When adding an IPv4 address, kernel will also add a device-route.
We don't want that route because it has the wrong metric. Instead,
we add our own route (with a different metric) and remove the
kernel-added one.
This could be avoided if kernel would support an IPv4 address flag
IFA_F_NOPREFIXROUTE like it does for IPv6 (see related bug rh#1221311).
One important thing is, that we want don't want to manage the
device-route on assumed devices. Note that this is correct behavior
if "assumed" means "do-not-touch".
If "assumed" means "seamlessly-takeover", then this is wrong.
Imagine we get a new DHCP address. In this case, we would not manage
the device-route on the assumed device. This cannot be fixed without
splitting unmanaged/assumed with related bug bgo 746440.
This is no regression as we would also not manage device-routes
for assumed devices previously.
We also don't want to remove the device-route if the user added
it externally. Note that here we behave wrongly too, because we
don't record externally added kernel routes in update_ip_config().
This still needs fixing.
Let IPv4 device-routes also be managed by NMRouteManager. NMRouteManager
has a list of all routes and can properly add, remove, and restore
the device route as needed.
One problem is, that the device-route does not get added immediately
with the address. It only appears some time later. This is solved
by NMRouteManager watching platform and if a matchin device-route shows up
within a short time after configuring addresses, remove it.
If the route appears after the short timeout, assume they were added for
other reasons (e.g. by the user) and don't remove them.
https://bugzilla.gnome.org/show_bug.cgi?id=751264https://bugzilla.redhat.com/show_bug.cgi?id=1211287
Change nm_platform_link_get() to return the cached NMPlatformLink
instance. Now what all our implementations (fake and linux) have such a
cache internal object, let's just expose it directly.
Note that the lifetime of the exposed link object is possibly quite
short. A caller must copy the returned value if he intends to preserve
it for later.
Also add nm_platform_link_get_by_ifname() and modify nm_platform_link_get_by_address()
to return the instance.
Certain functions, such as nm_platform_link_get_name(),
nm_platform_link_get_ifindex(), etc. are solely implemented based
on looking at the returned NMPlatformLink object. No longer implement
them as virtual functions but instead implement them in the base class
(nm-platform.c).
This removes code and eliminates the redundancy of the exposed
NMPlatformLink instance and the nm_platform_link_get_*() accessors.
Thereby also fix a bug in NMFakePlatform that tracked the link address
in a separate "address" field, instead of using "link.addr". That was
a case where the redundancy actually led to a bug in fake platform.
Also remove some stub implementations in NMFakePlatform that just
bail out. Instead allow for a missing virtual functions and perform
the "default" action in the accessor.
An example for that is nm_platform_link_get_permanent_address().
For NMPlatform instances we had an error reporting mechanism
which stores the last error reason in a private field. Later we
would check it via nm_platform_get_error().
Remove this. It was not used much, and it is not a great way
to report errors.
One problem is that at the point where the error happens, you don't
know whether anybody cares about an error code. So, you add code to set
the error reason because somebody *might* need it (but in realitiy, almost
no caller cares).
Also, we tested this functionality which is hardly used in non-testing code.
While this was a burden to maintain in the tests, it was likely still buggy
because there were no real use-cases, beside the tests.
Then, sometimes platform functions call each other which might overwrite the
error reason. So, every function must be cautious to preserve/set
the error reason according to it's own meaning. This can involve storing
the error code, calling another function, and restoring it afterwards.
This is harder to get right compared to a "return-error-code" pattern, where
every function manages its error code independently.
It is better to return the error reason whenever due. For that we already
have our common glib patterns
(1) gboolean fcn (...);
(2) gboolean fcn (..., GError **error);
In few cases, we need more details then a #gboolean, but don't want
to bother constructing a #GError. Then we should do instead:
(3) NMPlatformError fcn (...);
Later remove nm_platform_get_error() and signal errors via return
error codes.
Also, fix nm_platform_infiniband_partition_add() and
nm_platform_vlan_add() to check the type of the existing link
and fail with WRONG_TYPE otherwise.
- rename "NONE" to "SUCCESS", what it really is.
- change the to-string result not to contain spaces
and being closer the name of the enum value.
- add new error reasons "UNSPECIFIED" and "BUG".
- remove the code comments around the enum definition.
They add no further description about why this error
happens and only paraphrase the name of the enum.
- reserve negative integers for 'errno'. This is neat
because if we get a system error we can pass on the
underlying errno as cause.
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.
Add a construct-only property NM_PLATFORM_REGISTER_SINGLETON to NMPlatform.
When set to TRUE, the constructor will self-register to nm_platform_setup().
The reason for this is that the _LOG() macro in NMLinuxPlatform logs the
self pointer if the instance is not the singleton instance.
During construction, we already have many log lines due to initialization
of the instance. These lines all end up qualified with the self pointer.
By earlier self-registering, printing the pointer value is omitted.
Yes, this patch is really just to prettify logging.
NMPObject is a simple "object" implemenation around NMPlatformObject.
They are ref-counted and have a class-pointer. Several basic functions
like equality, hash, to-string are implemented.
NMPCache is can be used to store the NMPObject. Objects are indexed
via their primary id, but there is also multi-lookup via NMCacheId
and NMMultiIndex.
Part of the implementation is inside "nm-linux-platform.c",
because it depends on utility functions from there.
Cache the scope as part of the NMPlatformIP4Route and
no longer read it from libnl object when needed. Later
there will be no more libnl objects around, and we need
to scope when deleting an IPv4 route.
nm_platform_query_devices() would raise an 'added' signal
for all its links. That is bad style because it could
confuse other listeners for platform signals which don't
expect such artificial change signals.
The public API of NMPlatform already gives NMManager the ability
to 'pull' all the links and iterate them itself.
Before, nm_platform_query_devices() would also initialize udev
devices, so there was a more compelling reason for this function.
We already populate the netlink cache in constructed(). No need
to wait with udev devices until nm_platform_query_devices(). Just
do it right away.
Add a hack to keep 'lo' default-unmanaged. Now that we load
udev devices earlier, we end up clearing the default-unmanged
flag on 'lo', which has bad consequences.
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.
Always intern string from udev_get_driver().
We use the result of udev_get_driver() for setting NMPlatformLink.driver.
In all other cases, we already set that value to an interned string,
which simplifies memory handling.
As it was, the lifetime of that string was tied to the lifetime of the
GUdevDevice.
This is not a stelar solution, but we assume that the overall numbers
of different drivers is limited so we don't leak large amounts of
memory.
link_extract_type() would return the NMLinkType and a
@type_name string. If the type was unknown, this string
was rtnl_link_get_type() (IFLA_INFO_KIND).
Split up this behavior and treat those values independently.
link_extract_type() now only detects the NMLinkType. Most users
don't care about unknown types and can just use nm_link_type_to_string()
to get a string represenation.
Only nm_platform_link_get_type_name() (and NMDeviceGeneric:type_description)
cared about a more descriptive type. For that, modify link_get_type_name()
to return nm_link_type_to_string() if NMLinkType could be detected.
As fallback, return rtnl_link_get_type().
Also, rename the field NMPlatformLink:link_type to "kind". For now this
field is mostly unused. It will be used later when refactoring platform
caching.
Given the name nm_link_type_to_string(), we would not expect
to find it in nm-linux-platform.c. It either should be named
nm_platform_link_type_to_string() and be put in a new
nm-platform-utils.c file, or it should be named
nm_utils_link_type_to_string() and be put in NetworkManagerUtils.h.
For now, just leave it here.
Instead of having a nm_platform_free() function, use NM_DEFINE_SINGLETON_WEAK_REF()
and register a weak reference. That way, users who want to free the platform
instance can just unref it.
We already have nm_*_platform_setup() that gets specified
via -DSETUP. This SETUP() hook gives us all the flexiblity
we need to customize our singleton, so just do any required
setup there.
Also, it would be easier to add an alternative (hypotetical)
nm_fake_platform_setup_custom() to customize the singleton then to
parametrize the NMPlatform:setup() implementation. So this virtual
function is less flexible and redundant.
We have two hooks to modify setup of the platform singleton:
nm_linux_platform_setup() and the virtual setup() function.
On the other hand, nm_platform_setup() limits us by accepting
only a GType, instead of a prepeared platform instance.
Make the nm_platform_setup() method more flexible, so that we can
later drop the setup() hook.
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".
Some out of tree drivers add Ethernet devices that are supposed to be managed
by other their tooling, e.g. VirtualBox or VMWare.
Rather than hardcoding their drivers (at least VirtualBox doesn't even set a
"driver" property in sysfs) or hardcoding a logic that identifies such devices
let's just add a possibility to blacklist them in udev. This makes it possible
for whoever who ships such a driver to ship rules that prevent NetworkManager
from managing the device itself.
Furthermore it makes it possible for the user with special needs leverage the
flexibility of udev rules to override the defaults. In the end the user can
decide to let NetworkManager manage default-unmanaged interfaces such as VEth
or turn on default-unmanaged for devices on a particular bus.
An udev rule for VirtualBox would look like this:
SUBSYSTEM=="net", ENV{INTERFACE}=="vboxnet[0-9]*", ENV{NM_UNMANAGED}="1"
Create a NMRouteManager singleton.
Refactor, no functional changes apart from change of log domain from
LOGD_PLATFORM to LOGD_CORE.
Subsequent commit will keep track of the conflicting routes, avoid overwriting
older ones with newer ones and apply the new ones when the old ones go away.
