lanefu changed the topic of #armbian-rockchip to: Armbian - Linux for ARM development boards | Rockchip SoC | www.armbian.com | This channel is relayed to the equivalent Discord channel | this channel is logged
montjoie has quit [Ping timeout: 272 seconds]
montjoie has joined #armbian-rockchip
DC-IRC has quit [Remote host closed the connection]
DC-IRC has joined #armbian-rockchip
<DC-IRC>
<Werner> I consider Nanopi R4S Bookworm minimal image **broken** since SSH does not properly start up. https://paste.armbian.com/utafukaram
<DC-IRC>
<Werner> From journalctl: `Jul 16 07:31:01 nanopi-r4s sshd[1589]: sshd: no hostkeys available -- exiting`
<DC-IRC>
<Werner> However there ARE keys available (`/etc/ssh/ssh_host_*`).
<DC-IRC>
<Werner> Anyway deleting them and using `dpkg-reconfigure openssh-server` creates a new set which then allows ssh to start properly.
<DC-IRC>
<igorpec> i'll recreate those images
<DC-IRC>
<Werner> There seems to be an issue with overlays as well
<DC-IRC>
<Werner> Hm...is this additional `rockchip-` in the file name too much? I think prefix defines the folder or is this actually correct as it is?
<DC-IRC>
<Werner> (the 3rd line I just assembled the path to the file for comparison and is not present in the armbianenv.txt)
troglodito has quit [*.net *.split]
<DC-IRC>
<c0rnelius77> I'm fairly certain that overclocking overlays have been broken for a while on 6.1 and up. It isn't just a rockchip specific thing. I stopped using them in my personal builds and moved to using turbo-mode along with a service to OC the boards.
<DC-IRC>
<Tonymac32> Hmm
<DC-IRC>
<Tonymac32> Turbo requires a proper opp table though as well
<DC-IRC>
<c0rnelius77> yeap. just add it to the table and there is waits. ```
<DC-IRC>
<c0rnelius77> Doing it this way also doesn't require a reboot and can just be done on the fly.
<DC-IRC>
<Tonymac32> Sure, I had the TB set up that way for a while, but inherently the overlays should work, something just got shifted in the dts or perhaps the max speed set by the Armbian configs is wrong thanks to adding some more dissimilar hardware to the mix and not checking the effects of the variables changed. Haven't looked