_interrupt_handler, dev);
> >-
> >-	/* Remove all flows */
> >-	while ((p_flow = TAILQ_FIRST(&pf->flow_list))) {
> >-		TAILQ_REMOVE(&pf->flow_list, p_flow, node);
> >-		ice_free_switch_filter_rule(p_flow->rule);
> >-		rte_free(p_flow);
> >-	}
> >-
> > 	return 0;
> > }
> >
> >diff --git a/drivers/net/ice/ice_ethdev.h
> >b/drivers/net/ice/ice_ethdev.h index 9bf5de08d..d1d07641d 100644
> >--- a/drivers/net/ice/ice_ethdev.h
> >+++ b/drivers/net/ice/ice_ethdev.h
> >@@ -241,16 +241,14 @@ struct ice_vsi {
> > 	bool offset_loaded;
> > };
> >
> >-extern const struct rte_flow_ops ice_flow_ops;
> >-
> >-/* Struct to store flow created. */
> >-struct rte_flow {
> >-	TAILQ_ENTRY(rte_flow) node;
> >-	void *rule;
> >-};
> >
> >+struct rte_flow;
> > TAILQ_HEAD(ice_flow_list, rte_flow);
> >
> >+
> >+struct ice_flow_parser;
> >+TAILQ_HEAD(ice_parser_list, ice_flow_parser);
> >+
> > struct ice_pf {
> > 	struct ice_adapter *adapter; /* The adapter this PF associate to */
> > 	struct ice_vsi *main_vsi; /* pointer to main VSI structure */ @@
> >-278,6 +276,9 @@ struct ice_pf {
> > 	bool offset_loaded;
> > 	bool adapter_stopped;
> > 	struct ice_flow_list flow_list;
> >+	struct ice_parser_list rss_parser_list;
> >+	struct ice_parser_list perm_parser_list;
> >+	struct ice_parser_list dist_parser_list;
> > };
> >
> > /**
> >diff --git a/drivers/net/ice/ice_generic_flow.c
> >b/drivers/net/ice/ice_generic_flow.c
> >index 1c0adc779..aa11d6170 100644
> >--- a/drivers/net/ice/ice_generic_flow.c
> >+++ b/drivers/net/ice/ice_generic_flow.c
> >@@ -17,7 +17,22 @@
> >
> > #include "ice_ethdev.h"
> > #include "ice_generic_flow.h"
> >-#include "ice_switch_filter.h"
> >+
> >+/**
> >+ * Non-pipeline mode, fdir and swith both used as distributor,
> 
> s/swith/switch

OK, will fix it in v2.

> 
> >+ * fdir used first, switch used as fdir's backup.
> >+ */
> >+#define ICE_FLOW_CLASSIFY_STAGE_DISTRIBUTOR_ONLY 0 /*Pipeline mode,
> >+switch used at permission stage*/ #define
> >+ICE_FLOW_CLASSIFY_STAGE_PERMISSION 1 /*Pipeline mode, fdir used at
> >+distributor stage*/ #define ICE_FLOW_CLASSIFY_STAGE_DISTRIBUTOR 2
> >+
> >+static int ice_pipeline_stage =
> >+		ICE_FLOW_CLASSIFY_STAGE_DISTRIBUTOR_ONLY;
> >+
> >+static struct ice_engine_list engine_list =
> >+		TAILQ_HEAD_INITIALIZER(engine_list);
> >
> > static int ice_flow_validate(struct rte_eth_dev *dev,
> > 		const struct rte_flow_attr *attr,
> >@@ -34,17 +49,153 @@ static int ice_flow_destroy(struct rte_eth_dev *dev,
> > 		struct rte_flow_error *error);
> > static int ice_flow_flush(struct rte_eth_dev *dev,
> > 		struct rte_flow_error *error);
> >+static int ice_flow_query_count(struct rte_eth_dev *dev,
> >+		struct rte_flow *flow,
> >+		const struct rte_flow_action *actions,
> >+		void *data,
> >+		struct rte_flow_error *error);
> >
> > const struct rte_flow_ops ice_flow_ops = {
> > 	.validate = ice_flow_validate,
> > 	.create = ice_flow_create,
> > 	.destroy = ice_flow_destroy,
> > 	.flush = ice_flow_flush,
> >+	.query = ice_flow_query_count,
> > };
> >
> >+
> >+void
> >+ice_register_flow_engine(struct ice_flow_engine *engine) {
> >+	TAILQ_INSERT_TAIL(&engine_list, engine, node); }
> >+
> >+int
> >+ice_flow_init(struct ice_adapter *ad)
> >+{
> >+	int ret = 0;
> >+	struct ice_pf *pf = &ad->pf;
> >+	void *temp;
> >+	struct ice_flow_engine *engine = NULL;
> >+
> >+	TAILQ_INIT(&pf->flow_list);
> >+	TAILQ_INIT(&pf->rss_parser_list);
> >+	TAILQ_INIT(&pf->perm_parser_list);
> >+	TAILQ_INIT(&pf->dist_parser_list);
> >+
> >+	TAILQ_FOREACH_SAFE(engine, &engine_list, node, temp) {
> >+		if (engine->init == NULL)
> >+			return -EINVAL;
> >+
> >+		ret = engine->init(ad);
> >+		if (ret)
> >+			return ret;
> >+	}
> >+	return 0;
> >+}
> >+
> >+void
> >+ice_flow_uninit(struct ice_adapter *ad) {
> >+	struct ice_pf *pf = &ad->pf;
> >+	struct ice_flow_engine *engine;
> >+	struct rte_flow *p_flow;
> >+	struct ice_flow_parser *p_parser;
> >+	void *temp;
> >+
> >+	TAILQ_FOREACH_SAFE(engine, &engine_list, node, temp) {
> >+		if (engine->uninit)
> >+			engine->uninit(ad);
> >+	}
> >+
> >+	/* Remove all flows */
> >+	while ((p_flow = TAILQ_FIRST(&pf->flow_list))) {
> >+		TAILQ_REMOVE(&pf->flow_list, p_flow, node);
> >+		if (p_flow->engine->free)
> >+			p_flow->engine->free(p_flow);
> >+		rte_free(p_flow);
> >+	}
> >+
> >+	/* Cleanup parser list */
> >+	while ((p_parser = TAILQ_FIRST(&pf->rss_parser_list)))
> >+		TAILQ_REMOVE(&pf->rss_parser_list, p_parser, node);
> >+
> >+	while ((p_parser = TAILQ_FIRST(&pf->perm_parser_list)))
> >+		TAILQ_REMOVE(&pf->perm_parser_list, p_parser, node);
> >+
> >+	while ((p_parser = TAILQ_FIRST(&pf->dist_parser_list)))
> >+		TAILQ_REMOVE(&pf->dist_parser_list, p_parser, node); }
> >+
> >+int
> >+ice_register_parser(struct ice_flow_parser *parser,
> >+		struct ice_adapter *ad)
> >+{
> >+	struct ice_parser_list *list = NULL;
> >+	struct ice_pf *pf = &ad->pf;
> >+
> >+	switch (parser->stage) {
> >+	case ICE_FLOW_STAGE_RSS:
> >+		list = &pf->rss_parser_list;
> >+		break;
> >+	case ICE_FLOW_STAGE_PERMISSION:
> >+		list = &pf->perm_parser_list;
> >+		break;
> >+	case ICE_FLOW_STAGE_DISTRIBUTOR:
> >+		list = &pf->dist_parser_list;
> >+		break;
> >+	default:
> >+		return -EINVAL;
> >+	}
> >+
> >+	if (ad->devargs.pipeline_mode_support)
> >+		TAILQ_INSERT_TAIL(list, parser, node);
> >+	else {
> >+		if (parser->engine->type == ICE_FLOW_ENGINE_SWITCH
> >+			|| parser->engine->type == ICE_FLOW_ENGINE_HASH)
> >+			TAILQ_INSERT_TAIL(list, parser, node);
> >+		else if (parser->engine->type == ICE_FLOW_ENGINE_FDIR)
> >+			TAILQ_INSERT_HEAD(list, parser, node);
> >+		else
> >+			return -EINVAL;
> >+	}
> >+	return 0;
> >+}
> >+
> >+void
> >+ice_unregister_parser(struct ice_flow_parser *parser,
> >+		struct ice_adapter *ad)
> >+{
> >+	struct ice_pf *pf = &ad->pf;
> >+	struct ice_parser_list *list;
> >+	struct ice_flow_parser *p_parser;
> >+	void *temp;
> >+
> >+	switch (parser->stage) {
> >+	case ICE_FLOW_STAGE_RSS:
> >+		list = &pf->rss_parser_list;
> >+		break;
> >+	case ICE_FLOW_STAGE_PERMISSION:
> >+		list = &pf->perm_parser_list;
> >+		break;
> >+	case ICE_FLOW_STAGE_DISTRIBUTOR:
> >+		list = &pf->dist_parser_list;
> >+		break;
> >+	default:
> >+		return;
> >+	}
> 
> The switch blocks in above functions are the same, it's better to use a common
> function to reduce the duplicated code.

The switch blocks in the above two functions have little difference in the default behavior, one is return -EINVAL, the other is just return, for register/unregister funcs have different return value types. So, Can I just keep this format?

> 
> Thanks,
> Xiaolong

Thanks,
Ying