The initscripts do this:
MATCH='^.+\.[0-9]{1,4}$'
if [[ "${DEVICE}" =~ $MATCH ]]; then
VID=$(echo "${DEVICE}" | LC_ALL=C sed 's/^.*\.\([0-9]\+\)/\1/')
PHYSDEV=${DEVICE%.*}
fi
MATCH='^vlan[0-9]{1,4}?'
if [[ "${DEVICE}" =~ $MATCH ]]; then
VID=$(echo "${DEVICE}" | LC_ALL=C sed 's/^vlan0*//')
# PHYSDEV should be set in ifcfg-vlan* file
if test -z "$PHYSDEV"; then
net_log $"PHYSDEV should be set for device ${DEVICE}"
exit 1
fi
fi
which means that if the VLAN name starts with "vlan" then
PHYSDEV must be set, otherwise the parent interface cannot
be determined.
Since PHYSDEV, if set, reflects the explicit intentions of the
user instead of assuming the name from DEVICE, make PHYSDEV
take precedence over determining the parent interface from
heuristics.
A couple functions depended on the passed-in error being !NULL to
correctly report errors, and we can't depend on that because it might
not be true. So fix up those functions' call chain to ensure that
errors get reported regardless of whether 'error' is !NULL.
We have to copy the UUID (key) because otherwise the pointer can be invalidated
when the connection is destroyed and problems will start.
The issue showed up as an unability to delete a conenction via D-Bus.
Reproducer:
$ nmcli con add type eth con-name AAA ifname blah
$ nmcli con delete AAA
$ nmcli con add type eth con-name AAA ifname blah
$ nmcli con delete AAA
-- here the connection is not removed from NM
(even though ifcfg- file) was removed --
Add a "monitor-connection-files" config option, which can be set to
"false" to disable automatic reloading of connections on file change.
To go with this, add a new ReloadConnections method on
o.fd.NM.Settings that can be used to manually reload connections, and
add an nm-cli command to call it.
Some plugins may emit :new-connection or :unmanaged-specs-changed
while reading connections, so don't connect to those signals until
after the initial load_connections() (and just unconditionally emit
:unmanaged-specs-changed at that point).
In ifcfg-rh's get_unmanaged_specs(), don't bother to try to read the
connections first; if they haven't been read yet, just return NULL;
NMSettings will call it again after the connections have been read.
ifcfg-rh didn't let you unmanage an InfiniBand device by hardware
address because it was recording the hardware address with uppercase
letters, while nm_match_spec_hwaddr() required lowercase. Fix this by
making nm_match_spec_hwaddr() match case-insensitively (and remove the
manual lowercasing that several other places were doing to work around
this.)
keyfile didn't let you unmanage an InfiniBand device by hardware
address because it only accepted ARPHRD_ETHER hardware addresses. Fix
that by using nm_utils_hwaddr_valid() instead.
We don't always want to immediately write new connections to disk, to
facilitate "runtime" or "temporary" connections where an interface's
runtime config isn't backed by on-disk config. Also, just because
an interface's configuration is changed doesn't necessarily mean
that new configuration should be written to disk either.
Add D-Bus methods for adding new connections and for updating existing
connections that don't immediately save the connection to disk.
Also add infrastructure to indicate to plugins that the new connection
shouldn't be immediately saved if the connection was added with the
new method.
Use the new NMConnection 'changed' signal to mark connections
as dirty/unsaved, and reset that when they get flushed to disk.
Previously, the 'Updated' signal was emitted only when the
connection was changed and flushed to disk, but now we have
more granular needs, and the signal is emitted whenever the
connection actually *is* changed, regardless of whether its
flushed to disk or not.
Although having different parts of NM in different subdirectories
keeps the source tree neat, it has made the build messy, particularly
because of cross-dependencies between the subdirs.
Reorganize to build all of the pieces of the NetworkManager binary
from src/Makefile, and only use recursive make for test programs,
helper binaries, and plugins.
As part of this, get rid of all the per-directory convenience
libraries, and switch to building a single top-level
libNetworkManager.la, containing everything except main.c, which all
of the test programs can then link against.
STP defaults to yes in NetworkManager (and the initscripts), so a missing
STP value in an ifcfg file means yes/on. Calling svSetValue(STP, NULL)
clears that line from the ifcfg, and thus STP gets interpreted as yes.
Explicitly set stp to "no" so that the value actually gets saved.
For settings corresponding to devices that have a :carrier property
(ie bond, bridge, infiniband, vlan, and wired), add a :carrier-detect
property specifying how that affects the connection:
yes: The connection can only be activated when the device
has carrier, and will be deactivated if the device loses
carrier (for more than 4 seconds).
no: The connection ignores carrier on the device; it can be
activated when there is no carrier, and stays activated
when carrier is lost.
on-activate: The connection can only be activated when the
device has carrier, but it will not be deactivated if the
device loses carrier.
https://bugzilla.gnome.org/show_bug.cgi?id=688284
GObject creation cannot normally fail, except for types that implement
GInitable and take a GError in their _new() method. Some NM types
override constructor() and return NULL in some cases, but these
generally only happen in the case of programmer error (eg, failing to
set a mandatory property), and so crashing is reasonable (and most
likely inevitable anyway).
So, remove all NULL checks after calls to g_object_new() and its
myriad wrappers.
https://bugzilla.gnome.org/show_bug.cgi?id=693678
g_malloc(), etc, never return NULL, by API contract. Likewise, by
extension, no other glib function ever returns NULL due to lack of
memory. So remove lots of unnecessary checks (the vast majority of
which would have immediately crashed had they ever run anyway, since
g_set_error(), g_warning(), and nm_log_*() all need to allocate
memory).
https://bugzilla.gnome.org/show_bug.cgi?id=693678
Allows to attach any connection to a bridge using the BRIDGE= key.
IP configuration is optional for bridge components but not
prohibited. Test case included.
This patch makes DHCPv6 support more or less equivalent to that
one of IPv4 DHCP.
(dcbw: fix some formatting, rearrange code so it's less convoluted,
fix up writing hostname to ifcfg files)
Use --enable-doc and --enable-tests instead of --with-docs and
--with-tests. This is consistent with other features and with
--enable-gtk-doc option. Support current variants as fallback.
Don't build tests unless --enable-tests is specified.
When determining the system hostname, /etc/hostname should override
/etc/sysconfig/network, so monitor both files.
When setting the hostname, set it in /etc/hostname, and delete the
/etc/sysconfig/network HOSTNAME entry if present.
https://bugzilla.redhat.com/show_bug.cgi?id=831735
We had separate checks for glib-2.0, gobject-2.0, gmodule-2.0, and
gio-unix-2.0. It doesn't make sense to link a binary against all 4
because gio-unix-2.0 depends on glib-2.0 and gobject-2.0. Doing this
actually breaks things in unusual circumstances.
Generally, few bits of NM actually just use glib, and not gio. We
might as well coalesce those requirements together, even if it means
in some cases we "overlink". Additionally, I chose for now to fold
gmodule-2.0 in as well, even though many fewer programs need it. The
cost of overlinking is quite small.
The benefit of this is less repeated junk in Makefile.am, as well as
more centralized control over GLib. A followup patch will allow us to
set -DGLIB_VERSION_MIN_REQUIRED in just one place, rather than having
to replicate it 4 times.
The NM configure is still suboptimal - for example, libpolkit-1
depends on gio-2.0, so really we should determine the compiler flags
all in one pass. But it doesn't matter too much for now.
This functionality is (mostly) obsoleted by the newer
GLIB_VERSION_MIN_REQUIRED and GLIB_VERSION_MAX_ALLOWED defines. With
this, your build doesn't all of a sudden blow up if we deprecate
something in GLib - you have to explicitly opt-in to the newer
version.
G_DISABLE_DEPRECATED does still apply for macros and things that can't
take __attribute__((deprecated)), but it's not really worth the pain
and cargo culting around just for that.
VLAN connections can have "hardware" settings in addition to the
VLAN-specific ones. ifcfg-rh was reading in wired settings for VLANs,
but was not writing them back out.
