eth: fbnic: Add support for HDS configuration

Add support for configuring the header data split threshold.
For fbnic, the tcp data split support is enabled all the time.

Fbnic supports a maximum buffer size of 4KB. However, the reservation
for the headroom, tailroom, and padding reduce the max header size
accordingly.

  ethtool_hds -g eth0
  Ring parameters for eth0:
  Pre-set maximums:
  ...
  HDS thresh:		3584
  Current hardware settings:
  ...
  HDS thresh:		1536

Verify hds tests in ksft-net-drv are passing

  ksft-net-drv]# ./drivers/net/hds.py
  TAP version 13
  1..13
  ok 1 hds.get_hds
  ok 2 hds.get_hds_thresh
  ok 3 hds.set_hds_disable # SKIP disabling of HDS not supported by ...
  ...
  ...
  ok 12 hds.ioctl_set_xdp
  ok 13 hds.ioctl_enabled_set_xdp
  \# Totals: pass:12 fail:0 xfail:0 xpass:0 skip:1 error:0

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Mohsin Bashir <mohsin.bashr@gmail.com>
Link: https://patch.msgid.link/20250813221319.3367670-2-mohsin.bashr@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>

drivers/net/ethernet/meta/fbnic/fbnic_ethtool.c

@@ -2,6 +2,7 @@
 /* Copyright (c) Meta Platforms, Inc. and affiliates. */
 
 #include <linux/ethtool.h>
+#include <linux/ethtool_netlink.h>
 #include <linux/netdevice.h>
 #include <linux/pci.h>
 #include <net/ipv6.h>
@@ -160,6 +161,7 @@ static void fbnic_clone_swap_cfg(struct fbnic_net *orig,
 	swap(clone->num_rx_queues, orig->num_rx_queues);
 	swap(clone->num_tx_queues, orig->num_tx_queues);
 	swap(clone->num_napi, orig->num_napi);
+	swap(clone->hds_thresh, orig->hds_thresh);
 }
 
 static void fbnic_aggregate_vector_counters(struct fbnic_net *fbn,
@@ -277,15 +279,21 @@ fbnic_get_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring,
 	ring->rx_mini_pending = fbn->hpq_size;
 	ring->rx_jumbo_pending = fbn->ppq_size;
 	ring->tx_pending = fbn->txq_size;
+
+	kernel_ring->tcp_data_split = ETHTOOL_TCP_DATA_SPLIT_ENABLED;
+	kernel_ring->hds_thresh_max = FBNIC_HDS_THRESH_MAX;
+	kernel_ring->hds_thresh = fbn->hds_thresh;
 }
 
 static void fbnic_set_rings(struct fbnic_net *fbn,
-			    struct ethtool_ringparam *ring)
+			    struct ethtool_ringparam *ring,
+			    struct kernel_ethtool_ringparam *kernel_ring)
 {
 	fbn->rcq_size = ring->rx_pending;
 	fbn->hpq_size = ring->rx_mini_pending;
 	fbn->ppq_size = ring->rx_jumbo_pending;
 	fbn->txq_size = ring->tx_pending;
+	fbn->hds_thresh = kernel_ring->hds_thresh;
 }
 
 static int
@@ -316,8 +324,13 @@ fbnic_set_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring,
 		return -EINVAL;
 	}
 
+	if (kernel_ring->tcp_data_split == ETHTOOL_TCP_DATA_SPLIT_DISABLED) {
+		NL_SET_ERR_MSG_MOD(extack, "Cannot disable TCP data split");
+		return -EINVAL;
+	}
+
 	if (!netif_running(netdev)) {
-		fbnic_set_rings(fbn, ring);
+		fbnic_set_rings(fbn, ring, kernel_ring);
 		return 0;
 	}
 
@@ -325,7 +338,7 @@ fbnic_set_ringparam(struct net_device *netdev, struct ethtool_ringparam *ring,
 	if (!clone)
 		return -ENOMEM;
 
-	fbnic_set_rings(clone, ring);
+	fbnic_set_rings(clone, ring, kernel_ring);
 
 	err = fbnic_alloc_napi_vectors(clone);
 	if (err)
@@ -1678,6 +1691,8 @@ fbnic_get_rmon_stats(struct net_device *netdev,
 static const struct ethtool_ops fbnic_ethtool_ops = {
 	.supported_coalesce_params	= ETHTOOL_COALESCE_USECS |
 					  ETHTOOL_COALESCE_RX_MAX_FRAMES,
+	.supported_ring_params		= ETHTOOL_RING_USE_TCP_DATA_SPLIT |
+					  ETHTOOL_RING_USE_HDS_THRS,
 	.rxfh_max_num_contexts		= FBNIC_RPC_RSS_TBL_COUNT,
 	.get_drvinfo			= fbnic_get_drvinfo,
 	.get_regs_len			= fbnic_get_regs_len,

drivers/net/ethernet/meta/fbnic/fbnic_netdev.c

@@ -695,6 +695,10 @@ struct net_device *fbnic_netdev_alloc(struct fbnic_dev *fbd)
 	fbn->rx_usecs = FBNIC_RX_USECS_DEFAULT;
 	fbn->rx_max_frames = FBNIC_RX_FRAMES_DEFAULT;
 
+	/* Initialize the hds_thresh */
+	netdev->cfg->hds_thresh = FBNIC_HDS_THRESH_DEFAULT;
+	fbn->hds_thresh = FBNIC_HDS_THRESH_DEFAULT;
+
 	default_queues = netif_get_num_default_rss_queues();
 	if (default_queues > fbd->max_num_queues)
 		default_queues = fbd->max_num_queues;

drivers/net/ethernet/meta/fbnic/fbnic_netdev.h

@@ -31,6 +31,8 @@ struct fbnic_net {
 	u32 ppq_size;
 	u32 rcq_size;
 
+	u32 hds_thresh;
+
 	u16 rx_usecs;
 	u16 tx_usecs;
 

drivers/net/ethernet/meta/fbnic/fbnic_txrx.c

@@ -2232,13 +2232,22 @@ static void fbnic_enable_rcq(struct fbnic_napi_vector *nv,
 {
 	struct fbnic_net *fbn = netdev_priv(nv->napi.dev);
 	u32 log_size = fls(rcq->size_mask);
-	u32 rcq_ctl;
+	u32 hds_thresh = fbn->hds_thresh;
+	u32 rcq_ctl = 0;
 
 	fbnic_config_drop_mode_rcq(nv, rcq);
 
-	rcq_ctl = FIELD_PREP(FBNIC_QUEUE_RDE_CTL1_PADLEN_MASK, FBNIC_RX_PAD) |
-		   FIELD_PREP(FBNIC_QUEUE_RDE_CTL1_MAX_HDR_MASK,
-			      FBNIC_RX_MAX_HDR) |
+	/* Force lower bound on MAX_HEADER_BYTES. Below this, all frames should
+	 * be split at L4. It would also result in the frames being split at
+	 * L2/L3 depending on the frame size.
+	 */
+	if (fbn->hds_thresh < FBNIC_HDR_BYTES_MIN) {
+		rcq_ctl = FBNIC_QUEUE_RDE_CTL0_EN_HDR_SPLIT;
+		hds_thresh = FBNIC_HDR_BYTES_MIN;
+	}
+
+	rcq_ctl |= FIELD_PREP(FBNIC_QUEUE_RDE_CTL1_PADLEN_MASK, FBNIC_RX_PAD) |
+		   FIELD_PREP(FBNIC_QUEUE_RDE_CTL1_MAX_HDR_MASK, hds_thresh) |
 		   FIELD_PREP(FBNIC_QUEUE_RDE_CTL1_PAYLD_OFF_MASK,
 			      FBNIC_RX_PAYLD_OFFSET) |
 		   FIELD_PREP(FBNIC_QUEUE_RDE_CTL1_PAYLD_PG_CL_MASK,

drivers/net/ethernet/meta/fbnic/fbnic_txrx.h

@@ -53,7 +53,6 @@ struct fbnic_net;
 #define FBNIC_RX_HROOM \
 	(ALIGN(FBNIC_RX_TROOM + NET_SKB_PAD, 128) - FBNIC_RX_TROOM)
 #define FBNIC_RX_PAD			0
-#define FBNIC_RX_MAX_HDR		(1536 - FBNIC_RX_PAD)
 #define FBNIC_RX_PAYLD_OFFSET		0
 #define FBNIC_RX_PAYLD_PG_CL		0
 
@@ -61,6 +60,12 @@ struct fbnic_net;
 #define FBNIC_RING_F_CTX		BIT(1)
 #define FBNIC_RING_F_STATS		BIT(2)	/* Ring's stats may be used */
 
+#define FBNIC_HDS_THRESH_MAX \
+	(4096 - FBNIC_RX_HROOM - FBNIC_RX_TROOM - FBNIC_RX_PAD)
+#define FBNIC_HDS_THRESH_DEFAULT \
+	(1536 - FBNIC_RX_PAD)
+#define FBNIC_HDR_BYTES_MIN		128
+
 struct fbnic_pkt_buff {
 	struct xdp_buff buff;
 	ktime_t hwtstamp;