**** BEGIN LOGGING AT Tue Jan 22 02:59:58 2013 Jan 22 03:00:21 a bit bummed. Was hoping to get some more mileage out of that slide set. Had some expanded demo I wanted to do. Oh well Jan 22 03:01:09 I'm going to try to the vendor night or whatever it's called. Haven't seen an email about that yet. Of course I didn't get an email about registration being open either, and missed early-reg... argh Jan 22 03:07:11 we are sure to troll your booth Jan 22 03:17:29 koen: so does power over usb not work with your 3.8 branch? Jan 22 03:19:04 mranostay: trying to implement http://www.ietf.org/rfc/rfc3251.txt over usb networking? Jan 22 03:19:22 ah crap i know what i did Jan 22 03:19:44 heh kinda.. :) Jan 22 03:25:51 mru: someone seems to have taken that rfc seriously: http://spectrum.ieee.org/energy/the-smarter-grid/japans-digital-grid-scheme Jan 22 03:36:28 mranostay: you missed the sillyness this morning Jan 22 04:00:27 that's why we have logs Jan 22 04:01:52 have you got a mill? Jan 22 04:01:55 for your logs. Jan 22 04:02:41 its the new industry, to move to, once you release your bitmine is going to send you broke, becuase you are spending more on electricy than you are earning Jan 22 04:12:21 never happen Jan 22 04:12:40 there's a limit on how many bitcoins are allowed in circulation. Jan 22 04:29:48 there is a limit on a fictional thing? wow! Jan 22 04:46:41 sigh... profs are sure lazy thesedays Jan 22 04:49:51 hmmm well there goes using my PRUSS VFD for the demo.. too much power :) Jan 22 04:52:59 how so? Jan 22 05:05:52 Is there anything special I need to do to get sound working on the beagleboard XM with the 3.2 kernel Jan 22 05:05:54 ? Jan 22 05:08:25 Dunearhp_: set the mixer settings Jan 22 05:12:23 mranostay, what do you mean "too much power"? Jan 22 05:12:33 the PRUSS draw 21 Jigawatts? Jan 22 05:27:54 <_av500_> damn Jan 22 05:27:58 <_av500_> my usb rtc woke me up Jan 22 05:33:16 _av500_: wow. where did you get your musb drivers? :D Jan 22 05:35:10 ds2: didn't say it was on time Jan 22 05:35:43 so it is really a usb unreal TC? :D Jan 22 05:41:08 he probably requested a 20amp outlet for his demo Jan 22 06:07:54 hey, scale 11x came out with their schedule too Jan 22 06:24:47 Is it possible to run the beagleboard's micro usb port in gadget mode while powering it? Jan 22 06:25:08 I understand you can't use it in host mode while powering it... but how would it even work with the boot sequence as a gadget port Jan 22 06:27:05 sure Jan 22 06:28:40 sure as in it's possible and should work -- when I get my driver working? I'm new to this whole thing, but doesnt the host send the initial message, so if the device is still booting up while it recieves that... won't it not work? Jan 22 06:28:52 no Jan 22 06:29:01 per default it should be in device mode Jan 22 06:51:34 So then, during the boot process, when the computer sends the signal via usb to my beagleboard, it should respond correctly? Even though it's being powered by ut? Jan 22 06:51:36 it^ Jan 22 06:52:30 yes Jan 22 06:52:34 like a mobile phone Jan 22 06:52:44 that you charge and connect to at the same time Jan 22 06:54:52 hey mr av500 Jan 22 06:55:00 hey mr mranostay Jan 22 07:00:02 what does this DT warning mean "has invalid length (4 bytes) (#address-cells == 2, #size-cells == 1" Jan 22 07:15:42 mranostay: people report that it only works for a short while Jan 22 07:15:57 mranostay: I bet playing with the cpufreq governor would help Jan 22 07:51:17 anyone; can someone point me at the logs for the channel Jan 22 07:51:19 ? Jan 22 07:51:25 sure Jan 22 07:51:42 logs.nslu2-linux.org/livelogs Jan 22 07:51:51 We have a unit that says "card removed" and throws I/O errors at us when it gets cold. I disabled card detect (gpio_cd = -EINVAL), but still get I/O pain. Anyone else have bad experiences with MMC in cold environments? Jan 22 07:52:49 **** BEGIN LOGGING AT Tue Jan 22 02:59:58 2013 Jan 22 07:52:51 tasslehoff, ya, too many Jan 22 07:52:51 and thats it Jan 22 07:53:03 okay let me get you the backlog Jan 22 07:53:04 jas Jan 22 07:53:53 mrpackethead_: why not make your own logs? Jan 22 07:53:58 http://logs.nslu2-linux.org/livelogs/beagle.txt Jan 22 07:54:31 cause i'm a leech. Jan 22 07:54:40 i'll suck someone else logs. Jan 22 07:54:49 guess Ishould setup the rotator Jan 22 07:55:45 lol. Jan 22 07:55:50 just goign to move. Jan 22 07:55:50 tasslehoff, I had to put mine *inside* the transmitter to keep it warm enough. Jan 22 07:56:14 ka6sox: both good and bad to know. more to say on the issue? some of our units misbehave when tested (in a fridge), others don't. people want me to find out if sw can "fix" it. Jan 22 07:56:48 I've not found one...I often wondered if it was mechanical or electrical. Jan 22 07:57:00 tasslehoff: SW can make it hot Jan 22 07:57:06 put it inside a 5kW transmitter...problem solved. Jan 22 08:19:25 av500: by making everything else on the board work hard? Jan 22 08:22:58 tasslehoff: make me think of https://plus.google.com/100242854243155306943/posts/9VbpSWoZJkp (pin mux) Jan 22 08:28:57 keesj: hehe. wish that could solve this problem as well Jan 22 08:38:03 ka6sox: 5kW? that is quite a lot :) Jan 22 08:38:28 tasslehoff, its a baby Jan 22 08:59:10 Now my beagleboard XM has a kernel lockup problem Jan 22 08:59:28 Linux 3.2 Jan 22 08:59:43 has anyone encountered that? Jan 22 09:00:43 you mean has a kernel ever locked up? Jan 22 09:01:16 mine did...but then i was trying something different. Jan 22 09:04:05 * thurbad brings in the amazing karnac Jan 22 09:04:19 I suspect it has something to do with a large ntp time differential Jan 22 09:05:24 without background we really can't constructively comment Jan 22 09:05:47 but that won't stop us from commenting ~.~ Jan 22 09:07:07 No problem, just asking if it is a known issue. It is a new install and when I leave the machine for a few minutes I come back to a blank screen, no network activity and the caps lock leds inactive Jan 22 09:07:32 the last thin in the journal is a message about a large ntp time slew Jan 22 09:07:49 just ran ntpdate, will see if it is more stable now Jan 22 09:09:51 are you in a console or gui Jan 22 09:10:03 was in a gui Jan 22 09:10:41 and it's not a screen saver? Jan 22 09:11:15 Might be caused by one, but my ssh connectrions drop as well Jan 22 09:11:34 eww Jan 22 11:47:46 hi,guys Jan 22 11:47:56 anybody there? Jan 22 12:08:12 no Jan 22 12:15:31 still not back from lunch? Jan 22 12:22:04 gm all Jan 22 12:22:14 gm one Jan 22 12:22:47 gm Jan 22 12:22:48 yikes. 1degree F outside. Jan 22 12:23:32 how much is that in nubbins per square pegs? Jan 22 12:24:00 42 Jan 22 12:25:58 jkridner: same here, watch my kid drop his usb stick (filled with speech slideware) getting on the bus...after digging in the snow and getting it he was clutching his wind-swept icy hands as the bus left Jan 22 12:26:04 darn sneaker-net Jan 22 12:26:50 speech slideware is what you call warez these days? Jan 22 12:27:05 warez+pr0n Jan 22 12:27:30 panto: sorry, pron and kids is a taboo Jan 22 12:28:13 oh sorry, I guess they'll have to wait until they're 18 Jan 22 12:55:48 we don't have pr0n in The Heartland(tm) Jan 22 12:55:51 real america ;) Jan 22 12:58:36 only hentai? Jan 22 13:00:45 lol Jan 22 13:04:13 ka6sox: bof landed at 1030 thursday Jan 22 13:06:16 if nobody shows up I can go heck^H^H^H^Hsee joel_ and denix Jan 22 13:07:35 panto: did you see kevinsc's session? Jan 22 13:10:04 kevinsc has a session at ELC? Jan 22 13:10:10 yes Jan 22 13:10:25 there's a funny back story on it Jan 22 13:10:58 I was cced on a comment from marketing, "if ELC accepts only one presentation from TI, let's be sure it's this one" Jan 22 13:11:00 LOL Jan 22 13:11:06 heh, WBU gets nuked, ASP finally gets a chance to present Jan 22 13:11:25 mdp, yes Jan 22 13:11:32 which is even funnier given how many TIers are on the schedule Jan 22 13:13:44 mmm...a repeat of linusw on pinctrl Jan 22 13:14:06 did anybody else go to that one in 2012? it was interesting Jan 22 13:14:06 russ, denix, joel_, mdp, panto, mranostay, prpplague, koen, kevinsc, wmat, mru_ ... pretty excellent #beagle representation! Jan 22 13:14:49 jkridner: nice strategy...buying up all the "airtime" at conferences ;) Jan 22 13:16:15 * mdp so wants a "z" in mru's session title...oooh. Jan 22 13:16:57 why? Jan 22 13:17:49 I'm a product of my environment...my brain simply requires that word to have a "z" Jan 22 13:18:01 dezign? Jan 22 13:18:09 yes! Jan 22 13:18:51 * mdp checks his shoe sise. Jan 22 13:22:49 Hi! Using PSP (3.2) kernel for am335x. When inserting thumb drive into USB1 (host only) it gets detected (lsusb), but no /dev/sda device created. With OTG port - no problem. With Koen's kernel (https://github.com/koenkooi/linux/tree/beaglebone-3.2) both ports are working as expected. Any idea? Jan 22 13:27:40 with the advanced test on the beaglebone, is that programmed into the chip/eeprom or is it part of the bootloader? Jan 22 13:27:42 koen > PSP Jan 22 13:28:51 but koen is evil, remember that Jan 22 13:29:06 evil vendor Jan 22 13:29:09 evil vendor tree Jan 22 13:29:33 yegorich: what are you trying to accomplish? why not just use what works for you? Jan 22 13:31:15 jkridner: I need everything ;-) For example dual MAC mode, fb_set_par support etc. This stuff is not present in the other kernel.รค Jan 22 13:31:36 making a custom board? Jan 22 13:31:54 jkridner: yes Jan 22 13:32:38 I wasn't aware the psp kernel didn't have working USB, but for the 3.2 kernel, http://github.com/beagleboard/kernel has patches on top of the psp kernel rather than mainline. Jan 22 13:32:56 so, you can look at the patches and figure out the fix. Jan 22 13:33:20 for a custom board I'd switch to the http://github.com/beagleboard/kernel 3.8 kernel Jan 22 13:33:25 be sure to look at the beaglebone-3.2 branch. Jan 22 13:33:28 that has dual mac and DRM Jan 22 13:33:42 no USB though, but you can nag jkridner about that Jan 22 13:34:13 not that it'd do you any good. I just turn around and bug kevinsc about it. Jan 22 13:34:51 you heard the man, go nag kevinsc about it instead :) Jan 22 13:35:03 O.K. :-) Jan 22 13:35:15 it's funny that the DT portion of dual-phy musb is in mainline, but not the actual driver code needed Jan 22 13:35:26 usually it's the reverse with TI am33xx stuff Jan 22 13:35:37 driver code without DT bindings, "tested" on a beaglebone Jan 22 13:36:23 sounds challenging Jan 22 13:38:05 What make me crazy is that other devices like FTDI or 3G modem have no problem. Only mass storage Jan 22 13:38:46 what does dmesg say? Jan 22 13:39:10 @av500: look here http://e2e.ti.com/support/dsp/sitara_arm174_microprocessors/f/791/t/240727.aspx Jan 22 13:40:45 USB drivers are compiled in Jan 22 13:41:52 add more nand, so mass storage isn't needed :) Jan 22 13:42:54 keyboard is also working Jan 22 13:44:24 pastebin a diff between the 2 kernels :) Jan 22 13:45:59 @av500: I'm digging through the commits right now :-( Jan 22 14:03:08 jkridner: I made the usb gadget port appear, no host mode Jan 22 14:03:32 jkridner: and the good news is that the usb patchset under review at l-o ml breaks it completely on am33xx Jan 22 14:07:42 I can confirm Jan 22 14:07:56 b0rked Jan 22 14:09:09 * koen stabs his ISP Jan 22 14:25:38 http://www.beertone.ch/ Jan 22 14:35:15 yegorich: :) Jan 22 14:35:57 @av500: I need a bottle of beer otherwise I can't solve USB issue ;-) Jan 22 14:36:03 :) Jan 22 14:37:14 @av500: Franziskaner Hefe-Weisse Dunkel please ;-) Jan 22 14:37:36 yegorich: tasty beverage Jan 22 14:40:48 jackmitchell: oh yes! Jan 22 14:42:07 agree Jan 22 14:44:00 jkridner: can you give http://dominion.thruhere.net/koen/angstrom/beaglebone/1ghz/uImage-dtb.am335x-bonelt-3.8 a try? Jan 22 14:45:13 k. Jan 22 14:48:12 koen: any idea how much power the bone consumes at 1GHz? Jan 22 14:48:37 Jacmet: not yet, need to hook it up to the PSU that supports digital readout Jan 22 14:48:43 koen: ok Jan 22 14:49:33 koen: the latest kernel build seems to have some latency spikes in the kernel somewhere, I can only assume it is something to do with the new PM patchset, have you noticed anything? The latency spikes seem to be ~600ms Jan 22 14:50:52 when I say latest, I refer to 3.8: add PM patchset Jan 22 14:51:37 jackmitchell: I haven't been doing latency measurements, but I wouldn't be suprised by those latency spikes Jan 22 14:51:38 koen, panto, robclark: is this necessary: https://gist.github.com/4595201 ? Jan 22 14:52:14 I don't think that will actually work Jan 22 14:52:30 jkridner, I agree with koen Jan 22 14:52:31 jkridner, not for hdmi or dvi.. Jan 22 14:52:47 k. well, fbset was giving me odd info. Jan 22 14:53:02 the only way I got HDMI to work was with hardcoding a res (720p reduced blanking) Jan 22 14:53:04 timings 0 0 0 0 0 0 0 Jan 22 14:53:06 jkridner, tilcdc will set the disp-pll appropriately.. and tfp410 and nxp output drivers will use edid for timings Jan 22 14:53:15 and even then the image was 'wavy' Jan 22 14:53:18 koen: they weren't present in earlier kernels so I will keep an eye and see if I can get to the bottom of it. It will probably be a temporary deadlock somewhere Jan 22 14:53:36 panto, fwiw, latest version on my bone3 branch introduces a bandwidth limit.. Jan 22 14:53:57 panto, so none of the outputs will try and pick a resolution higher than what the bone can handle Jan 22 14:54:06 robclark, nice Jan 22 14:54:17 (otherwise it would happily try 1080p which seems to be more than what the memory bandwidth can handle) Jan 22 14:54:18 what about the clocking? Jan 22 14:54:36 * jkridner starts boot of koen's new kernel build Jan 22 14:54:44 robclark, I would take a look at it too Jan 22 14:54:46 clocking will be picked based on pixel clock from the edid Jan 22 14:55:05 perhaps you can't generate an accurate clock at a high enough freq Jan 22 14:55:05 it will try to set disp-pll at 2x the pixel clock (since minimum divider in lcdc seems to be 2) Jan 22 14:55:15 hmm... tilcd ...: timeout waiting for framedone Jan 22 14:55:24 lots of eMMC timeouts (expected) Jan 22 14:55:54 jkridner, fwiw, I have to remove the eMMC entry from DT.. otherwise I think systemd happily tries to mount the device Jan 22 14:57:50 jkridner, oh, and the framedone thing should be harmless Jan 22 14:57:58 k, koen's image gives me HDMI output on my TV now. Jan 22 14:58:19 fbset gives bogus timing values, but resolution is right. Jan 22 14:58:41 image is overscanned, cropping off the menus. Jan 22 14:58:49 hmm, I'll have to check w/ fbset on my end.. Jan 22 14:59:16 fwiw, the fbdev part is handled all by the cma helpers in drm, but perhaps they aren't setting some fields in the var/fix structs.. Jan 22 14:59:54 koen: lsusb fails with unable to initialize libusb: -99 (probably expected) Jan 22 14:59:59 (for host) Jan 22 15:02:01 koen: I need modules to try out gadget drivers. Jan 22 15:04:41 koen: do you think you could make any type of adjustment to account for overscan on TVs? Jan 22 15:06:30 jkridner: there's afaik no scaler to do anything like that Jan 22 15:06:45 jkridner, your tv should have an option to disable overscan Jan 22 15:06:58 probably, but I couldn't find it. Jan 22 15:07:00 something like pure pixel for example Jan 22 15:07:06 and I suspect others will have similar issues. Jan 22 15:09:12 koen: is this the same kernel that is on github so that I can build the modules myself? Jan 22 15:10:20 jkridner: yes, I only uncommented the 1GHz OPPs in am33xx-common.dtsi Jan 22 15:11:47 cpufreq-info shows hardware limits of 275MHz - 720MHz Jan 22 15:18:18 using that kernel? Jan 22 15:18:48 jkridner: trying building one yourself with the 1GHz and 800MHz OPP uncommented Jan 22 15:18:57 k Jan 22 15:18:59 yegorich: I will happily join you for that Jan 22 15:19:00 koen; I'm going to try building that beaglebone image again :) Jan 22 15:19:22 * mdp stabs time warner Jan 22 15:22:16 * Russ hands mdp some verizon fios Jan 22 15:22:31 Russ, my ISP side is fine as usual ;) Jan 22 15:33:10 D'oh. No luck. Jan 22 15:33:59 is twisted available on beagleboard or beaglebone? Jan 22 15:34:11 i'd prefer to use twisted rather than node.js Jan 22 15:38:07 jimboy: sure, just opkg install the twisted modules you want Jan 22 15:38:17 and get help, using pything is not healthy Jan 22 15:38:29 k Jan 22 15:38:31 thanks Jan 22 15:38:54 i'm waiting patiently for my beaglebone to arrive atm Jan 22 15:39:10 Looks like I can't use Narcissus for my WM; I guess I'll have to resort to doing the install manually with opkg Jan 22 15:47:31 jkridner: this commit seems to break mass storage stuff: http://arago-project.org/git/projects/?p=linux-am33x.git;a=commit;h=76c5325b7710a3a81e0684bb322cf8023c1191b3 Jan 22 15:47:57 jkridner: not that I could see why .... Jan 22 15:49:39 my guess would be that some clock doesn't get set correctly in some situation Jan 22 15:56:53 mru: O.K. will dig this tomorrow Jan 22 15:57:25 I say that because the commit changes how the musb clocks are managed Jan 22 16:03:34 koen, did you get my patch for the bbone reboot? Jan 22 16:06:13 jackmitchell, fyi - the rcu fixes are in rafael's next branch http://git.kernel.org/?p=linux/kernel/git/rafael/linux-pm.git;a=shortlog;h=refs/heads/linux-next Jan 22 16:06:26 jackmitchell, just closing the loop on the rcu warnings Jan 22 16:06:58 jsabeaudry: haven't seen it yet, sorry Jan 22 16:07:04 jsabeaudry: where/how did you send it? Jan 22 16:07:09 jackmitchell, thanks for reporting the bug and verifying the fix. Jan 22 16:07:33 koen, here is it: http://pastebin.ca/2305540 Jan 22 16:08:30 jsabeaudry, probably belongs to l-o for upstream merge? Jan 22 16:08:56 NishanthMenon: no problem Jan 22 16:09:13 jsabeaudry, i would recommend a BUG() instead of while(1) when it falls through OCP barrier Jan 22 16:09:35 ugh. how long does 'git gc' take! :( Jan 22 16:09:37 jsabeaudry: can you pastebin the git-format-patch'ed version? Jan 22 16:09:50 jkridner: a few minutes, it's cpu bound, single threaded Jan 22 16:10:01 NishanthMenon, No idea, I'm really new to all this stuff, I just did copy paste from a couple of places Jan 22 16:10:13 it does about 2.5MiB/s on my ivy bridge i7 Jan 22 16:10:38 jsabeaudry, posting a patch out is rather easy. Jan 22 16:10:57 jsabeaudry, just do a git add files Jan 22 16:11:00 git commit -s Jan 22 16:11:15 git format-patch -M -o . -1 Jan 22 16:11:56 http://stackoverflow.com/questions/5203436/how-to-set-git-commit-author-properly-from-rake-running-under-localsystem-acco -> just do a git config with your name and email ID Jan 22 16:12:06 prior to doing a git commit :) Jan 22 16:12:21 there's a 'fix author' switch to --amend Jan 22 16:12:26 doesn't fix SOB, though Jan 22 16:12:57 git commit --author="John Doe " --amend Jan 22 16:12:59 is an example Jan 22 16:15:28 jsabeaudry, one more thing v |= 1; you might want to be defining and using a macro -> some SoCs have cold reset and warm reset bits in RSTCTRL Jan 22 16:15:53 so a v |= 1; forces reviewers to read through TRM to figure out what you are attempting there Jan 22 16:17:05 wtf is arm_1103_le-gcc Jan 22 16:18:00 where did you find that? Jan 22 16:18:02 NishanthMenon, good points, I'll fix them and figure out the git format-patch Jan 22 16:18:04 related question - RSTTIME might need to be setup based on SoC - am not really too familiar with am335x to know for sure, but on some platforms, using the default reset values caused havoc depending on pmic.. (e.g. omap5) Jan 22 16:18:31 Oh that I wouldnt know Jan 22 16:18:51 jsabeaudry, i did not know either till i could not get reset to work ;) Jan 22 16:19:31 jsabeaudry, nothing directly related to your patch ofcourse. that is standalone and seems valid Jan 22 16:20:14 mru: its a tell you, kill you thing Jan 22 16:20:40 https://github.com/apc-io/apc-8950 Jan 22 16:21:12 not bad Jan 22 16:22:19 WTF are they changing CROSS in a sub make file???? Jan 22 16:24:01 it's the kind of thing the kind of people who do that kind of makefiles do Jan 22 16:25:41 ah no Jan 22 16:25:45 Is it normal that the IRQ numbers in /proc/interrupts differ from one kernel to another although it is the same GPIO ? Jan 22 16:25:55 they have out of kernel module makefile *in* the kernel source tree Jan 22 16:26:06 KERNELDIR=../../../../ Jan 22 16:26:31 jsabeaudry: DT and non DT kernel? Jan 22 16:26:40 XorA|gone, yup Jan 22 16:26:45 jsabeaudry: yes then Jan 22 16:26:57 jsabeaudry: DT stuff allocates as needed Jan 22 16:30:52 and/or sparse irqs Jan 22 16:32:45 jsabeaudry, minor note, see arch/arm/kernel/process.c machine_restart function, when pm_restart fails, -> , it will fall through and be handled by standard handlers. while(1); in your patch is not needed Jan 22 16:34:33 koen: I think omap still does not support sparse IRQ Jan 22 16:35:45 koen: didn't take as long as I thought... stupid ssh connection had stalled. Jan 22 16:37:04 koen: memsize returns 384MB Jan 22 16:37:40 cat /proc/meminfo gives 509940kB free Jan 22 16:37:53 free looks good Jan 22 16:38:20 I was concerned b/c the .dtsi has 256MB in it. Jan 22 16:40:09 panto is there a syscall or sysfs entry for loading a new dtbo? Jan 22 16:40:38 there's a sysfs entry Jan 22 16:40:56 it's the slots attribute in the bone cape driver Jan 22 16:41:14 echo PART-NUMBER[:REVISION] > slots Jan 22 16:41:35 should request PART-NUMBER-[REVISION|00A0].dtbo Jan 22 16:41:51 sweet. thanks!! Jan 22 16:42:37 np Jan 22 16:42:49 is there a doc I should look for in the future? overlay-notes.txt ? Jan 22 16:45:03 the documents describe the overlay format mostly Jan 22 16:45:37 there is a document in the patches describing the bindings of the driver Jan 22 16:45:47 I see lots of those documents on the bindings. Jan 22 16:45:49 other than that, there's nothing else to configure (only this slots thing) Jan 22 16:45:56 anything documenting the command-line compiler? Jan 22 16:46:10 the dtc compiler you mean? Jan 22 16:46:14 yes Jan 22 16:46:38 dtc development is separate, you get drop in the kernel from time to time Jan 22 16:46:48 the dtc tree is at (/me looks) Jan 22 16:46:51 I'm thinking of doing a short write-up on providing run-time support for a cape not in the kernel/firmware. Jan 22 16:47:23 git://www.jdl.com/software/dtc.git Jan 22 16:48:09 there's a Documentation directory there that describes dtc pretty thoroughly Jan 22 16:54:17 jkridner: if you update the 3.8 tree, there are now a few extra patches in hacks/ Jan 22 16:54:28 k, thanks. Jan 22 16:55:09 jkridner: apply 0003 and the lt will get 1ghz support in its DTS Jan 22 16:56:13 jkridner: FWIW, that OPP table is incorrect for the lower OPPs, it only adds the new 800 and 1GHz ones Jan 22 16:56:37 jkridner: if you have time, plz fix the dts to have the correct table so we burn less power add the lower ones :) Jan 22 16:56:52 I've seen a lot of beer talk lately. Can I assume everyone has already seen http://www.internetofbeer.com then? Jan 22 16:59:55 NishanthMenon, since the while(1) is superfluous, am I correct to think that the barrier is also superfluous? Jan 22 17:00:15 jsabeaudry, barrier is mandatory Jan 22 17:00:36 else you cannot guarentee that the write to register has gone all the way and registered with PRCM Jan 22 17:00:48 on L3, there are multiple intitiators with different priorities Jan 22 17:00:53 some writes can get delayed Jan 22 17:00:57 hence the need for barrier Jan 22 17:01:40 udelays (even though present in generic code) is not trustworthy as a barrier Jan 22 17:03:07 jkridner, indeed, very cool Jan 22 17:03:46 jkridner, my brother-in-law was working in Flint yesterday...now on his way to my house with two growlers full of local Flint brew :) I love traveling beer friends ;) Jan 22 17:06:06 lots of good beer up here. my last one turned out pretty well until bottling. Bottles bottle-fermented at basement temp turned out fine. Some were bottle-fermented in a cooler per instructions we had, but those soured. :( Jan 22 17:07:00 hopefully the yeast will get back to work, but those beers are probably gone. Jan 22 17:08:16 jkridner, ugh..I had an excellent local Porter (I have to drink those) at The Henry when the GENIVI f2f was in Dearborn a few years back Jan 22 17:09:07 I should have logged the name properly like real pros do with their wine labels Jan 22 17:09:18 Woo, rootfs deployment worked Jan 22 17:09:19 now I can't remember the brewery that made it :( Jan 22 17:09:38 I just don't wanna do all the opkg stuff by hand :P Jan 22 17:18:20 NishanthMenon, koen, Have a look at this and tell me what you think: https://www.refheap.com/paste/e1f2940c8fe75302d55342ec1 Jan 22 17:18:22 CAn anyone tell me about accessing ETB registers (Coresight) (jo JTAG - just direct access) Jan 22 17:19:18 jsabeaudry, kernel geeks have seeing '//' Jan 22 17:19:49 * NishanthMenon checks to see if there is an rmw api Jan 22 17:21:11 jsabeaudry, am33xx_prm_rmw_reg_bits(AM33XX_GLOBAL_WARM_SW_RST_MASK, AM33XX_GLOBAL_WARM_SW_RST_MASK, AM33XX_PRM_DEVICE_MOD, AM33XX_PRM_RSTCTRL_OFFSET)? Jan 22 17:21:34 you can drop that entire v variable and usage Jan 22 17:21:58 ocp barrier could be (void) am33xx_prm_read_reg(AM33XX_PRM_DEVICE_MOD, Jan 22 17:23:14 other than that, add a commit message. $subject should probably look like: ARM: AM33xx: Add SoC restart function Jan 22 17:24:02 jsabeaudry, dont forget to add your SOB (Signed-off-by:) tag Jan 22 17:24:07 git commit --amend -s Jan 22 17:24:12 for adding it in Jan 22 17:36:08 morning Jan 22 17:36:26 I feel like I'm doing something horribly wrong... I have my microSD, I have two partitions... one vfat with MLO, u-boot.img, and uImage... and one ext3 with the extracted rootfs from the tgz I got off narcissus Jan 22 17:36:40 I throw it in the beaglebone, nothing Jan 22 17:36:52 It just sits there with its lights on Jan 22 17:36:58 nothing on serial? Jan 22 17:37:11 No null modem cable; i'm relying on the built-in LCD :\ Jan 22 17:37:25 I thought bone was usb->serial inbuilt Jan 22 17:37:49 Hmmmmmmmmmmmmmm Jan 22 17:39:30 I know its unusual for an engineer, but did you read the docs :-D Jan 22 17:39:36 >.> Jan 22 17:39:37 <.< Jan 22 17:39:48 lookin shifty there :-D Jan 22 17:39:52 >.< Jan 22 17:39:54 Feelin shifty Jan 22 17:39:54 =/ Jan 22 17:40:13 I'm not used to the hardware I'm working with actually being that smart =/ Jan 22 17:40:26 not using the right hardware Jan 22 17:42:14 agreed Jan 22 17:42:23 Not my call though, I'm just the engineer at the company. :P Jan 22 17:42:48 Em`Zee, have u tried putting an uEnv.txt on the FAT partition? Jan 22 17:43:08 my beagle would do nothing but boot the kernel and lock without it Jan 22 17:43:13 Hmm Jan 22 17:43:17 still not used to the new boot env Jan 22 17:43:18 Just blank? Jan 22 17:43:24 i cam from boot.scr days Jan 22 17:43:37 it has 2 lines in it Jan 22 17:43:41 Em`Zee: the bone is the almost ideal bit of hardware :-D Jan 22 17:43:42 dont recall what they were Jan 22 17:44:36 Bleah Jan 22 17:44:57 XorA; I have the drivers installed, plug in the USB, and Windows says there's a device called "SUBARCTIC" and doesn't find a driver for it Jan 22 17:45:22 Em`Zee: get yourself vmware player :-D Jan 22 17:45:50 also the uImage on fat is apparently not used anymore Jan 22 17:45:59 X/ Jan 22 17:46:01 seems uboot now pulls it from the ext3 partions Jan 22 17:46:38 other issues i ran into Jan 22 17:46:56 is that it seems u-boot requires the fat partition to be marked active in the partion table Jan 22 17:47:07 about would not load MLO/uboot without it set Jan 22 17:47:32 rather boot rom Jan 22 17:47:43 or wtf ever is in charge at that point Jan 22 17:47:47 could have been any of the 3 Jan 22 17:49:15 looks like the subarctic issue is common Jan 22 17:49:27 disclaimer: i have beagleboard c3 NOT beaglebone Jan 22 17:52:59 crashovrd: the marked-for-boot issue is indeed a requirement of the dm3730/omap3 boot ROM. Jan 22 17:53:27 on dm3730/omap3 (beagleboard-xm/beagleboard), you also have to have MLO be unfragmented and the first directory entry. Jan 22 17:53:44 on am335x (beaglebone), the first directory entry issue is gone. Jan 22 17:54:25 XorA; Nothing at all on the serial console Jan 22 17:54:42 Oh, take that back. I got a bunch of Cs Jan 22 17:55:32 At least you didn't get Fs Jan 22 17:57:09 I'd rather an A+ Jan 22 17:57:09 :P Jan 22 17:57:52 Em`Zee: you used omap3-mkcard.sh to make your cad? Jan 22 17:58:09 I used mkcard.txt from the demo images page, along with those boot files Jan 22 17:58:18 then I just used the tgz rootfs from koen's narcissus Jan 22 18:00:02 console=tty0 & dvimode=hd720 Jan 22 18:00:11 thats whats in my uEnv.txt Jan 22 18:00:26 * Em`Zee pulls out microSD for another go at mkcard Jan 22 18:00:27 im not kidding, it just hangs after decompressing kernel Jan 22 18:00:37 if i dont have a uEnv.txt Jan 22 18:00:49 havent reduced it to whether the diff is having it or not Jan 22 18:00:55 a P is sufficient. Jan 22 18:02:18 XorA; Crap, I just realized that mkcard has some error coming back from sfdisk :| Jan 22 18:02:40 sfdisk: ERROR: sector 0 does not have an msdos signature Jan 22 18:02:48 then /dev/sdb: unrecognized partition table type Jan 22 18:03:21 morning Jan 22 18:03:34 I mean, that's not surprising... it's dd-ing the drive first, why would it have that signature? Jan 22 18:04:18 Em`Zee: it should not have msdos signature as first step in mkcard is to kill that Jan 22 18:04:18 Em`Zee: you forgetting to run as root? Jan 22 18:04:28 Nope, I'm running as root Jan 22 18:04:33 sec, I'll pastebin Jan 22 18:04:52 running a crazy distro with an automounter? Jan 22 18:04:55 like ubuntu? Jan 22 18:05:07 Ubuntu yes, but no partitions are mounted Jan 22 18:05:41 lemme just turn that whole thing off Jan 22 18:05:42 sec Jan 22 18:06:07 * XorA has found that when automounter tries to mount partitions we havent formatted yet it all goes a bit wrong Jan 22 18:07:17 Ahhh dconf-tools, why aren't you included by default Jan 22 18:08:20 hmmm... a CfP for papers on Case-Based Reasoning sent to beagle ml... Jan 22 18:08:23 reject. Jan 22 18:08:55 XorA; http://pastebin.com/AUZFzKj3 Jan 22 18:09:08 as a data point (mostly for the logs) - the ubuntu openjdk binaries work fine on ARM Jan 22 18:10:00 XorA; And fyi, I did just check -- mtab says none of that is mounted Jan 22 18:10:11 (brb) Jan 22 18:10:16 Em`Zee: mkfs.vfat: unable to open /dev/sdb1: Device or resource busy Jan 22 18:10:28 something is using it which is not the script Jan 22 18:14:04 XorA; Which is weird because if I run it after the script executes, it runs fine Jan 22 18:15:17 put a big old sleep in there :-D Jan 22 18:15:33 but if you run the same command by hand that is the same effect Jan 22 18:17:49 remove card Jan 22 18:17:51 plug it back in. Jan 22 18:17:55 rerun script Jan 22 18:20:42 Yeah, doing that after I figure out why udisks won't listen to its own config Jan 22 18:23:34 Kernel Oops in check_preempt_curr? http://pastebin.ca/2305983 Jan 22 18:23:55 i think you are exposing a bug/assumption of the script Jan 22 18:24:26 jkridner: fwiw, I have a 1024x768 picture out of the hdmi port now with the latest 3.8 Jan 22 18:24:40 for some reason in your setup, the partition table didn't quite re-read correctly Jan 22 18:24:44 intentionally? Jan 22 18:24:52 wonder if you can force the re-read ioctl to sync it up Jan 22 18:25:10 my compile is still running. beagleboard.org is a super slow machine and I've booted into OSX on my desktop. :( Jan 22 18:25:46 wish the clocking was more consistant amount the sitaras :) Jan 22 18:25:47 :( Jan 22 18:25:58 *** No rule to make target `firmware/am335x-pm-firmware.bin' is back Jan 22 18:26:06 jkridner: have you read the README.md? Jan 22 18:26:14 it tells you to download it Jan 22 18:26:42 ds2: sfdisk kicks the re-read ioctl when finished Jan 22 18:26:58 adding a sfdisk -l will kick it again Jan 22 18:27:53 koen: evil vendor closed-source firmware? Jan 22 18:28:00 XorA: that might be a more appropriate sync up then a plain sleep Jan 22 18:28:26 koen: quick review my patches before panto breaks them Jan 22 18:28:36 turning off gnome automount also fixes the issue Jan 22 18:28:43 normally Jan 22 18:29:06 oh bastards Jan 22 18:29:12 I killed udisks Jan 22 18:29:15 no more automount Jan 22 18:29:19 still won't run cleanly Jan 22 18:30:05 * XorA never runs with a desktop on build machine anyway Jan 22 18:30:31 XorA: you could never work for Canonical :) Jan 22 18:30:47 mranostay: there is an ubuntu server edition Jan 22 18:31:03 ds2; You suggest adding an "sfdisk -l"? Where? Jan 22 18:31:16 Em`Zee: after the sleep Jan 22 18:31:30 n/m that if you got cruft in the background Jan 22 18:31:52 I would make that sleep real long and see if that fixes it :-D like 30s or so Jan 22 18:32:08 wow, that fubars it worse Jan 22 18:32:17 it tries to search on /dev/mapper Jan 22 18:32:39 you remeber to call it on same device? Jan 22 18:32:50 grabbing pastebin Jan 22 18:33:14 http://pastebin.com/7PGH7TVP Jan 22 18:33:17 sfdisk -l $DRIVE Jan 22 18:33:20 oh Jan 22 18:33:22 herp derp Jan 22 18:33:24 thanks Jan 22 18:33:45 real fix would be a way to check $DRIVE is not in use, not sure how to do that from a script Jan 22 18:34:21 wonder if fuser will tell you that Jan 22 18:35:03 still tossing out device or resource busy, but all the automounter crap is 100% stopped; processes aren't even running Jan 22 18:35:07 and things aren't mounting Jan 22 18:35:11 *mounted Jan 22 18:35:52 sleep 10; sfdisk -l $DRIVE; sleep 10 Jan 22 18:36:07 I but udev is holding the devices or something Jan 22 18:36:26 s/but/bet/ Jan 22 18:38:33 XorA; Any idea how to kill that? Jan 22 18:39:03 Does address 0xE584215C ring a bell for anyone? Jan 22 18:39:48 dont kill udev, thats bad (tm) Jan 22 18:40:03 Not literally killing udev :P Jan 22 18:40:08 did you try with the extended sleeps? Jan 22 18:40:08 I meant how to kill those bindings Jan 22 18:40:39 jsabeuadry: device, userland or kernel? Jan 22 18:41:42 Okay Jan 22 18:41:43 ds2, kernel paging request that seems to happen in check_preempt_curr Jan 22 18:41:50 Just put in the device, fresh reboot, no automounting Jan 22 18:41:51 good Jan 22 18:43:13 XorA; No go. Jan 22 18:43:33 Also I still get the sfdisk "sector 0 does not have an msdos signature" at the beginning Jan 22 18:43:47 sounds like your card is dead Jan 22 18:43:53 or usb card read/writer Jan 22 18:44:01 Howso? I can partition it and play with it if I'm not using this script Jan 22 18:44:07 jsabeaudr: oh Jan 22 18:44:11 I've done so 3 or 4 times today already Jan 22 18:44:24 Em`Zee: run the same commands by hand then, something odd is goign on Jan 22 18:44:33 lemme give her a reboot quick and I will Jan 22 18:44:57 what distro? Jan 22 18:45:03 mint Jan 22 18:45:06 as your the first report Jan 22 18:45:15 Linux Mint 14 Jan 22 18:45:24 so thats ubuntu 12.10? Jan 22 18:45:40 Linux MintVM 3.5.0-17-generic #28-Ubuntu SMP Tue Oct 9 19:32:08 UTC 2012 i686 i686 i686 GNU/Linux Jan 22 18:45:52 Quantal, yeah Jan 22 18:46:06 bizarre I think a lot of people have run script on that Jan 22 18:46:13 is this in a VM? or native? Jan 22 18:46:22 VM Jan 22 18:46:25 I wonder if that could be it Jan 22 18:46:39 But again, without that script I can still do these operations Jan 22 18:46:45 I'm going to do them one by one here Jan 22 18:46:56 one by one would indicate a timing issue Jan 22 18:47:13 sticking a sleep 10 between every command would indicate that probablu Jan 22 18:47:39 I have used script before on VirtualBox and VMWare but I did notice with later kernels timing issues with other devices Jan 22 18:50:42 Hm. Jan 22 18:50:45 Now that's interesting Jan 22 18:51:25 Oh, nvm, fatfinger :) Jan 22 18:52:52 Russ: interested in Lunch? :) Jan 22 18:57:31 * djlewis just finished pizza pizza ... :) Jan 22 18:58:57 Oy, you're right -- *something* is trapping it Jan 22 18:59:02 and I have no friggin clue what that is Jan 22 18:59:08 lsof is no help Jan 22 18:59:36 XorA: "timing issues" Jan 22 18:59:53 XorA: I used to work with a guy who said that all the time when he really meant "race condition" Jan 22 19:00:25 glad to be off that particular project :) Jan 22 19:01:03 djlewis: hot and ready $5 is hard to beat Jan 22 19:01:46 Em`Zee: Im not too worried, once you have a valid card you rarely need another :-D Jan 22 19:01:54 XorA; I'm doing it in two passes Jan 22 19:01:59 one pass for the partition table Jan 22 19:02:04 one for the filesystems Jan 22 19:02:09 that seems to work fine Jan 22 19:02:23 It's something about sfdisk that's seeming to cause udev to wet itself Jan 22 19:03:02 I wonder if there's a bug in it about rereading partition tables without an insert event Jan 22 19:04:28 alright. Partitions are made. Now to copy stuff over Jan 22 19:09:27 wow Jan 22 19:09:33 XorA; Got it working Jan 22 19:09:39 but holy crap is there an issue with the rootfs Jan 22 19:10:05 It's looping tightly over starting D-Bus Jan 22 19:10:58 koen; Halp :P Jan 22 19:12:05 when will the shootings end. Jan 22 19:12:30 another episode in TX Jan 22 19:12:40 :\ Jan 22 19:13:12 If I could go back to the original Narcissus install I would Jan 22 19:13:14 mrpackethead_: obviously once there are armed guards EVERYWHERE... Jan 22 19:13:24 but I don't know if there's anything in there that will be compatible with the bone Jan 22 19:13:37 or at least that's what some people like to halucinate Jan 22 19:17:35 This is frustrating as hell... Jan 22 19:17:46 I feel like I'm the only one trying to use Narcissus on the bone Jan 22 19:17:47 :P Jan 22 19:22:09 Em`Zee: i felt the same way bout a clock yesterday Jan 22 19:23:17 I have a production crew waiting for me to give a thumbs up or thumbs down on the bone and I can't get anything besides the demo install working Jan 22 19:23:32 It's really not a good thing, because I like the hardware, but the software is driving me up the wall Jan 22 19:23:43 same boat Jan 22 19:24:15 well on the software issues Jan 22 19:24:24 nobody is waiting on me to approve a device Jan 22 19:24:26 :P Jan 22 19:24:28 mmm.. i've been able to use the bone to do some pretty usefu stuff so far Jan 22 19:25:42 whats preventing you from getting stuff to work Jan 22 19:26:06 Right now, I'm using Koen's new Narcissus and the rootfs it's giving me won't load dbus Jan 22 19:26:18 and the system gets stuck in an infinite loop Jan 22 19:26:45 plus that narcissus install won't embed xfce Jan 22 19:26:45 for me it just kernel panics a lot Jan 22 19:27:28 that was from narcissus Jan 22 19:27:42 what power source and board? Jan 22 19:27:43 in the demo images, opkg 404's Jan 22 19:28:16 Mine is a Beaglebone A6, running a 2A/5V DC power supply Jan 22 19:28:38 * XorA steals Em`Zee board Jan 22 19:28:55 Take it; maybe you'll have a better shot at making it work :P Jan 22 19:29:14 for me Beagle C3, Sony PS3 game pad charger Jan 22 19:29:17 Em`Zee: where do you call home? Jan 22 19:29:31 XorA; Wherever I hang my coat. Jan 22 19:29:35 Just kidding; Rochester NY Jan 22 19:29:49 crashovrd: is that microusb? Jan 22 19:30:24 if so that's probably the source of your crashes Jan 22 19:30:31 its got 2 full size USB and i have a cable to the mini usb on the beagle Jan 22 19:30:41 Em`Zee: bah, too much to walk for low value theft :-D Jan 22 19:30:42 power through the barrel connector Jan 22 19:32:13 XorA; Haha Jan 22 19:32:25 Looks like you're in UK -- wouldn't that be more like, too much to swim? :P Jan 22 19:32:46 Em`Zee: in Cupertino, CA Jan 22 19:33:03 XorA; Ahh... probably have better weather in Cupertino than we've got up here Jan 22 19:33:11 I woke up to it being -4F outside Jan 22 19:34:08 * XorA only works in SI temperatures :-( Jan 22 19:34:37 ah -20 Jan 22 19:34:37 crashovrd: i'm using an older ( squeeze ) version of debian.. but its solid and reliable, and predicatble. Jan 22 19:34:39 -20C Jan 22 19:34:42 nice Jan 22 19:34:54 crashovrd: it works nicely Jan 22 19:34:55 only gets that cold in Edinburgh once every 10 years or so Jan 22 19:35:14 yeah, there really needs to be a debian based distro for omapX Jan 22 19:35:27 and now I know google search will convert units :-D Jan 22 19:35:29 this open embedded stuff is just not cutting it Jan 22 19:35:34 crashovrd: how about debian? Jan 22 19:35:43 im fine with debian Jan 22 19:35:49 so run that Jan 22 19:36:02 need kernel and omap3 drivers Jan 22 19:36:09 for ? Jan 22 19:36:10 debian Jan 22 19:36:13 and the TI support stuff for dsp Jan 22 19:36:20 I was running Debian on it, then I realized that I couldn't do it because I was using the camera Jan 22 19:36:22 =/ Jan 22 19:36:30 u guys do headless Jan 22 19:36:31 any distro can run on any kernel Jan 22 19:36:34 within reason Jan 22 19:36:38 i need gles3, openvg, and openmax Jan 22 19:36:42 only a fool uses the distro-supplied kernel anyway Jan 22 19:36:47 oops Jan 22 19:36:50 gles2 Jan 22 19:36:56 mru: thats a pretty big call Jan 22 19:44:42 What's the most lightweight UI in the 2012.05 feed at this point? Xfce? Jan 22 19:44:48 I can't find IceWM or LXDE in there Jan 22 19:44:48 bash Jan 22 19:44:54 Heh, I mean X11 UI Jan 22 19:44:58 koen: no ksh? Jan 22 19:45:00 twm Jan 22 19:45:04 or xfce Jan 22 19:45:13 but xfce is not that much faster than the angstrom gnome2 Jan 22 19:45:26 mru: no, I'm fan of bash Jan 22 19:45:28 twm it is then Jan 22 19:45:33 or rather, I dont't know better Jan 22 19:46:33 bash is a reasonable interactive shell, sure Jan 22 19:46:50 and I use it Jan 22 19:47:30 but if the world weren't so full of idiots unable to write standard shell scripts, using something else for /bin/sh would be preferable Jan 22 19:47:59 and if you're _really_ tight on resources, you'll want something else for the interactive shell too Jan 22 19:48:36 connect the POSIX shell error routine to a cattle prod connected to someones delicates :-D Jan 22 19:52:32 dont worry, yesterday it was decided to add LUA to the kernel so you wont have to shell anymore! Jan 22 19:52:35 :) Jan 22 19:52:49 D'oh... base systemd image doesn't find the LCD Jan 22 19:52:56 two steps forward, one step back Jan 22 19:52:58 =/ Jan 22 19:53:26 Em`Zee, if u dont mind my asking, what is the use case for the hardware? Jan 22 19:53:57 crashovrd; embedded pill dispense system with visual verification of both user and pill type Jan 22 19:54:09 awesome Jan 22 19:54:12 So, my bone is bone+camera+rs232+lcd Jan 22 19:54:34 like a drug atm! Jan 22 19:54:36 Unfortunately, that's a lot of hardware to make work with one image; I generally get any 3 of the 4 working Jan 22 19:54:40 pretty much Jan 22 19:54:53 now i want one Jan 22 19:55:04 I'll mention that to the sales guy Jan 22 19:55:06 :) Jan 22 19:55:38 Em'Zee, great app, i'll ahve one working this week Jan 22 19:55:48 and we'll go into competition with you Jan 22 19:55:53 haha Jan 22 19:55:53 drug-cape Jan 22 19:56:02 hehe Jan 22 19:56:02 it will work well with baconcape Jan 22 19:56:05 and beercape Jan 22 19:56:09 I'd buy a baconcape Jan 22 19:56:25 you need a bitcoincape as well to pay for your baconcape Jan 22 19:56:29 Hahaha Jan 22 19:56:43 want to see ugly cape Jan 22 19:56:46 It'll be a moot point if I can't get this thing working :\ Jan 22 19:57:07 https://fbcdn-sphotos-g-a.akamaihd.net/hphotos-ak-snc7/428302_10151369241637661_66993423_n.jpg?dl=1 Jan 22 19:57:13 opps Jan 22 19:57:16 https://fbcdn-sphotos-g-a.akamaihd.net/hphotos-ak-snc7/428302_10151369241637661_66993423_n.jpg Jan 22 19:57:22 thats ugly cape Jan 22 19:57:31 sfw? Jan 22 19:58:00 yes.. very bording Jan 22 19:58:04 boring. Jan 22 19:58:13 it has nude electronics Jan 22 19:58:23 not even a solder mask Jan 22 19:58:26 Neat! Jan 22 19:58:31 how obsene is that Jan 22 19:59:48 nothing very clver there, just something to do some buffering Jan 22 19:59:49 koen: on the latest tree when I build, I'm back to getting "Invalid Format" on the HDMI. :( Jan 22 20:00:15 fbset shows the 1024x768, instead of the desired 1280x720. Jan 22 20:00:40 I do now have the 800MHz and 1GHz OPPs Jan 22 20:00:54 thanks to the hacks/0003 Jan 22 20:03:03 mrpackethead_: looks pretty clean to me... you should have seen mdp's 6502 nest-of-wires! Jan 22 20:03:32 is a ext4 file system a robust jorneled file system? Jan 22 20:04:02 Is there a list of bitbake targets anywhere for the angstrom build? Jan 22 20:04:17 Em'Zee, why are you using Angstrom Jan 22 20:04:50 mrpackethead_; Angstrom is the only distro I've found with support for the camera... I was in here about a month ago having this argument with the devs Jan 22 20:04:51 :P Jan 22 20:05:00 lol Jan 22 20:05:19 which camera Jan 22 20:05:29 3.2MP Camera cape Jan 22 20:05:32 right. Jan 22 20:05:42 theres lots of thigns that don't work with the beaglebone. Jan 22 20:05:47 its a "DEV" system remember Jan 22 20:05:50 I'm finding that out. Jan 22 20:05:53 Yeah Jan 22 20:06:25 i bet the camera woudl work with debian. Jan 22 20:06:36 you will jsut need to put the resources in to get it too work. Jan 22 20:07:42 Damn shame the client doesn't want to pay me for "putting in the resources" Jan 22 20:07:52 This project is very oxymoronic Jan 22 20:08:20 im unsure how to help you then. Jan 22 20:08:34 this square peg does not work with this round hole Jan 22 20:08:38 something has to change Jan 22 20:08:40 It's alright. At this point I'm sticking with angstrom, just trying to muck with getting X to run Jan 22 20:08:46 that'll be good enough for them :P Jan 22 20:09:11 its alright, i've passed your idea onto my chinese collegues Jan 22 20:09:17 Oh good Jan 22 20:09:18 they will have it running by lunchtime. Jan 22 20:09:30 Allwinner A10 !!!11111!!!! Jan 22 20:13:05 mrpackethead_; Let me know how that goes Jan 22 20:13:05 :P Jan 22 20:13:09 I'd like to buy one Jan 22 20:19:13 but your client has no desire to pay any money for it Jan 22 20:20:19 mrpackethead_; Always more work than there is money, these days. Jan 22 20:21:35 jkridner: I get 1024 if it can't read the EDID Jan 22 20:21:48 jkridner: and I get 720p if it can read it Jan 22 20:22:11 jkridner: I updated the drm patchset partially to include the bw limit robclark mentioned Jan 22 20:22:53 jkridner: fwiw, the 1024 mode works with my monitor, first time I get a "decent" picture Jan 22 20:23:04 koen, btw, I mentioned to jkridner.. not sure if I remembered to mention to you.. later versions of tilcdc changed driver name, s/lcdc/tilcdc/.. Jan 22 20:23:17 yes, I have those already Jan 22 20:23:22 ok, cool.. Jan 22 20:23:26 and used some sed -i magic to fix not-capebus Jan 22 20:23:33 ahh, perfect Jan 22 20:23:47 I only left out the new panels patch since not-capebus uses the old OF videomode helpers Jan 22 20:23:54 is there some history to the name 'not-capebus'? Jan 22 20:24:08 yes, it's not a bus and it's not just for capes :) Jan 22 20:24:28 'capebus' turned into a big bikeshed about the name Jan 22 20:24:43 to panto named the next version 'not-capebus' :) Jan 22 20:24:48 aaah Jan 22 20:24:49 haha Jan 22 20:25:00 is the plan for it to be mainlined? Jan 22 20:25:03 big bikesheds are full of many dutch? Jan 22 20:25:08 pllleeeaaase let it be called not-capebus Jan 22 20:25:22 koen, hmm... I did have a version that used older of helpers somewhere.. although that might have been pre s/lcdc/tilcdc/ so might be a bit of a mess to pull together.. Jan 22 20:25:29 XorA: not, the shed is uphill, unreachable by dutchies Jan 22 20:25:36 * XorA laughs Jan 22 20:25:47 need that bru NRG :-D Jan 22 20:26:22 robclark: basically drm/tilcdc: add support for LCD panels (v4) introduced the new OF helper dependency Jan 22 20:26:30 * mdp notes that the parc55 ELC rate is higher than off-the-street rate Jan 22 20:26:34 the lcdc drm driver itself doesn't seem to care Jan 22 20:26:55 koen, yup Jan 22 20:27:08 xanium4332, if $DEITY^Wmaintainers willing sure Jan 22 20:27:17 looks like people get useful work done with it Jan 22 20:27:25 so just dropping the panel output part is fine if lcd cape not needed urgently Jan 22 20:27:41 what state is mainline 3.7/3.8 for the beaglebone in? Jan 22 20:27:47 panto, capevan! Jan 22 20:27:59 robclark: I use drm/tilcdc: add support for LCD panels (v3) :) Jan 22 20:28:07 ie, should I abandon my 3.2 hacking for a new cape Jan 22 20:28:11 koen, ahh, ok, perfect Jan 22 20:28:16 that should work Jan 22 20:28:20 it does Jan 22 20:28:22 DVI works Jan 22 20:28:26 hdmi 'works' Jan 22 20:28:30 and lcd capes work Jan 22 20:28:35 \o/ Jan 22 20:28:36 mdp: the street is even cheper Jan 22 20:28:41 xanium4332: it's wonderful! Jan 22 20:28:46 unless they have gpio-keys, the PM patchset seems to cause troubles there Jan 22 20:28:47 usb host also works (but it doesn't pick up card insertions easily) Jan 22 20:28:48 McASP (a.k.a. audio in)? Jan 22 20:28:52 I would use mainline...again and again! Jan 22 20:28:55 xanium4332: no! Jan 22 20:29:00 :( Jan 22 20:29:00 mranostay: a lot of people sleep in the street in that area Jan 22 20:29:01 []_[] Jan 22 20:29:11 koen: yes i know Jan 22 20:29:12 xanium4332: I'm be posting patches for that RSN Jan 22 20:29:19 s/I'm/I'll/ Jan 22 20:29:22 xanium4332, audio does work Jan 22 20:29:39 mdp: awesome, I have a 4-channel audio input cape I'd like to get working Jan 22 20:29:54 good night ppl Jan 22 20:30:18 panto: including recording? Jan 22 20:30:32 recording should work, but it picks up nothing Jan 22 20:30:34 edit nm: night time has arrived Jan 22 20:30:44 I suspect some other board related problem Jan 22 20:30:50 I only have one hacked up board to test Jan 22 20:31:03 this is in 3.8? Jan 22 20:31:03 panto, he did say mainline, not the evil 3.8 vendor kernel ;) Jan 22 20:31:15 a PSP 3.8 kernel? Jan 22 20:31:17 panto is talking about the beagleboard/kernel tree Jan 22 20:31:21 aaah Jan 22 20:31:29 there's no such thing as PSP at TI any longer Jan 22 20:31:33 ok Jan 22 20:31:45 is the new kernel going to support a real time clock? Jan 22 20:31:53 so the beagleboard/kernel 3.8 branch has working audio Jan 22 20:32:44 xanium4332: there's at least some not-ready-for-mainline stuff in panto's not-capebus branch v15ish or so Jan 22 20:32:51 does the old 3.2 kernel still have the audio recording dropout bug? Jan 22 20:33:05 I have a separate clean version using simple-card/DT for mainline submission Jan 22 20:33:14 xanium4332: I've heard it does Jan 22 20:33:23 but never used audio on 3.2 before Jan 22 20:34:11 mmm Jan 22 20:34:49 trying to decide which kernel tree to attempt to add support for my cape to (which is a 4-channel audio recording cape) Jan 22 20:35:56 hey guys, ( im using the beagleboneA6 with the cam-cape and lcd3 cape) once there was a bug mentioned about the v4l drivers, output was dma_alloc_coherent of size 0 failed, at my own application, with the cheese application from the newest angstrom-img there comes the same message, just with a hugh number like 6298... instead of 0, does anybody knows if they fixed it? Jan 22 20:37:34 alexxd: no fix yet Jan 22 20:38:15 alexxd: so large framesizes (>6MB) are hitting an edma programming limitation, the '0' error is a buggy v4l2 implementation Jan 22 20:38:21 alexxd: both are know, no fix yet Jan 22 20:38:38 so i guess i have to use a webcam through the usbhost ? Jan 22 20:39:04 there's PSP/3.2, beagleboards/3.8, koen's (which version?), and pantos/not-capebus-v15 Jan 22 20:39:06 alexxd: I'd try to find out what gstreamer does to make it work Jan 22 20:39:29 'my' version is just the output of beagleboards/3.8 Jan 22 20:39:39 and pantos/not-capebus-v15 is pretty much the same as pantos/not-capebus-v15 Jan 22 20:39:40 ehm Jan 22 20:39:46 the same as beagleboards/3.8 Jan 22 20:39:48 koen: allright, thx Jan 22 20:41:48 xanium4332: if you are starting *right now* you want to use 3.8 Jan 22 20:41:54 #so beagleboards/3.8 sounds like the best target? There's been a huge amount of dev between 3.2 and 3.8, anybody Jan 22 20:41:59 do'h Jan 22 20:42:27 put hash in terminal prompt -> save for later. Put hash in irc prompt -> sends anyway :( Jan 22 20:43:05 Yeah I'm starting 'right now', but looking at 3.8, I've got a hold load of not-capebus and device tree stuff to learn, any pointers? Jan 22 20:43:51 how much coffee do you have on hand? Jan 22 20:44:30 sorry dumb question.. doesnt audio alrady work? Jan 22 20:44:45 not for my custom fabbed board Jan 22 20:44:53 ahh Jan 22 20:44:54 ok, Jan 22 20:45:02 what Audio IC are you using Jan 22 20:46:02 AD1937 Jan 22 20:46:19 ok.. Jan 22 20:46:24 i cheated a bit Jan 22 20:46:25 it appears the kernel has some support for it. The choice of audio IC was quite limited, as I wanted 4-channel input Jan 22 20:46:32 i've used the TLV320 IC from TI Jan 22 20:46:33 uhoh Jan 22 20:46:40 aaah Jan 22 20:46:41 but i only needed 1 audio in. Jan 22 20:46:53 does it work for you though? Jan 22 20:47:07 yes, fine Jan 22 20:47:32 im using it with ALSA and Jack Jan 22 20:48:13 here's a pic of the board: http://imgur.com/FnyMSFA Jan 22 20:48:40 nice. Jan 22 20:48:48 you using the beagleboard/3.8 kernel Jan 22 20:48:54 no. Jan 22 20:49:18 which one? Jan 22 20:49:20 3.2.33-psp26 Jan 22 20:50:01 mmm ok, I thought there was meant to be an audio bug in recording, but I don't know any of the specifics Jan 22 20:50:24 its working fine for me Jan 22 20:50:34 not seen any issues. Jan 22 20:51:11 might try it then, the whole multichannel bit is where things are going to get difficult Jan 22 20:51:30 that TLV chip supports 4 line ins Jan 22 20:51:58 i've litterally just replicated the Audio Cape Jan 22 20:52:08 it only has 2 audio ins. Jan 22 20:52:09 got the full part no Jan 22 20:52:45 TLV320AIC3106 Jan 22 20:53:36 mmm it only has a stereo ADC I think Jan 22 20:53:55 actuyaly you are right Jan 22 20:54:01 its got multiple inputs Jan 22 20:54:08 but they are muxed to the ADC Jan 22 20:55:04 I'm sampling hydrophones, and need 4-channels sampled together. That chip DOES have preamps, which I've had to put on my board separately (aka the four lines of components on my pcb) Jan 22 20:55:42 I don't think the codec will matter too much, at least if the ad1937 driver is up to scratch, I should be able to get it to work Jan 22 20:56:16 listening to whales? Jan 22 20:56:30 put two on the board? Jan 22 20:56:40 Close, calculating the bearing of an underwater pinger Jan 22 20:57:05 they need exactly the sampling clock, as I want to perform correlation on the data Jan 22 20:57:07 how far away are the hydrophones? Jan 22 20:57:17 Can edma slots be reused or am I supposed to alloc/free the slot at each transfer? Jan 22 20:57:22 from each other, yet to be decided, but currently 30-40cm Jan 22 20:57:34 how far away is the pinger? Jan 22 20:57:45 from the pinger, 1-60m Jan 22 20:58:00 oh quite close Jan 22 20:58:06 the pinger is *loud* as well Jan 22 20:58:13 yeah SNR won't be an issue I think Jan 22 20:58:24 nice.. Jan 22 20:58:28 can you tell us what it is? Jan 22 20:58:42 yeah, this is for a university engineering project Jan 22 20:58:51 an autonomous underwater vehicle Jan 22 20:58:56 ahh Jan 22 20:58:56 (AUV). Jan 22 20:59:14 koen: do you know if this bug (dma_alloc_coherent ...) exists only with angstrom (and opencv) or does the same happens if i will try it with opencv on a linux (ubuntu) os, finally both are linux systems Jan 22 20:59:16 We go to a competition every year, and one of the tasks is to home-in on an underwater pinger Jan 22 20:59:40 ahh, drive your sub to the pinger, and pick it up? Jan 22 21:00:04 The pinger is attached to a large remote controlled boat, which we need to surface under Jan 22 21:01:20 We're about to outfit our submarine with entirely ARM-based computers, one of which will be a beaglebone, for which this soundcard is designed for Jan 22 21:02:35 I've tried to change my hostname by replacing beagleboard in /etc/hostname and by using hostname, when I log on locally, the console shows user@newhostname when i ssh in it shows user@oldhostname. Can anyone explain this? Jan 22 21:03:01 cached in sshd probably Jan 22 21:03:18 even after multiple reboots? Jan 22 21:03:24 and a hard shutdown Jan 22 21:03:24 mdp: I have lots of coffee, would you mind giving a summary of what I should know to switch development from 3.2 -> 3.8? I guess I need to relearn how capes are being implemented. Looks like I need to learn a lot of device tree stuff, and not-capebus? Jan 22 21:03:33 spacecolonyone: probably not then Jan 22 21:03:47 alexxd: it will happen on all Jan 22 21:03:57 you need to talk to panto about the state-of-the-art for capes, he understands that stuff as the author of it all Jan 22 21:04:00 alexxd: on ubuntu it will happen later, since it's so slow on the 'bone Jan 22 21:04:03 xanium4332: ^^ Jan 22 21:04:17 prpplague: DiP right now on bbc1 Jan 22 21:04:21 koen, "edma programming limitation" :P Jan 22 21:05:20 hm, k, thx guys Jan 22 21:05:43 mdp: and with regards to OE workflow and 3.8? I'm currently using Angstrom 2012.05, I've seen the 2012.12 branch, but should I move forward from 2012.05 if I want to use the beaglebone/3.8 kernel? Jan 22 21:06:01 not sure, I don't use OE directly Jan 22 21:06:16 Alternatively, are there any meta- layers that have a recipe for the 3.8 kernel. I know meta-beagleboard exists, but last time I checked it only included mainline stuff Jan 22 21:06:36 I use Angstrom armv5/v7 rootfses + opkg + cross/native gcc to produce what I need Jan 22 21:06:39 for userspace Jan 22 21:06:51 I leave OE to smart people Jan 22 21:07:16 While you guys are going on about updating to 3.8, did the smart reflex 1GHz BBxM stuff finally make it into mainline Jan 22 21:07:50 koen: hehe Jan 22 21:08:06 koen: i'll have to get a copy from my backup service this evening Jan 22 21:08:34 mdp: mmm ok, cheers for the info Jan 22 21:08:55 there's some knowledgeable OE folks lurking here Jan 22 21:09:05 OE/yocto/angstrom/poky/etc Jan 22 21:09:10 I'm still doing all my stuff on 3.2 and the only things that I really miss (that I know of) are proper power off (I understand the BBxM just enters a while loop forever at halt), wake on pan support (the usb ethernet gadget supports it, I think the board does as well at a hardware level), and 1GHz operation Jan 22 21:10:41 systemd was on my list, (not a kernel thing, I know) but just recently is was updated to 189 so I was happy about that. Koen, if that was you, thanks Jan 22 21:10:58 koen; If I'm not mistaken, it was you doing work on the LCD4; was there anything specific I need to put in uEnv.txt to make it play nice? Jan 22 21:11:06 I've been jumping into OE and Angstrom etc..., but coming up with a sane workflow for kernel development, OE development etc... is proving hard Jan 22 21:11:35 i've decided its ok, to live on a older kernel Jan 22 21:11:48 maybe it prevents me from being quite so cutting edge Jan 22 21:11:50 xanium4332, I hear you. It took me a while to get the hang of it. Jan 22 21:11:59 but it does mean i can work with stuff that is functional. Jan 22 21:12:15 I still run up against stuff that bites me Jan 22 21:12:34 mrpackethead_: I'm fine with being on 3.2, I just need to have working audio, which you say you have achieved :) Jan 22 21:12:53 yes, but not with that IC you have Jan 22 21:12:58 #spacecolonyone: any tips for kernel workflow, atm I'm adding patches into the OE kernel build Jan 22 21:13:01 as in I don't know how to do it, but I ended up creating my own layer as an overlay and using the angstrom setup scripts and it has really made it much better Jan 22 21:13:13 create your own layer Jan 22 21:13:22 mrpackethead_: of course not, I'm fully prepared for kernel hacking, just I thought the McASp driver was broken Jan 22 21:13:23 xanium4332: it pretty much worked out of the box for me under debian squeeze Jan 22 21:13:40 just had to add some kernal modules Jan 22 21:13:57 did have some issues with jackd Jan 22 21:14:07 it quite port nicely to arm Jan 22 21:14:08 for instance I'm having to patch for a few different things, https://github.com/baileyji/m2fs-layer Jan 22 21:14:14 but is now working fine. Jan 22 21:14:22 jackd works a treat. Jan 22 21:14:24 spacecolonyone: I'm doing the exact same thing with layers, but mainly for angstrom customisation. I've not made my own kernel bitbake recipe Jan 22 21:15:57 yea, I've appended the mainline 3.2 recipe with the patches I need. the next thing on my list is to figure out how to add to my layer so when I build my image it does things line placing install files where I want them and creates users for me automatically Jan 22 21:15:59 sidetrack: ooo you have a patch for an aptina sensor Jan 22 21:16:05 yea Jan 22 21:16:43 isn't there a base-files recipe you can extend to add files to the image (ie login moto) Jan 22 21:16:49 s/moto/motd Jan 22 21:16:51 not mine, I pilfered it, as you can see, but I needed it as a proof of concept for the custom imager I'm making Jan 22 21:17:26 did you fork angstrom's setup-scripts and add your branch in? Jan 22 21:17:36 there is, but I haven't taken the time to find it yet. Jan 22 21:17:43 no maybe I should have. Jan 22 21:17:57 only if you're using angstrom... Jan 22 21:18:17 how do you get away with using the mainline 3.2 kernel? don't you need the PSP release? Jan 22 21:18:19 I haven't rebuilt or pulled in months, so I'm actually kind of terrified that all hell is going to break loose next time I pull and rebuild Jan 22 21:18:35 I am using 2012.05 Jan 22 21:19:53 i'm annoyed. Jan 22 21:20:18 i dont' ahve a DS1339 sitting in the proto parts draw. Jan 22 21:20:38 digikey, overnight Jan 22 21:21:02 are your patches to add support for the DS1339? Jan 22 21:21:20 If so, you may have just solved a *major* headache I was trying to solve earlier today Jan 22 21:21:36 me? Jan 22 21:21:56 By switching all our machines to ARM within the submarine, we've lost all the RTCs we have. We need good logs, so was trying to come up with an RTC solution Jan 22 21:22:15 I meant mrpackethead_ , who mentioned the DS1339 Jan 22 21:23:26 like you, my machines have to run 'disconected' Jan 22 21:23:40 and there is no guarrentee that they will stay powered up Jan 22 21:24:40 dumb question, if i want to employee someone in the USA Jan 22 21:24:43 but i'm not a US company Jan 22 21:24:51 how do you pay them Jan 22 21:24:55 and sort out their tax Jan 22 21:25:34 send gold bars Jan 22 21:26:22 set them up with an offshore bank account :) Jan 22 21:27:08 bitcoin Jan 22 21:27:15 * xanium4332 ducks for cover Jan 22 21:27:38 * mru covers xanium4332 in ducks Jan 22 21:29:59 mrpackethead_ do it not depend the country your company is located? Jan 22 21:31:50 might be more complicated when you are located in north korea Jan 22 21:35:07 mrpackethead_: you just get them to invoice you, it is upto them to sort out tax (if your in the EU) Jan 22 21:36:03 XorA, is being too serious, but this is the best approach Jan 22 21:36:59 yo Crofton|work Jan 22 21:37:12 hey Jan 22 21:37:16 T-11 until beer days :-D Jan 22 21:37:17 freezing here Jan 22 21:37:21 yes Jan 22 21:37:35 and I have a flaky central heting unit Jan 22 21:37:43 crofton will be cold as last year in brussel Jan 22 21:37:49 get on that bike and peddle for heat Jan 22 21:37:51 bother Jan 22 21:37:53 I did Jan 22 21:38:10 I am hoping it warms up for FOSDEM Jan 22 21:38:33 yeah a repeat of the sun and t-shrts your would be nice :-D Jan 22 21:38:41 s/your/year/ Jan 22 21:39:52 I ahve OE stickers Jan 22 21:46:48 Crofton|work: I can see this going wrong after a few beers Jan 22 21:49:20 #oeallthethings? Jan 22 21:49:29 no good ever came from a room of drunk engineers Jan 22 21:49:57 next year I want to get temporary tattoos Jan 22 21:52:31 Is there a way to set a default window manager without using .xinitrc? (Since I can't seem to find xinit or startx in the angstrom feed) Jan 22 21:54:32 why did you install more than one wm? Jan 22 21:55:10 I didn't Jan 22 21:55:15 I have twm installed Jan 22 21:55:23 but it doesn't autoload when I just run "X" Jan 22 21:57:15 uhm we have a twm recipe Jan 22 21:58:23 latest beagleboard image from angstrom/meta-ti -- aplay -l does not list any sound cards :-( Jan 22 22:00:03 woglinde; So far, none of the precanned stuff has worked for me. I've been at this for over a month. Jan 22 22:00:18 I said screw it, downloaded the base systemd image, and have just been hacking away at that Jan 22 22:00:44 perhaps modules need loaded or something ... Jan 22 22:01:57 snd_soc_twl4030 is loaded -- hmmm Jan 22 22:06:02 hiya cbrake Jan 22 22:06:20 Crofton|work: UT-OH!!!! Jan 22 22:07:28 cbrake: module dependencies for sound are probably still broke so you probably need more modules Jan 22 22:08:55 Crofton|work: #OETheWorld ! Jan 22 22:08:56 ka6sox: Hello Jan 22 22:09:06 Bitbake the Planet! Jan 22 22:09:21 XorA: yes, seems like there should be a beagle soc module or something ... Jan 22 22:10:55 hah kernel-module-snd-soc-omap3beagle Jan 22 22:10:59 * cbrake tries Jan 22 22:12:17 davest: i liked the other profile pic :) Jan 22 22:12:37 did he change the one on the elc website? Jan 22 22:13:00 yes Jan 22 22:13:10 he's such a stick in the mud Jan 22 22:13:43 and soc-omap, soc-omap-mcbsp ... Jan 22 22:17:15 ok... so the question is........ Jan 22 22:17:25 how many TI developers does it take to....... Jan 22 22:17:38 prpplague haha Jan 22 22:17:51 catch two squirrels running free inside the forest lane campuses building... Jan 22 22:18:36 depends on how bored everyone is with what the _ought_ to be doing. Jan 22 22:18:38 catch them with what? Jan 22 22:18:59 owls Jan 22 22:19:13 ...are not what they seem Jan 22 22:19:35 yes - all offices should be provisioned with owls in case of an outbreak of small furry creatures. Jan 22 22:20:37 prpplague: you mean nobody has pulled out six-shooter yet? Jan 22 22:20:48 youcantbetoocareful/thinkofthechildren/etc Jan 22 22:20:50 did anyone sign off on the owl spec? Jan 22 22:21:14 mranostay: hey, this is Texas we're talking about, not like your wild west of Californ... oh wait.. nevermind Jan 22 22:21:39 that's my next project - owlcape. Jan 22 22:21:52 emeb: owls with lasers/ Jan 22 22:21:52 ? Jan 22 22:22:03 and LUA Jan 22 22:22:08 mranostay: hehe no, but someone did have a static discharge gun Jan 22 22:22:13 here's a prototype: http://www.youtube.com/watch?v=qIYZiTqenJs Jan 22 22:22:53 prpplague: i think that will just get them to bite you Jan 22 22:23:11 honeybadger don't care Jan 22 22:23:33 how many shots for rabies again? Jan 22 22:23:38 emeb: way cool vid. Jan 22 22:23:38 friday? Jan 22 22:24:07 alan_o: that would be a fun DIY project - w/o the explosives of course. Jan 22 22:27:25 prpplague, again? Jan 22 22:27:39 prpplague, we had a squirrel run last week or so Jan 22 22:27:46 NishanthMenon: this time two Jan 22 22:28:08 prpplague, awesome.. mebbe we should switch to squirrel mail ;) Jan 22 22:28:52 * NishanthMenon considers it a *sign* :P Jan 22 22:29:47 omen of a squirrel uprising? Jan 22 22:30:00 prpplague: can you design us an owlboard? Jan 22 22:30:22 too many daisy guns :P Jan 22 22:30:26 nuke it from orbit, its the only way to be sure Jan 22 22:30:27 Pi-cape Jan 22 22:30:46 somrhing i have not seen yet, is i-cape Jan 22 22:30:54 ie, a docking station for your i-phone Jan 22 22:30:57 i'm really suprized Jan 22 22:30:59 so now I'm looking through the beaglebone/3.8 kernel, but I can't find any code for capes, all I can find is devicetree stuff and a single geiger cape file. Are devicetree files really all that are needed to make the existing capes work? Jan 22 22:31:01 still no smeagolboard, count me as disappointed. Jan 22 22:31:09 troll-cape Jan 22 22:31:12 hobbit-cape Jan 22 22:31:16 dwart-cape Jan 22 22:31:19 wizard-cape Jan 22 22:31:26 Mortgoth-cape Jan 22 22:31:27 mru: hehe Jan 22 22:31:29 get the set, and you can go on a enchanted jornery Jan 22 22:31:31 -t Jan 22 22:31:37 NishanthMenon: well it is an on going thing with the squirrells Jan 22 22:31:46 NishanthMenon: the attacked me the parking lot the other day Jan 22 22:31:55 i just have to finish bacon-cape first Jan 22 22:32:02 mrpackethead_: you'd need the special apple chip to do that Jan 22 22:32:04 NishanthMenon: and on friday they were throwing acorns at me from one of the trees Jan 22 22:32:17 NishanthMenon: so today i find one hiding in my cube Jan 22 22:32:36 http://hackaday.com/2012/03/20/birdwatching-meets-a-computer-controlled-water-cannon-awesomeness-ensues/ Jan 22 22:33:04 prpplague: thats the new Android developers Jan 22 22:33:16 use your opposable thumbs Jan 22 22:33:26 XorA: TI is paying them with acorns? Jan 22 22:33:30 instead of peanuts? Jan 22 22:33:36 prpplague: cost cutting Jan 22 22:33:46 whoa acrons are high class Jan 22 22:33:51 *acorns even Jan 22 22:34:02 says a man never tasted one :-D Jan 22 22:34:49 prpplague, opencv + watergun + squirells + beaglebone + E-wing/parking lot ;) Jan 22 22:35:00 you guys are nuts Jan 22 22:35:08 please tell me they get quality white oak acorns...choosy wildlife prefer the low tannic acids Jan 22 22:35:26 mru: +1 Jan 22 22:35:54 and being thus, WATCH OUT for them squirrels Jan 22 22:37:42 NishanthMenon: i've wanted to make a few capes like that with questionable legality :) Jan 22 22:37:49 http://en.wikipedia.org/wiki/Squirrel_fishing Jan 22 22:38:04 squirrel fur capes :_D Jan 22 22:38:40 Crofton|work: TIL... Jan 22 22:44:00 gn Jan 22 22:48:53 well, after yesterdays debate on real time clocks Jan 22 22:48:58 it seems really easy. Jan 22 22:49:22 im still worried i'm issing something Jan 22 22:53:19 debate on RTCs? Jan 22 22:53:29 HUSH! Jan 22 22:53:36 we dont mention that anymore Jan 22 22:53:43 * crashovrd scrubs the logs Jan 22 22:53:51 we've moved on to squirrels now Jan 22 22:54:07 i was left confused, bruisd and upset Jan 22 22:54:26 <_av500_> I left just in time Jan 22 22:54:30 <_av500_> in real time Jan 22 22:54:33 i did'nt think a RTC was such a painful subject Jan 22 22:54:51 _av500_: you have a JIT clock? Jan 22 22:55:12 so in the absence of a sensible answer Jan 22 22:55:25 i have just decided to do what i thought i'd do Jan 22 22:55:31 and will see how it works. Jan 22 22:55:46 <_av500_> mru: a lot of JITter Jan 22 22:56:08 how could there be anything about RTC to talk about? Jan 22 22:56:28 i think it was all about kernel stuff Jan 22 22:56:31 but it was way over my head. Jan 22 22:56:59 <_av500_> we concluded to put user space in the kernel Jan 22 22:57:08 <_av500_> then everybody wins Jan 22 22:57:20 ah Jan 22 22:57:31 well, some partys concluded to do that Jan 22 22:57:48 <_av500_> mrpackethead_: we call it the "ruling party" Jan 22 22:57:58 the side with the most cluebats wins Jan 22 22:58:12 A conventional computer operating system usually segregates virtual memory into kernel space and user space. Kernel space is strictly reserved for running the kernel, kernel extensions, and most device drivers. In contrast, user space is the memory area where all user mode applications work and this memory can be swapped out when necessary. Jan 22 22:58:22 is that not true? Jan 22 22:58:28 that is not true Jan 22 22:58:32 <_av500_> for some definition of true, yes Jan 22 22:58:54 kernel/user-space is about privilege levels, not address ranges Jan 22 22:59:00 if ( 1 == 2 ) Jan 22 22:59:40 of course there will be memory areas only accessible in kernel-space Jan 22 22:59:57 and in some cases there is also hyperspace Jan 22 23:00:00 so, because i really don't understnad whats going on. Jan 22 23:00:17 i'm simply bolting a RTC onto the I2C bus Jan 22 23:00:26 <_av500_> mru: do not forget white space Jan 22 23:00:27 writing an init.d script in the write place Jan 22 23:00:38 reading the time and setting the sytem clock. Jan 22 23:00:42 seems easy Jan 22 23:00:53 <_av500_> how mundane Jan 22 23:00:57 hwclock --hctosys Jan 22 23:01:10 i guess i will live or die on my deicison Jan 22 23:01:15 <_av500_> even pedestrian Jan 22 23:01:16 but at least i'll find out Jan 22 23:01:34 or just enable CONFIG_RTC_HCTOSYS Jan 22 23:01:58 <_av500_> but as long as DT itself does not allow to implement I2C by bitbanging in the .dt file.... what can we do... Jan 22 23:02:40 why do i need to mess with that kernel stuff though Jan 22 23:02:53 this seems to be pretyt straight forward, and will acehive the desired result. Jan 22 23:03:00 a machine that knows approximately what time it is. Jan 22 23:03:04 and here we go again Jan 22 23:03:07 :P Jan 22 23:03:32 there must be a simple answer to this..? Jan 22 23:04:01 im hearing that there is a way it will work; Jan 22 23:04:09 and i'm hearing that theres a way that it might work better.. Jan 22 23:05:34 _av500_: we need that forth interpreter Jan 22 23:05:55 <_av500_> what happened to the third? Jan 22 23:06:42 same thing that happened to ipv3 Jan 22 23:07:05 <_av500_> political backstabbery Jan 22 23:08:10 well from a hardware perspective, i don't think its going to make any differnece Jan 22 23:08:36 so, i'll just stick the RTC on the board appropraitely Jan 22 23:08:48 and make the software that -----> guy's problem. Jan 22 23:21:58 \o/ someone else's problem = win! Jan 22 23:25:02 since i pay that -----> guy Jan 22 23:25:11 i guess it is my problem still. Jan 22 23:28:13 hmmmmm Jan 22 23:29:10 ---> he understands this stuff much better than i do. Jan 22 23:29:33 what is this tempest in a teapot about RTCs? Jan 22 23:29:46 tempest at teatime Jan 22 23:40:34 good evening all. cant do initial boot on xm, please help Jan 22 23:41:13 reformatted sd to fat32, have all files from ext2 part on fat32 alonf with xxxx_jb.tar Jan 22 23:42:02 running win 7 x64 as host Jan 22 23:44:21 any takers? Jan 22 23:49:21 you're not gonna have links and such working on fat32.... Jan 22 23:50:21 permissions will be wrong as well Jan 22 23:54:49 heh the days when we used to be able to keep linux rootfs on fat are I think gone Jan 22 23:55:35 in that case, i am lost Jan 22 23:55:42 Guest18943: dont understand what you mean, what is random file xxx_jb.tar? Jan 22 23:55:49 i hate trees Jan 22 23:56:08 buy more books Jan 22 23:56:26 sry beagleboard-xm-jb.tar Jan 22 23:56:51 how to actually flash the image? Jan 22 23:56:58 is jb jelly bean? Jan 22 23:57:04 mru: fight the pollen though books? Jan 22 23:57:15 also do you have a serial connection to the board? Jan 22 23:57:18 jb is jelly bean Jan 22 23:57:33 * XorA presses the eject button now, never tried Android on beagle Jan 22 23:57:40 no serial at the moment. power and usb Jan 22 23:58:11 seriel is essential for all embedded guys, you learn to live on 80x25 at 9600 baud Jan 22 23:58:27 nah Jan 22 23:58:30 115200 baby! Jan 22 23:58:30 would this be easier if i had linux running somewhere? Jan 22 23:58:35 this is the 2010's Jan 22 23:58:42 ok, it's gonna be tough to diagnose if you get anything wrong in the boot setup even if you're close without a serial connection Jan 22 23:59:06 none of the wiki's say use serial Jan 22 23:59:08 yes, it would make things easier, even if it's in a virtual machine Jan 22 23:59:09 ds2: those crazy marvell chaps using 38400 :-D Jan 22 23:59:36 ok, i will linux in my virual box Jan 23 00:00:25 how does sd card need to be set up? 128mb ext2 and fat32 for the rest or all fat? Jan 23 00:00:47 fat32 just big enough for boot files, the rest whatever android wants Jan 23 00:00:56 if all fat, all files uImage etc, go there? Jan 23 00:02:07 fat32 for the 1st partition Jan 23 00:02:42 ok. where do I put the uImage, MLO and uboot? Jan 23 00:02:53 ext2 or higher for the second, the sizes don't matter a lot as long as you have enough Jan 23 00:03:14 MLO and uboot go on the first partition, in that order Jan 23 00:03:41 ok, cool Jan 23 00:04:25 and the flashing part? Jan 23 00:04:59 do I need x-loader? Jan 23 00:05:17 MLO = x-loader Jan 23 00:05:26 ohhh Jan 23 00:05:37 your xm doesn't have have nand, if that's what you're talking about flashing Jan 23 00:06:21 ok, I guess thats my confusion. I'm good with android devices(phones)...not so much here Jan 23 00:07:18 thanks all Jan 23 00:07:27 will be around... Jan 23 00:07:29 always good to stretch out into new areas Jan 23 00:11:51 time to head home Jan 23 01:09:06 anyone know if its possible to detect if power is being applied from the USB or from teh 5V rail? Jan 23 01:09:16 or is avaialbel on either. Jan 23 01:10:33 through software or physically? Jan 23 01:11:07 i was just thinking, if i could plug a battery pack into the USB port Jan 23 01:11:11 as a backup Jan 23 01:11:17 if the main 5V goes down. Jan 23 01:11:30 or rig up a battery pack for the barrel jack Jan 23 01:11:31 if the main 5V goes down, then i can shutdown gracefully. Jan 23 01:11:46 there are vendors selling that sort of cables Jan 23 01:11:59 signal pin to gpio Jan 23 01:12:18 i looked at teh scheamtics Jan 23 01:12:24 i could'nt see a way that it could work as it sat Jan 23 01:12:31 but this a room full of engineers! we can do better! you could send the signal over USB Jan 23 01:12:44 needs some kind of external signalling. Jan 23 01:13:07 engineers? where.. i can only see trolls, dwarfs and hobbits Jan 23 01:13:38 as a hobbit, I resent that :P Jan 23 01:13:46 and you could even re-use the kernel USB UPS drivers! Jan 23 01:14:37 and we could take back the arm trust zone! and run an OS on it that monitors that Jan 23 01:14:42 SECURELY! Jan 23 01:14:48 LETS BUILD IT!!! Jan 23 01:15:05 crap! we probably need a RTC Jan 23 01:15:13 oh well, it was a good plan guys Jan 23 01:15:24 heh Jan 23 01:16:20 I don't know that it switches from barrel to usb fast enough to avoid a power hiccup/reboot you'd need to test that before even attempting Jan 23 01:17:06 power is power, why not just use the barrel for backup power Jan 23 01:17:12 no need to route it over usb Jan 23 01:17:33 and like i said, hook the power sense line up to a gpio to notify failure Jan 23 01:27:21 hi mr pfefferz Jan 23 01:29:36 hi mranostay Jan 23 01:29:46 anyone remember the name of the X server that lets you reattach to it? like VNC but all done using X Jan 23 01:30:12 need a charger for a LiPO battery Jan 23 01:30:34 but that shoud'tn bee too hard. Jan 23 01:31:11 ds2: NX Jan 23 01:31:21 ds2: freenx for GPL version Jan 23 01:31:25 hmm didn't seem like it but I'll look it up Jan 23 01:31:45 its an X protocol proxy Jan 23 01:31:52 there is Xrdp Jan 23 01:32:07 i'm just wanting to bomb proof this thing a bit Jan 23 01:32:11 google dide some version of NX as well, cant rmeember the name Jan 23 01:32:50 yo prpplague Jan 23 01:32:57 hey Jan 23 01:35:24 ho Jan 23 01:43:33 ds2: Xpra? Jan 23 01:43:47 ds2: as used in winswitch? Jan 23 01:46:28 XorA: that still doesn't quite seem right Jan 23 01:46:41 maybe it was some obscure server Jan 23 01:47:20 I can emulate it with xmove and xvfb Jan 23 02:50:27 is there any particular reason to favor node.js over twisted on the beaglebone? Jan 23 02:50:55 i'd prefer to use python rather than javascript Jan 23 02:51:03 so, does it matter what i use? Jan 23 02:51:22 or umm..what are the advantages disadvantages to using either Jan 23 02:51:27 or is that question too broad to ask here Jan 23 02:51:35 just looking for pointers Jan 23 02:53:38 no reason other then there are people using it Jan 23 02:53:56 more people using it == more people to share stories with Jan 23 02:54:38 k Jan 23 02:54:43 yah Jan 23 02:54:55 that is good reason to use node.js Jan 23 02:55:13 as good as any Jan 23 02:55:18 some of us prefer to work directly with C :) Jan 23 02:55:23 i've heard its got a slight performance advantage Jan 23 02:55:37 just slightly Jan 23 02:55:38 i want to work with python Jan 23 02:55:42 primarily Jan 23 02:56:05 unless i'm directly targeting a specific platform Jan 23 02:56:16 then obviously c Jan 23 02:56:36 but ya know, i don't really care to develop my javascript skills Jan 23 02:56:41 i have better things to do Jan 23 02:56:45 with my time **** ENDING LOGGING AT Wed Jan 23 02:59:59 2013