**** BEGIN LOGGING AT Tue May 08 03:00:02 2012
May 08 04:41:59 <corbomite> what servers does everyone use other than freenode?
May 08 07:15:10 <LetoThe2nd> attention, german: http://www.pro-linux.de/news/1/18342/magpi-monatliches-magazin-rund-um-den-raspberry-pi.html
May 08 07:15:13 <LetoThe2nd> *facepalm*
May 08 07:41:13 <koen> LetoThe2nd: *facepalm* indeed
May 08 07:41:13 <koen> monthly magazine
May 08 07:44:09 <LetoThe2nd> koen: you need to keep the hype alive.
May 08 07:48:41 <aholler> av500 had posted that already
May 08 07:50:46 <aholler> It's like those magazines offered at the doctor, meant to be read while you are waiting ;)
May 08 07:52:29 * aholler is awaiting an ivy-bridge magazine ;)
May 08 07:52:57 <tomba> what's the relationship between omap3xxx and dm3xxx SoCs? dm3 is a davinci soc, and there's arch/arm/mach-davinci, but I also see arch/arm/mach-omap2/board-omap3beagle.c mentioning BB-xM, which has dm3... what am I missing?
May 08 07:53:52 <aholler> confusion is the relation
May 08 07:56:07 <aholler> every department of ti wants it's own soc
May 08 07:56:19 <aholler> and every manager a name
May 08 07:57:15 <aholler> so you can use the part names to build an organizational structure of ti
May 08 07:57:38 <av500> LetoThe2nd: the real fail is that it is in Flash......
May 08 07:58:45 <tomba> aholler: so I take it your answer is: "I don't know" =)
May 08 07:59:01 <aholler> tomba: no, omap is from the mobile department
May 08 07:59:11 <LetoThe2nd> av500: i resisted and did not click it.
May 08 07:59:19 <aholler> dm is a catalog part, if I'm correct
May 08 07:59:51 <koen> tomba: omap36xx == dm37xx
May 08 07:59:58 <tomba> well, I'm more interested in the techical side, than who sells the soc =)
May 08 08:00:15 <tomba> koen: okay, so dm37xx is not mach-davinci, but mach-omap2?
May 08 08:00:27 <tomba> when talking about kernel
May 08 08:00:39 <koen> tomba: omap34xx == omap35xx
May 08 08:00:39 <koen> tomba: 34xx,36xx are wbu parts, 35xx,37xx are catalog parts
May 08 08:00:39 <koen> tomba: correct
May 08 08:01:05 <koen> tomba: take a 3630 and print "dm3730" on top of it, noone will notice the difference
May 08 08:01:35 <tomba> somebody at TI should make a web page that explains what TI's SoCs really mean ;)
May 08 08:01:38 <koen> tomba: but note that am35xx is not the same as omap35xx, it rips out the dsp and replaces sdrc with an emif
May 08 08:02:05 <koen> tomba: that would be an internal only webpage, since the partnumbers are used for segmentation :)
May 08 08:03:17 <tomba> koen: luckily I'm positioned in Finland, if I was located at the same place as the HW/marketing/whoever guys, I'd have a big urge to give them a good slap on the back of the head
May 08 08:03:24 <koen> tomba: oh, and am37xx is dm37xx without dsp, so the the same as 36xx without dsp
May 08 08:04:20 <koen> tomba: true story: 'omap' was perceived as difficult due to all the dsp nonsense, so catalog renamed their chips. A while later customers said the 'new' chips were a lot easier to use
May 08 08:05:13 <av500> tomba:
May 08 08:05:16 <av500> [14:56:33] <av500> I think TI needs one more website
May 08 08:05:17 <av500> [14:56:42] <av500> timarketingnames.com
May 08 08:05:58 <av500> koen: and there are 35xx and 36xx without DSP
May 08 08:06:02 <av500> 3505
May 08 08:06:05 <av500> 3611
May 08 08:06:07 <av500> etc..
May 08 08:06:17 <koen> right
May 08 08:06:29 <av500> and 3611 is even no lpddr iirc
May 08 08:06:30 <koen> I keep forgetting the exact numbers
May 08 08:06:34 <koen> so I say "without dsp" :)
May 08 08:09:26 <koen> wtf?
May 08 08:09:34 <koen> only between 8pm and 6am?
May 08 08:11:05 <av500> koen: yep
May 08 08:11:11 <av500> wtf indeed
May 08 08:11:25 <av500> koen: but actaully, that is not the worst part
May 08 08:11:33 <av500> the worst is that thay have to forget the content
May 08 08:11:46 <av500> because private TV claims it will kill them otherwise
May 08 08:12:03 <av500> so we pay this with TV fees and then others that we do not pay claim we cannot see it
May 08 08:12:28 <av500> and yes, private TV is a gem that needs preservation
May 08 08:13:21 * koen grew up with schimanski
May 08 08:13:36 * av500 too
May 08 08:13:53 <koen> I actually learned most of my german for tatort and der alte
May 08 08:13:57 <koen> and a bit of derrick
May 08 08:14:33 <koen> s/for/from/
May 08 08:15:50 <av500> koen: but in the end, you dont need to download Tatort at all, with all the 3rd channels, there is a (older) Tatort on every night
May 08 08:19:10 <dm8tbr> der alte is very popular in finland too i noticed
May 08 08:19:50 <koen> av500: yeah, we get a dose of tatort on the dutch public channels as well
May 08 08:20:10 <av500> dm8tbr: well, since german crime TV is being taken over by scandinian crime.....
May 08 08:20:35 <av500> you now get more soko göteborg than 5113
May 08 08:20:46 <koen> my father-in-law is addicted to scandinavian crime series
May 08 08:21:11 <dm8tbr> there even was a "finnish" tatort ;-)
May 08 08:21:17 <av500> yep
May 08 08:21:29 <koen> forbrydelsen, bron, etc
May 08 09:08:03 <lpi> Hey ya! Is it possible to compile and debug with CodeSourcery on a Beagleboard xM under Windows?
May 08 09:10:14 <aholler> windows doesn't run on a beagleboard
May 08 09:19:13 <lpi> aholler: Sorry, I mean CodeSourcery under windows, of course!
May 08 09:20:39 <koen> lpi: is what you're running on the beagle built with codesourcery under windows?
May 08 09:20:57 <koen> lpi: the answer to that is the same answer as the one to your question
May 08 09:27:08 <lpi> koen: codesourcery is installed on a windows host machine. I want to cross compile and automatically upload binaries to the BB, and debug it
May 08 09:27:37 <koen> lpi: that's not an answer to my question
May 08 09:27:53 <koen> lpi: what OS or distro is running on your beagle?
May 08 09:28:06 <koen> is that built with codesourcery under windows?
May 08 09:28:28 * koen wonders how many different way he'll need to ask that question before it gets understood
May 08 09:31:22 <lpi> koen: sorry, BBxM is tunnign Angstrom
May 08 09:32:04 <lpi> and i'd installed codesourcery for windows on my host PC
May 08 09:54:33 <koen> lpi: so answer this: is angstrom built with codesourcery under windows?
May 08 10:03:48 <av500> koen: remind me, is gst-ti abandonware or not?
May 08 10:04:08 <mianos> my beaglebone V4 boots angstrom but can't start up ubuntu :(
May 08 10:05:45 <mianos> 11.10 or 12.04 using the guide on http://elinux.org/BeagleBoardUbuntu
May 08 10:07:16 * av500 looks up "can't start up" in Wikipedia
May 08 10:10:51 <koen> av500: it is
May 08 10:11:16 <mianos> hmm, ok, loaders runs, boots the kernel, mounts the filesystem and then for 12.05 sort of stops in init
May 08 10:11:32 <mianos> and 11.10 getty stops
May 08 10:11:36 <koen> av500: dspbridge is enabled in the 3.2 kernel for beagleboard, only the userspace is missing
May 08 10:11:43 <mianos> INFO: task getty:596 blocked for more than 120 seconds.
May 08 10:11:44 <mianos> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
May 08 10:11:46 <koen> av500: so people can choose which abandonware to run
May 08 10:44:58 <nickjohnson> Okay, PCB order for a prototype stepper shield for the beaglebone ordered :)
May 08 10:47:04 <nickjohnson> By the way, is anyone aware of a Python library for the beaglebone that doesn't try to pretend to be an Arduino?
May 08 10:47:37 <nickjohnson> Having easy ways to read and write from GPIO pins is great, but I'm not looking for something to provide me with 'setup' and 'main' hooks so I can pretend I'm writing Arduino code in Python.
May 08 10:53:53 <koen> nickjohnson: if you're serious about those steppers, python will be made of fail for that
May 08 10:54:07 <koen> nickjohnson: you want a kernel module for the timing you need
May 08 10:54:36 <koen> anything interpreted will fail
May 08 10:58:58 <nickjohnson> koen: I don't agree. Do you have any idea just how many clock cycles per step we have on something like the beaglebone?
May 08 10:59:35 <nickjohnson> But in case someone does want high rate constant-speed stepping, the step pins are both on PWM enabled outputs
May 08 11:01:44 <koen> do you have any idea how bad the sysfs gpio interface is?
May 08 11:01:44 <koen> this is not a microcontroller
May 08 11:38:14 <aholler> explaining the chaos theory by stepping from userspace ;)
May 08 11:39:15 <LetoThe2nd> aholler: explaining the user theory by space from chaotic stepping?
May 08 11:39:34 <av500> Nuke them from orbit!
May 08 11:40:01 <LetoThe2nd> <attention, rating r>: av500: nudes into orbit?!?
May 08 11:40:40 <mru> that could get chaotic
May 08 11:42:06 * LetoThe2nd is ATM reviewing chaotic user code.
May 08 11:42:49 <mru> LetoThe2nd: strip it naked
May 08 11:42:54 <mru> and spank it
May 08 11:42:58 <mru> the user, that is
May 08 11:43:09 <LetoThe2nd> mru: not a pretty user.
May 08 11:43:26 <mru> nor pretty code
May 08 11:43:38 <LetoThe2nd> indeed
May 08 11:44:29 <av500> give it the hose
May 08 11:44:51 <LetoThe2nd> the metal singer in my headphones telling me something about asgard and valkyries also is not particularly helpful. damn.
May 08 11:46:13 <av500> if programmers are converting coffee into code, maybe LetoThe2nd is converting metal into rage code....
May 08 11:47:46 <nickjohnson> koen: Who said I was planning to use sysfs? That's a horrible hack if you're writing code.
May 08 11:47:53 <nickjohnson> Much easier and better to use /dev/mem
May 08 11:48:16 <LetoThe2nd> ok. and i thought my user code is disgusting.
May 08 11:48:20 * av500 googles to see if the definition of "hack" changed
May 08 11:48:39 <av500> ah no, its still minced meat..
May 08 11:48:41 <nickjohnson> LetoThe2nd: That's the point of having a library that abstracts it away
May 08 11:49:23 <LetoThe2nd> someone poking /dev/mem from python and trying to keep timing contraints in the menawhile should better not qualify other ways as "hack". but thats my $.02
May 08 11:49:36 <nickjohnson> Besides, writing to simulated files with the filesystem API is a _much_ bigger hack than manipulating the registers created for the purpose
May 08 11:49:53 <nickjohnson> LetoThe2nd: What timing constraints? I'm just flipping GPIO bits and maybe mux settings.
May 08 11:50:14 <nickjohnson> Or are you talking about stepping intervals? In which case, see my earlier comment RE code execution speed
May 08 11:50:39 <LetoThe2nd> nickjohnson: nevermind me. i have not said anything. i am not even here. *jedihandmove*
May 08 11:51:23 <nickjohnson> At 2k steps per second, I have 360k clock cycles per step to work with. I think 'even' an interpreted language can manage that
May 08 11:51:29 <av500> in this channel we welcome all kinds of hackery
May 08 11:51:54 <nickjohnson> Admittedly context switching is a bigger concern, albeit one that's present regardless of language
May 08 11:52:10 <av500> SCHED_FIFO will prevent that
May 08 11:52:23 <nickjohnson> It'll certainly help, but linux isn't an RTOS :)
May 08 11:52:46 <av500> nickjohnson: as you said, many clocks per step..
May 08 11:53:00 <nickjohnson> yup
May 08 11:53:12 <mru> linux does not promise it won't go off and do something for a million cycles
May 08 11:53:32 * nickjohnson checks what the typical timeslice size is in linux
May 08 11:53:38 <mru> that's irrelevant
May 08 11:53:44 <mru> it's tickless anyway
May 08 11:53:52 <nickjohnson> mru: And if I were, say, sequencing an x-ray emitter for neutron therapy, that would be relevant
May 08 11:54:04 <nickjohnson> But I'm not, so I'm prepared for the slim possibility of that happening
May 08 11:54:06 <mru> and even if it weren't, there's no guarantee that you'll evet a slice at all
May 08 11:54:36 <aholler> python is a pretty accurate programming language
May 08 11:54:46 <av500> 1+1 = 2=
May 08 11:54:47 <mru> if I were building some dangerous machine, I'd use nothing more more complex than an arm9
May 08 11:54:47 <av500> 1+1 = 2?
May 08 11:54:56 <av500> pah, bunch of 74xx...
May 08 11:55:06 <koen> nickjohnson: /dev/mem *is* a hack
May 08 11:55:12 <koen> nickjohnson: nothing in userspace should be using it
May 08 11:55:15 <LetoThe2nd> av500: only occaionally. sometimes it is also 1.9999994
May 08 11:55:21 <av500> thats fine
May 08 11:55:50 <nickjohnson> koen: gpio via sysfs is also a hack. What would you suggest?
May 08 11:55:59 <koen> lastlog kernel
May 08 11:56:06 <koen> 12:54 < koen> nickjohnson: you want a kernel module for the timing you need
May 08 11:56:07 <mru> write a kernel driver, perhaps
May 08 11:56:08 <koen> there
May 08 11:56:19 <nickjohnson> Ever heard the phrase 'the perfect is the enemy of the good'?
May 08 11:56:32 <koen> ok, you're the expert on all this
May 08 11:56:35 <mru> sysfs is nice for casually playing around
May 08 11:56:36 <nickjohnson> Developing a kernel module any time you want to poke at GPIO pins isn't a practical solution for most people, myself included
May 08 11:56:39 <koen> we don't know what we are talking about
May 08 11:56:48 <mru> we're just guessing
May 08 11:56:49 * koen joins LetoThe2nd in *jedihandwave*
May 08 11:57:09 <LetoThe2nd> koen: hey, now we are two. we can join the dark side! :)
May 08 11:57:20 <LetoThe2nd> koen: rumour is they have cookies....
May 08 11:57:23 <nickjohnson> mru: I concur. But there's room for a category between 'casually playing around' and 'write your own kernel modules'
May 08 11:57:26 <koen> LetoThe2nd: cookies!
May 08 11:57:38 <av500> casually write a kernel module...
May 08 11:57:49 <nickjohnson> koen: Sorry, I have something set up to filter out passive-aggressive bullshit like that.
May 08 11:58:17 * LetoThe2nd hands koen a pssive-aggressive dark side cookie.
May 08 11:59:28 <nickjohnson> Besides, isn't there anyone here who thinks writing your very much user-level code as a kernel module is _also_ a hack?
May 08 12:00:28 <mru> sounds like you're sitting on a hack without proper separation of concerns
May 08 12:00:34 <av500> writing a /dev/stepper might not be a hack
May 08 12:00:46 <mru> hey, I've got it! put it all in systemd!
May 08 12:00:52 <mru> how could that possibly go wrong?
May 08 12:00:55 <av500> mru: get in the line!
May 08 12:01:06 <av500> the merge queue is already booked until 2015
May 08 12:01:20 <LetoThe2nd> mru: fork it and call it jedid.
May 08 12:01:24 <jannau> mru: that won't help after systemd is integrated into the kernel
May 08 12:01:34 <av500> jannau: other way round
May 08 12:01:37 <aholler> must be /usr/stepper
May 08 12:01:41 <mru> jannau: hopefully qnx will be open source by then
May 08 12:02:01 <av500> once systemd becomes selfaware we are all doomed anyway
May 08 12:02:20 <LetoThe2nd> hm, now the metal singer proclaims: "the dark times return..."
May 08 12:03:04 * LetoThe2nd runs off for $MEETING
May 08 12:03:34 <nickjohnson> av500: Well, if you're doing constant rate stepping, you don't really need one - you can use PWM
May 08 12:04:03 <nickjohnson> But I'm not stepping at a constant rate, or even at a constantly changing rate, so I don't really see a way to separate concerns without putting my fairly specific code into the module
May 08 12:04:16 <av500> nickjohnson: yeah, it all depends
May 08 12:04:57 <mru> LetoThe2nd: http://www.volkside.com/wp/wp-content/uploads/2009/03/hold-a-meeting.gif
May 08 12:05:01 <nickjohnson> (For context, I'm drawing straight lines in polar coordinates, at a constant linear velocity. If anyone knows of a better way to do that than to calculate which motors to step every timeslice, by all means let me know)
May 08 12:05:48 <av500> nickjohnson: you could give the kernel module a "render list"
May 08 12:05:58 <av500> and it would notify you once it reached the end
May 08 12:06:28 <nickjohnson> av500: That's a possibility, yeah. But it's putting a lot of work into the kernel's hands, that has to be done asynchronously from the process calling it
May 08 12:06:56 <nickjohnson> Which means the kernel has to schedule itself for that. I wonder how that's handled?
May 08 12:07:04 * nickjohnson bets it uses the same scheduler as user-space processes
May 08 12:08:50 <aholler> you don't a stepper to check how accurate your python-sw will be
May 08 12:08:54 <aholler> +need
May 08 12:09:11 <av500> define "accurate"
May 08 12:10:17 <nickjohnson> For that matter, are there other existing GPIO kernel modules that provide a better API than sysfs? Like a device we can mmap with only the relevant fields exposed, or a sysctl type interface?
May 08 12:11:06 <av500> I dont get your aversion to sysfs
May 08 12:11:21 <aholler> accurate = +- 10s ;)
May 08 12:12:05 <nickjohnson> av500: 1) It's a really terrible API for what comes down to bitbashing. A file handle per GPIO pin and a write operation per toggle is an enormous amount of overhead, 2) I haven't benchmarked, but I suspect it's glacially slow.
May 08 12:12:40 <nickjohnson> aholler: I'm not sure I understand what you're saying
May 08 12:12:45 <av500> me neither
May 08 12:13:31 <phh> i think he means, you could set the gpio 10 seconds too early
May 08 12:15:09 * mru would use a uC to do the actual stepping
May 08 12:15:40 <mru> from a 'render list' computed by the AP
May 08 12:15:56 <koen> like the PRU?
May 08 12:15:58 <nickjohnson> mru: Then you have the problem I currently have: Coordinate transforms in software floating point on an AVR are _really_ slow
May 08 12:16:11 <nickjohnson> Admittedly, they're fast enough for the stepping rate I currently need, but only barely
May 08 12:16:15 <phh> nickjohnson: hence the "render list computed by the AP"
May 08 12:16:19 <mru> FLOATING POINT????
May 08 12:16:21 <mru> are you mad?
May 08 12:16:24 <nickjohnson> Yes, I started writing my answer before I read that
May 08 12:16:32 <phh> fair enough
May 08 12:16:34 <phh> and erm
May 08 12:16:43 <phh> stm32f4 are also considered as uCs :D
May 08 12:16:44 <mru> I also said nothing about using an avr specifically
May 08 12:16:45 <nickjohnson> mru: I tried a fixed point implementation. It was actually slower, probably because their division op was less optimized than gcc's.
May 08 12:16:56 <mru> you could use a cortex-m4 with fpu if you insist on using floats
May 08 12:17:02 <nickjohnson> mru: No, but it's the one I used, so I can speak specifically to my experience with it
May 08 12:17:21 <av500> if you have a render list, you dont need a separate uC
May 08 12:17:32 <av500> a 1GHz arm can handle that just fine too
May 08 12:17:47 <mru> av500: not if the non-rtos decides to not play ball
May 08 12:18:07 <av500> you dont need an OS to toogle gpios
May 08 12:18:11 <phh> pff, just takes an interrupt the os doesn't take with highest priority
May 08 12:18:17 <av500> yep
May 08 12:18:21 <av500> FIQ ftw
May 08 12:18:21 <mru> you could probably get the dma controller to do most of the work, actually
May 08 12:18:33 <av500> gptimer12 to FIQ
May 08 12:18:34 <mru> but that might count as a hack
May 08 12:18:39 <av500> a neat one
May 08 12:18:50 <nickjohnson> Well, like I said, I'm not designing reactors or nuclear therapy, or anything even remotely important or dangerous. I'm prepared to take the risk that SCHED_FIFO or a low nice level will very occasionally cause glitches in the timing.
May 08 12:18:58 <phh> mru: using dma, hack ? wat ?
May 08 12:19:09 <av500> hack in the original sense
May 08 12:19:14 <aholler> it isn't only the scheduler.
May 08 12:19:19 <mru> phh: using dma transfer lists to reprogram the dma controller would be a hack
May 08 12:19:35 <phh> i can't see how it would be a hack ...
May 08 12:19:42 <phh> but i'll wait for av500 to define me what a hack is :p
May 08 12:20:20 <av500> my high school CS teacher defined "hacking" as banging loud on the keyboard keys - she did not like that.....
May 08 12:20:29 <av500> she had no idea about CS either...
May 08 12:21:08 <mru> funny, neither did mine
May 08 12:21:19 <av500> phh: http://www.urbandictionary.com/define.php?term=hack
May 08 12:21:22 <mru> to the point she asked me to take over some classes
May 08 12:21:23 <nickjohnson> av500: And yes, I could use render lists. The original reason I didn't is because I'm communicating with the AVR over bluetooth, and bandwidth can often be an issue with that. And I'm using bluetooth because I don't have room for both an MCU board and a more complex one such as the beagleboard together (yes, I could design a stepper board with onboard MCU; I'm trying to keep my count of 'parts onl
May 08 12:21:23 <nickjohnson> y I will ever use' to a minimum)
May 08 12:22:09 <phh> "To program a computer in a clever, virtuosic, and wizardly manner." hack, using the dma is hacking.
May 08 12:22:23 <phh> s/hack, /ok, /
May 08 12:23:02 <av500> now, using the DMA to hack into another computers, that's a HACK!
May 08 12:23:15 <aholler> intel does that
May 08 12:23:23 <aholler> and apple
May 08 12:23:53 <aholler> hackers they are ;)
May 08 12:24:37 <av500> phh: and since 1. and 3. are somewhat opposites you can call a lot of things a hack :)
May 08 12:25:16 <phh> indeed
May 08 12:25:25 <phh> av500: well, in all definitions, it means that it works :p
May 08 12:25:32 <av500> true that
May 08 12:26:21 * nickjohnson is now pondering if he should've put a very low power MCU on the stepper board and communicated over UART or I2C, though
May 08 12:27:01 <phh> nickjohnson: it's possible without it, but noone dares doing so
May 08 12:27:26 <nickjohnson> phh: Well, as I said, if you want constant speed (which a lot of stepper applications do), you can always use the PWM support
May 08 12:27:56 <aholler> but you will miss the start and end points
May 08 12:28:05 <av500> nickjohnson: so you have a stepper controller with a "step" pin?
May 08 12:28:08 <phh> nickjohnson: how do you drive your steppers ?
May 08 12:28:18 <phh> .
May 08 12:28:19 <av500> but dont you need to count the steps?
May 08 12:28:30 <nickjohnson> I'm using pololu's A4988 stepper boards, which have 'step' and 'dir' pins
May 08 12:28:52 <GregoryGL> Can someone provide me a working angstrom build with opencv and toolchain? I made one with the online builder but can't get it to boot :(
May 08 12:28:52 <phh> ok
May 08 12:28:53 <av500> but then you have no microsteps.... omg
May 08 12:28:56 <phh> that sounds totally weird to me
May 08 12:28:58 <nickjohnson> av500: Yes. Most PWMs can be configured to stop at the end of each count, or to increment a timer, or trigger an interrupt
May 08 12:29:10 <nickjohnson> av500: No, they have microstep selector pins too
May 08 12:29:25 <nickjohnson> Though on the board I just sent off, they're broken out as solder jumpers, so you have to decide on microstepping level ahead of time
May 08 12:30:20 <nickjohnson> phh: It's convenient. :)
May 08 12:30:30 <phh> nickjohnson: indeed
May 08 12:30:54 <av500> nickjohnson: ic
May 08 12:31:03 <nickjohnson> And yes, I agree, trying to control them via an H-Bridge from GPIOs, particularly with microstepping, would be insane.
May 08 12:31:37 <phh> two h bridges*  (at least)
May 08 12:31:45 <phh> so 4 gpio outputs at least
May 08 12:31:46 <phh> anyway.
May 08 12:32:04 <av500> 2 gpio at least
May 08 12:32:20 <phh> nickjohnson: use timers that you retrigger every time and set next time you'll have to change something ?
May 08 12:32:22 <nickjohnson> yes
May 08 12:32:40 <nickjohnson> phh: What's your suggestion in reference to?
May 08 12:34:26 <av500> nickjohnson: since you already have stepper controllers, I dont think sysfs is so bad to start with
May 08 12:34:45 <nickjohnson> av500: I still hate the overhead and ugly API of it
May 08 12:34:53 <av500> nonsense
May 08 12:35:07 <nickjohnson> I think it's great for poking around on the command line, but I want something better and less kludgy as an API
May 08 12:35:08 <av500> overhead maybe if there is one you can measure
May 08 12:35:35 <av500> for API what do you care, you write a gpio_set( int on ) call and thats it
May 08 12:35:44 <nickjohnson> Measuring it is a good idea. I'll get back to you on that. ;)
May 08 12:35:48 <av500> and its only 3 lines
May 08 12:35:54 <nickjohnson> Well, what if I want to set more than one bit at the same time?
May 08 12:36:06 <av500> if you really need same.time
May 08 12:36:11 <av500> then yes, its lacking
May 08 12:37:47 <nickjohnson> Anyway
May 08 12:37:49 * nickjohnson -> bed
May 08 12:38:02 <nickjohnson> Thank you for your insights. Benchmarks forthcoming. :)
May 08 12:43:50 <isbric> Evning, i need some help getting my uart2 up and running on my bone.
May 08 12:45:00 <isbric> im hucked up to pins 21,22 but /dev/ttyO1 doesnt get anything.
May 08 12:47:14 <Crofton|work> anyone know much about disconnects on hard drives attahed top the ehci port (with a 3.0 vintage kernel)?
May 08 12:48:34 <bradfa> isbric, set the pin mux correctly?
May 08 12:51:32 <isbric> bradfa: i dont know how, my /sys/ tree looks different from the angstroms and i cant find anything on arch linux
May 08 12:51:49 <bradfa> mount debugfs?
May 08 12:52:03 <isbric> also  i read that uart2 shuld be muxed by default, but that might have been on angstromg aswell.
May 08 12:52:21 <bradfa> if you have debugfs mounted, you should be able to take a look-see
May 08 12:53:39 <bradfa> debugfs is kernel thing, not arch / angstrom / etc thing
May 08 12:54:53 <bradfa> cat /sys/kernel/debug/omap_mux/uart1*
May 08 12:56:59 <isbric> i dont have it mounted, how does one use debugfs..
May 08 12:58:27 <bradfa> https://en.wikipedia.org/wiki/Debugfs
May 08 12:58:33 <av500> one poke around in it
May 08 12:58:35 <av500> pokes
May 08 12:58:59 <isbric> it was to easy
May 08 12:59:06 <isbric> thx though
May 08 12:59:20 <isbric> mount -t debugfs none /sys/kernel/debug/
May 08 12:59:26 <bradfa> mount it then you can cat and echo to files in it, it makes things happen
May 08 13:00:11 <isbric> are the state changes saved on reboot or do i need to do that every time i have booted?
May 08 13:00:27 <bradfa> not saved
May 08 13:00:46 <bradfa> if you want it to come up with a particular mux, either rebuild the kernel with that mux or make a script to set it at boot time
May 08 13:02:30 <bradfa> isbric, this might be helpful: https://webcache.googleusercontent.com/search?q=cache:K1xhzOX_l80J:www.gigamegablog.com/2012/01/22/beaglebone-coding-101-using-the-serial-and-analog-pins/
May 08 13:02:40 <bradfa> google cache, site isn't responding for me
May 08 13:02:53 <isbric> yeah i was there, database down
May 08 13:03:10 <isbric> thx
May 08 14:46:57 <ncbas> Decided to show some pictures of his 3-D printer interface cape, using the PRUSS for step pulse generation:
May 08 14:47:01 <ncbas> http://imageshack.us/photo/my-images/855/20120501dsc2172topexpl.png
May 08 14:50:16 <stuk_gen> hi all :) i'm trying to configure qt embedded with -plugin-gfx-powervr with openembedded. I change the line QT_GLFLAGS to "-opengl es2 -depths 16,24,32 -plugin-gfx-powervr" but receive error pvrqwswsegl.c:42:30: fatal error: GLES/eglplatform.h: No such file or directory
May 08 14:51:05 <stuk_gen> when i building manually the packet i have to set where graphic sdk from ti where installed...how can i set this in openembedded?
May 08 14:52:09 <av500> ncbas: driving the popolus how?
May 08 14:52:15 <av500> ah, pruss
May 08 14:52:17 <av500> right
May 08 14:52:34 <ncbas> Yup, tested upto 60 kHz
May 08 14:54:07 <koen> ncbas: nice!
May 08 14:54:07 <koen> ncbas: have you considered allowing >12V for the steppers?
May 08 14:54:07 <koen> ncbas: I run my pololus at 24V, gives me more torque, less heat
May 08 14:54:18 <aholler> eh, there is a big cap, that won't be usable for decades ;)
May 08 14:54:53 <ncbas> Koen: Pololus have separate supply.
May 08 14:55:22 <koen> ah, top right corner
May 08 14:55:24 <ncbas> aholler: Nichicon long life cap
May 08 14:55:53 <aholler> ncbas: was just a joke, regarding a discussion yesterday or so ;)
May 08 14:56:31 <av500> aholler: I guess there is a special circuit to make the cap explode too
May 08 14:56:57 <ncbas> aholler: No joke, there are no other alu or tantal capacitors present. It's a rigid design.
May 08 14:57:54 <aholler> will tantal caps expire too?
May 08 14:58:23 <ncbas> aholler: mostly they explode ;-)
May 08 14:58:36 <aholler> never seen such
May 08 14:58:53 <LetoThe2nd> oO( big badaboom? )
May 08 14:58:55 <bradfa> just don't overvolt tant caps
May 08 14:59:19 <bradfa> then they less likely to xplode
May 08 14:59:27 <ncbas> I have, also seen burnt traces by exploding /short circuiting tantal caps.
May 08 14:59:29 <mru> much more fun to overvolt huge electrolytic caps
May 08 14:59:40 * LetoThe2nd likes badaboom.
May 08 14:59:47 <bradfa> for various definitions of the word 'fun' :)
May 08 14:59:58 <mru> anything that goes boom
May 08 15:00:07 <mru> bigger boom, more fun
May 08 15:00:36 <bradfa> beagle makes little tiny boom when it dies
May 08 15:00:44 <bradfa> no fun
May 08 15:01:24 * mru spent a summer working for these guys: http://www.profoto.se/
May 08 15:01:28 <mru> they have large caps
May 08 15:01:52 <bradfa> photo flashes are awesom
May 08 15:02:14 <LetoThe2nd> mru: hehe. kaboom.
May 08 15:02:23 <mru> when those caps accidentally discharged into the control board, the results were interesting
May 08 15:02:34 <bradfa> what kind of voltage on those flash controllers use?
May 08 15:02:34 <LetoThe2nd> s/interesting/entertaining/
May 08 15:02:45 <koen> ncbas: what's your plan for building more boards?
May 08 15:02:46 <mru> bradfa: those run at 1kV
May 08 15:02:51 <bradfa> yey!
May 08 15:03:10 <mru> 2.4kJ per discharge
May 08 15:03:23 <bradfa> joules are such an odd unit, no one has any idea how much one is
May 08 15:03:31 <mru> which takes a few milliseconds
May 08 15:05:28 <ncbas> koen: I've just ordered a new batch of pcb's. Using two of the shown bords myself, porting the software right now.
May 08 15:05:36 <av500> bradfa: about one BigMac
May 08 15:05:53 <bradfa> av500, from US or EU? :)
May 08 15:06:09 <koen> ncbas: did you do any work on teacup?
May 08 15:06:12 <av500> US it seems
May 08 15:07:10 <ncbas> koen: yeah, used it with my printer in the past. Too hard to change the software to my requirements. And then I saw the BeagleBone...
May 08 15:07:25 <koen> ncbas: ok, so you're the same Bas :)
May 08 15:07:37 <ncbas> koen: Ja!
May 08 15:18:47 <koen> ncbas: any fancy things like coordinated acceleration planned?
May 08 15:20:10 <ncbas> koen: Just normal ramping (fixed accel and decel up to now). It that what you're referring to?
May 08 15:22:45 <Crofton|work> I'm impressed koen knows about that :)
May 08 15:23:44 <mru> he's probably just name dropping
May 08 15:27:48 <koen> ncbas: matching the accel with the extruder axis
May 08 15:31:34 <ncbas> koen: IIR that's already the case.
May 08 15:31:46 <mru> spotify decided to play 'the idiots are taking over', is it trying to tell me something?
May 08 15:32:26 <aholler> they are years behind
May 08 15:33:18 <mru> and now it's playing 'I fought the law'
May 08 15:34:09 <aholler> afterwards they play "the ph0rum song"
May 08 15:36:05 <av500> dont forget the "firefox anthem"
May 08 15:36:52 <aholler> and the "web2zero opera"
May 08 15:38:29 <av500> https://firefoxflicks.mozilla.org/en-US/video/589
May 08 16:10:30 <av500> mru: what about a gpiod that python apps talk to via dbus?
May 08 16:10:54 <koen> don't forget systemd
May 08 16:11:08 <mru> we already had that
May 08 16:11:10 <av500> well, it will be merged into it eventually
May 08 16:13:00 <serekson> I am trying to apply enable SPI using OE using the patch here: http://elinux.org/BeagleBoard/SPI as reference.  But I don't understand how to apply the patch
May 08 16:13:47 <av500> serekson: can you recompile the kernel?
May 08 16:14:04 <serekson> yeah, that works fine
May 08 16:14:56 <av500> you apply pathces with "patch"
May 08 16:15:02 <av500> man patch
May 08 16:15:09 <mru> really, not with git am?
May 08 16:16:31 <serekson> if I use patch then how does git know what i have done?
May 08 16:16:53 <Russ> git status
May 08 16:17:44 <serekson> so then if I git -rebase it will try and merge my changes with the updated main repository?
May 08 16:18:49 <serekson> and should I branch before I apply the patch?
May 08 16:22:54 <Russ> I don't know the context, but its usually best to work on a branch
May 08 16:22:57 <av500> patch is one thing, a git commit another one
May 08 16:23:25 <av500> and yes, rebase will replay your commit on top of the updated main repository
May 08 16:23:58 <av500> serekson: patch only modifies the file contents, same as vi does
May 08 16:30:00 <serekson> so fi I patch and/or modify  the appropriate files in the build/tmp-Anstrom directory and then do "bitbake virtual/kernel" they will be incorporated into the new kernel image?
May 08 16:32:39 <av500> serekson: with angstrom its more difficult
May 08 16:32:52 <av500> google "angstrom kernel workflow"
May 08 16:33:45 <mdp> favorite e2e thread of the week: http://e2e.ti.com/support/embedded/linux/f/354/p/186504/672485.aspx#672485
May 08 16:34:43 <Sgarr> serekson: http://www.slimlogic.co.uk/2011/05/openembeddedangstrom-kernel-workflow/   <—I found this helpful, probably what av500's search will lead you to
May 08 16:35:11 <Russ> mdp, hey, cool, now there doesn't need to be any new threads
May 08 16:35:23 <mdp> Russ, we can only hope
May 08 16:35:26 <av500> Sgarr: correct
May 08 16:36:18 <mdp> Russ, I like the slight implication in the one post that maybe ajay could write his driver for him since he has a tight schedule
May 08 16:36:57 <av500> mdp: I think Ajay should write drivers for everybody!
May 08 16:37:10 <mdp> more drivers for you!
May 08 16:38:08 <mranostay> koen: nice g+ post
May 08 16:39:18 <nemik> hello, so according to http://processors.wiki.ti.com/index.php/AM35x_EMAC/MDIO_Module the beaglebone's main IC can't even really support gigabit rates, correct?
May 08 16:40:38 <nemik> reason i'm asking is i want to build a 'router-like' platform with the Beaglebone with a WAN RJ45 port and a LAN 4-port switch, so trying to figure out how to go about doing that while possible supporting gigabit rates from the LANs to the WAN
May 08 16:41:20 <mru> why would you use a beaglebone for that?
May 08 16:41:37 <mru> if you want to build a router, go get a board meant for that
May 08 16:41:49 <av500> last time I checked, the bone CPU did not have a 4 port LAN interface....
May 08 16:42:07 <av500> but maybe one call will one into existence
May 08 16:42:34 <nemik> mru: i want the CPU and PRU in the beaglebone but need to support ethernet-connected 'clients'
May 08 16:42:46 <av500> buy a $15 switch
May 08 16:43:32 <nemik> i can't do WAN and LAN simultaneously from a $15 switch
May 08 16:43:58 <mru> then get a proper router system
May 08 16:44:48 <nemik> this is custom hardware we're trying to fit into one board so space is a consideration
May 08 16:45:05 <mru> well, the bone can't do what you want
May 08 16:45:10 <nemik> yea i know that
May 08 16:45:11 <mru> so you have to find something else
May 08 16:45:22 <nemik> but the AM35x chip on it can, that's what i'm asking about
May 08 16:45:30 <av500> can what?
May 08 16:45:40 <mru> where does it have a 4-port ethernet interface?
May 08 16:45:52 <av500> add on over USB
May 08 16:45:57 <mru> ugh
May 08 16:46:10 <Russ> it does can
May 08 16:46:11 <av500> but then you dont need gigabit any more :)
May 08 16:46:24 <nemik> it supports two GMII/RMII interfaces, right? so one could be to a WAN regular PHY chip and the other to a LAN switch, no?
May 08 16:46:31 <mru> Russ: it can can?
May 08 16:46:55 <Russ> I don't know how to respond to that with any more can
May 08 16:46:56 <mdp> no can do
May 08 16:47:10 <nemik> no, can do
May 08 16:48:02 <nemik> alright, so i take it no one has experience with adding a second ethernet interface to the Bone or AM35x?
May 08 16:48:58 <mdp> nemik, yes, you could do that…been done many times on other parts
May 08 16:49:36 <mdp> s/AM35x/AM335x/
May 08 16:50:16 <nemik> ah so the AM335x would be more appropriate for that than BeagleBone's AM35x?
May 08 16:50:21 <av500> wasnt there some issue with the gbit ports not doing gbit actually?
May 08 16:50:32 <mdp> BeagleBone has AM335x on board
May 08 16:50:32 <LetoThe2nd> av500: there was.
May 08 16:50:46 <nemik> shit, that's right. i'm all mixed up here, sorry
May 08 16:51:09 <nemik> av500: ah so it won't really work?
May 08 16:51:11 <mdp> av500, on the early evm
May 08 16:51:30 <nemik> was it a problem with the chip or just layout of the board?
May 08 16:51:44 <mdp> layout
May 08 16:51:52 <nemik> ah ok, thanks mdp
May 08 16:52:54 <Mojito> Later revisions fixed that
May 08 16:53:39 <nemik> ah i think that's why BeagleBone used the MII instead of the RMII interface too, because it needs a faster clock and the board layout wasn't cooperating
May 08 16:53:52 <nemik> s/it/RMII
May 08 17:22:30 <aravind> Hi
May 08 17:22:59 <aravind> I'm trying to connect a pico projector to my beagle bone ...can anyone help me ?.
May 08 17:34:49 <mranostay> aravind: less vague please
May 08 17:45:12 <jsabeaudry> news about the zombie beaglebone: A timing issue during the communication with my fpga cause me to read invalid data and do a AB-sync dma with a negative CIDX thus overwriting kernel memory, all is under control now.
May 08 17:47:23 <djlewis> He was in the right place, but it must have been the wrong time... Sing along ;)
May 08 17:59:11 <mdp> my head is in a bad place, but I'm having such good time...
May 08 18:31:14 <isbric> anyone know what to make out of this: zd1211rw 1-1:1.0: error ioread32(CR_REG1): -19
May 08 18:31:35 <isbric> bone on external power, works just fine on usb power.
May 08 18:57:02 <CanyonMan> I am getting a TON of these 'babble' error messages from musb and it is making me want to cry
May 08 18:57:10 <CanyonMan> the thign that's weird is it seems to be at exactly a 22 second interval
May 08 18:57:14 <CanyonMan> What the hell is that about?
May 08 19:02:02 <mdp> usb guy says that's a h/w bug
May 08 19:09:46 <mru> gptimer22?
May 08 19:11:14 <mranostay> heh musb strikes again
May 08 19:12:25 <mru> hasn't it had its 3 strikes yet?
May 08 19:12:45 <ogra_> sysctl -e kernel.printk=0
May 08 19:12:47 <ogra_> ;)
May 08 19:13:34 <mru> that is an ubuntu-style fix
May 08 19:13:50 <ogra_> heh
May 08 19:14:02 <ogra_> at least it keeps your logs small
May 08 19:16:46 <isbric> http://pastebin.com/ThkMqSTn
May 08 19:17:13 <isbric> i dont even know where to start :(
May 08 19:27:29 <mdp> s/usb guy/usb sw guy/
May 08 19:29:52 <koen> rumour has it that the babble is a hw problem
May 08 19:29:59 <koen> but the jury is still out on that one
May 08 19:31:57 <_av500_> out drinking to forget
May 08 19:34:08 <koen> av500, mru: what's the plan for linuxtag?
May 08 19:34:19 <_av500_> none yet here
May 08 19:34:41 <_av500_> I might pop in for one day
May 08 19:34:55 * mru will be there
May 08 19:34:59 <mru> same hotel as last year
May 08 19:36:07 <koen> the russian one?
May 08 19:37:35 <mru> abendstern
May 08 19:37:39 <mru> yeah, the russian one
May 08 19:37:48 <mru> cheap and clean enough to sleep in
May 08 19:38:01 <mru> and close to the s/u-bahn
May 08 19:42:36 <CanyonMan> mdp: does the h/w bug = 22 seconds?
May 08 19:44:53 <spacecolonyone> hey guys any advice on boot faliure with the xMrevC after opkg update/upgrade.
May 08 19:45:16 <_av500_> dont opkg upgrade
May 08 19:46:20 <spacecolonyone> I'd just rebuilt a clean task-base-extended with bb and upgraded. looks like the upgrade tried moving me from 3.0.28->3.2.16 (which excites me) but without booting love.
May 08 19:46:29 <_av500_> otherwise, pastebin boot logs etc..
May 08 19:46:29 <mdp> CanyonMan: noise on the lines is what a babble interrupt is
May 08 19:47:25 <spacecolonyone> kk av500 in a few, caffeine first
May 08 19:48:16 <damian0815> hey guys
May 08 19:48:19 <CanyonMan> huh.
May 08 19:48:25 <CanyonMan> This is a usb cell card modem
May 08 19:48:28 <CanyonMan> it doesn't even have a cable
May 08 19:48:30 <CanyonMan> it plugs right into the bone
May 08 19:48:41 <mdp> CanyonMan: mmm hmmm….
May 08 19:48:44 <CanyonMan> I noticed that the 5V and 3.3V are fairly noisy
May 08 19:48:48 <mdp> it has traces
May 08 19:48:59 <mdp> and so does the bone
May 08 19:49:01 <CanyonMan> I wonder if cleaning those up some would help
May 08 19:49:16 <mdp> ask on the list, see what gerald says
May 08 19:49:59 <mdp> fwiw, jkridner already raised another babble related issue internally…more issues on the list might generate some concern
May 08 19:50:28 <isbric> CanyonMan: clening them up with somthing like a 1uf capacitor?
May 08 19:50:37 <CanyonMan> yeah
May 08 19:50:54 <isbric> let me know if you get some results.
May 08 19:51:00 <CanyonMan> k
May 08 19:55:02 <tasslehoff> koen: fyi: ogre still works fine in the 2011.03 maintenance branch of oe-classic. I just messed things up every way I could :p
May 08 19:57:54 <damian0815> hey, does anyone if the beagle/omap3 i2c interface supports dma?
May 08 19:58:28 <damian0815> ^ anyone know if
May 08 19:58:41 <prpplague> damian0815: do you really need dma to transfer at 100KHz?
May 08 19:58:54 <prpplague> or even at 400Khz?
May 08 19:59:02 <damian0815> no, but i don't really know what i'm doing so i'm trying to probe all the options ;-)
May 08 19:59:24 <prpplague> damian0815: use what works first, then if there is a problem start looking at buzz words
May 08 20:00:47 <damian0815> well, there's a problem. i need to read 8 12-bit DAC inputs and make them available to an audio DSP chain running on the CPU. i want a fresh DAC input reading for every audio buffer processed, ie (44100/64) reads per second
May 08 20:01:04 <damian0815> i'm struggling with synchronization.
May 08 20:01:15 <prpplague> damian0815: and you want to do this over i2c?
May 08 20:01:34 <damian0815> not necessarily. but the DACs i've been testing so far have been i2c dacs.
May 08 20:01:40 <Russ> damian0815, didn't I already do the math for you
May 08 20:01:53 <Russ> damian0815, and wasn't it already explained that i2c is the control interface?
May 08 20:02:02 <damian0815> Russ: i'm sorry?
May 08 20:02:18 <spacecolonyone> _av500_: http://pastebin.com/mfm0K4Vn
May 08 20:02:25 * prpplague sees this has already been discussed and goes back to work
May 08 20:02:37 <damian0815> Russ: are you saying that i2c is not intended for use as a data transfer bus?
May 08 20:03:14 <damian0815> apologies for not already knowing stuff!
May 08 20:03:18 <damian0815> :-(
May 08 20:04:07 <tasslehoff> normally you setup stuff by writing registers over i2c, and cool audio data starts flowing on some other pins from the dac
May 08 20:04:22 <damian0815> i'm not reading audio data
May 08 20:04:42 <damian0815> it's control voltage inputs from LFOs.
May 08 20:04:56 <damian0815> only needs to be read at 1000hz or so.
May 08 20:05:34 <spacecolonyone> as a side: I don't understand how u-boot now knows to go get the kernel image at /boot/uImage on the ext3 partition instead of using the one in boot partition. Was there a behind-the-scenes change in where u-boot looks over the past few months?
May 08 20:05:39 <Russ> damian0815, did you actually read the datasheet for your dac?
May 08 20:05:46 <damian0815> Russ: yes i did.
May 08 20:05:49 <Russ> what dac?
May 08 20:06:17 * tasslehoff worked with audio before, and his mind is hardwired to think audio samples when he hears about a dac
May 08 20:06:47 <damian0815> max11614
May 08 20:07:20 <damian0815> 8ch 12bit i2c
May 08 20:07:41 <damian0815> to be clear:
May 08 20:08:09 <Russ> ok, so that one isn't like an audio dac
May 08 20:08:12 <Russ> the data interface is i2c
May 08 20:08:20 <damian0815> Russ: yes, i know.
May 08 20:08:23 <mdp> thx for posting CanyonMan
May 08 20:08:36 <damian0815> Russ: 'wasn't it already explained' :-P
May 08 20:08:48 <Russ> you want 44.1k reads per second?
May 08 20:09:00 <damian0815> no, i want 1k reads per second, for each channel
May 08 20:09:01 <CanyonMan> mdp: I'm not really even sure that adding more caps is going to help in any way.  I'm going to go tell my guy working on this to try an extension though and get the radio far away from the bone and see if that helps him.
May 08 20:09:08 <agmlego> Why are we reading a DAC?
May 08 20:09:10 <damian0815> + there'll be a similar adc with 8 outs
May 08 20:09:21 <damian0815> we are making an analog modular synth module
May 08 20:09:30 <agmlego> OK, but a DAC is an output device.
May 08 20:09:34 <agmlego> No reason to read it.
May 08 20:09:41 <agmlego> And an ADC is an input device.
May 08 20:09:52 <damian0815> right yes sorry. adc/dac .
May 08 20:09:53 <agmlego> YOu read an ADC, you write to a DAC.
May 08 20:09:58 <damian0815> i swapped them over by mistake
May 08 20:09:58 <Russ> you you want to read from 8 devices?
May 08 20:10:19 <damian0815> i want 8 control voltage inputs, and i want 8 control voltage outputs
May 08 20:10:43 <Russ> so you want to read from 8 devices, and write to 8 devices
May 08 20:10:50 <damian0815> yes
May 08 20:10:52 <Russ> so 16k updates per sec
May 08 20:11:03 <damian0815> if that's achievable, yes
May 08 20:11:18 <agmlego> I2C can run at 100kHz or 400kHz.
May 08 20:11:31 <damian0815> i've set the i2c bus to 400khz
May 08 20:11:36 <agmlego> Which is also the datarate in bits per second, including packet overhead.
May 08 20:11:40 <damian0815> the issue is not i2c bus bandwidth
May 08 20:11:48 <agmlego> So what is the issue?
May 08 20:12:06 <damian0815> synchronizing the DAC process loop to the ADC inputs
May 08 20:12:06 <Russ> you want a 16 bit number from/to each one?
May 08 20:12:16 <damian0815> Russ: yes
May 08 20:12:28 <Russ> 16 * 16 * 1000 / 256k
May 08 20:12:35 <Russ> hold on, address
May 08 20:12:37 <damian0815> i'm happy to drop down to 4 in/ 4 out
May 08 20:12:50 <Russ> 384,000
May 08 20:12:57 <Russ> hold on, start, stop, ack
May 08 20:13:19 <Russ> 448,000 bits per second
May 08 20:14:04 <damian0815>  i'm happy to drop down to 4 in / 4 out and do the updates at 500hz or even 200 hz
May 08 20:14:12 <damian0815> but this isn't really the issue
May 08 20:14:21 <Russ> have you considered using one of the other serial buses for the data?
May 08 20:14:39 <agmlego> SPI, or putting the two devices on different I2C busses.
May 08 20:14:50 <agmlego> There are multiple I2C busses.
May 08 20:15:04 <Russ> then what is the issue
May 08 20:15:56 <damian0815> synchronisation. at the moment i'm using ioctl with a loop in userspace
May 08 20:16:08 <agmlego> Synchronisation how?
May 08 20:16:23 <agmlego> You want a packet in to trigger a packet out, or something like that?
May 08 20:16:37 <jsabeaudry> damian0815, use a simultaneous sampling ADC if you want to have stuff synchronized (MAX11046 for example)
May 08 20:16:44 <Russ> oh wait, hold on, let me factor in the possible clock stretch
May 08 20:16:52 <Russ> 8.3uS * 16 * 1000
May 08 20:17:12 <mdp> damian0815: fwiw, omap3 i2c supports dma, but the linux driver doesn't bother
May 08 20:17:28 <Russ> so clock stretch adds another 132.8ms per second
May 08 20:17:51 <agmlego> Russ: Who says the clock is being stretched?
May 08 20:17:55 <agmlego> And by how much?
May 08 20:17:56 <mdp> CanyonMan: yeah, I can't comment further..just passing along info..I've never used usb on Bone before, not interested in it
May 08 20:18:04 <damian0815> agmlego: ok, so the DSP chain processes 64-frame audio chunks, but it processes them asynchronously. i was thinking of having a kernel module that filled a buffer once every 1ms (or 2ms, depending on bandwidth), and then allow the DSP access to this buffer via shared memory as it needed to.
May 08 20:18:11 <agmlego> Russ: I think you are derailing the discussion here, to be honest.
May 08 20:18:23 <agmlego> damian0815: OK...
May 08 20:18:26 <damian0815> but obviously i don't want to be polling/sleeping in a kernel module. hence the idea of DMA. perhaps i don't understand it, but i was just looking for some pointers.
May 08 20:18:57 <agmlego> damian0815: Ah, I see. Sleeping should work fine in the kernel--that is how a lot of daemons run, works fine.
May 08 20:19:16 <Russ> damian0815, i2c does its work for you and then gives an interrupt when it's done
May 08 20:20:05 <damian0815> agmlego: the regularity is the important thing here. i want the ADC samples to be taken in sync with real-world time, i was assuming i'd have to manage that from the module itself.
May 08 20:20:52 <agmlego> damian0815: That may be, or use the RTC on the board and use the ALARM signal on Linux to get super-precise timing.
May 08 20:21:14 <agmlego> damian0815: Or if it is *really* critical, write the daemon as a realtime task and use the realtime extensions to the kernel.
May 08 20:21:18 <damian0815> if it helps, the overview is to build a module that will sit in a modular analog synthesis setup, that accepts control voltage signals from external (analog) LFOs and processes audio in-sync in realtime.
May 08 20:21:24 <Russ> damian0815, have you instrumented the omap i2c driver yet?
May 08 20:21:32 <damian0815> ALARM signal. thanks.
May 08 20:22:02 <damian0815> Russ: yes i have. works fine reading only one ADC channel, has a habit of annihilating the sound driver on buffer underrun.
May 08 20:22:14 <damian0815> esp. with a usb wlan device.
May 08 20:22:31 <Russ> "annihilating"?
May 08 20:22:38 <damian0815> yeah. audio just stops.
May 08 20:22:55 <Russ> you don't need to instrument every xfer
May 08 20:23:07 <Russ> just have it do every few thousand
May 08 20:23:13 <damian0815> what does 'instrument' mean in this context?
May 08 20:23:26 <Russ> printk with timing information
May 08 20:23:29 <agmlego> I am confused about that myself.
May 08 20:23:42 <damian0815> err, i'm not printking anything
May 08 20:23:43 <Russ> he seems to be very concerned with variable latency
May 08 20:23:55 <Russ> then have have you instrumented the omap i2c driver?
May 08 20:24:15 <damian0815> i don't understand your question, i'm sorry.
May 08 20:24:34 <agmlego> damian0815: I think Russ means by "instrumenting the driver", have you done timing analysis of the driver?
May 08 20:24:54 <agmlego> As in, have you attempted to figure out where the latencies you are having issues with are coming from?
May 08 20:24:58 <damian0815> agmlego: thanks for the clarification. no i haven't.
May 08 20:25:00 <Russ> you seem very concerned about timing, but you haven't really said what your timing requirements or problems are
May 08 20:25:11 <Russ> just vague statements
May 08 20:25:18 <agmlego> damian0815: Then how do you know you have a solvable timing issue?
May 08 20:25:24 <agmlego> Or one at all?
May 08 20:25:32 <Russ> "in sync with real-world time" isn't something that can be met
May 08 20:25:41 <Russ> you will always have jitter
May 08 20:26:00 <agmlego> Russ: You can get pretty close with either a signal-edge trigger, or an RTC.
May 08 20:26:12 <agmlego> Depending on what "real-world time" means.
May 08 20:26:14 <Russ> yahoo, I got it within a second, I'm done!
May 08 20:26:21 <Russ> or did you want femtoseconds?
May 08 20:26:30 <Russ> you need a jitter and latency budget
May 08 20:26:59 <damian0815> by 'in sync with real world' i meant 'sampled at regular intervals'
May 08 20:27:12 <Russ> again, meaningless
May 08 20:27:18 <Russ> what is "regular"
May 08 20:27:18 <agmlego> So, not actually in sync with real-world time at all.
May 08 20:27:34 <agmlego> Just sampled at the same time interval.
May 08 20:27:38 <Russ> "regular" is a measure of your jitter
May 08 20:27:46 <agmlego> Use the RTC, and the ALARM signal.
May 08 20:27:46 <Russ> and "sampled" is a measure of your latency
May 08 20:28:00 <Russ> agmlego, why the hell would you use RTC and ALARM?
May 08 20:28:05 <Russ> why not just clock_gettime
May 08 20:28:23 <agmlego> Russ: Because that has to use a busy loop.
May 08 20:28:29 <Russ> damian0815, if each sample jitters by 500ms, is that regular enough?
May 08 20:28:39 <agmlego> The ALARM signal is a timer interrupt, far more accurate.
May 08 20:28:56 <damian0815> Russ: no, 10ms is the max jitter
May 08 20:28:57 <agmlego> And it means the daemon can be idle, off CPU while waiting.
May 08 20:29:00 <Russ> agmlego, it isn't a timer interrupt, its just a signal
May 08 20:29:09 <Russ> damian0815, hooray, we have a jitter budget
May 08 20:29:21 <Russ> damian0815, is that end to end?
May 08 20:29:26 <agmlego> Russ: Sorry, I accidentally a word. The word "like" should be in there.
May 08 20:29:45 <agmlego> As min, the signal can be used *like* a timer interrupt.
May 08 20:29:56 <Russ> agmlego, the kernel is pretty free to add a lot of jitter to that
May 08 20:30:08 <agmlego> Russ: Not as much as 10ms though.
May 08 20:30:10 <Russ> the man page lists the time as a minimum, nothing about the maximum
May 08 20:30:15 <agmlego> And especially not with the RT extensions.
May 08 20:30:32 <agmlego> And I am talking about personal experience here on several projects.
May 08 20:30:55 <Russ> I think you'd get the same thing out of select and you wouldn't need signals
May 08 20:31:00 <damian0815> Russ: 'end to end' i'm going to assume that 'yes' is the correct answer to that question.
May 08 20:31:19 <agmlego> Russ: "select"?
May 08 20:31:31 <Russ> or poll, or pselect
May 08 20:31:50 <agmlego> Sure, Tim Toady always shows up.
May 08 20:32:09 <agmlego> Was I supposed to enumerate a comprehensive list of ways to accomplish this?
May 08 20:33:52 <Russ> signals are no fun
May 08 20:33:57 <Russ> select is much much easier
May 08 20:34:07 <agmlego> If you say so. ;-P
May 08 20:34:10 <damian0815> ok thanks for the pointers folks... i'm going to go research the ALARM timer and see if i can get something out of that.
May 08 20:34:15 <mru> Russ: epoll!
May 08 20:34:18 <Russ> course nanosleep is even easier
May 08 20:34:22 * agmlego deals with low-level stuff, and so signals are second nature.
May 08 20:34:30 <agmlego> damian0815: ALARm *signal*
May 08 20:34:39 <mru> SIGIRQ
May 08 20:34:51 <jsabeaudry> Russ, won't sleep induce drift?
May 08 20:34:52 <damian0815> agmlego: right you are. thanks.
May 08 20:35:01 <Russ> or clock_nanosleep
May 08 20:35:05 <mru> jsabeaudry: yes, it will
May 08 20:35:46 <Russ> ah, maybe timer_create() then
May 08 20:35:54 <mru> clock_nanosleep can sleep until an absolute time
May 08 20:36:13 <mru> as can sem_timedwait
May 08 20:36:24 <mru> which is usually available even if clock_nanosleep is not
May 08 20:36:29 <Russ> just make sure you aren't letting the kernel add slack
May 08 20:37:03 <mru> if you increment an absolute time in fixed steps, you're good
May 08 20:37:12 <mru> you might get some jitter but no drift
May 08 20:37:45 <mru> you should guard against over-sleeping by a full period or more though
May 08 20:44:10 <spacecolonyone> guys any ideas why I wouldn't be seening the linux 3.2 kernel recipe in my sources despite being up to date with oebb.sh
May 08 20:46:04 <spacecolonyone> could this be a branch issue of some sort? the changes look pretty recent (5 days) OTOH there is already a 3.2.16 package in the repository
May 08 20:57:47 <jsabeaudry> while (time_before(jiffies, j1)) { schedule(); } with a decent CONFIG_HZ and no dynamitc ticks would be very simple to get under 10ms jitter
May 08 21:00:14 <mru> unless something with higher priority gets in your way
May 08 21:00:59 <Russ> jsabeaudry, that's only good for kernel anyway
May 08 21:01:23 <mru> and there's always gptimer12 and fiq
May 08 21:04:09 <aleek> le fu! u-boot doesnt implement tftp command :<
May 08 21:06:14 <Russ> ?
May 08 21:07:36 <aleek> Im used to load kernel images via tftp while developing - Ive just found out, bb-xm's uboot doesnt provide tftp command :(
May 08 21:08:04 <Russ> I use it all the time
May 08 21:08:12 <Russ> maybe download a newer u-boot version for xm?
May 08 21:08:20 <aleek> maybe :>
May 08 21:08:24 <prpplague> or turn on the compile flag
May 08 21:09:01 <aleek> prpplague: ?
May 08 21:09:12 <Russ> aleek, did you build your own u-boot?
May 08 21:09:25 <aleek> not yet, board just arrived to me
May 08 21:09:29 <prpplague> aleek: network commands such as the tftp can be enabled or disabled depending on your configuration
May 08 21:09:45 <Russ> aleek, then download a newer pre-build binary
May 08 21:09:54 <aleek> ok ok :)
May 08 21:10:09 <Russ> it just goes on the mmc card
May 08 21:21:00 <aleek> uboot for -xM is the same as for regular bb?
May 08 21:21:28 <mru> can be
May 08 21:21:43 <mru> if it's a recent build, it should work on either
May 08 21:25:01 * xxiao is watching tizen live stream
May 08 21:25:52 <spacecolonyone> Any advice on double checking that I am indeed hanging at Uncompressing Linux... done, booting the kernel.
May 08 21:26:26 <aleek> spacecolonyone: what power do you use?
May 08 21:26:29 <aleek> usb?
May 08 21:26:40 <spacecolonyone> Heartbeat LED  doesn't and I've got console=ttyO2,115200n8
May 08 21:26:46 <spacecolonyone> nah, DC adapter
May 08 21:26:59 <spacecolonyone> xMrevC btw
May 08 21:27:30 <aleek> Ibe read, that otg can hang or panic linux, as it give less power than it needs
May 08 21:27:33 <aleek> Ive*
May 08 21:27:37 <spacecolonyone> this started right after opkg upgrade on a clean task-base-extended build
May 08 21:28:04 <spacecolonyone> I'll keep that in mind
May 08 21:28:57 <aleek> its interesing that all in all 30% of my boots went well
May 08 21:29:04 <aleek> by powering via usb
May 08 21:35:12 <spacecolonyone> well I think it is safe to say it really is hanging, simply changing the symlink from uImage-3.2.16 back to 3.0.28 and the boot ran normally
May 08 21:36:00 <spacecolonyone> well, barring any issues one would expect booting to a 3.0.28 kernel after upgrading with opkg to 3.2.16 (which I can't weigh in on)
May 08 21:39:10 <xxiao> will angstrom do hob someday
May 08 21:42:44 <mdp> spacecolonyone: did I understand correctly? 3.0.28 boots consistently and 3.2.16 doesn't?
May 08 21:44:06 <spacecolonyone> yes
May 08 21:44:40 <spacecolonyone> to be more specific: 3.2.16 has not ever booted past Uncompressing Linux... done, booting the kernel.
May 08 21:46:25 <spacecolonyone> I've been doing everything in bitbake, then opkg update/upgrade once I'm up running
May 08 21:47:28 <spacecolonyone> this last time it pulled down the 3.2.16 kernel and then no more booting until I changed the symlink back to 3.0.28
May 08 21:50:22 <aleek> btw, why hdmi cable cannot be plugged while bb is on?
May 08 21:51:15 <spacecolonyone> aleek: I've wondered that as well. I can say with some experience that plug/unplug with it on at least a few times has yet to do any damage
May 08 21:52:12 <aleek> maybe it is because hdmi plug on thee board is not grounded
May 08 21:57:40 <movax> how is pin 19 configured?
May 08 21:57:43 <movax> (on HDMI)
May 08 22:03:08 <serekson> when I do "bitbake virtual/kernel" the tarball only has the lib directory in it is this correct?
May 08 22:16:26 <spacecolonyone> mdp: Don't tease me, what are you thinking?
May 08 22:18:49 <Russ> spacecolonyone, turn on the PHYS_VIRT thing
May 08 22:18:53 <Russ> (in kernel config)
May 08 22:20:35 <spacecolonyone> Russ: I'm not familiar with that, is there a command line arg or do I need to reconfigure/rebuild?
May 08 22:21:38 <spacecolonyone> If it is the latter, then I'll first need to figure out how to get bitbake to do the 3.2.16 kernel instead of 3.0.16
May 08 22:21:39 <Russ> make sure your kernel is configured with CONFIG_ARM_PATCH_PHYS_VIRT=y
May 08 22:22:30 <spacecolonyone> How do I check that when I pulled the kernel from the angsrtom repo with opkg
May 08 22:31:22 <serekson> does "bitbake virtual/kernel" only provide you with the /lib directory and the uImage?
May 08 22:34:04 <spacecolonyone> Russ: I really have no clue how to tell what this particular kernel build config was I got with opkg. That said, I think I found the defconfig that may have been used in building in the repo referenced by koen at www.mail-archive.com/meta-ti@yoctoproject.org/msg00317.html
May 08 22:34:30 <spacecolonyone> at least in that defconfig CONFIG_ARM_PATCH_PHYS_VIRT=y
May 08 22:34:33 <spacecolonyone> is set
May 08 22:35:03 <Russ> if you do get it to boot, you can zcat /proc/config.gz | grep PHYS_VIRT
May 08 22:35:54 <spacecolonyone> thanks, I'll remember that. I'm not even sure how to approach diagnosing why it won't boot, though.
May 08 22:36:05 <Russ> did you say it boots sometimes?
May 08 22:36:58 <spacecolonyone> I figure I've got other fish to fry so I'll table it for now. I figure this is just too close to the bleeding edge.
May 08 22:37:01 <spacecolonyone> No
May 08 22:37:03 <spacecolonyone> never boots
May 08 22:37:10 <spacecolonyone> not with 3.2
May 08 22:37:37 <spacecolonyone> hangs at Uncompressing Linux... done, booting the kernel. No heartbeat LED either
May 08 22:38:50 <spacecolonyone> I'm just happy that the python haslib patch got incorporated into the appropriate repo and it was magically fixed in this recent bitbake build I did.
May 08 22:39:07 <spacecolonyone> now I've actaully got ipython running
May 08 22:40:26 <spacecolonyone> if only I could get the USB ethernet adapter to come up automatically. that has continued to stump me for a few months of on and off work
May 08 22:41:18 <spacecolonyone> I'm probably going to end up writing an simply script that runs after bootup that simply executes ifup eth1 for me
May 08 22:44:17 <Russ> auto usb0
May 08 22:44:21 <Russ> allow-hotplug usb0
May 08 22:44:31 <Russ> iface usb0 inet dhcp
May 08 22:46:05 <spacecolonyone> auto eth1
May 08 22:46:14 <spacecolonyone> iface eth1 inet dhcp
May 08 22:46:22 <Russ> the allow-hotplug is the one that you are missing
May 08 22:46:35 <spacecolonyone> in interfaces?
May 08 22:47:14 * spacecolonyone tries it
May 08 22:48:24 <Russ> yes
May 08 22:48:45 <Jetro> I'm a newbie to beaglebone and trying to load DEbian Squeeze on it but running in Fdisk issue
May 08 22:49:19 <spacecolonyone> misplaced- option allow-hotplug
May 08 22:49:29 <Jetro> Here's the error message "Sorry, this script currently doesn't work with GNU Fdisk"
May 08 22:50:09 <Jetro> So add in the script "sudo ./mk_mmc.sh --mmc /dev/sdd --uboot bone --distro squeeze"? Thanks
May 08 22:50:49 <Russ> spacecolonyone, where did you put it?
May 08 22:51:18 <spacecolonyone> it's own line, between auto eth1 and iface eth1 ...
May 08 22:51:26 <Russ> that's where I put mine
May 08 22:51:38 <Russ> and you put a 'allow-hotplug eth1', right?
May 08 22:51:53 <spacecolonyone> oddly if I comment out auto eth1 ifup eth1 works again
May 08 22:52:11 <spacecolonyone> without the '', yes
May 08 22:52:22 <spacecolonyone> sadly still isn't automatic
May 08 22:53:28 <Russ> I needed both the auto and allow-hotplug for the usb network adapter attached to my server
May 08 22:56:30 <spacecolonyone> I just double checked, 'auto eth1' XOR 'allow-hotplug eth1' and 'ifup eth1' works
May 08 22:56:35 <spacecolonyone> not both
May 08 22:56:57 <Russ> hu
May 08 22:56:58 <Russ> h
May 08 22:56:58 <spacecolonyone> but neither causes it to come up and request an IP at boot
May 08 22:57:46 <spacecolonyone> nor does allow-hotplug cause it to request one after booting if I pull the adapter and reinsert
May 08 22:59:31 <Russ> /lib/udev/net.agent is the start of the chain
May 08 22:59:56 <Russ> it should eventually call 'exec ifup --allow=hotplug $INTERFACE'
May 08 22:59:59 <Russ> is lo up?
May 08 23:00:41 <spacecolonyone> yea
May 08 23:00:46 <spacecolonyone> lo is up
May 08 23:01:14 <spacecolonyone> eth0 (the onboard ethernet) doesn't come up automatically either
May 08 23:01:20 <Russ> what happens if you do ifup --allow=hotplug eth1?
May 08 23:02:06 <spacecolonyone> there is no /lib/udev/net.agent
May 08 23:02:59 <Russ> /sbin/udevd --version?
May 08 23:03:33 <spacecolonyone> Russ that causes busybox to gripe about an unrecognized option
May 08 23:03:56 <spacecolonyone> no such file
May 08 23:04:40 <spacecolonyone> found it in lib/udev/udevd
May 08 23:04:43 <spacecolonyone> checking
May 08 23:05:03 <spacecolonyone> 182
May 08 23:06:07 <spacecolonyone> I'm using systemd, in case that affects your way of thinking
May 08 23:08:08 <Russ> spacecolonyone, then I'm not sure
May 08 23:08:46 <Russ> I think you need an eth1.service
May 08 23:08:54 <spacecolonyone> no worries mate, thanks for the help
May 08 23:08:54 <Russ> and the command is systemctl start ifup@eth1.service
May 08 23:09:12 <spacecolonyone> hrm
May 08 23:10:01 <Russ> I think the /etc/network/interfaces might be compatibility fluff on a systemd system
May 08 23:10:25 <spacecolonyone> yea, I was starting to get a hint of that
May 08 23:10:39 <spacecolonyone> ok, well I'll look into getting that or creating it. the command itself currently gives no such file
May 08 23:12:37 <Russ> https://wiki.archlinux.org/index.php/Systemd#Network
May 08 23:14:46 <spacecolonyone> ok, now to find the appropriate files
May 08 23:16:34 <spacecolonyone> busybox-udpcpc.service looks promising
May 08 23:18:23 <spacecolonyone> hahaha: busybox-udpcpc.service -> /dev/null
May 08 23:18:33 <spacecolonyone> thats never going to work
May 08 23:20:25 <spacecolonyone> alright, well I'll check in in the AM to see if anyone has had suggestions about this or about the kernel booting issue
May 08 23:22:28 <arugh> is there a stable tag to checkout for building angstrom for the beaglebone? I'm having trouble with the image produced with a current checkout.
May 08 23:28:49 <Russ> spacecolonyone, anything with 'udhcpc' in it *has* to be awesome
May 09 00:08:38 <nickjohnson> I just realised, an ATMega328 costs significantly less than these bidirectional level converters. So a shield cape should probably use one of those...
May 09 01:12:13 <jaykay> Hello everybody!
May 09 01:12:39 <jaykay> I am hopelessly stuck with a certain issue and am desperately in need of help
May 09 01:12:47 <jaykay> anybody up for helping me out?
May 09 01:15:06 <jaykay> nobody here?
May 09 01:17:35 <strata> ok so i got a beaglebone today. coming from arduino, where can i find decent docs/tutorials on doing stuff like i2c and pwm on the beagle?
May 09 01:17:54 <strata> oh wait. oops. is this a dev chan or a regular user chan? lol
May 09 01:18:55 <mdp> both
May 09 01:18:56 <mdp> http://www.gigamegablog.com/2012/01/05/beaglebone-coding-101-blinking-an-led/
May 09 01:19:31 <strata> mdp: yea. already did that :)
May 09 01:20:05 <mdp> http://www.gigamegablog.com/2012/03/16/beaglebone-coding-101-buttons-and-pwm/
May 09 01:20:47 <strata> mdp: did that too :)
May 09 01:20:56 <jaykay> anybody experienced out there? im having trouble with my beagleboard...help please!
May 09 01:20:58 <strata> i guess it's like.. where do i start? maybe i should install the cloud9 ide?
May 09 01:21:22 <mdp> strata, question answered then
May 09 01:21:25 <strata> mdp: i was just playing it by ssh'ing in and using python lol
May 09 01:21:38 <strata> but i suppose i'll dig into the IDE
May 09 01:21:58 <mdp> I don't use the ide, sounds like an adventure
May 09 01:22:03 * mru wonders how long before people try running eclipse on the beagle
May 09 01:22:12 <mdp> is there some expected value to the IDE?
May 09 01:22:17 <strata> *shrug*
May 09 01:23:50 <mdp> I suggest to start here: http://cwraig.id.au/?p=516
May 09 01:25:38 <mdp> then read the beagleboard list archives to see how other people are coding i2c, spi, pwm, and gpio…in those posts you'll see people learning about how pinmuxing works on one of these parts and how you manage mux settings for whatever feature you are using
May 09 01:26:10 <mdp> google beaglebone and FOO and you'll find tons of blog entries of people that got most everything working
May 09 01:26:17 <mdp> that's about it
May 09 01:26:57 <mdp> mru, I tell those people to go get an R-Pi, then hope they don't return
May 09 01:27:09 <mru> :)
May 09 01:28:35 <strata> i got a bunch of different devices working on an arduino uno. but now i shall try it on a beaglebone :)
May 09 01:28:55 <strata> anyways. first i must learn linux and C as per mdp's failtroll.
May 09 01:29:12 <mdp> you could use starterware
May 09 01:29:17 <mdp> or roll your own
May 09 01:29:21 <mdp> no need to use linux
May 09 01:29:22 <strata> can i run BASIC too?
May 09 01:29:27 <mdp> if you like
May 09 01:29:34 <meklort> I go with the roll your own option
May 09 01:29:34 <strata> cool!
May 09 01:30:56 <mdp> strata, in any case, I did show a nice starting point to get the native sdk and start building stuff
May 09 01:31:03 <mdp> there's also some nice youtube videos
May 09 01:31:05 <mdp> have fun
May 09 01:31:26 <jaykay> hey everybody, i need to find out if my board is dead, its not booting, only leds d13, d14 flashes and then d5 is on
May 09 01:31:35 <jaykay> can somebody help me determine if i need a new board
May 09 01:33:22 <rcn-ee> jaykay, have you tried what gerald asked?
May 09 01:33:40 <jaykay> hold on i must have missed it
May 09 01:34:07 <KeatonT> beaglebone coming tomorrow :)
May 09 01:34:15 <jaykay> what did gerald ask!!?!
May 09 01:34:17 <jaykay> geraaaald?!?!?1
May 09 01:34:28 <rcn-ee> although, i've accidentely ejected the mmc many times while it's been running, worst case, it wrecks the sd card and you just have to rebuild..
May 09 01:34:51 <rcn-ee> jaykay check the beagleboard group email...
May 09 01:34:54 <jaykay> I have done that, following instructions from circuitco
May 09 01:35:07 <jaykay> but still no success
May 09 01:35:14 <jaykay> and I was beginning to think maybe my board is a goner
May 09 01:35:19 <rcn-ee> do you have a 2nd sd card to try?
May 09 01:35:23 <jaykay> yes
May 09 01:35:26 <jaykay> i did
May 09 01:35:47 <jaykay> and besides, shouldn't i be getting some output at least even if i powerup without an sd card
May 09 01:35:54 <jaykay> i have nothing on DVI no matter what
May 09 01:35:56 <rcn-ee> nope...
May 09 01:36:22 <rcn-ee> the bootrom looks for the bootloader on the microsd card, if it doesn't find it... well not much happens on the xM..
May 09 01:36:35 <meklort> if it doesn't have it, it tries booting form the usb port
May 09 01:36:36 <jaykay> hmmmm
May 09 01:36:38 <meklort> err, serial port
May 09 01:36:53 <jaykay> ok 1 sec id like to share with u the instructions i followed
May 09 01:37:07 <jaykay> http://circuitco.com/support/index.php?title=BeagleBoard-xM#Image_Files
May 09 01:37:25 <jaykay> so i found these instructions saying it was the image which the beagleboard came with
May 09 01:37:36 <jaykay> and it was very simple instructions
May 09 01:37:44 <jaykay> so i created a new sd card
May 09 01:37:51 <jaykay> but still no success =/
May 09 01:38:15 <rcn-ee> then, you'll have to just reply to gerald, it'll probally have to rma'ed..
May 09 01:38:25 <jaykay> moreover, when i do "screen /dev/ttyUSB0 115200" and power up my beagle after the blank screen
May 09 01:38:27 <jaykay> i still get nothing
May 09 01:38:45 <jaykay> i can get it replaceD?
May 09 01:39:03 <rcn-ee> i'm just a user. ;)
May 09 01:39:25 <jaykay> ah i see
May 09 01:39:44 <jaykay> i should hang out here more often
May 09 01:40:01 <jaykay> im new to embedded linux
May 09 01:42:22 <jaykay> what nickname does gerald go by?
May 09 01:42:35 <rcn-ee> he's not on irc..
May 09 01:44:50 <jaykay> rcn-ee do u have a case for your beagleboard?
May 09 01:45:27 <rcn-ee> nope, just stacked 4 of them into a small case with plastic standoffs
May 09 01:45:47 <jaykay> you have 4 bb's ? :D
May 09 01:46:20 <rcn-ee> and bunch more other arm boards.. ;)
May 09 01:47:18 <jaykay> r u an academic or an engineer/enthusiast
May 09 01:51:45 <mranostay> s/r/are/
May 09 01:51:49 <mranostay> s/u/you/
May 09 02:32:00 <jaykay> long live supreme leader kim jong un!
May 09 02:34:44 <Russ> thats not funny
May 09 02:44:21 <strata> oo nice. latest angstrom doesn't boot.
May 09 02:44:22 <strata> hmmm.
May 09 02:44:33 <strata> common problem?
May 09 02:54:54 <jstearns74> strata: for beagleboard or beaglebone?  Did you download the latest image?
May 09 02:56:26 <arugh> the archlinux image for the beagle bone works really well so far
May 09 02:56:48 <strata> jstearns74: it's the website lol
May 09 02:56:49 <arugh> I couldn't build a version of angstrom that was stable
May 09 02:56:53 <strata> it won't download the full file
May 09 02:57:49 <jstearns74> I had that problem on my work network, but came home and had no problem.  I had better luck using the command prompt & wget.
May 09 02:57:57 <strata> yea using wget now.
May 09 02:58:17 <strata> their server must be running centos *cough* ;)
**** ENDING LOGGING AT Wed May 09 02:59:58 2012