be2net: Add functionality to support RoCE driver

- Increase MSI-X vectors by 5 for RoCE traffic.
- Add macro to check roce support on a device.
- Add device-specific doorbell and MSI-X vector fields shared with NIC
  functionality.
- Provide RoCE driver registration and deregistration functions.
- Add support functions which will be invoked on adapter add/remove
  and port up/down events.
- Traverse through the list of adapters to invoke callback functions.
Signed-off-by: Parav Pandit <parav.pandit@emulex.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Roland Dreier <roland@purestorage.com>

drivers/net/ethernet/emulex/benet/Makefile

@@ -4,4 +4,4 @@
 
 obj-$(CONFIG_BE2NET) += be2net.o
 
-be2net-y :=  be_main.o be_cmds.o be_ethtool.o
+be2net-y :=  be_main.o be_cmds.o be_ethtool.o be_roce.o

drivers/net/ethernet/emulex/benet/be.h

@@ -32,6 +32,7 @@
 #include <linux/u64_stats_sync.h>
 
 #include "be_hw.h"
+#include "be_roce.h"
 
 #define DRV_VER			"4.2.116u"
 #define DRV_NAME		"be2net"
@@ -102,7 +103,8 @@ static inline char *nic_name(struct pci_dev *pdev)
 #define MAX_RX_QS		(MAX_RSS_QS + 1) /* RSS qs + 1 def Rx */
 
 #define MAX_TX_QS		8
-#define MAX_MSIX_VECTORS	MAX_RSS_QS
+#define MAX_ROCE_EQS		5
+#define MAX_MSIX_VECTORS	(MAX_RSS_QS + MAX_ROCE_EQS) /* RSS qs + RoCE */
 #define BE_TX_BUDGET		256
 #define BE_NAPI_WEIGHT		64
 #define MAX_RX_POST		BE_NAPI_WEIGHT /* Frags posted at a time */
@@ -382,6 +384,17 @@ struct be_adapter {
 	u8 transceiver;
 	u8 autoneg;
 	u8 generation;		/* BladeEngine ASIC generation */
+	u32 if_type;
+	struct {
+		u8 __iomem *base;	/* Door Bell */
+		u32 size;
+		u32 total_size;
+		u64 io_addr;
+	} roce_db;
+	u32 num_msix_roce_vec;
+	struct ocrdma_dev *ocrdma_dev;
+	struct list_head entry;
+
 	u32 flash_status;
 	struct completion flash_compl;
 
@@ -413,6 +426,10 @@ struct be_adapter {
 #define lancer_chip(adapter)	((adapter->pdev->device == OC_DEVICE_ID3) || \
 				 (adapter->pdev->device == OC_DEVICE_ID4))
 
+#define be_roce_supported(adapter) ((adapter->if_type == SLI_INTF_TYPE_3 || \
+				adapter->sli_family == SKYHAWK_SLI_FAMILY) && \
+				(adapter->function_mode & RDMA_ENABLED))
+
 extern const struct ethtool_ops be_ethtool_ops;
 
 #define msix_enabled(adapter)		(adapter->num_msix_vec > 0)
@@ -577,10 +594,29 @@ static inline bool be_is_wol_excluded(struct be_adapter *adapter)
 	}
 }
 
+static inline bool be_type_2_3(struct be_adapter *adapter)
+{
+	return (adapter->if_type == SLI_INTF_TYPE_2 ||
+		adapter->if_type == SLI_INTF_TYPE_3) ? true : false;
+}
+
 extern void be_cq_notify(struct be_adapter *adapter, u16 qid, bool arm,
 		u16 num_popped);
 extern void be_link_status_update(struct be_adapter *adapter, u8 link_status);
 extern void be_parse_stats(struct be_adapter *adapter);
 extern int be_load_fw(struct be_adapter *adapter, u8 *func);
 extern bool be_is_wol_supported(struct be_adapter *adapter);
+
+/*
+ * internal function to initialize-cleanup roce device.
+ */
+extern void be_roce_dev_add(struct be_adapter *);
+extern void be_roce_dev_remove(struct be_adapter *);
+
+/*
+ * internal function to open-close roce device during ifup-ifdown.
+ */
+extern void be_roce_dev_open(struct be_adapter *);
+extern void be_roce_dev_close(struct be_adapter *);
+
 #endif				/* BE_H */

drivers/net/ethernet/emulex/benet/be_cmds.h

@@ -1056,6 +1056,7 @@ struct be_cmd_resp_modify_eq_delay {
 /* The HW can come up in either of the following multi-channel modes
  * based on the skew/IPL.
  */
+#define RDMA_ENABLED				0x4
 #define FLEX10_MODE				0x400
 #define VNIC_MODE				0x20000
 #define UMC_ENABLED				0x1000000

drivers/net/ethernet/emulex/benet/be_hw.h

@@ -98,11 +98,13 @@
 #define SLI_INTF_REV_SHIFT			4
 #define SLI_INTF_FT_MASK			0x00000001
 
+#define SLI_INTF_TYPE_2		2
+#define SLI_INTF_TYPE_3		3
 
 /* SLI family */
 #define BE_SLI_FAMILY		0x0
 #define LANCER_A0_SLI_FAMILY	0xA
-
+#define SKYHAWK_SLI_FAMILY      0x2
 
 /********* ISR0 Register offset **********/
 #define CEV_ISR0_OFFSET 			0xC18

drivers/net/ethernet/emulex/benet/be_main.c

@@ -2103,10 +2103,17 @@ static uint be_num_rss_want(struct be_adapter *adapter)
 static void be_msix_enable(struct be_adapter *adapter)
 {
 #define BE_MIN_MSIX_VECTORS		1
-	int i, status, num_vec;
+	int i, status, num_vec, num_roce_vec = 0;
 
 	/* If RSS queues are not used, need a vec for default RX Q */
 	num_vec = min(be_num_rss_want(adapter), num_online_cpus());
+	if (be_roce_supported(adapter)) {
+		num_roce_vec = min_t(u32, MAX_ROCE_MSIX_VECTORS,
+					(num_online_cpus() + 1));
+		num_roce_vec = min(num_roce_vec, MAX_ROCE_EQS);
+		num_vec += num_roce_vec;
+		num_vec = min(num_vec, MAX_MSIX_VECTORS);
+	}
 	num_vec = max(num_vec, BE_MIN_MSIX_VECTORS);
 
 	for (i = 0; i < num_vec; i++)
@@ -2123,7 +2130,17 @@ static void be_msix_enable(struct be_adapter *adapter)
 	}
 	return;
 done:
-	adapter->num_msix_vec = num_vec;
+	if (be_roce_supported(adapter)) {
+		if (num_vec > num_roce_vec) {
+			adapter->num_msix_vec = num_vec - num_roce_vec;
+			adapter->num_msix_roce_vec =
+				num_vec - adapter->num_msix_vec;
+		} else {
+			adapter->num_msix_vec = num_vec;
+			adapter->num_msix_roce_vec = 0;
+		}
+	} else
+		adapter->num_msix_vec = num_vec;
 	return;
 }
 
@@ -2282,6 +2299,8 @@ static int be_close(struct net_device *netdev)
 	struct be_eq_obj *eqo;
 	int i;
 
+	be_roce_dev_close(adapter);
+
 	be_async_mcc_disable(adapter);
 
 	if (!lancer_chip(adapter))
@@ -2390,6 +2409,7 @@ static int be_open(struct net_device *netdev)
 	if (!status)
 		be_link_status_update(adapter, link_status);
 
+	be_roce_dev_open(adapter);
 	return 0;
 err:
 	be_close(adapter->netdev);
@@ -3122,6 +3142,24 @@ static void be_unmap_pci_bars(struct be_adapter *adapter)
 		iounmap(adapter->csr);
 	if (adapter->db)
 		iounmap(adapter->db);
+	if (adapter->roce_db.base)
+		pci_iounmap(adapter->pdev, adapter->roce_db.base);
+}
+
+static int lancer_roce_map_pci_bars(struct be_adapter *adapter)
+{
+	struct pci_dev *pdev = adapter->pdev;
+	u8 __iomem *addr;
+
+	addr = pci_iomap(pdev, 2, 0);
+	if (addr == NULL)
+		return -ENOMEM;
+
+	adapter->roce_db.base = addr;
+	adapter->roce_db.io_addr = pci_resource_start(pdev, 2);
+	adapter->roce_db.size = 8192;
+	adapter->roce_db.total_size = pci_resource_len(pdev, 2);
+	return 0;
 }
 
 static int be_map_pci_bars(struct be_adapter *adapter)
@@ -3130,11 +3168,18 @@ static int be_map_pci_bars(struct be_adapter *adapter)
 	int db_reg;
 
 	if (lancer_chip(adapter)) {
-		addr = ioremap_nocache(pci_resource_start(adapter->pdev, 0),
-			pci_resource_len(adapter->pdev, 0));
-		if (addr == NULL)
-			return -ENOMEM;
-		adapter->db = addr;
+		if (be_type_2_3(adapter)) {
+			addr = ioremap_nocache(
+					pci_resource_start(adapter->pdev, 0),
+					pci_resource_len(adapter->pdev, 0));
+			if (addr == NULL)
+				return -ENOMEM;
+			adapter->db = addr;
+		}
+		if (adapter->if_type == SLI_INTF_TYPE_3) {
+			if (lancer_roce_map_pci_bars(adapter))
+				goto pci_map_err;
+		}
 		return 0;
 	}
 
@@ -3159,14 +3204,19 @@ static int be_map_pci_bars(struct be_adapter *adapter)
 	if (addr == NULL)
 		goto pci_map_err;
 	adapter->db = addr;
-
+	if (adapter->sli_family == SKYHAWK_SLI_FAMILY) {
+		adapter->roce_db.size = 4096;
+		adapter->roce_db.io_addr =
+				pci_resource_start(adapter->pdev, db_reg);
+		adapter->roce_db.total_size =
+				pci_resource_len(adapter->pdev, db_reg);
+	}
 	return 0;
 pci_map_err:
 	be_unmap_pci_bars(adapter);
 	return -ENOMEM;
 }
 
-
 static void be_ctrl_cleanup(struct be_adapter *adapter)
 {
 	struct be_dma_mem *mem = &adapter->mbox_mem_alloced;
@@ -3272,6 +3322,8 @@ static void __devexit be_remove(struct pci_dev *pdev)
 	if (!adapter)
 		return;
 
+	be_roce_dev_remove(adapter);
+
 	unregister_netdev(adapter->netdev);
 
 	be_clear(adapter);
@@ -3350,17 +3402,27 @@ static int be_dev_family_check(struct be_adapter *adapter)
 		break;
 	case BE_DEVICE_ID2:
 	case OC_DEVICE_ID2:
-	case OC_DEVICE_ID5:
 		adapter->generation = BE_GEN3;
 		break;
 	case OC_DEVICE_ID3:
 	case OC_DEVICE_ID4:
 		pci_read_config_dword(pdev, SLI_INTF_REG_OFFSET, &sli_intf);
+		adapter->if_type = (sli_intf & SLI_INTF_IF_TYPE_MASK) >>
+						SLI_INTF_IF_TYPE_SHIFT;
 		if_type = (sli_intf & SLI_INTF_IF_TYPE_MASK) >>
 						SLI_INTF_IF_TYPE_SHIFT;
-
 		if (((sli_intf & SLI_INTF_VALID_MASK) != SLI_INTF_VALID) ||
-			if_type != 0x02) {
+			!be_type_2_3(adapter)) {
+			dev_err(&pdev->dev, "SLI_INTF reg val is not valid\n");
+			return -EINVAL;
+		}
+		adapter->sli_family = ((sli_intf & SLI_INTF_FAMILY_MASK) >>
+					 SLI_INTF_FAMILY_SHIFT);
+		adapter->generation = BE_GEN3;
+		break;
+	case OC_DEVICE_ID5:
+		pci_read_config_dword(pdev, SLI_INTF_REG_OFFSET, &sli_intf);
+		if ((sli_intf & SLI_INTF_VALID_MASK) != SLI_INTF_VALID) {
 			dev_err(&pdev->dev, "SLI_INTF reg val is not valid\n");
 			return -EINVAL;
 		}
@@ -3620,6 +3682,8 @@ static int __devinit be_probe(struct pci_dev *pdev,
 	if (status != 0)
 		goto unsetup;
 
+	be_roce_dev_add(adapter);
+
 	dev_info(&pdev->dev, "%s: %s port %d\n", netdev->name, nic_name(pdev),
 		adapter->port_num);
 

drivers/net/ethernet/emulex/benet/be_roce.c

+/*
+ * Copyright (C) 2005 - 2011 Emulex
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation. The full GNU General
+ * Public License is included in this distribution in the file called COPYING.
+ *
+ * Contact Information:
+ * linux-drivers@emulex.com
+ *
+ * Emulex
+ * 3333 Susan Street
+ * Costa Mesa, CA 92626
+ */
+
+#include <linux/mutex.h>
+#include <linux/list.h>
+#include <linux/netdevice.h>
+#include <linux/module.h>
+
+#include "be.h"
+#include "be_cmds.h"
+
+static struct ocrdma_driver *ocrdma_drv;
+static LIST_HEAD(be_adapter_list);
+static DEFINE_MUTEX(be_adapter_list_lock);
+
+static void _be_roce_dev_add(struct be_adapter *adapter)
+{
+	struct be_dev_info dev_info;
+	int i, num_vec;
+	struct pci_dev *pdev = adapter->pdev;
+
+	if (!ocrdma_drv)
+		return;
+	if (pdev->device == OC_DEVICE_ID5) {
+		/* only msix is supported on these devices */
+		if (!msix_enabled(adapter))
+			return;
+		/* DPP region address and length */
+		dev_info.dpp_unmapped_addr = pci_resource_start(pdev, 2);
+		dev_info.dpp_unmapped_len = pci_resource_len(pdev, 2);
+	} else {
+		dev_info.dpp_unmapped_addr = 0;
+		dev_info.dpp_unmapped_len = 0;
+	}
+	dev_info.pdev = adapter->pdev;
+	if (adapter->sli_family == SKYHAWK_SLI_FAMILY)
+		dev_info.db = adapter->db;
+	else
+		dev_info.db = adapter->roce_db.base;
+	dev_info.unmapped_db = adapter->roce_db.io_addr;
+	dev_info.db_page_size = adapter->roce_db.size;
+	dev_info.db_total_size = adapter->roce_db.total_size;
+	dev_info.netdev = adapter->netdev;
+	memcpy(dev_info.mac_addr, adapter->netdev->dev_addr, ETH_ALEN);
+	dev_info.dev_family = adapter->sli_family;
+	if (msix_enabled(adapter)) {
+		/* provide all the vectors, so that EQ creation response
+		 * can decide which one to use.
+		 */
+		num_vec = adapter->num_msix_vec + adapter->num_msix_roce_vec;
+		dev_info.intr_mode = BE_INTERRUPT_MODE_MSIX;
+		dev_info.msix.num_vectors = min(num_vec, MAX_ROCE_MSIX_VECTORS);
+		/* provide start index of the vector,
+		 * so in case of linear usage,
+		 * it can use the base as starting point.
+		 */
+		dev_info.msix.start_vector = adapter->num_evt_qs;
+		for (i = 0; i < dev_info.msix.num_vectors; i++) {
+			dev_info.msix.vector_list[i] =
+			    adapter->msix_entries[i].vector;
+		}
+	} else {
+		dev_info.msix.num_vectors = 0;
+		dev_info.intr_mode = BE_INTERRUPT_MODE_INTX;
+	}
+	adapter->ocrdma_dev = ocrdma_drv->add(&dev_info);
+}
+
+void be_roce_dev_add(struct be_adapter *adapter)
+{
+	if (be_roce_supported(adapter)) {
+		INIT_LIST_HEAD(&adapter->entry);
+		mutex_lock(&be_adapter_list_lock);
+		list_add_tail(&adapter->entry, &be_adapter_list);
+
+		/* invoke add() routine of roce driver only if
+		 * valid driver registered with add method and add() is not yet
+		 * invoked on a given adapter.
+		 */
+		_be_roce_dev_add(adapter);
+		mutex_unlock(&be_adapter_list_lock);
+	}
+}
+
+void _be_roce_dev_remove(struct be_adapter *adapter)
+{
+	if (ocrdma_drv && ocrdma_drv->remove && adapter->ocrdma_dev)
+		ocrdma_drv->remove(adapter->ocrdma_dev);
+	adapter->ocrdma_dev = NULL;
+}
+
+void be_roce_dev_remove(struct be_adapter *adapter)
+{
+	if (be_roce_supported(adapter)) {
+		mutex_lock(&be_adapter_list_lock);
+		_be_roce_dev_remove(adapter);
+		list_del(&adapter->entry);
+		mutex_unlock(&be_adapter_list_lock);
+	}
+}
+
+void _be_roce_dev_open(struct be_adapter *adapter)
+{
+	if (ocrdma_drv && adapter->ocrdma_dev &&
+	    ocrdma_drv->state_change_handler)
+		ocrdma_drv->state_change_handler(adapter->ocrdma_dev, 0);
+}
+
+void be_roce_dev_open(struct be_adapter *adapter)
+{
+	if (be_roce_supported(adapter)) {
+		mutex_lock(&be_adapter_list_lock);
+		_be_roce_dev_open(adapter);
+		mutex_unlock(&be_adapter_list_lock);
+	}
+}
+
+void _be_roce_dev_close(struct be_adapter *adapter)
+{
+	if (ocrdma_drv && adapter->ocrdma_dev &&
+	    ocrdma_drv->state_change_handler)
+		ocrdma_drv->state_change_handler(adapter->ocrdma_dev, 1);
+}
+
+void be_roce_dev_close(struct be_adapter *adapter)
+{
+	if (be_roce_supported(adapter)) {
+		mutex_lock(&be_adapter_list_lock);
+		_be_roce_dev_close(adapter);
+		mutex_unlock(&be_adapter_list_lock);
+	}
+}
+
+int be_roce_register_driver(struct ocrdma_driver *drv)
+{
+	struct be_adapter *dev;
+
+	mutex_lock(&be_adapter_list_lock);
+	if (ocrdma_drv) {
+		mutex_unlock(&be_adapter_list_lock);
+		return -EINVAL;
+	}
+	ocrdma_drv = drv;
+	list_for_each_entry(dev, &be_adapter_list, entry) {
+		struct net_device *netdev;
+		_be_roce_dev_add(dev);
+		netdev = dev->netdev;
+		if (netif_running(netdev) && netif_oper_up(netdev))
+			_be_roce_dev_open(dev);
+	}
+	mutex_unlock(&be_adapter_list_lock);
+	return 0;
+}
+EXPORT_SYMBOL(be_roce_register_driver);
+
+void be_roce_unregister_driver(struct ocrdma_driver *drv)
+{
+	struct be_adapter *dev;
+
+	mutex_lock(&be_adapter_list_lock);
+	list_for_each_entry(dev, &be_adapter_list, entry) {
+		if (dev->ocrdma_dev)
+			_be_roce_dev_remove(dev);
+	}
+	ocrdma_drv = NULL;
+	mutex_unlock(&be_adapter_list_lock);
+}
+EXPORT_SYMBOL(be_roce_unregister_driver);

drivers/net/ethernet/emulex/benet/be_roce.h

+/*
+ * Copyright (C) 2005 - 2011 Emulex
+ * All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License version 2
+ * as published by the Free Software Foundation. The full GNU General
+ * Public License is included in this distribution in the file called COPYING.
+ *
+ * Contact Information:
+ * linux-drivers@emulex.com
+ *
+ * Emulex
+ * 3333 Susan Street
+ * Costa Mesa, CA 92626
+ */
+
+#ifndef BE_ROCE_H
+#define BE_ROCE_H
+
+#include <linux/pci.h>
+#include <linux/netdevice.h>
+
+struct ocrdma_dev;
+
+enum be_interrupt_mode {
+	BE_INTERRUPT_MODE_MSIX	= 0,
+	BE_INTERRUPT_MODE_INTX	= 1,
+	BE_INTERRUPT_MODE_MSI	= 2,
+};
+
+#define MAX_ROCE_MSIX_VECTORS   16
+struct be_dev_info {
+	u8 __iomem *db;
+	u64 unmapped_db;
+	u32 db_page_size;
+	u32 db_total_size;
+	u64 dpp_unmapped_addr;
+	u32 dpp_unmapped_len;
+	struct pci_dev *pdev;
+	struct net_device *netdev;
+	u8 mac_addr[ETH_ALEN];
+	u32 dev_family;
+	enum be_interrupt_mode intr_mode;
+	struct {
+		int num_vectors;
+		int start_vector;
+		u32 vector_list[MAX_ROCE_MSIX_VECTORS];
+	} msix;
+};
+
+/* ocrdma driver register's the callback functions with nic driver. */
+struct ocrdma_driver {
+	unsigned char name[32];
+	struct ocrdma_dev *(*add) (struct be_dev_info *dev_info);
+	void (*remove) (struct ocrdma_dev *);
+	void (*state_change_handler) (struct ocrdma_dev *, u32 new_state);
+};
+
+enum {
+	BE_DEV_UP	= 0,
+	BE_DEV_DOWN	= 1
+};
+
+/* APIs for RoCE driver to register callback handlers,
+ * which will be invoked when device is added, removed, ifup, ifdown
+ */
+int be_roce_register_driver(struct ocrdma_driver *drv);
+void be_roce_unregister_driver(struct ocrdma_driver *drv);
+
+/* API for RoCE driver to issue mailbox commands */
+int be_roce_mcc_cmd(void *netdev_handle, void *wrb_payload,
+		    int wrb_payload_size, u16 *cmd_status, u16 *ext_status);
+
+#endif /* BE_ROCE_H */