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: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

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: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

This patch converts the hdev->link_keys list to be protected through
RCU, thereby eliminating the need to hold the hdev lock while accessing
the list.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

In __hci_cmd_sync_ev() and __hci_req_sync() if the hci_req_run() call
fails and we return from the functions we should ensure that the state
doesn't remain in TASK_INTERRUPTIBLE that we just set it to. This patch
fixes missing calls to set_current_state(TASK_RUNNING) in both places.
Reported-by: NKirill A. Shutemov <kirill@shutemov.name>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Tested-by: NKirill A. Shutemov <kirill@shutemov.name>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Doing things like hci_conn_hash_flush() while holding the hdev lock is
risky since its synchronous pending work cancellation could cause the
L2CAP layer to try to reacquire the hdev lock. Right now there doesn't
seem to be any obvious places where this would for certain happen but
it's already enough to cause lockdep to start warning against the hdev
and the work struct locks being taken in the "wrong" order:

[  +0.000373] mgmt-tester/1603 is trying to acquire lock:
[  +0.000292]  ((&conn->pending_rx_work)){+.+.+.}, at: [<c104266d>] flush_work+0x0/0x181
[  +0.000270]
but task is already holding lock:
[  +0.000000]  (&hdev->lock){+.+.+.}, at: [<c13b9a80>] hci_dev_do_close+0x166/0x359
[  +0.000000]
which lock already depends on the new lock.

[  +0.000000]
the existing dependency chain (in reverse order) is:
[  +0.000000]
-> #1 (&hdev->lock){+.+.+.}:
[  +0.000000]        [<c105ea8f>] lock_acquire+0xe3/0x156
[  +0.000000]        [<c140c663>] mutex_lock_nested+0x54/0x375
[  +0.000000]        [<c13d644b>] l2cap_recv_frame+0x293/0x1a9c
[  +0.000000]        [<c13d7ca4>] process_pending_rx+0x50/0x5e
[  +0.000000]        [<c1041a3f>] process_one_work+0x21c/0x436
[  +0.000000]        [<c1041e3d>] worker_thread+0x1be/0x251
[  +0.000000]        [<c1045a22>] kthread+0x94/0x99
[  +0.000000]        [<c140f801>] ret_from_kernel_thread+0x21/0x30
[  +0.000000]
-> #0 ((&conn->pending_rx_work)){+.+.+.}:
[  +0.000000]        [<c105e158>] __lock_acquire+0xa07/0xc89
[  +0.000000]        [<c105ea8f>] lock_acquire+0xe3/0x156
[  +0.000000]        [<c1042696>] flush_work+0x29/0x181
[  +0.000000]        [<c1042864>] __cancel_work_timer+0x76/0x8f
[  +0.000000]        [<c104288c>] cancel_work_sync+0xf/0x11
[  +0.000000]        [<c13d4c18>] l2cap_conn_del+0x72/0x183
[  +0.000000]        [<c13d8953>] l2cap_disconn_cfm+0x49/0x55
[  +0.000000]        [<c13be37a>] hci_conn_hash_flush+0x7a/0xc3
[  +0.000000]        [<c13b9af6>] hci_dev_do_close+0x1dc/0x359
[  +0.012038]        [<c13bbe38>] hci_unregister_dev+0x6e/0x1a3
[  +0.000000]        [<c12d33c1>] vhci_release+0x28/0x47
[  +0.000000]        [<c10dd6a9>] __fput+0xd6/0x154
[  +0.000000]        [<c10dd757>] ____fput+0xd/0xf
[  +0.000000]        [<c1044bb2>] task_work_run+0x6b/0x8d
[  +0.000000]        [<c1001bd2>] do_notify_resume+0x3c/0x3f
[  +0.000000]        [<c140fa70>] work_notifysig+0x29/0x31
[  +0.000000]
other info that might help us debug this:

[  +0.000000]  Possible unsafe locking scenario:

[  +0.000000]        CPU0                    CPU1
[  +0.000000]        ----                    ----
[  +0.000000]   lock(&hdev->lock);
[  +0.000000]                                lock((&conn->pending_rx_work));
[  +0.000000]                                lock(&hdev->lock);
[  +0.000000]   lock((&conn->pending_rx_work));
[  +0.000000]
 *** DEADLOCK ***

Fully fixing this would require some quite heavy refactoring to change
how the hdev lock and hci_conn instances are handled together. A simpler
solution for now which this patch takes is to try ensure that the hdev
workqueue is empty before proceeding with the various cleanup calls,
including hci_conn_hash_flush().
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

The common short form of "randomizer" is "rand" in many places
(including the Bluetooth specification). The shorter version also makes
for easier to read code with less forced line breaks. This patch renames
all occurences of "randomizer" to "rand" in the Bluetooth subsystem
code.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

This patch set converts the hdev->identity_resolving_keys list to use
RCU to eliminate the need to use hci_dev_lock/unlock.

An additional change that must be done is to remove use of
CRYPTO_ALG_ASYNC for the hdev-specific AES crypto context. The reason is
that this context is used for matching RPAs and the loop that does the
matching is under the RCU read lock, i.e. is an atomic section which
cannot sleep.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

This patch set converts the hdev->long_term_keys list to use RCU to
eliminate the need to use hci_dev_lock/unlock.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

The debufs entry for the BR/EDR whitelist is confusing since there is
a controller debugfs entry with the name white_list and both are two
different things.

With the BR/EDR whitelist, the actual interface in use is the device
list and thus just include all values from the internal BR/EDR whitelist
in the device_list debugfs entry.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

Some Bluetooth drivers require to reset the upper stack. To avoid having
all drivers send HCI Hardware Error events, provide a generic function
to wrap the reset functionality.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

__hci_cmd_sync_ev(), __hci_req_sync() could miss wake_up_interrupt from
hci_req_sync_complete() because hci_cmd_work() workqueue and its response
could be completed before they are ready to get the signal through
add_wait_queue(), set_current_state(TASK_INTERRUPTIBLE).
Signed-off-by: NChan-yeol Park <chanyeol.park@samsung.com>
Signed-off-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

Use spin_lock_bh() as the code is called from softirq in networking subsystem.
This is needed to prevent deadlocks when 6lowpan link is in use.
Signed-off-by: NJukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Fix spelling errors in comments.
Signed-off-by: NStephen Hemminger <stephen@networkplumber.org>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

The hci_recv_fragment function is no longer used by any driver and thus
do not export it. In fact it is not even needed by the core and it can
be removed altogether.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

The Bluetooth core already does processing of the HCI command header
and puts it together before sending it to the driver. It is not really
efficient for the driver to look at the HCI command header again in
case it has to make certain decisions about certain commands. To make
this easier, just provide the opcode as part of the SKB control buffer
information. The extra information about the opcode is optional and
only provided for HCI commands.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

The hci_update_random_address will clear the RPA_EXPIRED flag and
proceed with setting a new one if the flag was set. However, the
set_random_addr() function that is called may choose to defer the update
to a later moment. In such a case the flag would incorrectly remain
unset unless set_random_addr() re-sets it. This patch fixes the issue.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

The necessary steps for freeing connection paramaters have grown quite a
bit so we can simplify the code by factoring it out into its own
function.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Wherever we keep hci_conn pointers around we should be using
hci_conn_get/put to ensure that they stay valid. This patch fixes
all places violating against the principle currently.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Recently the LE passive scanning and auto-connections feature was
introduced. It uses the hci_connect_le() API which returns a hci_conn
along with a reference count to that object. All previous users would
tie this returned reference to some existing object, such as an L2CAP
channel, and there'd be no leaked references this way. For
auto-connections however the reference was returned but not stored
anywhere, leaving established connections with one higher reference
count than they should have.

Instead of playing special tricks with hci_conn_hold/drop this patch
associates the returned reference from hci_connect_le() with the object
that in practice does own this reference, i.e. the hci_conn_params
struct that caused us to initiate a connection in the first place. Once
the connection is established or fails to establish this reference is
removed appropriately.

One extra thing needed is to call hci_pend_le_actions_clear() before
calling hci_conn_hash_flush() so that the reference is cleared before
the hci_conn objects are fully removed.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Now that we have per-adapter SMP data thanks to the root SMP L2CAP
channel we can take advantage of it and attach the AES crypto context
(only used for SMP) to it. This means that the smp_irk_matches() and
smp_generate_rpa() function can be converted to internally handle the
AES context.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

As preparation for moving SMP to use l2cap_chan infrastructure we need
to move the (de)initialization functions to smp.c (where they'll
eventually need access to the local L2CAP channel callbacks).
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

First of all, it's wasteful to initialize SMP if it's never going to be
used (e.g. on non-LE controllers). Second of all, when we move to use
l2cap_chan we need to know the real local address, meaning we must have
completed at least part of the HCI init. This patch moves the SMP
initialization to after the HCI init procedure and makes it depend on
whether the controller actually supports LE.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

As preparation for converting SMP to use the l2cap_chan infrastructure
refactor the (de)initialization into separate functions.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

If the AES crypto has not been initialized properly we should cleanly
return from the hci_find_irk_by_rpa() function. Right now this will not
happen in practice, but once (in subsequent patches) SMP init is moved
to after the HCI init procedure it is possible that the pointer is NULL.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

If the AES crypto context is not available we cannot generate new RPAs.
We should therefore cleanly return an error from the function
responsible for updating the random address.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

When we're not connectable and all whitelisted (BR/EDR) devices are
connected it doesn't make sense to keep page scan enabled. This patch
adds code to check for any disconnected whitelist devices and if there
are none take the appropriate action in the hci_update_page_scan()
function to disable page scan.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Similar to our hci_update_background_scan() function we can simplify a
lot of code by creating a unified helper function for doing page scan
updates. This patch adds such a function to hci_core.c and updates all
the relevant places to use it.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

The HCI_PAIRABLE flag isn't actually controlling whether we're pairable
but whether we're bondable. Therefore, rename it accordingly.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

With the Bluetooth 4.1 specification the Simultaneous LE and BR/EDR
controller option has been deprecated. It shall be set to zero and
ignored otherwise.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

Expose the default values for minimum and maximum LE advertising
interval via debugfs for testing purposes.
Signed-off-by: NGeorg Lukas <georg@op-co.de>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Store the default values for minimum and maximum advertising interval
with all the other controller defaults. These vaules are sent to the
adapter whenever advertising is (re)enabled.
Signed-off-by: NGeorg Lukas <georg@op-co.de>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Devices using resolvable private addresses are required to provide
an identity resolving key. These devices can not be found using
the current controller white list support. This means if the kernel
knows about any devices with an identity resolving key, the white
list filtering must be disabled.

However so far the kernel kept identity resolving keys around even
for devices that are not using resolvable private addresses. The
notification to userspace clearly hints to not store the key and
so it is best to just remove the key from the kernel as well at
that point.

With this it easy now to detect when using the white list is
possible or when kernel side resolving of addresses is required.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

The Bluetooth controller can use a white list filter when scanning
to avoid waking up the host for devices that are of no interest.

Devices marked as reporting, direct connection (incoming) or general
connection are now added to the controller white list. The update of
the white list happens just before enabling passive scanning.

In case the white list is full and can not hold all devices, the
white list is not used and the filter policy set to accept all
advertisements.

Using the white list for scanning allows for power saving with
controllers that do not handle the duplicate filtering correctly.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

When adding remote devices to the kernel using the Add Device management
command, these devices are explicitly allowed to connect. This kind of
incoming connections are possible even when the controller itself is
not connectable.

For BR/EDR this distinction is pretty simple since there is only one
type of incoming connections. With LE this is not that simple anymore
since there are ADV_IND and ADV_DIRECT_IND advertising events.

The ADV_DIRECT_IND advertising events are send for incoming (slave
initiated) connections only. And this is the only thing the kernel
should allow when adding devices using action 0x01. This meaning
of incoming connections is coming from BR/EDR and needs to be
mapped to LE the same way.

Supporting the auto-connection of devices using ADV_IND advertising
events is an important feature as well. However it does not map to
incoming connections. So introduce a new action 0x02 that allows
the kernel to connect to devices using ADV_DIRECT_IND and in addition
ADV_IND advertising reports.

This difference is represented by the new HCI_AUTO_CONN_DIRECT value
for only connecting to ADV_DIRECT_IND. For connection to ADV_IND and
ADV_DIRECT_IND the old value HCI_AUTO_CONN_ALWAYS is used.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

If the Bluetooth controller supports Get MWS Transport Layer
Configuration command, then issue it during initialization.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

If the Bluetooth controller supports Read Local Supported Codecs
command, then issue it during initialization so that the list of
codecs is known.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

It's safer practice to use sizeof(*ptr) instead of sizeof(ptr_type) when
allocating memory in case the type changes. This also fixes the
following style of warnings from static analyzers:

CHECK: Prefer kzalloc(sizeof(*ie)...) over kzalloc(sizeof(struct inquiry_entry)...)
+	ie = kzalloc(sizeof(struct inquiry_entry), GFP_KERNEL);
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

When HCI_CONNECTABLE is set the code has been enabling passive scanning
in order to be consistent with BR/EDR and accept connections from any
device doing directed advertising to us. However, some hardware
(particularly CSR) can get very noisy even when doing duplicates
filtering, making this feature waste resources.

Considering that the feature is for fairly corner-case use (devices
who'd use directed advertising would likely be in the whitelist anyway)
it's better to disable it for now. It may still be brought back later,
possibly with a better implementation (e.g. through improved scan
parameters).
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

The current "wait_on_bit" interface requires an 'action'
function to be provided which does the actual waiting.
There are over 20 such functions, many of them identical.
Most cases can be satisfied by one of just two functions, one
which uses io_schedule() and one which just uses schedule().

So:
 Rename wait_on_bit and        wait_on_bit_lock to
        wait_on_bit_action and wait_on_bit_lock_action
 to make it explicit that they need an action function.

 Introduce new wait_on_bit{,_lock} and wait_on_bit{,_lock}_io
 which are *not* given an action function but implicitly use
 a standard one.
 The decision to error-out if a signal is pending is now made
 based on the 'mode' argument rather than being encoded in the action
 function.

 All instances of the old wait_on_bit and wait_on_bit_lock which
 can use the new version have been changed accordingly and their
 action functions have been discarded.
 wait_on_bit{_lock} does not return any specific error code in the
 event of a signal so the caller must check for non-zero and
 interpolate their own error code as appropriate.

The wait_on_bit() call in __fscache_wait_on_invalidate() was
ambiguous as it specified TASK_UNINTERRUPTIBLE but used
fscache_wait_bit_interruptible as an action function.
David Howells confirms this should be uniformly
"uninterruptible"

The main remaining user of wait_on_bit{,_lock}_action is NFS
which needs to use a freezer-aware schedule() call.

A comment in fs/gfs2/glock.c notes that having multiple 'action'
functions is useful as they display differently in the 'wchan'
field of 'ps'. (and /proc/$PID/wchan).
As the new bit_wait{,_io} functions are tagged "__sched", they
will not show up at all, but something higher in the stack.  So
the distinction will still be visible, only with different
function names (gds2_glock_wait versus gfs2_glock_dq_wait in the
gfs2/glock.c case).

Since first version of this patch (against 3.15) two new action
functions appeared, on in NFS and one in CIFS.  CIFS also now
uses an action function that makes the same freezer aware
schedule call as NFS.
Signed-off-by: NNeilBrown <neilb@suse.de>
Acked-by: David Howells <dhowells@redhat.com> (fscache, keys)
Acked-by: Steven Whitehouse <swhiteho@redhat.com> (gfs2)
Acked-by: NPeter Zijlstra <peterz@infradead.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Steve French <sfrench@samba.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: http://lkml.kernel.org/r/20140707051603.28027.72349.stgit@notabene.brownSigned-off-by: NIngo Molnar <mingo@kernel.org>

To make the code more understandable it makes sense to use the new HCI
defines for connection role instead of a "bool master" parameter. This
makes it immediately clear when looking at the function calls what the
last parameter is describing.
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>