**** BEGIN LOGGING AT Tue Sep 21 02:59:57 2010
Sep 21 02:59:59 <jo-erlend> rsalveti, can you be a little more spesific? :)
Sep 21 03:00:05 <rsalveti> or wait until I get my package uploaded, so you can install it by using qemu
Sep 21 03:00:24 <rsalveti> jo-erlend: which ubuntu release are you using at your host?
Sep 21 03:00:30 <jo-erlend> maverick.
Sep 21 03:00:31 <rsalveti> if you're using ubuntu :-)
Sep 21 03:00:35 <rsalveti> cool
Sep 21 03:00:56 <rsalveti> so you could get inside your partition with chroot and qemu, to install the needed packages
Sep 21 03:01:05 <rsalveti> like running it at your arm machine
Sep 21 03:01:33 <rsalveti> mount your sd card at your host computer
Sep 21 03:01:40 <rsalveti> then get inside the rootfs partition
Sep 21 03:01:42 <rsalveti> cp /usr/bin/qemu-arm-static usr/bin/
Sep 21 03:01:48 <rsalveti> mount --bind /dev dev
Sep 21 03:01:59 <rsalveti> mount -t proc proc proc/
Sep 21 03:02:14 <rsalveti> and then you can just give chroot /media/<rootfs>
Sep 21 03:02:21 <devilhorns> persia, that's the idea :)
Sep 21 03:02:36 <rsalveti> now you're running an arm environment, and able to give apt-get normally
Sep 21 03:03:00 <rsalveti> then all you need to do is to get my kernel package and give dpkg -i on it
Sep 21 03:03:25 <rsalveti> umount everything, boot your arm board and be happy with it
Sep 21 03:04:24 <persia> One may well have to install qemu-kvm-extras-static to make that procedure work...
Sep 21 03:04:46 <rsalveti> yup, was too lazy to find the package for it :-)
Sep 21 03:06:30 <jo-erlend> rsalveti, I'm in an arm chroot with the rootfs from my memorycard.
Sep 21 03:07:52 <jo-erlend> ... now what? :)
Sep 21 03:09:28 <rsalveti> jo-erlend: create the correct kernel directory for the kernel I sent you
Sep 21 03:09:50 <rsalveti> mkdir /lib/modules/2.6.35.4+
Sep 21 03:09:51 <rsalveti> I guess
Sep 21 03:10:05 <rsalveti> then run depmod with correct parameter, one sec
Sep 21 03:12:11 <rsalveti> depmod -a 2.6.35.4+
Sep 21 03:12:14 <rsalveti> please try it
Sep 21 03:12:53 <jo-erlend> uh, how? :)
Sep 21 03:13:15 <rsalveti> at your chroot
Sep 21 03:13:19 <rsalveti> mkdir /lib/modules/2.6.35.4+
Sep 21 03:13:23 <rsalveti> depmod -a 2.6.35.4+
Sep 21 03:13:27 <jo-erlend> I did that. No response
Sep 21 03:13:33 <rsalveti> that's fine
Sep 21 03:13:43 <rsalveti> ls -l /lib/modules/2.6.35.4+ should show you some files
Sep 21 03:13:49 <rsalveti> now exit your chroot
Sep 21 03:13:51 <rsalveti> umount dev
Sep 21 03:13:52 <rsalveti> umount proc
Sep 21 03:13:56 <rsalveti> and umount the rootfs
Sep 21 03:14:03 <rsalveti> put your card at your board and boot it
Sep 21 03:14:16 <jo-erlend> it does.
Sep 21 03:14:50 <jo-erlend> you mean I exit my chroot by issuing those commands?
Sep 21 03:15:55 <rsalveti> no, after you get out your chroot
Sep 21 03:16:06 <jo-erlend> how do I do that? "exit"?
Sep 21 03:16:10 <rsalveti> yep
Sep 21 03:16:13 <rsalveti> or control + d
Sep 21 03:20:28 <jo-erlend> rsalveti, well. Some of the errors have gone, but I'm stuck with one: "mount: unknown filesystem type 'binfmt_misc'
Sep 21 03:20:47 <jo-erlend> and that's where it stops.
Sep 21 03:21:04 <rsalveti> hm, ok, so it's really missing some modules
Sep 21 03:21:17 <rsalveti> jo-erlend: http://people.canonical.com/~rsalveti/kernel/linux-image-2.6.35.4+_2.6.35.4+-43_armel.deb
Sep 21 03:21:27 <rsalveti> you can find all kernel modules you need on it
Sep 21 03:22:07 <jo-erlend> then I enter a chroot again, download that package and install it?
Sep 21 03:22:30 <rsalveti> jo-erlend: yup
Sep 21 03:22:41 <rsalveti> jo-erlend: but I'm afraid you'll get fs errors
Sep 21 03:22:48 <rsalveti> let's see
Sep 21 03:23:03 <rsalveti> just dpkg -i <deb>
Sep 21 03:27:12 <jo-erlend> hmm. it seems I don't have networking in the chroot?
Sep 21 03:27:33 <jo-erlend> oh. nvm
Sep 21 03:28:19 <rsalveti> you have it, could be a problem with your resolv.conf file
Sep 21 03:30:42 <persia> jo-erlend, You have qemu-kvm-extras-static installed?
Sep 21 03:32:29 <jo-erlend> possibly. Why?
Sep 21 03:32:53 <rsalveti> persia: he was able to get inside the chroot before
Sep 21 03:33:03 <rsalveti> so it should be able to get inside it again
Sep 21 03:33:09 <jo-erlend> rsalveti, when installing the package inside the chroot, I got _lots_ of "Unsupported ioctl: cmd=0xc020660b"
Sep 21 03:33:16 <rsalveti> jo-erlend: that's normal
Sep 21 03:33:50 <persia> Well, rather, it's a bug, but it doesn't break installing a kernel
Sep 21 03:34:59 <jo-erlend> rsalveti: dpkg: error prosessing linux-image-blabla: failed in write on buffer copy for backend dpkg-deb during /lib/modues/2.6.35.4+/kernel/drivers/media/common/tuners/tda18271.ko": no space left on device
Sep 21 03:35:10 <rsalveti> jo-erlend: :-(
Sep 21 03:35:20 <rsalveti> that's the fs error I was afraid of
Sep 21 03:35:31 <rsalveti> this is because your kernel never got the resize
Sep 21 03:35:52 <rsalveti> jo-erlend: so you need to start from beginning
Sep 21 03:36:03 <jo-erlend> how much from the beginning?
Sep 21 03:36:05 <rsalveti> dd the daily build
Sep 21 03:36:12 <rsalveti> hm
Sep 21 03:36:17 <jo-erlend> oh..
Sep 21 03:36:29 <rsalveti> probably by copying the uInitrd from the daily image
Sep 21 03:36:33 <rsalveti> and uImage
Sep 21 03:36:57 <rsalveti> here's the problem, you need to at least boot the image using the original uImage and uInitrd once
Sep 21 03:37:26 <rsalveti> it'll just resize it (with some additional things)
Sep 21 03:37:30 <rsalveti> and reboot
Sep 21 03:37:35 <jo-erlend> oh. The guide from igeps wiki told me to delete those.
Sep 21 03:37:45 <rsalveti> jo-erlend: yep, the guide is wrong :-(
Sep 21 03:37:45 <jo-erlend> well, everything except uImage.
Sep 21 03:38:16 <rsalveti> after the first boot and reboot, you're able to do whatever you need/want with your image
Sep 21 03:38:27 <jo-erlend> ok. Is it sufficient to copy the uInitrd back into the boot partition and run it?
Sep 21 03:38:29 <rsalveti> without facing these weird fs errors
Sep 21 03:38:34 <rsalveti> probably
Sep 21 03:39:34 <jo-erlend> do I need MLO and u-boot.bin?
Sep 21 03:39:38 <rsalveti> nops
Sep 21 03:39:41 <rsalveti> yours should be fine
Sep 21 03:40:42 <rsalveti> jo-erlend: just make sure you have the right commands to load uInitrd and uImage at your boot.scr or boot.ini
Sep 21 03:40:46 <rsalveti> depends on what you have
Sep 21 03:40:50 <rsalveti> http://paste.ubuntu.com/497451/
Sep 21 03:40:53 <rsalveti> example from beagle
Sep 21 03:41:00 <rsalveti>         fatload mmc 0:1 0x80000000 uImage
Sep 21 03:41:01 <rsalveti>         fatload mmc 0:1 0x81600000 uInitrd
Sep 21 03:41:08 <rsalveti> than your boot cmd line
Sep 21 03:41:14 <rsalveti> and:
Sep 21 03:41:15 <rsalveti>         bootm 0x80000000 0x81600000
Sep 21 03:41:23 <rsalveti> with that you should be able to load uImage and uInitrd
Sep 21 03:41:45 <rsalveti> the uInitrd is the responsible for calling jasper, the one responsible for resizing it
Sep 21 03:42:37 <jo-erlend> oh, ok.
Sep 21 03:43:06 <rsalveti> jo-erlend: well, need to go now, quite late already
Sep 21 03:43:19 <rsalveti> jo-erlend: please try using uInitrd and let me know what happens
Sep 21 03:43:43 <rsalveti> if not, the recommended step would be to start from the beginning, but without removing the uInitrd file
Sep 21 03:43:44 <jo-erlend> will do.
Sep 21 03:43:51 <rsalveti> jo-erlend: just don't give up :-)
Sep 21 03:44:00 <rsalveti> I'm able to help you again tomorrow
Sep 21 03:44:05 <jo-erlend> I've seen text! That gives me courage. :)
Sep 21 03:44:11 <rsalveti> jo-erlend: :-)
Sep 21 03:44:23 <jo-erlend> thank you very much :)
Sep 21 03:44:26 <rsalveti> jo-erlend: if you're able to test, please update bug 607250
Sep 21 03:44:27 <ubot2> Launchpad bug 607250 in linux (Ubuntu) "omapdss: VDDA_DAC regulator on IGEPv2 (affects: 2) (dups: 1) (heat: 24)" [Undecided,Confirmed] https://launchpad.net/bugs/607250
Sep 21 03:44:41 <rsalveti> once we have a successful test, I'll send to the kernel team
Sep 21 03:45:01 <jo-erlend> one more thing...
Sep 21 03:45:08 <rsalveti> the kernel patches, if you're interested: http://kernel.ubuntu.com/git?p=rsalveti/ubuntu-maverick.git;a=shortlog;h=refs/heads/fix-igepv2-display
Sep 21 03:45:18 <rsalveti> sure
Sep 21 03:45:18 <jo-erlend> in that fatload directive, you use some numbers. Are they ok to use?
Sep 21 03:45:35 <rsalveti> yup, should be fine
Sep 21 03:46:50 <jo-erlend> what about those other lines?
Sep 21 03:47:07 <jo-erlend> "fixrtc" should I add that? Should I add the bootm directive?
Sep 21 03:47:07 <rsalveti> the third one should be your own cmd line
Sep 21 03:47:12 <rsalveti> that you were using before
Sep 21 03:47:24 <rsalveti> you could, should cause no harm
Sep 21 03:47:45 <rsalveti> bootm 0x80000000 0x8200000 is just to load from the memory location
Sep 21 03:48:09 <rsalveti> fatload mmc 0:1 0x80000000 uImage means to load uImage from mmc0 partition 1 to memory addr 0x80000000
Sep 21 03:48:25 <rsalveti> the bootm says that you should now boot from this memory location
Sep 21 03:48:28 <Martyn> ah, the joys of uboot
Sep 21 03:48:30 <jo-erlend> ah.
Sep 21 03:48:56 <persia> Martyn, You have a grub port yet?
Sep 21 03:49:16 <rsalveti> jo-erlend: time to go now, anything else I can do to help?
Sep 21 03:49:55 <jo-erlend> I guess not today, but if you're around tomorrow, I'm sure I will have something. :)
Sep 21 03:50:03 <jo-erlend> thanks again for your help.
Sep 21 03:50:42 <rlameiro> wow that was a tough ride
Sep 21 03:50:55 <rsalveti> jo-erlend: sure, np!
Sep 21 03:50:57 <rsalveti> see ya
Sep 21 03:51:06 <Martyn> persia : There will be no grub port.  Once we added Jason and Rob to the SStone team, we went for u-boot as the solution
Sep 21 03:51:06 * rsalveti out for bed
Sep 21 03:51:11 <rlameiro> hours of debbuging
Sep 21 03:51:16 <jo-erlend> rsalveti, sleep well. :)
Sep 21 03:51:25 <jo-erlend> rlameiro, I got more normal text now! Improvement. :)
Sep 21 03:51:29 <persia> Martyn, Awww..  but...  :(
Sep 21 03:51:30 <Martyn> persia : So we're concentrating on that, and adding what we need to u-boot to make the platform work
Sep 21 03:52:10 <Martyn> persia : No buts .. it came down to return for the time spent on it .. and grub2 was going to be too much work to be worth it.
Sep 21 03:52:36 <Martyn> persia : Michael's software bootloader also went by the wayside .. it's always the same story.
Sep 21 03:52:54 <Martyn> BUT .. we got pxe working .. and that's something
Sep 21 03:52:59 <persia> That's something.
Sep 21 03:53:14 <persia> Now I just have to find another group that's more idealistic than underfunded :)
Sep 21 03:53:45 <rlameiro> lol
Sep 21 03:54:23 <rlameiro> is uboot beeing customized for ubuntu?
Sep 21 03:54:44 <rlameiro> detection routines?
Sep 21 04:00:18 <persia> rlameiro, I believe the plan is to mostly use linaro's uboot, and leave the coordination between the many uboot trees to the linaro project.
Sep 21 04:00:24 <Martyn> yep
Sep 21 04:00:39 <Martyn> and eventually probably move to UEFI, but who knows when that will work...
Sep 21 04:00:42 <rlameiro> nice
Sep 21 04:00:56 <persia> We're good at making a distribution, but we're less good at making individual bits of software good for everyone.  Some of us work upstream in various ways, but as a team, it's better to focus on integration.
Sep 21 04:01:07 <rlameiro> I never got that linaro heads stuff
Sep 21 04:01:10 <persia> Martyn, UEFI works without a second-stage?
Sep 21 04:01:35 <Martyn> UEFI works when device tree is finished
Sep 21 04:01:53 <persia> rlameiro, I don't understand linaro "heads" stuff either, but the linaro project does seem to be putting a lot of effort into collaboration between different trees, making it *lots* easier for us to select the right thing when building a general-purpose distribution.
Sep 21 04:02:02 <Martyn> and yeah .. UEFI is a complete little boot shell
Sep 21 04:02:08 <Martyn> 'little'
Sep 21 04:02:34 <rlameiro> kinda lilo ? hehe
Sep 21 04:02:36 <persia> So, what was the deal with all the amd64 grub2 stuff this cycle to make it work for UEFI?
Sep 21 04:03:11 <Martyn> persia : Because I went to LinuxCon:Boston .. my boss won't let me to go LinuxCon Japan :( :(
Sep 21 04:03:15 <Martyn> I'm a sad panda
Sep 21 04:03:41 <persia> Clearly you shouldn't have attended LC:B :)
Sep 21 04:04:02 <persia> MIssed you last time, but if you come visit when I'm actually on this side of the world, be fun to do something.
Sep 21 04:04:14 <Martyn> Yeah :)
Sep 21 04:04:18 <Martyn> Coming to UDS:Orlando?
Sep 21 04:08:24 <persia> That's the current plan.
Sep 21 04:11:42 <Martyn> Great :)  See you there :()
Sep 21 04:11:43 <Martyn> :)
Sep 21 04:14:12 <DanaG1> oopsie, that was me having two computers with pidgin open.
Sep 21 04:16:49 <persia> DanaG, Thanks for sorting it quickly :)  By the way, you might be interested in using smuxi or bip or similar sorts of things if you often want to IRC from multiple computers.
Sep 21 04:18:07 * GrueMaster uses quasselcore on a server and connects to it from multiple laptops.
Sep 21 04:19:44 <GrueMaster> persia: While i am thinking of it.  I filed two regression bugs; Bug 644028 & Bug 644037
Sep 21 04:19:49 <ubot2> Launchpad bug 644028 in linux-meta-mvl-dove (Ubuntu) "Audio regression on Marvell Dove images (affects: 1) (heat: 8)" [High,New] https://launchpad.net/bugs/644028
Sep 21 04:19:50 <ubot2> Launchpad bug 644037 in pulseaudio (Ubuntu) "defaults need adjustment for Dove A0 for audible audio (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/644037
Sep 21 04:20:14 <persia> GrueMaster, Why is 644037 a pulse bug?
Sep 21 04:20:35 <persia> Surely it's best to fix in ALSA or udev or the kernel.
Sep 21 04:20:40 <GrueMaster> It is a duplicate of Bug #451635
Sep 21 04:20:42 <ubot2> Launchpad bug 451635 in pulseaudio (Ubuntu Karmic) (and 8 other projects) "defaults need adjustment on dove X0 for audible audio (affects: 1) (heat: 11)" [Undecided,New] https://launchpad.net/bugs/451635
Sep 21 04:20:51 <GrueMaster> That was fixed in Lucid.
Sep 21 04:21:17 <persia> But fixed with a dirty hack, and nobody went back to clean it up, and now we have a regression :(
Sep 21 04:21:19 <Martyn> So much for getting 10.10 installed on an imac 27" ... display port detection is broken in the current kernel
Sep 21 04:21:31 <Martyn> I was kind of hoping I could convert this thing to a useful machine :)
Sep 21 04:22:08 <persia> GrueMaster, Grab a kernel guy, and make alsa/udev see the right thing and report the right thing to pulse :)
Sep 21 04:22:42 <persia> A good fix for 644028 might also fix 644037
Sep 21 04:22:44 <GrueMaster> persia: That may be part of the fix needed for the first bug.
Sep 21 04:23:38 <GrueMaster> I'll test the Beta image on beagle tomorrow to see if it regressed as well.
Sep 21 04:24:06 <GrueMaster> At any rate, wife calls.  Must go veg in front of the TV before bed.
Sep 21 04:34:59 <DanaG> I still have yet to see where you get one of those "dove" boards.
Sep 21 04:35:32 <Martyn> ddon't bother
Sep 21 04:35:47 <Martyn> the panda board is as good for development :)
Sep 21 04:36:26 <Martyn> the Marvell boards were useful stepping stones, but I see the panda board as being a good measuring stick
Sep 21 04:37:05 <DanaG> I mean, I've never even seen so much as a picture of said "dove" boards... the only things I've seen retail are the old Kirkwood.
Sep 21 04:40:02 <persia> I don't believe there are any retail dove boards.
Sep 21 04:40:24 <persia> I believe the only retail stuff on which Ubuntu is easily available is imx51 or omap3 currently.
Sep 21 04:42:37 <Martyn> yep
Sep 21 04:42:59 <Martyn> the Marvell boards I got were direct from the engineering group
Sep 21 04:43:19 <Martyn> the whole point of the Dove boards is like the Lange boards .. engineering testing and development
Sep 21 07:23:56 <hrw> morning
Sep 21 11:01:22 <ogra> bah, sigh
Sep 21 11:01:43 <ogra> two openoffice build. one linux-linaro and more linux builds in the queue
Sep 21 11:02:08 * ogra guesses we wont see much from the buildds today
Sep 21 11:05:38 <amitk> why two OO.org?
Sep 21 11:06:09 <ogra> maverick and a lucid SRU
Sep 21 11:10:42 <persia> maverick is a good thing.  This one should build.
Sep 21 11:16:02 <ogra> its the same source for both
Sep 21 11:22:48 <mk0> is it possible to install ubuntu-arm on very old dell axim x5?
Sep 21 11:23:35 <ogra> what CPU is in it ?
Sep 21 11:23:54 <hrw> it is not
Sep 21 11:23:59 <hrw> mk0: its armv5
Sep 21 11:24:06 <ogra> should be fine with jaunty
Sep 21 11:24:08 <hrw> pxa27x iirc
Sep 21 11:24:12 <ogra> but nothing newer
Sep 21 11:24:29 <hrw> ogra: jaunty works on 64MB ram?
Sep 21 11:24:30 <ogra> (userspace at least, kernel and bootloader is up to you)
Sep 21 11:24:33 <persia> I'd recommend installing Debian on an ARMv5 device.
Sep 21 11:24:41 <persia> hrw, Sufficiently stripped, yeah.
Sep 21 11:24:51 <ogra> hrw, yes, we offered an NSLU2 image in jaunty
Sep 21 11:25:06 <ogra> its super painful, but works ... kind of
Sep 21 11:25:51 <mk0> oh shi... too advanced answers
Sep 21 11:25:55 <mk0> for me
Sep 21 11:27:12 <hrw> ogra: nslu2 can have 256MB ram ;D
Sep 21 11:27:25 <hrw> all depends on soldering skills
Sep 21 11:27:26 <ogra> we only used the default
Sep 21 11:31:36 <hrw> I want to drop my beagleboard to floor and jump on it few times
Sep 21 11:31:53 <persia> You'd end up having to get a new one ...
Sep 21 11:32:02 <hrw> persia: I have other C3 here
Sep 21 11:32:16 <hrw> but it will not change situation. even worse
Sep 21 11:32:23 <persia> mk0, Short answer: you'd do better to install Debian.  Ubuntu support for that device happened once, but isn't continuing.
Sep 21 11:32:24 <hrw> both my C3 have usb problem
Sep 21 11:32:45 <hrw> but one which I use has expansion header soldered so usb fix was easier
Sep 21 11:33:06 <mk0> persia, thanks) will lurk towards deb
Sep 21 11:34:16 <ogra> but even there you might need to do a kernel and bootloader setup on your own
Sep 21 11:35:11 <persia> bootloader at least.  I believe there to be pxa2xx kernels in Debian
Sep 21 11:38:25 <zumbi> persia: what?
Sep 21 11:38:40 <zumbi> pxa2xx kernels in debian :?
Sep 21 11:39:03 <zumbi> persia: i use pxa270 board
Sep 21 11:39:04 <persia> zumbi, Maybe it was just leftover (unused) config in a kernel source I read once.  I haven't looked recently.  I'll trust your word over my memory.
Sep 21 11:39:59 <zumbi> persia: debian arm sadly does not provide pxa2xx support :/
Sep 21 11:40:31 <zumbi> persia: i am fighting kernel guys and see if we are lucky to include mv78xxx machines
Sep 21 11:40:51 <persia> It's always a matter of having kernel folk to support stuff.
Sep 21 11:40:56 <zumbi> but kernel build takes so long in native that it is a blocker for adding more platforms
Sep 21 11:41:12 <persia> Clearly you need more, faster, buildds :)
Sep 21 11:41:34 <zumbi> well.. now we have them, so we'll see if more platforms are added
Sep 21 11:42:00 <zumbi> but with the freeze, all development is somehow stalled on bugfixing
Sep 21 11:42:15 <persia> This is a good thing :)
Sep 21 11:42:47 <zumbi> well, it is good yes, for release (I hate releases :) )
Sep 21 11:43:55 <zumbi> btw, are you guys playing on omap4?
Sep 21 11:44:46 <persia> Some folk have dev boards.
Sep 21 11:45:14 <hrw> 10.10 on arm require 768MB ram or 512MB is enough?
Sep 21 11:45:19 <ogra> zumbi, well, we release images for it :)
Sep 21 11:45:38 <zumbi> ogra: omap4 seems to be a beast
Sep 21 11:45:44 <ogra> hrw, 256M should be enough, 512M is recommended
Sep 21 11:46:18 <persia> 384 is the regular requirement.  256 may work, depending.
Sep 21 11:46:31 <ogra> right
Sep 21 11:46:42 <ogra> 256M should kind of work
Sep 21 11:46:51 <ogra> with -fun :)
Sep 21 11:46:58 <persia> Just don't try to do too many things at once :)
Sep 21 11:47:04 <hrw> I need to kill plymouth, ureadahead and then maybe it will finally boot
Sep 21 11:47:09 <ogra> +fun starts at 512M and above :)
Sep 21 11:47:38 <ogra> hrw, ureadahead is supposed to be fixed in the recent images
Sep 21 11:47:39 <persia> +fun is *supposed* to happen at 384MB.  Do the requirements need a bump?  Do we need a general bloat reduction push again?
Sep 21 11:47:52 <ogra> nah
Sep 21 11:47:57 <ogra> +ok starts at 384
Sep 21 11:48:04 <ogra> +fun starts at 512 :)
Sep 21 11:48:18 <persia> Ah.
Sep 21 11:48:26 <hrw> ogra: first I need to boot to be able to run apt
Sep 21 11:48:40 <ogra> yeah, booting is such a thing ...
Sep 21 11:49:34 * ogra is really happy that all his ubuntu probs vanished when he started to use the tegra netbook
Sep 21 11:49:52 <hrw> ogra: how much it costs?
Sep 21 11:50:06 <ogra> 380-400 €
Sep 21 11:50:11 <persia> What do you mean "all ubuntu probs vanished"?  You aren't using Ubuntu anymore?
Sep 21 11:50:23 <ogra> persia, no, i have only android probs now
Sep 21 11:50:23 <hrw> ogra: too much ;)
Sep 21 11:50:42 <persia> heh.  Ubuntu problems *exist* but aren't quite as intense.
Sep 21 11:50:44 <hrw> nice... bb hang on
Sep 21 11:50:49 <hrw> beagle login: hrw
Sep 21 11:51:02 <suihkulokki> ogra: bought from where?
Sep 21 11:51:13 <ogra> suihkulokki, media markt in germany
Sep 21 11:51:18 <ogra> they sell it off the shelf here
Sep 21 11:52:00 <suihkulokki> interesting
Sep 21 11:52:03 <ogra> but its locked down and runs android ... i managed to unlock it and can run ubuntu chroots just fine for working ... but there is no way to access the bootloader to change the preinstalled os
Sep 21 11:52:35 <suihkulokki> did you write up hacking notes somewhere?
Sep 21 11:52:43 <ogra> someone did
Sep 21 11:53:50 <zumbi> suihkulokki: pandaboard is very much near if you want A9 hw
Sep 21 11:54:11 <zumbi> near in time
Sep 21 11:54:42 <hrw> nice... /var/lib/dpkg/status trashed
Sep 21 11:55:44 <ogra> suihkulokki, http://tosh-ac100.wetpaint.com/
Sep 21 11:57:24 <ogra> wow, the rooting howto is completely worng :P
Sep 21 11:57:47 <ogra>  /data is mounted noexec there is no way to run binaries in it ... funny
Sep 21 11:58:22 <ogra> i wonder if people test such stuff if they write howtos
Sep 21 11:59:42 <hrw> in glibc systems noexec has no meaning iirc
Sep 21 12:01:47 <NekoSchool> ouch, ac100 needs "rooting"? :D
Sep 21 12:02:00 <NekoSchool> damn toshiba :D
Sep 21 12:03:08 <ogra> NekoSchool, its android
Sep 21 12:03:17 <dcordes> ogra: I am sure there will be ways eventually
Sep 21 12:03:33 <NekoSchool> yeah sure but there isn't some nice standard bootloader in there like redboot or uboot with a serial console or something and some pads on the board for a serial cable?
Sep 21 12:03:42 <ogra> i dont think anyone ever shipped an andriod based device off the shelf with root enabled
Sep 21 12:03:59 <ogra> NekoSchool, nope, nothing like that
Sep 21 12:04:19 <ogra> its a patches down android fastboot bootloader it seems
Sep 21 12:04:20 <NekoSchool> even if you ship a completely locked down software on a device, this isn't a phone.. running ubuntu on it should be a nice feature for those who have the right SD card or something :D
Sep 21 12:04:39 <ogra> dcordes, ?
Sep 21 12:04:51 <dcordes> ogra: do boot your own kernel/initramfs
Sep 21 12:04:53 <ogra> dcordes, ways to install ubuntu you mean ?
Sep 21 12:04:58 <ogra> ah, likely
Sep 21 12:05:17 <ogra> the more people complain to toshiba the more likelier it gets
Sep 21 12:05:26 <dcordes> DIY
Sep 21 12:05:40 <ogra> it seems to use a raw partition for the bootloader stuff
Sep 21 12:05:49 * ogra made a dd copy from the eMMC
Sep 21 12:05:50 <dcordes> They don't have to let you boot your own kernel so I think it does not matter how many people will complain.
Sep 21 12:06:13 <ogra> first partition only starts at block 33 so there is "free space" before it
Sep 21 12:06:15 <dcordes> ogra: what's eMMC? Any documentation ?
Sep 21 12:06:36 <ogra> a builtin flash like device that exposes itself as MMC
Sep 21 12:06:36 <persia> dcordes, It's the same as MMC, except you can't remove it.
Sep 21 12:06:44 <dcordes> ok
Sep 21 12:07:02 <ogra> i think the e stands for embedded (not sure though)
Sep 21 12:07:13 <dcordes> ogra: can you provide the dd ?
Sep 21 12:07:29 <ogra> not sure if thats legal, so not publically, no
Sep 21 12:07:49 <dcordes> ok
Sep 21 12:08:29 <dcordes> Are any system upgrades anounced by toshiba ?
Sep 21 12:10:54 <ogra> apparently there is an andriod 2.2 one expected in the next 4-6 weeks
Sep 21 12:11:02 <ogra> oh !
Sep 21 12:11:09 <hrw> eMMC describes an architecture comprised of an embedded storage solution with MMC interface, flash memory and controller, all in a small ball grid array (BGA) package.
Sep 21 12:11:12 <hrw> It is based upon the industry-standard MMC system specification version 4.1/4.2 and JEDEC BGA packaging standards
Sep 21 12:11:15 * ogra just found a proper recovey mode by accident !
Sep 21 12:11:35 <hrw> thats from JEDEC informations
Sep 21 12:11:42 <dcordes> hrw: thanks
Sep 21 12:12:20 <dcordes> AC100 looks real nice. Checking some videos
Sep 21 12:13:06 <zumbi> hrw: hey! i got that jedec document :P
Sep 21 12:14:00 <persia> ogra, You mean that next time you brick it, you can recover without help?
Sep 21 12:14:20 <ogra> persia, not sure, depends if the bootloader is in two parts :)
Sep 21 12:14:33 * persia continues to hold off on a purchase decision
Sep 21 12:14:37 * ogra doesnt know where that recovery mode lives ... first or second stage
Sep 21 12:14:50 <ogra> and i wont play with it right now
Sep 21 12:15:03 * ogra has work stuff to do :)
Sep 21 12:15:07 <persia> Of course :)
Sep 21 12:17:25 <dcordes> ogra: Did you think about possibility to boot new kernel from within running system ?
Sep 21 12:17:52 <ogra> dcordes, that would require kexec support in the running kernel
Sep 21 12:17:57 <persia> dcordes, Most kernels provide a userspace environment that prohibits that.
Sep 21 12:19:37 <dcordes> grep -i kexec ac100_defconfig
Sep 21 12:19:43 <dcordes> CONFIG_KEXEC=y
Sep 21 12:21:36 <dcordes> ogra: j/k :>
Sep 21 12:21:42 <dcordes> ogra: you have a 3g device ?
Sep 21 12:22:12 <ogra> yes, but no sim
Sep 21 12:24:44 <dcordes> ogra: I have a hard time finding the ac100 kernel source code. Do you have a link ?
Sep 21 12:25:36 <ogra> dcordes, i dont think its public, toshiba will liekly ship it to you if you give them the product id
Sep 21 12:26:23 * ogra goes for some late lunch
Sep 21 12:26:54 <dcordes> That might apply to the MIT licensed userspace portions..
Sep 21 12:28:05 <dcordes> I want to see the ac100 kernel sources and I don't have the device
Sep 21 12:31:16 <rsalveti> morning
Sep 21 12:32:04 <jo-erlend> rsalveti, good morning! :)
Sep 21 12:32:37 <rsalveti> jo-erlend: hey!
Sep 21 12:32:44 <jo-erlend> rsalveti, I re-dd
Sep 21 12:32:46 <persia> dcordes, Just convince ogra to make a kernel source request, and share it with you :)
Sep 21 12:33:09 <jo-erlend> uh.. I re-dd'd my memorycard and booted it with uInitrd. It did change the partitions, but now I'm stuck. Where do I go from here?
Sep 21 12:34:25 <rsalveti> jo-erlend: cool, you just need to boot it once
Sep 21 12:34:52 <rsalveti> to get the resize, then mount it at your host, and use chroot to install my kernel package
Sep 21 12:35:04 <dcordes> persia: I don't want to have to convince anybody to anything when all I want is to read kernel source
Sep 21 12:35:13 <dcordes> that is GPL licensed
Sep 21 12:35:20 <persia> Why?
Sep 21 12:35:25 <jo-erlend> rsalveti, ah. I just replaced the uImage with the one you sent me. That didn't work.
Sep 21 12:35:27 <persia> There's no requirement for GPL code to be public.
Sep 21 12:35:51 <rsalveti> jo-erlend: it'll break the same way as before, because of missing modules
Sep 21 12:36:09 <persia> There exist licenses that require code to be public, but most folk don't like them, and they fail the desert-island-test of freedom.
Sep 21 12:36:39 <persia> (because if you're on a desert island with no connectivity, you'd be restricted from modifying the code, since you couldn't make the modifications public)
Sep 21 12:37:05 <ogra_cmpc> GPL only7 requires that you give the source to a user you sold the binary to
Sep 21 12:37:15 <dcordes> ok
Sep 21 12:37:27 <ogra_cmpc> no requirement to make it public at all
Sep 21 12:37:38 <ogra_cmpc> at least GPL v2
Sep 21 12:37:46 <jo-erlend> rsalveti, I still have two partitions, but now the second partition contains /boot and has some files in it. Is that the way it's supposed to be?
Sep 21 12:37:49 <ogra_cmpc> i think it changed a bit for v3
Sep 21 12:37:51 <persia> ogra, No, GPL requires you give the source to anyone who has your source or binaries, regardless of whether they use it or whether they paid for it.
Sep 21 12:37:58 <rsalveti> jo-erlend: yep, that's fine
Sep 21 12:38:04 <persia> and there's still no public requirement for GPLv3
Sep 21 12:38:13 <rsalveti> jo-erlend: because u-boot uses the files from the first partition, so it just ignores the /boot
Sep 21 12:38:29 <ogra_cmpc> how would one get the binaries without buying the device ?
Sep 21 12:38:34 <ogra_cmpc> thats a moot point
Sep 21 12:39:08 <rsalveti> jo-erlend: you should have vmlinuz and initrd.img from your kernel
Sep 21 12:39:12 <persia> ogra, Maybe one got a developer sample.  Maybe one downloaded a recovery image.  Maybe someone posted a copy of the kernel on the internet.  Lots of ways.
Sep 21 12:39:27 <dcordes> ogra: 'CONFIG_IKCONFIG_PROC=y' ?
Sep 21 12:39:40 <dcordes> ogra: you have /proc/config.gz file ?
Sep 21 12:39:45 <ogra_cmpc> sure
Sep 21 12:40:21 <dcordes> can you pastebin ?
Sep 21 12:41:20 <dcordes> persia: ogra mentioned they are planning an upgrade to new android version. as android makes many depencies on kernel level this will likely involve new kernel binary
Sep 21 12:41:46 <dcordes> persia: the question is if it will be some OTR thing or also password restricted for the device buyers
Sep 21 12:41:49 <persia> dcordes, So, when that happens, if it's an open download, download that binary.  Once you have it, request source.
Sep 21 12:42:24 <persia> Could be either way.  Doesn't really matter.  the kernel source can be made available as soon as someone posts it.  there's no restriction on posting except motivation of folks.
Sep 21 12:42:36 <persia> Hence the value of convincing folk to do stuff :)
Sep 21 12:42:41 <dcordes> Actually I don't like the fact such things must be discussed.. I am just recovering of the DELL Streak gpl violation thing
Sep 21 12:43:36 <dcordes> persia: so a 'leaked' binary can also be jurisdical base to enforce gpl ?
Sep 21 12:43:47 * persia refuses to provide a legal opinion
Sep 21 12:44:11 <persia> That said, it probably depends on how the leak happened, etc.  Quite possibly.
Sep 21 12:44:31 <persia> the GPL requires source to be made available to anyone to whom a binary is made available.
Sep 21 12:45:04 <persia> But you'd have to make a GPL claim against the leaker, which may have repurcussions, rather than against the original patch authors whose code was included in the leak.
Sep 21 12:45:49 <persia> So if the leaker was a friend, you might not want to do that.  If the leaker was a whistleblowing organisation, you don't help their cause by making claims against them.
Sep 21 12:46:35 <persia> GPL only gives you the right to request source code from the same folk that provided the binary.
Sep 21 12:46:53 <suihkulokki> yep
Sep 21 12:46:55 <dcordes> how does nokia handle it with n900 and friends ?
Sep 21 12:47:07 <suihkulokki> with a written offer
Sep 21 12:47:10 <dcordes> the whole source-for-device-owner-only thing is new to me
Sep 21 12:47:10 <persia> I believe they offer code for allthe GPL stuff.
Sep 21 12:47:21 <ogra_cmpc> with good lawyers
Sep 21 12:47:28 <rsalveti> they offer in the ftp servers
Sep 21 12:47:31 <rsalveti> easier way
Sep 21 12:47:32 <jo-erlend> rsalveti, I do. They point to vmlinuz-2.6.35-19-omap
Sep 21 12:47:39 <rsalveti> instead of waiting requests from everybody
Sep 21 12:47:47 <persia> dcordes, It's not source-for-device-owner-only!  It's that the source only needs to be made available to folks that receive the binaries.  Most large projects end up with a public repo, but that's just a matter of convenience.
Sep 21 12:47:53 <rsalveti> jo-erlend: sure, the default kernel
Sep 21 12:48:02 <rsalveti> jo-erlend: now just install mine with dpkg -i
Sep 21 12:48:05 <persia> dcordes, once any owner receives the source, they are free to post to the internet anywhere they like.
Sep 21 12:48:10 <rsalveti> you should get new vmlinuz and initrd.img
Sep 21 12:48:18 <ogra_cmpc> i think toshiba plans that too, there is a URL in the license agreement
Sep 21 12:48:21 <rsalveti> then you generate your own uImage and uInitrd from it
Sep 21 12:48:25 <ogra_cmpc> (which doesnt work yet)
Sep 21 12:48:49 <dcordes> persia: yeah that's what I mean. Matter of convenience. I haven't seen that toshiba behaviour before
Sep 21 12:49:36 * persia has often had such limits in contracts, because the client didn't want their code distributed, and I wanted to write GPL code.
Sep 21 12:49:55 <dcordes> ogra: can you pastebin the /proc/config.gz ?
Sep 21 12:50:05 <ogra_cmpc> dcordes, toshiba hasnt provided android devices before
Sep 21 12:50:49 <dcordes> persia: that's interesting. I once got a contract offering from people who wanted to make shanzai android phone.. it was same story
Sep 21 12:51:56 <persia> dcordes, Not uncommon: lots of companies want to time their releases.  Doesn't make the code less GPL when users get it.
Sep 21 12:52:37 <dcordes> in case of HTC you also see such delays but it's always in an acceptable frame
Sep 21 12:52:48 <suihkulokki> you need to sell early publicing of code as "free QA review from community"
Sep 21 12:53:04 <ogra_cmpc> the ac100 got on the market 10 dys ago or so
Sep 21 12:53:27 * ogra_cmpc wouldnt see that as inacceptable if it happens at some point soon
Sep 21 12:53:58 <jo-erlend> rsalveti, how do I do that?
Sep 21 12:54:03 <persia> suihkulokki, I think that doing so requires a bit more stature than being able to require GPL which requires more than random coding.
Sep 21 12:54:23 <rsalveti> jo-erlend: what exactly? generate uImage and uInitrd?
Sep 21 12:54:26 <jo-erlend> yes
Sep 21 12:54:34 <rsalveti> with mkimage
Sep 21 12:54:38 <dcordes> ogra_cmpc: if you understood it, I did not mean to compare to toshiba there
Sep 21 12:54:40 <suihkulokki> persia: well yes of course
Sep 21 12:55:05 <jo-erlend> rsalveti, little more details please? :=
Sep 21 12:55:05 <rsalveti> mkimage -A arm -O linux -T ramdisk -C none -a 0 -e 0 -n initramfs -d initrd.img uInitrd
Sep 21 12:55:22 <rsalveti> mkimage -A arm -O linux -T kernel -C none -a 0x80008000 -e 0x80008000 -n "Linux" -d vmlinuz uImage
Sep 21 12:55:28 <dcordes> Here that was ridiculous story http://laforge.gnumonks.org/weblog/2010/09/13/#20100913-dell_streak_sources
Sep 21 12:55:32 <dcordes> and it's not over yet.
Sep 21 12:55:38 <rsalveti> jo-erlend: install package uboot-mkimage
Sep 21 12:55:57 <jo-erlend> rsalveti, I have. I used it to create the boot.ini file.
Sep 21 12:56:04 <rsalveti> after installing my package, run these commands with the new initrd.img and vmlinuz
Sep 21 12:56:21 <rsalveti> then you'll have a new uInitrd and uImage
Sep 21 12:56:35 <jo-erlend> they'll be placed in rootfs/boot?
Sep 21 12:56:36 <rsalveti> copy them to the first partition, recreating it first
Sep 21 12:57:23 <rsalveti> jo-erlend: after installing the package you can find them at rootfs/boot, but after generating the uI* files you need to copy them to the first partition
Sep 21 12:57:38 <jo-erlend> right.
Sep 21 13:00:57 <jo-erlend> rsalveti, I got the same error today when installing the package, although on a different file. No space left on device.
Sep 21 13:01:27 <rsalveti> jo-erlend: this is weird, if you got the resize working, you should have enough space
Sep 21 13:01:41 <rsalveti> how much space left do you have at your sd card?
Sep 21 13:01:59 <ogra> GrueMaster, how was the oo.o testing ?
Sep 21 13:02:05 <ogra> (doko just asked)
Sep 21 13:02:09 <jo-erlend> rsalveti, well. Nautilus shows it as a 7.9GB volume, but from within the chroot, df -h displays it as 2.0GB.
Sep 21 13:02:31 <GrueMaster> I didn't get to it.  Swamped with alsa testing.
Sep 21 13:02:40 <ogra> GrueMaster, well
Sep 21 13:02:47 <ogra> the packages are in the archive now
Sep 21 13:02:51 <rsalveti> hm, so it was probably not resized correctly, werid
Sep 21 13:02:56 <GrueMaster> ok
Sep 21 13:03:14 <GrueMaster> I had also downloaded them to my server.  Just never got a chance.
Sep 21 13:03:22 <rsalveti> jo-erlend: did you boot it once with the original uImage and uInitrd, right?
Sep 21 13:03:27 <jo-erlend> yes
Sep 21 13:03:50 <rsalveti> jo-erlend: did you see if you got the reboot? or how long did you wait before turning the device off?
Sep 21 13:03:55 <jo-erlend> rsalveti, I might have made a mistake along the way. It was 6am when I tried the last time, so I was a bit sleepy. I think I'll start over.
Sep 21 13:04:02 <rsalveti> you should wait at least 5 minutes
Sep 21 13:04:06 <jo-erlend> oh, ok.
Sep 21 13:04:10 <jo-erlend> I don't think I waited that long.
Sep 21 13:04:14 <rsalveti> :-)
Sep 21 13:16:26 <jo-erlend> rsalveti, should I use the uImage you sent me for the first boot?
Sep 21 13:16:47 <rsalveti> jo-erlend: use the original one, with original uInitrd
Sep 21 13:16:53 <rsalveti> to be sure we don't face any issues
Sep 21 13:17:00 <rsalveti> just boot and wait for 5,8 minutes
Sep 21 13:17:02 <persia> The original initrd has some extra code to resize the filesystem, etc.
Sep 21 13:17:15 <rsalveti> persia: that's why we're doing this
Sep 21 13:54:07 <berco> does anyone have a beagleboard up and running who can quickly do "cat /proc/asound/cards" and report me the output?
Sep 21 13:54:31 <jo-erlend> rsalveti, this is very strange. I booted the board with the original uImage and uInitrd and left it running for 15-20 minutes. When I mounted it on my laptop, Nautilus reports it as a 7.9GB filesystem, but also reports that is has only 298,7MB of free space.
Sep 21 13:54:58 <rsalveti> jo-erlend: sigh
Sep 21 13:55:16 <rsalveti> jo-erlend: for some reason the resize is not working, I believe
Sep 21 13:55:26 <jo-erlend> can I do it manually?
Sep 21 13:55:50 <rsalveti> probably
Sep 21 13:55:59 <rsalveti> let me try to find the command
Sep 21 13:57:54 <persia> berco, http://paste.ubuntu.com/497723/
Sep 21 13:58:17 <berco> persia, thanks
Sep 21 13:58:43 <berco> persia, I think we are missing the equivalent of that twl4030 field on pandaboard
Sep 21 13:59:03 <berco> this is why my alsa/SDP4030.conf file doesn't get read
Sep 21 13:59:21 <rsalveti> jo-erlend: http://bazaar.launchpad.net/~ogra/jasper-initramfs/trunk/annotate/head:/scripts/local-premount/jasper_growroot#L82
Sep 21 13:59:28 <rsalveti> quite a few steps
Sep 21 14:00:07 <rsalveti> resize_partitions to change the disk size
Sep 21 14:00:30 <persia> berco, What does the pandaboard report?
Sep 21 14:00:46 <rsalveti> then e2fsck and resize2fs
Sep 21 14:01:48 <jo-erlend> rsalveti, the disk size seems to have changed.
Sep 21 14:02:15 <rsalveti> fdisk should show you
Sep 21 14:02:28 <rsalveti> but if you get it right at nautilus, it's probably changed
Sep 21 14:02:41 <berco> persia, pandaboard reports this http://paste.ubuntu.com/497729/
Sep 21 14:03:11 <persia> Yeah, that's missing the machine configuration entirely.
Sep 21 14:03:35 <berco> persia: when I use iwatch I can see the file is not accessed whereas when I do the test on my PC, the .conf is accessed
Sep 21 14:04:46 <jo-erlend> rsalveti, palimpsest shows /dev/sdb2 as a 7.9GB Ext3 filesystem+
Sep 21 14:06:03 <ogra> jo-erlend, what about df -h
Sep 21 14:06:09 <jo-erlend> it shows 2.0GB.
Sep 21 14:06:13 <ogra> right
Sep 21 14:06:17 <mpoirier> rsalveti: in the meeting you talked about "a bug for the kernel and another for u-boot", it this related to sound ?
Sep 21 14:06:24 <rsalveti> mpoirier: nops, LED
Sep 21 14:06:30 <mpoirier> ok.
Sep 21 14:06:30 <ogra> mpoirier, no, to new panda HW
Sep 21 14:06:36 <ogra> oh, that one
Sep 21 14:06:40 <ogra> ignore me :P
Sep 21 14:06:54 <mpoirier> ogra: I can do that !
Sep 21 14:06:58 <GrueMaster> mpoirier: bug 626795
Sep 21 14:06:59 <ubot2> Launchpad bug 626795 in linux-ti-omap4 (Ubuntu) "Missing LED support for Pandaboard (affects: 1) (heat: 8)" [Low,Fix released] https://launchpad.net/bugs/626795
Sep 21 14:07:11 <GrueMaster> Just for clarity.
Sep 21 14:07:20 <persia> mpoirier, Do you have any recommendations for berco?  He's working on a userspace workaround in case you don't find a kernelspace fix, but the kernel isn't reporting enough to dynamically load the right configuration :(
Sep 21 14:07:21 <jo-erlend> fdisk sais it's a 8GB disk...
Sep 21 14:07:45 <ogra> jo-erlend, fdisk and the gnome tool report partition sizes
Sep 21 14:07:52 <ogra> jo-erlend, df shows filesystem size
Sep 21 14:07:59 <rsalveti> so no resize
Sep 21 14:08:05 <mpoirier> persia: berco: I'm just getting started - give me a few days and I'll have an opinion.
Sep 21 14:08:12 <ogra> well, resize_partitions has run
Sep 21 14:08:19 <ogra> but not resize2fs
Sep 21 14:08:24 <rsalveti> yup
Sep 21 14:08:33 <jo-erlend> ah. Then palimpsest was confusing, since it sais "7.9GB ext3"
Sep 21 14:08:44 <rsalveti> weird, but without console and display working, no much to say, unless checking jasper logs
Sep 21 14:16:20 <jo-erlend> rsalveti, I won't have to manually create the /lib/modules directory like I did yesterday?
Sep 21 14:16:44 <rsalveti> jo-erlend: nops, because now you installed by kernel package
Sep 21 14:16:55 <rsalveti> *my
Sep 21 14:17:14 <rsalveti> or, you will install it when you're able to resize it
Sep 21 14:21:31 <jo-erlend> yes, it's being installed.
Sep 21 14:22:12 <ogra> GrueMaster, hmm, do you see update-manager popping up automatically on the images ?
Sep 21 14:22:30 * ogra juts ran an apt-get update, that should trigger u-m i think 
Sep 21 14:22:39 <GrueMaster> I haven't looked.  Had been focusing on a lot of other issues lately.
Sep 21 14:23:31 <GrueMaster> will look on omap4 now.
Sep 21 14:23:37 <rsalveti> jo-erlend: so, did you resize it by hand?
Sep 21 14:25:57 <GrueMaster> ogra: apt-get update usually will trigger it within a few minutes.  Will let you know if I see any activity.
Sep 21 14:27:16 <GrueMaster> I know it works on Lucid.  It shows up at least weekly on my babbage.
Sep 21 14:27:28 <ogra> k
Sep 21 14:27:55 <ogra> no hurry, was just a question out of interest
Sep 21 14:28:43 * ogra takes a break
Sep 21 14:35:43 <GrueMaster> ogra: The "Install OMAP4 addons" icon is in "Other" on 20100920, but the cmdline fails.
Sep 21 14:38:51 <jo-erlend> rsalveti, does it matter if I create uInitrd and uImage from the chroot or not?
Sep 21 14:39:07 <rsalveti> jo-erlend: nops
Sep 21 14:39:08 <jo-erlend> rsalveti, I resized by hand, yes.
Sep 21 14:39:11 <rsalveti> cool
Sep 21 14:39:26 <rsalveti> so you should now have enough space to play with it later :-)
Sep 21 14:39:40 <jo-erlend> yes, and the kernel installed properly. :)
Sep 21 14:39:51 <rsalveti> nice :-)
Sep 21 14:41:10 <ogra_ac> GrueMaster, yeah, its partially fixed in 21.1
Sep 21 14:41:20 <GrueMaster> ok.
Sep 21 14:41:24 <ogra_ac> but needs a sw-center fix first
Sep 21 14:41:38 <ogra_ac> which is in the tree but wasnt uploaded
Sep 21 14:41:51 <jo-erlend> rsalveti, I should use initrd.img-2.6.35.4+?
Sep 21 14:42:04 * ogra_ac wishes he could find an IRC client for android with a lag meter
Sep 21 14:42:11 <GrueMaster> Just so you know, I really have a hard time testing dailys and apps and special requests on a daily basis.
Sep 21 14:42:22 <GrueMaster> I'm getting overwhelmed.
Sep 21 14:42:51 <ogra_ac> why is that ? because your boards are busy doing other stuff ?
Sep 21 14:43:10 <GrueMaster> Boards are slow, and some of the tests can take hours.
Sep 21 14:43:31 <ogra_ac> well, omap4 should be rather fine for quick tests
Sep 21 14:43:42 <ogra_ac> i agee for beagle and dove
Sep 21 14:43:55 <rsalveti> jo-erlend: yep
Sep 21 14:44:12 <GrueMaster> Even just running checkbox with only known supported tests can take a day.
Sep 21 14:44:13 <rsalveti> sorry, seems my notification system is not working properly
Sep 21 14:44:31 <persia> ogra, An 8-way 5Ghz 32Gb system would be slow for some tests.
Sep 21 14:45:09 <ogra_ac> sigh
Sep 21 14:45:23 <GrueMaster> This is part of why I haven't jumped on open office yet.  Install alone will take 20 minutes.
Sep 21 14:45:26 <rsalveti> ogra_ac: what happens with your client?
Sep 21 14:45:38 * ogra_ac pokes android in the eye
Sep 21 14:45:52 <ogra_ac> client is fine, the wifi is unstable after some time
Sep 21 14:46:00 <ogra_ac> ralink NIC
Sep 21 14:46:09 <ogra_ac> with binary driver it seems
Sep 21 14:46:42 <rsalveti> hm, ok
Sep 21 14:47:03 <ogra_ac> i have to connect to teh Ap with teh worst signal to make it work at all
Sep 21 14:47:26 <ogra_ac> might be my local setup or some channel interference between the tree APs i have
Sep 21 14:47:45 <ogra_ac> seems the card is more fragile than my others
Sep 21 14:48:23 <ogra_ac> the annoying part is that i cant find a client that shows me the lag so i dont even notice i'm disconnected
Sep 21 14:51:46 <ndec> ogra: hi! i have a question for you...
Sep 21 14:51:56 <ogra_ac> ndec, shoot
Sep 21 14:52:16 <ndec> ogra: i want to mount the ext3 partition from the .img in qemu on my laptop... what is the magic command line?
Sep 21 14:52:25 <ndec> ogra: qemu arm i mean
Sep 21 14:52:57 <ogra_ac> i guuess you need to loop mount it, then use /dev/loop* for aemu
Sep 21 14:53:02 <ogra_ac> *qemu
Sep 21 14:53:23 <jo-erlend> rsalveti, I have graphics! Welcome to Ubuntu :)
Sep 21 14:53:31 <ogra_ac> for the -hda option
Sep 21 14:53:38 <GrueMaster> ndec: First, run "file <img file>" to find out the partition info.  Then run "sudo mount <img file> <mount point> -o loop,offset=$((512*<partiotion start sector>))"
Sep 21 14:55:46 <ndec> ogra: GrueMaster: thanks. I will try this. I didn't realize I could a mount point to -hda, i was looking to a way to pass the offset to qemu.
Sep 21 14:55:47 <GrueMaster> ogra_ac: Looks like update-notifier is not working at all.  I also tried it on my netbook running maverick.
Sep 21 14:55:58 <ogra_ac> bah
Sep 21 14:56:16 <ogra_ac> update-manager works though, i just did an update
Sep 21 14:56:31 <ogra_ac> anything in .xsession-errors ?
Sep 21 14:58:32 <ogra_ac> probably the behavior changed and we just dont know
Sep 21 14:58:49 <rsalveti> jo-erlend: nice :-)
Sep 21 14:58:53 <ogra_ac> (the default i mean)
Sep 21 15:01:53 <GrueMaster> ogra_ac: I'm not seeing anything since it's startup note.
Sep 21 15:02:08 <ogra_ac> k
Sep 21 15:02:39 * ogra_ac will ask mvo in #ubuntu-devel if there are deliberate changes
Sep 21 15:04:20 <rsalveti> jo-erlend: so it did fixed, can you please then update the bug report?
Sep 21 15:07:46 <ogra_ac> GrueMaster, update-notifier --debug-update seems to be the magic to find out whats wrong
Sep 21 15:08:53 <GrueMaster> Looks like the update interval is set ot 7 days in gconf.
Sep 21 15:09:03 <ogra_ac> yes, thats normal
Sep 21 15:09:12 <ogra_ac> was like that in lucid too
Sep 21 15:09:21 <ogra_ac> (and karmic)
Sep 21 15:09:26 <GrueMaster> Ok, so...
Sep 21 15:10:12 <ogra_ac> its just the cron timer, but u-n should run in the session too and pick up changes if you run apt-get update
Sep 21 15:10:48 <GrueMaster> It also says "/usr/lib/update-notifier/apt-check returned 38 (security: 0)"
Sep 21 15:11:08 <ogra_ac> that sounds right, weird that it doesnt pop up
Sep 21 15:11:30 <ogra_ac> 38 normal packages, 0 security updates
Sep 21 15:11:54 <GrueMaster> Yes, that sounds accurate since I ran apt-get update before restarting update-notifier.
Sep 21 15:12:38 <ogra_ac> i would think we miss something on arm in teh efl session, but if you see it too on x86 ...
Sep 21 15:15:04 <GrueMaster> Just restarted on x86-64 desktop (my netbook).  apt-check returns 12 updates (0 security).  No notification there either.
Sep 21 15:16:11 * GrueMaster grumbles and files yet another bug that should have been picked up elsewhere.
Sep 21 15:17:16 <ogra_ac> gruemaster ... why elsewhere ?
Sep 21 15:17:53 <GrueMaster> If it exists on x86, it should have been tested & reported there.
Sep 21 15:18:32 <ogra_ac> well, its a bug ... we are just the lucky ones having teh idea to test it :)
Sep 21 15:23:20 <jo-erlend> rsalveti, I will. However, I see no signs of networking, either cabled or wireless.
Sep 21 15:24:58 <rsalveti> jo-erlend: hm, then need first to find the chip responsible for networking
Sep 21 15:27:05 <GrueMaster> ogra_ac: Bug 644459 filed.
Sep 21 15:27:07 <ubot2> Launchpad bug 644459 in update-notifier (Ubuntu) "update-notifier not reporting updates available (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/644459
Sep 21 15:27:59 <jo-erlend> rsalveti, and I get a very low resolution, and I can't change it in System > Preferences > Monitors. All the controls are deactivated.
Sep 21 15:28:11 <ogra_ac> GrueMaster, thanks
Sep 21 15:28:36 <rsalveti> jo-erlend: you can't, you need to specify the correct one by giving it to the omapfb parameter
Sep 21 15:29:08 <jo-erlend> oh.
Sep 21 15:29:52 <rsalveti> argh, seems I'm getting random temperatures report from my acpi
Sep 21 15:30:00 <rsalveti> and sometimes my kernel decides to halt to protect me
Sep 21 15:30:17 <rsalveti> but it's not hot, and I'm tracing the temperature, weird
Sep 21 15:30:27 <rsalveti> must be a bios issue...
Sep 21 15:31:29 <jo-erlend> rsalveti, which omapfb mode should I use for 1920x1080?
Sep 21 15:31:44 <rsalveti> jo-erlend: nops, omap 3 doesn't support it
Sep 21 15:31:53 <jo-erlend> hmm?
Sep 21 15:33:06 <jo-erlend> it's supposed to be able to use much higher resolutions than that
Sep 21 15:34:14 <ogra_ac> at 1bit depth ?
Sep 21 15:34:34 <rsalveti> 1280x1024@50 or 1280x720@60 I guess
Sep 21 15:34:40 <ogra_ac> yeah
Sep 21 15:34:44 <jo-erlend> they say it supports up to 2048x2048.
Sep 21 15:34:49 <jcrigby> ogra_ac, who do I ping about build machine weirdness
Sep 21 15:34:56 <rsalveti> you could get more than that, but need to adjust the timings and etc
Sep 21 15:35:04 <ogra_ac> jcrigby, lamont
Sep 21 15:35:14 <rsalveti> jo-erlend: hm, if igepv2 is the same chip we have for beagle, probably not
Sep 21 15:35:16 <ogra_ac> jcrigby, whats wrong ?
Sep 21 15:35:24 <rsalveti> the pixel clock can't get that higher
Sep 21 15:36:06 <rsalveti> jo-erlend: using 16bit
Sep 21 15:36:15 <jcrigby> ogra_ac, we are waiting for a kernel build.  It starts up building, runs for a couple minutes then goes idle.  Seems to keep bouncing between hubbard and hawthorn
Sep 21 15:36:27 <ogra_ac> weird
Sep 21 15:36:29 <jcrigby> I pinged lamont on #is but he is away
Sep 21 15:36:57 <ogra_ac> well, he is the master of arm buildds and knows the queue software
Sep 21 15:37:09 <ogra_ac> not sure whom to ping else on such an issue
Sep 21 15:37:23 <jcrigby> ok, thanks
Sep 21 15:40:09 <jo-erlend> rsalveti, ok. What do I use to get 1280x720?
Sep 21 15:40:23 <rsalveti> jo-erlend: vram=12M omapfb.mode=dvi:1280x720MR-16@60
Sep 21 15:40:50 <jo-erlend> thanks.
Sep 21 15:42:50 <lag> ogra_ac: Stop using my name in vain - jerk
Sep 21 15:42:51 <lag> ;)
Sep 21 15:45:40 * ogra_ac thinks lag meter is a valid device though
Sep 21 15:45:51 <jo-erlend> rsalveti, hmm. The bootpartition is suddenly empty.
Sep 21 15:46:03 * lag is |-..........................|
Sep 21 15:46:06 <rsalveti> jo-erlend: ouch, shouldn't be
Sep 21 15:46:06 <jo-erlend> the boot fs, I mean.
Sep 21 15:46:09 * lag is |-----......................|
Sep 21 15:46:14 * lag is |---------..................|
Sep 21 15:46:24 * lag is |-------------..............|
Sep 21 15:46:30 * lag is |-----------------..........|
Sep 21 15:46:33 <ogra_ac> heh
Sep 21 15:46:35 * lag is |---------------------......|
Sep 21 15:46:37 <lag> :D
Sep 21 15:46:44 <ogra_ac> how is asia ?
Sep 21 15:46:55 <lag> Hot and sticky
Sep 21 15:47:12 <ogra_ac> heh, nothing new then
Sep 21 15:47:24 <lag> Have you been here?
Sep 21 15:47:38 <ogra_ac> only once on my flight to the second UDS
Sep 21 15:47:49 <ogra_ac> i had a smoke break in singapore
Sep 21 15:49:46 <ogra_ac> lag, btw, ac 100 rooted :)
Sep 21 15:49:52 <lag> Where was the second UDS?
Sep 21 15:49:54 <ogra_ac> and i bricked one already
Sep 21 15:50:00 <lag> Well done
Sep 21 15:50:04 <lag> And not so well done
Sep 21 15:50:05 <ogra_ac> that was in sydney
Sep 21 15:50:08 <lag> How did you do it?
Sep 21 15:50:12 <lag> Cold card?
Sep 21 15:50:19 <ogra_ac> well, got an immediate replacement at the shop
Sep 21 15:50:29 <lag> Wow - when's the next Aus UDS?
Sep 21 15:50:31 <ogra_ac> no, more embarrasing
Sep 21 15:50:35 <lag> Go on?
Sep 21 15:50:47 <ogra_ac> i accidnetially removed /bin/sh
Sep 21 15:50:59 <lag> HA
Sep 21 15:51:08 <lag> Isn't it just a simlink?
Sep 21 15:51:10 <ogra_ac> androids init calls shell scripts
Sep 21 15:51:15 <ogra_ac> yes
Sep 21 15:51:19 <lag> So?
Sep 21 15:51:25 <ogra_ac> i wanted to link it to busybox
Sep 21 15:51:31 <lag> So do it
Sep 21 15:51:32 <ogra_ac> but messed it up
Sep 21 15:51:48 <ogra_ac> it didnt boot anymore
Sep 21 15:51:53 <ogra_ac> no way to access it
Sep 21 15:52:02 <lag> Oh, I see
Sep 21 15:52:07 <lag> That was the first one?
Sep 21 15:52:11 <lag> What about the new one?
Sep 21 15:52:17 <ogra_ac> so i carried it back to the shop "doesnt boot !!"
Sep 21 15:52:25 <ogra_ac> they just handed me a new one
Sep 21 15:52:29 <ogra_ac> without words
Sep 21 15:52:45 <ogra_ac> the new one is rooted and i run an ubuntu chroot on SD now
Sep 21 15:53:07 <ogra_ac> dont want to brick it again so i wait until someone else figures out teh bootloader
Sep 21 15:53:23 <lag> So, how did you root it?
Sep 21 15:53:29 <GrueMaster> Odd that they make it impossible to recover.
Sep 21 15:53:30 <ogra_ac> the chroot is fine for compiling etc
Sep 21 15:54:05 <ogra_ac> there is a tool called rageagainstthecage that uses a kernel exploit by opening endless amounts of file descriptors
Sep 21 15:54:11 <ogra_ac> works quite well
Sep 21 15:54:27 <ogra_ac> you run it, the next shell you open is a rootshell
Sep 21 15:54:48 <ogra_ac> then you can make busybox suid root and use the busybox su command in the future
Sep 21 15:55:26 <ogra_ac> GrueMaster, it apparently has a recovery mode i discovered today
Sep 21 15:55:35 <ogra_ac> but its kind of a blind flight
Sep 21 15:55:44 <GrueMaster> heh
Sep 21 15:55:48 <ogra_ac> and requires a specisl nvflash to access the device
Sep 21 15:56:04 <GrueMaster> What did you expect, a bios prompt & boot menu?
Sep 21 15:56:05 <ogra_ac> which isnt in teh tegr4linux tools
Sep 21 15:56:29 <ogra_ac> i'd lobe to have USB serial, yeah
Sep 21 15:56:33 <ogra_ac> *love
Sep 21 15:56:34 <GrueMaster> What bootloader are they running?
Sep 21 15:56:40 <ogra_ac> fastboot i think
Sep 21 15:56:51 <ogra_ac> its android after all
Sep 21 15:57:25 <ogra_ac> the normal recovery mode can only reset to factory or install update.zip
Sep 21 15:57:25 <ogra_ac> nothing more
Sep 21 15:57:26 <GrueMaster> btw:  linaro call in 3 minutes.
Sep 21 15:57:34 <ogra_ac> and update.zip requires a toshiba kesy
Sep 21 15:57:45 <ogra_ac> oh, thanks i forgot about it
Sep 21 15:57:47 <ogra_ac> as usual
Sep 21 15:57:56 * ogra_ac goes upstairs
Sep 21 16:00:25 <ogra> bug 642117
Sep 21 16:00:26 <ubot2> Launchpad bug 642117 in pencil (Ubuntu) (and 4 other projects) "[armel] QFloat / double / float confusion (affects: 1) (heat: 8)" [Medium,Confirmed] https://launchpad.net/bugs/642117
Sep 21 16:00:46 <devilhorns> now this is much better: http://www.pastebin.ca/1945897
Sep 21 16:01:06 <devilhorns> tho have to bring the Res down still
Sep 21 16:03:37 <ogra> devilhorns, WOW!
Sep 21 16:03:46 <devilhorns> ogra, hmm ?
Sep 21 16:03:56 <ogra> awesome numbers
Sep 21 16:04:00 <devilhorns> :)
Sep 21 16:04:04 <devilhorns> getting there
Sep 21 16:04:17 <devilhorns> still a bit more work todo ... want to trim that Res down some more
Sep 21 16:04:41 <devilhorns> ideally I'd like to get to <= 10m
Sep 21 16:05:21 <ogra> well, we still run half of gnome ... even these values are awesome
Sep 21 16:05:49 <devilhorns> but already trimmed roughly 10,000 from the Virt, and 1m from the shared
Sep 21 16:06:36 <devilhorns> ogra, half of gnome ? why ? ... wrt this new one, the only thing that will need to be running is a "window manager" to draw borders, handle minimize, etc, etc
Sep 21 16:07:18 <ogra> devilhorns, we need to run the indicator applets and all the other ubuntu stuff
Sep 21 16:07:32 <ogra> everything that hooks into the system needs to work
Sep 21 16:07:47 <ogra> and all the old things are turned into the new indicator things
Sep 21 16:10:01 <ogra> (volume control, network management, time, app indicators etc)
Sep 21 16:11:28 <devilhorns> ahh yea, will still have to run the gnome-panel
Sep 21 16:11:34 <devilhorns> not much I can do about that tho :/
Sep 21 16:28:23 <ndec> ogra: not arm related problem... but I am trying to upgrade my laptop to maverick. i am running a custom lucid (i am running maverick lts backport kernel + a maverick thunderdird), when I try to upgrade the tool refuses to continue with 'E:Error, pkgProblemResolver::Resolve generated breaks, this may be caused by held packages.'. is it safe to replace lucid by maverick in my source.list and run dist-upgrade? or is the upgrade tool doing more than 
Sep 21 16:28:58 <ogra> its doing more
Sep 21 16:29:20 <ogra> you should talk to mvo in #ubuntu-devel, he maintains the upgrade tools
Sep 21 16:29:58 <ogra> he will surely be intrested in your situation since he loves to cover all corner cases for upgrading
Sep 21 16:51:25 <Neko> guys any clue where the Applications Places System menu is defined in gnome or xfce or whatever config?
Sep 21 16:51:40 <Neko> we have a button size problem here and I'd like to go about hacking in a fix but it seems to be automagically created
Sep 21 16:53:32 <GrueMaster> Neko: What are you referring to?
Sep 21 16:53:52 <GrueMaster> Looks like an Ubuntu Desktop issue.
Sep 21 16:54:15 <GrueMaster> (not armel).
Sep 21 16:54:58 <ogra> yeah, likely a question for #ubuntu-desktop
Sep 21 16:55:00 <Neko> http://launchpadlibrarian.net/55758568/screenshot.png
Sep 21 16:55:02 <Neko> like that
Sep 21 16:55:14 <Neko> it happens on GNOME too just so you know, but some people don't see it
Sep 21 16:56:26 <GrueMaster> Interesting.  I'm not seeing it on my netbook, and it is running Maverick amd64.
Sep 21 16:57:11 <GrueMaster> But you might want to bring it up in #ubuntu-desktop.  They could probably help more on this.
Sep 21 16:57:20 <ogra> right
Sep 21 16:57:21 <Neko> I just did and they said they are in a meeting
Sep 21 16:58:13 <ogra> yeah, they have a silly policy to hold their meetings in their channel instead of using #ubuntu-meeting like all other teams
Sep 21 16:58:31 * ogra got trapped by that several times already
Sep 21 17:06:38 <ogra_ac> oh, sweet, seems someone found acces to the ac100 bootloader
Sep 21 17:12:16 <davidm> ogra, sweet
Sep 21 17:12:52 <ogra_ac> no way to replace kernel etc yet, but they got access ... so linux is just a step away
Sep 21 17:13:22 * ogra_ac got careful with his, i dont want to have to replace it again so i'll wait for someone else to figure it out
Sep 21 17:13:27 <davidm> Yep very nice
Sep 21 17:13:53 <prpplague> ogra_ac: ac100?
Sep 21 17:14:15 <ogra_ac> prpplague, yeah, toshiba ac100 ... tegra2 based
Sep 21 17:14:58 <prpplague> ahh
Sep 21 17:15:02 <ogra_ac> sadly it comes with locked down android preinstalled
Sep 21 17:15:34 <ogra_ac> but i rooted mine and use an ubuntu chroot for testbuilding packages for omap already ;)
Sep 21 17:16:21 <ogra_ac> panda is better though ... but less portable
Sep 21 17:17:17 <prpplague> indeed
Sep 21 17:17:37 <prpplague> ogra_ac: hopefully we can remedy the portable portion soon
Sep 21 17:18:18 <ogra_ac> yeah, a nice light netbook based on panda would be awesome
Sep 21 17:20:05 <ogra_ac> (and indeed with ubuntu preinstalled)
Sep 21 17:20:10 <ogra_ac> ;)
Sep 21 17:25:52 <ogra_ac> some stable wlan would be cool though
Sep 21 17:26:08 * ogra_ac sighs about teh ralink adapter
Sep 21 17:33:47 <fredim> Hi, how can I copy the image ubuntu-4.10-server-armel omap.img for SDcard through the windows?
Sep 21 17:35:43 <GrueMaster> ubuntu-4.10?
Sep 21 17:38:37 <fredim> I can not mount the SDcard from Linux, I need to know the exact modules.
Sep 21 17:38:39 <fredim> or
Sep 21 17:38:42 <fredim> or
Sep 21 17:38:44 <fredim> learn how to copy the image through the windows
Sep 21 17:38:57 <GrueMaster> prpplague: I hadn't had a chance to test it before now, but panda ES2 works on my hdmi switch now.  Way cool.  I might even be able to reduce monitor count back to 5 or 6.
Sep 21 17:39:32 <prpplague> hehe
Sep 21 17:39:41 <prpplague> GrueMaster: yea robclark and i spent some time testing
Sep 21 17:39:56 <Neko> okay here's a question a bit generic but relevant to arm all the same
Sep 21 17:40:06 <Neko> how do I magically invoke the thing that makes an initrd for me
Sep 21 17:40:17 <Neko> seems if I installed a kernel package it'd do it for me but I don't have one yet
Sep 21 17:40:19 <Neko> and I need it
Sep 21 17:40:38 <GrueMaster> update-initramfs I believe.
Sep 21 17:41:49 <GrueMaster> prpplague: Thanks a lot for your efforts.  It really does help me a lot.
Sep 21 17:41:57 <prpplague> :)
Sep 21 17:42:53 <GrueMaster> Uh, oh.  May have spoken too soon.  It just came up in X for oem-config.  Looks aweful on my 1440x900 monitor.
Sep 21 17:43:13 <GrueMaster> Will reboot.  switch had lost focus for a bit.
Sep 21 17:43:14 <prpplague> doh
Sep 21 17:43:58 <GrueMaster> splash looks ok.
Sep 21 17:44:12 <Neko> GrueMaster, with no arguments or so?
Sep 21 17:44:25 <ogra_ac> just use teh right glasses :P
Sep 21 17:44:31 <GrueMaster> prpplague: Ok, non-issue.
Sep 21 17:44:43 <GrueMaster> ogra_ac: I only have one pair.
Sep 21 17:45:17 <ogra_ac> ah, thats the prob then :)
Sep 21 17:45:43 <GrueMaster> Neko: There are only a few parameters.  use update-initramfs -h to list.
Sep 21 17:46:08 <ogra_ac> for initial creation you want -c and -k
Sep 21 17:48:08 <GrueMaster> Now I can effectively test both es2 boards in parallel.  (as if I didn't have enough to do).  :P
Sep 21 17:48:23 <Neko> and in theory it just copies a bunch of stuff from the modules directory into a ramfs plus some userspace stuff?
Sep 21 17:48:38 <Neko> so I can create a ramfs for another system as long as I have the modules around (and it's set to "most" or "all" or something?)
Sep 21 17:49:53 <ogra_ac> hrm
Sep 21 17:52:56 <ogra_ac> Neko, no, initramfs-tools provides a lot of files in /usr/share/initramfs-tools
Sep 21 17:53:23 <ogra_ac> and every package can put scripts or hooks in there
Sep 21 17:53:36 <ogra_ac> they all end up in the initramfs
Sep 21 17:54:45 <dcordes> Hi again
Sep 21 17:55:24 <dcordes> ogra_ac: looked at some more ac100 stuff and the device is really nice. What is the canonical attitude towards adding images for additional machines ?
Sep 21 17:55:38 <ogra_ac> there is none
Sep 21 17:56:08 <dcordes> ogra_ac: So if you would be up to adding ac100 images you could just put them ?
Sep 21 17:56:25 <ogra_ac> i wouldnt
Sep 21 17:56:29 <Neko> .. I just want to know if I can generate an initramfs without booting that particular system
Sep 21 17:56:42 <Neko> I mean all the stuff for an exactly the same userspace should basically be common right?
Sep 21 17:56:45 <Neko> except the modules?
Sep 21 17:56:51 <Neko> and I specify which one it trawls with -k
Sep 21 17:56:58 <ogra_ac> dcordes, such situations were the reason why i wrote rootstock
Sep 21 17:57:27 <ogra_ac> Neko, just chroot into your SD card and run update-initramfs ether
Sep 21 17:57:32 <ogra_ac> *there
Sep 21 17:58:40 <ogra_ac> then you can be sure all scripts and hokks needed by installed packages are in your initrd
Sep 21 17:59:12 <Neko> yes but imagine I am in rootstock and I have a bunch of kernel files there extracted somehow, and I run update-initramfs is it going to work
Sep 21 17:59:24 <Neko> considering this userspace will work on 20 different systems if I only put the right modules in place :D
Sep 21 17:59:39 <Neko> armv7 is armv7 at the end of the day and there are no board-specific stuff required
Sep 21 17:59:42 <Neko> except the modules
Sep 21 17:59:55 <ogra_ac> i think rcn-ee once added an option for rolling initramfses inside rootstock
Sep 21 18:00:21 <GrueMaster> Neko: Do you have an armv7 that you can use to modify SD cards on the fly?  That would be easiest.
Sep 21 18:00:29 <ogra_ac> there might be no HW specific suff, but system specific stuff
Sep 21 18:00:48 <Neko> yeah I do
Sep 21 18:00:51 <ogra_ac> GrueMaster chroot under x86 would be even easier
Sep 21 18:00:53 <Neko> I can boot the fs on my efikamx
Sep 21 18:01:09 <ogra_ac> thats why we have qemu-arm-static
Sep 21 18:01:09 <Neko> but I want to generate the one for the netbook without having to boot it on the netbook too
Sep 21 18:01:21 <Neko> then I can work out how to do it in a chroot on my x86_64 box
Sep 21 18:01:53 <GrueMaster> ogra_ac: I was thinking of something more stable.  Like a babbage or something.
Sep 21 18:02:03 <ogra_ac> apt-get install qemu-arm-static .... mount SD ... copy the qemu-arm-static binary into SD
Sep 21 18:02:11 <ogra_ac> chroot /sd-mountpoint
Sep 21 18:02:19 <ogra_ac> thats it
Sep 21 18:03:08 <GrueMaster> ogra_ac: Thats assuming you are running maverick, which isn't fully stable yet.  Otherwise you get an older version of qemu.
Sep 21 18:03:36 * ogra_cmpc gibes up on 6teh ac100 wlan
Sep 21 18:03:45 <ogra_cmpc> GrueMaster, no, we have it since karmic
Sep 21 18:03:59 <ogra_ac> works fine without any probs
Sep 21 18:03:59 <ogra_ac> and a lot faster than natively
Sep 21 18:04:00 <ogra_cmpc> and the static chroot part was always stable
Sep 21 18:04:19 <ogra_cmpc> apart from running mono
Sep 21 18:04:33 <GrueMaster> mono is another beast entirely.
Sep 21 18:04:34 <ogra_cmpc> which is why we dont use it for image builds
Sep 21 18:04:54 <ogra_cmpc> update-initramfs is just a script
Sep 21 18:05:14 <ogra_cmpc> using the chroot method your scripts run at host speed
Sep 21 18:05:38 <Neko> no offense but the desktop on ubuntu is a horrible, horrible mess
Sep 21 18:05:44 <ogra_cmpc> compiling runs at qemu speed though... thats why we dont use the chroot method for compiling either
Sep 21 18:06:07 <ogra_cmpc> Neko, ? why
Sep 21 18:06:46 <Neko> I have deep disdain for any system that has to dereference 20 different things to finally get down to a default, and then find that it is called "/usr/share/backgrounds/warty-final-ubuntu.png"
Sep 21 18:07:04 <ogra_ac> lol
Sep 21 18:07:13 <ogra_ac> thats historically based
Sep 21 18:07:27 <Neko> why didn't they just make it called default-backdrop.png
Sep 21 18:07:36 <ogra_ac> file a bug
Sep 21 18:07:44 <ogra_ac> i think nobody ever cared
Sep 21 18:08:03 <Neko> oh god I can't be bothered, nobody fixes this stuff. By the time it gets looked at, it'll be WONTFIX and someone will have posted "does this still happen in Narwhal?"
Sep 21 18:08:41 <ogra_ac> well, you are free to use whatever value for teh gconf key anyway in a custom image
Sep 21 18:09:09 <Neko> what I can't find is what is the canonical (pun intended) place to set that gconf key and how on earth I do it without dicking around as the "gdm" user in a logged in system
Sep 21 18:09:11 <ogra_ac> so just call it ... umm ... default-backdrop.png for example :)
Sep 21 18:09:47 <ogra_ac> you do it as all system wide gconf keys under /usr/share/gconf/default
Sep 21 18:10:10 <Neko> the basic idea is that I can change something in a chroot or just in /mnt/usr/share/gconf or something and the system will be fine when it boots
Sep 21 18:10:24 <ogra_ac> you can
Sep 21 18:10:32 <Neko> and that script is how I do it?
Sep 21 18:10:33 <ogra_ac> thats how we customize images
Sep 21 18:10:40 <Neko> but that is only run when update-gconf-defaults is run
Sep 21 18:10:48 <Neko> I can't run that I can "only" edit text files
Sep 21 18:11:03 <ogra_ac> no, it needs to generate the xml
Sep 21 18:11:19 <ogra_ac> you could probably do that too by hand indeed
Sep 21 18:11:22 <Neko> so if I edit that, where do I edit the xml?
Sep 21 18:11:31 <Neko> to match and so that it works first time and then forever after? :D
Sep 21 18:11:43 <dcordes> is it possible to enable on screen keyboard in oem-config via configuration ? I am looking for a workaround to bug 626055 as it seems it is a 'wontfix'
Sep 21 18:11:44 <ubot2> Launchpad bug 626055 in ubiquity (Ubuntu) "oem-config: make on-screen keyboard available (affects: 1) (heat: 159)" [Low,New] https://launchpad.net/bugs/626055
Sep 21 18:11:45 <ogra_ac> read rthe gconf source ... no idea
Sep 21 18:11:55 <Neko> ughhhhh
Sep 21 18:12:11 <ogra_ac> or do it properly
Sep 21 18:12:16 <ogra_ac> and roll a package
Sep 21 18:12:19 <Neko> I give up for now we'll just ship with the horrible placeholder backdrop :D
Sep 21 18:12:22 <ogra_ac> that carries your defaults
Sep 21 18:14:07 <dcordes> GrueMaster: ping.. can you help with the above ?
Sep 21 18:15:23 <GrueMaster> dcordes: Not really at this time.  I would have to learn how oem-config works, and I am really only a QA guy.
Sep 21 18:17:50 <dcordes> Ok thanks anyway
Sep 21 18:20:36 <Neko> ogra, rsalveti
Sep 21 18:20:41 <Neko> what is this line for in rootstock?
Sep 21 18:20:41 <Neko> rm /usr/lib/ubiquity/plugins/ubi-tasks.py*
Sep 21 18:21:17 <rsalveti> Neko: ogra_ac should know it
Sep 21 18:24:43 <Neko> potentially it's the bit that sorta goes "which desktop do you want, dude?"
Sep 21 18:24:50 <Neko> which is kinda pointless then
Sep 21 18:24:57 <Neko> but then again so is partition management and suchlike
Sep 21 18:25:15 <ogra_ac> it disables tasksel in oem-config
Sep 21 18:25:34 <Neko> because of?
Sep 21 18:25:37 <Neko> a bug?
Sep 21 18:25:42 <ogra_ac> no
Sep 21 18:26:08 <ogra_ac> because you dont want to run tasksel if you already provided your tasks at buildtime
Sep 21 18:26:13 <Neko> oem-config-prepare does not such thing so I would assume that it lets it stay in by default everywhere except a rootstock
Sep 21 18:26:30 <ogra_ac> other images use preseed files
Sep 21 18:26:35 <ogra_ac> so tasksel isnt run
Sep 21 18:26:57 <Neko> ok
Sep 21 18:27:31 <ogra_ac> it might not even be in teh maverick oem-config anymore
Sep 21 18:27:52 <ogra_ac> since ubiquity was nearly completely rewritten
Sep 21 18:28:36 <Neko> that explains why it never f**king works :D
Sep 21 18:28:44 <ogra_ac> it does
Sep 21 18:28:55 <ogra_ac> in lucid and karmic
Sep 21 18:30:13 <Neko> regardless of what I do, if I run oem-config-prepare or just manually touch that /var/lib/oem-config/run thing..
Sep 21 18:30:27 <Neko> all I get is back to the gdm login prompt
Sep 21 18:30:30 <ogra_ac> you shoudnt run -prepare
Sep 21 18:30:42 <ogra_ac> just use teh touch
Sep 21 18:30:56 <Neko> all -prepare does is touch, rm -f persistent net rule, pop up a zenity dialog unless you said quiet
Sep 21 18:31:16 <ogra_ac> iirc it also creates a user
Sep 21 18:31:17 <Neko> and remove the oem-config-prepare-gtk.desktop from /usr/share/applications iirc
Sep 21 18:31:20 <Neko> no not here
Sep 21 18:31:40 <ogra_ac> ah, then thats the bit that runs during teh live env
Sep 21 18:31:49 <ogra_ac> -prepare is in two parts iirc
Sep 21 18:31:56 <ogra_ac> just dont use it
Sep 21 18:32:12 <ogra_ac> touching teh file after creating teh dir should be just fine
Sep 21 18:32:21 <Neko> sigh, it doesn't :(
Sep 21 18:32:30 <ogra_ac> well, works on our aily images
Sep 21 18:32:33 <Neko> it never runs. I have no initramfs but then I don't see how that makes any difference at all
Sep 21 18:32:36 <ogra_ac> *daily
Sep 21 18:32:49 <Neko> the file is not in the ramfs, and what is.. is just shit like cryptsetup
Sep 21 18:33:06 <ogra_ac> udev and dbus .... etc etc
Sep 21 18:33:08 <Neko> if you run rootstock that's not in the system anyway
Sep 21 18:33:15 <Neko> udev is the thing that runs it?
Sep 21 18:33:25 <ogra_ac> right, rootstock never gets you the same as an install
Sep 21 18:33:29 <ogra_ac> no
Sep 21 18:33:34 <ogra_ac> udev is in iniramfs
Sep 21 18:33:56 <Neko> but they're also in the rootfs so udev and dbus don't matter then
Sep 21 18:35:33 <Neko> friggin' thing it's supposed to start this upstart job but it's never run
Sep 21 18:35:59 <Neko> gdm runs before it
Sep 21 18:36:14 <Neko> so since gdm is running.. it seems not to run since the display manager is taken
Sep 21 18:36:22 <Neko> ubuntu hate++
Sep 21 18:37:00 * ogra_ac gives up ... its no fun to help you if you offed all teh software we work on
Sep 21 18:37:08 <Neko> offed?
Sep 21 18:39:31 <Neko> it's no fun to work on this because everything is undocumented and somehow some internal canonical secret that doesn't exist on any other linux distribution.. all I can say is I am only glad this isn't SuSE because I would have ripped my eyes out before working on YaST plugins to do this :D
Sep 21 18:42:40 <GrueMaster> Neko: Unfortunately, complaining and ranting doesn't help.  While ogra_ac has been trying to help, it is hard to work through all the rants.  And it keeps others from helping or getting help as well.  If you know anything about development cycles, you'd know that 30% is spent on development and 70% is spent on documentation.  We simply can't afford that 70% at the moment.
Sep 21 18:43:14 <GrueMaster> If you would like to help, the wiki is open for adding documentation.  Launchpad is open fir filing bugs.
Sep 21 18:43:28 <GrueMaster> And we would certainly welcome the assistance.
Sep 21 18:43:31 <rsalveti> and it's not a internal canonical secrect
Sep 21 18:43:48 <Neko> I have a good idea oem services charges for the stuff we're trying to do here and I know what you get at the end is a frozen, no security updates repository, with still no documentation for the money paid... I'm not sure what you mean by "afford" here
Sep 21 18:44:27 <GrueMaster> Well, we are not oem services, for one.
Sep 21 18:45:13 <Neko> I don't see what me adding to the wiki solves. I don't know what I am doing, or where I am looking, which is why I ask in here.. if I knew already why would I be looking for support? If I did know you can be sure I'd put it on a wiki or a blog somewhere
Sep 21 18:45:33 <Neko> I don't even know what I'd file a bug against or what the description is besides "nobody knows how this works, but somehow it's broken"
Sep 21 18:46:07 <GrueMaster> We are a small handful of engineers (and one tester) trying to get working images on prerelease hardware.  We spend days trying to fix issues like broken kernels and video corruption.
Sep 21 18:46:27 <GrueMaster> Not much time left in the day for documenting how to roll your own images on unsupported hardware.
Sep 21 18:46:49 <ogra_ac> there is a ton of documentation on th ewiki
Sep 21 18:47:07 <ogra_ac> especially for cutomizing stuff
Sep 21 18:47:10 <Neko> well, where I work, I do all that here too.. that is why I am here asking. We have one guy, which is me, no testers, besides me.. no experience, besides mine, which is admittedly tantamount to zero on this stuff
Sep 21 18:47:26 <Neko> unfortunately the hardware is *retail* and was shipping 4 weeks ago
Sep 21 18:47:32 <Neko> or 8 months ago for the efikamx
Sep 21 18:48:00 <Neko> I don't have the luxury of letting it break for another release, it is soul destroying to run karmic on an mx51
Sep 21 18:48:49 <GrueMaster> So if something is shipping we should arbitrarily just support it?  Intel has been shipping Poulsbo based systems for 2 years, but even they don't support it in Linux.
Sep 21 18:51:15 * GrueMaster wanders off to do more testing.
Sep 21 19:01:02 <Neko> no I just don't get to spend all my days fixing little issues like video corruption because we can't even get oem-config to work
Sep 21 19:01:09 <Neko> I found a description of what is going on and I'm going to report it
Sep 21 19:02:22 <armin76> feature :D
Sep 21 19:04:18 <Neko> do you think that it is possible that oem-config systems are meant to run tasksel such that gdm is installed after creating a user?
Sep 21 19:13:54 <Neko> bug 644638
Sep 21 19:13:55 <ubot2> Launchpad bug 644638 in ubiquity (Ubuntu) "maverick oem-config gtk frontend does not start on first boot because gdm runs before it (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/644638
Sep 21 19:14:30 <robclark> ogra: I tried the cdimage.. very slick :-)
Sep 21 19:14:33 <robclark> but: mem=460M@0x80000000 mem=256M@0xA0000000
Sep 21 19:14:45 <robclark> is there a reason for omitting the last 256mb?
Sep 21 19:15:09 <GrueMaster> ogra_ac: ^^^
Sep 21 19:19:21 <rsalveti> robclark: the weird mem corruption bug we're facing with highmem
Sep 21 19:19:30 <rsalveti> there's a bug for it, one sec
Sep 21 19:20:03 <rsalveti> bug 633227
Sep 21 19:20:05 <robclark> rsalveti: ahh.. so you are using 768 without highmem?
Sep 21 19:20:05 <ubot2> Launchpad bug 633227 in linux-ti-omap4 (Ubuntu) "instabilities with highmem activated (affects: 2) (heat: 508)" [Undecided,New] https://launchpad.net/bugs/633227
Sep 21 19:20:23 <rsalveti> robclark: we have highmem activated, but not using the last 256mb
Sep 21 19:20:39 <rsalveti> when we use it, we can easily get memory issues
Sep 21 19:20:41 <robclark> hmm.. ok, so it is last 256mb that is the issue?
Sep 21 19:20:44 <rsalveti> just try to build a kernel
Sep 21 19:20:50 <robclark> yeah, I've seen the issues..
Sep 21 19:21:00 <rsalveti> yep..
Sep 21 19:21:43 <rsalveti> we got a patch that improved a little bit, but it's still broken
Sep 21 19:22:27 <robclark> ok, yeah.. I saw that.  I just wasn't expecting skipping last 256m would work around the issue.. interesting
Sep 21 19:22:42 <rsalveti> robclark: any idea if x-loader could be the culprit?
Sep 21 19:22:56 <rsalveti> we wanted to test current upstream, but mmc is broken
Sep 21 19:23:48 <robclark> not sure.. I assumed it was highmem, but I don't know enough about highmem to know why it would make a difference, 768 vs 1024..
Sep 21 19:25:47 <rsalveti> robclark: yep...
Sep 21 19:26:28 <rsalveti> from http://lkml.org/lkml/2010/9/8/425 it shows that this issue was happening with others too
Sep 21 19:26:38 <rsalveti> but then with the proposed fix we're still facing issues
Sep 21 19:26:45 <rsalveti> better, but not yet fixed
Sep 21 19:30:35 * rsalveti brb
Sep 21 21:33:24 <robclark> rsalveti: w/ the cdimage filesystem (with initrd and all that), are there any special instructions to use my own (cross-compiled) kernel?
Sep 21 21:33:38 <robclark> ie. does the initrd need to be updated w/ new modules or anything like that?
Sep 21 21:35:51 <rsalveti> robclark: I'm not that sure if you'll need a new uInitrd, as I generally get the new generated one when I install my package
Sep 21 21:36:21 <rsalveti> if you're running your board already, just install your new kernel package, that's enough to regenerate the initrd.img
Sep 21 21:36:35 <rsalveti> than I believe flash-kernel could do the work of updating it for you
Sep 21 21:36:39 <rsalveti> if the links are correct at /boot
Sep 21 21:36:45 <robclark> hmm.. and if I'm building the kernel on my laptop?
Sep 21 21:36:59 <robclark> ok..  so scp over my uImage and modules, and then run flash-kernel?
Sep 21 21:37:33 <rsalveti> robclark: that's ok, what I mean is that if you plan to install it at your running panda or planning to mount the sd card at your host and then install the package
Sep 21 21:38:09 <rsalveti> oh, what you're planning? to copy the modules by hand or to install the generated deb file?
Sep 21 21:38:24 <rsalveti> I generally run: make -j 6 ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- CONFIG_DEBUG_SECTION_MISMATCH=y deb-pkg
Sep 21 21:38:25 <robclark> I could do it either way.. whatever is easier..
Sep 21 21:38:41 <robclark> hmm, ok.. there is a deb-pkg make target?
Sep 21 21:38:48 <rsalveti> then scp the kernel-image, dpkg -i, check the link and flash-kernel
Sep 21 21:38:59 <rsalveti> robclark: yep :-)
Sep 21 21:38:59 <robclark> I guess that is something on the ubuntu kernel?  Or is that in upstream kernel makefiles?
Sep 21 21:39:06 <rsalveti> robclark: upstream
Sep 21 21:39:12 <robclark> oh.. cool
Sep 21 21:40:59 * robclark learns something new every day
Sep 21 21:42:07 <rsalveti> hm..., raining a lot today, if I get off-line that should be the reason
Sep 21 21:45:13 <GrueMaster> swim faster.
Sep 21 21:45:16 <robclark> weather.com says that we might get some rain tomorrow..   then again, the weather rarely does what the forecast says here
Sep 21 21:46:04 * GrueMaster gets weather updates from Dr. Quid B. Wong.
Sep 21 21:46:37 * robclark usually finds out about the weather after the fact
Sep 21 21:49:38 * rsalveti uses the gnome-weather applet :-)
Sep 21 22:31:13 <persia> devilhorns, Earlier you said we had to run gnome-panel: would it not be possible to host indicators and/or menus somewhere else?
Sep 21 22:31:44 <devilhorns> persia, ideally yes ... gnome-panel is very heavy :(
Sep 21 22:31:59 <devilhorns> tho the only other thing I know if is lxde-panel
Sep 21 22:32:03 <devilhorns> or perhaps xfce-panel
Sep 21 22:32:33 <persia> Those share a common ancestor :)
Sep 21 22:32:41 <devilhorns> yea :/
Sep 21 22:34:09 <devilhorns> persia, I'd like to be able to make an efl panel that would be able to host them ... but that means I'd have to dig into the core applet system and figure out how they are expected to be hosted, etc, etc
Sep 21 22:34:14 <devilhorns> lots of issues there :/
Sep 21 22:35:29 <davidm> persia, devilhorns lets not go too crazy here please
Sep 21 22:35:34 <persia> Indeed.
Sep 21 22:35:39 <devilhorns> davidm, hehe indeed :)
Sep 21 22:35:49 <devilhorns> hence why I was just gonna forget about it for now :)
Sep 21 22:35:55 <davidm> that sounds like a natty thing (if at all) :-)
Sep 21 22:35:58 <persia> davidm, No worries.  Just idly chatting about lean and fast :)
Sep 21 22:36:08 <persia> Maybe even natty+
Sep 21 22:36:18 <persia> Very much not maverick
Sep 21 22:36:37 <davidm> persia, don't forget to remind folks that we need blueprint ideas this week, I've told everyone but ....
Sep 21 22:37:31 <GrueMaster> persia: alsa update.  I am able to play audio on beagle using paplay, but not any of the GUI apps (totem, rhythmbox).  I think it is mainly due to memory limitations.
Sep 21 22:38:19 <persia> GrueMaster, Heh.  Thanks for testing that with paplay though.
Sep 21 23:23:22 <GrueMaster> rsalveti: Where is this script for omap4 audio?
Sep 21 23:23:42 <GrueMaster> It isn't in bug 628217.
Sep 21 23:23:43 <ubot2> Launchpad bug 628217 in linux-ti-omap4 (Ubuntu Maverick) (and 1 other project) "No audio output from analog ports on panda (omap4) (affects: 1) (heat: 252)" [High,Confirmed] https://launchpad.net/bugs/628217
Sep 21 23:24:35 <robclark> GrueMaster: amixer.sh?
Sep 21 23:25:06 <GrueMaster> I guess.  I keep hearing about a script to make audio work on the panda analog ports.
Sep 21 23:25:42 <robclark> if rsalveti doesn't beat me to it, I'll fwd in a couple min
Sep 21 23:31:14 <robclark> GrueMaster: http://paste.ubuntu.com/498061/
Sep 21 23:31:21 <robclark> run amixer.sh -a
Sep 21 23:32:56 <GrueMaster> thanks.
Sep 21 23:34:05 <robclark> np
Sep 21 23:34:34 <rlameiro> evening everyone
Sep 21 23:39:33 <GrueMaster> robclark: Still no audio.  Ran sudo sh amixer.sh -a.  Rebooted.  Now nothing shows up in gnome volume control.
Sep 21 23:39:54 <GrueMaster> am I missing a driver?
Sep 21 23:39:59 <robclark> GrueMaster: oh, ok.. maybe still the updated default.pa is needed?
Sep 21 23:40:05 <robclark> hang on..
Sep 21 23:40:12 <GrueMaster> This is a base image
Sep 21 23:41:29 <robclark> GrueMaster: http://paste.ubuntu.com/498064/ is /etc/pulse/default.pa.. and one more coming
Sep 21 23:42:28 <robclark> GrueMaster: http://paste.ubuntu.com/498068/ is daemon.conf
Sep 21 23:42:46 <GrueMaster> Ok.  Will test then add these to the bug report.
Sep 21 23:42:48 <robclark> there is some TI package, I think, that configures everything..
Sep 21 23:43:12 <robclark> and still I think open point about how to do it without overwriting system files.. but that is a known issue I suppose
Sep 21 23:43:58 <GrueMaster> Well, having these attached to the bug report will give the respective developers something to work with.
Sep 21 23:44:08 <robclark> yup
Sep 21 23:47:48 <rlameiro> audio problems too?
Sep 21 23:47:49 <rlameiro> lol
Sep 21 23:47:55 <rlameiro> they are comming
Sep 21 23:48:13 <rlameiro> Evening GrueMaster  :D
Sep 21 23:48:32 <GrueMaster> hello, and welcome back.
Sep 21 23:48:47 <robclark> there is lot of ongoing work for ASoC..  which is making audio a bit painful right now
Sep 21 23:49:02 <robclark> but should make things better for the future
Sep 21 23:50:06 <GrueMaster> yea, I've seen a flood of traffic on alsa-devel mailing list.
Sep 21 23:51:19 <rlameiro> well, i see that i have some serious bugs on the audio setup on igepv2
Sep 21 23:51:29 <rlameiro> but seems to be the same problem with the beagle
Sep 21 23:54:03 <GrueMaster> rlameiro: On beagle, there are a few steps you can take in alsamixer to get audio working.  Mainly unmute DAC2 Analog, HeadsetL Mixer AudioL2 and HeadsetR Mixer AudioR2, and increase volume for Headset and DAC2 Analog.
Sep 21 23:54:37 <rlameiro> i did that on my igep
Sep 21 23:55:10 <GrueMaster> Does "speaker-test -c 2 -t wav" produce audio with those settings?
Sep 21 23:55:26 <rlameiro> but even like that, the driver seems very buggy, as it cant handle big "WORKLOAD"
Sep 21 23:55:48 <rlameiro> GrueMaster: never tried it out with those settings but it should
Sep 21 23:55:57 <rlameiro> my problem is runnig puredata on it
Sep 22 00:08:00 <persia> GrueMaster, robclark: the daemon.conf changes should not be needed.  default.pa is (unless either we find a way to recognise the chip for berco's conf fragment, or the kernel gets fixed)
Sep 22 00:08:37 <robclark> ahh, ok
Sep 22 00:26:47 <persia> Hrm.  Actually, might be able to do something with udev.
Sep 22 00:26:53 * persia reads more docs
Sep 22 00:51:13 <GrueMaster> Gah.  death by alsa.  After copying the default.pa and running the amixer.sh script, I am now getting kernel segfaults in alsa.
Sep 22 00:55:47 <persia> \o/
Sep 22 00:55:59 * GrueMaster bangs head in silence.
Sep 22 00:56:05 <persia> So it's obviously a kernel bug.  The kernel isn't supposed to segfault, regardless of what we do with userspace :)
Sep 22 00:56:57 <GrueMaster> I know.
Sep 22 00:57:04 <GrueMaster> Bug 644828
Sep 22 00:57:06 <ubot2> Launchpad bug 644828 in linux-ti-omap4 (Ubuntu) "BUG: scheduling while atomic: alsa-source/2058/0x00000002 (affects: 1) (heat: 8)" [Undecided,New] https://launchpad.net/bugs/644828
Sep 22 00:57:56 <GrueMaster> I think we are going to continue seeing these atomic issues in SOC systems until Cortex A9 is out in force and developers can fix their code.
Sep 22 00:58:40 <GrueMaster> Until now, no SOC was multicore, and some of the SOC specific code got a little free with the atomic stuff.
Sep 22 01:00:24 <GrueMaster> It's kind of the same growing pains that plagued x86 before SMP code was completed, just not as severe due to the lack of specific devices.
Sep 22 01:32:06 <GrueMaster> persia: I can verify that with the changes to default.pa and running amixer.sh from ti, I can now playback music in rhythmbox on panda.  Not bad.
Sep 22 01:33:30 <persia> amixer.sh makes up for the fact that 1) the kernel doesn't actually know what to do with it and 2) doesn't even provide enough information to userspace to use a userspace config file.
Sep 22 01:34:14 <persia> default.pa is because 1) the kernel doesn't actually know what to do with it and 2) there currently exist no udev rules to explain (although this may be impacted by the same 2) as for amixer.sh: this remains to be determined)
Sep 22 01:35:06 <persia> Unfortunately, because of the 2) for amixer.sh, we are forced to fix this in the kernel somehow, as the userspace config trick won't work (unless we expect every single user to manually run amixer.sh)
Sep 22 01:35:49 <persia> pulse is differently complicated, but because of how udev works, it may be possible to have a special "panda-pulse-enable" package that handles that.
Sep 22 01:36:23 <persia> (and. absolutely worst cast, that hack package could run amixer.sh in postinst or something)
Sep 22 01:37:07 <GrueMaster> What I have found is that I have had to run amixer.sh after each reboot.
Sep 22 01:37:24 <GrueMaster> Even running it sudo doesn't save state properly.
Sep 22 01:38:04 <persia> Ugh.  That's too painful for words.
Sep 22 01:38:08 <persia> Right.  Kernel patch!
Sep 22 01:38:14 <robclark> GrueMaster: I noticed same issue.. it used to work (saving settings on shutdown) with older internal fs that I had
Sep 22 01:38:28 <robclark> but now I have to re-run script too :-(
Sep 22 01:40:32 <GrueMaster> well, it is after 6:30pm here.  I could be doing this while drinking a beer. brb
Sep 22 01:59:12 <persia> Bother.  Making pulse's module-udev-detect work requires properly poplating sysfs.
Sep 22 01:59:49 <persia> Could someone with a panda check /sys/class/sound/card.../pcmC...../pcm_class and tell me what it contains?
Sep 22 02:01:13 <GrueMaster> generic for all.
Sep 22 02:01:18 <persia> berco, /lib/udev/rules.d/78-sound-card.rules might be interesting to you (let's hope it isn't)
Sep 22 02:01:29 <persia> That's not helpful :(
Sep 22 02:02:11 <persia> Yeah, I think we just don't have enough information in userspace to detect that we want to do special stuff right now :(
Sep 22 02:03:48 <GrueMaster> Anything else I can do for you tonight?  If not I'm calling it a day.  12 hours is enough.
Sep 22 02:04:03 <robclark> persia: http://paste.ubuntu.com/498132/
Sep 22 02:04:33 <persia> GrueMaster, Have a good night.
Sep 22 02:05:34 <persia> robclark, Yep.  That just doesn't give us the information we want to be able to automatically populate something equivalent to http://paste.ubuntu.com/498064/
Sep 22 02:06:17 <persia> specifically the "load-module module-alsa-{sink,source} device=..." lines
Sep 22 02:16:24 <ajay> hi all, i have IGEP board unfortunately i saved wrong uboot parameter on it,now i am not able to boot from flash,sdcard as well not getting prompt on minicom
Sep 22 02:18:31 <persia> You probably have to perform some sort of reset on uboot.  Check the docs for your specific board for low-level recovery instructions.
Sep 22 02:18:57 <persia> For example, some boards allow you to hold down some special button to boot from an alternate bootloader
Sep 22 02:19:50 <ajay> persia, thanks but this board doesnot have any type of button as input
Sep 22 02:20:16 <ajay> there is no recovery mode inside flash
Sep 22 02:21:24 <persia> No idea then.  If you can't boot anything, and have no input/output (serial or console), I'm not sure if we can help much.
Sep 22 02:21:51 <persia> I really think you'd do better with the IGEP distributors (although I may be mistaken, and there are other folk here who have similar boards)
Sep 22 02:22:12 <ajay> persia,ok
Sep 22 02:25:05 <davidm> ajay, it may require the use of a hardware debugger
Sep 22 02:25:08 <davidm> jtag
Sep 22 02:25:47 <ajay> davidm, i have onle serial port console that is minuicm
Sep 22 02:25:51 <ajay> minicom
Sep 22 02:27:01 <davidm> ajay, if your board is: http://www.igep-platform.com/index.php?option=com_content&view=article&id=46&Itemid=55
Sep 22 02:27:43 <davidm> Then it has holes for a JTAG connector and that with special instructions will allow you to reflash the boot system
Sep 22 02:28:45 <robclark> will IGEP boot off of MMC before NAND?
Sep 22 02:29:13 <robclark> (or are there some DIP switch or jumpers to make that so)
Sep 22 02:29:13 <ajay> robbiew, yes it tryies to boot from mmc first
Sep 22 02:29:26 <ajay> then flash
Sep 22 02:29:50 <robclark> then you should be able to put a good bootloader on an MMC card and recover that way
Sep 22 02:31:29 <ajay> i did that as well for this i am getting white screen in LCD and no message over minicom
Sep 22 02:32:13 <robclark> hmm, is that different behavior than booting without MMC card?
Sep 22 02:32:21 <robclark> if so, then you know MMC comes first..
Sep 22 02:32:55 <robclark> then it is "just" a matter of getting a correctly formatted MMC card with good MLO and u-boot.bin..
Sep 22 02:33:28 <robclark> (there are scripts floating around to correctly format the MMC card under linux..)
**** ENDING LOGGING AT Wed Sep 22 02:59:57 2010