**** BEGIN LOGGING AT Tue Jul 19 02:59:58 2016 Jul 19 03:00:40 well it is running as we speak although the pmic is a tad warm lol Jul 19 03:02:37 so maybe my pmic has had enough and let the majic smoke out ( i honestly cant remember what i was doing with it before and its been almost half a yr since i used it last) Jul 19 03:08:26 so zmatt any ideas ? :) is it worth sending it in for repair/diagnostics ? Jul 19 03:10:21 if the pmic were dead the board would be dead Jul 19 03:10:28 it's needed to provide all cpu supplies Jul 19 03:10:54 and I have no idea Jul 19 03:11:17 ok so tomorrow i will remove the barel connector and see if i still hav a short and work from there Jul 19 03:11:29 you said usb didn't work either though? Jul 19 03:12:31 usb works but only once its powered the way im running it for example if i now remove power from p7 on the p9 header it will still run however it doesnt turn on when you insert the usb alone Jul 19 03:13:10 have you also been powering it through sys5v on P9 before the problems started? Jul 19 03:13:47 npe i always used the barrel connector before that i remember as i said i cant be 100% sure though as its a long time since i used it last Jul 19 03:13:54 since that might actually be a way you could damage the pmic, not sure though Jul 19 03:13:56 hm Jul 19 03:14:34 have any way to check current consumption? Jul 19 03:14:47 its acting like supplying power to sysv5v is kicking the pmic into life and then it does what its sposed to Jul 19 03:15:19 the board comes up immediately or do you have to press the power button? Jul 19 03:15:25 i will check that tomorrow as its 4am here and i need my ugly sleep lol Jul 19 03:15:56 comes up straight away only if i uses p7 otherwise it wont start at all Jul 19 03:17:00 hm, that's not like the battery terminals, odd Jul 19 03:17:16 well, go sleep, tomorrow's another day :P Jul 19 03:17:33 yep ty for your help :) Jul 19 03:17:39 gnite Jul 19 04:56:53 anyone know if unused ram refreshes? Jul 19 05:09:17 ayjay: the SDRAM controller refreshes all RAM Jul 19 06:20:28 hi, anyone please let me know what is the difference between BeagleBoard-x15 Rev A2 and Rev B1? Jul 19 06:21:18 fixes Jul 19 06:30:02 @: are the fixes important? big updates? Jul 19 06:30:57 presumably they are to fix issues preventing the boards to pass "FCC" testing Jul 19 06:31:25 Nobody in their right mind would change anything 'big' at this point Jul 19 10:01:22 Question about autoloading the dtbo file for a custom cape. I created a dts file for my cape. If I force load it with echo CAPE_NAME > $SLOTS, it loads fine. Now that I have my first hardware cape ready, I burned its EEPROM (appropriately I think). And when I reboot, I see the cape in SLOT 3. But it didn't seem to load the dtbo file. Jul 19 10:02:29 And if after booting, I do the echo > $SLOTS, it works fine. So my real question is why didn't it auto-load? Jul 19 10:44:19 Hi, have a problem with my BBB, I configured the network, after I turned off the BBB and restarting I can not access via ssh, I've tried with several IP and don't get it, someone else happens this?, thanks Jul 19 10:46:23 ‮dddd Jul 19 10:46:31 acess it over the debug UART or the USB UART and check it out yourself Jul 19 12:52:45 hi Jul 19 12:53:17 Please send me the official mail id for customer services and sales Jul 19 12:53:31 anybody here? Jul 19 12:54:16 nrj: well this is not an official contact channel. What are you trying to do? Maybe someone can suggest whom to contact. Jul 19 12:55:15 I want to ask for quotations of the products Jul 19 12:55:34 Its kind of a official thing Jul 19 12:56:01 Thats why I needed the support mail ID but I dont see any Jul 19 13:15:19 tbr: can you give me official ID, please? Jul 19 13:15:58 * tbr flashes KotH ... ... his ID Jul 19 13:16:42 ugh! i didnt want to see THAT Jul 19 13:17:46 you asked for it! Jul 19 13:39:28 Hiya People Jul 19 14:22:52 ayjay: ddr3 controllers normally don't even refresh specific parts of ram, they send "refresh" commands to ram which itself keeps an internal counter to ensure all of ram gets refreshed. the only thing the ram controller needs to do is send the refresh commands sufficiently often, the interval of which is given by the ddr3 standard Jul 19 14:26:22 the only case where only part of ram is refreshed is if you enable "partial array self-refresh" and then enter the ram into self-refresh mode (suspend), which could be used to e.g. make half of ram non-volatile while in suspend and half of ram volatile. (e.g. if half of it is available to linux and half of it volatile storage needed for some processing tasks that are disabled when going into suspend) Jul 19 14:32:13 hello everyone, I got a really stupid question: how can I download a file from bbb with cloud9 to my computer? rightclick on a file in the workspace doesn't show download neither the file menue ;( Jul 19 14:37:54 WHAT IS cloud9? Jul 19 14:38:01 upsi, caplock. Jul 19 14:40:33 simonJ: Right click on the file in the file system tree view and select "Download" Jul 19 14:41:34 It took me a second to find it too. You can also upload files from your machine from the File menu. Jul 19 14:41:47 (from your machine to the BBB) Jul 19 15:26:34 MathOnMapkins: i am ok with the right click, i thought that too. but there is no download... i know this is a really silly question.. Jul 19 15:27:03 MathOnNapkins: i am ok with the right click, i thought that too. but there is no download... i know this is a really silly question.. Jul 19 15:28:46 simonJ: I don't konw what to tell you, other than I am accessing it from the "Workspace" view Jul 19 15:52:45 hi, I would like to know if all bone kernel versions have uio_pruss ? I think of 3.18.13-bone37. Jul 19 15:59:55 dirkk: the old ones do, mainline recent ones use a different structure Jul 19 16:01:36 veremit: ok thanks. Do you mean I cannot use pru functions that I coded from recent version ? Jul 19 16:01:47 dirkk: I mean kernel 4+ Jul 19 16:05:28 veremit: ok, but the procedure with c loader and assembly remains the same, right ? Jul 19 16:06:36 dirkk: yes, all 3.8.xxx should be the same Jul 19 16:07:01 veremit: thanks Jul 19 16:12:54 the idea of the numbering system, is that no major changes occur in minor versions .. :) Jul 19 16:33:26 dirkk: 4.1.19-bone20 has uio_pruss. Jul 19 16:34:31 afaik all bone versions Jul 19 16:35:20 kernel 4+ also Jul 19 16:36:01 maybe at most some sysfs paths changed slightly Jul 19 16:36:40 zmatt: I thought the pru interface changed ?? Jul 19 16:36:51 in -ti kernels yes Jul 19 16:36:55 that remoteproc crap? Jul 19 16:36:57 aha Jul 19 16:36:58 not in -bone kernels Jul 19 16:37:02 ohh okies Jul 19 16:37:06 RN ftw Jul 19 16:37:36 actually I think rcn includes both uio_pruss and remoteproc in both -bone and -ti, but the default DT nodes are different Jul 19 16:41:51 lovely Jul 19 16:44:23 python Jul 19 16:45:07 woops...wrong channel sorry. Jul 19 16:55:03 lolz Jul 19 16:57:48 That's why I'm using a bone kernel - "that remoteproc crap". (grin) Jul 19 16:59:09 I suspect its very much a case of what you're used to .. if you start from the ground up with remoteproc, you're gonna wonder wtf that 'pruss' crap is :D Jul 19 17:00:10 I started from the ground up with remoteproc. It's an atrocioius chainsaw abortion in my book. I use uio_pruss. (shrug) Jul 19 17:00:35 ah ok :P Jul 19 20:22:23 stupid question, but what do I need to do to use jtag with the 4.4 beaglebone linux kernel? Jul 19 20:22:40 from here: https://github.com/beagleboard/linux.git Jul 19 20:23:26 jtag works just fine until the kernel boots, and from there I cannot talk to the APB-AP any more Jul 19 20:24:56 My research indicates it's not possible to use JTAG with Linux. One must do bare metal, or use "machinekit", which is a JTAG-friendly OS. Jul 19 20:25:23 Ragnorok: ?? Jul 19 20:26:04 Ragnorok: you mean, you cannot use jtag to debug linux in general? Jul 19 20:26:10 Correct. Jul 19 20:26:13 oh Jul 19 20:26:30 and why? Jul 19 20:27:43 Because JTAG is a low-level hardware debugger, not an OS debugger, or something like that. Dunno really. Just know when I started BBB dev I got a JTAG thinking I'd need it, and I can't use it. (shrug) There it is. Jul 19 20:29:13 there's no need to jtag linux .. lol. Jtag isn't for that. Jul 19 20:29:33 veremit: ? Jul 19 20:29:35 vagrant and such will be sufficient to explore memory, etc. Jul 19 20:30:27 or the gnu dbugger .. Jul 19 20:31:25 erm.. lots of options for in-linux debugging ... jtag is only really useful if you need to probe individual processor registers .. and if it's working .. then there is unlikely to be a problem. However, you can do baremetal probing with jtag .. and I think zmatt has done some register debugging from linux unless I'm mistaken Jul 19 20:35:05 veremit: Ragnorok: guys, but I think you're off by at least one, if not two AU. Jul 19 20:36:01 linux really won't funciton in "single-step" mode lol Jul 19 20:36:39 what has single-stepping to do with that? Jul 19 20:37:33 JTAG is a hardware level single-step and CPU examination thing. What's that NOT got to do with it? Jul 19 20:41:34 Ragnorok: no, i mean, why would linux not work with single-stepping? Jul 19 20:43:03 The entire OS would come to halt while JTAG is active, effectively destroying all functionality as an OS. How can it single-step? Jul 19 20:43:29 Ragnorok: yes, but isn't that the point? Jul 19 20:44:00 Ragnorok: you halt the system, poke around in memory, step, set breakpoints, etc Jul 19 20:44:09 Apparently not, or JTAG would work. I'm not getting what you're not getting, but it's clear I'm not helping, so I'll stop now. Jul 19 20:46:09 thinkfat: I haven't checked schematics or anything, but how about pinmuxing? That could perhaps disconnect JTAG Jul 19 20:46:18 adj: indeed Jul 19 20:47:14 adj: but what I see it that actually concerns JTAG does work, i.e. everything IEEE 1149.1 works Jul 19 20:47:48 adj: I guess the kernel switches off the debug subsystem of the chip when it boots Jul 19 20:48:52 adj: I was hoping there would be a simple config option to keep debug clocks enabled, I fail to find them Jul 19 20:50:09 just out of curiosity, what would you use JTAG for if you get it working under linux? Jul 19 20:50:38 for low level device driver debugging JTAG would be useful also under linux, i guess Jul 19 20:50:41 adj: debugging some power management issues on the bbb Jul 19 20:52:58 Kernel Hacking ---> Enable JTAG clock for debugger connectivity Jul 19 20:53:14 how do I load omap-serial on beaglebone black? I only see ttyS* devices and not ttyO*? Jul 19 20:53:16 adj: that's what i was looking for Jul 19 20:53:24 adj: but it seems to be missing on my kernel Jul 19 20:53:51 damn, that was on kernel 3.12 Jul 19 20:55:55 Ragnorok: jtag should definitely be usable for kernel debugging, although I've never done so Jul 19 20:56:22 yeah linux does tend to disable the debug subsystem clock Jul 19 20:56:33 vkushner: should be in the kernel config options .. but yeah, some time it switched from ttyOx to ttySx .. you could try an lsmod to see what's loaded? or lspci perhaps? Jul 19 20:56:43 zmatt: trying with "clk_ignore_unused" now Jul 19 20:56:44 which normally the debugger should be able to override, but power management via ICEPick is broken in the am335x Jul 19 20:56:52 you can easily reenable it though Jul 19 20:57:48 zmatt: yes, I noticed, i can "ask" to power up the debug system, but it just pretends and doesn't actually do it Jul 19 20:57:50 zmatt: I just used kern.log for that LKM debugging. Was vastly simpler than trying to hack in JTAG. Jul 19 20:58:00 https://github.com/dutchanddutch/jbang grab that, make a file that contains #include "ti/subarctic/prcm.h" int main() { prcm.mod_dbgss.enable(); return 0; } Jul 19 20:58:04 compile and run Jul 19 20:58:08 debug clock will be enabled again Jul 19 20:58:55 zmatt: probably just going to hack the kernel to keep that hwmod enabled Jul 19 20:59:14 I think you mayeven be able to force it using DT alone Jul 19 20:59:44 zmatt: yeah, just need to find the place that disables all the hwmods, that might give a hint Jul 19 21:00:01 yeah uh somewhere in arch/arm/mach-omap2/ Jul 19 21:01:24 another kernel config option that could be useful when using jtag is CONFIG_PID_IN_CONTEXTIDR Jul 19 21:01:27 veremit: it definitely not loaded. Jul 19 21:02:09 veremit: when trying to modprobe it it says "Module omap_serial not found."... I guess thats not really the module name ;). Jul 19 21:02:16 zmatt: definitely, maybe not with openocd in particular Jul 19 21:02:30 oh openocd Jul 19 21:02:54 I think I've seen actual linux-debugging related options in CCS, but no idea in openocd Jul 19 21:03:20 zmatt: works just fine as long as you debug the kernel only Jul 19 21:03:33 zmatt: but stands no chance if you halt in userspace Jul 19 21:03:41 hehe Jul 19 21:04:00 well, it won't crash, but it won't do anything useful either Jul 19 21:04:10 also don't forget to check the debug resource manager registers related to emulation suspend signals to peripherals Jul 19 21:04:15 to ensure the watchdog halts and such Jul 19 21:04:27 yeah, that I have taken care of already Jul 19 21:04:46 vkushner: I suspect you're seein the 8250 serial driver .. hence the ttyS* - its been the default for a short while Jul 19 21:05:03 it's back to omap-serial? Jul 19 21:05:12 when did that happen? Jul 19 21:05:56 zmatt: wrong way round :P~ Jul 19 21:06:34 veremit: well since you said "its been the default for a short while" Jul 19 21:06:52 zmatt: quote veremit " the 8250 serial driver " Jul 19 21:06:58 yes Jul 19 21:07:15 if it's been the default "for a short while" that implies it's no longer the default Jul 19 21:07:15 the 8250 is not omap .. and generates ttyS* nodes : right?! Jul 19 21:07:27 yes Jul 19 21:07:29 no .. that implie it hasn't always been theefault .. haa Jul 19 21:07:41 argh keyboard fil Jul 19 21:07:54 ok, ambiguous sentence I guess Jul 19 21:07:55 fail* Jul 19 21:08:07 short while != long time :D Jul 19 21:09:22 well even if I reread it I can't parse it as anything but "it was briefly the default" Jul 19 21:09:32 hence my surprise/confusion Jul 19 21:10:15 thinkfat: it's really stupid, normally ICEPick controls should always take precedence Jul 19 21:10:38 zmatt: :P Jul 19 21:11:30 thinkfat: I have an omap5 to play with now, I'll hopefully get around to trying jtag on it too in the nearby future... I expect it should work spotlessly there Jul 19 21:11:44 since low-power debugging would have been rather important in the mobile market segment Jul 19 21:12:49 zmatt: haha Jul 19 21:13:32 ok maybe not necessarily spotlessly, but probably a lot better than on the am335x :P Jul 19 21:13:48 which, I'll grant, is not a very difficult goal to reach Jul 19 21:14:30 zmatt: where did the O5 come from? Jul 19 21:15:07 it was a trap, I got lured into the Pyra project :P Jul 19 21:15:08 zmatt: yup, you're right Jul 19 21:15:31 zmatt: there's a DT property "ti,no-idle-on-init" Jul 19 21:15:37 :) Jul 19 21:16:46 veremit: so.. how do I load omap instead? or it no longer being the default means I shouldn't use it? :) Jul 19 21:17:27 its just that all info I can find on bbb and rs485 talk about omap and gpio_pin in the rs485_serial struct, etc.. Jul 19 21:18:03 yeah rs485 is a **** Jul 19 21:18:11 vkushner: ^^ Jul 19 21:18:58 zmatt: bingo Jul 19 21:19:12 ***? : Jul 19 21:19:14 ***? :) Jul 19 21:19:48 thinkfat: possibly you could also have connected to the M3 (if it was released from reset) and use its AHP-AP to enable the debug subsystem clock Jul 19 21:20:04 vkushner: "difficult" Jul 19 21:20:24 which is irritating. I was going to do some testing myself .. but that has never happened yet Jul 19 21:20:25 zmatt: not a bad idea Jul 19 21:21:08 zmatt: but it's easier to just hack the hwmod data and set the HWMOD_INIT_NO_IDLE and HWMOD_INIT_NO_RESET flags Jul 19 21:21:31 thinkfat: if it's still held in reset however that path doesn't work... you can reach the PPB but not anything outside Jul 19 21:22:10 probably that reset signal also applies to the ahb-to-l4wkup bridge Jul 19 21:24:10 why would you hack the hwmod data when a DT patch suffices? Jul 19 21:24:47 zmatt: because with my current debug setup it's way easier to exchange the kernel binary than to exchange the device tree Jul 19 21:25:01 lol ok Jul 19 21:25:06 zmatt: ok that's not true, but I didn't yet bother to do it Jul 19 21:26:13 zmatt: the setup is basically: reset the board via SRST, halt it in rom code, set a hw breakpoint at the load address of the zImage, resume until u-boot hits the breakpoint, then upload a new zImage via JTAG and jump into it Jul 19 21:27:41 zmatt: and of course I could just as easily upload an new dtb, had I ever bothered to find out to which address Jul 19 21:28:30 So all it really takes is a DTB tweak and it works? Jul 19 21:28:42 Ragnorok: yes Jul 19 21:28:53 Interesting! Jul 19 21:29:19 Prolly just as well I never went down that road, but good to know. Thanks! Jul 19 21:30:18 Ragnorok: just whish you'll never have to. Because then, shit really hit the fan. Jul 19 21:30:34 rofl Jul 19 21:31:12 thinkfat: why would you do such a complicated and slow procedure instead of having uboot grab the kernel and DT from your computer via tftp ? Jul 19 21:31:45 zmatt: because I can ;) Jul 19 21:31:58 I mean, jtag uploads are terribly slow Jul 19 21:32:15 zmatt: roughly 700kB/s, not that slow Jul 19 21:34:13 zmatt: the main advantage right now is that I never need to leave the IDE, it's all automated Jul 19 21:34:55 zmatt: admitted, I could just as well script your suggestion. but what I currently have is good enough Jul 19 21:35:27 well the solution you already have working is always the easiest ;) Jul 19 21:36:21 I've already used tftp-based booting quite a bit (including having ROM netboot baremetal code) so that solution would be the first I'd consider Jul 19 21:37:01 zmatt: well, if all you have is a hammer, suddenly there are a lot of nails around to hit Jul 19 21:37:34 anyway, time for bed now, and thanks for the support Jul 19 21:38:32 you're welcome Jul 19 21:38:35 time to head to work Jul 19 21:52:32 how do I load /lib/firmware/BB-UART4-RS485-00A0.dtbo ? tried to echo BB-UART4-RS485-00A0, BB-UART4-RS485, and BB-UART4-RS485:00A0 into /sys/devices/platform/bone_capemgr/slots. all give me errors Jul 19 21:53:33 what errors? what appears in dmesg? (dmesg -w is a good command!) Jul 19 21:54:34 iirc its 'slots.*' also .. Jul 19 21:54:58 checking ;) Jul 19 21:56:26 oh.. "slot #11: BB-UART4-RS485 conflict P9.13 (#0:cape-CBB-Serial)" Jul 19 21:56:49 tada! Jul 19 21:57:19 yes, and no. trying to follow this cape's manual: https://docs.google.com/document/d/1y0aKYqKd3_5HuzUvIepQmC5AjbFIqBlBak-Jtv0YB7Q/edit# Jul 19 21:57:58 it says that cape-CBB-Serial is supposed to be loaded Jul 19 21:58:31 ah yes .. that one I've seen before .. Jul 19 21:59:27 so yeah, you can't load both :D Jul 19 22:00:09 can I unload one? or I need to disable it in uEnv.txt ? Jul 19 22:01:03 its likely the CBB one auto-loads via the EEPROM method Jul 19 22:01:29 so you don't need the other one, unless you want to override the former (cape_mgr.disable etc) Jul 19 22:03:34 well, according to the manual BB-UART4-RS485 is the one that enables rs485 to use RTS, and from what I understand its supposed to be loaded ontop of cape-CBB-Serial Jul 19 22:04:01 that would be unusual .. Jul 19 22:04:22 the only problem is that the manual seems to be a bit outdated. talks about slots being at /sys/devices/bone_capemgr.*/slots and I have it at /sys/devices/platform/bone_capemgr/slots Jul 19 22:05:27 and it also talks about ttyO* devices, using omap_serial Jul 19 22:06:14 yeah that is about right .. that's the only configuration which really works Jul 19 22:06:18 bat this seems to only be needed to use gpio pin instead of RTS Jul 19 22:06:40 true .. because the gpio code is in the omap driver, and not the 8250 Jul 19 22:06:51 veremit: which one? with gpio instead of RTS? Jul 19 22:07:03 but you're likely to find the timing of an RTS signal to be rather sloppy .. depending on your application Jul 19 22:07:03 I mean "configuration that works" ;) Jul 19 22:07:41 in all my dealings, I've ended up using an FTDI usb<->serial bridge, which has hardware TXEN provided Jul 19 22:07:55 veremit: well, this seems to be my current problem. I got logic analyzer and I see data going out, but I control gpio from user space, and it seems to be too slow. Jul 19 22:07:59 (either FT232 or the cheaper FT230X) Jul 19 22:08:32 there were some patches to the mainline 8250 omap implementation using gpio, but I dunno whether they ever got merged Jul 19 22:09:00 you'd be looking at kernel 4.2+ for that iirc Jul 19 22:09:17 it has always been a 'bodge' :/ Jul 19 22:09:22 from what I get on the net it was merged long time ago, into omap driver. but its not loaded ;). Jul 19 22:09:44 so I can either a) try to get omap to load or b) try to work with it w/o gpio. e.g. using RTS Jul 19 22:09:47 which is stupid really .. rs485 is an industrial control protocol .. on an industrial cpu .. running a totally non-industrial OS .. sadly. Jul 19 22:10:13 yeah the omap_serial still exists afaik .. but you wil need to compile a different kernel to pull it in Jul 19 22:10:22 clearly your default config has the later 8250 Jul 19 22:10:24 ok, I'm going to try to disable cape-CBB and see if it works ;) Jul 19 22:12:45 Hello, I need to add an overlay for an LCD cape. where can I get help with doing that? Jul 19 22:14:48 john3727: what is your display, and what do the instructions state? Jul 19 22:23:06 It is actually a projector and I need to connect it through a 24-bit RGB interface along with the other LCD lines Jul 19 22:24:01 the BBB only uses 18-bit, but this shouldn't be too much of an issue .. I've used a LVDS serializer from the BBB RGB pins no problem to a TFT display Jul 19 22:27:46 I created a .dts to enable the pins for the 24-bit interface and adjust the resolution and video parameters. But I've had problems in adding this overlay using the capemgr Jul 19 22:29:36 john3727: I left everything as-is, because the on-board hdmi framer chips uses these pins anyway, and are appropriately configured to do so .. Jul 19 22:30:00 you can force the video mode on the kernel command-lne via uEnv.txt Jul 19 22:30:36 the framer will simply ignore anything it can't cope with .. but the signals should come out ok on the pins nontheless Jul 19 22:42:41 veremit: I am also completely new to the BBB and embedded development by the way, but am I not supposed do disable the HDMI in the BBB in order to use 24-Bit mode? Jul 19 22:59:56 trying to disable a cape overlay, have this in uEnv: "optargs optargs=capemgr.disable_partno=cape-CBB-Serial", but after reboot its still loaded: "0: P---L- 0 cape-CBB-Serial,r01,Logic Supply,cape-CBB-Serial" Jul 19 23:08:42 vkushner: you need an :r1 I -think- from what the instructions seem to say Jul 19 23:09:44 I tried :r01 Jul 19 23:09:48 will try :r1 now Jul 19 23:09:49 ;) Jul 19 23:11:06 echo cape-CBB-Serial:r01 > /sys/devices/bone_capemgr.*/slots Jul 19 23:11:24 I can't remember how you unload .. but there is a way .. Jul 19 23:13:13 veremit: I'm on 4.x, so there is no /sys/devices/bone_capemgr.*/slots Jul 19 23:13:33 and I don't want to enable it, I want to disable it, or unload Jul 19 23:15:06 mhmm Jul 19 23:17:02 vkushner: # cat /sys/devices/bone_capemgr*/slots Jul 19 23:17:13 # echo -6 >/sys/devices/bone_capemgr*/slots Jul 19 23:17:17 should remove #6 Jul 19 23:17:35 without rebooting .. theoretically Jul 19 23:17:52 ha... just tried to remove slot 0, it crashed ;) Jul 19 23:18:31 so I really need to prevent it from loading in the first place Jul 19 23:18:59 yeah you should be able to do that with a cape_mgr.disable_partnos= kinda line (don't quote me on that one) Jul 19 23:19:07 in uEnv.txt Jul 19 23:20:15 veremit: this is what I'm trying to do. Jul 19 23:20:49 have this now: optargs=capemgr.disable_partno=cape-CBB-Serial:r1 Jul 19 23:21:04 tried with :r01 and w/o :r.. Jul 19 23:22:11 reading docs on capemgr now ;) Jul 19 23:22:16 overlays suck, just disable the capemgr and include stuff you want in the main dts :P Jul 19 23:23:36 shup zmatt lol Jul 19 23:23:50 its not the universal one :P Jul 19 23:24:56 the universal one sucks *more*, but overlays in general often seem more hassle to me than just customizing the main dts Jul 19 23:25:32 overlays only make sense if you really intend to dynamically swap them in and out at runtime Jul 19 23:25:48 the fact that unloading frequently results in a kernel panic however doesn't help that use case much :P Jul 19 23:27:48 zmatt: I thnk we've had the discussio why that wuld be an unusual use case.. Jul 19 23:28:53 well typically a DT change accompanies a hw change Jul 19 23:29:27 though even then, rebooting is a sufficiently minor inconvenience that I never really bother with overlays Jul 19 23:46:58 where can I find cape manager docs for 4.4 kernel? can't find proper docs for how to disable cape-CBB-Serial Jul 19 23:55:55 * zykotick9 doesn't use video on his BBB due to non-free software requirements :( but as an "always on, light server" he think the BBB kicks ass! Jul 19 23:56:21 zykotick9: there's no "non-free software requirements" unless you want to use the 3d gpu Jul 19 23:58:11 zmatt: i'm just avoiding the entire issue, and sshing into it... i have heard (from fsf single-boared suggestions page) that some video might be possible in freedom - but i _know_ no video is safe ;) thanks for the hint! Jul 19 23:58:28 ehm Jul 19 23:58:45 you have to put in a lot of effort to get the non-free driver working if you *wanted* to Jul 19 23:58:49 the bbb doesn't ship with it Jul 19 23:59:50 ahhh, bbb's default debian image does have non-free stuff :( /me used the minimal image to remove any non-free and hasn't had any issue with his ssh'ing only install :) Jul 20 00:00:02 what non-free stuff? Jul 20 00:00:35 I start with the minimal image too but mostly since I want to avoid unnecessary crap Jul 20 00:00:42 ? you tell me. the crappy/shitty vrms will tell ya. <- assuming debian. Jul 20 00:01:02 well you're the one making random claims here Jul 20 00:01:06 zmatt: +1 boat sucks! Jul 20 00:02:08 but if you want graphics just install x11, or wayland, or use qt5 applications with the linuxfb backend, whatever floats your boat Jul 20 00:02:21 zmatt: the only non-debian main only thing i have left on my bbb-server is the repo-rcn-ee.com repo... Jul 20 00:02:57 * zykotick9 doesn't need graphics on either of his bbb right now <- though, at one time, thought it might be neat Jul 20 00:04:43 the sgx530 is the only component on the am335x which needs a closed-source driver, or more precisely it needs closed-source userspace libraries (the kernel components are GPL. horrid, but GPL nevertheless) Jul 20 00:05:01 "tried to boot from mmc" "bad magic" Jul 20 00:05:04 and like I said they're not included by default afaik and don't even work under X11 Jul 20 00:05:57 zmatt: i _know_ there is either contrib and/or non-free in some of the default BBB-debian images. Jul 20 00:06:19 *shrug* Jul 20 00:06:27 if you say so. which packages are you concerned about? Jul 20 00:07:44 zmatt: i'm not. using the "minimal" bbb image, doesn't have the problem. i installed what i wanted, and everything was great <- read "free" IMO Jul 20 00:11:39 well you were just complaining you didn't use video due to "non-free software requirements" and that "bbb's default debian image does have non-free stuff", so then when I ask which packages you are concerned about and replying "i'm not" is inconsistent Jul 20 00:23:56 zmatt: is something about bbb on this page inaccurate http://www.fsf.org/resources/hw/single-board-computers <- i wouldn't know. based on that page's info, i've just avoided using video on my bbb entirely, and after installing the crappy vrms on the default image, and finding either contrib/non-free - i used the minimal image and just avoided video altogether (i don't need it personally) Jul 20 00:25:52 zykotick9: that page has no info on the beaglebone Jul 20 00:26:07 * zykotick9 has two bbb systems, one with the default image (for using for flashing bios of lenovo system with libreboot) and one with a "minimal" bbb image, as a home server (which he finds VERY fuckin' handy) Jul 20 00:26:46 zmatt: ahhh, first paragraph under "Single-board computers with serious flaws" Jul 20 00:27:44 zykotick9: and even then the page mentioned it's just 3d (and on the beagleboard video decoding acceleration) that is the problem, with the workaround being "do these jobs on the CPU" Jul 20 00:27:56 zmatt: nice Jul 20 00:28:27 zykotick9: the beaglebone is not the beagleboard, and the am335x is not an omap Jul 20 00:29:45 veremit: what? Jul 20 00:30:56 zmatt: "do decoding on the cpu" .. because yeah, plenty of spare capacity there lol Jul 20 00:31:01 zykotick9: and like I said, that non-free blob isn't even included with the standard debian distro Jul 20 00:32:41 veremit: was IVA actually working in the standard software releases for Jul 20 00:32:50 beagleboard Jul 20 00:32:55 duno Jul 20 00:33:28 zykotick9: quite often the closed source blobs are sufficient trouble to get working that most people don't even bother :P Jul 20 00:34:54 maybe it was working, dunno, there's nothing like iva on the beaglebone anyhow so that's a non-issue Jul 20 00:36:27 zykotick9: the fsf page seems rather full of hyperbole Jul 20 00:36:35 * zykotick9 love his BBB, but _would_ switch, if a totally free SBC became available... Jul 20 00:36:49 like "The Raspberry Pi requires nonfree software to start up." Jul 20 00:36:54 zmatt: i hear ya. but that hyperbole is important (to some people, like me) Jul 20 00:37:01 hey, you mean like, basically every PC in the world (BIOS) Jul 20 00:37:29 thus the reason i use libreboot on my lenovo systems... Jul 20 00:37:51 what about the closed source microcode of your cpu? Jul 20 00:37:52 s/lenovo/started as lenovo/ Jul 20 00:39:01 zmatt: i do use systems pre-https://libreboot.org/faq/#intelme Jul 20 00:40:09 zmatt: honestly, i'm NOT trying to insult you (or your computing needs)! Jul 20 00:40:17 gotta go. Jul 20 00:40:40 zykotick9: yet you make unsubstantiated claims about the bbb based on a unverified misreading of some fsf page Jul 20 00:40:56 do your homework better next time Jul 20 00:49:21 heh, funny, apparently all recent AMD processors have an ARM core on die Jul 20 00:51:55 okay, libreboot sounds like a bunch of idiots Jul 20 01:25:13 uboot remains the Standrd Jul 20 01:26:04 even if ts not perfect Jul 20 02:03:50 where does the "cape_universal=enable" at the end of the kernel command line comes from? its not in uEnv.txt Jul 20 02:44:44 Does anybody have any luck with max14830 or any of the other max chips supported by max310x kernel driver on 4.4 kernel? Jul 20 02:46:57 or do i just not understand how the dts files work? Jul 20 02:50:46 vkushner: eww eww Jul 20 02:51:35 laurittr: you need to get some more debug info, either from dmesg, or by adding debug info to the max driver when it loads .. so that you can see what's happening Jul 20 02:52:48 veremit: dmesg just says the overlay is loaded, cat /sys/bus/spi/devices/*/modalias -> spi:max14830, but the device is not visible in /dev/ Jul 20 02:53:41 hmm that could mean a lot of things Jul 20 02:56:25 http://lxr.free-electrons.com/source/Documentation/devicetree/bindings/serial/maxim,max310x.txt says i need to define a clock, but i dont understand where? I have connected an external crystal to my max chip **** ENDING LOGGING AT Wed Jul 20 02:59:58 2016