- 22 Aug, 2013 6 commits
-
-
Gabor Juhos authored
According to the Ralink driver, there is an MCU command which can be used to send the frequency offset value directly to the USB device without going through the RFCSR writing sequence. Based on the DPO_RT5572_LinuxSTA_2.6.0.1_20120629 driver. Reference: RTMPAdjustFrequencyOffset function in common/rt_rf.c Signed-off-by:
Gabor Juhos <juhosg@openwrt.org> Signed-off-by:
John W. Linville <linville@tuxdriver.com>
-
Gabor Juhos authored
Don't write the new value into the register if it is the same as the old value to avoid unncessary USB bus traffic with USB devices. The change also saves a few cycle on MMIO based devices. Based on the DPO_RT5572_LinuxSTA_2.6.0.1_20120629 driver. Reference: RTMPAdjustFrequencyOffset function in common/rt_rf.c Signed-off-by:
-
Gabor Juhos authored
The current code in the 'rt2800_adjust_freq_offset' function limits the device specific frequency offset value to FREQ_BOUND but ignores the fact that the uppermost bit is not part of the frequency offset value. As the result, the driver always uses the FREQ_BOUND value if the uppermost bit is set. Update the code to use the correct source value for calculating the boundary. Based on the DPO_RT5572_LinuxSTA_2.6.0.1_20120629 driver. Reference: RTMPAdjustFrequencyOffset function in common/rt_rf.c Signed-off-by:
-
Gabor Juhos authored
Instead of precomputing the beacon base in each caller, pass the beacon index to the 'rt2800_clear_beacon_register' function and compute the beacon base in there. This allows to simplify the caller functions a bit. Signed-off-by:
-
Gabor Juhos authored
The name of the HW_BEACON_OFFSET macro is a bit confusing. It returns with one of the HW_BEACON_BASE* values, so rename the macro to HW_BEACON_BASE to reflect that. The patch contains no functional changes. Signed-off-by:
-
-
- 21 Aug, 2013 19 commits
-
-
Frédéric Dalleau authored
When initiating a transparent eSCO connection, make use of T2 settings at first try. T2 is the recommended settings from HFP 1.6 WideBand Speech. Upon connection failure, try T1 settings. When CVSD is requested and eSCO is supported, try to establish eSCO connection using S3 settings. If it fails, fallback in sequence to S2, S1, D1, D0 settings. To know which setting should be used, conn->attempt is used. It indicates the currently ongoing SCO connection attempt and can be used as the index for the fallback settings table. These setting and the fallback order are described in Bluetooth HFP 1.6 specification p. 101. Signed-off-by:
Frédéric Dalleau <frederic.dalleau@linux.intel.com> Signed-off-by:
Johan Hedberg <johan.hedberg@intel.com> Signed-off-by:
Gustavo Padovan <gustavo.padovan@collabora.co.uk>
-
Frédéric Dalleau authored
Synchronous Connection Complete event can return error "Connection Rejected due to Limited resources (0x10)". Handling this error is required for SCO connection fallback. This error happens when the server tried to accept the connection but failed to negotiate settings. This error code has been verified experimentally by sending a T2 request to a T1 only SCO listener. Client dump follows : < HCI Command (0x01|0x0028) plen 17 [hci0] 3.696064 Handle: 12 Transmit bandwidth: 8000 Receive bandwidth: 8000 Max latency: 13 Setting: 0x0003 Retransmission effort: Optimize for link quality (0x02) Packet type: 0x0380 > HCI Event (0x0f) plen 4 [hci0] 3.697034 Setup Synchronous Connection (0x01|0x0028) ncmd 1 Status: Success (0x00) > HCI Event (0x2c) plen 17 [hci0] 3.736059 Status: Connection Rejected due to Limited Resources (0x0d) Handle: 0 Address: xx:xx:xx:xx:xx:AB (OUI 70-F3-95) Link type: eSCO (0x02) Transmission interval: 0x0c Retransmission window: 0x06 RX packet length: 60 TX packet length: 60 Air mode: Transparent (0x03) Server dump follows : > HCI Event (0x04) plen 10 [hci0] 4.741513 Address: xx:xx:xx:xx:xx:D9 (OUI 20-68-9D) Class: 0x620100 Major class: Computer (desktop, notebook, PDA, organizers) Minor class: Uncategorized, code for device not assigned Networking (LAN, Ad hoc) Audio (Speaker, Microphone, Headset) Telephony (Cordless telephony, Modem, Headset) Link type: eSCO (0x02) < HCI Command (0x01|0x0029) plen 21 [hci0] 4.743269 Address: xx:xx:xx:xx:xx:D9 (OUI 20-68-9D) Transmit bandwidth: 8000 Receive bandwidth: 8000 Max latency: 13 Setting: 0x0003 Retransmission effort: Optimize for link quality (0x02) Packet type: 0x03c1 > HCI Event (0x0f) plen 4 [hci0] 4.745517 Accept Synchronous Connection (0x01|0x0029) ncmd 1 Status: Success (0x00) > HCI Event (0x2c) plen 17 [hci0] 4.749508 Status: Connection Rejected due to Limited Resources (0x0d) Handle: 0 Address: xx:xx:xx:xx:xx:D9 (OUI 20-68-9D) Link type: eSCO (0x02) Transmission interval: 0x0c Retransmission window: 0x06 RX packet length: 60 TX packet length: 60 Air mode: Transparent (0x03) Signed-off-by:Marcel Holtmann <marcel@holtmann.org> Signed-off-by:
-
Frédéric Dalleau authored
Older Bluetooth devices may not support Setup Synchronous Connection or SCO transparent data. This is indicated by the corresponding LMP feature bits. It is not possible to know if the adapter support these features before setting BT_VOICE option since the socket is not bound to an adapter. An adapter can also be added after the socket is created. The socket can be bound to an address before adapter is plugged in. Thus, on a such adapters, if user request BT_VOICE_TRANSPARENT, outgoing connections fail on connect() and returns -EOPNOTSUPP. Incoming connections do not fail. However, they should only be allowed depending on what was specified in Write_Voice_Settings command. EOPNOTSUPP is choosen because connect() system call is failing after selecting route but before any connection attempt. Signed-off-by:
-
Frédéric Dalleau authored
This patch defines constants and macro for transparent data LMP features. It refers to Bluetooth Core V4.0 specification, Part C, Chap 3.3 which defines LMP feature mask. Signed-off-by:
-
Frédéric Dalleau authored
In order to establish a transparent SCO connection, the correct settings must be specified in the Setup Synchronous Connection request. For that, a setting field is added to ACL connection data to set up the desired parameters. The patch also removes usage of hdev->voice_setting in CVSD connection and makes use of T2 parameters for transparent data. Signed-off-by:
-
Frédéric Dalleau authored
When an incoming eSCO connection is requested, check the selected voice setting and reply appropriately. Voice setting should have been negotiated previously. For example, in case of HFP, the codec is negotiated using AT commands on the RFCOMM channel. This patch only changes replies for socket with deferred setup enabled. Signed-off-by:
-
Frédéric Dalleau authored
This patch defines constants for SCO airmode from SCO voice setting. It refers to Bluetooth Core V4.0 specification, Part E, Chap 6.12 which describe SCO voice setting format. Signed-off-by:
-
Frédéric Dalleau authored
This patch extends the current Bluetooth socket options with BT_VOICE. This is intended to choose voice data type at runtime. It only applies to SCO sockets. Incoming connections shall be setup during deferred setup. Outgoing connections shall be setup before connect(). The desired setting is stored in the SCO socket info. This patch declares needed members, modifies getsockopt() and setsockopt(). Signed-off-by:
-
Frédéric Dalleau authored
From Bluetooth Core v4.0 specification, 7.1.8 Accept Connection Request Command "When accepting synchronous connection request, the Role parameter is not used and will be ignored by the BR/EDR Controller." Signed-off-by:
-
Frédéric Dalleau authored
hci_connect is a super function for connecting hci protocols. But the voice_setting parameter (introduced in subsequent patches) is only needed by SCO and security requirements are not needed for SCO channels. Thus, it makes sense to have a separate function for SCO. Signed-off-by:
-
Gianluca Anzolin authored
In rfcomm_tty_cleanup we purge the dlc->tx_queue which may contain socket buffers referencing the tty_port and thus preventing the tty_port destruction. Signed-off-by:
Gianluca Anzolin <gianluca@sottospazio.it> Reviewed-by:
Peter Hurley <peter@hurleysoftware.com> Signed-off-by:
-
Gianluca Anzolin authored
The tty_port can be released in two cases: when we get a HUP in the functions rfcomm_tty_hangup() and rfcomm_dev_state_change(). Or when the user releases the device in rfcomm_release_dev(). In these cases we set the flag RFCOMM_TTY_RELEASED so that no other function can get a reference to the tty_port. The use of !test_and_set_bit(RFCOMM_TTY_RELEASED) ensures that the 'initial' tty_port reference is only dropped once. The rfcomm_dev_del function is removed becase it isn't used anymore. Signed-off-by:
-
Gianluca Anzolin authored
Implement .activate, .shutdown and .carrier_raised methods of tty_port to manage the dlc, moving the code from rfcomm_tty_install() and rfcomm_tty_cleanup() functions. At the same time the tty .open()/.close() and .hangup() methods are changed to use the tty_port helpers that properly call the aforementioned tty_port methods. Signed-off-by:
-
Gianluca Anzolin authored
Move the tty_struct initialization from rfcomm_tty_open() to rfcomm_tty_install() and do the same for the cleanup moving the code from rfcomm_tty_close() to rfcomm_tty_cleanup(). Add also extra error handling in rfcomm_tty_install() because, unlike .open()/.close(), .cleanup() is not called if .install() fails. Signed-off-by:
-
Gianluca Anzolin authored
The current code removes the device from the device list in several places. Do it only in the destructor instead and in the error path of rfcomm_add_dev() if the device couldn't be initialized. Signed-off-by:
-
Gianluca Anzolin authored
In net/bluetooth/rfcomm/tty.c the struct tty_struct is used without taking references. This may lead to a use-after-free of the rfcomm tty. Fix this by taking references properly, using the tty_port_* helpers when possible. The raw assignments of dev->port.tty in rfcomm_tty_open/close are addressed in the later commit 'rfcomm: Implement .activate, .shutdown and .carrier_raised methods'. Signed-off-by:
-
Marcel Holtmann authored
In case of a Low Energy only controller it makes no sense to configure the full BR/EDR event mask. It will just enable events that can not be send anyway and there is no guarantee that such a controller will accept this value. Use event mask 0x90 0xe8 0x04 0x02 0x00 0x80 0x00 0x20 for LE-only controllers which enables the following events: Disconnection Complete Encryption Change Read Remote Version Information Complete Command Complete Command Status Hardware Error Number of Completed Packets Data Buffer Overflow Encryption Key Refresh Complete LE Meta This is according to Core Specification, Part E, Section 3. Signed-off-by: -
Johan Hedberg authored
When a socket is in deferred state there does actually exist an underlying connection even though the connection state is not yet BT_CONNECTED. In the deferred state it should therefore be allowed to get socket options that usually depend on a connection, such as SCO_OPTIONS and SCO_CONNINFO. This patch fixes the behavior of some user space code that behaves as follows without it: $ sudo tools/btiotest -i 00:1B:DC:xx:xx:xx -d -s accept=2 reject=-1 discon=-1 defer=1 sec=0 update_sec=0 prio=0 voice=0x0000 Listening for SCO connections bt_io_get(OPT_DEST): getsockopt(SCO_OPTIONS): Transport endpoint is not connected (107) Accepting connection Successfully connected to 60:D8:19:xx:xx:xx. handle=43, class=000000 The conditions that the patch updates the if-statements to is taken from similar code in l2cap_sock.c which correctly handles the deferred state. Signed-off-by:
-
Andy Shevchenko authored
There are two places where DIV_ROUND_UP may be used. It makes code a bit clearer. Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by:
-
- 16 Aug, 2013 15 commits
-
-
-
-
Gabor Juhos authored
The 2011_1007_RT5390_RT5392_Linux_STA_V2.5.0.3_DPO driver enables PCIe wakeup for these chips as well. Do the same in rt2x00. References: rt28xx_init in common/rtmp_init_intf.c RTMPInitPCIeLinkCtrlValue in os/linux/rt_rbus_pci_drv.c Signed-off-by:
-
Gabor Juhos authored
The rt2800pci driver uses the same [RT]XWI size for all chipsets, however some chips requires different values. The size of the [RT]XWI structures is a constant value for a given chipset and it does not depend on the underlying interface. Add a helper function which returns the correct values for the actual chipset and use the new helper both in the rt2800usb and in the rt2800pci drivers. This ensures that both drivers are using the correct values. Signed-off-by:
-
Simon Wunderlich authored
CSA is only enabled for one interface, but the same limitation applies for mac80211 too. It checks whether the beacon has been sent (different approaches for non-EDMA-enabled and EDMA-enabled devices), and completes the channel switch after that. Signed-off-by:
Simon Wunderlich <siwu@hrz.tu-chemnitz.de> Signed-off-by:
Mathias Kretschmer <mathias.kretschmer@fokus.fraunhofer.de> Signed-off-by:
-
Simon Wunderlich authored
Signed-off-by:
-
Simon Wunderlich authored
Signed-off-by:
-
Simon Wunderlich authored
Signed-off-by:
-
Simon Wunderlich authored
Signed-off-by:
-
Simon Wunderlich authored
Signed-off-by:
-
Simon Wunderlich authored
Signed-off-by:
-
Simon Wunderlich authored
To enable support for 5/10 MHz, some internal functions must be converted from using the (old) channel_type to chandef. This is a good chance to change all remaining occurences. Signed-off-by:
-
Simon Wunderlich authored
5/10 MHz channels should always use SIFS times as defined in IEEE 802.11-2012 18.4.4 (OFDM PHY characteristics). This makes it compatible to ath5k, which does the same. Signed-off-by:
-
Johannes Berg authored
In order to be able to (securely) keep connections alive after the system was suspended for WoWLAN, we need some additional APIs. We already have API (ieee80211_gtk_rekey_notify) to tell wpa_supplicant about the new replay counter if GTK rekeying was done by the device while the host was asleep, but that's not sufficient. If GTK rekeying wasn't done, we need to tell the host about sequence counters for the GTK (and PTK regardless of rekeying) that was used while asleep, add ieee80211_set_key_rx_seq() for that. If GTK rekeying was done, then we need to be able to disable the old keys (with ieee80211_remove_key()) and allocate the new GTK key(s) in mac80211 (with ieee80211_gtk_rekey_add()). If protocol offload (e.g. ARP) is implemented, then also the TX sequence counter for the PTK must be updated, using the new ieee80211_set_key_tx_seq() function. Signed-off-by:Johannes Berg <johannes.berg@intel.com>
-
Simon Wunderlich authored
Channel Switch will later require to generate beacons without setting them immediately. Therefore split the presp generation in an own function. Splitting the original very long function might be a good idea anyway. Signed-off-by:
-
