**** BEGIN LOGGING AT Mon Jul 18 02:59:58 2016 Jul 18 03:45:46 zmatt: well, 2016.07-rc3 works fine, but 2017.07 doesn't :( Jul 18 04:03:25 that's weird Jul 18 04:04:05 I'm also using rc3 Jul 18 04:39:33 help Jul 18 04:43:16 Anyway to pull out message history here? Jul 18 08:35:34 * plaw sits on jbdatko Jul 18 17:27:32 Hi all Jul 18 17:29:41 I tried to build kernel from https://github.com/RobertCNelson/ti-linux-kernel-dev.git. I followed the steps to build branch ti-linux-rt-4.4.y and replace the kernel on my board. Jul 18 17:30:14 It can not boot. The board was loaded with image downloaded via BBB-eMMC-flasher-ubuntu-16.04-console-armhf-2016-06-09-2gb.img Jul 18 17:31:01 what's the UART output? Jul 18 17:31:50 Starting kernel ... [ 0.001260] clocksource_probe: no matching clocksources found [ 1.661665] wkup_m3_ipc 44e11324.wkup_m3_ipc: could not get rproc handle [ 1.836304] omap_voltage_late_init: Voltage driver support not added [ 1.844118] PM: Cannot get wkup_m3_ipc handle [ 1.936813] bone_capemgr bone_capemgr: slot #0: No cape found [ 1.980806] bone_capemgr bone_capemgr: slot #1: No cape found [ 2.024805] bone_cape Jul 18 17:32:15 [ 2.068806] bone_capemgr bone_capemgr: slot #3: No cape found Jul 18 17:32:18 That is all. Jul 18 17:32:25 The system stops. Jul 18 17:33:26 I didn't change anything from the repo. I just ran the build_kernel.sh. Jul 18 17:35:29 can you pastebin that please? Jul 18 17:36:06 pastebin? Jul 18 17:36:36 Let me do it line by line. Jul 18 17:36:38 pastebin.mozilla.org gist.github.com pastebin.com - choose one Jul 18 17:38:09 https://pastebin.mozilla.org/8884981 Jul 18 17:39:54 I just found Robert has updated the branch 3 days ago. Let me try to rebuild it again. Jul 18 17:43:33 hmm, looks like the kernel manages to put out at least some messages, before it possibly stalls Jul 18 17:43:45 not familiar with that sort of problem Jul 18 17:43:47 Yes. Jul 18 17:44:20 I also tried to build a version that matches the image I have. It can not even start to build. Jul 18 17:44:46 It failed at patch.sh when it tried to backport patches. Jul 18 17:46:38 It seems the image I downloaded can not be built from the source. It's weird. Jul 18 17:47:17 The image I have is prebuilt. Jul 18 18:00:35 I set up a Debian build environment and it works fine for building the kernel I'm using. Well I think it does. I've only ever used the build tree for DTB & LKM, so maybe it doesn't. It doesn't error though. Jul 18 18:02:06 My kernel was loaded and starts to run. But, it failed at beginning. Jul 18 18:02:26 Ragnorok, can you please take a look at https://pastebin.mozilla.org/8884981 ? Jul 18 18:02:45 This is my serial port log. Jul 18 18:04:41 I did. I don't know what's up with it. Sadly I'm not a real kernel guru, I just play one at home. Jul 18 18:05:02 Got it, thanks. Jul 18 22:42:31 none of those messages look like serious errors though Jul 18 23:01:00 hey everyone! Jul 18 23:01:04 long time no type! Jul 18 23:01:15 yow Jul 18 23:01:22 * yates snickers Jul 18 23:02:00 it's lookin' like i might be getting a new BBB-based project. Jul 18 23:02:24 this one has to have a user interface via a touchscreen Jul 18 23:02:28 this time without usb crossing a split ground plane? ;) Jul 18 23:02:44 damn right! Jul 18 23:04:34 is it reasonable to just run a desktop environment (probably something light like LXDE) and then run a user app like a Qt-based GUI on top? Jul 18 23:05:01 (forced to full-screen) Jul 18 23:05:11 and sysinit'ed to run at startup. Jul 18 23:05:59 you can also run q qt-based app without X Jul 18 23:06:08 oh really?!? Jul 18 23:06:22 that would be even lighter... Jul 18 23:06:25 TI actually dropped X11 support entirely, they've gone full wayland Jul 18 23:06:46 TI? Jul 18 23:06:56 but direct on fb is also possible with the linuxfb or egl backends for Qt5 Jul 18 23:07:12 i was going to run debian. Jul 18 23:07:26 linuxfb backend definitely works on debian Jul 18 23:07:36 QT_QPA_PLATFORM=linuxfb ./your-qt-app Jul 18 23:07:41 so one way or the other, definitely doable Jul 18 23:08:27 are there any plans (or talk) of TI end-of-lifing the AM335x anytime soon? Jul 18 23:08:37 nothing of the sort Jul 18 23:08:41 good. Jul 18 23:09:23 if softbank revokes..... Jul 18 23:09:24 ;) Jul 18 23:09:53 also, do you happen to know if the absolute latest BBB hardware (rev C?) is available as an Altium project? our board guy says he only sees rev b-xyz stuff on the site. Jul 18 23:10:00 TI actually released a major new sdk a week ago Jul 18 23:10:31 rev C is the same schematic and pcb as the latest rev b, just a BOM change Jul 18 23:10:40 ah. Jul 18 23:10:47 yates: what are you doing that wants SGX accel? Jul 18 23:11:08 did i say i wanted SGX accel? Jul 18 23:11:20 you didn't, although it may be desirable regardless Jul 18 23:11:24 then why do you care about the SDK? Jul 18 23:11:35 who says he does? Jul 18 23:11:39 ds2: did i say i cared about the SDK? Jul 18 23:11:52 zmatt was just informing me about it Jul 18 23:11:55 I just mentioned it as the opposite of intent to abandon Jul 18 23:12:00 right. Jul 18 23:12:12 ah misunderstood Jul 18 23:12:20 been there, done that... :) Jul 18 23:12:23 thought you were concerned about move to wayland Jul 18 23:12:44 although the sdk might be convenient to see if SGX is useful for application performance Jul 18 23:13:01 older SDK should suffice Jul 18 23:13:01 what is SGX? Silicon Graphics? Jul 18 23:13:02 you can probably get it working on debian too but might be less out-of-the-box Jul 18 23:13:14 SGX530 is the GPU on the am335x Jul 18 23:13:56 i thought getting a desktop environment gui up was old hat on the BBB. Jul 18 23:14:06 sure Jul 18 23:14:09 without gpu acceleration Jul 18 23:14:38 there isn't an AM335x SGX driver in vanilla linux? Jul 18 23:15:12 or even in chocolate... Jul 18 23:15:17 there's no sgx driver in mainline linux nor any hope the current driver will get mainlined Jul 18 23:15:17 what are you doing that needs a desktop GUI ? Jul 18 23:15:57 user interface screens on a 7-inch touchscreen. so we may not need SGX if the non-accelerated driver runs decently. Jul 18 23:16:59 ds2: but like zmatt suggested, apparently i can run qt with just a linuxfb Jul 18 23:17:05 I was positively surprised though to discover the new TI sdk uses kernel 4.4, gcc 5.3, and other such not-obsolete versions of stuff Jul 18 23:17:44 u-boot 2016.05 Jul 18 23:17:48 systemd Jul 18 23:17:49 zmatt: are you talking about this? http://www.ti.com/tool/PROCESSOR-SDK-AM335X Jul 18 23:19:03 yeah sdk v3 Jul 18 23:20:15 ok, well thanks for the update. nice to talk with everyone again. Jul 18 23:21:02 haven't actually tried it yet though, just judged it by its cover Jul 18 23:21:17 aaand he's gone Jul 18 23:21:19 -.- Jul 18 23:51:06 hi guys i have a beaglebone white rev A6 and it only wants to run if i power it directly from the P9 header so im wondering if ive fried the PMIC or if its something between the barrel con and usb con Jul 19 01:25:53 * zmatt grabs the schematic for the white Jul 19 01:27:50 same as on the bbb, P9 pins 5-6 and the barrel connector are tied directly together, either both or neither should work unless e.g. you managed to physically damage the barrel connector Jul 19 01:28:33 usb is a different story, although I've never heard of the pmic being damaged, let alone only its usb input (without taking out the whole thing) Jul 19 01:29:46 one possibility though: if the board is drawing a lot of current immediately on power-up then it may trip the overcurrent limit, which by default is much lower when powered via usb than when powered via barrel/P9 Jul 19 01:30:51 if there's no power-hungry external hardware connected, a damaged IC might be the cause of such current draw Jul 19 01:56:01 sorry zmatt i didnt see that you responded and then my net burped Jul 19 02:11:31 http://pastebin.com/raw/9sAtNYig Jul 19 02:15:00 hmmm just dbl checked with a meter and when i plug in the barrel connector it gives a direct short Jul 19 02:19:37 hmmmmm to WiFi or not WiFi Jul 19 02:20:10 also doesnt explain why it wont boot on usb alone when it used to Jul 19 02:20:10 ds2: if you mean the bbg-wireless that thing has some really stupid design Jul 19 02:20:57 zmatt: how so? Jul 19 02:21:00 do elaborate :D Jul 19 02:21:48 ds2: they ripped off ethernet, but then failed to notice that all the pins needed for the wifi/bt module are available there as mux options and sacrificed a whole bunch of expansion I/O pins instead Jul 19 02:22:14 zmatt: oh that little detail...HE HE HE... but no, I wasn't talking about that Jul 19 02:22:35 zmatt: trying to decide on having WiFi or not on my board, the VKS Jul 19 02:22:44 (plus if they had used the ethernet pins they could have powered VDDHV5 at 1.8V and avoided the need for level shifters) Jul 19 02:23:20 can't help you decide that Jul 19 02:24:11 zmatt: antenna location is always suboptimal as the board shrinks :( Jul 19 02:24:52 H-H-H: well like I said, pins 5-6 on P9 are connected straight to the barrel connector, so if you can power it via the former but not the latter I'm really not sure what could be the cause other than physical damage Jul 19 02:25:00 noidea about the usb though Jul 19 02:25:02 zmatt: another issue on that BBGW design is that module is only rated for off body use Jul 19 02:25:24 and you like hugging your beaglebone? Jul 19 02:25:34 doesn't the barrel connector have a mechanical switch? Jul 19 02:25:39 no Jul 19 02:25:46 not that I'm aware of? Jul 19 02:25:48 zmatt: wearables Jul 19 02:26:10 zmatt: the part has 3 pins and the schematics suggest it has a switch... breaking that switch could cause weird stuff to happen Jul 19 02:26:18 I wouldn't actually worry about RF... the official limits are afaik extremely conservative Jul 19 02:26:22 hm, good point Jul 19 02:26:44 zmatt: that's fine if you don't ever sell it Jul 19 02:27:00 selling it requires keeping the man happy Jul 19 02:27:54 the barel connector has a switch internally that is operated as the connector is pushed in (atleast according tp the pdf im looking at) Jul 19 02:27:56 or just being oblivious, probably also a common strategy Jul 19 02:28:50 http://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwi6m825vf7NAhViDsAKHcdvAi8QFggcMAA&url=http%3A%2F%2Fbeagleboard.org%2Fstatic%2Fbeaglebone%2Flatest%2FDocs%2FHardware%2FBONE_SRM.pdf&usg=AFQjCNH3_L2FsZMxK79Am5obKdcLmncEDg&sig2=4NjdO8_-fdRjYLU6GLoSHA&bvm=bv.127178174,d.ZGg Jul 19 02:29:00 bah sorry for the stupidly long ur; Jul 19 02:29:06 url* Jul 19 02:30:20 put a meter across the pins and monitor for a short as you insert the jack. (do this w/o power) Jul 19 02:30:20 p33 fig 9 Jul 19 02:30:32 i.e. leave the power cube unplugged Jul 19 02:33:15 yep dead short hmmm wierdness Jul 19 02:44:53 i wonder if its worth sending it in lol Jul 19 02:54:19 also i just noticed you assumed i was using pins 5,6 on the p9 header but im using p1 as ground and p7 to power it Jul 19 02:57:09 which is sys5v not vdd5v Jul 19 02:57:56 and that actually worked? o.O Jul 19 02:59:09 hm I guess it would act similarly to powering via the battery terminals Jul 19 02:59:50 but vsys was never meant as a way of providing power, it is a power output **** ENDING LOGGING AT Tue Jul 19 02:59:58 2016