**** BEGIN LOGGING AT Mon Nov 12 02:59:58 2012 Nov 12 04:29:56 anybody home? Nov 12 05:23:33 i'm back Nov 12 05:40:28 were you gone? Nov 12 05:41:19 well i slept past my flight to start the day Nov 12 05:42:16 So, anyone here know about the tps65217 backlight driver? Nov 12 05:42:42 you mean the wled? Nov 12 05:42:51 Yeah Nov 12 05:43:55 what is your question? Nov 12 05:44:13 I'm trying to figure out how to get the driver for it so I can drive my backlight Nov 12 05:44:27 there was a patch floating around i used Nov 12 05:44:28 And I've found some information. Nov 12 05:44:48 Really? I've been looking. Nov 12 05:45:07 I found the driver in the kernel.org kernel. Nov 12 05:45:11 I've got a pretty basic question: I did bitbake virtual/kernel -c config; make menuconfig; copied .config to defconfig then did bitbake virtual/kernel. It says all tasks ran successfully but I don't see the .ko files for some of the modules I opted to build. Am I missing a step? Nov 12 05:46:12 But I can't find the driver patch for angstrom. Nov 12 05:46:27 Do you know where I can find the patch? Nov 12 05:46:47 Been googling for it all day. Nov 12 05:57:08 df___: you can check to see if a patch is overwritting your defconfig settings by doing a bitbake virtual/kernel -c devshell and then attaching to the devshell and running a make menuconfig Nov 12 05:57:48 df___: if so you probably want to make your config settings a kernel patch Nov 12 05:58:34 Nevermind, just found the patch you were talking about. Going to try it now. Nov 12 05:58:42 df___: how to make a kernel patch: http://communistcode.co.uk/blog/blogPost.php?blogPostID=3 Nov 12 06:20:59 JViz: Thanks, I'll give that a shot (sorry was mindlessly watching the bits bake) Nov 12 06:24:55 Another somewhat unrelated question: Can I use spi channel 0 on the beaglebone? Seemed like I'd read some stuff that 0 wasn't available (I think only 3 is available via spidev, and 2 can be enabled with a custom kernel) is that right? Nov 12 06:33:49 did you check the SRM on this? Nov 12 06:34:48 dm8tbr: trolling! Nov 12 06:41:16 df___: are you using any capes? Nov 12 06:41:39 df___: if so, both SPIs may or may not be occupied Nov 12 06:42:24 df___: i'm using an LCD7 and very unsuccessfully trying to get spidev2 to work. Nov 12 06:42:26 mranostay: a troll a day, keeps the doctor away! Nov 12 06:43:04 df___: without any capes, and on an older kernel, i had no problems getting spidev1 working Nov 12 06:46:56 alright i had a beer left in my fridge Nov 12 06:47:22 * mranostay tries to find something to do with this pandaboard Nov 12 06:48:21 mranostay: use it as a doorstop Nov 12 06:48:40 why'd you update the kernel if the old one worked? Nov 12 06:49:24 the LCD7 didn't work with the old kernel Nov 12 06:49:42 .. that's a good reason Nov 12 06:49:54 yeah Nov 12 06:49:55 dm8tbr: so Munich opened my bag and asked what i had in the box Nov 12 06:50:09 although sometimes you can backport drivers more successfully than updating the entire kernel Nov 12 06:50:18 mranostay: oooh :) Nov 12 06:50:41 probably because that was the departure point for US Nov 12 06:51:01 i just need to get a working spidev2.0 in my /dev folder Nov 12 06:51:35 mranostay: hehe, i was carrying two pandaboard cardboxes, one containing a panda, the other two PSUs. after xraying, the pointed to the box with the PSUs and told me to get it out the next time. Nov 12 06:53:56 i physically disconnected the button on gpio3_16, i edited the board-am335x.c to populate the spidev2 pinmux, i disabled the keys from initing, i changed the pin modes, and i have SPIDEV=y in my kernel config, i don't know what i could be missing Nov 12 06:58:39 the SFO line for re-entering the country was quite long Nov 12 07:00:47 mranostay: it's very long for non-residents and seemed always short for residents... Nov 12 07:01:14 dm8tbr: the inverse this time Nov 12 07:01:22 fascinating Nov 12 07:01:34 dm8tbr: sarcasm? :) Nov 12 07:01:41 did you notice how they are jamming all sorts of frequencies down there? Nov 12 07:02:08 probably since you aren't suppose to be using your phone until exiting the Customs area Nov 12 07:02:38 still, the first time I saw active jamming/interference Nov 12 07:02:42 * mranostay wonders who's email he has in iphone notes... Nov 12 07:03:00 anybody have any ideas as to what i'm missing? Nov 12 07:03:16 my dmesg comes up clean Nov 12 07:03:31 as far as i can tell Nov 12 07:11:08 do i have to set a mode in the board file for spidev to work? Nov 12 07:16:37 hello Nov 12 07:17:04 hello Nov 12 07:17:12 I have a problem Nov 12 07:17:36 why "sudo apt-get install" not work in beaglebone? Nov 12 07:17:46 thanks in advance Nov 12 07:17:59 frodo: are you running ubuntu on your beaglebone, or debian? Nov 12 07:18:36 default system Nov 12 07:18:57 frodo: which is angstrom. then, go figure. Nov 12 07:19:14 frodo: you want to read up on opkg, young padawan. Nov 12 07:19:14 I use eclipse in ubuntu connect to beaglebone Nov 12 07:19:58 thank you yoda Nov 12 07:20:51 Why "reboot" command not work? Nov 12 07:21:53 shutdown -r now Nov 12 07:22:42 how to do that, I am a rookie, sorry Nov 12 07:24:47 -sh: sudo command not found Nov 12 07:25:20 angstrom is not debian based Nov 12 07:25:24 frodo: do you see the little "root" at the left end of your command line? think about what that means for using sudo... Nov 12 07:25:52 frodo: and please, really google for some angstrom first steps or such before asking more meaningless debian/ubuntu things. Nov 12 07:26:33 ok Nov 12 07:26:38 <_av500_> LetoThe2nd: Do you sell hubcaps for a '72 Pinto hatchback? Nov 12 07:27:09 sudo sell --target _av500_ --object $WHATEVERHEWANTS Nov 12 07:27:46 frodo: there are ubuntu images for the beagles Nov 12 07:28:16 frodo: it just doesn't come with them by default Nov 12 07:28:49 ok, I see Nov 12 07:29:29 frodo: mainly because ubuntu takes up a lot of space, and angstrom is geared more towards embedded development Nov 12 07:30:46 angstrom assumes you're going to be making an application based system, and not a full computer, so it installs the bare minimum by default, the whole install is like 14MB Nov 12 07:32:20 what timezone is koen in? Nov 12 07:33:04 i have a feeling he's the only one that can answer my questions Nov 12 07:33:48 perfect Nov 12 07:34:09 What is the default system in bb? Nov 12 07:34:15 bb? Nov 12 07:35:22 beaglebone.. Nov 12 07:35:43 angstrom? what do you mean by system? Nov 12 07:35:57 OS? Nov 12 07:36:15 Did you mean install angstrom in beaglebone? Nov 12 07:36:53 i mean that angstrom comes by default on the beagle. when you buy it, you the sdcard you receive has angstrom on it Nov 12 07:37:18 got it Nov 12 07:37:38 I am very freshman on linux and ubuntu Nov 12 07:37:52 and beaglebone Nov 12 07:38:42 what are you trying to use the beaglebone for? if you don't mind me asking Nov 12 07:38:52 But why "sudo" not working on beaglebone, is that a ubuntu command? Nov 12 07:39:04 make a robot Nov 12 07:39:36 it's an extended command, sudo isn't needed by default, because the only user, by default is root Nov 12 07:39:54 I have some experience on MSP430 Nov 12 07:41:04 ok, thank you. Shoud I need switch user? Nov 12 07:41:20 cloud9 is an attempt to make the beaglebone function more like a microcontroller Nov 12 07:41:52 that's cool Nov 12 07:42:14 * LetoThe2nd bites keyboard and tries not to cite linus torvalds. Nov 12 07:42:16 if you're robot is a hobby and you're the only one that's going to be accessing it, then no, i wouldn't worry about using root Nov 12 07:42:47 though, it's usually a good idea to secure your environment Nov 12 07:44:15 I mean how can I use "sudo" on beaglebone, can not figure out what is the relationship with root Nov 12 07:44:40 normally, on say, ubuntu, you are running as a normal user Nov 12 07:44:45 frodo: please, if you do not even know that, pleasepleaseplease gogole it! Nov 12 07:45:03 ok Nov 12 07:45:06 sudo makes it so that when you run your command, it gets it's permissions elevated to root status Nov 12 07:45:21 on the beagle, by default, you already have root status Nov 12 07:45:24 so you don't need it Nov 12 07:45:27 so they didn't install it Nov 12 07:45:43 got you Nov 12 07:45:45 frodo: this is so utterly essential that it even has an extensive wikipedia article: http://en.wikipedia.org/wiki/Sudo Nov 12 07:45:54 thank you Nov 12 07:46:53 if I want install gdbserver, may I input "apt-get install gdbserver"? Nov 12 07:47:14 opkg install gdbserver Nov 12 07:47:16 frodo: did you even listen to the two very first lines i told you? Nov 12 07:47:21 but first you need to do an Nov 12 07:47:23 opkg update Nov 12 07:47:40 already did that. Nov 12 07:48:07 opkg is angstrom's version of apt Nov 12 07:48:32 * LetoThe2nd runs.... erm, really flees now. Nov 12 07:48:40 if you want to find packages Nov 12 07:48:48 opkg list|grep blah Nov 12 07:49:14 so, this is a angstrom's command? Nov 12 07:49:20 yes Nov 12 07:49:36 cool , thank you JViz Nov 12 07:49:38 LetoThe2nd: do you know anything about spidev? Nov 12 07:50:19 JViz: partially yes. but not about pinmuxing on the beagle, which i think is your problem. Nov 12 07:50:37 JViz: if i knew something valuable for you, i already would have told you, no worries ;) Nov 12 07:50:48 ah, thank you Nov 12 07:51:04 you think my problem is with pinmuxing? Nov 12 07:51:16 JViz: didn't you say that? Nov 12 07:51:20 can pinmuxing stop a kernel module from loading? Nov 12 07:51:39 er, bad pinmuxing, rather Nov 12 07:52:09 JViz: not the pinmuxing itself, but the claiming of the pins should, IMHO. Nov 12 07:52:30 well, i think i have the pins claimed Nov 12 07:52:42 have you seen this tutorial http://www.youtube.com/watch?v=vFv_-ykLppo&feature=relmfu? Nov 12 07:52:53 <_av500_> download GDBSERVER.ZIP, unpack and click on SETUP.EXE Nov 12 07:52:55 JViz: yeah, and if something else has claimed them earlier, loading will fail. Nov 12 07:53:16 about 36mins , he use the sudo in beaglebone Nov 12 07:53:31 well, the way spidev2 works is only if the cape init function thinks the pins are free Nov 12 07:53:50 <_av500_> 36min? I have only a patience for 30s on youtube Nov 12 07:53:53 so i only edited the one part that i thought would claim them Nov 12 07:54:00 and it shows up in dmesg Nov 12 07:54:32 [ 1.325256] BeagleBone cape: exporting SPI pins as spidev Nov 12 07:54:50 it *shouldn't* get that far if something else is claiming them, theoretically Nov 12 07:55:01 JViz: agreed Nov 12 07:55:25 i watched that tutorial once, i should probably go back and watch it again though Nov 12 07:55:35 JViz: can you try hand-muxing it through debugfs? Nov 12 07:55:52 wouldn't that require exporting the pins as gpio? Nov 12 07:56:14 don't think so. Nov 12 07:56:26 even if i did that, if i got the settings right, would spidev2 just magically show up? Nov 12 07:57:45 possible - like i said, its not my field of expertise, just guessing. Nov 12 07:58:33 if i pastebin my kernel patch, could you take a look at it and see if anything jumps out at you? Nov 12 07:58:38 it's not very big Nov 12 07:59:11 i can, but thats no guarantee that it actually is correct. dump it to a pastebin and post the link here. Nov 12 08:00:18 http://pastebin.com/1eMpUMzA Nov 12 08:03:45 i'm not using spi flash, so i turned the pullups off Nov 12 08:04:00 i'm just connecting directly to a micro Nov 12 08:08:20 JViz: one a first glance it looks ok... have you also tried with the old 12M speed? Nov 12 08:09:19 no, i haven't Nov 12 08:10:19 wouldn't that just make the communications fail if it was talking too fast? Nov 12 08:10:43 why would that stop spidev from populating? Nov 12 08:15:08 i know why i haven't watched that video all the way through, i'm using qt-creator with qt4-embedded 4.8 Nov 12 08:17:00 blah Nov 12 08:17:28 i think i Nov 12 08:19:12 hi Nov 12 08:19:13 does CONFIG_SPI_MASTER conflict with CONFIG_SPI_SPIDEV? Nov 12 08:19:49 hmmm Nov 12 08:20:08 no, i just changed it from m to y, so i can't seee that as being the problem Nov 12 08:20:49 i had Beaglboard -xm Expansion V2 from chipsee. and how to let touchscreen working at kernel 2.6.37? Nov 12 08:23:40 anyone know? Nov 12 08:24:11 fox1213: are you getting any input? what are you trying to do that isn't working? Nov 12 08:24:55 what happens when you hexdump /dev/input/event0? Nov 12 08:24:58 JViz: sry, i'm out (no more ideas + work here) Nov 12 08:25:25 LetoThe2nd: thank you! Nov 12 08:26:51 i refer the 2.6.32 board-omap3beagle.c to modify 2.6.37 same file, but when i reboot the kernel, it shows cant detected device Nov 12 08:27:53 was it working on 2.6.32? Nov 12 08:28:18 2.6.32 is working correct Nov 12 08:29:35 i don't have either of those versions to look at, but there are a few different files in that directory that could be effecting your screen Nov 12 08:31:07 after i find this issue on the net, i think maybe SPI initilize has problem. but i dont know which file need to change. Nov 12 08:32:35 like this. static void __init omap3_beagle_config_mcspi3_mux(void) { // NOTE: Clock pins need to be in input mode omap_mux_init_signal("sdmmc2_clk.mcspi3_clk", OMAP_PIN_INPUT); omap_mux_init_signal("sdmmc2_dat3.mcspi3_cs0", OMAP_PIN_OUTPUT); omap_mux_init_signal("sdmmc2_dat2.mcspi3_cs1", OMAP_PIN_OUTPUT); omap_mux_init_signal("sdmmc2_cmd.mcspi3_simo", OMAP_PIN_OUTPUT); omap_mux_init_signal("sdmmc2_dat0.mcspi3_so Nov 12 08:33:03 spi initialization is in board-am335xevm.c Nov 12 08:33:20 oh wait Nov 12 08:33:24 i'm thinking of the beaglebone Nov 12 08:33:38 you're on a beagleboard-xm right? Nov 12 08:33:44 yes Nov 12 08:35:25 my kernel source from TI_android_gingerbread_234 Nov 12 08:38:19 on my lcd7, all the spi pins use a pullup Nov 12 08:38:23 I couldn’t find the full BeagleBone pin information in a machine readable format anywhere, so i did it myself: https://github.com/ion1/beagle_bone_pins Nov 12 08:40:08 ion: thank you Nov 12 08:43:02 no one porting TI gingerbread 2_3_4 to beagleboard+Expansion Board v2?? Nov 12 08:44:00 fox1213: if you don't mind me asking, why do you need that? Nov 12 08:46:11 i cant find android 2.3 source , but TI 2.3.4 Nov 12 08:57:24 Or anyone can tell me where can download android 2.3 source with kernel 2.6.32? Nov 12 09:01:43 fox1213: you don't want to use 4.0? Nov 12 09:02:30 er, i think i understand your question now Nov 12 09:02:32 lol Nov 12 09:05:14 fox1213: if you want android 2.3 source, you have to get it out of google's git repo Nov 12 09:05:51 the TI 2.3.4 version is the one tailored for beagle Nov 12 09:06:44 you're not going to find any non-mainlined board support packages in google's sources Nov 12 09:07:23 i've never looked at google's code, so i don't know what they support internally, but i seriously doubt you're going to find anything applicable to the beagle Nov 12 09:10:11 JViz i think upgrade to 4.0 will meet same problem. Nov 12 09:11:23 i still dont know how to let TP working correct. Nov 12 09:11:36 yeah, i just didn't know why you were looking for android, at first, and then i realized you were trying to fix that touchscreen problem you were talking about earlier Nov 12 09:12:05 is it showing up as an input in /dev/input? Nov 12 09:12:11 er Nov 12 09:12:13 event Nov 12 09:12:21 is it showing up as an event in /dev/input? Nov 12 09:12:26 oh Nov 12 09:12:28 damnit Nov 12 09:12:36 i forgot what we were talking about earlier Nov 12 09:13:49 you were saying that you were getting a message saying it's not detecting Nov 12 09:15:33 koen: ping Nov 12 09:15:52 did dosfstools get dropped from the feed? Nov 12 09:16:50 yes, ads7846 spi3.0: no device detected, test read result was 0x00000FFF Nov 12 09:18:11 fox1213: hah, then you're having the same or similiar problem to me Nov 12 09:18:53 you cann't use TP too? Nov 12 09:19:12 i'm can't get spidev2 to populate /dev Nov 12 09:19:53 it seems like half of the people that come in here are having spi problems Nov 12 09:20:11 ........... Nov 12 09:21:20 it seems like you might just need to change your pins to pullups though Nov 12 09:23:00 mranostay: you are talking about angstrom package feeds, right? why would one get dropped? a maintainer left? Nov 12 09:26:25 im not sure SPI3 is correct setting, because i check the expansion board schematic, the TP pins are not connect to SPI3. It use mmc3_data3,GPIO16,3_4,mcbsp1_fsr,GPIO15,mcbsp1_clkx Nov 12 09:32:02 and enters Russ-o Nov 12 09:33:36 Russ-o? Nov 12 09:36:29 * av500 looks at the BBC micro accessory on his desk and wonders if that makes me want an r-pi :) Nov 12 09:37:01 ha, computing history museum has one too Nov 12 09:37:10 r-pi is uncloneable Nov 12 09:39:44 anyway gnite Nov 12 09:41:36 mranostay: pong Nov 12 09:41:56 nm Nov 12 09:42:11 av500: 100k credit card? Nov 12 09:43:07 sure, why not? Nov 12 09:43:11 its 65k only Nov 12 09:43:18 35k for shopping Nov 12 09:44:41 only Nov 12 09:44:44 clearly a bargain! Nov 12 09:45:15 let me go get that spare 100k i have Nov 12 09:50:57 Russ: sigh, it was too easy Nov 12 09:53:33 http://www.computinghistory.org.uk/det/12506/Marconi-RB2-Tracker-Ball/ Nov 12 09:53:56 mine is converted from Beeb to serial, need to upgrade it to USB now Nov 12 09:54:11 in fact, an old Genius Mouse lives inside it :) Nov 12 10:01:20 av500: by now probably an actual mouse lives inside it Nov 12 10:03:01 mru: nope Nov 12 10:03:02 * mranostay hides from panto Nov 12 10:03:02 I checked Nov 12 10:03:18 hi mranostay Nov 12 10:03:20 h av500 Nov 12 10:03:33 hi even Nov 12 10:04:00 hi odd Nov 12 10:04:35 panto is sure odd Nov 12 10:04:45 and proud of it Nov 12 10:05:04 (cause it makes people even more uncomfortable) Nov 12 10:08:29 panto: you do managed to keep your wife from meeting the real weirdos i noticed :) Nov 12 10:08:41 no offense to the weirdos here :) Nov 12 10:09:14 mranostay: weird, and proud of it? Nov 12 10:09:25 or proud and weird about it? Nov 12 10:10:17 *weird pride* Nov 12 10:10:49 *troll pride* Nov 12 11:38:03 https://www.pcworld.com/article/2013622/tiny-57-pc-is-like-the-raspberry-pi-but-faster-and-fully-open.html Nov 12 11:39:12 blasphemy Nov 12 11:40:16 hey, its fully open Nov 12 11:40:27 like the A13 public TRM that you can download *everywhere* Nov 12 11:40:35 no? Nov 12 11:40:54 koen: but in fact, there is not a 500page A10 TRM Nov 12 11:40:57 is now* Nov 12 11:41:08 mdp was muy impressado Nov 12 11:52:05 av500: 500 pages, so they're ~10% of the way there Nov 12 11:52:36 a modern soc is simply not possible to describe in less than a few k pages Nov 12 11:53:59 mru: A10% then :) Nov 12 11:58:44 we're the 99%? ;) Nov 12 11:59:53 yep Nov 12 12:00:25 I liked the comparison between A10 power sequencing and omap4 power sequencing Nov 12 12:00:55 surely ti didn't design it like that just for fun Nov 12 12:02:02 the cynic in me suggests it was designed that way to sell more PMICs Nov 12 12:02:16 yep Nov 12 12:02:33 at least, SoC and PMIC co-evolve Nov 12 12:03:44 whatever, the future is mips+sgx Nov 12 12:04:17 yes Nov 12 12:04:23 running a chinese android clone Nov 12 12:04:30 not tizen? Nov 12 12:04:41 and RMS will endorse it Nov 12 12:04:58 because sgx driver will be in EEPROM with the /WR pin cut Nov 12 12:05:24 mru: last I heard, tizen will run bada soon Nov 12 12:13:22 Hi koen. Did you see my message about 12½ hours ago? Nov 12 12:18:46 no, I was fast asleep at midnight Nov 12 12:19:56 koen: What’s the best place to get the sources for the same kernel version and set of kernel patches that the default Beagle Bone distro uses? The binary build from works with Debian, but i’d need to enable CONFIG_PPS and CONFIG_PPS_CLIENT_GPIO. All of , Nov 12 12:19:58 and exhibit the same problem: when powered over USB, CPU load causes the host computer to cut power to the Bone. Nov 12 12:24:38 https://github.com/beagleboard/kernel/tree/beaglebone-3.2 is the canonical repo Nov 12 12:24:50 (not the shuttleboy canonical) Nov 12 12:25:35 cloudboy Nov 12 12:25:47 slayer of red hats Nov 12 12:26:00 how much is a shuttle worth these days, decommissioned and all? Nov 12 12:26:56 koen: Ok, thanks. Is the branch at the “update to 3.2.28” commit roughly similar to the binary build at ? Perhaps i should try that one. Nov 12 12:40:35 hi alan_o Nov 12 12:42:05 hi panto! back in the "office?" Nov 12 12:42:17 back in the office :) Nov 12 12:42:19 how were the last couple days? Nov 12 12:42:22 morning commute was a killer Nov 12 12:42:33 alan_o, sightseeing mostly Nov 12 12:42:51 and trying to keep up with the rest of the guys on the alcoholic consumption dpt. Nov 12 12:43:08 hehe, yeah, they don't mess around :) Nov 12 12:43:57 tell you what though, mondays suck Nov 12 12:44:34 aww... how bad can it be, you're hacking the kernel in your house? That's the good life :) Nov 12 12:45:06 that said though, I'm sitting here procrastinating..... Nov 12 12:45:58 alan_o, I'm sitting at home, trying to finally start hacking the kernel... Nov 12 12:46:10 ahh.. yes, the email pile-up Nov 12 12:46:28 more like a highway car-crash pile-up Nov 12 12:47:12 I spent a few hours blowing through that on Saturday. Was not fun. Never enough time to keep up with that stuff when on travel. Especially when IMAP:SSL port is blocked by hotel internet. Nov 12 12:47:39 could have gone downstairs to the LF network, but to lazy to take my laptop down there. I'd have just gotten distracted by people I know anyway.... Nov 12 13:23:10 * koen is almost caught up with email Nov 12 13:30:49 I hate the post-conf backToRealitySucks mood Nov 12 13:31:03 * mdp uses studlyCaps to emphasize Nov 12 13:31:04 :) Nov 12 13:31:23 all that excitement and inspiration...nearly gone Nov 12 13:32:12 mdp: that's why you need a conference every 3-4 months Nov 12 13:32:20 hi mdp Nov 12 13:33:09 mru, maybe I should take dsaxena's advice and show up at Connects just to keep my momentum ;) Nov 12 13:33:29 hello panto Nov 12 13:33:33 next one is in march, a week after elc Nov 12 13:33:45 mdp: conf all year long? Nov 12 13:33:48 post conf blues? Nov 12 13:34:22 panto, I made the mistake of dialing into our sortaalmostaunifiedteammeeting friday Nov 12 13:34:35 panto, and somebody said something that has left me annoyed Nov 12 13:35:04 can you share? I feel that I should be annoyed all day today too Nov 12 13:35:38 +1 Nov 12 13:35:39 * alan_o likes having things to be annoyed about Nov 12 13:35:49 I'm only half annoyed Nov 12 13:35:52 alan_o: that's what the internet is for, no? Nov 12 13:36:03 mru: yes, and news Nov 12 13:36:14 alan_o: do you know why zigbee is such a fail for power consumption? Nov 12 13:36:18 isn't it pr0n? Nov 12 13:36:21 compared to e.g. BLE Nov 12 13:36:46 panto: that part is obvious, I'm talking about additional uses Nov 12 13:37:31 av500: Zig and 802.15.4 pre-date BLE by like 8 years. that's the best answer I can give really. I don't know much at all about BLE. Nov 12 13:37:42 alan_o: yeah Nov 12 13:38:08 BLE has slots per default and that seems to help a lot Nov 12 13:46:09 so if you add signals BLE can run Qt, no? Nov 12 13:48:29 koen: yes Nov 12 13:48:44 koen is making my head hurt.... Nov 12 13:49:08 alan_o: marcel said he'd try to get a few intel people assigned to fix 6lowpan-over-BLE Nov 12 13:49:10 panto, just a snarky comment...mgmt wouldn't catch it, of course so they get away with it Nov 12 13:49:16 alan_o: but he's on sabatical till january Nov 12 13:49:29 koen: oh, interesting. Nov 12 13:49:36 panto, makes me want to focus on something like parport_gpio today ;) Nov 12 13:49:58 as long as it's bitbanging parport Nov 12 13:50:05 what else? Nov 12 13:50:12 heh Nov 12 13:50:30 You can start writing parport_pru Nov 12 13:50:55 to get all the 1284 modes timing accurate Nov 12 13:51:07 the world will beat a path to my doorstep Nov 12 13:51:15 actually not, nobody will visit here ;) Nov 12 13:51:16 with fistfuls of cash Nov 12 13:51:22 oooh! Nov 12 13:51:29 bitbang the zombie apocalypse Nov 12 13:51:38 wait, I will just open an office in the bay area to accept all that cash Nov 12 13:52:07 guys, have you ever seen a non-optical mouse with ps2? Nov 12 13:52:14 as in one with a ball inside? Nov 12 13:52:17 av500: yes. Nov 12 13:52:19 er, with USB Nov 12 13:52:22 not ps2 Nov 12 13:52:22 av500: I'm using one right this minute Nov 12 13:52:23 oh Nov 12 13:52:31 not that Nov 12 13:52:35 all I find is optical ones that have usb Nov 12 13:52:44 it seems the switch to optical was befroe the one to usb Nov 12 13:53:10 av500, I have vintage ones with the big mouse balls, yes Nov 12 13:53:11 av500: mine is usb and has a ball on top, does that count? Nov 12 13:53:21 thats the other option Nov 12 13:53:25 but no Nov 12 13:53:32 I need something with quadrature encoders Nov 12 13:53:48 this is sounding like a serious project Nov 12 13:53:49 mdp: and usb? Nov 12 13:53:49 for tinkering or actual mousing? Nov 12 13:53:52 that's a big word! Nov 12 13:54:04 the trackball I posted Nov 12 13:54:06 av500, let me browse my storage cabinet..sec Nov 12 13:54:16 I converted it to serial a decade ago Nov 12 13:54:23 now I need to upgrade to usb Nov 12 13:54:42 so I need a ball mouse to hook the encoders too Nov 12 13:54:59 hm. Nov 12 13:55:01 other option is ps2 mouse and ps2 2 usb Nov 12 13:55:13 relucatantly Nov 12 13:58:02 ok, I lied..this will not be the last time Nov 12 13:58:28 14 mice in the box..8 optical usb and 6 ps/2 mechanical Nov 12 13:58:36 av500: I have a ball based usb mouse, Nov 12 13:58:41 good Nov 12 13:58:45 I think I threw out the one I'm thinking of Nov 12 13:59:16 av500, you know, you can just do a beaglebone ps/2 -> usb converter Nov 12 13:59:20 yes Nov 12 13:59:21 bitbang ps/2 Nov 12 13:59:25 one on each pru Nov 12 13:59:35 several years old and cheap but with heavy ball Nov 12 13:59:38 mdp: can one remoteproc talk to another? Nov 12 13:59:55 av500, "It's just software, we can do anything." Nov 12 14:00:00 then that could run and leave the A8 free for other things Nov 12 14:00:38 mdp: hey, thats my line Nov 12 14:01:24 av500, my part is usually, "...but I seriously doubt you have the budget for 'anything'" Nov 12 14:01:54 av500: i've heard that line way before. Nov 12 14:02:21 phh, it's a great one as mgmt invariably asks if something can be done given no constraints Nov 12 14:02:28 duh Nov 12 14:02:35 yeah Nov 12 14:02:56 av500: http://www.amazon.de/gp/product/B000K7BJZ8/ref=olp_product_details?ie=UTF8&me=&seller=&tag=aflattrcom369-21 Nov 12 14:03:18 janne: yes, I know that one :) Nov 12 14:04:36 I guess optic mice took the world over Nov 12 14:05:58 mdp: I saw your team in the press: http://arstechnica.com/gadgets/2012/11/remember-ti-84-calculators-theyre-finally-getting-color-screens/ Nov 12 14:06:10 lol Nov 12 14:06:13 I was thinking that ;) Nov 12 14:06:18 my product! Nov 12 14:06:36 You wrote that SPI display driver.... now we know what it was _really_ for :) Nov 12 14:07:01 alan_o, it will continue to be the perfect answer for people here in the Heartland asking, "now, exactly what is it that you do?" Nov 12 14:07:23 "You need color calculators!" Nov 12 14:07:29 alan_o: damn, I always want to click +1 on irc lines :) Nov 12 14:08:14 alan_o, you'd actually be impressed as to what the real customer was/is for the other spi fbdev driver I wrote for the dayjob Nov 12 14:08:19 av500: hehe, one day I'll have to get me on The Google's Plus-One Nov 12 14:08:38 mdp: oh, that was day job? Nov 12 14:08:54 alan_o, I can buy you an enterprise edition of G+ to entice you. Nov 12 14:08:57 mdp: let me guess, building a secure communicator for high ranking CIA peeps? Nov 12 14:09:12 av500, heh, fortunately no Nov 12 14:09:26 more geek oriented, hopefully they release RSN Nov 12 14:09:29 slow slow slow Nov 12 14:11:25 alan_o: I cleaned it for public consumption..you can tell it was for work as it's against a crusty old PSP kernel: https://github.com/ohporter/linux/tree/st7586fb Nov 12 14:12:24 mdp: I really hope they release RSN Nov 12 14:12:30 been waiting on that product for a while now Nov 12 14:12:35 mdp: ah, I know Nov 12 14:12:36 RSN? Nov 12 14:12:45 koen, me too, "I have all their albums!!!" :) Nov 12 14:13:14 LetoThe2nd: real soon now Nov 12 14:13:15 LetoThe2nd: snarky Real Soon Now Nov 12 14:13:20 ah. Nov 12 14:13:41 * koen caught a spanish flu bug in barcelona Nov 12 14:14:11 not spanish fly? Nov 12 14:14:21 koen: oh no!! I came home sick too. Mine was just from too many nights of no sleep I think. You sure yours wasn't from those snails? Nov 12 14:14:25 koen: at least the spanish flu mug did not catch you in barcelona. Nov 12 14:14:45 they gave out flu mugs ? Nov 12 14:14:48 thats mean ! Nov 12 14:14:57 the way they had this display supported at first was they reimplemented the davinci spi driver in a char driver...and then they used write() to pass framebuffers to the driver Nov 12 14:15:21 it was fairly humorous when they were complaining about performance and that they needed dma support Nov 12 14:15:39 alan_o: it started on thursday, before the snails Nov 12 14:15:41 mdp: why didn't they spi-gpio and spidev? Nov 12 14:15:46 alan_o: but those snails were yummie Nov 12 14:15:58 alan_o: it's surprising that they didn't ;) Nov 12 14:16:16 alan_o, it was a classic rtos approach..they go some sample codez from sitronix Nov 12 14:16:25 alan_o, either way they would have failed ;) Nov 12 14:16:42 mdp: now that's starting to sound about right. "Let's port this code to Linux now...." Nov 12 14:16:44 alan_o: the davinci kernel from TI does bitbang i2c Nov 12 14:16:54 alan_o: it muxes the i2c pins to gpio and then uses i2c-gpio on those Nov 12 14:16:55 koen, not any longer! Nov 12 14:16:57 stop that Nov 12 14:17:05 alan_o: I found out when the customer asked for the "dma version" Nov 12 14:17:35 koen, btw, I had to backport the 2.6.37/upstream version of the davinci spi driver to 2.6.33 *cough* to make this work Nov 12 14:17:50 samle all teh bits and bang teh codez.. ah no, the other way round.... Nov 12 14:17:59 mdp: .33rc + rt? Nov 12 14:18:15 mdp: is that for the follow-up product? Nov 12 14:18:20 of a certain company? Nov 12 14:18:33 I'm glad one of my former customers didn't upgrade to am335x Nov 12 14:18:41 av500, yes, but you're being so specific you'll give it away! Nov 12 14:18:47 they required a review of every non-mainline commit in the kernel tree Nov 12 14:19:05 and PSP can't say how they built the git tree... Nov 12 14:19:24 mdp: my lisp is sealed Nov 12 14:19:28 koen: that actually sounds like a good idea. I'd probably ask for the same if someone was hacking kernel for me. Nov 12 14:19:31 (((((((((lisp)))))))))))9 Nov 12 14:19:46 koen: do they also ask to morally justify mainline patches? Nov 12 14:19:52 were they developed in the open? Nov 12 14:19:57 koen, v2.6.33-rc4_DAVINCIPSP_03.20.00.14 is the magic "production" kernel Nov 12 14:19:57 by free will Nov 12 14:20:01 luckily not Nov 12 14:20:10 that will change Nov 12 14:20:59 koen, I was horrified when I started here and saw -rc releases used as the basis for supported releases Nov 12 14:21:08 never heard of such a thing before Nov 12 14:21:11 alan_o: I had to sit throught a morning of perforce gui on beamer before my statement of "we're not using mainline, we're using linux-davinci" was proven right Nov 12 14:21:11 we cant have open source just being "open" and "source" Nov 12 14:21:14 yes, perforce Nov 12 14:21:52 per farce? Nov 12 14:22:06 mdp: do I understand correctly that cpu_dai->name and cpu_dai_driver->name have nothing in common, i.e. the first refers to McAsp interface and the second to operation mode (IIS or DIT)? Nov 12 14:22:08 "who is this kevin hilman and why is he patching clocks?" Nov 12 14:22:15 koen, it seems like the wicked witch of the east had a lot to do with the culture of failure during those times Nov 12 14:22:21 LetoThe2nd: http://en.wikipedia.org/wiki/Medieval_hunting Nov 12 14:22:41 mdp: glad to see that my nickname for her stuck :) Nov 12 14:22:53 yegorich: tbh, I'm not very familiar with that driver Nov 12 14:22:54 av500: hrhrhr Nov 12 14:23:15 yegorich: you might check with joelagnel when he's around as he's working with it on a daily basis Nov 12 14:23:28 mdp: thanks Nov 12 14:24:01 it seems he's victim assignee for dealing with that driver ;) Nov 12 14:24:03 mdp: I just wonder if http://arago-project.org/git/projects/?p=linux-am33x.git;a=shortlog;h=refs/heads/v3.2-staging has functional mcasp.1 Nov 12 14:24:25 mdp: DMA stuff Nov 12 14:24:30 mdp: etc. Nov 12 14:25:38 in theory it should work Nov 12 14:26:15 can't vouch for ToT but the releases pulled off that tree pass system test where obvious stuff like playback/capture are tested on the 335x evm Nov 12 14:26:22 with the aic codec Nov 12 14:26:51 well Nov 12 14:27:01 capture doesn't work properly on beaglebone Nov 12 14:27:02 from what I hear, the driver is typical in that it has some narrow assumptions made and it will fail if you deviate from those Nov 12 14:27:06 mdp: O.K. I have a diffecerent codec though, but the problem is mpg123 hangs if I try to play sound file Nov 12 14:27:11 koen, did I say beaglebone? ;) Nov 12 14:27:12 and noone wants to go on record that it works on the evm Nov 12 14:27:40 the failure mode is that it only records for ~1s and then cuts out Nov 12 14:27:52 so in TI speak: it works! Nov 12 14:27:52 mdp: I don't need input, just output to the 9022 HDMI chip Nov 12 14:28:11 BH90210? Nov 12 14:28:48 yegorich: joelagnel is working on nxp hdmi support Nov 12 14:28:55 has a tree for that stuffs Nov 12 14:28:58 on github Nov 12 14:29:28 koen: almost ;-) Nov 12 14:29:39 yegorich: look through the branches at https://github.com/joelagnel/linux-kernel/tree/bonelt-hdmi Nov 12 14:29:51 some of them may be more current on his linux-kernel tree Nov 12 14:30:33 I think Gururaja and he have had a thread too where joelagnel was trying to get something working with Gururaja's patches hosted somewhere on gitorious Nov 12 14:31:03 iirc, too, daniel mack has mcasp working on mainline with his custom board Nov 12 14:33:51 mpd: I saw your thread on alsa mailing list Nov 12 14:37:05 ok, basically gururaja has reappeared and joelagnel is working on this hdmi thing so they'll have to most knowledge on how to actually get it working Nov 12 14:37:19 mdp: thanks Nov 12 14:37:53 I plan to stay far away from it...except that I do some regression tests on da8xx to make sure my series doesn't break access to the the private edma api on that platform. :) Nov 12 14:38:03 koen, I know you hate that I test...I'm sorry Nov 12 14:39:48 mdp: it builds, ship it Nov 12 14:39:58 cpsw ftw Nov 12 14:40:06 koen: that's "unit test complete," right? Nov 12 14:40:48 alan_o, with the right peers in your review, a unit test of "It build?" will be the complete unit test cadre ;) Nov 12 14:41:23 mdp: you mean you don't do TDD? Nov 12 14:42:26 total demolition derby? Nov 12 14:42:45 alan_o: your merchandise was a great way to anonymize our ELC badges Nov 12 14:43:01 +1 Nov 12 14:43:08 alan_o, you're making me think of that project I mentioned where we never updated our massive code and unit test document for the govt Nov 12 14:43:32 alan_o, rewrote everything and they said, "don't worry about it, just get it working" Nov 12 14:44:57 anyone seen a Nexus 4 yet? Nov 12 14:45:03 are LG phones any good? Nov 12 14:45:49 * koen loads http://www.signal11.us/oss/elce2012/ into chrome Nov 12 14:46:31 Hey guys. I want to integrate a ultrasonic sensor with the beagleboard C4 using a microcontroller. The micro must determine the distance to objects and send the value to the BeagleBoard. So far all i am sure of is that i would need a logic converter if micro is 5V. The question is, what is the simple fast way to get the micro to communicate with the beagle. I2C? Nov 12 14:46:46 * mdp sets alarm for 23:59 on 14nov to begin harassing alan_o Nov 12 14:48:00 Crofton|work: judging by their discontinued tablets, no Nov 12 14:48:11 koen: port it into chrome Nov 12 14:48:25 Odi: why a micro? Nov 12 14:48:30 Crofton|work: I'd like to know as well Nov 12 14:48:40 I wonder if google made them do a good job on the Nexus 4 Nov 12 14:48:50 online reviewers seem to think it's decent Nov 12 14:49:06 We need to find a troll with the exact answer Nov 12 14:49:22 or exact change Nov 12 14:49:27 using exact for everything should catch on Nov 12 14:49:37 Change we can all believe in Nov 12 14:49:42 you could always try and find out Nov 12 14:49:42 exactly Nov 12 14:50:44 Crofton|work: but than, at that price its a steal Nov 12 14:50:45 av500: hw or sw complaints about lg discontinued tablets Nov 12 14:50:51 and next year there will be another one Nov 12 14:51:07 is it possible to swap SDRC_nCS0 and SDRC_nCS1 in software before I had to cut the trace and fly some wires on omap3530/3730? Nov 12 14:52:00 My wife is using the nexus 1 Nov 12 14:52:16 I could upgrade and let her use 3 Nov 12 14:52:50 Crofton|work: I mentioned the usrp e-100 as a usecase for the DT stuffs Nov 12 14:54:17 in the sense of reading from eeprom to know what drivers are needed? Nov 12 14:55:21 xxiao possibly using mem=128M@0x88000000 as a kernel argument? Nov 12 14:55:30 Crofton|work: yes Nov 12 14:55:45 Crofton|work: and solve the conflicting LEDs :) Nov 12 14:56:00 ah yes Nov 12 14:56:20 the fpga+arm people need to get on board also Nov 12 14:56:31 iirc, they replied Nov 12 14:56:47 Odi: ask in the channel Nov 12 14:56:54 pm is metered Nov 12 14:57:11 oh Nov 12 14:57:32 I stopped the clock, there will be no charges :) Nov 12 14:57:43 lol tnx Nov 12 14:57:48 heh Nov 12 14:57:59 48 hrs of non-response and you can close the ticket Nov 12 14:58:01 what is thread called again? Nov 12 14:58:22 thurbad: tried that...kernel dead very early Nov 12 14:58:31 thurbad: just a few lines after uncompress Nov 12 14:58:43 :( Nov 12 14:58:59 ok so i want to connect an ultrasonic sensor to the beagleboard. its 5V. its output is a waveform and i need to measure pulse width. I know how to use a micrcontroller and interrupts but dont know how to go about doing that on a beagleboard. Nov 12 14:59:05 Crofton|work: "[RFC] Device Tree Overlays Proposal ..." Nov 12 14:59:39 koen: hehe yeah, they're badge-sized. Really by acciedent. Nov 12 15:00:10 thurbad: there might be some areas that is hard-coded to 0x80000000 as the physical starting address for omap35x Nov 12 15:00:13 alan_o, you're lying! you have your marketing hat on so it's all lies now Nov 12 15:00:37 mdp: I'm sure you'll be calling me on the phone at midnight..... "plzhaz teh codes! ur prmisd!" Nov 12 15:00:38 alan_o, *very* clever Nov 12 15:00:39 thurbad: i'm checking to see if i can modify MLO to change it to 0x88000000 instead, to avoid a tract-cut Nov 12 15:00:45 s/tract/trace/ Nov 12 15:01:07 xxiao the boot address is probably specified Nov 12 15:01:17 in boot.scr Nov 12 15:01:25 alan_o, and asking if I need an 802.15.4 radio to get it working Nov 12 15:01:31 alan_o, watz? Nov 12 15:01:52 thurbad: you mean bootm? no this is about phy starting address, it's now 80000000 Nov 12 15:02:10 i want to move it to 88000000, bypass the first 128M totally Nov 12 15:02:18 kernel will be loaded to 8a000000 Nov 12 15:02:27 'k Nov 12 15:02:32 alan_o: what was that SPI module breakout board again? Nov 12 15:03:49 koen: the mrf24j40ma? Nov 12 15:03:52 are beagleboard timers and interrupts easy to access? Nov 12 15:04:26 alan_o: that's the one, thanks! Nov 12 15:04:45 alan_o, btw, I was looking at an X-10 appnote from Microchip and I thought of another useless task for the PRU. Nov 12 15:05:02 mdp: yeah? Nov 12 15:05:34 no need to pay for a $.50 X-10 module to hook to beaglebone if you can make on in the PRU Nov 12 15:06:01 mdp: the X-10 wireless stuff? Nov 12 15:06:08 and X-10 is soooo old that it becomes naturally interesting Nov 12 15:06:35 no, normal X-10 protocol over power systems Nov 12 15:06:52 oh yeah.... Nov 12 15:07:00 normal people would just attach one of the free firecracker or CM-11a modules to their beaglebone Nov 12 15:07:20 you'd have your beaglebone plugged directly into the outlet :) Nov 12 15:07:34 alan_o, *ahem* what could go wrong? Nov 12 15:07:56 alan_o, I'll be sure to exclaim, "Here, hold my beer" before plugging it in Nov 12 15:07:59 hey, if mranostay can generate 500v and put it on bare wires...... Nov 12 15:09:49 * alan_o tries to figure out what all needs to get put online from his demo Nov 12 15:10:07 alan_o, all those commercial modules have the same directly coupled design to be cheap, fwiw Nov 12 15:11:53 completely transformerless...almost scary except it's got a 130V VDR to handle spikes in the +5V supply circuit Nov 12 15:13:55 bbl errands,, arrrrgh.... Nov 12 15:17:35 av500, btw, I was only impressed by the A10 datasheet because I've learned to set my expectations low and have them exceeded. Nov 12 15:18:27 mdp: wow, I have _no_ expectations and they're still not exceeded Nov 12 15:18:48 mru, you are difficult to please...do not change one bit. Nov 12 15:23:53 speaking of alternative community-oriented boards, did anybody discuss http://wandboard.org/ yet? Nov 12 15:25:54 even the price is magical Nov 12 15:26:48 community-oriented in the sense of "probably doesn't work and we don't give a damn"? Nov 12 15:27:14 ahh, you picked up on the subtle hint, excellent! Nov 12 15:28:31 I suppose that's different than beagle* where we give a damn but can't do anything about it Nov 12 15:28:40 it's the thought that counts, really Nov 12 15:33:16 FYI, new rev G of am335x TRM at http://www.ti.com/lit/ug/spruh73g/spruh73g.pdf Nov 12 15:48:35 guys is there a way to boot kernel from 128M@0x88000000 instead of 128M@0x80000000? tried to change uImage -a -e to 0x88008000 did not help Nov 12 15:53:07 <_av500_> mdp: btw, that linux toolchain for LM4, where is it? Nov 12 15:53:17 <_av500_> i took it with me on the train, might hacka bit tonight Nov 12 16:12:09 _av500_: exact steps... https://github.com/utzig/lm4tools Nov 12 16:12:43 <_av500_> thx Nov 12 16:13:21 they don't mention it, but I believe https://launchpad.net/gcc-arm-embedded will be fine as well as the other three toolchains they mention Nov 12 16:13:38 hello Nov 12 16:14:05 <_av500_> mdp: nice Nov 12 16:14:33 can someone help me with crontab on angstrom? Nov 12 16:14:40 it isnt working Nov 12 16:16:27 it is searching /etc/init.d/functions Nov 12 16:40:44 hello everyone, Im using BB-xM running Angstrom. When I plug the usb-webcam, the board found as /dev/video1 Nov 12 16:41:06 But I cannot use this device to take a photo using opencv Nov 12 16:41:26 Any wrong with my Board? Thank you so much Nov 12 16:42:05 <_av500_> try another app like e.g. cheese Nov 12 16:42:12 <_av500_> or a v4l2 demo app Nov 12 16:43:06 Yes, but I control my KIT via Serial port, not use the DVI display port Nov 12 16:43:14 how can I use cheese? Nov 12 16:43:34 this is my simple code: http://pastebin.com/1UFptq8s Nov 12 16:44:31 it always returns "No camera" Nov 12 16:44:36 sounds more like an opencv or driver issue, not a problem wirh the biard. Nov 12 16:44:53 i suggest yiy ask in the channels for those projects. Nov 12 16:45:03 *you Nov 12 16:45:13 <_av500_> why 2? Nov 12 16:45:14 agmlego: yes, thank you. Nov 12 16:45:24 the board can detect the webcam: Nov 12 16:45:57 when I typed: dmesg Nov 12 16:46:02 right, which is why i said you shoukd ask the opencv folk or tge driver folk for help. Nov 12 16:46:15 lots of opencv apps only try /dev/video0 Nov 12 16:46:21 http://pastebin.com/Gqbr2fYr Nov 12 16:46:31 so set the index to -1 or hardcode 1 in your code Nov 12 16:46:57 thank you Koen, let me try. Nov 12 16:51:42 Koen: it still returns "No camera Nov 12 16:51:57 I think openCV cannot use the camera. Nov 12 16:52:18 <_av500_> why 2? Nov 12 16:52:20 Do I need install any driver? Nov 12 16:53:01 iav_500: I used the index from -1,0,1,2 Nov 12 16:53:12 it returned No camera Nov 12 16:53:39 <_av500_> try some v4l2 example app Nov 12 16:53:42 I use 2 because in the dev, I see the video2 if I plug in the cam Nov 12 16:54:34 thank you av500, Im trying. Nov 12 17:01:24 mdp: I gave cscope a go last night. Good stuff. my cscope.out is 318M, but it's really fast. I guess I'm done for LXR for now. Thanks for the PROTIP :) Nov 12 17:09:44 What bus speed is the Beaglebone DDR2 running at (both 500MHz usb pwr'd & 720MHz ac adapter pwr'd)? I've looked but can't figure it out. Thx. Nov 12 17:23:32 xrivainc, 266 MHz I believe Nov 12 17:31:54 bradfa, i'm seeing better different benchmark numbers at 720mhz vs 500mhz, thought there might be a difference in memory bus clock. Nov 12 17:32:38 xrivainc, dunno, 266 is what u-boot sets it to, the OPP modes _probably_ don't change it but I've not looked closely Nov 12 17:33:27 xrivainc, got a scope? :) Nov 12 17:33:36 which bus? Nov 12 17:34:04 oh, the ddr2 Nov 12 17:34:16 the memory chip is a 400MHz one fwiw Nov 12 17:35:01 aka DDR2-800 Nov 12 17:37:01 xrivainc, see page 82 in the am335x data sheet Nov 12 17:37:03 yes, max ddr2 on am335x is 266, the chip is 400. I'm seeing 16% memcopy speed improvement when the CPU is 44% improved (500 vs 720) Nov 12 17:37:05 shows opp modes and ddr speeds Nov 12 17:37:35 data sheet says only OPP50 slows the ddr clock as far as am335x silicon is concerned Nov 12 17:37:50 and opp50 has errata so don't use it :) Nov 12 17:37:58 i've got a 100mhz scope, but haven't tried to look at it yet Nov 12 17:38:11 xrivainc, 100 MHz scope won't show you much Nov 12 17:38:23 that's why i haven't tried to look at it yet Nov 12 17:38:26 :) Nov 12 17:38:41 didn't think i'd get an answer that way Nov 12 17:40:50 it might show you nice 50MHz waveform :) Nov 12 17:41:52 okay, i looked at the opp vs ddr table. i'll presume that is the answers I'm looking for (266 for all opp except opp50). Nov 12 17:43:20 xrivainc, if you're seeing only 16% better memcpy speed with a 44% cpu increase, I'd say the CPU is the limit and even at 500 MHz it's rather close to saturating the ddr bus Nov 12 17:43:42 are you running in Linux or bare metal or u-boot? Nov 12 17:43:58 regardless, interesting findings Nov 12 17:45:55 this is the beaglebone off-the-shelf linux (angstrom?). I agree it probably shows that at 500mhz cpu, it can't saturate the ddr2, so that's why 720 can improve it a bit. Nov 12 17:50:10 so it seems at 720mhz, the beaglebone rev a6 has similar benchmarks to 720mhz beagleboard (rev c4). The memcopy is almost the same (beagleboard is omap3, 32bit 166mhz lpddr1). Nov 12 17:50:59 xrivainc, seems reasonable Nov 12 17:51:07 the bone dram is 16-bit wide Nov 12 17:51:26 beagle is 32-bit Nov 12 17:54:28 yes, comparison of 16 bit ddr2 @266 vs 32 bit lpddr1 @166. Nov 12 17:54:59 what speed is the L3 on the am33x? Nov 12 17:59:17 page 82 of datasheet says 200mhz on l3 for all opp except opp50 Nov 12 17:59:35 opp50 is 100 on l3 Nov 12 18:00:28 on omap3 L3 and ddr share clock Nov 12 18:02:20 Anyone used the TPS65217 backlight driver? Nov 12 18:02:25 I've patched in this (https://groups.google.com/forum/?fromgroups=#!msg/rowboat/Uzt_QDofrMo/gTpbnaFRqIsJ) Nov 12 18:02:39 Which is the rowboat tps65217 backlight driver patch. Nov 12 18:02:44 But I can't figure out how to enable it. Nov 12 18:03:05 Can anyone offer some guidance? Nov 12 18:07:33 edahling_, depends on the kernel you use Nov 12 18:07:43 is it the mainline kernel? Nov 12 18:07:49 Using the TI-am33x-3.2 kernel Nov 12 18:08:21 you need to add platform data for it then Nov 12 18:08:48 Could you elaborate, or give me something to read / google? I'm kind of new to this. Nov 12 18:08:58 Where would I add platform data? Nov 12 18:09:09 sigh... Nov 12 18:09:16 ok, hang on a bit Nov 12 18:09:33 Sorry, and thanks. Nov 12 18:11:04 I think I have a vauge idea of what you're saying. Nov 12 18:11:26 I've added a struct tps65217_bl_pdata to my board-am335xevm.c Nov 12 18:11:40 yes Nov 12 18:11:45 I'm just not 100% sure how this all works yet. Nov 12 18:11:58 you need to also add a bl_pdata field Nov 12 18:12:12 to beaglebone_tps65217_info Nov 12 18:12:22 and point to the that structure Nov 12 18:12:25 however Nov 12 18:13:00 be advised that any pwm backlight value for the beaglebone not 0% or 100%, will cause the ethernet MII to fail Nov 12 18:13:30 Well that sucks. =P Nov 12 18:13:34 Why is that? Nov 12 18:14:49 something wrong with the EMI, or shielding I guess Nov 12 18:16:28 So these TPS65217_BL_FDIM_xxxHZ, those are what I use to set the backlight value? Nov 12 18:17:21 edahling_, no idea TBH, I've only used the DT versions Nov 12 18:17:35 but yeah that sounds right Nov 12 18:17:42 Alright. Sounds right to me too. Nov 12 18:17:55 Thanks for the input, I think I have a better idea of what's going on now. Nov 12 18:18:32 np Nov 12 18:19:50 I also have a post on googlegroups about the lcd I'm trying to interface, if anyone wants to take a glance at my issue. I'm quite stumped. Nov 12 18:19:56 https://groups.google.com/forum/#!topic/beagleboard/IEGjhZ0z-5M Nov 12 18:28:05 So I'm probably just doing something wrong here, but I tried bitbake virtual/kernel -c devshell and then in the devshell did make menuconfig and then make modules and ld is complaining about unrecognized options for fs/9p/9p.o -- trying to do make modules from devshell wrong or is something else going on? Nov 12 18:30:07 after make menuconfig you need to exit Nov 12 18:30:46 and bitbake -f -c compilevirtual/kernel; bitbake -c deploy virtual/kernel Nov 12 18:31:22 (assuming you want your work not to all go away after compiling Nov 12 18:31:24 ) Nov 12 18:36:09 thurbad: Ok, I'll give it a shot. Nov 12 18:56:24 alan_o, let's post the Q&A to stackoverflow.com so I can build some cred! Nov 12 18:56:47 alan_o, glad you like it..much more favorable to the typical cli workflow Nov 12 18:58:09 mdp: yeah, I like it. I also tried cbrowser, which is nice too, but for that one, it'd be nice if I could click on the text inside the display window and then re-search for that. Nov 12 18:58:25 I actually don't have an SO account Nov 12 18:58:40 but I'm open to making one Nov 12 18:59:07 I wonder how many people have multiple SO accounts and just ask themselves questions..... Nov 12 18:59:36 "What do I do if I have no SPI peripheral?" .... "Bit-bang it of course" Nov 12 18:59:45 * prpplague slaps alan_o around Nov 12 18:59:47 * alan_o marks accepted answer Nov 12 18:59:47 alan_o, I don't either..SO account Nov 12 19:00:36 if I knew how to build/use OE, I could be happy with just handling questions here though Nov 12 19:00:40 hey, prpplague should be the one to give that answer... he gave a whole talk on bitbanging :) Nov 12 19:01:03 indeed, he's the supreme bitbanger Nov 12 19:01:22 yeah, I told koen that if he helped me understand some stuff about OE, that I'd update his documentation for him. Nov 12 19:01:37 Saying it here so that it's on the record, and I can't weasel out :) Nov 12 19:02:03 careful Nov 12 19:02:24 saying something here forced me to test/use bradfa's u-boot am335x serial console patches ;) Nov 12 19:02:42 bitbake virtual/kernel -c devshell (in new shell: make menuconfig -> select modules -> exit and save; exit devshell with Alt+d); bitbake -f -c compile virtual/kernel; bitbake virtual/kernel -c deploy --> But my module.ko files don't show up? Nov 12 19:02:45 yeah... I have a way of biting off more than I can chew, with good intentions. Nov 12 19:03:02 df___: exactly Nov 12 19:03:17 alan_o: see, endless opportunities to be a helpdesk hero Nov 12 19:03:30 yeah, if I only knew the answers :( Nov 12 19:03:32 alan_o, you could have the highest rate of ticket closures on the team Nov 12 19:03:41 well, yeah, that does suck Nov 12 19:03:49 alan_o: That's expected? Nov 12 19:04:05 maybe we can bump our metrics just by giving some answer however inaccurate Nov 12 19:04:10 df___: well, I'm saying it fits with our discussion of how there's only one guy here who seems to know OE. Nov 12 19:04:21 unless the rest just hide :) Nov 12 19:04:23 * dm8tbr will demand exact steps before accepting answers Nov 12 19:04:26 alan_o: Ah, I see. Nov 12 19:05:32 mdp: oh, koen also told me that he has a tree on his github which reflects his mainline + patches script tree. Nov 12 19:05:34 let me find it Nov 12 19:05:46 that' will be good for when people come on here and want to build a kernel. Nov 12 19:05:59 dm8tbr: there's any number of blogs with exact steps...98% of them are wrong exact steps...we aren't going to demand correctness Nov 12 19:06:21 dm8tbr: yeah, don't be picky :) Nov 12 19:06:32 So basically if I read something that didn't come from koen or wasn't linked by koen it probably doesn't work? ;) Nov 12 19:06:46 :D Nov 12 19:07:05 df___: There's a lot of stuff that's wrong. There's also two different kernels. The one that comes on your bone, and one that tracks mainline. Nov 12 19:07:16 The one on your bone is based on a TI kernel which has been heavily patched Nov 12 19:07:22 based on 3.2 Nov 12 19:07:22 you have to save the config file with the right filename Nov 12 19:08:03 alan_o: Right. Nov 12 19:08:11 has anyone fight with crontab and angstrom on beagle? Nov 12 19:08:36 jpg, not personally Nov 12 19:09:15 koen has an RPi tree on his github... We'll have to give him a hard time for that :) Nov 12 19:09:36 jajajja Nov 12 19:10:10 It's OE related so maybe he gets a pass :) Nov 12 19:10:10 mdp, you didn't volunteer for my Ethernet PHY patches though :( Nov 12 19:10:19 the problem is that crond call /etc/init.d/functions and this file doesnt exist on angstrom Nov 12 19:10:19 s/patches/patch/ Nov 12 19:10:34 alan_o: you know that darn pandaboard display bag got lost Nov 12 19:10:35 bradfa: I'm out of time! :P Nov 12 19:10:40 mdp, you can't make more? Nov 12 19:10:44 Ok, well here's another question then. In this comment (http://goo.gl/vz4no) on the beagleboard list/group koen describes the steps I just did except after make menuconfig he copies .config elsewhere (/OE/openembedded/recipes/linux/linux-omap-//defconfig) I used setup-scripts to setup my env. Does anyone know where that recipes dir might be using setup-scripts? Nov 12 19:10:48 alan_o: it showed up at my house on sunday afternoon Nov 12 19:10:59 prpplague: did you pick it up in MIA? Nov 12 19:11:07 or lost from MIA to DFW? Nov 12 19:11:07 alan_o: yea Nov 12 19:11:09 ah Nov 12 19:11:17 bradfa, alas, no Nov 12 19:11:44 df__: depends on your kernel, but if you're using the TI one, it's under source/meta-ti/ Nov 12 19:11:52 bradfa, I tried to bitbake time, and it blew up with some weird recipe dependencies..and I can't find any hits on bing to solve the problem Nov 12 19:12:11 it says I have to install something called, "python" Nov 12 19:12:13 edahling_: Ok, yeah I am (beaglebone). Thanks. Nov 12 19:12:18 mdp, you need be in UTC for that to work, else you have to bit bang python Nov 12 19:12:42 yeah, I saw a guy with some pru code to implement this python thing Nov 12 19:12:51 true story Nov 12 19:13:09 I think interpreted languages are the software equivalent of bit-banging. Nov 12 19:13:20 alan_o: hehe Nov 12 19:13:27 completely agree Nov 12 19:13:28 alan_o: speaking bit banging Nov 12 19:13:34 * prpplague joins #raspberrypi Nov 12 19:13:43 hah Nov 12 19:13:59 I've lurked on there before. They never say anything interesting when I'm on though. Nov 12 19:14:16 "other fruity boards" Nov 12 19:14:36 alan_o, there's only one hardwarebug quote worthy comment there every few days. Nov 12 19:15:10 and today, everybody from that channel is standing in line for the midnight blackops release...they'll be playing all night long in their parents' basement Nov 12 19:15:12 mdp: dollar general stores are selling a 5lb reeses peanut butter cup, shall we order a bunch of them and have them delivered to jkridner ? Nov 12 19:17:38 prpplague: wait, wha??? Nov 12 19:17:39 You can buy 5lb reese's cups? Nov 12 19:17:54 edahling_, in texas, that's the standard size Nov 12 19:17:59 it goes up from there Nov 12 19:18:02 heh Nov 12 19:18:02 Haha. Nov 12 19:18:15 I guess everything is bigger in texas. Nov 12 19:18:34 that's why it's on the bottom of the map... Nov 12 19:18:39 dinner Nov 12 19:18:53 they're like 10mg portions in michigan ;) Nov 12 19:18:54 edahling: hehe they have a snickers bar this size of a regular loaf of bread Nov 12 19:19:01 yes, even texas is bigger in texas Nov 12 19:19:19 Whoa, you're kidding. I've never heard of this. Nov 12 19:19:43 right, there's this full scale map of texas, and I have no idea where that things is stored. Nov 12 19:20:09 edahling_: you've never been to texas Nov 12 19:20:59 texas https://www.youtube.com/watch?feature=player_embedded&v=xCEmb4XZohU Nov 12 19:24:25 I'm still trying to figure out one of jkridner's g+ posts Nov 12 19:25:07 which one? Nov 12 19:25:27 "Another way not to add a driver to BeagleBone" Nov 12 19:25:56 summary, somebody tried em28xx on bone and it didn't work Nov 12 19:26:20 mdp, if you want, I can write all sorts of ways to write bad drivers and put them on g+ Nov 12 19:26:23 em28xx isn't exactly stable Nov 12 19:26:53 edahling: http://www.candyblog.net/blog/item/snickers_slice_n_share Nov 12 19:27:04 mdp: which one? Nov 12 19:27:24 * jkridner should read better Nov 12 19:27:25 https://plus.google.com/u/0/104712705716996155416/posts/D4k8L1Q9gng Nov 12 19:27:42 ok, not jkridner but bhoundleroy Nov 12 19:27:43 mdp: sounds like you figured it out. Nov 12 19:28:36 not sure, actually ;) Nov 12 19:29:24 Those are my favorite posts. Titled something like "Adding a Kernel module to beaglebone" then it goes on with step-by-step directions for two pages and concludes with: "I never did get it to work, maybe I can compile it on the beaglebone?" Nov 12 19:30:02 df___: yeah, you should have to pass a test before they give you a blog :) Nov 12 19:31:33 jkridner, I suppose the idea could be that we can all learn from failures Nov 12 19:33:44 alan_o, before you institute that test, I'm doing a blog entry about how I failed to get my wintv 950q working with cppi 4.1 dma on bone Nov 12 19:33:54 * mdp scampers off. Nov 12 19:34:57 Prpplague: that is one huge snickers. Nov 12 19:36:03 mdp: hehe. sometimes those can be good, if the point of the blog is up front about what problems the person had, and that the post is about the problems. As long as it's not just whining, and as long as the person is also working upstream. Nov 12 19:36:32 how many conditions can you place in one statement? hehe Nov 12 19:36:35 Greg K-H said in one of his videos, "tell us when there's a problem; don't just whine about it on your blog" Nov 12 19:37:05 heh, hardly surprising that a driver that barely works on x86 has issues on arm :-D Nov 12 19:37:06 I was just left wanting for more Nov 12 19:37:10 mdp: yeah, that was a lot of conditions. I'll be checking out your post with a clipboard later..... Nov 12 19:37:57 mdp: with giant checkboxes on it, like on those drug commericals. Nov 12 19:38:05 alan_o, and btw, contiki OS has completely ruined me...thanks to you Nov 12 19:38:20 mdp: ruined you? Nov 12 19:38:33 yes, reading too contiki-os code now...bah Nov 12 19:38:40 nice! Nov 12 19:39:23 I feel a 6502-based mdp board coming in our future, with full 6LoWPAN support :) Nov 12 19:40:13 alan_o, that radio is spi-based, yes? Nov 12 19:43:44 yep Nov 12 19:43:55 There's already support for the TI radio in there too, iirc. Nov 12 19:44:01 CC2530? Nov 12 19:44:19 edahling: that is just the 1lb one Nov 12 19:44:25 edahling: you should see the 5lb one Nov 12 19:44:43 ds2, mdp: cc2420 Nov 12 19:44:51 alan_o, one reason I was going to put a cpld on the cape is to allow having this: http://www.6502.org/users/andre/hwinfo/spi/spi65b/index.html Nov 12 19:45:56 ti has too many radios Nov 12 19:46:11 ds2, if they would quit buying companies... Nov 12 19:46:30 I can't imagine the opamp portfolio with national's parts mixed in now Nov 12 19:46:53 mdp: oooh.... That's what you need man, a full-up bus going on. Nov 12 19:48:26 alan_o, with a hw spi peripheral and all those slave chip selects...it makes for some capable interfacing Nov 12 19:48:28 mdp: these all have CC in their name Nov 12 19:48:31 so... :P Nov 12 19:48:46 ds2, true, I do marvel at all the acquired CC parts Nov 12 19:49:40 hi Nov 12 19:51:31 can someone take a look at this kernel patch and tell me if they see anything wrong with it? Nov 12 19:51:53 JViz: looks good to me so far Nov 12 19:51:58 heh Nov 12 19:52:08 alan_o, it's also a nice PIC...between it and a 65c22 via, you get a complete system...given that the via has the timer/counters/gpios one typically wants to do real work Nov 12 19:52:09 http://pastebin.com/1eMpUMzA Nov 12 19:52:34 i'm trying to enable /dev/spidev2.0 Nov 12 19:52:43 i have an lcd7 Nov 12 19:52:51 ewwwwwwwwwwww lcd7 Nov 12 19:52:59 eww? Nov 12 19:53:06 ds2, lcd8 is better? Nov 12 19:53:15 JViz: does it work? Nov 12 19:53:23 yes. lcd8.14159265354 is much better Nov 12 19:53:34 well, as far as i can tell, no Nov 12 19:53:47 JViz: well ,then I don't like your patch :) Nov 12 19:54:13 JViz: you're missing a step somewhere. Nov 12 19:54:21 spi1_pin_mux needs to acutally get used. Nov 12 19:54:44 JViz: I looked at the TowerCAN cape as an example. Nov 12 19:54:58 [ 1.325256] BeagleBone cape: exporting SPI pins as spidev Nov 12 19:54:58 Follow everything it does, and you should be able to connect your driver to SPI1 Nov 12 19:55:03 in dmesg Nov 12 19:55:28 it's doing the export Nov 12 19:55:40 is the mux getting used the way you want? Nov 12 19:55:52 Why did you change the pulling, etc? Nov 12 19:56:08 because the configuration i need doesn't use pullups Nov 12 19:56:57 also, INPUT and OUTPUT are somewhat of a misnomer there. Nov 12 19:57:08 Those map to "Receiver Enable" Nov 12 19:57:19 and best I could tell, those have to be on for all for SPI lines Nov 12 19:57:33 "Receiver Enable"? Nov 12 19:57:39 not sure why, but when it started working, I stopped screwing with it Nov 12 19:57:52 yes, those map to the MUX registers in the chip Nov 12 19:58:12 and AM33XX_PIN_INPUT maps to the Receiver Enable bit Nov 12 19:58:27 i'm trying to make them just plain non-pulled inputs and outputs Nov 12 19:58:32 what do i need to call them? Nov 12 19:58:32 Look at the TRM, section 9.3.1.50 Nov 12 19:58:33 they are named to induce confusion versus the TRM Nov 12 19:58:42 see if you are all paying attention Nov 12 20:00:04 JViz: Mine are all set to 0x33 Nov 12 20:00:08 (I'm using device tree) Nov 12 20:00:29 If I were you, I'd use what's already there, and just take off the pull Nov 12 20:00:43 also, pull is inverse logic Nov 12 20:00:47 0 is enable pull Nov 12 20:01:08 so taking pull off of there didn't change anything. Nov 12 20:01:32 swell Nov 12 20:01:49 yes. Read that section of the TRM Nov 12 20:01:53 does the pinmuxing prevent spidev from populating though? Nov 12 20:01:56 er Nov 12 20:01:58 bad pinmuxing? Nov 12 20:02:07 probably not Nov 12 20:02:11 it just prevents it from working Nov 12 20:03:03 so spidev not populating /dev/ is an unrelated problem? Nov 12 20:03:11 hah Nov 12 20:03:22 oh, well probably Nov 12 20:03:26 yes Nov 12 20:03:36 great Nov 12 20:03:45 look at the TowerCAN cape. Do everything it does Nov 12 20:03:52 except change the driver to spidev Nov 12 20:03:57 spidev not populating means you either 1) did not build in the module or have the module present that matches your current 2) did not register the spidev device in your board file (or dts) Nov 12 20:04:19 oh yeah, make sure it's built in :) Nov 12 20:04:27 s/current/current kernel/ Nov 12 20:04:30 i have CONFIG_SPI_SPIDEV=y in my config Nov 12 20:04:40 well, did you fulfill #2? Nov 12 20:04:41 JViz: ok, so it's number 2 Nov 12 20:04:57 again, TowerCAN.... go! Nov 12 20:05:06 your patch shows no indication of the spidev device being registered Nov 12 20:05:15 right Nov 12 20:05:24 it's right in koen's stock kernel board file too Nov 12 20:05:43 I recall he made spidev work by default for some cases or two Nov 12 20:05:58 yeah, well, the line right under exporting the pinmux is registering the spidev board info Nov 12 20:06:10 I was saying earlier that I didn't see it in there, but I thought he said that it was showing up. I apparently mis-read. Nov 12 20:06:29 so 1 and 2 should be being met Nov 12 20:06:35 does spidev2 match with SPI1 on that kernel? Nov 12 20:06:53 it's on bus 2 Nov 12 20:07:28 the line under the pinmux does not register anything Nov 12 20:07:30 bus 2 is correct Nov 12 20:07:36 .bus_num = 2, Nov 12 20:07:37 that's what I used anyway, with that kernel Nov 12 20:08:00 you have a structure...with spidev board info..did you check that it is actually used? Nov 12 20:08:36 do you see the difference? Nov 12 20:08:51 spi_register_board_info(bone_spidev2_info, ARRAY_SIZE(bone_spidev2_info)); Nov 12 20:08:53 mdp: see the new linux foundation event schedule for 2013? Nov 12 20:09:02 JViz: Serious dude.... Look at what it does for TowerCAN. That's what I did. Everything it does, I did, and it works. Nov 12 20:09:22 JViz: did you confirm that executes? Nov 12 20:09:23 towerCAN uses spidev? Nov 12 20:09:32 JViz: no..... Nov 12 20:09:42 mdp it shows up in my dmesg with the rest of the process block Nov 12 20:09:56 JViz: Do everything that does, but use the spidev as the driver name instead Nov 12 20:10:10 JViz: it serves as an example of how to register an SPI device. Nov 12 20:10:29 JViz: but you're not registering an mcp2515, you're registering an spidev, but it's all the same stuff. Nov 12 20:11:30 JViz: why don't you paste your dmesg Nov 12 20:11:57 i did Nov 12 20:12:40 [ 1.325256] BeagleBone cape: exporting SPI pins as spidev Nov 12 20:12:55 or do you mean the whole thing? Nov 12 20:13:35 the towerCAN cape uses the default spi1 pinmux Nov 12 20:13:39 whole thing Nov 12 20:13:52 so that lines doesn't show me that you registered the board info Nov 12 20:14:14 well, actually, I take that back..had to look again at this old kernel ;) Nov 12 20:16:50 JViz: one more option 3) spidev probe() fails Nov 12 20:17:00 http://pastebin.com/fUiZTTwc Nov 12 20:18:43 3) is tough to manage for spidev as it would have to fail some trivial core device model stuff... Nov 12 20:19:00 interesting..sadly, it works fine stock for koen Nov 12 20:19:16 JViz: you're going to need to put some printk() in there and make sure it's doing what you want it to do. Nov 12 20:19:42 JViz: good old fashioned debugging and learning. You have all the tools, you have examples in the code that work. Now you just need to put it together. Nov 12 20:19:52 alan_o, can DT help here? Nov 12 20:20:14 mdp: yes, but then we'd be debugging his Device Tree :) Nov 12 20:20:29 JViz: you have mcspi enabled? Nov 12 20:21:08 all i did was that patch i showed you and add SPIDEV=y to the config Nov 12 20:21:36 i haven't read all of the code, it's all default Nov 12 20:22:11 check that CONFIG_SPI_OMAP24XX=y Nov 12 20:23:36 1586:CONFIG_SPI_OMAP24XX=y Nov 12 20:24:13 i noticed koen had code in there that was supposed to enable spidev2, so i thought i could just enable that Nov 12 20:24:34 the presence of the lcd7 is supposed to disable it Nov 12 20:24:44 i commented out that line Nov 12 20:25:05 and i physically disconnected the button that was attached to gpio3_16 Nov 12 20:25:36 nfc, there's a lot going on in that board file, unfortunately Nov 12 20:25:52 that's an understatement Nov 12 20:25:53 you ruled out all the common stuff now Nov 12 20:26:26 yeah, not sure how one board file for a bunch of different boards made sense. Nov 12 20:26:43 I thought that was the point of the board file, to pin down what's on this board. Nov 12 20:27:30 in the last kernel version, my edit worked Nov 12 20:27:38 spidev2 came on Nov 12 20:28:13 but the pullups were makiing it fail to function properly Nov 12 20:28:24 JViz: those pullups are still on Nov 12 20:28:35 alan_o, I was overruled on that one because this aligned us with DT ;) Nov 12 20:28:36 yeah, but now it's not even populating /dev Nov 12 20:28:53 mdp: what? Nov 12 20:28:54 s/overruled/shouted down/ Nov 12 20:29:00 What does that have to do with DT? Nov 12 20:29:07 alan_o, nfc Nov 12 20:29:21 oh Nov 12 20:29:32 I did correct that to "shouted down" Nov 12 20:29:54 yeah, someone really thinks that the beaglebone is just another EVM Nov 12 20:30:09 even though the marketing is completely different, and the support from TI is completely different Nov 12 20:30:15 it's just takes an eeprom Nov 12 20:30:21 or a village Nov 12 20:31:27 bed time is here Nov 12 20:31:45 panto, lucky, cya Nov 12 20:31:53 later panto ! Nov 12 20:31:53 but from the way that is built, it is just another EVM Nov 12 20:32:08 supported by an approved vendor designated proxy Nov 12 20:32:15 third party interests are ignored. Nov 12 20:33:04 ds2, I was asked one day if we expect all customer boards to have an eeprom Nov 12 20:33:14 heheh Nov 12 20:33:18 can't say who...but it was humorous (to me) Nov 12 20:33:25 but of course you do Nov 12 20:33:36 we all know you're staging things to acquire a EEPROM company ;) Nov 12 20:33:58 ds2, yes, I said, "it depends..some people expect that...but they don't live in my world" Nov 12 20:34:14 'lo Nov 12 20:37:55 time to experiement Nov 12 20:45:16 mdp, all boards don't have eeproms? Nov 12 20:45:36 :) Nov 12 20:45:59 really, I was surprised too Nov 12 20:46:17 bradfa, this would be an excellent rant^H^H^H^Hblog entry for you Nov 12 20:51:35 hehe Nov 12 20:53:16 on a lighter note, I have power button press and usb/ac gain/loss detection working on bone Nov 12 20:53:24 now just need to do something useful Nov 12 20:53:55 bradfa: power button like from USB HID? Nov 12 20:54:27 alan_o, power button like the one on the bone header Nov 12 20:54:36 alan_o, I'm making tps65217 do my bidding Nov 12 20:54:37 finally! Nov 12 20:55:08 * mdp EOLs tps65217 Nov 12 20:55:08 now need to learn about input devices and turn the tps65217 into one Nov 12 20:55:10 bradfa: oh, I didn't look that closely at the pins apparently. Nov 12 20:55:26 mdp, tps65910 looks like similar Nov 12 20:55:33 is actually what I'm basing my driver off Nov 12 20:55:42 so I don't mind much if EOL 65217 happens :) Nov 12 20:55:52 * mdp delivers bradfa a tps65218 with a completely different register set Nov 12 20:55:59 * mdp EOLs 65910 too Nov 12 20:56:05 ha! Nov 12 20:56:05 * mdp chases bradfa's designs Nov 12 20:56:06 :_ Nov 12 20:56:07 ) Nov 12 20:56:09 tps65217 2.0, different pinout Nov 12 20:56:29 * mdp leaves bradfa 600mil DIP packages only Nov 12 20:57:29 I used to hate the tps65217, but now I understand it, and it's not so bad Nov 12 20:57:58 * koen installs a new OTP on bradfas TPS chips Nov 12 20:58:09 he's been infected! Nov 12 20:58:33 * bradfa dies of tps infection Nov 12 21:00:02 * dm8tbr demands TPS reports! hands out cover sheets and memos about them Nov 12 21:04:52 Does anyone know of a simple way to figure out which defconfig file is being used by bitbake virtual/kernel -c compile? I tried tacking -D on and looking through the log file but I still didn't see it. Nov 12 21:06:46 When I run compile it says it's using linux-ti33x-psp_3.2.bb as the recipe, so I copied my defconfig file to linux-ti33x-psp-3.2 but doesn't seem to be getting picked up. Nov 12 21:07:16 (all under sources/meta-ti/recipes-kernel/linux/) Nov 12 21:08:34 bradfa, it's probably not that it's not so bad..it's just that you decided the cure is worse than the tps disease Nov 12 21:09:18 mdp, I've now accepted the tps disease into my life. I am now at peace Nov 12 21:09:38 it's not a bad lifestyle change Nov 12 21:09:50 mdp, and it's legal in NY! Nov 12 21:09:54 yes! Nov 12 21:10:49 * bradfa goes home and has a beer, ta ta! Nov 12 21:14:11 dfarmer: doesn't the bb file reference a defconfig? Nov 12 21:15:45 I'm trying to use can on the bone, and I can't get can0 to show up. Nov 12 21:16:14 I'm muxing and running am33xx_d_can_init(1) Nov 12 21:17:08 edahling: did you look at the TowerCAN example? It sets up the onboard CAN and also the external SPI CAN Nov 12 21:17:24 * alan_o sets TowerCAN to be his default support answer Nov 12 21:17:25 I did, and I'm pretty sure I'm doing everything exactly the same. Nov 12 21:17:37 Which is why I'm so confused. Nov 12 21:18:28 did you do ifconfig can0 up? Nov 12 21:19:18 * mdp uses TowerCAN to fix DMA Nov 12 21:19:36 yes we CAN! Nov 12 21:19:39 mdp: TowerCAN and DT.... the SOLUTION! Nov 12 21:19:57 alan_o: no, just bitbang it Nov 12 21:20:27 squirt dt in via ftdi jtag Nov 12 21:20:34 Maybe that's my next ELC talk, "Solving the world's problems with TowerCAN, DT, and BitBanging (with Exact Steps)" Nov 12 21:20:37 four more years! Nov 12 21:20:57 djlewis: I have to listen to TowerCAN for 4 more years!?!? Nov 12 21:21:05 intolerable Nov 12 21:21:08 4 more ears! Nov 12 21:21:09 It's looking thtat way Nov 12 21:21:40 we say, 4 more years until those bottom-feeders come back to Ohio (really 3, but...) Nov 12 21:23:53 When I do ifconfig can0 up I get some errors. Nov 12 21:24:01 bittiming not defined... Nov 12 21:24:15 Where do I define bit timing? Nov 12 21:25:19 with ip Nov 12 21:25:33 look at Documentation/net/can.txt Nov 12 21:25:34 ah, but when I run that it gives me a different error. Nov 12 21:25:37 ifconfig should die Nov 12 21:25:37 or something like that Nov 12 21:26:04 mru: 10 years they've been saying that and it hasn't. If ip weren't so difficult to use, maybe it would have. Nov 12 21:26:17 'ip' is _much_ easier to use Nov 12 21:26:36 mru: think so? I know the man page is harder Nov 12 21:26:39 the man page even includes a formal syntax Nov 12 21:26:45 exactly :) Nov 12 21:26:55 nothing could be easier Nov 12 21:27:49 we'll continue using ifconfig in .us, thank you Nov 12 21:28:05 hehe.... Nov 12 21:28:12 ah, it's an imperial/metric thing Nov 12 21:28:13 I require my exact steps in both units Nov 12 21:28:33 I did say to someone one time, "IP is designed to be like Cisco Router syntax" (which is what I read sometime). Nov 12 21:28:34 but ifconfig also talks about metric... Nov 12 21:28:52 mru, quit trying to confuse people with the facts Nov 12 21:28:54 and the response I got was, "doesn't Cisco make that hard on purpose so they can sell training certs?" Nov 12 21:28:55 cisco syntax actually makes sense Nov 12 21:29:01 if you understand what you're doiing Nov 12 21:29:21 can we align on NetworkManager only? Nov 12 21:29:34 the problem is that most people don't understand basic networking concepts Nov 12 21:30:18 ever seen a network configuration gui with dns settings tied to specific interfaces? Nov 12 21:30:24 that's just sick... Nov 12 21:30:41 mru: I'm a firm believer in making the easy/common things easy. in the case of ifconfig/ip, setting IP/mask/gw is the common case ,and should be the easiest, and called out specifically probably near the beginning of the man page. Nov 12 21:31:10 gw should also not be tied to an interface Nov 12 21:31:14 that's not how networks work Nov 12 21:31:32 right Nov 12 21:31:42 but, those 3 things are the things you'll want to set up Nov 12 21:31:49 and what's so hard about "ip addr add a.b.c.d/len dev eth0"? Nov 12 21:31:59 or "ip link up/down" Nov 12 21:32:09 mru: probably nothing Nov 12 21:32:13 or "ip route add a.b.c.d/len via e.f.g.h" Nov 12 21:33:03 and if those three things were called out as examples in the man page, we might all have a little less gray hair. Nov 12 21:33:22 if you only ever have one interface with one address and one route to everything else, you might not notice the difference Nov 12 21:34:08 start doing things with multiple interfaces, interfaces with multiple addresses, multiple routes etc, and ifconfig and friends fail quickly Nov 12 21:34:19 So yeah, here is my problem. I run ip link set can0 type can bitrate 50000 triple-sampling on and it tells me my type is garbage. Nov 12 21:35:25 we don't serve your time 'round here Nov 12 21:35:35 edahling_: you need more than bitrate Nov 12 21:35:39 you need the proper timing Nov 12 21:35:43 look up Microchip AN754 Nov 12 21:36:01 mru: you're right about that. Nov 12 21:36:14 I've been largely following http://processors.wiki.ti.com/index.php/AM335X_DCAN_Driver_Guide Nov 12 21:36:15 mru: and it's largely because ifconfig uses the old interfaces to the kernel. Nov 12 21:36:52 edahling_: I don't know that document, but you need tq, phase-seg-1 phase-seg-2, and sjw Nov 12 21:37:10 see that can.txt document I referenced earlier Nov 12 21:37:28 I'm reading through that now. Nov 12 21:37:33 edahling_: also, make sure that the timing you specify is possible with the oscillator settings you are using. Nov 12 21:37:55 alan_o: for extra fun, add some tunnel devices Nov 12 21:37:59 Well, I don't know what oscillator settings I'm using... I'm just gonna do some more reading and come back later. Nov 12 21:38:03 Thanks for the guidance. Nov 12 21:38:20 edahling_: seriously checkout that AN754 doc. It explains all the timing Nov 12 21:38:28 and that info isn't on that TI page you linked Nov 12 21:38:30 * KotH wonders what all the ruckus is about Nov 12 21:38:43 KotH: people doing Work! Nov 12 21:38:53 alan_o: unpossible Nov 12 21:39:01 Russ, ;) Nov 12 21:39:37 KotH: can you describe the ruckus? Nov 12 21:39:38 alan_o: sounds more like some grumpy old man bitching about how their way of life is better than other peoples Nov 12 21:39:44 KotH: and you have yet to see a proper friday here Nov 12 21:40:06 mru: i've been on #panda for a while... that was quite enough :-) Nov 12 21:40:18 it's much better here Nov 12 21:40:21 alan_o, yeah: It says file://defconfig in the bb file. It also has file://am33something.bin -- so I put defconfig in linux-ti33x-psp-3.2 (which has the firmware) but it doesn't seem to get used. Nov 12 21:40:26 mru: i can already see that Nov 12 21:40:27 theres' barely been 10 lines on #panda all day Nov 12 21:40:35 panda, that's the one with the Amazon processor on it? Nov 12 21:40:36 alan_o: Thanks, gonna read through both articles. Nov 12 21:41:02 dfarmer: that references a file which exists. Nov 12 21:41:23 if you dump a defconfig somewhere and didn't overwrite something, you put it in the wrong place Nov 12 21:41:59 why do people talk about "a defconfig" when what they mean is "a config"? Nov 12 21:42:24 defconfigs are those files shipped in the kernel tree as defaults for various systems Nov 12 21:42:29 mru: i hope you do not really expect an answer to that? Nov 12 21:42:38 anything else is by definition not the default any more Nov 12 21:42:46 mru: tell that to OE :) Nov 12 21:43:06 mru: but I agree with your lamentation of the watering-down of terminology. Nov 12 21:44:25 so grab a shotgun and help me fix it Nov 12 21:44:42 mru: is that because you aren't allowed to? :D Nov 12 21:44:43 alan_o: i hope, you've a raised floor as well Nov 12 21:44:58 alan_o: no need for a short temper though... Nov 12 21:45:12 KotH: ?? Nov 12 21:45:34 alan_o: I have a t-shirt that reads "I have a shotgun, a short temper, and a raised floor in my server room" Nov 12 21:45:58 sounds like exact steps to me Nov 12 21:46:14 * mru should remember to wear it at next elc Nov 12 21:46:55 anyways.. gtg and get some sleep Nov 12 21:46:58 night boys Nov 12 21:52:50 Alright, I'm understanding the timing stuff, but when i run '# ip link set can0 type can help' it returns 'ip: either "dev" is duplicate, or "type" is garbage' Nov 12 21:53:04 I believe it's not supposed to do that. Nov 12 21:55:04 I read somewhere that the /bin/ip version doesn't work because of some association with busybox?.., but I don't have /sbin/ip Nov 12 21:55:36 edahling_: oh... that could be Nov 12 21:55:57 How can I go about remedying that? Nov 12 21:56:03 install actual IP Nov 12 21:56:06 err... ip Nov 12 21:56:21 with opkg...? Nov 12 21:56:51 edahling_: my can experience is limited. I only did it on a board using the default image that came with it. In fact, one of my tasks this week is to build a rootfs for it. Nov 12 21:57:11 edahling: not sure how to do it with Angstrom Nov 12 21:57:21 might could do it that way. Nov 12 21:57:30 or build an angstrom image that isn't busybox based Nov 12 21:57:31 Mmkday. Nov 12 21:57:35 or build ip from source Nov 12 21:57:40 Alright. Nov 12 21:57:44 shouldn't be that hard to do the latter Nov 12 21:57:45 I can figure this out. Nov 12 21:58:47 edahling_: when you say "some association with busybox" do you know what you mean there? Nov 12 21:58:57 about what busybox is? Nov 12 21:59:14 (I never know who knows what things) Nov 12 21:59:34 This (http://www.armadeus.com/wiki/index.php?title=CAN_bus_Linux_driver) Nov 12 21:59:38 alan_o: assume nobody knows anything and you'll be close to the truth Nov 12 21:59:49 Yeah, I really don't know anything. Nov 12 22:00:25 ok, so busybox is a multi-call binary. somewhere there's a symlink for /bin/ip which points to /bin/busybox Nov 12 22:00:33 /bin/ls points to /bin/busybox Nov 12 22:00:34 etc Nov 12 22:00:42 (/etc doesn't point anywhere) Nov 12 22:00:47 but regardless..... Nov 12 22:01:05 the behavior busybox gives, is determined by the name of the symlink it's called by, get it? Nov 12 22:01:14 is a symlink to /dev/null a null pointer? Nov 12 22:01:31 whoa Nov 12 22:01:39 * alan_o is thinking about what a universe might look like where that is true... Nov 12 22:01:58 is there some relation to evil spock here too? Nov 12 22:02:19 Ah, so why would anyone want that? Nov 12 22:02:20 evil spock only barely rings a bell.... Nov 12 22:02:42 edahling_: Excellent question.... it's because having only one binary means a lot less overhead. Nov 12 22:02:51 Alright, good answer. Nov 12 22:03:10 edahling_: busybox installs are really small, and it's only one exe that's loaded into memory for as many times as you run it concurrently. Nov 12 22:03:38 it also lets all the little tools share code for things like option parsing without using shared libs Nov 12 22:03:49 and less overhead than shared libs would have too Nov 12 22:04:02 it's double-plus good Nov 12 22:04:41 except the evil license Nov 12 22:04:45 And it's GPL too, so it's one more thing to spin people up :) Nov 12 22:04:48 hahah Nov 12 22:04:52 jinx Nov 12 22:04:56 toybox ftw Nov 12 22:05:00 ugh... Nov 12 22:05:05 I think that's a stupid idea Nov 12 22:05:17 "let's re-invent it with a different license" Nov 12 22:05:19 but Nov 12 22:05:38 I do understand that there are license trolls who use busybox to come after you for other unrelated things. Nov 12 22:06:13 yes, what a shame that people are expected to follow a license agreement Nov 12 22:06:18 darn trolls Nov 12 22:06:31 having a bsd-licensed set of system utils is not a bad idea Nov 12 22:06:49 mru, no complaints Nov 12 22:07:07 always nice to see people _do_ instead of license whine Nov 12 22:07:14 The thing is, it's not even the system utils that anyone cares about. Something I read on LWN said that one of those trolling companies has yet to find one line of modification to busybox. Nov 12 22:07:27 mdp: I guess that's somewhat true. Nov 12 22:08:05 I'm opposed to everything going BSD. for sysutils, who cares, but for example my android phone, I get a phone with an "open source" operating system that I can't modify. Nov 12 22:08:40 if people want to bsd-license code they wrote, they have the right to do so Nov 12 22:08:40 BSD is great if you're the first one getting the code. Everyone downstream from the first guy that closes it, is screwed. Nov 12 22:09:13 mru: no doubt about it. And I have BSD code out there too Nov 12 22:09:47 those who take bsd code and close it up wouldn't use gpl code in the first place Nov 12 22:09:54 so there's really nothing to discuss Nov 12 22:11:15 mru: that's true, and part of the argument for Android. If Anroid was GPL, would we see it on our phones? The answer is maybe. We see GPL Linux on our phones because it's so good. I think the same may have happened for Android, had it been GPL, but maybe to a lesser extent. Nov 12 22:11:52 there's a gpl kernel in android because it does the job and allows non-gpl userland Nov 12 22:12:08 if linus were an arse and tried to prevent non-gpl userland, linux would be dead Nov 12 22:12:32 oh yeah, no doubt. Nov 12 22:13:15 but android vendors (and even google) would no doubt much prefer a BSD-licensed kernel. Unfortunately, none of the BSDs could do the job. Nov 12 22:13:25 I don't see why Nov 12 22:13:38 they'd gain nothing by closing the kernel Nov 12 22:13:42 it's just a kernel Nov 12 22:13:57 That's what Karim says anyway. He says that they (google) evaluated the BSDs and they didn't have the features they needed. Nov 12 22:14:14 it's not about closing the kernel, it's about avoiding GPL lawsuit headache. Nov 12 22:14:22 sure Nov 12 22:14:34 the same as with the userland of Android Nov 12 22:15:42 besides...busybox is bloated Nov 12 22:16:10 or featureful, depending on your perspective :) Nov 12 22:16:21 depends on your application Nov 12 22:16:45 if there is no room, then it doesn't matter what your perspective is Nov 12 22:17:12 yeah, it'd be fun to get in there and make an option for _everything_ Nov 12 22:17:33 or a few levels of functionality Nov 12 22:17:40 I think there's something like that iirc. Nov 12 22:17:56 if you take a look at the klibc stuff, busybox is just gigantic Nov 12 22:18:35 ds2, you mean with lots of busybox options on..or even within functionality units? Nov 12 22:18:49 mdp: I mean just overall binary size Nov 12 22:18:56 I mean, sure, most distro builds of busybox are ginormous Nov 12 22:19:07 well, it can be very very small Nov 12 22:19:22 I question if you can make it smaller then what klibc can do Nov 12 22:19:29 no Nov 12 22:19:36 I agree klibc is more useless Nov 12 22:19:37 challenge accepted! Nov 12 22:20:05 userland exists solely to demostrate kernel features ;) Nov 12 22:20:18 ds2, klibc is definitely focused on being slimmer and minimal, it succeeds Nov 12 22:20:37 ds2, userland exists solely as a demonstration of poor coding Nov 12 22:20:50 klibc is the bare minimum needed to load some modules and mount a real filesystem Nov 12 22:21:03 they have a shell and basic commands Nov 12 22:21:06 enough to do diags Nov 12 22:21:24 btw, great dave johnson session at an old OLS on that..bad userspace code Nov 12 22:22:07 ds2: sure, but it's not meant to be the primary tools of a system Nov 12 22:22:20 mru: for embedded systems, why not? Nov 12 22:22:46 if it does what you need, go ahead and use if of course Nov 12 22:22:50 *it Nov 12 22:23:02 hence my mentioning that klibc is more useless Nov 12 22:23:36 size matters! Nov 12 22:23:44 indeed Nov 12 22:24:07 we're all working with 8M of the system memory and XIPing from NOR. Nov 12 22:24:16 right? Nov 12 22:24:26 obviously it has to fit within the size limits of the systems Nov 12 22:24:27 8M, who the hell died and made you rich Nov 12 22:24:32 8M?! what luxury Nov 12 22:25:09 yes, 8K ought to be enough for any 6502 Nov 12 22:25:51 why have userland at all? Just have the kernel start your one-and-only userspace app as its init Nov 12 22:25:54 well, 1 or 2M for embedded linux systems used to be typical Nov 12 22:26:02 alan_o, I'm sure lennart is working on it Nov 12 22:26:15 Russ, back in the day ;) Nov 12 22:26:48 my big iron PPC had 16M Nov 12 22:26:51 why have a userspace init Nov 12 22:27:05 why have a kernel? Nov 12 22:27:06 ds2: just put it all in the kernel :) Nov 12 22:27:17 put it all in mlo Nov 12 22:27:18 we can run apps in u-boot Nov 12 22:27:23 and eliminate DT too Nov 12 22:27:25 yeah, my_app.S Nov 12 22:27:26 u-boot, such bloat! Nov 12 22:27:50 Busyboot (BSD version) Nov 12 22:27:57 get TI to put it into the ROM Nov 12 22:28:17 alan_o, too many TPS reports to put something in the ROM Nov 12 22:28:36 alan_o, a lot of people in robes sitting on high thrones review that stuff Nov 12 22:28:46 just run everything on the pmic Nov 12 22:29:00 get rid of that pesky omap Nov 12 22:29:02 we EOLed those earlier today when we found that bradfa figured them out Nov 12 22:29:18 can stellaris help here? Nov 12 22:29:45 solaris? Nov 12 22:29:48 if we could get the PRU out of its SoC prison then we might have a solution Nov 12 22:29:56 yes, slowaris for PRU Nov 12 22:30:14 prularis Nov 12 22:30:23 pluralis Nov 12 22:30:37 one for each core of course Nov 12 22:30:53 there shall only be a plurality Nov 12 22:30:58 no way man, SMP Solaris on dual PRU Nov 12 22:31:04 get porting! Nov 12 22:31:15 who will give us a C compiler?!? Nov 12 22:31:27 write a sparc emulator for pru Nov 12 22:31:31 surely not TI...but then who? Nov 12 22:31:45 mru knows C, maybe he can write one for us Nov 12 22:31:52 he claims to know it Nov 12 22:31:53 fun weekend project for him Nov 12 22:32:18 why would anyone want a c compiler for pru? Nov 12 22:32:32 because assembly is too difficult Nov 12 22:32:34 because we need to port solaris Nov 12 22:32:47 but then C is too difficult too Nov 12 22:32:47 and what we really want is python, but baby steps. Nov 12 22:32:51 mdp: good luck predicting the timing of c code Nov 12 22:32:54 how much do you know about Ruby? Nov 12 22:33:04 alan_o: exact baby? Nov 12 22:33:11 * Russ notes that it's time to uninstall mcafee Nov 12 22:33:14 mru, bah, your obsession with time is unnerving Nov 12 22:33:34 did you look at sdcc? Nov 12 22:33:38 mdp: did you know that time progresses in exact steps? Nov 12 22:33:40 mru, please give us exact steps Nov 12 22:33:47 mru, metric units? Nov 12 22:33:57 mru, those scientifical ones? Nov 12 22:34:05 microsteps Nov 12 22:34:16 yoctosteps? Nov 12 22:34:25 CareBear\: iirc, I think sdcc was what panto threw out once Nov 12 22:34:48 I like sdcc. Nov 12 22:35:14 helped me to get going on 8051. In a few years it'll be analogous to mdp's 6502 :) Nov 12 22:35:16 cc65 is growing on me, mostly cause it's what I've been working with lately Nov 12 22:35:34 mdp: ever try out lunix? Nov 12 22:35:35 mdp : ack Nov 12 22:35:39 * Russ tries to track down his hc11 code Nov 12 22:36:19 alan_o, 8051 shall ever be the mcu where I fist encountered PL/M-80 :( Nov 12 22:36:26 alan_o: no Nov 12 22:37:03 is cc65 related to the ancient deepblueC? Nov 12 22:37:27 came from SmallC Nov 12 22:37:35 so did deepblueC Nov 12 22:37:42 different forks then Nov 12 22:37:49 mdp: PL/M yikes... I had to look it up Nov 12 22:37:56 does cc65 output real 6502 code? Nov 12 22:38:30 IIRC, deepblueC did some sort of byte code thing Nov 12 22:38:43 alan_o, koen saw an atm reboot into OS/2 at elc .. dunno if you heard that...those were the new units when I co-oped at that company...previous units I coded on was all custom code in PL/M-80 on the various processors Nov 12 22:38:45 yuck Nov 12 22:38:56 ds2: you're gonna have to watch his video from ELC-E.... the best moment of ELC-E was during his talk, IMO Nov 12 22:39:13 'his'? Nov 12 22:39:22 p *'his' Nov 12 22:39:22 mdp's Nov 12 22:39:27 oh Nov 12 22:39:30 "you're all following this, right? 0x61 store something something something" Nov 12 22:39:31 lol, alan_o, not sure it was quite that great, but ok ;) Nov 12 22:39:42 mdp: I'm still laughing Nov 12 22:39:57 mdp: multiple people said something like "best of show" about your beef demo Nov 12 22:40:13 oh, the machine code Nov 12 22:40:19 seemed fairly normal to me Nov 12 22:40:47 I shortened it because I had a bug in the backspace handling of the wozmon port and if I typoed it didn't remove the character from the stream Nov 12 22:41:04 it was a demo only geeks could handle Nov 12 22:41:29 but it was an exact demo Nov 12 22:41:40 there were exact steps, indeed. Nov 12 22:42:34 everytime I look at the png output from fritzing, that's what makes me laugh Nov 12 22:42:44 it's harder to follow than looking at the breadboard itself Nov 12 22:43:24 I added USR0…USR3, as well some TODO items. Any patches would be appreciated, but i’ll try to get around to the TODO items at some point myself. https://github.com/ion1/beagle_bone_pins Nov 12 23:07:38 ds2 : yes cc65 outputs real 6502 code. it's what the entire c64 demoscene uses Nov 13 00:24:03 CareBear\: There's a c64 demo scene still? I need to get me one of those..... Nov 13 00:27:07 alan_o : it is booming. this year was 30th anniversary of the c64. the major c64 demo party was in the netherlands three weeks ago. the competition entries were better than ever before. all other demo scenes are decreasing, c64 is increasing. Nov 13 00:28:24 alan_o : http://youtu.be/zes4VwV_hIw http://youtu.be/f5UlXq4gUB0 Nov 13 00:28:36 1st and 2nd place Nov 13 00:31:59 it must take forever to make that stuff Nov 13 00:32:46 years Nov 13 00:35:29 turn it up to HD for the true experience :p Nov 13 00:36:10 alan_o : we talked about http://www.nordicsemi.com/ 2.4GHz radios http://www.nordicsemi.com/eng/Products/2.4GHz-RF like the nRF24L01+ (transceiver only) or nRF24LU1+ (SoC with flash and usb) Nov 13 00:37:11 alan_o : their "ShockBurst" thingy has MTU=32 Nov 13 00:37:48 yikes! Nov 13 00:38:08 note that like all other simple radios they are susceptible to the packets-in-packets attack Nov 13 00:39:11 so make sure to add crypto Nov 13 00:39:11 man, now you've got me looking at C64's on ebay..... Nov 13 00:39:15 haha Nov 13 00:39:24 they're not all that easy to come by anymore Nov 13 00:46:55 vice seems to do a good job Nov 13 00:51:56 Is there a place I can see all of the images that I can build for the beaglebone? Nov 13 00:52:04 (bitbake) Nov 13 00:55:39 For beaglebone is board-am335xevm.h the right board file to edit to change e.g., SPI setup? Nov 13 00:55:49 err... .c Nov 13 01:04:02 Nevermind meta-angstrom/recipes-images/angstrom Nov 13 01:49:41 Hah. I found an error in . On page 34 it says that uart5_rxd is an output. Nov 13 01:52:14 ion, ahh, but did you then click the "Submit Documentation Feedback" at the bottom to report it? Nov 13 01:52:32 mdp: Not yet, but i obviously will. Nov 13 01:52:42 great! thx Nov 13 01:53:06 sadly, many people don't even see that link in the footer of each page of the trms/datasheets Nov 13 01:56:23 gah it feels like 10a Nov 13 02:19:53 help - trying to build the angstrom kernel for beaglebone and the build process is failing Nov 13 02:20:23 for some reason it can't git clone the meta-smartphone repo so the whole build fails. Nov 13 02:30:14 fun Nov 13 02:41:41 yeah - basically all of OE is broken because one branch is inaccessible Nov 13 02:46:15 but I thought git was distributed Nov 13 02:46:22 they told me this could *never* happen Nov 13 02:51:30 heh Nov 13 02:51:44 never is shorter than you think Nov 13 02:51:57 well, it's back to rcs then Nov 13 02:52:10 a bit extreme? Nov 13 02:52:34 or you know, you could just have a clone of the OE git repos around somewhere Nov 13 02:52:40 but I think rcs is the answer Nov 13 02:54:57 emeb_mac, you might have better luck asking in #oe, but it is late and people are afk Nov 13 02:56:45 mmmm beer **** ENDING LOGGING AT Tue Nov 13 02:59:57 2012