The TXD_W0_CIPHER field is a 1-bit field. It only acts as boolean value
to indicate if the frame must be encrypted or not.

The way rt2x00_set_field32() worked it would grab the least signifcant bit
from txdesc->cipher and use that as value. Because of that WEP 64 and TKIP
worked since they had odd-numbered values, while WEP 128 and AES were
even numbers and didn't work.

Correctly booleanize the txdecs->cipher value to allow the hardware to
encrypt the outgoing data. After this we can enable HW crypto by default again.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This patch fixes a panic that might occur, if the device is part of a mesh
and tries to send with a higher rate index than "0".

kernel BUG at net/mac80211/rate.c:239!
invalid opcode: 0000 [#1] SMP
[...]
 Call Trace:
 <IRQ> <0> ? invoke_tx_handlers+0x474/0xb57 [mac80211]
 ? __ieee80211_tx_prepare+0x260/0x2a8 [mac80211]
 ? ieee80211_master_start_xmit+0x300/0x43a [mac80211]
 ? __qdisc_run+0xde/0x1da
 ? net_tx_action+0xb4/0x102
 Kernel panic - not syncing: Fatal exception in interrupt
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Should return NETDEV_TX_{OK,BUSY} instead of 0,-1 (this doesn't change
any current functionality).

Changes-licensed-under: 3-Clause-BSD
Reported-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Bob Copeland <me@bobcopeland.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Due to misunderstanding of the returned values allowed for the tx callback
of mac80211, rtl8187 was using skb's that had been freed. This problem was
triggered when the module was sujected to a rmmod/insmod cycle.

After that was fixed, the modules would not work after the rmmod/insmod cycle
until the USB device was reset.
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This patch hopefully fixes a mac80211<->p54 interaction problem, which was
described by Larry Finger (ref: http://marc.info/?l=linux-wireless&m=123009889327707 )

I guess the warning was triggered by pending frames in the receive queue,
while we're doing a band change 5GHz.
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Changing mode on an interface is not allowed if IBSS is disabled for the
current channel.  That restriction should only apply when switching to
the ad-hoc mode, as it was prior to "cfg80211: handle SIOCGIWMODE/SIOCSIWMODE".
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This corrects usage of AR5K_CFG_ADHOC introduced in
"ath5k: Update PCU code". Also,
the name of the indicator is changed to AR5K_CFG_IBSS to more
accurately reflect its function. This change restores
beaconing in AP and mesh modes.
Signed-off-by: Steve Brown <sbrown@cortland.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This patch reverts "ath9k: Fix TX status reporting for retries and MCS index"
because that change ended up breaking ath9k rate control. While the
MCS index reporting to mac80211 was indeed fixed by the patch, it did
not take into account that the ath9k rate control algorithm was
updating private tables based on this index and the index comes
through the rate control API call, i.e., based on mac80211 TX status
call. In addition, it looks like the "fix" to remove +1 from TX status
'count' field was not correct based on ieee80211_tx_status()
implementation that counts the total of count values, but starting
from -1, not 0.

The TX status reporting for frames using MCS needs to be fixed
somehow, but it does not look like there is any easy fix for the ath9k
rate control algorithm, so the best option now seems to be to revert the
change and bring it back once the rate control code is cleaned up to
handle this better.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This patch fixes the following WARNING (caused by rix_to_ndx): "
>WARNING: at net/mac80211/rc80211_minstrel.c:69 minstrel_rate_init+0xd2/0x33a [mac80211]()
>[...]
>Call Trace:
> warn_on_slowpath+0x51/0x75
> _format_mac_addr+0x4c/0x88
> minstrel_rate_init+0xd2/0x33a [mac80211]
> print_mac+0x16/0x1b
> schedule_hrtimeout_range+0xdc/0x107
> ieee80211_add_station+0x158/0x1bd [mac80211]
> nl80211_new_station+0x1b3/0x20b [cfg80211]

The reason is that I'm experimenting with "g" only mode on a 802.11 b/g card.

Therefore rate_lowest_index returns 4 (= 6Mbit, instead of usual 0 = 1Mbit).
Since mi->r array is initialized with zeros in minstrel_alloc_sta,
rix_to_ndx has a hard time to find the 6Mbit entry and will trigged the WARNING.
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This patch fixes an issue with the sequence numbers of unaggregated
QoS frames, because of which the frames are handled in a different order
at the AP and resulted in MLME REPLAYFAILURE.
Signed-off-by: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

rt2500usb doesn't strip the IV/ICV data from received frames,
so we don't need to set the RX_FLAG_IV_STRIPPED flag.
We do need to set the RX_FLAG_MMIC_STRIPPED flag for all
encryption types since the MMIC has been removed from the frame.

After this patch TKIP Hardware crypto works for rt2500usb.
WEP and AES are still failing.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

To be compatible with mac80211 following "mac80211: only create
default STA interface if supported", rtl8180 needs to set
NL80211_IFTYPE_STATION in interface_modes.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Reported-by: Fabio Rossi <rossi.f@inwind.it>
Tested-by: Piter PUNK <piterpunk@unitednerds.org>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

My first (minor) patch, hopefully this is correct.

Fix a typo in iwl-commands.h for CCK rates which needs 7 bits and not 4.
Signed-off-by: Daniel Wu <dyqith@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

The queue_end() macro points to 1 position after the
queue, which means that if we want to know if queue
is at the end of the queue we should first increment
the position and then check if it is a valid entry.

This fixes a segmentation fault which only occurs when
the device has enough endpoints to provide a dedicated
endpoint for all TX queues (which likely won't happen
for rt2500usb and rt73usb, but will happen for rt2800usb).
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

All LM87 firmwares need a explicit termination "packet",
in oder to finish the pending transfer properly.
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

This patch fixes two small flaws:
 - restore the original TKIP IV if we altered it.
 - reserve & initialize ICV with zeros.
   This is actually only necessary for some obsolete p54usb firmwares.
   But we don't know yet, if all devices are compatible with the new revisions.
Signed-off-by: Christian Lamparter <chunkeey@web.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Reported by Michael Jarosch <mitsch@riotmusic.de>
Signed-off-by: Pavel Roskin <proski@gnu.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

0x1b is a 3945 specific command, we should print it too when debugging.
Signed-off-by: Samuel Ortiz <samuel.ortiz@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

The chainmasks have to be updated before setting the channel,
since the HW reset routine uses them to set the appropriate registers.
Signed-off-by: Sujith <Sujith.Manoharan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Fix the warning reproduced below.

We add to rx_list in interrupt context and remove elements in tasklet
context. While removing elements we need to prevent the interrupt
modifying the list.

Note that "orinoco: Process bulk of receive interrupt in a tasklet" did not
preserve locking semantics on what is now orinoco_rx.

This patch reinstates the locking semantics and ensures it covers
rx_list as well. This leads to additional cleanup required in
free_orinocodev.

[89479.105038] WARNING: at lib/list_debug.c:30 __list_add+0x8f/0xa0()
[89479.105058] list_add corruption. prev->next should be next (dddb3568), but was cbc28978. (prev=dddb3568).
[89479.106002] Pid: 15746, comm: X Not tainted 2.6.28-1avb #26
[89479.106020] Call Trace:
[89479.106062]  [<c011d3b0>] warn_slowpath+0x60/0x80
[89479.106104]  [<c01073d0>] ? native_sched_clock+0x20/0x70
[89479.106194]  [<c013d825>] ? lock_release_holdtime+0x35/0x200
[89479.106218]  [<c018d9f0>] ? __slab_alloc+0x550/0x560
[89479.106254]  [<c02f9c9d>] ? _spin_unlock+0x1d/0x20
[89479.106270]  [<c018d9f0>] ? __slab_alloc+0x550/0x560
[89479.106302]  [<c01ff2a7>] ? delay_tsc+0x17/0x24
[89479.106319]  [<c01ff221>] ? __const_udelay+0x21/0x30
[89479.106376]  [<dfa8b1e2>] ? hermes_bap_seek+0x112/0x1e0 [hermes]
[89479.106396]  [<c013d7eb>] ? trace_hardirqs_off+0xb/0x10
[89479.106418]  [<c018e307>] ? __kmalloc_track_caller+0xb7/0x110
[89479.106448]  [<c028eefc>] ? dev_alloc_skb+0x1c/0x30
[89479.106465]  [<c028eefc>] ? dev_alloc_skb+0x1c/0x30
[89479.106482]  [<c020e13f>] __list_add+0x8f/0xa0
[89479.106551]  [<dfd0fcae>] orinoco_interrupt+0xcae/0x16c0 [orinoco]
[89479.106574]  [<c013b0e3>] ? tick_dev_program_event+0x33/0xb0
[89479.106594]  [<c01073d0>] ? native_sched_clock+0x20/0x70
[89479.106613]  [<c013d825>] ? lock_release_holdtime+0x35/0x200
[89479.106662]  [<c013d7eb>] ? trace_hardirqs_off+0xb/0x10
[89479.106892]  [<dfe7faa7>] ? usb_hcd_irq+0x97/0xa0 [usbcore]
[89479.106926]  [<c015ba79>] handle_IRQ_event+0x29/0x60
[89479.106947]  [<c015cf89>] handle_level_irq+0x69/0xe0
[89479.106963]  [<c015cf20>] ? handle_level_irq+0x0/0xe0
[89479.106977]  <IRQ>  [<c02ca933>] ? tcp_v4_rcv+0x633/0x6e0
[89479.107025]  [<c0103f0c>] ? common_interrupt+0x28/0x30
[89479.107057]  [<c02a0000>] ? sk_run_filter+0x320/0x7a0
[89479.107078]  [<c020e041>] ? list_del+0x21/0x90
[89479.107106]  [<dfd0d24e>] ? orinoco_rx_isr_tasklet+0x2ce/0x480 [orinoco]
[89479.107131]  [<c01402e0>] ? __lock_acquire+0x160/0x1650
[89479.107151]  [<c01073d0>] ? native_sched_clock+0x20/0x70
[89479.107169]  [<c013d825>] ? lock_release_holdtime+0x35/0x200
[89479.107200]  [<c012249a>] ? irq_enter+0xa/0x60
[89479.107217]  [<c0104e52>] ? do_IRQ+0xd2/0x130
[89479.107518]  [<c010342c>] ? restore_nocheck_notrace+0x0/0xe
[89479.107542]  [<c0122830>] ? __do_softirq+0x0/0x110
[89479.107561]  [<c013f7b4>] ? trace_hardirqs_on_caller+0x74/0x140
[89479.107583]  [<c01ff678>] ? trace_hardirqs_on_thunk+0xc/0x10
[89479.107602]  [<c0122087>] ? tasklet_action+0x27/0x90
[89479.107620]  [<c013f7b4>] ? trace_hardirqs_on_caller+0x74/0x140
[89479.107638]  [<c01220a3>] ? tasklet_action+0x43/0x90
[89479.107655]  [<c012289f>] ? __do_softirq+0x6f/0x110
[89479.107674]  [<c0122830>] ? __do_softirq+0x0/0x110
[89479.107685]  <IRQ>  [<c015cf20>] ? handle_level_irq+0x0/0xe0
[89479.107715]  [<c012246d>] ? irq_exit+0x5d/0x80
[89479.107732]  [<c0104e52>] ? do_IRQ+0xd2/0x130
[89479.107747]  [<c0103337>] ? sysenter_exit+0xf/0x16
[89479.107765]  [<c013f83d>] ? trace_hardirqs_on_caller+0xfd/0x140
[89479.107782]  [<c0103f0c>] ? common_interrupt+0x28/0x30
[89479.107797] ---[ end trace a1fc0a52df4a729d ]---
Reported-by: Andrey Borzenkov <arvidjaar@mail.ru>
Signed-off-by: David Kilroy <kilroyd@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

CONFIG_ATH9K=y results in build issues if CONFIG_RFKILL=m since ath9k
does not depend on rfkill in kconfig (i.e., CONFIG_RFKILL is used to
select whether to enable rfkill in ath9k), but uses its functions if
rfkill is enabled. Enforce ath9k to be build as a module if
CONFIG_RFKILL=m to avoid this invalid configuration.
Signed-off-by: Jouni Malinen <jouni.malinen@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Add the USB ID for Thomson Speedtouch 121g to p54usb.
Signed-off-by: Michiel <michiel@ettema.net>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

The short preamble mode was not correctly detected during TX,
rt2x00 used the rate->hw_value_short field but mac80211 is not
using this field that way.
Instead the flag IEEE80211_TX_RC_USE_SHORT_PREAMBLE should be
used to determine if the frame should be send out using
short preamble or not.
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

add USB ID for the Linksys WUSB200 Wireless-G Business USB Adapter to
rt73usb.
Signed-off-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Cc: stable <stable@kernel.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

On x86_64 architecture with 4K page size and SLUB debugging enabled, stress
testing on p54usb has resulted in skb allocation failures of O(1) and extreme
page fragmentation. Reducing rx_mtu fixes this problem by reducing the size of
all receive skb allocations to be of O(0). This change does not impact
performance in any way.
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Since "rt2x00: Fix LED state handling", rt2x00leds_led_radio wrongly
checks that the LED type is LED_TYPE_ASSOC. This patch makes it check
for LED_TYPE_RADIO once again.
Signed-off-by: Andrew Price <andy@andrewprice.me.uk>
Acked-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

Raytheon cards use 2.4 GHz, not 2.4 MHz.
See http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/Linux.Wireless.drivers.html#RaylinkSigned-off-by: Erik Ekman <erik@kryo.se>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

The UCC Event Register (UCCE) already has unambigous macro definitions in qe.h,
so we should not be defining our own in the UCC Ethernet driver.

Removed unused local variable 'dev' from ucc_geth_poll(), which fixes
a warning caused by commit 908a7a16
("net: Remove unused netdev arg from some NAPI interfaces.").

Replaced in_be/out_be pairs with setbits32 or clrbits32, where applicable.
Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

The recent dmaengine rework removed the capability to remove dma device
driver modules while net_dma is active.  Rather than notify
dmaengine-clients that channels are trying to be removed, we now rely on
clients to notify dmaengine when they no longer have a need for
channels.  Teach net_dma to release channels by taking dmaengine
references at netdevice open and dropping references at netdevice close.
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

I have a system with a Chelsio adapter (driven by cxgb3) whose ports are
part of a Linux bridge.  Recently I updated the kernel and discovered
that things stopped working because cxgb3 was doing LRO on packets that
were passed into the bridge code for forwarding.  (Incidentally, this
problem manifested itself in a strange way that made debugging a bit
interesting -- for some reason, the skb_warn_if_lro() check in bridge
didn't trigger and these LROed packets were forwarded out a forcedeth
interface, and caused the forcedeth transmit path to get stuck)

This is because cxgb3 has no way of keeping state for the LRO flag until
the interface is brought up, so if the bridging code disables LRO while
the interface is down, then cxgb3_up() will just reenable LRO, and on my
Debian system at least, the init scripts add interfaces to a bridge
before bringing the interfaces up.

Fix this by keeping track of each interface's LRO state in cxgb3 so that
when bridge disables LRO, it stays disabled in cxgb3_up() when the
interface is brought up.  I did this by changing the rx_csum_offload
flag into a pair of bit flags; the effect of this on the rx_eth() fast
path is miniscule enough that it should be fine (eg on x86, a cmpb
instruction becomes a testb instruction).
Signed-off-by: Roland Dreier <rolandd@cisco.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Only the SFX7101 requires software power control.  This was
incorrectly being applied to the SFT9001 rev A as well.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Thanks to Wei and Arnaldo for pointing out the correct
new reference for CCID-3.
Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

This change lets "cat /proc/interrupts" show the name of the ethernet
device (e.g. eth0) rather than the driver name (smsc911x).
Signed-off-by: Steve Glendinning <steve.glendinning@smsc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

if this code path is ever hit, the platform_data struct isn't properly
configured with a bus width flag so the device won't work (hence the
BUG()).

This patch adds a dummy return statement to eliminate this compiler
warning:

drivers/net/smsc911x.c: In function 'smsc911x_reg_read':
drivers/net/smsc911x.c:148: warning: control reaches end of non-void function
Signed-off-by: Steve Glendinning <steve.glendinning@smsc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch fixes a potential race condition between scheduling napi and
completing napi poll. The call to netif_rx_schedule should be under
protection of the lock (as is the completion), otherwise, interrupts
could be masked off.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Removed the __exit annotation of tfrc_lib_exit(), in order to suppress the following section mismatch messages:

WARNING: net/dccp/dccp.o(.text+0xd9): Section mismatch in reference from the function ccid_cleanup_builtins() to the function .exit.text:tfrc_lib_exit()
The function ccid_cleanup_builtins() references a function in an exit section.
Often the function tfrc_lib_exit() has valid usage outside the exit section
and the fix is to remove the __exit annotation of tfrc_lib_exit.

WARNING: net/dccp/dccp.o(.init.text+0x48): Section mismatch in reference from the function ccid_initialize_builtins() to the function .exit.text:tfrc_lib_exit()
The function __init ccid_initialize_builtins() references
a function __exit tfrc_lib_exit().
This is often seen when error handling in the init function
uses functionality in the exit path.
The fix is often to remove the __exit annotation of
tfrc_lib_exit() so it may be used outside an exit section.
Signed-off-by: Leonardo Potenza <lpotenza@inwind.it>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch removes the feature flag for mgmt unit as it is not used for
this chipset.
Signed-off-by: Ayaz Abdulla <aabdulla@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

There is no need to sleep while waiting for the hardware
semaphore to become available.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Support for dev id 8000 is pushed out until 2.6.30.
Signed-off-by: Ron Mercer <ron.mercer@qlogic.com>
Signed-off-by: David S. Miller <davem@davemloft.net>