**** BEGIN LOGGING AT Sun Jan 29 02:59:58 2012 Jan 29 03:06:08 load a register with the address and move it to the pc? Jan 29 03:30:54 quiet here... too quiet. Jan 29 03:40:42 i got dropbear working Jan 29 03:40:56 but i cant log in as root cause it wont except 'blank' as a password Jan 29 03:54:24 got it working Jan 29 04:07:48 anyone get login incorrect on narcissus angstrom build with xfce..works with console only builds Jan 29 05:00:43 can beagleboard do everything beaglebone can do? Jan 29 05:04:35 that's rather general Jan 29 05:04:53 i'm new to this too but i thought beaglebone was supposed to be a BB stripped down Jan 29 05:05:10 not quite Jan 29 05:05:27 hey again Jan 29 05:05:36 way more expandable, I/O breakout headers Jan 29 05:05:40 there are some small hardware change Jan 29 05:05:56 ok fair enough, i was referring more to the low level stuff.. Jan 29 05:05:56 its a diff CPU too, very close but different Jan 29 05:05:59 forgot to consider that Jan 29 05:06:09 one have different fpu or something? Jan 29 05:06:13 no flash Jan 29 05:06:31 no nand doesn't really mean it can't do any less does it Jan 29 05:06:40 it's just a different method Jan 29 05:06:45 am3359 vs. OMAP3530 Jan 29 05:07:08 I know, just mentioning the differences I know about Jan 29 05:07:16 512MB vs 1G (on xM) Jan 29 05:07:17 my mistake, thought they were the same soc Jan 29 05:07:31 slower clock Jan 29 05:07:36 yes but Jan 29 05:07:37 ~700Mhz vs 1Ghz Jan 29 05:07:46 i think he asked if the beagleboard can't do something the beaglebone can.. Jan 29 05:08:01 not the reverse lol Jan 29 05:08:18 it depends on the project Jan 29 05:08:38 agreed though, the pin headers yea..missed that Jan 29 05:08:46 if you want to wire up stuff, then the I/O headers might be the thing, otherwise nah Jan 29 05:08:50 i'm still trying to get logged in with x11..not working Jan 29 05:09:46 well I've got a 7.5MB rootfs that boots in <2s and I can ssh in. I just got htop and rsync on there as well. That's about all I need in userland. Jan 29 05:09:54 Now to start kernel hacking Jan 29 05:10:46 do you know anything about OpenOCD support? Jan 29 05:12:27 more specifically, the JTAG over USB stuff Jan 29 05:15:38 my beagleboard is rev c4 Jan 29 05:15:42 i saw node.js was available to install on beagleboard Jan 29 05:15:48 from narcissus image builder Jan 29 05:16:00 less than 2 seconds hmm... Jan 29 05:16:16 how the hell did u manage that Jan 29 05:16:46 no udev, no sysvinit Jan 29 05:17:15 the pin headers on the beaglebone are probably level shifted to 3.3v or 5v Jan 29 05:17:23 3.3v Jan 29 05:17:26 ah Jan 29 05:19:59 they're going for an arduino thing where there will be add-on boards that are plug n play Jan 29 05:20:16 yah...i think thats a great idea Jan 29 05:20:32 each board has a ROM and the system can reconfigure the pins to what it needs on bootup Jan 29 05:21:07 there's already a DVI board I think Jan 29 05:21:53 its got video out already Jan 29 05:22:01 built in Jan 29 05:22:09 for beaglebone Jan 29 05:22:13 yah Jan 29 05:22:20 no it doesnt Jan 29 05:22:26 o Jan 29 05:22:54 ports are: usb host, usb gadget, microsd, rj45, and DC IN Jan 29 05:23:00 I didn't see a video out in the ref manual for the bone... Jan 29 05:23:13 don't have one, so I can't say with certainty Jan 29 05:23:39 there's a detailed hardware reference manual Jan 29 05:23:43 but I also have one right here Jan 29 05:24:01 http://boardzoo.com/product_info.php?products_id=86 Jan 29 05:24:29 HDMI out Jan 29 05:24:36 (addon) Jan 29 05:27:58 hey mrenouf, do you know how to get rid of all the unecessary stuff from the linux-omap git Jan 29 05:28:13 like you said udev is pretty heavyweight Jan 29 05:28:38 it was a manual process Jan 29 05:28:40 arm-linux-gnueabi-g++ with make file and defconfig's. I imagine it would be in the defconfig? Jan 29 05:28:56 care to elaborate? Jan 29 05:29:06 well, i enabled devtmpfs in the kernel Jan 29 05:29:15 that takes over /dev Jan 29 05:29:50 interesting Jan 29 05:29:50 the rest I built up from scratch mostly. just build busybox, and dropbear (ssh) and a few other little tools Jan 29 05:30:07 my startup script just does a few things Jan 29 05:30:34 this wouldn't be suitable for general use Jan 29 05:30:53 i'm sure it wouldn't Jan 29 05:31:20 there are just so many things i don't need in the linux-omap driver and i'm pretty new to this whole kernel hacking scene Jan 29 05:31:40 evidence by the huge devconfig with stuff i don't need Jan 29 05:31:48 by 'linux-omap driver' are you referring to the kernel? Jan 29 05:32:12 I didn't take much out of the kernel config Jan 29 05:32:33 except disabled video, sound, framebuffer Jan 29 05:33:05 it's tons of trial and error unfortunately Jan 29 05:33:58 i would stick with what you have, it's fairly lean already. if you want to boot into desktop then its going to take the extra time Jan 29 05:34:05 did i say driver Jan 29 05:34:25 lol sorry i'm trying to figure out how to find this driver right now Jan 29 05:35:16 yea my kernel stuff is around 10 seconds. to gui isn't much longer for console only Jan 29 05:35:21 x11 tho Jan 29 05:35:22 heh Jan 29 05:35:35 about a minute or 2 Jan 29 05:40:24 yep. thats about right Jan 29 05:41:12 could could probably get it to under a minute with lots of careful tweaks and trimming Jan 29 05:41:39 I would start with only X, and some minimal wm like fvwm Jan 29 05:47:25 i tried out the matchbox package Jan 29 05:47:35 that works Jan 29 05:47:47 but no root access through that wm Jan 29 05:48:16 i jsut don't see how console only will work but with any wm it just doesn't let me log in Jan 29 05:53:39 mrenouf..that was interesting..my network decided to kill itself; Jan 29 06:00:55 anyway you mentioned fvwm; i know this is going to be a bit ignorant but how would I go about getting it compiled for arm then installing the package Jan 29 06:16:12 compiling it yourself is too much to explain over IRC, but you could try pulling a package from debian Jan 29 06:16:28 thats what I do for a lot of stuff Jan 29 07:00:26 Heya Jan 29 07:03:27 Has anyone obtained more information from TI on the OMAP5 series? Despite there being a working devkit (for internal useā€¦), they don't release technical info. I doubt this stuff hasn't been written yet. Jan 29 07:07:32 <_av500_> too early Jan 29 07:08:08 Yeah :< Jan 29 07:08:58 How quick does TI expect business to move though? If things are expected to be in devices by the third quarter, TI must be working with companies in 'secret' Jan 29 07:10:05 like they always do Jan 29 07:10:11 <_av500_> omg, secrets Jan 29 07:16:54 I actually enclosed it in quotes to show I wasn't being paranoid or suggesting dodgy business. But I suppose it went the other way. Jan 29 07:26:52 * _av500_ makes note "doublebeta, not paranoid" Jan 29 07:28:53 * doublebeta coughs. Jan 29 09:26:07 woo. Jan 29 09:35:09 <_av500_> free energy? Jan 29 09:51:27 any day now. Jan 29 09:52:39 as you well know, having prototypes /= mass production. Jan 29 09:53:57 <_av500_> right Jan 29 09:54:07 <_av500_> chosing a color scheme is tough Jan 29 09:56:52 where do the camera data ports map to on the os? (ie how gpio pins map under /dev Jan 29 10:03:01 <_av500_> they dont Jan 29 10:03:07 <_av500_> unless you write a driver for it Jan 29 10:14:37 <_av500_> oops, wrong channel Jan 29 10:14:48 <_av500_> relax: they would be in /dev/videoX Jan 29 10:14:57 <_av500_> as in v4l2 Jan 29 10:15:12 <_av500_> assuming there is a driver of course Jan 29 10:24:25 have other members of community managed to get other cameras working via hacking the aptina drivers? Jan 29 10:54:50 / Jan 29 11:32:24 http://www.bbc.co.uk/news/uk-england-dorset-16754531 - distinctly Fortean. Jan 29 12:45:56 i saw 3 diff windows last time Jan 29 12:46:05 like beagleboard Jan 29 12:46:22 beaglebone Jan 29 12:46:22 and beagle Jan 29 12:46:28 wat happ y are those closed in IRC? Jan 29 15:16:30 hi Jan 29 15:22:00 yo Jan 29 17:01:43 <_av500_> bro Jan 29 17:05:23 sup Jan 29 17:46:02 not much. what i typed 6 hours ago is still on screen. Jan 29 20:30:01 Hello, I'm using latest angstrom version and I'm getting the following error when loading omap3-sgx-modules: PVR_K: (FAIL) SGXInit: Incompatible HW core rev (10205) and SW core rev (10201). Jan 29 20:58:41 Hi, quick question: Does anybody know how to I disable g_mass_storage on the beagle(bone) angstrom image? I want it to bring up the g_ether immediatly after booting w/o having to eject the mass storage device on the host machine. Jan 29 20:59:20 (uh, sorry for the typos) Jan 29 21:01:55 mattzz: been a while since I played with that, but I seem to recall there's a config file that sets the default. Jan 29 21:03:09 tasslehoff: thanks, I tried blacklisting g_mass_storage but that did not help. I'll keep searching for that config file ;) Jan 29 21:03:23 usb-gadget.conf maybe Jan 29 21:03:47 mattzz: perhaps http://cgit.openembedded.org/openembedded/tree/recipes/usb-gadget-mode/usb-gadget-mode.bb?id=a7756ec75b735de3900627f7786a74f9d91d3971 points you to some of the files you need Jan 29 21:05:23 tasslehoff: Ah, thanks a lot. That will definitely help. Appreciate it. Jan 29 21:06:15 * etheretic1 observes (as a non-programmer/hw hacker who sees a certain morally righteous commercial potential in the bb hw/community) that the bb faq is a hard hat area, apparently. Jan 29 21:08:03 (i.e morphing like an altarian changeloid-on-metasteroids) Jan 29 21:09:18 tasslehoff: found ./lib/systemd/system/storage-gadget-init.service. (A brute force 'find / -name "*gadget*" -print' helped...) Jan 29 21:11:28 Project: An OpenCPN/Linux-based navigation/sensor control- and monitoring system that'll blast the pants off more commercial offerings. Jan 29 21:11:33 Hello, I'm using latest angstrom version and I'm getting the following error when loading omap3-sgx-modules: PVR_K: (FAIL) SGXInit: Incompatible HW core rev (10205) and SW core rev (10201). Jan 29 21:14:02 mattzz: ah, yeah. forgot that systemd has replaced sysvinit :) Jan 29 21:14:47 ah, good old sysvinit days... Now off to the manpages of systemd. Jan 29 21:16:39 felipealmeida: appears to be a version issue according to http://www.beagleboard.org/irclogs/index.php?date=2011-07-11 . just a guess. Jan 29 21:18:14 etheretic1: thanks Jan 29 21:19:15 felipealmeida: search.yippy.com is my friend. Jan 29 21:20:09 beats scroogle/google as to actual usefulness. Jan 29 21:21:08 did the log answer your q? Jan 29 21:21:52 etheretic1: I don't think so. But I think I may have found the issue Jan 29 21:22:22 The libgles-omap3_4.05.00.03.bb recipe has the following instruction: BINLOCATION_omap3 = "${S}/gfx_rel_es3.x" Jan 29 21:22:31 And beagleboard xM uses ES 5.0 Jan 29 21:23:10 good. what are you up to with the bb? nature of project? Jan 29 21:23:25 search.yippy.com is unavailable to me Jan 29 21:23:43 etheretic1: I just want to run angstrom from master branch. Jan 29 21:24:02 the project actually is just a program that uses directfb 1.6 (from linaro with ARM NEON patches) over ansgtrom Jan 29 21:24:05 angstrom Jan 29 21:24:28 I'm using the gfxdriver omap, and so I needed the omapfb driver Jan 29 21:25:01 felipealmeida: it's quantum to me. 8) Jan 29 21:25:06 lol Jan 29 21:27:18 i used to use acorn (arm-based) comps for graphic design and layout in the 90's, and were involved in the network computer flop. Jan 29 21:28:49 so i became very excited to see an arm design being the engine in an open-hardware project. Jan 29 21:41:52 etheretic1, that and about 95% of the mobile phone and tablet industry Jan 29 21:42:51 mrenouf: aye. wilson and furber are gods. 8) Jan 29 21:43:44 whatever hapoened to amulet, the asynchrounous arm? Jan 29 21:44:43 (big powersaving issue) Jan 29 21:46:59 I guess it hit some issue that made it unpractical and they incorporated some things they learned into more normal designs Jan 29 21:47:46 * mru doesn't feel like digging through the corporate document archives today Jan 29 21:49:03 mru: http://apt.cs.man.ac.uk/people/sfurber/ Jan 29 21:51:07 brb... Jan 29 21:51:16 links to http://apt.cs.man.ac.uk/projects/processors/amulet/ which yields http://apt.cs.man.ac.uk/projects/processors/amulet/AMULET3i.php .it became quiet over 10 years ago. Jan 29 21:51:52 yes, presumably because there are some non-obvious problems with it Jan 29 21:53:19 "Extensive testing of AMULET3i has revealed only two bugs and simple patches have been identified for both of these." Jan 29 21:53:42 that's not the kind of problems I meant Jan 29 21:54:08 what, then? Jan 29 21:54:13 there could be problems with high-volume fabrication, interfacing with other things, sensitivity to interference, just about anything Jan 29 21:54:44 or simply that regular, clocked cores surpassed it Jan 29 21:54:58 it's a neat idea, it was probably worth exploring it Jan 29 21:55:02 your 2nd point sounds the likeliest, if any. Jan 29 21:56:17 modern cores are very asynchronous too, btw Jan 29 21:56:52 and extensive clock gating means unused parts use very little power Jan 29 21:59:57 as long as one is dealing with electrons and not photons, interference will be an issue at the macro-quantum level. but the async-ness of the amulet is "neural" in a way, much like oure brains work. we're not known for short-circuiting. 8) Jan 29 22:01:01 snaakje: 'evenin'. Jan 29 22:01:11 Mornin', eth :) Jan 29 22:01:17 Went to the boat today. Jan 29 22:01:23 Ff'in cold! Jan 29 22:01:31 how is she? Jan 29 22:01:32 On flipflops and without socks. Jan 29 22:01:35 White. Jan 29 22:01:38 :) Jan 29 22:01:43 snow? Jan 29 22:01:45 that cold Jan 29 22:08:40 anyone familiar with the beaglebone and it's expansion port capabilities? Jan 29 22:10:21 the 15 mm square version of the am3359 and am3358 has a two port 10/100/1000 ethernet switch on it. Is it possible to add a second phy to the board using a cape? Jan 29 22:12:35 hmm, I don't know what the cape-abilities of the board are Jan 29 22:21:25 etheretic1: where do you get all those sayings from yhou throw around here? Jan 29 22:22:04 er... sources. Jan 29 22:22:29 highly reliable sources Jan 29 22:22:31 is it dnc to u? Jan 29 22:22:32 I guess too much time, buy a bb and go on building the sw, that will need much more time than the hw Jan 29 22:22:42 sources who'll gladly tell you there's free energy to be had Jan 29 22:24:16 mru: well in banking, 1+1=whatever-they-want-it-to-be so who's to say "they" don't dictate "reality"? Jan 29 22:25:14 oh, you mean like this: http://www.physics.nyu.edu/sokal/transgress_v2/transgress_v2_singlefile.html Jan 29 22:27:02 Back to BB: any systemd wizzards around? I systemctl disabled a servie, I linked the service to /dev/null in basic.target.wants. Damn g_mass_storage keeps beeing loaded on reboot. Jan 29 22:33:09 What is the proper way to blacklist a kernel module (i.e. prevent auto loading) in angstrom? Jan 29 22:33:38 I'm sure deleting it will do the trick Jan 29 22:33:51 if you don't want it, don't build it Jan 29 22:35:29 well, I want the module but I dont like it to be autoloaded on boot. g_mass_storage might be handy at some point but it annoys me on boot-up because it prevents g_ether from being loaded at boot time. Jan 29 22:35:55 that is surely not autoloaded Jan 29 22:36:06 so grep in etc what loads it Jan 29 22:37:17 My guess was that systemd loads it. noting in etc about storage Jan 29 22:38:16 again, I'm using the demo image of anstrom for bb Jan 29 22:39:03 even if systemd loads it, something instructs it to do so. But maybe systemd uses a binary config ;) Jan 29 22:39:43 search for g_ in /etc, there can't be that many Jan 29 22:40:20 I assume it's a combined gadget with storage and ether Jan 29 22:42:19 but maybe loading of that is hidden in an unnecessary initrd. don't know if angstrom uses such. Jan 29 22:42:45 nothing in /etc and /etc/* regarding storage. storage and ether are two separate kernel modules. I will try renaming the module just for fun. Jan 29 22:43:11 I said grep for g_, not storage ;) Jan 29 22:44:02 g_ did not result in anything related to either storage or either. Jan 29 22:44:03 there are distinct modules for everything and combined ones. I don't remember how the combined ones are called, I just know g_* ;) Jan 29 22:45:56 g_multi is it Jan 29 22:46:45 there are distinct .ko files in /lib/module for ether, storage etc.. g_multi is one of them. Jan 29 22:50:48 ok, renaming did the trick. ugly solution, though... Kind of brute force Jan 29 22:55:18 Device refence manual mentions A/D converts on the BeagleBone, but I can't find mention of them in the AM335x manual. Anyone know what to look for in there? Jan 29 22:55:46 "Seven 100K sample per second A to D converters are available on the expansion header." Jan 29 22:56:21 they're probably part of the pmic Jan 29 22:56:58 oh, found 'em, they're labelled AIN[0:6] Jan 29 22:57:08 I bet, will keep diggin' Jan 29 23:00:14 nope, those are pins on the am335x Jan 29 23:06:38 ahh, got it: 12-Bit Successive Approximation Register (SAR) AD Jan 29 23:12:30 It's part of the touchscreen controller Jan 29 23:12:41 I have other uses in mind :-) Jan 30 00:24:00 http://www.amazon.com/ARM-System-Chip-Architecture-2nd/dp/0201675196 - the beagle bible? Jan 30 00:24:30 ooh, an antique Jan 30 00:26:28 pride o'th' shelf. Jan 30 00:28:16 issued 2000, last rewiew 2009. Jan 30 00:28:50 'night. Jan 30 00:30:32 in 2000, arm9 was cutting-edge Jan 30 00:40:36 what's wrong with ARM ARM? Jan 30 00:40:43 nothing whatsoever **** ENDING LOGGING AT Mon Jan 30 02:59:59 2012