Merge branch 'marvell-cn10k'

Geetha sowjanya says:

====================
Add Marvell CN10K support

The current admin function (AF) driver and the netdev driver supports
OcteonTx2 silicon variants. The same OcteonTx2's
Resource Virtualization Unit (RVU) is carried forward to the next-gen
silicon ie OcteonTx3, with some changes and feature enhancements.

This patch set adds support for OcteonTx3 (CN10K) silicon and gets
the drivers to the same level as OcteonTx2. No new OcteonTx3 specific
features are added.

Changes cover below HW level differences
- PCIe BAR address changes wrt shared mailbox memory region
- Receive buffer freeing to HW
- Transmit packet's descriptor submission to HW
- Programmable HW interface identifiers (channels)
- Increased MTU support
- A Serdes MAC block (RPM) configuration

v5-v6
Rebased on top of latest net-next branch.

v4-v5
Fixed sparse warnings.

v3-v4
Fixed compiler warnings.

v2-v3
Reposting as a single thread.
Rebased on top latest net-next branch.

v1-v2
Fixed check-patch reported issues.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

MAINTAINERS

@@ -10719,6 +10719,8 @@ M:	Sunil Goutham <sgoutham@marvell.com>
 M:	Linu Cherian <lcherian@marvell.com>
 M:	Geetha sowjanya <gakula@marvell.com>
 M:	Jerin Jacob <jerinj@marvell.com>
+M:	hariprasad <hkelam@marvell.com>
+M:	Subbaraya Sundeep <sbhatta@marvell.com>
 L:	netdev@vger.kernel.org
 S:	Supported
 F:	Documentation/networking/device_drivers/ethernet/marvell/octeontx2.rst

drivers/net/ethernet/marvell/octeontx2/af/Makefile

@@ -4,10 +4,10 @@
 #
 
 ccflags-y += -I$(src)
-obj-$(CONFIG_OCTEONTX2_MBOX) += octeontx2_mbox.o
-obj-$(CONFIG_OCTEONTX2_AF) += octeontx2_af.o
+obj-$(CONFIG_OCTEONTX2_MBOX) += rvu_mbox.o
+obj-$(CONFIG_OCTEONTX2_AF) += rvu_af.o
 
-octeontx2_mbox-y := mbox.o rvu_trace.o
-octeontx2_af-y := cgx.o rvu.o rvu_cgx.o rvu_npa.o rvu_nix.o \
+rvu_mbox-y := mbox.o rvu_trace.o
+rvu_af-y := cgx.o rvu.o rvu_cgx.o rvu_npa.o rvu_nix.o \
 		  rvu_reg.o rvu_npc.o rvu_debugfs.o ptp.o rvu_npc_fs.o \
-		  rvu_cpt.o rvu_devlink.o
+		  rvu_cpt.o rvu_devlink.o rpm.o rvu_cn10k.o

drivers/net/ethernet/marvell/octeontx2/af/cgx.c

@@ -21,47 +21,11 @@
 #include <linux/of_net.h>
 
 #include "cgx.h"
+#include "rvu.h"
+#include "lmac_common.h"
 
-#define DRV_NAME	"octeontx2-cgx"
-#define DRV_STRING      "Marvell OcteonTX2 CGX/MAC Driver"
-
-/**
- * struct lmac
- * @wq_cmd_cmplt:	waitq to keep the process blocked until cmd completion
- * @cmd_lock:		Lock to serialize the command interface
- * @resp:		command response
- * @link_info:		link related information
- * @event_cb:		callback for linkchange events
- * @event_cb_lock:	lock for serializing callback with unregister
- * @cmd_pend:		flag set before new command is started
- *			flag cleared after command response is received
- * @cgx:		parent cgx port
- * @lmac_id:		lmac port id
- * @name:		lmac port name
- */
-struct lmac {
-	wait_queue_head_t wq_cmd_cmplt;
-	struct mutex cmd_lock;
-	u64 resp;
-	struct cgx_link_user_info link_info;
-	struct cgx_event_cb event_cb;
-	spinlock_t event_cb_lock;
-	bool cmd_pend;
-	struct cgx *cgx;
-	u8 lmac_id;
-	char *name;
-};
-
-struct cgx {
-	void __iomem		*reg_base;
-	struct pci_dev		*pdev;
-	u8			cgx_id;
-	u8			lmac_count;
-	struct lmac		*lmac_idmap[MAX_LMAC_PER_CGX];
-	struct			work_struct cgx_cmd_work;
-	struct			workqueue_struct *cgx_cmd_workq;
-	struct list_head	cgx_list;
-};
+#define DRV_NAME	"Marvell-CGX/RPM"
+#define DRV_STRING      "Marvell CGX/RPM Driver"
 
 static LIST_HEAD(cgx_list);
 
@@ -77,22 +41,45 @@ static int cgx_fwi_link_change(struct cgx *cgx, int lmac_id, bool en);
 /* Supported devices */
 static const struct pci_device_id cgx_id_table[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_OCTEONTX2_CGX) },
+	{ PCI_DEVICE(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_CN10K_RPM) },
 	{ 0, }  /* end of table */
 };
 
 MODULE_DEVICE_TABLE(pci, cgx_id_table);
 
-static void cgx_write(struct cgx *cgx, u64 lmac, u64 offset, u64 val)
+static bool is_dev_rpm(void *cgxd)
+{
+	struct cgx *cgx = cgxd;
+
+	return (cgx->pdev->device == PCI_DEVID_CN10K_RPM);
+}
+
+bool is_lmac_valid(struct cgx *cgx, int lmac_id)
+{
+	return cgx && test_bit(lmac_id, &cgx->lmac_bmap);
+}
+
+struct mac_ops *get_mac_ops(void *cgxd)
 {
-	writeq(val, cgx->reg_base + (lmac << 18) + offset);
+	if (!cgxd)
+		return cgxd;
+
+	return ((struct cgx *)cgxd)->mac_ops;
 }
 
-static u64 cgx_read(struct cgx *cgx, u64 lmac, u64 offset)
+void cgx_write(struct cgx *cgx, u64 lmac, u64 offset, u64 val)
 {
-	return readq(cgx->reg_base + (lmac << 18) + offset);
+	writeq(val, cgx->reg_base + (lmac << cgx->mac_ops->lmac_offset) +
+	       offset);
 }
 
-static inline struct lmac *lmac_pdata(u8 lmac_id, struct cgx *cgx)
+u64 cgx_read(struct cgx *cgx, u64 lmac, u64 offset)
+{
+	return readq(cgx->reg_base + (lmac << cgx->mac_ops->lmac_offset) +
+		     offset);
+}
+
+struct lmac *lmac_pdata(u8 lmac_id, struct cgx *cgx)
 {
 	if (!cgx || lmac_id >= MAX_LMAC_PER_CGX)
 		return NULL;
@@ -136,6 +123,20 @@ void *cgx_get_pdata(int cgx_id)
 	return NULL;
 }
 
+void cgx_lmac_write(int cgx_id, int lmac_id, u64 offset, u64 val)
+{
+	struct cgx *cgx_dev = cgx_get_pdata(cgx_id);
+
+	cgx_write(cgx_dev, lmac_id, offset, val);
+}
+
+u64 cgx_lmac_read(int cgx_id, int lmac_id, u64 offset)
+{
+	struct cgx *cgx_dev = cgx_get_pdata(cgx_id);
+
+	return cgx_read(cgx_dev, lmac_id, offset);
+}
+
 int cgx_get_cgxid(void *cgxd)
 {
 	struct cgx *cgx = cgxd;
@@ -186,8 +187,10 @@ static u64 mac2u64 (u8 *mac_addr)
 int cgx_lmac_addr_set(u8 cgx_id, u8 lmac_id, u8 *mac_addr)
 {
 	struct cgx *cgx_dev = cgx_get_pdata(cgx_id);
+	struct mac_ops *mac_ops;
 	u64 cfg;
 
+	mac_ops = cgx_dev->mac_ops;
 	/* copy 6bytes from macaddr */
 	/* memcpy(&cfg, mac_addr, 6); */
 
@@ -206,8 +209,11 @@ int cgx_lmac_addr_set(u8 cgx_id, u8 lmac_id, u8 *mac_addr)
 u64 cgx_lmac_addr_get(u8 cgx_id, u8 lmac_id)
 {
 	struct cgx *cgx_dev = cgx_get_pdata(cgx_id);
+	struct mac_ops *mac_ops;
 	u64 cfg;
 
+	mac_ops = cgx_dev->mac_ops;
+
 	cfg = cgx_read(cgx_dev, 0, CGXX_CMRX_RX_DMAC_CAM0 + lmac_id * 0x8);
 	return cfg & CGX_RX_DMAC_ADR_MASK;
 }
@@ -216,15 +222,16 @@ int cgx_set_pkind(void *cgxd, u8 lmac_id, int pkind)
 {
 	struct cgx *cgx = cgxd;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 
 	cgx_write(cgx, lmac_id, CGXX_CMRX_RX_ID_MAP, (pkind & 0x3F));
 	return 0;
 }
 
-static inline u8 cgx_get_lmac_type(struct cgx *cgx, int lmac_id)
+static u8 cgx_get_lmac_type(void *cgxd, int lmac_id)
 {
+	struct cgx *cgx = cgxd;
 	u64 cfg;
 
 	cfg = cgx_read(cgx, lmac_id, CGXX_CMRX_CFG);
@@ -238,10 +245,10 @@ int cgx_lmac_internal_loopback(void *cgxd, int lmac_id, bool enable)
 	u8 lmac_type;
 	u64 cfg;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 
-	lmac_type = cgx_get_lmac_type(cgx, lmac_id);
+	lmac_type = cgx->mac_ops->get_lmac_type(cgx, lmac_id);
 	if (lmac_type == LMAC_MODE_SGMII || lmac_type == LMAC_MODE_QSGMII) {
 		cfg = cgx_read(cgx, lmac_id, CGXX_GMP_PCS_MRX_CTL);
 		if (enable)
@@ -263,11 +270,13 @@ int cgx_lmac_internal_loopback(void *cgxd, int lmac_id, bool enable)
 void cgx_lmac_promisc_config(int cgx_id, int lmac_id, bool enable)
 {
 	struct cgx *cgx = cgx_get_pdata(cgx_id);
+	struct mac_ops *mac_ops;
 	u64 cfg = 0;
 
 	if (!cgx)
 		return;
 
+	mac_ops = cgx->mac_ops;
 	if (enable) {
 		/* Enable promiscuous mode on LMAC */
 		cfg = cgx_read(cgx, lmac_id, CGXX_CMRX_RX_DMAC_CTL0);
@@ -325,7 +334,7 @@ int cgx_get_rx_stats(void *cgxd, int lmac_id, int idx, u64 *rx_stat)
 {
 	struct cgx *cgx = cgxd;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 	*rx_stat =  cgx_read(cgx, lmac_id, CGXX_CMRX_RX_STAT0 + (idx * 8));
 	return 0;
@@ -335,12 +344,17 @@ int cgx_get_tx_stats(void *cgxd, int lmac_id, int idx, u64 *tx_stat)
 {
 	struct cgx *cgx = cgxd;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 	*tx_stat = cgx_read(cgx, lmac_id, CGXX_CMRX_TX_STAT0 + (idx * 8));
 	return 0;
 }
 
+u64 cgx_features_get(void *cgxd)
+{
+	return ((struct cgx *)cgxd)->hw_features;
+}
+
 static int cgx_set_fec_stats_count(struct cgx_link_user_info *linfo)
 {
 	if (!linfo->fec)
@@ -400,7 +414,7 @@ int cgx_lmac_rx_tx_enable(void *cgxd, int lmac_id, bool enable)
 	struct cgx *cgx = cgxd;
 	u64 cfg;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 
 	cfg = cgx_read(cgx, lmac_id, CGXX_CMRX_CFG);
@@ -417,7 +431,7 @@ int cgx_lmac_tx_enable(void *cgxd, int lmac_id, bool enable)
 	struct cgx *cgx = cgxd;
 	u64 cfg, last;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 
 	cfg = cgx_read(cgx, lmac_id, CGXX_CMRX_CFG);
@@ -432,13 +446,16 @@ int cgx_lmac_tx_enable(void *cgxd, int lmac_id, bool enable)
 	return !!(last & DATA_PKT_TX_EN);
 }
 
-int cgx_lmac_get_pause_frm(void *cgxd, int lmac_id,
-			   u8 *tx_pause, u8 *rx_pause)
+static int cgx_lmac_get_pause_frm_status(void *cgxd, int lmac_id,
+					 u8 *tx_pause, u8 *rx_pause)
 {
 	struct cgx *cgx = cgxd;
 	u64 cfg;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (is_dev_rpm(cgx))
+		return 0;
+
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 
 	cfg = cgx_read(cgx, lmac_id, CGXX_SMUX_RX_FRM_CTL);
@@ -449,13 +466,16 @@ int cgx_lmac_get_pause_frm(void *cgxd, int lmac_id,
 	return 0;
 }
 
-int cgx_lmac_set_pause_frm(void *cgxd, int lmac_id,
-			   u8 tx_pause, u8 rx_pause)
+static int cgx_lmac_enadis_pause_frm(void *cgxd, int lmac_id,
+				     u8 tx_pause, u8 rx_pause)
 {
 	struct cgx *cgx = cgxd;
 	u64 cfg;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (is_dev_rpm(cgx))
+		return 0;
+
+	if (!is_lmac_valid(cgx, lmac_id))
 		return -ENODEV;
 
 	cfg = cgx_read(cgx, lmac_id, CGXX_SMUX_RX_FRM_CTL);
@@ -479,11 +499,12 @@ int cgx_lmac_set_pause_frm(void *cgxd, int lmac_id,
 	return 0;
 }
 
-static void cgx_lmac_pause_frm_config(struct cgx *cgx, int lmac_id, bool enable)
+static void cgx_lmac_pause_frm_config(void *cgxd, int lmac_id, bool enable)
 {
+	struct cgx *cgx = cgxd;
 	u64 cfg;
 
-	if (!cgx || lmac_id >= cgx->lmac_count)
+	if (!is_lmac_valid(cgx, lmac_id))
 		return;
 	if (enable) {
 		/* Enable receive pause frames */
@@ -541,6 +562,9 @@ void cgx_lmac_ptp_config(void *cgxd, int lmac_id, bool enable)
 	if (!cgx)
 		return;
 
+	if (is_dev_rpm(cgx))
+		return;
+
 	if (enable) {
 		/* Enable inbound PTP timestamping */
 		cfg = cgx_read(cgx, lmac_id, CGXX_GMP_GMI_RXX_FRM_CTL);
@@ -563,7 +587,7 @@ void cgx_lmac_ptp_config(void *cgxd, int lmac_id, bool enable)
 }
 
 /* CGX Firmware interface low level support */
-static int cgx_fwi_cmd_send(u64 req, u64 *resp, struct lmac *lmac)
+int cgx_fwi_cmd_send(u64 req, u64 *resp, struct lmac *lmac)
 {
 	struct cgx *cgx = lmac->cgx;
 	struct device *dev;
@@ -611,8 +635,7 @@ static int cgx_fwi_cmd_send(u64 req, u64 *resp, struct lmac *lmac)
 	return err;
 }
 
-static inline int cgx_fwi_cmd_generic(u64 req, u64 *resp,
-				      struct cgx *cgx, int lmac_id)
+int cgx_fwi_cmd_generic(u64 req, u64 *resp, struct cgx *cgx, int lmac_id)
 {
 	struct lmac *lmac;
 	int err;
@@ -885,12 +908,16 @@ static inline bool cgx_event_is_linkevent(u64 event)
 
 static irqreturn_t cgx_fwi_event_handler(int irq, void *data)
 {
+	u64 event, offset, clear_bit;
 	struct lmac *lmac = data;
 	struct cgx *cgx;
-	u64 event;
 
 	cgx = lmac->cgx;
 
+	/* Clear SW_INT for RPM and CMR_INT for CGX */
+	offset     = cgx->mac_ops->int_register;
+	clear_bit  = cgx->mac_ops->int_ena_bit;
+
 	event = cgx_read(cgx, lmac->lmac_id, CGX_EVENT_REG);
 
 	if (!FIELD_GET(EVTREG_ACK, event))
@@ -926,7 +953,7 @@ static irqreturn_t cgx_fwi_event_handler(int irq, void *data)
 	 * Ack the interrupt register as well.
 	 */
 	cgx_write(lmac->cgx, lmac->lmac_id, CGX_EVENT_REG, 0);
-	cgx_write(lmac->cgx, lmac->lmac_id, CGXX_CMRX_INT, FW_CGX_INT);
+	cgx_write(lmac->cgx, lmac->lmac_id, offset, clear_bit);
 
 	return IRQ_HANDLED;
 }
@@ -970,14 +997,16 @@ int cgx_get_fwdata_base(u64 *base)
 {
 	u64 req = 0, resp;
 	struct cgx *cgx;
+	int first_lmac;
 	int err;
 
 	cgx = list_first_entry_or_null(&cgx_list, struct cgx, cgx_list);
 	if (!cgx)
 		return -ENXIO;
 
+	first_lmac = find_first_bit(&cgx->lmac_bmap, MAX_LMAC_PER_CGX);
 	req = FIELD_SET(CMDREG_ID, CGX_CMD_GET_FWD_BASE, req);
-	err = cgx_fwi_cmd_generic(req, &resp, cgx, 0);
+	err = cgx_fwi_cmd_generic(req, &resp, cgx, first_lmac);
 	if (!err)
 		*base = FIELD_GET(RESP_FWD_BASE, resp);
 
@@ -1056,10 +1085,11 @@ static int cgx_fwi_link_change(struct cgx *cgx, int lmac_id, bool enable)
 
 static inline int cgx_fwi_read_version(u64 *resp, struct cgx *cgx)
 {
+	int first_lmac = find_first_bit(&cgx->lmac_bmap, MAX_LMAC_PER_CGX);
 	u64 req = 0;
 
 	req = FIELD_SET(CMDREG_ID, CGX_CMD_GET_FW_VER, req);
-	return cgx_fwi_cmd_generic(req, resp, cgx, 0);
+	return cgx_fwi_cmd_generic(req, resp, cgx, first_lmac);
 }
 
 static int cgx_lmac_verify_fwi_version(struct cgx *cgx)
@@ -1092,8 +1122,8 @@ static void cgx_lmac_linkup_work(struct work_struct *work)
 	struct device *dev = &cgx->pdev->dev;
 	int i, err;
 
-	/* Do Link up for all the lmacs */
-	for (i = 0; i < cgx->lmac_count; i++) {
+	/* Do Link up for all the enabled lmacs */
+	for_each_set_bit(i, &cgx->lmac_bmap, MAX_LMAC_PER_CGX) {
 		err = cgx_fwi_link_change(cgx, i, true);
 		if (err)
 			dev_info(dev, "cgx port %d:%d Link up command failed\n",
@@ -1113,12 +1143,77 @@ int cgx_lmac_linkup_start(void *cgxd)
 	return 0;
 }
 
+static void cgx_lmac_get_fifolen(struct cgx *cgx)
+{
+	u64 cfg;
+
+	cfg = cgx_read(cgx, 0, CGX_CONST);
+	cgx->mac_ops->fifo_len = FIELD_GET(CGX_CONST_RXFIFO_SIZE, cfg);
+}
+
+static int cgx_configure_interrupt(struct cgx *cgx, struct lmac *lmac,
+				   int cnt, bool req_free)
+{
+	struct mac_ops *mac_ops = cgx->mac_ops;
+	u64 offset, ena_bit;
+	unsigned int irq;
+	int err;
+
+	irq      = pci_irq_vector(cgx->pdev, mac_ops->lmac_fwi +
+				  cnt * mac_ops->irq_offset);
+	offset   = mac_ops->int_set_reg;
+	ena_bit  = mac_ops->int_ena_bit;
+
+	if (req_free) {
+		free_irq(irq, lmac);
+		return 0;
+	}
+
+	err = request_irq(irq, cgx_fwi_event_handler, 0, lmac->name, lmac);
+	if (err)
+		return err;
+
+	/* Enable interrupt */
+	cgx_write(cgx, lmac->lmac_id, offset, ena_bit);
+	return 0;
+}
+
+int cgx_get_nr_lmacs(void *cgxd)
+{
+	struct cgx *cgx = cgxd;
+
+	return cgx_read(cgx, 0, CGXX_CMRX_RX_LMACS) & 0x7ULL;
+}
+
+u8 cgx_get_lmacid(void *cgxd, u8 lmac_index)
+{
+	struct cgx *cgx = cgxd;
+
+	return cgx->lmac_idmap[lmac_index]->lmac_id;
+}
+
+unsigned long cgx_get_lmac_bmap(void *cgxd)
+{
+	struct cgx *cgx = cgxd;
+
+	return cgx->lmac_bmap;
+}
+
 static int cgx_lmac_init(struct cgx *cgx)
 {
 	struct lmac *lmac;
+	u64 lmac_list;
 	int i, err;
 
-	cgx->lmac_count = cgx_read(cgx, 0, CGXX_CMRX_RX_LMACS) & 0x7;
+	cgx_lmac_get_fifolen(cgx);
+
+	cgx->lmac_count = cgx->mac_ops->get_nr_lmacs(cgx);
+	/* lmac_list specifies which lmacs are enabled
+	 * when bit n is set to 1, LMAC[n] is enabled
+	 */
+	if (cgx->mac_ops->non_contiguous_serdes_lane)
+		lmac_list = cgx_read(cgx, 0, CGXX_CMRX_RX_LMACS) & 0xFULL;
+
 	if (cgx->lmac_count > MAX_LMAC_PER_CGX)
 		cgx->lmac_count = MAX_LMAC_PER_CGX;
 
@@ -1132,24 +1227,25 @@ static int cgx_lmac_init(struct cgx *cgx)
 			goto err_lmac_free;
 		}
 		sprintf(lmac->name, "cgx_fwi_%d_%d", cgx->cgx_id, i);
-		lmac->lmac_id = i;
+		if (cgx->mac_ops->non_contiguous_serdes_lane) {
+			lmac->lmac_id = __ffs64(lmac_list);
+			lmac_list   &= ~BIT_ULL(lmac->lmac_id);
+		} else {
+			lmac->lmac_id = i;
+		}
+
 		lmac->cgx = cgx;
 		init_waitqueue_head(&lmac->wq_cmd_cmplt);
 		mutex_init(&lmac->cmd_lock);
 		spin_lock_init(&lmac->event_cb_lock);
-		err = request_irq(pci_irq_vector(cgx->pdev,
-						 CGX_LMAC_FWI + i * 9),
-				   cgx_fwi_event_handler, 0, lmac->name, lmac);
+		err = cgx_configure_interrupt(cgx, lmac, lmac->lmac_id, false);
 		if (err)
 			goto err_irq;
 
-		/* Enable interrupt */
-		cgx_write(cgx, lmac->lmac_id, CGXX_CMRX_INT_ENA_W1S,
-			  FW_CGX_INT);
-
 		/* Add reference */
-		cgx->lmac_idmap[i] = lmac;
-		cgx_lmac_pause_frm_config(cgx, i, true);
+		cgx->lmac_idmap[lmac->lmac_id] = lmac;
+		cgx->mac_ops->mac_pause_frm_config(cgx, lmac->lmac_id, true);
+		set_bit(lmac->lmac_id, &cgx->lmac_bmap);
 	}
 
 	return cgx_lmac_verify_fwi_version(cgx);
@@ -1173,12 +1269,12 @@ static int cgx_lmac_exit(struct cgx *cgx)
 	}
 
 	/* Free all lmac related resources */
-	for (i = 0; i < cgx->lmac_count; i++) {
-		cgx_lmac_pause_frm_config(cgx, i, false);
+	for_each_set_bit(i, &cgx->lmac_bmap, MAX_LMAC_PER_CGX) {
 		lmac = cgx->lmac_idmap[i];
 		if (!lmac)
 			continue;
-		free_irq(pci_irq_vector(cgx->pdev, CGX_LMAC_FWI + i * 9), lmac);
+		cgx->mac_ops->mac_pause_frm_config(cgx, lmac->lmac_id, false);
+		cgx_configure_interrupt(cgx, lmac, lmac->lmac_id, true);
 		kfree(lmac->name);
 		kfree(lmac);
 	}
@@ -1186,6 +1282,37 @@ static int cgx_lmac_exit(struct cgx *cgx)
 	return 0;
 }
 
+static void cgx_populate_features(struct cgx *cgx)
+{
+	if (is_dev_rpm(cgx))
+		cgx->hw_features =  (RVU_MAC_RPM | RVU_LMAC_FEAT_FC);
+	else
+		cgx->hw_features = (RVU_LMAC_FEAT_FC | RVU_LMAC_FEAT_PTP);
+}
+
+static struct mac_ops	cgx_mac_ops    = {
+	.name		=       "cgx",
+	.csr_offset	=       0,
+	.lmac_offset    =       18,
+	.int_register	=       CGXX_CMRX_INT,
+	.int_set_reg	=       CGXX_CMRX_INT_ENA_W1S,
+	.irq_offset	=       9,
+	.int_ena_bit    =       FW_CGX_INT,
+	.lmac_fwi	=	CGX_LMAC_FWI,
+	.non_contiguous_serdes_lane = false,
+	.rx_stats_cnt   =       9,
+	.tx_stats_cnt   =       18,
+	.get_nr_lmacs	=	cgx_get_nr_lmacs,
+	.get_lmac_type  =       cgx_get_lmac_type,
+	.mac_lmac_intl_lbk =    cgx_lmac_internal_loopback,
+	.mac_get_rx_stats  =	cgx_get_rx_stats,
+	.mac_get_tx_stats  =	cgx_get_tx_stats,
+	.mac_enadis_rx_pause_fwding =	cgx_lmac_enadis_rx_pause_fwding,
+	.mac_get_pause_frm_status =	cgx_lmac_get_pause_frm_status,
+	.mac_enadis_pause_frm =		cgx_lmac_enadis_pause_frm,
+	.mac_pause_frm_config =		cgx_lmac_pause_frm_config,
+};
+
 static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 {
 	struct device *dev = &pdev->dev;
@@ -1199,6 +1326,12 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	pci_set_drvdata(pdev, cgx);
 
+	/* Use mac_ops to get MAC specific features */
+	if (pdev->device == PCI_DEVID_CN10K_RPM)
+		cgx->mac_ops = rpm_get_mac_ops();
+	else
+		cgx->mac_ops = &cgx_mac_ops;
+
 	err = pci_enable_device(pdev);
 	if (err) {
 		dev_err(dev, "Failed to enable PCI device\n");
@@ -1220,7 +1353,7 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 		goto err_release_regions;
 	}
 
-	nvec = CGX_NVEC;
+	nvec = pci_msix_vec_count(cgx->pdev);
 	err = pci_alloc_irq_vectors(pdev, nvec, nvec, PCI_IRQ_MSIX);
 	if (err < 0 || err != nvec) {
 		dev_err(dev, "Request for %d msix vectors failed, err %d\n",
@@ -1244,6 +1377,10 @@ static int cgx_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	cgx_link_usertable_init();
 
+	cgx_populate_features(cgx);
+
+	mutex_init(&cgx->lock);
+
 	err = cgx_lmac_init(cgx);
 	if (err)
 		goto err_release_lmac;
@@ -1267,8 +1404,10 @@ static void cgx_remove(struct pci_dev *pdev)
 {
 	struct cgx *cgx = pci_get_drvdata(pdev);
 
-	cgx_lmac_exit(cgx);
-	list_del(&cgx->cgx_list);
+	if (cgx) {
+		cgx_lmac_exit(cgx);
+		list_del(&cgx->cgx_list);
+	}
 	pci_free_irq_vectors(pdev);
 	pci_release_regions(pdev);
 	pci_disable_device(pdev);

drivers/net/ethernet/marvell/octeontx2/af/cgx.h

@@ -13,6 +13,7 @@
 
 #include "mbox.h"
 #include "cgx_fw_if.h"
+#include "rpm.h"
 
  /* PCI device IDs */
 #define	PCI_DEVID_OCTEONTX2_CGX		0xA059
@@ -42,12 +43,12 @@
 #define CGXX_CMRX_RX_ID_MAP		0x060
 #define CGXX_CMRX_RX_STAT0		0x070
 #define CGXX_CMRX_RX_LMACS		0x128
-#define CGXX_CMRX_RX_DMAC_CTL0		0x1F8
+#define CGXX_CMRX_RX_DMAC_CTL0		(0x1F8 + mac_ops->csr_offset)
 #define CGX_DMAC_CTL0_CAM_ENABLE	BIT_ULL(3)
 #define CGX_DMAC_CAM_ACCEPT		BIT_ULL(3)
 #define CGX_DMAC_MCAST_MODE		BIT_ULL(1)
 #define CGX_DMAC_BCAST_MODE		BIT_ULL(0)
-#define CGXX_CMRX_RX_DMAC_CAM0		0x200
+#define CGXX_CMRX_RX_DMAC_CAM0		(0x200 + mac_ops->csr_offset)
 #define CGX_DMAC_CAM_ADDR_ENABLE	BIT_ULL(48)
 #define CGXX_CMRX_RX_DMAC_CAM1		0x400
 #define CGX_RX_DMAC_ADR_MASK		GENMASK_ULL(47, 0)
@@ -55,6 +56,7 @@
 #define CGXX_SCRATCH0_REG		0x1050
 #define CGXX_SCRATCH1_REG		0x1058
 #define CGX_CONST			0x2000
+#define CGX_CONST_RXFIFO_SIZE	        GENMASK_ULL(23, 0)
 #define CGXX_SPUX_CONTROL1		0x10000
 #define CGXX_SPUX_LNX_FEC_CORR_BLOCKS	0x10700
 #define CGXX_SPUX_LNX_FEC_UNCORR_BLOCKS	0x10800
@@ -86,7 +88,6 @@
 #define CGX_CMD_TIMEOUT			2200 /* msecs */
 #define DEFAULT_PAUSE_TIME		0x7FF
 
-#define CGX_NVEC			37
 #define CGX_LMAC_FWI			0
 
 enum  cgx_nix_stat_type {
@@ -157,5 +158,11 @@ int cgx_get_fec_stats(void *cgxd, int lmac_id, struct cgx_fec_stats_rsp *rsp);
 int cgx_get_phy_fec_stats(void *cgxd, int lmac_id);
 int cgx_set_link_mode(void *cgxd, struct cgx_set_link_mode_args args,
 		      int cgx_id, int lmac_id);
-
+u64 cgx_features_get(void *cgxd);
+struct mac_ops *get_mac_ops(void *cgxd);
+int cgx_get_nr_lmacs(void *cgxd);
+u8 cgx_get_lmacid(void *cgxd, u8 lmac_index);
+unsigned long cgx_get_lmac_bmap(void *cgxd);
+void cgx_lmac_write(int cgx_id, int lmac_id, u64 offset, u64 val);
+u64 cgx_lmac_read(int cgx_id, int lmac_id, u64 offset);
 #endif /* CGX_H */

drivers/net/ethernet/marvell/octeontx2/af/cgx_fw_if.h

@@ -204,6 +204,7 @@ enum cgx_cmd_own {
  * CGX_STAT_SUCCESS
  */
 #define RESP_FWD_BASE		GENMASK_ULL(56, 9)
+#define RESP_LINKSTAT_LMAC_TYPE                GENMASK_ULL(35, 28)
 
 /* Response to cmd ID - CGX_CMD_LINK_BRING_UP/DOWN, event ID CGX_EVT_LINK_CHANGE
  * status can be either CGX_STAT_FAIL or CGX_STAT_SUCCESS

drivers/net/ethernet/marvell/octeontx2/af/common.h

@@ -155,6 +155,8 @@ enum nix_scheduler {
 #define	NIC_HW_MIN_FRS			40
 #define	NIC_HW_MAX_FRS			9212
 #define	SDP_HW_MAX_FRS			65535
+#define CN10K_LMAC_LINK_MAX_FRS		16380 /* 16k - FCS */
+#define CN10K_LBK_LINK_MAX_FRS		65535 /* 64k */
 
 /* NIX RX action operation*/
 #define NIX_RX_ACTIONOP_DROP		(0x0ull)
@@ -191,6 +193,9 @@ enum nix_scheduler {
 #define NIX_LINK_LBK(a)			(12 + (a))
 #define NIX_CHAN_CGX_LMAC_CHX(a, b, c)	(0x800 + 0x100 * (a) + 0x10 * (b) + (c))
 #define NIX_CHAN_LBK_CHX(a, b)		(0 + 0x100 * (a) + (b))
+#define NIX_CHAN_SDP_CH_START		(0x700ull)
+
+#define SDP_CHANNELS			256
 
 /* NIX LSO format indices.
  * As of now TSO is the only one using, so statically assigning indices.

drivers/net/ethernet/marvell/octeontx2/af/lmac_common.h

+/* SPDX-License-Identifier: GPL-2.0 */
+/*  Marvell OcteonTx2 RPM driver
+ *
+ * Copyright (C) 2020 Marvell.
+ */
+
+#ifndef LMAC_COMMON_H
+#define LMAC_COMMON_H
+
+#include "rvu.h"
+#include "cgx.h"
+/**
+ * struct lmac
+ * @wq_cmd_cmplt:	waitq to keep the process blocked until cmd completion
+ * @cmd_lock:		Lock to serialize the command interface
+ * @resp:		command response
+ * @link_info:		link related information
+ * @event_cb:		callback for linkchange events
+ * @event_cb_lock:	lock for serializing callback with unregister
+ * @cmd_pend:		flag set before new command is started
+ *			flag cleared after command response is received
+ * @cgx:		parent cgx port
+ * @lmac_id:		lmac port id
+ * @name:		lmac port name
+ */
+struct lmac {
+	wait_queue_head_t wq_cmd_cmplt;
+	/* Lock to serialize the command interface */
+	struct mutex cmd_lock;
+	u64 resp;
+	struct cgx_link_user_info link_info;
+	struct cgx_event_cb event_cb;
+	/* lock for serializing callback with unregister */
+	spinlock_t event_cb_lock;
+	bool cmd_pend;
+	struct cgx *cgx;
+	u8 lmac_id;
+	char *name;
+};
+
+/* CGX & RPM has different feature set
+ * update the structure fields with different one
+ */
+struct mac_ops {
+	char		       *name;
+	/* Features like RXSTAT, TXSTAT, DMAC FILTER csrs differs by fixed
+	 * bar offset for example
+	 * CGX DMAC_CTL0  0x1f8
+	 * RPM DMAC_CTL0  0x4ff8
+	 */
+	u64			csr_offset;
+	/* For ATF to send events to kernel, there is no dedicated interrupt
+	 * defined hence CGX uses OVERFLOW bit in CMR_INT. RPM block supports
+	 * SW_INT so that ATF triggers this interrupt after processing of
+	 * requested command
+	 */
+	u64			int_register;
+	u64			int_set_reg;
+	/* lmac offset is different is RPM */
+	u8			lmac_offset;
+	u8			irq_offset;
+	u8			int_ena_bit;
+	u8			lmac_fwi;
+	u32			fifo_len;
+	bool			non_contiguous_serdes_lane;
+	/* RPM & CGX differs in number of Receive/transmit stats */
+	u8			rx_stats_cnt;
+	u8			tx_stats_cnt;
+	/* Incase of RPM get number of lmacs from RPMX_CMR_RX_LMACS[LMAC_EXIST]
+	 * number of setbits in lmac_exist tells number of lmacs
+	 */
+	int			(*get_nr_lmacs)(void *cgx);
+	u8                      (*get_lmac_type)(void *cgx, int lmac_id);
+	int                     (*mac_lmac_intl_lbk)(void *cgx, int lmac_id,
+						     bool enable);
+	/* Register Stats related functions */
+	int			(*mac_get_rx_stats)(void *cgx, int lmac_id,
+						    int idx, u64 *rx_stat);
+	int			(*mac_get_tx_stats)(void *cgx, int lmac_id,
+						    int idx, u64 *tx_stat);
+
+	/* Enable LMAC Pause Frame Configuration */
+	void			(*mac_enadis_rx_pause_fwding)(void *cgxd,
+							      int lmac_id,
+							      bool enable);
+
+	int			(*mac_get_pause_frm_status)(void *cgxd,
+							    int lmac_id,
+							    u8 *tx_pause,
+							    u8 *rx_pause);
+
+	int			(*mac_enadis_pause_frm)(void *cgxd,
+							int lmac_id,
+							u8 tx_pause,
+							u8 rx_pause);
+
+	void			(*mac_pause_frm_config)(void  *cgxd,
+							int lmac_id,
+							bool enable);
+};
+
+struct cgx {
+	void __iomem		*reg_base;
+	struct pci_dev		*pdev;
+	u8			cgx_id;
+	u8			lmac_count;
+	struct lmac		*lmac_idmap[MAX_LMAC_PER_CGX];
+	struct			work_struct cgx_cmd_work;
+	struct			workqueue_struct *cgx_cmd_workq;
+	struct list_head	cgx_list;
+	u64			hw_features;
+	struct mac_ops		*mac_ops;
+	unsigned long		lmac_bmap; /* bitmap of enabled lmacs */
+	/* Lock to serialize read/write of global csrs like
+	 * RPMX_MTI_STAT_DATA_HI_CDC etc
+	 */
+	struct mutex		lock;
+};
+
+typedef struct cgx rpm_t;
+
+/* Function Declarations */
+void cgx_write(struct cgx *cgx, u64 lmac, u64 offset, u64 val);
+u64 cgx_read(struct cgx *cgx, u64 lmac, u64 offset);
+struct lmac *lmac_pdata(u8 lmac_id, struct cgx *cgx);
+int cgx_fwi_cmd_send(u64 req, u64 *resp, struct lmac *lmac);
+int cgx_fwi_cmd_generic(u64 req, u64 *resp, struct cgx *cgx, int lmac_id);
+bool is_lmac_valid(struct cgx *cgx, int lmac_id);
+struct mac_ops *rpm_get_mac_ops(void);
+
+#endif /* LMAC_COMMON_H */

drivers/net/ethernet/marvell/octeontx2/af/mbox.c

@@ -20,9 +20,9 @@ static const u16 msgs_offset = ALIGN(sizeof(struct mbox_hdr), MBOX_MSG_ALIGN);
 
 void __otx2_mbox_reset(struct otx2_mbox *mbox, int devid)
 {
-	void *hw_mbase = mbox->hwbase + (devid * MBOX_SIZE);
 	struct otx2_mbox_dev *mdev = &mbox->dev[devid];
 	struct mbox_hdr *tx_hdr, *rx_hdr;
+	void *hw_mbase = mdev->hwbase;
 
 	tx_hdr = hw_mbase + mbox->tx_start;
 	rx_hdr = hw_mbase + mbox->rx_start;
@@ -56,12 +56,9 @@ void otx2_mbox_destroy(struct otx2_mbox *mbox)
 }
 EXPORT_SYMBOL(otx2_mbox_destroy);
 
-int otx2_mbox_init(struct otx2_mbox *mbox, void *hwbase, struct pci_dev *pdev,
-		   void *reg_base, int direction, int ndevs)
+static int otx2_mbox_setup(struct otx2_mbox *mbox, struct pci_dev *pdev,
+			   void *reg_base, int direction, int ndevs)
 {
-	struct otx2_mbox_dev *mdev;
-	int devid;
-
 	switch (direction) {
 	case MBOX_DIR_AFPF:
 	case MBOX_DIR_PFVF:
@@ -121,7 +118,6 @@ int otx2_mbox_init(struct otx2_mbox *mbox, void *hwbase, struct pci_dev *pdev,
 	}
 
 	mbox->reg_base = reg_base;
-	mbox->hwbase = hwbase;
 	mbox->pdev = pdev;
 
 	mbox->dev = kcalloc(ndevs, sizeof(struct otx2_mbox_dev), GFP_KERNEL);
@@ -129,11 +125,27 @@ int otx2_mbox_init(struct otx2_mbox *mbox, void *hwbase, struct pci_dev *pdev,
 		otx2_mbox_destroy(mbox);
 		return -ENOMEM;
 	}
-
 	mbox->ndevs = ndevs;
+
+	return 0;
+}
+
+int otx2_mbox_init(struct otx2_mbox *mbox, void *hwbase, struct pci_dev *pdev,
+		   void *reg_base, int direction, int ndevs)
+{
+	struct otx2_mbox_dev *mdev;
+	int devid, err;
+
+	err = otx2_mbox_setup(mbox, pdev, reg_base, direction, ndevs);
+	if (err)
+		return err;
+
+	mbox->hwbase = hwbase;
+
 	for (devid = 0; devid < ndevs; devid++) {
 		mdev = &mbox->dev[devid];
 		mdev->mbase = mbox->hwbase + (devid * MBOX_SIZE);
+		mdev->hwbase = mdev->mbase;
 		spin_lock_init(&mdev->mbox_lock);
 		/* Init header to reset value */
 		otx2_mbox_reset(mbox, devid);
@@ -143,6 +155,35 @@ int otx2_mbox_init(struct otx2_mbox *mbox, void *hwbase, struct pci_dev *pdev,
 }
 EXPORT_SYMBOL(otx2_mbox_init);
 
+/* Initialize mailbox with the set of mailbox region addresses
+ * in the array hwbase.
+ */
+int otx2_mbox_regions_init(struct otx2_mbox *mbox, void **hwbase,
+			   struct pci_dev *pdev, void *reg_base,
+			   int direction, int ndevs)
+{
+	struct otx2_mbox_dev *mdev;
+	int devid, err;
+
+	err = otx2_mbox_setup(mbox, pdev, reg_base, direction, ndevs);
+	if (err)
+		return err;
+
+	mbox->hwbase = hwbase[0];
+
+	for (devid = 0; devid < ndevs; devid++) {
+		mdev = &mbox->dev[devid];
+		mdev->mbase = hwbase[devid];
+		mdev->hwbase = hwbase[devid];
+		spin_lock_init(&mdev->mbox_lock);
+		/* Init header to reset value */
+		otx2_mbox_reset(mbox, devid);
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(otx2_mbox_regions_init);
+
 int otx2_mbox_wait_for_rsp(struct otx2_mbox *mbox, int devid)
 {
 	unsigned long timeout = jiffies + msecs_to_jiffies(MBOX_RSP_TIMEOUT);
@@ -175,9 +216,9 @@ EXPORT_SYMBOL(otx2_mbox_busy_poll_for_rsp);
 
 void otx2_mbox_msg_send(struct otx2_mbox *mbox, int devid)
 {
-	void *hw_mbase = mbox->hwbase + (devid * MBOX_SIZE);
 	struct otx2_mbox_dev *mdev = &mbox->dev[devid];
 	struct mbox_hdr *tx_hdr, *rx_hdr;
+	void *hw_mbase = mdev->hwbase;
 
 	tx_hdr = hw_mbase + mbox->tx_start;
 	rx_hdr = hw_mbase + mbox->rx_start;

drivers/net/ethernet/marvell/octeontx2/af/mbox.h

@@ -52,6 +52,7 @@
 
 struct otx2_mbox_dev {
 	void	    *mbase;   /* This dev's mbox region */
+	void	    *hwbase;
 	spinlock_t  mbox_lock;
 	u16         msg_size; /* Total msg size to be sent */
 	u16         rsp_size; /* Total rsp size to be sure the reply is ok */
@@ -98,6 +99,9 @@ void otx2_mbox_destroy(struct otx2_mbox *mbox);
 int otx2_mbox_init(struct otx2_mbox *mbox, void __force *hwbase,
 		   struct pci_dev *pdev, void __force *reg_base,
 		   int direction, int ndevs);
+int otx2_mbox_regions_init(struct otx2_mbox *mbox, void __force **hwbase,
+			   struct pci_dev *pdev, void __force *reg_base,
+			   int direction, int ndevs);
 void otx2_mbox_msg_send(struct otx2_mbox *mbox, int devid);
 int otx2_mbox_wait_for_rsp(struct otx2_mbox *mbox, int devid);
 int otx2_mbox_busy_poll_for_rsp(struct otx2_mbox *mbox, int devid);
@@ -155,6 +159,9 @@ M(CGX_GET_PHY_FEC_STATS, 0x212, cgx_get_phy_fec_stats, msg_req, msg_rsp) \
 M(CGX_FW_DATA_GET,	0x213, cgx_get_aux_link_info, msg_req, cgx_fw_data) \
 M(CGX_SET_LINK_MODE,	0x214, cgx_set_link_mode, cgx_set_link_mode_req,\
 			       cgx_set_link_mode_rsp)	\
+M(CGX_FEATURES_GET,	0x215, cgx_features_get, msg_req,		\
+			       cgx_features_info_msg)			\
+M(RPM_STATS,		0x216, rpm_stats, msg_req, rpm_stats_rsp)	\
  /* NPA mbox IDs (range 0x400 - 0x5FF) */				\
 /* NPA mbox IDs (range 0x400 - 0x5FF) */				\
 M(NPA_LF_ALLOC,		0x400, npa_lf_alloc,				\
@@ -244,6 +251,9 @@ M(NIX_BP_ENABLE,	0x8016, nix_bp_enable, nix_bp_cfg_req,	\
 				nix_bp_cfg_rsp)	\
 M(NIX_BP_DISABLE,	0x8017, nix_bp_disable, nix_bp_cfg_req, msg_rsp) \
 M(NIX_GET_MAC_ADDR, 0x8018, nix_get_mac_addr, msg_req, nix_get_mac_addr_rsp) \
+M(NIX_CN10K_AQ_ENQ,	0x8019, nix_cn10k_aq_enq, nix_cn10k_aq_enq_req, \
+				nix_cn10k_aq_enq_rsp)			\
+M(NIX_GET_HW_INFO,	0x801a, nix_get_hw_info, msg_req, nix_hw_info)
 
 /* Messages initiated by AF (range 0xC00 - 0xDFF) */
 #define MBOX_UP_CGX_MESSAGES						\
@@ -361,7 +371,7 @@ struct get_hw_cap_rsp {
 
 struct cgx_stats_rsp {
 	struct mbox_msghdr hdr;
-#define CGX_RX_STATS_COUNT	13
+#define CGX_RX_STATS_COUNT	9
 #define CGX_TX_STATS_COUNT	18
 	u64 rx_stats[CGX_RX_STATS_COUNT];
 	u64 tx_stats[CGX_TX_STATS_COUNT];
@@ -478,6 +488,25 @@ struct cgx_set_link_mode_rsp {
 	int status;
 };
 
+#define RVU_LMAC_FEAT_FC		BIT_ULL(0) /* pause frames */
+#define RVU_LMAC_FEAT_PTP		BIT_ULL(1) /* precison time protocol */
+#define RVU_MAC_VERSION			BIT_ULL(2)
+#define RVU_MAC_CGX			BIT_ULL(3)
+#define RVU_MAC_RPM			BIT_ULL(4)
+
+struct cgx_features_info_msg {
+	struct mbox_msghdr hdr;
+	u64    lmac_features;
+};
+
+struct rpm_stats_rsp {
+	struct mbox_msghdr hdr;
+#define RPM_RX_STATS_COUNT		43
+#define RPM_TX_STATS_COUNT		34
+	u64 rx_stats[RPM_RX_STATS_COUNT];
+	u64 tx_stats[RPM_TX_STATS_COUNT];
+};
+
 /* NPA mbox message formats */
 
 /* NPA mailbox error codes
@@ -632,6 +661,39 @@ struct nix_lf_free_req {
 	u64 flags;
 };
 
+/* CN10K NIX AQ enqueue msg */
+struct nix_cn10k_aq_enq_req {
+	struct mbox_msghdr hdr;
+	u32  qidx;
+	u8 ctype;
+	u8 op;
+	union {
+		struct nix_cn10k_rq_ctx_s rq;
+		struct nix_cn10k_sq_ctx_s sq;
+		struct nix_cq_ctx_s cq;
+		struct nix_rsse_s   rss;
+		struct nix_rx_mce_s mce;
+	};
+	union {
+		struct nix_cn10k_rq_ctx_s rq_mask;
+		struct nix_cn10k_sq_ctx_s sq_mask;
+		struct nix_cq_ctx_s cq_mask;
+		struct nix_rsse_s   rss_mask;
+		struct nix_rx_mce_s mce_mask;
+	};
+};
+
+struct nix_cn10k_aq_enq_rsp {
+	struct mbox_msghdr hdr;
+	union {
+		struct nix_cn10k_rq_ctx_s rq;
+		struct nix_cn10k_sq_ctx_s sq;
+		struct nix_cq_ctx_s cq;
+		struct nix_rsse_s   rss;
+		struct nix_rx_mce_s mce;
+	};
+};
+
 /* NIX AQ enqueue msg */
 struct nix_aq_enq_req {
 	struct mbox_msghdr hdr;
@@ -896,6 +958,12 @@ struct nix_bp_cfg_rsp {
 	u8	chan_cnt; /* Number of channel for which bpids are assigned */
 };
 
+struct nix_hw_info {
+	struct mbox_msghdr hdr;
+	u16 max_mtu;
+	u16 min_mtu;
+};
+
 /* NPC mbox message structs */
 
 #define NPC_MCAM_ENTRY_INVALID	0xFFFF

drivers/net/ethernet/marvell/octeontx2/af/ptp.c

@@ -21,6 +21,9 @@
 #define PCI_SUBSYS_DEVID_OCTX2_95XX_PTP		0xB300
 #define PCI_SUBSYS_DEVID_OCTX2_LOKI_PTP		0xB400
 #define PCI_SUBSYS_DEVID_OCTX2_95MM_PTP		0xB500
+#define PCI_SUBSYS_DEVID_CN10K_A_PTP		0xB900
+#define PCI_SUBSYS_DEVID_CNF10K_A_PTP		0xBA00
+#define PCI_SUBSYS_DEVID_CNF10K_B_PTP		0xBC00
 #define PCI_DEVID_OCTEONTX2_RST			0xA085
 
 #define PCI_PTP_BAR_NO				0
@@ -234,6 +237,15 @@ static const struct pci_device_id ptp_id_table[] = {
 	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_OCTEONTX2_PTP,
 			 PCI_VENDOR_ID_CAVIUM,
 			 PCI_SUBSYS_DEVID_OCTX2_95MM_PTP) },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_OCTEONTX2_PTP,
+			 PCI_VENDOR_ID_CAVIUM,
+			 PCI_SUBSYS_DEVID_CN10K_A_PTP) },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_OCTEONTX2_PTP,
+			 PCI_VENDOR_ID_CAVIUM,
+			 PCI_SUBSYS_DEVID_CNF10K_A_PTP) },
+	{ PCI_DEVICE_SUB(PCI_VENDOR_ID_CAVIUM, PCI_DEVID_OCTEONTX2_PTP,
+			 PCI_VENDOR_ID_CAVIUM,
+			 PCI_SUBSYS_DEVID_CNF10K_B_PTP) },
 	{ 0, }
 };
 

drivers/net/ethernet/marvell/octeontx2/af/rpm.c

+// SPDX-License-Identifier: GPL-2.0
+/*  Marvell OcteonTx2 RPM driver
+ *
+ * Copyright (C) 2020 Marvell.
+ *
+ */
+
+#include "cgx.h"
+#include "lmac_common.h"
+
+static struct mac_ops	rpm_mac_ops   = {
+	.name		=       "rpm",
+	.csr_offset     =       0x4e00,
+	.lmac_offset    =       20,
+	.int_register	=       RPMX_CMRX_SW_INT,
+	.int_set_reg    =       RPMX_CMRX_SW_INT_ENA_W1S,
+	.irq_offset     =       1,
+	.int_ena_bit    =       BIT_ULL(0),
+	.lmac_fwi	=	RPM_LMAC_FWI,
+	.non_contiguous_serdes_lane = true,
+	.rx_stats_cnt   =       43,
+	.tx_stats_cnt   =       34,
+	.get_nr_lmacs	=	rpm_get_nr_lmacs,
+	.get_lmac_type  =       rpm_get_lmac_type,
+	.mac_lmac_intl_lbk =    rpm_lmac_internal_loopback,
+	.mac_get_rx_stats  =	rpm_get_rx_stats,
+	.mac_get_tx_stats  =	rpm_get_tx_stats,
+	.mac_enadis_rx_pause_fwding =	rpm_lmac_enadis_rx_pause_fwding,
+	.mac_get_pause_frm_status =	rpm_lmac_get_pause_frm_status,
+	.mac_enadis_pause_frm =		rpm_lmac_enadis_pause_frm,
+	.mac_pause_frm_config =		rpm_lmac_pause_frm_config,
+};
+
+struct mac_ops *rpm_get_mac_ops(void)
+{
+	return &rpm_mac_ops;
+}
+
+static void rpm_write(rpm_t *rpm, u64 lmac, u64 offset, u64 val)
+{
+	cgx_write(rpm, lmac, offset, val);
+}
+
+static u64 rpm_read(rpm_t *rpm, u64 lmac, u64 offset)
+{
+	return	cgx_read(rpm, lmac, offset);
+}
+
+int rpm_get_nr_lmacs(void *rpmd)
+{
+	rpm_t *rpm = rpmd;
+
+	return hweight8(rpm_read(rpm, 0, CGXX_CMRX_RX_LMACS) & 0xFULL);
+}
+
+void rpm_lmac_enadis_rx_pause_fwding(void *rpmd, int lmac_id, bool enable)
+{
+	rpm_t *rpm = rpmd;
+	u64 cfg;
+
+	if (!rpm)
+		return;
+
+	if (enable) {
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+	} else {
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg |= RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+	}
+}
+
+int rpm_lmac_get_pause_frm_status(void *rpmd, int lmac_id,
+				  u8 *tx_pause, u8 *rx_pause)
+{
+	rpm_t *rpm = rpmd;
+	u64 cfg;
+
+	if (!is_lmac_valid(rpm, lmac_id))
+		return -ENODEV;
+
+	cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+	*rx_pause = !(cfg & RPMX_MTI_MAC100X_COMMAND_CONFIG_RX_P_DISABLE);
+
+	cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+	*tx_pause = !(cfg & RPMX_MTI_MAC100X_COMMAND_CONFIG_TX_P_DISABLE);
+	return 0;
+}
+
+int rpm_lmac_enadis_pause_frm(void *rpmd, int lmac_id, u8 tx_pause,
+			      u8 rx_pause)
+{
+	rpm_t *rpm = rpmd;
+	u64 cfg;
+
+	if (!is_lmac_valid(rpm, lmac_id))
+		return -ENODEV;
+
+	cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+	cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_RX_P_DISABLE;
+	cfg |= rx_pause ? 0x0 : RPMX_MTI_MAC100X_COMMAND_CONFIG_RX_P_DISABLE;
+	cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE;
+	cfg |= rx_pause ? 0x0 : RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE;
+	rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+	cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+	cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_TX_P_DISABLE;
+	cfg |= tx_pause ? 0x0 : RPMX_MTI_MAC100X_COMMAND_CONFIG_TX_P_DISABLE;
+	rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+	cfg = rpm_read(rpm, 0, RPMX_CMR_RX_OVR_BP);
+	if (tx_pause) {
+		cfg &= ~RPMX_CMR_RX_OVR_BP_EN(lmac_id);
+	} else {
+		cfg |= RPMX_CMR_RX_OVR_BP_EN(lmac_id);
+		cfg &= ~RPMX_CMR_RX_OVR_BP_BP(lmac_id);
+	}
+	rpm_write(rpm, 0, RPMX_CMR_RX_OVR_BP, cfg);
+	return 0;
+}
+
+void rpm_lmac_pause_frm_config(void *rpmd, int lmac_id, bool enable)
+{
+	rpm_t *rpm = rpmd;
+	u64 cfg;
+
+	if (enable) {
+		/* Enable 802.3 pause frame mode */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_PFC_MODE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+		/* Enable receive pause frames */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_RX_P_DISABLE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+		/* Enable forward pause to TX block */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+		/* Enable pause frames transmission */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg &= ~RPMX_MTI_MAC100X_COMMAND_CONFIG_TX_P_DISABLE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+		/* Set pause time and interval */
+		cfg = rpm_read(rpm, lmac_id,
+			       RPMX_MTI_MAC100X_CL01_PAUSE_QUANTA);
+		cfg &= ~0xFFFFULL;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_CL01_PAUSE_QUANTA,
+			  cfg | RPM_DEFAULT_PAUSE_TIME);
+		/* Set pause interval as the hardware default is too short */
+		cfg = rpm_read(rpm, lmac_id,
+			       RPMX_MTI_MAC100X_CL01_QUANTA_THRESH);
+		cfg &= ~0xFFFFULL;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_CL01_QUANTA_THRESH,
+			  cfg | (RPM_DEFAULT_PAUSE_TIME / 2));
+
+	} else {
+		/* ALL pause frames received are completely ignored */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg |= RPMX_MTI_MAC100X_COMMAND_CONFIG_RX_P_DISABLE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+		/* Disable forward pause to TX block */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg |= RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+
+		/* Disable pause frames transmission */
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG);
+		cfg |= RPMX_MTI_MAC100X_COMMAND_CONFIG_TX_P_DISABLE;
+		rpm_write(rpm, lmac_id, RPMX_MTI_MAC100X_COMMAND_CONFIG, cfg);
+	}
+}
+
+int rpm_get_rx_stats(void *rpmd, int lmac_id, int idx, u64 *rx_stat)
+{
+	rpm_t *rpm = rpmd;
+	u64 val_lo, val_hi;
+
+	if (!rpm || lmac_id >= rpm->lmac_count)
+		return -ENODEV;
+
+	mutex_lock(&rpm->lock);
+
+	/* Update idx to point per lmac Rx statistics page */
+	idx += lmac_id * rpm->mac_ops->rx_stats_cnt;
+
+	/* Read lower 32 bits of counter */
+	val_lo = rpm_read(rpm, 0, RPMX_MTI_STAT_RX_STAT_PAGES_COUNTERX +
+			  (idx * 8));
+
+	/* upon read of lower 32 bits, higher 32 bits are written
+	 * to RPMX_MTI_STAT_DATA_HI_CDC
+	 */
+	val_hi = rpm_read(rpm, 0, RPMX_MTI_STAT_DATA_HI_CDC);
+
+	*rx_stat = (val_hi << 32 | val_lo);
+
+	mutex_unlock(&rpm->lock);
+	return 0;
+}
+
+int rpm_get_tx_stats(void *rpmd, int lmac_id, int idx, u64 *tx_stat)
+{
+	rpm_t *rpm = rpmd;
+	u64 val_lo, val_hi;
+
+	if (!rpm || lmac_id >= rpm->lmac_count)
+		return -ENODEV;
+
+	mutex_lock(&rpm->lock);
+
+	/* Update idx to point per lmac Tx statistics page */
+	idx += lmac_id * rpm->mac_ops->tx_stats_cnt;
+
+	val_lo = rpm_read(rpm, 0, RPMX_MTI_STAT_TX_STAT_PAGES_COUNTERX +
+			    (idx * 8));
+	val_hi = rpm_read(rpm, 0, RPMX_MTI_STAT_DATA_HI_CDC);
+
+	*tx_stat = (val_hi << 32 | val_lo);
+
+	mutex_unlock(&rpm->lock);
+	return 0;
+}
+
+u8 rpm_get_lmac_type(void *rpmd, int lmac_id)
+{
+	rpm_t *rpm = rpmd;
+	u64 req = 0, resp;
+	int err;
+
+	req = FIELD_SET(CMDREG_ID, CGX_CMD_GET_LINK_STS, req);
+	err = cgx_fwi_cmd_generic(req, &resp, rpm, 0);
+	if (!err)
+		return FIELD_GET(RESP_LINKSTAT_LMAC_TYPE, resp);
+	return err;
+}
+
+int rpm_lmac_internal_loopback(void *rpmd, int lmac_id, bool enable)
+{
+	rpm_t *rpm = rpmd;
+	u8 lmac_type;
+	u64 cfg;
+
+	if (!rpm || lmac_id >= rpm->lmac_count)
+		return -ENODEV;
+	lmac_type = rpm->mac_ops->get_lmac_type(rpm, lmac_id);
+	if (lmac_type == LMAC_MODE_100G_R) {
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_PCS100X_CONTROL1);
+
+		if (enable)
+			cfg |= RPMX_MTI_PCS_LBK;
+		else
+			cfg &= ~RPMX_MTI_PCS_LBK;
+		rpm_write(rpm, lmac_id, RPMX_MTI_PCS100X_CONTROL1, cfg);
+	} else {
+		cfg = rpm_read(rpm, lmac_id, RPMX_MTI_LPCSX_CONTROL1);
+		if (enable)
+			cfg |= RPMX_MTI_PCS_LBK;
+		else
+			cfg &= ~RPMX_MTI_PCS_LBK;
+		rpm_write(rpm, lmac_id, RPMX_MTI_LPCSX_CONTROL1, cfg);
+	}
+
+	return 0;
+}

drivers/net/ethernet/marvell/octeontx2/af/rpm.h

+/* SPDX-License-Identifier: GPL-2.0 */
+/*  Marvell OcteonTx2 RPM driver
+ *
+ * Copyright (C) 2020 Marvell.
+ *
+ */
+
+#ifndef RPM_H
+#define RPM_H
+
+#include <linux/bits.h>
+
+/* PCI device IDs */
+#define PCI_DEVID_CN10K_RPM		0xA060
+
+/* Registers */
+#define RPMX_CMRX_SW_INT                0x180
+#define RPMX_CMRX_SW_INT_W1S            0x188
+#define RPMX_CMRX_SW_INT_ENA_W1S        0x198
+#define RPMX_CMRX_LINK_CFG		0x1070
+#define RPMX_MTI_PCS100X_CONTROL1       0x20000
+#define RPMX_MTI_LPCSX_CONTROL1         0x30000
+#define RPMX_MTI_PCS_LBK                BIT_ULL(14)
+#define RPMX_MTI_LPCSX_CONTROL(id)     (0x30000 | ((id) * 0x100))
+
+#define RPMX_CMRX_LINK_RANGE_MASK	GENMASK_ULL(19, 16)
+#define RPMX_CMRX_LINK_BASE_MASK	GENMASK_ULL(11, 0)
+#define RPMX_MTI_MAC100X_COMMAND_CONFIG	0x8010
+#define RPMX_MTI_MAC100X_COMMAND_CONFIG_RX_P_DISABLE	BIT_ULL(29)
+#define RPMX_MTI_MAC100X_COMMAND_CONFIG_TX_P_DISABLE	BIT_ULL(28)
+#define RPMX_MTI_MAC100X_COMMAND_CONFIG_PAUSE_IGNORE	BIT_ULL(8)
+#define RPMX_MTI_MAC100X_COMMAND_CONFIG_PFC_MODE	BIT_ULL(19)
+#define RPMX_MTI_MAC100X_CL01_PAUSE_QUANTA		0x80A8
+#define RPMX_MTI_MAC100X_CL01_QUANTA_THRESH		0x80C8
+#define RPM_DEFAULT_PAUSE_TIME			0xFFFF
+#define RPMX_CMR_RX_OVR_BP		0x4120
+#define RPMX_CMR_RX_OVR_BP_EN(x)	BIT_ULL((x) + 8)
+#define RPMX_CMR_RX_OVR_BP_BP(x)	BIT_ULL((x) + 4)
+#define RPMX_MTI_STAT_RX_STAT_PAGES_COUNTERX 0x12000
+#define RPMX_MTI_STAT_TX_STAT_PAGES_COUNTERX 0x13000
+#define RPMX_MTI_STAT_DATA_HI_CDC            0x10038
+
+#define RPM_LMAC_FWI			0xa
+
+/* Function Declarations */
+int rpm_get_nr_lmacs(void *rpmd);
+u8 rpm_get_lmac_type(void *rpmd, int lmac_id);
+int rpm_lmac_internal_loopback(void *rpmd, int lmac_id, bool enable);
+void rpm_lmac_enadis_rx_pause_fwding(void *rpmd, int lmac_id, bool enable);
+int rpm_lmac_get_pause_frm_status(void *cgxd, int lmac_id, u8 *tx_pause,
+				  u8 *rx_pause);
+void rpm_lmac_pause_frm_config(void *rpmd, int lmac_id, bool enable);
+int rpm_lmac_enadis_pause_frm(void *rpmd, int lmac_id, u8 tx_pause,
+			      u8 rx_pause);
+int rpm_get_tx_stats(void *rpmd, int lmac_id, int idx, u64 *tx_stat);
+int rpm_get_rx_stats(void *rpmd, int lmac_id, int idx, u64 *rx_stat);
+#endif /* RPM_H */

drivers/net/ethernet/marvell/octeontx2/af/rvu.c

@@ -22,7 +22,7 @@
 
 #include "rvu_trace.h"
 
-#define DRV_NAME	"octeontx2-af"
+#define DRV_NAME	"rvu_af"
 #define DRV_STRING      "Marvell OcteonTX2 RVU Admin Function Driver"
 
 static int rvu_get_hwvf(struct rvu *rvu, int pcifunc);
@@ -78,6 +78,9 @@ static void rvu_setup_hw_capabilities(struct rvu *rvu)
 		if (is_rvu_96xx_A0(rvu))
 			hw->cap.nix_rx_multicast = false;
 	}
+
+	if (!is_rvu_otx2(rvu))
+		hw->cap.per_pf_mbox_regs = true;
 }
 
 /* Poll a RVU block's register 'offset', for a 'zero'
@@ -852,6 +855,31 @@ static int rvu_setup_cpt_hw_resource(struct rvu *rvu, int blkaddr)
 	return rvu_alloc_bitmap(&block->lf);
 }
 
+static void rvu_get_lbk_bufsize(struct rvu *rvu)
+{
+	struct pci_dev *pdev = NULL;
+	void __iomem *base;
+	u64 lbk_const;
+
+	pdev = pci_get_device(PCI_VENDOR_ID_CAVIUM,
+			      PCI_DEVID_OCTEONTX2_LBK, pdev);
+	if (!pdev)
+		return;
+
+	base = pci_ioremap_bar(pdev, 0);
+	if (!base)
+		goto err_put;
+
+	lbk_const = readq(base + LBK_CONST);
+
+	/* cache fifo size */
+	rvu->hw->lbk_bufsize = FIELD_GET(LBK_CONST_BUF_SIZE, lbk_const);
+
+	iounmap(base);
+err_put:
+	pci_dev_put(pdev);
+}
+
 static int rvu_setup_hw_resources(struct rvu *rvu)
 {
 	struct rvu_hwinfo *hw = rvu->hw;
@@ -1003,6 +1031,10 @@ static int rvu_setup_hw_resources(struct rvu *rvu)
 		rvu_scan_block(rvu, block);
 	}
 
+	err = rvu_set_channels_base(rvu);
+	if (err)
+		goto msix_err;
+
 	err = rvu_npc_init(rvu);
 	if (err)
 		goto npc_err;
@@ -1018,10 +1050,14 @@ static int rvu_setup_hw_resources(struct rvu *rvu)
 	if (err)
 		goto npa_err;
 
+	rvu_get_lbk_bufsize(rvu);
+
 	err = rvu_nix_init(rvu);
 	if (err)
 		goto nix_err;
 
+	rvu_program_channels(rvu);
+
 	return 0;
 
 nix_err:
@@ -1936,41 +1972,105 @@ static inline void rvu_afvf_mbox_up_handler(struct work_struct *work)
 	__rvu_mbox_up_handler(mwork, TYPE_AFVF);
 }
 
+static int rvu_get_mbox_regions(struct rvu *rvu, void **mbox_addr,
+				int num, int type)
+{
+	struct rvu_hwinfo *hw = rvu->hw;
+	int region;
+	u64 bar4;
+
+	/* For cn10k platform VF mailbox regions of a PF follows after the
+	 * PF <-> AF mailbox region. Whereas for Octeontx2 it is read from
+	 * RVU_PF_VF_BAR4_ADDR register.
+	 */
+	if (type == TYPE_AFVF) {
+		for (region = 0; region < num; region++) {
+			if (hw->cap.per_pf_mbox_regs) {
+				bar4 = rvu_read64(rvu, BLKADDR_RVUM,
+						  RVU_AF_PFX_BAR4_ADDR(0)) +
+						  MBOX_SIZE;
+				bar4 += region * MBOX_SIZE;
+			} else {
+				bar4 = rvupf_read64(rvu, RVU_PF_VF_BAR4_ADDR);
+				bar4 += region * MBOX_SIZE;
+			}
+			mbox_addr[region] = (void *)ioremap_wc(bar4, MBOX_SIZE);
+			if (!mbox_addr[region])
+				goto error;
+		}
+		return 0;
+	}
+
+	/* For cn10k platform AF <-> PF mailbox region of a PF is read from per
+	 * PF registers. Whereas for Octeontx2 it is read from
+	 * RVU_AF_PF_BAR4_ADDR register.
+	 */
+	for (region = 0; region < num; region++) {
+		if (hw->cap.per_pf_mbox_regs) {
+			bar4 = rvu_read64(rvu, BLKADDR_RVUM,
+					  RVU_AF_PFX_BAR4_ADDR(region));
+		} else {
+			bar4 = rvu_read64(rvu, BLKADDR_RVUM,
+					  RVU_AF_PF_BAR4_ADDR);
+			bar4 += region * MBOX_SIZE;
+		}
+		mbox_addr[region] = (void *)ioremap_wc(bar4, MBOX_SIZE);
+		if (!mbox_addr[region])
+			goto error;
+	}
+	return 0;
+
+error:
+	while (region--)
+		iounmap((void __iomem *)mbox_addr[region]);
+	return -ENOMEM;
+}
+
 static int rvu_mbox_init(struct rvu *rvu, struct mbox_wq_info *mw,
 			 int type, int num,
 			 void (mbox_handler)(struct work_struct *),
 			 void (mbox_up_handler)(struct work_struct *))
 {
-	void __iomem *hwbase = NULL, *reg_base;
-	int err, i, dir, dir_up;
+	int err = -EINVAL, i, dir, dir_up;
+	void __iomem *reg_base;
 	struct rvu_work *mwork;
+	void **mbox_regions;
 	const char *name;
-	u64 bar4_addr;
+
+	mbox_regions = kcalloc(num, sizeof(void *), GFP_KERNEL);
+	if (!mbox_regions)
+		return -ENOMEM;
 
 	switch (type) {
 	case TYPE_AFPF:
 		name = "rvu_afpf_mailbox";
-		bar4_addr = rvu_read64(rvu, BLKADDR_RVUM, RVU_AF_PF_BAR4_ADDR);
 		dir = MBOX_DIR_AFPF;
 		dir_up = MBOX_DIR_AFPF_UP;
 		reg_base = rvu->afreg_base;
+		err = rvu_get_mbox_regions(rvu, mbox_regions, num, TYPE_AFPF);
+		if (err)
+			goto free_regions;
 		break;
 	case TYPE_AFVF:
 		name = "rvu_afvf_mailbox";
-		bar4_addr = rvupf_read64(rvu, RVU_PF_VF_BAR4_ADDR);
 		dir = MBOX_DIR_PFVF;
 		dir_up = MBOX_DIR_PFVF_UP;
 		reg_base = rvu->pfreg_base;
+		err = rvu_get_mbox_regions(rvu, mbox_regions, num, TYPE_AFVF);
+		if (err)
+			goto free_regions;
 		break;
 	default:
-		return -EINVAL;
+		return err;
 	}
 
 	mw->mbox_wq = alloc_workqueue(name,
 				      WQ_UNBOUND | WQ_HIGHPRI | WQ_MEM_RECLAIM,
 				      num);
-	if (!mw->mbox_wq)
-		return -ENOMEM;
+	if (!mw->mbox_wq) {
+		err = -ENOMEM;
+		goto unmap_regions;
+	}
 
 	mw->mbox_wrk = devm_kcalloc(rvu->dev, num,
 				    sizeof(struct rvu_work), GFP_KERNEL);
@@ -1986,23 +2086,13 @@ static int rvu_mbox_init(struct rvu *rvu, struct mbox_wq_info *mw,
 		goto exit;
 	}
 
-	/* Mailbox is a reserved memory (in RAM) region shared between
-	 * RVU devices, shouldn't be mapped as device memory to allow
-	 * unaligned accesses.
-	 */
-	hwbase = ioremap_wc(bar4_addr, MBOX_SIZE * num);
-	if (!hwbase) {
-		dev_err(rvu->dev, "Unable to map mailbox region\n");
-		err = -ENOMEM;
-		goto exit;
-	}
-
-	err = otx2_mbox_init(&mw->mbox, hwbase, rvu->pdev, reg_base, dir, num);
+	err = otx2_mbox_regions_init(&mw->mbox, mbox_regions, rvu->pdev,
+				     reg_base, dir, num);
 	if (err)
 		goto exit;
 
-	err = otx2_mbox_init(&mw->mbox_up, hwbase, rvu->pdev,
-			     reg_base, dir_up, num);
+	err = otx2_mbox_regions_init(&mw->mbox_up, mbox_regions, rvu->pdev,
+				     reg_base, dir_up, num);
 	if (err)
 		goto exit;
 
@@ -2015,25 +2105,36 @@ static int rvu_mbox_init(struct rvu *rvu, struct mbox_wq_info *mw,
 		mwork->rvu = rvu;
 		INIT_WORK(&mwork->work, mbox_up_handler);
 	}
-
+	kfree(mbox_regions);
 	return 0;
+
 exit:
-	if (hwbase)
-		iounmap((void __iomem *)hwbase);
 	destroy_workqueue(mw->mbox_wq);
+unmap_regions:
+	while (num--)
+		iounmap((void __iomem *)mbox_regions[num]);
+free_regions:
+	kfree(mbox_regions);
 	return err;
 }
 
 static void rvu_mbox_destroy(struct mbox_wq_info *mw)
 {
+	struct otx2_mbox *mbox = &mw->mbox;
+	struct otx2_mbox_dev *mdev;
+	int devid;
+
 	if (mw->mbox_wq) {
 		flush_workqueue(mw->mbox_wq);
 		destroy_workqueue(mw->mbox_wq);
 		mw->mbox_wq = NULL;
 	}
 
-	if (mw->mbox.hwbase)
-		iounmap((void __iomem *)mw->mbox.hwbase);
+	for (devid = 0; devid < mbox->ndevs; devid++) {
+		mdev = &mbox->dev[devid];
+		if (mdev->hwbase)
+			iounmap((void __iomem *)mdev->hwbase);
+	}
 
 	otx2_mbox_destroy(&mw->mbox);
 	otx2_mbox_destroy(&mw->mbox_up);
@@ -2653,8 +2754,6 @@ static void rvu_enable_afvf_intr(struct rvu *rvu)
 	rvupf_write64(rvu, RVU_PF_VFME_INT_ENA_W1SX(1), INTR_MASK(vfs - 64));
 }
 
-#define PCI_DEVID_OCTEONTX2_LBK 0xA061
-
 int rvu_get_num_lbk_chans(void)
 {
 	struct pci_dev *pdev;

drivers/net/ethernet/marvell/octeontx2/af/rvu.h

@@ -19,12 +19,15 @@
 #include "common.h"
 #include "mbox.h"
 #include "npc.h"
+#include "rvu_reg.h"
 
 /* PCI device IDs */
 #define	PCI_DEVID_OCTEONTX2_RVU_AF		0xA065
+#define	PCI_DEVID_OCTEONTX2_LBK			0xA061
 
 /* Subsystem Device ID */
 #define PCI_SUBSYS_DEVID_96XX                  0xB200
+#define PCI_SUBSYS_DEVID_CN10K_A	       0xB900
 
 /* PCI BAR nos */
 #define	PCI_AF_REG_BAR_NUM			0
@@ -303,6 +306,8 @@ struct hw_cap {
 	bool	nix_shaping;		 /* Is shaping and coloring supported */
 	bool	nix_tx_link_bp;		 /* Can link backpressure TL queues ? */
 	bool	nix_rx_multicast;	 /* Rx packet replication support */
+	bool	per_pf_mbox_regs; /* PF mbox specified in per PF registers ? */
+	bool	programmable_chans; /* Channels programmable ? */
 };
 
 struct rvu_hwinfo {
@@ -311,14 +316,20 @@ struct rvu_hwinfo {
 	u16	max_vfs_per_pf; /* Max VFs that can be attached to a PF */
 	u8	cgx;
 	u8	lmac_per_cgx;
+	u16	cgx_chan_base;	/* CGX base channel number */
+	u16	lbk_chan_base;	/* LBK base channel number */
+	u16	sdp_chan_base;	/* SDP base channel number */
+	u16	cpt_chan_base;	/* CPT base channel number */
 	u8	cgx_links;
 	u8	lbk_links;
 	u8	sdp_links;
+	u8	cpt_links;	/* Number of CPT links */
 	u8	npc_kpus;          /* No of parser units */
 	u8	npc_pkinds;        /* No of port kinds */
 	u8	npc_intfs;         /* No of interfaces */
 	u8	npc_kpu_entries;   /* No of KPU entries */
 	u16	npc_counters;	   /* No of match stats counters */
+	u32	lbk_bufsize;	   /* FIFO size supported by LBK */
 	bool	npc_ext_set;	   /* Extended register set */
 
 	struct hw_cap    cap;
@@ -476,6 +487,59 @@ static inline bool is_rvu_96xx_B0(struct rvu *rvu)
 		(pdev->subsystem_device == PCI_SUBSYS_DEVID_96XX);
 }
 
+/* REVID for PCIe devices.
+ * Bits 0..1: minor pass, bit 3..2: major pass
+ * bits 7..4: midr id
+ */
+#define PCI_REVISION_ID_96XX		0x00
+#define PCI_REVISION_ID_95XX		0x10
+#define PCI_REVISION_ID_LOKI		0x20
+#define PCI_REVISION_ID_98XX		0x30
+#define PCI_REVISION_ID_95XXMM		0x40
+
+static inline bool is_rvu_otx2(struct rvu *rvu)
+{
+	struct pci_dev *pdev = rvu->pdev;
+
+	u8 midr = pdev->revision & 0xF0;
+
+	return (midr == PCI_REVISION_ID_96XX || midr == PCI_REVISION_ID_95XX ||
+		midr == PCI_REVISION_ID_LOKI || midr == PCI_REVISION_ID_98XX ||
+		midr == PCI_REVISION_ID_95XXMM);
+}
+
+static inline u16 rvu_nix_chan_cgx(struct rvu *rvu, u8 cgxid,
+				   u8 lmacid, u8 chan)
+{
+	u64 nix_const = rvu_read64(rvu, BLKADDR_NIX0, NIX_AF_CONST);
+	u16 cgx_chans = nix_const & 0xFFULL;
+	struct rvu_hwinfo *hw = rvu->hw;
+
+	if (!hw->cap.programmable_chans)
+		return NIX_CHAN_CGX_LMAC_CHX(cgxid, lmacid, chan);
+
+	return rvu->hw->cgx_chan_base +
+		(cgxid * hw->lmac_per_cgx + lmacid) * cgx_chans + chan;
+}
+
+static inline u16 rvu_nix_chan_lbk(struct rvu *rvu, u8 lbkid,
+				   u8 chan)
+{
+	u64 nix_const = rvu_read64(rvu, BLKADDR_NIX0, NIX_AF_CONST);
+	u16 lbk_chans = (nix_const >> 16) & 0xFFULL;
+	struct rvu_hwinfo *hw = rvu->hw;
+
+	if (!hw->cap.programmable_chans)
+		return NIX_CHAN_LBK_CHX(lbkid, chan);
+
+	return rvu->hw->lbk_chan_base + lbkid * lbk_chans + chan;
+}
+
+static inline u16 rvu_nix_chan_cpt(struct rvu *rvu, u8 chan)
+{
+	return rvu->hw->cpt_chan_base + chan;
+}
+
 /* Function Prototypes
  * RVU
  */
@@ -612,9 +676,16 @@ void npc_enable_mcam_entry(struct rvu *rvu, struct npc_mcam *mcam,
 void npc_read_mcam_entry(struct rvu *rvu, struct npc_mcam *mcam,
 			 int blkaddr, u16 src, struct mcam_entry *entry,
 			 u8 *intf, u8 *ena);
+bool is_mac_feature_supported(struct rvu *rvu, int pf, int feature);
+u32  rvu_cgx_get_fifolen(struct rvu *rvu);
+
 /* CPT APIs */
 int rvu_cpt_lf_teardown(struct rvu *rvu, u16 pcifunc, int lf, int slot);
 
+/* CN10K RVU */
+int rvu_set_channels_base(struct rvu *rvu);
+void rvu_program_channels(struct rvu *rvu);
+
 #ifdef CONFIG_DEBUG_FS
 void rvu_dbg_init(struct rvu *rvu);
 void rvu_dbg_exit(struct rvu *rvu);

drivers/net/ethernet/marvell/octeontx2/af/rvu_cgx.c

@@ -14,6 +14,7 @@
 
 #include "rvu.h"
 #include "cgx.h"
+#include "lmac_common.h"
 #include "rvu_reg.h"
 #include "rvu_trace.h"
 
@@ -42,6 +43,20 @@ static struct _req_type __maybe_unused					\
 MBOX_UP_CGX_MESSAGES
 #undef M
 
+bool is_mac_feature_supported(struct rvu *rvu, int pf, int feature)
+{
+	u8 cgx_id, lmac_id;
+	void *cgxd;
+
+	if (!is_pf_cgxmapped(rvu, pf))
+		return 0;
+
+	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_id, &lmac_id);
+	cgxd = rvu_cgx_pdata(cgx_id, rvu);
+
+	return  (cgx_features_get(cgxd) & feature);
+}
+
 /* Returns bitmap of mapped PFs */
 static u16 cgxlmac_to_pfmap(struct rvu *rvu, u8 cgx_id, u8 lmac_id)
 {
@@ -92,9 +107,10 @@ static int rvu_map_cgx_lmac_pf(struct rvu *rvu)
 {
 	struct npc_pkind *pkind = &rvu->hw->pkind;
 	int cgx_cnt_max = rvu->cgx_cnt_max;
-	int cgx, lmac_cnt, lmac;
 	int pf = PF_CGXMAP_BASE;
+	unsigned long lmac_bmap;
 	int size, free_pkind;
+	int cgx, lmac, iter;
 
 	if (!cgx_cnt_max)
 		return 0;
@@ -125,14 +141,17 @@ static int rvu_map_cgx_lmac_pf(struct rvu *rvu)
 	for (cgx = 0; cgx < cgx_cnt_max; cgx++) {
 		if (!rvu_cgx_pdata(cgx, rvu))
 			continue;
-		lmac_cnt = cgx_get_lmac_cnt(rvu_cgx_pdata(cgx, rvu));
-		for (lmac = 0; lmac < lmac_cnt; lmac++, pf++) {
+		lmac_bmap = cgx_get_lmac_bmap(rvu_cgx_pdata(cgx, rvu));
+		for_each_set_bit(iter, &lmac_bmap, MAX_LMAC_PER_CGX) {
+			lmac = cgx_get_lmacid(rvu_cgx_pdata(cgx, rvu),
+					      iter);
 			rvu->pf2cgxlmac_map[pf] = cgxlmac_id_to_bmap(cgx, lmac);
 			rvu->cgxlmac2pf_map[CGX_OFFSET(cgx) + lmac] = 1 << pf;
 			free_pkind = rvu_alloc_rsrc(&pkind->rsrc);
 			pkind->pfchan_map[free_pkind] = ((pf) & 0x3F) << 16;
 			rvu_map_cgx_nix_block(rvu, pf, cgx, lmac);
 			rvu->cgx_mapped_pfs++;
+			pf++;
 		}
 	}
 	return 0;
@@ -154,8 +173,10 @@ static int rvu_cgx_send_link_info(int cgx_id, int lmac_id, struct rvu *rvu)
 				&qentry->link_event.link_uinfo);
 	qentry->link_event.cgx_id = cgx_id;
 	qentry->link_event.lmac_id = lmac_id;
-	if (err)
+	if (err) {
+		kfree(qentry);
 		goto skip_add;
+	}
 	list_add_tail(&qentry->evq_node, &rvu->cgx_evq_head);
 skip_add:
 	spin_unlock_irqrestore(&rvu->cgx_evq_lock, flags);
@@ -251,6 +272,7 @@ static void cgx_evhandler_task(struct work_struct *work)
 
 static int cgx_lmac_event_handler_init(struct rvu *rvu)
 {
+	unsigned long lmac_bmap;
 	struct cgx_event_cb cb;
 	int cgx, lmac, err;
 	void *cgxd;
@@ -271,7 +293,8 @@ static int cgx_lmac_event_handler_init(struct rvu *rvu)
 		cgxd = rvu_cgx_pdata(cgx, rvu);
 		if (!cgxd)
 			continue;
-		for (lmac = 0; lmac < cgx_get_lmac_cnt(cgxd); lmac++) {
+		lmac_bmap = cgx_get_lmac_bmap(cgxd);
+		for_each_set_bit(lmac, &lmac_bmap, MAX_LMAC_PER_CGX) {
 			err = cgx_lmac_evh_register(&cb, cgxd, lmac);
 			if (err)
 				dev_err(rvu->dev,
@@ -349,6 +372,7 @@ int rvu_cgx_init(struct rvu *rvu)
 
 int rvu_cgx_exit(struct rvu *rvu)
 {
+	unsigned long lmac_bmap;
 	int cgx, lmac;
 	void *cgxd;
 
@@ -356,7 +380,8 @@ int rvu_cgx_exit(struct rvu *rvu)
 		cgxd = rvu_cgx_pdata(cgx, rvu);
 		if (!cgxd)
 			continue;
-		for (lmac = 0; lmac < cgx_get_lmac_cnt(cgxd); lmac++)
+		lmac_bmap = cgx_get_lmac_bmap(cgxd);
+		for_each_set_bit(lmac, &lmac_bmap, MAX_LMAC_PER_CGX)
 			cgx_lmac_evh_unregister(cgxd, lmac);
 	}
 
@@ -381,6 +406,7 @@ static bool is_cgx_config_permitted(struct rvu *rvu, u16 pcifunc)
 
 void rvu_cgx_enadis_rx_bp(struct rvu *rvu, int pf, bool enable)
 {
+	struct mac_ops *mac_ops;
 	u8 cgx_id, lmac_id;
 	void *cgxd;
 
@@ -390,11 +416,12 @@ void rvu_cgx_enadis_rx_bp(struct rvu *rvu, int pf, bool enable)
 	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_id, &lmac_id);
 	cgxd = rvu_cgx_pdata(cgx_id, rvu);
 
+	mac_ops = get_mac_ops(cgxd);
 	/* Set / clear CTL_BCK to control pause frame forwarding to NIX */
 	if (enable)
-		cgx_lmac_enadis_rx_pause_fwding(cgxd, lmac_id, true);
+		mac_ops->mac_enadis_rx_pause_fwding(cgxd, lmac_id, true);
 	else
-		cgx_lmac_enadis_rx_pause_fwding(cgxd, lmac_id, false);
+		mac_ops->mac_enadis_rx_pause_fwding(cgxd, lmac_id, false);
 }
 
 int rvu_cgx_config_rxtx(struct rvu *rvu, u16 pcifunc, bool start)
@@ -426,10 +453,11 @@ int rvu_mbox_handler_cgx_stop_rxtx(struct rvu *rvu, struct msg_req *req,
 	return 0;
 }
 
-int rvu_mbox_handler_cgx_stats(struct rvu *rvu, struct msg_req *req,
-			       struct cgx_stats_rsp *rsp)
+static int rvu_lmac_get_stats(struct rvu *rvu, struct msg_req *req,
+			      void *rsp)
 {
 	int pf = rvu_get_pf(req->hdr.pcifunc);
+	struct mac_ops *mac_ops;
 	int stat = 0, err = 0;
 	u64 tx_stat, rx_stat;
 	u8 cgx_idx, lmac;
@@ -440,28 +468,47 @@ int rvu_mbox_handler_cgx_stats(struct rvu *rvu, struct msg_req *req,
 
 	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_idx, &lmac);
 	cgxd = rvu_cgx_pdata(cgx_idx, rvu);
+	mac_ops = get_mac_ops(cgxd);
 
 	/* Rx stats */
-	while (stat < CGX_RX_STATS_COUNT) {
-		err = cgx_get_rx_stats(cgxd, lmac, stat, &rx_stat);
+	while (stat < mac_ops->rx_stats_cnt) {
+		err = mac_ops->mac_get_rx_stats(cgxd, lmac, stat, &rx_stat);
 		if (err)
 			return err;
-		rsp->rx_stats[stat] = rx_stat;
+		if (mac_ops->rx_stats_cnt == RPM_RX_STATS_COUNT)
+			((struct rpm_stats_rsp *)rsp)->rx_stats[stat] = rx_stat;
+		else
+			((struct cgx_stats_rsp *)rsp)->rx_stats[stat] = rx_stat;
 		stat++;
 	}
 
 	/* Tx stats */
 	stat = 0;
-	while (stat < CGX_TX_STATS_COUNT) {
-		err = cgx_get_tx_stats(cgxd, lmac, stat, &tx_stat);
+	while (stat < mac_ops->tx_stats_cnt) {
+		err = mac_ops->mac_get_tx_stats(cgxd, lmac, stat, &tx_stat);
 		if (err)
 			return err;
-		rsp->tx_stats[stat] = tx_stat;
+		if (mac_ops->tx_stats_cnt == RPM_TX_STATS_COUNT)
+			((struct rpm_stats_rsp *)rsp)->tx_stats[stat] = tx_stat;
+		else
+			((struct cgx_stats_rsp *)rsp)->tx_stats[stat] = tx_stat;
 		stat++;
 	}
 	return 0;
 }
 
+int rvu_mbox_handler_cgx_stats(struct rvu *rvu, struct msg_req *req,
+			       struct cgx_stats_rsp *rsp)
+{
+	return rvu_lmac_get_stats(rvu, req, (void *)rsp);
+}
+
+int rvu_mbox_handler_rpm_stats(struct rvu *rvu, struct msg_req *req,
+			       struct rpm_stats_rsp *rsp)
+{
+	return rvu_lmac_get_stats(rvu, req, (void *)rsp);
+}
+
 int rvu_mbox_handler_cgx_fec_stats(struct rvu *rvu,
 				   struct msg_req *req,
 				   struct cgx_fec_stats_rsp *rsp)
@@ -554,6 +601,9 @@ static int rvu_cgx_ptp_rx_cfg(struct rvu *rvu, u16 pcifunc, bool enable)
 	u8 cgx_id, lmac_id;
 	void *cgxd;
 
+	if (!is_mac_feature_supported(rvu, pf, RVU_LMAC_FEAT_PTP))
+		return 0;
+
 	/* This msg is expected only from PFs that are mapped to CGX LMACs,
 	 * if received from other PF/VF simply ACK, nothing to do.
 	 */
@@ -640,17 +690,47 @@ int rvu_mbox_handler_cgx_get_linkinfo(struct rvu *rvu, struct msg_req *req,
 	return err;
 }
 
+int rvu_mbox_handler_cgx_features_get(struct rvu *rvu,
+				      struct msg_req *req,
+				      struct cgx_features_info_msg *rsp)
+{
+	int pf = rvu_get_pf(req->hdr.pcifunc);
+	u8 cgx_idx, lmac;
+	void *cgxd;
+
+	if (!is_pf_cgxmapped(rvu, pf))
+		return 0;
+
+	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_idx, &lmac);
+	cgxd = rvu_cgx_pdata(cgx_idx, rvu);
+	rsp->lmac_features = cgx_features_get(cgxd);
+
+	return 0;
+}
+
+u32 rvu_cgx_get_fifolen(struct rvu *rvu)
+{
+	struct mac_ops *mac_ops;
+	int rvu_def_cgx_id = 0;
+	u32 fifo_len;
+
+	mac_ops = get_mac_ops(rvu_cgx_pdata(rvu_def_cgx_id, rvu));
+	fifo_len = mac_ops ? mac_ops->fifo_len : 0;
+
+	return fifo_len;
+}
+
 static int rvu_cgx_config_intlbk(struct rvu *rvu, u16 pcifunc, bool en)
 {
-	int pf = rvu_get_pf(pcifunc);
+	struct mac_ops *mac_ops;
 	u8 cgx_id, lmac_id;
 
 	if (!is_cgx_config_permitted(rvu, pcifunc))
 		return -EPERM;
 
-	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_id, &lmac_id);
+	mac_ops = get_mac_ops(rvu_cgx_pdata(cgx_id, rvu));
 
-	return cgx_lmac_internal_loopback(rvu_cgx_pdata(cgx_id, rvu),
+	return mac_ops->mac_lmac_intl_lbk(rvu_cgx_pdata(cgx_id, rvu),
 					  lmac_id, en);
 }
 
@@ -673,7 +753,12 @@ int rvu_mbox_handler_cgx_cfg_pause_frm(struct rvu *rvu,
 				       struct cgx_pause_frm_cfg *rsp)
 {
 	int pf = rvu_get_pf(req->hdr.pcifunc);
+	struct mac_ops *mac_ops;
 	u8 cgx_id, lmac_id;
+	void *cgxd;
+
+	if (!is_mac_feature_supported(rvu, pf, RVU_LMAC_FEAT_FC))
+		return 0;
 
 	/* This msg is expected only from PF/VFs that are mapped to CGX LMACs,
 	 * if received from other PF/VF simply ACK, nothing to do.
@@ -682,13 +767,16 @@ int rvu_mbox_handler_cgx_cfg_pause_frm(struct rvu *rvu,
 		return -ENODEV;
 
 	rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_id, &lmac_id);
+	cgxd = rvu_cgx_pdata(cgx_id, rvu);
+	mac_ops = get_mac_ops(cgxd);
 
 	if (req->set)
-		cgx_lmac_set_pause_frm(rvu_cgx_pdata(cgx_id, rvu), lmac_id,
-				       req->tx_pause, req->rx_pause);
+		mac_ops->mac_enadis_pause_frm(cgxd, lmac_id,
+					      req->tx_pause, req->rx_pause);
 	else
-		cgx_lmac_get_pause_frm(rvu_cgx_pdata(cgx_id, rvu), lmac_id,
-				       &rsp->tx_pause, &rsp->rx_pause);
+		mac_ops->mac_get_pause_frm_status(cgxd, lmac_id,
+						  &rsp->tx_pause,
+						  &rsp->rx_pause);
 	return 0;
 }
 

drivers/net/ethernet/marvell/octeontx2/af/rvu_cn10k.c

+// SPDX-License-Identifier: GPL-2.0
+/*  Marvell RPM CN10K driver
+ *
+ * Copyright (C) 2020 Marvell.
+ */
+
+#include <linux/bitfield.h>
+#include <linux/pci.h>
+#include "rvu.h"
+#include "cgx.h"
+#include "rvu_reg.h"
+
+int rvu_set_channels_base(struct rvu *rvu)
+{
+	struct rvu_hwinfo *hw = rvu->hw;
+	u16 cpt_chan_base;
+	u64 nix_const;
+	int blkaddr;
+
+	blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_NIX, 0);
+	if (blkaddr < 0)
+		return blkaddr;
+
+	nix_const = rvu_read64(rvu, blkaddr, NIX_AF_CONST);
+
+	hw->cgx = (nix_const >> 12) & 0xFULL;
+	hw->lmac_per_cgx = (nix_const >> 8) & 0xFULL;
+	hw->cgx_links = hw->cgx * hw->lmac_per_cgx;
+	hw->lbk_links = (nix_const >> 24) & 0xFULL;
+	hw->cpt_links = (nix_const >> 44) & 0xFULL;
+	hw->sdp_links = 1;
+
+	hw->cgx_chan_base = NIX_CHAN_CGX_LMAC_CHX(0, 0, 0);
+	hw->lbk_chan_base = NIX_CHAN_LBK_CHX(0, 0);
+	hw->sdp_chan_base = NIX_CHAN_SDP_CH_START;
+
+	/* No Programmable channels */
+	if (!(nix_const & BIT_ULL(60)))
+		return 0;
+
+	hw->cap.programmable_chans = true;
+
+	/* If programmable channels are present then configure
+	 * channels such that all channel numbers are contiguous
+	 * leaving no holes. This way the new CPT channels can be
+	 * accomodated. The order of channel numbers assigned is
+	 * LBK, SDP, CGX and CPT.
+	 */
+	hw->sdp_chan_base = hw->lbk_chan_base + hw->lbk_links *
+				((nix_const >> 16) & 0xFFULL);
+	hw->cgx_chan_base = hw->sdp_chan_base + hw->sdp_links * SDP_CHANNELS;
+
+	cpt_chan_base = hw->cgx_chan_base + hw->cgx_links *
+				(nix_const & 0xFFULL);
+
+	/* Out of 4096 channels start CPT from 2048 so
+	 * that MSB for CPT channels is always set
+	 */
+	if (cpt_chan_base <= 0x800) {
+		hw->cpt_chan_base = 0x800;
+	} else {
+		dev_err(rvu->dev,
+			"CPT channels could not fit in the range 2048-4095\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+#define LBK_CONNECT_NIXX(a)		(0x0 + (a))
+
+static void __rvu_lbk_set_chans(struct rvu *rvu, void __iomem *base,
+				u64 offset, int lbkid, u16 chans)
+{
+	struct rvu_hwinfo *hw = rvu->hw;
+	u64 cfg;
+
+	cfg = readq(base + offset);
+	cfg &= ~(LBK_LINK_CFG_RANGE_MASK |
+		 LBK_LINK_CFG_ID_MASK | LBK_LINK_CFG_BASE_MASK);
+	cfg |=	FIELD_PREP(LBK_LINK_CFG_RANGE_MASK, ilog2(chans));
+	cfg |=	FIELD_PREP(LBK_LINK_CFG_ID_MASK, lbkid);
+	cfg |=	FIELD_PREP(LBK_LINK_CFG_BASE_MASK, hw->lbk_chan_base);
+
+	writeq(cfg, base + offset);
+}
+
+static void rvu_lbk_set_channels(struct rvu *rvu)
+{
+	struct pci_dev *pdev = NULL;
+	void __iomem *base;
+	u64 lbk_const;
+	u8 src, dst;
+	u16 chans;
+
+	/* To loopback packets between multiple NIX blocks
+	 * mutliple LBK blocks are needed. With two NIX blocks,
+	 * four LBK blocks are needed and each LBK block
+	 * source and destination are as follows:
+	 * LBK0 - source NIX0 and destination NIX1
+	 * LBK1 - source NIX0 and destination NIX1
+	 * LBK2 - source NIX1 and destination NIX0
+	 * LBK3 - source NIX1 and destination NIX1
+	 * As per the HRM channel numbers should be programmed as:
+	 * P2X and X2P of LBK0 as same
+	 * P2X and X2P of LBK3 as same
+	 * P2X of LBK1 and X2P of LBK2 as same
+	 * P2X of LBK2 and X2P of LBK1 as same
+	 */
+	while (true) {
+		pdev = pci_get_device(PCI_VENDOR_ID_CAVIUM,
+				      PCI_DEVID_OCTEONTX2_LBK, pdev);
+		if (!pdev)
+			return;
+
+		base = pci_ioremap_bar(pdev, 0);
+		if (!base)
+			goto err_put;
+
+		lbk_const = readq(base + LBK_CONST);
+		chans = FIELD_GET(LBK_CONST_CHANS, lbk_const);
+		dst = FIELD_GET(LBK_CONST_DST, lbk_const);
+		src = FIELD_GET(LBK_CONST_SRC, lbk_const);
+
+		if (src == dst) {
+			if (src == LBK_CONNECT_NIXX(0)) { /* LBK0 */
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_X2P,
+						    0, chans);
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_P2X,
+						    0, chans);
+			} else if (src == LBK_CONNECT_NIXX(1)) { /* LBK3 */
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_X2P,
+						    1, chans);
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_P2X,
+						    1, chans);
+			}
+		} else {
+			if (src == LBK_CONNECT_NIXX(0)) { /* LBK1 */
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_X2P,
+						    0, chans);
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_P2X,
+						    1, chans);
+			} else if (src == LBK_CONNECT_NIXX(1)) { /* LBK2 */
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_X2P,
+						    1, chans);
+				__rvu_lbk_set_chans(rvu, base, LBK_LINK_CFG_P2X,
+						    0, chans);
+			}
+		}
+		iounmap(base);
+	}
+err_put:
+	pci_dev_put(pdev);
+}
+
+static void __rvu_nix_set_channels(struct rvu *rvu, int blkaddr)
+{
+	u64 nix_const = rvu_read64(rvu, blkaddr, NIX_AF_CONST);
+	u16 cgx_chans, lbk_chans, sdp_chans, cpt_chans;
+	struct rvu_hwinfo *hw = rvu->hw;
+	int link, nix_link = 0;
+	u16 start;
+	u64 cfg;
+
+	cgx_chans = nix_const & 0xFFULL;
+	lbk_chans = (nix_const >> 16) & 0xFFULL;
+	sdp_chans = SDP_CHANNELS;
+	cpt_chans = (nix_const >> 32) & 0xFFFULL;
+
+	start = hw->cgx_chan_base;
+	for (link = 0; link < hw->cgx_links; link++, nix_link++) {
+		cfg = rvu_read64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link));
+		cfg &= ~(NIX_AF_LINKX_BASE_MASK | NIX_AF_LINKX_RANGE_MASK);
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_RANGE_MASK, ilog2(cgx_chans));
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_BASE_MASK, start);
+		rvu_write64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link), cfg);
+		start += cgx_chans;
+	}
+
+	start = hw->lbk_chan_base;
+	for (link = 0; link < hw->lbk_links; link++, nix_link++) {
+		cfg = rvu_read64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link));
+		cfg &= ~(NIX_AF_LINKX_BASE_MASK | NIX_AF_LINKX_RANGE_MASK);
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_RANGE_MASK, ilog2(lbk_chans));
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_BASE_MASK, start);
+		rvu_write64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link), cfg);
+		start += lbk_chans;
+	}
+
+	start = hw->sdp_chan_base;
+	for (link = 0; link < hw->sdp_links; link++, nix_link++) {
+		cfg = rvu_read64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link));
+		cfg &= ~(NIX_AF_LINKX_BASE_MASK | NIX_AF_LINKX_RANGE_MASK);
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_RANGE_MASK, ilog2(sdp_chans));
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_BASE_MASK, start);
+		rvu_write64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link), cfg);
+		start += sdp_chans;
+	}
+
+	start = hw->cpt_chan_base;
+	for (link = 0; link < hw->cpt_links; link++, nix_link++) {
+		cfg = rvu_read64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link));
+		cfg &= ~(NIX_AF_LINKX_BASE_MASK | NIX_AF_LINKX_RANGE_MASK);
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_RANGE_MASK, ilog2(cpt_chans));
+		cfg |=	FIELD_PREP(NIX_AF_LINKX_BASE_MASK, start);
+		rvu_write64(rvu, blkaddr, NIX_AF_LINKX_CFG(nix_link), cfg);
+		start += cpt_chans;
+	}
+}
+
+static void rvu_nix_set_channels(struct rvu *rvu)
+{
+	int blkaddr = 0;
+
+	blkaddr = rvu_get_next_nix_blkaddr(rvu, blkaddr);
+	while (blkaddr) {
+		__rvu_nix_set_channels(rvu, blkaddr);
+		blkaddr = rvu_get_next_nix_blkaddr(rvu, blkaddr);
+	}
+}
+
+static void __rvu_rpm_set_channels(int cgxid, int lmacid, u16 base)
+{
+	u64 cfg;
+
+	cfg = cgx_lmac_read(cgxid, lmacid, RPMX_CMRX_LINK_CFG);
+	cfg &= ~(RPMX_CMRX_LINK_BASE_MASK | RPMX_CMRX_LINK_RANGE_MASK);
+
+	/* There is no read-only constant register to read
+	 * the number of channels for LMAC and it is always 16.
+	 */
+	cfg |=	FIELD_PREP(RPMX_CMRX_LINK_RANGE_MASK, ilog2(16));
+	cfg |=	FIELD_PREP(RPMX_CMRX_LINK_BASE_MASK, base);
+	cgx_lmac_write(cgxid, lmacid, RPMX_CMRX_LINK_CFG, cfg);
+}
+
+static void rvu_rpm_set_channels(struct rvu *rvu)
+{
+	struct rvu_hwinfo *hw = rvu->hw;
+	u16 base = hw->cgx_chan_base;
+	int cgx, lmac;
+
+	for (cgx = 0; cgx < rvu->cgx_cnt_max; cgx++) {
+		for (lmac = 0; lmac < hw->lmac_per_cgx; lmac++) {
+			__rvu_rpm_set_channels(cgx, lmac, base);
+			base += 16;
+		}
+	}
+}
+
+void rvu_program_channels(struct rvu *rvu)
+{
+	struct rvu_hwinfo *hw = rvu->hw;
+
+	if (!hw->cap.programmable_chans)
+		return;
+
+	rvu_nix_set_channels(rvu);
+	rvu_lbk_set_channels(rvu);
+	rvu_rpm_set_channels(rvu);
+}

drivers/net/ethernet/marvell/octeontx2/af/rvu_nix.c

@@ -16,6 +16,7 @@
 #include "rvu.h"
 #include "npc.h"
 #include "cgx.h"
+#include "lmac_common.h"
 
 static void nix_free_tx_vtag_entries(struct rvu *rvu, u16 pcifunc);
 static int rvu_nix_get_bpid(struct rvu *rvu, struct nix_bp_cfg_req *req,
@@ -214,6 +215,7 @@ static bool is_valid_txschq(struct rvu *rvu, int blkaddr,
 static int nix_interface_init(struct rvu *rvu, u16 pcifunc, int type, int nixlf)
 {
 	struct rvu_pfvf *pfvf = rvu_get_pfvf(rvu, pcifunc);
+	struct mac_ops *mac_ops;
 	int pkind, pf, vf, lbkid;
 	u8 cgx_id, lmac_id;
 	int err;
@@ -233,17 +235,19 @@ static int nix_interface_init(struct rvu *rvu, u16 pcifunc, int type, int nixlf)
 				"PF_Func 0x%x: Invalid pkind\n", pcifunc);
 			return -EINVAL;
 		}
-		pfvf->rx_chan_base = NIX_CHAN_CGX_LMAC_CHX(cgx_id, lmac_id, 0);
+		pfvf->rx_chan_base = rvu_nix_chan_cgx(rvu, cgx_id, lmac_id, 0);
 		pfvf->tx_chan_base = pfvf->rx_chan_base;
 		pfvf->rx_chan_cnt = 1;
 		pfvf->tx_chan_cnt = 1;
 		cgx_set_pkind(rvu_cgx_pdata(cgx_id, rvu), lmac_id, pkind);
 		rvu_npc_set_pkind(rvu, pkind, pfvf);
 
+		mac_ops = get_mac_ops(rvu_cgx_pdata(cgx_id, rvu));
 		/* By default we enable pause frames */
 		if ((pcifunc & RVU_PFVF_FUNC_MASK) == 0)
-			cgx_lmac_set_pause_frm(rvu_cgx_pdata(cgx_id, rvu),
-					       lmac_id, true, true);
+			mac_ops->mac_enadis_pause_frm(rvu_cgx_pdata(cgx_id,
+								    rvu),
+						      lmac_id, true, true);
 		break;
 	case NIX_INTF_TYPE_LBK:
 		vf = (pcifunc & RVU_PFVF_FUNC_MASK) - 1;
@@ -262,10 +266,10 @@ static int nix_interface_init(struct rvu *rvu, u16 pcifunc, int type, int nixlf)
 		 * loopback channels.Therefore if odd number of AF VFs are
 		 * enabled then the last VF remains with no pair.
 		 */
-		pfvf->rx_chan_base = NIX_CHAN_LBK_CHX(lbkid, vf);
+		pfvf->rx_chan_base = rvu_nix_chan_lbk(rvu, lbkid, vf);
 		pfvf->tx_chan_base = vf & 0x1 ?
-					NIX_CHAN_LBK_CHX(lbkid, vf - 1) :
-					NIX_CHAN_LBK_CHX(lbkid, vf + 1);
+					rvu_nix_chan_lbk(rvu, lbkid, vf - 1) :
+					rvu_nix_chan_lbk(rvu, lbkid, vf + 1);
 		pfvf->rx_chan_cnt = 1;
 		pfvf->tx_chan_cnt = 1;
 		rvu_npc_install_promisc_entry(rvu, pcifunc, nixlf,
@@ -1000,6 +1004,14 @@ int rvu_mbox_handler_nix_aq_enq(struct rvu *rvu,
 	return rvu_nix_aq_enq_inst(rvu, req, rsp);
 }
 #endif
+/* CN10K mbox handler */
+int rvu_mbox_handler_nix_cn10k_aq_enq(struct rvu *rvu,
+				      struct nix_cn10k_aq_enq_req *req,
+				      struct nix_cn10k_aq_enq_rsp *rsp)
+{
+	return rvu_nix_aq_enq_inst(rvu, (struct nix_aq_enq_req *)req,
+				  (struct nix_aq_enq_rsp *)rsp);
+}
 
 int rvu_mbox_handler_nix_hwctx_disable(struct rvu *rvu,
 				       struct hwctx_disable_req *req,
@@ -2535,6 +2547,43 @@ static int nix_af_mark_format_setup(struct rvu *rvu, struct nix_hw *nix_hw,
 	return 0;
 }
 
+static void rvu_get_lbk_link_max_frs(struct rvu *rvu,  u16 *max_mtu)
+{
+	/* CN10K supports LBK FIFO size 72 KB */
+	if (rvu->hw->lbk_bufsize == 0x12000)
+		*max_mtu = CN10K_LBK_LINK_MAX_FRS;
+	else
+		*max_mtu = NIC_HW_MAX_FRS;
+}
+
+static void rvu_get_lmac_link_max_frs(struct rvu *rvu, u16 *max_mtu)
+{
+	/* RPM supports FIFO len 128 KB */
+	if (rvu_cgx_get_fifolen(rvu) == 0x20000)
+		*max_mtu = CN10K_LMAC_LINK_MAX_FRS;
+	else
+		*max_mtu = NIC_HW_MAX_FRS;
+}
+
+int rvu_mbox_handler_nix_get_hw_info(struct rvu *rvu, struct msg_req *req,
+				     struct nix_hw_info *rsp)
+{
+	u16 pcifunc = req->hdr.pcifunc;
+	int blkaddr;
+
+	blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_NIX, pcifunc);
+	if (blkaddr < 0)
+		return NIX_AF_ERR_AF_LF_INVALID;
+
+	if (is_afvf(pcifunc))
+		rvu_get_lbk_link_max_frs(rvu, &rsp->max_mtu);
+	else
+		rvu_get_lmac_link_max_frs(rvu, &rsp->max_mtu);
+
+	rsp->min_mtu = NIC_HW_MIN_FRS;
+	return 0;
+}
+
 int rvu_mbox_handler_nix_stats_rst(struct rvu *rvu, struct msg_req *req,
 				   struct msg_rsp *rsp)
 {
@@ -3099,6 +3148,7 @@ int rvu_mbox_handler_nix_set_hw_frs(struct rvu *rvu, struct nix_frs_cfg *req,
 	u64 cfg, lmac_fifo_len;
 	struct nix_hw *nix_hw;
 	u8 cgx = 0, lmac = 0;
+	u16 max_mtu;
 
 	blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_NIX, pcifunc);
 	if (blkaddr < 0)
@@ -3108,7 +3158,12 @@ int rvu_mbox_handler_nix_set_hw_frs(struct rvu *rvu, struct nix_frs_cfg *req,
 	if (!nix_hw)
 		return -EINVAL;
 
-	if (!req->sdp_link && req->maxlen > NIC_HW_MAX_FRS)
+	if (is_afvf(pcifunc))
+		rvu_get_lbk_link_max_frs(rvu, &max_mtu);
+	else
+		rvu_get_lmac_link_max_frs(rvu, &max_mtu);
+
+	if (!req->sdp_link && req->maxlen > max_mtu)
 		return NIX_AF_ERR_FRS_INVALID;
 
 	if (req->update_minlen && req->minlen < NIC_HW_MIN_FRS)
@@ -3168,7 +3223,8 @@ int rvu_mbox_handler_nix_set_hw_frs(struct rvu *rvu, struct nix_frs_cfg *req,
 
 	/* Update transmit credits for CGX links */
 	lmac_fifo_len =
-		CGX_FIFO_LEN / cgx_get_lmac_cnt(rvu_cgx_pdata(cgx, rvu));
+		rvu_cgx_get_fifolen(rvu) /
+		cgx_get_lmac_cnt(rvu_cgx_pdata(cgx, rvu));
 	cfg = rvu_read64(rvu, blkaddr, NIX_AF_TX_LINKX_NORM_CREDIT(link));
 	cfg &= ~(0xFFFFFULL << 12);
 	cfg |=  ((lmac_fifo_len - req->maxlen) / 16) << 12;
@@ -3208,23 +3264,40 @@ int rvu_mbox_handler_nix_set_rx_cfg(struct rvu *rvu, struct nix_rx_cfg *req,
 	return 0;
 }
 
+static u64 rvu_get_lbk_link_credits(struct rvu *rvu, u16 lbk_max_frs)
+{
+	/* CN10k supports 72KB FIFO size and max packet size of 64k */
+	if (rvu->hw->lbk_bufsize == 0x12000)
+		return (rvu->hw->lbk_bufsize - lbk_max_frs) / 16;
+
+	return 1600; /* 16 * max LBK datarate = 16 * 100Gbps */
+}
+
 static void nix_link_config(struct rvu *rvu, int blkaddr)
 {
 	struct rvu_hwinfo *hw = rvu->hw;
 	int cgx, lmac_cnt, slink, link;
+	u16 lbk_max_frs, lmac_max_frs;
 	u64 tx_credits;
 
+	rvu_get_lbk_link_max_frs(rvu, &lbk_max_frs);
+	rvu_get_lmac_link_max_frs(rvu, &lmac_max_frs);
+
 	/* Set default min/max packet lengths allowed on NIX Rx links.
 	 *
 	 * With HW reset minlen value of 60byte, HW will treat ARP pkts
 	 * as undersize and report them to SW as error pkts, hence
 	 * setting it to 40 bytes.
 	 */
-	for (link = 0; link < (hw->cgx_links + hw->lbk_links); link++) {
+	for (link = 0; link < hw->cgx_links; link++) {
 		rvu_write64(rvu, blkaddr, NIX_AF_RX_LINKX_CFG(link),
-			    NIC_HW_MAX_FRS << 16 | NIC_HW_MIN_FRS);
+				((u64)lmac_max_frs << 16) | NIC_HW_MIN_FRS);
 	}
 
+	for (link = hw->cgx_links; link < hw->lbk_links; link++) {
+		rvu_write64(rvu, blkaddr, NIX_AF_RX_LINKX_CFG(link),
+			    ((u64)lbk_max_frs << 16) | NIC_HW_MIN_FRS);
+	}
 	if (hw->sdp_links) {
 		link = hw->cgx_links + hw->lbk_links;
 		rvu_write64(rvu, blkaddr, NIX_AF_RX_LINKX_CFG(link),
@@ -3236,7 +3309,8 @@ static void nix_link_config(struct rvu *rvu, int blkaddr)
 	 */
 	for (cgx = 0; cgx < hw->cgx; cgx++) {
 		lmac_cnt = cgx_get_lmac_cnt(rvu_cgx_pdata(cgx, rvu));
-		tx_credits = ((CGX_FIFO_LEN / lmac_cnt) - NIC_HW_MAX_FRS) / 16;
+		tx_credits = ((rvu_cgx_get_fifolen(rvu) / lmac_cnt) -
+			       lmac_max_frs) / 16;
 		/* Enable credits and set credit pkt count to max allowed */
 		tx_credits =  (tx_credits << 12) | (0x1FF << 2) | BIT_ULL(1);
 		slink = cgx * hw->lmac_per_cgx;
@@ -3250,7 +3324,7 @@ static void nix_link_config(struct rvu *rvu, int blkaddr)
 	/* Set Tx credits for LBK link */
 	slink = hw->cgx_links;
 	for (link = slink; link < (slink + hw->lbk_links); link++) {
-		tx_credits = 1000; /* 10 * max LBK datarate = 10 * 100Gbps */
+		tx_credits = rvu_get_lbk_link_credits(rvu, lbk_max_frs);
 		/* Enable credits and set credit pkt count to max allowed */
 		tx_credits =  (tx_credits << 12) | (0x1FF << 2) | BIT_ULL(1);
 		rvu_write64(rvu, blkaddr,
@@ -3381,14 +3455,6 @@ static int rvu_nix_block_init(struct rvu *rvu, struct nix_hw *nix_hw)
 	if (err)
 		return err;
 
-	/* Set num of links of each type */
-	cfg = rvu_read64(rvu, blkaddr, NIX_AF_CONST);
-	hw->cgx = (cfg >> 12) & 0xF;
-	hw->lmac_per_cgx = (cfg >> 8) & 0xF;
-	hw->cgx_links = hw->cgx * hw->lmac_per_cgx;
-	hw->lbk_links = (cfg >> 24) & 0xF;
-	hw->sdp_links = 1;
-
 	/* Initialize admin queue */
 	err = nix_aq_init(rvu, block);
 	if (err)
@@ -3623,10 +3689,14 @@ static int rvu_nix_lf_ptp_tx_cfg(struct rvu *rvu, u16 pcifunc, bool enable)
 {
 	struct rvu_hwinfo *hw = rvu->hw;
 	struct rvu_block *block;
-	int blkaddr;
+	int blkaddr, pf;
 	int nixlf;
 	u64 cfg;
 
+	pf = rvu_get_pf(pcifunc);
+	if (!is_mac_feature_supported(rvu, pf, RVU_LMAC_FEAT_PTP))
+		return 0;
+
 	blkaddr = rvu_get_blkaddr(rvu, BLKTYPE_NIX, pcifunc);
 	if (blkaddr < 0)
 		return NIX_AF_ERR_AF_LF_INVALID;

drivers/net/ethernet/marvell/octeontx2/af/rvu_npc.c

@@ -102,9 +102,9 @@ int npc_mcam_verify_channel(struct rvu *rvu, u16 pcifunc, u8 intf, u16 channel)
 			return -EINVAL;
 	} else {
 		rvu_get_cgx_lmac_id(rvu->pf2cgxlmac_map[pf], &cgx_id, &lmac_id);
-		base = NIX_CHAN_CGX_LMAC_CHX(cgx_id, lmac_id, 0x0);
+		base = rvu_nix_chan_cgx(rvu, cgx_id, lmac_id, 0x0);
 		/* CGX mapped functions has maximum of 16 channels */
-		end = NIX_CHAN_CGX_LMAC_CHX(cgx_id, lmac_id, 0xF);
+		end = rvu_nix_chan_cgx(rvu, cgx_id, lmac_id, 0xF);
 	}
 
 	if (channel < base || channel > end)

drivers/net/ethernet/marvell/octeontx2/af/rvu_reg.h

@@ -44,6 +44,11 @@
 #define RVU_AF_PFME_INT_W1S                 (0x28c8)
 #define RVU_AF_PFME_INT_ENA_W1S             (0x28d0)
 #define RVU_AF_PFME_INT_ENA_W1C             (0x28d8)
+#define RVU_AF_PFX_BAR4_ADDR(a)             (0x5000 | (a) << 4)
+#define RVU_AF_PFX_BAR4_CFG                 (0x5200 | (a) << 4)
+#define RVU_AF_PFX_VF_BAR4_ADDR             (0x5400 | (a) << 4)
+#define RVU_AF_PFX_VF_BAR4_CFG              (0x5600 | (a) << 4)
+#define RVU_AF_PFX_LMTLINE_ADDR             (0x5800 | (a) << 4)
 
 /* Admin function's privileged PF/VF registers */
 #define RVU_PRIV_CONST                      (0x8000000)
@@ -100,6 +105,8 @@
 #define RVU_PF_MSIX_VECX_ADDR(a)            (0x000 | (a) << 4)
 #define RVU_PF_MSIX_VECX_CTL(a)             (0x008 | (a) << 4)
 #define RVU_PF_MSIX_PBAX(a)                 (0xF0000 | (a) << 3)
+#define RVU_PF_VF_MBOX_ADDR                 (0xC40)
+#define RVU_PF_LMTLINE_ADDR                 (0xC48)
 
 /* RVU VF registers */
 #define	RVU_VF_VFPF_MBOX0		    (0x00000)
@@ -399,12 +406,16 @@
 #define NIX_AF_RX_NPC_MIRROR_RCV	(0x4720)
 #define NIX_AF_RX_NPC_MIRROR_DROP	(0x4730)
 #define NIX_AF_RX_ACTIVE_CYCLES_PCX(a)	(0x4800 | (a) << 16)
+#define NIX_AF_LINKX_CFG(a)		(0x4010 | (a) << 17)
 
 #define NIX_PRIV_AF_INT_CFG		(0x8000000)
 #define NIX_PRIV_LFX_CFG		(0x8000010)
 #define NIX_PRIV_LFX_INT_CFG		(0x8000020)
 #define NIX_AF_RVU_LF_CFG_DEBUG		(0x8000030)
 
+#define NIX_AF_LINKX_BASE_MASK		GENMASK_ULL(11, 0)
+#define NIX_AF_LINKX_RANGE_MASK		GENMASK_ULL(19, 16)
+
 /* SSO */
 #define SSO_AF_CONST			(0x1000)
 #define SSO_AF_CONST1			(0x1008)
@@ -637,4 +648,17 @@
 		(0x00F00 | (a) << 5 | (b) << 4)
 #define NDC_AF_BANKX_HIT_PC(a)		(0x01000 | (a) << 3)
 #define NDC_AF_BANKX_MISS_PC(a)		(0x01100 | (a) << 3)
+
+/* LBK */
+#define LBK_CONST			(0x10ull)
+#define LBK_LINK_CFG_P2X		(0x400ull)
+#define LBK_LINK_CFG_X2P		(0x408ull)
+#define LBK_CONST_CHANS			GENMASK_ULL(47, 32)
+#define LBK_CONST_DST			GENMASK_ULL(31, 28)
+#define LBK_CONST_SRC			GENMASK_ULL(27, 24)
+#define LBK_CONST_BUF_SIZE		GENMASK_ULL(23, 0)
+#define LBK_LINK_CFG_RANGE_MASK		GENMASK_ULL(19, 16)
+#define LBK_LINK_CFG_ID_MASK		GENMASK_ULL(11, 6)
+#define LBK_LINK_CFG_BASE_MASK		GENMASK_ULL(5, 0)
+
 #endif /* RVU_REG_H */

drivers/net/ethernet/marvell/octeontx2/nic/Makefile

@@ -3,11 +3,11 @@
 # Makefile for Marvell's OcteonTX2 ethernet device drivers
 #
 
-obj-$(CONFIG_OCTEONTX2_PF) += octeontx2_nicpf.o
-obj-$(CONFIG_OCTEONTX2_VF) += octeontx2_nicvf.o
+obj-$(CONFIG_OCTEONTX2_PF) += rvu_nicpf.o
+obj-$(CONFIG_OCTEONTX2_VF) += rvu_nicvf.o
 
-octeontx2_nicpf-y := otx2_pf.o otx2_common.o otx2_txrx.o otx2_ethtool.o \
-		     otx2_ptp.o otx2_flows.o
-octeontx2_nicvf-y := otx2_vf.o
+rvu_nicpf-y := otx2_pf.o otx2_common.o otx2_txrx.o otx2_ethtool.o \
+		     otx2_ptp.o otx2_flows.o cn10k.o
+rvu_nicvf-y := otx2_vf.o
 
 ccflags-y += -I$(srctree)/drivers/net/ethernet/marvell/octeontx2/af

drivers/net/ethernet/marvell/octeontx2/nic/cn10k.c

+// SPDX-License-Identifier: GPL-2.0
+/* Marvell OcteonTx2 RVU Physcial Function ethernet driver
+ *
+ * Copyright (C) 2020 Marvell.
+ */
+
+#include "cn10k.h"
+#include "otx2_reg.h"
+#include "otx2_struct.h"
+
+static struct dev_hw_ops	otx2_hw_ops = {
+	.sq_aq_init = otx2_sq_aq_init,
+	.sqe_flush = otx2_sqe_flush,
+	.aura_freeptr = otx2_aura_freeptr,
+	.refill_pool_ptrs = otx2_refill_pool_ptrs,
+};
+
+static struct dev_hw_ops cn10k_hw_ops = {
+	.sq_aq_init = cn10k_sq_aq_init,
+	.sqe_flush = cn10k_sqe_flush,
+	.aura_freeptr = cn10k_aura_freeptr,
+	.refill_pool_ptrs = cn10k_refill_pool_ptrs,
+};
+
+int cn10k_pf_lmtst_init(struct otx2_nic *pf)
+{
+	int size, num_lines;
+	u64 base;
+
+	if (!test_bit(CN10K_LMTST, &pf->hw.cap_flag)) {
+		pf->hw_ops = &otx2_hw_ops;
+		return 0;
+	}
+
+	pf->hw_ops = &cn10k_hw_ops;
+	base = pci_resource_start(pf->pdev, PCI_MBOX_BAR_NUM) +
+		       (MBOX_SIZE * (pf->total_vfs + 1));
+
+	size = pci_resource_len(pf->pdev, PCI_MBOX_BAR_NUM) -
+	       (MBOX_SIZE * (pf->total_vfs + 1));
+
+	pf->hw.lmt_base = ioremap(base, size);
+
+	if (!pf->hw.lmt_base) {
+		dev_err(pf->dev, "Unable to map PF LMTST region\n");
+		return -ENOMEM;
+	}
+
+	/* FIXME: Get the num of LMTST lines from LMT table */
+	pf->tot_lmt_lines = size / LMT_LINE_SIZE;
+	num_lines = (pf->tot_lmt_lines - NIX_LMTID_BASE) /
+			    pf->hw.tx_queues;
+	/* Number of LMT lines per SQ queues */
+	pf->nix_lmt_lines = num_lines > 32 ? 32 : num_lines;
+
+	pf->nix_lmt_size = pf->nix_lmt_lines * LMT_LINE_SIZE;
+	return 0;
+}
+
+int cn10k_vf_lmtst_init(struct otx2_nic *vf)
+{
+	int size, num_lines;
+
+	if (!test_bit(CN10K_LMTST, &vf->hw.cap_flag)) {
+		vf->hw_ops = &otx2_hw_ops;
+		return 0;
+	}
+
+	vf->hw_ops = &cn10k_hw_ops;
+	size = pci_resource_len(vf->pdev, PCI_MBOX_BAR_NUM);
+	vf->hw.lmt_base = ioremap_wc(pci_resource_start(vf->pdev,
+							PCI_MBOX_BAR_NUM),
+				     size);
+	if (!vf->hw.lmt_base) {
+		dev_err(vf->dev, "Unable to map VF LMTST region\n");
+		return -ENOMEM;
+	}
+
+	vf->tot_lmt_lines = size / LMT_LINE_SIZE;
+	/* LMTST lines per SQ */
+	num_lines = (vf->tot_lmt_lines - NIX_LMTID_BASE) /
+			    vf->hw.tx_queues;
+	vf->nix_lmt_lines = num_lines > 32 ? 32 : num_lines;
+	vf->nix_lmt_size = vf->nix_lmt_lines * LMT_LINE_SIZE;
+	return 0;
+}
+EXPORT_SYMBOL(cn10k_vf_lmtst_init);
+
+int cn10k_sq_aq_init(void *dev, u16 qidx, u16 sqb_aura)
+{
+	struct nix_cn10k_aq_enq_req *aq;
+	struct otx2_nic *pfvf = dev;
+	struct otx2_snd_queue *sq;
+
+	sq = &pfvf->qset.sq[qidx];
+	sq->lmt_addr = (__force u64 *)((u64)pfvf->hw.nix_lmt_base +
+			       (qidx * pfvf->nix_lmt_size));
+
+	/* Get memory to put this msg */
+	aq = otx2_mbox_alloc_msg_nix_cn10k_aq_enq(&pfvf->mbox);
+	if (!aq)
+		return -ENOMEM;
+
+	aq->sq.cq = pfvf->hw.rx_queues + qidx;
+	aq->sq.max_sqe_size = NIX_MAXSQESZ_W16; /* 128 byte */
+	aq->sq.cq_ena = 1;
+	aq->sq.ena = 1;
+	/* Only one SMQ is allocated, map all SQ's to that SMQ  */
+	aq->sq.smq = pfvf->hw.txschq_list[NIX_TXSCH_LVL_SMQ][0];
+	/* FIXME: set based on NIX_AF_DWRR_RPM_MTU*/
+	aq->sq.smq_rr_weight = pfvf->netdev->mtu;
+	aq->sq.default_chan = pfvf->hw.tx_chan_base;
+	aq->sq.sqe_stype = NIX_STYPE_STF; /* Cache SQB */
+	aq->sq.sqb_aura = sqb_aura;
+	aq->sq.sq_int_ena = NIX_SQINT_BITS;
+	aq->sq.qint_idx = 0;
+	/* Due pipelining impact minimum 2000 unused SQ CQE's
+	 * need to maintain to avoid CQ overflow.
+	 */
+	aq->sq.cq_limit = ((SEND_CQ_SKID * 256) / (pfvf->qset.sqe_cnt));
+
+	/* Fill AQ info */
+	aq->qidx = qidx;
+	aq->ctype = NIX_AQ_CTYPE_SQ;
+	aq->op = NIX_AQ_INSTOP_INIT;
+
+	return otx2_sync_mbox_msg(&pfvf->mbox);
+}
+
+#define NPA_MAX_BURST 16
+void cn10k_refill_pool_ptrs(void *dev, struct otx2_cq_queue *cq)
+{
+	struct otx2_nic *pfvf = dev;
+	u64 ptrs[NPA_MAX_BURST];
+	int num_ptrs = 1;
+	dma_addr_t bufptr;
+
+	/* Refill pool with new buffers */
+	while (cq->pool_ptrs) {
+		if (otx2_alloc_buffer(pfvf, cq, &bufptr)) {
+			if (num_ptrs--)
+				__cn10k_aura_freeptr(pfvf, cq->cq_idx, ptrs,
+						     num_ptrs,
+						     cq->rbpool->lmt_addr);
+			break;
+		}
+		cq->pool_ptrs--;
+		ptrs[num_ptrs] = (u64)bufptr + OTX2_HEAD_ROOM;
+		num_ptrs++;
+		if (num_ptrs == NPA_MAX_BURST || cq->pool_ptrs == 0) {
+			__cn10k_aura_freeptr(pfvf, cq->cq_idx, ptrs,
+					     num_ptrs,
+					     cq->rbpool->lmt_addr);
+			num_ptrs = 1;
+		}
+	}
+}
+
+void cn10k_sqe_flush(void *dev, struct otx2_snd_queue *sq, int size, int qidx)
+{
+	struct otx2_nic *pfvf = dev;
+	int lmt_id = NIX_LMTID_BASE + (qidx * pfvf->nix_lmt_lines);
+	u64 val = 0, tar_addr = 0;
+
+	/* FIXME: val[0:10] LMT_ID.
+	 * [12:15] no of LMTST - 1 in the burst.
+	 * [19:63] data size of each LMTST in the burst except first.
+	 */
+	val = (lmt_id & 0x7FF);
+	/* Target address for LMTST flush tells HW how many 128bit
+	 * words are present.
+	 * tar_addr[6:4] size of first LMTST - 1 in units of 128b.
+	 */
+	tar_addr |= sq->io_addr | (((size / 16) - 1) & 0x7) << 4;
+	dma_wmb();
+	memcpy(sq->lmt_addr, sq->sqe_base, size);
+	cn10k_lmt_flush(val, tar_addr);
+
+	sq->head++;
+	sq->head &= (sq->sqe_cnt - 1);
+}

drivers/net/ethernet/marvell/octeontx2/nic/cn10k.h

+/* SPDX-License-Identifier: GPL-2.0
+ * Marvell OcteonTx2 RVU Ethernet driver
+ *
+ * Copyright (C) 2020 Marvell.
+ */
+
+#ifndef CN10K_H
+#define CN10K_H
+
+#include "otx2_common.h"
+
+void cn10k_refill_pool_ptrs(void *dev, struct otx2_cq_queue *cq);
+void cn10k_sqe_flush(void *dev, struct otx2_snd_queue *sq, int size, int qidx);
+int cn10k_sq_aq_init(void *dev, u16 qidx, u16 sqb_aura);
+int cn10k_pf_lmtst_init(struct otx2_nic *pf);
+int cn10k_vf_lmtst_init(struct otx2_nic *vf);
+#endif /* CN10K_H */

drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.c

@@ -15,6 +15,7 @@
 #include "otx2_reg.h"
 #include "otx2_common.h"
 #include "otx2_struct.h"
+#include "cn10k.h"
 
 static void otx2_nix_rq_op_stats(struct queue_stats *stats,
 				 struct otx2_nic *pfvf, int qidx)
@@ -229,7 +230,6 @@ int otx2_hw_set_mtu(struct otx2_nic *pfvf, int mtu)
 		return -ENOMEM;
 	}
 
-	pfvf->max_frs = mtu +  OTX2_ETH_HLEN;
 	req->maxlen = pfvf->max_frs;
 
 	err = otx2_sync_mbox_msg(&pfvf->mbox);
@@ -526,6 +526,26 @@ static int otx2_alloc_rbuf(struct otx2_nic *pfvf, struct otx2_pool *pool,
 	return ret;
 }
 
+int otx2_alloc_buffer(struct otx2_nic *pfvf, struct otx2_cq_queue *cq,
+		      dma_addr_t *dma)
+{
+	if (unlikely(__otx2_alloc_rbuf(pfvf, cq->rbpool, dma))) {
+		struct refill_work *work;
+		struct delayed_work *dwork;
+
+		work = &pfvf->refill_wrk[cq->cq_idx];
+		dwork = &work->pool_refill_work;
+		/* Schedule a task if no other task is running */
+		if (!cq->refill_task_sched) {
+			cq->refill_task_sched = true;
+			schedule_delayed_work(dwork,
+					      msecs_to_jiffies(100));
+		}
+		return -ENOMEM;
+	}
+	return 0;
+}
+
 void otx2_tx_timeout(struct net_device *netdev, unsigned int txq)
 {
 	struct otx2_nic *pfvf = netdev_priv(netdev);
@@ -585,8 +605,8 @@ int otx2_txschq_config(struct otx2_nic *pfvf, int lvl)
 	/* Set topology e.t.c configuration */
 	if (lvl == NIX_TXSCH_LVL_SMQ) {
 		req->reg[0] = NIX_AF_SMQX_CFG(schq);
-		req->regval[0] = ((OTX2_MAX_MTU + OTX2_ETH_HLEN) << 8) |
-				   OTX2_MIN_MTU;
+		req->regval[0] = ((pfvf->netdev->max_mtu + OTX2_ETH_HLEN) << 8)
+				  | OTX2_MIN_MTU;
 
 		req->regval[0] |= (0x20ULL << 51) | (0x80ULL << 39) |
 				  (0x2ULL << 36);
@@ -728,9 +748,6 @@ void otx2_sqb_flush(struct otx2_nic *pfvf)
 #define RQ_PASS_LVL_AURA (255 - ((95 * 256) / 100)) /* RED when 95% is full */
 #define RQ_DROP_LVL_AURA (255 - ((99 * 256) / 100)) /* Drop when 99% is full */
 
-/* Send skid of 2000 packets required for CQ size of 4K CQEs. */
-#define SEND_CQ_SKID	2000
-
 static int otx2_rq_init(struct otx2_nic *pfvf, u16 qidx, u16 lpb_aura)
 {
 	struct otx2_qset *qset = &pfvf->qset;
@@ -764,11 +781,48 @@ static int otx2_rq_init(struct otx2_nic *pfvf, u16 qidx, u16 lpb_aura)
 	return otx2_sync_mbox_msg(&pfvf->mbox);
 }
 
+int otx2_sq_aq_init(void *dev, u16 qidx, u16 sqb_aura)
+{
+	struct otx2_nic *pfvf = dev;
+	struct otx2_snd_queue *sq;
+	struct nix_aq_enq_req *aq;
+
+	sq = &pfvf->qset.sq[qidx];
+	sq->lmt_addr = (__force u64 *)(pfvf->reg_base + LMT_LF_LMTLINEX(qidx));
+	/* Get memory to put this msg */
+	aq = otx2_mbox_alloc_msg_nix_aq_enq(&pfvf->mbox);
+	if (!aq)
+		return -ENOMEM;
+
+	aq->sq.cq = pfvf->hw.rx_queues + qidx;
+	aq->sq.max_sqe_size = NIX_MAXSQESZ_W16; /* 128 byte */
+	aq->sq.cq_ena = 1;
+	aq->sq.ena = 1;
+	/* Only one SMQ is allocated, map all SQ's to that SMQ  */
+	aq->sq.smq = pfvf->hw.txschq_list[NIX_TXSCH_LVL_SMQ][0];
+	aq->sq.smq_rr_quantum = DFLT_RR_QTM;
+	aq->sq.default_chan = pfvf->hw.tx_chan_base;
+	aq->sq.sqe_stype = NIX_STYPE_STF; /* Cache SQB */
+	aq->sq.sqb_aura = sqb_aura;
+	aq->sq.sq_int_ena = NIX_SQINT_BITS;
+	aq->sq.qint_idx = 0;
+	/* Due pipelining impact minimum 2000 unused SQ CQE's
+	 * need to maintain to avoid CQ overflow.
+	 */
+	aq->sq.cq_limit = ((SEND_CQ_SKID * 256) / (pfvf->qset.sqe_cnt));
+
+	/* Fill AQ info */
+	aq->qidx = qidx;
+	aq->ctype = NIX_AQ_CTYPE_SQ;
+	aq->op = NIX_AQ_INSTOP_INIT;
+
+	return otx2_sync_mbox_msg(&pfvf->mbox);
+}
+
 static int otx2_sq_init(struct otx2_nic *pfvf, u16 qidx, u16 sqb_aura)
 {
 	struct otx2_qset *qset = &pfvf->qset;
 	struct otx2_snd_queue *sq;
-	struct nix_aq_enq_req *aq;
 	struct otx2_pool *pool;
 	int err;
 
@@ -805,40 +859,13 @@ static int otx2_sq_init(struct otx2_nic *pfvf, u16 qidx, u16 sqb_aura)
 	sq->sqe_thresh = ((sq->num_sqbs * sq->sqe_per_sqb) * 10) / 100;
 	sq->aura_id = sqb_aura;
 	sq->aura_fc_addr = pool->fc_addr->base;
-	sq->lmt_addr = (__force u64 *)(pfvf->reg_base + LMT_LF_LMTLINEX(qidx));
 	sq->io_addr = (__force u64)otx2_get_regaddr(pfvf, NIX_LF_OP_SENDX(0));
 
 	sq->stats.bytes = 0;
 	sq->stats.pkts = 0;
 
-	/* Get memory to put this msg */
-	aq = otx2_mbox_alloc_msg_nix_aq_enq(&pfvf->mbox);
-	if (!aq)
-		return -ENOMEM;
-
-	aq->sq.cq = pfvf->hw.rx_queues + qidx;
-	aq->sq.max_sqe_size = NIX_MAXSQESZ_W16; /* 128 byte */
-	aq->sq.cq_ena = 1;
-	aq->sq.ena = 1;
-	/* Only one SMQ is allocated, map all SQ's to that SMQ  */
-	aq->sq.smq = pfvf->hw.txschq_list[NIX_TXSCH_LVL_SMQ][0];
-	aq->sq.smq_rr_quantum = DFLT_RR_QTM;
-	aq->sq.default_chan = pfvf->hw.tx_chan_base;
-	aq->sq.sqe_stype = NIX_STYPE_STF; /* Cache SQB */
-	aq->sq.sqb_aura = sqb_aura;
-	aq->sq.sq_int_ena = NIX_SQINT_BITS;
-	aq->sq.qint_idx = 0;
-	/* Due pipelining impact minimum 2000 unused SQ CQE's
-	 * need to maintain to avoid CQ overflow.
-	 */
-	aq->sq.cq_limit = ((SEND_CQ_SKID * 256) / (sq->sqe_cnt));
-
-	/* Fill AQ info */
-	aq->qidx = qidx;
-	aq->ctype = NIX_AQ_CTYPE_SQ;
-	aq->op = NIX_AQ_INSTOP_INIT;
+	return pfvf->hw_ops->sq_aq_init(pfvf, qidx, sqb_aura);
 
-	return otx2_sync_mbox_msg(&pfvf->mbox);
 }
 
 static int otx2_cq_init(struct otx2_nic *pfvf, u16 qidx)
@@ -942,7 +969,7 @@ static void otx2_pool_refill_task(struct work_struct *work)
 			}
 			return;
 		}
-		otx2_aura_freeptr(pfvf, qidx, bufptr + OTX2_HEAD_ROOM);
+		pfvf->hw_ops->aura_freeptr(pfvf, qidx, bufptr + OTX2_HEAD_ROOM);
 		cq->pool_ptrs--;
 	}
 	cq->refill_task_sched = false;
@@ -1186,6 +1213,11 @@ static int otx2_pool_init(struct otx2_nic *pfvf, u16 pool_id,
 
 	pool->rbsize = buf_size;
 
+	/* Set LMTST addr for NPA batch free */
+	if (test_bit(CN10K_LMTST, &pfvf->hw.cap_flag))
+		pool->lmt_addr = (__force u64 *)((u64)pfvf->hw.npa_lmt_base +
+						 (pool_id * LMT_LINE_SIZE));
+
 	/* Initialize this pool's context via AF */
 	aq = otx2_mbox_alloc_msg_npa_aq_enq(&pfvf->mbox);
 	if (!aq) {
@@ -1274,7 +1306,7 @@ int otx2_sq_aura_pool_init(struct otx2_nic *pfvf)
 		for (ptr = 0; ptr < num_sqbs; ptr++) {
 			if (otx2_alloc_rbuf(pfvf, pool, &bufptr))
 				return -ENOMEM;
-			otx2_aura_freeptr(pfvf, pool_id, bufptr);
+			pfvf->hw_ops->aura_freeptr(pfvf, pool_id, bufptr);
 			sq->sqb_ptrs[sq->sqb_count++] = (u64)bufptr;
 		}
 	}
@@ -1324,8 +1356,8 @@ int otx2_rq_aura_pool_init(struct otx2_nic *pfvf)
 		for (ptr = 0; ptr < num_ptrs; ptr++) {
 			if (otx2_alloc_rbuf(pfvf, pool, &bufptr))
 				return -ENOMEM;
-			otx2_aura_freeptr(pfvf, pool_id,
-					  bufptr + OTX2_HEAD_ROOM);
+			pfvf->hw_ops->aura_freeptr(pfvf, pool_id,
+						   bufptr + OTX2_HEAD_ROOM);
 		}
 	}
 
@@ -1604,6 +1636,46 @@ void otx2_set_cints_affinity(struct otx2_nic *pfvf)
 	}
 }
 
+u16 otx2_get_max_mtu(struct otx2_nic *pfvf)
+{
+	struct nix_hw_info *rsp;
+	struct msg_req *req;
+	u16 max_mtu;
+	int rc;
+
+	mutex_lock(&pfvf->mbox.lock);
+
+	req = otx2_mbox_alloc_msg_nix_get_hw_info(&pfvf->mbox);
+	if (!req) {
+		rc =  -ENOMEM;
+		goto out;
+	}
+
+	rc = otx2_sync_mbox_msg(&pfvf->mbox);
+	if (!rc) {
+		rsp = (struct nix_hw_info *)
+		       otx2_mbox_get_rsp(&pfvf->mbox.mbox, 0, &req->hdr);
+
+		/* HW counts VLAN insertion bytes (8 for double tag)
+		 * irrespective of whether SQE is requesting to insert VLAN
+		 * in the packet or not. Hence these 8 bytes have to be
+		 * discounted from max packet size otherwise HW will throw
+		 * SMQ errors
+		 */
+		max_mtu = rsp->max_mtu - 8 - OTX2_ETH_HLEN;
+	}
+
+out:
+	mutex_unlock(&pfvf->mbox.lock);
+	if (rc) {
+		dev_warn(pfvf->dev,
+			 "Failed to get MTU from hardware setting default value(1500)\n");
+		max_mtu = 1500;
+	}
+	return max_mtu;
+}
+EXPORT_SYMBOL(otx2_get_max_mtu);
+
 #define M(_name, _id, _fn_name, _req_type, _rsp_type)			\
 int __weak								\
 otx2_mbox_up_handler_ ## _fn_name(struct otx2_nic *pfvf,		\

drivers/net/ethernet/marvell/octeontx2/nic/otx2_common.h

@@ -50,6 +50,9 @@ enum arua_mapped_qtypes {
 #define NIX_LF_ERR_VEC				0x81
 #define NIX_LF_POISON_VEC			0x82
 
+/* Send skid of 2000 packets required for CQ size of 4K CQEs. */
+#define SEND_CQ_SKID	2000
+
 /* RSS configuration */
 struct otx2_rss_ctx {
 	u8  ind_tbl[MAX_RSS_INDIR_TBL_SIZE];
@@ -190,7 +193,6 @@ struct otx2_hw {
 	u8			lso_tsov6_idx;
 	u8			lso_udpv4_idx;
 	u8			lso_udpv6_idx;
-	u8			hw_tso;
 
 	/* MSI-X */
 	u8			cint_cnt; /* CQ interrupt count */
@@ -208,6 +210,16 @@ struct otx2_hw {
 	u64			cgx_fec_uncorr_blks;
 	u8			cgx_links;  /* No. of CGX links present in HW */
 	u8			lbk_links;  /* No. of LBK links present in HW */
+#define HW_TSO			BIT_ULL(0)
+#define CN10K_MBOX		BIT_ULL(1)
+#define CN10K_LMTST		BIT_ULL(2)
+	unsigned long		cap_flag;
+
+#define LMT_LINE_SIZE		128
+#define NIX_LMTID_BASE		72 /* RX + TX + XDP */
+	void __iomem		*lmt_base;
+	u64			*npa_lmt_base;
+	u64			*nix_lmt_base;
 };
 
 struct otx2_vf_config {
@@ -266,9 +278,18 @@ struct otx2_flow_config {
 	struct list_head	flow_list;
 };
 
+struct dev_hw_ops {
+	int	(*sq_aq_init)(void *dev, u16 qidx, u16 sqb_aura);
+	void	(*sqe_flush)(void *dev, struct otx2_snd_queue *sq,
+			     int size, int qidx);
+	void	(*refill_pool_ptrs)(void *dev, struct otx2_cq_queue *cq);
+	void	(*aura_freeptr)(void *dev, int aura, u64 buf);
+};
+
 struct otx2_nic {
 	void __iomem		*reg_base;
 	struct net_device	*netdev;
+	struct dev_hw_ops	*hw_ops;
 	void			*iommu_domain;
 	u16			max_frs;
 	u16			rbsize; /* Receive buffer size */
@@ -317,6 +338,10 @@ struct otx2_nic {
 
 	/* Block address of NIX either BLKADDR_NIX0 or BLKADDR_NIX1 */
 	int			nix_blkaddr;
+	/* LMTST Lines info */
+	u16			tot_lmt_lines;
+	u16			nix_lmt_lines;
+	u32			nix_lmt_size;
 
 	struct otx2_ptp		*ptp;
 	struct hwtstamp_config	tstamp;
@@ -341,6 +366,25 @@ static inline bool is_96xx_B0(struct pci_dev *pdev)
 		(pdev->subsystem_device == PCI_SUBSYS_DEVID_96XX_RVU_PFVF);
 }
 
+/* REVID for PCIe devices.
+ * Bits 0..1: minor pass, bit 3..2: major pass
+ * bits 7..4: midr id
+ */
+#define PCI_REVISION_ID_96XX		0x00
+#define PCI_REVISION_ID_95XX		0x10
+#define PCI_REVISION_ID_LOKI		0x20
+#define PCI_REVISION_ID_98XX		0x30
+#define PCI_REVISION_ID_95XXMM		0x40
+
+static inline bool is_dev_otx2(struct pci_dev *pdev)
+{
+	u8 midr = pdev->revision & 0xF0;
+
+	return (midr == PCI_REVISION_ID_96XX || midr == PCI_REVISION_ID_95XX ||
+		midr == PCI_REVISION_ID_LOKI || midr == PCI_REVISION_ID_98XX ||
+		midr == PCI_REVISION_ID_95XXMM);
+}
+
 static inline void otx2_setup_dev_hw_settings(struct otx2_nic *pfvf)
 {
 	struct otx2_hw *hw = &pfvf->hw;
@@ -349,10 +393,10 @@ static inline void otx2_setup_dev_hw_settings(struct otx2_nic *pfvf)
 	pfvf->hw.cq_ecount_wait = CQ_CQE_THRESH_DEFAULT;
 	pfvf->hw.cq_qcount_wait = CQ_QCOUNT_DEFAULT;
 
-	hw->hw_tso = true;
+	__set_bit(HW_TSO, &hw->cap_flag);
 
 	if (is_96xx_A0(pfvf->pdev)) {
-		hw->hw_tso = false;
+		__clear_bit(HW_TSO, &hw->cap_flag);
 
 		/* Time based irq coalescing is not supported */
 		pfvf->hw.cq_qcount_wait = 0x0;
@@ -363,6 +407,10 @@ static inline void otx2_setup_dev_hw_settings(struct otx2_nic *pfvf)
 		pfvf->hw.rq_skid = 600;
 		pfvf->qset.rqe_cnt = Q_COUNT(Q_SIZE_1K);
 	}
+	if (!is_dev_otx2(pfvf->pdev)) {
+		__set_bit(CN10K_MBOX, &hw->cap_flag);
+		__set_bit(CN10K_LMTST, &hw->cap_flag);
+	}
 }
 
 /* Register read/write APIs */
@@ -471,10 +519,51 @@ static inline u64 otx2_atomic64_add(u64 incr, u64 *ptr)
 }
 
 #else
-#define otx2_write128(lo, hi, addr)
+#define otx2_write128(lo, hi, addr)		writeq((hi) | (lo), addr)
 #define otx2_atomic64_add(incr, ptr)		({ *ptr += incr; })
 #endif
 
+static inline void __cn10k_aura_freeptr(struct otx2_nic *pfvf, u64 aura,
+					u64 *ptrs, u64 num_ptrs,
+					u64 *lmt_addr)
+{
+	u64 size = 0, count_eot = 0;
+	u64 tar_addr, val = 0;
+
+	tar_addr = (__force u64)otx2_get_regaddr(pfvf, NPA_LF_AURA_BATCH_FREE0);
+	/* LMTID is same as AURA Id */
+	val = (aura & 0x7FF) | BIT_ULL(63);
+	/* Set if [127:64] of last 128bit word has a valid pointer */
+	count_eot = (num_ptrs % 2) ? 0ULL : 1ULL;
+	/* Set AURA ID to free pointer */
+	ptrs[0] = (count_eot << 32) | (aura & 0xFFFFF);
+	/* Target address for LMTST flush tells HW how many 128bit
+	 * words are valid from NPA_LF_AURA_BATCH_FREE0.
+	 *
+	 * tar_addr[6:4] is LMTST size-1 in units of 128b.
+	 */
+	if (num_ptrs > 2) {
+		size = (sizeof(u64) * num_ptrs) / 16;
+		if (!count_eot)
+			size++;
+		tar_addr |=  ((size - 1) & 0x7) << 4;
+	}
+	memcpy(lmt_addr, ptrs, sizeof(u64) * num_ptrs);
+	/* Perform LMTST flush */
+	cn10k_lmt_flush(val, tar_addr);
+}
+
+static inline void cn10k_aura_freeptr(void *dev, int aura, u64 buf)
+{
+	struct otx2_nic *pfvf = dev;
+	struct otx2_pool *pool;
+	u64 ptrs[2];
+
+	pool = &pfvf->qset.pool[aura];
+	ptrs[1] = buf;
+	__cn10k_aura_freeptr(pfvf, aura, ptrs, 2, pool->lmt_addr);
+}
+
 /* Alloc pointer from pool/aura */
 static inline u64 otx2_aura_allocptr(struct otx2_nic *pfvf, int aura)
 {
@@ -486,11 +575,12 @@ static inline u64 otx2_aura_allocptr(struct otx2_nic *pfvf, int aura)
 }
 
 /* Free pointer to a pool/aura */
-static inline void otx2_aura_freeptr(struct otx2_nic *pfvf,
-				     int aura, u64 buf)
+static inline void otx2_aura_freeptr(void *dev, int aura, u64 buf)
 {
-	otx2_write128(buf, (u64)aura | BIT_ULL(63),
-		      otx2_get_regaddr(pfvf, NPA_LF_AURA_OP_FREE0));
+	struct otx2_nic *pfvf = dev;
+	void __iomem *addr = otx2_get_regaddr(pfvf, NPA_LF_AURA_OP_FREE0);
+
+	otx2_write128(buf, (u64)aura | BIT_ULL(63), addr);
 }
 
 static inline int otx2_get_pool_idx(struct otx2_nic *pfvf, int type, int idx)
@@ -645,6 +735,10 @@ void otx2_ctx_disable(struct mbox *mbox, int type, bool npa);
 int otx2_nix_config_bp(struct otx2_nic *pfvf, bool enable);
 void otx2_cleanup_rx_cqes(struct otx2_nic *pfvf, struct otx2_cq_queue *cq);
 void otx2_cleanup_tx_cqes(struct otx2_nic *pfvf, struct otx2_cq_queue *cq);
+int otx2_sq_aq_init(void *dev, u16 qidx, u16 sqb_aura);
+int cn10k_sq_aq_init(void *dev, u16 qidx, u16 sqb_aura);
+int otx2_alloc_buffer(struct otx2_nic *pfvf, struct otx2_cq_queue *cq,
+		      dma_addr_t *dma);
 
 /* RSS configuration APIs*/
 int otx2_rss_init(struct otx2_nic *pfvf);
@@ -704,5 +798,5 @@ int otx2_del_macfilter(struct net_device *netdev, const u8 *mac);
 int otx2_add_macfilter(struct net_device *netdev, const u8 *mac);
 int otx2_enable_rxvlan(struct otx2_nic *pf, bool enable);
 int otx2_install_rxvlan_offload_flow(struct otx2_nic *pfvf);
-
+u16 otx2_get_max_mtu(struct otx2_nic *pfvf);
 #endif /* OTX2_COMMON_H */

drivers/net/ethernet/marvell/octeontx2/nic/otx2_pf.c

@@ -22,10 +22,11 @@
 #include "otx2_txrx.h"
 #include "otx2_struct.h"
 #include "otx2_ptp.h"
+#include "cn10k.h"
 #include <rvu_trace.h>
 
-#define DRV_NAME	"octeontx2-nicpf"
-#define DRV_STRING	"Marvell OcteonTX2 NIC Physical Function Driver"
+#define DRV_NAME	"rvu_nicpf"
+#define DRV_STRING	"Marvell RVU NIC Physical Function Driver"
 
 /* Supported devices */
 static const struct pci_device_id otx2_pf_id_table[] = {
@@ -585,9 +586,17 @@ static int otx2_pfvf_mbox_init(struct otx2_nic *pf, int numvfs)
 	if (!pf->mbox_pfvf_wq)
 		return -ENOMEM;
 
-	base = readq((void __iomem *)((u64)pf->reg_base + RVU_PF_VF_BAR4_ADDR));
-	hwbase = ioremap_wc(base, MBOX_SIZE * pf->total_vfs);
+	/* On CN10K platform, PF <-> VF mailbox region follows after
+	 * PF <-> AF mailbox region.
+	 */
+	if (test_bit(CN10K_MBOX, &pf->hw.cap_flag))
+		base = pci_resource_start(pf->pdev, PCI_MBOX_BAR_NUM) +
+		       MBOX_SIZE;
+	else
+		base = readq((void __iomem *)((u64)pf->reg_base +
+					      RVU_PF_VF_BAR4_ADDR));
 
+	hwbase = ioremap_wc(base, MBOX_SIZE * pf->total_vfs);
 	if (!hwbase) {
 		err = -ENOMEM;
 		goto free_wq;
@@ -1042,7 +1051,7 @@ static int otx2_pfaf_mbox_init(struct otx2_nic *pf)
 	 * device memory to allow unaligned accesses.
 	 */
 	hwbase = ioremap_wc(pci_resource_start(pf->pdev, PCI_MBOX_BAR_NUM),
-			    pci_resource_len(pf->pdev, PCI_MBOX_BAR_NUM));
+			    MBOX_SIZE);
 	if (!hwbase) {
 		dev_err(pf->dev, "Unable to map PFAF mailbox region\n");
 		err = -ENOMEM;
@@ -1279,6 +1288,33 @@ static void otx2_free_sq_res(struct otx2_nic *pf)
 	}
 }
 
+static int otx2_get_rbuf_size(struct otx2_nic *pf, int mtu)
+{
+	int frame_size;
+	int total_size;
+	int rbuf_size;
+
+	/* The data transferred by NIX to memory consists of actual packet
+	 * plus additional data which has timestamp and/or EDSA/HIGIG2
+	 * headers if interface is configured in corresponding modes.
+	 * NIX transfers entire data using 6 segments/buffers and writes
+	 * a CQE_RX descriptor with those segment addresses. First segment
+	 * has additional data prepended to packet. Also software omits a
+	 * headroom of 128 bytes and sizeof(struct skb_shared_info) in
+	 * each segment. Hence the total size of memory needed
+	 * to receive a packet with 'mtu' is:
+	 * frame size =  mtu + additional data;
+	 * memory = frame_size + (headroom + struct skb_shared_info size) * 6;
+	 * each receive buffer size = memory / 6;
+	 */
+	frame_size = mtu + OTX2_ETH_HLEN + OTX2_HW_TIMESTAMP_LEN;
+	total_size = frame_size + (OTX2_HEAD_ROOM +
+		     OTX2_DATA_ALIGN(sizeof(struct skb_shared_info))) * 6;
+	rbuf_size = total_size / 6;
+
+	return ALIGN(rbuf_size, 2048);
+}
+
 static int otx2_init_hw_resources(struct otx2_nic *pf)
 {
 	struct nix_lf_free_req *free_req;
@@ -1295,9 +1331,9 @@ static int otx2_init_hw_resources(struct otx2_nic *pf)
 	hw->sqpool_cnt = hw->tx_queues;
 	hw->pool_cnt = hw->rqpool_cnt + hw->sqpool_cnt;
 
-	/* Get the size of receive buffers to allocate */
-	pf->rbsize = RCV_FRAG_LEN(OTX2_HW_TIMESTAMP_LEN + pf->netdev->mtu +
-				  OTX2_ETH_HLEN);
+	pf->max_frs = pf->netdev->mtu + OTX2_ETH_HLEN + OTX2_HW_TIMESTAMP_LEN;
+
+	pf->rbsize = otx2_get_rbuf_size(pf, pf->netdev->mtu);
 
 	mutex_lock(&mbox->lock);
 	/* NPA init */
@@ -1490,6 +1526,14 @@ int otx2_open(struct net_device *netdev)
 	if (!qset->rq)
 		goto err_free_mem;
 
+	if (test_bit(CN10K_LMTST, &pf->hw.cap_flag)) {
+		/* Reserve LMT lines for NPA AURA batch free */
+		pf->hw.npa_lmt_base = (__force u64 *)pf->hw.lmt_base;
+		/* Reserve LMT lines for NIX TX */
+		pf->hw.nix_lmt_base = (__force u64 *)((u64)pf->hw.npa_lmt_base +
+				      (NIX_LMTID_BASE * LMT_LINE_SIZE));
+	}
+
 	err = otx2_init_hw_resources(pf);
 	if (err)
 		goto err_free_mem;
@@ -2328,6 +2372,8 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 		goto err_free_netdev;
 	}
 
+	otx2_setup_dev_hw_settings(pf);
+
 	/* Init PF <=> AF mailbox stuff */
 	err = otx2_pfaf_mbox_init(pf);
 	if (err)
@@ -2353,7 +2399,9 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (err)
 		goto err_detach_rsrc;
 
-	otx2_setup_dev_hw_settings(pf);
+	err = cn10k_pf_lmtst_init(pf);
+	if (err)
+		goto err_detach_rsrc;
 
 	/* Assign default mac address */
 	otx2_get_mac_from_af(netdev);
@@ -2408,7 +2456,7 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 
 	/* MTU range: 64 - 9190 */
 	netdev->min_mtu = OTX2_MIN_MTU;
-	netdev->max_mtu = OTX2_MAX_MTU;
+	netdev->max_mtu = otx2_get_max_mtu(pf);
 
 	err = register_netdev(netdev);
 	if (err) {
@@ -2438,6 +2486,8 @@ static int otx2_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 err_ptp_destroy:
 	otx2_ptp_destroy(pf);
 err_detach_rsrc:
+	if (hw->lmt_base)
+		iounmap(hw->lmt_base);
 	otx2_detach_resources(&pf->mbox);
 err_disable_mbox_intr:
 	otx2_disable_mbox_intr(pf);
@@ -2597,6 +2647,9 @@ static void otx2_remove(struct pci_dev *pdev)
 	otx2_ptp_destroy(pf);
 	otx2_mcam_flow_del(pf);
 	otx2_detach_resources(&pf->mbox);
+	if (pf->hw.lmt_base)
+		iounmap(pf->hw.lmt_base);
+
 	otx2_disable_mbox_intr(pf);
 	otx2_pfaf_mbox_destroy(pf);
 	pci_free_irq_vectors(pf->pdev);

drivers/net/ethernet/marvell/octeontx2/nic/otx2_reg.h

@@ -44,6 +44,8 @@
 #define RVU_PF_MSIX_VECX_ADDR(a)            (0x000 | (a) << 4)
 #define RVU_PF_MSIX_VECX_CTL(a)             (0x008 | (a) << 4)
 #define RVU_PF_MSIX_PBAX(a)                 (0xF0000 | (a) << 3)
+#define RVU_PF_VF_MBOX_ADDR                 (0xC40)
+#define RVU_PF_LMTLINE_ADDR                 (0xC48)
 
 /* RVU VF registers */
 #define	RVU_VF_VFPF_MBOX0		    (0x00000)
@@ -57,6 +59,7 @@
 #define	RVU_VF_MSIX_VECX_ADDR(a)	    (0x000 | (a) << 4)
 #define	RVU_VF_MSIX_VECX_CTL(a)		    (0x008 | (a) << 4)
 #define	RVU_VF_MSIX_PBAX(a)		    (0xF0000 | (a) << 3)
+#define RVU_VF_MBOX_REGION                  (0xC0000)
 
 #define RVU_FUNC_BLKADDR_SHIFT		20
 #define RVU_FUNC_BLKADDR_MASK		0x1FULL
@@ -91,6 +94,7 @@
 #define NPA_LF_QINTX_INT_W1S(a)         (NPA_LFBASE | 0x318 | (a) << 12)
 #define NPA_LF_QINTX_ENA_W1S(a)         (NPA_LFBASE | 0x320 | (a) << 12)
 #define NPA_LF_QINTX_ENA_W1C(a)         (NPA_LFBASE | 0x330 | (a) << 12)
+#define NPA_LF_AURA_BATCH_FREE0         (NPA_LFBASE | 0x400)
 
 /* NIX LF registers */
 #define	NIX_LFBASE			(BLKTYPE_NIX << RVU_FUNC_BLKADDR_SHIFT)

drivers/net/ethernet/marvell/octeontx2/nic/otx2_struct.h

@@ -142,7 +142,9 @@ struct nix_rx_parse_s {
 	u64 vtag0_ptr    : 8; /* W5 */
 	u64 vtag1_ptr    : 8;
 	u64 flow_key_alg : 5;
-	u64 rsvd_383_341 : 43;
+	u64 rsvd_359_341 : 19;
+	u64 color	 : 2;
+	u64 rsvd_383_362 : 22;
 	u64 rsvd_447_384;     /* W6 */
 };
 
@@ -218,7 +220,8 @@ struct nix_sqe_ext_s {
 	u64 vlan1_ins_tci : 16;
 	u64 vlan0_ins_ena : 1;
 	u64 vlan1_ins_ena : 1;
-	u64 rsvd_127_114  : 14;
+	u64 init_color    : 2;
+	u64 rsvd_127_116  : 12;
 };
 
 struct nix_sqe_sg_s {
@@ -237,7 +240,8 @@ struct nix_sqe_sg_s {
 /* NIX send memory subdescriptor structure */
 struct nix_sqe_mem_s {
 	u64 offset        : 16; /* W0 */
-	u64 rsvd_52_16    : 37;
+	u64 rsvd_51_16    : 36;
+	u64 per_lso_seg   : 1;
 	u64 wmem          : 1;
 	u64 dsz           : 2;
 	u64 alg           : 4;

drivers/net/ethernet/marvell/octeontx2/nic/otx2_txrx.c

@@ -17,6 +17,7 @@
 #include "otx2_struct.h"
 #include "otx2_txrx.h"
 #include "otx2_ptp.h"
+#include "cn10k.h"
 
 #define CQE_ADDR(CQ, idx) ((CQ)->cqe_base + ((CQ)->cqe_size * (idx)))
 
@@ -199,7 +200,8 @@ static void otx2_free_rcv_seg(struct otx2_nic *pfvf, struct nix_cqe_rx_s *cqe,
 		sg = (struct nix_rx_sg_s *)start;
 		seg_addr = &sg->seg_addr;
 		for (seg = 0; seg < sg->segs; seg++, seg_addr++)
-			otx2_aura_freeptr(pfvf, qidx, *seg_addr & ~0x07ULL);
+			pfvf->hw_ops->aura_freeptr(pfvf, qidx,
+						   *seg_addr & ~0x07ULL);
 		start += sizeof(*sg);
 	}
 }
@@ -255,12 +257,11 @@ static bool otx2_check_rcv_errors(struct otx2_nic *pfvf,
 		/* For now ignore all the NPC parser errors and
 		 * pass the packets to stack.
 		 */
-		if (cqe->sg.segs == 1)
-			return false;
+		return false;
 	}
 
 	/* If RXALL is enabled pass on packets to stack. */
-	if (cqe->sg.segs == 1 && (pfvf->netdev->features & NETIF_F_RXALL))
+	if (pfvf->netdev->features & NETIF_F_RXALL)
 		return false;
 
 	/* Free buffer back to pool */
@@ -275,9 +276,14 @@ static void otx2_rcv_pkt_handler(struct otx2_nic *pfvf,
 				 struct nix_cqe_rx_s *cqe)
 {
 	struct nix_rx_parse_s *parse = &cqe->parse;
+	struct nix_rx_sg_s *sg = &cqe->sg;
 	struct sk_buff *skb = NULL;
+	void *end, *start;
+	u64 *seg_addr;
+	u16 *seg_size;
+	int seg;
 
-	if (unlikely(parse->errlev || parse->errcode || cqe->sg.segs > 1)) {
+	if (unlikely(parse->errlev || parse->errcode)) {
 		if (otx2_check_rcv_errors(pfvf, cqe, cq->cq_idx))
 			return;
 	}
@@ -286,9 +292,19 @@ static void otx2_rcv_pkt_handler(struct otx2_nic *pfvf,
 	if (unlikely(!skb))
 		return;
 
-	otx2_skb_add_frag(pfvf, skb, cqe->sg.seg_addr, cqe->sg.seg_size, parse);
-	cq->pool_ptrs++;
-
+	start = (void *)sg;
+	end = start + ((cqe->parse.desc_sizem1 + 1) * 16);
+	while (start < end) {
+		sg = (struct nix_rx_sg_s *)start;
+		seg_addr = &sg->seg_addr;
+		seg_size = (void *)sg;
+		for (seg = 0; seg < sg->segs; seg++, seg_addr++) {
+			otx2_skb_add_frag(pfvf, skb, *seg_addr, seg_size[seg],
+					  parse);
+			cq->pool_ptrs++;
+		}
+		start += sizeof(*sg);
+	}
 	otx2_set_rxhash(pfvf, cqe, skb);
 
 	skb_record_rx_queue(skb, cq->cq_idx);
@@ -304,7 +320,6 @@ static int otx2_rx_napi_handler(struct otx2_nic *pfvf,
 {
 	struct nix_cqe_rx_s *cqe;
 	int processed_cqe = 0;
-	dma_addr_t bufptr;
 
 	while (likely(processed_cqe < budget)) {
 		cqe = (struct nix_cqe_rx_s *)CQE_ADDR(cq, cq->cq_head);
@@ -330,28 +345,23 @@ static int otx2_rx_napi_handler(struct otx2_nic *pfvf,
 
 	if (unlikely(!cq->pool_ptrs))
 		return 0;
-
 	/* Refill pool with new buffers */
+	pfvf->hw_ops->refill_pool_ptrs(pfvf, cq);
+
+	return processed_cqe;
+}
+
+void otx2_refill_pool_ptrs(void *dev, struct otx2_cq_queue *cq)
+{
+	struct otx2_nic *pfvf = dev;
+	dma_addr_t bufptr;
+
 	while (cq->pool_ptrs) {
-		if (unlikely(__otx2_alloc_rbuf(pfvf, cq->rbpool, &bufptr))) {
-			struct refill_work *work;
-			struct delayed_work *dwork;
-
-			work = &pfvf->refill_wrk[cq->cq_idx];
-			dwork = &work->pool_refill_work;
-			/* Schedule a task if no other task is running */
-			if (!cq->refill_task_sched) {
-				cq->refill_task_sched = true;
-				schedule_delayed_work(dwork,
-						      msecs_to_jiffies(100));
-			}
+		if (otx2_alloc_buffer(pfvf, cq, &bufptr))
 			break;
-		}
 		otx2_aura_freeptr(pfvf, cq->cq_idx, bufptr + OTX2_HEAD_ROOM);
 		cq->pool_ptrs--;
 	}
-
-	return processed_cqe;
 }
 
 static int otx2_tx_napi_handler(struct otx2_nic *pfvf,
@@ -438,7 +448,8 @@ int otx2_napi_handler(struct napi_struct *napi, int budget)
 	return workdone;
 }
 
-static void otx2_sqe_flush(struct otx2_snd_queue *sq, int size)
+void otx2_sqe_flush(void *dev, struct otx2_snd_queue *sq,
+		    int size, int qidx)
 {
 	u64 status;
 
@@ -796,7 +807,7 @@ static void otx2_sq_append_tso(struct otx2_nic *pfvf, struct otx2_snd_queue *sq,
 		sqe_hdr->sizem1 = (offset / 16) - 1;
 
 		/* Flush SQE to HW */
-		otx2_sqe_flush(sq, offset);
+		pfvf->hw_ops->sqe_flush(pfvf, sq, offset, qidx);
 	}
 }
 
@@ -805,8 +816,6 @@ static bool is_hw_tso_supported(struct otx2_nic *pfvf,
 {
 	int payload_len, last_seg_size;
 
-	if (!pfvf->hw.hw_tso)
-		return false;
 
 	/* HW has an issue due to which when the payload of the last LSO
 	 * segment is shorter than 16 bytes, some header fields may not
@@ -820,6 +829,9 @@ static bool is_hw_tso_supported(struct otx2_nic *pfvf,
 	if (last_seg_size && last_seg_size < 16)
 		return false;
 
+	if (!test_bit(HW_TSO, &pfvf->hw.cap_flag))
+		return false;
+
 	return true;
 }
 
@@ -914,7 +926,7 @@ bool otx2_sq_append_skb(struct net_device *netdev, struct otx2_snd_queue *sq,
 	netdev_tx_sent_queue(txq, skb->len);
 
 	/* Flush SQE to HW */
-	otx2_sqe_flush(sq, offset);
+	pfvf->hw_ops->sqe_flush(pfvf, sq, offset, qidx);
 
 	return true;
 }

drivers/net/ethernet/marvell/octeontx2/nic/otx2_txrx.h

@@ -24,7 +24,6 @@
 
 #define	OTX2_ETH_HLEN		(VLAN_ETH_HLEN + VLAN_HLEN)
 #define	OTX2_MIN_MTU		64
-#define	OTX2_MAX_MTU		(9212 - OTX2_ETH_HLEN)
 
 #define OTX2_MAX_GSO_SEGS	255
 #define OTX2_MAX_FRAGS_IN_SQE	9
@@ -114,6 +113,7 @@ struct otx2_cq_poll {
 struct otx2_pool {
 	struct qmem		*stack;
 	struct qmem		*fc_addr;
+	u64			*lmt_addr;
 	u16			rbsize;
 };
 
@@ -156,4 +156,10 @@ static inline u64 otx2_iova_to_phys(void *iommu_domain, dma_addr_t dma_addr)
 int otx2_napi_handler(struct napi_struct *napi, int budget);
 bool otx2_sq_append_skb(struct net_device *netdev, struct otx2_snd_queue *sq,
 			struct sk_buff *skb, u16 qidx);
+void cn10k_sqe_flush(void *dev, struct otx2_snd_queue *sq,
+		     int size, int qidx);
+void otx2_sqe_flush(void *dev, struct otx2_snd_queue *sq,
+		    int size, int qidx);
+void otx2_refill_pool_ptrs(void *dev, struct otx2_cq_queue *cq);
+void cn10k_refill_pool_ptrs(void *dev, struct otx2_cq_queue *cq);
 #endif /* OTX2_TXRX_H */

include/linux/soc/marvell/octeontx2/asm.h

@@ -22,8 +22,16 @@
 			 : [rs]"r" (ioaddr));           \
 	(result);                                       \
 })
+#define cn10k_lmt_flush(val, addr)			\
+({							\
+	__asm__ volatile(".cpu  generic+lse\n"		\
+			 "steor %x[rf],[%[rs]]"		\
+			 : [rf]"+r"(val)		\
+			 : [rs]"r"(addr));		\
+})
 #else
 #define otx2_lmt_flush(ioaddr)          ({ 0; })
+#define cn10k_lmt_flush(val, addr)	({ addr = val; })
 #endif
 
 #endif /* __SOC_OTX2_ASM_H */