**** BEGIN LOGGING AT Thu Mar 07 02:59:58 2013 Mar 07 03:21:49 NishanthMenon: ! Mar 07 03:21:58 mranostay, yo Mar 07 03:22:35 troll! Mar 07 03:23:50 beer sweet beer Mar 07 03:24:02 * mru prefers it bitter Mar 07 03:24:08 matches my personality Mar 07 03:24:13 heh Mar 07 03:24:28 mru: http://21st-amendment.com/beers/bitter-american/ Mar 07 03:27:16 * NishanthMenon gets pissed off with old beagle_recover scripts :( Mar 07 03:33:24 NishanthMenon: group hug? Mar 07 03:33:35 mranostay, :D Mar 07 03:33:44 group beer ;) Mar 07 03:34:25 that'll do Mar 07 03:42:12 ok. looks like it was a hub that pusb does not like :) Mar 07 03:43:21 aah bus powered hub Mar 07 03:43:27 ok. that kind of explains now Mar 07 03:43:29 :) Mar 07 04:50:09 roast-beef Mar 07 04:50:15 anyone want dinner? Mar 07 05:40:12 more ELC videos up Mar 07 05:40:41 * mranostay checks Mar 07 06:01:51 ? Mar 07 06:02:09 Is this the right forum to ask for Beaglebone support? Mar 07 06:10:32 maybe if you stayed longer than 1 minute :P Mar 07 06:10:52 :> Mar 07 06:27:44 heh Mar 07 06:27:54 jason is probably sleeping Mar 07 06:37:23 mranostay: trolls don't sleep Mar 07 06:39:36 jason isn't a troll Mar 07 06:39:47 usually Mar 07 06:40:31 he can certainly troll if he wants to Mar 07 06:40:43 that makes him a troll Mar 07 06:42:02 heh Mar 07 06:42:16 i did like his depmod troll last week Mar 07 06:59:21 morn' Mar 07 06:59:29 moanin' Mar 07 07:00:58 guten moanin'? Mar 07 07:01:31 * panto moanining Mar 07 07:06:15 up too late Mar 07 07:48:58 morning Mar 07 07:50:16 moin Mar 07 07:57:01 yo Mar 07 09:12:07 which repo is used for linux kernel development for the beagle bone? Mar 07 09:13:18 there are quite a few trees Mar 07 09:13:56 which one do you recommend? Mar 07 09:16:12 any that works Mar 07 09:18:29 https://github.com/beagleboard/kernel Mar 07 09:18:51 3.8 branch or so Mar 07 09:19:29 for kernel development? Mar 07 09:21:58 Russ: who'd be crazy enought to even consider that? Mar 07 09:22:39 well, there was that former finnish student... Mar 07 09:23:34 how about this one? http://git.ti.com/ti-linux-kernel Mar 07 09:23:42 I just mean who would you send the patches to? Mar 07 09:24:00 f11f12, what kernel development are you doing? Mar 07 09:24:17 btw, if anyone is working on the grove motor controllers let me know Mar 07 09:24:32 I have a very simple kernel driver that works but I'd like to know what else you might need Mar 07 09:25:38 I'd just like to have USB drivers that use DMA, if they are available. Mar 07 09:26:15 panto: grove? Mar 07 09:27:33 http://www.seeedstudio.com/wiki/GROVE_System Mar 07 09:27:50 this one specifically http://www.seeedstudio.com/wiki/Grove_-_I2C_Motor_Driver_V1.3 Mar 07 09:28:28 f11f12, so you don't want to do kernel development, you want a kernel to use Mar 07 09:29:27 av500: justa L298 Mar 07 09:31:21 panto: why does this need a kernel driver? Mar 07 09:31:23 its i2c, no? Mar 07 09:31:38 bitbang it! Mar 07 09:38:28 Russ: true Mar 07 09:38:53 have you tried the https://github.com/beagleboard/kernel tree and found it to be wanting? Mar 07 09:42:35 Russ: I'll check it out. Thanks! Mar 07 09:56:54 hmm, how do I bind tftpd to usb rndis? Mar 07 10:03:14 gm ogra Mar 07 10:03:18 gm av500 Mar 07 10:03:24 gm woglinde Mar 07 10:03:46 av500 setup up the interface and restart tftpd? Mar 07 10:04:22 nah Mar 07 10:04:27 needs to be automatic Mar 07 10:04:47 lets see if that suse networkdamager can do it Mar 07 10:05:37 hm Mar 07 10:05:43 do not think so Mar 07 10:05:43 anyone here with msp430 knowledge? i have a strange phenomena: i configured the timerA to give me an interrupt on the timerA0 vector, but what i get is an interrupt on timerA1... the datasheet and the errata do not contain any explenation for that. does anyone have an idea? Mar 07 10:06:02 koth you hit a hw bug? Mar 07 10:06:20 maybe... but i would expect it in the errata..which it isnt Mar 07 10:06:23 av500 never heard that network-manager can do it Mar 07 10:06:33 it isnt like it is a easy to miss bug Mar 07 10:06:37 av500 but ifup/down with post-up should do the trick Mar 07 10:06:49 av500 but you maybee need a little time for the client Mar 07 10:07:42 av500 possible too that ha-tftpd works out of the box when running it on 0.0.0.0 Mar 07 10:08:51 get away from network-manager. just use an udev-rule Mar 07 10:09:30 lol Mar 07 10:27:54 av500, cause banging stuff on the i2c without a driver is gross Mar 07 10:28:42 I've seen the way people use it with i2cset and it's an accident waiting to happen Mar 07 10:28:50 linux is not an arduino people Mar 07 10:29:26 panto: so write a lib Mar 07 10:29:35 that abstracts the motor thing and talks I2C Mar 07 10:29:58 it still would be userland Mar 07 10:30:02 bs Mar 07 10:30:12 it's a stupid idea Mar 07 10:30:29 a driver is the right thing Mar 07 10:31:26 you can then abstract the driver behind a lib, but nothing is easier than not needing a lib at all Mar 07 10:32:19 and let's not talk about weird problems when trying to have more than one process access the same i2c peripheral Mar 07 10:32:34 now you have to arbitrate in user space Mar 07 10:33:30 what do you want to control? Mar 07 10:35:08 panto: I can totally see 2 processes control a stepper motor, one per phase Mar 07 10:35:26 with linux in userland? Mar 07 10:35:59 even in kernel-space that is risky Mar 07 10:36:08 even it was a joke Mar 07 10:36:23 *g* Mar 07 10:36:54 I still would use an avr or similiar to control a motor. Mar 07 10:37:32 right tool for the right problem Mar 07 10:37:47 just letting linux doing higher level things which don't have timing restrictions Mar 07 10:37:50 woglinde: no hw bug: stupidity on my side Mar 07 10:39:02 aholler: why? Mar 07 10:39:26 you have HW running ~1GHz and it cannot control a stepper motor? Mar 07 10:39:31 but needs an AVR Mar 07 10:39:33 thats poor Mar 07 10:39:39 thats linux Mar 07 10:40:13 koth exact steps please Mar 07 10:40:26 can't you pwm it? Mar 07 10:41:02 av500 nobody said van neumann should control motors Mar 07 10:41:09 ups von Mar 07 10:41:27 what has that got to do with it? Mar 07 10:41:59 woglinde: i didnt read the datasheet carefully enough and enabled the general timerA irq instead of the timerA0 irq. and the general irq goes to the same vector as the timerA1 irq... Mar 07 10:42:25 poor koth Mar 07 10:42:52 hm I could update java today Mar 07 10:43:08 and get an ask.com toolbar! Mar 07 10:43:15 what confused me was, that in the mode i'm running the timer the general irq worked exactly as i wanted it, just going to the wrong vector Mar 07 10:43:16 ? Mar 07 10:43:28 russ they are doing since 5 years Mar 07 10:43:38 dont know why now all yell at this Mar 07 10:43:52 because there have been so many updates lately Mar 07 10:43:54 maybee they see much often Mar 07 10:43:59 but no Mar 07 10:44:13 meta-java do installs 3rd taskbars Mar 07 10:44:20 +not Mar 07 10:44:28 +party Mar 07 10:55:16 panto: so for grove alone, there will be 76 kernel drivers? Mar 07 10:55:41 av500, why not batch them together in one (for the weird ones) Mar 07 10:56:05 the drivers that fall into something standard should be different Mar 07 10:56:30 my point is that an i2c driver is not hard to create Mar 07 10:56:36 sure Mar 07 10:56:42 I am mostly pulling your leg Mar 07 10:56:46 and it does help making things clear Mar 07 10:56:51 aha, trolling then? Mar 07 10:56:54 no Mar 07 10:56:58 would not call it that Mar 07 10:57:06 troll toll troll Mar 07 10:57:06 but if you insist Mar 07 10:58:31 the driver for that stupid motor controller was just a few hours of hacking Mar 07 10:59:03 why not do the right thing if you can? Mar 07 10:59:29 sure Mar 07 10:59:31 pastebin? Mar 07 11:02:05 http://pastebin.com/bfbXY89e Mar 07 11:03:25 again this "OF" Mar 07 11:03:31 does that replace DT? Mar 07 11:04:26 re Mar 07 11:05:41 * panto awards 6 troll points Mar 07 11:15:19 eh, sprintf Mar 07 11:36:44 * av500 feels we should use the word "Community" more often Mar 07 11:39:11 dm8tbr: maybe if they had given everybody free phones and tablets? Mar 07 11:39:21 that worked for nokia Mar 07 11:39:22 what should we throw while chanting 'community, community, community' Mar 07 11:40:03 hmm, segfaults? Mar 07 11:40:04 'they'? Mar 07 11:40:14 yes Mar 07 11:40:18 C Mar 07 11:40:40 oh, you meant my comment Mar 07 11:40:47 no Mar 07 11:41:01 I just did not provide context Mar 07 11:41:06 what comment? Mar 07 11:41:15 ok Mar 07 11:41:26 54! Mar 07 11:41:41 everybody should just chip in 26 Mar 07 11:41:52 then we'll get at least 54 or a multiple of it Mar 07 11:47:01 panto, morning and LOL Mar 07 11:47:05 anyone tried perf one the bone yet? Mar 07 11:47:14 s/one/on/ Mar 07 11:47:27 panto, saw the existing hunk for my gpevt test driver in your patch ;) Mar 07 11:47:35 mdp, heh Mar 07 11:47:46 yeah, never said I was about to post it to mainline yet :) Mar 07 11:47:46 panto, more evidence that koen will take *anything* into his *evil* vendor tree Mar 07 11:47:53 hahaha Mar 07 11:48:21 panto, I spotted the help message in Kconfig immediately ;) Mar 07 11:48:33 hehe evil Mar 07 11:49:13 I've seen much more evil in real evil vendor kernels :) Mar 07 11:49:26 his evil is harmless at best :) Mar 07 11:50:53 then its not evil Mar 07 11:51:01 a harmless evil, right Mar 07 11:51:11 ARMless Mar 07 11:52:54 panto, did you consider just doing an i2c-dev driver? or because of dynamic needs you need a unique kernel driver? Mar 07 11:53:30 unique kernel driver is the best Mar 07 11:53:41 shoragan: jackmitchell said perf crashes on exit with 3.8 and is plain broken on 3.2 Mar 07 11:53:54 I can just declare everything in the DT fragment, and bam, all is ready when it loads Mar 07 12:00:49 panto, I figured that's why you would do something this trivial as a kernel driver Mar 07 12:01:15 DT is all about describing hardware Mar 07 12:01:25 a userspace driver would be an interesting use case to support ;) Mar 07 12:01:39 doing it that way, everything is nice and really fits into place Mar 07 12:01:56 sure, you just made me consider that it's forced the most trivial things to be implemented as kernel drivers Mar 07 12:01:58 i2c-dev makes sense when you don't want to mess with board files Mar 07 12:02:04 are we talking about this grove stuff? Mar 07 12:02:05 not a horrible thing..just an observation Mar 07 12:02:19 mdp, define trivial Mar 07 12:02:52 panto, trivial is in the eye of the beholder Mar 07 12:02:58 it's a hardware peripheral, have a driver handle it and present a more abstract interface to user-space Mar 07 12:03:35 panto: so you hide 3 lines of libi2c Mar 07 12:03:46 by using printf on sysfs Mar 07 12:03:56 the above is my point Mar 07 12:04:00 av500, not really Mar 07 12:04:05 since we know it can be done in far less lines Mar 07 12:04:24 however, those far less lines don't support dynamic fragment loading, so they are *useless* in some cases Mar 07 12:04:27 i2c_open( adrr ) Mar 07 12:04:42 i2c_write( addr, whatever) Mar 07 12:04:49 i2c_close() Mar 07 12:04:58 sprinkle some void context over it Mar 07 12:05:30 and if I use the grove driver for a stepper motor, I still have all the timing in user space Mar 07 12:05:30 av500, sure, now instead of solenoid control, you have to control something like a stepper motor, or something that needs semi-accurate timing Mar 07 12:05:33 stir or use the Kitchenaid ? Mar 07 12:05:47 your user-space method is now useless Mar 07 12:05:56 panto: yes, so where is speed/direction and step count? :) Mar 07 12:06:14 av500, I said it's about solenoids in my application :) Mar 07 12:06:27 low hanging fruit :) Mar 07 12:06:39 my point is that use an interface that hides the transport Mar 07 12:06:52 now, move the control to something like the PRU Mar 07 12:06:58 you are creating a motor subsystem? ;) Mar 07 12:07:03 YES! Mar 07 12:07:06 instead of the piddly AVR Mar 07 12:07:06 vrooom! Mar 07 12:07:13 gti Mar 07 12:07:13 do that in user-space :P Mar 07 12:07:25 mdp, dunno, sounds like fun :) Mar 07 12:07:34 panto, it's my idea! Mar 07 12:07:52 not anymore! I stole it! I stole it! hahaha... Mar 07 12:07:56 * panto runs away with the idea Mar 07 12:08:15 cite me as your inspiration when you post your to-be-ignored patches Mar 07 12:08:46 mdp: is there and "ignored-by:" line? Mar 07 12:08:48 * panto doesn't really have time for that :) Mar 07 12:08:49 an* Mar 07 12:09:02 av500: that would be nice Mar 07 12:21:46 thath would make uncountable lines Mar 07 12:22:41 getting ingored is the norm Mar 07 12:23:23 DIN norm? Mar 07 12:26:38 linux is already bougt by cooperations Mar 07 12:27:18 +h Mar 07 12:33:16 linux is da big money! Mar 07 14:05:35 jackmitchell: u-boot is in now: https://github.com/beagleboard/meta-beagleboard/commit/0b4d5ed55961b0ffb1edd9e2f74dee035fb26cdc Mar 07 14:05:54 koen, the crash on exist is also what i see here Mar 07 14:05:59 (regarding perf) Mar 07 14:06:14 koen: heh, these patches look like what I applied a few days ago Mar 07 14:06:19 koen: cheers koen Mar 07 14:06:30 shoragan: I get crashes on perf exit, is this what you're seeing? Mar 07 14:06:37 koen: oh, DFU Mar 07 14:06:39 nice Mar 07 14:06:41 jackmitchell, yes, on 3.8.2 Mar 07 14:06:43 does it work? Mar 07 14:06:57 koen: you should resend that ext4 patch series Mar 07 14:07:07 Jacmet: I know, Tartarus told me the same Mar 07 14:07:10 shoragan: I get the same, I haven't had chance to really debug it yet Mar 07 14:07:33 it's a pain in the neck because it doesn't give out any back trace or dump at all Mar 07 14:07:45 av500: I think DFU nearly works as well in there as it does in mainline Mar 07 14:07:47 which is to say, not quite Mar 07 14:07:53 jackmitchell, do you know if it should basically work, or are there drivers (or something) missing in 3.8? Mar 07 14:08:08 av500: I have a usb3 controller in my laptop, so activating DFU crashes the OS Mar 07 14:08:13 av500: both OSX and linux Mar 07 14:08:17 Fixing an issue with raw atm Mar 07 14:08:18 lol Mar 07 14:08:20 nice Mar 07 14:08:28 Tartarus: ic Mar 07 14:08:46 right now I'm going over all the patches I did while travelling and pushing those out Mar 07 14:08:58 And I should re-re-prepare this SD card with a small FAT partition 1 and ext* 2 and still leave room up front for RAW writing images Mar 07 14:10:09 shoragan: I don't I'm afraid Mar 07 14:12:35 Tartarus: I did some tests the other day, and the am335x rom seems to be happy booting raw even though you have a pc mbr at sector 0 and MLO data at 128kb Mar 07 14:12:59 Jacmet: yup, same "trick" as all the rest of the TI platforms :) Mar 07 14:13:10 Tartarus: ahh ;) Mar 07 14:13:21 What makes me sad atm is that for RAW mode we hard-code placing U-Boot into the 3rd spot ROM will check for a valid boot file Mar 07 14:13:43 (ROM will try at 0x0, then 0x100, 0x200 and 0x300 for a valid header) Mar 07 14:14:25 Tartarus: yes, I also dislike all this hardcoding. I haven't found time yet, but would like to add the possibility for the platform to define the highlevel spl logic (E.G. location) Mar 07 14:14:54 Jacmet: Well, for RAW you do have to hard code where to look at some level. But it's too late to change where we look now :( Mar 07 14:15:11 Tartarus: we often need to handle failsafe upgrades and chose between 2 images in the bootloader based on some flag Mar 07 14:15:18 Yeah Mar 07 14:15:28 There's some plumbing to be done is all, really Mar 07 14:15:34 esp for am335x where we have "a lot" of memory Mar 07 14:15:50 BOOTCOUNT or similar just needs to be made SPL aware Mar 07 14:16:04 Tartarus: I believe that's a fairly common usecase, so it imho makes sense to add a weak function driving the spl logic and allow boards to override it with their own logic Mar 07 14:16:31 Jacmet: Right, what I'm saying is there's already a general facility, at the U-Boot level for doing this Mar 07 14:16:36 It's just not been made SPL-aware Mar 07 14:17:21 Tartarus: true Mar 07 14:17:55 Tartarus: the code becomes kind of hairy with all the options though Mar 07 14:18:34 Jacmet: It shouldn't be too bad. But yes, some weak functions can help make it cleaner Mar 07 14:18:44 I'm just saying, please take and extend BOOTCOUNT :) Mar 07 14:19:12 (ok, CONFIG_BOOTCOUNT_LIMIT) Mar 07 14:19:22 Tartarus: I will have a look, first I'll need to do a bit more work on the ti816x stuff Mar 07 14:19:34 Yeah, no rush Mar 07 14:19:39 And looking forward to reviewing those too :) Mar 07 14:19:45 Tartarus: ;) Mar 07 14:20:09 Tartarus: the delta from 814x is not too bad Mar 07 14:20:34 Yeah, that was my goal with the prodding I gave mdp Mar 07 14:20:46 I'd spent some time in an older vendor tree with all 3 platforms Mar 07 14:21:09 hrm, wha? Mar 07 14:21:23 ahh, yes Mar 07 14:22:07 Jacmet: I'm working on the kernel side now..hoping to leverage your 816x sweat in u-boot as I do have that platform to test too Mar 07 14:23:29 Tartarus: ;) Mar 07 14:24:10 mdp: it's mainly Antoine's work, I just want to use it and have helped a bit with the emac stuff and testing Mar 07 14:24:20 Tartarus, some of us are still trolling those vendor trees for bits :( Mar 07 14:25:04 mdp: some of us have nearby deadlines that force us to do products based on those trees :( Mar 07 14:25:08 Jacmet: I'm just happy to see it about to go upstream Mar 07 14:25:16 Jacmet: I feel your pain Mar 07 14:25:59 I've been a victim of TI vendor trees too..had to support a customer onsite with PM issues on 2.6.32 ;) Mar 07 14:26:11 a most horrid week Mar 07 14:28:25 mdp: :/ Mar 07 14:28:35 * KotH hands mdp a piece of chocolate Mar 07 14:29:08 where were you that week? :) Mar 07 14:41:04 mdp: probably in switzerland Mar 07 14:42:01 yes Mar 07 14:45:05 jackmitchell, could you try it with 'nohlt' in the commandline? then perf works for me :) Mar 07 14:46:43 mdp: unless it was somewhen when i was in japan, reading manga ;) Mar 07 14:47:07 shoragan: how would I put that in the command line? Mar 07 14:47:12 perf nohlt ./app Mar 07 14:47:17 kernel command line Mar 07 14:47:32 ah I'll give it a go Mar 07 14:48:11 it disables the WFI (wait-for-interrupt) instruction Mar 07 14:50:11 is it bad to not mention to annoying colleagues about the buffer overflows in their network applications and instead setup cronjobs to trigger them on a regular basis? ;) Mar 07 14:51:40 shoragan: yep, that fixes it Mar 07 14:52:24 Jacmet: I don't know about bad, but it sounds like mischievous fun ;) Mar 07 14:52:32 jackmitchell: ;) Mar 07 14:52:52 jackmitchell: now I'll just have to wait until he needs to demo it to somebody important :P Mar 07 14:53:09 maybe some TI ppl know how WFI would interfere with perf/PMU? Mar 07 14:53:37 shoragan: hmm, I just tried perf record ./app, perf report, and it gave an empty file... Mar 07 14:53:52 the perf.data file has no sample Mar 07 14:54:51 maybe the sample clock has something to do with WFI and therefore perf is working because it is never actually invoked? Mar 07 14:55:05 and therefore not working ;_ Mar 07 14:55:06 ;) Mar 07 14:56:37 try perf stat md5sum /boot/uImage Mar 07 14:56:42 or something like that Mar 07 14:56:48 it shows plausible values Mar 07 14:57:35 shoragan: it does Mar 07 14:58:29 but for perf record something is still strange Mar 07 15:00:13 does 'perf test' work for you? Mar 07 15:00:43 shoragan: all fail bar 3 Mar 07 15:00:57 for me as well Mar 07 15:07:56 bah oe-core systemd is a sad state Mar 07 15:13:19 jackmitchell: it is Mar 07 15:13:50 jackmitchell: the TSC has been discussing that, but it's only a symptom of a bigger problem Mar 07 15:14:19 koen: yeah, what do you think the root of it is? Mar 07 15:15:33 jackmitchell: @intel.com and @windriver.com addresses get to live by different rules Mar 07 15:15:41 I mean, the systemd patch has NAK'ed by multiple people Mar 07 15:15:46 and it still went in Mar 07 15:16:08 koen: what, are you ruled my a spaceman too? Mar 07 15:16:12 and even if they get to live by the same rules, the perception is that they don't Mar 07 15:16:14 by a* Mar 07 15:17:59 koen: hmmm, it will be interesting to see the solution that comes by 1.4 Mar 07 15:18:58 to this date I still can't even get systemd booting using my custom distro; i've had some ideas as to why but the whole INITMAN/backfill makes it a winding road Mar 07 15:19:08 jackmitchell: exactly Mar 07 15:19:42 jackmitchell: I had a few minutes today to spend on 1.4 and I 'wasted' in on fixing other layers so the metadata would actually parse :( Mar 07 15:20:02 :/ Mar 07 15:20:32 jackmitchell: I think it will come down to otavio, martin and me making meta-systemd override oe-core and make it work again Mar 07 15:20:47 I dislike that solution for obvious reasons Mar 07 15:21:10 koen: it doesn't sound ideal Mar 07 15:22:43 it downright sucks Mar 07 15:22:46 anyway Mar 07 15:22:56 I also found out I need python3 for my target Mar 07 15:24:24 * koen hasn't contributed towards a solution for systemd or python3 and has to shut up now Mar 07 15:24:55 ;) Mar 07 15:28:05 a great plague has come upon us Mar 07 15:28:12 indeed Mar 07 15:30:47 is it purple? Mar 07 15:33:06 of course Mar 07 15:35:01 * KotH paints it red Mar 07 15:35:11 hello koen, panto and prpplague: saviors of all things #beagle. :) Mar 07 15:35:36 jkridner++ Mar 07 15:35:37 * prpplague is doing intel things atm Mar 07 15:35:38 hello beaglemom Mar 07 15:35:39 jkridner, I added something to the pinmux-helper driver that is of interest to you Mar 07 15:36:00 got word from felipe/ravi this morning that they should be giving priority to addressing the USB drivers on 3.8 in the short term. Mar 07 15:36:08 you can define more than one pinmux states for it, and switch among them in runtime Mar 07 15:36:13 * jkridner kick prpplague Mar 07 15:36:21 boot to the head! Mar 07 15:36:33 jkridner: fixing the crash on shutdown would be a nice start, replugging a nice second Mar 07 15:36:43 jkridner: but you have to admire the timing Mar 07 15:36:47 jkridner, required for the TI 3.8 tree...good job "aligning" Mar 07 15:36:48 I had those in opposite priority. Mar 07 15:37:04 jkridner: usb replug didn't work for white and now it doesn't work for the new one Mar 07 15:37:14 now its a feature Mar 07 15:37:19 you have to keep it that way Mar 07 15:37:22 koen: is g_multi working well for you now? Mar 07 15:37:26 working on mac? Mar 07 15:38:08 mac running linux, yes Mar 07 15:38:18 ah. :-| Mar 07 15:39:12 I pushed the g_multi changes to the angstrom meta-ti repo Mar 07 15:39:19 please test and report back Mar 07 15:43:20 koen, is plugging and replugging stuff really a requirement for USB? Mar 07 15:43:27 heh Mar 07 15:43:35 which recipe should I build? cloud9-image Mar 07 15:44:26 jkridner: whichever you want that includes the g_multi stuff Mar 07 15:44:45 koen: just drop USB.. problem solved Mar 07 15:44:51 you could, you know, just build the recipe and install the ipk Mar 07 15:44:54 instead of a complete image Mar 07 15:44:58 crazy talk, I know Mar 07 15:45:04 might actually get things done and all Mar 07 15:45:44 k. Mar 07 15:45:55 do you know which one? Mar 07 15:46:13 yes Mar 07 15:46:27 but I feel childish and leave it as an exercise to the reader Mar 07 15:47:04 * koen has been dealing with sw sweatshops again today and is very cranky Mar 07 15:47:20 * KotH hands koen a bar of white chocolate Mar 07 15:48:00 KotH: I am cranky too Mar 07 15:48:03 mmmm... tasty. but not "real" chocolate. Mar 07 15:48:07 argh. crap. 2012.05 is still the default. :( Mar 07 15:59:48 I'm thinking of that xkcd comic depicting happiness over time last xorg.conf has been opened Mar 07 15:59:53 same goes for uboot Mar 07 16:00:51 http://xkcd.com/963/ Mar 07 16:01:36 * jkridner issues 'MACHINE=beaglebone ./oebb.sh bitbake gadget-init' and hopes it builds the right .ipk. Mar 07 16:03:24 * Tartarus gets depressed, flings feces towards gqq Mar 07 16:03:48 heh Mar 07 16:04:26 * gqq is already covered in his own feces; sanity was last seen when cloning the uboot repo Mar 07 16:06:20 gqq: where did you end up job wise? Mar 07 16:06:23 uni finished? Mar 07 16:07:12 av500, yep with half much more credits than actually needed ... looks like I was worried for nothing :P Mar 07 16:07:34 been working with the TrustZone for 6 months, it was really fun Mar 07 16:07:58 but they had no money to keep me so now I'm 'securing' an aircraft IFE system Mar 07 16:08:03 urgh. oebb.sh keeps creating .oe/environment-angstromv2012.05 even though I'm on the 2012.12 branch Mar 07 16:08:13