Fixes #1444. Implements the same fix that issue ermal made in #1522 to the console menu
This function takes two parameters, fix PHP error. Reported at http://forum.pfsense.org/index.php/topic,36648.0.html
Ticket #CZH-831780. If gif(4) is part of a bridge and its mtu is smaller than 1500(ethernet standard) do not consider it in finding the smaller mtu because we have a patch to allow gif(4) be member of a bridge with smaller mtu. See https://rcs.pfsense.org/projects/pfsense-tools/repos/mainline/commits/67d3135722db4a3c911761ead5c881ccaef02c65 for details.
Fixes #1394. Create a function get_itnerface_default_mtu and use it for resetting the mtu of a interface to default when needed. This adds the overhead of fetching the interface mtu and comparing with the default one every interface configuration run.
Correct saving of qinq specified members and also correctly destroy parent vlan when deleteing the interfaces. Also take care of attaching to netgraph now that we detach by default.
Make sure that openvpn tunnels are not impacted by hitting 'Save' on the Interface->Configuration page when assigned.
Use the needed variable here so hitting 'Save' from Interface->Configuration section does not leave the assigned gif interfaces without tunnel addresses.
Some configurations might have gre/gif on top of carp. Make sure to handle this configurations and to bring the tunnel correctly up.
Actually call interfaces_carp_setup after the carp interfaces are created so carp traffic can only flow after we have all vips up and running. This prevents premption more early than necessary. Ticket #1432.
Block instead of allowing proto carp/pfsync during bootup since this may cause issues. Ticket #1432
Do not error out for pptp/l2tp localip setting not set since its legitimate.
Do more strict checking if an ppp type interface is assigned before starting the mpd process behind it. Trigered-by: http://forum.pfsense.org/index.php/topic,34377.0.html
Add cas(4)
Pass the -a parameters to pgrep to be certain we search ancestors as well. The side effects might be inoquos from the pfSense context.
Allow queues on top of bridge. Though more investigation is needed on its correct meaning.
Resolves #1288. Add alc(4) to altq(4) supported list.
Comment out this code since it currently segfaults and get_real_interface could cause major slowdown here for some configurations.
Remove extra call to get_real_interface.
Use the first element of the array instead of the array itself for the comparison.
Revert "I think this is supposed to be get_real_interface too, not get_parent." - This function should not call get_real_interface here to avoid slowdown from recursion.
This reverts commit 54ac5d9080c2ea6669af07aa49a5ce660f2ede76.
Make get_parent_interface return an array to handle MLPPP and make it find vlan parents too.
Also, update interface_netgraph_needed to handle MLPPP on vlans.
Add NULL check to wireless functions.
The old interface_translate_type_to_real function used previously would return OPTXif it was passed "OPTX" and OPTX existed but was unassigned, whereas get_real_interface returns NULL, so now wecheck for NULL.
I think this is supposed to be get_real_interface too, not get_parent.
Revert if I mis-understood.
Change name of function "interface_translate_type_to_real" to match what it's doing (or should be doing.)
Next commits will change functionality of this function because before nowit's identical to "get_real_interface" function (because of bug mentioned below.)...
We don't want to detach netgraph nodes from interfaces that are used byany PPPoE/PPTP/L2TP configurations.
Ticket #259 trim the \n from the command output and return only the numeric part of it.
Really do not deactivate netgraph if the underlying interface is needed by it!
When the parent interface of a pppoe is assigned do not remove the interface from ng_ether since this will break ppp clients of netgraph with this interface as parent. This solves http://forum.pfsense.org/index.php/topic,31247.75.html
Remove uncessesary function call.
This function call is extraneous. mpd will automatically shut down existing mpd processesgracefully because we call it with the "-k" command line switch.
In addition, this call causes mpd to be started twice in rapid succession when users have...
Reshuffle code a bit. Prevent potential race condition between one mpd instanceshutting down and the second one starting.
There's no need to kill mpd before restarting it since we always start it withthe "-k" flag which means "wait for existing mpd process to exit gracefully before starting."
Hanlde the case when advbase is not present especially on previous 2.0 setups.
Ticket #757. Use correct pidfile.
nuke trailing carriage returns
Do not do useless work.
attach to ng_ether(4) only on pppoe/l2tp/pptp links no others.
Unbreak pppoe clients. Pointy-hat: myself
Do not attach ng_etther(4) to every system interface. Instead do a search if netgraph is needed on single/every interface during interface configuration. Also enable netgraph support for interface as needed when enabling pptp/l2tp/pppoe/... . This should prevent the netgraph queue to slow down network performance on fast links.
Teach convert_friendly_interface_to_friendly_descr about carp vips. This helps cases like Ticket #1086.
Make this at least right though it might not be correct.
Use the array index for this.
Expose advbase option of carp(4) to alleviate advanced configs.
Move all dynamic dns update processes under the same even 'reload dyndns' since it makes sense to do so.
Add lem(4) to our drivers list.
Fix vip descriptions in openvpn and ipsec screens. Ticket #1042
Remove unused binary from list
cleanup other bogus code.
Remove bogus code and use proper function here.
Actually honor the mtu/mac spoofing option in the interfaces.php page even while the type is set to other than dhcp/static/none. For this inhance the interface_translate_type_to_real to return the real hardware interface for ppp* types.
Do not reload twice carp and vips in general during an interface reload.
Correct code to reconfigure an interface for groups/gre/gif tunnels to take into account that an interface might be part of many instances of this. Also reconfigure interface groups for every newwanip trigger on interfaces to make sure dynamic interfaces retain their membership on groups.
Do not call filter_configure here the caller does. Also remove old code of the times when check_reload_status used files as trigers.
The caller of this function is responsible for this and usually all the callers do this from what i can tell. Remove redundant call.
Restore spoofed MAC after running hostapd to keep behavior consistent with other interfaces. Ticket #841
Restore the original MAC address before running hostapd to work around a hostapd bug. Ticket #841
Use correct variable name. Also related to Ticket #847.
Ticket #1047. Make the flags the same for each member interface before adding them to lagg(4)/bridge(4)
Set net.inet.carp.log=1. Ticket #1038
Nuke trailing c/r
Send errors to 2>
Fix test for altq on vlans and wlan.
Ticket #904. Hmm fix the interface_has_gateway() too.
Ticket #904. Actually correctly handle the assigned openvpn client as a dynamic gateway rather than breaking the behaviour of the system. Strange nobody has noticed broken gateway behaviour with openvpn assigned!
Actually was coorect before. 3rd parameter is length not index.
Revert "Correct this to make it actually work. This is also mentioned in Ticket #904 though it was already implemented."
This reverts commit 6f2cc3a680f984ccbb387301a26d022e6969e665.
Remove trailing carriage return
Correct this to make it actually work. This is also mentioned in Ticket #904 though it was already implemented.
Use a shell script rather than bad hack to execute php code for pppoe periodic reset.
Protect from strange situations on bootup by testing for is_array(). Do not add anymore the 127.0.0.2 route its not needed anymore. Also during bootup bring up all interfaces so the assignment process can deal with them(Possibly should be done in another code flow!).
Ticket #927. Increase timeout to gice mpd the time needed to exit gracefully.
Resolves #944. Actually bring down the vlan interface if it existed previously. This is a regression from the ppp dance/requests/whatever.
This one looks safe, and might improve speed. Add it back.
This ref should be safe too
Whoops this reference was good (for speed) and should stay.
More unneeded references.
Remove unneeded reference.
Remove support for this legacy code. The upgrade path handles this.
Restart dnsmasq first because dyndns might depend on it. Possibly these all should be sent through evnets to check_reload_status
Add as a search domain the domains suggested by dhcp.
Ticket #867. Flush the cache when restarting apinger and also when rc.newwanip is called. also restart apinger earlier in to be on latest address asap.
Use the new events mechanisms to dispatch events.
Do this check earlier rather than later, instead of moving the interface_bring_down call.
Revert "Rearrange function calls for wireless to prevent killing hostapd/wpa_supplicant at the wrong time."
This reverts commit e93385d06ddf0e2888b3f5ee5764992a19f99120.
Revert "Ticket #867. Do not destory interfaces other than preventing a loop for ppp interfaces during interface configuration."
This reverts commit 990fb7a9a9b03f744c56792011f1c1e2950ce764.
Protect against wpa_supplicant dying when bringing down interfaces to change wireless regulatory configuration.
Rearrange function calls for wireless to prevent killing hostapd/wpa_supplicant at the wrong time.
- Moved interface_wireless_configure call to after the new location of the interface_bring_down call to prevent it from killing hostapd and wpa_supplicant right after they start....
Ticket #843. Disable for now bringing down the interface unless we want to destroy it. While this is not all correct this seems to help alleviate problems when the parent is shared on cloned types.
Ticket #866 #632. Save old settings or actions to be taken for reconfiguring a route/vip on a tmp file and make use that information when apply settings buttons is clicked. This makes the gui behave as expected.
Ticket #867. Do not destory interfaces other than preventing a loop for ppp interfaces during interface configuration.
Correct binaries and find_dhclient_process
Make this more readble.
Ticket #860. Honor the parameter if passed.
Ticket #860. Correct code for proxyarp/choparp starting. Not sure why php did not bail at some code here missing curlies :).
Actually it should test the other way.
Ticket #846. Do not try to add an interface to the bridge during an event of LINK_UP.
Allow carp interfaces to sit on top of aliases. Also add safety belts to ipalias deletion for this to avoid breakage. Trigered-by: http://forum.pfsense.org/index.php/topic,27834.0.html
Fixes #847. Check even interface groups before allowing deletion of an interface. Also try to readd to a group an interface when configuring.
Revert "Use get_real_interface here and compare to that to make it more generic instead of doing this wireless-specific stuff." which probably ended up doing some recursion in various cases, especially slowing down as it goes down a larger list in this foreach loop....
Tiny change to boot message to remove redundant info and make it pretty.
Add run(4) to known wireless nics. Reported-by: http://forum.pfsense.org/index.php/topic,27744.0.html
Add bwn to interface lists too.
When booting up do not try to configure ovpn[c|s] interfaces if present also try to configure openvpn eariler just after the interfaces have been setup.