[net 00/13][pull request] Intel Wired LAN Driver Updates

[net 00/13][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series contains updates to e1000, e1000e, igb, igbvf, ixgb, ixgbe,
ixgbevf and i40evf.

Mark fixes an issue with ixgbe and ixgbevf by adding a bit to indicate
when workqueues have been initialized.  This permits the register read
error handling from attempting to use them prior to that, which also
generates warnings.  Checking for a detected removal after initializing
the work queues allows the probe function to return an error without
getting the workqueue involved.  Further, if the error_detected
callback is entered before the workqueues are initialized, exit without
recovery since the device initialization was so truncated.

Francois Romieu provides several patches to all the drivers to remove
the open coded skb_cow_head.

Jakub Kicinski provides a fix for igb where last_rx_timestamp should be
updated only when Rx time stamp is read.

Mitch provides a fix for i40evf where a recent change broke the RSS LUT
programming causing it to be programmed with all 0's.

The following are changes since commit 6d39d589bb76ee8a1c6cde6822006ae0053decff:
  net: core: don't account for udp header size when computing seglen
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net master

Francois Romieu (8):
  ixgbe: remove open-coded skb_cow_head
  ixgbevf: remove open-coded skb_cow_head
  e1000: remove open-coded skb_cow_head
  e1000e: remove open-coded skb_cow_head
  igb: remove open-coded skb_cow_head
  igbvf: remove open-coded skb_cow_head
  ixgb: remove open-coded skb_cow_head
  i40evf: remove open-coded skb_cow_head

Jakub Kicinski (2):
  e1000: remove debug messages with function names
  igb: fix last_rx_timestamp usage

Mark Rustad (2):
  ixgbe: Add bit to mark service task initialization
  ixgbevf: Add bit to mark work queue initialization

Mitch A Williams (1):
  i40evf: program RSS LUT correctly

 drivers/net/ethernet/intel/e1000/e1000_hw.c       | 124 ----------------------
 drivers/net/ethernet/intel/e1000/e1000_main.c     |  11 +-
 drivers/net/ethernet/intel/e1000e/netdev.c        |  10 +-
 drivers/net/ethernet/intel/i40evf/i40e_txrx.c     |  10 +-
 drivers/net/ethernet/intel/i40evf/i40evf_main.c   |  20 +++-
 drivers/net/ethernet/intel/igb/igb.h              |  16 +--
 drivers/net/ethernet/intel/igb/igb_main.c         |  13 +--
 drivers/net/ethernet/intel/igb/igb_ptp.c          |  14 +--
 drivers/net/ethernet/intel/igbvf/netdev.c         |  16 ++-
 drivers/net/ethernet/intel/ixgb/ixgb_main.c       |  10 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe.h          |   1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c     |  28 +++--
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |   1 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c |  20 ++--
 14 files changed, 91 insertions(+), 203 deletions(-)

-- 
1.9.0

[net 01/13] ixgbe: Add bit to mark service task initialization

[Jeff Kirsher]

From: Mark Rustad <mark.d.rustad@intel.com>

There needs to be an indication when the service task has been
initialized. This is because register access prior to that time
can detect a removal and attempt to schedule the service task.
Adding the __IXGBE_SERVICE_INITED bit allows this to be checked
and if not set prevent the service task scheduling. By checking
for a removal right after initialization, the probe can be failed
at that point without getting the service task involved.

Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe.h      |  1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 15 ++++++++++++++-
 2 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index 55c53a1..1a12c1d 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -811,6 +811,7 @@ enum ixgbe_state_t {
 	__IXGBE_DISABLED,
 	__IXGBE_REMOVING,
 	__IXGBE_SERVICE_SCHED,
+	__IXGBE_SERVICE_INITED,
 	__IXGBE_IN_SFP_INIT,
 	__IXGBE_PTP_RUNNING,
 	__IXGBE_PTP_TX_IN_PROGRESS,
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 8436c65..00c9308 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -297,7 +297,8 @@ static void ixgbe_remove_adapter(struct ixgbe_hw *hw)
 		return;
 	hw->hw_addr = NULL;
 	e_dev_err("Adapter removed\n");
-	ixgbe_service_event_schedule(adapter);
+	if (test_bit(__IXGBE_SERVICE_INITED, &adapter->state))
+		ixgbe_service_event_schedule(adapter);
 }
 
 void ixgbe_check_remove(struct ixgbe_hw *hw, u32 reg)
@@ -8023,6 +8024,10 @@ static int ixgbe_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	/* EEPROM */
 	memcpy(&hw->eeprom.ops, ii->eeprom_ops, sizeof(hw->eeprom.ops));
 	eec = IXGBE_READ_REG(hw, IXGBE_EEC);
+	if (ixgbe_removed(hw->hw_addr)) {
+		err = -EIO;
+		goto err_ioremap;
+	}
 	/* If EEPROM is valid (bit 8 = 1), use default otherwise use bit bang */
 	if (!(eec & (1 << 8)))
 		hw->eeprom.ops.read = &ixgbe_read_eeprom_bit_bang_generic;
@@ -8185,7 +8190,12 @@ skip_sriov:
 	setup_timer(&adapter->service_timer, &ixgbe_service_timer,
 		    (unsigned long) adapter);
 
+	if (ixgbe_removed(hw->hw_addr)) {
+		err = -EIO;
+		goto err_sw_init;
+	}
 	INIT_WORK(&adapter->service_task, ixgbe_service_task);
+	set_bit(__IXGBE_SERVICE_INITED, &adapter->state);
 	clear_bit(__IXGBE_SERVICE_SCHED, &adapter->state);
 
 	err = ixgbe_init_interrupt_scheme(adapter);
@@ -8494,6 +8504,9 @@ static pci_ers_result_t ixgbe_io_error_detected(struct pci_dev *pdev,
 
 skip_bad_vf_detection:
 #endif /* CONFIG_PCI_IOV */
+	if (!test_bit(__IXGBE_SERVICE_INITED, &adapter->state))
+		return PCI_ERS_RESULT_DISCONNECT;
+
 	rtnl_lock();
 	netif_device_detach(netdev);
 
-- 
1.9.0

[net 02/13] ixgbe: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 00c9308..c4c526b 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -6510,6 +6510,7 @@ static int ixgbe_tso(struct ixgbe_ring *tx_ring,
 	struct sk_buff *skb = first->skb;
 	u32 vlan_macip_lens, type_tucmd;
 	u32 mss_l4len_idx, l4len;
+	int err;
 
 	if (skb->ip_summed != CHECKSUM_PARTIAL)
 		return 0;
@@ -6517,11 +6518,9 @@ static int ixgbe_tso(struct ixgbe_ring *tx_ring,
 	if (!skb_is_gso(skb))
 		return 0;
 
-	if (skb_header_cloned(skb)) {
-		int err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-		if (err)
-			return err;
-	}
+	err = skb_cow_head(skb, 0);
+	if (err < 0)
+		return err;
 
 	/* ADV DTYP TUCMD MKRLOC/ISCSIHEDLEN */
 	type_tucmd = IXGBE_ADVTXD_TUCMD_L4T_TCP;
@@ -7078,8 +7077,8 @@ netdev_tx_t ixgbe_xmit_frame_ring(struct sk_buff *skb,
 					IXGBE_TX_FLAGS_VLAN_PRIO_SHIFT;
 		if (tx_flags & IXGBE_TX_FLAGS_SW_VLAN) {
 			struct vlan_ethhdr *vhdr;
-			if (skb_header_cloned(skb) &&
-			    pskb_expand_head(skb, 0, 0, GFP_ATOMIC))
+
+			if (skb_cow_head(skb, 0))
 				goto out_drop;
 			vhdr = (struct vlan_ethhdr *)skb->data;
 			vhdr->h_vlan_TCI = htons(tx_flags >>
-- 
1.9.0

[net 03/13] ixgbevf: Add bit to mark work queue initialization

[Jeff Kirsher]

From: Mark Rustad <mark.d.rustad@intel.com>

An indication of work queue initialization is needed. This is
because register accesses prior to that time can detect a removal
and attempt to schedule the watchdog task. Adding the
__IXGBEVF_WORK_INIT bit allows this to be checked and if not
set prevent the watchdog task scheduling. By checking for a
removal right after initialization, the probe can be failed
at that point without getting the watchdog task involved.

Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |  1 +
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 11 ++++++++++-
 2 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
index e7e7d69..a0a1de9 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf.h
@@ -421,6 +421,7 @@ enum ixbgevf_state_t {
 	__IXGBEVF_DOWN,
 	__IXGBEVF_DISABLED,
 	__IXGBEVF_REMOVING,
+	__IXGBEVF_WORK_INIT,
 };
 
 struct ixgbevf_cb {
diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 4ba139b..905c26c 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -107,7 +107,8 @@ static void ixgbevf_remove_adapter(struct ixgbe_hw *hw)
 		return;
 	hw->hw_addr = NULL;
 	dev_err(&adapter->pdev->dev, "Adapter removed\n");
-	schedule_work(&adapter->watchdog_task);
+	if (test_bit(__IXGBEVF_WORK_INIT, &adapter->state))
+		schedule_work(&adapter->watchdog_task);
 }
 
 static void ixgbevf_check_remove(struct ixgbe_hw *hw, u32 reg)
@@ -3573,8 +3574,13 @@ static int ixgbevf_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	adapter->watchdog_timer.function = ixgbevf_watchdog;
 	adapter->watchdog_timer.data = (unsigned long)adapter;
 
+	if (IXGBE_REMOVED(hw->hw_addr)) {
+		err = -EIO;
+		goto err_sw_init;
+	}
 	INIT_WORK(&adapter->reset_task, ixgbevf_reset_task);
 	INIT_WORK(&adapter->watchdog_task, ixgbevf_watchdog_task);
+	set_bit(__IXGBEVF_WORK_INIT, &adapter->state);
 
 	err = ixgbevf_init_interrupt_scheme(adapter);
 	if (err)
@@ -3667,6 +3673,9 @@ static pci_ers_result_t ixgbevf_io_error_detected(struct pci_dev *pdev,
 	struct net_device *netdev = pci_get_drvdata(pdev);
 	struct ixgbevf_adapter *adapter = netdev_priv(netdev);
 
+	if (!test_bit(__IXGBEVF_WORK_INIT, &adapter->state))
+		return PCI_ERS_RESULT_DISCONNECT;
+
 	rtnl_lock();
 	netif_device_detach(netdev);
 
-- 
1.9.0

[net 04/13] ixgbevf: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
index 905c26c..d0799e8 100644
--- a/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
+++ b/drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c
@@ -2839,6 +2839,7 @@ static int ixgbevf_tso(struct ixgbevf_ring *tx_ring,
 	struct sk_buff *skb = first->skb;
 	u32 vlan_macip_lens, type_tucmd;
 	u32 mss_l4len_idx, l4len;
+	int err;
 
 	if (skb->ip_summed != CHECKSUM_PARTIAL)
 		return 0;
@@ -2846,11 +2847,9 @@ static int ixgbevf_tso(struct ixgbevf_ring *tx_ring,
 	if (!skb_is_gso(skb))
 		return 0;
 
-	if (skb_header_cloned(skb)) {
-		int err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-		if (err)
-			return err;
-	}
+	err = skb_cow_head(skb, 0);
+	if (err < 0)
+		return err;
 
 	/* ADV DTYP TUCMD MKRLOC/ISCSIHEDLEN */
 	type_tucmd = IXGBE_ADVTXD_TUCMD_L4T_TCP;
-- 
1.9.0

[net 05/13] e1000: remove debug messages with function names

[Jeff Kirsher]

From: Jakub Kicinski <kubakici@wp.pl>

e1000_hw.c contains a lot of debug messages which print
name of invoked function and contain no new line character
at the end.  Remove them as equivalent information can be
nowadays obtained using function tracer.

Reported-by: Joe Perches <joe@perches.com>
Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/e1000/e1000_hw.c | 124 ----------------------------
 1 file changed, 124 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000/e1000_hw.c b/drivers/net/ethernet/intel/e1000/e1000_hw.c
index 2879b96..c1d3fdb 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_hw.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_hw.c
@@ -115,8 +115,6 @@ static DEFINE_SPINLOCK(e1000_phy_lock);
  */
 static s32 e1000_set_phy_type(struct e1000_hw *hw)
 {
-	e_dbg("e1000_set_phy_type");
-
 	if (hw->mac_type == e1000_undefined)
 		return -E1000_ERR_PHY_TYPE;
 
@@ -159,8 +157,6 @@ static void e1000_phy_init_script(struct e1000_hw *hw)
 	u32 ret_val;
 	u16 phy_saved_data;
 
-	e_dbg("e1000_phy_init_script");
-
 	if (hw->phy_init_script) {
 		msleep(20);
 
@@ -253,8 +249,6 @@ static void e1000_phy_init_script(struct e1000_hw *hw)
  */
 s32 e1000_set_mac_type(struct e1000_hw *hw)
 {
-	e_dbg("e1000_set_mac_type");
-
 	switch (hw->device_id) {
 	case E1000_DEV_ID_82542:
 		switch (hw->revision_id) {
@@ -365,8 +359,6 @@ void e1000_set_media_type(struct e1000_hw *hw)
 {
 	u32 status;
 
-	e_dbg("e1000_set_media_type");
-
 	if (hw->mac_type != e1000_82543) {
 		/* tbi_compatibility is only valid on 82543 */
 		hw->tbi_compatibility_en = false;
@@ -415,8 +407,6 @@ s32 e1000_reset_hw(struct e1000_hw *hw)
 	u32 led_ctrl;
 	s32 ret_val;
 
-	e_dbg("e1000_reset_hw");
-
 	/* For 82542 (rev 2.0), disable MWI before issuing a device reset */
 	if (hw->mac_type == e1000_82542_rev2_0) {
 		e_dbg("Disabling MWI on 82542 rev 2.0\n");
@@ -566,8 +556,6 @@ s32 e1000_init_hw(struct e1000_hw *hw)
 	u32 mta_size;
 	u32 ctrl_ext;
 
-	e_dbg("e1000_init_hw");
-
 	/* Initialize Identification LED */
 	ret_val = e1000_id_led_init(hw);
 	if (ret_val) {
@@ -683,8 +671,6 @@ static s32 e1000_adjust_serdes_amplitude(struct e1000_hw *hw)
 	u16 eeprom_data;
 	s32 ret_val;
 
-	e_dbg("e1000_adjust_serdes_amplitude");
-
 	if (hw->media_type != e1000_media_type_internal_serdes)
 		return E1000_SUCCESS;
 
@@ -730,8 +716,6 @@ s32 e1000_setup_link(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 eeprom_data;
 
-	e_dbg("e1000_setup_link");
-
 	/* Read and store word 0x0F of the EEPROM. This word contains bits
 	 * that determine the hardware's default PAUSE (flow control) mode,
 	 * a bit that determines whether the HW defaults to enabling or
@@ -848,8 +832,6 @@ static s32 e1000_setup_fiber_serdes_link(struct e1000_hw *hw)
 	u32 signal = 0;
 	s32 ret_val;
 
-	e_dbg("e1000_setup_fiber_serdes_link");
-
 	/* On adapters with a MAC newer than 82544, SWDP 1 will be
 	 * set when the optics detect a signal. On older adapters, it will be
 	 * cleared when there is a signal.  This applies to fiber media only.
@@ -1051,8 +1033,6 @@ static s32 e1000_copper_link_preconfig(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_copper_link_preconfig");
-
 	ctrl = er32(CTRL);
 	/* With 82543, we need to force speed and duplex on the MAC equal to
 	 * what the PHY speed and duplex configuration is. In addition, we need
@@ -1112,8 +1092,6 @@ static s32 e1000_copper_link_igp_setup(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_copper_link_igp_setup");
-
 	if (hw->phy_reset_disable)
 		return E1000_SUCCESS;
 
@@ -1254,8 +1232,6 @@ static s32 e1000_copper_link_mgp_setup(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_copper_link_mgp_setup");
-
 	if (hw->phy_reset_disable)
 		return E1000_SUCCESS;
 
@@ -1362,8 +1338,6 @@ static s32 e1000_copper_link_autoneg(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_copper_link_autoneg");
-
 	/* Perform some bounds checking on the hw->autoneg_advertised
 	 * parameter.  If this variable is zero, then set it to the default.
 	 */
@@ -1432,7 +1406,6 @@ static s32 e1000_copper_link_autoneg(struct e1000_hw *hw)
 static s32 e1000_copper_link_postconfig(struct e1000_hw *hw)
 {
 	s32 ret_val;
-	e_dbg("e1000_copper_link_postconfig");
 
 	if ((hw->mac_type >= e1000_82544) && (hw->mac_type != e1000_ce4100)) {
 		e1000_config_collision_dist(hw);
@@ -1473,8 +1446,6 @@ static s32 e1000_setup_copper_link(struct e1000_hw *hw)
 	u16 i;
 	u16 phy_data;
 
-	e_dbg("e1000_setup_copper_link");
-
 	/* Check if it is a valid PHY and set PHY mode if necessary. */
 	ret_val = e1000_copper_link_preconfig(hw);
 	if (ret_val)
@@ -1554,8 +1525,6 @@ s32 e1000_phy_setup_autoneg(struct e1000_hw *hw)
 	u16 mii_autoneg_adv_reg;
 	u16 mii_1000t_ctrl_reg;
 
-	e_dbg("e1000_phy_setup_autoneg");
-
 	/* Read the MII Auto-Neg Advertisement Register (Address 4). */
 	ret_val = e1000_read_phy_reg(hw, PHY_AUTONEG_ADV, &mii_autoneg_adv_reg);
 	if (ret_val)
@@ -1707,8 +1676,6 @@ static s32 e1000_phy_force_speed_duplex(struct e1000_hw *hw)
 	u16 phy_data;
 	u16 i;
 
-	e_dbg("e1000_phy_force_speed_duplex");
-
 	/* Turn off Flow control if we are forcing speed and duplex. */
 	hw->fc = E1000_FC_NONE;
 
@@ -1939,8 +1906,6 @@ void e1000_config_collision_dist(struct e1000_hw *hw)
 {
 	u32 tctl, coll_dist;
 
-	e_dbg("e1000_config_collision_dist");
-
 	if (hw->mac_type < e1000_82543)
 		coll_dist = E1000_COLLISION_DISTANCE_82542;
 	else
@@ -1970,8 +1935,6 @@ static s32 e1000_config_mac_to_phy(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_config_mac_to_phy");
-
 	/* 82544 or newer MAC, Auto Speed Detection takes care of
 	 * MAC speed/duplex configuration.
 	 */
@@ -2049,8 +2012,6 @@ s32 e1000_force_mac_fc(struct e1000_hw *hw)
 {
 	u32 ctrl;
 
-	e_dbg("e1000_force_mac_fc");
-
 	/* Get the current configuration of the Device Control Register */
 	ctrl = er32(CTRL);
 
@@ -2120,8 +2081,6 @@ static s32 e1000_config_fc_after_link_up(struct e1000_hw *hw)
 	u16 speed;
 	u16 duplex;
 
-	e_dbg("e1000_config_fc_after_link_up");
-
 	/* Check for the case where we have fiber media and auto-neg failed
 	 * so we had to force link.  In this case, we need to force the
 	 * configuration of the MAC to match the "fc" parameter.
@@ -2337,8 +2296,6 @@ static s32 e1000_check_for_serdes_link_generic(struct e1000_hw *hw)
 	u32 status;
 	s32 ret_val = E1000_SUCCESS;
 
-	e_dbg("e1000_check_for_serdes_link_generic");
-
 	ctrl = er32(CTRL);
 	status = er32(STATUS);
 	rxcw = er32(RXCW);
@@ -2449,8 +2406,6 @@ s32 e1000_check_for_link(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_check_for_link");
-
 	ctrl = er32(CTRL);
 	status = er32(STATUS);
 
@@ -2632,8 +2587,6 @@ s32 e1000_get_speed_and_duplex(struct e1000_hw *hw, u16 *speed, u16 *duplex)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_get_speed_and_duplex");
-
 	if (hw->mac_type >= e1000_82543) {
 		status = er32(STATUS);
 		if (status & E1000_STATUS_SPEED_1000) {
@@ -2699,7 +2652,6 @@ static s32 e1000_wait_autoneg(struct e1000_hw *hw)
 	u16 i;
 	u16 phy_data;
 
-	e_dbg("e1000_wait_autoneg");
 	e_dbg("Waiting for Auto-Neg to complete.\n");
 
 	/* We will wait for autoneg to complete or 4.5 seconds to expire. */
@@ -2866,8 +2818,6 @@ s32 e1000_read_phy_reg(struct e1000_hw *hw, u32 reg_addr, u16 *phy_data)
 	u32 ret_val;
 	unsigned long flags;
 
-	e_dbg("e1000_read_phy_reg");
-
 	spin_lock_irqsave(&e1000_phy_lock, flags);
 
 	if ((hw->phy_type == e1000_phy_igp) &&
@@ -2894,8 +2844,6 @@ static s32 e1000_read_phy_reg_ex(struct e1000_hw *hw, u32 reg_addr,
 	u32 mdic = 0;
 	const u32 phy_addr = (hw->mac_type == e1000_ce4100) ? hw->phy_addr : 1;
 
-	e_dbg("e1000_read_phy_reg_ex");
-
 	if (reg_addr > MAX_PHY_REG_ADDRESS) {
 		e_dbg("PHY Address %d is out of range\n", reg_addr);
 		return -E1000_ERR_PARAM;
@@ -3008,8 +2956,6 @@ s32 e1000_write_phy_reg(struct e1000_hw *hw, u32 reg_addr, u16 phy_data)
 	u32 ret_val;
 	unsigned long flags;
 
-	e_dbg("e1000_write_phy_reg");
-
 	spin_lock_irqsave(&e1000_phy_lock, flags);
 
 	if ((hw->phy_type == e1000_phy_igp) &&
@@ -3036,8 +2982,6 @@ static s32 e1000_write_phy_reg_ex(struct e1000_hw *hw, u32 reg_addr,
 	u32 mdic = 0;
 	const u32 phy_addr = (hw->mac_type == e1000_ce4100) ? hw->phy_addr : 1;
 
-	e_dbg("e1000_write_phy_reg_ex");
-
 	if (reg_addr > MAX_PHY_REG_ADDRESS) {
 		e_dbg("PHY Address %d is out of range\n", reg_addr);
 		return -E1000_ERR_PARAM;
@@ -3129,8 +3073,6 @@ s32 e1000_phy_hw_reset(struct e1000_hw *hw)
 	u32 ctrl, ctrl_ext;
 	u32 led_ctrl;
 
-	e_dbg("e1000_phy_hw_reset");
-
 	e_dbg("Resetting Phy...\n");
 
 	if (hw->mac_type > e1000_82543) {
@@ -3189,8 +3131,6 @@ s32 e1000_phy_reset(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_phy_reset");
-
 	switch (hw->phy_type) {
 	case e1000_phy_igp:
 		ret_val = e1000_phy_hw_reset(hw);
@@ -3229,8 +3169,6 @@ static s32 e1000_detect_gig_phy(struct e1000_hw *hw)
 	u16 phy_id_high, phy_id_low;
 	bool match = false;
 
-	e_dbg("e1000_detect_gig_phy");
-
 	if (hw->phy_id != 0)
 		return E1000_SUCCESS;
 
@@ -3301,7 +3239,6 @@ static s32 e1000_detect_gig_phy(struct e1000_hw *hw)
 static s32 e1000_phy_reset_dsp(struct e1000_hw *hw)
 {
 	s32 ret_val;
-	e_dbg("e1000_phy_reset_dsp");
 
 	do {
 		ret_val = e1000_write_phy_reg(hw, 29, 0x001d);
@@ -3333,8 +3270,6 @@ static s32 e1000_phy_igp_get_info(struct e1000_hw *hw,
 	u16 phy_data, min_length, max_length, average;
 	e1000_rev_polarity polarity;
 
-	e_dbg("e1000_phy_igp_get_info");
-
 	/* The downshift status is checked only once, after link is established,
 	 * and it stored in the hw->speed_downgraded parameter.
 	 */
@@ -3414,8 +3349,6 @@ static s32 e1000_phy_m88_get_info(struct e1000_hw *hw,
 	u16 phy_data;
 	e1000_rev_polarity polarity;
 
-	e_dbg("e1000_phy_m88_get_info");
-
 	/* The downshift status is checked only once, after link is established,
 	 * and it stored in the hw->speed_downgraded parameter.
 	 */
@@ -3487,8 +3420,6 @@ s32 e1000_phy_get_info(struct e1000_hw *hw, struct e1000_phy_info *phy_info)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_phy_get_info");
-
 	phy_info->cable_length = e1000_cable_length_undefined;
 	phy_info->extended_10bt_distance = e1000_10bt_ext_dist_enable_undefined;
 	phy_info->cable_polarity = e1000_rev_polarity_undefined;
@@ -3527,8 +3458,6 @@ s32 e1000_phy_get_info(struct e1000_hw *hw, struct e1000_phy_info *phy_info)
 
 s32 e1000_validate_mdi_setting(struct e1000_hw *hw)
 {
-	e_dbg("e1000_validate_mdi_settings");
-
 	if (!hw->autoneg && (hw->mdix == 0 || hw->mdix == 3)) {
 		e_dbg("Invalid MDI setting detected\n");
 		hw->mdix = 1;
@@ -3551,8 +3480,6 @@ s32 e1000_init_eeprom_params(struct e1000_hw *hw)
 	s32 ret_val = E1000_SUCCESS;
 	u16 eeprom_size;
 
-	e_dbg("e1000_init_eeprom_params");
-
 	switch (hw->mac_type) {
 	case e1000_82542_rev2_0:
 	case e1000_82542_rev2_1:
@@ -3770,8 +3697,6 @@ static s32 e1000_acquire_eeprom(struct e1000_hw *hw)
 	struct e1000_eeprom_info *eeprom = &hw->eeprom;
 	u32 eecd, i = 0;
 
-	e_dbg("e1000_acquire_eeprom");
-
 	eecd = er32(EECD);
 
 	/* Request EEPROM Access */
@@ -3871,8 +3796,6 @@ static void e1000_release_eeprom(struct e1000_hw *hw)
 {
 	u32 eecd;
 
-	e_dbg("e1000_release_eeprom");
-
 	eecd = er32(EECD);
 
 	if (hw->eeprom.type == e1000_eeprom_spi) {
@@ -3920,8 +3843,6 @@ static s32 e1000_spi_eeprom_ready(struct e1000_hw *hw)
 	u16 retry_count = 0;
 	u8 spi_stat_reg;
 
-	e_dbg("e1000_spi_eeprom_ready");
-
 	/* Read "Status Register" repeatedly until the LSB is cleared.  The
 	 * EEPROM will signal that the command has been completed by clearing
 	 * bit 0 of the internal status register.  If it's not cleared within
@@ -3974,8 +3895,6 @@ static s32 e1000_do_read_eeprom(struct e1000_hw *hw, u16 offset, u16 words,
 	struct e1000_eeprom_info *eeprom = &hw->eeprom;
 	u32 i = 0;
 
-	e_dbg("e1000_read_eeprom");
-
 	if (hw->mac_type == e1000_ce4100) {
 		GBE_CONFIG_FLASH_READ(GBE_CONFIG_BASE_VIRT, offset, words,
 		                      data);
@@ -4076,8 +3995,6 @@ s32 e1000_validate_eeprom_checksum(struct e1000_hw *hw)
 	u16 checksum = 0;
 	u16 i, eeprom_data;
 
-	e_dbg("e1000_validate_eeprom_checksum");
-
 	for (i = 0; i < (EEPROM_CHECKSUM_REG + 1); i++) {
 		if (e1000_read_eeprom(hw, i, 1, &eeprom_data) < 0) {
 			e_dbg("EEPROM Read Error\n");
@@ -4112,8 +4029,6 @@ s32 e1000_update_eeprom_checksum(struct e1000_hw *hw)
 	u16 checksum = 0;
 	u16 i, eeprom_data;
 
-	e_dbg("e1000_update_eeprom_checksum");
-
 	for (i = 0; i < EEPROM_CHECKSUM_REG; i++) {
 		if (e1000_read_eeprom(hw, i, 1, &eeprom_data) < 0) {
 			e_dbg("EEPROM Read Error\n");
@@ -4154,8 +4069,6 @@ static s32 e1000_do_write_eeprom(struct e1000_hw *hw, u16 offset, u16 words,
 	struct e1000_eeprom_info *eeprom = &hw->eeprom;
 	s32 status = 0;
 
-	e_dbg("e1000_write_eeprom");
-
 	if (hw->mac_type == e1000_ce4100) {
 		GBE_CONFIG_FLASH_WRITE(GBE_CONFIG_BASE_VIRT, offset, words,
 		                       data);
@@ -4205,8 +4118,6 @@ static s32 e1000_write_eeprom_spi(struct e1000_hw *hw, u16 offset, u16 words,
 	struct e1000_eeprom_info *eeprom = &hw->eeprom;
 	u16 widx = 0;
 
-	e_dbg("e1000_write_eeprom_spi");
-
 	while (widx < words) {
 		u8 write_opcode = EEPROM_WRITE_OPCODE_SPI;
 
@@ -4274,8 +4185,6 @@ static s32 e1000_write_eeprom_microwire(struct e1000_hw *hw, u16 offset,
 	u16 words_written = 0;
 	u16 i = 0;
 
-	e_dbg("e1000_write_eeprom_microwire");
-
 	/* Send the write enable command to the EEPROM (3-bit opcode plus
 	 * 6/8-bit dummy address beginning with 11).  It's less work to include
 	 * the 11 of the dummy address as part of the opcode than it is to shift
@@ -4354,8 +4263,6 @@ s32 e1000_read_mac_addr(struct e1000_hw *hw)
 	u16 offset;
 	u16 eeprom_data, i;
 
-	e_dbg("e1000_read_mac_addr");
-
 	for (i = 0; i < NODE_ADDRESS_SIZE; i += 2) {
 		offset = i >> 1;
 		if (e1000_read_eeprom(hw, offset, 1, &eeprom_data) < 0) {
@@ -4394,8 +4301,6 @@ static void e1000_init_rx_addrs(struct e1000_hw *hw)
 	u32 i;
 	u32 rar_num;
 
-	e_dbg("e1000_init_rx_addrs");
-
 	/* Setup the receive address. */
 	e_dbg("Programming MAC Address into RAR[0]\n");
 
@@ -4553,8 +4458,6 @@ static s32 e1000_id_led_init(struct e1000_hw *hw)
 	u16 eeprom_data, i, temp;
 	const u16 led_mask = 0x0F;
 
-	e_dbg("e1000_id_led_init");
-
 	if (hw->mac_type < e1000_82540) {
 		/* Nothing to do */
 		return E1000_SUCCESS;
@@ -4626,8 +4529,6 @@ s32 e1000_setup_led(struct e1000_hw *hw)
 	u32 ledctl;
 	s32 ret_val = E1000_SUCCESS;
 
-	e_dbg("e1000_setup_led");
-
 	switch (hw->mac_type) {
 	case e1000_82542_rev2_0:
 	case e1000_82542_rev2_1:
@@ -4678,8 +4579,6 @@ s32 e1000_cleanup_led(struct e1000_hw *hw)
 {
 	s32 ret_val = E1000_SUCCESS;
 
-	e_dbg("e1000_cleanup_led");
-
 	switch (hw->mac_type) {
 	case e1000_82542_rev2_0:
 	case e1000_82542_rev2_1:
@@ -4714,8 +4613,6 @@ s32 e1000_led_on(struct e1000_hw *hw)
 {
 	u32 ctrl = er32(CTRL);
 
-	e_dbg("e1000_led_on");
-
 	switch (hw->mac_type) {
 	case e1000_82542_rev2_0:
 	case e1000_82542_rev2_1:
@@ -4760,8 +4657,6 @@ s32 e1000_led_off(struct e1000_hw *hw)
 {
 	u32 ctrl = er32(CTRL);
 
-	e_dbg("e1000_led_off");
-
 	switch (hw->mac_type) {
 	case e1000_82542_rev2_0:
 	case e1000_82542_rev2_1:
@@ -4889,8 +4784,6 @@ static void e1000_clear_hw_cntrs(struct e1000_hw *hw)
  */
 void e1000_reset_adaptive(struct e1000_hw *hw)
 {
-	e_dbg("e1000_reset_adaptive");
-
 	if (hw->adaptive_ifs) {
 		if (!hw->ifs_params_forced) {
 			hw->current_ifs_val = 0;
@@ -4917,8 +4810,6 @@ void e1000_reset_adaptive(struct e1000_hw *hw)
  */
 void e1000_update_adaptive(struct e1000_hw *hw)
 {
-	e_dbg("e1000_update_adaptive");
-
 	if (hw->adaptive_ifs) {
 		if ((hw->collision_delta *hw->ifs_ratio) > hw->tx_packet_delta) {
 			if (hw->tx_packet_delta > MIN_NUM_XMITS) {
@@ -5114,8 +5005,6 @@ static s32 e1000_get_cable_length(struct e1000_hw *hw, u16 *min_length,
 	u16 i, phy_data;
 	u16 cable_length;
 
-	e_dbg("e1000_get_cable_length");
-
 	*min_length = *max_length = 0;
 
 	/* Use old method for Phy older than IGP */
@@ -5231,8 +5120,6 @@ static s32 e1000_check_polarity(struct e1000_hw *hw,
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_check_polarity");
-
 	if (hw->phy_type == e1000_phy_m88) {
 		/* return the Polarity bit in the Status register. */
 		ret_val = e1000_read_phy_reg(hw, M88E1000_PHY_SPEC_STATUS,
@@ -5299,8 +5186,6 @@ static s32 e1000_check_downshift(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 phy_data;
 
-	e_dbg("e1000_check_downshift");
-
 	if (hw->phy_type == e1000_phy_igp) {
 		ret_val = e1000_read_phy_reg(hw, IGP01E1000_PHY_LINK_HEALTH,
 					     &phy_data);
@@ -5411,8 +5296,6 @@ static s32 e1000_config_dsp_after_link_change(struct e1000_hw *hw, bool link_up)
 	s32 ret_val;
 	u16 phy_data, phy_saved_data, speed, duplex, i;
 
-	e_dbg("e1000_config_dsp_after_link_change");
-
 	if (hw->phy_type != e1000_phy_igp)
 		return E1000_SUCCESS;
 
@@ -5546,8 +5429,6 @@ static s32 e1000_set_phy_mode(struct e1000_hw *hw)
 	s32 ret_val;
 	u16 eeprom_data;
 
-	e_dbg("e1000_set_phy_mode");
-
 	if ((hw->mac_type == e1000_82545_rev_3) &&
 	    (hw->media_type == e1000_media_type_copper)) {
 		ret_val =
@@ -5594,7 +5475,6 @@ static s32 e1000_set_d3_lplu_state(struct e1000_hw *hw, bool active)
 {
 	s32 ret_val;
 	u16 phy_data;
-	e_dbg("e1000_set_d3_lplu_state");
 
 	if (hw->phy_type != e1000_phy_igp)
 		return E1000_SUCCESS;
@@ -5699,8 +5579,6 @@ static s32 e1000_set_vco_speed(struct e1000_hw *hw)
 	u16 default_page = 0;
 	u16 phy_data;
 
-	e_dbg("e1000_set_vco_speed");
-
 	switch (hw->mac_type) {
 	case e1000_82545_rev_3:
 	case e1000_82546_rev_3:
@@ -5872,7 +5750,6 @@ static s32 e1000_polarity_reversal_workaround(struct e1000_hw *hw)
  */
 static s32 e1000_get_auto_rd_done(struct e1000_hw *hw)
 {
-	e_dbg("e1000_get_auto_rd_done");
 	msleep(5);
 	return E1000_SUCCESS;
 }
@@ -5887,7 +5764,6 @@ static s32 e1000_get_auto_rd_done(struct e1000_hw *hw)
  */
 static s32 e1000_get_phy_cfg_done(struct e1000_hw *hw)
 {
-	e_dbg("e1000_get_phy_cfg_done");
 	msleep(10);
 	return E1000_SUCCESS;
 }
-- 
1.9.0

[net 06/13] e1000: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/e1000/e1000_main.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
index 46e6544..27058df 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
@@ -2682,14 +2682,13 @@ static int e1000_tso(struct e1000_adapter *adapter,
 	u32 cmd_length = 0;
 	u16 ipcse = 0, tucse, mss;
 	u8 ipcss, ipcso, tucss, tucso, hdr_len;
-	int err;
 
 	if (skb_is_gso(skb)) {
-		if (skb_header_cloned(skb)) {
-			err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-			if (err)
-				return err;
-		}
+		int err;
+
+		err = skb_cow_head(skb, 0);
+		if (err < 0)
+			return err;
 
 		hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
 		mss = skb_shinfo(skb)->gso_size;
-- 
1.9.0

[net 07/13] e1000e: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/e1000e/netdev.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index dce377b..d50c91e 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -5100,16 +5100,14 @@ static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb)
 	u32 cmd_length = 0;
 	u16 ipcse = 0, mss;
 	u8 ipcss, ipcso, tucss, tucso, hdr_len;
+	int err;
 
 	if (!skb_is_gso(skb))
 		return 0;
 
-	if (skb_header_cloned(skb)) {
-		int err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-
-		if (err)
-			return err;
-	}
+	err = skb_cow_head(skb, 0);
+	if (err < 0)
+		return err;
 
 	hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
 	mss = skb_shinfo(skb)->gso_size;
-- 
1.9.0

[net 08/13] igb: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/igb/igb_main.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)

diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 3019818..720ad12 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -4605,6 +4605,7 @@ static int igb_tso(struct igb_ring *tx_ring,
 	struct sk_buff *skb = first->skb;
 	u32 vlan_macip_lens, type_tucmd;
 	u32 mss_l4len_idx, l4len;
+	int err;
 
 	if (skb->ip_summed != CHECKSUM_PARTIAL)
 		return 0;
@@ -4612,11 +4613,9 @@ static int igb_tso(struct igb_ring *tx_ring,
 	if (!skb_is_gso(skb))
 		return 0;
 
-	if (skb_header_cloned(skb)) {
-		int err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-		if (err)
-			return err;
-	}
+	err = skb_cow_head(skb, 0);
+	if (err < 0)
+		return err;
 
 	/* ADV DTYP TUCMD MKRLOC/ISCSIHEDLEN */
 	type_tucmd = E1000_ADVTXD_TUCMD_L4T_TCP;
-- 
1.9.0

[net 09/13] igb: fix last_rx_timestamp usage

[Jeff Kirsher]

From: Jakub Kicinski <kubakici@wp.pl>

last_rx_timestamp should be updated only when rx time stamp is
read. Also it's only used with NICs that have per-interface time
stamping resources so it can be moved to adapter structure and
set in igb_ptp_rx_rgtstamp().

Signed-off-by: Jakub Kicinski <kubakici@wp.pl>
Acked-by: Matthew Vick <matthew.vick@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/igb/igb.h      | 16 +---------------
 drivers/net/ethernet/intel/igb/igb_main.c |  4 +++-
 drivers/net/ethernet/intel/igb/igb_ptp.c  | 14 +++++++-------
 3 files changed, 11 insertions(+), 23 deletions(-)

diff --git a/drivers/net/ethernet/intel/igb/igb.h b/drivers/net/ethernet/intel/igb/igb.h
index 7fbe1e9..2713006 100644
--- a/drivers/net/ethernet/intel/igb/igb.h
+++ b/drivers/net/ethernet/intel/igb/igb.h
@@ -241,7 +241,6 @@ struct igb_ring {
 		struct igb_tx_buffer *tx_buffer_info;
 		struct igb_rx_buffer *rx_buffer_info;
 	};
-	unsigned long last_rx_timestamp;
 	void *desc;			/* descriptor ring memory */
 	unsigned long flags;		/* ring specific flags */
 	void __iomem *tail;		/* pointer to ring tail register */
@@ -437,6 +436,7 @@ struct igb_adapter {
 	struct hwtstamp_config tstamp_config;
 	unsigned long ptp_tx_start;
 	unsigned long last_rx_ptp_check;
+	unsigned long last_rx_timestamp;
 	spinlock_t tmreg_lock;
 	struct cyclecounter cc;
 	struct timecounter tc;
@@ -533,20 +533,6 @@ void igb_ptp_rx_hang(struct igb_adapter *adapter);
 void igb_ptp_rx_rgtstamp(struct igb_q_vector *q_vector, struct sk_buff *skb);
 void igb_ptp_rx_pktstamp(struct igb_q_vector *q_vector, unsigned char *va,
 			 struct sk_buff *skb);
-static inline void igb_ptp_rx_hwtstamp(struct igb_ring *rx_ring,
-				       union e1000_adv_rx_desc *rx_desc,
-				       struct sk_buff *skb)
-{
-	if (igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TS) &&
-	    !igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TSIP))
-		igb_ptp_rx_rgtstamp(rx_ring->q_vector, skb);
-
-	/* Update the last_rx_timestamp timer in order to enable watchdog check
-	 * for error case of latched timestamp on a dropped packet.
-	 */
-	rx_ring->last_rx_timestamp = jiffies;
-}
-
 int igb_ptp_set_ts_config(struct net_device *netdev, struct ifreq *ifr);
 int igb_ptp_get_ts_config(struct net_device *netdev, struct ifreq *ifr);
 #ifdef CONFIG_IGB_HWMON
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 720ad12..fb98d46 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -6954,7 +6954,9 @@ static void igb_process_skb_fields(struct igb_ring *rx_ring,
 
 	igb_rx_checksum(rx_ring, rx_desc, skb);
 
-	igb_ptp_rx_hwtstamp(rx_ring, rx_desc, skb);
+	if (igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TS) &&
+	    !igb_test_staterr(rx_desc, E1000_RXDADV_STAT_TSIP))
+		igb_ptp_rx_rgtstamp(rx_ring->q_vector, skb);
 
 	if ((dev->features & NETIF_F_HW_VLAN_CTAG_RX) &&
 	    igb_test_staterr(rx_desc, E1000_RXD_STAT_VP)) {
diff --git a/drivers/net/ethernet/intel/igb/igb_ptp.c b/drivers/net/ethernet/intel/igb/igb_ptp.c
index 2cca8fd..9209d65 100644
--- a/drivers/net/ethernet/intel/igb/igb_ptp.c
+++ b/drivers/net/ethernet/intel/igb/igb_ptp.c
@@ -427,10 +427,8 @@ static void igb_ptp_overflow_check(struct work_struct *work)
 void igb_ptp_rx_hang(struct igb_adapter *adapter)
 {
 	struct e1000_hw *hw = &adapter->hw;
-	struct igb_ring *rx_ring;
 	u32 tsyncrxctl = rd32(E1000_TSYNCRXCTL);
 	unsigned long rx_event;
-	int n;
 
 	if (hw->mac.type != e1000_82576)
 		return;
@@ -445,11 +443,8 @@ void igb_ptp_rx_hang(struct igb_adapter *adapter)
 
 	/* Determine the most recent watchdog or rx_timestamp event */
 	rx_event = adapter->last_rx_ptp_check;
-	for (n = 0; n < adapter->num_rx_queues; n++) {
-		rx_ring = adapter->rx_ring[n];
-		if (time_after(rx_ring->last_rx_timestamp, rx_event))
-			rx_event = rx_ring->last_rx_timestamp;
-	}
+	if (time_after(adapter->last_rx_timestamp, rx_event))
+		rx_event = adapter->last_rx_timestamp;
 
 	/* Only need to read the high RXSTMP register to clear the lock */
 	if (time_is_before_jiffies(rx_event + 5 * HZ)) {
@@ -540,6 +535,11 @@ void igb_ptp_rx_rgtstamp(struct igb_q_vector *q_vector,
 	regval |= (u64)rd32(E1000_RXSTMPH) << 32;
 
 	igb_ptp_systim_to_hwtstamp(adapter, skb_hwtstamps(skb), regval);
+
+	/* Update the last_rx_timestamp timer in order to enable watchdog check
+	 * for error case of latched timestamp on a dropped packet.
+	 */
+	adapter->last_rx_timestamp = jiffies;
 }
 
 /**
-- 
1.9.0

[net 10/13] igbvf: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/igbvf/netdev.c | 16 +++++++---------
 1 file changed, 7 insertions(+), 9 deletions(-)

diff --git a/drivers/net/ethernet/intel/igbvf/netdev.c b/drivers/net/ethernet/intel/igbvf/netdev.c
index b7ab03a..d608599 100644
--- a/drivers/net/ethernet/intel/igbvf/netdev.c
+++ b/drivers/net/ethernet/intel/igbvf/netdev.c
@@ -1910,20 +1910,18 @@ static int igbvf_tso(struct igbvf_adapter *adapter,
                      struct sk_buff *skb, u32 tx_flags, u8 *hdr_len)
 {
 	struct e1000_adv_tx_context_desc *context_desc;
-	unsigned int i;
-	int err;
 	struct igbvf_buffer *buffer_info;
 	u32 info = 0, tu_cmd = 0;
 	u32 mss_l4len_idx, l4len;
+	unsigned int i;
+	int err;
+
 	*hdr_len = 0;
 
-	if (skb_header_cloned(skb)) {
-		err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-		if (err) {
-			dev_err(&adapter->pdev->dev,
-			        "igbvf_tso returning an error\n");
-			return err;
-		}
+	err = skb_cow_head(skb, 0);
+	if (err < 0) {
+		dev_err(&adapter->pdev->dev, "igbvf_tso returning an error\n");
+		return err;
 	}
 
 	l4len = tcp_hdrlen(skb);
-- 
1.9.0

[net 11/13] ixgb: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgb/ixgb_main.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgb/ixgb_main.c b/drivers/net/ethernet/intel/ixgb/ixgb_main.c
index f42c201..6080127 100644
--- a/drivers/net/ethernet/intel/ixgb/ixgb_main.c
+++ b/drivers/net/ethernet/intel/ixgb/ixgb_main.c
@@ -1220,17 +1220,15 @@ ixgb_tso(struct ixgb_adapter *adapter, struct sk_buff *skb)
 	unsigned int i;
 	u8 ipcss, ipcso, tucss, tucso, hdr_len;
 	u16 ipcse, tucse, mss;
-	int err;
 
 	if (likely(skb_is_gso(skb))) {
 		struct ixgb_buffer *buffer_info;
 		struct iphdr *iph;
+		int err;
 
-		if (skb_header_cloned(skb)) {
-			err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-			if (err)
-				return err;
-		}
+		err = skb_cow_head(skb, 0);
+		if (err < 0)
+			return err;
 
 		hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
 		mss = skb_shinfo(skb)->gso_size;
-- 
1.9.0

[net 12/13] i40evf: remove open-coded skb_cow_head

[Jeff Kirsher]

From: Francois Romieu <romieu@fr.zoreil.com>

Signed-off-by: Francois Romieu <romieu@fr.zoreil.com>
Cc: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40e_txrx.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
index 53be5f4..b9f50f4 100644
--- a/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
+++ b/drivers/net/ethernet/intel/i40evf/i40e_txrx.c
@@ -1114,20 +1114,18 @@ static int i40e_tso(struct i40e_ring *tx_ring, struct sk_buff *skb,
 		    u64 *cd_type_cmd_tso_mss, u32 *cd_tunneling)
 {
 	u32 cd_cmd, cd_tso_len, cd_mss;
+	struct ipv6hdr *ipv6h;
 	struct tcphdr *tcph;
 	struct iphdr *iph;
 	u32 l4len;
 	int err;
-	struct ipv6hdr *ipv6h;
 
 	if (!skb_is_gso(skb))
 		return 0;
 
-	if (skb_header_cloned(skb)) {
-		err = pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
-		if (err)
-			return err;
-	}
+	err = skb_cow_head(skb, 0);
+	if (err < 0)
+		return err;
 
 	if (protocol == htons(ETH_P_IP)) {
 		iph = skb->encapsulation ? inner_ip_hdr(skb) : ip_hdr(skb);
-- 
1.9.0

[net 13/13] i40evf: program RSS LUT correctly

[Jeff Kirsher]

From: Mitch A Williams <mitch.a.williams@intel.com>

A recent change broke the RSS LUT programming, causing it to be
programmed with all 0. Correct this by actually assigning the
incremented value back to the counter variable so that the increment
will be remembered by the calling function.

While we're at it, add a proper kernel-doc function comment to our
helper function.

Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/i40evf/i40evf_main.c | 20 ++++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

diff --git a/drivers/net/ethernet/intel/i40evf/i40evf_main.c b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
index e35e66f..2797548 100644
--- a/drivers/net/ethernet/intel/i40evf/i40evf_main.c
+++ b/drivers/net/ethernet/intel/i40evf/i40evf_main.c
@@ -1412,6 +1412,14 @@ restart_watchdog:
 	schedule_work(&adapter->adminq_task);
 }
 
+/**
+ * i40evf_configure_rss - increment to next available tx queue
+ * @adapter: board private structure
+ * @j: queue counter
+ *
+ * Helper function for RSS programming to increment through available
+ * queus. Returns the next queue value.
+ **/
 static int next_queue(struct i40evf_adapter *adapter, int j)
 {
 	j += 1;
@@ -1451,10 +1459,14 @@ static void i40evf_configure_rss(struct i40evf_adapter *adapter)
 	/* Populate the LUT with max no. of queues in round robin fashion */
 	j = adapter->vsi_res->num_queue_pairs;
 	for (i = 0; i <= I40E_VFQF_HLUT_MAX_INDEX; i++) {
-		lut = next_queue(adapter, j);
-		lut |= next_queue(adapter, j) << 8;
-		lut |= next_queue(adapter, j) << 16;
-		lut |= next_queue(adapter, j) << 24;
+		j = next_queue(adapter, j);
+		lut = j;
+		j = next_queue(adapter, j);
+		lut |= j << 8;
+		j = next_queue(adapter, j);
+		lut |= j << 16;
+		j = next_queue(adapter, j);
+		lut |= j << 24;
 		wr32(hw, I40E_VFQF_HLUT(i), lut);
 	}
 	i40e_flush(hw);
-- 
1.9.0

Re: [net 00/13][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Fri, 11 Apr 2014 06:52:50 -0700

> This series contains updates to e1000, e1000e, igb, igbvf, ixgb, ixgbe,
> ixgbevf and i40evf.
> 
> Mark fixes an issue with ixgbe and ixgbevf by adding a bit to indicate
> when workqueues have been initialized.  This permits the register read
> error handling from attempting to use them prior to that, which also
> generates warnings.  Checking for a detected removal after initializing
> the work queues allows the probe function to return an error without
> getting the workqueue involved.  Further, if the error_detected
> callback is entered before the workqueues are initialized, exit without
> recovery since the device initialization was so truncated.
> 
> Francois Romieu provides several patches to all the drivers to remove
> the open coded skb_cow_head.
> 
> Jakub Kicinski provides a fix for igb where last_rx_timestamp should be
> updated only when Rx time stamp is read.
> 
> Mitch provides a fix for i40evf where a recent change broke the RSS LUT
> programming causing it to be programmed with all 0's.

Pulled, thanks a lot Jeff.