IPVS: Allow boot time change of hash size

I was very frustrated about the fact that I have to recompile the kernel
to change the hash size. So, I created this patch.

If IPVS is built-in you can append ip_vs.conn_tab_bits=?? to kernel
command line, or, if you built IPVS as modules, you can add
options ip_vs conn_tab_bits=??.

To keep everything backward compatible, you still can select the size at
compile time, and that will be used as default.

It has been about a year since this patch was originally posted
and subsequently dropped on the basis of insufficient test data.

Mark Bergsma has provided the following test results which seem
to strongly support the need for larger hash table sizes:

We do however run into the same problem with the default setting (212 =
4096 entries), as most of our LVS balancers handle around a million
connections/SLAB entries at any point in time (around 100-150 kpps
load). With only 4096 hash table entries this implies that each entry
consists of a linked list of 256 connections *on average*.

To provide some statistics, I did an oprofile run on an 2.6.31 kernel,
with both the default 4096 table size, and the same kernel recompiled
with IP_VS_CONN_TAB_BITS set to 18 (218 = 262144 entries). I built a
quick test setup with a part of Wikimedia/Wikipedia's live traffic
mirrored by the switch to the test host.

With the default setting, at ~ 120 kpps packet load we saw a typical %si
CPU usage of around 30-35%, and oprofile reported a hot spot in
ip_vs_conn_in_get:

samples  %        image name               app name
symbol name
1719761  42.3741  ip_vs.ko                 ip_vs.ko      ip_vs_conn_in_get
302577    7.4554  bnx2                     bnx2          /bnx2
181984    4.4840  vmlinux                  vmlinux       __ticket_spin_lock
128636    3.1695  vmlinux                  vmlinux       ip_route_input
74345     1.8318  ip_vs.ko                 ip_vs.ko      ip_vs_conn_out_get
68482     1.6874  vmlinux                  vmlinux       mwait_idle

After loading the recompiled kernel with 218 entries, %si CPU usage
dropped in half to around 12-18%, and oprofile looks much healthier,
with only 7% spent in ip_vs_conn_in_get:

samples  %        image name               app name
symbol name
265641   14.4616  bnx2                     bnx2         /bnx2
143251    7.7986  vmlinux                  vmlinux      __ticket_spin_lock
140661    7.6576  ip_vs.ko                 ip_vs.ko     ip_vs_conn_in_get
94364     5.1372  vmlinux                  vmlinux      mwait_idle
86267     4.6964  vmlinux                  vmlinux      ip_route_input

[ horms@verge.net.au: trivial up-port and minor style fixes ]
Signed-off-by: Catalin(ux) M. BOIE <catab@embedromix.ro>
Cc: Mark Bergsma <mark@wikimedia.org>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Patrick McHardy <kaber@trash.net>

netfilter: xtables: obtain random bytes earlier, in checkentry

We can initialize the random hash bytes on checkentry. This is
preferable since it is outside the hot path.

Reference: http://bugzilla.netfilter.org/show_bug.cgi?id=621
Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>

netfilter: xtables: do not grab random bytes at __init

"It is deliberately not done in the init function, since we might not
have sufficient random while booting."

Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>

netfilter: xt_recent: save 8 bytes per htable

Moving rnd_inited into the hole after the uint8 lets go of the uint32
rnd_inited was using, plus the padding that would follow the int group.

Signed-off-by: Jan Engelhardt <jengelh@medozas.de>
Signed-off-by: Patrick McHardy <kaber@trash.net>

netfilter: SNMP NAT: correct the size argument to kzalloc

obj has type struct snmp_object **, not struct snmp_object *.  But indeed
it is not even clear why kmalloc is needed.  The memory is freed by the end
of the function, so the local variable of pointer type should be sufficient.

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@disable sizeof_type_expr@
type T;
T **x;
@@

  x =
  <+...sizeof(
- T
+ *x
  )...+>
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: Patrick McHardy <kaber@trash.net>

axnet_cs: remove unnecessary spin_unlock_irqrestore

axnet_cs:
    remove unnecessary spin_unlock_irqrestore,spin_lock_irqsave.

Signed-off-by: Ken Kawasaki <ken_kawasaki@spring.nifty.jp>
Signed-off-by: David S. Miller <davem@davemloft.net>

tipc: use kconfig to limit numeric ranges

We can rely on kconfig to limit these numbers,
no need to limit them at compile time/run time.

Users who modify these numbers manually should
be responsible for themself. :)

Signed-off-by: WANG Cong <amwang@redhat.com>
Cc: Per Liden <per.liden@ericsson.com>
Cc: Jon Maloy <jon.maloy@ericsson.com>
Cc: Allan Stephens <allan.stephens@windriver.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

can/netlink: add CAN_CTRLMODE_ONE_SHOT

This patch adds the flag CAN_CTRLMODE_ONE_SHOT. It is used as mask
or flag in the "struct can_ctrlmode".

It allows userspace via netlink to set a CAN controller into the special
"one-shot" mode. In this mode, if supported by the CAN controller, it
tries only once to deliver a CAN frame and aborts it if an error
(e.g.: arbitration lost) happens.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Wolfgang Grandegger <wg@grandegger.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

can: Speed up CAN frame receiption by using ml_priv

this patch removes the hlist that contains the CAN receiver filter lists.
It uses the 'midlayer private' pointer ml_priv and links the filters directly
to the CAN netdevice, which allows to omit the walk through the complete CAN
devices hlist for each received CAN frame.

This patch is tested and does not remove any locking.

Signed-off-by: Oliver Hartkopp <oliver@hartkopp.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/cxgb3: Use kzalloc for allocating only one thing

Use kzalloc rather than kcalloc(1,...)

The semantic patch that makes this change is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
@@

- kcalloc(1,
+ kzalloc(
          ...)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Acked-by: Divy Le Ray <divy@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

bonding: allow arp_ip_targets on separate vlans to use arp validation

This allows a bond device to specify an arp_ip_target as a host that is
not on the same vlan as the base bond device and still use arp
validation.  A configuration like this, now works:

BONDING_OPTS="mode=active-backup arp_interval=1000 arp_ip_target=10.0.100.1 arp_validate=3"

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    inet6 ::1/128 scope host
       valid_lft forever preferred_lft forever
2: eth1: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond0 qlen 1000
    link/ether 00:13:21:be:33:e9 brd ff:ff:ff:ff:ff:ff
3: eth0: <BROADCAST,MULTICAST,SLAVE,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast master bond0 qlen 1000
    link/ether 00:13:21:be:33:e9 brd ff:ff:ff:ff:ff:ff
8: bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue
    link/ether 00:13:21:be:33:e9 brd ff:ff:ff:ff:ff:ff
    inet6 fe80::213:21ff:febe:33e9/64 scope link
       valid_lft forever preferred_lft forever
9: bond0.100@bond0: <BROADCAST,MULTICAST,MASTER,UP,LOWER_UP> mtu 1500 qdisc noqueue
    link/ether 00:13:21:be:33:e9 brd ff:ff:ff:ff:ff:ff
    inet 10.0.100.2/24 brd 10.0.100.255 scope global bond0.100
    inet6 fe80::213:21ff:febe:33e9/64 scope link
       valid_lft forever preferred_lft forever

Ethernet Channel Bonding Driver: v3.6.0 (September 26, 2009)

Bonding Mode: fault-tolerance (active-backup)
Primary Slave: None
Currently Active Slave: eth1
MII Status: up
MII Polling Interval (ms): 0
Up Delay (ms): 0
Down Delay (ms): 0
ARP Polling Interval (ms): 1000
ARP IP target/s (n.n.n.n form): 10.0.100.1

Slave Interface: eth1
MII Status: up
Link Failure Count: 1
Permanent HW addr: 00:40:05:30:ff:30

Slave Interface: eth0
MII Status: up
Link Failure Count: 0
Permanent HW addr: 00:13:21:be:33:e9

Signed-off-by: Andy Gospodarek <andy@greyhouse.net>
Signed-off-by: Jay Vosburgh <fubar@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'master' of git://git./linux/kernel/git/linville/wireless-next-2.6

Merge git://git./linux/kernel/git/linville/wireless-2.6

Conflicts:
drivers/net/wireless/libertas/scan.c

Subject: drivers/net/sh_eth.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/r8169.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/octeon/octeon_mgmt.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/smc911x.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/sunvnet.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/usb/catc.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/s390/net/qeth_l2_main.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/xilinx_emaclite.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/via-velocity.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/usb/rtl8150.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/igbvf/netdev.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/lib82596.c: use %pM to shown MAC address

Use the %pM kernel extension to display the MAC address.

The only difference in the output is that the MAC address is
shown in the usual colon-separated hex notation instead of
space-separated.

Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Revert "b43: Enforce DMA descriptor memory constraints"

This reverts commit 9bd568a50c446433038dec2a5186c5c57c3dbd23.

That commit is shown to cause allocation failures during initialization
on some machines.

http://bugzilla.kernel.org/show_bug.cgi?id=14844

Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: annotate sleeping driver ops

To make it easier to notice cases of calling sleeping ops in atomic context,
annotate driver-ops.h with appropiate might_sleep() calls. At the same time,
also document in mac80211.h the op functions with missing contexts.

mac80211 doesn't seem to use get_tx_stats anywhere currently. Just to be on
the safe side, I documented it to be atomic, but hopefully the op can be
removed in the future.

Compile-tested only.

Signed-off-by: Kalle Valo <kalle.valo@iki.fi>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: remove requeue from work

There's no need to be requeueing the work struct
since we check for the scan after removing items
due to possible timeouts.

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

mac80211: remove struct ieee80211_if_init_conf

All its members (vif, mac_addr, type) are now available
in the vif struct directly, so we can pass that instead
of the conf struct. I generated this patch (except the
mac80211 and header file changes) with this semantic
patch:

@@
identifier conf, fn, hw;
type tp;
@@
tp fn(struct ieee80211_hw *hw,
-struct ieee80211_if_init_conf *conf)
+struct ieee80211_vif *vif)
{
<...
(
-conf->type
+vif->type
|
-conf->mac_addr
+vif->addr
|
-conf->vif
+vif
)
...>
}

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

mac80211/cfg80211: add station events

When, for instance, a new IBSS peer is found, userspace
wants to be notified. Add events for all new stations
that mac80211 learns about.

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

mac80211: make off-channel work generic

This changes mac80211 to allow being off-channel for
any type of work, not just the 'remain-on-channel'
work. This also helps fast transition to a BSS on a
different channel.

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

mac80211: support remain-on-channel command

This implements the new remain-on-channel cfg80211
command in mac80211, extending the work interface.

Also change the work purge code to be able to clean
up events properly (pretending they timed out.)

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

cfg80211: add remain-on-channel command

Add new commands for requesting the driver to remain awake
on a specified channel for the specified amount of time
(and another command to cancel such an operation). This
can be used to implement userspace-controlled off-channel
operations, like Public Action frame exchange on another
channel than the operation channel.

The off-channel operation should behave similarly to scan,
i.e. the local station (if associated) moves into power
save mode to request the AP to buffer frames for it and
then moves to the other channel to allow the off-channel
operation to be completed. The duration parameter can be
used to request enough time to receive a response from
the target station.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: Generalize off-channel operation helpers from scan code

The off-channel operations for going into power save mode (station
mode) or stop beaconing (AP/IBSS) are not limited to scanning. Move
these into a separate file and allow them to be used for other
purposes, too.

Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: proper bss private data handling

cfg80211 offers private data for each BSS struct,
which mac80211 uses. However, mac80211 uses internal
and external (cfg80211) BSS pointers interchangeably
and has a hack to put the cfg80211 bss struct into
the private struct.

Remove this hack, properly converting between the
pointers wherever necessary.

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

mac80211: split up and insert custom IEs correctly

Currently, we insert all user-specified IEs before the HT
IE for association, and after the HT IE for probe requests.
For association, that's correct only if the user-specified
IEs are RSN only, incorrect in all other cases including
WPA. Change this to split apart the user-specified IEs in
two places for association: before the HT IE (e.g. RSN),
after the HT IE (generally empty right now I think?) and
after WMM (all other vendor-specific IEs). For probes,
split the IEs in different places to be correct according
to the spec.

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

mac80211: refactor association

Refactor the code to reserve an skb of the right size
(instead of hoping 200 bytes are enough forever), and
also put HT IE generation into an own function.

Additionally, put the HT IE before the vendor-specific
WMM IE. This still leaves things not quite ordered
correctly, due to user-specified IEs, add a note about
that for now.

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

mac80211: rewrite a few work messages

The station we're authenticating/associating with
may not always be an AP in the sense that word is
mostly understood, so print only the MAC address
of the peer instead.

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

mac80211: generalise work handling

In order to use auth/assoc for different purposes
other than MLME, it needs to be split up. For other
purposes, a generic work handling (potentially on
another channel) will be useful.

To achieve that, this patch moves much of the MLME
work handling out of mlme into a new work API. The
API can currently handle probing a specific AP,
authentication and association. The MLME previously
handled probe/authentication as one step and will
continue to do so, but they are separate in the new
work handling.

Work items are RCU-managed to be able to check for
existence of an item for a specific frame in the RX
path, but they can be re-used which the MLME right
now will do for its combined probe/auth step.

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

mac80211: generalise management work a bit

As a first step of generalising management work,
this renames a few things and puts more information
directly into the struct so that auth/assoc need
not access the BSS pointer as often -- in fact it
can be removed from auth completely. Also since the
previous patch made sure a new work item is used
for association, we can make the different data a
union.

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

mac80211: let cfg80211 manage auth state

mac80211 currently hangs on to the auth state by
keeping it on the work list. That can lead to
confusing behaviour like rejecting scans while
authenticated to any AP (but not yet associated.)
It also means that it needs to keep track of the
work struct while associated for when it gets
disassociated (or disassociates.)

Change this to free the work struct after the
authentication completed successfully and
allocate a new one for associating, thereby
letting cfg80211 manage the auth state. Another
change necessary for this is to tell cfg80211
about all unicast deauth frames sent to mac80211
since now it can no longer check the auth state,
but that check was racy anyway.

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

mac80211: introduce flush operation

We've long lacked a good confirmation that frames
have really gone out, e.g. before going off-channel
for a scan. Add a flush() operation that drivers
can implement to provide that confirmation, and use
it in a few places:
 * before scanning sends the nullfunc frames
 * after scanning sends the nullfunc frames, if any
 * when going idle, to send any pending frames

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

mac80211: add ieee80211_sdata_running

Instead of always using netif_running(sdata->dev)
use ieee80211_sdata_running(sdata) now which is
just an inline containing netif_running() for now.

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

rtl8187: remove priv->mode

It is checked in add_interface, but there it is easily replaced with
a check of priv->vif.  If that information should become necessary,
it is available in vif->type anyway.

It is also checked in led_turn_on and led_turn_off, where I made the
substitutions as described above.  Of course, these checks seem to
have been incorrect since the driver was using NL80211_IFTYPE_MONITOR
to indicate no interface rather than NL80211_IFTYPE_UNSPECIFIED.
Anyway, I think these checks may be extraneous...?

Signed-off-by: John W. Linville <linville@tuxdriver.com>

rtl8180: remove priv->mode

It is only checked in add_interface, and there it is easily replaced
with a check of priv->vif.  If that information should become necessary,
it is available in vif->type anyway.

Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Cleanup chip handling helper functions.

Let each of them take a struct rt2x00_dev pointer as argument instead of
a mixture of struct rt2x00_chip and struct rt2x00_dev pointers.
Preparation for further clean ups in the rt2x00 chip handling, especially
for rt2800 devices.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: convert RT2800PCI_PCI and RT2800PCI_SOC Kconfig symbols to booleans.

There is no need for Kconfig symbols RT2800PCI_PCI and RT2800PCI_SOC to be
tristates, as they are only used to check whether RT2800 PCI or SOC support
is to be compiled in.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Let rt2800lib check CONFIG_RT2X00_LIB_USB instead of CONFIG_RT2800USB

rt2800lib currently checks whether RT2800USB is enabled in the configuration.
Strictly speaking this is not necessary, it only needs to know whether the
generic rt2x00usb library functions are available. Therefore check for
RT2X00_LIB_USB instead.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2x00: Fix checks for rt2800 SOC support.

Fix checking for SOC support in rt2800pci. The wrong config (an unexisting
one) was checked.

Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wireless: remove CONFIG_WIRELESS_OLD_REGULATORY

This is no longer needed with the availability of
CONFIG_CFG80211_INTERNAL_REGDB.

Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Prevent performing "join" before association

There is a minor bug in the code causing a "join" to be performed before
there is an intention to associate. Fix this.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Check vif for NULL when indicating beacon-loss

Because the interface is started and the vif are created and destroyed
separately, there is a slim possibility beacon-loss indications occur while
there is no vif - causing a kernel-oops unless checked.

Add checking for the vif.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Add support for acx_pm_config

This acx configures host clock parameters in correspondence with the clock
request line - the settling time of the clock, and whether fast wake-up is
supported.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Change booleans in struct wl1271 into a flags bitmask

For cleaner implementation, change the bunch of booleans in the struct wl1271
structure into a flags bitmask.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Fix supported rate management

Previously, only basic rates were used for data transmission - resulting in
reduced transfer rates. This patch takes enables the firmware to take advantage
of the full set of data rates supported by the AP.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Change rates configured for templates

Previously a "firmware chooses" value was used for the rates of all control
message templates set to the firmware. This resulted in a too high rate to be
chosen to transmit those messages. Change this by configuring a fixed low rate
for the templates.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Add rudimentary ad-hoc support

This patch adds rudimentary a-hoc support for the driver. It will allow
setting up ad-hoc, and for other devices to scan and join. The beacon and probe
response template setting is slightly dirty, and need to be properly
implemented with support from mac80211. Also, the SSID is not configured to the
firmware - the FW will not be able to respond to probe requests autonomously.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Remove beacon-loss-ind from PSM entry failure handling

Remove the beacon-loss indication to stack from PSM entry failure handling -
this will cause more problems than it will solve due to the design of the
mac80211.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Add pre-power-on sleep

This patch adds a short delay before powering on the wl1271. Normally, it is
not needed, but if the wl1271 has been powered off shortly before, for reliable
firmware-booting this small stabilization delay is required.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Implement chipset boot retry

The wl1271 has a hardware bug which causes it to randomly (very infrequently)
to fail powering up properly. This patch implements retry for the chipset boot
sequence, to work around the hardware problem.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: prevent power save entry while not associated

The mac80211 sometimes requests power save entry while not associated - this
will cause problems, so prevent it if not associated. Go to powersave once
association is complete.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Fix event acknowledging functionality

In reference source, events are acknowledged separately - fix the driver to
do the same.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Use slow rates for association messages

While not associated, default the data rates to 1 and 2mbps, so that only
those rates will be used for association related message transfer. Once
associated, configure the full rate-set supported by the AP.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Remove smart reflex ACX

Remove the smart-reflex ACX - the associated parameters are now configured
in the radio parameters config.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: Configure smart-reflex paramter values.

Configure correct values to be used with the smart-reflex configuration of the
firmware.

Signed-off-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: use channel 1 when configuring the data path

In the data path configuration, one of the parameters is the channel.  We
have been setting it to wl->channel, which is not correct in this case.  This
channel has nothing to do with the channel we're currently tuned to, since it
is only used for calibration during this phase.  Hardcoded the channel to 1,
according to the reference driver.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: some new configuration values according to new reference

In the new reference driver, some of the firmware configuration values have
been changed.  This patch changes them in the wl1271 driver accordingly.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: upload only the first 468 bytes from the NVS file

We were uploading the whole NVS file, but that is wrong, because the same
file also contains the initialization values.  For the latest firmwares, we
should upload only the initial 468 bytes from the file.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: add gpio_power file in debugfs to power the chip on and off

Some debugging tools require the chip to be powered on before they can work.
With these tools, we shouldn't upload the firmware nor boot the firmware
ourselves, so this debufs file is provided.  It always contains the gpio
power setting (0 = off, 1 = on).  To change the power setting, just write 0
or 1 to the file.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Reviewed-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: fix one typo in the rx_rssi_and_proc_compens values

There was a typo in one of the values in the rx_rssi_and_proc_compens elemt
of the Radio Parameters struct.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: implement dco itrim parameters setting

Newer firmwares require the dco itrim parameters to be set during
initialization.  This patch implements the new ACX function and calls it.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: use the correct macro when setting the basic rates

We were using CONF_TX_RATE_MASK_ALL when calling wl1271_acx_rate_policies()
during init.  We should use WL1271_DEFAULT_BASIC_RATE_SET instead.  The
values are the same, but the latter is just the correct macro to use.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: set null data template when BSSID is known

The call to wl1271_cmd_build_null_data() was missing when we got associated,
this was causing PS to fail.  This patch adds the call and now PS seems to
work.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: check result code from the join command

We were not checking the return value from the call to wl1271_cmd_join().
Added a check to make things more reliable.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: limit TX power to 25dBm for every channel

The wl1271 firmware supports maximun 25.5dBm, so the driver was returning
-EINVALID to anything above that.  This patch uses the channel max_power
option to limit the TX power to 25dBm.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: remove workaround for disconnection

Now we're using a the idle information coming from mac80211 to decide when to
disconnect.  If we have joined (ie. we're listening to a channel), whenever
the interface goes to idle, we will issue a disconnect command.  So the
workaround to send a disconnect command before joining is not needed anymore.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: use join command with dummy BSSID

When we need to change the channel before association, we have to send a join
command with a valid BSSID.  With this patch we use 0baddeadbeef as the
BSSID.  There are ongoing discussions with TI to get this done in a cleaner
way.

When we go back to idle, we issue a CMD_DISCONNECT to make sure the firmware
stops listening to the channel and cleans things up internally.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: added radio parameters configuration values newer firmwares

Add new radio parameters for new structures based on firmware revision
6.1.0.0.288.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: update radio and general parameters values

There were some changes in the values we have to use for these settings.  This
patches updates them.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: updated general parameters structure for newer firmwares

In revision 6.1.0.0.288 the general parameters structure has changed.  This
patch updates the driver code accordingly.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1271: updated radio parameters structure for newer firmwares

In revision 6.1.0.0.288 the radio parameters structure has changed.  This
patch updates the driver code accordingly.

Signed-off-by: Luciano Coelho <luciano.coelho@nokia.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

b43: Allow PIO mode to be selected at module load

If userencounter the "Fatal DMA Problem" with a BCM43XX device, and
still wish to use b43 as the driver, their only option is to rebuild
the kernel with CONFIG_B43_FORCE_PIO. This patch removes this option and
allows PIO mode to be selected with a load-time parameter for the module.
Note that the configuration variable CONFIG_B43_PIO is also removed.

Once the DMA problem with the BCM4312 devices is solved, this patch will
likely be reverted.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Tested-by: John Daiker <daikerjohn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: remove duplicate local per-vif copy of ieee80211_bss_conf

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: remove unused mwl8k_vif::priv

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: fix up AP vs. STA firmware image receive descriptor handling

The receive descriptor ops that are currently marked as being for
8687 only are actually used for all STA firmware images, whereas the
receive descriptor ops marked as 8366 are only used on 8366 when an
AP firmware image is in use.

Rename the receive descriptor ops to reflect this, use the STA ops
unconditionally if the firmware image loaded reported the STA ready
code, and rename the mwl8k_device_info::rxd_ops member to ap_rxd_ops
to indicate that it should only be used if we are running on AP
firmware.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: do rx/tx ring initialisation after loading firmware

Whether the firmware we have loaded is AP or STA firmware decides
which receive descriptor format we have to use.  Therefore, move
rx/tx ring initialisation to be after firmware loading.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: get rid of the struct mwl8k_firmware abstraction

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: add 2.4GHz channels 12, 13 and 14

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: initialize the mwl8k_info_tbl table using the MWL* enums

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: inline qos field manipulation functions

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: get rid of the AMSDU check in the transmit path

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: hw is never NULL in mwl8k_set_radio_preamble()

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mwl8k: firmware command code cleanup

Sort firmware commands by command code, get rid of the 802_11 substring
in all command names, and make sure that the command functions match the
firmware command names.

Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

iwmc3200wifi: fix array out-of-boundary access

Allocate priv->rx_packets[IWM_RX_ID_HASH + 1] because the max array
index is IWM_RX_ID_HASH according to IWM_RX_ID_GET_HASH().

Cc: stable@kernel.org
Signed-off-by: Zhu Yi <yi.zhu@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

wl1251: timeout one too soon in wl1251_boot_run_firmware()

`loop' reaches INIT_LOOP + 1 after the loop. so if ACX_INTR_INIT_COMPLETE
occurs in the last iteration the write occurs but also the error out as if a
timeout occurred. This is probably very unlikely to ever occur.

Signed-off-by: Roel Kluin <roel.kluin@gmail.com>
Acked-by: Kalle Valo <kalle.valo@nokia.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: fix propagation of failed hardware reconfigurations

mac80211 does not propagate failed hardware reconfiguration
requests. For suspend and resume this is important due to all
the possible issues that can come out of the suspend <-> resume
cycle. Not propagating the error means cfg80211 will assume
the resume for the device went through fine and mac80211 will
continue on trying to poke at the hardware, enable timers,
queue work, and so on for a device which is completley
unfunctional.

The least we can do is to propagate device start issues and
warn when this occurs upon resume. A side effect of this patch
is we also now propagate the start errors upon harware
reconfigurations (non-suspend), but this should also be desirable
anyway, there is not point in continuing to reconfigure a
device if mac80211 was unable to start the device.

For further details refer to the thread:

http://marc.info/?t=126151038700001&r=1&w=2

Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

mac80211: fix race with suspend and dynamic_ps_disable_work

When mac80211 suspends it calls a driver's suspend callback
as a last step and after that the driver assumes no calls will
be made to it until we resume and its start callback is kicked.
If such calls are made, however, suspend can end up throwing
hardware in an unexpected state and making the device unusable
upon resume.

Fix this by preventing mac80211 to schedule dynamic_ps_disable_work
by checking for when mac80211 starts to suspend and starts
quiescing. Frames should be allowed to go through though as
that is part of the quiescing steps and we do not flush the
mac80211 workqueue since it was already done towards the
beginning of suspend cycle.

The other mac80211 issue will be hanled in the next patch.

For further details see refer to the thread:

http://marc.info/?t=126144866100001&r=1&w=2

Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: fix missed error codes in the tx status check

My previous change added in:

 commit 815833e7ecf0b9a017315cae6aef4d7cd9517681
    ath9k: fix tx status reporting

was not checking all possible tx error conditions. This could possibly
lead to throughput issues due to slow rate control adaption or missed
retransmissions of failed A-MPDU frames.

This patch adds a mask for all possible error conditions and uses it
in the xmit ok check.

Cc: stable@kernel.org
Reported-by: Björn Smedman <bjorn.smedman@venatech.se>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: wake hardware during AMPDU TX actions

AMDPDU actions poke hardware for TX operation, as such
we want to turn hardware on for these actions. AMDPU RX operations
do not require hardware on as nothing is done in hardware for
those actions. Without this we cannot guarantee hardware has
been programmed correctly for each AMPDU TX action.

Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

ath9k: wake hardware for interface IBSS/AP/Mesh removal

When we remove a IBSS/AP/Mesh interface we stop DMA
but to do this we should ensure hardware is on. Awaken
the device prior to these calls. This should ensure
DMA is stopped upon suspend and plain device removal.

Cc: stable@kernel.org
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>