dsa: Remove unnecessary exports

I mistakenly exported functions from slave.c that are only called from
dsa.c, part of the same module.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Acked-by: Lennert Buytenhek <buytenh@wantstofly.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'for-davem' of git://git./linux/kernel/git/linville/wireless-next

tg3: Scale back code that modifies MRRS

Tg3 normally gets a performance boost by increasing the PCI Maximum Read
Request Size (MRRS) to 4k.  Unfortunately, this is causing some problems
on particular hardware platforms.  This patch removes all code that
modifies the MRRS except for one case.

As part of a solution to fix an internal FIFO problem on the 5719, the
driver artificially capped the MRRS to 2k for the entire 5719, and later
5720, ASIC revs.  This was overly aggressive and only really needed to
be done for the 5719 A0.  In the spirit of the rest of this patch, the
driver will only reprogram the MRRS for this device if the value exceeds
the 2k cap.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tg3: Fix TSO CAP for 5704 devs w / ASF enabled

On the earliest TSO capable devices, TSO was accomplished through
firmware.  The TSO cannot coexist with ASF management firmware though.
The tg3 driver determines whether or not ASF is enabled by calling
tg3_get_eeprom_hw_cfg(), which checks a particular bit of NIC memory.
Commit dabc5c670d3f86d15ee4f42ab38ec5bd2682487d, entitled "tg3: Move
TSO_CAPABLE assignment", accidentally moved the code that determines
TSO capabilities earlier than the call to tg3_get_eeprom_hw_cfg().  As a
consequence, the driver was attempting to determine TSO capabilities
before it had all the data it needed to make the decision.

This patch fixes the problem by revisiting and reevaluating the decision
after tg3_get_eeprom_hw_cfg() is called.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sch_sfb: use skb_flow_dissect()

Current SFB double hashing is not fulfilling SFB theory, if two flows
share same rxhash value.

Using skb_flow_dissect() permits to really have better hash dispersion,
and get tunnelling support as well.

Double hashing point was mentioned by Florian Westphal

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

cls_flow: use skb_flow_dissect()

Instead of using a custom flow dissector, use skb_flow_dissect() and
benefit from tunnelling support.

This lack of tunnelling support was mentioned by Dan Siemon.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: use skb_flow_dissect() in __skb_get_rxhash()

No functional changes.

This uses the code we factorized in skb_flow_dissect()

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: introduce skb_flow_dissect()

We use at least two flow dissectors in network stack, with known
limitations and code duplication.

Introduce skb_flow_dissect() to factorize this, highly inspired from
existing dissector from __skb_get_rxhash()

Note : We extensively use skb_header_pointer(), this permits us to not
touch skb at all.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Change value comparison order

Change comparison order such that the variable will come before the compared value.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Cosmetic changes

Fix spelling, alignment, empty lines, relocate the is_4_port_mode function, and split bnx2x_link_status_update function.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Fix self test of BCM57800

Fix the MAC test of the 1G port of the BCM57800 to use the UMAC instead of the XMAC.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Add known PHY type check

The populate function will fail in case an unknown external PHY is detected.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Change Warpcore MDIO work around mode

This patch enables the usage of simpler MDC/MDIO work-around when accessing Warpcore registers.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Fix BCM84833 link and LED behavior

This patch contain several fixes for the BCM84833. This PHY is still not in bnx2x production, hence this patch can be considered as enhancement.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Warpcore HW reset following fan failure

Put Warpcore in low power mode in case of fan failure to reduce heat.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: ETS changes

Fix a problem when new traffic class is created with 0% BW, the ETS is not conforming.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: PFC changes

Change BRB to work in per class guaranteed mode and handle cases for BW 0%.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: tcp_sendmsg() wrong access to sk_route_caps

Now sk_route_caps is u64, its dangerous to use an integer to store
result of an AND operator. It wont work if NETIF_F_SG is moved on the
upper part of u64.

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

net/irda: convert drivers/net/irda/* to use module_platform_driver()

This patch converts the drivers in drivers/net/irda/* to use the
module_platform_driver() macro which makes the code smaller and a bit
simpler.

Cc: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: skip cwnd moderation in TCP_CA_Open in tcp_try_to_open

The problem: Senders were overriding cwnd values picked during an undo
by calling tcp_moderate_cwnd() in tcp_try_to_open().

The fix: Don't moderate cwnd in tcp_try_to_open() if we're in
TCP_CA_Open, since doing so is generally unnecessary and specifically
would override a DSACK-based undo of a cwnd reduction made in fast
recovery.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: allow undo from reordered DSACKs

Previously, SACK-enabled connections hung around in TCP_CA_Disorder
state while snd_una==high_seq, just waiting to accumulate DSACKs and
hopefully undo a cwnd reduction. This could and did lead to the
following unfortunate scenario: if some incoming ACKs advance snd_una
beyond high_seq then we were setting undo_marker to 0 and moving to
TCP_CA_Open, so if (due to reordering in the ACK return path) we
shortly thereafter received a DSACK then we were no longer able to
undo the cwnd reduction.

The change: Simplify the congestion avoidance state machine by
removing the behavior where SACK-enabled connections hung around in
the TCP_CA_Disorder state just waiting for DSACKs. Instead, when
snd_una advances to high_seq or beyond we typically move to
TCP_CA_Open immediately and allow an undo in either TCP_CA_Open or
TCP_CA_Disorder if we later receive enough DSACKs.

Other patches in this series will provide other changes that are
necessary to fully fix this problem.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: use SACKs and DSACKs that arrive on ACKs below snd_una

The bug: When the ACK field is below snd_una (which can happen when
ACKs are reordered), senders ignored DSACKs (preventing undo) and did
not call tcp_fastretrans_alert, so they did not increment
prr_delivered to reflect newly-SACKed sequence ranges, and did not
call tcp_xmit_retransmit_queue, thus passing up chances to send out
more retransmitted and new packets based on any newly-SACKed packets.

The change: When the ACK field is below snd_una (the "old_ack" goto
label), call tcp_fastretrans_alert to allow undo based on any
newly-arrived DSACKs and try to send out more packets based on
newly-SACKed packets.

Other patches in this series will provide other changes that are
necessary to fully fix this problem.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: use DSACKs that arrive when packets_out is 0

The bug: Senders ignored DSACKs after recovery when there were no
outstanding packets (a common scenario for HTTP servers).

The change: when there are no outstanding packets (the "no_queue" goto
label), call tcp_fastretrans_alert() in order to use DSACKs to undo
congestion window reductions.

Other patches in this series will provide other changes that are
necessary to fully fix this problem.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: make is_dupack a parameter to tcp_fastretrans_alert()

Allow callers to decide whether an ACK is a duplicate ACK. This is a
prerequisite to allowing fastretrans_alert to be called from new
contexts, such as the no_queue and old_ack code paths, from which we
have extra info that tells us whether an ACK is a dupack.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4_en: bug fix for the case of vlan id 0 and UP 0

When using vlan 0 and UP 0, vlan header wasn't placed.

Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4_en: adding loopback support

Device must be in promiscuous mode or DMAC must be same as the host MAC, or
else packet will be dropped by the HW rx filtering.

Signed-off-by: Amir Vadai <amirv@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4_en: fix WOL handlers were always looking at port2 capability bit

There are 2 capability bits for WOL, one for each port.
WOL handlers were looking only on the second bit, regardless of the port.

Signed-off-by: Oren Duer <oren@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4_en: using non collapsed CQ on TX

Moving to regular Completion Queue implementation (not collapsed)
Completion for each transmitted packet is written to new entry.

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4_en: fix sparse warning on a cast which truncates bits from constant value

the MLX4_EN_WOL_DO_MODIFY flag which is defined through enum targets
bit 63, this triggers a "cast truncate bits from constant value
(8000000000000000 becomes 0)" warning from sparse, fix that by using
define instead of enum.

Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4: fix UDP RSS related settings

Using RSS which takes into account UDP headers is controlled by
a module param, fix the setting of the HW RSS context to align
with that scheme. So far it was uncoditionally allowing hashing
on the UDP headers.

Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/mlx4: move RSS related definitions to be global

Towards adding RSS support for IB drivers/application who use
the mlx4 HW, make the RSS related definitions global and change
the mlx4_en driver to use them.

Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: Shlomo Pongratz <shlomop@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

isdn/gigaset: report ISDN4Linux interface only once

Move the "ISDN4Linux interface" message from device registration,
where it is emitted for each device, to driver registration, where
it is emitted only once, for consistency with the CAPI variant.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: David S. Miller <davem@davemloft.net>

ifenslave: Fix unused variable warnings.

Documentation/networking/ifenslave.c: In function ‘if_getconfig’:
Documentation/networking/ifenslave.c:508:14: warning: variable ‘mtu’ set but not used [-Wunused-but-set-variable]
Documentation/networking/ifenslave.c:508:6: warning: variable ‘metric’ set but not used [-Wunused-but-set-variable]

The purpose of this function is to simply print out the values
it probes, so...

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

atm: eliminate atm_guess_pdu2truesize()

Signed-off-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

virtio_net: return already tracked tx_fifo_errors via virtnet_getstats()

Tx_fifo_errors are tracked in start_xmit_ for virtio_net, but not
reported in the tallies returned by virtnet_stats().  Return them
as the rx "sub-stats" rx_length_errors and rx_frame_errors are.

Signed-off-by: Rick Jones <rick.jones2@hp.com>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

ray_cs: Fix array bounds warnings.

rx_msg is defined to have a 1 entry array at the end, so gcc warns:

drivers/net/wireless/ray_cs.c: In function ‘rx_authenticate’:
drivers/net/wireless/ray_cs.c:2436:3: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2436:3: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2436:3: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2436:3: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2436:3: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2439:15: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2452:16: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2453:18: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/ray_cs.c:2453:32: warning: array subscript is above array bounds [-Warray-bounds]

Use a zero length array and rename to "ray_rx_msg" to make sure we hit all
of the necessary cases.

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

airo: Fix array bounds warning when moving packet payload.

drivers/net/wireless/airo.c: In function ‘encapsulate’:
drivers/net/wireless/airo.c:1421:15: warning: array subscript is above array bounds [-Warray-bounds]
drivers/net/wireless/airo.c: In function ‘decapsulate’:
drivers/net/wireless/airo.c:1509:16: warning: array subscript is above array bounds [-Warray-bounds]

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

KSZ8851-SNL: Add ethtool support for EEPROM via eeprom_93cx6

Add ethtool EEPROM read/write support using the eeprom_93cx6
library instead of open-coding the functions.

Depends on eeprom_93cx6 driver getting EEPROM write support.

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Simtec Linux Team <linux@simtec.co.uk>
[sboyd@codeaurora.org: Removed previous eeprom implementation]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

KSZ8851-SNL: Fix MAC address change problem

When device is off it is under power saving mode. Changing the MAC address
in that situation will result in the device not communicating as the first
write to the MAC address register is not executed.

Signed-off-by: Tristram Ha <Tristram.Ha@micrel.com>
[ben@simtec.co.uk: cleaned up header]
Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

KSZ8851-SNL: Add support for EEPROM MAC address

Add support for reading the MAC address from the system registers if there
is an EEPROM present. This involves caching the KS_CCR register for later
use (will also be useful for ETHTOOL support) and adding a print to say
that there is an EEPROM present.

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

eeprom_93cx6: Add write support

Add support for writing data to EEPROM.

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Cc: Wolfram Sang <w.sang@pengutronix.de>
Cc: Jean Delvare <khali@linux-fr.org>
Cc: Linux Kernel <linux-kernel@vger.kernel.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

eeprom_93cx6: Add data direction control.

Some devices need to know if the data is to be output or read, so add a
data direction into the eeprom structure to tell the driver whether the
data line should be driven.

The user in this case is the Micrel KS8851 which has a direction
control for the EEPROM data line and thus needs to know whether
to drive it (writing) or to tristate it for receiving.

Signed-off-by: Ben Dooks <ben@simtec.co.uk>
Cc: Wolfram Sang <w.sang@pengutronix.de>
Cc: Jean Delvare <khali@linux-fr.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Fix non utilization of RX queues

When non power of two MSIX vectors are given to driver, some RX queues
are not utilized. Program RSS table in such a way that all queues
are utilized.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Use new hash key

This new hash key gives better distribution of packets across RX
queues.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Add error handling for Lancer

Detect error in Lancer by polling a HW register and
recover from this error if it is recoverable.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Fix error recovery paths

When TX queues are created again after error recovery,
netif_set_real_num_tx_queues() is invoked to update number of real
TX queues created. rtnl lock needs to be held when invoking this routine.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Move to new SR-IOV implementation in Lancer

SR-IOV implementation is Lancer has changed in following ways -
1)PF driver assigns one MAC addresses for VF using COMMON_SET_IFACE_MAC_LIST.
2)VF driver queries its MAC address using COMMON_GET_IFACE_MAC_LIST command
and assigns it to its interface.

Signed-off-by: Mammatha Edhala <mammatha.edhala@emulex.com>
Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bonding: Remove obsolete source file 'bond_ipv6.c'

This file is now unused and should have been removed by commit
7c89943236750537d26421d9bbb6f6575e2d1e1b ("bonding, ipv4, ipv6, vlan:
Handle NETDEV_BONDING_FAILOVER like NETDEV_NOTIFY_PEERS").

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

net: Use IS_ENABLED() in netdevice.h as appropriate

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

dsa: Allow core and drivers to be built as modules

Change the kconfig types to tristate and adjust the condition for
declaring net_device::dsa_ptr to allow for this.

Adjust the makefile so that if NET_DSA_MV88E6123_61_65=y and
NET_DSA_MV88E6131=m or vice versa then both drivers are built-in.  We
could leave these options as bool and make NET_DSA_MV88E6XXX a
user-selected option, but that would break existing configurations.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

dsa: Define module author, description, license and aliases for drivers

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

mv88e6xxx: Combine mv88e6131 and mv88e612_61_65 drivers

These drivers share a lot of code, so if we make them modular they
should be built into the same module.  Therefore, link them together
and merge their respective module init and exit functions.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

dsa: Combine core and tagging code

These files have circular dependencies, so if we make DSA modular then
they must be built into the same module.  Therefore, link them
together and merge their respective module init and exit functions.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

dsa: Export functions from core to modules

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

dsa: Change dsa_uses_{dsa, trailer}_tags() into inline functions

eth_type_trans() will use these functions if DSA is enabled, which
blocks building DSA as a module.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge git://git./linux/kernel/git/davem/net

Conflicts:
net/ipv4/inet_diag.c

net: Revert ARCNET and PHYLIB to tristate options

commit 88491d8103498a6166f70d5999902fec70924314 ("drivers/net: Kconfig
& Makefile cleanup") changed the type of these options to bool, but
they select code that could (and still can) be built as modules.

Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Don't use the cached pmtu informations for input routes

The pmtu informations on the inetpeer are visible for output and
input routes. On packet forwarding, we might propagate a learned
pmtu to the sender. As we update the pmtu informations of the
inetpeer on demand, the original sender of the forwarded packets
might never notice when the pmtu to that inetpeer increases.
So use the mtu of the outgoing device on packet forwarding instead
of the pmtu to the final destination.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

route: struct rtable can be const in rt_is_input_route and rt_is_output_route

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: Move mtu handling down to the protocol depended handlers

We move all mtu handling from dst_mtu() down to the protocol
layer. So each protocol can implement the mtu handling in
a different manner.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: Rename the dst_opt default_mtu method to mtu

We plan to invoke the dst_opt->default_mtu() method unconditioally
from dst_mtu(). So rename the method to dst_opt->mtu() to match
the name with the new meaning.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

route: Use the device mtu as the default for blackhole routes

As it is, we return null as the default mtu of blackhole routes.
This may lead to a propagation of a bogus pmtu if the default_mtu
method of a blackhole route is invoked. So return dst->dev->mtu
as the default mtu instead.

Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'for_david' of git://git.open-mesh.org/linux-merge

netns: fix proxy ARP entries listing on a netns

Skip entries from foreign network namespaces.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/netlabel: copy and paste bug in netlbl_cfg_unlbl_map_add()

This was copy and pasted from the IPv4 code.  We're calling the
ip4 version of that function and map4 is NULL.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: mv643xx_eth: fix build error

Fix below build error:
  CC      drivers/net/ethernet/marvell/mv643xx_eth.o
drivers/net/ethernet/marvell/mv643xx_eth.c: In function 'mv643xx_eth_get_drvinfo':
drivers/net/ethernet/marvell/mv643xx_eth.c:1505: error: 'info' undeclared (first use in this function)
drivers/net/ethernet/marvell/mv643xx_eth.c:1505: error: (Each undeclared identifier is reported only once
drivers/net/ethernet/marvell/mv643xx_eth.c:1505: error: for each function it appears in.)
make[4]: *** [drivers/net/ethernet/marvell/mv643xx_eth.o] Error 1

Signed-off-by: Axel Lin <axel.lin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4: Save nexthop address of LSRR/SSRR option to IPCB.

We can not update iph->daddr in ip_options_rcv_srr(), It is too early.
When some exception ocurred later (eg. in ip_forward() when goto
sr_failed) we need the ip header be identical to the original one as
ICMP need it.

Add a field 'nexthop' in struct ip_options to save nexthop of LSRR
or SSRR option.

Signed-off-by: Li Wei <lw@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ehea: Use round_jiffies_relative to align workqueue

Use round_jiffies_relative to align the ehea workqueue and avoid
extra wakeups.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

ehea: Reduce memory usage in buffer pools

Now that we enable multiqueue by default the ehea driver is using
quite a lot of memory for its buffer pools. With 4 queues we
consume 64MB in the jumbo packet ring, 16MB in the medium packet
ring and 16MB in the tiny packet ring.

We should only fill the jumbo ring once the MTU is increased but
for now halve it's size so it consumes 32MB. Also reduce the tiny
packet ring, with 4 queues we had 16k entries which is overkill.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

qlge: fix size of external list for TX address descriptors

When transmiting a fragmented skb, qlge fills a descriptor with the
fragment addresses, after DMA-mapping them. If there are more than eight
fragments, it will use the eighth descriptor as a pointer to an external
list. After mapping this external list, called OAL to a structure
containing more descriptors, it fills it with the extra fragments.

However, considering that systems with pages larger than 8KiB would have
less than 8 fragments, which was true before commit a715dea3c8e, it
defined a macro for the OAL size as 0 in those cases.

Now, if a skb with more than 8 fragments (counting skb->data as one
fragment), this would start overwriting the list of addresses already
mapped and would make the driver fail to properly unmap the right
addresses on architectures with pages larger than 8KiB.

Besides that, the list of mappings was one size too small, since it must
have a mapping for the maxinum number of skb fragments plus one for
skb->data and another for the OAL. So, even on architectures with page
sizes 4KiB and 8KiB, a skb with the maximum number of fragments would
make the driver overwrite its counter for the number of mappings, which,
again, would make it fail to unmap the mapped DMA addresses.

Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Fix 5461x LED

Fix port identify test on 5461x PHY by driving LEDs through MDIO.

Signed-off-by: Yaniv Rosner <yanivr@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: treewide use of RCU_INIT_POINTER

rcu_assign_pointer(ptr, NULL) can be safely replaced by
RCU_INIT_POINTER(ptr, NULL)

(old rcu_assign_pointer() macro was testing the NULL value and could
omit the smp_wmb(), but this had to be removed because of compiler
warnings)

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv4 : igmp : fix error handle in ip_mc_add_src()

When add sources to interface failure, need to roll back the sfcount[MODE]
to before state. We need to match it corresponding.

Acked-by: David L Stevens <dlstevens@us.ibm.com>
Acked-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Jun Zhao <mypopydev@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: tcp: fix tcp_v6_conn_request()

Since linux 2.6.26 (commit c6aefafb7ec6 : Add IPv6 support to TCP SYN
cookies), we can drop a SYN packet reusing a TIME_WAIT socket.

(As a matter of fact we fail to send the SYNACK answer)

As the client resends its SYN packet after a one second timeout, we
accept it, because first packet removed the TIME_WAIT socket before
being dropped.

This probably explains why nobody ever noticed or complained.

Reported-by: Jesse Young <jlyo@jlyo.org>
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

b44: Use dev_kfree_skb_irq() in b44_tx()

Reported issues when using dev_kfree_skb() on UP systems and
systems with low numbers of cores.  dev_kfree_skb_irq() will
properly save IRQ state before freeing the skb.

Tested on 3.1.1 and 3.2_rc2

Example of reproducible trace of kernel 3.1.1
------------[ cut here ]------------
   WARNING: at kernel/softirq.c:159 local_bh_enable+0x32/0x79()
   ...
   Pid: 0, comm: swapper Not tainted 3.1.1-gentoo #1
   Call Trace:
    [<c1022970>] warn_slowpath_common+0x65/0x7a
    [<c102699e>] ? local_bh_enable+0x32/0x79
    [<c1022994>] warn_slowpath_null+0xf/0x13
    [<c102699e>] local_bh_enable+0x32/0x79
    [<c134bfd8>] destroy_conntrack+0x7c/0x9b
    [<c134890b>] nf_conntrack_destroy+0x1f/0x26
    [<c132e3a6>] skb_release_head_state+0x74/0x83
    [<c132e286>] __kfree_skb+0xb/0x6b
    [<c132e30a>] consume_skb+0x24/0x26
    [<c127c925>] b44_poll+0xaa/0x449
    [<c1333ca1>] net_rx_action+0x3f/0xea
    [<c1026a44>] __do_softirq+0x5f/0xd5
    [<c10269e5>] ? local_bh_enable+0x79/0x79
    <IRQ>  [<c1026c32>] ? irq_exit+0x34/0x8d
    [<c1003628>] ? do_IRQ+0x74/0x87
    [<c13f5329>] ? common_interrupt+0x29/0x30
    [<c1006e18>] ? default_idle+0x29/0x3e
    [<c10015a7>] ? cpu_idle+0x2f/0x5d
    [<c13e91c5>] ? rest_init+0x79/0x7b
    [<c15c66a9>] ? start_kernel+0x297/0x29c
    [<c15c60b0>] ? i386_start_kernel+0xb0/0xb7
   ---[ end trace 583f33bb1aa207a9 ]---

Signed-off-by: Xander Hover <LKML@hover.be>
Signed-off-by: David S. Miller <davem@davemloft.net>

netfilter: Remove NOTRACK/RAW dependency on NETFILTER_ADVANCED.

Distributions are using this in their default scripts, so don't hide
them behind the advanced setting.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

netprio_cgroup: Fix build break

I broke the build with the addition of netprio_cgroups if CONFIG_CGROUPS=n.
This patch corrects it by moving the offending struct into an ifdef
CONFIG_CGROUPS block.  Also clean up a few needless defines and inline functions
that don't get called if CONFIG_CGROUPS isn't defined while Im at it.

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: tcp: fix panic in SYN processing

commit 72a3effaf633bc ([NET]: Size listen hash tables using backlog
hint) added a bug allowing inet6_synq_hash() to return an out of bound
array index, because of u16 overflow.

Bug can happen if system admins set net.core.somaxconn &
net.ipv4.tcp_max_syn_backlog sysctls to values greater than 65536

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: fix a bug in ndisc_send_redirect

Release skb when transmit rate limit _not_ allow

Signed-off-by: Li Wei <lw@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

net: remove netdev_alloc_page and use __GFP_COLD

Given we dont use anymore the struct net_device *dev argument, and this
interface brings litle benefit, remove netdev_{alloc|free}_page(), to
debloat include/linux/skbuff.h a bit.

(Some drivers used a mix of these interfaces and alloc_pages())

When allocating a page given to device for DMA transfer (device to
memory), it makes sense to use a cold one (__GFP_COLD)

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
CC: Dimitris Michailidis <dm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Sweep away N/A fw_version dustbunnies from the .get_drvinfo routine of a number of drivers

Per discussion with Ben Hutchings and David Miller, go through and
remove assignments of "N/A" to fw_version in various drivers'
.get_drvinfo routines.  While there clean-up some use of bare
constants and such.

Signed-off-by: Rick Jones <rick.jones2@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: remove ipv6_addr_copy()

C assignment can handle struct in6_addr copying.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Revert "rt2800pci: handle spurious interrupts"

This reverts commit 4ba7d9997869d25bd223dea7536fc1ce9fab3b3b.

The original patch was a misguided attempt to improve performance on
some hardware that is apparently prone to spurious interrupt generation.

Signed-off-by: John W. Linville <linville@tuxdriver.com>

tc: comment spelling fixes

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Revert "rt2x00: handle spurious pci interrupts"

This reverts commit 23085d5796561625db4143a671f1de081f66ef08.

The original patch was a misguided attempt to improve performance on
some hardware that is apparently prone to spurious interrupt generation.

Signed-off-by: John W. Linville <linville@tuxdriver.com>

net: correct comments of skb_shift

when skb_shift, we want to shift paged data from skb to tgt frag area.
Original comments revert the shift order

Signed-off-by: Feng King <kinwin2008@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

atm: Allow MSG_PEEK for atm sockets

Now that the vcc backends do the right thing with respect the receive
queue on registration, allow MSK_PEEK for atm sockets.

This allows a userspace program to inspect the packets and decide what
backend to use to handle them.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

atm: Introduce vcc_process_recv_queue

This function moves the implementation found in the clip and br2684
modules to common code, correctly unlinks the skb from the queue
before pushing it and makes pppoatm use it.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

atm: clip: move clip_devs check to clip_push

This will allow further cleanup.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

atm: clip: Don't move counters backwards

I don't see the point on substracting the skb len from the netdev
stats.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

atm: br2684: Do not move counters backwards

This snippet has caused several bugs in the past, and I don't see the
point on substracting the skb len from netdev stats.

Signed-off-by: Jorge Boncompte [DTI2] <jorge@dti2.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

net-netlink: fix diag to export IPv4 tos for dual-stack IPv6 sockets

Signed-off-by: Maciej Żenczykowski <maze@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tg3: Add MDI-X reporting

This patch adds MDI-X state reporting.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tg3: Restrict large prod ring cap devices

Future devices may or may not be capable of supporting larger rx
producer rings.  This patch changes the code so that this flag is set on
an ASIC rev to ASIC rev basis.  Also, this patch changes a place where
the LRG_PROD_RING_CAP flag was not controlling how the rx standard
producer ring size was set.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tg3: Adjust BD replenish thresholds

The BD replenish thresholds for the 57765 and newer ASIC revs are a
little strict.  They were tuned for a mode that is currently unused.
This patch relaxes the thresholds so that they are set to values more
inline with the resources available.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tg3: Make 1000Base-X FC resolution look like 1000T

This patch changes tg3's 1000Base-X flow control resolution to look like
the 1000Base-T flow control resolution code.

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

CDC NCM: Use kzalloc rather than kmalloc followed by memset with 0

This considers some simple cases that are common and easy to validate
Note in particular that there are no ...s in the rule, so all of the
matched code has to be contiguous

The semantic patch that makes this change is available
in scripts/coccinelle/api/alloc/kzalloc-simple.cocci.

Signed-off-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

xfrm: optimize ipv4 selector matching

Current addr_match() is errh, under-optimized.

Compiler doesn't know that memcmp() branch doesn't trigger for IPv4.
Also, pass addresses by value -- they fit into register.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>