- All internal source files (except "examples", which are not internal)
should include "config.h" first. As also all internal source
files should include "nm-default.h", let "config.h" be included
by "nm-default.h" and include "nm-default.h" as first in every
source file.
We already wanted to include "nm-default.h" before other headers
because it might contains some fixes (like "nm-glib.h" compatibility)
that is required first.
- After including "nm-default.h", we optinally allow for including the
corresponding header file for the source file at hand. The idea
is to ensure that each header file is self contained.
- Don't include "config.h" or "nm-default.h" in any header file
(except "nm-sd-adapt.h"). Public headers anyway must not include
these headers, and internal headers are never included after
"nm-default.h", as of the first previous point.
- Include all internal headers with quotes instead of angle brackets.
In practice it doesn't matter, because in our public headers we must
include other headers with angle brackets. As we use our public
headers also to compile our interal source files, effectively the
result must be the same. Still do it for consistency.
- Except for <config.h> itself. Include it with angle brackets as suggested by
https://www.gnu.org/software/autoconf/manual/autoconf.html#Configuration-Headers
nm_supplicant_manager_iface_get() would cache and reuse the supplicant
interface. But no ref-counting was in place so that the first user returning
the interface via nm_supplicant_manager_iface_release() would destroy the
instance for others.
This is broken for a very long time. Which shows that we hardly ever
have a cache-hit and usually create a new instance. So, instead of
letting nm_supplicant_manager_create_interface() check for existing
supplicant interface, always create a new instance. This also makes
sense, because we would expect that per ifname only one instance is
requested at a time. Also add an assertion that we don't return
multiple supplicant interface instances for the same ifname.
Drop nm_supplicant_manager_iface_release() in favor of requiring users
to unref the returned instance.
Also, use a GSList instead of a GHashTable for the cache.
Also, previously callers would pass @is_wireless to nm_supplicant_manager_iface_get(),
but the cache lookup did not consider that value. That doesn't matter
now as we always create a new instance.
https://bugzilla.redhat.com/show_bug.cgi?id=1298007
All implementations of NMDevice:setup_start() in derived classes
invoke the parent implementation first. Enforce that by moving
NMDevice:setup_start() to realize_start_setup() and only notify
derived classes afterwards via NMDevice:realize_start_notify().
In certain situations, ethernet links first appear with a zero MAC
address and then the MAC changes some time later. Currently NM does
not deal correctly with this scenario since it initializes wrong
@initial_hwaddr and @permanent_hwaddr on the device and tries to
immediately activate it.
To fix this, initialize the device's addresses only when the MAC
becomes valid and make the device available only at that point.
We inconsistently use gulong,guint,int types to store signal handler
id, but the type returned by g_signal_connect() is a gulong.
This has no practical consequences because a int/guint is enough to
store the value, however it is better to use a consistent type, also
because nm_clear_g_signal_handler() accepts a pointer to the signal id
and thus it must be always called with the same pointer type.
NMPlatformLink is a plain struct (not a GObject, for which we usually
don't use const). We certainly don't want the functions to modify the
passed-in data.
Ensure the platform link with the same interface name as the
NMDevice is actually compatible with it before using the link
for initialization of device properties. If not, remove the
NMDevice and create a new one since there are kernel resources
with a different type.
Unrealized devices aren't backed by kernel resources and so won't know
all of their attributes. That means three things:
1) they must update their attributes when they become realized
2) they must clear those attributes when unrealized
3) they must be looser in checking compatible connections until
they are realized
This requires that the setup() function be split into two parts, start & finish,
because finish must be run after add_device()
Also, we can simplify whether to pay attention to 'recheck-assume', which
is now dependent on priv->is_nm_owned, because the only case where NM should
*not* listen for the 'recheck-assume' signal is when the device is a
software device created by NM itself. That logic was previously spread
across the callers of add_device() but is now consolidated into
nm-manager.c::device_realized() and nm-device.c::nm_device_create_and_realize().
The nm_supplicant_config_add_*() functions used to log failures
themselves. As also the caller was logging the failure this resulted
in duplicate logging lines like:
<warn> MAC address randomization is not supported
<error> [1447867727.909185] [nm-device-wifi.c:2238] build_supplicant_config(): (wlp3s0): Couldn't add 802-11-wireless setting to supplicant config.
<error> [1447867727.909261] [nm-device-wifi.c:2472] act_stage2_config(): (wlp3s0): Activation: (wifi) couldn't build wireless configuration.
Instead, propagate the error reason back to the caller where there
is more context to log one single concise message.
Now you'd see only:
<error> [1447935996.859371] [nm-device-wifi.c:2475] act_stage2_config(): (wlp3s0): Activation: (wifi) couldn't build wireless configuration: 802-11-wireless: cannot enable mac-randomization due to missing supplicant support
Previously most objects were implicitly unexported when they were
destroyed, but since refcounts may make the object live longer than
intended, we should explicitly unexport them when they should no
longer be present on the bus.
This means we can assume that objects will always be un-exported
already when they are destroyed, *except* when quitting where most
objects will live until exit because NM leaves interfaces up and
running on quit.
Add a new 'ignore' option to NMSettingWired.wake-on-lan which disables
management of wake-on-lan by NetworkManager (i.e. the pre-existing
option will not be touched). Also, change the default behavior to be
'ignore' instead of 'disabled'.
https://bugzilla.gnome.org/show_bug.cgi?id=755182
NetworkManager set wpa_supplicant's fragment_size option to 1300. But if MTU
was lower, wpa_supplicant failed with "l2_packet_send - sendto: Message too
long" due to fragmentation of EAP-TLS or EAP-PEAP packets.
Actually, MTU has to be 14 bytes bigger than the "fragment_size" parameter.
Ideally, wpa_supplicant would take MTU in the account and adjust the
fragmentation limit accordingly. See discussion in
http://lists.shmoo.com/pipermail/hostap/2015-August/033546.htmlhttps://bugzilla.gnome.org/show_bug.cgi?id=755145
Clone the connection upon activation. This makes it safe for the user
to modify the original connection while it is activated.
This involves several changes:
- NMActiveConnection gets @settings_connection and @applied_connection.
To support add-and-activate, we constructing a NMActiveConnection with
no connection set. Previously, we would set the "connection" field to
a temporary NMConnection. Now NMManager piggybacks this temporary
connection as object-data (TAG_ACTIVE_CONNETION_ADD_AND_ACTIVATE).
- get rid of the functions nm_active_connection_get_connection_type()
and nm_active_connection_get_connection_uuid(). From their names
it is unclear whether this returns the settings or applied connection.
The (few) callers should figure that out themselves.
- rename nm_active_connection_get_id() to
nm_active_connection_get_settings_connection_id(). This function
is only used internally for logging.
- dispatcher calls now get two connections as well. The
applied-connection is used for the connection data, while
the settings-connection is used for the connection path.
- needs special handling for properties that apply immediately
when changed (nm_device_reapply_settings_immediately()).
Co-Authored-By: Thomas Haller <thaller@redhat.com>
https://bugzilla.gnome.org/show_bug.cgi?id=724041
Refactor agent-manager to always invoke the complete function for
nm_agent_manager_get_secrets().
In general, the complete function is always invoked asnychronously
when starting the operation. On the other hand, when cancelling the
operation or disposing the manager with pending operations, we now
(always) synchronously invoke the callback.
This makes it simpler for the user to reliably cancel the request
and perform potential cleanup.
This behavior bubbles up through NMSettingsConnection and NMActRequest,
and other callers that make directly or indicrectly make use of
nm_agent_manager_get_secrets().
Instead of having the call_id of type guint32, make it an (opaque)
pointer type.
This has the advantage of strong typing and avoids the possiblity
of reusing an invalid integer (or overflow of the call-id counter).
OTOH, it has the disadvantage, that after a call_id is disposed,
it might be reused for future invocations (because malloc might
reuse the memory).
In fact, it is always an error to use a call_id that is already
completed. This commit also adds assertions to the cancel() calls
that the provided call_id is a pending call. Hence, such a bug
will be uncovered by assertions (that only might not tigger in
certain unlikely cases where a call-id got reused).
Note that for NMAgentManager, save_secrets() and delete_secrets()
both returned a call_id. But they didn't also provide a callback when
the operation completes. So the user trying to cancel such a call,
cannot know whether the operation is still in process and he cannot
avoid triggering an assertion.
Fix that by not returning a call-id for these operations. No caller
cared about it anyway.
For NMSettingsConnection, also track the internally scheduled requests
for so that we can cancel them on dispose.
The localization headers are now included via "nm-default.h".
Also fixes several places, where we wrongly included <glib/gi18n-lib.h>
instead of <glib/gi18n.h>. For example under "clients/" directory.
Future patches will create devices long before they are backed by
kernel resources, so we need to split NMDevice object creation from
actual setup based on the backing resources.
This patch combines the NMDeviceFactory's new_link() and
create_virtual_device_for_connection() class methods into a single
create_device() method that simply creates an unrealized NMDevice
object; this method is not expected to fail unless the device is
supposed to be ignored. This also means that the NMDevice
'platform-device' property is removed, because a platform link
object may not be available at NMDevice object creation time.
After the device is created, it is then "realized" at some later
time from a platform link (for existing/hardware devices via the
realize() method) or from an NMConnection (for newly created software
devices via the create_and_realize() NMDeviceClass methods).
https://bugzilla.gnome.org/show_bug.cgi?id=737458
Add NMExportedObject, make it the base class of all D-Bus-exported
types, and move the nm-properties-changed-signal logic into it. (Also,
make NMSettings use the same properties-changed code as everything
else, which it was not previously doing, presumably for historical
reasons).
(This is mostly just shuffling code around at this point, but
NMExportedObject will be more important in the gdbus port, since
gdbus-codegen doesn't do a very good job of supporting objects that
export multiple interfaces [as each NMDevice subclass does, for
example], so we will need more glue/helper code in NMExportedObject
then.)
Rather than randomly including one or more of <glib.h>,
<glib-object.h>, and <gio/gio.h> everywhere (and forgetting to include
"nm-glib-compat.h" most of the time), rename nm-glib-compat.h to
nm-glib.h, include <gio/gio.h> from there, and then change all .c
files in NM to include "nm-glib.h" rather than including the glib
headers directly.
(Public headers files still have to include the real glib headers,
since nm-glib.h isn't installed...)
Also, remove glib includes from header files that are already
including a base object header file (which must itself already include
the glib headers).
The @udi field is not a static string, so any user of a NMPlatformLink
instance must make sure not to use the field beyond the lifetime of the
NMPlatformLink instance.
As we pass on the platform link instance during platform changed events,
this is hard to ensure for the subscriber of the signal -- because a
call back into platform could invalidate/modify the object.
Just not expose this field as part of the link instance. The few callers
who actually needed it should instead call nm_platform_get_uid(). With
that, the lifetime of the returned 'const char *' pointer is clearly
defined.
Try to set the MTU of the parent Ethernet interface to match the
requested PPP MTU and MRU. This allows the negotiation of a PPP MTU
and MRU greater than 1492.
Instead of hacky stuff in the Manager, let plugins themselves indicate
which links should be ignored (because they are really child links that
are controlled by a different device that the plugin handles).
Ethernet, WiFi, and VLAN used the same implementation for initial address.
Ethernet and WiFi used the same implementation (and duplicated code) for
permanent MAC address, plus they both used ethtool in what should be
generic code, which is better done in the platform.
Instead of looping over all plugins and asking each plugin whether it
can handle a link or a connection, have them advertise the link and
connection types they support, and use that when creating new devices.
Most nm_platform_*() functions operate on the platform
singleton nm_platform_get(). That made sense because the
NMPlatform instance was mainly to hook fake platform for
testing.
While the implicit argument saved some typing, I think explicit is
better. Especially, because NMPlatform could become a more usable
object then just a hook for testing.
With this change, NMPlatform instances can be used individually, not
only as a singleton instance.
Before this change, the constructor of NMLinuxPlatform could not
call any nm_platform_*() functions because the singleton was not
yet initialized. We could only instantiate an incomplete instance,
register it via nm_platform_setup(), and then complete initialization
via singleton->setup().
With this change, we can create and fully initialize NMPlatform instances
before/without setting them up them as singleton.
Also, currently there is no clear distinction between functions
that operate on the NMPlatform instance, and functions that can
be used stand-alone (e.g. nm_platform_ip4_address_to_string()).
The latter can not be mocked for testing. With this change, the
distinction becomes obvious. That is also useful because it becomes
clearer which functions make use of the platform cache and which not.
Inside nm-linux-platform.c, continue the pattern that the
self instance is named @platform. That makes sense because
its type is NMPlatform, and not NMLinuxPlatform what we
would expect from a paramter named @self.
This is a major diff that causes some pain when rebasing. Try
to rebase to the parent commit of this commit as a first step.
Then rebase on top of this commit using merge-strategy "ours".
Extend nm_match_spec_*() to support an "except:" prefix to negate
the result of a match. "except:" only works when followed by
an exact match type, for example "except:interface-name:vboxnet0",
but not "except:vboxnet0".
A matching "except:" spec always wins, regardless of other positive
matchings.
With this change, NMConfig is really immutable and all
modifyable parts migrated to NMConfigData.
Another advantage is that components can now subscribe to
NMConfig changes to pickup changes to no-auto-default.
nm_device_get_hw_address() may return NULL and nm_platform_link_get_type may
return NM_LINK_TYPE_NONE. While it might be a good idea to check for such cases
at the init time it seems easier to just ignore it and prevent blowing up in
subsequent deactivation.
A quick test case:
# while :; do ip link add moo0 type veth peer moo1; ip link del moo0 ; done
Yields:
NetworkManager:ERROR:devices/nm-device-ethernet.c:268:constructor:
assertion failed: (link_type == NM_LINK_TYPE_ETHERNET ||
link_type == NM_LINK_TYPE_VETH)
nm_device_set_hw_addr: assertion 'addr != NULL' failed
https://bugzilla.gnome.org/show_bug.cgi?id=740992
config.h should be included from every .c file, and it should be
included before any other include. Fix that.
(As a side effect of how I did this, this also changes us to
consistently use "config.h" rather than <config.h>. To the extent that
it matters [which is not much], quotes are more correct anyway, since
we're talking about a file in our own build tree, not a system
include.)
Add nm-core-types.h, typedefing all of the GObject types in
libnm-core; this is needed so that nm-setting.h can reference
NMConnection in addition to nm-connection.h referencing NMSetting.
Removing the cross-includes from the various headers causes lots of
fallout elsewhere. (In particular, nm-utils.h used to include
nm-connection.h, which included every setting header, so any file that
included nm-utils.h automatically got most of the rest of libnm-core
without needing to pay attention to specifics.) Fix this up by
including nm-core-internal.h from those files that are now missing
includes.
