This usually indicates that the driver missed beacons from the AP, due to driver bugs
or faulty power-save management. It doesn't mean that the PSK is wrong.
Previously, we would require a @self argument and the @call_id in
nm_act_request_cancel_secrets(), although the @call_id already has
a pointer to @self.
In principle that is not necessary, but it makes the API a bit
more robust as you need to care about the lifetime of the @req
as well.
However it is a bit inconvenient, because it requires that caller to
track both the activation request and the call-id.
Now, allow nm_act_request_get_secrets() to instruct the call-id to
take an additional reference to @self. Later on, we would allow to omit
the argument during cancelling. We only allow this, if the call-id
takes a reference to @self.
Generating "src/NetworkManager.ver" each time seems to work well.
Thus, src/NetworkManager.ver-orig is unused an gets easily out
of date. Just remove it. It's not useful anymore.
We used MASTER, BRIDGE and TEAM_MASTER keys for a differnet purpose than the
network.service did, confusing the legacy tooling. Let's do our best to write
compatible configuration files:
* Add *_UUID properties that won't clash with initscripts
* Ignore non-*_UUID keys on read if *_UUID is present
* If the connection.master is an UUID of a connection with a
connection.interface-name, write the uuid into the *_UUID key while setting
the non-*_UUID key to the interface name for compatibility
https://bugzilla.redhat.com/show_bug.cgi?id=1369091
Currently ipv4 and ipv6 settings share the ifcfg plugin vars for the
dhcp-hostname and dhcp-send-hostname options (DHCP_HOSTNAME and
DHCP_SEND_HOSTNAME). Add brand new vars DHCPV6_HOSTNAME and
DHCPV6_SEND_HOSTNAME to have var-property one to one mappings.
When those values are missing they are already considered as TRUE.
As they are written also when DHCP is not enabled, write them only
if they have a value different from the default.
If a connection is ACTIVATED and another one is ACTIVATING but there
is no global connectivity, we currently set the manager state to
CONNECTING and start a connectivity check to verify whether the
manager state can be promoted to CONNECTED_GLOBAL.
If this connectivity check fails, we shouldn't promote a CONNECTING
state to CONNECTED_SITE.
Fixes: 084da69a30
The manager state is already computed every time an active connection
changes state, it is not necessary to call nm_manager_update_state()
also when the property is read.
Moreover, nm_manager_update_state() emits a "notify::state" signal
which causes a re-read of the property by the nm-exported-object,
resulting in a nested execution of nm_manager_update_state().
We call nmp_utils_ethtool_get_driver_info() twice when receiving a
netlink message, but we don't need a clone of the string values.
Instead, expose a data structure that should be stack allocated
by the caller.
The ioctl APIs ethtool/mii require an interface ifname. That is inherrently
racy as interfaces can be renamed. This cannot be fixed, we can only
minimize the time between verifying the ifname and calling ioctl.
We already had problems with that when ethtool would access an interface
by name that didn't exists. See commit ab41c13b06 .
Checking for an existing interface only helps avoiding races when an interface
gets deleted. It does not help against renaming.
Go one step further, and instead of checking whether such an ifname
exists, try to get the ifname based on the ifindex immediately before
we need it.
This brings an additional overhead for each ethtool access.
A reimplementation of g_file_get_contents() to overcome two limitations:
- nm_utils_file_get_contents() accepts a @dirfd argument to open the
file relative using openat().
- nm_utils_fd_get_contents() allows to read the content from a file
filedescriptor.
- both support a max_length argument, to fail gracefully if we
get tricked into loading a huge file.
- use nm_auto_close cleanup attribute
- optionally, return the found ifname
- don't stat "phy80211". If such an entity can be opened,
just assume it's a directory.
Since function nmp_utils_open_sysctl() can avoid race condition, use it
in wifi_utils_is_wifi() to open sysfs and correctly check if it's a wifi
device.
https://bugzilla.gnome.org/show_bug.cgi?id=775613
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
A race condition may happen when NetworkManager opens sysfs and udev
renames interface name at the same time. Thomas Haller provides a new
function [1] which can avoid the race condition when opening sysfs.
This patch is a direct copy from [1].
[1] https://mail.gnome.org/archives/networkmanager-list/2016-December/msg00004.html
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
It would be nice that our logging functions are guaranteed to
preserve errno. We are currently not very consistent about handling
errno, let's improve on that.
update_dns() is rather complicated. The part where we collect
resolv-conf data can be moved out. That is nice, because
_collect_resolv_conf_data() has no side-effects (except logging)
and it's only purpose is to return the out-arguments.
