**** BEGIN LOGGING AT Wed Mar 08 03:00:02 2017 Mar 08 03:55:42 Hi Mar 08 03:56:29 I'm trying to using beaglebone black wireless , but my iphone is not able to scan the beaglebone Mar 08 03:57:21 I ssh to the device , and power the bluetooth , and only can discover in Android phone Mar 08 03:57:47 Does any one has the same issue like mine ? Mar 08 05:05:28 is there any prebuilt image for debian stretch console bbb Mar 08 05:05:54 Or where can i download debian stretch rootfs Mar 08 05:40:10 prem: rcn has a few, such as stretch-iot and stretch-wayland. or just grab a jessie console image and upgrade it to stretch Mar 08 05:40:26 as long as there aren't many packages installed, such an upgrade is fairly trivial Mar 08 05:42:22 you can find the latest (jessie) console and stretch-iot here -> https://rcn-ee.com/rootfs/bb.org/testing/2017-03-07/ Mar 08 05:50:15 I just realized I ran into another "fun" issue with qt5... it requests the framebuffer to be XRGB8888, hardcoded. Since the BBB is wired for 16-bit color this results in swapped red/blue channels (or, if the new-ish blue-and-red-wiring property of lcdc has been set correctly in DT, you simply get a failure) Mar 08 05:54:44 it's quite easy to "fix" -> http://pastebin.com/8BpBmc18 Mar 08 05:55:28 but I didn't want to find out how to recompile this stuff. fortunately the constants were very easy to locate in the disassembly, so I did a binary patch :) Mar 08 08:56:21 zmatt: failed to build sgx_build_modules.sh error in bb-kernel 4.9.13-bone4, fatal error No such file or directory Mar 08 09:18:08 did you apply the patches? Mar 08 09:19:58 I'll send an email to rcn to ask him what his status is on supporting the new sgx libs, I get the impression he may not have noticed my post on the list Mar 08 09:39:27 No , i thought the build_kernel script will handle the patches. Now trying the build with 4.4 kernel Mar 08 09:45:31 zmatt: can you check your privmsg? Mar 08 09:49:29 heh Mar 08 09:59:31 heh heh Mar 08 10:13:44 cuba_: I can, and I always do, no need to poke me via multiple routes Mar 08 10:14:29 I've sent him a message that you miss him very very much Mar 08 10:15:41 prem: my point was that I linked to a new set of patches i added Mar 08 10:16:31 thanks so much :) Mar 08 10:18:06 prem: as I said earlier, I don't think rcn has seen my post Mar 08 10:18:21 otherwise they would probably have been included in the -bone kernels Mar 08 10:24:13 Hi all! Mar 08 10:24:19 Long time, no IRC. Mar 08 12:45:55 Good day all Mar 08 12:47:25 Zmatt: Hi, don't know if you remaember me but yesterday you gave me some insight about the beaglebone black. thanks for that. I got some stuff going with the SSH Mar 08 12:49:24 Unfortunatlty sine yeterday i have been trying to detect my sensor connected to the BBB in vain. i tried to connect other sensors. I still can't detect the added devices Mar 08 12:51:32 Aryan: what distro? Mar 08 12:51:48 Aryan: is in your BBB? Mar 08 12:52:34 I detected the i2c bus i2c-0 i2c OMAP I2C adapter I2C adapter i2c-1 i2c OMAP I2C adapter I2C adapter Mar 08 12:52:55 i2c-0 i2c OMAP I2C adapter I2C adapter i2c-1 i2c OMAP I2C adapter I2C adapter Mar 08 12:54:06 Analucia:i detected the i2c bus on the BBB but i can't detect the devices plugged to it Mar 08 12:54:23 that's where i'm stuck Mar 08 12:55:23 Aryan: sorry, i don't know about that. i was only asking you what operating system you are using. Mar 08 12:57:29 Aryan: is this similar to your problem? http://raspberrypi.stackexchange.com/questions/10566/i2c-devices-not-detected Mar 08 12:57:49 tbr: I think I figured it all out. Thanks for your pointers! Mar 08 12:58:46 Analucia: oh ok, I'm on windows, but downloaded puTTy to allow me to work with my BBB directly as if i was working with linux Mar 08 12:59:16 Aryan: maybe this you should read beaglebone.cameon.net/home/i2c-devices Mar 08 13:01:55 Analucia: seem like the problem I'm having Mar 08 13:02:36 Analucia: Will have a look thanks Mar 08 13:06:05 Analucia: i already read it. Mar 08 13:06:11 thanks though Mar 08 13:10:39 Analucia: The thing is that according to what i have been reading one of the minus yoou see in the big table should be replaced by number (adresses) indicating the adress of the connected device to the BBB Mar 08 13:12:15 Analucia: in my case; Device connected or not i still just get the table with the minus and UU (reserved or skipped adresses)) Mar 08 13:19:33 GeofC: good :) Mar 08 13:25:06 Aryan: what big table? Mar 08 13:25:58 Zmatt: I'm refering to this Mar 08 13:26:04 0 1 2 3 4 5 6 7 8 9 a b c d e f 00: -- -- -- -- -- -- -- -- -- -- -- -- -- 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 20: -- -- -- -- UU -- -- -- -- -- -- -- -- -- -- -- 30: -- -- -- -- UU -- -- -- -- -- -- -- -- -- -- -- 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 50: UU -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 70: UU -- -- -- -- -- -- Mar 08 13:26:36 please don't paste such multi-line content into the channel Mar 08 13:26:46 0 1 2 3 4 5 6 7 8 9 a b c d e f 00: -- -- -- -- -- -- -- -- -- -- -- -- -- 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 20: -- -- -- -- UU -- -- -- -- -- -- -- -- -- -- -- 30: -- -- -- -- UU -- -- -- -- -- -- -- -- -- -- -- 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 50: UU -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 70: UU -- -- -- -- -- -- Mar 08 13:26:48 use a paste service like pastebin.com, hastebin.com Mar 08 13:27:05 hold on please Mar 08 13:27:11 and I have no idea what I'm looking at Mar 08 13:27:19 gist.github.com is also decent Mar 08 13:27:25 can you just explain what you're doing? Mar 08 13:28:41 you should connect your sensor to the i2c-1 bus and then perform suitable i2c transactions to its slave address, which you'll find in the datasheet Mar 08 13:29:45 i2c-1 is enabled by default on pins P9.19 (SCL) and P9.20 (SDA) Mar 08 13:31:04 Hello my bbb was working fine until yesterday but now when I am trying to ssh into it ,its not allowing .I tried nmap and arp and it dosent show any host up,basically its not detecting the hardware anywhere.It has Debian 8.7 on it.SO now how can I fix it/? Mar 08 13:31:11 0 1 2 3 4 5 6 7 8 9 a b c d e f Mar 08 13:32:54 Aryan: also btw it's worth mentioning that the MPU-9250 also supports SPI instead of I2C for communication, with much much higher performance (20 MHz spi vs 400 kHz i2c) Mar 08 13:33:52 guitarman_: hard to diagnose based on this description Mar 08 13:34:07 guitarman_: I does it still seem to be optional? Mar 08 13:34:26 *operational Mar 08 13:36:17 Operational .It gets detected by the pc,Also the usr leds are lighting up Mar 08 13:36:52 are you trying to ssh via usb networking or ethernet? Mar 08 13:37:30 it gets detected by the pc... earlier you said "basically its not detecting the hardware anywhere" :) Mar 08 13:40:11 usb Mar 08 13:40:27 its not detecting the hardware over the network Mar 08 13:40:29 note that even if networking isn't working, you can still get in via the console to diagnose the situation... if it shows up via usb you may be able to reach the console that way (it shows up as a serial port), otherwise the real serial console can be lifesaver Mar 08 13:41:09 I think I will have to go for the serial console .Does it require some extra wire? Mar 08 13:41:14 serial to usb/? Mar 08 13:41:31 yeah you need a 3.3V serial cable thingy, see e.g. http://elinux.org/Beagleboard:BeagleBone_Black_Serial Mar 08 13:41:50 Also I had earlier sort of messed up my beaglebone file system.Can you help me with tahht too Mar 08 13:42:27 (those friction-fit things near the bottom of the wiki page are really awesome if you want to be able to debug a beaglebone that has a cape attached) Mar 08 13:43:04 again, problem statement is way too vague.... we're not magicians nor psychic :) Mar 08 13:43:38 Zmatt: What i meant is that; when we use the command i2cdetect -y -r 0. We should be able to detect the connected device adress,isn't? Mar 08 13:43:43 generally having access to the serial cable makes debugging issue a lot easier, since you can see what's going on Mar 08 13:43:55 Aryan: don't use i2cdetect Mar 08 13:44:06 look in the datasheet for the correct address to use Mar 08 13:45:53 also, you were probing the wrong bus Mar 08 13:47:19 What i had done is that i think I created a partition in the BBB and made it into a bootable.I thought that I ended up wiping out the entire data but luckily my data was intact as I was able to ssh into it and allmy codes were still there.How can I fix it now as when its detected by the file system it shows all the live bootable files.I tried formatting that part but it dosent allow me to Mar 08 13:47:22 oh ok Mar 08 13:47:49 I actually probed both buses Mar 08 13:48:10 that's was just to show you what i meant Mar 08 13:48:12 1. don't. probe. Mar 08 13:48:45 i2cdetect is a program that shouldn't really exist Mar 08 13:49:18 nor do you need it Mar 08 13:49:51 guitarman_: ehh Mar 08 13:50:51 ok Zmatt let me look into it first. may be i can get it working Mar 08 13:51:05 guitarman_: I don't really quite understand what you're describing, but it doesn't sound good Mar 08 13:51:34 Aryan: check that you connected it in an appropriate way to the right pins, supply voltage, etc Mar 08 13:52:24 if you want to connect it to anything other than i2c-1, you'll also need to setup the beaglebone correctly for that Mar 08 13:53:25 Zmatt: the connections seem fine Mar 08 13:54:27 datasheet says its i2c address is 0x68 or 0x69, depending on whether the AD0 signal is tied low or tied high Mar 08 13:55:33 Zmatt:GND to pin 1, VCC to pin 3, scl to pin 19 and SDA to pin 20 Mar 08 13:56:29 zmatt: Aryan: remember, linux i2c addresses are >>1 Mar 08 13:56:33 uhh, not on the MPU-9250... are you using some kind of module? Mar 08 13:56:48 thinkfat: uh, no they're not Mar 08 13:57:19 the read/write bit is not part of the address Mar 08 13:57:48 zmatt: yes, but in the datasheet normally the addresses are given including r/w bit Mar 08 13:58:05 zmatt: or, not normally, but frequently Mar 08 13:58:09 on rare occasion I see that, but not normally Mar 08 13:58:29 zmatt: i'm using the MPU-9250 breakout from sparkfun Mar 08 14:01:46 zmatt:what i meant is that i conected MPU GND to BBB pin 1, MPU vcc to BBB pin 3, MPU SCL to BBB pin 19 and MPU SDA to BBB pin 20 Mar 08 14:03:21 ohh, you meant of p9. please then say that explicitly, there is no "BBB pin 1" Mar 08 14:03:34 also, that means you probably fried your MPU-9250 Mar 08 14:03:39 no way Mar 08 14:03:42 8awwaiat Mar 08 14:03:43 dsfjhdksjh Mar 08 14:03:45 sorry Mar 08 14:03:57 I was thinking wrong, pin 3 is 3.3V not 5V Mar 08 14:04:50 (my mind was briefly thinking "third row") Mar 08 14:04:52 yes Mar 08 14:05:56 and pin 1 on P9 is ground, isnt? Mar 08 14:06:03 yes Mar 08 14:06:50 so my connection diagram is correct Mar 08 14:07:10 then i don't know what i'm missing here Mar 08 14:10:37 i2cget -y 1 0x68 0x75 Mar 08 14:11:04 Zmatt: any suggestion of something you think i should try? Mar 08 14:11:07 yes Mar 08 14:16:34 ... and? Mar 08 14:16:54 does that command work? Mar 08 14:20:08 Zmatt: read failled Mar 08 14:20:29 that's the response i'm getting Mar 08 14:20:43 which debian version are you running the beaglebone? Mar 08 14:22:04 just to be extra sure you use my 'show-pins' utility to verify that i2c-1 is correctly configured -> https://github.com/mvduin/bbb-pin-utils Mar 08 14:22:28 ohhhhhh Mar 08 14:22:38 do you have pull-ups on the bus? Mar 08 14:23:18 oh the sparkfun thing has those Mar 08 14:24:33 (kinda rude since that could be a problem if you want multiple devices on the same bus.... *shrug* ) Mar 08 14:25:05 yes ut does Mar 08 14:26:20 then I don't know... you can use my util to confirm that those header pins are indeed configured for i2c-1 Mar 08 14:26:32 you can try powering off and on Mar 08 14:27:14 the datasheet of MPU seemed to suggest it is a bit fussy about power-up, turning on the bbb while the MPU is connected should result in a nice power-up for it too Mar 08 14:28:00 ok Mar 08 14:28:10 the debian version is 8 Mar 08 14:28:10 maybe try hooking up some stupid device like an eeprom just to confirm the bus is working Mar 08 14:29:17 ok Mar 08 15:08:50 hello. Mar 08 15:09:21 What should i do to make a LCD cape working? it just show background light Mar 08 15:09:42 on a BBB with the latest Jessie image linked from the site Mar 08 15:11:39 I remember i had to change something in the old Debian image i used two years ago, but i can't remember or find what it was Mar 08 15:13:44 zmatt: do you know anything about wakeup_m3 software for current BBB images, wherefrom, which tag? Mar 08 15:15:57 thinkfat: yeah it's in a git repo somewhere, one sec Mar 08 15:17:53 ah beagleboard.org has a mirror of it on github even... https://github.com/beagleboard/am33x-cm3 Mar 08 15:18:02 ok, that I know Mar 08 15:18:19 where's the upstream, arago-project or git.ti.com? Mar 08 15:18:51 http://arago-project.org/git/projects/?p=am33x-cm3.git;a=summary Mar 08 15:19:16 the github repo is pretty outdated Mar 08 15:20:18 I really still need to play a bit more with power management on the am335x and the interaction between a8 and m3... I did a bit of testing a while back, with a small baremetal prog on the a8 and jtag access into the m3 Mar 08 15:20:25 it doesn't work like I thought it did Mar 08 15:21:01 I've been toying with that, too. but I never managed to suspend and wakeup, I guess I didn't manage to meet all the requirements Mar 08 15:21:29 you mean with linux? oh I've tested it a few times and it worked for me Mar 08 15:21:40 no, in-house OS Mar 08 15:21:46 ah Mar 08 15:22:15 so, the funny thing is, the a8 is not at all as dependent on the m3 as I imagined Mar 08 15:23:40 that's the current stuff: http://git.ti.com/processor-firmware/ti-amx3-cm3-pm-firmware Mar 08 15:23:56 both github and arago are cold Mar 08 15:24:53 zmatt: I thought the a8 needed the m3 for wakeup? and for putting the main oscillator to sleep? Mar 08 15:24:56 originally I thought that when the a8 sets its own modulemode to 0 and then enters wfi to acknowledge the idle request from prcm that that was it for the a8, the m3 is woken up by irq and it's in charge until it reenables the a8 and itself goes to sleep Mar 08 15:25:14 that would also have made a nice race-condition-free interlock Mar 08 15:25:29 zmatt: yes, that's what I thought, too. it's not the case? Mar 08 15:26:50 no, you can turn the modulemode, put the clock domain in software sleep.... an irq will happily barge through all of that Mar 08 15:27:02 it seems the a8 intc generates a wakeup request Mar 08 15:27:31 probably not the deepest sleep possible Mar 08 15:27:33 iirc even if the mpu power down is off (the aintc is part of core power domain I think) Mar 08 15:28:02 and iirc every wfi sends the m3 that irq Mar 08 15:28:18 so this makes the interaction much much messier Mar 08 15:29:11 no I'm sure the m3 can take the SoC into lower power modes Mar 08 15:29:31 simply because it is part of the wakeup power domain Mar 08 15:30:49 hmmm, no. I remember I debugged that with jtag, if you don't set the modulemode of the cpu to 0, the wfi will not generate an irq Mar 08 15:31:12 ok then I misremember that part Mar 08 15:31:59 although the value of that gating is dubious if the mpu can wake itself up from idle while the m3 is handling the irq Mar 08 15:34:05 I'd be interested to know if you're able to reproduce this observation... I didn't spend that much time on it back then, maybe I was doing something silly Mar 08 15:35:25 zmatt: sure. I've just blown the dust off my bbb and trying if my old oocd scripts still work for simultaneous debugging of the cm3 and the a8 Mar 08 15:35:30 zmatt: turns out they do Mar 08 15:35:38 of course you don't even need the m3 for this Mar 08 15:36:52 if my observations are right then e.g. my tiny irq-driven led demo would run correctly even if mpu modulemode is zero, just with a bit higher irq latency Mar 08 15:37:45 zmatt: I think for just cpu modulemode 0 you don't need the m3 Mar 08 15:38:21 and maybe even with power state request set to "OFF", provided you ensure that before you call wfi any registers of importance are pushed onto stack and the resume vector for ROM points to a suitable place Mar 08 15:38:22 zmatt: you only need it of you e.g. want to shut down the main oscillator, too, for maximum power saving Mar 08 15:39:12 does that really save a lot if most logic is already clock-gated? Mar 08 15:39:59 zmatt: no, only a few mW Mar 08 15:40:28 anyway, I really need to focus on other stuff... I don't really care that much about power management on the am335x... (only on the omap5 ;) Mar 08 15:40:40 zmatt: mostly, you want to power-off most of the chip to avoid the leakage currents Mar 08 15:42:45 the cpu can probably power off a lot though... including itself possibly :) Mar 08 15:43:53 but for some reason, when the beaglebone is integrated right next to a 1550 W amplifier, I have trouble bringing myself to care about power management Mar 08 15:44:34 ;) Mar 08 15:44:49 zmatt: in fact the cm3 software isn't doing a lot, I think it mostly cares about DDR and some few power domain states, and of course the main oscillator Mar 08 15:45:35 I think it got less involved over time? Mar 08 15:48:05 hmm, looking at the commit history, I'd rather say it got more and more involved Mar 08 15:48:20 okay Mar 08 15:48:28 I don't follow it that well Mar 08 15:49:32 thanks for the url update btw (I just realized I thought it but didn't type it) Mar 08 15:51:35 the m3 firmware could probably be so much better Mar 08 15:52:42 instead of long lists of steps it should use some sort of state machine, to allow it to begin resuming again at any point while busy entering a low power mode Mar 08 16:04:34 zmatt: I'm still not sure if it was supposed to be acting during normal cpu-idle modes, there's a bug in the firmware that prevents these commands from functioning Mar 08 16:10:57 thinkfat: git.ti.com/processor-firmware is the latest Mar 08 16:11:45 the repo got moved and renamed few times, hence the old copies... Mar 08 16:17:05 denix: yes, got it, but not a lot has changed since firmware version 190 Mar 08 16:17:59 denix: do you know if the corresponding linux code still needs a function executing from sram that puts the DDR memory into self-refresh? Mar 08 16:20:13 thinkfat: not sure. you can check the corresponding kernel at git.ti.com/ti-linux-kernel Mar 08 16:27:02 gitorious so much sucks Mar 08 16:27:20 how do I browse all the branches, not just the first 5? Mar 08 16:33:01 denix: yes, still does Mar 08 18:04:09 that sounds a bit odd though, you can configure emif to automatically enter selfrefresh whenever it's idle for a bit Mar 08 18:05:51 but indeed if fiddling with EMIF settings is ever needed for whatever reason, that's obviously something the A8 has to do Mar 08 18:22:51 Question about booting up a "virgin" BeagleBoneGreen-like clone. If I make a Flasher-SD using /opt/scripts/tools/eMMC/beaglebone-black-make-microSD-flasher-from-eMMC.sh, it does not appear to have the U-SPL required to boot up from a virgin system -- where I mean both eMMC and EEPROM are blank. How do I add U-Boot-SPL to the beginning of the SD card to allow this? Mar 08 18:25:13 From what I can tell, my flasher script will set the EEPROM -- if I can get Linux booted off the SD. Mar 08 18:37:53 And from what I can see: If the boardID EEPROM is blank (FF's), the U-SPL won't start with a flasher made by /opt/scripts/tools/eMMC/beaglebone-black-make-microSD-flasher-from-eMMC.sh. Is that true? Mar 08 18:38:57 It appears that the ROM bootloader is looking for some magic from the SD care that I don't have. Mar 08 18:48:12 GeofC: did you add https://github.com/RobertCNelson/Bootloader-Builder/blob/master/patches/v2017.03-rc3/0002-NFM-Production-eeprom-assume-device-is-BeagleBone-Bl.patch to your u-boot to ignore the EEPROM? Mar 08 18:48:40 GeofC: yes, it is true. u-boot expects values in the EEPROM so you have to patch it like ^^^ Mar 08 18:50:56 No, that's the bit of magic that I was looking for. Is there a doc there I assume that will show me how to add it? Mar 08 18:54:05 Does this assume that I'm building my own U-boot? Could I just copy the uboot file off of an RCN flasher (that seems to have the patch already)? That would be simpler for me. :-) Mar 08 22:16:00 hi Mar 08 22:16:15 I am working on beaglebonegreen wireless Mar 08 22:16:38 I want to customize the linux and add more kernel configurations also some packages Mar 08 22:16:48 how can i do it? Mar 08 22:17:05 I am searching for relevant links for long time now Mar 09 02:01:07 GeofC: rebuilding u-boot is fairly trivial. I don't think anyone sees much reason to make this use-case easier since generally one would expect that those who can make beaglebone-like devices have some understanding of what they're doing ;P **** ENDING LOGGING AT Thu Mar 09 03:00:02 2017