pbuckleyfor compiling on panda, mount /tmp as tmpfs definitely seems like a win00:11
pbuckley(vs having it write out to an sd card as in my setup)00:12
GrueMasteryes, that also helps.  :P00:19
GrueMasterAlthough a USB drive is still better overall.00:19
pbuckleyyeh.. ill bring one in tomorrow00:20
pbuckleywas hoping to be able to survive off just the sd00:20
pbuckleybut thats not really being realistic at this point00:20
pbuckleymight try the iscsi thing tonight00:20
GrueMasterPersonally, I think the SD is good for running shiny demos.  Not really fast enough for serious development work.00:20
GrueMasterNFS-Root also works w/o the iSCSI overhead.00:21
pbuckleyaye.. i think my experience agrees with your statement00:21
pbuckleyk.. ill try that first00:21
pbuckleythank you :)00:21
=== Lopi__ is now known as Lopi
=== jkridner_ is now known as jkridner
Person987Does anyone know if it could be harmful to plug a monitor into an already running pandaboard?04:25
twbI have no idea, but I would hope not!04:28
Person987I'm embedding one in a robot and I'd like to be able to just hook up a monitor if I need to without turning it off.  Normally it won't have a monitor on it.04:30
dioxinGrueMaster: you there? I'm getting an error with my local repo, apt-get appears to be looking in the wrong place :(10:01
dioxindeb oneiric main restricted universe multiverse is whats in my sources.list10:01
=== chuck_ is now known as zul
=== rbelem_ is now known as rbelem
=== Pratik is now known as Pratik_away
=== dduffey_afk is now known as dduffey
=== KRF_ is now known as KRF
=== shadeslayer_ is now known as shadeslayer
=== Ursinha is now known as Ursinha-lunch
GrueMasterdioxin: Here now.  You can start checking your server configuration by using a browser from another system to see if that address shows up.  If you don't see anything at something is wrong with your server config.15:35
Snarkcould people check if they are hit by https://bugs.launchpad.net/ubuntu/+source/gcc-4.6/+bug/926855 too ?15:41
ubot2Launchpad bug 926855 in gcc-4.6 "[ARMhf] gcc produces assembler it can't compile" [Undecided,New]15:41
GrueMasterSnark: It compiles fine here with "-marm", which I believe is default on all of our projects.  janimo, is that correct?15:46
SnarkGrueMaster, yes, that is in a comment15:46
Snarkthe problem is that it doesn't compile without, while it did before15:46
GrueMasterYes, I know.  I read the bug.  I think -marm is a required setting (which is why I asked janimo for confirmation).15:47
SnarkGrueMaster, uh... if it were required, why isn't it on by default?15:56
GrueMasterSnark: I don't know the specifics, just repeating from memory.  Mainly I run the tests, thats why I pinged another member of our team for more detail.15:57
Snarkthe same source code was building on oneiric, and stops building on precise, that is why I think it's a bug on the compiler side...15:58
GrueMasterSnark: It could easily be an armhf specific issue too.  Can you try in precise armel?  (if not, I can in a little bit).15:59
SnarkGrueMaster, I don't have a precise armel install16:02
GrueMasterOk.  I can reimage a spare panda here.16:03
* Snark doesn't have spares16:05
SnarkI make experiments with the same device I use for my work...16:05
dioxinGrueMaster: the website shows up, but it only has ls-lR.gz and pool/16:37
GrueMasterThen something happened and it isn't getting dists/16:37
GrueMasterCheck your ubumirror.conf16:38
dioxinwhich one?16:38
dioxinGrueMaster: my ubumirror.conf file in /etc/ubumirror.conf is a carbon copy of yours16:41
GrueMasterCheck your latest /var/log/ubumirror/ubuports.log* to see if it indicates any issues.  The app is supposed to pull updates from ./pool and then everything but ./pool.16:42
GrueMasterYou should see this:16:43
GrueMastertobin@mirror:/server/data/www$ ls ubuntu-ports16:43
GrueMasterdists  indices  ls-lR.gz  pool  project16:43
=== Ursinha-lunch is now known as Ursinha
GrueMasterWhich version of ubumirror are you running?  dpkg-query -l ubumirror16:46
dioxin[receiver] io timeout after 601 seconds -- exiting16:47
dioxinfor version of ubumirror16:47
GrueMasterlet me check mine and make sure I didn't modify it locally.16:48
GrueMasterThe only script I have changed was ubucdimage (has nothing to do with your issue).16:57
GrueMasterAnd 0.3ubuntu3 has been the latest since 10.04.16:57
dioxinrunning the following command as root atm /usr/bin/ubuports >> /var/log/ubumirror/cron.log16:58
dioxinand I've got the ubuport.log tail-f 'ed16:58
dioxincurrently at pool/universe/t16:58
GrueMasterAfter it finishes with pool/*, it should show: Ubuntu ports pool sync completed.16:59
GrueMasterFollowed by:  Initiating Ubuntu ports dists and indices sync...17:00
GrueMasterSnark: I updated the bug report with my test results.17:07
GrueMasterCan you ping someone in #linaro and see what they think?  They do a lot of the arm optimization to gcc.17:09
SnarkGrueMaster, I see17:28
=== jkridner__ is now known as jkridner
=== Pratik_away is now known as Pratik
SnarkGrueMaster, I was pointed to Michael Hope, which wasn't available, so I sent him a mail18:09
GrueMasterSorry about the run around.  I'm just trying to get you hooked up with someone that would know more about this that I do.18:10
SnarkI'm used to getting run around with a cannon... you're a soft one!18:13
janimoGrueMaster, -mthumb is our default, -marm only for a few cases where package got miscompiled18:18
GrueMasterAh, there you are.  What would you suggest in this case where it builds on oneiric but not precise w/o -marm?  (I haven't tested with -mthumb)18:19
janimoSnark, IIRC I saw similar errors where huge assembly code was generated by -funroll-loops, so that in thumb mode it could not be directly jumped too. I only vaguely remember though18:20
janimoGrueMaster, it probably is a gcc issue18:20
janimogood that Linaro GCC is a bugtask on this18:20
GrueMasterYea, failed with -mthumb,18:20
janimoSnark, as a work-around if you need the package on precise ASAP build with either -marm or try without funroll-loops18:21
janimootherwise we may need to wait for a gcc fix18:21
janimoSnark, or try forcing a build with gcc-4.5 which was the default in oneiric to see it is indeed a regression18:21
janimobut in either case these are workarounds and a proper fix in gcc is needed I'd say18:22
dioxinGrueMaster: the repo appears to be working now, no idea what was wrong with it :(18:34
GrueMasterI'm not sure why, but I had to run ubuports 3 times before I had a working mirror.  Glad to hear you are up and running.  I still recommend a 2 hour update period.18:35
GrueMasterEspecially during times were we rebuild the entire pool (which I think is currently happening).18:36
jomHi everybody! owner of a Pandaboard Rev.A4 I get stuck with a ubuntu11.10. Monitor, mouse & kbd are connected and detected but it seems that a problem while resising and mounting the filesystem append I can logging in in manual but what I am suppose to do to fix it? Does the activation of the console help finding a way through this issue? thanks!20:23
GrueMasterjom: Which image are you using?20:24
jomthis one : ubuntu-11.10-preinstalled-desktop-armel+omap4.img20:27
GrueMasterjom: Did you check the md5sum of the image?  What is the error you are seeing?20:28
kurtankIm a bit new to linux, and need a little help. Im using Ubuntu 10.10 running via a 4GB image in chroot next to android on a samsung Galaxy SII.20:28
kurtankI was wondering if there was any place i could get ARM-specific applications. Most of the app center doesnt work.20:29
GrueMasterkurtank: You should be able to apt-get install any application in the ubuntu pool.  I would recommend upgrading to at least 11.10, although you will see a marked performance increase if you were to start with 12.04 armhf.20:31
GrueMasterYou can download a pre-built core chroot environment from http://cdimage.ubuntu.com/ubuntu-core.20:31
kurtankAbout upgrading: many users report that 11.10 runs slowly and breaks the image.20:31
kurtankI think its also worth mentioning that the only way to interact with ubuntu with this method is through VNC.20:32
GrueMasterWell, I am not very familiar with your usage model unfortunately.20:33
kurtankIm not surprised.20:33
GrueMasterAh, similar to the "nookbuntu" image.20:33
kurtankIts more of a hack than anythint.20:33
jomThe MD5 was correct. The error is "Errors were found while checking the disk drive for/." I zeroed the card with 'dd' ans unfortunatly it doesn't work better at this point20:35
GrueMasterWhat is the card brand/size/class?  You may have a bad card.20:36
jomYes you right I think it could have been that but I previously run an Angstrom distro on it and everything was Ok. The card is a microSD SANDISK/4GB/420:38
GrueMasterAngstrom doesn't do much except basic diag.  It doesn't resize the image, and isn't as complete.20:39
kurtankhttp://forum.xda-developers.com/showthread.php?t=1223128 This is the method and the image i used. Oddly, xchat irc downloaded ans installes fine.20:39
kurtank*and installed20:39
kurtankSomeone also said it would be possible to get sound output, and USB host functions apparantly work. Autodetect will never work though and requires manual detection. Ive been using linux for about only a week and i have no idea how to do that.20:42
jomI agree with you GrueMaster about angstrom distro. I don't want to start here a Troll but Angstrom it really nice when you need just a console coz it has a very low memory sized and very efficient. A tool for every needs ;o)20:42
jomCould you explain why the image is then resized I don't get the point on that20:43
GrueMasterOh, don't get me wrong, it is an excellent diagnostic image.20:43
GrueMasterThe image is pregenerated on our servers.  We needed a way to get the image on a variety of SD cards w/o knowing the size of the card in advance.  This setup will boot the kernel and initrd, and a script in the initrd will repartition the rootfs and resize it to fill your SD (4G-32G tested).20:45
GrueMasterEasier than having one image for each size SD on the market.20:45
dioxin_GrueMaster, do you know of any images for a beaglebone?20:46
=== Pratik is now known as Pratik_away
GrueMasterdioxin_: I don't have a beaglebone, but I have been told that the omap images should work.20:47
GrueMasterI think that system is designed for headless, so I would recommend either the preinstalled-server or netboot (exercise your mirror a little).20:47
dioxin_I havent managed to get any display out of the serial port tho :(20:48
GrueMasterNot even with their stock angstrom image?20:48
dioxin_yes with the stock, but not with the ubuntu 11.10 server omap320:49
GrueMasterdioxin_: Try one of the images at http://cdimage.ubuntu.com/ubuntu-server/daily-preinstalled for omap (not omap4).20:52
GrueMasterOdd that you don't see anything, not even u-boot.20:52
dioxin_http://elinux.org/BeagleBoardUbuntu#Demo_Image trying this atm, but I'll let you know about the link you gave20:52
GrueMasterUnfortunately, I don't have one (and no budget for one atm).20:52
dioxin_I'll keep you posted :D20:53
GrueMasterNote that I won't be able to help at all on images we don't produce on cdimage.ubuntu.com.20:55
jomBut when I enter the Manual recovery I can see that the '/' directory is mounted on the SD card and a 'mount' cmd shows all the necessary diretories any idea?20:55
dioxin_i know :)20:55
GrueMasterLooking over that link briefly, it looks like beaglebone support was added in linux 3.2 (which our precise images will have).  Can't say about u-boot though.20:56
jomwhen we are on the manual recovery mode what can we do to make the system finish is resizing properly?21:01
=== jkridner___ is now known as jkridner
GrueMasterjom, I would try reflashing and start over.21:19
GrueMasterThe whole process shouldn't take more than 10 minutes on a 4G card.21:19
jomHum I did it already 2 times and I always step down on the same issue. I will try to find another 4GB card.21:21
GrueMasterWhat method are you using to flash the image to SD.  Keep in mind the image is gzipped to begin with.21:23
GrueMasterI have a script that essentially runs "gunzip < precise-preinstalled-desktop-armhf+omap4.img.gz |sudo dd of=/dev/sdf bs=4M".  The 4M helps with the speed of writing to a flash device (and reduces the issues you seem to be seeing).21:25
jomI 1st unzip it and I transfert it to the SD device not mounted using 'dd'21:26
jomit's exactly what I do21:26
GrueMasterI have been testing these images on a variety of SD cards from 4G to 16G, class 4, 6, and 10.  That's why I ask.21:27
jomI try to do it once again else I will have to buy another card tomorrow. Do you use uSD HC card? And does it work the same?21:44
GrueMasterI have a few uSD cards, yes.  I use them mainly for beagleXM and Freescale Quickstart testing, but the image is identical (except for kernels).21:48
dioxinGrueMaster: I finally managed to boot into ubuntu on the beagleBone, but I used the elinux link21:52
dioxinI think the problem lies in the tty output21:52
GrueMasterWhat does dmesg show for UARTS?21:53
dioxinI dont think it supports UARTS, I'm using a mini-usb cable for the console access21:53
GrueMasterThe serial console "should" be /dev/ttyO2 on almost all omap hw.21:53
GrueMasterOh, interesting.21:54
GrueMasterWhat is /proc/cmdline say for console?21:54
dioxinttyO0 I think21:59
GrueMasterAh, then you should be able to modify the boot.scr on the normal ubuntu image to point to that instead of ttyO2.21:59
GrueMasterYou will need to strip the first 72 bytes from boot.scr, then edit the bootargs and rerun mkimage to regenerate boot.scr.  You can't just edit the file.22:00
dioxinGrueMaster: do you have a guide for that process?22:25
GrueMasterNot exactly.  Here are the steps (assuming the boot partition is mounted on /mnt):22:27
GrueMasterdd bs=72 skip=1 if=/mnt/boot.scr of=boot.script22:27
TheMusodioxin: dd should be able to strip it for you.22:27
GrueMasteredit boot.script22:27
GrueMastersudo mkimage -A arm -O linux -T script -C none -d boot.script /mnt/boot.scr22:28
GrueMaster(I leave the "edit boot.script" part for you to figure out).22:28
jomabout Ubuntu 11.10 it felt on the same issue!22:31
dioxinok, my first attempts at editing the boot.script failed... so I'm going to cheat :D22:42
dioxinhmm... the elinux imaging process doesnt have a boot.script22:47
jom:o)) get it works! my SANDISK SD card is on a very bad shape the TDK one make it works! So that was just a hardware problem [SOLVED]23:40
jomBye & thanks to you GrueMaster for your help!23:40
TheMusoogra_: In your request for the tegra ALC patches for alsa-lib, you mentioned related patches. What do you mean by this exactly? There is the patch you linked to, but I don't see any others in that thread...23:58

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!