A cpu executing the network receive path sheds packets when its input
queue grows to netdev_max_backlog. A single high rate flow (such as a
spoofed source DoS) can exceed a single cpu processing rate and will
degrade throughput of other flows hashed onto the same cpu.

This patch adds a more fine grained hashtable. If the netdev backlog
is above a threshold, IRQ cpus track the ratio of total traffic of
each flow (using 4096 buckets, configurable). The ratio is measured
by counting the number of packets per flow over the last 256 packets
from the source cpu. Any flow that occupies a large fraction of this
(set at 50%) will see packet drop while above the threshold.

Tested:
Setup is a muli-threaded UDP echo server with network rx IRQ on cpu0,
kernel receive (RPS) on cpu0 and application threads on cpus 2--7
each handling 20k req/s. Throughput halves when hit with a 400 kpps
antagonist storm. With this patch applied, antagonist overload is
dropped and the server processes its complete load.

The patch is effective when kernel receive processing is the
bottleneck. The above RPS scenario is a extreme, but the same is
reached with RFS and sufficient kernel processing (iptables, packet
socket tap, ..).
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Since commit ab78029e (drivers/pinctrl: grab default handles from device core)
we can rely on device core for handling pinctrl, so remove
devm_pinctrl_get_select_default() from the driver.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

We should correctly free related resources (grant ref, memory page, evtchn)
when setup_netfront fails.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add support for the VIA Velocity network driver to be bound to a
OF created platform device.
Signed-off-by: Tony Prisk <linux@prisktech.co.nz>
Signed-off-by: David S. Miller <davem@davemloft.net>

Remove the pci_* dma functions and replace with the more generic
versions.

In preparation of adding platform support, a new struct device *dev
is added to struct velocity_info which can be used by both the pci
and platform code.
Signed-off-by: Tony Prisk <linux@prisktech.co.nz>
Signed-off-by: David S. Miller <davem@davemloft.net>

Improve the clarity of the code in preparation for converting the
dma functions to generic versions, which require a struct device *.

This makes it possible to store a 'struct device *dev' in the
velocity_info structure.
Signed-off-by: Tony Prisk <linux@prisktech.co.nz>
Signed-off-by: David S. Miller <davem@davemloft.net>

It's suboptimal to invoke quite complex VORTEX_PCI() macro every time we want
to get a 'struct pci_dev *' when we already have it in a variable...
Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

These 2 places are the only matches for is_eisa in the whole tree.
Signed-off-by: Rolf Eike Beer <eike-kernel@sf-tec.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2x_set_features() forces a driver reload if GRO setting is changed.

A reload makes the ethernet port unresponsive for about 5 seconds.

This is not needed in the common case LRO is enabled, as LRO
(TPA_ENABLE_FLAG) has precedence over GRO (GRO_ENABLE_FLAG)

Tested:
 Verified that "ethtool -K eth0 gro {on|off}" doesn't blackout
 the NIC anymore

Google-Bug-Id: 8440442
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Dmitry Kravkov <dmitry@broadcom.com>
Acked-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Nithin Nayak Sujir says:

====================
This series adds support for modifying EEE settings via ethtool. Since this can
impact Link Flap Avoidance, the driver pulls the current hardware settings if
LFA is enabled. This is similar to how we do the link settings to avoid a flap.

v2: Fixes pointed out by Ben Hutchings.
 - Use MDIO_AN_EEE_LPABLE to set the lp_advertised field.
 - Check that tx_lpi_timer is within valid range.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Reviewed-by: Ben Li <benli@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

eee_config_ok() was checking only for mismatch in advertised settings.
This patch expands the scope of eee_config_ok() to check for mismatch in
the other eee settings. On mismatch we will require a call to
tg3_setup_eee() to push the configured settings to the hardware.
Reviewed-by: Ben Li <benli@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add tg3_eee_pull_config() to pull the settings from the hardware and
populate the eee structure.

If Link Flap Avoidance is enabled, we pull the eee settings from the hw
so as not to cause a phy reset on eee config mismatch later. This
requires moving down tg3_setup_eee() below the tg3_pull_config() to not
trample existing settings.
Reviewed-by: Ben Li <benli@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Add an eee structure and update it with eee settings. This will be used
for set/get_eee operations. Add common function tg3_setup_eee() that
will be used in the subsequent patches.
Reviewed-by: Ben Li <benli@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Do not leak starting address of BPF JIT code for non root users,
as it might help intruders to perform an attack.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Ben Hutchings <bhutchings@solarflare.com>
Cc: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

hpa bringed into my attention some security related issues
with BPF JIT on x86.

This patch makes sure the bpf generated code is marked read only,
as other kernel text sections.

It also splits the unused space (we vmalloc() and only use a fraction of
the page) in two parts, so that the generated bpf code not starts at a
known offset in the page, but a pseudo random one.

Refs:
http://mainisusuallyafunction.blogspot.com/2012/11/attacking-hardened-linux-systems-with.htmlReported-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Daniel Borkmann <dborkman@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp_timeout_skb() was intended to trigger fast recovery on timeout,
unfortunately in reality it often causes spurious retransmission
storms during fast recovery. The particular sign is a fast retransmit
over the highest sacked sequence (SND.FACK).

Currently the RTO timer re-arming (as in RFC6298) offers a nice cushion
to avoid spurious timeout: when SND.UNA advances the sender re-arms
RTO and extends the timeout by icsk_rto. The sender does not offset
the time elapsed since the packet at SND.UNA was sent.

But if the next (DUP)ACK arrives later than ~RTTVAR and triggers
tcp_fastretrans_alert(), then tcp_timeout_skb() will mark any packet
sent before the icsk_rto interval lost, including one that's above the
highest sacked sequence. Most likely a large part of scorebard will be
marked.

If most packets are not lost then the subsequent DUPACKs with new SACK
blocks will cause the sender to continue to retransmit packets beyond
SND.FACK spuriously. Even if only one packet is lost the sender may
falsely retransmit almost the entire window.

The situation becomes common in the world of bufferbloat: the RTT
continues to grow as the queue builds up but RTTVAR remains small and
close to the minimum 200ms. If a data packet is lost and the DUPACK
triggered by the next data packet is slightly delayed, then a spurious
retransmission storm forms.

As the original comment on tcp_timeout_skb() suggests: the usefulness
of this feature is questionable. It also wastes cycles walking the
sack scoreboard and is actually harmful because of false recovery.

It's time to remove this.
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Acked-by: Nandita Dukkipati <nanditad@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch removes mentioning the sysfsf net_device weight attribute
(class/net/<device>/weight)
in Documentation/sysctl/net.txt, since the net sysfs weight attribute
was removed by the following patch:

[NET]: Make NAPI polling independent of struct net_device objects
 bea3348eSigned-off-by: Rami Rosen <ramirose@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds the support of peer address for IPv6. For example, it is
possible to specify the remote end of a 6inY tunnel.
This was already possible in IPv4:
 ip addr add ip1 peer ip2 dev dev1

The peer address is specified with IFA_ADDRESS and the local address with
IFA_LOCAL (like explained in include/uapi/linux/if_addr.h).
Note that the API is not changed, because before this patch, it was not
possible to specify two different addresses in IFA_LOCAL and IFA_REMOTE.
There is a small change for the dump: if the peer is different from ::,
IFA_ADDRESS will contain the peer address instead of the local address.
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

module_free()/vfree() takes care of details, we no longer need a wrapper
and a work_struct.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This was added by commit 59b9997b (Revert "net: maintain namespace
isolation between vlan and real device").
In fact, before the initial commit - the one that is reverted -, this
statement was not present.
'skb->dev = dev' is already done in eth_type_trans(), which is call just
after.
Spotted-by: Alain Ritoux <alain.ritoux@6wind.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch enables user to unload netback module, which is useful when user
wants to upgrade to a newer netback module without rebooting the host.

Netfront cannot handle netback removal event. As we cannot fix all possible
frontends we add module get / put along with vif get / put to avoid
mis-unloading of netback. To unload netback module, user needs to shutdown all
VMs or migrate them to another host or unplug all vifs before hand.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>¬
Signed-off-by: David S. Miller <davem@davemloft.net>

The array mmap_pages is never touched in the initialization function. This is
remnant of mapping mechanism, which does not exist upstream. In current
upstream code this array only tracks usage of pages inside netback. Those
pages are allocated when contructing a SKB and passed directly to network
subsystem.
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

It looks like we can call module_free()/vfree() from softirq context,
so no longer need a wrapper and a work_struct.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

module_usb_driver() eliminates boilerplate and simplifies the code.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Cc: Realtek linux nic maintainers <nic_swsd@realtek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

version.h header inclusion is not necessary as detected by
checkversion.pl.
Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

The commit 823aa873
  Author: stephen hemminger <stephen@networkplumber.org>
  Date:   Sat Apr 27 11:31:57 2013 +0000

    vxlan: allow choosing destination port per vxlan

introduced per-vxlan UDP port configuration but only did half of the
necessary work.  It added per vxlan destination for sending, but
overlooked the handling of multiple ports for incoming traffic.

This patch changes the listening port management to handle multiple
incoming UDP ports. The earlier per-namespace structure is now a hash
list per namespace.

It is also now possible to define the same virtual network id
but with different UDP port values which can be useful for migration.
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Clean up the code a bit to initialize the variables directly when
defining them.
Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>

Clean up the code a bit to initialize the variables directly when
defining them.
Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>

Clean up the code a bit to initialize the variables directly when
defining them.
Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>

Clean up the code a bit to initialize the variables directly when
defining them.
Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: David S. Miller <davem@davemloft.net>

Marc Kleine-Budde says:

====================
this is a pull-request for net-next/master. It consists of 4 patches by
Jingoo Han, which remove the unnecessary platform_set_drvdata() and a
patch by Laurent Navet converting the grcan driver to use
devm_ioremap_resource().
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch removes unnecessary #if 0 code from 3c509.c
Signed-off-by: govindarajulu.v <govindarajulu90@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

devm_ioremap_resource does sanity checks on the given resource. No need to
duplicate this in the driver.
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp_fixup_rcvbuf() contains a loop to estimate initial socket
rcv space needed for a given mss. With large MTU (like 64K on lo),
we can loop ~500 times and consume a lot of cpu cycles.

perf top of 200 concurrent netperf -t TCP_CRR

5.62%  netperf  [kernel.kallsyms]  [k] tcp_init_buffer_space
1.71%  netperf  [kernel.kallsyms]  [k] _raw_spin_lock
1.55%  netperf  [kernel.kallsyms]  [k] kmem_cache_free
1.51%  netperf  [kernel.kallsyms]  [k] tcp_transmit_skb
1.50%  netperf  [kernel.kallsyms]  [k] tcp_ack

Lets use a 100% factor, and remove the loop.

100% is needed anyway for tcp_adv_win_scale=1
default value, and is also the maximum factor.

Refs: commit b49960a0
      ("tcp: change tcp_adv_win_scale and tcp_rmem[2]")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d063
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d063
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d063
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

The driver core clears the driver data to NULL after device_release
or on probe failure, since commit 0998d063
(device-core: Ensure drvdata = NULL when no driver is bound).
Thus, it is not needed to manually clear the device driver data to NULL.
Signed-off-by: Jingoo Han <jg1.han@samsung.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>

Replace a call to deprecated devm_request_and_ioremap by devm_ioremap_resource.
dev_err() message is no more needed since it's already displayed in
devm_ioremap_resource().
Signed-off-by: Laurent Navet <laurent.navet@gmail.com>
Acked-by: Andreas Larsson <andreas@gaisler.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>