/srv/irclogs.ubuntu.com/2011/11/21/#ubuntu-arm.txt

Netham46I'm typing up a tutorial for installing ubuntu on a tablet, can anyone recall what the file name format is for a rootstock image?01:17
twbDefine "rootstock"01:21
Netham46twb, the program on ubuntu that creates an image for ARM01:34
twbOh, right, that's the arm-specific equivalent of debootstrap; that rings a bell01:35
twbDunno the answer to your question, though.01:35
Netham46Yup.01:35
Netham46and kk01:35
infinityNetham46: rootstock is deprecated, we don't generally encourage its use.02:08
twbinfinity: what's the recommended solution?02:09
infinityNetham46: We tend to prefer people start with ubuntu-core and tailor it to their needs, or if you really need to build images from the ground up, use live-build.02:09
twbCool02:09
twbinfinity: if you have write access to help.ubuntu.com or wherever the rootstock page was, please add a deprecation notice to it02:09
twbhttps://wiki.ubuntu.com/ARM/RootStock02:10
infinitytwb: Yeah, we have a lot of docs to fix, afaik.02:10
infinityI might add a notice to that page now, though.02:10
twbI'd do it myself but ICBF going through the account creation bollocks02:11
infinitytwb: Does that look reasonable for now (until we have time to tidy more)?02:13
twbinfinity: yeah fine02:13
Netham46infinity, hmm, okay02:36
Netham46I didn't realize that it was antiquated02:36
twbNot your fault02:39
An-iSociaLoh hello again04:51
Hanmac1is ubuntu-arm ready for Nvidia "Tegra 3"? or should i wait?08:15
lilstevieHanmac1: it will be08:31
lilstevie"ready08:31
lilstevie"08:31
lilstevieyou will need to work the kernel side08:31
lilsteviebut the userland part will work just fine, well providing they don't have any nasty bugs like the Tegra2 and the cp15 bug08:32
Hanmac1i mean exist an image that i can use to install it or should i build one on my own?08:43
infinityHanmac1: No images, since we have no kernels (and no specific devices we're targetting)08:54
Hanmac1hm what is needed to build the kernels? i wanted to use "Asus Eee Pad Transformer Prime" with ubuntu09:02
ogra_infinity, well, he could start from ubuntu-core, make a kernel and hack the bootloader and be done09:06
ogra_Hanmac1, find the kernel source for your device ... take the linux-ac100 kernel config as a base and adjust your config for the tegra3 to match everything but the HW bits needed for tegra3 ... build it ...09:08
ogra_then make sure you have full access to the bootloader, flash your kernel and boot into a modified ubuntu-core rootfs on MMC or SD09:09
ogra_if you dont understand any of the above i said above. you should better not do it though09:09
lilstevieHanmac1: not enough is known about the prime yet09:10
lilstevieHanmac1: biggest issue I forsee is the SBK09:10
=== popey_ is now known as popey
Hanmac1hm i does not know if i will have full access to the bootloader (i have not buy the pad jet), i have played enough with bootloaders (i know how to install it to the MBR), but what did you mean with SBK?09:13
lilstevieogra_: well given my exp with the tf101 I would say, if you have nvflash access, not bootloader, cause asus bootloaders are non interactive09:13
lilstevieand also, installing a bootsector to MBR is not the same thing09:14
* ogra_ would see nvflash and full bootloader access as the same thing ...09:14
ogra_if you can replace it you have full access ;)09:15
lilstevieheh I wouldn't09:15
lilstevieyou could argue razorclaw gives you "full access" to the bootloader09:15
lilsteviebecause you can write to the staging partition09:15
lilstevieand that will flash pretty much everything09:16
lilstevieexcept BCT09:16
lilstevieand it really is full access, just risk of brickage increases exponantially as you flash more09:16
=== Quintasan_ is now known as Quintasan
S0NiChi14:40
S0NiCi try to crosscompile a litte c++-programm for my pandaboard which is running ubuntu 10.10. my hostsystem runs also ubuntu 10.10 and i tried "arm -linux-gnueabi-cpp main.cpp" and the error was "fatal error: QTGui/QWidget: no such file or directory"... has anyone an idea?14:44
S0NiCthanks in advance14:44
tgall_foodoko, ping14:44
tgall_fooS0NiC, sounds like you're missing either a path to the qt includes or they aren't installed15:34
dokotgall_foo, pong15:46
tgall_foohi doko, just wanted to catch up with you on the libjpeg8 / libjpeg-turbo project we had talked about at the last UDS15:56
tgall_foodoko, as of this morning I've got the majority of pkgs with deps on libjpeg-dev built in my PPAs against the libjpeg-turbo's version of the libjpeg-dev for precise15:57
tgall_fooand by majority I mean majority in the main archive,  universe is another matter15:58
tgall_fooif I hadn't run out of space in my PPA main would be complete ?  sigh15:58
tgall_foodoko, for the sake of trivia https://wiki.linaro.org/TomGall/LibjpegUse  lists the debs in universe and main for libjpeg62, libjpeg8, libjpeg-dev etc15:59
tgall_foodoko, make sense ?16:00
dokotgall_foo, can we move this discussion to #ubuntu-devel?16:10
tgall_foodoko, sure16:11
=== ericm|ubuntu is now known as ericm-Zzz
An-iSociaLstill working on a kernel that will boot on a Stingray board16:56
ogra_good luck :)16:56
An-iSociaLty16:56
ppisatibug 88856916:56
ubot2Launchpad bug 888569 in kernel-sru-workflow/verification-testing "linux-ti-omap4: 2.6.35-903.27 -proposed tracker" [Undecided,Fix released] https://launchpad.net/bugs/88856916:56
ppisatiGrueMaster: ^^16:57
An-iSociaLi think i can utilize the ac100 kernel config as a base then add the stingray stuff16:57
ppisatiGrueMaster: wrt the failed tests, is there a way i can rerun these tests?16:57
ppisatiGrueMaster: actually it should just a matter of flipping some configs, but you never know...16:57
ogra_An-iSociaL, note that the ac100 kernel is based on chromeos16:57
An-iSociaLnoted, i suspect that just compiling the stingray kernel with the right things enabled will work, using the ac100 as a template for what needs to be enabled could do the trick16:59
ogra_sure, but there are ac100 specific bits you will likely need to clean up17:00
ogra_like the nvec controller17:00
An-iSociaLill need some time to do the research but tell me if im wrong, having things enabled which do not actually exist wont hinder anything will it?17:02
An-iSociaLas long as the things which do are enabled17:02
ogra_no idea, might make your buold fall over17:02
ogra_*build17:03
An-iSociaLlol quite likely, ive tried so many things already it couldnt be any worse17:03
An-iSociaLwhat kernel version is this for anyway?17:03
An-iSociaLi havent seen that part yet17:03
ogra_ac100 ?17:03
An-iSociaLoh nvm17:03
ogra_thats 2.6.3817:03
An-iSociaLlol duh, config-2.6.3817:03
An-iSociaLsorry17:04
An-iSociaLspeedway in my head had a pile up, traffic's been diverted and the wreck has been cleared17:04
GrueMasterppisati: Sorry, didn't see your ping earlier.  Yes, you can run the failing tests.  Either pull the entire bzr tree for qrt (bzr branch lp:qa-regression-testing), or I can post tarballs of just the failing tests somewhere.18:49
GrueMasterI'm remote this week, but it should still be possible for me to post these.18:49
An-iSociaLok comparisons are done19:15
An-iSociaLnot nearly as bad as i thought19:16
An-iSociaL10-15 unused which includes aroud 10 which are android19:16
An-iSociaL13 which are unset in ac100 but tagged as yes in stingray19:17
An-iSociaL7 of which are debug flags19:17
An-iSociaLill write up a report when i get a moment19:17
An-iSociaLsee how this goes first19:17
An-iSociaLseveral marked for modules in ac100 were marked as yes in stingray so i just marked them yes in ac10019:19
An-iSociaLlikely could go back to modules but i wanted to make sure they were available in case something went wrong with accessing the modules19:19
ppisatiGrueMaster: i'll pull from bzr, thanks19:45
GrueMasterppisati: Ok.  On the test-kernel-security.py, the only test that "should" fail is /dev/mem.  Still looking at it.19:46
GrueMasterAlso, test-kernel-[aslr-collisions|root-ops] both fail and are being worked on. I don't think they are kernel issues, rather kernel differences (vs x86)19:48
ppisatiGrueMaster: ok, and in case one of the option there is not implemented, to whom shall i report it?19:49
ppisatiGrueMaster: i.e. SECCOMP was not supported on arm back then19:49
GrueMasterYes, it was.  The test passes on earlier arm kernels.19:50
GrueMasteriMX5/Dove.19:50
GrueMasterAlso, I think the config just has it disabled in the omap4 kernel.19:50
ppisatiGrueMaster: not for omap4, perhaps had some stuff backported as part of the marvell BSP19:51
ppisati*perhaps mvl-dove19:51
GrueMasterIt should be in the 2.6.35 kernels.  There was a CVE filed for the x86 kernel that added it.19:53
GrueMasterI think.  Will have to check my remote system (if possible - I'm away from home this week).19:53
GrueMasterppisati: I have remote access to my panda with Maverick installed.  Will look it over again for SECCOMP.20:13
=== gary_ is now known as Guest71968
An-iSociaLhm20:31
An-iSociaLno go on compiling20:31
An-iSociaLthe sources is all kinds of outdated20:31
An-iSociaLpatches were released for most of my problems20:31
An-iSociaLwould it be possible to import the mach-tegra to another kernel?20:34
An-iSociaLwith the stingray sources*20:34
steev_An-iSociaL: possible, yes, but it would be a lot of work21:04
An-iSociaLhm21:04
steev_clear your calendar if you plan to :)21:05
An-iSociaLyea thats not going to happen then21:05
An-iSociaLperhaps i can get the newer kernel from android 4.021:05
An-iSociaLfor the moment though, i do have it compiled, about to test21:05
An-iSociaLi think i was able to keep support for ac10021:05
An-iSociaLimpressive it actually booted21:15
An-iSociaLnow to actually put this on an sdcard21:16
An-iSociaLlet me revice, the kernel booted, as in didnt freak out or freeze up21:16
An-iSociaLrevise*21:16
GrueMasterWell, that is always a good sign.21:40
An-iSociaLwell i setup the sdcard... now all im getting is checksum file missing21:41
An-iSociaLnever seen that one before21:41
An-iSociaLoh great and google doesnt have anything on it :|21:42
An-iSociaLuhm need a little help with this one21:43
An-iSociaL"Error: Checksum file missing"21:44
An-iSociaL"Spawning a shell within the initramfs"21:44
An-iSociaLNo results found for "Error: Checksum file missing".21:45
GrueMasterIs this on our image?21:46
An-iSociaLfrom the kernel i compiled using the ac100 kernel config and using the ac100 prebuilt filesystem21:46
GrueMasterinfinity: Do you know how to create this checksum?21:47
GrueMasterI think it is done at image creation time and added to the initrd in some weird fashion21:47
An-iSociaLhm21:47
An-iSociaLi created the initrd from chroot inside the prebuilt rootfs21:48
GrueMasterI think the md5sum is added after that step.  Not sure how.21:48
An-iSociaLwell im right there with you21:48
GrueMasterBut it shouldbe an md5sum of the rootfs tarball (iirc)21:49
An-iSociaLwell thats gonna be a problem as i extracted it onto an sdcard to boot it21:49
GrueMasterSo you are trying to dual boot in essence?21:50
An-iSociaLkind of, im working with a motorola xoom, the kernel+initrd were put into a img file then flashed to the recovery partition from fastboot21:51
An-iSociaLnever had a problem with that before21:51
GrueMasterthere is probably a script in the initrd that can either be deleted or modified.  But  I'm remote from my test bench this week, otherwise I could look.21:51
An-iSociaLok i found it21:52
An-iSociaL-> /usr/share/initramfs-tools/scripts/installer21:52
An-iSociaLregenerating21:52
An-iSociaLer21:55
An-iSociaLi guess i didnt read something i should have21:55
An-iSociaLi did the install myself21:55
An-iSociaLthis could get interesting21:57
An-iSociaLok proper way to remove: dpkg --purge ac100-tarball-installer21:58
* GrueMaster crosses fingers/toes in anticipation.21:59
* An-iSociaL does too while he fights with his microsd reader22:00
An-iSociaLand here we go22:06
An-iSociaLstarting services22:07
An-iSociaLwaiting for network configuration22:07
An-iSociaLerr22:07
An-iSociaLno desktop22:07
An-iSociaLlocalhost.localdomain login:22:08
An-iSociaLwhich i cant do22:08
An-iSociaLno keyboard22:08
An-iSociaLok let me get adbd onto it22:08
An-iSociaLuhm22:15
An-iSociaLwhats going on in this filesystem?22:15
An-iSociaLno desktop started up22:15
An-iSociaLbut i dont even see one to be started22:15
An-iSociaLoh22:16
An-iSociaLlightdm22:16
GrueMasterProbably more magic that the install script handles.22:17
An-iSociaLyea22:18
An-iSociaLkinda hard to install from my sdcard to my sdcard though22:18
An-iSociaLi guess if its what i have to do i can work out how to put it on internal storage22:18
An-iSociaLjust partitions after all22:18
GrueMasterI think the image is supposed to launch oem-config on startup.  I think you can have it autostart with "sudo touch /var/lib/oem-config/run"22:19
An-iSociaLok lightdm enabled and booting again22:19
An-iSociaLstill no22:21
An-iSociaLwhoa22:26
An-iSociaLits freaking out22:26
An-iSociaLkeeps switching ttys, guessing trying to start lightdm and failing22:26
An-iSociaLtrying to get adbd up and running so i can actually do something22:27
An-iSociaLwelp lets check the logs22:27
An-iSociaL[    89.326] (EE) open /dev/fb0: Device or resource busy22:28
An-iSociaLnegative :(22:28
An-iSociaLok time to try the nvidia drivers22:30
An-iSociaLits a bust22:39
=== Gary_ is now known as Guest96366
An-iSociaLwelp ill try some more stuff later22:43
An-iSociaLpretty much concrete the nvidia drivers do not work22:43

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