**** BEGIN LOGGING AT Wed May 02 02:59:58 2012 May 02 04:05:48 Anyone had any luck using CAN on the beagle bone ? May 02 05:27:34 Hi, somebody May 02 05:27:57 you May 02 05:27:58 he May 02 05:27:59 she May 02 05:27:59 it May 02 05:43:37 <_tasslehoff_> koen: the patches you mentioned yesterday, do I pull them off the mailing list, or is there a place to cherry-pick them from? May 02 05:54:05 _tasslehoff_: https://github.com/Angstrom-distribution/meta-ti/commits/beagleboard-update-v2012.05 May 02 06:05:00 ds2: drinking?.. May 02 06:15:30 av500: big change May 02 06:16:11 blame my isp May 02 06:17:04 so, yoctizen released.... May 02 06:17:07 damn you ipconnect! May 02 06:17:34 * mranostay stares at empty IPA bottles May 02 06:17:47 what have i done?!?! :) May 02 06:18:40 you have *not* taken out the trash.... May 02 06:18:49 and remember to "recycle" or so.... May 02 06:19:33 i look like an alcoholic when i take out my recycling in my Yocto bag :) May 02 06:23:42 <_tasslehoff_> koen: https://github.com/Angstrom-distribution/meta-angstrom/commit/d3ed55d4e647138382930ab95fce66c28544a4bf <- forgot to commit var-run.conf? May 02 06:25:04 <_tasslehoff_> hum, no. wrong diagnose May 02 06:25:48 <_tasslehoff_> a case of my bbappend colliding with yours, I think :) May 02 06:28:00 hmm, TI is still keen on selling me an MP3 player from the 90's May 02 06:28:59 av500: linky-poo? May 02 06:32:50 error May 02 06:32:52 oops May 02 06:32:53 :) May 02 06:32:58 http://www.ti.com/tool/430boost-c55audio1 May 02 06:34:02 heh why is that still listed? :) May 02 06:34:32 I don't think its for an "mp3 player" May 02 06:34:45 more of something that can play back audio files in response to stimulus May 02 06:36:48 right, now I am convenced :) May 02 06:36:54 convinced even May 02 06:43:54 <_tasslehoff_> dell.com told me the other day that they recommend usb pendrives rather than floppies. controversial advice. May 02 06:45:07 who the fuck uses floppies? :) May 02 06:50:51 My Agilent spectrum analyzer uses floppies. May 02 06:53:59 what does a floppy run this days? May 02 06:54:57 don't even know where to buy them. I had to dig a bunch of used ones out of box in the garage. May 02 06:55:24 Mmmmm spec'ans May 02 06:55:37 wonder if those SM floppy adapters will work with that May 02 07:12:29 heh i forgot about 2.88M floppies May 02 07:21:07 <_tasslehoff_> koen: the sgx demos still fail with the patches. /etc/powervr.ini says WindowSystem=libpvrPVR2D_FRONTWSEGL.so.1 May 02 07:23:15 the raw demoes still fail? May 02 07:27:06 <_tasslehoff_> koen: yes. and the pvr-init.service fails to start. May 02 07:27:50 <_tasslehoff_> did a cleansstate and will rebuild to make sure I didn't fudge something up May 02 07:28:01 re-install windows to make sure May 02 07:30:41 <_tasslehoff_> ok. I will do that since I have doubt. May 02 07:37:08 _tasslehoff_: what sgx demo? May 02 07:37:20 the bouncing red and white ball..... May 02 07:37:50 * _tasslehoff_ prefers the chameleon man May 02 07:41:02 _tasslehoff_: what was the error? May 02 07:42:50 _tasslehoff_: I'm loading a fresh build onto my xM right now May 02 07:49:27 <_tasslehoff_> koen: fresh image now. no error on boot, but when starting a demo: http://pastebin.com/VAwTF69s May 02 07:50:22 <_tasslehoff_> koen: then on second boot pvr-init fails May 02 07:55:29 'The Polytechnic Simulator building is all clear. Fire determined the smell was a sewer odor only. You may now return to the area.' May 02 07:55:37 dude, I'll head over there RIGHT AWAY! May 02 07:56:46 good to know it's only gas, not fire May 02 07:57:12 sorry, suppose some context would help, the previous message was 'Please stay away from the area of the Simulator building on the Polytechnic campus. Fire is investigating a potential natural gas leak.' May 02 07:57:23 oh its natural May 02 07:58:46 new vulcano May 02 08:04:39 _tasslehoff_: its because /etc/pvr-init.sh have a $cputype wrong May 02 08:05:04 stuk_gen: you keep saying that, but it still reports it correctly on all the boards May 02 08:05:48 koen: have you try on bb xM? May 02 08:05:52 yes May 02 08:05:57 what image? May 02 08:06:07 doesn't matter what image May 02 08:06:17 there's only one recipe for the pvr stuff May 02 08:07:02 koen: can you give me your pvr-init.sh script? May 02 08:07:28 it's in meta-ti May 02 08:09:07 koen: meta-ti/recipe-ti/ ? May 02 08:09:26 making natural gas smell like shit may have been short sighted May 02 08:11:03 you mean making us smelling it like that was short sighted? May 02 08:12:23 it might have a reason we evolved to such. ;) May 02 08:15:34 <_tasslehoff_> nothing wrong with $cputype here May 02 08:21:48 koen: found the problem May 02 08:22:15 the version in git are https://github.com/Angstrom-distribution/meta-ti/blob/master/recipes-graphics/libgles/libgles-omap3/cputype my version of cputype is different May 02 08:22:38 and haven't this row -e 's/0x00000E00/OMAP3530/' May 02 08:24:00 aholler, natural gas is odorless May 02 08:24:07 aholler, the smell is an added sulfur compound May 02 08:50:39 koen, I was playing angry birds in that 3.3 seconds, that's what May 02 08:51:20 :D May 02 08:51:32 and it does that regardless of whether or not USB is going to work May 02 08:51:48 so its some other xM problem May 02 08:52:16 it just jumped out to me in your bootlog May 02 09:29:47 <_tasslehoff_> koen: did your xM like the fresh build? May 02 10:02:08 _tasslehoff_: I think I found the bug May 02 10:03:29 _tasslehoff_: where does your /usr/lib/libsrv_um.so point to? May 02 10:10:23 <_tasslehoff_> koen: /usr/lib/libsrv_um.so -> libsrv_um_dri.so.1.6.16.4117 May 02 10:11:03 _tasslehoff_: cd /usr/lib ; ln -sf libsrv_um.so.1.6.16.4117 libsrv_um.so May 02 10:13:48 <_tasslehoff_> koen: "PVRShell: EGL 1.4 initialized", so that's good. May 02 10:13:57 <_tasslehoff_> No chameleon man on my display though May 02 10:14:21 using x? and if so, is /etc/powervr.ini correct? May 02 10:14:28 and is DSIPLAY set? May 02 10:14:36 and does 'xhost +' help? May 02 10:15:00 <_tasslehoff_> not using X May 02 10:15:18 <_tasslehoff_> it took down the whole thing (hearbeat led stopped). rebooting now May 02 10:15:53 <_tasslehoff_> WindowSystem=libpvrPVR2D_FRONTWSEGL.so.1 May 02 10:16:25 omap3530? May 02 10:16:57 <_tasslehoff_> koen: yep May 02 10:18:21 you're hitting the bug sakoman and I are seeing as well May 02 10:18:28 instant lockup on omap3530 ES3.x May 02 10:29:46 <_tasslehoff_> ouch May 02 11:02:42 Has anyone seen this error when compiling for cloud9-image? NOTE: Running task 3389 of 4600 (ID: 1750, virtual:native:/home/michael/recon_aero/BEAGLEBONE/setup-scripts/sources/meta-openembedded/meta-oe/recipes-devtools/nodejs/nodejs_0.6.14.bb, do_populate_sysroot) NOTE: package nodejs-native-0.6.14-r0: task do_populate_sysroot: Started ERROR: Error executing a python function in /home/michael/recon_aero/BEAGLEBONE/setup-scripts/sour May 02 11:05:56 ooops, guess there's a char limit. Anyway, it goes on to say ....setup-scripts/build/tmp-angstrom_v2012_05-eglibc/sysroots/i686-linux: Cannot chdir: No such file or directory tar: Error is not recoverable: exiting now tar: This does not look like a tar archive tar: Exiting with failure status due to previous errors May 02 11:07:15 you're missing a file May 02 11:08:27 actually, it says there's no directory... May 02 11:08:39 so if you run tar -zxvf file -C ./dir May 02 11:08:53 it will change directory to that directory to expand the tar file... May 02 11:09:18 so you can probably get around it by creating the directory... or it could be that you're not running as root, so it cant make the directory May 02 11:10:46 tnx, I'll try running as root. I followed the directions for "git"ing the setup-scripts. So if I'm missing anything I'd think everyone would see this error May 02 11:12:28 * koen headdesks May 02 11:12:44 running a build as root May 02 11:12:46 worst May 02 11:12:47 idea May 02 11:12:48 ever May 02 11:13:06 gabeagle: is your home directory encrypted? May 02 11:14:02 double.rot13 May 02 11:14:20 av500: you cant build on encfs AFAIK May 02 11:16:11 or ecryptfs or whatever kernel version is called May 02 11:16:50 luks? May 02 11:18:26 dmcrypt is fine as that is block level May 02 11:21:05 http://www.realtimetechnosoft.com/ May 02 11:21:45 https://help.ubuntu.com/11.04/serverguide/ecryptfs.html May 02 11:21:48 *bingo* May 02 11:22:17 bingo? May 02 11:22:30 34 69 23 45 68 May 02 11:22:39 "eCryptfs is a POSIX-compliant enterprise-class stacked cryptographic filesystem for Linux" May 02 11:22:51 bullshit bingo May 02 11:23:10 bullshit? May 02 11:23:14 buzzword May 02 11:23:32 jannau: damn, it misses "synergy", so no jackpot for you.... May 02 11:23:54 AFAIK my home dir is not encrypted. Newb? How would I tell if it is? May 02 11:23:55 also no eco system May 02 11:24:09 I have seen at least two people complain of similar build problems due to ecryptfs having path depth limits May 02 11:24:31 gabeagle: its an option when you install May 02 11:24:53 XorA: path depth limit? wtf.... May 02 11:24:56 gabeagle: 'mount | grep home' should tell May 02 11:25:09 av500: thats what I thought :-D May 02 11:25:14 it's a real enterprise solution May 02 11:26:09 av500: it cant be enterprise its not in Java May 02 11:27:43 mount | grep home gvfs-fuse-daemon on /home//.gvfs type fuse.gvfs-fuse-daemon (rw,nosuid,nodev,user=) May 02 11:29:10 gabeagle: well thats ruled out that potential issue May 02 11:29:10 av500: do you've found another innovativ company which can teach you new levels in embedded carma design? May 02 11:29:13 gabeagle: that looks fine / not using ecryptfs May 02 11:30:04 BTW...compiling as root complains about needing PERSISTENT_DIR or CACHE set May 02 11:30:14 FFS! May 02 11:30:17 gabeagle: are there related errors in dmesg? May 02 11:30:18 13:12 < koen> running a build as root May 02 11:30:18 13:12 < koen> worst May 02 11:30:18 13:12 < koen> idea May 02 11:30:18 13:12 < koen> ever May 02 11:30:22 Anyone know if mc (midnight commander) is on some repository somehow? "opkg mc" doesn't know midnight commander May 02 11:30:45 aholler: well, somebody from that company think he needs to teach me how to behave May 02 11:31:10 educa: opkg update May 02 11:31:32 by telling him teh exact steps? May 02 11:31:43 jannau....not as far as I can tell May 02 11:31:48 http://www.angstrom-distribution.org/repo/?pkgname=mc May 02 11:32:31 ok, did update but "mc" still not known May 02 11:33:07 aholler: https://groups.google.com/forum/?fromgroups#!topic/beagleboard/7G3Z-YsEhZs May 02 11:33:27 gee, he even deleted his original mail May 02 11:33:40 too bad I full-quoted it... May 02 11:34:02 now he is telling me in PM that I should shut up since I dont know his answer May 02 11:34:17 or rather, I should work on the answer to him May 02 11:37:13 av500 I went to that url you gave and it shows package details, but even after opkg update it still is a unknown package May 02 11:38:05 educa: no idea, I dont use angstrom May 02 11:38:10 koen might know May 02 11:38:43 IC May 02 11:38:45 mc will show up later today May 02 11:39:09 I'm beginner I know sorry, but I just got my bone today and want to start using it May 02 11:39:12 thanks koen May 02 11:39:24 wtf, mips: https://developer.android.com/sdk/ndk/index.html May 02 11:46:54 are there any google certified mips based android devices? or any larger mips based android deployments? May 02 11:47:17 are there any larger mips deployments of anything these days? May 02 11:47:24 printers? May 02 11:47:29 or washing machines? May 02 11:47:47 jannau: they cant be certified until there is support for it :) May 02 11:47:53 toasters May 02 11:48:25 at nds I worked on a mips-based set top box nick-named "the toaster" May 02 11:48:31 it ran a bit hot May 02 11:48:59 mips might become owned by china May 02 11:49:01 but the STBs are all going arm now May 02 11:49:08 aholler: rumours, rumours May 02 11:49:28 china has been doing that loongson thing for a while already May 02 11:49:35 * mru is not impressed May 02 11:50:18 aholler: what does china gain by owning mips? they "have" the IP they need anyway May 02 11:50:46 don't know. just read that last week or so May 02 11:51:03 I read all sorts of crazy things about china last week May 02 11:51:08 but I assume they "have" every IP ;) May 02 11:52:26 anyway my router is still mips. there you have a large deployment May 02 11:52:59 yes, existing devices will not be spontaneously transmuted to arm May 02 11:53:30 maybe av500 should help with a hand waving ;) May 02 11:55:21 sorry, my transmuting powers are gone May 02 11:55:31 hm, looking at the Table Of Hardware at OpenWRT's wiki it's > 90% of mips May 02 11:55:46 as mru said May 02 11:57:04 ynezz: and I bet that Cisco's marketing is constantly bugging the engineers to add android to the routers May 02 11:57:11 bradfa: ping May 02 11:59:02 hmm, that company does everything, even mainframes and sap. May 02 11:59:16 hm, I can't remember any ARM soc with WiFi MAC and 5-port integrated switch/PHY and gigabit MAC May 02 11:59:25 aholler: sap for android? May 02 11:59:48 utilizing ti-dsps May 02 12:00:09 ynezz: if such things don't already exist, they probably will soon enough May 02 12:00:28 mips is for all intents and purposes dead May 02 12:00:29 I'm not aware about such SoC May 02 12:00:38 nor am I May 02 12:01:03 but I'm not in the habit of memorising the catalogues of all silicon vendors May 02 12:01:40 well, my catalogue is Linux kernel source code :) May 02 12:01:55 what's not in there, doesn't exist for me May 02 12:01:59 http://techcrunch.com/2011/12/05/99-chinese-tablet-is-mips-based-runs-android-4-0/ May 02 12:02:17 so, soon it will be certified May 02 12:02:32 and buyers will find out that no games that have native code work.... May 02 12:02:36 ynezz: 1) not all devices run linux, 2) many that do, run only with patched vendor kernels May 02 12:04:08 koen, hola May 02 12:04:19 bradfa: can you give http://dominion.thruhere.net/koen/angstrom/beaglebone/preempt/ a try? May 02 12:04:34 bradfa: I got passed some secret info on preempt problems May 02 12:04:40 oooo :) May 02 12:04:48 what tree can I apply them to? May 02 12:05:32 bradfa: any 3.2 based psp tree, but I based it on https://github.com/koenkooi/linux/tree/linux-ti33x-psp-3.2-r10d+gitre8004dad869568692ca2a45c04e8464ce48cc4d7 May 02 12:05:46 'linux-ti33x-psp-3.2-r10d+gitre8004dad869568692ca2a45c04e8464ce48cc4d7' is the branch name May 02 12:06:04 nice name May 02 12:06:21 koen, OK, will try that out today and let you know May 02 12:06:31 thanks! May 02 12:06:42 I also responded to Bas' mail on the list May 02 12:07:26 he's searching for the same limits in git as they are in eCryptfs May 02 12:07:48 _tasslehoff_: https://github.com/Angstrom-distribution/meta-ti/commit/81183b6f630616d7b61f45c861fbc965da7310ea May 02 12:08:10 koen, OK, will reply to list with results May 02 12:08:23 _tasslehoff_: in the end I just put the whole fs under git control to find out what changed between working and non-working May 02 12:08:28 bradfa: thanks! May 02 12:09:16 bradfa: I don't think it will fix preempt completely, but it should improve things May 02 12:09:36 I wonder how many minutes such a mips-based tablet will run without power cord May 02 12:10:16 koen, improvements are good :) May 02 12:10:55 koen, ok, improvements are very welcomed :) May 02 12:11:24 bradfa: AIUI irq handling moved from asm to C in between 3.1-psp and 3.2-psp May 02 12:11:33 bradfa: the conversion missed some things May 02 12:12:59 koen, ok, good to know May 02 12:13:15 <_tasslehoff_> koen: thanks. have you got any more info on that instant lockup? May 02 12:13:22 koen, so are you saying that arago 3.1 patches on top of 3.1-stable should NOT show issues with PREEMPT? May 02 12:14:11 _tasslehoff_: sadly not, but hopefully the 1.7 based ddk will fix it May 02 12:14:36 bradfa: I'm saying 3.2 is the "how can this have ever worked?" camp May 02 12:14:43 3.2-psp, that is May 02 12:19:40 funny name for a camp May 02 12:20:36 is it the bearded camp? May 02 12:21:14 koen: .32-psp and now 3.2-psp, I see a pattern May 02 12:22:04 There was one internal patch fixing this irq, recently created. May 02 12:33:02 Hi, do you know when I should expect to get my BBxm from digikey? I got a mail saying they had a 24week backlog. May 02 12:33:38 hvaibhav, If it's different than koen's, can you share? May 02 12:34:49 I am not aware of Koen's patch May 02 12:35:11 but I can share that with you. I will ask to push it to Arago May 02 12:35:15 hvaibhav, 2 patches -> http://dominion.thruhere.net/koen/angstrom/beaglebone/preempt/ May 02 12:35:33 hvaibhav, pushing to Arago would be very helpful, thanks! :) May 02 12:35:52 it's so sad that people still have to ask to get things pushed to arago May 02 12:36:25 but hey, if TI doesn't want help testing patches, so be it May 02 12:37:20 Is there a community visible mailing list for Arago kernels? I'm on arago-devel but most mails on that list relate to OE type issues and not kernels May 02 12:37:26 bradfa: hah! May 02 12:37:32 :) May 02 12:37:35 No not really May 02 12:37:36 bradfa: I asked about that May 02 12:37:47 bradfa: it's too much work, use the forums May 02 12:37:56 hvaibhav, If you can pull the ears of those uppers at TI who can make something like that happen, it'd be handy May 02 12:38:03 it is under internal review and I have seen today itself comment saying that "Acked by" May 02 12:38:19 koen, which forums? (and are we going to get back into the conversation about who has forums over mailing lists? :) May 02 12:38:26 beagle? May 02 12:38:56 there's linux-omap list and tree, but am33xx isn't quite up in there yet May 02 12:39:46 bradfa: see PM May 02 12:39:55 bradfa: its bug...in the baseport. May 02 12:40:37 and just FYI, the internal review which i am talking about, is internal mailing list for our group. May 02 12:41:08 hvaibhav, theoretical question, let's say I give TI 1 million dollars, could I get on that mailing list? :) May 02 12:41:58 bradfa: Ha...Ha... No ;) May 02 12:43:10 hvaibhav, worth asking :) May 02 12:43:29 bradfa: I looked at the Koen's patch, it is more or less same May 02 12:43:50 only thing is, our version is more generic, taking OMAP3 into account May 02 12:44:18 hvaibhav, ok, I'll watch arago tree for updates, thanks for your help May 02 12:44:57 Sending email to push it to Arago ASAP.... May 02 12:48:20 hvaibhav, thanks! May 02 12:48:54 Most welcome :) May 02 12:49:31 thanks hvaibhav May 02 12:50:51 jkridner: You are most welcome. May 02 13:00:35 koen: fyi, next update to bonescript will have an interactive javascript terminal in the browser and access to all the basic arduino-like functions, but with asynchronous returns. May 02 13:00:49 jkridner: nice May 02 13:01:09 I'm moving the pinmux read into a getPinMode function. May 02 13:01:11 jkridner: I'm working with the gateone author to fix the issues with that one as well May 02 13:01:23 cool. May 02 13:01:28 so hopefully we'll have 2 functional terminals in the end May 02 13:02:30 I'm on a call w/ a customer having challenges setting UUID on Bluetooth. May 02 13:03:20 koen, when you said mc would become available this afternoon, did you mean that I would simply be able to do "opkg install mc" ? or should I first do a "opkg update" ? Currently it still fails but afternoon is long of course :) May 02 13:03:43 update first May 02 13:03:50 ds2: have you ever set the UUID? May 02 13:04:02 ok, but uts perfectly normal that at this moment it doesn't work yet? May 02 13:13:06 koen: the code for eglGetDisplay is not something I can get to, right? May 02 13:13:45 ogre manages to open an X display, but eglGetDisplay fails, and I wonder why May 02 13:13:59 educa: try 'opkg install http://www.angstrom-distribution.org/feeds/v2012.05/ipk/eglibc/armv7a/base/mc_4.8.1-r3_armv7a.ipk' May 02 13:14:12 educa: the mirror network should have it in a few hours May 02 13:22:17 I am using beagle XM ver A3. In the expansion header It is shhowing a short between pin 1(VIO_1V8) and GND May 02 13:22:47 But board is working still May 02 13:23:09 When I try to read from the LCD controller CTRL (0x4830E004) register I get a segmentation fault, but if read from the USBSS REVREG (0x47400000) I don't. What am I missing? May 02 13:24:16 clocks May 02 13:24:30 Koen, this worked flawlessly, thank you May 02 13:24:37 I am using beagle XM ver A3. In the expansion header It is shhowing a short between pin 1(VIO_1V8) and GND. But the board is still working. Can some expalin the reason May 02 13:27:40 av500, sounds about right May 02 13:28:51 mdp: you hear the clocks ticking May 02 13:28:55 or not ticking May 02 13:29:12 I suppose May 02 13:30:45 How do I set the LCD_CLK? May 02 13:31:25 there is a clock framwork in the kernel May 02 13:32:32 So it's not a register? May 02 13:33:47 thomas_fogh: see http://www.mjmwired.net/kernel/Documentation/driver-model/platform.txt#96 May 02 13:34:06 it briefly mentions how clocks are kept off until requested by a driver May 02 13:34:50 if you look at drivers/video/da8xx-fb.c, you'll also see how that api is used to enable the lcd controller clocks May 02 13:35:56 av500, sorry to jump in, just that doc pointer was fresh in my mind. May 02 13:36:10 mdp: Thanks. I've been looking at da8xx-fb, but I was hoping I could test my display by writing directly to the HW registers... May 02 13:37:42 mdp: Not used to having an OS... ;) May 02 13:38:29 mdp: sorry for what? May 02 13:40:49 thomas_fogh: you *can* whack the ick/fck on manually too if you are just poking around. av500 and I were just making sure you understood that there's a framework that manages this for you…and is tightly couple with PM support May 02 13:41:36 thomas_fogh: implied within that is a warning that you can really hose things up if you don't use that framework since it thinks it has completely control of managing clocks May 02 13:42:16 av500, nm, had a brief bout with being overly polite…it doesn't happen often May 02 13:42:20 mdp: I'll try the correct way. Thanks! May 02 13:43:41 thomas_fogh, you could make a simple kernel module to clk_get the right right clocks…then poke around after that..to make your life easier May 02 13:43:45 have fun May 02 14:28:21 mdp: that kickstarter is done by the guy behind the displaylink linux support May 02 14:28:29 mdp: you might have run into him during ELC May 02 14:28:52 coolio May 02 14:29:12 koen: ah, that makes an instant business case then May 02 14:29:35 yeah, makes things much clearer May 02 14:30:43 usb makes me sad though May 02 14:30:46 instantly sad May 02 14:30:57 I have a few of his old docks to improve multiseat in angstrom May 02 14:32:06 koen: its funny, up until a few days ago, I associated "seats" with what you multiply an insanely high price with when licensing commercial SW May 02 14:34:02 Hello, I've been working with the BeagleBoard-xM for a little while now with a pre-built kernel image. I now would like to create my own kernel image. I have been following: http://blog.galemin.com/author/mgalemin/ When I hit make, I end up getting a uImage, however, it ends with an "EPOLL_CLOEXEC" undeclared error and errors recursively. The uImage works, but is slightly smaller than expected and I believe it lacks some of the d May 02 14:35:05 I am using the codesourcery toolchain arm-2009q1 to build. Any advice? I'm unfamiliar with this error and not quite sure what to do next to debug the situation. May 02 14:35:17 koen, I just keep seeing my Uni sparcstation slc diskless configuration when I look at that system May 02 14:35:22 * mdp feels old now May 02 14:35:44 koen, further proof that there are no new ideas :) May 02 14:37:04 too bad that today even a single website can eat all your CPU, so hosting 16 seats is daring :) May 02 14:38:09 heh May 02 14:38:34 Marc_SSL: erm, a 3 year old toolchain? May 02 14:38:59 some people drink 30 year old whiskey May 02 14:39:28 av500: every year it tastes less like meths as the alcohol evaporates :-D May 02 14:39:32 XorA: I had the same doubts, but on his website, he said any toolchain could work if the setup is reconfigured, which I did. Do you think updating to 2011 would solve the problem? May 02 14:40:13 Marc_SSL: it does sound like that might possibly a new feature, if your on ubuntu try their excellent toolchain May 02 14:43:13 mdp: I use hp envizex x terminals to elevate the espresso machine to the proper height nowadays :) May 02 14:43:24 XorA: Will do, I'll look into it. Though most sources I have read up on have recommended the CodeSourcery toolchain. (Stupid question:) Can the Ubuntu native toolchain cross-compile to an ARM device? May 02 14:43:36 koen: what I dont get, why does he not just simply use r-pis? May 02 14:44:16 Marc_SSL: apt-cache search eabi and you will see their cross toolchain May 02 14:45:59 I think TI might own shares in CodeSourcery or something, seem to write broken code that requires their toolchains May 02 14:46:16 * XorA looks at TIers for confirmation :-D May 02 14:46:53 Haha, alright. I entered that command and didn't see any cross-compilers, I might have to download the package May 02 14:47:20 Marc_SSL: on an old LTS release? May 02 14:48:03 10.04.2 LTS May 02 14:48:20 ah right, possibly might be a PPA for that, not sure May 02 14:48:32 I think the cross toolchains appeared in 11.04 May 02 14:48:46 maybe 10.10 May 02 14:49:04 Hmm, alright, I think I might try updating CodeSourcery before updating my kernel :D May 02 14:49:08 haha May 02 14:49:22 I ue 2010q1-202 a lot May 02 14:49:32 For BBxM development? May 02 14:49:41 for omap3/4 stuffs May 02 14:50:11 Does it have any noticeable bugs? May 02 14:51:14 Marc_SSL: not that I fallen over yet May 02 14:52:02 XorA: Alright, I'll have to check it out. Thanks XorA for the great help, I'm still pretty new to Linux and the OMAP, and every day is still learning experience. May 02 14:52:16 Marc_SSL: https://launchpad.net/~linaro-maintainers/+archive/toolchain your PPA May 02 14:53:38 XorA: Do you think this would work better than CodeSourcery? May 02 14:54:05 poking yourself in the eye works better than codesourcery May 02 14:54:12 Marc_SSL: I always use this toolchain for mainstream kernel work! May 02 14:54:48 koen: Hahaha May 02 14:54:49 Marc_SSL: and its that famous ex OE guy hrw who maintains it :-D May 02 14:55:16 XorA: Alright, I think I'll give it a go! Now, can I ask one more really stupid question? May 02 14:55:30 Marc_SSL: we charge extra for the really stupid ones :-D May 02 14:56:59 XorA: Hahaha, alright, well on the LeopardBoard 365, you had to re-make when you added code to the board, on the Beagleboard-xM, or OMAP system I suppose, is it possible to add code to the file system without having to rebuild a kernel image? Perhaps build the code natively if the tools are loaded in the image? May 02 14:59:07 motorola saves germany from windows 7 :-D May 02 14:59:20 yippie May 02 15:02:26 XorA: you wish May 02 15:03:00 av500: :-D May 02 15:03:45 Marc_SSL: if your using Angstrom or Ubuntu you can install new packages, or install native toolchain on the board and compile there May 02 15:03:57 Marc_SSL: I wouldnt compile anything large on an XM though May 02 15:08:42 XorA: I'm using Angstrom. I'm guessing to install new packages, that would have to be over ethernet? I'm planning on compiling only some C code, probably nothing very large, a hundred or so lines of code at most. May 02 15:09:02 that should be all cool May 02 15:09:40 cool code May 02 15:10:32 Coolness May 02 15:10:37 Thanks again XorA May 02 15:10:51 This has been awesome advice May 02 15:11:11 * av500 renames this the Cool Club May 02 15:18:34 first rule of Cool Club... May 02 15:19:48 don't talk about Cool Club? May 02 15:20:00 bradfa: you are out May 02 15:20:02 oh, I broke the rul May 02 15:51:40 koen, looking at the patch you posted on the preempt thread, does that mean irqs from bank 3 were not processed at all? May 02 15:52:09 jsabeaudry: I actually have no idea, I just made the patch based on a email sent to me :) May 02 15:52:23 jsabeaudry: yeah, that looks very strange May 02 15:52:51 jsabeaudry: I'll go with "I wouldnt be suprised if that were the case" May 02 15:53:21 I can't seem to find in the TRM what maps to INTC_PENDING_IRQ3 May 02 16:06:35 well the good news is, ethernet works with a preempt kernel now May 02 16:07:07 ship it May 02 16:07:34 bradfa, we can say "Coolness Club" and not break the rule though..but you're already out May 02 16:17:26 mdp, I'll make my own club, maybe I can be in the "Still uses GCC 4.4" club :) May 02 16:20:18 * mdp shuns bradfa May 02 16:20:44 that compiler is starting to really stink May 02 16:20:55 you probably IPL your beaglebone with a big bank of toggle switches and blinky lights too? :) May 02 16:24:23 windows rt is starting to look better and better May 02 16:27:39 Is there some kind of performancecounter on the beaglebone which is accessible from C ? May 02 16:32:13 I _LOVE_ toggle switches! :) May 02 16:32:46 don't we all :) May 02 16:32:59 and blinkenlights May 02 16:33:02 hehe May 02 16:33:15 educa, what do you mean by performance counter? May 02 16:33:39 you should watch some of the adafruit new product videos…it's hilarious to see Limor get excited over a new SPDT switch when they add it to the store. May 02 16:33:58 "We've had push button switches, but now we have a breadboard capable toggle switch!!!!!" May 02 16:34:07 that's now on my todo list! May 02 16:34:08 * mdp bounces May 02 16:34:23 great people, it's just kinda funny May 02 16:34:38 mdp: trying to seem bigger? May 02 16:34:45 hehe May 02 16:44:33 I'm getting a wierd situation on my beaglebone, basically after a few minutes of operation, the heartbeat lead stops blinking, the dma stop working (but I still get DMA_COMPLETE interrupt but i see nothing on the scope), cat /proc/kmsg doesn't print anything anymore May 02 16:44:50 does that ring a bell to anyone regarding what could be affected May 02 16:46:38 jsabeaudry, no. What kernel are you running? May 02 17:06:16 bradfa, I mean that I would like to have some kind of high resolution counter which I can read from a gcc program on the beaglebone. I found that they call this performancecounters which count the number of cpu ticks elapsed. Based on that you can do precise time measurement between 2 periods by getting tick count, do some stuff and get tick count again. Difference in ticks is high resolution duration of your routine May 02 17:24:55 bradfa, koen's r10d May 02 17:25:56 educa, getnanotime? May 02 17:26:10 thats available ? May 02 17:26:20 would be very nice to have, I'll check May 02 17:28:51 clock_gettime May 02 17:29:13 what resolution does that have ? May 02 17:29:54 it gives you access to low level stuff May 02 17:31:38 So regarding my between-life-and-death beaglebone status, both top and htop do not work anymore, but I can still communicate with the bone through ssh May 02 17:34:26 jsabeaudry, hmm, I dunno, I haven't gotten to run koen's r10d yet, but that does sound awfully weird May 02 17:34:39 when I do, I'll let it run for a while and see if I can reproduce May 02 17:34:45 how long after boot do you see this? May 02 17:35:19 bradfa, i still have a hard time to reproduce it, sometimes 5 minutes sometimes 1 hour May 02 17:36:05 Russ, maybe stupid question but I'm not in tme.h looking at the clock_getres and clock_gettime which seems to return an INT, but I don't really understand the parameters of that function. It asks for a clockID and such? Or are these all optional params ? May 02 17:36:43 educa, you haven't found manpages yet? Here, let me make your life much much easier May 02 17:36:46 man clock_gettime May 02 17:36:49 I meant I'm NOW in time.h May 02 17:36:50 only happens if I use it at high capacity, (basically my process dma's from gpmc to memory then sends the data over http at around 2MB/s) May 02 17:39:34 oops, -sh: man: command not found May 02 17:39:56 I run on the sd that came with my beagleboard revA5A May 02 17:40:04 beagleBONE I meant May 02 17:41:01 jsabeaudry, Ok, not sure I have a test case like that, but I'll try to heavily load a Bone with koen's kernel May 02 17:42:21 Hi, culd anyone tell me why opkg update; opg upgrade broke the beaglebone (aka not booting)? May 02 17:42:46 educa, http://linux.die.net/man/3/clock_gettime May 02 17:42:49 i thought we had a repository for the bone :D May 02 17:43:16 isbric, how do you know it's not booting? May 02 17:43:19 hmm, angstrom shipswithout manpages? May 02 17:43:48 I tend to avoid opkg upgrade... tends to cause problems May 02 17:44:14 angstrom ships without man afaik May 02 17:44:17 jsabeaudry: ah, smart question, it boots but all the network interfaces sems to have some problem May 02 17:44:58 so, i realy shuld stick to the prebuild images or even create a new image myself eveytime i want some updates? May 02 17:45:11 a system without manpages is unusable May 02 17:45:35 aholler: what manpage are you looking for? May 02 17:45:48 I don't, I have them all ;) May 02 17:46:01 if you need a specific update grab it, just don't do a blanket upgrade May 02 17:47:15 isbric, depends on your needs, what do you do with the beaglebone? May 02 17:47:25 is there any slim way to interface with the bone once blank upgrade has been done? May 02 17:47:50 isbric, the serial usb? May 02 17:48:06 jsabeaudry: im going to use it with a robot, but mostly interfacing with microcontrollers over seral link May 02 17:48:27 jsabeaudry: ftdi driver doesnt recognise it anymore. May 02 17:48:40 not as a disk nor a network interface. May 02 17:48:54 <_av500_> dont update anything unless you need it May 02 17:49:33 i needed somw extender wireless driver support, was to lazy to find out what pkg to update May 02 17:50:00 isbric, try out arch linux arm you might like it May 02 17:51:50 jsabeaudry: got a link? May 02 17:53:07 isbric, sure: http://lmgtfy.com/?q=arch+linux+arm&l=1 May 02 17:53:20 xD May 02 17:53:33 have you tryed rowboat? May 02 17:55:25 life is but a dream May 02 17:56:14 http://code.google.com/p/rowboat/ May 02 17:58:00 rowboat is old and unmaintained May 02 17:59:13 oh, i never got around to try it May 02 17:59:51 jsabeaudry: arch linux sounds just like what im looking for. May 02 18:03:51 even the web server works on my zombie beaglebone May 02 18:04:29 zombie? May 02 18:05:14 the heartbeat led has stopped flashing and a lot of stuff is not working anymore, (If I reboot it all is fine), and some stuff still work May 02 18:06:19 jsabeaudry: dont tell me made love to it, i sertenly had to force myself not to :) such qute hardware! May 02 18:06:32 cpufreq-set hangs! May 02 18:06:37 jsabeaudry: dmesg? May 02 18:07:50 burrito! May 02 18:08:06 aholler, it would seem that the beginning of dmesg is truncated May 02 18:08:38 increase the buffer-size, there is a kernel-option for that May 02 18:09:34 aholler, oh thanks, I thought the truncation was corruption of the buffer or something May 02 18:09:46 ham + swiss cheese + lettuce and a peanut butter sandwich for dessert :) May 02 18:09:58 jsabeaudry: why do you think it is? May 02 18:10:37 is the buffer in his terminal or on dmesg? May 02 18:11:25 the log behind dmesg uses a circular buffer May 02 18:12:31 and some configs have sizes from times when the kernel wasn't really communicative May 02 18:13:17 or from people which wanted to save the last kb ;) May 02 18:13:18 increased it from 128kb to 1mb, now I just need to reproduce the issue again May 02 18:15:12 also the reboot command wont acually reboot the system when I'm in zombie mode May 02 18:15:28 jsabeaudry: aim for the head May 02 18:16:02 jsabeaudry: that i have experianced aswell, with Rev A5 May 02 18:16:41 actually, peanut butter + honey sandwich. May 02 18:18:46 I'm trying to get Ogre up and running, but eglGetDisplay fails connecting to the X-display that seemed to open ok. Any EGL/PowerVR wizards in da house? May 02 18:32:11 backup plan: koen: do you remember a date or commit when ogre was working in oe-classic? May 02 18:32:30 years ago, ogre was working fine May 02 18:34:36 fee fie foe fum May 02 18:34:55 dum dee da da dum May 02 18:35:08 drum roll . . . May 02 18:36:30 jsabeaudry, when you get into zombie mode, do you get an oops? May 02 18:37:21 bradfa, I couldnt see because my log buffer wasnt big enough, I made it bigger now I'm trying to reproduce May 02 18:37:29 k May 02 18:38:16 I'm going to start some testing now with koen's patches. I want to reproduce my kernel panics on shutdown first, not sure how successful I'll be with that... May 02 18:38:24 we'll see May 02 18:39:22 bradfa, are you Bas Laarhoven? May 02 18:39:31 no, I'm Andrew Bradford May 02 18:39:40 Bas Laarhoven is ncbas May 02 18:39:42 ah, so it was an ogre jack met May 02 18:39:54 he's not online May 02 18:39:54 whois is friendly :) May 02 18:44:03 bradfa, koen's patches fixed the ethernet in my case very curious what kind of results you get May 02 18:44:49 jsabeaudry, OK, good. I'm building some older Arago kernel right now, I want to find the kernel config that caused me all sorts of issues, then use same config for koen's May 02 18:44:54 tasslehoff: correlate it with my youtube vids :) May 02 18:45:55 koen: good thinking. I was just about to check when the recipes came to life. figured it was working then :) May 02 18:47:40 can't find my ogre SD image to give you a better idea of the date May 02 18:53:58 Is there a quick way of grabbing the 3.0.8 kernel source for Ångström? I'm staring at this link - http://www.angstrom-distribution.org/repo/?pkgname=kernel-3.0.8%2B May 02 19:07:11 ds2: thanks anyway. I'll try to find some other sw that uses the sgx and see if ogre is holding it wrong May 02 19:25:25 ok, so I can reproduce dhcp not receiving DHCP Offers with PREEMPT on Arago (my original problem) May 02 19:38:39 oh uh it is kkeller May 02 19:39:00 hey mranostay May 02 20:01:50 bradfa, I don't get any oops when I get into zombie mode May 02 20:02:00 ok May 02 20:02:20 have you tried with koen's patches? May 02 20:02:32 jsabeaudry, not yet, still working on arago stuff May 02 20:02:48 jsabeaudry, I can reproduce the not accepting dhcp offers when PREEMPT is enabled May 02 20:02:55 about to build koen kernel May 02 20:03:12 Good thing that it still reproes May 02 20:04:52 I'm making gists of my arago non working kernels May 02 20:05:08 I can't reproduce any oops though, which kind of stinks May 02 20:10:48 koen's patches don't fix me not accepting DHCP offer :( May 02 20:10:55 same behavior as without May 02 20:11:35 I can't get an oops like ncbas in recent list pastebin, though May 02 20:11:48 could nfs root play into it at all? May 02 20:12:00 I have root on sd card May 02 20:20:30 wait a moment, disregard my statement about koen's patches not working. I built wrong kernel May 02 20:20:38 * bradfa trying again with proper sources May 02 20:26:14 I apologize in advance, but I can't seem to find on google how to upgrade the version of Angstom I am running on my beagleboard-xm. I want to install cloud9, but from what I have found I need to update to version v2012.01 (I have 2010.7-test). I build the image using Narcissus. I tried 'opkg upgrade', but that did nothing. Advice? May 02 20:27:26 jsabeaudry, try opkg update before you upgrade May 02 20:27:59 Yes - I did do that first. should have mentioned that. May 02 20:29:34 you cant update 2010-2012 May 02 20:30:53 *confused* ok - so how does one get v2012? May 02 20:31:06 create a new image from narcissus May 02 20:34:29 What do I need to select in order to get a newer version? I am assuming complexity:advanced then version 2011.03 (which I assume is stable), core-eglibc, & next. -> select either core-eglibc or next? May 02 20:35:52 next-eglibc I think May 02 20:36:08 probably should kick koen to update the labels with new release :-D May 02 20:36:19 ok - I will give it a shot! Thanks for your help XorA! May 02 20:37:11 What could cause top and htop to stop working? May 02 20:37:24 /proc not mounted May 02 20:44:42 XorA, it's still mounted May 02 20:45:04 no idea then May 02 20:46:37 jsabeaudry, top uses /proc, no? May 02 20:46:56 kernel provides /proc May 02 20:47:03 maybe? May 02 20:47:57 jsabeaudry, I don't have time to run stress tests today, but tl;dr is that koen's patches don't fix my DHCP offer acceptance. Mail going to list shortly. May 02 20:52:45 bradfa, that is so surprising there must be 2 issues May 02 20:53:50 jsabeaudry: when the heartbeat stops, something serious is broken, this is just a timer May 02 20:54:38 that thing even used to run after a shutdown May 02 20:54:43 aholler, yup seomthing very serious is broken, yet so hard to figure out what or why May 02 20:57:30 I can't afford to spend that much more time on this this week. Hopefully next week will be better. May 02 20:59:36 Can someone post their /proc/timer_list on beaglebone, (i'd rather not reboot since it's so hard to get in zombie mode) mine looks broken http://pastebin.ca/2143278 May 02 21:02:42 I don't know if the heartbeat led is based on gp_timer but there are no more interrupts on that timer May 02 21:04:51 I went back to Narcissus to build an Angstrom version based on core-eglibc. I followed my instructions when I used the '2011-03' stable version. With the core-eglibc version, I end up with a u-boot.img file & not a u-boot.bin file. I copied over the .img file, but when I attempt to boot up the beagleboard-xm, it hangs with a error - 'can't find u-boot.bin'. Any suggestions? May 02 21:05:20 newer MLO loads u-boot.img I think May 02 21:05:42 but renaming .img to .bin should also work May 02 21:06:00 I will try that.... brb. May 02 21:07:58 dang - it just hangs at "Loading u-boot.bin from mmc". May 02 21:09:32 BTW - I am using the MLO that gets generated along side everything else by selecting: Bootloader Files (x-load/u-boot/scripts). Maybe I need to go find a newer MLO file from somewhere else?!? May 02 21:09:49 confusing then May 02 21:10:07 * XorA hasnt played with his bone much recently May 02 21:10:18 something could be wrong with image I guess May 02 21:10:30 or my XM May 02 21:10:32 I want to say that I had this problem in the past which is why I went back to the 2011.03 branch. May 02 21:11:17 jsabeaudry, http://pastebin.com/6Dfw5iY6 May 02 21:14:04 mdp, thank you, my value for next_event looks like it is broken May 02 21:14:55 There is a 199.4485 factor between yours and mine May 02 21:16:37 Is there a way to restart gp_timer or soemthing similar? May 02 21:32:04 mdp, have you had more crashes with the dma? May 02 21:32:59 it's not my project so I'm not sure…I mostly just internally consult :) May 02 21:33:09 and I haven't heard back from the customer May 02 21:33:17 that means everything's great! May 02 21:38:31 or the customer is locked and loaded ;) May 02 21:39:49 mdp: maybe they are using a AT91SAM part now ;) May 02 21:42:39 by "customer" I mean team member..so I hope he isn't using AT91SAM! :) May 02 21:42:40 haha May 02 21:43:34 let's just say that the call has gone into design verification for test vectors…to me, that's usually a negative indicator May 02 21:56:55 What is susceptible to break the gp_timer? May 02 22:00:05 * XorA is a dimwit May 02 22:16:39 Could overheating lead to a gp_timer that does not work anymore? May 02 22:17:04 <_av500_> timer12? May 02 22:18:09 that one does not exist May 02 22:40:40 echo "22" > /sys/class/gpio/export May 03 02:51:09 I can't get my beaglebone to connect to the internet via an ethernet to my computer May 03 02:51:35 I have Angstrom on the beaglebone and ubuntu on the computer **** ENDING LOGGING AT Thu May 03 03:00:01 2012