[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

The following series contains updates to e1000e only.  Most of these
changes contain fixup/cleanups/conversions.

The following are changes since commit e55684fadbc5c9c69cad1b3bae228c8374b99b85:
  infiniband: nes: Convert nes_addr_resolve_neigh() over to dst_neigh_lookup().
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

  e1000e: cleanup Rx checksum offload code
  e1000e: add Receive Packet Steering (RPS) support
  e1000e: re-enable alternate MAC address for all devices which support
    it
  e1000e: convert head, tail and itr_register offsets to __iomem
    pointers
  e1000e: pass pointer to ring struct instead of adapter struct
  e1000e: re-factor ethtool get/set ring parameter
  e1000e: default IntMode based on kernel config & available hardware
    support
  e1000e: always set transmit descriptor control registers the same
  e1000e: 82579: workaround for link drop issue
  e1000e: use default settings for Tx Inter Packet Gap timer
  e1000e: use hardware default values for Transmit Control register
  e1000e: 82574/82583 Tx hang workaround

 drivers/net/ethernet/intel/e1000e/80003es2lan.c |    3 +-
 drivers/net/ethernet/intel/e1000e/82571.c       |    4 +
 drivers/net/ethernet/intel/e1000e/defines.h     |    8 +
 drivers/net/ethernet/intel/e1000e/e1000.h       |   29 +-
 drivers/net/ethernet/intel/e1000e/ethtool.c     |  186 +++++++----
 drivers/net/ethernet/intel/e1000e/hw.h          |    9 +-
 drivers/net/ethernet/intel/e1000e/ich8lan.c     |   22 ++
 drivers/net/ethernet/intel/e1000e/lib.c         |    7 +-
 drivers/net/ethernet/intel/e1000e/netdev.c      |  423 +++++++++++++----------
 drivers/net/ethernet/intel/e1000e/param.c       |   48 +++-
 10 files changed, 471 insertions(+), 268 deletions(-)

-- 
1.7.7.6

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Wed, 25 Jan 2012 23:21:43 -0800

> The following are changes since commit e55684fadbc5c9c69cad1b3bae228c8374b99b85:
>   infiniband: nes: Convert nes_addr_resolve_neigh() over to dst_neigh_lookup().
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
> 
>   e1000e: cleanup Rx checksum offload code
>   e1000e: add Receive Packet Steering (RPS) support
>   e1000e: re-enable alternate MAC address for all devices which support
>     it
>   e1000e: convert head, tail and itr_register offsets to __iomem
>     pointers
>   e1000e: pass pointer to ring struct instead of adapter struct
>   e1000e: re-factor ethtool get/set ring parameter
>   e1000e: default IntMode based on kernel config & available hardware
>     support
>   e1000e: always set transmit descriptor control registers the same
>   e1000e: 82579: workaround for link drop issue
>   e1000e: use default settings for Tx Inter Packet Gap timer
>   e1000e: use hardware default values for Transmit Control register
>   e1000e: 82574/82583 Tx hang workaround

Pulled, thanks Jeff.

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: David Miller <davem@davemloft.net>
Date: Thu, 26 Jan 2012 13:49:13 -0500 (EST)

> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> Date: Wed, 25 Jan 2012 23:21:43 -0800
> 
>> The following are changes since commit e55684fadbc5c9c69cad1b3bae228c8374b99b85:
>>   infiniband: nes: Convert nes_addr_resolve_neigh() over to dst_neigh_lookup().
>> and are available in the git repository at:
>>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
 ...
> Pulled, thanks Jeff.

I had to apply the following patch to fix the build on non-x86
architectures.

--------------------
e1000e: Need to include vmalloc.h

Otherwise (on sparc64):

drivers/net/ethernet/intel/e1000e/ethtool.c:657:3: error: implicit declaration of function 'vmalloc' [-Werror=implicit-function-declaration]

Signed-off-by: David S. Miller <davem@davemloft.net>
---
 drivers/net/ethernet/intel/e1000e/ethtool.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/net/ethernet/intel/e1000e/ethtool.c
index ffb6c14..1ea317f 100644
--- a/drivers/net/ethernet/intel/e1000e/ethtool.c
+++ b/drivers/net/ethernet/intel/e1000e/ethtool.c
@@ -34,6 +34,7 @@
 #include <linux/pci.h>
 #include <linux/slab.h>
 #include <linux/delay.h>
+#include <linux/vmalloc.h>
 
 #include "e1000.h"
 
-- 
1.7.7.6

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series of patches contains additions/cleanups to igb and ixgbe.
There 2 patches for igb & ixgbe which add FX-ALL feature flag and
sending of custom Ethernet FCS from Ben Greear.

The remaining patches in the series is part three of three to update
ixgbe.  Although it looks like there will be at least one follow on
patch series complete the update/cleanup of ixgbe.

The following are changes since commit 126a3fd251b244eabd9ab9dcb32b8b6f999c1b91:
  eni: fix driver remove function and driver probe error path.
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Alexander Duyck (8):
  ixgbe: Replace standard receive path with a page based receive
  ixgbe: cleanup logic in ixgbe_change_mtu
  ixgbe: Make certain that all frames fit minimum size requirements
  ixgbe: Modify setup of descriptor flags to avoid conditional jumps
  ixgbe: Use packets to track Tx completions instead of a separate
    value
  ixgbe: Place skb on first buffer_info structure to avoid using stack
    space
  ixgbe: Write gso_segs and bytcount to the ring sooner
  ixgbe: always write DMA for single_mapped value with skb

Ben Greear (4):
  igb: Support sending custom Ethernet FCS.
  igb: Support RX-ALL feature flag.
  ixgbe: Support sending custom Ethernet FCS.
  ixgbe: Support RX-ALL feature flag.

 drivers/net/ethernet/intel/igb/e1000_defines.h   |    2 +
 drivers/net/ethernet/intel/igb/igb_main.c        |   37 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe.h         |   60 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c |   23 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_fcoe.c    |   17 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c    | 1261 ++++++++++++----------
 6 files changed, 772 insertions(+), 628 deletions(-)

-- 
1.7.7.6

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Sat, 17 Mar 2012 01:50:48 -0700

> This series of patches contains additions/cleanups to igb and ixgbe.
> There 2 patches for igb & ixgbe which add FX-ALL feature flag and
> sending of custom Ethernet FCS from Ben Greear.
> 
> The remaining patches in the series is part three of three to update
> ixgbe.  Although it looks like there will be at least one follow on
> patch series complete the update/cleanup of ixgbe.
> 
> The following are changes since commit 126a3fd251b244eabd9ab9dcb32b8b6f999c1b91:
>   eni: fix driver remove function and driver probe error path.
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Pulled, thanks Jeff.

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series of patches contains updates for igb and ixgbe.
Special note ixgbe patches contain the addition of PTP support from
Jacob Keller.

The following are changes since commit 8feedbb4a710784d2858acba5c90e903e93e36eb:
  dsa: Convert compare_ether_addr to ether_addr_equal
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Alexander Duyck (3):
  ixgbe: Exit on error case in VF message processing
  ixgbe: Clean up priority based flow control
  ixgbe: Set Drop_EN bit when multiple Rx queues are present w/o flow
    control

Don Skidmore (2):
  ixgbe: cleanup the hwmon function calls
  ixgbe: update version number

Greg Rose (1):
  ixgbe: Fix bogus error message

Jacob E Keller (1):
  ixgbe: Enable timesync clock-out feature for PPS support on X540

Jacob Keller (4):
  ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)
  ixgbe: correct disable_rx_buff timeout
  ixgbe: add support for get_ts_info
  ixgbe: support software timestamping

Koki Sanagi (1):
  igb: output register's information related to RX/TX queue[4-15]

 drivers/net/ethernet/intel/Kconfig                 |   11 +
 drivers/net/ethernet/intel/igb/igb_ethtool.c       |   38 +-
 drivers/net/ethernet/intel/ixgbe/Makefile          |    2 +
 drivers/net/ethernet/intel/ixgbe/ixgbe.h           |   37 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c    |    2 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82598.c |   69 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.c |   93 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c    |   98 +--
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c   |   71 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c      |  198 ++++-
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c       |  900 ++++++++++++++++++++
 drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c     |   13 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_sysfs.c     |   36 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_type.h      |   47 +-
 14 files changed, 1370 insertions(+), 245 deletions(-)
 create mode 100644 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c

-- 
1.7.7.6

[net-next 01/12] igb: output register's information related to RX/TX queue[4-15]

[Jeff Kirsher]

From: Koki Sanagi <sanagi.koki@jp.fujitsu.com>

Current igb outputs registers related to TX/RX queues(ex. RDT, RDH, TDT, TDH).
But it thinks the number of RX/TX queues is 4. But 82576 has 16 RX/TX queues.
This patch modifies igb to output the rest of the registers if the device is
82576.

Signed-off-by: Koki Sanagi <sanagi.koki@jp.fujitsu.com>
Acked-by: Carolyn Wyborny <carolyn.wyborny@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_ethtool.c |   38 +++++++++++++++++++++++++-
 1 files changed, 37 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/intel/igb/igb_ethtool.c b/drivers/net/ethernet/intel/igb/igb_ethtool.c
index e10821a..faadf23 100644
--- a/drivers/net/ethernet/intel/igb/igb_ethtool.c
+++ b/drivers/net/ethernet/intel/igb/igb_ethtool.c
@@ -335,7 +335,7 @@ static void igb_set_msglevel(struct net_device *netdev, u32 data)
 
 static int igb_get_regs_len(struct net_device *netdev)
 {
-#define IGB_REGS_LEN 551
+#define IGB_REGS_LEN 739
 	return IGB_REGS_LEN * sizeof(u32);
 }
 
@@ -556,6 +556,42 @@ static void igb_get_regs(struct net_device *netdev,
 	regs_buff[552] = adapter->stats.b2ospc;
 	regs_buff[553] = adapter->stats.o2bspc;
 	regs_buff[554] = adapter->stats.b2ogprc;
+
+	if (hw->mac.type != e1000_82576)
+		return;
+	for (i = 0; i < 12; i++)
+		regs_buff[555 + i] = rd32(E1000_SRRCTL(i + 4));
+	for (i = 0; i < 4; i++)
+		regs_buff[567 + i] = rd32(E1000_PSRTYPE(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[571 + i] = rd32(E1000_RDBAL(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[583 + i] = rd32(E1000_RDBAH(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[595 + i] = rd32(E1000_RDLEN(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[607 + i] = rd32(E1000_RDH(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[619 + i] = rd32(E1000_RDT(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[631 + i] = rd32(E1000_RXDCTL(i + 4));
+
+	for (i = 0; i < 12; i++)
+		regs_buff[643 + i] = rd32(E1000_TDBAL(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[655 + i] = rd32(E1000_TDBAH(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[667 + i] = rd32(E1000_TDLEN(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[679 + i] = rd32(E1000_TDH(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[691 + i] = rd32(E1000_TDT(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[703 + i] = rd32(E1000_TXDCTL(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[715 + i] = rd32(E1000_TDWBAL(i + 4));
+	for (i = 0; i < 12; i++)
+		regs_buff[727 + i] = rd32(E1000_TDWBAH(i + 4));
 }
 
 static int igb_get_eeprom_len(struct net_device *netdev)
-- 
1.7.7.6

[net-next 02/12] ixgbe: Exit on error case in VF message processing

[Jeff Kirsher]

From: Alexander Duyck <alexander.h.duyck@intel.com>

Previously we would get a mailbox error and still process the message.
Instead we should exit on error.

In addition we should also be flushing the ACK of the message so that we
can guarantee that the other end is aware we have received the message
while we are processing it.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c |    7 ++++++-
 1 files changed, 6 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c
index 3985637..b64f1fc 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c
@@ -544,13 +544,18 @@ static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf)
 
 	retval = ixgbe_read_mbx(hw, msgbuf, mbx_size, vf);
 
-	if (retval)
+	if (retval) {
 		pr_err("Error receiving message from VF\n");
+		return retval;
+	}
 
 	/* this is a message we already processed, do nothing */
 	if (msgbuf[0] & (IXGBE_VT_MSGTYPE_ACK | IXGBE_VT_MSGTYPE_NACK))
 		return retval;
 
+	/* flush the ack before we write any messages back */
+	IXGBE_WRITE_FLUSH(hw);
+
 	/*
 	 * until the vf completes a virtual function reset it should not be
 	 * allowed to start any configuration.
-- 
1.7.7.6

[net-next 03/12] ixgbe: Clean up priority based flow control

[Jeff Kirsher]

From: Alexander Duyck <alexander.h.duyck@intel.com>

This change cleans up the logic in the priority based flow control
configuration routines.  Both the 82599 and 82598 based routines perform
similar functions however they are both arranged completely differently.
This patch goes over both of them to clean up the code.

In addition I am dropping the ixgbe_fc_pfc flow control mode and instead
just replacing it with checks for if priority flow control is enabled.
This allows us to maintain some of the link flow control information which
allows for an easier transition between link and priority flow control.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82598.c |   69 +++++++--------
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.c |   93 +++++++++----------
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c    |   95 +++++++-------------
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c   |   30 ++-----
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c      |   70 ++++++++------
 drivers/net/ethernet/intel/ixgbe/ixgbe_type.h      |    3 -
 6 files changed, 156 insertions(+), 204 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82598.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82598.c
index d3695ed..87592b4 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82598.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82598.c
@@ -191,53 +191,46 @@ s32 ixgbe_dcb_config_tx_data_arbiter_82598(struct ixgbe_hw *hw,
  */
 s32 ixgbe_dcb_config_pfc_82598(struct ixgbe_hw *hw, u8 pfc_en)
 {
-	u32 reg;
+	u32 fcrtl, reg;
 	u8  i;
 
-	if (pfc_en) {
-		/* Enable Transmit Priority Flow Control */
-		reg = IXGBE_READ_REG(hw, IXGBE_RMCS);
-		reg &= ~IXGBE_RMCS_TFCE_802_3X;
-		/* correct the reporting of our flow control status */
-		reg |= IXGBE_RMCS_TFCE_PRIORITY;
-		IXGBE_WRITE_REG(hw, IXGBE_RMCS, reg);
-
-		/* Enable Receive Priority Flow Control */
-		reg = IXGBE_READ_REG(hw, IXGBE_FCTRL);
-		reg &= ~IXGBE_FCTRL_RFCE;
-		reg |= IXGBE_FCTRL_RPFCE;
-		IXGBE_WRITE_REG(hw, IXGBE_FCTRL, reg);
-
-		/* Configure pause time */
-		for (i = 0; i < (MAX_TRAFFIC_CLASS >> 1); i++)
-			IXGBE_WRITE_REG(hw, IXGBE_FCTTV(i), 0x68006800);
+	/* Enable Transmit Priority Flow Control */
+	reg = IXGBE_READ_REG(hw, IXGBE_RMCS);
+	reg &= ~IXGBE_RMCS_TFCE_802_3X;
+	reg |= IXGBE_RMCS_TFCE_PRIORITY;
+	IXGBE_WRITE_REG(hw, IXGBE_RMCS, reg);
 
-		/* Configure flow control refresh threshold value */
-		IXGBE_WRITE_REG(hw, IXGBE_FCRTV, 0x3400);
-	}
+	/* Enable Receive Priority Flow Control */
+	reg = IXGBE_READ_REG(hw, IXGBE_FCTRL);
+	reg &= ~(IXGBE_FCTRL_RPFCE | IXGBE_FCTRL_RFCE);
 
-	/*
-	 * Configure flow control thresholds and enable priority flow control
-	 * for each traffic class.
-	 */
-	for (i = 0; i < MAX_TRAFFIC_CLASS; i++) {
-		int enabled = pfc_en & (1 << i);
+	if (pfc_en)
+		reg |= IXGBE_FCTRL_RPFCE;
 
-		reg = hw->fc.low_water << 10;
+	IXGBE_WRITE_REG(hw, IXGBE_FCTRL, reg);
 
-		if (enabled == pfc_enabled_tx ||
-		    enabled == pfc_enabled_full)
-			reg |= IXGBE_FCRTL_XONE;
+	fcrtl = (hw->fc.low_water << 10) | IXGBE_FCRTL_XONE;
+	/* Configure PFC Tx thresholds per TC */
+	for (i = 0; i < MAX_TRAFFIC_CLASS; i++) {
+		if (!(pfc_en & (1 << i))) {
+			IXGBE_WRITE_REG(hw, IXGBE_FCRTL(i), 0);
+			IXGBE_WRITE_REG(hw, IXGBE_FCRTH(i), 0);
+			continue;
+		}
+
+		reg = (hw->fc.high_water[i] << 10) | IXGBE_FCRTH_FCEN;
+		IXGBE_WRITE_REG(hw, IXGBE_FCRTL(i), fcrtl);
+		IXGBE_WRITE_REG(hw, IXGBE_FCRTH(i), reg);
+	}
 
-		IXGBE_WRITE_REG(hw, IXGBE_FCRTL(i), reg);
+	/* Configure pause time */
+	reg = hw->fc.pause_time * 0x00010001;
+	for (i = 0; i < (MAX_TRAFFIC_CLASS / 2); i++)
+		IXGBE_WRITE_REG(hw, IXGBE_FCTTV(i), reg);
 
-		reg = hw->fc.high_water[i] << 10;
-		if (enabled == pfc_enabled_tx ||
-		    enabled == pfc_enabled_full)
-			reg |= IXGBE_FCRTH_FCEN;
+	/* Configure flow control refresh threshold value */
+	IXGBE_WRITE_REG(hw, IXGBE_FCRTV, hw->fc.pause_time / 2);
 
-		IXGBE_WRITE_REG(hw, IXGBE_FCRTH(i), reg);
-	}
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.c
index 65913c5..4eac80d 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.c
@@ -211,24 +211,42 @@ s32 ixgbe_dcb_config_tx_data_arbiter_82599(struct ixgbe_hw *hw,
  */
 s32 ixgbe_dcb_config_pfc_82599(struct ixgbe_hw *hw, u8 pfc_en, u8 *prio_tc)
 {
-	u32 i, j, reg;
+	u32 i, j, fcrtl, reg;
 	u8 max_tc = 0;
 
-	for (i = 0; i < MAX_USER_PRIORITY; i++)
+	/* Enable Transmit Priority Flow Control */
+	IXGBE_WRITE_REG(hw, IXGBE_FCCFG, IXGBE_FCCFG_TFCE_PRIORITY);
+
+	/* Enable Receive Priority Flow Control */
+	reg = IXGBE_READ_REG(hw, IXGBE_MFLCN);
+	reg |= IXGBE_MFLCN_DPF;
+
+	/*
+	 * X540 supports per TC Rx priority flow control.  So
+	 * clear all TCs and only enable those that should be
+	 * enabled.
+	 */
+	reg &= ~(IXGBE_MFLCN_RPFCE_MASK | IXGBE_MFLCN_RFCE);
+
+	if (hw->mac.type == ixgbe_mac_X540)
+		reg |= pfc_en << IXGBE_MFLCN_RPFCE_SHIFT;
+
+	if (pfc_en)
+		reg |= IXGBE_MFLCN_RPFCE;
+
+	IXGBE_WRITE_REG(hw, IXGBE_MFLCN, reg);
+
+	for (i = 0; i < MAX_USER_PRIORITY; i++) {
 		if (prio_tc[i] > max_tc)
 			max_tc = prio_tc[i];
+	}
+
+	fcrtl = (hw->fc.low_water << 10) | IXGBE_FCRTL_XONE;
 
 	/* Configure PFC Tx thresholds per TC */
-	for (i = 0; i < MAX_TRAFFIC_CLASS; i++) {
+	for (i = 0; i <= max_tc; i++) {
 		int enabled = 0;
 
-		if (i > max_tc) {
-			reg = 0;
-			IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(i), reg);
-			IXGBE_WRITE_REG(hw, IXGBE_FCRTH_82599(i), reg);
-			continue;
-		}
-
 		for (j = 0; j < MAX_USER_PRIORITY; j++) {
 			if ((prio_tc[j] == i) && (pfc_en & (1 << j))) {
 				enabled = 1;
@@ -236,50 +254,29 @@ s32 ixgbe_dcb_config_pfc_82599(struct ixgbe_hw *hw, u8 pfc_en, u8 *prio_tc)
 			}
 		}
 
-		reg = hw->fc.low_water << 10;
-
-		if (enabled)
-			reg |= IXGBE_FCRTL_XONE;
-		IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(i), reg);
+		if (enabled) {
+			reg = (hw->fc.high_water[i] << 10) | IXGBE_FCRTH_FCEN;
+			IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(i), fcrtl);
+		} else {
+			reg = IXGBE_READ_REG(hw, IXGBE_RXPBSIZE(i)) - 32;
+			IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(i), 0);
+		}
 
-		reg = hw->fc.high_water[i] << 10;
-		if (enabled)
-			reg |= IXGBE_FCRTH_FCEN;
 		IXGBE_WRITE_REG(hw, IXGBE_FCRTH_82599(i), reg);
 	}
 
-	if (pfc_en) {
-		/* Configure pause time (2 TCs per register) */
-		reg = hw->fc.pause_time | (hw->fc.pause_time << 16);
-		for (i = 0; i < (MAX_TRAFFIC_CLASS / 2); i++)
-			IXGBE_WRITE_REG(hw, IXGBE_FCTTV(i), reg);
-
-		/* Configure flow control refresh threshold value */
-		IXGBE_WRITE_REG(hw, IXGBE_FCRTV, hw->fc.pause_time / 2);
-
-
-		reg = IXGBE_FCCFG_TFCE_PRIORITY;
-		IXGBE_WRITE_REG(hw, IXGBE_FCCFG, reg);
-		/*
-		 * Enable Receive PFC
-		 * 82599 will always honor XOFF frames we receive when
-		 * we are in PFC mode however X540 only honors enabled
-		 * traffic classes.
-		 */
-		reg = IXGBE_READ_REG(hw, IXGBE_MFLCN);
-		reg &= ~IXGBE_MFLCN_RFCE;
-		reg |= IXGBE_MFLCN_RPFCE | IXGBE_MFLCN_DPF;
-
-		if (hw->mac.type == ixgbe_mac_X540) {
-			reg &= ~IXGBE_MFLCN_RPFCE_MASK;
-			reg |= pfc_en << IXGBE_MFLCN_RPFCE_SHIFT;
-		}
+	for (; i < MAX_TRAFFIC_CLASS; i++) {
+		IXGBE_WRITE_REG(hw, IXGBE_FCRTL_82599(i), 0);
+		IXGBE_WRITE_REG(hw, IXGBE_FCRTH_82599(i), 0);
+	}
 
-		IXGBE_WRITE_REG(hw, IXGBE_MFLCN, reg);
+	/* Configure pause time (2 TCs per register) */
+	reg = hw->fc.pause_time * 0x00010001;
+	for (i = 0; i < (MAX_TRAFFIC_CLASS / 2); i++)
+		IXGBE_WRITE_REG(hw, IXGBE_FCTTV(i), reg);
 
-	} else {
-		hw->mac.ops.fc_enable(hw);
-	}
+	/* Configure flow control refresh threshold value */
+	IXGBE_WRITE_REG(hw, IXGBE_FCRTV, hw->fc.pause_time / 2);
 
 	return 0;
 }
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
index a09d6b4..7f2fa69 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
@@ -338,6 +338,8 @@ static void ixgbe_dcbnl_devreset(struct net_device *dev)
 static u8 ixgbe_dcbnl_set_all(struct net_device *netdev)
 {
 	struct ixgbe_adapter *adapter = netdev_priv(netdev);
+	struct ixgbe_dcb_config *dcb_cfg = &adapter->dcb_cfg;
+	struct ixgbe_hw *hw = &adapter->hw;
 	int ret = DCB_NO_HW_CHG;
 	int i;
 
@@ -350,32 +352,6 @@ static u8 ixgbe_dcbnl_set_all(struct net_device *netdev)
 	if (!adapter->dcb_set_bitmap)
 		return ret;
 
-	if (adapter->dcb_cfg.pfc_mode_enable) {
-		switch (adapter->hw.mac.type) {
-		case ixgbe_mac_82599EB:
-		case ixgbe_mac_X540:
-			if (adapter->hw.fc.current_mode != ixgbe_fc_pfc)
-				adapter->last_lfc_mode =
-				                  adapter->hw.fc.current_mode;
-			break;
-		default:
-			break;
-		}
-		adapter->hw.fc.requested_mode = ixgbe_fc_pfc;
-	} else {
-		switch (adapter->hw.mac.type) {
-		case ixgbe_mac_82598EB:
-			adapter->hw.fc.requested_mode = ixgbe_fc_none;
-			break;
-		case ixgbe_mac_82599EB:
-		case ixgbe_mac_X540:
-			adapter->hw.fc.requested_mode = adapter->last_lfc_mode;
-			break;
-		default:
-			break;
-		}
-	}
-
 	if (adapter->dcb_set_bitmap & (BIT_PG_TX|BIT_PG_RX)) {
 		u16 refill[MAX_TRAFFIC_CLASS], max[MAX_TRAFFIC_CLASS];
 		u8 bwg_id[MAX_TRAFFIC_CLASS], prio_type[MAX_TRAFFIC_CLASS];
@@ -388,23 +364,19 @@ static u8 ixgbe_dcbnl_set_all(struct net_device *netdev)
 			max_frame = max(max_frame, IXGBE_FCOE_JUMBO_FRAME_SIZE);
 #endif
 
-		ixgbe_dcb_calculate_tc_credits(&adapter->hw, &adapter->dcb_cfg,
-					       max_frame, DCB_TX_CONFIG);
-		ixgbe_dcb_calculate_tc_credits(&adapter->hw, &adapter->dcb_cfg,
-					       max_frame, DCB_RX_CONFIG);
+		ixgbe_dcb_calculate_tc_credits(hw, dcb_cfg, max_frame,
+					       DCB_TX_CONFIG);
+		ixgbe_dcb_calculate_tc_credits(hw, dcb_cfg, max_frame,
+					       DCB_RX_CONFIG);
 
-		ixgbe_dcb_unpack_refill(&adapter->dcb_cfg,
-					DCB_TX_CONFIG, refill);
-		ixgbe_dcb_unpack_max(&adapter->dcb_cfg, max);
-		ixgbe_dcb_unpack_bwgid(&adapter->dcb_cfg,
-				       DCB_TX_CONFIG, bwg_id);
-		ixgbe_dcb_unpack_prio(&adapter->dcb_cfg,
-				      DCB_TX_CONFIG, prio_type);
-		ixgbe_dcb_unpack_map(&adapter->dcb_cfg,
-				     DCB_TX_CONFIG, prio_tc);
+		ixgbe_dcb_unpack_refill(dcb_cfg, DCB_TX_CONFIG, refill);
+		ixgbe_dcb_unpack_max(dcb_cfg, max);
+		ixgbe_dcb_unpack_bwgid(dcb_cfg, DCB_TX_CONFIG, bwg_id);
+		ixgbe_dcb_unpack_prio(dcb_cfg, DCB_TX_CONFIG, prio_type);
+		ixgbe_dcb_unpack_map(dcb_cfg, DCB_TX_CONFIG, prio_tc);
 
-		ixgbe_dcb_hw_ets_config(&adapter->hw, refill, max,
-					bwg_id, prio_type, prio_tc);
+		ixgbe_dcb_hw_ets_config(hw, refill, max, bwg_id,
+					prio_type, prio_tc);
 
 		for (i = 0; i < IEEE_8021QAZ_MAX_TCS; i++)
 			netdev_set_prio_tc_map(netdev, i, prio_tc[i]);
@@ -413,20 +385,19 @@ static u8 ixgbe_dcbnl_set_all(struct net_device *netdev)
 	}
 
 	if (adapter->dcb_set_bitmap & BIT_PFC) {
-		u8 pfc_en;
-		u8 prio_tc[MAX_USER_PRIORITY];
-
-		ixgbe_dcb_unpack_map(&adapter->dcb_cfg,
-				     DCB_TX_CONFIG, prio_tc);
-		ixgbe_dcb_unpack_pfc(&adapter->dcb_cfg, &pfc_en);
-		ixgbe_dcb_hw_pfc_config(&adapter->hw, pfc_en, prio_tc);
-		if (ret != DCB_HW_CHG_RST)
-			ret = DCB_HW_CHG;
+		if (dcb_cfg->pfc_mode_enable) {
+			u8 pfc_en;
+			u8 prio_tc[MAX_USER_PRIORITY];
+
+			ixgbe_dcb_unpack_map(dcb_cfg, DCB_TX_CONFIG, prio_tc);
+			ixgbe_dcb_unpack_pfc(dcb_cfg, &pfc_en);
+			ixgbe_dcb_hw_pfc_config(hw, pfc_en, prio_tc);
+		} else {
+			hw->mac.ops.fc_enable(hw);
+		}
+		ret = DCB_HW_CHG;
 	}
 
-	if (adapter->dcb_cfg.pfc_mode_enable)
-		adapter->hw.fc.current_mode = ixgbe_fc_pfc;
-
 #ifdef IXGBE_FCOE
 	/* Reprogam FCoE hardware offloads when the traffic class
 	 * FCoE is using changes. This happens if the APP info
@@ -647,7 +618,9 @@ static int ixgbe_dcbnl_ieee_setpfc(struct net_device *dev,
 				   struct ieee_pfc *pfc)
 {
 	struct ixgbe_adapter *adapter = netdev_priv(dev);
+	struct ixgbe_hw *hw = &adapter->hw;
 	u8 *prio_tc;
+	int err;
 
 	if (!(adapter->dcbx_cap & DCB_CAP_DCBX_VER_IEEE))
 		return -EINVAL;
@@ -659,16 +632,16 @@ static int ixgbe_dcbnl_ieee_setpfc(struct net_device *dev,
 			return -ENOMEM;
 	}
 
-	if (pfc->pfc_en) {
-		adapter->last_lfc_mode = adapter->hw.fc.current_mode;
-		adapter->hw.fc.current_mode = ixgbe_fc_pfc;
-	} else {
-		adapter->hw.fc.current_mode = adapter->last_lfc_mode;
-	}
-
 	prio_tc = adapter->ixgbe_ieee_ets->prio_tc;
 	memcpy(adapter->ixgbe_ieee_pfc, pfc, sizeof(*adapter->ixgbe_ieee_pfc));
-	return ixgbe_dcb_hw_pfc_config(&adapter->hw, pfc->pfc_en, prio_tc);
+
+	/* Enable link flow control parameters if PFC is disabled */
+	if (pfc->pfc_en)
+		err = ixgbe_dcb_hw_pfc_config(hw, pfc->pfc_en, prio_tc);
+	else
+		err = hw->mac.ops.fc_enable(hw);
+
+	return err;
 }
 
 static int ixgbe_dcbnl_ieee_setapp(struct net_device *dev,
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
index cca3e9c..082446d 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
@@ -391,11 +391,6 @@ static void ixgbe_get_pauseparam(struct net_device *netdev,
 	} else if (hw->fc.current_mode == ixgbe_fc_full) {
 		pause->rx_pause = 1;
 		pause->tx_pause = 1;
-#ifdef CONFIG_DCB
-	} else if (hw->fc.current_mode == ixgbe_fc_pfc) {
-		pause->rx_pause = 0;
-		pause->tx_pause = 0;
-#endif
 	}
 }
 
@@ -404,21 +399,14 @@ static int ixgbe_set_pauseparam(struct net_device *netdev,
 {
 	struct ixgbe_adapter *adapter = netdev_priv(netdev);
 	struct ixgbe_hw *hw = &adapter->hw;
-	struct ixgbe_fc_info fc;
+	struct ixgbe_fc_info fc = hw->fc;
 
-#ifdef CONFIG_DCB
-	if (adapter->dcb_cfg.pfc_mode_enable ||
-		((hw->mac.type == ixgbe_mac_82598EB) &&
-		(adapter->flags & IXGBE_FLAG_DCB_ENABLED)))
+	/* 82598 does no support link flow control with DCB enabled */
+	if ((hw->mac.type == ixgbe_mac_82598EB) &&
+	    (adapter->flags & IXGBE_FLAG_DCB_ENABLED))
 		return -EINVAL;
 
-#endif
-	fc = hw->fc;
-
-	if (pause->autoneg != AUTONEG_ENABLE)
-		fc.disable_fc_autoneg = true;
-	else
-		fc.disable_fc_autoneg = false;
+	fc.disable_fc_autoneg = (pause->autoneg != AUTONEG_ENABLE);
 
 	if ((pause->rx_pause && pause->tx_pause) || pause->autoneg)
 		fc.requested_mode = ixgbe_fc_full;
@@ -426,14 +414,8 @@ static int ixgbe_set_pauseparam(struct net_device *netdev,
 		fc.requested_mode = ixgbe_fc_rx_pause;
 	else if (!pause->rx_pause && pause->tx_pause)
 		fc.requested_mode = ixgbe_fc_tx_pause;
-	else if (!pause->rx_pause && !pause->tx_pause)
-		fc.requested_mode = ixgbe_fc_none;
 	else
-		return -EINVAL;
-
-#ifdef CONFIG_DCB
-	adapter->last_lfc_mode = fc.requested_mode;
-#endif
+		fc.requested_mode = ixgbe_fc_none;
 
 	/* if the thing changed then we'll update and use new autoneg */
 	if (memcmp(&fc, &hw->fc, sizeof(struct ixgbe_fc_info))) {
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index ea3cb71..0915e77 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -610,39 +610,50 @@ void ixgbe_unmap_and_free_tx_resource(struct ixgbe_ring *ring,
 	/* tx_buffer must be completely set up in the transmit path */
 }
 
-static void ixgbe_update_xoff_received(struct ixgbe_adapter *adapter)
+static void ixgbe_update_xoff_rx_lfc(struct ixgbe_adapter *adapter)
 {
 	struct ixgbe_hw *hw = &adapter->hw;
 	struct ixgbe_hw_stats *hwstats = &adapter->stats;
-	u32 data = 0;
-	u32 xoff[8] = {0};
 	int i;
+	u32 data;
 
-	if ((hw->fc.current_mode == ixgbe_fc_full) ||
-	    (hw->fc.current_mode == ixgbe_fc_rx_pause)) {
-		switch (hw->mac.type) {
-		case ixgbe_mac_82598EB:
-			data = IXGBE_READ_REG(hw, IXGBE_LXOFFRXC);
-			break;
-		default:
-			data = IXGBE_READ_REG(hw, IXGBE_LXOFFRXCNT);
-		}
-		hwstats->lxoffrxc += data;
+	if ((hw->fc.current_mode != ixgbe_fc_full) &&
+	    (hw->fc.current_mode != ixgbe_fc_rx_pause))
+		return;
 
-		/* refill credits (no tx hang) if we received xoff */
-		if (!data)
-			return;
+	switch (hw->mac.type) {
+	case ixgbe_mac_82598EB:
+		data = IXGBE_READ_REG(hw, IXGBE_LXOFFRXC);
+		break;
+	default:
+		data = IXGBE_READ_REG(hw, IXGBE_LXOFFRXCNT);
+	}
+	hwstats->lxoffrxc += data;
 
-		for (i = 0; i < adapter->num_tx_queues; i++)
-			clear_bit(__IXGBE_HANG_CHECK_ARMED,
-				  &adapter->tx_ring[i]->state);
+	/* refill credits (no tx hang) if we received xoff */
+	if (!data)
 		return;
-	} else if (((adapter->dcbx_cap & DCB_CAP_DCBX_VER_CEE) &&
-		    !(adapter->dcb_cfg.pfc_mode_enable)) ||
-		   ((adapter->dcbx_cap & DCB_CAP_DCBX_VER_IEEE) &&
-		    adapter->ixgbe_ieee_pfc &&
-		    !(adapter->ixgbe_ieee_pfc->pfc_en)))
+
+	for (i = 0; i < adapter->num_tx_queues; i++)
+		clear_bit(__IXGBE_HANG_CHECK_ARMED,
+			  &adapter->tx_ring[i]->state);
+}
+
+static void ixgbe_update_xoff_received(struct ixgbe_adapter *adapter)
+{
+	struct ixgbe_hw *hw = &adapter->hw;
+	struct ixgbe_hw_stats *hwstats = &adapter->stats;
+	u32 xoff[8] = {0};
+	int i;
+	bool pfc_en = adapter->dcb_cfg.pfc_mode_enable;
+
+	if (adapter->ixgbe_ieee_pfc)
+		pfc_en |= !!(adapter->ixgbe_ieee_pfc->pfc_en);
+
+	if (!(adapter->flags & IXGBE_FLAG_DCB_ENABLED) || !pfc_en) {
+		ixgbe_update_xoff_rx_lfc(adapter);
 		return;
+	}
 
 	/* update stats for each tc, only valid with PFC enabled */
 	for (i = 0; i < MAX_TX_PACKET_BUFFERS; i++) {
@@ -4403,9 +4414,6 @@ static int __devinit ixgbe_sw_init(struct ixgbe_adapter *adapter)
 	/* default flow control settings */
 	hw->fc.requested_mode = ixgbe_fc_full;
 	hw->fc.current_mode = ixgbe_fc_full;	/* init for ethtool output */
-#ifdef CONFIG_DCB
-	adapter->last_lfc_mode = hw->fc.current_mode;
-#endif
 	ixgbe_pbthresh_setup(adapter);
 	hw->fc.pause_time = IXGBE_DEFAULT_FCPAUSE;
 	hw->fc.send_xon = true;
@@ -6542,15 +6550,17 @@ int ixgbe_setup_tc(struct net_device *dev, u8 tc)
 
 	if (tc) {
 		netdev_set_num_tc(dev, tc);
-		adapter->last_lfc_mode = adapter->hw.fc.current_mode;
 		adapter->flags |= IXGBE_FLAG_DCB_ENABLED;
 		adapter->flags &= ~IXGBE_FLAG_FDIR_HASH_CAPABLE;
 
-		if (adapter->hw.mac.type == ixgbe_mac_82598EB)
+		if (adapter->hw.mac.type == ixgbe_mac_82598EB) {
+			adapter->last_lfc_mode = adapter->hw.fc.requested_mode;
 			adapter->hw.fc.requested_mode = ixgbe_fc_none;
+		}
 	} else {
 		netdev_reset_tc(dev);
-		adapter->hw.fc.requested_mode = adapter->last_lfc_mode;
+		if (adapter->hw.mac.type == ixgbe_mac_82598EB)
+			adapter->hw.fc.requested_mode = adapter->last_lfc_mode;
 
 		adapter->flags &= ~IXGBE_FLAG_DCB_ENABLED;
 		adapter->flags |= IXGBE_FLAG_FDIR_HASH_CAPABLE;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
index 5e64c77..9559c03 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
@@ -2573,9 +2573,6 @@ enum ixgbe_fc_mode {
 	ixgbe_fc_rx_pause,
 	ixgbe_fc_tx_pause,
 	ixgbe_fc_full,
-#ifdef CONFIG_DCB
-	ixgbe_fc_pfc,
-#endif
 	ixgbe_fc_default
 };
 
-- 
1.7.7.6

[net-next 04/12] ixgbe: Set Drop_EN bit when multiple Rx queues are present w/o flow control

[Jeff Kirsher]

From: Alexander Duyck <alexander.h.duyck@intel.com>

The drop enable bit can be used to improve the performance of the adapter
in the case of multiple queues being present.  This performance gain is due
to the fact that some slower CPUs can cause the FIFO to backfill preventing
faster CPUs from receiving additional work.  By setting the drop enable bit
we prevent this and instead just drop the packets that would have been
bound for the slower CPU.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@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_dcb_nl.c |    5 ++
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c   |   59 ++++++++++++++++++++++-
 3 files changed, 64 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index 425f864..2262bf7 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -652,6 +652,7 @@ extern void ixgbe_atr_compute_perfect_hash_82599(union ixgbe_atr_input *input,
 						 union ixgbe_atr_input *mask);
 extern void ixgbe_set_rx_mode(struct net_device *netdev);
 #ifdef CONFIG_IXGBE_DCB
+extern void ixgbe_set_rx_drop_en(struct ixgbe_adapter *adapter);
 extern int ixgbe_setup_tc(struct net_device *dev, u8 tc);
 #endif
 extern void ixgbe_tx_ctxtdesc(struct ixgbe_ring *, u32, u32, u32, u32);
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
index 7f2fa69..5164a21 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c
@@ -395,6 +395,9 @@ static u8 ixgbe_dcbnl_set_all(struct net_device *netdev)
 		} else {
 			hw->mac.ops.fc_enable(hw);
 		}
+
+		ixgbe_set_rx_drop_en(adapter);
+
 		ret = DCB_HW_CHG;
 	}
 
@@ -641,6 +644,8 @@ static int ixgbe_dcbnl_ieee_setpfc(struct net_device *dev,
 	else
 		err = hw->mac.ops.fc_enable(hw);
 
+	ixgbe_set_rx_drop_en(adapter);
+
 	return err;
 }
 
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 0915e77..1693ec3 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -2771,6 +2771,61 @@ static void ixgbe_configure_tx(struct ixgbe_adapter *adapter)
 		ixgbe_configure_tx_ring(adapter, adapter->tx_ring[i]);
 }
 
+static void ixgbe_enable_rx_drop(struct ixgbe_adapter *adapter,
+				 struct ixgbe_ring *ring)
+{
+	struct ixgbe_hw *hw = &adapter->hw;
+	u8 reg_idx = ring->reg_idx;
+	u32 srrctl = IXGBE_READ_REG(hw, IXGBE_SRRCTL(reg_idx));
+
+	srrctl |= IXGBE_SRRCTL_DROP_EN;
+
+	IXGBE_WRITE_REG(hw, IXGBE_SRRCTL(reg_idx), srrctl);
+}
+
+static void ixgbe_disable_rx_drop(struct ixgbe_adapter *adapter,
+				  struct ixgbe_ring *ring)
+{
+	struct ixgbe_hw *hw = &adapter->hw;
+	u8 reg_idx = ring->reg_idx;
+	u32 srrctl = IXGBE_READ_REG(hw, IXGBE_SRRCTL(reg_idx));
+
+	srrctl &= ~IXGBE_SRRCTL_DROP_EN;
+
+	IXGBE_WRITE_REG(hw, IXGBE_SRRCTL(reg_idx), srrctl);
+}
+
+#ifdef CONFIG_IXGBE_DCB
+void ixgbe_set_rx_drop_en(struct ixgbe_adapter *adapter)
+#else
+static void ixgbe_set_rx_drop_en(struct ixgbe_adapter *adapter)
+#endif
+{
+	int i;
+	bool pfc_en = adapter->dcb_cfg.pfc_mode_enable;
+
+	if (adapter->ixgbe_ieee_pfc)
+		pfc_en |= !!(adapter->ixgbe_ieee_pfc->pfc_en);
+
+	/*
+	 * We should set the drop enable bit if:
+	 *  SR-IOV is enabled
+	 *   or
+	 *  Number of Rx queues > 1 and flow control is disabled
+	 *
+	 *  This allows us to avoid head of line blocking for security
+	 *  and performance reasons.
+	 */
+	if (adapter->num_vfs || (adapter->num_rx_queues > 1 &&
+	    !(adapter->hw.fc.current_mode & ixgbe_fc_tx_pause) && !pfc_en)) {
+		for (i = 0; i < adapter->num_rx_queues; i++)
+			ixgbe_enable_rx_drop(adapter, adapter->rx_ring[i]);
+	} else {
+		for (i = 0; i < adapter->num_rx_queues; i++)
+			ixgbe_disable_rx_drop(adapter, adapter->rx_ring[i]);
+	}
+}
+
 #define IXGBE_SRRCTL_BSIZEHDRSIZE_SHIFT 2
 
 static void ixgbe_configure_srrctl(struct ixgbe_adapter *adapter,
@@ -5276,8 +5331,10 @@ static void ixgbe_watchdog_update_link(struct ixgbe_adapter *adapter)
 	if (adapter->ixgbe_ieee_pfc)
 		pfc_en |= !!(adapter->ixgbe_ieee_pfc->pfc_en);
 
-	if (link_up && !((adapter->flags & IXGBE_FLAG_DCB_ENABLED) && pfc_en))
+	if (link_up && !((adapter->flags & IXGBE_FLAG_DCB_ENABLED) && pfc_en)) {
 		hw->mac.ops.fc_enable(hw);
+		ixgbe_set_rx_drop_en(adapter);
+	}
 
 	if (link_up ||
 	    time_after(jiffies, (adapter->link_check_timeout +
-- 
1.7.7.6

[net-next 05/12] ixgbe: Fix bogus error message

[Jeff Kirsher]

From: Greg Rose <gregory.v.rose@intel.com>

If the VF sends a MACVLAN request with index of zero then it is not
actually trying to add a filter.  Check the index value and only
indicate that operation is not allowed when the VF is actually trying
to add a filter.

Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c |    6 +++---
 1 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c
index b64f1fc..2d971d1 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c
@@ -640,14 +640,14 @@ static int ixgbe_rcv_msg_from_vf(struct ixgbe_adapter *adapter, u32 vf)
 		}
 		break;
 	case IXGBE_VF_SET_MACVLAN:
-		if (adapter->vfinfo[vf].pf_set_mac) {
+		index = (msgbuf[0] & IXGBE_VT_MSGINFO_MASK) >>
+			IXGBE_VT_MSGINFO_SHIFT;
+		if (adapter->vfinfo[vf].pf_set_mac && index > 0) {
 			e_warn(drv, "VF %d requested MACVLAN filter but is "
 				    "administratively denied\n", vf);
 			retval = -1;
 			break;
 		}
-		index = (msgbuf[0] & IXGBE_VT_MSGINFO_MASK) >>
-			IXGBE_VT_MSGINFO_SHIFT;
 		/*
 		 * If the VF is allowed to set MAC filters then turn off
 		 * anti-spoofing to avoid false positives.  An index
-- 
1.7.7.6

[net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Jeff Kirsher]

From: Jacob Keller <jacob.e.keller@intel.com>

This patch enables hardware timestamping for use with PTP software by
extracting a ns counter from an arbitrary fixed point cycles counter.
The hardware generates SYSTIME registers using the DMA tick which
changes based on the current link speed. These SYSTIME registers are
converted to ns using the cyclecounter and timecounter structures
provided by the kernel. Using the SO_TIMESTAMPING api, software can
enable and access timestamps for PTP packets.

The SO_TIMESTAMPING API has space for 3 different kinds of timestamps,
SYS, RAW, and SOF. SYS hardware timestamps are hardware ns values that
are then scaled to the software clock. RAW hardware timestamps are the
direct raw value of the ns counter. SOF software timestamps are the
software timestamp calculated as close as possible to the software
transmit, but are not offloaded to the hardware. This patch only
supports the RAW hardware timestamps due to inefficiency of the SYS
design.

This patch also enables the PHC subsystem features for atomically
adjusting the cycle register, and adjusting the clock frequency in
parts per billion. This frequency adjustment works by slightly
adjusting the value added to the cycle registers each DMA tick. This
causes the hardware registers to overflow rapidly (approximately once
every 34 seconds, when at 10gig link). To solve this, the timecounter
structure is used, along with a timer set for every 25 seconds. This
allows for detecting register overflow and converting the cycle
counter registers into ns values needed for providing useful
timestamps to the network stack.

Only the basic required clock functions are supported at this time,
although the hardware supports some ancillary features and these could
easily be enabled in the future.

Note that use of this hardware timestamping requires modifying daemon
software to use the SO_TIMESTAMPING API for timestamps, and the
ptp_clock PHC framework for accessing the clock. The timestamps have
no relation to the system time at all, so software must use the posix
clock generated by the PHC framework instead.

Signed-off-by: Jacob E Keller <jacob.e.keller@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/Kconfig            |   11 +
 drivers/net/ethernet/intel/ixgbe/Makefile     |    2 +
 drivers/net/ethernet/intel/ixgbe/ixgbe.h      |   32 ++
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |   53 ++-
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c  |  732 +++++++++++++++++++++++++
 drivers/net/ethernet/intel/ixgbe/ixgbe_type.h |   32 ++
 6 files changed, 861 insertions(+), 1 deletions(-)
 create mode 100644 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c

diff --git a/drivers/net/ethernet/intel/Kconfig b/drivers/net/ethernet/intel/Kconfig
index 546efe3..79b07ec 100644
--- a/drivers/net/ethernet/intel/Kconfig
+++ b/drivers/net/ethernet/intel/Kconfig
@@ -220,6 +220,17 @@ config IXGBE_DCB
 
 	  If unsure, say N.
 
+config IXGBE_PTP
+	bool "PTP Clock Support"
+	default n
+	depends on IXGBE && PTP_1588_CLOCK
+	---help---
+	  Say Y here if you want support for 1588 Timestamping with a
+	  PHC device, using the PTP 1588 Clock support. This is
+	  required to enable timestamping support for the device.
+
+	  If unsure, say N.
+
 config IXGBEVF
 	tristate "Intel(R) 82599 Virtual Function Ethernet support"
 	depends on PCI_MSI
diff --git a/drivers/net/ethernet/intel/ixgbe/Makefile b/drivers/net/ethernet/intel/ixgbe/Makefile
index 0708d7e..0bdf06b 100644
--- a/drivers/net/ethernet/intel/ixgbe/Makefile
+++ b/drivers/net/ethernet/intel/ixgbe/Makefile
@@ -39,4 +39,6 @@ ixgbe-objs := ixgbe_main.o ixgbe_common.o ixgbe_ethtool.o \
 ixgbe-$(CONFIG_IXGBE_DCB) +=  ixgbe_dcb.o ixgbe_dcb_82598.o \
                               ixgbe_dcb_82599.o ixgbe_dcb_nl.o
 
+ixgbe-$(CONFIG_IXGBE_PTP) += ixgbe_ptp.o
+
 ixgbe-$(CONFIG_FCOE:m=y) += ixgbe_fcoe.o
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index 2262bf7..c90fbd2 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -36,6 +36,12 @@
 #include <linux/aer.h>
 #include <linux/if_vlan.h>
 
+#ifdef CONFIG_IXGBE_PTP
+#include <linux/clocksource.h>
+#include <linux/net_tstamp.h>
+#include <linux/ptp_clock_kernel.h>
+#endif /* CONFIG_IXGBE_PTP */
+
 #include "ixgbe_type.h"
 #include "ixgbe_common.h"
 #include "ixgbe_dcb.h"
@@ -96,6 +102,7 @@
 #define IXGBE_TX_FLAGS_FCOE		(u32)(1 << 5)
 #define IXGBE_TX_FLAGS_FSO		(u32)(1 << 6)
 #define IXGBE_TX_FLAGS_TXSW		(u32)(1 << 7)
+#define IXGBE_TX_FLAGS_TSTAMP		(u32)(1 << 8)
 #define IXGBE_TX_FLAGS_VLAN_MASK	0xffff0000
 #define IXGBE_TX_FLAGS_VLAN_PRIO_MASK	0xe0000000
 #define IXGBE_TX_FLAGS_VLAN_PRIO_SHIFT  29
@@ -458,6 +465,7 @@ struct ixgbe_adapter {
 #define IXGBE_FLAG2_FDIR_REQUIRES_REINIT        (u32)(1 << 7)
 #define IXGBE_FLAG2_RSS_FIELD_IPV4_UDP		(u32)(1 << 8)
 #define IXGBE_FLAG2_RSS_FIELD_IPV6_UDP		(u32)(1 << 9)
+#define IXGBE_FLAG2_OVERFLOW_CHECK_ENABLED	(u32)(1 << 10)
 
 	/* Tx fast path data */
 	int num_tx_queues;
@@ -545,6 +553,17 @@ struct ixgbe_adapter {
 	u32 interrupt_event;
 	u32 led_reg;
 
+#ifdef CONFIG_IXGBE_PTP
+	struct ptp_clock *ptp_clock;
+	struct ptp_clock_info ptp_caps;
+	unsigned long last_overflow_check;
+	spinlock_t tmreg_lock;
+	struct cyclecounter cc;
+	struct timecounter tc;
+	u32 base_incval;
+	u32 cycle_speed;
+#endif /* CONFIG_IXGBE_PTP */
+
 	/* SR-IOV */
 	DECLARE_BITMAP(active_vfs, IXGBE_MAX_VF_FUNCTIONS);
 	unsigned int num_vfs;
@@ -689,4 +708,17 @@ static inline struct netdev_queue *txring_txq(const struct ixgbe_ring *ring)
 	return netdev_get_tx_queue(ring->netdev, ring->queue_index);
 }
 
+#ifdef CONFIG_IXGBE_PTP
+extern void ixgbe_ptp_init(struct ixgbe_adapter *adapter);
+extern void ixgbe_ptp_stop(struct ixgbe_adapter *adapter);
+extern void ixgbe_ptp_overflow_check(struct ixgbe_adapter *adapter);
+extern void ixgbe_ptp_tx_hwtstamp(struct ixgbe_q_vector *q_vector,
+				  struct sk_buff *skb);
+extern void ixgbe_ptp_rx_hwtstamp(struct ixgbe_q_vector *q_vector,
+				  struct sk_buff *skb);
+extern int ixgbe_ptp_hwtstamp_ioctl(struct ixgbe_adapter *adapter,
+				    struct ifreq *ifr, int cmd);
+extern void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter);
+#endif /* CONFIG_IXGBE_PTP */
+
 #endif /* _IXGBE_H_ */
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 1693ec3..9a83c40 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -789,6 +789,13 @@ static bool ixgbe_clean_tx_irq(struct ixgbe_q_vector *q_vector,
 		total_bytes += tx_buffer->bytecount;
 		total_packets += tx_buffer->gso_segs;
 
+#ifdef CONFIG_IXGBE_PTP
+		if (unlikely(tx_buffer->tx_flags &
+			     IXGBE_TX_FLAGS_TSTAMP))
+			ixgbe_ptp_tx_hwtstamp(q_vector,
+					      tx_buffer->skb);
+
+#endif
 		/* free the skb */
 		dev_kfree_skb_any(tx_buffer->skb);
 
@@ -1389,6 +1396,11 @@ static void ixgbe_process_skb_fields(struct ixgbe_ring *rx_ring,
 
 	ixgbe_rx_checksum(rx_ring, rx_desc, skb);
 
+#ifdef CONFIG_IXGBE_PTP
+	if (ixgbe_test_staterr(rx_desc, IXGBE_RXDADV_STAT_TS))
+		ixgbe_ptp_rx_hwtstamp(rx_ring->q_vector, skb);
+#endif
+
 	if (ixgbe_test_staterr(rx_desc, IXGBE_RXD_STAT_VP)) {
 		u16 vid = le16_to_cpu(rx_desc->wb.upper.vlan);
 		__vlan_hwaccel_put_tag(skb, vid);
@@ -5387,6 +5399,11 @@ static void ixgbe_watchdog_link_is_up(struct ixgbe_adapter *adapter)
 		flow_rx = false;
 		break;
 	}
+
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_start_cyclecounter(adapter);
+#endif
+
 	e_info(drv, "NIC Link is Up %s, Flow Control: %s\n",
 	       (link_speed == IXGBE_LINK_SPEED_10GB_FULL ?
 	       "10 Gbps" :
@@ -5424,6 +5441,10 @@ static void ixgbe_watchdog_link_is_down(struct ixgbe_adapter *adapter)
 	if (ixgbe_is_sfp(hw) && hw->mac.type == ixgbe_mac_82598EB)
 		adapter->flags2 |= IXGBE_FLAG2_SEARCH_FOR_SFP;
 
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_start_cyclecounter(adapter);
+#endif
+
 	e_info(drv, "NIC Link is Down\n");
 	netif_carrier_off(netdev);
 }
@@ -5723,6 +5744,9 @@ static void ixgbe_service_task(struct work_struct *work)
 	ixgbe_watchdog_subtask(adapter);
 	ixgbe_fdir_reinit_subtask(adapter);
 	ixgbe_check_hang_subtask(adapter);
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_overflow_check(adapter);
+#endif
 
 	ixgbe_service_event_complete(adapter);
 }
@@ -5873,6 +5897,11 @@ static __le32 ixgbe_tx_cmd_type(u32 tx_flags)
 	if (tx_flags & IXGBE_TX_FLAGS_HW_VLAN)
 		cmd_type |= cpu_to_le32(IXGBE_ADVTXD_DCMD_VLE);
 
+#ifdef CONFIG_IXGBE_PTP
+	if (tx_flags & IXGBE_TX_FLAGS_TSTAMP)
+		cmd_type |= cpu_to_le32(IXGBE_ADVTXD_MAC_TSTAMP);
+#endif
+
 	/* set segmentation enable bits for TSO/FSO */
 #ifdef IXGBE_FCOE
 	if (tx_flags & (IXGBE_TX_FLAGS_TSO | IXGBE_TX_FLAGS_FSO))
@@ -6263,6 +6292,13 @@ netdev_tx_t ixgbe_xmit_frame_ring(struct sk_buff *skb,
 		tx_flags |= IXGBE_TX_FLAGS_SW_VLAN;
 	}
 
+#ifdef CONFIG_IXGBE_PTP
+	if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP)) {
+		skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
+		tx_flags |= IXGBE_TX_FLAGS_TSTAMP;
+	}
+#endif
+
 #ifdef CONFIG_PCI_IOV
 	/*
 	 * Use the l2switch_enable flag - would be false if the DMA
@@ -6415,7 +6451,14 @@ static int ixgbe_ioctl(struct net_device *netdev, struct ifreq *req, int cmd)
 {
 	struct ixgbe_adapter *adapter = netdev_priv(netdev);
 
-	return mdio_mii_ioctl(&adapter->hw.phy.mdio, if_mii(req), cmd);
+	switch (cmd) {
+#ifdef CONFIG_IXGBE_PTP
+	case SIOCSHWTSTAMP:
+		return ixgbe_ptp_hwtstamp_ioctl(adapter, req, cmd);
+#endif
+	default:
+		return mdio_mii_ioctl(&adapter->hw.phy.mdio, if_mii(req), cmd);
+	}
 }
 
 /**
@@ -7202,6 +7245,10 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev,
 
 	device_set_wakeup_enable(&adapter->pdev->dev, adapter->wol);
 
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_init(adapter);
+#endif /* CONFIG_IXGBE_PTP*/
+
 	/* save off EEPROM version number */
 	hw->eeprom.ops.read(hw, 0x2e, &adapter->eeprom_verh);
 	hw->eeprom.ops.read(hw, 0x2d, &adapter->eeprom_verl);
@@ -7330,6 +7377,10 @@ static void __devexit ixgbe_remove(struct pci_dev *pdev)
 	set_bit(__IXGBE_DOWN, &adapter->state);
 	cancel_work_sync(&adapter->service_task);
 
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_stop(adapter);
+#endif
+
 #ifdef CONFIG_IXGBE_DCA
 	if (adapter->flags & IXGBE_FLAG_DCA_ENABLED) {
 		adapter->flags &= ~IXGBE_FLAG_DCA_ENABLED;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
new file mode 100644
index 0000000..0b6553e
--- /dev/null
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
@@ -0,0 +1,732 @@
+/*******************************************************************************
+
+  Intel 10 Gigabit PCI Express Linux driver
+  Copyright(c) 1999 - 2012 Intel Corporation.
+
+  This program is free software; you can redistribute it and/or modify it
+  under the terms and conditions of the GNU General Public License,
+  version 2, as published by the Free Software Foundation.
+
+  This program is distributed in the hope it will be useful, but WITHOUT
+  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+  more details.
+
+  You should have received a copy of the GNU General Public License along with
+  this program; if not, write to the Free Software Foundation, Inc.,
+  51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+
+  The full GNU General Public License is included in this distribution in
+  the file called "COPYING".
+
+  Contact Information:
+  e1000-devel Mailing List <e1000-devel@lists.sourceforge.net>
+  Intel Corporation, 5200 N.E. Elam Young Parkway, Hillsboro, OR 97124-6497
+
+*******************************************************************************/
+#include "ixgbe.h"
+#include <linux/export.h>
+
+/*
+ * The 82599 and the X540 do not have true 64bit nanosecond scale
+ * counter registers. Instead, SYSTIME is defined by a fixed point
+ * system which allows the user to define the scale counter increment
+ * value at every level change of the oscillator driving the SYSTIME
+ * value. For both devices the TIMINCA:IV field defines this
+ * increment. On the X540 device, 31 bits are provided. However on the
+ * 82599 only provides 24 bits. The time unit is determined by the
+ * clock frequency of the oscillator in combination with the TIMINCA
+ * register. When these devices link at 10Gb the oscillator has a
+ * period of 6.4ns. In order to convert the scale counter into
+ * nanoseconds the cyclecounter and timecounter structures are
+ * used. The SYSTIME registers need to be converted to ns values by use
+ * of only a right shift (division by power of 2). The following math
+ * determines the largest incvalue that will fit into the available
+ * bits in the TIMINCA register.
+ *
+ * PeriodWidth: Number of bits to store the clock period
+ * MaxWidth: The maximum width value of the TIMINCA register
+ * Period: The clock period for the oscillator
+ * round(): discard the fractional portion of the calculation
+ *
+ * Period * [ 2 ^ ( MaxWidth - PeriodWidth ) ]
+ *
+ * For the X540, MaxWidth is 31 bits, and the base period is 6.4 ns
+ * For the 82599, MaxWidth is 24 bits, and the base period is 6.4 ns
+ *
+ * The period also changes based on the link speed:
+ * At 10Gb link or no link, the period remains the same.
+ * At 1Gb link, the period is multiplied by 10. (64ns)
+ * At 100Mb link, the period is multiplied by 100. (640ns)
+ *
+ * The calculated value allows us to right shift the SYSTIME register
+ * value in order to quickly convert it into a nanosecond clock,
+ * while allowing for the maximum possible adjustment value.
+ *
+ * These diagrams are only for the 10Gb link period
+ *
+ *           SYSTIMEH            SYSTIMEL
+ *       +--------------+  +--------------+
+ * X540  |      32      |  | 1 | 3 |  28  |
+ *       *--------------+  +--------------+
+ *        \________ 36 bits ______/  fract
+ *
+ *       +--------------+  +--------------+
+ * 82599 |      32      |  | 8 | 3 |  21  |
+ *       *--------------+  +--------------+
+ *        \________ 43 bits ______/  fract
+ *
+ * The 36 bit X540 SYSTIME overflows every
+ *   2^36 * 10^-9 / 60 = 1.14 minutes or 69 seconds
+ *
+ * The 43 bit 82599 SYSTIME overflows every
+ *   2^43 * 10^-9 / 3600 = 2.4 hours
+ */
+#define IXGBE_INCVAL_10GB 0x66666666
+#define IXGBE_INCVAL_1GB  0x40000000
+#define IXGBE_INCVAL_100  0x50000000
+
+#define IXGBE_INCVAL_SHIFT_10GB  28
+#define IXGBE_INCVAL_SHIFT_1GB   24
+#define IXGBE_INCVAL_SHIFT_100   21
+
+#define IXGBE_INCVAL_SHIFT_82599 7
+#define IXGBE_INCPER_SHIFT_82599 24
+#define IXGBE_MAX_TIMEADJ_VALUE  0x7FFFFFFFFFFFFFFFULL
+
+#define IXGBE_OVERFLOW_PERIOD    (HZ * 30)
+
+/**
+ * ixgbe_ptp_read - read raw cycle counter (to be used by time counter)
+ * @cc - the cyclecounter structure
+ *
+ * this function reads the cyclecounter registers and is called by the
+ * cyclecounter structure used to construct a ns counter from the
+ * arbitrary fixed point registers
+ */
+static cycle_t ixgbe_ptp_read(const struct cyclecounter *cc)
+{
+	struct ixgbe_adapter *adapter =
+		container_of(cc, struct ixgbe_adapter, cc);
+	struct ixgbe_hw *hw = &adapter->hw;
+	u64 stamp = 0;
+
+	stamp |= (u64)IXGBE_READ_REG(hw, IXGBE_SYSTIML);
+	stamp |= (u64)IXGBE_READ_REG(hw, IXGBE_SYSTIMH) << 32;
+
+	return stamp;
+}
+
+/**
+ * ixgbe_ptp_adjfreq
+ * @ptp - the ptp clock structure
+ * @ppb - parts per billion adjustment from base
+ *
+ * adjust the frequency of the ptp cycle counter by the
+ * indicated ppb from the base frequency.
+ */
+static int ixgbe_ptp_adjfreq(struct ptp_clock_info *ptp, s32 ppb)
+{
+	struct ixgbe_adapter *adapter =
+		container_of(ptp, struct ixgbe_adapter, ptp_caps);
+	struct ixgbe_hw *hw = &adapter->hw;
+	u64 freq;
+	u32 diff, incval;
+	int neg_adj = 0;
+
+	if (ppb < 0) {
+		neg_adj = 1;
+		ppb = -ppb;
+	}
+
+	smp_mb();
+	incval = ACCESS_ONCE(adapter->base_incval);
+
+	freq = incval;
+	freq *= ppb;
+	diff = div_u64(freq, 1000000000ULL);
+
+	incval = neg_adj ? (incval - diff) : (incval + diff);
+
+	switch (hw->mac.type) {
+	case ixgbe_mac_X540:
+		IXGBE_WRITE_REG(hw, IXGBE_TIMINCA, incval);
+		break;
+	case ixgbe_mac_82599EB:
+		IXGBE_WRITE_REG(hw, IXGBE_TIMINCA,
+				(1 << IXGBE_INCPER_SHIFT_82599) |
+				incval);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+/**
+ * ixgbe_ptp_adjtime
+ * @ptp - the ptp clock structure
+ * @delta - offset to adjust the cycle counter by
+ *
+ * adjust the timer by resetting the timecounter structure.
+ */
+static int ixgbe_ptp_adjtime(struct ptp_clock_info *ptp, s64 delta)
+{
+	struct ixgbe_adapter *adapter =
+		container_of(ptp, struct ixgbe_adapter, ptp_caps);
+	unsigned long flags;
+	u64 now;
+
+	spin_lock_irqsave(&adapter->tmreg_lock, flags);
+
+	now = timecounter_read(&adapter->tc);
+	now += delta;
+
+	/* reset the timecounter */
+	timecounter_init(&adapter->tc,
+			 &adapter->cc,
+			 now);
+
+	spin_unlock_irqrestore(&adapter->tmreg_lock, flags);
+	return 0;
+}
+
+/**
+ * ixgbe_ptp_gettime
+ * @ptp - the ptp clock structure
+ * @ts - timespec structure to hold the current time value
+ *
+ * read the timecounter and return the correct value on ns,
+ * after converting it into a struct timespec.
+ */
+static int ixgbe_ptp_gettime(struct ptp_clock_info *ptp, struct timespec *ts)
+{
+	struct ixgbe_adapter *adapter =
+		container_of(ptp, struct ixgbe_adapter, ptp_caps);
+	u64 ns;
+	u32 remainder;
+	unsigned long flags;
+
+	spin_lock_irqsave(&adapter->tmreg_lock, flags);
+	ns = timecounter_read(&adapter->tc);
+	spin_unlock_irqrestore(&adapter->tmreg_lock, flags);
+
+	ts->tv_sec = div_u64_rem(ns, 1000000000ULL, &remainder);
+	ts->tv_nsec = remainder;
+
+	return 0;
+}
+
+/**
+ * ixgbe_ptp_settime
+ * @ptp - the ptp clock structure
+ * @ts - the timespec containing the new time for the cycle counter
+ *
+ * reset the timecounter to use a new base value instead of the kernel
+ * wall timer value.
+ */
+static int ixgbe_ptp_settime(struct ptp_clock_info *ptp,
+			     const struct timespec *ts)
+{
+	struct ixgbe_adapter *adapter =
+		container_of(ptp, struct ixgbe_adapter, ptp_caps);
+	u64 ns;
+	unsigned long flags;
+
+	ns = ts->tv_sec * 1000000000ULL;
+	ns += ts->tv_nsec;
+
+	/* reset the timecounter */
+	spin_lock_irqsave(&adapter->tmreg_lock, flags);
+	timecounter_init(&adapter->tc, &adapter->cc, ns);
+	spin_unlock_irqrestore(&adapter->tmreg_lock, flags);
+
+	return 0;
+}
+
+/**
+ * ixgbe_ptp_enable
+ * @ptp - the ptp clock structure
+ * @rq - the requested feature to change
+ * @on - whether to enable or disable the feature
+ *
+ * enable (or disable) ancillary features of the phc subsystem.
+ * our driver does not support any of these features
+ */
+static int ixgbe_ptp_enable(struct ptp_clock_info *ptp,
+			    struct ptp_clock_request *rq, int on)
+{
+	return -ENOTSUPP;
+}
+
+/**
+ * ixgbe_ptp_overflow_check - delayed work to detect SYSTIME overflow
+ * @work: structure containing information about this work task
+ *
+ * this work function is scheduled to continue reading the timecounter
+ * in order to prevent missing when the system time registers wrap
+ * around. This needs to be run approximately twice a minute when no
+ * PTP activity is occurring.
+ */
+void ixgbe_ptp_overflow_check(struct ixgbe_adapter *adapter)
+{
+	unsigned long elapsed_jiffies = adapter->last_overflow_check - jiffies;
+	struct timespec ts;
+
+	if ((adapter->flags2 & IXGBE_FLAG2_OVERFLOW_CHECK_ENABLED) &&
+	    (elapsed_jiffies >= IXGBE_OVERFLOW_PERIOD)) {
+		ixgbe_ptp_gettime(&adapter->ptp_caps, &ts);
+		adapter->last_overflow_check = jiffies;
+	}
+}
+
+/**
+ * ixgbe_ptp_tx_hwtstamp - utility function which checks for TX time stamp
+ * @q_vector: structure containing interrupt and ring information
+ * @skb: particular skb to send timestamp with
+ *
+ * if the timestamp is valid, we convert it into the timecounter ns
+ * value, then store that result into the shhwtstamps structure which
+ * is passed up the network stack
+ */
+void ixgbe_ptp_tx_hwtstamp(struct ixgbe_q_vector *q_vector,
+			   struct sk_buff *skb)
+{
+	struct ixgbe_adapter *adapter;
+	struct ixgbe_hw *hw;
+	struct skb_shared_hwtstamps shhwtstamps;
+	u64 regval = 0, ns;
+	u32 tsynctxctl;
+	unsigned long flags;
+
+	/* we cannot process timestamps on a ring without a q_vector */
+	if (!q_vector || !q_vector->adapter)
+		return;
+
+	adapter = q_vector->adapter;
+	hw = &adapter->hw;
+
+	tsynctxctl = IXGBE_READ_REG(hw, IXGBE_TSYNCTXCTL);
+	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_TXSTMPL);
+	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_TXSTMPH) << 32;
+
+	/*
+	 * if TX timestamp is not valid, exit after clearing the
+	 * timestamp registers
+	 */
+	if (!(tsynctxctl & IXGBE_TSYNCTXCTL_VALID))
+		return;
+
+	spin_lock_irqsave(&adapter->tmreg_lock, flags);
+	ns = timecounter_cyc2time(&adapter->tc, regval);
+	spin_unlock_irqrestore(&adapter->tmreg_lock, flags);
+
+	memset(&shhwtstamps, 0, sizeof(shhwtstamps));
+	shhwtstamps.hwtstamp = ns_to_ktime(ns);
+	skb_tstamp_tx(skb, &shhwtstamps);
+}
+
+/**
+ * ixgbe_ptp_rx_hwtstamp - utility function which checks for RX time stamp
+ * @q_vector: structure containing interrupt and ring information
+ * @skb: particular skb to send timestamp with
+ *
+ * if the timestamp is valid, we convert it into the timecounter ns
+ * value, then store that result into the shhwtstamps structure which
+ * is passed up the network stack
+ */
+void ixgbe_ptp_rx_hwtstamp(struct ixgbe_q_vector *q_vector,
+			   struct sk_buff *skb)
+{
+	struct ixgbe_adapter *adapter;
+	struct ixgbe_hw *hw;
+	struct skb_shared_hwtstamps *shhwtstamps;
+	u64 regval = 0, ns;
+	u32 tsyncrxctl;
+	unsigned long flags;
+
+	/* we cannot process timestamps on a ring without a q_vector */
+	if (!q_vector || !q_vector->adapter)
+		return;
+
+	adapter = q_vector->adapter;
+	hw = &adapter->hw;
+
+	tsyncrxctl = IXGBE_READ_REG(hw, IXGBE_TSYNCRXCTL);
+	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_RXSTMPL);
+	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_RXSTMPH) << 32;
+
+	/*
+	 * If this bit is set, then the RX registers contain the time stamp. No
+	 * other packet will be time stamped until we read these registers, so
+	 * read the registers to make them available again. Because only one
+	 * packet can be time stamped at a time, we know that the register
+	 * values must belong to this one here and therefore we don't need to
+	 * compare any of the additional attributes stored for it.
+	 *
+	 * If nothing went wrong, then it should have a skb_shared_tx that we
+	 * can turn into a skb_shared_hwtstamps.
+	 */
+	if (!(tsyncrxctl & IXGBE_TSYNCRXCTL_VALID))
+		return;
+
+	spin_lock_irqsave(&adapter->tmreg_lock, flags);
+	ns = timecounter_cyc2time(&adapter->tc, regval);
+	spin_unlock_irqrestore(&adapter->tmreg_lock, flags);
+
+	shhwtstamps = skb_hwtstamps(skb);
+	shhwtstamps->hwtstamp = ns_to_ktime(ns);
+}
+
+/**
+ * ixgbe_ptp_hwtstamp_ioctl - control hardware time stamping
+ * @adapter: pointer to adapter struct
+ * @ifreq: ioctl data
+ * @cmd: particular ioctl requested
+ *
+ * Outgoing time stamping can be enabled and disabled. Play nice and
+ * disable it when requested, although it shouldn't case any overhead
+ * when no packet needs it. At most one packet in the queue may be
+ * marked for time stamping, otherwise it would be impossible to tell
+ * for sure to which packet the hardware time stamp belongs.
+ *
+ * Incoming time stamping has to be configured via the hardware
+ * filters. Not all combinations are supported, in particular event
+ * type has to be specified. Matching the kind of event packet is
+ * not supported, with the exception of "all V2 events regardless of
+ * level 2 or 4".
+ */
+int ixgbe_ptp_hwtstamp_ioctl(struct ixgbe_adapter *adapter,
+			     struct ifreq *ifr, int cmd)
+{
+	struct ixgbe_hw *hw = &adapter->hw;
+	struct hwtstamp_config config;
+	u32 tsync_tx_ctl = IXGBE_TSYNCTXCTL_ENABLED;
+	u32 tsync_rx_ctl = IXGBE_TSYNCRXCTL_ENABLED;
+	u32 tsync_rx_mtrl = 0;
+	bool is_l4 = false;
+	bool is_l2 = false;
+	u32 regval;
+
+	if (copy_from_user(&config, ifr->ifr_data, sizeof(config)))
+		return -EFAULT;
+
+	/* reserved for future extensions */
+	if (config.flags)
+		return -EINVAL;
+
+	switch (config.tx_type) {
+	case HWTSTAMP_TX_OFF:
+		tsync_tx_ctl = 0;
+	case HWTSTAMP_TX_ON:
+		break;
+	default:
+		return -ERANGE;
+	}
+
+	switch (config.rx_filter) {
+	case HWTSTAMP_FILTER_NONE:
+		tsync_rx_ctl = 0;
+		break;
+	case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
+		tsync_rx_ctl |= IXGBE_TSYNCRXCTL_TYPE_L4_V1;
+		tsync_rx_mtrl = IXGBE_RXMTRL_V1_SYNC_MSG;
+		is_l4 = true;
+		break;
+	case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
+		tsync_rx_ctl |= IXGBE_TSYNCRXCTL_TYPE_L4_V1;
+		tsync_rx_mtrl = IXGBE_RXMTRL_V1_DELAY_REQ_MSG;
+		is_l4 = true;
+		break;
+	case HWTSTAMP_FILTER_PTP_V2_SYNC:
+	case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
+	case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
+		tsync_rx_ctl |= IXGBE_TSYNCRXCTL_TYPE_L2_L4_V2;
+		tsync_rx_mtrl = IXGBE_RXMTRL_V2_SYNC_MSG;
+		is_l2 = true;
+		is_l4 = true;
+		config.rx_filter = HWTSTAMP_FILTER_SOME;
+		break;
+	case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
+	case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
+	case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
+		tsync_rx_ctl |= IXGBE_TSYNCRXCTL_TYPE_L2_L4_V2;
+		tsync_rx_mtrl = IXGBE_RXMTRL_V2_DELAY_REQ_MSG;
+		is_l2 = true;
+		is_l4 = true;
+		config.rx_filter = HWTSTAMP_FILTER_SOME;
+		break;
+	case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
+	case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
+	case HWTSTAMP_FILTER_PTP_V2_EVENT:
+		tsync_rx_ctl |= IXGBE_TSYNCRXCTL_TYPE_EVENT_V2;
+		config.rx_filter = HWTSTAMP_FILTER_PTP_V2_EVENT;
+		is_l2 = true;
+		is_l4 = true;
+		break;
+	case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
+	case HWTSTAMP_FILTER_ALL:
+	default:
+		/*
+		 * register RXMTRL must be set, therefore it is not
+		 * possible to time stamp both V1 Sync and Delay_Req messages
+		 * and hardware does not support timestamping all packets
+		 * => return error
+		 */
+		return -ERANGE;
+	}
+
+	if (hw->mac.type == ixgbe_mac_82598EB) {
+		if (tsync_rx_ctl | tsync_tx_ctl)
+			return -ERANGE;
+		return 0;
+	}
+
+	/* define ethertype filter for timestamped packets */
+	if (is_l2)
+		IXGBE_WRITE_REG(hw, IXGBE_ETQF(3),
+				(IXGBE_ETQF_FILTER_EN | /* enable filter */
+				 IXGBE_ETQF_1588 | /* enable timestamping */
+				 ETH_P_1588));     /* 1588 eth protocol type */
+	else
+		IXGBE_WRITE_REG(hw, IXGBE_ETQF(3), 0);
+
+#define PTP_PORT 319
+	/* L4 Queue Filter[3]: filter by destination port and protocol */
+	if (is_l4) {
+		u32 ftqf = (IXGBE_FTQF_PROTOCOL_UDP /* UDP */
+			    | IXGBE_FTQF_POOL_MASK_EN /* Pool not compared */
+			    | IXGBE_FTQF_QUEUE_ENABLE);
+
+		ftqf |= ((IXGBE_FTQF_PROTOCOL_COMP_MASK /* protocol check */
+			  & IXGBE_FTQF_DEST_PORT_MASK /* dest check */
+			  & IXGBE_FTQF_SOURCE_PORT_MASK) /* source check */
+			 << IXGBE_FTQF_5TUPLE_MASK_SHIFT);
+
+		IXGBE_WRITE_REG(hw, IXGBE_L34T_IMIR(3),
+				(3 << IXGBE_IMIR_RX_QUEUE_SHIFT_82599 |
+				 IXGBE_IMIR_SIZE_BP_82599));
+
+		/* enable port check */
+		IXGBE_WRITE_REG(hw, IXGBE_SDPQF(3),
+				(htons(PTP_PORT) |
+				 htons(PTP_PORT) << 16));
+
+		IXGBE_WRITE_REG(hw, IXGBE_FTQF(3), ftqf);
+
+		tsync_rx_mtrl |= PTP_PORT << 16;
+	} else {
+		IXGBE_WRITE_REG(hw, IXGBE_FTQF(3), 0);
+	}
+
+	/* enable/disable TX */
+	regval = IXGBE_READ_REG(hw, IXGBE_TSYNCTXCTL);
+	regval &= ~IXGBE_TSYNCTXCTL_ENABLED;
+	regval |= tsync_tx_ctl;
+	IXGBE_WRITE_REG(hw, IXGBE_TSYNCTXCTL, regval);
+
+	/* enable/disable RX */
+	regval = IXGBE_READ_REG(hw, IXGBE_TSYNCRXCTL);
+	regval &= ~(IXGBE_TSYNCRXCTL_ENABLED | IXGBE_TSYNCRXCTL_TYPE_MASK);
+	regval |= tsync_rx_ctl;
+	IXGBE_WRITE_REG(hw, IXGBE_TSYNCRXCTL, regval);
+
+	/* define which PTP packets are time stamped */
+	IXGBE_WRITE_REG(hw, IXGBE_RXMTRL, tsync_rx_mtrl);
+
+	IXGBE_WRITE_FLUSH(hw);
+
+	/* clear TX/RX time stamp registers, just to be sure */
+	regval = IXGBE_READ_REG(hw, IXGBE_TXSTMPH);
+	regval = IXGBE_READ_REG(hw, IXGBE_RXSTMPH);
+
+	return copy_to_user(ifr->ifr_data, &config, sizeof(config)) ?
+		-EFAULT : 0;
+}
+
+/**
+ * ixgbe_ptp_start_cyclecounter - create the cycle counter from hw
+ * @adapter - pointer to the adapter structure
+ *
+ * this function initializes the timecounter and cyclecounter
+ * structures for use in generated a ns counter from the arbitrary
+ * fixed point cycles registers in the hardware.
+ *
+ * A change in link speed impacts the frequency of the DMA clock on
+ * the device, which is used to generate the cycle counter
+ * registers. Therefor this function is called whenever the link speed
+ * changes.
+ */
+void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter)
+{
+	struct ixgbe_hw *hw = &adapter->hw;
+	u32 incval = 0;
+	u32 shift = 0;
+	u32 cycle_speed;
+	unsigned long flags;
+
+	/**
+	 * Determine what speed we need to set the cyclecounter
+	 * for. It should be different for 100Mb, 1Gb, and 10Gb. Treat
+	 * unknown speeds as 10Gb. (Hence why we can't just copy the
+	 * link_speed.
+	 */
+	switch (adapter->link_speed) {
+	case IXGBE_LINK_SPEED_100_FULL:
+	case IXGBE_LINK_SPEED_1GB_FULL:
+	case IXGBE_LINK_SPEED_10GB_FULL:
+		cycle_speed = adapter->link_speed;
+		break;
+	default:
+		/* cycle speed should be 10Gb when there is no link */
+		cycle_speed = IXGBE_LINK_SPEED_10GB_FULL;
+		break;
+	}
+
+	/* Bail if the cycle speed didn't change */
+	if (adapter->cycle_speed == cycle_speed)
+		return;
+
+	/**
+	 * Scale the NIC cycle counter by a large factor so that
+	 * relatively small corrections to the frequency can be added
+	 * or subtracted. The drawbacks of a large factor include
+	 * (a) the clock register overflows more quickly, (b) the cycle
+	 * counter structure must be able to convert the systime value
+	 * to nanoseconds using only a multiplier and a right-shift,
+	 * and (c) the value must fit within the timinca register space
+	 * => math based on internal DMA clock rate and available bits
+	 */
+	switch (cycle_speed) {
+	case IXGBE_LINK_SPEED_100_FULL:
+		incval = IXGBE_INCVAL_100;
+		shift = IXGBE_INCVAL_SHIFT_100;
+		break;
+	case IXGBE_LINK_SPEED_1GB_FULL:
+		incval = IXGBE_INCVAL_1GB;
+		shift = IXGBE_INCVAL_SHIFT_1GB;
+		break;
+	case IXGBE_LINK_SPEED_10GB_FULL:
+		incval = IXGBE_INCVAL_10GB;
+		shift = IXGBE_INCVAL_SHIFT_10GB;
+		break;
+	}
+
+	/**
+	 * Modify the calculated values to fit within the correct
+	 * number of bits specified by the hardware. The 82599 doesn't
+	 * have the same space as the X540, so bitshift the calculated
+	 * values to fit.
+	 */
+	switch (hw->mac.type) {
+	case ixgbe_mac_X540:
+		IXGBE_WRITE_REG(hw, IXGBE_TIMINCA, incval);
+		break;
+	case ixgbe_mac_82599EB:
+		incval >>= IXGBE_INCVAL_SHIFT_82599;
+		shift -= IXGBE_INCVAL_SHIFT_82599;
+		IXGBE_WRITE_REG(hw, IXGBE_TIMINCA,
+				(1 << IXGBE_INCPER_SHIFT_82599) |
+				incval);
+		break;
+	default:
+		/* other devices aren't supported */
+		return;
+	}
+
+	/* reset the system time registers */
+	IXGBE_WRITE_REG(hw, IXGBE_SYSTIML, 0x00000000);
+	IXGBE_WRITE_REG(hw, IXGBE_SYSTIMH, 0x00000000);
+	IXGBE_WRITE_FLUSH(hw);
+
+	/* store the new cycle speed */
+	adapter->cycle_speed = cycle_speed;
+
+	ACCESS_ONCE(adapter->base_incval) = incval;
+	smp_mb();
+
+	/* grab the ptp lock */
+	spin_lock_irqsave(&adapter->tmreg_lock, flags);
+
+	memset(&adapter->cc, 0, sizeof(adapter->cc));
+	adapter->cc.read = ixgbe_ptp_read;
+	adapter->cc.mask = CLOCKSOURCE_MASK(64);
+	adapter->cc.shift = shift;
+	adapter->cc.mult = 1;
+
+	/* reset the ns time counter */
+	timecounter_init(&adapter->tc, &adapter->cc,
+			 ktime_to_ns(ktime_get_real()));
+
+	spin_unlock_irqrestore(&adapter->tmreg_lock, flags);
+}
+
+/**
+ * ixgbe_ptp_init
+ * @adapter - the ixgbe private adapter structure
+ *
+ * This function performs the required steps for enabling ptp
+ * support. If ptp support has already been loaded it simply calls the
+ * cyclecounter init routine and exits.
+ */
+void ixgbe_ptp_init(struct ixgbe_adapter *adapter)
+{
+	struct net_device *netdev = adapter->netdev;
+
+	switch (adapter->hw.mac.type) {
+	case ixgbe_mac_X540:
+	case ixgbe_mac_82599EB:
+		snprintf(adapter->ptp_caps.name, 16, "%pm", netdev->dev_addr);
+		adapter->ptp_caps.owner = THIS_MODULE;
+		adapter->ptp_caps.max_adj = 250000000;
+		adapter->ptp_caps.n_alarm = 0;
+		adapter->ptp_caps.n_ext_ts = 0;
+		adapter->ptp_caps.n_per_out = 0;
+		adapter->ptp_caps.pps = 0;
+		adapter->ptp_caps.adjfreq = ixgbe_ptp_adjfreq;
+		adapter->ptp_caps.adjtime = ixgbe_ptp_adjtime;
+		adapter->ptp_caps.gettime = ixgbe_ptp_gettime;
+		adapter->ptp_caps.settime = ixgbe_ptp_settime;
+		adapter->ptp_caps.enable = ixgbe_ptp_enable;
+		break;
+	default:
+		adapter->ptp_clock = NULL;
+		return;
+	}
+
+	spin_lock_init(&adapter->tmreg_lock);
+
+	ixgbe_ptp_start_cyclecounter(adapter);
+
+	/* (Re)start the overflow check */
+	adapter->flags2 |= IXGBE_FLAG2_OVERFLOW_CHECK_ENABLED;
+
+	adapter->ptp_clock = ptp_clock_register(&adapter->ptp_caps);
+	if (IS_ERR(adapter->ptp_clock)) {
+		adapter->ptp_clock = NULL;
+		e_dev_err("ptp_clock_register failed\n");
+	} else
+		e_dev_info("registered PHC device on %s\n", netdev->name);
+
+	return;
+}
+
+/**
+ * ixgbe_ptp_stop - disable ptp device and stop the overflow check
+ * @adapter: pointer to adapter struct
+ *
+ * this function stops the ptp support, and cancels the delayed work.
+ */
+void ixgbe_ptp_stop(struct ixgbe_adapter *adapter)
+{
+	/* stop the overflow check task */
+	adapter->flags2 &= ~IXGBE_FLAG2_OVERFLOW_CHECK_ENABLED;
+
+	if (adapter->ptp_clock) {
+		ptp_clock_unregister(adapter->ptp_clock);
+		adapter->ptp_clock = NULL;
+		e_dev_info("removed PHC on %s\n",
+			   adapter->netdev->name);
+	}
+}
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
index 9559c03..87d54ca 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
@@ -1879,6 +1879,36 @@ enum {
 #define IXGBE_RXDCTL_RLPML_EN   0x00008000
 #define IXGBE_RXDCTL_VME        0x40000000  /* VLAN mode enable */
 
+#define IXGBE_TSYNCTXCTL_VALID		0x00000001 /* Tx timestamp valid */
+#define IXGBE_TSYNCTXCTL_ENABLED	0x00000010 /* Tx timestamping enabled */
+
+#define IXGBE_TSYNCRXCTL_VALID		0x00000001 /* Rx timestamp valid */
+#define IXGBE_TSYNCRXCTL_TYPE_MASK	0x0000000E /* Rx type mask */
+#define IXGBE_TSYNCRXCTL_TYPE_L2_V2	0x00
+#define IXGBE_TSYNCRXCTL_TYPE_L4_V1	0x02
+#define IXGBE_TSYNCRXCTL_TYPE_L2_L4_V2	0x04
+#define IXGBE_TSYNCRXCTL_TYPE_EVENT_V2	0x0A
+#define IXGBE_TSYNCRXCTL_ENABLED	0x00000010 /* Rx Timestamping enabled */
+
+#define IXGBE_RXMTRL_V1_CTRLT_MASK	0x000000FF
+#define IXGBE_RXMTRL_V1_SYNC_MSG	0x00
+#define IXGBE_RXMTRL_V1_DELAY_REQ_MSG	0x01
+#define IXGBE_RXMTRL_V1_FOLLOWUP_MSG	0x02
+#define IXGBE_RXMTRL_V1_DELAY_RESP_MSG	0x03
+#define IXGBE_RXMTRL_V1_MGMT_MSG	0x04
+
+#define IXGBE_RXMTRL_V2_MSGID_MASK		0x0000FF00
+#define IXGBE_RXMTRL_V2_SYNC_MSG		0x0000
+#define IXGBE_RXMTRL_V2_DELAY_REQ_MSG		0x0100
+#define IXGBE_RXMTRL_V2_PDELAY_REQ_MSG		0x0200
+#define IXGBE_RXMTRL_V2_PDELAY_RESP_MSG		0x0300
+#define IXGBE_RXMTRL_V2_FOLLOWUP_MSG		0x0800
+#define IXGBE_RXMTRL_V2_DELAY_RESP_MSG		0x0900
+#define IXGBE_RXMTRL_V2_PDELAY_FOLLOWUP_MSG	0x0A00
+#define IXGBE_RXMTRL_V2_ANNOUNCE_MSG		0x0B00
+#define IXGBE_RXMTRL_V2_SIGNALING_MSG		0x0C00
+#define IXGBE_RXMTRL_V2_MGMT_MSG		0x0D00
+
 #define IXGBE_FCTRL_SBP 0x00000002 /* Store Bad Packet */
 #define IXGBE_FCTRL_MPE 0x00000100 /* Multicast Promiscuous Ena*/
 #define IXGBE_FCTRL_UPE 0x00000200 /* Unicast Promiscuous Ena */
@@ -2008,6 +2038,7 @@ enum {
 #define IXGBE_RXDADV_STAT_FCSTAT_NODDP  0x00000010 /* 01: Ctxt w/o DDP */
 #define IXGBE_RXDADV_STAT_FCSTAT_FCPRSP 0x00000020 /* 10: Recv. FCP_RSP */
 #define IXGBE_RXDADV_STAT_FCSTAT_DDP    0x00000030 /* 11: Ctxt w/ DDP */
+#define IXGBE_RXDADV_STAT_TS		0x00010000 /* IEEE 1588 Time Stamp */
 
 /* PSRTYPE bit definitions */
 #define IXGBE_PSRTYPE_TCPHDR    0x00000010
@@ -2285,6 +2316,7 @@ struct ixgbe_adv_tx_context_desc {
 /* Adv Transmit Descriptor Config Masks */
 #define IXGBE_ADVTXD_DTALEN_MASK      0x0000FFFF /* Data buf length(bytes) */
 #define IXGBE_ADVTXD_MAC_LINKSEC      0x00040000 /* Insert LinkSec */
+#define IXGBE_ADVTXD_MAC_TSTAMP	      0x00080000 /* IEEE 1588 Time Stamp */
 #define IXGBE_ADVTXD_IPSEC_SA_INDEX_MASK   0x000003FF /* IPSec SA index */
 #define IXGBE_ADVTXD_IPSEC_ESP_LEN_MASK    0x000001FF /* IPSec ESP length */
 #define IXGBE_ADVTXD_DTYP_MASK  0x00F00000 /* DTYP mask */
-- 
1.7.7.6

[net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Jeff Kirsher]

From: Jacob E Keller <jacob.e.keller@intel.com>

This patch enables the PPS system in the PHC framework, by enabling
the clock-out feature on the X540 device. Causes the SDP0 to be set as
a 1Hz clock. Also configures the timesync interrupt cause in order to
report each pulse to the PPS via the PHC framework, which can be used
for general system clock synchronization. (This allows a stable method
for tuning the general system time via the on-board SYSTIM register
based clock.)

Signed-off-by: Jacob E Keller <jacob.e.keller@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe.h      |    2 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c |    6 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c  |  170 ++++++++++++++++++++++++-
 drivers/net/ethernet/intel/ixgbe/ixgbe_type.h |   12 ++
 4 files changed, 189 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index c90fbd2..cba6ff4 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -466,6 +466,7 @@ struct ixgbe_adapter {
 #define IXGBE_FLAG2_RSS_FIELD_IPV4_UDP		(u32)(1 << 8)
 #define IXGBE_FLAG2_RSS_FIELD_IPV6_UDP		(u32)(1 << 9)
 #define IXGBE_FLAG2_OVERFLOW_CHECK_ENABLED	(u32)(1 << 10)
+#define IXGBE_FLAG2_PTP_PPS_ENABLED		(u32)(1 << 11)
 
 	/* Tx fast path data */
 	int num_tx_queues;
@@ -719,6 +720,7 @@ extern void ixgbe_ptp_rx_hwtstamp(struct ixgbe_q_vector *q_vector,
 extern int ixgbe_ptp_hwtstamp_ioctl(struct ixgbe_adapter *adapter,
 				    struct ifreq *ifr, int cmd);
 extern void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter);
+extern void ixgbe_ptp_check_pps_event(struct ixgbe_adapter *adapter, u32 eicr);
 #endif /* CONFIG_IXGBE_PTP */
 
 #endif /* _IXGBE_H_ */
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 9a83c40..1ad6e2a 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -2322,6 +2322,9 @@ static irqreturn_t ixgbe_msix_other(int irq, void *data)
 	}
 
 	ixgbe_check_fan_failure(adapter, eicr);
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_check_pps_event(adapter, eicr);
+#endif
 
 	/* re-enable the original interrupt state, no lsc, no queues */
 	if (!test_bit(__IXGBE_DOWN, &adapter->state))
@@ -2514,6 +2517,9 @@ static irqreturn_t ixgbe_intr(int irq, void *data)
 	}
 
 	ixgbe_check_fan_failure(adapter, eicr);
+#ifdef CONFIG_IXGBE_PTP
+	ixgbe_ptp_check_pps_event(adapter, eicr);
+#endif
 
 	/* would disable interrupts here but EIAM disabled it */
 	napi_schedule(&q_vector->napi);
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
index 0b6553e..ddc6a4d 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
@@ -96,6 +96,10 @@
 
 #define IXGBE_OVERFLOW_PERIOD    (HZ * 30)
 
+#ifndef NSECS_PER_SEC
+#define NSECS_PER_SEC 1000000000ULL
+#endif
+
 /**
  * ixgbe_ptp_read - read raw cycle counter (to be used by time counter)
  * @cc - the cyclecounter structure
@@ -252,15 +256,154 @@ static int ixgbe_ptp_settime(struct ptp_clock_info *ptp,
  * @on - whether to enable or disable the feature
  *
  * enable (or disable) ancillary features of the phc subsystem.
- * our driver does not support any of these features
+ * our driver only supports the PPS feature on the X540
  */
 static int ixgbe_ptp_enable(struct ptp_clock_info *ptp,
 			    struct ptp_clock_request *rq, int on)
 {
+	struct ixgbe_adapter *adapter =
+		container_of(ptp, struct ixgbe_adapter, ptp_caps);
+
+	/**
+	 * When PPS is enabled, unmask the interrupt for the ClockOut
+	 * feature, so that the interrupt handler can send the PPS
+	 * event when the clock SDP triggers. Clear mask when PPS is
+	 * disabled
+	 */
+	if (rq->type == PTP_CLK_REQ_PPS) {
+		switch (adapter->hw.mac.type) {
+		case ixgbe_mac_X540:
+			if (on)
+				adapter->flags2 |= IXGBE_FLAG2_PTP_PPS_ENABLED;
+			else
+				adapter->flags2 &=
+					~IXGBE_FLAG2_PTP_PPS_ENABLED;
+			return 0;
+		default:
+			break;
+		}
+	}
+
 	return -ENOTSUPP;
 }
 
 /**
+ * ixgbe_ptp_check_pps_event
+ * @adapter - the private adapter structure
+ * @eicr - the interrupt cause register value
+ *
+ * This function is called by the interrupt routine when checking for
+ * interrupts. It will check and handle a pps event.
+ */
+void ixgbe_ptp_check_pps_event(struct ixgbe_adapter *adapter, u32 eicr)
+{
+	struct ixgbe_hw *hw = &adapter->hw;
+	struct ptp_clock_event event;
+
+	event.type = PTP_CLOCK_PPS;
+
+	/* Make sure ptp clock is valid, and PPS event enabled */
+	if (!adapter->ptp_clock ||
+	    !(adapter->flags2 & IXGBE_FLAG2_PTP_PPS_ENABLED))
+		return;
+
+	switch (hw->mac.type) {
+	case ixgbe_mac_X540:
+		if (eicr & IXGBE_EICR_TIMESYNC)
+			ptp_clock_event(adapter->ptp_clock, &event);
+		break;
+	default:
+		break;
+	}
+}
+
+/**
+ * ixgbe_ptp_enable_sdp
+ * @hw - the hardware private structure
+ * @shift - the clock shift for calculating nanoseconds
+ *
+ * this function enables the clock out feature on the sdp0 for the
+ * X540 device. It will create a 1second periodic output that can be
+ * used as the PPS (via an interrupt).
+ *
+ * It calculates when the systime will be on an exact second, and then
+ * aligns the start of the PPS signal to that value. The shift is
+ * necessary because it can change based on the link speed.
+ */
+static void ixgbe_ptp_enable_sdp(struct ixgbe_hw *hw, int shift)
+{
+	u32 esdp, tsauxc, clktiml, clktimh, trgttiml, trgttimh;
+	u64 clock_edge = 0;
+	u32 rem;
+
+	switch (hw->mac.type) {
+	case ixgbe_mac_X540:
+		esdp = IXGBE_READ_REG(hw, IXGBE_ESDP);
+
+		/*
+		 * enable the SDP0 pin as output, and connected to the native
+		 * function for Timesync (ClockOut)
+		 */
+		esdp |= (IXGBE_ESDP_SDP0_DIR |
+			 IXGBE_ESDP_SDP0_NATIVE);
+
+		/*
+		 * enable the Clock Out feature on SDP0, and allow interrupts
+		 * to occur when the pin changes
+		 */
+		tsauxc = (IXGBE_TSAUXC_EN_CLK |
+			  IXGBE_TSAUXC_SYNCLK |
+			  IXGBE_TSAUXC_SDP0_INT);
+
+		/* clock period (or pulse length) */
+		clktiml = (u32)(NSECS_PER_SEC << shift);
+		clktimh = (u32)((NSECS_PER_SEC << shift) >> 32);
+
+		clock_edge |= (u64)IXGBE_READ_REG(hw, IXGBE_SYSTIML);
+		clock_edge |= (u64)IXGBE_READ_REG(hw, IXGBE_SYSTIMH) << 32;
+
+		/*
+		 * account for the fact that we can't do u64 division
+		 * with remainder, by converting the clock values into
+		 * nanoseconds first
+		 */
+		clock_edge >>= shift;
+		div_u64_rem(clock_edge, NSECS_PER_SEC, &rem);
+		clock_edge += (NSECS_PER_SEC - rem);
+		clock_edge <<= shift;
+
+		/* specify the initial clock start time */
+		trgttiml = (u32)clock_edge;
+		trgttimh = (u32)(clock_edge >> 32);
+
+		IXGBE_WRITE_REG(hw, IXGBE_CLKTIML, clktiml);
+		IXGBE_WRITE_REG(hw, IXGBE_CLKTIMH, clktimh);
+		IXGBE_WRITE_REG(hw, IXGBE_TRGTTIML0, trgttiml);
+		IXGBE_WRITE_REG(hw, IXGBE_TRGTTIMH0, trgttimh);
+
+		IXGBE_WRITE_REG(hw, IXGBE_ESDP, esdp);
+		IXGBE_WRITE_REG(hw, IXGBE_TSAUXC, tsauxc);
+
+		IXGBE_WRITE_REG(hw, IXGBE_EIMS, IXGBE_EICR_TIMESYNC);
+		break;
+	default:
+		break;
+	}
+}
+
+/**
+ * ixgbe_ptp_disable_sdp
+ * @hw - the private hardware structure
+ *
+ * this function disables the auxiliary SDP clock out feature
+ */
+static void ixgbe_ptp_disable_sdp(struct ixgbe_hw *hw)
+{
+	IXGBE_WRITE_REG(hw, IXGBE_EIMC, IXGBE_EICR_TIMESYNC);
+	IXGBE_WRITE_REG(hw, IXGBE_TSAUXC, 0);
+}
+
+/**
  * ixgbe_ptp_overflow_check - delayed work to detect SYSTIME overflow
  * @work: structure containing information about this work task
  *
@@ -557,6 +700,9 @@ int ixgbe_ptp_hwtstamp_ioctl(struct ixgbe_adapter *adapter,
  * the device, which is used to generate the cycle counter
  * registers. Therefor this function is called whenever the link speed
  * changes.
+ *
+ * This function also turns on the SDP pin for clock out feature (X540
+ * only), because this is where the shift is first calculated.
  */
 void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter)
 {
@@ -588,6 +734,9 @@ void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter)
 	if (adapter->cycle_speed == cycle_speed)
 		return;
 
+	/* disable the SDP clock out */
+	ixgbe_ptp_disable_sdp(hw);
+
 	/**
 	 * Scale the NIC cycle counter by a large factor so that
 	 * relatively small corrections to the frequency can be added
@@ -640,6 +789,10 @@ void ixgbe_ptp_start_cyclecounter(struct ixgbe_adapter *adapter)
 	IXGBE_WRITE_REG(hw, IXGBE_SYSTIMH, 0x00000000);
 	IXGBE_WRITE_FLUSH(hw);
 
+	/* now that the shift has been calculated and the systime
+	 * registers reset, (re-)enable the Clock out feature*/
+	ixgbe_ptp_enable_sdp(hw, shift);
+
 	/* store the new cycle speed */
 	adapter->cycle_speed = cycle_speed;
 
@@ -676,6 +829,19 @@ void ixgbe_ptp_init(struct ixgbe_adapter *adapter)
 
 	switch (adapter->hw.mac.type) {
 	case ixgbe_mac_X540:
+		snprintf(adapter->ptp_caps.name, 16, "%pm", netdev->dev_addr);
+		adapter->ptp_caps.owner = THIS_MODULE;
+		adapter->ptp_caps.max_adj = 250000000;
+		adapter->ptp_caps.n_alarm = 0;
+		adapter->ptp_caps.n_ext_ts = 0;
+		adapter->ptp_caps.n_per_out = 0;
+		adapter->ptp_caps.pps = 1;
+		adapter->ptp_caps.adjfreq = ixgbe_ptp_adjfreq;
+		adapter->ptp_caps.adjtime = ixgbe_ptp_adjtime;
+		adapter->ptp_caps.gettime = ixgbe_ptp_gettime;
+		adapter->ptp_caps.settime = ixgbe_ptp_settime;
+		adapter->ptp_caps.enable = ixgbe_ptp_enable;
+		break;
 	case ixgbe_mac_82599EB:
 		snprintf(adapter->ptp_caps.name, 16, "%pm", netdev->dev_addr);
 		adapter->ptp_caps.owner = THIS_MODULE;
@@ -720,6 +886,8 @@ void ixgbe_ptp_init(struct ixgbe_adapter *adapter)
  */
 void ixgbe_ptp_stop(struct ixgbe_adapter *adapter)
 {
+	ixgbe_ptp_disable_sdp(&adapter->hw);
+
 	/* stop the overflow check task */
 	adapter->flags2 &= ~IXGBE_FLAG2_OVERFLOW_CHECK_ENABLED;
 
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
index 87d54ca..204848d 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_type.h
@@ -824,6 +824,8 @@ struct ixgbe_thermal_sensor_data {
 #define IXGBE_TRGTTIMH0  0x08C28 /* Target Time Register 0 High - RW */
 #define IXGBE_TRGTTIML1  0x08C2C /* Target Time Register 1 Low - RW */
 #define IXGBE_TRGTTIMH1  0x08C30 /* Target Time Register 1 High - RW */
+#define IXGBE_CLKTIML    0x08C34 /* Clock Out Time Register Low - RW */
+#define IXGBE_CLKTIMH    0x08C38 /* Clock Out Time Register High - RW */
 #define IXGBE_FREQOUT0   0x08C34 /* Frequency Out 0 Control register - RW */
 #define IXGBE_FREQOUT1   0x08C38 /* Frequency Out 1 Control register - RW */
 #define IXGBE_AUXSTMPL0  0x08C3C /* Auxiliary Time Stamp 0 register Low - RO */
@@ -1309,6 +1311,7 @@ enum {
 #define IXGBE_EICR_LINKSEC      0x00200000 /* PN Threshold */
 #define IXGBE_EICR_MNG          0x00400000 /* Manageability Event Interrupt */
 #define IXGBE_EICR_TS           0x00800000 /* Thermal Sensor Event */
+#define IXGBE_EICR_TIMESYNC     0x01000000 /* Timesync Event */
 #define IXGBE_EICR_GPI_SDP0     0x01000000 /* Gen Purpose Interrupt on SDP0 */
 #define IXGBE_EICR_GPI_SDP1     0x02000000 /* Gen Purpose Interrupt on SDP1 */
 #define IXGBE_EICR_GPI_SDP2     0x04000000 /* Gen Purpose Interrupt on SDP2 */
@@ -1326,6 +1329,7 @@ enum {
 #define IXGBE_EICS_MAILBOX      IXGBE_EICR_MAILBOX   /* VF to PF Mailbox Int */
 #define IXGBE_EICS_LSC          IXGBE_EICR_LSC       /* Link Status Change */
 #define IXGBE_EICS_MNG          IXGBE_EICR_MNG       /* MNG Event Interrupt */
+#define IXGBE_EICS_TIMESYNC     IXGBE_EICR_TIMESYNC  /* Timesync Event */
 #define IXGBE_EICS_GPI_SDP0     IXGBE_EICR_GPI_SDP0  /* SDP0 Gen Purpose Int */
 #define IXGBE_EICS_GPI_SDP1     IXGBE_EICR_GPI_SDP1  /* SDP1 Gen Purpose Int */
 #define IXGBE_EICS_GPI_SDP2     IXGBE_EICR_GPI_SDP2  /* SDP2 Gen Purpose Int */
@@ -1344,6 +1348,7 @@ enum {
 #define IXGBE_EIMS_LSC          IXGBE_EICR_LSC       /* Link Status Change */
 #define IXGBE_EIMS_MNG          IXGBE_EICR_MNG       /* MNG Event Interrupt */
 #define IXGBE_EIMS_TS           IXGBE_EICR_TS        /* Thermel Sensor Event */
+#define IXGBE_EIMS_TIMESYNC     IXGBE_EICR_TIMESYNC  /* Timesync Event */
 #define IXGBE_EIMS_GPI_SDP0     IXGBE_EICR_GPI_SDP0  /* SDP0 Gen Purpose Int */
 #define IXGBE_EIMS_GPI_SDP1     IXGBE_EICR_GPI_SDP1  /* SDP1 Gen Purpose Int */
 #define IXGBE_EIMS_GPI_SDP2     IXGBE_EICR_GPI_SDP2  /* SDP2 Gen Purpose Int */
@@ -1361,6 +1366,7 @@ enum {
 #define IXGBE_EIMC_MAILBOX      IXGBE_EICR_MAILBOX   /* VF to PF Mailbox Int */
 #define IXGBE_EIMC_LSC          IXGBE_EICR_LSC       /* Link Status Change */
 #define IXGBE_EIMC_MNG          IXGBE_EICR_MNG       /* MNG Event Interrupt */
+#define IXGBE_EIMC_TIMESYNC     IXGBE_EICR_TIMESYNC  /* Timesync Event */
 #define IXGBE_EIMC_GPI_SDP0     IXGBE_EICR_GPI_SDP0  /* SDP0 Gen Purpose Int */
 #define IXGBE_EIMC_GPI_SDP1     IXGBE_EICR_GPI_SDP1  /* SDP1 Gen Purpose Int */
 #define IXGBE_EIMC_GPI_SDP2     IXGBE_EICR_GPI_SDP2  /* SDP2 Gen Purpose Int */
@@ -1501,8 +1507,10 @@ enum {
 #define IXGBE_ESDP_SDP4 0x00000010 /* SDP4 Data Value */
 #define IXGBE_ESDP_SDP5 0x00000020 /* SDP5 Data Value */
 #define IXGBE_ESDP_SDP6 0x00000040 /* SDP6 Data Value */
+#define IXGBE_ESDP_SDP0_DIR     0x00000100 /* SDP0 IO direction */
 #define IXGBE_ESDP_SDP4_DIR     0x00000004 /* SDP4 IO direction */
 #define IXGBE_ESDP_SDP5_DIR     0x00002000 /* SDP5 IO direction */
+#define IXGBE_ESDP_SDP0_NATIVE  0x00010000 /* SDP0 Native Function */
 
 /* LEDCTL Bit Masks */
 #define IXGBE_LED_IVRT_BASE      0x00000040
@@ -1879,6 +1887,10 @@ enum {
 #define IXGBE_RXDCTL_RLPML_EN   0x00008000
 #define IXGBE_RXDCTL_VME        0x40000000  /* VLAN mode enable */
 
+#define IXGBE_TSAUXC_EN_CLK   0x00000004
+#define IXGBE_TSAUXC_SYNCLK   0x00000008
+#define IXGBE_TSAUXC_SDP0_INT 0x00000040
+
 #define IXGBE_TSYNCTXCTL_VALID		0x00000001 /* Tx timestamp valid */
 #define IXGBE_TSYNCTXCTL_ENABLED	0x00000010 /* Tx timestamping enabled */
 
-- 
1.7.7.6

[net-next 08/12] ixgbe: correct disable_rx_buff timeout

[Jeff Kirsher]

From: Jacob Keller <jacob.e.keller@intel.com>

The current value of the udelay timeout for ixgbe_disable_rx_buff is too
short. This causes the security path to not not be properly disabled during
the section that is meant to have it turned off. The end result causes a race
condition that results in RX issues.

Signed-off-by: Jacob Keller <jacob.e.keller@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_common.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
index c7e51b8..77ac41f 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_common.c
@@ -2561,7 +2561,7 @@ s32 ixgbe_disable_rx_buff_generic(struct ixgbe_hw *hw)
 			break;
 		else
 			/* Use interrupt-safe sleep just in case */
-			udelay(10);
+			udelay(1000);
 	}
 
 	/* For informational purposes only */
-- 
1.7.7.6

[net-next 09/12] ixgbe: add support for get_ts_info

[Jeff Kirsher]

From: Jacob Keller <jacob.e.keller@intel.com>

This patch adds support for the ethtool get_ts_info operation, which enables
access of available timestamp/timesync support for that device. It can query
which ptp clock device is associated with the particular port.

Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c |   41 ++++++++++++++++++++++
 1 files changed, 41 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
index 082446d..3178f1e 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
@@ -2696,6 +2696,46 @@ static int ixgbe_set_rxnfc(struct net_device *dev, struct ethtool_rxnfc *cmd)
 	return ret;
 }
 
+static int ixgbe_get_ts_info(struct net_device *dev,
+			     struct ethtool_ts_info *info)
+{
+	struct ixgbe_adapter *adapter = netdev_priv(dev);
+
+	switch (adapter->hw.mac.type) {
+#ifdef CONFIG_IXGBE_PTP
+	case ixgbe_mac_X540:
+	case ixgbe_mac_82599EB:
+		info->so_timestamping =
+			SOF_TIMESTAMPING_TX_HARDWARE |
+			SOF_TIMESTAMPING_RX_HARDWARE |
+			SOF_TIMESTAMPING_RAW_HARDWARE;
+
+		if (adapter->ptp_clock)
+			info->phc_index = ptp_clock_index(adapter->ptp_clock);
+		else
+			info->phc_index = -1;
+
+		info->tx_types =
+			(1 << HWTSTAMP_TX_OFF) |
+			(1 << HWTSTAMP_TX_ON);
+
+		info->rx_filters =
+			(1 << HWTSTAMP_FILTER_NONE) |
+			(1 << HWTSTAMP_FILTER_PTP_V1_L4_SYNC) |
+			(1 << HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ) |
+			(1 << HWTSTAMP_FILTER_PTP_V2_SYNC) |
+			(1 << HWTSTAMP_FILTER_PTP_V2_DELAY_REQ) |
+			(1 << HWTSTAMP_FILTER_PTP_V2_EVENT) |
+			(1 << HWTSTAMP_FILTER_SOME);
+		break;
+#endif /* CONFIG_IXGBE_PTP */
+	default:
+		return ethtool_op_get_ts_info(dev, info);
+		break;
+	}
+	return 0;
+}
+
 static const struct ethtool_ops ixgbe_ethtool_ops = {
 	.get_settings           = ixgbe_get_settings,
 	.set_settings           = ixgbe_set_settings,
@@ -2724,6 +2764,7 @@ static const struct ethtool_ops ixgbe_ethtool_ops = {
 	.set_coalesce           = ixgbe_set_coalesce,
 	.get_rxnfc		= ixgbe_get_rxnfc,
 	.set_rxnfc		= ixgbe_set_rxnfc,
+	.get_ts_info		= ixgbe_get_ts_info,
 };
 
 void ixgbe_set_ethtool_ops(struct net_device *netdev)
-- 
1.7.7.6

[net-next 10/12] ixgbe: support software timestamping

[Jeff Kirsher]

From: Jacob Keller <jacob.e.keller@intel.com>

Kernel software timestamping requires that the driver calls skb_tx_timestamp
just before passing the skb to the MAC, in order to provide the best software
timestamps. This patch adds this call for that support.

Signed-off-by: Jacob Keller <jacob.e.keller@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 |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 1ad6e2a..05f66a2 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -6298,6 +6298,8 @@ netdev_tx_t ixgbe_xmit_frame_ring(struct sk_buff *skb,
 		tx_flags |= IXGBE_TX_FLAGS_SW_VLAN;
 	}
 
+	skb_tx_timestamp(skb);
+
 #ifdef CONFIG_IXGBE_PTP
 	if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_HW_TSTAMP)) {
 		skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
-- 
1.7.7.6

[net-next 11/12] ixgbe: cleanup the hwmon function calls

[Jeff Kirsher]

From: Don Skidmore <donald.c.skidmore@intel.com>

When the hwmon code was initially added it was with the assumption that a
sysfs patch would be also coming soon.  Since that isn't the case some
clean up needs to be done.  This patch does that.

Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
---
 drivers/net/ethernet/intel/ixgbe/ixgbe.h       |    2 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c  |    4 ++
 drivers/net/ethernet/intel/ixgbe/ixgbe_sysfs.c |   36 ++---------------------
 3 files changed, 10 insertions(+), 32 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index cba6ff4..3ef3c52 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -677,8 +677,10 @@ extern int ixgbe_setup_tc(struct net_device *dev, u8 tc);
 #endif
 extern void ixgbe_tx_ctxtdesc(struct ixgbe_ring *, u32, u32, u32, u32);
 extern void ixgbe_do_reset(struct net_device *netdev);
+#ifdef CONFIG_IXGBE_HWMON
 extern void ixgbe_sysfs_exit(struct ixgbe_adapter *adapter);
 extern int ixgbe_sysfs_init(struct ixgbe_adapter *adapter);
+#endif /* CONFIG_IXGBE_HWMON */
 #ifdef IXGBE_FCOE
 extern void ixgbe_configure_fcoe(struct ixgbe_adapter *adapter);
 extern int ixgbe_fso(struct ixgbe_ring *tx_ring,
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 05f66a2..863c9b7 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -7344,8 +7344,10 @@ static int __devinit ixgbe_probe(struct pci_dev *pdev,
 	e_dev_info("%s\n", ixgbe_default_device_descr);
 	cards_found++;
 
+#ifdef CONFIG_IXGBE_HWMON
 	if (ixgbe_sysfs_init(adapter))
 		e_err(probe, "failed to allocate sysfs resources\n");
+#endif /* CONFIG_IXGBE_HWMON */
 
 	return 0;
 
@@ -7397,7 +7399,9 @@ static void __devexit ixgbe_remove(struct pci_dev *pdev)
 	}
 
 #endif
+#ifdef CONFIG_IXGBE_HWMON
 	ixgbe_sysfs_exit(adapter);
+#endif /* CONFIG_IXGBE_HWMON */
 
 #ifdef IXGBE_FCOE
 	if (adapter->flags & IXGBE_FLAG_FCOE_ENABLED)
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_sysfs.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_sysfs.c
index f81c166..1d80b1c 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_sysfs.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_sysfs.c
@@ -37,12 +37,7 @@
 #include <linux/netdevice.h>
 #include <linux/hwmon.h>
 
-/*
- * This file provides a sysfs interface to export information from the
- * driver.  The information presented is READ-ONLY.
- */
 #ifdef CONFIG_IXGBE_HWMON
-
 /* hwmon callback functions */
 static ssize_t ixgbe_hwmon_show_location(struct device *dev,
 					 struct device_attribute *attr,
@@ -162,17 +157,13 @@ static int ixgbe_add_hwmon_attr(struct ixgbe_adapter *adapter,
 
 	return rc;
 }
-#endif /* CONFIG_IXGBE_HWMON */
 
 static void ixgbe_sysfs_del_adapter(struct ixgbe_adapter *adapter)
 {
-#ifdef CONFIG_IXGBE_HWMON
 	int i;
-#endif /* CONFIG_IXGBE_HWMON */
 
 	if (adapter == NULL)
 		return;
-#ifdef CONFIG_IXGBE_HWMON
 
 	for (i = 0; i < adapter->ixgbe_hwmon_buff.n_hwmon; i++) {
 		device_remove_file(&adapter->pdev->dev,
@@ -183,12 +174,6 @@ static void ixgbe_sysfs_del_adapter(struct ixgbe_adapter *adapter)
 
 	if (adapter->ixgbe_hwmon_buff.device)
 		hwmon_device_unregister(adapter->ixgbe_hwmon_buff.device);
-#endif /* CONFIG_IXGBE_HWMON */
-
-	if (adapter->info_kobj != NULL) {
-		kobject_put(adapter->info_kobj);
-		adapter->info_kobj = NULL;
-	}
 }
 
 /* called from ixgbe_main.c */
@@ -200,32 +185,19 @@ void ixgbe_sysfs_exit(struct ixgbe_adapter *adapter)
 /* called from ixgbe_main.c */
 int ixgbe_sysfs_init(struct ixgbe_adapter *adapter)
 {
-#ifdef CONFIG_IXGBE_HWMON
 	struct hwmon_buff *ixgbe_hwmon = &adapter->ixgbe_hwmon_buff;
 	unsigned int i;
 	int n_attrs;
-#endif /* CONFIG_IXGBE_HWMON */
-	struct net_device *netdev = adapter->netdev;
 	int rc = 0;
 
-	/* create info kobj and attribute listings in kobj */
-	adapter->info_kobj = kobject_create_and_add("info", &netdev->dev.kobj);
-	if (adapter->info_kobj == NULL) {
-		rc = -ENOMEM;
-		goto err;
-	}
-
-#ifdef CONFIG_IXGBE_HWMON
 	/* If this method isn't defined we don't support thermals */
 	if (adapter->hw.mac.ops.init_thermal_sensor_thresh == NULL) {
-		rc = -EPERM;
-		goto err;
+		goto exit;
 	}
 
 	/* Don't create thermal hwmon interface if no sensors present */
-	rc = adapter->hw.mac.ops.init_thermal_sensor_thresh(&adapter->hw);
-	if (rc)
-		goto err;
+	if (adapter->hw.mac.ops.init_thermal_sensor_thresh(&adapter->hw))
+		goto exit;
 
 	/*
 	 * Allocation space for max attributs
@@ -261,7 +233,6 @@ int ixgbe_sysfs_init(struct ixgbe_adapter *adapter)
 		if (rc)
 			goto err;
 	}
-#endif /* CONFIG_IXGBE_HWMON */
 
 	goto exit;
 
@@ -270,4 +241,5 @@ err:
 exit:
 	return rc;
 }
+#endif /* CONFIG_IXGBE_HWMON */
 
-- 
1.7.7.6

[net-next 12/12] ixgbe: update version number

[Jeff Kirsher]

From: Don Skidmore <donald.c.skidmore@intel.com>

Update version number to better match the version of the out of tree
driver with similar functionality.

Signed-off-by: Don Skidmore <donald.c.skidmore@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 |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 863c9b7..bf20457 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -63,8 +63,8 @@ static char ixgbe_default_device_descr[] =
 			      "Intel(R) 10 Gigabit Network Connection";
 #endif
 #define MAJ 3
-#define MIN 8
-#define BUILD 21
+#define MIN 9
+#define BUILD 15
 #define DRV_VERSION __stringify(MAJ) "." __stringify(MIN) "." \
 	__stringify(BUILD) "-k"
 const char ixgbe_driver_version[] = DRV_VERSION;
-- 
1.7.7.6

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Richard Cochran]

Mostly, this looks very good. I do have one concern and a nit, though.

On Wed, May 09, 2012 at 11:46:47PM -0700, Jeff Kirsher wrote:
> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> index 1693ec3..9a83c40 100644
> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> @@ -789,6 +789,13 @@ static bool ixgbe_clean_tx_irq(struct ixgbe_q_vector *q_vector,
>  		total_bytes += tx_buffer->bytecount;
>  		total_packets += tx_buffer->gso_segs;
>  
> +#ifdef CONFIG_IXGBE_PTP
> +		if (unlikely(tx_buffer->tx_flags &
> +			     IXGBE_TX_FLAGS_TSTAMP))
> +			ixgbe_ptp_tx_hwtstamp(q_vector,
> +					      tx_buffer->skb);

This looks strangely wrapped.

> +
> +#endif
>  		/* free the skb */
>  		dev_kfree_skb_any(tx_buffer->skb);
>  

...

> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
> new file mode 100644
> index 0000000..0b6553e
> --- /dev/null
> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c

...

> +/**
> + * ixgbe_ptp_rx_hwtstamp - utility function which checks for RX time stamp
> + * @q_vector: structure containing interrupt and ring information
> + * @skb: particular skb to send timestamp with
> + *
> + * if the timestamp is valid, we convert it into the timecounter ns
> + * value, then store that result into the shhwtstamps structure which
> + * is passed up the network stack
> + */
> +void ixgbe_ptp_rx_hwtstamp(struct ixgbe_q_vector *q_vector,
> +			   struct sk_buff *skb)
> +{
> +	struct ixgbe_adapter *adapter;
> +	struct ixgbe_hw *hw;
> +	struct skb_shared_hwtstamps *shhwtstamps;
> +	u64 regval = 0, ns;
> +	u32 tsyncrxctl;
> +	unsigned long flags;
> +
> +	/* we cannot process timestamps on a ring without a q_vector */
> +	if (!q_vector || !q_vector->adapter)
> +		return;
> +
> +	adapter = q_vector->adapter;
> +	hw = &adapter->hw;
> +
> +	tsyncrxctl = IXGBE_READ_REG(hw, IXGBE_TSYNCRXCTL);
> +	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_RXSTMPL);
> +	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_RXSTMPH) << 32;
> +
> +	/*
> +	 * If this bit is set, then the RX registers contain the time stamp. No
> +	 * other packet will be time stamped until we read these registers, so
> +	 * read the registers to make them available again. Because only one
> +	 * packet can be time stamped at a time, we know that the register
> +	 * values must belong to this one here and therefore we don't need to
> +	 * compare any of the additional attributes stored for it.

I suspect that this assumption is wrong. What happens if the time
stamping logic locks a value but the packet is lost because the ring
is full?

BTW, the IGB driver also has this defect.

Thanks,
Richard

Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Richard Cochran]

On Wed, May 09, 2012 at 11:46:48PM -0700, Jeff Kirsher wrote:
> From: Jacob E Keller <jacob.e.keller@intel.com>
> 
> This patch enables the PPS system in the PHC framework, by enabling
> the clock-out feature on the X540 device. Causes the SDP0 to be set as
> a 1Hz clock. Also configures the timesync interrupt cause in order to
> report each pulse to the PPS via the PHC framework, which can be used
> for general system clock synchronization. (This allows a stable method
> for tuning the general system time via the on-board SYSTIM register
> based clock.)

Glad to see the PPS output and internal PPS support.

> diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> index 9a83c40..1ad6e2a 100644
> --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c

...

>  /**
> + * ixgbe_ptp_check_pps_event
> + * @adapter - the private adapter structure
> + * @eicr - the interrupt cause register value
> + *
> + * This function is called by the interrupt routine when checking for
> + * interrupts. It will check and handle a pps event.
> + */
> +void ixgbe_ptp_check_pps_event(struct ixgbe_adapter *adapter, u32 eicr)
> +{
> +	struct ixgbe_hw *hw = &adapter->hw;
> +	struct ptp_clock_event event;
> +
> +	event.type = PTP_CLOCK_PPS;
> +
> +	/* Make sure ptp clock is valid, and PPS event enabled */
> +	if (!adapter->ptp_clock ||
> +	    !(adapter->flags2 & IXGBE_FLAG2_PTP_PPS_ENABLED))
> +		return;
> +
> +	switch (hw->mac.type) {
> +	case ixgbe_mac_X540:
> +		if (eicr & IXGBE_EICR_TIMESYNC)

Since this function is called in every interrupt, I would check this
flag first thing.

> +			ptp_clock_event(adapter->ptp_clock, &event);
> +		break;
> +	default:
> +		break;
> +	}
> +}

Thanks,
Richard

RE: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 7:11 AM
> To: Kirsher, Jeffrey T
> Cc: davem@davemloft.net; Keller, Jacob E; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware
> Clock (PHC)
> 
> Mostly, this looks very good. I do have one concern and a nit, though.
> 
> On Wed, May 09, 2012 at 11:46:47PM -0700, Jeff Kirsher wrote:
> > diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > index 1693ec3..9a83c40 100644
> > --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > @@ -789,6 +789,13 @@ static bool ixgbe_clean_tx_irq(struct ixgbe_q_vector
> *q_vector,
> >  		total_bytes += tx_buffer->bytecount;
> >  		total_packets += tx_buffer->gso_segs;
> >
> > +#ifdef CONFIG_IXGBE_PTP
> > +		if (unlikely(tx_buffer->tx_flags &
> > +			     IXGBE_TX_FLAGS_TSTAMP))
> > +			ixgbe_ptp_tx_hwtstamp(q_vector,
> > +					      tx_buffer->skb);
> 
> This looks strangely wrapped.
> 
> > +
> > +#endif
> >  		/* free the skb */
> >  		dev_kfree_skb_any(tx_buffer->skb);
> >
> 
> ...
> 
> > diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
> > b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
> > new file mode 100644
> > index 0000000..0b6553e
> > --- /dev/null
> > +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c
> 
> ...
> 
> > +/**
> > + * ixgbe_ptp_rx_hwtstamp - utility function which checks for RX time
> > +stamp
> > + * @q_vector: structure containing interrupt and ring information
> > + * @skb: particular skb to send timestamp with
> > + *
> > + * if the timestamp is valid, we convert it into the timecounter ns
> > + * value, then store that result into the shhwtstamps structure which
> > + * is passed up the network stack
> > + */
> > +void ixgbe_ptp_rx_hwtstamp(struct ixgbe_q_vector *q_vector,
> > +			   struct sk_buff *skb)
> > +{
> > +	struct ixgbe_adapter *adapter;
> > +	struct ixgbe_hw *hw;
> > +	struct skb_shared_hwtstamps *shhwtstamps;
> > +	u64 regval = 0, ns;
> > +	u32 tsyncrxctl;
> > +	unsigned long flags;
> > +
> > +	/* we cannot process timestamps on a ring without a q_vector */
> > +	if (!q_vector || !q_vector->adapter)
> > +		return;
> > +
> > +	adapter = q_vector->adapter;
> > +	hw = &adapter->hw;
> > +
> > +	tsyncrxctl = IXGBE_READ_REG(hw, IXGBE_TSYNCRXCTL);
> > +	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_RXSTMPL);
> > +	regval |= (u64)IXGBE_READ_REG(hw, IXGBE_RXSTMPH) << 32;
> > +
> > +	/*
> > +	 * If this bit is set, then the RX registers contain the time stamp. No
> > +	 * other packet will be time stamped until we read these registers, so
> > +	 * read the registers to make them available again. Because only one
> > +	 * packet can be time stamped at a time, we know that the register
> > +	 * values must belong to this one here and therefore we don't need to
> > +	 * compare any of the additional attributes stored for it.
> 
> I suspect that this assumption is wrong. What happens if the time stamping
> logic locks a value but the packet is lost because the ring is full?
> 
> BTW, the IGB driver also has this defect.
>

Note how I read the rx registers first? So it will always clear the value.
That should unlock the value for the next rx stamp packet.
 
> Thanks,
> Richard

RE: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 7:17 AM
> To: Kirsher, Jeffrey T
> Cc: davem@davemloft.net; Keller, Jacob E; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS
> support on X540
> 
> On Wed, May 09, 2012 at 11:46:48PM -0700, Jeff Kirsher wrote:
> > From: Jacob E Keller <jacob.e.keller@intel.com>
> >
> > This patch enables the PPS system in the PHC framework, by enabling
> > the clock-out feature on the X540 device. Causes the SDP0 to be set as
> > a 1Hz clock. Also configures the timesync interrupt cause in order to
> > report each pulse to the PPS via the PHC framework, which can be used
> > for general system clock synchronization. (This allows a stable method
> > for tuning the general system time via the on-board SYSTIM register
> > based clock.)
> 
> Glad to see the PPS output and internal PPS support.
> 
> > diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > index 9a83c40..1ad6e2a 100644
> > --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> 
> ...
> 
> >  /**
> > + * ixgbe_ptp_check_pps_event
> > + * @adapter - the private adapter structure
> > + * @eicr - the interrupt cause register value
> > + *
> > + * This function is called by the interrupt routine when checking for
> > + * interrupts. It will check and handle a pps event.
> > + */
> > +void ixgbe_ptp_check_pps_event(struct ixgbe_adapter *adapter, u32
> > +eicr) {
> > +	struct ixgbe_hw *hw = &adapter->hw;
> > +	struct ptp_clock_event event;
> > +
> > +	event.type = PTP_CLOCK_PPS;
> > +
> > +	/* Make sure ptp clock is valid, and PPS event enabled */
> > +	if (!adapter->ptp_clock ||
> > +	    !(adapter->flags2 & IXGBE_FLAG2_PTP_PPS_ENABLED))
> > +		return;
> > +
> > +	switch (hw->mac.type) {
> > +	case ixgbe_mac_X540:
> > +		if (eicr & IXGBE_EICR_TIMESYNC)
> 
> Since this function is called in every interrupt, I would check this flag
> first thing.
> 
Not sure what you mean? Check this before checking the other interrupts?
I can do that.

> > +			ptp_clock_event(adapter->ptp_clock, &event);
> > +		break;
> > +	default:
> > +		break;
> > +	}
> > +}
> 
> Thanks,
> Richard

RE: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Keller, Jacob E]

> -----Original Message-----
> From: Keller, Jacob E
> Sent: Thursday, May 10, 2012 2:55 PM
> To: 'Richard Cochran'; Kirsher, Jeffrey T
> Cc: davem@davemloft.net; netdev@vger.kernel.org; gospo@redhat.com;
> sassmann@redhat.com
> Subject: RE: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS
> support on X540
> 
> 
> 
> > -----Original Message-----
> > From: Richard Cochran [mailto:richardcochran@gmail.com]
> > Sent: Thursday, May 10, 2012 7:17 AM
> > To: Kirsher, Jeffrey T
> > Cc: davem@davemloft.net; Keller, Jacob E; netdev@vger.kernel.org;
> > gospo@redhat.com; sassmann@redhat.com
> > Subject: Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature
> > for PPS support on X540
> >
> > On Wed, May 09, 2012 at 11:46:48PM -0700, Jeff Kirsher wrote:
> > > From: Jacob E Keller <jacob.e.keller@intel.com>
> > >
> > > This patch enables the PPS system in the PHC framework, by enabling
> > > the clock-out feature on the X540 device. Causes the SDP0 to be set
> > > as a 1Hz clock. Also configures the timesync interrupt cause in
> > > order to report each pulse to the PPS via the PHC framework, which
> > > can be used for general system clock synchronization. (This allows a
> > > stable method for tuning the general system time via the on-board
> > > SYSTIM register based clock.)
> >
> > Glad to see the PPS output and internal PPS support.
> >
> > > diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > > b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > > index 9a83c40..1ad6e2a 100644
> > > --- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> > > +++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
> >
> > ...
> >
> > >  /**
> > > + * ixgbe_ptp_check_pps_event
> > > + * @adapter - the private adapter structure
> > > + * @eicr - the interrupt cause register value
> > > + *
> > > + * This function is called by the interrupt routine when checking
> > > +for
> > > + * interrupts. It will check and handle a pps event.
> > > + */
> > > +void ixgbe_ptp_check_pps_event(struct ixgbe_adapter *adapter, u32
> > > +eicr) {
> > > +	struct ixgbe_hw *hw = &adapter->hw;
> > > +	struct ptp_clock_event event;
> > > +
> > > +	event.type = PTP_CLOCK_PPS;
> > > +
> > > +	/* Make sure ptp clock is valid, and PPS event enabled */
> > > +	if (!adapter->ptp_clock ||
> > > +	    !(adapter->flags2 & IXGBE_FLAG2_PTP_PPS_ENABLED))
> > > +		return;
> > > +
> > > +	switch (hw->mac.type) {
> > > +	case ixgbe_mac_X540:
> > > +		if (eicr & IXGBE_EICR_TIMESYNC)
> >
> > Since this function is called in every interrupt, I would check this
> > flag first thing.
> >
> Not sure what you mean? Check this before checking the other interrupts?
> I can do that.
> 
> > > +			ptp_clock_event(adapter->ptp_clock, &event);
> > > +		break;
> > > +	default:
> > > +		break;
> > > +	}
> > > +}
> >
> > Thanks,
> > Richard

Oops stupid mail program sent that on accident. Anyways: I think you might be
right, Richard. We don't read those timestamp values unless the stat err bit
for timestamps is set on the descriptor. But I am not sure what happens when the
tjmestamped packet is dropped off the end of the ring. What would you propose
here? How can we detect if this timestamp doesn't match the packet? I can look
into using the extra hardware features for matching timestamps. That might be
a more useful, in that it would help prevent this case.

- Jake

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Wed,  9 May 2012 23:46:41 -0700

> This series of patches contains updates for igb and ixgbe.
> Special note ixgbe patches contain the addition of PTP support from
> Jacob Keller.
> 
> The following are changes since commit 8feedbb4a710784d2858acba5c90e903e93e36eb:
>   dsa: Convert compare_ether_addr to ether_addr_equal
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Pulled, thanks Jeff.

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Richard Cochran]

On Thu, May 10, 2012 at 09:53:18PM +0000, Keller, Jacob E wrote:
> > > +	/*
> > > +	 * If this bit is set, then the RX registers contain the time stamp. No
> > > +	 * other packet will be time stamped until we read these registers, so
> > > +	 * read the registers to make them available again. Because only one
> > > +	 * packet can be time stamped at a time, we know that the register
> > > +	 * values must belong to this one here and therefore we don't need to
> > > +	 * compare any of the additional attributes stored for it.
> > 
> > I suspect that this assumption is wrong. What happens if the time stamping
> > logic locks a value but the packet is lost because the ring is full?
> > 
> > BTW, the IGB driver also has this defect.
> >
> 
> Note how I read the rx registers first? So it will always clear the value.
> That should unlock the value for the next rx stamp packet.

1. Hw recognizes ptp event packet, locks time stamp
2. Hw drops packet because queue is full
3. No more time stamps are ever generated

Can this happen? The docs seems to say it can.

Richard

Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Richard Cochran]

On Thu, May 10, 2012 at 09:54:52PM +0000, Keller, Jacob E wrote:
> > Since this function is called in every interrupt, I would check this flag
> > first thing.
> > 
> Not sure what you mean? Check this before checking the other interrupts?
> I can do that.

I only meant that, assuming that the other interrupt sources are much
more frequent than the PPS, the normal case will be that the PPS flag
is not set.

It would be more efficient (that is, shorter ISR code path in normal
case) to check the flag first, perhaps like this.

ixgbe_intr()
{
	...
#ifdef CONFIG_IXGBE_PTP
	if (eicr & IXGBE_EICR_TIMESYNC)
		ixgbe_ptp_check_pps_event(adapter, eicr);
#endif
	...
}

Thanks,
Richard

Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Richard Cochran]

On Thu, May 10, 2012 at 10:08:44PM +0000, Keller, Jacob E wrote:
> 
> Oops stupid mail program sent that on accident. Anyways: I think you might be
> right, Richard. We don't read those timestamp values unless the stat err bit
> for timestamps is set on the descriptor. But I am not sure what happens when the
> tjmestamped packet is dropped off the end of the ring. What would you propose
> here? How can we detect if this timestamp doesn't match the packet? I can look
> into using the extra hardware features for matching timestamps. That might be
> a more useful, in that it would help prevent this case.

[ Talking about the Rx time stamping locking from other patch... ]

The IGB provides some PTP event packet identification fields (seqNum,
etc) just for the purpose of matching time stamps to packets. Some of
the other PHC drivers (ixp4xx, dp83640) have code that does the
matching.

HTH,
Richard

RE: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 10:15 PM
> To: Keller, Jacob E
> Cc: Kirsher, Jeffrey T; davem@davemloft.net; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware
> Clock (PHC)
> 
> On Thu, May 10, 2012 at 09:53:18PM +0000, Keller, Jacob E wrote:
> > > > +	/*
> > > > +	 * If this bit is set, then the RX registers contain the time
> stamp. No
> > > > +	 * other packet will be time stamped until we read these
> registers, so
> > > > +	 * read the registers to make them available again. Because only
> one
> > > > +	 * packet can be time stamped at a time, we know that the
> register
> > > > +	 * values must belong to this one here and therefore we don't
> need to
> > > > +	 * compare any of the additional attributes stored for it.
> > >
> > > I suspect that this assumption is wrong. What happens if the time
> > > stamping logic locks a value but the packet is lost because the ring is
> full?
> > >
> > > BTW, the IGB driver also has this defect.
> > >
> >
> > Note how I read the rx registers first? So it will always clear the value.
> > That should unlock the value for the next rx stamp packet.
> 
> 1. Hw recognizes ptp event packet, locks time stamp 2. Hw drops packet because
> queue is full 3. No more time stamps are ever generated
> 
> Can this happen? The docs seems to say it can.
> 
> Richard

It might be possible. I'm curious what's the best method to solve this. I don't
think this generation of hardware can get an interrupt when the RX timestamp is
locked, and I don't know how easy it would be to process the sequence number within
the driver. It might be the case that the next packet would still be marked in the
descriptor and therefor the previous timestamp would apply to the next packet. I am
not sure, but I know we haven't seen issues where timestamps drop completely. This
is something I am not sure what's the best solution.

- Jake

RE: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 10:36 PM
> To: Keller, Jacob E
> Cc: Kirsher, Jeffrey T; davem@davemloft.net; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS
> support on X540
> 
> On Thu, May 10, 2012 at 09:54:52PM +0000, Keller, Jacob E wrote:
> > > Since this function is called in every interrupt, I would check this
> > > flag first thing.
> > >
> > Not sure what you mean? Check this before checking the other interrupts?
> > I can do that.
> 
> I only meant that, assuming that the other interrupt sources are much more
> frequent than the PPS, the normal case will be that the PPS flag is not set.
> 
> It would be more efficient (that is, shorter ISR code path in normal
> case) to check the flag first, perhaps like this.
> 
> ixgbe_intr()
> {
> 	...
> #ifdef CONFIG_IXGBE_PTP
> 	if (eicr & IXGBE_EICR_TIMESYNC)
> 		ixgbe_ptp_check_pps_event(adapter, eicr); #endif
> 	...
> }

Ok. This makes sense. :) Check the flag before checking the hardware type.

- Jake

> 
> Thanks,
> Richard

RE: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS support on X540

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 10:41 PM
> To: Keller, Jacob E
> Cc: Kirsher, Jeffrey T; davem@davemloft.net; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 07/12] ixgbe: Enable timesync clock-out feature for PPS
> support on X540
> 
> On Thu, May 10, 2012 at 10:08:44PM +0000, Keller, Jacob E wrote:
> >
> > Oops stupid mail program sent that on accident. Anyways: I think you
> > might be right, Richard. We don't read those timestamp values unless
> > the stat err bit for timestamps is set on the descriptor. But I am not
> > sure what happens when the tjmestamped packet is dropped off the end
> > of the ring. What would you propose here? How can we detect if this
> > timestamp doesn't match the packet? I can look into using the extra
> > hardware features for matching timestamps. That might be a more useful, in
> that it would help prevent this case.
> 
> [ Talking about the Rx time stamping locking from other patch... ]
> 
> The IGB provides some PTP event packet identification fields (seqNum,
> etc) just for the purpose of matching time stamps to packets. Some of the
> other PHC drivers (ixp4xx, dp83640) have code that does the matching.
> 

Ixgbe has the sequence number also. I'll take a look at the PHC drivers
that match already, and see how difficult it would be to perform this check.

- Jake

> HTH,
> Richard
> 

RE: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 10:15 PM
> To: Keller, Jacob E
> Cc: Kirsher, Jeffrey T; davem@davemloft.net; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware
> Clock (PHC)
> 
> On Thu, May 10, 2012 at 09:53:18PM +0000, Keller, Jacob E wrote:
> > > > +	/*
> > > > +	 * If this bit is set, then the RX registers contain the time
> stamp. No
> > > > +	 * other packet will be time stamped until we read these
> registers, so
> > > > +	 * read the registers to make them available again. Because only
> one
> > > > +	 * packet can be time stamped at a time, we know that the
> register
> > > > +	 * values must belong to this one here and therefore we don't
> need to
> > > > +	 * compare any of the additional attributes stored for it.
> > >
> > > I suspect that this assumption is wrong. What happens if the time
> > > stamping logic locks a value but the packet is lost because the ring is
> full?
> > >
> > > BTW, the IGB driver also has this defect.
> > >
> >
> > Note how I read the rx registers first? So it will always clear the value.
> > That should unlock the value for the next rx stamp packet.
> 
> 1. Hw recognizes ptp event packet, locks time stamp 2. Hw drops packet because
> queue is full 3. No more time stamps are ever generated
> 
> Can this happen? The docs seems to say it can.
> 
> Richard

I believe this very rare case might be possible, but I don't think that checking the ptp seqid will fix anything. In normal cases, hardware latches Rx packet timestamp, then the ptp packet goes into the queue and we process it shortly after. Before we process that packet there will never be another packet in the queue that needs a timestamp. We know this because the hardware stops timestamping until we unlatch the RX registers. This should mean we don't need to check the sequence ID, and spending time doing it would never fix the issue you are talking about.

The issue is for when a packet is timestamped and then never reaches the queue. Then the rx stamp registers are locked for good, because we never clear them, and hardware would never timestamp another receive packet. I don't know a good solution to this, except to clear the registers periodically. Do you have any suggestions?

- Jake

RE: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Keller, Jacob E]

> -----Original Message-----
> From: Richard Cochran [mailto:richardcochran@gmail.com]
> Sent: Thursday, May 10, 2012 10:15 PM
> To: Keller, Jacob E
> Cc: Kirsher, Jeffrey T; davem@davemloft.net; netdev@vger.kernel.org;
> gospo@redhat.com; sassmann@redhat.com
> Subject: Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware
> Clock (PHC)
> 
> On Thu, May 10, 2012 at 09:53:18PM +0000, Keller, Jacob E wrote:
> > > > +	/*
> > > > +	 * If this bit is set, then the RX registers contain the time
> stamp. No
> > > > +	 * other packet will be time stamped until we read these
> registers, so
> > > > +	 * read the registers to make them available again. Because only
> one
> > > > +	 * packet can be time stamped at a time, we know that the
> register
> > > > +	 * values must belong to this one here and therefore we don't
> need to
> > > > +	 * compare any of the additional attributes stored for it.
> > >
> > > I suspect that this assumption is wrong. What happens if the time
> > > stamping logic locks a value but the packet is lost because the ring is
> full?
> > >
> > > BTW, the IGB driver also has this defect.
> > >
> >
> > Note how I read the rx registers first? So it will always clear the value.
> > That should unlock the value for the next rx stamp packet.
> 
> 1. Hw recognizes ptp event packet, locks time stamp 2. Hw drops packet because
> queue is full 3. No more time stamps are ever generated
> 
> Can this happen? The docs seems to say it can.
> 
> Richard

Sorry for the spam here, but I looked at the ixgbe code, and found a solution. When ptp4l discovers a missing rx timestamp, it faults and then waits for 15 seconds until the fault is cleared. After this, it reopens the socket, and reruns the hwtstamp ioctl. This function actually does clear the tx/rx timestamps (just in case) So after a fault the values should end up being reset. Is this good enough?

- Jake

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Richard Cochran]

On Fri, May 11, 2012 at 07:34:12PM +0000, Keller, Jacob E wrote:
> 
> Sorry for the spam here, but I looked at the ixgbe code, and found a
> solution. When ptp4l discovers a missing rx timestamp, it faults and
> then waits for 15 seconds until the fault is cleared. After this, it
> reopens the socket, and reruns the hwtstamp ioctl. This function
> actually does clear the tx/rx timestamps (just in case) So after a
> fault the values should end up being reset. Is this good enough?

Yes, for ptp4l, that will work, but this program is acting extremely
defensively in that situation. Not every user space program will
necessarily behave in the same way. The classic ptpd just ignores
packets with missing time stamps, IRRC.

Richard

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Richard Cochran]

On Fri, May 11, 2012 at 07:23:44PM +0000, Keller, Jacob E wrote:
> 
>
> I believe this very rare case might be possible, but I don't think
> that checking the ptp seqid will fix anything. In normal cases,
> hardware latches Rx packet timestamp, then the ptp packet goes into
> the queue and we process it shortly after. Before we process that
> packet there will never be another packet in the queue that needs a
> timestamp. We know this because the hardware stops timestamping
> until we unlatch the RX registers. This should mean we don't need to
> check the sequence ID, and spending time doing it would never fix
> the issue you are talking about.
>
> The issue is for when a packet is timestamped and then never reaches
> the queue. Then the rx stamp registers are locked for good, because
> we never clear them, and hardware would never timestamp another
> receive packet. I don't know a good solution to this, except to
> clear the registers periodically. Do you have any suggestions?

Well, one solution would be to check every received packet with the
BPF in ptp_classify.h (whenever Rx time stamping is enabled).

When the driver finds an event packet in the Rx queue, and
TSYNCRXCTL[RXTT] is set, it reads out the time stamp along with
RXSATRL/H. If the fields match, then add the time stamp to the skb.

[ Or perhaps instead of using RXSATRL/H, just use the descriptor bit.
  If *not* set, then the time stamp does not belong to this packet. ]

HTH,
Richard

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Jacob Keller]

On 05/11/2012 10:34 PM, Richard Cochran wrote:
> On Fri, May 11, 2012 at 07:23:44PM +0000, Keller, Jacob E wrote:
>>
>>
>> I believe this very rare case might be possible, but I don't think
>> that checking the ptp seqid will fix anything. In normal cases,
>> hardware latches Rx packet timestamp, then the ptp packet goes into
>> the queue and we process it shortly after. Before we process that
>> packet there will never be another packet in the queue that needs a
>> timestamp. We know this because the hardware stops timestamping
>> until we unlatch the RX registers. This should mean we don't need to
>> check the sequence ID, and spending time doing it would never fix
>> the issue you are talking about.
>>
>> The issue is for when a packet is timestamped and then never reaches
>> the queue. Then the rx stamp registers are locked for good, because
>> we never clear them, and hardware would never timestamp another
>> receive packet. I don't know a good solution to this, except to
>> clear the registers periodically. Do you have any suggestions?
>
> Well, one solution would be to check every received packet with the
> BPF in ptp_classify.h (whenever Rx time stamping is enabled).
>
> When the driver finds an event packet in the Rx queue, and
> TSYNCRXCTL[RXTT] is set, it reads out the time stamp along with
> RXSATRL/H. If the fields match, then add the time stamp to the skb.
>
> [ Or perhaps instead of using RXSATRL/H, just use the descriptor bit.
>    If *not* set, then the time stamp does not belong to this packet. ]
>
> HTH,
> Richard

Ok, this sounds like a good plan. Considering that the device already 
doesn't allow timestamping of other types of packets, so it doesn't need 
to be general purpose.

Am I correct in thinking all I need to do is check the type and if it 
matches the currently configured rx timestamp mode, then double check 
the bit for whether a timestamp is available, and whether the descriptor 
had a timestamp bit enabled?

Thanks

- Jake

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Jacob Keller]

On 05/11/2012 10:34 PM, Richard Cochran wrote:
> On Fri, May 11, 2012 at 07:23:44PM +0000, Keller, Jacob E wrote:
>>
>>
>> I believe this very rare case might be possible, but I don't think
>> that checking the ptp seqid will fix anything. In normal cases,
>> hardware latches Rx packet timestamp, then the ptp packet goes into
>> the queue and we process it shortly after. Before we process that
>> packet there will never be another packet in the queue that needs a
>> timestamp. We know this because the hardware stops timestamping
>> until we unlatch the RX registers. This should mean we don't need to
>> check the sequence ID, and spending time doing it would never fix
>> the issue you are talking about.
>>
>> The issue is for when a packet is timestamped and then never reaches
>> the queue. Then the rx stamp registers are locked for good, because
>> we never clear them, and hardware would never timestamp another
>> receive packet. I don't know a good solution to this, except to
>> clear the registers periodically. Do you have any suggestions?
>
> Well, one solution would be to check every received packet with the
> BPF in ptp_classify.h (whenever Rx time stamping is enabled).
>
> When the driver finds an event packet in the Rx queue, and
> TSYNCRXCTL[RXTT] is set, it reads out the time stamp along with
> RXSATRL/H. If the fields match, then add the time stamp to the skb.
>
> [ Or perhaps instead of using RXSATRL/H, just use the descriptor bit.
>    If *not* set, then the time stamp does not belong to this packet. ]
>
> HTH,
> Richard

sk_run_filter is always returning PTP_CLASS_NONE (0), instead of the 
correct type. Do you know what I am doing wrong? I can send you the 
patch I am working with...

- Jake

Re: [net-next 06/12] ixgbe: Hardware Timestamping + PTP Hardware Clock (PHC)

[Richard Cochran]

On Mon, May 14, 2012 at 03:42:03PM -0700, Jacob Keller wrote:
> 
> sk_run_filter is always returning PTP_CLASS_NONE (0), instead of the
> correct type. Do you know what I am doing wrong? I can send you the
> patch I am working with...

The filter need to run with a pointer to the DST MAC. Perhaps you are
passing it the address of padding or Layer 3?

I will take a look at the patch...

Thanks,
Richard

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series contains updates to ixgbe only.  Only change to this series
is I dropped the "ixgbe: Add support for pipeline reset" due to
change requested by Martin Josefsson.

The following are changes since commit d94ce9b283736a876b2e6dec665c68e5e8b5d55e:
  ipv4: 16 slots in initial fib_info hash table
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Alexander Duyck (7):
  ixgbe: Add support for IPv6 and UDP to ixgbe_get_headlen
  ixgbe: Add support for tracking the default user priority to SR-IOV
  ixgbe: Add support for GET_QUEUES message to get DCB configuration
  ixgbe: Enable support for VF API version 1.1 in the PF.
  ixgbevf: Add VF DCB + SR-IOV support
  ixgbe: Drop unnecessary addition from ixgbe_set_rx_buffer_len
  ixgbe: Fix possible memory leak in ixgbe_set_ringparam

Don Skidmore (1):
  ixgbe: Add function ixgbe_reset_pipeline_82599

Emil Tantilov (1):
  ixgbe: add WOL support for new subdevice id

Jacob Keller (1):
  ixgbe: (PTP) refactor init, cyclecounter and reset

Tushar Dave (1):
  ixgbe: Correcting small packet padding

Wei Yongjun (1):
  ixgbe: using is_zero_ether_addr() to simplify the code

 drivers/net/ethernet/intel/ixgbe/ixgbe.h          |   6 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_82599.c    |  44 ++++++
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.c   |   3 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_common.h   |   1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c  | 102 +++++++-------
 drivers/net/ethernet/intel/ixgbe/ixgbe_main.c     |  66 +++++++--
 drivers/net/ethernet/intel/ixgbe/ixgbe_mbx.h      |  10 ++
 drivers/net/ethernet/intel/ixgbe/ixgbe_ptp.c      | 109 +++++++--------
 drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c    | 124 +++++++++++++----
 drivers/net/ethernet/intel/ixgbe/ixgbe_type.h     |   1 +
 drivers/net/ethernet/intel/ixgbevf/defines.h      |   7 +-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |   4 +-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c | 159 +++++++++++++++++++++-
 drivers/net/ethernet/intel/ixgbevf/mbx.h          |  10 ++
 drivers/net/ethernet/intel/ixgbevf/vf.c           |  58 ++++++++
 drivers/net/ethernet/intel/ixgbevf/vf.h           |   2 +
 16 files changed, 539 insertions(+), 167 deletions(-)

-- 
1.7.11.7

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Mon, 22 Oct 2012 21:26:13 -0700

> This series contains updates to ixgbe only.  Only change to this series
> is I dropped the "ixgbe: Add support for pipeline reset" due to
> change requested by Martin Josefsson.
> 
> The following are changes since commit d94ce9b283736a876b2e6dec665c68e5e8b5d55e:
>   ipv4: 16 slots in initial fib_info hash table
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Pulled, thanks Jeff.

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series contains updates to e1000e, ixgbe and ixgbevf.

Majority of the changes are against e1000e (from Bruce Allan).
Bruce adds additional error handling on PHY register access, as
well as improve slow performance on 82579 when connected to a
10Mbit hub.  In addition, fixes LED blink logic for cathode
LED design.  Most notable is added EEE support which is enabled
by default and the added support for LTR on I217/I218.

The ixgbe and ixgbevf from Greg Rose changes the VM so that if a user
does not assign a MAC address, the MAC address is set to all zeros
instead of a random MAC address.  This ensures that we always know when
we have a random address and udev won't get upset about it.

The following are changes since commit 8b49a4c75965ed157e21450d23dcadd6b27c1aa3:
  bnx2x: fix compilation without CONFIG_BNX2X_SRIOV
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Bruce Allan (10):
  e1000e: additional error handling on PHY register accesses
  e1000e: slow performance between two 82579 connected via 10Mbit hub
  e1000e: fix LED blink logic for designs with LEDs driven by cathode
  e1000e: long access timeouts when I217/I218 MAC and PHY are out of
    sync
  e1000e: EEE capability advertisement not set/disabled as required
  e1000e: enable EEE by default
  e1000e: add support for LTR on I217/I218
  e1000e: cleanup unused defines
  e1000e: increase driver version number
  e1000e: fix scheduling while atomic bugs

Greg Rose (2):
  ixgbe: Don't give VFs random MAC addresses
  ixgbevf: Adjust to handle unassigned MAC address from PF

 drivers/net/ethernet/intel/e1000e/defines.h       |   2 +
 drivers/net/ethernet/intel/e1000e/e1000.h         |   3 +
 drivers/net/ethernet/intel/e1000e/ethtool.c       |  63 ++-----
 drivers/net/ethernet/intel/e1000e/ich8lan.c       | 213 ++++++++++++++++++----
 drivers/net/ethernet/intel/e1000e/ich8lan.h       |  11 +-
 drivers/net/ethernet/intel/e1000e/mac.c           |  27 ++-
 drivers/net/ethernet/intel/e1000e/netdev.c        |  38 +++-
 drivers/net/ethernet/intel/e1000e/phy.c           |  20 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_sriov.c    |  23 ++-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c |  21 ++-
 drivers/net/ethernet/intel/ixgbevf/vf.c           |   7 +-
 11 files changed, 315 insertions(+), 113 deletions(-)

-- 
1.7.11.7

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Thu, 28 Mar 2013 02:00:03 -0700

> This series contains updates to e1000e, ixgbe and ixgbevf.
> 
> Majority of the changes are against e1000e (from Bruce Allan).
> Bruce adds additional error handling on PHY register access, as
> well as improve slow performance on 82579 when connected to a
> 10Mbit hub.  In addition, fixes LED blink logic for cathode
> LED design.  Most notable is added EEE support which is enabled
> by default and the added support for LTR on I217/I218.
> 
> The ixgbe and ixgbevf from Greg Rose changes the VM so that if a user
> does not assign a MAC address, the MAC address is set to all zeros
> instead of a random MAC address.  This ensures that we always know when
> we have a random address and udev won't get upset about it.
> 
> The following are changes since commit 8b49a4c75965ed157e21450d23dcadd6b27c1aa3:
>   bnx2x: fix compilation without CONFIG_BNX2X_SRIOV
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Pulled, thanks Jeff.

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series contains updates to e1000e, igb and ixgbe.

Bruce Allan provide 2 minor cleanups for e1000e to resolve whitespace
issues and build warnings about unused parameters.

Carolyn provides a couple of fixes for igb, one being a fix for a
possible panic when the interface is down and receive traffic
arrives.  The second fix resolves an issue on newer parts which have
multiple checksum fields and set_ethtool was only checking to update
the first checksum of the NVM image.

Akeem provides majority of the changes in this patch set.  Akeem
provides a fix for e1000e on an issue reported from the community to
resolve the issue of unlocking swflag_mutex for 82574 and 82583
devices even if the hardware semaphore was successfully acquired.
The other patches from Akeem are against igb, where he adds support
SFP module discovery, LED blink mechanism for devices using cathodes,
LED support for i210/i211 parts and cleanup of a i2c function which
was not being used.

Matthew provides an update for igb to support a more accurate check
for a PTP RX hang.

Amir provides a patch for ixgbe to set the software prio_tc values at
initialization to the hardware setting to remove the need to reset the
device at the first time we call ixgbe_dcbnl_ieee_setets.

The following are changes since commit 1e18583adccfc122b5d6415cfe4bf1826c370f4e:
  ThunderLAN: remove is_eisa flag
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Akeem G. Abodunrin (6):
  e1000e: Release mutex lock only if it has been initially acquired
  igb: Changed LEDs blink mechanism to include designs using cathode
  igb: Support for SFP modules discovery
  igb: SerDes flow control setting
  igb: Implementation of i210/i211 LED support
  igb: Removed unused i2c function

Amir Hanania (1):
  IXGBE: Set the SW prio_tc values at initialization to the HW setting.

Bruce Allan (2):
  e1000e: cleanup whitespace
  e1000e: prevent warning from -Wunused-parameter

Carolyn Wyborny (2):
  igb: Fix set_ethtool function to call update nvm for entire image
  igb: Fix possible panic caused by Rx traffic arrival while interface
    is down

Matthew Vick (1):
  igb: Add update to last_rx_timestamp in Rx rings

 drivers/net/ethernet/intel/e1000e/80003es2lan.c    |  24 ++--
 drivers/net/ethernet/intel/e1000e/82571.c          |  30 ++---
 drivers/net/ethernet/intel/e1000e/ethtool.c        |  34 +++---
 drivers/net/ethernet/intel/e1000e/hw.h             |  34 +++---
 drivers/net/ethernet/intel/e1000e/ich8lan.c        |  62 +++++------
 drivers/net/ethernet/intel/e1000e/netdev.c         |  55 +++++----
 drivers/net/ethernet/intel/e1000e/nvm.c            |   1 -
 drivers/net/ethernet/intel/e1000e/phy.c            |  22 ++--
 drivers/net/ethernet/intel/igb/e1000_82575.c       | 120 +++++++++++++++++++-
 drivers/net/ethernet/intel/igb/e1000_defines.h     |  36 +++---
 drivers/net/ethernet/intel/igb/e1000_hw.h          |   2 +
 drivers/net/ethernet/intel/igb/e1000_i210.h        |   6 +-
 drivers/net/ethernet/intel/igb/e1000_mac.c         |  45 ++++++--
 drivers/net/ethernet/intel/igb/e1000_phy.c         | 124 +++++++++++++++++++++
 drivers/net/ethernet/intel/igb/e1000_phy.h         |  20 ++++
 drivers/net/ethernet/intel/igb/igb.h               |  14 +--
 drivers/net/ethernet/intel/igb/igb_ethtool.c       |  74 ++++++------
 drivers/net/ethernet/intel/igb/igb_main.c          |   9 +-
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.c       |  23 ++++
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb.h       |   2 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_82599.h |   1 +
 drivers/net/ethernet/intel/ixgbe/ixgbe_dcb_nl.c    |   3 +
 22 files changed, 526 insertions(+), 215 deletions(-)

-- 
1.7.11.7

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Tue, 21 May 2013 15:03:08 -0700

> This series contains updates to e1000e, igb and ixgbe.
> 
> Bruce Allan provide 2 minor cleanups for e1000e to resolve whitespace
> issues and build warnings about unused parameters.
> 
> Carolyn provides a couple of fixes for igb, one being a fix for a
> possible panic when the interface is down and receive traffic
> arrives.  The second fix resolves an issue on newer parts which have
> multiple checksum fields and set_ethtool was only checking to update
> the first checksum of the NVM image.
> 
> Akeem provides majority of the changes in this patch set.  Akeem
> provides a fix for e1000e on an issue reported from the community to
> resolve the issue of unlocking swflag_mutex for 82574 and 82583
> devices even if the hardware semaphore was successfully acquired.
> The other patches from Akeem are against igb, where he adds support
> SFP module discovery, LED blink mechanism for devices using cathodes,
> LED support for i210/i211 parts and cleanup of a i2c function which
> was not being used.
> 
> Matthew provides an update for igb to support a more accurate check
> for a PTP RX hang.
> 
> Amir provides a patch for ixgbe to set the software prio_tc values at
> initialization to the hardware setting to remove the need to reset the
> device at the first time we call ixgbe_dcbnl_ieee_setets.
> 
> The following are changes since commit 1e18583adccfc122b5d6415cfe4bf1826c370f4e:
>   ThunderLAN: remove is_eisa flag
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Pulled, excellent merge commit message, thanks!

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series contains updates to e100 and e1000e.

The e100 patch from Andy simply updates the netif_printk() to use
%*ph to dump small buffers.

The changes to e1000e include a fix from Dean Nelson to resolve a
issue where a pci_clear_master() was accidentally dropped during a
conflict resolution. Wei Young provides 2 patches, one removes an
assignment of the default ring size because it was a duplicate. The
second changes the packet split receive structure to use
PS_PAGE_BUFFERS macro for the length so that problems won't occur
when the length is changed.

The remaining patches for e1000e are from Bruce Allan, where he
provides a number of fixes and updates for I218.  In addition, a
fix for 82583 which can disappear off the PCIe bus, to resolve this,
disable ASPM L1.  Bruce also provides a fix to a previous commit
(commit e60b22c5b7 e1000e: fix accessing to suspended device) so that
devices are only taken out of runtime power management for those
ethtool operations that must access device registers.

The following are changes since commit dcfe8048de66c3468060c8a2ec2c04ae3725d002:
  bonding: remove bond_resend_igmp_join_requests read_unlock leftover
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Andy Shevchenko (1):
  e100: dump small buffers via %*ph

Bruce Allan (8):
  e1000e: disable ASPM L1 on 82583
  e1000e: iAMT connections drop on driver unload when jumbo frames
    enabled
  e1000e: low throughput using 4K jumbos on I218
  e1000e: Tx hang on I218 when linked at 100Half and slow response at
    10Mbps
  e1000e: ethtool unnecessarily takes device out of RPM suspend
  e1000e: enable support for new device IDs
  e1000e: do not resume device from RPM suspend to read PHY status
    registers
  e1000e: fix I217/I218 PHY initialization flow

Dean Nelson (1):
  e1000e: restore call to pci_clear_master()

Wei Yang (2):
  e1000e: Remove duplicate assignment of default rx/tx ring size
  e1000e: Use marco instead of digit for defining
    e1000_rx_desc_packet_split

 drivers/net/ethernet/intel/e100.c           |  15 ++-
 drivers/net/ethernet/intel/e1000e/82571.c   |   1 +
 drivers/net/ethernet/intel/e1000e/e1000.h   |   3 -
 drivers/net/ethernet/intel/e1000e/ethtool.c | 105 +++++++++++++++------
 drivers/net/ethernet/intel/e1000e/hw.h      |  10 +-
 drivers/net/ethernet/intel/e1000e/ich8lan.c | 140 +++++++++++++++++++---------
 drivers/net/ethernet/intel/e1000e/ich8lan.h |   6 ++
 drivers/net/ethernet/intel/e1000e/netdev.c  |  38 ++++----
 8 files changed, 217 insertions(+), 101 deletions(-)

-- 
1.7.11.7

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Sun, 28 Jul 2013 04:41:04 -0700

> This series contains updates to e100 and e1000e.
> 
> The e100 patch from Andy simply updates the netif_printk() to use
> %*ph to dump small buffers.
> 
> The changes to e1000e include a fix from Dean Nelson to resolve a
> issue where a pci_clear_master() was accidentally dropped during a
> conflict resolution. Wei Young provides 2 patches, one removes an
> assignment of the default ring size because it was a duplicate. The
> second changes the packet split receive structure to use
> PS_PAGE_BUFFERS macro for the length so that problems won't occur
> when the length is changed.
> 
> The remaining patches for e1000e are from Bruce Allan, where he
> provides a number of fixes and updates for I218.  In addition, a
> fix for 82583 which can disappear off the PCIe bus, to resolve this,
> disable ASPM L1.  Bruce also provides a fix to a previous commit
> (commit e60b22c5b7 e1000e: fix accessing to suspended device) so that
> devices are only taken out of runtime power management for those
> ethtool operations that must access device registers.
> 
> The following are changes since commit dcfe8048de66c3468060c8a2ec2c04ae3725d002:
>   bonding: remove bond_resend_igmp_join_requests read_unlock leftover
> and are available in the git repository at:
>   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Pulled, thanks a lot Jeff.

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

[-- Attachment #1: Type: text/plain, Size: 1764 bytes --]

On Sun, 2013-07-28 at 13:21 -0700, David Miller wrote:
> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> Date: Sun, 28 Jul 2013 04:41:04 -0700
> 
> > This series contains updates to e100 and e1000e.
> > 
> > The e100 patch from Andy simply updates the netif_printk() to use
> > %*ph to dump small buffers.
> > 
> > The changes to e1000e include a fix from Dean Nelson to resolve a
> > issue where a pci_clear_master() was accidentally dropped during a
> > conflict resolution. Wei Young provides 2 patches, one removes an
> > assignment of the default ring size because it was a duplicate. The
> > second changes the packet split receive structure to use
> > PS_PAGE_BUFFERS macro for the length so that problems won't occur
> > when the length is changed.
> > 
> > The remaining patches for e1000e are from Bruce Allan, where he
> > provides a number of fixes and updates for I218.  In addition, a
> > fix for 82583 which can disappear off the PCIe bus, to resolve this,
> > disable ASPM L1.  Bruce also provides a fix to a previous commit
> > (commit e60b22c5b7 e1000e: fix accessing to suspended device) so that
> > devices are only taken out of runtime power management for those
> > ethtool operations that must access device registers.
> > 
> > The following are changes since commit dcfe8048de66c3468060c8a2ec2c04ae3725d002:
> >   bonding: remove bond_resend_igmp_join_requests read_unlock leftover
> > and are available in the git repository at:
> >   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
> 
> Pulled, thanks a lot Jeff.

It looks like you pulled my net tree instead of my net-next tree.
Fortunately, I did not have anything applied to my net tree that had not
already been sent out.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Sun, 28 Jul 2013 16:11:38 -0700

> On Sun, 2013-07-28 at 13:21 -0700, David Miller wrote:
>> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
>> Date: Sun, 28 Jul 2013 04:41:04 -0700
>> 
>> > This series contains updates to e100 and e1000e.
>> > 
>> > The e100 patch from Andy simply updates the netif_printk() to use
>> > %*ph to dump small buffers.
>> > 
>> > The changes to e1000e include a fix from Dean Nelson to resolve a
>> > issue where a pci_clear_master() was accidentally dropped during a
>> > conflict resolution. Wei Young provides 2 patches, one removes an
>> > assignment of the default ring size because it was a duplicate. The
>> > second changes the packet split receive structure to use
>> > PS_PAGE_BUFFERS macro for the length so that problems won't occur
>> > when the length is changed.
>> > 
>> > The remaining patches for e1000e are from Bruce Allan, where he
>> > provides a number of fixes and updates for I218.  In addition, a
>> > fix for 82583 which can disappear off the PCIe bus, to resolve this,
>> > disable ASPM L1.  Bruce also provides a fix to a previous commit
>> > (commit e60b22c5b7 e1000e: fix accessing to suspended device) so that
>> > devices are only taken out of runtime power management for those
>> > ethtool operations that must access device registers.
>> > 
>> > The following are changes since commit dcfe8048de66c3468060c8a2ec2c04ae3725d002:
>> >   bonding: remove bond_resend_igmp_join_requests read_unlock leftover
>> > and are available in the git repository at:
>> >   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
>> 
>> Pulled, thanks a lot Jeff.
> 
> It looks like you pulled my net tree instead of my net-next tree.
> Fortunately, I did not have anything applied to my net tree that had not
> already been sent out.

No, really, I did this:

git pull git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

[-- Attachment #1: Type: text/plain, Size: 2425 bytes --]

On Sun, 2013-07-28 at 17:20 -0700, David Miller wrote:
> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> Date: Sun, 28 Jul 2013 16:11:38 -0700
> 
> > On Sun, 2013-07-28 at 13:21 -0700, David Miller wrote:
> >> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> >> Date: Sun, 28 Jul 2013 04:41:04 -0700
> >> 
> >> > This series contains updates to e100 and e1000e.
> >> > 
> >> > The e100 patch from Andy simply updates the netif_printk() to use
> >> > %*ph to dump small buffers.
> >> > 
> >> > The changes to e1000e include a fix from Dean Nelson to resolve a
> >> > issue where a pci_clear_master() was accidentally dropped during a
> >> > conflict resolution. Wei Young provides 2 patches, one removes an
> >> > assignment of the default ring size because it was a duplicate. The
> >> > second changes the packet split receive structure to use
> >> > PS_PAGE_BUFFERS macro for the length so that problems won't occur
> >> > when the length is changed.
> >> > 
> >> > The remaining patches for e1000e are from Bruce Allan, where he
> >> > provides a number of fixes and updates for I218.  In addition, a
> >> > fix for 82583 which can disappear off the PCIe bus, to resolve this,
> >> > disable ASPM L1.  Bruce also provides a fix to a previous commit
> >> > (commit e60b22c5b7 e1000e: fix accessing to suspended device) so that
> >> > devices are only taken out of runtime power management for those
> >> > ethtool operations that must access device registers.
> >> > 
> >> > The following are changes since commit dcfe8048de66c3468060c8a2ec2c04ae3725d002:
> >> >   bonding: remove bond_resend_igmp_join_requests read_unlock leftover
> >> > and are available in the git repository at:
> >> >   git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master
> >> 
> >> Pulled, thanks a lot Jeff.
> > 
> > It looks like you pulled my net tree instead of my net-next tree.
> > Fortunately, I did not have anything applied to my net tree that had not
> > already been sent out.
> 
> No, really, I did this:
> 
> git pull git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Disregard, it was me misreading my git log in my net tree after updating
it against your latest tree.

When will your kernel.org net-next be updated?  I ask because I have a
series for ixgbe I am putting together and wanted to make sure it is
against the latest net-next.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Sun, 28 Jul 2013 19:08:55 -0700

> When will your kernel.org net-next be updated?  I ask because I have a
> series for ixgbe I am putting together and wanted to make sure it is
> against the latest net-next.

Are you asking when I will next merge 'net' into 'net-next'?  If so,
I'll likely do it some time tomorrow.

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

[-- Attachment #1: Type: text/plain, Size: 1010 bytes --]

On Mon, 2013-07-29 at 00:34 -0700, David Miller wrote:
> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> Date: Sun, 28 Jul 2013 19:08:55 -0700
> 
> > When will your kernel.org net-next be updated?  I ask because I have a
> > series for ixgbe I am putting together and wanted to make sure it is
> > against the latest net-next.
> 
> Are you asking when I will next merge 'net' into 'net-next'?  If so,
> I'll likely do it some time tomorrow.

I am not worried about that.  I just cloned your net-next tree and did
not see your pull of my net-next tree which had the e100 and e1000e
patches.  I am working on creating another series against net-next and
the tip on kernel.org is:

commit 9d4a0314642918cbda9ed4012df51e8df608fce6
Author: Hannes Frederic Sowa <hannes@stressinduktion.org>
Date:   Fri Jul 26 17:05:16 2013 +0200

    ipv4, ipv6: send igmpv3/mld packets with TC_PRIO_CONTROL

which is several days old.  So I was not sure if I just make my series
against my net-next tip.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Mon, 29 Jul 2013 01:34:45 -0700

> On Mon, 2013-07-29 at 00:34 -0700, David Miller wrote:
>> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
>> Date: Sun, 28 Jul 2013 19:08:55 -0700
>> 
>> > When will your kernel.org net-next be updated?  I ask because I have a
>> > series for ixgbe I am putting together and wanted to make sure it is
>> > against the latest net-next.
>> 
>> Are you asking when I will next merge 'net' into 'net-next'?  If so,
>> I'll likely do it some time tomorrow.
> 
> I am not worried about that.  I just cloned your net-next tree and did
> not see your pull of my net-next tree which had the e100 and e1000e
> patches.  I am working on creating another series against net-next and
> the tip on kernel.org is:
> 
> commit 9d4a0314642918cbda9ed4012df51e8df608fce6
> Author: Hannes Frederic Sowa <hannes@stressinduktion.org>
> Date:   Fri Jul 26 17:05:16 2013 +0200
> 
>     ipv4, ipv6: send igmpv3/mld packets with TC_PRIO_CONTROL
> 
> which is several days old.  So I was not sure if I just make my series
> against my net-next tip.

Weird, it should have all of your changes.

Tip is currently:

http://git.kernel.org/cgit/linux/kernel/git/davem/net-next.git/commit/?id=fe6f700d6cbb7e8a61711e325f53d9c9e0a42a4c

net/mlx4_core: Respond to operation request by firmware

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

[-- Attachment #1: Type: text/plain, Size: 1541 bytes --]

On Mon, 2013-07-29 at 01:46 -0700, David Miller wrote:
> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> Date: Mon, 29 Jul 2013 01:34:45 -0700
> 
> > On Mon, 2013-07-29 at 00:34 -0700, David Miller wrote:
> >> From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
> >> Date: Sun, 28 Jul 2013 19:08:55 -0700
> >> 
> >> > When will your kernel.org net-next be updated?  I ask because I have a
> >> > series for ixgbe I am putting together and wanted to make sure it is
> >> > against the latest net-next.
> >> 
> >> Are you asking when I will next merge 'net' into 'net-next'?  If so,
> >> I'll likely do it some time tomorrow.
> > 
> > I am not worried about that.  I just cloned your net-next tree and did
> > not see your pull of my net-next tree which had the e100 and e1000e
> > patches.  I am working on creating another series against net-next and
> > the tip on kernel.org is:
> > 
> > commit 9d4a0314642918cbda9ed4012df51e8df608fce6
> > Author: Hannes Frederic Sowa <hannes@stressinduktion.org>
> > Date:   Fri Jul 26 17:05:16 2013 +0200
> > 
> >     ipv4, ipv6: send igmpv3/mld packets with TC_PRIO_CONTROL
> > 
> > which is several days old.  So I was not sure if I just make my series
> > against my net-next tip.
> 
> Weird, it should have all of your changes.
> 
> Tip is currently:
> 
> http://git.kernel.org/cgit/linux/kernel/git/davem/net-next.git/commit/?id=fe6f700d6cbb7e8a61711e325f53d9c9e0a42a4c
> 
> net/mlx4_core: Respond to operation request by firmware

All looks fine now, thanks!

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

[net-next 00/12][pull request] Intel Wired LAN Driver Updates

[Jeff Kirsher]

This series contains updates to i40e, igb, ixgbe and ixgbevf.

Shannon provides a couple of i40e patches, first restricts the ethtool
diag test messages by using netif_info() macro to when the hardware
bit is enabled in the message level netdev message mask.  Second
provides a fix for when there is an out-of-range descriptor request.

Kamil provides a fix for i40e by updating the loopback enum types and
add information about the current loopback mode to data returned from
get_link_info().

Jesse provides a fix for i40e define name that was being mis-used.
I40E_ITR_NONE was being used as an ITRN register index by accident
because it was easily associated with the i40e Rx ITR and friends
defines, when it should be associated with the DYN_CTL register sets.

Jacob provides an update for ixgbevf Kconfig description since the VF
driver supports more than just the 82599 device.

Don and Alex provide a cleanup patch for ixgbe to make it where head,
tail, next to clean and next to use are all reset in a single function
for both Tx and Rx path.  Before, the code for this was spread out over
several areas which made it difficult to track what the values were for
each of the values.

Carolyn provides two igb patches to add a media switching feature for
i354 PHY's and new Media Auto Sense for 82580 devices only.

Aaron Sierra provides a fix for igb to resolve an issue with the 64-bit
PCI addresses being truncated because the return values of
pci_resource_start() and pci_resouce_end() were being cast to unsigned
long.

Guenter Roeck provides two igb patches, first simplifies the code by
attaching the hwmon sysfs attributes to hwmon device instead of the
PCI device.  Second fixes the temperature sensor attribute index by
setting it to 1 instead of 0 (per hwmon ABI).

The following are changes since commit 5824d2d16d39a9c57aa3bacf955fbd0c2134061f:
  bgmac: connect to PHY and make use of PHY device
and are available in the git repository at:
  git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-next master

Aaron Sierra (1):
  igb: Support ports mapped in 64-bit PCI space

Carolyn Wyborny (2):
  igb: Add media switching feature for i354 PHY's
  igb: Add new feature Media Auto Sense for 82580 devices only

Catherine Sullivan (1):
  i40e: Bump version number

Don Skidmore (1):
  ixgbe: Focus config of head, tail ntc, and ntu all into a single
    function

Guenter Roeck (2):
  igb: Convert to use devm_hwmon_device_register_with_groups
  igb: Start temperature sensor attribute index with 1

Jacob Keller (1):
  ixgbevf: update Kconfig description

Jesse Brandeburg (1):
  i40e: remove and fix confusing define name

Kamil Krawczyk (1):
  i40e: loopback info and set loopback fix

Shannon Nelson (2):
  i40e: restrict diag test messages
  i40e: complain about out-of-range descriptor request

 drivers/net/ethernet/intel/Kconfig                |   8 +-
 drivers/net/ethernet/intel/i40e/i40e_common.c     |   1 +
 drivers/net/ethernet/intel/i40e/i40e_diag.h       |   8 +-
 drivers/net/ethernet/intel/i40e/i40e_ethtool.c    |  41 +++--
 drivers/net/ethernet/intel/i40e/i40e_main.c       |   2 +-
 drivers/net/ethernet/intel/i40e/i40e_txrx.h       |  21 ++-
 drivers/net/ethernet/intel/i40e/i40e_type.h       |   1 +
 drivers/net/ethernet/intel/igb/e1000_82575.c      |  89 +++++++++
 drivers/net/ethernet/intel/igb/e1000_defines.h    |  16 ++
 drivers/net/ethernet/intel/igb/e1000_hw.h         |   3 +
 drivers/net/ethernet/intel/igb/igb.h              |  19 +-
 drivers/net/ethernet/intel/igb/igb_ethtool.c      |   4 +
 drivers/net/ethernet/intel/igb/igb_hwmon.c        | 108 ++++++-----
 drivers/net/ethernet/intel/igb/igb_main.c         | 209 +++++++++++++++++++++-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf.h      |   4 +-
 drivers/net/ethernet/intel/ixgbevf/ixgbevf_main.c |  55 ++----
 16 files changed, 453 insertions(+), 136 deletions(-)

-- 
1.8.3.1

Re: [net-next 00/12][pull request] Intel Wired LAN Driver Updates

[David Miller]

From: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Date: Tue, 10 Dec 2013 02:22:29 -0800

> This series contains updates to i40e, igb, ixgbe and ixgbevf.

Looks fine, pulled, thanks Jeff.