| Message ID | 20220504090423.13896-1-hacks@slashdirt.org |
|---|---|
| State | New |
| Headers | show |
| Series | ath79: consistent network configuration of 2-port APs | expand |
The sender domain has a DMARC Reject/Quarantine policy which disallows sending mailing list messages using the original "From" header. To mitigate this problem, the original message has been wrapped automatically by the mailing list software. Em 04/05/2022 06:04, Thibaut VARÈNE escreveu: > In ath79 we have two groups of indoor/outdoor 2-port APs: one where > both ports are assigned to the lan interface, and one where an > arbitrarily chosen port is assigned to the wan interface. > > From a look at the device pictures, it seems none of them make that > distinction on their casing, so this patch moves the second group in > line with the first. IMO, this is going to expose anyone who upgrades into a firmware that has these changes [possibly after a factory-reset] to something that is very dangerous[1]. So, IMHO, this sort of change *MUST* be listed in the next stable RELEASE NOTES as an "incompatible change". Also IMO, the only way for this to work is to have an in-tree list of what-needs-to-be-in-next-RELEASE_NOTES, and enforce that any such changes landing in the tree also update that file (and, when desired/required, bump the sysupgrade image config version guard). Note that I am not opposing the change. I just think we should at some point (preferably now) ensure that such changes to be less of a pitfall to downstream and users. [1] as in: when it goes from WAN to LAN, a port that is *physically* connected to the WAN circuit suddenly becomes a bridge, and connects that WAN circuit to the LAN circuit. And yes, obviously applies to VLANs as well. Switching a port from LAN to WAN is also bad, and it can be just as dangerous depending on several factors.
diff --git a/target/linux/ath79/generic/base-files/etc/board.d/02_network b/target/linux/ath79/generic/base-files/etc/board.d/02_network index a4eb2ae06b..dede15a924 100644 --- a/target/linux/ath79/generic/base-files/etc/board.d/02_network +++ b/target/linux/ath79/generic/base-files/etc/board.d/02_network @@ -124,11 +124,32 @@ ath79_setup_interfaces() ;; alfa-network,n2q|\ alfa-network,n5q|\ + comfast,cf-e110n-v2|\ + comfast,cf-e120a-v3|\ + comfast,cf-e314n-v2|\ + compex,wpj531-16m|\ devolo,dvl1200e|\ devolo,dvl1750e|\ engenius,enstationac-v1|\ engenius,ews511ap|\ ocedo,ursus|\ + openmesh,a40|\ + openmesh,a60|\ + openmesh,om2p-v1|\ + openmesh,om2p-v4|\ + openmesh,om2p-hs-v4|\ + plasmacloud,pa300|\ + plasmacloud,pa300e|\ + tplink,cpe210-v1|\ + tplink,cpe220-v2|\ + tplink,cpe220-v3|\ + tplink,cpe510-v1|\ + tplink,wbs210-v1|\ + tplink,wbs210-v2|\ + tplink,wbs510-v1|\ + tplink,wbs510-v2|\ + ubnt,nanostation-m|\ + ubnt,routerstation|\ ubnt,unifi-ap-outdoor-plus) ucidef_set_interface_lan "eth0 eth1" ;; @@ -193,29 +214,6 @@ ath79_setup_interfaces() ucidef_add_switch "switch0" \ "0@eth0" "1:lan:1" "3:lan:4" "4:lan:3" "5:lan:2" "2:wan" ;; - comfast,cf-e110n-v2|\ - comfast,cf-e120a-v3|\ - comfast,cf-e314n-v2|\ - compex,wpj531-16m|\ - openmesh,a40|\ - openmesh,a60|\ - openmesh,om2p-v1|\ - openmesh,om2p-v4|\ - openmesh,om2p-hs-v4|\ - plasmacloud,pa300|\ - plasmacloud,pa300e|\ - tplink,cpe210-v1|\ - tplink,cpe220-v2|\ - tplink,cpe220-v3|\ - tplink,cpe510-v1|\ - tplink,wbs210-v1|\ - tplink,wbs210-v2|\ - tplink,wbs510-v1|\ - tplink,wbs510-v2|\ - ubnt,nanostation-m|\ - ubnt,routerstation) - ucidef_set_interfaces_lan_wan "eth1" "eth0" - ;; comfast,cf-e375ac) ucidef_add_switch "switch0" \ "0@eth0" "1:wan" "2:lan"
In ath79 we have two groups of indoor/outdoor 2-port APs: one where both ports are assigned to the lan interface, and one where an arbitrarily chosen port is assigned to the wan interface. From a look at the device pictures, it seems none of them make that distinction on their casing, so this patch moves the second group in line with the first. Signed-off-by: Thibaut VARÈNE <hacks@slashdirt.org> --- .../generic/base-files/etc/board.d/02_network | 44 +++++++++---------- 1 file changed, 21 insertions(+), 23 deletions(-)