**** BEGIN LOGGING AT Sat May 26 02:59:58 2012 May 26 03:01:23 me too May 26 03:01:47 Well $14 bucks lost if I did it wrong. May 26 03:02:10 Just placed an order on seeedstudio May 26 03:45:04 Anyone got some cool beaglebone projects May 26 03:55:18 I was think of interfacing it with a minecraft server May 26 03:58:41 To do what? May 26 03:59:42 Not quite sure yet May 26 03:59:49 lol cool. May 26 04:00:07 I just made a custom PCB to interface alarm sensors to the beaglebone. May 26 04:00:21 Basically it's a low pass filter May 26 04:00:24 http://dl.dropbox.com/u/7323462/Screen%20Shot%202012-05-25%20at%2010.12.25%20PM.png May 26 04:01:09 I'm going to make a home alarm system with logging and statistical analysis all with the bone. May 26 04:33:22 huh May 26 04:33:28 raspberry ordered May 26 04:33:54 i didnt know that it dosnt support newest android :( May 26 04:34:18 maybe i should order bone and pandaboard es May 26 05:10:10 this isn't #r-pi May 26 05:14:32 YEAH May 26 05:14:40 those XBMC faries May 26 13:48:02 that might be a nice gimmick for the bone: http://gcc.godbolt.org/ May 26 13:51:06 <_troll_> why the bone? May 26 13:51:40 <_troll_> even that site has arm already May 26 13:51:52 because it already has a web-gui May 26 13:53:09 yes, changing march and you get output May 26 13:53:24 funny thing May 26 13:54:28 at least for those who like looking at asm May 26 13:54:59 <_troll_> I don't like looking at asm as output by gcc May 26 13:55:23 one could change it to actually produce exectuables from single-source-files May 26 13:55:27 <_troll_> compiling to .o and disassembling gives more readable output May 26 13:56:30 riskable: my clean build got all the way to the rootfs stage for cloud9-image May 26 13:56:42 hmm, I wonder is that is hack-proof. ;) May 26 13:56:50 s/is/if/ May 26 13:57:09 <_troll_> as long as it doesn't _run_ the code it should be fine May 26 13:57:36 <_troll_> unless they were _really_ stupid May 26 13:57:43 it calls the compiler with arbitrary arguments May 26 13:58:07 <_troll_> sandboxing the compiler should be easy May 26 13:58:41 anyway, nice gimmick May 26 14:00:25 riskable: note, by oe-build-core setup is pulling submodules from a couple days back, so it seems there is some issue with your setup-scripts tree May 26 14:13:34 cbrake: Yeah, my build of cloud9-image is almost done and there's no errors so far... The problem lives somewhere in the whole setup-scripts/oebb.sh process May 26 14:13:49 Your oe-build-core method is much better :) May 26 14:15:17 FYI: It wouldn't be much effort at all to build an interface like gcc.godbolt.org using Gate One May 26 14:19:43 riskable no problem is that some recipes in oe-core got removed May 26 14:20:03 and the other layers did not remove the overrides appends for it May 26 14:20:31 I am sure koen will fix it when he see it May 26 14:20:56 woglinde: I don't understand why that would break the oebb.sh method but not oe-build-core May 26 14:21:57 As I understand it, oebb.sh just does a 'git clone' of all the requisite repos whereas oe-build-core uses git remotes inside a single local repo May 26 14:22:12 ...but that shouldn't make a difference, really May 26 14:23:59 Build just finished... Getting an error: ERROR: Task 8 (/storage/beagleboard/oe-build-core/sources/meta-ti/recipes-misc/images/cloud9-image.bb, do_rootfs) failed with exit code '1' May 26 14:24:26 ...but looking at the output I suspect I just need to add a loopback filesystem May 26 14:24:37 Easy enough May 26 14:24:46 riskable oebb.sh sets you up oe-core and the layers koen likes May 26 14:25:02 you can always edit to your fashion May 26 14:28:43 ...but shouldn't 'cloud9-image' build a pre-defined set of layers and whatnot? May 26 14:31:27 Regarding the PREFERRED_PROVIDER variable... Is that something I should set in my environment or should I set it somewhere else? May 26 14:31:30 o.O May 26 14:31:45 no layers are repositories for recipes May 26 14:31:48 ntp-ssl sounds cool (always just used regular ntp or rdate =) May 26 14:32:10 RIght, but the recipes are defined by the cloud9-image.bb, right? Unless the layers override things? May 26 14:32:16 what you means are packages which got compiled while cloud9-image has it as depencendies May 26 14:32:37 no the recipes are defined on ther own May 26 14:32:44 I see May 26 14:33:05 you have layers, distrubition, recipes, images May 26 14:33:20 where images are sort of recipes May 26 14:39:33 I just haven't poked around enough May 26 14:40:00 I'll figure all this stuff out. I only just got my Beaglebone on Monday but I didn't start playing around with the build process or recipes until yesterday May 26 14:41:03 sure May 26 14:42:58 I am playing with the stuff since 2004 May 26 14:43:39 Once I've got the hang of things I'll be able to contribute more May 26 14:44:08 sure May 26 14:44:19 I am the sloppy maintainer of meta-java May 26 14:44:34 I used to mess around with OpenWRT a lot May 26 14:44:53 In fact, I made an entire OpenWRT-based distribution called PyCI May 26 14:45:10 openwrt is okay May 26 14:45:15 (it was for the RouterStation Challenge contest which I won) May 26 14:45:25 uhm May 26 14:45:30 what was the goal? May 26 14:45:46 I wrote my master thesis about wlan location stuff May 26 14:45:49 They wanted a good-looking GUI for configuring OpenWRT for their RouterStation platform May 26 14:45:56 lol May 26 14:46:13 okay May 26 14:46:26 The requirements list was a mile long. I suspect the primary reason I one was because my entry was the only one that actually met all the requirements May 26 14:46:36 s/one/won May 26 14:47:10 I thought of a technical challenge May 26 14:47:33 I also wrote a layer 2 telnet server/client. I've been meaning to re-visit that to make it A) secure and B) more robust May 26 14:48:01 uhm why layer2 May 26 14:48:09 That was one of the requirements: A layer 2 client/server so you could rescue a router with a hosed IP or firewall config May 26 14:48:11 <_troll_> secure telnet? is it contradiction day already? May 26 14:48:23 Well, I'd change it so it was more like SSH over layer 2 May 26 14:48:48 <_troll_> and yes, I'm aware of thinks like kerberised telnet May 26 14:48:54 <_troll_> *things May 26 14:49:01 Yeah, no one uses that these days May 26 14:49:40 <_troll_> they use ssh with kerberos authentication instead May 26 14:49:44 <_troll_> almost the same thing May 26 14:50:04 I hope someone will step up todo gssapi for libssh2 May 26 14:50:39 Kerberized telnet is still insecure. It is just the authentication part that isn't :) May 26 14:50:41 so I can use it for x2go from windows May 26 14:51:07 <_troll_> riskable: depends on configuration May 26 14:51:10 Well GSSAPI won't work without a Kerberos implementation... Does OE have that? I haven't looked May 26 14:51:41 * riskable is assuming that you mean GSSAPI Kerberos auth because no one uses GSSAPI for anything else :) May 26 14:52:09 My day job involves a lot of Kerberos work May 26 14:52:56 riskable sure kerberos May 26 14:53:02 and kerberos is on oe May 26 14:53:10 damn May 26 14:53:18 forgot to patch mkbc again May 26 14:53:21 or call May 26 14:53:33 halfhour useless compiling May 26 14:53:33 I'm one of the few people at my (enormous) employer that knows anything at all about Kerberos so I end up being consulted/having to troubleshoot everything Kerberos-related May 26 14:53:55 riskable hm ask your boss to put the support in libssh May 26 14:54:05 args libssh2 May 26 14:54:27 I am to slow to do it May 26 14:54:28 I'm not much of a C/C++ guy May 26 14:54:35 hm okay May 26 14:54:59 The only thing I've used C for recently is Arduino/microcontroller stuff May 26 14:55:19 find . -name krb5* --> Returns no results in the sources dir May 26 14:55:47 find . -name kerberos* --> Only returns some stuff related to javax May 26 14:55:52 <_troll_> maybe it's called something else May 26 14:56:11 git grep May 26 14:56:22 * _troll_ lost track of kerberos implementations some years ago May 26 14:56:35 there is only heimdal and mit May 26 14:56:41 Nothing turns up for just krb* either... 'git grep kerberos' and 'git grep krb' return nothing May 26 14:56:49 hm May 26 14:57:09 <_troll_> oh, so nothing changed in 10 yeaars May 26 14:57:11 so maybee we dont have it May 26 14:57:20 troll right May 26 14:57:30 2003 kerberos came to windows May 26 14:57:54 <_troll_> iirc it was an option a bit before that May 26 14:57:57 Microsoft's Kerberos implementation is half-assed May 26 14:58:51 Not only that but all of Microsoft's own client components to Kerberos automatically fall back to NTLM by default May 26 14:59:21 So 99% of the time Windows SAs will set up IIS thinking Kerberos and proper Single Sign-On are working when they're really not May 26 14:59:27 They almost always forget to add SPNs May 26 14:59:32 _troll_ sure but not integrated May 26 14:59:40 was third party May 26 14:59:54 ...and debugging such things with IE is a waste of time since it provides absolutely zero useful information about auth May 26 15:00:11 <_troll_> I recall kerberos being used for windows auth at my university and I spent very little time there after 2003 May 26 15:00:25 <_troll_> and certainly not on the windows systems May 26 15:00:36 By default when you join a host to Active Directory it uses Kerberos May 26 15:00:37 I was happy when I got putty working with kerberos support to an linux machine which was part of windows-domain May 26 15:00:57 From that point on the host will use a keytab to authorize itself with AD for performing queries May 26 15:01:22 woglinde: Yeah, it's fun the first time. Wait until you have to troubleshoot that on a day-to-day basis. Then it sucks! May 26 15:01:24 riskable only ticket forwarding I did not understand well May 26 15:01:42 woglinde: Yeah, 'delegation' is a bit confusing May 26 15:01:45 riskable I spent some weeks on it May 26 15:02:05 and I wrote some stuff so you could use it without samba May 26 15:02:14 woglinde: If you enable ticket forwarding on all your hosts you can SSH into one and from that host SSH into another without having to re-enter your password. Your ticket will just be forwarded along until it expires May 26 15:02:40 riskable sure that I understood May 26 15:02:49 but not how it technical works May 26 15:03:36 Think of your ticket like an SSL certificate. When you create it you can attach certain flags. One of those flags is a boolean, "forwardable" May 26 15:03:52 I did a loud lol when I setup my first kerberos realm and was able to mount some user homes at ethz May 26 15:04:12 hi emeb May 26 15:04:43 Hi woglinde. Wassup? May 26 15:04:51 nothing May 26 15:05:00 Kerberos is wicked complicated and all the well-defined error messages (with specific codes) are as cryptic as can be May 26 15:05:08 only gossiping about kerberos May 26 15:05:32 Never thought much about kerberos. May 26 15:05:32 riskable yes May 26 15:05:40 emeb sso May 26 15:05:42 Here's a tip... Write this down for reference when you're pulling your hair out troubleshooting a, "wrong principal" error: FQDN *must* come first in /etc/hosts! Anything else is a misconfiguration May 26 15:05:53 when you want single sign on May 26 15:06:01 riskable sure I know this May 26 15:06:34 It seems that every couple of weeks at my day job I have to debate some SA about why the FQDN must come first. Drives me batty! May 26 15:06:48 making the right certs with the right crypto algo is fun too May 26 15:06:58 I need to start an advertising campaign about it, "Fight the short hostname!" May 26 15:07:16 woglinde: Nothing says "Fun!" like trying to get any given application working with TLS 1.2! May 26 15:08:51 WIth all the SSL vulnerabilities as of late only TLS 1.2 is considered secure May 26 15:09:05 Not only that but anything less than 4096-bit RSA is also considered insecure May 26 15:09:25 ECC is recommended but support for ECC is still lacking in most apps May 26 15:10:10 Hi, I'm trying to boot my beaglebone via UART. I'm sending the bootloader image via XMODEM, but I'm only getting a 'Give your local XMODEM receive command now'. Should I be doing something else to complete the transfer? May 26 15:10:29 Though, honestly, ECC is vastly superior on all fronts. I love how tiny ECC certs and keys are :) May 26 15:10:34 I'm using minicom May 26 15:11:48 riskable whats ecc? May 26 15:11:59 Elliptic Curve Cryptography May 26 15:12:04 ieehks May 26 15:12:17 snake oil ellliptic May 26 15:12:41 ECC is an mathematical process like RSA is... But there's more nuance. May 26 15:13:29 <_troll_> all public key crypto is based on functions for which the inverse is hard to compute May 26 15:13:38 Yeah, P != NP May 26 15:16:00 What's the proper way to run menuconfig from inside oe-build-core? bitbake... Something I assume? May 26 15:16:29 riskable you want do kernel config? May 26 15:16:48 Yep, I want to add some additional USB drivers that don't seem to be present in the default kernel May 26 15:17:08 hm you need to copy the .config file afterwards for persistence May 26 15:17:19 bitbake -c menuconfig linux should work May 26 15:18:00 Thanks May 26 15:18:19 I'll do it once this cloud9-image build is done... Been running forever just doing the image part May 26 15:18:54 opkg-cl has been running a long time but it isn't using up much CPU or memory. Just a little bit here and there. I wonder what the bottleneck is? Hmm May 26 15:22:30 yeah opkg is a bit fishy sometimes May 26 15:23:46 Is there a list of layers somewhere? I mean, ones I could add May 26 15:24:11 on the oe wiki site May 26 15:24:28 but koen has added all I know off in setup-scripts May 26 15:25:16 Thanks May 26 15:25:59 only this is missing May 26 15:26:00 https://gitorious.org/sysmocom-openembedded May 26 15:34:36 Thanks May 26 15:34:36 I'm heading out. I'll be back later tonight May 26 16:37:33 uhh how do you apply patches like this http://code.google.com/p/android-x86/issues/attachmentText?id=670&aid=6700000000&name=0001-Faytech-Automotive-Touchscreen&token=-QwYU1OnpIAMLdo1iEXVtVnZmpI%3A1338049897888 ? May 26 16:41:50 Borillion: it's a git patch, but diff will understand it. May 26 16:42:08 how do I do it just diff May 26 16:42:42 cd kernelfolderhere; diff -p1 < filenamehere May 26 16:43:03 The -p1 will strip the 1st folder name from every path (to get rid of the filler a/ and b/) May 26 16:43:53 kk thanks May 26 16:44:03 If it says it rejected the patch, your kernel source is too distanced from the source the author of your patch worked on, and you'll need to either really dig in or find this author's source. May 26 16:46:20 The patch is from 2009 so I can't imagine you'll have success May 26 16:46:45 2012, wow, my eyes really tricked me. Nvm May 26 16:46:53 hmm diff -p1 < fayatech May 26 16:46:53 diff: missing operand after `-p1' May 26 16:46:53 diff: Try `diff --help' for more information. May 26 16:47:01 sorry multipost May 26 16:49:00 Oh my gosh. patch, not diff. Sorry. May 26 16:49:12 its ok May 26 16:52:51 nm got it May 26 16:52:55 thanks javajake May 26 16:53:07 now for building ics :P May 26 16:54:50 Nice :) May 26 17:05:57 getting barked at about "The correct version is: Java SE 1.6." May 26 17:06:24 I have 1.6.0_24 shouldn't that be sufficient? May 26 17:06:42 Yep May 26 17:06:57 how do you get around that? May 26 17:07:02 not finding anything May 26 17:08:38 I haven't worked with ics source, so I can't say. You may have the wrong implementation of Java, or it may not be finding the right Java in your system. May 26 17:08:57 If you're using Sun Java I can imagine it might whine ;) May 26 17:10:43 hmmm May 26 17:10:57 Borillion: http://forum.xda-developers.com/showthread.php?t=1386615&page=10 <-- this guy had similar issues and noted that ics wanted OpenJDK, not Sun. May 26 17:11:25 Oh, interesting, the official build instructions say Sun Java is usable. May 26 17:11:27 I don't really know. May 26 17:27:36 I always use Oracle/Sun java over OpenJDK Java. OpenJDK seems to corupt my minecraft server maps May 26 17:40:34 raymohi intressting May 26 17:44:57 hm thats looks good May 26 17:44:58 NOTE: package openjdk-7-jre-03b147-2.1+2.11-devel+hg1+ae30f070f936-r0.0: task do_package_write_ipk: Started May 26 17:58:50 javaJake, appears that openjdk 7 works just fine with ICS but isn't either recommended or just not on googles instructions May 26 17:59:52 scratch that, lots of places say it doesn't work May 26 18:25:46 Hello! I'm running Ubuntu on Beaglebaord B4 and trying to get usb storage device working. I have both Beagleboard and USB hub externally powered, still I'm not able to get access to usb storage device. http://paste.ubuntu.com/1008505/ Am I missing something? May 26 18:47:40 orated whats lsusb saying? May 26 18:48:56 Good afternoon beaglers May 26 18:49:40 woglinde: I've already included that in the paste :) Still, it says May 26 18:49:41 $ lsusb May 26 18:49:43 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub May 26 18:49:44 Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub May 26 18:50:27 oh sorry May 26 18:51:01 No problem :) What do you suggest? May 26 18:51:36 orated try to connect a keyboard May 26 18:51:44 and see if it is regconized May 26 18:52:29 Yes, I've to try that but I don't have any usb keyboard/mouse at the moment. Only one wireless mouse :/ May 26 18:53:16 o.O May 26 18:53:21 connect the wireless mouse May 26 18:54:22 wireless mouse?.. its gonna give driver issues I guess. Moreover, how will I test mouse? Its all cli at the moment.. or only to see if it detects using dmesg/lsusb? May 26 18:54:39 sure May 26 18:54:58 but so we can see if something is recognized at all May 26 18:55:07 I got something like this - http://upload.wikimedia.org/wikipedia/commons/6/69/Microsoft-wireless-mouse.jpg May 26 18:55:11 Sure, just a momnet May 26 18:58:35 woglinde: Connected the wireless mouse - http://paste.ubuntu.com/1008558/ And the mouse is trying to get connected, still blinking led May 26 18:59:33 Hub is getting detected, so maybe do I need to modprobe for usb storage devices? I'm not sure May 26 18:59:39 as you see no usb device is reconized at all May 26 18:59:52 try to connect the mice directly May 26 19:02:18 No, no change - http://paste.ubuntu.com/1008563/ May 26 19:03:34 The BB B4 got a USB OTG port which I have connected to USB type B female to female adapter which is then connected to USB hub/mice May 26 19:03:56 I also got USB-ethernet adapter, do you want me to test that? May 26 19:07:26 hm ah bb b4 May 26 19:07:36 okay I think it dont works on this May 26 19:08:18 I mentioned B4 initially :) There is no separate USB port on B4 May 26 19:09:03 But the manual suggests to connect in the way I said above May 26 19:09:05 yes May 26 19:10:45 woglinde: Is there a way to test usb-ethernet on the same? May 26 19:17:57 Okay, thanks for your input woglinde May 26 19:43:39 Hey, anyone here know of a tablet-esque handheld or device that is capable of being carried around (and obviously can have Linux installed)? May 26 19:44:19 did you even try google? May 26 19:45:58 javaJake: How about literally any Android tablet or phone? May 26 19:46:03 Android is based on Linux. May 26 19:46:09 Yes, I'm finding a lot of older, underpowered devices. I'm still looking but oftentimes there's that quiet hobbyist corner that did something cool :) May 26 19:46:33 There used to be Chumby. May 26 19:46:49 THere exist modules for the BeagleBoard and PandaBoard for LCDs and touchscreens. May 26 19:46:52 agmlego: I've heard bad things about Android tablets but I will take a look. Now that I think about it, the negative part of the reviews usually have to do with the vendor software. May 26 19:47:15 Also, there is no such thing as underpowered, until a use case is given. May 26 19:47:22 True :) May 26 20:16:51 Hi May 26 20:17:07 I have a question regarding the BBxM May 26 20:17:22 Ask. May 26 20:17:26 javaJake: the Android tablet problems are 99.9% software May 26 20:17:39 I've got one, and I've installed almost every OS I can on it, but I can't seem to get any video playback May 26 20:17:54 Azertify: hardware-accelerated, you mean? May 26 20:17:59 Yes May 26 20:18:37 mind clarifying the symptoms? May 26 20:18:53 mplayer won't show video, but does audio May 26 20:18:58 hmm May 26 20:19:45 I don't understand why, is there a driver of some sorts for hardware accelerated video playback? May 26 20:19:54 Yeah. May 26 20:19:55 probably May 26 20:20:01 Oh May 26 20:20:02 Hello! Is right command to configure ethernet if usb-ethernet adapter is used? May 26 20:20:09 Azertify: Did you have the same issues on Angstrom? May 26 20:20:35 Azertify: try running glxgears May 26 20:20:40 I overwrote angstrom before I noticed it was on it. May 26 20:20:44 and pastebin the output of lsmod May 26 20:21:08 Okay, I'll try to reinstall angstrom on it, then I'll give you what I find. May 26 20:21:22 Should I use the demo image? May 26 20:21:44 Azertify: Grab the latest Angstrom image off the site and put it on a card, rather than dealing with narcissus. May 26 20:21:47 Yeah. May 26 20:22:54 Right. Do I need to download the MLO, uboot, etc or are they in the tar file? May 26 20:23:26 Unsure. May 26 20:23:32 Download and see? May 26 20:23:34 I'll take a look May 26 20:23:58 Crappy connection so it'll take me a while to download, I'll take a shower whilst I wait. May 26 20:24:11 (yes it's that bad) May 26 20:24:19 I understand completely. May 26 20:57:12 a handy ARM log2(r0): clz r0, r0; rsb r0, r0, #31 May 26 21:02:04 <_troll_> XMPPwocky: if you're using gcc, __builtin_clz() is your friend May 26 21:02:33 _troll_: pure assembly, actually May 26 21:07:30 rootfs stage on cloud9-image is taking a long time. looks like opkg-cl is still working ... May 26 21:08:56 cbrake no its a bug somewhere May 26 21:09:23 * cbrake looks at the log file ... May 26 21:12:11 that it takes so long May 26 21:12:27 Hello! I'm now running Angstrom on Beagleboard B4 and unable to access usb storage. lsusb returns - Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub twice and fdisk -l on root@beagleboard:~# prompt gives nothing. Am I missing something? May 26 21:16:16 orated sudo May 26 21:16:29 add sudo to before fdisk May 26 21:17:09 sudo in Angstrom? -sh: sudo: not found May 26 21:17:32 KeatonT: Angstrom runs as root by default. May 26 21:17:52 I thought sudo was installed. May 26 21:18:38 But you are running as root. May 26 21:18:48 sudo is not necessary if you are already root. May 26 21:18:51 It is a no-op. May 26 21:19:19 I've almost finished installing Angstrom now May 26 21:19:51 God I hope this works, I followed the instructions right but the boot loader always seems to end up going wrong for me. May 26 21:19:56 KeatonT no its a kernel or hw problem no usb stuff is regconized so sudo will not work May 26 21:20:19 ls when in /media returns - card/ hdd/ mmcblk0p1/ net/ realroot/ cf/ j��1/ mmcblk0p2/ ram/ union/ ... Is there anything like modprobe usb_storage command required? May 26 21:20:32 orated whats your kernel? May 26 21:20:57 Linux beagleboard 2.6.32 #3 PREEMPT Tue Jul 26 10:56:56 CEST 201�1 armv7l unknown May 26 21:21:09 hm thats old May 26 21:21:27 Latest is 3.2.2 isn't it? May 26 21:21:51 Then again, I haven't checked for a while, I may be wrong. May 26 21:22:00 Azertify: Not on ARM, I do not think. May 26 21:22:04 orated what was your rev again? May 26 21:22:13 woglinde: Same, BB B4 May 26 21:22:20 Oh yeah, I forgot that the versions are different. May 26 21:22:41 The only difference now is that I'm trying Angstrom instead of Ubuntu, woglinde May 26 21:23:30 I would never run ubuntu on beagle May 26 21:23:47 Do you recommend angstrom? May 26 21:23:55 Yes. May 26 21:24:10 Because it is more or less built for the OMAP boards. May 26 21:24:17 Awesome May 26 21:24:20 And whatever the bone is using. May 26 21:24:43 Ubuntu runs OK, but is compiled for the ARM in general, not the ARM7 in particular. May 26 21:25:15 Really, if you are going to go off Angstrom, I would do something like gentoo, and use a different machine for the cross-compilation of packages. May 26 21:25:40 Is there a full list of package names for beagleboard xm drivers? May 26 21:25:46 http://elinux.org/BeagleBoard#USB May 26 21:26:12 orated you need to shorten some pins so the otg gets the right mode May 26 21:26:30 Ah, "usb EHCI has been removed because of a hardware defect" May 26 21:28:08 orated newer kernel may help May 26 21:28:10 let me see May 26 21:28:33 woglinde: HUB and the usb-ethernet connected to it is detected May 26 21:28:45 dmesg | grep hub May 26 21:28:47 er May 26 21:29:38 woglinde: log file: https://gist.github.com/2795399 May 26 21:29:40 brb May 26 21:30:02 woglinde: IMAGE_FSTYPES is out of control May 26 21:30:33 doh, my local.conf is the culprit :-\ May 26 21:30:46 thought I had already fixed, but that was another build dir May 26 21:30:53 * nextime uses debian May 26 21:31:11 ( armhf ) May 26 21:32:49 Yeah, that is even further from the ARM7 than Ubuntu May 26 21:37:13 orated hm I would the latest angstroem distro with kernel 3.2 May 26 21:37:17 +try May 26 21:39:24 woglinde: I wasn't able to get proper direction on starting with Angstrom that I tried basic narcissus image first, now should I try http://downloads.angstrom-distribution.org/demo/beagleboard/ for latest kernel? May 26 21:42:59 orated: That's a really old version, from january 2011. May 26 21:44:01 orated: Almost always, you want to avoid narcissus. THis from a friend who is on the beaglebone project--it is just too hard for most of us to get a useful image out of. May 26 21:45:03 Azertify: Are you trying http://treyweaver.blogspot.co.uk/2010/10/installing-angstrom-on-beagleboard-xm.html , as you said ? May 26 21:45:38 Yes May 26 21:45:52 But replace narcissus with the demo image May 26 21:45:56 Azertify: Those instructions asks using narcissus which gave Linux beagleboard 2.6.32 here May 26 21:46:02 Ok, better then May 26 21:46:37 It works better this way anyway because the instructions later use a bz2 (the demo image) whereas narcissus makes a gz file. May 26 21:50:57 woglinde: And regarding shorting some pins in OTG, were you referring to - "Specifically, Pin 4 (ID) of the OTG connector needs to be shorted to Pin 5 (GND) by using a 5-pin USB Mini-A plug which shorts these pins together in the plug." ? I already have http://trisoft.de/pics/ZHost.JPG connected as the wiki says May 26 21:54:55 I've got the board connected via hdmi to a tv, but the tv isn't displaying anything. I think that the bootloader might not have worked... again. May 26 21:58:33 Shit, the MLO file is 0 bytes. May 26 22:03:29 woglinde: Even Angstrom-Beagleboard-demo-image-glibc-ipk-2011.1-beagleboard.rootfs.tar.bz2 from http://downloads.angstrom-distribution.org/demo/beagleboard/ extracts /boot/uImage-2.6.32 May 26 22:06:30 orated 2011 we have 2012 May 26 22:06:59 er May 26 22:07:45 woglinde: You mean Angstrom-systemd-image-eglibc-ipk-v2012.05-beagleboard.rootfs.tar.bz2 file from the same link? May 26 22:07:48 I'm not sure May 26 22:07:55 yes May 26 22:07:57 try this May 26 22:08:04 it should have kernel 3.2 May 26 22:08:32 Okay May 26 22:08:49 if not you maybee need to build it on your own May 26 22:09:22 woglinde: And regarding OTG, were you referring to what I asked above^? May 26 22:15:36 afternoon May 26 22:33:03 good nite May 26 23:50:47 Trying to build cloud9-image I'm getting this error: May 26 23:50:47 ERROR: ld.so: object 'libpseudo.so' from LD_PRELOAD cannot be preloaded: ignored. May 26 23:51:27 Seems to have made a Angstrom-Cloud9-IDE-eglibc-ipk-v2012.05-core-beaglebone.rootfs.tar.bz2 image anyway though May 26 23:57:02 I also want to know why the instructions want me to zero out my entire microsd every time I make a new image. Seems like just mkfs would be fine May 27 00:11:25 i've noticed camera people are paranoid about flash in the same way May 27 00:14:20 anyone have a clue what this means cc1plus: all warnings being treated as errors May 27 00:14:20 make: *** [out/host/linux-x86/obj/EXECUTABLES/obbtool_intermediates/Main.o] Error 1 May 27 00:14:20 May 27 00:14:31 sorry about multi-line May 27 01:01:13 So if I have an improved /etc/nsswitch.conf to include in the default angstrom image what recipe should I patch? May 27 01:09:31 Oh, and I'd love to know how to make that cool .xz image that Koen originally pointed me to... It was basically an entire dd image of the microsd card May 27 01:19:26 riskable: I'm sure the bits for that are in OE somewhere ... May 27 01:19:43 riskable: IMAGE_FSTYPES="tar.gz tar.bz2 sdimg" -- perhaps grep on sdimg? May 27 01:19:55 cbrake: Well, I'm not sure if my nsswitch.conf would be relevant to anything but the Angstrom default image (since it uses Avahi) May 27 01:20:11 riskable: systemd-image rootfs takes 2m to build, cloud9-image takes 32m and errors out May 27 01:20:29 cbrake: Hah, I actually got cloud9-image to build May 27 01:20:35 I loaded it onto my SD card but now it won't boot May 27 01:20:48 Hangs at "Hit any key to stop autoboot: 0 " May 27 01:20:56 ...but I only *just* started investigating that May 27 01:21:13 Not at the point where I need to ask for help on that one yet. I could've goofed up something obvious :) May 27 01:22:23 riskable: looks like I have images generated even though rootfs is throwing an error May 27 01:23:25 Wait: Was it this error? ERROR: ld.so: object 'libpseudo.so' from LD_PRELOAD cannot be preloaded: ignored. May 27 01:23:59 riskable: ./openembedded-core/meta/recipes-core/base-files/base-files/nsswitch.conf May 27 01:24:09 riskable: just a guess based on a quick grep May 27 01:24:26 riskable: send a patch to the openembedded-core maillist May 27 01:24:54 riskable: if you change is Angstrom specific, then perhaps a bbappend in the angstrom layer May 27 01:25:10 Does openembedded-core use avahi in all their recipes by default? My nsswitch.conf would only apply if Avahi is included in something. Then again, perhaps I should modify the avahi package to automatically update nsswitch.conf May 27 01:25:45 Not terribly important. I'll figure out the best place to put my change May 27 01:26:00 Do you know what the MLO file is? May 27 01:26:06 riskable: I'm not sure about oe-core generic -- I never use it without Angstrom May 27 01:26:18 riskable: yes, that is the first stage bootloader May 27 01:26:35 Yeah, that's my concern... The change adds the mdns4 stuff to /etc/nsswitch.conf under the 'hosts' line. But you don't want that if you're not using zeroconf May 27 01:26:47 Would MLO change if I just built cloud9-image? May 27 01:26:59 riskable: or 2nd depending how you look at it -- a boot ROM in the CPU loads the MLO. the MLO loads u-boot. u-boot loads the kernel ... May 27 01:27:18 riskable: MLO should be independent of image built May 27 01:27:36 Oh, OK. I was wondering that maybe my Beaglebone won't boot because I forgot to copy the MLO over... But the old MLO is still there. I doubt it changeed May 27 01:27:46 riskable: agreed May 27 01:28:04 I should've done an md5sum before I overwrote it May 27 01:35:31 What's the difference between the uImage that goes on the first partition and the uImage that goes on the second partition in /boot? May 27 01:36:02