[07:55] <lool> Daviey: Hey, what went wront with the netboot installer images?
[07:55] <lool> Daviey: Oh in qemu
[07:55] <lool> Daviey: qemu doesn't support dove nor imx51
[07:56] <lool> Daviey: We stopped doing versatile because we moved karmic to armv6 and versatile in the upstream kernel implies armv5 -- note that it can perfectly work with v6 or v7 though
[08:13] <Daviey> lool: So does that mean that it's currently note possible to virtualise arm in karmic?
[08:17] <lool> Daviey: It means that there's no prebuilt kernel to run in our qemu indeed
[08:20] <lool> Daviey: While it's an useful thing and we have a lot of people running qemu which hang around here, it's not a requirement; you could ask for such a flavour to the kernel team, I personally think it would be useful
[08:20] <lool> I didnt because I've felt I needed them to focus on the flavours supporting our hardware first
[09:20] <ogra_> Daviey, http://people.canonical.com/~ogra/arm/qemu/vmlinuz-2.6.31-rc3versatile1-cortex-a8 might be of some help
[09:20] <ogra_> note it's quite old, but enough to get a VM up
[09:21] <ogra_> (the rootstock builder from /topic uses it to create rootfs'es and qemu images for arm)
[09:26] <Daviey> ogra_: thanks.. i'll give it a go.
[09:26] <ogra_> note there are no modules for it or anything ...
[09:26] <Daviey> ogra_: incidently, the upgrade path cannot be canculated from jaunty to karmic. :(
[09:27]  * ogra_ has no time/resources to build and patch a fresh one 
[09:27] <Daviey> :(
[09:27] <ogra_> in a jaunty qemu vm you mean ?
[09:27] <Daviey> yah
[09:28] <ogra_> thats intentional, though i thought i added a proper error message
[09:28] <ogra_> do a cat /proc/cpuinfo in your vm
[09:28] <ogra_> karmic only supports v6 and v7
[09:29] <Daviey> ogra_: do-release-upgrade craps out as it tries to remove ubuntu-minimal, but that is blacklisted and marked for non-removal
[09:29] <ogra_> u-m should pop up an error message telling you that pre-v6 isnt supported
[09:30] <ogra_> oh, not sure do-release-upgrade shows the same errors u-m does
[09:30] <Daviey> ogra_: just booting now.
[09:32] <Daviey> lool: Sorry for not replying, missed your response.  I understand :)
[09:33] <Daviey> ogra_: Processor: ARM926EJ-S rev 5 (v5l) :(
[09:33] <ogra_> right
[09:33] <ogra_> my kernel fixes that
[09:33] <ogra_> it has a patch lool developed to fake a cortex-a8
[09:35] <Daviey> hmm, made qemu core dump :(
[09:37] <ogra_> you indeed need the right -M
[09:39] <ogra_> err
[09:39] <ogra_> -cpu, not -M
[09:39] <ogra_> -cpu cortex-a8
[09:39] <ogra_> and -M versatilepb
[09:40] <Daviey> yah, ogra_ you rock. :)
[09:40] <ogra_> :)
[10:03] <erikcorry> I finally got around to filing http://gcc.gnu.org/bugzilla/show_bug.cgi?id=41354
[10:03] <ubot4> gcc.gnu.org bug 41354 in c++ "g++: Inlining constructors puts wrong vtable in objects" [Normal,Unconfirmed]
[10:03] <erikcorry> clever bot
[11:58] <rabeeh> j #openplug
[13:13] <mike^> ogra, thanks to you I have omap3evm with ubuntu :)
[13:13] <ogra> :)
[13:14] <mike^> it's somewhat slow on app startup, but even 3D works :)
[13:14] <mike^> http://2.bp.blogspot.com/_dhW-lh4q-sk/Sqli478fQGI/AAAAAAAAAYo/zHk7-iRgQ1k/s1600-h/ubuntu-greeter3-small.png
[13:14] <mike^> http://4.bp.blogspot.com/_dhW-lh4q-sk/SqlivloG3fI/AAAAAAAAAYg/3-K5pHMwEa8/s1600-h/xfce2.png
[13:16] <mike^> ogra, do you know by chance is anybody working on the DSP stuff for OMAP?
[13:16] <ogra> nobody from canonical at least
[13:17] <ogra> not sure if there are any community people
[13:17] <siji> hey mike^, still am struggling with the same issue
[13:18] <siji>  checking GLES/egl.h usability... no
[13:18] <siji>  checking GLES/egl.h presence... no
[13:18] <siji>  checking for GLES/egl.h... no
[13:18] <siji>  configure: error: Unable to locate required GLES headers
[13:25] <lool> ogra: I was thinking of something: libc-vfp was dropped but wont be removed on upgrades which mean people might run an out of date libc
[13:25] <lool> ogra: Would you be interested in an update-manager snippet to drop it on upgade?
[13:25] <ogra> another u-m hack ?
[13:25] <lool> *upgrade
[13:25] <lool> Yeah that's what I had in mind
[13:25] <lool> Would check with mvo what he things though
[13:25] <ogra> sure, why not
[13:25] <lool> ogra: I'll file that now while I think of it and assign it to you then; thanks
[13:27] <ogra> gar
[13:27]  * ogra messed up his bug 
[13:28] <lool> ogra: Actually thinking about it it's more of an eglibc bug, it should conflict/replace
[13:29] <ogra> yup
[13:30] <lool> ogra: there's an = dep on libc6 which is why it gets removed on upgrades
[13:30] <lool> I guess that's ok
[13:32] <ogra> god, pushing an ubiquity branch takes a century
[13:33] <lool> Yeah would need an upgrade
[13:34] <ogra> i thought its under ubuntu-core-dev but it isnt ... so i need to create a brandnew branch for merging
[13:34] <ogra> OO.o looks better this time ... didnt fail yet
[13:34] <lool> Well LP uses stacking for modern .bzr branches
[13:34] <ogra> ah
[13:34] <ogra> doesnt help me right now though :)
[13:34] <ogra> but my initramfs-tools install prob is solved
[13:35] <lool> Cool
[13:35] <lool> the unr and umr livefs breakages are awful just before A6
[13:40] <mike^> siji, native as well?
[13:40] <mike^> siji, I haven't tried to cross-compile... and apparently won't for some time
[13:40] <ogra> lool, well, i didnt have armel livefses since a few days either
[13:44] <lool> ogra: Do they build now?
[13:44] <lool> ogra: I gave back eds over the WE as I saw it was blocking but didn't look since
[13:45] <ogra> e-indicator and e-couchdb were ftbfs as well
[13:45] <ogra> but they should be fine now
[13:45] <ogra> i'll trigger a testbuild for dove and imx51
[13:57] <ogra> lool, do you know if the dove installer bits are ready so far ?
[13:57]  * ogra thinks there is still something missing, but not sure
[14:13] <tasslehoff> I'm trying to get my beagleboard to send 1680x1050 to my display (default it is 640x480). I followed the "Xorg omapfb Drivers"-section on http://elinux.org/BeagleBoardUbuntu, but afterwards X wouldn't start.
,bad luck
[14:15] <siji> :(
[14:16] <siji> have u updated the kernel
[14:16] <siji> ?
[14:18] <tasslehoff> siji: no, I have just created my own rootFS following the wiki, and started working/installing on top of that
, you have to add some patches in the kernel
[14:20] <lool> ogra: NCommander said the boot.scr doesn't get generated; probaby a simple bug
[14:20] <siji> for getting SGX Video Acceleration
[14:21] <ogra> lool, but flash-kernel and friends are ready ?
[14:22] <lool> ogra: Well it's a bug in flash-kernel-installer.postinst
[14:22] <ogra> yeah
[14:23] <ogra> i meant the rest of the changes there
[14:23] <ogra> ubiquity surely needs changes too for the bootloader parts
[14:23] <ogra> i didnt see him touching anything there yte
[14:23] <ogra> *yet
[14:23] <tasslehoff> siji: ok, I'll try to follow the wiki all the way to the bottom before asking any more :)
[14:24] <siji> yes
[14:24] <siji> it's there in the wiki
[15:06] <erikcorry> Does -mcpu=vfp get me a binary that uses floating point instructions?
[15:19] <tasslehoff> On my Beagle Board Xubuntu install update-apt-xapian-index is using almost all the cpu, and all I have set running myself is an xterm..
[15:21] <lool> erikcorry: That sounds wrong; you want fpu
[15:22] <lool> erikcorry: if you're using karmic, armv6 + vfp (v2) is the default
[15:22] <lool> So no need for any flag
[15:22] <lool> If using jaunty you want --with-float=softfp --with-fpu=vfp
[15:22] <lool> Sorry these are gcc build flags
[15:23] <lool> You want -mfloat-abi=softfp and -mfpu=vfp
[15:28] <erikcorry> Isn't the float ABI softfp by default?
[15:28] <erikcorry> Or does the mfpu option change that?
[15:30] <tasslehoff> Is the OMAP35x SDK needed to get SGX Video Acceleration on the Beagleboard?
[15:34] <lool> erikcorry: Our jaunty ABI is soft
[15:35] <lool> not softfp
[15:42] <erikcorry> What I want is a program that uses fp hardware, but is ABI compatible with the Karmic ABI.
[15:42] <erikcorry> So I think -mfpu=vfp is about right.  No change to the ABI, but use vfp inside functions.
[15:42] <erikcorry> Unless I am misunderstanding something.
[15:45] <lool> erikcorry: what's your dist?
[15:45] <lool> jaunty or karmic?
[15:45] <erikcorry> Karmic
[15:45] <lool> erikcorry: So you dont need -mfpu=vfp
[15:45] <lool> it's the default
[15:45] <erikcorry> So Karmic doesn't work on a machine with no vfp?
[15:45] <lool> No
[15:45] <erikcorry> Ah!
[15:46] <erikcorry> Good to know!
[15:46] <lool> And for jaunty you do need both
[15:46] <lool> Because one tells gcc that it may use vfp instructions in binaries but not in the ABI and actually should and the other tells which fpu instructions to use
[15:46] <erikcorry> Do Jaunty binaries run on a Karmic machine?
[15:46] <lool> Yes
[15:47] <erikcorry> Thanks
[15:49] <erikcorry> And Karmic also requires ARMv6?
[15:50] <ogra> yes
[15:52] <Daviey> Would someone be really kind, and test some packages cleanly install on armel for me?  It's a dkms package, and i want to test it also works of armel before getting it uploaded?
[15:52] <Daviey> s/of/on
[15:53] <ogra> all macines i could test on are busy for the next few hours
[15:53]  * Daviey has an n810.. but possibly not the best native arm device to test on :)
[15:53] <Daviey> ogra: oh okay, thanks.
[15:53] <ogra> especially for dkms stuff :)
[15:54] <Daviey> i've now got a qemu machine, that is almost on karmic.. but i'm not sure it's a fair test, considering how the kernel lives outside.
[15:56] <Daviey> i'm wondering if my toaster would run faster than qemu. :)
[15:57] <ogra> likely