**** BEGIN LOGGING AT Sun Apr 22 03:00:05 2018 Apr 22 09:27:59 Hi there Apr 22 09:29:15 I am phasing problem while trying to flash new image for beaglebone black wireless, need help Apr 22 09:37:28 noone can possibly help you if you don't explain what problem you're facing Apr 22 10:05:41 What is the difference between the ti and the bone kernel? Apr 22 10:44:48 Hi Apr 22 10:45:10 I updated my kernel to use Xenomai by executing "sudo ./update-kernel.sh --bone-xenomai-kernel" Apr 22 10:45:35 But now my BBB's LEDs stay on and I am unable to establish an SSH connection Apr 22 10:47:00 a xenomai kernel? o.O Apr 22 10:47:12 are those actually still being maintained? Apr 22 10:47:47 huh, it seems they are Apr 22 10:48:06 I like to experiment with Xenomai and the PRU Apr 22 10:48:47 still, if you're going to experiment with such an obscure kernel variant, you might want to have a serial console cable handy to debug boot issues ;P Apr 22 10:49:17 ah, ti-xenomai is maintained Apr 22 10:49:20 bone-xenomai is not Apr 22 10:49:28 you installed a 3.8 kernel Apr 22 10:50:02 Aha, good to know Apr 22 10:50:32 We installed the Bone variant because of the inclusion of the UIO drivers Apr 22 10:50:42 uio is in both -ti and -bone kernels Apr 22 10:51:09 there have been some issues with it in some -ti kernel versions, but those issues should be resolved Apr 22 10:51:30 (with uio-pruss I mean, uio itself has always worked fine) Apr 22 11:00:56 Where did you find the information concerning what variant is maintained? Apr 22 11:01:38 I just checked what versions were listed as latest for the two branches via https://rcn-ee.com/repos/latest/stretch-armhf/ Apr 22 11:01:47 (this ia also what the update-kernel.sh script uses) Apr 22 11:01:49 *is Apr 22 11:02:35 Aha, thanksss Apr 22 11:37:01 What package contains PASM? Apr 22 11:37:34 https://github.com/beagleboard/am335x_pru_package/tree/master/pru_sw/utils Apr 22 11:38:23 isn't there an official package that you can simply apt-get install? Apr 22 11:38:28 dunno Apr 22 11:38:29 as clpru is already present Apr 22 12:53:50 hi. i have a bbb rev. a5b, and i'm trying to flash its eMMC flash with openbsd6.3. i can speak to it over ssh and www (if i let angstrom boot), have it hooked up to dedicated USB power and listening on the serial debugging header. Apr 22 12:54:04 i can't input via the serial tho (i believe this is expected) Apr 22 12:54:23 uhh Apr 22 12:54:43 why would that be expected? Apr 22 12:55:13 now, when i dd the miniroom image onto the sd, and try to boot it (holding down the BOOT/USER key and resetting the board, and releasing only after the first USER LED lights up) Apr 22 12:55:48 (you can let go of the boot button once the power led turns on, no need to wait for the first user led) Apr 22 12:55:50 uhm, i won't tell you where exactly i got it off the top of my head, but i found either a blog post or a wiki entry saying this was something they encountered Apr 22 12:56:15 malloc: what type of input would you expect to do over serial? Apr 22 12:56:27 i've been mainly looking at tedu's post on this from back in 2015, when he did it with a rev c board (https://www.tedunangst.com/flak/post/OpenBSD-on-BeagleBone-Black) Apr 22 12:57:02 thinkfat: regular console, to either 1) log into the angstrom distro or 2) enter commands at the uboot prompt Apr 22 12:57:22 however, what tedu describes is not quite in line with what i'm seeing Apr 22 12:58:08 he mentions the uboot issue when he's trying to flash it, but uboot doesn't pick up the bsd parititon when i try simply booting off the SD Apr 22 12:58:13 malloc: uboot would work of course, being independent from openbsd, but for being able to login via serial, should there not be at least a getty process running? Apr 22 12:58:46 yeah, that would be the case i'd expect when angstrom boots from the eMMC flash. however, it doesn't accept input either Apr 22 12:58:47 hmm, it's possible that those early pcb revisions may be afflicted by a power sequencing issue that could lead to the io cell of the serial console rxd pin being destroyed (after a certain number of power-cycles while the rxd header pin is held high) Apr 22 12:58:55 hold on, let me reboot it and see for sure Apr 22 12:59:27 btw, i have logs, if that helps Apr 22 12:59:39 malloc: provided you have wired up your serial interface correctly, you should see u-boot output Apr 22 13:00:08 yeah, i do. i can see both uboot in action and the angstrom kernel being decompressed and launched Apr 22 13:01:19 also, when i look at the openbsd miniroot image (both when simply mounting or inspecting with guestfish), the /boot doesn't contain a uenv.txt Apr 22 13:02:05 and uboot does kinda appear to shout that the second partition is invalid, as it can't find the txt file Apr 22 13:02:37 so, this is slightly uncomfy low waters for me, and im not sure im getting everything right. Apr 22 13:02:53 does this openbsd image not include its own u-boot ? Apr 22 13:03:17 it does Apr 22 13:03:41 so, by your question, im judging that it's irrelevant whether the image contains the uenv configuration Apr 22 13:04:02 it may use some different form of configuration, I don't know Apr 22 13:04:31 the webpage you linked seems to be implying it's using a file named "uenv.txt" on a fat partition Apr 22 13:04:34 zmatt: see the article at the link i posted. it mentions similar behavior under 'internal install' Apr 22 13:04:59 rather than a /boot/uEnv.txt file on the root filesystem Apr 22 13:05:41 overall openbsd seems to use a rather different bootloading setup compared to the debian one Apr 22 13:07:46 when using a very uncommon OS on an embedded system you should expect there's a good chance to have an "interesting" time getting it to work properly :) Apr 22 13:08:09 there are occasionally other people using openbsd here, but they're few and far between Apr 22 13:08:16 kremlin: ping Apr 22 13:08:34 zmatt: pong Apr 22 13:09:11 malloc: your lucky day :) Apr 22 13:11:06 malloc: so, to summarize what's your question? (other than serial input not working, which sounds like a hw issue to me) Apr 22 13:11:57 reading Apr 22 13:12:40 here's the miniroot image listing: https://pastebin.com/iwEHmWDd Apr 22 13:12:47 let me get the uboot logs, one moment Apr 22 13:13:10 err, s/uboot/serial/ Apr 22 13:14:44 malloc: can you roughly summarize the issue you're having with openbsd? Apr 22 13:14:58 i should be able to lend a hand Apr 22 13:17:00 malloc: btw in general when interfacing BBB revisions older than A6 to external hardware, be very cautious since the vdd_3v3b (which is what's available on the P9 header) is powered up 1ms before the 3.3v io supply of the processor itself is powered up, which creates a risk of driving external signals into unpowered i/o pins, which is bad Apr 22 13:18:17 unfortunately, the isolation buffer of the serial console is also powered from vdd_3v3b, which means the processor's uart 0 rxd pin gets abused every time you power up with the rxd pin high (i.e. if you power up with a serial cable attached) Apr 22 13:18:43 kremlin: i'm following tedu's writeup from 2015 when he got the rev C to run 5.5 Apr 22 13:19:24 of the bbb? Apr 22 13:19:29 now, when i write the miniroot to the sd card and try to boot from it, it appears to not find the (right) partition that the u-boot image is on Apr 22 13:20:29 here is the serial output when booting the thing without an inserted SD card or interacting with it in any way: https://pastebin.com/bkM8jGNd Apr 22 13:20:58 here's the log when the SD card is inserted, and i hold down the BOOT button and reset the device Apr 22 13:21:01 https://pastebin.com/NE70RgV8 Apr 22 13:21:19 afterwards, if i simply reset the device, and dont hold down the boot button, it fails in exactly the same manner Apr 22 13:21:25 malloc: uhh, it's loading the same u-boot Apr 22 13:21:49 you mentioned something about angstrom earlier? or has that already been wiped? Apr 22 13:21:54 yeah, so what im assuming is that it doesn't find the uboot on the sd card and falls back to the emmc flash Apr 22 13:22:00 no, that's still on the board Apr 22 13:22:12 that's in the 2nd pastebin link i posted Apr 22 13:22:19 there's no fallback if you power-on with the boot button pressed Apr 22 13:22:39 doesn't the boot device order simply change? Apr 22 13:22:52 if you power-on with the boot button held down and no sd card inserted, it should fail to boot at all (only periodic CCC on the serial console as it attempts to perform xmodem boot) Apr 22 13:23:06 eMMC is removed entirely from the boot order Apr 22 13:23:15 hold on, let's see this happen Apr 22 13:23:27 oh Apr 22 13:23:28 oh Apr 22 13:23:30 hold BOOT3 Apr 22 13:23:35 find the button BOOT3 and hold it while you boot Apr 22 13:23:41 "boot3" ? there's no such button Apr 22 13:23:52 it's the boot button or "S2" button Apr 22 13:23:53 hold on let me grab myboard Apr 22 13:24:14 it's the button closest to the card slot, somewhat away from the power and reset buttons Apr 22 13:24:40 note that the boot button is only sampled during power-up, not if you reset/reboot Apr 22 13:24:40 S3. hold S3 while you plug it in Apr 22 13:24:50 i know :) Apr 22 13:25:01 i literally remeber the button by the way it feels on my thumb not the name Apr 22 13:25:07 uhh, S3 is the power button Apr 22 13:25:12 i have done this exact. same. manuever. 1200 times Apr 22 13:25:33 https://pastebin.com/KCxGqcBh, this is what it looks like when i try user booting w/ no sd card Apr 22 13:26:14 malloc: then either you're holding down the wrong button, not pressing it properly, or not holding it down at the right time Apr 22 13:27:38 you need to hold the S2 button while applying power to the system (you can let go once the power led turns on) Apr 22 13:27:42 hi how can i use eMMc pins as a gpio input Apr 22 13:28:04 i disabled eMMC in uEnv.txt Apr 22 13:28:30 that's all you need to do Apr 22 13:30:06 but P8_3 to P8_6 and P8_20 to P8_25 its gpio input is always showing HIGH but gpio output condition is working Apr 22 13:30:43 i tested /sys/class/gpio38 after i was tested gpio IN directions ia always showing 1 Apr 22 13:31:03 value is Apr 22 13:31:03 hold on, let me get you guys a picture Apr 22 13:31:16 reddy: install my show-pins script: https://github.com/mvduin/bbb-pin-utils/#show-pins Apr 22 13:31:35 reddy: then tell me what it says if you run sudo show-pins | grep P8.25 Apr 22 13:31:40 also, mind you, this is a A5B revision board Apr 22 13:31:52 malloc: board revision is not important for this Apr 22 13:32:07 zmatt: later ones fixed the sequencer issue, iirc Apr 22 13:32:21 just thought it'd point it out, as you mentioned it Apr 22 13:32:27 re: the serial input Apr 22 13:32:37 yes that's what I said earlier, A6 fixes that Apr 22 13:32:44 but that's not important for this Apr 22 13:34:12 note that as alternative to holding down the S2 button you can also connect a wire between P8.43 and ground (ground pins are P8.01-02, P9.01-02, P9.43-46), or just wipe eMMC Apr 22 13:35:21 now i was boot os in SD card and inserted on BBB i was not flash the OS in eMMC. Apr 22 13:36:35 reddy: ehm, obviously since you can't boot from eMMC if you want to use those pins as gpios Apr 22 13:37:31 zmatt: Apr 22 13:37:32 # show-pins | grep P8.25 Apr 22 13:37:33 P8.25 / eMMC d0 0 fast rx up 1 mmc 1 d0 11@mmc (pinmux_emmc2_pins) Apr 22 13:37:35 yes that time you told than after i did what you told Apr 22 13:37:50 malloc: I was asking reddy, not you :P Apr 22 13:37:51 after i get to access all gpio pins Apr 22 13:38:02 hah Apr 22 13:38:15 but problem is now all eMMc pins input is always showing HIGH Apr 22 13:38:20 kremlin: hold on, let me find out if s3 is really power or the user button Apr 22 13:38:27 reddy: you're repeating yourself Apr 22 13:38:56 reddy: I'm waiting for your response to what I asked you a few minutes ago: Apr 22 13:39:14 15:31 < zmatt> reddy: install my show-pins script: https://github.com/mvduin/bbb-pin-utils/#show-pins Apr 22 13:39:16 s3 is power Apr 22 13:39:23 15:31 < zmatt> reddy: then tell me what it says if you run sudo show-pins | grep P8.25 Apr 22 13:39:35 yes iwas installed above link Apr 22 13:39:38 sorry, they're right, i mean S2 Apr 22 13:39:46 malloc: s1 is reset, s2 is boot, s3 is power Apr 22 13:40:35 no please give me 1 minute i will run sir Apr 22 13:40:58 i have it enclosed in a case, so it's covering the identifiers Apr 22 13:41:13 P8.25 / eMMC d0 0 fast rx down 7 gpio 1.00 << hi P8_25 (pinmux_P8_25_default_pin) Apr 22 13:41:19 malloc: s1 and s3 are close together, s3 is some distance away Apr 22 13:41:24 eh Apr 22 13:41:26 s2 is some distance away Apr 22 13:41:43 yes Apr 22 13:41:47 but s1 and s3 are next to each other (s3 being on the side of the network port) and s2 is on the other end of the board, next to the sd card slot Apr 22 13:41:57 reddy: okay, connect a wire between P8.25 and ground, then run the command again Apr 22 13:42:00 zmatt: exactly Apr 22 13:42:08 yes it is some distance near the SD slot Apr 22 13:42:37 ok okay, connect a wire between P8.25 and ground, then run the command i will run Apr 22 13:42:41 OH CRUD! Apr 22 13:42:52 i unplugged it and then held down S3, and then added power Apr 22 13:42:57 and it booted to the openbsd installer! Apr 22 13:43:06 malloc: S3 ? S2 ? Apr 22 13:43:11 great, there i go, with my shortcuts Apr 22 13:43:15 sorry, s2 Apr 22 13:43:26 note to self: reset != unplug/plug Apr 22 13:43:26 lol, so what were you doing previously? Apr 22 13:43:30 yes that's what I said Apr 22 13:43:36 it's not sampled at reset Apr 22 13:43:38 god dang it! Apr 22 13:43:38 only at power-on Apr 22 13:43:40 :D Apr 22 13:43:51 man, and i've been at this for the second day straight Apr 22 13:43:53 :D Apr 22 13:43:57 which is quite practical, by the way Apr 22 13:44:11 in fact, it will now continue to be in force-sd-boot (even across reboots) until you power-cycle again without the s2 button pressed Apr 22 13:44:50 so be mindful of that if you install to emmc... be sure to power cycle after that, not merely reboot Apr 22 13:45:59 OS is booted by the SD card i was using the tool is WIN32 disk imager. and while inserted the SD card on BBB i was not pressed S2 button but it is working fine. Apr 22 13:46:22 reddy: why are you telling us this? Apr 22 13:46:57 sorry sir Apr 22 13:47:03 you seem to be getting confused because there are two discussions going on at the same time Apr 22 13:48:14 okay, so here's obstacle number two: there's no input over serial, so i can't do anything with the installer Apr 22 13:48:17 have you tried the command again yet with P8.25 connected to ground? Apr 22 13:49:22 malloc: yup, you'll need something you can ssh into Apr 22 13:49:26 no sir i am searched for the jumper cable please give me time once its done i will ping you Apr 22 13:50:37 kremlin: suggestions for installing without serial console access? Apr 22 13:52:02 hold on, i have a usb keyboard and a hdmi-micro to hdmi cable Apr 22 13:52:05 maybe i can do it that way Apr 22 13:53:55 @zmatt sir tested its showing the result is P8.25 / eMMC d0 0 fast rx down 7 gpio 1.00 << lo P8_25 (pinmux_P8_25_default_pin) Apr 22 13:54:14 reddy: okay, so it's working perfectly fine as input Apr 22 13:54:22 previously it was showing << hi, now it's << lo Apr 22 13:54:40 yes high its showing previously Apr 22 13:55:08 so the pins are working fine as inputs Apr 22 13:57:21 with out give ground pin to P8_25 it will not work . is there any another way i have to access Apr 22 13:58:13 I don't understand what you're asking Apr 22 13:58:22 this test shows the input is working correctly Apr 22 13:58:46 beacause of i was put the ground wire to P8_25 than after how can i use that pin as for I/O Apr 22 13:59:01 the ground wire was to test it was working as input Apr 22 13:59:08 it was initially showing high Apr 22 13:59:21 pulling it down using a ground wire resulted it in showing low Apr 22 13:59:50 therefore, the input is working correctly Apr 22 14:00:52 ok thank you thanks very much sir. Apr 22 14:09:04 zmatt: kremlin: the hdmi doesn't output a bit, and the usb keyboard isn't an option, as i can't see what im inputting Apr 22 14:09:26 i...know Apr 22 14:09:26 and the serial, just as previously, takes no input Apr 22 14:09:33 i wrote amdisplay Apr 22 14:09:38 i wrote that driver. Apr 22 14:09:47 the usb controller is not ehci compliant Apr 22 14:09:57 it has some OTG function that screws things up Apr 22 14:10:20 there's still an entire USB driver to write & im wokring on that with a local friend Apr 22 14:10:23 hold one Apr 22 14:10:26 pm ( Apr 22 14:10:28 on ** Apr 22 14:10:33 so, is the only difference that my serial doesn't take input? Apr 22 14:10:34 is there no way to make an installer image with ssh daemon? Apr 22 14:11:28 alternatively, can you install it from a host system onto a usb mass storage device? since if so, I have a solution Apr 22 14:12:13 qemu the installer and install onto the sd? Apr 22 14:12:20 here, i explained it in this mail here: https://marc.info/?l=openbsd-tech&m=149231570317514&w=2 Apr 22 14:12:24 flash some bootable linux image onto sd card, boot the beaglebone from that, and use g_mass_storage to export the eMMC as usb mass storage device Apr 22 14:12:40 look at "The USB host controller..." paragraph Apr 22 14:13:04 then the BBB's eMMC appears like a usb stick for the host system you've plugged the bbb into Apr 22 14:13:09 and you can install onto that Apr 22 14:23:56 kremlin: do you agree with zmatt? option worth giving a shot? Apr 22 14:31:27 zmatt: can you suggest a live distro for this? Apr 22 14:32:24 bbb.io/latest-images first download link Apr 22 14:32:35 or maybe the console image suffices actually Apr 22 14:33:08 yeah on second though use the console image, https://elinux.org/Beagleboard:BeagleBoneBlack_Debian#Stretch_Snapshot_console Apr 22 14:33:31 but, I'm afk now **** ENDING LOGGING AT Mon Apr 23 03:00:11 2018