**** BEGIN LOGGING AT Wed Oct 02 02:59:58 2013 Oct 02 06:24:46 Hi... Oct 02 06:25:11 is there a device tree concept in beaglebone... Oct 02 06:25:58 no windows, no gates, no concept. Oct 02 06:30:01 then what does this file implies "am335x-bone-common.dtsi".. Oct 02 06:31:45 <_av500_> yes Oct 02 06:31:50 <_av500_> the bone uses DT Oct 02 06:33:48 so my BBB now has 3.12 on it (from arch update), is this the big daddy update of kernels that TI has been pouring stuff into for the lst months? Oct 02 06:34:17 <_av500_> pouring? Oct 02 06:34:28 <_av500_> I thoight it was mre like trickle charging it Oct 02 06:34:31 ok.. maybe a bad choice of words ;) Oct 02 06:35:00 well everytime I asked for something like "will there be suspend/resume support" etc it was.. "maybe in 3.12, we're concentrating on that" Oct 02 06:35:30 or maybe it was 3.10 or something else, which is why I asked Oct 02 06:36:12 <_av500_> I dont know what TI concentrates on Oct 02 06:38:53 next product of course Oct 02 06:40:54 ^^ best answer Oct 02 06:41:02 Blueberry Rho $15 is going to be tough competition :p Oct 02 06:41:18 yes well obviously they are not putting huge amounts of energy into BBB as a board but I'm mostly interested in the support for the am335x SoC and PMIC which is a platform they will likely keep around for a while Oct 02 06:41:50 and power management is a big thing in mobile devices etc that they are targeting Oct 02 06:41:59 are they? Oct 02 06:42:07 mobile devices died in omapgeddon Oct 02 06:42:12 I didn't think the am335x was targeting mobile Oct 02 06:42:13 ok :) Oct 02 06:42:28 thought it was more vending machines / industrial machines Oct 02 06:42:40 ok so solar powered coke machines then ;) Oct 02 06:42:48 JUST GIVE ME MY SLEEPMODE! Oct 02 06:43:04 haha Oct 02 06:43:05 LPM4; Oct 02 06:43:24 I'd prefer usb hotplug working correctly Oct 02 06:43:41 well yes and uSD hotplug too Oct 02 06:43:56 though, they did deliver some drivers for that... but doesn't completely work it seems Oct 02 06:46:10 and if I can make requests for the BBB^2, 2 usb ports, HDMI 1080p60 and analog audio out. thanks! Oct 02 06:46:40 I don't personally need them but most of the people using RasPis complain about those on the BBB. Oct 02 06:46:53 Vaizki: if you want low power, go msp430 Oct 02 06:47:49 I actually have a MSP430 and BBB hooked up together so that the MSP wakes up the BBB when it's needed. but currently that means going through a whole boot cycle. Oct 02 06:53:14 Because of the Federal Goverment SHUTDOWN, All National IRC channels ARE closed. Oct 02 06:53:59 Please take your keyboards and go home, thank you. Oct 02 06:58:42 ynezz: good thing i'm european, we dont shutdown everything just because a few spoiled brats got into a sandbox fight Oct 02 07:04:44 hi Oct 02 07:07:17 www.nasa.gov made me laugh Oct 02 07:07:31 throw(toys) Oct 02 07:08:51 hi I can't write /sys/devices/bone_capemgr.*/slots. I try echo enable-uart5 > /sys/devices/bone_capemgr.*/slots. This code return -bash: echo: write error: Invalid argument Oct 02 07:09:37 check your syslog Oct 02 07:13:15 How can I check? ls /var/log btmp faillog journal lastlog old pacman.log wtmp Xorg.0.log Xorg.0.log.old Xorg.1.log Xorg.1.log.old ? Oct 02 07:15:24 ooookeeeey Oct 02 07:16:35 try journalctl -n 100 for example :) Oct 02 07:19:22 thank u Vaizki please look http://paste.laravel.com/VUa ? Oct 02 07:27:25 interesting. can't really help you.. looks like your kernel is missing stuff referenced by the device tree. Oct 02 07:30:45 okey thank you so much. Best regards. Oct 02 07:36:29 KotH: well, It doesn't necessarily means, that it couldn't happen here also :p Oct 02 07:37:02 but you're safe in your .ch sandbox Oct 02 08:35:18 hi! Oct 02 10:02:21 hey! Oct 02 10:02:51 does anybody know if there is a driver for BBB CANBus under SYS/BIOS? Oct 02 11:53:54 hi Oct 02 12:23:07 I'm playing a bit with a Li-Po powered BB trying to connect a 5V signal to VDD_5V pins. My aim is to see if a 5V signal provided here could let me use the USB Host Oct 02 12:23:28 there's a discussion about the BBB here https://groups.google.com/forum/#!searchin/beagleboard/5V$20boost$20converter/beagleboard/6hQ7voyk_Xw/z4ssmKMBkkQJ Oct 02 12:24:13 as far as i understand it should work however it doesn't Oct 02 12:24:27 the 5V at VDD_5V pins doesn't seem to be enough Oct 02 12:24:41 anybody with experience on this could share some thoughts? Oct 02 12:29:13 vmayoral, take a picture of the connections Oct 02 12:30:16 Hi! I'm desperately searching for a *1GByte* RAM + $some NAND PoP chip for DM3730. I know such chips must exist, since Nokia N9 comes with 1GB of RAM. Any hints? Oct 02 12:30:38 digi-key Oct 02 12:30:44 haha Oct 02 12:30:47 :P Oct 02 12:30:59 part number? Oct 02 12:31:17 ill take a look at the stock Oct 02 12:31:24 part # Oct 02 12:32:03 there "don't exist" chips with >512MB RAM Oct 02 12:32:34 MT41K256M16HA-125 Oct 02 12:32:41 hi! I am trying to load my BBB rootfs from my Ubuntu workstation but it does not work. I have added the following to my uEnv.txt: mmcargs=setenv bootargs console=${console} ${optargs} root=/dev/nfs nfsroot=192.168.0.23:${rootpath},${nfsopts} rw ip=dhcp Oct 02 12:34:13 bmatusiak_: thanks, but aiui this is a 4Gb (aka 512MB) chip Oct 02 12:34:26 error is: [ 3.937390] Kernel panic - not syncing: No init found. Oct 02 12:34:50 Anguel: are you sure your rootfs is good ? Oct 02 12:35:08 not 100% but I also see [ 0.779509] net eth0: phy 4a101000.mdio:01 not found on slave 1 Oct 02 12:35:13 is this normal? Oct 02 12:35:33 DocScrutinizer05, you have to find something compareable to that Oct 02 12:35:48 well, that's what I'm trying to do ;-) Oct 02 12:36:20 I thought I could just overwrite the mmcargs=...... line through my uEnv.txt. Is this ok? Oct 02 12:37:42 I already wondered if we could design a tiny PCB to solder on top of DM3730, which in turn holds 2 of those 512MB RAM chips Oct 02 12:38:04 DocScrutinizer05 MT41K512M16 Oct 02 12:38:17 search that on digikey Oct 02 12:39:33 already did, thanks Oct 02 12:39:39 seems it lacks NAND Oct 02 12:40:14 I don't need much NAND, but a few MB for booting would be really nice Oct 02 12:43:27 aiui the DM3730CBP package has NAND balls on both top and bottom, so I might go for that 8Gb RAM PoP plus "discrete" NAND, right? Oct 02 12:44:01 http://processors.wiki.ti.com/index.php/Memory_Vendor_Selection_Guide Oct 02 12:44:38 but I *really* would prefer a PoP that comes with 1GB RAM *plus* some NAND Oct 02 12:44:39 bmatusiak_: here it is https://dl.dropboxusercontent.com/u/10096191/20131002_144159.jpg Oct 02 12:47:41 you behing a 5v regulator? Oct 02 12:48:19 yes, the 5V signal is being provided by this booster https://www.sparkfun.com/products/11231 Oct 02 12:48:50 send me a link to your SRM Oct 02 12:49:35 i think i see the problem... Oct 02 12:49:43 http://beagleboard.org/static/beaglebone/latest/Docs/Hardware/BONE_SRM.pdf Oct 02 12:52:16 vmayoral, p9 pin 7 and 8 is your SYS_5V Oct 02 12:53:29 looks like you connected to VDD_5V Oct 02 12:54:16 yes, you're right, didn't notice that Oct 02 12:54:23 weird Oct 02 12:54:36 the schematics say it's the other way around Oct 02 12:55:06 mmm Oct 02 12:55:10 not all right Oct 02 12:55:11 my mistake Oct 02 12:55:23 so it working now? Oct 02 12:55:23 you're right, 7 and 8 are SYS_5V Oct 02 12:56:39 they're connected for now to VDD_5V (5 and 6) and it doesn't work and if I do connect 5V to SYS_5V... Oct 02 12:56:59 that is wierd Oct 02 12:57:25 doesn't work either Oct 02 12:57:45 i even tried putting 5V in both SYS_5V and VDD_5V Oct 02 12:58:04 nothing still Oct 02 12:59:13 it seems this works on the BBB (https://groups.google.com/forum/#!searchin/beagleboard/5V$20boost$20converter/beagleboard/6hQ7voyk_Xw/z4ssmKMBkkQJ), i'll try it tonight once i'm home but as far as i understand it should also work on the BB Oct 02 13:03:43 vvu|Log: thanks, it was indeed the rootfs. Now I get to the login prompt but then -> beaglebone login: [ 120.199181] INFO: task systemd-journal:85 blocked for more than 60 seconds. Oct 02 13:04:38 and then a kernel panic Oct 02 13:04:48 anguel hm maybee network-manager waiting for ip adress Oct 02 13:05:28 Anguel so there is maybe a correaltion with your kernel message Oct 02 13:05:32 bmatusiak_: don't know if it's related but: SYS_5V=3.65 V (battery level), VDD_5V=4.65V (boost converter output) Oct 02 13:06:16 seems low Oct 02 13:06:43 you mean the output from the boost converter? Oct 02 13:07:05 woglinde: aha, you mean connman is trying to get a new ip and that hangs the nfs? Oct 02 13:07:12 vmayoral, no sys_5v Oct 02 13:07:42 anguel hm Oct 02 13:08:01 did not regonize you are using nfsroot Oct 02 13:08:43 bmatusiak_: well it's adopting the battery voltage level. Maybe that's the problem? Oct 02 13:09:17 yes, I added the following to my uEnv.txt to make BBB load rootfs from nfs: mmcargs=setenv bootargs console=${console} ${optargs} root=/dev/nfs nfsroot=192.168.0.23:${rootpath},${nfsopts} rw ip=dhcp Oct 02 13:10:34 seems to boot fine but hangs at login prompt and spits out that [ 120.199181] INFO: task systemd-journal:85 blocked for more than 60 seconds. Oct 02 13:12:07 ynezz: nope, it wont easily happen Oct 02 13:12:26 ynezz: for one thing, we have more than two parties, hence one party cannot force a discussion to fail Oct 02 13:12:57 woglinde: removed connman.service from rootfs, can login now :) Oct 02 13:13:10 vmayoral, VDD_5V is the only pin safe to supply voltage too Oct 02 13:13:13 ynezz: for another, the infrastructure is independent of the parliament. as long as the parliament is working on some laws/regulations, the old ones are still in place and thus work continues Oct 02 13:13:55 anguel hm okay Oct 02 13:13:57 ;) Oct 02 13:14:12 vmayoral, The SYS_5V rail is the main rail for the regulators on the main board Oct 02 13:15:57 woglinde: yes, thank you for the hint, I remembered this link: http://wiki.beyondlogic.org/index.php/BeagleBoneBlack_Building_Kernel#NFS_boot_failures Oct 02 13:16:19 woglinde: not very elegant but works... Oct 02 13:17:19 bmatusiak_: i see, thanks for the clarification. I'll read the TPS datasheet more carefully and see if i find the reason why SYS_5V is at the battery level and not at 5V Oct 02 13:19:18 vmayoral, also i think you need a 100k resistor to BAT_TEMP.. (confident but not really sure) Oct 02 13:20:10 bmatusiak_: yeap it's needed to simulate the battery temp. sensor but i believe that's just necessary for charging Oct 02 13:20:36 you can power the board from a Li-Po cell without the 100k resistor :) Oct 02 13:25:17 anybody ever heard of a PoP chip called "110J1H0K" (combo memory on top of OMAP 8Gb SDRAM 4Gb NAND M3) Oct 02 13:25:54 http://wstaw.org/m/2013/10/02/plasma-desktopFk3743.png Oct 02 13:26:56 ...or anything compatible to that chip Oct 02 13:28:31 don't tell me you're trying to design that franken900 based on some OMAP? Oct 02 13:28:50 sure Oct 02 13:29:03 on what else would we base it? Oct 02 13:29:15 hi dm8tbr :-) Oct 02 13:29:29 dunno, something a bit more future proof? Oct 02 13:29:47 the problem: anything more future proof is less maemo proof Oct 02 13:29:56 particularly fremantle proof Oct 02 13:30:26 major design goal: binary compatibility Oct 02 13:30:28 although if price is not an issue, try to get OMAP5 chips, the design might even be open for the UEVM Oct 02 13:30:55 so the chip need so have broken Thumb2 too? Oct 02 13:31:06 OMAP5 still hard to source, and I'm afraid it wouldn't play too nice with fremantle Oct 02 13:31:27 HAHA, no, we will happily use working thumb2 Oct 02 13:32:10 compatible improvements welcome Oct 02 13:32:12 nuhu, good luck then. best you'll get is a 4470 then I guess. Oct 02 13:32:22 no major design changes though Oct 02 13:33:07 the whole project based on existing and debigged&working GTA04 board Oct 02 13:33:15 debugged* Oct 02 13:33:41 and that's DM3730CBPD100 Oct 02 13:34:14 we just try to squeeze as much as possible from it Oct 02 13:34:41 so: 1TGB RAM + >=256MB NAND PoP would be a dream Oct 02 13:34:46 -T Oct 02 13:35:30 we already pondered NAND-less design, but that for sure breaks lots of properties relevant for fremantle Oct 02 13:36:11 and users mad for RAM, we probably won't get away with "only" 512MB Oct 02 13:36:24 seems 1GB RAM is a key feature Oct 02 13:36:56 bmatusiak_: solved! After checking that DC_IN was 5V it got really weird so i decided to review the the booster datasheet. That was it. It seems that i was not enough current. I'll dig a bit more though. Thanks! Oct 02 13:37:08 dm8tbr: this is a ultra-low funded project, no way to go for leete new stuff Oct 02 13:38:00 dm8tbr: quite unusual for me, but this time I have to pick components not by specs but by availability and compatibility to an OS Oct 02 13:38:28 dm8tbr: it's not like I was too dull to figure a really leete modern design Oct 02 13:38:28 vmayoral, YAY! i knew it was low Oct 02 13:39:26 dm8tbr: omap4 already doesn't fit into available volume, mechanically Oct 02 13:40:42 (plus I don't really like OMAP4 after all the secret SiErr I heard of, during my time at ST-E) Oct 02 13:41:34 "WONTFIX in OMAP$, scheduled fix OMAP5" X-P Oct 02 13:41:42 s/$/4/ Oct 02 13:50:56 DocScrutinizer05: same went for omap3, lots of stuff there never fixed, but addressed in omap4. it's always a n+1 game Oct 02 13:51:31 yeah, but the SiErr in OMAP3 were worked around in N900 design and fremantle Oct 02 13:51:47 while the SiErr in OMAP4 are... terribly nasty Oct 02 13:52:05 and basically we found no real workaround for them Oct 02 13:52:46 then you'd need to go back to a 3430 design and not 3630 Oct 02 13:52:51 when hw doesn't behave according to a interface protocol, all you can do is break the protocol specs, and that's not really an option Oct 02 13:53:06 as 3630 had quite different errata Oct 02 13:53:16 yeah, well Oct 02 13:53:26 I hope it won't bite us Oct 02 13:54:03 at which point, latest, your whole exercise starts to look really weird, as there is essentially no gain in it Oct 02 13:54:04 anyway, omap4 is no option, omap5 not even a possibility Oct 02 13:54:39 I'd rather be interested in finding that phantom chip with 1GB of RAM Oct 02 13:55:12 ask the usual vendors. micron etc Oct 02 13:55:19 dm8tbr: the gain is that we get new N900 Oct 02 13:55:30 everything on top is a bonus Oct 02 13:55:48 that can be had cheaper off ebay I'd argue Oct 02 13:55:50 this is NOT about evolution Oct 02 13:55:53 new N900? Oct 02 13:55:56 where? Oct 02 13:56:23 with android support? please? Oct 02 13:56:50 dm8tbr: we're not arguing the rationale for doing that project. seems there's a market for it, and I personally would like such a device Oct 02 13:57:10 on a 3430 I'd not try anything beyond Gingerbread Oct 02 13:57:44 GTA04 already exists, we slightly modify it to cram it into a case that'S actually available. End of rationale Oct 02 13:58:10 honestly: we don't support anything, we build hw Oct 02 13:58:45 honestly: we don't have a sw team, we couldn't afford any Oct 02 13:58:49 as long as that hw has firmware and drivers that properly support linux I'm game Oct 02 13:58:59 http://talk.maemo.org/showthread.php?t=91142 Oct 02 13:59:04 DocScrutinizer05: so, you are doing over the wall engineering? Oct 02 13:59:06 <_av500_> OMAP3 based Android stuff does exist Oct 02 13:59:17 <_av500_> its just not fun any more Oct 02 13:59:42 <_av500_> but sure, your xterm witl blink a cursor happily Oct 02 13:59:54 KotH: no, we do open engineering Oct 02 14:00:16 DocScrutinizer05: wow, that looks like a cool project Oct 02 14:00:59 DocScrutinizer05: in my experience, if a hw team has no sw guy to look over the design, it will usually contain lots of quirks in sw that could have been very simple to solve in hw Oct 02 14:01:33 <_av500_> dm8tbr: somehow I am am reminded of how the PMA4 was Zaurus binary compatible Oct 02 14:01:37 <_av500_> +00 Oct 02 14:02:01 _av500_: ugh... Oct 02 14:02:10 port-a-potty galore! Oct 02 14:02:24 <_av500_> ipk hacking > xda hacking Oct 02 14:02:39 yeah, but not much more Oct 02 14:03:11 openPMA was essentially a custom rom before it was cool. that's how hipster we were ;) Oct 02 14:03:32 although we did fix that nasty HDD bug... Oct 02 14:05:03 KotH: that guy will be me Oct 02 14:07:21 KotH: actually that's my main professional competence: ambassador between hw and sw Oct 02 14:08:46 don't waste money on complicated chips when same thing is easy to implement in sw. Don't miss to take care of hw bits that significantly simplify sw development and system performance Oct 02 14:09:44 like e.g. IRQ driven design, for the latter Oct 02 14:27:33 anybody here know how to create a device tree? Oct 02 14:30:43 take a bunch of devices and nail them together in the shape of a tree Oct 02 14:32:22 a little less sarcasm Oct 02 14:32:31 fine, use device seeds Oct 02 14:32:49 there we go Oct 02 14:32:50 bmatusiak_: i've done some stuff with them, i'd recommend you to take a look at https://docs.google.com/document/d/17P54kZkZO_-JtTjrFuVz-Cp_RMMg7GB_8W9JK9sLKfA/edit?hl=en&forcehl=1#heading=h.j4ega7pcz5c Oct 02 14:33:17 bmatusiak_: just read the docs, use the files that are around as examples and you should do that yourself Oct 02 14:33:50 bmatusiak_: if you have a little experience with linux, you'll be ready in an hour or two, with no experience it takes about half a day Oct 02 14:34:00 bmatusiak_: but you have to read them _yourself_ Oct 02 14:34:46 panto: didnt your DT doc go to some more prominent place on elinux or so? Oct 02 14:34:53 understood i have a dts already i dont know if its currectly setup Oct 02 14:35:05 or if im missing anything Oct 02 14:35:23 if you have a DT, then it means it's setup Oct 02 14:35:24 bmatusiak_: take a look at Jason Kridner's example then https://github.com/vmayoral/validation-scripts/tree/master/test-capemgr Oct 02 14:35:24 just find our Oct 02 14:35:37 out even Oct 02 14:35:45 bmatusiak_: you have to read the DT and see whether it contains the stuff you need or not Oct 02 14:35:54 bmatusiak_: and for that you need to understand its structure Oct 02 14:38:06 hm, now that I have mounted my rootfs via nfs, how can I access my uEnv.txt that is on my eMMC? I tried mounting mmcblk0p2 but I don't see any files in its /media/BEAGLEBONE directory Oct 02 14:40:42 maybe it's not the right partition? Oct 02 14:40:47 or the mount failed Oct 02 14:41:01 check the output of `mount` (without parameters) Oct 02 14:42:24 KotH: yes, solved it, the uEnv.txt is on mmcblk0p1 Oct 02 14:42:31 thanks Oct 02 14:44:28 hi crofton Oct 02 14:44:52 hm oh only reconnect Oct 02 14:46:30 yeah Oct 02 14:53:06 hi woglinde Oct 02 14:59:09 bmatusiak_: aiui from http://www.micron.com/~/media/Documents/Products/Data%2520Sheet/DRAM/DDR3L-RS_8Gb_x16_2CS_TwinDie_V80A.pdf%2BMT41K512 this chip is DDR3, while DM3730 needs LPDDR1 ( http://processors.wiki.ti.com/index.php/Memory_Vendor_Selection_Guide ). Would that chip be compatible nevertheless to DM3730? Oct 02 15:00:23 not too exactly sure Oct 02 15:00:46 thanks Oct 02 15:07:27 I have built a qt-embedded sdk, however the resulting rootfs under /sysroots/beaglebone seems incomplete. is this meant to be merged with the systemd rootfs? Oct 02 15:11:18 o_0 Oct 02 15:11:25 you only have one rootfs Oct 02 15:11:39 and systemd is just a daemon, not a system definition Oct 02 15:11:47 and definitly nothing that provides a rootfs Oct 02 15:12:04 woglinde: btw: bist du ende monat in edinburgh? Oct 02 15:17:24 KotH: sorry, I meant this rootfs: Angstrom-systemd-image-eglibc-ipk-v2012.12-beaglebone-2013.09.12.rootfs.tar.xz Oct 02 15:18:13 Anguel: uhmm... Oct 02 15:18:23 Anguel: i dont know whether this is the right approach Oct 02 15:19:06 Anguel: qt apps are c++ apps, this means you need to have the libs with the same version installe, that have been compiled with the same g++ version you have used for your application. otherwise you cannot ensure that your application is working correctly Oct 02 15:19:43 Anguel: i.e. you want to integrate your application into the build system of your meta-distro you are using, be it angstr�m, buildroot, debian or whatever you want to use Oct 02 15:19:58 Anguel: i usually recomend buildroot, because it's the most simple to get up and running Oct 02 15:20:50 KotH: I built the Qt SDK with bitbake exactly as described here: http://www.angstrom-distribution.org/toolchains/ but this gives me an incomplete rootfs in /tmp-angstrom_v2012_12-eglibc/sysroots/beaglebone Oct 02 15:21:19 Anguel: sorry, i dont know anything about angstr�m Oct 02 15:21:56 koth no Oct 02 15:22:05 I am learning Oct 02 15:22:19 KotH: vielen dank trotzdem :) Oct 02 15:23:10 kein problem :) Oct 02 15:25:17 btw, where is koen? :) Oct 02 15:25:26 <_av500_> honeymoon Oct 02 15:26:45 omg, we are lost - did they all get married? zubair who is doing the adc drivers is also doing the same Oct 02 15:27:17 <_av500_> yes Oct 02 15:27:21 <_av500_> group mariage Oct 02 15:27:28 <_av500_> +r Oct 02 15:27:39 <_av500_> Reverend Moon presiding Oct 02 15:27:45 :) Oct 02 15:29:01 and the irc logfiles are also not working Oct 02 15:31:24 KotH: what linux are you using? Oct 02 15:31:44 the one from torvalds :) Oct 02 15:32:05 Anguel: on embedded systems? buildroot Oct 02 15:32:57 KotH: seems to be nicer than openembedded Oct 02 15:34:00 Anguel: it is less fancy Oct 02 15:34:14 Anguel: OE does a lot of stuff for you, while you do that by hand in buildroot Oct 02 15:34:25 Anguel: but at the same time, it makes buildroot easier to modify for your needs Oct 02 15:34:39 Anguel: both have their target audiences Oct 02 15:34:57 well, oe throws results in a temp dir and you can digg through it an guess what is where and why Oct 02 15:35:15 hmm? Oct 02 15:35:18 sounds counter intuitive Oct 02 15:35:24 i'm sure it ends up somehwere more prominent Oct 02 15:36:37 not from what I have seen or read so far Oct 02 15:37:55 maybe koen can give some hints when he is back Oct 02 15:38:05 dont expect him to be back soon Oct 02 15:38:54 soon? months? Oct 02 15:39:33 years Oct 02 15:40:21 looks like there are no other angstrom / oe experts around here Oct 02 15:40:37 or they don't talk much :) Oct 02 15:40:49 there are very few experts here, in general :) Oct 02 15:40:51 OE is pretty cool but I still like ptxdist better Oct 02 15:41:04 At least for the moment Oct 02 15:41:26 Its just so much easier to tell WTF its doing Oct 02 15:41:53 georgem: I am new to everything so I thought it would be the best to stick with Angstrom and OE :/ Oct 02 15:42:07 I am currently thinking about using the BeagleBone black along with some accelorometers, are there any problems with the resolution of the data? I would be using the triple axis accelerometer Oct 02 15:42:24 Anguel: Yeah. Its probably best to stick with it just because OE is becoming quite mainstream Oct 02 15:43:27 Anguel: I was using ptxdist when OE was still in diapers so I've been reluctant to transition Oct 02 15:43:45 georgem: it actually seems to work quite well, has not refused building any of the recipes I tried, just don't know what exactly it produces and how you can combine that with other recipes Oct 02 15:44:13 scottk201: bbb does not care about your data Oct 02 15:44:24 scottk201: you can use whatever resolution you want to use Oct 02 15:44:48 Anguel: you are inspecting the beaglebone sysroot used during build :/ The SDK should be under /deploy ;) Oct 02 15:44:58 Anguel: It does work well. The biggest problem I have is trying to figure out exactly what its doing. The chained together repositories / recipes make for quite a mess to untangle Oct 02 15:46:22 KotH: the accelerometers would output 12 bits of resolution, I was informed that BBB only has a 10 bit input Oct 02 15:47:07 For instance one time after building Angstrom successfully I thought. Let me see if I can find where the linux kernel version being used is specified. I think it took me like 2 hours to find it. In ptxdist its set in a text file in the root directory of your workspace. Oct 02 15:47:38 And you can change it with Kconfig Oct 02 15:48:17 And if I had to go find it again it would probably still take me an hour to find Oct 02 15:49:03 ant_work: you mean this one /usr/local/oecore-x86_64/sysroots/armv7a-vfp-neon-angstrom-linux-gnueabi ? Oct 02 15:49:26 scottk201: this information is wrong Oct 02 15:50:01 scottk201: your accelerometer uses probably i2c or spi, the bbb does not care what kind of data you have, it doesnt even care how long it is, you just have to write the software to handle it correctly Oct 02 15:50:13 Anguel: I mean the image is under /tmp-angstrom_v2012_12-eglibc/deploy Oct 02 15:50:40 scottk201: for more information read the TRM on spi and i2c and the linux kernel documentation on i2c and spi Oct 02 15:50:49 KotH: would you say writing the software would be too difficult for a novice? Oct 02 15:51:02 scottk201: depends on what kind of novice you are Oct 02 15:51:24 ant_work: I am looking for the rootfs Oct 02 15:52:24 you are confusing things Oct 02 15:52:29 on ething is the sdk Oct 02 15:52:36 scottk201: i know guys who've been writing c code for >15 years and still cannot write a ring buffer without dozens of bugs. on the other hand, i know guys who didnt know programming at all and started writing video codecs, and were successful Oct 02 15:52:37 other are the 'images' = rootfs Oct 02 15:53:39 KotH: that is somewhat helpful thanks for the help. Leaving now Oct 02 15:53:48 scottk201: have fun coding :) Oct 02 15:54:44 ant_work: ok, I am looking for the files :) Oct 02 15:55:26 ant_work: to access via nfs Oct 02 16:02:17 ant_work: so the simple question is: where do I find the directories that I need to copy to my exported NFS directory Oct 02 16:07:26 Anguel: sry I have to go. bb later. Check the artifacts under /deploy Oct 02 16:20:17 I'm trying to find a way to access the BeagleBone's U Oct 02 16:20:32 *USB port. Can anyone point me in the right direction of how that might be done? Oct 02 16:20:57 * mru points at the short side of the bord Oct 02 16:21:04 *board Oct 02 16:22:34 KotH: ot question - I see you're +O in accesslist (sth that freenode deprecates) but not +o in the channel. How that? Oct 02 16:25:54 KotH: tip: you can drop own +O on accesslist, even when you're not +f Oct 02 16:27:14 don't mess it up and accidentally drop +o on accesslist though! ;-) Oct 02 16:33:27 DocScrutinizer05: i dont like to be +o when not needed Oct 02 16:35:17 DocScrutinizer05: and if you had a look at the access list, you would have noticed that i cannot change it Oct 02 16:36:40 I had a look at access list, that's why I wondered why you are not +o in chan despite +O (auto-op) in accesslist. And yes, you *can* drop own privileges on access list, always Oct 02 16:37:41 actually freedone suggests as well _not_ to use +O Oct 02 16:37:49 freenode even Oct 02 16:39:40 /cs help flags Oct 02 16:40:16 >>+o grants +O<<< Oct 02 16:40:55 means as long as you're +o on accesslist, you can set and unset +O on your own Oct 02 16:44:21 /msg ChanServ FLAGS #beagle KotH -O Oct 02 16:51:02 DocScrutinizer05: You are not authorized to execute this command. Oct 02 16:51:10 DocScrutinizer05: read the docu on the access list commands again Oct 02 16:51:52 orly? I already wondered why it didn't work ;-P Nah, I posted it for your convenience Oct 02 16:52:25 since _you_ are allowed to execute that command Oct 02 16:53:28 * KotH nods Oct 02 16:53:43 DocScrutinizer05: just for the record, i'm not new to irc Oct 02 16:53:43 [2013-10-02 18:39:11] [Notice] -ChanServ- If you are not a founder, you may only manipulate flags you Oct 02 16:53:45 [2013-10-02 18:39:11] [Notice] -ChanServ- have yourself, and may not edit users that have flags you Oct 02 16:53:46 [2013-10-02 18:39:11] [Notice] -ChanServ- don't have. For this purpose, +v grants +V, +o grants +O Oct 02 16:53:48 [2013-10-02 18:39:11] [Notice] -ChanServ- and +r grants +b. Oct 02 16:53:51 DocScrutinizer05: neither am i new to freenodes chanserv Oct 02 16:54:03 * DocScrutinizer05 neither Oct 02 16:56:25 I just wondered why you have all +O on accesslist while obviously you don't want to be always +o in channel Oct 02 16:57:03 (which is what freenode suggests) Oct 02 16:59:50 because we happend to end up with +O Oct 02 17:00:09 it's like why we all speak english instead of japanese Oct 02 17:01:46 DocScrutinizer05: this channel doesn't adhere to freenode convention in that regard Oct 02 17:02:44 ophat is used to mark on-duty staff Oct 02 17:02:50 wait Oct 02 17:02:53 fuck Oct 02 17:02:56 I'm in the wrong channel Oct 02 17:03:03 * KotH nods Oct 02 17:03:15 now please leave before you mess up the n00bs ;-) Oct 02 17:03:34 lol Oct 02 17:07:01 nevermind, it's just I been worried about your comfort, not to deop yourself each time you join. Not that it matters in any regard Oct 02 17:23:43 DocScrutinizer05: uhmm... unless freenode kicks me out, i'm online Oct 02 17:23:54 and dont worry about my confort Oct 02 17:24:03 things like this are of no concern Oct 02 17:24:27 customers or suppliers on the other hand... i wouldnt mind if you could get rid of some of those Oct 02 17:24:50 hehe Oct 02 17:58:55 bmatusiak_: aiui the MT41K512 isn't available in 168p CBP PoP package? Oct 02 18:00:39 mabe think about building a cape to add another 2gb? Oct 02 18:00:49 o wait Oct 02 18:00:50 memory Oct 02 18:01:05 yea and memory Oct 02 18:01:23 that would be inpossible Oct 02 18:05:06 I already pondered building a piggyback PCB that goes in between OMAP and PoP and extends the memory bus to place another 512MB RAM on the 2nd CS signal. But that won't filt into the case - too high :-/ Oct 02 18:06:15 I'm also worried about reflow for such contraption Oct 02 18:29:12 actually it seems there *might* be a chip that offers 1GByte RAM on PoP-168-vfbga: NT6TL256T32AS - I'm still wondering if it really meets the DM3730 memory bus specs Oct 02 18:29:52 <_av500_> check the TRM Oct 02 18:29:58 <_av500_> and try it Oct 02 18:30:04 <_av500_> we found that combination to work best Oct 02 18:30:06 lol@#2 Oct 02 18:30:20 yeah, sure Oct 02 18:30:42 check TRM is a tedious work Oct 02 18:31:02 I know there's no way around it in the end Oct 02 18:31:28 except some other users hollering "WFM" Oct 02 18:31:55 <_av500_> seriously, we found that we have to qualify all the RAM we used Oct 02 18:32:07 you always have to Oct 02 18:32:25 wonderful world of component sourcing Oct 02 18:32:49 a breeze Oct 02 18:33:00 :-S Oct 02 18:33:32 couldn't TI put a few more chips on their QML Oct 02 18:33:33 <_av500_> also be prepared for any POP to go away overnight Oct 02 18:33:39 <_av500_> because the vendor got bored Oct 02 18:33:48 yeah, been there seen that Oct 02 18:33:51 <_av500_> or you did not order 100k Oct 02 18:34:01 *cough* Oct 02 18:34:05 -k Oct 02 18:34:44 or you order 500k and get the chip custom made like that friggin phantom chip Nokia using in their N9 Oct 02 18:35:28 <_av500_> or that Oct 02 18:35:30 110J1H0K Oct 02 18:35:44 nobody ever heard of that chip Oct 02 18:35:52 <_av500_> but hey, as a dead horse beater, you knew what you were up against... Oct 02 18:35:54 nevertheless it exists Oct 02 18:36:03 yep :-P Oct 02 18:36:54 I seriously pondered buying broken/used N9 and recovering chips from them ;-D Oct 02 18:37:26 for the batch volume we aim at, it would be actually kinda feasible Oct 02 18:37:41 contact nokia and offer to buy their surplus ;) Oct 02 18:37:46 but then, reball those chips, eeew Oct 02 18:38:01 dm8tbr: hey, that's actually no bad idea Oct 02 18:38:28 (for the chips that must still rot somewhere in the fab) Oct 02 18:38:40 <_av500_> scrapped long ago Oct 02 18:38:44 prolly Oct 02 18:38:48 <_av500_> somebody ebay'ed the box Oct 02 18:38:53 :D Oct 02 18:38:56 storage shelf space is expensive Oct 02 18:39:57 damn it, I'm temped to say "screw those 1GB of RAM demands" Oct 02 18:40:25 "512MB it is, take it or leave it" Oct 02 18:41:31 <_av500_> n900 had 1GB? Oct 02 18:41:42 but hey, 1GB RAM PoP NT6TL256T32AS plus discrete NAND sounds feasible - if the RAM works Oct 02 18:41:56 _av500_: had 256MB X-P Oct 02 18:42:12 _av500_: but you know users Oct 02 18:42:32 "WUT???!? *ONLY* 512MB RAM??" Oct 02 18:42:34 <_av500_> omap3 can drive 1GB? Oct 02 18:42:37 yep Oct 02 18:42:42 <_av500_> you sure? Oct 02 18:42:48 <_av500_> as in somebody tested it ever? Oct 02 18:42:50