**** BEGIN LOGGING AT Mon Feb 27 02:59:58 2012 Feb 27 05:31:33 so it looks only thing in the beacon board you can't use on the Bone is the EEPROM Feb 27 07:13:07 / Feb 27 07:17:55 hi, which usb wlan chipset is the best for an beaglebone with linux? Feb 27 07:19:10 one which works Feb 27 07:20:50 :/ Feb 27 07:33:09 jegade: any chip which works with x86-linux should work with the bone too. Feb 27 07:40:33 hello! Could anyone help me about beagleboard? My problem is the beagleboard doesn't appear on LAN. I can login over USB only, and when i reboot the beaglebone, it autodisconnect from usb and initialize eth0. After that it appears on the LAN. I would like to see the beagleboard on LAN without USB login and reboot. Feb 27 09:16:58 Is it too much to expect that ifup/ifdown should understand from my /etc/network/interfaces file that an interface is static and does not need a dhcp client? Feb 27 09:25:17 "...The first, was Telefónica’s support for the creation of open Web devices based on our ”Boot to Gecko” project. Also supporting the project are industry leaders, Adobe and Qualcomm. ..." Feb 27 09:25:26 way to go mozilla, way to go Feb 27 09:25:34 * koen buys av500 some gecko boots Feb 27 09:59:49 gah. nevermind. pebkac yet again. Feb 27 10:32:51 hi guys, i just want to share that, i fixed the problem which i ask couple days ago: Feb 27 10:32:52 i have beagleboard b7 and angstrom 2.6.32, so i use usb otg port which is connected to externally powered usb hub via usb mini B (pin4 and pin5 short circuited) cable. But i cant see usb devices.. even if loading g_ether or g_zero modules after boot up, i cant see. But when i plug out and plug in cable, then modules automatically been loaded, and i can see usb devices. So in this case whats the problem? Feb 27 10:34:02 The problem was beagle suppose itself, connected to a host machine. I force beagle to being host by: "echo host> /sys/devices/platform/musb_hdrc/mode" Feb 27 12:04:13 hello Feb 27 12:04:45 is anybody there who be familiar with beaglebone and ccs5? Feb 27 12:34:53 Koen? What the hell is Vaibhav talking about? What kernel? Feb 27 12:40:38 SilicaGel2: the e2e kernel Feb 27 12:40:52 That's T.I. ? Feb 27 12:40:55 a special kernel that is hosted on a forum, not a git Feb 27 12:41:07 well sheesh Feb 27 12:41:12 maybe in order to get around a company firewall... Feb 27 12:41:23 Who uses that kernel? Feb 27 12:41:42 pspeople Feb 27 12:41:44 His e-mail kind of had the tone I should have known about this but for crap's sake Feb 27 12:42:18 there are how many f'ing kernels, and whatever he's talking about wasn't in the T.I. linuxsdk nor angstrom, and somebody from T.I. told me "this is what you have to do" so we we f'ing did it Feb 27 12:42:24 How I feel like a complete retard Feb 27 12:50:19 ok. i answered/asked. Feb 27 12:50:52 it took me hours to figure out that the reset was the problem. It's definitely possible that somebody else fixed it, and it's quite cerain that his way of fixing it is correct and mine is wrong. I concede that. Feb 27 12:51:29 but in the linux-3.1.0-psp04.06.00.03.sdk kernel whatever he's talking about certainly was NOT there. Feb 27 12:51:56 SilicaGel2: vaibhav is trying to maintain a kernel tree that tracks upstream and only uses patches that have already been sent out for review to their upstream maintainers Feb 27 12:52:12 so is his the 'root' ? Feb 27 12:52:14 SilicaGel2: as you can imagine it is lacking support for pretty much everything at this time Feb 27 12:52:25 * SilicaGel2 reads that again Feb 27 12:53:29 koen: I hope jwinnebeck doesn't kill me for making him go through all this, figuring out how to make a proper patch, only to have vaibhav point out that it was completely unnecessary :-( Feb 27 12:54:03 oh, oh, unnecessary AND, apparently, a stupid way to do it, even though someone else at T.I. told me to fix it that way :( Feb 27 12:57:28 i have to get my arse to work, i'm late. Back in 30 minutes. Feb 27 12:58:33 SilicaGel2: it's unnecessary in vaibhavs kernel, that doesn't mean it's unnecessary in the kernel the beaglebone ships with Feb 27 13:01:48 SilicaGel2: you need to read vaibhavs comment as "You need that PRU clock patch on the kernel you are using. But if you switch to my upstream-vaibhav kernel, it is not needed" Feb 27 13:08:04 hello Feb 27 13:08:27 is anybody there who is familiar with beaglabone and ccs5 Feb 27 13:08:30 ? Feb 27 13:09:34 I can not coonect to the board? Feb 27 13:12:11 amin_: not many people use ccs5 Feb 27 13:12:16 even fewer use it with the bone Feb 27 13:12:21 you might be the first one Feb 27 13:12:27 (here) Feb 27 13:14:02 koen: you are right. I guess I am not very good at reading. Feb 27 13:14:11 dammit Feb 27 13:14:18 his way is better. Feb 27 13:14:50 That's it. I have had enough. I need to understand clkmod hwmod and power domain stuff. Now. I'll spend time on that today. Feb 27 13:30:57 SilicaGel2: for the long run, vaibhavs kernel is better Feb 27 13:31:25 SilicaGel2: but right now it lacks too much features to be considered for being the default beaglebone kernel Feb 27 13:31:31 so, the bone team is heading into the wrong direction with its kernel.... Feb 27 13:31:43 you guys should like upstream :) Feb 27 13:32:17 Oh you guys are talking about the hvaibhav kernel on github? Yeah I'm trying to look at that now to see if the reset code is in there Feb 27 13:32:23 let me put it like this Feb 27 13:32:28 He seemed to suggest that it is? Feb 27 13:32:31 the ppc/ppc64 -> powerpc merge took 3 years Feb 27 13:32:43 and that with with 4 platforms that weren't too different Feb 27 13:33:04 ls arch/arm/ | grep mach | wc -l 64 Feb 27 13:33:13 64 wildly different arm platforms Feb 27 13:33:21 Why does it take so long? Bureaucracy? Strict standards from upstream? Feb 27 13:33:31 lack of effort from vendor Feb 27 13:33:50 rmk Feb 27 13:34:27 all of the aboce Feb 27 13:34:29 above Feb 27 13:34:50 So koen, someone (ka6sox) said I can generate a pull request on github for meta-ti? That's not possible is it because meta-ti isn't on github? Feb 27 13:35:11 there's a meta-ti repo on github Feb 27 13:35:15 "generate a pull request"? what happened to posting patches on mailing listS? Feb 27 13:35:19 or even forums Feb 27 13:35:29 Well that's what I did av500... Feb 27 13:35:30 but as I keep saying: read the README, follow its instructions Feb 27 13:35:38 I did read the README Feb 27 13:35:40 koen: so e2e *and* github are now both requirements? Feb 27 13:35:45 it said send pul lrequests Feb 27 13:35:49 av500: and anonymous ftp Feb 27 13:35:51 if you said it's on github I will look again Feb 27 13:35:52 what else? facebook account? Feb 27 13:36:00 being left-handed? Feb 27 13:36:04 av500: paste patch in excel sheet, print it, place on wooden table, photograph it, have photo printed, scan photo, paste into word doc, place on shared ftp Feb 27 13:36:14 Now there are 3 kernels though, there is the beaglebone one in meta-ti, koen's kernel, and now Hiremath's Feb 27 13:36:15 mru: of course, thats normal Feb 27 13:36:25 but asking for stuff on top of that is insane Feb 27 13:36:59 koen: the readme is empty, at least from my perspective Feb 27 13:37:03 * koen heads against desk Feb 27 13:37:11 mru: looking at it sideways? Feb 27 13:37:22 jwinnebeck: I keep saying: 'my' kernel is the same as the on in meta-ti Feb 27 13:37:27 try to turn it around Feb 27 13:37:29 av500: no, from that perspective it doesn't even exist Feb 27 13:38:14 I bet it does not even use CJK doublebyte ascii chars Feb 27 13:38:16 So https://github.com/Angstrom-distribution/meta-ti is the same as git://git.angstrom-distribution.org/meta-texasinstruments and so I should "fork" the meta-ti and commit to my repo and issue a pull request from that? Feb 27 13:38:34 Because I'm just looking at where the source comes from Feb 27 13:38:51 and according to the layers.txt I got from angstrom, it's git://git.angstrom-distribution.org/meta-texasinstruments and not github Feb 27 13:39:24 * av500 wonders how Linus got his 1st patch into linux when there was neither github nor git Feb 27 13:39:59 maybe he brided A.Cox Feb 27 13:40:02 bribed :) Feb 27 13:41:20 braided? Feb 27 13:41:28 eternally Feb 27 13:42:26 OK so that github that I found it is not up to date with respect to the version I'm changing, it only goes to Feb 7 but there are commits up to Feb 24 in the angstrom git. Feb 27 13:42:47 So I still can't believe that's what you want me to work with? Feb 27 13:45:23 jwinnebeck: I said: follow the readme Feb 27 14:12:33 * SilicaGel whistles Feb 27 14:13:14 the RPi thread on the mailing list is interesting Feb 27 14:19:12 rpi is great because it can pipe gigE to /dev/null Feb 27 14:20:21 and because as long it is still hyped and not yet bubbled, everybody talking about it can feel like an ambedded geek who really knows what he is talking about. Feb 27 14:33:39 * SilicaGel chuckles Feb 27 14:35:05 people who have tried ubuntu on the beagle have found it sluggish Feb 27 14:35:17 so, what is to be expected from fedora on the rpi? Feb 27 14:36:24 demos/screenshots are more impressive than reality Feb 27 14:36:25 av500, can you point me towards some references speaking to why people found Ubuntu to be sluggish on the xM? I've not personally tried, but I'd like to learn Feb 27 14:37:12 bradfa: search #beagle irc logs Feb 27 14:37:49 is there a wiki somewhere for putting up notes/howto's? Spent a bunch of time cross compiling a kernel module for beaglebone and I would like to document what i did Feb 27 14:38:50 Red1: elinux.org and/or omappedia.org are good places to dump knowledge :) Feb 27 14:40:13 av500, Was there any one root cause that accounted for a large amount of the slowness? For example, the SD cards that ship with Beagles are rather poor for use as a file system. Was that a root cause? Feb 27 14:40:23 I am searching the IRC logs as well Feb 27 14:40:25 ubuntu is "designed" to run on typical desktop systems with multi-GHz intel CPUs, piles of RAM, and high-performance GPU Feb 27 14:40:30 av500: technically ubuntu doesn't run on pi, since it's an armv6 cpu Feb 27 14:40:37 sd cards in general are slow Feb 27 14:40:41 of course it's sluggish on a 720MHz, 256MB mobile system Feb 27 14:40:57 bradfa: all of the above Feb 27 14:41:09 av500, OK, thanks Feb 27 14:41:30 koen: I did not mention ubuntu on rpi Feb 27 14:41:34 thurbad, They're not all slow, some are much MUCH faster than others, especially for use as Linux file systems Feb 27 14:41:43 I said fedora, but we all know ubuntu > fedora Feb 27 14:41:52 simply put, the beagle is under-specced for ubuntu by a few orders of magnitude Feb 27 14:42:05 android runs fine from the same sd cards that supposedly make ubuntu crawl Feb 27 14:42:07 av500: right, I misread Feb 27 14:42:11 SD? or are you including compact flash, ad other flash variants? Feb 27 14:42:19 av500: I've been reading #pandaboard too much today :) Feb 27 14:42:20 because android is designed to run on such systems Feb 27 14:42:22 thurbad: does not matter Feb 27 14:42:26 thurbad, not including anything except SD and microSD in my assessment Feb 27 14:42:39 I'm not saying they're fast, just that some really really suck compared to others Feb 27 14:42:57 The Kingston cards are not very good, in general Feb 27 14:43:04 for Linux use with ext3 Feb 27 14:43:13 lets not make this into a discussion about card speed Feb 27 14:43:20 OK, sorry av500 Feb 27 14:43:23 np Feb 27 14:44:54 even android is slow to do things it was not desinged for Feb 27 14:45:03 like e.g. manage 30k songs in its database Feb 27 14:47:29 In my experience android is painfully slow on the XM Feb 27 14:48:16 ubuntu is snappy though, that is until I startx ;-) Feb 27 14:48:17 missing the sgx drivers? Feb 27 14:48:24 smplman: :) Feb 27 14:48:24 av500: possibly Feb 27 14:48:53 I saw a talk that claimed that ubuntu wont even move a mouse pointer without SGX, so I believe you... Feb 27 14:49:33 that's true Feb 27 14:49:51 ubuntu without hw opengl is unusable Feb 27 14:50:07 ok, the mouse pointer does move Feb 27 14:50:27 Ubuntu is getting rid of KDE, so they are DEAD TO ME!!! Feb 27 14:50:28 it just takes a minute or so Feb 27 14:50:54 SilicaGel: wait till they get rid of gnome too Feb 27 14:50:55 and lydia is switching to gnome :) Feb 27 14:51:06 av500: maybe if they do that I will like them again Feb 27 14:51:15 Android doesn't use X or anything like that, right? They invented their own thing from scratch? Feb 27 14:51:28 yes Feb 27 14:51:52 mru: what, no cursing about the missing network layer? Feb 27 14:51:57 Wondering what they would think about wayland, but that didn't exist when Android started (nor did OpenJDK and I wonder what they think about that) Feb 27 14:52:09 av500: I don't need that on a phone Feb 27 14:52:24 jwinnebeck: the motto is: don't think, ship Feb 27 14:52:42 We have a similar phrase around here. If it compiles, ship it! Feb 27 14:53:04 jwinnebeck: over here it's: CTRL-S succeeded, ship it Feb 27 14:53:13 wow that's hardcore Feb 27 14:53:24 We're not that hardcore, we at least run it through the compiler :) Feb 27 14:53:31 for hw it's: it draws power, ship it Feb 27 14:53:42 jwinnebeck: maybe they use a hex editor Feb 27 14:53:54 cut out the compiler, improve time to market Feb 27 14:54:13 one less thing to go wrong Feb 27 14:55:47 true Feb 27 15:49:21 Wow. The PRU timer example uses a hardware timer. It configures said timer (from userspace) by opening /dev/mem, mmapping it, and tweaking the timer directly. can you say yuck????? Feb 27 15:49:46 yuck Feb 27 15:49:51 duck Feb 27 15:49:55 luck Feb 27 15:49:57 much Feb 27 15:49:58 There must be a mechanism for allocating timers in the kernel Feb 27 15:49:59 nope Feb 27 15:50:02 * koen fails Feb 27 15:50:30 This is probably goign to be the thing that pushes me to use my own driver and not use uio_pruss. Feb 27 15:50:32 koen: did you sign a NBA with TI? Feb 27 15:50:49 Woah, like Dennis Rodman??? Feb 27 15:51:02 av500: I sang one a few years ago Feb 27 15:51:11 non bashing agreement? Feb 27 15:51:19 * ogra_ wasnt aware TI was into basketball at all ! Feb 27 15:52:25 koen: so when is it time for champagne this afternoon? Feb 27 15:52:47 * av500 had his already Feb 27 15:52:52 somewhere around now I guess Feb 27 15:52:56 LetoThe2nd: when are you at EW? Feb 27 15:53:02 What are we celebrating? Feb 27 15:53:02 * koen is waiting for clint to get into the office Feb 27 15:53:21 jsabeaudry: koen working for a real Texan Feb 27 15:53:22 av500: wednesday afternoon, and evening still around nuremberg Feb 27 15:53:25 boots and hat Feb 27 15:54:14 * ogra_ sends a "yee haw" to koen Feb 27 15:54:24 koen: and you mean "riding into the office"? Feb 27 15:54:32 hoooooooowdy! Feb 27 15:54:35 sweet I'll get my whip, my lasso and my revolvers Feb 27 15:54:44 branding iron Feb 27 15:55:12 * av500 wonders if there are magnastat tips for that Feb 27 15:55:28 av500: when will you be there? Feb 27 15:55:34 so far not at all Feb 27 15:55:42 ah yes Feb 27 15:55:53 lets check trains Feb 27 15:59:27 hrm when i was running T.I. linuxsdk I knew why one of these gpio/user LEDs blinked, beacuse it showed up in ps ... in this angstrom build I don't see what's making it blink Feb 27 15:59:45 kernel led trigger Feb 27 15:59:57 really Feb 27 16:00:07 i think so Feb 27 16:00:26 haha somebody else asked this question I think on th emailing list about an hour and a 45 ago Feb 27 16:01:55 I noticed with angstrom it appeared to me that the LEDs blink based on IO/network/etc, I just figured it was a "feature" implemented there Feb 27 16:02:31 yes, kernel led trigger connect to network activity Feb 27 16:03:33 hmm no the LEDs on my beaglebone don't appear to be related to I/O or network I just started a heavy I/O operation and pipe it all to SSH and lights didn't change Feb 27 16:03:45 except LED 1 came on solid Feb 27 16:03:59 LED 0 blinks at a steady rythm though Feb 27 16:04:06 SilicaGel: btw, it's 'TI', not 'T.I.' Feb 27 16:04:14 Thank you, I'll try to remember that Feb 27 16:04:24 saves typing :) Feb 27 16:04:29 short blink, short delay, short blink, long delay, repeat Feb 27 16:04:47 jwinnebeck: it's a heartbeat, done by the kernel Feb 27 16:05:06 Is Texas Instruments like AT&T where it's no longer "Texas Instruments" but it's just "TI" now Feb 27 16:05:17 everyone seems to be doing that now Feb 27 16:05:22 sorry, it's not AT&T anymore, it's at&t Feb 27 16:05:33 because capital letters are no longer cool Feb 27 16:06:06 koen: thanks I'll have to remember that heartbeat if it ever freezes for some reason, but latest systemd images seem to be very stable for me Feb 27 16:07:05 who is Thara is that someone we know? Feb 27 16:07:59 Thara & Tarun Feb 27 16:08:15 koen: do you know what LED1 is doing? LED0 is the heartbeat, LED1 seems to sometimes come on and stay on few a second or so and stay that way Feb 27 16:08:25 or is it just another heartbeat from another place in the kernel Feb 27 16:08:29 mmc access Feb 27 16:08:38 OK well it's not actually tied to mmc access Feb 27 16:08:41 look at /sys/class/leds/*/trigger Feb 27 16:08:51 it is tied to mmc access Feb 27 16:08:57 Well I run a dd and it doesn't come on Feb 27 16:09:00 as well as find at / Feb 27 16:09:14 buffers Feb 27 16:09:23 buffering... with dd ? Feb 27 16:09:39 sorry I mean Feb 27 16:09:44 dd if=/dev/mmcblk0p1 of=/dev/zero bs=4096 count=1000 Feb 27 16:09:58 I thought accessing the dev devices directly won't use buffers Feb 27 16:10:15 Ok I'm sorry I take it all back Feb 27 16:10:21 it seems to be turning on the light now Feb 27 16:10:46 http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=af8350c756cb48a738474738f7bf8c0e572fa057 Feb 27 16:10:53 it hooks into the mmc core Feb 27 16:11:13 yeah I believe you now Feb 27 16:11:25 I must have messed up my test, I think count=1000 ran too fast Feb 27 16:11:31 LED was off before I could observe it Feb 27 16:11:48 but it does set, count=10000 makes it even easier to see Feb 27 16:22:04 so koen, will you be coming to the uk before your last TI day? we should do a pub crawl Feb 27 16:24:01 mru: I think it's over, he sent his jar of marmite back Feb 27 16:24:19 and all the mighty white Feb 27 16:24:37 awesome the file omap_hwmod.c has a crapload of clear documentation at the beginning of the file!!! Feb 27 16:24:37 tetracel white? Feb 27 16:25:31 mru: last day was friday, but I should indeed schedule a trip for a pub crawl :) Feb 27 16:25:46 did you /quit ? Feb 27 16:25:52 t.i. Feb 27 16:25:53 yep Feb 27 16:25:59 koen: no notice period? Feb 27 16:26:03 * mdp hears "pub crawl" and perks up Feb 27 16:26:05 do you have new and exciting plans? Feb 27 16:26:12 mru: yes, but I don't need to work during that Feb 27 16:26:24 mru: so the next weeks are paid, but no TI work Feb 27 16:26:32 koen: untaken holiday? Feb 27 16:26:56 koen: you're not going to leave us, are you? :~( Feb 27 16:27:34 SilicaGel: he's going to work on Ubuntu for RPi FT Feb 27 16:27:47 mru: I'm should be signing my contract with cco today Feb 27 16:27:51 mdp: on fedora Feb 27 16:28:03 mdp: and getting rid of KDE? Feb 27 16:28:09 I heard fedora is running fully optimized Feb 27 16:28:21 saw a screenshot Feb 27 16:28:26 nothing is fully optimised Feb 27 16:28:43 * mdp assaults mru with zzzzzs Feb 27 16:28:46 optimisered Feb 27 16:28:51 I think /dev/null might be pretty close to fully optimized Feb 27 16:29:45 mru, I find we never quite even start to optimize, partially functional is good enough Feb 27 16:30:04 https://plus.google.com/106692013977020487319/posts Feb 27 16:30:07 * mdp like his zzzs Feb 27 16:30:12 mdp: then the next chip generation comes out Feb 27 16:30:26 oh no, zach again Feb 27 16:30:30 /undo Feb 27 16:30:41 av500: what? you think it should be "sach" ? Feb 27 17:10:58 hmmm Feb 27 17:11:06 +1 Feb 27 17:11:07 backlight control on a PMIC Feb 27 17:11:12 run Feb 27 17:11:13 is that a regulator or not? Feb 27 17:12:07 koen: in Arkansas it is: http://www.buchtips.net/rez780-die-regulatoren-in-arkansas.htm Feb 27 17:12:37 dunno about TX Feb 27 17:13:31 some idiot put a 1.8K resistor in the 1.0K bottle! Feb 27 17:13:49 bootle? Feb 27 17:13:53 bottle? Feb 27 17:14:08 yeah I keep all the SMT resistors and capacitors in little glass vials Feb 27 17:14:11 1000 Ohm ± 80% Feb 27 17:14:30 resistor in a bottle Feb 27 17:14:37 I think there's a song about that Feb 27 17:15:06 99 bottles of R on the wall Feb 27 17:15:09 is the song "When Will You Die" by They Might Be Giants? Feb 27 17:15:10 oh. Feb 27 17:15:27 oh, wait, it was a "message", nm Feb 27 17:15:34 av500, that one sounds better Feb 27 17:15:36 resistor in a message? Feb 27 17:15:45 messaging via resistors Feb 27 17:15:49 SMS or RMS? Feb 27 17:15:50 I have sent many messages that were received with great resistance Feb 27 17:16:00 LMS Feb 27 17:17:55 whew, I thought I lost my diodes, but i was able to rectify the situation. Feb 27 17:19:00 *cymbal crash* Feb 27 17:19:05 haha Feb 27 17:19:18 No way one of those RPi users woudl get THAT!!! fake embedded wannabees!!! Feb 27 17:20:05 I'm going to run WebOS on my RPi Feb 27 17:20:27 and then maybe Windows using OnLive Feb 27 17:21:06 I'll post screenshots Feb 27 17:22:14 SilicaGel: stay tuned for an RPi cape :) Feb 27 17:22:25 one cape to end all capes Feb 27 17:23:18 it'll be a plain paper printed picture of an RPi teetering on standoffs above a flyingbone cape Feb 27 17:23:27 genius Feb 27 17:24:34 one cape to rule them all Feb 27 17:37:44 OK, anemometer interface board is complete. https://plus.google.com/photos/108380917869049424169/albums/posts/5713870833708401682 Feb 27 18:38:49 Is using the GPMC just a matter of setting pin muxes, ioremap, and then using the memory map as if it was RAM ? Feb 27 18:39:22 yes Feb 27 18:40:24 Crofton|work, Wow, I can't believe it's that simple! Feb 27 18:41:37 well, you need to setup the gpmc timings Feb 27 18:41:40 everything is simple Feb 27 18:41:50 but the default settings will do something Feb 27 18:41:59 once you have figured it out Feb 27 18:42:05 indeed Feb 27 18:42:21 Crofton|work: gpmc was a piece of cake for you, I remember Feb 27 18:42:21 no doing it fast is different Feb 27 18:42:29 now doing it fast is different Feb 27 18:42:54 SilicaGel, This link leads me to a page with photos of a house and a bird, where is the anemometer interface board? Feb 27 18:43:17 Crofton|work, What do you mean by doing it fast? DMA? Feb 27 18:43:34 dma, bus timings, fpga code etc Feb 27 18:43:35 Crofton|work: I found that figuring out BCH, Chien search algorithm was difficult while under the gun of a production stop issue Feb 27 18:44:13 dma easy, caching not as easy :) Feb 27 18:44:28 that tends to ruin most otherwise interesting new things Feb 27 18:45:24 having the gpmc support code in the kernel correct must not be easy, btw, because it's still not done Feb 27 18:45:27 Crofton|work, Do you mean disabling caching? Feb 27 18:45:36 no Feb 27 18:46:04 Crofton|work, Why do you want to cache what the FPGA gives you? Feb 27 18:46:45 you want to cache invalidate what your read from the fpga Feb 27 18:46:52 and flush what you write to it Feb 27 18:48:04 jsabeaudry: weird. um. https://plus.google.com/108380917869049424169/posts Feb 27 18:49:01 salut, i have a problem with my beagleboard xm. I used to connect it at work with hdmi-dvi and everything worked fine. Now i tried to connect it to my monitor and home and I can't do it. The monitor still gives me a warning saying that its entering the power save mode and i have to move my mouse or press sth on keyboard (what Im doing, but still no reaction). What can be wrong? I would be really grateful for any help. Feb 27 18:50:02 Crofton|work, Is your GPMC FPGA work open? Feb 27 18:50:37 rofl Feb 27 18:56:35 jsa: that's bad too isn't it, I mus tsuck at internet. how about https://lh6.googleusercontent.com/-ZS6hVylUXr8/T0u_P_enhBI/AAAAAAAAIsA/2UipEOIZ_DM/w402/12%2B-%2B1 Feb 27 18:58:23 jsabeaudry, what are you trying to do? Feb 27 18:59:14 SilicaGel: is it because your post isn't public? Feb 27 18:59:18 The link worked for me Feb 27 18:59:22 hmmmmm Feb 27 18:59:25 but it's not a public post or image Feb 27 18:59:30 it should ahve been i thought hrm. Feb 27 18:59:52 click where it says "limited" it shows you who can see it Feb 27 18:59:54 it's not a public post Feb 27 19:20:40 SilicaGel, That last one works Feb 27 19:22:20 oh ok. So what that board does is pretty simple, takes the signal from my anemometer, puts some series impedance to limit the current, a couple of pull-up resistors, and over/under voltage protection diodes, then routes it to either the bone or my avr, not sure which yet Feb 27 19:22:27 Crofton|work, Right now just trying to plan course of action to take when I receive FPGA board, so I'm gathering info about what will need to be done and also looking for potential collaborators who are working on something similar which is hooking a fpga to the beaglebone as if it were a nand flash device Feb 27 19:23:24 hi guys can you help me make a speech ccrambler for my class project, please send me exactly how you would do it thanks Feb 27 19:23:41 I kind of wish taht wasn't the 1st thing on the beagle board haha Feb 27 19:23:59 why are you treating it as nand flash? Feb 27 19:24:31 lol Feb 27 19:25:14 SilicaGel: which one is that? Feb 27 19:25:36 the one where the kid is asking for help on what's obviously a class assignment and him geting pounced on Feb 27 19:25:39 which i approve of Feb 27 19:25:44 but Feb 27 19:26:07 Crofton|work, From what I could understand from the FPGA guys, this would be one of the simplest options, but I'll clarify that with them wednesday. Is there a better option? Feb 27 19:26:31 it kind of reminds me of how when ancient indiginous cultures killed their enemies, they put their heads on a stick at the entrance to the village as a warning for all other asshats who might want to come in and start something Feb 27 19:26:43 Which I also approve of. So I'm not sure what I'm wanting here. Feb 27 19:27:22 well, we just did nor falsh Feb 27 19:28:16 neither nor flash? Feb 27 19:32:18 Crofton|work, Ah yes, I'm not too knowledgeable about what the differences would be between those two Feb 27 19:32:27 We use ext3 on our uSD, and currently investigate hangs/pauses/hiccups when writing to it. We mount with async, but when data is written back we experience delays of 1-2 seconds. This is on a Class 6 Sandisk that should perform good. Is ext3 a bad choice? Feb 27 19:34:04 http://blogofterje.wordpress.com/2012/01/14/optimizing-fs-on-sd-card/ suggests that ext4 with no journal could be clever Feb 27 19:35:01 Having no journal would make it faster, but at a cost Feb 27 19:35:48 or with journal on a seperate partition Feb 27 19:35:54 you can always tune2fs the journal off though even with what he has now Feb 27 19:36:01 just to see what happens Feb 27 19:40:07 will do that. it is mostly fast enough, but the 1-2 second hiccups are not fun Feb 27 20:09:44 Hello... I configured my environment to cross compile for beagleboard xM through eclipse... but, when I put the binary file at BB and ran it, I get this message "sh: ./hellobeagle: not found"... when I run file hellobeagle, I get: "ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), not stripped" Feb 27 20:10:00 can anyone help me with this problem? Feb 27 20:10:41 chmod +x ? Feb 27 20:10:44 yes Feb 27 20:10:52 and, after 777 Feb 27 20:10:58 just for test Feb 27 20:17:20 ldd hellobeagle Feb 27 20:18:01 sh: ldd: not found Feb 27 20:18:21 install a real os ;) Feb 27 20:18:42 maybe I have to update some basic libraries at angstrom... is it needed? Feb 27 20:19:11 I am working at angstrom Feb 27 20:19:25 host: ubuntu --- target: angstrom Feb 27 20:19:41 what libraries are you linking against? Feb 27 20:20:55 I think these: stdio and stdlib Feb 27 20:21:07 my code is here: (just a hello world) Feb 27 20:21:18 #include #include int main(void) { //puts("!!!Hello Embedded World!!!"); /* prints !!!Hello Embedded World!!! */ return 0; } Feb 27 20:24:32 build it static (with -s) Feb 27 20:24:55 do you know how do I set this in eclipse? Feb 27 20:30:04 in project properties->c/c++build->settings->c++ linker->general Feb 27 20:30:22 but it sayd -s is omit symbols, not sure that'swhat you want Feb 27 20:31:09 oh, sorry, -static I meant Feb 27 20:38:00 hmm, seeing those long-options with only one minus, hows the arm support from llvm? ;) Feb 27 20:45:02 i'd like to backup my beagle SD card image, will these commands do what i think (copy MBR & disk): Feb 27 20:45:07 dd if=/dev/sdb of=boot.bin bs=512 count=1; dd if=/dev/sdb1 of=filesys.bin Feb 27 20:46:35 userx-: er why not just copy the whole thing? Feb 27 20:46:43 how? Feb 27 20:47:02 mranostay: ah, i got it Feb 27 20:47:18 but i don't want to copy empty blocks Feb 27 20:47:28 it's only about 30% full Feb 27 20:48:29 then just copy the files or use tar Feb 27 20:49:22 and how will i make a similar bootable SD card afterwards Feb 27 20:49:54 sfdisk -d /dev/sdb1 >sfdisk.txt will save the partition-layout Feb 27 20:50:56 to recreate it sfdisk that should read "sfdisk -d /dev/sdb >sfdisk.txt" and "sfdisk /dev/sdb the second one is what might kill your hd, if sdb isn't the sd-card Feb 27 20:52:56 hmm, too many corner cases... Feb 27 20:53:03 and formating the partition(s) is still needed after sfdisk . Feb 27 20:53:11 what's the most straightforward way? Feb 27 20:53:27 preferrably 1-2 commands backup, 1-2 for restore Feb 27 20:53:49 buy a sd-card backup-sw Feb 27 20:56:52 too time-consuming Feb 27 20:57:33 i also tried this: dd if=/dev/sdb | gzip image.gz Feb 27 20:57:41 *gzip > Feb 27 20:57:58 but i didn't seem to get much compression out of it Feb 27 20:58:08 well that would be slower Feb 27 20:59:53 slower yes, but more space efficient, right? Feb 27 21:05:11 you would get amazing compression out of it if you ran zerofs first Feb 27 21:05:22 unless it's very full Feb 27 21:05:36 zerofs? Feb 27 21:05:52 it writes zeros to all places on disk where there is not a file Feb 27 21:05:55 yeah, it's a tool for ext partitions that takes all the unallocated blocks and zeros them Feb 27 21:05:57 yup yup Feb 27 21:06:05 so it lets gzip and friends trivially compress the long runs of zeros Feb 27 21:06:22 mount /dev/foo /foo;dd if=/dev/null of=/foo/zero;rm /foo/zero; umount /foo Feb 27 21:06:52 maybe a sync inbetween Feb 27 21:06:54 yeah well Feb 27 21:06:59 zerofs is an actual program though Feb 27 21:07:03 not that yorus isn't a program Feb 27 21:07:04 :) Feb 27 21:07:12 that doesn't do anything? Feb 27 21:07:18 it copies /dev/null to Feb 27 21:07:20 n/m Feb 27 21:07:21 haha Feb 27 21:07:26 he makes a ginormous file of zeros Feb 27 21:07:29 but that already are more than 2 commands, so too much corner cases and no time left Feb 27 21:07:29 yeah Feb 27 21:07:35 I thought it should say /dev/zero Feb 27 21:07:49 yes Feb 27 21:08:02 i'm sorry Feb 27 21:08:08 i told you the wrong thing, the program's name is: zerofree Feb 27 21:08:18 aholler, thurbad... thanks a lot! it worked! see you... Feb 27 21:10:06 hahah Feb 27 21:20:01 and because it's an sd-card, tr '\000' '\377' < /dev/zero >/foo/zero would be better than dd. Feb 27 21:23:22 does anybody use a bdi2/3k on BBxM? Feb 27 21:43:45 i'm not getting what's the kernel level interface I'm supposed to use to grab a gp_timer ... since i'm not supposed to call the stuff in hwmod directly, and it SEEMS to have the data structures in there to let me get what I want Feb 27 21:43:53 the only help I've found seems to point to using pm_something Feb 27 21:43:59 but this isn't really a power management function Feb 27 21:45:03 gptimer12 Feb 27 21:45:05 oops Feb 27 21:45:09 that one is secret Feb 27 21:45:12 12 would be fine Feb 27 21:45:22 it's too late, i already knew about 12 Feb 27 21:45:23 (that's an omap3 in-joke) Feb 27 21:45:35 oh Feb 27 21:45:37 DAMMIT Feb 27 21:45:41 then I don't get it Feb 27 21:45:55 .... but I responded anyway and now I look like a dumbass Feb 27 21:46:34 SilicaGel: it was used as a workaround in the kernel Feb 27 21:46:52 SilicaGel: but since it's a "secure" timer it's undocumented Feb 27 21:47:06 what makes it secure, the lack of documentation? Feb 27 21:47:14 SilicaGel: so no TI person can publicly talk about it without violating an NDA Feb 27 21:47:21 oh i see Feb 27 21:47:41 which was incredibly funny when the patch came up again on the linux-omap mailinglist Feb 27 21:47:56 well as long as it won't do something like send a message out that says "WELCOME NEW ALIEN OVERLORDS" I guess I'm OK with it Feb 27 21:48:00 because I, for one, do NOT welcome them. Feb 27 21:48:14 <_av500_> they are fine Feb 27 21:48:20 no they aren't Feb 27 21:48:36 when it was E.T. they were fine but he's just one alien. Contrast him with what you saw in Independence Day Feb 27 21:48:43 What if we end up with THOSE bastards Feb 27 21:49:25 if they read the us-patents, they all must think we are totally stupid Feb 27 21:50:24 it'll be like Childhood's End Feb 27 21:52:59 yeah Feb 27 21:53:05 that was the one where the aliens looked like devils right Feb 27 21:53:09 yeah Feb 27 21:53:20 that was a fantastic and creepy one Feb 27 21:54:26 but djlewis will see them before they come down to us, so at least we are warned ;) Feb 27 21:54:49 s/we are/we will be/ Feb 27 21:55:19 or maybe like ender's game... at war with a species we've never even seen Feb 27 21:55:49 or maybe it'll be like 3rd Rock from the Sun, where they are all zany, and one of them is the radio Feb 27 21:55:54 but I doubt it. Feb 27 21:56:55 btw. there is a directory named timers in Documentation, if you can live with gp_* Feb 27 21:57:03 s/with/without/ Feb 27 21:57:51 I need an actual hardware timer, I need to generate fairly rapid events that the PRU can observe Feb 27 21:59:37 hm i found somethign called dmtimer which is "OMAP Dual-Mode Timers" what is that i wonder Feb 27 21:59:38 * SilicaGel googles Feb 27 22:01:03 Uh oh, koen. I just found a blog that says "With 12 timers available on the OMAP 3530 there are quite a few real-time tasks that can be accomplished..." Feb 27 22:01:07 Do you need to find/kill that guy? Feb 27 22:04:00 http://www.kunen.org/uC/beagle/omap_dmtimer.html Feb 27 22:05:44 heh Feb 27 22:06:00 depends on how the timers are numbers Feb 27 22:06:06 does gptimer start at 0 or 1? Feb 27 22:27:10 Hi... Feb 27 22:28:15 I'm trying to compile angstrom for the beaglebone and I have some errors like: ERROR: Layer dependency intel of layer sugarbay not found Feb 27 22:28:36 Anybody know how to fix this sort of thing? Feb 27 22:29:04 I'm trying to get an image that has both the A/D converter and Dallas 1 wire drivers enabled. Feb 27 22:29:11 unless I missed something earlier, i say "insufficient data" Feb 27 22:30:52 I was able to successfully compile the demo distro. When I tried again with this line: MACHINE=beaglebone ./oebb.sh bitbake systemd-image. I get the ERROR: dependency response for sugerbar, crownbay, etc. Feb 27 22:33:09 <_av500_> SilicaGel: with 12 timers you can be real time only half of the day, you need 24 for real real time Feb 27 22:34:57 and a good cable, not like those used by cern Feb 27 22:35:40 <_av500_> oh, it was a bad cable in the end? Feb 27 22:36:06 they think so, but still are testing Feb 27 22:37:04 <_av500_> should have uses a wifi cable Feb 27 22:37:07 <_av500_> used Feb 27 22:38:55 type e² ;) Feb 27 22:39:21 eh, no, mc² Feb 27 22:55:36 hmm, i still wonder when the first omap5 tablet will come to light Feb 27 22:56:54 no news about that from mwc as it seems Feb 27 22:59:21 so I assume not befor christmas :/ Feb 27 23:03:40 and asus gets crazy now using qualcomm too, besides tegra3 Feb 27 23:03:57 Just in case this helps somebody with the same errors, I was able to start compiling by deleting those machine specific recipes under BSPLAYERS in conf/bblayers.conf. Feb 27 23:04:30 aholler: using qualcomm isn't all that crazy if you have the volumes required for them to talk to you Feb 27 23:04:48 how are they? Feb 27 23:07:23 are they cortex-a9 too? Feb 27 23:07:38 they have an a9-class cpu of their own making Feb 27 23:10:59 anyway, the availability of drivers are which will make the difference. at least for me. ;) Feb 27 23:13:53 so some mali-thingy would seem to be the best choice. but I assume no one uses that. :/ Feb 27 23:14:45 mali is used here and there Feb 27 23:15:10 but driver availability is not a problem if all you want to do is build a few million tablets Feb 27 23:15:18 you'll have your drivers Feb 27 23:15:36 it won't be hacker-friendly, but that was probably not the intention Feb 27 23:15:47 isn't mali competitive or why is that used that seldom? Feb 27 23:16:08 I don't know Feb 27 23:20:05 an omap5 without graphics but with sata would be nice too ;) Feb 27 23:20:19 aholler: samsung uses mali Feb 27 23:20:32 and st-e Feb 27 23:20:57 aholler: omap5 has sata Feb 27 23:21:30 then they should offer a version without graphics ;) Feb 27 23:21:32 <_av500_> mali is rising I would say Feb 27 23:21:42 <_av500_> all the cheap chinese are using mali now Feb 27 23:22:11 <_av500_> since they shop at arm for the MPU core, I guess they just click the "add mali" checkbox Feb 27 23:23:46 _av500_: you miss/forget to uncheck it Feb 27 23:23:51 i would like to build a moohshot type system with omap5s without graphics ;) Feb 27 23:24:22 aholler: does the presence of a gpu bother you? Feb 27 23:24:34 yes, the price Feb 27 23:25:14 not sure how much that is in relation to the rest Feb 27 23:25:20 <_av500_> you want to shoot it to the moon? Feb 27 23:25:27 * _av500_ is puzzled Feb 27 23:25:36 the saturn v did not have a gpu Feb 27 23:25:44 <_av500_> ah right Feb 27 23:25:55 na, server-farm from hp Feb 27 23:25:56 nor sata Feb 27 23:26:22 <_av500_> aholler: hp invented a server far? Feb 27 23:26:24 <_av500_> farm Feb 27 23:26:45 http://h17007.www1.hp.com/ca/en/iss/110111.aspx Feb 27 23:27:28 <_av500_> gee Feb 27 23:28:01 that combines both things i like, networking and (embedded) arm Feb 27 23:28:03 <_av500_> they copied http://www.flickr.com/photos/av500/5125778540/ Feb 27 23:28:49 <_av500_> ok, I omitted the shoe box around, seems to be important Feb 27 23:29:02 sd-cards aren't that usable for a server-farm ;) Feb 27 23:29:10 <_av500_> aholler: neither are buzzwords Feb 27 23:29:28 <_av500_> so, putting a few cpus together if the cpus are ARM suddenly makes it exiting? Feb 27 23:30:14 it makes arm competitive in the server market Feb 27 23:30:27 <_av500_> to be proven Feb 27 23:31:16 if the costs are less than x86-stuff and they get the 64bit barrier Feb 27 23:31:51 <_av500_> as i said, to be proven Feb 27 23:31:59 <_av500_> and many server jobs dont even need 32bit Feb 27 23:32:09 I don't see why it couldn't work Feb 27 23:32:47 if you are an isp or datafoo it makes a difference if you can but some hundreds servers in one rack or only some dozens Feb 27 23:33:11 besides the cost for power and cooling Feb 27 23:33:22 <_av500_> yes Feb 27 23:33:29 <_av500_> I am not saying it wont work Feb 27 23:33:42 s/but/put/ Feb 27 23:33:56 <_av500_> its just that the UK laws of physics are the same as the US ones Feb 27 23:34:02 <_av500_> (mostly) Feb 27 23:34:08 ? Feb 27 23:34:46 <_av500_> moving a certain amount of information per second needs a certain amount of energy Feb 27 23:35:06 I still thinks arms are more energy efficient than x86s with all their legacy stuff Feb 27 23:35:24 <_av500_> what legacy stuff do you need in a server? Feb 27 23:35:31 x86 Feb 27 23:35:32 <_av500_> you need sata for both Feb 27 23:35:41 <_av500_> you need maybe pci-e for both Feb 27 23:35:44 I mean the instrucion sets Feb 27 23:35:45 <_av500_> you need a memory bus Feb 27 23:35:49 <_av500_> aholler: gee Feb 27 23:36:01 <_av500_> this gets boring Feb 27 23:36:18 the instruction decoder isn't all that large compared to the rest of the chip Feb 27 23:36:25 <_av500_> exactly Feb 27 23:36:43 ok, so I await a intel* without a fan for less than $100 Feb 27 23:36:48 the x86 ISA is terribly annoying, but it's not to blame for lack of power efficiency Feb 27 23:36:54 <_av500_> and since nobody ever built a set of SOCs that were identical except the ISA, I dont buy any of that talk Feb 27 23:37:18 <_av500_> aholler: here: http://androidcommunity.com/orange-intel-santa-clara-android-phone-hands-on-20120227/ Feb 27 23:37:36 you can find atom boards at a decent price Feb 27 23:37:48 <_av500_> I am sure, no fan and less than $100 for the cpu/SoC Feb 27 23:38:15 <_av500_> we paid intel like ~30$ or so and we did not have a fan in the A9 Feb 27 23:38:30 <_av500_> over one year ago Feb 27 23:38:38 <_av500_> and it did full HD Feb 27 23:38:49 more like 2 years Feb 27 23:38:53 <_av500_> yes Feb 27 23:39:02 and eats still much more power than an arm Feb 27 23:39:09 <_av500_> sure Feb 27 23:39:13 <_av500_> it runs windows Feb 27 23:39:31 no wakelocks Feb 27 23:40:10 no clocks Feb 27 23:40:22 <_av500_> but that was not an SoC, it was a std atom+chipset Feb 27 23:40:23 anyway, ok, av500 is right, arm is dead Feb 27 23:40:28 <_av500_> no Feb 27 23:40:45 <_av500_> washin machines wont go x86 soon Feb 27 23:41:01 arm is dead? Feb 27 23:41:08 nor will phones, even if someone manages to make one Feb 27 23:41:13 only when i sleep on mine Feb 27 23:41:15 and even if they then manage to sell it Feb 27 23:42:17 why by an arm if you can some x86-compatible from intel which is as power efficient as an arm, as av500 tries to tell me? Feb 27 23:42:53 if they are equally power efficient, that aspect is no longer an argument for either architecture Feb 27 23:43:32 x86-compatibility is an aspect Feb 27 23:43:36 no Feb 27 23:43:47 not in your land, ask some admins Feb 27 23:44:09 if you're building linux servers running mostly custom software, cpu architecture does not matter Feb 27 23:44:47 linux servers with custom software? Feb 27 23:44:56 how about price point? Feb 27 23:44:59 google, facebook, etc etc Feb 27 23:45:32 between the first two and etc etc is a big big gap Feb 27 23:46:07 <_av500_> aholler: what I am trying to tell you is that 99.9% of servers run x86 and ARM has just started, yet intel will be dead because they are not "power efficient" Feb 27 23:46:14 <_av500_> as you say Feb 27 23:46:27 i never said that Feb 27 23:46:33 <_av500_> so, thats what you are tryin to tell e Feb 27 23:46:34 x86 compat only matters if you expect to be able to buy a DVD with 3rd-party software and run it Feb 27 23:46:34 <_av500_> me Feb 27 23:46:43 you afre trying to turn the words in my mouth Feb 27 23:46:47 <_av500_> :) Feb 27 23:47:14 mru: dvd? I thought Apple killed those plastic discs... Feb 27 23:47:17 <_av500_> not trying to do that Feb 27 23:47:42 <_av500_> aholler: but you said intel cannot be power efficinet due to x86 Feb 27 23:47:49 <_av500_> which I think is nonsense Feb 27 23:48:05 <_av500_> intel is atm not power efficient because nobody trolled them into being it Feb 27 23:48:22 I think they learned the lesson the hard way... Feb 27 23:48:24 <_av500_> like asus trolled M$ to sell XP cheap, by putting linux on netbooks Feb 27 23:49:04 <_av500_> dwery: who? Feb 27 23:49:12 Intel Feb 27 23:49:12 mru: no, as i said, you don't know admins. they get bad dreams if they have to exchange a seven year old apache against something new. So to bring arm in the server market you need some arguments, and several hundred servers in one rack is exactly such an argument Feb 27 23:49:22 _av500_: we trolled intel pretty good at elc Feb 27 23:49:23 <_av500_> dwery: because they are doing badly atm? Feb 27 23:49:44 aholler: if you're building a new system, why should you care? Feb 27 23:49:48 Ubuntui everywhere! Feb 27 23:49:54 I think they are perfectly conscious that they have to be power efficient and are working into that Feb 27 23:50:00 I'm not expecting anyone to _replace_ existing intel systems with arm Feb 27 23:50:01 <_av500_> dwery: of course they are Feb 27 23:50:19 <_av500_> they are since a long time Feb 27 23:50:46 I almost burned my fingers on the atom they were demoing at elc... Feb 27 23:50:57 mru: so how long you want to test your newly baked sw until you say it's production ready? Feb 27 23:51:00 <_av500_> mru: you can burn your finger on an omap4 as well Feb 27 23:51:11 <_av500_> try 4460 at 1.5ghz Feb 27 23:51:16 aholler: whatever the figure, it'll be the same on intel and arm Feb 27 23:51:35 _av500_: haven't tried that, but the samsung exynos does run pretty hot at 1.2 Feb 27 23:51:49 but that's without a giant heatsink Feb 27 23:51:50 no, x86 is available since a long time and you can still use the old binaries Feb 27 23:52:06 but then you are not building a new system Feb 27 23:52:18 nobody expects arm to replace intel, but there's definitely a move toward a server market (arm) and a move toward being power efficient (intel) Feb 27 23:52:36 there is a reason why debilian updates only every 3 year or so. Feb 27 23:52:50 <_av500_> yes, they are debian Feb 27 23:52:55 they're chickens? Feb 27 23:52:57 bawk bawk bawk! Feb 27 23:53:03 nobody said you have to update more often on arm Feb 27 23:53:18 yes, just sw you baked 3a ago ;) Feb 27 23:53:22 while true ; do sleep 1 ; opkg update ; done Feb 27 23:53:36 keyword, I go too sleep Feb 27 23:55:39 <_av500_> ok, so arm on server will take 3 more years, due to debian Feb 27 23:55:42 if you intend to run 3y old software, you are not building a *new* system BY DEFINITION Feb 27 23:56:06 i've just learned arm is dead for servers, no need to discuss more Feb 27 23:56:12 <_av500_> ... Feb 27 23:58:26 and I know why if server people have to discuss with embedded people they will get crazy ;) Feb 28 00:00:20 is this the right place to look for help building openembedded for the beaglebone? Feb 28 00:00:29 <_av500_> yes and no Feb 28 00:00:32 <_av500_> just ask :) Feb 28 00:00:42 is this the right place to be on-topic? Feb 28 00:00:55 hehee Feb 28 00:01:02 :) Feb 28 00:02:02 Been trying to get a build to work following the directions based on cloning the setup-scripts repo from angstrom and running into a variety of errors Feb 28 00:03:04 I'm trying to figure out if the problem is that I'm running it in a vm, or that the vm is debian 6 amd64 or that there's something else I'm getting wrong Feb 28 00:04:12 on an ubuntu machine, I was able to get openembedded-core to build, but I'm a bit fuzzy on to configure the bitbake to build for the beaglebone Feb 28 00:09:31 Oren__, did you use the angstrom setup script? Feb 28 00:10:55 yes, I ran the setup script - on the vm when I try a "bitbake nano" it just hangs after getting to about 95% done with the parsing scripts Feb 28 00:15:40 the embedded application I want to use the beaglebone for doesn't need many of the things that seem to get built with the default setup - eg x11 - it seemed that even with the console-image that is part of the build Feb 28 00:22:59 Oren - have you seen this link: http://omappedia.org/wiki/OMAP_Angstrom The commands seem simple on the angstrom site (which appears to be back up) but I experience all kinds of inconsistent errors. I'm sure it's me, but it's a grand puzzle trying to figure out how to solve them. Feb 28 00:25:03 I haven't seen that one. But I also have found it to be a giant puzzle - it's nice to know I'm not the only one. Feb 28 00:28:21 ok, well, I suppose it's not too surprising, but increasing my vm's RAM to 2GB from 512MB seems to have improved well things are working (now it takes 13 minutes to parse, but it's now actually building things for the first time) Feb 28 00:49:50 d Feb 28 00:50:13 :q! Feb 28 00:51:55 8! Feb 28 00:52:08 * SilicaGell is now known as The Human Tripod Feb 28 00:52:47 i think I decided that, while cool, Xbee Programmable is not for me. Feb 28 00:53:19 SDCC for 'HC08 and 'HCS08 do NOT seem to be well supported AT ALL, and I see stuff about it generating, quote, "bad code, I mean BAD" Feb 28 00:53:34 and CodeWarrior only runs on billgatesware so I disavow it Feb 28 01:09:13 http://www.ti.com/tool/tmdxice3359 isn't this basically the beaglebone? Feb 28 01:11:29 tlab: the AM3359 is the same Feb 28 01:12:06 but i'll bet that thing costs 3x at least Feb 28 01:13:04 99 bucks Feb 28 01:13:36 hmmm Feb 28 01:15:06 yea I'm interested in the industrial communication stuff, seems the beaglebone is lacking the interface, but still has the options to add it Feb 28 01:15:29 wonder what all I can get working together Feb 28 01:16:15 very little documentation yet for that board Feb 28 01:16:16 it says it's $99.00 **** ENDING LOGGING AT Tue Feb 28 02:59:59 2012