When an UDP application switches from AF_INET to AF_INET6 sockets, we
have a small performance degradation for IPv4 communications because of
extra cache line misses to access ipv6only information.

This can also be noticed for TCP listeners, as ipv6_only_sock() is also
used from __inet_lookup_listener()->compute_score()

This is magnified when SO_REUSEPORT is used.

Move ipv6only into struct sock_common so that it is available at
no extra cost in lookups.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates 2014-07-01

This series contains updates to i40e, i40evf, igb and ixgbe.

Shannon adds the Base Address High and Low to the admin queue structure
to simplify the logic in the configuration routines.  Also adds code to
clear all queues and interrupts to help clean up after a PXE or other
early boot activity.

Kevin fixes mask assignment value since -1 cannot be used for unsigned
integer types.

Mitch fixes an issue where in some circumstances the reply from the PF
would come back before we were able to properly modify the admin queue
pending and required flags.  This would mess up the flags and put the
driver in an indeterminate state, so fix this by simply setting the flags
before sending the request to the admin queue.  Also changes the branding
string for i40evf to reduce confusion and to match up with our other
marketing materials.

Kamil adds a new variable defining admin send queue (ASQ) command write
back timeout to allow for dynamic modification of this timeout.

Anjali fix a bug in the flow director filter replay logic, so that we
call a replay after a sideband reset correctly.

Jesse adds code to initialize all members of the context descriptor to
prevent possible stale data.

Christopher fixes i40e to prevent writing to reserved bits, since the
queue index is only 0-127.

Jacob removes the unneeded header export.h from the i40e PTP code.
Fixes ixgbe PTP code where the PPS signal was not correct, as it
generates a one half HZ clock signal, it only generates one level
change per second.  To generate a full clock, we need two level changes
per second.

Todd provides a fix for igb to bring up link when the PHY has powered
up, which was reported by Jeff Westfahl.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This limitation maybe had some reason in the past, but now there is not
one -> removing this.
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Acked-by: NVeaceslav Falico <vfalico@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Hariprasad Shenai says:

====================
cxgb4: Fix for PCI passthrough and some Misc. fixes

This patch series fixes probe failure in VM when PF is exposed through PCI
Passthrough. Adds support to use firmware interface to get BAR0 value.
Replace the backdoor mechanism to access the HW memory with PCIe Window method
which fixes memory I/O. Also adds device ID of few more adapters for cxgb4 and
cxgb4vf driver.

The patches series is created against 'net-next' tree.
And includes patches on cxgb4, cxgb4vf and iw_cxgb4 driver.

Since this patch-series contains mainly cxgb4 related changes, we would like to
request this patch series to get merged via David Miller's 'net-next' tree.

We have included all the maintainers of respective drivers. Kindly review the
change and let us know in case of any review comments.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Rip out a bunch of redundant PCI-E Memory Window Read/Write routines,
collapse the more general purpose routines into a single routine
thereby eliminating the need for a large stack frame (and extra data
copying) in the outer routine, change everything to use the improved
routine t4_memory_rw.

Based on origninal work by Casey Leedom <leedom@chelsio.com> and
Steve Wise <swise@opengridcomputing.com>
Signed-off-by: NCasey Leedom <leedom@chelsio.com>
Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use the firmware interface to get the BAR0 value since we really don't want
to use the PCI-E Configuration Space Backdoor access which is owned by the
firmware.

Set up PCI-E Memory Window registers using the true values programmed into
BAR registers.  When the PF4 "Master Function" is exported to a Virtual
Machine, the values returned by pci_resource_start() will be for the
synthetic PCI-E Configuration Space and not the real addresses. But we need
to program the PCI-E Memory Window address decoders with the real addresses
that we're going to be using in order to have accesses through the Memory
Windows work.

Based on origninal work by Casey Leedom <leedom@chelsio.com>
Signed-off-by: NCasey Leedom <leedom@chelsio.com>
Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Change logic which determines our Physical Function at PCI Probe time.
Now we read the PL_WHOAMI register and get the Physical Function.

Pass Physical Function to Upper Layer Drivers in lld_info structure in the
new field "pf" added to lld_info.  This is useful for the cases where the
PF, say PF4, is attached to a Virtual Machine via some form of "PCI
Pass Through" technology and the PCI Function shows up as PF0 in the VM.

Based on original work by Casey Leedom <leedom@chelsio.com>
Signed-off-by: NCasey Leedom <leedom@chelsio.com>
Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Stefan Sørensen says:

====================
dp83640: Increase support perout pins

This patch series increases the number of periodic output pins supported
on the dp83640 to 7, and allows for reprogramming the calibration pin.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The ptp pin function programming does not allow calibration pin to change
function. This is problematic on hardware that uses the default calibration
pin for other purposes.

Removing this limitation does not impact calibration if userspace does not
reprogram the calibration pin.
Signed-off-by: NStefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For consistency, use the ptp_find_pin function to get the calibration pin,
not gpio_tab.
Signed-off-by: NStefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This constraints the pin assignment to not allow the calibration function to
be reassigned and only allow reassigning the calibratin pin if only one phy is
connected.
Signed-off-by: NStefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch increases the number of supported periodic output pins from
1 to 7. The last pin is reserved for sync.
Signed-off-by: NStefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Periodic output triggers 0 and 1 of the dp83640 has a programmable
duty-cycle which is controlled by the Pulsewidth2 field of the trigger
data register.  This field is not documented in the datasheet, but it
is described in the "PHYTER Software Development Guide" section
3.1.4.1. Failing to set the field can result in unstable/no trigger
output.

Add programming of the Pulsewidth2 field, setting it to the same value
as the Pulsewidth field for a 50% duty cycle.
Signed-off-by: NStefan Sørensen <stefan.sorensen@spectralink.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Yuval Mintz says:

====================
bnx2x: Enhancement patch series

This patch series introduces the ability to propagate link parameters
to VFs as well as control the VF link via hypervisor.

In addition, it contains 2 small improvements [one IOV-related and the
other improves performance on machines with short cache lines].

Please consider applying these patches to `net-next'.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There are linux distributions where the inbox bnx2x driver contains SRIOV
support but doesn't contain the changes introduced in b9871bcf
"bnx2x: VF RSS support - PF side".

A VF in a VM running that distribution over a new hypervisor will access
incorrect addresses when trying to transmit packets, causing an attention
in the hypervisor and making that VF inactive until FLRed.

The driver in the VM has to ne upgraded [no real way to overcome this], but
due to the HW attention currently arising upgrading the driver in the VM
would not suffice [since the VF needs also be FLRed if the previous driver
was already loaded].

This patch causes the PF to fail the acquire message from a VF running an
old problematic driver; The VF will then gracefully fail it's probe preventing
the HW attention [and allow clean upgrade of driver in VM].
Signed-off-by: NYuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: NAriel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This improves the performance of driver on machine with L1_CACHE_SHIFT of at
most 32 bytes [HW was planned for 64-byte aligned fastpath data].
Signed-off-by: NDmitry Kravkov <Dmitry.Kravkov@qlogic.com>
Signed-off-by: NYuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: NAriel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Until now VFs were oblvious to the actual configured link parameters.
This patch does 2 things:

  1. It enables a PF to inform its VF using the bulletin board of the link
     configured, and allows the VF to present that information.

  2. It adds support of `ndo_set_vf_link_state', allowing the hypervisor
     to set the VF link state.
Signed-off-by: NDmitry Kravkov <Dmitry.Kravkov@qlogic.com>
Signed-off-by: NYuval Mintz <Yuval.Mintz@qlogic.com>
Signed-off-by: NAriel Elior <Ariel.Elior@qlogic.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jesper Dangaard Brouer says:

====================
Optimizing pktgen for single CPU performance

This series focus on optimizing "pktgen" for single CPU performance.

V2-series:
 - Removed some patches
 - Doc real reason for TX ring buffer filling up

NIC tuning for pktgen:
 http://netoptimizer.blogspot.dk/2014/06/pktgen-for-network-overload-testing.html

General overload setup according to:
 http://netoptimizer.blogspot.dk/2014/04/basic-tuning-for-network-overload.html

Hardware:
 System: CPU E5-2630
 NIC: Intel ixgbe/82599 chip

Testing done with net-next git tree on top of
 commit 6623b419 ("Merge branch 'master' of...jkirsher/net-next")

Pktgen script exercising race condition:
 https://github.com/netoptimizer/network-testing/blob/master/pktgen/unit_test01_race_add_rem_device_loop.sh

Tool for measuring LOCK overhead:
 https://github.com/netoptimizer/network-testing/blob/master/src/overhead_cmpxchg.c
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The if_lock()/if_unlock() in next_to_run() adds a significant
overhead, because its called for every packet in busy loop of
pktgen_thread_worker().  (Thomas Graf originally pointed me
at this lock problem).

Removing these two "LOCK" operations should in theory save us approx
16ns (8ns x 2), as illustrated below we do save 16ns when removing
the locks and introducing RCU protection.

Performance data with CLONE_SKB==100000, TX-size=512, rx-usecs=30:
 (single CPU performance, ixgbe 10Gbit/s, E5-2630)
 * Prev   : 5684009 pps --> 175.93ns (1/5684009*10^9)
 * RCU-fix: 6272204 pps --> 159.43ns (1/6272204*10^9)
 * Diff   : +588195 pps --> -16.50ns

To understand this RCU patch, I describe the pktgen thread model
below.

In pktgen there is several kernel threads, but there is only one CPU
running each kernel thread.  Communication with the kernel threads are
done through some thread control flags.  This allow the thread to
change data structures at a know synchronization point, see main
thread func pktgen_thread_worker().

Userspace changes are communicated through proc-file writes.  There
are three types of changes, general control changes "pgctrl"
(func:pgctrl_write), thread changes "kpktgend_X"
(func:pktgen_thread_write), and interface config changes "etcX@N"
(func:pktgen_if_write).

Userspace "pgctrl" and "thread" changes are synchronized via the mutex
pktgen_thread_lock, thus only a single userspace instance can run.
The mutex is taken while the packet generator is running, by pgctrl
"start".  Thus e.g. "add_device" cannot be invoked when pktgen is
running/started.

All "pgctrl" and all "thread" changes, except thread "add_device",
communicate via the thread control flags.  The main problem is the
exception "add_device", that modifies threads "if_list" directly.

Fortunately "add_device" cannot be invoked while pktgen is running.
But there exists a race between "rem_device_all" and "add_device"
(which normally don't occur, because "rem_device_all" waits 125ms
before returning). Background'ing "rem_device_all" and running
"add_device" immediately allow the race to occur.

The race affects the threads (list of devices) "if_list".  The if_lock
is used for protecting this "if_list".  Other readers are given
lock-free access to the list under RCU read sections.

Note, interface config changes (via proc) can occur while pktgen is
running, which worries me a bit.  I'm assuming proc_remove() takes
appropriate locks, to assure no writers exists after proc_remove()
finish.

I've been running a script exercising the race condition (leading me
to fix the proc_remove order), without any issues.  The script also
exercises concurrent proc writes, while the interface config is
getting removed.
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Avoid calling set_current_state() inside the busy-loop in
pktgen_thread_worker().  In case of pkt_dev->delay, then it is still
used/enabled in pktgen_xmit() via the spin() call.

The set_current_state(TASK_INTERRUPTIBLE) uses a xchg, which implicit
is LOCK prefixed.  I've measured the asm LOCK operation to take approx
8ns on this E5-2630 CPU.  Performance increase corrolate with this
measurement.

Performance data with CLONE_SKB==100000, rx-usecs=30:
 (single CPU performance, ixgbe 10Gbit/s, E5-2630)
 * Prev:  5454050 pps --> 183.35ns (1/5454050*10^9)
 * Now:   5684009 pps --> 175.93ns (1/5684009*10^9)
 * Diff:  +229959 pps -->  -7.42ns
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using pktgen I'm seeing the ixgbe driver "push-back", due TX ring
running full.  Thus, the TX ring is artificially limiting pktgen.
(Diagnose via "ethtool -S", look for "tx_restart_queue" or "tx_busy"
counters.)

Using ixgbe, the real reason behind the TX ring running full, is due
to TX ring not being cleaned up fast enough. The ixgbe driver combines
TX+RX ring cleanups, and the cleanup interval is affected by the
ethtool --coalesce setting of parameter "rx-usecs".

Do not increase the default NIC TX ring buffer or default cleanup
interval.  Instead simply document that pktgen needs special NIC
tuning for maximum packet per sec performance.

Performance results with pktgen with clone_skb=100000.
TX ring size 512 (default), adjusting "rx-usecs":
 (Single CPU performance, E5-2630, ixgbe)
 - 3935002 pps - rx-usecs:  1 (irqs:  9346)
 - 5132350 pps - rx-usecs: 10 (irqs: 99157)
 - 5375111 pps - rx-usecs: 20 (irqs: 50154)
 - 5454050 pps - rx-usecs: 30 (irqs: 33872)
 - 5496320 pps - rx-usecs: 40 (irqs: 26197)
 - 5502510 pps - rx-usecs: 50 (irqs: 21527)

TX ring size adjusting (ethtool -G), "rx-usecs==1" (default):
 - 3935002 pps - tx-size:  512
 - 5354401 pps - tx-size:  768
 - 5356847 pps - tx-size: 1024
 - 5327595 pps - tx-size: 1536
 - 5356779 pps - tx-size: 2048
 - 5353438 pps - tx-size: 4096

Notice after commit 6f25cd47 (pktgen: fix xmit test for BQL enabled
devices) pktgen uses netif_xmit_frozen_or_drv_stopped() and ignores
the BQL "stack" pause (QUEUE_STATE_STACK_XOFF) flag.  This allow us to put
more pressure on the TX ring buffers.

It is the ixgbe_maybe_stop_tx() call that stops the transmits, and
pktgen respecting this in the call to netif_xmit_frozen_or_drv_stopped(txq).
Signed-off-by: NJesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This stub now allows userspace to see IFLA_INFO_KIND for ovs master and
IFLA_INFO_SLAVE_KIND for slave.
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

So far, it is assumed that ops->setup is filled up. But there might be
case that ops might make sense even without ->setup. In that case,
forbid to newlink and dellink.

This allows to register simple rtnl link ops containing only ->kind.
That allows consistent way of passing device kind (either device-kind or
slave-kind) to userspace.
Signed-off-by: NJiri Pirko <jiri@resnulli.us>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In commit 37112105("net:
QDISC_STATE_RUNNING dont need atomic bit ops") the
__QDISC_STATE_RUNNING is renamed to __QDISC___STATE_RUNNING,
but the old names existing in comment are not replaced with
the new name completely.
Signed-off-by: NYing Xue <ying.xue@windriver.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This can be used in virtual networking applications, and
may have other uses as well.  The option is disabled by
default.

A specific use case is setting up virtual routers, bridges, and
hosts on a single OS without the use of network namespaces or
virtual machines.  With proper use of ip rules, routing tables,
veth interface pairs and/or other virtual interfaces,
and applications that can bind to interfaces and/or IP addresses,
it is possibly to create one or more virtual routers with multiple
hosts attached.  The host interfaces can act as IPv6 systems,
with radvd running on the ports in the virtual routers.  With the
option provided in this patch enabled, those hosts can now properly
obtain IPv6 addresses from the radvd.
Signed-off-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This is disabled by default, just like similar debug info
already in this module.  But, makes it easier to find out
why RA is not being accepted when debugging strange behaviour.
Signed-off-by: NBen Greear <greearb@candelatech.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The PPS signal is not correct, as it generates a one half HZ clock
signal, as it only generates one level change per second. To generate a
full clock, we need two level changes per second. Also, change the name
of the #define, in order to prevent confusion between it and
NSEC_PER_SEC which is not guaranteed to be a 64bit value.
Signed-off-by: NJacob Keller <jacob.e.keller@intel.com>
Tested-by: NPhil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Call igb_setup_link() when the PHY is powered up.
Signed-off-by: NTodd Fujinaka <todd.fujinaka@intel.com>
Reported-by: NJeff Westfahl <jeff.westfahl@ni.com>
Tested-by: NAaron Brown <aaron.f.brown@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Bump versions.

Change-ID: Id5082d7c3995fbddd22b3e303d804c86fcd240a3
Signed-off-by: NCatherine Sullivan <catherine.sullivan@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Add a slash to the branding string to reduce confusion and match up with
our other marketing materials.

Change-ID: I8229e8c3e43083b7a29c859a250f8d2d4dc46b9e
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

We don't need the export.h header so we can just go ahead and remove it.

Change-ID: I9057396b141ee449d8299409081358b9270a7c4d
Signed-off-by: NJacob Keller <jacob.e.keller@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Prevent writing to reserved bits, queue index is 0-127

Change-ID: Ic923e1c92012a265983414acd8f547c4bdac2e34
Signed-off-by: NChristopher Pau <christopher.pau@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Driver needs to initialize all members of context descriptor. Stale
data is possible otherwise.

Change-ID: Idc6b53af45583509da42d5ec0824cbaf78aee64f
Signed-off-by: NJesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

With the auto_disable flags added there was a bug that was causing the
replay logic to not work correctly.
This patch fixes the issue so that we call a replay after a sideband
reset correctly.

Change-ID: I005fe1ac361188ee5b19517a83c922038cba1b00
Signed-off-by: NAnjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Add new variable defining ASQ command write back timeout to allow for
dynamic modification of this timeout. Initialize it on AQ initialize
routine with default value, vary it on device ID.

Change-ID: I5c9908f9d7c5455634353b694a986d6f146d1b9d
Signed-off-by: NKamil Krawczyk <kamil.krawczyk@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

In some circumstances, the firmware could beat us to the punch, and the
reply from the PF would come back before we were able to properly modify
the aq_pending and aq_required flags. This would mess up the flags and
put the driver in an indeterminate state, much like Schrödinger's cat.
However, unlike the cat, the driver is definitely dead.

To fix this, simply set the flags before sending the request to the AQ.
This way, it won't matter if the interrupt comes back too soon.

Change-ID: I9784655e475675ebcb3140cc7f36f4a96aaadce5
Signed-off-by: NMitch Williams <mitch.a.williams@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Make mask value of all 1s.  Value of -1 can't be used for u32 type.

Change-ID: I49d58b77639939fe7447a229dbf1f4a1bf7419ce
Signed-off-by: NKevin Scott <kevin.c.scott@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>

Per a recent HW designer comment, this code is for ripping through the
queues and interrupts to fully disable them on driver init, specifically
to help clean up after a PXE or other early boot activity.

Change-ID: I32ed452021a1c2b06dace1969976f882a37b9741
Signed-off-by: NShannon Nelson <shannon.nelson@intel.com>
Signed-off-by: NJeff Kirsher <jeffrey.t.kirsher@intel.com>