Merge branch 'bnxt_en-XDP_REDIRECT'

Michael Chan says:

====================
bnxt_en: Add XDP_REDIRECT support.

This patch series adds XDP_REDIRECT support by Andy Gospodarek.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/Kconfig

@@ -199,6 +199,7 @@ config BNXT
 	select FW_LOADER
 	select LIBCRC32C
 	select NET_DEVLINK
+	select PAGE_POOL
 	---help---
 	  This driver supports Broadcom NetXtreme-C/E 10/25/40/50 gigabit
 	  Ethernet cards.  To compile this driver as a module, choose M here:

drivers/net/ethernet/broadcom/bnxt/bnxt.c

@@ -54,6 +54,7 @@
 #include <net/pkt_cls.h>
 #include <linux/hwmon.h>
 #include <linux/hwmon-sysfs.h>
+#include <net/page_pool.h>
 
 #include "bnxt_hsi.h"
 #include "bnxt.h"
@@ -668,19 +669,20 @@ static void bnxt_tx_int(struct bnxt *bp, struct bnxt_napi *bnapi, int nr_pkts)
 }
 
 static struct page *__bnxt_alloc_rx_page(struct bnxt *bp, dma_addr_t *mapping,
+					 struct bnxt_rx_ring_info *rxr,
 					 gfp_t gfp)
 {
 	struct device *dev = &bp->pdev->dev;
 	struct page *page;
 
-	page = alloc_page(gfp);
+	page = page_pool_dev_alloc_pages(rxr->page_pool);
 	if (!page)
 		return NULL;
 
 	*mapping = dma_map_page_attrs(dev, page, 0, PAGE_SIZE, bp->rx_dir,
 				      DMA_ATTR_WEAK_ORDERING);
 	if (dma_mapping_error(dev, *mapping)) {
-		__free_page(page);
+		page_pool_recycle_direct(rxr->page_pool, page);
 		return NULL;
 	}
 	*mapping += bp->rx_dma_offset;
@@ -716,7 +718,8 @@ int bnxt_alloc_rx_data(struct bnxt *bp, struct bnxt_rx_ring_info *rxr,
 	dma_addr_t mapping;
 
 	if (BNXT_RX_PAGE_MODE(bp)) {
-		struct page *page = __bnxt_alloc_rx_page(bp, &mapping, gfp);
+		struct page *page =
+			__bnxt_alloc_rx_page(bp, &mapping, rxr, gfp);
 
 		if (!page)
 			return -ENOMEM;
@@ -1989,6 +1992,9 @@ static int __bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 		}
 	}
 
+	if (event & BNXT_REDIRECT_EVENT)
+		xdp_do_flush_map();
+
 	if (event & BNXT_TX_EVENT) {
 		struct bnxt_tx_ring_info *txr = bnapi->tx_ring;
 		u16 prod = txr->tx_prod;
@@ -2254,9 +2260,23 @@ static void bnxt_free_tx_skbs(struct bnxt *bp)
 
 		for (j = 0; j < max_idx;) {
 			struct bnxt_sw_tx_bd *tx_buf = &txr->tx_buf_ring[j];
-			struct sk_buff *skb = tx_buf->skb;
+			struct sk_buff *skb;
 			int k, last;
 
+			if (i < bp->tx_nr_rings_xdp &&
+			    tx_buf->action == XDP_REDIRECT) {
+				dma_unmap_single(&pdev->dev,
+					dma_unmap_addr(tx_buf, mapping),
+					dma_unmap_len(tx_buf, len),
+					PCI_DMA_TODEVICE);
+				xdp_return_frame(tx_buf->xdpf);
+				tx_buf->action = 0;
+				tx_buf->xdpf = NULL;
+				j++;
+				continue;
+			}
+
+			skb = tx_buf->skb;
 			if (!skb) {
 				j++;
 				continue;
@@ -2343,7 +2363,7 @@ static void bnxt_free_rx_skbs(struct bnxt *bp)
 				dma_unmap_page_attrs(&pdev->dev, mapping,
 						     PAGE_SIZE, bp->rx_dir,
 						     DMA_ATTR_WEAK_ORDERING);
-				__free_page(data);
+				page_pool_recycle_direct(rxr->page_pool, data);
 			} else {
 				dma_unmap_single_attrs(&pdev->dev, mapping,
 						       bp->rx_buf_use_size,
@@ -2480,6 +2500,8 @@ static void bnxt_free_rx_rings(struct bnxt *bp)
 		if (xdp_rxq_info_is_reg(&rxr->xdp_rxq))
 			xdp_rxq_info_unreg(&rxr->xdp_rxq);
 
+		rxr->page_pool = NULL;
+
 		kfree(rxr->rx_tpa);
 		rxr->rx_tpa = NULL;
 
@@ -2494,6 +2516,26 @@ static void bnxt_free_rx_rings(struct bnxt *bp)
 	}
 }
 
+static int bnxt_alloc_rx_page_pool(struct bnxt *bp,
+				   struct bnxt_rx_ring_info *rxr)
+{
+	struct page_pool_params pp = { 0 };
+
+	pp.pool_size = bp->rx_ring_size;
+	pp.nid = dev_to_node(&bp->pdev->dev);
+	pp.dev = &bp->pdev->dev;
+	pp.dma_dir = DMA_BIDIRECTIONAL;
+
+	rxr->page_pool = page_pool_create(&pp);
+	if (IS_ERR(rxr->page_pool)) {
+		int err = PTR_ERR(rxr->page_pool);
+
+		rxr->page_pool = NULL;
+		return err;
+	}
+	return 0;
+}
+
 static int bnxt_alloc_rx_rings(struct bnxt *bp)
 {
 	int i, rc, agg_rings = 0, tpa_rings = 0;
@@ -2513,9 +2555,26 @@ static int bnxt_alloc_rx_rings(struct bnxt *bp)
 
 		ring = &rxr->rx_ring_struct;
 
+		rc = bnxt_alloc_rx_page_pool(bp, rxr);
+		if (rc)
+			return rc;
+
 		rc = xdp_rxq_info_reg(&rxr->xdp_rxq, bp->dev, i);
-		if (rc < 0)
+		if (rc < 0) {
+			page_pool_free(rxr->page_pool);
+			rxr->page_pool = NULL;
+			return rc;
+		}
+
+		rc = xdp_rxq_info_reg_mem_model(&rxr->xdp_rxq,
+						MEM_TYPE_PAGE_POOL,
+						rxr->page_pool);
+		if (rc) {
+			xdp_rxq_info_unreg(&rxr->xdp_rxq);
+			page_pool_free(rxr->page_pool);
+			rxr->page_pool = NULL;
 			return rc;
+		}
 
 		rc = bnxt_alloc_ring(bp, &ring->ring_mem);
 		if (rc)
@@ -10233,6 +10292,7 @@ static const struct net_device_ops bnxt_netdev_ops = {
 	.ndo_udp_tunnel_add	= bnxt_udp_tunnel_add,
 	.ndo_udp_tunnel_del	= bnxt_udp_tunnel_del,
 	.ndo_bpf		= bnxt_xdp,
+	.ndo_xdp_xmit		= bnxt_xdp_xmit,
 	.ndo_bridge_getlink	= bnxt_bridge_getlink,
 	.ndo_bridge_setlink	= bnxt_bridge_setlink,
 	.ndo_get_devlink_port	= bnxt_get_devlink_port,

drivers/net/ethernet/broadcom/bnxt/bnxt.h

@@ -26,6 +26,8 @@
 #include <net/xdp.h>
 #include <linux/dim.h>
 
+struct page_pool;
+
 struct tx_bd {
 	__le32 tx_bd_len_flags_type;
 	#define TX_BD_TYPE					(0x3f << 0)
@@ -590,12 +592,18 @@ struct nqe_cn {
 #define BNXT_RX_EVENT		1
 #define BNXT_AGG_EVENT		2
 #define BNXT_TX_EVENT		4
+#define BNXT_REDIRECT_EVENT	8
 
 struct bnxt_sw_tx_bd {
+	union {
 		struct sk_buff		*skb;
+		struct xdp_frame	*xdpf;
+	};
 	DEFINE_DMA_UNMAP_ADDR(mapping);
+	DEFINE_DMA_UNMAP_LEN(len);
 	u8			is_gso;
 	u8			is_push;
+	u8			action;
 	union {
 		unsigned short		nr_frags;
 		u16			rx_prod;
@@ -793,6 +801,7 @@ struct bnxt_rx_ring_info {
 	struct bnxt_ring_struct	rx_ring_struct;
 	struct bnxt_ring_struct	rx_agg_ring_struct;
 	struct xdp_rxq_info	xdp_rxq;
+	struct page_pool	*page_pool;
 };
 
 struct bnxt_cp_ring_info {

drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c

@@ -2799,7 +2799,7 @@ static int bnxt_run_loopback(struct bnxt *bp)
 		dev_kfree_skb(skb);
 		return -EIO;
 	}
-	bnxt_xmit_xdp(bp, txr, map, pkt_size, 0);
+	bnxt_xmit_bd(bp, txr, map, pkt_size);
 
 	/* Sync BD data before updating doorbell */
 	wmb();

drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.c

@@ -15,12 +15,14 @@
 #include <linux/bpf.h>
 #include <linux/bpf_trace.h>
 #include <linux/filter.h>
+#include <net/page_pool.h>
 #include "bnxt_hsi.h"
 #include "bnxt.h"
 #include "bnxt_xdp.h"
 
-void bnxt_xmit_xdp(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
-		   dma_addr_t mapping, u32 len, u16 rx_prod)
+struct bnxt_sw_tx_bd *bnxt_xmit_bd(struct bnxt *bp,
+				   struct bnxt_tx_ring_info *txr,
+				   dma_addr_t mapping, u32 len)
 {
 	struct bnxt_sw_tx_bd *tx_buf;
 	struct tx_bd *txbd;
@@ -29,7 +31,6 @@ void bnxt_xmit_xdp(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
 
 	prod = txr->tx_prod;
 	tx_buf = &txr->tx_buf_ring[prod];
-	tx_buf->rx_prod = rx_prod;
 
 	txbd = &txr->tx_desc_ring[TX_RING(prod)][TX_IDX(prod)];
 	flags = (len << TX_BD_LEN_SHIFT) | (1 << TX_BD_FLAGS_BD_CNT_SHIFT) |
@@ -40,30 +41,67 @@ void bnxt_xmit_xdp(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
 
 	prod = NEXT_TX(prod);
 	txr->tx_prod = prod;
+	return tx_buf;
+}
+
+static void __bnxt_xmit_xdp(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
+			    dma_addr_t mapping, u32 len, u16 rx_prod)
+{
+	struct bnxt_sw_tx_bd *tx_buf;
+
+	tx_buf = bnxt_xmit_bd(bp, txr, mapping, len);
+	tx_buf->rx_prod = rx_prod;
+	tx_buf->action = XDP_TX;
+}
+
+static void __bnxt_xmit_xdp_redirect(struct bnxt *bp,
+				     struct bnxt_tx_ring_info *txr,
+				     dma_addr_t mapping, u32 len,
+				     struct xdp_frame *xdpf)
+{
+	struct bnxt_sw_tx_bd *tx_buf;
+
+	tx_buf = bnxt_xmit_bd(bp, txr, mapping, len);
+	tx_buf->action = XDP_REDIRECT;
+	tx_buf->xdpf = xdpf;
+	dma_unmap_addr_set(tx_buf, mapping, mapping);
+	dma_unmap_len_set(tx_buf, len, 0);
 }
 
 void bnxt_tx_int_xdp(struct bnxt *bp, struct bnxt_napi *bnapi, int nr_pkts)
 {
 	struct bnxt_tx_ring_info *txr = bnapi->tx_ring;
 	struct bnxt_rx_ring_info *rxr = bnapi->rx_ring;
+	bool rx_doorbell_needed = false;
 	struct bnxt_sw_tx_bd *tx_buf;
 	u16 tx_cons = txr->tx_cons;
 	u16 last_tx_cons = tx_cons;
-	u16 rx_prod;
 	int i;
 
 	for (i = 0; i < nr_pkts; i++) {
+		tx_buf = &txr->tx_buf_ring[tx_cons];
+
+		if (tx_buf->action == XDP_REDIRECT) {
+			struct pci_dev *pdev = bp->pdev;
+
+			dma_unmap_single(&pdev->dev,
+					 dma_unmap_addr(tx_buf, mapping),
+					 dma_unmap_len(tx_buf, len),
+					 PCI_DMA_TODEVICE);
+			xdp_return_frame(tx_buf->xdpf);
+			tx_buf->action = 0;
+			tx_buf->xdpf = NULL;
+		} else if (tx_buf->action == XDP_TX) {
+			rx_doorbell_needed = true;
 			last_tx_cons = tx_cons;
+		}
 		tx_cons = NEXT_TX(tx_cons);
 	}
 	txr->tx_cons = tx_cons;
-	if (bnxt_tx_avail(bp, txr) == bp->tx_ring_size) {
-		rx_prod = rxr->rx_prod;
-	} else {
+	if (rx_doorbell_needed) {
 		tx_buf = &txr->tx_buf_ring[last_tx_cons];
-		rx_prod = tx_buf->rx_prod;
+		bnxt_db_write(bp, &rxr->rx_db, tx_buf->rx_prod);
 	}
-	bnxt_db_write(bp, &rxr->rx_db, rx_prod);
 }
 
 /* returns the following:
@@ -88,19 +126,19 @@ bool bnxt_rx_xdp(struct bnxt *bp, struct bnxt_rx_ring_info *rxr, u16 cons,
 		return false;
 
 	pdev = bp->pdev;
-	txr = rxr->bnapi->tx_ring;
 	rx_buf = &rxr->rx_buf_ring[cons];
 	offset = bp->rx_offset;
 
+	mapping = rx_buf->mapping - bp->rx_dma_offset;
+	dma_sync_single_for_cpu(&pdev->dev, mapping + offset, *len, bp->rx_dir);
+
+	txr = rxr->bnapi->tx_ring;
 	xdp.data_hard_start = *data_ptr - offset;
 	xdp.data = *data_ptr;
 	xdp_set_data_meta_invalid(&xdp);
 	xdp.data_end = *data_ptr + *len;
 	xdp.rxq = &rxr->xdp_rxq;
 	orig_data = xdp.data;
-	mapping = rx_buf->mapping - bp->rx_dma_offset;
-
-	dma_sync_single_for_cpu(&pdev->dev, mapping + offset, *len, bp->rx_dir);
 
 	rcu_read_lock();
 	act = bpf_prog_run_xdp(xdp_prog, &xdp);
@@ -132,10 +170,34 @@ bool bnxt_rx_xdp(struct bnxt *bp, struct bnxt_rx_ring_info *rxr, u16 cons,
 		*event = BNXT_TX_EVENT;
 		dma_sync_single_for_device(&pdev->dev, mapping + offset, *len,
 					   bp->rx_dir);
-		bnxt_xmit_xdp(bp, txr, mapping + offset, *len,
+		__bnxt_xmit_xdp(bp, txr, mapping + offset, *len,
 				NEXT_RX(rxr->rx_prod));
 		bnxt_reuse_rx_data(rxr, cons, page);
 		return true;
+	case XDP_REDIRECT:
+		/* if we are calling this here then we know that the
+		 * redirect is coming from a frame received by the
+		 * bnxt_en driver.
+		 */
+		dma_unmap_page_attrs(&pdev->dev, mapping,
+				     PAGE_SIZE, bp->rx_dir,
+				     DMA_ATTR_WEAK_ORDERING);
+
+		/* if we are unable to allocate a new buffer, abort and reuse */
+		if (bnxt_alloc_rx_data(bp, rxr, rxr->rx_prod, GFP_ATOMIC)) {
+			trace_xdp_exception(bp->dev, xdp_prog, act);
+			bnxt_reuse_rx_data(rxr, cons, page);
+			return true;
+		}
+
+		if (xdp_do_redirect(bp->dev, &xdp, xdp_prog)) {
+			trace_xdp_exception(bp->dev, xdp_prog, act);
+			page_pool_recycle_direct(rxr->page_pool, page);
+			return true;
+		}
+
+		*event |= BNXT_REDIRECT_EVENT;
+		break;
 	default:
 		bpf_warn_invalid_xdp_action(act);
 		/* Fall thru */
@@ -149,6 +211,56 @@ bool bnxt_rx_xdp(struct bnxt *bp, struct bnxt_rx_ring_info *rxr, u16 cons,
 	return true;
 }
 
+int bnxt_xdp_xmit(struct net_device *dev, int num_frames,
+		  struct xdp_frame **frames, u32 flags)
+{
+	struct bnxt *bp = netdev_priv(dev);
+	struct bpf_prog *xdp_prog = READ_ONCE(bp->xdp_prog);
+	struct pci_dev *pdev = bp->pdev;
+	struct bnxt_tx_ring_info *txr;
+	dma_addr_t mapping;
+	int drops = 0;
+	int ring;
+	int i;
+
+	if (!test_bit(BNXT_STATE_OPEN, &bp->state) ||
+	    !bp->tx_nr_rings_xdp ||
+	    !xdp_prog)
+		return -EINVAL;
+
+	ring = smp_processor_id() % bp->tx_nr_rings_xdp;
+	txr = &bp->tx_ring[ring];
+
+	for (i = 0; i < num_frames; i++) {
+		struct xdp_frame *xdp = frames[i];
+
+		if (!txr || !bnxt_tx_avail(bp, txr) ||
+		    !(bp->bnapi[ring]->flags & BNXT_NAPI_FLAG_XDP)) {
+			xdp_return_frame_rx_napi(xdp);
+			drops++;
+			continue;
+		}
+
+		mapping = dma_map_single(&pdev->dev, xdp->data, xdp->len,
+					 DMA_TO_DEVICE);
+
+		if (dma_mapping_error(&pdev->dev, mapping)) {
+			xdp_return_frame_rx_napi(xdp);
+			drops++;
+			continue;
+		}
+		__bnxt_xmit_xdp_redirect(bp, txr, mapping, xdp->len, xdp);
+	}
+
+	if (flags & XDP_XMIT_FLUSH) {
+		/* Sync BD data before updating doorbell */
+		wmb();
+		bnxt_db_write(bp, &txr->tx_db, txr->tx_prod);
+	}
+
+	return num_frames - drops;
+}
+
 /* Under rtnl_lock */
 static int bnxt_xdp_set(struct bnxt *bp, struct bpf_prog *prog)
 {

drivers/net/ethernet/broadcom/bnxt/bnxt_xdp.h

@@ -10,12 +10,15 @@
 #ifndef BNXT_XDP_H
 #define BNXT_XDP_H
 
-void bnxt_xmit_xdp(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
-		   dma_addr_t mapping, u32 len, u16 rx_prod);
+struct bnxt_sw_tx_bd *bnxt_xmit_bd(struct bnxt *bp,
+				   struct bnxt_tx_ring_info *txr,
+				   dma_addr_t mapping, u32 len);
 void bnxt_tx_int_xdp(struct bnxt *bp, struct bnxt_napi *bnapi, int nr_pkts);
 bool bnxt_rx_xdp(struct bnxt *bp, struct bnxt_rx_ring_info *rxr, u16 cons,
 		 struct page *page, u8 **data_ptr, unsigned int *len,
 		 u8 *event);
 int bnxt_xdp(struct net_device *dev, struct netdev_bpf *xdp);
+int bnxt_xdp_xmit(struct net_device *dev, int num_frames,
+		  struct xdp_frame **frames, u32 flags);
 
 #endif