[Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Dave Taht]

I am not in a position to test this release at all - I gave my last
wndr3800s to a deserving student in Paris, and the only other ones I
have are in the yurtlab and I'm 3000 miles from there.

So do not under any circumstances install this one on your default
gateway as yet.

If you are daring, and bold, and have had perhaps a bit too much
eggnog, please feel free to try out this release.

I AM happy to announce that thx to stephen walker polishing up the
last bits of bug http://www.bufferbloat.net/issues/362 and then
getting them into ceropackages, that at least in theory we have
working upnp and mini-dnssd support. I don't have a device that uses
this stuff, so I'd like to know if they work, particularly on an xbox.
Or if I still missed something... I would certainly like more people's
xboxes working by christmas.

There is also the first version of Steven Barth's new ipv6-support
package - supporting 6in4, 6to4, 6rd, and native ipv6, AND *dhcp-pd*,
interfacing with a totally bleeding edge and highly beta version of
Simon Kelley's dnsmasq, which does saner ipv6 naming of slaac and
dhcpv6 assigned names, and dnsmasq also does dhcp-v6 services, and
handles RA announcements, in addition to it's also eminent duties as a
dns and dhcp server.

Expect bugs! but aside from some new fq_codel work and getting npt66
to work this is rather close to "Feature complete" for the next stable
version of cerowrt...

And... radvd is DEAD. Yea!

The ipv6-support package is rather under documented as yet, as you
might imagine. The GUI needs updating... the new options to it and
dnsmasq I have in my mailbox...



-- 
Dave Täht

Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Dave Taht]

Oops... forgot the link.

http://snapon.lab.bufferbloat.net/~cero2/cerowrt/3.6/wndr/3.6.11-2/

Back to the eggnog, for me, I guess.

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Dave Taht]

Oh, and optionally available and installable via opkg is are the "THC"
ipv6 test tools

(talked about here: http://www.thc.org/thc-ipv6/ )

Now you too can share my nightmares about making ipv6 universally available.

On Sun, Dec 23, 2012 at 3:09 PM, Dave Taht <dave.taht@gmail.com> wrote:
> Oops... forgot the link.
>
> http://snapon.lab.bufferbloat.net/~cero2/cerowrt/3.6/wndr/3.6.11-2/
>
> Back to the eggnog, for me, I guess.



-- 
Dave Täht

Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Maciej Soltysiak]

[-- Attachment #1: Type: text/plain, Size: 2191 bytes --]

Fantastic! I'll test the build with a upnp client and an ssdp device mid
next week.

Happy holidays!
Maciej
On 23 Dec 2012 21:08, "Dave Taht" <dave.taht@gmail.com> wrote:

> I am not in a position to test this release at all - I gave my last
> wndr3800s to a deserving student in Paris, and the only other ones I
> have are in the yurtlab and I'm 3000 miles from there.
>
> So do not under any circumstances install this one on your default
> gateway as yet.
>
> If you are daring, and bold, and have had perhaps a bit too much
> eggnog, please feel free to try out this release.
>
> I AM happy to announce that thx to stephen walker polishing up the
> last bits of bug http://www.bufferbloat.net/issues/362 and then
> getting them into ceropackages, that at least in theory we have
> working upnp and mini-dnssd support. I don't have a device that uses
> this stuff, so I'd like to know if they work, particularly on an xbox.
> Or if I still missed something... I would certainly like more people's
> xboxes working by christmas.
>
> There is also the first version of Steven Barth's new ipv6-support
> package - supporting 6in4, 6to4, 6rd, and native ipv6, AND *dhcp-pd*,
> interfacing with a totally bleeding edge and highly beta version of
> Simon Kelley's dnsmasq, which does saner ipv6 naming of slaac and
> dhcpv6 assigned names, and dnsmasq also does dhcp-v6 services, and
> handles RA announcements, in addition to it's also eminent duties as a
> dns and dhcp server.
>
> Expect bugs! but aside from some new fq_codel work and getting npt66
> to work this is rather close to "Feature complete" for the next stable
> version of cerowrt...
>
> And... radvd is DEAD. Yea!
>
> The ipv6-support package is rather under documented as yet, as you
> might imagine. The GUI needs updating... the new options to it and
> dnsmasq I have in my mailbox...
>
>
>
> --
> Dave Täht
>
> Fixing bufferbloat with cerowrt:
> http://www.teklibre.com/cerowrt/subscribe.html
> _______________________________________________
> Cerowrt-devel mailing list
> Cerowrt-devel@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/cerowrt-devel
>

[-- Attachment #2: Type: text/html, Size: 2876 bytes --]

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Dave Taht]

On Sun, Dec 23, 2012 at 3:41 PM, Maciej Soltysiak <maciej@soltysiak.com> wrote:
> Fantastic! I'll test the build with a upnp client and an ssdp device mid
> next week.

I'd appreciate some confirmation that it actually booted from someone.
I just tried to update the only machine in the yurtlab I dared
reinstall remotely, and lost it from the network (saw it only briefly
on the mesh, via a very looong 500 yard last hop. Scratching my head
now about why it's not being picked up by a closer machine...)

I'll do a pico station build next and see what happens.

I'll be within a few yards dec 29th, but between now and then expect
no real work from me.

> Happy holidays!
> Maciej
>
> On 23 Dec 2012 21:08, "Dave Taht" <dave.taht@gmail.com> wrote:
>>
>> I am not in a position to test this release at all - I gave my last
>> wndr3800s to a deserving student in Paris, and the only other ones I
>> have are in the yurtlab and I'm 3000 miles from there.
>>
>> So do not under any circumstances install this one on your default
>> gateway as yet.
>>
>> If you are daring, and bold, and have had perhaps a bit too much
>> eggnog, please feel free to try out this release.
>>
>> I AM happy to announce that thx to stephen walker polishing up the
>> last bits of bug http://www.bufferbloat.net/issues/362 and then
>> getting them into ceropackages, that at least in theory we have
>> working upnp and mini-dnssd support. I don't have a device that uses
>> this stuff, so I'd like to know if they work, particularly on an xbox.
>> Or if I still missed something... I would certainly like more people's
>> xboxes working by christmas.
>>
>> There is also the first version of Steven Barth's new ipv6-support
>> package - supporting 6in4, 6to4, 6rd, and native ipv6, AND *dhcp-pd*,
>> interfacing with a totally bleeding edge and highly beta version of
>> Simon Kelley's dnsmasq, which does saner ipv6 naming of slaac and
>> dhcpv6 assigned names, and dnsmasq also does dhcp-v6 services, and
>> handles RA announcements, in addition to it's also eminent duties as a
>> dns and dhcp server.
>>
>> Expect bugs! but aside from some new fq_codel work and getting npt66
>> to work this is rather close to "Feature complete" for the next stable
>> version of cerowrt...
>>
>> And... radvd is DEAD. Yea!
>>
>> The ipv6-support package is rather under documented as yet, as you
>> might imagine. The GUI needs updating... the new options to it and
>> dnsmasq I have in my mailbox...
>>
>>
>>
>> --
>> Dave Täht
>>
>> Fixing bufferbloat with cerowrt:
>> http://www.teklibre.com/cerowrt/subscribe.html
>> _______________________________________________
>> Cerowrt-devel mailing list
>> Cerowrt-devel@lists.bufferbloat.net
>> https://lists.bufferbloat.net/listinfo/cerowrt-devel



-- 
Dave Täht

Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Richard Brown]

[-- Attachment #1.1: Type: text/plain, Size: 8063 bytes --]

Initial Impressions - Field Report on 3.6.11-2 Modena build

Dave: Congratulations on the new build. There's a lot to like here. I have a bunch of fussy comments...

- Installed without incident using procedure in the wiki
- DNS came up quickly (shortly after I had an IPv4 address)
- I didn't get DHCP address on Ethernet (se00) until I disconnected the Ethernet connector and re-connected it. I had disabled WiFi on my laptop, and used the ethernet link to burn the firmware (using address 192.168.1.2 as suggested in the wiki). After the tftp completed, I switched back to DHCP addressing, and waited for the router to come up. A couple minutes later, I still had a link local address (169.254…). I disconnected Ethernet and reconnected, and immediately got a 172.30.42.x address. Not sure what the proper behavior is, but I expected to have DHCP provide a 172.30.42.x address automatically.
- The web gui reports that the firmware is "CeroWrt Modena 3.6.10-3" (in two places), although the kernel is listed as "3.6.11" See attached screen shot.


[cid:29994FE1-5771-4593-88BD-41D1404C3842@home.lan]

- I have seen other reports about mDNS naming when I switch from the wired interface to using wireless. I get a GUI message in OSX that says, basically, 'Name Conflict: Updated Local Hostname from "Richs-MacBook-Pro-9.local" to "Richs-MacBook-Pro-10.local"' Here are the system log file entries that contain "mdns" at the time. NB: I am using CeroWrt as a secondary router, letting the gw00 interface connect to the LAN side of my primary router

---
12/25/12 9:58:30.349 PM mDNSResponderHelper[31716]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:13.069 PM mDNSResponderHelper[32051]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:13.069 PM mDNSResponderHelper[32051]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:20.782 PM mDNSResponderHelper[32051]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:20.783 PM mDNSResponderHelper[32051]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:24.981 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Received from FDBD:E109:C5B4:0004:0000:0000:0000:0001:5353    4 Richs-MacBook-Pro-9.local. Addr 172.30.42.25
12/25/12 10:15:24.981 PM mDNSResponder[51]: mDNSCoreReceiveResponse: ProbeCount 2; will deregister   16 Richs-MacBook-Pro-9.local. AAAA FE80:0000:0000:0000:E2F8:47FF:FE11:C796
12/25/12 10:15:24.982 PM mDNSResponder[51]: Local Hostname Richs-MacBook-Pro-9.local already in use; will try Richs-MacBook-Pro-10.local instead
12/25/12 10:15:24.982 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Ignoring response received before we even began probing:   34 Richs-MacBook-Pro-9\.local._intermapper._tcp.local. SRV 0 0 8181 Richs-MacBook-Pro-10.local.
12/25/12 10:15:24.982 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Ignoring response received before we even began probing:   34 PasswordWallet4\032Sync\032(CBC6BE41-B4A5-41AB-82C0-979B7119E187)._passwordwallet._tcp.local. SRV 0 0 1025 Richs-MacBook-Pro-10.local.
12/25/12 10:15:24.983 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Ignoring response received before we even began probing:   34 Rich’s\032MacBook\032Pro\032(richb)._syncmate._tcp.local. SRV 0 0 27384 Richs-MacBook-Pro-10.local.
12/25/12 10:15:25.011 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Ignoring response received before we even began probing:   34 Richs-MacBook-Pro-9\.local._intermapper._tcp.local. SRV 0 0 8181 Richs-MacBook-Pro-10.local.
12/25/12 10:15:25.012 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Ignoring response received before we even began probing:   34 PasswordWallet4\032Sync\032(CBC6BE41-B4A5-41AB-82C0-979B7119E187)._passwordwallet._tcp.local. SRV 0 0 1025 Richs-MacBook-Pro-10.local.
12/25/12 10:15:25.012 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Ignoring response received before we even began probing:   34 Rich’s\032MacBook\032Pro\032(richb)._syncmate._tcp.local. SRV 0 0 27384 Richs-MacBook-Pro-10.local.
12/25/12 10:15:25.904 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Received from 172.30.42.97:5353   33 Richs-MacBook-Pro-9\.local._intermapper._tcp.local. SRV 0 0 8181 Richs-MacBook-Pro-9.local.
12/25/12 10:15:25.904 PM mDNSResponder[51]: mDNSCoreReceiveResponse: ProbeCount 0; will deregister   34 Richs-MacBook-Pro-9\.local._intermapper._tcp.local. SRV 0 0 8181 Richs-MacBook-Pro-10.local.
12/25/12 10:15:25.904 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Received from 172.30.42.97:5353   33 PasswordWallet4\032Sync\032(CBC6BE41-B4A5-41AB-82C0-979B7119E187)._passwordwallet._tcp.local. SRV 0 0 1025 Richs-MacBook-Pro-9.local.
12/25/12 10:15:25.904 PM mDNSResponder[51]: mDNSCoreReceiveResponse: ProbeCount 0; will deregister   34 PasswordWallet4\032Sync\032(CBC6BE41-B4A5-41AB-82C0-979B7119E187)._passwordwallet._tcp.local. SRV 0 0 1025 Richs-MacBook-Pro-10.local.
12/25/12 10:15:25.905 PM mDNSResponder[51]: mDNSCoreReceiveResponse: Received from 172.30.42.97:5353   33 Rich’s\032MacBook\032Pro\032(richb)._syncmate._tcp.local. SRV 0 0 27384 Richs-MacBook-Pro-9.local.
12/25/12 10:15:25.905 PM mDNSResponder[51]: mDNSCoreReceiveResponse: ProbeCount 0; will deregister   34 Rich’s\032MacBook\032Pro\032(richb)._syncmate._tcp.local. SRV 0 0 27384 Richs-MacBook-Pro-10.local.
12/25/12 10:15:26.438 PM mDNSResponderHelper[32051]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:26.438 PM mDNSResponderHelper[32051]: do_mDNSInterfaceAdvtIoctl: ioctl call SIOCGIFINFO_IN6 failed - error (22) Invalid argument
12/25/12 10:15:26.695 PM mDNSResponder[51]: Name Conflict: Updated Local Hostname from "Richs-MacBook-Pro-9.local" to "Richs-MacBook-Pro-10.local"
---

- The UPNP configuration page gives an error. See attached image.

---
This page contains the following errors:
error on line 111 at column 33: xmlParseEntityRef: no name
Below is a rendering of the page up to the first error.
---

[cid:BA5C8F54-C3FB-449E-A371-9F82E3A4CBEF@home.lan]

- IPv6: Without doing any configuration (and no external IPv6 addressing set up yet), my Mac reports an IPv6 of fdbd:e109:c5b4::3e07:54ff:fe3f:daff at the same time I got the DHCP address.
- I configured a Hurricane Electric tunnel using the GUI and the information from my tunnelbroker.net<http://tunnelbroker.net> page. CeroWrt successfully got the expected 2001:470… IPv6 address and could ping V6 hosts. None of the LAN interfaces of CeroWrt got a 2001… addresses, although they had fe80:... and fdbd:… addresses. For example, ifconfig showed:

...
se00      Link encap:Ethernet  HWaddr C6:3D:C7:9D:E3:9A
          inet addr:172.30.42.1  Bcast:172.30.42.31  Mask:255.255.255.224
          inet6 addr: fdbd:e109:c5b4::1/64 Scope:Global
          inet6 addr: fe80::c43d:c7ff:fe9d:e39a/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:16854 errors:0 dropped:0 overruns:12 frame:0
          TX packets:16612 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:4969550 (4.7 MiB)  TX bytes:7208592 (6.8 MiB)
          Interrupt:4

sw00      Link encap:Ethernet  HWaddr C4:3D:C7:9D:E3:9A
          inet addr:172.30.42.65  Bcast:172.30.42.95  Mask:255.255.255.224
          inet6 addr: fdbd:e109:c5b4:2::1/64 Scope:Global
          inet6 addr: fe80::c63d:c7ff:fe9d:e39a/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:456 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:32
          RX bytes:0 (0.0 B)  TX bytes:138408 (135.1 KiB)
… etc...

I will keep playing with IPv6 configuration following the OpenWrt ipv6-support wiki page…

Rich

[-- Attachment #1.2: Type: text/html, Size: 10254 bytes --]

[-- Attachment #2: CeroWrt-3.6.11-2.png --]
[-- Type: image/png, Size: 27182 bytes --]

[-- Attachment #3: CeroWrt 3.6.11-2-UPNP error.png --]
[-- Type: image/png, Size: 22574 bytes --]

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Dave Taht]

[-- Attachment #1: Type: text/plain, Size: 327 bytes --]

As for the MDNS issue, we've filed a bug on it, and demonstrated to
multiple people "that should know", and no progress on fixing it as yet.

I hope it's merely a bug in avahi rather than the protocol. It's a real
PITA.

-- 
Dave Täht

Fixing bufferbloat with cerowrt:
http://www.teklibre.com/cerowrt/subscribe.html

[-- Attachment #2: Type: text/html, Size: 453 bytes --]

Re: [Cerowrt-devel] CeroWrt 3.6.11-2 is out with ipv6-support package, bleeding edge dnsmasq, upnp,

[Dave Taht]

I note that the bufferbloat.net domain filters out ip addresses in
message bodies, so richard's original message probably didn't make the
list. I have to find a fix for that - so 172 dot X dot Y dot Z goes
through.

...

Thx for the complete report!

Configuration name problem noted and fixed. (hopefully). Have to automate
that, I keep doing it...

An update to the luci upnp gui landed a day or two ago for upnp, it's my
hope that addresses the bug there, but as I am still 3000 miles away from
the nearest router I dare reflash I won't be in a position to test til
saturday. I did do a build with that fix in it:

http://snapon.lab.bufferbloat.net/~cero2/cerowrt/wndr/3.6.11-3/

Which is so trivial as you can do a sysupgrade to it without harming your
existing config. Unless I broke something else.

The default configuration defaults to generating ULAs which are local ipv6
addresses in the fd:: range. While this is useful for bringing up a ipv6
network, it's not so useful for connecting to the outside world with it, as
you've discovered. My hope is with the linux 3.7 release to be able to
enable npt66 by default in this case, (which was the source of some
controversy earlier this month), but using 3.7 as a base is a bit of a long
way out. I'm trying to get on top of the 3.7 issue (and some other gui
stuff) by doing an x86 virtual build (which so far, doesn't boot)

As for not distributing the 2001:: (and disabling the ula entirely) I don't
know - that package is so bleeding edge as yet! Happy hacking.




-- 
Dave Täht

Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html