While forwarding we should not use the protocol path mtu to calculate
the mtu for a forwarded packet but instead use the interface mtu.

We mark forwarded skbs in ip_forward with IPSKB_FORWARDED, which was
introduced for multicast forwarding. But as it does not conflict with
our usage in unicast code path it is perfect for reuse.

I moved the functions ip_sk_accept_pmtu, ip_sk_use_pmtu and ip_skb_dst_mtu
along with the new ip_dst_mtu_maybe_forward to net/ip.h to fix circular
dependencies because of IPSKB_FORWARDED.

Because someone might have written a software which does probe
destinations manually and expects the kernel to honour those path mtus
I introduced a new per-namespace "ip_forward_use_pmtu" knob so someone
can disable this new behaviour. We also still use mtus which are locked on a
route for forwarding.

The reason for this change is, that path mtus information can be injected
into the kernel via e.g. icmp_err protocol handler without verification
of local sockets. As such, this could cause the IPv4 forwarding path to
wrongfully emit fragmentation needed notifications or start to fragment
packets along a path.

Tunnel and ipsec output paths clear IPCB again, thus IPSKB_FORWARDED
won't be set and further fragmentation logic will use the path mtu to
determine the fragmentation size. They also recheck packet size with
help of path mtu discovery and report appropriate errors.

Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: David Miller <davem@davemloft.net>
Cc: John Heffner <johnwheffner@gmail.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

This is to be compatible with the use of "get_time" (i.e. default
time unit in us) in iproute2 patch for HHF as requested by Stephen.

Signed-off-by: Terry Lam <vtlam@google.com>
Acked-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

vmalloc is a limited resource.  Don't use it unnecessarily.

It seems this allocation should work with kcalloc.

Remove unnecessary memset(,0,) of buf as it's completely
overwritten as the previously only unset field in
struct qlcnic_pci_func_cfg is now set to 0.

Use kfree instead of vfree.
Use ETH_ALEN instead of 6.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

That code has been around for ages without being used.

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>

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>

It's a huge mess currently, that is really hard to read. This cleanup
doesn't touch the logic at all, it only breaks easy-to-fix long lines and
updates comment styles.

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>

Sabrina Dubroca says:

====================
alx: add statistics

Currently, the alx driver doesn't support statistics [1,2]. The
original alx driver [3] that Johannes Berg modified provided
statistics. This patch is an adaptation of the statistics code from
the original driver to the alx driver included in the kernel.

v4:
 - modified the assignements of hw stats to netstats (Ben Hutchings)
 - added comments to describe the stats fields (copied from atlx)

v3:
 - renamed __alx_update_hw_stats to alx_update_hw_stats (Stephen Hemminger)

v2:
 - use u64 instead of unsigned long  (Ben Hutchings)
 - implement ndo_get_stats64 instead of ndo_get_stats (Ben Hutchings)
 - use EINVAL instead of ENOTSUPP  (Ben Hutchings)
 - add BUILD_BUG_ON to check the size of the stats (Johannes Berg, Ben
   Hutchings)
 - add a comment regarding persistence of the stats (Stephen Hemminger)
 - align assignments in __alx_update_hw_stats

[1] https://bugzilla.kernel.org/show_bug.cgi?id=63401
[2] http://www.spinics.net/lists/netdev/msg245544.html
[3] https://github.com/mcgrof/alx


====================

Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Reviewed-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
Intel Wired LAN Driver Updates

This series contains updates to i40e and now i40evf.

Most notable is Jacob's patch to add PTP support to i40e.

Mitch cleans up additional memcpy's and use struct assignment instead.
Then fixes long lines to appease checkpatch.pl.  Mitch then provides
a fix to keep us from spamming the log with confusing errors.  If you
use ip to change the MAC address of a VF while the VF driver is loaded,
closing the VF interface or unloading the VF driver will cause the VF
driver to remove the MAC filter for its original (now invalid) MAC
address.

Jesse cleans up macros which are no longer needed or used.

I (Jeff) cleanup function header comments to ensure Doxygen/kdoc works
correctly to generate documentation without warnings.

Anjali fixes a bug where ethtool set-channels would return failure when
configuring only one Rx queue.  Then fixes a bug where the driver was
erroneously exiting the driver unload path if one part of the unload
failed.

Shannon fixes if the IPV6EXADD but is set in the Rx descriptor status,
there was an optional extension header with an alternate IP address
detected and the hardware checksum was not handling the alternate IP
address correctly.  Then adjusts the ITR max and min values to match
the hardware max value and recommended min value.  Shannon makes sure
to clear the PXE mode after the adminq is initialized.

v2:
 - fix patch 14 "i40e: enable PTP" to address Richard Cochran's spelling
   catch and Ben Hutchings Kconfig, SIOCGHWTSTAMP and sizeof() suggestions
 - added Paul Gortmaker's i40evf fix patch
v3:
 - fix patch 14 "i40e: enable PTP" to address Ben Hutchings concerns about
   a race with PTP init and cleanup and i40e_get_ts_info().
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

As of commit 7f12ad74 ("i40evf: transmit
and receive functionality") the s390 builds (allyesconfig) fail with:

drivers/net/ethernet/intel/i40evf/i40e_txrx.c: In function 'i40e_clean_rx_irq':
drivers/net/ethernet/intel/i40evf/i40e_txrx.c:818:3: error: implicit declaration of function 'prefetch'
make[5]: *** [drivers/net/ethernet/intel/i40evf/i40e_txrx.o] Error 1

due to an implicit assumption that the prototype from linux/prefetch.h
will be present.

Cc: Mitch Williams <mitch.a.williams@intel.com>
Cc: Greg Rose <gregory.v.rose@intel.com>
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Update the driver version to 0.3.28-k.

Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Change the redundant "vsi VSI" to VSI.

Change-ID: Ic16ea5820a99abc7831713cde39e7d032a7ba4d3
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

New feature: Enable PTP support in the i40e driver.

Change-ID: I6a8e799f582705191f9583afb1b9231a8db96cc8
Cc: Richard Cochran <richardcochran@gmail.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Matthew Vick <matthew.vick@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

In the latest firmware the clear_pxe_mode function will use the
AdminQ request, so call this after AdminQ is set up rather than
relying on i40e_pf_reset() to clear the PXE mode.

Change-ID: Ice8cba2e9cbc3c7bde0a0bcf8eaf5009abef040b
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Make sure the "new" qtx_head[q] register is cleared before
enabling the Tx queue.

Change-ID: I0c7a12815e343a5ae68807af172a35d6c6857935
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Set the ITR max and min values to match the hardware max value
and the recommended min value.  These values are shifted right
one bit because the register counts in 2 usec units, so leave
a comment to explain.

Change-ID: I289c27955cf6c566a6d21b95c3110b88cbb15dad
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

If the IPV6EXADD bit is set in the Rx descriptor status, there
was an optional extension header with an alternate IP address
detected.  The HW checksum offload doesn't handle the alternate
IP address correctly so likely comes up with the wrong answer.
Thus, if the bit is set we ignore the checksum offload value.

Change-ID: I70ff8d38cdcddccf44107691cae13d0c07c284c8
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

If you use ip to change the MAC address of a VF while the VF
driver is loaded, closing the VF interface or unloading the VF
driver will cause the VF driver to remove the MAC filter for its
original (now invalid) MAC address. This would cause the PF
driver to kick an error message to the log, and back to the VF
driver.

Since the VF driver has not really done anything naughty, let's
not punish it. Don't check for MAC address overrides on the
delete operation, just make sure it's a valid address. This keeps
us from spamming the log with confusing errors.

Change-ID: I1f051bd4014e50855457d928c9ee8b0766981b2f
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Fix a bug where the driver was erroneously exiting the driver unload
path if one part of the unload failed.  Instead of the original way
the driver should always continue when disabling and be sure to disable
all queues.

Change-ID: Ib8c81c596bc87c31d8e9ca97ebf871168475279d
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Fix a bug where ethtool set-channels would return failure when configuring
only one Rx queue.

Change-ID: Id833c48c17d71e352b30f3249f6acf9e7aaec57e
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

These changes make Doxygen/kdoc work correctly without warnings.

Change-ID: I2941f38860be805ff7548d84dae35754c83f1d62
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>

Avoid over-length lines in order to appease checkpatch.

Change-ID: I63820a710acf798f49d2f85c610228711af84f72
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Update driver version to 0.3.27-k

Signed-off-by: Catherine Sullivan <catherine.sullivan@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The current driver will warn the user if the NVM version
is out of date, this raises the bar to a newer version.

Change-ID: I5ec21d8efa4e7c3fdacb56f85d310bb2229b1483
Signed-off-by: Anjali Singhai Jain <anjali.singhai@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

A previous commit removed any need for these macros, so remove
them too.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

These instances were found by coccinelle/spatch, and can
use struct assignment instead of memcpy.

Change-ID: Idc23c3599241bf8a658bda18c80417af3fbfee66
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Tested-by: Kavindya Deegala <kavindya.s.deegala@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Included changes:
- substitute FSF address with URL
- deselect current bat-GW when GW-client mode gets deactivated
- send every DHCP packet using bat-unicast messages when GW-client mode is
  enabled
- implement the Extended Isolation mechanism (it is an enhancement of the
  already existing batman-AP-isolation). This mechanism allows the user to drop
  packets exchanged by selected clients by using netfilter marks.
- fix typ0 in header guard
- minor code cleanups

Signed-off-by: David S. Miller <davem@davemloft.net>

Christoph Paasch says:

====================
Make tcp-metrics source-address aware

Currently tcp-metrics only stores per-destination addresses. This brings
problems, when a host has multiple interfaces (e.g., a smartphone having
WiFi/3G):

For example, a host contacting a server over WiFi will store the tcp-metrics
per destination IP. If then the host contacts the same server over 3G, the
same tcp-metrics will be used, although the path-characteristics are completly
different (e.g., the ssthresh is probably not the same).

In case of TFO this is not a problem, as the server will provide us a new cookie
once he saw our SYN+DATA with an incorrect cookie.
It may be (in case of carrier-grade NAT), that we keep the same public IP but
have a different private IP. Thus, we better reuse the old cookie even if our
source-IP has changed. However, this scenario is probably very uncommon, as
carriers try to provide the same src-IP to the clients behind their CGN.

Patches 1 + 2 add the source-IP to the tcp metrics.

Patches 3 to 5 modify the netlink-api to support the source-IP. From now on,
when using the command "ip tcp_metrics delete address ADDRESS" all entries
which match this destination IP will be deleted.

Today's iproute2 will complain when doing "ip tcp_metrics flush PREFIX" if
several entries are present for the same destination-IP but with different
source-IPs:

root@client:~/test# ip tcp_metrics
10.2.1.2 age 3.640sec rtt 16250us rttvar 15000us cwnd 10
10.2.1.2 age 4.030sec rtt 18750us rttvar 15000us cwnd 10
root@client:~/test# ip tcp_metrics flush 10.2.1.2/16
Failed to send flush request
: No such process

Follow-up patches will modify iproute2 to handle this correctly and allow
specifying the source-IP in the get/del commands.

v2: Added the patch that allows to selectively get/del of tcp-metrics based
    on src-IP and moved the patch that adds the new netlink attribute before
    the other patches.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

We want to be able to get/del tcp-metrics based on the src IP. This
patch adds the necessary parsing of the netlink attribute and if the
source address is set, it will match on this one too.

Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

As we now can have multiple entries per destination-IP, the "ip
tcp_metrics delete address ADDRESS" command deletes all of them.

Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds a new netlink attribute for the source-IP and appends it
to the netlink reply. Now, iproute2 can have access to the source-IP.

Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

We add the source-address to the tcp-metrics, so that different metrics
will be used per source/destination-pair. We use the destination-hash to
store the metric inside the hash-table. That way, deleting and dumping
via "ip tcp_metrics" is easy.

Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

As we will add also the source-address, we rename all accesses to the
tcp-metrics address to use "daddr".

Signed-off-by: Christoph Paasch <christoph.paasch@uclouvain.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

John W. Linville says:

====================
Please pull these updates for the 3.14 stream!

For the mac80211 bits, Johannes says:

"Felix adds some helper functions for P2P NoA software tracking, Joe
fixes alignment (but as this apparently never caused issues I didn't
send it to 3.13), Kyeyoon/Jouni add QoS-mapping support (a Hotspot 2.0
feature), Weilong fixed a bunch of checkpatch errors and I get to play
fire-fighter or so and clean up other people's locking issues. I also
added nl80211 vendor-specific events, as we'd discussed at the wireless
summit."

For the iwlwifi bits, Emmanuel says:

"I have here a rework of the interrupt handling to meet RT kernel
requirements - basically we don't take any lock in the primary interrupt
handler. This gave me a good reason to clean things up a bit on the way.
There is also a fix of the QoS mapping along with a few workarounds for
hardware / firmware issues that are hard to hit.
Three fixes suggested by static analyzers, and other various stuff.
Most importantly, I update the Copyright note to include the new year."

For the bluetooth bits, Gustavo says:

"More patches to 3.14. The bulk of changes here is the 6LoWPAN support for
Bluetooth LE Devices. The commits that touches net/ieee802154/ are already
acked by David Miller. Other than that we have some RFCOMM fixes and
improvements plus fixes and clean ups all over the tree."

Beyond that, ath9k, brcmfmac, mwifiex, and wil6210 get their usual
level of attention.  The wl1251 driver gets a number of updates,
and there are a handful of other bits here and there.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

====================
This batch contains one single patch with the l2tp match
for xtables, from James Chapman.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-next into for-davem

Conflicts:
	net/ieee802154/6lowpan.c