**** BEGIN LOGGING AT Wed Mar 09 02:59:58 2016
Mar 09 09:31:35 <Shapeshifter> Hi. I'm trying to use tethering via USB with my N900 (providing the laptop with a 3G internet connection). I don't have NetworkManager. I connected the N900 via USB and set it to PC Suite Mode. Two network devices show up in ip link: usbpn0 and enp0s20u1i8. I tried using dhcpcd on both but it's not working, I get a time out.
Mar 09 09:32:11 <L29Ah> ifconfig them up ffs
Mar 09 09:33:16 <Shapeshifter> L29Ah: I tried it. ip link set usbpn0 up has no effect. It stays in "UNKNOWN" mode. usbpn0: <POINTOPOINT,NOARP,UP,LOWER_UP> mtu 65541 qdisc fq_codel state UNKNOWN mode DEFAULT group default qlen 3. The other device, enp0s20u1i8 stays in DOWN mode, i.e. ip link set enp0s20u1i8 up has no effect
Mar 09 09:33:26 <Shapeshifter>  enp0s20u1i8: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc fq_codel state DOWN mode DEFAULT group default qlen 1000
Mar 09 09:35:54 <Maxdamantus> Shapeshifter: you'd need to either bridge the devices or forward them (either through routing or netfilter rules) and probably do NAT somewhere.
Mar 09 09:36:19 <Maxdamantus> er, to be clear, the NAT is probably required in both cases.
Mar 09 09:36:36 <Maxdamantus> I just do: iptables --table nat --append POSTROUTING --out-interface gprs0 -j MASQUERADE && iptables --append FORWARD --in-interface usb0 -j ACCEPT
Mar 09 09:37:23 <L29Ah> Mar 09 12:35:55 [kernel] [912041.004420] cdc_ether 1-1.5.4.2:1.8 usb0: register 'cdc_ether' at usb-0000:00:1a.0-1.5.4.2, CDC Ethernet Device, ea:82:70:48:30:4
Mar 09 09:37:31 <L29Ah> Nokia-N900:~# ifconfig usb0 up 192.168.66.6
Mar 09 09:37:37 <L29Ah> l29ah-x201 /usr/src/linux/drivers/hid ∞ ifconfig usb0 up 192.168.66.5
Mar 09 09:37:38 <L29Ah> l29ah-x201 /usr/src/linux/drivers/hid ∞ ping 192.168.66.6
Mar 09 09:37:38 <L29Ah> PING 192.168.66.6 (192.168.66.6) 56(84) bytes of data.
Mar 09 09:37:38 <L29Ah> 64 bytes from 192.168.66.6: icmp_seq=1 ttl=64 time=1.46 ms
Mar 09 09:37:38 <L29Ah> 64 bytes from 192.168.66.6: icmp_seq=2 ttl=64 time=1.00 ms
Mar 09 09:37:39 <L29Ah> meh
Mar 09 09:38:17 <L29Ah> dunno if n900 calls a dhcp client when it connects to you, you can check it with wireshark
Mar 09 09:38:33 <Maxdamantus> I don't think it does.
Mar 09 09:38:59 <Maxdamantus> but it's not hard to do.
Mar 09 09:39:03 <L29Ah> anyway you can ask your udev to do the proper init when it's connected
Mar 09 09:39:23 <Maxdamantus> Presumably in that situation though you're running a DHCP server on the device it's tethered to.
Mar 09 09:39:38 <Maxdamantus> (or have the usb network interface bridged to something that does)
Mar 09 09:57:12 <Shapeshifter> When you plug in the n900 in PC Suite mode, do you also get a message in dmesg saying "cdc_ether 1-1:1.8 enp0s20u1i8: kevent 12 may have been dropped" ?
Mar 09 09:58:07 <L29Ah> no
Mar 09 09:58:10 <Shapeshifter> http://lpaste.net/6600118746759561216 
Mar 09 09:58:18 <L29Ah> also i get only one ethernet
Mar 09 09:58:29 <L29Ah> Mar 09 12:35:55 [kernel] [912041.002204] cdc_acm 1-1.5.4.2:1.6: ttyACM0: USB ACM device
Mar 09 09:58:29 <L29Ah> Mar 09 12:35:55 [kernel] [912041.004420] cdc_ether 1-1.5.4.2:1.8 usb0: register 'cdc_ether' at usb-0000:00:1a.0-1.5.4.2, CDC Ethernet Device, ea:82:70:48:30:4e
Mar 09 09:58:32 <Shapeshifter> It says "Gained carrier", then that error, then "Lost carrier". Might be the reason why it stays down
Mar 09 10:03:36 <Maxdamantus> Shapeshifter: imo try just using g_ether instead of g_nokia.
Mar 09 10:05:44 <Maxdamantus> Hm. If you've run dhcpcd, it might be hanging around and setting the interface down.
Mar 09 10:06:16 <Maxdamantus> dhcpcd simply isn't going to work though unless you're running a DHCP server on the other end.
Mar 09 10:07:07 <Maxdamantus> Mär 09 10:57:32 tachychineta systemd-networkd[18178]: enp0s20u1i8: Lost carrier
Mar 09 10:07:20 <Maxdamantus> Yeah, that message is coming from something in systemd, probably dhclient.
Mar 09 10:09:00 <DocScrutinizer05> Shapeshifter: http://paste.opensuse.org/68231912
Mar 09 10:09:08 <DocScrutinizer05> ^^^ PC side
Mar 09 10:12:40 <Maxdamantus> That `route` line seems pointless.
Mar 09 10:13:36 <Maxdamantus> 192.168.0.202/32 is already part of the 192.168.0.0/24 subnet, which Linux normally adds a route for by default when you configure the interface using that `ifconfig` command.
Mar 09 10:13:52 <Maxdamantus> 192.168.1.0/24 dev brznet  proto kernel  scope link  src 192.168.1.42 
Mar 09 10:18:53 <Shapeshifter> It's not helping though. I can set a static ip but the interface remains down http://lpaste.net/5433001060979441664
Mar 09 10:19:43 <Shapeshifter> If I plug it in, the very first time I run ip link set usb0 up I get the messages "Gained carrier", "kevent 12 may have been dropped", "Lost carrier". I think it may be a driver bug or something
Mar 09 10:19:52 <DocScrutinizer05> this script is like 7 years old and eventually got adapted to new interface naming once (usb0 -> eth4). Otherwise it always worked, given the network doesn't collide with your regular PC ethernet
Mar 09 10:21:58 <DocScrutinizer05> ooh, and I usually had to stop the damn network manager on PC
Mar 09 10:22:41 <DocScrutinizer05> network manager tries to act smart and thus actually acts terribly stupid
Mar 09 10:23:24 <Shapeshifter> I'm not using NM. I tried it manually and with systemd-networkd and udev rules. Anyway, I don't think it's a config issue at this point. The interface just can't go up.
Mar 09 10:23:43 <DocScrutinizer05> ohmy, systemd
Mar 09 10:23:45 <Maxdamantus> You're using systemd, which does network management.
Mar 09 10:24:26 <Maxdamantus> on my Debian machine at work I have this: https://gist.github.com/Maxdamantus/57c2765f2782c0932986
Mar 09 10:24:27 <Shapeshifter> Maxdamantus: yeah but I tried it with systemd-networkd disabled as well.
Mar 09 10:24:29 <DocScrutinizer05> systemd is worse than NM
Mar 09 10:24:52 <Maxdamantus> in /etc/network/interfaces.d/n900usb
Mar 09 10:25:10 <Maxdamantus> I suspect that's a general systemd configuration.
Mar 09 10:26:10 <Shapeshifter> Maxdamantus: I don't think so. systemd-networkd config is in /etc/systemd/network and uses a different format.
Mar 09 10:26:46 <Maxdamantus> Ah, so it does.
Mar 09 10:28:19 <DocScrutinizer05> isn't it nice how the never actually seen problems that systemd "solves" are creating new problems all over the place, in stuff that worked before? Dang, what does that remind me to... AAH PolypAudio, and AVAHI
Mar 09 10:28:57 <Shapeshifter> I'm no fan of systemd either but it's here to stay
Mar 09 10:29:15 <DocScrutinizer05> like HAL
Mar 09 10:31:17 <DocScrutinizer05> also see devuan
Mar 09 10:31:21 <Maxdamantus> and Wayland.
Mar 09 10:31:33 <DocScrutinizer05> http://devuan.org
Mar 09 10:31:55 <DocScrutinizer05> actually...
Mar 09 10:31:58 <DocScrutinizer05> ~nosystemd
Mar 09 10:31:59 <infobot> [nosystemd] https://devuan.org http://without-systemd.org/wiki/index.php/Arguments_against_systemd http://blog.darknedgy.net/technology/2015/10/11/0/
Mar 09 10:32:11 <DocScrutinizer05> ~systemd
Mar 09 10:32:11 <infobot> systemd cabal: a bunch of people (Lennart Poettering, Kay Sievers, H. Hoyer, D. Mack, T. Gundersen, D. Herrmann) who want to turn linux into their wet dream perverted version of windows-me-too: http://0pointer.net/blog/revisiting-how-we-put-together-linux-systems.html -- Rumor has it that 2016 systemd will have replaced kernel. See ~nosystemd"
Mar 09 10:49:13 <DocScrutinizer05> also note that maemo uses another incompatible concept of cgroups and thus it will be virtually impossible to migrate maemo to systemd
Mar 09 10:50:34 <Maxdamantus> I imagine maemo's uses of cgroups could probably be merged into systemd's uses of cgroups.
Mar 09 10:50:47 <Wizzup> DocScrutinizer05: I hope there are no plans to.
Mar 09 10:50:47 <DocScrutinizer05> not even get me started about "usr/?bin must live on rootfs"
Mar 09 10:50:50 <Wizzup> Openrc is just fine.
Mar 09 10:51:06 <Maxdamantus> in any case, I wouldn't consider them particularly part of maemo.
Mar 09 10:51:31 <Maxdamantus> not any moreso than Linux 2.6.28 is part of Maemo.
Mar 09 10:51:44 <Maxdamantus> it's just a random implementation detail in how the maemo software is run.
Mar 09 10:52:26 <Wizzup> 'random'
Mar 09 10:52:37 <Maxdamantus> non-maemo-specific.
Mar 09 10:53:13 <DocScrutinizer05> Wizzup: 'random'?
Mar 09 10:53:29 <Wizzup> you're not seeing what Maxdamantus is saying
Mar 09 10:53:37 <DocScrutinizer05> ooh, nevermind
Mar 09 10:53:40 <Wizzup> yep.
Mar 09 10:54:01 <DocScrutinizer05> I guess he disagrees with me as usual
Mar 09 10:54:14 <Maxdamantus> If you ran hildon-desktop, hildon-status-menu, maemo-xinput-sounds, browserd, etc outside of any cgroups, I'd still call it maemo.
Mar 09 10:54:32 <Maxdamantus> just as if you ran all those things on a kernel other than 2.6.28
Mar 09 10:55:16 <Maxdamantus> Its use of cgroups is only an implementation detail in how it's currently run.
Mar 09 10:56:46 <Maxdamantus> I don't necessarily usually disagree with you, but I'm more likely to voice my opinions about disagreements than agreements.
Mar 09 10:57:11 <DocScrutinizer05> Wizzup: ((openrc)) well, iirc OOM (with PA etc) use cgroups in maemo, so you have a szable implementation of phonecalls etc. You don't want to port all that stuff to openrc
Mar 09 10:57:25 <Maxdamantus> I don't like systemd either, but there isn't much to add by saying that I also dislike it.
Mar 09 10:57:49 <Wizzup> DocScrutinizer05: openrc supports linux cgroups, and I don't think it will be too hard to port it over
Mar 09 10:58:13 <DocScrutinizer05> hmm, it's quite obscure, and iirc parts of OOM are even closed
Mar 09 10:58:33 <Maxdamantus> parts of OOM?
Mar 09 10:59:06 <Wizzup> Time to compile v4.5-rc5-n900
Mar 09 10:59:07 <Maxdamantus> I'm sure it's a standard feature of cgroups to make processes in certain cgroups killed first by the OOM killer.
Mar 09 10:59:55 <Maxdamantus> You don't even need cgroups to do it in a way that would be suitable.
Mar 09 11:00:24 <DocScrutinizer05> Wizzup: I din't really like upstart but it's waaaaay better than systemd, and I don't see why we should migrate maemo to any other init system
Mar 09 11:00:33 <Wizzup> because upstart is dead
Mar 09 11:00:40 <Wizzup> openrc is a very sensible alt.
Mar 09 11:00:47 <DocScrutinizer05> it works just fine on all my N900
Mar 09 11:00:48 <Wizzup> I don't want to get into this discussion more fyi
Mar 09 11:00:57 <Wizzup> sure, but many things works well on n900, but they are very outdated
Mar 09 11:00:59 <Maxdamantus> Chromium has been setting something on its child processes that makes them usually the first to be killed.
Mar 09 11:01:09 <Wizzup> and to keep a system up to date and healthy, you cannot maintain all components yourself
Mar 09 11:01:15 <Wizzup> so you have to find upstream that does work for you
Mar 09 11:01:18 <Maxdamantus> (without cgroups, afaik)
Mar 09 11:01:56 <DocScrutinizer05> Wizzup: well, that's maybe true for the usual PC linux system. Maemo is its own upstream
Mar 09 11:02:06 <Wizzup> DocScrutinizer05: that is a problem :) 
Mar 09 11:02:15 <DocScrutinizer05> and CSSU is what's all about keeping it alive
Mar 09 11:02:25 <Maxdamantus> echo 300 >/proc/$$/oom_score
Mar 09 11:02:26 <Wizzup> you're not getting my point, I think
Mar 09 11:02:37 <Wizzup> I am simply saying: you cannot, ever, find enough manpower do to that
Mar 09 11:02:54 <DocScrutinizer05> you even less can find the manpower to migrate
Mar 09 11:03:06 <Wizzup> perhaps
Mar 09 11:05:14 <DocScrutinizer05> since upstart is designed reasonably decent (unlike systemd), the effort to keep it alive is marginal, there hardly will show up any new flaws or bugs
Mar 09 11:05:55 <Wizzup> ... okay.
Mar 09 11:06:10 <DocScrutinizer05> the true flaw is apps depending on the init system, I.E. systemd
Mar 09 11:06:27 <DocScrutinizer05> no app really depends on upstart
Mar 09 11:07:06 <DocScrutinizer05> except for an occasional "systemcl start lircd" or sth like that
Mar 09 11:07:54 <DocScrutinizer05> rather "sudo start lircd"
Mar 09 11:09:26 <bencoh> no, the *true* flow is no fact that with no real upstream we'll keep drifting with old ABIs
Mar 09 11:09:31 <bencoh> flaw*
Mar 09 11:09:52 <bencoh> s/no/the/ bleh 
Mar 09 11:10:36 <DocScrutinizer05> well, if you wanna change the ABI, you need a new major release anyway, since all your programs need recompiling
Mar 09 11:11:39 <DocScrutinizer05> that's maemo7 then ;-)
Mar 09 11:12:00 <bencoh> DocScrutinizer05: recompiling wouldn't be the main issue (it is for closed-source stuff)
Mar 09 11:12:09 <bencoh> APIs change too
Mar 09 11:13:16 <Maxdamantus> usually the ABI compatibility is broken by an API incompatibility.
Mar 09 11:13:19 <DocScrutinizer05> sure, you can migrate / clone all infra incl extras repio to a maemo7 and recompile everything and hope that nothing breaks from the ABI change
Mar 09 11:13:38 <Wizzup> or rip out closed parts
Mar 09 11:13:48 <Maxdamantus> except maybe libpng. I'm not sure what the problem is with that.
Mar 09 11:14:02 <DocScrutinizer05> that won't help, you still need to hope that nothing breaks
Mar 09 11:15:09 <DocScrutinizer05> I think we're quite far already with ripping out the closed bits
Mar 09 11:15:35 <DocScrutinizer05> the whole linux telephony stack is the major issue afaik
Mar 09 11:16:11 <DocScrutinizer05> our RE guys boggle from that
Mar 09 11:17:32 <Wizzup> I know
Mar 09 11:17:44 <Wizzup>  I have yet to live up to my promise to help out there.
Mar 09 11:18:18 <DocScrutinizer05> well, luckily most of it is obsolete anyway on Neo900
Mar 09 11:18:51 <DocScrutinizer05> it's just the damn API that's not really documented
Mar 09 11:20:35 <DocScrutinizer05> so operating telephony is no problem on Neo900, integrating it with the rest of the system might be
Mar 09 11:22:36 <DocScrutinizer05> I could write a bare bones dialer-ui in just 2 days or less. However that won't integrate nicely with phonebook, messaging, status bar, flight mode settings, 911-calls even from lockscreen, etc pp
Mar 09 11:28:19 <DocScrutinizer05> dang ISI, I wish they had used ofono or even fsogsmd
Mar 09 11:30:05 <Shapeshifter> Just to be clear, when you set a static ip on the usb0 device, the interface is UP, yes?
Mar 09 11:30:14 <Shapeshifter> i.e. there's no point to any of this if I can't get the interface UP
Mar 09 11:33:47 <DocScrutinizer05> ask ifconfig if it's UP
Mar 09 11:35:06 <Shapeshifter> DocScrutinizer05: it's not, that's the problem. I can't set it UP
Mar 09 11:35:43 <DocScrutinizer05> on PC?
Mar 09 11:35:54 <Shapeshifter> DocScrutinizer05: yes.
Mar 09 11:44:07 <DocScrutinizer05> sorry, doesn't work on my PC either, prolly wrong cgroup, thanks systemd
Mar 09 11:45:06 <Shapeshifter> DocScrutinizer05: what does that mean? Can I fix it?
Mar 09 11:45:27 <DocScrutinizer05> http://paste.opensuse.org/84826295
Mar 09 11:45:47 <DocScrutinizer05> I don't know how, without learning about that damn systemd
Mar 09 11:47:05 <Shapeshifter> I'm not sure how systemd might mess with this though. Could you paste ip l show enp0s29u1 ?
Mar 09 11:47:43 <DocScrutinizer05> please rephrase
Mar 09 11:47:54 <DocScrutinizer05> 2nd half
Mar 09 11:48:09 <Shapeshifter> Could you please run ip l show enp0s29u1 and paste the first line in here?
Mar 09 11:48:30 <DocScrutinizer05> saturn:~ # ip l show enp0s29u1
Mar 09 11:48:31 <DocScrutinizer05> Device "enp0s29u1" does not exist.
Mar 09 11:48:45 <Shapeshifter> It's weird though that it appears in ifconfig.
Mar 09 11:48:53 <Shapeshifter> So is it listed in ip l=
Mar 09 11:48:56 <Shapeshifter> ?
Mar 09 11:49:18 <DocScrutinizer05> 5: enp0s29u1u8i8: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast state DOWN mode DEFAULT qlen 1000
Mar 09 11:49:48 <Shapeshifter> Ah
Mar 09 11:50:03 <DocScrutinizer05> saturn:~ # ip l show enp0s29u1u8i8
Mar 09 11:50:04 <DocScrutinizer05> 5: enp0s29u1u8i8: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast state DOWN mode DEFAULT qlen 1000
Mar 09 11:50:19 <Shapeshifter> I took the device id from the paste you linked. So does ip l set enp0s29u1u8i8 up work for you?
Mar 09 11:50:29 <Maxdamantus> jove ~ # ip l show usb0
Mar 09 11:50:29 <Maxdamantus> 192: usb0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN mode DEFAULT group default qlen 1000
Mar 09 11:50:30 <DocScrutinizer05> ^^^
Mar 09 11:50:47 <Maxdamantus> jove > saturn
Mar 09 11:50:52 <DocScrutinizer05> yep
Mar 09 11:53:08 <DocScrutinizer05> Shapeshifter: I guess ifconfig truncates the name
Mar 09 11:53:42 <DocScrutinizer05> saturn:~ # ifconfig enp0s29u1u8i8
Mar 09 11:53:44 <DocScrutinizer05> enp0s29u1 Link encap:Ethernet  HWaddr 2E:66:C8:C5:75:F3  
Mar 09 11:55:37 <DocScrutinizer05> saturn:~ # ifconfig enp0s29u1u8i8 192.168.0.200 netmask 255.255.255.0
Mar 09 11:55:38 <DocScrutinizer05> saturn:~ # ifconfig enp0s29u1u8i8
Mar 09 11:55:40 <DocScrutinizer05> enp0s29u1 Link encap:Ethernet  HWaddr 2E:66:C8:C5:75:F3  
Mar 09 11:55:41 <DocScrutinizer05>           inet addr:192.168.0.200  Bcast:192.168.0.255  Mask:255.255.255.0
Mar 09 11:56:02 <DocScrutinizer05> that's prolly not a ttyACM or USB interface though
Mar 09 11:56:36 <Maxdamantus> after doing: ip addr add 192.168.30.1/24 dev usb0; ip link set dev usb0 up
Mar 09 11:56:47 <Maxdamantus> 64 bytes from 192.168.30.2: icmp_req=2 ttl=64 time=0.323 ms
Mar 09 11:57:04 <Maxdamantus> since it was already configured on the N900 side.
Mar 09 11:57:23 <Maxdamantus> (ifconfig usb0 192.168.30.2 netmask 255.255.255.0)
Mar 09 11:57:27 <DocScrutinizer05> crw-rw---- 1 root dialout 166,  0 Mar  9 12:39 /dev/ttyACM0
Mar 09 11:57:59 <DocScrutinizer05> ttyACM however is not what we need
Mar 09 12:01:36 <Shapeshifter> Maxdamantus: so the usb0 is definitely UP?
Mar 09 12:01:45 <Shapeshifter> Maxdamantus: What do I need to configure on the N900 side?
Mar 09 12:05:28 <DocScrutinizer05> Shapeshifter: you seen http://wiki.maemo.org/N900_USB_networking ?
Mar 09 12:07:16 <Shapeshifter> DocScrutinizer05: yes but in the first sentence it says " to share your PC's internet to N900 via USB"
Mar 09 12:07:23 <Shapeshifter> I need the other way around
Mar 09 12:07:39 <Shapeshifter> http://wiki.maemo.org/N900_USB_router I'm trying this now
Mar 09 12:07:55 <DocScrutinizer05> you want the N900 internet access to get shared to PC? 
Mar 09 12:08:07 <parazyd> what is tethering?
Mar 09 12:08:23 <Shapeshifter> DocScrutinizer05: yes. 
Mar 09 12:08:24 <DocScrutinizer05> parazyd: that :-)
Mar 09 12:08:34 <Shapeshifter> Is this jeopardy?
Mar 09 12:08:53 <DocScrutinizer05> Shapeshifter: ideally you use NM or something similar to configure N900 as MODEM
Mar 09 12:10:47 * Wizzup just does /etc/init.d/net.usb0 start
Mar 09 12:10:57 <DocScrutinizer05> http://wiki.maemo.org/USB_networking
Mar 09 12:11:26 <Shapeshifter> oh.
Mar 09 12:11:43 <Shapeshifter> So we've been talking about the other thing for the past hour
Mar 09 12:11:47 <Shapeshifter> xD
Mar 09 12:13:17 <DocScrutinizer05> http://wiki.maemo.org/Debian-Tethering
Mar 09 12:14:05 <DocScrutinizer05> well, that's for DUN (BT)
Mar 09 12:15:03 <Shapeshifter> Is kernel-power still OK with latest CSSU etc?
Mar 09 12:15:40 <DocScrutinizer05> yep, afaik
Mar 09 12:16:11 <DocScrutinizer05> for (usb)tethering you don't need KP though
Mar 09 12:16:36 <Shapeshifter> DocScrutinizer05: the article says that if I use the n900 as a "modem", it won't have internet access itself anymore
Mar 09 12:16:43 <DocScrutinizer05> tethering actually works in PC-mode
Mar 09 12:16:55 <DocScrutinizer05> yes, that's true
Mar 09 12:17:23 <DocScrutinizer05> tehtering completely exports the modem, like mass-storage exposts whole storage partitions
Mar 09 12:17:24 <Shapeshifter> http://wiki.maemo.org/N900_USB_router this looks simple enough, so I'm trying it
Mar 09 12:17:49 <DocScrutinizer05> sure, sounds ok
Mar 09 12:21:35 <DocScrutinizer05> see last section on http://wiki.maemo.org/N900_USB_router - it clearly lists the differences between tethering and networking
Mar 09 12:24:49 <Shapeshifter> https://wiki.maemo.org/N900_USB_networking this one says that the N900 provides a default network interface for usb0
Mar 09 12:24:57 <Shapeshifter> But I don't have such an interface on the N900.
Mar 09 12:25:33 <Shapeshifter> I see the config in /etc/network/interfaces but I only have gprs0, lo, phonet0 in ifconfig on the N900.
Mar 09 12:26:13 <Shapeshifter> and ifub usb0 gives me No such device sure enough
Mar 09 12:30:04 <DocScrutinizer05> Wizzup: /me lols @ cat /etc/init.d/networking
Mar 09 12:30:39 <Shapeshifter> If I set PC Suite mode I also get an upnlink0
Mar 09 12:30:42 <Shapeshifter> but no usb0
Mar 09 12:31:01 <Shapeshifter> oh wait
Mar 09 12:31:09 <Shapeshifter> I get it. 1) PC Suite mode, 2) ifup usb0
Mar 09 12:31:51 <DocScrutinizer05> yes, sounds reasonable
Mar 09 12:33:57 <Wizzup> DocScrutinizer05: on what system?
Mar 09 12:34:04 <DocScrutinizer05> on N900
Mar 09 12:34:30 <DocScrutinizer05> it has a few funny details
Mar 09 12:35:13 <DocScrutinizer05> like >>         if ifdown -a; then             log_action_end_msg $?;         else             log_action_end_msg $?;         fi   <<
Mar 09 12:35:56 <DocScrutinizer05> or the whole doopt()
Mar 09 12:45:14 <DocScrutinizer05> rotfl, bug in /lib/lsb/init-functions
Mar 09 12:45:29 <DocScrutinizer05> [ "x$TERM" != "" ]
Mar 09 12:46:43 <DocScrutinizer05> could you please check if recent versions of /lib/lsb/init-functions on debian have that fixed already?
Mar 09 12:47:04 <DocScrutinizer05> it's in log_use_fancy_output ()
Mar 09 12:47:29 <DocScrutinizer05> #Copyright (c) 2002-08 Chris Lawrence
Mar 09 12:49:50 <bencoh>     [ "x${TERM:-}" != "x" ] &&
Mar 09 12:49:50 <bencoh>     [ "x${TERM:-}" != "xdumb" ] &&
Mar 09 12:50:02 <bencoh> (weezy)
Mar 09 12:50:12 <bencoh> 3+h
Mar 09 12:51:21 <DocScrutinizer05> bencoh: ta
Mar 09 12:52:48 <DocScrutinizer05> Shapeshifter: success?
Mar 09 13:43:33 <Shapeshifter> DocScrutinizer05: not really. I can get a working connection between the laptop and N900 via USB if I configure a static address on the laptop, but so far I can't get an internet connection on the laptop via the N900
Mar 09 13:43:58 <Shapeshifter> According to http://wiki.maemo.org/N900_USB_router dhcp should work as well but it doesn't
Mar 09 13:44:23 <DocScrutinizer05> hmm, you don't need a dhcp
Mar 09 13:44:49 <Shapeshifter> DocScrutinizer05: I don't need one, true. Static ip works, dhcp would be convenient I guess. But I haven't figured out yet hot to get internet access on the laptop.
Mar 09 13:45:11 <DocScrutinizer05> when you already have a working connection and ip-tables forwarding on N900, you just need a route on your PC
Mar 09 13:45:36 <Shapeshifter> DocScrutinizer05: like the one in your paste? i.e. /sbin/route add -host 192.168.0.202/32 dev $DEV$
Mar 09 13:45:51 <Shapeshifter> What IP do I need to put there?
Mar 09 13:46:09 <DocScrutinizer05> similar, yes, however for default route
Mar 09 13:46:58 <DocScrutinizer05> default         192.168.4.1     0.0.0.0         UG    0      0        0 eth0
Mar 09 13:47:08 <DocScrutinizer05> 192.168.4.0     *               255.255.255.0   U     1      0        0 eth0
Mar 09 13:47:32 <bencoh> Shapeshifter: http://pastebin.notk.org/pastebin.php?show=d39d5e3db
Mar 09 13:47:37 <bencoh> here is what I use on my n900
Mar 09 13:48:02 <DocScrutinizer05> I assumed he already has that
Mar 09 13:48:12 <bencoh> oh and, do use mtr on your laptop to try and debug that kind of routing issues:)
Mar 09 13:48:48 <DocScrutinizer05> first his laptop needs to know that Neo900 is the default gateway
Mar 09 13:49:05 <bencoh> n900* unless I missed something ;)
Mar 09 13:49:20 <DocScrutinizer05> or he does >> ping 8.8.8.8 $ip-of-N900
Mar 09 13:49:33 <DocScrutinizer05> oops yep
Mar 09 13:49:37 <bencoh> dont try to ping n900
Mar 09 13:49:55 <bencoh> once you connect to cellular data it disables icmp answers
Mar 09 13:49:57 <DocScrutinizer05> no, $ip-of-N900 is the gateway to use, here
Mar 09 13:50:15 <DocScrutinizer05> might as well be the NIC name
Mar 09 13:50:16 <bencoh> icmp echo reply*
Mar 09 13:50:38 <DocScrutinizer05> bencoh: huh?
Mar 09 13:50:46 <DocScrutinizer05> only with tethering, no?
Mar 09 13:51:06 <Shapeshifter> It works
Mar 09 13:51:09 <Shapeshifter> thanks guys
Mar 09 13:51:11 <DocScrutinizer05> :-)
Mar 09 13:51:30 <Shapeshifter> ip r add to default via 192.168.2.15
Mar 09 13:51:46 <DocScrutinizer05> :-))
Mar 09 13:52:04 <bencoh> DocScrutinizer05: no
Mar 09 13:52:32 <DocScrutinizer05> Shapeshifter: could you check >> ping 192.168.2.15 << now?
Mar 09 13:52:41 <bencoh> /etc/network/if-post-down.d/zz_enable_icmp_echo_reply:19:echo 0 > /proc/sys/net/ipv4/icmp_echo_ignore_all
Mar 09 13:52:44 <bencoh> /etc/network/if-up.d/00_disable_icmp_echo_reply:25:echo 1 > /proc/sys/net/ipv4/icmp_echo_ignore_all
Mar 09 13:52:48 <Shapeshifter> DocScrutinizer05: ping doesn't work
Mar 09 13:52:50 <DocScrutinizer05> duh!
Mar 09 13:53:03 <bencoh> haha :)
Mar 09 13:54:06 <bencoh> I think they didnt want to do it through netfilter at first, but that lead them to do it system-wide because linux doesn't provide a way to do it per-interface without netfilter iirc
Mar 09 13:54:42 <DocScrutinizer05> ack
Mar 09 16:35:43 <Sicelo> 15:48 < DocScrutinizer05> first his laptop needs to know that Neo900 is the default gateway
Mar 09 16:35:54 <Sicelo> someone has Neo900? :p
Mar 09 16:42:05 <kerio> neo900 spotted in the wild, turns out it's just a reskinned 3g wifi access point
Mar 09 16:52:33 <parazyd> lol
Mar 09 18:02:11 <DocScrutinizer05> ((<bencoh> once you connect to cellular data it disables icmp answers)) alas this helps only marginally, since even with ICMP answers disabled, the GPRS afaik still ACKs the ICMP ping inbound data paket
Mar 09 18:02:48 <DocScrutinizer05> IOW GPRS enables TX even when data is strictly inbound-only
Mar 09 18:04:36 <DocScrutinizer05> which sucks big time since you can't stop a battery drain from 'rogue' inbound pakets, no matter what you do. Unless you disable GPRS completely
Mar 09 18:21:52 <DocScrutinizer05> https://github.com/systemd/systemd/issues/2402
Mar 09 18:21:59 <DocScrutinizer05> ECHAN
Mar 09 18:59:35 <bencoh> DocScrutinizer05: indeed
Mar 09 21:28:44 <Wizzup> Linux gentoo900 4.5.0-rc5+ #1 PREEMPT Wed Mar 9 18:58:02 CET 2016 armv7l Nokia RX-51 board GNU/Linux
Mar 09 21:28:48 <Wizzup> :)
Mar 09 21:32:46 <ecc3g> emerge -e world
Mar 09 21:50:52 <Pali> oo :-)
Mar 09 22:36:07 <L29Ah> 00:28:46]<Wizzup> Linux gentoo900 4.5.0-rc5+ #1 PREEMPT Wed Mar 9 18:58:02 CET 2016 armv7l Nokia RX-51 board GNU/Linux
Mar 09 22:36:07 <L29Ah> did you bring up distcc?
Mar 09 22:45:38 <Wizzup> L29Ah: no, why
Mar 09 23:10:51 <L29Ah> Wizzup: cuz not much ram to run g++ in
Mar 09 23:12:09 <Wizzup> L29Ah: I compile in a chroot on my chromebook (arm)
Mar 09 23:13:09 <L29Ah> oh ok
Mar 09 23:13:35 <L29Ah> are you using some portage magic for this or just throw binary packages or just throw filesystem images?
Mar 09 23:14:23 <L29Ah> anyways poke me when you get an usable ui :]
Mar 09 23:14:30 <Wizzup> $ arch
Mar 09 23:14:30 <Wizzup> armv7l
Mar 09 23:14:32 <Wizzup> ^laptop
Mar 09 23:14:34 <Wizzup> ;-)
Mar 09 23:14:41 <L29Ah> i've tried to package some maemo gui stuff but failed at lots of binary-only deps
Mar 09 23:14:56 <L29Ah> the ones i got are in ::booboo
Mar 10 00:03:56 <ecc3g> wizzup: curious, how long does the armv7l take to just run "emerge" with no arguments "time emerge"?  Also what's the clock speed on it?
Mar 10 00:05:30 <L29Ah> http://www.hardkernel.com/main/products/prdt_info.php?g_code=G145457216438 /me thinks about hooking his video glasses to this thing
Mar 10 00:05:35 <L29Ah> too bad only 2GB of RAM
Mar 10 00:23:20 <Oksana> Curious... Missing dependencies? http://maemo.org/packages/package_instance/view/fremantle_extras-devel_free_armel/navit/0.5.0+dfsg.1-1maemo1~6534/
Mar 10 01:02:29 <ecc3g> temptation...to buy another n900...
Mar 10 01:40:27 <Wizzup> ecc3g: I didn't run emerge on the n900, I only run it from a diff device
Mar 10 01:40:31 <Wizzup> the n900 doesn't even have a portage tree
Mar 10 01:40:40 <Wizzup> but on my laptop -- it's quite fast
Mar 10 01:41:02 <Wizzup> real	0m2.858s
Mar 10 02:06:44 <Luke-Jr> ecc3g: real    0m24.342s user    0m10.469s sys 0m0.680s
Mar 10 02:22:58 <Oksana> Anybody knows why aptitude segfaults on N900? /Actually seeing package dependencies tree would be nice/
**** ENDING LOGGING AT Thu Mar 10 02:59:58 2016