Commit 547669d4 ("tcp: xps: fix reordering issues") added
unexpected reorders in case netem is used in a MQ setup for high
performance test bed.

ETH=eth0
tc qd del dev $ETH root 2>/dev/null
tc qd add dev $ETH root handle 1: mq
for i in `seq 1 32`
do
 tc qd add dev $ETH parent 1:$i netem delay 100ms
done

As all tcp packets are orphaned by netem, TCP stack believes it can
set skb->ooo_okay on all packets.

In order to allow producers to send more packets, we want to
keep sk_wmem_alloc from reaching sk_sndbuf limit.

We can do that by accounting one byte per skb in netem queues,
so that TCP stack is not fooled too much.

Tested:

With above MQ/netem setup, scaling number of concurrent flows gives
linear results and no reorders/retransmits

lpq83:~# for n in 1 10 20 30 40 50 60 70 80 90 100
 do echo -n "n:$n " ; ./super_netperf $n -H 10.7.7.84; done
n:1 198.46
n:10 2002.69
n:20 4000.98
n:30 6006.35
n:40 8020.93
n:50 10032.3
n:60 12081.9
n:70 13971.3
n:80 16009.7
n:90 17117.3
n:100 17425.5
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Current net name space has only one genid for both IPv4 and IPv6, it has below
drawbacks:

- Add/delete an IPv4 address will invalidate all IPv6 routing table entries.
- Insert/remove XFRM policy will also invalidate both IPv4/IPv6 routing table
  entries even when the policy is only applied for one address family.

Thus, this patch attempt to split one genid for two to cater for IPv4 and IPv6
separately in a fine granularity.
Signed-off-by: Fan Du <fan.du@windriver.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The RFE interrupt is enabled for the r8a7790 but isn't handled,
resulting in the interrupts core noticing unhandled interrupts, and
eventually disabling the ethernet IRQ.

Fix it by adding RFE to the bitmask of error interrupts to be handled
for r8a7790.
Signed-off-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
This series contains updates to ixgbe and pci.

The first patch for ixgbe from Greg Rose is the second submission.  The
first submission of "ixgbe: Retain VLAN filtering in promiscuous + VT
mode" had a typo, which Joe Perches pointed out and is fixed in this
submission.

Alex updates the ixgbe driver to use the generic helper pci_vfs_assigned
instead of the driver specific function ixgbe_vfs_are_assigned.

Don Skidmore provides 4 patches for ixgbe, the first being a fix for
flow control ethtool reporting.  Originally ixgbe_device_supports_autoneg_fc()
was expected to be called by only copper devices, which lead to false
information being displayed via ethtool.  Two other patches add support
for fixed fiber for SFP+ devices and the addition of a quad-port x520
adapter.  The last patch simply bumps the driver version.

Emil Tantilov provides 3 fixes for ixgbe, two of which resolve
semaphore lock issues.  The third fix resolves several issues in the
previous implementation of the SFF data dumps of SFP+ modules.

The remaining ixgbe and pci patches are from Jacob Keller.  The pci
patches exposes bus speed, link speed and bus width so that drivers
can take advantage of this information.  In addition, adds a pci function
which obtains minimum link width and speed.  Jacob also provides the
ixgbe patch to incorporate the pci function. He provides a patch that
fixes a lockdep issue created due to ixgbe_ptp_stop always running
cancel_work_sync even if the work item had not been created properly with
INIT_WORK. This issue was found and reported by Stephen Hemminger.

-v2-
* fix patch 3 to be a bool function based on David Miller's feedback
* fix patch 4 debug message based on David Miller's feedback
* fix patch 8 moved the extern declarations to pci.h based on Bjorn
  Helgaas's feedback
* fix patch 11 update the error message to include encoding loss based
* fix patch 8/9/10 title based on Bjorn's feedback
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove declaration, 4 defines and confusing comment that are no longer used
since 1a2c6181 ("tcp: Remove TCPCT").
Signed-off-by: Dmitry Popov <dp@highloadlab.com>
Acked-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

This is a x520 based quad-port (4x10Gbps) NIC with a single QSFP+
connector.  Changes were required to our identify functions due to
different eeprom address which is also included here.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch changes the error code path in ixgbe_acquire_swfw_sync() to deal
with cases where acquiring SW semaphore times out.

In cases where the SW/FW semaphore bits were set (i.e. due to a crash) the
driver will hang on load. With this patch the driver will clear
the stuck bits if the semaphore was not acquired in the allotted time.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch renames the stats introduced by the busy poll feature so that they
are more inline with the current statistics naming schemes.
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>

This patch fixes a lockdep issue created due to ixgbe_ptp_stop always running
cancel_work_sync even if the work item had not been created properly with
INIT_WORK. This is caused because ixgbe_ptp_stop did not check to actually
ensure PTP was running first. The new implementation introduces a state in the
&adapter->state field which is used to indicate that PTP is running. (This
replaces the IXGBE_FLAG2_PTP_ENABLED field). This state will use the atomic
set_bit, test_bit, and test_and_clear_bit functions. ixgbe_ptp_stop will check
to ensure that PTP was enabled, (and if not, it will not attempt to do any
cleanup work from ixgbe_ptp_init). This resolves the lockdep annotation warning
found by Stephen Hemminger
Reported-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-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>

This patch uses the new pcie_get_minimum_link function to perform a check to
ensure that the adapter is hooked into a slot which is capable of providing the
necessary bandwidth. This check supersedes the original method which only
checked the current pci device. The new method is capable of determining the
minimum speed and link of an entire PCI chain.

-v2-
* update the error message to include encoding loss

CC: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

A PCI Express device can potentially report a link width and speed which it will
not properly fulfill due to being plugged into a slower link higher in the
chain. This function walks up the PCI bus chain and calculates the minimum link
width and speed of this entire chain. This can be useful to enable a device to
determine if it has enough bandwidth for optimum functionality.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

"ifa->ifa_label" is an array inside the in_ifaddr struct.  It can never
be NULL so we can remove this check.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>

Support IPPROTO_IPV6 similar to IPPROTO_IPIP
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Explicitly set proto to ETH_P_IP and jump directly to ip processing.
Signed-off-by: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

pcie_link_width is the enum used to define the link width values for a pcie
device. This enum should not be contained solely in pci_hotplug.h, and this
patch moves it next to pci_bus_speed in pci.h
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

pcie_link_speed and pcix_bus_speed are arrays used by probe.c to correctly
convert lnksta register values into the pci_bus_speed enum. These static arrays
are useful outside probe for this purpose. This patch makes these defines into
conist arrays and exposes them with an extern header in drivers/pci/pci.h

-v2-
* move extern declarations to drivers/pci/pci.h

CC: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This patch fixes several issues with the previous implementation of the
SFF data dump of SFP+ modules:

- removed the __IXGBE_READ_I2C flag - I2C access locking is handled in the
  HW specific routines

- fixed the read loop to read data from ee->offset to ee->len

- the reads fail if __IXGBE_IN_SFP_INIT is set in the process - this is
  needed because on some HW I2C operations can take long time and disrupt
  the SFP and link detection process
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Reported-by: Ben Hutchings <bhutchings@solarflare.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

ixgbe_read/write_i2c_phy_82598() does not hold the SWFW_SYNC
semaphore for the entire function. Instead the lock is held only
during the phy.ops.read/write_reg operations. As result when the
function is being called simultaneously the I2C read/writes can
be corrupted.

The following patch introduces the SWFW_SYNC semaphore for the
entire ixgbe_read/write_i2c_phy_82598() function. To accomplish
this I had to create 2 separate functions:

ixgbe_read_phy_reg_mdi()
ixgbe_write_phy_reg_mdi()

Those functions are identical to ixgbe_read/write_phy_reg_generic()
sans the locking, and can be used in ixgbe_read/write_i2c_phy_82598()
with the SWFW_SYNC semaphore being held.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Bump the version number to better match with a similar version of the
out of tree driver.
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>

This patch adds support for a new media type fiber_fixed.  This is useful
to avoid all the SFP+ hot plug support path on devices who's fix fiber need
not worry about such things.  This patch is needed for a following patch
that adds support for "fiber_fixed" devices.

v2: cleaned up logging message based on feedback from David Miller
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Jiri Pirko says:

====================
This patchset is based on patch by Narendra_K@Dell.com
Once device which can change phys port id during its lifetime adopts this,
NETDEV_CHANGEPHYSPORTID event will be added and driver will call
call_netdevice_notifiers(NETDEV_NETDEV_CHANGEPHYSPORTID, dev) to propagate
the change to userspace.

v1->v2: as suggested by Ben, handle -EOPNOTSUPP in rtnl code (wrapped up ndo call)
v2->v3: adjusted patch 1 commit message
v3->v4: used "%phN" for sysfs printf as suggested by DaveM
        added igb/igbvf implementation as requested by Or Gerlitz
v4->v5: used prandom_u32 to generate id in igb_probe
        removed duplicate code in ibgvf_probe
        pushed dev_err string into one line in igbvf_refresh_ppid
v5->v6: use uuid_le_gen for generating 16-byte phys port id for igb/igbvf
	as suggested by BenH

1) Why do we need this, and why do existing facilities fail to provide
   a way to accomplish this?

Currenty there's very hard to tell if two netdevs are using the same physical
port. For sr-iov this can be get by sysfs. For other mechanisms, like NPAR
there's very hard to do it (one must learn it from NIC BIOS). But even for
sr-iov there's no way to say if two netdevs are using the same phys port when
these are passed through to virtual guests.

This patchset provides the generic way of letting this information know to
userspace. This info can be used by apps like NetworkManager, teamd, Wicked,
ovs daemon, etc, to do smarter bonding decisions.

2) Why is the physical port ID defined as a 32 byte opaque cookie?
   What formats and layouts need to be accomodated, and which
   influenced the design of the ID?

For user to distinguish if two netdevs are using the same port, he only needs
to compare their phys port ids. Nothing else is needed. This id has no
structure for security reasons. VF should not know anything about PF.

3) Are IDs globally unique?  Why or why not?  If IDs should be
   globally unique, but only in certain cases, what exactly are those
   cases.

Most of the time only uniqueness needed is in scope of single machine.
There might be case when the id should be unique between couple of machines
in virtualization environment. Given that for example for igb/igbvf 16B uuid
is used, there is no problem for this case as well. But each driver can
implement this differently focusing the hw capabilities and needs.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Narendra K <narendra_k@dell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Narendra K <narendra_k@dell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds a ndo for getting physical port of the device. Driver
which is aware of being virtual function of some physical port should
implement this ndo. This is applicable not only for IOV, but for other
solutions (NPAR, multichannel) as well. Basically if there is possible
to have multiple netdevs on the single hw port.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Acked-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Originally ixgbe_device_supports_autoneg_fc() was only expected to
be called by copper devices.  This would lead to false information
to be displayed via ethtool.

v2: changed ixgbe_device_supports_autoneg_fc() to a bool function,
    it returns bool.  Based on feedback from David Miller
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

This change makes it so that the ixgbe driver uses the generic helper
pci_vfs_assigned instead of the ixgbe specific function
ixgbe_vfs_are_assigned.
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

When using the new bridge FDB interface to allow SR-IOV virtual function
network devices to communicate with SW bridged network devices the
physical function is placed into promiscuous mode and hardware VLAN
filtering is disabled.  This defeats the ability to use VLAN tagging
to isolate user networks.  When the device is in promiscuous mode and
VT mode simultaneously ensure that VLAN hardware filtering remains
enabled.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Stephen Ko <stephen.s.ko@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Use the "swap descriptor" feature of the hardware to properly swap the
descriptors when running in big endian mode. Since the swapping occurs
on 64 bits words, we also need to provide a separate structure layout
for the DMA descriptors between little endian and big endian mode,
like is done in the mv643xx_eth driver.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The macros used for the various fields of the RX and TX descriptions
are currently declared next to those fields within the structure
definitions of the RX and TX descriptors.

However, in order to support big endian, we'll have to use the "swap
descriptors" features of the hardware, which swaps every byte within
each 64 bits word of the descriptors. This requires a separate
definition of the RX and TX descriptor structures for little and big
endian, as is done in the mv643xx_eth. Those macros can therefore no
longer be defined inside those structures.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Unlike for IPv6, the IPv4 checksum functions are only available
if CONFIG_INET is set.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Thomas Graf <tgraf@suug.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>

| If you want to test which effects syncookies have to your
| network connections you can set this knob to 2 to enable
| unconditionally generation of syncookies.

Original idea and first implementation by Eric Dumazet.

Cc: Florian Westphal <fw@strlen.de>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Adding support for setting MAC address to cpsw device via ndo_set_mac_address
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch reverts commit

80b45261

which was implementing a 'cancelled' functionality to notify that
a cancelled request will not be replied.

This implementation was not used anywhere and therefore removed.
Signed-off-by: Andi Shyti <andi@etezian.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver uses dev_err when enabling of AER fails (e.g. PCIe AER is not
supported). The dev_info is more appropriate to avoid console pollution.

Cc: sathya.perla@emulex.com
Cc: subbu.seetharaman@emulex.com
Cc: ajit.khaparde@emulex.com
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The 5762 devices sometimes incorrectly treat udp fragments as RMCP
packets and route to the APE. This patch sets the RX_MODE_IPV4_FRAG_FIX
bit for these devices which enables the proper behaviour.
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The PTP_CAPABLE tg3 devices have a gpio output that is toggled when the
free running counter matches a watchdog value. This patch adds support
to set the watchdog and enable this feature.

Since the output is controlled via bits in the EAV_REF_CLCK_CTL
register, we have to read-modify-write it when we stop/resume.

Cc: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Also remove the call to tg3_power_down_prepare() in tg3_power_down()
since tg3_close() calls it.
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Previously, when the interface was brought down, the driver would set
the power state to D3hot. In D3hot, we don't have access to the NVRAM.
This patch removes the call to set the power state to PCI_D3hot in
close. A following patch will implement the shutdown handler to properly
set the D3hot state when the system is going down.

Doing the above means that the TG3_PHYFLG_IS_LOW_POWER should not be
checked to validate access to the NVRAM.
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>