**** BEGIN LOGGING AT Sun Apr 16 03:00:03 2017 Apr 16 03:21:14 Hello... Apr 16 03:49:32 0.0004884004884004884 is my read-out from my TMP36 temp. sensor. I am obviously way off. Here is the pastebin.com software: https://pastebin.com/CKDPkrTV. Apr 16 03:49:33 ... Apr 16 03:49:47 Does anyone use the TMP36 sensor with the BBB? Apr 16 06:49:13 meh python is a scourage Apr 16 06:49:18 scourge Apr 16 06:49:22 C is the only language anyone needs Apr 16 06:51:07 ^ Apr 16 07:20:16 hi there! Apr 16 07:20:47 who Apr 16 08:35:11 Set_, Change console.log(value) to console.log(f) Apr 16 17:20:48 ykchavan_: or temp_c for celsius :) although I suspect more problems unless it really is -50 ͏°C there... the readout he gave is 1 on scale 0-2047 Apr 16 19:22:56 zmatt: gm Apr 16 19:23:26 zmatt: so, ds0 transition now works, so does wakeup, but on the second attempt to enter ds0, the A8 hangs on first access to OCMCRAM Apr 16 19:24:19 zmatt: ahb access through jtag also gives an error, and I cannot access the target agent for OCMC port 0 either Apr 16 19:25:41 zmatt: OCMCRAM hwmod shows modulemode "enable" and idlest says "functional" Apr 16 19:28:04 zmatt: I compared all prcm registers between a good and no-good attempt, nothing obvious turns up. basically on the failed attempt all modules and clock domains that are expected to be on are on but that's about it Apr 16 19:28:55 zmatt: only weird thing is the MPU clock domain, Apr 16 19:29:40 zmatt: not clock domain but hwmod: according to clkctrl, module is "enabled", "functional" but "stby" Apr 16 19:30:40 hi Apr 16 19:30:50 I need some help with networking on BBB Apr 16 19:31:22 after installing new ubuntu with RT kernel, ethernet over USB doesn't work Apr 16 19:34:18 rogi: you previously had a version of ubuntu where it did work? Apr 16 19:34:26 yes Apr 16 19:34:33 hmm Apr 16 19:35:06 I know very little about the ubuntu images, almost noone seems to use them Apr 16 19:35:46 and what are the others using? debian? Apr 16 19:35:50 or angstrom? Apr 16 19:35:51 thinkfat: is the l3 interconnect even up? Apr 16 19:35:53 debian Apr 16 19:36:15 these distros should be pretty similar Apr 16 19:36:47 in some senses yes Apr 16 19:37:29 but this is a piece of non-default system configuration, which however has been specifically setup on the debian images for bbb Apr 16 19:37:54 hence I wondered if the ubuntu images even have such a thing Apr 16 19:39:06 yes, previous image had that configured and working, this one didn't and after two days of trying to make it work, I have a feeling that I have messed things not only on BBB but in my host too... Apr 16 19:39:11 where did you find a download link? Apr 16 19:39:39 well the host normally requires no configuration... linux or windows? Apr 16 19:39:45 I have installed after following instructions from eewiki https://eewiki.net/display/linuxonarm/BeagleBone+Black Apr 16 19:39:57 host = ubuntu Apr 16 19:40:02 16.04 Apr 16 19:42:18 zmatt: PER_L3_CLKCTRL == 0x2, PER_L3_CLKSTCTRL == 0x16 Apr 16 19:42:35 thinkfat: can you reach other l3 targets via debug? Apr 16 19:42:37 brb pizza Apr 16 19:42:49 zmatt: no, fails just as well Apr 16 19:42:54 zmatt: nothing on l3 seems to work Apr 16 19:50:49 hmm, rogi left Apr 16 19:51:57 thinkfat: ok, that makes more sense since I've never seen an L3 target agent being unreachable just because the module is disabled/unpowered/missing Apr 16 19:53:43 btw, you should really consider blogging about this or something, if you aren't already :) Apr 16 19:56:10 make people respect suspend/resume a bit more when it does work Apr 16 19:58:41 zmatt: could actually make a nice talk out of it for ELCE or something Apr 16 20:00:00 I'm not sure if I have any useful advice other than "check everything" .. maybe dump prcm and control module immediately after the first resume vs immediately after the second resume Apr 16 20:00:12 and then diff those Apr 16 20:00:21 zmatt: yeah, second resume never happens ;) Apr 16 20:01:06 zmatt: since second deep sleep entry hangs already Apr 16 20:01:16 ah Apr 16 20:01:46 since you mentioned a8 access I assume it was in the resume path already Apr 16 20:01:51 *assumed Apr 16 20:01:55 no, was on the suspend path Apr 16 20:02:27 after doing what exactly? Apr 16 20:03:00 from the pattern of still enabled hwmods I infer that it hangs before doing anything meaningful Apr 16 20:03:20 e.g. EMIF is still on, MPU modulemode is still 0x2 Apr 16 20:03:22 that sounds like the first resume simply didn't work Apr 16 20:04:10 hm, what aspect of it didn't work? Apr 16 20:04:21 not being able to reach the l3 interconnect? Apr 16 20:04:46 there must be a bug in my logic Apr 16 20:04:46 wait that makes no sene Apr 16 20:04:49 *sense Apr 16 20:04:58 it can't get anywhere without the l3 Apr 16 20:05:01 not even through bootrom Apr 16 20:05:36 yeah, at some point L3 must sill work, since ticker works and timer interrupts work and scheduling in the OS works Apr 16 20:05:45 EMIF works, too, obviously Apr 16 20:06:18 UART0 works, too Apr 16 20:06:25 so, it can't be in standby at that time :) Apr 16 20:06:52 (do verify debug access to l3 works in that post-resume state, just in case it's two separate issues you're looking at) Apr 16 20:18:37 zmatt: it must be something I'm doing on the way to suspend... Apr 16 20:25:59 have you ever tried getting ETM trace (to ETB) to work? having an execution trace would probably help a lot in cases like these Apr 16 20:29:55 i've got the display hardware on the beagle working on openbsd if anyone is interested: https://marc.info/?l=openbsd-tech&m=149231570317514&w=2 Apr 16 20:30:21 zmatt: no, never tried that Apr 16 20:32:30 zmatt: should not be too difficult, but I'll still poke around a bit more before I go on another quest into the unknown :) Apr 16 20:34:14 zmatt: hmm, I think I have an idea Apr 16 20:35:08 zmatt: it could be some interaction with the M3 Apr 16 20:38:03 indy: at home since last thursday Apr 16 20:38:31 indy: but boy still needs a little oxigen during the night and constant monitoring of oxigenation level Apr 16 20:38:48 indy: getting better, but slowly Apr 16 20:44:33 oops, wrong channel Apr 16 21:00:14 zmatt: there you go :-) Apr 16 21:05:46 kremlin: hey, haven't seen you in a while :) Apr 16 21:06:15 kremlin: on very rare occasion someone stumbles in here in questions regarding openbsd on the bbb Apr 16 21:23:36 zmatt: so, looks like there was an interrupt still pending on the m3 and as soon as I sent the "prepare ds0" message it took that irq and did "something" Apr 16 21:24:07 zmatt: linux kernel sends a "reset" message to M3, always wondered why, now I know Apr 16 21:25:36 zmatt: not much better though, second attempts fails now too, just differently :) Apr 16 21:34:20 zmatt: ill have to catch them Apr 16 21:34:26 ive been doing a lot of openbsd/am335x work Apr 16 21:37:20 well one of them waited a whole 7 minutes before leaving again :P Apr 16 21:47:49 thinkfat: lol Apr 16 21:47:53 thinkfat: nice and robust that m3 code Apr 16 21:51:40 thinkfat: note btw that besides ETM trace there's also STM trace, which is useful since the m3 can also send trace messages to it Apr 16 21:52:21 (and the a8, pru, edma, various hw sources like the l3 statistics collectors and "ocp watchpoint") Apr 16 21:56:32 zmatt: ah, ok Apr 16 21:56:49 zmatt: it's not the second suspend that is failing, it's the second resume that fails Apr 16 22:02:47 o.O Apr 16 22:02:52 anyway, bbl Apr 17 01:05:41 argh, asshats... Apr 17 01:06:04 "This ABI is deprecated and will be removed after 2020. It is replaced with the GPIO character device." -- https://www.kernel.org/doc/Documentation/ABI/obsolete/sysfs-gpio Apr 17 01:06:46 how on earth does that guy think his crappy character device is a useful replacement for the sysfs interface **** ENDING LOGGING AT Mon Apr 17 03:00:01 2017