8 years agoBluetooth: Use delayed work for advertisiment cache timeout
Gustavo F. Padovan [Mon, 20 Jun 2011 19:39:29 +0000]
Bluetooth: Use delayed work for advertisiment cache timeout

As HCI rx path is now done in process context it makes sense to do all the
timer in process context as well.

Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Use delayed_work for connection timeout
Gustavo F. Padovan [Fri, 17 Jun 2011 16:03:21 +0000]
Bluetooth: Use delayed_work for connection timeout

Bluetooth rx task runs now in a workqueue, so it a good approach run any
timer that share locking with process context code also in a workqueue.

Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Replace spin_lock by mutex in hci_dev
Gustavo F. Padovan [Fri, 17 Jun 2011 16:03:21 +0000]
Bluetooth: Replace spin_lock by mutex in hci_dev

Now we run everything in HCI in process context, so it's a better idea use
mutex instead spin_lock. The macro remains hci_dev_lock() (and I got rid
of hci_dev_lock_bh()), of course.

Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Process recv path in a workqueue instead of a tasklet
Marcel Holtmann [Mon, 9 Aug 2010 03:06:53 +0000]
Bluetooth: Process recv path in a workqueue instead of a tasklet

Run recv process in workqueue helps a lot with our processing as the recv
path will also be in the process context, i.e., now all our tx and rx are
in process context.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: make hci_conn_enter_sniff_mode static
Gustavo F. Padovan [Wed, 7 Dec 2011 15:24:33 +0000]
Bluetooth: make hci_conn_enter_sniff_mode static

It isn't used outside hci_conn.c

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Use list_for_each_entry in hci_conn_hash_flush()
Gustavo F. Padovan [Thu, 15 Dec 2011 00:52:31 +0000]
Bluetooth: Use list_for_each_entry in hci_conn_hash_flush()

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Use chan instead of l2cap_pi macro
Gustavo F. Padovan [Fri, 9 Dec 2011 06:40:10 +0000]
Bluetooth: Use chan instead of l2cap_pi macro

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Initialize LE connection count
Gustavo F. Padovan [Wed, 14 Dec 2011 22:56:12 +0000]
Bluetooth: Initialize LE connection count

le_num needs to be set to zero.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: remove lock from struct conn_hash
Gustavo F. Padovan [Wed, 14 Dec 2011 17:10:41 +0000]
Bluetooth: remove lock from struct conn_hash

It isn't used anywhere.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: remove struct hci_chan_hash
Gustavo F. Padovan [Wed, 14 Dec 2011 15:02:51 +0000]
Bluetooth: remove struct hci_chan_hash

Only the list member of the struct was used, so we now fold it into
hci_conn.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: trivial: correct check for LMP version
Andrei Emeltchenko [Thu, 1 Dec 2011 12:33:28 +0000]
Bluetooth: trivial: correct check for LMP version

Make sure that code match exactly what comment says about pre 1.2
bluetooth version. Since this is HCI detail lmp_ver changed to hci_ver.

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Remove magic bluetooth version numbers
Andrei Emeltchenko [Thu, 1 Dec 2011 12:33:27 +0000]
Bluetooth: Remove magic bluetooth version numbers

Use bluetooth names instead of BT SIG assigned numbers

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoMerge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/padovan/bluetooth
Gustavo F. Padovan [Fri, 2 Dec 2011 23:26:25 +0000]
Merge branch 'master' of git://git./linux/kernel/git/padovan/bluetooth

8 years agoBluetooth: Correct version check in hci_setup
Andrei Emeltchenko [Wed, 23 Nov 2011 09:23:34 +0000]
Bluetooth: Correct version check in hci_setup

Check for hci_ver instead of lmp_ver

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agobtusb: fix a memory leak in btusb_send_frame()
Cong Wang [Tue, 22 Nov 2011 01:32:57 +0000]
btusb: fix a memory leak in btusb_send_frame()

This patch fixes the following memory leak reported by kmemleak:

unreferenced object 0xffff880060a53840 (size 192):
  comm "softirq", pid 0, jiffies 4320571771 (age 1406.569s)
  hex dump (first 32 bytes):
    01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<ffffffff81138a1c>] create_object+0x187/0x28b
    [<ffffffff814be12e>] kmemleak_alloc+0x73/0x98
    [<ffffffff811289d3>] __kmalloc+0xfc/0x123
    [<ffffffff81386546>] usb_alloc_urb+0x1e/0x48
    [<ffffffffa0130274>] btusb_send_frame+0x86/0x385 [btusb]
    [<ffffffffa02d8230>] hci_send_frame+0xa0/0xa5 [bluetooth]
    [<ffffffffa02d8a4e>] hci_cmd_task+0xa0/0xfb [bluetooth]
    [<ffffffff81058548>] tasklet_action+0x8f/0xef
    [<ffffffff81058a4c>] __do_softirq+0xf4/0x1db
    [<ffffffff81058bb7>] run_ksoftirqd+0x84/0x129
    [<ffffffff8106f1c4>] kthread+0xa0/0xa8
    [<ffffffff814dd144>] kernel_thread_helper+0x4/0x10
    [<ffffffffffffffff>] 0xffffffffffffffff

The problem is that when inc_tx() returns non-zero, we forgot
to call usb_free_urb().

Cc: Marcel Holtmann <marcel@holtmann.org>
Cc: "Gustavo F. Padovan" <padovan@profusion.mobi>
Signed-off-by: WANG Cong <amwang@redhat.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: remove old code
Andrei Emeltchenko [Thu, 1 Dec 2011 12:42:08 +0000]
Bluetooth: remove old code

Remove old code not touched for several years.

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: LE Set Scan Parameter Command
Andre Guedes [Fri, 2 Dec 2011 12:13:31 +0000]
Bluetooth: LE Set Scan Parameter Command

This patch adds the parameter struct and the command complete event
handler to the LE Set Scan Parameter HCI command.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add dev_flags to struct hci_dev
Andre Guedes [Fri, 25 Nov 2011 23:53:38 +0000]
Bluetooth: Add dev_flags to struct hci_dev

This patch adds the dev_flags field to struct hci_dev. This new
flags variable should be used to define flags related to BR/EDR
and/or LE controller itself. It should be used to define flags
which represents states from the controller. The dev_flags is
cleared in case the controller sends a Reset Command Complete
Event to the host.

Also, this patch adds the HCI_LE_SCAN flag which was created to
track if the controller is performing LE scan or not. The flag
is set/cleared when the controller starts/stops scanning.

This is an initial effort to stop using hdev->flags to define
internal flags since it is exported to userspace by an ioctl.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add HCI Read Flow Control Mode function
Andrei Emeltchenko [Thu, 24 Nov 2011 12:52:02 +0000]
Bluetooth: Add HCI Read Flow Control Mode function

Upstream Code Aurora function with minor trivial fixes.
Origin: git://codeaurora.org/kernel/msm.git

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Centralize SMP pairing failure handling
Brian Gix [Wed, 23 Nov 2011 16:28:36 +0000]
Bluetooth: Centralize SMP pairing failure handling

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Cleanup blkcipher on SMP termination
Brian Gix [Wed, 23 Nov 2011 16:28:35 +0000]
Bluetooth: Cleanup blkcipher on SMP termination

The blkcipher must be freed to avoid memory leak.

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add HCI User Passkey Req Evt handling
Brian Gix [Wed, 23 Nov 2011 16:28:34 +0000]
Bluetooth: Add HCI User Passkey Req Evt handling

Some MITM scenarios require handling of the User Passkey Request event,
by querying the user, and passing the response back.

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add User Passkey Response handling
Brian Gix [Wed, 23 Nov 2011 16:28:33 +0000]
Bluetooth: Add User Passkey Response handling

For some MITM protection pairing scenarios, the user is
required to enter or accept a 6 digit passkey.

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add user readable debug for state changes
Gustavo F. Padovan [Wed, 23 Nov 2011 22:11:46 +0000]
Bluetooth: Add user readable debug for state changes

I did this as a part of a testing course at university, but it might be
useful upstream as well.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Use queue in the device list
Andrei Emeltchenko [Wed, 16 Nov 2011 15:30:20 +0000]
Bluetooth: Use queue in the device list

Use queue instead of stack discipline for device list. When processing
dev_list with list_for_each* devices will be prosessed in order they
were added (Usually BR/EDR first and AMP later).

Also output from hciconfig looks nicer :-)

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Differentiate LE User Pairing Responses
Brian Gix [Wed, 16 Nov 2011 21:53:14 +0000]
Bluetooth: Differentiate LE User Pairing Responses

Low Energy (LE) pairing responses must be recognized and handled
differently from BR/EDR pairing responses. BR/EDR responses are
handled via HCI commands by the LMP layer, and LE responses are
handled by the Host.

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: User Pairing Response restructuring
Brian Gix [Wed, 16 Nov 2011 21:53:13 +0000]
Bluetooth: User Pairing Response restructuring

There are 4 possible User Responses to pairing requests,
and they all share the same checks and handling. This
restructures the handling of the two Confirm responses
in preperation for the second two.

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: making enable_hs independent from L2CAP
Andrei Emeltchenko [Fri, 18 Nov 2011 11:35:42 +0000]
Bluetooth: making enable_hs independent from L2CAP

Fixes bluetooth compiling when CONFIG_BT_L2CAP is not enabled

net/built-in.o: In function `hci_dev_open':
(.text+0xdce9a): undefined reference to `enable_hs'

Reported-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add MGMT event for Passkey Entry
Brian Gix [Wed, 16 Nov 2011 21:53:12 +0000]
Bluetooth: Add MGMT event for Passkey Entry

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: bnep: Fix module reference
David Herrmann [Sat, 19 Nov 2011 12:23:33 +0000]
Bluetooth: bnep: Fix module reference

We cannot call module_put(THIS_MODULE) if this is our last reference. Otherwise,
this call may cleanup our module before it returns.

Gladly, the kthread API provides a simple wrapper for us. So lets use
module_put_and_exit() to avoid a race condition with the module cleanup code.

Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: cmtp: Fix module reference
David Herrmann [Sat, 19 Nov 2011 12:23:32 +0000]
Bluetooth: cmtp: Fix module reference

We cannot call module_put(THIS_MODULE) if this is our last reference. Otherwise,
this call may cleanup our module before it returns.

Gladly, the kthread API provides a simple wrapper for us. So lets use
module_put_and_exit() to avoid a race condition with the module cleanup code.

Signed-off-by: David Herrmann <dh.herrmann@googlemail.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: btmrvl: support Marvell Bluetooth device SD8797
Bing Zhao [Thu, 17 Nov 2011 04:40:42 +0000]
Bluetooth: btmrvl: support Marvell Bluetooth device SD8797

The SD8797 firmware image is shared with mwifiex driver.
Whoever gets loaded first will be responsible for firmware
downloading.

Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: Frank Huang <frankh@marvell.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Fix some checkpatch.pl errors and warnings
Szymon Janc [Wed, 16 Nov 2011 08:32:21 +0000]
Bluetooth: Fix some checkpatch.pl errors and warnings

Signed-off-by: Szymon Janc <szymon.janc@tieto.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Simplify __l2cap_global_chan_by_addr
Szymon Janc [Wed, 16 Nov 2011 08:32:22 +0000]
Bluetooth: Simplify __l2cap_global_chan_by_addr

Make __l2cap_global_chan_by_addr similar to other find functions.

Signed-off-by: Szymon Janc <szymon.janc@tieto.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Refactor loop in l2cap_retransmit_one_frame
Szymon Janc [Wed, 16 Nov 2011 08:32:20 +0000]
Bluetooth: Refactor loop in l2cap_retransmit_one_frame

This make it easier to see what is the real reason for loop to exit.
skb_queue_next return valid skb or garbage, not NULL.

Signed-off-by: Szymon Janc <szymon.janc@tieto.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Simplify l2cap_add_to_srej_queue
Szymon Janc [Wed, 16 Nov 2011 08:32:19 +0000]
Bluetooth: Simplify l2cap_add_to_srej_queue

Make it easier to see what is loop break condition.
skb_queue_next return valid skb or garbage, not NULL.

Signed-off-by: Szymon Janc <szymon.janc@tieto.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Fix possible NULL pointer derefence in l2cap code
Szymon Janc [Wed, 16 Nov 2011 08:32:18 +0000]
Bluetooth: Fix possible NULL pointer derefence in l2cap code

Due to ERTM reliability L2CAP channel needs to be disconnected if
adding to srej list failed.

Signed-off-by: Szymon Janc <szymon.janc@tieto.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Allow to set AMP type for virtual HCI
Andrei Emeltchenko [Mon, 14 Nov 2011 10:42:48 +0000]
Bluetooth: Allow to set AMP type for virtual HCI

Type can be changed during re-opening device /dev/vhci.

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Do not set HCI_RAW when HS enabled
Andrei Emeltchenko [Fri, 11 Nov 2011 15:02:15 +0000]
Bluetooth: Do not set HCI_RAW when HS enabled

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Move scope of kernel parameter enable_hs
Andrei Emeltchenko [Fri, 11 Nov 2011 15:02:14 +0000]
Bluetooth: Move scope of kernel parameter enable_hs

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add MGMT opcodes for Passkey Entry
Brian Gix [Sun, 13 Nov 2011 06:01:12 +0000]
Bluetooth: Add MGMT opcodes for Passkey Entry

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add HCI defines for User Passkey entry
Brian Gix [Sun, 13 Nov 2011 06:01:11 +0000]
Bluetooth: Add HCI defines for User Passkey entry

Signed-off-by: Brian Gix <bgix@codeaurora.org>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Pass all message parameters to mgmt_start_discovery
Johan Hedberg [Sat, 12 Nov 2011 09:58:22 +0000]
Bluetooth: Pass all message parameters to mgmt_start_discovery

The mgmt_start_discovery command contains the type of discovery that
should be started so this should be passed to the start_discovery
function. This patch doesn't yet add any action depending on the type of
the requested discovery.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add missing cmd_complete for mgmt_load_link_keys
Johan Hedberg [Fri, 11 Nov 2011 14:18:54 +0000]
Bluetooth: Add missing cmd_complete for mgmt_load_link_keys

The command complete event was completely missing for this command.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Create a unique mgmt error code hierarchy
Johan Hedberg [Fri, 11 Nov 2011 16:10:00 +0000]
Bluetooth: Create a unique mgmt error code hierarchy

The management protocol uses a single byte for error codes (aka command
status). In some places this value is directly copied from HCI and in
other a POSIX error number is used. This makes it impossible for
user-space to uniquily decipher the meaning of an error.

To solve this issue a new mgmt-specific set of error codes is added
along with a conversion table for HCI status values.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Return success instead of EALREADY for mgmt commands
Johan Hedberg [Fri, 11 Nov 2011 14:18:52 +0000]
Bluetooth: Return success instead of EALREADY for mgmt commands

When the adapter state is already what is requested it's more friendly
to user-space to simply report success than to send a EALREADY error
message.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Fix mgmt_pair_device imediate error responses
Johan Hedberg [Thu, 10 Nov 2011 22:07:35 +0000]
Bluetooth: Fix mgmt_pair_device imediate error responses

When possible cmd_complete should be returned instead of cmd_status
since it contains the remote address (this helps user-space track what
exactly failed).

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add address type to mgmt_pair_device
Johan Hedberg [Thu, 10 Nov 2011 22:07:34 +0000]
Bluetooth: Add address type to mgmt_pair_device

The kernel needs to know whether it should connect to a device over
BR/EDR or over LE. This is particularly important in the future when
dual-mode device may be connectable also over LE. It is also important
if/when we decide to move the LE advertisement cache from the kernel
into user-space. Adding the type to the mgmt command also ensures
conformance with the latest mgmt API spec.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: mgmt_stop_discovery_failed()
Andre Guedes [Wed, 9 Nov 2011 20:14:26 +0000]
Bluetooth: mgmt_stop_discovery_failed()

This patches creates mgmt_stop_discovery_failed() which removes
pending MGMT_OP_STOP_DISCOVERY commands and sends proper command
status events.

This patch also fixes the MGMT_OP_STOP_DISCOVERY command leak in
case cancel inquiry fails.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Rename mgmt_inquiry_failed()
Andre Guedes [Wed, 9 Nov 2011 20:14:25 +0000]
Bluetooth: Rename mgmt_inquiry_failed()

This patch renames the function mgmt_inquiry_failed() to
mgmt_start_discovery_failed(). This function is more related
to MGMT_OP_START_DISCOVERY command handling than to inquiry.
Besides, this functions will be reused by LE based discovery
procedures in case of failure.

Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
Acked-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Only set ack_timer if we didn't send and ack
Gustavo F. Padovan [Thu, 23 Jun 2011 22:30:48 +0000]
Bluetooth: Only set ack_timer if we didn't send and ack

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Allow L2CAP to increase the security level
Gustavo F. Padovan [Sat, 5 Nov 2011 21:58:31 +0000]
Bluetooth: Allow L2CAP to increase the security level

Some incomming connections needs to increase the security level by
requesting encryption for example (HID keyboard case). This change allows
the userspace to change it through setsockopt with defer_setup enabled.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Rename l2cap_check_security()
Gustavo F. Padovan [Sat, 5 Nov 2011 21:54:24 +0000]
Bluetooth: Rename l2cap_check_security()

rename to l2cap_chan_check_security() to make it consistent with other
l2cap_exported functions. This function will be exported in a later
commit.

Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add status parameter to mgmt_disconnect response
Johan Hedberg [Thu, 10 Nov 2011 13:54:39 +0000]
Bluetooth: Add status parameter to mgmt_disconnect response

Since disconnecting may fail the status needs to be communicated to user
space. This also updates the implementation to match the latest mgmt API
specification.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add proper response to mgmt_remove_keys command
Johan Hedberg [Thu, 10 Nov 2011 13:54:38 +0000]
Bluetooth: Add proper response to mgmt_remove_keys command

Since the command can fail we need to have a proper response with the
remote address and a failure status for it. This also updates it to
conform to the latest mgmt API spec.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Fix deadlock with mgmt_pair_device
Johan Hedberg [Thu, 10 Nov 2011 13:54:37 +0000]
Bluetooth: Fix deadlock with mgmt_pair_device

The hci_conn callbacks are called with the hci_dev lock already held so
no locking should be attempted in them.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agoBluetooth: Add public/random LE address information to mgmt messages
Johan Hedberg [Wed, 9 Nov 2011 11:58:58 +0000]
Bluetooth: Add public/random LE address information to mgmt messages

It's necessary to know the distinction between public and random LE
addresses so the mgmt interface also needs to distinguish between them.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Gustavo F. Padovan <padovan@profusion.mobi>

8 years agomwifiex: prevent corruption instead of just warning
Dan Carpenter [Tue, 8 Nov 2011 03:31:47 +0000]
mwifiex: prevent corruption instead of just warning

Probably we never hit this condition, but in case we do, we may as
well put a return here instead of just printing a warning message and
then corrupting memory.  The caller doesn't check the return code.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomwifiex: enable SDIO multiport aggregation
Amitkumar Karwar [Tue, 8 Nov 2011 03:31:46 +0000]
mwifiex: enable SDIO multiport aggregation

By default SDIO multiport aggregation is disabled. It's useful to
get good throughput results. This patch enables it.

Signed-off-by: Amitkumar Karwar <akarwar@marvell.com>
Signed-off-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomwifiex: remove an unneeded NULL check
Dan Carpenter [Tue, 8 Nov 2011 03:31:45 +0000]
mwifiex: remove an unneeded NULL check

We dereference adapter in the error handling code so this needed to
be fixed.  This function is always called like:
adapter->if_ops.host_to_card(adapter, ...);
so adapter can never be NULL and I've removed the NULL check.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomwifiex: remove unneeded kfree(NULL);
Dan Carpenter [Tue, 8 Nov 2011 03:31:44 +0000]
mwifiex: remove unneeded kfree(NULL);

The previous if statement means this that pointer is NULL so there
is no need to free it.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Bing Zhao <bzhao@marvell.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agonet/rfkill/core.c: use kstrtoul, etc
Julia Lawall [Sun, 6 Nov 2011 13:26:49 +0000]
net/rfkill/core.c: use kstrtoul, etc

Use kstrtoul, etc instead of the now deprecated strict_strtoul, etc.

A semantic patch rule for the kstrtoul case is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression a,b;
{int,long} *c;
@@

-strict_strtoul
+kstrtoul
 (a,b,c)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agonet/mac80211/debugfs.c: use kstrtoul, etc
Julia Lawall [Sun, 6 Nov 2011 13:26:48 +0000]
net/mac80211/debugfs.c: use kstrtoul, etc

Use kstrtoul, etc instead of the now deprecated strict_strtoul, etc.

A semantic patch rule for the kstrtoul case is as follows:
(http://coccinelle.lip6.fr/)

// <smpl>
@@
expression a,b;
{int,long} *c;
@@

-strict_strtoul
+kstrtoul
 (a,b,c)
// </smpl>

Signed-off-by: Julia Lawall <julia@diku.dk>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomac80211: implement wifi TX status
Johannes Berg [Sun, 6 Nov 2011 13:13:34 +0000]
mac80211: implement wifi TX status

Implement the socket wifi TX status error
queue reflection in mac80211.

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

8 years agonl80211: advertise socket TX status capability
Johannes Berg [Sun, 6 Nov 2011 13:13:33 +0000]
nl80211: advertise socket TX status capability

The new wifi socket TX capability should be
supported by wifi drivers, let them advertise
whether they do or not.

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

8 years agomac80211: Also report the STA's TDLS flag via nl80211
Helmut Schaa [Sat, 5 Nov 2011 13:15:24 +0000]
mac80211: Also report the STA's TDLS flag via nl80211

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

8 years agobrcm80211: fmac: optimize chip core info management
Franky Lin [Fri, 4 Nov 2011 21:23:42 +0000]
brcm80211: fmac: optimize chip core info management

Prepare for adding backplane interconnect type support

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: replace private SB macros with ssb_regs version
Franky Lin [Fri, 4 Nov 2011 21:23:41 +0000]
brcm80211: fmac: replace private SB macros with ssb_regs version

Use SSB macros in order to clean up brcmfmac code

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move chip drive strength related code to sdio_chip.c
Franky Lin [Fri, 4 Nov 2011 21:23:40 +0000]
brcm80211: fmac: move chip drive strength related code to sdio_chip.c

This patch is part of the abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move chip detach function to sdio_chip.c
Franky Lin [Fri, 4 Nov 2011 21:23:39 +0000]
brcm80211: fmac: move chip detach function to sdio_chip.c

This patch is part of the abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move chip reset core function to sdio_chip.c
Franky Lin [Fri, 4 Nov 2011 21:23:38 +0000]
brcm80211: fmac: move chip reset core function to sdio_chip.c

This patch is part of the abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: abstract chip core revision function
Franky Lin [Fri, 4 Nov 2011 21:23:37 +0000]
brcm80211: fmac: abstract chip core revision function

Prepare for adding backplane interconnect type support

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: abstract chip iscoreup function
Franky Lin [Fri, 4 Nov 2011 21:23:36 +0000]
brcm80211: fmac: abstract chip iscoreup function

Prepare for adding backplane interconnect type support

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: chip attach code flow clean up
Franky Lin [Fri, 4 Nov 2011 21:23:35 +0000]
brcm80211: fmac: chip attach code flow clean up

Merged brcmf_sdbrcm_chip_attach into brcmf_sdio_chip_attach
for better readability.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: remove duplicate regiter set in chip attach path
Franky Lin [Fri, 4 Nov 2011 21:23:34 +0000]
brcm80211: fmac: remove duplicate regiter set in chip attach path

Same register writes have been done in brcmf_sdbrcm_probe_init
which is earlier in the same code path.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move dongle gpio reset code to chip attach function
Franky Lin [Fri, 4 Nov 2011 21:23:33 +0000]
brcm80211: fmac: move dongle gpio reset code to chip attach function

This patch is part of the abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: disable dongle arm core in bus core setup function
Franky Lin [Fri, 4 Nov 2011 21:23:32 +0000]
brcm80211: fmac: disable dongle arm core in bus core setup function

This will provide a better code flow that fits the logic

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move core disable function to sdio_chip.c
Franky Lin [Fri, 4 Nov 2011 21:23:31 +0000]
brcm80211: fmac: move core disable function to sdio_chip.c

This patch is part of the abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: abstract chip buscore setup function
Franky Lin [Fri, 4 Nov 2011 21:23:30 +0000]
brcm80211: fmac: abstract chip buscore setup function

This patch is part of the abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move bus core prep code to sdio_chip.c
Franky Lin [Fri, 4 Nov 2011 21:23:29 +0000]
brcm80211: fmac: move bus core prep code to sdio_chip.c

This patch is part of abstracting chip backplane handle code
series.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: move chip recognition function to sdio_chip.c
Franky Lin [Fri, 4 Nov 2011 21:23:28 +0000]
brcm80211: fmac: move chip recognition function to sdio_chip.c

Currently backplane handle code is scatterd around dhd_sdio.c which
is not good for maintenance and adding new backplane interconnect
type support. This patch and the follow up patches are going to
abstract all chip backplane control code specific for sdio bus
into this new sdio_chip.c

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agobrcm80211: fmac: remove unnecessary 4329 chip specific code
Franky Lin [Fri, 4 Nov 2011 21:23:27 +0000]
brcm80211: fmac: remove unnecessary 4329 chip specific code

4329 with chiprev 0 can not be found on any product. The code can
be removed.

Reviewed-by: Arend van Spriel <arend@broadcom.com>
Reviewed-by: Roland Vossen <rvossen@broadcom.com>
Signed-off-by: Franky Lin <frankyl@broadcom.com>
Signed-off-by: Arend van Spriel <arend@broadcom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agocfg80211/mac80211: allow management TX to not wait for ACK
Johannes Berg [Fri, 4 Nov 2011 10:18:21 +0000]
cfg80211/mac80211: allow management TX to not wait for ACK

For probe responses it can be useful to not wait for ACK to
avoid retransmissions if the station that sent the probe is
already on the next channel, so allow userspace to request
not caring about the ACK with a new nl80211 flag.

Since mac80211 needs to be updated for the new function
prototype anyway implement it right away -- it's just a
few lines of code.

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

8 years agomac80211: send unexpected 4addr event
Johannes Berg [Fri, 4 Nov 2011 10:18:20 +0000]
mac80211: send unexpected 4addr event

Implement the cfg80211 notification but only send
one event per associated station to avoid having
tons of events if the station thinks it should be
allowed to use 4addr frames but it isn't.

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

8 years agocfg80211: add event for unexpected 4addr frames
Johannes Berg [Fri, 4 Nov 2011 10:18:19 +0000]
cfg80211: add event for unexpected 4addr frames

The frames are used by AP/STA WDS mode, and hostapd
needs to know when such a frame was received to set
up the VLAN appropriately to allow using it.

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

8 years agomac80211: report OBSS beacons
Johannes Berg [Fri, 4 Nov 2011 10:18:18 +0000]
mac80211: report OBSS beacons

If there's an interface in AP mode, OBSS beacons
are needed by hostapd/wpa_s to implement logic to
enable/disable protection etc. Report the frames
and set the capability flag.

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

8 years agocfg80211: allow registering to beacons
Johannes Berg [Fri, 4 Nov 2011 10:18:17 +0000]
cfg80211: allow registering to beacons

Add the ability to register to received beacon frames
to allow implementing OLBC logic in userspace. The
registration is per wiphy since there's no point in
receiving the same frame multiple times.

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

8 years agomac80211: support client probe
Johannes Berg [Fri, 4 Nov 2011 10:18:16 +0000]
mac80211: support client probe

Support probing clients with null data frames
in AP mode.

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

8 years agonl80211: add API to probe a client
Johannes Berg [Fri, 4 Nov 2011 10:18:15 +0000]
nl80211: add API to probe a client

When the AP SME in hostapd is used it wants to
probe the clients when they have been idle for
some time. Add explicit API to support this.

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

8 years agonl80211: advertise device AP SME
Johannes Berg [Mon, 7 Nov 2011 11:39:33 +0000]
nl80211: advertise device AP SME

Add the ability to advertise that the device
contains the AP SME and what features it can
support. There are currently no features in
the bitmap -- probe response offload will be
advertised by a few patches Arik is working
on now (who took over from Guy Eilam) and a
device with AP SME will typically implement
and require response offload.

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

8 years agomac80211: support spurious class3 event
Johannes Berg [Fri, 4 Nov 2011 10:18:13 +0000]
mac80211: support spurious class3 event

Add support for the spurious class3 frame event
to mac80211 to enable AP w/o monitor mode.

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

8 years agonl80211: allow subscribing to unexpected class3 frames
Johannes Berg [Fri, 4 Nov 2011 10:18:12 +0000]
nl80211: allow subscribing to unexpected class3 frames

To implement AP mode without monitor interfaces we
need to be able to send a deauth to stations that
send frames without being associated. Enable this
by adding a new nl80211 event for such frames that
an application can subscribe to.

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

8 years agomac80211: add support for control port protocol in AP mode
Johannes Berg [Fri, 4 Nov 2011 10:18:11 +0000]
mac80211: add support for control port protocol in AP mode

This will allow us to support dynamic WEP with 802.1X
properly in mac80211 by not encrypting outgoing and
accepting unencrypted incoming frames.

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

8 years agomac80211: add helper to free TX skb
Johannes Berg [Fri, 4 Nov 2011 10:18:10 +0000]
mac80211: add helper to free TX skb

Drivers that need to drop a frame before it
can be transmitted will usually simply free
that frame. This is currently fine, but in
the future it'll be needed to tell mac80211
about this case, so add a new routine that
frees a TX skb.

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

8 years agomac80211: simplify mesh frame queue mapping and QoS
Javier Cardona [Fri, 4 Nov 2011 04:11:13 +0000]
mac80211: simplify mesh frame queue mapping and QoS

We only need to set the skb queue twice:

1. by the netdev, on local TX.
2. when forwarding a mesh frame.

We only need to set the qos header twice:

1. by mac80211, on local TX.
2. when putting a frame on the mpath->frame_queue

We also don't need the RA in order to set the proper queue mapping since
all mesh STAs are QoS, indicate this and do it once when the frame is
received. Also fixes an issue where the QoS header and queue mapping was not
set for unicast forwarded frames.

Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomac80211: check if frame is really part of this BA
Thomas Pedersen [Fri, 4 Nov 2011 04:11:12 +0000]
mac80211: check if frame is really part of this BA

There was an an implicit assumption that any QoS data frame received
from a STA/TID with an active BA session was sent to this vif as part of
a BA.  This is not true if IFF_PROMISC is enabled and the frame was
destined for a different peer, for example. Don't treat these frames as
part of a BA from the sending STA.

Signed-off-by: Thomas Pedersen <thomas@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomac80211: QoS multicast frames have No Ack policy
Thomas Pedersen [Fri, 4 Nov 2011 04:11:11 +0000]
mac80211: QoS multicast frames have No Ack policy

Previously QoS multicast frames had the Normal Acknowledgment QoS
control bits set. This would cause broadcast frames to be discarded by
peers with which we have a BA session, since their sequence number would
fall outside the allowed range. Set No Ack QoS control bits on multicast
QoS frames and filter these in de-aggregation code.

Signed-off-by: Thomas Pedersen <thomas@cozybit.com>

v2: Use proper QoS Ack Policy ctl field mask (Christian)

v3: Clean up conditional (Johannes)
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomac80211: Avoid filling up mesh preq queue with redundant requests
Javier Cardona [Fri, 4 Nov 2011 04:11:10 +0000]
mac80211: Avoid filling up mesh preq queue with redundant requests

Don't accept redundant PREQs for a given destination. This fixes a
problem under high load:

kernel: [20386.250913] mesh_queue_preq: 235 callbacks suppressed
kernel: [20386.253335] Mesh HWMP (mesh0): PREQ node queue full
kernel: [20386.253352] Mesh HWMP (mesh0): PREQ node queue full
(...)

The 802.11s protocol has a provision to limit the rate of path requests
(PREQs) are transmitted (dot11MeshHWMPpreqMinInterval) but there was no
limit on the rate at which PREQs were being queued up.  There is a valid
reason for queuing PREQs: this way we can even out PREQ bursts.  But
queueing multiple PREQs for the same destination is useless.

Reported-by: Pedro Larbig <pedro.larbig@carhs.de>
Signed-off-by: Javier Cardona <javier@cozybit.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agoath9k: Improve debugfs printout for stations.
Ben Greear [Thu, 3 Nov 2011 18:33:13 +0000]
ath9k: Improve debugfs printout for stations.

Add interface address so it can be mapped to a local
interface.  Add max-ampdu and mpdu-density.

Print out the tid->baw_size

Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>

8 years agomac80211: verify virtual interfaces in driver API
Johannes Berg [Thu, 3 Nov 2011 13:41:13 +0000]
mac80211: verify virtual interfaces in driver API

The driver is never informed about monitor or
AP_VLAN interfaces, so whenever we pass those
to it later this is a bug. Verify we don't as
there are some cases where this could happen.

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