**** BEGIN LOGGING AT Wed Nov 14 03:00:00 2012 Nov 14 04:29:03 hi all Nov 14 04:33:35 lo mranostay Nov 14 04:34:39 * mranostay starts writing his first driver in about 3 years Nov 14 04:35:59 rusty? Nov 14 04:36:38 oh yeah Nov 14 04:38:11 if i submit my paper to the US ELC i should be able to have a nixie tube for a display :) Nov 14 04:38:51 80x25 nixie tube? Nov 14 04:39:00 har Nov 14 04:39:33 it is the old school answer to the beagle wall Nov 14 04:40:27 mranostay: US ELC will be a little closer to you Nov 14 04:40:34 emeb: a lot :) Nov 14 04:40:45 don't have to worry about loosing the demo to security Nov 14 04:40:48 What's that - like 15 minutes up the freeway? Nov 14 04:40:55 ds2: actually a 9 digit VFD Nov 14 04:41:03 emeb: more like an hour Nov 14 04:41:07 9 digit? Nov 14 04:41:10 that is disappointing Nov 14 04:41:21 I got a pile of 40x4 VFDs lying around here Nov 14 04:41:39 http://www.decadecounter.com/vta/articleview.php?item=408 Nov 14 04:42:48 that's no challenge Nov 14 04:43:55 did you see that 1-digit nixie clock some guy did wired free-form? Nov 14 04:44:17 http://dangerousprototypes.com/2012/10/10/freeform-nixie-clock/ Nov 14 04:44:53 the HV supply is a switcher that's controlled in software! Nov 14 04:46:00 given that they used to make geiger counter using a pushbutton to make/break a coil... Nov 14 04:46:18 heh Nov 14 04:46:44 it is some challenge at least :) Nov 14 04:46:58 more to have another capebus example Nov 14 06:30:33 my connection feels so damned slow now http://arstechnica.com/business/2012/11/google-fiber-is-live-in-kansas-city-real-world-speeds-at-700-mbps/ Nov 14 06:32:42 Russ: are you planning a DDoS scheme? Nov 14 06:33:08 with my paltry 35Mbps uplink? are you kidding? Nov 14 06:33:31 depends on who you are DDoS'ing Nov 14 06:33:57 the bastards with the 700Mbps async connection Nov 14 06:34:07 hahaha Nov 14 06:34:19 * mranostay hides from ds2 with his irc cloak Nov 14 06:34:21 but they probally have 699Mbps of torrant traffics already ;) Nov 14 06:37:29 the porn isn't downloading itself you know Nov 14 06:38:31 it is if you use the gdrive ;) Nov 14 06:38:45 porn symlinks ;) Nov 14 07:56:20 jkridner1: and noone uses omap2_plusdefconfig Nov 14 07:56:33 jkridner1: at least noone that wants to use a rootfs from after 2008 Nov 14 08:01:18 is it bad i've been up 22 hours and don't feel tired? Nov 14 08:04:16 jet lag is a b**** Nov 14 08:07:13 last part of Hunger Games should knock me out Nov 14 08:59:12 look at that Nov 14 08:59:26 european day for action and solidarity Nov 14 08:59:52 or as the south of europe calls it: strike! Nov 14 09:07:02 mranostay: Blümchen, nice :) Nov 14 09:14:31 hi everyone, do you know any software to check a webcam on BBxm running Angstrom? Nov 14 09:14:35 Thank you. Nov 14 09:24:16 cheese Nov 14 09:27:08 yes, I tried opkg install cheese Nov 14 09:27:21 but cannot install Nov 14 09:27:28 packet not found Nov 14 09:34:54 av500_: I asked about cheese last minuted! my PC just restarted Nov 14 09:35:24 I install cheese by this command: opkg install cheese Nov 14 09:35:39 the board said: Unknow packet Nov 14 09:35:53 How can I correct this? Thank you Nov 14 09:41:20 opkg update Nov 14 09:41:57 http://www.angstrom-distribution.org/repo/?pkgname=cheese Nov 14 09:44:36 So, I just download & run it on my board? Nov 14 09:45:40 opkg install or so Nov 14 09:46:01 I use buildroot, so I would not know :) Nov 14 09:47:40 yes, I use: opkg update & opkg install cheese but still unknown packet" Nov 14 09:48:11 So, I downloaded the .ipk file. How can I install this file, av500 Nov 14 09:48:28 ipkg Nov 14 09:48:34 opkg Nov 14 09:48:35 google Nov 14 09:48:39 I win Nov 14 09:49:58 av500: we need to google bomb for 'exact steps' or 'exact steps'+$animalboardnames Nov 14 09:50:24 dm8tbr: you want to outpanda panda with panda? Nov 14 09:52:18 dm8tbr: ozelotboard? Nov 14 09:55:52 'exact steps' for xyz on 'pandaborad' :) Nov 14 10:10:20 av500: I did: opkg install cheese_2.30.1-r1.9_armv7a.ipk but still cannot install cheese. Nov 14 10:10:31 Do you know another software? Nov 14 10:10:35 thank you. Nov 14 10:10:50 cannot why? Nov 14 10:13:12 sorry, I forgot go to the directory Nov 14 10:13:21 let me check it again Nov 14 10:26:31 av500_ : I got these errors: Nov 14 10:27:17 http://pastebin.com/2sDUtzRc Nov 14 10:28:27 yes Nov 14 10:28:54 and Nov 14 10:29:00 it can get no satisfaction Nov 14 10:29:04 how to I update the Nov 14 10:29:16 libfreetype6 Nov 14 10:29:32 mine just 2.4.8 Nov 14 10:29:48 through opkg I guess Nov 14 10:30:02 [10:46:00] I use buildroot, so I would not know :) Nov 14 10:30:15 if I was you, I would find out why the package feed is not working for you Nov 14 10:30:23 opkg list Nov 14 10:30:25 etc.. Nov 14 10:30:37 or use a newer angstrom image and work from there Nov 14 10:30:46 yeah.. Nov 14 10:31:11 yes, thank you. Nov 14 10:32:20 do angstrom have another software can replace for cheese Nov 14 10:35:12 mplayer? Nov 14 10:35:14 gstreamer? Nov 14 10:35:15 vlc? Nov 14 10:38:39 thanks koen, Nov 14 10:38:45 I try it Nov 14 10:43:42 koen: I installed gstreamer, so, how can I run it from the terminal Nov 14 10:45:00 is google down today? Nov 14 10:45:04 yes Nov 14 10:45:41 is bing up? Nov 14 10:46:17 let me check lycos Nov 14 10:46:48 Ted_: to be blunt, you way of conducting business here is bordeline annoying Nov 14 10:46:50 +r Nov 14 10:50:01 Sorry all. Nov 14 10:56:13 greetings fellow travelers Nov 14 10:59:34 iho Nov 14 10:59:37 hiho Nov 14 11:00:57 afk Nov 14 11:09:29 * koen stabs linux Nov 14 11:09:52 I finally hooked up the board+cape to a scope to see what is happening Nov 14 11:09:57 and why the down button doesn't work Nov 14 11:10:06 and I booted 3.7 Nov 14 11:10:22 where pinctrl informs me I have a mux conflict with spi_d1/gpio3_16 Nov 14 11:10:36 evil vendor 3.2 gives no such warning Nov 14 11:12:20 just edit the board file... Nov 14 11:12:43 that's what I'm doing right now Nov 14 11:13:12 while waiting for the build to complete I'm butchering some more xmas tree songs to be about DT Nov 14 11:14:23 down up right left? Nov 14 11:16:42 up up down down left right left right b a start Nov 14 11:17:04 back Nov 14 11:20:17 front Nov 14 11:23:22 koen: I mean the labeling on the LCD cape Nov 14 11:23:33 normally one puts left before right Nov 14 11:23:35 just saying Nov 14 11:23:43 maybe it was developed in .il Nov 14 11:28:57 .tx Nov 14 11:29:47 cue full metal jacket quote Nov 14 11:30:08 steers and queers Nov 14 11:30:30 http://marc.info/?l=openbsd-tech&m=135221032423220&w=2 Nov 14 11:32:03 nice quote, but I fail to see why lennart is relevant Nov 14 11:32:14 noone can force you to use his code or projects Nov 14 11:32:18 it's all open source Nov 14 11:32:23 if you don't like it, patch it Nov 14 11:32:46 or use ubuntu, that lacks most lennartware Nov 14 11:32:57 and everyone knows unity never crashes Nov 14 11:33:03 * XorA|gone laughs Nov 14 11:33:35 koen: ah come on, dont be so reasonable Nov 14 11:52:58 koen: please stop Nov 14 11:53:20 kay & lennart with the aid of a billion-dollar company are hell-bent on forcing their crap onto everybody Nov 14 11:53:26 and they're being pretty darn successful Nov 14 11:54:57 anyway, I have a retrocomputing challenge for the likes of mdp Nov 14 11:55:07 hook up a beaglebone to this: https://lh6.googleusercontent.com/-AacYF28_m1w/T__mLkILEWI/AAAAAAAAB80/veoHjkE6D24/s0/ Nov 14 11:59:05 mru: what is it? Nov 14 11:59:05 koen: did you see that my comment on your blog was classified as spam? Nov 14 11:59:55 keesj: it's a mechanical calculator Nov 14 12:00:03 never seen one of those? Nov 14 12:00:11 I like a quote later in the thread much better: "If you want people to gain traction you will need to Nov 14 12:00:12 kids these days... Nov 14 12:00:16 reduce some standards... Nov 14 12:00:23 that's very true Nov 14 12:00:52 popularity is as always a race to the intellectual bottom Nov 14 12:01:11 it's just as true in high school as it is on kickstarter Nov 14 12:01:24 and I can hardly imagine lower than lennart Nov 14 12:01:33 I don't know about that Nov 14 12:01:44 that two-faced lying weasel Nov 14 12:01:45 I think that's a separate discussion actually Nov 14 12:02:03 I am a strong proponent of systemd Nov 14 12:02:09 why? Nov 14 12:02:33 tell me _one_ thing it does better than the best alternative Nov 14 12:03:09 and give me _one_ reason for bundling all manner of functionality together like that into a single process Nov 14 12:03:09 it does service supervision more thoroughly than alternatives Nov 14 12:03:19 meaning? Nov 14 12:03:44 I use daemontools everywhere, since way over a decade Nov 14 12:04:25 I wanted to write something exactly like supervise, but then I found daemontools, and am a very happy user Nov 14 12:04:56 but daemontools is too rough and minimal for a very large group of people who might otherwise benefit tremendously from using Linux Nov 14 12:06:38 systemd does a much more thorough job at service supervision, and in addition it comes with a GUI Nov 14 12:07:03 though I couldn't care less about the GUI for myself, that in itself is reason enough to support it imo Nov 14 12:07:44 even windows nt 3 had a GUI for managing services Nov 14 12:08:18 again, I am not a GUI person, but the masses are Nov 14 12:09:17 anything that allows me to ditch silicon vendor provided initscripts is a win Nov 14 12:10:59 koen: silicon vendor supplied init scripts??? Nov 14 12:13:43 KotH: yes, as part of the 'SDK' Nov 14 12:14:01 KotH: that should raise some red flags :) Nov 14 12:19:16 koen: i dont use SDKs unless i'm testing a eval board for the first time Nov 14 12:19:24 koen: after that, it's buildroot all the way :) Nov 14 12:22:08 CareBear\: the very moment, when you are not able anymore to debug your system because of the complexity of one of its core components, is the day you lose control over what your computer does Nov 14 12:22:19 CareBear\: and systemd (actually all lennartware) is going that way Nov 14 12:22:39 CareBear\: freedesktop.org/gnome also does a very good job at that Nov 14 12:23:24 CareBear\: and mind you, i dont care whether john doe can use a linux system or not. i care about whether i can use it to solve my daily problems or not Nov 14 12:23:50 CareBear\: and with lennartware/fd.org/gnome/... it's becomming more and more difficult every day Nov 14 12:41:18 I am trying to figure out what a "normal" core frequency should be on the beagle-xm. TI's code composer studio will start the omap code clock at 332 Mhz. (When I read the cycle counter I get about 340 Mh instructions per second). Nov 14 12:41:25 is that slow or normal? Nov 14 12:42:27 no idea? what does your application need? power or performance? Nov 14 12:42:35 Nurse, I need more trolling on #raspberrypi, stat! Nov 14 12:42:38 I saw u-boot does use some different diviers but I quite can't figure out the end result. (This is not on Linux) Nov 14 12:42:56 bradfa: what are they up to now? Nov 14 12:43:01 currently performance :P Nov 14 12:43:03 putty problems Nov 14 12:43:04 bitbanging something, I suppose.. Nov 14 12:43:28 mru, mainline bit banging? Nov 14 12:43:44 No, I am just trying to figure out if the u-boot defaults are sane. Nov 14 12:43:48 * bradfa wonders when keystone will get mainline... Nov 14 12:43:49 are they bitbanging wireless yet? Nov 14 12:44:45 keesj: which clock is 332MHz? Nov 14 12:44:54 dm8tbr: wireless bits? Nov 14 12:45:14 mru: yes, someone should suggest that to them Nov 14 12:45:33 'can I bitbang $wireless_protocol?' Nov 14 12:45:51 sure, just toggle the bit fast enough Nov 14 12:46:06 dm8tbr: didn't you say the n900 bitbangs usb? Nov 14 12:46:30 mru: essentially, yes. they put the USB controller in test mode and implement most of the protocol themselves. Nov 14 12:46:31 core dpll(3) is at 664Mhz and core clk at 332Mhz Nov 14 12:46:48 dm8tbr: 0_o Nov 14 12:47:20 that's because if you attach a host cable the additional usb phy will do weird stuff and not allow the musb to go to host mode Nov 14 12:47:25 keesj: and L3 is half again of that? Nov 14 12:48:01 yes (166) Nov 14 12:48:26 that should be 200 on the xm Nov 14 12:48:41 but stock u-boot/kernel get it wrong Nov 14 12:51:41 I have been reading the u-boot code but so far not 100% confident about where to look.(I am mainly poking registers at this point). Nov 14 12:52:19 but 340 M clock cycles / s is not "very slow" right? Nov 14 12:52:38 depends on which clock Nov 14 12:53:29 (this is reading the performance counter CCNT) Nov 14 12:57:24 that one counts cpu cycles Nov 14 12:58:12 and when measuring clock frequency with that counter, you have to make sure it's kept ticking all the time Nov 14 12:58:24 if you sleep the clock can be gated Nov 14 13:00:09 also in something like wfi? Nov 14 13:07:23 is there a more "easy" way to measure the cpu frequency? Nov 14 13:08:04 bradfa: mmmm...mainline keystones Nov 14 13:08:16 bradfa, tomorrow, the next day? Nov 14 13:10:12 keesj: yes, read the cycle counter twice with a second or so between Nov 14 13:10:51 just dont measure the second using the same counter Nov 14 13:11:10 gettimeofday() works fine Nov 14 13:12:25 I output to the serial after 5 seconds and use "ts" to be shure i also programmer the timmer correctly Nov 14 13:13:02 thanks Nov 14 13:13:16 lunch Nov 14 13:14:07 keesj: http://hardwarebug.org/files/clock.c Nov 14 13:15:23 what will this do for me? Nov 14 13:17:59 nice Nov 14 13:23:53 It will estimate the amout of intructions the cpu can run per second Nov 14 13:24:27 android drops blueZ for BCM stack Nov 14 13:24:33 does it mean it drops dbus too? Nov 14 13:24:41 (but you need to enable access to these counters from userland) Nov 14 13:25:13 ah Nov 14 13:25:17 yes Nov 14 13:26:17 (http://stackoverflow.com/questions/3247373/how-to-measure-program-execution-time-in-arm-cortex-a8-processor ) Nov 14 14:06:57 * bradfa just placed order for beaglebones and got a free #raspberrypi! :) Nov 14 14:10:24 bradfa: omg where? how many bones do I have to order and scrap to get a pi? Nov 14 14:11:10 adafruit Nov 14 14:12:06 $350 orders get a free model b Nov 14 14:12:11 wow Nov 14 14:12:11 bradfa, you'll be banished from the kingdom Nov 14 14:12:37 at least he does not get to wear a tutu Nov 14 14:12:42 stop that! Nov 14 14:12:46 mdp, I wasn't already? :) Nov 14 14:13:25 bradfa, I guess we both live in geographical exile anyway ;) Nov 14 14:13:27 mdp, I joined #raspberrypi yesterday, haven't gotten banned yet, mru said I was failed human Nov 14 14:13:32 yes Nov 14 14:22:46 ha, where is NTD3 when one needs him: https://lkml.org/lkml/2012/11/13/531 Nov 14 14:28:24 alan is so wise Nov 14 14:28:46 I immediately thought of the skeletonfb.c driver when I saw this thread Nov 14 14:29:15 a driver that is full of stale and deprecated interfaces since it's not a real driver ;) Nov 14 14:30:00 aka a repository of bad habits ready for cargo-culting Nov 14 14:31:25 hehe Nov 14 14:32:13 mru: yes, instead of checking a template driver in, the kernel should write a template driver on demand Nov 14 14:32:32 A good, maintained example would be great Nov 14 14:32:40 since there are so many crap ones out there Nov 14 14:32:42 node.js could help here Nov 14 14:32:43 Crofton|work: pick 2 Nov 14 14:32:45 well volunteered Crofton|work :-D Nov 14 14:33:05 anything I do is based on the crappy templates Nov 14 14:33:18 write a template writer Nov 14 14:33:26 Crofton|work: the unwritten thing here is that even amongst real drivers, you can easily pick ones with deprecated interfaces or uses in them too Nov 14 14:33:26 python-cheetah Nov 14 14:33:38 yes, that is exactly yhe issue Nov 14 14:33:55 just base you driver off /dev/bitbang Nov 14 14:33:57 but, unfortunately, this example things ends up being worse in practice Nov 14 14:33:57 +r Nov 14 14:34:34 it would be interesting to see it with *everything* fixed up Nov 14 14:34:48 use of regmap, etc. Nov 14 14:35:22 he should publish it in a dead-tree book Nov 14 14:35:38 and DT bindings Nov 14 14:35:47 templatebus Nov 14 14:35:58 I don't like that name! Nov 14 14:35:58 I mean, what if I need to stack 2 templates? Nov 14 14:36:05 tutubus? Nov 14 14:36:11 there we go Nov 14 14:38:07 Crofton|work: it would be great with a seriously committed maintainer owning it Nov 14 14:38:39 mdp, +1 Nov 14 14:38:53 that's never going to happen Nov 14 14:39:00 sounds like a horribly thankless task ;) Nov 14 14:39:03 mru: certain guy just volunteered Nov 14 15:00:22 SSL peer was unable to negotiate an acceptable set of security parameters. Nov 14 15:09:13 av500, you volunteered? thx Nov 14 15:09:24 fame and no fortune await you Nov 14 15:10:07 did av500 sign up for the front lines? Nov 14 15:11:05 lol Nov 14 15:12:40 mdp: I have a coworker that always says "it would be great" Nov 14 15:13:17 av500, we have people here that use, "there ought to be a law!" Nov 14 15:13:26 as if we don't have enough already ;) Nov 14 15:13:57 dont you all want less laws? Nov 14 15:14:23 * mranostay waves guns and beer bottles in the air Nov 14 15:14:30 mdp: so, the federal government should maintain that driver Nov 14 15:14:37 av500, only ~1% of the popular vote went for less laws ;) Nov 14 15:14:57 we need obamakernel Nov 14 15:15:35 mranostay, we need to dry you out until elc-us Nov 14 17:19:39 hi all, I wanna know where can I get the google SketchUp model for the beaglebone? Nov 14 17:20:17 Hi guys, trying to figure out how to troubleshoot beaglebone not booting up with a custom image. I built a custom kernel and systemd image using setup-scripts/bitbake. partitioned SD card into FAT boot and ext3 angstrom partitions and put MLO / u-boot.img on boot/ and uncompressed the deploy bz2 to the other partition. Plugged into bone and don't get led heartbeat. Since I can't connect I'm kinda stuck on where to start debuggi Nov 14 17:20:17 ng? Nov 14 17:26:20 no output on serial? Nov 14 17:41:45 koen, you mean serial over USB? I guess I hadn't thought to try because I figured it wouldn't working. Nov 14 17:45:15 yeah, not able to connect to /dev/ttyUSB1 or 0 with the thing powered by 5v and connected to USB Nov 14 17:59:47 looks like requests for free homework on the mail list Nov 14 18:14:18 how can I generate a PWM signal on a beagleboard-xm? Nov 14 18:14:59 angs, by using the available PWM generators of course Nov 14 18:15:40 panto: could you refer me a document to read how to use it? Nov 14 18:16:14 which kernel version? Nov 14 18:16:18 I generate pwm by continuously setting 1 and 0 to a GPIO, however it generates a bad PWM and cause CPU load Nov 14 18:16:20 the 3.2 TI tree, or mainline? Nov 14 18:16:27 * panto facepalms Nov 14 18:16:37 Linux ubuntu 3.2.0-23-omap #36-Ubuntu Tue Apr 10 20:24:21 UTC 2012 armv7l armv7l armv7l GNU/Linux Nov 14 18:16:42 angs, obviously, that's why the nice hardware blocks can do it much better Nov 14 18:16:59 http://beaglebone.cameon.net/home/using-pwm-outputs Nov 14 18:17:04 first link on google Nov 14 18:17:05 panto: thank you Nov 14 18:17:28 mdp, another support ticket closed Nov 14 18:17:57 I use beagleboard Nov 14 18:18:03 xm Nov 14 18:18:14 err, damn Nov 14 18:18:36 similar method, but I can't recall if the PWMs are brought out to the connectors Nov 14 18:18:45 I cannot see pwm under /sys/class Nov 14 18:19:40 http://elinux.org/BeagleBoardPWM Nov 14 18:20:12 of course this method is totally evil Nov 14 18:20:42 * panto makes a note about checking pwms on the beagleboard Nov 14 18:21:23 I read it before but I will give a try to understand it again Nov 14 18:30:33 * mdp updates panto's support ticket metrics on the big board Nov 14 18:36:08 Maybe I should back up a step, if I built a kernel using the setup-scripts beaglebone machine config and just added some modules then built systemd image should I expect the LED heartbeat? Or is that something special in the cloud9-image setup? Nov 14 18:38:03 panto, not evil, omap3 pwms are intimately tied to dm timers Nov 14 18:38:23 ehrpwm has its own quirkiness Nov 14 18:40:33 I'll ask the designers to add a 3rd IP to the next chips as soon as everybody figures out these two Nov 14 18:40:41 gotta keep on your toes ;) Nov 14 19:01:01 dfarmer, if you don't even load an SD card in the bone and connect to the USB-serial, you should see "CCCCCC" printed which is the boot rom looking for a serial load. Nov 14 19:01:31 dfarmer, if you have a valid u-boot and SPL (the MLO file) on your FAT, can you get a u-boot prompt over the usb-serial? Nov 14 19:04:27 dfarmer, if not, you have problems to solve before you figure out your heartbeat LED problem :) Nov 14 19:05:15 bradfa, Ah ok. That helps. Connected to ttyUSB1 and reset and I see "itch \n OMAP" Nov 14 19:06:12 bradfa, Yeah, I can plug in the original SD card and it boots right up and everything's fine. Just something screwy with how I made my SD card I'm sure. Nov 14 19:09:16 The thing that I'm sort of stuck on is that I compared the stock SD card with the one I made (e.g, partitions) and they look the same. Obviously I have less files on my image, but I've got MLO and u-boot.bin in the boot partition and the rest of the system on Angstrom partition Nov 14 19:09:41 I did just notice there is also a modules.bz2 zip file that the tutorial I was looking at didn't mention though Nov 14 19:09:45 dfarmer, if you hold the reset button on the bone down while you launch your term, then release, you'll see the whole boot serial output Nov 14 19:10:31 and while you do this, say a curse for the people who put the evil ftdi thing on there Nov 14 19:10:35 "ubootswitch \n OMAP" Nov 14 19:10:37 dfarmer, or you can use a different UART, but then you won't see the "CCCCC" bootrom load output Nov 14 19:10:38 hehehe Nov 14 19:10:46 mru, oh wait... Nov 14 19:11:19 dfarmer, in u-boot mainline do a "make am335x_evm_uart1_config" and use a real serial port ;) Nov 14 19:11:44 * bradfa curses FTDI and the people who like them! Nov 14 19:12:43 dfarmer, follow mdp's guide -> http://engineersofthecorn.blogspot.com/2012/10/homebrew-beaglebone-rs-232-cape.html Nov 14 19:12:56 just don't use those patches, they're already in mainline u-boot Nov 14 19:13:33 mdp, someone wrote a blog about rs232 in u-boot on the bone and their exact steps don't work for me any more, git complains! Nov 14 19:14:42 bradfa, haven't I told you that there are tons of blog entries on doing stuff on beaglebone...98% are anywhere from mostly incorrect to completely incorrect Nov 14 19:15:04 Ok, will have to try that later when I can get to electronics store. Nov 14 19:15:12 mdp, can e2e help here? Nov 14 19:15:23 what about dt? Nov 14 19:15:39 Hooked up ac adapter and reset and it says "No daughter" Nov 14 19:15:42 * bradfa doesn't yet have hair added to his chest due to device tree... is sad Nov 14 19:15:48 mdp: just put it in but don't put it in the docs. once people figure out the previous ones, bang, instant new riddle announcement! Nov 14 19:16:13 dfarmer, you're going to have to be more descriptive than that... pastebin? Nov 14 19:16:31 the overall scheme seems to keep the questions coming Nov 14 19:16:59 djlewis, isn't that how we make money? Lose money on each unit but make it up in question volume? Nov 14 19:18:09 Sure, http://pastebin.com/zi1n6BQz Nov 14 19:18:10 bradfa, I noticed I have 587 unread notifications in my e2e mailbox..from our linux/sitara ph0rums..I'm really on top of e2e Nov 14 19:18:28 but that's literally all I've got. Nov 14 19:19:27 mdp, you have an e2e mailbox! I'm so jealous Nov 14 19:19:48 mdp, I did find a helpful e2e thread though, I'm afraid the world might end. Can I take that up with your management? Nov 14 19:19:58 bradfa, it used to be required reading/response for work Nov 14 19:20:09 dfarmer, you have problems Nov 14 19:20:44 dfarmer, boot should look something like https://gist.github.com/3552559 Nov 14 19:21:08 but probably with different u-boot git commit count and sha and compile date Nov 14 19:21:08 bradfa, there's quite a bit that is useful on there...the trick involves searching the proverbial haystack Nov 14 19:21:29 mdp, some quality pointers I found helping with tps65217 power button development Nov 14 19:21:40 of which I owe patch sometime RSN Nov 14 19:21:57 bradfa, yeah, I just booted up my good card and got a "normal" output. Nov 14 19:22:09 like that am335x errata I forgot to read where the nnmi interrupt pin polarity is backwards in the silicon Nov 14 19:22:30 dfarmer, can you just use that card's u-boot and put your kenrel on there Nov 14 19:22:32 ? Nov 14 19:22:36 s/kenrel/kernel/ Nov 14 19:22:47 That's what I'm considering, yeah. Nov 14 19:22:58 dfarmer, do that Nov 14 19:23:11 dfarmer, then you'll know if your kernel is the problem of if your u-boot is Nov 14 19:23:23 sounds like your u-boot could use some love Nov 14 19:24:52 yeah, I'm sure I couldn't do it manually but that's the unfortunate thing with bitbake is the black magic aspect. I have no idea what it did, it just spit out an MLO and u-boot.img that apparently don't work :/ Nov 14 19:30:18 dfarmer, just build u-boot mainline, use the 'am335x_evm_config' Nov 14 19:30:31 it'll build the MLO and u-boot.img for you Nov 14 19:30:40 any decent arm cross compiler will work Nov 14 19:30:43 Ok Nov 14 19:31:18 make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- {distclean,am335x_evm_config,all} Nov 14 19:31:30 replace the arm-linux-gnueabi- with what ever prefix your cross compiler uses Nov 14 19:31:31 bradfa, I really appreciate the help. Nov 14 19:31:38 dfarmer, any time Nov 14 19:57:52 speaking of bitbake magic, what is the target to get it to spit out new images in deploy/images/beaglebone? I reconfigured my kernel with menuconfig and rebuilt everything using -f and compile/deploy but it doesn't make new image or module archive in deploy Nov 14 20:00:15 kernel-modules is the target for modules and look for recipes in meta-angstrom/recipes-images for images Nov 14 20:18:14 Woohoo, rebuilding u-boot from mainline did the trick. Nov 14 20:24:49 amazing how small the core patchset for arm hibernate is Nov 14 20:27:01 W1N9Zr0, when I try "bb virtual/kernel -c kernel-models -f" I get "task do_kernel-models" doesn't exist for target virtual/kernel" likewise with targets "modules" and "console-image" "image" etc? including the stuff I found in meta-angstrom/recipes-images Nov 14 20:28:16 I can only compile the kernel and build images once, otherwise I have to change the version number in linux-ti33x-psp_3.2bb to get bitbake to do it again Nov 14 20:28:52 I change MACHINE_KERNEL_PR_append but there has to be a way to get bitbake to generate the images and modules again Nov 14 20:31:22 excuse me for stating the obvious: have you tried bitbaking the images again? Nov 14 20:32:22 what do you mean? Nov 14 20:32:58 it's just "bitbake THING" not virtual/kernel -c THING, where THING is sytemd-image or kernel-modules in your case Nov 14 20:33:42 OK but it rebuilt the modules during the deploy target, it just didn't create new image and module archive files in the deploy folder like it did the first time Nov 14 20:34:44 W1N9Zr0, sorry, "bitbake kernel-modules" gives me "Nothing PROVIDES 'kernel-modules'" Nov 14 20:36:34 am I calling bitbake from the wrong folder or something? I have no trouble building the kernel/modules and installing them but when I make a change I am unable to get bitbake to create the images again without changing the version number in the main .bb Nov 14 20:38:31 try -f Nov 14 20:39:05 sorry kernel-modules is what it's called when you include it in the image, i don't remember how to bitbake it. does not not give you a modules tar when you build kernel? Nov 14 20:39:38 *does it not Nov 14 20:40:42 it gives me the modules tar the first time but subsequent times it won't, and it does actually recompile the kernel and modules when I use -f Nov 14 20:43:38 what are you changing that you need to recompile it? Nov 14 20:46:19 preemptive scheduling, recompiling with preempt worked except the etherenet coudln't send/receive anything so I had been using the RT_PREEMPT patch under which ethernet works but my wifi driver crashes Nov 14 20:46:43 are you chaging it through the recipe or are you editing stuff in the work directory? Nov 14 20:47:14 my application is an audio recorder and without preemptive scheduling I've been getting buffer overruns source code is here https://github.com/drue/King-James Nov 14 20:47:56 I added the RT patch as another patch in the stack but I did do a manual menuconfig since I haven't looked yet to see where config changes Nov 14 20:48:20 i.e. I did bitbake menuconfig and changed the scheduling config Nov 14 20:50:56 I'm trying to do it the right way please correct me if I'm not Nov 14 20:51:08 if you do the config through the recipe as well, and increment the recipe version every change bitbake should be rebuilding it for you correctly Nov 14 20:53:16 it does but then it appears to redo the entire kernel including patches, I was hoping there would be better way but I can't seem to get it to create the build products again Nov 14 20:54:41 I realize that changing the kernel config necessitates a recompile of everything, but if I'm hacking the wireless driver then rebuilding everything for every change is going to be tedious Nov 14 20:55:22 you've got a fast multi-core cpu to compile on, right? :) Nov 14 20:56:02 W1N9Zr0, yes (and I was using an extra-large instance on EC2 for that before I got a modern computer) but it still takes a while? Nov 14 20:56:36 did you configure PARALLEL_MAKE and BB_NUMBER_THREADS in conf/local.conf ? Nov 14 20:57:10 also, I have this driver that lives outside the linux tree and I'm not sure how to build it using this setup Nov 14 20:57:57 it's a ralink driver Nov 14 21:13:17 burris in oe we have serval recipes for out of tree modules Nov 14 21:14:15 woglinde, any info or pointers you could provide would be very helpful as I'm really a newbie with oe/bb Nov 14 21:15:02 hm try git grep inherit | grep kernel Nov 14 21:15:07 in oe core Nov 14 21:17:43 the driver comes with it's own crazy hw manufacturer makefiles, it might be easiest just to point it at the work directory Nov 14 21:18:27 W1N9Zr0, building everything takes 20 minutes on my box Nov 14 21:19:07 you have a fast machine if you can build angström in just 20min Nov 14 21:19:25 I don't think it is angstrom, just the kernel and modules and hey no build products agai Nov 14 21:19:30 :-( Nov 14 21:19:45 last time i build a new kernel, it took 3h Nov 14 21:19:52 date Nov 14 21:20:02 today, right now Nov 14 21:20:24 sorry wrong window :-) did find the build products Nov 14 21:21:27 yeah it used to take me hours and hours on my 4 year old laptop so I was using an ec2 extra large instance to compile but recently I got a new box Nov 14 21:21:49 o_0 Nov 14 21:22:21 .o0(enough money to go into the cloud, not enough money for a good computer) Nov 14 21:22:50 are you using an SSD to get it down to 20 minutes? or is that on spinning rust? Nov 14 21:22:51 it's $0.75 to save you a few hours, your time must not be worth much Nov 14 21:23:34 W1N9Zr0, spinning rust but its a raidz2 zfs volume with L2ARC and ZIL on an ssd Nov 14 21:24:16 dual core i3, the cheapest/lowest-power one I could get six months ago Nov 14 21:26:42 that's zfs-on-linux btw, so probably very suboptimal fs performance Nov 14 21:33:32 0.o Nov 14 21:34:15 KotH: \o/ Nov 14 21:39:01 mranostay: ? Nov 14 21:41:29 http://finance.yahoo.com/news/ti-reduce-costs-wireless-business-213000647.html Nov 14 21:42:16 http://www.streetinsider.com/Corporate+News/TI+%28TXN%29+to+Cut+1,700+Jobs,+to+Focus+More+on+Wireless+Business/7879446.html Nov 14 21:43:57 cost reduction has a softer spin to it Nov 14 21:44:08 anyone we know? Nov 14 21:44:29 mranostay, I see no names in the PR :) Nov 14 21:45:24 cute Nov 14 21:46:41 <_av500_> ha, I once had lunch with Delagi :))) Nov 14 21:46:47 <_av500_> my BFF Nov 14 21:47:29 BFF with ponies Nov 14 21:47:34 <_av500_> yep Nov 14 21:47:42 <_av500_> i should have snapped pics Nov 14 21:48:47 "It was also commented that Android really is the future, even possibly Nov 14 21:48:47 in server, so it behooves us all to ensure that its as easy and open Nov 14 21:48:47 to work with as we can help make it. Android on the server is coming, Nov 14 21:48:47 are we ready?" Nov 14 21:49:03 lol, wut? Nov 14 21:49:16 wait, is android based on threads? Nov 14 21:49:24 coincidentally from pfefferz as he enters the room ;) Nov 14 21:50:22 Russ, it's more like yarn Nov 14 21:51:17 uh oh jkridner-web Nov 14 21:51:45 hide the pasm patches. ;-) Nov 14 21:51:56 wait for it...jkridner-node.js-rockstar is next to enter Nov 14 21:52:35 node.js is kick-ass rock star technology! Nov 14 21:52:45 did you say lisp? Nov 14 21:52:57 with a lisp? Nov 14 21:53:15 that isn't PC Nov 14 21:54:00 elisp Nov 14 21:54:08 lisp was my 4th programming language. BASIC, Fortran, assembler and then Lisp Nov 14 21:54:15 then C. Nov 14 21:54:39 Everything beyond C is a compromise. Nov 14 21:54:45 JavaScript just happens to be a good compromise. Nov 14 21:54:49 :-) Nov 14 21:55:20 modula-2 and ada83/95 will forever be superior to everything else Nov 14 21:55:25 I'm just forced to work in C Nov 14 21:55:55 * Russ tries to suppress memories of ada95 on arm Nov 14 21:56:03 mdp: so you are a wirthian? Nov 14 21:57:19 Russ, we have so much in common from that industry ;) Nov 14 21:57:29 jkridner-web: good? :) Nov 14 21:57:32 KotH: not strictly, the theory is interesting Nov 14 21:57:42 its not WSMS, is it? Nov 14 21:58:18 mdp: wirth had nice ideas, but his theories went into one way, when the rest of the world went another Nov 14 21:58:44 mdp: these days people are glad he is retired, and wait for the day the last of his students leave the eth Nov 14 22:01:02 <_av500_> KotH: lol Nov 14 22:01:36 hehe Nov 14 22:04:26 is it just me, or are all the links on my.ti.com broken? 'The requested URL /cgi-bin/redirect.pl was not found on this server.' Nov 14 22:09:17 mdp, huh? Nov 14 22:11:29 pfefferz, just quoting from your g+ posting here Nov 14 22:12:15 mdp, :) Nov 14 22:13:02 pfefferz: btw, remind me to tell you what happened at ELCE with regards to "the good, the bad, and the ugly" Nov 14 22:13:15 prpplague, aye Nov 14 22:26:22 here http://41j.com/blog/2011/09/beagleboard-gpio-input-driverless/ it says pinconf[0x2168/4] = 0x011C011C; configures pin as input. can anyone tell me how to find out to use "0x011C011C" value? or if I want to use it as output what value do I need to assign? Nov 14 22:33:02 angs : read processor documentation over and over until you understand every bit of the register Nov 14 22:33:55 yikes Nov 14 22:34:31 CareBear: it is 3674 pages, could you tell me what chapter/subchapter explains it? Nov 14 22:34:42 g'night Nov 14 22:36:28 I've not tried memmapping, but I've heard lots of people say to avoid it when possible with gpio Nov 14 22:37:17 thurbad: mmap() and bit-bang. Device tree can help, but I can't give exact steps. Nov 14 23:03:27 uhm.. I'm getting a lot of "hub 1-0:1.0: over-current condition on port 1" on my beaglebone.. I guess I have damaged something... Nov 14 23:13:18 dwery, I got stuff like that after someone else put 12V into a bunch of connected stuff that wanted 5V Nov 14 23:13:58 dwery: board still worked, but gave that same error iirc. Nov 14 23:14:15 oh dear, 'there are no "trained" boot-loader-writing monkeys.' Nov 14 23:14:18 it seems USB is not working Nov 14 23:14:35 Russ: don't show that to tartarus Nov 14 23:14:58 dwery: now that I think about it, I think only some ports worked. It was an xM Nov 14 23:15:20 I should have gotten it from them. I bet it went in the trash :( still would have been good for something. Nov 14 23:15:27 or partly working.. getting strange results Nov 14 23:17:48 probably has something to do with usb voltage regulation Nov 14 23:18:13 dwery: so you let _some_ of the smoke out Nov 14 23:18:35 so does x86 u-boot actually work? Nov 14 23:18:36 I might have briefly short circuited a couple of pins Nov 14 23:18:55 dwery: will neither confirm nor deny? :) Nov 14 23:19:09 nope :D Nov 14 23:21:22 odd, I've heard of people short circuiting pins on the beagleboard with solder blobs and everything working after the solder was removed Nov 14 23:21:41 thurbad: I'm sure it depends on the pins Nov 14 23:21:49 mranostay: you should try it and get back to us Nov 14 23:21:52 probably, voltages too Nov 14 23:22:30 I'm sure it did happen in the past too Nov 14 23:22:41 mranostay: I like syslinux/extlinux on PC. Nice and easy. The world is so much simpler when you have a BIOS. Nov 14 23:22:44 it probably depends on which pins you get :D Nov 14 23:24:54 i wouldn't call it simpler Nov 14 23:25:23 mranostay: it's simpler for everything _after_ the BIOS runs Nov 14 23:25:33 code that comes with the board that does all the hard work for you. Nov 15 00:23:44 mranostay : u-boot has been used as coreboot payload on x86 Nov 15 00:24:18 mranostay : celeron chromebooks do ship with coreboot+u-boot Nov 15 00:24:24 mranostay : so that would be a yes :) Nov 15 02:13:19 jkridner1: got the nxp driver to build with mainline Nov 15 02:13:36 have to apply lcdc patches now, few conflicts with that Nov 15 02:14:01 then will add DT bindings and boot through DT **** ENDING LOGGING AT Thu Nov 15 02:59:58 2012