**** BEGIN LOGGING AT Sun Jan 16 02:59:58 2011
Jan 16 03:22:55 <builder> build #51 of ifxmips is complete: Success [build successful]  Build details are at http://tksite.gotdns.org:8010/builders/ifxmips/builds/51
Jan 16 13:07:39 <CIA-77> bud * r25020 /packages/utils/ntfs-3g/Makefile: bump ntfs-3g to latest stable
Jan 16 14:06:25 <builder> build #54 of uml is complete: Failure [failed shell]  Build details are at http://tksite.gotdns.org:8010/builders/uml/builds/54
Jan 16 15:40:31 <mazilo> xMff: Lots of SPAM posts by Pritam1192 (https://forum.openwrt.org/profile.php?id=44946).
Jan 16 16:02:24 <nbd> mazilo: fixed
Jan 16 17:10:58 <builder> build #61 of gemini is complete: Failure [failed shell_9]  Build details are at http://tksite.gotdns.org:8010/builders/gemini/builds/61
Jan 16 21:21:05 <philipp64|laptop> xMff: ping
Jan 16 21:22:00 <philipp64|laptop> I'm looking at http://wiki.openwrt.org/doc/uci/network#options.valid.for.all.protocol.types ... and wondering, "how does one specify the bridge name"?
Jan 16 21:25:34 <xMff> can't be specified
Jan 16 21:25:53 <xMff> bridge names are always br-foo where foo is the name of the corresponding config interface section
Jan 16 21:26:13 <xMff> ports are given in the ifname fields
Jan 16 21:26:18 <xMff> *field
Jan 16 21:40:04 <philipp64|laptop> ah... so I can have multiple bridges in a vlan environment. that's a loss.
Jan 16 21:40:09 <philipp64|laptop> *can't
Jan 16 21:40:29 <xMff> why not?
Jan 16 21:41:28 <xMff> http://paste.openwrt.org/d17d64cd3 - or do you mean something else?
Jan 16 21:43:39 <philipp64|laptop> was looking at http://wiki.openwrt.org/doc/uci/network#bridge.without.ip and it wasn't clear that
Jan 16 21:43:49 <philipp64|laptop> that 'example' becomes the bridge name.
Jan 16 21:46:24 <xMff> ok
Jan 16 21:58:24 <swalker> updated openwrt/upstream, https://home.comcast.net/~sdwalker/uscan/uscan.shtml
Jan 16 22:00:33 <philipp64|laptop> btw, where is using uci-defaults to pre-configure your network documented?
Jan 16 22:04:56 <xMff> there is no dedicated documentation for it.
Jan 16 22:05:12 <xMff> its basically throwing some uci batch commands into an /etc/uci-defaults/foo
Jan 16 22:05:27 <xMff> it gets executed once on firstboot and is then deleted by the system
Jan 16 22:14:56 <philipp64|laptop> is there a way to build host binary for uci so that I could run it on a script file in my checkout and see what file it would touch and how... without actually modifying a host file, of course.
Jan 16 22:15:20 <philipp64|laptop> it would be faster than repeatedly burning CF's and rebooting. :-)
Jan 16 22:16:18 <xMff> yes, just clone uci.git from nbd.name
Jan 16 22:16:22 <xMff> and compile it
Jan 16 22:16:36 <xMff> you might need to patch it for another directory root
Jan 16 22:16:59 <xMff> unless you want to install an /etc/config/ symlink on your host
Jan 16 22:17:21 <philipp64|laptop> hmm... taking a --destdir=/tmp arg would be handy.
Jan 16 22:17:28 <xMff> it has this afaik
Jan 16 22:17:38 <philipp64|laptop> why not just build host binaries with the version in-tree?
Jan 16 22:18:01 <xMff> you can do that too
Jan 16 22:18:26 <xMff> ah, there it is: uci -c some_dir/etc/config ...
Jan 16 22:18:54 <xMff> stage is always in /tmp/.uci though
Jan 16 22:19:07 <xMff> and state always in /var/state
Jan 16 22:19:08 <philipp64|laptop> who owns uci? I'd like to request that they have the option to build it for the host...
Jan 16 22:20:28 <OutBackDingo> philipp64|laptop: uci builds on linux
Jan 16 22:20:57 <philipp64|laptop> I know... I'm saying I'd like the makefiles for it to build both target and host versions.
Jan 16 22:21:33 <OutBackDingo> philipp64|laptop: because u want to run it on ur desktop/servers ?
Jan 16 22:21:53 <CIA-77> jow * r25021 /trunk/tools/ipkg-utils/ (Makefile patches/180-add_installed_size.patch): tools/ipkg-utils: calculate installed size from embedded data.tar.gz, makes opkg space checking more precise
Jan 16 22:22:24 <philipp64|laptop> bbecause I want to be able to debug my target/linux/x86/geos/default-files/etc/uci/wireless
Jan 16 22:23:38 <CIA-77> jow * r25022 /trunk/include/package-ipkg.mk: [include] set Installed-Size to 0 when preparing control, its substituted by ipkg-build later on
Jan 16 22:24:25 <philipp64|laptop> how do I generate a wifi-iface section without an instance name using UCI?
Jan 16 22:24:36 <xMff> uci add wireless wifi-iface
Jan 16 22:24:49 <xMff> uci set wireless.@wifi-iface[-1].foo=bar
Jan 16 22:24:50 <xMff> ,,,
Jan 16 22:25:24 <xMff> or: id=$(uci add wireless wifi-iface); uci set wireless.$id.foo=bar
Jan 16 22:25:35 <xMff> but thats not suitable for uci batch
Jan 16 22:28:25 <philipp64|laptop> no wonder I couldn't figure it out. :-(
Jan 16 22:28:45 <philipp64|laptop> get a chance to review the cs5535 patch?
Jan 16 22:28:55 <xMff> not yet
Jan 16 22:30:41 <CIA-77> jow * r25023 /branches/backfire/tools/ipkg-utils/ (Makefile patches/180-add_installed_size.patch): [backfire] merge r25021
Jan 16 22:32:04 <CIA-77> jow * r25024 /branches/backfire/include/package-ipkg.mk: [backfire] backport r25022
Jan 16 22:49:33 <philipp64|laptop> and "uci delete network lan" do delete the default lan interface section?
Jan 16 22:53:02 <xMff> yes
Jan 16 22:53:09 <xMff> but you have to specify it with dot
Jan 16 22:53:14 <xMff> network.lan
Jan 16 22:57:22 <philipp64|laptop> weird...  my etc/uci-defaults/wireless file had no effect.
Jan 16 22:57:30 <xMff> was it executable ?
Jan 16 22:58:37 <philipp64|laptop> in the source tree it wasn't. but neither was my "network" file, which was executed.
Jan 16 22:58:42 <xMff> true
Jan 16 22:58:45 <philipp64|laptop> since it got deleted, I can't tell if it was...
Jan 16 22:58:55 <xMff> should still be in /rom
Jan 16 22:59:13 <xMff> however, keep in mind that the wifi config is generated after uci-defaults ran
Jan 16 22:59:22 <xMff> unless it exists already
Jan 16 22:59:49 <xMff> so you should create an /etc/config/wireless from your uci-defaults/wireless
Jan 16 23:02:54 <philipp64|laptop> what's the point of generating later?
Jan 16 23:04:32 <xMff> it is not known in advance which wifi might be present on the system
Jan 16 23:15:14 <philipp64|laptop> xMff: gpio loaded, and I'm running 2.6.37 on geos.
Jan 16 23:24:14 <philipp64|laptop> hmm... how come there's no /var/log/messages file?
Jan 16 23:24:39 <xMff> syslog is using an in-memory ring buffer by default
Jan 16 23:29:42 <philipp64|laptop> no output
Jan 16 23:29:53 <xMff> see logread
Jan 16 23:31:24 <philipp64|laptop> nothing.
Jan 16 23:31:50 <philipp64|laptop> did I need to enable anything to build ppp?
Jan 16 23:32:21 <xMff> pppd is a default package but ppp-mod-pppoa is not
Jan 16 23:32:26 <xMff> did you select it?
Jan 16 23:32:47 <philipp64|laptop> yup.
Jan 16 23:33:15 <xMff> so no info in logread and no running pppd ?
Jan 16 23:33:21 <xMff> and no error in ifup either?
Jan 16 23:34:40 <philipp64|laptop> correct.
Jan 16 23:35:19 <xMff> thats odd
Jan 16 23:35:32 <philipp64|laptop> want to look in the box?
Jan 16 23:35:40 <xMff> would help, yes
Jan 16 23:35:45 <philipp64|laptop> I'm wondering if our PPP changes broke something...
Jan 16 23:36:02 <xMff> I never had access to a pppoa setup
Jan 16 23:36:49 <philipp64|laptop> do I have your ssh public key?
Jan 16 23:56:50 <philipp64|laptop> working on it.
Jan 16 23:58:13 <philipp64|laptop> what's the default root password?
Jan 16 23:58:20 <xMff> its empty
Jan 16 23:58:27 <xMff> telnet
Jan 16 23:58:32 <xMff> passwd
Jan 16 23:58:43 <philipp64|laptop> eh?
Jan 16 23:58:46 <philipp64|laptop> I'm trying to ssh.
Jan 16 23:58:54 <xMff> there is no root password
Jan 16 23:59:02 <xMff> you need to telnet first set one with passwd
Jan 16 23:59:08 <xMff> this disables telnet and allows you to ssh
Jan 17 00:01:46 <philipp64|laptop> ok now.
Jan 17 00:02:26 <xMff> okay. connection works but now it wants a apssword
Jan 17 00:03:40 <philipp64|laptop> not sure why it's ignoring the .ssh/authorized_keys file.
Jan 17 00:05:51 <xMff> /etc/dropbear/authorized_keys
Jan 17 00:05:56 <xMff> for root
Jan 17 00:11:35 <xMff> found it
Jan 17 00:11:51 <xMff> we did indeed broke it
Jan 17 00:11:55 <xMff> but no biggie
Jan 17 00:14:21 <philipp64|laptop> what do I need to do?
Jan 17 00:14:36 <xMff> I patched it locally, send you a diff in a few seconds
Jan 17 00:14:50 <xMff> and I lost access, might be due to pppoa being established
Jan 17 00:15:54 <xMff> patch: http://paste.openwrt.org/d1d8ef965
Jan 17 00:16:08 <xMff> we need to fake a .device option since "ifup" needs it
Jan 17 00:17:41 <xMff> one could workaround it by adding "option device foo" in the network config but thats ugly
Jan 17 00:17:50 <xMff> therfore I synthesize it in the script
Jan 17 00:17:59 <xMff> its set to a real value later on if pppd is done
Jan 17 00:18:41 <philipp64|laptop> I've been arguing with Karl that dsl needs to always have a kernel net_device... for various reasons... including this.
Jan 17 00:18:57 <philipp64|laptop> plus it's nice to have a net_device to send netlink messages from.
Jan 17 00:19:23 <CIA-77> jow * r25025 /trunk/package/ppp/ (Makefile files/pppoa.sh): [package] ppp: synthesize a .device option to make ifup work on pppoa interfaces again
Jan 17 00:20:55 <xMff> well, the 6in4, 6to4 etc. packages have the same problem
Jan 17 00:21:26 <xMff> they create a layer 3 tunnel which is not connected to any physical iface, so they have no netdevice either and need to fake one too to make the script machinery happy
Jan 17 00:21:53 <xMff> thats the whole purpose of the scan_$proto() fook
Jan 17 00:21:55 <xMff> *hook
Jan 17 00:22:05 <philipp64|laptop> foo+hook = fook
Jan 17 00:24:14 <philipp64|laptop> ok, now if I just knew why wireless wasn't coming up.
Jan 17 00:24:56 <philipp64|laptop> how do I force my bridge to be renamed from br-br1 to 'lan'?
Jan 17 00:25:29 <xMff> config interface br1 -> config interface lan
Jan 17 00:26:39 <CIA-77> jow * r25026 /branches/backfire/package/ppp/ (Makefile files/etc/hotplug.d/atm/20-atm-modem files/pppoa.sh): [backfire] merge r23530, r23531, r23551 and r25025
Jan 17 00:26:43 <philipp64|laptop> the bridge code seems to want to add a prefix.
Jan 17 00:26:53 <xMff> yes
Jan 17 00:27:08 <philipp64|laptop> grr.
Jan 17 00:27:25 <philipp64|laptop> same with the pppoa stuff.
Jan 17 00:27:29 <philipp64|laptop> pppoa-$name
Jan 17 00:28:17 <xMff> yes, it simplifies a lot of scripts. otherwise one needs to lookup arbritary names every time
Jan 17 00:29:04 <philipp64|laptop> sigh. it makes other configs need to be more topology-aware, though.
Jan 17 00:29:23 <xMff> true
Jan 17 00:29:37 <philipp64|laptop> anyway, how do burn a new flash image onto a running system?
Jan 17 00:30:10 <xMff> on x86 you can usually just dd it from within /tmp to /dev/sda or however the storage presents itself
Jan 17 00:30:47 <philipp64|laptop> I'm using openwrt-x86-geos-combined-squashfs.img...  so I can just ssh host "dd if=- of=/dev/sda ..."
Jan 17 00:31:00 <xMff> yes
Jan 17 00:31:06 <xMff> leave out the if=-
Jan 17 00:31:17 <xMff> I'm unsure whether busybox dd supports this notation
Jan 17 00:31:22 <xMff> it will read from stdin by default
Jan 17 00:35:51 <philipp64|laptop> what do I need to do to fix that?
Jan 17 00:36:10 <xMff> iptables -I INPUT -p 22 -j ACCEPT  for example
Jan 17 00:37:03 <xMff> the "official" way is here: http://wiki.openwrt.org/doc/uci/firewall - but you're free to stick arbritary commands into /tc/firewall.user
Jan 17 00:38:12 <philipp64|laptop> ok, so no config knobs for the firewall.
Jan 17 00:38:19 <xMff> see link
Jan 17 00:38:32 <xMff> http://wiki.openwrt.org/doc/uci/firewall#opening.ports
Jan 17 00:39:22 <xMff> unless I misunderstood your remark
Jan 17 00:48:09 <philipp64|laptop> yup, that did it. opened the port and burning a new image.
Jan 17 00:48:51 <xMff> indeed, key works too now
Jan 17 00:49:03 <xMff> it looks fine to me
Jan 17 00:50:13 <philipp64|laptop> yeah... forgot to add uci-defaults file for firewall to image... rebuilding and reburning...
Jan 17 00:50:20 <xMff> heh
Jan 17 00:50:42 <philipp64|laptop> seems wifi isn't up...
Jan 17 00:51:29 <xMff> yep
Jan 17 00:52:05 <xMff> you need to set option type mac80211
Jan 17 00:52:11 <xMff> and option phy phy0
Jan 17 00:52:23 <xMff> you specified atheros but thats only valid for madwifi
Jan 17 00:52:50 <philipp64|laptop> ah.
Jan 17 00:54:37 <philipp64|laptop> I knew what you meant. btw, I rebuild via:
Jan 17 00:59:44 <builder> build #66 of ubicom32 is complete: Failure [failed compile_4]  Build details are at http://tksite.gotdns.org:8010/builders/ubicom32/builds/66
Jan 17 01:03:43 <swalker> ^ anyone want to terminate that struct?
Jan 17 01:04:38 <xMff> psst, qanyone is busy ignoring this error :P
Jan 17 01:06:01 <swalker> yep
Jan 17 01:11:40 <swalker> ep93xx: missing 2.6.36 KEYBOARD_EP93XX symbol, gemini: transient prep_logs.sh failure, cris: gcc 4.3.5 assembler complexity, pxcab, ps3: gcc 4.3.5 ICE, ubicom: missing pata_artop struct pci_device_id termination, uml: rm -fvr non zero return
Jan 17 01:13:41 <xMff> some of them can be only fixed by thepeople
Jan 17 01:14:03 <xMff> I already suggested to protect the scripting steps with || true
Jan 17 01:17:56 <swalker> das-Iro: expect is missing on your first buildslave
Jan 17 01:27:50 <swalker> and the second one and thepeople's buildslave2 fwiw
Jan 17 02:18:12 <CIA-77> jow * r25027 /trunk/package/ppp/files/pppoa.sh: [package] ppp: remove pppoa coldplug, the hotplug handler has been confirmed for ueagle and solos
Jan 17 02:18:42 <xMff> gr +working
Jan 17 02:19:37 <CIA-77> jow * r25028 /trunk/package/ppp/Makefile: [package] ppp: bump pkg revision after latest changes
Jan 17 02:20:34 <CIA-77> jow * r25029 /branches/backfire/package/ppp/ (Makefile files/pppoa.sh): [backfire] merge r25027 and r25028
Jan 17 02:50:47 <mazilo> xMff: The lib/xmlrpc_c package failes to compile on r25018. Here is the compilation dump (http://pastebin.com/WbbWPPZ2). See line #352
**** ENDING LOGGING AT Mon Jan 17 02:59:57 2011