ncharles | anybody know the default username/password on the 11.10 omap install? | 02:21 |
---|---|---|
infinity | ncharles: There isn't one until you install it. | 02:22 |
infinity | ncharles: At which point, it's the one you set up. | 02:22 |
ncharles | this is with the armel minimal image | 02:25 |
ncharles | I found it, it was ubuntu/temppwd | 02:25 |
twb | You're probably still in the oem phase then | 02:25 |
infinity | Err. | 02:27 |
infinity | Or not an actual ubuntu image at all. | 02:27 |
ncharles | it was the r0 image I had floating around | 02:27 |
infinity | "r0"? | 02:27 |
ncharles | ubuntu-11.10-r0-minimal-armel | 02:28 |
infinity | Yeah, that's not an ubuntu image. | 02:28 |
infinity | Not an official one. | 02:28 |
infinity | As in, not built by us. | 02:28 |
ncharles | oh ok, it was the one i could get the ethernet working with | 02:29 |
* infinity isn't sure how much he trusts rcn-ee.net | 02:29 | |
infinity | But okay. | 02:29 |
ncharles | what is the prefered image for a beagleboard-xm? | 02:31 |
infinity | desktop or server? | 02:31 |
ncharles | server | 02:31 |
infinity | http://cdimage.ubuntu.com/ubuntu-server/daily-preinstalled/current/ | 02:32 |
ncharles | ok well maybe I'll try that image then | 02:36 |
ncharles | does that work well with the beaglebone? | 02:36 |
infinity | Not sure, don't have one. | 02:37 |
infinity | I haven't heard excessive complaints, though. | 02:37 |
ncharles | i have one coming tomorrow so i'll have to give it a shot | 02:38 |
=== kenws_ is now known as kenws | ||
=== mckoan|away is now known as mckoan | ||
ndec | ogra_: hi... need an advice on ubuntu/debian pkg ;-) | 09:05 |
ndec | we have several flavor of our MM firmware, and I thought we should use update-alternatives to manage them all. | 09:05 |
ndec | is that a good idea? | 09:05 |
ndec | in the first place? | 09:05 |
infinity | ndec: Oh, speaking of packaging. | 09:06 |
ndec | infinity: hi! | 09:06 |
infinity | ndec: You guys are really going to want to rebuild your binary drivers with our latest toolchain. | 09:06 |
ndec | why that? | 09:06 |
infinity | ndec: The mali folks just discovered yesterday that not rebuilding caused them pain when they tried to package. | 09:06 |
ndec | i don't get it. | 09:07 |
infinity | ndec: dpkg-shlibdeps can't handle binaries with the old linker path. And, frankly, I don't care enough to mangle it, because in every non-binary-blob case, dpkg-shlibdeps (obviously) is operating on freshly-built binaries. | 09:07 |
ndec | what kind of problems? you uploaded new gcc and it breaks something? | 09:07 |
ndec | argh... | 09:07 |
infinity | ndec: Oh, you missed the drama, I guess. | 09:07 |
infinity | ndec: Linker path for armhf got bikeshedded to death, I was awake all weekend re-tooling. | 09:08 |
ndec | actually, no. i enjoyed reading the emails every day ;-) | 09:08 |
infinity | ndec: Your old packages still work fine. | 09:08 |
infinity | ndec: But rebuilding them will break. | 09:08 |
infinity | ndec: (As in, rebuilding the Debian packages will break, because of the blobs) | 09:08 |
infinity | ndec: So, just a quick rebuild of the blobs with the current toolchain, and life's good. | 09:08 |
ndec | when you say the old package works, did you try any application or just libs? | 09:09 |
infinity | Oh, it all works. | 09:09 |
infinity | All old binaries work fine. | 09:09 |
ndec | ok. | 09:09 |
infinity | This problem is isolated specifically to dpkg-shlibdeps, and my unwillingness to "fix" it for a use-case that I don't think matters. ;) | 09:09 |
infinity | So, you've been warned. | 09:09 |
ndec | is the new toolchain in the archive now? | 09:09 |
infinity | If you need to rev the Debian package, you'll need to rev your binaries first, that's all. | 09:09 |
infinity | ndec: Yeah, the archive is all good to go. | 09:10 |
ndec | ok. thx. | 09:10 |
infinity | ndec: Same for Debian, if you build for them too, I dunno. I got Debian all fixed up yesterday/today. | 09:10 |
infinity | (Though, I'm guessing you build your binaries in such a way that they'd work on both distros, but I've never taken a close look) | 09:11 |
ndec | infinity: we never really tried them on debian... | 09:16 |
ndec | and to be honest, we never explicitely tried to make sure they would work on debian... so if they work, it's pure coincidence... | 09:17 |
infinity | Fair enough. | 09:17 |
infinity | Well, if they link dynamically with glibc, they definitely won't work on Debian right now. | 09:17 |
lilstevie | glibc abi break? | 09:19 |
infinity | lilstevie: Eh? Well, we're ahead of Debian, that's all. 2.15 versus 2.13. | 09:19 |
infinity | So, if their build accidentally picks up any 2.14/2.15 symbols, no workie on Debian. | 09:19 |
infinity | Not a big deal. | 09:20 |
lilstevie | ah | 09:20 |
infinity | Maybe if I find some spare time post-release, I'll wrangle Debian to 2.15 as well. | 09:20 |
infinity | But I don't look forward to making sure all the weird (k*bsd, gah) patches forward-port cleanly. | 09:21 |
lilstevie | lol, yeah that does not sound fun at all | 09:21 |
lilstevie | by the way have you looked at the universal kernel thing at all for tegra? | 09:22 |
infinity | Haven't had any time to play, really. :/ | 09:23 |
infinity | It obviously won't happen for precise, with release being 1.5 weeks away. :P | 09:23 |
lilstevie | heh yeah sure | 09:23 |
lilstevie | more for future | 09:24 |
lilstevie | gives time to work on a proper solution | 09:24 |
lilstevie | we have some hacks in the tf201 kernel that may help expand it out a bit though | 09:25 |
lilstevie | https://github.com/EnJens/kernel_tf201_stock/commit/cbdf1403453c6a6304c25df94b4f371d229b9030 | 09:25 |
rbasak | What's the easiest/recommended way of telling armel and armhf apart from ubuntu userspace? | 09:40 |
rbasak | I know of dpkg-architecture but that needs it to be installed | 09:40 |
rbasak | ldd /bin/sh or something perhaps? | 09:41 |
infinity | rbasak: dpkg --print-architecture. | 09:41 |
infinity | rbasak: Anything else could be a false positive, due to multi-arch. | 09:42 |
rbasak | awesome, thanks! | 09:42 |
infinity | rbasak: (but if it's not a multi-arch system, testing for linkers and such works, yeah) | 09:42 |
infinity | rbasak: I would have assumed you'd already know of dpkg --print-architecture :P | 09:43 |
rbasak | I knew that dpkg knew the answer, just didn't know about that flag. | 09:43 |
infinity | Don't tell me that people are using uname hacks or something to determine i386/amd64. | 09:43 |
infinity | Which is, again, wrong. | 09:43 |
rbasak | They probably are | 09:44 |
rbasak | dpkg isn't cross-distro! | 09:44 |
infinity | Err, no, but you use the right tools on the right targets. | 09:44 |
infinity | userspace and kernel arch don't relate at all. :/ | 09:44 |
infinity | I really don't want to look at cobbler and maas and all that, do I? | 09:44 |
infinity | I'll just end up crying. | 09:45 |
rbasak | :) | 09:45 |
rbasak | cobbler is quite RH biased. | 09:45 |
nimesh_accenture | how do I put the pandaboard in fastboot mode? i'm basically stuck on : out/host/linux-x86/bin/fastboot oem format , it says waiting fro devices | 11:24 |
* ogra_ thinks thats better asked in an android channel :) | 11:26 | |
lilstevie | infinity: I have used the tag off uname in my installer shell script for ubuntu for the transformer cause I couldn't find a better way | 11:42 |
lilstevie | that works cross-platform | 11:42 |
ogra_ | lilstevie, not cross platform, but cleaner: use archdetect | 11:43 |
lilstevie | ok :) | 11:44 |
lilstevie | well to be honest when I get some time to tackle Qt being a PITA I will be migrating over to a GUI anyway | 11:45 |
infinity | lilstevie: Why would you need uname for a device-specific installer? It'll always be the same, surely? :P | 11:45 |
* ogra_ thought he had seen infinity say good night a while ago in a differnt channel ... | 11:52 | |
ogra_ | :) | 11:52 |
infinity | ogra_: I failed. | 11:53 |
ogra_ | as usual ... | 11:53 |
lilstevie | infinity: there is an executable that will only work in its flavor (i386 or x86_64) | 12:18 |
infinity | lilstevie: So, I can't install to transformer from my PPC machine? Sad. | 12:19 |
infinity | lilstevie: But yeah, uname's the wrong answer for that. I could have an x86_64 kernel and an i386 userspace, and the binary would still fail. | 12:20 |
lilstevie | true | 12:20 |
lilstevie | and no, you cannot because there isn't an nvflash for ppc | 12:20 |
xranby | how about using qemu-user and a i386 userspace on the ppc to run nvflash? | 12:21 |
lilstevie | well that could work | 12:22 |
lilstevie | however an upcoming program should handle that | 12:22 |
lilstevie | we have spent the past few months reversing the protocal | 12:23 |
xranby | how nice of you, setting up qemu-user are quite a pain :) | 12:25 |
xranby | i have mostly used it to run some of the android sdk binarys on my ac100 | 12:25 |
lilstevie | heh | 12:25 |
lilstevie | it doesn't help that there are 2 very different versions from HC to ICS though | 12:27 |
=== doko_ is now known as doko | ||
djszapi | ogra_: Hey, interesting that, ubuntu does not have /usr/include/linux/leds.h for the pandaboard | 13:48 |
ogra_ | did you look under /usr/include/arm-linux-gnueabihf/ ? | 13:50 |
ogra_ | and do you have the necessary -dev packages installed that contain the headers | 13:51 |
djszapi | root@linaro-ubuntu-desktop:~# find /usr/include/arm-linux-gnueabi/ -name \*led\* | 13:52 |
djszapi | root@linaro-ubuntu-desktop:~# | 13:52 |
djszapi | find /usr/include/ -name \*led\* -> returns nothing | 13:53 |
djszapi | interesting patch from the near past: | 13:54 |
djszapi | http://lwn.net/Articles/492249/ | 13:54 |
djszapi | http://us.generation-nt.com/answer/patch-arm-omap4-pandaboard-turn-phy-reference-clock-at-init-help-201703292.html -> linux/leds.h should exist, but it does not apparently exist for some reasons. :/ | 13:55 |
ogra_ | it definitely exists in some package | 13:56 |
ogra_ | install apt-file and search for it ;) | 13:56 |
djszapi | I would hope for its availability by default on the pandaboard... | 13:57 |
ogra_ | not if its in i.e. linux-ti-omap4-dev or some such | 13:58 |
ogra_ | we dont install dev packages by default | 13:58 |
djszapi | uncool on a dev board | 13:58 |
ogra_ | well, we do, but only a selected minimal set | 13:58 |
ogra_ | we also dont special case our images for "dev boards" :) | 13:58 |
ogra_ | thats something linaro does | 13:59 |
djszapi | well, it is a linaro-ubuntu-desktop after all. | 13:59 |
ogra_ | well, if you used a linaro image, ask them about the header ;) | 13:59 |
djszapi | #linaro ? | 13:59 |
ogra_ | ubuntu just tries to build as close to the other areches as possible if it comes to images | 14:00 |
djszapi | hmm, apt-file update downloads a lot of stuff... | 14:00 |
ogra_ | its like apt-get ... carries its own db though | 14:01 |
ndec | ogra_: hi. is 12.04 LTS for ARM? | 14:02 |
ogra_ | only on server afaik | 14:02 |
ogra_ | havent heard any final words about LTS yet though | 14:02 |
ndec | ok. | 14:02 |
ndec | btw, what does 'only server' mean? you can install any package even on server, no? | 14:03 |
ogra_ | sure, but only packages in the server set will be cared for during the 5 year period in that case | 14:04 |
djszapi | ogra_: the newest packages are "linux-headers-3.0.0-17" and "linux-headers-3.0.0-17-omap". | 14:04 |
ogra_ | desktop gets updates from the x86 uploads indeed, but nobody would look into arm specific issues after 18 months | 14:04 |
djszapi | ...and I am having this: uname -r | 14:04 |
djszapi | 3.1.1-26-linaro-lt-omap | 14:04 |
ogra_ | djszapi, you look for omap4 | 14:04 |
ogra_ | and *not* a linaro kernel (unless thats what you currently use) | 14:05 |
djszapi | I do ? | 14:05 |
djszapi | linux-headers-3.0.0-1208-omap4 -> this is the newest I found. | 14:05 |
ogra_ | yeah, that sounds right | 14:05 |
ogra_ | though there might be a 3.2 one as well | 14:05 |
ogra_ | or should | 14:05 |
djszapi | not after apt-get update | 14:06 |
ogra_ | linux-ti-omap4 3.2.0-1412.16 was the last upload to 12.04 | 14:06 |
djszapi | it is oneiric... | 14:07 |
ogra_ | oh, then that could be right | 14:07 |
ogra_ | infinity, https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2012-April/013579.html did that accidentially change ? | 14:27 |
gsnedders | Hey, is 12.04 going to be LTS on ARM? | 15:15 |
GrueMaster | gsnedders: That is the working theory, except it will probably only apply to the pool, not to any specific platform image. | 15:26 |
gsnedders | GrueMaster: k, thx! | 15:27 |
ogra_ | gsnedders, only server packages will be LTS on arm | 15:27 |
ogra_ | and only on omap4 (though that doesnt matter much for userspace) | 15:28 |
ogra_ | https://wiki.ubuntu.com/PrecisePangolin/ReleaseManifest | 15:28 |
=== mckoan is now known as mckoan|away | ||
gsnedders | ogra_: I guess things like libGL aren't server? | 15:30 |
ogra_ | right | 15:30 |
jimerickson | todays image works. updated and installed pvr sgx driver. all is well. | 15:54 |
pnphi | How to build package for ARM in chroot ? | 15:57 |
smplman | jimerickson: what hardware>? | 15:57 |
jimerickson | oh sorry the pandaboard ES | 15:57 |
pnphi | ARM ubuntu | 15:57 |
=== zyga is now known as zyga-food | ||
smplman | jimerickson: i can't get sec working on my beagle board xm | 16:01 |
smplman | sgx* | 16:01 |
nimesh_accenture | ok guys.. so i've succesfully managed to chroot ubuntu-core 11.10 on android ICS 4.03 ... gtg now ... and thx for helping me out! | 16:02 |
ogra_ | smplman, please file a bug ... i think we missedf to add autodetection support to jockey for omap3 ... while omap4 landed (so you get the sgx driver automatically after first boot) | 16:03 |
ogra_ | (file it against the "jockey" package) | 16:03 |
smplman | how are you all pulling the binary bits from TI? | 16:04 |
ogra_ | they hand them to linaro, linaro prepares a package, some ubuntu dev uploads to the ubuntu archive | 16:04 |
pnphi | How to build package for ARM in chroot ? | 16:04 |
=== Quintasan_ is now known as Quintasan | ||
djszapi | ogra_: Hey! Is it normal if the Status Led 1 still keeps blinking after running a "sync(); reboot(LINUX_REBOOT_CMD_POWER_OFF);" from my application so the board is essentially powered off. | 16:35 |
djszapi | ogra_: what twl6030 power features have you put in there ? | 16:44 |
djszapi | this is the "uname -a" output: Linux linaro-ubuntu-desktop 3.1.1-26-linaro-lt-omap #26~lt~ci~20120325001352+1332635991~4f6ec49b-Ubuntu SMP PREEMPT armv7l armv7l armv7l GNU/Linux | 16:44 |
GrueMaster | djszapi: This is probably best asked on #linaro, as it is their kernel, not the stock Ubuntu kernel. | 16:54 |
ogra_ | yeah, thats not our kernel | 17:01 |
djszapi | perhaps not | 17:01 |
=== Matt_O2 is now known as Matt_O | ||
=== zyga-food is now known as zyga-afk | ||
=== prpplague^2 is now known as prpplague | ||
infinity | ogra_: --as-needed has been the default since oneiric. Not sure where this person got "a few days ago". | 19:49 |
micahg | hehe, we even fiddled with it during the natty dev cycle and reverted at the last minute | 19:51 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!