[10:40] hi [10:40] lool: can i disturb you? [10:41] lool: yesterday you told me to add console= to the command line [10:41] but should i add only "console=" or should i put something else after it? [10:46] goshawk: See the sample command line I pasted [10:46] goshawk: 20:44 < lool> something like http://people.ubuntu.com/~lool/qemu-buildd/run [10:46] wow [10:46] can i add it to the rootfromscratch page? [10:46] in the ubuntu wiki? [10:46] i surfed on google and a lot of people have troubles [10:47] with this console issue [10:48] it works! [10:48] goshawk: The rootfs from scratch page is about creating a rootfs not starting a qemu [10:48] thanks a lot lool! [10:48] lool: there is a section in starting qemu too [10:48] wait [10:48] goshawk: You're welcome to start a wiki page about qemu console options [10:49] goshawk: That's probably where you want to add the information then [10:49] lool: Using a qemu image [10:49] in https://wiki.ubuntu.com/ARM/RootfsFromScratch [10:49] yep [10:49] i want to add this info there :) [10:50] goshawk: Oh sure, that makes sense then [10:50] Only the console arg is missing actually [10:51] yes [10:51] apart from the fact that rootfs from scrtch doesnt enable a serial console and the qemu command is supposed to run an SDL login [10:51] ogra: i was thinking to add... if you wanna run it in a console do bla bla bla [10:51] can you add a completely new section with a proper qemu commandline as well as instructions how to ser up a serial tty ? [10:52] yep [10:52] there is no serial tty configured by default [10:52] though i was planning to att it in the next release of https://launchpad.net/project-rootstock :) [10:52] lool: there is a problem, after i see kernel message i just see "Freeing init memory: 136K" and nothing more, no login prompt [10:53] thats what i meant :) [10:53] goshawk: -nographic only supports the serial console, not vts [10:53] upstart isnt configured to start a serial tty [10:54] * goshawk remember that in a debian he just added -nographic to an armel image and it worked... uhm [10:54] ogra: so upstart should start a serial tty for it to work [10:55] right you need to edit the image [10:55] * goshawk removes the -nographic option and takes the console= from lool [10:55] see the /etc/event.d/ttyS0 in https://help.ubuntu.com/community/SerialConsoleHowto ... thats what you need to add [10:56] goshawk: Use alt-shift 3 to go to serial console [10:57] lool: thx, but i need to exit from the screen program, cuz alt-shift is got by its [10:58] ogra: uhm.. so the script to generate the image should add the SerialConsole which is in that page, isn't it? [10:58] well or you loop mount the image and add it [10:59] the next generation of the script (at https://launchpad.net/project-rootstock) will have an option for it [10:59] ogra: is it already implemented? [10:59] nope [11:02] ok thanks for all the suggestions [11:02] now i gotta go [11:02] thanks again :) [11:04] echo "start on runlevel 2" >/etc/event.d/ttyS0 [11:04] echo "start on runlevel 3" >>/etc/event.d/ttyS0 [11:04] echo "start on runlevel 4" >>/etc/event.d/ttyS0 [11:04] echo "start on runlevel 5" >>/etc/event.d/ttyS0 [11:04] echo "stop on runlevel 0" >>/etc/event.d/ttyS0 [11:04] echo "stop on runlevel 1" >>/etc/event.d/ttyS0 [11:04] echo "stop on runlevel 6" >>/etc/event.d/ttyS0 [11:04] echo "respawn" >>/etc/event.d/ttyS0 [11:04] echo "exec /sbin/getty 115200 ttyS0" >>/etc/event.d/ttyS0 [11:04] ^^^^ [11:04] add that between the "echo LANG=${NEWLOCALE} >>/etc/environment" and the first sed command [11:04] then it will create /etc/event.d/ttyS0 by default on your next build [11:05] ah :) [11:05] good! [11:06] thx a lot ogra [11:06] :) [11:07] ogra: what aobut editing /etc/inittab? [11:07] and add #T0:23:respawn:/sbin/getty -L ttyS0 9600 vt100 [11:07] its just a quick hack ... project-rootstock will have it cleaner [11:07] no [11:07] there is no inittab in ubuntu since 6.06 anymore [11:07] just add the above and you will have a serial tty [11:07] ah so the page about serial console is outdated [11:07] no [11:07] you read it to sloppy :) [11:08] --- Configuring inittab (Dapper and older) --- [11:08] sorry [11:08] i didn't read this line [11:08] ;) [11:08] :) [11:08] should be a proper header then it would be bold [11:10] Oh Debian has instructions to cross debootstrap too [11:10] http://wiki.debian.org/ArmEabiHowto didn't see these until now [11:10] * goshawk creates a new image [11:11] lool: i used it before [11:11] is has networking configured too [11:11] on qemu [11:11] we were based on it on mer development [11:11] is https://help.ubuntu.com/community/SerialConsoleHowto clearer now ? [11:12] ogra: thx, now it's better [11:12] good ... [11:12] the evil is that it needs a completely new section for karmic since yesterday [11:12] so it will become more confusing [11:12] what happened yesterday? [11:12] upstart changed completely [11:12] new upstart version? [11:12] no more /etc/event.d [11:13] * goshawk hopes it changed better [11:19] ogra: upstart is madness [11:19] once it works as it should it will be sexy i guess [11:19] no more event.d? what does it use -then-? [11:20] the years long transition is madness, i agree [11:20] /etc/init? I'm not sure [11:20] /etc/init/ [11:20] and a new file format [11:20] yes [11:20] lovely [11:20] files have .conf suffixes now and the command syntax in them changed [11:21] ogra@osiris:/var/build$ cat /etc/init/tty1.conf |grep -v ^# [11:21] start on stopped rc RUNLEVEL=[2345] [11:21] stop on runlevel [!2345] [11:21] respawn [11:21] exec /sbin/getty -8 38400 tty1 [11:21] it's quite the same [11:22] time to add to the serial console page? [11:23] well, even though i assume you only need to change the getty line it should be tested before being added to the wiki :) [11:27] ogra: I think uboot also supports ext2 (marvell's does); could we perhaps use that to avoid a separate /boot if the host is installed with ext2/3/4? [11:27] well, we default to ext4 [11:28] Yes [11:28] * goshawk I: Unpacking the base system... [11:28] is that ext2 backwards compatible ? [11:29] like ext3 ? [11:29] * ogra thought it wasnt [11:29] as far as i know there is a tool to switch from ext3 to ext4 [11:30] but i've never used it [11:30] yes, but can you mount etx4 as ext2 ? [11:30] dunno [11:30] ogra: I don't know [11:30] you can mount any ext2 or ext3 filesystem as ext4 without any changes [11:30] from google [11:30] ogra: But perhaps uboot logic will work with ext2+? [11:31] no idea :) [11:31] ogra: Might be worth investigating for your fsl spec perhaps? [11:31] http://www.plugcomputer.org/plugwiki/index.php/U-Boot_Quick_Reference#Load_kernel_.2Fboot.2FuImage.sheeva.20090319_from_ext2.2Fext3.2Fext4_filesystem_.28first_partition.29_on_USB_device [11:31] sheva seems to be able to use ext4 [11:32] ogra: I think we also want to find a way to allow upgrades from jaunty [11:32] so it might be possible [11:32] ugh, shudder [11:32] that might get hairy [11:32] replacing the botloader and all [11:32] I think we can manage with flash-kernel.conf being present or not [11:32] we dont do that on x86 [11:33] I don't think we want to replace it [11:33] though grub1 will still work there [11:33] oh, you want to boot .31 on B1 with old redboot ? [11:33] hmm, i wonder if that will work [11:33] Yes [11:35] It depends whether the proto (which should be fairly stable) changed or not between 28 and 31 [11:35] Dunno what broke between 26 and 28 [11:35] amitk, did you try booting .31 on B1 ? [11:36] ogra: Could you cleanup the whiteboard in https://bugs.launchpad.net/ubuntu/karmic/+source/banshee/+bug/391588 [11:36] Launchpad bug 391588 in banshee "banshee fails to run on arm" [High,New] [11:36] release team meeting today [11:37] you mean rip out the stuff from the description ? [11:37] sure [11:37] ogra: Make it an attachment or something [11:38] ogra: Don't forget this stuff is copied on ALL emails [11:38] It would also be nice if you could get a new backtrace [11:38] ripped out [11:38] i'll attach the LVDS to my pegatron after lunch and try to get a proper backtrace together [11:39] sadly you cant run banshee remotely ... [11:39] dies becaue of dbus [11:55] well ogra, after the changes you told me [11:55] i built a new image [11:56] and booted with [11:56] sudo qemu-system-arm -M versatilepb -kernel vmlinuz-2.6.28-versatile -cpu arm926 -hda qemu-armel-200907101210.img -m 256M -append "root=/dev/sda mem=256M" -nographic [11:56] still no login [11:56] ogra: yes I did. Fail. But please feel free to try it again. [11:56] adding the console= settings from lool [11:56] yeah, you need to tell the kernel to use a serial console [11:57] goshawk: Use graphics or setup a serial console [11:57] goshawk: Which upstart version in your image? [11:57] lool: wait i see in the log [11:58] by default the script still builds jaunty [11:58] ogra: Are you still planning to deal with https://bugs.launchpad.net/bug-zapper/+bug/392094 ? I don't see an assignee, but if you don't intend to work on it that's fine [11:58] Launchpad bug 392094 in bug-zapper "needs search option for bugs a team is subscribed to" [Undecided,Confirmed] [11:58] lool, i will, but its not on my high prio list [11:58] lool: yep it's jaunty version [11:58] though given the babbage situation i might have a lot of spare time :P [11:59] ogra: Hmm you're assigned to the karmic task in https://bugs.launchpad.net/gnome-keyring/+bug/328167 [11:59] Launchpad bug 328167 in gnome-keyring "[arm] gnome-keyring-daemon eating 100% CPU at login in Jaunty" [High,Fix released] [12:00] right, i need to set up a jaunty on my B1 [12:00] its just a confirmation for upstream though [12:00] not sure we need to target it to karmic [12:00] amitk, bjf_afk: Whenever you're updating imx51's udebs, could you please address https://bugs.launchpad.net/ubuntu/+source/linux/+bug/359049 ? [12:01] Launchpad bug 359049 in linux "imx51 udeb hardcodes linux version in vmlinuz binary name" [High,Won't fix] [12:01] * goshawk creates a new image with --seed openssh-server since he can ping the last image but he can't connect. [12:02] did the serial console not come up ? [12:02] ogra: no [12:02] same behaviour as before [12:02] not even with proper console option? [12:02] neither with console=ttyAMA0 console=tty0 [12:03] and -nographic of course [12:03] console=ttyAMA0,115200n8 [12:03] ok trying [12:03] thats what the build script uses :) [12:04] yep i'm starting to see something more [12:04] so if you see it building after "I: Switching to Virtual Machine for second stage processing" that should work [12:04] reeing init memory: 136K [12:04] * Filesystem type 'fusectl' is not supported. Skipping mount. [12:04] * Setting preliminary keymap.. [12:04] perfect [12:04] yep it's working [12:04] :) [12:04] :) [12:04] well finally we have a working version ^__^ [13:07] ogra: poke [13:07] ogra: I reviewed https://wiki.ubuntu.com/Specs/ArmelSubarchitectureImageBuilds and it looked ok [13:07] * ogra falls over [13:07] cjwatson gave his ok as well already, david approved it [13:08] rootstock just grew a --serial and a --dist option [13:08] :) [13:14] :-) [13:30] hmm, i made 10 commits to the rootstock branchbut LP only gives me 7 points as top committer [13:31] * ogra wonders how these numbers correlate [13:34] hmm [13:34] did anyone of us ever try to run wine on armel ? [13:34] sqrt(commits) + 4 [13:35] should be impossible [13:35] would be funny if it worked ... [13:35] because wine is x86 only [13:35] to run office 2000 install ubuntu and wine on your arm device :) [13:35] yeah, likely not working [13:36] maybe you can start a project: wine embedded [13:36] why embedded ? [13:36] we dont do any embedded stuff :) [13:36] or mobile, ... [13:36] * ogra wishes that team name could be changed since a long time [13:37] if you want to run wine on arm the acronym fails ( wine is not an emulator) [13:38] you can ask sun to port virtaul box to arm [13:39] nah, they would probably only give me qemu-sparc :P === cbrake_away is now known as cbrake [15:01] DEC had a x86 emulation layer for the Alpha's, if there was one for ARM cortex 8, the x86 apps would probably run about K6 speeds === Nicke_ is now known as Nicke === bjf_afk is now known as bjf === cbrake is now known as cbrake_away