efx_mcdi_poll() uses get_seconds() to read the current time and to
implement a polling timeout.  The use of this function was chosen
partly because it could easily be replaced in a co-sim environment
with a macro that read the simulated time.

Unfortunately the real get_seconds() returns the system time (real
time) which is subject to adjustment by e.g. ntpd.  If the system time
is adjusted forward during a polled MCDI operation, the effective
timeout can be shorter than the intended 10 seconds, resulting in a
spurious failure.  It is also possible for a backward adjustment to
delay detection of a areal failure.

Use jiffies instead, and change MCDI_RPC_TIMEOUT to be denominated in
jiffies.  Also correct rounding of the timeout: check time > finish
(or rather time_after(time, finish)) and not time >= finish.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

The assertion of netif_device_present() at the top of
efx_hard_start_xmit() may fail if we don't do this.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

We sometimes hit a "failed to flush" timeout on some TX queues, but the
flushes have completed and the flush completion events seem to go missing.
In this case, we can check the TX_DESC_PTR_TBL register and drain the
queues if the flushes had finished.

[bwh: Minor fixes to coding style]
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

If the MC reboots then the stats it reports to us will have been
reset.  We need to reset ours to get efx_update_diff_stat() working
properly.

(Ideally we would maintain stats across the reboot, but as this should
only happen immediately after a firmware upgrade it's not really worth
the trouble.)
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

Currently we initialise the newly allocated buffer to all-1s, which is
important for event queues but not for descriptor queues.  And since
we also do that in efx_nic_init_eventq(), it is completely pointless
to do it here.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

efx_writed_table() uses a step of 16 bytes but efx_readd_table() uses
a step of 4 bytes.  Why are they different?

Firstly, register access is asymmetric:

- The EVQ_RPTR table and RX_INDIRECTION_TBL can (or must?) be written
  as dwords even though they have a step size of 16 bytes, unlike
  most other CSRs.
- In general, a read of any width is valid for registers, so long as
  it does not cross register boundaries.  There is also no latching
  behaviour in the BIU, contrary to rumour.

We write to the EVQ_RPTR table with efx_writed_table() but never read
it back as it's write-only.  We write to the RX_INDIRECTION_TBL with
efx_writed_table(), but only read it back for the register dump, where
we use efx_reado_table() as for any other table with step size of 16.

We read MC_TREG_SMEM with efx_readd_table() for the register dump, but
normally read and write it with efx_readd() and efx_writed() using
offsets calculated in bytes.

Since these functions are trivial and have few callers, it's clearer
to open-code them at the call sites.  While we're at it, update the
comments on the BIU behaviour again.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

efx_mcdi_rpc_start() returns a negative value on error or zero on
success.  However one caller that can't properly handle failure then
does WARN_ON(rc > 0).  Change it to WARN_ON(rc < 0).
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

Receiving pause frames can block TX queue flushes.  Earlier changes
work around this by reconfiguring the MAC during flushes for VFs, but
during flushes for the PF we would only change the fc_disable counter.
Unless the MAC is reconfigured for some other reason during the flush
(which I would not expect to happen) this had no effect at all.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

Add necessary cast when setting a bogus checksum.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

sparse has got a bit more picky since I last ran it over this.  Add
forced casts for use of ~0 as a big-endian value.  Undo the pointless
optimisation of parameter validation with '|'; using '||' avoids these
warnings.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>

This patch changes three methods to be static and removes their
EXPORT_SYMBOLs in core/dev.c and their external declaration in
netdevice.h. The methods, dev_gro_receive(), napi_frags_finish() and
napi_skb_finish(), which are in the GRO rx path, are not used
outside core/dev.c.
Signed-off-by: Rami Rosen <ramirose@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Make the description of how tcp_ecn works a bit more explicit and clear.
Signed-off-by: Rick Jones <rick.jones2@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

Pull networking fixes from David Miller:
 "Some more fixes trickled in over the past few days:

   1) PIM device names can overflow the IFNAMSIZ buffer unless we
      properly limit the allowed indexes, fix from Eric Dumazet.

   2) Under heavy load we can OOPS in icmp reply processing due to an
      unchecked inet_putpeer() call.  Fix from Neal Cardwell.

   3) SCTP round trip calculations need to use 64-bit math to avoid
      overflows, fix from Schoch Christian.

   4) Fix a memory leak and an error return flub in SCTP and IRDA
      triggerable by userspace.  Fix from Tommi Rantala and found by the
      syscall fuzzer (trinity).

   5) MLX4 driver gives bogus size to memcpy() call, fix from Amir
      Vadai.

   6) Fix length calculation in VHOST descriptor translation, from
      Michael S Tsirkin.

   7) Ambassador ATM driver loops forever while loading firmware, fix
      from Dan Carpenter.

   8) Over MTU packets in openvswitch warn about wrong device, fix from
      Jesse Gross.

   9) Netfilter IPSET's netlink code can overrun a string buffer because
      it's not properly limited to IFNAMSIZ.  Fix from Florian Westphal.

  10) PCAN USB driver sets wrong timestamp in SKB, from Oliver Hartkopp.

  11) Make sure the RX ifindex always has a valid value in the CAN BCM
      driver, even if we haven't received a frame yet.  Fix also from
      Oliver Hartkopp."

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net:
  team: fix hw_features setup
  atm: forever loop loading ambassador firmware
  vhost: fix length for cross region descriptor
  irda: irttp: fix memory leak in irttp_open_tsap() error path
  net: qmi_wwan: add Huawei E173
  net/mlx4_en: Can set maxrate only for TC0
  sctp: Error in calculation of RTTvar
  sctp: fix -ENOMEM result with invalid user space pointer in sendto() syscall
  sctp: fix memory leak in sctp_datamsg_from_user() when copy from user space fails
  net: ipmr: limit MRT_TABLE identifiers
  ipv4: avoid passing NULL to inet_putpeer() in icmpv4_xrlim_allow()
  can: bcm: initialize ifindex for timeouts without previous frame reception
  can: peak_usb: fix hwtstamp assignment
  netfilter: ipset: fix netiface set name overflow
  openvswitch: Store flow key len if ARP opcode is not request or reply.
  openvswitch: Print device when warning about over MTU packets.

Two small openswitch fixes from Jesse Gross.
Signed-off-by: David S. Miller <davem@davemloft.net>

CC      drivers/net/ethernet/ti/cpsw.o
drivers/net/ethernet/ti/cpsw.c: In function 'cpsw_ndo_ioctl':
drivers/net/ethernet/ti/cpsw.c:881:20: warning: unused variable 'priv'

The build warning is generated when CPTS is not selected in Kernel Build.
Fixing by passing the net_device pointer to cpts IOCTL instead of passing priv
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

John W. Linville says:

====================
This pull request is intended for the 3.8 stream.  It is a bit large
-- I guess Thanksgiving got me off track!  At least the code got to
spend some time in linux-next... :-)

This includes the usual batch of pulls for Bluetooth, NFC, and mac80211
as well as iwlwifi.  Also here is an ath6kl pull, and a new driver
in the rtlwifi family.  The brcmfmac, brcmsmac, ath9k, and mwl8k get
their usual levels of attention, and a handful of other updates tag
along as well.

For more detail on the pulls, please see below...

On Bluetooth, Gustavo says:

"Another set of patches for integration in wireless-next. There are two big set
of changes in it: Andrei Emeltchenko and Mat Martineau added more patches
towards a full Bluetooth High Speed support and Johan Hedberg improve the
single mode support for Bluetooth dongles. Apart from that we have small fixes
and improvements."

...and:

"A few patches to 3.8. The majority of the work here is from Andrei on the High
Speed support. Other than that Johan added support for setting LE advertising
data. The rest are fixes and clean ups and small improvements like support for
a new broadcom hardware."

On mac80211, Johannes says:

"This is for mac80211, for -next (3.8). Plenty of changes, as you can see
below. Some fixes for previous changes like the export.h include, the
beacon listener fix from Ben Greear, etc. Overall, no exciting new
features, though hwsim does gain channel context support for people to
try it out and look at."

...and...:

"This one contains the mac80211-next material. Apart from a few small new
features and cleanups I have two fixes for the channel context code. The
RX_END timestamp support will probably be reworked again as Simon Barber
noted the calculations weren't really valid, but the discussions there
are still going on and it's better than what we had before."

...and:

"Please pull (see below) to get the following changes:
 * a fix & a debug aid in IBSS from Antonio,
 * mesh cleanups from Marco,
 * a few bugfixes for some of my previous patches from Arend and myself,
 * and the big initial VHT support patchset"

And on iwlwifi, Johannes says:

"In addition to the previous four patches that I'm not resending,
we have a number of cleanups, message reduction, firmware error
handling improvements (yes yes... we need to fix them instead)
and various other small things all over."

...and:

"In his quest to try to understand the current iwlwifi problems (like
stuck queues etc.) Emmanuel has first cleaned up the PCIe code, I'm
including his changes in this pull request. Other than that I only have
a small cleanup from Sachin Kamat to remove a duplicate include and a
bugfix to turn off MFP if software crypto is enabled, but this isn't
really interesting as MFP isn't supported right now anyway."

On NFC, Samuel says:

"With this one we have:

- A few HCI improvements in preparation for an upcoming HCI chipset support.
- A pn544 code cleanup after the old driver was removed.
- An LLCP improvement for notifying user space when one peer stops ACKing I
  frames."

On ath6kl, Kalle says:

"Major changes this time are firmware recover support to gracefully
handle if firmware crashes, support for changing regulatory domain and
support for new ar6004 hardware revision 1.4. Otherwise there are just
smaller fixes or cleanups from different people."

Thats about it... :-)  Please let me know if there are problems!
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit ff33c0e1 ('net: Remove bogus
dependencies on INET') wrongly removed this dependency.  cxgb3 uses
the arp_send() function defined in net/ipv4/arp.c.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Do this in the same way bonding does. This fixed setup resolves performance
issues when using some cards with certain offloading.
Signed-off-by: Jiri Pirko <jiri@resnulli.us>
Signed-off-by: David S. Miller <davem@davemloft.net>

There was a forever loop introduced here when we converted this to
request_firmware() back in 2008.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Chas Williams <chas@cmf.nrl.navy.mil>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch reports the change made by Stephen Hemminger in ipip and gre[6] in
commit eccc1bb8 (tunnel: drop packet if ECN present with not-ECT).

Goal is to handle RFC6040, Section 4.2:

Default Tunnel Egress Behaviour.
 o If the inner ECN field is Not-ECT, the decapsulator MUST NOT
      propagate any other ECN codepoint onwards.  This is because the
      inner Not-ECT marking is set by transports that rely on dropped
      packets as an indication of congestion and would not understand or
      respond to any other ECN codepoint [RFC4774].  Specifically:

      *  If the inner ECN field is Not-ECT and the outer ECN field is
         CE, the decapsulator MUST drop the packet.

      *  If the inner ECN field is Not-ECT and the outer ECN field is
         Not-ECT, ECT(0), or ECT(1), the decapsulator MUST forward the
         outgoing packet with the ECN field cleared to Not-ECT.

The patch takes benefits from common function added in net/inet_ecn.h.

Like it was done for Xin4 tunnels, it adds logging to allow detecting broken
systems that set ECN bits incorrectly when tunneling (or an intermediate
router might be changing the header). Errors are also tracked via
rx_frame_error.

CC: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

On BE2 chip, an interrupt may be raised even when EQ is in un-armed state.
As a result be_intx()::events_get() and be_poll:events_get() can race and
notify an EQ wrongly.

Fix this by counting events only in be_poll(). Commit 0b545a62 fixes
the same issue in the MSI-x path.

But, on Lancer, INTx can be de-asserted only by notifying num evts. This
is not an issue as the above BE2 behavior doesn't exist/has never been
seen on Lancer.
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove an outdated comment, that should have been removed in the
patch named "MODULE_PARM conversions" from early 2005.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>

Building ewrk3.o triggers this GCC warning:
    drivers/net/ethernet/dec/ewrk3.c: In function '__check_irq':
    drivers/net/ethernet/dec/ewrk3.c:1915:1: warning: return from incompatible pointer type [enabled by default]

This can be trivially fixed by changing the 'irq' parameter from int to
byte (which is an alias for unsigned char for module parameters).
Signed-off-by: Paul Bolle <pebolle@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>

Marc Kleine-Budde says:

====================
this is pull request is for net-next. Contains a patch by Andreas
Larsson, which enables the sja1000 of driver to work under sparc.
AnilKumar Ch contributed a patch to improve the c_can support under
omap, Olivier Sobrie's patch brings support for the CAN/USB dongles
from Kvaser. In a bunch of patches by me missing MODULE_ALIAS and/or
MODULE_DEVICE_TABLE entries were added to the CAN drivers.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

An interface name overflow fix in netfilter via Pablo Neira Ayuso.
Signed-off-by: David S. Miller <davem@davemloft.net>

If all slaves of a balance-rr bond with ARP monitor are enslaved
with down link state, bond keeps down state even after slaves
go up.

This is caused by bond_enslave() setting curr_active_slave to
first slave not taking into account its link state. As
bond_loadbalance_arp_mon() uses curr_active_slave to identify
whether slave's down->up transition should update bond's link
state, bond stays down even if slaves are up (until first slave
goes from up to down at least once).

Before commit f31c7937 "bonding: start slaves with link down for
ARP monitor", this was masked by slaves always starting in UP
state with ARP monitor (and MII monitor not relying on
curr_active_slave being NULL if there is no slave up).
Signed-off-by: Michal Kubecek <mkubecek@suse.cz>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

If a single descriptor crosses a region, the
second chunk length should be decremented
by size translated so far, instead it includes
the full descriptor length.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Cleanup the memory we allocated earlier in irttp_open_tsap() when we hit
this error path. The leak goes back to at least 1da177e4
("Linux-2.6.12-rc2").

Discovered with Trinity (the syscall fuzzer).
Signed-off-by: Tommi Rantala <tt.rantala@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The Huawei E173 is a QMI/wwan device which normally appear
as 12d1:1436 in Linux. The descriptors displayed in that
mode will be picked up by cdc_ether.  But the modem has
another mode with a different device ID and a slightly
different set of descriptors. This is the mode used by
Windows like this:

3Modem:      USB\VID_12D1&PID_140C&MI_00\6&3A1D2012&0&0000
Networkcard: USB\VID_12D1&PID_140C&MI_01\6&3A1D2012&0&0001
Appli.Inter: USB\VID_12D1&PID_140C&MI_02\6&3A1D2012&0&0002
PC UI Inter: USB\VID_12D1&PID_140C&MI_03\6&3A1D2012&0&0003
Reported-by: Thomas Schäfer <tschaefer@t-online.de>
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch turns QFQ into QFQ+, a variant of QFQ that provides the
following two benefits: 1) QFQ+ is faster than QFQ, 2) differently
from QFQ, QFQ+ correctly schedules also non-leaves classes in a
hierarchical setting. A detailed description of QFQ+, plus a
performance comparison with DRR and QFQ, can be found in [1].

[1] P. Valente, "Reducing the Execution Time of Fair-Queueing Schedulers"
http://algo.ing.unimo.it/people/paolo/agg-sched/agg-sched.pdfSigned-off-by: Paolo Valente <paolo.valente@unimore.it>
Signed-off-by: David S. Miller <davem@davemloft.net>

As pointed out by Bjorn Mork, the generic "usb" driver sets this
for us so no need to directly set it in this driver.
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
Cc: Bjorn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds support for USB dynamic autosuspend to the
smsc75xx driver.  This saves virtually no power in the USB
device but enables power savings in upstream hosts and
the host CPU.

Note currently Linux doesn't automatically enable this
functionality by default for devices so to test this:

 echo auto > /sys/bus/usb/devices/2-1.2/power/control

where 2-1.2 is the USB bus address of the LAN7500.
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch ensures that if we fail to suspend the LAN7500 device
we call usbnet_resume before returning failure, instead of
leaving the usbnet driver in an unusable state.
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch enables LAN7500 family devices to wake from suspend
on either link up or link down events.

It also adds _nopm versions of mdio access functions, so we can
safely call them from suspend and resume functions

Updated patch to add newlines to printk messages
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch splits out the logic for entering suspend modes
to separate functions, to reduce the complexity of the
smsc75xx_suspend function.
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds a missing check and error message if smsc75xx_reset
fails.
Signed-off-by: Steve Glendinning <steve.glendinning@shawell.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

Had a typo in memcpy.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>