Merge branch 'master' of /linux/kernel/git/davem/net-2.6

ixgbe: Allow Priority Flow Control settings to survive a device reset

When changing DCB parameters, ixgbe needs to have the MAC reset.  The way
the flow control code is setup today, PFC will be disabled on a reset.
This patch adds a new flow control type for PFC, and then has the netlink
layer take care of toggling which type of flow control to enable.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: core: remove unneeded include in net/core/utils.c.

Signed-off-by: Rami Rosen <ramirose@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000e: update version number

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000e: fix close interrupt race

As noticed by Alan Cox, it is possible for e1000e to exit its interrupt
handler or NAPI with interrupts enabled even when the driver is unloading or
being configured administratively down.

fix related to fix for: http://bugzilla.kernel.org/show_bug.cgi?id=12876

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000e: fix loss of multicast packets

e1000e (and e1000, igb, ixgbe, ixgb) all do a series of operations each
time a multicast address is added.  The flow goes something like

1) stack adds one multicast address
2) stack passes whole current list of unicast and multicast addresses to
   driver
3) driver clears entire list in hardware
4) driver programs each multicast address using iomem in a loop

This was causing multicast packets to be lost during the reprogramming
process.

reference with test program:
http://kerneltrap.org/mailarchive/linux-netdev/2009/3/14/5160514/thread

Thanks to Dave Boutcher for his report and test program.

This driver fix prepares an array all at once in memory and programs it in
one shot to the hardware, not requiring an "erase" cycle.  It would still
be possible for packets to be dropped while the receiver is off during
reprogramming.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
CC: Dave Boutcher <daveboutcher@gmail.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000e: commonize tx cleanup routine to match e1000 & igb

This change updates the e1000e tx cleanup routine to more closely match
what already exists in igb and e1000.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

netfilter: fix nf_logger name in ebt_ulog.

This patch renames the ebt_ulog nf_logger from "ulog" to "ebt_ulog" to
be in sync with other modules naming. As this name was currently only
used for informational purpose, the renaming should be harmless.

Signed-off-by: Eric Leblond <eric@inl.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

netfilter: fix warning in ebt_ulog init function.

The ebt_ulog module does not follow the fixed convention about function
return. Loading the module is triggering the following message:

sys_init_module: 'ebt_ulog'->init suspiciously returned 1, it should follow 0/-E convention
sys_init_module: loading module anyway...
Pid: 2334, comm: modprobe Not tainted 2.6.29-rc5edenwall0-00883-g199e57b #146
Call Trace:
 [<c0441b81>] ? printk+0xf/0x16
 [<c02311af>] sys_init_module+0x107/0x186
 [<c0202cfa>] syscall_call+0x7/0xb

The following patch fixes the return treatment in ebt_ulog_init()
function.

Signed-off-by: Eric Leblond <eric@inl.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

netfilter: fix warning about invalid const usage

This patch fixes the declaration of the logger structure in ebt_log
and ebt_ulog: I forgot to remove the const option from their declaration
in the commit ca735b3aaa945626ba65a3e51145bfe4ecd9e222 ("netfilter:
use a linked list of loggers").

Pointed-out-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Eric Leblond <eric@inl.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000: fix close race with interrupt

this is in regards to
http://bugzilla.kernel.org/show_bug.cgi?id=12876

where it appears that e1000 can leave its interrupt enabled after
exiting the driver.  Fix the bug by making the interrupt enable
paths more aware of the driver exiting.

Thanks to Alan Cox for the poke and initial investigation.

CC: Alan Cox <alan@lxorguk.ukuu.org.uk>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000: cleanup clean_tx_irq routine so that it completely cleans ring

The tx cleanup routine was stopping after 64 packets and this was causing
issues resulting in the ring not being completely cleaned.

This change updates the driver to clean the entire ring and if it doesn't
it then will retry on the next pass.

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000: fix tx hang detect logic and address dma mapping issues

This patch changes the dma mapping to better support
skb_dma_map/skb_dma_unmap and addresses and redefines the tx hang logic to
be based off of time stamp instead of if the dma field is populated

Signed-off-by: Alexander Duyck <alexander.h.duyck@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bridge: bad error handling when adding invalid ether address

This fixes an crash when empty bond device is added to a bridge.
If an interface with invalid ethernet address (all zero) is added
to a bridge, then bridge code detects it when setting up the forward
databas entry. But the error unwind is broken, the bridge port object
can get freed twice: once when ref count went to zeo, and once by kfree.
Since object is never really accessible, just free it.

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

bonding: select current active slave when enslaving device for mode tlb and alb

I've hit an issue on my system when I've been using RealTek RTL8139D cards in
bonding interface in mode balancing-alb. When I enslave a card, the current
active slave (bond->curr_active_slave) is not set and the link is therefore
not functional.

----
# cat /proc/net/bonding/bond0
Ethernet Channel Bonding Driver: v3.5.0 (November 4, 2008)

Bonding Mode: adaptive load balancing
Primary Slave: None
Currently Active Slave: None
MII Status: up
MII Polling Interval (ms): 100
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth1
MII Status: up
Link Failure Count: 0
Permanent HW addr: 00:1f:1f:01:2f:22
----

The thing that gets it right is when I unplug the cable and then I put it back
into the NIC. Then the current active slave is set to eth1 and link is working
just fine. Here is dmesg log with bonding DEBUG messages turned on:
----
ADDRCONF(NETDEV_UP): bond0: link is not ready
event_dev: bond0, event: 1
IFF_MASTER
event_dev: bond0, event: 8
IFF_MASTER
bond_ioctl: master=bond0, cmd=35216
slave_dev=cac5d800:
slave_dev->name=eth1:
eth1: ! NETIF_F_VLAN_CHALLENGED
event_dev: eth1, event: 8
eth1: link up, 100Mbps, full-duplex, lpa 0xC5E1
event_dev: eth1, event: 1
event_dev: eth1, event: 8
IFF_SLAVE
Initial state of slave_dev is BOND_LINK_UP
bonding: bond0: enslaving eth1 as an active interface with an up link.
ADDRCONF(NETDEV_CHANGE): bond0: link becomes ready
event_dev: bond0, event: 4
IFF_MASTER
bond0: no IPv6 routers present

<<<<cable unplug>>>>

eth1: link down
event_dev: eth1, event: 4
IFF_SLAVE
bonding: bond0: link status definitely down for interface eth1, disabling it
event_dev: bond0, event: 4
IFF_MASTER

<<<<cable plug>>>>

eth1: link up, 100Mbps, full-duplex, lpa 0xC5E1
event_dev: eth1, event: 4
IFF_SLAVE
bonding: bond0: link status definitely up for interface eth1.
bonding: bond0: making interface eth1 the new active one.
event_dev: eth1, event: 8
IFF_SLAVE
event_dev: eth1, event: 8
IFF_SLAVE
bonding: bond0: first active interface up!
event_dev: bond0, event: 4
IFF_MASTER
----

The current active slave is set by calling bond_select_active_slave() function
from bond_miimon_commit() function when the slave (eth1) link goes to state up.

I also tested this on other machine with Broadcom NetXtreme II BCM5708
1000Base-T NIC and there all works fine. The thing is that this adapter is down
and goes up after few seconds after it is enslaved.

This patch calls bond_select_active_slave() in bond_enslave() function for modes
alb and tlb and makes sure that the current active slave is set up properly even
when the slave state is already up. Tested on both systems, works fine.

Notice: The same problem can maybe also occrur in mode 8023AD but I'm unable to
test that.

Signed-off-by: Jiri Pirko <jpirko@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

gianfar: reallocate skb when headroom is not enough for fcb

Gianfar uses a hardware header FCB for offloading.  However when used
with bridging or IP forwarding, TX skb might not have enough headroom
for the FCB.  Reallocate skb for such cases.

Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Bump release date to 25Mar2009 and version to 0.22

This patch bumps the driver release date to March 25th 2009
and release version to 0.22.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

r6040: Fix second PHY address

This patch fixes the second PHY address which is strapped
to be at PHY address 3 instead of 2.

Signed-off-by: Florian Fainelli <florian@openwrt.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: fix wait_event_timeout handling

wait_event_timeout just takes the numnber of jiffies to wait as
an argument. That value does not include jiffies itself.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: check for completion of a running recovery

When a recovery is started for a qeth device, additional invocations
to change a mac address, to configure a VLAN interface on top, or to
add multicast addresses should wait till recovery is finished,
otherwise recovery might fail.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: unregister MAC addresses during recovery.

qeth: Unregister MAC addresses from device (layer 2) during
recovery cycle. When the device is set online the MAC
addresses are registered again on the device.

Signed-off-by: Klaus-Dieter Wacker <kdwacker@de.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: remove EDDP

Performance measurements showed EDDP does not lower CPU costs but increase
them. So we dump out EDDP code from qeth driver.

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: add statistics for tx csum

Add statistics counter for software tx checksumming.

Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: struct device - replace bus_id with dev_name(), dev_set_name()

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Acked-by: Greg Kroah-Hartman <gregkh@suse.de>
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Frank Blaschka <frank.blaschka@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

rndis_wlan: Fix build with netdev_ops compat disabled.

Instead of storing a private ->set_multicast_list, just
have a private netdev ops.

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

ax88796: Add method to take MAC from platform data

Implement a way to provide the MAC address for ax88796 devices from
their platform data. Boards might decide to set the address
programmatically, taken from boot tags or other sources.

Signed-off-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ax88796.c: take IRQ flags from platform_device

This patch adds support to the ax88796 ethernet driver to take IRQ flags
given by the platform_device definition.

Signed-off-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

e1000e: add support for 82574 device ID 0x10F6

Add device ID for a new variant of the 82574 adapter.

Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Acked-by: John Ronciak <john.ronciak@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: Fix conflict resolutions during ipv6 binding

The ipv6 version of bind_conflict code calls ipv6_rcv_saddr_equal()
which at times wrongly identified intersections between addresses.
It particularly broke down under a few instances and caused erroneous
bind conflicts.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: Make v4-mapped bindings consistent with IPv4

Binding to a v4-mapped address on an AF_INET6 socket should
produce the same result as binding to an IPv4 address on
AF_INET socket.  The two are interchangable as v4-mapped
address is really a portability aid.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: Allow ipv4 wildcard binds after ipv6 address binds

The IPv4 wildcard (0.0.0.0) address does not intersect
in any way with explicit IPv6 addresses.  These two should
be permitted, but the IPv4 conflict code checks the ipv6only
bit as part of the test.  Since binding to an explicit IPv6
address restricts the socket to only that IPv6 address, the
side-effect is that the socket behaves as v6-only.  By
explicitely setting ipv6only in this case, allows the 2 binds
to succeed.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6: Disallow binding to v4-mapped address on v6-only socket.

A socket marked v6-only, can not receive or send traffic to v4-mapped
addresses.  Thus allowing binding to v4-mapped address on such a
socket makes no sense.

Signed-off-by: Vlad Yasevich <vladislav.yasevich@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: cleanup rx/tx rate calculations

Hi, Pls accept this patch to cleanup rx/tx rate calculations as follows:
- check for jiffies wraparound
- remove typecast of a denominator
- do rate calculation only in workqueue context periodically

Signed-off-by: Sathya Perla <sathyap@serverengines.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'master' of git://git./linux/kernel/git/kaber/nf-2.6

netlink: add NETLINK_NO_ENOBUFS socket flag

This patch adds the NETLINK_NO_ENOBUFS socket flag. This flag can
be used by unicast and broadcast listeners to avoid receiving
ENOBUFS errors.

Generally speaking, ENOBUFS errors are useful to notify two things
to the listener:

a) You may increase the receiver buffer size via setsockopt().
b) You have lost messages, you may be out of sync.

In some cases, ignoring ENOBUFS errors can be useful. For example:

a) nfnetlink_queue: this subsystem does not have any sort of resync
method and you can decide to ignore ENOBUFS once you have set a
given buffer size.

b) ctnetlink: you can use this together with the socket flag
NETLINK_BROADCAST_SEND_ERROR to stop getting ENOBUFS errors as
you do not need to resync (packets whose event are not delivered
are drop to provide reliable logging and state-synchronization).

Moreover, the use of NETLINK_NO_ENOBUFS also reduces a "go up, go down"
effect in terms of performance which is due to the netlink congestion
control when the listener cannot back off. The effect is the following:

1) throughput rate goes up and netlink messages are inserted in the
receiver buffer.
2) Then, netlink buffer fills and overruns (set on nlk->state bit 0).
3) While the listener empties the receiver buffer, netlink keeps
dropping messages. Thus, throughput goes dramatically down.
4) Then, once the listener has emptied the buffer (nlk->state
bit 0 is set off), goto step 1.

This effect is easy to trigger with netlink broadcast under heavy
load, and it is more noticeable when using a big receiver buffer.
You can find some results in [1] that show this problem.

[1] http://1984.lsi.us.es/linux/netlink/

This patch also includes the use of sk_drop to account the number of
netlink messages drop due to overrun. This value is shown in
/proc/net/netlink.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

myri10ge: update firmware headers to 1.4.41

Update myri10ge firmware headers to firmware version 1.4.41.

Signed-off-by: Brice Goglin <brice@myri.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

macb: fix warning "warning: unused variable `dev' "

Removed unused variable dev

Signed-off-by: vibi sreenivasan <vibi_sreenivasan@cms.com>
Signed-off-by: Haavard Skinnemoen <haavard.skinnemoen@atmel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ucc_geth: Convert to net_device_ops

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

claw: fix minor findings from code analysis tool

This patch fixes two problems in the claw driver identified by
static code analysis:
o Change in case differentiation of received sense codes
o Use correct data length in claw hard_start_xmit routine

Signed-off-by: Andrew H. Richter <richtera@us.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ctcm: fix minor findings from code analysis tool

From: Ursula Braun <ursula.braun@de.ibm.com>

This patch fixes problems in the ctcm driver identified by
static code analysis:
o remove an unnecessary always true condition in ctcm_unpack_skb
o remove duplicate assignment in ctc_mpc_alloc_channel
o remove an unnecessary always true condition in ctcmpc_send_sweep_resp
o remove duplicate initialization in ctcmpc_unpack_skb
o shorten if condition in mpc_action_go_inop
o remove INOP event if mpc group is undefined in mpc_action_doxid7

Signed-off-by: Joel A. Fowler <fowlerja@us.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ctcm: avoid wraparound in length of incoming data

Since the receive code should tolerate any incoming garbage, it
should be protected against a potential wraparound when manipulating
length values within incoming data.
block_len is unsigned, so a too large subtraction will cause a
wraparound.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ctcm: invalid return code from hard_start_xmit

Avoid kernel warning by using the correct hard_start_xmit return
code NETDEV_TX_BUSY for skb requeuing.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

claw: invalid return codes from hard_start_xmit

Avoid kernel warnings by using the correct hard_start_xmit return
code NETDEV_TX_BUSY for skb requeuing.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

netiucv: invalid return code from hard_start_xmit

Avoid kernel warning by using the correct hard_start_xmit return
code NETDEV_TX_BUSY for skb requeuing.

Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

lcs: invalid return codes from hard_start_xmit.

Lcs hard_start_xmit routine issued return codes other than
defined for this interface. Now lcs returns only either
NETDEV_TX_OK or NETDEV_TX_BUSY.

Signed-off-by: Klaus-Dieter Wacker <kdwacker@de.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use kthread instead of kernel_thread

Lcs uses low-level kernel_thread implementation.
All drivers should use <linux/kthread.h> API instead.

Signed-off-by: Klaus-Dieter Wacker <kdwacker@de.ibm.com>
Signed-off-by: Ursula Braun <ursula.braun@de.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: Add dependent headers to trace/skb.h

The tracing header needs to include definitions for the macros used and
the types referenced.  This lets automated tracing tools like SystemTap
make use of the tracepoint without any specific knowledge of its
meaning (leaving that to the user).

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

arp_tables: ifname_compare() can assume 16bit alignment

Arches without efficient unaligned access can still perform a loop
assuming 16bit alignment in ifname_compare()

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

netfilter: trivial Kconfig spelling fixes

Supplements commit 67c0d57930ff9a24c6c34abee1b01f7716a9b0e2.

Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'master' of git://git./linux/kernel/git/kaber/nf-next-2.6

dnet: drivers/net/dnet.c needs <linux/io.h>

On m68k:
| drivers/net/dnet.c: In function 'dnet_readw_mac':
| drivers/net/dnet.c:36: error: implicit declaration of function 'writel'
| drivers/net/dnet.c:43: error: implicit declaration of function 'readl'
| drivers/net/dnet.c: In function 'dnet_probe':
| drivers/net/dnet.c:873: error: implicit declaration of function 'ioremap'
| drivers/net/dnet.c:873: warning: assignment makes pointer from integer without a cast
| drivers/net/dnet.c:939: error: implicit declaration of function 'iounmap'

Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

ucc_geth: Fix build breakage caused by a merge

This patch fixes following build error:

  CC      ucc_geth.o
ucc_geth.c: In function 'ucc_geth_probe':
ucc_geth.c:3644: error: implicit declaration of function 'uec_mdio_bus_name'
make[2]: *** [ucc_geth.o] Error 1

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

netfilter: nf_conntrack: Reduce conntrack count in nf_conntrack_free()

We use RCU to defer freeing of conntrack structures. In DOS situation, RCU might
accumulate about 10.000 elements per CPU in its internal queues. To get accurate
conntrack counts (at the expense of slightly more RAM used), we might consider
conntrack counter not taking into account "about to be freed elements, waiting
in RCU queues". We thus decrement it in nf_conntrack_free(), not in the RCU
callback.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Tested-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Signed-off-by: Patrick McHardy <kaber@trash.net>

ucc_geth: Fix merge error.

I left a merge failure unresolved, noticed by Stephen
Rothwell.

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

fsl_pq_mdio: Use proper address translation

Currently the driver just read "reg" property for constructing MDIO
bus IDs, but this won't work when we'll start using "ranges = <>" in
the device tree, so this will pop up:

Freescale PowerQUICC MII Bus: probed
sysfs: duplicate filename 'mdio@520' can not be created
------------[ cut here ]------------
Badness at c00cb6b8 [verbose debug info unavailable]
NIP: c00cb6b8 LR: c00cb6b8 CTR: c001271c
REGS: cf82fc10 TRAP: 0700   Not tainted  (2.6.29-rc7-03702-g7ccd10f)
MSR: 00029032 <EE,ME,CE,IR,DR>  CR: 42044022  XER: 20000000
TASK = cf81fbd0[1] 'swapper' THREAD: cf82e000
GPR00: c00cb6b8 cf82fcc0 cf81fbd0 0000003b 00000e42 ffffffff 00004000 00000e42
GPR08: c03cb0fc c03bfbdc 00000e42 c03cac50 22044022 1006a2bc 0ffcb000 00000000
GPR16: 0ffc04b0 0ffc5a40 00000000 0ffc79a8 0f7863a8 00000004 00000000 00000000
GPR24: c033a6a8 d1014520 cf85e840 cf82fd08 cf87cf2c cf82fcd8 cf85dea8 ffffffef
NIP [c00cb6b8] sysfs_add_one+0x4c/0x54
LR [c00cb6b8] sysfs_add_one+0x4c/0x54
Call Trace:
[cf82fcc0] [c00cb6b8] sysfs_add_one+0x4c/0x54 (unreliable)
[cf82fcd0] [c00cbc18] create_dir+0x58/0xc0
[cf82fd00] [c00cbcc0] sysfs_create_dir+0x40/0x70
[cf82fd20] [c0159388] create_dir+0x28/0x78
[cf82fd30] [c0159824] kobject_add_internal+0x98/0x13c
[cf82fd50] [c0159e98] kobject_add+0x60/0x98
[cf82fd80] [c018a480] device_add+0x98/0x2ac
[cf82fda0] [c01a2380] mdiobus_register+0xbc/0x1c0
[cf82fdc0] [c019f31c] fsl_pq_mdio_probe+0x284/0x2a0
[cf82fe00] [c0223814] of_platform_device_probe+0x5c/0x84
...

This patch fixes the issue by translating the "reg" property to a full
address, and thus avoids the duplicate names.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

udp: Wrong locking code in udp seq_file infrastructure

Reading zero bytes from /proc/net/udp or other similar files which use
the same seq_file udp infrastructure panics kernel in that way:

=====================================
[ BUG: bad unlock balance detected! ]
-------------------------------------
read/1985 is trying to release lock (&table->hash[i].lock) at:
[<ffffffff81321d83>] udp_seq_stop+0x27/0x29
but there are no more locks to release!

other info that might help us debug this:
1 lock held by read/1985:
 #0:  (&p->lock){--..}, at: [<ffffffff810eefb6>] seq_read+0x38/0x348

stack backtrace:
Pid: 1985, comm: read Not tainted 2.6.29-rc8 #9
Call Trace:
 [<ffffffff81321d83>] ? udp_seq_stop+0x27/0x29
 [<ffffffff8106dab9>] print_unlock_inbalance_bug+0xd6/0xe1
 [<ffffffff8106db62>] lock_release_non_nested+0x9e/0x1c6
 [<ffffffff810ef030>] ? seq_read+0xb2/0x348
 [<ffffffff8106bdba>] ? mark_held_locks+0x68/0x86
 [<ffffffff81321d83>] ? udp_seq_stop+0x27/0x29
 [<ffffffff8106dde7>] lock_release+0x15d/0x189
 [<ffffffff8137163c>] _spin_unlock_bh+0x1e/0x34
 [<ffffffff81321d83>] udp_seq_stop+0x27/0x29
 [<ffffffff810ef239>] seq_read+0x2bb/0x348
 [<ffffffff810eef7e>] ? seq_read+0x0/0x348
 [<ffffffff8111aedd>] proc_reg_read+0x90/0xaf
 [<ffffffff810d878f>] vfs_read+0xa6/0x103
 [<ffffffff8106bfac>] ? trace_hardirqs_on_caller+0x12f/0x153
 [<ffffffff810d88a2>] sys_read+0x45/0x69
 [<ffffffff8101123a>] system_call_fastpath+0x16/0x1b
BUG: scheduling while atomic: read/1985/0xffffff00
INFO: lockdep is turned off.
Modules linked in: cpufreq_ondemand acpi_cpufreq freq_table dm_multipath kvm ppdev snd_hda_codec_analog snd_hda_intel snd_hda_codec snd_hwdep snd_seq_dummy snd_seq_oss snd_seq_midi_event arc4 snd_s
eq ecb thinkpad_acpi snd_seq_device iwl3945 hwmon sdhci_pci snd_pcm_oss sdhci rfkill mmc_core snd_mixer_oss i2c_i801 mac80211 yenta_socket ricoh_mmc i2c_core iTCO_wdt snd_pcm iTCO_vendor_support rs
rc_nonstatic snd_timer snd lib80211 cfg80211 soundcore snd_page_alloc video parport_pc output parport e1000e [last unloaded: scsi_wait_scan]
Pid: 1985, comm: read Not tainted 2.6.29-rc8 #9
Call Trace:
 [<ffffffff8106b456>] ? __debug_show_held_locks+0x1b/0x24
 [<ffffffff81043660>] __schedule_bug+0x7e/0x83
 [<ffffffff8136ede9>] schedule+0xce/0x838
 [<ffffffff810d7972>] ? fsnotify_access+0x5f/0x67
 [<ffffffff810112d0>] ? sysret_careful+0xb/0x37
 [<ffffffff8106be9c>] ? trace_hardirqs_on_caller+0x1f/0x153
 [<ffffffff8137127b>] ? trace_hardirqs_on_thunk+0x3a/0x3f
 [<ffffffff810112f6>] sysret_careful+0x31/0x37
read[1985]: segfault at 7fffc479bfe8 ip 0000003e7420a180 sp 00007fffc479bfa0 error 6
Kernel panic - not syncing: Aiee, killing interrupt handler!

udp_seq_stop() tries to unlock not yet locked spinlock. The lock was lost
during splitting global udp_hash_lock to subsequent spinlocks.

Signed-off by: Vitaly Mayatskikh <v.mayatskih@gmail.com>
Acked-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

phy: add new LAN8710 and LAN8720 device ids to smsc phy driver

LAN8710 and LAN8720 are two new 10/100 ethernet PHY models. The two
share the same phy id, this patch adds it to the smsc phy driver.

Signed-off-by: Steve Glendinning <steve.glendinning@smsc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

spider_net: Add missing .ndo_validate_addr

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

myri_sbus/sunbmac/sunlance/sunqe: Add missing net_device_ops entries.

Noticed by Stephen Hemminger.

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

Merge branch 'master' of /linux/kernel/git/davem/net-2.6

Conflicts:
drivers/net/ucc_geth.c

sunqe: Convert to net_device_ops.

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

sunlance: Convert to net_device_ops.

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

sunbmac: Convert to net_device_ops.

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

myri_sbus: Convert to net_device_ops.

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

netfilter: nf_conntrack_tcp: fix unaligned memory access in tcp_sack

This patch fixes an unaligned memory access in tcp_sack while reading
sequence numbers from TCP selective acknowledgement options.  Prior to
applying this patch, upstream linux-2.6.27.20 was occasionally
generating messages like this on my sparc64 system:

  [54678.532071] Kernel unaligned access at TPC[6b17d4] tcp_packet+0xcd4/0xd00

Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Patrick McHardy <kaber@trash.net>

nefilter: nfnetlink: add nfnetlink_set_err and use it in ctnetlink

This patch adds nfnetlink_set_err() to propagate the error to netlink
broadcast listener in case of memory allocation errors in the
message building.

Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Patrick McHardy <kaber@trash.net>

netfilter: sysctl support of logger choice

This patchs adds support of modification of the used logger via sysctl.
It can be used to change the logger to module that can not use the bind
operation (ipt_LOG and ipt_ULOG). For this purpose, it creates a
directory /proc/sys/net/netfilter/nf_log which contains a file
per-protocol. The content of the file is the name current logger (NONE if
not set) and a logger can be setup by simply echoing its name to the file.
By echoing "NONE" to a /proc/sys/net/netfilter/nf_log/PROTO file, the
logger corresponding to this PROTO is set to NULL.

Signed-off-by: Eric Leblond <eric@inl.fr>
Signed-off-by: Patrick McHardy <kaber@trash.net>

atl1c: remove duplicated #include

Remove duplicated #include in drivers/net/atl1c/atl1c.h.

Signed-off-by: Huang Weiyi <weiyi.huang@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

spider_net: Fix build.

Based upon a report by Stephen Rothwell.

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

tcp: Discard segments that ack data not yet sent

Discard incoming packets whose ack field iincludes data not yet sent.
This is consistent with RFC 793 Section 3.9.

Change tcp_ack() to distinguish between too-small and too-large ack
field values.  Keep segments with too-large ack fields out of the fast
path, and change slow path to discard them.

Reported-by: Oliver Zheng <mailinglists+netdev@oliverzheng.com>
Signed-off-by: John Dykstra <john.dykstra1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ucc_geth: Fix oops when using fixed-link support

commit b1c4a9dddf09fe99b8f88252718ac5b357363dc4 ("ucc_geth: Change
uec phy id to the same format as gianfar's") introduced a regression
in the ucc_geth driver that causes this oops when fixed-link is used:

Unable to handle kernel paging request for data at address 0x00000000
Faulting instruction address: 0xc0151270
Oops: Kernel access of bad area, sig: 11 [#1]
TMCUTU
NIP: c0151270 LR: c0151270 CTR: c0017760
REGS: cf81fa60 TRAP: 0300   Not tainted  (2.6.29-rc8)
MSR: 00009032 <EE,ME,IR,DR>  CR: 24024042  XER: 20000000
DAR: 00000000, DSISR: 20000000
TASK = cf81cba0[1] 'swapper' THREAD: cf81e000
GPR00: c0151270 cf81fb10 cf81cba0 00000000 c0272e20 c025f354 00001e80
cf86b08c
GPR08: d1068200 cffffb74 06000000 d106c200 42024042 10085148 0fffd000
0ffc81a0
GPR16: 00000001 00000001 00000000 007ffeb0 00000000 0000c000 cf83f36c
cf83f000
GPR24: 00000030 cf83f360 cf81fb20 00000000 d106c200 20000000 00001e80
cf83f360
NIP [c0151270] ucc_geth_open+0x330/0x1efc
LR [c0151270] ucc_geth_open+0x330/0x1efc
Call Trace:
[cf81fb10] [c0151270] ucc_geth_open+0x330/0x1efc (unreliable)
[cf81fba0] [c0187638] dev_open+0xbc/0x12c
[cf81fbc0] [c0187e38] dev_change_flags+0x8c/0x1b0

This patch fixes the issue by removing offending (and somewhat
duplicate) code from init_phy() routine, and changes _probe()
function to use uec_mdio_bus_name().

Also, since we fully construct phy_bus_id in the _probe() routine,
we no longer need ->phy_address and ->mdio_bus fields in
ucc_geth_info structure.

I wish the patch would be a bit shorter, but it seems like the only
way to fix the issue in a sane way. Luckily, the patch has been
tested with real PHYs and fixed-link, so no further regressions
expected.

Reported-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Tested-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Signed-off-by: David S. Miller <davem@davemloft.net>

dm9000: locking bugfix

This fixes a locking bug in the dm9000 driver.  It calls
request_irq() without setting IRQF_DISABLED ... which is
correct for handlers that support IRQ sharing, since that
behavior is not guaranteed for shared IRQs.  However, its
IRQ handler then wrongly assumes that IRQs are blocked.
So the fix just uses the right spinlock primitives in the
IRQ handler.

NOTE:  this is a classic example of the type of bug which
lockdep currently masks by forcibly setting IRQF_DISABLED
on IRQ handlers that did not request that flag.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x: Adding licensing to bnx2x_init_values.h

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

net: update dnet.c for bus_id removal

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

netdev: expose net_device_ops compat as config option

Now that most network device drivers in (all but one in x86_64 allmodconfig)
support net_device_ops. Expose it as a configuration parameter. Still
need to address even older 32 bit drivers, and other arch before
compatiablity can be scheduled for removal in some future release.

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

hostap: convert to net_device_ops

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

hostap: convert to internal net_device_stats

Use pre-existing net_device_stats in network_device struct.

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

ipw2200: convert to net_device_ops

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

ipw2100: convert to net_device_ops

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

ipw2x00: convert infrastructure for use by net_device_ops

Expose routines so drivers can hook. Only set ptrs in netdev
if using old compat code.

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

ipw2x00: convert to internal net_device_stats

Replace struct in ieee with current net_device_stats, so no longer
need get_stats hook

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

libertas: convert to net_device_ops

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

libertas: convert to internal net_device_stats

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

prism54: convert to internal net_device_stats

Also, make ethtool_ops const as it should be, and get rid
of useless cast.

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

prism54: convert to net_device_ops

Also, make ethtool_ops const as it should be, and get rid
of useless cast.

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

mac80211_hwsim: convert to net_device_ops

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

zd1201: convert to net_device_ops

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

zd1201: convert to internal net_device_stats

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

wl3501: convert to net_device_ops

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

wl3501: convert to internal net_device_stats

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

raylan: convert to net_device_ops

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

atmel: convert to net_device_ops

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

airo: convert to net_device_ops

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

wavelan: convert to net_device_ops

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

wavelan: convert to internal net_device_stats

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

strip: convert to net_device_ops

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

netwave: convert to net_device_ops

Also get rid of unneeded cast

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

netwave: convert to internal net_device_stats

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

sbni: convert to net_device_ops

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

sbni: use internal net_device_stats

Convert to use existing net_device_stats.
This driver,
      has bad style,
       of using commas,
   when brackets should be used...

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