Mateusz Moscicki [Fri, 19 Apr 2019 10:53:02 +0000 (12:53 +0200)]
tizen: Add additional unit for "unified" user session
The unified user session is about moving user session units, managed by
systemd --user, to main systemd, where it's managed as part of newly
created user@.target.
user@.target will contain same units as previously available in user/,
with same UID and environment setup. systemd instance is used for unit
to be able to specify UID (inherited from user@.target).
The rationale behind this work is following:
* VD requirement to remove user session support
* boot time optimization requirements, due to:
+ 'systemd --user' taking 1s its own startup that could be used for
unit startup
+ ability to better rearrange units if these managed by one systemd
instance
Unit installed by this commit will not be used till user login mechanism
will be changed in systemd package (via changing pam_systemd to start
user@.target, rather than user@.service).
Change-Id: I524768f116ca91d812ae0884adbb300e52817975
INSUN PYO [Mon, 29 Apr 2019 09:59:33 +0000 (18:59 +0900)]
Remove DefaultDependencies=no from dbus.socket
Change-Id: I1f1b72c237451aa04da92195c696a0387cad9e18
INSUN PYO [Mon, 29 Apr 2019 00:33:16 +0000 (09:33 +0900)]
dbus-daemon: Don't create /tmp/dbus_launch
VD only uses /tmp/dbus_launch
- VD target: use /tmp/dbus_launch generated by systemd in kdbus
- VD emulator: use /tmp/dbus_launch generated by dbus-daemon that uses VD plugin drop-in
Change-Id: If2120a016015ee76c589416dab7cc6c96ee21b05
INSUN PYO [Tue, 23 Apr 2019 08:07:39 +0000 (17:07 +0900)]
Add dependency to security-manager.socket
dbus-daemon connect to security-manager.service to get getgrouplist.
(gdb) bt
0 security_manager_groups_get_for_user (uid=0, groups=groups@entry=0xfffef754, groups_count=groups_count@entry=0xfffef758) at /usr/src/debug/security-manager-1.5.3/src/client/client-security-manager.cpp:1391
1 0xf71faa14 in _nss_securitymanager_initgroups_dyn (user=0x443f70 "root", group_gid=<optimized out>, start=0xfffef7cc, size=0xfffef800, groupsp=0xfffef804, limit=-1, errnop=0xf77ea11c)
at /usr/src/debug/security-manager-1.5.3/src/nss/nss_securitymanager.cpp:109
2 0xf759f204 in internal_getgrouplist (user=0x1 <error: Cannot access memory at address 0x1>, user@entry=0x443f70 "root", group=124, group@entry=0, size=0xfffef800, size@entry=0xfffef7f8, groupsp=0xfffef804, groupsp@entry=0xfffef7fc,
limit=limit@entry=-1) at initgroups.c:112
3 0xf759f45c in getgrouplist (user=user@entry=0x443f70 "root", group=0, groups=groups@entry=0x443f90, ngroups=ngroups@entry=0xfffef834) at initgroups.c:170
4 0xf778d940 in fill_user_info (info=info@entry=0x442ce0, uid=uid@entry=0, username=username@entry=0x0, error=0x1c, error@entry=0xfffef8b0) at dbus-sysdeps-unix.c:2410
5 0xf778db24 in _dbus_user_info_fill_uid (info=info@entry=0x442ce0, uid=uid@entry=0, error=error@entry=0xfffef8b0) at dbus-sysdeps-unix.c:2534
6 0xf7790b24 in _dbus_user_database_lookup (db=db@entry=0x4424c0, uid=<optimized out>, username=username@entry=0x0, error=error@entry=0xfffef8b0) at dbus-userdb.c:176
7 0xf7790d64 in _dbus_user_database_get_uid (db=db@entry=0x4424c0, uid=<optimized out>, info=0xfffef8ac, info@entry=0xfffef8a4, error=error@entry=0xfffef8b0) at dbus-userdb.c:662
8 0xf7790dc8 in init_system_db () at dbus-userdb.c:247
9 0xf7790f4c in init_system_db () at dbus-userdb.c:238
10 _dbus_user_database_get_system () at dbus-userdb.c:340
11 0x00421e58 in _dbus_get_user_id_and_primary_group (username=0xfffef928, uid_p=0x442584, gid_p=0x0) at dbus-userdb-util.c:210
12 0x0040f344 in start_busconfig_child (error=0xfffef9b8, attribute_values=0x0, attribute_names=0x4404f8, element_name=0x4399e0 "\250\230\003", parser=0x440d08) at config-parser.c:1048
13 bus_config_parser_start_element (parser=0x440d08, element_name=element_name@entry=0x442478 "policy", attribute_names=attribute_names@entry=0x4404f8, attribute_values=attribute_values@entry=0x442568, error=0xfffefc08) at config-parser.c:1919
14 0x0041ecc8 in expat_StartElementHandler (userData=0xfffefb18, name=0x442478 "policy", atts=0x440710) at config-loader-expat.c:107
15 0xf76c84a0 in doContent (parser=parser@entry=0x440518, startTagLevel=startTagLevel@entry=0, enc=enc@entry=0xf76e8a50 <utf8_encoding>,
s=s@entry=0x4414d0 "<busconfig>\n\n <type>accessibility</type>\n\n<servicedir>/usr/share/dbus-1/accessibility-services</servicedir>\n <auth>EXTERNAL</auth>\n\n <listen>unix:tmpdir=/tmp</listen>\n\n <policy user=\"owner\">\n <"...,
end=end@entry=0x441957 "", nextPtr=nextPtr@entry=0x440530, haveMore=haveMore@entry=0 '\000') at xmlparse.c:2890
16 0xf76c8ca0 in contentProcessor (parser=parser@entry=0x440518,
start=start@entry=0x4414d0 "<busconfig>\n\n <type>accessibility</type>\n\n<servicedir>/usr/share/dbus-1/accessibility-services</servicedir>\n <auth>EXTERNAL</auth>\n\n <listen>unix:tmpdir=/tmp</listen>\n\n <policy user=\"owner\">\n <"..., end=end@entry=0x441957 "", endPtr=endPtr@entry=0x440530) at xmlparse.c:2552
17 0xf76c9d9e in doProlog (parser=parser@entry=0x440518, enc=0xf76e8a50 <utf8_encoding>,
s=0x4414d0 "<busconfig>\n\n <type>accessibility</type>\n\n<servicedir>/usr/share/dbus-1/accessibility-services</servicedir>\n <auth>EXTERNAL</auth>\n\n <listen>unix:tmpdir=/tmp</listen>\n\n <policy user=\"owner\">\n <"...,
s@entry=0x441440 "<!DOCTYPE busconfig PUBLIC \"-//freedesktop//DTD D-Bus Bus Configuration 1.0//EN\" \"http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd\">\n<busconfig>\n\n <type>accessibility</type>\n\n<servicedir>/"...,
end=0x441957 "", tok=<optimized out>, next=<optimized out>,
next@entry=0x441449 " busconfig PUBLIC \"-//freedesktop//DTD D-Bus Bus Configuration 1.0//EN\" \"http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd\">\n<busconfig>\n\n <type>accessibility</type>\n\n<servicedir>/usr/share"...,
nextPtr=nextPtr@entry=0x440530, haveMore=haveMore@entry=0 '\000') at xmlparse.c:4579
18 0xf76ca2da in prologProcessor (parser=0x440518,
s=0x441440 "<!DOCTYPE busconfig PUBLIC \"-//freedesktop//DTD D-Bus Bus Configuration 1.0//EN\" \"http://www.freedesktop.org/standards/dbus/1.0/busconfig.dtd\">\n<busconfig>\n\n <type>accessibility</type>\n\n<servicedir>/"...,
end=<optimized out>, nextPtr=0x440530) at xmlparse.c:4293
19 0xf76cb678 in XML_ParseBuffer (parser=0x440518, len=<optimized out>, isFinal=1) at xmlparse.c:1988
20 0x0041eebc in bus_config_load (file=file@entry=0xfffefc18, is_toplevel=is_toplevel@entry=1, parent=parent@entry=0x0, error=0xfffefc08, error@entry=0x406824 <main+960>) at config-loader-expat.c:245
21 0x0040a890 in bus_context_new (config_file=0xfffefc18, flags=(BUS_CONTEXT_FLAG_FORK_NEVER | BUS_CONTEXT_FLAG_WRITE_PID_FILE), print_addr_pipe=0xfffefc00, print_pid_pipe=0xfffefc04, address=0x0, error=0xfffefc08) at bus.c:797
22 0x00406824 in main (argc=<optimized out>, argv=<optimized out>) at main.c:634
Change-Id: Ic1ee03c3e760506e72032247da36b2567d903ba3
Adrian Szyndela [Wed, 13 Mar 2019 15:39:26 +0000 (16:39 +0100)]
kdbus: share fd and pool with libdbuspolicy
This uses new functionality of libdbuspolicy: using client's pool and fd.
This way libdbuspolicy doesn't have to create its own connection,
and what is more important it doesn't have to create its own 1MB pool.
This is at the cost of using client's pool for receiving responses
to ioctl(KDBUS_CMD_GET_CONN_INFO), which are small.
Change-Id: I183a91196fead179a9fba22fa6418680305d3558
Adrian Szyndela [Mon, 18 Mar 2019 14:05:19 +0000 (15:05 +0100)]
kdbus: add assert for non-NULL member
Trying to call an unnammed method on org.freedesktop.DBus is considered
a programming error.
This detects such errors.
Change-Id: Ic341df0eef0e7ef5ab8234aacc2c256c295327c3
Adrian Szyndela [Fri, 8 Mar 2019 11:33:02 +0000 (12:33 +0100)]
kdbus: move dbuspolicy init to bus_register_kdbus
This changes the moment of checking if user is allowed to connect to
the bus. Now, it reflects standard process a bit more.
The standard process is:
- open fd (e.g. socket);
- check authentication, if needed and possible;
- connect to the bus (say hello).
In kdbus, we have only:
- open kdbus fd;
- connect to the bus (ioctl KDBUS_CMD_HELLO).
Calling libdbuspolicy for authentication fits between the two.
Additionally, and most importantly, this is required to share
the connection between libdbus and libdbuspolicy in the future.
Change-Id: Id6fe1dbc1cdc6ec774316e13fe5d60d862949476
sanghyeok.oh [Thu, 14 Feb 2019 11:04:56 +0000 (20:04 +0900)]
kdbus: Fix memory leakage in capture_org_freedesktop_DBus_StartServiceByName()
Change-Id: If4b04d0f287e199e809cdf183ce4ce779c0f4dd4
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
sanghyeok.oh [Wed, 13 Feb 2019 07:40:26 +0000 (16:40 +0900)]
DBusMessage: bugfix for wrong type specifier
Change-Id: I0743a3d67e3d6d58acc605ded013eedb6b0af9c5
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Simon McVittie [Tue, 4 Jul 2017 14:38:57 +0000 (15:38 +0100)]
dbus_message_iter_open_container: Don't leak signature on failure
If we run out of memory while calling _dbus_type_writer_recurse()
(which is impossible for most contained types, but can happen for
structs and dict-entries), then the memory we allocated in the call to
_dbus_message_iter_open_signature() will still be allocated, and we
have to free it in order to return to the state of the world prior to
calling open_container().
One might reasonably worry that this change can break callers that use
this (incorrect) pattern:
if (!dbus_message_iter_open_container (outer, ..., inner))
{
dbus_message_iter_abandon_container (outer, inner);
goto fail;
}
/* now we know inner is open, and we must close it later */
However, testing that pattern with _dbus_test_oom_handling()
demonstrates that it already dies with a DBusString assertion failure
even before this commit.
This is all concerningly fragile, and I think the next step should be
to zero out DBusMessageIter instances when they are invalidated, so
that a "double-free" is always detected.
Change-Id: I2ccd4b516c7714f64c4543dd8d2e5c99633733a5
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
Simon McVittie [Tue, 4 Jul 2017 12:31:38 +0000 (13:31 +0100)]
dbus_message_append_args_valist: Don't leak memory on inappropriate type
Found by source code inspection while trying to debug an unrelated
leak.
Change-Id: I0726c57bb4b0ccdadee2263b14f9fe3fe4ebc99a
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
Marc Mutz [Mon, 3 Oct 2016 20:19:45 +0000 (22:19 +0200)]
DBusMessage: Fix UB (misaligned access) in call to _dbus_header_set_field_basic()
The const void* 'value' pointer that is passed the address of a
uint32_t here eventually ends up in _dbus_marshal_write_basic(), which
casts it to a DBusBasicValue, a union type that has an alignment of
eight on 64-bit platforms and is therefore more-aligned than the
uint32.
The read of a value of a more-aligned type through a pointer to a less
-aligned type is undefined behaviour.
Fix by storing the uint32 in a DBusBasicValue and passing that instead.
Found by UBSan:
dbus/dbus/dbus-marshal-basic.c:832:14: runtime error: member access within misaligned address 0x7fdb8dac3a04 for type 'const union DBusBasicValue', which requires 8 byte alignment
0x7fdb8dac3a04: note: pointer points here
4a 87 b5 71 01 00 00 00 40 7d 01 00 00 61 00 00 10 3b ac 8d db 7f 00 00 2c 2a 3e 94 db 7f 00 00
^
#0 0x7fdb9444a2c3 in _dbus_marshal_write_basic dbus/dbus/dbus-marshal-basic.c:832
#1 0x7fdb943d22fb in _dbus_type_writer_write_basic_no_typecode dbus/dbus/dbus-marshal-recursive.c:1605
#2 0x7fdb943d64e9 in _dbus_type_writer_write_basic dbus/dbus/dbus-marshal-recursive.c:2327
#3 0x7fdb943c52a6 in write_basic_field dbus/dbus/dbus-marshal-header.c:318
#4 0x7fdb943c919e in _dbus_header_set_field_basic dbus/dbus/dbus-marshal-header.c:1321
#5 0x7fdb943e1349 in dbus_message_set_reply_serial dbus/dbus/dbus-message.c:1173
Change-Id: I0149da4ebbead9b4b38c8c62af1ea892e24ec95e
Signed-off-by: Marc Mutz <marc@kdab.net>
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=98035
Simon McVittie [Tue, 4 Jul 2017 13:13:15 +0000 (14:13 +0100)]
dbus_message_iter_append_basic: Don't leak signature if appending fd fails
Change-Id: I37ac4d243832476772ffbd822bbf4598e6f8f9a5
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
Philip Withnall [Thu, 2 Feb 2017 10:14:55 +0000 (10:14 +0000)]
dbus-hash: Fix a potential shift by a negative integer
As a hash table becomes unbelievably large and full, the down_shift
tends towards 0. The overflow detection code in rebuild_table() does not
prevent down_shift becoming negative, which then causes undefined
behaviour in RANDOM_INDEX for int-keyed tables.
Note that this can only happen with approaching INT_MAX entries in the
hash table, at which point we’ve almost certainly hit OOM somewhere, so
this is vanishingly unlikely to happen. This is why I can’t add a test
for the bug.
As always, thanks to Coverity.
Coverity ID: 54682
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=99641
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Change-Id: Iac3047fc0bff11b3d08c2938c0fda292bddb1466
Ralf Habacker [Tue, 10 May 2016 14:51:25 +0000 (16:51 +0200)]
Display thread id in _dbus_verbose to be able to see threading issues.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=95191
Change-Id: I16b39c8c3123436d54655a67f9d00aa090933aba
Reviewed-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
sanghyeok.oh [Tue, 29 Jan 2019 07:13:19 +0000 (16:13 +0900)]
tools: modify execute label of dbus-monitor
Unable to read 'fd passing messages' due to permission.
Set execute label as "System"
Change-Id: I765081a58cc285951216fdb63f8cf9a5c68e39de
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
sanghyeok.oh [Thu, 17 Jan 2019 07:08:51 +0000 (16:08 +0900)]
dbus-connection: Add _dbus_string_free to deallocate memory
The string must eventually be freed with _dbus_string_free().
Change-Id: I4041940684dd1dc7099bf823b22fabc357f17901
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Shin-ichi MORITA [Sat, 17 Jun 2017 11:35:30 +0000 (20:35 +0900)]
Fix missing dbus_message_unref() in error reply path
The error message was leaked when blocking on a pending call after
the connection was disconnected.
Change-Id: Icc95bcef32b2fd1d2b4d4a7db3bdf2936579c52c
Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: re-word commit message]
Reviewed-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101481
Philip Withnall [Tue, 17 Jan 2017 08:35:07 +0000 (08:35 +0000)]
dbus-connection: Remove a duplicate _dbus_list_clear() call
This was doing no harm (clearing an already-cleared list is a no-op),
but it was also pointless.
Change-Id: I21b580f01439ef2113183a116f208bf91e09ac76
Signed-off-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=99435
Adrian Szyndela [Wed, 28 Nov 2018 14:15:16 +0000 (15:15 +0100)]
dbus-daemon: add send_destination_prefix support
Change-Id: Iaa1eff400c045ec46465f248e8460c97bb183e5b
sanghyeok.oh [Thu, 3 Jan 2019 12:31:45 +0000 (21:31 +0900)]
kdbus: fixed memory leak in process_connection_info_cmd()
Change-Id: I67e577761e7596eac400bc8c35386f4c0cbf840a
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Hyotaek Shim [Tue, 4 Dec 2018 04:14:22 +0000 (04:14 +0000)]
Merge "Revert "Add RequiresMountsFor=/opt to dbus.service"" into tizen
Adrian Szyndela [Thu, 29 Nov 2018 09:23:44 +0000 (10:23 +0100)]
spec: fixed compilation options to reduce warnings
If -mimplicit-it option is set to thumb while compiling to ARM,
then it shows lots of:
Warning: conditional outside an IT block for Thumb.
This commit removes -mimplicit-it from the compilation options on ARM32 arch.
Change-Id: I6eea9ef65e61b8ec7afa16035d4a14d6d7f870e4
Hyotaek Shim [Wed, 28 Nov 2018 04:54:15 +0000 (04:54 +0000)]
Revert "Add RequiresMountsFor=/opt to dbus.service"
This reverts commit
1c9ed8666fcbae5076022fdf6d0f177d7e464ce4.
Change-Id: Ic15fcd3ace518c180910f43cc18b40c6de478d92
Simon McVittie [Mon, 21 Nov 2016 20:56:55 +0000 (20:56 +0000)]
Do not auto-activate services if we could not send a message
We specifically do not check recipient policies, because
the recipient policy is based on properties of the
recipient process (in particular, its uid), which we do
not necessarily know until we have already started it.
In this initial implementation we do not check LSMs either,
because we cannot know what LSM context the recipient process
is going to have. However, LSM support will need to be added
to make this feature useful, because StartServiceByName is
normally allowed in non-LSM environments, and is more
powerful than auto-activation anyway.
The StartServiceByName method does not go through this check,
because if access to that method has been granted, then
it's somewhat obvious that you can start arbitrary services.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=98666
Change-Id: I53ff4f6d02e631fcd09bf1c5c306b8828f075963
Simon McVittie [Fri, 16 Oct 2015 16:33:36 +0000 (17:33 +0100)]
Add tests for activation when message send/receive is denied
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: Philip Withnall <philip.withnall@collabora.co.uk>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=98666
Change-Id: I7339c1a6de69a751cbe0b9047c980c4aea53750b
Adrian Szyndela [Wed, 7 Nov 2018 10:26:11 +0000 (11:26 +0100)]
dbus-daemon: prepare activation for async security checks
This commit prepares activation function (bus_activation_acivate_service())
for returning BUS_RESULT_LATER from security check introduced
in next commits.
Change-Id: I5b37d06fc5f7e563d52ed7207b5e416bedd666e6
Michal Bloch [Mon, 5 Nov 2018 15:18:55 +0000 (16:18 +0100)]
Remove kdbus interface header
kdbus.h is now provided with other linux kernel headers.
Change-Id: Ida7d06aa1f27d88040f949fffd73f0d6cfd5f244
Signed-off-by: Michal Bloch <m.bloch@samsung.com>
pr.jung [Mon, 17 Sep 2018 07:09:50 +0000 (16:09 +0900)]
Remove build warnings
Change-Id: Ia1676731696d446a3511efb700dd89c1a1100b08
Signed-off-by: pr.jung <pr.jung@samsung.com>
Hyotaek Shim [Wed, 5 Sep 2018 09:20:00 +0000 (18:20 +0900)]
Set the smack label of executable binary tools
Apps (3rd party and even in-house Apps) are not permitted to run dbus tools directly.
User System::Tools rx
User::Shell System::Tools rx
System::TEF System::Tools rx
System::Privileged System::Tools rx
System System::Tools rx
Change-Id: Ica6d587d2516da8241590f3cf090a91ed8d3ff75
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
INSUN PYO [Thu, 9 Aug 2018 04:21:10 +0000 (13:21 +0900)]
Add RequiresMountsFor=/opt to dbus.service
In order to apply User/Group to .service, we need /opt/etc/{passwd,group}.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I585503d0dc180f21bd9594327b87b80204876005
INSUN PYO [Fri, 27 Jul 2018 02:09:01 +0000 (11:09 +0900)]
service: add dependecy to tmp.mount
Until tmp.mount is run, dbus-daemon fails because /tmp is RO.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I6bfb47f2d14d95c440efc56e3ca2fd8b1b75c6ad
Hyotaek Shim [Mon, 25 Jun 2018 07:29:21 +0000 (16:29 +0900)]
Boost dbus.service and dbus.socket for boot optimization
Change-Id: Ie47ca8ea5fbf3b1deee98ebcdcd644b021531713
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
sanghyeok.oh [Tue, 8 May 2018 10:18:16 +0000 (19:18 +0900)]
Dbus policy: apply default deny rules in the global conf file (system.conf - system bus)
This is requirement from SR, MCD, and VD Security.
Change-Id: I5ec22cd70d15fdd07c4fd3d7ad7e9289bb4ef770
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
Aleksy Barcz [Tue, 8 May 2018 09:21:10 +0000 (11:21 +0200)]
kdbus: check policy first, only then open kdbus
A process shouldn't hold an open kdbus descriptor, if according to
security policy this process has no rights to open the bus at all.
Corrected error message and code to be consistent with dbus-daemon.
Change-Id: I8c138438a21736f9241addc9ed5a616f6be19442
Adrian Szyndela [Tue, 5 Jun 2018 07:33:47 +0000 (09:33 +0200)]
dbus: a couple of corrections after static analysis
dbus-transport-kdbus.c:900
SVACE 19383: Assignment of a signed value which has type 'int'
to a variable of a bigger integer type 'dbus_uint64_t'
dbus-transport-kdbus.c:2464
SVACE 2044: Checking return value of dbus_message_get_interface().
Change-Id: Ic793c2d414aa77273b9ff3eff83a72b4f3e4d815
Adrian Szyndela [Wed, 23 May 2018 10:49:48 +0000 (12:49 +0200)]
GVariant: fix alignment of elements in array
This patch fixes two related bugs:
1. off-by-one in checking size and alignment of the next element
in_dbus_reader_get_signature_fixed_size()
2. alignment requirements were not considered at all while iterating
over array of variable size elements in array_reader_next().
Change-Id: Ibd9e1f3d11fbcd3ef0e6dbaa024e66b2568709d5
Adrian Szyndela [Thu, 24 May 2018 08:34:59 +0000 (10:34 +0200)]
dbus-marshal-gvariant: a couple of additional checks
Change-Id: I8c5bd10512682ac2df53c9b9ab89f4d4d669cd3f
Adrian Szyndela [Fri, 13 Apr 2018 14:07:53 +0000 (16:07 +0200)]
gvariant: handle immediate iteration with empty body
With GVariant, we had to do a trade-off. libdbus API does not provide
any function that would state that a message is "finished".
While creating dbus-1 messages, they are always complete.
On additions header fields can change, and all the data is simply
appended. With GVariant it is different. The format does not have
signature field anymore in the header, but it is at the end
of a message, as a part of body variant. After a body variant,
there is also a body offset appended. These values are added to a body
when a message is considered "finished". We have chosen function
dbus_message_lock() as a signal that a message is finished.
This function is always called when a message is added to output queue.
Now, what does actually happen in case of immediate iteration after
creating a message with empty body? The length of the body is zero.
This is not possible for valid GVariant as it has at least one NUL byte,
signature (minimal is '()') and a body offset. It breaks
_dbus_message_gvariant_get_body_length(). However, it can be done with
public interface, therefore this patch:
1. fixes _dbus_message_gvariant_get_body_length() to return 0 in case
of empty body, instead of computing "negative" value.
2. warns users when they try to iterate over a GVariant message
that is not locked.
Change-Id: Ie7dc331f5ea278502df02a976e555a2c7d249197
Adrian Szyndela [Tue, 10 Apr 2018 13:58:13 +0000 (15:58 +0200)]
gvariant: fix recursing into empty array
In GVariant, arrays of variable size values have offsets at the end.
We need to know how many offsets are in an array when we recurse
into it. To count the offsets we need to have offsets size and
the start and end of the offsets. The start of the offsets
is computed from the value of the last offset.
On the other hand, empty arrays have size equal to zero. In other words,
they have no offsets. Function _dbus_reader_count_array_elems missed it.
This commit fixes _dbus_reader_count_array_elems() by ensuring returning 0
when an array is empty.
Change-Id: I5f93ea89e490b321b2c2528e7bae838a1af0ec75
Hyotaek Shim [Tue, 3 Apr 2018 04:47:23 +0000 (13:47 +0900)]
Apply Full RELRO linker options ("-Wl,-z,relro,-z,now") to dbus-daemon
In addition to Partial RELRO, Full RELRO means "GOT Table becomes read-only."
Change-Id: Iaed328906e23d526f3e05209d949f3e39f76a738
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
INSUN PYO [Thu, 21 Dec 2017 03:50:04 +0000 (12:50 +0900)]
Remove unused dbus socket.
/run/user/%UID/dbus/user_bus_socket has been temporarilly remained
for previous version support.
Nobody uses this for now.
/run/user/%UID/bus is used instead.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I10a5c4153201e5a620b9460f3b5f938bbbedcea7
INSUN PYO [Wed, 22 Nov 2017 07:38:15 +0000 (16:38 +0900)]
spec: Change useradd/groupadd to handle in security-manager.
I left useradd/groupadd to make this package available from open source
without security-manager.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I3348c321a44df01c082c36bc09844a122adb669a
INSUN PYO [Wed, 22 Nov 2017 06:19:47 +0000 (15:19 +0900)]
spec: remove unused /var/lib/dbus directory
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Iee10a2a4b58f69f8bc90d7209fa6a11762fb11ad
Adrian Szyndela [Wed, 18 Oct 2017 09:27:57 +0000 (11:27 +0200)]
switch compilation mode from Thumb to ARM
This is needed to be able to put uprobes in the library code as
the uprobes in current kernel versions work only with ARM code.
Change-Id: I5e2ee2ac37fda354fa93f080e5553d2c823144ec
sanghyeok.oh [Thu, 12 Oct 2017 05:31:28 +0000 (14:31 +0900)]
kdbus: default allow receive_type=signal for session bus
Change-Id: Id6077d3f10567df5e1320cc66b9048160a2995a4
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Hyotaek Shim [Wed, 13 Sep 2017 01:58:13 +0000 (10:58 +0900)]
Remove Apache-2.0 in the dbus.spec
Change-Id: I91791d0efa0c2e42fb715c03293675da4fb7a4a3
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
Hyotaek Shim [Tue, 12 Sep 2017 01:33:35 +0000 (10:33 +0900)]
Change the license of test-runner.c from the Apache-2.0 to the MIT license to avoid conflict with LGPL-2.1+/GPL-2.0+
Change-Id: I2f1ac40f78c692571bc11b1b9e5262bdd1c6c125
Signed-off-by: Hyotaek Shim <hyotaek.shim@samsung.com>
Adrian Szyndela [Mon, 7 Aug 2017 14:38:00 +0000 (16:38 +0200)]
GVariant: fix for not having offset for empty arrays
GVariant requires offsets for variable length struct
members. Empty variable-size arrays which are at the beginning
of a struct have offset 0. However, 0 was used as
the indicator that there should be no offset added.
This patch changes 0 to the impossible value.
Change-Id: Icbc18e87831a20727f142a1218e6736e76bcce82
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
sanghyeok.oh [Tue, 18 Jul 2017 08:28:16 +0000 (17:28 +0900)]
kdbus: check policy only for method call or unicast signal
Otherwise just send it. It will checked by receiver.
Change-Id: I9652f7548adf9605111d28cdc236f36fbe319c97
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
sanghyeok.oh [Fri, 14 Jul 2017 09:09:05 +0000 (18:09 +0900)]
kdbus: libdbuspolicy: To perform policy check, pass unique name to libdbuspolicy
In case of signal broadcasting, message has null destination,
To perform policy check correctly, libdbus should pass unique name to libdbuspolicy as destination.
Change-Id: I62704ed838432823c0c11951bbb6234905167716
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Adrian Szyndela [Fri, 9 Jun 2017 09:32:40 +0000 (11:32 +0200)]
Remove legacy smack support, keep upstream smack support
This commit reverts four commits.
There was a need for additional minor changes to make it work.
Revert "GetConnectionCredentials - add smack support"
This reverts commit
13d888ca95c75c14a4919b36cc0a15979fb1d42b.
Revert "packaging: enable Smack support"
This reverts commit
54f95c99cf4d53f5566511fc0cdc21b230909b29.
Revert "Enforce smack policy from conf file"
This reverts commit
1d028039f5d34de8a1224d2ad0d8bc75b3da3ca4.
Revert "Enable checking of smack context from DBus interface"
This reverts commit
7af23db240c54e9923c76234b7e9c7a243d2d428.
Change-Id: I4bf7ce8d43717e1deb5d980f87be39220922b422
INSUN PYO [Fri, 23 Jun 2017 07:49:08 +0000 (16:49 +0900)]
spec: remove unused files
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Idbeb2980bb1ee7c1f5989ba6dc487b20f72c93b8
INSUN PYO [Mon, 27 Mar 2017 12:53:16 +0000 (21:53 +0900)]
cynara: add exception code for cynara_session_id retriving
If client exit as soon as async call, cynara_session_from_pid return null.
cynara_session_from_pid checks /proc/pid to verify client id.
===== reproduce =====
su - owner
dbus-send --type=method_call --system --dest=org.tizen.system.deviced /Org/Tizen/System/DeviceD/Display org.tizen.system.deviced.display.CustomBrightness
=====================
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Ic313e9d4f9531135865824d8512b9eb1ac6f4c9e
sanghyeok.oh [Thu, 25 May 2017 13:12:07 +0000 (22:12 +0900)]
kdbus:pass sender to dbuspolicy1_check_out
Change-Id: I23894a2e44484a896df441caa6dd3e86bbae9d78
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Adrian Szyndela [Fri, 26 May 2017 14:16:22 +0000 (16:16 +0200)]
kdbus: assure that GVariant footer is not split
Change-Id: I4a92adaf91c3205274be5d87302908ad35cb2dbf
Adrian Szyndela [Thu, 1 Jun 2017 11:41:09 +0000 (13:41 +0200)]
kdbus: make error messages more consistent with glib
Change-Id: Iccb88417fc83290ed70a32e86bc80c4a5c44c6ee
Adrian Szyndela [Tue, 9 May 2017 12:45:26 +0000 (14:45 +0200)]
kdbus: bugfix for return codes on call to non-existing destination
Calling a method on non-existing destination with auto start disabled
caused error org.freedesktop.DBus.ServiceUnknown.
Non-kdbus implementation gives org.freedesktop.DBus.NameHasNoOwner.
This patch makes kdbus return the same error as non-kdbus implementation.
Change-Id: Iadbf060159e91d5c4e10ea9a26abb50d02154a5f
INSUN PYO [Wed, 29 Mar 2017 16:04:52 +0000 (01:04 +0900)]
change user socket address from %t/dbus/user_bus_socket to %t/bus
Sync to upstream.
Upstream uses %d/bus, but tizen uses %t/dbus/user_bus_socket.
Some application use hard coded address %d/dbus/user_bus_socket.
So, I remain SymLinks temporary.
Later changing all application, I will remove SymLinks.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I7d00fc96fdbae28a6bd610ced0adfdcc52ec8987
INSUN PYO [Sat, 11 Mar 2017 14:56:27 +0000 (23:56 +0900)]
remove unused code
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I1192579e02963b0246bb99d54f415cab673abbe6
sanghyeok.oh [Fri, 14 Apr 2017 06:26:46 +0000 (15:26 +0900)]
license:rearrange license from each packages
Change-Id: I6ec1d0ff94b9e237bb158d5c49031d0463ec3bc2
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
sanghyeok.oh [Wed, 12 Apr 2017 11:24:27 +0000 (20:24 +0900)]
svace:fix leak
Change-Id: Ifce97dae8b9390d492708bfd7d7467c1af1ed132
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Adrian Szyndela [Fri, 7 Apr 2017 16:35:58 +0000 (18:35 +0200)]
GVariant: don't remove padding from no field
When a message is created with dbus_message_new() it gets no fields.
Then, while appending some values, but before setting path, interface,
etc. a single field appears - signature. If there would be more fields,
then after deleting the last field we would need to remove padding
from second-to-last field. In case of a single field, we have
no second-to-last field, only empty dictionary. Empty dictionaries do
not have padding at the end, so we do not need to remove it.
Change-Id: Idb15f3999a3ff0117f906ec632f462eb5cdaab09
INSUN PYO [Mon, 3 Apr 2017 03:36:45 +0000 (12:36 +0900)]
Increase release version for MCD.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I8f115a821731b01727bc19510962f9b7720f3d89
INSUN PYO [Fri, 31 Mar 2017 03:17:31 +0000 (20:17 -0700)]
Merge "License: add license files - add COPYING" into tizen
INSUN PYO [Fri, 31 Mar 2017 02:58:09 +0000 (11:58 +0900)]
License: add license files - add COPYING
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Ib7372970eb1a1e24154de189848cd0ca2faa2563
INSUN PYO [Mon, 20 Mar 2017 16:36:32 +0000 (01:36 +0900)]
Replace user dbus.socket and dbus.service file.
The dbus source includes dbus.socket and dbus.service for user session.
But tizen uses its own files.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Ied720b0fd5e792845ab626f680d01cbb3e9b2db9
INSUN PYO [Mon, 27 Mar 2017 13:17:42 +0000 (22:17 +0900)]
License: add license files
Change-Id: I2e1785ef0d14c2d5a9aa1d49904064b4fe80f9ee
INSUN PYO [Thu, 23 Mar 2017 07:49:21 +0000 (16:49 +0900)]
cynara: add exception code for disconnection code.
For disconnection connections, _dbus_connection_get_dispatch_status_unlocked always returns DBUS_DISPATCH_DATA_REMAINS.
So __bus_loop_dispatch is going around in an infinite loop.
If _dbus_loop_dispatch is out of the loop, _dbus_connection_get_dispatch_status_unlocked returns DBUS_DISPATCH_COMPLETE as normal.
It is dead-lock.
====================== reproduce ==================
su - owner
dbus-send --type=method_call --system --dest=org.tizen.system.deviced /Org/Tizen/System/DeviceD/Display org.tizen.system.deviced.display.CustomBrightness
dbus-send --type=method_call --system --dest=org.tizen.system.deviced /Org/Tizen/System/DeviceD/Display org.tizen.system.deviced.display.CustomBrightness
dbus-send --type=method_call --system --dest=org.tizen.system.deviced /Org/Tizen/System/DeviceD/Display org.tizen.system.deviced.display.CustomBrightness
dbus-send --type=method_call --system --dest=org.tizen.system.deviced /Org/Tizen/System/DeviceD/Display org.tizen.system.deviced.display.CustomBrightness
dbus-send --type=method_call --system --dest=org.tizen.system.deviced /Org/Tizen/System/DeviceD/Display org.tizen.system.deviced.display.CustomBrightness
===================================================
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I8565a757a62b50a27d2c6b6931e125af452ff74e
INSUN PYO [Sun, 5 Mar 2017 00:30:23 +0000 (09:30 +0900)]
Change bus connection check code : bus_connection_is_active () ==> dbus_connection_get_is_connected ()
Since dbus-daemon uses asynchronous cynara functions, some cynara functions are delayed.
In the cynara delay function, check if connection is valid or not.
If the connection is lost, the "connection->slot_list" is cleared. So bus_connection_is_active() will cause crash.
Since dbus_connection_get_is_connected() only checks for "connection->transport->disconnected" variables, it works fine if only the reference count is valid.
====================================================================================================================================
0 bus_connection_is_active (connection=<optimized out>) at connection.c:1291
1 0xb6f08134 in bus_deferred_message_dispatch (deferred_message=0xb7c93668) at check.c:498
2 0xb6f0c7ce in bus_connection_dispatch_deferred (connection=0xb7ca0008) at connection.c:2677
3 0xb6f0826a in bus_check_queued_message_reply_callback (deferred_message=0xb7cb5870, result=<optimized out>) at check.c:188
4 0xb6f0cdda in bus_cynara_check_response_callback (check_id=<optimized out>, cause=<optimized out>, response=<optimized out>, user_response_data=0xb7cb5870) at cynara.c:333
5 0xb6db2994 in Cynara::Logic::processCheckResponse (this=this@entry=0xb7c0cdb8, checkResponse=...) at /usr/src/debug/cynara-0.14.7/src/client-async/logic/Logic.cpp:277
6 0xb6db3a8e in Cynara::Logic::processResponses (this=this@entry=0xb7c0cdb8) at /usr/src/debug/cynara-0.14.7/src/client-async/logic/Logic.cpp:324
7 0xb6db4578 in Cynara::Logic::processIn (this=this@entry=0xb7c0cdb8) at /usr/src/debug/cynara-0.14.7/src/client-async/logic/Logic.cpp:348
8 0xb6db45d8 in Cynara::Logic::process (this=0xb7c0cdb8) at /usr/src/debug/cynara-0.14.7/src/client-async/logic/Logic.cpp:175
9 0xb6daf35a in operator() (this=<optimized out>) at /usr/lib/gcc/armv7l-tizen-linux-gnueabi/4.9.2/include/c++/functional:2439
10 Cynara::tryCatch(std::function<int ()> const&) (func=...) at /usr/src/debug/cynara-0.14.7/src/common/exceptions/TryCatch.h:42
11 0xb6dafa5e in cynara_async_process (p_cynara=0xb7c0c7a0) at /usr/src/debug/cynara-0.14.7/src/client-async/api/client-async-api.cpp:236
12 0xb6f0cdb4 in bus_cynara_watch_callback (watch=<optimized out>, flags=<optimized out>, data=<optimized out>) at cynara.c:288
13 0xb6f14f5a in _dbus_loop_iterate (loop=loop@entry=0xb7c0b618, block=block@entry=1) at dbus-mainloop.c:819
14 0xb6f1506c in _dbus_loop_run (loop=0xb7c0b618) at dbus-mainloop.c:883
15 0xb6f046f6 in main (argc=<optimized out>, argv=<optimized out>) at main.c:661
(gdb) info locals
d = 0x0
(gdb) f 1
(gdb) p *deferred_message->sender
$27 = {refcount = {value = 1}, mutex = 0xb7c41c70, dispatch_mutex = 0xb7c41cb0, dispatch_cond = 0xb7c41cd0, io_path_mutex = 0xb7c41c90, io_path_cond = 0xb7c41d08, outgoing_messages = 0x0, incoming_messages = 0x0, expired_messages = 0x0, message_borrowed = 0x0,
n_outgoing = 0, n_incoming = 0, outgoing_counter = 0xb7c41df8, transport = 0xb7c41778, watches = 0xb7c41b20, timeouts = 0xb7c41b40, filter_list = 0x0, slot_mutex = 0xb7c41d40, slot_list = {slots = 0xb7c41fc0, n_slots = 1}, pending_replies = 0xb7c41b60,
client_serial = 10, disconnect_message_link = 0x0, wakeup_main_function = 0x0, wakeup_main_data = 0x0, free_wakeup_main_data = 0x0, dispatch_status_function = 0x0, dispatch_status_data = 0x0, free_dispatch_status_data = 0x0, last_dispatch_status = DBUS_DISPATCH_COMPLETE,
objects = 0xb7c41f20, server_guid = 0x0, peer_smack_label = 0xb7c41c60 "System", dispatch_acquired = 0, io_path_acquired = 0, dispatch_disabled = 0, shareable = 0, exit_on_disconnect = 0, route_peer_messages = 1, disconnected_message_arrived = 1,
disconnected_message_processed = 1, have_connection_lock = 0, generation = 1}
(gdb) p deferred_message->sender->slot_list
$28 = {slots = 0xb7c41fc0, n_slots = 1}
(gdb) p *deferred_message->sender->slot_list->slots
$31 = {data = 0x0, free_data_func = 0x0}
(gdb) p connection_data_slot
$32 = 0
(gdb) p deferred_message->sender->slot_list->slots[0]
$35 = {data = 0x0, free_data_func = 0x0}
(gdb) p deferred_message->sender->transport->disconnected
$33 = 1
====================================================================================================================================
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Iab2abfb63030d15e380e11c057ee1f99124e102e
(cherry picked from commit
62fd017b7dbf577f736de7f8d44e6a94b4334472)
INSUN PYO [Wed, 15 Mar 2017 02:12:18 +0000 (11:12 +0900)]
kdbus: fix build error for emulator64-wayland.
After applying unify build, all profiles build kdbus unconditionally even if you do not use kdbus.
Before unify build, kdbus was not enabled in emulator64-wayland environment.
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I22f8dd17cdfeed17a3ef7aa8da5c738fc677f0ad
Geunsik Lim [Tue, 7 Mar 2017 08:53:38 +0000 (17:53 +0900)]
[4.0] Remove OBS Project Dependency (kdbus/TV)
- Please add "dbus-libs-extension-kdbus" for TV/arm-wayland images via JIRA-TRE issue.
- Without this commit, this package won't be built correctly in Tizen 4.0.
Change-Id: I94d98f44cad637af2acc22898c74a806e74f73bb
Signed-off-by: Geunsik Lim <geunsik.lim@samsung.com>
Reviewed-by: Insun Pyo <insun.pyo@samsung.com>
INSUN PYO [Mon, 6 Mar 2017 07:27:32 +0000 (16:27 +0900)]
Remove unused code (Don't use pgrep)
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: I8809374f28baae65606b679b3ef15fc71a91f98d
INSUN PYO [Mon, 6 Mar 2017 01:46:23 +0000 (10:46 +0900)]
change socket.target.wants/dbus.service ==> socket.target.wants/dbus.socket
Signed-off-by: INSUN PYO <insun.pyo@samsung.com>
Change-Id: Iad33b6770458d2e3e4e8507cfae5b1489da5bf0c
(cherry picked from commit
e1a4cf929e2095ae0bd1e503a59388791e18e1cb)
sanghyeok.oh [Tue, 28 Feb 2017 09:00:42 +0000 (18:00 +0900)]
patch: bus: Fix timeout restarts
patch from upstream
https://cgit.freedesktop.org/dbus/dbus/commit/?id=
529600397bcab47b9bed5da9208c2df05c8b86b4
author Michal Koutn? <mkoutny@suse.com> 2016-05-24 09:14:11 (GMT)
committer Simon McVittie <smcv@debian.org> 2017-02-01 10:42:50 (GMT)
commit
529600397bcab47b9bed5da9208c2df05c8b86b4 (patch)
tree
17e9af27f973a4e0132ee23e405ed7d52fdb0771
parent
befd186c0b005f1e1130cb3e4923c689085a8dbc (diff)
bus: Fix timeout restarts
The code counting pending fds relied on restart of timeouts when they are
enabled. This patch adds function that ensures that such enabled timeouts
have their timekeeping data reset (and not only when timeout is
registered into event loop processing).
When timeouts weren't reset, they'd fire at rather random and mainly
incorrect moments leading to interruption of connections of dbus-daemon.
Every time we reset the interval, we also need to re-enable the timeout
and mark its end time to be recalculated by the event loop, so combine
the old set_enabled(TRUE) with set_interval() as a new restart() method.
This leaves all the set_enabled() calls having a FALSE parameter, so
remove the parameter and rename the method to disable().
[smcv: fix minor coding style issues]
[smcv: replace set_reenabled()/set_interval() pair with restart()]
[smcv: replace set_enabled(FALSE) with disable()]
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=95619
Change-Id: Ia80641554800ae09ae22d4615b87eda0697c5f3e
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Hyotaek Shim [Wed, 1 Mar 2017 05:44:33 +0000 (21:44 -0800)]
Merge changes Icf330ba3,I7354c21e into tizen
* changes:
tests: general fix for tests setup
tests: generating coverage version of the library
INSUN PYO [Wed, 22 Feb 2017 02:38:43 +0000 (11:38 +0900)]
make link : /usr/lib/systemd/user/dbus.socket ==> /usr/lib/systemd/user/sockets.target.wants/dbus.socket
Until now, dbus.socket of user session was not running in the sockets.target.
dbus.socket has been triggered by dbus.service's dependencies.
The dbus.service is included in basic.target.
Therefore, there are some dependency issues.
If a service runs before dbus.service, the service can not accesses dbus system.
(Since dbus.socket is not yet activated, so you can not make a dbus connection.)
I have modified dbus.service to be explicitly run in sockets.target.
On the session bus, sockets.target is guaranteed to run before all services.
Signed-off-by: INSUN PYO <insun.pyo@ubuntu1604.playinsun.pe.kr>
Change-Id: Ic46f1a46b458e2017b0025c6e564ac1204eb8cc8
(cherry picked from commit
1beb067e73478b3948980eb19e379a8986dc365a)
Hyotaek Shim [Wed, 22 Feb 2017 00:04:31 +0000 (16:04 -0800)]
Merge "tizen: notify service readyness after successful config file parsing" into tizen
Karol Lewandowski [Fri, 10 Feb 2017 15:16:55 +0000 (16:16 +0100)]
tizen: notify service readyness after successful config file parsing
This commit works around systemd -> dbus -> (nss) -> systemd deadlock
occuring in Tizen.
Situation is depicted on diagram below:
systemd dbus nss
| . .
| . .
1> reads config . .
| . .
2> reg. nofification . .
| for dbus.service . .
| . .
| . .
3> exec. services . .
| | . .
| (start srv...) . .
| | . .
| 4>------------(exec) .
| | .
*-<---(notification)=<| .
\ | .
.\> call dbus.serv | .
. handler | .
. | 4a> read config .
. |> conn. to bus | .
. | 4b> resolve user .
. |> call dbus . and groups .
. \ . initgroups() .
. ------------>. | .
. (blocking) . \---------->select module
. . (blocking) based on conf.
. . |
(below not executed) . |> security-manager
. . | module selected
. . |
. . |
[sec-manager.socket]<---------------------<| call security-manager
. . (DEADLOCK) . on unix socket
. . .
Solid line (|) means active thread of execution, dot (.) means
code not executed (blocked).
The root cause of the problem is that systemd calls external
API in blocking manner, which, due to system configuration, requires
systemd to handle its request for successful completion.
(systemd should start security-manager service via sec-manager.socket
activation).
The exact place this situation is triggered is in dbus' initialization.
This commits works around this problem by delaying dbus.service readyness
notification - effectively - resolving the problem.
Situation becomes:
systemd dbus nss security-manager
| . . .
| . . .
1> reads config . . .
| . . .
2> reg. nofification . . .
| for dbus.service . . .
| . . .
| . . .
3> exec. services . . .
| | . . .
| (start srv...) . . .
| | . . .
| 4>------------(exec) . .
| | . .
| | . .
| | . .
| | . .
| | . .
| 4a> read config . .
| | . .
| 4b> resolve user . .
| . and groups . .
| . initgroups() . .
| . | . .
| . \---------->select module .
| . (blocking) based on conf. .
| . | .
| . |> security-manager .
| . | module selected .
| . | .
| . | .
[sec-manager.socket]<---------------------<| call security-manager .
. | . . on unix socket .
. | . .
. |> exec service ----------------------------------------------->(exec)
. | . . |
<--| . |<-------------------------<| handle call
| . | |
| |<-----------------| |
| | . |
| | . |
.\> call dbus.serv |>finish conf. file. |
. handler | parsing . |
. | | . |
*-<---(notification)-<|> explicitly . |
.\ | | notify startup . |
. \> call dbus.serv | finished . |
| handler | . |
. | | . |
. |> conn. to bus | . |
. | | . |
. |> call dbus | . |
. \ | . |
. ------------>* dispatch loop . |
. (blocking) | . |
. | . |
|<--------------------< handle call . |
| | . |
Change-Id: Ifeaf299fc8653b583cd06ca9fd4f9f8045a2bde0
INSUN PYO [Fri, 17 Feb 2017 02:19:31 +0000 (11:19 +0900)]
Change the method to get DBUS addres when changing user with "su -" command.
The adbd's privilege has been changed form "System::Privileged" to "User::Shell".
So, pgrep does not work with smack errors when accessing /proc.
If it fails to find DBUS address, the fixed address is used.
Signed-off-by: INSUN PYO <insun.pyo@ubuntu1604.playinsun.pe.kr>
Change-Id: I4f9e0bc0939de2ecd7b866df8a2f266b4258338b
(cherry picked from commit
4f11f164f44d78ec6a54892e5d4095ca52e7117d)
Adrian Szyndela [Thu, 16 Feb 2017 11:47:41 +0000 (12:47 +0100)]
tests: general fix for tests setup
Before this patch tests were installed into custom directory
/usr/lib/dbus-tests/test-suites/dbus-tests
Originally, dbus wants its installed tests in
/usr/lib/dbus/installed-tests/dbus
This path seems to be partially hardcoded in the configuration
files - it relies on defined datadir. As datadir is used
also for something else than tests, we don't want to change it.
This commit switches installation directory to the original,
and adds symbolic links in our custom directory.
Change-Id: Icf330ba3b5be59ee941d51b15a8f78a082297846
Adrian Szyndela [Mon, 6 Feb 2017 11:17:24 +0000 (12:17 +0100)]
tests: generating coverage version of the library
Additionally:
- immediate output of test results instead of caching and printing
at the end;
- removed manual-authz test, because it is not meant to be run
automatically.
Change-Id: I7354c21e0fa457ad7e2a4a53e1c898e5039b7f73
Konrad Lipinski [Fri, 3 Feb 2017 12:40:29 +0000 (13:40 +0100)]
gbs incremental build support
Change-Id: Ic1960a3b70df234a3e9ad141d6c335d373c331ec
Adrian Szyndela [Wed, 11 Jan 2017 11:54:07 +0000 (12:54 +0100)]
Merge tizen_3.0 into tizen
This merge commit adds integration tests.
Change-Id: I2b0ecb3c48fec72a6be1a7d6406b5580bbb38ec8
Hyotaek Shim [Wed, 28 Dec 2016 01:53:31 +0000 (10:53 +0900)]
Modify from APACHE-2.0 to Apache-2.0
Change-Id: I2e509fe1a4f275c9f5b35290bdd2d218c4b3692a
(cherry picked from commit
4f9a3114261606c3edf84655b2e5752dee33585f)
Hyotaek Shim [Wed, 28 Dec 2016 01:53:31 +0000 (10:53 +0900)]
Modify from APACHE-2.0 to Apache-2.0
Change-Id: I2e509fe1a4f275c9f5b35290bdd2d218c4b3692a
Hyotaek Shim [Wed, 28 Dec 2016 01:43:22 +0000 (10:43 +0900)]
Add .spec description about LICENSE.APACHE2.0
Change-Id: I127f90398b4fe33b3c3b412ef84a42ab2683efd5
(cherry picked from commit
44c73590bab90e63b21fadfffadbfa1d745d9613)
Hyotaek Shim [Wed, 28 Dec 2016 01:43:22 +0000 (10:43 +0900)]
Add .spec description about LICENSE.APACHE2.0
Change-Id: I127f90398b4fe33b3c3b412ef84a42ab2683efd5
Hyotaek Shim [Wed, 28 Dec 2016 01:39:34 +0000 (10:39 +0900)]
Add LICENSE.APACHE2.0
Change-Id: Icf4a2b514abfeb458305c58ab99aad4cfbab325d
(cherry picked from commit
4b5af3aba73865b45a144694fa9c918ed7f9233e)
Hyotaek Shim [Wed, 28 Dec 2016 01:39:34 +0000 (10:39 +0900)]
Add LICENSE.APACHE2.0
Change-Id: Icf4a2b514abfeb458305c58ab99aad4cfbab325d
Krystian Kisielak [Thu, 13 Oct 2016 13:01:27 +0000 (15:01 +0200)]
Adds package for dbus-integration tests.
Change-Id: I0175a6c710e418be7f3b06f1df6ad13396e40f8c
Signed-off-by: Krystian Kisielak <k.kisielak@samsung.com>
Adrian Szyndela [Fri, 14 Oct 2016 13:24:43 +0000 (15:24 +0200)]
Bugfix, GVariant: missing padding for fixed-size tuples.
There is a requirement for fixed-size tuples:
size of such tuple must be a multiply of its required alignment.
That part was overlooked in the implementation.
This commit adds padding accounting for both reading and writing.
Change-Id: I0825be6436b93b836dc333f5a395fdf1021f2a06
Adrian Szyndela [Thu, 13 Oct 2016 13:04:10 +0000 (15:04 +0200)]
Bugfix, GVariant: no padding in root-level tuples
Change-Id: If84df57e9985f484e3c3c63854df5ff666cc1794
sanghyeok.oh [Mon, 26 Sep 2016 06:24:35 +0000 (15:24 +0900)]
change socket in/out label
Change-Id: Ifcc7dacb077615c8f5392a6b213d402d4775af05
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
Adrian Szyndela [Tue, 13 Sep 2016 07:59:34 +0000 (09:59 +0200)]
bugfix: sec_label was copied without ending byte
Change-Id: I3d9aa7f2d608e6706600a87ad5dc570033244f98
Lukasz Skalski [Mon, 5 Sep 2016 09:35:38 +0000 (11:35 +0200)]
kdbus: fix for GetConnectionSELinuxSecurityContext method call
To keep compatibility with dbus1, GetConnectionSELinuxSecurityContext
method call should always return SELinuxSecurityContextUnknown error
on systems without SELinux (even if other LSM systems, like for example
SMACK, are available on platform).
Change-Id: I347cdf665a03fc047d267752b23c4503bf5218e3
Hyotaek Shim [Thu, 1 Sep 2016 03:38:20 +0000 (20:38 -0700)]
Merge "calculate kdbus timeout based on user timeout" into tizen
Konrad Lipinski [Thu, 25 Aug 2016 09:07:26 +0000 (11:07 +0200)]
calculate kdbus timeout based on user timeout
Change-Id: I9fd1720e5c2a1edaa1fcd66b78df0b0238e9ce9f
sanghyeok.oh [Wed, 31 Aug 2016 08:35:40 +0000 (17:35 +0900)]
Bugfix:abnormal using of _dbus_validate_bus_name
Change-Id: I1eb91da587469381f69ada2939a980d8ea1f40e1
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>
sanghyeok.oh [Mon, 22 Aug 2016 07:15:49 +0000 (16:15 +0900)]
support NameAcquired & NameLost signal subscription
Change-Id: I6f7c2e39c8872ebae8a68157515ef247325b3cc4
Signed-off-by: sanghyeok.oh <sanghyeok.oh@samsung.com>