**** BEGIN LOGGING AT Thu Nov 12 02:59:58 2015 Nov 12 03:00:01 BBBlfs is blowing my mind Nov 12 03:00:54 least ur getting blown Nov 12 03:00:58 (Trying to understand how it all works. Is the Beaglebone Black some massive composite USB device? Why are there all these different VID / PIDs available) Nov 12 03:01:25 YES Nov 12 03:01:28 10 points to uudruid-74 Nov 12 03:01:44 Whooo Hooo Nov 12 03:01:54 I think every time I move the magnet away from the connected device, I get an e-mail. "Further testing is in order." Nov 12 03:01:56 What can I buy with them? Nov 12 03:02:41 I think it has to be away from the connected device for over 60 seconds. I will have to check the code. Nov 12 03:02:45 since the points are made up, you could conceivably make up anything to buy them with Nov 12 03:03:00 *with them Nov 12 03:03:22 anything? ANYTHING???? Nov 12 03:03:39 hmmm Nov 12 03:03:45 * uudruid-74 thinks Nov 12 03:03:49 You can buy my in unwavering loyalty for about 5 milliseconds Nov 12 03:03:54 *my unwavering Nov 12 03:04:15 * GenTooMan puts an relaxation oscilator next to Set_ open door detector should be good for a few grand worth of messages. Nov 12 03:04:35 No one comes by...I should get no messages. Nov 12 03:04:49 But...in case. I get robbed or people hate me one day. Nov 12 03:04:53 Yep...I will know! Nov 12 03:05:02 Ha! Nov 12 03:05:40 My dooralert.py file does not work. It shows no problems but it makes PuTTY go aloof. Nov 12 03:05:55 I think I got an issue. Back to work! Nov 12 03:06:15 Actually it's an SoC for 'industrial' type stuff. Although most industrial doesn't need or want something on a BGA like that due to humidity vibration and rocks. Nov 12 03:06:45 well Set_ we were going to tell you about it but ... Nov 12 03:06:52 Wait...what? Nov 12 03:07:01 your "issue" :D Nov 12 03:07:14 How do you know my issue? Nov 12 03:07:21 Just asking. Nov 12 03:07:35 * GenTooMan tries not to fall out of his chair snickering. Nov 12 03:08:00 Come on...I need to go to work on the BBB and its multi-use capabilities. Nov 12 03:08:18 It is time to scare my family for spying on me when I am not home. Nov 12 03:08:26 Go for it.. although door alarm probably is a bit pricey way to use it. Nov 12 03:08:37 Oh... Nov 12 03:08:42 What would you do? Nov 12 03:08:45 Instead? Nov 12 03:09:10 I am talking...I have the 10" sub already! Nov 12 03:09:10 You have IO connect them to a use number of things and use it to 'watch' the house? Nov 12 03:09:23 I know. Nov 12 03:09:38 I think not knowing the reaction until confronted is far better. Nov 12 03:09:49 Boom...BOom....Boom! Nov 12 03:10:09 you'll have pictures of things like "the mail was delivered 2 minutes early" Nov 12 03:10:14 Have some, "Booty Mix '96" blaring out of the sub. Nov 12 03:10:28 Hhahahaha. Nov 12 03:10:48 Clap it like Donovan...boom, boom, boom! Nov 12 03:10:48 Why not have some "house sounds" like a door closing etc? Nov 12 03:10:59 Oh...that is okay. Nov 12 03:11:15 Hey...what do you all use for mic and speaker settings via BBB? Nov 12 03:11:28 what online program? Nov 12 03:11:41 I tried alsamixer but I am having trouble with it. Nov 12 03:11:54 I would use Audacity for recording and editing sounds. Nov 12 03:12:03 Okay...thank you. Nov 12 03:12:22 I've edited things up to 90 minutes long without issue. Nov 12 03:12:22 I saw something about audacity a while back but I did not look into it. Nov 12 03:12:28 Cool, Nov 12 03:12:52 It seems to only import wave but can export a number of formats. Nov 12 03:12:55 I will look into it. Nov 12 03:13:13 that is alsamixer too. It uses .wav files. Nov 12 03:13:41 I found that program by using a Packtpub book. Nov 12 03:13:47 Audacity is specifically for editing and modifying sound you can have any number of tracks with it. Nov 12 03:14:12 so, apt-get audacity Nov 12 03:14:12 ? Nov 12 03:14:41 apt-get install audacity Nov 12 03:14:42 ? Nov 12 03:15:17 You could use it on windows edit the files then store the files. For playing you probably could use "play" which is part of sox. it will use your current sound device under linux. Nov 12 03:15:49 okay... Nov 12 03:15:56 timidity++ is handy if you want to play something like beethovens fifth or the anvil chorus. Nov 12 03:16:33 Hey Gen...I am on smoke break. I need to go for a smoke. I will right back. Nov 12 03:18:31 letting the magic smoke out? Nov 12 03:19:37 up in smoke ... Nov 12 03:22:30 good night people. brain is already off. Nov 12 03:22:42 Okay... Nov 12 03:22:43 later Nov 12 03:22:56 Thank you for the adice. Nov 12 03:23:08 adice = advice Nov 12 03:24:23 ds2...nope. Just tobacco. Nov 12 03:26:42 not what he meant Nov 12 03:27:02 electronics work because they have magic smoke inside Nov 12 03:27:14 Oh. Nov 12 03:27:29 I am not the smartest person. Nov 12 03:27:41 If you let the magic smoke out if those little boxes called transistors, then they wont work anymore Nov 12 03:27:50 Aw! Nov 12 03:27:54 Eureka! Nov 12 03:27:56 Hhahahahaha. Nov 12 03:28:08 I get it now. Nov 12 03:28:26 Old EE joke my electronics teacher used to tell Nov 12 03:28:57 This code is bunk. It checks on the door being open for a whole minute. Nov 12 03:29:02 I have to change it. Nov 12 03:29:12 No one leaves a door open for a minute. Nov 12 03:29:22 Open and close..."ya mean?" Nov 12 03:29:27 Hhahaah. Nov 12 03:29:44 Please hold. Nov 12 03:35:13 The code is supposed to keep e-mailing me if the door is open more than seconds but it will not. Nov 12 03:35:28 I have to keep working. Nov 12 03:35:33 Sheesh. Nov 12 03:37:21 Well, you have the code. Change it! And I bet the guys in #python can give you plenty of pointers Nov 12 03:37:51 I changed it. I am waiting to see if it works or not. Nov 12 03:37:55 I am about to run it. Nov 12 03:38:31 I get a message every 11 seconds. Either it says blah, blah, blah or something else. Nov 12 03:38:44 Depending on if the door stay open more than 2 seconds or not. Nov 12 03:41:30 You've got to set a timer to keep checking it Nov 12 03:41:55 Depending on exactly how you want it to behave Nov 12 03:43:38 Oh. Nov 12 03:43:39 Okay. Nov 12 03:43:44 I have the timer set. Nov 12 03:43:51 I think. Nov 12 03:46:17 I need to check my wiring now. Nov 12 03:46:37 I think I could have made a mistake. Nov 12 03:47:25 Nope...I got that. Nov 12 03:51:11 I think I may have complications due to the amount of time it takes for my hard drive mail to expose e-mails. Nov 12 03:51:37 i.e. e-mails from gmail.com are automatic. Nov 12 03:52:02 But...it takes time for the hard drive e-mail to show e-mails. Nov 12 03:56:43 hard drive emails??? Nov 12 03:57:18 yea...um. What is it called. Outlook. Nov 12 03:57:32 yea...Outlook. Nov 12 03:58:11 I thought since my Outlook had different upload times, I could have a timing issue. Nov 12 03:58:55 The remote virus shell Nov 12 03:59:15 I have a remote shell. Nov 12 03:59:22 Yes... Nov 12 03:59:49 but my gmail.com account uploads quickly and my Outlook takes uploads every five minutes. Nov 12 04:00:17 * uudruid-74 hands to Set_ a 'n00b' sticker Nov 12 04:00:30 Yes sir! Nov 12 04:02:01 The code is running continuously but the output is not working like thought. Nov 12 04:02:07 Yikes! Nov 12 04:03:29 1 - Figure out what you want it to do, step by step Nov 12 04:03:51 2 - Find how to code each step. Nov 12 04:05:58 For example, when I said Timer I mean that you seem to only want it to send mails at a certain rate. So you set a flag when the mail I set, then clear the flag after the timer expires. If the flag is set, do not send mail. Nov 12 04:06:41 it does that but it also does something else. Nov 12 04:06:57 As to how to do that, the guys in #python could help better than us. Nov 12 04:07:06 It shows that if it is the sensor is not attached for "n" seconds, send another e-mail. Nov 12 04:07:13 Okay. Nov 12 04:07:16 No problem. Nov 12 04:07:26 I will not harass you further about it. Nov 12 04:08:21 That sounds like the opposite if what you want, repeated emails over and over. :-) Nov 12 04:09:44 I know...I figured I would just type out this code to send off an e-mail every time the door is opened and if the door stays open for more than 2 secs. Nov 12 04:09:51 And its not harassment, just bound to get better advice from python specific stuff in the python channel since your issue isn't specific to beaglebone Nov 12 04:10:04 Okay. Nov 12 04:10:20 I will look into #python soon. Nov 12 04:11:14 And if you want to log into your bone from an Android phone, JuiceSSH is awesome. I hate putty ... I'd rather use my phone than use Windows Nov 12 04:11:31 Okay. Nov 12 04:13:33 Okay udruid. I am taking off now. I hope to see you again. Have a good day. Nov 12 04:13:36 Later for now. Nov 12 04:13:40 I popped in to ask about power settings. I have a 750GB drive on my BBB and it didn't have power saving. It was actually overheating Nov 12 04:13:43 cya Nov 12 04:14:41 Anyway ... whoever shot me that link, its working GREAT. Drive is now COLD to the touch and not too hot to touch! Nov 12 04:28:50 uudruid-74: when forced to windows I use msys so I can have a proper shell and ssh and stuff Nov 12 05:50:10 hi everybody, is this a chatroom where we can get help about beaglebone black? Nov 12 05:50:50 sure Nov 12 05:51:08 ask technical questions, wait patiently for answers... Nov 12 05:55:48 so a simple question: I have the BBB connected to USB power (from my laptop) and connect via micro-hdmi to DVI to a monitor. This should theoretically display video right? Nov 12 05:56:46 I've looked up up online, and all sources so far have said I don't necessarily need the DC power for this. This is correct right? Nov 12 05:59:38 works for me Nov 12 06:01:12 as long as the usb port can supply enough current Nov 12 06:01:42 (and keep in mind that current may vary considerably based on activity, so "works" need not imply "works reliably") Nov 12 06:01:53 I have no extra peripherals, so I think the current is sufficient! Nov 12 06:02:14 I'm just worried because I'm getting to output on my monitor whatsoever! Nov 12 06:02:44 which kernel version? Nov 12 06:03:46 are you able to ssh to the BBB ? Nov 12 06:06:39 some kernels still have broken pixel clock frequency calculation, which effectively means that whether or not a particular video mode is supported is decided by fair dice roll ;-) Nov 12 06:07:20 sorry I disconnected, but yes I ssh'ed into the BBB and using uname -r, I got 3.8.13-bone70 Nov 12 06:12:34 ah, that old... that predates the tilcdc driver entirely, it still used an old different one, so I don't know in that case Nov 12 06:13:51 so I should consider updating this kernel? Nov 12 06:15:15 you can try running my lcd-util http://gerbil.xs4all.nl/lcd-util.tgz (after it displays the settings and says press enter to change polarity, press control-C instead) Nov 12 06:15:29 if lcdc is badly misconfigured it'll tell you Nov 12 06:15:35 also check the kernel log (dmesg) Nov 12 06:16:40 (lcd-util directly inspects the state of the hardware hence should be kernel-independent) Nov 12 06:16:56 zmatt: (OT) do you know if lcdc can do interlaced output? Nov 12 06:18:39 Thanks! I'll check the kernel log Nov 12 06:19:13 ds2: no... I mean, you could feed it successive fields of course, but you'd need to make the even/odd field signal somehow Nov 12 06:19:51 'k Nov 12 06:20:08 so much for a quick experiment in superesolution Nov 12 06:20:08 ds2: it's possible to make PRU run synchronous to the lcd clock rather than the core clock, and use it to generate the desired synchronization signals Nov 12 06:20:27 you want more than 1080p ? Nov 12 06:22:50 note that the hdmi framer is very flexible in reconstructing sync signals so it may be possible, *if* the framer supports such insane rates and you can establish initial even/odd synchronization Nov 12 06:23:03 *insane resolutions Nov 12 06:23:51 but what did you have in mind? considering monitors afaik normally don't do interlaced, and I don't think TVs normally do more than 1080p Nov 12 06:25:51 1920x1080R@30 is the highest I've ever run it at (and it worked) Nov 12 06:34:34 hey matt, I extracted the tgz file and ran the code, and it said that the LCD controller clock source: 0. And then main osc: 24 MHz and also "### fatal: LCDC not enabled" Nov 12 06:34:57 try rebooting the BBB and running the tool shortly after boot Nov 12 06:35:04 Okay! Nov 12 06:39:28 Okay, I've tried it 3 times, and they all result in LCDC not enabled Nov 12 06:43:49 anything in kernel log? sounds like it doesn't understand the monitor at all, e.g. an EDID problem Nov 12 06:44:34 znattL bi bi Nov 12 06:44:37 zmatt: no no Nov 12 06:45:07 zmatt: Was reading about how you can double the resolution of a projector by applying a displacement Nov 12 06:45:35 it isn't the absolute resolution; was looking for a way of experimenting with the technique Nov 12 06:46:40 TeddyTea: trying a new kernel may be worthwhile, although if you have an old system it may be wiser to just reflash a new one than update it (although usually updating just the kernel of an old system does work I think) Nov 12 06:46:52 ds2: you mean physically displace the monitor in the vertical plane? Nov 12 06:47:33 to create effective interlacing Nov 12 06:47:35 hmm Nov 12 06:48:34 that could work, but you'd have timing issues anyway since the projector in that case doesn't know it's an interlaced signal, rather the field data needs to go to whatever actuator is being used Nov 12 06:49:40 it needs to sync to when the data is actually *displayed* by the projector, not when it's delivered to it, so I wouldn't expect a normal field-output to work even if lcdc had one Nov 12 06:50:12 zmatt: well, a projector Nov 12 06:50:12 zmatt: Does this mean the hdmi cape is disabled? (From kernel log) "command line: console=ttyO0,115200n8 capemgr.disable_partn o=BB-BONELT-HDMI," Nov 12 06:50:19 more like this Nov 12 06:50:27 TeddyTea: yes Nov 12 06:50:29 project odd, display vertically, project even, repeat Nov 12 06:50:33 TeddyTea: or wait, no Nov 12 06:50:39 TeddyTea: ie means hdmi audio is disabled Nov 12 06:50:43 *it Nov 12 06:50:45 it should double the resolution Nov 12 06:51:06 ds2: s/display/displace/ ? Nov 12 06:51:15 yes, displace Nov 12 06:51:17 ds2: and that's exactly what I meant Nov 12 06:51:20 can't type tonight Nov 12 06:51:37 as long as the LCDC can rip the display into even and odd pieces, that's the hard work Nov 12 06:51:57 interlaced video is natively always transmitted as separate fields Nov 12 06:52:09 generating the signals to sync things can be done with an external FPGA. w/o LCDC interlacing, I'd have to buffer Nov 12 06:52:17 zmatt: yes, that is exactly what I want Nov 12 06:52:29 that's the normal way interlaced video is stored already Nov 12 06:52:32 make projector think even field is a full frame Nov 12 06:52:49 all you need to do is refrain from deinterlacing it Nov 12 06:53:00 hmm? Nov 12 06:53:25 keep in mind interlaced video fields aren't just even/odd lines, they're even/odd lines of *different frames* Nov 12 06:53:34 i want memory to be L1L2L3L4L5L6L7L8 but LCD to send out L1L3L5L7 VSYNC L2L4L6L8 VSYNC .... Nov 12 06:53:38 even lines of frame 0, odd lines of frame 1, even lines of frame 2, etc Nov 12 06:54:03 why would you want memory to be like that? Nov 12 06:54:06 yes but for a static image, frame N == frame N+1 Nov 12 06:54:19 so the app don't have to rip up the image Nov 12 06:54:29 ok you're not actually displaying video Nov 12 06:54:51 yes, trying a simple science experiment Nov 12 06:56:08 so, even if lcdc natively supported interlaced video, it still wouldn't do that for you Nov 12 06:56:22 since native interlaced video isn't stored or transmitted like that Nov 12 06:56:34 (like that = your "i want memory to be") Nov 12 06:56:59 i don't understand why it wouldn't be done that way Nov 12 06:57:31 are you saying it is stored as L1L3L5L7L2L4L6L8 in memory? Nov 12 06:57:58 I say your line numbering is wrong Nov 12 06:58:13 or well Nov 12 06:58:19 no they aren't I geuss Nov 12 06:58:29 but the odd lines belong to odd fields Nov 12 06:58:32 and even lines to even fields Nov 12 06:58:42 and video is a sequence of fields Nov 12 06:58:42 okay Nov 12 06:59:14 okay. no disagreement there. maybe there is a terminology difference Nov 12 06:59:49 interleaving them in the way you suggest makes no sense for the way interlaced video is normally handled Nov 12 07:00:32 i.e. think of a CRT monitor (which is the only kind of monitor natively supporting interlaced video): it does exactly what you're trying to do with the projector Nov 12 07:00:49 the electron beam traces all odd lines for odd fields and all even lines for even fields Nov 12 07:00:52 okay Nov 12 07:01:00 note that every field is transmitted as a frame, i.e. with a vsync Nov 12 07:01:13 yes and that is exactly what I want Nov 12 07:01:25 that's the native format of interlaced video Nov 12 07:01:34 the only detail is syncing up with the even/odd Nov 12 07:01:50 otherway you'd get ... funny output Nov 12 07:01:53 yes Nov 12 07:02:05 NTSC has a sync pattern and half line to do that Nov 12 07:02:24 yeah, that's the trick used when the total number of lines is odd Nov 12 07:02:35 or it's not even really a trick Nov 12 07:02:54 but I don't see why there can't be a O/E signal like what the NTSC video decoder chips provide Nov 12 07:03:09 (if the number of lines is odd, divided by two, and vsync is still period, is necessary ends up halfway a line) Nov 12 07:03:11 or prehaps we are talking about the same thing with different words and going in circles :D Nov 12 07:03:13 they do Nov 12 07:03:22 commonly called 'fld' Nov 12 07:03:48 lcdc lacks it... but it's sync timing isn't standards-compliant anyway Nov 12 07:04:20 the hdmi framer is programmed to trigged on their edge and reconstruct correctly timed syncs Nov 12 07:04:32 *its Nov 12 07:06:01 but as I said, if the project doesn't understand interlacing and you're creating the displacement with an actuator, you may need the fld output to lag reality anyway, depending on the display latency of the projector Nov 12 07:06:06 wonder if the dispc can do it then Nov 12 07:06:22 yes that's a different "fun" Nov 12 07:07:14 so my suggestion of running pru in sync with lcdc (switch pruss clock over to the display PLL) and have it create a field output with adjustable timing may not be a bad idea Nov 12 07:07:48 requires pixel clock rate <= 100 MHz though to avoid running PRU out of spec Nov 12 07:07:52 yes but that doesn't help sort out the memory Nov 12 07:08:15 and w/o that, I might as well do it in a CLPD outside Nov 12 07:08:45 so, since you have two alternating fields you're displaying video whether you want to or not Nov 12 07:08:48 tell the LCDC it is 2x the Y resolution and have the CPLD generate the real VSYNC Nov 12 07:09:16 but your idea of a PRUSS sync'ed to video is interesting Nov 12 07:09:37 any thoughts on how to sync to frame start on the PRUSS? since L4 access could take too long Nov 12 07:09:38 still requires sorting the lines, unless the CPLD can buffer a full frame of video Nov 12 07:10:28 CPLD just needs to start on VSYNC, count HSYNC and twiddle the real VSYNC as needed Nov 12 07:10:54 or maybe tell LCDC it is 2*Y + vsync length Nov 12 07:11:00 make the image into continuous video, run it through mencoder or something to convert it to interlaced, then find some way to display it without getting it deinterlaced again Nov 12 07:11:25 note that lcdc is actually perfect for alternating between two images since it does that natively Nov 12 07:11:43 but I'm not sure how to take advantage of that with the normal kernel driver Nov 12 07:12:03 bah... all the display drivers are crap :D when in doubt, rework it Nov 12 07:12:34 look at all the revs of the dispc driver Nov 12 07:12:41 just make the framebuffer twice as high as you said, but program the two halves into lcdc's two frame buffer pointers Nov 12 07:12:52 and disable irqs, you won't need them anymore ;) Nov 12 07:13:03 lcdc will alternate between the two frames Nov 12 07:13:19 I like that Nov 12 07:13:43 userspace will still need to sort the lines Nov 12 07:13:50 *nod* Nov 12 07:13:52 but there's no escape from that Nov 12 07:13:56 well Nov 12 07:14:06 it's a shame lcdc's dma controller doesn't support line stride Nov 12 07:14:19 you could I suppose setup another framebuffer and have edma perform the copy Nov 12 07:14:25 actually there is another trick Nov 12 07:14:36 tell LCDC X resolution is 2X Nov 12 07:14:46 and fix up HSYNC externally Nov 12 07:14:55 that helps.. how exactly? Nov 12 07:15:10 then have LCDC alternate at FB start and FB start + line offset Nov 12 07:15:24 the first X pixels will alternative automatically Nov 12 07:15:34 I am back! Nov 12 07:15:35 but your line length will still be twice the desired value Nov 12 07:15:38 Whoo HA! Nov 12 07:15:54 yes, hence the need for HSYNC to be fixed up externally Nov 12 07:16:04 but your line length will still be twice the desired value Nov 12 07:16:13 hsync doesn't change that timing aspect Nov 12 07:16:20 so the CPLD will count out X PCLKs and send a real HSYNC Nov 12 07:16:23 you'd need to change the surplus into horizontal blanking Nov 12 07:16:36 no need for a cpld, the hdmi framer already does this for you right now Nov 12 07:16:49 since, as I said, the hsync timing is broken anyway Nov 12 07:16:50 yes, when I say HSYNC, I mean HSYNC signal + any blanking Nov 12 07:16:59 and porches Nov 12 07:17:03 Okay...my LEDs are moving from right to left and back. I am getting a new image. Nov 12 07:17:04 yes but then you have a HUGE amount of horizontal blanking Nov 12 07:17:12 and your pixel clock rate must be twice as high as needed Nov 12 07:17:17 which may get you into trouble Nov 12 07:17:19 When do I know that the image is done? Nov 12 07:17:31 Set_: all leds stay on Nov 12 07:17:34 well.. the stuff I want to experiemnt with is like 640x480 Nov 12 07:17:42 2x PCLK is not an issue Nov 12 07:17:42 Okay. Nov 12 07:17:44 Thank you. Nov 12 07:17:46 ds2: @ which framerate? Nov 12 07:17:58 zmatt: 60 or 30 Nov 12 07:18:04 field rate I should say Nov 12 07:18:13 not blinking right? Not all four LEDs blinking? Nov 12 07:18:26 no, steady on Nov 12 07:18:31 Okay. Thank you. Nov 12 07:18:35 Sorry to interrupt. Nov 12 07:18:46 zmatt: reading a book on various display technologys and manipulations so there is an interest in testing out ideas Nov 12 07:19:33 Do you all know of alsamixer and how to configure it with a Daffodil USB thing with a BBB? Nov 12 07:20:00 I can only get input into my BBB with alsamixer and no output. Nov 12 07:20:04 I think my speaker is junk. Nov 12 07:20:08 Who knows? Nov 12 07:21:12 I get feedback only. Nov 12 07:21:45 I think the Daffodil USB stick can only produce voice input with alsamixer. I do not think it is the BBB. Nov 12 07:21:59 ds2: but yeah, if the projector is okay with such a deviant amount of hblank you could abuse this to get the desired line stride... I think for a static image, doing it in userspace makes more sense Nov 12 07:22:18 or, as I said, edma can perform the job in a single transfer Nov 12 07:23:23 zmatt: this is the book in case you have read it - http://www.displaysbook.info/ Nov 12 07:23:47 I haven't, I'm not really into video stuff Nov 12 07:24:09 it isn't just video... talks about 2D displays, 3D displays, etc Nov 12 07:24:23 Damn...the dude from alsamixer says, and I quote, "It is there own damn fault." Nov 12 07:24:26 ebook is free Nov 12 07:24:45 He is speaking of terminal emulators that do not allow for ncurses/alsamixer. Nov 12 07:24:53 It could be that PuTTY does not work. Nov 12 07:25:02 putty is pretty broken Nov 12 07:25:05 Oh. Nov 12 07:25:15 also, by default reports a wrong value of TERM Nov 12 07:25:25 What are other people using these days. Nov 12 07:25:27 it should be putty to correctly reflects its capabilities Nov 12 07:25:32 but often claims xterm Nov 12 07:25:55 which will bite it in the ass, since xterm is a highly capable terminator emulator, much more so than putty Nov 12 07:26:05 Oh. Nov 12 07:26:10 xterm? Nov 12 07:26:17 I guess I can look it up. Nov 12 07:26:29 fixing the TERM variable may suffice to make it work Nov 12 07:26:40 Sorry...I am interrupting again. Nov 12 07:27:02 there's a setting in putty for it, but you can also try doing "export TERM=putty" before starting alsamixer Nov 12 07:27:12 okay. Nov 12 07:27:47 How would I go about fixing this TERM variable? Nov 12 07:27:51 though I didn't think putty was too crappy to display ncurses Nov 12 07:27:58 by doing exactly what I just said Nov 12 07:28:02 Okay. Nov 12 07:28:16 I wanted to go into PuTTY and configure it. Nov 12 07:28:25 I do not want to use my BBB to configure it. Nov 12 07:28:35 note that I haven't verified whether terminfo for putty exists but I said this mentioned on the xterm site recently Nov 12 07:28:50 Oh... Nov 12 07:28:52 Okay. Nov 12 07:29:05 I thank you. Nov 12 07:29:11 You are very helpful. Nov 12 07:29:12 hence try it first on the BBB, if it works then locate the putty setting and change it Nov 12 07:29:25 what should I change it to? Nov 12 07:29:27 I don't know where in putty it is since I don't use putty if you can avoid it Nov 12 07:29:34 Ha! Nov 12 07:29:41 *if I can Nov 12 07:29:52 Alright. I am going to try. Nov 12 07:30:01 I have to wait until my new image is done. Nov 12 07:30:15 I am using an older BBB for the bot I am making. Nov 12 07:30:29 This one, the one I am signed on with, is my newest one. Nov 12 07:30:32 I just checked, terminfo for "putty" does indeed exist Nov 12 07:30:37 yea! Nov 12 07:31:14 so locate the setting for the TERM variable (i.e. which terminal it should claim to be) and change it from "xterm" or whatever to "putty" Nov 12 07:31:30 Okay. Nov 12 07:32:11 alsamixer works fine with TERM=putty btw (even in xterm, which makes sense since putty is probably just a subset of xterm) Nov 12 07:32:44 If that is the case, I cannot figure out what is wrong. Nov 12 07:33:06 I did exactly what was explained and I thought that the info. was correct. Nov 12 07:33:16 well, life is hard sometimes Nov 12 07:33:20 Yep. Nov 12 07:33:40 I can only get voice input and I have to have output for command returns. Nov 12 07:33:46 Dang! Nov 12 07:34:42 ds2: anyway, I think you have plenty of options on the BBB itself :) Nov 12 07:35:58 ds2: I must admit I'm slightly sceptical about doing this with an lcd projector since I think it projects a continuous image, so I fear that although the resolution may be increased in theory, it will be negated by motion blur Nov 12 07:38:39 there is a new version released the last days Nov 12 07:38:44 perhaps rty that Nov 12 07:38:47 try Nov 12 07:39:03 uavcam: I think Set_ is now, hopefully, flashing a recent image Nov 12 07:41:06 ds2: some vertical pre-sharpening might help to compensate, if the dynamic range of the image leaves some headroom for that Nov 12 07:50:13 dang... Nov 12 07:50:22 That flashing is taking forever. Nov 12 07:50:53 It is blinking all four LEDs now and has been for as long as I have been aloof. Nov 12 07:51:19 I just sent the people at alsamixer some input about what hardware I am using. Nov 12 07:51:45 I asked them if they are having any complications with certain USB configurable devices. Nov 12 07:52:07 e.g. specifically the Daffodil USB Sound Adapter. Nov 12 07:52:43 I don't know exactly how long flashing a 4gb image takes, but I think I've seen the number "45 minutes" here (but I may remember incorrectly, and it'll depend on whether your BBB has Kingston or Micron eMMC) Nov 12 07:53:01 Oh... Nov 12 07:53:12 I just downloaded this image the other day to my newer model. Nov 12 07:53:17 It took 15 minutes. Nov 12 07:53:31 The older model has been having complications with this image. Nov 12 07:53:36 your newer model... you mean this one isn't a rev C ? Nov 12 07:54:03 I did get another, newer image, I think 3/2015, to work on the older model. Nov 12 07:54:15 Oh no...I am on Rev. C now while online. Nov 12 07:54:25 The one that is downloading the image is an older model. Nov 12 07:54:38 It says A5C. Nov 12 07:54:40 in other words, you're trying to flash a 4 GB image onto a 2 GB eMMC Nov 12 07:54:46 Ha. Nov 12 07:54:48 I guess. Nov 12 07:54:54 surprise, that won't work :P Nov 12 07:55:08 It worked with the newer 03/2015 image. Nov 12 07:55:20 I will probably have to resort back to that one. Nov 12 07:55:23 that will probably have been a 2gb image Nov 12 07:55:24 no Nov 12 07:55:34 No? Nov 12 07:55:35 get the 2gb version of latest one Nov 12 07:55:38 Okay. Nov 12 07:55:47 I guess I was not paying any attention. Nov 12 07:55:49 which I recommended originally anyway, remember Nov 12 07:55:54 yea. Nov 12 07:56:10 but you wanted "more crap", and opted for the 4 gb one Nov 12 07:56:19 I just remembered. Nov 12 07:56:22 good Nov 12 07:56:25 You thought that. Nov 12 07:56:32 I was unaware of the difference. Nov 12 07:56:35 well, your previous model doesn't have space for more crap Nov 12 07:56:37 :) Nov 12 07:56:40 I did not know that they had options. Nov 12 07:56:50 Okay, Okay. I get it now. Nov 12 07:56:53 only rev C has 5 GB Nov 12 07:56:55 4 GB Nov 12 07:56:57 typo Nov 12 07:57:06 all previous revisions had 2 GB Nov 12 07:57:19 I thought something was up. Nov 12 07:57:49 I need to go find a 2GB card. Nov 12 07:58:00 you can use a 4 GB card, no problem Nov 12 07:58:04 I just formatted this large SD card. Now...I have to go and find a small one. Nov 12 07:58:07 Okay. Nov 12 07:58:07 larger card is not a problem Nov 12 07:58:11 Oh. Nov 12 07:58:12 Okay. Nov 12 07:58:33 it'll just leave the extra space unused on the card Nov 12 07:58:44 I am like, "Why will 4GB of info not fit into 2GB?" Nov 12 07:58:46 Ha! Nov 12 08:01:28 I bet you that I put the 2gb version on my 4gb BBB. Nov 12 08:01:58 Damn...no wonder it would not show me a screen when I tried to hook it up to the monitor. Nov 12 08:02:55 2 gb version works fine on a 4 gb BBB Nov 12 08:03:22 Oh. Nov 12 08:03:25 it just means you get less preinstalled packages Nov 12 08:03:36 I wanted to get the OS too (GUI). Just to check it out. Nov 12 08:03:39 which I see as a pro, not as a con Nov 12 08:03:41 Oh. Nov 12 08:03:45 both have a GUI Nov 12 08:03:48 Really? Nov 12 08:03:52 the 2gb just less applications Nov 12 08:04:07 however the old image may have trouble with some displays Nov 12 08:04:07 Hmmm...I am having some issues then. Nov 12 08:04:36 Like...a new mirco HDMI to DVI monitor? Nov 12 08:04:38 which kernel are you on now? Nov 12 08:04:48 I will check. Nov 12 08:05:12 I do not know. Nov 12 08:05:16 uname -r Nov 12 08:05:22 I am not on that BBB right now. Nov 12 08:05:24 I am on another one. Nov 12 08:05:31 I am on the new one. Nov 12 08:05:32 the new one yes Nov 12 08:05:37 Oh...okay. Nov 12 08:05:40 Please hold. Nov 12 08:05:48 (the 2gb version of the same image will have the same kernel) Nov 12 08:05:51 3.8.13 when you are using the 3/1/15 image Nov 12 08:05:59 uavcam: that's ancient Nov 12 08:06:04 uavcam: both the image and the kernel Nov 12 08:06:48 yeah. but the new images will have the same kernel. Nov 12 08:06:55 I really hope not Nov 12 08:06:57 if he uses the prebuild ones Nov 12 08:06:58 Yea...3.8.13 bone 79 Nov 12 08:07:01 wtf Nov 12 08:07:05 heh? Nov 12 08:07:17 I just downloaded the new image. Nov 12 08:07:48 Set_: which one? jessie (which I recommended) or "release" (which I warned was still ancient crap, even the latest release) Nov 12 08:07:48 yes. but the kernel is quite oldschool. that what matt is wtfing about Nov 12 08:08:23 Oh. Nov 12 08:08:24 the kernel is old enough that I have no idea what does or doesn't work there Nov 12 08:08:27 I am using Debian. Nov 12 08:08:39 yes, all of these are debian Nov 12 08:08:47 *which* one :P Nov 12 08:09:01 Which Debian? I think Debian 7. Nov 12 08:09:20 wheezy, officially declared obsolete by debian themselvs Nov 12 08:09:21 https://rcn-ee.com/rootfs/bb.org/testing/2015-11-08/lxqt-2gb/BBB-eMMC-flasher-debian-8.2-lxqt-2gb-armhf-2015-11-08-2gb.img.xz Nov 12 08:09:25 get this one Nov 12 08:09:33 Well...I typed uname -r and 3.8.13 came back. Nov 12 08:09:43 Okay. Nov 12 08:10:01 I just downloaded an old version again. Sheesh. Nov 12 08:10:07 that's the latest one indeed, thanks for grabbing the link Nov 12 08:10:19 yes its the debian 7 image with 3.8.13 kernel. the snapshot image uses debian8 with 4.1.whatever Nov 12 08:10:30 Hold on. I am going to the link. Oh. Nov 12 08:10:35 Okay... Nov 12 08:10:45 uavcam: 4.1.whatever-ti specifically Nov 12 08:11:07 that bit is relevant since at least 4.2.whatever-bone still had a broken tilcdc driver Nov 12 08:11:22 the fixes were in -ti but somehow hadn't managed to reach upstream yet Nov 12 08:11:58 machinekit? Nov 12 08:12:12 or the 2gb one? Nov 12 08:12:15 Set_: you received a direct download link from uavcam Nov 12 08:12:17 use it :P Nov 12 08:12:20 Okay. Nov 12 08:13:00 I am getting now. Nov 12 08:14:08 Dang...I thought you guys wanted to get me a newer Debian image. Nov 12 08:14:19 I did not understand that you wanted me to get the Jessie image. Nov 12 08:14:32 jessie is also debian Nov 12 08:14:40 Heh? Nov 12 08:14:41 jessie is the newest debian you can get. its denian8 Nov 12 08:14:48 Okay. Nov 12 08:14:53 uavcam: no it's not. it's "debian stable" Nov 12 08:14:53 I understand now. Nov 12 08:15:02 Augh! Nov 12 08:15:05 uavcam: I use stretch on all BBBs here Nov 12 08:15:10 Oh. Nov 12 08:15:21 Set_: not talking to you, go download the image and flash it :P Nov 12 08:15:25 Hi - I want to get Emacs but apt-get install emacs can't find it - any ideas? Nov 12 08:15:26 Okay. Nov 12 08:15:31 it should be compatible with both your older and newer BBB Nov 12 08:15:35 yea...do not use it. Nov 12 08:15:45 use Nano. It is so simple. Nov 12 08:15:48 you maneged to get rid of systemd? Nov 12 08:16:19 I cannot use emacs either. I learned nano instead. Nov 12 08:16:35 uavcam: ?!? why would you want to do such a thing? I actually replaced more stuff (ifupdown, dhcp) by their systemd equivalents Nov 12 08:16:51 systemd-networkd isn't really usable yet in jessie, it is in stretch Nov 12 08:17:02 considerably reduced boot time Nov 12 08:17:56 Guest80260: it should be in the repos, have you tried "apt update" first? (or "apt-get update", same thing) Nov 12 08:18:46 Guest80260: ah, there are many 'emacs' packages Nov 12 08:18:52 variants Nov 12 08:19:05 oh, nm there is an 'emacs' metapackage also Nov 12 08:19:33 ... which I seem to have installed on my laptop for some reason, even though I never used emacs in my life Nov 12 08:20:30 Set_: nano sucks really badly Nov 12 08:20:33 emacs is Vim but nano is easier. Nov 12 08:20:35 Oh. Nov 12 08:20:39 What do you use? Nov 12 08:20:40 vim Nov 12 08:20:44 Augh. Nov 12 08:20:44 vim Nov 12 08:20:46 Okay... Nov 12 08:20:49 Vim heh? Nov 12 08:20:54 I have not used that one yet. Nov 12 08:21:03 does nano even have any kind of syntax highlighting? Nov 12 08:21:17 (I never explored it beyond "HOW DO I GET OUT") Nov 12 08:21:27 I have been making all my scripts in nano and running the old school way, i.e. python XXX.py. Nov 12 08:21:31 (followed by "who the fuck installed this") Nov 12 08:21:32 Yea. Nov 12 08:21:38 Hhahaha. Nov 12 08:21:46 zmatt - I did update, but emacs still isn't there. Nov 12 08:22:00 what mirror do you use? Nov 12 08:22:31 <<<<< smoke break! Nov 12 08:22:38 I will downloading the image soon. Nov 12 08:23:11 i tried wheezy and jessie. both have metapkg for emacs Nov 12 08:23:47 ditto stretch and sid of course Nov 12 08:23:52 Okay...the image is doing the do! Nov 12 08:23:56 oops - apt-get install emacs works Nov 12 08:24:05 Smoke break and then I will see if that image was success. Nov 12 08:24:08 Set_: you know you don't have to narrate every single step you do Nov 12 08:24:14 Oh. Nov 12 08:24:16 Cool. Nov 12 08:24:45 it does not add much value for us to know that you're downloading it, or going for a smoke break, it's just spam, and you do it constantly Nov 12 08:28:30 Okay...I will not do it anymore. Nov 12 08:28:44 announcing your departure makes sense if you're in the middle of a conversation with someone. but you weren't in a conversation with anyone, we were just hoping it would this time with a bit or luck take more than a few minutes before you announce your next problem Nov 12 08:29:03 Okay. Nov 12 08:29:12 (sorry if I'm a bit harsh, but this has been going on for quite a while now and my supply of politeness just ran out) Nov 12 08:29:16 You made yourself clear. Nov 12 08:29:51 I thought I could chat and have fun. I know this is a serious BBB only chat now. Nov 12 08:29:53 I am okay with it. Nov 12 08:30:14 "Only BBB concerns," is all I will say. Nov 12 08:31:55 people socialize too, but that doesn't work well yet when I see your nickname appear and have to brace myself for the next tiresome exchange involving you having a problem, most likely caused by overlooking or ignoring advice given previously Nov 12 08:32:20 I get it man. I will chill out and read more. Nov 12 08:32:26 sorry Nov 12 08:32:32 No issues. Nov 12 08:33:39 Set_: just... try to actually read and follow advice from us. Or if you like learning the hard way feel free to try things your way, but then don't narrate them here or ask us to deal with the resulting problems Nov 12 08:35:15 Okay...got it. Nov 12 08:35:20 good :) Nov 12 12:23:56 can anybody help me regarding gstreamer porting on beaglebone XM Nov 12 12:27:48 I suggest simply asking the question you have (and be patient, it may take time before someone with knowledge in that particular area sees your question) Nov 12 12:54:42 Hello Nov 12 12:55:44 Anyone here that can help me with beagleboard black revc? Nov 12 12:56:27 hello? Nov 12 12:56:38 what is your problem with it? Nov 12 12:57:11 all GPIOs of P9 or P8 are activated once i activate only a single GPIO of either Nov 12 12:58:52 I'm having problems connceting LED matrix for that reason. Nov 12 13:00:33 any insight would be helpful Nov 12 13:01:15 how do you activate them? what is connected to the GPIOs? Nov 12 13:01:29 currently an LED Nov 12 13:02:02 I want to have +ve on one GPIO and -ve on the other and connect them with LED Nov 12 13:04:35 well, drive one low and the other high. Nov 12 13:04:46 thats the problem Nov 12 13:05:10 If i take two GPIOs both from P9 or P8 , they both are high Nov 12 13:05:54 and you muxed them properly? Nov 12 13:06:01 i have to take one from p9 and set it high , another from p8 and set it low. But by that , all the other pins of p9 also gets high and all the other of p8 also gets low Nov 12 13:06:24 what do you mean by muxing? Nov 12 13:06:25 how do you set the gpios high or low? Nov 12 13:07:02 GPIO.setup("pin name", GPIO.OUT) Nov 12 13:07:15 GPIO.output("pin name", GPIO.HIGH) Nov 12 13:08:00 what language/software/environment is this? Nov 12 13:08:16 python Nov 12 13:08:31 and what library is that? Nov 12 13:08:36 Im currently running programs on cloud9 IDE Nov 12 13:08:48 Adafruit_BBIO.GPIO Nov 12 13:10:35 If you want to take a look at the program here's the link Nov 12 13:10:35 http://pastebin.com/9FgbyN0y Nov 12 13:10:52 I've never used that library, sorry Nov 12 13:11:25 I have just started programming in it, earlier in raspberry pi i've used python , hence I'm using python it here too Nov 12 13:11:46 Well I also tried bonescript. It also gives the same problem Nov 12 13:41:38 P8.3 and P8.5 ?! those are eMMC pins! Nov 12 13:41:41 oh, he left Nov 12 13:45:14 I really can't even begin to imagine what he could possibly have been doing Nov 12 13:46:21 adafruit gpio is the pinnacle of horrible inefficiency, but should work fine other than that Nov 12 13:47:41 well, and other than the fact that you can't switch a pin from input to output-high without briefly driving it low Nov 12 13:48:30 but I'll admit it's not quite intuitive on how to do that with sysfs :P Nov 12 13:50:35 Hello... Nov 12 13:50:53 Is it #python-unregistered I am looking for? Nov 12 13:51:12 ? Nov 12 13:51:31 Someone said to get to #pyton with my antics. I was looking for #python but I only found #python-unregistered. Nov 12 13:51:54 possibly #python lives on a different ircnet? Nov 12 13:51:58 I had more python questions outside of the BBB. Nov 12 13:51:59 Oh. Nov 12 13:52:02 Okay. Nov 12 13:52:09 https://www.python.org/community/irc/ Nov 12 13:52:13 Augh. Nov 12 13:52:21 hmm, freenode Nov 12 13:52:34 freenode! Nov 12 13:52:35 so it's right here after all Nov 12 13:52:44 Right, what is right? Nov 12 13:53:16 I just tried /join #python ... it's there, what made you think it isn't? Nov 12 13:53:38 Um...I tried and it brought me to python-unregistered. Nov 12 13:54:03 zmatt: inefficient software? on emebedded systems? done by hobbyists? unpossiburu! Nov 12 13:54:07 ahh, you need to have your nick registered Nov 12 13:54:17 probably to avoid spammers Nov 12 13:54:20 Oh. Nov 12 13:54:32 "How day do dat?" Nov 12 13:54:42 How do I register? Nov 12 13:55:01 I just pasted a link, it mentions the fact that you need to register and has a link to nickname setup instructions Nov 12 13:55:10 Oh...Okay. Nov 12 13:55:32 KotH: I know right! It's baffling! Nov 12 13:56:28 KotH: I also have no idea why they wrote this as a C-extension if they're only using sysfs anyway... they could do that just as easily in python Nov 12 13:57:03 there is so much stupidity and ignorance going on in the whole electronics (and related) field these days, it hurts Nov 12 13:57:11 i whish i had the time to educate people Nov 12 13:58:02 but it's also very frustrating that a damn lot have no inclination to learn the least bit, they just want to get their shit working by the fix they think is right, no matter how wrong it is Nov 12 14:04:11 heh, nice way to start a flame war... someone suggesting to make clang-tidy "fix" (in his view) some particular code style where you can easily see coming that other people would rather see the "fix" going the other direction Nov 12 14:06:46 I'm in the second camp, except with a twist ;) http://lists.llvm.org/pipermail/cfe-dev/2015-November/046038.html Nov 12 14:07:36 * zmatt puts on a flame-retardant suit as a precaution Nov 12 14:13:28 I have to go. I hope everyone has a great day. Nov 12 14:28:17 Hi Nov 12 14:28:31 can anyone tell me how to free more gpios in beaglebone black? Nov 12 14:29:43 disable hdmi Nov 12 14:31:27 how? Nov 12 14:32:38 I've tried mount /dev/mmcblk0p1 /mnt/card Nov 12 14:32:43 mount isnt working Nov 12 14:32:50 mount: mount point /mnt/card does not exist Nov 12 14:33:27 mkdir /mnt/card Nov 12 14:38:18 Thanks. it worked Nov 12 14:38:33 Btw. can i also disable eMMC pins? Nov 12 14:38:46 yes Nov 12 14:40:20 which kernel do you have? Nov 12 14:49:29 @nerienna angstorm linux Nov 12 14:50:51 * KotH nods Nov 12 14:50:57 a perfect description of the kernel used Nov 12 14:52:49 I have debian, kernel 3.8.13 - that's the version I know Nov 12 14:53:35 that counts as prehistoric, even for debian ^^' Nov 12 14:53:44 I know Nov 12 14:53:53 but it works Nov 12 14:54:05 yeah.. the kernel was done, back then Nov 12 15:11:54 a newer version would mean troublr because we have no time to make new device tree overlays and test software Nov 12 15:36:33 so i downloaded this jessi image and have it on my bbb .. Nov 12 15:37:01 it has standard on it ttyO4 and stuff Nov 12 15:37:23 but i cant read data from it for some reason Nov 12 15:39:29 i configure it with stty /dev/ttyO4 raw and stty /dev/ttyO4 9600 Nov 12 15:39:43 -F Nov 12 15:41:17 since i have the ttyO4 at start up i assume the device overlay and everything is done for me Nov 12 15:41:31 its enabled because i see it in the list? Nov 12 15:46:01 super_guest, permissions, flow-control? Nov 12 15:46:18 ls -al /dev/ttyO* Nov 12 15:46:22 yes its in there Nov 12 15:46:28 although, raw mode should fix flow :) Nov 12 15:46:34 i got it working with the older debian version Nov 12 15:46:46 and indeed when i was on the older debian version, did the device overlay, used raw mode, i could recieve messages Nov 12 15:47:07 ill try echoing the bb-uart in the firmware thing Nov 12 15:47:23 to the cape manager Nov 12 15:48:33 egh it says i dont have a cape manger Nov 12 15:58:12 ok it worked ;( Nov 12 15:58:14 such a relief Nov 12 15:58:26 you can do that in uEnv.txt too Nov 12 15:58:31 yes i did that Nov 12 15:58:40 with the enable command thing Nov 12 16:00:02 great ;D Nov 12 16:00:51 only my linux doenst want to ssh to my bone anymore think its cause the hash or what not changed Nov 12 16:02:48 super_guest: ssh-keygen -R will remove the old hash Nov 12 16:02:53 Hashes change when you reflash the OS Nov 12 16:03:05 O: Nov 12 16:03:17 D: Nov 12 16:03:26 <(^_^<) Nov 12 16:03:36 nah i just typed .1 instead of .2 Nov 12 16:03:40 so dumb ;x Nov 12 16:03:46 thank u guyz ;p Nov 12 16:04:04 Hah Nov 12 16:04:10 That could also be it; P Nov 12 16:59:49 rcn-ee, is there a package list for the debian images anywhere? Nov 12 17:00:19 stt_michael, yeah... in the script.. ;) Nov 12 17:00:38 zmatt, what problems were we having with dtb's and overlays Nov 12 17:01:07 rcn-ee, there doesn't seem to be lib/firmware files for the uarts/peripherals in the bare console images .. is there a package for them, or can we/one create one? Nov 12 17:01:21 rcn-ee, can we link them on the downloads page?! :P Nov 12 17:01:31 [the package lists] Nov 12 17:01:57 there was a question of what's in the 4gb images vs the 2gb ones Nov 12 17:02:56 stt_michael, https://github.com/beagleboard/bb.org-overlays/blob/master/readme.md Nov 12 17:03:11 4gb vs 2gb, usually chrome gone.. Nov 12 17:36:32 argh, why is the kernel refusing to enable lcdc ... Nov 12 17:36:41 (first tests with our own panel) Nov 12 17:36:56 own hardware with lcd panel that is Nov 12 17:37:47 zmatt: You didn't bribe it with alcohol. Nov 12 18:47:32 hello, i was wondering how does the am37xx chip on the beaglebone black interface ( uses) ethernet?? Nov 12 18:48:20 the cpu on the bbb is the am3359 Nov 12 18:48:41 the ethernet PHY chip is connected via rmii, iirc. Further details are on the schematics Nov 12 18:49:48 hardware files: http://elinux.org/Beagleboard:BeagleBoneBlack#LATEST_PRODUCTION_FILES_.28C.29 Nov 12 18:51:24 i know that the beaglebone black uses the am3359 but it is a very similar processor so i am trying to figure out hoe the am37xx would use ethernet. did some research and was able to find that it connects to the ethernet via gpmc but with no explanation Nov 12 18:53:06 how* Nov 12 18:55:03 hi I have a question about chardrv driver on BBB Nov 12 18:55:32 Is here the correct place to make the question ? Nov 12 18:57:15 Hi? Nov 12 19:02:46 stt_michael: am3358 (only am3359 in some early devices but you can't use the extra functionality of the 9 anyway) Nov 12 19:02:58 nerienna: the am37xx is not similar to the am335x Nov 12 19:03:10 the am37xx is an omap3 Nov 12 19:04:23 the am335x is a descendent of the dm81xx series, which itself is omap4-generation (though still quite different from an omap4) Nov 12 19:13:33 I have 335x Nov 12 19:14:09 3357 with the higher temperature range Nov 12 19:14:34 the am335x has an integrated gigabit ethernet switch subsystem with two external mii/rmii/rgmii ports Nov 12 19:15:10 the omap3 (including am/dm37xx) has no ethernet at all Nov 12 19:16:12 do you want to talk to me? I have no problems with ethernet, just with gpio0 Nov 12 19:16:22 I already am talking to you Nov 12 19:16:43 I think, I have some fault in the initialization of the registers Nov 12 19:17:13 ah wait I see I now confused you with another nick Nov 12 19:17:22 who already left Nov 12 19:17:24 sorry Nov 12 19:17:27 I'm jus wondering if you are talking of someone elses problem Nov 12 19:17:39 ^^ Nov 12 19:18:18 yeah "nickki_" .. which isn't actually all that similar to your nick so I really wasn't paying attention Nov 12 19:18:56 as I mentioned - what is the difference between gpio0 and the other gpio banks? Nov 12 19:19:42 I can do anything I like with gpio 1 to 3 with the registers - gpio0 doesn't work Nov 12 19:19:47 gpio0 is one of the peripherals that have been moved to a separate interconnect on the am335x, the L4WK Nov 12 19:20:21 how do IO get it to work? Nov 12 19:20:28 so it can be used to wake up the device from deep sleep modes in which the main power domain is shut off (but the L4WK is always-on) Nov 12 19:20:33 using c and the registers Nov 12 19:20:35 same as the other controllers really Nov 12 19:21:09 I did that Nov 12 19:21:19 you got the right address? Nov 12 19:21:24 yes Nov 12 19:21:33 checked the datasheet Nov 12 19:22:36 what exactly is the problem you observe? are you getting a bus error trying to reach the peripheral, or.. ? Nov 12 19:23:13 I can't set directions and high or low Nov 12 19:23:48 if I try, I can read from the mapped registers the correct values, but not from the pins Nov 12 19:24:04 you're running linux? Nov 12 19:24:37 if I map all GPIO0 as input, even sysfs in the shell does not get the GPIOs to output Nov 12 19:24:48 so I think, the address is correct Nov 12 19:24:53 yes Nov 12 19:24:57 debian Nov 12 19:25:06 3.8.13-bone70 Nov 12 19:25:33 I know, it's ancient Nov 12 19:25:50 :) Nov 12 19:26:02 I don't know if my util works on that kernel version, but you can try... Nov 12 19:26:09 http://gerbil.xs4all.nl/show-pins-v2.pl.gz Nov 12 19:26:44 slightly beaglebone-oriented though... you have custom hardware? Nov 12 19:26:44 zmatt: the am35x does, IIRC Nov 12 19:26:48 ;) Nov 12 19:26:55 ds2: the am35xx is a whole different story Nov 12 19:27:13 it looks omap3 at first sight, until you look closer Nov 12 19:27:24 then you go blind Nov 12 19:27:28 lol Nov 12 19:28:14 but yeah it has ethernet iirc, it also has a different memory controller and other changes to adapt it to broadmarket use instead of mobile Nov 12 19:28:25 the am37xx is just an omap3 though Nov 12 19:28:35 my beaglebones are modified, but they have am335x Nov 12 19:28:36 the memory controller is same as teh DM37x Nov 12 19:28:41 3357 Nov 12 19:28:50 nerienna: ok Nov 12 19:28:54 nerienna: then the util works fine Nov 12 19:29:00 if it works with your kernel version Nov 12 19:29:02 never tried it Nov 12 19:29:34 I'll try it as soon as my homing procedure works the way I want it to Nov 12 19:30:04 just send myself an email with the link Nov 12 19:30:59 ds2: no, dm37xx has SDRC and supports mSDR and mDDR (lpddr1) Nov 12 19:31:25 I'll tell you if it works Nov 12 19:31:27 ds2: am35xx has EMIF4 and supports mddr and ddr2 Nov 12 19:33:07 is the emif4 also in the omap4? Nov 12 19:33:21 i know i was able to crib code from another chip to setup the am35x Nov 12 19:35:29 slightly different flavors of emif4 are omap4, dm81xx, am335x, etc Nov 12 19:35:33 *are in Nov 12 19:38:28 hi Nov 12 19:47:56
? Nov 12 19:48:05
list Nov 12 21:28:07 hello, was wondering how the am37xx processor interface(uses) with ethernet?? Nov 12 21:41:47 zmatt .. its back. Nov 12 21:55:39 Any ideas on why my beaglebone black's rtc scratch registers are read only now that I updated from 3.19.8 to 4.2.6? Nov 12 21:56:18 It looks like there is some type of "KICK" write locking mechanism that might affect them, but it isn't clear what code change has caused them to now be read-only by default Nov 12 22:01:26 Ah, looks like a patch was added that locks & unlocks them. Blast. Nov 12 22:01:51 now I need to figure out how to have bootcount & rtc co-operate for access to that shared register space :( Nov 12 22:02:01 or I could just revert the patch :) Nov 12 22:08:40 blame systemd-timesyncd :P Nov 12 22:24:20 fishey1: what are you trying to use them for? Nov 12 22:24:40 nickii: the am37xx is an omap3 and has no ethernet itself Nov 12 22:24:52 nickii: not to be confused with the am335x which is a totally different processor btw Nov 12 22:25:53 nickii: the omap3 (including am/dm37xx) is essentially designed for smartphones... and those don't usually have an ethernet connector :P Nov 12 22:26:38 veremit: why on earth would you blame timesyncd :P Nov 12 22:27:20 zmatt.. just cos :) Nov 12 22:27:20 so why didn't they invent a micro ethernet connector with POE for smart phones then ;) Nov 12 22:27:29 GenTooMan: +1 Nov 12 22:27:48 I'm still waiting for macro-USB Nov 12 22:28:12 hey did you see USB-type C? 100W or 60W continuous LOL Nov 12 22:28:45 "so now we have to what 20V 3A .... what the heck." Nov 12 22:29:15 * zmatt finally got the lcd to work Nov 12 22:29:27 you mean with the BBB? Nov 12 22:30:23 yeah... the timings config was wrong, in the sense that some values were out of range for the lcdc... of course it gives you an excellent and informative error message about that in the kernel log.... or not :P Nov 12 22:30:33 the beagleboard-xm is open hardware and has an AM37x with Ethernet, so im trying to find out the interface between them Nov 12 22:30:49 nickii: I think some usb->ethernet thing like on the rpi, but I'm not sure Nov 12 22:31:14 and don't use <> to address people, it looks like you're quoting them Nov 12 22:32:27 zmatt: kernel 4.2? Nov 12 22:33:36 paranic: hm? Nov 12 22:33:55 you made BBView work on kernel 4.2 ? Nov 12 22:34:50 what's bbview? and no since there's no 4.2-ti kernel yet afaik Nov 12 22:35:14 thanks...i found this article that says they interface via gpmc but not sure how accurate is that Nov 12 22:35:15 sigh.... Nov 12 22:35:22 ubuntu makes me want to switch to Windows Nov 12 22:35:31 nickii: also possible yes. again, don't use <> to address people, it looks like you're quoting them Nov 12 22:36:18 alright sorry Nov 12 22:36:36 zmatt: bbview is the circuito lcd display module and there is a jessie image for bbb with 4.2 i think Nov 12 22:36:37 ds2: why are you using ubuntu? Nov 12 22:36:52 the one rcn make Nov 12 22:37:00 cuz the jack asses made it the default build env Nov 12 22:37:08 ds2: that's like using windows only with a direct feed to amazon.com with everything you do. Nov 12 22:37:27 paranic: the latest -ti kernel-image rcn-ee released is still 4.1.something Nov 12 22:37:32 ds2: I comiserate. Nov 12 22:37:39 that can be fixed with a proper router Nov 12 22:37:46 zmatt: then i mean 4.1 :-) Nov 12 22:37:48 the defaults are just outright annoying Nov 12 22:38:14 paranic: there's 4.2 and 4.3 -bone kernels, but tilcdc is badly broken in the non-ti kernels, or at least was, dunno if it's fixed by now Nov 12 22:38:23 (by "fixed" I mean "less broken") Nov 12 22:38:34 doese anyone ever played with coin acceptors? Nov 12 22:39:25 i want to 'talk' with a coin acceptor with cctalk protocol but this protocol is single line half duplex and i don't know where to start to connect it directly Nov 12 22:41:03 paranic: 4.2.5-bone2 still has the broken pixel clock divider calculation Nov 12 22:41:34 pos is design to screw you as much as it can Nov 12 22:41:47 how many pixels broken ? :-) Nov 12 22:41:50 terminal color is enabled in multiple places that ignore your wishes Nov 12 22:42:15 paranic: I think you have about 50% chance a particular pixel clock rate will work Nov 12 22:43:29 Hmmm so If I include a LCD on an expansion board I need to be sure I have the right bone kernel? :D Nov 12 22:44:00 a -ti kernel Nov 12 22:44:01 GenTooMan: or maybe compile your own Nov 12 22:44:09 or apply the trivial patch yourself Nov 12 22:44:13 it depends on your LCD Nov 12 22:44:17 I don't understand why it's still not mainlined Nov 12 22:44:24 not really Nov 12 22:44:44 maybe the powers to be are pulling purse strings to keep down ARM Nov 12 22:44:45 ;) Nov 12 22:45:07 i recently bought an BBVIEW 7 inch and the farnell site had its own kernel and support files for the display to work Nov 12 22:46:21 so the I2C device descriptor can at least have the right PINS available but not guarentee the rest of the BBB will work correctly (sigh) Nov 12 22:47:00 the broken calculation requests 2 * target_pclk_freq as clock and then uses (2 * target_pclk_freq) / actual_clock as pixel clock divider Nov 12 22:47:59 since that / rounds down that means you get either 2 or 1 as pclk divider, depending on whether the actual clock is slightly higher or lower than requested Nov 12 22:48:06 1 is an invalid pclk divider Nov 12 22:48:20 whoever wrote it was a moron Nov 12 22:48:23 lolo Nov 12 22:48:45 1 is an invalid divider?! Nov 12 22:49:19 ds2: yes, lcdc requires the pixel clock divider to be >= 2 for active-matrix raster mode Nov 12 22:49:42 didn't the older dispc just offset it by 1? Nov 12 22:49:49 seem recall registers defined like that Nov 12 22:49:58 I'm talking about the actual divider, not the programmed value Nov 12 22:50:03 it's indeed programmed off-by-one iirc Nov 12 22:50:05 Ohhh Nov 12 22:50:06 ds2: Not all values are valid however it probably SHOULD assume 1 as 2 instead. However I didn't design the HW. Nov 12 22:50:21 the the layer on layer of garbage is finally catching up? Nov 12 22:50:32 GenTooMan: 1 is a valid divider for the LCDC, just not in raster mode Nov 12 22:50:58 heh.. fun stuff. Nov 12 22:51:02 but the whole act of performing a division there is inane Nov 12 22:51:11 since the result is already known in advance: approximately 2 Nov 12 22:51:26 zmatt Dear Sire madam or whatever the case may be "you are inane!" Nov 12 22:52:24 * veremit slips GenTooMan an 's' Nov 12 22:52:41 inanity all over the place :-) Nov 12 22:53:35 sorry, I meant actual_clock / target_pclk_freq earlier Nov 12 22:53:44 rest of story is still accurate Nov 12 22:53:49 Well I've done some dumb things before... most times because of incomplete documentation. It's like some processors DMA stops working if you have debugging enabled for example. However it's not IN the documentation INSTEAD it's in a diagram placed somewhere, however no where in the documentation it STATES specifically that happens. Nov 12 22:53:52 (the division as written would be approximately 1 ) Nov 12 22:54:13 the fix is really just Nov 12 22:54:22 - div = lcd_clk / (crtc->mode.clock * 1000); Nov 12 22:54:26 + div = 2; Nov 12 22:55:02 although the actual patch in the -ti branch did a bit more cleanup Nov 12 22:55:15 what is still a mystery to me is... why on earth is it still not upstream Nov 12 22:55:54 approval seems to be completely .. missing .. Nov 12 22:56:13 rcn-ee has known about this situation for a while now, I would have guessed he'd poked the right people by now... then again maybe he's been too focussed on the x15 and forgot Nov 12 22:56:44 like many people who have their plate full add something new and stuff falls off the edge Nov 12 22:57:22 I wonder whether why of the kernels also include a fix for the buffer flip race condition, although I personally haven't been affected yet Nov 12 22:58:46 zmatt perhaps making cool animations is more important. You know you flip more buffers when animating things or displaying video. Nov 12 23:00:46 GenTooMan: it sure makes animations... I don't think the people affected thought it was a cool one though -> http://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/791/Screen-Shot-2015_2D00_10_2D00_12-at-10.08.30-PM.png Nov 12 23:01:56 looks like a nice graphical memory dump Nov 12 23:19:34 Hello... Nov 12 23:19:57 Does anyone know of the Electric Imp? Nov 12 23:20:00 It is real! Nov 12 23:20:40 I want to use it with my board, BBB, but the Elec. Imp is too large. It is a full SD card and not a micro SD card. Nov 12 23:20:41 Boo! Nov 12 23:20:52 Okay...before I make that one guy hate me more. Later! Nov 12 23:23:10 ... Nov 12 23:25:25 rofl Nov 12 23:26:05 "that one guy" ... how hard is it to remember my name Nov 12 23:28:08 he can download the E-I image and burn to a uSD card surely .. Nov 12 23:28:45 oh wait .. its a platform Nov 13 00:03:44 Surely their is an "adapter" LOL :D Nov 13 00:12:11 :P Nov 13 00:12:18 ooo .. Orange-Pi .. Nov 13 00:12:45 serioualy .. Im' gonna collect every kinda Pi they create :D Nov 13 00:28:03 hi Nov 13 00:30:42 ordered .. now time to ZZZ .. nite folks Nov 13 02:24:11 the amount of XXX and TODO comments in the power management firmware running on the cortex-m3 really inspires confidence... Nov 13 02:24:36 including things like "Why is this required here?" Nov 13 02:39:39 hello! Nov 13 02:40:07 im hoping someone is feeling generous with some advice. Nov 13 02:40:15 any takers? Nov 13 02:40:31 here we go.. Nov 13 02:40:47 just ask, don't ask to ask :P Nov 13 02:40:56 beaglebone black, revision C. i toasted my original one and bought a new one Nov 13 02:41:19 put in the sd card to flash with the latest version of the OS from beagleboard Nov 13 02:41:34 accidentally copied over my fstab with the old controller's UUID Nov 13 02:41:45 rebooted, and now she's stuck. Nov 13 02:42:12 when i debug the UART port, everything boots fine until Loading, please wait... systemd-fsck[210]: rootfs: clean, 85135/233856 files, 483096/933632 blocks Nov 13 02:42:20 it sits for a while Nov 13 02:42:37 and says welcome to emergency mode, control d and stuff to continue Nov 13 02:42:57 no response from my terminal client, or a usb keyboard Nov 13 02:43:18 it should respond to the serial console Nov 13 02:43:20 and when i hold down the user boot button to initiate a boot from sd card, it doesnt boot Nov 13 02:43:38 the serial console slowly prints "C" when i try to bood from the sdcard Nov 13 02:43:46 i agree zmatt, haha Nov 13 02:43:57 that means ROM doesn't consider your SD card to be bootable Nov 13 02:44:16 it tries several boot devices Nov 13 02:44:33 hm. would you reccomend trying another sdcard then? Nov 13 02:44:45 this sdcard is what ive been using for the past year... Nov 13 02:44:47 without SD button it's { eMMC, SD, uart, usb } Nov 13 02:45:13 with SD button it's { spi, SD, usb, uart } Nov 13 02:45:27 the CCCC you see is an attempt to boot via the uart Nov 13 02:45:41 zmatt: I use the third rtc scratch register to keep a "bootcount" in co-operation with u-boot. This allows me to have u-boot fallback to a different kernel/rootfs if the bootcount grows too large (due to not being reset by userspace) Nov 13 02:45:44 aah... so every other attempt has failed. gotcha Nov 13 02:45:44 if you see CCCP, you got a clone Nov 13 02:45:51 :D Nov 13 02:46:00 lukesaber: and it then cycles around Nov 13 02:46:14 fishey1: ah, right Nov 13 02:46:23 awesome, youve gotten me further than i would have gotten myself. Nov 13 02:47:00 fishey1: if the rtc driver locks the kick registers then the safest would be to implement the functionality to reset the bootcount in the rtc driver Nov 13 02:47:10 e.g. make a sysfs thingy for it or something Nov 13 02:47:26 so if i get myself a better quality sd card, do you think my success at using a flasher rom would be expectable? Nov 13 02:47:39 yep, long term I'll look at that. Right now it's a seperate driver that exposes a sysfs file. Driver is from the u-boot folks. Nov 13 02:47:58 ill be honest, what you two are discussing is way over my head. :) Nov 13 02:48:10 fishey1: yeah but you can't safely share the RTC if its driver handles the kicker Nov 13 02:48:32 zmatt: correct, which is why I disabled the kicking Nov 13 02:48:45 by reverting the patch that started using it Nov 13 02:49:00 I have to admit I don't see the benefit either Nov 13 02:49:02 still not very clean, but at least I'll be able to limp by Nov 13 02:49:37 zmatt: do you think the problem is my sdcard itself? Nov 13 02:49:47 I mean, it protects the RTC for... random writes? I wouldn't recommend doing those anyway, and the chance of hitting the RTC is very small :P Nov 13 02:49:53 lukesaber: hard to say Nov 13 02:49:55 ya, I'm not entirely sure whether having the write locking on the rtc is solving a real problem. Nov 13 02:50:03 lukesaber: is your host system linux or windows? Nov 13 02:50:06 (or mac) Nov 13 02:50:26 zmatt: i am using windows primarily, as its the only thing with an sdcard slot Nov 13 02:50:26 it might just be an imagined one, or something that comes from the land of systems with no mmu Nov 13 02:51:41 fishey1: the RTC looks distinctly "foreign" to me so it may indeed Nov 13 02:52:08 the term kick-register/kicker is also unusual, I think I've seen that one more commonly in TI's DSPs Nov 13 02:52:35 thanks for the help, ill try another sdcard and hope. take care! Nov 13 02:52:45 lukesaber: I really doubt it's the card Nov 13 02:52:48 lukesaber: but wait Nov 13 02:53:04 no wait, you can't recover it via ums since you don't have a linux host, hmm Nov 13 02:53:18 that would have been the quickest fix Nov 13 02:54:00 no wait, you can still fix it from initramfs... the main question is why isn't your console port responding Nov 13 02:54:06 are you sure hardware handshake is disabled? **** ENDING LOGGING AT Fri Nov 13 02:59:59 2015