[04:34] <gatotkaca> how can I use ubuntu-arm with v2.6.29 with only fully serial console?
[04:35] <gatotkaca> I don't understand what is the meaning of: configure upstart to spawn a serial tty event
[04:36] <gatotkaca> I found it based on this: http://irclogs.ubuntu.com/2009/05/14/%23ubuntu-arm.txt
[04:37] <gatotkaca> I've got the same problem; with following logs line: [ OK ]figuring network interfaces...
[04:37] <gatotkaca> [ OK ]ting up console font and keymap...
[04:37] <gatotkaca> [ OK ]rting system log daemon...
[04:37] <gatotkaca> [ OK ]rting kernel log daemon...
[04:44] <Martyn-> gatoacka : Go into the filesystem you are using, and change /etc/inittab to include a ttyAMA0 (or ttyS0, or whatever)
[04:44] <Martyn-> don't forget to change /etc/securetty
[04:46] <gatotkaca> Martyn : Thanks -- /etc/securetty? What should I change?
=== shenki_ is now known as shenki
[09:38] <gatotkaca> anyone working on DSPLink on ubuntu-arm?
[09:38] <gatotkaca> beagleboard?
[09:54] <lool> gatotkaca: We don't do OMAP ATM
[09:55] <lool> gatotkaca: There's no inittab in Ubuntu, we use upstart instead; you just need to create an /etc/event.d file
[09:56] <lool> gatotkaca: https://lists.ubuntu.com/archives/ubuntu-devel/2009-June/028351.html
[10:03] <gatotkaca> lool:how about alsa?
[10:03] <gatotkaca> lool:what should I give to the --seed argument?
[10:03] <gatotkaca> --seed alsa-utils?
[10:12] <lool> gatotkaca: It depends what you need; try it out and see if that's what you want?
[10:23] <gatotkaca> lool:ok I'll try it
[14:29] <rjune_wrk> What is the name of the inexpensive development board avaialbe?
[14:29] <lool> _the_?  :)
[14:30] <lool> rjune_wrk: For Ubuntu, there isn't really one perfect board yet, but in general for ARM v6+ (v7 actually) development, the beagleboard is a nice device
[14:30] <rjune_wrk> that's the one I was thinking of, thanks.
[14:31] <Martyn-> lool : We really really have to get an official branch of Ubuntu 9.04 on the beagle.  I think it was a mistake to not have one on launchpad...
[14:32] <Martyn-> The more I think of it, the more it kinda sucks to have to tell people "yeah, you can run it, but you have to do these six steps to build a rootfs, then wrap it up in a bow this way, in order to get it working... "
[14:38] <rjune_wrk> Martyn-: so is that a project you'll be working on? or will you be delegating that?
[14:46] <armin76> Martyn: thats what a dev board is for :)
[14:49] <kblin> Martyn-: yeah, and so people can start to file kernel bugs ;)
[14:52] <Martyn-> rjune_wrk : Well, considering I've now updated my own beagleboard build seven times .. I guess I just volunteered.
[14:53] <Martyn-> armin76 : That's just it though, the beagleboard is more than just a dev board.  It's a nicely packaged micro motherboard.
[14:53] <Martyn-> armin76 ; One of the few 'dev' boards that's instantly useful out of the box.
[14:53] <Martyn-> plus I just got a build of UEFI for it...
[14:54] <Martyn-> thinks
[14:56] <rjune_wrk> *U* EFI?
[14:57] <Martyn-> Yep.
[14:57] <Martyn-> Better bootloader than any we've had so far
[14:57] <rjune_wrk> what is U ?
[14:58] <Martyn-> http://www.uefi.org/home
[14:58] <rjune_wrk> I thought EFI replaced the BIOS
[14:58] <rjune_wrk> yeah, I went and googled it
[14:58] <Martyn-> 'unified'
[14:58] <Martyn-> it does, and it provides a flexible extensible framework to abstract and access the hardware just like the BIOS
[14:58] <rjune_wrk>  Better bootloader than any we've had so far
[14:58] <Martyn-> so that you only have to write bootloader hooks once
[14:59] <rjune_wrk> shouldn't that be better BIOS then any we've had?
[14:59] <Martyn-> ARM doesn't have a BIOS
[14:59] <Martyn-> in fact, only x86 arch has a bios
[15:00] <Martyn-> Linux is forced to act like one, for just about all the architectures.   If you look at the crazyness in the linux kernel branch, with all the weirdnesses that each mach- and plat- sub-branch need...
[15:02] <Martyn-> So, while UEFI doens't handle POST, and CPU rom bootstrap setup, it does provide everything else.  That includes the ability to boot a kernel that matches the EFI 1.10 spec
[15:04] <Martyn-> Apple is graciously providing source code to their UEFI implementation
[15:13] <lool> I didn't know of EFI implementations for BB, cool
[15:13] <rjune_wrk> I'm guessing BB does not refer to busybox
[15:14] <Martyn-> lool : I just got it on Monday
[15:14] <Martyn-> lool : With much whispering and sup3r s3kr3t handshakes (I wish I was kidding)
[15:15] <Martyn-> however, after reading the code, it's under a dual license until a given date
[15:15] <Martyn-> after that date, its good ol open source.
[15:15] <Martyn-> (and I think there will be much rejoycing)
[15:39] <armin76> Martyn: hrm...it doesn't have eth
=== thschee_ is now known as thschee
[15:56] <Martyn-> armin76 : So?  It has full USB and usb OTG
[15:56] <Martyn-> not having ethernet is as easy as plugging an ethernet adapter in
[15:56] <Martyn-> and poof .. you have ethernet
[15:57] <Martyn-> most (useful) beagle implementations I've seen involve plugging the beagle into a micro USB hub, and then putting peripherals on the hub
[15:57] <Martyn-> With the C3 revision of the board (256MB ram) and the C4 around the corner (512MB ram) .. it's a useful little computer
[16:21] <kblin> Martyn-: on the revB, USB is a bit sucky
[16:21] <kblin> Martyn-: at least on ogra's ubuntu kernels, it's pretty unstable
[16:22] <lool> Martyn-: Oh do you have infos on C4?  512m would be gorgious
[16:22] <kblin> I can have the kernel watchdog kick out the ethernet if I cause simultaneous high-speed USB traffic, e.g. to a usb hdd
[16:26] <Martyn-> lool : that's the C4's big upgrade
[16:26] <Martyn-> lool : They are kicking around putting ethernet on the board, but it may be too expensive to lay down the PHY
[16:27] <Martyn-> both in $$ an on the power budget
[16:27] <lool> I see
[16:27] <Martyn-> ETH PHY can absorb 20W by itself
[16:29] <Martyn-> er .. 2W, sorry
[16:29] <Martyn-> order of mag
[16:29] <rjune_wrk> I was gonna say, 20W for eth seemed a bit high
[16:30] <Martyn-> it is
[16:30] <Martyn-> I keep getting that one wrong
[16:30] <Martyn-> okay, I'm heading to work, cause I need the lab
[16:31] <Martyn-> back online in a bit
[16:38] <ogra> kblin, my ubuntu kernels ?
[16:40] <ogra> kblin, you really shouldnt use any kernels i do ;) ... http://www.rcn-ee.com/deb/kernel/beagle/ has properly packaged ones
[16:41] <ogra> kblin, if i do kernels they are usually for a first test
[16:55] <kblin> ogra: now that you remind me, that's what I'm using in my current set-up..
[16:55] <kblin> or rather, the ubuntu-update-kernel.sh script does
[16:56] <ogra> yeah, thats not mine ... i only did a very early 2.6.27 build from linux-omap ... doesnt even properly support the beagle framebuffer
[16:56] <ogra> just to get a ssh server up on the board to tinker with it
[16:57] <ogra> these kernels are done by robert nelson, he works for digikey afaik
[16:58] <ogra> his mail address is in http://www.rcn-ee.com/deb/kernel/changes.txt
[16:58] <ogra> in case you want to contact him
[17:04] <kblin> ogra: I might have used your kernel in one of my earlier tries.. I wouldn't notice the framebuffer missing, no monitor connected to my BB
[17:04] <ogra> heh
[17:08] <kblin> besides, this seems to be a known issue, but upstream closed the bug with wontfix as it can be worked around by forcing the high-speed usb device to full-speed usb
[17:09] <kblin> and I currently don't have the time to test and argue :)
[17:10] <ogra> and you dont use angstrom :)
[17:10] <kblin> yeah, bad me
[17:10] <ogra> while koen and friends are very helpful, they dont like people running ubuntu on the beagle
[17:11] <ogra> so the help you can get is rather marginal beyond education and info from them ...
[17:11] <kblin> I'm pretty sure it's a kernel problem, so I'd bet I can replicate it with angstrom, didn't have time to play with that yet
[17:11] <ogra> (i really like #beagle, but they dont really focus on ubuntu :) )
[17:12] <kblin> I'm working on getting a demo working for the storage developers conference in sept, and I felt more comfortable keeping the distro for development given that I changed the architecture already
[17:12] <kblin> for the demo, I'm fine with going at USB 1.1 speeds
[17:27] <Martyn> re
[17:28] <ogra> hmm, i remember i had decent USB speed when i used a xubuntu build with roberts kernels ...
[17:28] <ogra> even swapping to a USB disk to run ubuntu-desktop on the 128M board i have worked in a usable way
[17:29] <ogra> but its quite a while ago that i even booted my beagle
[17:32] <armin76> Martyn: you have a beagleboard, right? can you tftpboot from u-boot with an usb eth? :)
[17:33] <armin76> if it just had eth :P
[17:35] <kblin> ogra: yeah, I can either have high-speed usb or full-speed usb, but both seems to cause trouble
[18:54] <lool> NCommander: https://blueprints.launchpad.net/ubuntu/+spec/mobile-unr-karmic-other-arches is obsolete, I'm dropping it from /Roadmap
[18:55] <ogra> was that about porting to mips ?
[18:55] <lool> same for https://blueprints.launchpad.net/ubuntu/+spec/mobile-unr-karmic-desktop-switcher-improvements and https://blueprints.launchpad.net/ubuntu/+spec/mobile-karmic-moblin2
[18:55] <ogra> oh, lpia
[18:56] <NCommander> ogra, I wish :-/
[19:01] <Martyn> re
[19:01] <Martyn> armin76 : I can
[19:01] <Martyn> armin76 : however, I had to extend u-boot to handle USB ethernet booting, and it only works with two adapters
[19:01] <Martyn> armin76 : The -right- way to do USB booting is to use the linux kernel's ability to do a kexec
[19:01] <Martyn> USB-ethernet that is
[19:02]  * lool &
[19:02] <ogra> the right way to do netbooting is to add an ethernetport to the damned hardware :)
[19:02] <Martyn> armin76 : What you do is use u-boot to boot a small linux kernel from NAND with a cramfs at the end.  It's entire job is to bring up USB and networking, then attempt to do a DHCP/tftp boot of another kernel
[19:03] <Martyn> ogra : Hey, add your voice to the C4's request list!
[19:03] <Martyn> -grin-
[19:03] <kblin> ogra: actually, I'd probably just use other damned hardware
[19:03] <ogra> i doubt i have to, i think thats been requested often enough :)
[19:03] <Martyn> i think the big deal right now is the cost of the PHY, since there is a mac already present in the OMAP
[19:03] <Martyn> kblin : name one other piece of ARM hardware with as much capability as the beagleboard, for anywhere near the price...
[19:03] <ogra> if the beagle had an ethernet port i'D write an LTSP implementation for ubuntu-arm within two days
[19:04] <Martyn> ogra : Post that then :)
[19:04] <ogra> and the beagle would be the next generation thin client ruling the world
[19:04] <Martyn> ogra : That's a threat you can put teeth on :)
[19:04] <ogra> heh
[19:04] <Martyn> I'm not kidding.
[19:04] <Martyn> Post that, and make that level of commitment (minus the whole two days thing) .. and you'd get a very strong response
[19:05] <Martyn> I don't have that kind of spare time, and couldn't get my company to sponsor my time to do it
[19:05] <kblin> Martyn: depends on what you want to with it
[19:05] <ogra> why not the two day thing ... i wrote most of the recent LTSP ... it wouldnt take me more :P
[19:05] <kblin> Martyn: I'm using my beagle as network server, and these days I'd probably get a sheevaplug..
[19:05] <Martyn> ogra : Well, because it sounds like a boast, really.
[19:06] <Martyn> ogra : Which would probably be recieved with a bit of "oh yea?  Pfft."  which is bad :)
[19:06] <Martyn> (even if it's true .. -heh-)
[19:06] <ogra> well ...
[19:06] <kblin> Martyn: of course if you want the whole display GUI thing, I don't know that part
[19:06] <ogra> i'll just wait who comes first with an ethernet port built into the board :)
[19:07] <Martyn> kblin : I have a shivaplug, and what they did in the mach-kirkwood subtree is an abomination.   Their implementation of cpuidle.c (which DOESN'T have cpu_idle_wait, damnit) is semi broken for anyone else
[19:07] <Martyn> so I can't reuse the code for Cortex-A9, which is a shame and a half.
[19:07] <Martyn> kblin : I have usb-display working on my shivaplug :)
[19:07] <kblin> Martyn: possibly. I don't have one of the things
[19:07] <Martyn> kblin : X on a shivaplug is .. very .. very .. slow .. but pretty
[19:08] <Martyn> ogra : Even if it's not built in .. why not support USB ethernet?
[19:08] <kblin> Martyn: all I know is that the BB (revB is what I could test) sucks as a storage server
[19:08] <ogra> Martyn, way more work ...
[19:08] <Martyn> kblin : Yep
[19:08] <Martyn> Seriously, just WAIT until we have our chip done :)
[19:08] <kblin> I mean, I can do domain logons to it from my local XP machines
[19:08] <Martyn> Or even when TI and Freescale have theirs done
[19:09] <Martyn> (samsung's Cortex-A9 run is going to be completely bought out by Apple.  fuck 'em)
[19:09] <kblin> but as soon as I want to transfer any amount of data, usb 1.1 speed really bites
[19:12] <Martyn> kblin : 1.1?  Why just 1.1?
[19:12] <Martyn> I thought there was a full 2.0 in there .. it's just 1.1?
[19:12] <kblin> http://code.google.com/p/beagleboard/issues/detail?id=31
[19:13] <kblin> I've yet to try a revC board, but I figure if this was fixed, the hardware would be a pretty decent SOHO network server
[19:14] <kblin> a real ethernet port would help, of course
[19:14] <suihkulokki> ogra: gumstix overo has a breakout board with ethernet
[19:14] <suihkulokki> ethernet over usb is just flaky (especially if the usb in general is as flaky as on omaps..)
[19:15] <ogra> yes
[19:15] <ogra> and for LTSP you put a lot load on the NIC
[19:18] <Martyn> true
[19:18] <Martyn> Well, tell you what
[19:18] <Martyn> I'm meeting with Gordon Freeburg next week... i'll ask if he has anything beagle-ish in the works based around the OMAP
[19:19] <Martyn> i know the last time I talked with him, he had a plan for a less gumstick-y board that could be more of a general development and prototyping board.
[19:19] <Martyn> i.e. chip on the board, no risers or expensive mezzanine connectors
[19:23] <suihkulokki> seems there are already beagle derivates with ethernet: http://www.igep-platform.com/index.php?option=com_content&view=article&id=46&Itemid=55
[19:50] <armin76> Martyn: thats a mess :)
[20:02] <Martyn> And where can I _buy_ an IGEPv2 board?
[20:02] <Martyn> As far as I can see, they only produced it as a concept
=== ogra__ is now known as ogra_