Age | Commit message (Collapse) | Author | Files | Lines |
|
|
|
The previous take did not contain the stuff that initially motivated
rewriting the quoting function.
The code of both PPP connection types is still sub-par, but at least
functionally, it is a little better now.
|
|
The most important change is that we now bring the interface down
correctly in pppoe_down.
|
|
|
|
The tuntap connection type requires both a User and a Group to be
defined. However, when configuring a tuntap device in the kernel, the
device's uid and gid must both match for someone to be able to use the
device. The only way to create a tuntap device accessible to a group of
users is therefore to avoid specifying the user argument to
"ip tuntap add".
Reported by: A Web
|
|
Use 'wpa' to refer to the 802.11i and 802.1x client side software
(wpa_supplicant) in general.
|
|
Router advertisements can tell clients to not generate a public address automatically (AdvAutonomous off).
Reported by: Jonne Haß
|
|
Split them into different variables for different clients.
Also, IP6 gets its own, so that it can have options different from the
IP4 invocation.
|
|
|
|
|
|
Also:
- Replaced echos with invocations of cat, where possible.
- Fixed PIDFILE to point to the file ppp creates. This ensures the
interface can be brought down.
- Fixed setting of InterfaceRoot.
- Removed vim syntax highlighting from ethernet connection.
|
|
This allows future connection types to use interfaces other than
networking interfaces, for instance to use USB interfaces.
|
|
It had grown a little diffuse.
|
|
|
|
The up/down functions don't take arguments as they did in netcfg.
|
|
|
|
This allows connection scripts to provide defaults, among other things.
|
|
Strings that look like arguments, as well as no string at all, were
causing trouble.
Reported by: Thomas Bächler
|
|
wpa_supplicant would not immediately terminate with an attached ctrl monitor.
|
|
Remove distribution-specific paths (and allow overrides).
|
|
This fixes the use of all sorts of characters in profile names/interface names.
|
|
${#array[@]} requires "[@]" to count the array members. Otherwise it will
return the length of the first element which will hardly ever be 1.
Signed-off-by: Florian Pritz <bluewind@xinu.at>
|
|
The unit corresponding to a profile named $p is netctl@$p.service.
We should not drop the suffix, since $p could contain a valid unit
suffix (although .service is not allowed).
We take care of this uniformly by wrapping around systemctl.
Additionally, `systemctl list-units` does not accept an '--active'
parameter, so we fix it to be consistent with `netctl list`.
|
|
Redirecting inside a do_debug call is, of course, impossible.
|
|
Custom interface names like "eth-phone" or "eth-lte" are not uncommon,
and since they are used in pathes inside systemd units, dashes and unprintable
characters shall be escaped per unit file rules.
So replace "-" with "\x2d" in interface names.
|
|
Old location: /etc/network.d
New location: /etc/netctl
This should ease migration from netcfg and is a better path anyway.
|
|
This is meant to reduce confusion where netctl should be used instead of
systemctl.
|
|
A default netmask makes no sense (except possibly /32).
Spotted by: Thomas Bächler <thomas@archlinux.org>
|
|
BindsTo= is not enough, we also need After=.
|
|
If netctl is already connected to a network and wifi-menu is run, then
it will fail to connect. Having wifi-menu call `netctl switch-to`
instead of `netctl restart` causes wifi-menu to behave like netctl
switch-to: It will connect if not connected, and if already connected on
that interface it will put the connected profile down the put up the
selected profile.
This fixes github issue #6.
Signed-off-by: William Giokas <1007380@gmail.com>
|
|
After the systemd 197 device name changes, wlan0 no longer exists (it
does for people who mask the rule, but new users will have issues with
this). This just searches the /sys/class/net directory for wireless
names. If there is more than one, then report an
"Invalid interface specification", not a "Missing" interface (as there
could be multiple).
A quick thanks to Dave Reisner for his help in the channel.
Altered by: Jouke Witteveen <j.witteveen@gmail.com>
|
|
Reduced unnecessary quoting and some rewording.
|
|
- Stop dhcpcd also when DHCPClient is not specified
- Exit successfully in ifplugd on going down
|
|
This settles a style for 'case' statements:
The label and closing colons are indented with two spaces,
The body is indented with two more, totalling four spaces.
|
|
As ifplugd brings the interface up, we need to use ForceConnect.
|
|
|
|
Curtis Shimamoto noted it did not work.
|
|
- The services now bind to their interfaces correctly.
- Documentation of ExcludeAuto is added.
|
|
It was quite useless.
|
|
When starting/stopping multiple profiles, check whether we don't try to
start/stop no profiles at all.
|
|
|
|
This commit contains the refactoring and rewriting of code.
|
|
This commit contains the moving of files.
|
|
This is an errata to f1598, in which I introduced an error.
|
|
Example
DNS=('8.8.8.8' '8.8.4.4')
DNS_OPTIONS=('rotate' 'timeout:1')
Will create the following /etc/resolv.conf
nameserver 8.8.8.8
nameserver 8.8.4.4
options rotate
options timeout:1
(patch edited by Jouke Witteveen)
|
|
Implement the current kernel ABI and support sysfs paths encountered on current systems.
|
|
Check "$WPA_CONF_DIR/$INTERFACE" for existence; exit if it's absent
(that is, wpa_supplicant isn't running).
Just avoids an error message from wpa_cli.
Signed-off-by: Ivan Shapovalov <intelfx100@gmail.com>
|
|
Reported by coin3d.
|
|
Executing "dhclient -x" also needs option "-pf" and a PID-file.
Signed-off-by: Ivan Shapovalov <intelfx100@gmail.com>
|
|
This commit removes an undocumented feature where IFACE_{UP,DOWN} would
be evaluated when the interface is brought up or down.
This functionality did not work as expected since the interface could
potentially be brought up/down more than once in one netcfg invocation.
This is apparent in the changes to the bonding code.
|