[PATCH net-next 0/13] bnx2x: patch series

[PATCH net-next 0/13] bnx2x: patch series

[Yuval Mintz]

Hi Dave,

This patch series contains several changes:

 1. SR-IOV related patches - allow setting and getting the number of VFs
    via sysfs, configuring and querying VFs via iproute2 callbacks and
    properly implement for VFs only the ethtool ops which they can safely
    perform.

 2. Advance the FW version used by the driver. This allows 57840 boards to
    use FCoE, as well as fixing various issues.

 3. Several small link enhancements.

 4. Several fixes of rare corner-cases.

Please consider applying this patch-series to 'net-next'.

Thanks,
Yuval Mintz

[PATCH net-next 01/13] bnx2x: fix vlan-mac memory leak

[Yuval Mintz]

Release (previously leaking) memory when elements are removed from pending
execution lists in the bnx2x driver.

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
index 7306416..9f2637c 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
@@ -1854,6 +1854,7 @@ static int bnx2x_vlan_mac_del_all(struct bnx2x *bp,
 				return rc;
 			}
 			list_del(&exeq_pos->link);
+			bnx2x_exe_queue_free_elem(bp, exeq_pos);
 		}
 	}
 
-- 
1.7.7.2

[PATCH net-next 02/13] bnx2x: Set ethtool ops for vfs

[Yuval Mintz]

From: Ariel Elior <ariele@broadcom.com>

Virtual functions don't have access to HW registers, therefore most ethtool ops
are forbidden to them. Instead of checking in each op whether the device being
driven is a virtual function or a physical function, this patch creates a
separate ethtool ops struct for virtual functions and uses it to initialize
the ethtool ops of the driver in case it is driving a virtual function device.

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x.h        |    2 +-
 .../net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c    |   29 ++++++++++++++++++-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c   |    2 +-
 3 files changed, 29 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 9577cce..8ddc78b 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -2285,7 +2285,7 @@ static const u32 dmae_reg_go_c[] = {
 	DMAE_REG_GO_C12, DMAE_REG_GO_C13, DMAE_REG_GO_C14, DMAE_REG_GO_C15
 };
 
-void bnx2x_set_ethtool_ops(struct net_device *netdev);
+void bnx2x_set_ethtool_ops(struct bnx2x *bp, struct net_device *netdev);
 void bnx2x_notify_link_changed(struct bnx2x *bp);
 
 #define BNX2X_MF_SD_PROTOCOL(bp) \
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
index edfa67a..324d691 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_ethtool.c
@@ -3232,7 +3232,32 @@ static const struct ethtool_ops bnx2x_ethtool_ops = {
 	.get_ts_info		= ethtool_op_get_ts_info,
 };
 
-void bnx2x_set_ethtool_ops(struct net_device *netdev)
+static const struct ethtool_ops bnx2x_vf_ethtool_ops = {
+	.get_settings		= bnx2x_get_settings,
+	.set_settings		= bnx2x_set_settings,
+	.get_drvinfo		= bnx2x_get_drvinfo,
+	.get_msglevel		= bnx2x_get_msglevel,
+	.set_msglevel		= bnx2x_set_msglevel,
+	.get_link		= bnx2x_get_link,
+	.get_coalesce		= bnx2x_get_coalesce,
+	.get_ringparam		= bnx2x_get_ringparam,
+	.set_ringparam		= bnx2x_set_ringparam,
+	.get_sset_count		= bnx2x_get_sset_count,
+	.get_strings		= bnx2x_get_strings,
+	.get_ethtool_stats	= bnx2x_get_ethtool_stats,
+	.get_rxnfc		= bnx2x_get_rxnfc,
+	.set_rxnfc		= bnx2x_set_rxnfc,
+	.get_rxfh_indir_size	= bnx2x_get_rxfh_indir_size,
+	.get_rxfh_indir		= bnx2x_get_rxfh_indir,
+	.set_rxfh_indir		= bnx2x_set_rxfh_indir,
+	.get_channels		= bnx2x_get_channels,
+	.set_channels		= bnx2x_set_channels,
+};
+
+void bnx2x_set_ethtool_ops(struct bnx2x *bp, struct net_device *netdev)
 {
-	SET_ETHTOOL_OPS(netdev, &bnx2x_ethtool_ops);
+	if (IS_PF(bp))
+		SET_ETHTOOL_OPS(netdev, &bnx2x_ethtool_ops);
+	else /* vf */
+		SET_ETHTOOL_OPS(netdev, &bnx2x_vf_ethtool_ops);
 }
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index e81a747..14a7784 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -11953,7 +11953,7 @@ static int bnx2x_init_dev(struct bnx2x *bp, struct pci_dev *pdev,
 	dev->watchdog_timeo = TX_TIMEOUT;
 
 	dev->netdev_ops = &bnx2x_netdev_ops;
-	bnx2x_set_ethtool_ops(dev);
+	bnx2x_set_ethtool_ops(bp, dev);
 
 	dev->priv_flags |= IFF_UNICAST_FLT;
 
-- 
1.7.7.2

[PATCH net-next 03/13] bnx2x: Take chip version from MFW

[Yuval Mintz]

In latest boards, the CHIP_METAL register contains an incorrect
revision value, so the correct one needs to be obtained in a
different manner.

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c |    8 ++++++--
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h  |    6 ++----
 2 files changed, 8 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 14a7784..82f2e96 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -10034,8 +10034,12 @@ static void bnx2x_get_common_hwinfo(struct bnx2x *bp)
 	id = ((val & 0xffff) << 16);
 	val = REG_RD(bp, MISC_REG_CHIP_REV);
 	id |= ((val & 0xf) << 12);
-	val = REG_RD(bp, MISC_REG_CHIP_METAL);
-	id |= ((val & 0xff) << 4);
+
+	/* Metal is read from PCI regs, but we can't access >=0x400 from
+	 * the configuration space (so we need to reg_rd)
+	 */
+	val = REG_RD(bp, PCICFG_OFFSET + PCI_ID_VAL3);
+	id |= (((val >> 24) & 0xf) << 4);
 	val = REG_RD(bp, MISC_REG_BOND_ID);
 	id |= (val & 0xf);
 	bp->common.chip_id = id;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
index 791eb2d..d22bc40 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_reg.h
@@ -1491,10 +1491,6 @@
 /* [R 4] This field indicates the type of the device. '0' - 2 Ports; '1' - 1
    Port. */
 #define MISC_REG_BOND_ID					 0xa400
-/* [R 8] These bits indicate the metal revision of the chip. This value
-   starts at 0x00 for each all-layer tape-out and increments by one for each
-   tape-out. */
-#define MISC_REG_CHIP_METAL					 0xa404
 /* [R 16] These bits indicate the part number for the chip. */
 #define MISC_REG_CHIP_NUM					 0xa408
 /* [R 4] These bits indicate the base revision of the chip. This value
@@ -6331,6 +6327,8 @@
 #define PCI_PM_DATA_B					0x414
 #define PCI_ID_VAL1					0x434
 #define PCI_ID_VAL2					0x438
+#define PCI_ID_VAL3					0x43c
+
 #define GRC_CONFIG_REG_PF_INIT_VF		0x624
 #define GRC_CR_PF_INIT_VF_PF_FIRST_VF_NUM_MASK	0xf
 /* First VF_NUM for PF is encoded in this register.
-- 
1.7.7.2

[PATCH net-next 04/13] bnx2x: Prevent "Unknown MF" print in SF mode

[Yuval Mintz]

From: Ariel Elior <ariele@broadcom.com>

When using a chip operating in Single Function mode, when the chip is probed
the bnx2x would print a message warning of an unknown Multi Function mode.

This patch prevents said message.

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 82f2e96..423b5a07 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -11060,6 +11060,9 @@ static int bnx2x_get_hwinfo(struct bnx2x *bp)
 				} else
 					BNX2X_DEV_INFO("illegal OV for SD\n");
 				break;
+			case SHARED_FEAT_CFG_FORCE_SF_MODE_FORCED_SF:
+				bp->mf_config[vn] = 0;
+				break;
 			default:
 				/* Unknown configuration: reset mf_config */
 				bp->mf_config[vn] = 0;
-- 
1.7.7.2

[PATCH net-next 05/13] bnx2x: Add iproute2 support for vfs

[Yuval Mintz]

From: Ariel Elior <ariele@broadcom.com>

This patch adds support for iproute2 callbacks allowing querying a physical
function as to its child virtual functions, and setting the macs and vlans
of said virtual functions.

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x.h       |    1 +
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h   |    3 +
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c  |   38 +++-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c    |   24 +--
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h    |    6 +-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c |  239 +++++++++++++++++++--
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h |    3 +
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h  |    9 +-
 8 files changed, 277 insertions(+), 46 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 8ddc78b..d62d037 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -1214,6 +1214,7 @@ enum {
 	BNX2X_SP_RTNL_ENABLE_SRIOV,
 	BNX2X_SP_RTNL_VFPF_MCAST,
 	BNX2X_SP_RTNL_VFPF_STORM_RX_MODE,
+	BNX2X_SP_RTNL_HYPERVISOR_VLAN,
 };
 
 
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
index 8d158d8..4620fa5 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.h
@@ -496,7 +496,10 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev);
 /* setup_tc callback */
 int bnx2x_setup_tc(struct net_device *dev, u8 num_tc);
 
+int bnx2x_get_vf_config(struct net_device *dev, int vf,
+			struct ifla_vf_info *ivi);
 int bnx2x_set_vf_mac(struct net_device *dev, int queue, u8 *mac);
+int bnx2x_set_vf_vlan(struct net_device *netdev, int vf, u16 vlan, u8 qos);
 
 /* select_queue callback */
 u16 bnx2x_select_queue(struct net_device *dev, struct sk_buff *skb);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 423b5a07..9be9b03 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -75,8 +75,6 @@
 #define FW_FILE_NAME_E1H	"bnx2x/bnx2x-e1h-" FW_FILE_VERSION ".fw"
 #define FW_FILE_NAME_E2		"bnx2x/bnx2x-e2-" FW_FILE_VERSION ".fw"
 
-#define MAC_LEADING_ZERO_CNT (ALIGN(ETH_ALEN, sizeof(u32)) - ETH_ALEN)
-
 /* Time in jiffies before concluding the transmitter is hung */
 #define TX_TIMEOUT		(5*HZ)
 
@@ -3227,16 +3225,29 @@ static void bnx2x_drv_info_ether_stat(struct bnx2x *bp)
 {
 	struct eth_stats_info *ether_stat =
 		&bp->slowpath->drv_info_to_mcp.ether_stat;
+	struct bnx2x_vlan_mac_obj *mac_obj =
+		&bp->sp_objs->mac_obj;
+	int i;
 
 	strlcpy(ether_stat->version, DRV_MODULE_VERSION,
 		ETH_STAT_INFO_VERSION_LEN);
 
-	bp->sp_objs[0].mac_obj.get_n_elements(bp, &bp->sp_objs[0].mac_obj,
-					DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED,
-					ether_stat->mac_local);
-
+	/* get DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED macs, placing them in the
+	 * mac_local field in ether_stat struct. The base address is offset by 2
+	 * bytes to account for the field being 8 bytes but a mac address is
+	 * only 6 bytes. Likewise, the stride for the get_n_elements function is
+	 * 2 bytes to compensate from the 6 bytes of a mac to the 8 bytes
+	 * allocated by the ether_stat struct, so the macs will land in their
+	 * proper positions.
+	 */
+	for (i = 0; i < DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED; i++)
+		memset(ether_stat->mac_local + i, 0,
+		       sizeof(ether_stat->mac_local[0]));
+	mac_obj->get_n_elements(bp, &bp->sp_objs[0].mac_obj,
+				DRV_INFO_ETH_STAT_NUM_MACS_REQUIRED,
+				ether_stat->mac_local + MAC_PAD, MAC_PAD,
+				ETH_ALEN);
 	ether_stat->mtu_size = bp->dev->mtu;
-
 	if (bp->dev->features & NETIF_F_RXCSUM)
 		ether_stat->feature_flags |= FEATURE_ETH_CHKSUM_OFFLOAD_MASK;
 	if (bp->dev->features & NETIF_F_TSO)
@@ -3258,8 +3269,7 @@ static void bnx2x_drv_info_fcoe_stat(struct bnx2x *bp)
 	if (!CNIC_LOADED(bp))
 		return;
 
-	memcpy(fcoe_stat->mac_local + MAC_LEADING_ZERO_CNT,
-	       bp->fip_mac, ETH_ALEN);
+	memcpy(fcoe_stat->mac_local + MAC_PAD, bp->fip_mac, ETH_ALEN);
 
 	fcoe_stat->qos_priority =
 		app->traffic_type_priority[LLFC_TRAFFIC_TYPE_FCOE];
@@ -3361,8 +3371,8 @@ static void bnx2x_drv_info_iscsi_stat(struct bnx2x *bp)
 	if (!CNIC_LOADED(bp))
 		return;
 
-	memcpy(iscsi_stat->mac_local + MAC_LEADING_ZERO_CNT,
-	       bp->cnic_eth_dev.iscsi_mac, ETH_ALEN);
+	memcpy(iscsi_stat->mac_local + MAC_PAD, bp->cnic_eth_dev.iscsi_mac,
+	       ETH_ALEN);
 
 	iscsi_stat->qos_priority =
 		app->traffic_type_priority[LLFC_TRAFFIC_TYPE_ISCSI];
@@ -9525,6 +9535,10 @@ sp_rtnl_not_reset:
 		bnx2x_vfpf_storm_rx_mode(bp);
 	}
 
+	if (test_and_clear_bit(BNX2X_SP_RTNL_HYPERVISOR_VLAN,
+			       &bp->sp_rtnl_state))
+		bnx2x_pf_set_vfs_vlan(bp);
+
 	/* work which needs rtnl lock not-taken (as it takes the lock itself and
 	 * can be called from other contexts as well)
 	 */
@@ -11798,6 +11812,8 @@ static const struct net_device_ops bnx2x_netdev_ops = {
 	.ndo_setup_tc		= bnx2x_setup_tc,
 #ifdef CONFIG_BNX2X_SRIOV
 	.ndo_set_vf_mac		= bnx2x_set_vf_mac,
+	.ndo_set_vf_vlan        = bnx2x_set_vf_vlan,
+	.ndo_get_vf_config	= bnx2x_get_vf_config,
 #endif
 #ifdef NETDEV_FCOE_WWNN
 	.ndo_fcoe_get_wwn	= bnx2x_fcoe_get_wwn,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
index 9f2637c..6b03acd 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
@@ -30,8 +30,6 @@
 
 #define BNX2X_MAX_EMUL_MULTI		16
 
-#define MAC_LEADING_ZERO_CNT (ALIGN(ETH_ALEN, sizeof(u32)) - ETH_ALEN)
-
 /**** Exe Queue interfaces ****/
 
 /**
@@ -444,30 +442,21 @@ static bool bnx2x_put_credit_vlan_mac(struct bnx2x_vlan_mac_obj *o)
 }
 
 static int bnx2x_get_n_elements(struct bnx2x *bp, struct bnx2x_vlan_mac_obj *o,
-				int n, u8 *buf)
+				int n, u8 *base, u8 stride, u8 size)
 {
 	struct bnx2x_vlan_mac_registry_elem *pos;
-	u8 *next = buf;
+	u8 *next = base;
 	int counter = 0;
 
 	/* traverse list */
 	list_for_each_entry(pos, &o->head, link) {
 		if (counter < n) {
-			/* place leading zeroes in buffer */
-			memset(next, 0, MAC_LEADING_ZERO_CNT);
-
-			/* place mac after leading zeroes*/
-			memcpy(next + MAC_LEADING_ZERO_CNT, pos->u.mac.mac,
-			       ETH_ALEN);
-
-			/* calculate address of next element and
-			 * advance counter
-			 */
+			memcpy(next, &pos->u, size);
 			counter++;
-			next = buf + counter * ALIGN(ETH_ALEN, sizeof(u32));
+			DP(BNX2X_MSG_SP, "copied element number %d to address %p element was:\n",
+			   counter, next);
+			next += stride + size;
 
-			DP(BNX2X_MSG_SP, "copied element number %d to address %p element was %pM\n",
-			   counter, next, pos->u.mac.mac);
 		}
 	}
 	return counter * ETH_ALEN;
@@ -2013,6 +2002,7 @@ void bnx2x_init_vlan_obj(struct bnx2x *bp,
 		vlan_obj->check_move        = bnx2x_check_move;
 		vlan_obj->ramrod_cmd        =
 			RAMROD_CMD_ID_ETH_CLASSIFICATION_RULES;
+		vlan_obj->get_n_elements    = bnx2x_get_n_elements;
 
 		/* Exe Queue */
 		bnx2x_exe_queue_init(bp,
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
index ff90760..ac57e63 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
@@ -313,8 +313,9 @@ struct bnx2x_vlan_mac_obj {
 	 *
 	 * @return number of copied bytes
 	 */
-	int (*get_n_elements)(struct bnx2x *bp, struct bnx2x_vlan_mac_obj *o,
-			      int n, u8 *buf);
+	int (*get_n_elements)(struct bnx2x *bp,
+			      struct bnx2x_vlan_mac_obj *o, int n, u8 *base,
+			      u8 stride, u8 size);
 
 	/**
 	 * Checks if ADD-ramrod with the given params may be performed.
@@ -842,6 +843,7 @@ enum bnx2x_q_type {
 #define BNX2X_MULTI_TX_COS_E3B0			3
 #define BNX2X_MULTI_TX_COS			3 /* Maximum possible */
 
+#define MAC_PAD (ALIGN(ETH_ALEN, sizeof(u32)) - ETH_ALEN)
 
 struct bnx2x_queue_init_params {
 	struct {
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
index 6adfa20..7b234e4 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
@@ -20,7 +20,9 @@
 #include "bnx2x.h"
 #include "bnx2x_init.h"
 #include "bnx2x_cmn.h"
+#include "bnx2x_sp.h"
 #include <linux/crc32.h>
+#include <linux/if_vlan.h>
 
 /* General service functions */
 static void storm_memset_vf_to_pf(struct bnx2x *bp, u16 abs_fid,
@@ -958,6 +960,12 @@ op_err:
 	BNX2X_ERR("QSETUP[%d:%d] error: rc %d\n", vf->abs_vfid, qid, vfop->rc);
 op_done:
 	case BNX2X_VFOP_QSETUP_DONE:
+		vf->cfg_flags |= VF_CFG_VLAN;
+		smp_mb__before_clear_bit();
+		set_bit(BNX2X_SP_RTNL_HYPERVISOR_VLAN,
+			&bp->sp_rtnl_state);
+		smp_mb__after_clear_bit();
+		schedule_delayed_work(&bp->sp_rtnl_task, 0);
 		bnx2x_vfop_end(bp, vf, vfop);
 		return;
 	default:
@@ -3029,6 +3037,88 @@ void bnx2x_enable_sriov(struct bnx2x *bp)
 		DP(BNX2X_MSG_IOV, "sriov enabled\n");
 }
 
+void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp)
+{
+	int vfidx;
+	struct pf_vf_bulletin_content *bulletin;
+
+	DP(BNX2X_MSG_IOV, "configuring vlan for VFs from sp-task\n");
+	for_each_vf(bp, vfidx) {
+	bulletin = BP_VF_BULLETIN(bp, vfidx);
+		if (BP_VF(bp, vfidx)->cfg_flags & VF_CFG_VLAN)
+			bnx2x_set_vf_vlan(bp->dev, vfidx, bulletin->vlan, 0);
+	}
+}
+
+static int bnx2x_vf_ndo_sanity(struct bnx2x *bp, int vfidx,
+			       struct bnx2x_virtf *vf)
+{
+	if (!IS_SRIOV(bp)) {
+		BNX2X_ERR("vf ndo called though sriov is disabled\n");
+		return -EINVAL;
+	}
+
+	if (vfidx >= BNX2X_NR_VIRTFN(bp)) {
+		BNX2X_ERR("vf ndo called for uninitialized VF. vfidx was %d BNX2X_NR_VIRTFN was %d\n",
+			  vfidx, BNX2X_NR_VIRTFN(bp));
+		return -EINVAL;
+	}
+
+	if (!vf) {
+		BNX2X_ERR("vf ndo called but vf was null. vfidx was %d\n",
+			  vfidx);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int bnx2x_get_vf_config(struct net_device *dev, int vfidx,
+			struct ifla_vf_info *ivi)
+{
+	struct bnx2x *bp = netdev_priv(dev);
+	struct bnx2x_virtf *vf = BP_VF(bp, vfidx);
+	struct bnx2x_vlan_mac_obj *mac_obj = &bnx2x_vfq(vf, 0, mac_obj);
+	struct bnx2x_vlan_mac_obj *vlan_obj = &bnx2x_vfq(vf, 0, vlan_obj);
+	struct pf_vf_bulletin_content *bulletin = BP_VF_BULLETIN(bp, vfidx);
+	int rc;
+
+	/* sanity */
+	rc = bnx2x_vf_ndo_sanity(bp, vfidx, vf);
+	if (rc)
+		return rc;
+
+	ivi->vf = vfidx;
+	ivi->qos = 0;
+	ivi->tx_rate = 10000; /* always 10G. TBA take from link struct */
+	ivi->spoofchk = 1; /*always enabled */
+	if (vf->state == VF_ENABLED) {
+		/* mac and vlan are in vlan_mac objects */
+		mac_obj->get_n_elements(bp, mac_obj, 1, (u8 *)&ivi->mac,
+					0, ETH_ALEN);
+		vlan_obj->get_n_elements(bp, vlan_obj, 1, (u8 *)&ivi->vlan,
+					 0, VLAN_HLEN);
+	} else {
+		/* mac */
+		if (bulletin->valid_bitmap & (1 << MAC_ADDR_VALID))
+			/* mac configured by ndo so its in bulletin board */
+			memcpy(&ivi->mac, bulletin->mac, ETH_ALEN);
+		else
+			/* funtion has not been loaded yet. Show mac as 0s */
+			memset(&ivi->mac, 0, ETH_ALEN);
+
+		/* vlan */
+		if (bulletin->valid_bitmap & (1 << VLAN_VALID))
+			/* vlan configured by ndo so its in bulletin board */
+			memcpy(&ivi->vlan, &bulletin->vlan, VLAN_HLEN);
+		else
+			/* funtion has not been loaded yet. Show vlans as 0s */
+			memset(&ivi->vlan, 0, VLAN_HLEN);
+	}
+
+	return 0;
+}
+
 /* New mac for VF. Consider these cases:
  * 1. VF hasn't been acquired yet - save the mac in local bulletin board and
  *    supply at acquire.
@@ -3044,23 +3134,19 @@ void bnx2x_enable_sriov(struct bnx2x *bp)
  * VF to configure any mac for itself except for this mac. In case of a race
  * where the VF fails to see the new post on its bulletin board before sending a
  * mac configuration request, the PF will simply fail the request and VF can try
- * again after consulting its bulletin board
+ * again after consulting its bulletin board.
  */
-int bnx2x_set_vf_mac(struct net_device *dev, int queue, u8 *mac)
+int bnx2x_set_vf_mac(struct net_device *dev, int vfidx, u8 *mac)
 {
 	struct bnx2x *bp = netdev_priv(dev);
-	int rc, q_logical_state, vfidx = queue;
+	int rc, q_logical_state;
 	struct bnx2x_virtf *vf = BP_VF(bp, vfidx);
 	struct pf_vf_bulletin_content *bulletin = BP_VF_BULLETIN(bp, vfidx);
 
-	/* if SRIOV is disabled there is nothing to do (and somewhere, someone
-	 * has erred).
-	 */
-	if (!IS_SRIOV(bp)) {
-		BNX2X_ERR("bnx2x_set_vf_mac called though sriov is disabled\n");
-		return -EINVAL;
-	}
-
+	/* sanity */
+	rc = bnx2x_vf_ndo_sanity(bp, vfidx, vf);
+	if (rc)
+		return rc;
 	if (!is_valid_ether_addr(mac)) {
 		BNX2X_ERR("mac address invalid\n");
 		return -EINVAL;
@@ -3085,7 +3171,7 @@ int bnx2x_set_vf_mac(struct net_device *dev, int queue, u8 *mac)
 	if (vf->state == VF_ENABLED &&
 	    q_logical_state == BNX2X_Q_LOGICAL_STATE_ACTIVE) {
 		/* configure the mac in device on this vf's queue */
-		unsigned long flags = 0;
+		unsigned long ramrod_flags = 0;
 		struct bnx2x_vlan_mac_obj *mac_obj = &bnx2x_vfq(vf, 0, mac_obj);
 
 		/* must lock vfpf channel to protect against vf flows */
@@ -3106,14 +3192,133 @@ int bnx2x_set_vf_mac(struct net_device *dev, int queue, u8 *mac)
 		}
 
 		/* configure the new mac to device */
-		__set_bit(RAMROD_COMP_WAIT, &flags);
+		__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
 		bnx2x_set_mac_one(bp, (u8 *)&bulletin->mac, mac_obj, true,
-				  BNX2X_ETH_MAC, &flags);
+				  BNX2X_ETH_MAC, &ramrod_flags);
 
 		bnx2x_unlock_vf_pf_channel(bp, vf, CHANNEL_TLV_PF_SET_MAC);
 	}
 
-	return rc;
+	return 0;
+}
+
+int bnx2x_set_vf_vlan(struct net_device *dev, int vfidx, u16 vlan, u8 qos)
+{
+	struct bnx2x *bp = netdev_priv(dev);
+	int rc, q_logical_state;
+	struct bnx2x_virtf *vf = BP_VF(bp, vfidx);
+	struct pf_vf_bulletin_content *bulletin = BP_VF_BULLETIN(bp, vfidx);
+
+	/* sanity */
+	rc = bnx2x_vf_ndo_sanity(bp, vfidx, vf);
+	if (rc)
+		return rc;
+
+	if (vlan > 4095) {
+		BNX2X_ERR("illegal vlan value %d\n", vlan);
+		return -EINVAL;
+	}
+
+	DP(BNX2X_MSG_IOV, "configuring VF %d with VLAN %d qos %d\n",
+	   vfidx, vlan, 0);
+
+	/* update PF's copy of the VF's bulletin. No point in posting the vlan
+	 * to the VF since it doesn't have anything to do with it. But it useful
+	 * to store it here in case the VF is not up yet and we can only
+	 * configure the vlan later when it does.
+	 */
+	bulletin->valid_bitmap |= 1 << VLAN_VALID;
+	bulletin->vlan = vlan;
+
+	/* is vf initialized and queue set up? */
+	q_logical_state =
+		bnx2x_get_q_logical_state(bp, &bnx2x_vfq(vf, 0, sp_obj));
+	if (vf->state == VF_ENABLED &&
+	    q_logical_state == BNX2X_Q_LOGICAL_STATE_ACTIVE) {
+		/* configure the vlan in device on this vf's queue */
+		unsigned long ramrod_flags = 0;
+		unsigned long vlan_mac_flags = 0;
+		struct bnx2x_vlan_mac_obj *vlan_obj =
+			&bnx2x_vfq(vf, 0, vlan_obj);
+		struct bnx2x_vlan_mac_ramrod_params ramrod_param;
+		struct bnx2x_queue_state_params q_params = {NULL};
+		struct bnx2x_queue_update_params *update_params;
+
+		memset(&ramrod_param, 0, sizeof(ramrod_param));
+
+		/* must lock vfpf channel to protect against vf flows */
+		bnx2x_lock_vf_pf_channel(bp, vf, CHANNEL_TLV_PF_SET_VLAN);
+
+		/* remove existing vlans */
+		__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
+		rc = vlan_obj->delete_all(bp, vlan_obj, &vlan_mac_flags,
+					  &ramrod_flags);
+		if (rc) {
+			BNX2X_ERR("failed to delete vlans\n");
+			return -EINVAL;
+		}
+
+		/* send queue update ramrod to configure default vlan and silent
+		 * vlan removal
+		 */
+		__set_bit(RAMROD_COMP_WAIT, &q_params.ramrod_flags);
+		q_params.cmd = BNX2X_Q_CMD_UPDATE;
+		q_params.q_obj = &bnx2x_vfq(vf, 0, sp_obj);
+		update_params = &q_params.params.update;
+		__set_bit(BNX2X_Q_UPDATE_DEF_VLAN_EN_CHNG,
+			  &update_params->update_flags);
+		__set_bit(BNX2X_Q_UPDATE_SILENT_VLAN_REM_CHNG,
+			  &update_params->update_flags);
+
+		if (vlan == 0) {
+			/* if vlan is 0 then we want to leave the VF traffic
+			 * untagged, and leave the incoming traffic untouched
+			 * (i.e. do not remove any vlan tags).
+			 */
+			__clear_bit(BNX2X_Q_UPDATE_DEF_VLAN_EN,
+				    &update_params->update_flags);
+			__clear_bit(BNX2X_Q_UPDATE_SILENT_VLAN_REM,
+				    &update_params->update_flags);
+		} else {
+			/* configure the new vlan to device */
+			__set_bit(RAMROD_COMP_WAIT, &ramrod_flags);
+			ramrod_param.vlan_mac_obj = vlan_obj;
+			ramrod_param.ramrod_flags = ramrod_flags;
+			ramrod_param.user_req.u.vlan.vlan = vlan;
+			ramrod_param.user_req.cmd = BNX2X_VLAN_MAC_ADD;
+			rc = bnx2x_config_vlan_mac(bp, &ramrod_param);
+			if (rc) {
+				BNX2X_ERR("failed to configure vlan\n");
+				return -EINVAL;
+			}
+
+			/* configure default vlan to vf queue and set silent
+			 * vlan removal (the vf remains unaware of this vlan).
+			 */
+			update_params = &q_params.params.update;
+			__set_bit(BNX2X_Q_UPDATE_DEF_VLAN_EN,
+				  &update_params->update_flags);
+			__set_bit(BNX2X_Q_UPDATE_SILENT_VLAN_REM,
+				  &update_params->update_flags);
+			update_params->def_vlan = vlan;
+		}
+
+		/* Update the Queue state */
+		rc = bnx2x_queue_state_change(bp, &q_params);
+		if (rc) {
+			BNX2X_ERR("Failed to configure default VLAN\n");
+			return rc;
+		}
+
+		/* clear the flag indicating that this VF needs its vlan
+		 * (will only be set if the HV configured th Vlan before vf was
+		 * and we were called because the VF came up later
+		 */
+		vf->cfg_flags &= ~VF_CFG_VLAN;
+
+		bnx2x_unlock_vf_pf_channel(bp, vf, CHANNEL_TLV_PF_SET_VLAN);
+	}
+	return 0;
 }
 
 /* crc is the first field in the bulletin board. compute the crc over the
@@ -3165,6 +3370,10 @@ enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp)
 		memcpy(bp->dev->dev_addr, bulletin.mac, ETH_ALEN);
 	}
 
+	/* the vlan in bulletin board is valid and is new */
+	if (bulletin.valid_bitmap & 1 << VLAN_VALID)
+		memcpy(&bulletin.vlan, &bp->old_bulletin.vlan, VLAN_HLEN);
+
 	/* copy new bulletin board to bp */
 	bp->old_bulletin = bulletin;
 
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
index b405017..33d4951 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
@@ -193,6 +193,7 @@ struct bnx2x_virtf {
 #define VF_CFG_TPA		0x0004
 #define VF_CFG_INT_SIMD		0x0008
 #define VF_CACHE_LINE		0x0010
+#define VF_CFG_VLAN		0x0020
 
 	u8 state;
 #define VF_FREE		0	/* VF ready to be acquired holds no resc */
@@ -757,6 +758,7 @@ static inline int bnx2x_vf_headroom(struct bnx2x *bp)
 {
 	return bp->vfdb->sriov.nr_virtfn * BNX2X_CLIENTS_PER_VF;
 }
+void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp);
 
 #else /* CONFIG_BNX2X_SRIOV */
 
@@ -804,6 +806,7 @@ static inline enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp
 
 static inline int bnx2x_vf_map_doorbells(struct bnx2x *bp) {return 0; }
 static inline int bnx2x_vf_pci_alloc(struct bnx2x *bp) {return 0; }
+static inline void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp) {}
 
 #endif /* CONFIG_BNX2X_SRIOV */
 #endif /* bnx2x_sriov.h */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
index bfc80ba..41708fa 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_vfpf.h
@@ -328,9 +328,15 @@ struct pf_vf_bulletin_content {
 #define MAC_ADDR_VALID		0	/* alert the vf that a new mac address
 					 * is available for it
 					 */
+#define VLAN_VALID		1	/* when set, the vf should not access
+					 * the vfpf channel
+					 */
 
 	u8 mac[ETH_ALEN];
-	u8 padding[2];
+	u8 mac_padding[2];
+
+	u16 vlan;
+	u8 vlan_padding[6];
 };
 
 union pf_vf_bulletin {
@@ -353,6 +359,7 @@ enum channel_tlvs {
 	CHANNEL_TLV_LIST_END,
 	CHANNEL_TLV_FLR,
 	CHANNEL_TLV_PF_SET_MAC,
+	CHANNEL_TLV_PF_SET_VLAN,
 	CHANNEL_TLV_MAX
 };
 
-- 
1.7.7.2

[PATCH net-next 06/13] bnx2x: Control number of vfs dynamically

[Yuval Mintz]

From: Ariel Elior <ariele@broadcom.com>

1. Support sysfs interface for getting the maximal number of virtual functions
   of a given physical function.
2. Support sysfs interface for getting and setting the current number of
   virtual functions.

Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x.h       |    2 +
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c  |   36 ++-------
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c |   87 +++++++++++++++++----
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h |   10 ++-
 4 files changed, 91 insertions(+), 44 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index d62d037..33fbdfd 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -1281,6 +1281,8 @@ struct bnx2x {
 	dma_addr_t		pf2vf_bulletin_mapping;
 
 	struct pf_vf_bulletin_content	old_bulletin;
+
+	u16 requested_nr_virtfn;
 #endif /* CONFIG_BNX2X_SRIOV */
 
 	struct net_device	*dev;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index 9be9b03..f685d2e 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -9546,8 +9546,10 @@ sp_rtnl_not_reset:
 
 	/* enable SR-IOV if applicable */
 	if (IS_SRIOV(bp) && test_and_clear_bit(BNX2X_SP_RTNL_ENABLE_SRIOV,
-					       &bp->sp_rtnl_state))
+					       &bp->sp_rtnl_state)) {
+		bnx2x_disable_sriov(bp);
 		bnx2x_enable_sriov(bp);
+	}
 }
 
 static void bnx2x_period_task(struct work_struct *work)
@@ -11423,26 +11425,6 @@ static int bnx2x_init_bp(struct bnx2x *bp)
  * net_device service functions
  */
 
-static int bnx2x_open_epilog(struct bnx2x *bp)
-{
-	/* Enable sriov via delayed work. This must be done via delayed work
-	 * because it causes the probe of the vf devices to be run, which invoke
-	 * register_netdevice which must have rtnl lock taken. As we are holding
-	 * the lock right now, that could only work if the probe would not take
-	 * the lock. However, as the probe of the vf may be called from other
-	 * contexts as well (such as passthrough to vm failes) it can't assume
-	 * the lock is being held for it. Using delayed work here allows the
-	 * probe code to simply take the lock (i.e. wait for it to be released
-	 * if it is being held).
-	 */
-	smp_mb__before_clear_bit();
-	set_bit(BNX2X_SP_RTNL_ENABLE_SRIOV, &bp->sp_rtnl_state);
-	smp_mb__after_clear_bit();
-	schedule_delayed_work(&bp->sp_rtnl_task, 0);
-
-	return 0;
-}
-
 /* called with rtnl_lock */
 static int bnx2x_open(struct net_device *dev)
 {
@@ -12498,13 +12480,8 @@ static int bnx2x_init_one(struct pci_dev *pdev,
 			goto init_one_exit;
 	}
 
-	/* Enable SRIOV if capability found in configuration space.
-	 * Once the generic SR-IOV framework makes it in from the
-	 * pci tree this will be revised, to allow dynamic control
-	 * over the number of VFs. Right now, change the num of vfs
-	 * param below to enable SR-IOV.
-	 */
-	rc = bnx2x_iov_init_one(bp, int_mode, 0/*num vfs*/);
+	/* Enable SRIOV if capability found in configuration space */
+	rc = bnx2x_iov_init_one(bp, int_mode, BNX2X_MAX_NUM_OF_VFS);
 	if (rc)
 		goto init_one_exit;
 
@@ -12820,6 +12797,9 @@ static struct pci_driver bnx2x_pci_driver = {
 	.suspend     = bnx2x_suspend,
 	.resume      = bnx2x_resume,
 	.err_handler = &bnx2x_err_handler,
+#ifdef CONFIG_BNX2X_SRIOV
+	.sriov_configure = bnx2x_sriov_configure,
+#endif
 };
 
 static int __init bnx2x_init(void)
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
index 7b234e4..df930e3 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.c
@@ -1467,7 +1467,6 @@ static u8 bnx2x_vf_is_pcie_pending(struct bnx2x *bp, u8 abs_vfid)
 		return bnx2x_is_pcie_pending(dev);
 
 unknown_dev:
-	BNX2X_ERR("Unknown device\n");
 	return false;
 }
 
@@ -1972,8 +1971,10 @@ int bnx2x_iov_init_one(struct bnx2x *bp, int int_mode_param,
 	if (iov->total == 0)
 		goto failed;
 
-	/* calculate the actual number of VFs */
-	iov->nr_virtfn = min_t(u16, iov->total, (u16)num_vfs_param);
+	iov->nr_virtfn = min_t(u16, iov->total, num_vfs_param);
+
+	DP(BNX2X_MSG_IOV, "num_vfs_param was %d, nr_virtfn was %d\n",
+	   num_vfs_param, iov->nr_virtfn);
 
 	/* allocate the vf array */
 	bp->vfdb->vfs = kzalloc(sizeof(struct bnx2x_virtf) *
@@ -3020,21 +3021,47 @@ void bnx2x_unlock_vf_pf_channel(struct bnx2x *bp, struct bnx2x_virtf *vf,
 	vf->op_current = CHANNEL_TLV_NONE;
 }
 
-void bnx2x_enable_sriov(struct bnx2x *bp)
+int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs_param)
 {
-	int rc = 0;
 
-	/* disbale sriov in case it is still enabled */
-	pci_disable_sriov(bp->pdev);
-	DP(BNX2X_MSG_IOV, "sriov disabled\n");
+	struct bnx2x *bp = netdev_priv(pci_get_drvdata(dev));
 
-	/* enable sriov */
-	DP(BNX2X_MSG_IOV, "vf num (%d)\n", (bp->vfdb->sriov.nr_virtfn));
-	rc = pci_enable_sriov(bp->pdev, (bp->vfdb->sriov.nr_virtfn));
-	if (rc)
+	DP(BNX2X_MSG_IOV, "bnx2x_sriov_configure called with %d, BNX2X_NR_VIRTFN(bp) was %d\n",
+	   num_vfs_param, BNX2X_NR_VIRTFN(bp));
+
+	/* HW channel is only operational when PF is up */
+	if (bp->state != BNX2X_STATE_OPEN) {
+		BNX2X_ERR("VF num configurtion via sysfs not supported while PF is down");
+		return -EINVAL;
+	}
+
+	/* we are always bound by the total_vfs in the configuration space */
+	if (num_vfs_param > BNX2X_NR_VIRTFN(bp)) {
+		BNX2X_ERR("truncating requested number of VFs (%d) down to maximum allowed (%d)\n",
+			  num_vfs_param, BNX2X_NR_VIRTFN(bp));
+		num_vfs_param = BNX2X_NR_VIRTFN(bp);
+	}
+
+	bp->requested_nr_virtfn = num_vfs_param;
+	if (num_vfs_param == 0) {
+		pci_disable_sriov(dev);
+		return 0;
+	} else {
+		return bnx2x_enable_sriov(bp);
+	}
+}
+
+int bnx2x_enable_sriov(struct bnx2x *bp)
+{
+	int rc = 0, req_vfs = bp->requested_nr_virtfn;
+
+	rc = pci_enable_sriov(bp->pdev, req_vfs);
+	if (rc) {
 		BNX2X_ERR("pci_enable_sriov failed with %d\n", rc);
-	else
-		DP(BNX2X_MSG_IOV, "sriov enabled\n");
+		return rc;
+	}
+	DP(BNX2X_MSG_IOV, "sriov enabled (%d vfs)\n", req_vfs);
+	return req_vfs;
 }
 
 void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp)
@@ -3050,6 +3077,11 @@ void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp)
 	}
 }
 
+void bnx2x_disable_sriov(struct bnx2x *bp)
+{
+	pci_disable_sriov(bp->pdev);
+}
+
 static int bnx2x_vf_ndo_sanity(struct bnx2x *bp, int vfidx,
 			       struct bnx2x_virtf *vf)
 {
@@ -3087,6 +3119,10 @@ int bnx2x_get_vf_config(struct net_device *dev, int vfidx,
 	rc = bnx2x_vf_ndo_sanity(bp, vfidx, vf);
 	if (rc)
 		return rc;
+	if (!mac_obj || !vlan_obj || !bulletin) {
+		BNX2X_ERR("VF partially initialized\n");
+		return -EINVAL;
+	}
 
 	ivi->vf = vfidx;
 	ivi->qos = 0;
@@ -3405,3 +3441,26 @@ alloc_mem_err:
 		       sizeof(union pf_vf_bulletin));
 	return -ENOMEM;
 }
+
+int bnx2x_open_epilog(struct bnx2x *bp)
+{
+	/* Enable sriov via delayed work. This must be done via delayed work
+	 * because it causes the probe of the vf devices to be run, which invoke
+	 * register_netdevice which must have rtnl lock taken. As we are holding
+	 * the lock right now, that could only work if the probe would not take
+	 * the lock. However, as the probe of the vf may be called from other
+	 * contexts as well (such as passthrough to vm failes) it can't assume
+	 * the lock is being held for it. Using delayed work here allows the
+	 * probe code to simply take the lock (i.e. wait for it to be released
+	 * if it is being held). We only want to do this if the number of VFs
+	 * was set before PF driver was loaded.
+	 */
+	if (IS_SRIOV(bp) && BNX2X_NR_VIRTFN(bp)) {
+		smp_mb__before_clear_bit();
+		set_bit(BNX2X_SP_RTNL_ENABLE_SRIOV, &bp->sp_rtnl_state);
+		smp_mb__after_clear_bit();
+		schedule_delayed_work(&bp->sp_rtnl_task, 0);
+	}
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
index 33d4951..a10bdb2 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sriov.h
@@ -753,12 +753,15 @@ static inline int bnx2x_vf_ustorm_prods_offset(struct bnx2x *bp,
 enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp);
 void bnx2x_vf_map_doorbells(struct bnx2x *bp);
 int bnx2x_vf_pci_alloc(struct bnx2x *bp);
-void bnx2x_enable_sriov(struct bnx2x *bp);
+int bnx2x_enable_sriov(struct bnx2x *bp);
+void bnx2x_disable_sriov(struct bnx2x *bp);
 static inline int bnx2x_vf_headroom(struct bnx2x *bp)
 {
 	return bp->vfdb->sriov.nr_virtfn * BNX2X_CLIENTS_PER_VF;
 }
 void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp);
+int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs);
+int bnx2x_open_epilog(struct bnx2x *bp);
 
 #else /* CONFIG_BNX2X_SRIOV */
 
@@ -781,7 +784,8 @@ static inline void bnx2x_iov_init_dmae(struct bnx2x *bp) {}
 static inline int bnx2x_iov_init_one(struct bnx2x *bp, int int_mode_param,
 				     int num_vfs_param) {return 0; }
 static inline void bnx2x_iov_remove_one(struct bnx2x *bp) {}
-static inline void bnx2x_enable_sriov(struct bnx2x *bp) {}
+static inline int bnx2x_enable_sriov(struct bnx2x *bp) {return 0; }
+static inline void bnx2x_disable_sriov(struct bnx2x *bp) {}
 static inline int bnx2x_vfpf_acquire(struct bnx2x *bp,
 				     u8 tx_count, u8 rx_count) {return 0; }
 static inline int bnx2x_vfpf_release(struct bnx2x *bp) {return 0; }
@@ -807,6 +811,8 @@ static inline enum sample_bulletin_result bnx2x_sample_bulletin(struct bnx2x *bp
 static inline int bnx2x_vf_map_doorbells(struct bnx2x *bp) {return 0; }
 static inline int bnx2x_vf_pci_alloc(struct bnx2x *bp) {return 0; }
 static inline void bnx2x_pf_set_vfs_vlan(struct bnx2x *bp) {}
+static inline int bnx2x_sriov_configure(struct pci_dev *dev, int num_vfs) {return 0; }
+static inline int bnx2x_open_epilog(struct bnx2x *bp) {return 0; }
 
 #endif /* CONFIG_BNX2X_SRIOV */
 #endif /* bnx2x_sriov.h */
-- 
1.7.7.2

Re: [PATCH net-next 06/13] bnx2x: Control number of vfs dynamically

[David Miller]

From: "Yuval Mintz" <yuvalmin@broadcom.com>
Date: Mon, 11 Mar 2013 17:17:46 +0200

> From: Ariel Elior <ariele@broadcom.com>
> 
> 1. Support sysfs interface for getting the maximal number of virtual functions
>    of a given physical function.
> 2. Support sysfs interface for getting and setting the current number of
>    virtual functions.
> 
> Signed-off-by: Ariel Elior <ariele@broadcom.com>
> Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
> Signed-off-by: Eilon Greenstein <eilong@broadcom.com>

I don't like these sysfs knobs at all, what are other drivers
doing?

RE: [PATCH net-next 06/13] bnx2x: Control number of vfs dynamically

[Ariel Elior]

> -----Original Message-----
> From: David Miller [mailto:davem@davemloft.net]
> Sent: Tuesday, March 12, 2013 11:28 AM
> To: Yuval Mintz
> Cc: netdev@vger.kernel.org; Eilon Greenstein; Ariel Elior
> Subject: Re: [PATCH net-next 06/13] bnx2x: Control number of vfs
> dynamically
> 
> From: "Yuval Mintz" <yuvalmin@broadcom.com>
> Date: Mon, 11 Mar 2013 17:17:46 +0200
> 
> > From: Ariel Elior <ariele@broadcom.com>
> >
> > 1. Support sysfs interface for getting the maximal number of virtual
> functions
> >    of a given physical function.
> > 2. Support sysfs interface for getting and setting the current number of
> >    virtual functions.
> >
> > Signed-off-by: Ariel Elior <ariele@broadcom.com>
> > Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
> > Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
> 
> I don't like these sysfs knobs at all, what are other drivers
> doing?

Intel's ixgbe and igb already implement these callbacks.
Other drivers are using proprietary module parameters.
This is a generic sysfs interface which any driver can use.

Yuval Mintz from Broadcom started the
"New commands to configure IOV features" thread
http://marc.info/?l=linux-netdev&m=133638974914452&w=2
As the discussion developed it was realized that virtual functions
have a wider scope than networking.
The discussion culminated with a submission to the pci tree:
commit 1789382a72a537447d65ea4131d8bcc1ad85ce7b
Author: Donald Dutile <ddutile@redhat.com>
Date:   Mon Nov 5 15:20:36 2012 -0500

Adding sriov_numvfs and sriov_totalvfs to all SR-IOV capable pci devices.
I would like to add that sysfs is actually a very convenient place to have 
these knobs, because the virtfn entries there are a simple reliable way to
know which vf devices sprouted from which PF devices and what are their
names, which is very handy when scripting with VFs.

Thanks,
Ariel

Re: [PATCH net-next 06/13] bnx2x: Control number of vfs dynamically

[David Miller]

From: "Ariel Elior" <ariele@broadcom.com>
Date: Tue, 12 Mar 2013 10:51:13 +0000

> Yuval Mintz from Broadcom started the
> "New commands to configure IOV features" thread
> http://marc.info/?l=linux-netdev&m=133638974914452&w=2
> As the discussion developed it was realized that virtual functions
> have a wider scope than networking.
> The discussion culminated with a submission to the pci tree:
> commit 1789382a72a537447d65ea4131d8bcc1ad85ce7b
> Author: Donald Dutile <ddutile@redhat.com>
> Date:   Mon Nov 5 15:20:36 2012 -0500
> 
> Adding sriov_numvfs and sriov_totalvfs to all SR-IOV capable pci devices.
> I would like to add that sysfs is actually a very convenient place to have 
> these knobs, because the virtfn entries there are a simple reliable way to
> know which vf devices sprouted from which PF devices and what are their
> names, which is very handy when scripting with VFs.

Oh that's right, thanks for reminding me about that discussion.
I rescing my objection.

[PATCH net-next 07/13] bnx2x: Get gso_segs from FW

[Yuval Mintz]

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c |    9 +++++----
 1 files changed, 5 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index a923bc4..cd74ee5 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -451,7 +451,8 @@ static void bnx2x_tpa_start(struct bnx2x_fastpath *fp, u16 queue,
  * Compute number of aggregated segments, and gso_type.
  */
 static void bnx2x_set_gro_params(struct sk_buff *skb, u16 parsing_flags,
-				 u16 len_on_bd, unsigned int pkt_len)
+				 u16 len_on_bd, unsigned int pkt_len,
+				 u16 num_of_coalesced_segs)
 {
 	/* TPA aggregation won't have either IP options or TCP options
 	 * other than timestamp or IPv6 extension headers.
@@ -480,8 +481,7 @@ static void bnx2x_set_gro_params(struct sk_buff *skb, u16 parsing_flags,
 	/* tcp_gro_complete() will copy NAPI_GRO_CB(skb)->count
 	 * to skb_shinfo(skb)->gso_segs
 	 */
-	NAPI_GRO_CB(skb)->count = DIV_ROUND_UP(pkt_len - hdrs_len,
-					       skb_shinfo(skb)->gso_size);
+	NAPI_GRO_CB(skb)->count = num_of_coalesced_segs;
 }
 
 static int bnx2x_alloc_rx_sge(struct bnx2x *bp,
@@ -537,7 +537,8 @@ static int bnx2x_fill_frag_skb(struct bnx2x *bp, struct bnx2x_fastpath *fp,
 	/* This is needed in order to enable forwarding support */
 	if (frag_size)
 		bnx2x_set_gro_params(skb, tpa_info->parsing_flags, len_on_bd,
-				     le16_to_cpu(cqe->pkt_len));
+				     le16_to_cpu(cqe->pkt_len),
+				     le16_to_cpu(cqe->num_of_coalesced_segs));
 
 #ifdef BNX2X_STOP_ON_ERROR
 	if (pages > min_t(u32, 8, MAX_SKB_FRAGS) * SGE_PAGES) {
-- 
1.7.7.2

[PATCH net-next 08/13] bnx2x: Add RJ45 SFP module detection

[Yuval Mintz]

From: Yaniv Rosner <yanivr@broadcom.com>

Add RJ45 SFP module detection. In case the user set 10G link speed, and the
module doesn't support it, then force the speed to 1G and notify the user.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c |   21 +++++++++++++--------
 1 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
index 31c5787..600f850 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
@@ -152,6 +152,7 @@
 #define SFP_EEPROM_CON_TYPE_ADDR		0x2
 	#define SFP_EEPROM_CON_TYPE_VAL_LC	0x7
 	#define SFP_EEPROM_CON_TYPE_VAL_COPPER	0x21
+	#define SFP_EEPROM_CON_TYPE_VAL_RJ45	0x22
 
 
 #define SFP_EEPROM_COMP_CODE_ADDR		0x3
@@ -8049,20 +8050,24 @@ static int bnx2x_get_edc_mode(struct bnx2x_phy *phy,
 		break;
 	}
 	case SFP_EEPROM_CON_TYPE_VAL_LC:
+	case SFP_EEPROM_CON_TYPE_VAL_RJ45:
 		check_limiting_mode = 1;
 		if ((val[1] & (SFP_EEPROM_COMP_CODE_SR_MASK |
 			       SFP_EEPROM_COMP_CODE_LR_MASK |
 			       SFP_EEPROM_COMP_CODE_LRM_MASK)) == 0) {
-			DP(NETIF_MSG_LINK, "1G Optic module detected\n");
+			DP(NETIF_MSG_LINK, "1G SFP module detected\n");
 			gport = params->port;
 			phy->media_type = ETH_PHY_SFP_1G_FIBER;
-			phy->req_line_speed = SPEED_1000;
-			if (!CHIP_IS_E1x(bp))
-				gport = BP_PATH(bp) + (params->port << 1);
-			netdev_err(bp->dev, "Warning: Link speed was forced to 1000Mbps."
-			      " Current SFP module in port %d is not"
-			      " compliant with 10G Ethernet\n",
-			 gport);
+			if (phy->req_line_speed != SPEED_1000) {
+				phy->req_line_speed = SPEED_1000;
+				if (!CHIP_IS_E1x(bp)) {
+					gport = BP_PATH(bp) +
+					(params->port << 1);
+				}
+				netdev_err(bp->dev,
+					   "Warning: Link speed was forced to 1000Mbps. Current SFP module in port %d is not compliant with 10G Ethernet\n",
+					   gport);
+			}
 		} else {
 			int idx, cfg_idx = 0;
 			DP(NETIF_MSG_LINK, "10G Optic module detected\n");
-- 
1.7.7.2

[PATCH net-next 09/13] bnx2x: Add EEE support for BCM84834

[Yuval Mintz]

From: Yaniv Rosner <yanivr@broadcom.com>

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
index 600f850..fa027d8 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
@@ -10284,7 +10284,8 @@ static u8 bnx2x_848xx_read_status(struct bnx2x_phy *phy,
 				LINK_STATUS_LINK_PARTNER_10GXFD_CAPABLE;
 
 		/* Determine if EEE was negotiated */
-		if (phy->type == PORT_HW_CFG_XGXS_EXT_PHY_TYPE_BCM84833)
+		if ((phy->type == PORT_HW_CFG_XGXS_EXT_PHY_TYPE_BCM84833) ||
+		    (phy->type == PORT_HW_CFG_XGXS_EXT_PHY_TYPE_BCM84834))
 			bnx2x_eee_an_resolve(phy, params, vars);
 	}
 
-- 
1.7.7.2

[PATCH net-next 10/13] bnx2x: Control SFP+ tap values via nvm config

[Yuval Mintz]

From: Yaniv Rosner <yanivr@broadcom.com>

Configure SFP+ tap values to optimize link signal according to NVRAM setup.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h  |   17 +++++-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c |   77 ++++++++++++---------
 2 files changed, 60 insertions(+), 34 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
index 037860e..a7a3504 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
@@ -508,7 +508,22 @@ struct port_hw_cfg {		    /* port 0: 0x12c  port 1: 0x2bc */
 	#define PORT_HW_CFG_PAUSE_ON_HOST_RING_DISABLED               0x00000000
 	#define PORT_HW_CFG_PAUSE_ON_HOST_RING_ENABLED                0x00000001
 
-	u32 reserved0[6];				    /* 0x178 */
+	/* SFP+ Tx Equalization: NIC recommended and tested value is 0xBEB2
+	 * LOM recommended and tested value is 0xBEB2. Using a different
+	 * value means using a value not tested by BRCM
+	 */
+	u32 sfi_tap_values;                                 /* 0x178 */
+	#define PORT_HW_CFG_TX_EQUALIZATION_MASK                      0x0000FFFF
+	#define PORT_HW_CFG_TX_EQUALIZATION_SHIFT                     0
+
+	/* SFP+ Tx driver broadcast IDRIVER: NIC recommended and tested
+	 * value is 0x2. LOM recommended and tested value is 0x2. Using a
+	 * different value means using a value not tested by BRCM
+	 */
+	#define PORT_HW_CFG_TX_DRV_BROADCAST_MASK                     0x000F0000
+	#define PORT_HW_CFG_TX_DRV_BROADCAST_SHIFT                    16
+
+	u32 reserved0[5];				    /* 0x17c */
 
 	u32 aeu_int_mask;				    /* 0x190 */
 
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
index fa027d8..2d5bac8 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_link.c
@@ -3630,6 +3630,16 @@ static u8 bnx2x_ext_phy_resolve_fc(struct bnx2x_phy *phy,
  * init configuration, and set/clear SGMII flag. Internal
  * phy init is done purely in phy_init stage.
  */
+#define WC_TX_DRIVER(post2, idriver, ipre) \
+	((post2 << MDIO_WC_REG_TX0_TX_DRIVER_POST2_COEFF_OFFSET) | \
+	 (idriver << MDIO_WC_REG_TX0_TX_DRIVER_IDRIVER_OFFSET) | \
+	 (ipre << MDIO_WC_REG_TX0_TX_DRIVER_IPRE_DRIVER_OFFSET))
+
+#define WC_TX_FIR(post, main, pre) \
+	((post << MDIO_WC_REG_TX_FIR_TAP_POST_TAP_OFFSET) | \
+	 (main << MDIO_WC_REG_TX_FIR_TAP_MAIN_TAP_OFFSET) | \
+	 (pre << MDIO_WC_REG_TX_FIR_TAP_PRE_TAP_OFFSET))
+
 static void bnx2x_warpcore_enable_AN_KR2(struct bnx2x_phy *phy,
 					 struct link_params *params,
 					 struct link_vars *vars)
@@ -3754,20 +3764,13 @@ static void bnx2x_warpcore_enable_AN_KR(struct bnx2x_phy *phy,
 	/* Set Transmit PMD settings */
 	lane = bnx2x_get_warpcore_lane(phy, params);
 	bnx2x_cl45_write(bp, phy, MDIO_WC_DEVAD,
-		      MDIO_WC_REG_TX0_TX_DRIVER + 0x10*lane,
-		     ((0x02 << MDIO_WC_REG_TX0_TX_DRIVER_POST2_COEFF_OFFSET) |
-		      (0x06 << MDIO_WC_REG_TX0_TX_DRIVER_IDRIVER_OFFSET) |
-		      (0x09 << MDIO_WC_REG_TX0_TX_DRIVER_IPRE_DRIVER_OFFSET)));
+			 MDIO_WC_REG_TX0_TX_DRIVER + 0x10*lane,
+			 WC_TX_DRIVER(0x02, 0x06, 0x09));
 	/* Configure the next lane if dual mode */
 	if (phy->flags & FLAGS_WC_DUAL_MODE)
 		bnx2x_cl45_write(bp, phy, MDIO_WC_DEVAD,
 				 MDIO_WC_REG_TX0_TX_DRIVER + 0x10*(lane+1),
-				 ((0x02 <<
-				 MDIO_WC_REG_TX0_TX_DRIVER_POST2_COEFF_OFFSET) |
-				  (0x06 <<
-				   MDIO_WC_REG_TX0_TX_DRIVER_IDRIVER_OFFSET) |
-				  (0x09 <<
-				MDIO_WC_REG_TX0_TX_DRIVER_IPRE_DRIVER_OFFSET)));
+				 WC_TX_DRIVER(0x02, 0x06, 0x09));
 	bnx2x_cl45_write(bp, phy, MDIO_WC_DEVAD,
 			 MDIO_WC_REG_CL72_USERB0_CL72_OS_DEF_CTRL,
 			 0x03f0);
@@ -3910,6 +3913,8 @@ static void bnx2x_warpcore_set_10G_XFI(struct bnx2x_phy *phy,
 {
 	struct bnx2x *bp = params->bp;
 	u16 misc1_val, tap_val, tx_driver_val, lane, val;
+	u32 cfg_tap_val, tx_drv_brdct, tx_equal;
+
 	/* Hold rxSeqStart */
 	bnx2x_cl45_read_or_write(bp, phy, MDIO_WC_DEVAD,
 				 MDIO_WC_REG_DSC2B0_DSC_MISC_CTRL0, 0x8000);
@@ -3953,23 +3958,33 @@ static void bnx2x_warpcore_set_10G_XFI(struct bnx2x_phy *phy,
 
 	if (is_xfi) {
 		misc1_val |= 0x5;
-		tap_val = ((0x08 << MDIO_WC_REG_TX_FIR_TAP_POST_TAP_OFFSET) |
-			   (0x37 << MDIO_WC_REG_TX_FIR_TAP_MAIN_TAP_OFFSET) |
-			   (0x00 << MDIO_WC_REG_TX_FIR_TAP_PRE_TAP_OFFSET));
-		tx_driver_val =
-		      ((0x00 << MDIO_WC_REG_TX0_TX_DRIVER_POST2_COEFF_OFFSET) |
-		       (0x02 << MDIO_WC_REG_TX0_TX_DRIVER_IDRIVER_OFFSET) |
-		       (0x03 << MDIO_WC_REG_TX0_TX_DRIVER_IPRE_DRIVER_OFFSET));
-
+		tap_val = WC_TX_FIR(0x08, 0x37, 0x00);
+		tx_driver_val = WC_TX_DRIVER(0x00, 0x02, 0x03);
 	} else {
+		cfg_tap_val = REG_RD(bp, params->shmem_base +
+				     offsetof(struct shmem_region, dev_info.
+					      port_hw_config[params->port].
+					      sfi_tap_values));
+
+		tx_equal = cfg_tap_val & PORT_HW_CFG_TX_EQUALIZATION_MASK;
+
+		tx_drv_brdct = (cfg_tap_val &
+				PORT_HW_CFG_TX_DRV_BROADCAST_MASK) >>
+			       PORT_HW_CFG_TX_DRV_BROADCAST_SHIFT;
+
 		misc1_val |= 0x9;
-		tap_val = ((0x0f << MDIO_WC_REG_TX_FIR_TAP_POST_TAP_OFFSET) |
-			   (0x2b << MDIO_WC_REG_TX_FIR_TAP_MAIN_TAP_OFFSET) |
-			   (0x02 << MDIO_WC_REG_TX_FIR_TAP_PRE_TAP_OFFSET));
-		tx_driver_val =
-		      ((0x03 << MDIO_WC_REG_TX0_TX_DRIVER_POST2_COEFF_OFFSET) |
-		       (0x02 << MDIO_WC_REG_TX0_TX_DRIVER_IDRIVER_OFFSET) |
-		       (0x06 << MDIO_WC_REG_TX0_TX_DRIVER_IPRE_DRIVER_OFFSET));
+
+		/* TAP values are controlled by nvram, if value there isn't 0 */
+		if (tx_equal)
+			tap_val = (u16)tx_equal;
+		else
+			tap_val = WC_TX_FIR(0x0f, 0x2b, 0x02);
+
+		if (tx_drv_brdct)
+			tx_driver_val = WC_TX_DRIVER(0x03, (u16)tx_drv_brdct,
+						     0x06);
+		else
+			tx_driver_val = WC_TX_DRIVER(0x03, 0x02, 0x06);
 	}
 	bnx2x_cl45_write(bp, phy, MDIO_WC_DEVAD,
 			 MDIO_WC_REG_SERDESDIGITAL_MISC1, misc1_val);
@@ -4106,15 +4121,11 @@ static void bnx2x_warpcore_set_20G_DXGXS(struct bnx2x *bp,
 	/* Set Transmit PMD settings */
 	bnx2x_cl45_write(bp, phy, MDIO_WC_DEVAD,
 			 MDIO_WC_REG_TX_FIR_TAP,
-			((0x12 << MDIO_WC_REG_TX_FIR_TAP_POST_TAP_OFFSET) |
-			 (0x2d << MDIO_WC_REG_TX_FIR_TAP_MAIN_TAP_OFFSET) |
-			 (0x00 << MDIO_WC_REG_TX_FIR_TAP_PRE_TAP_OFFSET) |
-			 MDIO_WC_REG_TX_FIR_TAP_ENABLE));
+			 (WC_TX_FIR(0x12, 0x2d, 0x00) |
+			  MDIO_WC_REG_TX_FIR_TAP_ENABLE));
 	bnx2x_cl45_write(bp, phy, MDIO_WC_DEVAD,
-		      MDIO_WC_REG_TX0_TX_DRIVER + 0x10*lane,
-		     ((0x02 << MDIO_WC_REG_TX0_TX_DRIVER_POST2_COEFF_OFFSET) |
-		      (0x02 << MDIO_WC_REG_TX0_TX_DRIVER_IDRIVER_OFFSET) |
-		      (0x02 << MDIO_WC_REG_TX0_TX_DRIVER_IPRE_DRIVER_OFFSET)));
+			 MDIO_WC_REG_TX0_TX_DRIVER + 0x10*lane,
+			 WC_TX_DRIVER(0x02, 0x02, 0x02));
 }
 
 static void bnx2x_warpcore_set_sgmii_speed(struct bnx2x_phy *phy,
-- 
1.7.7.2

[PATCH net-next 11/13] bnx2x: Avoid using zero MAC

[Yuval Mintz]

Prevent bnx2x devices which are used mainly for storage from using zero
MAC addresses as their primary MAC address.

Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c |   12 +++++-------
 1 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index f685d2e..e5662a1 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -10832,14 +10832,12 @@ static void bnx2x_get_cnic_mac_hwinfo(struct bnx2x *bp)
 			}
 		}
 
-		if (IS_MF_STORAGE_SD(bp))
-			/* Zero primary MAC configuration */
-			memset(bp->dev->dev_addr, 0, ETH_ALEN);
-
-		if (IS_MF_FCOE_AFEX(bp) || IS_MF_FCOE_SD(bp))
-			/* use FIP MAC as primary MAC */
+		/* If this is a storage-only interface, use SAN mac as
+		 * primary MAC. Notice that for SD this is already the case,
+		 * as the SAN mac was copied from the primary MAC.
+		 */
+		if (IS_MF_FCOE_AFEX(bp))
 			memcpy(bp->dev->dev_addr, fip_mac, ETH_ALEN);
-
 	} else {
 		val2 = SHMEM_RD(bp, dev_info.port_hw_config[port].
 				iscsi_mac_upper);
-- 
1.7.7.2

[PATCH net-next 12/13] bnx2x: use FW 7.8.17

[Yuval Mintz]

From: Dmitry Kravkov <dmitry@broadcom.com>

Update appropriate HSI files and adapt driver accordingly.

Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x.h        |    2 +-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c    |  122 +++++------
 .../net/ethernet/broadcom/bnx2x/bnx2x_fw_defs.h    |   87 ++++----
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h    |  235 +++++++++++++++++---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c   |    5 +-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c     |   29 ++-
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h     |    5 +-
 7 files changed, 338 insertions(+), 147 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
index 33fbdfd..f865ad5 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x.h
@@ -729,7 +729,7 @@ struct bnx2x_fastpath {
 #define SKB_CS(skb)		(*(u16 *)(skb_transport_header(skb) + \
 					  skb->csum_offset))
 
-#define pbd_tcp_flags(skb)	(ntohl(tcp_flag_word(tcp_hdr(skb)))>>16 & 0xff)
+#define pbd_tcp_flags(tcp_hdr)	(ntohl(tcp_flag_word(tcp_hdr))>>16 & 0xff)
 
 #define XMIT_PLAIN			0
 #define XMIT_CSUM_V4			0x1
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
index cd74ee5..9f7a379 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c
@@ -3086,11 +3086,11 @@ int bnx2x_poll(struct napi_struct *napi, int budget)
  * to ease the pain of our fellow microcode engineers
  * we use one mapping for both BDs
  */
-static noinline u16 bnx2x_tx_split(struct bnx2x *bp,
-				   struct bnx2x_fp_txdata *txdata,
-				   struct sw_tx_bd *tx_buf,
-				   struct eth_tx_start_bd **tx_bd, u16 hlen,
-				   u16 bd_prod, int nbd)
+static u16 bnx2x_tx_split(struct bnx2x *bp,
+			  struct bnx2x_fp_txdata *txdata,
+			  struct sw_tx_bd *tx_buf,
+			  struct eth_tx_start_bd **tx_bd, u16 hlen,
+			  u16 bd_prod)
 {
 	struct eth_tx_start_bd *h_tx_bd = *tx_bd;
 	struct eth_tx_bd *d_tx_bd;
@@ -3098,11 +3098,10 @@ static noinline u16 bnx2x_tx_split(struct bnx2x *bp,
 	int old_len = le16_to_cpu(h_tx_bd->nbytes);
 
 	/* first fix first BD */
-	h_tx_bd->nbd = cpu_to_le16(nbd);
 	h_tx_bd->nbytes = cpu_to_le16(hlen);
 
-	DP(NETIF_MSG_TX_QUEUED,	"TSO split header size is %d (%x:%x) nbd %d\n",
-	   h_tx_bd->nbytes, h_tx_bd->addr_hi, h_tx_bd->addr_lo, h_tx_bd->nbd);
+	DP(NETIF_MSG_TX_QUEUED,	"TSO split header size is %d (%x:%x)\n",
+	   h_tx_bd->nbytes, h_tx_bd->addr_hi, h_tx_bd->addr_lo);
 
 	/* now get a new data BD
 	 * (after the pbd) and fill it */
@@ -3131,7 +3130,7 @@ static noinline u16 bnx2x_tx_split(struct bnx2x *bp,
 
 #define bswab32(b32) ((__force __le32) swab32((__force __u32) (b32)))
 #define bswab16(b16) ((__force __le16) swab16((__force __u16) (b16)))
-static inline __le16 bnx2x_csum_fix(unsigned char *t_header, u16 csum, s8 fix)
+static __le16 bnx2x_csum_fix(unsigned char *t_header, u16 csum, s8 fix)
 {
 	__sum16 tsum = (__force __sum16) csum;
 
@@ -3146,7 +3145,7 @@ static inline __le16 bnx2x_csum_fix(unsigned char *t_header, u16 csum, s8 fix)
 	return bswab16(tsum);
 }
 
-static inline u32 bnx2x_xmit_type(struct bnx2x *bp, struct sk_buff *skb)
+static u32 bnx2x_xmit_type(struct bnx2x *bp, struct sk_buff *skb)
 {
 	u32 rc;
 
@@ -3254,8 +3253,8 @@ exit_lbl:
 }
 #endif
 
-static inline void bnx2x_set_pbd_gso_e2(struct sk_buff *skb, u32 *parsing_data,
-					u32 xmit_type)
+static void bnx2x_set_pbd_gso_e2(struct sk_buff *skb, u32 *parsing_data,
+				 u32 xmit_type)
 {
 	*parsing_data |= (skb_shinfo(skb)->gso_size <<
 			      ETH_TX_PARSE_BD_E2_LSO_MSS_SHIFT) &
@@ -3272,13 +3271,13 @@ static inline void bnx2x_set_pbd_gso_e2(struct sk_buff *skb, u32 *parsing_data,
  * @pbd:	parse BD
  * @xmit_type:	xmit flags
  */
-static inline void bnx2x_set_pbd_gso(struct sk_buff *skb,
-				     struct eth_tx_parse_bd_e1x *pbd,
-				     u32 xmit_type)
+static void bnx2x_set_pbd_gso(struct sk_buff *skb,
+			      struct eth_tx_parse_bd_e1x *pbd,
+			      u32 xmit_type)
 {
 	pbd->lso_mss = cpu_to_le16(skb_shinfo(skb)->gso_size);
 	pbd->tcp_send_seq = bswab32(tcp_hdr(skb)->seq);
-	pbd->tcp_flags = pbd_tcp_flags(skb);
+	pbd->tcp_flags = pbd_tcp_flags(tcp_hdr(skb));
 
 	if (xmit_type & XMIT_GSO_V4) {
 		pbd->ip_id = bswab16(ip_hdr(skb)->id);
@@ -3305,15 +3304,15 @@ static inline void bnx2x_set_pbd_gso(struct sk_buff *skb,
  * @parsing_data:	data to be updated
  * @xmit_type:		xmit flags
  *
- * 57712 related
+ * 57712/578xx related
  */
-static inline  u8 bnx2x_set_pbd_csum_e2(struct bnx2x *bp, struct sk_buff *skb,
-					u32 *parsing_data, u32 xmit_type)
+static u8 bnx2x_set_pbd_csum_e2(struct bnx2x *bp, struct sk_buff *skb,
+				u32 *parsing_data, u32 xmit_type)
 {
 	*parsing_data |=
 		((((u8 *)skb_transport_header(skb) - skb->data) >> 1) <<
-		ETH_TX_PARSE_BD_E2_TCP_HDR_START_OFFSET_W_SHIFT) &
-		ETH_TX_PARSE_BD_E2_TCP_HDR_START_OFFSET_W;
+		ETH_TX_PARSE_BD_E2_L4_HDR_START_OFFSET_W_SHIFT) &
+		ETH_TX_PARSE_BD_E2_L4_HDR_START_OFFSET_W;
 
 	if (xmit_type & XMIT_CSUM_TCP) {
 		*parsing_data |= ((tcp_hdrlen(skb) / 4) <<
@@ -3328,17 +3327,14 @@ static inline  u8 bnx2x_set_pbd_csum_e2(struct bnx2x *bp, struct sk_buff *skb,
 	return skb_transport_header(skb) + sizeof(struct udphdr) - skb->data;
 }
 
-static inline void bnx2x_set_sbd_csum(struct bnx2x *bp, struct sk_buff *skb,
-	struct eth_tx_start_bd *tx_start_bd, u32 xmit_type)
+static void bnx2x_set_sbd_csum(struct bnx2x *bp, struct sk_buff *skb,
+			       struct eth_tx_start_bd *tx_start_bd,
+			       u32 xmit_type)
 {
 	tx_start_bd->bd_flags.as_bitfield |= ETH_TX_BD_FLAGS_L4_CSUM;
 
-	if (xmit_type & XMIT_CSUM_V4)
-		tx_start_bd->bd_flags.as_bitfield |=
-					ETH_TX_BD_FLAGS_IP_CSUM;
-	else
-		tx_start_bd->bd_flags.as_bitfield |=
-					ETH_TX_BD_FLAGS_IPV6;
+	if (xmit_type & XMIT_CSUM_V6)
+		tx_start_bd->bd_flags.as_bitfield |= ETH_TX_BD_FLAGS_IPV6;
 
 	if (!(xmit_type & XMIT_CSUM_TCP))
 		tx_start_bd->bd_flags.as_bitfield |= ETH_TX_BD_FLAGS_IS_UDP;
@@ -3352,9 +3348,9 @@ static inline void bnx2x_set_sbd_csum(struct bnx2x *bp, struct sk_buff *skb,
  * @pbd:	parse BD to be updated
  * @xmit_type:	xmit flags
  */
-static inline u8 bnx2x_set_pbd_csum(struct bnx2x *bp, struct sk_buff *skb,
-	struct eth_tx_parse_bd_e1x *pbd,
-	u32 xmit_type)
+static u8 bnx2x_set_pbd_csum(struct bnx2x *bp, struct sk_buff *skb,
+			     struct eth_tx_parse_bd_e1x *pbd,
+			     u32 xmit_type)
 {
 	u8 hlen = (skb_network_header(skb) - skb->data) >> 1;
 
@@ -3482,7 +3478,7 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 			mac_type = MULTICAST_ADDRESS;
 	}
 
-#if (MAX_SKB_FRAGS >= MAX_FETCH_BD - 3)
+#if (MAX_SKB_FRAGS >= MAX_FETCH_BD - BDS_PER_TX_PKT)
 	/* First, check if we need to linearize the skb (due to FW
 	   restrictions). No need to check fragmentation if page size > 8K
 	   (there will be no violation to FW restrictions) */
@@ -3530,12 +3526,9 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	first_bd = tx_start_bd;
 
 	tx_start_bd->bd_flags.as_bitfield = ETH_TX_BD_FLAGS_START_BD;
-	SET_FLAG(tx_start_bd->general_data,
-		 ETH_TX_START_BD_PARSE_NBDS,
-		 0);
 
-	/* header nbd */
-	SET_FLAG(tx_start_bd->general_data, ETH_TX_START_BD_HDR_NBDS, 1);
+	/* header nbd: indirectly zero other flags! */
+	tx_start_bd->general_data = 1 << ETH_TX_START_BD_HDR_NBDS_SHIFT;
 
 	/* remember the first BD of the packet */
 	tx_buf->first_bd = txdata->tx_bd_prod;
@@ -3555,19 +3548,16 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		/* when transmitting in a vf, start bd must hold the ethertype
 		 * for fw to enforce it
 		 */
-#ifndef BNX2X_STOP_ON_ERROR
-		if (IS_VF(bp)) {
-#endif
+		if (IS_VF(bp))
 			tx_start_bd->vlan_or_ethertype =
 				cpu_to_le16(ntohs(eth->h_proto));
-#ifndef BNX2X_STOP_ON_ERROR
-		} else {
+		else
 			/* used by FW for packet accounting */
 			tx_start_bd->vlan_or_ethertype = cpu_to_le16(pkt_prod);
-		}
-#endif
 	}
 
+	nbd = 2; /* start_bd + pbd + frags (updated when pages are mapped) */
+
 	/* turn on parsing and get a BD */
 	bd_prod = TX_BD(NEXT_TX_IDX(bd_prod));
 
@@ -3579,21 +3569,22 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		memset(pbd_e2, 0, sizeof(struct eth_tx_parse_bd_e2));
 		/* Set PBD in checksum offload case */
 		if (xmit_type & XMIT_CSUM)
+			/* Set PBD in checksum offload case w/o encapsulation */
 			hlen = bnx2x_set_pbd_csum_e2(bp, skb,
 						     &pbd_e2_parsing_data,
 						     xmit_type);
 
-		if (IS_MF_SI(bp) || IS_VF(bp)) {
-			/* fill in the MAC addresses in the PBD - for local
-			 * switching
-			 */
-			bnx2x_set_fw_mac_addr(&pbd_e2->src_mac_addr_hi,
-					      &pbd_e2->src_mac_addr_mid,
-					      &pbd_e2->src_mac_addr_lo,
+		/* Add the macs to the parsing BD this is a vf */
+		if (IS_VF(bp)) {
+			/* override GRE parameters in BD */
+			bnx2x_set_fw_mac_addr(&pbd_e2->data.mac_addr.src_hi,
+					      &pbd_e2->data.mac_addr.src_mid,
+					      &pbd_e2->data.mac_addr.src_lo,
 					      eth->h_source);
-			bnx2x_set_fw_mac_addr(&pbd_e2->dst_mac_addr_hi,
-					      &pbd_e2->dst_mac_addr_mid,
-					      &pbd_e2->dst_mac_addr_lo,
+
+			bnx2x_set_fw_mac_addr(&pbd_e2->data.mac_addr.dst_hi,
+					      &pbd_e2->data.mac_addr.dst_mid,
+					      &pbd_e2->data.mac_addr.dst_lo,
 					      eth->h_dest);
 		}
 
@@ -3615,14 +3606,13 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	/* Setup the data pointer of the first BD of the packet */
 	tx_start_bd->addr_hi = cpu_to_le32(U64_HI(mapping));
 	tx_start_bd->addr_lo = cpu_to_le32(U64_LO(mapping));
-	nbd = 2; /* start_bd + pbd + frags (updated when pages are mapped) */
 	tx_start_bd->nbytes = cpu_to_le16(skb_headlen(skb));
 	pkt_size = tx_start_bd->nbytes;
 
 	DP(NETIF_MSG_TX_QUEUED,
-	   "first bd @%p  addr (%x:%x)  nbd %d  nbytes %d  flags %x  vlan %x\n",
+	   "first bd @%p  addr (%x:%x)  nbytes %d  flags %x  vlan %x\n",
 	   tx_start_bd, tx_start_bd->addr_hi, tx_start_bd->addr_lo,
-	   le16_to_cpu(tx_start_bd->nbd), le16_to_cpu(tx_start_bd->nbytes),
+	   le16_to_cpu(tx_start_bd->nbytes),
 	   tx_start_bd->bd_flags.as_bitfield,
 	   le16_to_cpu(tx_start_bd->vlan_or_ethertype));
 
@@ -3635,10 +3625,12 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 
 		tx_start_bd->bd_flags.as_bitfield |= ETH_TX_BD_FLAGS_SW_LSO;
 
-		if (unlikely(skb_headlen(skb) > hlen))
+		if (unlikely(skb_headlen(skb) > hlen)) {
+			nbd++;
 			bd_prod = bnx2x_tx_split(bp, txdata, tx_buf,
 						 &tx_start_bd, hlen,
-						 bd_prod, ++nbd);
+						 bd_prod);
+		}
 		if (!CHIP_IS_E1x(bp))
 			bnx2x_set_pbd_gso_e2(skb, &pbd_e2_parsing_data,
 					     xmit_type);
@@ -3728,9 +3720,13 @@ netdev_tx_t bnx2x_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	if (pbd_e2)
 		DP(NETIF_MSG_TX_QUEUED,
 		   "PBD (E2) @%p  dst %x %x %x src %x %x %x parsing_data %x\n",
-		   pbd_e2, pbd_e2->dst_mac_addr_hi, pbd_e2->dst_mac_addr_mid,
-		   pbd_e2->dst_mac_addr_lo, pbd_e2->src_mac_addr_hi,
-		   pbd_e2->src_mac_addr_mid, pbd_e2->src_mac_addr_lo,
+		   pbd_e2,
+		   pbd_e2->data.mac_addr.dst_hi,
+		   pbd_e2->data.mac_addr.dst_mid,
+		   pbd_e2->data.mac_addr.dst_lo,
+		   pbd_e2->data.mac_addr.src_hi,
+		   pbd_e2->data.mac_addr.src_mid,
+		   pbd_e2->data.mac_addr.src_lo,
 		   pbd_e2->parsing_data);
 	DP(NETIF_MSG_TX_QUEUED, "doorbell: nbd %d  bd %u\n", nbd, bd_prod);
 
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_fw_defs.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_fw_defs.h
index e5f8083..40f22c6 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_fw_defs.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_fw_defs.h
@@ -30,31 +30,31 @@
 	* IRO[138].m2) + ((sbId) * IRO[138].m3))
 #define CSTORM_IGU_MODE_OFFSET (IRO[157].base)
 #define CSTORM_ISCSI_CQ_SIZE_OFFSET(pfId) \
-	(IRO[316].base + ((pfId) * IRO[316].m1))
-#define CSTORM_ISCSI_CQ_SQN_SIZE_OFFSET(pfId) \
 	(IRO[317].base + ((pfId) * IRO[317].m1))
+#define CSTORM_ISCSI_CQ_SQN_SIZE_OFFSET(pfId) \
+	(IRO[318].base + ((pfId) * IRO[318].m1))
 #define CSTORM_ISCSI_EQ_CONS_OFFSET(pfId, iscsiEqId) \
-	(IRO[309].base + ((pfId) * IRO[309].m1) + ((iscsiEqId) * IRO[309].m2))
+	(IRO[310].base + ((pfId) * IRO[310].m1) + ((iscsiEqId) * IRO[310].m2))
 #define CSTORM_ISCSI_EQ_NEXT_EQE_ADDR_OFFSET(pfId, iscsiEqId) \
-	(IRO[311].base + ((pfId) * IRO[311].m1) + ((iscsiEqId) * IRO[311].m2))
+	(IRO[312].base + ((pfId) * IRO[312].m1) + ((iscsiEqId) * IRO[312].m2))
 #define CSTORM_ISCSI_EQ_NEXT_PAGE_ADDR_OFFSET(pfId, iscsiEqId) \
-	(IRO[310].base + ((pfId) * IRO[310].m1) + ((iscsiEqId) * IRO[310].m2))
+	(IRO[311].base + ((pfId) * IRO[311].m1) + ((iscsiEqId) * IRO[311].m2))
 #define CSTORM_ISCSI_EQ_NEXT_PAGE_ADDR_VALID_OFFSET(pfId, iscsiEqId) \
-	(IRO[312].base + ((pfId) * IRO[312].m1) + ((iscsiEqId) * IRO[312].m2))
+	(IRO[313].base + ((pfId) * IRO[313].m1) + ((iscsiEqId) * IRO[313].m2))
 #define CSTORM_ISCSI_EQ_PROD_OFFSET(pfId, iscsiEqId) \
-	(IRO[308].base + ((pfId) * IRO[308].m1) + ((iscsiEqId) * IRO[308].m2))
+	(IRO[309].base + ((pfId) * IRO[309].m1) + ((iscsiEqId) * IRO[309].m2))
 #define CSTORM_ISCSI_EQ_SB_INDEX_OFFSET(pfId, iscsiEqId) \
-	(IRO[314].base + ((pfId) * IRO[314].m1) + ((iscsiEqId) * IRO[314].m2))
+	(IRO[315].base + ((pfId) * IRO[315].m1) + ((iscsiEqId) * IRO[315].m2))
 #define CSTORM_ISCSI_EQ_SB_NUM_OFFSET(pfId, iscsiEqId) \
-	(IRO[313].base + ((pfId) * IRO[313].m1) + ((iscsiEqId) * IRO[313].m2))
+	(IRO[314].base + ((pfId) * IRO[314].m1) + ((iscsiEqId) * IRO[314].m2))
 #define CSTORM_ISCSI_HQ_SIZE_OFFSET(pfId) \
-	(IRO[315].base + ((pfId) * IRO[315].m1))
+	(IRO[316].base + ((pfId) * IRO[316].m1))
 #define CSTORM_ISCSI_NUM_OF_TASKS_OFFSET(pfId) \
-	(IRO[307].base + ((pfId) * IRO[307].m1))
+	(IRO[308].base + ((pfId) * IRO[308].m1))
 #define CSTORM_ISCSI_PAGE_SIZE_LOG_OFFSET(pfId) \
-	(IRO[306].base + ((pfId) * IRO[306].m1))
+	(IRO[307].base + ((pfId) * IRO[307].m1))
 #define CSTORM_ISCSI_PAGE_SIZE_OFFSET(pfId) \
-	(IRO[305].base + ((pfId) * IRO[305].m1))
+	(IRO[306].base + ((pfId) * IRO[306].m1))
 #define CSTORM_RECORD_SLOW_PATH_OFFSET(funcId) \
 	(IRO[151].base + ((funcId) * IRO[151].m1))
 #define CSTORM_SP_STATUS_BLOCK_DATA_OFFSET(pfId) \
@@ -114,7 +114,7 @@
 #define TSTORM_ISCSI_RQ_SIZE_OFFSET(pfId) \
 	(IRO[268].base + ((pfId) * IRO[268].m1))
 #define TSTORM_ISCSI_TCP_LOCAL_ADV_WND_OFFSET(pfId) \
-	(IRO[277].base + ((pfId) * IRO[277].m1))
+	(IRO[278].base + ((pfId) * IRO[278].m1))
 #define TSTORM_ISCSI_TCP_VARS_FLAGS_OFFSET(pfId) \
 	(IRO[264].base + ((pfId) * IRO[264].m1))
 #define TSTORM_ISCSI_TCP_VARS_LSB_LOCAL_MAC_ADDR_OFFSET(pfId) \
@@ -136,35 +136,32 @@
 #define USTORM_ASSERT_LIST_INDEX_OFFSET	(IRO[177].base)
 #define USTORM_ASSERT_LIST_OFFSET(assertListEntry) \
 	(IRO[176].base + ((assertListEntry) * IRO[176].m1))
-#define USTORM_CQE_PAGE_NEXT_OFFSET(portId, clientId) \
-	(IRO[205].base + ((portId) * IRO[205].m1) + ((clientId) * \
-	IRO[205].m2))
 #define USTORM_ETH_PAUSE_ENABLED_OFFSET(portId) \
 	(IRO[183].base + ((portId) * IRO[183].m1))
 #define USTORM_FCOE_EQ_PROD_OFFSET(pfId) \
-	(IRO[318].base + ((pfId) * IRO[318].m1))
+	(IRO[319].base + ((pfId) * IRO[319].m1))
 #define USTORM_FUNC_EN_OFFSET(funcId) \
 	(IRO[178].base + ((funcId) * IRO[178].m1))
 #define USTORM_ISCSI_CQ_SIZE_OFFSET(pfId) \
-	(IRO[282].base + ((pfId) * IRO[282].m1))
-#define USTORM_ISCSI_CQ_SQN_SIZE_OFFSET(pfId) \
 	(IRO[283].base + ((pfId) * IRO[283].m1))
+#define USTORM_ISCSI_CQ_SQN_SIZE_OFFSET(pfId) \
+	(IRO[284].base + ((pfId) * IRO[284].m1))
 #define USTORM_ISCSI_ERROR_BITMAP_OFFSET(pfId) \
-	(IRO[287].base + ((pfId) * IRO[287].m1))
+	(IRO[288].base + ((pfId) * IRO[288].m1))
 #define USTORM_ISCSI_GLOBAL_BUF_PHYS_ADDR_OFFSET(pfId) \
-	(IRO[284].base + ((pfId) * IRO[284].m1))
+	(IRO[285].base + ((pfId) * IRO[285].m1))
 #define USTORM_ISCSI_NUM_OF_TASKS_OFFSET(pfId) \
-	(IRO[280].base + ((pfId) * IRO[280].m1))
+	(IRO[281].base + ((pfId) * IRO[281].m1))
 #define USTORM_ISCSI_PAGE_SIZE_LOG_OFFSET(pfId) \
-	(IRO[279].base + ((pfId) * IRO[279].m1))
+	(IRO[280].base + ((pfId) * IRO[280].m1))
 #define USTORM_ISCSI_PAGE_SIZE_OFFSET(pfId) \
-	(IRO[278].base + ((pfId) * IRO[278].m1))
+	(IRO[279].base + ((pfId) * IRO[279].m1))
 #define USTORM_ISCSI_R2TQ_SIZE_OFFSET(pfId) \
-	(IRO[281].base + ((pfId) * IRO[281].m1))
+	(IRO[282].base + ((pfId) * IRO[282].m1))
 #define USTORM_ISCSI_RQ_BUFFER_SIZE_OFFSET(pfId) \
-	(IRO[285].base + ((pfId) * IRO[285].m1))
-#define USTORM_ISCSI_RQ_SIZE_OFFSET(pfId) \
 	(IRO[286].base + ((pfId) * IRO[286].m1))
+#define USTORM_ISCSI_RQ_SIZE_OFFSET(pfId) \
+	(IRO[287].base + ((pfId) * IRO[287].m1))
 #define USTORM_MEM_WORKAROUND_ADDRESS_OFFSET(pfId) \
 	(IRO[182].base + ((pfId) * IRO[182].m1))
 #define USTORM_RECORD_SLOW_PATH_OFFSET(funcId) \
@@ -190,39 +187,39 @@
 #define XSTORM_FUNC_EN_OFFSET(funcId) \
 	(IRO[47].base + ((funcId) * IRO[47].m1))
 #define XSTORM_ISCSI_HQ_SIZE_OFFSET(pfId) \
-	(IRO[295].base + ((pfId) * IRO[295].m1))
+	(IRO[296].base + ((pfId) * IRO[296].m1))
 #define XSTORM_ISCSI_LOCAL_MAC_ADDR0_OFFSET(pfId) \
-	(IRO[298].base + ((pfId) * IRO[298].m1))
-#define XSTORM_ISCSI_LOCAL_MAC_ADDR1_OFFSET(pfId) \
 	(IRO[299].base + ((pfId) * IRO[299].m1))
-#define XSTORM_ISCSI_LOCAL_MAC_ADDR2_OFFSET(pfId) \
+#define XSTORM_ISCSI_LOCAL_MAC_ADDR1_OFFSET(pfId) \
 	(IRO[300].base + ((pfId) * IRO[300].m1))
-#define XSTORM_ISCSI_LOCAL_MAC_ADDR3_OFFSET(pfId) \
+#define XSTORM_ISCSI_LOCAL_MAC_ADDR2_OFFSET(pfId) \
 	(IRO[301].base + ((pfId) * IRO[301].m1))
-#define XSTORM_ISCSI_LOCAL_MAC_ADDR4_OFFSET(pfId) \
+#define XSTORM_ISCSI_LOCAL_MAC_ADDR3_OFFSET(pfId) \
 	(IRO[302].base + ((pfId) * IRO[302].m1))
-#define XSTORM_ISCSI_LOCAL_MAC_ADDR5_OFFSET(pfId) \
+#define XSTORM_ISCSI_LOCAL_MAC_ADDR4_OFFSET(pfId) \
 	(IRO[303].base + ((pfId) * IRO[303].m1))
-#define XSTORM_ISCSI_LOCAL_VLAN_OFFSET(pfId) \
+#define XSTORM_ISCSI_LOCAL_MAC_ADDR5_OFFSET(pfId) \
 	(IRO[304].base + ((pfId) * IRO[304].m1))
+#define XSTORM_ISCSI_LOCAL_VLAN_OFFSET(pfId) \
+	(IRO[305].base + ((pfId) * IRO[305].m1))
 #define XSTORM_ISCSI_NUM_OF_TASKS_OFFSET(pfId) \
-	(IRO[294].base + ((pfId) * IRO[294].m1))
+	(IRO[295].base + ((pfId) * IRO[295].m1))
 #define XSTORM_ISCSI_PAGE_SIZE_LOG_OFFSET(pfId) \
-	(IRO[293].base + ((pfId) * IRO[293].m1))
+	(IRO[294].base + ((pfId) * IRO[294].m1))
 #define XSTORM_ISCSI_PAGE_SIZE_OFFSET(pfId) \
-	(IRO[292].base + ((pfId) * IRO[292].m1))
+	(IRO[293].base + ((pfId) * IRO[293].m1))
 #define XSTORM_ISCSI_R2TQ_SIZE_OFFSET(pfId) \
-	(IRO[297].base + ((pfId) * IRO[297].m1))
+	(IRO[298].base + ((pfId) * IRO[298].m1))
 #define XSTORM_ISCSI_SQ_SIZE_OFFSET(pfId) \
-	(IRO[296].base + ((pfId) * IRO[296].m1))
+	(IRO[297].base + ((pfId) * IRO[297].m1))
 #define XSTORM_ISCSI_TCP_VARS_ADV_WND_SCL_OFFSET(pfId) \
-	(IRO[291].base + ((pfId) * IRO[291].m1))
+	(IRO[292].base + ((pfId) * IRO[292].m1))
 #define XSTORM_ISCSI_TCP_VARS_FLAGS_OFFSET(pfId) \
-	(IRO[290].base + ((pfId) * IRO[290].m1))
+	(IRO[291].base + ((pfId) * IRO[291].m1))
 #define XSTORM_ISCSI_TCP_VARS_TOS_OFFSET(pfId) \
-	(IRO[289].base + ((pfId) * IRO[289].m1))
+	(IRO[290].base + ((pfId) * IRO[290].m1))
 #define XSTORM_ISCSI_TCP_VARS_TTL_OFFSET(pfId) \
-	(IRO[288].base + ((pfId) * IRO[288].m1))
+	(IRO[289].base + ((pfId) * IRO[289].m1))
 #define XSTORM_RATE_SHAPING_PER_VN_VARS_OFFSET(pfId) \
 	(IRO[44].base + ((pfId) * IRO[44].m1))
 #define XSTORM_RECORD_SLOW_PATH_OFFSET(funcId) \
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
index a7a3504..12f00a4 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_hsi.h
@@ -114,6 +114,10 @@ struct license_key {
 #define EPIO_CFG_EPIO30                     0x0000001f
 #define EPIO_CFG_EPIO31                     0x00000020
 
+struct mac_addr {
+	u32 upper;
+	u32 lower;
+};
 
 struct shared_hw_cfg {			 /* NVRAM Offset */
 	/* Up to 16 bytes of NULL-terminated string */
@@ -2836,8 +2840,8 @@ struct afex_stats {
 
 #define BCM_5710_FW_MAJOR_VERSION			7
 #define BCM_5710_FW_MINOR_VERSION			8
-#define BCM_5710_FW_REVISION_VERSION		2
-#define BCM_5710_FW_ENGINEERING_VERSION			0
+#define BCM_5710_FW_REVISION_VERSION		17
+#define BCM_5710_FW_ENGINEERING_VERSION		0
 #define BCM_5710_FW_COMPILE_FLAGS			1
 
 
@@ -3528,11 +3532,14 @@ struct client_init_tx_data {
 #define CLIENT_INIT_TX_DATA_BCAST_ACCEPT_ALL_SHIFT 2
 #define CLIENT_INIT_TX_DATA_ACCEPT_ANY_VLAN (0x1<<3)
 #define CLIENT_INIT_TX_DATA_ACCEPT_ANY_VLAN_SHIFT 3
-#define CLIENT_INIT_TX_DATA_RESERVED1 (0xFFF<<4)
-#define CLIENT_INIT_TX_DATA_RESERVED1_SHIFT 4
+#define CLIENT_INIT_TX_DATA_RESERVED0 (0xFFF<<4)
+#define CLIENT_INIT_TX_DATA_RESERVED0_SHIFT 4
 	u8 default_vlan_flg;
 	u8 force_default_pri_flg;
-	__le32 reserved3;
+	u8 tunnel_lso_inc_ip_id;
+	u8 refuse_outband_vlan_flg;
+	u8 tunnel_non_lso_pcsum_location;
+	u8 reserved1;
 };
 
 /*
@@ -3566,6 +3573,11 @@ struct client_update_ramrod_data {
 	__le16 silent_vlan_mask;
 	u8 silent_vlan_removal_flg;
 	u8 silent_vlan_change_flg;
+	u8 refuse_outband_vlan_flg;
+	u8 refuse_outband_vlan_change_flg;
+	u8 tx_switching_flg;
+	u8 tx_switching_change_flg;
+	__le32 reserved1;
 	__le32 echo;
 };
 
@@ -3635,7 +3647,8 @@ struct eth_classify_header {
  */
 struct eth_classify_mac_cmd {
 	struct eth_classify_cmd_header header;
-	__le32 reserved0;
+	__le16 reserved0;
+	__le16 inner_mac;
 	__le16 mac_lsb;
 	__le16 mac_mid;
 	__le16 mac_msb;
@@ -3648,7 +3661,8 @@ struct eth_classify_mac_cmd {
  */
 struct eth_classify_pair_cmd {
 	struct eth_classify_cmd_header header;
-	__le32 reserved0;
+	__le16 reserved0;
+	__le16 inner_mac;
 	__le16 mac_lsb;
 	__le16 mac_mid;
 	__le16 mac_msb;
@@ -3870,8 +3884,68 @@ struct eth_halt_ramrod_data {
 
 
 /*
- * Command for setting multicast classification for a client
+ * destination and source mac address.
+ */
+struct eth_mac_addresses {
+#if defined(__BIG_ENDIAN)
+	__le16 dst_mid;
+	__le16 dst_lo;
+#elif defined(__LITTLE_ENDIAN)
+	__le16 dst_lo;
+	__le16 dst_mid;
+#endif
+#if defined(__BIG_ENDIAN)
+	__le16 src_lo;
+	__le16 dst_hi;
+#elif defined(__LITTLE_ENDIAN)
+	__le16 dst_hi;
+	__le16 src_lo;
+#endif
+#if defined(__BIG_ENDIAN)
+	__le16 src_hi;
+	__le16 src_mid;
+#elif defined(__LITTLE_ENDIAN)
+	__le16 src_mid;
+	__le16 src_hi;
+#endif
+};
+
+/* tunneling related data */
+struct eth_tunnel_data {
+#if defined(__BIG_ENDIAN)
+	__le16 dst_mid;
+	__le16 dst_lo;
+#elif defined(__LITTLE_ENDIAN)
+	__le16 dst_lo;
+	__le16 dst_mid;
+#endif
+#if defined(__BIG_ENDIAN)
+	__le16 reserved0;
+	__le16 dst_hi;
+#elif defined(__LITTLE_ENDIAN)
+	__le16 dst_hi;
+	__le16 reserved0;
+#endif
+#if defined(__BIG_ENDIAN)
+	u8 reserved1;
+	u8 ip_hdr_start_inner_w;
+	__le16 pseudo_csum;
+#elif defined(__LITTLE_ENDIAN)
+	__le16 pseudo_csum;
+	u8 ip_hdr_start_inner_w;
+	u8 reserved1;
+#endif
+};
+
+/* union for mac addresses and for tunneling data.
+ * considered as tunneling data only if (tunnel_exist == 1).
  */
+union eth_mac_addr_or_tunnel_data {
+	struct eth_mac_addresses mac_addr;
+	struct eth_tunnel_data tunnel_data;
+};
+
+/*Command for setting multicast classification for a client */
 struct eth_multicast_rules_cmd {
 	u8 cmd_general_data;
 #define ETH_MULTICAST_RULES_CMD_RX_CMD (0x1<<0)
@@ -3889,7 +3963,6 @@ struct eth_multicast_rules_cmd {
 	struct regpair reserved3;
 };
 
-
 /*
  * parameters for multicast classification ramrod
  */
@@ -3898,7 +3971,6 @@ struct eth_multicast_rules_ramrod_data {
 	struct eth_multicast_rules_cmd rules[MULTICAST_RULES_COUNT];
 };
 
-
 /*
  * Place holder for ramrods protocol specific data
  */
@@ -3962,11 +4034,14 @@ struct eth_rss_update_ramrod_data {
 #define ETH_RSS_UPDATE_RAMROD_DATA_IPV6_TCP_CAPABILITY_SHIFT 4
 #define ETH_RSS_UPDATE_RAMROD_DATA_IPV6_UDP_CAPABILITY (0x1<<5)
 #define ETH_RSS_UPDATE_RAMROD_DATA_IPV6_UDP_CAPABILITY_SHIFT 5
+#define ETH_RSS_UPDATE_RAMROD_DATA_EN_5_TUPLE_CAPABILITY (0x1<<6)
+#define ETH_RSS_UPDATE_RAMROD_DATA_EN_5_TUPLE_CAPABILITY_SHIFT 6
 #define ETH_RSS_UPDATE_RAMROD_DATA_UPDATE_RSS_KEY (0x1<<7)
 #define ETH_RSS_UPDATE_RAMROD_DATA_UPDATE_RSS_KEY_SHIFT 7
 	u8 rss_result_mask;
 	u8 rss_mode;
-	__le32 __reserved2;
+	__le16 udp_4tuple_dst_port_mask;
+	__le16 udp_4tuple_dst_port_value;
 	u8 indirection_table[T_ETH_INDIRECTION_TABLE_SIZE];
 	__le32 rss_key[T_ETH_RSS_KEY];
 	__le32 echo;
@@ -4130,6 +4205,23 @@ enum eth_tpa_update_command {
 	MAX_ETH_TPA_UPDATE_COMMAND
 };
 
+/* In case of LSO over IPv4 tunnel, whether to increment
+ * IP ID on external IP header or internal IP header
+ */
+enum eth_tunnel_lso_inc_ip_id {
+	EXT_HEADER,
+	INT_HEADER,
+	MAX_ETH_TUNNEL_LSO_INC_IP_ID
+};
+
+/* In case tunnel exist and L4 checksum offload,
+ * the pseudo checksum location, on packet or on BD.
+ */
+enum eth_tunnel_non_lso_pcsum_location {
+	PCSUM_ON_PKT,
+	PCSUM_ON_BD,
+	MAX_ETH_TUNNEL_NON_LSO_PCSUM_LOCATION
+};
 
 /*
  * Tx regular BD structure
@@ -4181,8 +4273,8 @@ struct eth_tx_start_bd {
 #define ETH_TX_START_BD_FORCE_VLAN_MODE_SHIFT 4
 #define ETH_TX_START_BD_PARSE_NBDS (0x3<<5)
 #define ETH_TX_START_BD_PARSE_NBDS_SHIFT 5
-#define ETH_TX_START_BD_RESREVED (0x1<<7)
-#define ETH_TX_START_BD_RESREVED_SHIFT 7
+#define ETH_TX_START_BD_TUNNEL_EXIST (0x1<<7)
+#define ETH_TX_START_BD_TUNNEL_EXIST_SHIFT 7
 };
 
 /*
@@ -4231,15 +4323,10 @@ struct eth_tx_parse_bd_e1x {
  * Tx parsing BD structure for ETH E2
  */
 struct eth_tx_parse_bd_e2 {
-	__le16 dst_mac_addr_lo;
-	__le16 dst_mac_addr_mid;
-	__le16 dst_mac_addr_hi;
-	__le16 src_mac_addr_lo;
-	__le16 src_mac_addr_mid;
-	__le16 src_mac_addr_hi;
+	union eth_mac_addr_or_tunnel_data data;
 	__le32 parsing_data;
-#define ETH_TX_PARSE_BD_E2_TCP_HDR_START_OFFSET_W (0x7FF<<0)
-#define ETH_TX_PARSE_BD_E2_TCP_HDR_START_OFFSET_W_SHIFT 0
+#define ETH_TX_PARSE_BD_E2_L4_HDR_START_OFFSET_W (0x7FF<<0)
+#define ETH_TX_PARSE_BD_E2_L4_HDR_START_OFFSET_W_SHIFT 0
 #define ETH_TX_PARSE_BD_E2_TCP_HDR_LENGTH_DW (0xF<<11)
 #define ETH_TX_PARSE_BD_E2_TCP_HDR_LENGTH_DW_SHIFT 11
 #define ETH_TX_PARSE_BD_E2_IPV6_WITH_EXT_HDR (0x1<<15)
@@ -4251,8 +4338,51 @@ struct eth_tx_parse_bd_e2 {
 };
 
 /*
- * The last BD in the BD memory will hold a pointer to the next BD memory
+ * Tx 2nd parsing BD structure for ETH packet
  */
+struct eth_tx_parse_2nd_bd {
+	__le16 global_data;
+#define ETH_TX_PARSE_2ND_BD_IP_HDR_START_OUTER_W (0xF<<0)
+#define ETH_TX_PARSE_2ND_BD_IP_HDR_START_OUTER_W_SHIFT 0
+#define ETH_TX_PARSE_2ND_BD_IP_HDR_TYPE_OUTER (0x1<<4)
+#define ETH_TX_PARSE_2ND_BD_IP_HDR_TYPE_OUTER_SHIFT 4
+#define ETH_TX_PARSE_2ND_BD_LLC_SNAP_EN (0x1<<5)
+#define ETH_TX_PARSE_2ND_BD_LLC_SNAP_EN_SHIFT 5
+#define ETH_TX_PARSE_2ND_BD_NS_FLG (0x1<<6)
+#define ETH_TX_PARSE_2ND_BD_NS_FLG_SHIFT 6
+#define ETH_TX_PARSE_2ND_BD_TUNNEL_UDP_EXIST (0x1<<7)
+#define ETH_TX_PARSE_2ND_BD_TUNNEL_UDP_EXIST_SHIFT 7
+#define ETH_TX_PARSE_2ND_BD_IP_HDR_LEN_OUTER_W (0x1F<<8)
+#define ETH_TX_PARSE_2ND_BD_IP_HDR_LEN_OUTER_W_SHIFT 8
+#define ETH_TX_PARSE_2ND_BD_RESERVED0 (0x7<<13)
+#define ETH_TX_PARSE_2ND_BD_RESERVED0_SHIFT 13
+	__le16 reserved1;
+	u8 tcp_flags;
+#define ETH_TX_PARSE_2ND_BD_FIN_FLG (0x1<<0)
+#define ETH_TX_PARSE_2ND_BD_FIN_FLG_SHIFT 0
+#define ETH_TX_PARSE_2ND_BD_SYN_FLG (0x1<<1)
+#define ETH_TX_PARSE_2ND_BD_SYN_FLG_SHIFT 1
+#define ETH_TX_PARSE_2ND_BD_RST_FLG (0x1<<2)
+#define ETH_TX_PARSE_2ND_BD_RST_FLG_SHIFT 2
+#define ETH_TX_PARSE_2ND_BD_PSH_FLG (0x1<<3)
+#define ETH_TX_PARSE_2ND_BD_PSH_FLG_SHIFT 3
+#define ETH_TX_PARSE_2ND_BD_ACK_FLG (0x1<<4)
+#define ETH_TX_PARSE_2ND_BD_ACK_FLG_SHIFT 4
+#define ETH_TX_PARSE_2ND_BD_URG_FLG (0x1<<5)
+#define ETH_TX_PARSE_2ND_BD_URG_FLG_SHIFT 5
+#define ETH_TX_PARSE_2ND_BD_ECE_FLG (0x1<<6)
+#define ETH_TX_PARSE_2ND_BD_ECE_FLG_SHIFT 6
+#define ETH_TX_PARSE_2ND_BD_CWR_FLG (0x1<<7)
+#define ETH_TX_PARSE_2ND_BD_CWR_FLG_SHIFT 7
+	u8 reserved2;
+	u8 tunnel_udp_hdr_start_w;
+	u8 fw_ip_hdr_to_payload_w;
+	__le16 fw_ip_csum_wo_len_flags_frag;
+	__le16 hw_ip_id;
+	__le32 tcp_send_seq;
+};
+
+/* The last BD in the BD memory will hold a pointer to the next BD memory */
 struct eth_tx_next_bd {
 	__le32 addr_lo;
 	__le32 addr_hi;
@@ -4267,6 +4397,7 @@ union eth_tx_bd_types {
 	struct eth_tx_bd reg_bd;
 	struct eth_tx_parse_bd_e1x parse_bd_e1x;
 	struct eth_tx_parse_bd_e2 parse_bd_e2;
+	struct eth_tx_parse_2nd_bd parse_2nd_bd;
 	struct eth_tx_next_bd next_bd;
 };
 
@@ -4678,10 +4809,10 @@ enum common_spqe_cmd_id {
 	RAMROD_CMD_ID_COMMON_STOP_TRAFFIC,
 	RAMROD_CMD_ID_COMMON_START_TRAFFIC,
 	RAMROD_CMD_ID_COMMON_AFEX_VIF_LISTS,
+	RAMROD_CMD_ID_COMMON_SET_TIMESYNC,
 	MAX_COMMON_SPQE_CMD_ID
 };
 
-
 /*
  * Per-protocol connection types
  */
@@ -4878,7 +5009,7 @@ struct vf_flr_event_data {
  */
 struct malicious_vf_event_data {
 	u8 vf_id;
-	u8 reserved0;
+	u8 err_id;
 	u16 reserved1;
 	u32 reserved2;
 	u32 reserved3;
@@ -4984,10 +5115,10 @@ enum event_ring_opcode {
 	EVENT_RING_OPCODE_CLASSIFICATION_RULES,
 	EVENT_RING_OPCODE_FILTERS_RULES,
 	EVENT_RING_OPCODE_MULTICAST_RULES,
+	EVENT_RING_OPCODE_SET_TIMESYNC,
 	MAX_EVENT_RING_OPCODE
 };
 
-
 /*
  * Modes for fairness algorithm
  */
@@ -5025,14 +5156,18 @@ struct flow_control_configuration {
  */
 struct function_start_data {
 	u8 function_mode;
-	u8 reserved;
+	u8 allow_npar_tx_switching;
 	__le16 sd_vlan_tag;
 	__le16 vif_id;
 	u8 path_id;
 	u8 network_cos_mode;
+	u8 dmae_cmd_id;
+	u8 gre_tunnel_mode;
+	u8 gre_tunnel_rss;
+	u8 nvgre_clss_en;
+	__le16 reserved1[2];
 };
 
-
 struct function_update_data {
 	u8 vif_id_change_flg;
 	u8 afex_default_vlan_change_flg;
@@ -5042,14 +5177,19 @@ struct function_update_data {
 	__le16 afex_default_vlan;
 	u8 allowed_priorities;
 	u8 network_cos_mode;
+	u8 lb_mode_en_change_flg;
 	u8 lb_mode_en;
 	u8 tx_switch_suspend_change_flg;
 	u8 tx_switch_suspend;
 	u8 echo;
-	__le16 reserved1;
+	u8 reserved1;
+	u8 update_gre_cfg_flg;
+	u8 gre_tunnel_mode;
+	u8 gre_tunnel_rss;
+	u8 nvgre_clss_en;
+	u32 reserved3;
 };
 
-
 /*
  * FW version stored in the Xstorm RAM
  */
@@ -5076,6 +5216,22 @@ struct fw_version {
 #define __FW_VERSION_RESERVED_SHIFT 4
 };
 
+/* GRE RSS Mode */
+enum gre_rss_mode {
+	GRE_OUTER_HEADERS_RSS,
+	GRE_INNER_HEADERS_RSS,
+	NVGRE_KEY_ENTROPY_RSS,
+	MAX_GRE_RSS_MODE
+};
+
+/* GRE Tunnel Mode */
+enum gre_tunnel_type {
+	NO_GRE_TUNNEL,
+	NVGRE_TUNNEL,
+	L2GRE_TUNNEL,
+	IPGRE_TUNNEL,
+	MAX_GRE_TUNNEL_TYPE
+};
 
 /*
  * Dynamic Host-Coalescing - Driver(host) counters
@@ -5239,6 +5395,26 @@ enum ip_ver {
 	MAX_IP_VER
 };
 
+/*
+ * Malicious VF error ID
+ */
+enum malicious_vf_error_id {
+	VF_PF_CHANNEL_NOT_READY,
+	ETH_ILLEGAL_BD_LENGTHS,
+	ETH_PACKET_TOO_SHORT,
+	ETH_PAYLOAD_TOO_BIG,
+	ETH_ILLEGAL_ETH_TYPE,
+	ETH_ILLEGAL_LSO_HDR_LEN,
+	ETH_TOO_MANY_BDS,
+	ETH_ZERO_HDR_NBDS,
+	ETH_START_BD_NOT_SET,
+	ETH_ILLEGAL_PARSE_NBDS,
+	ETH_IPV6_AND_CHECKSUM,
+	ETH_VLAN_FLG_INCORRECT,
+	ETH_ILLEGAL_LSO_MSS,
+	ETH_TUNNEL_NOT_SUPPORTED,
+	MAX_MALICIOUS_VF_ERROR_ID
+};
 
 /*
  * Multi-function modes
@@ -5383,7 +5559,6 @@ struct protocol_common_spe {
 	union protocol_common_specific_data data;
 };
 
-
 /*
  * The send queue element
  */
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index e5662a1..c7df223 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -2953,14 +2953,15 @@ static unsigned long bnx2x_get_common_flags(struct bnx2x *bp,
 	__set_bit(BNX2X_Q_FLG_ACTIVE, &flags);
 
 	/* tx only connections collect statistics (on the same index as the
-	 *  parent connection). The statistics are zeroed when the parent
-	 *  connection is initialized.
+	 * parent connection). The statistics are zeroed when the parent
+	 * connection is initialized.
 	 */
 
 	__set_bit(BNX2X_Q_FLG_STATS, &flags);
 	if (zero_stats)
 		__set_bit(BNX2X_Q_FLG_ZERO_STATS, &flags);
 
+	__set_bit(BNX2X_Q_FLG_PCSUM_ON_PKT, &flags);
 
 #ifdef BNX2X_STOP_ON_ERROR
 	__set_bit(BNX2X_Q_FLG_TX_SEC, &flags);
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
index 6b03acd..66ab259 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.c
@@ -476,7 +476,8 @@ static int bnx2x_check_mac_add(struct bnx2x *bp,
 
 	/* Check if a requested MAC already exists */
 	list_for_each_entry(pos, &o->head, link)
-		if (!memcmp(data->mac.mac, pos->u.mac.mac, ETH_ALEN))
+		if (!memcmp(data->mac.mac, pos->u.mac.mac, ETH_ALEN) &&
+		    (data->mac.is_inner_mac == pos->u.mac.is_inner_mac))
 			return -EEXIST;
 
 	return 0;
@@ -509,7 +510,9 @@ static int bnx2x_check_vlan_mac_add(struct bnx2x *bp,
 	list_for_each_entry(pos, &o->head, link)
 		if ((data->vlan_mac.vlan == pos->u.vlan_mac.vlan) &&
 		    (!memcmp(data->vlan_mac.mac, pos->u.vlan_mac.mac,
-			     ETH_ALEN)))
+				  ETH_ALEN)) &&
+		    (data->vlan_mac.is_inner_mac ==
+		     pos->u.vlan_mac.is_inner_mac))
 			return -EEXIST;
 
 	return 0;
@@ -527,7 +530,8 @@ static struct bnx2x_vlan_mac_registry_elem *
 	DP(BNX2X_MSG_SP, "Checking MAC %pM for DEL command\n", data->mac.mac);
 
 	list_for_each_entry(pos, &o->head, link)
-		if (!memcmp(data->mac.mac, pos->u.mac.mac, ETH_ALEN))
+		if ((!memcmp(data->mac.mac, pos->u.mac.mac, ETH_ALEN)) &&
+		    (data->mac.is_inner_mac == pos->u.mac.is_inner_mac))
 			return pos;
 
 	return NULL;
@@ -562,7 +566,9 @@ static struct bnx2x_vlan_mac_registry_elem *
 	list_for_each_entry(pos, &o->head, link)
 		if ((data->vlan_mac.vlan == pos->u.vlan_mac.vlan) &&
 		    (!memcmp(data->vlan_mac.mac, pos->u.vlan_mac.mac,
-			     ETH_ALEN)))
+			     ETH_ALEN)) &&
+		    (data->vlan_mac.is_inner_mac ==
+		     pos->u.vlan_mac.is_inner_mac))
 			return pos;
 
 	return NULL;
@@ -759,6 +765,8 @@ static void bnx2x_set_one_mac_e2(struct bnx2x *bp,
 	bnx2x_set_fw_mac_addr(&rule_entry->mac.mac_msb,
 			      &rule_entry->mac.mac_mid,
 			      &rule_entry->mac.mac_lsb, mac);
+	rule_entry->mac.inner_mac =
+		cpu_to_le16(elem->cmd_data.vlan_mac.u.mac.is_inner_mac);
 
 	/* MOVE: Add a rule that will add this MAC to the target Queue */
 	if (cmd == BNX2X_VLAN_MAC_MOVE) {
@@ -775,6 +783,9 @@ static void bnx2x_set_one_mac_e2(struct bnx2x *bp,
 		bnx2x_set_fw_mac_addr(&rule_entry->mac.mac_msb,
 				      &rule_entry->mac.mac_mid,
 				      &rule_entry->mac.mac_lsb, mac);
+		rule_entry->mac.inner_mac =
+			cpu_to_le16(elem->cmd_data.vlan_mac.
+						u.mac.is_inner_mac);
 	}
 
 	/* Set the ramrod data header */
@@ -963,7 +974,8 @@ static void bnx2x_set_one_vlan_mac_e2(struct bnx2x *bp,
 	bnx2x_set_fw_mac_addr(&rule_entry->pair.mac_msb,
 			      &rule_entry->pair.mac_mid,
 			      &rule_entry->pair.mac_lsb, mac);
-
+	rule_entry->pair.inner_mac =
+		cpu_to_le16(elem->cmd_data.vlan_mac.u.vlan_mac.is_inner_mac);
 	/* MOVE: Add a rule that will add this MAC to the target Queue */
 	if (cmd == BNX2X_VLAN_MAC_MOVE) {
 		rule_entry++;
@@ -980,6 +992,9 @@ static void bnx2x_set_one_vlan_mac_e2(struct bnx2x *bp,
 		bnx2x_set_fw_mac_addr(&rule_entry->pair.mac_msb,
 				      &rule_entry->pair.mac_mid,
 				      &rule_entry->pair.mac_lsb, mac);
+		rule_entry->pair.inner_mac =
+			cpu_to_le16(elem->cmd_data.vlan_mac.u.
+						vlan_mac.is_inner_mac);
 	}
 
 	/* Set the ramrod data header */
@@ -4417,6 +4432,10 @@ static void bnx2x_q_fill_init_tx_data(struct bnx2x_queue_sp_obj *o,
 	tx_data->force_default_pri_flg =
 		test_bit(BNX2X_Q_FLG_FORCE_DEFAULT_PRI, flags);
 
+	tx_data->tunnel_non_lso_pcsum_location =
+		test_bit(BNX2X_Q_FLG_PCSUM_ON_PKT, flags) ? PCSUM_ON_PKT :
+								  PCSUM_ON_BD;
+
 	tx_data->tx_status_block_id = params->fw_sb_id;
 	tx_data->tx_sb_index_number = params->sb_cq_index;
 	tx_data->tss_leading_client_id = params->tss_leading_cl_id;
diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
index ac57e63..064dba2 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_sp.h
@@ -100,6 +100,7 @@ struct bnx2x_raw_obj {
 /************************* VLAN-MAC commands related parameters ***************/
 struct bnx2x_mac_ramrod_data {
 	u8 mac[ETH_ALEN];
+	u8 is_inner_mac;
 };
 
 struct bnx2x_vlan_ramrod_data {
@@ -108,6 +109,7 @@ struct bnx2x_vlan_ramrod_data {
 
 struct bnx2x_vlan_mac_ramrod_data {
 	u8 mac[ETH_ALEN];
+	u8 is_inner_mac;
 	u16 vlan;
 };
 
@@ -825,7 +827,8 @@ enum {
 	BNX2X_Q_FLG_TX_SEC,
 	BNX2X_Q_FLG_ANTI_SPOOF,
 	BNX2X_Q_FLG_SILENT_VLAN_REM,
-	BNX2X_Q_FLG_FORCE_DEFAULT_PRI
+	BNX2X_Q_FLG_FORCE_DEFAULT_PRI,
+	BNX2X_Q_FLG_PCSUM_ON_PKT
 };
 
 /* Queue type options: queue type may be a compination of below. */
-- 
1.7.7.2

[PATCH net-next 13/13] bnx2x: Restore FCoE 4-port devices support

[Yuval Mintz]

From: Dmitry Kravkov <dmitry@broadcom.com>

bnx2x FW 1.78.17 properly supports DCBX configuration for 4-port devices,
enabling FCoE support on 57840 boards.

Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
---
 drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c |   10 ----------
 1 files changed, 0 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
index c7df223..04d123f 100644
--- a/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
+++ b/drivers/net/ethernet/broadcom/bnx2x/bnx2x_main.c
@@ -12492,16 +12492,6 @@ static int bnx2x_init_one(struct pci_dev *pdev,
 	if (CHIP_IS_E1x(bp))
 		bp->flags |= NO_FCOE_FLAG;
 
-	/* disable FCOE for 57840 device, until FW supports it */
-	switch (ent->driver_data) {
-	case BCM57840_O:
-	case BCM57840_4_10:
-	case BCM57840_2_20:
-	case BCM57840_MFO:
-	case BCM57840_MF:
-		bp->flags |= NO_FCOE_FLAG;
-	}
-
 	/* Set bp->num_queues for MSI-X mode*/
 	bnx2x_set_num_queues(bp);
 
-- 
1.7.7.2

Re: [PATCH net-next 0/13] bnx2x: patch series

[David Miller]

From: "Yuval Mintz" <yuvalmin@broadcom.com>
Date: Mon, 11 Mar 2013 17:17:40 +0200

> This patch series contains several changes:
> 
>  1. SR-IOV related patches - allow setting and getting the number of VFs
>     via sysfs, configuring and querying VFs via iproute2 callbacks and
>     properly implement for VFs only the ethtool ops which they can safely
>     perform.
> 
>  2. Advance the FW version used by the driver. This allows 57840 boards to
>     use FCoE, as well as fixing various issues.
> 
>  3. Several small link enhancements.
> 
>  4. Several fixes of rare corner-cases.
> 
> Please consider applying this patch-series to 'net-next'.

Series applied, thanks.