**** BEGIN LOGGING AT Sat Apr 30 23:59:57 2005
May 01 01:17:24 <rwhitby> well, that's over.
May 01 02:01:56 <rwhitby> glc: any updates on that linuxrc you sent?
May 01 02:02:18 <rwhitby> isn't /mnt/newroot always there in Unslung firmware?  Why does it need to be created?
May 01 03:45:25 <rwhitby> glc: I have to rebuild my unslung image, so I won't have a chance to test out that new linuxrc for a couple of hours at least.
May 01 04:43:02 <rwhitby> k_velikov-away: did you get MRTG to compile?  Are you going to contribute it to the Unslung CVS repository ?
May 01 05:06:47 <dyoung> morning
May 01 05:07:00 <rwhitby> g'day
May 01 05:07:12 <pTweety> morning
May 01 06:46:35 <glc> rwhitby: newroot was in ramdisk but not jffs2. Only takes 1 line to make sure it's there.
May 01 06:47:17 <dyoung> we should probably add it to the unslung image directly.
May 01 06:47:24 <glc> I'm happily running nfsroot rightnow. 
May 01 06:47:30 <dyoung> cool!
May 01 06:47:36 <dyoung> Are the instructions still valid?
May 01 06:48:05 <rwhitby> or use an existing /mnt point which is in jffs2
May 01 06:48:15 <glc> Yes dyoung, but it was much easier for me to add it to the linuxrc and not have to wait 2 hours for a rebuild.
May 01 06:48:42 <glc> rwhitby: That too.
May 01 06:49:27 <glc> I was looking for minimum changes.
May 01 06:49:28 <rwhitby> I think /mnt/tmpmnt will do
May 01 06:52:14 <glc> dyoung: I only needed to include the 2 variables in .nfsroot.  All the other junk is in the linuxrc now.
May 01 06:55:27 <glc> It works very well as long as you mount nolock. If you want locking you need to include portmap.
May 01 06:55:42 <glc> Even then locking doesn't work well.
May 01 06:57:45 <glc> Initrd still can't be umounted, but it doesn't matter now, as it doesn't consume ram.
May 01 06:58:52 <glc> The kernel is running processes keeping /initrd/dev/console open.
May 01 06:59:06 <dyoung> wait a minute.
May 01 06:59:32 <dyoung> Oh, right, youre doing nfs.
May 01 06:59:57 <dyoung> my brain is all hash already.
May 01 07:00:17 <glc> Wondering how I could try to umount /initrd?
May 01 07:00:32 <rwhitby> glc: pushing new version now, please review to make sure I have included functionally equivalent changes.
May 01 07:00:59 <dyoung> is the initrd ro or rw?
May 01 07:01:10 <rwhitby> consistency check takes so long ....
May 01 07:01:52 <glc> O.K. You have the fallback to /bin/init on the tail?
May 01 07:03:33 <CIA-8> 03rwhitby 07 * r1.3199.1.1 10openembedded/packages/nslu2-binary-only/unslung-rootfs-2.3r25/linuxrc: Added glc's fixes to Unslung linuxrc
May 01 07:03:54 <rwhitby> yep
May 01 07:04:03 <glc> It is rw.
May 01 07:04:26 <dyoung> thats either a bug or a feature.
May 01 07:05:01 <dyoung> on one hand its nice to have the jffs2 available post pivot; but on the other it  may be tempting for Joe Guy to screw with it.
May 01 07:05:12 <glc> Yes. A feature if you want to edit. A bug if you want to protect.
May 01 07:05:30 <glc> Let me try quicky to remount ro before the pivot.
May 01 07:05:36 <rwhitby> Joe Guy won't know what "initrd" is, so won't be tempted.
May 01 07:06:27 <rwhitby> glc: I would have thought we would have to mount ro in kernel args, and then remount rw in /etc/rc or something ...
May 01 07:08:37 <glc> It remounts just fine after the chroot, so it can be added to one of the rc.xxxx scripts if desired.
May 01 07:10:00 <glc> Pulling.
May 01 07:10:35 <dyoung> morning [g2]
May 01 07:12:48 <rwhitby> [g2]: How's OpenRevvy going?
May 01 07:23:37 <[g2]> morning!
May 01 07:24:29 <[g2]> rwhitby, I've still got serial issues, but it's not really need until I do something drastic
May 01 07:24:52 <[g2]> I had switchbox 4.11 executing in a chroot environ night before last
May 01 07:25:01 <rwhitby> not able to discern the pinout?
May 01 07:25:26 <[g2]> The grounds are all clear
May 01 07:25:33 <[g2]> I think the VCCs are all clear
May 01 07:26:34 <[g2]> I've tried probing for the other lines and I'm not sure if it's my soldering or the pins
May 01 07:27:36 <[g2]> while searching the web, I ran across the network backdoors to the bootloader and telnet (and subsequent shell access) so I've been in and all over the local fs
May 01 07:28:29 <[g2]> rv042 is 8MB flash, 32MB mem @ 266 on IXP425
May 01 07:28:48 <[g2]> rv082 is 16MB flash, 32MB mem @ 533 on IXP525
May 01 07:28:51 <[g2]> 425
May 01 07:45:40 <glc> rwhitby: what stamp and work directory do I need to kill to rebuild without doing it from scratch?
May 01 07:50:15 <rwhitby> unslung-rootfs-*
May 01 07:54:03 <glc> Thanks.
May 01 07:56:55 <[g2]> rwhitby, have you ever played with nbd by chance ?
May 01 07:57:15 <rwhitby> no, but the guys on mpeg-playcenter.com have it running on Unslung
May 01 07:57:42 <[g2]> THX... I'll have to check that out
May 01 07:57:59 <[g2]> I tried mounting an NFS swap the other day and I couldn't get it to work
May 01 07:58:15 <[g2]> Someone said nbd was the way to go for mounting swap
May 01 07:59:08 <[g2]> A fattened rv082 (to 128 or 256MB) and nbd would make a excellent native compile environ
May 01 08:11:37 <glc>     /sbin/ifconfig lo up 127.0.0.1
May 01 08:11:38 <glc>     /sbin/ifup
May 01 08:12:03 <glc> rwhitby: need ifup for nfsroot.
May 01 08:12:38 <rwhitby> just before route add ?
May 01 08:12:59 <rwhitby> oh, sorry - just saw it
May 01 08:13:07 <rwhitby> Line 92, right?
May 01 08:14:22 <ka6sox-away> morning
May 01 08:19:16 <glc> Perhaps it should be:
May 01 08:19:26 <glc>     /sbin/ifconfig ixp0 up $ipaddr netmask $netmask
May 01 08:19:27 <glc>     /sbin/ifconfig lo up 127.0.0.1
May 01 08:19:27 <glc>     /sbin/ifup
May 01 08:19:27 <glc>     /sbin/route add default gw $gateway
May 01 08:20:29 <rwhitby> too late - I just pushed the other one
May 01 08:20:37 <CIA-8> 03rwhitby 07 * r1.3202 10openembedded/packages/nslu2-binary-only/unslung-rootfs-2.3r25/linuxrc: Added glc's ifup fix to Unslung linuxrc
May 01 08:20:54 <rwhitby> why is the ifup required if the other two have up already?
May 01 08:23:02 <glc> I don't know, but without ifup, the interface is not up.
May 01 08:23:28 <rwhitby> what does the "up" parameter in ifconfig do then ... ?
May 01 08:24:49 <CIA-8> 03ptweety * 10unslung/make/cyrus-imapd.mk: removed strip for /opt/bin/* in cyrus-imapd.mk
May 01 08:25:12 <glc> Let me check. It may be a routing problem.
May 01 08:32:27 <glc> Yes, no routes.
May 01 08:39:28 <glc> Oops ixp0 has no ip address.
May 01 08:46:35 <glc> #    /sbin/ifconfig ixp0 up $ipaddr netmask $netmask
May 01 08:46:36 <glc> #    /sbin/ifconfig lo up 127.0.0.1
May 01 08:46:36 <glc> #    /sbin/route add default gw $gateway
May 01 08:46:36 <glc>     /sbin/ifup
May 01 08:47:10 <glc> Seems to work just fine.  Let me try with dhcp instead of static.
May 01 08:48:10 <jp30> pTweety, ping?
May 01 08:48:27 <pTweety> jp30, pong
May 01 08:48:39 <jp30> did you get my email about cyrus-imapd?
May 01 08:48:40 <pTweety> I'm right now answering you mail
May 01 08:48:43 <pTweety> :-)
May 01 08:48:49 <jp30> great
May 01 08:49:59 <pTweety> jp30, mail is sent
May 01 08:55:00 <jp30> pTweety, thanks...
May 01 08:56:48 <pTweety> np
May 01 09:40:18 <glc> dhcp doesn't seem to work, so ifup may not be too good an idea. Still checking.
May 01 10:07:42 <glc> Hmm. reset sysconf and reflashed with stock firmware, and dhcp doesn't work there either.
May 01 10:36:06 <CIA-8> 03jp30 * 10unslung/make/man.mk: man depends on less
May 01 10:41:17 <jp30> jeanfabrice_, i trust you've had no trouble with the new thread-safe php?
May 01 10:41:38 <glc> rwhitby: You are right. We do not need ifup. We also don't need the space after "ipaddr="
May 01 10:47:42 <jp30> rwhitby: i noticed that you removed the part of libxml2.mk that stages xml2-config into staging/bin
May 01 10:48:41 <jp30> ...if you're here, we should have a chat about the reason it was there
May 01 10:48:49 <NAiL> Anyone know which USB->Parallel adapters work with linux? I'm gonna stick one inside the slug so I get a lpt-port for my LCD-display :)
May 01 10:54:01 <CIA-8> 03jp30 * 10unslung/make/libxslt.mk: build libxslt without having to have xml2-config in the path
May 01 11:07:48 <glc> rwhitby: exit 0 should never be reached. If it is, kernel will panic. /bin/sh would be a better safetynet for those with serial ports.
May 01 11:53:31 <caplink811-away> jeanfabrice_: for the logs, tcl builds fine native, but there's a problem building the ipkg, so don't try it on a prductiv slug, the installer installs at /opt/ and not at TCL_IPK_DIR
May 01 12:45:22 <jeanfabrice_> caplink811-away: I buid tcl cross
May 01 12:46:25 <caplink811-away> jeanfabrice: are you able to build an ipk paket?
May 01 12:47:29 <jeanfabrice> let me see
May 01 12:47:40 <jeanfabrice> humm.. yep
May 01 12:48:26 <jeanfabrice> but I needed to patch too files
May 01 12:48:51 <caplink811-away> one is build_dir/unix/Makefile, and the other=
May 01 12:49:00 <caplink811-away> s/=/?
May 01 12:49:43 <jeanfabrice> one is configure.in and the other is strstr.c as it lacks #define NULL 0
May 01 12:50:01 <caplink811-away> ah, right, have seen this, if i'vr tried cross
May 01 12:50:39 <caplink811-away> in configure .in you've patch install_root?
May 01 12:50:49 <jeanfabrice> I used make  INSTALL_ROOT=$(TCL_IPK_DIR) install
May 01 12:51:18 <caplink811-away> aha, let my try
May 01 12:51:54 <jeanfabrice> But I got a compilation error (twice defined function), so that's the patch I made to configure.in
May 01 12:52:48 <jeanfabrice> Can you check during ad configure time what does it says about "checking for Solaris2.4/Tru64 strtod bugs""
May 01 12:53:24 <caplink811-away> native or cross?
May 01 12:53:28 <jeanfabrice> cross
May 01 12:53:51 <caplink811-away> will try it tomorrow
May 01 12:55:53 <jeanfabrice> coz for me it says "buggy" and then configure makefile to compile and link against fixstrtod.c. The problem is that it still link against strtod.c making the same function defined twice
May 01 13:20:17 <CIA-8> 03caplink811 * 10unslung/make/tcl.mk: compiles now native, and building an ipk packet works fine
May 01 13:22:14 <CIA-8> 03caplink811 * 10unslung/Makefile: added tcl to the native_ready_for_testing section
May 01 13:25:41 <jeanfabrice> caplink811-away: ping
May 01 13:26:51 <caplink811-away> jeanfabrice: pong
May 01 13:27:16 <jeanfabrice> caplink811-away: can't tcl be cross compiled ?
May 01 13:28:23 <caplink811-away> dunno, haven_t spend much time on this, found the problem with undefined NULL at afternoon, but haven't tried more, cause it was a fine sunny day ;-)
May 01 13:28:44 <jeanfabrice> okay
May 01 13:33:43 <jeanfabrice> As I made some change on it locally to have it compiled cross, could you tell me change change you made ?
May 01 13:35:00 <jeanfabrice> anyway. just a question : did you make a symbolic link of /opt/lib/libtcl8.4.so to /opt/lib/litcl.so ?
May 01 13:35:14 <jeanfabrice> because make install don't create taht link
May 01 14:49:24 <rwhitby> jp30-work: ping
May 01 14:52:18 <rwhitby> jeanfabrice: we prefer cross compiles over native compiles, so if you are able to compile tcl cross, that is great.
May 01 14:53:12 <rwhitby> glc: checking in new linuxrc now ...
May 01 14:54:31 <jeanfabrice> does it make sense creating a link from a libxxx.<version>.so to libxxx.so ?
May 01 14:57:54 <rwhitby> Is that links normally there on a desktop system?
May 01 14:58:04 <jeanfabrice> dunno
May 01 14:58:22 <jeanfabrice> I would prefer discuss tcl.mk with caplink as I create some patches
May 01 14:58:59 <rwhitby> indeed.  I hope you two can work out together how to cross compile tcl
May 01 14:59:41 <jeanfabrice> np
May 01 15:00:12 <dyoung> usually you need to do that.  because when something else comes along, it cant just do -ltcl instead of -ltcl8.4
May 01 15:01:17 <jeanfabrice> dyoung: agreed. I was wondering why make install in tcl doesn't create this link
May 01 15:01:59 <CIA-8> 03rwhitby 07 * r1.3203 10openembedded/packages/nslu2-binary-only/unslung-rootfs-2.3r25/linuxrc: Fixed the fall-back, and removed the ifup in Unslung linuxrc
May 01 15:48:19 <jeanfabrice> 'nite all
May 01 15:56:34 <glc> rwhitby: Pulling.
May 01 15:56:42 <ka6sox-away> hi
May 01 15:57:16 <ka6sox-away> anything I should know about?
May 01 17:31:21 <Christopher-bbl> anyone around?
May 01 17:32:48 <jacques> yah
May 01 17:34:04 <Christopher> my /dev/sdb1 always comes up read only when I reboot
May 01 17:34:11 <Christopher> where would I go to fix this?
May 01 17:35:05 <jacques> hmm, anything in dmesg that tells why?
May 01 17:35:17 <jacques> I'm assuming it's not specified ro in /etc/fstab
May 01 17:36:27 <Christopher> no
May 01 17:36:38 <Christopher> it seems to be performing some sort of check on the drive
May 01 17:36:51 <Christopher> it deletes several unreference inodes
May 01 17:36:59 <Christopher> unreferenced inodes
May 01 17:37:22 <jacques> it dies this every time it boots?
May 01 17:37:28 <Christopher> it eventually remounts itself rw
May 01 17:37:29 <jacques> can you umount and fsck ?
May 01 17:37:31 <jacques> ah
May 01 17:37:42 <Christopher> but then I have to start all the processes with logs manually
May 01 17:37:52 <Christopher> because they can't write to their logs so they don't start
May 01 17:38:23 <jacques> is this a recent unslung ?
May 01 17:38:27 <Christopher> 3.18
May 01 17:38:53 <jacques> is sdb a hard drive or flash ?
May 01 17:38:57 <Christopher> hd
May 01 17:39:20 <jacques> hmm. I would really try fscking /dev/sdb1 (after umounting it)
May 01 17:39:33 <Christopher> ok
May 01 17:39:34 <jacques> or boot without it attached and then attach and fsck
May 01 17:39:41 <jacques> since it may not umount
May 01 17:40:07 <Christopher> yeah, it didn't
May 01 17:40:17 <Christopher> i'll try that after dinner and see what happens
May 01 17:40:24 <Christopher> bbl, thanks
May 01 17:40:31 <jacques> good luck
May 01 17:48:10 <dyoung> Christopher, its probably running the quotas when its ro.
May 01 17:48:41 <dyoung> and may be getting stuffed up because it needs fscking.
May 01 18:14:16 <Christopher> fscking...
May 01 20:27:02 <glc> rwhitby-away: Looks good!
May 01 21:10:33 <CIA-8> 03bzhou * 10unslung/ (make/libtiff.mk sources/libtiff/control): upstream upgrade from 3.7.1 to 3.7.2; generates control file
**** ENDING LOGGING AT Sun May 01 23:59:56 2005