OpenWrt 25.12.0-rc4 - Release Candidate

Yes indeed

3 Likes

Great solution, mate! Though I’ll wait for a patch since I’m a bit lazy and not much of a coding guy. I’m primarily running Netduma, just keeping an eye on how the OpenWrt changes are progressing and checking for bugs, etc.

1 Like

Tried today on Edgerouter X.

It failed miserably…. dnsmasq complained of problems in ā€œ/var/etc/dnsmasq.conf.cfg01411cā€ file (created by himself I think, cause never seen that file previously) and hangup.

I had to reflash 24.10.5 and everything ok.

little note:

since r1 there seems to be issue with dnsmasq and untill now r4
some devices, their firmware downloaded without even dnsmasq itself
then no dhcp acquired over wan, no dhcp served on lan neither, if on exroot with cheap usb and doing upgrade, some ghost file make it ā€œseemsā€ to work but in fact, some core parts missing.

for two same devices one firmware download can be with and the other without dnsmasq included!

happen to me on various gl-mt3000, r7000, zyxel nwa 50 ax, some tp-link and other netgear devices
same firmware, custom build, for each specific devices, some have and some does not have dnsmasq
scratch my head before noticing the issue as 2 first time it happen to me was with devices with exroot

3 arcadian upgrade all downloaded with the dnsmasq though.

beyond that, the release is very stable

Thank you all dev!
:+1:

1 Like

Ok thanks you

In my case dnsmasq seemed to be present, but faulty…

TP-Link Archer C20, works.

I filed https://github.com/openwrt/openwrt/issues/21880 for the issue of upgrades of the MR52 within 25.12-branch builds causing the ethernet devices to get reversed.

2 Likes

ER605v2 (mt7621) updated from 24.10.5 to 25.12.0-rc4, no issues.

DAP-2610 (ipq40xx), working, but no scan in 5 Ggz. band, always getting the same channel.

Is Rockchip (my Nanopi R4S) expected to be stabilised in the 25.12 final release?

Just buildbot barfed, no problem per se.

Yes, in fact it's already in 25.12-snapshot and main snapshot, it was just the -rc4 where that target was broken.

OpenWrt 25.12.0 using kernel 6.12? Happy days!

6.12 brings Intel Thread Detector support, this matters a lot for me, I have a proxmox server with Intel Pentium Gold 8505 (1P+4E cores).

I believe it is finally time to try OpenWrt on x86.

Any feedback on how QEMU agent and igc network drivers are behaving in the latest OpenWrt releases?

2 Likes

@RHBH Yea 6.12 there is also tickless kernel support for x86-64 which went in to main over a year ago but I don’t think was ever backported to 24.10. So you should find x64 in good shape for 25.12 from what I’ve read. Give a try and please post your experiences when you can. (I only use a Filogic 830 device now so haven’t tested myself.)

@CryptoWizard if you are looking for a new build but don’t want to wait for rc5 you can grab a 25.12-snapshot rockchip is back being built on those. Just do that with caution as snapshots are moving targets day to day.

2 Likes

Someone added it as an open issue: https://github.com/openwrt/openwrt/issues/21876

1 Like

I’m testing OpenWrt 25.12.0-rc4 on my D-Link DIR-878 A1, and a cron job to reboot the router daily is not working.

The cron entry I’m using is:

30 4 * * * sleep 70 && touch /etc/banner && reboot

Ref https://pubs.opengroup.org/onlinepubs/9699919799/utilities/crontab.html#tag_20_25_07
ie check `sh -c "sleep .... echo reboot"

This works when tested directly in the terminal:

sh -c "sleep 70 && touch /etc/banner && /sbin/reboot"

However, this does not work when added directly in LuCI as a cron job:

19 5 * * * sh -c 'sleep 70 && touch /etc/banner && /sbin/reboot'

The router does not reboot when the cron job runs.

remove sh -c from there - it uses sh -c to parse the line.

So, I went back to LuCI and added the cron job again like this, but it still did not work:

42 5 * * * sleep 70 && touch /etc/banner && /sbin/reboot

The cron job did not run and the router did not reboot.