**** BEGIN LOGGING AT Thu Apr 05 03:00:05 2018 Apr 05 03:00:16 I have an elf file. This file needs to be compiled, I guess. Apr 05 03:00:29 On the bone. It will take forever. Apr 05 03:01:17 uhh, you mean you have source code that needs to be compiled? (an elf executable is the output) Apr 05 03:01:28 Oh. Apr 05 03:01:46 That is it. The commands told me to scp the file to the BBBlue. Apr 05 03:01:55 I did it. It is done. Apr 05 03:02:02 But...it is just an elf file. Apr 05 03:02:14 i.e. a program Apr 05 03:02:19 I guess. Apr 05 03:02:30 I have no clue. I cannot read it b/c it...you know. Apr 05 03:02:56 ? Apr 05 03:03:11 Yep. That is what I am typing right now. ? Apr 05 03:03:14 Huh? Apr 05 03:03:16 Hahhaah. Apr 05 03:03:37 I downloaded the file. Do I uncompress it or what? Apr 05 03:03:44 I am at a loss. Apr 05 03:04:43 what file? downloaded from where? what do you get when you run the command "file" with the filename as argument (on the beaglebone) ? Apr 05 03:05:01 HOld please. Apr 05 03:05:08 I will get all that info. Apr 05 03:06:02 https://github.com/mirkix/ardupilotblue where it states prebuild ardupilot and it has as link ending in "blue." Apr 05 03:06:22 I will run the file w/ file. Apr 05 03:06:27 Please hold. Apr 05 03:07:54 https://pastebin.com/HubeAFSH is the output Apr 05 03:08:11 okay yeah that's simply an executable Apr 05 03:08:46 Oh. Apr 05 03:08:55 make sure it's marked as executable ( chmod a+x arducopter-3_5_4 ) and run it with appropriate arguments ( https://github.com/mirkix/ardupilotblue#run-ardupilot ) Apr 05 03:09:41 I got you. So, I just needed to make it executable. Apr 05 03:10:02 chmod will do wonders! Apr 05 03:10:55 yeah that should be added to the "Prebuild ArduPilot" section, along with fixing the title of that section to "Prebuilt ArduPilot" Apr 05 03:11:16 I always forget. Apr 05 03:11:50 I have done the chmod cmd many times but always fail when it is not in my face. Apr 05 03:12:01 or maybe "Download prebuilt ArduPilot" Apr 05 03:12:15 I did that already. I wanted to try another way. Apr 05 03:12:32 that doc should also make it clearer that Download prebuilt, Compile, and Cross-compile are three alternatives Apr 05 03:12:39 right now it looks like three sequential steps Apr 05 03:12:52 Yep. Apr 05 03:13:21 This is my mistake. I took the challenge and did not understand. Apr 05 03:14:37 I am getting cmd not found. Apr 05 03:14:40 Odd days. Apr 05 03:15:21 (There's also a third option: build on a beaglebone but use distcc to offload the actual compilation to one or more faster machines with suitable cross-compiler installed. This works great for projects which are normally difficult to cross-compile) Apr 05 03:15:37 the current directory is not in the default search path for commands Apr 05 03:16:02 I am using...oh. Apr 05 03:16:31 Where shall I look? Apr 05 03:16:38 either copy the executable into /usr/local/sbin/ to place it in the default search path for privileged commands, or invoke it with explicit path, e.g. ./arducopter instead of just arducopter Apr 05 03:16:54 Okay. Apr 05 03:18:10 (history lesson: traditionally unix systems were usually multi-user, so searching for commands in the current directory would be a security hazard. imagine putting a command named "ls" in some public directory :) Apr 05 03:19:18 I will just use cp instead of invoking it w/ an explicit path. I tried the latter. I received permission denied and command not found. Apr 05 03:19:35 I tried it w/ sudo and w/out. Apr 05 03:19:39 uh what Apr 05 03:19:45 Yep. Apr 05 03:20:23 if you're in the directory where the file is located, and the file is executable, sudo ./filename cannot possibly give permission dnied Apr 05 03:20:28 *denied Apr 05 03:20:32 it did. Apr 05 03:20:53 then it's an error produced by the program itself Apr 05 03:20:57 I'd guess Apr 05 03:21:01 Oh. Apr 05 03:21:02 Okay. Apr 05 03:21:04 No issue. Apr 05 03:21:41 just in case, can you pastebin a log of your terminal session where you invoke the command? (also include an invocation of ls -l filename) Apr 05 03:21:52 I already compiled the software earlier anyway. Apr 05 03:22:04 zmatt: I am sorry. I just erased the file. Apr 05 03:22:16 rm -rf Apr 05 03:22:25 Sorry man. Apr 05 03:22:36 why -rf ? Apr 05 03:22:53 rm all the junk that came along w/ it. Apr 05 03:22:58 it was a single file Apr 05 03:23:01 Oh. Apr 05 03:23:06 Oops! Apr 05 03:23:23 zmatt: I can get it back and test what you want me to test if necessary. Apr 05 03:24:05 I was just assuming you wanted to get that program working Apr 05 03:24:36 I have most of it working except for spidev-2.0/2.1. Apr 05 03:24:42 For the BBBlue. Apr 05 03:24:43 what do you mean? Apr 05 03:24:52 I will show you. Please hold. Apr 05 03:27:19 https://github.com/mirkix/BBBMINI/blob/master/doc/software/software.md but I think that it has many commands that should not be followed one after another. Apr 05 03:28:47 I did it natively. It took forever and a day. Apr 05 03:29:33 and? Apr 05 03:29:51 zmatt: I cannot find what I found earlier that had me listening to the commands. Please hold a bit longer. Apr 05 03:30:00 there are no references to "spidev-2.0/2.1" on that page, what are you trying to show me? Apr 05 03:31:24 Forget it. Apr 05 03:31:29 I cannot find it again. Apr 05 03:31:35 I will keep looking if necessary. Apr 05 03:31:54 or just explain what problem you're having Apr 05 03:32:22 it also still sounds odd to me that the prebuilt version wasn't working for you Apr 05 03:33:29 Oh well. Poo Poo happens. Apr 05 03:33:42 I guess I am mistaken and maybe that fellow changed something. Apr 05 03:34:00 Who knows? I will check. Apr 05 03:35:18 oh duh I can't believe it took me so long to realize it's you Apr 05 03:35:25 why do you need changing your name? that's really annoying Apr 05 03:35:39 need=keep Apr 05 03:37:36 Some person keeps using my names. Apr 05 03:38:06 first set_, then fortnight, and I got fed up. Apr 05 03:38:13 I just started typing random names. Apr 05 03:38:21 why don't you register your nickname with nickserv? Apr 05 03:38:30 Oh. I thought I did. Apr 05 03:38:33 set_ Apr 05 03:38:40 but they keep stealing it. Apr 05 03:38:59 you can't steal a nickname when it's registered with nickserv, that's what nickserv is for Apr 05 03:39:34 also, 05:38 -NickServ- set_ is not registered. Apr 05 03:39:55 Hmm. Apr 05 03:40:09 Oh. I will do that then. Apr 05 03:40:27 I stole fortnight. It was used by someone else. Apr 05 03:41:01 there is someone using the nickname right now, but he's been connected for less than a day so most likely that person will disconnect again sooner or later Apr 05 03:41:10 oh wait Apr 05 03:41:14 What? Apr 05 03:41:23 unless it was because of freenode itself Apr 05 03:41:30 I just realized I got disconnected earlier too Apr 05 03:41:34 Maybe. Apr 05 03:41:47 Updated servers? Apr 05 03:42:24 anyway, if that person disconnects (or changes name) and hasn't registered set_ with nickserv in the mean time, you can grab it and register it Apr 05 03:42:59 Okay. Apr 05 03:43:12 I could just state that this is set_. Apr 05 03:43:19 <<<<<<<< set_ Apr 05 03:44:26 anyway, I don't actually care what nickname you use, but it would be nice if you just picked one instead of changing it all the time Apr 05 03:44:39 Okay. Apr 05 03:44:41 Done. Apr 05 03:45:17 You just want to know not to talk to me. Apr 05 03:45:18 I know. Apr 05 03:45:37 Sorry but ding dong. Apr 05 03:45:39 ... Apr 05 03:47:57 I try to answer coherent questions regardless of who asks them Apr 05 03:50:19 Okay. Fair enough. Apr 05 05:11:16 Any ideas where to begin investigating that the PWR button doesn't trigger shutdown? Apr 05 05:11:32 which board? Apr 05 05:12:28 pocket beagle. I've seen it work. Maybe i'll try with a freshly flashed card that I haven't customized Apr 05 05:13:15 you've seen it work? interesting Apr 05 05:13:34 on this board, not necessarily with this sd card and whats' on it Apr 05 05:13:37 always debian Apr 05 05:13:45 but I've modified some things that mgiht be interfering Apr 05 05:14:01 it's not a currently intermittent problem, it's consistent Apr 05 05:14:33 it's not currently an intermittent problem, to word it better Apr 05 05:14:54 since I was just browsing the schematic and noticed it lacks the rtc oscillator... but I guess maybe the rtc can be configured to run off an internal clock source instead Apr 05 05:15:27 in that case it's just a kernel driver issue, iirc the pmic driver was already bad but actually got worse at some point Apr 05 05:15:42 I think I actually did a patch in our kernels for that reason, hold on Apr 05 05:15:49 maybe I misremember Apr 05 05:15:54 no wait, never mind, that's unrelated Apr 05 05:16:16 k. my rtc is working fwiw Apr 05 05:16:38 you're using it? Apr 05 05:16:50 yes Apr 05 05:17:06 but I'm talking about a pwr button not rtc eh? Apr 05 05:17:08 oh yeah I guess it keeps the time across reboot Apr 05 05:17:22 I know, but the rtc is requires to power off the system (don't ask) Apr 05 05:17:23 yes Apr 05 05:17:31 oh interesting Apr 05 05:17:44 actually that could be connected Apr 05 05:18:03 it's quite possible i've not seen the pwr button work only when the rtc isn't configured yet Apr 05 05:18:04 so, 'sudo poweroff' works normally? Apr 05 05:18:29 I usually do sudo shutdown -h now which works Apr 05 05:18:51 I think there's something a bit different about my rtc.. Apr 05 05:19:01 never mind then Apr 05 05:19:22 if poweroff (or shutdown -h) works normally then the issue isn't related to what I just said Apr 05 05:19:28 k Apr 05 05:19:46 so either no event is being generated, or the event doesn't trigger poweroff Apr 05 05:19:52 which kernel? Apr 05 05:20:25 I was thinking what's different is that my clock is at /dev/rtc1 instead of /dev/rtc0 Apr 05 05:20:36 maybe not important Apr 05 05:20:43 oh you're talking about an external rtc Apr 05 05:20:47 yes Apr 05 05:20:48 I'm talking about the internal one Apr 05 05:21:21 I think this contains the kernel version? /boot/dtbs/4.9.82-ti-r102/ Apr 05 05:21:28 yes Apr 05 05:21:40 and I think you taught me that :) Apr 05 05:21:41 does /sys/bus/i2c/devices/0-0024/tps65217-pwrbutton exist? Apr 05 05:21:44 thanks again Apr 05 05:24:19 yes, it does Apr 05 05:26:02 check the name of its event device: echo /sys/bus/i2c/devices/0-0024/tps65217-pwrbutton/input/input*/event* Apr 05 05:26:07 just trying with a new sd for a sec Apr 05 05:26:12 unmodified distro Apr 05 05:26:42 then monitor that event device using: Apr 05 05:26:51 perl -C0 -E 'say unpack "H*", $_ while sysread STDIN, $_, 256' and press the power button to see if an event is generated Apr 05 05:29:41 lol i just tried to eject the spring-loaded sd card with one finger instead of counter squeezing the other side of the board and it super ejected the thing across the room. I had a bit of trouble finding it Apr 05 05:29:59 been there, done that Apr 05 05:30:06 really? I haven't yet Apr 05 05:30:14 ? Apr 05 05:30:32 i will try your suggestion shortly. getting there. Apr 05 05:30:33 I mean I've also managed to high-speed-eject an sd card from a beaglebone once Apr 05 05:30:57 that's what I havne't yet :) Apr 05 05:31:48 uhh, but you just said that happened? I'm really confused now Apr 05 05:32:25 until now, I have not yet had a super-eject happen Apr 05 05:32:32 only controlled ;) Apr 05 05:33:03 k trying ... Apr 05 05:34:13 what is echo /sys/bus/i2c/devices/0-0024/tps65217-pwrbutton/input/input*/event* supposed to do? I did literally that and it repeated what I echoed Apr 05 05:34:40 what? it is supposed to list the filename(s) matching that glob Apr 05 05:34:57 you have a tps65217-pwrbutton device but it has no input event device within it? Apr 05 05:35:07 no just gave me back /sys/bus/i2c/devices/0-0024/tps65217-pwrbutton/input/input*/event* Apr 05 05:35:27 i haven't heard of input event devices before now Apr 05 05:35:41 https://pastebin.com/raw/Mt73RWzD Apr 05 05:36:46 what is inside the tps65217-pwrbutton for you? Apr 05 05:36:58 in ...pwrbutton I have, driver_override modalias power subsystem uevent Apr 05 05:37:14 eh? it didn't load the driver? Apr 05 05:37:41 maybe I should take a step back and tell you the very few things I have modified Apr 05 05:37:49 in case something catches your attention. Apr 05 05:38:10 in /boot/uEnv.txt I added a line to enable the 4 uarts Apr 05 05:38:15 you said you just tested an unmodified distro? Apr 05 05:38:45 I added a ppp chat script and replaced the interfaces file with a version that brings up ppp at boot Apr 05 05:38:52 enabling the uarts requires no changes in /boot/uEnv.txt Apr 05 05:39:29 maybe I'm working from old information Apr 05 05:39:47 and I modified the default pin mode for P2_01 by recompiling the dtb and replacing it Apr 05 05:39:58 (from pwm to gpio) Apr 05 05:40:25 all those things seem to be working Apr 05 05:40:36 i'll try with a newly flashed card righ tnow Apr 05 05:40:58 wait, pocketbeagle... what exactly did you do for the uarts? since unlike the beaglebone, there shouldn't be any information out there that makes you meddle with uEnv.txt for uarts Apr 05 05:41:03 for the pocketbeagle Apr 05 05:41:28 overlays are basically not used at all on the pocketbeagle Apr 05 05:41:40 (you still can for select purposes, but) Apr 05 05:41:58 changing the default pin mode is a good example of something for which you should probably use an overlay Apr 05 05:42:13 to avoid your changes from being lost when you update the kernel Apr 05 05:42:19 I was working with a bbg and I think I was told they're basically compatible but I don't know if that's exact enough Apr 05 05:42:25 bbg == beaglebone green Apr 05 05:42:39 for pinmux purposes they aren't, not remotely, since they have totally different headers Apr 05 05:42:42 and pin numbers Apr 05 05:43:23 oh i think i edited the wrong uEnv.txt Apr 05 05:43:26 as far as overall method of configuring pins, that is the same provided you're not staring at old information for the beaglebone Apr 05 05:43:29 i was going from this btw http://beaglebone.cameon.net/home/serial-ports-uart Apr 05 05:43:44 yeah that's totally obsolete information Apr 05 05:44:24 in fact, that won't do anything whatsoever Apr 05 05:44:30 on kernel 4.x Apr 05 05:44:31 hm no my uEnv was in the only place I found it Apr 05 05:44:36 so anyway yeah, old info? Apr 05 05:44:52 interesting.. what's the new way? Apr 05 05:45:35 brb Apr 05 05:47:36 btw i just checked on the fresh card the pwr button. I was logged in and when I pushed it the lights started flickering in a different pattern then settled on only pwr light and i think usr2 (3rd) led light on solid. It logged me out. Apr 05 05:47:46 when I shutdown -h now it turns off all the usr lights Apr 05 05:48:00 and only the pwr light stays on when i do that shutdown command Apr 05 05:48:18 in other words it's not shutting down Apr 05 05:48:33 so you're sure you've seen shutdown working? Apr 05 05:48:54 I wonder what it is doing, it kind of seems like on this fresh card it is Apr 05 05:49:00 except for that one solid light Apr 05 05:49:09 it's halted probably Apr 05 05:49:11 or is that more of a halt or something less power donw Apr 05 05:49:13 k Apr 05 05:49:17 or actually Apr 05 05:49:24 on the beaglebone green all the lights would go out Apr 05 05:49:31 brb Apr 05 05:49:44 it might also be waiting for the shutdown it scheduled via the rtc Apr 05 05:49:51 depends a bit on what they did in DT Apr 05 05:50:02 lemme check Apr 05 05:53:57 so in this fresh card fwiw i don't have my external rtc configured or any non-default DT stuff Apr 05 05:54:11 and the light is still on from the last shutdown so it's hung if it's waiting Apr 05 06:00:31 oops sorry back Apr 05 06:00:50 found the bug I think Apr 05 06:00:53 ooo Apr 05 06:02:56 in dtb-rebuilder, try applying this patch and recompiling the dtb: https://pastebin.com/gw6CtX5c Apr 05 06:02:57 hey, second-last post on this thread seems to suggest that editing /boot/uEnv.txt as I have is still relevant https://groups.google.com/forum/#!msg/beaglebone/TIHllcAeluc/YYOu-7h3CAAJ Apr 05 06:03:29 ok Apr 05 06:03:33 no, he said he did those two things and then it worked Apr 05 06:03:45 that's because the second thing made it work Apr 05 06:04:53 will try patch... Apr 05 06:05:07 that guy's using config-pin but that's not going to help at boot Apr 05 06:05:16 unless I guess he puts them in a service but then that's too late eh? Apr 05 06:05:24 why is that too late? Apr 05 06:05:49 i guess i meant it would be much later in time, which could be too late for ones application Apr 05 06:06:04 might be ok for me, not sure Apr 05 06:06:15 uhh, then you haven't arranged things right Apr 05 06:06:28 obviously you can have a service be ordered before whatever launches your application Apr 05 06:06:45 it's true Apr 05 06:06:46 you can of course also change the default pinmux to really have it done as early as possible Apr 05 06:06:55 yeah Apr 05 06:07:02 either by modifying the base dts or using a tiny overlay Apr 05 06:07:09 just so I"m super confident, how do you know the uEnv.txt is old news? Apr 05 06:07:10 I'm checking right now what the most elegant way is Apr 05 06:07:38 is the patch command just "patch foo.patch targetfile ? Apr 05 06:07:39 uEnv.txt itself isn't old news, just not needed as often anymore. capemgr however *is* old news Apr 05 06:07:49 k Apr 05 06:07:56 yeah that's what i meant, adding the line in it Apr 05 06:08:16 I think so yes, or just patch it manually Apr 05 06:08:21 k Apr 05 06:08:57 actually, on the beaglebone capemgr is old news Apr 05 06:09:01 on the pocketbeagle it has never been Apr 05 06:09:20 it doesn't exist at all on the beaglebone and never has Apr 05 06:09:27 capes, hence capemgr, are purely a beaglebone thing Apr 05 06:10:14 also, I just realized that you can also just remove the two lines instead of patching them Apr 05 06:10:39 since the int-clk declaration is already in the defaults for the SoC Apr 05 06:15:02 ohh Apr 05 06:15:05 makes sense Apr 05 06:16:18 trying it Apr 05 06:19:32 and ext-clk doesn't pertain to my external 3rd party rtc? Apr 05 06:19:40 not related at all Apr 05 06:19:55 how are you enabling the rtc btw? Apr 05 06:22:38 i'm enabling the rtc via a service that calls a .sh script whose main line is "echo ds3231 0x68 >/sys/bus/i2c/devices/i2c-1/new_device" Apr 05 06:22:51 then hwclock -s -f /dev/rtc1 Apr 05 06:24:47 zmatt so after rebooting with the new dtb the pwr button doesn't appear to have any effect Apr 05 06:25:15 I am still and will continue to use the fresh sd card Apr 05 06:25:23 this wasn't for fixint that Apr 05 06:25:29 check if poweroff works now Apr 05 06:25:46 shutdown -h now or poweroff Apr 05 06:25:46 (or shutdown -h now, which is technically wrong but works nonetheless) Apr 05 06:25:53 :) ok Apr 05 06:26:16 shutdown works as before, all lights except pwr button go off Apr 05 06:26:21 shutdown command* Apr 05 06:28:03 hm Apr 05 06:30:10 hmmm, the kernel driver doesn't clear the flag if ext-clk is missing, so if u-boot sets the flag then the problem would still exist Apr 05 06:30:29 hold on Apr 05 06:30:59 are you looking at the kernel source code for how it's interpreting the dtb? Apr 05 06:31:05 yes Apr 05 06:31:11 drivers/rtc/rtc-omap.c Apr 05 06:31:14 bad-ass Apr 05 06:32:09 i am actually curious to look at that too – what is the repository url? Apr 05 06:32:58 https://github.com/RobertCNelson/linux-stable-rcn-ee/blob/4.9.82-ti-r102/drivers/rtc/rtc-omap.c Apr 05 06:34:49 I've never been here before, I can see torvalds (Linux' namesake) committed recently but who's RobertCNelson and why does the kernel get to be in his repo? or is his a fork for BB specifically? Apr 05 06:37:23 rcn maintains the kernels and debian images for beagleboard.org Apr 05 06:37:48 cool Apr 05 06:41:09 do you mean the has_ext_clk flag? Apr 05 06:43:53 so your patch solution was to remove the ext_clk Apr 05 06:44:20 yes, since the pocketbeagle doesn't have an external 32768 Hz oscillator for the rtc Apr 05 06:44:22 and that is an on-board rtc that you think/know the pwr btn shutdown relies on Apr 05 06:44:30 ok Apr 05 06:44:32 poweroff relies on it Apr 05 06:44:35 not the power button Apr 05 06:44:39 that's a separate issue Apr 05 06:44:42 power button triggers poweroff ? Apr 05 06:45:02 it's supposed to, except for some reason the power button driver didn't load for you Apr 05 06:45:14 hence no power button event Apr 05 06:45:18 hence no poweroff triggered Apr 05 06:45:37 can't remember, maybe that was on the customzied sd not the fresh one. .. i'll double check fresh one Apr 05 06:47:04 yeah still driver_override modalias power subsystem uevent Apr 05 06:47:37 I'm first making a little thingy to check the extrtc thing Apr 05 06:47:49 if poweroff doesn't work, making the power button work isn't helpful either :) Apr 05 06:47:51 if I had to I could connect a button to a gpio but it's not ideal at this point Apr 05 06:48:07 i see, cool Apr 05 06:49:14 I tried sudo poweroff and it behaves the same as sudo shutdown as far as logging me out and one remaining PWR button FYI Apr 05 06:49:46 shutdown -h invokes poweroff yes Apr 05 06:50:22 i see Apr 05 06:58:36 almost midnight here. Can I try your extrtc tool shortly or tomorrow? Apr 05 06:58:40 thanks btw Apr 05 07:03:14 zmatt I'll come back on after a show and ping ya in case there's anything new you wanted me to try. Cheers, Apr 05 07:10:56 jkridner[m]: if you (or anyone else) want to try if switching the rtc clock source to internal fixes the issue of the pocketbeagle refusing to poweroff: grab https://liktaanjeneus.nl/rtc-clksrc.tar.gz and run: ./rtc-clksrc --int Apr 05 07:11:00 (and then try poweroff) Apr 05 07:38:27 zmatt back Apr 05 07:41:33 windsurf_: https://liktaanjeneus.nl/rtc-clksrc.tar.gz Apr 05 07:41:47 grab that, try sudo ./rtc-clksrc --int and then poweroff Apr 05 07:45:24 cool, will do Apr 05 07:46:10 windsurf_: also, this probably helps for making the power button work: https://pastebin.com/cuecjbrR Apr 05 07:46:41 actually the status="okay" is superfluous Apr 05 07:48:07 (there's a tendency to add this out of habit, but it's actually only relevant to override a previous status = "disabled"; Hence, putting status = "okay"; in the declaration of a new node is pointless) Apr 05 07:51:06 zmatt yes, that turned all the lights off when it shut down cool Apr 05 07:52:33 cool. I've just mailed rcn about that Apr 05 07:52:48 nice work, what did you do? Apr 05 07:53:40 you mean what rtc-clksrc does? it just toggles the clock source select bit in the rtc Apr 05 07:53:50 note that it's not actually safe to do so while the rtc is enabled Apr 05 07:54:05 since this could glitch the clock and mess up the rtc state Apr 05 07:54:15 yes that's what i meant Apr 05 07:54:30 I've also found the reason why fixing the DT isn't sufficient Apr 05 07:54:59 u-boot actually enables the rtc and unconditionally sets the clock source to external Apr 05 07:55:21 without rtc-clksrc is the current behaviour of running poweroff benign? Apr 05 07:55:29 other than consuming power I assume Apr 05 07:55:48 well, presumably you have a reason for wanting to power off the system Apr 05 07:56:06 right now it's just... not Apr 05 07:56:15 btw does my second patch fix the power button issue? Apr 05 07:57:19 I have two concerns: one, being able to turn off the device without corrupting the sd card so that's why I want the proper shutdown. once I know it's safe, I have a basic physical power switch to disconnect it from the battery. So, even if poweroff gets it to the state it currently does where I think it's safe to turn off the power I think I'm ok with that Apr 05 07:57:24 I haven't tried the patch just yet Apr 05 07:57:44 ok yeah, halting suffices for that Apr 05 07:57:47 great Apr 05 07:58:10 we've got some more elaborate ideas for a managed power system for shutoff but it's not in scope atm Apr 05 07:58:51 to get proper poweroff, using rtc-clksrc as temporary workaround until rcn fixes the issue in u-boot is probably not very hazardous either, since you don't really care about the rtc state anyway. Apr 05 07:59:41 the only rtc state I care about is our 3rd party one as long as whatever might happen with the onboard one is benign Apr 05 07:59:54 exactly Apr 05 08:00:01 btw I added some pocketbeagle examples to https://github.com/mvduin/overlay-utils Apr 05 08:00:09 including one for changing the default pinmux Apr 05 08:00:33 nice Apr 05 08:00:58 this may be a desirable alternative to patching the main dts, since then you can upgrade/switch kernel versions without worry Apr 05 08:01:21 I don't know how to use this yet Apr 05 08:01:23 see the comments in uEnv.txt on how to enable overlays Apr 05 08:01:32 or, well Apr 05 08:02:04 dtb_overlay=/path/to/your/overlay.dtbo Apr 05 08:02:45 thanks a lot! Apr 05 08:03:29 I've obviously not tested these examples *at all* other than checking they compile :) (I don't have a pocketbeagle) Apr 05 08:03:49 ok Apr 05 08:10:19 fares36 /msg Apr 05 08:24:39 zmatt I get a parse error after patching Apr 05 08:25:06 this is my diff now https://pastebin.com/P08WFK64 Apr 05 08:25:38 huh, I copied that block literally from am335x-bone-common.dtsi Apr 05 08:26:19 ahh Apr 05 08:26:50 am335x-pocketbeagle-common.dtsi is missing (at the top): #include Apr 05 08:27:54 i.e. overall patch https://pastebin.com/4uWML005 Apr 05 08:28:39 gotcha. tryin' Apr 05 08:28:59 I mean, https://pastebin.com/f6xgPYRh :P Apr 05 08:31:13 zmatt I guess i'm missing that header Apr 05 08:31:19 file Apr 05 08:31:31 dt-bindings/mfd/tps65217.h Apr 05 08:31:59 I really appreciate your help but it's really late and I can't think straight anymore. mind if I see if I can find you tomorrow sometime? Apr 05 08:32:02 that's impossible Apr 05 08:32:10 it's right there => https://github.com/RobertCNelson/dtb-rebuilder/blob/4.9-ti/include/dt-bindings/mfd/tps65217.h Apr 05 08:32:24 I got fatal error: dt-bindings/mfd/tps65217.h: No such file or directory Apr 05 08:32:34 you are rebuilding the dtb using dtb-rebuilder right? Apr 05 08:32:37 branch 4.9-ti Apr 05 08:32:41 (to match your kernel) Apr 05 08:33:01 I just ran, make src/arm/am335x-pocketbeagle.dtb Apr 05 08:33:06 in.. ? Apr 05 08:33:13 the root of the cloned repo Apr 05 08:33:18 which branch? Apr 05 08:33:21 um. Apr 05 08:33:24 master i suppose Apr 05 08:33:32 oh Apr 05 08:33:44 git branch gives me 4.4-ti Apr 05 08:33:50 bingo, wrong branch Apr 05 08:33:57 ok cool Apr 05 08:34:19 anyway, good night :) Apr 05 08:34:28