**** BEGIN LOGGING AT Sat Jun 08 02:59:58 2013 Jun 08 03:01:00 Thihi, Prototype 3. Prototypes 1 and 2 worked great. Prototype 1 used an industrial PLC controller, prototype 2 used an arduino. If I am lucky prototype 3 will be on a BeagleBone Black, yet making progress on controlling the pins in C has been daunting. Jun 08 03:03:00 MH_ it'll all come in good time, and you'll be that much more happy when you do "get it" Jun 08 03:03:30 I cant wait to start tpying with the PRUs eventually Jun 08 03:03:35 toying* Jun 08 03:05:27 Yea, I am hoping if I keep pushing at it I will eventually get past these hurdles. I've got my head around the PRU's alright, the whole linux device tree has me pretty lost though 0_o. What languages are people using predominately for the Bone? Is the Java Bonescript the most popular? Jun 08 03:06:47 cant say im on the C/C++ bandwagon personally Jun 08 03:08:51 using the PRU's though we're forced into using asm it seems Jun 08 03:11:11 yes but that is also the pru's strength. The instruction set is small and it is really easy to use. Jun 08 03:11:13 MH_, ok, very interesting. What kind of milestones did you mean earlier btw. Like what do you want to get done *at least* before you publish? Jun 08 03:15:32 I would like to use the debug serial port to read data from an external device on the BBB. I am having trouble finding information. Jun 08 03:18:51 Thihi, I want to have confirmed that my development direction is the right path. I do not want to start publishing work on a design project that has shortcoming preventing its growth. Which means at a bare minimum getting past basic access of the beaglebone pins. My prototype 1 was limited by my PLC's communication capabilities. I want my 3D scanning software and electronics to be universally applicable for different 3D scanning mac Jun 08 03:19:30 As I have a long list of 3D scanning appliations I am trying to acheive:D Jun 08 03:25:13 Thihi, typically I like to have proven my concepts before I do any documentation. I think when my projects get to about 3/4 the way done then I like to work on documentation as I know the time will be well spent. Jun 08 03:53:48 hey emeb_mac Jun 08 03:53:57 hi ssi Jun 08 03:54:08 any luck getting spi in working? Jun 08 03:54:14 no, not really :) Jun 08 03:54:21 sadness Jun 08 03:54:21 all I know so far is "yep, that's not how your dts should look" Jun 08 03:54:22 hahah Jun 08 03:54:33 I really don't understand the stuff well enough to even know how to begin Jun 08 03:54:41 I'm reading up on it now Jun 08 03:54:42 I'm grokking it in bits and pieces Jun 08 03:54:55 I have a pretty good handle on the pinmux side, but I don't understand how to configure the mcspi driver Jun 08 03:54:58 just updated the eMMC on my BBB A4 to 5-27 Jun 08 03:55:14 the dts I wrote is for spidev driver, and I think I'd be better off using the omap2-mcspi driver Jun 08 03:55:46 spidev should sit between userspace and mcspi Jun 08 03:56:05 (I think) Jun 08 03:56:34 but then my knowledge is probably out of date Jun 08 03:56:35 you may be right Jun 08 03:56:45 anyway the mcspi has the concept of tx only vs txrx spi ports Jun 08 03:56:51 and I imagine that the way I have it configured is as a tx only port Jun 08 03:56:55 aha Jun 08 03:57:00 but I don't know what sort of sorcery is required to make it do properly Jun 08 03:57:11 also ka6sox tells me that my dts is missing necessary clock routing info Jun 08 03:57:30 he knows that better than I Jun 08 03:57:47 yes he knows better than most of us Jun 08 03:57:54 but apparently not quite well enough to be able to rattle off the solution :) Jun 08 03:58:23 it seems like 1% of the developers have 99% of the dt knowledge Jun 08 03:58:24 :D Jun 08 03:58:34 someone should start a protest movement Jun 08 04:00:39 occupyDT Jun 08 04:00:52 :D Jun 08 04:02:01 ka6sox has worked w/ panto who seems to be near the source of this stuff. Jun 08 04:04:40 hi Jun 08 04:05:18 ssi, you are correct....the 1%'ers do know...but they are trying to make it less of a mystery Jun 08 04:05:24 speak of the ka6sox and he shall appear Jun 08 04:05:33 hehe that's what I was hoping :D Jun 08 04:05:44 right now I've been dealing with a work emergency...still in progress Jun 08 04:05:49 understandable Jun 08 04:06:09 ssi: do you have your .dts somewhere I can see it? Jun 08 04:06:11 friday night emergencies are always fun Jun 08 04:06:20 he has a nice gits Jun 08 04:06:22 gist Jun 08 04:06:25 https://gist.github.com/ianmcmahon/5732144 Jun 08 04:06:30 that's the gist of it Jun 08 04:06:31 :D Jun 08 04:06:31 that sounded dirty Jun 08 04:06:46 ssi, I managed to really screw up the routing Jun 08 04:06:54 oh no Jun 08 04:07:01 took the studio guy about 45 minutes to unravel my mess Jun 08 04:07:17 he has *dialup* Jun 08 04:07:31 hah Jun 08 04:08:09 eventually he got in and took over the console and got things rerouted. Jun 08 04:09:07 he kept calling me and saying he couldn't get in...turns out I didn't log out right Jun 08 04:09:21 oh well. Jun 08 04:09:35 when things go wrong, they go wrong all the way Jun 08 04:09:52 you don't even know the 1/2 of it...it got "interesting" earlier Jun 08 04:10:01 because of the evac? Jun 08 04:10:08 ya Jun 08 04:10:12 yeah that's a rough one Jun 08 04:10:25 4 Jun 08 04:10:27 we had a fun one a few months back Jun 08 04:10:34 I'll bet Jun 08 04:10:38 HVAC contractor hit what he thought was an hvac disconnect Jun 08 04:10:45 uh oh Jun 08 04:10:51 was actually the emergency power shutdown for our cnn 8s datacenter Jun 08 04:10:58 whoops Jun 08 04:11:01 whoops indeed Jun 08 04:11:11 you never know just how bad your failover strategy is til someone hits the big red button Jun 08 04:11:18 he he Jun 08 04:11:37 I've been coordinating bugout stuff all night Jun 08 04:12:13 someone had the presence of mind to bring a nationwide show on a USB stick just after recording Jun 08 04:12:33 so it went out off a USB stick Jun 08 04:12:47 good planning Jun 08 04:13:16 we have shows that are filmed in ny that get broadcast from our facility, and pseudo-live Jun 08 04:13:19 like conan Jun 08 04:13:26 and we get tapes of the second segment while the first segment is airing Jun 08 04:13:34 they nailed up a backhaul to DC and pushed the stick in just before it was scheduled to air (like 5 seconds) Jun 08 04:13:53 JIT Jun 08 04:14:08 I had to design live second screen content recognition sync systems to be able to deal with that show Jun 08 04:14:20 oh my Jun 08 04:14:44 it basically gets fingerprinted on the playout system, which is delayed from air just long enough to get the fingerprints in a public facing database in time for clients to start making requests against it Jun 08 04:14:49 super fun Jun 08 04:14:57 all our preproduced content can be pre-fingerprinted Jun 08 04:14:58 but not conan Jun 08 04:14:58 no thanks Jun 08 04:15:02 ssi: did you look at cape-bone-adafruit-lcd.dtbo? Jun 08 04:15:10 emeb_mac: no, not yet Jun 08 04:15:13 is it in the kernel tree? Jun 08 04:15:23 it's in /lib/firmware Jun 08 04:15:31 run it backwards through dtc Jun 08 04:15:43 you'll get the fragment source - it uses spi1 Jun 08 04:15:54 to get all the satanic messages? Jun 08 04:15:58 has some good stuff in it on setting that up Jun 08 04:16:04 can I play a country song thru it and get my dog, truck and woman back? Jun 08 04:16:10 exactly Jun 08 04:16:24 dtc -I dtb -O dts Jun 08 04:16:31 I'm trying to get the bbb back online Jun 08 04:16:34 I just got home... long day :P Jun 08 04:16:47 now it's time for the fun. Jun 08 04:17:02 nothing better than hacking BBB on a friday night. ;) Jun 08 04:17:04 yep Jun 08 04:17:07 I'm ready for it too Jun 08 04:17:09 all jacked up on coffee Jun 08 04:17:19 * emeb_mac has a Hoegaarden Jun 08 04:17:41 it's in the linux src Jun 08 04:17:42 ./linux-src/firmware/cape-bone-adafruit-lcd-00A0.dtbo.gen.S Jun 08 04:17:47 emeb_mac, can you post that...I'm out of windows here with having all the network stuff up Jun 08 04:17:51 er maybe not Jun 08 04:17:59 sure... Jun 08 04:18:23 even the Pandasystem is full Jun 08 04:18:45 I've got monitors on 11 differnt sites up now... Jun 08 04:18:52 in case anything blows up Jun 08 04:19:09 http://pastebin.com/R9yR6WZe Jun 08 04:19:54 that's presumably for the SPI LCD that's been used here before Jun 08 04:20:03 I actually bought one of those LCDs Jun 08 04:20:05 have it sitting here Jun 08 04:20:09 but I believe it's RX only Jun 08 04:20:13 so this may not solve my problem :) Jun 08 04:20:15 whole lotta DEADBEEF Jun 08 04:20:45 yeah and I don't see where it monkeys with the spi driver at all Jun 08 04:21:19 fragment 3 Jun 08 04:21:22 fragment@3 Jun 08 04:21:33 including the frequencies Jun 08 04:21:43 and polarities, etc. Jun 08 04:21:46 right Jun 08 04:21:52 ok Jun 08 04:22:10 see I'm still missing something fundamental Jun 08 04:22:26 wazzat? Jun 08 04:22:33 I dunno exactly Jun 08 04:22:41 the target is through me Jun 08 04:22:44 this is for the little tiny one Jun 08 04:22:48 see I thought the target selected the node to modify Jun 08 04:22:58 but wtf is 0xdeadbeef Jun 08 04:23:02 is it a node in the effective tree? Jun 08 04:23:09 it's not in mine Jun 08 04:23:19 seems to show up in all the .dts files. maybe just a placeholder? Jun 08 04:23:36 this thing even does PWM Jun 08 04:23:44 for backlight control Jun 08 04:23:46 for brightness prolly Jun 08 04:23:56 ya Jun 08 04:24:17 it's a pretty nice little display. I've got a couple - hooked to STM32F4 systems. Jun 08 04:24:18 so when you say "including the frequencies" Jun 08 04:24:23 all I see is the spi-max-frequency Jun 08 04:24:26 is that all it takes? Jun 08 04:24:39 cause mine has that Jun 08 04:24:41 dunno how important that is Jun 08 04:25:01 I played with that back in the boardfile days and it doesn't do much. Jun 08 04:25:19 plus spidev gives you access to that parameter. Jun 08 04:25:26 right Jun 08 04:25:41 unfortunately panto isn't here till monday Jun 08 04:25:52 slacker Jun 08 04:25:56 maybe mdp will be around over the weekend Jun 08 04:26:06 ya, he only puts in 20hrs a day Jun 08 04:26:29 ya - too bad someone hasn't done a simple example of plain-old spidev on spi1 Jun 08 04:26:33 that's all we need. Jun 08 04:26:45 hm he was here earlier Jun 08 04:26:48 emeb_mac: we need spi0 Jun 08 04:26:54 derp Jun 08 04:27:03 spidev is a dirty word I'm told... Jun 08 04:27:08 why? Jun 08 04:27:09 I'd rather do it that way Jun 08 04:27:13 dunno.... Jun 08 04:27:21 its kind of like the UIO driver for PRUSS Jun 08 04:27:23 lets you do the hard stuff in userspace Jun 08 04:27:23 UBE Jun 08 04:27:30 yes, I know Jun 08 04:27:42 so no futzing with driver load/unload/etc Jun 08 04:27:42 kind of like the UIO driver Jun 08 04:27:51 I'm looking at static void of_register_spi_devices(struct spi_master *master) Jun 08 04:28:11 it gives me some idea of what the magic words in the dts are Jun 08 04:28:33 we are all SpiMasters I guess Jun 08 04:28:43 are we? :( Jun 08 04:29:55 I guess I gotta crack the TRM again to look @ McSpi Jun 08 04:30:10 bah Jun 08 04:30:12 but not till after I make the switch @ 10pm Jun 08 04:30:25 works fine on my old kernel with boardfile setup. Jun 08 04:30:35 why is DT so much harder? Jun 08 04:30:41 its new Jun 08 04:30:57 and improved! :P Jun 08 04:31:24 it's just poorly documented Jun 08 04:31:38 there's a lot of *magic happens* Jun 08 04:31:40 "I'm not bad - I'm just drawn that way" Jun 08 04:31:44 exactly ;) Jun 08 04:31:54 exact step example of an existing dts would probably help with detailed explanation Jun 08 04:32:06 * m_billybob shrugs Jun 08 04:32:13 again with the exact steps. :) Jun 08 04:32:14 the problem is that a dts for spi is drastically different than an spi for something else Jun 08 04:32:21 er, dts for something else Jun 08 04:32:24 words are hard Jun 08 04:32:32 ow, my brain Jun 08 04:32:32 ssi, its kind of like sysfs instantiation...its magic and you don't want to know how. Jun 08 04:32:39 yeah Jun 08 04:32:47 heh Jun 08 04:32:57 so how about some softballs Jun 08 04:33:03 what exactly is address-cells and size-cells Jun 08 04:33:10 I am just a simple caveman Jun 08 04:33:19 your trees and fragments... confuse me Jun 08 04:33:25 I do not understand your modern ways Jun 08 04:33:31 i havent truely applied myself with DT yet. Once i do i bet "beagleboard and the 3.8 kernel" will be enough Jun 08 04:33:42 m_billybob: haha good luck with that Jun 08 04:34:01 i already understand much of it just from reading the doc once Jun 08 04:34:12 so tell me how to do it! Jun 08 04:34:12 there are examples of a lot of common things...but spi seems to be one that isn't quite there yet Jun 08 04:34:16 hah - you _think_ Jun 08 04:34:20 I didnt say i knew everything Jun 08 04:34:26 :) Jun 08 04:34:36 I thought I understood target Jun 08 04:34:41 until the 0xdeadbeef curveball :P Jun 08 04:35:01 the steps covered in the text though show you how to compile, load and unload a dtbo Jun 08 04:35:08 yeah that's the easy part Jun 08 04:35:11 so break out the bbq and make some burgers :P Jun 08 04:35:13 I've had that working for awhile Jun 08 04:35:16 #exactsteps Jun 08 04:35:21 although unloading always barfs Jun 08 04:35:28 not for me Jun 08 04:35:30 and requires a reboot :) Jun 08 04:35:35 unloading hdmi always fails Jun 08 04:35:44 and barfs Jun 08 04:36:18 although form what panto told me all i need to do is configure those pins before the hdmi overloy is loaded and all should be golgen Jun 08 04:36:26 overlay* Jun 08 04:36:29 yeah that's easy if you do an eeprom Jun 08 04:36:43 i've got that working as well... I can get my dtbo to load in one of the "hard" slots Jun 08 04:37:03 3: 57:P---L BCC,00A0,Ian,BB-SPI0-01 Jun 08 04:37:03 4: ff:P-O-L BB-BONE-RST2,00A0,Texas Instrument,BB-BONE-RST2 Jun 08 04:37:03 Jun 08 04:37:16 easypeasy Jun 08 04:37:21 i tried the optargs= in uEnv.txt that hasnt worked yet Jun 08 04:37:21 programming the eeprom is even trivially easy Jun 08 04:37:29 but i think that has to do with the distro im using Jun 08 04:37:45 hardest part was writing a c program with a stuct that had the right structure and writing it out to a file :P Jun 08 04:38:21 oh im not saying ill get it first go im sure it will give me pause for thought, thats already happened with DT for me already Jun 08 04:39:37 I think someone already said in here, if you're goingot write a "howto" do it while you're learning. otherwise you'll just be glad its over and never look back Jun 08 04:39:41 or somethign to that effect. Jun 08 04:39:57 concerning DT Jun 08 04:40:41 right now, im just releshing in the techno geek aspect of getting a nice working envoirnment setup Jun 08 04:40:56 * m_billybob loves techno geeky stuff Jun 08 04:43:11 ka6sox: is this thing more properly an omap2-mcspi or an omap4-mcspi? Jun 08 04:46:07 omap2-mcspi Jun 08 04:46:33 okay let me see if this event fires @ 2200 Jun 08 04:46:33 I think I found where the magic properties are defined Jun 08 04:46:42 "ti,spi-num-cs" Jun 08 04:46:48 "ti,pindir-d0-out-d1-in" Jun 08 04:46:58 :) Jun 08 04:47:29 looks liek those are the only omap2-mcspi specific properties :/ Jun 08 04:53:45 grrr Jun 08 04:54:38 so how do I know what clock polarity and phase I want? Jun 08 04:54:50 I guess from the fpga and/or flash datasheet? :P Jun 08 04:55:14 the FPGA doesnt' care, and the flash probably does Jun 08 04:56:49 These devices can be driven by a microcontroller with its serial peripheral interface (SPI) running in either of the following two SPI modes: Jun 08 04:56:52 • CPOL=0, CPHA=0 • CPOL=1, CPHA=1 Jun 08 04:59:11 interesting Jun 08 04:59:25 bb5m Jun 08 05:00:22 when I set up spi on the earlier kernel I didn't mess with the POL and PHA stuff. Jun 08 05:00:35 presumably that means it defaulted to something - maybe 0 & 0? Jun 08 05:00:43 ssi: sounds like you're making some progress, after you get it working I'll be motivated to buy a heat gun to spin my FPGA. Jun 08 05:03:13 first thing thats gone right all day Jun 08 05:03:54 emeb_mac: yes it defaults to 0,0 Jun 08 05:04:06 and if you look at teh dts I've been using: Jun 08 05:04:09 http://pastebin.com/m1GYGDbp Jun 08 05:04:20 spidev1.0 is set to cpha=1, cpol = 0 Jun 08 05:04:24 (which is wrong) Jun 08 05:04:27 and 1.1 is set to 0,0 Jun 08 05:04:32 now I understand the comment in the second one Jun 08 05:05:03 learning dts is a zen thing Jun 08 05:05:16 when you have the masters in front of you and you ask a question, your mind is not yet ready to accept the answer Jun 08 05:05:33 it is only by traveling the path alone that one can gain the appreciation for the wisdom that the masters have to bestow Jun 08 05:06:05 ssi, once you get the spi0 running I'll add the gpmc as I've been playing with that Jun 08 05:06:47 ka6sox: so it looks like perhaps you have two levels of node here Jun 08 05:07:06 there's the outer spi@0xaddress node, which is compatible="ti,omap4-mcspi" Jun 08 05:07:06 yes Jun 08 05:07:20 and then within that node, there's the spi0_0 node which is compatible="spidev" Jun 08 05:07:24 if not spidev, then what Jun 08 05:07:40 we aren't omap4 Jun 08 05:07:47 (not this week) Jun 08 05:07:49 :) Jun 08 05:08:03 my effective dts shows omap4 Jun 08 05:08:07 but that's here nor there, its the same driver Jun 08 05:08:19 okay lucky us... Jun 08 05:08:23 same all round Jun 08 05:09:01 hi powered my beaglebone with 5v 4Amp battery and now it's not booting at all Jun 08 05:09:07 is there a way to fix it Jun 08 05:09:39 i think the NCP349 stopping it from powering up Jun 08 05:13:48 #address-cells - number of cells required to define a chip select Jun 08 05:13:49 11 address on the SPI bus. Jun 08 05:13:52 so what the heck is a cell Jun 08 05:14:55 OH Jun 08 05:14:56 OH OH OH Jun 08 05:14:57 I get it Jun 08 05:15:07 the slave nodes are what's on the other end of the bus Jun 08 05:15:16 in the case of the adafruit lcd cape, the slave node was lcd0 Jun 08 05:15:23 and it's compatible = "adafruit,tft-lcd-1.8-red", "sitronix,st7735"; Jun 08 05:15:39 but in my case, the slave node is spi0_0, and compatible="spidev" Jun 08 05:15:49 because I just want a userspace device node that I can read/write Jun 08 05:16:07 the alternative would be if I were always writing to the flash, I might have a custom driver for the flash Jun 08 05:16:37 but for the BCC, I don't think we want to lock it into anything, because sometimes it's fpga configuration, sometimes it's flash, sometimes it's comms with the program on the fpga Jun 08 05:16:42 all without chipselects, so all one device Jun 08 05:18:29 ok, so then what does SW101 do? Jun 08 05:18:58 just sets the I2C address of the eeprom Jun 08 05:19:01 that sets the cape slot Jun 08 05:19:16 if both are off, you're address 0x54, so you're in slot 0 Jun 08 05:19:22 both on is address 0x57, slot 3 Jun 08 05:19:26 ok got that's I2C, unrelated to spi Jun 08 05:19:30 yep Jun 08 05:19:40 the kernel probes i2c at those four addresses and interrogates any eeproms it finds Jun 08 05:19:57 wow that person last 4 minutes before they bailed Jun 08 05:20:13 new record ? Jun 08 05:20:17 lol, I think he may have had a reverse polarity issue Jun 08 05:20:17 couldn't handle my DTS knowledge bomb Jun 08 05:20:22 * m_billybob really doesnt care Jun 08 05:20:46 so I'm understanding more and more of this Jun 08 05:20:52 not at you mastiff that person. Jun 08 05:20:53 but still not enough to know why my MISO doesn't work Jun 08 05:21:16 I guess 4 minutes is patient in some corners of the world. Jun 08 05:21:35 MISO ? SPI ? Jun 08 05:21:54 Master in slave out Jun 08 05:22:17 ya Jun 08 05:22:27 I can transmit but I can't receive Jun 08 05:22:31 thought you were doing I2C my mistake Jun 08 05:22:50 is your device meant to send ? Jun 08 05:22:57 yes Jun 08 05:23:20 ok so this is a screen ? Jun 08 05:23:25 no, it's a flash Jun 08 05:23:41 the programming procedure, the first t hing it does is interrogate the flash and ask for its device id Jun 08 05:23:45 and that fails Jun 08 05:24:09 I've traced it with a scope and confirmed that the line is toggling all the way back to P9 on the MISO line Jun 08 05:24:15 SPI0_D0 I think it is Jun 08 05:24:24 but the spidev never sees it Jun 08 05:24:26 reports all 0 Jun 08 05:24:41 Yeah im limited in my SPI device knowledge but you've got me in troubleshoot mode now lol Jun 08 05:24:56 I've already been through the whole hardware troubleshooting rigamarole Jun 08 05:25:04 I'm 99% certain it's the driver/DTO Jun 08 05:25:06 Was thinking some devices like the nokia5110 take data and command arguements Jun 08 05:25:57 so forget the devices for a second Jun 08 05:26:03 I should be able to loopback SPI Jun 08 05:26:06 connect MOSI to MISO Jun 08 05:26:12 yes Jun 08 05:26:14 and write to the device and be able to read back the echo Jun 08 05:26:16 and that doesn't work Jun 08 05:26:32 there's a spidev-test program in the kernel tree, and it writes a chunk of data to it, and I receive back all 0s Jun 08 05:26:35 so the RX is somehow messed up Jun 08 05:26:37 heh, loopback spi Jun 08 05:26:43 it's a circular buffer Jun 08 05:27:30 Russ: will that not work the way I'm expecting? Jun 08 05:28:03 I suppose it detects certain failure modes Jun 08 05:28:37 I think the spidev has an rx buffer and a tx buffer Jun 08 05:29:00 ie if I'm writing the contents of the device out to screen, and writing to the device, I don't necessarily see the same thing on both sides Jun 08 05:29:12 unless I'm looping it back Jun 08 05:29:31 you are toggling CS correctly ? Jun 08 05:29:38 ssi: your scope doesn't have spi trigger? Jun 08 05:29:58 mastiff: eh it might, but the frontpanel is getting flaky and it's a whore to use Jun 08 05:30:06 and it's ancient, so it probably doesn't Jun 08 05:30:09 hmm...suppose I haven't checked spru to see if there are two different shift registers Jun 08 05:30:18 I have a logic analyzer, but I don't think it has SPI trigger either Jun 08 05:30:23 still might be useful Jun 08 05:30:39 I just haven't gotten to the point of cleaning off a bench and dragging it out Jun 08 05:31:11 I'm hoping eric will get fed up with my bitching and get to figuring this out for me ;) Jun 08 05:31:37 well i know someone who has already loaded spidev via DT and it works so you have to be doing something wrong ;) Jun 08 05:31:54 so why not send me that dts Jun 08 05:31:57 dont *all* SPI devices require CS ? Jun 08 05:32:05 no Jun 08 05:32:07 My cheap usb MSO-19 has spi trigger. Tomorrow if you can get me up to speed with what you've done so far, I can try to help troubleshoot with that. Jun 08 05:32:41 You're just working with the BBB right now not the BCC cape yet? Jun 08 05:32:46 correct Jun 08 05:33:02 although I'm not 100% certain that the loopback thing would work Jun 08 05:33:08 it definitely doesnt work as is now Jun 08 05:33:23 ssi hold on, browser loading pages now . . Jun 08 05:33:45 m_billybob: I've seen several examples of people using SPI to talk to write-only devices Jun 08 05:33:48 I can do that now Jun 08 05:35:13 http://forum.beaglefu.com/index.php?/topic/80-beaglebone-black-drive-a-ws2811-led-strip-with-spi/ Jun 08 05:35:24 yeah then that may not help Jun 08 05:36:33 Ah, dding directly to block devices makes me happy Jun 08 05:36:39 livin on the edge! Jun 08 05:37:02 m_billybob: yea he's not even pinmuxing the D0 line Jun 08 05:37:08 or CS for that matter Jun 08 05:37:22 no idea i havent read through his code fully yet Jun 08 05:39:28 * Shadyman sings Jun 08 05:39:36 You can't help that pin from faaaaaaallinnnn' Jun 08 05:42:38 it may be time to bust out the logic analyzer Jun 08 05:42:46 verify that what's actually on the MISO line is valid Jun 08 05:43:17 i used Narcissus build bbxm image. I attached LI-5M03 to bbxm. I can cat gain_val. cat basic_reg_val returns error -121 Jun 08 05:46:23 root@beagleboard:/sys/devices/platform/i2c_omap.2/i2c-2/2-0048# cat ./basic_reg_ Jun 08 05:46:23 addr Jun 08 05:46:23 0x8 Jun 08 05:46:23 root@beagleboard:/sys/devices/platform/i2c_omap.2/i2c-2/2-0048# cat ./basic_reg_ Jun 08 05:46:23 val Jun 08 05:46:24 [ 2505.748626] mt9p031 2-0048: read from offset 0x8 error -121 Jun 08 05:46:24 [ 2505.755432] mt9p031: Basic register read failedroot@beagleboard:/sys/devices/platform/i2c_omap.2/i2c-2/2-0048# Jun 08 05:46:32 anyone can help me? Jun 08 08:04:46 apparently if you get too far outside the speed range of the SPI interface it locks up Jun 08 08:11:00 hello Jun 08 08:11:01 hello Jun 08 08:30:06 hello Jun 08 08:30:07 hello Jun 08 08:40:23 <_av500_> hell Jun 08 08:40:24 <_av500_> o Jun 08 08:42:13 heaven Jun 08 08:42:14 o Jun 08 08:43:27 so freakin frustrated with this SPI mess :( Jun 08 08:43:36 oh. Jun 08 08:43:42 i'm about to start work on usiogn SPI Jun 08 08:43:51 i want to talk the FPGA on teh BCC cape Jun 08 08:43:57 that's what I'm working on Jun 08 08:44:04 I can configure the fpga over spi Jun 08 08:44:05 that works fine Jun 08 08:44:10 but I can't read anything back over SPI Jun 08 08:44:17 which means I can't program the config flash Jun 08 08:44:31 because it talks bidirectionally to the flash, and has to confirm its device id before it can program Jun 08 08:44:32 oh. Jun 08 08:45:04 hm I wonder if it's something extra stupid like a blown up pin on the am3359 Jun 08 08:45:04 ssi, your usin the BCC cape? Jun 08 08:45:06 yes Jun 08 08:45:22 i got mine assembled toay Jun 08 08:46:16 so grab eric's bcc_tool and my dto Jun 08 08:46:38 https://gist.github.com/ianmcmahon/5732144 Jun 08 09:12:07 ssi, what OS are you running on the BBB? Jun 08 09:12:20 debian Jun 08 09:12:24 oh cool Jun 08 09:12:30 3.8 kernel Jun 08 09:12:36 i probabl will need some help Jun 08 09:12:37 :-) Jun 08 09:12:48 which which part Jun 08 09:16:52 <_Sy_> Morning. Jun 08 09:56:44 morning Jun 08 10:00:44 ssi: are you using dev/spi ? Jun 08 10:02:22 have you tried the loopback test? Jun 08 10:10:55 * mranostay zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz Jun 08 11:35:35 hey, is it possible to install freebsd on the built in storage of the Beaglebone Black? Jun 08 11:35:50 I currently only now it for an sd card Jun 08 11:36:23 from a software point of view the emmc is just an sd card glued to the board Jun 08 11:36:24 you can think of the eMMC as a non-removable SD Jun 08 11:38:00 ok so it wont be any problem at all to just use the emmc fd instead of the sd one? There is no chance of "breaking" the board Jun 08 11:38:01 ? Jun 08 11:38:37 you can always just boot it from SD Jun 08 11:39:33 ok thank you for the advice Jun 08 11:39:48 btw. would you use a "normal" filesystem like ext2 for eMMC or a flash filesystem like ubifs? Jun 08 11:40:10 I would use the bsd one Jun 08 11:40:33 there's no need for a special filesystem Jun 08 11:40:49 ufs2 - zfs would be overpowered Jun 08 11:41:18 I am going to start a native nodejs addon library is there any interest on your side? Jun 08 11:41:21 UBIFS is for raw flash devices. Jun 08 11:41:30 emmc is not a raw flash device. Jun 08 11:41:40 It has a little 32 bit processor inside, managing things. Jun 08 11:41:50 ah, ok. good to know. Jun 08 11:41:53 who says it's 32-bit? Jun 08 11:42:03 mru: I vaguely remember some teardowns Jun 08 11:42:10 Including decapping Jun 08 11:42:17 so the processor takes care of the wear levelling? Jun 08 11:42:17 what's important is that there's a controller Jun 08 11:42:21 that does wear levellign Jun 08 11:42:27 :) Jun 08 11:43:15 mru: Yes. In an undocumented and untransparent manner. Jun 08 11:43:37 that's beside the point Jun 08 11:43:47 Fun issues arise with concealed wear leveling. Jun 08 11:43:48 doing your own on top won't help Jun 08 11:44:09 It at best seriously complicates it. Jun 08 12:01:51 when i go to the usb site, 192.168.7.2 with my browser, and am greeted with the startup page, where do those files live, and in Angstrom, what is the webserver that serves them? Jun 08 12:10:42 <_av500_> ps might tell you Jun 08 12:14:14 what i am trying to do is copy those files from one of the imgs i have downloaded. i have mounted, and chrooted to the Anstrom/Beagle img. So, the machine isn't actually running. Jun 08 12:16:09 <_av500_> see the services that systemd starts Jun 08 12:16:13 <_av500_> one should be a web server Jun 08 12:16:28 good idea. thanks Jun 08 12:22:22 <_Sy_> Hello koen ? Jun 08 12:24:12 or i suppose i could have just googled it: http://stackoverflow.com/questions/10818459/where-is-the-web-server-root-directory-on-angstrom-linux-beagleboard-beaglebo lol Jun 08 13:09:22 hi Jun 08 13:18:33 lainwir3d: you watched too many animes! Jun 08 13:18:48 lainwir3d: you should go out instead, and enjoy the sun! Jun 08 13:19:52 omg Jun 08 13:20:14 I have been discovered Jun 08 13:20:25 well after i manage to get this led to work Jun 08 13:20:35 maybe i'll go out and enjoy the sun ;) Jun 08 13:21:38 <_av500_> you 4 blue leds on the BBB, no ned for the sun Jun 08 13:21:45 <_av500_> you have* Jun 08 13:22:12 you BBB, no need for a verb Jun 08 13:23:08 i have the BB, so it's green leds Jun 08 13:23:18 :'( Jun 08 13:23:23 blue ftw Jun 08 13:23:54 if you want to build a navi, you have to buy a BBB Jun 08 13:23:57 a BB wont do Jun 08 13:24:19 yeah, it should have been ordered yesterday Jun 08 13:24:22 one or two Jun 08 13:24:23 ;) Jun 08 13:24:30 and dont forget the cooling system! Jun 08 13:24:46 i determined its a socket server, but still cant find the files. Jun 08 14:02:48 <_Sy_> Hello, I want to delay the start up of a service (mySQL) to allow the system NTP service to acquire the clock, how would I do this? Jun 08 14:13:26 <_Sy_> Is it possible to delay the start up of a service to allow enough time for another service to start up? Jun 08 14:14:26 make one depend on the other Jun 08 14:14:54 <_Sy_> how do you do that? Jun 08 14:15:24 <_Sy_> I would like to hold the startup of mysqld until ntpdate has acquired the current time/date Jun 08 14:21:38 _Sy_: have you looked at systemd documentation or configuration files? Jun 08 14:34:55 <_Sy_> I've done a few searches online, but didn't find anything, will try again. Jun 08 15:02:28 <_Sy_> Just downloaded the Angstrom Manual, no mention in it of systemd. Jun 08 15:25:51 lol, i see why angstrom fails to opkg upgrade Jun 08 15:26:03 it fills up /tmp with ipks and doesnt delete them till the end Jun 08 15:26:41 koen, ^^^ Jun 08 15:29:13 oh it extracts them in there also Jun 08 15:29:37 lol, deleted chromium and firefox once installed, so it should actually finish .. lets see if i caught it in time though Jun 08 15:36:35 Crofton|work, http://bugs.elinux.org/issues/13 Jun 08 15:37:56 thats not the solution Jun 08 15:37:59 there is a better solution Jun 08 15:38:06 upgrade firefox and chromium _first, then purge /tmp Jun 08 15:38:34 morn ka6sox Jun 08 15:38:41 maxinux, when people come here...they have *already* done "opkg upgrade" Jun 08 15:38:59 yes, but this fix could also be implemented in ipkg Jun 08 15:39:04 and they have already gotten the system in an unstable state Jun 08 15:39:39 unstable, unknown, 1/2baked state Jun 08 15:39:50 ka6sox: so it occurred to me last night.... maybe the SPI0_D0 pin on this board is just dead for some reason Jun 08 15:39:59 morning ssi Jun 08 15:40:18 easy enough to check, right? I have a second bbb Jun 08 15:40:27 but the second one won't boot correctly for some reason Jun 08 15:40:28 yes, thats one way :0 Jun 08 15:40:38 it'll boot a bbb flasher image off SD Jun 08 15:40:44 but it won't boot my debian flash Jun 08 15:40:51 hmmmm Jun 08 15:40:58 and I let angstrom fully install to emmc overnight Jun 08 15:41:00 the 6/6 image Jun 08 15:41:05 and it won't boot off emmc either Jun 08 15:41:06 good Jun 08 15:41:08 there is a 5/5 image? damnit Jun 08 15:41:10 err 6/6 Jun 08 15:41:11 I get CCCCCCCC from emmc Jun 08 15:41:12 ssi: can you set that pin as a GPIO and check it that way? Jun 08 15:42:03 and I get Did not find a recognized configuration, assuming General purpose EVM in Profile 0 with Daughter board Jun 08 15:42:06 from my debian SD Jun 08 15:42:12 mastiff: yeah I can Jun 08 15:42:15 mastiff: it's just more work :) Jun 08 15:42:56 it'd be super easy to just try it on the other board, but the other board is giving me fits :( Jun 08 15:43:24 got a bootable sd handy? Jun 08 15:43:29 yeah several Jun 08 15:43:33 could boot from an sd instead of the emmc Jun 08 15:43:49 booting from angstrom sd works Jun 08 15:43:52 from debian, I get: Did not find a recognized configuration, assuming General purpose EVM in Profile 0 with Daughter board Jun 08 15:44:00 the debian sd that works on one board fails on the other Jun 08 15:44:16 the debian sd which is where dtc and all my code lives Jun 08 15:44:28 angstrom seems to still not ship with dtc :/ Jun 08 15:44:54 dtc? Jun 08 15:45:02 ssi, really? Jun 08 15:45:02 angstrom is pretty much crap anyways Jun 08 15:45:03 huh? dtc is on my angstrom Jun 08 15:45:15 its on mine too Jun 08 15:45:16 where is it? Jun 08 15:45:27 which dtc Jun 08 15:45:33 whoops Jun 08 15:45:36 hehe Jun 08 15:46:19 my son is working with it now...he has it offline for some reason...lemme ask him to hook it back up. Jun 08 15:46:46 mastiff? Jun 08 15:46:47 it's in PATH so just typing dtc runs it, not sure the file location Jun 08 15:46:54 type "which dtc" Jun 08 15:47:22 Jun 08 15:47:34 heh yes that part is important :P Jun 08 15:47:53 onoes we segfaulted his brain Jun 08 15:48:14 I wish they put a build date in etc/issue or something Jun 08 15:48:22 they do Jun 08 15:48:28 2012.12? Jun 08 15:48:32 /etc/dogtag Jun 08 15:48:32 :P Jun 08 15:48:33 oh Jun 08 15:48:40 I don't have that Jun 08 15:48:48 debian? Jun 08 15:48:49 usr/bin/dtc Jun 08 15:48:54 no, angstrom Jun 08 15:49:00 wierd... Jun 08 15:49:01 okay Jun 08 15:49:08 yeah I have no /usr/bin/dtc Jun 08 15:49:11 no dtc on path Jun 08 15:49:29 what did you last flash into it? Jun 08 15:49:33 angstrom 6.6 Jun 08 15:50:04 on monday I'll see what we can do to make sure both of those are in there. Jun 08 15:50:22 on future images Jun 08 15:50:24 oh well I gotta run out for a bit Jun 08 15:50:26 back later :D Jun 08 15:50:30 laters ssi Jun 08 15:50:43 me too, gotta work on the sailboat hehe Jun 08 15:50:53 mastiff, enjoy Jun 08 15:51:00 later Jun 08 18:13:43 panto's connection is weird Jun 08 18:13:52 did the greeks forget to pay the internet bill Jun 08 18:16:17 austerity ya know Jun 08 18:26:26 <_Sy_> Can anyone help me...I've been round and round the same loop over and over, trying to get the toolchain installed in ubuntu linux virtual box, so that I can compile a simple hello world test application....but even from the folder where the gcc file exists it still says "No such file or directory", I've checked permissions and everything should be ok. Jun 08 18:26:51 <_Sy_> I installed the toolchain as per the instructions on http://www.eewiki.net/display/linuxonarm/BeagleBone+Black?showChildren=false Jun 08 18:27:40 <_Sy_> to test I'm in the same folder as the compiler and executing "./arm-linux-gnueabihf-gcc" results in: Jun 08 18:27:59 <_Sy_> bash: ./arm-linux-gnueabihf-gcc: No such file or directory Jun 08 18:28:36 <_Sy_> the permissions on the file are: lrwxrwxrwx Jun 08 18:30:29 <_Sy_> anyone? Jun 08 18:31:58 <_Sy_> Am I the only one here? Jun 08 18:32:57 _Sy_: could it be possible that this file is only a link? Jun 08 18:33:39 <_Sy_> It is a link to arm-linux-gnueabihf-gcc-4.7.3 Jun 08 18:34:03 <_Sy_> I've checked the file it links to and even if I run it directly it does the same thing, permissions on the file it links to are ok too. Jun 08 18:38:57 <_Sy_> much as I'd really like to get the set-up working so I can cross compile and remotely debug, its taking so long to get what should be pretty simple stuff working that I'm going to have to develop locally on the BBB. Jun 08 18:42:49 just don't hook it up to 12v Jun 08 18:48:35 ok will use the 9V Jun 08 19:08:34 I'm gonna go buy a BBW today Jun 08 19:08:38 maybe that will solve all my problems :P Jun 08 19:14:59 ssi: giving up on DT? Jun 08 19:20:55 monday we should talk to panto and get this resolved Jun 08 19:21:13 because he wants this to work.... Jun 08 19:22:16 ka6sox: hello kitty day to you! Jun 08 19:23:07 where is your picture....oh wait...its still early for you...you post @ 11:59:59 usually Jun 08 19:25:13 ssi: ka6sox: I guess I don't see how getting a BBW solves the problem Jun 08 19:25:14 mranostay, just stop it Jun 08 19:25:29 because it will work with a 3.2 kernel Jun 08 19:25:43 ok Jun 08 19:25:48 because you already made it work with a 3.2 kernel. Jun 08 19:26:13 ya, but then we're stuck in the past Jun 08 19:26:35 the solution is to let panto help us make this work. Jun 08 19:26:43 :) Jun 08 19:27:07 once we understand it then it may be better than sysfs instantiation.... Jun 08 19:27:13 help us obi-wan panto, you're our only hope. Jun 08 19:27:16 but I doubt it. Jun 08 19:28:23 ssi, were you able to get it to programme? Jun 08 19:28:36 <_Sy_> I've been looking for information on systemd and angstrom, I want to set-up a service (mysqld) to wait until ntpdate has acquired a date, because when mysqld starts with no date, the synchronisation gets screwed up...but so far I'm struggling to find anything helpful. Jun 08 19:30:58 it feels like "you put a <1> in that field....nobody knows why but it doesn't work till you do" Jun 08 19:37:16 emeb_mac, my problem is I don't know spidev Jun 08 19:37:29 so I'm no help at all :( Jun 08 19:37:33 ka6sox: what's to know? Jun 08 19:37:53 it's a userspace interface to the underlying unique spi hardware drivers. Jun 08 19:38:03 fairly simple api Jun 08 19:38:26 its not the api, its how to connect it up via DT Jun 08 19:38:34 right Jun 08 19:38:51 what we don't know is how DT wants spidev to be defined Jun 08 19:38:59 Bingo! Jun 08 19:39:54 of course there may also be issues with the omap spi drivers for BBB Jun 08 19:40:14 since ssi said there appear to be both tx-only and txrx options Jun 08 19:40:37 (in the newer kernels) Jun 08 19:44:38 ka6sox, in older kernel releases (prior to dt overlay support), there was a spidev example in the base dts Jun 08 19:46:12 like the 3.7 version? Jun 08 19:46:23 * ka6sox looks Jun 08 19:46:25 yes, 3.6 or 3.7 Jun 08 19:46:32 kk Jun 08 19:48:01 <_Sy_> Just been reading the Angstrom manual on http://www.linuxtogo.org/gowiki/AngstromManual Jun 08 19:48:16 that'll rot yer brain Jun 08 19:48:40 <_Sy_> The only locale that appears to in the image is locale-base-en-us, I tried setting opkg install locale-base-en-gb, but it isn't there Jun 08 19:49:09 ssi, https://github.com/beagleboard/kernel/blob/3.7/patches/spi/0001-spi-spidev-Add-device-tree-bindings.patch Jun 08 19:49:50 <_Sy_> I was hopping to find information on the start-up scripts and how to create a dependancy but there isn't much in the manual at all, and nothing to explain what I'm after. Jun 08 19:50:27 this your brain on DT Jun 08 19:50:51 _Sy_: have you searched the net for the documentation on systemd? Jun 08 19:51:42 <_Sy_> yes, there is a bit, but nothing angstrom specific...I had a look at several of the pages, but couldn't find anything about creating a dependancy Jun 08 19:51:43 mdp, thanks, that might help us..its not in the 3.8 stuff Jun 08 19:51:53 _Sy_: well, it's not angstrom specific knowledge Jun 08 19:51:56 I'll have to see if that patch is already in 3.8 Jun 08 19:53:32 ka6sox, did you guys read the manual? ;) Jun 08 19:53:48 which "the manual" :) Jun 08 19:53:59 the one panto wrote? Jun 08 19:54:01 <_Sy_> I don't have much linux knowledge Jun 08 19:54:13 no, panto's manual isn't the generic DT stuff Jun 08 19:54:25 it's pretty much specific to dynamic overlays on BBB Jun 08 19:54:47 <_Sy_> So is systemd the service manager that is used in Angstrom? Jun 08 19:54:54 ka6sox, extracted from a branch where I replaced it for my st7735 display: http://pastebin.com/hM4JAy4t Jun 08 19:56:08 that looks similar to what ssi is doing...jas Jun 08 19:56:32 http://pastebin.com/m1GYGDbp Jun 08 19:58:02 fwiw, works for me Jun 08 19:58:03 the manual is the binding description: http://www.mjmwired.net/kernel/Documentation/devicetree/bindings/spi/spi-bus.txt Jun 08 19:58:54 * m_billybob hides from the evil DT code pastes Jun 08 20:00:14 mdp, the paste that I sent you or something else? Jun 08 20:00:24 (what works I mean) ;) Jun 08 20:01:05 ka6sox, the thing that I believe is unwritten is that the old driver modalias stuff for spi slave drivers will be matched from the compatible string Jun 08 20:01:22 so, that's why you put "spidev" in there and how it gets probed Jun 08 20:01:36 k Jun 08 20:01:58 if it doesn't work in dynamic dt land, then there's something specific to that broken Jun 08 20:03:27 the semantics are already different from upstream in that it doesn't require the node to be marked "okay"..not sure what would be wrong in that fragment if it doesn't even probe the spidev driver Jun 08 20:06:39 hmm there is no spidev dt fragment? Jun 08 20:07:20 not in 3.8 Jun 08 20:07:47 let me see if that patch is in there already in 3.8 Jun 08 20:08:02 if not then I'll forward "port" that patch to 3.8 Jun 08 20:08:27 ka6sox, this is one of those "basic" fragments that should be in the rootfs for both spi0/1 IMHO Jun 08 20:08:46 ka6sox, I remember in 3.7 that koen specifically had that in the base dts to make life easier for people Jun 08 20:09:03 it got taken out of 3.8 for some reason Jun 08 20:09:17 So im kind of listening in here, the idea is that the spidev kernel module needs to be loaded ( properly configured ) as well as having the DT overlay setup proerly ? Jun 08 20:09:25 spi0 won't conflict with the HDMI Jun 08 20:09:41 and our use case doesn't cause an issue Jun 08 20:09:44 spi1 conflicts though? Jun 08 20:09:51 HDMI Jun 08 20:10:23 I wish I could remove hdmi from mine...no use to me Jun 08 20:10:50 mranostay the spidev kernel module is even placed in */ignore/* at least on the cross dev platform im using. Jun 08 20:10:52 mdp: +1 Jun 08 20:11:10 and wow apparently im lagging. Jun 08 20:11:40 mdp, we just need a overlay that takes precedence over the onboard and we are golden Jun 08 20:11:49 it will just "go away" Jun 08 20:13:02 Crofton: welcome aboard Jun 08 20:14:11 ka6sox, at some point I'll probably have to learn the practical aspects of overlays...it's hard to follow as it conflicts with the dayjob upstream world ;) Jun 08 20:14:30 mdp, understood Jun 08 20:14:48 but the arduino stuff requires that I grok it at some point ;) Jun 08 20:18:19 lets see, I have the patched source someplace around here. Jun 08 20:25:25 mdp, the 3.8 source has about 1/2 the patch right but only talks about a rohm device not .compatible = linux,spidev Jun 08 20:31:36 yay - recompile yer kenels! Jun 08 20:32:08 emeb_mac: I bought a BBW Jun 08 20:32:23 ssi: cheater. :) Jun 08 20:32:27 haha Jun 08 20:32:31 I need a known reference point Jun 08 20:32:38 good idea Jun 08 20:32:45 bbw is more useful anyway :P Jun 08 20:37:20 mdp: when you say that spi is working for you, are you able to do bidirectional transfers? Jun 08 20:37:33 mdp: it looked like you said you're using an LCD, and I figure there's lots of people using write-only devices Jun 08 20:37:40 I can write fine, I just can't read Jun 08 20:37:46 I get all 0x00 or all 0xFF Jun 08 20:37:54 and I've seen reports of other people having the same problem Jun 08 20:38:03 spidev_test in the kernel documentation tree doesn't work for me Jun 08 20:38:47 does anybody know whether XBMC works on BB ? Jun 08 20:41:59 azdem: which hw? Jun 08 20:44:02 BB black Jun 08 20:46:00 I mean whether someone has already ported it? R-Pi seems to have a ready to use Linux image with XBMC emphasis Jun 08 20:47:22 xbmc is the one thing rpi can actually do Jun 08 20:47:28 might as well use the rpi for it :D Jun 08 20:50:36 well the RPi is a media decoder with a tiny ARM core latched to it Jun 08 20:50:58 exactly Jun 08 20:53:53 j Jun 08 20:54:09 How can I see the spec of hardware? Jun 08 20:54:16 I found only the wiki Jun 08 20:55:01 Manual & google Jun 08 20:55:06 SRM Jun 08 20:55:31 I want to see the schematics Jun 08 20:55:37 is it available ? Jun 08 20:56:09 ugh I wish esx would put vms on the autostart list by default Jun 08 20:56:13 that always catches me :( Jun 08 20:56:35 zozo: should be there as well Jun 08 20:57:17 There are also docs at TI Jun 08 20:58:10 bleh, patching the 3.7 kernel now :P Jun 08 20:58:57 Someone had serious audio issues with bbb? Jun 08 20:58:57 ka6sox: my machine where I built my kernel has been down since my virt server rebooted apparently Jun 08 20:59:00 should be coming up now Jun 08 20:59:54 ssi, ya, switch to the 3.8 branch, Jun 08 21:00:15 I am going to figure out what we need to make that spidev.c patch more "generic" Jun 08 21:00:24 right now its specific to some rohm part Jun 08 21:00:25 :p[ Jun 08 21:03:48 I'm cloning my linuxcnc working image Jun 08 21:03:52 32g card to 32g card Jun 08 21:03:54 might take a minute Jun 08 21:03:55 heh Jun 08 21:04:03 eeek Jun 08 21:04:09 minute. hah Jun 08 21:04:20 ssi, so still rough going with spi rx ? Jun 08 21:04:37 yea no success yet Jun 08 21:04:43 considering how long it took to unzip the 5-27 image onto a 4G card last night... Jun 08 21:04:45 bummer :( Jun 08 21:04:58 emeb_mac: yeah I kicked it off because I need to go the store anyway Jun 08 21:05:03 I'll get this kernel building too before I go :) Jun 08 21:05:38 fwiw, ive got exact steps for iscsi /home, working on iscsi /root atm Jun 08 21:05:40 shouldn't take too long Jun 08 21:06:21 dont know about you all, but for me iscsi over 10/100 ethernet == twice as fast as uSD Jun 08 21:06:31 atleast for writes Jun 08 21:07:46 ka6sox: so I don't really grok what this patch in the 3.7 branch is exactly Jun 08 21:08:06 "This will allow to probe spidev from device tree Jun 08 21:08:07 its patching spidev to add dt bindings Jun 08 21:08:10 can't we do that already? Jun 08 21:08:15 not really Jun 08 21:08:20 my spidev works, it just only works one direction Jun 08 21:08:37 the patch is specific to a rohm device Jun 08 21:08:39 not generic Jun 08 21:08:48 so it doesn't surprise me Jun 08 21:09:38 oh this is that kernel tree with eight hundred thousand hours of patches Jun 08 21:09:43 I got off this tree for a reason Jun 08 21:10:00 we need to be *on* this tree :) Jun 08 21:10:19 I don't think I ever got it to work Jun 08 21:10:26 but it sure took its sweet time not working Jun 08 21:12:23 doesn't surprise me Jun 08 21:12:34 I'm getting the 3.7 tree and patching that too Jun 08 21:13:03 hmm 6/6 image seems borked Jun 08 21:13:19 or they changed something that cause X to fail to launch in my setup that worked on older images Jun 08 21:14:15 alright so I have the dd from hell going, and the patchset from hell going Jun 08 21:14:23 I'm gonna go to the store and get some ice so I can come home and get drunk. Jun 08 21:14:32 back in a flash Jun 08 21:18:18 I'm sure mranostay would drink to that Jun 08 21:47:09 could not login via ssh to the beaglebone black. Try to "ssh-keyscan 192.168.7.2" shows no keys. Any ideas to solve this problem? Jun 08 21:50:31 somebody in this channel?!? Jun 08 21:50:45 yes what is your error Jun 08 21:51:15 your error when you ssh in Jun 08 21:52:24 i try to "ssh 192.168.7.2", but the connection could not be established. The same the the "gateone". Jun 08 21:52:57 ssh-keyscan 192.168.7.2 shows that there is a problem with the "ssh-key-exchange" Jun 08 21:53:20 are you using the latest image? have you been able to ssh in before? Jun 08 21:53:37 ssh-keyscan -vv 192.168.7.2 debug2: fd 3 setting O_NONBLOCK 192.168.7.2: Connection closed by remote host Jun 08 21:54:15 my first steps with the beagleboard, got it 2 days ago Jun 08 21:54:36 dont use keyscan Jun 08 21:54:49 what else? Jun 08 21:54:50 flash the board to the latest angstrom image, that will be a good start Jun 08 21:55:14 ssh -vvv Jun 08 21:55:30 but it sounds like the image is broken and ssh is not doing anything Jun 08 21:55:39 and new image would be a wise step Jun 08 21:55:50 if you did opkg update it broke your image probably Jun 08 21:56:26 yes I remember opkg update causing a lot of problems for people Jun 08 21:57:02 the system is still in the "factory default state" :-( Jun 08 21:58:09 Hmm, doing non-trivial homework for someone = essentially, 'stealing' their education ;) Jun 08 21:59:02 ssh -vvv gives you only the same hint "ssh_exchange_identification: Connection closed by remote host" like the ssh-keyscan Jun 08 22:01:28 will flash the board tomorrow, hope that this will fix the login errors Jun 08 22:02:34 good plan Jun 08 22:04:32 back Jun 08 22:04:41 dd is still running :P Jun 08 22:07:01 has anyone had success toggling pins P8.27-P8.46. I am trying setup gpio pins using the terminal. Some pins I can export with # echo [PinNumber] > /sys/class/gpio/export yet this does not work for any of the pins P8.27-P8.46. Any ideas why I can't access them? I assume another resource has control of them.. Jun 08 22:29:15 MH_, a LOT of the P8 pins are owned by things like the DHIM and eMMC Jun 08 22:29:24 HDMI Jun 08 22:29:27 bleh Jun 08 22:30:52 http://www.elinux.org/BeagleBone_Black_Capes Jun 08 22:31:00 scroll down to the table Jun 08 22:31:06 ka6sox yea I saw in the schematic where P8.3-P8.25 says "Caution: used on board". Much of the remaining pins on P8 I think are in a lcd_data mode Jun 08 22:31:08 that shows the pins Jun 08 22:31:29 ssi: i was just looking at your ssi dts Jun 08 22:31:31 the pins are used by eMMC and HDMI(the LCD ones) Jun 08 22:31:54 any reason you are using spi0_0 for spi mode 1 and spi0_1 for spi mode 0? Jun 08 22:35:35 Rickta59: no reason whatsoever Jun 08 22:35:48 maybe you could add spi mode 2 and 3 too Jun 08 22:35:51 I started with someone else's stuff Jun 08 22:35:53 and I've since changed it Jun 08 22:36:07 but my impression is that spi0_1 would also use CS1 instead of CS0 Jun 08 22:36:12 and CS1 isn't broken out on the headers Jun 08 22:36:14 ka6sox yea I was thinking they were, I will go back to searching Jun 08 22:36:14 but I'm not sure about that Jun 08 22:36:48 the pullups mess with someone who wasnt to use mode 1 Jun 08 22:36:53 who wants Jun 08 22:37:04 I need to use mode 0 Jun 08 22:37:19 well I can use either pol 0, pha 0 or pol 1 pha 1 Jun 08 22:37:21 so 0 or 3 Jun 08 22:37:32 what are interacting with? Jun 08 22:38:05 a spi configuration flash on an fpga cape Jun 08 22:39:58 does it work? Jun 08 22:40:19 not exactly? Jun 08 22:40:26 i was thinking about wiring up one of those microchip 23k256 chips and seeing if i could get it going Jun 08 22:40:26 I can configure the fpga via SPI, that works fine Jun 08 22:40:45 when I try to program the flash, the first thing the tool does is interrogate the flash and ask for device id Jun 08 22:40:49 and that returns all zeros Jun 08 22:40:50 ssi, I think the spidev driver is hozed Jun 08 22:40:58 when I started digging into it, I realised that I have no RX at all Jun 08 22:41:00 ka6sox: how so? Jun 08 22:41:06 the compaible line reads: { .compatible = "rohm,dh2228fv" }, Jun 08 22:41:31 the main spidev driver? Jun 08 22:41:36 the compaible line reads: { .compatible = "linux,spidev" }, Jun 08 22:41:39 yes Jun 08 22:41:42 spidev.c Jun 08 22:41:48 so it does Jun 08 22:41:49 heh Jun 08 22:42:14 compare that to: Jun 08 22:42:24 I can't get on mine to see what my version of the kernel spidev had because I'm still doing that massive dd Jun 08 22:42:33 I wish dd had a progress bar :P Jun 08 22:43:01 https://github.com/beagleboard/kernel/blob/3.7/patches/spi/0001-spi-spidev-Add-device-tree-bindings.patch Jun 08 22:43:16 yea I follow you now Jun 08 22:43:49 so mainline was polluted :P Jun 08 22:44:05 https://patchwork.kernel.org/patch/1678101/ Jun 08 22:44:05 there Jun 08 22:44:26 I think the {}, at the end of that array is a catchall though Jun 08 22:45:12 EVT stuff making it into mainline without checking other use cases :P Jun 08 22:46:05 bb20m Jun 08 22:48:12 Is the angstrom linux 3d accelerated? how do I get glxgears to run? Jun 08 22:48:59 so I spent all that time and money running around getting bar supplies Jun 08 22:49:02 and I forgot to get dinner Jun 08 22:50:14 Rickta59: so yea my current spi overlay just has spi0_0, and sets it as mode 0 Jun 08 22:50:45 it would be nice if all that stuff with all modes was in the standard distro Jun 08 22:51:12 lots of things would be nice :P Jun 08 22:51:13 i don't know how you deal with the pull up pull downs though Jun 08 22:51:46 they're pinmuxable Jun 08 22:51:50 i tried unloading one and it crashed Jun 08 22:52:05 I haven't had any luck unloading overlays Jun 08 22:52:08 always crashes capemgc Jun 08 22:53:52 I wanna see if the BBW SPI1 port loopback works out of the box Jun 08 22:54:07 have you disabled your hdmi port? Jun 08 22:54:13 no Jun 08 22:54:23 well... sorta Jun 08 22:54:26 to get iat it i had to do that Jun 08 22:54:33 spi1 that is Jun 08 22:54:34 on the bbb I put a line in uenv to disable it Jun 08 22:54:42 yeah that is what i did Jun 08 22:54:43 but I'm using spi0 Jun 08 22:54:53 on the bbw, I'm about to test spi1, but there's no hdmi Jun 08 22:55:28 ssi: same problem, have been enjoying bar supplies and forgot to eat, now making crabcakes ;) Jun 08 22:55:40 I may order a pizza :) Jun 08 22:55:54 today's endeavor is to try to replicate longhorn's "perfect margarita" Jun 08 22:55:56 because it's damn good Jun 08 22:56:02 my first try is close, but not quite there Jun 08 22:56:57 aw crap, there's no spidev-test on BBW Jun 08 22:58:16 grab the source from the kernel source tree & compile Jun 08 22:58:29 will the version out of the 3.8 tree work? Jun 08 22:58:53 probably - doubt that it's changed much over the years. Jun 08 22:59:56 emeb_mac: make my dd finish! :( Jun 08 23:00:27 taking a long time? Jun 08 23:00:33 yea a bit Jun 08 23:00:47 there's a version of dd out there that gives a running progress bar Jun 08 23:00:55 I started it about 2 hours ago Jun 08 23:00:59 that's handy on these long SD writes Jun 08 23:01:48 according to disk activity in osx, I'm transferring at 5MB/sec to each card Jun 08 23:01:55 exactly 5.00 Jun 08 23:01:58 I wonder where that limit's coming from Jun 08 23:02:07 surely the cards are faster than that? Jun 08 23:02:08 maybe not Jun 08 23:02:58 maybe your USB -> SD card writer is fullspeed? Jun 08 23:03:05 not highspeed Jun 08 23:03:13 dunno Jun 08 23:03:28 I have one in my sd slot in a usd sd adapter Jun 08 23:03:36 the ther is in a usb adapter in my keyboard's port :P Jun 08 23:03:38 so god knows Jun 08 23:04:49 still - at 5MB/s you should take about 3min for 1GB. So 90min for a 32GB copy? Jun 08 23:05:07 if you are on a mac use rdisk Jun 08 23:05:09 instead of disk Jun 08 23:05:17 i.e. bash-3.2# xz random-9174968c-image-beagleboard-sd-4GiB.img.gz >/dev/rdisk7 Jun 08 23:05:23 that gives you raw unbuffered access Jun 08 23:05:47 imcmahon-home:lcd imcmahon$ sudo dd if=/dev/rdisk2 of=/dev/rdisk3 bs=10m Jun 08 23:05:49 way ahead of ya Jun 08 23:05:54 maybe too small a blocksize Jun 08 23:06:02 10m is fine Jun 08 23:06:09 are you using the internal sd card writer on a mini? Jun 08 23:06:12 ssi: you need real limes for perfect margarita Jun 08 23:06:15 http://linux.die.net/man/1/dcfldd <- dd /w progress bar Jun 08 23:06:18 no, it's internal on an imac Jun 08 23:06:32 mastiff: I bought real limes, but I haven't squeezed them yet... such a hassle :P Jun 08 23:06:36 and I don't think I bought enough Jun 08 23:06:51 ... the current gen of mac mini internal reader is borked, and slow, i have to use a usb reader for it to not be insanely slow... my gf's mini one gen back is better Jun 08 23:07:01 first one I made I used the little lime shaped squeeze bottle Jun 08 23:07:07 next one will be rose's sweetened lime juice Jun 08 23:07:16 if I have enough lime juice out of three limes I'll try one with fresh Jun 08 23:07:24 it's a hard job, trying all these margaritas Jun 08 23:07:30 That's decent, but get yourself a lime squeezer and make a batch of simple syrup. Jun 08 23:07:54 yea I can go that route, but I'm lazy :) Jun 08 23:09:00 PIZZA ORDERED Jun 08 23:09:01 it's gonna be a night Jun 08 23:10:14 heh - hacking BB + liquor = SUCCESS GUARANTEED Jun 08 23:10:20 exactly Jun 08 23:10:25 I should put a record on too Jun 08 23:10:30 oh I was trying to get spidev-test over to the bbw Jun 08 23:10:33 something loud Jun 08 23:10:38 yeah Jun 08 23:11:02 well everything's loud with 93dB/W speakers :P Jun 08 23:11:48 is anyone else trying this at home btw? Jun 08 23:11:51 trying to make spi work on bbb? Jun 08 23:11:54 or am I the only one Jun 08 23:12:14 I'm following with interest - does that count? Jun 08 23:12:27 plus, I've already got it going on my BBW w/ older kernel Jun 08 23:12:38 haha Jun 08 23:13:26 you saw the link to getting it on old kernel that I had on my webpage, right? Jun 08 23:13:30 yeah Jun 08 23:13:33 looks like a lot of work Jun 08 23:13:33 heh Jun 08 23:13:37 not really Jun 08 23:14:10 just recompile the kernel. I used OE Jun 08 23:14:21 it's not hard, just tedious. Jun 08 23:14:30 yeah, the tedious part is what I'm worried about Jun 08 23:14:36 (ie time consuming) Jun 08 23:14:52 oh NO Jun 08 23:14:55 ? Jun 08 23:14:59 dd: /dev/rdisk3: Input/output error Jun 08 23:14:59 3005+0 records in Jun 08 23:14:59 3004+0 records out Jun 08 23:14:59 31499223040 bytes transferred in 7579.579861 secs (4155801 bytes/sec) Jun 08 23:15:00 hahaha Jun 08 23:15:08 I guess one disk is a bit smaller than the other Jun 08 23:15:12 oops Jun 08 23:15:15 bad blocks? Jun 08 23:15:20 beats me Jun 08 23:15:34 lets see if she boots Jun 08 23:15:37 ssi: I just got uart working yesterday, doing I2C next, then SPI. Jun 08 23:15:39 hopefully unused blocks missing Jun 08 23:16:31 meh - /lib/firmware has all that gpio & pwm DT stuff - why not for spidev too? Jun 08 23:17:03 * emeb_mac would like to be spoiled. Jun 08 23:17:17 * maxinux would like someone to maintain a kernel and a reasonable distro for bbb :) Jun 08 23:17:22 spoil me lots ;D Jun 08 23:17:28 maxinux, :P Jun 08 23:17:46 emeb_mac, thats what we are working on today Jun 08 23:17:49 fixing spidev Jun 08 23:17:55 ya Jun 08 23:17:59 but its broken, and not just because of DT Jun 08 23:18:05 someone put in a bad patch Jun 08 23:18:09 upstream Jun 08 23:18:25 bad patch! No! Jun 08 23:18:30 https://patchwork.kernel.org/patch/1678101/ Jun 08 23:18:40 ka6sox: I have little confidence that's going to fix it Jun 08 23:18:42 but I sure hope so Jun 08 23:18:46 "don't rub his nose in it - you'll never teach him anything that way" Jun 08 23:18:48 its vendor specific and there is no "everyone else" Jun 08 23:19:06 ssi, the 3.7 patch is a "y'all" patch Jun 08 23:19:42 I need to look at the rest of spidev and see if he added any in there too. Jun 08 23:19:51 to make it very rohm specific Jun 08 23:19:54 so who's got the latest OE build set up & can rebuild the kernel quick-like-a-bunny? Jun 08 23:20:40 is anyone doing 3.8 with DT on BBW? Jun 08 23:21:07 nothing preventing you from running 3.8 on a BBW Jun 08 23:21:10 should act the same as on BBB, no? Jun 08 23:21:15 well I know Jun 08 23:21:17 but is anyone doing it? Jun 08 23:21:33 * ka6sox listens to the sound of crickets :) Jun 08 23:22:01 heheh Jun 08 23:22:03 that's what I figured Jun 08 23:22:38 yea my dd clone is borked :( Jun 08 23:25:48 bummers Jun 08 23:29:44 ssi, ya, the real diff is those lines Jun 08 23:29:58 just that compatible line? Jun 08 23:30:19 there are other things...but his change was from generic to rohm specific Jun 08 23:30:53 have you gotten that patch to apply against 3.8 yet? Jun 08 23:31:15 I'm working on it now Jun 08 23:32:00 but I had to diff the 3.7 first Jun 08 23:32:25 to see what was being done before Jun 08 23:32:57 basically change that line from rohm,dh2228f to linux,spidev Jun 08 23:33:04 and recompile Jun 08 23:35:21 Hi. I'm hoping someone can help me with this. I'm having a problem trying to get Kinect working with BBB and Ubuntu. I have the software all installed, but there seems to be an issue with the USB. If I plug in the Kinect, it is not recognized by the system. Furthermore, if I try unplugging and plugging in a flash drive after, it doesn't recognize the flash drive either (which it did before). If I reboot, the flash drive will be recogn Jun 08 23:38:46 What does lsusb show? Jun 08 23:40:14 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Jun 08 23:44:34 bbl Jun 08 23:44:40 noooooooo Jun 08 23:46:18 he will be back Jun 08 23:46:50 i know Jun 08 23:46:59 I'm in limbo a bit now anyway Jun 08 23:47:07 I think its just the line I showed you Jun 08 23:47:16 rahil: reboot with the kinect plugged in and see if lsusb output changes. Jun 08 23:47:18 I'm rebuilding now Jun 08 23:47:45 sounds like the Doggers...'its a rebuilding years" Jun 08 23:52:14 lets see if my dd fails Jun 08 23:53:11 mastiff: same thing Jun 08 23:53:22 though it took a lot longer to boot Jun 08 23:53:46 ka6sox: hopefully if yours is gonna fail it'll fail in less than 7579 seconds Jun 08 23:55:20 rahil: Don't know what to tell ya, every usb device I've tried plugging into BBB has at least been recognized by lsusb after reboot. Jun 08 23:55:27 mastiff: though, dmesg has lots of usb errors Jun 08 23:55:28 [ 87.799051] hub 1-0:1.0: port_wait_reset: err = -16 [ 87.799089] hub 1-0:1.0: port 1 not enabled, trying reset again... [ 87.799125] hub 1-0:1.0: Cannot enable port 1. Maybe the USB cable is bad? [ 87.861579] hub 1-0:1.0: port 1 not reset yet, waiting 50ms [ 87.914227] hub 1-0:1.0: port 1 not reset yet, waiting 200ms [ 88.117406] hub 1-0:1.0: port 1 not reset yet, waiting 200ms [ 88.320542] hub 1-0:1.0: port 1 not res Jun 08 23:55:58 Hmm, alright, not sure what it is. Works fine with my computer Jun 08 23:56:03 maybe check the kinect on other devices, I have not used one. Jun 08 23:57:45 Yeah, it works fine with both computers I have access to. Don't have any other embedded type devices to check with unfortunately. Jun 08 23:58:22 booting 3.8.13 build now Jun 08 23:58:32 it's definitely the new kernel because I have the led heartbeat back Jun 08 23:59:46 Linux arm 3.8.13-00641-g79d66fa-dirty #2 SMP Sat Jun 8 19:49:01 EDT 2013 armv7l GNU/Linux Jun 08 23:59:49 dirty dirty kernel Jun 09 00:01:20 ka6sox: no change :P Jun 09 00:04:56 k loopback definitely works on the BBW, 3.2, SPI1 Jun 09 00:09:05 ssi: do you hae a SPI pingroup as expected? what does this show: cat /sys/kernel/debug/pinctrl/44e10800.pinmux/pingroups Jun 09 00:09:51 gimme a sec Jun 09 00:11:34 group: pinctrl_spi0_pins Jun 09 00:11:34 pin 84 (44e10950) Jun 09 00:11:34 pin 85 (44e10954) Jun 09 00:11:34 pin 86 (44e10958) Jun 09 00:11:34 pin 87 (44e1095c) Jun 09 00:14:38 Ok so you have the SPI group, any chance the pin mode on the rx pin could be set wrong? Jun 09 00:14:56 pretty sure it's correct Jun 09 00:15:18 ssi, what pin is your rx pin spi d1? Jun 09 00:15:31 spi0 D1? Jun 09 00:15:42 and what pin # Jun 09 00:15:49 pin 84 (44e10950) 00000010 pinctrl-single Jun 09 00:15:53 pin 85 (44e10954) 00000030 pinctrl-single Jun 09 00:15:56 pin 86 (44e10958) 00000010 pinctrl-single Jun 09 00:15:59 pin 87 (44e1095c) 00000010 pinctrl-single Jun 09 00:16:07 ka6sox: it's 0x954 Jun 09 00:16:12 pin 22 I think it is? Jun 09 00:16:28 that is so readable :) Jun 09 00:16:33 no 18 Jun 09 00:16:52 17 is cs0, 18 is d1, 21 is d0, 22 is sclk Jun 09 00:17:10 I even went so far as to confirm the pad number in the am3359 datasheet Jun 09 00:17:17 to make sure it was'nt an SRM typo Jun 09 00:18:05 P8 or P8? Jun 09 00:18:08 P9 Jun 09 00:18:12 NONE OF THE ABOVE Jun 09 00:18:14 :D Jun 09 00:18:35 p9-22? Jun 09 00:18:38 p9-18 Jun 09 00:18:46 do you hae a pullup enabled? Jun 09 00:18:54 yes, it's mode 0x30 Jun 09 00:19:06 that should be mode 0, pullud-enable, pullup, receiever enabled Jun 09 00:19:10 0x158? Jun 09 00:19:16 0x154 Jun 09 00:19:30 er wait Jun 09 00:19:42 sorry back up Jun 09 00:19:44 d1 is output Jun 09 00:19:46 d0 is input Jun 09 00:19:50 d0 is pin 21 Jun 09 00:19:53 and it's 0x154 Jun 09 00:20:05 P9_18 0×158 I2C1_SDA spi0_d1 Jun 09 00:20:07 pinmux.tking.org agrees with you Jun 09 00:20:51 any thoughts on the slew rate? Jun 09 00:20:57 I'm assuming fast is correct Jun 09 00:23:21 ~wave~ Hello, all. Jun 09 00:23:55 good way to start Jun 09 00:26:42 Anyone know anyone who's gotten a LAMP stack running on the BeagleBone Black/Angstrom? I've been butting heads with getting it running, but as a Linux n00b, making little progress. So far googling hasn't turned up any notes for me from anyone having trod this path before. Jun 09 00:27:25 I haven't seen it, but it shouldn't be terribly hard to do Jun 09 00:29:38 someone is trying (and having issues with) HA Mysqld but thats related to not getting bring up order right yet. Jun 09 00:30:04 I was able to get bluetooth sniffing with ubertooth working on the BBB. Jun 09 00:30:26 I didn't figure it would be too hard either, but so far I've been foiled at each turn. opkg can't find a version of apache it likes, so I've started trying to compile it manually. Only the ./configure step gets hung up on finding apr. opkg CAN find a version of apr it likes, but not apr-util. So I fall back on trying to install both of those manually. That install looks fishy: Apache says all you need to do is extract the two files in Jun 09 00:30:56 Try opkg list | grep apache Jun 09 00:30:59 maybe you ought to start with a more... full-featured distro Jun 09 00:31:27 Yeah, I've tried opkg list | grep apache. Nothing. Jun 09 00:32:00 apache is too heavyweight for this Jun 09 00:32:01 If speed matters nginx is probably gonna be faster on the BBB. Jun 09 00:32:12 nginx or lighttpd Jun 09 00:32:16 (which is there Jun 09 00:32:17 And yeah, I've looked at installing ubuntu on the BBB, but all the docs I've found so far refer to installing it onto a separate SD card and then booting from that. I don't have a separate SD card. Jun 09 00:32:20 maybe cherokee is there. Jun 09 00:32:25 debian may be a betteroption for a LAMP Jun 09 00:32:27 heh Jun 09 00:32:39 2g emmc is going to be pretty tight Jun 09 00:32:44 might want to invest in a usd :) Jun 09 00:32:46 not at all Jun 09 00:33:08 just don't load up gnome/kde or that stuffa Jun 09 00:33:16 debian as configured by RCN come with apache already and is less than 400MB Jun 09 00:33:31 nginx and lighttpd aren't supported by the dev packages I'm looking to run on top of the LAMP stack. Jun 09 00:33:47 what dev packages? Jun 09 00:34:22 anahita Jun 09 00:35:22 I want to do an RPG style system, and I like the idea of using BBBs as a poor-man's blade server. Just add them as more capacity is needed. A social networking/CMS like Anahita is going to work well for the RPG type web-based game I have in mind. Jun 09 00:35:54 this is NOT a full size server...putting a bunch of stuff like this on top just *might* kill it. Jun 09 00:36:04 Why not use a vps? Jun 09 00:36:40 Well, I figure for the final system I'll offload a chunk of the work to Amazon Web Services or something like that. Jun 09 00:37:17 But the BBB, with a 1GHz proc and a few gigs of storage space is a neat little bundle of computational power to have running a load balanced front end. Jun 09 00:37:41 And at 50$ a pop, they're pretty economical to add more as needed. Jun 09 00:37:51 honestly, if you're talking about scaling them wide, the support plant costs are going to quickly overwhelm the cost benefits you think they have Jun 09 00:37:54 PHP Jun 09 00:38:00 are you kidding me? Jun 09 00:38:04 it's cheaper to run a virtualization server Jun 09 00:38:50 Hurt me Baby...make me run PHP on a BBB.... Jun 09 00:39:11 the expression "dog slow" was made for this... Jun 09 00:39:30 Dude, a BBB has horsepower to spare for PHP. It wasn't that long ago that a 1GHz proc was blazing. Jun 09 00:39:34 build a kernel on BBB and then come back and tell me you want to run a server on it Jun 09 00:40:55 PHP dates back to '95. Back in '95 we'd have killed for a 1GHz proc. Jun 09 00:40:58 GLHF...running Apache + mysql on a flash based system seems like a bad idea.... Jun 09 00:41:29 since apache's child processes take a LOT of ram for PHP apps. Jun 09 00:41:48 I don't recall exactly what state of the art in '95 was, but I remember hearing about overclocking procs to 666MHz in '98 and thinking that was crazy fast. Jun 09 00:41:50 With a vps you can scale at will, don't have to buy and configure more BBBs. Jun 09 00:42:13 By the standards when PHP came out, a 1 GHZ proc is a supercomputer. Jun 09 00:42:44 mastiff: not to mention power, network cable, install... Jun 09 00:43:00 baudot, you have 500MB of memory...period, and a Flash Based System...so wear on this app is going to be horrific on the Flash. Jun 09 00:43:04 i try to display video on fb0, but there is a cursor and some old image. how do I clear framebuffer from command line? Jun 09 00:44:36 ssi: also bandwidth and latency Jun 09 00:44:44 yep Jun 09 00:44:55 and virtual switches are MUCH MUCH MUCH cheaper :) Jun 09 00:45:38 what you could do that MIGHT make sense is to configure a BBB as a local backup to a VPS, but don't try to use it for production Jun 09 00:46:39 if he NFS boots this and contends with those issues it might work...as long as he does't try to let it use *any* VM Jun 09 00:47:10 yeah swap == death on this thing Jun 09 00:47:53 Might be fun as a covert server using McD's or starbucks wifi though.... Jun 09 00:48:44 oh sweet Jun 09 00:48:53 NFS boot it and don't use VM Jun 09 00:48:57 then its truely a Blade. Jun 09 00:49:38 only without the convenient lack of cabling :) Jun 09 00:51:07 so... Jun 09 00:51:10 I'm stuck :( Jun 09 00:51:30 I might should enable SPI1 and see if it loops back Jun 09 00:53:10 NFS booting BBBs does sound fun... Jun 09 00:53:17 ssi, is there any evidence that the adafruit folks were sucessful? Jun 09 00:53:42 baudot, I think m_billybob was interested in that or iscsi for doing something similar Jun 09 00:53:44 I haven't seen anyone that's talked about spi working that was anything but tx only Jun 09 00:55:55 ssi, let me diff the 3.7 vs 3.8 tree for spidev and see what else changed Jun 09 00:58:28 ssi: can you echo a value to the spi rx pin? Jun 09 00:59:31 I don't see why not Jun 09 00:59:37 er Jun 09 00:59:41 what are you asking exactly? Jun 09 00:59:49 I can echo to the device, but that writes out on the tx pin Jun 09 01:02:28 what does your test program look like ssi? Jun 09 01:03:28 got a copy of 3.8 source around? Jun 09 01:03:34 yes Jun 09 01:03:34 Documentation/spi/spidev_test.c Jun 09 01:04:16 and the other end is an eprom? Jun 09 01:04:25 doesn't it want a conversation? Jun 09 01:04:29 no, the other end is a loopback cable Jun 09 01:04:39 I confirmed that it works on the BBW Jun 09 01:04:46 and it doesn't on the BBB? Jun 09 01:04:51 when you run that program, it does a synchronous transaction, and writes a seto f bytes Jun 09 01:04:57 on the BBW, I get that set of bytes back printed to screen Jun 09 01:05:03 on BBB, I get either 0x00s or 0xFFs Jun 09 01:05:14 dts? pastebin? Jun 09 01:05:58 http://pastebin.com/vG2ssFS7 Jun 09 01:06:01 that's the current state Jun 09 01:06:50 i think clock is supposed to be input Jun 09 01:07:02 why would it? Jun 09 01:07:07 the mcspi generates the clock Jun 09 01:07:10 it's outputting it to the slaves Jun 09 01:07:12 read it somplace Jun 09 01:07:57 Well gents, nice chatting with you. Jun 09 01:08:43 0x150 0x20 /* spi0_sclk, FAST | INPUT_PULLDOWN | MODE1 , P9.22 */ Jun 09 01:08:45 Good Luck with the cluster ...let us know how it goes...I would be curious about nfsbooting Jun 09 01:08:52 that is how i have mine configured for mode 1 Jun 09 01:09:01 and i've looked at the clock on the scope Jun 09 01:09:49 ssi: I was asking if you cat the rx pin, and then try to echo the opposite value, when you cat the rx pin again does it change? Jun 09 01:10:00 http://forum.beaglefu.com/index.php?app=core&module=attach§ion=attach&attach_rel_module=post&attach_id=43 ssi Jun 09 01:10:13 yellow is the clock with the above setting Jun 09 01:10:30 I had mine on a logic analyzer last night and the clock is fine Jun 09 01:10:34 k Jun 09 01:11:25 mastiff: I don't know how ot "cat the rx pin" Jun 09 01:11:41 ka6sox been doign a bit of reading about nfs bootign even seems fairly straight forward. Jun 09 01:12:01 I've always had better luck with that Jun 09 01:12:20 iscsi and I have a hate/hate relationship Jun 09 01:12:38 ka6sox its pretty simple once you get some experience with it Jun 09 01:12:54 as for iscsi booting yeah it a bit newer technology than nfs booting so . . . Jun 09 01:13:01 Rickta59: why do I count 15 clock transitions per word there? Jun 09 01:13:07 (in your scope image) Jun 09 01:13:48 i'm sending 15 bits at a time Jun 09 01:13:50 ka6sox, either way, i have to recompile my kernel again lol. I made all my options as modules, where they need to be compiled into the kernel Jun 09 01:13:54 ok Jun 09 01:14:19 https://github.com/RickKimball/beaglebone/blob/master/ws2811_spi_test/main.c Jun 09 01:14:33 each bit is about ~84 nsec Jun 09 01:14:46 so 3 bits is ~250 .. 7 ~600 Jun 09 01:15:24 i actually only connected the clock so i could see how bad the gap is Jun 09 01:15:27 you're doing that on the BBB? Jun 09 01:15:32 yeah Jun 09 01:15:38 can you try the loopback test? Jun 09 01:15:54 that why asked for your code was going to give it a shot Jun 09 01:16:07 it's in the linux tree, but I can pastebin it Jun 09 01:16:34 i have it Jun 09 01:16:40 ok Jun 09 01:16:57 https://www.kernel.org/doc/Documentation/spi/spidev_test.c that one right? Jun 09 01:17:03 eah Jun 09 01:17:04 yeah Jun 09 01:19:02 can you paste how you are running it Jun 09 01:19:20 ./spidev_test -D /dev/spidev1.0 Jun 09 01:19:32 you have the pins wired together? Jun 09 01:19:36 yep Jun 09 01:19:46 and if I do that on spi2.0 on BBW with the stock kernel, it works Jun 09 01:19:55 on BBB on spi1.0, I get all zeros typically Jun 09 01:20:28 should be pins 18 and 21 Jun 09 01:22:28 ssi: you have a file that corresponds to the spi rx pin correct? Jun 09 01:24:30 if so, I don't know where Jun 09 01:24:38 At least with GPIO you can echo values to the file corresponding to the GPIO pin such as echo 0 > /sys/class/gpio/gpio38/value Jun 09 01:24:54 yeah, but spi isn't gpio Jun 09 01:25:08 and then you can read the value with cat /sys/class/gpio/gpio38/value Jun 09 01:25:28 spidev is the userspace interface to it, it's a device node, and the kernel handles turning reads and writes to/from into mcspi register reads/writes Jun 09 01:26:08 Since you were getting all zeros, was wondering if you can write a 1 to the file, but sounds like the GPIO stuff is different than SPI Jun 09 01:30:29 hmm .. not seeing the love Jun 09 01:30:56 ? Jun 09 01:31:28 getting back all zeros Jun 09 01:31:33 haha yep Jun 09 01:31:40 http://pinmux.tking.org/p9.html Jun 09 01:31:44 i tried that .. Jun 09 01:32:05 what exactly did you try? Jun 09 01:32:11 pin 21 Jun 09 01:32:33 seems to be right Jun 09 01:32:50 10 seems right Jun 09 01:33:12 should both of them be pulling high? Jun 09 01:33:16 yeah Jun 09 01:33:18 and it shuld be 30 Jun 09 01:33:22 to enable input receiver Jun 09 01:33:32 21 is 0x30, the rest are 0x10 Jun 09 01:33:55 i see the clock ... let me find some more wires i can put the scope on the tx rx Jun 09 01:34:13 by all means Jun 09 01:34:29 I've been down this road, and I'm pretty sure it's a problem in the driver or the DTO Jun 09 01:37:03 didn't ka6sox say it was borked? Jun 09 01:37:07 spidev.c Jun 09 01:37:19 well there was some speculation about the compatible = line in there Jun 09 01:37:27 but I tried changing it with no success Jun 09 01:38:44 i see the pull ups work Jun 09 01:38:54 little spikes in the clock gap Jun 09 01:38:58 yep Jun 09 01:39:33 but it looks like all zeros are being sent Jun 09 01:39:47 being sent from the tx pin? Jun 09 01:39:53 I was able to confirm the data that was being sent Jun 09 01:39:55 i only have one pin Jun 09 01:40:07 eh? Jun 09 01:40:07 right there is just a wire between tx->rx Jun 09 01:40:18 d1->d0 Jun 09 01:40:22 so remove the wire, scope pin 18 (d1) Jun 09 01:40:23 run the program Jun 09 01:40:30 you should see the data in the test block being sent Jun 09 01:40:38 if you wire them back together, you should be able to see the data on both lines Jun 09 01:40:39 :Q! Jun 09 01:40:42 since they're the same line Jun 09 01:40:54 but the test program RECEIVES all zeroes Jun 09 01:40:57 yeah it comes out as 0s Jun 09 01:41:04 on the wire? Jun 09 01:41:10 that's inconsistent with my results Jun 09 01:41:33 i'm also running ubuntu Jun 09 01:41:38 I'm on debian Jun 09 01:41:42 shouldn't matter Jun 09 01:41:43 let me switch over to angstrom Jun 09 01:50:04 hmm .. Jun 09 01:50:16 no success Jun 09 01:51:03 i misspoke before Jun 09 01:51:22 it is actually send a pattern Jun 09 01:51:33 it should be receiving the pattern based on the scope Jun 09 01:52:00 right Jun 09 01:52:10 that's consistent with my results Jun 09 01:52:18 a bunch of 1's some zeros Jun 09 01:52:24 more 1s Jun 09 01:52:31 all the stuff in tx[] Jun 09 01:52:43 I modified tx so the first six bytes are 01 02 03 04 05 06 Jun 09 01:52:50 easier to spot and identify on scope Jun 09 01:55:12 did you try spi1? Jun 09 01:55:29 not yet Jun 09 01:55:36 on the BBW did you run the same image? Jun 09 01:55:39 no Jun 09 01:55:44 same kernel? Jun 09 01:55:44 stock BBW image, 3.2 kernel Jun 09 01:55:46 nope Jun 09 01:55:52 3.2 so no DT stuff? Jun 09 01:56:03 right Jun 09 01:56:16 k .. well at least you have another same Jun 09 01:56:19 sample Jun 09 01:56:19 go get the bbw kernel and try it Jun 09 01:56:24 but you need to do 1 thing Jun 09 01:56:28 add a wire Jun 09 01:56:42 he had success there Jun 09 01:58:38 k .. well enough fun for me Jun 09 02:31:12 ssi: I think I just figure out why my usb bluetooth adapter won't work, the entire rfcomm.conf file is commented out Jun 09 02:31:27 woops Jun 09 02:32:36 maybe conman related, I gotta read up on that BS. conman wipes my network configs everytime I boot as well Jun 09 02:32:56 have to manually reenter the gateway Jun 09 02:33:55 apt-get update; apt-get remove conman; apt-get install debian Jun 09 02:50:39 hmm Can't init device hci0: Operation not possible due to RF-kill **** ENDING LOGGING AT Sun Jun 09 02:59:59 2013