Adapter property notification are send from multiple places so it make
sense to have helper for that. This is especially usefull for 'simple'
properties.
Remote device property notification will be send from multiple places
so it make sense to have helper for that. This will be especially
usefull for 'simple' properties.
Rename send_remote_device_name_prop to get_device_name and make it
accept struct device as parameter. Also return HAL status code.
This will allow to use this function also in get device property
command handler.
This match adapter properties handling functions with properties names.
Will make code easier to understand and avoid clashes with remote
device properties functions.
Currently eir_parse always return 0 but it is checked throughout the code
(in android/bluetooth code as well in src/adapteri, etc) for return value
(err < 0) which never happens. Make function eir_parse return void. This
fixes warnings from static analyzer tools.
There is no need to sending set_discoverable command if new settings
are not connectable as clearing connectable also clear discoverable.
Sending set_discoverable when not connectable result in 'rejected'
mgmt response.
This fix issue with sending invalid success response from several
places where ipc_send_rsp was used for reporting success. Instead of
using using ipc_send for success response, make helper handle that.
This makes function flow easier to follow and understand. Besides that
it also fix issue with sending to many bytes if some prop were not
present in EIR.
Before sending any ssp request or pin code request up to HAL library we
need to send bond state change with bonding state. Otherwise incoming
bonding is not correctly handled by Bluetooth.apk.
In this patch also device list has been added in order to e.g track
bonding state.
Note: For incoming paring (security mode 3) there is a need to send
HAL_EVE_REMOTE_DEVICE_PROPS before HAL_EV_PIN_REQUEST.
It is because Android will crash due to bug in pinRequestCallback
function in java. Android checks if device is already in HashMap and if
not then creates device, but forget to use that one, but instead do
operations on NULL. By sending HAL_BOND_STATE_BONDING event it works
better but we have race issue. It is because new device is added to
HashMap not in callback context but later after BONDING msg will be
received by BondStateMachine. If it happens before pin_request_cb hits
java then we are fine, otherwise not. So for that reason we send
HAL_EV_REMOTE_DEVICE_PROPS so in the java handler class new device will
be added to HashMap in the callback context.
In ssp case we don't have this problem as we send device found once acl
is created.
