Well, as a typical home router sits most of the time rather idle, so running the CPU all the time at 1700 MHz instead of 384 MHz seems wasteful and causes higher thermal load on the device.
No it does not, at least in my case. The temperature did not change at all.
What SSH command would I use to set the governor to performance? I tried looking it up but have been unsuccessful.
Some hints below:
root@Router1:~# ls /sys/devices/system/cpu/cpufreq/
ondemand policy0 policy1
root@Router1:~# ls /sys/devices/system/cpu/cpufreq/policy0/
affected_cpus related_cpus scaling_governor
cpuinfo_cur_freq scaling_available_frequencies scaling_max_freq
cpuinfo_max_freq scaling_available_governors scaling_min_freq
cpuinfo_min_freq scaling_cur_freq scaling_setspeed
cpuinfo_transition_latency scaling_driver stats
root@Router1:~# cat /sys/devices/system/cpu/cpufreq/policy0/scaling_governor
ondemand
root@Router1:~# cat /sys/devices/system/cpu/cpufreq/policy0/scaling_available_governors
ondemand performance
root@Router1:~# echo performance > /sys/devices/system/cpu/cpufreq/policy0/scaling_governor
root@Router1:~# cat /sys/devices/system/cpu/cpufreq/policy0/scaling_governor
performance
(and naturally the same for both cores 0 and 1)
Very cool thanks. Wishing my WRT32X had CPU scaling for heat/power efficiency too as it would probably save a couple watts. Right out of the box (Linksys' stock firmware based off OpenWrt) during boot says:
[ 0.090030] cpuidle: using governor ladder
[ 0.092202] mvebu-pmsu: CPU idle is currently broken on Armada 38x: disabling
You can test my version of the mvebu CPU frequency scaling driver:
1 Like
This one only scales cpu though, and not the voltages. R7800 is just a lot better for that.
Thank you very much for the instructions. Real transfers in my specific case have increased after wifi from 20MB / s to 30MB / s (link 866), and after a cable with about 40MB / s to 60MB / s it is increasing the real bandwidth by 50%. I am very happy with the results. In the CPU clock statistics I finally see twice after 1.7GHz. Thank you very much again. PS. we are slowly approaching the parameters of the netgear software 
FreeNAS and UnRaid. I use UnRaid and have been happy for almost 10 years now 
wow.....
I have my setup perfect for me but am curious about going forward which file type to just do an upgrade instead of a full blown install. Here is my version. How would I just update my current setup? Thanks
OpenWrt 18.06-SNAPSHOT r7157-8a0ad2660a / LuCI openwrt-18.06 branch (git-18.198.67944-94b6aa2)
Just a normal sysupgrade with a sysupgrade image. Nothing special needed.
1 Like
Thank you!
Got this router a week ago to replace my trusty Asus RT-N66u which ran Merlins FW.
Not to say the least but was not happy with the FW from Netgear.
I flashed the R7800-owrt1806-r7190-ce234299bc-20180730-2012-sqfs-factory.img and it was a bit of a challenge to find all the settings etc ... (yes, I've used dd-wrt, tomato etc in the wrt54 age)
After 2 days running I got complains that my 5GHz dropped out. In the logs I see this ... unfortunately it seems that I can't copy paste from the log so typing is the way to go. WLAN0 disconnected due toexcessive missing ACKs
Already fixed bug that only affects the default Bootstrap theme. You can install update luci theme from the download repo.
opkg update
opkg install luci-theme-bootstrap
(or you can switch to Material theme that is also installed in my build)
Wow that was fast.
Not sure if this is something, Should have gotten the time it happened
Wed Aug 1 19:50:03 2018 daemon.info dnsmasq-dhcp[23954]: DHCPACK(br-lan) 192.168.1.236 24:92:0e:26:c4:14 Galaxy-A5-2017
Wed Aug 1 19:50:08 2018 daemon.info hostapd: wlan0: STA f8:16:54:56:05:db IEEE 802.11: disconnected due to excessive missing ACKs
Wed Aug 1 19:50:10 2018 kern.warn kernel: [97069.480204] ath10k_pci 0000:01:00.0: Invalid peer id 51 peer stats buffer
hmm...router failed to come up after sysupgrade to r7729-f960490fc8, got stuck on flashing white-power light.
after a power-cycle it seemed to crash again (it had flashing white power, then went to orange power after a bit), then it came up properly. is there some persistent boot log where i can work out what went wrong? i've had this happen a couple of times before, after a sysupgrade. this is since master moved to kernel 4.14.x, before that i never had any problems.
In the script saveHNsettings.sh you put back your config.
May I ask what these files contain for which special settings?
config/dhcp (is it used for open dns?)
config/bcp38 (?)
/etc/config/dhcp contains the configuration for dnsmasq and odhcpd, you will find this on every OpenWrt device.
/etc/config/bcp38 contains the configuration for bcp38, a tool that helps you reject packets claiming to come from non-globally routable IP(v4)s from the WAN interface, without looking any further; see IETF BCP38.
There is nothing specific to hnyman's build with these files (other than the semi-automated fashion to restore his preferred configuration from an encrypted archive).
1 Like
Nothing really special. Just normal personal wifi settings, firewall settings, ssh keys etc.
Like slh said, the script just helps me to expand my settings from the encrypted archive file (/etc/HNsettings...) easily. I can't leave the settings openly there for other users of my build, and also do not want to e.g. set SQM enabled in its config for all users. Having the key settings stored that way on the flash also enables me to rather quickly recover from possible bricks.
There is no persistent log.
Boot problems happen every now and then. Have also happened earlier, I think.
The can be a bad flashing, or weak flash memory block, causing intermittent read errors.