Joined: 13 Mar 2014 Posts: 856 Location: Montreal, QC
Posted: Mon Apr 04, 2016 0:01 Post subject:
There was an update to the script that resolved an issue regarding restarting dnsmasq as ddwrt dnsmasq service now uses additional command line switches. I'll try and upload the update tonight or tomorrow. FYI the update has been tested on ddwrt versions up to v28800 but should work fine on current versions
A future update will remove portions of the script which are no longer needed.
Joined: 10 Jan 2015 Posts: 270 Location: Minnesota
Posted: Tue Apr 05, 2016 22:03 Post subject:
James, i've been using a piece of your simple IPv6 code you posted some while back and many thanks for providing this.
(In Services tab, under DNSMasq options)
dhcp-range=::1000,::FFFF,constructor:br0,ra-stateless,300
enable-ra
quiet-ra
quiet-dhcp
quiet-dhcp6
This has given me perfect IPv6 functionality with my Comcast connection. Until builds with kernel 4.4.2+ that is, when QoS is enabled... I currently went back to build 29085 in the meantime, but I wanted to ask either of you if you have had working IPv6 with QoS enabled using any build after 29085? _________________ LATEST DD-WRT FW IS LOCATED HERE: https://dd-wrt.com/support/other-downloads/?path=betas%2F
Joined: 13 Mar 2014 Posts: 856 Location: Montreal, QC
Posted: Wed Apr 06, 2016 18:58 Post subject:
@ MDA400
Unfortunately all routers I have configured with ipv6 and qos are running ~v28600 and I've been holding off upgrading until kong releases a stable k4 build
Is it by design, or a bug, that DD-WRT restarts the router whenever you change IPv6 settings?
Usually my IPv6 config works on the first cold boot but becomes hit or miss after that. I'm running Kong 29230 on an R7000 with Entware and YAMon but I've noticed this for some time now.
Is it by design, or a bug, that DD-WRT restarts the router whenever you change IPv6 settings?
Usually my IPv6 config works on the first cold boot but becomes hit or miss after that. I'm running Kong 29230 on an R7000 with Entware and YAMon but I've noticed this for some time now.
No this is not by design, but most likely a result of installing something (entware) that is by design not compatible with dd-wrt: Entware compiles packages with a toolchain that is not compatible with dd-wrt. This can cause segfaults and reboots.
My latest build includes 'bootstrap', that allows you to prepare for opkg and then install compatible and signed packages. _________________ KONG PB's: http://www.desipro.de/ddwrt/
KONG Info: http://tips.desipro.de/
I removed entware (or at least everything entware associated I could find) and still had the issue.
But rebooting and modifying IPv6 without my usb attached (/opt mounted) allowed modification with no reboot so there must be something else going on, although nothing jumped out me in the syslog or dmesg.
I'll completely wipe over the weekend and try again to see if I find anything.
Kong I know you're extremely busy, but if you can provide any guidance on where I can look for a potential culprit that would be appreciated. I hate giving up on problem without solving it.
I did upgrade to 29300 and used your bootstrap for opkg. Any chance of adding xinetd to your repo?
Joined: 10 Jan 2015 Posts: 270 Location: Minnesota
Posted: Sun Apr 10, 2016 2:03 Post subject:
JAMESMTL wrote:
@ MDA400
Unfortunately all routers I have configured with ipv6 and qos are running ~v28600 and I've been holding off upgrading until kong releases a stable k4 build
I see. I did actually find a way to make both work at the same time on latest builds but you need to be using RADVD and not DNSMasq unfortunately...
To any one that minds, with no code for IPv6 RA's in DNSMasq additional options reboot the router (note: My WAN address shows 0.0.0.0 when IPv6 is enabled so i have to replug the WAN cable everytime i boot my router for it to grab the addresses from the modem) enable RADVD then click Save (not applying settings or else it will reboot your router).
Replug your WAN cable and in a few seconds it should grab the IPv6 prefix from your modem. I know its a short term hassle till this gets solved but if you do this make sure you disable RADVD everytime you need to reboot the router to set this up again. _________________ LATEST DD-WRT FW IS LOCATED HERE: https://dd-wrt.com/support/other-downloads/?path=betas%2F
Joined: 13 Mar 2014 Posts: 856 Location: Montreal, QC
Posted: Sun Apr 10, 2016 4:21 Post subject:
Hmmm. This is way to early in the debug process but I been looking at a user's setup in the atheros forums who appears to be having issues with DHCPv6-PD address acquisition. It can take hours before dhcp6c assigns the PD even though ISP is sending it out immediately. Straight up DHCPv6 for WANIF is being assigned immediately.
Fwiw, he's also running a current version. At this point I've more or less ruled out carrier security so I'm beginning to question if there is a subtle issue affecting a small subset of users.
Unfortunately I can't look into it further until later on this week but if I come across anything potentially related to your issue I'll follow up here. Anyhow its early in the debug process so don't get your hopes up yet.
I removed entware (or at least everything entware associated I could find) and still had the issue.
But rebooting and modifying IPv6 without my usb attached (/opt mounted) allowed modification with no reboot so there must be something else going on, although nothing jumped out me in the syslog or dmesg.
I'll completely wipe over the weekend and try again to see if I find anything.
Kong I know you're extremely busy, but if you can provide any guidance on where I can look for a potential culprit that would be appreciated. I hate giving up on problem without solving it.
I did upgrade to 29300 and used your bootstrap for opkg. Any chance of adding xinetd to your repo?
There is still a problem with 6in4 tunnels in the latest builds. If you are running this type of tunnel, especially with HE.net or SixXS, can you try adding a firewall script that runs the following commands and let us know if it fixes the problem?
Code:
# Fix IPv6 Tunnel
ip -6 addr add <Assigned/Routed Prefix>::/64 dev br0
ip -6 addr add <Tunnel Client IPv6 Address>::2/64 dev ip6tun
echo 1480 > /proc/sys/net/ipv6/conf/br0/mtu
Joined: 13 Mar 2014 Posts: 856 Location: Montreal, QC
Posted: Fri Apr 15, 2016 1:04 Post subject:
@MDA400
if the issue reoccurs and your game for some debugging let me know. The case of the other user I mentioned above was ISP related and I'm looking to test how often this problem is occurring in the wild.
debugging would involve running the dhcpv6-pd (dhcp6c) client in debug mode while doing a tcpdump.
Joined: 10 Jan 2015 Posts: 270 Location: Minnesota
Posted: Fri Apr 15, 2016 8:24 Post subject:
JAMESMTL wrote:
@MDA400
if the issue reoccurs and your game for some debugging let me know. The case of the other user I mentioned above was ISP related and I'm looking to test how often this problem is occurring in the wild.
debugging would involve running the dhcpv6-pd (dhcp6c) client in debug mode while doing a tcpdump.
Thanks for the reply Kong... I'll update to 29300 and get rid of entware... and let you know what I see.
I'll also dig deeper into my logs to see if I see any indicator of a culprit.
So I dug into this some more and found the culprit.
Many moons ago, when OTRW2 was still fully functional on the latest builds (between 18000 and 22000) I updated my path in my .profile in the root so that the /opt folders came first. Apparently I still had so non-functioning binaries in some of those folders. I'd completely forgotten about it. Once I changed it so the /opt folders came afterwards, things worked as expected. I'm pleasantly surprised nothing else got broken as a result of that.
Oh and the reboot issue, was an ID10T error. Apparently in some script modification I'd posted a reboot call in my startup script which would have a very nasty side effect.
The long and short of it is, it was all user error. Thanks for the great work!
