**** BEGIN LOGGING AT Fri Oct 26 03:00:01 2012 Oct 26 04:35:43 does the pin configuration(in/out pull up/down) matter for a device if the pins are in the right mode? Oct 26 05:34:32 JViz: right mode? :) Oct 26 05:34:52 you can totally screw youself if you don't want you are doing Oct 26 06:13:10 can we please shoot people who insist on sending large attachments? Oct 26 06:14:21 mranostay: pin mux mode 0-7 Oct 26 06:18:08 ds2: i always down for a good shooting Oct 26 06:20:10 is it sane to want patchs pushed to git to a develop branch to get apply automagically to other xyz-vendor-branch? Oct 26 06:20:15 * mranostay starts the poll Oct 26 06:20:49 * mranostay thinks his must be some retarded thing perforce or clearcase does Oct 26 06:33:21 hi mdp Oct 26 06:42:54 <_av500_> gm Oct 26 06:45:13 _av500_: how is my favorite troll today? Oct 26 06:46:26 _av500_: what day you get in BCN? Oct 26 06:47:07 <_av500_> ? Oct 26 06:47:18 _av500_: for E-ELC Oct 26 06:48:19 <_av500_> sunday evening Oct 26 10:36:56 quiet friday Oct 26 11:37:47 mranostay, PMIC is my favoritest of all time, but close behind it is the U8 regulator on bone! Oct 26 12:37:36 mru: maybe servers will get energy ratings like light bulbs or fridges? Oct 26 12:58:53 Try to cross compile "top" (out of procps) - but fail. make lib64=lib CC=arm-none-linux-gnueabi-gcc gives: top.c:28: fatal error: curses.h: No such file or directory while adding the dir with make CC=arm-none-linux-gnueabi-gcc CPPFLAGS=-I/usr/include cc1: warning: include location "/usr/include" is unsafe for cross-compilation top.c:3385: error: impossible constraint in 'asm' I have no clue how to solve this issue. Can you help? Wh Oct 26 13:00:53 you are cross compiling Oct 26 13:01:00 so you cannot pass your host /usr/include Oct 26 13:02:01 you need a cross compile ncurses lib Oct 26 13:02:11 and that would come with the proper headers too Oct 26 13:06:48 I have the source for the libncurse on my PC. So they arn't compiled together? Oct 26 13:07:17 With source I mean the dev package in ubuntu Oct 26 13:08:49 So if I need libs to cross compile - what is the best way to generate them? I could imagine that some tools need some libs - that can become very painful ... Oct 26 13:13:35 Arno__, yes, it's painful :) Oct 26 13:13:42 if you're on debian, xapt is helpful Oct 26 13:13:51 multilib was supposed to fix this Oct 26 13:14:05 sort of Oct 26 13:14:30 Arno__, dev packages usually are just for the host arch, you need dev packages for the target arch Oct 26 13:15:08 Arno__: alternative is to not cross-compile and compile while on the target(or get the libs from the target and link against that) Oct 26 13:19:34 ok, just for the blind of us - just tried compiled the sources of ncurses with / ./configure CC=arm-none-linux-gnueabi-gcc --host=arm / make -> Oct 26 13:20:40 Arno__: normally you would use some kind of (cross) build system Oct 26 13:20:44 like angstrom/OE Oct 26 13:20:47 or buildroot Oct 26 13:20:49 or others Oct 26 13:21:01 of course you can compile all needed stuff by hand Oct 26 13:21:49 Oh I see - it will use wrong linker - thats wyh this fails : /usr/bin/g++ -o demo ../objects/demo.o -L.. ... Oct 26 13:22:48 What I have is codesourcery - so compiler and standard libs are there Oct 26 13:24:51 What do I have to set while configure, that not only compiler, but also the right linker or g++ is used ? Oct 26 13:26:56 arno thats why we have in angstroem external_toolchain stuff, its not perfect yet but should work Oct 26 13:27:09 and even an internal toolchain Oct 26 13:27:19 yes sure Oct 26 13:27:36 *g* but he talked about codesourcery Oct 26 13:29:33 harry potter toolchains Oct 26 13:29:50 ? Oct 26 13:30:25 koen *g* Oct 26 13:31:05 :( Oct 26 13:34:47 till later Oct 26 13:35:15 bye Oct 26 13:39:37 mmm...internal trees Oct 26 13:44:57 ./configure --host=arm-none-linux-gnueabi --target=arm-non-linux-gnueabi ?? Oct 26 13:57:19 Arno__, host is your host arch, target is your target arch, build usually is the same as host arch unless you're building a canadian cross compiler Oct 26 13:57:48 or building a target arch compiler Oct 26 14:03:45 What would be the right setting if I compile on a PC for embedded? Just the target setting? Oct 26 14:05:05 Arno__: google "cross compiling" Oct 26 14:05:13 you are not the first one to attempt that Oct 26 14:06:30 you are right Oct 26 14:45:15 re Oct 26 14:51:22 er Oct 26 14:58:00 be Oct 26 15:51:38 bradfa: ping Oct 26 15:52:13 mdp, pong Oct 26 15:52:38 hi bradfa, I was using your pin tables to look over something on my 6502 design.. Oct 26 15:53:00 and I noticed that pru mode5/6 settings are missing Oct 26 15:53:10 if that a "patches are welcome" situation? Oct 26 15:53:13 s/if/is/ Oct 26 15:53:20 yes Oct 26 15:53:51 ok, can I just complain the the community hasn't fixed this yet instead? Oct 26 15:54:03 "The Community(tm)" Oct 26 15:54:15 mdp, you can, but I'm pretty sure that'll just spark comments from the peanut gallery :) Oct 26 15:54:22 ;) Oct 26 15:55:16 I had to rewire a couple things and the nature of the pru pin routing and general availability was driving me batshit crazy Oct 26 15:55:50 I'll put it on my list Oct 26 15:56:01 the ever-growing list Oct 26 15:57:48 bradfa, I owe you a retest too Oct 26 15:58:20 mdp, for uarts? Oct 26 15:58:27 yeah Oct 26 15:58:42 v4 went to the list, also in my github if that's easier for you Oct 26 15:58:46 I'm already "deployed" with the v2/v3 or whatever series here ;) Oct 26 15:59:11 there's a v5 in github but I'm not going to send it to the list unless someone yells about CONFIG_SERIAL_MULTI and CONFIG_ENV_OVERWRITE, those are the only changes Oct 26 15:59:42 ok, yeah, I saw the couple comments go by to get you to v4..will try that in a bit Oct 26 15:59:50 thanks! :) Oct 26 16:01:13 meanwhile I can continue wishing we had 16 completely selectable in/out pins for each pru...I love wishes Oct 26 16:03:58 apologies for violating the normal friday policy here Oct 26 16:04:16 its fine Oct 26 16:04:22 we had 2 fridays already here this week Oct 26 16:04:30 dunno if I can handle one more Oct 26 16:04:31 *whew* Oct 26 16:04:48 imagine if we were sitting in the same office Oct 26 16:04:52 lots of fridays Oct 26 16:04:58 eternal friday Oct 26 16:06:55 if only we could get rid of the customers and still have revenue...things would be better Oct 26 16:07:23 * av500 puts on sinking hat Oct 26 16:08:55 mdp: +1 Oct 26 17:15:11 Anyone ever experience beaglebone getting stuck on systemd-fsck during boot? Oct 26 17:16:01 I am using a rootfs from http://dominion.thruhere.net/koen/narcissus/ Oct 26 17:16:31 heh no Oct 26 17:16:42 bad sd card? Oct 26 17:16:56 maybe Oct 26 17:17:02 let me give another one a shot Oct 26 17:18:31 hmmm... Parallella kickstarter has 29 hours to go and is $92k short Oct 26 17:19:24 they've got 28 hours to do it. It's gone up $15k in the last 20 min I've been watching. About $45k/h, so at that rate they'll make it. Oct 26 17:20:55 emeb: you could donate $92k :P Oct 26 17:21:18 I don't think that would get past the finance committee. Oct 26 17:52:11 <_av500_> kickstarters are like ebay auctions Oct 26 17:52:15 <_av500_> it all happens at the end Oct 26 17:52:40 <_av500_> except when they are totally out of whack Oct 26 17:52:46 <_av500_> then they fund in a day Oct 26 17:53:42 <_av500_> what the west coast time now? Oct 26 17:53:55 <_av500_> 11ush? Oct 26 17:53:57 <_av500_> ish Oct 26 17:54:11 ya Oct 26 17:54:32 <_av500_> good, keep at it Oct 26 17:54:49 are you writing this from happy hour? Oct 26 17:55:04 <_av500_> hmm Oct 26 17:55:14 the parallella one is madness Oct 26 17:55:35 how much money are they trying to raise? Oct 26 17:56:30 do not watch the latest video Oct 26 17:56:41 <_av500_> it has cyberspectrum? Oct 26 17:56:44 I have not watched any video Oct 26 17:57:06 apparently the latest way to raise fund is show ubuntu running on an arm, display on a tv, with a kid googling Oct 26 17:57:16 no content about the interesting part f the project Oct 26 17:57:24 <_av500_> and yelling "parallel, parallel, parallel" Oct 26 17:57:47 devfunks: it is always happy hour in Germany Oct 26 17:57:58 <_av500_> it could be a classroom full of britich kids Oct 26 17:58:01 I get the impression it's a small group of inexperienced amateurs trying to build on a shoestring budget a sophisticated SoC an established semiconductor company would need years to produce Oct 26 17:58:02 <_av500_> brtitish Oct 26 17:59:30 * _av500_ goes to pickup a kid in fact Oct 26 18:01:00 Crofton|work: they're trying to get $750k for a maskset. Only $80k short at this point. Oct 26 18:02:07 the $99 devboard that they're promising would be a smoking deal all by itself. Oct 26 18:02:34 w/ a Xilinx zynq chip - digilent sells a board with that for > $300. Oct 26 18:04:58 ah, so they the need the money to make themaks for the interesting bit Oct 26 18:05:24 pretty much. Oct 26 18:05:53 I don't see how they're going to fund the maskset *and* all the devboards for that amount though. Oct 26 18:06:16 yeah, that would be my concern Oct 26 18:08:49 Considering qty pricing on Xilinx Zynq chips is > $300/ea those devboards are a loss @ $99. Oct 26 18:09:45 correction: >$200/ea @ 1000pcs qty. Oct 26 18:31:12 I fail to understand Oct 26 18:46:48 Crofton|work: what don't you understand? For $99 pledge, Adapteva is promising a devboard with a Xilinx FPGA on it that currently lists for >$200, plus other goodies based on their own IP. Oct 26 18:48:25 Seems like there's a math problem there to me, but then I'm not an entrepreneur. :) Oct 26 18:48:26 I do not understand how tehy will deliver Oct 26 18:48:33 Nor I. Oct 26 18:51:16 I think I understand Oct 26 18:51:38 for 700K I buy a harme for a year, then apologize for failing Oct 26 18:52:03 It's happened before on KS. Oct 26 18:56:57 emeb: don't FPGAs generally have huge discounts for education, etc.? maybe they somehow managed to arrange a deal? Oct 26 18:59:17 General pricing for FPGAs is take the price Digikey shows and divide by anywhere between 3 and 10, depending on volume and other agreements Oct 26 18:59:53 Digikey shows $250-ish for 1 Zynq, so in volume they'll be a bit less than $99 Oct 26 19:00:14 <_av500_> emeb: aint it about delivering their SOC? Oct 26 19:00:30 <_av500_> the adapteva super duper? Oct 26 19:03:18 _av500_: I guess they need FPGA as a glue between their chip and ARM, and it does not look like they are adding FPGA as just an extra bonus feature Oct 26 19:13:12 ssvb: It's possible that they've got backing from Xilinx and they're getting the Zynq parts for a good price. Oct 26 19:15:27 emeb: New Lecroy scopes with 12-Bit vertical res. Oct 26 19:15:36 but at the moment, they're on the hook for ~3000 of these devboards, and assuming a low-ball BOM of $250, they're $450k in the hole before they've even funded the chip development. Oct 26 19:15:54 ouch Oct 26 19:16:18 _av500_: That's what you'd think, but it looks like most of the KS cash is going to building devboards at the moment. Oct 26 19:19:00 djlewis: cool. 4k points would smooth things out nicely. Oct 26 19:19:58 bradfa: the $227 pricing mentioned was based on Avnet's 1000 qty pricing. Oct 26 19:22:41 emeb, That's not an accurate price then, if Digikey will sell me 1 for $250, the 1k qty pricing isn't 4227 Oct 26 19:22:52 s/4227/$227/ Oct 26 19:23:18 I'd venture the 1k pricing they're getting puts the Zynq under $50 Oct 26 19:23:30 probably well under $50 Oct 26 19:23:42 * prpplague would disagree with that Oct 26 19:24:02 most SoC pricing doesnt break until after 24K units Oct 26 19:24:14 prpplague, fpga pricing breaks quite early in my experience Oct 26 19:24:28 high end fpgas don't go in high volume products Oct 26 19:24:46 but doesn't the zynq include an arm core? Oct 26 19:24:59 * prpplague tries to refresh his memory about the zynq Oct 26 19:25:01 prpplague, yeah, but $10 for the ARM core is reasonable Oct 26 19:25:07 even in low qty Oct 26 19:25:09 I'd bet Oct 26 19:25:15 bradfa: if it was just the arm core Oct 26 19:25:35 prpplague, I doubt the $99 price point makes them money Oct 26 19:25:36 bradfa: but when you start doing on die integration, the amount of work goes up, as well as the costs Oct 26 19:25:49 I'm just saying the Zynq isn't >$200 for them, probably Oct 26 19:26:01 prpplague, how much are the FPGAs with PowerPC inside? Oct 26 19:26:09 Xilinx has hard powerpc cores in some fpgas Oct 26 19:26:18 older stuff Oct 26 19:26:30 * prpplague is not familiar with powerpc pricing, but _is_ familiar with ARM pricing Oct 26 19:27:29 bradfa: Old Virtex4 parts had PPC cores and were in the $100-$200 range before they got eclipsed by V5 w/o the hard cores. Oct 26 19:28:42 emeb, ok, so maybe I'm talking out of my butt... Oct 26 19:28:45 But essentially what you're saying is that Avnet is quoting a price that's ~4x higher than what you actually pay if you close the door and work out a deal. Oct 26 19:28:47 could happen :) Oct 26 19:28:52 emeb, yes Oct 26 19:28:54 that Oct 26 19:29:24 And in the FPGA world that's not completely unlikely. Oct 26 19:31:27 summary: this KS is a risky investment? Oct 26 19:32:47 My fundamental argument is that they're on the hook for ~3000 devboards that they're not making money on. That's eating up enough of the funding that they're going to have a challenge to afford the chip development. Oct 26 19:33:16 So yeah. Risky. Oct 26 19:35:52 they probably have investors outside of kickstarter Oct 26 19:37:01 Likely even. Oct 26 19:42:07 mdp, I'm not buying one, would prefer to use $99 on Beagles or Arduinos Oct 26 19:42:20 same here Oct 26 19:42:59 four arduino's 1 beagle Oct 26 19:43:02 or Oct 26 19:43:04 bradfa, that's 14 6502s Oct 26 19:43:22 just saying... Oct 26 19:43:27 or beer Oct 26 19:43:33 friday! Oct 26 19:44:02 or 10 new arudino's programmable with a basic compiler :P Oct 26 19:47:33 or 20 lunch boxes! Oct 26 19:47:56 * bradfa 's lunch box is still in the static bag Oct 26 19:48:19 well, it's friday and almost beer o'clock, cyas! Oct 26 19:50:34 10 STM32 Discoveries! Oct 26 19:52:15 40 beers! Oct 26 20:03:18 * djlewis will raise you 2 lunch boxes, 2 Discoveries and a arduino R3 Oct 26 20:17:43 <_av500_> emeb: it says for 99$: You get everything in the SUPPORTER reward and an Epiphany-III based Parallella board Oct 26 20:17:57 <_av500_> epiphany is their chip, no? Oct 26 20:18:28 <_av500_> ah, there is that zync too Oct 26 20:18:54 <_av500_> this makes no sense Oct 26 20:19:00 <_av500_> why the fpga and the 16-core? Oct 26 20:22:38 probably so they have something that will run a real OS to talk to their weird stuff? Oct 26 20:23:28 <_av500_> emeb: yes, so why not buy a cheap arm core Oct 26 20:23:33 <_av500_> why the fpga? Oct 26 20:23:52 <_av500_> ah, because that would force them to desing a full SOC Oct 26 20:24:04 <_av500_> this way the offload all hte ugly glue to the fpga Oct 26 20:24:08 good question. Maybe they've got strange interfacing requirements? Still, a cheap external FPGA could do that. Oct 26 20:24:38 <_av500_> in order to develop for their chip, al they need is it on a pcie Oct 26 20:24:45 <_av500_> but that is not sexy Oct 26 20:25:05 <_av500_> so they build a 16+1 core r-pi Oct 26 20:25:54 it needs 16 binary blobs to boot? Oct 26 20:25:55 a lot about that doesn't make sense, given what they're willing to reveal. Oct 26 20:27:53 they'll probably design their cores into the FPGA fabric so that when the silicon has bugs they can still ship. Oct 26 20:31:57 koen, i fixed the appended dtb patch to apply Oct 26 20:32:07 sending pull req now with the audio stuff Oct 26 20:55:23 Anyone else having 0003-ARM-CUSTOM-Build-a-uImage-with-dtb-already-appended.patch fail to apply on beaglebone-3.7? Oct 26 20:55:42 Aw, don't append your dtb Oct 26 20:55:44 just pass it along Oct 26 20:56:23 I'm just trying to run patch.sh, figured it should probably complete without errors. Oct 26 20:56:58 * mranostay bangs head Oct 26 20:57:19 desk shrugs Oct 26 20:57:59 georgem: Need a bit more context then to help I guess, but don't append your dtb, just pass it on through Oct 26 20:58:20 And if you're using a U-Boot that doesn't support device trees for am335x, upgrade to mainline v2012.10 release Oct 26 21:00:24 Tartarus: I'm just trying to run patch.sh from the 3.7 branch of git://github.com/beagleboard/kernel.git. It tries to apply that patch I mentioned and fails. Oct 26 21:01:57 I'll start over from scratch and see if the same thing happens. Oct 26 21:15:07 Russ: howdy Oct 26 21:15:19 howdy Oct 26 21:15:33 finally ditched wireless and got a nice bump in connection speed Oct 26 21:15:38 http://i.imgur.com/Od8WU.png Oct 26 21:16:20 <_av500_> dug a trench a ran a cable? Oct 26 21:17:02 start the DDoSing! Oct 26 21:17:09 lots of cable pulling Oct 26 21:17:17 _av500_: he could be sending terabyte drives by bike messenger Oct 26 21:17:36 you can get amazing bandwidths that way Oct 26 21:17:43 latency is a bitch though Oct 26 21:17:47 hehe Oct 26 21:17:48 a van full of blu-rays Oct 26 21:18:09 I suspect hard drives have higher density Oct 26 21:18:24 a bluray is only 50G Oct 26 21:18:35 smaller Oct 26 21:18:45 really? Oct 26 21:18:55 probably lighter Oct 26 21:19:10 hrm, maybe not Oct 26 21:20:21 a bluray disk is 120mm diameter and ~1mm thick Oct 26 21:20:53 <_av500_> mru: using a tube with that diameter that would be a hell of a data conduit Oct 26 21:21:06 in drivermaker's measures, 20 of them hold a terabyte Oct 26 21:21:07 <_av500_> stuff it with blurays Oct 26 21:21:18 that's larger than a 1TB hard drive Oct 26 21:21:31 volume wise? Oct 26 21:21:37 and there are drives much much higher capacity these days Oct 26 21:21:50 mranostay: by my estimate, yes Oct 26 21:22:09 <_av500_> one could send hard drive platters Oct 26 21:22:17 <_av500_> no need for the eletronics to travel Oct 26 21:23:41 _av500_: platters like these? https://lh5.googleusercontent.com/-Y3MXWWiI83Q/UFn51LDeX1I/AAAAAAAAClw/3r3x6l_aucw/s919/2012-09-18_15-03-34_788.jpg Oct 26 21:24:04 _av500_: er isn't alignment a bitch on the other end? Oct 26 21:24:48 mranostay, single platter laptop drives Oct 26 21:25:29 <_av500_> mranostay: why? Oct 26 21:25:34 <_av500_> we send stuff to mars Oct 26 21:25:46 we crash into mars occasionally too Oct 26 21:25:48 <_av500_> we faked a perfect moon landing Oct 26 21:25:53 <_av500_> we can aling a platter Oct 26 21:26:00 _av500_: packet loss rate to mars is pretty bad though Oct 26 21:26:12 thurbad, that's just to show mars we mean business Oct 26 21:26:19 lol Oct 26 21:26:29 <_av500_> +1 Oct 26 21:27:03 <_av500_> nuke from orbit first, take rock sample later Oct 26 21:27:18 wow, this planet is really radioactive Oct 26 21:27:26 <_av500_> wonder why Oct 26 21:27:27 .... nope no life here Oct 26 21:27:33 <_av500_> lets move on Oct 26 21:27:36 <_av500_> next! Oct 26 21:29:18 * mranostay hurls gamma rays at mru Oct 26 21:29:56 * mru dons lead foil hat Oct 26 21:30:05 * _av500_ hurls himself into bed Oct 26 21:30:11 bed crumbles Oct 26 21:32:53 * mranostay makes some neutrons and fires at mru Oct 26 22:44:52 hello Oct 26 22:45:06 anybody is here? Oct 26 22:52:32 exit Oct 26 23:24:53 heh Oct 26 23:33:15 huh Oct 26 23:33:36 hah Oct 26 23:34:15 truly the three horsemen of the troll apocalypse Oct 26 23:34:27 i am not a troll! Oct 26 23:34:59 I might be, but I'm not saying for certain. Oct 26 23:35:10 <_troll_> I'm not Oct 26 23:35:17 hmmmm Oct 26 23:35:27 "Why do you contradict people?" Oct 26 23:35:46 contradiction is just a state of mind Oct 26 23:36:07 ask schrödinger's cat Oct 26 23:36:19 emeb must be schrödinger's troll Oct 26 23:38:06 I contradict myself until you lift the lid on the box. Oct 26 23:39:48 :<>] Oct 27 00:38:53 BB Xm autologin not working - just hangs there - any way to CTRL- something to exit display mode on bootup and go to command line? Oct 27 00:58:47 \o/ Oct 27 01:00:01 try escape meta alt control shift Oct 27 01:00:11 no wait, that launches the editor Oct 27 01:08:10 beer time is almost upon us! Oct 27 01:08:23 almost? it's 2am... Oct 27 01:08:48 why are you not blacked out? Oct 27 01:09:00 where's the fun in that? Oct 27 01:10:24 it sure is fun waking up the next morning and seeing the people you drunk texted :) Oct 27 01:11:19 I don't tend to do that Oct 27 01:12:36 i've finally have learned to handle my alcohol with years of practice though Oct 27 01:17:07 mru: can i put that as a LinkedIn skill? Oct 27 01:17:18 drunk texting? Oct 27 01:17:49 mru: beer Oct 27 01:27:26 no combination of escape meta alt control shift worked (that are the permutations on that?) Oct 27 02:25:32 i'm trying to compile u-boot for my beagleboard xm, but when i copy u-boot-2012.07 onto my debian-wheezy image, i get a "ramdisk image or corrupt or invalid" error Oct 27 02:25:54 if i use the u-boot from the debian image, everything works nicely Oct 27 02:26:17 oh, sorry, "Wrong Ramdisk Image Format", then "Ramdisk image is corrupt or invalid" Oct 27 02:26:28 any thoughts? **** ENDING LOGGING AT Sat Oct 27 03:00:01 2012