Tuesday, December 29, 2015< ^ >
Thomas Watteyne has set the subject to: IETF93 ROLL WG meeting
Room Configuration
Room Occupants

[01:10:53] <Michael Richardson> I think to worry about the various other 6lowpan solutions (like BACnet) should worry about things when we get there. If netdev folks would go for dev_addr2, then great.  
[01:11:10] <Michael Richardson> otherwise, I think 8+2 would be easier for the rest of the system.
[01:12:13] <Michael Richardson> G8859 has (RFC7428) 1-byte addresses.... I think it never uses the long form, as I recall.
[13:30:31] eintopf joins the room
[13:30:50] <eintopf> couple of people cancel the talk @ IoT devroom, FOSDEM
[13:30:57] <eintopf> I have the chance again :-)
[13:31:08] <eintopf> If there are free slots, everybody has the chance
[13:38:24] <eintopf> ([01:12:13] <Michael Richardson> G8859 has (RFC7428) 1-byte addresses.... I think it never uses the long form, as I recall.) - no idea. I have currently the overview about subsystems which has the linux kernel and it has 6LoWPAN support -> 802.15.4 and BTLE
[13:38:39] <eintopf> possible new would be nfc, because linux has a nfc subsystem
[13:39:36] <eintopf> and the one link layer which running out of traditional mac address handling stuff is 802.15.4 with two different mac types which can be used at the same time. :S
[13:54:14] <eintopf> and I have no idea how bluetooth works, but I tested some time in a very unstable environment that btle is terrible slow
[13:54:16] <eintopf> btle 6lowpan
[13:58:28] <eintopf> but my talk... I am not happy to talking about that... but the most people on the mailinglist are asking "how can I build rpi mainline kernel, etc..."
[13:58:38] <eintopf> and have different issues why it's not working
[15:08:29] <Michael Richardson> true... ML can be annoying --- too many noobs...
[15:10:48] <eintopf> it's also some sign to me that it needs more easy-to-use stuff
[15:12:29] <eintopf> my talk is about board support packages which will build mainline kernel and then wip patches for radvd to support some new optionfields, etc. it's also good for development
[15:12:46] <eintopf> for common platforms rpi/rpi2/beaglebone
[15:13:04] <eintopf> also qemu with fakelb driver
[15:13:22] <eintopf> should be fine then when I answer "use xy BSP"
[15:13:32] <eintopf> WPANKit is the name :-)
[15:14:13] <eintopf> "easy-to-use" is also not possible for 802.15.4 because missing usb dongle support...
[15:14:26] <eintopf> but openlabs with this BSP, it should work _mostly_ out of box
[15:14:53] <eintopf> networking setup is another question, there we need support inside network managers
[15:15:56] <eintopf> and to talk about that, it's my job. I use tools from my company to do that (all open source stuff)
[16:02:04] <eintopf> btw: I will try to include unstrung into this BSP
[17:43:34] <eintopf> complete kernel freeze when RPi tries to load the btle usb dongle
[17:44:02] <eintopf> grml
[17:53:57] <eintopf> okay, the btle setup which I try to get working is always so much unstable
[17:54:07] <eintopf> I think the dongles are bad, but I don't have others...
[17:56:03] <eintopf> ok, doesn't matter I will say that the btle state is unstable in my expierence and works with qemu+usb support
[18:02:05] <eintopf> I think the beaglebone usb support is better
[22:38:43] <eintopf> 64 bytes from fe80::5ef3:70ff:fe6b:3b1f: icmp_seq=1000 ttl=64 time=140 ms
--- fe80::5ef3:70ff:fe6b:3b1f ping statistics ---
1000 packets transmitted, 1000 received, 0% packet loss, time 303309ms
rtt min/avg/max/mdev = 75.521/119.329/206.910/31.378 ms
[22:39:22] <eintopf> ~120 ms rtt for btle 6lowpan ping6 @ 64 bytes payload
[22:39:54] <eintopf> but I get it working now between qemu <-> beaglebone
[22:40:25] <eintopf> RPi USB support is a mess :-)
[22:41:05] <eintopf> reminds me to fix some btle 6lowpan bugs which are known :-/
[22:55:55] <eintopf> good night
[22:56:01] eintopf leaves the room: Disconnected: closed
Powered by ejabberd Powered by Erlang Valid XHTML 1.0 Transitional Valid CSS!