pfSense

Consider parent mtu when it's set instead of default one

Simplify a bit the logic

Pass parent interface to link_interface_to_vlans

Fix typo on var name

Fix typo s/assgined/assigned/

Be consistent on the check

Solve the issue when changing vlan parent mtu and leaving its childs with older mtu.

More fixes

Remove debug code

Correct code to apply the mtu correctly

Correctly handle mtu changing between vlans

Change parent only if its mtu is smaller

Skip vlan checks for being configured vlan as well

Only treat interesting interfaces

Make this work better and not always go over things for setting mtu

Hopefully is correct now

Set mtu to the correct interface

Correct which mtu value is used to be correct

Sprinkle some XXX just for being consistent

Correct this behaviour of properly honoring the configured mtu

Tune up a bit dhclient-script and call the kill state command only once its enough

You can have multiple 6rd tunnels now days

Remove old ip information even from the tracking interface as well so that track6 config code does not get confused. Fixes #2627

Remove old ip information so that track6 config does not get confused. Fixes #2627

Save ppp info and realif on interfaces.apply to be able to cleanup old configuration when changing interface type. Fixes #2758

Make parent interface and all VLANs use the same MTU. Fixes #2786

Create rules for grouped interfaces before regular ones. Fixes #2837

Fix is_altq_capable

Fixes traffic shaper determine which interfaces should be displayed and used.

simplify logic a bit

Set pfsync0 as down when it's not being used

Fixes on HA setup:

- Configure pfsync when configuration is changed
- Disable pfsync when it's disabled on GUI
- Remove syncpeer and back it to default multicast address when it's
removed from conf

Fixes #2491

Set if interface is enable before try to configure static arp. Fixes #1988

find_interface_ipv6 should return the first configured address, not the last one

A nic family name can contain numbers on it, remove only the index in the end. Fixes #2010

Remove dead code

Fix these tests, they will never succeed the way we set these variables using the old test. Fixes #1987

Ticket #1629 Another round of fixes related to state clearing

Commit missed global

Optimize get_real_itnerface to avoid a loop of doing a convertion from real back to friendly when the inforamtion is there already

Try to use standard interface to find the source for connecting a tunnel. Also if a ping host is defined for the configuration used that for remoteid since its known to be reachble

Teach dhcpd_configure about family optional parameter and use it to properly setup dhcpv6 on track6 case so that config is properly written

Actually do restart radvd and dhcp6d on all track6 interfaces

Use posix_kill here

Restart dhcp and radvd on change

Actually for SLAAC type interfaces being tracked the configuration is needed

Revert "Remove track6 dhcp configuration since dhcp6 will do this for us."

This reverts commit 6ebfa0ccfd7db500a4f85d2d45ebd74699a8805f.

Remove track6 dhcp configuration since dhcp6 will do this for us.

Remove this old code

Correctly set the mtu of the interface during configuration. Since for interfaces that are clonable previously the mtu was set than the interface destroy so the operation was useles. While here correct check for vlan interfaces.

Add check_subnetsv6_overlap()

find_virtual_ip_alias() add IPv6 support

find_ip_interface() add IPv6 support

When IP is not set, the field doesn't have the value 'none', use empty() to check instead

Fix get_interface_subnetv6() for VIPs

Removed unused block of code

Use the better -Fs modifies to pf to kill the states by interface. Also kill both sides on an interface when -k needs to be used

Load if_stf module if has not been done

Add missing parameter to interface_track6_configure() call

Fix a typo

Do delayed handling of track6 interfaces a bit earlier. More thoughts need to be done on the track6 configuration

Correct name as reported by phil-davis@github

Destroy the stf interface to avoid issues with ipv6 interface assignment. Also delete any previous address on track6 interfaces before assigning a new one.

Do not do any operation if its not necessary

Create link_interface_to_track6 to make code more readble and easily trackble. Also improve the generation of rules to avoid creating problems during filter reload and some optimizations with it

Merge pull request #362 from bcyrill/patch-4

Adapt regex pattern for new stp interface naming

Do more strict checks. Also silence some more log_error under debug. Also do some sanity checks on rc.newwanipv6 before doing any operation

Update etc/inc/interfaces.inc

With the new support for multiple stf interfaces their names have changed. Adapt the regex pattern.

Correct fetching ipv4 gateway by passing correct parameter

pfctl -b is gone just use std tools of pfctl

Make this regular expression a little more specific so it does not kill the wrong wpa_supplicant if there are double-digit interface numbers.

Corrected cmd line for hostapd

Revert "Use pid even for hostapd rather then trying to guess with regex"

This reverts commit b73a96b68d1d4e919a268d2e8e3a6ec902709e6d.

Use pfSense function where possible and rename stf0 to interface_stf to allow multiple instances of iot

I am not sure why this is needed but for sure its not needed during bootup

Fix wrong unset calls

Do not do so much work during boot since all interfaces will be called for configuration no need to go and find those. Also delay tracking interfaces during config to give a chance their parents to come up

Use pid even for hostapd rather then trying to guess with regex

If the interface is not enabled do not try to configure it

Make check more strong and readble. Also make dhcp6c use its pid file to make things easy

Use unescaped variable in if condition

Remove return value of interface_bridge_configure

Correct another parameter passed to pfsense functions

Fix holdcount/holdcnt parameter confusion

No need for escaping here since this is php internal functions

Do not use array_keys uselessly

Maintain the previous order of creating wireless interfaces at boot so additional virtual APs get the same generated MAC as before.

Actually even earlier since lagg can have wireless ifaces as member or vlan can reside on top of them

To avoid issues reported during bootup with wireless create clones early on bootup. Wireless can only have base interfaces and not clonable one so this should had been done since beggining.

Return the full list of vips to help with #1823

Mostly for testing name relation in github. No functional change

Put for now the code here in preparation of properly cleaning a recofnigured interface

Teach interface_bring_down to read a cofniguration passed in. Will be useful for Ticket #2758

Do not be so verbose by default this is just needed by a developer at best

Resolves #2482. Remove an ipv6 alias from a vip eevn though the pfsense module does not have support for that

Escape some variables used in mwexec() calls

Validate QinQ interface to be sure it's valid

Fix indent

Fix indent

Update etc/inc/interfaces.inc

As noted in http://forum.pfsense.org/index.php/topic,57849.0.html

Set flag setautosrc to replicate parent's mac address automatically when it changes, it's a better solution then dettach it. This should fix #2641