- 16 10月, 2013 7 次提交
-
-
由 Gustavo Padovan 提交于
l2cap_state_change_and_error() introduces the ability to update a l2cap_user with changes in channel's state and error code with just one call. The main reason for this is to avoid race conditions between and setting the state and then the error. Otherwise we would need to release the lock between both operations. This is another step of an ongoing work to make l2cap_core.c totally independent from l2cap's struct sock. Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Gustavo Padovan 提交于
Instead of creating an new function pointer to report errors we are just reusing state_change for that and there is a simple reason for this, one place in the l2cap_core.c code needs, in a locked sk, set both the sk_state and sk_err. If we create two different functions for this we would need to release the lock between the two operation putting the socket in non desired state. The change is transparent to the l2cap_core.c code, user that only needs to set the state won't need any modification. This is another step of an ongoing work to make l2cap_core.c totally independent from l2cap's struct sock. Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Marcel Holtmann 提交于
When the discoverable timeout triggers and limited discoverable mode was used, then the class of device needs to be updated to remove the limited discoverable bit. To keep the class of device logic in a central place, expose a new function mgmt_discoverable_timeout that can be called from the timeout callback. In case the class of device value needs updating, it will add the HCI command to the transaction. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The eir_get_length() function is only used from hci_event.c and so instead of having a public function move it to the location where it is used. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The eir_append_data() function is only used from mgmt.c and so instead of having a public function move it to the location where it is used. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The return value of mgmt_new_link_key() function is not used and so just change it to return void. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The limited discoverable mode should be used when a device is only discoverable for a certain amount of time and after that it returns back into being non-discoverable. This adds another option to the set discoverable management command to clearly distinguish limited discoverable from general discoverable mode. While the general discoverable mode can be set with a specific timeout or as permanent setting, the limited discoverable mode requires a timeout. The timeout is flexible and the kernel will not enforce any specific limitations. That GAP part of this is required by userspace to enforce according to the Bluetooth core specification. Devices in limited discoverable mode can still be found by the general discovery procedure. It is mandatory that a device sets both GIAC and LIAC when entering limited discoverable mode. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
- 15 10月, 2013 20 次提交
-
-
由 Marcel Holtmann 提交于
This patch just adds the HCI command structure for configuring the current IAC LAP setting. The length of the command is variable and supports more than two IAC. However since there is only general discoverable and limited discoverable modes, this can be limited to two possible IACs. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
When only the discoverable timeout gets updated, just cancel the current timeout, store the new timeout value. If the new timeout is valid, then arm the discoverable timeout again. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The discoverable timeout is currently armed from hci_event.c and causes some side effects when using HCI commands instead of the management interface. To make this clear, only arm the discoverable timeout from the management command complete handler. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
When the discoverable mode gets changed, ensure that the class of device value has the correct limited discoverable bit value set. Since the class of device HCI command will only be send to the controller when the value changes, it is safe to just always trigger the update. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The return value of mgmt_write_scan_failed() function is not used and so just change it to return void. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The return value of mgmt_connectable() function is not used and so just change it to return void. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The return value of mgmt_discoverable() function is not used and so just change it to return void. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
Add a new flag that can be set when in limited discoverable mode. This flag will cause the limited discoverable bit in the class of device value to bet set. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
Magically updating the advertising data when some random command enables advertising in the controller is not really a good idea. It also caused a bit of complicated code with the exported hci_udpate_ad function that is shared from many places. This patch consolidates the advertising data update into the management core. It also makes sure that when powering on with LE enabled or later on enabling LE the controller has a good default for advertising data. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
When the discoverable timeout triggers and it is time to turn inquiry scan back off, use the HCI request framework to do it. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
There is a minor coding style violation and so just fix it. No actual logic has changed. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
Read the current IAC LAP values when initializing the controller. The values are not used, but it is good to have them in the trace files for debugging purposes. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
-
由 Marcel Holtmann 提交于
When initializing a controller make sure to read out the number of supported IAC and store its result. This value is needed to determine if limited discoverable for BR/EDR can be configured or not. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NGustavo Padovan <gustavo.padovan@collabora.co.uk>
-
由 Marcel Holtmann 提交于
The scan window parameter for connection establishment and passive scanning needs to be smaller or equal than the scan interval. Instead of waiting for a controller to reject these values later on, just reject them right away. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Johan Hedberg 提交于
If a socket was bound to an address type other than BR/EDR (such as LE) we should reject trying to connect it to a BR/EDR address. The same applies for binding to BR/EDR and trying to connect to non-BR/EDR. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
We need to verify that the bdaddr type passed to connect() and bind() is within the set of valid values. If it is not we need to cleanly fail with EINVAL. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
This patch converts Set Discoverable to use an asynchronous request along with its own completion callback. This is necessary for splitting raw HCI socket use cases from mgmt, as well as for enabling the hooking up of Advertising parameters together with the HCI_DISCOVERABLE flag (coming in later patches). Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
Now that the connectable setting is also applicable for the LE side it's possible that the HCI_CONNECTABLE flag is already set when changing the BR/EDR setting from false to true while the controller is powered. In this situation we need to update the BR/EDR scan mode to reflect the setting. Additionally, since HCI_CONNECTABLE also applies to LE we must not clear the HCI_CONNECTABLE flag when disabling bredr. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
The set_bredr_scan() function will soon be needed by the set_bredr() function, so move it to a new location to avoid having to add a forward declaration. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
This patch updates the Set Connectable Management command to also update the LE advertising type to either connectable or non-connectable advertising. An extra helper function is needed for getting the right advertising type since we can not only rely on the HCI_CONNECTABLE flag but must also check for a pending Set Connectable command (in which case the flag does not yet have its final value). Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
- 14 10月, 2013 13 次提交
-
-
由 Johan Hedberg 提交于
We need to ensure that the advertising data is up-to-date whenever advertising is enabled, but when disabling advertising we do not need to worry about it (since it will eventually get fixed as soon as advertising is enabled again). This patch fixes this in the command complete callback for set_adv_enable. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
These functions will soon be used by set_connectable() so move them to a location in mgmt.c that doesn't require forward declarations. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
If the HCI commands related to the Set Connectable command fail we will get a non-zero status in the request completion callback. In such a case we must respond with the appropriate command status message to user space. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
This patch moves the responsibility of setting/clearing the HCI_CONNECTABLE flag to the request completion callback of the Set Connectable command. This will allow us to cleanly add support for LE Advertising hooks in later patches. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Johan Hedberg 提交于
This patch moves all the decisions of which HCI commands to send (or not to send) to the code between hci_req_init() and hci_req_run() this allows us to further extend the request with further commands but still keep the same logic of handling whether to return a direct mgmt response in the case that no HCI commands were sent. Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com> Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
-
由 Marcel Holtmann 提交于
Clearing the BT_SK_SUSPEND socket flag from the L2CAP core is causing a dependency on the socket. So intead of doing that, use a channel callback into the socket handling to resume. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The L2CAP core should not look into the socket flags to figure out the setting of defer setup. So introduce a L2CAP channel flag that mirrors the socket flag. Since the defer setup option is only set in one place this becomes a really easy thing to do. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The exposed socket information do not contain source or destination addresses. So adjust the header accordingly. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
There have been a lot of changes in the core Bluetooth handling lately. So it is a good idea to increase the module version. The module version is not used anywhere, but it makes debugging a little bit simpler if versions can be distinguished. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
The L2CAP connectionless channels use SOCK_DGRAM and recvmsg() and need to receive the remote BD_ADDR and PSM information via msg_name from the recvmsg() system call. So in case the L2CAP socket is for connectionless channels, provide a msg_name callback that can update the data. Also store the remote BD_ADDR and PSM in the skb so it can be extracted later on. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
This allows to add a per socket msg_name callback that can be used for updating the msg_name information for recvmsg() system calls. This feature is used by another patch to support address information on L2CAP connectionless channels. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
There are few places where it makes sense to use l2cap_pi(sk) directly instead of assigning it to temporary structure. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-
由 Marcel Holtmann 提交于
Every socket protocol now stores its own address information. So just remove the generic src and dst fields since they are no longer needed. Signed-off-by: NMarcel Holtmann <marcel@holtmann.org> Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
-