OpenWrt 25.12.5 - Service Release

Binaries of new odhcp6c were released yesterday. Just upgrade this one package. Those installing 25.12.5 via attended sysupgrade should get fixed version automatically.

3 Likes

Can't build an image using Firmware Selector. Is it broken for now? Is there a topic I can subscribe to see when it's fixed?

Some else had problems after installing wget?
Installing wget works fine

root@Huggins:~# apk add wget
(1/2) Installing libpcre2 (10.47-r1)
  Executing libpcre2-10.47-r1.post-install
(2/2) Installing wget-nossl (1.25.0-r2)
  Executing wget-nossl-1.25.0-r2.post-install
  * add alternative: /usr/bin/wget -> /usr/libexec/wget-nossl
OK: 49.8 MiB in 275 packages

but after installing I'm unable to update package cache or install new packages

root@Huggins:~# apk update
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/targets/mediatek/filogic/packages/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/base/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/targets/mediatek/filogic/kmods/6.12.94-1-5a6c1f71be683ae9980b15d3ce73e24d/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/luci/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/packages/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/routing/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/telephony/packages.adb: unexpected end of file
ERROR: wget: exited with error 1
WARNING: updating https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/video/packages.adb: unexpected end of file
 [https://downloads.openwrt.org/releases/25.12.5/targets/mediatek/filogic/packages/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/base/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/targets/mediatek/filogic/kmods/6.12.94-1-5a6c1f71be683ae9980b15d3ce73e24d/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/luci/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/packages/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/routing/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/telephony/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/video/packages.adb]
0 unavailable, 8 stale; 11094 distinct packages available

after removing wget all went back to normal

root@Huggins:~# apk del wget
(1/2) Purging wget-nossl (1.25.0-r2)
  Executing wget-nossl-1.25.0-r2.pre-deinstall
  * add alternative: /usr/bin/wget -> /bin/uclient-fetch
(2/2) Purging libpcre2 (10.47-r1)
  Executing libpcre2-10.47-r1.pre-deinstall
OK: 49.0 MiB in 273 packages
root@Huggins:~# apk update
 [https://downloads.openwrt.org/releases/25.12.5/targets/mediatek/filogic/packages/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/base/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/targets/mediatek/filogic/kmods/6.12.94-1-5a6c1f71be683ae9980b15d3ce73e24d/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/luci/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/packages/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/routing/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/telephony/packages.adb]
 [https://downloads.openwrt.org/releases/25.12.5/packages/aarch64_cortex-a53/video/packages.adb]
OK: 11094 distinct packages available

My Device: GL.iNet GL-MT6000


Solution: it was a bit of my mistake, install "wget" installing wget-nossl - which case the problems, install explicit wget-ssl did the trick

2 Likes

Thanks for your feedback, I did it, worked! :handshake:

Does OpenWrt do a similar aproach like TrueNAS? I only need to install releases for security and not individual patches?

Installed 25.12.5 on a pair of TP-Link Archer AX21 (v4)s, which are newly fully supported in this release. One using sysupgrade.bin image upgrading from a prior 25.12-SNAPSHOT build and one using factory.bin from the stock TP-Link firmware.

Both are working fine. I did have a problem with DNS forwarding, but that turned out to be due to a configuration error on my part that prior version of dnsmasq apparently handled differently.

In one sense, OpenWrt is even more monolithic and "appliance-like" than TrueNAS. Each release has the kernel, kmods and a subset of packages built specifically together and they will never be changed; you must do a full upgrade to get any changes made to these between releases. There are no "hot fixes" or patches or whatnot in OpenWrt, it's an all-or-nothing approach.

On the other hand, OpenWrt's package feeds are constantly updating, so if there's a security fix in something like wget or adblock, then there's no need to jump a version to get it (although the safest way to install any of these is to do an "upgrade to the same version").

As to when you should perform an upgrade, well, that's a matter of choice. For my production devices, I only upgrade when a new dot release comes out; they're all at 25.12.5 now and won't be re-flashed until 25.12.6 comes out; when 26.x gets to RC status, I'll wait for 26.x.0 to be released before I upgrade. So, for my edge facing router, it's always on latest supported release, with whatever packages were available at the time of that release, I never use any -SNAPSHOTs or -RCs for them.

For my test devices, I'm all over the place, I've got 23.05, 24.10, 25.12 and main SNAPSHOT running on something somewhere nearby...

9 Likes

First, thanks as always to the OpenWRT devs for such an excellent job.

Regarding IPv6 prefix delegation, I also observed with my provider (Verizon FIOS) that after upgrade 25.12.4 --> 25.12.5 that an IPv6 was no longer provisioned. However, after waiting for the IPv6 lease timeout (ca. 2 hr for FiOS), 25.12.5 was then able to pickup IPv6 for prefix delegation as before.

I didn't save the pcap files, but believe the difference came from the scripts change to the IA_PD value (which in 25.12.4 was different than IA_NA value, but in 25.12.5 is the same by https://github.com/openwrt/openwrt/commit/7029a38c93637613385f52539a391c438e8021dd ), which in turn caused the ISP to not release a delegated prefix until the prior lease (to a different IA_NA) expired.

Note: though I didn't test it here, from prior similar things, on some ISPs turning off power to the ONT for a period of time is sufficient to reset things even before the lease time is up.

4 Likes

Has anyone heard anything?

https://x.com/hackerfantastic/status/2074871544984064163 - this is the way to post things like that. People at least can have date, context...

P.S. Are you the guy behind that Twitter account?

3 Likes

@hauke can you confirm that information from @zeropi2021 ?

1 Like

My mistake, I apologize, I'll remember that in the future.

No, sir. I have nothing to do with this particular account on X.

Generally speaking, the manner and statements of this person and declarations do not convince me and normally I would rather consider it BS, but since this person also mentioned OpenWRT, I thought I would ask if anyone has heard anything...

But there is a very good chance that this is some kind of troll post or other AI with claims about the holy grail.

4 Likes

This person also wrote it in that thread...

I can confirm, that for Deutsche Glasfaser after 3h 15 minutes (router's uptime, so maybe less than that) I can receive a PD as well with restarting wan6. Thanks for posting your solution @tmcqueen-materials and big kudos to the developers!

Installed smoothly on GL.inet GL-MT6000, thanks again to all the devs

1 Like

25.12.5 running on a Netgear WNDR3700 v4 can't connect to some IOT devices. My Thermomaven G2-0R0A could connect just fine in 25.12.4 but not on in 25.12.5 . I even created a separate SSID just for the ESP32 devices using 2.4 Ghz only and even switched from 802.11 N to "legacy", but still no connection. Some of my older ESP32 smart plugs connect fine but a newer replacement smart plug won't connect. The issue isn't limited to a single device from a single manufacturer.

As another data point, I'm running main SNAPSHOT on my Quantum Fiber W1700K's and the Thermomaven couldn't connect to that either.

Disable 802.11w Management Frame Protection and Operating Channel Validation in Wireless Security tab.

Or... share your wireless config.

1 Like

Thank you! That was it. Management Frame Protection was gumming up the works.

1 Like

Not very important, but days later I just noticed that a 10/100 device during the router update went to 10, instead of being re-negotiated back up to 100.

Taking the link down and back up on the device auto-negotiated it back to up to 100.

So I guess look at the switch page to see what speeds are set for each port just in case.

Hi everyone,

Life(TM) has kept me busier than usual as of late, so this has taken longer: when trying to use 25.12.5 to test the Ansible collection community.openwrt it seems that the docker images (x86_64 at least) for it were not published.

Currently https://github.com/ansible-collections/community.openwrt/pull/246 is failing when trying to pull the image (from ghcr.io).

Running docker pull ghcr.io/openwrt/rootfs:x86_64-25.12.5 locally is giving me the same "manifest unknown" results also from quay.io and from docker hub itself.

Is that just a glitch in the publishing process for this version or was that on purpose? Thanks