[06:55] <error404notfound> Hi!
[07:11] <error404notfound> how easy it is to install ubuntu on an arm machine with no usb port, no cd rom, no ethernet, just wifi and 3g?
[07:14] <persia> Not very.  You'd need some way to download enough to boot, and then be able to boot off that.
[07:15] <error404notfound> persia: hmmm, actually I am really confused that if it doesn't have a usb port, a cdrom, or ethernet, how can even copy anything from network to it. I don't know from where should I even start
[07:16] <persia> I'd recommend starting by looking for someone else who has gotten any linux running on your device.
[07:17] <persia> You can probably adjust those instructions as required to have the install be Ubuntu.
[07:18] <error404notfound> I am following your first recommendation by joining this channel...
[07:20] <persia> In that case, you might want to mention the specific device name :)
[07:21] <error404notfound> hmmm, wanna know the truth? I still haven't got one, its stuck for clearance on airport and I am waiting for it... may take couple of days.. :P
[07:46] <Stskeeps> sometimes i want to shoot whoever chose right click as paste.
[07:47] <persia> Oughtn't be right-click.  Ought be middle-click.
[07:47] <Stskeeps> yeah
[08:02] <tomsh> is it normal when using build-arm-rootfs
[08:02] <tomsh> * Setting kernel variables (/etc/sysctl.d/10-network-security.conf)...          error: "net.ipv4.tcp_syncookies" is an unknown key
[08:02] <tomsh> * Setting kernel variables (/etc/sysctl.d/10-process-security.conf)...          error: "vm.mmap_min_addr" is an unknown key
[08:02] <tomsh> ?
[08:33] <persia> tomsh: It depends on your kernel configuration, really.  It's been reported before.
[08:56] <lool> persia: Do you know whether we have bug for these?  It shouldn't happen with the versatile kernel as it's a distro kernel
[08:57] <lool> And the imx51 one is as well now
[08:59] <persia> We don't have a bug for it: it's never been reported against a distro kernel (that I've heard).
[14:00] <lool> NCommander: Do you think you could report these issues as well as the babbage ones?
[14:01] <lool> NCommander: Tag them arm
[14:04] <ogra> lool, they are on my list of bugs for amitk_
[14:04] <ogra> net.ipv4.tcp_syncookies definately shows on babbage as well
[14:04] <lool> ogra: Which list?
[14:04] <ogra> my personal one i'nm assembling while testing the babbage kernel here
[14:05] <ogra> klogd not starting is another one ...
[14:06] <ogra> though currently i'm after finding out whats wrong with aufs/apparmor that makes all ttys die and X not start up
[14:06] <ogra> thats more important than cosmetical boormessages
[14:06] <ogra> *boot
[14:10] <ogra> lool, i wont manage to file these bugs right now, can you mention in the release meeting that there are imx51 bugs to come we will need to fix before beta ?
[14:11] <lool> ogra: No
[14:11] <lool> I'll give a general status update
[14:11] <ogra> ok
[14:11] <lool> I remember quite well the day were I was shot in place for mentionning bugs which I hadn't filed yet
[14:12] <ogra> well, they are all under investigation only atm ...
[14:13] <ogra> the imx51 kernel isnt 100% merged and i assume a lot will be solved once that happened
[14:20] <lool> NCommander: Please Cc: slangasek on the RedBoot build issues with native toolchain
[14:22] <obra> Hi. I'm seeing some weird behaviour with ubuntu (and debian) arm ports on qemu. It _appears_ that some of the time the SCSI device just stops responding.  It appears to be fairly random. When it happens, spurious "a"s get printed to my terminal.
[14:23] <obra> I'm seeing the behaviour across several versions of qemu on OSX and Linux.
[14:23] <obra> Is there an obvious FAQ I'm missing?
[14:40] <Stskeeps> versatile kernel or?
[14:41] <obra> versatile kernel.
[14:41] <obra> /usr/local/qemu/bin/qemu-system-arm -M versatilepb -m 256 -kernel vmlinuz-2.6.28-8-versatile -hda armel.img -append "mem=256M root=/dev/sda1"
[14:42] <obra> (Starting from the initrd images and doing an install)
[14:42] <obra> I get the same behaviour using http://people.debian.org/~aurel32/qemu/armel/
[16:25] <Martyn> Good morning.
[16:25] <Martyn> Ogra, look, amit .. any of you online?
[16:26] <Martyn> lool rather
[16:27] <lool> Martyn: I'm around, but I'm out for an hour or so
[16:27] <lool> I'll be back here later
[16:32] <Martyn> Allright.  I'm trying to do a first bringup of ubuntu on the board, and wanted to verify some settings
[16:33] <amitk_> Martyn: is this something quick?
[16:33]  * amitk_ is about to wrap up for the day
[16:39] <Martyn> amitk : Relatively.  The image that ogra put up (the SD card image) seems to have reboot on it, and the first partition is a few sectors from the beginning of the card...
[16:39] <ogra> right
[16:40] <ogra> i think i left 20M ... though 15M shoudl suffice
[16:40] <Martyn> When I put it in the board and power up, I do see the keyboard flash, but no console output or display output.. is that normal for this image?
[16:40] <Martyn> yep, about 20M
[16:40] <ogra> you should get console output on VGA
[16:40] <Martyn> dreck, I don't.
[16:40] <ogra> after redboot decompressed the kernel
[16:40] <ogra> probably an issue with your DIP switch settings
[16:40] <Martyn> would you mind going through a quick board comparison?  I want to make sure the DIP switches on mine match yours.
[16:41] <Martyn> the two that are together at the top, read from SATA->pushbutton are:
[16:41] <Martyn> 0-0-0-0-0-0-0-1 1-0-0-0-0-0-0-0
[16:41] <ogra> the middle ones are  on
[16:42] <ogra> right
[16:42] <Martyn> the one next to the SD card is (read from VGA to serial breakout)
[16:42] <Martyn> sorry, read from serial->vga direction
[16:42] <ogra> the second should be on
[16:42] <Martyn> 0-1-0-0-0-0-0-0
[16:42] <Martyn> Okay, those match
[16:42] <ogra> looks ok
[16:42] <Martyn> So, in theory it should have just worked
[16:43] <ogra> yes
[16:43] <Martyn> I'll try a different SD card then.   Darnit!
[16:43] <ogra> note the kernel doesnt support DVI
[16:43] <Martyn> Wait, I thought the kernel didn't support VGA...
[16:43] <ogra> you need to use VGA
[16:43] <ogra> no
[16:43] <Martyn> okay, switching connectors.
[16:43] <Martyn> and switching to a real SanDisk card
[16:44] <ogra> yup, sandisk ++
[16:44] <Martyn> okay, I'm going to re dd it to a sandisk 2Gb card.
[16:46] <Martyn> allright, once the DD is finished, I'll try again.
[16:46]  * ogra finally got the initramfs booting for amitk's latest kernel 
[16:47] <Martyn> congrats!
[16:47] <ogra> well, i dont see the bugs fixed yet ...
[16:47]  * ogra waits for casper to find the squashfs 
[16:48] <ogra> amitk, tty's still die :(
[16:48] <ogra> amitk, oh, but you added fusectl ....
[16:48] <ogra> at least it doesnt complain anymore
[16:49] <ogra> oh, and X now starts up but i get a "your session lasted less than 10 secs."+ message now
[16:49] <amitk> hmm
[16:49] <Martyn> ogra : At least X starts up.
[16:50] <ogra> hmm, that looks a bit like the former error though, i see attempts fopr autologin on tty1
[16:50] <amitk> ogra: this is live image?
[16:50] <ogra> yep
[16:50] <amitk> with union=aufs?
[16:50] <ogra> without union=
[16:50] <ogra> which defaults to aufs
[16:50] <ogra> i see it loading aufs
[16:50] <amitk> ok
[16:50] <ogra> switching to console i see multiple motd messages
[16:51] <ogra> something is wrong with the login, but its hard to debug since i konw if i boot into single mode i will see everything with the user is right
[16:52]  * ogra will boot into single and start executing the initscripts piece by piece ... time consuming, but probably i see where i hit the issue
[16:53] <ogra> oh
[16:53]  * Martyn chuckles, the dd is still going :)
[16:53] <Martyn> this isn't my first dance with new hardware, but I'm frankly amazed at how fast you guys at canonical are doing the bring-up of this one
[16:54] <ogra> well, we know our tools :)
[16:54] <ogra> so its just a matter of adapting them
[16:54] <Martyn> HUZZAH
[16:54] <Martyn> we're up.
[16:54] <ogra> the main issue was to get initramfs working with redboot
[16:55] <ogra> everything beyond that is just small adjustments
[16:55] <ogra> congrats !
[16:55] <Martyn> gets as far as sd
[16:55] <Martyn> then stops
[16:55] <ogra> give it time :)
[16:55] <Martyn> timeout issue?
[16:55] <ogra> it looks for the squashfs that takes some seconds ...
[16:56] <ogra> amitk, klogd still dies
[16:56] <Martyn> there it goes..
[16:56] <Martyn> I should probably disable SATA
[16:56] <Martyn> okay, I'm seing the tty messages
[16:56] <Martyn> nice X
[16:56] <Martyn> and there we go.  Thanks ogra :)
[16:57] <ogra> enjoy
[16:57] <ogra> note that its not installable
[16:57] <Martyn> This is a much faster bring-up than I usually have to do (kernel, then busybox, then ...)
[16:57] <Martyn> That's fine.  This is enough for us to do some very, very important tests for the work we're doing.
[16:58] <ogra> but you could for example use debootstrap or the rootfs script from the channel topic and install a system to a usb disk
[16:58] <Martyn> *nod*
[16:58] <Martyn> Or sata?
[16:58] <ogra> then just change the cmdline in redboot to not use casper
[16:58] <Martyn> Or is boot-to-sata not quite yet ready for prime time?
[16:58] <ogra> and add root=UUID=<uuid of your disk partition>
[16:58] <ogra> i havent tried sata at all yet
[16:58] <amitk> usb disk should work well
[16:59] <ogra> and i think sata is just a fake ...
[16:59] <ogra> in the backend its attached to usb
[16:59] <Martyn> Oh, you think there's a bridge chip?
[16:59] <ogra> yes
[16:59] <Martyn> The mx51's spec shows a SATA controller though.
[16:59] <Martyn> well, doesn't matter :)
[16:59] <ogra> well, feel free to try it :)
[17:00] <ogra> (if amitk has the sata driver enabled)
[17:00] <Martyn> I'm pleased as punch to have a working install.   I'll track your builds from this point.
[17:00] <Martyn> It looked like it was at least trying to bring sata up during kernel initialization
[17:00] <amitk> ogra: no sata drivers yet IIRC
[17:00] <Martyn> amitk : Danke.
[17:00] <ogra> right, thats what i thought
[17:01] <amitk> Martyn: I'll flesh out support for _every_ storage driver as a module in the next upload
[17:02] <ogra> if we find one that doesnt make the consoles die :P
[17:02] <Martyn> amitk : Ooooo :)
[17:03] <Martyn> ogra : Just out of curiosity, where are the authentication token error messages coming from?
[17:03] <ogra> which one ? "your password is expired" ?
[17:04] <Martyn> yep
[17:04] <ogra> run "sudo hwclock --systohc" and reboot ...
[17:04] <ogra> (if your system clock is right at least)
[17:04] <ogra> that fixes it
[17:05] <ogra> useradd somehow sets the expiry time to zero if the system date is at 01 01 1970
[17:05] <ogra> setting the hwclock fixes it
[17:07] <Martyn> fixed, and thanks.
[17:07] <Martyn> worth adding to your README
[17:08] <Martyn> now I want the serial debug dongle ... -sigh-
[17:08] <ogra> worth fixing the bug rather :)
[17:09] <Martyn> being a unionfs based image, I'm assuming any changes I make to the FS will be lost at reboot?
[17:09] <Martyn> very much worth fixing the bug!  It's just more of an appnote for that very first release of the image
[17:14] <Martyn> thanks :) You guys just made my day.
[17:16] <ogra> amitk, hmm, klogd tries to access /dev/console
[17:16] <ogra> and then dies
[17:16] <Martyn> interesting!
[17:16] <Martyn> when the hwclock is set correctly, the tty's also come to life
[17:16] <ogra> yep
[17:17] <ogra> the autologin works then
[17:17] <Martyn> it does indeed
[17:17] <ogra> upstart makes them die if the user isnt set up correctly
[17:17] <Martyn> no modules.dep file?
[17:17] <ogra> nope, next image will have the kernel package inside
[17:17] <Martyn> heh, no modules at all :)
[17:17] <Martyn> *chuckle*
[17:17] <ogra> and have the modules.dep
[17:17] <Martyn> should I apt-get it?
[17:18] <ogra> apt-get install linux-imx51
[17:18] <Martyn> done
[17:18] <ogra> that will get you everything you need
[17:18] <ogra> note that the NIC driver is shaky
[17:18] <Martyn> Yeah, I know.  Works okay though.
[17:19] <Martyn> you guys should be very, very proud of this.   Most hardware first bring-ups are a hell of a lot more shaky than this.
[17:19] <Martyn> This is practically almost useable in comparison to some projects I've worked on.
[17:20] <ogra> :)
[17:23] <Martyn> Wow, everything is arm V7 clean?
[17:23] <Martyn> no v4 compiles?
[17:36] <Martyn> what's the process I have to kill to prevent the 100% cpu spinning?
[17:36] <Martyn> gnome-keyring?
[17:46] <Martyn> ogra, how do I get the live image to boot single user, if I don't have access to the serial console?
[17:55] <Stskeeps> Martyn: think it's armv5te targeted, not armv7
[18:05] <Martyn> Stskeeps : Thanks.
[18:06] <Martyn> I was wondering how far the system was going to be optimised...
[18:06] <Martyn> v5 is good enough
[18:06] <Stskeeps> think they're looking at HWCAP of libraries too, which makes sense
[18:14] <ogra> Martyn, we might switch to v7 completely with 9.10
[18:14] <ogra> its still being benchmarked what we gain from it and will be discussed at next UDS
[18:15] <Stskeeps> or loose in terms of supporting less devices, heh
[18:15] <Martyn> ogra : *nod*
[18:16] <Martyn> ogra : If I don't have access to the redboot monitor (thanks to having no serial debug board), how can I get the liveimage you've posted to come up single user?
[18:16] <ogra> not at all ...
[18:17]  * Martyn tried hexediting the redboot, to change the kernel bootparams, but that didn't work :)
[18:17] <ogra> lool is working on something we should see next week that will make it possible to dd changes into the fis partitions
[18:17] <Martyn> Is there a way to bring down X safely once the livecd has booted, without triggering a reboot?
[18:18] <ogra> you can switch to a console and sudo /etc/init.d/gdm stop i guess
[18:18] <Martyn> Tried that.  No joy.
[18:19] <ogra> well, both options you can set are cmdline options to get into single
[18:19] <ogra> and the cmdline is only accessible via serial atm
[18:19] <ogra> one would be to add "text" to it, the other would be "single"
[18:19] <Martyn> yep
[18:19] <ogra> text actually gets you a full system, while single gets you only a recovery console
[18:20] <Martyn> Oh well :)  I guess I have to be patient and wait until Monday when a serial debug dongle will arrive as if by magic.
[18:20]  * ogra was pondering to add a menu to the boot  ....
[18:20] <ogra> but i have to much to do to track down the kernel issues on the weekend ...
[18:21] <ogra> if i magically find the solution i might look into adding some bootmenu