Noticed by Stephen Rothwell:

	Today's linux-next build (x86_64 allmodconfig gcc-4.4.0)
	produced this warning:

	drivers/net/wan/dscc4.c: In function 'dscc4_rx_skb':
	drivers/net/wan/dscc4.c:670: warning: suggest parentheses around comparison in operand of '|'

	which actually points out a bug, I think.  It is doing
		(x & (y | z)) != y | z
	when it probably means
		(x & (y | z)) != (y | z)

	Introduced by commit 5de3fcab


	("WAN: bit and/or confusion").

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

Here is a patch which fixes an issue observed when using TCP over IPv6
and AH from IPsec.

When a connection gets closed the 4-way method and the last ACK from
the server gets dropped, the subsequent FINs from the client do not
get ACKed because tcp_v6_send_response does not set the transport
header pointer. This causes ah6_output to try to allocate a lot of
memory, which typically fails, so the ACKs never make it out of the
stack.

I have reproduced the problem on kernel 2.6.7, but after looking at
the latest kernel it seems the problem is still there.

Signed-off-by: Cosmin Ratiu <cratiu@ixiacom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

To unclutter probe() a little bit, put all device initialization code
in one spot and device deinit code in another spot.  Also remove unused
rq->buf_index variable/func.

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Nic firmware can return zero for port MTU, so check for non-zero value
before checking for change in port MTU.

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Deprecate some old APIa; change arguments to stats dump all API; add new
interrupt assert API

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Bug fix: enable VLAN filtering

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Provision for multiple Rx/Tx queues.  Max of 8 WQs and 8 RQs.  Max for
completion queue is 8+8=16 and max for interrupt resources is 8+8+2.

Add driver/firmware interface for setting up RSS secret key and indirection
table.

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Bug fix: included MAC drops in rx_dropped netstat.  Also track Rx trunctations
stat at the MAC

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Some driver -> nic firmware calls weren't guarded with a spinlock, exposing
the call i/f to a race between two threads

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Use netdev_alloc_skb rather than dev_alloc_skb

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

enic WQ desc supports a maximum 16K buf size, so split any send fragments
larger than 16K into several descs.

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

A0 revision ASIC has an erratum on the RQ desc cache on chip where the
cache can become corrupted causing pkt buf writes to wrong locations.  The s/w
workaround is to post a dummy RQ desc in the ring every 32 descs, causing a
flush of the cache.  A0 parts are not production, but there are enough of
these parts in the wild in test setups to warrant including workaround.  A1
revision ASIC parts fix erratum.

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Nic firmware can place resources (queues, intrs, etc) on multiple BARs, so
allow driver to discover/map resources beyond BAR0.

Signed-off-by: Scott Feldman <scofeldm@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Its hard to tell if vlans are dropping frames, since
every frame given to vlan_???_start_xmit() functions
is accounted as fully transmitted by lower device.

We can test dev_queue_xmit() return values to
properly account for dropped frames.

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

macvlan devices are currently not multi-queue capable.

We can do that defining rtnl_link_ops method,
get_tx_queues(), called from rtnl_create_link()

This new method gets num_tx_queues/real_num_tx_queues
from lower device.

macvlan_get_tx_queues() is a copy of vlan_get_tx_queues().

Because macvlan_start_xmit() has to update netdev_queue
stats only (and not dev->stats), I chose to change
tx_errors/tx_aborted_errors accounting to tx_dropped,
since netdev_queue structure doesnt define tx_errors /
tx_aborted_errors.

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

A few drivers still access the arguments to MDIO ioctls as an array of
u16.  Convert them to use struct mii_ioctl_data.

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

dev_ioctl() already checks capable(CAP_NET_ADMIN) before calling the
driver's implementation of MDIO ioctls.

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

The standard MDIO ioctl numbers are well-established and these should
no longer be needed.

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

While perusing vendor driver, I saw that it did not enable the Vaux
power unless device was able to wake from lan for D3cold.
This might help for Rene's power issue.

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

Fix a perpetual while() loop in unwinding partial
mapped tx skb on dma mapping failure.

Reported-by: "Juha Leppanen" <juha_motorsportcom@luukku.com>
Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove duplicate calls to netxen_napi_add().

Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Alloc 12k skbuffs so that firmware can aggregate more
packets into one buffer. This doesn't raise memory
consumption since 9k skbs use 16k slab cache anyway.

Signed-off-by: Dhananjay Phadke <dhananjay@netxen.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The FCoE DDP in 82599 can be used for both FCoE initiator as well as FCoE
target, depending on the indication of the exchange being the responder or
originator in the F_CTL (frame control) field in the encapsulated Fiber
Channel frame header (T10 Spec., FC-FS). For the initiator, OX_ID is used
for FCoE DDP, where for the target RX_ID is used for FCoE DDP.

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This adds a simple selection of a FCoE tx queue based on the current cpu id to
distribute transmission of FCoE traffic evenly among multiple FCoE transmit
queues.

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds support for multiple transmit queues to the Fiber Channel
over Ethernet (FCoE) feature found in 82599. Currently, FCoE has multiple
Rx queues available, along with a redirection table, that helps distribute
the I/O load across multiple CPUs based on the FC exchange ID. To make
this the most effective, we need to provide the same layout of transmit
queues to match receive.

Particularly, when Data Center Bridging (DCB) is enabled, the designated
traffic class for FCoE can have dedicated queues for just FCoE traffic,
while not affecting any other type of traffic flow.

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch updates things so that vlan tags are taken into account when
setting the receive large packet maximum length.  This allows the VF driver
to correctly receive full sized frames when vlans are enabled.

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>

The igb_irq_disable/enable calls were causing virtual functions associated
with the igb physical function to have their interrupts disabled.  In order
to prevent this from occuring we should only clear/set the bits related to
the physical function.

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>

This patch adds support for the set_rx_mode netdevice operation so that igb
can better support multiple unicast addresses.

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>

Remove a debugging aid I accidently left in previous 'cleanup' patch

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

vlan_dev_hard_start_xmit() & vlan_dev_hwaccel_hard_start_xmit()
select txqueue number 0, instead of using index provided by
skb_get_queue_mapping().

This is not correct after commit 2e59af3d


[vlan: multiqueue vlan device] because
txq->tx_packets  & txq->tx_bytes changes are performed on
a single location, and not the right locking.

Fix is to take the appropriate struct netdev_queue pointer

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

Pure style cleanup patch before surgery :)

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

atm/br2684: netif_stop_queue() when atm device busy and netif_wake_queue() when we can send packets again.

This patch removes the call to dev_kfree_skb() when the atm device is busy.
Calling dev_kfree_skb() causes heavy packet loss then the device is under
heavy load, the more correct behavior should be to stop the upper layers,
then when the lower device can queue packets again wake the upper layers.

Signed-off-by: Karl Hiramoto <karl@hiramoto.org>
Signed-off-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

fec_enet_mii, fec_enet_rx and fec_enet_tx are both only called by
fec_enet_interrupt in interrupt context.  So they must not use
spin_lock_irq/spin_unlock_irq.

This fixes:
	WARNING: at kernel/lockdep.c:2140 trace_hardirqs_on_caller+0x130/0x194()
	...

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Greg Ungerer <gerg@uclinux.org>
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Matt Waddel <Matt.Waddel@freescale.com>
Cc: netdev@vger.kernel.org
Cc: Tim Sander <tim01@vlsi.informatik.tu-darmstadt.de>
Acked-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

mii_discover_phy is only called by fec_enet_mii (via mip->mii_func).  So
&fep->mii_lock is already held and mii_discover_phy must not call
mii_queue which locks &fep->mii_lock, too.

This was noticed by lockdep:

	=============================================
	[ INFO: possible recursive locking detected ]
	2.6.31-rc8-00038-g37d0892 #109
	---------------------------------------------
	swapper/1 is trying to acquire lock:
	 (&fep->mii_lock){-.....}, at: [<c01569f8>] mii_queue+0x2c/0xcc

	but task is already holding lock:
	 (&fep->mii_lock){-.....}, at: [<c0156328>] fec_enet_interrupt+0x78/0x460

	other info that might help us debug this:
	2 locks held by swapper/1:
	 #0:  (rtnl_mutex){+.+.+.}, at: [<c0183534>] rtnl_lock+0x18/0x20
	 #1:  (&fep->mii_lock){-.....}, at: [<c0156328>] fec_enet_interrupt+0x78/0x460

	stack backtrace:
	Backtrace:
	[<c00226fc>] (dump_backtrace+0x0/0x108) from [<c01eac14>] (dump_stack+0x18/0x1c)
	 r6:c781d118 r5:c03e41d8 r4:00000001
	[<c01eabfc>] (dump_stack+0x0/0x1c) from [<c005bae4>] (__lock_acquire+0x1a20/0x1a88)
	[<c005a0c4>] (__lock_acquire+0x0/0x1a88) from [<c005bbac>] (lock_acquire+0x60/0x74)
	[<c005bb4c>] (lock_acquire+0x0/0x74) from [<c01edda8>] (_spin_lock_irqsave+0x54/0x68)
	 r7:60000093 r6:c01569f8 r5:c785e468 r4:00000000
	[<c01edd54>] (_spin_lock_irqsave+0x0/0x68) from [<c01569f8>] (mii_queue+0x2c/0xcc)
	 r7:c785e468 r6:c0156b24 r5:600a0000 r4:c785e000
	[<c01569cc>] (mii_queue+0x0/0xcc) from [<c0156b78>] (mii_discover_phy+0x54/0xa8)
	 r8:00000002 r7:00000032 r6:c785e000 r5:c785e360 r4:c785e000
	[<c0156b24>] (mii_discover_phy+0x0/0xa8) from [<c0156354>] (fec_enet_interrupt+0xa4/0x460)
	 r5:c785e360 r4:c077a170
	[<c01562b0>] (fec_enet_interrupt+0x0/0x460) from [<c0066674>] (handle_IRQ_event+0x48/0x120)
	[<c006662c>] (handle_IRQ_event+0x0/0x120) from [<c0068438>] (handle_level_irq+0x94/0x11c)
	...

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Greg Ungerer <gerg@uclinux.org>
Cc: Ben Hutchings <ben@decadent.org.uk>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Matt Waddel <Matt.Waddel@freescale.com>
Cc: netdev@vger.kernel.org
Cc: Tim Sander <tim01@vlsi.informatik.tu-darmstadt.de>
Acked-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The bpq ether driver is modifying the data art of the skb by first
dropping the KISS byte (a command byte for the radio) then prepending the
length + 4 of the remaining AX.25 packet to be transmitted as a little
endian 16-bit number.  If the high byte of the length has a different
value than the dropped KISS byte users of clones of the skb may observe
this as corruption.  This was observed with by running listen(8) -a which
uses a packet socket which clones transmit packets.  The corruption will
then typically be displayed for as a KISS "TX Delay" command for AX.25
packets in the range of 252..508 bytes or any other KISS command for
yet larger packets.

Fixed by using skb_cow to create a private copy should the skb be cloned.
Using skb_cow also allows us to cleanup the old logic to ensure sufficient
headroom in the skb.

While at it, replace a return of 0 from bpq_xmit with the proper constant
NETDEV_TX_OK which is now being used everywhere else in this function.

Affected: all 2.2, 2.4 and 2.6 kernels.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Reported-by: Jann Traschewski <jann@gmx.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

Request_irq() may fail in different ways, handle accordingly.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This fixed a lockdep warning which appeared when doing stress
memory tests over NFS:

	inconsistent {RECLAIM_FS-ON-W} -> {IN-RECLAIM_FS-W} usage.

	page reclaim => nfs_writepage => tcp_sendmsg => lock sk_lock

	mount_root => nfs_root_data => tcp_close => lock sk_lock =>
			tcp_send_fin => alloc_skb_fclone => page reclaim

David raised a concern that if the allocation fails in tcp_send_fin(), and it's
GFP_ATOMIC, we are going to yield() (which sleeps) and loop endlessly waiting
for the allocation to succeed.

But fact is, the original GFP_KERNEL also sleeps. GFP_ATOMIC+yield() looks
weird, but it is no worse the implicit sleep inside GFP_KERNEL. Both could
loop endlessly under memory pressure.

CC: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
CC: David S. Miller <davem@davemloft.net>
CC: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds support to flash a firmware image to a device using ethtool.
The driver gets the filename of the firmware image and flashes the image
using the request firmware path.

The region "on the chip" to be flashed can be specified by an option.
It is upto the device driver to enumerate the region number passed by ethtool,
to the region to be flashed.

The default behavior is to flash all the regions on the chip.

Signed-off-by: Ajit Khaparde <ajitk@serverengines.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The generic packet receive code takes care of setting
netdev->last_rx when necessary, for the sake of the
bonding ARP monitor.

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

Christoph Lameter pointed out that packet drops at qdisc level where not
accounted in SNMP counters. Only if application sets IP_RECVERR, drops
are reported to user (-ENOBUFS errors) and SNMP counters updated.

IP_RECVERR is used to enable extended reliable error message passing,
but these are not needed to update system wide SNMP stats.

This patch changes things a bit to allow SNMP counters to be updated,
regardless of IP_RECVERR being set or not on the socket.

Example after an UDP tx flood
# netstat -s 
...
IP:
    1487048 outgoing packets dropped
...
Udp:
...
    SndbufErrors: 1487048


send() syscalls, do however still return an OK status, to not
break applications.

Note : send() manual page explicitly says for -ENOBUFS error :

 "The output queue for a network interface was full.
  This generally indicates that the interface has stopped sending,
  but may be caused by transient congestion.
  (Normally, this does not occur in Linux. Packets are just silently
  dropped when a device queue overflows.) "

This is not true for IP_RECVERR enabled sockets : a send() syscall
that hit a qdisc drop returns an ENOBUFS error.

Many thanks to Christoph, David, and last but not least, Alexey !

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