[PATCH v1 0/3] raw/ntb: add AMD NTB support

[PATCH v1 0/3] raw/ntb: add AMD NTB support

[Raghavendra Ningoji]

This series adds support for the Non-Transparent Bridge (NTB) endpoints
integrated in AMD EPYC Embedded "Turin", "Genoa" and "Siena" processors
to the raw/ntb driver.

The NTB rawdev framework was originally written around the Intel
back-to-back topology and the built-in scratchpad handshake protocol.
Patch 1 generalizes it with a few vendor-neutral hooks so other vendors
can reuse the common code; the Intel driver is unaffected since the new
hooks fall back to the built-in paths when NULL. Patch 2 adds the AMD
hardware driver, and patch 3 documents the new support.

A few AMD-specific hardware details worth noting for reviewers:

- Primary/secondary topology. One endpoint enumerates as the primary
  (device ID 0x14c0) and the other as the secondary (0x14c3), rather
  than the Intel back-to-back model.

- Packed scratchpad handshake. The hardware exposes a single shared
  16-register scratchpad bank split into two disjoint 8-register sets
  (one per side), so the driver uses a packed handshake layout that
  fits in 8 registers, plugged in through the dev_handshake and
  read_peer_config hooks.

- Outbound translation window. AMD NTB uses an outbound translation
  window: a write to a BARxx memory-window offset is forwarded to
  (xlat_base | offset) in the peer's memory rather than
  (xlat_base + offset). The translation base must therefore be aligned
  to a power of two >= the window length so that no offset bit collides
  with a set bit in the base. This is reported to applications through a
  new mw_addr_align field in struct ntb_dev_info and honoured by the ntb
  example when reserving the memory window memzone.

- Secondary-side link status. On the secondary side the device's own
  PCIe link status does not reflect the true inter-host link, so the
  speed and width are read from the upstream switch port via sysfs.

Tested on a two-node setup (primary + secondary EPYC Embedded hosts)
with the dpdk-ntb example: the link comes up at the expected speed and
width on both sides, and file transfers are byte-correct end to end for
both small and large files.

Raghavendra Ningoji (3):
  raw/ntb: generalize framework for multiple vendors
  raw/ntb: add AMD NTB support
  doc: add AMD NTB support

 MAINTAINERS                            |   1 +
 doc/guides/rawdevs/ntb.rst             |  18 +-
 doc/guides/rel_notes/release_26_11.rst |   8 +
 drivers/raw/ntb/meson.build            |   3 +-
 drivers/raw/ntb/ntb.c                  | 115 ++--
 drivers/raw/ntb/ntb.h                  |  18 +
 drivers/raw/ntb/ntb_hw_amd.c           | 709 +++++++++++++++++++++++++
 drivers/raw/ntb/ntb_hw_amd.h           | 114 ++++
 drivers/raw/ntb/rte_pmd_ntb.h          |  10 +
 examples/ntb/ntb_fwd.c                 |  22 +-
 usertools/dpdk-devbind.py              |   4 +-
 11 files changed, 985 insertions(+), 37 deletions(-)
 create mode 100644 drivers/raw/ntb/ntb_hw_amd.c
 create mode 100644 drivers/raw/ntb/ntb_hw_amd.h

-- 
2.34.1

[PATCH v1 1/3] raw/ntb: generalize framework for multiple vendors

[Raghavendra Ningoji]

The NTB rawdev framework was written around the Intel back-to-back
topology and the built-in scratchpad handshake protocol. To allow
other vendors to plug into the same framework, add vendor-neutral
hooks and make the common code dispatch through them:

- Add NTB_TOPO_PRI/NTB_TOPO_SEC topology types for hardware that uses
  a primary/secondary topology instead of back-to-back.
- Add optional ntb_dev_ops hooks: interrupt_handler (vendor-specific
  MSI-X handler), dev_handshake (vendor-specific link handshake) and
  read_peer_config (vendor-specific peer-config read at start). When a
  hook is NULL the common code keeps using the existing built-in path,
  so the Intel driver is unaffected.
- Add a pmd_private pointer to struct ntb_hw for vendor-specific state.
- Guard the receive path against a malformed stream with no end-of-packet
  marker so it cannot overflow the descriptor ring.

Signed-off-by: Raghavendra Ningoji <raghavendra.ningoji@amd.com>
---
 drivers/raw/ntb/ntb.c | 94 ++++++++++++++++++++++++++++---------------
 drivers/raw/ntb/ntb.h | 18 +++++++++
 2 files changed, 80 insertions(+), 32 deletions(-)

diff --git a/drivers/raw/ntb/ntb.c b/drivers/raw/ntb/ntb.c
index d54f2fb783..3a6a299081 100644
--- a/drivers/raw/ntb/ntb.c
+++ b/drivers/raw/ntb/ntb.c
@@ -746,6 +746,11 @@ ntb_dequeue_bufs(struct rte_rawdev *dev,
 	for (nb_rx = 0; nb_rx < count; nb_rx++) {
 		i = 0;
 		while (true) {
+			if (unlikely(nb_mbufs >= rxq->nb_rx_desc)) {
+				NTB_LOG(ERR, "Malformed rx stream (no EOP); "
+					"aborting to avoid desc overflow.");
+				goto end_of_rx;
+			}
 			rx_item = rxq->rx_used_ring + rxq->last_used;
 			rxm_t = sw_ring[rxq->last_used].mbuf;
 			rxm_t->data_len = rx_item->len;
@@ -882,8 +887,13 @@ ntb_dev_configure(const struct rte_rawdev *dev, rte_rawdev_obj_t config,
 	hw->ntb_xstats_off = rte_zmalloc("ntb_xstats_off", xstats_num *
 					 sizeof(uint64_t), 0);
 
-	/* Start handshake with the peer. */
-	ret = ntb_handshake_work(dev);
+	/* Start handshake with the peer. Use the vendor-specific handshake
+	 * if provided, otherwise the built-in scratchpad protocol.
+	 */
+	if (hw->ntb_ops->dev_handshake != NULL)
+		ret = (*hw->ntb_ops->dev_handshake)(dev);
+	else
+		ret = ntb_handshake_work(dev);
 	if (ret < 0) {
 		rte_free(hw->rx_queues);
 		rte_free(hw->tx_queues);
@@ -929,35 +939,44 @@ ntb_dev_start(struct rte_rawdev *dev)
 		goto err_q_init;
 	}
 
-	if (hw->ntb_ops->spad_read == NULL) {
-		ret = -ENOTSUP;
-		goto err_up;
-	}
+	/* Read/validate peer config. Use the vendor-specific reader if
+	 * provided, otherwise the built-in scratchpad reads.
+	 */
+	if (hw->ntb_ops->read_peer_config != NULL) {
+		ret = (*hw->ntb_ops->read_peer_config)(dev);
+		if (ret < 0)
+			goto err_up;
+	} else {
+		if (hw->ntb_ops->spad_read == NULL) {
+			ret = -ENOTSUP;
+			goto err_up;
+		}
 
-	peer_val = (*hw->ntb_ops->spad_read)(dev, SPAD_Q_SZ, 0);
-	if (peer_val != hw->queue_size) {
-		NTB_LOG(ERR, "Inconsistent queue size! (local: %u peer: %u)",
-			hw->queue_size, peer_val);
-		ret = -EINVAL;
-		goto err_up;
-	}
+		peer_val = (*hw->ntb_ops->spad_read)(dev, SPAD_Q_SZ, 0);
+		if (peer_val != hw->queue_size) {
+			NTB_LOG(ERR, "Inconsistent queue size! (local: %u peer: %u)",
+				hw->queue_size, peer_val);
+			ret = -EINVAL;
+			goto err_up;
+		}
 
-	peer_val = (*hw->ntb_ops->spad_read)(dev, SPAD_NUM_QPS, 0);
-	if (peer_val != hw->queue_pairs) {
-		NTB_LOG(ERR, "Inconsistent number of queues! (local: %u peer:"
-			" %u)", hw->queue_pairs, peer_val);
-		ret = -EINVAL;
-		goto err_up;
-	}
+		peer_val = (*hw->ntb_ops->spad_read)(dev, SPAD_NUM_QPS, 0);
+		if (peer_val != hw->queue_pairs) {
+			NTB_LOG(ERR, "Inconsistent number of queues! (local: %u peer:"
+				" %u)", hw->queue_pairs, peer_val);
+			ret = -EINVAL;
+			goto err_up;
+		}
 
-	hw->peer_used_mws = (*hw->ntb_ops->spad_read)(dev, SPAD_USED_MWS, 0);
+		hw->peer_used_mws = (*hw->ntb_ops->spad_read)(dev, SPAD_USED_MWS, 0);
 
-	for (i = 0; i < hw->peer_used_mws; i++) {
-		peer_base_h = (*hw->ntb_ops->spad_read)(dev,
-				SPAD_MW0_BA_H + 2 * i, 0);
-		peer_base_l = (*hw->ntb_ops->spad_read)(dev,
-				SPAD_MW0_BA_L + 2 * i, 0);
-		hw->peer_mw_base[i] = (peer_base_h << 32) + peer_base_l;
+		for (i = 0; i < hw->peer_used_mws; i++) {
+			peer_base_h = (*hw->ntb_ops->spad_read)(dev,
+					SPAD_MW0_BA_H + 2 * i, 0);
+			peer_base_l = (*hw->ntb_ops->spad_read)(dev,
+					SPAD_MW0_BA_L + 2 * i, 0);
+			hw->peer_mw_base[i] = (peer_base_h << 32) + peer_base_l;
+		}
 	}
 
 	dev->started = 1;
@@ -1057,8 +1076,13 @@ ntb_dev_close(struct rte_rawdev *dev)
 	rte_intr_disable(intr_handle);
 
 	/* Unregister callback func to eal lib */
-	rte_intr_callback_unregister(intr_handle,
-				     ntb_dev_intr_handler, dev);
+	if (hw->ntb_ops->interrupt_handler != NULL)
+		rte_intr_callback_unregister(intr_handle,
+					     hw->ntb_ops->interrupt_handler,
+					     dev);
+	else
+		rte_intr_callback_unregister(intr_handle,
+					     ntb_dev_intr_handler, dev);
 
 	return 0;
 }
@@ -1409,9 +1433,15 @@ ntb_init_hw(struct rte_rawdev *dev, struct rte_pci_device *pci_dev)
 	(*hw->ntb_ops->db_clear)(dev, hw->db_valid_mask);
 
 	intr_handle = pci_dev->intr_handle;
-	/* Register callback func to eal lib */
-	rte_intr_callback_register(intr_handle,
-				   ntb_dev_intr_handler, dev);
+	/* Register callback func to eal lib. Use the vendor-specific handler
+	 * if provided, otherwise fall back to the built-in handler.
+	 */
+	if (hw->ntb_ops->interrupt_handler != NULL)
+		rte_intr_callback_register(intr_handle,
+					   hw->ntb_ops->interrupt_handler, dev);
+	else
+		rte_intr_callback_register(intr_handle,
+					   ntb_dev_intr_handler, dev);
 
 	ret = rte_intr_efd_enable(intr_handle, hw->db_cnt);
 	if (ret)
diff --git a/drivers/raw/ntb/ntb.h b/drivers/raw/ntb/ntb.h
index 8c7a2230f9..270f9f9046 100644
--- a/drivers/raw/ntb/ntb.h
+++ b/drivers/raw/ntb/ntb.h
@@ -42,6 +42,9 @@ enum ntb_topo {
 	NTB_TOPO_NONE = 0,
 	NTB_TOPO_B2B_USD,
 	NTB_TOPO_B2B_DSD,
+	/* Primary/secondary topology (e.g. AMD NTB). */
+	NTB_TOPO_PRI,
+	NTB_TOPO_SEC,
 };
 
 enum ntb_link {
@@ -100,6 +103,8 @@ enum ntb_spad_idx {
  * for those db bits.
  * @peer_db_set: Set doorbell bit to generate peer interrupt for that bit.
  * @vector_bind: Bind vector source [intr] to msix vector [msix].
+ * @interrupt_handler: Vendor-specific interrupt handler. If NULL, the
+ * built-in handler is used.
  */
 struct ntb_dev_ops {
 	int (*ntb_dev_init)(const struct rte_rawdev *dev);
@@ -119,6 +124,16 @@ struct ntb_dev_ops {
 	int (*peer_db_set)(const struct rte_rawdev *dev, uint8_t db_bit);
 	int (*vector_bind)(const struct rte_rawdev *dev, uint8_t intr,
 			   uint8_t msix);
+	void (*interrupt_handler)(void *param);
+	/* Optional vendor-specific handshake. If NULL, the built-in
+	 * scratchpad handshake is used. Used by hardware (e.g. AMD) whose
+	 * scratchpad layout differs from the built-in protocol.
+	 */
+	int (*dev_handshake)(const struct rte_rawdev *dev);
+	/* Optional vendor-specific peer-config read at device start. If NULL,
+	 * the built-in scratchpad reads are used.
+	 */
+	int (*read_peer_config)(const struct rte_rawdev *dev);
 };
 
 struct ntb_desc {
@@ -208,6 +223,9 @@ struct ntb_hw {
 
 	const struct ntb_dev_ops *ntb_ops;
 
+	/* Vendor-specific hardware private data. */
+	void *pmd_private;
+
 	struct rte_pci_device *pci_dev;
 	char *hw_addr;
 
-- 
2.34.1

[PATCH v1 2/3] raw/ntb: add AMD NTB support

[Raghavendra Ningoji]

Add support for the NTB endpoints integrated in AMD EPYC Embedded
"Turin", "Genoa" and "Siena" processors to the raw/ntb driver.

The AMD NTB uses a primary/secondary topology: one endpoint enumerates
as the primary (device ID 0x14c0) and the other as the secondary
(device ID 0x14c3). The hardware exposes two memory windows (BAR23 and
BAR45), 16 doorbells and a single shared 16-register scratchpad bank.
The scratchpad bank is split into two disjoint 8-register sets, one per
side, so the driver uses a packed handshake layout that fits in 8
registers, plugged in through the framework's dev_handshake and
read_peer_config hooks. A vendor-specific MSI-X interrupt handler is
provided through the interrupt_handler hook.

AMD NTB uses an outbound translation window: a write to a BARxx memory
window offset is forwarded to (xlat_base | offset) in the peer's memory
rather than (xlat_base + offset). For that to be correct the
translation base must be aligned to a power of two >= the window length
so that no offset bit collides with a set bit in the base; the XLAT
register also requires at least 4K alignment. Report this requirement
to applications through a new mw_addr_align field in struct
ntb_dev_info, reject a misaligned base in amd_ntb_mw_set_trans, and
honour the field when reserving the memzone in the ntb example.

On the secondary side the device's own PCIe link status does not
reflect the true inter-host link, so the link speed and width are read
from the upstream switch port via sysfs.

Signed-off-by: Raghavendra Ningoji <raghavendra.ningoji@amd.com>
---
 drivers/raw/ntb/meson.build   |   3 +-
 drivers/raw/ntb/ntb.c         |  21 +
 drivers/raw/ntb/ntb_hw_amd.c  | 709 ++++++++++++++++++++++++++++++++++
 drivers/raw/ntb/ntb_hw_amd.h  | 114 ++++++
 drivers/raw/ntb/rte_pmd_ntb.h |  10 +
 examples/ntb/ntb_fwd.c        |  22 +-
 usertools/dpdk-devbind.py     |   4 +-
 7 files changed, 879 insertions(+), 4 deletions(-)
 create mode 100644 drivers/raw/ntb/ntb_hw_amd.c
 create mode 100644 drivers/raw/ntb/ntb_hw_amd.h

diff --git a/drivers/raw/ntb/meson.build b/drivers/raw/ntb/meson.build
index 9096f2b25a..d7a8f2d1ed 100644
--- a/drivers/raw/ntb/meson.build
+++ b/drivers/raw/ntb/meson.build
@@ -3,5 +3,6 @@
 
 deps += ['rawdev', 'mbuf', 'mempool', 'pci', 'bus_pci']
 sources = files('ntb.c',
-                'ntb_hw_intel.c')
+                'ntb_hw_intel.c',
+                'ntb_hw_amd.c')
 headers = files('rte_pmd_ntb.h')
diff --git a/drivers/raw/ntb/ntb.c b/drivers/raw/ntb/ntb.c
index 3a6a299081..b87141e4f4 100644
--- a/drivers/raw/ntb/ntb.c
+++ b/drivers/raw/ntb/ntb.c
@@ -20,12 +20,15 @@
 #include <rte_rawdev_pmd.h>
 
 #include "ntb_hw_intel.h"
+#include "ntb_hw_amd.h"
 #include "rte_pmd_ntb.h"
 #include "ntb.h"
 
 static const struct rte_pci_id pci_id_ntb_map[] = {
 	{ RTE_PCI_DEVICE(NTB_INTEL_VENDOR_ID, NTB_INTEL_DEV_ID_B2B_SKX) },
 	{ RTE_PCI_DEVICE(NTB_INTEL_VENDOR_ID, NTB_INTEL_DEV_ID_B2B_ICX) },
+	{ RTE_PCI_DEVICE(NTB_AMD_VENDOR_ID, NTB_AMD_DEV_ID_PRI) },
+	{ RTE_PCI_DEVICE(NTB_AMD_VENDOR_ID, NTB_AMD_DEV_ID_SEC) },
 	{ .vendor_id = 0, /* sentinel */ },
 };
 
@@ -846,6 +849,20 @@ ntb_dev_info_get(struct rte_rawdev *dev, rte_rawdev_obj_t dev_info,
 	info->mw_size_align = (uint8_t)(hw->pci_dev->id.vendor_id ==
 					NTB_INTEL_VENDOR_ID);
 
+	/**
+	 * AMD NTB uses an outbound translation window: writes to a BARxx
+	 * memory window are forwarded to the peer via the XLAT registers,
+	 * whose base must be 4K aligned. If the mw memzone base is not
+	 * aligned, the low bits are dropped and all window writes land at
+	 * the wrong offset. Report the required alignment so the memzone is
+	 * reserved correctly. Intel uses mw_size_align (a superset), so this
+	 * only matters for non-Intel vendors.
+	 */
+	if (hw->pci_dev->id.vendor_id == NTB_AMD_VENDOR_ID)
+		info->mw_addr_align = RTE_PGSIZE_4K;
+	else
+		info->mw_addr_align = 0;
+
 	if (!hw->queue_size || !hw->queue_pairs) {
 		NTB_LOG(ERR, "No queue size and queue num assigned.");
 		return -EAGAIN;
@@ -1406,6 +1423,10 @@ ntb_init_hw(struct rte_rawdev *dev, struct rte_pci_device *pci_dev)
 	case NTB_INTEL_DEV_ID_B2B_ICX:
 		hw->ntb_ops = &intel_ntb_ops;
 		break;
+	case NTB_AMD_DEV_ID_PRI:
+	case NTB_AMD_DEV_ID_SEC:
+		hw->ntb_ops = &amd_ntb_ops;
+		break;
 	default:
 		NTB_LOG(ERR, "Not supported device.");
 		return -EINVAL;
diff --git a/drivers/raw/ntb/ntb_hw_amd.c b/drivers/raw/ntb/ntb_hw_amd.c
new file mode 100644
index 0000000000..9861dcee57
--- /dev/null
+++ b/drivers/raw/ntb/ntb_hw_amd.c
@@ -0,0 +1,709 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2026 Advanced Micro Devices, Inc.
+ */
+
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <limits.h>
+
+#include <rte_io.h>
+#include <rte_eal.h>
+#include <rte_pci.h>
+#include <bus_pci_driver.h>
+#include <rte_rawdev.h>
+#include <rte_rawdev_pmd.h>
+#include <rte_malloc.h>
+#include <rte_memzone.h>
+
+#include "ntb.h"
+#include "ntb_hw_amd.h"
+
+static enum amd_ntb_bar amd_ntb_bar[] = {
+	AMD_NTB_BAR23,
+	AMD_NTB_BAR45,
+};
+
+static int
+amd_ntb_dev_init(const struct rte_rawdev *dev)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw;
+	uint32_t sideinfo;
+	int i, bar;
+
+	if (ntb == NULL) {
+		NTB_LOG(ERR, "Invalid device.");
+		return -EINVAL;
+	}
+
+	amd_hw = rte_zmalloc("amd_ntb_hw", sizeof(struct amd_ntb_hw), 0);
+	if (amd_hw == NULL) {
+		NTB_LOG(ERR, "Failed to allocate memory for amd_ntb_hw.");
+		return -ENOMEM;
+	}
+	ntb->pmd_private = amd_hw;
+
+	ntb->hw_addr = (char *)ntb->pci_dev->mem_resource[0].addr;
+	amd_hw->self_mmio = ntb->hw_addr;
+	amd_hw->peer_mmio = (char *)amd_hw->self_mmio + AMD_PEER_OFFSET;
+	amd_hw->int_mask = AMD_EVENT_INTMASK;
+
+	/* Bit 0 of the link status register carries the topology. */
+	sideinfo = rte_read32((char *)ntb->hw_addr + AMD_SIDEINFO_OFFSET);
+	if (sideinfo & AMD_SIDE_MASK)
+		ntb->topo = NTB_TOPO_SEC;
+	else
+		ntb->topo = NTB_TOPO_PRI;
+
+	NTB_LOG(INFO, "Device topology: %s (sideinfo 0x%" PRIx32 ")",
+		ntb->topo == NTB_TOPO_SEC ? "secondary" : "primary", sideinfo);
+
+	ntb->mw_cnt = AMD_MW_COUNT;
+	ntb->db_cnt = AMD_DB_COUNT;
+	/* The 16-register scratchpad bank is split into two halves, so only
+	 * AMD_SPAD_PER_SIDE registers are usable per side.
+	 */
+	ntb->spad_cnt = AMD_SPAD_PER_SIDE;
+
+	ntb->mw_size = rte_zmalloc("ntb_mw_size",
+				   ntb->mw_cnt * sizeof(uint64_t), 0);
+	if (ntb->mw_size == NULL) {
+		NTB_LOG(ERR, "Cannot allocate memory for mw size.");
+		rte_free(amd_hw);
+		ntb->pmd_private = NULL;
+		return -ENOMEM;
+	}
+	for (i = 0; i < ntb->mw_cnt; i++) {
+		bar = amd_ntb_bar[i];
+		ntb->mw_size[i] = ntb->pci_dev->mem_resource[bar].len;
+		NTB_LOG(INFO, "mw[%d] bar%d size 0x%" PRIx64, i, bar,
+			ntb->mw_size[i]);
+	}
+
+	/* The 16-register scratchpad bank is shared between both sides, so
+	 * split it into two disjoint sets to avoid clobbering: one side owns
+	 * offset 0, the other offset AMD_SPAD_SET_OFFSET.
+	 */
+	if (ntb->topo == NTB_TOPO_PRI) {
+		amd_hw->self_spad = 0;
+		amd_hw->peer_spad = AMD_SPAD_SET_OFFSET;
+	} else {
+		amd_hw->self_spad = AMD_SPAD_SET_OFFSET;
+		amd_hw->peer_spad = 0;
+	}
+
+	/* Reserve the last 2 scratchpad registers for application use. */
+	for (i = 0; i < NTB_SPAD_USER_MAX_NUM; i++)
+		ntb->spad_user_list[i] = ntb->spad_cnt;
+	ntb->spad_user_list[0] = ntb->spad_cnt - 2;
+	ntb->spad_user_list[1] = ntb->spad_cnt - 1;
+
+	return 0;
+}
+
+static void *
+amd_ntb_get_peer_mw_addr(const struct rte_rawdev *dev, int mw_idx)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+
+	if (mw_idx < 0 || mw_idx >= ntb->mw_cnt) {
+		NTB_LOG(ERR, "Invalid memory window index (0 - %u).",
+			ntb->mw_cnt - 1);
+		return NULL;
+	}
+
+	return ntb->pci_dev->mem_resource[amd_ntb_bar[mw_idx]].addr;
+}
+
+static int
+amd_ntb_mw_set_trans(const struct rte_rawdev *dev, int mw_idx,
+		     uint64_t addr, uint64_t size)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+	uint32_t xlat_off, limit_off;
+	uint64_t mw_size, reg_val;
+	uint8_t bar;
+
+	if (mw_idx < 0 || mw_idx >= ntb->mw_cnt) {
+		NTB_LOG(ERR, "Invalid memory window index (0 - %u).",
+			ntb->mw_cnt - 1);
+		return -EINVAL;
+	}
+
+	bar = amd_ntb_bar[mw_idx];
+	mw_size = ntb->pci_dev->mem_resource[bar].len;
+	if (size > mw_size) {
+		NTB_LOG(ERR, "Set translation size 0x%" PRIx64 " exceeds mw "
+			"size 0x%" PRIx64, size, mw_size);
+		return -EINVAL;
+	}
+
+	/*
+	 * The NTB uses an outbound translation window (see AMD doc 63939):
+	 * a write to a BAR window offset is forwarded to (xlat_base | offset)
+	 * in the peer's memory, rather than (xlat_base + offset). This is only
+	 * correct when the base has no set bits within the window's offset
+	 * range, i.e. the base must be aligned to a power of two >= the window
+	 * size. Otherwise offset bits collide with base bits and writes
+	 * silently alias to the wrong location. The memzone is reserved with
+	 * that alignment (see mw_addr_align in ntb_dev_info); reject a
+	 * misaligned base early.
+	 */
+	if (addr & (rte_align64pow2(size) - 1)) {
+		NTB_LOG(ERR, "mw%d translation base 0x%" PRIx64 " is not "
+			"aligned to a power of two >= size 0x%" PRIx64
+			"; window writes would alias.", mw_idx, addr, size);
+		return -EINVAL;
+	}
+
+	/* Program the peer's outbound translation window (S* registers) so
+	 * that peer writes into its BAR land in our local memory at 'addr'.
+	 */
+	if (mw_idx == 0) {
+		xlat_off = AMD_BAR23_XLAT_OFFSET;
+		limit_off = AMD_BAR23_LIMIT_OFFSET;
+	} else {
+		xlat_off = AMD_BAR45_XLAT_OFFSET;
+		limit_off = AMD_BAR45_LIMIT_OFFSET;
+	}
+
+	rte_write64(addr, (char *)amd_hw->peer_mmio + xlat_off);
+	reg_val = rte_read64((char *)amd_hw->peer_mmio + xlat_off);
+	if (reg_val != addr) {
+		NTB_LOG(ERR, "Failed to set mw%d translation.", mw_idx);
+		rte_write64(0, (char *)amd_hw->peer_mmio + xlat_off);
+		return -EIO;
+	}
+
+	rte_write64(size, (char *)amd_hw->peer_mmio + limit_off);
+	reg_val = rte_read64((char *)amd_hw->peer_mmio + limit_off);
+	if (reg_val != size) {
+		NTB_LOG(ERR, "Failed to set mw%d limit.", mw_idx);
+		rte_write64(0, (char *)amd_hw->peer_mmio + xlat_off);
+		rte_write64(0, (char *)amd_hw->peer_mmio + limit_off);
+		return -EIO;
+	}
+
+	return 0;
+}
+
+static void *
+amd_ntb_ioremap(const struct rte_rawdev *dev, uint64_t addr)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	void *mapped = NULL;
+	void *base;
+	int i;
+
+	for (i = 0; i < ntb->peer_used_mws; i++) {
+		if (addr >= ntb->peer_mw_base[i] &&
+		    addr <= ntb->peer_mw_base[i] + ntb->mw_size[i]) {
+			base = amd_ntb_get_peer_mw_addr(dev, i);
+			mapped = (void *)(size_t)(addr - ntb->peer_mw_base[i] +
+						  (size_t)base);
+			break;
+		}
+	}
+
+	return mapped;
+}
+
+/* Read link speed/width from the device's PCIe capability link status. */
+static int
+amd_ntb_read_pcie_link_status(struct ntb_hw *ntb, uint16_t *link_status)
+{
+	struct rte_pci_device *pci_dev = ntb->pci_dev;
+	uint8_t pos, cap_id, next;
+	uint16_t status;
+	int ret;
+
+	ret = rte_pci_read_config(pci_dev, &status, sizeof(status),
+				  RTE_PCI_STATUS);
+	if (ret != sizeof(status))
+		return -EIO;
+	if (!(status & RTE_PCI_STATUS_CAP_LIST))
+		return -ENOTSUP;
+
+	ret = rte_pci_read_config(pci_dev, &pos, sizeof(pos),
+				  RTE_PCI_CAPABILITY_LIST);
+	if (ret != sizeof(pos))
+		return -EIO;
+
+	while (pos) {
+		ret = rte_pci_read_config(pci_dev, &cap_id, sizeof(cap_id),
+					  pos);
+		if (ret != sizeof(cap_id))
+			return -EIO;
+		ret = rte_pci_read_config(pci_dev, &next, sizeof(next),
+					  pos + 1);
+		if (ret != sizeof(next))
+			return -EIO;
+		if (cap_id == RTE_PCI_CAP_ID_EXP) {
+			ret = rte_pci_read_config(pci_dev, link_status,
+						  sizeof(*link_status),
+						  pos + RTE_PCI_EXP_LNKSTA);
+			if (ret != sizeof(*link_status))
+				return -EIO;
+			return 0;
+		}
+		pos = next;
+	}
+
+	return -ENOTSUP;
+}
+
+/* Read the PCIe link status (LNKSTA) from an arbitrary device's config space
+ * exposed via sysfs. Used to query bridges that are not bound to this driver
+ * (e.g. the upstream switch above a secondary-side NTB), which rte_pci_*
+ * cannot access directly.
+ */
+static int
+amd_ntb_read_lnksta_sysfs(const char *config_path, uint16_t *link_status)
+{
+	uint8_t pos, cap_id, next;
+	uint16_t status;
+	int fd, ret = -EIO;
+
+	fd = open(config_path, O_RDONLY);
+	if (fd < 0)
+		return -errno;
+
+	if (pread(fd, &status, sizeof(status), RTE_PCI_STATUS) !=
+	    sizeof(status))
+		goto out;
+	if (!(status & RTE_PCI_STATUS_CAP_LIST))
+		goto out;
+
+	if (pread(fd, &pos, sizeof(pos), RTE_PCI_CAPABILITY_LIST) !=
+	    sizeof(pos))
+		goto out;
+
+	while (pos) {
+		if (pread(fd, &cap_id, sizeof(cap_id), pos) != sizeof(cap_id))
+			goto out;
+		if (pread(fd, &next, sizeof(next), pos + 1) != sizeof(next))
+			goto out;
+		if (cap_id == RTE_PCI_CAP_ID_EXP) {
+			if (pread(fd, link_status, sizeof(*link_status),
+				  pos + RTE_PCI_EXP_LNKSTA) ==
+			    sizeof(*link_status))
+				ret = 0;
+			goto out;
+		}
+		pos = next;
+	}
+out:
+	close(fd);
+	return ret;
+}
+
+/* On the secondary side the NTB device's own PCIe link status does not reflect
+ * the inter-host link. Mirror the Linux amd_ntb behaviour by walking up two
+ * bridge levels (device -> downstream switch -> upstream switch) and reading
+ * the upstream switch port's link status instead.
+ */
+static int
+amd_ntb_read_upstream_link_status(struct ntb_hw *ntb, uint16_t *link_status)
+{
+	char path[PATH_MAX];
+	char real[PATH_MAX];
+	char *p;
+	int i;
+
+	snprintf(path, sizeof(path),
+		 "/sys/bus/pci/devices/%04x:%02x:%02x.%x",
+		 ntb->pci_dev->addr.domain, ntb->pci_dev->addr.bus,
+		 ntb->pci_dev->addr.devid, ntb->pci_dev->addr.function);
+
+	if (realpath(path, real) == NULL)
+		return -errno;
+
+	/* Strip two trailing path components to reach the upstream switch. */
+	for (i = 0; i < 2; i++) {
+		p = strrchr(real, '/');
+		if (p == NULL || p == real)
+			return -ENOENT;
+		*p = '\0';
+	}
+
+	snprintf(path, sizeof(path), "%s/config", real);
+	return amd_ntb_read_lnksta_sysfs(path, link_status);
+}
+
+static int
+amd_ntb_get_link_status(const struct rte_rawdev *dev)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+	uint16_t link_status = 0;
+	uint32_t sideinfo;
+	int ret;
+
+	/* The link is usable once the peer has set its SIDE_READY bit. */
+	sideinfo = rte_read32((char *)amd_hw->peer_mmio + AMD_SIDEINFO_OFFSET);
+	ntb->link_status = !!(sideinfo & AMD_SIDE_READY);
+
+	if (!ntb->link_status) {
+		ntb->link_speed = NTB_SPEED_NONE;
+		ntb->link_width = NTB_WIDTH_NONE;
+		return 0;
+	}
+
+	/* The primary reads its own PCIe link status; the secondary must read
+	 * the upstream switch port above it. If the upstream read fails, fall
+	 * back to the local device so speed/width is still best-effort.
+	 */
+	if (ntb->topo == NTB_TOPO_SEC) {
+		ret = amd_ntb_read_upstream_link_status(ntb, &link_status);
+		if (ret != 0)
+			ret = amd_ntb_read_pcie_link_status(ntb, &link_status);
+	} else {
+		ret = amd_ntb_read_pcie_link_status(ntb, &link_status);
+	}
+
+	if (ret == 0) {
+		ntb->link_speed = AMD_LNK_STA_SPEED(link_status);
+		ntb->link_width = AMD_LNK_STA_WIDTH(link_status);
+	} else {
+		NTB_LOG(WARNING, "Failed to read PCIe link status (%d).", ret);
+		ntb->link_speed = NTB_SPEED_NONE;
+		ntb->link_width = NTB_WIDTH_NONE;
+	}
+
+	return 0;
+}
+
+static int
+amd_ntb_set_link(const struct rte_rawdev *dev, bool up)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+	void *mmio = amd_hw->self_mmio;
+	uint32_t reg;
+
+	reg = rte_read32((char *)mmio + AMD_SIDEINFO_OFFSET);
+	if (up) {
+		if (!(reg & AMD_SIDE_READY)) {
+			reg |= AMD_SIDE_READY;
+			rte_write32(reg, (char *)mmio + AMD_SIDEINFO_OFFSET);
+		}
+		reg = rte_read32((char *)mmio + AMD_CNTL_OFFSET);
+		reg |= (AMD_PMM_REG_CTL | AMD_SMM_REG_CTL);
+		rte_write32(reg, (char *)mmio + AMD_CNTL_OFFSET);
+	} else {
+		if (reg & AMD_SIDE_READY) {
+			reg &= ~AMD_SIDE_READY;
+			rte_write32(reg, (char *)mmio + AMD_SIDEINFO_OFFSET);
+		}
+		reg = rte_read32((char *)mmio + AMD_CNTL_OFFSET);
+		reg &= ~(AMD_PMM_REG_CTL | AMD_SMM_REG_CTL);
+		rte_write32(reg, (char *)mmio + AMD_CNTL_OFFSET);
+	}
+
+	return 0;
+}
+
+static uint32_t
+amd_ntb_spad_read(const struct rte_rawdev *dev, int spad, bool peer)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+	uint32_t offset;
+
+	if (spad < 0 || spad >= ntb->spad_cnt) {
+		NTB_LOG(ERR, "Invalid scratchpad index.");
+		return 0;
+	}
+
+	offset = peer ? amd_hw->peer_spad : amd_hw->self_spad;
+
+	return rte_read32((char *)ntb->hw_addr + AMD_SPAD_OFFSET + offset +
+			  (spad << 2));
+}
+
+static int
+amd_ntb_spad_write(const struct rte_rawdev *dev, int spad,
+		   bool peer, uint32_t spad_v)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+	uint32_t offset;
+
+	if (spad < 0 || spad >= ntb->spad_cnt) {
+		NTB_LOG(ERR, "Invalid scratchpad index.");
+		return -EINVAL;
+	}
+
+	offset = peer ? amd_hw->peer_spad : amd_hw->self_spad;
+
+	rte_write32(spad_v, (char *)ntb->hw_addr + AMD_SPAD_OFFSET + offset +
+		    (spad << 2));
+
+	return 0;
+}
+
+static uint64_t
+amd_ntb_db_read(const struct rte_rawdev *dev)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+
+	return (uint64_t)rte_read16((char *)amd_hw->self_mmio +
+				    AMD_DBSTAT_OFFSET);
+}
+
+static int
+amd_ntb_db_clear(const struct rte_rawdev *dev, uint64_t db_bits)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+
+	rte_write16((uint16_t)db_bits, (char *)amd_hw->self_mmio +
+		    AMD_DBSTAT_OFFSET);
+
+	return 0;
+}
+
+static int
+amd_ntb_db_set_mask(const struct rte_rawdev *dev, uint64_t db_mask)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+
+	if (db_mask & ~ntb->db_valid_mask)
+		return -EINVAL;
+
+	ntb->db_mask |= db_mask;
+	rte_write16((uint16_t)ntb->db_mask, (char *)amd_hw->self_mmio +
+		    AMD_DBMASK_OFFSET);
+
+	return 0;
+}
+
+static int
+amd_ntb_peer_db_set(const struct rte_rawdev *dev, uint8_t db_idx)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+
+	if (((uint64_t)1 << db_idx) & ~ntb->db_valid_mask) {
+		NTB_LOG(ERR, "Invalid doorbell.");
+		return -EINVAL;
+	}
+
+	rte_write16((uint16_t)1 << db_idx, (char *)ntb->hw_addr +
+		    AMD_DBREQ_OFFSET);
+
+	return 0;
+}
+
+static int
+amd_ntb_vector_bind(const struct rte_rawdev *dev __rte_unused,
+		    uint8_t intr __rte_unused, uint8_t msix __rte_unused)
+{
+	/* Each doorbell/event maps to its MSI-X vector by default. */
+	return 0;
+}
+
+/* Handshake: advertises the local configuration to the peer using the
+ * packed 8-register scratchpad layout, programs the peer's outbound
+ * translation windows and rings doorbell 0 to signal readiness.
+ */
+static int
+amd_dev_handshake(const struct rte_rawdev *dev)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	uint32_t info;
+	uint64_t base;
+	int i, ret;
+
+	info = (ntb->mw_cnt & 0xff) |
+	       ((uint32_t)(ntb->queue_pairs & 0xff) << 8) |
+	       ((uint32_t)(ntb->used_mw_num & 0xff) << 16);
+	ret = amd_ntb_spad_write(dev, AMD_SPAD_CNT_INFO, 1, info);
+	if (ret < 0)
+		return ret;
+
+	ret = amd_ntb_spad_write(dev, AMD_SPAD_QUEUE_SZ, 1, ntb->queue_size);
+	if (ret < 0)
+		return ret;
+
+	for (i = 0; i < ntb->used_mw_num; i++) {
+		/* Advertise the memzone virtual base (used by ioremap on the
+		 * peer) and program the translation with the IOVA.
+		 */
+		base = (uint64_t)(size_t)ntb->mz[i]->addr;
+		ret = amd_ntb_spad_write(dev, AMD_SPAD_MW0_BA_L + 2 * i, 1,
+					 (uint32_t)base);
+		if (ret < 0)
+			return ret;
+		ret = amd_ntb_spad_write(dev, AMD_SPAD_MW0_BA_H + 2 * i, 1,
+					 (uint32_t)(base >> 32));
+		if (ret < 0)
+			return ret;
+	}
+
+	for (i = 0; i < ntb->used_mw_num; i++) {
+		ret = amd_ntb_mw_set_trans(dev, i, ntb->mz[i]->iova,
+					   ntb->mz[i]->len);
+		if (ret < 0)
+			return ret;
+	}
+
+	/* Ring doorbell 0 to tell the peer the device is ready. */
+	return amd_ntb_peer_db_set(dev, 0);
+}
+
+/* Peer-config read at device start. Validates the peer's queue
+ * configuration and records the peer memory-window base addresses.
+ */
+static int
+amd_read_peer_config(const struct rte_rawdev *dev)
+{
+	struct ntb_hw *ntb = dev->dev_private;
+	uint32_t info, peer_qps, peer_qsz, lo, hi;
+	int i;
+
+	info = amd_ntb_spad_read(dev, AMD_SPAD_CNT_INFO, 0);
+	peer_qps = (info >> 8) & 0xff;
+	if (peer_qps != ntb->queue_pairs) {
+		NTB_LOG(ERR, "Inconsistent number of queues! (local: %u peer: %u)",
+			ntb->queue_pairs, peer_qps);
+		return -EINVAL;
+	}
+
+	peer_qsz = amd_ntb_spad_read(dev, AMD_SPAD_QUEUE_SZ, 0);
+	if (peer_qsz != ntb->queue_size) {
+		NTB_LOG(ERR, "Inconsistent queue size! (local: %u peer: %u)",
+			ntb->queue_size, peer_qsz);
+		return -EINVAL;
+	}
+
+	ntb->peer_used_mws = (info >> 16) & 0xff;
+	for (i = 0; i < ntb->peer_used_mws; i++) {
+		lo = amd_ntb_spad_read(dev, AMD_SPAD_MW0_BA_L + 2 * i, 0);
+		hi = amd_ntb_spad_read(dev, AMD_SPAD_MW0_BA_H + 2 * i, 0);
+		ntb->peer_mw_base[i] = ((uint64_t)hi << 32) | lo;
+	}
+
+	return 0;
+}
+
+static void
+amd_ntb_dev_interrupt_handler(void *param)
+{
+	struct rte_rawdev *dev = (struct rte_rawdev *)param;
+	struct ntb_hw *ntb = dev->dev_private;
+	struct amd_ntb_hw *amd_hw = ntb->pmd_private;
+	uint32_t event, ack, info;
+	uint64_t db_bits;
+	uint32_t peer_mw_cnt;
+
+	db_bits = amd_ntb_db_read(dev);
+
+	/* Doorbell 0: peer device is ready. */
+	if (db_bits & 1) {
+		amd_ntb_db_clear(dev, 1);
+		if (ntb->peer_dev_up)
+			return;
+
+		info = amd_ntb_spad_read(dev, AMD_SPAD_CNT_INFO, 0);
+		peer_mw_cnt = info & 0xff;
+		if (peer_mw_cnt != ntb->mw_cnt) {
+			NTB_LOG(ERR, "Peer mw cnt %u != local mw cnt %u.",
+				peer_mw_cnt, ntb->mw_cnt);
+			return;
+		}
+
+		ntb->peer_dev_up = 1;
+
+		/* Re-run the handshake so the device that came up second does
+		 * not miss the first doorbell (scratchpad/mw programming only
+		 * takes effect once both sides are up).
+		 */
+		if (amd_dev_handshake(dev) < 0) {
+			NTB_LOG(ERR, "Handshake work failed.");
+			return;
+		}
+
+		(*ntb->ntb_ops->get_link_status)(dev);
+		NTB_LOG(INFO, "Peer device up. Link speed %u width %u.",
+			ntb->link_speed, ntb->link_width);
+		return;
+	}
+
+	/* Doorbell 1: peer device is going down. */
+	if (db_bits & (1 << 1)) {
+		NTB_LOG(INFO, "DB1: Peer device is down.");
+		amd_ntb_db_clear(dev, (1 << 1));
+		ntb->peer_dev_up = 0;
+		(*ntb->ntb_ops->peer_db_set)(dev, 2);
+		return;
+	}
+
+	/* Doorbell 2: peer acknowledged our device-down request. */
+	if (db_bits & (1 << 2)) {
+		NTB_LOG(INFO, "DB2: Peer agrees device to be down.");
+		amd_ntb_db_clear(dev, (1 << 2));
+		ntb->peer_dev_up = 0;
+		return;
+	}
+
+	/* Any remaining doorbells. */
+	if (db_bits)
+		amd_ntb_db_clear(dev, db_bits);
+
+	/* Handle link/power-management events and acknowledge the SMU. */
+	event = rte_read32((char *)amd_hw->self_mmio + AMD_INTSTAT_OFFSET);
+	event &= AMD_EVENT_INTMASK;
+	if (event == 0)
+		return;
+
+	switch (event) {
+	case AMD_PEER_FLUSH_EVENT:
+		NTB_LOG(INFO, "Peer flush event.");
+		break;
+	case AMD_PEER_D0_EVENT:
+		ack = rte_read32((char *)amd_hw->self_mmio + AMD_PMESTAT_OFFSET);
+		if (ack & 0x1)
+			NTB_LOG(INFO, "D0 wakeup completed for NTB.");
+		/* fall through to ack the SMU */
+		/* Falls through. */
+	case AMD_PEER_RESET_EVENT:
+	case AMD_LINK_DOWN_EVENT:
+	case AMD_PEER_D3_EVENT:
+	case AMD_PEER_PMETO_EVENT:
+	case AMD_LINK_UP_EVENT:
+		ack = rte_read32((char *)amd_hw->self_mmio + AMD_SMUACK_OFFSET);
+		ack |= event;
+		rte_write32(ack, (char *)amd_hw->self_mmio + AMD_SMUACK_OFFSET);
+		break;
+	default:
+		NTB_LOG(ERR, "Unknown interrupt event 0x%" PRIx32, event);
+		break;
+	}
+}
+
+const struct ntb_dev_ops amd_ntb_ops = {
+	.ntb_dev_init		= amd_ntb_dev_init,
+	.get_peer_mw_addr	= amd_ntb_get_peer_mw_addr,
+	.mw_set_trans		= amd_ntb_mw_set_trans,
+	.ioremap		= amd_ntb_ioremap,
+	.get_link_status	= amd_ntb_get_link_status,
+	.set_link		= amd_ntb_set_link,
+	.spad_read		= amd_ntb_spad_read,
+	.spad_write		= amd_ntb_spad_write,
+	.db_read		= amd_ntb_db_read,
+	.db_clear		= amd_ntb_db_clear,
+	.db_set_mask		= amd_ntb_db_set_mask,
+	.peer_db_set		= amd_ntb_peer_db_set,
+	.vector_bind		= amd_ntb_vector_bind,
+	.interrupt_handler	= amd_ntb_dev_interrupt_handler,
+	.dev_handshake		= amd_dev_handshake,
+	.read_peer_config	= amd_read_peer_config,
+};
diff --git a/drivers/raw/ntb/ntb_hw_amd.h b/drivers/raw/ntb/ntb_hw_amd.h
new file mode 100644
index 0000000000..dafd62e30b
--- /dev/null
+++ b/drivers/raw/ntb/ntb_hw_amd.h
@@ -0,0 +1,114 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2026 Advanced Micro Devices, Inc.
+ */
+
+#ifndef _NTB_HW_AMD_H_
+#define _NTB_HW_AMD_H_
+
+#include <stdint.h>
+
+/* NTB vendor and device IDs (EPYC Embedded Turin/Genoa/Siena). */
+#define NTB_AMD_VENDOR_ID		0x1022
+#define NTB_AMD_DEV_ID_PRI		0x14c0	/* Primary NTB endpoint. */
+#define NTB_AMD_DEV_ID_SEC		0x14c3	/* Secondary NTB endpoint. */
+
+/* Device data (spec Table 12). */
+#define AMD_MW_COUNT			2
+#define AMD_DB_COUNT			16
+#define AMD_SPAD_COUNT			16
+#define AMD_MSIX_VECTOR_COUNT		24
+
+/* Peer register window: peer register = primary offset + 0x400. */
+#define AMD_PEER_OFFSET			0x400
+
+/* Primary-side MMIO register offsets (BAR0), spec Table 1. */
+#define AMD_CNTL_OFFSET			0x200	/* Link Control. */
+#define AMD_SPADMUTEX_OFFSET		0x20C	/* Scratchpad mutex. */
+#define AMD_SPAD_OFFSET			0x210	/* Scratchpad registers. */
+#define AMD_SIDEINFO_OFFSET		0x408	/* Link Status (PSIDE_INFO). */
+#define AMD_BAR23_LIMIT_OFFSET		0x418	/* MW BAR23 size. */
+#define AMD_BAR45_LIMIT_OFFSET		0x420	/* MW BAR45 size. */
+#define AMD_BAR23_XLAT_OFFSET		0x438	/* MW BAR23 translation. */
+#define AMD_BAR45_XLAT_OFFSET		0x440	/* MW BAR45 translation. */
+#define AMD_DBFM_OFFSET			0x450	/* Doorbell flush mode. */
+#define AMD_DBREQ_OFFSET		0x454	/* Doorbell request. */
+#define AMD_DBMASK_OFFSET		0x45C	/* Doorbell mask. */
+#define AMD_DBSTAT_OFFSET		0x460	/* Doorbell status. */
+#define AMD_INTMASK_OFFSET		0x470	/* Interrupt mask. */
+#define AMD_INTSTAT_OFFSET		0x474	/* Interrupt status. */
+#define AMD_PMESTAT_OFFSET		0x480	/* PME status. */
+#define AMD_SMUACK_OFFSET		0x4A0	/* SMU control (PSMU_ACK). */
+
+/* Link Status Register (0x408) bits, spec Table 5. */
+#define AMD_SIDE_MASK			(1 << 0) /* 0: primary, 1: secondary. */
+#define AMD_SIDE_READY			(1 << 1) /* Side link ready. */
+
+/* Link Control Register (0x200) bits, spec Table 2. */
+#define AMD_SMM_REG_CTL			(1 << 20)
+#define AMD_PMM_REG_CTL			(1 << 21)
+
+/* Interrupt Status/Mask event bits, spec Table 7. */
+#define AMD_PEER_FLUSH_EVENT		(1 << 0)
+#define AMD_PEER_RESET_EVENT		(1 << 1)
+#define AMD_PEER_D3_EVENT		(1 << 2)
+#define AMD_PEER_PMETO_EVENT		(1 << 3)
+#define AMD_PEER_D0_EVENT		(1 << 4)
+#define AMD_LINK_UP_EVENT		(1 << 5)
+#define AMD_LINK_DOWN_EVENT		(1 << 6)
+#define AMD_EVENT_INTMASK		(AMD_PEER_FLUSH_EVENT | \
+					 AMD_PEER_RESET_EVENT | \
+					 AMD_PEER_D3_EVENT | \
+					 AMD_PEER_PMETO_EVENT | \
+					 AMD_PEER_D0_EVENT | \
+					 AMD_LINK_UP_EVENT | \
+					 AMD_LINK_DOWN_EVENT)
+
+/* PCIe link status decoding (link speed/width). */
+#define AMD_LNK_STA_SPEED_MASK		0x000f
+#define AMD_LNK_STA_WIDTH_MASK		0x03f0
+#define AMD_LNK_STA_SPEED(x)		((x) & AMD_LNK_STA_SPEED_MASK)
+#define AMD_LNK_STA_WIDTH(x)		(((x) & AMD_LNK_STA_WIDTH_MASK) >> 4)
+
+/* The 16-register scratchpad bank is shared between both sides (there is no
+ * separate peer-scratchpad window). It is split into two disjoint 8-register
+ * sets so each side owns one half (offset 0 for one side, 0x20 for the other)
+ * and no mutex is required. Because only 8 registers are available per side,
+ * the driver uses its own packed handshake layout instead of the built-in
+ * protocol.
+ */
+#define AMD_SPAD_SET_OFFSET		0x20
+#define AMD_SPAD_PER_SIDE		(AMD_SPAD_COUNT >> 1)
+
+/* Packed scratchpad handshake layout (indices within an 8-register set).
+ * Indices 6 and 7 are reserved for application user scratchpads.
+ */
+enum amd_spad_idx {
+	AMD_SPAD_CNT_INFO = 0,	/* num_mws | num_qps<<8 | used_mws<<16 */
+	AMD_SPAD_QUEUE_SZ,	/* queue_size */
+	AMD_SPAD_MW0_BA_L,	/* mw0 base address, low 32 bits */
+	AMD_SPAD_MW0_BA_H,	/* mw0 base address, high 32 bits */
+	AMD_SPAD_MW1_BA_L,	/* mw1 base address, low 32 bits */
+	AMD_SPAD_MW1_BA_H,	/* mw1 base address, high 32 bits */
+};
+
+enum amd_ntb_bar {
+	AMD_NTB_BAR23 = 2,
+	AMD_NTB_BAR45 = 4,
+};
+
+/* Hardware private data. */
+struct amd_ntb_hw {
+	void *self_mmio;	/* BAR0 primary register window. */
+	void *peer_mmio;	/* self_mmio + AMD_PEER_OFFSET. */
+
+	uint32_t self_spad;	/* Byte offset of local scratchpad set. */
+	uint32_t peer_spad;	/* Byte offset of peer scratchpad set. */
+
+	uint32_t int_mask;
+	uint32_t peer_status;
+	uint32_t ctl_status;
+};
+
+extern const struct ntb_dev_ops amd_ntb_ops;
+
+#endif /* _NTB_HW_AMD_H_ */
diff --git a/drivers/raw/ntb/rte_pmd_ntb.h b/drivers/raw/ntb/rte_pmd_ntb.h
index 76da3be026..59a2ad6849 100644
--- a/drivers/raw/ntb/rte_pmd_ntb.h
+++ b/drivers/raw/ntb/rte_pmd_ntb.h
@@ -27,6 +27,16 @@ struct ntb_dev_info {
 	uint8_t mw_size_align;
 	uint8_t mw_cnt;
 	uint64_t *mw_size;
+	/**< Minimum alignment (bytes) required for the mw translation base
+	 * address, and a flag that the base must additionally be aligned to a
+	 * power of two >= the mw length. 0 means no extra alignment beyond
+	 * cache line. AMD NTB uses an outbound translation window that forms
+	 * the target as (xlat_base | offset) instead of (xlat_base + offset),
+	 * so the base must be size-aligned to avoid offset bits colliding with
+	 * base bits; it also requires at least 4K alignment for the XLAT
+	 * register.
+	 */
+	uint64_t mw_addr_align;
 };
 
 struct ntb_dev_config {
diff --git a/examples/ntb/ntb_fwd.c b/examples/ntb/ntb_fwd.c
index 33f3c1ef17..7cc4e22147 100644
--- a/examples/ntb/ntb_fwd.c
+++ b/examples/ntb/ntb_fwd.c
@@ -1146,8 +1146,10 @@ ntb_mbuf_pool_create(uint16_t mbuf_seg_size, uint32_t nb_mbuf,
 		if (!left_sz)
 			break;
 		snprintf(mz_name, sizeof(mz_name), "ntb_mw_%d", mz_id);
-		align = ntb_info.mw_size_align ? ntb_info.mw_size[mz_id] :
-			RTE_CACHE_LINE_SIZE;
+		if (ntb_info.mw_size_align)
+			align = ntb_info.mw_size[mz_id];
+		else
+			align = RTE_CACHE_LINE_SIZE;
 		/* Reserve ntb header space on memzone 0. */
 		max_mz_len = mz_id ? ntb_info.mw_size[mz_id] :
 			     ntb_info.mw_size[mz_id] - ntb_info.ntb_hdr_size;
@@ -1155,6 +1157,22 @@ ntb_mbuf_pool_create(uint16_t mbuf_seg_size, uint32_t nb_mbuf,
 			(max_mz_len / total_elt_sz * total_elt_sz);
 		if (!mz_len)
 			continue;
+		/*
+		 * Some NTB hardware (e.g. AMD) uses an outbound translation
+		 * window that forms the target as (xlat_base | offset) rather
+		 * than (xlat_base + offset). For that to be correct the memzone
+		 * base must be aligned to a power of two >= its length, so that
+		 * no offset bit collides with a set bit in the base address.
+		 * Honour that requirement when the driver reports mw_addr_align.
+		 */
+		if (ntb_info.mw_addr_align) {
+			uint64_t pow2_align = rte_align64pow2(mz_len);
+
+			if (pow2_align > align)
+				align = pow2_align;
+			if (ntb_info.mw_addr_align > align)
+				align = ntb_info.mw_addr_align;
+		}
 		mz = rte_memzone_reserve_aligned(mz_name, mz_len, socket_id,
 					RTE_MEMZONE_IOVA_CONTIG, align);
 		if (mz == NULL) {
diff --git a/usertools/dpdk-devbind.py b/usertools/dpdk-devbind.py
index e72f238aba..cf5747b003 100755
--- a/usertools/dpdk-devbind.py
+++ b/usertools/dpdk-devbind.py
@@ -78,6 +78,8 @@
                  'SVendor': None, 'SDevice': None}
 intel_ntb_icx = {'Class': '06', 'Vendor': '8086', 'Device': '347e',
                  'SVendor': None, 'SDevice': None}
+amd_ntb = {'Class': '06', 'Vendor': '1022', 'Device': '14c0,14c3',
+           'SVendor': None, 'SDevice': None}
 
 cnxk_sso = {'Class': '08', 'Vendor': '177d', 'Device': 'a0f9,a0fa',
             'SVendor': None, 'SDevice': None}
@@ -105,7 +107,7 @@
 regex_devices = [cn9k_ree]
 ml_devices = [cnxk_ml]
 misc_devices = [cnxk_bphy, cnxk_bphy_cgx, cnxk_inl_dev,
-                intel_ntb_skx, intel_ntb_icx,
+                intel_ntb_skx, intel_ntb_icx, amd_ntb,
                 virtio_blk]
 
 # global dict ethernet devices present. Dictionary indexed by PCI address.
-- 
2.34.1

[PATCH v1 3/3] doc: add AMD NTB support

[Raghavendra Ningoji]

Document AMD EPYC Embedded NTB support in the raw/ntb guide, add a
release note entry, and claim maintainership of the AMD NTB support.

Signed-off-by: Raghavendra Ningoji <raghavendra.ningoji@amd.com>
---
 MAINTAINERS                            |  1 +
 doc/guides/rawdevs/ntb.rst             | 18 +++++++++++++++++-
 doc/guides/rel_notes/release_26_11.rst |  8 ++++++++
 3 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/MAINTAINERS b/MAINTAINERS
index e99a65d197..13306345fd 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -1613,6 +1613,7 @@ F: doc/guides/rawdevs/cnxk_rvu_lf.rst
 
 NTB
 M: Jingjing Wu <jingjing.wu@intel.com>
+M: Bhagyada Modali <bhagyada.modali@amd.com>
 F: drivers/raw/ntb/
 F: doc/guides/rawdevs/ntb.rst
 F: examples/ntb/
diff --git a/doc/guides/rawdevs/ntb.rst b/doc/guides/rawdevs/ntb.rst
index 3721b4880e..c89eaccac8 100644
--- a/doc/guides/rawdevs/ntb.rst
+++ b/doc/guides/rawdevs/ntb.rst
@@ -37,10 +37,26 @@ then reboot.
 - Set ``PCIe PLL SSC (Spread Spectrum Clocking)`` as ``Disabled``, on both hosts.
   This is a hardware requirement when using Re-timer Cards.
 
+AMD EPYC Embedded NTB
+---------------------
+
+The driver also supports the NTB endpoints integrated in AMD EPYC Embedded
+"Turin", "Genoa" and "Siena" processors. These use a primary/secondary
+topology rather than the Intel back-to-back topology: one endpoint is
+enumerated as the primary (device ID ``0x14c0``) and the other as the
+secondary (device ID ``0x14c3``). The BIOS on both systems performs NTB link
+training; no additional NTB-specific BIOS options are required beyond enabling
+the NTB endpoints.
+
+The AMD NTB hardware exposes two memory windows (BAR23 and BAR45), 16
+doorbells and a single shared 16-register scratchpad bank. The scratchpad
+bank is split into two disjoint 8-register sets, one owned by each side, so
+the driver uses a packed handshake layout that fits within 8 registers.
+
 Device Setup
 ------------
 
-The Intel NTB devices need to be bound to a DPDK-supported kernel driver
+The NTB devices need to be bound to a DPDK-supported kernel driver
 to use, i.e. igb_uio, vfio. The ``dpdk-devbind.py`` script can be used to
 show devices status and to bind them to a suitable kernel driver. They will
 appear under the category of "Misc (rawdev) devices".
diff --git a/doc/guides/rel_notes/release_26_11.rst b/doc/guides/rel_notes/release_26_11.rst
index c8cc86295d..0b364ccbd4 100644
--- a/doc/guides/rel_notes/release_26_11.rst
+++ b/doc/guides/rel_notes/release_26_11.rst
@@ -55,6 +55,14 @@ New Features
      Also, make sure to start the actual text at the margin.
      =======================================================
 
+* **Added AMD NTB support to the NTB rawdev driver.**
+
+  Added support for the NTB endpoints integrated in AMD EPYC Embedded
+  "Turin", "Genoa" and "Siena" processors to the ``raw/ntb`` driver.
+  The NTB rawdev framework was generalized to support multiple vendors,
+  with AMD-specific hardware access, a primary/secondary topology and a
+  packed scratchpad handshake.
+
 
 Removed Items
 -------------
-- 
2.34.1

Re: [PATCH v1 0/3] raw/ntb: add AMD NTB support

[Raghavendra Ningoji]

Hi all,

Adding David, Bruce, Sachin, Hemant  to Cc for visibility on this series.

The full series is on the list here:
https://inbox.dpdk.org/dev/?q=raw%2Fntb%3A+add+AMD+NTB+support

Thanks,
Raghavendra