**** BEGIN LOGGING AT Sun Jul 17 23:59:57 2005
Jul 18 04:54:05 <NAiL> yay
Jul 18 04:54:18 * NAiL got a replacement drive for the broken S-Ata drive
Jul 18 04:54:52 <NAiL> Should start taking backups though...
Jul 18 06:10:57 * kolla_ pats his slug
Jul 18 06:11:09 <kolla_> good thing is compiling new gcc and needs some patting
Jul 18 06:15:33 <DaKa2> compiling gcc, sounds fun
Jul 18 06:15:55 <kolla_> sure is :)
Jul 18 06:23:05 <kolla_> DaKa2: it did it on saturday too, this is second round in the bootstrapping process :)
Jul 18 06:23:30 <DaKa2> hehe, lots of fun.. but, why?
Jul 18 06:23:48 <kolla_> gentoo/armeb-uclibc
Jul 18 06:24:15 <DaKa2> ahh...
Jul 18 06:25:22 <VoodooZ_Work> morning [g2]
Jul 18 06:25:30 <[g2]> morning VoodooZ_Work 
Jul 18 06:25:39 <DaKa2> morning, almost
Jul 18 06:25:43 <VoodooZ_Work> hehe
Jul 18 06:25:55 <[g2]> hey sunshine!
Jul 18 06:26:05 <VoodooZ_Work> DaKa2, guess what I managed to break this weekend?
Jul 18 06:26:20 <VoodooZ_Work> I blew up my I2C!! Yes! Again!
Jul 18 06:26:23 <DaKa2> another i2c pin on the slug?
Jul 18 06:26:24 <DaKa2> ahh :-)
Jul 18 06:26:37 <[g2]> I'm up to jam building on that latest full compile
Jul 18 06:26:45 <DaKa2> well, you have a few leds to go..
Jul 18 06:26:48 <VoodooZ_Work> I could fix it again by rewiring the only LED pin left but I decided to move all my sensors/motor to USB.
Jul 18 06:27:03 <VoodooZ_Work> Should be interesting and way cooler.
Jul 18 06:27:25 <VoodooZ_Work> I2C and serial is so 90s anyways! ;)
Jul 18 06:27:29 <DaKa2> ...
Jul 18 06:27:31 <DaKa2> hey!
Jul 18 06:28:00 <DaKa2> Ive been thinking about going usb, but i2c works damn well, 
Jul 18 06:28:01 <VoodooZ_Work> so I'm thinking of trying that software AVRUSB thing first as it's cheap.
Jul 18 06:28:38 <VoodooZ_Work> I'll still be using I2C but I'll have an AVR collecting all the I2C data and then the ARM will fetch it via USB.
Jul 18 06:29:02 <DaKa2> ah
Jul 18 06:29:08 <DaKa2> that would work
Jul 18 06:29:10 <VoodooZ_Work> It will only do 1.5Mbit/s but that's plenty fast compared to 100Khz I2C.
Jul 18 06:29:38 <VoodooZ_Work> check it out here: http://www.obdev.at/products/avrusb/
Jul 18 06:29:41 <DaKa2> And you probably doesn't even need the full 100Khz I2C
Jul 18 06:30:14 <VoodooZ_Work> The original guy who did that is here: http://www.cesko.host.sk/
Jul 18 06:30:29 <VoodooZ_Work> latency is a killer though.
Jul 18 06:30:51 <DaKa2> hm.. nice.. software usb
Jul 18 06:31:06 <VoodooZ_Work> And with the software I2C module I could feel it when doing some vision stuff. like 3-4 fps drop
Jul 18 06:31:18 <rwhitby> VoodooZ_Work: there's a post on the mailing list especially for you :-)
Jul 18 06:31:31 <VoodooZ_Work> If It doesn't do the job I'll simply have to bite the bullet and buy those new USB AVRs.
Jul 18 06:31:48 <rwhitby> [g2]: the poster might be a future customer of yours too
Jul 18 06:31:56 <VoodooZ_Work> rwhitby, really? Does it make fun of my special ability to break things? :)
Jul 18 06:31:58 <VoodooZ_Work> customer?
Jul 18 06:32:02 <VoodooZ_Work> let me check!!!
Jul 18 06:32:55 <VoodooZ_Work> you mean the I2C one from Simon?
Jul 18 06:33:00 <rwhitby> yep
Jul 18 06:33:35 <VoodooZ_Work> cool. I'll reply.
Jul 18 06:35:31 <rwhitby> night all
Jul 18 06:37:12 <[g2]> rwhitby-asleep, nite
Jul 18 06:37:24 <VoodooZ_Work> night
Jul 18 08:38:57 <DaKa2> [g2]: ping?
Jul 18 08:39:04 <[g2]> DaKa2, pong
Jul 18 08:39:15 <DaKa2> how should we do about the busybox thing?
Jul 18 08:39:21 <DaKa2> update-altenatives
Jul 18 08:39:26 <DaKa2> http://david.thg.se/saker/openslug/conflicting-files
Jul 18 08:40:17 <DaKa2> one problem is coreutils.. that places stuff in /usr/bin where busybox places in /bin
Jul 18 08:40:46 <[g2]> I doubt we are the first running across this bridge
Jul 18 08:41:10 <DaKa2> yep...
Jul 18 08:41:34 <[g2]> I think we need to chat with ppl in OE and see if some other distros solved it and see whether we like that approach or not
Jul 18 08:42:11 <DaKa2> hm.. yes, probably
Jul 18 08:43:05 <[g2]> did you see the updated OpenSlug/Packages page ?
Jul 18 08:43:11 <DaKa2> yep
Jul 18 08:43:28 <[g2]> everything built cleanly from the latest pull
Jul 18 08:43:39 <DaKa2> I look at AllRecentChanges from time to time
Jul 18 08:43:55 <[g2]> yeah me too :)
Jul 18 08:44:27 <DaKa2> :) well, do you feel like taking to the oe ppl?
Jul 18 08:44:39 <[g2]> oh sure....
Jul 18 08:44:57 <[g2]> I've wanted to look at the update-alternative
Jul 18 08:45:25 <[g2]> however, the fix for the reboot is *really* important
Jul 18 08:45:49 <DaKa2> hehe, yes
Jul 18 08:45:55 <DaKa2> I can agree on that
Jul 18 08:46:19 <[g2]> I'll check with jbowler-away and NAiL but here are my thoughts on the next binary release
Jul 18 08:47:03 <[g2]> 1) test the reboot fix
Jul 18 08:47:33 <[g2]> 2) patch the kernel to disable the irq26:noboby cared issue as with thing is a red herring
Jul 18 08:48:04 <[g2]> 3) work out the update-alternatives issue for the native install
Jul 18 08:48:54 <[g2]> 3a have perl/python natively built for the stable feed
Jul 18 08:49:14 <[g2]> 3a  is kinda optional for the release but would be nice
Jul 18 08:49:22 <DaKa2> really nice
Jul 18 08:49:38 <[g2]> it's quite likely that 3 and 3a will be done in a similar time frame
Jul 18 08:50:15 <[g2]> jbowler-away and I are the co-leads on OpenSlug
Jul 18 08:50:24 <NAiL> [g2]: That's more or less my thoughts too
Jul 18 08:50:33 <NAiL> That, and some small changes to the feed
Jul 18 08:50:42 <DaKa2> do we really need the shared perl lib? otherwise half of 3a is done
Jul 18 08:52:11 <[g2]> I think so
Jul 18 08:52:17 <[g2]> but I don't really know
Jul 18 08:52:26 <NAiL> [g2]: Going to change the feeds to feeds/openslug/[cross|native]/[$DISTRO_VERSION|unstable]/ before the next release
Jul 18 08:52:28 <[g2]> have you installed file ?
Jul 18 08:53:10 <DaKa2> hm.. no...
Jul 18 08:53:13 <[g2]> NAiL, that's sounds like a good plan -- Not that I've ever used the feeds though 
Jul 18 08:53:33 <[g2]> file is needed for creating the python .so
Jul 18 08:53:56 <[g2]> maybe an updated libtoo (haven't verified if just file does it)
Jul 18 08:54:04 <[g2]> libtool
Jul 18 08:54:33 <DaKa2> Ill try that
Jul 18 08:55:59 <DaKa2> well, in some time..
Jul 18 09:28:07 <jbowler-away> The fix for /disk/network boot with static ip needs to go in - I have it but haven't tested it yet
Jul 18 09:28:29 <jbowler-away> Any changes to the feed must go in.
Jul 18 09:28:38 <NAiL> yes
Jul 18 09:28:45 <jbowler-away> If busybox is going to set the alternatives that must go in.
Jul 18 09:29:12 <jbowler-away> The irq26 is optional because it is a simple kernel upgrade.
Jul 18 09:29:47 <NAiL> hmm? Is it fixed anywhere? Or are you talking "if there is a irq26 fix, just flash a new kernel?"
Jul 18 09:29:57 <jbowler-away> (I.e. it doesn't cause any problems if someone doesn't have the fix).
Jul 18 09:30:42 <jbowler-away> The suggestion on irq26 was to disable the message somehow, I meant the latter - if it comes in after a binary release and a user cares just reflash the kernel.
Jul 18 09:30:43 <NAiL> 4.
Jul 18 09:31:11 <NAiL> ok
Jul 18 09:31:45 <jbowler-away> If someone was talking about actually fixing the problem I might think differently ;-)
Jul 18 09:34:42 <NAiL> There was talk a while back, but I don't think anything is gonna happen before the release
Jul 18 09:35:03 <NAiL> I've got no idea how to even try to figure out how to fix it
Jul 18 09:36:23 <jbowler-away> It's a matter of understanding why that irq is generated and whether it matters that nothing apparently handles it (maybe there is a bug and it really is handled...)
Jul 18 09:36:55 <jbowler-away> It's only interesting if it is a symptom of a real problem - but there is no evidence yet of any real problem.
Jul 18 09:37:12 <NAiL> The only thing I've noticed is that it slowed down samba on my slug. But that was apparently *only* on my slug 
Jul 18 09:37:56 <jbowler-away> Generation of a lot of kernel messages will do that - particularly if they are logged to disk - the suggested fix (disabling the message) would fix that problem.
Jul 18 09:38:26 <[g2]> jbowler-away, I'll forward the e-mail exchange with lennert on the subject
Jul 18 09:38:49 <[g2]> it was based on his discussions with Deepak
Jul 18 09:38:55 <jbowler-away> Still, my gentoo system generates fantastic amounts of 'atomic counter underflow' messages (because I built an mm kernel with preemptive scheduling) and I see no problems.
Jul 18 09:39:12 <jbowler-away> [g2]: that would be interesting
Jul 18 09:42:12 <[g2]> jbowler-away, sent
Jul 18 09:43:17 <[g2]> jbowler-away, when beewoolie and I like at the irq26 problem in the driver no other interrupts are pending
Jul 18 09:43:29 <[g2]> s/like/looked/
Jul 18 09:48:50 <jbowler-away> What I never understood was what driver should handle irq26 - or is it that the irq's are shared now (like on a PC?)  so there might be several drivers?
Jul 18 09:50:00 <jacques>  26:     409942   ehci_hcd:usb1
Jul 18 09:50:01 <[g2]> I think the EHCI driver just registers for irq26
Jul 18 09:50:43 <[g2]> I'd imagine it's actually coming in on the PCI INT(A|B|C) line
Jul 18 09:51:35 <[g2]> ehci_hcd 0000:00:01.2: irq 26, io mem 0x48002000
Jul 18 09:51:55 <[g2]> ohci_hcd 0000:00:01.0: irq 28, io mem 0x48000000
Jul 18 09:52:11 <[g2]> ohci_hcd 0000:00:01.1: irq 27, io mem 0x48001000
Jul 18 09:56:30 <jbowler-away> dyoung pointed out some strangeness in the defconfig in this area too..
Jul 18 10:05:04 <[g2]> jbowler-away, NAiL after a little testing, fixing/masking the irq26 issue, and working out the update-alternative for all the components and perl and python in the ipkg feed.....
Jul 18 10:05:33 <[g2]> How about a real, as in non-beta, OpenSlug release ?
Jul 18 10:10:25 <jbowler-away> I thought the theory was beta == not-supported and that there is no support.
Jul 18 10:11:19 <[g2]> I unslug beta ?
Jul 18 10:11:21 <[g2]> is
Jul 18 10:11:31 <jbowler-away> USB_OHCI_LITTLE_ENDIAN is set to 'y' because of:
Jul 18 10:11:33 <jbowler-away>         default n if STB03xxx || PPC_MPC52xx
Jul 18 10:11:33 <jbowler-away>         default y
Jul 18 10:12:07 <jbowler-away> DISTRO_NAME = "Unslung"
Jul 18 10:12:08 <jbowler-away> DISTRO_VERSION = "5.5-beta"
Jul 18 10:12:08 <jbowler-away> DISTRO_TYPE = "beta"
Jul 18 10:12:16 <jbowler-away> From unslung.conf
Jul 18 10:13:10 <[g2]> I'm not sure I fully understand what "supported" means from an open-source project like this
Jul 18 10:15:44 <jbowler-away> IMO it means something where if you have an issue there is some route for someone other than you to fix it.
Jul 18 10:16:33 <jbowler-away> As opposed to one where a reasonable response is "yes, that seems to be a problem, will you update the wiki when you find out how to fix it?"
Jul 18 10:18:24 <[g2]> so all the sf projects in stable/production are really beta then because there's no real support for them
Jul 18 10:19:13 <[g2]> and clearly with a dev project on sf with 80+ dev's we're probably a lot closer to the largest projects there
Jul 18 10:20:52 <[g2]> ok.. 77 devs
Jul 18 10:21:42 <jbowler-away> Yes, by this definition.
Jul 18 10:26:36 <jbowler-away> So the irq26 issue would simply seem to be that the code doesn't handle the fact that it can't reliably clear an asynchronous interrupt - so it may see the same interrupt twice.
Jul 18 10:26:58 <jbowler-away> That's a well know problem isn't it?  Related to the halting problem or something like that?
**** BEGIN LOGGING AT Mon Jul 18 10:51:32 2005
Jul 18 11:24:05 <[g2]> hanjo, hey!
Jul 18 11:24:16 <hanjo> hi
Jul 18 11:24:44 <hanjo> g2, is there a new openslug binary release in the pipeline?
Jul 18 11:24:52 <[g2]> yup :)
Jul 18 11:25:39 <hanjo> with a fixed shudown -r ?
Jul 18 11:25:47 <[g2]> yup :)
Jul 18 11:26:01 <hanjo> great
Jul 18 11:26:09 <[g2]> thank jbowler-away and dyoung 
Jul 18 11:27:25 <hanjo> does the fix require a recompile?
Jul 18 11:27:39 <hanjo> or is it a script change, etc.
Jul 18 11:27:45 <[g2]> it's a kernel fix
Jul 18 11:27:51 <hanjo> oh, sh...
Jul 18 11:28:34 <[g2]> are you running 2.6.11.2 ?
Jul 18 11:28:58 <hanjo> I beleive so. I am at home now, without access to my slug farm :)
Jul 18 11:29:07 <[g2]> ok..
Jul 18 11:29:34 <[g2]> We're on 2.6.12.2 and dyoung had fixed the console issue
Jul 18 11:29:53 <hanjo> I don't have serial so...
Jul 18 11:30:05 <[g2]> nm then
Jul 18 11:30:36 <[g2]> the native compile environ is pretty much all sorted out
Jul 18 11:30:40 <hanjo> So what would be the best way to update to the new relase in your view, under the assumption that you can not manualy reset the slugs?
Jul 18 11:31:23 <[g2]> do you have the shutdown -h  ?
Jul 18 11:31:26 <hanjo> g2, do you remember who was playing with the libusb?
Jul 18 11:31:32 <[g2]> DaKa2, 
Jul 18 11:31:35 <[g2]> it's fixed
Jul 18 11:31:51 <VoodooZ_Work> good as I'll need it!
Jul 18 11:32:02 <hanjo> I have ssh access to the slugs, but they are distributed in the building and I don't have the keys from all the rooms :)
Jul 18 11:32:03 <[g2]> well don't take my word for it :)
Jul 18 11:32:20 <VoodooZ_Work> I'll know soon enough!
Jul 18 11:33:02 <VoodooZ_Work> It's probably going to a pita to deal with endianess when transferring word size data from my devices.
Jul 18 11:33:15 <[g2]> hanjo, don't you have power control over those boxes ?
Jul 18 11:33:15 <hanjo> VoodooZ, are you speaking about the libusb?
Jul 18 11:33:25 <hanjo> g2, no I don't
Jul 18 11:33:44 <VoodooZ_Work> indirectly as I'm starting to convert my I2C based decice on my robot to USB.
Jul 18 11:33:57 <VoodooZ_Work> So I might try using libusb to start instead of writing a kernel driver.
Jul 18 11:34:14 <[g2]> hanjo, what happens if the boxes crash ?
Jul 18 11:34:20 <[g2]> as in lock-up
Jul 18 11:34:53 <[g2]> you're running from external usb right ?
Jul 18 11:34:53 <hanjo> g2, if I have to, I can get a master key for one afternoon, and do the reset. You can see why I'm so happy to hear that we can reset cleanly
Jul 18 11:35:08 <[g2]> so far so good
Jul 18 11:35:13 <VoodooZ_Work> on that note check out this project: http://www.obdev.at/products/avrusb/powerswitch.html
Jul 18 11:35:32 <VoodooZ_Work> usb based power controller.
Jul 18 11:35:33 <[g2]> I've reboot several times with the -r and an external device
Jul 18 11:35:51 <[g2]> hanjo, is a trade-off in time 
Jul 18 11:35:55 <[g2]> it is
Jul 18 11:36:50 <hanjo> VodooZ, that will not work for me, as the slugs are connected over ethernet with the control station
Jul 18 11:36:55 <[g2]> I think you could figure out the magic and ssh in to the devices and basically backport the fix in loadable module
Jul 18 11:37:31 <[g2]> but depending on how big the building is and whether you have the shutdown -h fix
Jul 18 11:38:00 <[g2]> it'd be a lot easier to just upgrade then shutdown -h and then go power the boxes back on
Jul 18 11:38:01 <hanjo> If I know for sure that the the reset is going to work from this point on, I will do this last "manual" step
Jul 18 11:38:16 <hanjo> That's how I plan to do it
Jul 18 11:38:29 <[g2]> well it's in the repo as of last night
Jul 18 11:38:54 <hanjo> or send a mail to my colleagues to press the reset for me:)
Jul 18 11:39:00 <[g2]> sure
Jul 18 11:39:00 <DaKa2> libusb works great, both sane and gphoto2 is tested with it and both works
Jul 18 11:39:18 <hanjo> I use libusb also and it has worked fine
Jul 18 11:39:45 <DaKa2> a small patch was all that was needed
Jul 18 11:40:04 <hanjo> the only problem that I could not figure out is to swig -wrap the library for use in python
Jul 18 11:40:20 <jbowler-away> hanjo: a work round without the kernel fix is to mount -o remount,ro /dev/mtdblock4 /initrd
Jul 18 11:40:39 <jbowler-away> It seems the flash chip is left in write mode and this means the hardware reset hangs.
Jul 18 11:40:50 <hanjo> when I was using a PC infrastructure (insted of the slugs), we had libusb wrapped using swig
Jul 18 11:41:04 <hanjo> jbowler: that's all?
Jul 18 11:41:09 <jbowler-away> So putting the initrd into ro should be sufficient to ensure that the shutdown -r works.
Jul 18 11:41:43 <jbowler-away> hanjo: I think that's sufficient, but I haven't got any good way of verifying it.
Jul 18 11:41:49 <hanjo> jbowler, have you tested this on nfs-root systems also?
Jul 18 11:42:00 <jbowler-away> The kernel fix has been checked in for all the kernel builds.
Jul 18 11:42:16 <jbowler-away> hanjo: no, that's the next step, I've only verified for disk.
Jul 18 11:42:45 <hanjo> my slugs are on nfs, but I'll try this tomorrow when I get back to work
Jul 18 11:42:53 <jbowler-away> (And I now have no way of verifying the non-kernel work round - because all the kernels have the fix)
Jul 18 11:43:03 <hanjo> :)
Jul 18 11:43:34 <jbowler-away> I changed /etc/init.d/umountinitrd.sh to do the remount if the umount of /mnt fails (that's the case where initrd is mounted rw)
Jul 18 11:44:00 <jbowler-away> It's also possible to just do the remount by hand to test whether it is sufficient.
Jul 18 11:44:42 <jbowler-away> BTW there is a change to /boot/network too, which now uses the bootproto setting from /etc/default/sysconf.
Jul 18 11:45:04 <jbowler-away> It doesn't do dhcp if the setting is static, but I had previously changed the default (raw NSLU2 setup) to be dhcp.
Jul 18 11:45:05 <hanjo> so is dhcp on by default?
Jul 18 11:45:21 <jbowler-away> Yes, if you flash an out-of-the-box system.
Jul 18 11:45:39 <jbowler-away> But if you upgrade and you have an old /etc/default/sysconf you may encounter problems.
Jul 18 11:46:07 <jbowler-away> It is sufficient to change bootproto in /etc/default/sysconf to 'dhcp' if it is still 'static'.
Jul 18 11:46:19 <hanjo> my sysconf was "cleaned-up" using turnup -i so I don't expect much problems.
Jul 18 11:46:55 <jbowler-away> Well, turnup init will definately set it correctly, but I would check...  (It's the one in the flash file system which matters).
Jul 18 11:51:14 <hanjo> hmm
Jul 18 11:51:44 <jbowler-away> One important thing...  A shutdown -r on a system without the kernel fix after a reflash will probably hang.
Jul 18 11:52:05 <jbowler-away> That's because the reflash writes to the flash and the write is the last thing it does.
Jul 18 11:52:21 <hanjo> I did the fix on an ancient  2.6.11.2 with nfs root and it hang
Jul 18 11:53:18 <jbowler-away> Hum...  I was afraid of that.
Jul 18 11:56:42 <jbowler-away> What's more my nfs root system has booted but doesn't let me ssh in (the auth succeeds but then I get an immediate exit)
Jul 18 12:00:20 <[g2]> beewoolie-afk, hey! welcome !
Jul 18 12:01:36 <jbowler-away> My nfs boot problem is because passwd is 'preserve' in /etc/default/conffiles but the old root home has been removed.
Jul 18 12:01:59 <jbowler-away> That has to be fixed.
Jul 18 12:02:23 <[g2]> I was just noticing I was in /root too
Jul 18 12:03:23 <jbowler-away> A clean flash is fine.  A reflash is fine (because it doesn't zap the disk or nfs root), but after that point a turnup -i will produce an unbootable partition.
Jul 18 12:03:48 <jbowler-away> Oh, and the original flash partition isn't usable either...  Well, they boot, but login is impossible.
Jul 18 12:05:44 <jbowler-away> Slugbug 205
Jul 18 12:06:39 <hanjo> what is /etc/default/conffiles ?
Jul 18 12:07:09 <hanjo> Is that how you tell turnup what to keep in tmpfs ?
Jul 18 12:07:21 <jbowler-away> It's a list of files which need special attention on reflash - it says whether to preserve the one in the old image, overwrite, or consult the user (on reflash).
Jul 18 12:08:00 <jbowler-away> tmpfs (/var) handling is in /etc/default/volatiles
Jul 18 12:08:38 <hanjo> I see. I don't have those on this slug at home. It is pre openslug-beta.
Jul 18 12:09:15 <jbowler-away> There are quite a lot of changes, in fact iirc those too are post openslug-2.0
Jul 18 12:09:41 <jbowler-away> Anyway, the kernel fix seems to work on an nfs system - i.e. I just did a shutdown -r with an NFS root without any problem.
Jul 18 12:10:51 <jbowler-away> hanjo: I believe we will have an official new binary release soon, within a few days, so you might want to wait for that.
Jul 18 12:11:12 <hanjo> Yes, I will definitely wait for that for my production slugs
Jul 18 12:11:30 <hanjo> I have decided to keep them on the release + some small changes that I need
Jul 18 12:11:49 <jbowler-away> For the non-production ones it's worth upgrading now - then we can fix any problems before release.
Jul 18 12:11:53 <hanjo> but I will test the head with this one I have at home
Jul 18 12:12:26 <hanjo> what version of monotone are we using now?
Jul 18 12:13:49 <jbowler-away> 0.21
Jul 18 12:14:01 <jbowler-away> You must have either 0.20 or 0.21 to connect to the server.
Jul 18 12:14:16 <hanjo> ok. I need to upgrade. I have .18
Jul 18 12:14:53 <hanjo> I'll report back when I'm done with the build and have tried shutdown -r with a nfs root system. bbl
Jul 18 13:41:03 <dyoung> so should USB_OHCI_LITTLE_ENDIAN really be y and USB_OHCI_BIG_ENDIAN=n?  Or the reverse?
Jul 18 13:43:16 <DaKa2> isn't the nec usb chip little endian or something? (/me have no idea)
Jul 18 13:55:38 <hanjo> Is the http://www.nslu2-linux.org/Makefile still current? I get connection refused on monotone.vanille.de
Jul 18 13:58:42 <jbowler> That's something temporary, it has been happening for an hour or so.
Jul 18 13:59:14 <jbowler> It's irrelevant though - skip over it and the Makefile will get the data from nslu2-linux.org
Jul 18 14:01:54 <hanjo> I am trying to setup monotone environment for the first time on my home machine. What do you mean by "skip it over" ? make setup-openembedded fails because the nslu2 branch is empty
Jul 18 14:03:45 <hanjo> never mind, I'll try a bit later
Jul 18 14:04:02 <DaKa2> [g2]: Ive managed to setup a native packagebuild environ without compiling anything nativly first :-)
Jul 18 14:05:30 <DaKa2> except bitbake, if you would call that compiling
Jul 18 15:02:30 <beewoolie-afk> [g2]: Hey man.  Audio was off, so I didn't here the ping.
Jul 18 15:02:44 <beewoolie-afk> Someone posted a HOWTO for installing Debian on th eslug.
Jul 18 15:03:01 <kolla_> yes
Jul 18 15:03:49 <kolla_> jacmet at #nslu2-linux
Jul 18 15:26:40 <[g2]> beewoolie-afk, hey!
Jul 18 15:26:48 <beewoolie-afk> Hey man
Jul 18 15:26:59 <[g2]> where's the howto ?
Jul 18 15:27:15 <beewoolie-afk> http://peter.korsgaard.com/articles/debian-nslu2.php
Jul 18 15:27:32 <beewoolie-afk> Mind you, it's LE, so it requires a USB ethernet device.
Jul 18 15:29:33 <[g2]> is that Jacmet ?
Jul 18 15:29:38 <dyoung> Yes
Jul 18 15:29:48 <[g2]> yeah
Jul 18 15:30:10 <dyoung> So Beewoolie...
Jul 18 15:30:20 <[g2]> I'll go congratulate him in nslu2-linux
Jul 18 15:30:21 <dyoung> I see your pawprints in ohci.h .
Jul 18 15:31:17 <dyoung> I was wondering if we really are supposed to be CONFIG_USB_OHCI_LITTLE_ENDIAN=y
Jul 18 15:31:45 <beewoolie-afk> dyoung: let me see if I can find what you are looking at.
Jul 18 15:32:27 <[g2]> doesn't debian have a BE target ?
Jul 18 15:32:36 <beewoolie-afk> [g2]: Nope.
Jul 18 15:32:40 <dyoung> Our defconfig says CONFIG_USB_OHCI_LITTLE_ENDIAN=y and CONFIG_USB_OHCI_BIG_ENDIAN=n
Jul 18 15:32:43 <[g2]> luser's :)
Jul 18 15:32:54 <beewoolie-afk> If it did, we'd be able to pull their packages.
Jul 18 15:33:10 <beewoolie-afk> Well, give'em a break.  We could make a target if we wanted to.
Jul 18 15:33:32 <beewoolie-afk> dyoung: I don't know about this.  I don't think I changed it...on purpose at least.
Jul 18 15:33:56 <[g2]> reminds me I've not talked to spanky in a while
Jul 18 15:33:58 <beewoolie-afk> Seems like it wouldn't work at all if it were wrong.
Jul 18 15:34:23 <dyoung> Okay, I was just poking through our files and looking for things that JDLR.
Jul 18 15:34:29 <dyoung> It may well be right.
Jul 18 15:35:08 <beewoolie-afk> I don't see it in the ixp42x defconfig.
Jul 18 15:35:50 <dyoung> right, but the defautl ixp42x doesnt have a USB HCI either. :-)
Jul 18 15:36:17 <beewoolie-afk> CONFIG_USB_ARCH_HAS_OHCI=y... this is from the ixp42x config
Jul 18 15:36:59 <[g2]> wouldn't it be funny if that's where the irq26's were coming from :)
Jul 18 15:37:24 <beewoolie-afk> I thought we knew where the irq26's were coming from.
Jul 18 15:38:23 <[g2]> we pretty much do
Jul 18 15:38:49 <dyoung> Right, that HAS_OHCI is for the client I think.  Theres another section for USB HOST stuff.
Jul 18 15:39:06 <dyoung> where for nslu2, 
Jul 18 15:39:07 <dyoung> CONFIG_USB_OHCI_HCD=y
Jul 18 15:39:07 <dyoung> # CONFIG_USB_OHCI_BIG_ENDIAN is not set
Jul 18 15:39:07 <dyoung> CONFIG_USB_OHCI_LITTLE_ENDIAN=y
Jul 18 15:39:32 <beewoolie-afk> Well, you can try it the other way.  :-)
Jul 18 15:39:39 <dyoung> Yep, I could.
Jul 18 15:39:53 <dyoung> I was just hoping someone would have the answer ahead of time.
Jul 18 15:39:55 * dyoung lazy
Jul 18 15:47:18 <[g2]> jbowler, I make a fresh install (full reflash) of the new image with the /root change
Jul 18 15:47:47 <[g2]> since I did turnup with that image I'm fine because the /root in /etc/passwd matches the system
Jul 18 15:50:52 <jbowler> Correct, that scenario is no problem.
Jul 18 15:53:28 <[g2]> jbowler, ok thx.... I was going to cut over FatTurbo
Jul 18 15:54:04 <jbowler> dyoung: you can't change it, look in the Kconfig for drivers/usb - it's set that way because of the chips (if I understood it)
Jul 18 15:54:44 <dyoung> Okay, np.  It was JDLR for me, so just want to make sure iit really is right.
Jul 18 15:56:22 <jbowler> From my limited understanding it might be wrong if the chipset support is incorrect, but I suspect it probably gets ignored - after all the system supports both big and little endian.
Jul 18 16:35:37 <[g2]> jbowler, DUDE I've shutdown -r maybe 10 times so far with no problems!
Jul 18 16:35:41 <[g2]> you rock!
**** ENDING LOGGING AT Mon Jul 18 23:59:57 2005