Replace the per network namespace shared atomic "mem" accounting
variable, in the fragmentation code, with a lib/percpu_counter.

Getting percpu_counter to scale to the fragmentation code usage
requires some tweaks.

At first view, percpu_counter looks superfast, but it does not
scale on multi-CPU/NUMA machines, because the default batch size
is too small, for frag code usage.  Thus, I have adjusted the
batch size by using __percpu_counter_add() directly, instead of
percpu_counter_sub() and percpu_counter_add().

The batch size is increased to 130.000, based on the largest 64K
fragment memory usage.  This does introduce some imprecise
memory accounting, but its does not need to be strict for this
use-case.

It is also essential, that the percpu_counter, does not
share cacheline with other writers, to make this scale.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This change is primarily a preparation to ease the extension of memory
limit tracking.

The change does reduce the number atomic operation, during freeing of
a frag queue.  This does introduce a some performance improvement, as
these atomic operations are at the core of the performance problems
seen on NUMA systems.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Fragmentation code cacheline adjusting of struct inet_frag_queue.

Take advantage of the size of struct timer_list, and move all but
spinlock_t lock, below the timer struct.  On 64-bit 'lru_list',
'list' and 'refcnt', fits exactly into the next cacheline, and a
new cacheline starts at 'fragments'.

The netns_frags *net pointer is moved to the end of the struct,
because its used in a compare, with "next/close-by" elements of
which this struct is embedded into.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The globally shared rwlock, of struct inet_frags, shares
cacheline with the 'rnd' number, which is used by the hash
calculations.  Fix this, as this obviously is a bad idea, as
unnecessary cache-misses will occur when accessing the 'rnd'
number.

Also small note that, moving function ptr (*match) up in struct,
is to avoid it lands on the next cacheline (on 64-bit).
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This small cacheline adjustment of struct netns_frags improves
performance significantly for the fragmentation code.

Struct members 'lru_list' and 'mem' are both hot elements, and it
hurts performance, due to cacheline bouncing at every call point,
when they share a cacheline.  Also notice, how mem is placed
together with 'high_thresh' and 'low_thresh', as they are used in
the compare operations together.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

just as it should have been. It also helps
removing the, now unnecessary, workqueue.
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When allocating memory for neighbour cache entry, if
tbl->entry_size is not set, we always calculate
sizeof(struct neighbour) + tbl->key_len, which is common
in the same table.

With this change, set tbl->entry_size during the table
initialization phase, if it was not set, and use it in
neigh_alloc() and neighbour_priv().

This change also allow us to have both of protocol private
data and device priate data at tha same time.

Note that the only user of prototcol private is DECnet
and the only user of device private is ATM CLIP.
Since those are exclusive, we have not been facing issues
here.
Signed-off-by: YOSHIFUJI Hideaki <yoshfuji@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

    I found if we write a larger than 4GB value to some sysctl
variables, the sending syscall will hang up forever, because these
variables are 32 bits, such large values make them overflow to 0 or
negative.

    This patch try to fix overflow or prevent from zero value setup
of below sysctl variables:

net.core.wmem_default
net.core.rmem_default

net.core.rmem_max
net.core.wmem_max

net.ipv4.udp_rmem_min
net.ipv4.udp_wmem_min

net.ipv4.tcp_wmem
net.ipv4.tcp_rmem
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Li Yu <raise.sail@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Switch to use ndo_get_stats64 to get 64bit statistics.
Signed-off-by: Jamie Gloudon <jamie.gloudon@gmail.com>
Tested-by: Jamie Gloudon <jamie.gloudon@gmail.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Summary of changes:
.Newly added phys
	-KSZ8081/KSZ8091, which has some phy ids.
	-KSZ8061
	-KSZ9031, which is Gigabit phy.
	-KSZ886X, which has a switch function.
	-KSZ8031, which has a same phy ids with KSZ8021.
Signed-off-by: David J. Choi <david.choi@micrel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch adds the minimal driver to manage the
Realtek RTL8211E 10/100/1000 Transceivers.
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This will allow us to setup netconsole in a different namespace
rather than where init_net is.

Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ipv6_addr_equal() is faster.

Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

dev->npinfo is protected by RCU.

This fixes the following sparse warnings:

net/core/netpoll.c:177:48: error: incompatible types in comparison expression (different address spaces)
net/core/netpoll.c:200:35: error: incompatible types in comparison expression (different address spaces)
net/core/netpoll.c:221:35: error: incompatible types in comparison expression (different address spaces)
net/core/netpoll.c:327:18: error: incompatible types in comparison expression (different address spaces)

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

John W. Linville says:

====================
Included is an NFC pull.  Samuel says:

"It brings the following goodies:

- LLCP socket timestamping (To be used e.g with the recently released nfctool
  application for a more efficient skb timestamping when sniffing).
- A pretty big pn533 rework from Waldemar, preparing the driver to support
  more flavours of pn533 based devices.
- HCI changes from Eric in preparation for the microread driver support.
- Some LLCP memory leak fixes, cleanups and slight improvements.
- pn544 and nfcwilink move to the devm_kzalloc API.
- An initial Secure Element (SE) API.
- An nfc.h license change from the original author, allowing non GPL
  application code to safely include it."

Also included are a pair of mac80211 pulls.  Johannes says:

"We found two bugs in the previous code, so I'm sending you a pull
request again this soon.

This contains two regulatory bug fixes, some of Thomas's hwsim beacon
timer work and a documentation fix from Bob."

"Another pull request for mac80211-next. This time, I have a number of
things, the patches are mostly self-explanatory. There are a few fixes
from Felix and myself, and random cleanups & improvements. The biggest
thing is the partial patchset from Marco preparing for mesh powersave."

Additionally, there are a pair of iwlwifi pulls.  Johannes says:

"For iwlwifi-next, I have a few cleanups/improvements as well as a few
not very important fixes and more preparations for new devices."

"Please pull a few updates for iwlwifi. These are just some cleanups and
a debug improvement."

On top of that, there is a slew of driver updates.  This includes
brcmfmac, mwifiex, ath9k, carl9170, and mwl8k as well as a handful
of others.  The bcma and ssb busses get some attention as well.
Still, I don't see any big headliners here.

Also included is a pull of the wireless tree, in order to resolve
some merge conflicts.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Jeff Kirsher says:

====================
This series contains updates to e1000e, ixgbevf, igb and igbvf.
Majority of the patches are code cleanups of e1000e where code
is removed (Yeah!).  The other two e1000e patches are fixes.  The
first is to fix the maximum frame size for 82579 devices.  The second
fix is to resolve an issue with devices other than 82579 that suffer
from dropped transactions on platforms with deep C-states when
jumbo frames are enabled.

The ixgbevf patch is to ensure that the driver fetches the correct,
refreshed value for link status and speed when the values have changed.

The igb and igbvf patches are a solution to an issue Stefan Assmann
reported, where when the PF is up and igbvf is loaded, the MAC address
is not generated using eth_hw_addr_random().
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

Added accessor and skb_reserve helpers for struct can_skb_priv.
Removed pointless skb_headroom() check.
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
CC: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

Conflicts:
	drivers/net/wireless/ath/ath9k/main.c
	drivers/net/wireless/iwlwifi/dvm/tx.c

Tighten up some of the code surrounding MAC addresses. Since the PF is
now giving all zeros instead of a random address, check for this case
and generate a random address. This ensures that we always know when we
have a random address and udev won't get upset about it.

Additionally, tighten up some of the log messages and clean up the
formatting.
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
CC: Andy Gospodarek <andy@greyhouse.net>
CC: Stefan Assmann <sassmann@kpanic.de>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Tested-by: Stefan Assmann <sassmann@redhat.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

If the user has not assigned a MAC address to a VM, then don't give it a
random one. Instead, just give it zeros and let it figure out what to do
with them.
Signed-off-by: Mitch Williams <mitch.a.williams@intel.com>
CC: Andy Gospodarek <andy@greyhouse.net>
CC: Stefan Assmann <sassmann@kpanic.de>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Tested-by: Stefan Assmann <sassmann@redhat.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

A recent change makes it necessary to set get_link_status to ensure that
the driver fetches the correct, refreshed value for link status and speed
when it has changed in the physical function device.
Signed-off-by: Greg Rose <gregory.v.rose@intel.com>
Tested-by: Sibai Li <sibai.li@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Code was removed but the applicable comments were not.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Remove unnecessary #include, forward prototype of struct e1000_adapter and
an empty comment; fix a comment which mentions "static data for the MAC"
which is not applicable to the following struct; and cleanup some
whitespace issues.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

All references to E1000_ERT_2048 have been removed.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

It has been found that devices other than 82579 (a.k.a. e1000_pch2lan)
suffer from dropped transactions on platforms with deep C-states when
jumbo frames are enabled.  For example, LOMs on ICH9- and ICH10-based
platforms which recently had early-receive de-featured (for stability
reasons) suffer from this.  To resolve this for all devices, when jumbo
frames are enabled set the PM QoS DMA latency request based on the size
of the receive packet buffer less one full frame.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

The largest jumbo frame supported by the 82579 hardware is 9018.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Remove the function e1000e_commit_phy() and replace the few calls to it
with the same function pointer that it would call.  The function pointer is
almost always set for the devices that access these code paths so there is
no risk of a NULL pointer dereference; for the few instances where the
function pointer might not be set (i.e. can be called for the few devices
which do not have this function pointer set), check for a valid function
pointer.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Jeff Pieper <jeffrey.e.pieper@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Remove the function e1000_get_cable_length() and replace the two calls
to it with the same function pointer that it would call.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Remove the function e1000_get_phy_cfg_done() and replace the single call
to it with the same function pointer that it would call.  The function
pointer is always set so there is no risk of a NULL pointer dereference.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

In keeping with the e1000e driver function naming convention, the subject
function is renamed to indicate it is generic, i.e. it is applicable to
more than just a single MAC family (e.g. 80003es2lan, 82571, ich8lan).
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Remove the function e1000_force_speed_duplex() and replace the single call
to it with the same function pointer that it would call.  The function
pointer is always set so there is no risk of a NULL pointer dereference.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Replace the function e1000_set_d0_lplu_state() with the contents of it
coded in place of the single call to the function.
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com>
Tested-by: Aaron Brown <aaron.f.brown@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>

Pravin Shelar mentioned that GSO could potentially generate
wrong TX checksum if skb has fragments that are overwritten
by the user between the checksum computation and transmit.

He suggested to linearize skbs but this extra copy can be
avoided for normal tcp skbs cooked by tcp_sendmsg().

This patch introduces a new SKB_GSO_SHARED_FRAG flag, set
in skb_shinfo(skb)->gso_type if at least one frag can be
modified by the user.

Typical sources of such possible overwrites are {vm}splice(),
sendfile(), and macvtap/tun/virtio_net drivers.

Tested:

$ netperf -H 7.7.8.84
MIGRATED TCP STREAM TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to
7.7.8.84 () port 0 AF_INET
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 87380  16384  16384    10.00    3959.52

$ netperf -H 7.7.8.84 -t TCP_SENDFILE
TCP SENDFILE TEST from 0.0.0.0 (0.0.0.0) port 0 AF_INET to 7.7.8.84 ()
port 0 AF_INET
Recv   Send    Send
Socket Socket  Message  Elapsed
Size   Size    Size     Time     Throughput
bytes  bytes   bytes    secs.    10^6bits/sec

 87380  16384  16384    10.00    3216.80

Performance of the SENDFILE is impacted by the extra allocation and
copy, and because we use order-0 pages, while the TCP_STREAM uses
bigger pages.
Reported-by: Pravin Shelar <pshelar@nicira.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Marc Kleine-Budde says:

====================
this is a pull-request for net-next/master. There is are 9 patches by
Fabio Baltieri and Kurt Van Dijck which add LED infrastructure and
support for CAN devices. Bernd Krumboeck adds a driver for the USB CAN
adapter from 8 devices. Oliver Hartkopp improves the CAN gateway
functionality. There are 4 patches by me, which clean up the CAN's
Kconfig.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

sock->sk_dst_cache is protected by RCU.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sock->sk_dst_cache is protected by RCU, therefore we should
use __sk_dst_get() to deref it once we lock the sock.

This fixes several sparse warnings.

Cc: linux-decnet-user@lists.sourceforge.net
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Cong Wang <amwang@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

filters_lock might have been used while it was re-initialized.
Moved filters_lock and filters_list initialization to init_netdev instead of
alloc_resources which is called every time the device is configured.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There is a possible race where the TX completion handler can clean the
entire TX queue between the decision that the queue is full and actually
closing it. To avoid this situation, check again if the queue is really
full, if not, reopen the transmit and continue with sending the packet.

CC: Eric Dumazet <edumazet@google.com>
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.com>
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Returning 0 (success) when in fact we are aborting the load, leads to kernel
panic when unloading the module. Fix that by returning the actual error code.
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>