pandora-kernel.git
9 years agoBluetooth: Refactor ioctl scan state update to its own function
Johan Hedberg [Thu, 10 Jul 2014 09:09:07 +0000 (12:09 +0300)]
Bluetooth: Refactor ioctl scan state update to its own function

With subsequent patches we'll also need to update the discoverable
state. As the code grows bigger it's better to move this out from the
switch statement into its own function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unneeded mgmt_connectable function
Johan Hedberg [Thu, 10 Jul 2014 09:09:06 +0000 (12:09 +0300)]
Bluetooth: Remove unneeded mgmt_connectable function

The mgmt_connectable function has been used to ensure that the right
actions to HCI_CONNECTABLE are taken when the HCI_Write_Scan_Enable
command is triggered by something else than mgmt. The only other user
that we really care about is the HCISETSCAN ioctl code, so we can
actually more simply perform the needed changes there instead.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix using test_and_clear instead of test_and_set
Johan Hedberg [Thu, 10 Jul 2014 08:50:27 +0000 (11:50 +0300)]
Bluetooth: Fix using test_and_clear instead of test_and_set

The code for updating the HCI_CONNECTABLE flag was incorrectly using
test_and_set_bit instead of test_and_clear_bit when HCI_CONNECTABLE is
to be cleared.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluteooth: Reject SMP bonding if HCI_PAIRABLE is not set
Johan Hedberg [Thu, 10 Jul 2014 08:02:07 +0000 (11:02 +0300)]
Bluteooth: Reject SMP bonding if HCI_PAIRABLE is not set

If the remote device tries to initiate bonding with us and we don't have
HCI_PAIRABLE set we should just flat out reject the request. This brings
SMP in line with how the flag is used for BR/EDR SSP.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix advertising parameter update when toggling connectable
Johan Hedberg [Thu, 10 Jul 2014 07:51:27 +0000 (10:51 +0300)]
Bluetooth: Fix advertising parameter update when toggling connectable

When we change the connectable state and have advertising enabled we
should update the advertising parameters no matter what. The code was
incorrectly only updating them if advertising was not already active.
This patch fixes the issue.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix incorrect clearing of SMP_FLAG_INITIATOR
Johan Hedberg [Wed, 9 Jul 2014 16:18:11 +0000 (19:18 +0300)]
Bluetooth: Fix incorrect clearing of SMP_FLAG_INITIATOR

When the SMP context is created all flags default to zero. To determine
that we are the initiators it's therefore best to simply change the flag
value when we know we're sending the first SMP PDU. Clearing the flag
when receiving a Pairing Request is not correct since the request may be
a response to a previous Security Request from us (for which we would
already have correctly set the flag). Same goes for receiving a Security
Request which may be coming after us already starting pairing by sending
a Pairing Request.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix forcing SMP just-works with no-bonding
Johan Hedberg [Wed, 9 Jul 2014 16:18:10 +0000 (19:18 +0300)]
Bluetooth: Fix forcing SMP just-works with no-bonding

Whether we bond or not should not have any impact on the user
interaction model. This patch removes an incorrect fall-back from
JUST_CFM to JUST_WORKS in case we're not bonding.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Always confirm incoming SMP just-works requests
Johan Hedberg [Wed, 9 Jul 2014 16:18:09 +0000 (19:18 +0300)]
Bluetooth: Always confirm incoming SMP just-works requests

For incoming requests we want to let the user know that pairing is
happening since otherwise there could be access to MEDIUM security
services without any user interaction at all. Therefore, set the
selected method to JUST_CFM instead of JUST_WORKS and let it be
converted back to JUST_WORKS later if we are the initators.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Don't send ERTM configuration option when disabled
Marcel Holtmann [Wed, 9 Jul 2014 09:53:35 +0000 (11:53 +0200)]
Bluetooth: Don't send ERTM configuration option when disabled

When ERTM support is disabled, then do not even send ERTM configuration
option even if the remote side supports it.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Fix setting HCI_CONNECTABLE from ioctl code
Johan Hedberg [Wed, 9 Jul 2014 10:28:26 +0000 (13:28 +0300)]
Bluetooth: Fix setting HCI_CONNECTABLE from ioctl code

When the white list is in use the code would not update the
HCI_CONNECTABLE flag if it gets changed through the ioctl code (e.g.
hciconfig hci0 pscan). Since the flag is important for properly
accepting incoming connections add code to fix it up if necessary and
emit a New Settings mgmt event.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Add white list lookup for incoming connection requests
Johan Hedberg [Wed, 9 Jul 2014 09:59:19 +0000 (12:59 +0300)]
Bluetooth: Add white list lookup for incoming connection requests

This patch adds support for looking up entries in the white list when
HCI_CONNECTABLE is not set. The logic is fairly simple: if we're
connectable check the black list, if we're not connectable check the
white list.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Enable page scan also if there are white list entries
Johan Hedberg [Wed, 9 Jul 2014 09:59:18 +0000 (12:59 +0300)]
Bluetooth: Enable page scan also if there are white list entries

Page scan should be enabled either if the connectable setting is set or
if there are any entries in the BR/EDR white list. This patch implements
such behavior by updating the two places that were making decisions on
whether to enable page scan or not.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Refactor connection request handling
Johan Hedberg [Wed, 9 Jul 2014 09:59:17 +0000 (12:59 +0300)]
Bluetooth: Refactor connection request handling

The conditions for accepting an incoming connections are already
non-trivial and will become more so once a white list is added. This
patch breaks up the checks for when to reject the request by creating a
helper function for it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix incorrectly setting HCI_CONNECTABLE
Johan Hedberg [Wed, 9 Jul 2014 09:59:16 +0000 (12:59 +0300)]
Bluetooth: Fix incorrectly setting HCI_CONNECTABLE

Since page scan might be enabled by Add Device we should not implicitly
set connectable whenever something else than Set Connectable changes it.
This patch makes sure that we don't set HCI_CONNECTABLE for these cases
if there are any entries in the white list.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Update page scan when necessary for Add/Remove Device
Johan Hedberg [Wed, 9 Jul 2014 09:59:15 +0000 (12:59 +0300)]
Bluetooth: Update page scan when necessary for Add/Remove Device

When we're removing the last item in the white list or adding the first
one to it and HCI_CONNECTABLE is not set we need to update the current
page scan. This patch adds a simple helper function for the purpose and
calls it from the respective mgmt command handlers.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Introduce a whitelist for BR/EDR devices
Johan Hedberg [Wed, 9 Jul 2014 09:59:14 +0000 (12:59 +0300)]
Bluetooth: Introduce a whitelist for BR/EDR devices

This patch extends the Add/Remove device commands by letting user space
pass BR/EDR addresses to them. The resulting entries get stored in a new
hdev->whitelist list. The idea is that we can now selectively accept
connections from devices in the list even though HCI_CONNECTABLE is not
set (the actual implementation of this is coming in a subsequent patch).

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Unify helpers for bdaddr_list manipulations
Johan Hedberg [Wed, 9 Jul 2014 09:59:13 +0000 (12:59 +0300)]
Bluetooth: Unify helpers for bdaddr_list manipulations

We already have several lists with struct bdaddr_list entries, and there
will be more in the future. Since the operations for adding, removing,
looking up and clearing entries in these lists are exactly the same it
doesn't make sense to define new functions for every single list. This
patch unifies the functions by passing the list_head to them instead of
a hci_dev pointer.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix enabling Authenticated Payload Timeout Expired event
Marcel Holtmann [Wed, 9 Jul 2014 07:49:05 +0000 (09:49 +0200)]
Bluetooth: Fix enabling Authenticated Payload Timeout Expired event

The Authenticated Payload Timeout Expired event is valid for
controllers with BR/EDR Secure Connections support, but also for
LE only controllers supporting LE Ping feature. When either of them
is available enable this event. Previous it was not enabled when
the controller was only supporting LE operation.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for Acer [13D3:3432]
Anantha Krishnan [Tue, 8 Jul 2014 13:55:08 +0000 (19:25 +0530)]
Bluetooth: Add support for Acer [13D3:3432]

Add support for the QCA6174 chip.

    T:  Bus=04 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#= 30 Spd=12  MxCh= 0
    D:  Ver= 1.10 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
    P:  Vendor=13d3 ProdID=3432 Rev=00.02
    C:  #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
    I:  If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
    I:  If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb

Signed-off-by: Anantha Krishnan <ananthk@codeaurora.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix clearing HCI_LE_ADV for LE connections
Johan Hedberg [Tue, 8 Jul 2014 14:21:51 +0000 (17:21 +0300)]
Bluetooth: Fix clearing HCI_LE_ADV for LE connections

All LE controllers always implicitly stop advertising when establishing
connections. Therefore, be sure to clear the flag in the event handler
for new LE connections.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix toggling background scan when changing connectable state
Johan Hedberg [Tue, 8 Jul 2014 13:35:34 +0000 (16:35 +0300)]
Bluetooth: Fix toggling background scan when changing connectable state

If the connectable state change doesn't require any special HCI commands
the set_connectable_update_settings() function is used instead of the
set_connectable_complete() function. We must therefore make sure to call
hci_update_background_scan() there as well. This code path is used also
when we're powered off, but that's fine since hci_update_background_scan()
has the necessary checks for it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix connectable and discoverable supported settings values
Johan Hedberg [Tue, 8 Jul 2014 13:07:34 +0000 (16:07 +0300)]
Bluetooth: Fix connectable and discoverable supported settings values

The connectable and discoverable mgmt settings are supported both for LE
and BR/EDR controllers so they do not belong behind a lmp_bredr_capable()
condition. This patch fixes the issue in get_supported_settings().

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix setting STOPPING state for discovery
Johan Hedberg [Tue, 8 Jul 2014 13:05:06 +0000 (16:05 +0300)]
Bluetooth: Fix setting STOPPING state for discovery

If any of the HCI commands from the hci_stop_discovery function were
successfully sent we need to set the discovery state to STOPPING. The
Stop Discovery code was already handling this, but the code in
clean_up_hci_state was not. This patch updates the hci_stop_discovery to
return a bool to indicate whether it queued any commands and the
clean_up_hci_state() function respectively to look at the return value
and call hci_discovery_set_state() if necessary.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix check for re-enabling advertising
Johan Hedberg [Tue, 8 Jul 2014 13:05:05 +0000 (16:05 +0300)]
Bluetooth: Fix check for re-enabling advertising

There are many different places that can disable LE scanning but we only
want to re-enable advertising in hci_cc_le_set_scan_enable() for a very
specific use case, which is when the active scanning part of Start
Discovery is complete. Because of this, fix the discovery state check to
test for the exact state.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix advertising and active scanning co-existence
Johan Hedberg [Tue, 8 Jul 2014 12:07:53 +0000 (15:07 +0300)]
Bluetooth: Fix advertising and active scanning co-existence

Many controllers allow simultaneous active scanning and advertising
(e.g. Intel and Broadcom) but some do not (e.g. CSR). It's therefore
safest to implement mutual exclusion of these states in the kernel.

This patch ensures that the two states are never entered simultaneously.
Extra precaution needs to be taken for outgoing connection attempts in
slave role (i.e. through directed advertising) in which case the
operation that came first has precedence and the one that comes after
gets a rejection.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Stop advertising always before initiating a connection
Johan Hedberg [Tue, 8 Jul 2014 12:07:52 +0000 (15:07 +0300)]
Bluetooth: Stop advertising always before initiating a connection

Most controllers do not support advertising while initiating an LE
connection. We also have to first disable current advertising if the
initiation is going to happen through direct advertising. Therefore,
simply stop advertising as the first thing when starting to issue
commands to establish an LE connection.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Use the correct flag to decide to disable advertising
Johan Hedberg [Tue, 8 Jul 2014 12:07:51 +0000 (15:07 +0300)]
Bluetooth: Use the correct flag to decide to disable advertising

When deciding to call disable_advertising() we're interested in the real
state instead of the mgmt setting. Use therefore HCI_LE_ADV instead of
the HCI_ADVERTISING flag.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Simplify usage of the enable_advertising function
Johan Hedberg [Tue, 8 Jul 2014 12:07:50 +0000 (15:07 +0300)]
Bluetooth: Simplify usage of the enable_advertising function

By adding support for disabling advertising when necessary and doing the
checks for existing LE connections inside the enable_advertising
function we can simplify the calling code quite a lot.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Use real advertising state to random address update decision
Johan Hedberg [Tue, 8 Jul 2014 12:07:49 +0000 (15:07 +0300)]
Bluetooth: Use real advertising state to random address update decision

Now that we have a flag for tracking the real advertising state we
should use that to determine whether it's safe to update the random
address or not. The couple of places that were clearing the flag due to
a pending request need to be updated too.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unnecessary mgmt_advertising function
Johan Hedberg [Tue, 8 Jul 2014 12:07:48 +0000 (15:07 +0300)]
Bluetooth: Remove unnecessary mgmt_advertising function

Since the real advertising state is now tracked with its own flag we can
simply set/unset the HCI_ADVERTISING flag in the
set_advertising_complete function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Add flag to track the real advertising state
Johan Hedberg [Tue, 8 Jul 2014 12:07:47 +0000 (15:07 +0300)]
Bluetooth: Add flag to track the real advertising state

Having a single HCI_ADVERTISING flag is problematic since it tries to
track both the real advertising state and the corresponding mgmt
setting. To make the logic simpler and more reliable add a new flag that
only tracks the actual advertising state that has been written to the
controller.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Don't try background scanning if LE is not enabled
Johan Hedberg [Mon, 7 Jul 2014 12:19:50 +0000 (15:19 +0300)]
Bluetooth: Don't try background scanning if LE is not enabled

For adapters that do not support LE and ones where LE hasn't been
enabled we shouldn't be trying to initiate background scanning. This
patch adds an extra check to the hci_update_background_scan() to ensure
that we bail out if HCI_LE_ENABLED is not set.

Since we do allow user space to feed the kernel with LE connection
parameters even when LE is not enabled we now need to also call
hci_update_background_scan() as soon as LE gets enabled so that scanning
gets started if necessary.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Pass desired connection role to hci_connect_le()
Johan Hedberg [Mon, 7 Jul 2014 12:02:28 +0000 (15:02 +0300)]
Bluetooth: Pass desired connection role to hci_connect_le()

If we have both LE scanning and advertising simultaneously enabled we
need a way to tell hci_connect_le() in which role to initiate a
connection. This patch adds a new parameter to the function to give it
the necessary information. For auto-connect and mgmt_pair_device we
always use master role, whereas for L2CAP users (in practice sockets) we
use slave role whenever HCI_ADVERTISING is set and master role
otherwise.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove auth_type parameter from hci_connect_le()
Johan Hedberg [Mon, 7 Jul 2014 12:02:27 +0000 (15:02 +0300)]
Bluetooth: Remove auth_type parameter from hci_connect_le()

The auth_type value which gets assigned to hci_conn->auth_type is
something that's only used for BR/EDR connections and is of no value for
LE connections. It makes therefore little sense to pass it to the
hci_connect_le() function. This patch removes the parameter from the
function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Enable passive scanning whenever we're connectable
Johan Hedberg [Mon, 7 Jul 2014 11:40:22 +0000 (14:40 +0300)]
Bluetooth: Enable passive scanning whenever we're connectable

Enabling passive scanning always when we're connectable aligns us with
the BR/EDR page scanning. This is also consistent with the fact that the
code dealing with passive scanning results will actively try to connect
any direct advertising event when we're connectable.

This patch implements the feature by adding the connectable condition to
hci_update_background_scan() checks for starting scanning and by calling
hci_update_background_scan() whenever the connectable state changes.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Don't let background scanning interfering with discovery
Johan Hedberg [Mon, 7 Jul 2014 10:24:59 +0000 (13:24 +0300)]
Bluetooth: Don't let background scanning interfering with discovery

If we have an active discovery going on we shouldn't do any changes to
LE scanning when hci_update_background_scan() is called (a call which
can happen for many different reasons). This patch fixes the issue by
returning from the function if the discovery state is anything else
except DISCOVERY_STOPPED.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Update discovery state earlier in hci_discovery_set_state
Johan Hedberg [Mon, 7 Jul 2014 10:24:58 +0000 (13:24 +0300)]
Bluetooth: Update discovery state earlier in hci_discovery_set_state

In a subsequent patch the hci_update_background_scan() function will
depend on being able to know the current discovery state. For this to be
possible we need to set the new state early in the function. Since we
also need to check what the old state was this patch introduces an extra
variable for tracking it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unnecessary return value from check_pending_le_conn
Johan Hedberg [Mon, 7 Jul 2014 09:45:55 +0000 (12:45 +0300)]
Bluetooth: Remove unnecessary return value from check_pending_le_conn

Since the only caller of this function doesn't care about the return
value anymore let's just remove it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove redundant IRK lookup in mgmt_device_found()
Johan Hedberg [Mon, 7 Jul 2014 09:45:54 +0000 (12:45 +0300)]
Bluetooth: Remove redundant IRK lookup in mgmt_device_found()

Now that we have the process_adv_report() function doing the IRK lookup
and updating the bdaddr we don't need to do this anymore in mgmt.c in
the mgmt_device_found() function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix connecting devices during LE device discovery
Johan Hedberg [Mon, 7 Jul 2014 09:45:53 +0000 (12:45 +0300)]
Bluetooth: Fix connecting devices during LE device discovery

If we have devices set as to be connected we should connect to them even
during normal discovery if we get a connectable advertising event. If we
also have HCI_CONNECTABLE set we should connect ADV_DIRECT_IND events
even to devices that we don't have in our pend_le_conns list. This patch
implements such behavior by passing the advertising report type to
check_pending_le_conn() and calls that function regardless of what type
of scanning we are doing.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Use BTUSB_BROKEN_ISOC flag for CSR USB sniffer devices
Marcel Holtmann [Sun, 6 Jul 2014 22:12:04 +0000 (00:12 +0200)]
Bluetooth: Use BTUSB_BROKEN_ISOC flag for CSR USB sniffer devices

Instead of setting data->isoc manually, use BTUSB_BROKEN_ISOC to
indicate that isochronous endpoints are not needed for CSR USB
sniffer devices.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Enforce providing hdev->send driver callback
Marcel Holtmann [Sun, 6 Jul 2014 13:50:27 +0000 (15:50 +0200)]
Bluetooth: Enforce providing hdev->send driver callback

The hdev->send driver callback is mandatory to be provided by a driver
before calling hci_register_dev. So enforce it and return EINVAL in
case it is not available.

All existing drivers are providing this callback anyway, so this is
just an extra sanity check.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Fix memory leaking when hdev->send returns an error
Marcel Holtmann [Sun, 6 Jul 2014 13:36:15 +0000 (15:36 +0200)]
Bluetooth: Fix memory leaking when hdev->send returns an error

The drivers are allowed to just return an error from hdev->send callback
and in that case the driver does not own the SKB. Which means that the
caller has to free the SKB.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Ignore isochronous endpoints for Intel USB bootloader
Marcel Holtmann [Sun, 6 Jul 2014 12:53:55 +0000 (14:53 +0200)]
Bluetooth: Ignore isochronous endpoints for Intel USB bootloader

The isochronous endpoints are not valid when the Intel Bluetooth
controller boots up in bootloader mode. So just mark these endpoints
as broken and then they will not be configured.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Handle Intel USB bootloader with buggy interrupt
Marcel Holtmann [Sun, 6 Jul 2014 12:53:54 +0000 (14:53 +0200)]
Bluetooth: Handle Intel USB bootloader with buggy interrupt

The interrupt interface for the Intel USB bootloader devices is only
enabled after receiving SetInterface(0, AltSetting=0). When this USB
command is not send, then no HCI events will be received.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Fix updating background scan for LE connect complete
Johan Hedberg [Sun, 6 Jul 2014 12:44:23 +0000 (15:44 +0300)]
Bluetooth: Fix updating background scan for LE connect complete

When we get an LE connection complete event we should restart background
scanning if there are any devices needing it. So far the code was only
making the decision based on whether the completed connection had any
stored parameters or not. This patch ensures that we trigger background
scanning always when necessary.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove module parameters for ignoring USB devices
Marcel Holtmann [Sun, 6 Jul 2014 12:27:39 +0000 (14:27 +0200)]
Bluetooth: Remove module parameters for ignoring USB devices

The module parameters to ignore devices based on USB VID/PID are not
needed at all. So just remove them.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Skip unconfigured init procedure for raw-only devices
Marcel Holtmann [Sun, 6 Jul 2014 11:43:20 +0000 (13:43 +0200)]
Bluetooth: Skip unconfigured init procedure for raw-only devices

When the driver sets HCI_QUIRK_RAW_DEVICE, the controller will be set
as unconfigured. However running the unconfigured init procecure is
not useful since raw-only devices are not allowed to change its
configuration. This change skips the init procedure and just allows
user channel operation for this device.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for Intel bootloader devices
Marcel Holtmann [Sun, 6 Jul 2014 11:29:58 +0000 (13:29 +0200)]
Bluetooth: Add support for Intel bootloader devices

Intel Bluetooth devices that boot up in bootloader mode can not
be used as generic HCI devices, but their HCI transport is still
valuable and so bring that up as raw-only devices.

T:  Bus=02 Lev=02 Prnt=03 Port=00 Cnt=01 Dev#= 14 Spd=12   MxCh= 0
D:  Ver= 1.10 Cls=ff(vend.) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=8087 ProdID=0a5a Rev= 0.00
S:  Manufacturer=Intel(R) Corporation
S:  Product=Intel(R) Wilkins Peak 2x2
S:  SerialNumber=001122334455 WP_A0
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=100mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=81(I) Atr=03(Int.) MxPS=  64 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=ff(vend.) Sub=00 Prot=00 Driver=(none)
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Use lower timeout for LE auto-connections
Johan Hedberg [Sun, 6 Jul 2014 10:41:15 +0000 (13:41 +0300)]
Bluetooth: Use lower timeout for LE auto-connections

When we establish connections as a consequence of receiving an
advertising report it makes no sense to wait the normal 20 second LE
connection timeout. This patch modifies the hci_connect_le function to
take an extra timeout value and uses a lower 2 second timeout for the
auto-connection case. This timeout is intentionally chosen to be just a
bit higher than the 1.28 second timeout that High Duty Cycle Advertising
uses.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Clear HCI_RAW flag when controller becomes configured
Marcel Holtmann [Sun, 6 Jul 2014 10:11:16 +0000 (12:11 +0200)]
Bluetooth: Clear HCI_RAW flag when controller becomes configured

When an unconfigured controllers reaches the configured state, it is
important to change the HCI_RAW flag. It indicates to userspace that
the controller is fully operational.

External configuration allows to bring the controller back into an
unconfigured state. In that case make sure HCI_RAW flag is set again.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for changing the public device address
Marcel Holtmann [Sun, 6 Jul 2014 10:11:15 +0000 (12:11 +0200)]
Bluetooth: Add support for changing the public device address

This adds support for changing the public device address. This feature
is required by controllers that do not provide a public address and
have HCI_QUIRK_INVALID_BDADDR set.

Even if a controller has a public device address, this is useful when
an embedded system wants to use its own value. As long as the driver
provides the set_bdaddr callback, this allows changing the device
address before powering on the controller.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Run controller setup after external configuration
Marcel Holtmann [Sun, 6 Jul 2014 10:11:14 +0000 (12:11 +0200)]
Bluetooth: Run controller setup after external configuration

When the external configuration triggers the switch to a configured
controller, it means the setup needs to be run. Controllers that start
out unconfigured have only run limited set of HCI commands. This is
not enough for complete operation and thus run the setup procedure
before announcing the new controller index.

This introduces HCI_CONFIG flag as companion to HCI_SETUP flag. The
HCI_SETUP flag is only used once for the initial setup procedure. And
during that procedure hdev->setup driver callback is called. With the
new HCI_CONFIG the switch from unconfigured to configured state is
triggering the same setup procedure just without hdev->setup. This
is required since bringing a controller back to unconfigured state
from configured state is possible.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Fix sending Device Removed when clearing all parameters
Johan Hedberg [Sun, 6 Jul 2014 10:06:51 +0000 (13:06 +0300)]
Bluetooth: Fix sending Device Removed when clearing all parameters

When calling Device Remove with BDADDR_ANY we should in a similar way
emit Device Removed events as we do when removing a single device. Since
we have to iterate the list and call device_removed() the dedicated
hci_conn_params_clear_enabled() is not really useful anymore. This patch
removes the helper function and does the event emission and list item
removal in a single loop.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Restrict blocked device check in l2cap_recv_frame to LE
Johan Hedberg [Sun, 6 Jul 2014 08:03:36 +0000 (11:03 +0300)]
Bluetooth: Restrict blocked device check in l2cap_recv_frame to LE

BR/EDR has the connection request and connection request rejection, but
LE doesn't have anything similar. We still request LE connections to
blocked devices to be disconnected but it's possible that ACL data slips
through before that. The check in l2cap_recv_frame really only needs to
be for LE and not BR/EDR because of this.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Store Bluetooth address from controller setup
Marcel Holtmann [Sat, 5 Jul 2014 08:48:03 +0000 (10:48 +0200)]
Bluetooth: Store Bluetooth address from controller setup

During the setup phase of a controller, the Bluetooth address will be
read and to have that original address available for later use, store
it as setup address.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Run special init procedure for unconfigured controllers
Marcel Holtmann [Sat, 5 Jul 2014 08:48:02 +0000 (10:48 +0200)]
Bluetooth: Run special init procedure for unconfigured controllers

For an unconfigured controller it is required to read at least the
local version information. If the set_bdaddr driver callback is
provideded, then also the local Bluetooth address will be read.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Use a more simpler style for HCI event callbacks
Marcel Holtmann [Sat, 5 Jul 2014 08:48:01 +0000 (10:48 +0200)]
Bluetooth: Use a more simpler style for HCI event callbacks

The HCI event callbacks have grown over the last years and some
functions handle status checking different than others. For the
simple ones, check the status at the beginning and exit if an
error with the HCI command occured.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for New Configuration Options management event
Marcel Holtmann [Fri, 4 Jul 2014 17:06:23 +0000 (19:06 +0200)]
Bluetooth: Add support for New Configuration Options management event

When one or more of the missing configuration options change, then send
this even to all the other management interface clients.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Move mgmt_event helper function to different location
Marcel Holtmann [Fri, 4 Jul 2014 17:06:22 +0000 (19:06 +0200)]
Bluetooth: Move mgmt_event helper function to different location

Move the mgmt_event function higher up in the code so that no forward
declaration is needed.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for Set External Configuration management command
Marcel Holtmann [Fri, 4 Jul 2014 16:11:55 +0000 (18:11 +0200)]
Bluetooth: Add support for Set External Configuration management command

The Set External Configuration management command allows for switching
between configured and unconfigured start if HCI_QURIK_EXTERNAL_CONFIG
is set by the transport driver.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Support HCI_QUIRK_EXTERNAL_CONFIG for hci_vhci driver
Marcel Holtmann [Fri, 4 Jul 2014 15:23:35 +0000 (17:23 +0200)]
Bluetooth: Support HCI_QUIRK_EXTERNAL_CONFIG for hci_vhci driver

This adds support for configuring the hci_vhci virtual controllers
to require a setup stage using HCI_QUIRK_EXTERNAL_CONFIG. With this
option the virtual controller will start out as unconfigured.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Fix quirks that are valid during setup driver callback
Marcel Holtmann [Fri, 4 Jul 2014 15:23:34 +0000 (17:23 +0200)]
Bluetooth: Fix quirks that are valid during setup driver callback

For the quirks that are allow to be set during setup callback, the
check needs to be modified so that they are applied even if no
setup callback provided by the driver.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add quirk for external configuration requirement
Marcel Holtmann [Fri, 4 Jul 2014 15:23:33 +0000 (17:23 +0200)]
Bluetooth: Add quirk for external configuration requirement

When a controller requires external configuration, then setting this
quirk will allow indicating this.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Allow reading configuration info when unconfigured
Marcel Holtmann [Fri, 4 Jul 2014 14:54:40 +0000 (16:54 +0200)]
Bluetooth: Allow reading configuration info when unconfigured

Reading the controller configuration information is a valid management
command when the controller is unconfigured. Allow this command, but
return invalid index on all other commands.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Set HCI_QUIRK_INVALID_BADDR for BCM20702A0 default address
Marcel Holtmann [Fri, 4 Jul 2014 14:54:39 +0000 (16:54 +0200)]
Bluetooth: Set HCI_QUIRK_INVALID_BADDR for BCM20702A0 default address

When the Broadcom USB controller has a default address, then set the quirk
so the Bluetooth core knows that controller configuration is required.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Set HCI_QUIRK_INVALID_BADDR for Intel USB default address
Marcel Holtmann [Fri, 4 Jul 2014 14:54:38 +0000 (16:54 +0200)]
Bluetooth: Set HCI_QUIRK_INVALID_BADDR for Intel USB default address

When the Intel USB controller has a default address, then set the quirk
so the Bluetooth core knows that controller configuration is required.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add quirk for invalid controller address setting
Marcel Holtmann [Fri, 4 Jul 2014 14:54:37 +0000 (16:54 +0200)]
Bluetooth: Add quirk for invalid controller address setting

When a Bluetooth controller does not have a valid public Bluetooth
address, then allow the driver to indicate this. If the quirk is
set, the Bluetooth core will switch to unconfigured state first
and will allow userspace to configure the address before starting
the full initialization of the controller.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Document the existing device quirks
Marcel Holtmann [Fri, 4 Jul 2014 14:08:02 +0000 (16:08 +0200)]
Bluetooth: Document the existing device quirks

The current existing device quirks are not documented. So instead of
spreading bits and pieces somewhere in the code, add proper comments
on where these quirks can be used and what behavior they change.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Use list_del when freeing the list entry
Johan Hedberg [Fri, 4 Jul 2014 13:17:23 +0000 (16:17 +0300)]
Bluetooth: Use list_del when freeing the list entry

It's wasteful to use list_del_init (which re-initializes the list_head)
if we're just about to free the element and never use it again.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Don't take actions on blocked devices when scanning
Johan Hedberg [Fri, 4 Jul 2014 13:15:49 +0000 (16:15 +0300)]
Bluetooth: Don't take actions on blocked devices when scanning

If a found device is marked as blocked while doing passive LE scanning,
neither report it nor try to connect to it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix constant for public address configuration
Marcel Holtmann [Fri, 4 Jul 2014 11:23:38 +0000 (13:23 +0200)]
Bluetooth: Fix constant for public address configuration

The public address configuration option is value 0x02 since the generic
external configuration is value 0x01. So adjust this accordingly and
also add the value 0x01 to the list.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Use hci_pend_le_action_lookup to look up report entries
Johan Hedberg [Fri, 4 Jul 2014 09:37:27 +0000 (12:37 +0300)]
Bluetooth: Use hci_pend_le_action_lookup to look up report entries

Instead of looking through the entire list of entries we can more
efficiently use the new hci_pend_le_action_lookup() function to look up
entries specifically in the pend_le_reports list. Since the search is
now limited to the right list we can also remove an unnecessary check
for list_empty() before the lookup.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Make hci_pend_le_conn_lookup more general purposed
Johan Hedberg [Fri, 4 Jul 2014 09:37:26 +0000 (12:37 +0300)]
Bluetooth: Make hci_pend_le_conn_lookup more general purposed

In some circumstances we need to look up entries in pend_le_conns and in
other in pend_le_reports. This patch converts the existing lookup
function for pend_le_conns to something that can be used for both lists.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unused hci_pend_le_conn_add function
Johan Hedberg [Fri, 4 Jul 2014 09:37:25 +0000 (12:37 +0300)]
Bluetooth: Remove unused hci_pend_le_conn_add function

Since there are no more users of this function we can simply go ahead
and remove it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unnecessary usage of hci_pend_le_conn_add
Johan Hedberg [Fri, 4 Jul 2014 09:37:24 +0000 (12:37 +0300)]
Bluetooth: Remove unnecessary usage of hci_pend_le_conn_add

This is the last place using hci_pend_le_conn_add() and we can more
just as simply manipulate the list directly here.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix clearing and restarting all LE actions on power cycle
Johan Hedberg [Fri, 4 Jul 2014 09:37:23 +0000 (12:37 +0300)]
Bluetooth: Fix clearing and restarting all LE actions on power cycle

When powering off (hci_dev_do_close) we should clear both the
pend_le_reports and pend_le_conns types of entries. When powering on
respectively we should populate both lists. This patch converts the
hci_pend_le_conns_clear() function into hci_pend_le_actions_clear()
(which can now be static) and converts the restart_le_auto_conns()
function into restart_le_actions().

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unused hci_pend_le_conn_del() function
Johan Hedberg [Fri, 4 Jul 2014 09:37:22 +0000 (12:37 +0300)]
Bluetooth: Remove unused hci_pend_le_conn_del() function

Now that there are no-longer any users of the hci_pend_le_conn_del()
function we can simply go ahead and remove it.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Simplify hci_conn_params->action list usage
Johan Hedberg [Fri, 4 Jul 2014 09:37:21 +0000 (12:37 +0300)]
Bluetooth: Simplify hci_conn_params->action list usage

Since params->action is used for both the pend_le_conns and
pend_le_reports lists we can simplify the adding and deleting of the
lists considerably. For example, when deleting entries in most
situations we no-longer need to check the auto_connect value but can
directly proceed with calling list_del_init on param->action (which is
safe even if the entry is not part of any list).

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Don't bother doing anything if auto_connect doesn't change
Johan Hedberg [Fri, 4 Jul 2014 09:37:20 +0000 (12:37 +0300)]
Bluetooth: Don't bother doing anything if auto_connect doesn't change

When hci_conn_params_set() is called if the new auto_connect value is
the same as the old one we don't need to take any action. Simply return
success from the function in this case.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Simplify use of hci_pend_le_conns_clear()
Johan Hedberg [Fri, 4 Jul 2014 09:37:19 +0000 (12:37 +0300)]
Bluetooth: Simplify use of hci_pend_le_conns_clear()

Now that pend_le_connections is a list of hci_conn_params entries we can
simply remove items from that list as we iterate through the global list
of le_conn_params. This also moves the responsibility of calling
hci_update_background_scan() to the functions that were previously
calling hci_pend_le_conns_clear(). The only user that's left for
hci_pend_le_conns_clear() is hci_dev_do_close() which anyway does not
need to call hci_update_background_scan().

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Convert pend_le_reports into a list
Johan Hedberg [Fri, 4 Jul 2014 09:37:18 +0000 (12:37 +0300)]
Bluetooth: Convert pend_le_reports into a list

To simplify manipulation and lookup of hci_conn_params entries of the
type HCI_AUTO_CONN_REPORT it makes sense to store them in their own
list. The new action list_head in hci_conn_params is used for this
purpose.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Convert pend_le_conn list to a generic action list
Johan Hedberg [Fri, 4 Jul 2014 09:37:17 +0000 (12:37 +0300)]
Bluetooth: Convert pend_le_conn list to a generic action list

In preparation to store also HCI_AUTO_CONN_REPORT entries in a list it
makes sense to convert the existing pend_le_conn list head of
hci_conn_params into a more generically named "action". This makes sense
because a parameter entry will never participate in more than one action
list.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix missing return statement in process_adv_report
Johan Hedberg [Fri, 4 Jul 2014 09:37:16 +0000 (12:37 +0300)]
Bluetooth: Fix missing return statement in process_adv_report

If we're doing passive scanning we shouldn't proceed with any of the
code that deals with active scanning (pending reports, etc.). This patch
fixes a missing return statement for the passive scanning section in the
process_adv_report() function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Add support for controller configuration info command
Marcel Holtmann [Thu, 3 Jul 2014 22:46:56 +0000 (00:46 +0200)]
Bluetooth: Add support for controller configuration info command

The Read Controller Configuration Information command allows retrieving
details about possible configurations option. The supported options are
returned and also the missing options (if any).

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add identity address check in param lookup functions
Johan Hedberg [Thu, 3 Jul 2014 16:33:51 +0000 (19:33 +0300)]
Bluetooth: Add identity address check in param lookup functions

Since we only store entries with identity addresses in the
le_conn_params and pend_le_conns lists we can avoid unnecessary lookups
by checking for an identity address before diving into the lists
themselves.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove unnecessary checks for auto-connected devices
Johan Hedberg [Thu, 3 Jul 2014 16:33:50 +0000 (19:33 +0300)]
Bluetooth: Remove unnecessary checks for auto-connected devices

If a device is in the pend_le_conns list it cannot at the same time also
have the need to be notified through mgmt_device_found. By making
check_pending_le_conn return whether it found an entry or not we can
avoid unnecessary checks in process_adv_report().

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Use hci_conn_params in pend_le_conns
Johan Hedberg [Thu, 3 Jul 2014 16:33:49 +0000 (19:33 +0300)]
Bluetooth: Use hci_conn_params in pend_le_conns

Since the connection parameters are always a basis for adding entries to
hdev->pend_le_conns (so far of type bdaddr_list) it's simpler and more
efficient to have the parameters themselves be the entries in the
pend_le_conns list. We do this by adding another list_head to the
hci_conn_params struct.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Remove redundant IRK lookup
Johan Hedberg [Thu, 3 Jul 2014 16:33:48 +0000 (19:33 +0300)]
Bluetooth: Remove redundant IRK lookup

When processing passive scanning results we need the resolved identity
address both in check_pending_le_conn() as well as later in
process_adv_report(). Since process_adv_report() calls
check_pending_le_conn() we can simply resolve the IRK earlier in the
function and thereby eliminate a second lookup.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix missing update of pend_le_reports
Johan Hedberg [Thu, 3 Jul 2014 16:33:47 +0000 (19:33 +0300)]
Bluetooth: Fix missing update of pend_le_reports

When calling Remove Device for an entry using HCI_AUTO_CONN_REPORT we
need to decrement the pend_le_reports value correspondingly. This patch
fixes one such missing action in the Remove Device command handler.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Fix buffer overflow with variable length commands
Johan Hedberg [Thu, 3 Jul 2014 10:52:27 +0000 (13:52 +0300)]
Bluetooth: Fix buffer overflow with variable length commands

The handler for variable length commands were trying to calculate the
expected length of the command based on the given parameter count, and
then comparing that with the received data. However, the expected count
was stored in a u16 which can easily overflow. With a carefully crafted
command this can then be made to match the given data even though the
parameter count is actually way too big, resulting in a buffer overflow
when parsing the parameters.

This patch fixes the issue by calculating a per-command maximum
parameter count and returns INVALID_PARAMS if it is exceeded.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Support scanning for devices using RPA
Johan Hedberg [Wed, 2 Jul 2014 20:09:24 +0000 (23:09 +0300)]
Bluetooth: Support scanning for devices using RPA

When we're scanning for specific devices that use an RPA we need to
convert the RPA to the identity address before looking up the entry in
the connection parameters. This patch adds the necessary code to do this
in the process_adv_report() function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Add support for background LE scanning
Johan Hedberg [Wed, 2 Jul 2014 19:42:02 +0000 (22:42 +0300)]
Bluetooth: Add support for background LE scanning

If we have one or more devices with HCI_AUTO_CONN_REPORT we should do
background scanning and emit mgmt_device_found events. This patch
modifies the hci_update_background_scan() function to extend the
conditions needed to trigger scanning, and adds the necessary code to
process_adv_report() to emit mgmt_device_found events.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Allow mgmt_device_found events for kernel-side scanning
Johan Hedberg [Wed, 2 Jul 2014 19:42:01 +0000 (22:42 +0300)]
Bluetooth: Allow mgmt_device_found events for kernel-side scanning

When the kernel is doing LE scanning because of one or more devices
added with action 0x00 through the Add Device command we do want to let
mgmt_device_found() to proceed with sending an event. This kind of
devices are tracked with hdev->pend_le_reports, so check this value
before bailing out from the function.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Track number of added devices with HCI_AUTO_CONN_REPORT
Johan Hedberg [Wed, 2 Jul 2014 19:42:00 +0000 (22:42 +0300)]
Bluetooth: Track number of added devices with HCI_AUTO_CONN_REPORT

To be able to make the right choice of whether to start passive scanning
or to send out a mgmt_device_found event we need to know if there are
any devices in the le_conn_params list with the auto_connect value set
to HCI_AUTO_CONN_REPORT. This patch adds a counter for this kind of
devices.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
9 years agoBluetooth: Support HCI_QUIRK_RAW_DEVICE for hci_vhci driver
Marcel Holtmann [Wed, 2 Jul 2014 23:35:10 +0000 (01:35 +0200)]
Bluetooth: Support HCI_QUIRK_RAW_DEVICE for hci_vhci driver

This adds support for configuring the hci_vhci virtual controllers
as a raw-only device using HCI_QUIRK_RAW_DEVICE. This is useful for
testing the kernel internal infrastructure.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for Read Unconfigured Index List command
Marcel Holtmann [Wed, 2 Jul 2014 20:10:52 +0000 (22:10 +0200)]
Bluetooth: Add support for Read Unconfigured Index List command

This command allows to get the list of currently known controller that
are in unconfigured state.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for Unconfigured Index Removed events
Marcel Holtmann [Wed, 2 Jul 2014 19:30:55 +0000 (21:30 +0200)]
Bluetooth: Add support for Unconfigured Index Removed events

When a controller in an unconfigured state gets removed, then send
Unconfigured Index Removed events.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Add support for Unconfigured Index Added events
Marcel Holtmann [Wed, 2 Jul 2014 19:30:54 +0000 (21:30 +0200)]
Bluetooth: Add support for Unconfigured Index Added events

When a controller is in unconfigured state it is currently hidden
from the management interface. This change now announces the new
controller with an Unconfigured Index Added event and allows clients
to easily detect the controller.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
9 years agoBluetooth: Introduce unconfigured controller state
Marcel Holtmann [Wed, 2 Jul 2014 17:10:33 +0000 (19:10 +0200)]
Bluetooth: Introduce unconfigured controller state

With the new unconfigured controller state it is possible to provide a
fully functional HCI transport, but disable the higher level operations
that would normally happen. This way userspace can try to configure the
controller before releases the unconfigured state.

The internal state is represented by HCI_UNCONFIGURED. This replaces the
HCI_QUIRK_RAW_DEVICE quirk as internal state representation. This is now
a real state and drivers can use the quirk to actually trigger this
state. In the future this will allow a more fine grained switching from
unconfigured state to configured state for controller inititialization.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>