**** BEGIN LOGGING AT Sun May 14 09:59:56 2006
May 14 13:00:54 <likewise> Is there a small rootfs that inits IXP eth0 and boots through a rootfs over NFS?
May 14 13:02:47 <EvilDevil> likewise, openslug can do that. flash it, and use turnup nfs
May 14 13:03:47 <likewise> OK, going to try that.
May 14 13:06:35 <EvilDevil> likewise, http://www.nslu2-linux.org/wiki/HowTo/LoadRootfsFromNFS
May 14 13:07:05 <likewise> yes found it now. I first need to flash OpenSlug - hold on.
May 14 13:11:40 <likewise> EvilDevil: my slug is still virgin - I have no password configured yet. Its seems I cannot get into the web interface to upload openslug.
May 14 13:12:19 <likewise> EvilDevil: never mind, default is admin it seems
May 14 13:12:36 <EvilDevil> okay ;) likewise 
May 14 13:13:56 <EvilDevil> i hope you don't mind loosing your webinterface if you're running openslug.
May 14 13:15:07 <EvilDevil> to reflash the stock linksys firmware again, you have to use this method http://www.nslu2-linux.org/wiki/HowTo/UseTheResetButtonToEnterUpgradeMode
May 14 13:16:58 <likewise> I can intercept RedBoot.
May 14 13:17:07 <likewise> or otherwise, I will do it using JTAG
May 14 13:17:32 <likewise> it's just that I have no serial cable connected, so I do most thing blindly.
May 14 13:20:49 <EvilDevil> the redboot partition will not be touched during flashing the device (unless you're using the linksys' erase-all utility), so, you're safe
May 14 13:33:32 <likewise> The argument to 'nfs' or 'disk' must be an empty partition
May 14 13:33:34 <likewise>   of sufficient size to hold the root file system
May 14 13:33:49 <likewise> OK, that's easy; I can start with an emty NFS root?
May 14 13:34:15 <EvilDevil> if the wiki says so ... yes
May 14 13:34:44 <EvilDevil>    1. Ensure you have an NFS server running with an empty directory exported with the options rw,no_root_squash.
May 14 13:34:44 <EvilDevil>    2. as described in turnup help, run turnup nfs -i <hostname>:<export_name>.
May 14 13:34:44 <EvilDevil>    3. shutdown -r now. It should come up in the new nfs area. 
May 14 13:40:45 <likewise> EvilDevil: Do you know how OpenSlug does this? As far as I know, the ethernet (IXP) module cannot be in the kernel, so it needs a first bootstrap rootfs first.
May 14 13:42:35 <EvilDevil> the nslu2 loads the init ramdisk from flash (+ kernel modules) and executes /linuxrc from the initrd. this executes /boot/nfs in your case
May 14 13:42:56 <likewise> EvilDevil: thanks
May 14 13:43:23 <likewise> I was about to make such a rootfs myself as I wrongly thought OpenSlug only did NFS server functionality, not client.
May 14 13:43:40 * likewise was wrong :-)
May 14 13:44:40 <EvilDevil> and you can still boot the nslu2 even if the nfs server is unavailable
May 14 13:51:14 <likewise> EvilDevil: Excellent. I am being amazed by OpenSlug every day :-)
May 14 13:51:53 <likewise> I am interested in bringing this initial ramdisk over to our own IXP4xx platform for development purposes.
May 14 14:09:59 <lcars8467> I have a samba3 problem on unslung 6.8 beta: cannot become connected user, I followed the guide on NSLU-linux but still getting this error
May 14 14:11:59 <EvilDevil> likewise, take a look at www.nslu2-linux.org/wiki/Development/MasterMakefile
May 14 17:56:33 <bofh> wtf.
May 14 17:56:44 <bofh> Okay, the Intel flash chip is a 64Mb one
May 14 17:56:57 <bofh> Then howcome only 8Mb are used?!
May 14 18:03:53 <dwery> Mb == megabit
May 14 18:04:04 <dwery> right?
May 14 18:10:20 <bofh> Megabyte, I think?
May 14 18:10:31 <bofh> Not sure, actuallt
May 14 18:10:33 <bofh> actually*
May 14 18:10:38 <bofh> Lemme reread the specs
May 14 18:11:26 <likewise> bofh: what platform?
May 14 18:11:57 <bofh> Yeah, the StrataFlash is 64Mbit
May 14 18:12:12 <likewise> chip specs are always in bits AFAIK
May 14 18:12:33 <bofh> likewise: any, I'm referring to the internal flash chip
May 14 18:13:22 <bofh> So therefore a 128MByte flash chip (my current hdd attached to the slug) is 1GBit?
May 14 18:13:41 <likewise> yes
May 14 18:13:46 <bofh> Hmm
May 14 18:14:01 <likewise> but are you sure it's only 1 chip?
May 14 18:14:03 <bofh> "The IXP420 has 24 address lines on the expansion bus and the flash also has 24 address lines. Unfortunately the 24th address line on the IXP420 is not connected to the flash at all. This limits us to a maximum of 128Mbit."
May 14 18:14:26 <bofh> So wouldn't it be possible to connect the 24th addr line to the new flash chip?
May 14 18:15:00 <likewise> yes, but if the line is not available on the PCB, you are quite lost, or need excellent drilling skills.
May 14 18:15:35 <likewise> and you probably would need chip select logic, otherwise aliasing occurs.
May 14 18:15:47 <bofh> hm...
May 14 18:22:27 <likewise> bofh: what do you want to achieve?
May 14 18:26:03 <likewise> OpenSlug Makefile: Read 0x00103344 bytes from "vmlinuz" into <Kernel> (9 blocks / 8 blocks) --> the kernel has become toooooo big :-/
May 14 18:26:21 <likewise> padBytes error: data (1061716) is longer than total_len (1048576) at /home/leon/sandbox/openslug/openslug/tmp/staging/x86_64-linux/bin/slugimage line 58.
May 14 18:28:35 <bofh> swap internal flash from the 8MByte to a 128MByte one
May 14 18:28:58 <bofh> (not anytime soon, mind you, 263 day uptime on the slug and I want to first get a year on it)
May 14 18:29:19 <likewise> bofh: what distro are you running then?
May 14 18:31:01 <bofh> openslug
May 14 18:31:15 <bofh> (openembedded would be the distro, then?)
May 14 18:31:24 <likewise> probably yes
May 14 18:50:25 <Caramba_> Hi folks. I've run into something that looks kinda serious. My Unslung slug  (V2.3R29-uNSLUng-5.5-beta) quit responding to all outside attempts to contact it (except for the webinterface). I managed to login using the telnet-backdoor, and found that all the software I've installed is gone. When I do "ls -l" in the top-directory (/), I get a segmentation fault. I suspect that the slug doesn't...
May 14 18:50:26 <Caramba_> ...read my external hard drive on power up, but...it allows me to browse the stuff on it from the web interface. How should I start troubleshooting this situation? Thx for any help!
May 14 18:56:29 <marceln> Caramba_: ?
May 14 18:56:38 <Caramba_> Yep, marcel
May 14 18:57:00 <marceln> What happend before this?
May 14 18:57:41 <marceln> Is it possible that someone pulled the usb cable?
May 14 18:57:49 <marceln> What is your setup?
May 14 18:57:59 <Caramba_> Hmm...I don't know. I was away for the weekend, when I came back and tried to contact it with my Squeezebox (SlimServer) I didn't get a response. I started to investigate, and noticed that all contact (ssh and so on) failed.
May 14 18:58:25 <Caramba_> What particular setup are you refering to? And no, nobody should have been able to pull the cable. 
May 14 18:59:00 <marceln> How did you connect your disks?
May 14 18:59:14 <Caramba_> Only one disk, plugged into port 1.
May 14 18:59:31 <marceln> Is that disk stil running?
May 14 19:00:00 <Caramba_> Yep. And the slug reads it through the web interface! I can download stuff from it, but I can't find it from the telnet prompt.
May 14 19:00:14 <Caramba_> And obviously the slug doesn't use the disk in the boot process.
May 14 19:03:03 <marceln> Did you check the kernel messages (dmesg)?
May 14 19:03:46 <Caramba_> nope...scuse a slightly green linux-user, where do i find it?
May 14 19:04:31 <marceln> Just type "dmesg"
May 14 19:04:50 <Caramba_> well, typing the command gives me a long list of "!!!abort_handler addr = 0x40015000, isr = 0x3, status = 0x2a0". SOme variations on the theme, but this is basically it.
May 14 19:08:49 <Caramba_> Should I worry?
May 14 19:09:23 <marceln> There are a lot of references to  abort_handler on google.
May 14 19:09:35 <marceln> I just try to see the interesting ones
May 14 19:21:20 <marceln> I can't find specific information.
May 14 19:22:04 <marceln> Did you try to reset the slug?
May 14 19:22:56 <Caramba_> Yep - rebooted it quite a few times.
May 14 19:23:49 <marceln> Do you have a linux system?
May 14 19:24:06 <Caramba_> The situation is *so* odd. The biggest paradox is the fact that the web interface allows me to browse the disk, but I can't see it from the prompt.
May 14 19:24:11 <marceln> In that case i would run a "fsck" on that disk from the linux system.
May 14 19:24:29 <Caramba_> And check the integrity of the file system, you mean?
May 14 19:24:34 <marceln> Yes
May 14 19:24:50 <Caramba_> OK. Can I do it from a Knoppix boot disk?
May 14 19:25:04 <marceln> No problem
May 14 19:25:40 <Caramba_> Guess I'll do that, then. I'm rather worried, I've gone to a lot of trouble to rip all my CD:s and put them on the disk
May 14 19:26:23 <marceln> Maybe you should look at "http://www.nslu2-linux.org/wiki/HowTo/TroubleshootADriveMissingCondition"
May 14 19:26:50 <Caramba_> I'll read that one, thx. Maybe I can contribute something if I solve the problem!
May 14 19:29:35 <Caramba_> while i'm asking: I've tried to do a Scandisk a few times, but the log always tells me that the disk is in use, and the scan cannot be done. Is this because of unslung using  the disk all the time?
May 14 19:38:02 <caplink811-log> yes, you couldn't run a fsck on a mounted disk
May 14 20:16:29 <likewise> Read 0x001014F8 bytes from "vmlinuz" into <Kernel> (9 blocks / 8 blocks)
May 14 20:16:42 <likewise> darn, my openembedded kernel is still 14f8 bytes too big...
May 14 20:17:56 <EvilDevil> likewise, try ucslugc ;)
May 14 20:18:20 <likewise> evildevil: doesn't make a diff for the kernel
May 14 20:18:39 <likewise> or is it a non-MMU kernel on the IXP/
May 14 20:18:40 <likewise> ?
May 14 20:21:00 <likewise> what RTC chip is in the NSLU?
May 14 20:21:05 <EvilDevil> likewise, it has less kernel options enabled by default
May 14 20:24:09 <EvilDevil> likewise, how do i find out what rtc is in the nslu?
May 14 20:24:27 <likewise> I can open it. Never mind, I think it is a X1205
May 14 20:26:11 <likewise> http://www.nslu2-linux.org/wiki/Info/RTCChipInfo
May 14 20:26:19 <likewise> two chips possible (rumor)
May 14 20:28:20 <EvilDevil> CONFIG_RTC_DRV_X1205=y
May 14 20:28:20 <EvilDevil> CONFIG_RTC_DRV_DS1672=y
May 14 20:28:20 <EvilDevil> CONFIG_RTC_DRV_PCF8563=y
May 14 20:28:20 <EvilDevil> CONFIG_RTC_DRV_RS5C372=y
May 14 20:28:33 <EvilDevil> (from the debianslug .config)
May 14 20:28:56 <EvilDevil> this catches all ;)
May 14 21:08:25 <CIA-5> 03marceln * 10unslung/ (make/xinetd.mk sources/xinetd/postinst): Upstream update that fixes xinetd hang after tftp
May 14 21:18:19 <mwester-treo> marceln: thanks for getting xinetd fixed - that bug has been around for a while!
May 14 21:23:00 <marceln> It was a lucky shot.
May 14 21:23:14 <marceln> I started with an upstream upgrade and it worked.
May 14 22:19:40 <likewise> How does openslug select the package versions of openembedded to be built?
May 14 22:39:21 <EvilDevil> likewise, mompls
May 14 22:40:45 <likewise> EvilDevil: sorry, I am being off-topic here I guess.
May 14 22:41:12 <EvilDevil> ~/slug/openembedded/conf/distro/slugos.conf (PREFERRED_VERSION_package-name="x.xx")
May 14 22:45:15 <likewise> EvilDevil: I saw that, but the non-listed packages are always promoted to use the highest version available?
May 14 22:50:49 <likewise> thanks EvilDevil for your time. I am off to bed now.
**** ENDING LOGGING AT Mon May 15 09:59:56 2006