When activation of the connection fails with no-secrets, we block
autoconnect due to that. However, NMPolicy also unblocks such
autoconnect, whenever a new secret-agent registers. The reason
is obviously, that the new secret-agent might be able to provide
the previously missing secrets.
However, there is a race between
- making the secret request, failing activation and blocking autoconnect
- new secret-agent registers
If the secret-agent registers after making the request, but before we
block autoconnect, then autoconnect stays blocked.
[1511468634.5759] device (wlp4s0): state change: config -> need-auth (reason 'none', sys-iface-state: 'managed')
[1511468634.5772] device (wlp4s0): No agents were available for this request.
[1511468638.4082] agent-manager: req[0x55ea7e58a5d0, :1.32/org.kde.plasma.networkmanagement/1000]: agent registered
[1511468638.4082] policy: re-enabling autoconnect for all connections with failed secrets
[1511468664.6280] device (wlp4s0): state change: need-auth -> failed (reason 'no-secrets', sys-iface-state: 'managed')
[1511468664.6287] policy: connection 'tuxmobil' now blocked from autoconnect due to no secrets
Note the long timing between making the secret request and the
activation failure. This race already existed before, but now with
WPS push-button method enabled by default, the duraction of the
activation is much longer and the race is easy to hit.
https://bugzilla.gnome.org/show_bug.cgi?id=790571
Extend the enum and API to use flags for the blocked reasons.
A connection is blocked from autoconnect if it has any reason
set.
There is no behavioral change in this patch beyond that, because
where we previously would set blocked-reason NONE, we would still
clear all flags, and not only a particular one.
Later of course, we want to set and clear individual flags
independently.
It's not ever going to change(*), and NMPolicy calls reset() a lot.
No need to lookup the configuration in the GKeyFile every time.
(*) per NMConfigData instance. The config may be reloaded, in which
case NMConfig creates a new NMConfigData instance, but the NMConfigData
instance itself is immutable.
NMPolicy would at various time call nm_settings_connection_autoconnect_retries_reset()
followed by nm_settings_connection_autoconnect_retries_get().
This resulted in two logging messages, first to indicate that the value
was unset, and then reset it to the value from configuration. While that
is correct, it causes a lot of verbose logging. Especially for all connections
which autoconnect retry counter didn't actually change.
The advantage of that was, that we only loaded the actual value when we
need it the first time (during get()). That means, the user could reload
the configuration, and the value would be loaded and cached at a later
pointer.
However, the duplicate logging was annoying, but we still want to see
a message about the resetting.
So, now during reset load the value setting from NetworkManager.conf
and set it right away. Skip the intermediate UNSET value. In most
cases nothing changed now, and we don't log anything for most
connections.
Note that for the
if (nm_device_state_reason_check (reason) == NM_DEVICE_STATE_REASON_NO_SECRETS)
case we no longer do the
if (nm_settings_connection_autoconnect_retries_get (connection) == 0)
check. But that is fine, because we only skip schedling a reset_connections_retries()
action. But note, that that previously we also would never actually
scheudle a new timeout, because
- either nm_settings_connection_autoconnect_retries_get (connection) != 0
- or the retries count was zero, in which case we already have a
reset_connections_retries action pending (from the time when we
set it to zero.
So, there is no change in behavior at all except dropping of a redundant
logging line.
NMSettingsConnection has 3 properties that are related to autoconnect:
- autoconnect_retries
- autoconnect_blocked_until
- autoconnect_blocked_reason
autoconnect_blocked_reason is entirely independent from the other two.
A connection have have autoconnect blocked via a blocked-reason, but the
retry count is not affected by that. The retry count is an independent
mechanism, that may additionally prevent autoconnect.
However autoconnect_retries and autoconnect_retries_blocked_until are
strongly related. The latter is set if and only if autoconnect_retries is
at zero.
Rename to reflect that better.
We cannot just call g_object_set() with an integer that is out of bound.
Otherwise, glib will warn. We can use nm_g_object_set_property*() to return
an error without asserting.
Currently both bridge.mac-address and ethernet.cloned-mac-address get
written to the same MACADDR ifcfg-rh variable; the ethernet property
wins if both are present.
When one property is set and the connection is saved (and thus reread)
both properties are populated with the same value. This is wrong
because, even if the properties have the same meaning, the setting
plugin should not read something different from what was written. Also
consider that after the following steps:
$ nmcli con mod c ethernet.cloned-mac-address 00:11:22:33:44:55
$ nmcli con mod c ethernet.cloned-mac-address ""
the connection will still have the new mac address set in the
bridge.mac-address property, which is certainly unexpected.
In general, mapping multiple properties to the same variable is
harmful and must be avoided. Therefore, let's use a different variable
for bridge.mac-address. This changes behavior, but not so much:
- connections that have MACADDR set will behave as before; the only
difference will be that the MAC will be present in the wired
setting instead of the bridge one;
- initscripts compatibility is not relevant because MACADDR for
bridges was a NM extension;
- if someone creates a new connection and sets bridge.mac-address NM
will set the BRIDGE_MACADDR property instead of MACADDR. But this
shouldn't be a big concern as bridge.mac-address is documented as
deprecated and should not be used for new connections.
https://bugzilla.redhat.com/show_bug.cgi?id=1516659
Settings plugins now return the connection that was reread from file
when adding a connection, which means that any agent-owned secret is
lost. Ensure that we don't forget agent-owned secrets by caching them
and readding them to the new connection returned by plugins.
Fixes: 8a1d483ca8
Fixes: b4594af55ehttps://bugzilla.gnome.org/show_bug.cgi?id=789383
Bond options are stored in a hash table and the order in which they
are returned by the API is not guaranteed. Sort them alphabetically so
that a connection will always be written in the same way, even if the
internal implementation of the hash table or the hashing function
changes, as it did in commit a6be2f4aa9 ("all: use nm_str_hash()
instead of g_str_hash()").
We also do this for libnm, where it causes visible changes
in behavior. But if somebody would rely on the hashing implementation
for hash tables, it would be seriously flawed.
GHashTable optimizes a NULL equality function to use direct pointer
comparison. That saves the overhead of calling g_direct_equal().
This is also documented behavior for g_hash_table_new().
While at it, also don't pass g_direct_hash() but use the default
of %NULL. The behavior is the same, but consistently don't use
g_direct_hash().
Next we will use siphash24() instead of the glib version g_direct_hash() or
g_str_hash(). Hence, the "nm-utils/nm-hash-utils.h" header becomes very
fundamental and will be needed basically everywhere.
Instead of requiring the users to include them, let it be included via
"nm-default.h" header.
Comparing @secrets_keys indicates to coverity that it might be NULL.
Below, we access @secrets_keys without check, and coverity doesn't realize
that this cannot crash, because secrets_keys_n would be zero too.
Anyway, this way we safe the sorting, in case we only have
one element.
Kernel doesn't support it for IPv6.
This is especially useful, if you combine static routes
with DHCP. In that case, you might want to get the device-route
to the gateway automatically, but add a static-route for it.
Currently the ifcfg-rh plugin doesn't explicitly store the connection
type for team slaves and is only able to read back ethernet and vlan
connections.
Leave this unchanged for ethernet and vlan slaves, but store the TYPE
variable for other connection types (Wi-Fi and Infiniband) so that we
can properly determine their type when the connection is read.
The number of authentication retires is useful also for passwords aside
802-1x settings. For example, src/devices/wifi/nm-device-wifi.c also has
a retry counter and uses a hard-coded value of 3.
Move the setting, so that it can be used in general. Although it is still
not implemented for other settings.
This is an API and ABI break.
