Aaron Brown says:

====================
Mark updates ixgbe for LER / adapter removal.  He restores the HW
address in the recovery path so the device is not perpetually removed,
fixes up some removed state ethtool results and adds checks related to
config space access.

Jacob adds support for the new SIOCGHWTSTAMP ioctl.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds support for the new SIOCGHWTSTAMP ioctl, which enables a
process to determine the current timestamp configuration. In order to
implement this, store a copy of the timestamp configuration. In
addition, we can remove the 'int cmd' parameter as the new set_ts_config
function doesn't use it. I also fixed a typo in the function
description.

-v2
* Only save the settings after validating them
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Configuration space reads should also be checked for removal. So
add some checks related to config space accesses.

v2:
* Fixed indent
Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some ethtool tests returned apparently good results when the
adapter was in a removed state. Fix that by checking for removal.
This also fixes two paths that could return uninitialized memory
in data[4].
Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The hw_addr needs to be restored in the pcie recovery path or
else the device will be perpetually removed. Also restore the
value in the resume path.
Signed-off-by: Mark Rustad <mark.d.rustad@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently we're only sending arp requests if we have a route to the target
(and, thus, can find out the source ip address).

There are some use cases, however, where we don't want/need to set an ip
address (or set up a specific route) for bonding to use arp monitoring *for
traffic generation*. We can easily send arp probes (arp requests with src
ip == 0) to generate arp broadcast responses from the target ip and use
them for determining if the target is up.

This, obviously, won't work with arp validation - because we don't have the
ip address set and, thus, will filter out the responses. So in that case -
print a warning.

CC: François CACHEREUL <f.cachereul@alphalink.fr>
CC: Zhenjie Chen <zhchen@redhat.com>
CC: Jay Vosburgh <fubar@us.ibm.com>
CC: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Veaceslav Falico <vfalico@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Alexander Aring says:

====================
6lowpan: reimplementation of fragmentation handling

this patch series reimplementation the fragmentation handling of 6lowpan
accroding to rfc4944 [1].

The first big note is, that the current fragmentation behaviour isn't rfc
complaint. The main issue is a wrong datagram_size value which needs to be:
datagram_size = ipv6_payload + ipv6 header + (maybe compressed transport header,
                currently only udp is supported)

but the current datagram_size value is calculated as:
datagram_size = ipv6_payload

Fragmentation work in a linux<->linux communication only.

Why reimplementation?

I reimplemted the reassembly side only. The current behaviour is to allocate a
skb with the reassembled size and hold all fragments in a list, protected by a
spinlock. After we received all fragments (detected by the sum of all fragments,
it begins to place all fragments into the allocated skb).

This reassembly implementation has some race condition. Additional I make it more
rfc complaint. The current implementation match on the tag value inside the frag
header only, but rfc4944 says we need to match on dst addr(mac), src addr(mac),
tag value, datagram_size value. [2]

The new reassembly handling use the inet_frag api (I mean the callback interface
of ipv6 and ipv4 reassembly). I looked into ipv6 and wanted to see how ipv6 is
dealing with reassembly, so I based my code on this implementation.

On the sending side to generate the fragments I improved the current code to use
the nearest 8 divided payload. (We can do that, because the mac layer has a
dynamic size, so it depends on mac_header how big we can do the payload).

Of course I fix also the reassembly/sending side to be rfc complaint now.

Regards
Alexander Aring

[1] http://tools.ietf.org/html/rfc4944
[2] http://tools.ietf.org/html/rfc4944#section-5.3

changes since v2:
 - rework checkpatch code style issue patch.
   Merge two pr_debugs into one pr_debug.

changes since v3:
 - rename 6lowpan.ko to 6lowpan_rtnl.c in commit msg of patch 5/8.

changes since v4:
 - Add a new patch 2/8 to introduce lowpan_uncompress_size function. Also
   improving this function a little bit.
 - Add a new patch 4/8 to change tag value to __be16.
 - use skb_header_reset function on FRAG1 only, which should have the
   lowpan header. See lowpan_get_frag_info function. (slightly improving
   of fragmentation header parsing).
 - changes types of variables to u16 in lowpan_skb_fragmentation.
 - use lowpan_uncompress_size instead of storing necessary information
   in skb control block, this can be destroyed after dev_queue_xmit call.
   Thanks David for this hint.
 - remove Tested-by: Martin Townsend <martin.townsend@xsilon.com>, because
   too many funcionality change.

changes since v5:
 - handle lowpan_addr_mode_size with lookup table.

changes since v6:
 - remove unnecessary parameter in lowpan_frag_queue.
 - fix commit message in patch 8/8 which included a describtion of adding the
   lownpan_uncompress_size function. This was splitted in a seperate patch.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch drops the current way of 6lowpan fragmentation on receiving
side and replace it with a implementation which use the inet_frag api.
The old fragmentation handling has some race conditions and isn't
rfc4944 compatible. Also adding support to match fragments on
destination address, source address, tag value and datagram_size
which is missing in the current implementation.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds necessary ieee802154 6lowpan namespace to provide the
inet_frag information. This is a initial support for handling 6lowpan
fragmentation with the inet_frag api.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Detected with:

./scripts/checkpatch.pl --strict -f net/ieee802154/6lowpan_rtnl.c
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We have a 6lowpan.c file and 6lowpan.ko file. To avoid confusing we
should move 6lowpan.c to 6lowpan_rtnl.c. Then we can support multiple
source files for 6lowpan module.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fix the fragmentation on sending side according to rfc4944.

Also add improvement to use the full payload of a PDU which calculate
the nearest divided to 8 payload length for the fragmentation datagram
size attribute.

The main issue is that the datagram size of fragmentation header use the
ipv6 payload length, but rfc4944 says it's the ipv6 payload length inclusive
network header size (and transport header size if compressed).
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch add a lookup function for uncompressed 6LoWPAN header
size. This is needed to estimate the real size after uncompress the
6LoWPAN header.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds a 6lowpan fragmentation struct into cb of skb which
is necessary to hold fragmentation information.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use devm_ioremap_resource() in order to make the code simpler.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use devm_ioremap_resource() in order to make the code simpler.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 57f89bfa ("network: Allow af_packet to transmit +4 bytes
for VLAN packets.") added the possibility for non-mmaped frames to
send extra 4 byte for VLAN header so the MTU increases from 1500 to
1504 byte, for example.

Commit cbd89acb ("af_packet: fix for sending VLAN frames via
packet_mmap") attempted to fix that for the mmap part but was
reverted as it caused regressions while using eth_type_trans()
on output path.

Lets just act analogous to 57f89bfa and add a similar logic
to TX_RING. We presume size_max as overcharged with +4 bytes and
later on after skb has been built by tpacket_fill_skb() check
for ETH_P_8021Q header on packets larger than normal MTU. Can
be easily reproduced with a slightly modified trafgen in mmap(2)
mode, test cases:

 { fill(0xff, 12) const16(0x8100) fill(0xff, <1504|1505>) }
 { fill(0xff, 12) const16(0x0806) fill(0xff, <1500|1501>) }

Note that we need to do the test right after tpacket_fill_skb()
as sockets can have PACKET_LOSS set where we would not fail but
instead just continue to traverse the ring.
Reported-by: Mathias Kretschmer <mathias.kretschmer@fokus.fraunhofer.de>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Ben Greear <greearb@candelatech.com>
Cc: Phil Sutter <phil@nwl.cc>
Tested-by: Mathias Kretschmer <mathias.kretschmer@fokus.fraunhofer.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Aaron Brown says:

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

Don provides an update to change a hard coded timeout interval to
a system-wide timeout one, collects AUTOC register functions into
one place and fixes some firmware bit handling.

Emil resolves a tx handling error introduced in a recent commit and
adds check for CHECKSUM_PARTIAL to avoid an skb_is_gso check
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds check for CHECKSUM_PARTIAL to avoid the skb_is_gso check
in ixgbevf_tso(). It should reduce overhead for workloads that are not using
TSO or checksum offloads. It is the same as in ixgbe.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch resolves an issue introduced by:
commit 7ad1a093
ixgbevf: make the first tx_buffer a repository for most of the skb info

Incorrect check for the result of ixgbevf_tso() can lead to calling
ixgbevf_tx_csum() which can spawn 2 context descriptors and result in
performance degradation and/or corrupted packets.
Signed-off-by: Emil Tantilov <emil.s.tantilov@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver will now honor the MNG FW veto bit in blocking link resets.
This patch will affect x520 and x540 systems.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The current code doesn't toggle the correct bit to reset the data pipeline
on Restart_AN assertion.  This patch corrects that.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When reading or writing to the AUTOC register on 82599 devices we need to
preform various operations that aren't needed for other MAC types.  This
patch will collect all of that code into one place to minimize MAC checks
in common code paths.

While doing this I also clean up some cases where we weren't holding the
SW/FW semaphore during a read/modify/write of AUTOC.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Currently we were just always polling for a hard coded 80 ms and not
respecting the system-wide timeout interval.  Since up until now all
devices have been tested with this 80ms value we continue to use this
value as a hard minimum.
Signed-off-by: Don Skidmore <donald.c.skidmore@intel.com>
Tested-by: Phil Schmitt <phillip.j.schmitt@intel.com>
Signed-off-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Avoid the following sparse __CHECK_ENDIAN__ warnings:

 include/net/addrconf.h:318:25: warning: restricted __be64 degrades to integer
 include/net/addrconf.h:318:70: warning: restricted __be64 degrades to integer
 include/net/addrconf.h:330:25: warning: restricted __be64 degrades to integer
 include/net/addrconf.h:330:70: warning: restricted __be64 degrades to integer
 include/net/addrconf.h:347:25: warning: restricted __be64 degrades to integer
 include/net/addrconf.h:348:26: warning: restricted __be64 degrades to integer
 include/net/addrconf.h:349:18: warning: restricted __be64 degrades to integer

The warnings are false but they make it harder to spot real
bugs.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>

Because those following if conditions will not be matched.
Signed-off-by: Duan Jiong <duanj.fnst@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Steffen Klassert says:

====================
This is the rework of the IPsec virtual tunnel interface
for ipv4 to support inter address family tunneling and
namespace crossing. The only change to the last RFC version
is a compile fix for an odd configuration where CONFIG_XFRM
is set but CONFIG_INET is not set.

1) Add and use a IPsec protocol multiplexer.

2) Add xfrm_tunnel_skb_cb to the skb common buffer
   to store a receive callback there.

3) Make vti work with i_key set by not including the i_key
   when comupting the hash for the tunnel lookup in case of
   vti tunnels.

4) Update ip_vti to use it's own receive hook.

5) Remove xfrm_tunnel_notifier, this is replaced by the IPsec
   protocol multiplexer.

6) We need to be protocol family indepenent, so use the on xfrm_lookup
   returned dst_entry instead of the ipv4 rtable in vti_tunnel_xmit().

7) Add support for inter address family tunneling.

8) Check if the tunnel endpoints of the xfrm state and the vti interface
   are matching and return an error otherwise.

8) Enable namespace crossing tor vti devices.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Luis R. Rodriguez says:

====================
net: start kdoc'ifying net_device

While working on extending some functionality I felt restricted
with the amount of documentation I can add. Part of this is that
the existing style on the header files don't let me be verbose.
This starts addressing that by using kdoc for the net_device
flags, and as Ben noted, the priv_flags can be moved out from
UAPI.

Luis R. Rodriguez (2):
  net: kdoc struct net_device flags and priv_flags
  net: move net_device priv_flags out from UAPI
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

These are private to userspace, and they're unstable
anyway and can be shuffled at will (see 080e4130)
so any userspace application relying on them is on crack.

Test compiled with allyesconfig.

mcgrof@drvbp1 /pub/mem/mcgrof/net-next (git::master)$ make allyesconfig
mcgrof@drvbp1 /pub/mem/mcgrof/net-next (git::master)$ time make -j 20
...
  BUILD   arch/x86/boot/bzImage
Setup is 16992 bytes (padded to 17408 bytes).
System is 56153 kB
CRC 721d2751
Kernel: arch/x86/boot/bzImage is ready  (#1)
real    19m35.744s
user    280m37.984s
sys     27m54.104s

Cc: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We have documentation for these flags but they're scattered
all over the place. #defines don't allow documentation to be
written easily so to help to start bringing some documentation
together use the enums kdoc practice but keep the defines to
allow userspace to be able to #ifdef them.

I've verified the same values are assigned before and after
with a simple userspace test program [0] and checksumming the
output.

[0] http://drvbp1.linux-foundation.org/~mcgrof/kdoc/netdev_flags/

mcgrof@gnat ~/tmp $ ./check-flags | sha1sum
0ec5b6b1840aa3bb9ce464e61c564820871c92c3  -

Cc: netdev@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Luis R. Rodriguez <mcgrof@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We are trying to finally kill off interruptible_sleep_on_timeout.
the two uses in the nicstar driver can be trivially replaced
with wait_event_interruptible_lock_irq_timeout, which prevents the
wake-up race and is able to check the buffer state with scq->lock
held.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

Upcoming congestion controls for TCP require usec resolution for RTT
estimations. Millisecond resolution is simply not enough these days.

FQ/pacing in DC environments also require this change for finer control
and removal of bimodal behavior due to the current hack in
tcp_update_pacing_rate() for 'small rtt'

TCP_CONG_RTT_STAMP is no longer needed.

As Julian Anastasov pointed out, we need to keep user compatibility :
tcp_metrics used to export RTT and RTTVAR in msec resolution,
so we added RTT_US and RTTVAR_US. An iproute2 patch is needed
to use the new attributes if provided by the kernel.

In this example ss command displays a srtt of 32 usecs (10Gbit link)

lpk51:~# ./ss -i dst lpk52
Netid  State      Recv-Q Send-Q   Local Address:Port       Peer
Address:Port
tcp    ESTAB      0      1         10.246.11.51:42959
10.246.11.52:64614
         cubic wscale:6,6 rto:201 rtt:0.032/0.001 ato:40 mss:1448
cwnd:10 send
3620.0Mbps pacing_rate 7240.0Mbps unacked:1 rcv_rtt:993 rcv_space:29559

Updated iproute2 ip command displays :

lpk51:~# ./ip tcp_metrics | grep 10.246.11.52
10.246.11.52 age 561.914sec cwnd 10 rtt 274us rttvar 213us source
10.246.11.51

Old binary displays :

lpk51:~# ip tcp_metrics | grep 10.246.11.52
10.246.11.52 age 561.914sec cwnd 10 rtt 250us rttvar 125us source
10.246.11.51

With help from Julian Anastasov, Stephen Hemminger and Yuchung Cheng
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Larry Brakmo <brakmo@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>

ktime_get() is too expensive on some cases, and we'd like to get
usec resolution timestamps in TCP stack.

This patch adds a light weight facility using a combination of
local_clock() and jiffies samples.

Instead of :

        u64 t0, t1;

        t0 = ktime_get();
        // stuff
        t1 = ktime_get();
        delta_us = ktime_us_delta(t1, t0);

use :
        struct skb_mstamp t0, t1;

        skb_mstamp_get(&t0);
        // stuff
        skb_mstamp_get(&t1);
        delta_us = skb_mstamp_us_delta(&t1, &t0);

Note : local_clock() might have a (bounded) drift between cpus.

Do not use this infra in place of ktime_get() without understanding the
issues.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Larry Brakmo <brakmo@google.com>
Cc: Julian Anastasov <ja@ssi.bg>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for the led-mode property for the following PHYs
which have a single LED mode configuration value.

KSZ8001 and KSZ8041 which both use register 0x1e bits 15,14 and
KSZ8021, KSZ8031 and KSZ8051 which use register 0x1f bits 5,4
to control the LED configuration.
Signed-off-by: Ben Dooks <ben.dooks@codethink.co.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

The isdn core code uses a couple of wait queues with
interruptible_sleep_on, which is racy and about to get
removed from the kernel. Fortunately, we know for each case
what we are waiting for, so they can all be converted to
the better wait_event_interruptible interface.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

These two drivers use identical code for their procfs status
file handling, which contains a small race against status
data becoming available while reading the file.

This uses wait_event_interruptible instead to fix this
particular race and eventually get rid of all sleep_on
instances. There seems to be another race involving
multiple concurrent readers of the same procfs file, which
I don't try to fix here.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

The state machine code in the elsa driver uses interruptible_sleep_on
to wait for state changes, which is racy. A closer look at the possible
states reveals that it is always used to wait for getting back into
ARCOFI_NOP, so we can use wait_event_interruptible instead.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

interruptible_sleep_on is racy and going away. In case of pcbit,
the driver would run into a timeout if the card is initialized
before we start waiting for it. This uses wait_event to fix the
race. In order to do this, the state machine handling for the
timeout case has to get trivially reorganized so we actually know
whether the timeout has occorred or not.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Karsten Keil <isdn@linux-pingi.de>
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>

interruptible_sleep_on is racy and going away. This replaces the one use
in the firestream driver with the appropriate wait_event_interruptible
variant.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Chas Williams <chas@cmf.nrl.navy.mil>
Cc: linux-atm-general@lists.sourceforge.net
Cc: netdev@vger.kernel.org
Signed-off-by: David S. Miller <davem@davemloft.net>