**** BEGIN LOGGING AT Sat Mar 11 03:00:02 2017 Mar 11 15:02:41 hi all Mar 11 15:02:57 is there anyone familiarised with this error? Mar 11 15:02:59 File "analog_pins_reading.py", line 6, in potVal=ADC.read(analogPin) IOError: [Errno 11] Resource temporarily unavailable: 'Error while reading AIN port. Invalid or locked AIN file.' Mar 11 15:38:11 what is the thread model for the BBB? Mar 11 18:34:45 hy all Mar 11 18:34:45 i have a pb with my beagle bones black Mar 11 18:34:45 i create a qt app but the red color and blue is iverted Mar 11 18:34:45 only on BBB Mar 11 18:34:45 i tried on rpi all it fine Mar 11 18:34:46 you know whats happened ?and why ? Mar 11 19:09:10 Akex_: that will depend a lot on how you are running the app and on which distribution Mar 11 19:09:48 i am on ubuntu 16.04 kernel 4.4bones 21 of rcn Mar 11 19:09:57 and launch my app with Mar 11 19:10:10 xinit my app -- vt1 Mar 11 19:42:50 tbr: i think is that http://www.spinics.net/lists/dri-devel/msg116413.html Mar 11 20:37:16 Akex_: framebuffer depth :) Mar 11 20:37:27 if you're using x11 then it's a simple fix in xorg.conf Mar 11 20:37:52 i try thi Mar 11 20:37:54 +s Mar 11 20:38:04 depth 16 or 24 or 32 ? Mar 11 20:38:07 that it ? Mar 11 20:38:08 16 Mar 11 20:38:16 yes not work Mar 11 20:38:35 i try with otrer OS now Mar 11 20:38:38 standard images for the beaglebone already have an xorg.conf with the right setting Mar 11 20:38:47 i think is a kernel bug Mar 11 20:38:51 no Mar 11 20:38:55 well Mar 11 20:39:18 is not a standart image Mar 11 20:39:28 is a rober nelson image Mar 11 20:39:32 on recent kernels you can put a property in DT that ensures the kernel will reject requests for unsupported pixel formats Mar 11 20:39:44 robert nelson makes the standard images for bbb Mar 11 20:40:46 i have this image https://rcn-ee.com/rootfs/2017-03-09/microsd/bone-ubuntu-16.04.2-console-armhf-2017-03-09-2gb.img.xz Mar 11 20:41:04 you need DefaultDepth 16 in the "Screen" section of your xorg.conf Mar 11 20:41:28 hmm, weird if it's a bone image that has this problem... sounds like a bug report Mar 11 20:41:31 https://debian.beagleboard.org/images/bone-debian-8.6-iot-armhf-2016-12-09-4gb.img.xz now i try this Mar 11 20:41:56 unless the unusual way you start your application causes config files to be ignored, I don't know that much about X11 Mar 11 20:42:22 note: -iot images don't have X11 installed by default I think, although you can install it of course Mar 11 20:42:35 yep i will do that Mar 11 20:42:39 btw, if it's a single-window-fullscreen Qt app then you can also run it without X11 Mar 11 20:42:54 how can i do that ? Mar 11 20:44:12 https://www.irccloud.com/pastebin/YNe2Spnm/ Mar 11 20:44:58 DefaultDepth 24 must be 16 Mar 11 20:45:13 i take a photo with 16 Mar 11 20:45:20 2 min Mar 11 20:46:15 uh, I know 16 should work fine, we've used it Mar 11 20:47:27 http://pastebin.com/raw/VdfhMJWx <-- this is the default xorg.conf Mar 11 20:51:32 it the same Mar 11 20:51:47 http://imgur.com/8peQpg5 Mar 11 20:52:04 With your xorg.conf Mar 11 20:52:40 can the actual framebuffer depth being used? Mar 11 20:52:44 *can you check Mar 11 20:53:33 (I think xdpyinfo but I'm not sure) Mar 11 20:53:43 how can i do that please ? Mar 11 20:53:56 okay i will shearch Mar 11 20:54:26 xdpyinfo: unable to open display "". Mar 11 20:54:28 :( Mar 11 20:54:36 set environment var DISPLAY=:0 Mar 11 20:56:28 https://www.irccloud.com/pastebin/mSzgoVHI/ Mar 11 20:57:56 I'm trying to remember btw how to do Qt without X11... specifically how to set the framebuffer depth when using QT_QPA_PLATFORM=linuxfb . if you want gpu acceleration the easiest is just waiting a bit since I've been mailing about my results with robert nelson and he's working on supporting it in the default images Mar 11 20:58:32 yeah ;) i wait no pb Mar 11 20:58:37 yeah your X11 is 24-bit Mar 11 20:58:51 so either it ignored your config file, or your application made x11 change depth Mar 11 21:01:53 i don't know why Mar 11 21:04:13 maybe there's anything of interest in your Xorg.0.log ? Mar 11 21:06:03 try appending -fbbpp 16 to your xinit line Mar 11 21:07:01 ahhh Mar 11 21:07:04 yes Mar 11 21:07:17 that worked? Mar 11 21:07:21 nop Mar 11 21:07:33 but that it on xinit line Mar 11 21:07:41 i will delete that now Mar 11 21:07:54 hmmwhat? Mar 11 21:08:07 * zmatt doesn't understand Mar 11 21:10:16 i run a script Mar 11 21:10:25 for launch my ui Mar 11 21:10:41 and in this script there are the xinit line Mar 11 21:10:58 and i don't remember i add more option in this line Mar 11 21:12:08 http://imgur.com/8peQpg5 Mar 11 21:12:20 With depth 24 Mar 11 21:13:48 http://imgur.com/hl1hi6W Mar 11 21:13:55 With 16 :( Mar 11 21:14:17 lol, try -depth 16 in addition to -fbbpp 16 Mar 11 21:14:23 seriously, that's x11 being quite idiotic Mar 11 21:14:24 It for that i say you it not worked Mar 11 21:14:59 in my config file ? Mar 11 21:15:11 yes it looks like it rendered rgb565 data into an xbgr8888 framebuffer Mar 11 21:15:25 as arguments to xinit Mar 11 21:15:29 ok Mar 11 21:17:43 xinit /root/start2 -- vt1 -depth 16 -fbbpp 16 & Mar 11 21:17:57 the same Mar 11 21:18:03 of the last picture Mar 11 21:18:19 *sigh* x11.... Mar 11 21:18:26 though, this is via hdmi right? Mar 11 21:18:39 yes Mar 11 21:19:03 that's weird, it looks like the kernel should actually support xrgb8888 then in addition to xbgr8888 because the hdmi framer can swap red/blue Mar 11 21:19:21 i am sure is that yes Mar 11 21:19:45 although making x11 use rgb565 (16-bit color) is still the better solution Mar 11 21:19:46 i try the stadart image with lxde and the color is good Mar 11 21:19:54 standart* Mar 11 21:20:06 ok so it *is* a problem with xinit ignoring your xorg.conf file Mar 11 21:20:28 it's for that i want try with a stadart iot image Mar 11 21:20:35 iot doesn't include x11 Mar 11 21:20:38 afaik Mar 11 21:21:21 https://www.irccloud.com/pastebin/KJoRf4XD/ Mar 11 21:21:59 yeah that's 16-bit Mar 11 21:22:40 yep give that http://imgur.com/hl1hi6W Mar 11 21:22:52 "depths" lists 16 first, "depth of root window" is 16, default visual is 16-bit Mar 11 21:23:17 yes apparently those options force it at the wrong level Mar 11 21:23:41 the fbdev driver still allocated a 32bpp buffer Mar 11 21:24:01 and -fbbpp just told x11 to *pretend* it's 16bpp Mar 11 21:24:17 instead of making fbdev actually use 16bpp Mar 11 21:25:00 but like you said, if you use lxde in standard image the problem doesn't exist, so it is in fact a problem with how x11 is being started Mar 11 21:26:24 I just don't really know much about these ancient crufty internals of x11... I just found those options in the manpage :) Mar 11 21:26:46 reading x11 documentation always feels like I'm doing archeology Mar 11 21:27:12 ^^ Mar 11 21:34:19 i re said Mar 11 21:34:49 with robert image there are pb with my ui and lxde desktop Mar 11 21:35:31 on this https://debian.beagleboard.org/images/bone-debian-8.6-lxqt-4gb-armhf-2016-11-06-4gb.img.xz the color is good on desktop Mar 11 21:38:43 zmatt: if you remeber how run app without X HL me ;) Mar 11 21:39:13 well, I know how to do it with eglfs_kms and sgx, but it's not simple Mar 11 21:40:10 aie Mar 11 21:42:02 but like I said, with a bit of luck it will work in a near future image from rcn :) Mar 11 22:07:46 ok well Mar 11 22:07:58 it work with the last image Mar 11 22:09:37 Akex_: without X11 (but also without gpu) should work with QT_QPA_PLATFORM=linuxfb Mar 11 22:09:47 may require more environment variables to setup input correctly Mar 11 22:10:19 http://imgur.com/DRyVsK4 Mar 11 22:10:32 This is the right color Mar 11 22:10:36 yay :) Mar 11 22:10:41 https://debian.beagleboard.org/images/bone-debian-8.6-iot-armhf-2016-12-09-4gb.img.xz with this image Mar 11 22:10:57 looks like that gradient could use some dithering Mar 11 22:11:16 yes the gradient is ugly on BBB Mar 11 22:11:32 on RPI is very fine Mar 11 22:11:42 ok well zmatt what the pb ? Mar 11 22:11:46 the rcn image ? Mar 11 22:12:03 rpi has a stronger focus on graphics Mar 11 22:12:07 bbb on hardware I/O Mar 11 22:12:23 yes i want a godd mix ! :) Mar 11 22:12:30 good* Mar 11 22:13:34 to be honest 16-bpp can look fine if proper dithering is used Mar 11 22:14:20 i am on 24 Mar 11 22:14:49 what the conclusion of that pb zmatt please ? Mar 11 22:15:02 you shouldn't be, since the hardware doesn't support it Mar 11 22:15:07 if it work with the beaglebone website image Mar 11 22:15:14 and not with rcn image Mar 11 22:15:20 that's also an rcn image Mar 11 22:15:32 Oo Mar 11 22:15:42 rcn makes the debian images for beaglebone Mar 11 22:15:49 this is a kernel pb ? Mar 11 22:15:55 what's "pb" ? Mar 11 22:16:03 yes but what the diferance both 2 Mar 11 22:16:09 no idea Mar 11 22:16:12 some setting, somewhere Mar 11 22:16:14 probably Mar 11 22:16:21 the pb is inverted color Mar 11 22:16:48 on this image https://rcn-ee.com/rootfs/2017-03-09/microsd/bone-ubuntu-16.04.2-console-armhf-2017-03-09-2gb.img.xz Mar 11 22:16:52 oh I finally get that you're using "pb" as abbreviation for "problem" Mar 11 22:17:15 and normal color on this image https://debian.beagleboard.org/images/bone-debian-8.6-iot-armhf-2016-12-09-4gb.img.xz Mar 11 22:17:20 yes Mar 11 22:17:23 sorry :( Mar 11 22:17:29 pb = problem Mar 11 22:17:29 like I said, some setting, somewhere Mar 11 22:17:44 but if it's still 24bpp then that's still not good Mar 11 22:17:54 it should be 16 Mar 11 22:18:45 using more bpp just wastes memory bandwidth and lies to software about the true color depth (which also prevents correct dithering) Mar 11 22:19:55 with that Mar 11 22:19:58 xinit /root/start2 -- vt1 -depth 16 -fbbpp 16 &? Mar 11 22:20:03 xinit /root/start2 -- vt1 -depth 16 -fbbpp 16 & ? Mar 11 22:20:12 eh no you showed that didn't work Mar 11 22:20:31 but it's done right if you start x11 normally on lxde/lxqt images Mar 11 22:20:50 thanks to the default xorg.conf Mar 11 22:21:03 I have no idea why your setup seems to ignore the xorg.conf Mar 11 22:27:01 whats the most expensive part of the beaglebone black Mar 11 22:27:30 it's the 58 right? Mar 11 22:28:13 or maybe the 6 layer board Mar 11 22:28:24 ayjay_t: I think there are a few things that add up Mar 11 22:28:39 I just, it's a $55 dollar board, that's a lot Mar 11 22:29:17 Could you made a am3352 board so that cost-of-goods was < $20? Mar 11 22:29:17 it's not a $55 for them obviously :P Mar 11 22:29:27 i mean, i don't know ho much of a margin they have though right Mar 11 22:29:38 because they're mission was always sales > margin = huge user community Mar 11 22:29:53 no idea, but it will greatly depend on quantity Mar 11 22:30:22 yeah i meant at wherever it flatlines Mar 11 22:30:50 I have no idea when it does, that could easily be > 1M qty Mar 11 22:31:11 i.e. raspberry pi quantities Mar 11 22:31:18 well that's a different animal to me Mar 11 22:31:36 all of my knowledge comes from digikey/octopart pricebreak tables Mar 11 22:31:36 well that's why the rpi is so cheap Mar 11 22:31:58 i'm assuming at over 1M you need enterprise level negotiations Mar 11 22:32:30 zmatt: work fine in 16bpp with this image Mar 11 22:32:43 I guess actually yeah you have a lot on the beaglebone black, i wonder if there is a BOM anywhere Mar 11 22:32:44 it s very strange Mar 11 22:32:52 i know jason posted one on twitter for the pocketbone Mar 11 22:32:54 ayjay_t: the rpi has a SoC that's custom-produced for them :P Mar 11 22:33:12 yes the BOM is online Mar 11 22:33:17 github maybe Mar 11 22:33:20 check the hw git repo Mar 11 22:33:20 yes Mar 11 22:33:31 how are you doing zmatt? Mar 11 22:33:45 I'm okay :) the usual busy Mar 11 22:33:51 i haven't been here in a while. i switched from hexchat to irssi and i'm not as comfortable on irc Mar 11 22:33:57 i need to increase the font and maybe add some colors i think Mar 11 22:35:36 having "fun" with my (omap5) Pyra prototype: trying to figure out what's happening around (the many varieties of) reset, understanding the PMIC, trying to solve ddr3 leveling issues Mar 11 22:36:53 oh the ddr3 leveling issues sound interesting Mar 11 22:37:47 i'm trying to find the financial information for the beagleboard.org foundation haha Mar 11 22:37:53 it should be public its a non-profit but no luck yet Mar 11 22:39:04 yeah, "interesting" .. buggy hw leveling \o/ Mar 11 22:39:16 I'm going to try manual leveling instead Mar 11 22:39:50 I don't really see the benefit of hw leveling given that incremental leveling is still not supported (due to bugs) hence it won't be able to adjust the leveling adaptively anyway Mar 11 22:40:12 so you're talking about incremental leveling via setting registers? Mar 11 22:40:27 when you say hw leveling Mar 11 22:41:05 hw leveling as in, hw will automatically determine correct settings Mar 11 22:41:54 (except for those cases where it locks onto a false edge and ends up right on the transition boundary instead of at the optimal midpoint) Mar 11 22:43:41 so what do you expect the benefits to be Mar 11 22:43:49 ber? component lifetime? Mar 11 22:43:55 working ram Mar 11 22:43:59 oh its that bad Mar 11 22:44:34 zmatt do you do any social media other than irc? you can pm if you want Mar 11 22:44:36 sigh, second person in a few days who thinks ddr3 leveling means wear-leveling or something like that? Mar 11 22:44:57 i was keeping an open mind Mar 11 22:45:16 google ddr3 leveling Mar 11 22:48:51 yeah theres a lot of stuff on the bbb Mar 11 22:48:57 i wonder how much my customs gonna cost Mar 11 22:49:48 i still need to pick a power ic but i'm hesistant to go with the bbbs because of what i've heard Mar 11 22:52:42 well for the most part it usually still works. if you want alternatives, iirc the requirements for the am335x aren't all that complicated Mar 11 22:55:05 of course the osd335x exists exactly to simplify the stuff needed, at the price of being stuck with their choices... and it's not cheap Mar 11 22:57:41 no i'm not gonna do it Mar 11 22:58:09 it's like, i didn't spend years working iwth the BBB and then think "let me get something simpler" Mar 11 22:58:20 i worked with the BBB and become a better embedded engineer and now i'm customizing a board Mar 11 22:58:33 you know what i mean? the osd335x is like a solution without a problem Mar 11 22:58:40 :) Mar 11 22:58:53 it's a really really cool thing though so it works for the maker community Mar 11 22:58:58 becuase we all like cools htings Mar 11 22:59:44 unfortunately they took the beaglebone as basis, as-is, without investigating whether perhaps there are things that need improvement Mar 11 23:12:04 ayjay_t: anyway, it's pretty cool that you're designing an am335x-based board :) after that it's time for the am572x right? only 760 balls ;-) Mar 12 01:57:23 hey, anybody ever use a beagle bone black as an aggregator for iot devices **** ENDING LOGGING AT Sun Mar 12 03:00:00 2017