Setting and clearing of HCI_RUNNING flag in each and every driver is
just duplicating the same code all over the place. So instead of having
the driver do it in their hdev->open and hdev->close callbacks, set it
globally in the core transport handling.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

In all callbacks for hdev->send the status of HCI_RUNNING is checked. So
instead of repeating that code in every driver, move the check into the
hci_send_frame function before calling hdev->send.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

The commit 7bee8b08 allows the Read Verbose Config Info to fail
gracefully and not cause the controller setup to abort. It seems the
reason that command failed in the first place was the missing HCI Reset
to bring the controller in full Bluetooth mode.

Apple Bluetooth controllers start out in HID mode and when in that mode
the Read Verbose Config Info command is not allowed. Sending HCI Reset
switches the controller into full HCI mode.
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>
Signed-off-by: NJohan Hedberg <johan.hedberg@intel.com>

Add regmap ibt to support Intel Bluetooth silicon register access
over HCI. Intel BT/FM combo chip allows to read/write some registers
(e.g. FM registers) via its HCI interface.

Read/Write operations are performed via a HCI transaction composed of
a HCI command (host->controller) followed by a HCI command complete
event (controller->host). Read/Write Command opcodes can be specified
to the regmap init function.
We define data formats which are intel/vendor specific.

Register Read/Write HCI command payload (Host):
Field: | REG ADDR | MODE | DATA_LEN | DATA... |
size:  |   32b    |  8b  |    8b    |  8b*    |

Register Read HCI command complete event payload (Controller):
Field: | CMD STATUS | REG ADDR | DATA... |
size:  |     8b     |   32b    |  8b*    |

Register Write HCI command complete event payload (Controller):
Field: | CMD_STATUS |
size:  |     8b     |

Since this payload is HCI encapsulated, Little Endian byte order is
used.

Write/Read Example:

If we write 0x0000002a at address 0x00008c04, with opcode_write 0xfc5d,
The resulting transaction is (btmon trace):

< HCI Command (0x3f|0x005d) plen 10 [hci0]
        04 8c 00 00 02 04 2a 00 00 00
> HCI Event (0x0e) plen 4
      Unknown (0x3f|0x005d) ncmd 1
        00

Then, if we read the same register with opcode_read 0xfc5e:

< HCI Command (0x3f|0x005e) plen 6 [hci0]
        04 8c 00 00 02 04
> HCI Event (0x0e) plen 12 [hci0]
      Unknown (0x3f|0x005e) ncmd 1
        00 04 8c 00 00 2a 00 00 00
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

This device has always ACPI companion because driver supports only ACPI
enumeration. Therefore there is no need to test it in bcm_acpi_probe() and
we can pass it directly to acpi_dev_get_resources() (which will return
-EINVAL in case of NULL argument is passed).
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Tree wide grep for "hci_bcm" doesn't reveal there is any code registering
this platform device and "struct acpi_device_id" use for passing the
platform data looks a debug/test code leftover to me.

I'm assuming this driver effectively supports only ACPI enumeration and
thus test for ACPI_HANDLE() and platform data can be removed.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

There is no need to call acpi_match_device() in driver's probe path and
verify does it find a match to given ACPI _HIDs in .acpi_match_table as
driver/platform/acpi core code has found the match prior calling the probe.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Driver doesn't handle possible error from acpi_dev_get_resources(). Test it
and return the error code in case of error.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Caller of acpi_dev_get_resources() should free the constructed resource
list by calling the acpi_dev_free_resource_list() in order to avoid memory
leak.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

There is some unneeded code in "hci_intel" probing. First
acpi_match_device() call is needless as driver/platform/acpi core code has
already done the matching before calling the probe and the driver does not
use the returned pointer to matching _HID other than checking is it NULL.

Then tree wide grep for "hci_intel" doesn't reveal that there is any code
registering this platform device so it looks this device is always backed
with ACPI companion so also ACPI_HANDLE() test can be removed.
Signed-off-by: NJarkko Nikula <jarkko.nikula@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

This patch fixes checkpatch warnings:
- Comparison to NULL could be re-written
- no space required after a cast
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Use <linux/io.h> instead of <asm/io.h>, fixes checkpatch
Warning;
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

'retransmit' being set in HCI_IBS_TX_WAKING case, using bool would be
efficient. Initialize local bool to false.
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

NOT NULL comparison modified to be readable, reported
by checkpatch.
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

NOT NULL comparison modified to be readable, reported
by checkpatch.
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

NOT NULL comparison modified to be readable, reported
by checkpatch.
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

replaced 'not null' comparison that is readable, reported by checkpatch.
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Adds autosuspend runtime functionality to BCM UART driver.
Autosuspend is enabled at end of bcm_setup.

bcm_device_lock is used for system sleep functions as they can be
called at any time.
bcm_device_lock is not held for runtime suspend functions as this
is only enabled as long as platform device is opened.
Signed-off-by: NFrederic Danis <frederic.danis@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

rx_lock spinlock is only used in hci_uart_tty_receive() which is the
receive_buf ldisc callback.

hci_uart_tty_receive() is protected from re-entrance by its only
caller (flush_to_ldisc() in drivers/tty/tty_buffer.c) which held a
mutex (buf->lock) for this section.
This lock allows "safe use of the line discipline's receive_buf()
method by excluding the buffer work and any pending flush from using
the flip buffer." (comments from tty_buffer_lock_exclusive() in
drivers/tty/tty_buffer.c)

So, no need to double protect this resource with rx_lock.
Signed-off-by: NFrederic Danis <frederic.danis@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Change some CONFIG_PM_SLEEP to CONFIG_PM as hu and is_suspended parameters
will be used during PM runtime callbacks.

Add bcm_suspend_device() and bcm_resume_device() which performs link
management for PM callbacks.
These functions will be used for runtime management.
Signed-off-by: NFrederic Danis <frederic.danis@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

ACPI table for BCM2E39 of T100TA is not correct.
Set correct irq_polarity for this device.
Signed-off-by: NFrederic Danis <frederic.danis@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

This patch adds support for Marvell's new chipset SD8997.
Register offsets and supported feature flags are updated.
Signed-off-by: NZhaoyang Liu <liuzy@marvell.com>
Signed-off-by: NCathy Luo <cluo@marvell.com>
Signed-off-by: NAmitkumar Karwar <akarwar@marvell.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Coding style fix, extra spaces are removed to make casting
consistent.
Signed-off-by: NAmitkumar Karwar <akarwar@marvell.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

First firmware dump attempt from user works fine, but firmware goes
into bad state after this. Subsequent attempts fails.

As required by the firmware dump implementation, this change writes
FW_DUMP_READ_DONE value to dump ctrl register to address this issue.
Signed-off-by: NNachiket Kukade <kukaden@marvell.com>
Signed-off-by: NAmitkumar Karwar <akarwar@marvell.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Cleanup patch to fix spaces required reported by checkpatch
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

simple_strtoul is obsolete, use kstrtoul instead reported by checkpatch.
Signed-off-by: NPrasanna Karthik <mkarthi3@visteon.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

SCO packet reassembler may have a fragment of SCO packet, from
previous connection, cached and not removed when SCO connection
is ended. Packets from new SCO connection are then going to be
attached to that fragment, creating an invalid SCO packets.

Controllers like Intel's WilkinsPeak are always fragmenting
SCO packet into 3 parts (#1, #2, #3). Packet #1 contains
SCO header and audio data, others just audio data. if there is
a fragment cached from previous connection, i.e. #1, first
SCO packet from new connection is going to be attached to it
creating packet consisting of fragments #1-#1-#2. This will
be forwarded to upper layers. After that, fragment #3 is going
to be used as a starting point for another SCO packet.
It does not contain a SCO header, but the code expects it,
casts a SCO header structure on it, and reads whatever audio
data happens to be there as SCO packet length and handle.
From that point on, we are assembling random data into SCO
packets. Usually it recovers quickly as initial audio data
contains mostly zeros (muted stream), but setups of over
4 seconds were observed.
Issue manifests itself by printing on the console:
Bluetooth: hci0 SCO packet for unknown connection handle 48
Bluetooth: hci0 SCO packet for unknown connection handle 2560
Bluetooth: hci0 SCO packet for unknown connection handle 12288
It may also show random handles if audio data was non-zeroed.
Hcidump shows SCO packets with random length and handles.

Few messages with handle 0 at connection creation are OK
for some controllers (like WilkinsPeak), as there are SCO packets
with zeroed handle at the beginning (possible controller bug).
Few of such messages at connection end, with a handle looking
sane (around 256, 512, 768 ...) is also OK, as these are last
SCO packets that were assembled and sent up, before connection
was ended, but were not handled in time.

This issue may still manifest itself on WilkinsPeak as it sometimes,
at SCO connection creation, does not send third fragment of first
SCO packet (#1-#2-#1-#2-#3...). This is a firmware bug and this
patch does not address it.
Signed-off-by: NKuba Pawlak <kubax.t.pawlak@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

We need to explicitly enable the IRQ wakeup mode to let the controller
wake the system from sleep states (like suspend-to-ram).
PM suspend/resume callbacks now call the generic intel device PM
functions after enabling/disabling IRQ wake.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Change the way to insert LPM packets into the txq.
Use skb_queue_head instead of skb_queue_tail to always prioritise LPM
packets over potential tx queue content.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Retrieve the Interruption used by BCM device, which can be declared
as Interruption or GpioInt in the ACPI table.
Retrieve IRQ polarity from the ACPI table to use it for host_wake_active
parameter of Setup Sleep vendor specific command.
Configure BCM device to wake-up the host.
Enable IRQ wake while suspended.
Signed-off-by: NFrederic Danis <frederic.danis@linux.intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

btintel_load_ddc_config is now part of btintel.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Apply DDC parameters once controller is in operational mode.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

btintel_load_ddc_config retrieves the ddc file and sends its content
via DDC commands (opcode 0xfc8b).

The ddc file should contain one or more DDC structures.
A DDC structure is composed of the folowing fields:
field: | DDC LEN |    DDC ID    |   DDC VALUE   |
size:  | 1 byte  |    2 bytes   |  DDC LEN - 2  |
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

intel_lpm_suspend/resume are only used in case of CONFIG_PM.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Fix compilation the following compilation warning, which happens when
CONFIG_BT_INTEL is not set:

  drivers/bluetooth/btintel.h:98:13: warning: ‘btintel_version_info’
  defined but not used [-Wunused-function]
   static void btintel_version_info(struct hci_dev *hdev,
                                    struct intel_version *ver)
               ^
Signed-off-by: NVincent Stehlé <vincent.stehle@laposte.net>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

if btmrvl_tx_pkt() is called, and the branch
  if (skb_headroom(skb) < BTM_HEADER_LEN)
evaluates positive, a new skb is allocated via skb_realloc_headroom.

The original skb is stored in a tmp variable, before being free'd.
However on success, the new skb, is not free'd, nor is it
returned to the caller which will then double-free the original skb.

This issue exists from the original driver submission in
 commit: #132ff4e5

If this code path had been alive, it would have been noted from the
double-free causing a panic.

All skb's here should be allocated through bt_skb_alloc which
adds 8 bytes as headroom, which is plenty against the 4 bytes
pushed on by this driver.

This code path is dead, and buggy at the same time, so the cleanest
approach is to remove the affected branch.

Reported by coverity (CID 113422)
Signed-off-by: NKieran Bingham <kieranbingham@gmail.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Don't hide this packet size error.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Implement runtime PM suspend/resume callbacks.
If LPM supported, controller is put into supsend after a delay of
inactivity (1s). Inactivity is based on LPM idle notification and
host TX traffic.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Add PM suspend/resume callbacks which call lpm_suspend/resume.
Add LPM ack in threaded IRQ handler to notify the controller that
resume is complete.
Protect hci_uart against concurrent removing during suspend/resume.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>

Add LPM PM suspend/resume/host_wake LPM functions.
A LPM transaction is composed with a LPM request and ack/response.
Host can send a LPM suspend/resume request to the controller which
should respond with a LPM ack.
If resume is requested by the controller (irq), host has to send a LPM
ack once resumed.
Signed-off-by: NLoic Poulain <loic.poulain@intel.com>
Signed-off-by: NMarcel Holtmann <marcel@holtmann.org>