net: ethernet: ti: am65-cpsw: Introduce multi queue Rx

am65-cpsw can support up to 8 queues at Rx.
Use a macro AM65_CPSW_MAX_RX_QUEUES to indicate that.
As there is only one DMA channel for RX traffic, the
8 queues come as 8 flows in that channel.

By default, we will start with 1 flow as defined by the
macro AM65_CPSW_DEFAULT_RX_CHN_FLOWS.

User can change the number of flows by ethtool like so
'ethtool -L ethx rx <N>'

All traffic will still come on flow 0. To get traffic on
different flows the Classifiers will need to be set up.
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/ti/am65-cpsw-ethtool.c

@@ -427,9 +427,9 @@ static void am65_cpsw_get_channels(struct net_device *ndev,
 {
 	struct am65_cpsw_common *common = am65_ndev_to_common(ndev);
 
-	ch->max_rx = AM65_CPSW_MAX_RX_QUEUES;
-	ch->max_tx = AM65_CPSW_MAX_TX_QUEUES;
-	ch->rx_count = AM65_CPSW_MAX_RX_QUEUES;
+	ch->max_rx = AM65_CPSW_MAX_QUEUES;
+	ch->max_tx = AM65_CPSW_MAX_QUEUES;
+	ch->rx_count = common->rx_ch_num_flows;
 	ch->tx_count = common->tx_ch_num;
 }
 
@@ -447,9 +447,8 @@ static int am65_cpsw_set_channels(struct net_device *ndev,
 	if (common->usage_count)
 		return -EBUSY;
 
-	am65_cpsw_nuss_remove_tx_chns(common);
-
-	return am65_cpsw_nuss_update_tx_chns(common, chs->tx_count);
+	return am65_cpsw_nuss_update_tx_rx_chns(common, chs->tx_count,
+						chs->rx_count);
 }
 
 static void
@@ -913,80 +912,64 @@ static void am65_cpsw_get_mm_stats(struct net_device *ndev,
 	s->MACMergeHoldCount = readl(base + AM65_CPSW_STATN_IET_TX_HOLD);
 }
 
-static int am65_cpsw_get_coalesce(struct net_device *ndev, struct ethtool_coalesce *coal,
-				  struct kernel_ethtool_coalesce *kernel_coal,
-				  struct netlink_ext_ack *extack)
-{
-	struct am65_cpsw_common *common = am65_ndev_to_common(ndev);
-	struct am65_cpsw_tx_chn *tx_chn;
-
-	tx_chn = &common->tx_chns[0];
-
-	coal->rx_coalesce_usecs = common->rx_pace_timeout / 1000;
-	coal->tx_coalesce_usecs = tx_chn->tx_pace_timeout / 1000;
-
-	return 0;
-}
-
 static int am65_cpsw_get_per_queue_coalesce(struct net_device *ndev, u32 queue,
 					    struct ethtool_coalesce *coal)
 {
 	struct am65_cpsw_common *common = am65_ndev_to_common(ndev);
+	struct am65_cpsw_rx_flow *rx_flow;
 	struct am65_cpsw_tx_chn *tx_chn;
 
-	if (queue >= AM65_CPSW_MAX_TX_QUEUES)
+	if (queue >= AM65_CPSW_MAX_QUEUES)
 		return -EINVAL;
 
 	tx_chn = &common->tx_chns[queue];
-
 	coal->tx_coalesce_usecs = tx_chn->tx_pace_timeout / 1000;
 
+	rx_flow = &common->rx_chns.flows[queue];
+	coal->rx_coalesce_usecs = rx_flow->rx_pace_timeout / 1000;
+
 	return 0;
 }
 
-static int am65_cpsw_set_coalesce(struct net_device *ndev, struct ethtool_coalesce *coal,
+static int am65_cpsw_get_coalesce(struct net_device *ndev, struct ethtool_coalesce *coal,
 				  struct kernel_ethtool_coalesce *kernel_coal,
 				  struct netlink_ext_ack *extack)
 {
-	struct am65_cpsw_common *common = am65_ndev_to_common(ndev);
-	struct am65_cpsw_tx_chn *tx_chn;
-
-	tx_chn = &common->tx_chns[0];
-
-	if (coal->rx_coalesce_usecs && coal->rx_coalesce_usecs < 20)
-		return -EINVAL;
-
-	if (coal->tx_coalesce_usecs && coal->tx_coalesce_usecs < 20)
-		return -EINVAL;
-
-	common->rx_pace_timeout = coal->rx_coalesce_usecs * 1000;
-	tx_chn->tx_pace_timeout = coal->tx_coalesce_usecs * 1000;
-
-	return 0;
+	return am65_cpsw_get_per_queue_coalesce(ndev, 0, coal);
 }
 
 static int am65_cpsw_set_per_queue_coalesce(struct net_device *ndev, u32 queue,
 					    struct ethtool_coalesce *coal)
 {
 	struct am65_cpsw_common *common = am65_ndev_to_common(ndev);
+	struct am65_cpsw_rx_flow *rx_flow;
 	struct am65_cpsw_tx_chn *tx_chn;
 
-	if (queue >= AM65_CPSW_MAX_TX_QUEUES)
+	if (queue >= AM65_CPSW_MAX_QUEUES)
 		return -EINVAL;
 
 	tx_chn = &common->tx_chns[queue];
-
-	if (coal->tx_coalesce_usecs && coal->tx_coalesce_usecs < 20) {
-		dev_info(common->dev, "defaulting to min value of 20us for tx-usecs for tx-%u\n",
-			 queue);
-		coal->tx_coalesce_usecs = 20;
-	}
+	if (coal->tx_coalesce_usecs && coal->tx_coalesce_usecs < 20)
+		return -EINVAL;
 
 	tx_chn->tx_pace_timeout = coal->tx_coalesce_usecs * 1000;
 
+	rx_flow = &common->rx_chns.flows[queue];
+	if (coal->rx_coalesce_usecs && coal->rx_coalesce_usecs < 20)
+		return -EINVAL;
+
+	rx_flow->rx_pace_timeout = coal->rx_coalesce_usecs * 1000;
+
 	return 0;
 }
 
+static int am65_cpsw_set_coalesce(struct net_device *ndev, struct ethtool_coalesce *coal,
+				  struct kernel_ethtool_coalesce *kernel_coal,
+				  struct netlink_ext_ack *extack)
+{
+	return am65_cpsw_set_per_queue_coalesce(ndev, 0, coal);
+}
+
 const struct ethtool_ops am65_cpsw_ethtool_ops_slave = {
 	.begin			= am65_cpsw_ethtool_op_begin,
 	.complete		= am65_cpsw_ethtool_op_complete,

drivers/net/ethernet/ti/am65-cpsw-nuss.h

@@ -21,9 +21,7 @@ struct am65_cpts;
 
 #define HOST_PORT_NUM		0
 
-#define AM65_CPSW_MAX_TX_QUEUES	8
-#define AM65_CPSW_MAX_RX_QUEUES	1
-#define AM65_CPSW_MAX_RX_FLOWS	1
+#define AM65_CPSW_MAX_QUEUES	8	/* both TX & RX */
 
 #define AM65_CPSW_PORT_VLAN_REG_OFFSET	0x014
 
@@ -58,7 +56,7 @@ struct am65_cpsw_port {
 	struct am65_cpsw_qos		qos;
 	struct devlink_port		devlink_port;
 	struct bpf_prog			*xdp_prog;
-	struct xdp_rxq_info		xdp_rxq;
+	struct xdp_rxq_info		xdp_rxq[AM65_CPSW_MAX_QUEUES];
 	/* Only for suspend resume context */
 	u32				vid_context;
 };
@@ -94,16 +92,27 @@ struct am65_cpsw_tx_chn {
 	u32 rate_mbps;
 };
 
+struct am65_cpsw_rx_flow {
+	u32 id;
+	struct napi_struct napi_rx;
+	struct am65_cpsw_common	*common;
+	int irq;
+	bool irq_disabled;
+	struct hrtimer rx_hrtimer;
+	unsigned long rx_pace_timeout;
+	struct page_pool *page_pool;
+	struct page **pages;
+	char name[32];
+};
+
 struct am65_cpsw_rx_chn {
 	struct device *dev;
 	struct device *dma_dev;
 	struct k3_cppi_desc_pool *desc_pool;
 	struct k3_udma_glue_rx_channel *rx_chn;
-	struct page_pool *page_pool;
-	struct page **pages;
 	u32 descs_num;
 	unsigned char dsize_log2;
-	int irq;
+	struct am65_cpsw_rx_flow flows[AM65_CPSW_MAX_QUEUES];
 };
 
 #define AM65_CPSW_QUIRK_I2027_NO_TX_CSUM BIT(0)
@@ -145,16 +154,12 @@ struct am65_cpsw_common {
 	u32			tx_ch_rate_msk;
 	u32			rx_flow_id_base;
 
-	struct am65_cpsw_tx_chn	tx_chns[AM65_CPSW_MAX_TX_QUEUES];
+	struct am65_cpsw_tx_chn	tx_chns[AM65_CPSW_MAX_QUEUES];
 	struct completion	tdown_complete;
 	atomic_t		tdown_cnt;
 
+	int			rx_ch_num_flows;
 	struct am65_cpsw_rx_chn	rx_chns;
-	struct napi_struct	napi_rx;
-
-	bool			rx_irq_disabled;
-	struct hrtimer		rx_hrtimer;
-	unsigned long		rx_pace_timeout;
 
 	u32			nuss_ver;
 	u32			cpsw_ver;
@@ -203,8 +208,8 @@ struct am65_cpsw_ndev_priv {
 #define am65_common_get_host(common) (&(common)->host)
 #define am65_common_get_port(common, id) (&(common)->ports[(id) - 1])
 
-#define am65_cpsw_napi_to_common(pnapi) \
-	container_of(pnapi, struct am65_cpsw_common, napi_rx)
+#define am65_cpsw_napi_to_rx_flow(pnapi) \
+	container_of(pnapi, struct am65_cpsw_rx_flow, napi_rx)
 #define am65_cpsw_napi_to_tx_chn(pnapi) \
 	container_of(pnapi, struct am65_cpsw_tx_chn, napi_tx)
 
@@ -215,8 +220,8 @@ struct am65_cpsw_ndev_priv {
 extern const struct ethtool_ops am65_cpsw_ethtool_ops_slave;
 
 void am65_cpsw_nuss_set_p0_ptype(struct am65_cpsw_common *common);
-void am65_cpsw_nuss_remove_tx_chns(struct am65_cpsw_common *common);
-int am65_cpsw_nuss_update_tx_chns(struct am65_cpsw_common *common, int num_tx);
+int am65_cpsw_nuss_update_tx_rx_chns(struct am65_cpsw_common *common,
+				     int num_tx, int num_rx);
 
 bool am65_cpsw_port_dev_check(const struct net_device *dev);