summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Bluetooth: Fix SMP debug key handlingJohan Hedberg2014-12-032-11/+14
| | | | | | | | | | | We need to keep debug keys around at least until the point that they are used - otherwise e.g. slave role behavior wouldn't work as there'd be no key to be looked up. The correct behavior should therefore be to return any stored keys but when we clean up the SMP context to remove the key from the hdev list if keeping debug keys around hasn't been requestsed. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Organize SMP crypto functions to logical sectionsJohan Hedberg2014-12-031-63/+71
| | | | | | | | | This patch organizes the various SMP crypto functions so that the LE SC functions appear in one section and the legacy SMP functions in a separate one. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Fix missing const declarations in SMP functionsJohan Hedberg2014-12-032-10/+13
| | | | | | | | | Several SMP functions take read-only data. This patch fixes the declaration of these parameters to use the const specifier as appropriate. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Introduce SMP_DBG macro for low-level debugingJohan Hedberg2014-12-031-28/+40
| | | | | | | | | | | | The various inputs & outputs of the crypto functions as well as the values of the ECDH keys can be considered security sensitive. They should therefore not end up in dmesg by mistake. This patch introduces a new SMP_DBG macro which requires explicit compilation with -DDEBUG to be enabled. All crypto related data logs now use this macro instead of BT_DBG. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add basic LE SC OOB support for remote OOB dataJohan Hedberg2014-12-031-2/+36
| | | | | | | | | | | | | This patch adds basic OOB pairing support when we've received the remote OOB data. This includes tracking the remote r value (in smp->rr) as well as doing the appropriate f4() call when needed. Previously the OOB rand would have been stored in smp->rrnd however these are actually two independent values so we need separate variables for them. Na/Nb in the spec maps to smp->prnd/rrnd and ra/rb maps to smp->rr with smp->pr to come once local OOB data is supported. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Set SMP OOB flag if OOB data is availableJohan Hedberg2014-12-031-6/+25
| | | | | | | | | If we have OOB data available for the remote device in question we should set the OOB flag appropriately in the SMP pairing request or response. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for adding remote OOB data for LEJohan Hedberg2014-12-031-3/+11
| | | | | | | | | This patch adds proper support for passing LE OOB data to the hci_add_remote_oob_data() function. For LE the 192-bit values are not valid and should therefore be passed as NULL values. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Store address type with OOB dataJohan Hedberg2014-12-034-20/+29
| | | | | | | | | To be able to support OOB data for LE pairing we need to store the address type of the remote device. This patch extends the relevant functions and data types with a bdaddr_type variable. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Unify remote OOB data functionsJohan Hedberg2014-12-033-39/+22
| | | | | | | | | There's no need to duplicate code for the 192 vs 192+256 variants of the OOB data functions. This is also helpful to pave the way to support LE SC OOB data where only 256 bit data is provided. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add SC-only mode support for SMPJohan Hedberg2014-12-031-0/+9
| | | | | | | | | | | When Secure Connections-only mode is enabled we should reject any pairing command that does not have Secure Connections set in the authentication requirements. This patch adds the appropriate logic for this to the command handlers of Pairing Request/Response and Security Request. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add full SMP BR/EDR supportJohan Hedberg2014-12-032-22/+212
| | | | | | | | | | | When doing SMP over BR/EDR some of the routines can be shared with the LE functionality whereas others needs to be split into their own BR/EDR specific branches. This patch implements the split of BR/EDR specific SMP code from the LE-only code, making sure SMP over BR/EDR works as specified. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add skeleton for BR/EDR SMP channelJohan Hedberg2014-12-033-18/+74
| | | | | | | | This patch adds the very basic code for creating and destroying SMP L2CAP channels for BR/EDR connections. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add debugfs switch for forcing SMP over BR/EDRJohan Hedberg2014-12-032-0/+48
| | | | | | | | | | | To make it possible to use LE SC functionality over BR/EDR with pre-4.1 controllers (that do not support BR/EDR SC links) it's useful to be able to force LE SC operations even over a traditional SSP protected link. This patch adds a debugfs switch to force a special debug flag which is used to skip the checks for BR/EDR SC support. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add hci_conn flag for new link key generationJohan Hedberg2014-12-032-0/+4
| | | | | | | | | For LE Secure Connections we want to trigger cross transport key generation only if a new link key was actually created during the BR/EDR connection. This patch adds a new flag to track this information. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Use debug keys for SMP when HCI_USE_DEBUG_KEYS is setJohan Hedberg2014-12-031-9/+18
| | | | | | | | | | The HCI_USE_DEBUG_KEYS flag is intended to force our side to always use debug keys for pairing. This means both BR/EDR SSP as well as SMP with LE Secure Connections. This patch updates the SMP code to use the debug keys instead of generating a random local key pair when the flag is set. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add dummy handler for LE SC keypress notificationJohan Hedberg2014-12-031-0/+14
| | | | | | | | | Since we don not actively try to clear the keypress notification bit we might get these PDUs. To avoid failing the pairing process add a simple dummy handler for these for now. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Fix DHKey Check sending order for slave roleJohan Hedberg2014-12-031-17/+43
| | | | | | | | | | | According to the LE SC specification the initiating device sends its DHKey check first and the non-initiating devices sends its DHKey check as a response to this. It's also important that the non-initiating device doesn't send the response if it's still waiting for user input. In order to synchronize all this a new flag is added. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add passkey entry support for LE SCJohan Hedberg2014-12-031-11/+180
| | | | | | | | | | | | | | | | | | | The passkey entry mechanism involves either both sides requesting the user for a passkey, or one side requesting the passkey while the other one displays it. The behavior as far as SMP PDUs are concerned are considerably different from numeric comparison and therefore requires several new functions to handle it. In essence passkey entry involves both sides gradually committing to each bit of the passkey which involves 20 rounds of pairing confirm and pairing random PDUS being sent in both directions. This patch adds a new smp->passkey_round variable to track the current round of the passkey commitment and reuses the variables already present in struct hci_conn for the passkey and entered key count. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Fix BR/EDR Link Key type when derived through LE SCJohan Hedberg2014-12-032-3/+28
| | | | | | | | | | We need to set the correct Link Key type based on the properties of the LE SC pairing that it was derived from. If debug keys were used the type should be a debug key, and the authenticated vs unauthenticated information should be set on what kind of security level was reached. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for SC just-works pairingJohan Hedberg2014-12-031-12/+27
| | | | | | | | | If the just-works method was chosen we shouldn't send anything to user space but simply proceed with sending the DHKey Check PDU. This patch adds the necessary code for it. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Set correct LTK type and authentication for SCJohan Hedberg2014-12-031-1/+12
| | | | | | | | After generating the LTK we should set the correct type (normal SC or debug) and authentication information for it. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add check for accidentally generating a debug keyJohan Hedberg2014-12-031-3/+11
| | | | | | | | It is very unlikely, but to have a 100% guarantee of the generated key type we need to reject any keys which happen to match the debug key. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Detect SMP SC debug keysJohan Hedberg2014-12-031-0/+27
| | | | | | | | | | | We need to be able to detect if the remote side used a debug key for the pairing. This patch adds the debug key defines and sets a flag to indicate that a debug key was used. The debug private key (debug_sk) is also added in this patch but will only be used in a subsequent patch when local debug key support is implemented. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add selection of the SC authentication methodJohan Hedberg2014-12-031-5/+69
| | | | | | | | | | This patch adds code to select the authentication method for Secure Connections based on the local and remote capabilities. A new DSP_PASSKEY method is also added for displaying the passkey - something that is not part of legacy SMP pairing. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Track authentication method in SMP contextJohan Hedberg2014-12-031-15/+17
| | | | | | | | | | For Secure Connections we'll select the authentication method as soon as we receive the public key, but only use it later (both when actually triggering the method as well as when determining the quality of the resulting LTK). Store the method therefore in the SMP context. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for LE SC key generationJohan Hedberg2014-12-031-0/+57
| | | | | | | | | | | As the last step of the LE SC pairing process it's time to generate and distribute keys. The generation part is unique to LE SC and so this patch adds a dedicated function for it. We also clear the distribution bits for keys which are not distributed with LE SC, so that the code shared with legacy SMP will not go ahead and try to distribute them. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for LE SC DHKey check PDUJohan Hedberg2014-12-031-0/+56
| | | | | | | | Once we receive the DHKey check PDU it's time to first verify that the value is correct and then proceed with encrypting the link. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for handling LE SC user responseJohan Hedberg2014-12-031-0/+154
| | | | | | | | | | | | With LE SC, once the user has responded to the numeric comparison it's time to send DHKey check values in both directions. The DHKey check value is generated using new smp_f5 and smp_f6 cryptographic functions. The smp_f5 function is responsible for generating the LTK and the MacKey values whereas the smp_f6 function takes the MacKey as input and generates the DHKey Check value. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for LE SC numeric comparisonJohan Hedberg2014-12-031-1/+70
| | | | | | | | | | | After the Pairing Confirm and Random PDUs have been exchanged in LE SC it's time to generate a numeric comparison value using a new smp_g2 cryptographic function (which also builds on AES-CMAC). This patch adds the smp_g2 implementation and updates the Pairing Random PDU handler to proceed with the value genration and user confirmation. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add LE SC support for responding to Pairing Confirm PDUJohan Hedberg2014-12-031-0/+22
| | | | | | | | | | When LE SC is being used we should always respond to it by sending our local random number. This patch adds a convenience function for it which also contains a check for the pre-requisite public key exchange completion Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add support for sending LE SC Confirm valueJohan Hedberg2014-12-031-0/+96
| | | | | | | | | | | Once the public key exchange is complete the next step is for the non-initiating device to send a SMP Pairing Confirm PDU to the initiating device. This requires the use of a new smp_f4 confirm value generation function which in turn builds on the AES-CMAC cryptographic function. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add handler function for receiving LE SC public keyJohan Hedberg2014-12-031-0/+44
| | | | | | | | | This patch adds a handler function for the LE SC SMP Public Key PDU. When we receive the key we proceed with generating the shared DHKey value from the remote public key and local private key. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add basic support for sending our LE SC public keyJohan Hedberg2014-12-031-1/+44
| | | | | | | | | | | | | When the initial pairing request & response PDUs have been exchanged and both have had the LE SC bit set the next step is to generate a ECDH key pair and to send the public key to the remote side. This patch adds basic support for generating the key pair and sending the public key using the new Public Key SMP PDU. It is the initiating device that sends the public key first and the non-initiating device responds by sending its public key respectively (in a subsequent patch). Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add ECC library for LE Secure ConnectionsJohan Hedberg2014-12-033-1/+871
| | | | | | | | | | | | | | This patch adds a simple ECC library that will act as a fundamental building block for LE Secure Connections. The library has a simple API consisting of two functions: one for generating a public/private key pair and another one for generating a Diffie-Hellman key from a local private key and a remote public key. The code has been taken from https://github.com/kmackay/easy-ecc and modified to conform with the kernel coding style. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add basic support for AES-CMACJohan Hedberg2014-12-032-0/+11
| | | | | | | | | | Most of the LE Secure Connections SMP crypto functions build on top of the AES-CMAC function. This patch adds access to AES-CMAC in the kernel crypto subsystem by allocating a crypto_hash handle for it in a similar way that we have one for AES-CBC. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Set link key generation bit if necessary for LE SCJohan Hedberg2014-12-031-0/+10
| | | | | | | | | Depending on whether Secure Connections is enabled or not we may need to add the link key generation bit to the key distribution. This patch does the necessary modifications to the build_pairing_cmd() function. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Rename hci_find_ltk_by_addr to hci_find_ltkJohan Hedberg2014-12-034-11/+8
| | | | | | | | | Now that hci_find_ltk_by_addr is the only LTK lookup function there's no need to keep the long name anymore. This patch shortens the function name to simply hci_find_ltk. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Remove unused hci_find_ltk functionJohan Hedberg2014-12-032-30/+0
| | | | | | | | | Now that LTKs are always looked up based on bdaddr (with EDiv/Rand checks done after a successful lookup) the hci_find_ltk function is not needed anymore. This patch removes the function. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Update LTK lookup to correctly deal with SC LTKsJohan Hedberg2014-12-032-9/+36
| | | | | | | | | | | | | | LTKs derived from Secure Connections based pairing are symmetric, i.e. they should match both master and slave role. This patch updates the LTK lookup functions to ignore the desired role when dealing with SC LTKs. Furthermore, with Secure Connections the EDiv and Rand values are not used and should always be set to zero. This patch updates the LTK lookup to first use the bdaddr as key and then do the necessary verifications of EDiv and Rand based on whether the found LTK is for SC or not. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add mgmt_set_secure_conn support for any LE adapterJohan Hedberg2014-12-031-9/+8
| | | | | | | | | | Since LE Secure Connections is a purely host-side feature we should offer the Secure Connections mgmt setting for any adapter with LE support. This patch updates the supported settings value and the set_secure_conn command handler accordingly. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Use custom macro for testing BR/EDR SC enabledJohan Hedberg2014-12-034-7/+7
| | | | | | | | | | | | Since the HCI_SC_ENABLED flag will also be used for controllers without BR/EDR Secure Connections support whenever we need to check specifically for SC for BR/EDR we also need to check that the controller actually supports it. This patch adds a convenience macro for check all the necessary conditions and converts the places in the code that need it to use it. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Set the correct security level for SC LTKsJohan Hedberg2014-12-031-2/+6
| | | | | | | | | When the looked-up LTK is one generated by Secure Connections pairing the security level it gives is BT_SECURITY_FIPS. This patch updates the LTK request event handler to correctly set this level. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add mgmt support for LE Secure Connections LTK typesJohan Hedberg2014-12-033-7/+42
| | | | | | | | | We need a dedicated LTK type for LTK resulting from a Secure Connections based SMP pairing. This patch adds a new define for it and ensures that both the New LTK event as well as the Load LTKs command supports it. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Update SMP security level to/from auth_req for SCJohan Hedberg2014-12-031-4/+14
| | | | | | | | | This patch updates the functions which map the SMP authentication request to a security level and vice-versa to take into account the Secure Connections feature. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add SMP flag for SC and set it when necessary.Johan Hedberg2014-12-031-0/+7
| | | | | | | | | This patch adds a new SMP flag for tracking whether Secure Connections is in use and sets the flag when both remote and local side have elected to use Secure Connections. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Make auth_req mask dependent on SC enabled or notJohan Hedberg2014-12-031-7/+10
| | | | | | | | | If we haven't enabled SC support on our side we should use the same mask for the authentication requirement as we were using before SC support was added, otherwise we should use the extended mask for SC. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Add basic SMP defines for LE Secure ConnectionsJohan Hedberg2014-12-031-1/+24
| | | | | | | | This patch adds basic SMP defines for commands, error codes and PDU definitions for the LE Secure Connections feature. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Simplify Link Key Notification event handling logicJohan Hedberg2014-12-031-11/+14
| | | | | | | | | | | | | | | When we get a Link Key Notification HCI event we should already have a hci_conn object. This should have been created either in the Connection Request event handler, the hci_connect_acl() function or the hci_cs_create_conn() function (if the request was not sent by the kernel). Since the only case that we'd end up not having a hci_conn in the Link Key Notification event handler would be essentially broken hardware it's safe to simply bail out from the function if this happens. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Track both local and remote L2CAP fixed channel maskJohan Hedberg2014-12-022-21/+20
| | | | | | | | | | | | | | | | To pave the way for future fixed channels to be added easily we should track both the local and remote mask on a per-L2CAP connection (struct l2cap_conn) basis. So far the code has used a global variable in a racy way which anyway needs fixing. This patch renames the existing conn->fixed_chan_mask that tracked the remote mask to conn->remote_fixed_chan and adds a new variable conn->local_fixed_chan to track the local mask. Since the HS support info is now available in the local mask we can remove the conn->hs_enabled variable. Signed-off-by: Johan Hedberg <johan.hedberg@intel.com> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
* Bluetooth: Automatically flushable packets aren't allowed on LE linksSteven Walter2014-11-271-3/+11
| | | | | | | | The Bluetooth spec states that automatically flushable packets may not be sent over a LE-U link. Signed-off-by: Steven Walter <stevenrwalter@gmail.com> Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
OpenPOWER on IntegriCloud