**** BEGIN LOGGING AT Tue Jun 14 02:59:59 2016 Jun 14 07:19:52 Hello, i have a question. Does someone know the longevity of a beaglebone? Jun 14 07:21:42 Or else the best way to contact the person who can answer this question. Jun 14 07:30:29 Guest97809: what do you mean by this? Jun 14 07:33:40 life expectancy for beagle? Jun 14 07:33:51 does that mean for how long one can get replacements for specific board? Jun 14 07:36:41 or how long it will take to wear one out? Jun 14 10:26:23 entirely dependent on how you abuse it :P Jun 14 10:28:46 the processor is officially rated for 100000 power-on hours anyway Jun 14 10:31:02 Hi People Jun 14 12:45:23 eyyyy Jun 14 12:56:47 hi, I'm wondering if any gpio could be used as pwm outputs (except reserved io) ? Jun 14 12:59:46 explain myself : all pwm outputs is used. Could I use PRU for that ? Pru outputs will be LOW or HIGH following a timer Jun 14 13:00:07 beaglebone black user Jun 14 13:00:27 yup pru is probably the most convenient way to make software pwm Jun 14 13:03:10 ok, outputs limitation in term of frequency is the same as "real" pwm outputs Jun 14 13:10:28 I didn't mention PRU because I'm of the impression that *all* GPIO are not available to them. But the ones that are should be good to go. I would expect the limit on frequency to be how well written your PRU code is. Jun 14 13:15:31 Ragnorok: yeah, just accessible pru outputs Jun 14 14:38:04 Is the instruction set for the AM18xx and AM335x PRUs the same? If not, can anyone tell me where I can find info on AM335x PRU instruction set which is similar to this info: http://processors.wiki.ti.com/index.php/PRU_Assembly_Instructions Jun 14 14:50:09 I have a "am335x_PRU_ASM_Guide.pdf". Perhaps searching for that will help. Jun 14 14:51:59 Ragnorok: ah! I think that PDF confirms that the ti wiki page does actually apply to the am335x, too. The instructions look the same. Thanks! :) Jun 14 14:52:04 I also have a migrate guide from AM18x to AM335x, implying they are not the same. Jun 14 14:52:17 One is happy to be of service. (wink) Jun 14 14:54:45 The migration guide says the AM335x PRUSSv2 is based on the AM18x PRUSS. The primary differences are more resources. It specifially says the instruction set is backward compatible. Jun 14 14:55:39 Ragnorok bradfa I'm reading this perhaps it could help https://github.com/beagleboard/am335x_pru_package/ Jun 14 14:56:28 Ragnorok: ah, ok, thanks! :) Jun 14 14:57:03 Ragnorok: the whole PRUSS v1, v2, v3 differences is not entirely made clear by TI. It'd be nice if TI had a reference manual linked from the technical docs on the product page for am335x parts. Jun 14 14:57:58 lol. Yeah. I spend a LOT of time floundering around getting the PRUs working. I settled in C, though, instead of ASM, just to remove one more level of complexity, but I was starting from scratch so I had that option. Jun 14 14:59:43 Ragnorok: I just need to know if logical right shift of arbitrary number of bits is single cycle or not, seems it is :) Jun 14 15:00:02 \o/ Jun 14 15:00:06 Ragnorok: thanks for your help! :) Jun 14 15:00:20 Ragnorok: when you're bit banging a parallel bus at high speed, things like this get fun :) Jun 14 15:00:33 One is happy to be of service! I got enough from #beagle, it's good to give back. Jun 14 15:03:07 What kind of speed are you getting? I don't need a parallel bus, I'm just curious. Jun 14 15:12:03 meh i turned on spi and ran a sample program but my probes aren't picking it up Jun 14 15:12:09 does it have to be connected? Jun 14 15:12:17 can i just leave the pins open and probe it? Jun 14 15:39:43 Ragnorok: not sure yet, code is a long ways off from being written and will only get written if SPI isn't fast enough. Ask me again in like December :) Jun 14 15:45:01 lol Jun 14 16:02:29 :-/ Jun 14 17:07:49 I'm trying to load a device tree fragment that enables spi with 6 gpio-chip selects. I was told that it doesn't work for kernel 3.8, so I loaded the current Debian with kernel 4.4 to a micro sd card Jun 14 17:09:09 However it doesn't contain /sys/device/bone_capemgr/..., so I can't see what device tree fragments have actually been already loaded Jun 14 17:10:15 * /sys/devices/bone_capemgr Jun 14 17:17:07 ah, it is now located in /sys/devices/platform/bone_capemgr =) Jun 14 17:21:23 hey pyBlob i'm actually working on spi right now, i didn't realize it was gonna be usch a headache Jun 14 17:21:34 i thought the scripts woudl work out of the box :-( Jun 14 17:21:43 let me know if you have any luck Jun 14 17:23:12 well, just firing up spi was super easy (at least for the linux with kernel 3.8) Jun 14 17:23:44 it should work similar with the current kernel, however some things have moved and I'm currently searching for the right places to make changes Jun 14 17:24:41 so i'm 4.4 i used configfs (not cape manager) to load BB-SPIDEV0-0010.dtbo Jun 14 17:24:56 sorry that should be an A Jun 14 17:25:37 and i can see my /dev/spi1.{0,1}. I can also ioctl for max_speed and bit_width or w/e etc Jun 14 17:26:32 i get no errors but when i try to tx and rx with SPI_IO_IOCMESSAGE(1) i can't see anything on my scope. i have nothing hooked up right now because i was just hoping to see the clock Jun 14 17:26:57 the rx buffer does get stuffed with 0's tho... Jun 14 17:27:14 now i'm digging through the TRM and driver code, vom Jun 14 17:27:24 i was hoping for a before lunch SPI success haha Jun 14 17:27:33 you can try using just "echo blub > /dev/spi.0", this should at least trigger the clock and send some data ^^ Jun 14 17:27:53 * /dev/spi1.0 Jun 14 17:31:42 yeah i just tried it again Jun 14 17:31:45 i see anothing Jun 14 17:33:37 i'm gonna try and yes > /dev/spi1.0 Jun 14 17:33:53 actually it's spidev1.0 gith? Jun 14 17:33:55 right* Jun 14 17:34:11 I know that, but sorry I'm also just looking to make it work somehow xD Jun 14 17:34:19 ayjay: if you're running the stock debian image with 4.4 you should have the universal-io cape loaded by default, which enables the 2 SPI buses at boot in order. So in that case /dev/spidev0.x would correspond to SPI0 Jun 14 17:34:33 alexhiam: i've already disabled that universal Jun 14 17:35:07 cape_universal=disable Jun 14 17:35:09 ayjay: but you're gettin /dev/spidev1.x? so theres also a /dev/spidev0.x? Jun 14 17:35:19 theres no spidev0 Jun 14 17:35:53 i coudl try loading the SPIDEV1-00A0 and see what pops up... Jun 14 17:35:55 ohh, does Linux bus numbering start at 1?? Jun 14 17:36:14 yeah its really strange Jun 14 17:36:16 i think thats it Jun 14 17:36:25 it depends, on some boards it starts with 1, on others with 0 Jun 14 17:36:29 there were some complains about that in the dts Jun 14 17:37:20 yeah, starts at 1, I always forget that... Jun 14 17:37:25 yeah i enabled spi0 && spi1 and now i have a devspi{1,2} Jun 14 17:37:35 naturally Jun 14 17:37:39 ayjay: gotta ask, you sure you got the right pins? Jun 14 17:38:49 ive checked it quite a few times Jun 14 17:38:52 i've been doing this for a cpl hours Jun 14 17:39:24 i'm probing 31 and 32 on P9 Jun 14 17:39:44 acutally alexhiam i'm gonna reenable the universal cape and boot up Jun 14 17:39:48 that'll let me know if its a dt problem Jun 14 17:40:25 P9_31 and P9_32 are for SPI1 Jun 14 17:40:34 sorry i meant 22 and 31 Jun 14 17:40:42 i'm prbing both clocks and trying both spi's Jun 14 17:40:46 i've probed the datas too Jun 14 17:40:54 weird Jun 14 17:41:07 yeah, I'd try with universal-io Jun 14 17:41:19 you'll need to mux the pins with config-pin as well Jun 14 17:41:59 with universal-io? Jun 14 17:42:09 is that a userspace entry am335x-pinmux? Jun 14 17:42:45 essentially Jun 14 17:42:58 hmm never used it Jun 14 17:43:17 AH does disabling hdmi with enabling emmc have something to do with it? Jun 14 17:43:22 AHHHH Jun 14 17:43:24 works with universal-io to allow pinmuxing from userspace Jun 14 17:43:48 sorry but when i renabled universal, i got no spidev, but when i reenabled hdmi, i did Jun 14 17:43:53 let me see if i can see the clock... Jun 14 17:44:25 lemme see about that config-pin.. Jun 14 17:44:36 oh right, SPI0 conflicts with the eMMC Jun 14 17:44:59 wait I thought eMMC was our 4gb? Jun 14 17:45:37 yeah Jun 14 17:46:01 so if you're booting from the eMMC you can't use SPI0, only SPI1 Jun 14 17:46:46 if ur booting from it, i see. but can u sill access it? Jun 14 17:47:44 I'm sure you could mount it when booting from an sd card, but I've never tried Jun 14 17:48:22 i've actually never seen that in the docs, alexhiam. that spi0 conflicts with emmc Jun 14 17:48:36 i've only see that spi1 conflicts with hdmi's mcasp audio Jun 14 17:48:54 ayjay: oops, I lied, I was looking at P8 Jun 14 17:49:55 config-pin is cool tho, you can load overlays with it? howd i not know Jun 14 17:50:13 yeah, it's new on 4.x Jun 14 17:50:31 just discovered it too =) Jun 14 17:50:57 i'd really like to see announcements like these on the beaglebone website Jun 14 17:51:27 i should just google "github beagleboneblack" every so often Jun 14 17:52:01 rcn usually announces new stuff like that on the ml Jun 14 17:52:40 :-( i'm not good at mailing lists yet Jun 14 17:52:46 but i don't think i'm going to get away with that Jun 14 17:52:55 also, the config-pin utility is changing my universe Jun 14 17:52:59 would be good to make sure that stuff makes it in the newsletter as well Jun 14 17:53:15 shouldn't the am335x-pinmux fragment on the SPIDEV0 dtbo have done mode setting? Jun 14 17:53:32 yeah Jun 14 17:53:35 https://github.com/RobertCNelson/bb.org-overlays/blob/master/src/arm/BB-SPIDEV0-00A0.dts Jun 14 17:55:00 okay so i guess that works Jun 14 17:55:13 i can definitely see activity with yes>/dev/spi1.0 Jun 14 17:55:26 but it does need a significant amount of bytes for me to catch it on my scope Jun 14 17:55:27 with universal-io? Jun 14 17:56:45 it defaults to 16MHz, and there's probably a comparatively long gap between 'y's Jun 14 17:56:46 yeah plus pinconfig or configpin Jun 14 17:58:06 lots of help tho hiam. i got RGB detached from HDMI yesterday with the tilcdc driver after reading the source and i went to go use but found out that the lcd-controller wants its mode set with SPI Jun 14 17:58:29 Hi guys Jun 14 17:58:39 i was gonna write some doc for using tilcdc when i got the change Jun 14 17:58:40 chance* Jun 14 17:59:19 I have a problem. I have connected 2 beagle bone black to laptop, i dunno how to access them separately via IDE Jun 14 17:59:56 swaroop: you'll have to change one of their host names Jun 14 17:59:59 when i connect to http://192.168.7.2:3000/ide.html , i know that one beagle bone is connected. I have no clue how to access the other beagle bone, is there anything i need to change in the url ? Jun 14 18:00:51 oh ok thanks, will look into it Jun 14 18:01:14 swaroop: can you plug them in over ethernet instead? that'll be easier because that network-over-usb IP address doesn't use DHCP Jun 14 18:01:39 ayjay: what LCD? Jun 14 18:02:00 Another question I had was, I want to establish offline communication between beagle bone, I mean i just ethernet switch (no router or internet), I want send data from one board to other, is it possible ? Jun 14 18:03:20 Just put in static IP addresses (e.g. call one 10.0.0.1 and the other 10.0.0.2 and they can talk. Jun 14 18:03:40 just make sure they're on the same subnet Jun 14 18:04:16 alexhiam: I see, as i just posted other question, I can plug into ethernet switch. But they dont have router to assign auto ip address. I have to do it manually. Ok got it Jun 14 18:04:45 dirkx: oh i see, thanks, will try these things and get back to you guys. Jun 14 18:06:16 I am just getting started with BB. My end aim is, There is one ethernet switch. I will connect 5 beagle bones to that ethernet switch and they should be able to communicate with each other. Jun 14 18:06:36 I just got my beagle bone 1 week back. :) Jun 14 18:08:50 you could set one up as a DHCP server, but that would be a bit more complicated Jun 14 18:09:11 it's a new lcd distributor Jun 14 18:09:24 they want to give bbb's with their lcd's as gifts to generate sales i guess Jun 14 18:10:02 cool Jun 14 18:10:05 how big? Jun 14 18:10:25 1.8-7 Jun 14 18:10:39 nice Jun 14 18:10:40 about 10, some r-touch, some no touch Jun 14 18:10:45 10 different models Jun 14 18:10:58 three separate controllers Jun 14 18:11:11 depends on size Jun 14 18:11:22 theres literally hundreds of possible protocols Jun 14 18:12:44 personally i like high precision electronics more. i'd like to build an SEM or a electron-PVD but I'll need time for that Jun 14 18:13:51 well that would be cool Jun 14 18:13:58 theres a kid who did it at my university Jun 14 18:14:06 built an electron beam PVD Jun 14 18:31:51 hi, do you people use this to interface PRU: https://github.com/beagleboard/am335x_pru_package ? Jun 14 18:32:47 If it uses uio_pruss, that's the route I took; remoteproce may be a fine and dandy deployment mechanism but for development it's hard to imagine something less suitable. Jun 14 18:33:15 it uses uio_pruss Jun 14 18:33:25 but I am having some problems with it Jun 14 18:33:38 https://github.com/beagleboard/am335x_pru_package/blob/master/pru_sw/app_loader/interface/prussdrv.c#L81 Jun 14 18:33:42 on this line Jun 14 18:34:08 it tries to open /sys/class/uio/uio0/maps/map0/addr Jun 14 18:34:19 i am getting no such file or dir Jun 14 18:34:29 Most likely uio_pruss hasn't loaded correctly. Jun 14 18:34:40 maciejjo: what kernel version? Jun 14 18:34:49 If you cat /sys/class/uio it should have stuff in it. Jun 14 18:35:01 Sorry. ls, not cat. Jun 14 18:35:23 Linux beaglebone 4.4.12-g51452e4664 Jun 14 18:35:32 Ragnorok: I do have stuff Jun 14 18:35:37 Interesting. Jun 14 18:35:51 I don't thing 4.x has uio_pross anymore Jun 14 18:35:58 I patched it Jun 14 18:36:08 that'll have the new TI rproc stuff Jun 14 18:36:13 Does the stuff look like this: uio0 -> ../../devices/platform/ocp/4a300000.pruss/uio/uio0 Jun 14 18:36:21 yes it does Jun 14 18:36:26 Interesting. Jun 14 18:36:27 8 times this Jun 14 18:36:35 Yup. 0 to 7. Jun 14 18:36:40 exactly Jun 14 18:36:52 I am trying to make uio-pruss work on linux-ti Jun 14 18:36:52 Seems like uio_pruss is working then. Jun 14 18:37:52 If you drive to that dir, is addr there? And if you cat that, does have a value? Jun 14 18:37:56 maciejjo: have you looked at the latest stuff from TI? it's really not too bad anymore Jun 14 18:38:11 https://gist.github.com/alexanderhiam/2c4187c710b2c409d8dde8c4015fe007 Jun 14 18:38:13 Mine is present and says 0x4a300000, which is the PRU0 Data RAM base address. Jun 14 18:39:00 alexhiam: I did and I plan to learn how to use it sometime in the future, but now I wanted to do this quick Jun 14 18:39:14 fair enough Jun 14 18:39:38 Ragnorok: no, addr is not present Jun 14 18:39:53 Odd. Jun 14 18:40:00 so I figure something went wrong when creating these devices Jun 14 18:40:01 I have addr, name, offset, size. Jun 14 18:40:09 Seems like it. dmesg? Jun 14 18:40:18 Or kern.log. Jun 14 18:40:23 pruss_uio 4a300000.pruss: pins are not configured from the driver Jun 14 18:40:39 Aha. You didn't use the overlay then. Jun 14 18:40:53 no, but I patched the main DT Jun 14 18:41:04 i do not have overlay support on this kernel Jun 14 18:41:08 Hmm. Mine is as well. Are you sure that's being done correctly? Jun 14 18:41:27 I think there's some way to check pin mappings. Jun 14 18:43:25 http://pastebin.com/raw/ds0EG43y Jun 14 18:43:33 here is roughly what I did in DT Jun 14 18:44:08 Let me load mine and have a peek. I couldn't remember to win Jeopardy. Jun 14 18:46:41 Huh. Sorting by date doesn't have the change on top. Jun 14 18:53:13 Still looking. I didn't do this, so I may not have it close to hand. I've pinged the guy who did. Jun 14 18:54:19 Ok, thanks! Jun 14 19:04:16 He says this: http://pastebin.com/mJzwXeUr Jun 14 19:05:05 I think my patch does more or less the same... Jun 14 19:07:19 It's just a lot more verbose. Jun 14 19:07:23 ok, so I've managed to load my device tree fragment on the new kernel 4.4, this time it also doesn't crash like 3.8. However I now can only see 2 spi devices (instead of 6). This is the device tree file I use: http://pastebin.com/FWEK55qc Jun 14 19:07:57 @maciejjo, did you install uio_pruss? Jun 14 19:08:45 He has /sys/class/uio stuff, but no addr for instance under /sys/class/uio/uio0/maps/map0 Jun 14 19:09:36 Browce: as in kernel module? I do have it Jun 14 19:10:23 http://pastebin.com/raw/BrLNcp95 Jun 14 19:10:30 Yeah, that's what I meant. I wasn't sure where you were at. Jun 14 19:10:40 this is what I have in uio0 Jun 14 19:11:24 Browce: it is present and loaded on startup Jun 14 19:11:49 just prints one warning "pruss_uio 4a300000.pruss: pins are not configured from the driver" but I do not know if that is related Jun 14 19:12:54 Sounds possibly related, I'll take a closer look at your dtb, one sec Jun 14 19:12:55 maciejjo: If I follow the symlink to /sys/devices/.../uio0 I see the same thing as your pastebin. Jun 14 19:13:15 Ah! No I don't. I have "maps" and you don't. Jun 14 19:13:56 Ragnorok: what kernel are you using? Jun 14 19:14:48 4.1.19-bone20. It came with uio_pruss otb. Jun 14 19:14:49 maciejjo, I don't see this part in your patch anywhere Jun 14 19:14:54 "&pruss { Jun 14 19:14:54 status = "okay"; Jun 14 19:14:54 };" Jun 14 19:15:08 it is set to okay in the main node Jun 14 19:15:20 We're going to move to 4.4-something at some point, iirc, but we haven't. Jun 14 19:15:35 If it was disabled the driver would not load Jun 14 19:15:48 Thats a good point Jun 14 19:16:24 do you guys have some pinmux setup for PRU? Jun 14 19:16:46 maybe that's what i'm missing Jun 14 19:18:04 We should just have what's in the pastebin, which doesn't seem to have any pinmux settings. We aren't specifically doing anything outside the dtb. Jun 14 19:18:07 Not that I explicitly added, maybe it was already in the source Jun 14 19:23:01 if you do "dmesg | grep uio" do you see the warning I am seeing? Jun 14 19:27:02 @maciejjo, I'll go power up my beagle and check, one sec Jun 14 19:29:37 [ 8.042387] pruss_uio 4a300000.pruss: pins are not configured from the driver Jun 14 19:29:37 [ 8.042420] pruss_uio 4a300000.pruss: pruss_probe() get_resource ok Jun 14 19:30:33 Ragnorok, what about you? Same deal? Jun 14 19:37:20 That in kern.log? Jun 14 19:38:39 I don't see the pruss_probe() line. Jun 14 19:39:19 In kern.log or dmesg. Jun 14 19:40:35 dmesg Jun 14 19:40:41 dmesg | grep uio Jun 14 20:26:46 zmatt: btw: if you know someone with decent combination of EE and CS knowledge who is looking for a cool place to do a PhD, please let me know Jun 14 20:27:01 zmatt: we are looking for someone to expand our team Jun 14 21:03:01 Do a PhD. lol Jun 14 21:03:14 hello Jun 14 21:03:23 I need a little help Jun 14 21:03:58 I bricked my BBGW by messing up uEnv.txt Jun 14 21:04:47 I am trying to uBoot from the SD card with a minimal uEnv.txt and the latest Deb 8.4 img but not getting very far Jun 14 21:06:17 On a normal mmc boot, I am unable to ssh and the wifi is not working. Jun 14 21:06:57 On SD boot, I just get a power and BT leds, also can't SSH in Jun 14 21:17:48 Have you tried using the serial console, say with a FTDI cable? Jun 14 21:18:32 join #beagle Jun 14 21:18:38 hello? Jun 14 21:33:22 lol Jun 14 21:33:25 hello **** ENDING LOGGING AT Wed Jun 15 02:59:58 2016