We must set these compiler flags independent as to whether this
is a release build or a debug build.
In most cases, we don't differentiate between release and debug build
anyway. Granted, we have "-D more_asserts=100" and set "-O" CFLAGS,
but that is more granular and not a simple "buildtype".
In particular, these compiler flags apply to all kinds of builds.
This is important, because otherwise we get build failures, because
also in release build we want to build with `-Werror` and `-Wall`.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/692
(cherry picked from commit c0c6470e4d)
(cherry picked from commit 6b316a1991)
(cherry picked from commit 805c9fa8d4)
We also do that for the autotools implementation.
(cherry picked from commit fd57e9665c)
(cherry picked from commit c807e77271)
(cherry picked from commit 63f2d73b97)
With glib2-2.67.0-1.fc34.x86_64.rpm, clang-11.0.0-2.fc34.x86_64.rpm, the
generated code emits a compiler warning:
libnm-core/tests/nm-core-tests-enum-types.c:17:7: error: passing 'typeof (*(&g_define_type_id__volatile)) *' (aka 'volatile unsigned long *') to parameter of type 'gsize *' (aka 'unsigned long *') discards qualifiers [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
if (g_once_init_enter (&g_define_type_id__volatile))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gthread.h:260:7: note: expanded from macro 'g_once_init_enter'
(!g_atomic_pointer_get (location) && \
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gatomic.h:112:38: note: expanded from macro 'g_atomic_pointer_get'
__atomic_load (gapg_temp_atomic, &gapg_temp_newval, __ATOMIC_SEQ_CST); \
^~~~~~~~~~~~~~~~~
libnm-core/tests/nm-core-tests-enum-types.c:40:7: error: passing 'typeof (*(&g_define_type_id__volatile)) *' (aka 'volatile unsigned long *') to parameter of type 'gsize *' (aka 'unsigned long *') discards qualifiers [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
if (g_once_init_enter (&g_define_type_id__volatile))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gthread.h:260:7: note: expanded from macro 'g_once_init_enter'
(!g_atomic_pointer_get (location) && \
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gatomic.h:112:38: note: expanded from macro 'g_atomic_pointer_get'
__atomic_load (gapg_temp_atomic, &gapg_temp_newval, __ATOMIC_SEQ_CST); \
^~~~~~~~~~~~~~~~~
libnm-core/tests/nm-core-tests-enum-types.c:63:7: error: passing 'typeof (*(&g_define_type_id__volatile)) *' (aka 'volatile unsigned long *') to parameter of type 'gsize *' (aka 'unsigned long *') discards qualifiers [-Werror,-Wincompatible-pointer-types-discards-qualifiers]
if (g_once_init_enter (&g_define_type_id__volatile))
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gthread.h:260:7: note: expanded from macro 'g_once_init_enter'
(!g_atomic_pointer_get (location) && \
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/usr/include/glib-2.0/glib/gatomic.h:112:38: note: expanded from macro 'g_atomic_pointer_get'
__atomic_load (gapg_temp_atomic, &gapg_temp_newval, __ATOMIC_SEQ_CST); \
^~~~~~~~~~~~~~~~~
We could pass "-Wincompatible-pointer-types-discards-qualifiers" as CFLAGS
when building this file. However, we have a workaround in our "nm-glib-aux/nm-glib.h",
so we can instead include "nm-default.h". At first glance, that might look like
the less preferable solution. However, this file is only there for unit tests,
and we also include "nm-default.h" for other sources that are generated with
"glib-mkenums". So, doing it also for our tests becomes the preferable solution.
(cherry picked from commit 755d97d38c)
(cherry picked from commit 294efba18f)
(cherry picked from commit 2cff3f369b)
When including <glib.h>, it will always define G_LOG_DOMAIN if it
is not yet defined.
Usually we want to include "nm-default.h" as very first header. In that
case, <glib.h> is not yet included. Then the previous check #error works
well.
However, if we include "nm-default.h" in sources generated by
glib-mkenums, then the generator first already includes <glib.h>,
and thus defines G_LOG_DOMAIN. It does so for "libnm-core/nm-core-enum-types.c"
and "libnm/nm-enum-types.c", where the #error would not trigger.
But we will also include "nm-default.h" for "libnm-core/tests/nm-core-tests-enum-types.c".
That will start triggering this #error.
While in general we want to include "nm-default.h" first, we also need
to support cases where <glib.h> gets included first. Thus this error is
not useful. Remove it.
(cherry picked from commit 42fa8f3d27)
(cherry picked from commit a1f3cebbec)
(cherry picked from commit 61d78ed333)
It's not strictly necessary, because contrary to g_atomic_pointer_get()
and g_atomic_pointer_compare_and_exchange(), glib's variant for the
setter is mostly fine.
Still, reimplement it, because we use typeof() eagerly and can thus add
more static checks than glib.
(cherry picked from commit 7c60e984b6)
(cherry picked from commit 6ded463f36)
(cherry picked from commit 976b358be6)
The dispatcher code does not use the generated introspection sources
(anymore). Don't add a dependency.
(cherry picked from commit e0a3a5e2f8)
(cherry picked from commit 6ba600cb38)
(cherry picked from commit b8a172a622)
It seems it can happen that the service is not yet unregistered from the
D-Bus broker, even if we already reaped the PID.
/builds/NetworkManager/NetworkManager/tools/run-nm-test.sh --called-from-make /builds/NetworkManager/NetworkManager/build --launch-dbus=auto /builds/NetworkManager/NetworkManager/build/libnm/tests/test-nm-client
--- stdout ---
/libnm/device-added:
nmtst: initialize nmtst_get_rand() with NMTST_SEED_RAND=0
--- stderr ---
**
test:ERROR:../shared/nm-test-utils-impl.c:216:nmtstc_service_cleanup: assertion failed: (!name_exists(info->bus, "org.freedesktop.NetworkManager"))
Workaround by waiting a bit.
We now iterate the main GMainContext, unlike before. But that
should not cause any problems for the test.
(cherry picked from commit 1b8ccacc5d)
(cherry picked from commit d10d14d7ba)
(cherry picked from commit d34e6193da)
nmtst_main_context_iterate_until() is a macro, and we don't want to restrict the
valid integer type (or range) of the "timeout_msec" argument.
In particular, if the user calculates a timeout with "timestamp_msec -
now_msec", the resulting "timeout_msec" might be a negative gint64.
We should handle that gracefully, and not let it be cast to a huge
unsigned int.
(cherry picked from commit 6cb6888404)
(cherry picked from commit 4d572bea7e)
(cherry picked from commit 95b74a3bde)
gcc-11.0.0-0.7.fc34 warns here:
CC libnm-core/libnm_core_la-nm-setting-team.lo
libnm-core/nm-setting-team.c: In function ‘nm_team_link_watcher_new_ethtool’:
libnm-core/nm-setting-team.c:127:33: error: array subscript ‘NMTeamLinkWatcher[0]’ is partly outside array bounds of ‘unsigned char[16]’ [-Werror=array-bounds]
127 | watcher->ref_count = 1;
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~
libnm-core/nm-setting-team.c:125:15: note: referencing an object of size 16 allocated by ‘g_malloc’
125 | watcher = g_malloc(nm_offsetofend(NMTeamLinkWatcher, ethtool));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
libnm-core/nm-setting-team.c:128:33: error: array subscript ‘NMTeamLinkWatcher[0]’ is partly outside array bounds of ‘unsigned char[16]’ [-Werror=array-bounds]
128 | watcher->type = LINK_WATCHER_ETHTOOL;
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
libnm-core/nm-setting-team.c:125:15: note: referencing an object of size 16 allocated by ‘g_malloc’
125 | watcher = g_malloc(nm_offsetofend(NMTeamLinkWatcher, ethtool));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
libnm-core/nm-setting-team.c:129:33: error: array subscript ‘NMTeamLinkWatcher[0]’ is partly outside array bounds of ‘unsigned char[16]’ [-Werror=array-bounds]
129 | watcher->ethtool.delay_up = delay_up;
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~
libnm-core/nm-setting-team.c:125:15: note: referencing an object of size 16 allocated by ‘g_malloc’
125 | watcher = g_malloc(nm_offsetofend(NMTeamLinkWatcher, ethtool));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
libnm-core/nm-setting-team.c:130:33: error: array subscript ‘NMTeamLinkWatcher[0]’ is partly outside array bounds of ‘unsigned char[16]’ [-Werror=array-bounds]
130 | watcher->ethtool.delay_down = delay_down;
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~
libnm-core/nm-setting-team.c:125:15: note: referencing an object of size 16 allocated by ‘g_malloc’
125 | watcher = g_malloc(nm_offsetofend(NMTeamLinkWatcher, ethtool));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Maybe we should not use this trick and just malloc() a struct of the
intended size, however:
- the code below does a similar thing, doing it differently for ethtool
watcher is confusing.
- the NMTeamLinkWatcher is a union which cannot alter its type. In no
case is it correct to access the fields of the wrong union type. By
allocating a smaller chunk, valgrind might catch such bugs.
Also, NMTeamLinkWatcher's definition is private to the C source file,
in no case must anybody assume that the rest of the buffer actually
exists.
Hence, workaround the warning by suppressing it.
(cherry picked from commit e5699dbcb7)
(cherry picked from commit 221547bc21)
(cherry picked from commit 8f3cf4f3e8)
We cannot drop the reference count to zero while having
no lock. Otherwise, another thread might race doing
s = nm_ref_string_new("...");
nm_ref_string_unref(s);
and already successfully delete the instance.
Hitting this race should be rather difficult, especially because
we tend to use NMRefString only from one thread. But still, access
to global variables must be race free.
Fixes: 908fadec96 ('shared: add NMRefString')
(cherry picked from commit 3490a09a7d)
(cherry picked from commit 1a68a54471)
(cherry picked from commit 71f56a5d76)
GObject's dispose() functions may be called multiple times
to break reference cycles.
As dispose() calls _init_release_all(), the object might
already be partially destroyed.
Fixes: ce0e898fb4 ('libnm: refactor caching of D-Bus objects in NMClient')
(cherry picked from commit ef6edd8dbd)
(cherry picked from commit 9659db281f)
(cherry picked from commit a744397ad5)
wpa_supplicant has a property "scanning" and a "state=scanning".
Previously, NetworkManager considered both parts to indicate whether
supplicant is currently scanning (if either the property or the state
indicated scanning, it took that as indication for scanning).
If NetworkManager thinks that supplicant is scanning, it suppresses
explicit "Scan" requests. That alone is not severe, because the "Scan"
request is only to trigger a scan in supplicant (which supplicant
possibly is already doing in state "scanning").
However, what is severe is that NetworkManager will also block autoconnect
while supplicant is scanning. That is because NetworkManager wants to get
a complete scan result before deciding which network to connect to.
It seems that wpa_supplicant can get into "state=scanning" and stay
there indefinitely. This prevents NetworkManager from autoactivating
a profile.
Fix that, to only honor the "scanning" property.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/597
Fixes: b83f07916a ('supplicant: large rework of wpa_supplicant handling')
(cherry picked from commit 8cadfed2fe)
(cherry picked from commit fbe6240a7d)
(cherry picked from commit 2bec392e56)
Also, never update the value to %NULL. If the current
message does not contain a UUID, keep the previous one.
Fixes: 830a5a14cb ('device: add support for OpenVSwitch devices')
(cherry picked from commit 609b08e2eb)
(cherry picked from commit 0464c36a02)
(cherry picked from commit e1424f264d)
In connection_removed we use the id.name that was being g_freed a few
lines further down.
Fixes: bea6c40367 ('wifi/iwd: handle forgetting connection profiles')
(cherry picked from commit c1ff06e119)
(cherry picked from commit 03b63a893f)
(cherry picked from commit 09c4fa5a6c)
For hidden networks, we usually don't have an SSID. We try to match
and fill the SSID based on the profiles that we have:
<debug> [1603798852.9918] device[6b383dca267b6878] (wlp2s0): matched hidden AP AA:BB:CC:DD:EE:FF => "SSID"
However, we should not clear that value again on the next update:
<trace> [1603798856.5724] sup-iface[66c1a0883a262394,0,wlp2s0]: BSS /fi/w1/wpa_supplicant1/Interfaces/0/BSSs/3 updated
<debug> [1603798856.5726] device[6b383dca267b6878] (wlp2s0): wifi-ap: updated AA:BB:CC:DD:EE:FF (none)
Once we have a SSID, we can only update it to a better value,
but not clear it.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/438
Fixes: b83f07916a ('supplicant: large rework of wpa_supplicant handling')
(cherry picked from commit eb36380335)
(cherry picked from commit 40edd49025)
(cherry picked from commit 42d28ae2e8)
With "connection.multi-connect", a profile can be activated multiple
times on a device with `nmcli connection show`. Also, a profile may be
in the process of deactivating on one device, while activating on
another one. So, in general it's possible that `nmcli connection show`
lists the same profile on multiple lines (reflecting their multiple
activation states).
If the user requests no fields that are part of the activation state,
then the active connections are ignored. For example with `nmcli
-f UUID,NAME connection show`. In that case, each profile is listed only
once.
On the other hand, with `nmcli -g UUID,NAME,DEVICE connection show` the
user again requested also to see the activation state, and a profile can
appear multiple times.
To handle that, we need to consider which fields were requested.
There was a bug where the "ACTIVE" field was not treated as part of the
activation state. That results in `nmcli -f UUID,NAME,ACTIVE connection
show` always returning "no". Fix that.
Fixes: a1b25a47b0 ('cli: rework printing of `nmcli connection` for multiple active connections')
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/547https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/642
(cherry picked from commit 4eb3b5b9dd)
(cherry picked from commit bb802507e4)
(cherry picked from commit 8bd6d2a63e)
Fixes: b83f07916a ('supplicant: large rework of wpa_supplicant handling')
(cherry picked from commit 3cf8620294)
(cherry picked from commit 4b11029df9)
(cherry picked from commit bd5da809ca)
In the previous form, NM_STR_BUF_INIT() was a macro. That makes sense,
however it's not really possible to make that a macro without evaluating
the reservation length multiple times. That means,
NMStrBuf strbuf = NM_STR_BUF_INIT (nmtst_get_rand_uint32 () % 100, FALSE);
leads to a crash. That is unfortunate, so instead make it an inline
function that returns a NMStrBut struct. Usually, we avoid functions
that returns structs, but here we do it.
(cherry picked from commit c6809df4cd)
(cherry picked from commit 3ed95f308f)
Previously, for simplicity, NMStrBuf did not support buffers without any
data allocated. However, supporting that has very little
overhead/complexity, so do it.
Now you can initialize buffers to have no data allocated, and when
appending data, it will automatically grow.
(cherry picked from commit 83c79bc7a8)
(cherry picked from commit 5216e5c012)
g_hash_table_iter_next() wants a (gpointer *), not an (int *).
Fixes: f70ee67058 ('dns: sd-resolved: reset interface configuration on deactivation')
(cherry picked from commit 526b484be1)
(cherry picked from commit dedcba61ef)
(cherry picked from commit b42696d70b)
On Fedora 33 when compiling NetworkManager it is failing with the
following error:
```
ERROR: files left in build directory after distclean:
./docs/libnm/libnm.actions
make[1]: *** [Makefile:18427: distcleancheck] Error 1
make[1]: Leaving directory '/builddir/nm-build/NetworkManager/NetworkManager-1.27.90/_build/sub'
make: *** [Makefile:18356: distcheck] Error 1
Error make distcheck
```
Adding the file to the DISTCLEANFILES will enforce the removal of this
file.
Signed-off-by: Fernando Fernandez Mancera <ffmancera@riseup.net>
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/659
(cherry picked from commit 9a935e516e)
(cherry picked from commit d28133ad0d)
(cherry picked from commit e1c1369eef)
This results in the args of 'nm_utils_user_data_unpack'
containing random data potentially also from the
previous stack-frame which is really really bad.
Signed-off-by: Antonio Cardace <acardace@redhat.com>
Fixes: b50702775f ('device: implement auth-request as async operation nm_manager_device_auth_request()')
(cherry picked from commit b6a18e0593)
(cherry picked from commit 50adaf7414)
(cherry picked from commit 72a2e34b42)
The gtk-doc text that the tool receives is not XML, it's a plain text.
When setting the plain text as XML attribute, we need to properly escape
it. The previous XML escape code was naive, and didn't cover for a
plain ampersand.
[thaller@redhat.com: adjusted patch during backport from nm-1-26 to nm-1-24]
(cherry picked from commit 1641cc1d03)
(cherry picked from commit 5b7d39f8e1)
There isn't any 'host-name' option for DHCPv6. Read instead the
'fqdn-fqdn' option that carries the FQDN assigned by the server to the
client.
(cherry picked from commit 1f74ea52f5)
(cherry picked from commit 4e1da002a9)
The dhclient backend already exports all the option passed by
dhclient, including the FDQN. Export it also for the systemd backend.
(cherry picked from commit 1621a6ddb1)
(cherry picked from commit c6a7618f2b)
Parse option 39 (Client Fully Qualified Domain Name, RFC 4704) from the DHCP
reply, which specifies the FQDN assigned by the server to the client.
c43eea9f2e
(cherry picked from commit 813fb7d64e)
nm_setting_ip_config_next_valid_dns_option() API was added in libnm 1.2, but
it was never exported in the ABI of libnm. It thus was unusable, and any user
trying to link against it would have been unable to do so.
Hide the API now entirely. It doesn't seem a very nice API. If we want to
allow the user to validate option names, we should expose such a function
to validate an option (not to fetch the next valid option from a
profile).
Fixes: 019943bb5d ('libnm-core: add dns-options property to NMSettingIPConfig')
(cherry picked from commit e8e5c12480)
(cherry picked from commit 04946f71ea)
Fixes: 5e71f01605 ('device: merge stage3 and stage4 ip-config function for IPv4 and IPv6')
(cherry picked from commit a017936223)
(cherry picked from commit 0adfcadc9d)
First I wanted to fix
test:ERROR:../src/ndisc/tests/test-ndisc-fake.c:373:test_preference_changed_cb: assertion failed (_a->timestamp == (data->timestamp1 + 3)): (9 == 10)
but that leads to a different failure:
test:ERROR:../src/ndisc/tests/test-ndisc-fake.c:375:test_preference_changed_cb: assertion failed (_a->lifetime == (9)): (10 == 9)
Instead, the start and end times must match exact (in their duration),
we only allow them to be shifted by up to one second.
Fixes: 8209095ee1 ('ndisc/tests: relax the assertion in "test-ndisc-fake.c"')
(cherry picked from commit b2f03544a7)
(cherry picked from commit 838777a891)
