freifunk-franken
/
firmware
9
7
Fork 13
Code Issues 48 Pull Requests 7 Releases 17 Wiki Activity

Rework network configuration #209

Closed
fbl wants to merge 4 commits from fbl:network-overhaul into master
pull from: fbl:network-overhaul
merge into: freifunk-franken:master
Conversation 2 Commits 4 Files Changed 3 +149 -144
fbl commented 2022-02-27 01:26:41 +01:00
Owner
Copy Link

The configurenetwork script has been a big, hard to maintain monolith for a long time now, even though its functionality can be broken down into rather simple, easy-to-follow scripts. Also, due to some leftovers, it is rather slow and might also cause a race conditions (see #203).

This patchset aims to improve readabilty, maintainability and speed of our network configuration by removing manual changes to the kernel and separating static and dynamic configuration.

It might also fix the race condition mentioned in #203.

This patchset has been runtime-tested on a TP-Link WDR4900 using the node variant. One- and Two-Port devices, as well as devices using our DSA configuration still have to be tested.

Due to the rather weird dependencies of the layer3 variant to the base configuration established by the fff-network package, it should be tested for any regressions before applying the patchset.

The configurenetwork script has been a big, hard to maintain monolith for a long time now, even though its functionality can be broken down into rather simple, easy-to-follow scripts. Also, due to some leftovers, it is rather slow and might also cause a race conditions (see #203). This patchset aims to improve readabilty, maintainability and speed of our network configuration by removing manual changes to the kernel and separating static and dynamic configuration. It might also fix the race condition mentioned in #203. This patchset has been runtime-tested on a TP-Link WDR4900 using the node variant. One- and Two-Port devices, as well as devices using our DSA configuration still have to be tested. Due to the rather weird dependencies of the layer3 variant to the base configuration established by the fff-network package, it should be tested for any regressions before applying the patchset.
fbl added this to the next-bugfix milestone 2022-02-27 01:26:41 +01:00
fbl self-assigned this 2022-02-27 01:26:41 +01:00
fbl added 3 commits 2022-02-27 01:26:42 +01:00
1d3ed9953e fff-uradvd: Add reload trigger for network 
fff-uradvd depends on values stored in the uci network config.
Therefore, a reload is required if the uci network config is modified.

As fff-uradvd does not support reloads, and possibly changing values are
given as command parameters, do a restart on procd reload.

Signed-off-by: Fabian Bläse <fabian@blaese.de>
8cd45781a9 fff-network: Replace runtime configuration with reload_config 
Instead of both setting values in the uci configuration and applying
them manually, only set the values in the uci configuration and do a
combined reload_config afterward.

This not only ensures that the changes to the configuration are actually
correct, but also decreases the runtime of configurenetwork.
b66987e5bd fff-network: Remove unnecessary restart of fff-uradvd 
The explicit restat of fff-uradvd is not necessary anymore, as
appropriate procd reload triggers have been added, and reload_config is
executed after every run of configurenetwork.

Signed-off-by: Fabian Bläse <fabian@blaese.de>
rohammer reviewed 2022-02-27 21:35:19 +01:00
src/packages/fff/fff-network/files/etc/uci-defaults/22-network-base Outdated
@ -0,0 +49,4 @@
set network.bat0.fragmentation='1'
set network.bat0.orig_interval='1000'
set network.bat0.distributed_arp_table='1'
set network.bat0.hop_penalty='30'
rohammer commented 2022-02-27 21:35:19 +01:00
Member
Copy Link

Hi Fabian,
der Block kommt doch mit fff-batman-adv rein. Sollte nicht default sein. Oder moechtest du den auch in der l3 haben?
Robert

Hi Fabian, der Block kommt doch mit fff-batman-adv rein. Sollte nicht default sein. Oder moechtest du den auch in der l3 haben? Robert
fbl commented 2022-02-28 00:15:02 +01:00
Author
Owner
Copy Link

Ich möchte die l3 config langfristig komplett unabhängig von dieser Basiskonfiguration machen. Kurzfristig würde ich diesen Umstand einfach - wie bisher - hinnehmen.

Ich möchte die l3 config langfristig komplett unabhängig von dieser Basiskonfiguration machen. Kurzfristig würde ich diesen Umstand einfach - wie bisher - hinnehmen.
rohammer commented 2022-02-28 00:59:45 +01:00
Member
Copy Link

So ist es halt doppelt. Einmal hier und im package fff-batman-adv.

So ist es halt doppelt. Einmal hier und im package fff-batman-adv.
fbl commented 2022-02-28 15:45:14 +01:00
Author
Owner
Copy Link

Ah, an mir ist total vorbei gegangen, was du eigentlich meinst. Das war vorher ja auch nicht da und gehört hier entsprechend auch überhaupt nicht hin.

Ich hab da einen relativ alten Patch von mir wiederverwertet, möglicherweise ist da was übrig geblieben.

Würde ich entsprechend entfernen, gut aufgepasst!

Ah, an mir ist total vorbei gegangen, was du eigentlich meinst. Das war vorher ja auch nicht da und gehört hier entsprechend auch überhaupt nicht hin. Ich hab da einen relativ alten Patch von mir wiederverwertet, möglicherweise ist da was übrig geblieben. Würde ich entsprechend entfernen, gut aufgepasst!
fbl marked this conversation as resolved
rohammer reviewed 2022-02-27 21:47:09 +01:00
src/packages/fff/fff-network/files/etc/uci-defaults/22-network-base Outdated
@ -0,0 +143,4 @@
echo "MAC for client is set already"
else
echo "Fixing MAC on br-client"
sleep 10
rohammer commented 2022-02-27 21:47:09 +01:00
Member
Copy Link

Ich vermute, dass man hier das sleep nicht braucht. Im alten configurenetwork war da direkt davor ein network-restart. Auf den musste wohl gewartet werden. Hier gibt es nur ein uci commit.

Im Rest von configurenetwork ist auch noch so ein sleep uebrig, wo vorher ein network-restart war.

Ich vermute, dass man hier das sleep nicht braucht. Im alten configurenetwork war da direkt davor ein network-restart. Auf den musste wohl gewartet werden. Hier gibt es nur ein uci commit. Im Rest von configurenetwork ist auch noch so ein sleep uebrig, wo vorher ein network-restart war.
fbl commented 2022-02-28 00:16:31 +01:00
Author
Owner
Copy Link

Guter Punkt, der sollte nicht mehr nötig sein. Ebenso gilt dies für Fallunterscheidung, die kann komplett raus. Für das konfigurieren der fdff:: Adressen hatte ich das ja auch so gemacht, hier hab ich das dann wohl übersehen.

Guter Punkt, der sollte nicht mehr nötig sein. Ebenso gilt dies für Fallunterscheidung, die kann komplett raus. Für das konfigurieren der fdff:: Adressen hatte ich das ja auch so gemacht, hier hab ich das dann wohl übersehen.
fbl marked this conversation as resolved
fbl force-pushed network-overhaul from aa776ecbdf to ae0cfb7068 2022-02-28 17:28:14 +01:00 Compare
fbl commented 2022-02-28 17:28:28 +01:00
Author
Owner
Copy Link

Changes:

  • Remove bat0 interface configuration from 22-network-base (created in fff-batadv)
  • Remove unnecessary sleeps, uci commits and conditions for setting ETHMESHMAC and ROUTERMAC
Changes: - Remove bat0 interface configuration from 22-network-base (created in fff-batadv) - Remove unnecessary sleeps, uci commits and conditions for setting ETHMESHMAC and ROUTERMAC
rohammer approved these changes 2022-02-28 23:39:54 +01:00
rohammer left a comment
Member
Copy Link
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
``` Reviewed-by: Robert Langhammer <rlanghammer@web.de> ```
fbl modified the milestone from next-bugfix to 20220405-beta 2022-03-03 15:42:54 +01:00
fbl added a new dependency 2022-03-03 21:55:59 +01:00
#212 treewide: reorder layer3-specific uci-default scripts
fbl added a new dependency 2022-03-03 22:28:24 +01:00
#213 fff-network: remove hardcoded network config
fbl commented 2022-03-05 19:25:38 +01:00
Author
Owner
Copy Link

Auf meinen staging tree applied.

Auf meinen staging tree applied.
fbl closed this pull request 2022-03-05 19:25:38 +01:00
fbl deleted branch network-overhaul 2022-03-05 19:25:41 +01:00

Pull request closed

This pull request cannot be reopened because the branch was deleted.
Sign in to join this conversation.
Reviewers
No reviewers
rohammer
Labels
Clear labels   
RFC

   
RFT

   
WIP

   
blocked

   
bsp

touching bsp files or adding devices

  
bug

Something is not working

  
build/scripts/tools

touching buildscript or other tools

  
duplicate

This issue or pull request already exists

  
feature

New feature

  
fixed

Issue that got resolved

  
layer3

only for layer3 FW

  
mantis

Imported from mantis

  
more details required

Reported information is insufficient to debug the issue

  
needs changes

   
node

only for node FW

  
packages/fff

touching our packages

  
rejected

Rejected or won't fix

  
security

   
trivial

simple non-functional changes that are mostly matter of taste

  
upstream

touching upstream version, build or feed patches

No Label
RFC
RFT
WIP
blocked
bsp
bug
build/scripts/tools
duplicate
feature
fixed
layer3
mantis
more details required
needs changes
node
packages/fff
rejected
security
trivial
upstream
Milestone
Clear milestone
No items
No Milestone
20220405-beta
Assignees
Clear assignees
No Assignees
fbl
2 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Blocks
#212 treewide: reorder layer3-specific uci-default scripts
freifunk-franken/firmware
#213 fff-network: remove hardcoded network config
freifunk-franken/firmware
Reference: freifunk-franken/firmware#209
No description provided.
Delete Branch "fbl:network-overhaul"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?