Partial revert "Basic kernel memory functionality for the Memory Controller"

This reverts commit e5671dfae59b165e2adfd4dfbdeab11ac8db5bda.

After a follow up discussion with Michal, it was agreed it would
be better to leave the kmem controller with just the tcp files,
deferring the behavior of the other general memory.kmem.* files
for a later time, when more caches are controlled. This is because
generic kmem files are not used by tcp accounting and it is
not clear how other slab caches would fit into the scheme.

We are reverting the original commit so we can track the reference.
Part of the patch is kept, because it was used by the later tcp
code. Conflicts are shown in the bottom. init/Kconfig is removed from
the revert entirely.

Signed-off-by: Glauber Costa <glommer@parallels.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
CC: Kirill A. Shutemov <kirill@shutemov.name>
CC: Paul Menage <paul@paulmenage.org>
CC: Greg Thelen <gthelen@google.com>
CC: Johannes Weiner <jweiner@redhat.com>
CC: David S. Miller <davem@davemloft.net>
Conflicts:

Documentation/cgroups/memory.txt
mm/memcontrol.c
Signed-off-by: David S. Miller <davem@davemloft.net>

irda: use msecs_to_jiffies() rather than manual calculation

Also use mod_timer() instead of direct assignment to "expires".

Signed-off-by: Xi Wang <xi.wang@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

stmmac: update the driver's documentation (Dec-2011)

Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

stmmac: add the experimental PCI support

This patch adds the PCI support (as EXPERIMENTAL)
this has been also tested on XLINX XC2V3000 FF1152AMT0221
D1215994A VIRTEX FPGA board.
To support the PCI bus the main part has been reworked
and both the platform and the PCI specific parts have
been moved into different files.

Signed-off-by: Rayagond Kokatanur <rayagond@vayavyalabs.com>
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sch_sfq: rehash queues in perturb timer

A known Out Of Order (OOO) problem hurts SFQ when timer changes
perturbation value, since all new packets delivered to SFQ enqueue might
end on different slots than previous in-flight packets.

With round robin delivery, we can thus deliver packets in a different
order.

Since SFQ is limited to small amount of in-flight packets, we can rehash
packets so that this OOO problem is fixed.

This rehashing is performed only if internal flow classifier is in use.

We now store in skb->cb[] the "struct flow_keys" so that we dont call
skb_flow_dissect() again while rehashing.

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

net: ethernet: xilinx: Don't use NO_IRQ in xilinx

Fix ll_temac and emaclite drivers. Only Microblaze and Xilinx PPC
use then and both use NO_IRQ as 0. It will be removed in near future.

Signed-off-by: Michal Simek <monstr@monstr.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>

tcp: Replace constants with #define macros

to record the state of SACK/FACK and DSACK for better readability and maintenance.

Signed-off-by: Vijay Subramanian <subramanian.vijay@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: doc: fix many typos in scaling.txt

Fix some trivial typos.

Signed-off-by: Shan Wei <shanwei88@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qeth: recovery through asynchronous delivery

If recovery is triggered in presence of pending asynchronous
deliveries of storage blocks we do a forced cleanup after
the corresponding tasklets are completely stopped and trigger
appropriate notifications for the correspondingerror state.

Signed-off-by: Einar Lueck <elelueck@de.ibm.com>
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: improve recovery during resource shortage

In case there are no system resources to run a recovery we have to clear
recovery bitmasks so a further automatic or manual driven recovery can
fix up the device.

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

netiucv: allow multiple interfaces to same peer

The NETIUCV device driver allows to connect a Linux guest on z/VM to
another z/VM guest based on the z/VM communication facility IUCV.
Multiple output paths to different guests are possible, as well as
multiple input paths from different guests.
With this feature, you can configure multiple point-to-point NETIUCV
interfaces between your Linux on System z instance and another z/VM
guest.

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: forbid recovery during shutdown

A recovery does not make sense during shutdown and may even cause an
error like this:

qeth 0.0.f503: A recovery process has been started for the device
Badness at drivers/s390/cio/qdio_main.c:1156
Modules linked in: autofs4 sunrpc dm_multipath scsi_dh scsi_mod qeth_l3 ipv6 vmu
r qeth qdio ccwgroup ext3 jbd mbcache dasd_eckd_mod dasd_mod dm_mirror dm_region
_hash dm_log dm_mod [last unloaded: scsi_wait_scan]
CPU: 3 Not tainted 2.6.32-202.el6.s390x #1
Process qeth_recover (pid: 1498, task: 000000003efe2040, ksp: 000000003d5e3b80)
Krnl PSW : 0404200180000000 000003c000be6da8 (qdio_int_handler+0x88/0x43c [qdio]
)
           R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:0 CC:2 PM:0 EA:3
Krnl GPRS: ffffffffffff3bac 0000000000000005 0000000000000000 fffffffffffffff4
           0000000000000000 000000000000000c 0000000000000000 000000003ca97000
           0000000000000380 fffffffffffffff4 000000003f22d800 000000003f22c478
           000003c000bdf000 000003c000bea270 000000003f447e10 000000003f447db0
Krnl Code: 000003c000be6d9a: c21f00000004       clfi    %r1,4
           000003c000be6da0: a7c40021           brc     12,3c000be6de2
           000003c000be6da4: a7f40001           brc     15,3c000be6da6
          >000003c000be6da8: e320a0080004       lg      %r2,8(%r10)
           000003c000be6dae: a7390003           lghi    %r3,3
           000003c000be6db2: a72b0178           aghi    %r2,376
           000003c000be6db6: a7490001           lghi    %r4,1
           000003c000be6dba: a7590000           lghi    %r5,0
Call Trace:
([<000000000080ee80>] __per_cpu_offset+0x0/0x200)
[<00000000003d90e8>] ccw_device_call_handler+0x70/0xcc
[<00000000003d83a2>] ccw_device_irq+0x82/0x180
[<00000000003cc6a8>] do_IRQ+0x16c/0x1ec
[<0000000000118abe>] io_return+0x0/0x8
[<000003c000d04c74>] qeth_determine_capabilities+0x208/0x5cc [qeth]
([<000003c000d04c4a>] qeth_determine_capabilities+0x1de/0x5cc [qeth])
[<000003c000d0a6e0>] qeth_core_hardsetup_card+0x160/0x1258 [qeth]
[<000003c000f49f56>] __qeth_l3_set_online+0x132/0xb14 [qeth_l3]
[<000003c000f4ac70>] qeth_l3_recover+0x168/0x224 [qeth_l3]
[<000000000016e210>] kthread+0xa4/0xac
[<0000000000109c6e>] kernel_thread_starter+0x6/0xc
[<0000000000109c68>] kernel_thread_starter+0x0/0xc

The patch forbids start of a recovery once qeth shutdown is running.

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: suspicious rcu_dereference_check in recovery

qeth layer3 recovery invokes its set_multicast_list function, which
invokes function __vlan_find_dev_deep requiring rcu_read_lock or
rtnl lock. This causes kernel messages:

kernel: [ INFO: suspicious rcu_dereference_check() usage. ]
kernel: ---------------------------------------------------
kernel: net/8021q/vlan_core.c:70 invoked rcu_dereference_check() without protection!

kernel: stack backtrace:
kernel: CPU: 0 Not tainted 3.1.0 #9
kernel: Process qeth_recover (pid: 2078, task: 000000007e584680, ksp: 000000007e3e3930)
kernel: 000000007e3e3d08 000000007e3e3c88 0000000000000002 0000000000000000
kernel:       000000007e3e3d28 000000007e3e3ca0 000000007e3e3ca0 00000000005e77ce
kernel:       0000000000000000 0000000000000001 ffffffffffffffff 0000000000000001
kernel:       000000000000000d 000000000000000c 000000007e3e3cf0 0000000000000000
kernel:       0000000000000000 0000000000100a18 000000007e3e3c88 000000007e3e3cc8
kernel: Call Trace:
kernel: ([<0000000000100926>] show_trace+0xee/0x144)
kernel: [<00000000005d395c>] __vlan_find_dev_deep+0xb0/0x108
kernel: [<00000000004acd3a>] qeth_l3_set_multicast_list+0x976/0xe38
kernel: [<00000000004ae0f4>] __qeth_l3_set_online+0x75c/0x1498
kernel: [<00000000004aefec>] qeth_l3_recover+0xc4/0x1d0
kernel: [<0000000000185372>] kthread+0xa6/0xb0
kernel: [<00000000005ed4c6>] kernel_thread_starter+0x6/0xc
kernel: [<00000000005ed4c0>] kernel_thread_starter+0x0/0xc

The patch makes sure the rtnl lock is held once qeth recovery invokes
its set_multicast_list function.

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>

af_iucv: get rid of state IUCV_SEVERED

af_iucv differs unnecessarily between state IUCV_SEVERED and
IUCV_DISCONN. This patch removes state IUCV_SEVERED.
While simplifying af_iucv, this patch removes the 2nd invocation of
cpcmd as well.

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>

af_iucv: remove unused timer infrastructure

af_iucv contains timer infrastructure which is not exploited.
This patch removes the timer related code parts.

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>

af_iucv: release reference to HS device

For HiperSockets transport skbs sent are bound to one of the
available HiperSockets devices. Add missing release of reference to
a HiperSockets device before freeing an skb.

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>

af_iucv: accelerate close for HS transport

Closing an af_iucv socket may wait for confirmation of outstanding
send requests. This patch adds confirmation code for the new
HiperSockets transport.

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>

af_iucv: support ancillary data with HS transport

The AF_IUCV address family offers support for ancillary data.
This patch enables usage of ancillary data with the new
HiperSockets transport.

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>

macvtap: Fix macvtap_get_queue to use rxhash first

It was reported that the macvtap device selects a
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlx4_en: FIX: Setting default_qpn before using it

When UDP RSS is enabled, we use same QPN for TCP and UDP ranges
The bug is that the default_qpn was used base UDP qpn before it
was set.
Fixes bug introduced in commit: 1202d460b1df3a77fda66f4ba5f90ae3527dd42f

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: unix -- Add missing module.h inclusion

Otherwise getting

 | net/unix/diag.c:312:16: error: expected declaration specifiers or ‘...’ before string constant
 | net/unix/diag.c:313:1: error: expected declaration specifiers or ‘...’ before string constant

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

module_param: make bool parameters really bool (net & drivers/net)

module_param(bool) used to counter-intuitively take an int.  In
fddd5201 (mid-2009) we allowed bool or int/unsigned int using a messy
trick.

It's time to remove the int/unsigned int option.  For this version
it'll simply give a warning, but it'll break next kernel version.

(Thanks to Joe Perches for suggesting coccinelle for 0/1 -> true/false).

Cc: "David S. Miller" <davem@davemloft.net>
Cc: netdev@vger.kernel.org
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: fix assignment of 0/1 to bool variables.

DaveM said:
   Please, this kind of stuff rots forever and not using bool properly
   drives me crazy.

Joe Perches <joe@perches.com> gave me the spatch script:

@@
bool b;
@@
-b = 0
+b = false
@@
bool b;
@@
-b = 1
+b = true

I merely installed coccinelle, read the documentation and took credit.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

phy: Micrel KS8995MA 5-ports 10/100 managed Ethernet switch support added

Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
Signed-off-by: Frederic Lambert <frdrc66@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

de4x5/PCI: use list_for_each_entry() for bus->devices traversal

Replace open-coded list traversal with list_for_each_entry().

CC: Grant Grundler <grundler@parisc-linux.org>
CC: netdev@vger.kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sunhme/PCI: use list_for_each_entry() for bus->devices traversal

Replace open-coded list traversal with list_for_each_entry().

CC: David S. Miller <davem@davemloft.net>
CC: netdev@vger.kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

qla3xxx: Adding Maintainer.

Signed-off-by: Jitendra Kalsaria <jitendra.kalsaria@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: two vzalloc() cleanups

We can use vzalloc() helper now instead of __vmalloc() trick

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

tg3: Make the RSS indir tbl admin configurable

This patch adds the ethtool callbacks necessary to change the rss
indirection table from userspace.  Should the number of interrupts
change (e.g.  across a close / open call, or through a reset) and
any one of the indirection table values fall out-of-range, the driver
will reset the indirection table to a default layout.

[Integrated many suggestions made by Ben Hutchings.]

Changes since v3

* Removed TG3_FLAG_SUPPORT_MSIX checks at the start of
  tg3_get_rxfh_indir() and tg3_set_rxfh_indir().

Signed-off-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Benjamin Li <benli@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bnx2: Update driver to use new mips firmware.

bnx2-mips-06-6.2.3 and bnx2-mips-09-6.2.1.b

New firmware fixes iSCSI problems with some LeftHand targets that don't
set TTT=0xffffffff for Data-In according to spec.  Firmware generates
exception warnings for this condition and becomes very slow.  This is
fixed by suppressing these warnings when using default error mask.

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Fix INTx processing for Lancer

Lancer does not have HW registers to indicate the EQ causing the INTx
interrupt. As a result EQE entries of one EQ may be consumed when interrupt
is caused by another EQ. Fix this by arming CQs at the end of NAPI poll
routine to regenerate the EQEs.

Signed-off-by: Padmanabh Ratnakar <padmanabh.ratnakar@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

myri10ge: Fix typo of 'VMware' in comment.

Signed-off-by: Vinson Lee <vlee@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Revert "net: Remove unused neighbour layer ops."

This reverts commit 5c3ddec73d01a1fae9409c197078cb02c42238c3.

S390 qeth driver actually still uses the setup ops.

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

mlx4: Fixing wrong error codes in communication channel

The communication channel is HW interface from PF point of view
So the command return status should be stored as HW error code
and only then translated to errno values.

Reporetd-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlx4: not using spin_lock_irq when getting vf by resource.

The function is always called from irq context, changing the call
to spin_lock().

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlx4_en: nullify cached multicast address list after cleanup

Solves an issue where we tried to free the same page twice after
the port has been opened and closed.

Signed-off-by: Alexander Guller <alexg@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlx4_core: Changing link sensing logic

New FW can give clues to driver regarding default port type
and whether or not we should default to link sensing on the port.

2 bits are added to QUERY_PORT command:
1. suggested_type: This bit gives a hint whether the default port type should be
   IB or Ethernet.
   The driver will use this hint in case the user didn't specify explicitly the link layer
   type he wants to set.
2. default_sense: If this bit is set, we would sense the port type on start-up
   and default the port to link sensing

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlx4: capability for link sensing

For ConnectX3 devices, we allow link sensing only if FW explicitly
reported it supports the feature.
For older versions (ConnectX1 and 2), if the card supports both link layer types
(Ethenet and Infiniband), link sensing is supported.

Signed-off-by: Yevgeny Petrilin <yevgenyp@mellanox.co.il>
Signed-off-by: David S. Miller <davem@davemloft.net>

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

wimax/i2400m: remove an unused variable

"result" isn't used.  We ignore errors here because there is not much we
can do about them.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net:bridge: use IS_ENABLED

Use IS_ENABLED(CONFIG_FOO)
instead of defined(CONFIG_FOO) || defined (CONFIG_FOO_MODULE)

Signed-off-by: Igor Maravić <igorm@etf.rs>
Signed-off-by: David S. Miller <davem@davemloft.net>

net:netfilter: use IS_ENABLED

Use IS_ENABLED(CONFIG_FOO)
instead of defined(CONFIG_FOO) || defined (CONFIG_FOO_MODULE)

Signed-off-by: Igor Maravić <igorm@etf.rs>
Signed-off-by: David S. Miller <davem@davemloft.net>

net:x25: use IS_ENABLED

Use IS_ENABLED(CONFIG_FOO)
instead of defined(CONFIG_FOO) || defined (CONFIG_FOO_MODULE)

Signed-off-by: Igor Maravić <igorm@etf.rs>
Signed-off-by: David S. Miller <davem@davemloft.net>

net:core: use IS_ENABLED

Use IS_ENABLED(CONFIG_FOO)
instead of defined(CONFIG_FOO) || defined (CONFIG_FOO_MODULE)

Signed-off-by: Igor Maravić <igorm@etf.rs>
Signed-off-by: David S. Miller <davem@davemloft.net>

be2net: Add support for Skyhawk cards

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sch_gred: prefer GFP_KERNEL allocations

In control path, its better to use GFP_KERNEL allocations where
possible.

Before taking qdisc spinlock, we preallocate memory just in case we'll
need it in gred_change_vq()

This is a followup to commit 3f1e6d3fd37b (sch_gred: should not use
GFP_KERNEL while holding a spinlock)

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

net: fix compilation with !CONFIG_NET

Reported-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Glauber Costa <glommer@parallels.com>
CC: Hiroyouki Kamezawa <kamezawa.hiroyu@jp.fujitsu.com>
CC: David S. Miller <davem@davemloft.net>
CC: Eric Dumazet <eric.dumazet@gmail.com>
CC: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: fix sleeping while atomic problem in sock mem_cgroup.

We can't scan the proto_list to initialize sock cgroups, as it
holds a rwlock, and we also want to keep the code generic enough to
avoid calling the initialization functions of protocols directly,

Convert proto_list_lock into a mutex, so we can sleep and do the
necessary allocations. This lock is seldom taken, so there shouldn't
be any performance penalties associated with that

Signed-off-by: Glauber Costa <glommer@parallels.com>
CC: Hiroyouki Kamezawa <kamezawa.hiroyu@jp.fujitsu.com>
CC: David S. Miller <davem@davemloft.net>
CC: Eric Dumazet <eric.dumazet@gmail.com>
CC: Stephen Rothwell <sfr@canb.auug.org.au>
CC: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlx4: Fix compile error when driver is comiled-in

This patch fixes a compile error that occurs when the driver
is compile into the kernel and not as a module.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'for-davem' of git://git./linux/kernel/git/bwh/sfc-next

batman-adv: Fix merge error.

I didn't resolve the merge properly during the last pull of the net
tree into net-next.

The code in the final resolution should set flags to TT_CLIENT_ROAM
not TT_CLIENT_PENDING.

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

sfc: Use skb_fill_page_desc() to simplify passing of page buffers to GRO

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

ethtool: Define and apply a default policy for RX flow hash indirection

All drivers that support modification of the RX flow hash indirection
table initialise it in the same way: RX rings are assigned to table
entries in rotation.  Make that default policy explicit by having them
call a ethtool_rxfh_indir_default() function.

In the ethtool core, add support for a zero size value for
ETHTOOL_SRXFHINDIR, which resets the table to this default.

Partly-suggested-by: Matt Carlson <mcarlson@broadcom.com>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Acked-by: Shreyas N Bhatewara <sbhatewara@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ethtool: Centralise validation of ETHTOOL_{G, S}RXFHINDIR parameters

Add a new ethtool operation (get_rxfh_indir_size) to get the
indirectional table size.  Use this to validate the user buffer size
before calling get_rxfh_indir or set_rxfh_indir.  Use get_rxnfc to get
the number of RX rings, and validate the contents of the new
indirection table before calling set_rxfh_indir.  Remove this
validation from drivers.

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

ethtool: Clarify use of size field for ETHTOOL_GRXFHINDIR

In order to find out the device's RX flow hash table size, ethtool
initially uses ETHTOOL_GRXFHINDIR with a buffer size of zero.  This
must be supported, but it is not necessary to support any other user
buffer size less than the device table size.

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

unix_diag: Write it into kbuild

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Receive queue lenght NLA

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Pending connections IDs NLA

When establishing a unix connection on stream sockets the
server end receives an skb with socket in its receive queue.

Report who is waiting for these ends to be accepted for
listening sockets via NLA.

There's a lokcing issue with this -- the unix sk state lock is
required to access the peer, and it is taken under the listening
sk's queue lock. Strictly speaking the queue lock should be taken
inside the state lock, but since in this case these two sockets
are different it shouldn't lead to deadlock.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Unix peer inode NLA

Report the peer socket inode ID as NLA. With this it's finally
possible to find out the other end of an interesting unix connection.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Unix inode info NLA

Actually, the socket path if it's not anonymous doesn't give
a clue to which file the socket is bound to. Even if the path
is absolute, it can be unlinked and then new socket can be
bound to it.

With this NLA it's possible to check which file a particular
socket is really bound to.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Unix socket name NLA

Report the sun_path when requested as NLA. With leading '\0' if
present but without the leading AF_UNIX bits.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Dumping exact socket core

The socket inode is used as a key for lookup. This is effectively
the only really unique ID of a unix socket, but using this for
search currently has one problem -- it is O(number of sockets) :(

Does it worth fixing this lookup or inventing some other ID for
unix sockets?

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Dumping all sockets core

Walk the unix sockets table and fill the core response structure,
which includes type, state and inode.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

unix_diag: Basic module skeleton

Includes basic module_init/_exit functionality, dump/get_exact stubs
and declares the basic API structures for request and response.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

af_unix: Export stuff required for diag module

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sock_diag: Generalize requests cookies managements

The sk address is used as a cookie between dump/get_exact calls.
It will be required for unix socket sdumping, so move it from
inet_diag to sock_diag.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sock_diag: Fix module netlink aliases

I've made a mistake when fixing the sock_/inet_diag aliases :(

1. The sock_diag layer should request the family-based alias,
   not just the IPPROTO_IP one;
2. The inet_diag layer should request for AF_INET+protocol alias,
   not just the protocol one.

Thus fix this.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

sock_diag: Move the SOCK_DIAG_BY_FAMILY cmd declaration

It should belong to sock_diag, not inet_diag.

Signed-off-by: Pavel Emelyanov <xemul@parallels.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

ath9k: Support RSN Mesh

Signed-off-by: Chun-Yeow Yeoh <yeohchunyeow@gmail.com>
Acked-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

cfg80211: Restore orig channel values upon disconnect

When we restore regulatory settings the world regulatory domain
is properly reset on cfg80211 (or user prefered regulatory domain)
but we were never setting back channel values for drivers that use
WIPHY_FLAG_CUSTOM_REGULATORY. Set these values up again by using
the orig_ channel parameters.

This fixes restoring custom regulatory settings upon disconnect
events.

Cc: compat@orbit-lab.org
Cc: Paul Stewart <pstew@google.com>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Cc: Senthilkumar Balasubramanian <senthilb@qca.qualcomm.com>
Signed-off-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

cfg80211: allow following country IE power for custom regdom cards

By definition WIPHY_FLAG_STRICT_REGULATORY was intended to allow the
wiphy to adjust itself to the country IE power information if the
card had no regulatory data but we had no way to tell cfg80211 that if
the card also had its own custom regulatory domain (these are typically
custom world regulatory domains) that we want to follow the country IE's
noted values for power for each channel. We add support for this and
document it.

This is not a critical fix but a performance optimization for cards
with custom regulatory domains that associate to an AP with sends
out country IEs with a higher EIRP than the one on the custom
regulatory domain. In practice the only driver affected right now
are the Atheros drivers as they are the only drivers using both
WIPHY_FLAG_STRICT_REGULATORY and WIPHY_FLAG_CUSTOM_REGULATORY --
used on cards that have an Atheros world regulatory domain. Cards
that have been programmed to follow a country specifically will not
follow the country IE power. So although not a stable fix distributions
should consider cherry picking this.

Cc: compat@orbit-lab.org
Cc: Paul Stewart <pstew@google.com>
Cc: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Cc: Senthilkumar Balasubramanian <senthilb@qca.qualcomm.com>
Reported-by: Rajkumar Manoharan <rmanohar@qca.qualcomm.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211_hwsim: fix wmediumd_pid

Fix a few minor issues with wmediumd_pid:
a) make static
b) use u32 to match the snd_pid type
c) use ACCESS_ONCE since we don't lock it
d) don't explicitly initialize to 0

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: remove ATH9K_HW_CAP_CST

its not used anywhere in the current code

Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: validate for non-zero BSSID

before concluding that the recieved beacon is for us, let us make sure
that the BSSID is non-zero. when I configured ad-hoc mode as creator and
left it for some time without joining I found we recieved few frames whose
BSSID is zero, which we concluded wrongly as 'my_beacons'

Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Merge git://git./linux/kernel/git/davem/net

Conflicts:
drivers/net/ethernet/freescale/fsl_pq_mdio.c
net/batman-adv/translation-table.c
net/ipv6/route.c

b43: N-PHY: check for bustype before touching BCMA CC PLLs

Reported-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcm80211: smac: precendence bug in wlc_phy_attach()

Negate has higher precendence than compare and since neither zero nor
one are equal to four or eight the original condition is always false.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: add DFS radar pulse processing

This initial DFS module provides basic functionality to deal
with radar pulses reported by the Atheros DFS HW pulse detector.

The reported data is evaluated and basic plausibility checks
are performed to filter false pulses. Passing radar pulses are
forwarded to pattern detectors which are not yet implemented.

(Some modifications to actually use ATH9K_DFS_DEBUGFS based on comments
from Julian Calaby <julian.calaby@gmail.com>. -- JWL)

Signed-off-by: Zefir Kurtisi <zefir.kurtisi@neratec.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k_hw: add DFS testing check

In order to enable DFS upstream we want to be sure
DFS has been tested for each chipset. Push for public
documentation of the requirements we want in place and
allow for enabling each chipset through a single upstream
commit.

Signed-off-by: Zefir Kurtisi <zefir.kurtisi@neratec.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath: add a debug level for DFS

This can later be used by other drivers that implement
DFS support.

Signed-off-by: Zefir Kurtisi <zefir.kurtisi@neratec.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

brcmsmac: Replace kmalloc/memset with kzalloc

In ai_attach(), space is allocated for an si_info struct. Immediately
after the allocation, routine ai_doattach() is called and that allocated
space is set to zero. As no other routine calls ai_doattach(), kzalloc()
can be utilized.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Acked-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath: document ATH_DBG_MCI

A debug level was added to the ath module for printing
MCI messages but no documentation was provided. Clarify that
MCI is the Message Coexistence Interface, a private protocol
used exclusively for WLAN-BT coexistence starting from
AR9462.

Cc: wtsao@qca.qualcomm.com
Cc: rmanohar@qca.qualcomm.com
Cc: mohammed@qca.qualcomm.com
Cc: senthilb@qca.qualcomm.com
Signed-off-by: Luis R. Rodriguez <mcgrof@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: avoid retransmitting aggregation frames that a BAR was sent for

The receiver will discard them anyway.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: simplify tx locking

Instead of releasing and taking back the lock over and over again in the
tx path, hold the lock a bit longer, requiring much fewer lock/unlock pairs.
This makes locking much easier to review and should not have any noticeable
performance/latency impact.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: remove bogus sequence number increment

tid->seq_next is initialized on A-MPDU start anyway, and the comment next
to this chunk of code seems to be bogus as well.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: reduce indentation level in a few places

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: reduce the number of unnecessary BAR tx packets

When processing A-MPDU tx status, only send a BAR for the failed packet
with the highest sequence number.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: change maximum software retransmission handling

Instead of limiting a subframe to 10 A-MPDU software transmission attempts,
count hardware retransmissions as well and raise the limit a bit. That way
there will be fewer software retransmission attempts when traffic suffers
from lots of hardware retransmissions.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: Fix power save in change interface

we found that power save is not getting enabled when we do
change interface in this order STA->IBSS->STA. this is
because ieee80211_setup_sdata clears type-dependent union

Reported-by: Leela Kella <leela@qca.qualcomm.com>
Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: remove an unnecessary paraenthesis

Signed-off-by: Mohammed Shafi Shajakhan <mohammed@qca.qualcomm.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: Use appropriate TID for sending BAR, ADDBA and DELBA frames

Currently BAR, ADDBA and DELBA frames are always sent using AC_VO. If
the TID for which a BA session is established is assigned to a different
queue BAR, ADDBA and DELBA frames can "overtake" frames of the according
BA session.

Hence, always put BA session related frames into the same queue as the
BA sessions data frames.

Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: reduce station management complexity

Now that IBSS no longer needs to insert stations
from atomic context, we can get rid of all the
special cases for that, and even get rid of the
sta_lock (though it needs to stay as tim_lock.)

This makes the station management code much more
straight-forward.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: delay IBSS station insertion

In order to notify drivers and simplify the station
management code, defer IBSS station insertion to a
work item and don't do it directly while receiving
a frame.

This increases the complexity in IBSS a little bit,
but it's pretty straight forward and it allows us
to reduce the station management complexity (next
patch) considerably.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: make address arguments to sta_info_alloc const

No real changes, just note that they are const.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: count authorized stations per BSS

Currently, each AP interface will send multicast
traffic if any interface has a station entry even
if that station entry is allocated only. With the
new station state management we can easily fix it
by adding a counter that counts each authorized
station only and send multicast traffic only when
the correct interface has at least one authorized
station.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: refactor station state transitions

Station entries can have various states, the most
important ones being auth, assoc and authorized.
This patch prepares us for telling the driver about
these states, we don't want to confuse drivers with
strange transitions, so with this we enforce that
they move in the right order between them (back and
forth); some transitions might happen before the
driver even knows about the station, but at least
runtime transitions will be ordered correctly.

As a consequence, IBSS and MESH stations will now
have the ASSOC flag set (so they can transition to
AUTHORIZED), and we can get rid of a special case
in TX processing.

When freeing a station, unwind the state so that
other parts of the code (or drivers later) can rely
on the transitions.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: use station mutex in configuration

There's no need to use RCU here, we can just lock
the station mutex instead. This allows the code
to sleep, which is necessary for later patches.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: remove duplicate TDLS peer verification

This is already checked in cfg80211, so no need
to repeat the checks here.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

cfg80211: validate nl80211 station handling better

The nl80211 station handling code is a bit messy
and doesn't do a lot of validation. It seems like
this could be an issue for drivers that don't use
mac80211 to validate everything.

As cfg80211 doesn't keep station state, move the
validation of allowing supported_rates to change
for TDLS only in station mode to mac80211.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

nl80211: add TDLS peer flag to policy

This was evidently missed in the TDLS patch (07ba55d7).

Cc: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>