pwnguin | is there a list of ftbfs for arm? | 00:51 |
---|---|---|
pwnguin | the wiki has one last updated in like november | 00:52 |
prpplague | davidm: ping | 01:14 |
prpplague | Mirell: greetings | 02:08 |
Mirell | 'allo | 02:52 |
=== pwnguin_ is now known as jldugger | ||
ogra | jldugger, http://qa.ubuntuwire.com/ftbfs/ | 08:17 |
amitk | ogra: just finished reading the massive scrollback from yesterdays apex debugging. Impressive work! | 08:26 |
amitk | ogra: So it seems it isn't a kernel problem? | 08:27 |
* amitk starts to read up on apex a bit | 08:27 | |
ogra | amitk, well, upstream answered my last mail on debian arm ... where i said "hello, my kernel panics if i choose more than 5M for the ramdisk size" ... with "yes, you can use up to 8M for the ramdisk" | 08:32 |
ogra | its not all solved yet, but i have a workaround for now | 08:32 |
ogra | whih i intend to upload so we get working daily builds as soon as there is a kernel ... but its something we surely need to revisit | 08:33 |
ogra | *which | 08:33 |
amitk | ogra: re: as soon as there is a kernel? | 08:34 |
ogra | the kernel corruption thing itself is solved ... | 08:34 |
ogra | amitk, our endiness is wrong | 08:34 |
ogra | *endianess | 08:34 |
amitk | ogra: ah right, i'll get that fixed | 08:34 |
amitk | anything else you need in the ixp4xx image for this upload besides endianness? | 08:34 |
ogra | and if i use the kernel not swappped with devio it stops after unpacking (not sure thats related to endianess mistmatches with apex, kernel and initramfs but i suspected so | 08:35 |
ogra | only the NIC firmware ... doing a netinstall without NIC is quite tricky ;) | 08:36 |
amitk | and that will be required for d-i too | 08:36 |
ogra | right, but d-i builds udebs from linux-firmware | 08:37 |
ogra | just a matter of adding the right udeb to the netinstall image | 08:37 |
ogra | i might have kernel changes afterwards though, i couldnt test our kernel at all yet | 08:37 |
ogra | (i.e. i dont know how it bevhaves in an installed system ...) | 08:39 |
ogra | lool got it to boot directly from redboot during the sprint (which is why i suspect the non booting to just be an endianess issue in d-i) but we werent able to change the cmdline in redboot as everything at that stage is readonly | 08:40 |
ogra | amitk, did you get your serianl port to work ? | 08:41 |
ogra | *serial | 08:41 |
amitk | ogra: removing the solder blobs is being a problem, i've got to go and get a desoldering braid | 08:43 |
ogra | yeah, though davidm managed to just blow it out of the holes, i had to use a pump | 08:44 |
amitk | blowing it sounds dangerous, though he has a ton more experience than I do | 08:44 |
alesan | hi | 17:56 |
alesan | I have this board with an ARM, 128MB of memory, USB, ethernet, serial, VGA, audio, PS/2; I wrote all the drivers. Can I run Ubuntu on it? | 17:58 |
ogra | sure | 18:07 |
ogra | see the last link in the topic | 18:07 |
alesan | well | 18:08 |
alesan | I think my board is not a ready target | 18:08 |
alesan | it will require the proper drivers for example | 18:09 |
alesan | even an headless image would require my ethernet drivers | 18:10 |
alesan | or - maybe I'm mistaken and this rootfs does not contain the kernel | 18:10 |
ogra | right, its only the userspace | 18:12 |
ogra | if yu have a kernel, use that, if you can push your drivers upstream to kernel.org we can fully support your board in jaunty+1 | 18:13 |
alesan | yeah :) that would be nice | 18:45 |
alesan | just as an information, when is likely to be the "freeze" for j+1 ? | 18:45 |
ogra | hard to predict, the j+1 schedule will only be finalized at the next developer summit ... which is in may | 18:48 |
ogra | jaunty releases in april ... feature freeze for jaunty is next week | 18:49 |
ogra | feature freeze is usually the point where stuff like that needs to be in ... given that j+1 will happen in october it might be around august | 18:50 |
alesan | yeah but it is hard by that time my drivers will be in an official kernel release | 18:58 |
alesan | it's a lot of stuff in many subsystems | 18:59 |
ogra | well, if you can get them into .29 or .30 we'll get them for free from upstream | 19:00 |
ogra | then supporting your HW is a non issue, just a matter of enabling the .config | 19:00 |
alesan | yeah | 19:01 |
alesan | listen as a side question | 19:01 |
alesan | openoffice is actually running on ARM? | 19:02 |
ogra | on some HW, yes | 19:02 |
ogra | i doubt you will get it running properly on 128M though | 19:02 |
alesan | do you have an idea of the requirements? | 19:02 |
ogra | it will *run* but not be fun | 19:02 |
alesan | or an hardware that is known to run it? just as comparison with what I have here on this board | 19:03 |
ogra | if you want to support ubuntu-desktop including firefox and openoffice your HW should have above 256M and above 500MHz | 19:03 |
alesan | yeah :) it's a pity, I remember when I was able to run office97 on a 16MB PC :( nowadays 128MB are too little | 19:03 |
ogra | i doubt there is any HW public yet | 19:03 |
alesan | ah ok | 19:03 |
ogra | http://www.notebooks.com/2009/01/07/new-generation-of-netbooks-199-and-299-eight-hour-battery-sexy-design/ | 19:05 |
ogra | something like that will likely run ubuntu-desktop incl. openoffic fine | 19:05 |
alesan | by the way, what about Koffice, it used to be lighter than OOo | 19:09 |
* ogra hasnt tried any QT based stuff on arm yet | 19:09 | |
ogra | i know xubuntu and gcoffice work fine though | 19:09 |
ogra | even on the beagleboard i have if i add some swap and a USB disk | 19:09 |
ogra | (so that IO times are not to huge) | 19:10 |
alesan | gcoffice = ?? I only found a german cheese office with that name :) | 19:11 |
ogra | abiword and gnumeric | 19:11 |
ogra | the actual so called "gnome office" | 19:12 |
alesan | the moment you create this rootfs... what format is that? it is likely the way to mount it is with a USB key right? | 19:46 |
ogra | alesan, its a tgz that you csn untar on any kind of partition (given it has the right size) | 21:59 |
ogra | *can | 21:59 |
ogra | if you use the --notarball option it will create a qemu image instead | 21:59 |
ogra | so you can play with it in qemu-system-arm | 21:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!