platform/upstream/bluez.git
2 years agolib/mgmt: Adding Add Adv Patterns Monitor RSSI opcode
Archie Pusaka [Fri, 15 Jan 2021 11:50:34 +0000 (19:50 +0800)]
lib/mgmt: Adding Add Adv Patterns Monitor RSSI opcode

The new op is to utilize RSSI in advertisement monitor

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/crypto: Transform bt_cryto in a singleton
Luiz Augusto von Dentz [Fri, 15 Jan 2021 18:52:42 +0000 (10:52 -0800)]
shared/crypto: Transform bt_cryto in a singleton

This makes bt_crypto always return the same instance thus avoiding the
problem of having multiple PF sockets opened which may hit the file
descriptor limit.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoemulator: Fix sending CIS Requested to Create CIS device
Luiz Augusto von Dentz [Thu, 14 Jan 2021 20:54:13 +0000 (12:54 -0800)]
emulator: Fix sending CIS Requested to Create CIS device

CIS Requested shall be send to the remote peer not the originating one.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agogatt: Fix assuming service changed has been subscribed
Luiz Augusto von Dentz [Fri, 8 Jan 2021 21:08:25 +0000 (13:08 -0800)]
gatt: Fix assuming service changed has been subscribed

Unfortunately assuming service changed has been subscribed may cause
indication to time out in some peripherals (Logitech M720 Triathlon, Mx
Anywhere 2, Lenovo Mice N700, RAPOO BleMouse and Microsoft Designer
Mouse) even though the expect actually mandates that the client responds
with confirmation these peripherals just ignores it completely which
leads them to be disconnected whenever bluetoothd is restarted or the
system reboots.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodevice: Enable ATT layer debugging
Luiz Augusto von Dentz [Fri, 8 Jan 2021 18:57:04 +0000 (10:57 -0800)]
device: Enable ATT layer debugging

This uses bt_att_set_debug to enable ATT debugging which is useful for
detecting error such as an ATT transaction timing out.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/att: Add debug level to bt_att_set_debug
Luiz Augusto von Dentz [Fri, 8 Jan 2021 18:54:24 +0000 (10:54 -0800)]
shared/att: Add debug level to bt_att_set_debug

This creates different levels of debugging which can be passed to
bt_att_set_debug as depending on the application it may not need to
print everything which can pollute the logs quite a bit.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoutil: Introduce util_debug_va
Luiz Augusto von Dentz [Fri, 8 Jan 2021 18:51:58 +0000 (10:51 -0800)]
util: Introduce util_debug_va

This introduces util_debug_va which can take a va_list that enables
callers to create wrapper functions if they need to.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/gatt-server: Fix not accepting multiple requests
Luiz Augusto von Dentz [Wed, 6 Jan 2021 23:22:41 +0000 (15:22 -0800)]
shared/gatt-server: Fix not accepting multiple requests

With introduction of EATT it is possible to send multiple requests using
different channels therefore bt_gatt_server shall not attempt to check
if there is already an outstanding request.

gh issue: https://github.com/bluez/bluez/issues/75

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoadapter: Remove temporary devices before power off
Archie Pusaka [Wed, 6 Jan 2021 09:26:58 +0000 (17:26 +0800)]
adapter: Remove temporary devices before power off

If adapter is powered off when a currently connected device is
being removed, there is a possibility that we haven't finish waiting
for the disconnection but the adapter is already powered down.

When this happens, the kernel would fail to clean the device's
information, for example the pairing information. This causes
disagreement between the user space and the kernel about whether the
device is already paired, because the device is successfully removed
from the user space's perspective.

This patch enforces the removal of such devices before allowing the
adapter to power off.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agogatt: Make use of gatt_db_attribute_set_fixed_length
Luiz Augusto von Dentz [Wed, 6 Jan 2021 00:45:37 +0000 (16:45 -0800)]
gatt: Make use of gatt_db_attribute_set_fixed_length

This makes use of gatt_db_attribute_set_fixed_length so the database is
aware of the length of the values and perform bounds checking.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/gatt-db: Introduce gatt_db_attribute_set_fixed_length
Luiz Augusto von Dentz [Wed, 6 Jan 2021 00:38:03 +0000 (16:38 -0800)]
shared/gatt-db: Introduce gatt_db_attribute_set_fixed_length

This enables user to inform if an attribute has a fixed length so it can
automatically perform bounds checking.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/mgmt: Fix memory leak in mgmt_tlv_list
Howard Chung [Tue, 5 Jan 2021 03:12:57 +0000 (11:12 +0800)]
shared/mgmt: Fix memory leak in mgmt_tlv_list

This patch freed the mgmt_tlv properly in mgmt_tlv_list_free.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtmgmt: Add command set-sysconfig
Howard Chung [Tue, 5 Jan 2021 03:12:56 +0000 (11:12 +0800)]
btmgmt: Add command set-sysconfig

Add command set-sysconfig in btmgmt

Example usage:
[mgmt]# set-sysconfig -h
Parameters:
        -v <type:length:value>...
e.g.:
        set-sysconfig -v 001a:2:1234 001f:1:00
[mgmt]# set-sysconfig -v 8:2:abcd 1:02:0100 0016:2:0600

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtmgmt: Add command read-sysconfig
Howard Chung [Tue, 5 Jan 2021 03:12:55 +0000 (11:12 +0800)]
btmgmt: Add command read-sysconfig

Add command read-sysconfig in btmgmt

Example usage:
localhost ~ # btmgmt
[mgmt]# read-sysconfig
Type: 0x0000    Length: 02      Value: 0000
Type: 0x0001    Length: 02      Value: 0008
...
Type: 0x001f    Length: 01      Value: 01

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/mgmt: Add supports of parsing mgmt tlv list
Howard Chung [Tue, 5 Jan 2021 03:12:54 +0000 (11:12 +0800)]
shared/mgmt: Add supports of parsing mgmt tlv list

Response from Read System Default Configuration is a list of mgmt_tlv,
which requires further processing to get the values of each parameters.

This adds APIs for parsing response into mgmt_tlv_list, retrieving
parameter from mgmt_tlv_list.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agogatt: Fix potential buffer out-of-bound
Luiz Augusto von Dentz [Mon, 4 Jan 2021 18:38:31 +0000 (10:38 -0800)]
gatt: Fix potential buffer out-of-bound

When client features is read check if the offset is within the cli_feat
bounds.

Fixes: https://github.com/bluez/bluez/issues/70
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agocfg: Add check for newly used function in ELL
Brian Gix [Mon, 4 Jan 2021 18:33:53 +0000 (10:33 -0800)]
cfg: Add check for newly used function in ELL

ELL has a new dependency on rawmemchr() which needs to be accounted for
at configure time.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomesh: Update AppKeys on transition to Phase 0
Michael N. Moran [Mon, 4 Jan 2021 03:48:37 +0000 (22:48 -0500)]
mesh: Update AppKeys on transition to Phase 0

At the end of the mesh Key Refresh procedure when a subnet
transitions to Phase 0, local AppKeys that were updated were
not updating until the bluetooth-meshd daemon was restarted.

This patch iterates the AppKeys at the end of mesh Key Refresh
when the subnet transitions to Phase 0, setting the new state
of each updated AppKey.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/mgmt-tester: Adds new test data with updated tx power
Tedd Ho-Jeong An [Wed, 23 Dec 2020 00:26:33 +0000 (16:26 -0800)]
tools/mgmt-tester: Adds new test data with updated tx power

This patch adds new test data for the ext advertising test cases that
presets the advertising parameter and toggles the connectable flags.

When the adv instance is created, it updates the default tx power with
the value read from the controller. After toggling the connectable flag,
the host uses the tx_power stored in the adv list cache, which is the
value read from the controller, instead of the default tx power
value(127).

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoemulator/btdev: Add support LE_READ_TX_POWER command
Tedd Ho-Jeong An [Wed, 23 Dec 2020 00:26:32 +0000 (16:26 -0800)]
emulator/btdev: Add support LE_READ_TX_POWER command

This patch adds support for LE_READ_TX_POWER command.
It uses a random values for min and max tx power since these values is
read from the controller.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/mgmt-tester: Update sample data for adv features
Tedd Ho-Jeong An [Wed, 23 Dec 2020 00:26:31 +0000 (16:26 -0800)]
tools/mgmt-tester: Update sample data for adv features

This patch updates the expected sample data for advertising feature to
include the changes for tx power.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtmon: fix buffer bound checks
Jacob Siverskog [Tue, 22 Dec 2020 20:45:47 +0000 (21:45 +0100)]
btmon: fix buffer bound checks

index_list is of size MAX_INDEX - correct the checks that is meant to
catch out-of-bounds access.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodoc/mgmt-api: Add opcode for adding advertisement monitor with RSSI
Archie Pusaka [Tue, 15 Dec 2020 10:10:36 +0000 (18:10 +0800)]
doc/mgmt-api: Add opcode for adding advertisement monitor with RSSI

This is to leverage the filtering by RSSI feature on those controllers
which supports advertisement packet filtering. To avoid changing the
existing API and breaking it, a new opcode is required.

Reviewed-by: Manish Mandlik <mmandlik@chromium.org>
Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
Reviewed-by: Yun-Hao Chung <howardchung@google.com>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobthost: Add support for address types other then public
Luiz Augusto von Dentz [Thu, 17 Dec 2020 23:18:29 +0000 (15:18 -0800)]
bthost: Add support for address types other then public

This properly initialize the address type according to the connection
address.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agohciemu: Add support for multiple clients
Luiz Augusto von Dentz [Wed, 25 Nov 2020 21:13:04 +0000 (13:13 -0800)]
hciemu: Add support for multiple clients

This adds support for creating multiple clients (bthost).

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtdev: Add support for multiple connections
Luiz Augusto von Dentz [Thu, 19 Nov 2020 23:58:53 +0000 (15:58 -0800)]
btdev: Add support for multiple connections

This adds support for assigning different handles for connection
instead of always using fixed handles.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobthost: Always use conn->prnd for Pairing Random
Luiz Augusto von Dentz [Thu, 17 Dec 2020 18:24:47 +0000 (10:24 -0800)]
bthost: Always use conn->prnd for Pairing Random

Since it is now properly set with random bytes it shall be used when
responding.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest: add test app for Advertisement Monitor API
Manish Mandlik [Tue, 8 Dec 2020 22:58:47 +0000 (14:58 -0800)]
test: add test app for Advertisement Monitor API

The python test app to test working of the Advertisement Monitor API.

This app:
- registers itself with bluez by invoking the RegisterMonitor with the
  app root path
- exposes multiple monitor objects with both valid and invalid monitor
  parameters
- implements Activate/Release/DeviceFound/DeviceLost methods on the
  monitor object

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobthost: Use bt_crypto_random_bytes to generate a Pairing Random
Luiz Augusto von Dentz [Thu, 17 Dec 2020 00:28:00 +0000 (16:28 -0800)]
bthost: Use bt_crypto_random_bytes to generate a Pairing Random

This makes it more consistent with how stacks are implemented instead
of setting the response all zero which is not secure.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomonitor: Add option to disable time offset
Luiz Augusto von Dentz [Wed, 16 Dec 2020 22:58:08 +0000 (14:58 -0800)]
monitor: Add option to disable time offset

This adds option -N/--no-time which disables the time offset which
sometimes is useful when diffing traces where timing information is not
relevant.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtdev: Fix not responding with the same CIG ID
Luiz Augusto von Dentz [Tue, 15 Dec 2020 21:47:00 +0000 (13:47 -0800)]
btdev: Fix not responding with the same CIG ID

BT_HCI_CMD_LE_SET_CIG_PARAMS shall be responded using the same CIG ID.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoinput/hog: Cache the HID report map
Sonny Sasaka [Mon, 14 Dec 2020 22:15:02 +0000 (14:15 -0800)]
input/hog: Cache the HID report map

To optimize BLE HID devices reconnection response, we can cache the
report map so that the subsequent reconnections do not need round trip
time to read the report map.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoRevert "input/hog: Remove HID device after HoG device disconnects"
Sonny Sasaka [Mon, 14 Dec 2020 22:15:01 +0000 (14:15 -0800)]
Revert "input/hog: Remove HID device after HoG device disconnects"

This reverts commit d6cafa1f0c3ea1989f8a580e52f418b0998a3552.

In commit d6cafa1f0c3e ("input/hog: Remove HID device after HoG device
disconnects"), the bt_hog structure is destroyed in order to fix a bug
where the UHID connection is not destroyed. This fix has the cost of
increasing reconnection time because every reconnection would need to
re-read the report map again. An improvement to this fix is, instead of
removing the bt_hog structure, we can just destroy the UHID with
UHID_DESTROY event and use the existing bt_hog structure to keep the
cache of the report map to avoid re-reading the report map at
reconnection.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodoc: Fix qemu unable to mount root fs for test-runner
Tedd Ho-Jeong An [Wed, 25 Nov 2020 07:57:07 +0000 (23:57 -0800)]
doc: Fix qemu unable to mount root fs for test-runner

This patch fixes the kernel panic while loading the kernel image built
with the current config for test-runner.

 Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)
 CPU: 0 PID: 1 Comm: swapper Not tainted 5.10.0-rc2-702835-ged93facee36e #6
 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1 04/01/2014
 Call Trace:
  panic+0xee/0x27c
  mount_block_root+0x315/0x324
  ? rest_init+0x7a/0x7a
  prepare_namespace+0x128/0x157
  ? rest_init+0x7a/0x7a
  kernel_init+0x5/0xf6
  ret_from_fork+0x22/0x30
 Kernel Offset: 0x2a200000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbf)

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtdev: Fix not handling BT_HCI_CMD_HOST_NUM_COMPLETED_PACKETS
Luiz Augusto von Dentz [Tue, 8 Dec 2020 01:06:51 +0000 (17:06 -0800)]
btdev: Fix not handling BT_HCI_CMD_HOST_NUM_COMPLETED_PACKETS

System such as Zephyr do use BT_HCI_CMD_HOST_NUM_COMPLETED_PACKETS
since BT_HCI_CMD_SET_HOST_FLOW_CONTROL is also supported.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoadv_monitor: Update log messages
Manish Mandlik [Wed, 25 Nov 2020 20:18:46 +0000 (12:18 -0800)]
adv_monitor: Update log messages

Update log messages to avoid breaking of user-visible log strings
to make it easy to grep for them.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoclient: Use correct root path for RegisterMonitor
Manish Mandlik [Thu, 3 Dec 2020 03:46:52 +0000 (19:46 -0800)]
client: Use correct root path for RegisterMonitor

Bluetoothctl uses "/" as the app root path. So, use the same
app root path while invoking RegisterMonitor method. This will
also ensure that InterfacesAdded/InterfacesRemoved signals are
emitted on the correct app root path.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoadv_monitor: Register client app with app-root-path
Manish Mandlik [Thu, 3 Dec 2020 03:46:51 +0000 (19:46 -0800)]
adv_monitor: Register client app with app-root-path

When a client app is registered with g_dbus_client_new(), bluez root
path, i.e. "/", is used as the app root path and signal watches are
added at that root path.

Because of this, InterfacesAdded/InterfacesRemoved signals emitted by
app - while creating/removing advertisement monitor objects at the
app root path - are not received by bluetoothd.

Use g_dbus_client_new_full() to register a client app with the correct
app root path.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomain.conf: Fix using wrong field for PageTimeout
Luiz Augusto von Dentz [Tue, 1 Dec 2020 23:10:30 +0000 (15:10 -0800)]
main.conf: Fix using wrong field for PageTimeout

The correct field is btd_opts.defaults.br.page_timeout.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomain.conf: Use BR instead of BREDR as group name
Luiz Augusto von Dentz [Tue, 1 Dec 2020 22:24:50 +0000 (14:24 -0800)]
main.conf: Use BR instead of BREDR as group name

This align with latest version of main.conf that does use [BR] instead
of [BREDR].

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/mgmt-tester: Add tests for individual ext adv MGMT commands
Daniel Winkler [Tue, 24 Nov 2020 17:31:23 +0000 (09:31 -0800)]
tools/mgmt-tester: Add tests for individual ext adv MGMT commands

The new extended advertising MGMT commands are paired, requiring both a
Set Parameters and Set Data command to configure a new advertisement.
This patch adds the following new tests to test these APIs:

1. Verify that if the adapter is unpowered, MGMT_OP_ADD_EXT_ADV_PARAMS
will return a rejected status.

2. Verify that MGMT_OP_ADD_EXT_ADV_PARAMS will return an invalid
parameters status if the data payload is malformed.

3. Verify that MGMT_OP_ADD_EXT_ADV_PARAMS is successful on BT4 if
request is valid by checking MGMT return value.

4. Verify that MGMT_OP_ADD_EXT_ADV_PARAMS is successful on BT5 if
request is valid by checking MGMT return value and that HCI traffic
contains a correct advertising parameters request.

5. Verify that if a MGMT_OP_ADD_EXT_ADV_DATA request is received for an
advertising instance that is unknown, i.e. MGMT_OP_ADD_EXT_ADV_PARAMS
was not yet called, an invalid parameters status is returned.

Tests 6-8 first call MGMT_OP_ADD_EXT_ADV_PARAMS successfully in setup to
then validate the contents of the subsequent MGMT_OP_ADD_EXT_ADV_DATA
call.

6. Verify that if MGMT_OP_ADD_EXT_ADV_DATA is passed an invalid tlv
structure, an invalid parameters status is returned.

7. Verify that if MGMT_OP_ADD_EXT_ADV_DATA contains a valid structure,
the hci traffic contain a correctly formed Adv Data hci request.

8. Verify that if MGMT_OP_ADD_EXT_ADV_DATA contains a valid structure,
the hci traffic contain a correctly formed Adv Scan Response hci
request.

Reviewed-by: Manish Mandlik <mmandlik@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/mgmt-tester: Add mgmt-tester coverage for controller capabilities
Daniel Winkler [Tue, 24 Nov 2020 17:31:22 +0000 (09:31 -0800)]
tools/mgmt-tester: Add mgmt-tester coverage for controller capabilities

This patch adds two test cases for the new Read Controller
Capabilities MGMT command. Since the command itself is simple, the test
cases are simple as well:

1. Ensures that providing invalid parameters to the request will result
in a failed MGMT response.

2. Ensures that if the controller is BT5 compliant, the Read Controller
Capabilities command will return a valid TX power range as expected,
otherwise fail.

Reviewed-by: Manish Mandlik <mmandlik@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobattery: Implement Battery Provider API
Sonny Sasaka [Mon, 30 Nov 2020 21:56:02 +0000 (13:56 -0800)]
battery: Implement Battery Provider API

This patch implements the BatteryProvider1 and BatteryProviderManager1
API. This is a means for external clients to feed battery information to
BlueZ if they handle some profile and can decode battery reporting.

The battery information is then exposed externally via the existing
Battery1 interface. UI components can consume this API to display
Bluetooth peripherals' battery via a unified BlueZ API.

Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoadapter: Add a public function to find a device by path
Sonny Sasaka [Mon, 30 Nov 2020 21:56:01 +0000 (13:56 -0800)]
adapter: Add a public function to find a device by path

The public function is motivated by the Battery Provider API code which
needs to probe whether a device exists.

Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest: Add test app for Battery Provider API
Sonny Sasaka [Mon, 30 Nov 2020 21:56:00 +0000 (13:56 -0800)]
test: Add test app for Battery Provider API

The python test app simulates an application registering to BlueZ as a
Battery Provider providing three fake batteries drained periodically.

Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodoc: Add Battery Provider API doc
Sonny Sasaka [Mon, 30 Nov 2020 21:55:59 +0000 (13:55 -0800)]
doc: Add Battery Provider API doc

This patch add the documentation of the Battery Provider which lets
external clients feed battery information to BlueZ if they are able to
decode battery reporting via any profile. BlueZ UI clients can then use
the org.bluez.Battery1 API as a single source of battery information
coming from many different profiles.

Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobattery: Add Source property to Battery API
Sonny Sasaka [Mon, 30 Nov 2020 21:55:58 +0000 (13:55 -0800)]
battery: Add Source property to Battery API

As Battery API will be generalized for other battery reporting
protocols, the Source property is useful for diagnostics purposes.

Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoprofiles/battery: Refactor to use battery library
Sonny Sasaka [Mon, 30 Nov 2020 21:55:57 +0000 (13:55 -0800)]
profiles/battery: Refactor to use battery library

This refactors profiles/battery to use the internal battery library that
handles the D-Bus intricacies so that profiles/battery only handles the
GATT BAS concerns.

Reviewed-by: Daniel Winkler <danielwinkler@google.com>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobattery: Add the internal Battery API
Sonny Sasaka [Mon, 30 Nov 2020 21:55:56 +0000 (13:55 -0800)]
battery: Add the internal Battery API

This patch adds an API for internal BlueZ code to expose battery
information to org.bluez.Battery1 interface. The motivation behind this
is that there is going to be other places than GATT BAS handler that
exposes battery information, for example internal plugins and the
BatteryProvider1 D-Bus API for external clients.

Reviewed-by: Daniel Winkler <danielwinkler@google.com>
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoa2dp: Fix crash if setup is freed while reconfiguring
Luiz Augusto von Dentz [Mon, 30 Nov 2020 19:16:11 +0000 (11:16 -0800)]
a2dp: Fix crash if setup is freed while reconfiguring

This fixes the crash reported on:

https://github.com/bluez/bluez/issues/60
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomain.conf: use correct key for BREDR configuration
Ronan Pigott [Tue, 24 Nov 2020 22:46:58 +0000 (15:46 -0700)]
main.conf: use correct key for BREDR configuration

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agolib: Fix the unchecked return value
Tedd Ho-Jeong An [Fri, 20 Nov 2020 20:07:12 +0000 (12:07 -0800)]
lib: Fix the unchecked return value

This patch fixes the unchecked return value.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoprofile/bnep: Fix the unchecked return value
Tedd Ho-Jeong An [Fri, 20 Nov 2020 20:07:11 +0000 (12:07 -0800)]
profile/bnep: Fix the unchecked return value

This patch fixes the unchecked return value.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoemulator: Fix the unchecked return value
Tedd Ho-Jeong An [Fri, 20 Nov 2020 20:07:10 +0000 (12:07 -0800)]
emulator: Fix the unchecked return value

This patch fixes the unchecked return value.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtio: Fix the unchecked return value
Tedd Ho-Jeong An [Fri, 20 Nov 2020 20:07:09 +0000 (12:07 -0800)]
btio: Fix the unchecked return value

This patch fixes the unchecked return value.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomonitor: Fix the unchecked return value
Tedd Ho-Jeong An [Fri, 20 Nov 2020 20:07:08 +0000 (12:07 -0800)]
monitor: Fix the unchecked return value

This patch fixes the unchecked return value.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomonitor: Fix potential memory leak
Tedd Ho-Jeong An [Fri, 20 Nov 2020 20:07:07 +0000 (12:07 -0800)]
monitor: Fix potential memory leak

If the mainloop_add_fd() returns with failure, the destroy callback is
never called so any reosurces need to be released never freed/closed.

This potential leakage is checked with valgrind after failing the
mainloop_add_fd() function manually.

   ==258684== 1,500 bytes in 1 blocks are definitely lost in loss record 3 of 3
   ==258684==    at 0x483BB1A: calloc (vg_replace_malloc.c:760)
   ==258684==    by 0x123F1A: open_channel (control.c:1058)
   ==258684==    by 0x125B09: control_tracing (control.c:1540)
   ==258684==    by 0x122764: main (main.c:255)
   ==258684==
   ==258684== LEAK SUMMARY:
   ==258684==    definitely lost: 1,500 bytes in 1 blocks
   ==258684==    indirectly lost: 0 bytes in 0 blocks
   ==258684==      possibly lost: 0 bytes in 0 blocks
   ==258684==    still reachable: 48 bytes in 2 blocks
   ==258684==         suppressed: 0 bytes in 0 blocks

This patch frees/closes the resources if the function returns with
failure.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoa2dp: Fix crash when SEP codec has not been initialized
Luiz Augusto von Dentz [Mon, 23 Nov 2020 18:20:06 +0000 (10:20 -0800)]
a2dp: Fix crash when SEP codec has not been initialized

If SEP has not been properly discovered avdtp_get_codec may return NULL
thus causing crashes such as:

https://github.com/bluez/bluez/issues/57
Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agohog: Fix stripping off read report bytes
Dean Camera [Fri, 20 Nov 2020 01:42:46 +0000 (12:42 +1100)]
hog: Fix stripping off read report bytes

If the HID subsystem requests a HID report to be read from the
device, we currently incorrectly strip off the first byte of the
response, if the device has report IDs set in the HID report
descriptor.

This is incorrect; unlike USB HID, the report ID is *not* included
in the HOG profile's HID reports, and instead exists out of band
in a descriptor on the report's bluetooth characteristic in the
device.

In this patch, we remove the erroneous stripping of the first
byte of the report, and (if report IDs are enabled) prepend the
report ID to the front of the result. This makes the HID report
returned indentical in format to that of a USB HID report, so
that the upper HID drivers can consume HOG device reports in the
same way as USB.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agogatt: Fix duplicate free for GATT service includes
Pavel Maltsev [Thu, 19 Nov 2020 21:43:37 +0000 (21:43 +0000)]
gatt: Fix duplicate free for GATT service includes

Objects in the service->includes queue are obtained via
dbus_message_iter_get_basic call and according to the
contract for the value is that it is returned by the reference
and should not be freed thus we should make a copy.

This will fix the issue when the GATT service app is disconnected
(reproduced with gatt-service included in bluez), bluetoothd is crashing:

bluetoothd: src/gatt-database.c:gatt_db_service_removed() Local GATT service removed
bluetoothd: src/adapter.c:adapter_service_remove() /org/bluez/hci0
bluetoothd: src/adapter.c:remove_uuid() sending remove uuid command for index 0
bluetoothd: src/sdpd-service.c:remove_record_from_server() Removing record with handle 0x10006
bluetoothd: src/gatt-database.c:proxy_removed_cb() Proxy removed - removing service: /service1
munmap_chunk(): invalid pointer

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest-runner: Move code kill processes created
Luiz Augusto von Dentz [Tue, 17 Nov 2020 21:58:37 +0000 (13:58 -0800)]
test-runner: Move code kill processes created

This moves the code that kill processes outside of the loop iteration.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtdev: Fix not emitting Ext Connection Complete properly
Luiz Augusto von Dentz [Tue, 17 Nov 2020 20:31:28 +0000 (12:31 -0800)]
btdev: Fix not emitting Ext Connection Complete properly

le_meta_event must be using when emitting a meta event.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest-runner: Use multidevs=remap when mounting rootfs
Luiz Augusto von Dentz [Tue, 17 Nov 2020 19:01:57 +0000 (11:01 -0800)]
test-runner: Use multidevs=remap when mounting rootfs

This avoid inode collisions when the host has multiple fs to be mapped
to the rootfs on the guest.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest-runner: Change qemu output
Luiz Augusto von Dentz [Tue, 17 Nov 2020 18:51:11 +0000 (10:51 -0800)]
test-runner: Change qemu output

This changes the serial device to include mux=on which enables the
output of early kernel logs.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodevice: Update cache only if content changed
Szymon Janc [Tue, 17 Nov 2020 15:57:03 +0000 (16:57 +0100)]
device: Update cache only if content changed

This fix hammering storage device (eg sdcard) when doing scanning
and LE devices are around.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomain.conf: Add option to configure AVDP session/stream channel modes
Luiz Augusto von Dentz [Mon, 16 Nov 2020 22:01:34 +0000 (14:01 -0800)]
main.conf: Add option to configure AVDP session/stream channel modes

This adds a new group AVDTP where platform can configure the preferred
L2CAP channel modes for both session (signalling) and stream
(transport). For backward compatibility the both SessionMode and
StreamMode defaults to basic mode.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoavdtp: Fix connecting using streaming mode with signalling channel
Luiz Augusto von Dentz [Mon, 16 Nov 2020 19:08:16 +0000 (11:08 -0800)]
avdtp: Fix connecting using streaming mode with signalling channel

Signalling channel shall only use Basic or ERTM modes.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoadv_monitor: Fix remove monitor from the kernel
Manish Mandlik [Mon, 16 Nov 2020 18:41:08 +0000 (10:41 -0800)]
adv_monitor: Fix remove monitor from the kernel

A monitor is removed in the following scenarios:
- monitor dbus object removed by the app
- monitor removed by the kernel
- client app invokes UnregisterMonitor()
- client app is killed/disconnected
- AdvMonitorManager is destroyed

In the first case, we need to remove the corresponding monitor from the
kernel and free the bluez monitor object.

In the second case, we need to call the Release() method on the
corresponding dbus monitor object and free the bluez monitor object.

Kernel may remove all monitors and send MGMT_EV_ADV_MONITOR_REMOVED
event to bluez. In such case, we need to call Release() method on all
monitors from all registered apps, and free the bluez monitor objects.

In the third case, we need to call the Release() method on the monitor
objects created by the app, remove corresponding monitors from the
kernel and then free the bluez monitor object.

In the fourth case, since the app is not available, all the dbus monitor
objects created by that app are also unavailable. So, we just need to
remove corresponding monitors from the kernel and then free the bluez
monitor objects.

In the fifth case, we need to call Release() method on all monitors from
all registered apps, remove corresponding monitors from the kernel and
then free the bluez monitor objects.

When app exits or gets killed without removing the dbus monitor objects
and without invoking the UnregisterMonitor() method, a race condition
could happen between app_destroy and monitor_proxy_removed since dbus
objects hosted by the app are destroyed on app exit.

This patch fixes the first, second and fourth cases ensuring that
monitors from the kernel are removed correctly, Release() method is
invoked whenever necessary and bluez monitor objects are freed only
once.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest-runner: Add option to start monitor
Luiz Augusto von Dentz [Fri, 13 Nov 2020 23:07:03 +0000 (15:07 -0800)]
test-runner: Add option to start monitor

This adds an option (-m/--monitor) to start btmon before the command
running the command/tester which is convenient when a tests results in
HCI traffic which are decoded by hciemu as it only prints the hexdump.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobthost: Fix sending uninitilized data
Luiz Augusto von Dentz [Wed, 11 Nov 2020 22:42:36 +0000 (14:42 -0800)]
bthost: Fix sending uninitilized data

Syscall param writev(vector[...]) points to uninitialised byte(s)
   at 0x4A8A497: writev (in /usr/lib64/libc-2.31.so)
   by 0x1365C1: send_packet (bthost.c:509)
   by 0x1365C1: send_packet (bthost.c:492)
   by 0x1365C1: next_cmd (bthost.c:737)
   by 0x137578: evt_cmd_complete (bthost.c:882)
   by 0x137578: process_evt (bthost.c:1343)
   by 0x137578: bthost_receive_h4 (bthost.c:2414)
   by 0x12DF49: receive_bthost.part.0 (hciemu.c:134)
   by 0x48B978E: g_main_context_dispatch (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x48B9B17: ??? (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x48B9E32: g_main_loop_run (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x14A664: mainloop_run (mainloop-glib.c:66)
   by 0x14AA41: mainloop_run_with_signal (mainloop-notify.c:188)
   by 0x149A7B: tester_run (tester.c:871)
   by 0x129C77: main (mgmt-tester.c:10219)
 Address 0x4c48c05 is 21 bytes inside a block of size 280 alloc'd
   at 0x483CAE9: calloc (vg_replace_malloc.c:760)
   by 0x136487: queue_command (bthost.c:472)
   by 0x136487: send_command (bthost.c:715)
   by 0x13948A: bthost_set_ext_adv_enable (bthost.c:2581)
   by 0x12CEBF: trigger_device_found (mgmt-tester.c:8513)
   by 0x148C9A: wait_callback (tester.c:749)
   by 0x48BA330: ??? (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x48B978E: g_main_context_dispatch (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x48B9B17: ??? (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x48B9E32: g_main_loop_run (in /usr/lib64/libglib-2.0.so.0.6400.6)
   by 0x14A664: mainloop_run (mainloop-glib.c:66)
   by 0x14AA41: mainloop_run_with_signal (mainloop-notify.c:188)
   by 0x149A7B: tester_run (tester.c:871)

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agohciemu: Fix crash when test is aborted during pre setup
Luiz Augusto von Dentz [Wed, 11 Nov 2020 22:15:46 +0000 (14:15 -0800)]
hciemu: Fix crash when test is aborted during pre setup

If the tester_pre_setup_failed is called all timeout related to the
test must be cancelled as the test should have been freed by the
application and the next test is about to start.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Update expected advertisement flags
Luiz Augusto von Dentz [Wed, 11 Nov 2020 02:16:54 +0000 (18:16 -0800)]
mgmt-tester: Update expected advertisement flags

This updates the expected advertisement flags which now includes
MGMT_ADV_FLAG_CAN_SET_TX_POWER and MGMT_ADV_FLAG_HW_OFFLOAD.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Print unexpected command responses
Luiz Augusto von Dentz [Wed, 11 Nov 2020 02:15:06 +0000 (18:15 -0800)]
mgmt-tester: Print unexpected command responses

This prints the difference between the command response and the
expected command response to make it easier to debug.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Update supported-features to not include HS
Luiz Augusto von Dentz [Wed, 11 Nov 2020 02:09:56 +0000 (18:09 -0800)]
mgmt-tester: Update supported-features to not include HS

High speed has been disabled by default, also since it can be enabled
this turn checks of supported-features into warnings to allow it work
even if BT_HS has been enabled in the kernel.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoshared/tester: Add option to filter by string
Luiz Augusto von Dentz [Wed, 11 Nov 2020 02:06:00 +0000 (18:06 -0800)]
shared/tester: Add option to filter by string

This adds option to filter by string which does not require to enter
the entire prefix.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agosmp-tester: Fix not registering for MGMT_EV_USER_CONFIRM_REQUEST
Luiz Augusto von Dentz [Tue, 10 Nov 2020 23:10:09 +0000 (15:10 -0800)]
smp-tester: Fix not registering for MGMT_EV_USER_CONFIRM_REQUEST

For client test MGMT_EV_USER_CONFIRM_REQUEST shall be resgistered as
well as the kernel has been updated to request user confirmation even
when acting as initiator.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agol2cap-tester: Fix L2CAP LE Server - Success
Luiz Augusto von Dentz [Tue, 10 Nov 2020 22:57:10 +0000 (14:57 -0800)]
l2cap-tester: Fix L2CAP LE Server - Success

This test was failing due to the kernel now using different values for
MPS and credits.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobthost: Fix running request callbacks for remote requests
Luiz Augusto von Dentz [Tue, 10 Nov 2020 22:54:10 +0000 (14:54 -0800)]
bthost: Fix running request callbacks for remote requests

Ident are only unique per request/response pair so if a remote request
is initiated we shall not attempt to run the callback even if the ident
matches.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agobtdev: Rework command handling
Luiz Augusto von Dentz [Thu, 20 Aug 2020 00:53:55 +0000 (17:53 -0700)]
btdev: Rework command handling

This create command tables per version/supported command bits to make
it easier to keep them in sync.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomain: Don't warn for unset config option
Luiz Augusto von Dentz [Mon, 9 Nov 2020 22:57:56 +0000 (14:57 -0800)]
main: Don't warn for unset config option

Unset options shall not be printed if debug is not enabled.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoclient: Remove reference to g_steal_pointer()
Inga Stotland [Mon, 9 Nov 2020 19:27:42 +0000 (11:27 -0800)]
client: Remove reference to g_steal_pointer()

In order to avoid requiring higher version of GLib, replace
g_steal_pointer() with explicitly setting the pointer value
to NULL.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodoc/adapter-api: Add InProgess error to StartDiscovery
Michael Nosthoff [Mon, 9 Nov 2020 10:51:14 +0000 (11:51 +0100)]
doc/adapter-api: Add InProgess error to StartDiscovery

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoCancel GATT client operations when cleaning up GATT cache
Sonny Sasaka [Thu, 5 Nov 2020 22:49:23 +0000 (14:49 -0800)]
Cancel GATT client operations when cleaning up GATT cache

A crash is found when discovery_complete_op is invoked but
gatt_db_attribute objects have been freed. A solution is to always
cancel GATT client operations every time we clean GATT cache.

The crash is reproducible by connecting to an LE mouse and then calling
RemoveDevice immediately, triggering disconnection while GATT service
discovery is in progress.

Sample crash stack trace:
0  gatt_db_service_get_handles (service=0x1751130, service=0x1751130, end_handle=0x7ffcd600806e, start_handle=0x7ffcd600806c) at src/shared/gatt-db.c:569
1  gatt_db_attribute_get_service_data (attrib=<optimized out>, start_handle=0x7ffcd600806c, end_handle=0x7ffcd600806e, primary=0x0, uuid=0x0) at src/shared/gatt-db.c:1657
2  0x00000000004983a8 in discovery_op_complete (op=op@entry=0x173b320, success=<optimized out>, err=err@entry=10 '\n') at src/shared/gatt-client.c:406
3  0x000000000049a548 in discover_descs_cb (success=<optimized out>, att_ecode=<optimized out>, result=<optimized out>, user_data=0x173b320) at src/shared/gatt-client.c:915
4  0x00000000004a1d87 in discovery_op_complete (op=0x1748450, success=<optimized out>, ecode=<optimized out>) at src/shared/gatt-helpers.c:615
5  0x00000000004a2379 in discover_descs_cb (opcode=<optimized out>, pdu=0x174d551, length=<optimized out>, user_data=0x1748450) at src/shared/gatt-helpers.c:1465
6  0x00000000004966db in handle_rsp (pdu_len=4, pdu=<optimized out>, opcode=<optimized out>, chan=0x17483c0) at src/shared/att.c:814
7  can_read_data (io=<optimized out>, user_data=0x17483c0) at src/shared/att.c:1011
8  0x00000000004a0853 in watch_callback (channel=<optimized out>, cond=<optimized out>, user_data=<optimized out>) at src/shared/io-glib.c:157
9  0x00007fb3f2d7fe87 in g_main_context_dispatch () from /usr/lib64/libglib-2.0.so.0
10 0x00007fb3f2d80230 in ?? () from /usr/lib64/libglib-2.0.so.0
11 0x00007fb3f2d80542 in g_main_loop_run () from /usr/lib64/libglib-2.0.so.0
12 0x00000000004a0e25 in mainloop_run () at src/shared/mainloop-glib.c:66
13 0x00000000004a11f2 in mainloop_run_with_signal (func=func@entry=0x43f200 <signal_callback>, user_data=user_data@entry=0x0) at src/shared/mainloop-notify.c:188
14 0x000000000040c72e in main (argc=<optimized out>, argv=<optimized out>) at src/main.c:959

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoaudio/avdtp: Report failure in disconnected state
Archie Pusaka [Thu, 5 Nov 2020 06:40:34 +0000 (14:40 +0800)]
audio/avdtp: Report failure in disconnected state

A2DP are relying on the disconnected state callback to do cleanup.
If failure occurs when AVDTP are already in the disconnected state,
we didn't make any transition state, therefore A2DP would miss this
event.

This patch allows the transition to disconnected state, even though
we are previously already in the disconnected state.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoaudio: unref session when failed to setup
Archie Pusaka [Thu, 5 Nov 2020 06:40:33 +0000 (14:40 +0800)]
audio: unref session when failed to setup

There is a possibility to miss unref-ing when source/sink fails at
setup.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoservice: add adapter powered check before connecting
Archie Pusaka [Thu, 5 Nov 2020 06:40:32 +0000 (14:40 +0800)]
service: add adapter powered check before connecting

When policy_connect() is called, there might be a case where the
power is just turned off, hence the adapter is down. Add a check in
btd_service_connect to detect such condition.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoemulator: Return error for ext adv feature not supported
Luiz Augusto von Dentz [Wed, 4 Nov 2020 23:05:37 +0000 (15:05 -0800)]
emulator: Return error for ext adv feature not supported

This return an error if LE Set Extended Advertising Enable is used with
unsupported features.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Fix failing pairing tests
Luiz Augusto von Dentz [Wed, 4 Nov 2020 19:27:53 +0000 (11:27 -0800)]
mgmt-tester: Fix failing pairing tests

Just-works pairing does request user confirmation.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Don't run tests which command fails
Luiz Augusto von Dentz [Wed, 4 Nov 2020 19:25:15 +0000 (11:25 -0800)]
mgmt-tester: Don't run tests which command fails

This indicates that emulator don't support the test and until proper
support is added the test shall not be run.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Fix ext advertisement handle
Luiz Augusto von Dentz [Wed, 4 Nov 2020 19:06:14 +0000 (11:06 -0800)]
mgmt-tester: Fix ext advertisement handle

Handle has been change to 0x01 not 0x00:

< HCI Command: LE Set Extended Advertising Enable (0x08|0x0039) plen 6
        Extended advertising: Enabled (0x01)
        Number of sets: 1 (0x01)
        Entry 0
          Handle: 0x01
          Duration: 0 ms (0x00)
          Max ext adv events: 0

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Update set_connectable_off_adv_param
Luiz Augusto von Dentz [Tue, 3 Nov 2020 23:38:56 +0000 (15:38 -0800)]
mgmt-tester: Update set_connectable_off_adv_param

Parameters have been changed in the kernel which now uses the following:

< HCI Command: LE Set Advertising Parameters (0x08|0x0006) plen 15
        Min advertising interval: 62.500 msec (0x0064)
        Max advertising interval: 93.750 msec (0x0096)
        Type: Scannable undirected - ADV_SCAN_IND (0x02)
        Own address type: Random (0x01)
        Direct address type: Public (0x00)
        Direct address: 00:00:00:00:00:00 (OUI 00-00-00)
        Channel map: 37, 38, 39 (0x07)
        Filter policy: Allow Scan Request from Any, Allow Connect Request from Any (0x00)

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agomgmt-tester: Print out the HCI received and expected cmd value
Luiz Augusto von Dentz [Tue, 3 Nov 2020 23:36:28 +0000 (15:36 -0800)]
mgmt-tester: Print out the HCI received and expected cmd value

This make it simpler to debug when the kernel has been updated and
parameters no longer match:

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotest-runner: Fix using removed options
Luiz Augusto von Dentz [Tue, 3 Nov 2020 23:15:08 +0000 (15:15 -0800)]
test-runner: Fix using removed options

This removes -nodefconfig and -balloon which are not longer valid
options:

  qemu-system-x86_64: -nodefconfig: invalid option
  qemu-system-x86_64: -balloon: invalid option

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agodoc: Add tester.config
Luiz Augusto von Dentz [Tue, 3 Nov 2020 23:09:43 +0000 (15:09 -0800)]
doc: Add tester.config

This adds tester.config which can be used as a reference to create
kernel images to be run with test-runner.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agoadapter: Fix a crash caused by lingering discovery client pointer
Miao-chen Chou [Mon, 2 Nov 2020 22:30:03 +0000 (14:30 -0800)]
adapter: Fix a crash caused by lingering discovery client pointer

This cleans up the lingering pointer, adapter->client, during powering
off the adapter. The crash occurs when a D-Bus client set Powered
property to false and immediately calls StopDiscovery() when there is
ongoing discovery. As a part of powering off the adapter,
adapter->discovery_list gets cleared, and given that adapter->client
refers to one of the clients in adapter->discovery_list, adapter->client
should be cleared along with it.

(1) Connect to a BT audio device from BT system tray.
(2) Once the audio device is connected, power off BT and immediately
power off the audio device.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/smp-tester: Enable hciemu debug
Luiz Augusto von Dentz [Mon, 2 Nov 2020 19:09:12 +0000 (11:09 -0800)]
tools/smp-tester: Enable hciemu debug

This enable hciemu debug when debug is enabled.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/sco-tester: Enable hciemu debug
Luiz Augusto von Dentz [Mon, 2 Nov 2020 19:08:39 +0000 (11:08 -0800)]
tools/sco-tester: Enable hciemu debug

This enable hciemu debug when debug is enabled.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/rfcomm-tester: Enable hciemu debug
Luiz Augusto von Dentz [Mon, 2 Nov 2020 19:08:12 +0000 (11:08 -0800)]
tools/rfcomm-tester: Enable hciemu debug

This enable hciemu debug when debug is enabled.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/bnep-tester: Enable hciemu debug
Luiz Augusto von Dentz [Mon, 2 Nov 2020 19:07:19 +0000 (11:07 -0800)]
tools/bnep-tester: Enable hciemu debug

This enable hciemu debug when debug is enabled.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>
2 years agotools/mgmt-tester: Enable hciemu debug
Luiz Augusto von Dentz [Mon, 2 Nov 2020 19:06:35 +0000 (11:06 -0800)]
tools/mgmt-tester: Enable hciemu debug

This enable hciemu debug when debug is enabled.

Signed-off-by: Anuj Jain <anuj01.jain@samsung.com>
Signed-off-by: Ayush Garg <ayush.garg@samsung.com>