**** BEGIN LOGGING AT Sun Mar 04 02:59:59 2012 Mar 04 04:49:52 Hey all, I've got a BB xM rev C and am trying to get some basic GPIO going via c. I've got the CodeSourcery chain installed, but I've been searching around and can't figure out how to get the needed OMAP3(?) headers included. I've generated mux.h and pinmux.h from the TI utility, but they use OMAP34XX_CTRL_BASE, which doesn't seem to be defined anywhere Mar 04 04:49:58 (in headers that I can find) Mar 04 04:55:05 Hmm, I downloaded headers from kernel.org and it looks like they have the right stuff, I'm gonna try linking to those. If I cross-compile against these, will it work when I copy things over? Mar 04 04:57:54 whos a n00b here? Mar 04 04:58:09 Haha Mar 04 04:58:16 Did you see my questions? :P Mar 04 04:58:26 I haven't even gotten anything compiled yet Mar 04 05:01:51 wtf is with this lingo Mar 04 05:01:58 windowsfag here Mar 04 05:02:19 Which lingo? Most of that is general c-type lingo Mar 04 05:02:43 But maybe you're just coding in python or somesuch Mar 04 05:04:06 i finaly got xfce running on my BB <3 Mar 04 05:04:19 windows people don't cross compile apparently Mar 04 05:04:20 sadly, it's too slow for me Mar 04 05:05:40 rrrrrrrrrrrrrrrrrrrrrrrrreeeeeeeeeeeeeeeeeeeeeaaaaaaaaaaaaaaaalllllllllllllllllyyyyyyyyyyyyyyyyyyyy slow Mar 04 05:13:15 yay! <3 Mar 04 06:32:09 I'm trying to do GPIO in c, but all the examples I see just include , but my board/cross-compiler can't find it Mar 04 06:32:12 Anyone have any ideas? Mar 04 06:37:15 are you trying to control gpio from a device driver or from some user space application? Mar 04 06:37:38 Userspace - that's what I'm finding doing more reading, should I be using the file access stuff? Mar 04 06:37:43 It just seems rather clumsy Mar 04 06:39:40 well, i am a n00b but yeah i think you have to do the file access stuff Mar 04 06:39:45 Hrm gross Mar 04 06:39:54 this is the docs Mar 04 06:39:55 http://www.kernel.org/doc/Documentation/gpio.txt Mar 04 06:40:29 Yeah, but those are for device-driver side things yes? Mar 04 06:40:48 Oh I see, userspace is down below Mar 04 06:41:17 open one of the /sys/class/gpio/gpioN/ where N is the gio number Mar 04 06:41:28 yeah it is there too Mar 04 06:41:51 Hokay Mar 04 06:41:53 so read, write 0 or 1 to the file and control the gpio like that Mar 04 06:41:56 Right Mar 04 06:42:13 * fiveofoh hunts for some kind of wrapper for this Mar 04 06:42:16 what other way you expected it to work? Mar 04 06:43:04 Well I was hoping for something in userspace where I could do something like set_pin_direction(N, GPIO_OUT); Mar 04 06:43:06 Or some such Mar 04 06:43:14 oh ok you wanted to use something that wraps all the system calls Mar 04 06:43:16 And not have to deal directly with all this file voodoo Mar 04 06:43:16 Right Mar 04 06:43:35 I don't know if such thing exist Mar 04 06:48:07 here they have an example of using gpio from user space http://www.avrfreaks.net/wiki/index.php/Documentation:Linux/GPIO "Example of GPIO access from within a C program" Mar 04 06:48:12 no wrapper though Mar 04 06:50:03 I must admit that it is pretty nasty to do gpio using files. it simply not natural Mar 04 06:50:11 <_av500_> the file "voodoo" is 3 lines Mar 04 06:50:23 <_av500_> open, write, close Mar 04 06:50:33 sure but it is not intuitive Mar 04 06:50:38 <_av500_> sure Mar 04 06:50:43 <_av500_> but its not voodoo Mar 04 06:50:49 there should be an API for gpio from user space Mar 04 06:51:02 ok voodoo it is not Mar 04 06:51:22 nasty it is though Mar 04 06:52:14 * mranostay waves Mar 04 06:52:27 * _av500_ waveres Mar 04 06:52:32 * _av500_ wavers Mar 04 06:52:42 * _av500_ wafers Mar 04 06:52:45 heh there is a meeup group for everything Mar 04 06:52:45 * _av500_ waffles Mar 04 06:52:49 *meetup Mar 04 06:53:05 MeeGo has one. who would go to that? :) Mar 04 06:53:26 <_av500_> Tizen people? to not feel alone? Mar 04 06:53:34 Not voodoo, no Mar 04 06:54:14 But messy. I'm writing myself a little wrapper library - I don't want my pin initialization code all scattered with fopen() calles Mar 04 06:54:16 *calls Mar 04 06:54:43 <_av500_> write a function Mar 04 06:54:47 <_av500_> its like GOSUB Mar 04 06:55:14 yeah I hate having to read a program that does gpio opening stuff doing rea() write() Mar 04 06:56:15 _av500_: i should go troll that meetup :) Mar 04 06:56:21 "Android is the future!" :) Mar 04 06:56:43 lol that would be funny as hell Mar 04 06:57:07 Okay, important question though: I still don't know how to set the pinmux settings. Do I just have to do that in my uBoot, or can I do it in C? Mar 04 06:57:16 Or is that accessible via this file interface as well? Mar 04 06:57:38 yup everything is accessible through the file interface Mar 04 06:57:57 * fiveofoh goes and reads some more Mar 04 06:57:58 it's powerful for console scripts Mar 04 06:58:15 messy for everything else Mar 04 06:58:24 OrlandoT, its really not as big of a deal as you think Mar 04 06:58:41 OrlandoT, just imagine that you are using ioctl instead of read/write/seek, etc Mar 04 06:59:07 don't get me started with ioctls lol Mar 04 06:59:18 i always make an "sdk" layer on top of them Mar 04 07:01:20 Hmm Mar 04 07:01:41 So http://www.elinux.org/BeagleBoardPinMux seems to imply I can't do pinmux from userspace, I guess I'll just do the uboot method then? Mar 04 07:06:59 * mranostay notes drinking coffee at 11p never ends well Mar 04 07:14:10 OrlandoT, I hate it when people make APIs for APIs Mar 04 07:16:48 hey it is all about abstraction Mar 04 07:27:57 Russ: big fan of 'keep it simple' too? Mar 04 07:39:54 mranostay, yup Mar 04 08:23:59 <_av500_> APIfactoryPattern ftw Mar 04 08:25:05 Blinking light woo Mar 04 10:49:04 Hi, I'm a linux noob, but I have a beagle-XM rev C and I am trying to boot up the board with the Angstorm images from the beagleboard.org site on my personal SD-card. I've managed to do the partitions and get the system to boot up into Angstorm, but my USB keyboard and mouse are not detected. The same keyboard and mouse work fine with the preprogrammed SD-card that came with the board (which seems to be a Narcissus build) Could anyon Mar 04 10:49:40 debugging the USB host connections? Mar 04 11:12:23 hm, I'm quite lost, does anyone here know where's the latest linaro's kernel git tree for omaps(beagle)? Mar 04 11:12:40 there's quite a lot trees at git.linaro.org Mar 04 11:27:35 hi govind_ Mar 04 11:29:03 i would begin searching to discover what was different about the kernel versions and modules they have loaded at runtime Mar 04 12:00:13 hmm ok thanks, ill check that, and i have another question about the rootfs image on the angstorm site - do I need to unpack the modules.tar into the angstorm partition or is just the rootfs.tar enough for basic applications? Mar 04 13:34:39 koen, you awake? Mar 04 13:35:57 someone in need of trolling? Mar 04 13:36:25 not yet, just trying to make sure I under stand where the bone kernel source is Mar 04 13:37:00 good luck with that Mar 04 13:44:53 you want to understand ... where ... it is? Mar 04 13:45:39 well, ow branches in his git repo match Mar 04 13:45:49 mainly make sure I know what I am doing Mar 04 13:47:44 when you find the unicorn, you'll know you're getting close Mar 04 13:50:19 Crofton|work: yes Mar 04 13:50:55 Crofton|work: the canonical (no, not the ubuntu kind) source is the recipe Mar 04 13:51:25 Crofton|work: there's an exploded git tree at my github with branches for that match the recipe Mar 04 13:51:42 Crofton|work: but thats for people too lazy or too stupid to apply patches manually Mar 04 13:51:54 W: GPG error: http://ftp.uk.debian.org unstable InRelease: The following Mar 04 13:51:54 signatures couldn't be verified because the public key is not available: Mar 04 13:52:04 echan Mar 04 13:52:55 zumbi and? Mar 04 13:53:28 woglinde: wrong channel, sorry Mar 04 13:53:55 well, I am trying to make a patch Mar 04 13:54:00 and Mar 04 13:54:02 I am lazy Mar 04 13:54:31 bone uses the psp kernel? Mar 04 13:56:17 laziness is your greatest enemy Mar 04 13:57:16 the trick is to disguise it as patience Mar 04 13:57:26 :) Mar 04 13:57:58 one good way is to create insanely slow build systems, then invoke xkcd 303 Mar 04 13:58:11 you mean try to builds ona vm Mar 04 13:58:40 I asked for a headless dedicated box, and got a vm Mar 04 13:58:49 that every time I restart it, ssh is off Mar 04 13:59:21 o.O Mar 04 13:59:26 so enable sshd service Mar 04 14:01:21 Crofton|work: looking at the branches in github should give you a clue which kernel is the most recent one Mar 04 14:01:29 yeah Mar 04 14:01:33 r5a or something Mar 04 14:02:08 the bone kernele doesn't have the access perf counter from userspcae patch, right? Mar 04 14:02:12 grep suggests no Mar 04 14:03:38 I've been hesitant to add those Mar 04 14:03:44 yeah Mar 04 14:03:46 I understand Mar 04 14:04:01 I should make a module you can load to turn them on Mar 04 14:04:01 I don't want this kernel to work well Mar 04 14:04:06 what's the harm? Mar 04 14:04:11 I want people to help out with VHs kernel Mar 04 14:04:35 mru: it's not the perf counters I'm concerned about, it's investing time in a crap, dead end kernel Mar 04 14:04:50 VHS is already dead Mar 04 14:05:15 "VHs kernel"? Mar 04 14:05:45 maybe he wants you to move to a beta kernel Mar 04 14:06:36 all kernels except mainline are beta :) Mar 04 14:07:01 yes, but which one is max beta? Mar 04 14:07:10 beta max? Mar 04 14:13:52 Crofton|work: vaibhav hiremats kernel Mar 04 14:14:14 Crofton|work: you know, the one mentioned in the bone kernel roadmap email to the beagle list this week Mar 04 14:14:43 heh, I am behind on that list Mar 04 14:16:14 Crofton|work: remember the fuss about consolekit and polkit a while ago? Mar 04 14:16:20 yeah Mar 04 14:16:35 policykit is want linus wants Mar 04 14:16:37 * woglinde runs Mar 04 14:17:01 Crofton|work: oe-core upgrade maniacs broke it again with .103 -> .104 Mar 04 14:17:01 one day I need to to look them up and see what they do Mar 04 14:17:53 I've backported the patches from upstream that stop it from crashing and give an error message instead Mar 04 14:17:58 * mru read that as "lock them up" Mar 04 14:17:59 Crofton you want a short story? Mar 04 14:18:07 they are not for mru Mar 04 14:18:29 rofl Mar 04 14:19:24 what about spidev :) Mar 04 14:19:35 hm some kernel dev for my phone now or not Mar 04 14:20:34 Crofton|work: respond to the spidev thread and ask "so what clockspeed and polarity should we use for spidev?" Mar 04 14:21:05 haha the latest xkcd Mar 04 14:22:11 mind you, I think you can overrise all that from usersapce with spidev Mar 04 14:23:22 the vhs kernel is for all TI cpu's? omap etc? Mar 04 14:23:49 <_av500_> its for all of them Mar 04 14:23:56 <_av500_> even for Nvidia Mar 04 14:24:04 one tape to rule them all Mar 04 14:24:07 <_av500_> it cures cancer and fragmentation Mar 04 14:24:31 <_av500_> hmmn cancer is easy compared to linux arm Mar 04 14:24:46 especially omap Mar 04 14:24:50 omap is totally fucked Mar 04 14:24:52 rmk said so Mar 04 14:25:09 rmk is not the most reliable source Mar 04 14:25:17 news at 11 Mar 04 14:29:01 Crofton|work: VHs kernel is for am335x, it is linus + patches sent for review Mar 04 14:29:14 ok Mar 04 14:29:20 seems like a good approach Mar 04 14:29:26 it is Mar 04 14:29:42 he even accepts outside contributions Mar 04 14:30:49 his effort was a brief and unexpected outbreak of common sense Mar 04 14:31:56 I hope it succeeds and shows others at TI how to do things properly Mar 04 14:32:10 yes Mar 04 14:33:15 <_av500_> koen: is there a forum to send patches too? Mar 04 14:33:21 _av500_: heh Mar 04 14:33:36 _av500_: you know that joking about that will get you an official complaint, right? Mar 04 14:33:42 <_av500_> right Mar 04 14:33:50 <_av500_> next thinh I will post HDCP keys on e2e Mar 04 14:33:53 <_av500_> thing Mar 04 14:34:01 that was a major factor in why I quit TI a few weeks earlier that planned Mar 04 14:34:11 than* Mar 04 14:34:17 <_av500_> and antisec pastebins Mar 04 14:39:37 koen: btw, why is beagle still at 3.0.x? that sgx bug? Mar 04 14:39:48 I mean beagleboard Mar 04 14:42:13 git pull Mar 04 14:42:42 fscked unity focus crap Mar 04 14:44:07 ynezz, I think you need to follow that statement with, ', Batman!' Mar 04 14:44:47 ynezz: noone bothered to port all the PM crap to 3.1.x or 3.2.x and make it work Mar 04 14:45:05 ynezz: not to mention the camera stuff Mar 04 14:45:14 of course, devicetree will fix all that Mar 04 14:45:18 :D Mar 04 14:45:21 <_av500_> sure Mar 04 14:45:32 or maybe that Gary guy Mar 04 14:45:48 <_av500_> Powel? Mar 04 14:45:54 Thomas Mar 04 14:46:12 current pop star on the Angstrom mailing list Mar 04 14:46:36 <_av500_> ah, I meant Powers Mar 04 14:49:21 the thing is, that friend of mine has some crappy Samsung TV and poor beagle feeds it with some stupid resolution/frequencies, so I'm resurrecting that EDID thingie and found out via blind route omap_gpu/drm, that in the recent kernel, there's 'France V Ireland' Mar 04 14:49:33 oh, jesus no France V Ireland, but 'OMAP: DSS2: add panel-dvi driver' Mar 04 14:50:12 with edid support done previously by rsalveti Mar 04 14:51:56 ynezz: keep in mind that gary's consulting firm is blacklisted by at least one silicon vendor Mar 04 14:52:03 and I've started wondering why is beagle still on 3.0.x and found this https://github.com/RobertCNelson/stable-kernel/issues/8 Mar 04 14:52:14 but maybe it's just related to the ubuntu kernel? Mar 04 14:53:03 I think prabu fixed that in the 1.7.x based sgx stuff Mar 04 14:53:08 (which is DRI based) Mar 04 14:53:25 I suspect that will get released soon Mar 04 14:53:28 ok, apart from that PM and camera patches, no problems so far? Mar 04 14:53:28 ynezz, that's with the older 4_03_00_02 sgx... Mar 04 14:53:40 * koen makes not of another NDA that needs to get resigned Mar 04 14:53:47 just want some hint for a shortcut, don't want to just spend 2 days finding out what's broken :) Mar 04 14:54:01 I know, that it's hard in omap world, but... Mar 04 14:54:03 with 3.3, (kms/dri) i'm hopeing the newest sgx will fix that.. (fingers crossed) Mar 04 14:54:26 rcn-ee: 1.7.x is a *lot* better than 1.6.x Mar 04 14:55:17 yeah, so far i like it lot better, but it's only working in "non-x11" mode with v3.2/3.1 (can't use the xorg driver in the release, to old for).. ;) Mar 04 14:56:05 hm, my use case is just pure qt/qml on fb Mar 04 14:57:04 the 4.05.00.03's should work for that then.. Mar 04 14:57:34 4.06 is 1.7.x based Mar 04 14:57:39 4.05 is a bag of hurt Mar 04 14:58:03 :D Mar 04 14:58:38 looks at sgx link.. crap... not posted yet. ;) Mar 04 14:59:12 <_av500_> sgx works fine in android Mar 04 15:00:02 is that why mali usage is increasing? ;) Mar 04 15:00:12 <_av500_> rumours Mar 04 15:00:53 <_av500_> i guess TI will keep imgtec alive Mar 04 15:01:02 <_av500_> ti and aapl Mar 04 15:02:05 wish they didn't.. Mar 04 15:03:07 koen: and I also wonder why there's different mailing list just for meta-ti layer... Mar 04 15:03:16 <_av500_> meta-lists Mar 04 15:03:22 it's some kind of obfuscation? Mar 04 15:03:23 because every layer needs a mailinglist Mar 04 15:03:28 <_av500_> you get mails about mails Mar 04 15:03:37 it's listed in the README, no? Mar 04 15:03:57 my .procmailrc is getting longer then brooklyn bridge Mar 04 15:04:42 <_av500_> try .metaprocmailrc Mar 04 15:08:28 hm intel will not keep imgtech alive too? Mar 04 15:08:45 they are comming not along with there graphics soc's Mar 04 15:09:42 hm, and that magic VH's kernel tree is just for bone? Mar 04 15:10:08 i think he's doing it for all the am335 evm's too.. Mar 04 15:13:29 "especially targetted for BeagleBone and daughterboards" Mar 04 15:14:24 Hi there, I have a few questions about BeagleBone Mar 04 15:15:33 I'm new to Linux, so please be gentle :-) Where can I find what is started when the BeagleBone boots? Mar 04 15:16:18 And will it damage anything in the filesystem if the shutdown procedure is to just remove power? Mar 04 15:17:34 that really depends on how your file system is setup.. if your writing/accessing system files there's always a chance you'll do more damage then fsck can repair.. Mar 04 15:17:55 nfsroot ftw Mar 04 15:18:28 nfsroot over gsm yeah Mar 04 15:21:57 hmm, which board file does the bone use? Mar 04 15:24:12 interesting Mar 04 15:24:21 kconfig claims a beagle Mar 04 15:24:28 board-am335xevm.c? Mar 04 15:25:31 it's so easy to guess, isn't it? :p Mar 04 15:25:54 arg Mar 04 15:26:00 mltiple machines selected Mar 04 15:28:37 ynezz, I think you are right Mar 04 15:29:07 It's because my end application needs to be able to handle that the user just removes power. Mar 04 15:29:59 then leave root on read-only partiation Mar 04 15:30:03 I really don't know if I'll be accessing system files Mar 04 15:30:10 ro ^exactly. ;) Mar 04 15:31:27 My own application (written in C) will access IO, filesystem (saving some data) and use sockets Mar 04 15:32:15 use a initrd Mar 04 15:32:39 initrd, is that for starting programs at boot? Mar 04 15:32:56 Crofton|work: cp defconfig .config if you're using my git tree Mar 04 15:32:58 thats for loading rootfs into ram Mar 04 15:33:26 <_av500_> +1 Mar 04 15:33:50 well, that's a good suggestion on board with 256MB of RAM :) Mar 04 15:34:13 My BeagleBone has 256MB RAM Mar 04 15:34:13 ynezz you dont know the case Mar 04 15:34:16 koen, already done Mar 04 15:34:30 256 mb ram could be eniugh Mar 04 15:34:59 So using initrd will prevent powerloss from damaging the filesystem? Mar 04 15:35:03 no Mar 04 15:35:15 depends on the usecase you have Mar 04 15:35:26 "saving some data" is use case Mar 04 15:36:02 * woglinde still dont gets wyh the android vendors ship the wireless modules on /system and not on the initrd Mar 04 15:36:37 ynezz he didnt say where Mar 04 15:36:45 maybee saving them via network Mar 04 15:36:53 would be an option Mar 04 15:36:59 Perhaps it's a good idea to have a battery backup that takes over then power is lost, and a running service could detect power loss from an input and initiate a shutdown? Mar 04 15:37:26 Futterama thats sound like an good idea Mar 04 15:37:53 if you take care about the backup battery :) Mar 04 15:37:55 Sorry, I will be saving some userdata on the SD card somewhere - location is not important for me Mar 04 15:39:15 I'm thinking about using maybe a 2S LiPo battery, and then do a recharge when power is back on Mar 04 15:39:30 And ofcourse use a 5V regulator Mar 04 15:39:56 I would make one partition read-only for rootfs and the second one read-write for data Mar 04 15:40:22 ynezz that for sure Mar 04 15:40:25 and cross fingers, that some hickup to data partition won't take down whole SD card :p Mar 04 15:40:37 or cannt you connect a second card? Mar 04 15:41:07 yep, 2nd card or usb flash disk is much better solution Mar 04 15:41:19 usb flash is to big Mar 04 15:42:39 and usb on omap isn't synonym for stable either :) Mar 04 15:42:55 ven though I use another storage media for data, if I'm writing data while power is removed, filesystem on the 2nd storage could still currupt, right? Mar 04 15:43:14 Even though... Mar 04 15:44:20 if you poweroff in rw access case there is always a minimal changes to corrupt the fs Mar 04 15:44:23 its natural Mar 04 15:45:23 Ok, I think the backup battery is a better solution then. LiPo batteries are getting cheap these days Mar 04 15:45:57 again, depends on the use case, if it's outdoor then in winter it's quite useless... Mar 04 15:46:45 hehe Mar 04 15:46:51 temperature is another case Mar 04 15:47:22 but I think the omap is not build for -20 Mar 04 15:47:47 beagle was running fine in -20 this winter :) Mar 04 15:47:50 Temperature will be above 10 degrees C Mar 04 15:48:11 ynezz where do you use it outdoors? Mar 04 15:48:21 one mockup on forklift Mar 04 15:48:44 mockup? Mar 04 15:48:58 demo, proof-of-concept Mar 04 15:49:12 My application is actualy outdoor but not during winter Mar 04 15:49:32 ynezz ah okay Mar 04 15:49:35 till later Mar 04 15:49:36 has had the beagle survie -40F, (but the lead acid batteries didn't) Mar 04 15:50:20 would you put an ip-camera in -40 weather? Mar 04 15:50:47 mine survived -16 Mar 04 15:50:48 c Mar 04 15:50:55 well, yeah.. ;) it was a weather cam on the roof at work. .;) Mar 04 16:22:24 <_av500_> we did rootfs in ram with 64MB Mar 04 16:23:55 that is so cool Mar 04 16:24:04 is it generally useful? Mar 04 16:24:47 hm there must be a way to measure this Mar 04 17:06:38 <_av500_> well, we had just 2MB to boot from and a HDD that was off most of the time Mar 04 17:06:43 <_av500_> so RAM was the only option Mar 04 17:08:01 rcn-ee: -40 is a nice temp Mar 04 17:08:15 rcn-ee: it's the point where C and F are exactly the same :) Mar 04 17:08:43 <_av500_> koen: but which one is colder? Mar 04 17:15:44 * mranostay waves and yawns Mar 04 17:18:30 koen: i wouldn't call -40F or C a nice temp :) Mar 04 17:30:42 rcn-ee: is the xdriver issue at sgx because of the older x11 abi? Mar 04 17:31:12 if so then the only way to get it fixed is to get it built against the latest x11 used at ubuntu Mar 04 17:31:17 and that might not be that easy =\ Mar 04 18:18:34 ubuntu is like having a maid who doesn't speak my language Mar 04 18:18:34 wants to be helpful. gets in the way. Mar 04 18:18:57 hehe Mar 04 18:19:11 badezucker: can i quote you on that? Mar 04 18:19:24 yes sir Mar 04 18:20:03 i think it captures the problem succinctly does it not? Mar 04 18:20:12 i know how to sweep my floor Mar 04 18:20:22 i don't know how to tell the strange lady in the room how Mar 04 18:21:22 would violence solve the problem with ubuntu? Mar 04 18:24:29 sex maybe Mar 04 18:35:03 <_av500_> exorcism Mar 04 18:35:39 <_av500_> and hands off the maid Mar 04 18:50:43 rsalveti, yeah on the older sgx for the beagle, the builtin sgx xorg driver is looking for abi 6.. but i'm hoping robclark's omapdrm with 3.3 and his xorg driver should solve that problem.. ;) Mar 04 18:56:34 omapdrm in 3.3? Mar 04 18:56:52 or you mean on top of 3.3? Mar 04 18:57:19 it's in 3.3.. ;) it's just that it's main depencidy "cma-v23" is on top of that Mar 04 19:00:12 humm, this is going to be fun.. 3.3-rc5 is awesome on my panda_es, but my regular panda's can't stay up under load for a hour or two.. no serial error message.. they just 'disappear'.. Mar 04 19:07:14 I'm blind, I don't have drivers/gpu/drm/omap in 3.3-rc6 Mar 04 19:08:03 it's in staging.. Mar 04 19:08:38 yes, blind :) thanks Mar 04 19:09:07 no problem, there's a lot of crap in that dir. ;) easy to miss.. Mar 04 19:09:46 have to master grep more :p Mar 04 19:10:28 it's actually pretty cool.. reads the edid register and sets the resoultion... ;) Mar 04 19:31:02 ynezz git grep Mar 04 19:33:41 yes, I was talking about git grep Mar 04 20:13:01 "OMAP totally fucked?" t-shirt, anyone? :p Mar 04 20:15:23 av500 has one Mar 04 20:15:24 <_av500_> statin the obvious... Mar 04 20:15:27 for sure Mar 04 20:17:02 My beaglebone came with kernel 3.2.5+ and certain file system. What is the best way for me to be able to duplicate (meaning download and compile) this kernel + filesystem? Mar 04 20:17:39 I have tried Angstrom distro with OE but the recipes there are for kernel 3.1.2 or something like that Mar 04 20:18:09 I want to be able to create exactly what my beaglebone is currently running Mar 04 20:20:13 OrlandoT: I saw the eeprom mystery seems to be resolved, fantastic. If you figure out how to find the system image you're looking for could you let me know? One wouldn't imagine it'd be so problematic to keep the mirrors up to date with what's shipping on the boards, but maybe I'm missing something… Mar 04 20:23:08 OrlandoT you forgot to include meta-texasinstrument? Mar 04 20:24:18 * mranostay yawns again Mar 04 20:24:22 ah? Mar 04 20:24:56 Sgarr: Sure I will let you know but in reality i am looking for the sources so that I can make my own binary images Mar 04 20:25:14 woglinde: meta-ti? what? Mar 04 20:25:47 OrlandoT: yeah that's what I'll be looking for too. Mar 04 20:25:54 oh ok Mar 04 20:26:08 look at meta-ti/recipes-kernel/linux/linux-ti33x-psp_3.2.bb Mar 04 20:26:11 hey i got a question has anyone got autologin working with ubuntu 11.10 on their beagleboard xm? Mar 04 20:26:20 that's the "source" for that kernel Mar 04 20:26:51 or this https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.2-r5a+gitr09e9651bcf2ee8d86685f2a8075bc6557b1d3b91 Mar 04 20:27:00 ynezz: oh really? cool, how come the Angstrom scripts do not include that recipe? Mar 04 20:27:16 it does Mar 04 20:27:51 if you build kernel for machine=beaglebone it should pickup that kernel automagically Mar 04 20:28:09 require conf/machine/include/ti33x.inc Mar 04 20:28:10 that is exactly what I do Mar 04 20:28:25 that's in machine/beaglebone.conf Mar 04 20:28:35 PREFERRED_PROVIDER_virtual/kernel = "linux-ti33x-psp" Mar 04 20:28:44 and that include contains this line ^ Mar 04 20:28:48 and when i run my system it and do a -uname -r it says kernel 3.1.0 instead of 3.2.5+ Mar 04 20:29:31 if you do bitbake virtual/kernel, it should build that kernel Mar 04 20:29:46 ynezz: so you are telling me that Angstrom scripts already do all those steps that you are describing here? Mar 04 20:30:00 wow there are some horrible makefiles out there.. Mar 04 20:30:55 OrlandoT: yes, it should, but it's just my guess, since I don't have bone (yet) Mar 04 20:31:06 just interpreting the sources Mar 04 20:33:03 ok Mar 04 20:33:09 it will be probably something more recent than 3.2.5+ now, there's file://3.2.6/0087-Linux-3.2.6.patch Mar 04 20:33:13 so this is what i am doing, let me know where I go wrong Mar 04 20:33:36 git clone git://git.angstrom-distribution.org/setup-scripts Mar 04 20:33:37 cd setup-scripts Mar 04 20:33:50 MACHINE=beaglebone ./oebb.sh config beaglebone MACHINE=beaglebone ./oebb.sh update MACHINE=beaglebone ./oebb.sh bitbake virtual/kernel Mar 04 20:34:08 seems ok Mar 04 20:34:31 what about the meta-texasinstruments that you were talking about? Mar 04 20:34:38 it's meta-ti Mar 04 20:34:53 ls sources/meta-ti Mar 04 20:35:26 do i have to download that manually or that is taken care by the setup-scripts? Mar 04 20:35:41 it's taken care by the setup-scripts Mar 04 20:36:06 what git hash do you have in yours meta-ti? Mar 04 20:36:22 so after doing those steps and waiting like 8 hours for the stuff to compile i endup with version 3.1.0 of the kernel Mar 04 20:37:05 it has been a week now since last time i tried though Mar 04 20:37:10 maybe i should try again Mar 04 20:37:36 it has been exactly 7 days since last time i tried Mar 04 20:38:55 I am doing a rm -rf setup-scripts/ Mar 04 20:39:02 and trying again Mar 04 20:39:09 here goes another 8 hours of my life Mar 04 20:39:22 oh, why? Mar 04 20:39:23 :) Mar 04 20:39:44 <_av500_> get an i7, only 1h Mar 04 20:40:04 well, i'm running on a VM Mar 04 20:40:32 'git pull; ./oebb.sh update; source oe-environment; bitbake virtual/kernel' Mar 04 20:40:37 should be enough Mar 04 20:40:52 oh geez now you tell me lol Mar 04 20:41:00 you didn't asked :p Mar 04 20:41:02 too late Mar 04 20:41:09 i guess i didn't Mar 04 20:41:48 and you don't need to build it actually Mar 04 20:42:11 I don't? Mar 04 20:42:22 I mean to get the kernel version Mar 04 20:42:36 true Mar 04 20:42:49 it is a text file in the kernel dir Mar 04 20:43:06 i always forget the name but I will check after the downloading it again Mar 04 20:43:13 no, just run bitbake -e virtual/kernel Mar 04 20:43:22 there's also -v and -D debug option Mar 04 20:43:26 oh cool Mar 04 20:43:33 so you can see what it's doing etc. Mar 04 20:44:11 ok Mar 04 20:44:38 oebb.sh includes the line, DISTRO="angstrom-2010.x" if I leave that alone it builds 3.1.0+, but if I change the 2010.x to 2012.x it builds 3.2.0+. Mar 04 20:45:33 ah so Mar 04 20:46:07 oh really? Mar 04 20:46:39 so i have to manually edit the oebb.sh file to get 3.2.5+? Mar 04 20:46:47 no Mar 04 20:47:04 it's pinned to this version for some reason in this distro Mar 04 20:47:08 so what is Sgarr saying then? Mar 04 20:47:24 that it's pinned Mar 04 20:47:59 wow I am so confused right now Mar 04 20:48:19 to get latest do you just remove the DISTRO specification? Mar 04 20:48:41 OrlandoT: good, then _av500_ can go home Mar 04 20:49:53 lol Mar 04 20:50:28 I opened the oebb.sh that i just downloaded and it has DISTRO="angstrom-2010.x" Mar 04 20:50:41 <_av500_> mru: about to go to bed actually... Mar 04 20:51:18 OrlandoT: just add DEFAULT_PREFERENCE = "-1" to linux-ti33x-psp_3.1.bb Mar 04 20:51:27 so if that distro is pinned with version 3.1.0+ then doesn't that mean that i will not get version 3.2.5+? maybe i am getting confused with the term "pinned" english is my second language after all Mar 04 20:51:46 <_av500_> otoh, there's Red October on tele... Mar 04 20:54:57