**** BEGIN LOGGING AT Mon Sep 26 02:59:58 2016 Sep 26 04:29:19 lol the am3352 design specs in the datasheet are so specific i'm surprised we don't just get gerbers Sep 26 04:33:47 Hi. Can someone point me to the factory image for a BeagleBone Green Wireless? I tried bone-debian-8.4-lxqt-4gb-armhf-2016-05-13-4gb.img and it works, but its not the same as the image from Seeed. thanks. Sep 26 04:34:49 have you tried the seeed website? Sep 26 04:35:14 Looked around, yes. They mainly point to beaglebone.org/latest-images Sep 26 04:35:58 from here: http://wiki.seeed.cc/SeeedStudio_BeagleBone_Green_Wireless/#update-to-latest-software? Sep 26 04:38:44 That's what I tried first. It just links to the bone-debian-8.4 build. That build doesn't have the tethered wireless setup, or at least I don't know how to activate it from the regular build. Sep 26 05:58:04 mornin -.- Sep 26 06:28:47 Hi there Sep 26 06:40:23 Morning Sep 26 06:40:30 meh!ning Sep 26 06:55:58 Getting my new hdmi cable today Sep 26 06:57:07 doing science today! Sep 26 07:03:18 KotH, :-D Sep 26 08:17:14 KotH: SCIENCE! \o/ Sep 26 09:49:25 tommer: don't ask to ask, just ask. and don't unnecessarily use private messages, I'm not a personal assistant Sep 26 10:00:01 Hello Sep 26 10:00:09 Anyone able to run LXC on DRA7xx Sep 26 10:00:44 linux container should work basically on any linux system as far as I know Sep 26 10:00:53 there's nothing hardware-dependent in them Sep 26 10:02:04 but I am trying on Dra7xx BSP and always getting segfault Sep 26 10:02:17 that sounds pretty disturbing Sep 26 10:02:36 any idea why this leads to this problem Sep 26 10:03:08 no. does systemd-nspawn suffer the same problem? Sep 26 10:03:20 which kernel version are you on? Sep 26 10:03:28 3.14.49 Sep 26 10:03:48 how to verfiy systemd-nspwan Sep 26 10:04:07 ehm, maybe consider a not-ancient kernel? I know the latest sdk from ti uses kernel 4.4 Sep 26 10:04:36 correct but customer requirement is 3.14 Sep 26 10:04:45 I think it should work on it too Sep 26 10:05:06 it probably should Sep 26 10:05:21 I don't think older sdks use systemd so no systemd-nspawn available either Sep 26 10:05:57 we uses systemd Sep 26 10:05:57 I'd start out by stracing the command that fails Sep 26 10:06:09 yes that sounds like a good idea also Sep 26 10:06:25 and check kernel log just in case Sep 26 10:07:27 I tried with strace but no hinit Sep 26 10:07:43 gettimeofday({1473846850, 21345}, NULL) = 0 --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x8} --- +++ killed by SIGSEGV +++ Sep 26 10:08:40 https://www.stgraber.org/2013/12/23/lxc-1-0-some-more-advanced-container-usage/ here my detail logs last section Sep 26 10:09:17 just curious though, if you're using the dra7xx, why are you asking here? presumably TI supports automotive customers directly? Sep 26 10:09:50 do we have dra7xx #IRC Sep 26 10:09:59 I thought you guys can provide some hints Sep 26 10:11:10 we just did, but we're not free commercial support Sep 26 10:11:15 that Sep 26 10:11:44 I am working on research Sep 26 10:11:49 btw: I'd dig further into the backtrace Sep 26 10:12:08 and as I said, lxc itself isn't actually hardware-dependent, so either the problem lies with how you're using it or it's a manifestation of a problem with the particular sdk you're using Sep 26 10:13:30 if you're using systemd and have systemd-nspawn installed then you can try systemd-nspawn -D /path/to/rootfs to spawn a shell it in it Sep 26 10:13:43 add the --boot option to try to boot the system Sep 26 10:14:09 wait, that blogpost you mention. It goes into running qemu from lxc Sep 26 10:15:45 as one of its topics Sep 26 10:17:09 I tried systemd-nspawn -D /path/to/rootfs/bin/bash, it says Directory /rootfs_1bin/bash lacks the binary to execute Sep 26 10:17:27 systemd-nspawn -D /path/to/rootfs Sep 26 10:18:47 it saying Client died too early Sep 26 10:19:00 where to pass --boot option Sep 26 10:19:10 that doesn't sound good Sep 26 10:19:24 if not even spawning a shell works I wouldn't even bother trying --boot Sep 26 10:20:10 I am passing rootfs of debain compile for arm Sep 26 10:20:35 arm or armhf ? Sep 26 10:21:35 arm Sep 26 10:21:58 ELF-32-bit LSB executable, ARM EABI4 version Sep 26 10:22:11 can I pass same filesystem through which I booted Sep 26 10:22:31 a copy thereof Sep 26 10:23:42 what could be the rason of "Client died too early" Sep 26 10:24:14 no idea, but I'm suspicious of how you compiled the system. can you run the executables you compiled *outside* the container? Sep 26 10:26:08 on my beaglebone, running file on an executable says: Sep 26 10:26:09 ELF 32-bit LSB executable, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux-armhf.so.3 Sep 26 10:26:32 when I am running systemd-nspawan -D /roofs it says Timezone UTC does not exist in container, not updating container timezone Sep 26 10:26:37 that's okay Sep 26 10:26:58 or actually maybe it isn't okay Sep 26 10:27:16 since you showed you were getting a segfault after gettimeofday() Sep 26 10:27:38 missing timezone information really ought not result in a segfault though Sep 26 10:27:48 yes I have same understanding Sep 26 10:28:23 file lxc-attach says " ELF 32-bit LSB executable, ARM, EABI5 version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux-armhf.so.3" Sep 26 10:32:10 try grabbing e.g. https://rcn-ee.online/rootfs/bb.org/testing/2016-09-25/console/debian-8.6-console-armhf-2016-09-25.tar.xz and use systemd-nspawn on that tree Sep 26 10:33:00 ok let me try Sep 26 10:59:22 hi zmatt Sep 26 10:59:42 same result Sep 26 10:59:49 Client die too early Sep 26 11:02:03 any hint Sep 26 11:02:44 ¯\_(ツ)_/¯ Sep 26 11:03:22 what does it mean Sep 26 11:06:24 it means the child process created by nspawn died before even getting the chance to execute the target executable Sep 26 11:06:51 how to debug this Sep 26 11:06:52 it also means you're using an old version of systemd since the error message in current versions is "Child died too early" rather than "Client died too early" Sep 26 11:07:43 it saying Client dies too early Sep 26 11:08:41 I am using systemd 225 Sep 26 11:11:39 * ogra_ wonders if killing children is less barbarian than killing clients Sep 26 11:14:18 zmatt: my current SPI implementation uses ret = ioctl(spi, SPI_IOC_MESSAGE(1), &tr); to exec. spi com. on spi0 or spi1. I am often doing SPI read of 1024 bytes, which I would like the SPI0 and SPI1 to do in parallel. Now the IOCTL call is blocking Sep 26 11:15:48 tommer: I already explained: since the spi calls are blocking you need to use separate threads or processes to utilitize both buses in parallel Sep 26 11:16:05 *utilize Sep 26 11:18:17 yes, but wouldnt the IOCTL be blocking in the one thread before the context switch happens and the other spi starts the read? Sep 26 11:18:26 no Sep 26 11:19:07 especially not when using sufficiently large transfers Sep 26 11:19:46 the actual transfer is done using DMA, the process is just blocked until it completes but other processes can run at that time Sep 26 11:19:47 its 1024 bytes, but I cannot wait for it to finish before using the other spi periph. since my device has a high frame rate Sep 26 11:20:58 it's a bit annoying there's no async userspace API since the underlying kernel API in fact is asynchronous, but well that's how it is Sep 26 11:22:38 ok, then I'll try to split into two threads. Do you happen to know what the transfer size limit for DMA use is in the kernel API ? Sep 26 11:25:40 if there's a limit at all it's probably either 64KB or 64Kwords (where a word is the configured transfer size) Sep 26 11:27:02 ah no 4 KB, due to the userspace interface actually rather than the kernel api Sep 26 11:27:37 it's a kernel module parameter though so you could enlarge that Sep 26 11:29:48 hmm, so for transfers below this parameter the spi communication will be blocking? Since its not using the DMA for the transfers Sep 26 11:30:08 The default image for the X15 seems to be missing a few firmware files, can i just yoink them from the TI SDK or should i grab them from somewhere else? (the PRU firmwares) Sep 26 11:30:12 huh, no transfers larger than 4 KB will just return an error code Sep 26 11:30:40 DMA is used in all cases afaik (except maybe 1 or 2 word transfers) Sep 26 11:31:14 Velsiol: the PRU cores have no predefined function, you can load whatever firmware you want onto them Sep 26 11:31:48 thanks zmatt ! Sep 26 11:34:14 Ah all right! I was wondering since it threw some errors at startup. Same for dra7-ipu1? Sep 26 11:34:44 yeah, they're not really errors Sep 26 11:35:24 Haha i've spent the last few days wondering how to fix those, well at least i know now, thanks! Sep 26 11:35:52 yeah, just figure out something useful for each to do and write suitable firmware ;) Sep 26 11:44:27 Ah i thought the dra7-ipu1 missing had something to do with me not being able to run any dsp programs. If i try to run any of the opencl examples i get an error "Unable to allocate OCL MSMC memory from 0x40500000" Google gives me nothing :( Sep 26 11:45:14 no idea about that, but I don't see any way the IPUs would be involved in that Sep 26 11:45:25 I configure cmem with the arguments suggested by the examples (phys start 0xa0000000, phys end 0xc0000000, pools=1x536870912,) Sep 26 11:45:42 Ah allright, at least i know it isnt that, thanks! Sep 26 11:54:23 For my two SPI Read threads, how can I make sure that the kernel will not interrupt them? Do I need e.g. the real time patch preempt_rt ? Sep 26 11:55:02 ??? Sep 26 11:55:07 interrupt them with what? Sep 26 11:57:42 preempt_rt is only useful under special circumstances and usually requires that your code is written to take advantage of it. when in dobut, you don't need it Sep 26 11:59:07 Well, with a RTOS you get determinism, so a command will have a given response time no matter what, but without the response time can be varying Sep 26 11:59:38 and I really need the SPI reads to happen when I say it due to my systems sample rate Sep 26 12:03:33 it all depends on how tight those constraints are. if I'd really require SPI transfers in response to some external stimulus (e.g. an end-of-conversion signal from an ADC) I would hook that signal up to one of the external dma trigger pins and setup dma for that Sep 26 12:04:18 how many microseconds of timing margin do you have? Sep 26 12:06:25 ConnectionError: HTTPSConnectionPool(host='api.gitter.im', port=443): Max retries exceeded with url: /v1/rooms (Caused by NewConnectionError(': Failed to establish a new connection: [Errno 110] Connection timed out',)) (file "/usr/local/lib/python2.7/dist-packages/requests/adapters.py", line 437, in send) Sep 26 12:06:50 ok who is running this annoying bot here and *why* ? Sep 26 12:23:49 afk Sep 26 13:00:02 I have approx. 20 us margin Sep 26 13:09:03 Anyone there? Sep 26 13:09:11 I'm getting this error.. Sep 26 13:09:46 http://pastebin.com/8ayZA3QU Sep 26 13:10:17 tommer: then forget about timing this in linux userspace, I've measured irq latency to userspace to be about 40-90 us even with an RT kernel Sep 26 13:11:13 (RT doesn't improve the typical latency much, it mostly just helps to get rid of the outliers) Sep 26 13:11:30 hmm, would a kernel module have lower latency? Sep 26 13:12:04 probably, but I don't know whether it would have sufficiently lower latency. Sep 26 13:12:41 kiran4399: seems reasonably self-explanatory: you're trying to request a pin that is already in use Sep 26 13:12:45 kiran4399: I don't know dtb really, but it seems pretty clear you're attempting to use a pin that's already in use elsewhere. You have to find where it's being assigned and remove that, or change your pin to a different one. Sep 26 13:12:49 Poo. Sep 26 13:12:53 win Sep 26 13:12:54 :D Sep 26 13:13:08 ok, then I will have to settle for a lower frame rate. Maybe even try the Logibone with FPGA Sep 26 13:13:32 tommer: or use externally triggered dma Sep 26 13:13:43 that would be my first thought Sep 26 13:13:55 having PRU manage spi would be another option Sep 26 13:14:05 I'd try that one. Sep 26 13:14:44 however if you just need to transfer data in response to an external signal, using PRU seems an overly scenic route Sep 26 13:15:05 But look at the view! (grin) Sep 26 13:15:06 there are a few pins which can directly trigger an EDMA transfer Sep 26 13:17:44 hmm, using the PRU to control the SPI does not sound like a bad idea :D Sep 26 13:19:03 (the pins are P9.41 and .42 fwiw) Sep 26 13:25:18 bbl Sep 26 13:25:30 but can I set up externally triggered dma from userspace? Sep 26 13:27:05 any pointers on how to run a bbb image with qemu on my debian sid laptop? there are a lot of references around to different parameters to be used with qemu-system-arm, but none of them seem to run so far. most refer to qemu builds or ppas which are now unavailable Sep 26 14:36:51 zmatt: talk more about this tomorrow :) Sep 26 15:18:05 on reddit this morning: https://www.reddit.com/r/BeagleBone/ Sep 26 15:18:41 Hello, I needed some help on finding out how to add USB ports? Sep 26 15:19:00 actually my beaglebones have no USB ports at the moment, so I need to figure out if I can like Sep 26 15:19:05 maybe use GPIO to add ports Sep 26 15:19:10 or if there are any cheap capes Sep 26 15:29:42 kenshirriff has some nice PRU articles there. Which I'd have had them last January. (grin) Sep 26 15:31:03 Only "mistake" is the CCS compiler comes with the tool to convert the .out to to .bin for loading on the PRU, so it can be added as a post-build step. Sep 26 15:31:58 ogres: your bb has no usb ports? Sep 26 15:32:11 ogres: usb hubs are cheap Sep 26 15:32:30 <$10 from walmart Sep 26 15:32:49 I haven't even gotten a chance to play the PRU's yet meh Sep 26 15:32:57 my controller skills are weak anyway Sep 26 15:34:31 His assessment is accurate imo; very powerful, high learning curve. If you decide to wet your toes with PRU it looks like the two items I saw by kenshirriff would be an excellent tutorial. Sep 26 15:43:20 no usb ports ayjay_t Sep 26 15:46:30 that's my problem Sep 26 15:46:59 I need to either have USB ports or an extra ethernet port Sep 26 15:47:14 but my beaglebone has NO usb ports, and only one ethernet (of course) Sep 26 15:52:10 beagleboard.org shows no model that lacks USB. I presume it's been damaged somehow? Sep 26 15:55:14 ogres: every beaglebone has one host and one device port Sep 26 15:55:41 not this one Sep 26 15:55:53 in theory you might, with effort, be able to repurpose the device port as second host port Sep 26 15:56:00 its a custom beaglebone Sep 26 15:56:21 then complain to whoever made it? :P Sep 26 15:56:22 here's a link to one of t hem Sep 26 15:56:22 http://www.ebay.com/itm/Antminer-S4-S4-S5-S5-S7-BB-version-BeagleBone-Black-BB-Black-BB-board-/201675594453?hash=item2ef4cd5ad5:g:kBgAAOSwZVlXqc~- Sep 26 15:57:16 lovely how it still says usb host in the location where the usb port should have been Sep 26 15:57:43 Serious. Fly-by-night R Us. Sep 26 15:57:52 why would you use a frankenbeaglebone like this? Sep 26 15:58:47 oh nice and it's an AM3352 instead of an AM3358 Sep 26 15:59:20 huh? Sep 26 15:59:25 so if i solder a usb on there Sep 26 15:59:27 it would work? Sep 26 15:59:36 Regardless, the seller is the information point. If they won't help, then this is a lesson in buying off eBay. Sep 26 15:59:50 oh this is how it was designed Sep 26 16:00:08 there's many of these beagles out there Sep 26 16:00:16 Custom for 'bitcoin mining' Sep 26 16:00:21 but that's not what im doing Sep 26 16:00:30 so i need USBs :p Sep 26 16:00:34 ogres: I don't see the holes necessary for the usb port Sep 26 16:00:46 Then the designer is the information point. Because somone took the OSH and modded it doesn't mean it's a "BeagleBone". Sep 26 16:00:54 indeed Sep 26 16:01:03 hmm yeah me neither Sep 26 16:01:07 this is not a beaglebone black, it's a custom hackjob with the name "beaglebone black" printed on it Sep 26 16:01:14 wait its a 52? Sep 26 16:01:19 I recommend buying an unit with USB on it. Sep 26 16:01:19 i'm trying to break out a 52 now Sep 26 16:01:26 maybe they're schematics online Sep 26 16:01:41 so you can't add a USB thru gpio? Sep 26 16:01:44 no Sep 26 16:01:46 In the Arduino world a clone with their logo is illegal. Sep 26 16:01:47 lol Sep 26 16:02:04 I'm pretty sure this is illegal too isn't it? I mean beaglebone is probably trademarked, not that i care to look it up Sep 26 16:02:13 ogres: usb 2.0 is nearly 0.5 gigabit, you're not going to be able to emulate it with anything else Sep 26 16:02:14 I would expect yes. Sep 26 16:02:34 Illegal === we don't care if it works. Sep 26 16:02:44 I dont need that much speed zmatt Sep 26 16:02:48 I seen this: http://www.coolgear.com/product/beaglebone-black-usb-expansion-rs232-module?pk_campaign=googleshop&gclid=CK3txIu0rc8CFQiqaQodAvwEoQ Sep 26 16:02:53 a bit too expensive for me though Sep 26 16:03:12 honestly, just get a normal beaglebone black, or a beaglebone green Sep 26 16:03:13 i don't know i'm pretty interested in this Sep 26 16:03:38 ogres: see that usb cord on the "cape" you linked to: it needs an existing usb port Sep 26 16:03:46 oh yeah? Sep 26 16:03:49 ogres if you want to add the the usb port though i personally think that would be a fun job Sep 26 16:04:04 ayjay_t: yeah especially if the traces aren't there on the pcb Sep 26 16:04:05 i mean, you'll learn a ton Sep 26 16:04:18 zmatt: touche Sep 26 16:04:28 hmm Sep 26 16:04:37 but it looks like there using the BBB schematic Sep 26 16:04:40 Wire wrap wire directly on the SoC pins is "fun". Sep 26 16:04:41 theres no such thing as a Serial to USB is there? Sep 26 16:04:45 just without all the parts placed Sep 26 16:04:51 Serial to USB Female Sep 26 16:04:58 lmgtfy.com Sep 26 16:05:03 Serious. Sep 26 16:05:14 ayjay_t: no since like I said I don't see any holes where the through-hole usb connector is supposed to be Sep 26 16:05:41 but theres a sticker... Sep 26 16:06:07 ahh wait that's a sticker, not print Sep 26 16:06:13 yeah hey ogres Sep 26 16:06:17 check under that sticker, if you have the device Sep 26 16:06:17 yeah Sep 26 16:06:17 ogres: There are zillions. Most use FTDI or the older CP2012. If it's not female get a gender bender. Sep 26 16:06:24 dont have it yet Sep 26 16:06:32 oh well you can add a usb port if you want Sep 26 16:06:33 its not worth it Sep 26 16:06:43 i'd dispute the charge on your credit card and then buy a real bbb Sep 26 16:06:44 ogres: that cape you linked to is really bullshit btw Sep 26 16:06:49 yes agreed Sep 26 16:07:01 BUT if you are looking for a project and have a few weeks to burn Sep 26 16:07:06 oh Sep 26 16:07:09 if this was sold as a "beaglebone black" it is fraudulent Sep 26 16:07:17 I'd resell it on eBay and get a real BBB. It's not the sellers fault ogres bought the wrong thing. Sep 26 16:07:19 or the serial to usb might work Sep 26 16:07:29 Ragnorok: depends on how it was presented Sep 26 16:07:33 look at this Sep 26 16:07:34 http://www.hobbytronics.co.uk/usb-host-board-v2 Sep 26 16:07:34 I've never heard of serial-to-usb Sep 26 16:07:52 neither have i but i thought ragnorok suggested they exist Sep 26 16:08:05 no that i'd ever look for one Sep 26 16:08:08 zmatt: True. If it's illegal the seller deserves the shaft. Sep 26 16:08:28 yeah, like what ogres just linked to: containing a microcontroller with usb host support Sep 26 16:08:30 it's ambiguous. they don't make it hard to see that its not a BBB Sep 26 16:08:34 but they also say "BBB" Sep 26 16:08:45 but I seriously doubt linux would ever support one of those Sep 26 16:09:17 If it's not illegal then ogres deserves to eat whatever happens and chalk it up to education. Sep 26 16:09:34 haha ogres if you manage to get a USB port on the BBB you can probably make a lot more money in engineering than in bitcoin mining :-D Sep 26 16:09:41 hah Sep 26 16:09:43 :p Sep 26 16:09:44 i will see Sep 26 16:09:59 i should call it the non-BBB Sep 26 16:10:10 was this thing used for management of miners or what? Sep 26 16:10:34 Isn't FTDI serial to USB? I mean it claims it's "USB to serial", but it's bidirectional. It's just not host. Sep 26 16:10:46 tru Sep 26 16:11:18 Ragnorok: by FTDI most people mean their usb-to-serial converters Sep 26 16:11:24 wish there was a Ethernet to USB adapter lol.. but no such thing Sep 26 16:11:33 I'm not aware of the opposite Sep 26 16:11:40 That's what I mean as well. But it's still bidirectional. Sep 26 16:12:01 Ragnorok: usb is always bidirectional, that doesn't make a usb device a usb host Sep 26 16:12:05 You can't plug it into a hub because it's not host, but it will still provide USB connectivity. Sep 26 16:12:13 no it won't Sep 26 16:12:22 Host wasn't listed as a requirement. Only USB connection. Sep 26 16:12:23 usb is strictly asymmetric Sep 26 16:13:08 https://www.amazon.com/Insten-Female-Ethernet-Router-Adapter/dp/B00AF1PHZC Sep 26 16:13:46 shits probably fake Sep 26 16:14:58 yikes the horrors one finds on the internet Sep 26 16:15:32 amazon really ate dirt recently Sep 26 16:15:35 Particularly since Amazon has become eBay and filled with FBN sellers. Sep 26 16:15:35 http://www.newegg.com/Product/Product.aspx?Item=9SIA1NV0UF6114 Sep 26 16:15:47 yesh no kidding Sep 26 16:15:49 whats FBN? Sep 26 16:16:02 oh fictitious business name Sep 26 16:16:12 Fly By Night - shady - likely bogus Sep 26 16:16:22 ogres: I'm sure you can find adapters from any connector to any connector Sep 26 16:16:27 that doesn't mean they do anything useufl Sep 26 16:16:28 *useful Sep 26 16:16:35 it probably wont do the job yea Sep 26 16:16:45 ogres so why dont you want to get a real BBB? Sep 26 16:16:50 just out of curiosity Sep 26 16:16:52 And even if it does do the job I bet it requires a host connection. Sep 26 16:16:59 I bought these for $2 a piece Sep 26 16:17:13 Ah. Well I'd be interested for that price. (grin) Sep 26 16:17:34 ogres: well then you got good value for money I guess :P Sep 26 16:17:45 but maybe you should have considered your usb needs before buying them Sep 26 16:17:46 so why do you need usb? Sep 26 16:18:15 to add one more ethernet port Sep 26 16:18:27 why not just use an ethernet switch? Sep 26 16:18:34 Serious. Sep 26 16:18:36 there ya go Sep 26 16:18:41 the truth comes out Sep 26 16:18:50 But afaik the SoC has another phy on it if you can get to the pins. Sep 26 16:18:58 the SoC has no phy on it Sep 26 16:18:58 zmatt im trying to build firewalls Sep 26 16:19:06 so I can't use an ethernet switch I think Sep 26 16:19:16 ogres: use a switch with vlan support? Sep 26 16:19:17 I think I need one port for input and one for output Sep 26 16:19:18 am I right? Sep 26 16:19:31 I don't think you can use firewalls with vlan? Sep 26 16:19:36 i can't believe you got these for 2 dollars a chip... Sep 26 16:19:37 why not? Sep 26 16:19:43 i dont think its secure Sep 26 16:19:45 zmatt: So I used the wrong term. It has the ability to employ more than one Ethernet link. Sep 26 16:19:45 why not? Sep 26 16:19:52 $2 a piece for these 'beaglebones' ayjay_t Sep 26 16:20:03 Ragnorok: yeah it has two ethernet ports, and no you can't use the second one Sep 26 16:20:03 yeah but still, it must be clearance right? Sep 26 16:20:08 it overlaps other stuff Sep 26 16:20:10 its cheap yea Sep 26 16:20:11 Ok. Sep 26 16:20:30 what if Sep 26 16:20:42 I got a microsd male to usb female adapter Sep 26 16:20:47 and then got a USB hub Sep 26 16:20:50 that could probably work Sep 26 16:20:52 it's a bit sad, the am335x has a dual-port gigabit ethernet switch subsystem with vlan and QoS support, but all we get on the beaglebone is one 100mbit port Sep 26 16:20:57 but then again it would conflict with boot Sep 26 16:21:13 Ha. Custom design! Sep 26 16:21:24 ogres: you're getting more hilarious by the minute Sep 26 16:21:36 lol Sep 26 16:21:42 how lol Sep 26 16:21:46 Rube Goldberg would be proud. Sep 26 16:21:58 ayjat_t here's a pic of the difference between my beaglebone and the real beaglebone Sep 26 16:21:59 http://i.imgur.com/96N4U6b.jpg Sep 26 16:22:03 https://www.youtube.com/watch?v=i3RqDp7i308 Sep 26 16:23:17 custom designs based on the am335x is pretty straightforward Sep 26 16:23:41 ogres: the "whatever these are" look like components related to usb Sep 26 16:23:54 yeah they do Sep 26 16:23:56 you're also missing the usb power switch ic Sep 26 16:24:41 ooh wheres that Sep 26 16:25:10 ah wait no that's the 32768 Hz crystal Sep 26 16:25:15 you also have no resistors Sep 26 16:25:22 So, strip off everything but the Ethernet parts and figure out how to get both Gbe working, once there's no conflicting hardware. Or just get a device that will be firewall and use these for web servers or something. Sep 26 16:25:22 although given how they've been stripping this thing you're probably also missing the usb power switch Sep 26 16:25:54 getting GigE working is easy Sep 26 16:26:12 getting GigE PHYs is a bigger pain Sep 26 16:26:24 how does anyone feel about https://www.microsoft.com/en-us/research/project/project-catapult/ Sep 26 16:26:38 Knee jerk is avoid, it's ms. Sep 26 16:26:47 gbe? Sep 26 16:27:04 Ragnorok: I can't make it through reading it because its all SELL SELL SELL Sep 26 16:27:23 And man is it CHOCK full of buzzwords. Sep 26 16:27:44 i'm pretty sure they are making up new buzzwords Sep 26 16:27:48 "The distributed architecture deploys FPGAs as an addition to each datacenter server, rather than a bolt-on isolated cluster, to create an “acceleration fabric” throughout the datacenter. This elastic reconfigurable acceleration fabric provides the flexibility to harness an individual FPGA or up to thousands of FPGAs for a single service." Sep 26 16:28:01 does TI support an elastic reconfigurable acceleration fabric? Sep 26 16:28:06 because if not, I'm not having it Sep 26 16:28:48 Today in microsoft news: HEY GUYS LOOK AT THESE FPGAS THEY DO THINGS Sep 26 16:31:53 ugh Sep 26 16:32:03 if only i could do ethernet over GPIO Sep 26 16:32:56 o shit Sep 26 16:32:58 looks like i can use this Sep 26 16:32:58 http://www.microchip.com/wwwproducts/en/en022889 Sep 26 16:33:14 however it only can support max 10mbps Sep 26 16:33:15 which is oj Sep 26 16:33:15 ok Sep 26 16:33:37 spi-ethernet Sep 26 16:33:47 10e6 b/s Sep 26 16:33:49 *shrug* Sep 26 16:33:56 http://www.instructables.com/id/Connect-an-ESP8266-to-your-RaspberryPi/ Sep 26 16:34:18 alright back to work for me Sep 26 16:34:25 i have these weird touch screens that need to be configured still Sep 26 16:34:47 You can do Ethernet over GPIO. There's lots of Wiznet chips you interface, say with a PRU. Sep 26 16:34:53 okay Sep 26 16:34:59 can you link me to some? Sep 26 16:35:03 or is there a module somewhere? Sep 26 16:35:13 thats literally ALL I really need!! Sep 26 16:35:15 an ethernet port Sep 26 16:35:20 I dont need no fuckin USB :p Sep 26 16:35:27 really, or do you need one that's supported by linux? Sep 26 16:35:29 Not from this work machine. I'd search for WizNet, maybe Arduino, since they're big over there. Sep 26 16:35:50 i need one thats supported by linux of course Sep 26 16:36:06 okay i will look Sep 26 16:36:11 It is true you'd need some way to mash the GPIO based connection into Linux. That would be an exercise left to the reader. Sep 26 16:36:12 if it works thru raspberry pi gpio Sep 26 16:36:15 it should work for us here too right? Sep 26 16:36:28 *should* Sep 26 16:36:34 Dunno. Never done that. Wouldn't. I'd get something with two or more Ethernet. Sep 26 16:36:49 or get a managed switch, configure two ports on separate vlans and a trunk port going to the bbb Sep 26 16:37:19 ALIX would be much cheaper than a managed switch with VLAN support. Sep 26 16:37:23 firewalls can't support VLAN Sep 26 16:37:32 also 10mbps will work for me. Sep 26 16:38:25 ayjay_t: what do you do after you configure weird touch screens? Bask in the glory and then move on? Sep 26 16:38:52 your statement "firewalls can't support VLAN" is nonsensical, but whatever. as people have pointed out the best solution is to get a device that's actually suitable for the task Sep 26 16:39:15 :-P I suppose you share config files somewhere in case someone else uses said screen Sep 26 16:40:43 ericxdu: naw its for work Sep 26 16:40:51 everyone wants a touch screen Sep 26 16:41:16 Oh your job? That's a neat job Sep 26 16:41:39 i've gotten too used to it Sep 26 16:41:47 i wish i worked for a bigger firm sometimes so i got to play with big toys Sep 26 16:41:51 lol Sep 26 16:41:57 and had more people with more experience to help me out Sep 26 16:42:09 i worked in a semiconductor lab for two years but it sucked so i started freelancing Sep 26 16:43:09 Oooo. Freelance hardware interfacing. Sounds spiff. Sep 26 16:44:25