thank you very much, I will flash your firmware ASAP and let you know the result
edit: I see that this is "sysupgrade.bin" kind of firmware so I suppose I'll have to flash first factory.bin from any LEDE and only after that yours;
I am currently on stock firmware that's why I'm asking
A reminder, in my build I've replaced dnsmasq with unbound, and the DNS server used is quad9, you can change the server in: /etc/unbound/unbound_ext.conf . The commented lines are for enabling DNS over TLS, you can enable it uncommenting the lines and changing the quad9 port to 853, but DNS over TLS is slow with unbound.
Factory bin: https://drive.google.com/open?id=1k6_oZhC9e8qv9wsbYolFBFkjA_2wu8hQ
There are 3 factory builds:
factory.bin
factory-eu.bin
factory-us.bin
I've uploaded the factory.bin, if you need the eu or us just tell me
thanks again for your quick reply; I've downloaded it, but is there any difference between this one and the EU one? I live in EU, is it something that has to do with regulatory domains? I was using channel 149 and region US as reg domain to be able to squeeze out more power on the AP side
I think is related to the change that tp-link made to the bootloader to prevent flashing unauthorized OSes.
Edit: from the openwrt source:
**TP-Link has started providing US- and EU-specific stock firmwares that only
allow upgrading to firmwares with the same region code. Provide factory
images for both these regions.
To avoid confusing users outside these regions, we still provide a
"universal" factory.bin without a region code, although flashing either of
the US and EU images would work as well.**
factory-eu: https://drive.google.com/open?id=1x4OSkemzqvdDKxTzJzn7aNCEmbxb6mTV
@cliobrando, I flashed your firmware, but unfortunatelly I had no way of actually using it;
it obtained an IP on the WAN side but I had no DHCP leases active and no internet on clients side;
besides all these, Luci interface was so slow that I had to wait 15 seconds (delay) between toggling through the different menu interfaces;
anyway thanks for your efforts
reverted to last @r00t's nightly for some testing, download speeds are great but the LAG issue still remains...
@cliobrando, I post here the logread command's result as you instructed me a few days ago
this is with @r00t's last nightly;
I see a lot of DHCPV6 requests, but my ISP is sending only V4
edit: the board won't let me post so many characters, so I've uploaded the whole logread file here so maybe you can take a look :
I will paste here just a small part that I have encountered right now when trying to reach router's own webinterface (192.168.1.1) and I had a serious lag when did so
Tue Feb 6 08:22:32 2018 daemon.notice hostapd: wlan1: AP-STA-POLL-OK 04:4b:ed:17:d5:e0
Tue Feb 6 08:25:24 2018 daemon.info odhcpd[1559]: Using a RA lifetime of 0 seconds on br-lan
Tue Feb 6 08:25:28 2018 daemon.info odhcpd[1559]: Using a RA lifetime of 0 seconds on br-lan
Tue Feb 6 08:25:31 2018 daemon.info dnsmasq-dhcp[4101]: DHCPDISCOVER(br-lan) 74:d4:35:81:4e:ec
Tue Feb 6 08:25:31 2018 daemon.info dnsmasq-dhcp[4101]: DHCPOFFER(br-lan) 192.168.1.119 74:d4:35:81:4e:ec
Tue Feb 6 08:25:31 2018 daemon.info dnsmasq-dhcp[4101]: DHCPREQUEST(br-lan) 192.168.1.119 74:d4:35:81:4e:ec
Tue Feb 6 08:25:31 2018 daemon.info dnsmasq-dhcp[4101]: DHCPACK(br-lan) 192.168.1.119 74:d4:35:81:4e:ec user-PC
Tue Feb 6 08:25:32 2018 daemon.info odhcpd[1559]: Using a RA lifetime of 0 seconds on br-lan
Tue Feb 6 08:25:37 2018 daemon.notice odhcpd[1559]: Got DHCPv6 request
Tue Feb 6 08:25:37 2018 daemon.warn odhcpd[1559]: DHCPV6 SOLICIT IA_NA from 0001000120583e6d74d435814eec on br-lan: ok fdde:a959:54d5::4e2/128
Tue Feb 6 08:25:38 2018 daemon.notice odhcpd[1559]: Got DHCPv6 request
Tue Feb 6 08:25:38 2018 daemon.warn odhcpd[1559]: DHCPV6 SOLICIT IA_NA from 0001000120583e6d74d435814eec on br-lan: ok fdde:a959:54d5::4e2/128
Tue Feb 6 08:25:40 2018 daemon.notice odhcpd[1559]: Got DHCPv6 request
Tue Feb 6 08:25:40 2018 daemon.warn odhcpd[1559]: DHCPV6 SOLICIT IA_NA from 0001000120583e6d74d435814eec on br-lan: ok fdde:a959:54d5::4e2/128
Tue Feb 6 08:25:44 2018 daemon.notice odhcpd[1559]: Got DHCPv6 request
Tue Feb 6 08:25:44 2018 daemon.warn odhcpd[1559]: DHCPV6 SOLICIT IA_NA from 0001000120583e6d74d435814eec on br-lan: ok fdde:a959:54d5::4e2/128
Tue Feb 6 08:25:52 2018 daemon.notice odhcpd[1559]: Got DHCPv6 request
Tue Feb 6 08:25:52 2018 daemon.warn odhcpd[1559]: DHCPV6 SOLICIT IA_NA from 0001000120583e6d74d435814eec on br-lan: ok fdde:a959:54d5::4e2/128
Tue Feb 6 08:26:08 2018 daemon.notice odhcpd[1559]: Got DHCPv6 request
Tue Feb 6 08:26:08 2018 daemon.warn odhcpd[1559]: DHCPV6 SOLICIT IA_NA from 0001000120583e6d74d435814eec on br-lan: ok fdde:a959:54d5::4e2/128
Tue Feb 6 08:26:11 2018 daemon.info dnsmasq-dhcp[4101]: DHCPINFORM(br-lan) 192.168.1.119 74:d4:35:81:4e:ec
Tue Feb 6 08:26:11 2018 daemon.info dnsmasq-dhcp[4101]: DHCPACK(br-lan) 192.168.1.119 74:d4:35:81:4e:ec user-PC
edit2 : I have disabled IPv6 in network interface, but LAG still remains, I post a picture of my DHCP settings
Can you run htop in the console when the lag is ocurring?
I was reading the log, and I saw that you are using RADIUS to auth your clients, do you tried to use the router without RADIUS? just to test?
thanks for the hint, can you tell me how to turn this RADIUS off?
edit: meanwhile I turned off DHCPv6 and all IPv6 protocol in the Network configuration and also unticked "ignore resolv file" from the "resolv and hosts file" menu but unfortunatelly the LAG persist like before
This is my configuration:
everything is the same on my side also, only difference is the transmit power is set at 30 dbm and I cannot untick "enable legacy rates"
mmmm.. I don't know what can it be, there are messages in your log that doesn't appear in my log, like this:
"daemon.info procd: Instance sysntpd::instance1 s in a crash loop 6 crashes"
And messages about using RADIUS to auth your wireless clients (RADIUS is mainly used on enterprise networks).
I think that something is misconfigured, I never had a problem with the @r00t builds, but I prefer to do my own build because I don't like dnscrypt-proxy.
here is the htop command's print screens, I couldn't copy and paste
here is also my wifi settings page
problem is, that even if I reset to default values, or just after I flash a new version of LEDE, even if I connect with the default config without any wireless security, the lag is still there and I cannot acces certain pages because the message "cannot reach.try again later" comes up, or the webpage comes up but with a big delay
and like you see another user's comment up there, I am not the only one with this lag scenario
It's really weird, have you tried using the router with the wan cable disconnected?
but then I would have no internet acces, so what to test? only the luci interface responsiveness?
meanwhile, trying to acces Luci ...
this never happens witrh stock or dd-wrt, not even with openwrt or gargoyle, just with this firmware
yes, I just wanted to know if the LUCI interface was responsive without wan connection.
Hey @cliobrando! Using the most recent build you posted; works great. Quick question... does your build have all of the same patches at @root's build? I'm specifically interested in the FastPath patches?
@docbanzai
Hi, yes, I just made little changes to the patches (FastPath is enabled).
Changes to the packages: I think that more components can be removed from the firmware like SQM, because FastPath is not compatible with it (you can have one enabled, but not both). I have a download speed of 20MB/s and if I enable SQM to reduce bufferbloat my bandwidth is reduced by 3 - 5MB/s so is not an option.
I added haveged to increase entropy.
Removed DNSCrypt and replaced it with Unbound + odhcp to use DNS over TLS but unbound isn't optimized so I just disabled the SSL Forward and use unbound as a caching dns forwarder, now that Stubby + getdns is in the LEDE trunk I will replace unbound with Stubby (which works perfect with DNS over TLS).
Removed a bunch of drivers that are not needed and more stuff like that, but I think that my build is not much different from r00t build, just different package selection.
Great; this is pretty much exactly what I was looking for. I'm also happy to report that I'm not experiencing the Luci slowness / unresponsiveness that seems to be present in @root's most recent builds. Everything seems very zippy. Thanks!
I forgot,
An advice to everyone that replaced u-boot with Breed:
After replacing the bootloader, turn off your router, hold the reset button, and turn on your router with the reset button pressed, wait 5 seconds and release the button, then go to your PC and assign a manual IP to your wired NIC from the range 192.168.1.X (don't use 192.168.1.1 because is the router IP). Open a browser and go to http://192.168.1.1, and replace the FF:FF:FF:FF:FF:FF MAC with the one in your router sticker (see the picture attached), well you can put a random MAC, but if you don't change the default value (FF:FF:FF:FF:FF:FF) everytime your router start random macs will be generated for every interface (and that can be problematic in some cases).
My ISP blocked me because it detected that multiple devices were connected to the modem (my ISP allows 4 devices MAX) so everytime I restarted the router a new WAN MAC was generated and my ISP thought that it was a new device.
@cliobrando
You can already do this with any build of openwrt. Simply release the IP from your ISP, generate a unicast MAC address to set for WAN in network config, and restart network init script. This can often result in a new IP being leased to you as well. In my case, awhile ago I caught my ISP trying to throttle me during peak hours based on my router's MAC address, so I tend to change it every so often. I wrote a simple script that assists with the process.
