[PATCH 1/5] e1000e: cleanup several stats issues

[PATCH 1/5] e1000e: cleanup several stats issues

[Auke Kok]

From: Bruce Allan <bruce.w.allan@intel.com>

Several stats registers are completely unused and we just waste pci
bus time reading them. We also omit using the high 32 bits of the GORC/
GOTC counters. We can just read clear them and only read the low registers.

Mii-tool can also break es2lan if it executes a MII PHY register
ioctl while the device is in autonegotiation. Unfortunately it seems
that several applications and installations still perform this ioctl
call periodically and especially in this crucial startup time. We
can fool the ioctl by providing fail safe information that mimics
the "down" link state and only perform the dangerous PHY reads once
after link comes up to fill in the real values. As long as link
stays up the information will not change.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---

 drivers/net/e1000e/82571.c   |    6 --
 drivers/net/e1000e/defines.h |    2 +
 drivers/net/e1000e/e1000.h   |   25 +++++--
 drivers/net/e1000e/es2lan.c  |    2 -
 drivers/net/e1000e/ethtool.c |    6 +-
 drivers/net/e1000e/hw.h      |   12 +--
 drivers/net/e1000e/netdev.c  |  155 +++++++++++++++++++++++++-----------------
 7 files changed, 122 insertions(+), 86 deletions(-)

diff --git a/drivers/net/e1000e/82571.c b/drivers/net/e1000e/82571.c
index 01c8866..462351c 100644
--- a/drivers/net/e1000e/82571.c
+++ b/drivers/net/e1000e/82571.c
@@ -1326,12 +1326,10 @@ struct e1000_info e1000_82571_info = {
 	.mac			= e1000_82571,
 	.flags			= FLAG_HAS_HW_VLAN_FILTER
 				  | FLAG_HAS_JUMBO_FRAMES
-				  | FLAG_HAS_STATS_PTC_PRC
 				  | FLAG_HAS_WOL
 				  | FLAG_APME_IN_CTRL3
 				  | FLAG_RX_CSUM_ENABLED
 				  | FLAG_HAS_CTRLEXT_ON_LOAD
-				  | FLAG_HAS_STATS_ICR_ICT
 				  | FLAG_HAS_SMART_POWER_DOWN
 				  | FLAG_RESET_OVERWRITES_LAA /* errata */
 				  | FLAG_TARC_SPEED_MODE_BIT /* errata */
@@ -1347,12 +1345,10 @@ struct e1000_info e1000_82572_info = {
 	.mac			= e1000_82572,
 	.flags			= FLAG_HAS_HW_VLAN_FILTER
 				  | FLAG_HAS_JUMBO_FRAMES
-				  | FLAG_HAS_STATS_PTC_PRC
 				  | FLAG_HAS_WOL
 				  | FLAG_APME_IN_CTRL3
 				  | FLAG_RX_CSUM_ENABLED
 				  | FLAG_HAS_CTRLEXT_ON_LOAD
-				  | FLAG_HAS_STATS_ICR_ICT
 				  | FLAG_TARC_SPEED_MODE_BIT, /* errata */
 	.pba			= 38,
 	.get_variants		= e1000_get_variants_82571,
@@ -1365,11 +1361,9 @@ struct e1000_info e1000_82573_info = {
 	.mac			= e1000_82573,
 	.flags			= FLAG_HAS_HW_VLAN_FILTER
 				  | FLAG_HAS_JUMBO_FRAMES
-				  | FLAG_HAS_STATS_PTC_PRC
 				  | FLAG_HAS_WOL
 				  | FLAG_APME_IN_CTRL3
 				  | FLAG_RX_CSUM_ENABLED
-				  | FLAG_HAS_STATS_ICR_ICT
 				  | FLAG_HAS_SMART_POWER_DOWN
 				  | FLAG_HAS_AMT
 				  | FLAG_HAS_ERT
diff --git a/drivers/net/e1000e/defines.h b/drivers/net/e1000e/defines.h
index 572cfd4..4fb9d87 100644
--- a/drivers/net/e1000e/defines.h
+++ b/drivers/net/e1000e/defines.h
@@ -527,8 +527,10 @@
 #define PHY_ID2          0x03 /* Phy Id Reg (word 2) */
 #define PHY_AUTONEG_ADV  0x04 /* Autoneg Advertisement */
 #define PHY_LP_ABILITY   0x05 /* Link Partner Ability (Base Page) */
+#define PHY_AUTONEG_EXP  0x06 /* Autoneg Expansion Reg */
 #define PHY_1000T_CTRL   0x09 /* 1000Base-T Control Reg */
 #define PHY_1000T_STATUS 0x0A /* 1000Base-T Status Reg */
+#define PHY_EXT_STATUS   0x0F /* Extended Status Reg */
 
 /* NVM Control */
 #define E1000_EECD_SK        0x00000001 /* NVM Clock */
diff --git a/drivers/net/e1000e/e1000.h b/drivers/net/e1000e/e1000.h
index 5a89dff..4d3d1c2 100644
--- a/drivers/net/e1000e/e1000.h
+++ b/drivers/net/e1000e/e1000.h
@@ -147,6 +147,18 @@ struct e1000_ring {
 	struct e1000_queue_stats stats;
 };
 
+/* PHY register snapshot values */
+struct e1000_phy_regs {
+	u16 bmcr;		/* basic mode control register    */
+	u16 bmsr;		/* basic mode status register     */
+	u16 advertise;		/* auto-negotiation advertisement */
+	u16 lpa;		/* link partner ability register  */
+	u16 expansion;		/* auto-negotiation expansion reg */
+	u16 ctrl1000;		/* 1000BASE-T control register    */
+	u16 stat1000;		/* 1000BASE-T status register     */
+	u16 estatus;		/* extended status register       */
+};
+
 /* board specific private data structure */
 struct e1000_adapter {
 	struct timer_list watchdog_timer;
@@ -202,8 +214,8 @@ struct e1000_adapter {
 	/* Tx stats */
 	u64 tpt_old;
 	u64 colc_old;
-	u64 gotcl_old;
-	u32 gotcl;
+	u32 gotc;
+	u64 gotc_old;
 	u32 tx_timeout_count;
 	u32 tx_fifo_head;
 	u32 tx_head_addr;
@@ -227,8 +239,8 @@ struct e1000_adapter {
 	u64 hw_csum_err;
 	u64 hw_csum_good;
 	u64 rx_hdr_split;
-	u64 gorcl_old;
-	u32 gorcl;
+	u32 gorc;
+	u64 gorc_old;
 	u32 alloc_rx_buff_failed;
 	u32 rx_dma_failed;
 
@@ -250,6 +262,9 @@ struct e1000_adapter {
 	struct e1000_phy_info phy_info;
 	struct e1000_phy_stats phy_stats;
 
+	/* Snapshot of PHY registers */
+	struct e1000_phy_regs phy_regs;
+
 	struct e1000_ring test_tx_ring;
 	struct e1000_ring test_rx_ring;
 	u32 test_icr;
@@ -286,8 +301,6 @@ struct e1000_info {
 #define FLAG_HAS_CTRLEXT_ON_LOAD          (1 << 5)
 #define FLAG_HAS_SWSM_ON_LOAD             (1 << 6)
 #define FLAG_HAS_JUMBO_FRAMES             (1 << 7)
-#define FLAG_HAS_STATS_ICR_ICT            (1 << 9)
-#define FLAG_HAS_STATS_PTC_PRC            (1 << 10)
 #define FLAG_HAS_SMART_POWER_DOWN         (1 << 11)
 #define FLAG_IS_QUAD_PORT_A               (1 << 12)
 #define FLAG_IS_QUAD_PORT                 (1 << 13)
diff --git a/drivers/net/e1000e/es2lan.c b/drivers/net/e1000e/es2lan.c
index d59a99a..13a6f44 100644
--- a/drivers/net/e1000e/es2lan.c
+++ b/drivers/net/e1000e/es2lan.c
@@ -1231,12 +1231,10 @@ struct e1000_info e1000_es2_info = {
 	.mac			= e1000_80003es2lan,
 	.flags			= FLAG_HAS_HW_VLAN_FILTER
 				  | FLAG_HAS_JUMBO_FRAMES
-				  | FLAG_HAS_STATS_PTC_PRC
 				  | FLAG_HAS_WOL
 				  | FLAG_APME_IN_CTRL3
 				  | FLAG_RX_CSUM_ENABLED
 				  | FLAG_HAS_CTRLEXT_ON_LOAD
-				  | FLAG_HAS_STATS_ICR_ICT
 				  | FLAG_RX_NEEDS_RESTART /* errata */
 				  | FLAG_TARC_SET_BIT_ZERO /* errata */
 				  | FLAG_APME_CHECK_PORT_B
diff --git a/drivers/net/e1000e/ethtool.c b/drivers/net/e1000e/ethtool.c
index 6d1b257..c894a6f 100644
--- a/drivers/net/e1000e/ethtool.c
+++ b/drivers/net/e1000e/ethtool.c
@@ -46,8 +46,8 @@ struct e1000_stats {
 static const struct e1000_stats e1000_gstrings_stats[] = {
 	{ "rx_packets", E1000_STAT(stats.gprc) },
 	{ "tx_packets", E1000_STAT(stats.gptc) },
-	{ "rx_bytes", E1000_STAT(stats.gorcl) },
-	{ "tx_bytes", E1000_STAT(stats.gotcl) },
+	{ "rx_bytes", E1000_STAT(stats.gorc) },
+	{ "tx_bytes", E1000_STAT(stats.gotc) },
 	{ "rx_broadcast", E1000_STAT(stats.bprc) },
 	{ "tx_broadcast", E1000_STAT(stats.bptc) },
 	{ "rx_multicast", E1000_STAT(stats.mprc) },
@@ -83,7 +83,7 @@ static const struct e1000_stats e1000_gstrings_stats[] = {
 	{ "rx_flow_control_xoff", E1000_STAT(stats.xoffrxc) },
 	{ "tx_flow_control_xon", E1000_STAT(stats.xontxc) },
 	{ "tx_flow_control_xoff", E1000_STAT(stats.xofftxc) },
-	{ "rx_long_byte_count", E1000_STAT(stats.gorcl) },
+	{ "rx_long_byte_count", E1000_STAT(stats.gorc) },
 	{ "rx_csum_offload_good", E1000_STAT(hw_csum_good) },
 	{ "rx_csum_offload_errors", E1000_STAT(hw_csum_err) },
 	{ "rx_header_split", E1000_STAT(rx_hdr_split) },
diff --git a/drivers/net/e1000e/hw.h b/drivers/net/e1000e/hw.h
index 53f1ac6..a930e6d 100644
--- a/drivers/net/e1000e/hw.h
+++ b/drivers/net/e1000e/hw.h
@@ -592,10 +592,8 @@ struct e1000_hw_stats {
 	u64 bprc;
 	u64 mprc;
 	u64 gptc;
-	u64 gorcl;
-	u64 gorch;
-	u64 gotcl;
-	u64 gotch;
+	u64 gorc;
+	u64 gotc;
 	u64 rnbc;
 	u64 ruc;
 	u64 rfc;
@@ -604,10 +602,8 @@ struct e1000_hw_stats {
 	u64 mgprc;
 	u64 mgpdc;
 	u64 mgptc;
-	u64 torl;
-	u64 torh;
-	u64 totl;
-	u64 toth;
+	u64 tor;
+	u64 tot;
 	u64 tpr;
 	u64 tpt;
 	u64 ptc64;
diff --git a/drivers/net/e1000e/netdev.c b/drivers/net/e1000e/netdev.c
index c8dc47f..9d1143a 100644
--- a/drivers/net/e1000e/netdev.c
+++ b/drivers/net/e1000e/netdev.c
@@ -466,10 +466,10 @@ next_desc:
 	if (cleaned_count)
 		adapter->alloc_rx_buf(adapter, cleaned_count);
 
-	adapter->total_rx_packets += total_rx_packets;
 	adapter->total_rx_bytes += total_rx_bytes;
-	adapter->net_stats.rx_packets += total_rx_packets;
+	adapter->total_rx_packets += total_rx_packets;
 	adapter->net_stats.rx_bytes += total_rx_bytes;
+	adapter->net_stats.rx_packets += total_rx_packets;
 	return cleaned;
 }
 
@@ -606,8 +606,8 @@ static bool e1000_clean_tx_irq(struct e1000_adapter *adapter)
 	}
 	adapter->total_tx_bytes += total_tx_bytes;
 	adapter->total_tx_packets += total_tx_packets;
-	adapter->net_stats.tx_packets += total_tx_packets;
 	adapter->net_stats.tx_bytes += total_tx_bytes;
+	adapter->net_stats.tx_packets += total_tx_packets;
 	return cleaned;
 }
 
@@ -775,10 +775,10 @@ next_desc:
 	if (cleaned_count)
 		adapter->alloc_rx_buf(adapter, cleaned_count);
 
-	adapter->total_rx_packets += total_rx_packets;
 	adapter->total_rx_bytes += total_rx_bytes;
-	adapter->net_stats.rx_packets += total_rx_packets;
+	adapter->total_rx_packets += total_rx_packets;
 	adapter->net_stats.rx_bytes += total_rx_bytes;
+	adapter->net_stats.rx_packets += total_rx_packets;
 	return cleaned;
 }
 
@@ -2506,56 +2506,27 @@ void e1000e_update_stats(struct e1000_adapter *adapter)
 
 	adapter->stats.crcerrs += er32(CRCERRS);
 	adapter->stats.gprc += er32(GPRC);
-	adapter->stats.gorcl += er32(GORCL);
-	adapter->stats.gorch += er32(GORCH);
+	adapter->stats.gorc += er32(GORCL);
+	er32(GORCH); /* Clear gorc */
 	adapter->stats.bprc += er32(BPRC);
 	adapter->stats.mprc += er32(MPRC);
 	adapter->stats.roc += er32(ROC);
 
-	if (adapter->flags & FLAG_HAS_STATS_PTC_PRC) {
-		adapter->stats.prc64 += er32(PRC64);
-		adapter->stats.prc127 += er32(PRC127);
-		adapter->stats.prc255 += er32(PRC255);
-		adapter->stats.prc511 += er32(PRC511);
-		adapter->stats.prc1023 += er32(PRC1023);
-		adapter->stats.prc1522 += er32(PRC1522);
-		adapter->stats.symerrs += er32(SYMERRS);
-		adapter->stats.sec += er32(SEC);
-	}
-
 	adapter->stats.mpc += er32(MPC);
 	adapter->stats.scc += er32(SCC);
 	adapter->stats.ecol += er32(ECOL);
 	adapter->stats.mcc += er32(MCC);
 	adapter->stats.latecol += er32(LATECOL);
 	adapter->stats.dc += er32(DC);
-	adapter->stats.rlec += er32(RLEC);
 	adapter->stats.xonrxc += er32(XONRXC);
 	adapter->stats.xontxc += er32(XONTXC);
 	adapter->stats.xoffrxc += er32(XOFFRXC);
 	adapter->stats.xofftxc += er32(XOFFTXC);
-	adapter->stats.fcruc += er32(FCRUC);
 	adapter->stats.gptc += er32(GPTC);
-	adapter->stats.gotcl += er32(GOTCL);
-	adapter->stats.gotch += er32(GOTCH);
+	adapter->stats.gotc += er32(GOTCL);
+	er32(GOTCH); /* Clear gotc */
 	adapter->stats.rnbc += er32(RNBC);
 	adapter->stats.ruc += er32(RUC);
-	adapter->stats.rfc += er32(RFC);
-	adapter->stats.rjc += er32(RJC);
-	adapter->stats.torl += er32(TORL);
-	adapter->stats.torh += er32(TORH);
-	adapter->stats.totl += er32(TOTL);
-	adapter->stats.toth += er32(TOTH);
-	adapter->stats.tpr += er32(TPR);
-
-	if (adapter->flags & FLAG_HAS_STATS_PTC_PRC) {
-		adapter->stats.ptc64 += er32(PTC64);
-		adapter->stats.ptc127 += er32(PTC127);
-		adapter->stats.ptc255 += er32(PTC255);
-		adapter->stats.ptc511 += er32(PTC511);
-		adapter->stats.ptc1023 += er32(PTC1023);
-		adapter->stats.ptc1522 += er32(PTC1522);
-	}
 
 	adapter->stats.mptc += er32(MPTC);
 	adapter->stats.bptc += er32(BPTC);
@@ -2574,19 +2545,6 @@ void e1000e_update_stats(struct e1000_adapter *adapter)
 	adapter->stats.tsctc += er32(TSCTC);
 	adapter->stats.tsctfc += er32(TSCTFC);
 
-	adapter->stats.iac += er32(IAC);
-
-	if (adapter->flags & FLAG_HAS_STATS_ICR_ICT) {
-		adapter->stats.icrxoc += er32(ICRXOC);
-		adapter->stats.icrxptc += er32(ICRXPTC);
-		adapter->stats.icrxatc += er32(ICRXATC);
-		adapter->stats.ictxptc += er32(ICTXPTC);
-		adapter->stats.ictxatc += er32(ICTXATC);
-		adapter->stats.ictxqec += er32(ICTXQEC);
-		adapter->stats.ictxqmtc += er32(ICTXQMTC);
-		adapter->stats.icrxdmtc += er32(ICRXDMTC);
-	}
-
 	/* Fill out the OS statistics structure */
 	adapter->net_stats.multicast = adapter->stats.mprc;
 	adapter->net_stats.collisions = adapter->stats.colc;
@@ -2633,6 +2591,54 @@ void e1000e_update_stats(struct e1000_adapter *adapter)
 	spin_unlock_irqrestore(&adapter->stats_lock, irq_flags);
 }
 
+/**
+ * e1000_phy_read_status - Update the PHY register status snapshot
+ * @adapter: board private structure
+ **/
+static void e1000_phy_read_status(struct e1000_adapter *adapter)
+{
+	struct e1000_hw *hw = &adapter->hw;
+	struct e1000_phy_regs *phy = &adapter->phy_regs;
+	int ret_val;
+	unsigned long irq_flags;
+
+
+	spin_lock_irqsave(&adapter->stats_lock, irq_flags);
+
+	if ((er32(STATUS) & E1000_STATUS_LU) &&
+	    (adapter->hw.phy.media_type == e1000_media_type_copper)) {
+		ret_val  = e1e_rphy(hw, PHY_CONTROL, &phy->bmcr);
+		ret_val |= e1e_rphy(hw, PHY_STATUS, &phy->bmsr);
+		ret_val |= e1e_rphy(hw, PHY_AUTONEG_ADV, &phy->advertise);
+		ret_val |= e1e_rphy(hw, PHY_LP_ABILITY, &phy->lpa);
+		ret_val |= e1e_rphy(hw, PHY_AUTONEG_EXP, &phy->expansion);
+		ret_val |= e1e_rphy(hw, PHY_1000T_CTRL, &phy->ctrl1000);
+		ret_val |= e1e_rphy(hw, PHY_1000T_STATUS, &phy->stat1000);
+		ret_val |= e1e_rphy(hw, PHY_EXT_STATUS, &phy->estatus);
+		if (ret_val)
+			ndev_warn(adapter->netdev,
+				  "Error reading PHY register\n");
+	} else {
+		/*
+		 * Do not read PHY registers if link is not up
+		 * Set values to typical power-on defaults
+		 */
+		phy->bmcr = (BMCR_SPEED1000 | BMCR_ANENABLE | BMCR_FULLDPLX);
+		phy->bmsr = (BMSR_100FULL | BMSR_100HALF | BMSR_10FULL |
+			     BMSR_10HALF | BMSR_ESTATEN | BMSR_ANEGCAPABLE |
+			     BMSR_ERCAP);
+		phy->advertise = (ADVERTISE_PAUSE_ASYM | ADVERTISE_PAUSE_CAP |
+				  ADVERTISE_ALL | ADVERTISE_CSMA);
+		phy->lpa = 0;
+		phy->expansion = EXPANSION_ENABLENPAGE;
+		phy->ctrl1000 = ADVERTISE_1000FULL;
+		phy->stat1000 = 0;
+		phy->estatus = (ESTATUS_1000_TFULL | ESTATUS_1000_THALF);
+	}
+
+	spin_unlock_irqrestore(&adapter->stats_lock, irq_flags);
+}
+
 static void e1000_print_link_info(struct e1000_adapter *adapter)
 {
 	struct e1000_hw *hw = &adapter->hw;
@@ -2745,6 +2751,7 @@ static void e1000_watchdog_task(struct work_struct *work)
 		if (!netif_carrier_ok(netdev)) {
 			bool txb2b = 1;
 			/* update snapshot of PHY registers on LSC */
+			e1000_phy_read_status(adapter);
 			mac->ops.get_link_up_info(&adapter->hw,
 						   &adapter->link_speed,
 						   &adapter->link_duplex);
@@ -2842,10 +2849,10 @@ link_up:
 	mac->collision_delta = adapter->stats.colc - adapter->colc_old;
 	adapter->colc_old = adapter->stats.colc;
 
-	adapter->gorcl = adapter->stats.gorcl - adapter->gorcl_old;
-	adapter->gorcl_old = adapter->stats.gorcl;
-	adapter->gotcl = adapter->stats.gotcl - adapter->gotcl_old;
-	adapter->gotcl_old = adapter->stats.gotcl;
+	adapter->gorc = adapter->stats.gorc - adapter->gorc_old;
+	adapter->gorc_old = adapter->stats.gorc;
+	adapter->gotc = adapter->stats.gotc - adapter->gotc_old;
+	adapter->gotc_old = adapter->stats.gotc;
 
 	e1000e_update_adaptive(&adapter->hw);
 
@@ -3500,7 +3507,6 @@ static int e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr,
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct mii_ioctl_data *data = if_mii(ifr);
-	unsigned long irq_flags;
 
 	if (adapter->hw.phy.media_type != e1000_media_type_copper)
 		return -EOPNOTSUPP;
@@ -3512,13 +3518,40 @@ static int e1000_mii_ioctl(struct net_device *netdev, struct ifreq *ifr,
 	case SIOCGMIIREG:
 		if (!capable(CAP_NET_ADMIN))
 			return -EPERM;
-		spin_lock_irqsave(&adapter->stats_lock, irq_flags);
-		if (e1e_rphy(&adapter->hw, data->reg_num & 0x1F,
-				   &data->val_out)) {
-			spin_unlock_irqrestore(&adapter->stats_lock, irq_flags);
+		switch (data->reg_num & 0x1F) {
+		case MII_BMCR:
+			data->val_out = adapter->phy_regs.bmcr;
+			break;
+		case MII_BMSR:
+			data->val_out = adapter->phy_regs.bmsr;
+			break;
+		case MII_PHYSID1:
+			data->val_out = (adapter->hw.phy.id >> 16);
+			break;
+		case MII_PHYSID2:
+			data->val_out = (adapter->hw.phy.id & 0xFFFF);
+			break;
+		case MII_ADVERTISE:
+			data->val_out = adapter->phy_regs.advertise;
+			break;
+		case MII_LPA:
+			data->val_out = adapter->phy_regs.lpa;
+			break;
+		case MII_EXPANSION:
+			data->val_out = adapter->phy_regs.expansion;
+			break;
+		case MII_CTRL1000:
+			data->val_out = adapter->phy_regs.ctrl1000;
+			break;
+		case MII_STAT1000:
+			data->val_out = adapter->phy_regs.stat1000;
+			break;
+		case MII_ESTATUS:
+			data->val_out = adapter->phy_regs.estatus;
+			break;
+		default:
 			return -EIO;
 		}
-		spin_unlock_irqrestore(&adapter->stats_lock, irq_flags);
 		break;
 	case SIOCSMIIREG:
 	default:

[PATCH 2/5] e1000e: Add interrupt moderation run-time ethtool interface

[Auke Kok]

The ethtool -c / -C interface can now be used to modify the
irq moderation algorithm. This change does not require an
adapter reset and can thus be used at all times. The adapter
only supports changing/reading rx-usecs which has special
values for 0, 1 and 3:

0 - no irq moderation whatsoever
1 - normal moderation favoring regular mixed traffic (default)
3 - best attempt at low latency possible at cost of CPU

For values between 10 and 10000 the rx-usecs defines "the minimum
time between successive irqs" in usec, unlike the module parameter.

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---

 drivers/net/e1000e/e1000.h   |    3 +++
 drivers/net/e1000e/ethtool.c |   43 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 46 insertions(+), 0 deletions(-)

diff --git a/drivers/net/e1000e/e1000.h b/drivers/net/e1000e/e1000.h
index 4d3d1c2..79a426f 100644
--- a/drivers/net/e1000e/e1000.h
+++ b/drivers/net/e1000e/e1000.h
@@ -70,6 +70,9 @@ struct e1000_info;
 #define E1000_MAX_RXD			4096
 #define E1000_MIN_RXD			80
 
+#define E1000_MIN_ITR_USECS		10 /* 100000 irq/sec */
+#define E1000_MAX_ITR_USECS		10000 /* 100    irq/sec */
+
 /* Early Receive defines */
 #define E1000_ERT_2048			0x100
 
diff --git a/drivers/net/e1000e/ethtool.c b/drivers/net/e1000e/ethtool.c
index c894a6f..ce045ac 100644
--- a/drivers/net/e1000e/ethtool.c
+++ b/drivers/net/e1000e/ethtool.c
@@ -1770,6 +1770,47 @@ static int e1000_phys_id(struct net_device *netdev, u32 data)
 	return 0;
 }
 
+static int e1000_get_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *ec)
+{
+	struct e1000_adapter *adapter = netdev_priv(netdev);
+
+	if (adapter->itr_setting <= 3)
+		ec->rx_coalesce_usecs = adapter->itr_setting;
+	else
+		ec->rx_coalesce_usecs = 1000000 / adapter->itr_setting;
+
+	return 0;
+}
+
+static int e1000_set_coalesce(struct net_device *netdev,
+			      struct ethtool_coalesce *ec)
+{
+	struct e1000_adapter *adapter = netdev_priv(netdev);
+	struct e1000_hw *hw = &adapter->hw;
+
+	if ((ec->rx_coalesce_usecs > E1000_MAX_ITR_USECS) ||
+	    ((ec->rx_coalesce_usecs > 3) &&
+	     (ec->rx_coalesce_usecs < E1000_MIN_ITR_USECS)) ||
+	    (ec->rx_coalesce_usecs == 2))
+		return -EINVAL;
+
+	if (ec->rx_coalesce_usecs <= 3) {
+		adapter->itr = 20000;
+		adapter->itr_setting = ec->rx_coalesce_usecs;
+	} else {
+		adapter->itr = (1000000 / ec->rx_coalesce_usecs);
+		adapter->itr_setting = adapter->itr & ~3;
+	}
+
+	if (adapter->itr_setting != 0)
+		ew32(ITR, 1000000000 / (adapter->itr * 256));
+	else
+		ew32(ITR, 0);
+
+	return 0;
+}
+
 static int e1000_nway_reset(struct net_device *netdev)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
@@ -1845,6 +1886,8 @@ static const struct ethtool_ops e1000_ethtool_ops = {
 	.phys_id		= e1000_phys_id,
 	.get_ethtool_stats	= e1000_get_ethtool_stats,
 	.get_sset_count		= e1000e_get_sset_count,
+	.get_coalesce		= e1000_get_coalesce,
+	.set_coalesce		= e1000_set_coalesce,
 };
 
 void e1000e_set_ethtool_ops(struct net_device *netdev)

Re: [PATCH 2/5] e1000e: Add interrupt moderation run-time ethtool interface

[Andi Kleen]

Auke Kok <auke-jan.h.kok@intel.com> writes:

> The ethtool -c / -C interface can now be used to modify the
> irq moderation algorithm. This change does not require an
> adapter reset and can thus be used at all times. The adapter
> only supports changing/reading rx-usecs which has special
> values for 0, 1 and 3:
>
> 0 - no irq moderation whatsoever
> 1 - normal moderation favoring regular mixed traffic (default)
> 3 - best attempt at low latency possible at cost of CPU

It would be nice if ethtool instead of having this as a obscure magic
number had a higher level "--low-latency" or similar parameter
(that could also potentially set similar settings on other NICs)

You think there would be a clean way to implement this? 

-Andi

Re: [PATCH 2/5] e1000e: Add interrupt moderation run-time ethtool interface

[Rick Jones]

Andi Kleen wrote:
> Auke Kok <auke-jan.h.kok@intel.com> writes:
> 
> 
>>The ethtool -c / -C interface can now be used to modify the
>>irq moderation algorithm. This change does not require an
>>adapter reset and can thus be used at all times. The adapter
>>only supports changing/reading rx-usecs which has special
>>values for 0, 1 and 3:
>>
>>0 - no irq moderation whatsoever
>>1 - normal moderation favoring regular mixed traffic (default)
>>3 - best attempt at low latency possible at cost of CPU
> 
> 
> It would be nice if ethtool instead of having this as a obscure magic
> number had a higher level "--low-latency" or similar parameter
> (that could also potentially set similar settings on other NICs)
> 
> You think there would be a clean way to implement this? 

The biggest problems might be ones of interpretation as to whether a 
given setting was indeed:

--favor-latency
--favor-throutput
--favor-mixed

But it would indeed at one level be more "clear" to a random admin (IMO).

I use "--favor-mumble" above since what might be the --low-latency 
through one NIC might not be as low as the "--low-latency" though 
another NIC, and it also sort of implicitly suggests a trade-off is 
being made.

rick jones

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [PATCH 2/5] e1000e: Add interrupt moderation run-time ethtool interface

[Jeff Garzik]

Andi Kleen wrote:
> Auke Kok <auke-jan.h.kok@intel.com> writes:
> 
>> The ethtool -c / -C interface can now be used to modify the
>> irq moderation algorithm. This change does not require an
>> adapter reset and can thus be used at all times. The adapter
>> only supports changing/reading rx-usecs which has special
>> values for 0, 1 and 3:
>>
>> 0 - no irq moderation whatsoever
>> 1 - normal moderation favoring regular mixed traffic (default)
>> 3 - best attempt at low latency possible at cost of CPU
> 
> It would be nice if ethtool instead of having this as a obscure magic
> number had a higher level "--low-latency" or similar parameter
> (that could also potentially set similar settings on other NICs)

That would be fine with me -- that's one of the goals of ethtool:  make 
the setting of hardware-specific settings as generic and easy as possible.

	Jeff




-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [PATCH 2/5] e1000e: Add interrupt moderation run-time ethtool interface

[Andi Kleen]

Jeff Garzik wrote:
> Andi Kleen wrote:
>> Auke Kok <auke-jan.h.kok@intel.com> writes:
>>
>>> The ethtool -c / -C interface can now be used to modify the
>>> irq moderation algorithm. This change does not require an
>>> adapter reset and can thus be used at all times. The adapter
>>> only supports changing/reading rx-usecs which has special
>>> values for 0, 1 and 3:
>>>
>>> 0 - no irq moderation whatsoever
>>> 1 - normal moderation favoring regular mixed traffic (default)
>>> 3 - best attempt at low latency possible at cost of CPU
>>
>> It would be nice if ethtool instead of having this as a obscure magic
>> number had a higher level "--low-latency" or similar parameter
>> (that could also potentially set similar settings on other NICs)
> 
> That would be fine with me -- that's one of the goals of ethtool:  make
> the setting of hardware-specific settings as generic and easy as possible.

I think this would require new ethtool interfaces (NIC_POLICY ?) so that
the driver can decide what the right settings are.

Otherwise you would need to hard code a conversion table into the user
program, which is probably a bad idea.

-Andi

Re: [PATCH 2/5] e1000e: Add interrupt moderation run-time ethtool interface

[Kok, Auke]

Jeff Garzik wrote:
> Andi Kleen wrote:
>> Auke Kok <auke-jan.h.kok@intel.com> writes:
>>
>>> The ethtool -c / -C interface can now be used to modify the
>>> irq moderation algorithm. This change does not require an
>>> adapter reset and can thus be used at all times. The adapter
>>> only supports changing/reading rx-usecs which has special
>>> values for 0, 1 and 3:
>>>
>>> 0 - no irq moderation whatsoever
>>> 1 - normal moderation favoring regular mixed traffic (default)
>>> 3 - best attempt at low latency possible at cost of CPU
>>
>> It would be nice if ethtool instead of having this as a obscure magic
>> number had a higher level "--low-latency" or similar parameter
>> (that could also potentially set similar settings on other NICs)
> 
> That would be fine with me -- that's one of the goals of ethtool:  make
> the setting of hardware-specific settings as generic and easy as possible.

yes I think that's a good addition, and might hook better into power management
and other high-level configuration tools.

I'm all for that :)

Auke

[PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Auke Kok]

Fix TSO over VLAN's by propagating settings to our VLAN devices.

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---

 drivers/net/e1000e/ethtool.c |   16 ++++++++++++++++
 drivers/net/e1000e/netdev.c  |    8 ++++++++
 2 files changed, 24 insertions(+), 0 deletions(-)

diff --git a/drivers/net/e1000e/ethtool.c b/drivers/net/e1000e/ethtool.c
index ce045ac..23e9703 100644
--- a/drivers/net/e1000e/ethtool.c
+++ b/drivers/net/e1000e/ethtool.c
@@ -32,6 +32,7 @@
 #include <linux/ethtool.h>
 #include <linux/pci.h>
 #include <linux/delay.h>
+#include <linux/if_vlan.h>
 
 #include "e1000.h"
 
@@ -371,6 +372,8 @@ static int e1000_set_tx_csum(struct net_device *netdev, u32 data)
 static int e1000_set_tso(struct net_device *netdev, u32 data)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
+	int i;
+	struct net_device *v_netdev;
 
 	if (data) {
 		netdev->features |= NETIF_F_TSO;
@@ -378,8 +381,21 @@ static int e1000_set_tso(struct net_device *netdev, u32 data)
 	} else {
 		netdev->features &= ~NETIF_F_TSO;
 		netdev->features &= ~NETIF_F_TSO6;
+		/* disable TSO on all VLANs if they're present */
+		if (!adapter->vlgrp)
+			goto tso_out;
+		for (i = 0; i < VLAN_GROUP_ARRAY_LEN; i++) {
+			v_netdev = vlan_group_get_device(adapter->vlgrp, i);
+			if (!v_netdev)
+				continue;
+
+			v_netdev->features &= ~NETIF_F_TSO;
+			v_netdev->features &= ~NETIF_F_TSO6;
+			vlan_group_set_device(adapter->vlgrp, i, v_netdev);
+		}
 	}
 
+tso_out:
 	ndev_info(netdev, "TSO is %s\n",
 		  data ? "Enabled" : "Disabled");
 	adapter->flags |= FLAG_TSO_FORCE;
diff --git a/drivers/net/e1000e/netdev.c b/drivers/net/e1000e/netdev.c
index 9d1143a..146edf7 100644
--- a/drivers/net/e1000e/netdev.c
+++ b/drivers/net/e1000e/netdev.c
@@ -1432,6 +1432,7 @@ static void e1000_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
 	struct e1000_adapter *adapter = netdev_priv(netdev);
 	struct e1000_hw *hw = &adapter->hw;
 	u32 vfta, index;
+	struct net_device *v_netdev;
 
 	/* don't update vlan cookie if already programmed */
 	if ((adapter->hw.mng_cookie.status &
@@ -1443,6 +1444,13 @@ static void e1000_vlan_rx_add_vid(struct net_device *netdev, u16 vid)
 	vfta = E1000_READ_REG_ARRAY(hw, E1000_VFTA, index);
 	vfta |= (1 << (vid & 0x1F));
 	e1000e_write_vfta(hw, index, vfta);
+	/*
+	 * Copy feature flags from netdev to the vlan netdev for this vid.
+	 * This allows things like TSO to bubble down to our vlan device.
+	 */
+	v_netdev = vlan_group_get_device(adapter->vlgrp, vid);
+	v_netdev->features |= adapter->netdev->features;
+	vlan_group_set_device(adapter->vlgrp, vid, v_netdev);
 }
 
 static void e1000_vlan_rx_kill_vid(struct net_device *netdev, u16 vid)

Re: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Patrick McHardy]

Auke Kok wrote:
> Fix TSO over VLAN's by propagating settings to our VLAN devices.
>   

> @@ -378,8 +381,21 @@ static int e1000_set_tso(struct net_device *netdev, u32 data)
>  	} else {
>  		netdev->features &= ~NETIF_F_TSO;
>  		netdev->features &= ~NETIF_F_TSO6;
> +		/* disable TSO on all VLANs if they're present */
> +		if (!adapter->vlgrp)
> +			goto tso_out;
> +		for (i = 0; i < VLAN_GROUP_ARRAY_LEN; i++) {
> +			v_netdev = vlan_group_get_device(adapter->vlgrp, i);
> +			if (!v_netdev)
> +				continue;
> +
> +			v_netdev->features &= ~NETIF_F_TSO;
> +			v_netdev->features &= ~NETIF_F_TSO6;
> +			vlan_group_set_device(adapter->vlgrp, i, v_netdev);
>   

What exactly is this supposed to fix? If this simply wants
to propagate feature changes, I think it should use
netdev_feat_change and handle that within the VLAN code.

Re: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Kok, Auke]

Patrick McHardy wrote:
> Auke Kok wrote:
>> Fix TSO over VLAN's by propagating settings to our VLAN devices.
>>   
> 
>> @@ -378,8 +381,21 @@ static int e1000_set_tso(struct net_device
>> *netdev, u32 data)
>>      } else {
>>          netdev->features &= ~NETIF_F_TSO;
>>          netdev->features &= ~NETIF_F_TSO6;
>> +        /* disable TSO on all VLANs if they're present */
>> +        if (!adapter->vlgrp)
>> +            goto tso_out;
>> +        for (i = 0; i < VLAN_GROUP_ARRAY_LEN; i++) {
>> +            v_netdev = vlan_group_get_device(adapter->vlgrp, i);
>> +            if (!v_netdev)
>> +                continue;
>> +
>> +            v_netdev->features &= ~NETIF_F_TSO;
>> +            v_netdev->features &= ~NETIF_F_TSO6;
>> +            vlan_group_set_device(adapter->vlgrp, i, v_netdev);
>>   
> 
> What exactly is this supposed to fix? If this simply wants
> to propagate feature changes, I think it should use
> netdev_feat_change and handle that within the VLAN code.

I asked PJ and got this reply:

//
VLAN devices didn't get the parent's feature flags on creation.  I went
to fix this in the kernel, people pushed back that some devices couldn't
support both VLAN tag insertion offload and TSO, so I didn't push the
issue.  I worked around the issue by copying the flags in the driver.
The downside is when we turn off TSO using ethtool, we need to remove
TSO from all VLAN devices, since the hardware segmenter is no longer
available if the parent device doesn't have it enabled as a feature.  We
were seeing stack-based panics with gso_segment(), which was corrected
by removing the TSO flag from all VLAN devices.

I can't seem to find netdev_feat_change anywhere in the kernel, or
variants of that name, so I'm not sure what Patrick is pointing us at.

-PJ
//

Re: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Patrick McHardy]

Kok, Auke wrote:
> Patrick McHardy wrote:
>> Auke Kok wrote:
>>> Fix TSO over VLAN's by propagating settings to our VLAN devices.
>>>   
>> What exactly is this supposed to fix? If this simply wants
>> to propagate feature changes, I think it should use
>> netdev_feat_change and handle that within the VLAN code.
> 
> I asked PJ and got this reply:
> 
> //
> VLAN devices didn't get the parent's feature flags on creation.  I went
> to fix this in the kernel, people pushed back that some devices couldn't
> support both VLAN tag insertion offload and TSO, so I didn't push the
> issue.  I worked around the issue by copying the flags in the driver.
> The downside is when we turn off TSO using ethtool, we need to remove
> TSO from all VLAN devices, since the hardware segmenter is no longer
> available if the parent device doesn't have it enabled as a feature.  We
> were seeing stack-based panics with gso_segment(), which was corrected
> by removing the TSO flag from all VLAN devices.
> 
> I can't seem to find netdev_feat_change anywhere in the kernel, or
> variants of that name, so I'm not sure what Patrick is pointing us at.
> 
> -PJ
> //

That was a typo, I meant netdev_features_change(), which is invoked
by dev_ethtool() and calls the netdev notifier with NETDEV_FEAT_CHANGE,
on which the VLAN code could react. I wasn't aware of the TSO + VLAN
acceleration problems you've mentioned, do you have a pointer to more
information about this?

In any case I would prefer to avoid having drivers mess with VLAN
device flags. How about adding a device flag indicating that the
driver supports TSO + VLAN acceleration and using the
NETDEV_FEAT_CHANGE notification inside the VLAN code do adjust
the device's flags properly?

RE: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Waskiewicz Jr, Peter P]

> That was a typo, I meant netdev_features_change(), which is invoked
> by dev_ethtool() and calls the netdev notifier with
NETDEV_FEAT_CHANGE,
> on which the VLAN code could react. I wasn't aware of the TSO + VLAN
> acceleration problems you've mentioned, do you have a pointer to more
> information about this?

This was a conversation I had late last year on the list regarding TSO
and Tx checksum offload not getting turned on for VLAN devices, since
the VLAN device creation code doesn't copy those device feature flags.
I had proposed sending a patch, and the statement was made that some
devices might not support both of those hardware features at the same
time.  I'll try to dig up those mails.

> In any case I would prefer to avoid having drivers mess with VLAN
> device flags. How about adding a device flag indicating that the
> driver supports TSO + VLAN acceleration and using the
> NETDEV_FEAT_CHANGE notification inside the VLAN code do adjust
> the device's flags properly?

I hear you loud and clear.  I wanted to do this the "right" way in the
drivers if I couldn't do it in the kernel, and your suggestions sound
like the right way.  Let me take the suggestions here and respin this
part of the code, since I have this change in e1000, e1000e, igb, and
ixgbe, and want it to be correct.

Thanks Patrick for the feedback here.  Hopefully I'll soon have
something for Auke to resend.

Cheers,
-PJ Waskiewicz

RE: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Waskiewicz Jr, Peter P]

> That was a typo, I meant netdev_features_change(), which is invoked
> by dev_ethtool() and calls the netdev notifier with 
> NETDEV_FEAT_CHANGE,
> on which the VLAN code could react. I wasn't aware of the TSO + VLAN
> acceleration problems you've mentioned, do you have a pointer to more
> information about this?
> 
> In any case I would prefer to avoid having drivers mess with VLAN
> device flags. How about adding a device flag indicating that the
> driver supports TSO + VLAN acceleration and using the
> NETDEV_FEAT_CHANGE notification inside the VLAN code do adjust
> the device's flags properly?

I've been looking into this, and I agree this makes sense when creating
a new VLAN to get the device flags to the VLAN device.  However, we need
to get the TSO and/or Tx checksum offload feature flags off all VLAN
devices when the parent device has it turned off.  I don't see any clean
way of doing this with netdev_features_change().  I'd really love to
implement this correctly, so if you have any ideas on how to handle this
latter case, I'm all ears.  :-)

Thanks Patrick,
-PJ Waskiewicz

Re: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Patrick McHardy]

Waskiewicz Jr, Peter P wrote:
>> In any case I would prefer to avoid having drivers mess with VLAN
>> device flags. How about adding a device flag indicating that the
>> driver supports TSO + VLAN acceleration and using the
>> NETDEV_FEAT_CHANGE notification inside the VLAN code do adjust
>> the device's flags properly?
> 
> I've been looking into this, and I agree this makes sense when creating
> a new VLAN to get the device flags to the VLAN device.  However, we need
> to get the TSO and/or Tx checksum offload feature flags off all VLAN
> devices when the parent device has it turned off.  I don't see any clean
> way of doing this with netdev_features_change().  I'd really love to
> implement this correctly, so if you have any ideas on how to handle this
> latter case, I'm all ears.  :-)


Whats the problem with doing:

- on VLAN creation: copy VLAN flag from lower device if lowerdev
   has NETIF_F_<flag indicating non-broken TSO+VLAN accel>

- on change of TSO/checksumming flags of lower device:
   use netdevice notifier to adjust the settings for all
   VLANs on that device

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

Re: [PATCH 3/5] e1000e: Allow TSO to trickle down to VLAN device

[Patrick McHardy]

Patrick McHardy wrote:
> Waskiewicz Jr, Peter P wrote:
>>> In any case I would prefer to avoid having drivers mess with VLAN
>>> device flags. How about adding a device flag indicating that the
>>> driver supports TSO + VLAN acceleration and using the
>>> NETDEV_FEAT_CHANGE notification inside the VLAN code do adjust
>>> the device's flags properly?
>>
>> I've been looking into this, and I agree this makes sense when creating
>> a new VLAN to get the device flags to the VLAN device.  However, we need
>> to get the TSO and/or Tx checksum offload feature flags off all VLAN
>> devices when the parent device has it turned off.  I don't see any clean
>> way of doing this with netdev_features_change().  I'd really love to
>> implement this correctly, so if you have any ideas on how to handle this
>> latter case, I'm all ears.  :-)
> 
> 
> Whats the problem with doing:
> 
> - on VLAN creation: copy VLAN flag from lower device if lowerdev
                            ^^^^ TSO/checksumming flags
>   has NETIF_F_<flag indicating non-broken TSO+VLAN accel>
> 
> - on change of TSO/checksumming flags of lower device:
>   use netdevice notifier to adjust the settings for all
>   VLANs on that device

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

[PATCH 4/5] e1000e: Fix HW Error on es2lan, ARP capture issue by BMC

[Auke Kok]

From: David Graham <david.graham@intel.com>

Several components to this complex fix. The es2lan cards occasionally
gave a "HW Error" especially when forcing speed. Some users also
reported that the BMC stole ARP packets.

The fixes include setting the proper SW_FW bits to tell the BMC
that we're active and not do any un-initialization at all, so the
setup routine is largely changed.

Signed-off-by: David Graham <david.graham@intel.com>
Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---

 drivers/net/e1000e/defines.h |    1 
 drivers/net/e1000e/e1000.h   |    2 +
 drivers/net/e1000e/es2lan.c  |  127 +++++++++++++++++++++++++++++++-----------
 drivers/net/e1000e/phy.c     |   73 +++++++++++++-----------
 4 files changed, 137 insertions(+), 66 deletions(-)

diff --git a/drivers/net/e1000e/defines.h b/drivers/net/e1000e/defines.h
index 4fb9d87..2a53875 100644
--- a/drivers/net/e1000e/defines.h
+++ b/drivers/net/e1000e/defines.h
@@ -184,6 +184,7 @@
 #define E1000_SWFW_EEP_SM   0x1
 #define E1000_SWFW_PHY0_SM  0x2
 #define E1000_SWFW_PHY1_SM  0x4
+#define E1000_SWFW_CSR_SM   0x8
 
 /* Device Control */
 #define E1000_CTRL_FD       0x00000001  /* Full duplex.0=half; 1=full */
diff --git a/drivers/net/e1000e/e1000.h b/drivers/net/e1000e/e1000.h
index 79a426f..f6835c3 100644
--- a/drivers/net/e1000e/e1000.h
+++ b/drivers/net/e1000e/e1000.h
@@ -449,6 +449,8 @@ extern s32 e1000e_read_kmrn_reg(struct e1000_hw *hw, u32 offset, u16 *data);
 extern s32 e1000e_phy_has_link_generic(struct e1000_hw *hw, u32 iterations,
 			       u32 usec_interval, bool *success);
 extern s32 e1000e_phy_reset_dsp(struct e1000_hw *hw);
+extern s32 e1000e_read_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 *data);
+extern s32 e1000e_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data);
 extern s32 e1000e_check_downshift(struct e1000_hw *hw);
 
 static inline s32 e1000_phy_hw_reset(struct e1000_hw *hw)
diff --git a/drivers/net/e1000e/es2lan.c b/drivers/net/e1000e/es2lan.c
index 13a6f44..dc552d7 100644
--- a/drivers/net/e1000e/es2lan.c
+++ b/drivers/net/e1000e/es2lan.c
@@ -41,6 +41,7 @@
 #define E1000_KMRNCTRLSTA_OFFSET_FIFO_CTRL	 0x00
 #define E1000_KMRNCTRLSTA_OFFSET_INB_CTRL	 0x02
 #define E1000_KMRNCTRLSTA_OFFSET_HD_CTRL	 0x10
+#define E1000_KMRNCTRLSTA_OFFSET_MAC2PHY_OPMODE	 0x1F
 
 #define E1000_KMRNCTRLSTA_FIFO_CTRL_RX_BYPASS	 0x0008
 #define E1000_KMRNCTRLSTA_FIFO_CTRL_TX_BYPASS	 0x0800
@@ -48,6 +49,7 @@
 
 #define E1000_KMRNCTRLSTA_HD_CTRL_10_100_DEFAULT 0x0004
 #define E1000_KMRNCTRLSTA_HD_CTRL_1000_DEFAULT	 0x0000
+#define E1000_KMRNCTRLSTA_OPMODE_E_IDLE		 0x2000
 
 #define E1000_TCTL_EXT_GCEX_MASK 0x000FFC00 /* Gigabit Carry Extend Padding */
 #define DEFAULT_TCTL_EXT_GCEX_80003ES2LAN	 0x00010000
@@ -85,6 +87,9 @@
 /* Kumeran Mode Control Register (Page 193, Register 16) */
 #define GG82563_KMCR_PASS_FALSE_CARRIER		 0x0800
 
+/* Max number of times Kumeran read/write should be validated */
+#define GG82563_MAX_KMRN_RETRY  0x5
+
 /* Power Management Control Register (Page 193, Register 20) */
 #define GG82563_PMCR_ENABLE_ELECTRICAL_IDLE	 0x0001
 					   /* 1=Enable SERDES Electrical Idle */
@@ -270,6 +275,7 @@ static s32 e1000_acquire_phy_80003es2lan(struct e1000_hw *hw)
 	u16 mask;
 
 	mask = hw->bus.func ? E1000_SWFW_PHY1_SM : E1000_SWFW_PHY0_SM;
+	mask |= E1000_SWFW_CSR_SM;
 
 	return e1000_acquire_swfw_sync_80003es2lan(hw, mask);
 }
@@ -286,6 +292,8 @@ static void e1000_release_phy_80003es2lan(struct e1000_hw *hw)
 	u16 mask;
 
 	mask = hw->bus.func ? E1000_SWFW_PHY1_SM : E1000_SWFW_PHY0_SM;
+	mask |= E1000_SWFW_CSR_SM;
+
 	e1000_release_swfw_sync_80003es2lan(hw, mask);
 }
 
@@ -410,20 +418,27 @@ static s32 e1000_read_phy_reg_gg82563_80003es2lan(struct e1000_hw *hw,
 	u32 page_select;
 	u16 temp;
 
+	ret_val = e1000_acquire_phy_80003es2lan(hw);
+	if (ret_val)
+		return ret_val;
+
 	/* Select Configuration Page */
-	if ((offset & MAX_PHY_REG_ADDRESS) < GG82563_MIN_ALT_REG)
+	if ((offset & MAX_PHY_REG_ADDRESS) < GG82563_MIN_ALT_REG) {
 		page_select = GG82563_PHY_PAGE_SELECT;
-	else
+	} else {
 		/*
 		 * Use Alternative Page Select register to access
 		 * registers 30 and 31
 		 */
 		page_select = GG82563_PHY_PAGE_SELECT_ALT;
+	}
 
 	temp = (u16)((u16)offset >> GG82563_PAGE_SHIFT);
-	ret_val = e1000e_write_phy_reg_m88(hw, page_select, temp);
-	if (ret_val)
+	ret_val = e1000e_write_phy_reg_mdic(hw, page_select, temp);
+	if (ret_val) {
+		e1000_release_phy_80003es2lan(hw);
 		return ret_val;
+	}
 
 	/*
 	 * The "ready" bit in the MDIC register may be incorrectly set
@@ -433,20 +448,21 @@ static s32 e1000_read_phy_reg_gg82563_80003es2lan(struct e1000_hw *hw,
 	udelay(200);
 
 	/* ...and verify the command was successful. */
-	ret_val = e1000e_read_phy_reg_m88(hw, page_select, &temp);
+	ret_val = e1000e_read_phy_reg_mdic(hw, page_select, &temp);
 
 	if (((u16)offset >> GG82563_PAGE_SHIFT) != temp) {
 		ret_val = -E1000_ERR_PHY;
+		e1000_release_phy_80003es2lan(hw);
 		return ret_val;
 	}
 
 	udelay(200);
 
-	ret_val = e1000e_read_phy_reg_m88(hw,
-					 MAX_PHY_REG_ADDRESS & offset,
-					 data);
+	ret_val = e1000e_read_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset,
+					   data);
 
 	udelay(200);
+	e1000_release_phy_80003es2lan(hw);
 
 	return ret_val;
 }
@@ -467,20 +483,27 @@ static s32 e1000_write_phy_reg_gg82563_80003es2lan(struct e1000_hw *hw,
 	u32 page_select;
 	u16 temp;
 
+	ret_val = e1000_acquire_phy_80003es2lan(hw);
+	if (ret_val)
+		return ret_val;
+
 	/* Select Configuration Page */
-	if ((offset & MAX_PHY_REG_ADDRESS) < GG82563_MIN_ALT_REG)
+	if ((offset & MAX_PHY_REG_ADDRESS) < GG82563_MIN_ALT_REG) {
 		page_select = GG82563_PHY_PAGE_SELECT;
-	else
+	} else {
 		/*
 		 * Use Alternative Page Select register to access
 		 * registers 30 and 31
 		 */
 		page_select = GG82563_PHY_PAGE_SELECT_ALT;
+	}
 
 	temp = (u16)((u16)offset >> GG82563_PAGE_SHIFT);
-	ret_val = e1000e_write_phy_reg_m88(hw, page_select, temp);
-	if (ret_val)
+	ret_val = e1000e_write_phy_reg_mdic(hw, page_select, temp);
+	if (ret_val) {
+		e1000_release_phy_80003es2lan(hw);
 		return ret_val;
+	}
 
 
 	/*
@@ -491,18 +514,20 @@ static s32 e1000_write_phy_reg_gg82563_80003es2lan(struct e1000_hw *hw,
 	udelay(200);
 
 	/* ...and verify the command was successful. */
-	ret_val = e1000e_read_phy_reg_m88(hw, page_select, &temp);
+	ret_val = e1000e_read_phy_reg_mdic(hw, page_select, &temp);
 
-	if (((u16)offset >> GG82563_PAGE_SHIFT) != temp)
+	if (((u16)offset >> GG82563_PAGE_SHIFT) != temp) {
+		e1000_release_phy_80003es2lan(hw);
 		return -E1000_ERR_PHY;
+	}
 
 	udelay(200);
 
-	ret_val = e1000e_write_phy_reg_m88(hw,
-					  MAX_PHY_REG_ADDRESS & offset,
-					  data);
+	ret_val = e1000e_write_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset,
+					    data);
 
 	udelay(200);
+	e1000_release_phy_80003es2lan(hw);
 
 	return ret_val;
 }
@@ -882,10 +907,10 @@ static s32 e1000_copper_link_setup_gg82563_80003es2lan(struct e1000_hw *hw)
 	struct e1000_phy_info *phy = &hw->phy;
 	s32 ret_val;
 	u32 ctrl_ext;
-	u16 data;
+	u32 i = 0;
+	u16 data, data2;
 
-	ret_val = e1e_rphy(hw, GG82563_PHY_MAC_SPEC_CTRL,
-				     &data);
+	ret_val = e1e_rphy(hw, GG82563_PHY_MAC_SPEC_CTRL, &data);
 	if (ret_val)
 		return ret_val;
 
@@ -893,8 +918,7 @@ static s32 e1000_copper_link_setup_gg82563_80003es2lan(struct e1000_hw *hw)
 	/* Use 25MHz for both link down and 1000Base-T for Tx clock. */
 	data |= GG82563_MSCR_TX_CLK_1000MBPS_25;
 
-	ret_val = e1e_wphy(hw, GG82563_PHY_MAC_SPEC_CTRL,
-				      data);
+	ret_val = e1e_wphy(hw, GG82563_PHY_MAC_SPEC_CTRL, data);
 	if (ret_val)
 		return ret_val;
 
@@ -954,6 +978,18 @@ static s32 e1000_copper_link_setup_gg82563_80003es2lan(struct e1000_hw *hw)
 	if (ret_val)
 		return ret_val;
 
+	ret_val = e1000e_read_kmrn_reg(hw,
+				       E1000_KMRNCTRLSTA_OFFSET_MAC2PHY_OPMODE,
+				       &data);
+	if (ret_val)
+		return ret_val;
+	data |= E1000_KMRNCTRLSTA_OPMODE_E_IDLE;
+	ret_val = e1000e_write_kmrn_reg(hw,
+					E1000_KMRNCTRLSTA_OFFSET_MAC2PHY_OPMODE,
+					data);
+	if (ret_val)
+		return ret_val;
+
 	ret_val = e1e_rphy(hw, GG82563_PHY_SPEC_CTRL_2, &data);
 	if (ret_val)
 		return ret_val;
@@ -983,9 +1019,18 @@ static s32 e1000_copper_link_setup_gg82563_80003es2lan(struct e1000_hw *hw)
 		if (ret_val)
 			return ret_val;
 
-		ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &data);
-		if (ret_val)
-			return ret_val;
+		do {
+			ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL,
+					   &data);
+			if (ret_val)
+				return ret_val;
+
+			ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL,
+					   &data2);
+			if (ret_val)
+				return ret_val;
+			i++;
+		} while ((data != data2) && (i < GG82563_MAX_KMRN_RETRY));
 
 		data &= ~GG82563_KMCR_PASS_FALSE_CARRIER;
 		ret_val = e1e_wphy(hw, GG82563_PHY_KMRN_MODE_CTRL, data);
@@ -1074,7 +1119,8 @@ static s32 e1000_cfg_kmrn_10_100_80003es2lan(struct e1000_hw *hw, u16 duplex)
 {
 	s32 ret_val;
 	u32 tipg;
-	u16 reg_data;
+	u32 i = 0;
+	u16 reg_data, reg_data2;
 
 	reg_data = E1000_KMRNCTRLSTA_HD_CTRL_10_100_DEFAULT;
 	ret_val = e1000e_write_kmrn_reg(hw, E1000_KMRNCTRLSTA_OFFSET_HD_CTRL,
@@ -1088,9 +1134,16 @@ static s32 e1000_cfg_kmrn_10_100_80003es2lan(struct e1000_hw *hw, u16 duplex)
 	tipg |= DEFAULT_TIPG_IPGT_10_100_80003ES2LAN;
 	ew32(TIPG, tipg);
 
-	ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &reg_data);
-	if (ret_val)
-		return ret_val;
+	do {
+		ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &reg_data);
+		if (ret_val)
+			return ret_val;
+
+		ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &reg_data2);
+		if (ret_val)
+			return ret_val;
+		i++;
+	} while ((reg_data != reg_data2) && (i < GG82563_MAX_KMRN_RETRY));
 
 	if (duplex == HALF_DUPLEX)
 		reg_data |= GG82563_KMCR_PASS_FALSE_CARRIER;
@@ -1112,8 +1165,9 @@ static s32 e1000_cfg_kmrn_10_100_80003es2lan(struct e1000_hw *hw, u16 duplex)
 static s32 e1000_cfg_kmrn_1000_80003es2lan(struct e1000_hw *hw)
 {
 	s32 ret_val;
-	u16 reg_data;
+	u16 reg_data, reg_data2;
 	u32 tipg;
+	u32 i = 0;
 
 	reg_data = E1000_KMRNCTRLSTA_HD_CTRL_1000_DEFAULT;
 	ret_val = e1000e_write_kmrn_reg(hw, E1000_KMRNCTRLSTA_OFFSET_HD_CTRL,
@@ -1127,9 +1181,16 @@ static s32 e1000_cfg_kmrn_1000_80003es2lan(struct e1000_hw *hw)
 	tipg |= DEFAULT_TIPG_IPGT_1000_80003ES2LAN;
 	ew32(TIPG, tipg);
 
-	ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &reg_data);
-	if (ret_val)
-		return ret_val;
+	do {
+		ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &reg_data);
+		if (ret_val)
+			return ret_val;
+
+		ret_val = e1e_rphy(hw, GG82563_PHY_KMRN_MODE_CTRL, &reg_data2);
+		if (ret_val)
+			return ret_val;
+		i++;
+	} while ((reg_data != reg_data2) && (i < GG82563_MAX_KMRN_RETRY));
 
 	reg_data &= ~GG82563_KMCR_PASS_FALSE_CARRIER;
 	ret_val = e1e_wphy(hw, GG82563_PHY_KMRN_MODE_CTRL, reg_data);
diff --git a/drivers/net/e1000e/phy.c b/drivers/net/e1000e/phy.c
index 3a4574c..e102332 100644
--- a/drivers/net/e1000e/phy.c
+++ b/drivers/net/e1000e/phy.c
@@ -116,7 +116,7 @@ s32 e1000e_phy_reset_dsp(struct e1000_hw *hw)
 }
 
 /**
- *  e1000_read_phy_reg_mdic - Read MDI control register
+ *  e1000e_read_phy_reg_mdic - Read MDI control register
  *  @hw: pointer to the HW structure
  *  @offset: register offset to be read
  *  @data: pointer to the read data
@@ -124,7 +124,7 @@ s32 e1000e_phy_reset_dsp(struct e1000_hw *hw)
  *  Reads the MDI control register in the PHY at offset and stores the
  *  information read to data.
  **/
-static s32 e1000_read_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 *data)
+s32 e1000e_read_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 *data)
 {
 	struct e1000_phy_info *phy = &hw->phy;
 	u32 i, mdic = 0;
@@ -150,7 +150,7 @@ static s32 e1000_read_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 *data)
 	 * Increasing the time out as testing showed failures with
 	 * the lower time out
 	 */
-	for (i = 0; i < 64; i++) {
+	for (i = 0; i < (E1000_GEN_POLL_TIMEOUT * 3); i++) {
 		udelay(50);
 		mdic = er32(MDIC);
 		if (mdic & E1000_MDIC_READY)
@@ -170,14 +170,14 @@ static s32 e1000_read_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 *data)
 }
 
 /**
- *  e1000_write_phy_reg_mdic - Write MDI control register
+ *  e1000e_write_phy_reg_mdic - Write MDI control register
  *  @hw: pointer to the HW structure
  *  @offset: register offset to write to
  *  @data: data to write to register at offset
  *
  *  Writes data to MDI control register in the PHY at offset.
  **/
-static s32 e1000_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data)
+s32 e1000e_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data)
 {
 	struct e1000_phy_info *phy = &hw->phy;
 	u32 i, mdic = 0;
@@ -199,9 +199,13 @@ static s32 e1000_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data)
 
 	ew32(MDIC, mdic);
 
-	/* Poll the ready bit to see if the MDI read completed */
-	for (i = 0; i < E1000_GEN_POLL_TIMEOUT; i++) {
-		udelay(5);
+	/*
+	 * Poll the ready bit to see if the MDI read completed
+	 * Increasing the time out as testing showed failures with
+	 * the lower time out
+	 */
+	for (i = 0; i < (E1000_GEN_POLL_TIMEOUT * 3); i++) {
+		udelay(50);
 		mdic = er32(MDIC);
 		if (mdic & E1000_MDIC_READY)
 			break;
@@ -210,6 +214,10 @@ static s32 e1000_write_phy_reg_mdic(struct e1000_hw *hw, u32 offset, u16 data)
 		hw_dbg(hw, "MDI Write did not complete\n");
 		return -E1000_ERR_PHY;
 	}
+	if (mdic & E1000_MDIC_ERROR) {
+		hw_dbg(hw, "MDI Error\n");
+		return -E1000_ERR_PHY;
+	}
 
 	return 0;
 }
@@ -232,9 +240,8 @@ s32 e1000e_read_phy_reg_m88(struct e1000_hw *hw, u32 offset, u16 *data)
 	if (ret_val)
 		return ret_val;
 
-	ret_val = e1000_read_phy_reg_mdic(hw,
-					  MAX_PHY_REG_ADDRESS & offset,
-					  data);
+	ret_val = e1000e_read_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset,
+					   data);
 
 	hw->phy.ops.release_phy(hw);
 
@@ -258,9 +265,8 @@ s32 e1000e_write_phy_reg_m88(struct e1000_hw *hw, u32 offset, u16 data)
 	if (ret_val)
 		return ret_val;
 
-	ret_val = e1000_write_phy_reg_mdic(hw,
-					   MAX_PHY_REG_ADDRESS & offset,
-					   data);
+	ret_val = e1000e_write_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset,
+					    data);
 
 	hw->phy.ops.release_phy(hw);
 
@@ -286,18 +292,17 @@ s32 e1000e_read_phy_reg_igp(struct e1000_hw *hw, u32 offset, u16 *data)
 		return ret_val;
 
 	if (offset > MAX_PHY_MULTI_PAGE_REG) {
-		ret_val = e1000_write_phy_reg_mdic(hw,
-						   IGP01E1000_PHY_PAGE_SELECT,
-						   (u16)offset);
+		ret_val = e1000e_write_phy_reg_mdic(hw,
+						    IGP01E1000_PHY_PAGE_SELECT,
+						    (u16)offset);
 		if (ret_val) {
 			hw->phy.ops.release_phy(hw);
 			return ret_val;
 		}
 	}
 
-	ret_val = e1000_read_phy_reg_mdic(hw,
-					  MAX_PHY_REG_ADDRESS & offset,
-					  data);
+	ret_val = e1000e_read_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset,
+					   data);
 
 	hw->phy.ops.release_phy(hw);
 
@@ -322,18 +327,17 @@ s32 e1000e_write_phy_reg_igp(struct e1000_hw *hw, u32 offset, u16 data)
 		return ret_val;
 
 	if (offset > MAX_PHY_MULTI_PAGE_REG) {
-		ret_val = e1000_write_phy_reg_mdic(hw,
-						   IGP01E1000_PHY_PAGE_SELECT,
-						   (u16)offset);
+		ret_val = e1000e_write_phy_reg_mdic(hw,
+						    IGP01E1000_PHY_PAGE_SELECT,
+						    (u16)offset);
 		if (ret_val) {
 			hw->phy.ops.release_phy(hw);
 			return ret_val;
 		}
 	}
 
-	ret_val = e1000_write_phy_reg_mdic(hw,
-					   MAX_PHY_REG_ADDRESS & offset,
-					   data);
+	ret_val = e1000e_write_phy_reg_mdic(hw, MAX_PHY_REG_ADDRESS & offset,
+					    data);
 
 	hw->phy.ops.release_phy(hw);
 
@@ -420,7 +424,9 @@ s32 e1000e_copper_link_setup_m88(struct e1000_hw *hw)
 	if (ret_val)
 		return ret_val;
 
-	phy_data |= M88E1000_PSCR_ASSERT_CRS_ON_TX;
+	/* For newer PHYs this bit is downshift enable */
+	if (phy->type == e1000_phy_m88)
+		phy_data |= M88E1000_PSCR_ASSERT_CRS_ON_TX;
 
 	/*
 	 * Options:
@@ -463,7 +469,7 @@ s32 e1000e_copper_link_setup_m88(struct e1000_hw *hw)
 	if (ret_val)
 		return ret_val;
 
-	if (phy->revision < 4) {
+	if ((phy->type == e1000_phy_m88) && (phy->revision < 4)) {
 		/*
 		 * Force TX_CLK in the Extended PHY Specific Control Register
 		 * to 25MHz clock.
@@ -518,8 +524,11 @@ s32 e1000e_copper_link_setup_igp(struct e1000_hw *hw)
 		return ret_val;
 	}
 
-	/* Wait 15ms for MAC to configure PHY from NVM settings. */
-	msleep(15);
+	/*
+	 * Wait 100ms for MAC to configure PHY from NVM settings, to avoid
+	 * timeout issues when LFS is enabled.
+	 */
+	msleep(100);
 
 	/* disable lplu d0 during driver init */
 	ret_val = e1000_set_d0_lplu_state(hw, 0);
@@ -1152,9 +1161,7 @@ s32 e1000e_set_d3_lplu_state(struct e1000_hw *hw, bool active)
 
 	if (!active) {
 		data &= ~IGP02E1000_PM_D3_LPLU;
-		ret_val = e1e_wphy(hw,
-					     IGP02E1000_PHY_POWER_MGMT,
-					     data);
+		ret_val = e1e_wphy(hw, IGP02E1000_PHY_POWER_MGMT, data);
 		if (ret_val)
 			return ret_val;
 		/*

[PATCH 5/5] e1000e: lower ring minimum size to 64

[Auke Kok]

The lower limit of 80 descriptors in the ring is only valid for
one older 8254x chipset. All e1000e devices can use as low as
64 descriptors.

Signed-off-by: Auke Kok <auke-jan.h.kok@intel.com>
---

 drivers/net/e1000e/e1000.h |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/e1000e/e1000.h b/drivers/net/e1000e/e1000.h
index f6835c3..38bfd0d 100644
--- a/drivers/net/e1000e/e1000.h
+++ b/drivers/net/e1000e/e1000.h
@@ -64,11 +64,11 @@ struct e1000_info;
 /* Tx/Rx descriptor defines */
 #define E1000_DEFAULT_TXD		256
 #define E1000_MAX_TXD			4096
-#define E1000_MIN_TXD			80
+#define E1000_MIN_TXD			64
 
 #define E1000_DEFAULT_RXD		256
 #define E1000_MAX_RXD			4096
-#define E1000_MIN_RXD			80
+#define E1000_MIN_RXD			64
 
 #define E1000_MIN_ITR_USECS		10 /* 100000 irq/sec */
 #define E1000_MAX_ITR_USECS		10000 /* 100    irq/sec */