**** BEGIN LOGGING AT Mon Mar 07 02:59:58 2016 Mar 07 05:07:08 Please Help me in the project "Improving initial experience for novice developers" . I am not getting any information about the topic Mar 07 09:01:30 I wonder if Yatharth realized the irony here Mar 07 09:53:44 Can anyone here confirm, that instructions for mac os x from this guide http://beagleboard.org/getting-started work? Mar 07 09:55:01 After I install HoRNDIS package, I can't find any sign of it in the system. Not sure if it even installs. Mar 07 09:55:17 And FTDI drivers are not loaded Mar 07 09:55:31 When I load them by hand, nothing happens Mar 07 09:58:17 untitled: HoRNDIS is for ethernet-over-USB or something, right? Mar 07 09:58:28 yes Mar 07 09:58:36 how does this relate to FTDI Mar 07 09:58:53 you shouldn't need FTDI drivers on your host machine Mar 07 09:59:18 there are 2 sets of drivers on this getting-started guide and it says "install them both" Mar 07 09:59:27 does it lie to me then? Mar 07 10:00:09 you have an FTDI cable, yes? Mar 07 10:00:21 but anyways, I can't even install HoRNDIS :) the installation goes fine and then no sign of the driver anywhere Mar 07 10:00:33 I only have usb to miniUSB Mar 07 10:03:26 untitled: open up network in the system preferences app thing Mar 07 10:03:53 you should see new USB related connection mechanisms Mar 07 10:04:01 no, don't see it Mar 07 10:04:18 & you're going to need an FTDI cable before FTDI becomes relevant with whatever beagle-related things you're trying to do Mar 07 10:04:33 I searched through the system, there is no installed files, related to HoRNDIS Mar 07 10:04:34 you're tryna tether your phone or something to your mac right Mar 07 10:04:52 or, derp Mar 07 10:04:58 the beagle's USB/ethernet thing Mar 07 10:05:06 yes Mar 07 10:05:11 plug it in and power it on Mar 07 10:05:14 it is Mar 07 10:05:15 it should appear as an option Mar 07 10:05:24 and nothing appears :) Mar 07 10:05:32 something's wrong Mar 07 10:05:39 i've done this exact thing a bunch of times :) Mar 07 10:05:53 horndis comes as a .pkg right? Mar 07 10:06:29 yes Mar 07 10:06:52 and when you run it, it prompts you for a password right? Mar 07 10:06:56 I don't see no HoRNDIS.kext after install Mar 07 10:06:59 yes Mar 07 10:07:03 installation goes fine Mar 07 10:07:46 in the network settings thing, click the lock to unlock, then hit the little + Mar 07 10:08:03 you should find what you're looking for there Mar 07 10:08:18 there is nothing about USB there Mar 07 10:09:18 what about the little gear icon to the right of the +/- Mar 07 10:09:29 i'm looking at my macbook now & the story is different Mar 07 10:09:53 I have A5A revision, if it's important Mar 07 10:09:56 hi, i am doing Image Processing on Embedded Linux using OpenCV. i am following this link: http://derekmolloy.ie/beaglebone/beaglebone-video-capture-and-image-processing-on-embedded-linux-using-opencv/. but i am stuck in the first step. when i connect my board with adapter and ethernet cable. i am unable to make ssh connection. whereas when i reconnect the board through usb port it was normal. why is this so? and what should i do to solve Mar 07 10:09:57 bbb Mar 07 10:10:33 i'm not sure what's going on then :) Mar 07 10:10:37 sorry Mar 07 10:14:13 kremlin: but can you see any HoRNDIS.kext in your system_ Mar 07 10:14:16 ? Mar 07 10:14:25 yes Mar 07 10:14:36 well, wait i don't know Mar 07 10:14:45 ive never had a problem with it and haven't checked Mar 07 10:15:02 can you check now? Mar 07 10:16:10 i suppose Mar 07 10:16:12 one second Mar 07 10:20:16 untitled: no matches for that filename Mar 07 10:20:27 yet horndis is installed &* working Mar 07 11:07:23 hi, i am doing Image Processing on Embedded Linux using OpenCV. i am following this link: http://derekmolloy.ie/beaglebone/beaglebone-video-capture-and-image-processing-on-embedded-linux-using-opencv/. but i am stuck in the first step. when i connect my board with adapter and ethernet cable. i am unable to make ssh connection. whereas when i reconnect the board through usb port it was normal. why is this so? and what should i do to Mar 07 11:09:50 I don't see any steps on that page, it's rather generic Mar 07 11:12:04 actually this is to load camera on BBB Mar 07 11:12:21 but at first step we are trying to connect via ethernet Mar 07 11:12:26 but it never connects Mar 07 11:16:35 and how are you trying to connect? Mar 07 11:26:32 ssh root@192.168.7.2 Mar 07 11:26:57 do i need to configure etheret ports on host machine and BBB Mar 07 11:27:03 ?? Mar 07 11:27:18 that is obviously not going to work. you need the LAN IP of the BBB, not the USB interface IPv4 address Mar 07 11:27:38 where do you connect the ethernet cable from the BBB? Mar 07 11:28:15 one end is connected to BBB and other to host ubuntu machine Mar 07 11:29:20 i tried to configure network intergace with static ip Mar 07 11:29:27 but i am still unable to conenct Mar 07 11:31:17 you need to configure both sides with static IP addresses from within the same subnet Mar 07 11:31:48 i did on both end. but when i check it on my BBB it is not static one that i set Mar 07 11:31:51 any leads Mar 07 11:31:52 ?? Mar 07 11:33:09 any sample instrutions you might share for both end. i think i am missing some point while configuration Mar 07 11:33:26 how are you setting the IPv4 addresses? Mar 07 11:35:59 in network interface of BBB i am doing this Mar 07 11:36:03 auto eth0 iface eth0 inet static address 192.168.1.102 netmask 255.255.255.0 gateway 192.168.1.254 dns-nameservers 8.8.8.8 dns-nameservers 8.8.4.4 Mar 07 11:38:02 what's the output of this command on your laptop: Mar 07 11:38:05 ip addr|grep inet|grep -v inet6 Mar 07 11:40:01 inet 127.0.0.1/8 scope host lo inet 192.168.7.1/30 brd 192.168.7.3 scope global eth1 inet 192.168.1.6/24 brd 192.168.1.255 scope global wlan1 Mar 07 11:40:17 one thing i deleted that eth config from my laptop Mar 07 11:40:29 i am doing it and than will post output again Mar 07 11:40:31 right, so you have managed to create a network conflict Mar 07 11:40:49 do NOT under ANY circumstances assign the same subnet on two different interfaces! Mar 07 11:41:47 yes when i ifconfig i get Link encap:Ethernet HWaddr ac:16:2d:0e:4c:6a inet6 addr: fe80::ae16:2dff:fe0e:4c6a/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:51 errors:0 dropped:0 overruns:0 frame:0 TX packets:25 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:22728 (22.7 KB) TX bytes:4370 (4.3 KB) Int Mar 07 11:46:08 please don't paste multi-line content to the channel Mar 07 11:46:32 if you must, use something like pastebin.com or pastebin.mozilla.org or one of the many other services Mar 07 11:48:28 ix.io :) Mar 07 11:54:11 i have done this on my machines network interface: auto eth0 iface eth0 inet static address 192.168.1.254 netmask 255.255.0.0 Mar 07 11:55:03 output of ssh is: ssh: connect to host 192.168.1.102 port 22: No route to host Mar 07 11:55:22 as I told you before your configuration is wrong Mar 07 11:55:25 stop what you are doing Mar 07 11:55:33 select a DIFFERENT SUBNET Mar 07 11:56:40 that netmask is also bogus Mar 07 11:57:11 any suggestion Mar 07 11:57:13 ?? Mar 07 11:57:18 tea. Mar 07 11:57:23 255.255.255.0 Mar 07 11:57:24 ?? Mar 07 11:57:43 yeah sure zmatt. Tea... Mar 07 11:58:31 you don't have any dhcp server on your lan? (router, adsl/cable modem) Mar 07 11:59:08 on my network its 255.255.255.0 Mar 07 12:00:02 that I had noticed already. that answer has nothing to do with my question though Mar 07 12:00:50 yes we have router Mar 07 12:01:00 sorry but i am noob with networks Mar 07 12:01:16 then don't manually configure anything, just use dhcp Mar 07 12:01:26 also, what filesystem/image are you using? Mar 07 12:01:49 its debian Mar 07 12:02:00 11:40:49< tbr> do NOT under ANY circumstances assign the same subnet on two different interfaces! Mar 07 12:02:01 yes but which image Mar 07 12:02:24 tbr: there's a dhcp server, he shouldn't be assigning anything manually whatsoever Mar 07 12:02:25 you have on your laptop 192.168.1.0/24 on BOTH: wlan0 and eth0 Mar 07 12:02:31 zmatt: there is NOT Mar 07 12:02:37 13:00 < nonBeliever> yes we have router Mar 07 12:02:47 zmatt: direct cable from BBB to laptop, read scrollback Mar 07 12:02:51 zmatt: yes, a WIFI router Mar 07 12:02:59 tbr: oh ok Mar 07 12:03:03 that sucks Mar 07 12:03:16 since you can't bridge that either Mar 07 12:03:20 (iirc) Mar 07 12:03:24 you could, but meh Mar 07 12:03:30 just assign a different subnet Mar 07 12:03:39 if you need routing, add nat Mar 07 12:04:27 or don't bother configuring ipv4 at all, stick the cable in, connect to beaglebone.local Mar 07 12:04:34 and let avahi and ipv6 do the work for you Mar 07 12:04:43 that /might/ work, yes Mar 07 12:05:00 or figure out the bbb link local address and ssh into that Mar 07 12:05:03 in my experience it does Mar 07 12:05:13 Good day, I am new to BBB and looking for some advice with a booting problem. I have not been able to pick up my BBB when connecting it to my PC for the first time. My laptop is USB 3 and 3.1, i have tried all the ports. Mar 07 12:05:16 * tbr has done that often before, because screw IPv4 Mar 07 12:05:21 BBB-eMMC-flasher-debian-8.2-lxqt-4gb-armhf-2016-01-17-4gb is image i used in BBB Mar 07 12:05:35 nonBeliever: ok, good Mar 07 12:05:43 Datin: which LEDs light up? Mar 07 12:05:49 the only thing is, that uses connman and I know absolutely nothing about connman Mar 07 12:06:07 only the power LED tbr Mar 07 12:06:08 so i should stick to dhcp instead of static Mar 07 12:06:28 and my machine sould get it via ssh root@beaglebone.local Mar 07 12:06:28 ?? Mar 07 12:06:53 that should work yes, I'm *reasonably* sure avahi is installed by default (otherwise just install it) Mar 07 12:07:23 Datin: that's bad Mar 07 12:07:41 tbr: only the power LED. It stays solid all the time Mar 07 12:08:05 Datin: that usually means ROM can't find any bootable medium Mar 07 12:08:23 i.e. eMMC doesn't appear bootable to it Mar 07 12:08:54 Can I fix that? Mar 07 12:09:11 try reflashing Mar 07 12:10:09 I attempeted to load a new OS from the Micro SD card following the instructions on the BBB.org web page. Mar 07 12:10:33 what's the name of the image you used? Mar 07 12:10:54 you can find the latest images at http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_lxqt (I'd recommend the lxqt-2gb flasher) Mar 07 12:12:31 tbr: bone-debian-8.3-lxqt-4gb-armhf-2016-01-24-4gb Mar 07 12:12:47 I will try that zmatt. thx Mar 07 12:13:07 the -4gb should work too of course, provided you have BBB rev C Mar 07 12:14:42 even if you have a rev C (which has 4gb eMMC), the added benefit of the 4gb-flasher is limited (I think just the Chromium web browser) and using the 2gb-flasher leaves a lot more free space Mar 07 12:15:11 I will try the -2gb tonight and see how it goes. thx Mar 07 12:15:21 recently someone actually got a hosed system by running out of free space while installing updates... which is nasty Mar 07 12:17:23 it would be nice if there were a prebuilt image that omits GUI stuff but still includes stuff like bonescript, cloud9, etc that lets people be able to follow tutorials Mar 07 12:17:54 right now it's either "console" (which would be more appropriately called "minimal") or a full GUI, with nothing inbetween Mar 07 12:20:00 Thanks for the help guys, Have to get back to work. I will be back If it does not work.. lol Mar 07 12:20:19 zmatt: the IoT images seem to be also interesting. essentially it's the lxqt image but without all the graphical UI bits. Mar 07 12:20:57 tbr: I just noticed it in the list! so that's basically exactly what I was just praying for? Mar 07 12:21:07 yes Mar 07 12:21:11 cool Mar 07 12:21:28 Datin: so, if you don't need a gui, grab the iot image :) Mar 07 12:22:10 (if you need a few dedicated GUI bits you can always install them later) Mar 07 12:22:55 yes Mar 07 12:31:45 ssh from ethernt gave me this error: ssh: Could not resolve hostname beaglebone.local: Name or service not known Mar 07 12:31:51 i am installing avahi now Mar 07 12:32:22 ah, avahi-daemon wasn't installed? Mar 07 12:32:54 nope i guess Mar 07 12:36:39 I see the beagle bone black has an internal RTC, Is there anyway to configure and use it? Mar 07 12:36:54 As I dont see any external pins for it Mar 07 12:39:11 Guest94504: it is used automatically by the kernel Mar 07 12:40:51 do i need to use an external RTC for displaying time and date in my application? Mar 07 12:41:15 Or I can use the internal RTC itself for this purpose? Mar 07 12:43:33 the kernel maintains the current date and time, systemd-timesyncd (or ntpd on older images) will also synchronize it to an NTP server when possible. the RTC itself is mainly used to keep track of the time when normal timers cannot do this, e.g. across reboots Mar 07 12:45:10 you can obtain the current date/time using the 'date' utility, the gettimeofday() call in C/C++, and equivalents in other languages Mar 07 12:46:44 Ok, last thing, is it nessesary to be connected to the internet to maintain a display of proper time? Mar 07 12:47:23 if there's no way to automatically synchronize the current date/time, you can set it manually using the 'date' utility, but it will of course slowly start to drift away over time Mar 07 12:48:26 or you could hook it up to a gps time source Mar 07 12:48:29 there may be a way to at least calibrate the oscillator frequency to minimize this drift... I think leaving it connected to the internet (with NTP access) for a while will automatically do this and remember this frequency adjustment Mar 07 12:48:34 not even pps, but simple NMEA Mar 07 12:48:49 or what tbr said, you can connect some other source of timing info Mar 07 12:49:10 note that the BBB does *not* have any integrated battery to keep the RTC running when unpowered Mar 07 12:49:27 thats the problem Mar 07 12:49:40 I'd rather go for an external rtc Mar 07 12:50:26 And I am working on a standalone system, therefore it is not connected to internet Mar 07 12:50:50 if the accuracy of an RTC is 'good enough', then sure, you can do that Mar 07 12:51:00 yes, an external RTC with a standard 3V lithium battery, connected via I2C to the BBB is then probably the simplest solution Mar 07 12:51:48 yes will do that. Its simple. Mar 07 12:51:59 some of those have very decent accuracy over time and extremely low power consumption Mar 07 12:53:16 yes. each battery can last for upto 2 years and the current consumption is in nano amps Mar 07 12:53:38 assuming linux has a driver for the RTC (it has for lots of them), if you declare the RTC's existence in device tree then the kernel should automatically use it Mar 07 12:54:03 yes thats right. Mar 07 12:54:39 Anyways thanks for info folks. Good day! Mar 07 13:25:14 hey guys. i am back Mar 07 13:25:22 i tried to install avahi Mar 07 13:25:33 it said you already have latest version of avahi Mar 07 13:25:39 stuck agan Mar 07 14:59:17 where to consume news? not only for beaglebone but also intel edison and such? Mar 07 16:11:50 hi, any news on x15? Mar 07 17:32:02 It's me again, with the same problem (BBB not visible through USB in any OSs). Mar 07 17:32:16 I decided, that USB circuit is broken on this board Mar 07 17:32:55 Tried different cables, ofcourse Mar 07 17:34:52 while perhaps it's possible in theory, I don't think I've ever heard of such a thing before Mar 07 17:36:15 what system is flashed onto the BBB? do you have an alternate way to access it to inspect what's going on on the BBB side? (ssh via ethernet, or serial debug cable) Mar 07 17:37:10 I tried different images of debian and one of this sweedish weird distro Mar 07 17:37:22 noone of those exposed itself via usb Mar 07 17:37:43 you mentioned it's not visible though USB in any OS... do you have a linux host system available to test on? does anything usb-related appear in the kernel log at all when you plug it in? Mar 07 17:37:44 and I don't have other means to inspect the board, not even a monitor with hdmi Mar 07 17:38:17 yes, I have linux host system and no, nothing at all show up in /dev, lsusb or dmesg Mar 07 17:39:36 try powering on the BBB while holding the S2 button down (the one closest to the uSD card slot) with no card inserted... it won't boot in this mode, but it does (normally) enumerate as an RNDIS usb device in this state Mar 07 17:40:02 that last test would unambiguously exclude software issues on the BBB Mar 07 17:43:08 I see nothing in logs Mar 07 17:43:50 although, the board seems to power on normally and the leds flash according to documentation when flashing or working Mar 07 17:43:56 (with usb) Mar 07 17:44:05 funky. did you test with different host computers or (rather than one that has different OSes installed) Mar 07 17:44:21 you don't have any ethernet LAN available to plug it into? Mar 07 17:44:59 did test it with 2 macbooks Mar 07 17:45:01 :) Mar 07 17:45:19 and macos and linux hosts on them Mar 07 17:45:29 I do have LAN Mar 07 17:45:30 as a last resort you could try inserting an usb hub in the chain if you have one Mar 07 17:45:54 no, I'm a little short on devices now Mar 07 17:45:55 but to be honest, if the USB device port were broken on my BBB it would probably take months for me to even notice, I never use it XD Mar 07 17:46:06 I always connect via ethernet Mar 07 17:46:33 I think I probably even have usb device functionality disabled in my kernels Mar 07 17:47:17 I bought the damn thing just to practice in writing device drivers, so I was going to always plug it in with usb, ssh to it and start coding Mar 07 17:47:41 if you plug it into ethernet, it should be reachable as beaglebone.local Mar 07 17:48:11 you mean in the laptop or in the modem? Mar 07 17:48:22 because mac nowadays don't have ethernet Mar 07 17:49:14 assuming the modem is like a modern adsl/cable modem/router with both wireless and LAN ports, you can connect it to your modem and reach it from your laptop via wifi Mar 07 17:49:38 I think I tried that with a local port scanner Mar 07 17:49:41 it didn't find the board Mar 07 17:49:54 which image were you running? Mar 07 17:50:30 pre-installed angstrom Mar 07 17:50:38 it's a A5A revision with 2gb Mar 07 17:51:01 ok, angstrom is so ancient I have no idea what it may or may not be running Mar 07 17:51:11 but now there is debian Mar 07 17:51:14 some distros might also bork up if ethernet isn't plugged in already when you power it up Mar 07 17:51:43 yeah I recommend the latest jessie... if you're not using the GUI then you probably want the new "iot" image ( http://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Jessie_Snapshot_iot ) Mar 07 17:52:23 * zmatt will also quickly download it to see what rcn stuffed into that image Mar 07 17:52:51 that's 4 gb Mar 07 17:52:57 don't think it'll do Mar 07 17:53:25 given that it doesn't include any GUI stuff, I doubt it actually needs that much space... but hold on while I inspect Mar 07 17:54:17 well, I plugged the board to my router, to usb on the laptop and now what? Mar 07 17:54:54 which image are you running currently? Mar 07 17:55:49 you should be able to reach it using ssh debian@beaglebone.local if all is well Mar 07 17:56:10 http://beagleboard.org/latest-images Mar 07 17:56:20 Debian 7.5 (BeagleBone Black - 2GB eMMC) 2014-05-14 Mar 07 17:56:39 ick, ancient version Mar 07 17:56:51 I hope avahi is installed by default on that Mar 07 17:56:53 well, it worked when it was used :) Mar 07 17:57:19 yeah but I don't really know what is or isn't included in such ancient images, or how they are configured Mar 07 17:57:37 could not resolve hostname beaglebobe.local Mar 07 17:58:02 bobe? :P Mar 07 17:58:32 I have also used http://elinux.org/BeagleBoardDebian#Flasher Mar 07 17:58:43 and had same problem Mar 07 17:58:53 :D Mar 07 17:59:14 yeah, I typed it here myself) Mar 07 17:59:17 argh, annoying chicken and egg problems here... hard to diagnose what's going on when you're blind Mar 07 18:00:40 ok, thanks anyways, either I'm going to send this peace of ...hardware back to whoever sold it to me, or wait until uart-to-usb comes in post Mar 07 18:01:10 I'll be back :) Mar 07 18:01:12 I'll definitely poke rcn why he's making those iot images 4gb even though they would fit just fine on 2gb Mar 07 18:01:27 but Mar 07 18:01:29 you do that :) Mar 07 18:01:42 maybe I remember wrong about avahi being installed by default Mar 07 18:02:07 usually the web config of a modem shows which devices are on your network Mar 07 18:02:17 that's one last place you can check, assuming it has one Mar 07 18:03:20 though I can also understand if you're done with trying... RMA'ing does start to sound like a good option :) Mar 07 18:05:13 I'll never remember the username/password for it :D Mar 07 18:05:17 ok, bb Mar 07 19:51:08 layout guy to the software guy " i am sure u can read those cape eeprom on any of the i2c busses" .. me thinks -> thank u hardware layout collegues , u managed to break one of the view definitions we had for bbb capes. Mar 07 19:51:59 let me just redo the thing and make the kernel "detect" the stuff we need to "detect" the boards Mar 07 20:07:44 hello i have a question , anytime i try to configure the DT to make the PRUs working with some pins i have this from dmesg pruss_uio 4a300000.pruss: could not find pctldev for node /ocp/interrupt-controller@48200000, deferring probe Mar 07 20:07:55 anybody knows what is it? Mar 07 20:08:20 and no... google doesnt help Mar 07 20:09:10 [ 395.557356] bone-capemgr bone_capemgr.9: slot #7: #2 overlays Mar 07 20:09:10 [ 395.576601] omap_hwmod: pruss: failed to hardreset Mar 07 20:09:10 [ 395.583384] bone-capemgr bone_capemgr.9: slot #7: Applied #2 overlays. Mar 07 20:09:10 [ 395.623950] pruss_uio 4a300000.pruss: could not find pctldev for node /ocp/interrupt-controller@48200000, deferring probe Mar 07 20:09:10 [ 395.624009] pruss_uio 4a300000.pruss: pins are not configured from the driver Mar 07 20:09:27 kernel is 3.8 Mar 07 21:41:16 So I just flashed my emmc with the latest Jessie image, and now all 4 leds are flahing in unison - is that an error of some king or am I good to go? Mar 07 21:42:14 kind Mar 07 23:13:14 Hello... Mar 07 23:14:16 Is anyone working out of the book, "Getting Started with BeagleBone"? Mar 07 23:21:56 For instance...page 72, Example 6-3. Mar 07 23:30:21 Is there anyone in the house that knows Python for BBB/BBG and the Adafruit_BBIO.GPIO? Mar 07 23:36:15 I figured out my issue. It was a lower case "f" and then an upper case "F." Yikes! Damn Flask. Mar 07 23:51:50 That book has a lot of information. Most of it has nothing to do with "right now" in electronics with BBB/BBG and software. Mar 07 23:51:54 Day-um! Mar 07 23:52:42 Flask has changed, pip and setuptools for python have changed, and etc... Mar 07 23:52:43 Yikes! Mar 08 00:17:38 that's the big problem with putting anything computer-related into print Mar 08 00:17:45 blink your eyes and it's outdated Mar 08 00:19:01 Yep. Mar 08 00:19:26 I got code but I am getting a TemplateNotFound: main.html error in Flask. Mar 08 00:20:09 So, I went online after the initial error. I found out that something is wrong with my "Hello!" part of the code. I am just going to have to guess from here. Mar 08 00:20:17 not something I know anything about, sorry Mar 08 00:20:25 Cool, cool. That is okay. Mar 08 00:20:45 I went to the #pocoo page on this freenode thing. Mar 08 00:20:56 I am just waiting to get a response. Mar 08 00:21:32 zmatt: Did you get the update on the BBblue? Mar 08 00:22:49 ...something on holding off for a bit. Mar 08 00:25:03 hmm? Mar 08 00:25:38 Yep...I got one of the updates. It said something like, "As of now, we are on hold to produce the BBblue." Mar 08 00:25:54 I thought it was still in development Mar 08 00:26:17 Yep. Mar 08 00:26:41 It is...but the initial idea was that it would be towards the end of Feb. until the whole thing came together. Mar 08 00:26:44 This is what I thought. Mar 08 00:27:21 my understanding was May Mar 08 00:27:29 oh. I am probably wrong. Mar 08 00:27:33 Okay. Mar 08 00:29:09 I saw that it would have bluetooth and wifi. Mar 08 00:29:11 Cool! Mar 08 00:29:33 Easy for getting in and out of the shell. Cool! Mar 08 00:30:19 I would like to go outside with my BBG sometimes to get my GPS module up and running. Mar 08 00:30:38 I would have to have a very long Ethernet cable. Mar 08 00:31:48 I'm more curious about this thing -> http://beagleboard.org/blog/2015-11-23-beaglebone-enhanced/ Mar 08 00:34:30 I have not heard of SanCloud before. I will have to look into it. This model may help with my issues, i.e. more usb ports. Mar 08 00:36:43 More usb ports to me means that I can attach that USB powered sound card and still have accessible ports for a mouse and keyboard. Mar 08 00:39:37 zmatt, maybe you know something about this, so we are running an am335x@1GHz and DDR3 800MHz memory on our module, when we run a memset loop with a size between 25 to 2200 bytes we get flickering of the lcd screen, (above 2200 bytes memset the lcd works fine), got any ideas? as far as i'm aware, when we reduce the CPU speed it seems to happen less or not at all. my guess would be that calling memset in a tight loop (which at least one gstreamer Mar 08 00:39:37 1.6 plugin does) somehow clogs up the memory bus and the dma somehow is not working properly, but that's just a wild guess. Mar 08 00:41:32 filt3r: I'm assuming you mean DDR3-800 (i.e. 400 MHz) :P and there are lots of performance tuning knobs you can twiddle, e.g. to prioritize the LCD controller over the CPU Mar 08 00:42:04 if an underrun is occurring that should also show up in the kernel log Mar 08 00:42:15 yeh ddr3-800 :D, ok so you got any pointers, because we are seeing nothing in the kernel logs Mar 08 00:42:19 thats the strange thing Mar 08 00:42:24 hum Mar 08 00:43:55 that's kinda odd Mar 08 00:44:21 but then, I've seen all kinds of odd things happening with the lcd controller... dunno yet if it's hw or sw Mar 08 00:46:23 in any case, lcdc's dma controller has some tuning knobs you can configure via DT (dma-burst-sz, fifo-th) Mar 08 00:46:56 ok, i've seen them but could not get my hands on them yet Mar 08 00:47:14 and you can try twiddling its initiator pressure and/or MReqPrio, those are configurable via control module registers Mar 08 00:48:18 ok, that's something where me or some collegue can start looking into, thanks Mar 08 00:49:40 still, if an underrun is occurring it's weird that it's not reported Mar 08 00:53:24 that also confused me, but it seems like the data is there (on the screen) but just shifted Mar 08 00:53:50 ohey Mar 08 00:53:54 that's a familiar one Mar 08 00:54:00 but you're seeing that under X ? Mar 08 00:54:08 no we are using directfb Mar 08 00:54:12 ahh Mar 08 00:54:29 let me take a guess... the shift looks like it could be about 32 pixels? Mar 08 00:54:43 unfortunately no Mar 08 00:54:45 wait Mar 08 00:56:52 hmmmz, definitely different from the issues I've been seeing Mar 08 00:57:19 though 1:01 ... that shift does look like the one I'm referring to Mar 08 00:57:34 ok, so i might be something similar actually Mar 08 00:58:27 I've first seen the shifting issue on the framebuffer console, on various screens... rcn-ee reported he's also seen that issue all the way since the 3.8 kernels Mar 08 00:59:59 I'm currently using kernels with no fbcon at all, just direct linuxfb access... and I got an even weirder issue: the last pixel of each line is replaced by the one 32 pixels to the left Mar 08 01:00:16 consistently across numerous kernels Mar 08 01:00:20 and stably Mar 08 01:00:29 yet content of /dev/fb0 is correct Mar 08 01:01:23 since the same offset is involved I'm inclined to believe the issues are *somehow* related, but I'm truly mystified by it and can't even begin to form a mental model of what could be going on Mar 08 01:02:18 thats strange, for us the lcd is working nicely (expect the memset issue now) using direcfb and currently we are running something like 4.1.yy based on mainline Mar 08 01:03:02 ok, well we will take a look into the dt-options and the priorities and see if this helps Mar 08 01:03:30 yours is also interesting since I also see small jumps (that could be 32 pixels possibly), but it then degenerates into a complete mess Mar 08 01:04:30 see 0:17 - 0:20 Mar 08 01:04:57 ohh yes, this really looks like 32pixels Mar 08 01:05:40 so I think we have yet another flavor of the same underlying problem Mar 08 01:06:55 strange thing tho, if we clock the cpu down to 720MHz it does not happen anymore (although i haven't tested this personally) Mar 08 01:07:53 the fbcon variant is also unique since there the jumps are unpredictable except they're always in response to some event (e.g. typing something), never spontaneous Mar 08 01:09:41 that is interesting info, although so far it (like every observation so far) only seems to add to the mystery Mar 08 01:09:53 well the jumps in our issue are related to memset calls basically, in the video it's the wma decoder doing some memsets, but we have written a testprogramm where we run memset in a loop and it's basically always in the state like at 0:55 Mar 08 01:11:13 but when we add something like this after each memset call: { volatile int j; int i = 0; for (i = 0; i < 45; ++i) j = i + 2; } then the flickering only happens only occasionally increasing the loop to 50 seems to remove the flickering completly Mar 08 01:12:14 I have no idea if it could be related, but I have been wondering whether linux implements correct ocp barriers for memory coherency, e.g. to ensure data has arrived at EMIF before updating frame buffer pointers Mar 08 01:13:38 since doing so correctly is actually highly dependent on the SoC, in particular its interconnect topology (which is undocumented for the am335x!) Mar 08 01:15:07 though I don't see how exactly problems there could produce this issue Mar 08 01:15:36 ... but then again I don't yet see how anything whatsoever could manage to produce this issue Mar 08 01:17:27 I also participated in another lcdc-related issue thread on E2E that doesn't really seem related, but nevertheless still left me puzzled about what's going on with LCDC -> https://e2e.ti.com/support/arm/sitara_arm/f/791/p/446385/1653040 Mar 08 01:20:16 Hello everyone, JB here. I'm very frustrated with BBB at the moment. I'm trying to find a way to put the audio cape rev b and a LCD 7" touch screen on the BBB. Mar 08 01:20:35 I've used the 4D System 7" and it has a conflict. Mar 08 01:20:42 Any help or advice? Mar 08 01:25:35 no experience with off-the-shelf capes, sorry Mar 08 01:26:43 Thank you zmatt Mar 08 01:26:44 filt3r: it could be the 32-pixel shift and the total-bezerk things are different issues of course Mar 08 01:27:01 yes Mar 08 01:27:18 filt3r: but you're clearly having the 32-pixel shift too, a lot in fact Mar 08 01:27:42 ok i just quickly checked the tilcdc driver and it should give an kernel error on LCDC_SYNC_LOST and LCDC_FIFO_UNDERFLOW Mar 08 01:28:07 according to the other thread, LCDC_SYNC_LOST wasn't actually enabled, but maybe that's been fixed since then Mar 08 01:30:58 btw I'm not 100% sure what I did to get a copied-pixel issue instead of the shift-issue, but it may have been due to disabling fbcon entirely (CONFIG_VT=n) Mar 08 01:32:06 so you could add that to the list of experiments you could try Mar 08 01:32:08 :) Mar 08 01:32:35 sure Mar 08 01:33:25 but just one thing the thread uses the CONFIG_FB_DA8XX driver and we are using the CONFIG_DRM_TILCDC driver, which at least seems to check the sync and underflow flags in the isr Mar 08 01:33:39 and should print an error message Mar 08 01:34:00 I'm pretty sure the thread was also using tilcdc... or iirc he tried both of them Mar 08 01:34:07 ahh ok Mar 08 01:35:10 ok, thanks zmatt, at least i got some pointers now and let's see how this works out Mar 08 01:35:29 yeah I'm also debugging my flavor of the issue Mar 08 01:35:37 or rather, pondering where the hell to begin Mar 08 01:35:45 :D i'll keep you updated Mar 08 01:36:20 at least my variant has the useful property that it's perfectly stable... it's present even if I just cat a static image into /dev/fb0 Mar 08 01:37:12 ye that's one thing i wanted to try but had some more important stuff to do, but i'm like 90% sure the issue will still be the same Mar 08 01:41:43 btw, if you want to try fiddling with the priority settings in control module, beware that it ignores unprivileged writes so you actually need a bit of a hack to fiddle with those settings from userspace Mar 08 01:42:36 this hacky little project of mine actually shows a trick to do it -> https://github.com/dutchanddutch/jbang/ (it also contains a header file for the am335x control module) Mar 08 01:43:41 oh wow that sounds really hacky and cool "self-jtag without external hardware by manipulating the pinmux registers to make the processor think the pins are toggling." Mar 08 01:43:41 :D Mar 08 01:43:59 hehe, yeah Mar 08 01:44:29 and since pinmux registers are in the control module, I ran into aforementioned problem Mar 08 01:44:53 but I made a nice little helper class (src/privileged.h) which lets me do https://github.com/dutchanddutch/jbang/blob/master/src/hw-subarctic.cc#L9 Mar 08 01:47:52 ah so it's using process_vm_readv to read from memory and write to some other address (register) and since the syscall runs in the kernel it works Mar 08 01:48:02 yep Mar 08 01:48:16 ok, thanks this will come in handy Mar 08 01:49:26 https://github.com/dutchanddutch/jbang/blob/master/include/ti/subarctic/ctrl.h#L332 <-- and that register controls priority on the L3 interconnect (0 = low, 1 = medium, 3 = high) Mar 08 01:50:44 which bits? 0-1? Mar 08 01:50:51 2 bits per initiator Mar 08 01:51:19 cortex-a8 is bits 0-1, lcdc is bits 54-55 Mar 08 01:51:21 ahh i see it now lcdc 54-55 Mar 08 01:51:25 ye Mar 08 01:51:39 ok cool Mar 08 01:52:14 priotiziation inside EMIF itself is controlled by mreqprio, which for some initiators is controlled by a register a bit further down (offset 0x670) Mar 08 01:52:51 lcdc isn't there though, so its priority might be coming from lcdc itself... I recall it has a dma priority config field somewhere Mar 08 01:53:59 range is 0-7, don't remember whether lower value is lower or higher priority though Mar 08 01:59:41 I really wish C/C++ would allow things like an array of 2-bit integers :P Mar 08 02:01:16 a Nx2 boolean array? Mar 08 02:02:50 just u2 foo[32]; or something like that (stored in 64 bits) Mar 08 02:06:03 ok, zmatt your info and the TRM should be enough to twiddle with the priorities, thanks again :D Mar 08 02:06:29 you're welcome :) let me know how it goes Mar 08 02:06:32 sure Mar 08 02:08:33 btw, although the jbang README says it requires gcc 4.9, I think it may actually need gcc 5 currently... I never really test with older compilers Mar 08 02:15:33 zmatt, one thing i found in http://www.ti.com/lit/er/sprz360g/sprz360g.pdf look on page 30 "Advisory 1.0.27" Mar 08 02:16:22 but im not sure if this is related to the issue we are having Mar 08 02:16:25 LIDD mode isn't relevant here Mar 08 02:16:29 ahh ok Mar 08 02:16:31 that's for character displays Mar 08 02:17:00 LCDC is kinda two-peripherals-in-one Mar 08 02:17:10 LIDD and the raster controller are mutually exclusive Mar 08 02:17:21 ahh ok TIL Mar 08 02:17:38 "The LCDC peripheral cannot operate in LIDD mode and raster mode simultaneously, so only one type of display can be used at a time." Mar 08 02:17:56 it's still funny when I pointed out the existance of the LIDD controller to someone who had just implemented the same using GPIOs Mar 08 02:19:19 so lidd is for HD44780 style lcds? Mar 08 02:19:25 for example yes Mar 08 02:19:41 it supports three styles of bus interfaces Mar 08 02:19:52 see LCDC tab of my pins spreadsheet ( https://goo.gl/Jkcg0w ) Mar 08 02:20:22 well i don't know how good the support for lidd is from the kernel point of view, but i think for some people it might be faster to implement the protcol using gpios instead of writing a lidd driver from scratch (in case there is none in the kernel) Mar 08 02:20:37 he wrote a kernel driver Mar 08 02:20:56 and you could easily use LIDD from userspace directly Mar 08 02:20:58 (using uio) Mar 08 02:23:10 yes, but first you have to know that the hardware can do it :) Mar 08 02:23:16 true Mar 08 02:23:43 in general it's useful to at least roughly know what the chip you're working with can do ;) Mar 08 02:24:55 I've so often seen people do silly things with GPIOs for which the am335x has nice built-in peripherals Mar 08 02:27:49 ye, it's probably a good idea to check for built-in support before you start writing a gpio driver **** ENDING LOGGING AT Tue Mar 08 02:59:59 2016