**** BEGIN LOGGING AT Fri Oct 05 02:59:58 2012 Oct 05 04:54:34 <_av500_> mranostay: configure shrugs Oct 05 04:56:48 autoconf is evil Oct 05 04:57:37 <_av500_> autoconf is not evil, people are evil Oct 05 04:58:36 autoconf is evil... people just appear to be evil Oct 05 04:58:42 it is a reflection of autoconf Oct 05 05:16:37 heh Oct 05 05:31:21 ds2: appear? Oct 05 08:33:16 * mru trolls rmk Oct 05 08:34:09 with rmk the slow troll works best Oct 05 08:34:12 he's so predictable Oct 05 08:34:43 a few gentle nudges and you can have him exactly where you want him Oct 05 08:36:57 url? Oct 05 08:52:53 mru: url please Oct 05 08:53:17 I'm trying to enable use of a USB mouse on my Beagleboard xM rev C board. I'm basically starting with a stripped down linux-omap kernel, without USB or Input devices enabled. Oct 05 08:53:45 then you have to bitbang USB Oct 05 08:54:35 av500: I suspect http://lists.infradead.org/pipermail/linux-arm-kernel/2012-October/123977.html Oct 05 08:56:45 As a start I've enabled the Input layer and the Event interface, USB HID transport layer, USB support, EHCI HCD 2.0 support for OMAP3 devices Oct 05 09:03:56 but honestly I dont know what Im doing... Oct 05 09:05:27 then why are you doing it? Oct 05 09:05:37 why not use the stock omap3 config for beagle? Oct 05 09:06:02 he doesn't know why either Oct 05 09:06:09 they never do Oct 05 09:08:00 I do this because I want to learn, but I guess I need to read up on how USB works Oct 05 09:08:16 I'll try the stock omap3 config Oct 05 09:09:20 Just asking if there is any common or obvious mistake I should keep in mind Oct 05 09:10:39 hm if disable a feature, feature dont work, lesson learned Oct 05 09:13:03 well, it's going from "i want to be able to connect a usb-mouse to the beagleboard" to actually enabling it in the kernel thats hard for me Oct 05 09:19:58 woglinde: I can think of a manager or two to whom that's far from obvious Oct 05 09:21:49 kroon: well, then have a look at the default config and see what is enabled Oct 05 09:22:26 av500, yeah, using the defconfig actually works :-) so I'll go from there Oct 05 09:22:30 thanks Oct 05 09:34:37 koen: did hou rmk's latest reply? Oct 05 09:35:14 hou all the things! Oct 05 09:50:57 mru: the one about being too budy? Oct 05 09:53:50 yeah, didn't see that one coming :) Oct 05 10:07:06 From the omap 3530 errata: 3.1.1.130 only one usb dma channel (rx or tx) can be active. Anyone know if this is handled according to the possible workarounds in the kernel? My grep-fu is failing me. Oct 05 10:09:36 it was at some point Oct 05 10:09:48 then someone killed it again Oct 05 10:09:56 then I lost track Oct 05 10:11:14 I'm still running patched 2.6.32 on my classic beagles Oct 05 10:11:50 hm. when I setup a test with mucho nw-traffic both out and in, musb/network just stops working. No complaints, just the silent treatment. Oct 05 10:13:41 *yawn*...mmm drama on arm-kernel too Oct 05 10:13:48 fun couple days Oct 05 10:14:17 mru, you're going to have to do better to tear me away from "as the udev turns" Oct 05 10:14:33 udev drama is epic Oct 05 10:14:55 since devtmpsfs, what is the use of udev anyway? Oct 05 10:15:04 is it still going? Oct 05 10:15:46 the overall I meant Oct 05 10:18:09 av500: rules and policy Oct 05 10:18:33 av500: creating extra symlinks like /dev/input/something -> /dev/input/touchscreen0 Oct 05 10:18:41 oh Oct 05 10:18:43 av500: so you can do without if you don't need that Oct 05 10:18:54 thats a worthy task Oct 05 10:18:55 truly Oct 05 10:20:07 and you can't bitbang that one Oct 05 10:20:27 now with libkmod module loading happens differently which seems to trigger the firmware loading bugs Oct 05 10:20:37 TBH I think both sides in the argument are wrong Oct 05 10:20:49 the kernel should fix their shit and udev should add the workaround Oct 05 10:21:21 linus should aknowledge that you can trigger the same bug with the init=/bin/sh trick Oct 05 10:23:32 lennart and kay deserve some linus rage Oct 05 10:23:56 doesn't reslly matter over what Oct 05 10:47:31 http://processors.wiki.ti.com/index.php/UsbConfigControllerDriver says to disable dma on musb, and in meta-ti I see that beaglebone has done just that. same otg-shite on beagleboard, isn't it? Oct 05 10:48:12 tasslehoff: the recommendation to disable is based on the use of the cppi 4.1 dma block with musb on am335x Oct 05 10:48:31 am/dm37x don't integrate cppi 4.1 dma with musb Oct 05 10:48:43 so no, different otg-shite on beagleboard Oct 05 10:48:51 mdp: ok. a bit hotheaded here :) Oct 05 10:49:06 the 3530 can use sdma for one direction Oct 05 10:49:22 tasslehoff: np, it's a bit confusing the way we use sdma, mentor dma, cppi 4.1 dma here and there Oct 05 10:49:27 and the musb builtin dma for the other Oct 05 10:49:39 what mru says Oct 05 10:50:53 there's some sekrit internal hacks and benchmarks against 2.6.32 I once worked with that do various split sdma/mentor dma tricks with varying EP configurations to optimize performance Oct 05 10:51:10 typically used to show people a level of awesomeness on benchmarks Oct 05 10:51:53 2.6.32 was a good kernel Oct 05 10:53:13 not ours Oct 05 10:53:27 I'm trying to figure out if 2.6.39 does what mru said. Also running a test with PIO to see if it goes down then. Oct 05 10:53:49 but then 2.6.32 + thousands of patches usually ends up badly Oct 05 10:54:14 so that's why omap went multicore arm Oct 05 10:54:21 thousands of patches does that to anything Oct 05 10:54:26 to have a dedicated core for bitbanging usb Oct 05 10:54:57 tasslehoff: we also have cppi 4.1 dma on am180x/da850/l138 which is where I first noticed a bunch of these cppi 4.1 dma related failures Oct 05 10:58:10 tasslehoff, http://www.ti.com/lit/er/sprz360c/sprz360c.pdf, Advisory 1.0.13 Oct 05 10:59:34 SPRZ is the sound of a dying car engine Oct 05 11:00:30 tasslehoff: have a look at our frankenkernels, they should have a lot of musb related fixing and monkeypatching Oct 05 11:00:31 one wishes TI could change the naming convention adopted during 8+3 filename era. Oct 05 11:01:39 mdp: one wishes some people would not overuse the modern filename era Oct 05 11:02:34 i think i already lost miltiple hours on figuring out why i could not open a file in a program, before finding out that the 30 character filename plus the path to my desktop was way too long :D Oct 05 11:02:41 30* Oct 05 11:02:53 av500, your host port support, I assume it uses ehci? Oct 05 11:03:25 nope Oct 05 11:03:52 I see Oct 05 11:04:02 ehci was for e.g. hdds Oct 05 11:04:11 ok Oct 05 11:04:39 and between musb and omap ehci, I cannot say which one we hated more Oct 05 11:04:51 it's tough to choose Oct 05 11:05:13 ehci is broken across 4-5 chip revisions now Oct 05 11:05:42 in one omap5 related conf call we skaed if ehci was good now Oct 05 11:05:45 they all nodded Oct 05 11:05:52 I'm happy we've elected to focus on just one set of brokeness for usb on new parts Oct 05 11:05:54 then we opened the erata and lol Oct 05 11:05:56 'ed Oct 05 11:05:58 at least in my TI Oct 05 11:06:15 heh Oct 05 11:06:22 how does nodding work over the phone? Oct 05 11:06:40 ti has ways Oct 05 11:06:58 dm8tbr: usually somebody shouts out, "I'm nodding my head in agreement!!!" Oct 05 11:07:11 it's never subtle Oct 05 11:07:20 nodding is a suppored "use case" Oct 05 11:07:23 supported Oct 05 11:08:20 gah, this zippy2 schematic will blow my head if i continue trying to understand it >.< Oct 05 11:08:51 this is why it was discontinued Oct 05 11:08:54 it was too dangerous Oct 05 11:08:59 :D Oct 05 11:09:17 why in hell has no one tried and documented how to connect a second SD-card to a panda? >.< Oct 05 11:09:27 bamboo Oct 05 11:09:46 not sure if trolling... Oct 05 11:09:47 :D Oct 05 11:09:58 trolls? here? never ever. Oct 05 11:10:02 all i found was this santho guy: http://pandaboard.org/pbirclogs/index.php?date=2011-06-15#T09:57:08 Oct 05 11:10:17 Phrewfuf|work: if not bamboo then trainer Oct 05 11:10:20 and this av500 dude who just pointed him to the zippy2 schematics and docu Oct 05 11:10:26 trainer has no sd card Oct 05 11:10:34 zippy then Oct 05 11:10:36 whichever Oct 05 11:10:39 i forgot Oct 05 11:10:47 just stack the boards up Oct 05 11:10:51 yeah, zippy or bamboo Oct 05 11:11:04 XorA|gone: thatÄs the plan. Combine ALL the boards Oct 05 11:11:12 and make a new, awesome one Oct 05 11:11:27 throw prpplague some $$$ Oct 05 11:11:45 whom? Oct 05 11:11:48 he likes scooby snacks too Oct 05 11:12:12 prpplague, designed zippy/zippy2/trainer AFAIK Oct 05 11:12:24 av500: gen8? Oct 05 11:12:48 XorA|gone: oh hell no! he is a bad person! have you seen the zippy2 schematics? :D Oct 05 11:13:02 massive brainfuck Oct 05 11:13:12 tasslehoff: 3530 is more gen7 Oct 05 11:13:19 and gen6 Oct 05 11:13:29 gen8 is 3630 Oct 05 11:13:29 ack Oct 05 11:13:46 wear protective gear when going into these kernels Oct 05 11:13:57 Phrewfuf|work: sorry schematic looks pretty simple to me Oct 05 11:14:13 av500, *evil* vendor tree Oct 05 11:14:17 bamboo schematic for SDcard looks fine, but since there is no documentation for actually making the panda speak to it, it is somehow unpractical Oct 05 11:14:27 mdp: mvl4 ftw! Oct 05 11:14:33 lol Oct 05 11:14:39 XorA|gone: i'm looking at bamboo, trainer, Panda and Zippy2 schematics for days now+ Oct 05 11:14:46 btw, I just came across some CDs that have Monterey Linux scribbled on them Oct 05 11:14:55 Phrewfuf|work: I begin to think your EE and linux skills are now too hot Oct 05 11:15:12 Phrewfuf|work: making SD work is as simple as muxing it in then providing platform data Oct 05 11:15:55 the mux names are even on the schematic for you Oct 05 11:16:13 XorA|gone: the things that bug me about the schematics are: how can i know if an sdcard needs 1.8v or 3.3v? Oct 05 11:16:44 the zippy2 has a jumper for this and panda does autodetect (probably with some not-so-el-cheapo chip) Oct 05 11:16:58 I guess most you can buy are happy with 3.3.V Oct 05 11:16:59 av500, MVL4 seems like an eternity ago. Oct 05 11:17:00 Phrewfuf|work: throw in a levelshifter and always drive them with 3.3 Oct 05 11:17:05 yep Oct 05 11:17:34 the schematic has the level shifters on it Oct 05 11:17:41 ok then, 3.3V it is Oct 05 11:17:57 XorA|gone: guessed that, though i've never seen the schematic ;) Oct 05 11:18:04 XorA|gone: yeah, i know, but the SDMMC_VDD can be changed to 1.8 via jumper Oct 05 11:18:22 SD/MMC CARD VOLTAGE SELECT (DEFAULT = 3.3V) Oct 05 11:18:32 yes, use 3.3V Oct 05 11:18:33 but ok, i'll just leave it on 3.3 Oct 05 11:19:05 also i seem to have difficulties finding out how to set the MUX settings for the expansion ports...my google-fu is too weak Oct 05 11:20:38 the Panda_Board_Spec_REVEA1_04.pdf only tells something like: here's how you et display mux, here's how you set camera mux and down there are the possible MUX settings for almost everything Oct 05 11:20:48 but not how to actually set them >.< Oct 05 11:23:15 DT will fix that Oct 05 11:23:24 great! Oct 05 11:24:35 Oct 05 11:28:24 holy crap...is this really that simple? "echo 0x11b > /sys/kernel/debug/omap_mux/mcspi1_cs1" and suddenly pin10 on j3 is a GPIO with a pullup instead of default pulldown? Oct 05 11:28:48 yes, yes it is Oct 05 11:30:12 hmmm...interesting Oct 05 11:30:47 Phrewfuf|work: after careful consideration of how easy that is, that feature is not present in mainline Oct 05 11:30:51 what's with the beagle and trainer...the trainer has an eeprom on it which stores some magic configuration data which makes the beagle set the correct mux settings automagically Oct 05 11:31:32 mdp: you say, i can't use this? Oct 05 11:31:57 anyway: is it possible to make the panda read an eeprom and also set MUX automatically? Oct 05 11:32:19 Phrewfuf|work: I would never try to restrict you. ;) Oct 05 11:32:31 mdp: but it won't work :D Oct 05 11:32:41 there's just no support for it on the latest kernel Oct 05 11:32:51 Phrewfuf|work: anything is possible, but in involves writing code, not downloading some patch and clicking some button. Oct 05 11:34:42 mdp: see...whatever i find is either incomplete or doesn't work anymore >.< Oct 05 11:36:02 ? Oct 05 11:36:13 Phrewfuf|work: indeed Oct 05 11:36:15 but Oct 05 11:36:16 ... Oct 05 11:36:20 its all just linux Oct 05 11:36:31 Hello! I am struggling to understand how to install Xenomai on Angstrom for use in the BeagleBone, I've read the Xenomai instructions etc, but it's still very fuzzy in my head, does anyone know of a good online resource that is targetted at a novice like myself? Oct 05 11:38:08 xenomai again Oct 05 11:38:19 av500: i know...everything is a file...but why is the _all_ the info on how to do stuff on the beagle but none at all for the panda? Oct 05 11:39:02 seniority Oct 05 11:39:17 i think all i need to know is where to find the settings for the gpio mux on the panda Oct 05 11:39:23 panda is just a freshman Oct 05 11:39:27 in which file Oct 05 11:39:29 isn't it all in the TRM? Oct 05 11:39:45 mpdah, come one, it's more than one year old Oct 05 11:39:57 hm, type derp Oct 05 11:40:00 after all, reading pinmuxes out of TRMs has been standard novice knowledge for any embedded person for tens of years. Oct 05 11:40:11 woglinde: yeah, xenomai, I am developing a cheap wireless transmitter and need good latency and speed to the GPIO pins, also need to not have other processes interrupt -- so xenomai :) Oct 05 11:40:30 LetoThe2nd: well then i will learn (again) which muxmodes are available Oct 05 11:40:46 but still no idea on how to actually change them Oct 05 11:41:58 use debugfs, and if debugfs is not there, make it be there ;) Oct 05 11:42:14 this one looks good, but it's french http://www.blaess.fr/christophe/2012/06/04/gpio-pandaboard-et-temps-reel-5-le-multiplexage-des-gpio/ Oct 05 11:42:25 why, why on earth did i have to be french >,< Oct 05 11:42:28 it* Oct 05 11:42:56 heh Oct 05 11:43:00 temps-reel Oct 05 11:44:15 LetoThe2nd: do i really have to use debugfs or is there a more comfortable solution? Oct 05 11:45:00 board file Oct 05 11:45:07 Phrewfuf|work: what is uncomfortable about debugfs? first you complain about not knowing how to do it, then you complain about the way being uncomfortable.. Oct 05 11:45:31 you can declare a 2nd mmc in the board file and do the muxing there Oct 05 11:45:36 av500: we should hand that out as a project. changing boardfiles during runtime. Oct 05 11:45:58 LetoThe2nd: if you runtime multiplex the pins to mmc, it wont detect a new mmc port, will it? Oct 05 11:46:45 av500: hm, indeed. Oct 05 11:46:57 he did not ask about detection, though ;) Oct 05 11:47:08 Phrewfuf|work: since its all there for beagle, why dont see how its done for beagle? Oct 05 11:47:14 he just complained about pingmuxing ;) Oct 05 11:47:20 kernel-wise its very similar Oct 05 11:47:44 dvance to be honest all guys which I saw comming here to get xenomai working did not succeed Oct 05 11:47:52 av500: the problem is, that i tried multiple times to do things on behalf of "it should be similar" Oct 05 11:48:25 and i wasted multiple days on those things Oct 05 11:48:57 if you learned nothing during that time, yes wasted Oct 05 11:49:09 i did not succeed Oct 05 11:49:36 to do what? Oct 05 11:49:45 to make those things work Oct 05 11:49:58 ah, "things" Oct 05 11:50:31 so why learn from something that did not work? Oct 05 11:50:42 woglinde: Not very encouraging :) I know of one guy who's done it -- will try to get in touch through his project website or the message board, but I think I could not find an email so it might be a bit tricky Oct 05 11:50:44 a few of my favorite things Oct 05 11:51:05 Phrewfuf|work: because "learning" where to find things to copy-paste is not real learning. Oct 05 11:51:08 mdp what? Oct 05 11:51:31 Phrewfuf|work: real learning means that you understand things and can fix them yourself. which you obviously failed to learn. Oct 05 11:51:50 and again, 2nd mmc support is there for the beagle already Oct 05 11:52:03 so you have a template to follow Oct 05 11:52:13 LetoThe2nd: yeah, reading files and schematics all over just to go to some "community" to get told that i need to read files and schematics is not really encouraging Oct 05 11:52:41 Anyways, 10x for the help :] Oct 05 11:52:51 yes some "community" is really bad Oct 05 11:52:52 Phrewfuf|work: thats what community and open source means. pointers to possible fitting information. what you expect is called "paid support" Oct 05 11:53:00 Phrewfuf|work: pick a better community Oct 05 11:53:07 * av500 is starting to get pissed off Oct 05 11:53:16 oh no Oct 05 11:53:20 * LetoThe2nd hands av500 a big handkerchief Oct 05 11:53:20 * woglinde goes hiding Oct 05 11:53:26 * mru suggests rpi Oct 05 11:53:48 I'll come back when I make it work and maybe even post a tutorial somewhere :D Oct 05 11:54:08 dvance good luck Oct 05 11:54:29 Thanks! Oct 05 11:54:35 LetoThe2nd: paid support is when i pay peopple to do stuff for me without me even thinking about it Oct 05 11:55:06 * mru was not paid Oct 05 11:55:12 Phrewfuf|work: there are varying degress of that, depending on the complexity of the given task. Oct 05 11:55:27 in this case i _want_ to do stuff on my own and ask for a bit of help but all i get is "it may work in some cases if you're lucky with the stuff over there" Oct 05 11:55:49 Phrewfuf|work: so please tell me. "what do you expect?" Oct 05 11:55:50 and? Oct 05 11:55:55 life is hard Oct 05 11:55:57 always Oct 05 11:56:07 and the neweset fanciest shit might not work Oct 05 11:56:15 and then i will end up having to rewrite the kernel for the panda only to make a second SDcard run with it Oct 05 11:56:40 Phrewfuf|work: now please stop bi***ing around and tell me, honestly. what did you expect? Oct 05 11:57:19 working product on a silver platter Oct 05 11:57:26 mru: hell no Oct 05 11:57:33 Phrewfuf|work: so what did you expect? Oct 05 11:57:49 well let me point it this way: Oct 05 11:57:53 Phrewfuf|work: i've asked you three times now. if it is so easy what you want, why not just tell us? Oct 05 11:58:09 i found information to how set the mux settings Oct 05 11:58:17 Phrewfuf|work: don't put it any other way, and no lame excuses. what did you expect? Oct 05 11:58:30 i ask if it's correct and all i get as an answer is "it won't work" Oct 05 11:58:40 ? Oct 05 11:58:51 you got the an right answer Oct 05 11:58:57 its not in mainline kernel Oct 05 11:59:41 obviously any meaningful conversation is impossible already, as the asker has drowned in self-pity. too bad. who cares, then. Oct 05 11:59:45 woglinde: well what _is_ in mainline kernel that i can use to set the mux _and_ sdcard detection and mounting on the panda? Oct 05 12:00:19 LetoThe2nd: no need to get personal here Oct 05 12:00:23 Phrewfuf|work: answer: nothing, because it is board file dependent. Oct 05 12:00:31 Phrewfuf|work: that is your answer. Oct 05 12:00:37 Phrewfuf|work: that is waht you wanted. Oct 05 12:00:52 Phrewfuf|work: sorry if you do not like the actual truth. Oct 05 12:01:05 LetoThe2nd: your assumptions are not the truth Oct 05 12:01:28 * LetoThe2nd does not assume anything. Oct 05 12:01:33 *plonk. Oct 05 12:01:44 this is an assumption: "as the asker has drowned in self-pity." Oct 05 12:01:57 and atop of that it is offensive Oct 05 12:02:35 self-pity *and* entitlement Oct 05 12:03:23 anyway...board file...sounds interesting, tell me more about it? Oct 05 12:03:36 google linux devicetree Oct 05 12:05:21 /arch/arm/mach-omap2/board-*.c :) Oct 05 12:09:54 woglinde: do i understand this the right way: there is a bunch of files where all the hardware is described, right? Oct 05 12:10:15 * LetoThe2nd runs off for the weekend. seeya next time, folks. Oct 05 12:10:26 enjoy Oct 05 12:13:24 Phrewfuf yes with dt there are generics function for a soc/platform to setup all the stuff Oct 05 12:14:08 and then you have device descrition file for your board where all the specific adresses and gpios and so on are manifest Oct 05 12:14:28 but thats not the real solution for your problem Oct 05 12:17:08 Phrewfuf|work: dt is work in progress, the panda is not there yet Oct 05 12:17:21 for adding a 2nd mmc, see what was done for zippy on beagle and do that for panda Oct 05 12:17:31 * av500 is repeating himself a lot Oct 05 12:50:24 jonmasters: any luck so far? Oct 05 13:34:54 fresh testresults: after loading musb_hdrc with use_dma=n, heavy networking no longer seems to kill my board. Oct 05 13:37:39 speed is 50% of what it was, but that's a price I'll gladly pay if this works :) Oct 05 13:41:55 ok Oct 05 14:07:27 bradfa: coincidentally, I saw a request for what the pinmux tool should generate for DT today...just fyi Oct 05 14:24:02 hm poor musb Oct 05 14:24:13 but good to know that dma is broken Oct 05 14:42:19 what happens to the cpu use without dma? Oct 05 14:42:46 one guess! Oct 05 14:42:59 thurbad we need to ask tasslehof later Oct 05 14:43:02 but yes Oct 05 14:43:02 I know it's not good, just curious how bad Oct 05 14:43:18 thurbad but its bad usb crashes your network Oct 05 14:43:23 too Oct 05 14:43:28 true Oct 05 14:43:29 depends on your application Oct 05 14:44:08 maybee he can live with the higher cpu usage Oct 05 14:48:19 mdp, public info? Oct 05 14:48:37 bradfa, on what? Oct 05 14:48:45 mdp, pinmux tool request for dt Oct 05 14:48:54 sorry, was a while ago you mentioned Oct 05 14:48:56 I was afk Oct 05 14:49:00 no, just about the TI windows tool Oct 05 14:49:10 will it be outputting dt data? Oct 05 14:49:17 o.O Oct 05 14:49:35 woglinde, windows only tools that output files that can only be used on linux, TI has it all figured out! Oct 05 14:49:51 also, it can't print Oct 05 14:50:00 hit print button Oct 05 14:50:07 no such thing Oct 05 14:50:30 we're going paperless! Oct 05 14:50:43 * bradfa has had a bad morning Oct 05 14:50:55 * bradfa has said swear words about 100x as often as an average morning Oct 05 14:50:59 yes thats now our problem Oct 05 14:51:11 bradfa, well, I'm just saying that the people that implement that tool asked the linux people for the DT specs Oct 05 14:51:29 mdp, that's good Oct 05 14:51:33 thanks for fyi :) Oct 05 14:51:43 mdp it exist really? Oct 05 14:51:47 I just found it interesting given your work Oct 05 14:51:52 I thought it was a joke Oct 05 14:51:58 woglinde: the pinmux tool? Oct 05 14:52:03 well, my work so far is just making text files with pin mux for bone in them Oct 05 14:52:08 dt is a whole nother ball of wax Oct 05 14:52:25 which I will wrap my head around probably soon Oct 05 14:52:28 http://processors.wiki.ti.com/index.php/Pin_Mux_Utility_for_ARM_MPU_Processors Oct 05 14:52:58 written in .net? Oct 05 14:53:01 I'll let others decide if it's a joke or not Oct 05 14:53:14 I dunno Oct 05 14:54:13 you need wine or what? Oct 05 14:54:32 I still use the .net l138/da850 tool under mono to recover those boards Oct 05 14:54:46 woglinde: nfc Oct 05 14:56:12 nfc? Oct 05 14:56:32 I have no clue, never had a use for the tool Oct 05 14:57:45 I respect the concept of trying to help deal with all the conflicts visually though. Oct 05 14:58:14 sure but they could have made a qt application Oct 05 14:59:09 bradfa, btw, I often have fritzing open with my beaglebone model so I can hover over the pin header and see the muxes when doing a nest of wires project. Oct 05 14:59:27 woglinde: they might not know qt Oct 05 14:59:44 they would love it Oct 05 14:59:46 anyway Oct 05 14:59:55 depends on mangement too Oct 05 14:59:55 mdp, I'll give fritzing a try, haven't used it before Oct 05 15:00:06 to give users comfort Oct 05 15:00:21 woglinde, mdp, java gets no cross platform love? :) Oct 05 15:00:55 for all I know, they may be switching to something cross platform Oct 05 15:01:19 bradfa sure Oct 05 15:01:59 bradfa, http://engineersofthecorn.blogspot.com/2012/06/fritzing-part-for-beaglebone.html Oct 05 15:04:54 or eagle! Oct 05 15:06:51 but selling/providing a linux sdk with a windows tool Oct 05 15:07:49 I suspect nobody of importance has complained about it yet Oct 05 15:07:54 * bradfa busy debugging, bbl Oct 05 15:08:51 mdp yes maybee, maybee they tought he we ship eclipse with java so the customers will have windows too Oct 05 15:18:19 hi ! In the kerne from https://github.com/beagleboard/kernel.git there are no board specific config. I shall use omap2plus_defconfig ? Oct 05 15:20:25 bizulk, see ./configs/ in the tree Oct 05 15:21:42 mdp: I only have /patches/configs/empty Oct 05 15:22:00 checkout a branch Oct 05 15:22:30 oups sorru Oct 05 15:24:28 when calling 'git branch' I only see * beaglebone-3.2 Oct 05 15:24:28 master Oct 05 15:24:49 But there are other listed on the site (noob) Oct 05 15:25:38 git branch -a Oct 05 15:25:49 checkout a tracking branch Oct 05 15:25:52 then go to configs Oct 05 15:27:05 mdp: sorry If I try (git checkout origin/beagleboard-3.6 -b beagleboard-3.6) : error: Your local changes to the following files would be overwritten by checkout: Oct 05 15:27:05 patches/3.2.1/0001-MAINTAINERS-stable-Update-address.patch Oct 05 15:28:26 git stash Oct 05 15:28:42 tell you what, read some of the good git workflow tutorials out there Oct 05 15:28:57 it's all documented there Oct 05 15:29:07 ok ok sorry to disturb Oct 05 15:29:23 np, just helps to have the basics down first Oct 05 15:29:36 the key here for this tree is that nothing is on the master branch Oct 05 15:30:25 the owner only has stuff populated on various branches as it's using git in a way somewhat differently than a typical project that is concerned about merging to master over time Oct 05 15:30:40 that's not wrong...just "different" and useful in this case Oct 05 15:32:42 "othing is on the master branch" => that why the directory is empty ? I tought that master should keep common code. Oct 05 15:33:08 bizulk the tool is not reasonable for the misuse of the owner Oct 05 15:33:20 I just explained that this repo is not implemented that way Oct 05 15:33:33 you need to find out whats in master and whats in the other branches Oct 05 15:33:44 I promise you, if you follow my exact steps, you'll get at the configs you want Oct 05 15:37:41 mdp: Actually I did : git staches -> "git checkout origin/beagleboard-3.6 -b beagleboard-3.6" -> I get the configs for beagleboard -> this WE read some GIT tutos Oct 05 15:38:17 cool, nice job Oct 05 16:08:26 mdp hm the pinmux tool generates u-boot header so the kernel do not touches pinmux at all? Oct 05 16:09:04 that's a possible use for it Oct 05 16:18:22 well, with linux-stable 3.2.30 and the omapbeagleboard 3.6.0+ I can't manage to get the dsp decode working (using gst-dsp) : when launching video : BUG: scheduling while atomic: queue1:src/92/0x0000008e Oct 05 16:18:52 the dsp-test passes indeed Oct 05 16:18:56 av500 dsp customers Oct 05 16:19:03 :) Oct 05 16:19:14 customers -> victims Oct 05 16:19:32 hi djlewis Oct 05 16:19:51 that stuff seems to be very sensitive Oct 05 16:20:20 gm guys :) Oct 05 16:37:43 <_av500_> woglinde: hm? Oct 05 16:37:56 <_av500_> gst-dsp is for felipec Oct 05 16:38:12 <_av500_> he is the one not supproting that Oct 05 16:38:20 <_av500_> im not supporting gst-ti Oct 05 16:45:05 supporting or not supporting? Oct 05 16:45:16 anyway I would use gst-ti Oct 05 16:57:30 _av500_: he is the one not supporting that => ? Oct 05 16:57:49 Well I may also try gst-ti Oct 05 17:02:33 <_av500_> bizulk: sorry, no time atm Oct 05 17:21:26 bradfa: I was suggesting your beaglebone pinmux text list to somebody and something occurred to me Oct 05 17:21:54 bradfa, since the "signal name" pulled from the SRM is a completely fabricated concept..I would suggest removing that column Oct 05 17:22:32 it's meaningless and confuses people Oct 05 17:24:52 <_av500_> pinmux names should be 8.3 Oct 05 17:25:16 I am taking that recommendation to "those above" Oct 05 17:25:45 denix, infinite thanks to you for that phrase, btw Oct 05 17:25:56 <_av500_> also they should include soc name/number Oct 05 17:26:05 <_av500_> GPIO3530.A15 Oct 05 17:26:14 <_av500_> and be rueused on new socs Oct 05 17:26:20 agreed Oct 05 17:27:00 mdp: should I charge royalties for it use? Oct 05 17:27:14 s/it/its Oct 05 17:27:31 mdp, patch? :) Oct 05 17:27:48 patches are overrated! Oct 05 17:28:13 just paste the modified code fragments in a word doc Oct 05 17:28:19 and upload it to the shared ftp Oct 05 17:28:41 <_av500_> fax it Oct 05 17:28:58 <_av500_> to my dropbox Oct 05 17:29:02 fax is for sending money only! Oct 05 17:29:30 <_av500_> ok, then send a boy Oct 05 17:29:37 bye all Oct 05 17:33:20 boy via fax? Oct 05 17:35:08 bradfa, I will send you a .ppt Oct 05 17:35:19 sprz011.ppt Oct 05 17:35:47 I think koen gets patches in that format from some of his suppliers Oct 05 17:37:24 mdp, can you print it and take a picture of it on a wood table? Oct 05 17:38:30 I can Oct 05 17:38:42 I'm willing to do anything the maintainer demands! Oct 05 17:38:52 mdp, the NAME column is useful though, since some things on the headers aren't pins on the am335x Oct 05 17:38:53 just to be a part of your project Oct 05 17:38:54 like ground Oct 05 17:38:55 and power Oct 05 17:38:58 bradfa, yes Oct 05 17:39:13 gerald seems to have horribly abused things in the SRM Oct 05 17:39:30 NAME == what ever felt like something people might want to use that pin for Oct 05 17:39:34 rather than MODE 0 Oct 05 17:39:37 with those board specific pins being labeled under "signal name" which then confuses everything Oct 05 17:39:51 branch coming, one sec Oct 05 17:40:30 I would recommend "NAME" with the mode 0 string being reused there for all soc pins Oct 05 17:40:47 since that's what you would actually look up in the TRM CM regs Oct 05 17:41:20 signal just is completely wrong Oct 05 17:43:19 bradfa, your text file is so wonderful, I plan to dedicate some dead trees to it and enshrine it next to my despair.com posters Oct 05 17:43:24 :) Oct 05 17:43:38 mdp, what about blank NAME column for things connected to the SoC? Oct 05 17:43:47 but leave things like GND and PWR. Oct 05 17:43:59 link coming shortly, crafting nifty git commit message Oct 05 17:44:33 yes, except I think you *could* help people by duplicating the same thing in the mode 0 column into "NAME" Oct 05 17:44:53 since from a TRM-POV...the mode 0 name is the PIN NAME Oct 05 17:44:59 mdp, are you new here, we don't help people on Fridays Oct 05 17:45:14 that's how you look up the register in the TRM Oct 05 17:45:33 bradfa, I'm living on the edge today Oct 05 17:47:08 ok, I suck at git Oct 05 17:47:55 I'd say "join the club"...but if there were a club I couldn't mention it Oct 05 17:48:25 git push --force to the rescue! Oct 05 17:49:38 https://github.com/bradfa/beaglebone_pinmux_tables/tree/mdp-request Oct 05 17:50:39 mdp, I do see your point though, about mode 0 going in the NAME column Oct 05 17:50:55 that's how it's named in Linux and am335x docs Oct 05 17:51:01 exactly Oct 05 17:51:10 k, copy paste time Oct 05 17:54:16 mdp-request branch up to date Oct 05 17:54:18 yes, you want people to look at NAME and be able to search on "gpmc_ad0" in TRM Table 9-10 to get the right register offset Oct 05 17:54:23 for example Oct 05 17:54:29 register offsets! Oct 05 17:54:54 mdp, you know what would be really handy, speaking of register offsets, is if TI would put the damn base address on each register table! Oct 05 17:55:16 not all TI data sheets do this Oct 05 17:55:48 ;) Oct 05 17:55:49 it's like TI gave a bunch of monkeys coke and asked them to write tech documents Oct 05 17:56:08 oh, real fun one to look at, CC430 SPI slave picture in data sheet Oct 05 17:56:38 picture and comment don't match, one says 4 wire, other says 5 wire. No where does the word "chip select" show up Oct 05 17:56:39 bradfa, if you keep talking smack, I'll ask them to explode the TRM into separate 8+3 named files for each chapter...like am180x and the dsp parts ;) Oct 05 17:56:43 keep it up, pal Oct 05 17:56:47 :0 Oct 05 17:56:58 sprux893.ppt! Oct 05 17:57:23 (that might actually be a valid TI documentation file) Oct 05 17:57:56 the exploded TRM method has been a major pain point for me on am180x...what a pita to organize those pdfs Oct 05 17:58:03 I need to justify my columns in the text... Oct 05 17:58:15 Microchip provides some of their docs, 1 chapter per PDF. Oct 05 17:58:23 they rate limit you if you download too many at once Oct 05 17:58:43 I have this entire directory hierarchy so I can quickly pull up the spi chapter..I can't remember the 8 character magic name for each chapter ;) Oct 05 17:58:47 too dumb Oct 05 17:58:56 bradfa: limit by what? Oct 05 17:59:06 I have to prepend real names to my TI docs else I have no idea what they are fo Oct 05 17:59:13 won't let you download for time period Oct 05 17:59:19 at least last I tried Oct 05 17:59:26 but that was 1+ year ago Oct 05 17:59:30 glad to not deal with them any more Oct 05 17:59:43 bradfa, I quit doing that way because things were updated and I'd do it again.. Oct 05 17:59:52 yeah, that's a tradeoff Oct 05 17:59:53 just went to the directory thing Oct 05 18:00:24 ok, merging mdp-request branch into master, probably with a squash Oct 05 18:00:29 so far no objections! Oct 05 18:00:34 muhahahaha Oct 05 18:00:43 bradfa, I will ack! Oct 05 18:01:21 mdp, which email you want in the 'acked by' line? Oct 05 18:01:25 mdp@#beagle Oct 05 18:01:35 ? Oct 05 18:02:35 you should take all the credit here ;) Oct 05 18:02:47 ha, me take credit... Oct 05 18:02:50 all I did was chatter some comments ;) Oct 05 18:03:14 mdp, best part is my list is more accurate than bone SRM Oct 05 18:03:40 shhhh Oct 05 18:03:46 k k k Oct 05 18:08:24 mdp: blame is socialized, credit is personalized :) Oct 05 18:11:15 any recommendations on a good book that Packt Publishing has printed? I get a free one Oct 05 18:11:19 no idea what I'd want Oct 05 18:20:02 * bradfa needs a nap... Oct 05 18:20:18 haha Oct 05 18:37:42 Hello all. Just dusted out my bone, and it won't show up as a usbserial device in /dev/. I'm on Lion, and it shows up a an FTDI device in system profiler. Any suggestions? Oct 05 18:38:40 you have the driver loaded? Oct 05 18:40:03 Installed the FTDI driver from the micro SD card. Oct 05 18:50:13 sorry, battery died. Oct 05 18:50:39 I jsut plugged it into my linux machine, and it automagically mounted the partition. Oct 05 18:52:01 and I've got /dev/beaglebone, so it must just be an issue on my macBookPro. Oct 05 19:23:52 koen: are there daily images of github.com/beagleboard/kernel available somewhere? Oct 05 19:26:00 only if you build them Oct 05 19:26:15 and there aren't daily change to all branches Oct 05 19:27:54 koen: ok, no problem Oct 05 20:47:42 * mranostay bangs head Oct 05 20:48:15 against wall? Oct 05 20:49:03 3 do_fetch threads running. Oct 05 20:49:17 god this office has the worst internet connection ever Oct 05 20:50:27 make someone fix it? Oct 05 20:54:50 georgem: aren't we a dreamer? :) Oct 05 20:55:43 mranostay: maybe you need to work on being more annoying :) Oct 05 20:56:49 oh you clearly don't know me :) Oct 05 22:10:50 \ Oct 05 22:11:06 // Oct 05 22:11:33 \\\ Oct 05 22:11:40 /// Oct 05 22:13:10 joelangel: hey, how's life after HDMI? Oct 05 22:17:18 ds2, he's got life with mainline/DT now ;) Oct 05 22:32:03 god damn it... AWS has way too much IP space Oct 05 23:07:43 * mranostay curses Oct 05 23:18:28 * djlewis leaves for home :) Oct 06 02:39:53 is there a public repository for the code that gets grabbed with the "opkg upgrade"? Oct 06 02:41:26 you mean a feed? Oct 06 02:44:57 i don't know actually. Oct 06 02:45:07 i'm thinking of svn or git repository Oct 06 02:45:26 what's the definition of a "feed"? **** ENDING LOGGING AT Sat Oct 06 03:00:01 2012