Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking changes from David Miller:

1) Allow to dump, monitor, and change the bridge multicast database
   using netlink.  From Cong Wang.

2) RFC 5961 TCP blind data injection attack mitigation, from Eric
   Dumazet.

3) Networking user namespace support from Eric W. Biederman.

4) tuntap/virtio-net multiqueue support by Jason Wang.

5) Support for checksum offload of encapsulated packets (basically,
   tunneled traffic can still be checksummed by HW).  From Joseph
   Gasparakis.

6) Allow BPF filter access to VLAN tags, from Eric Dumazet and
   Daniel Borkmann.

7) Bridge port parameters over netlink and BPDU blocking support
   from Stephen Hemminger.

8) Improve data access patterns during inet socket demux by rearranging
   socket layout, from Eric Dumazet.

9) TIPC protocol updates and cleanups from Ying Xue, Paul Gortmaker, and
   Jon Maloy.

10) Update TCP socket hash sizing to be more in line with current day
    realities.  The existing heurstics were choosen a decade ago.
    From Eric Dumazet.

11) Fix races, queue bloat, and excessive wakeups in ATM and
    associated drivers, from Krzysztof Mazur and David Woodhouse.

12) Support DOVE (Distributed Overlay Virtual Ethernet) extensions
    in VXLAN driver, from David Stevens.

13) Add "oops_only" mode to netconsole, from Amerigo Wang.

14) Support set and query of VEB/VEPA bridge mode via PF_BRIDGE, also
    allow DCB netlink to work on namespaces other than the initial
    namespace.  From John Fastabend.

15) Support PTP in the Tigon3 driver, from Matt Carlson.

16) tun/vhost zero copy fixes and improvements, plus turn it on
    by default, from Michael S. Tsirkin.

17) Support per-association statistics in SCTP, from Michele
    Baldessari.

And many, many, driver updates, cleanups, and improvements.  Too
numerous to mention individually.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1722 commits)
  net/mlx4_en: Add support for destination MAC in steering rules
  net/mlx4_en: Use generic etherdevice.h functions.
  net: ethtool: Add destination MAC address to flow steering API
  bridge: add support of adding and deleting mdb entries
  bridge: notify mdb changes via netlink
  ndisc: Unexport ndisc_{build,send}_skb().
  uapi: add missing netconf.h to export list
  pkt_sched: avoid requeues if possible
  solos-pci: fix double-free of TX skb in DMA mode
  bnx2: Fix accidental reversions.
  bna: Driver Version Updated to 3.1.2.1
  bna: Firmware update
  bna: Add RX State
  bna: Rx Page Based Allocation
  bna: TX Intr Coalescing Fix
  bna: Tx and Rx Optimizations
  bna: Code Cleanup and Enhancements
  ath9k: check pdata variable before dereferencing it
  ath5k: RX timestamp is reported at end of frame
  ath9k_htc: RX timestamp is reported at end of frame
  ...

5 files changed, 156 insertions, 163 deletions

diff --git a/drivers/net/ethernet/realtek/8139cp.c b/drivers/net/ethernet/realtek/8139cp.c
index 609125a249d..cb6fc5a743c 100644
--- a/drivers/net/ethernet/realtek/8139cp.c
+++ b/drivers/net/ethernet/realtek/8139cp.c
@@ -648,6 +648,7 @@ static void cp_tx (struct cp_private *cp)
 {
 	unsigned tx_head = cp->tx_head;
 	unsigned tx_tail = cp->tx_tail;
+	unsigned bytes_compl = 0, pkts_compl = 0;
 
 	while (tx_tail != tx_head) {
 		struct cp_desc *txd = cp->tx_ring + tx_tail;
@@ -666,6 +667,9 @@ static void cp_tx (struct cp_private *cp)
 				 le32_to_cpu(txd->opts1) & 0xffff,
 				 PCI_DMA_TODEVICE);
 
+		bytes_compl += skb->len;
+		pkts_compl++;
+
 		if (status & LastFrag) {
 			if (status & (TxError | TxFIFOUnder)) {
 				netif_dbg(cp, tx_err, cp->dev,
@@ -697,6 +701,7 @@ static void cp_tx (struct cp_private *cp)
 
 	cp->tx_tail = tx_tail;
 
+	netdev_completed_queue(cp->dev, pkts_compl, bytes_compl);
 	if (TX_BUFFS_AVAIL(cp) > (MAX_SKB_FRAGS + 1))
 		netif_wake_queue(cp->dev);
 }
@@ -843,6 +848,8 @@ static netdev_tx_t cp_start_xmit (struct sk_buff *skb,
 		wmb();
 	}
 	cp->tx_head = entry;
+
+	netdev_sent_queue(dev, skb->len);
 	netif_dbg(cp, tx_queued, cp->dev, "tx queued, slot %d, skblen %d\n",
 		  entry, skb->len);
 	if (TX_BUFFS_AVAIL(cp) <= (MAX_SKB_FRAGS + 1))
@@ -937,6 +944,8 @@ static void cp_stop_hw (struct cp_private *cp)
 
 	cp->rx_tail = 0;
 	cp->tx_head = cp->tx_tail = 0;
+
+	netdev_reset_queue(cp->dev);
 }
 
 static void cp_reset_hw (struct cp_private *cp)
@@ -957,8 +966,38 @@ static void cp_reset_hw (struct cp_private *cp)
 
 static inline void cp_start_hw (struct cp_private *cp)
 {
+	dma_addr_t ring_dma;
+
 	cpw16(CpCmd, cp->cpcmd);
+
+	/*
+	 * These (at least TxRingAddr) need to be configured after the
+	 * corresponding bits in CpCmd are enabled. Datasheet v1.6 §6.33
+	 * (C+ Command Register) recommends that these and more be configured
+	 * *after* the [RT]xEnable bits in CpCmd are set. And on some hardware
+	 * it's been observed that the TxRingAddr is actually reset to garbage
+	 * when C+ mode Tx is enabled in CpCmd.
+	 */
+	cpw32_f(HiTxRingAddr, 0);
+	cpw32_f(HiTxRingAddr + 4, 0);
+
+	ring_dma = cp->ring_dma;
+	cpw32_f(RxRingAddr, ring_dma & 0xffffffff);
+	cpw32_f(RxRingAddr + 4, (ring_dma >> 16) >> 16);
+
+	ring_dma += sizeof(struct cp_desc) * CP_RX_RING_SIZE;
+	cpw32_f(TxRingAddr, ring_dma & 0xffffffff);
+	cpw32_f(TxRingAddr + 4, (ring_dma >> 16) >> 16);
+
+	/*
+	 * Strictly speaking, the datasheet says this should be enabled
+	 * *before* setting the descriptor addresses. But what, then, would
+	 * prevent it from doing DMA to random unconfigured addresses?
+	 * This variant appears to work fine.
+	 */
 	cpw8(Cmd, RxOn | TxOn);
+
+	netdev_reset_queue(cp->dev);
 }
 
 static void cp_enable_irq(struct cp_private *cp)
@@ -969,7 +1008,6 @@ static void cp_enable_irq(struct cp_private *cp)
 static void cp_init_hw (struct cp_private *cp)
 {
 	struct net_device *dev = cp->dev;
-	dma_addr_t ring_dma;
 
 	cp_reset_hw(cp);
 
@@ -992,17 +1030,6 @@ static void cp_init_hw (struct cp_private *cp)
 
 	cpw8(Config5, cpr8(Config5) & PMEStatus);
 
-	cpw32_f(HiTxRingAddr, 0);
-	cpw32_f(HiTxRingAddr + 4, 0);
-
-	ring_dma = cp->ring_dma;
-	cpw32_f(RxRingAddr, ring_dma & 0xffffffff);
-	cpw32_f(RxRingAddr + 4, (ring_dma >> 16) >> 16);
-
-	ring_dma += sizeof(struct cp_desc) * CP_RX_RING_SIZE;
-	cpw32_f(TxRingAddr, ring_dma & 0xffffffff);
-	cpw32_f(TxRingAddr + 4, (ring_dma >> 16) >> 16);
-
 	cpw16(MultiIntr, 0);
 
 	cpw8_f(Cfg9346, Cfg9346_Lock);
@@ -1197,18 +1224,16 @@ static void cp_tx_timeout(struct net_device *dev)
 	cp_clean_rings(cp);
 	rc = cp_init_rings(cp);
 	cp_start_hw(cp);
+	cp_enable_irq(cp);
 
 	netif_wake_queue(dev);
 
 	spin_unlock_irqrestore(&cp->lock, flags);
 }
 
-#ifdef BROKEN
 static int cp_change_mtu(struct net_device *dev, int new_mtu)
 {
 	struct cp_private *cp = netdev_priv(dev);
-	int rc;
-	unsigned long flags;
 
 	/* check for invalid MTU, according to hardware limits */
 	if (new_mtu < CP_MIN_MTU || new_mtu > CP_MAX_MTU)
@@ -1221,22 +1246,12 @@ static int cp_change_mtu(struct net_device *dev, int new_mtu)
 		return 0;
 	}
 
-	spin_lock_irqsave(&cp->lock, flags);
-
-	cp_stop_hw(cp);			/* stop h/w and free rings */
-	cp_clean_rings(cp);
-
+	/* network IS up, close it, reset MTU, and come up again. */
+	cp_close(dev);
 	dev->mtu = new_mtu;
-	cp_set_rxbufsize(cp);		/* set new rx buf size */
-
-	rc = cp_init_rings(cp);		/* realloc and restart h/w */
-	cp_start_hw(cp);
-
-	spin_unlock_irqrestore(&cp->lock, flags);
-
-	return rc;
+	cp_set_rxbufsize(cp);
+	return cp_open(dev);
 }
-#endif /* BROKEN */
 
 static const char mii_2_8139_map[8] = {
 	BasicModeCtrl,
@@ -1812,9 +1827,7 @@ static const struct net_device_ops cp_netdev_ops = {
 	.ndo_start_xmit		= cp_start_xmit,
 	.ndo_tx_timeout		= cp_tx_timeout,
 	.ndo_set_features	= cp_set_features,
-#ifdef BROKEN
 	.ndo_change_mtu		= cp_change_mtu,
-#endif
 
 #ifdef CONFIG_NET_POLL_CONTROLLER
 	.ndo_poll_controller	= cp_poll_controller,
diff --git a/drivers/net/ethernet/realtek/8139too.c b/drivers/net/ethernet/realtek/8139too.c
index 3ed7add23c1..5dc16163012 100644
--- a/drivers/net/ethernet/realtek/8139too.c
+++ b/drivers/net/ethernet/realtek/8139too.c
@@ -228,7 +228,7 @@ typedef enum {
 static const struct {
 	const char *name;
 	u32 hw_flags;
-} board_info[] __devinitconst = {
+} board_info[] = {
 	{ "RealTek RTL8139", RTL8139_CAPS },
 	{ "RealTek RTL8129", RTL8129_CAPS },
 };
@@ -748,7 +748,7 @@ static void rtl8139_chip_reset (void __iomem *ioaddr)
 }
 
 
-static __devinit struct net_device * rtl8139_init_board (struct pci_dev *pdev)
+static struct net_device *rtl8139_init_board(struct pci_dev *pdev)
 {
 	struct device *d = &pdev->dev;
 	void __iomem *ioaddr;
@@ -935,8 +935,8 @@ static const struct net_device_ops rtl8139_netdev_ops = {
 	.ndo_set_features	= rtl8139_set_features,
 };
 
-static int __devinit rtl8139_init_one (struct pci_dev *pdev,
-				       const struct pci_device_id *ent)
+static int rtl8139_init_one(struct pci_dev *pdev,
+			    const struct pci_device_id *ent)
 {
 	struct net_device *dev = NULL;
 	struct rtl8139_private *tp;
@@ -1103,7 +1103,7 @@ err_out:
 }
 
 
-static void __devexit rtl8139_remove_one (struct pci_dev *pdev)
+static void rtl8139_remove_one(struct pci_dev *pdev)
 {
 	struct net_device *dev = pci_get_drvdata (pdev);
 	struct rtl8139_private *tp = netdev_priv(dev);
@@ -1141,7 +1141,7 @@ static void __devexit rtl8139_remove_one (struct pci_dev *pdev)
 #define EE_READ_CMD		(6)
 #define EE_ERASE_CMD	(7)
 
-static int __devinit read_eeprom (void __iomem *ioaddr, int location, int addr_len)
+static int read_eeprom(void __iomem *ioaddr, int location, int addr_len)
 {
 	int i;
 	unsigned retval = 0;
@@ -2652,7 +2652,7 @@ static struct pci_driver rtl8139_pci_driver = {
 	.name		= DRV_NAME,
 	.id_table	= rtl8139_pci_tbl,
 	.probe		= rtl8139_init_one,
-	.remove		= __devexit_p(rtl8139_remove_one),
+	.remove		= rtl8139_remove_one,
 #ifdef CONFIG_PM
 	.suspend	= rtl8139_suspend,
 	.resume		= rtl8139_resume,
diff --git a/drivers/net/ethernet/realtek/atp.c b/drivers/net/ethernet/realtek/atp.c
index e02f04d7f3a..9f2d416de75 100644
--- a/drivers/net/ethernet/realtek/atp.c
+++ b/drivers/net/ethernet/realtek/atp.c
@@ -175,8 +175,7 @@ struct net_local {
     unsigned int tx_unit_busy:1;
     unsigned char re_tx,	/* Number of packet retransmissions. */
 		addr_mode,		/* Current Rx filter e.g. promiscuous, etc. */
-		pac_cnt_in_tx_buf,
-		chip_type;
+		pac_cnt_in_tx_buf;
 };
 
 /* This code, written by wwc@super.org, resets the adapter every
@@ -339,7 +338,6 @@ static int __init atp_probe1(long ioaddr)
 	write_reg_high(ioaddr, CMR1, CMR1h_RESET | CMR1h_MUX);
 
 	lp = netdev_priv(dev);
-	lp->chip_type = RTL8002;
 	lp->addr_mode = CMR2h_Normal;
 	spin_lock_init(&lp->lock);
 
@@ -852,7 +850,7 @@ net_close(struct net_device *dev)
  *	Set or clear the multicast filter for this adapter.
  */
 
-static void set_rx_mode_8002(struct net_device *dev)
+static void set_rx_mode(struct net_device *dev)
 {
 	struct net_local *lp = netdev_priv(dev);
 	long ioaddr = dev->base_addr;
@@ -864,58 +862,6 @@ static void set_rx_mode_8002(struct net_device *dev)
 	write_reg_high(ioaddr, CMR2, lp->addr_mode);
 }
 
-static void set_rx_mode_8012(struct net_device *dev)
-{
-	struct net_local *lp = netdev_priv(dev);
-	long ioaddr = dev->base_addr;
-	unsigned char new_mode, mc_filter[8]; /* Multicast hash filter */
-	int i;
-
-	if (dev->flags & IFF_PROMISC) {			/* Set promiscuous. */
-		new_mode = CMR2h_PROMISC;
-	} else if ((netdev_mc_count(dev) > 1000) ||
-		   (dev->flags & IFF_ALLMULTI)) {
-		/* Too many to filter perfectly -- accept all multicasts. */
-		memset(mc_filter, 0xff, sizeof(mc_filter));
-		new_mode = CMR2h_Normal;
-	} else {
-		struct netdev_hw_addr *ha;
-
-		memset(mc_filter, 0, sizeof(mc_filter));
-		netdev_for_each_mc_addr(ha, dev) {
-			int filterbit = ether_crc_le(ETH_ALEN, ha->addr) & 0x3f;
-			mc_filter[filterbit >> 5] |= 1 << (filterbit & 31);
-		}
-		new_mode = CMR2h_Normal;
-	}
-	lp->addr_mode = new_mode;
-    write_reg(ioaddr, CMR2, CMR2_IRQOUT | 0x04); /* Switch to page 1. */
-    for (i = 0; i < 8; i++)
-		write_reg_byte(ioaddr, i, mc_filter[i]);
-	if (net_debug > 2 || 1) {
-		lp->addr_mode = 1;
-		printk(KERN_DEBUG "%s: Mode %d, setting multicast filter to",
-			   dev->name, lp->addr_mode);
-		for (i = 0; i < 8; i++)
-			printk(" %2.2x", mc_filter[i]);
-		printk(".\n");
-	}
-
-	write_reg_high(ioaddr, CMR2, lp->addr_mode);
-    write_reg(ioaddr, CMR2, CMR2_IRQOUT); /* Switch back to page 0 */
-}
-
-static void set_rx_mode(struct net_device *dev)
-{
-	struct net_local *lp = netdev_priv(dev);
-
-	if (lp->chip_type == RTL8002)
-		return set_rx_mode_8002(dev);
-	else
-		return set_rx_mode_8012(dev);
-}
-
-
 static int __init atp_init_module(void) {
 	if (debug)					/* Emit version even if no cards detected. */
 		printk(KERN_INFO "%s", version);
diff --git a/drivers/net/ethernet/realtek/atp.h b/drivers/net/ethernet/realtek/atp.h
index 0edc642c2c2..040b1373994 100644
--- a/drivers/net/ethernet/realtek/atp.h
+++ b/drivers/net/ethernet/realtek/atp.h
@@ -16,8 +16,6 @@ struct rx_header {
 #define PAR_STATUS	1
 #define PAR_CONTROL 2
 
-enum chip_type { RTL8002, RTL8012 };
-
 #define Ctrl_LNibRead	0x08	/* LP_PSELECP */
 #define Ctrl_HNibRead	0
 #define Ctrl_LNibWrite	0x08	/* LP_PSELECP */
diff --git a/drivers/net/ethernet/realtek/r8169.c b/drivers/net/ethernet/realtek/r8169.c
index 927aa33d434..ed96f309bca 100644
--- a/drivers/net/ethernet/realtek/r8169.c
+++ b/drivers/net/ethernet/realtek/r8169.c
@@ -78,24 +78,18 @@ static const int multicast_filter_limit = 32;
 
 #define MAX_READ_REQUEST_SHIFT	12
 #define TX_DMA_BURST	7	/* Maximum PCI burst, '7' is unlimited */
-#define SafeMtu		0x1c20	/* ... actually life sucks beyond ~7k */
 #define InterFrameGap	0x03	/* 3 means InterFrameGap = the shortest one */
 
 #define R8169_REGS_SIZE		256
 #define R8169_NAPI_WEIGHT	64
 #define NUM_TX_DESC	64	/* Number of Tx descriptor registers */
 #define NUM_RX_DESC	256	/* Number of Rx descriptor registers */
-#define RX_BUF_SIZE	1536	/* Rx Buffer size */
 #define R8169_TX_RING_BYTES	(NUM_TX_DESC * sizeof(struct TxDesc))
 #define R8169_RX_RING_BYTES	(NUM_RX_DESC * sizeof(struct RxDesc))
 
 #define RTL8169_TX_TIMEOUT	(6*HZ)
 #define RTL8169_PHY_TIMEOUT	(10*HZ)
 
-#define RTL_EEPROM_SIG		cpu_to_le32(0x8129)
-#define RTL_EEPROM_SIG_MASK	cpu_to_le32(0xffff)
-#define RTL_EEPROM_SIG_ADDR	0x0000
-
 /* write/read MMIO register */
 #define RTL_W8(reg, val8)	writeb ((val8), ioaddr + (reg))
 #define RTL_W16(reg, val16)	writew ((val16), ioaddr + (reg))
@@ -456,6 +450,7 @@ enum rtl8168_registers {
 #define PWM_EN				(1 << 22)
 #define RXDV_GATED_EN			(1 << 19)
 #define EARLY_TALLY_EN			(1 << 16)
+#define FORCE_CLK			(1 << 15) /* force clock request */
 };
 
 enum rtl_register_content {
@@ -519,6 +514,7 @@ enum rtl_register_content {
 	PMEnable	= (1 << 0),	/* Power Management Enable */
 
 	/* Config2 register p. 25 */
+	ClkReqEn	= (1 << 7),	/* Clock Request Enable */
 	MSIEnable	= (1 << 5),	/* 8169 only. Reserved in the 8168. */
 	PCI_Clock_66MHz = 0x01,
 	PCI_Clock_33MHz = 0x00,
@@ -539,6 +535,7 @@ enum rtl_register_content {
 	Spi_en		= (1 << 3),
 	LanWake		= (1 << 1),	/* LanWake enable/disable */
 	PMEStatus	= (1 << 0),	/* PME status can be reset by PCI RST# */
+	ASPM_en		= (1 << 0),	/* ASPM enable */
 
 	/* TBICSR p.28 */
 	TBIReset	= 0x80000000,
@@ -687,6 +684,7 @@ enum features {
 	RTL_FEATURE_WOL		= (1 << 0),
 	RTL_FEATURE_MSI		= (1 << 1),
 	RTL_FEATURE_GMII	= (1 << 2),
+	RTL_FEATURE_FW_LOADED	= (1 << 3),
 };
 
 struct rtl8169_counters {
@@ -1816,8 +1814,7 @@ static int rtl8169_set_features(struct net_device *dev,
 }
 
 
-static inline u32 rtl8169_tx_vlan_tag(struct rtl8169_private *tp,
-				      struct sk_buff *skb)
+static inline u32 rtl8169_tx_vlan_tag(struct sk_buff *skb)
 {
 	return (vlan_tx_tag_present(skb)) ?
 		TxVlanTag | swab16(vlan_tx_tag_get(skb)) : 0x00;
@@ -2394,8 +2391,10 @@ static void rtl_apply_firmware(struct rtl8169_private *tp)
 	struct rtl_fw *rtl_fw = tp->rtl_fw;
 
 	/* TODO: release firmware once rtl_phy_write_fw signals failures. */
-	if (!IS_ERR_OR_NULL(rtl_fw))
+	if (!IS_ERR_OR_NULL(rtl_fw)) {
 		rtl_phy_write_fw(tp, rtl_fw);
+		tp->features |= RTL_FEATURE_FW_LOADED;
+	}
 }
 
 static void rtl_apply_firmware_cond(struct rtl8169_private *tp, u8 reg, u16 val)
@@ -2406,6 +2405,31 @@ static void rtl_apply_firmware_cond(struct rtl8169_private *tp, u8 reg, u16 val)
 		rtl_apply_firmware(tp);
 }
 
+static void r810x_aldps_disable(struct rtl8169_private *tp)
+{
+	rtl_writephy(tp, 0x1f, 0x0000);
+	rtl_writephy(tp, 0x18, 0x0310);
+	msleep(100);
+}
+
+static void r810x_aldps_enable(struct rtl8169_private *tp)
+{
+	if (!(tp->features & RTL_FEATURE_FW_LOADED))
+		return;
+
+	rtl_writephy(tp, 0x1f, 0x0000);
+	rtl_writephy(tp, 0x18, 0x8310);
+}
+
+static void r8168_aldps_enable_1(struct rtl8169_private *tp)
+{
+	if (!(tp->features & RTL_FEATURE_FW_LOADED))
+		return;
+
+	rtl_writephy(tp, 0x1f, 0x0000);
+	rtl_w1w0_phy(tp, 0x15, 0x1000, 0x0000);
+}
+
 static void rtl8169s_hw_phy_config(struct rtl8169_private *tp)
 {
 	static const struct phy_reg phy_reg_init[] = {
@@ -3096,6 +3120,23 @@ static void rtl8168e_1_hw_phy_config(struct rtl8169_private *tp)
 	rtl_writephy(tp, 0x0d, 0x0000);
 }
 
+static void rtl_rar_exgmac_set(struct rtl8169_private *tp, u8 *addr)
+{
+	const u16 w[] = {
+		addr[0] | (addr[1] << 8),
+		addr[2] | (addr[3] << 8),
+		addr[4] | (addr[5] << 8)
+	};
+	const struct exgmac_reg e[] = {
+		{ .addr = 0xe0, ERIAR_MASK_1111, .val = w[0] | (w[1] << 16) },
+		{ .addr = 0xe4, ERIAR_MASK_1111, .val = w[2] },
+		{ .addr = 0xf0, ERIAR_MASK_1111, .val = w[0] << 16 },
+		{ .addr = 0xf4, ERIAR_MASK_1111, .val = w[1] | (w[2] << 16) }
+	};
+
+	rtl_write_exgmac_batch(tp, e, ARRAY_SIZE(e));
+}
+
 static void rtl8168e_2_hw_phy_config(struct rtl8169_private *tp)
 {
 	static const struct phy_reg phy_reg_init[] = {
@@ -3178,6 +3219,11 @@ static void rtl8168e_2_hw_phy_config(struct rtl8169_private *tp)
 	rtl_w1w0_phy(tp, 0x19, 0x0000, 0x0001);
 	rtl_w1w0_phy(tp, 0x10, 0x0000, 0x0400);
 	rtl_writephy(tp, 0x1f, 0x0000);
+
+	r8168_aldps_enable_1(tp);
+
+	/* Broken BIOS workaround: feed GigaMAC registers with MAC address. */
+	rtl_rar_exgmac_set(tp, tp->dev->dev_addr);
 }
 
 static void rtl8168f_hw_phy_config(struct rtl8169_private *tp)
@@ -3250,6 +3296,8 @@ static void rtl8168f_1_hw_phy_config(struct rtl8169_private *tp)
 	rtl_writephy(tp, 0x05, 0x8b85);
 	rtl_w1w0_phy(tp, 0x06, 0x4000, 0x0000);
 	rtl_writephy(tp, 0x1f, 0x0000);
+
+	r8168_aldps_enable_1(tp);
 }
 
 static void rtl8168f_2_hw_phy_config(struct rtl8169_private *tp)
@@ -3257,6 +3305,8 @@ static void rtl8168f_2_hw_phy_config(struct rtl8169_private *tp)
 	rtl_apply_firmware(tp);
 
 	rtl8168f_hw_phy_config(tp);
+
+	r8168_aldps_enable_1(tp);
 }
 
 static void rtl8411_hw_phy_config(struct rtl8169_private *tp)
@@ -3354,6 +3404,8 @@ static void rtl8411_hw_phy_config(struct rtl8169_private *tp)
 	rtl_w1w0_phy(tp, 0x19, 0x0000, 0x0001);
 	rtl_w1w0_phy(tp, 0x10, 0x0000, 0x0400);
 	rtl_writephy(tp, 0x1f, 0x0000);
+
+	r8168_aldps_enable_1(tp);
 }
 
 static void rtl8168g_1_hw_phy_config(struct rtl8169_private *tp)
@@ -3439,21 +3491,19 @@ static void rtl8105e_hw_phy_config(struct rtl8169_private *tp)
 	};
 
 	/* Disable ALDPS before ram code */
-	rtl_writephy(tp, 0x1f, 0x0000);
-	rtl_writephy(tp, 0x18, 0x0310);
-	msleep(100);
+	r810x_aldps_disable(tp);
 
 	rtl_apply_firmware(tp);
 
 	rtl_writephy_batch(tp, phy_reg_init, ARRAY_SIZE(phy_reg_init));
+
+	r810x_aldps_enable(tp);
 }
 
 static void rtl8402_hw_phy_config(struct rtl8169_private *tp)
 {
 	/* Disable ALDPS before setting firmware */
-	rtl_writephy(tp, 0x1f, 0x0000);
-	rtl_writephy(tp, 0x18, 0x0310);
-	msleep(20);
+	r810x_aldps_disable(tp);
 
 	rtl_apply_firmware(tp);
 
@@ -3463,6 +3513,8 @@ static void rtl8402_hw_phy_config(struct rtl8169_private *tp)
 	rtl_writephy(tp, 0x10, 0x401f);
 	rtl_writephy(tp, 0x19, 0x7030);
 	rtl_writephy(tp, 0x1f, 0x0000);
+
+	r810x_aldps_enable(tp);
 }
 
 static void rtl8106e_hw_phy_config(struct rtl8169_private *tp)
@@ -3475,9 +3527,7 @@ static void rtl8106e_hw_phy_config(struct rtl8169_private *tp)
 	};
 
 	/* Disable ALDPS before ram code */
-	rtl_writephy(tp, 0x1f, 0x0000);
-	rtl_writephy(tp, 0x18, 0x0310);
-	msleep(100);
+	r810x_aldps_disable(tp);
 
 	rtl_apply_firmware(tp);
 
@@ -3485,6 +3535,8 @@ static void rtl8106e_hw_phy_config(struct rtl8169_private *tp)
 	rtl_writephy_batch(tp, phy_reg_init, ARRAY_SIZE(phy_reg_init));
 
 	rtl_eri_write(tp, 0x1d0, ERIAR_MASK_0011, 0x0000, ERIAR_EXGMAC);
+
+	r810x_aldps_enable(tp);
 }
 
 static void rtl_hw_phy_config(struct net_device *dev)
@@ -3708,33 +3760,19 @@ static void rtl8169_init_phy(struct net_device *dev, struct rtl8169_private *tp)
 static void rtl_rar_set(struct rtl8169_private *tp, u8 *addr)
 {
 	void __iomem *ioaddr = tp->mmio_addr;
-	u32 high;
-	u32 low;
-
-	low  = addr[0] | (addr[1] << 8) | (addr[2] << 16) | (addr[3] << 24);
-	high = addr[4] | (addr[5] << 8);
 
 	rtl_lock_work(tp);
 
 	RTL_W8(Cfg9346, Cfg9346_Unlock);
 
-	RTL_W32(MAC4, high);
+	RTL_W32(MAC4, addr[4] | addr[5] << 8);
 	RTL_R32(MAC4);
 
-	RTL_W32(MAC0, low);
+	RTL_W32(MAC0, addr[0] | addr[1] << 8 | addr[2] << 16 | addr[3] << 24);
 	RTL_R32(MAC0);
 
-	if (tp->mac_version == RTL_GIGA_MAC_VER_34) {
-		const struct exgmac_reg e[] = {
-			{ .addr = 0xe0, ERIAR_MASK_1111, .val = low },
-			{ .addr = 0xe4, ERIAR_MASK_1111, .val = high },
-			{ .addr = 0xf0, ERIAR_MASK_1111, .val = low << 16 },
-			{ .addr = 0xf4, ERIAR_MASK_1111, .val = high << 16 |
-								low  >> 16 },
-		};
-
-		rtl_write_exgmac_batch(tp, e, ARRAY_SIZE(e));
-	}
+	if (tp->mac_version == RTL_GIGA_MAC_VER_34)
+		rtl_rar_exgmac_set(tp, addr);
 
 	RTL_W8(Cfg9346, Cfg9346_Lock);
 
@@ -3796,7 +3834,7 @@ static void rtl_disable_msi(struct pci_dev *pdev, struct rtl8169_private *tp)
 	}
 }
 
-static void __devinit rtl_init_mdio_ops(struct rtl8169_private *tp)
+static void rtl_init_mdio_ops(struct rtl8169_private *tp)
 {
 	struct mdio_ops *ops = &tp->mdio_ops;
 
@@ -4048,7 +4086,7 @@ static void rtl_pll_power_up(struct rtl8169_private *tp)
 	rtl_generic_op(tp, tp->pll_power_ops.up);
 }
 
-static void __devinit rtl_init_pll_power_ops(struct rtl8169_private *tp)
+static void rtl_init_pll_power_ops(struct rtl8169_private *tp)
 {
 	struct pll_power_ops *ops = &tp->pll_power_ops;
 
@@ -4242,7 +4280,7 @@ static void r8168b_1_hw_jumbo_disable(struct rtl8169_private *tp)
 	RTL_W8(Config4, RTL_R8(Config4) & ~(1 << 0));
 }
 
-static void __devinit rtl_init_jumbo_ops(struct rtl8169_private *tp)
+static void rtl_init_jumbo_ops(struct rtl8169_private *tp)
 {
 	struct jumbo_ops *ops = &tp->jumbo_ops;
 
@@ -4683,7 +4721,7 @@ static u32 r8402_csi_read(struct rtl8169_private *tp, int addr)
 		RTL_R32(CSIDR) : ~0;
 }
 
-static void __devinit rtl_init_csi_ops(struct rtl8169_private *tp)
+static void rtl_init_csi_ops(struct rtl8169_private *tp)
 {
 	struct csi_ops *ops = &tp->csi_ops;
 
@@ -5015,8 +5053,6 @@ static void rtl_hw_start_8168e_2(struct rtl8169_private *tp)
 
 	RTL_W8(MaxTxPacketSize, EarlySize);
 
-	rtl_disable_clock_request(pdev);
-
 	RTL_W32(TxConfig, RTL_R32(TxConfig) | TXCFG_AUTO_FIFO);
 	RTL_W8(MCU, RTL_R8(MCU) & ~NOW_IS_OOB);
 
@@ -5025,7 +5061,8 @@ static void rtl_hw_start_8168e_2(struct rtl8169_private *tp)
 
 	RTL_W8(DLLPR, RTL_R8(DLLPR) | PFM_EN);
 	RTL_W32(MISC, RTL_R32(MISC) | PWM_EN);
-	RTL_W8(Config5, RTL_R8(Config5) & ~Spi_en);
+	RTL_W8(Config5, (RTL_R8(Config5) & ~Spi_en) | ASPM_en);
+	RTL_W8(Config2, RTL_R8(Config2) | ClkReqEn);
 }
 
 static void rtl_hw_start_8168f(struct rtl8169_private *tp)
@@ -5050,13 +5087,12 @@ static void rtl_hw_start_8168f(struct rtl8169_private *tp)
 
 	RTL_W8(MaxTxPacketSize, EarlySize);
 
-	rtl_disable_clock_request(pdev);
-
 	RTL_W32(TxConfig, RTL_R32(TxConfig) | TXCFG_AUTO_FIFO);
 	RTL_W8(MCU, RTL_R8(MCU) & ~NOW_IS_OOB);
 	RTL_W8(DLLPR, RTL_R8(DLLPR) | PFM_EN);
-	RTL_W32(MISC, RTL_R32(MISC) | PWM_EN);
-	RTL_W8(Config5, RTL_R8(Config5) & ~Spi_en);
+	RTL_W32(MISC, RTL_R32(MISC) | PWM_EN | FORCE_CLK);
+	RTL_W8(Config5, (RTL_R8(Config5) & ~Spi_en) | ASPM_en);
+	RTL_W8(Config2, RTL_R8(Config2) | ClkReqEn);
 }
 
 static void rtl_hw_start_8168f_1(struct rtl8169_private *tp)
@@ -5113,8 +5149,10 @@ static void rtl_hw_start_8168g_1(struct rtl8169_private *tp)
 	rtl_w1w0_eri(tp, 0xdc, ERIAR_MASK_0001, 0x01, 0x00, ERIAR_EXGMAC);
 
 	RTL_W8(ChipCmd, CmdTxEnb | CmdRxEnb);
-	RTL_W32(MISC, RTL_R32(MISC) & ~RXDV_GATED_EN);
+	RTL_W32(MISC, (RTL_R32(MISC) | FORCE_CLK) & ~RXDV_GATED_EN);
 	RTL_W8(MaxTxPacketSize, EarlySize);
+	RTL_W8(Config5, RTL_R8(Config5) | ASPM_en);
+	RTL_W8(Config2, RTL_R8(Config2) | ClkReqEn);
 
 	rtl_eri_write(tp, 0xc0, ERIAR_MASK_0011, 0x0000, ERIAR_EXGMAC);
 	rtl_eri_write(tp, 0xb8, ERIAR_MASK_0011, 0x0000, ERIAR_EXGMAC);
@@ -5330,6 +5368,9 @@ static void rtl_hw_start_8105e_1(struct rtl8169_private *tp)
 
 	RTL_W8(MCU, RTL_R8(MCU) | EN_NDP | EN_OOB_RESET);
 	RTL_W8(DLLPR, RTL_R8(DLLPR) | PFM_EN);
+	RTL_W8(Config5, RTL_R8(Config5) | ASPM_en);
+	RTL_W8(Config2, RTL_R8(Config2) | ClkReqEn);
+	RTL_W32(MISC, RTL_R32(MISC) | FORCE_CLK);
 
 	rtl_ephy_init(tp, e_info_8105e_1, ARRAY_SIZE(e_info_8105e_1));
 }
@@ -5355,6 +5396,9 @@ static void rtl_hw_start_8402(struct rtl8169_private *tp)
 
 	RTL_W32(TxConfig, RTL_R32(TxConfig) | TXCFG_AUTO_FIFO);
 	RTL_W8(MCU, RTL_R8(MCU) & ~NOW_IS_OOB);
+	RTL_W8(Config5, RTL_R8(Config5) | ASPM_en);
+	RTL_W8(Config2, RTL_R8(Config2) | ClkReqEn);
+	RTL_W32(MISC, RTL_R32(MISC) | FORCE_CLK);
 
 	rtl_ephy_init(tp, e_info_8402, ARRAY_SIZE(e_info_8402));
 
@@ -5376,7 +5420,10 @@ static void rtl_hw_start_8106(struct rtl8169_private *tp)
 	/* Force LAN exit from ASPM if Rx/Tx are not idle */
 	RTL_W32(FuncEvent, RTL_R32(FuncEvent) | 0x002800);
 
-	RTL_W32(MISC, (RTL_R32(MISC) | DISABLE_LAN_EN) & ~EARLY_TALLY_EN);
+	RTL_W32(MISC,
+		(RTL_R32(MISC) | DISABLE_LAN_EN | FORCE_CLK) & ~EARLY_TALLY_EN);
+	RTL_W8(Config5, RTL_R8(Config5) | ASPM_en);
+	RTL_W8(Config2, RTL_R8(Config2) | ClkReqEn);
 	RTL_W8(MCU, RTL_R8(MCU) | EN_NDP | EN_OOB_RESET);
 	RTL_W8(DLLPR, RTL_R8(DLLPR) & ~PFM_EN);
 }
@@ -5774,7 +5821,7 @@ static netdev_tx_t rtl8169_start_xmit(struct sk_buff *skb,
 	tp->tx_skb[entry].len = len;
 	txd->addr = cpu_to_le64(mapping);
 
-	opts[1] = cpu_to_le32(rtl8169_tx_vlan_tag(tp, skb));
+	opts[1] = cpu_to_le32(rtl8169_tx_vlan_tag(skb));
 	opts[0] = DescOwn;
 
 	rtl8169_tso_csum(tp, skb, opts);
@@ -6569,7 +6616,7 @@ static void rtl_shutdown(struct pci_dev *pdev)
 	pm_runtime_put_noidle(d);
 }
 
-static void __devexit rtl_remove_one(struct pci_dev *pdev)
+static void rtl_remove_one(struct pci_dev *pdev)
 {
 	struct net_device *dev = pci_get_drvdata(pdev);
 	struct rtl8169_private *tp = netdev_priv(dev);
@@ -6689,7 +6736,7 @@ DECLARE_RTL_COND(rtl_rxtx_empty_cond)
 	return (RTL_R8(MCU) & RXTX_EMPTY) == RXTX_EMPTY;
 }
 
-static void __devinit rtl_hw_init_8168g(struct rtl8169_private *tp)
+static void rtl_hw_init_8168g(struct rtl8169_private *tp)
 {
 	void __iomem *ioaddr = tp->mmio_addr;
 	u32 data;
@@ -6723,7 +6770,7 @@ static void __devinit rtl_hw_init_8168g(struct rtl8169_private *tp)
 		return;
 }
 
-static void __devinit rtl_hw_initialize(struct rtl8169_private *tp)
+static void rtl_hw_initialize(struct rtl8169_private *tp)
 {
 	switch (tp->mac_version) {
 	case RTL_GIGA_MAC_VER_40:
@@ -6736,7 +6783,7 @@ static void __devinit rtl_hw_initialize(struct rtl8169_private *tp)
 	}
 }
 
-static int __devinit
+static int
 rtl_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 {
 	const struct rtl_cfg_info *cfg = rtl_cfg_infos + ent->driver_data;
@@ -6987,20 +7034,9 @@ static struct pci_driver rtl8169_pci_driver = {
 	.name		= MODULENAME,
 	.id_table	= rtl8169_pci_tbl,
 	.probe		= rtl_init_one,
-	.remove		= __devexit_p(rtl_remove_one),
+	.remove		= rtl_remove_one,
 	.shutdown	= rtl_shutdown,
 	.driver.pm	= RTL8169_PM_OPS,
 };
 
-static int __init rtl8169_init_module(void)
-{
-	return pci_register_driver(&rtl8169_pci_driver);
-}
-
-static void __exit rtl8169_cleanup_module(void)
-{
-	pci_unregister_driver(&rtl8169_pci_driver);
-}
-
-module_init(rtl8169_init_module);
-module_exit(rtl8169_cleanup_module);
+module_pci_driver(rtl8169_pci_driver);