| 938a201a |
android: Add compat_includes to EXTRA_DIST
Those files are used only when building for Android and should be
part of release tarball. |
Szymon Janc |
10 years ago |
1 file, +3, -0 |
| b99c419f |
client/display: Fix compilation warning on Android
isprint requires ctype.h include.
target thumb C: btmgmt <= external/bluetooth/bluez/client/display.c
external/bluetooth/bluez/client/display.c: In function 'rl_hexdump':
external/bluetooth/bluez/client/display.c:81:3: warning: implicit
declaration of function 'isprint' [-Wimplicit-function-declaration]
str[(i % 16) + 51] = isprint(buf[i]) ? buf[i] : '.';
^ |
Szymon Janc |
10 years ago |
1 file, +1, -0 |
| 91994a70 |
tools/btmgmt: Fix compilation warning on Android
isspace requires ctype.h include.
external/bluetooth/bluez/tools/btmgmt.c: In function 'get_input':
external/bluetooth/bluez/tools/btmgmt.c:834:2: warning: implicit
declaration of function 'isspace' [-Wimplicit-function-declaration]
while (len && isspace(buf[len - 1]))
^ |
Szymon Janc |
10 years ago |
1 file, +1, -0 |
| 30e5bd9d |
android: Fix btmgmt build for Android
Android doesn't provide readline and wordexp libraries. Since
interactive mode is not necessary for now just provide stubbed
implementation.
Since those headers are loosly based on those libraries original
licence and copyright are present. |
Szymon Janc |
10 years ago |
4 files, +187, -0 |
| 52362c5f |
attrib: Fix condition check for attr delete
Checks handle value for non-zero as well as >= 0xffff, to avoid
infinite loop and deletion of unspecified attrib handles. |
Bharat Panda |
10 years ago |
1 file, +7, -1 |
| 191f4b66 |
tools/btgatt-client: Add support for set/get security on the link |
Lukasz Rymanowski |
10 years ago |
1 file, +64, -0 |
| c21c706a |
shared/gatt-client: Expose API to set/get security level on att |
Lukasz Rymanowski |
10 years ago |
2 files, +20, -0 |
| 64023666 |
shared/att: Add set/get security level |
Lukasz Rymanowski |
10 years ago |
2 files, +75, -0 |
| 0a657317 |
gatt-server: Fix mutlpile definition of same func
Removed 'put_uuid_le' func from gatt-server,
which was already defined in gatt-helpers. |
Gowtham Anandha Babu |
10 years ago |
3 files, +3, -22 |
| 812dea7b |
shared/gatt-client: Fix invalid read
The following backtrace can be reproduced when bt_gatt_client_unref is
with a pending request:
Invalid read of size 1
at 0x428D37: cancel_request (gatt-client.c:1766)
by 0x43235F: queue_remove_all (queue.c:387)
by 0x429C3D: bt_gatt_client_cancel_all (gatt-client.c:1786)
by 0x429C70: bt_gatt_client_free (gatt-client.c:1516)
by 0x429DF0: bt_gatt_client_unref (gatt-client.c:1639)
by 0x41F3AD: destroy_context (test-gatt.c:279)
by 0x41F3AD: context_quit (test-gatt.c:302)
by 0x41F817: test_read_cb (test-gatt.c:672)
by 0x427030: read_cb (gatt-client.c:1835)
by 0x42465B: handle_rsp (att.c:600)
by 0x42465B: can_read_data (att.c:732)
by 0x4312BA: watch_callback (io-glib.c:170)
by 0x4E7EAEA: g_main_context_dispatch (in /usr/lib64/libglib-2.0.so.0.4200.1)
by 0x4E7EE87: ??? (in /usr/lib64/libglib-2.0.so.0.4200.1)
Address 0x57e5748 is 8 bytes inside a block of size 40 free'd
at 0x4C2ACE9: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x428C8F: request_unref (gatt-client.c:153)
by 0x423CA6: cancel_att_send_op (att.c:204)
by 0x423CA6: bt_att_cancel (att.c:1069)
by 0x428D36: cancel_request (gatt-client.c:1764)
by 0x43235F: queue_remove_all (queue.c:387)
by 0x429C3D: bt_gatt_client_cancel_all (gatt-client.c:1786)
by 0x429C70: bt_gatt_client_free (gatt-client.c:1516)
by 0x429DF0: bt_gatt_client_unref (gatt-client.c:1639)
by 0x41F3AD: destroy_context (test-gatt.c:279)
by 0x41F3AD: context_quit (test-gatt.c:302)
by 0x41F817: test_read_cb (test-gatt.c:672)
by 0x427030: read_cb (gatt-client.c:1835)
by 0x42465B: handle_rsp (att.c:600)
by 0x42465B: can_read_data (att.c:732) |
Luiz Augusto von Dentz |
10 years ago |
1 file, +5, -2 |
| c16acfd5 |
obexd/client: Add check for invalid path
This adds a check for invalid paths in obc_session_setpath. |
Luiz Augusto von Dentz |
10 years ago |
1 file, +6, -0 |
| 5166b5de |
android/gatt: Remove redundant matching function
If NULL is passed as matching function, queue uses its internal
'direct_match()' thus there is no need to duplicate this function. |
Jakub Tyszkowski |
10 years ago |
1 file, +4, -9 |
| 2d927d8d |
android/gatt: Remove not needed checks and jumps
In the end we check the status anyway so its safe to break out of switch
statement. There is also no point in calling 'goto label' when the 'label'
is right after switch statement closing bracket. |
Jakub Tyszkowski |
10 years ago |
1 file, +1, -7 |
| 27c8b368 |
android/gatt: Improve send_dev_complete_response
For simple reads or writes there is no need to take the request out
of the queue as it is destroyed at the end anyway. We can peek and
check for errors even before switch statement. |
Jakub Tyszkowski |
10 years ago |
1 file, +11, -36 |
| 25c98225 |
android/gatt: Make struct destructors NULL proof
Some destructor functions do check for NULL and for some we need to
check manually. In few cases we forgot to check at all, which may lead to
NULL dereference. With this patch we are consistent about where this
check should be done. |
Jakub Tyszkowski |
10 years ago |
1 file, +17, -10 |
| 5a8f1719 |
android/gatt: Simplify app disconnection function call flow
We do not keep app_connections in disconnected state and disconnecting
is equal to app_connection destruction. Thus we can use more generic
'queue_remove()' with proper destructor instead of custom functions like
device_disconnect_client and app_disconnect_devices. This destructor will
notify app about disconnection. |
Jakub Tyszkowski |
10 years ago |
1 file, +20, -58 |
| 711e748e |
android/gatt: Destroy app connections before destroying apps
This is the right destruction order to avoid dangling pointers.
And it will actually matter when we make app connection destructor
automatically notify apps about disconnection. |
Jakub Tyszkowski |
10 years ago |
1 file, +3, -3 |
| 6efb2b3f |
android/gatt: Replace request state with boolean
As there are two states we can instead use boolean value. |
Jakub Tyszkowski |
10 years ago |
1 file, +7, -12 |
| 89f1181f |
android/pts: Update test notes for L2CAP
Daemon should be stopped for test purposes. |
Mariusz Skamra |
10 years ago |
1 file, +2, -3 |
| ec0f244c |
android/pts: Update HSP pixits, pics and test results for PTS 6.0
This patch updates test pixits, pics and results for HSP profile against
android 5.0 |
Grzegorz Kolodziejczyk |
10 years ago |
4 files, +10, -10 |
| 584afa71 |
tools/btmgmt: Fix end of array checks
all_cmd arrays is not terminated by zeroed element. |
Szymon Janc |
10 years ago |
1 file, +4, -4 |
| 0795eeb3 |
doc/test-coverage: Update l2cap-tester & mgmt-tester test count |
Johan Hedberg |
10 years ago |
1 file, +3, -3 |
| 470caaf8 |
doc/mgmt-api: Update error codes for Pair Device Command
This patch adds two error codes for Pair device command.
Reject error code which is used when requested transport is not enabled.
E.g. User wants to do LE pair but LE support is not enabled. Similar
with BREDR
Not Supported error code which is used if controller is not capable with
requested transport |
Lukasz Rymanowski |
10 years ago |
1 file, +8, -1 |
| a47fc1b5 |
tools/l2cap-tester: Add L2CAP LE Connection Reject test
This test checks if kernel correctly rejects LE connection when host is
not configured to support LE. |
Lukasz Rymanowski |
10 years ago |
1 file, +43, -2 |
| a7e2d72a |
tools/mgmt-tester: Add test for pair on not supported transport
If controller is not capable for giver transport, kernel will reply on
pair request with MGMT_STATUS_NOT_SUPPORTED |
Lukasz Rymanowski |
10 years ago |
1 file, +26, -0 |
| 7f310d59 |
tools/mgmt-tester: Add test for pair on not enabled transport
If transport is not enabled but controller is capable, kernel replies
with MGMT_STATUS_REJECTED |
Lukasz Rymanowski |
10 years ago |
1 file, +43, -2 |
| d9e800cb |
monitor: Use local includes for libbluetooth headers |
Marcel Holtmann |
10 years ago |
7 files, +12, -12 |
| afdd9634 |
emulator: Use local includes for libbluetooth headers |
Marcel Holtmann |
10 years ago |
9 files, +17, -17 |
| ab103f9f |
tools/btmgmt: Fix end of array checks in cmd_generator() |
Johan Hedberg |
10 years ago |
1 file, +5, -5 |
| 05f0a153 |
tools/btmgmt: Fix crash in completion in interactive mode
Use separate indexes while iterating over all_cmd and interactive_cmd.
Fix following crash:
[mgmt]# ==2224== Invalid read of size 1
==2224== at 0x4A092F2: strlen (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==2224== by 0x323C8860AD: strdup (in /usr/lib64/libc-2.18.so)
==2224== by 0x323EC1D550: rl_completion_matches (in /usr/lib64/libreadline.so.6.2)
==2224== by 0x402BBC: cmd_completion (btmgmt.c:3427)
==2224== by 0x323EC1D608: ??? (in /usr/lib64/libreadline.so.6.2)
==2224== by 0x323EC1D783: rl_complete_internal (in /usr/lib64/libreadline.so.6.2)
==2224== by 0x323EC156DD: _rl_dispatch_subseq (in /usr/lib64/libreadline.so.6.2)
==2224== by 0x323EC159FF: readline_internal_char (in /usr/lib64/libreadline.so.6.2)
==2224== by 0x323EC2AB6C: rl_callback_read_char (in /usr/lib64/libreadline.so.6.2)
==2224== by 0x4032E8: prompt_read (btmgmt.c:3551)
==2224== by 0x419048: io_callback (io-mainloop.c:123)
==2224== by 0x419842: mainloop_run (mainloop.c:157)
==2224== Address 0x68 is not stack'd, malloc'd or (recently) free'd |
Szymon Janc |
10 years ago |
1 file, +7, -6 |
| cd7eed33 |
android: Fix build for Android
mainloop.c was moved from monitor to shared folder. |
Szymon Janc |
10 years ago |
1 file, +4, -4 |
| 3d3f496f |
tools/btmgmt: Add support for 'hciX' form for specifying index |
Johan Hedberg |
10 years ago |
1 file, +2, -0 |
| beadb6d7 |
tools/btmgmt: Add index-dependent prompt |
Johan Hedberg |
10 years ago |
1 file, +23, -1 |
| 353c020e |
tools/btmgmt: Add support for switching selected index at runtime |
Johan Hedberg |
10 years ago |
1 file, +67, -38 |
| 649215b6 |
shared: Move mainloop implementation from monitor to shared |
Johan Hedberg |
10 years ago |
33 files, +485, -483 |
| fd6a3480 |
tools/btmgmt: Create separate table for interactive-only commands |
Johan Hedberg |
10 years ago |
1 file, +74, -41 |
| 02cfd662 |
tools/btmgmt: Remove unneeded bt_gap code
The bt_gap API is a useful one, but btmgmt really only needs to access
shared/mgmt, at least for now. |
Johan Hedberg |
10 years ago |
1 file, +0, -15 |
| 37f13066 |
tools/btmgmt: Fix checking for empty arguments list |
Johan Hedberg |
10 years ago |
1 file, +3, -0 |
| 36bf0bd0 |
tools/obexctl: Use wordexp() for parsing commands |
Johan Hedberg |
10 years ago |
1 file, +12, -11 |
| 5f2a5127 |
tools/btmgmt: Fix release_prompt() usage |
Johan Hedberg |
10 years ago |
1 file, +32, -29 |
| bc387af6 |
tools/btmgmt: Use new explicit exit status functions |
Johan Hedberg |
10 years ago |
1 file, +5, -6 |
| 725600b4 |
monitor/mainloop: Introduce functions for specifying explicit exit status |
Johan Hedberg |
10 years ago |
2 files, +14, -0 |
| c4a39397 |
monitor/mainloop: Introduce modifiable variable for exit status |
Johan Hedberg |
10 years ago |
1 file, +4, -1 |
| 0428eda7 |
monitor/mainloop: Use proper exit status code defines instead of hard-coded values |
Johan Hedberg |
10 years ago |
1 file, +4, -4 |
| f55a78fa |
android/pts: Minor style fix in GATT PTS results
This makes scripting easier. |
Szymon Janc |
10 years ago |
1 file, +2, -1 |
| 232f971f |
tools/btmgmt: Add proper readline-based user input support |
Johan Hedberg |
10 years ago |
1 file, +194, -98 |
| a24d1bf5 |
tools/btmgmt: Rename pending to pending_index for clarity |
Johan Hedberg |
10 years ago |
1 file, +4, -4 |
| 63e76a18 |
tools/btmgmt: Remove unnecessary 'monitor' command |
Johan Hedberg |
10 years ago |
1 file, +51, -110 |
| 28fd1d53 |
tools/btmgmt: Add initial readline support |
Johan Hedberg |
10 years ago |
2 files, +699, -551 |
| 853592f9 |
audioo/AVCTP: Set initial key repeat timeout
'AVC Panel Subunit Specification' mandates that:
To help the target to decide whether a pressed button is hold
or not, it is recommended that the controller have capability
of issuing a “release” command within 300 ms, after it receives
a response to the last “press” command. The target should not
judge the user operation to be “press and hold” within this
300ms.". The kernel's default timeout is currently set to 250ms
with 33ms between repeats which means it will issue one repeat
key before 300ms window elapses. |
Kuba Pawlak |
10 years ago |
1 file, +2, -0 |