| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
|
|
|
|
|
| |
Change-Id: I47d7d59c49d06149cbf27e0ac570188b451197f8
Signed-off-by: Paul Kocialkowski <contact@paulk.fr>
|
|
|
|
|
|
|
|
|
| |
Issue: AXIA-1459
Change-Id: Ie4cc5766446774a0bae3bbf7d9baa5f44e814f59
Signed-off-by: Ravindranath Doddi <ravindranathx.doddi@intel.com>
Signed-off-by: Krishnan V <krishnanx.vaidyanathan.venkitakrishnan@intel.com>
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Signed-off-by: Matt Gumbel <matthew.k.gumbel@intel.com>
|
|
|
|
|
|
|
| |
Initialization should not continue if we failed to open the hci transport.
Change-Id: Idf2008477f34ed216163ca47be637f415d640bf5
Signed-off-by: Matt Gumbel <matthew.k.gumbel@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
This lets device repos override some constants with their bdroid_buildcfg.h if
necessary.
This fixes Bluetooth issues (namely tethering) on endeavoru where
BTHC_USERIAL_READ_MEM_SIZE needs to be set to a higher value.
Change-Id: I8955963cdb2b34f1ba63f8160861699f7284814f
|
|
|
|
|
|
|
|
| |
Change nice to -16, same as android audio service. Affected tasks: btu,
gki timer, hci worker, media task, userial_read
bug 7267756
Change-Id: I0b5ecc4345240601ef6e5890240054dab193e4a0
|
|
|
|
|
|
|
|
| |
This port allows real time parsing of the btsnoop traces using an
external tool. Disabled this by default for production.
bug 7298743
Change-Id: I595e3e9fede532dbf1bd1383d9babe7c44503af1
|
|
|
|
|
|
|
|
|
|
|
| |
Root cause: HCI RX path event sometimes gets received at host HCI and
the num_hci_cmd_pkts counter gets updated before the TX path returned
from hci_mct_send_msg function. This results in next HCI command not being
sent to controller and eventually resulting in HCI command timeout
Bug 7223945, 7073814, 7150111
Change-Id: Idc6ea48d316385f99f95d9d9015d41b66cded5c8
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug: 7167816
Bug: 7136940
This could be fix for most of the hci timeout issues. The root cause seemed
like the while loop for sending hci message messed up when another thread try
to enque/deque a hci message at same time.
Change-Id: Ib7f6a68b49b700b0d575326cafae41a158483b27
Signed-off-by: Iliyan Malchev <malchev@google.com>
|
|
|
|
|
|
|
|
| |
First release under Apache source license with updated header files
and whitespace cleanups. Also fixed file permission to remove executable
bit from source files.
Change-Id: I7155903edfc2f491c1cbbd0db108f207852caaa0
|
|
|
|
|
|
|
|
| |
This reverts commit cc914cffd521b2e6e1be804c82aea2388b843ebf
More changes in the original CL after my cherry-pick.
Change-Id: Ic144dba96aeb8a60611baac7565c9f628ab8f9eb
|
|
|
|
|
|
|
| |
Also fixed file permission to remove executable bit
from source files.
Change-Id: I6954c2d16190bc35b1b0d38386543253696b1112
|
|
|
|
|
|
| |
Set thread names for easier debugging
Change-Id: I1beeaae24a7782ab97be22db57060c773721bc46
|
|
|
|
|
|
|
|
| |
Rather than using a text file and generating each buildcfg.h file
per device bdroid_buildcfg.h which supplies device specific values
and then bt_target.h will supply common values.
Change-Id: I12cb5408729c3e076562a292204f07a0c4b52628
|
|
|
|
|
|
| |
Revise BT Vendor Lib interface and add support of HCI multi-channel transport.
Change-Id: I5a7a7b64798bb30be60ed29bd564bc499bd5b5bc
|
|
|
|
|
|
|
|
|
| |
1. Move UART port configuration into vendor lib to respect the fact that each
vendor might have different UART port settings for its own chipsets.
2. Detail bt_vendor_lib.h header file with usage instructions.
Change-Id: I10dc44afd9b3cc1985769166a717089728281296
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The original baud rate change was done through closing the existing UART port
then reopening UART port at new baud rate. It was reported that this approach
would cause trouble on some platforms which spits out false start-bit signal
on UART bus during port close and open process.
Revise the way of doing baud rate change to use terminal I/O (termios) control
functions.
Change-Id: I18ab5a234576147dd9185eafd1e57d3b4e3fa2bd
Conflicts:
hci/src/userial.c
|
|
|
|
|
|
|
| |
We donnot need the extra function to return the interface.
bug 6853945
Change-Id: I693aa9f068a64ee5a5e9bc5ff3cc82ecf59d5de1
|
|
|
|
|
|
|
| |
libbt-vendor.so is in vendor/, and shouldn't be compile-time linked
against from external/. Use dlopen instead.
Change-Id: Iefc2d5232506ce8e60a6e5dd93328332549daa1e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Conflicts:
btif/src/btif_core.c
btif/src/btif_dm.c
btif/src/btif_hh.c
btif/src/btif_sock.c
gki/ulinux/gki_ulinux.c
hci/include/bt_hci_bdroid.h
hci/src/bt_hci_bdroid.c
hci/src/btsnoop.c
hci/src/hci_h4.c
hci/src/userial.c
vendor/libvendor/src/hardware.c
vendor/libvendor/src/upio.c
|
|
external/bluetooth/bluedroid).
1. Split the original external/bluetooth/bluedroid/vendor directory to two parts: external/bluetooth/bluedroid/hci and vendor/broadcom/libbt-vendor. Each part produces a .so library (libbt-hci.so and libbt-vendor.so).
2. libbt-hci.so contains:
+ HCI H4 send & receive functions
+ Userial read & write functions
+ LPM common logic operations
+ btsnoop functions.
3. No build-time (.txt) and run-time (.conf) configuration for libbt-hci.so.
4. libbt-vendor.so contains:
+ firmware patch download function
+ SCO configuration
+ LPM VSC and BT_WAKE control
+ UART port open
+ upio control for BT chip power on/off.
5. libbt-vendor.so has build-time and run-time configuration support.
Change-Id: I61d55c75c66d25459d80893a5f72bccce2b54770
|