/srv/irclogs.ubuntu.com/2011/03/01/#ubuntu-arm.txt

=== doko_ is now known as doko
=== asac_ is now known as asac
=== lag` is now known as lag
XorAmorning08:55
noah1989hi10:06
noah1989i want to run ubuntu on the beagleboard xM10:06
noah1989i#d prefer a prebuilt sd-card image10:07
noah1989is there anything recent that i can use?10:07
XorAthere is natty alpha-2 images10:08
janimonoah1989, Ubuntu 10.10 release or a few weeks old Natty image10:08
noah1989i found the 10.10 at http://cdimage.ubuntu.com/ubuntu-netbook/ports/releases/maverick/release/10:08
noah1989where do i find the natty images?10:08
janimothe omap one10:08
janimocdimage.ubuntu.con/ubuntu-netbook I think10:09
noah1989ah. found it, thanks10:09
noah1989is there a way to modify it so that it uses the s-video output by default?10:10
janimoI don't know10:10
noah1989i'm gonna buy a dvi-d cabe just in case...10:11
noah1989ah.. found it: http://elinux.org/BeagleBoardUbuntu#S-Video10:13
noah1989also i need a simple window manager10:16
noah1989preferably with all-fullscreen windows and simple switching10:17
noah1989side by side doesnt make sense at 640x48010:17
noah1989well maybe two xterms may be useful. i'll stick with awesome10:21
=== zul__ is now known as zul
dokoogra: ping12:04
ogradoko, yo12:10
ericb2hello12:11
dokoogra: I don't have a working babbage board anymore. could you have a look at bug #605042 (and test it)12:11
ubot2Launchpad bug 605042 in linux-fsl-imx51 "[armel] java fails to start with eglibc-2.12-0ubuntu4" [Critical,In progress] https://launchpad.net/bugs/60504212:11
ericb2just wondering. Is it possible, with natty, to remove the ... how to say ... new panel ?12:12
ericb2I'd like to have the background alone, and make the panel appear when I want. Said differently : not always visible12:13
ogradoko, hmm, my babbage didnt boot anymore last time i tried (though i didnt inspect it and might have done something wrong)12:13
ogradoko, can that wait until after A3 ?12:13
ericb2something like a shortcut, e.g. toggle visible /hidden would be great12:13
dokoogra: yes, I'll assign it to you for testing12:14
ograk12:15
ogradoko, note that i only have a 2.5 board, thats not what is in the datacenter12:15
ogralinaro has the 3.0 ones if anyone does12:16
dokodidn't we have 2.5 in the data center?12:16
ograno12:16
ogra3.012:16
ograthere are only minro differences iirc but they are not the same12:17
janimodoko, what is the issue with openjdk-6 on ARM? Not supported well by upstream yet?13:37
janimoogra, qt 4.7.2 released :)13:38
rsalvetiGrueMaster: new kernel for you to test, for omap 4: http://people.canonical.com/~roc/kernel/natty-ti-omap4-dev/ and http://people.canonical.com/~roc/kernel/natty-ti-omap4-dev-linaro/13:39
rsalvetithe first one is based on our tree + ti, the other also includes the linaro changes for 3813:40
rsalvetiGrueMaster: use the DVI as your default output: console=ttyO2,115200,n8 vram=32M mem=456M@0x80000000 mem=256M@0xA0000000root=/dev/mmcblk0p2 rootwait omapfb.mode=dvi:1280x1024MR-32@60omapdss.def_disp=dvi13:41
ograjanimo, yes, Riddell announced it this morning in -devel13:52
ograrsalveti, any idea how we handle upgrades ? (wrt cmdline)13:52
rsalvetiogra: don't think we're handling this now, the image still uses the boot.scr available from the installer13:54
ograright13:54
ograi'm wondering how to handle it properly13:54
rsalvetibefore using this as default we should for sure make a way to change the default boot args13:54
rsalvetithe final idea is to have the hdmi driver, so it's more a temporary solution13:55
rsalvetibut it'd be good to be prepared in case we have similar changes13:55
ograi mean, we cant really handle it ... the user needs to plug the monitorf into the other socket13:55
rsalvetilike once we fix the 1gb bug13:55
rsalvetiogra: yeah, at least this13:55
ograoh, there are plans for making hdmi work before release ?13:55
rsalvetibefore final release, yes13:55
* ogra understood that wont make natty13:56
ograah, good then13:56
ograbut we will still need to mangle boot.scr13:56
rsalvetiat least this was the plan, and they are heavily working on the hdmi driver atm13:56
ograk13:56
rsalvetiyup13:57
ogrado you think we could convince them to add a sane default for the resolution so we dont need to touch the bootargs ?13:57
rsalvetiyou mean, for dvi?13:57
ograno, for hdmi13:58
ograor is hdmi supposed to do edid detection again ?13:58
rsalvetifor hdmi it'll probe the edid13:58
rsalvetiand get the best resolution available13:58
ograah, sweet13:58
rsalvetiyes, that's why I'm also waiting some more for the edid work for omap 313:58
rsalvetibecause it'll basically be the same code13:58
rsalvetisame solution13:58
ograyup, saw that13:58
rsalvetibut we still document a way to update the boot args, when needed13:59
rsalvetimaybe post inst, not sure13:59
rsalvetionce we fix the 1g bug we'd like to change that for all the boards, when the kernel is updated14:00
rsalvetifor example14:00
ograi'd like to have some automation14:01
ogramaybe through u-m14:01
fairuzthe function flush_cache_all() flushes all cache or only L1 cache? it's from asm/cacheflush.h14:02
rsalvetiogra: could be14:05
rsalvetiwas thinking more like we have for grub14:05
ograthe legacy vs grub2 update script ?14:06
rsalvetiwith default args and etc, and once a tool is called the boot args would be updated14:06
ograthats not automated14:06
rsalvetiand then write the boot.scr file14:06
ograits tricky to handle it if your monitor goes black14:06
ograso i would rather like to force it in an automated way14:07
ograto make sure the user still sees something14:07
ograi'm less concerned about the 1G bit here14:07
rsalvetiit's just an example14:08
ograbut about users using milestones that have the DVI stuff and if the HDMI patch does *not* make it also for maverick->natty upgrades14:08
rsalvetiogra: we should have at least edid detection, even if it's just dvi14:10
ograin initrd ?14:11
rsalvetiogra: so then the problem would be the natty milestones, before this feature is in14:11
rsalvetino, in kernel14:11
ograor as a kernel fix14:11
ograah14:11
ograwell, if its only milestones we can probably get away with documentation14:11
rsalvetiok, that works, then I believe we're fine14:13
davidmogra, call time?15:00
ogradavidm, sure (why do you ping here)15:00
markos_ogra, you mentioned a free/open libz160 while in Cambridge, tried to find it but I didn't really end up with something, got a url somewhere?15:01
markos_the one I have from FSL is definitely not open :)15:02
XorAarm ppas work without special flags these days?15:04
dokojanimo gone?15:22
=== brendand_ is now known as brendand
ogramarkos_, asac had it iirc15:40
asacogra: its not free/open because of a stupid license header missing15:44
asacogra: current state is in the canonical-arm ppa iirc15:44
asacogra: i followed up one more time last cycle and never got a reply from Paul whether this is now resolved15:45
asaci can check with our fsl LT15:45
markos_asac, if it's the same one I have, it's *definitely* not open15:49
ograthe one we had was definitely open, just missing the license15:49
markos_afaik, fsl built only one libz16015:50
ograand we had been promised severeal times that would be fixed15:50
markos_could you point me to a url?15:50
ograi dont know where asac put it, i have ssen it first in a BSP which i cant make public15:51
markos_fsl bsp?15:51
ograyep15:51
markos_it's probably the same, hate to let you know, this one is not open15:51
ograhmm, they told us differently15:51
markos_probably some misunderstanding15:51
ograit had only 2d accel, EXA support and XV iirc15:52
ograno other acceleration15:52
ograand back when i touched it it was very crashy15:52
markos_Neko could comment more on that15:52
markos_afaik it's closed, I'd love to have it open, but I got a very firm *not open*15:53
markos_but yes, it's probably the same one15:53
markos_in fact the one I have here has a zero-sized file "DO_NOT_DISTRIBUTE"15:54
markos_looks closed to me :)15:54
markos_the headers are bsd-like licensed though :)15:55
ograhmm15:56
ograi dont think my version had that15:56
asacogra: its in canonical-arm-dev private iirc or fsl private ... we circled that back tot hem asking to add a COPYING15:56
markos_some other files in there have a very very firm "AMD_CONFIDENTIAL, DO NOT DISTRIBUTE"15:56
markos_I'm pretty sure it's not open15:56
ograthe adm bit is definately not open15:57
ograi was talking about the other one15:57
=== zul__ is now known as zul
markos_what other one, there is only one library libz16016:00
ograiirc it was called something with 14016:00
GrueMasterdoko: What do you need testing on babbage 3?  I have one still that I keep for SRU testing.16:02
DevilCodeHey All16:04
DevilCodehttps://wiki.ubuntu.com/ARM/OMAPMaverickInstall   <-- this the best place to find out how to put ubuntu on beagle board16:05
dokoGrueMaster:  bug #60504216:06
ubot2Launchpad bug 605042 in linux-fsl-imx51 "[armel] java fails to start with eglibc-2.12-0ubuntu4" [Critical,In progress] https://launchpad.net/bugs/60504216:06
dokoinstall the kernel from the bug report, the old eglibc packages (on a maverick chroot). with the old kernel, java -version segfaults, with the new kernel it should not16:07
GrueMasterI'll have a look.  Might not get to it until after A3, but I'll add it to the priority list.16:09
=== zyga is now known as zyga-food
GrueMasterMight take a while just to read through the bug log.  :P16:09
DevilCodeany one got a little to give me some noob guidance16:13
GrueMasterDevilCode: What more info do you need?16:14
DevilCodewell on the link:https://wiki.ubuntu.com/ARM/OMAPMaverickInstall16:14
DevilCodea little stuck on booting the image16:14
GrueMasterWhat system do you have?16:15
DevilCodewith xM rev B16:15
DevilCodemy SD card is just one partition /dev/sdd116:15
GrueMasterOk, Follow the instructions for BeagleBoard XM rev A3 and B.16:15
DevilCodewhats the second partition though16:16
GrueMasterYou need to flash the image to the sd card, not the partition.16:16
DevilCodegaaw]16:16
GrueMasterThe image has two partitions.16:16
DevilCodesudo dd bs=4M if=ubuntu-netbook-10.10-preinstalled-netbook-armel+<omap image>.img of=/dev/sdd  <--- instead of sdd116:16
GrueMasteryes16:16
DevilCode*blush*16:17
DevilCodeok16:17
ogra<omap image> is a placeholder16:17
ogrause the right name ;)16:17
DevilCodelol16:18
DevilCode*dances that i got that right*16:18
DevilCodeOn OMAP4 Blaze Mount the first SD card partition at your host PC and copy the following files to it:16:38
DevilCodehow do i know16:38
ogrado you own a blaze ?16:38
DevilCodeis that a rev V16:39
DevilCoderev B16:39
ogra(did you pay $2500 for your devboard)16:39
DevilCodelol no16:39
ograright and it doesnt have a case that looks like an elephant mobile phone either i guess :)16:39
DevilCodejust a cardboard box *sniffle*16:40
ograright, so you dont have a blaze16:40
ograand you said above you have a beagle16:40
ograwhich is omap316:40
ograso ignore all omap4 stuff :)16:40
DevilCodecrap16:40
DevilCodei went for the 4 lol16:40
DevilCode*start again*16:40
XorAheh, I think Linaro have most of the worlds blazes :-)16:40
ograhow many do they have ?16:42
ograwe have two in the team16:42
XorAogra: I dont think TI made very many at all, or 4430sdp16:43
XorAogra: I know just getting omap4 hardware for my work was difficult16:43
ograyeah16:44
ograwas difficult for everyone16:44
XorAogra: yeah but I was contracted by TI to work on it :-)16:44
XorAnot even enough for their own employees let alone anyone else16:44
ograyup, same prob we had16:45
* XorA is still waiting for his panda :-)16:46
DevilCodethey are slow creatures :P16:49
DevilCodeand slow cars16:50
DevilCodegenerally slow16:50
GrueMasterYea, but the boards are fast.16:50
DevilCodeand the animal is tasty16:50
DevilCode:P16:50
DevilCodehttp://www.tshirthell.com/funny-shirts/this-shirt-is-100-organic-65-baby-seal-25-panda-10-manatee/16:54
GrueMasterrsalveti: I am setting up to test the .38 kernels.  Will I be able to output via DVI?  My panda didn't have the DVI rework from the TI Rally.17:13
GrueMasterNevermind.  It's coming up17:14
rsalvetiGrueMaster: nice17:21
GrueMasterNo X, though.17:22
GrueMasterApparently I am not authorized to start it.17:22
GrueMasterMight have something to do with the TI drivers being installed.17:22
rsalveticould be17:24
rsalvetifor the drivers to work you also need a kernel module17:25
rsalvetithat for sure you don't have atm17:25
rsalvetiso just remove them17:25
GrueMasterright.17:26
ericb2janimo: pong ?17:30
GrueMasterrsalveti: I'm seeing major problems, but they are likely due to my hacked up image.  I'm going to wait until we have a new image to test these kernels with.18:08
rsalvetiGrueMaster: ok18:08
rsalvetiwe should have new images today18:08
GrueMasterI really don't want to start fresh with a two week old kernel and try to update it to current.18:08
GrueMasterYea, sure.  Believe it when I see it.  :P18:09
rsalvetihehe18:09
GrueMasterNew image build fail reports.  Wheee.18:10
DevilCodeanyone get BeagleBoard running with DirectFB18:14
rsalvetiDevilCode: try asking #beagle18:15
DevilCodekk18:15
ograGrueMaster, i guess that was pitti s manually killed build18:17
GrueMasterAh.18:17
* GrueMaster just notes yet another failed build attempt18:17
=== jhobbs_ is now known as jhobbs
=== zyga is now known as zyga-afk
=== zyga-afk is now known as zyga
=== zul_ is now known as zul
=== npitre_ is now known as npitre
GrueMasterHmm.  Looks like sources.list may not be populated.22:33
ogranot at all ?!?22:33
GrueMasterWell, that was according to syslog.  checking filesystem.22:36
* ogra checks jasper ...22:36
GrueMasterNo, it's ok.  No universe or multiverse though.22:37
ograyeah22:37
ogralongstabnding bug i have on my plate22:38
ograbug 65070323:32
ubot2Launchpad bug 650703 in ubiquity "oem-config-prepare works, but oem-config fails to start after reboot" [High,Confirmed] https://launchpad.net/bugs/65070323:32
prpplagueogra: hey, got a minute for me to pester you about an ubuntu desktop oddity ?23:45
ograprpplague, what is it, i'm trying to debug image breakage atm23:46
prpplagueogra: not important atm, go back to debugging23:46
ogra(for images we will have to release on thu, so i'm a bit under pressure)23:46
ograwell, you can still just throw in your question, probably its an easy task23:46
prpplagueogra: i've got 10.10 desktop on both a desktop pc and a netbook, on the netbook the task switch allows me to drag a window from one workspace to another, but on my desktop pc install it doesn't, just curious if there was something quirky about that23:48
prpplagueogra: not really important, just saw you online and thought i'd ask23:48
ograwell, is the netbook using the netbook UI ?23:49
prpplagueogra: negative, same for both23:50
prpplagueogra: used the same thumbdrive in fact23:50
ograthats indeed weird ... possibly worth a bug report23:50
rsalvetiprpplague: what are your video card at your desktop?23:52
rsalveticould be that on your desktop you're actually running with 2 x servers23:52
prpplaguersalveti: hmm that an install option?23:54
rsalvetinot install, but a runtime option23:55
rsalvetifor example, with nvidia driver you can select how your x11 will behave23:55
prpplaguersalveti: should i be able to grep the running processes for multiple x?23:55
rsalvetiprpplague: probably, also check your xorg logs23:57
prpplaguersalveti: intel chipset, but i only see one x running23:57
rsalvetiI remember I used to run the nvidia driver this way, having two different displays over each monitor23:57
rsalvetiand I couldn't move one window to the other23:57
rsalvetihm, so could be a bug23:57
prpplaguersalveti: i'll dig around some with that23:58
rsalvetiprpplague: compare both xorg logs23:58
rsalvetisee if anything pops up23:58
prpplaguersalveti: just an annoyance at this stage, since it _works_ with the latest fedora :P23:58
rsalvetiyou should try our alpha 3 image so ;-)23:59
rsalvetionce it's available hehe23:59

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