| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These buffer shall only be keep in cache for as long as they are
needed so this would cleanup any client cstates in the following
conditions:
- There is no cstate on the response
- No continuation can be found for cstate
- Different request opcode
- Respond with an error
- Client disconnect
Fixes: https://github.com/bluez/bluez/security/advisories/GHSA-3fqg-r8j5-f5xq
|
|
|
|
|
|
|
|
|
|
|
|
| |
BT core spec 5.3 promotes the usage of inclusive languages.
This CL replaces some terms with the more appropriate counterparts,
such as "central", "peripheral", "accept list", and "reject list".
Note that some suggestions come from
https://specificationrefs.bluetooth.com/language-mapping/Appropriate_Language_Mapping_Table.pdf
Also use "primary" to refer the global mgmt struct.
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch adds SPDX License Identifier and removes the license text.
-------------------------------------
License COUNT
-------------------------------------
LGPL-2.1-or-later : 62
GPL-2.0-or-later : 51
BSD-2-Clause : 2
License: LGPL-2.1-or-later
src/adv_monitor.c
src/adv_monitor.h
src/shared/mainloop-glib.c
src/shared/btp.c
src/shared/tester.c
src/shared/btsnoop.c
src/shared/gatt-helpers.h
src/shared/log.h
src/shared/util.h
src/shared/mainloop-ell.c
src/shared/hci-crypto.c
src/shared/mainloop.c
src/shared/hfp.h
src/shared/mainloop-notify.c
src/shared/io-mainloop.c
src/shared/btsnoop.h
src/shared/uhid.h
src/shared/util.c
src/shared/ringbuf.c
src/shared/btp.h
src/shared/shell.h
src/shared/io-glib.c
src/shared/queue.h
src/shared/gatt-server.h
src/shared/crypto.c
src/shared/crypto.h
src/shared/io-ell.c
src/shared/timeout-glib.c
src/shared/mgmt.c
src/shared/gatt-db.c
src/shared/mainloop-notify.h
src/shared/hfp.c
src/shared/timeout-ell.c
src/shared/shell.c
src/shared/hci.c
src/shared/log.c
src/shared/ringbuf.h
src/shared/gatt-helpers.c
src/shared/queue.c
src/shared/tty.h
src/shared/timeout-mainloop.c
src/shared/tester.h
src/shared/gap.h
src/shared/mgmt.h
src/shared/gatt-db.h
src/shared/uhid.c
src/shared/mainloop.h
src/shared/timeout.h
src/shared/att-types.h
src/shared/pcap.c
src/shared/gap.c
src/shared/att.c
src/shared/pcap.h
src/shared/hci.h
src/shared/gatt-client.h
src/shared/io.h
src/shared/gatt-client.c
src/shared/ad.c
src/shared/att.h
src/shared/gatt-server.c
src/shared/hci-crypto.h
src/shared/ad.h
License: GPL-2.0-or-later
src/attrib-server.h
src/advertising.h
src/textfile.c
src/attrib-server.c
src/uinput.h
src/log.h
src/eir.h
src/sdpd.h
src/sdpd-request.c
src/service.h
src/sdpd-database.c
src/adapter.h
src/plugin.h
src/uuid-helper.h
src/plugin.c
src/main.c
src/sdp-xml.c
src/error.h
src/oui.c
src/device.c
src/agent.h
src/sdp-client.c
src/oui.h
src/device.h
src/gatt-database.c
src/error.c
src/backtrace.c
src/sdpd-server.c
src/adapter.c
src/log.c
src/agent.c
src/backtrace.h
src/sdp-xml.h
src/hcid.h
src/dbus-common.h
src/uuid-helper.c
src/storage.c
src/profile.c
src/advertising.c
src/storage.h
src/dbus-common.c
src/profile.h
src/sdpd-service.c
src/eir.c
src/sdp-client.h
src/gatt-database.h
src/gatt-client.h
src/rfkill.c
src/gatt-client.c
src/service.c
src/textfile.h
License: BSD-2-Clause
src/shared/ecc.h
src/shared/ecc.c
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since commit b5f34f9420b50 "Bluetooth: Fix bt_sock_recvmsg return value"
in the kernel, Bluetooth sockets of type SOCK_SEQPACKET, when read()
will return the size of the packet received, which can be larger than
the the buffer passed by user space.
In this case the problem was causing a disconnection soon after
the reception of an SDP request.
Reported by: Alban Browaeys <prahal@yahoo.com>
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This allows to register Multi Profile Specification record with proper
MPSD bits set depending on currently registered services.
Service Name: Multi Profile
Service RecHandle: 0x10001
Service Class ID List:
"" (0x113b)
Profile Descriptor List:
"" (0x113a)
Version: 0x0100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch fixes processing of SDP requests which have wrong PDU size in header.
Every SDP request consists of two parts: header and data.
Header always contains 3 values: request code (identifies it's type),
transaction ID (in order to match response with request), and data's size.
According to Bluetooth specification, when the data's size is wrong,
SDP server should return SDP_ErrorResponse with "Invalid PDU Size" parameter.
But now the server doesn't respond at all, and corresponding socket on
client's side crashes.
The patch actually reverts commit 388761cdc8f8a1293bb0b1a5bd576b5fb41616ca,
because the size of request should be checked inside handle_request() function
in order to be able to respond with SDP_ErrorResponse, instead of stopping
the connection.
|
|
|
|
|
|
|
| |
This makes SDP code no longer depends on main_opts. DID record is now
registered from main() after sdp server was started. This is OK since
mainloop is not yet running and record will be present when first
request comes.
|
|
|
|
| |
Check that the full header is received in recv().
|
|
|
|
| |
Instead of checking for error check that correct number of bytes received.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 8a03376544b046a84301847d1594f6c3674983ff.
The patch needs to be split up and the gdbus/ changes were bogus
compared to the original commit message.
Conflicts:
Makefile.am
Makefile.obexd
profiles/cyclingspeed/cyclingspeed.c
profiles/heartrate/heartrate.c
src/error.c
|
|
|
|
|
|
|
|
|
|
| |
Instead of trying to include config.h in each file over the tree and
possibly forgetting to include it, give a "-include config.h" argument
to the compiler so it's guaranteed that a) it will be included for all
source files and b) it will be the first header included.
gdbus/ directory is left out, since it would break other projects using
it.
|
|
|
|
|
|
|
| |
Instead of parsing config DeviceID string on sdp server startup, parse
it in main while reading config file. This allow to store logical DID
values in main_opts (instead of raw config string) and make use of them
in bluetoothd code.
|
|
|
|
|
|
|
|
|
|
|
| |
The existing code relied on set_close_on_unref which doesn't work as
long as the io watch callbacks hold on to their own reference. To fix
this it's better to track the watch id's instead of the IO channels and
do g_source_remove instead of g_io_channel_unref.
This isn't a particularly critical fd leak since these are created once
on bluetoothd startup and closed on exit, however they do help clean up
the valgrind fd-leak report.
|
|
|
|
| |
Try to make log stuff more similar to ConnMan and oFono.
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|