=== Jack87 is now known as Jack87|Away | ||
reisei | hi, all! Please, advice me, how can I build kernel on the x86 and then replace it on the arm device? | 07:08 |
---|---|---|
reisei | I had try to follow https://wiki.ubuntu.com/Kernel/BuildYourOwnKernel But there is no such directory: debian in the the ti-omap4 branch ... | 07:09 |
twb | reisei: you're cross-compiling, that adds a layer of complexity | 07:10 |
twb | The debian/* files will only be present in git branches that are from debian/ubuntu, and even then maybe not. | 07:10 |
twb | When I'm *not* cross-compiling, I use the upstream-provided mechanism of "make deb-pkg" -- I don't know if this is recommended or if it works with cross-compilation. | 07:11 |
reisei | twb: thank you! I'll try this | 07:13 |
twb | If you don't have a cross-compile environment yet, you need to solve that problem first | 07:14 |
reisei | twb: btw, in the master branch there are the debian directory | 07:14 |
twb | master branch from where? | 07:15 |
reisei | twb: from kernel.ubuntu.com, for example ubuntu/ubuntu-maverick | 07:15 |
twb | Well I guess that is a "git branch from debian/ubuntu" | 07:16 |
twb | As opposed to linus-2.6 or some other upstream branch | 07:16 |
twb | I'm not a dev so my understanding of current best practice is pretty haze -- don't take anything I say as gospe | 07:17 |
twb | *gospel | 07:17 |
reisei | twb: yes, it is a "git branch from debian/ubuntu" :) I understand, that directory debian can be only in such gits, but I'am just interesting why in the ti-omap4 branche of this git there is no such directory :) | 07:20 |
twb | Probably the ti-omap4 branch comes straight from upstream or something | 07:21 |
reisei | got an error while make deb-pkg: dpkg-gencontrol: error: current host architecture 'armel' does not appear in package's architecture list (i386) | 07:39 |
twb | You probably need to do some magic dance | 07:41 |
twb | The other way you can do it is to just run arm code in a chroot on your x86 box, using qemu user-space emulation | 07:42 |
twb | What that does is basically leave everything as-is except the CPU instructions, so it's lightweight compared to a normal VM | 07:42 |
reisei | okay :) will go to learn some magic things | 07:43 |
twb | http://wiki.debian.org/QemuUserEmulation | 07:44 |
twb | Note you will probably need qemu 1.x due to http://bugs.debian.org/641270 | 07:47 |
reisei | twb: thanks a lot! | 07:48 |
twb | Eh, cross compilation is probably less sucky but I don't like grabbing unsigned tarballs from linaro or whatever | 07:49 |
reisei | twb: I just need recompile the kernel... | 07:53 |
twb | reisei: I mean: unsigned tarballs of the cross-compilation environment | 07:53 |
reisei | oh, okay | 07:53 |
twb | or maybe they're signed but not by ubuntu, etc. | 07:53 |
alum1 | Hi may I have a question about i.MX53 based board and Ubuntu? I have trouble to boot Oneiric | 08:37 |
alum1 | QuickStart image. | 08:37 |
alum1 | I have Nitrogen53 board with i.MX53. Originally board comes with Android on SD card, I re-write it with QuickStart Oneiric image | 08:39 |
alum1 | When I am trying to boot kernel and initrd, board always froze or (at best case says that is loading linux kernel and than froze) | 08:40 |
alum1 | What I am doing after board starts : | 08:41 |
alum1 | fatload mmc 0:2 0x70000000 uImage | 08:41 |
alum1 | fatload mmc 0:2 0x72000000 uInitrd | 08:41 |
alum1 | setenv bootargs console=tty0 console=ttymxc0,115200n8 root=/dev/mmcblk0p3 | 08:41 |
alum1 | bootm 0x70000000 0x72000000 | 08:41 |
ogra_ | we only support the original quickstart from freescale .... | 08:42 |
ogra_ | but if you can make your board work, we would love to get patches ;) | 08:42 |
alum1 | and is there any (technical) reason, why this board will not work? | 08:43 |
alum1 | I suppose, that problem is more in u-boot loader than in image itself | 08:43 |
ogra_ | either u-boot or a missing/superfluous kernel driver | 08:44 |
ogra_ | could be both, hard to guess without having the board | 08:44 |
alum1 | I'am not terrible experience in u-boot and arm architecture, so I am trying what's working and what's not | 08:44 |
alum1 | And how can I debug it and find out what it can be? | 08:45 |
ogra_ | well, the nitrogen is clearly different from the quickstart | 08:45 |
ogra_ | i would start with replacing the u-boot binary on the image with the one you got with the board (if you got one) and see if it works then ... (i.e. try to identify if its rather bootloader or kernel related) | 08:46 |
ogra_ | not sure if your android image actually has u-boot or something else | 08:46 |
alum1 | ok | 08:47 |
alum1 | Please correct me if I am wrong. I have u-boot on board flash and is is also on sd card at first partiton (made by zcat of ubuntu image}. So I need to take u0boot from board flash and place it on cd card, right. That's what you meant? | 08:49 |
GrueMaster | alum1: YYou can add earlyprintk=ttymx0,115200n8 to your boot cmdline to get early kernel output. | 08:54 |
GrueMaster | If your uboot is loading the kernel, thiswould be the next step to see where the kernel s breaking. | 08:54 |
ppisati | guys, why don't we ship images with "support" for serial console out of the box? e.g. console=ttyO2,foobar and /etc/init/ttyO2.conf? | 09:24 |
ogra_ | we do ... the server image ;) | 09:25 |
infinity | I think the argument is that "desktop" users will have a display and keyboard, an they might want to use the serial port for other things. | 09:25 |
infinity | Whereas "server" users are more likely to want a headless serial setup. | 09:25 |
infinity | But I dunno. It's fairly arbitrary. | 09:26 |
ogra_ | beyond that there is a longstanding upstart bug that it should automatically fire up a getty if a serial console is defined | 09:26 |
ogra_ | (which is simply not fixed because nobody uploaded the fix from the bug yet ) | 09:26 |
ogra_ | ppisati, infinity, bug 702574 .... someone needs to upload this :) | 09:31 |
ubot2` | Launchpad bug 702574 in upstart "getty should be started automatically on serial port when serial console is set on kernel command line" [Wishlist,New] https://launchpad.net/bugs/702574 | 09:31 |
ppisati | btw, did you see this one? http://www.alwaysinnovating.com/products/hdmidongle.htm | 09:42 |
infinity | ppisati: That's pretty slick looking. | 09:48 |
infinity | ppisati: Can our omap4 kernel boot it? :) | 09:48 |
ndec | ubuntu works on this dongle, yes. | 09:50 |
ndec | see http://groups.google.com/group/pandaboard/browse_thread/thread/dd881aa34e1a916e | 09:51 |
ndec | we work regularly with gregoire who did this, and he is using our Ubuntu releases a lot! | 09:51 |
infinity | ndec: Very cool. | 09:52 |
ndec | so should ubuntu TV ;-) | 09:52 |
infinity | ndec: So, slapping an Ubuntu preinstalled omap4 image on an SD for that dongle should "Just Work"? | 09:52 |
ndec | i asked gregoire to send me one so that I could try ... | 09:54 |
ndec | i don't know the details of the h/w and if that requires updates in bootloader/kernel. but 11.10 should be pretty close to work on the device | 09:54 |
infinity | Would definitely like to see 12.04 support it out of the box if 11.10 doesn't. | 09:55 |
infinity | I'd happily take a sample to play with. ;) | 09:55 |
ndec | no doubt about that ;-) | 09:57 |
ppisati | Ubuntu TV on a stick! :) | 10:09 |
ndec | no need to find TV makers... works on all TVs ! | 10:12 |
ppisati | ndec: are they using a 4430 or 4460? | 10:16 |
alum1 | GrueMaster: earlyprint didn't show anything. I can see just "Starting kernel ..." | 10:24 |
ndec | earlyprintk might help ;-) | 10:25 |
GrueMaster | So that means the kernel is not initializing. The output you see is u-boot forking into the kernel, but the kernel is not executing. | 10:25 |
ndec | GrueMaster: alum1: see ^^ you need a 'k' at the end | 10:25 |
GrueMaster | I assumed he copied what I had above. | 10:26 |
GrueMaster | [09:54:19] <GrueMaster> alum1: YYou can add earlyprintk=ttymx0,115200n8 to your boot cmdline to get early kernel output. | 10:26 |
suihkulokki | even without earlyprintk you should see "Uncompressing Linux..." | 10:29 |
alum1 | I see just "Starting kernel…", nothing more | 10:33 |
suihkulokki | alum1: that suggests that the location where uboot is jumping to does not have a valid kernel | 10:44 |
alum1 | suihkulokki: how can I check that? | 10:48 |
ogra_ | infinity, does that look sane to you ? http://paste.ubuntu.com/799233/ | 11:01 |
infinity | ogra_: Looks shockingly similar to what I uploaded 10 minutes ago. ;) | 11:04 |
ogra_ | :P | 11:04 |
infinity | Who needs nicotine? | 11:04 |
reisei | infinity: what about vodka? | 11:07 |
ppisati | has anyone tried to kexec into u-boot? | 11:12 |
NCommander | ppisati: I don't think anyone tried it as kexec sets up ATAGS and other voodoo to boot a kernel directly so you'd have to smack uboot to ignore that info and other fun stuff | 11:16 |
ppisati | NCommander: uhm, ok | 11:17 |
marvin24 | janimo: thanks for taking care of the new kernel so quickly | 11:32 |
janimo | marvin24, np :) | 11:32 |
marvin24 | looks like google found the bug in 3.0.13 | 11:32 |
janimo | it made evertyone's ac100 around here lock up every our or so otherwise | 11:32 |
janimo | oh great | 11:32 |
marvin24 | but I'll test it first | 11:32 |
janimo | any work going on on 3.2/3.3 on google's behalf? | 11:33 |
marvin24 | janimo: yeah, I guess that was a bad experience | 11:33 |
marvin24 | janimo: there are hacks for upstream | 11:33 |
janimo | well, development version of Ubuntu, anm assumed risk | 11:33 |
marvin24 | but we shouldn't use them | 11:33 |
marvin24 | janimo: last I heard was that someone at NV is working on a KMS driver | 11:34 |
marvin24 | but maybe at a low priority | 11:34 |
marvin24 | but that is the way to go for us | 11:34 |
marvin24 | just wait ~half a year | 11:34 |
janimo | as opposed to te binary l4t drivers? | 11:35 |
marvin24 | well, kms mean framebuffer only | 11:35 |
marvin24 | but maybe they will also do a 3d driver for it | 11:35 |
marvin24 | the current kernel implementation is definitely not for upstream | 11:35 |
marvin24 | in that case NV is way behing TI, Marvel or Samsung | 11:36 |
ogra_ | janimo, http://patchwork.ozlabs.org/patch/35501/ for quietening the ac100 boot | 11:59 |
pakattack | On which arm board, the beagleboardxm or the pandaboard es, will a ubuntu server be best for? | 13:26 |
k1l_ | Anyone tried smth with the HP Touchpad or a similar Tablet? | 13:28 |
pakattack | I am trying to build a microserver and want to know on which board ubuntu server runs better on, the pandaboard es or the beagleboard xm? | 13:29 |
pakattack | Anyone? | 13:40 |
=== Quintasan_ is now known as Quintasan | ||
OlivierN | yet another random guy who ask, nag and quit. | 14:41 |
k1l | welcome to the irc :) | 14:41 |
OlivierN | at least now I am checking that the guy is still there before preparing an answer | 14:42 |
pakattack | Which runs ubuntu server better, pandaboard es or beagleboardxm? | 16:12 |
infinity | pakattack: Panda. | 16:15 |
k1l | pakattack: i dont know the answer, but this channel is very quiet. so you have to bring some patience | 16:15 |
k1l | :( | 16:15 |
infinity | pakattack: If you missed my response, I said that the Panda will certainly run just about everything better than the Beagle. | 16:17 |
pakattack | Ok thanks | 16:19 |
Arcademan | May I ask where I might find what ISO I would need for my Archos device? | 17:55 |
sauerbraten | you have to find out which ARM platform it is based on I think. | 18:07 |
sauerbraten | Arcademan: ^ | 18:08 |
Arcademan | Ok, so like ARM Cortex A8 | 18:09 |
sauerbraten | what tablet is it? maybe I can find something for you then | 18:10 |
Arcademan | Please see https://wiki.kubuntu.org/ARM/Archos101it | 18:11 |
sauerbraten | OMAP3 then | 18:15 |
sauerbraten | Arcademan: choose your edition and versin here https://wiki.ubuntu.com/ARM/OMAP , then go for the omap3 build | 18:17 |
sauerbraten | also just "omap" in the file names | 18:17 |
sauerbraten | that should work I guess | 18:18 |
Arcademan | Ok, so I take it I must use the omap then I take it the devices does not support the desktop edition :) | 18:19 |
sauerbraten | wait, what? according to wikipedia, the archos 101 internet tablet fits the OMAP3 category, so I think any omap3/omap build should run | 18:21 |
sauerbraten | not omap4 of course | 18:21 |
sauerbraten | archos 101 g9 is omap4 thoug | 18:21 |
Arcademan | Ok may I ask one last question whats the real diff between desktop and server is the server one ubuntu server | 18:23 |
Arcademan | I see that on Wiki the Omap thing :) | 18:24 |
sauerbraten | I think the server one comes without X and has special packages preinstalled and repos configured | 18:36 |
sauerbraten | you can add X to server, though | 18:36 |
sauerbraten | you can also install all server packages on desktop of course | 18:36 |
sauerbraten | just a matter of what you think fits your needs better I guess | 18:37 |
Arcademan | Ok, sauerbraten I try it out this next week or so thanks :) | 18:41 |
sauerbraten | np, hope it goes well for you, since I got few problems with earlier versions on omap4 :D | 18:42 |
Arcademan | I will be playing around with it trying to figure out what desktop package works the best ex kde ect :) | 18:43 |
Arcademan | They have a debian release I may try and fork over some of them :P | 18:43 |
sauerbraten | eww, KDE :P I prefer XFCE or LXDE, some lighter stuff :) but well, matter of taste | 18:44 |
Arcademan | me to sauerbraten :P | 18:44 |
Arcademan | I just did not want to use the dang Android :P | 18:45 |
=== alexande1 is now known as Alex[sp3dev] | ||
kerute | hello | 22:28 |
kerute | any idea on how i can flash a new kernel on my eeepad transformer ? | 22:29 |
kerute | flashed ubuntu with OLife | 22:29 |
lilstevie | kerute: there are 2 ways | 23:10 |
lilstevie | kerute: one way is to replace ./kernels/2636zImage and ./kernels/initrd.img-2.6.36 in the OLiFE directories, then in the update device menu select update kernel. | 23:11 |
lilstevie | kerute: the other is to make a boot.img from your kernel and ramdisk, pack it in a nvblob and dd it to /dev/mmcblk0p4 | 23:13 |
kerute | lilstevie_: i think about nvflash ? | 23:13 |
kerute | nvblob ? | 23:14 |
kerute | first one seems easier :) | 23:14 |
kerute | lilstevie_: what about the 2.6.38 in OLiFE ? | 23:15 |
kerute | is that the CrOs kernel ? the script says the option is unavailable | 23:16 |
kerute | thanks for the advice ill try that | 23:26 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!