**** BEGIN LOGGING AT Sat Jun 04 23:59:56 2005
Jun 05 00:16:17 <ka6sox-away> k
Jun 05 02:02:23 <rwhitby-away> Now I get a too many open files error when compiling 3.4.3 - anyone know how to increase the max number of open files in ubuntu ?
Jun 05 02:04:18 <rwhitby-away> Robert Hammond makes a good argument on nslu2-linux for making it very hard to run ipkg when you are not unslung.  I think we'll do that too.
Jun 05 02:04:31 <jacques> rwhitby-away, hmmm, that's the error I got trying to build 3.4.4 - I wonder what's going on
Jun 05 02:05:03 <jacques> I'm not sure increasing the max open files will fix it
Jun 05 02:05:10 <jacques> it's probably some type of runaway
Jun 05 02:07:59 <rwhitby-away> nod
Jun 05 02:11:43 <rwhitby-away> And I'd like to add dropbear to the next version of Unslung, and enable root login by ssh by default with the uNSLUng password.  I've had enough of people not being able to follow instructions, so Unslung will have to dumb itself down to the lowest common denominator (which seems to be only slightly more intelligent than the creature our beloved NSLU2 is named after).  We will move to R29, remove 
Jun 05 02:11:44 <rwhitby-away> telnetd altogether, and enable SSH with a hard-coded well-known password.
Jun 05 02:14:29 <dyoung> Good Plan.
Jun 05 02:16:37 <dyoung> Of course going that route will lead to tons of "wheres telnet" questions to replace the ones on the list now
Jun 05 02:18:10 <jacques> yeah, "my windows doesn't have a ssh command"?
Jun 05 02:21:28 <rwhitby-away> maybe a custom telnetd which prints an appropriate "use ssh, stupid" echo response.
Jun 05 02:30:29 <dyoung> I never thought it would reach this point...
Jun 05 02:30:48 <dyoung> (the one where we *had* to stupidify the firmware.
Jun 05 02:30:50 <dyoung> )
Jun 05 02:34:39 <rwhitby-away> it's another 3 C's theorem
Jun 05 02:35:47 <rwhitby-away> Commoditisation Causes Cro-Magnumness
Jun 05 02:36:27 <rwhitby-away> The intelligence assumed by a firmware distribution is inversely proportional to the number of users.
Jun 05 02:37:08 <rwhitby-away> Or maybe we just turn on telnetd permanently with no password.
Jun 05 02:37:39 <rwhitby-away> Lock down OpenSlug and open up Unslung.
Jun 05 02:43:04 <jacques> heh
Jun 05 02:46:36 <dyoung> Thatas right, I forgot about the rwhitby 3 C's.
Jun 05 02:47:53 <dyoung> "the nslu2-linux development group takes no responsibility if you elect to connect the device to the open internet without adequate security precautions."
Jun 05 02:49:25 <jacques> one step close to my slug army we will be
Jun 05 02:50:05 <dyoung> heh heh
Jun 05 02:50:23 <dyoung> what nefarious purpose shall you use the 'army for?
Jun 05 02:54:29 <jacques> it needs a purpose ?
Jun 05 02:56:59 <rwhitby-away> jacques is going to get them all to send him paypal details so he can buy some new hiking boots ....
Jun 05 02:58:44 <jacques> yeah, the heels are totally worn out on these
Jun 05 03:07:29 <dyoung> Hmm, I should contract jacques to have some contributions for dyoung's shoe replacement too.
Jun 05 03:08:00 <dyoung> its cracked in the sole so puddles come pouring in.
Jun 05 03:36:00 <rwhitby-away> of course I can't even get unslung to compile at the moment, so no way to make all these changes ...
Jun 05 03:57:11 <rwhitby-away> Hmm - libpthread-0.9.so contains "GROUP ( libpthread.so.0 libpthread_nonshared.a )" - I don't think that's right
Jun 05 03:57:31 <rwhitby-away>  Too many open files
Jun 05 03:57:55 <rwhitby-away> "/home/unslung/tmp/cross/armeb-linux/bin/ld: cannot open linker script file /home/unslung/tmp/cross/armeb-linux/lib/libpthread.so.0:  Too many open files"
Jun 05 03:58:18 <jacques> rwhitby-away, yeah I'm talking to pb about it right now in #oe
Jun 05 04:08:35 <rwhitby> ok, pull now to get pb_'s fix
Jun 05 04:09:58 <jacques> rwhitby, bear in mind this might be unrelated to the compile error, it's just something he said "oops" about - he recently enabled javac for all archs other than arm and forgot about armeb
Jun 05 04:10:44 <jacques> I noticed because the compile error is when trying to link libgcjgc.so.1.0.1
Jun 05 04:11:41 <rwhitby> yeah, that's why I asked whether we knew if that was the cause
Jun 05 04:12:16 <jacques> ok, I just wanted that to be clear :-)
Jun 05 04:12:50 <rwhitby> ah right, for those who weren't following along in #oe
Jun 05 04:13:06 <jacques> yep
Jun 05 04:13:44 <jacques> I'm trying anyway - just removed the gcc-cross stamps and work dir
Jun 05 04:13:57 <rwhitby> yeah. me too
Jun 05 04:14:38 <jacques> but I really think it's some loop like what you found
Jun 05 04:15:02 <jacques> that's almost always the cause of too many open files, non-trivial symlink loops
Jun 05 04:15:22 <rwhitby> yeah, but  if nothing else in gcc uses pthreads, then it won't bit us
Jun 05 04:15:27 <rwhitby> s/bit/bite/
Jun 05 04:15:35 <jacques> ah lol I didn't even think of that
Jun 05 04:16:09 <rwhitby> "and the doctor replied: Don't do that"
Jun 05 04:17:24 <jacques> makes perfect sense - the only "big" change in the relavant parts of oe lately that pb could think of was the enabling of javac
Jun 05 04:18:22 <rwhitby> dyoung: at least the "where's telnet" questions can be easily answered by an FAQ, and we can make it a banning offence to ask that question ....
Jun 05 04:19:27 <rwhitby> jacques: you testing 3.4.3 or 3.4.4 ?
Jun 05 04:19:51 <dyoung> So this sounds like Unslung-5.x now eh
Jun 05 04:20:00 <jacques> rwhitby, 3.4.4
Jun 05 04:20:26 <rwhitby> dyoung: yeah, Unslung 5.x "Dumbed down for the masses"
Jun 05 04:20:31 <dyoung> Yeah.
Jun 05 04:20:40 <rwhitby> C'est la vie
Jun 05 04:21:04 <dyoung> Thats okay, if youre "hardcore" enough, you will already have serial.
Jun 05 04:22:44 <jacques> rwhitby, it got past gcc-cross
Jun 05 04:23:15 <jacques> still going, it's on unslung-kernel now
Jun 05 04:23:26 <rwhitby> looks promising then
Jun 05 04:23:32 <jacques> yep
Jun 05 04:24:37 <dyoung> btw, anyone mind a change to unslung-kernel_2.3r25.bb to have the CMDLINE_ROOT from a = to ?= construct?
Jun 05 04:25:23 <rwhitby> good plan
Jun 05 04:25:31 <jacques> cool
Jun 05 04:25:37 <rwhitby> we need a 2.3r29.bb too
Jun 05 04:25:48 <rwhitby> I put the tar file up on nslu.sf.net a long time ago
Jun 05 04:26:03 <jacques> what does r29 get us?
Jun 05 04:26:08 <rwhitby> we just need to add the kernel change
Jun 05 04:26:18 <dyoung> I want to be able to test fatslug configurations for unslung too.
Jun 05 04:26:21 <rwhitby> jacques: a new version number on the web pages
Jun 05 04:26:33 <rwhitby> and any mods that Linksys has done to it's binaries
Jun 05 04:26:37 <dyoung> The only thing about r29 which I oppose is the crappy Genesys patch.
Jun 05 04:26:56 <rwhitby> How about the 4000 -> 2000 interrupt rate patch?
Jun 05 04:27:05 <rwhitby> (I agree about the Genesys patch - we do that better)
Jun 05 04:27:56 <dyoung> I think the 4k vs 2k patch warrants furthur investigation.
Jun 05 04:28:09 <rwhitby> dyoung: ok, the next version is 5.1 - let's get started on it ....
Jun 05 04:29:22 <dyoung> I mean if we're getting to that level of hacking and modding, should we just patch it to not poll that way, and be interrupt driven like how we patche dopenslung?
Jun 05 04:29:43 <rwhitby> yeah - we can do that now that we have the click-through
Jun 05 04:32:09 <jacques> NOTE: build 200506050412: completed
Jun 05 04:32:59 <rwhitby> cool - will you let pb_ know?
Jun 05 04:33:09 <CIA-8> 03dereky 07 * r1.3422 10openembedded/packages/linux/unslung-kernel_2.3r25.bb: 
Jun 05 04:33:09 <CIA-8> unslung-kernel_2.3r25.bb:
Jun 05 04:33:09 <CIA-8>  change cmdline_root from = to ?= for ease of overriding.
Jun 05 04:34:35 <dyoung> is the nslu2-linux.bkbits.net repo current now?
Jun 05 04:34:41 <jacques> rwhitby, will do
Jun 05 04:34:50 <rwhitby> dyoung: yep
Jun 05 04:35:10 <dyoung> ok, I can fire off a build before going to sleep.  any preferences?
Jun 05 04:35:22 <rwhitby> dyoung: I'm creating r29 versions of stuff now ...
Jun 05 04:36:27 <dyoung> Okay, I'll build unslung for now.
Jun 05 04:40:12 <copperbeech> ~jbot, seen jp30
Jun 05 04:41:24 <dyoung> jp30 is on holiday.
Jun 05 04:41:42 <dyoung> oop, not holiday.  On travel.
Jun 05 04:41:57 <dyoung> and may be unavailable for a while.
Jun 05 04:42:22 <copperbeech> I think I'm destines to keep missing him :-)
Jun 05 04:42:30 <copperbeech> destines = destined
Jun 05 04:42:47 <jbot> jp30 <~jp30@jp30.nslu2-linux> was last seen on IRC in channel #nslu2-linux, 13h 47m 48s ago, saying: 'hi, copperbeech'.
Jun 05 04:43:00 * copperbeech chuckles
Jun 05 04:43:04 <rwhitby> ooh - that's spooky
Jun 05 04:43:44 <dyoung> Moment of truth.  gcc-cross....
Jun 05 04:44:01 * rwhitby learns never to Ctrl-C bk when it's bk mv'ing a directory of files, and starts a new bk clone of openembedded as a result ...
Jun 05 04:44:12 <jacques> :-\
Jun 05 04:44:54 <dyoung> I think I'm gonna have to shutdown prodslug-2
Jun 05 04:45:01 <dyoung> ts drive is driving me batty.
Jun 05 04:45:08 <dyoung> ^I
Jun 05 04:46:01 <dyoung> Im not sure whats worse.  The whine of the drive, or the whine of tinnitus (wearing earplugs)
Jun 05 04:47:53 <jacques> what did you do to your ears?
Jun 05 04:49:17 <dyoung> Hmm... I'm not sure what broke the camels back.
Jun 05 04:49:30 <dyoung> it was probably working in the commercial audio industry some years ago though.
Jun 05 04:49:40 <dyoung> Big Bass.
Jun 05 04:49:44 <dyoung> Boom Boom Boom.
Jun 05 04:50:11 <dyoung> Yo.   This right here is not a a test.  Yo.
Jun 05 04:50:58 <jacques> :-(
Jun 05 04:51:04 <dyoung> Either that, or the constant hum of A/C at the office...
Jun 05 04:51:09 <ew01> hi all
Jun 05 04:51:34 <dyoung> I've had it for a while though; but it only shows its ugly head when its totally silent.
Jun 05 04:54:23 <dyoung> Mr Hammond does make a good point.
Jun 05 04:54:24 <jacques> my left ear has a problem because my brother set off a firework too close to it one time by accident
Jun 05 04:55:19 <dyoung> I wonder if we should do something similar to Oleg, and make ipkg unavailble and have the unsling script ipkg-cl install ipkg or something.
Jun 05 04:55:55 <dyoung> jacques, Hmm, it coule be from that too...  Blowing stuff up without ear protection.....
Jun 05 04:56:12 <dyoung> But either way, now that I'm older I wish I didnt have the issue!
Jun 05 04:56:26 <jacques> the only time I really notice it is with really loud sounds, like gunshots
Jun 05 04:56:47 <jacques> these days I am pretty careful about my ears
Jun 05 04:56:52 <dyoung> gcc-cross finshed.
Jun 05 04:57:46 <dyoung> jacques, me too.  I try to wear the earplugs you stuff into your ear, PLUS the mickey mouse ears when going into sorta noisy places like engine rooms.
Jun 05 04:58:25 <jacques> i have some nice gell filled ear muffs I use
Jun 05 04:58:37 <jacques> 29dB attenuation IIRC
Jun 05 04:59:01 <dyoung> Nice.
Jun 05 05:00:55 <dyoung> most of my mickey mouse ears are Govt Surplus.
Jun 05 05:04:44 <dyoung> http://www.fg-ind.com/intro-handsignals.html becomes useful. :-)
Jun 05 05:05:53 <dyoung> I dont remember the other ones though....
Jun 05 05:06:09 <rwhitby> dyoung: just leaving off the ipkg->ipkg-cl link will be fine, and we'll make unsling add it to the external rootfs
Jun 05 05:06:21 <jacques> yeah I had to learn all those when I worked maintenance at the refinery
Jun 05 05:06:58 <dyoung> rwhitby, Okay, so the end-user will get what the firmware provides, and not neccessarily the latest/greatest?
Jun 05 05:07:25 <rwhitby> dyoung: what do you mean?
Jun 05 05:07:33 <rwhitby> oh, as far as ipkg version goes?
Jun 05 05:07:36 <dyoung> right.
Jun 05 05:10:28 * rwhitby hmms
Jun 05 05:11:13 <dyoung> I only bring this up because of the recent ipkg "oops".
Jun 05 05:11:17 * jacques can't keep his eyes open any longer and goes to bed.
Jun 05 05:11:35 <dyoung> night jacques.
Jun 05 05:11:39 <rwhitby> I don't think we can successfully ipkg upgrade ipkg for unslung
Jun 05 05:12:34 <rwhitby> and we haven't needed to so far
Jun 05 05:12:51 <rwhitby> the "oops" was just that, an "oops".  not a sign of anything broken
Jun 05 05:13:16 <rwhitby> night jacques, thx for the gcc help
Jun 05 05:14:25 <rwhitby> so for r29 stuff, should I just bk mv any foo-2.3r25 directories to just foo (so they will be searched by both _2.3r25.bb and _2.3r29.bb files) or make a complete copy of each foo-2.3r25 to foo-2.3r29 ?
Jun 05 05:17:21 <dyoung> based on Tiersten's diff, I think the former is probably ok.
Jun 05 05:18:16 <rwhitby> so we just have some extra diff files, which are only included in the kernel r29.bb ?
Jun 05 05:18:35 <dyoung> I was kind of thinking that a ipkg-cl upgrade ipkg might be a "feature" because it will ensure that the settings are already correct for when the user really does a ipkg update.  But I guess its kinda superfluous.
Jun 05 05:19:57 <dyoung> I suppose to be cleaner, iti might be better to copy the whole stuff to foo-2.3R29.
Jun 05 05:20:04 <dyoung> but I'm not sure.
Jun 05 05:20:04 <rwhitby> I'm not sure how we would introduce that kind of thing without having to separate feeds for old and new firmware (pre change and post change) - otherwise we get the "oops" happening again for the pre change firmware.
Jun 05 05:21:36 <rwhitby> I think I'll remove the version, cause I think as is the case for the genesys patch that we'd only ever add our versions of kernel patches
Jun 05 05:21:38 <dyoung> Oh yeah.  To keep it simple, the suggestion of generating the link in unsling is probably better.
Jun 05 05:24:00 <dyoung> I'm following jacques lead.
Jun 05 05:24:11 <rwhitby> night dyoung-logger
Jun 05 05:25:27 <dyoung-zzzz> My unslung build completed successfully.
Jun 05 05:32:02 <rwhitby> 5.1 will be released tonight.  It will probably be non-functional, but it's a place to start.
Jun 05 05:38:44 <rwhitby> http://www.nslu2-linux.org/wiki/Unslung/UnslungFeatures has the proposed 5.x features.
Jun 05 05:40:43 <rwhitby> Most will not be there tonight
Jun 05 05:47:03 <rwhitby> anyone remember where Tiersten put the R25->R29 kernel patch?
Jun 05 05:48:43 <rwhitby> nm - someone else can add it to the .bb while I sleep tonight :-)
Jun 05 05:53:57 <CIA-8> 03caplink811 * 10unslung/make/sysstat.mk: initial checkin
Jun 05 05:56:19 <CIA-8> 03caplink811 * 10unslung/sources/sysstat/ (CONFIG sysstat.crond): initial checkin
Jun 05 06:18:29 <rwhitby-away> ok, that's 5.1-alpha ready for others to bash on ... first thing is for someone to add Tiersten's kernel patch (if there is anything other than the Genesys patch or the IXP interrupt rate).  If not, then all we need to do is build the interrupt-driven IXP and add that.
Jun 05 06:44:29 * giel peeks around the corner
Jun 05 07:11:24 * rwhitby-away tries a Sercomm upgrade over wireless .....
Jun 05 07:12:30 <rwhitby> Tiersten: ping
Jun 05 07:20:03 <rwhitby> ok, don't try flashing 5.1-alpha, as it panics.
Jun 05 07:20:21 <Tiersten_> hey
Jun 05 07:20:23 <Tiersten_> whats up?
Jun 05 07:20:47 <rwhitby> just wondering where you put that R25 to R29 kernel patch?
Jun 05 07:21:11 <Tiersten_> http://www.trejan.com/projects/nslu2/linux-kernel-R25_to_R29.patch
Jun 05 07:21:15 <Tiersten_> Best to host that on SF or something
Jun 05 07:21:36 <Tiersten_> There really isn't anything particularly significant they've changed apart from the hack for the Genesys stuff
Jun 05 07:22:01 <Tiersten_> A couple more Maxtor OneTouch IDs and they've fiddled with the IXP driver slightly
Jun 05 07:22:10 <rwhitby> -bash-2.05b$ wget http://www.trejan.com/projects/nslu2/linux-kernel-R25_to_R29.patch
Jun 05 07:22:10 <rwhitby> --07:21:49--  http://www.trejan.com/projects/nslu2/linux-kernel-R25_to_R29.patch
Jun 05 07:22:10 <rwhitby>            => `linux-kernel-R25_to_R29.patch'
Jun 05 07:22:10 <rwhitby> Resolving www.trejan.com... 216.167.90.227
Jun 05 07:22:11 <rwhitby> Connecting to www.trejan.com[216.167.90.227]:80... failed: Connection refused.
Jun 05 07:22:18 <Tiersten_> odd
Jun 05 07:22:22 <rwhitby> that's from sf.net
Jun 05 07:23:17 <Tiersten_> That's SF
Jun 05 07:23:20 <Tiersten_> Nothing to do with me
Jun 05 07:23:35 <rwhitby> yeah, I just realised that.
Jun 05 07:23:38 <Tiersten_> It blocks other sites as well.  Just tried intel.com, nslu2-linux.org, google.com etc...
Jun 05 07:23:40 <Tiersten_> I'll upload it
Jun 05 07:24:05 <rwhitby> doing that now
Jun 05 07:24:08 <rwhitby> done
Jun 05 07:24:39 <rwhitby> http://nslu.sf.net/downloads/linux-kernel-R25_to_R29.patch
Jun 05 07:25:00 <Tiersten_> uh.  wget doesn't work :)
Jun 05 07:25:02 <Tiersten_> I'll upload it
Jun 05 07:25:10 <rwhitby> k
Jun 05 07:25:41 <Tiersten_> Okay.  Done
Jun 05 07:26:21 <Tiersten_> I'm curious about one of the IXP changes however
Jun 05 07:26:29 <Tiersten_> "Our board only one Ethernet port. And this port connect to PHY 2 while we call this port ixp0. So we need to cross port 1 and 2. "
Jun 05 07:27:11 <rwhitby> yeah
Jun 05 07:28:58 <rwhitby> wonder if that phy address swap fixes the errors that it prints on boot regarding MII stuff
Jun 05 07:30:37 <Tiersten> No.  It won't fix that
Jun 05 07:30:52 <Tiersten> There is a table of known (to the driver) PHY IDs
Jun 05 07:31:00 <Tiersten> The specific one that is used in the NSLU2 isn't in that list
Jun 05 07:31:34 <Tiersten> We could add it in and then put some of the PHY specific initialisations in but don't' know if it's worth the time
Jun 05 07:32:17 <rwhitby> Kernel panic: VFS: Unable to mount root fs on 01:00
Jun 05 07:32:19 <rwhitby> :-(
Jun 05 07:32:29 <Tiersten> What was changed in 5.x?
Jun 05 07:32:41 <rwhitby> I updated to R29
Jun 05 07:32:43 <Tiersten> Oh yeah.  I don't know if any of the Linksys binaries have changed between R25 and R29
Jun 05 07:32:51 <Tiersten> Updated what?
Jun 05 07:32:57 <rwhitby> just the ramdisk image
Jun 05 07:33:03 <rwhitby> (which includes all the binaries)
Jun 05 07:33:03 <Tiersten> ah
Jun 05 07:33:32 <Tiersten> It's not even able to load the ramdisk though?
Jun 05 07:33:47 <rwhitby> But we will be doing the following: http://www.nslu2-linux.org/wiki/Unslung/UnslungFeatures
Jun 05 07:34:18 <Tiersten> Yeah.  Saw your post easlier
Jun 05 07:34:20 <Tiersten> earlier
Jun 05 07:34:37 <rwhitby> Hmm - it shouldn't even be trying to boot from ramdisk.   I suspect PEBKAC.
Jun 05 07:34:58 <Tiersten> Let me find out what was changed in the binaries in R29
Jun 05 07:35:37 <rwhitby> ah - I think dyoung's CMDLINE_ROOT has broken it
Jun 05 07:36:37 <rwhitby> Kernel command line: console=ttyS0,115200n8 root=/dev/ram0 initrd=0x01000000,10M mem=32M@0x00000000
Jun 05 07:37:14 <rwhitby> that comes from nslu2-linksys-kernel_2.4.22.bb
Jun 05 07:37:26 <rwhitby> and is not overridden anymore due to the ?= that dyoung added.
Jun 05 07:39:23 <rwhitby> Fixed.  Wonder why CIA is not reporting bk pushes here ...
Jun 05 07:39:48 <Tiersten> It was working earlier?
Jun 05 07:40:15 <Tiersten> The last one from it in my logs was about 1:40 hours ago
Jun 05 07:40:26 <rwhitby> dyoung got one through
Jun 05 07:40:36 <rwhitby> last couple of mine haven't come through
Jun 05 07:40:43 <Tiersten> odd
Jun 05 07:41:25 <Tiersten> It boot now?
Jun 05 07:41:31 <rwhitby> still compiling
Jun 05 07:41:37 <Tiersten> k
Jun 05 07:47:51 <rwhitby> I'm starting to like this SerComm upgrade tool ...
Jun 05 07:49:19 <Tiersten> heh
Jun 05 07:49:39 <rwhitby> it's the little bar graph which makes it ..
Jun 05 07:49:50 <rwhitby> and the verify step
Jun 05 07:49:57 <rwhitby> it's so reassuring :-)
Jun 05 07:50:35 <rwhitby> upslug: no slugs found, turkey problems?
Jun 05 07:50:58 <rwhitby> upslug won't work through VMWare NAT and a wireless LAN
Jun 05 07:52:04 <Tiersten> It should work over wireless LAN but it won't work over NAT
Jun 05 07:53:29 <rwhitby> ixp425_eth: ixp0 is using the PHY at address 1
Jun 05 07:53:30 <rwhitby> ixp425_eth: ixp1 is using the PHY at address 0
Jun 05 07:53:44 <Tiersten> Yeah.  Thats what they changed.  Is it still using ixp0 though?
Jun 05 07:53:47 <rwhitby> I guess I've got that patch already, cause we use the Linksys binary ixp module
Jun 05 07:53:52 <Tiersten> ah
Jun 05 07:53:55 <Tiersten> yeah
Jun 05 07:53:57 <rwhitby> yep
Jun 05 07:54:16 <Tiersten> TBH we don't really need the R29 patch at all
Jun 05 07:54:43 <rwhitby> apart from the Maxtor ids
Jun 05 07:55:05 <Tiersten> Yeah
Jun 05 07:55:15 <rwhitby> which could be the reason why one guy on the mailing list said that Unslung didn't work for his disk
Jun 05 07:55:32 <Tiersten> No.  The OneTouch stuff is so it handles those buttons on the front
Jun 05 07:55:35 <rwhitby> in fact, I would suspect that is the only reason why R29 was released - to continue the Maxtor compatibility
Jun 05 07:55:56 <ka6sox> hi
Jun 05 07:55:59 <Tiersten> Otherwise it appears as a standard USB mass storage device
Jun 05 07:56:00 <rwhitby> ka6sox: hi
Jun 05 07:56:11 <ka6sox> everything going smooth?
Jun 05 07:56:14 <Tiersten> It adds some stuff to hook into the button (and LED?) on the front of the Maxtor
Jun 05 07:56:19 <Tiersten> Hi ka6sox
Jun 05 07:56:32 <ka6sox> morning Tiersten 
Jun 05 07:56:39 <Tiersten> R29 fixes large transfers as well apparently
Jun 05 07:56:40 <Tiersten> "Fixed large file lockup issue caused by files larger than 1GB being transfered across the network."
Jun 05 07:56:54 <Tiersten> That's all they've got in the release notes for R29 from R25
Jun 05 07:56:57 <rwhitby> that could only be the polling interval
Jun 05 07:57:22 <Tiersten> I'm more thinking one of the binaries
Jun 05 07:57:31 <Tiersten> dunno.  Need to check
Jun 05 07:57:48 <rwhitby> anyway, we have the new binaries in 5.1 now
Jun 05 07:57:51 <rwhitby> and it boots fine
Jun 05 07:57:56 <Tiersten> cool
Jun 05 07:58:16 <rwhitby> that was much easier than I thought it was going to be
Jun 05 07:58:36 <rwhitby> someone just needs to add the unusual_devices patch and we'll be golden.
Jun 05 08:00:11 <rwhitby> we need dropbear added to the base image, so that you can copy ipks to a slug which doesn't have net connectivity
Jun 05 08:00:31 <Tiersten> I'll split out the usual devices patch
Jun 05 08:00:34 <Tiersten> unusual_devices
Jun 05 08:00:59 <rwhitby> thx Tiersten
Jun 05 08:01:27 <Tiersten> hrm
Jun 05 08:01:30 <Tiersten> there are loads of differences
Jun 05 08:01:57 <rwhitby> Now that I've put down the infrastructure, I'll let others flesh it out while I sleep ....
Jun 05 08:02:06 <Tiersten> Night
Jun 05 08:02:12 <rwhitby-asleep> you mean in the binaries?
Jun 05 08:02:14 <ka6sox> we will have to see if UpSlug works on OSX Macs if we are going to remove Maintainence mode.
Jun 05 08:02:15 <Tiersten> yeah
Jun 05 08:02:19 <Tiersten> Busybox is different for one
Jun 05 08:02:22 <Tiersten> It's slightly bigger
Jun 05 08:02:23 <rwhitby-asleep> other than just a recompile maybe?
Jun 05 08:02:28 <Tiersten> Nearly all the binaries are different
Jun 05 08:02:36 <Tiersten> Not sure.  I need to compare version numbers
Jun 05 08:02:40 <rwhitby-asleep> yeah, I figured it was a recompile
Jun 05 08:02:50 <rwhitby-asleep> ka6sox: nod
Jun 05 08:02:52 <Tiersten> They've added something to handle the onetouch as well
Jun 05 08:02:59 <rwhitby-asleep> and removed telnetd
Jun 05 08:03:03 <Tiersten> Yeah
Jun 05 08:03:16 <rwhitby-asleep> heh - that means 5.1 doesn't have telnet and is serial-only at the moment :-)
Jun 05 08:03:28 <Tiersten> heh.  Easy enough to extract telnet
Jun 05 08:03:46 <rwhitby-asleep> just add it to slingbox
Jun 05 08:03:53 <rwhitby-asleep> or leave it out :-)
Jun 05 08:04:01 <ka6sox> too insecure
Jun 05 08:04:24 <rwhitby-asleep> Firmware Version:  V2.3R29-uNSLUng-5.1-alpha   
Jun 05 08:05:58 <root_> hi, was there something changed with the libc the last few days?
Jun 05 08:06:14 <rwhitby-asleep> ah, hang on.  I think I added telnetd to the R29 ramdisk image that I put up on sf.net, cause telnetd is there.
Jun 05 08:06:26 <Tiersten> root_: Not that I know of
Jun 05 08:06:44 <Tiersten> busybox is the same version and was compiled with the same version of GCC.  Still don't know why it's slightly bigger
Jun 05 08:07:07 <rwhitby-asleep> yeah, 5.1 is telnetd-enabled, same as before.
Jun 05 08:07:13 <root_> ups, wrong name, will be back with a better one :)
Jun 05 08:07:58 <mr_claus> Tiersten: where i can take a look into cvs?
Jun 05 08:08:22 <rwhitby-asleep> cvs.sf.net:/cvsroot/nslu
Jun 05 08:08:30 <rwhitby-asleep> co unslung
Jun 05 08:08:40 <rwhitby-asleep> (assuming you're looking for Unslung packages, not the firmware)
Jun 05 08:09:00 <mr_claus> is there a webclient available to take a look into cvs and the changes?
Jun 05 08:09:09 <rwhitby-asleep> sf.net
Jun 05 08:09:18 <rwhitby-asleep> projects/nslu
Jun 05 08:10:12 * rwhitby-asleep wanders off to bed ....
Jun 05 08:10:22 <mr_claus> ok, thx
Jun 05 08:10:25 <ka6sox> nitey
Jun 05 08:11:21 <mr_claus> ahh, jp30 checked in ldconfig :)
Jun 05 08:11:23 <mr_claus> nice
Jun 05 08:12:44 <Tiersten> Binaries that have changed are: /bin/busybox, /etc/rc.d/rc.bootbin, /home/httpd/html/index.cgi, /home/httpd/html/Management/setup.cgi, /lib/libhd.so, /lib/modules/2.4.22-xfs/kernel/drivers/net/ixp425_eth.o, /sbin/fdisk, /usr/sbin/QuickSet, /usr/sbin/upnpd and /usr/sbin/USB_Detect
Jun 05 08:13:15 <Tiersten> The IXP module differences are because they did that patch
Jun 05 08:14:01 <ka6sox> but not v1.5 of IXP?
Jun 05 08:14:06 <Tiersten> No
Jun 05 08:14:08 <Tiersten> Still 1.3
Jun 05 08:14:08 <ka6sox> k
Jun 05 08:14:23 <ka6sox> redboot still the same?
Jun 05 08:15:10 <Tiersten> I assume so.  It doesn't get flashed anyway
Jun 05 08:15:21 <ka6sox> yeah.
Jun 05 08:15:53 <mr_claus> Upgrading findutils on root from 4.1.20-r0 to 4.1.20-r1...
Jun 05 08:15:53 <mr_claus> Downloading http://ipkg.nslu2-linux.org/feeds/unslung/oe/findutils_4.1.20-r1_armeb.ipk
Jun 05 08:16:05 <mr_claus> ipkg_download: ERROR: Command failed with return value 1: `wget --passive-ftp  -q -P /tmp/ipkg-k2ho5m http://ipkg.nslu2-linux.org/feeds/unslung/oe/findutils_4.1.20-r1_armeb.ipk'
Jun 05 08:16:18 <mr_claus> any ideas why i get that error?
Jun 05 08:16:38 <mr_claus> i have run ipkg update before
Jun 05 08:16:56 <Tiersten> rwhitby posted about this on the mailing list.  It's a known issue.  Remove the OE feed
Jun 05 08:17:10 <mr_claus> ok
Jun 05 08:18:08 <mr_claus> whats up with packages i have installed from this source?
Jun 05 08:18:32 <Tiersten> *shrug* Look in the mailing list.  I've not been tracking the status of Unslung
Jun 05 08:18:58 <Tiersten> R29 adds /usr/sbin/onetouch_detect and modifies /etc/rc.d/rc.quickset to run it at start up
Jun 05 08:19:19 <Tiersten> R29 removes /home/httpd/html/Management/telnet.cgi, /home/httpd/html/Management/telnet.htm and /usr/sbin/telnetd which we already know about
Jun 05 08:19:39 <Tiersten> R29 fixes a typo in /home/httpd/html/home.htm
Jun 05 08:20:03 <Tiersten> And just generally updates a couple of files which have the version numbers in
Jun 05 08:20:35 <Tiersten> That along with the kernel changes are the sum total of R25 -> R29
Jun 05 08:21:39 <ka6sox> same kernel?
Jun 05 08:21:43 <ka6sox> 2.4.22?
Jun 05 08:21:52 <Tiersten> Same version yeah but with some minor patches
Jun 05 08:22:30 <Tiersten> Changes the poll rate in the IXP driver, swaps around the PHY addresses, adds two new Maxtor OneTouch IDs and a hack for the Genesys chips
Jun 05 08:22:56 <mr_claus> i have a structural question, it's nice to copy the / -> /share/hdd/data but it would nicer to copy it to /share/hdd/data/unslung and mount /share/hdd/data/unslung as root because all dirs in the data partition now mixed up with the data dirs
Jun 05 08:27:10 <Tiersten> mr_claus: Probably a better idea to post it to the mailing list
Jun 05 08:28:05 <CIA-8> 03caplink811 * 10unslung/sources/sysstat/rc.sysstat: initial checkin
Jun 05 08:28:34 <CIA-8> 03caplink811 * 10unslung/sources/sysstat/postinst: initial checkin
Jun 05 08:31:50 <CIA-8> 03caplink811 * 10unslung/make/sysstat.mk: makefile is now able to build the source, and the final ipk
Jun 05 08:37:07 <CompBrain> Anyone have a cross compiliation enviorment they'd be willing to compile MPD (musicpd.org) for me?
Jun 05 08:37:16 <CompBrain> I'd be willing to pay
Jun 05 08:42:11 <bofh_omega> CompBrain: -v (what system arch do you need, what libs are you using....)
Jun 05 08:43:44 <mr_claus> i think the nslu2 hardware
Jun 05 08:45:22 <bofh_omega> hehe, maybe good guess ;) Now we only need to guess the libs
Jun 05 08:57:36 <bofh_omega> next guess would be what should mpd support (which kind of audio files)
Jun 05 09:10:47 <bofh_omega> its compiled now (with mp3 and id3tag support)
Jun 05 09:12:11 <bofh_omega> whats my price ;) and where do you want me to send it to?
Jun 05 09:13:16 <bofh_omega> you'll need libnsl.so.0 libm.so.0 libz.so.1 libc.so.0 libgcc_s.so.1 ld-uClibc.so.0
Jun 05 09:42:49 <garpinc> Anyone a round?
Jun 05 09:47:44 <garpinc> I tried to do upgrade to 4.x but when I disconnect both hard drives and try to boot up I never get access to web interface. Does anyone have any ideas?
Jun 05 09:50:13 <bofh_omega> try ips 192.168.0.1 192.168.1.77
Jun 05 09:52:37 <garpinc> I did but I will try it again...
Jun 05 09:53:26 <bofh_omega> and for shure your regular IP ;)
Jun 05 09:53:47 <garpinc> The regular ip is 192.168.1.77
Jun 05 09:53:52 <bofh_omega> k ;)
Jun 05 09:54:40 <bofh_omega> if nothing else helps goto redboot and redo (or wait if someone with more glue tells you what to do)
Jun 05 09:55:49 <garpinc> I'm powering up now...
Jun 05 09:56:13 <garpinc> I also tried to get to redboot and couldn't...
Jun 05 09:56:28 <garpinc> Ok I got the beep
Jun 05 09:56:35 <garpinc> Let me try it..
Jun 05 09:57:52 <garpinc> neither ip works...
Jun 05 09:58:39 <garpinc> Actually when I try to go to 192.168.0.1 it trys to go to 10.9.95.1
Jun 05 09:58:46 <garpinc> Any ideas about that?
Jun 05 10:02:04 <bofh_omega> nop, sry but seems a bit wired to mee
Jun 05 10:03:35 <garpinc> If I show routing tables in router it shows:
Jun 05 10:03:57 <garpinc> 10.9.95.1 with subnetmast 255.255.255.255
Jun 05 10:04:10 <garpinc> 192.168.1.0 with mask 255.255.255.0
Jun 05 10:04:31 <garpinc> and 0.0.0.0 with mask 0.0.0
Jun 05 10:04:43 <garpinc> the first 2 have gateway 0.0.0.0
Jun 05 10:04:56 <garpinc> and the last has gateway 10.9.95.1
Jun 05 10:05:03 <garpinc> Does that seem right?
Jun 05 10:10:04 <caplink811> garpinc: we should step over to #nslu2-general where youre question seems more on topic
Jun 05 10:10:28 <garpinc> That's fine...
Jun 05 12:21:12 <bofh_omega> is it possible to check an openslug-image befor flashing it? Just to check if it would boot and if the net (dropbear) would work
Jun 05 12:49:24 <to-pse[1]> Hi everyone, I'm new to this irc-business - so please forgive me any initial wrong-doings ;-)
Jun 05 12:52:00 <to-pse[1]> Is there anyone around here, who can tell me if I can configure the toolchain "crosstool-native" to be able to generate little-endian libraries & executables?
Jun 05 12:52:17 <Tiersten> why do you want to do that?
Jun 05 12:52:34 <Tiersten> If you're not going to do native compilation then you might as well do it on your PC
Jun 05 12:53:00 <to-pse[1]> In addition to my NSLU2 which i have unslung to 4.2 (thanks everyone here involved for a cool job) I got myself a WMA11B media player
Jun 05 12:53:44 <to-pse[1]> I'd like to do a little work on the WMAMP replacement firmware which is also based on the same ARM architecture, but with small endian...
Jun 05 12:54:11 <Tiersten> You will need to recompile GCC & co if you want all of that
Jun 05 12:54:23 <Tiersten> crosstool-native is only for big endian
Jun 05 12:54:35 <Tiersten> You can't switch it on the fly
Jun 05 12:54:38 <to-pse[1]> ok - I thought it might just be some kind of run-time-switch
Jun 05 12:58:13 <Tiersten> GCC does have a switch but I've never used it
Jun 05 12:58:52 <Tiersten> Download crosstool from http://kegel.com/crosstool/
Jun 05 12:59:11 <Tiersten> The libraries are big endian anyway for crosstool-native
Jun 05 13:00:08 <to-pse[1]> Thanks for the link - I will take a look at the stuff...
Jun 05 13:03:46 <bofh_omega> have a nice day
Jun 05 15:17:41 <ka6sox> ~seen jbowler
Jun 05 15:17:44 <jbot> jbowler <~jbowler@jbowler.nslu2-linux> was last seen on IRC in channel #oe, 3d 6h 7m 33s ago, saying: 'etc/init.d/populate-var.sh ?'.
Jun 05 15:18:03 <ka6sox> ~seen jbowler-away 
Jun 05 15:18:04 <jbot> jbowler-away is currently on #oe (1d 16h 16m 26s)
Jun 05 15:23:01 <rwhitby> morning ka6sox
Jun 05 15:23:09 <rwhitby> anyone do any work on 5.1 overnight?
Jun 05 15:36:43 * jacques just got up.
Jun 05 15:47:12 <jacques> howdy Christopher 
Jun 05 15:47:20 <dyoung-zzzz> morning everyone.
Jun 05 15:47:23 <Christopher> howdy
Jun 05 15:47:29 <dyoung-zzzz> sorry for the cmdline breakage.
Jun 05 15:47:32 <jacques> howdy dyoung-zzzz 
Jun 05 15:47:43 <rwhitby> jacques: you are in a completely opposite timezone to me, but are keeping the same waking hours :-)
Jun 05 15:47:50 <rwhitby> dyoung: np
Jun 05 15:48:00 <jacques> rwhitby, lol I was just thinking the same thing
Jun 05 15:48:04 <dyoung> did I change the wrong one, or is it even possible to have a overridable cmdline ?
Jun 05 15:48:32 <jacques> my allergies are making me sleep more these days :-(
Jun 05 15:48:36 <Christopher> I keep getting this error since I upgraded to unslung 4.20 init: /bin/sh respawning too fast
Jun 05 15:48:46 <rwhitby> dyoung: the trouble is the include of the linksys kernel bb
Jun 05 15:49:57 <Christopher> any ideas?
Jun 05 15:51:09 <dyoung> Christopher, where are you getting that message?
Jun 05 15:51:18 <dyoung> (ie: telnet in, serial, ... ?)
Jun 05 15:52:00 <jacques> it usually means something init is configured to respawn is erroring out, and is getting respawned in a tight loop
Jun 05 15:52:26 <jacques> you made any changes to your inittab ?
Jun 05 15:52:39 <Christopher_> nope
Jun 05 15:54:21 <jacques> you do anything "strange" to your serial port?
Jun 05 15:55:34 <Christopher_> it is connected to my laptop through a USB serial dongle
Jun 05 15:56:24 <jacques> this same hardware worked in the past? connected same way?
Jun 05 15:56:44 <Christopher_> yes
Jun 05 15:56:59 <jacques> hmm, then offhand I dunno :-\
Jun 05 15:57:03 <Christopher_> ok, thanks
Jun 05 15:57:10 <Christopher_> i'll try to figure it out
Jun 05 15:57:16 <jacques> maybe some other clues in the logs
Jun 05 15:57:29 <Christopher_> it works if my hardisk os not plugged in to start with
Jun 05 15:57:36 <Christopher_> s /os /is
Jun 05 16:01:47 <ka6sox> Christopher_bbl, ping
Jun 05 16:01:53 <ka6sox> rats...missed him
Jun 05 16:02:40 * jacques goes to make a sandwich.
Jun 05 16:02:51 <rwhitby> dyoung: have you had any problems getting maintenance mode to work reliably on 4.20?
Jun 05 16:03:30 <rwhitby> It just refuses to do the upgrade for me, due to the memory calculation issue in the uprade.cgi binary
Jun 05 16:03:39 <ka6sox> Christopher_bbl, do you know if anyone has tested UpSlug with OSX?
Jun 05 16:03:48 <rwhitby> anyone here got comments on the maintenance mode debate?
Jun 05 16:10:05 <rwhitby> the channel goes strangely silent ....
Jun 05 16:11:22 <Tiersten> My opinion is that ideally we should keep the maint mode stuff but it would require writing our own versions of the Linksys apps which we really should do anyway.  The short term method I guess is to do what you propose and do away with it all.
Jun 05 16:12:51 <Tiersten> Getting rid of the ability to do a network upgrade without having to plug the NSLU2 in directly is a big annoyance
Jun 05 16:13:30 <rwhitby> Tiersten: I've done a number of upgrade mode upgrades in the last 24 hours wirelessly
Jun 05 16:13:49 <ka6sox> my concern is for the people who have OSX machines (and NO peecee's)
Jun 05 16:13:59 <Tiersten> Whats wrong with upslug?
Jun 05 16:14:03 <Tiersten> (Not tried it)
Jun 05 16:14:08 <ka6sox> dunno if it runs on OSX
Jun 05 16:14:09 <rwhitby> just haven't confirmed it yet
Jun 05 16:14:14 <ka6sox> correct
Jun 05 16:14:17 <rwhitby> should be no reason why it wouldn't work
Jun 05 16:14:18 <Tiersten> rwhitby: Wireless should be fine.  It's people trying to use NAT along the line that will break it
Jun 05 16:14:39 <ka6sox> Tiersten, that is every cable/dsl user
Jun 05 16:14:41 <Tiersten> It's unlikely that somebody will have NAT within their own network anyway unless you're doing something with VMware/VirtualPC etc...
Jun 05 16:14:50 <Tiersten> No.  Within their network
Jun 05 16:14:50 <rwhitby> ah, right.  yeah
Jun 05 16:14:54 <ka6sox> ah
Jun 05 16:14:55 <Tiersten> You have NAT inside your network??
Jun 05 16:15:01 <ka6sox> VPC
Jun 05 16:15:16 <Tiersten> If upslug works then we don't need VirtualPC/VMware etc
Jun 05 16:15:22 <ka6sox> true
Jun 05 16:15:38 <Tiersten> Probably could slap a GUI on it using WxWidgets then people won't be scared off by the command line interface
Jun 05 16:15:45 <ka6sox> as soon as I can get ahold of Christopher I'll ask
Jun 05 16:16:41 <rwhitby> back later
Jun 05 16:16:55 <ka6sox> and Linux PPC also.
Jun 05 16:17:38 <Tiersten> If he wrote it properly then it should work with no problems under any version of Linux that has an ethernet interface
Jun 05 16:18:14 <Tiersten> It's whether OSX has the same socket interface that is the issue.  I don't see why it wouldn't work.  He doesn't seem to be doing anything odd or special to Linux
Jun 05 16:20:40 <ka6sox> okay cool
Jun 05 16:21:41 <dyoung> it doesnt compile for lack of netpacket.h
Jun 05 16:21:56 <dyoung> netpacket/packet.h
Jun 05 16:22:21 <Tiersten> You got sockaddr_ll in OSX?
Jun 05 16:22:28 <dyoung> beats me.
Jun 05 16:22:34 <dyoung> how do I check?
Jun 05 16:22:48 <Tiersten> I'll do it
Jun 05 16:23:02 <ka6sox> Tiersten, you got your Mini?
Jun 05 16:23:11 <Tiersten> Got it ages ago
Jun 05 16:23:14 <ka6sox> k
Jun 05 16:24:54 <Tiersten> hmm.  how do you do AF_PACKET or equiv under a Mac...
Jun 05 16:25:11 <dyoung> I dunno, I did some grepping around and didnt find AF_PACKET anyplace.
Jun 05 16:25:14 <Tiersten> well.  without a working upslug then the getting rid of maint mode idea is dead
Jun 05 16:25:27 <Tiersten> unless we're happy to only allow Windows and Linux users to use Unslung...
Jun 05 16:25:36 <Tiersten> Or have to use redboot directly which is a bit much
Jun 05 16:25:37 <ka6sox> :P
Jun 05 16:25:56 <ka6sox> the telnet is tough to catch Redboot fast enough
Jun 05 16:26:10 <Tiersten> See.... Annoying Mac users! ;)
Jun 05 16:26:17 <ka6sox> heh
Jun 05 16:26:56 <ka6sox> there is enough *stuff* in the world that only Speaks Doze.
Jun 05 16:35:18 <ka6sox> other than OSX users not being able to program Unslung no other reason NOT to take out Maint mode.
Jun 05 16:35:46 <Tiersten> won't that annoy people who use a Mac?
Jun 05 16:40:03 <dyoung> looks like some contionals in upslug to use BPF calls may work.
Jun 05 16:40:12 <dyoung> but I dont really want to look at it right now.
Jun 05 16:40:22 <Tiersten> I was thinking about using pcap or another library to provide the abstraction layer
Jun 05 16:40:38 <Tiersten> It'd be a pain to have a billion different ifdefs just to handle everything
Jun 05 17:21:58 <jbowler-away> AF_PACKET is used by dhcp, Mac's do DHCP, therefore AF_PACKET must be supported at some level.
Jun 05 17:22:23 <Tiersten> It might use the BPF library
Jun 05 17:22:31 <jbowler-away> What's that?
Jun 05 17:22:52 <Tiersten> Berkley Packet Filter
Jun 05 17:23:11 <Tiersten> Basically it's a way of getting low level access in a portablish way
Jun 05 17:23:30 <Tiersten> There is a way of getting what we want on the Mac but it's not AF_PACKET because that doesn't exist on the Mac.  Not the way Linux does it anyway
Jun 05 17:24:22 <jbowler-away> Um, ok - what upslug needs is to send out (and receive) raw IP packets (if I understand the code correctly.)
Jun 05 17:24:31 <Tiersten> It's raw ethernet packets
Jun 05 17:24:45 <jbowler-away> Yes, sorry - you are correct.
Jun 05 17:25:19 <Tiersten> You must be able to do it under OSX.  Just don't know what interface they have on it.  It's best to just use a packet capture library which has all this stuff built in so we don't need to worry about the specifics
Jun 05 17:25:53 <jbowler-away> Does BPF run on OSX?
Jun 05 17:26:17 <Tiersten> Yes
Jun 05 17:26:48 <ka6sox> since its a BSD boxen.
Jun 05 17:26:54 <jbowler-away> Amusing acronym: British Plastics Federation, Buddhist Peace Fellowship...
Jun 05 17:26:59 <Tiersten> It's not really.  It's a hybrid
Jun 05 17:27:17 <ka6sox> right..because of MACHl.
Jun 05 17:27:20 <ka6sox> er MACH.
Jun 05 17:27:36 <Tiersten> It's Mach at the core but with a BSDish layer bolted on top from what I understand
Jun 05 17:27:48 <Tiersten> It's the reason for a recent bug actually.  Don't know if Apple have fixed it yet
Jun 05 17:29:45 <Tiersten> Pretty bad bug actually.  Let you cause a kernel panic as a normal user
Jun 05 17:36:11 <ka6sox> well...there are funny problems still and there *was* a recent upgrade.
Jun 05 17:37:22 <Tiersten> yeah
Jun 05 17:37:34 <Tiersten> Something is odd with colorsync in Tiger for one
Jun 05 17:38:08 <ka6sox> what is the deal with that????
Jun 05 17:38:23 <ka6sox> with the Lappy its no problem.
Jun 05 17:38:36 <Tiersten> No clue.  It's just messed up under Tiger from what I hear.  iPhoto screws up the colours under Tiger
Jun 05 17:38:54 <Tiersten> Does your other machines have a custom colorsync profile loaded specific to that display?
Jun 05 17:39:08 <Tiersten> If you don't have any profile then it doesn't break.  At least doesn't visibly break
Jun 05 17:49:35 <eno> just installed tiger today
Jun 05 17:51:06 <ka6sox> ha
Jun 05 17:51:09 <ka6sox> er ah
Jun 05 18:11:02 <jbowler-away> Anyone who does a bk pull will get updated compilers, openslug (and probably unslung) does not build with these.  I'm looking for a fix (for the right version to hardwire in).
Jun 05 18:12:48 <jacques> jbowler-away, when did you pull last?
Jun 05 18:13:04 <jacques> and which errors are you seeing? we fixed a problem with gcc-cross last night
Jun 05 18:13:07 <jbowler-away> Shortly after pb checked in the new compilers.
Jun 05 18:13:19 <jbowler-away> He checked them in an hour or so ago.
Jun 05 18:13:27 <jacques> oh
Jun 05 18:13:34 <jacques> even newer compilers? :-\
Jun 05 18:13:58 <jacques> ah, gcc 4.0.0
Jun 05 18:14:07 <jacques> I wouldn't expect things to build with that
Jun 05 18:14:09 <jbowler-away> The first problem was that the csl-arm one had a 'DEFAULT_PREFERENCE_armeb', but there's a 3.4.4 too and that doesn't work either.
Jun 05 18:14:32 <jacques> heck I've so far been unable to even build a armeb soft-float cross-toolchain including gcc 4.0.0
Jun 05 18:14:35 <jbowler-away> No, gcc 4.0.0 is not the issue - there is no gcc-cross-initial_4.0.0
Jun 05 18:14:56 <jacques> jbowler-away, ok we are going to have to drill down
Jun 05 18:15:45 <jacques> 3.4.4 doesn't do what? doesn't build? doesn't build X app ?
Jun 05 18:15:52 <jacques> unslung/openslug ?
Jun 05 18:15:59 <jbowler-away> 3.4.4 does not build glibc_2.3.2
Jun 05 18:16:04 <jacques> as OK
Jun 05 18:16:09 <jacques> ah
Jun 05 18:16:22 <jacques> I only tested unslung last night, and 3.4.4 built and the build completed
Jun 05 18:16:31 <jbowler-away> At this point I'm just trying to find the correct syntax to put it back to what it was an hour ago ;-)
Jun 05 18:20:08 <jacques> I just did a pull and see no changes to gcc3.x since last night
Jun 05 18:20:53 <jacques> I'm going to try an opensluyg build to ge the error
Jun 05 18:21:30 <jbowler-away> It's curious because unslung builds glibc too, and I got a failure right at the start.
Jun 05 18:22:30 <jacques> your openslug glibc build failure is right at the start of the glibc build ?
Jun 05 18:22:57 <jbowler-away> It gets multiple symbol definitions - not quite at the start.
Jun 05 18:23:14 <jbowler-away> My bitbake is out of date (was out of date), that may be a problem...
Jun 05 18:23:52 <jacques> hmm, mine isn'ty amazingly new either - tho it does have the RAM usage patches so it's not ancient
Jun 05 18:25:36 <jbowler-away> Well, I'm going to delete all my tmp directories, make sure I've not got any local changes, and try again...
Jun 05 18:26:09 <jacques> yeah, these days I always rm tmp and even .ccache
Jun 05 18:26:11 <jbowler-away> jacques: one question, which version of gcc-cross-initial do you have (i.e. in the unslung build, in work)
Jun 05 18:26:19 <jacques> checking
Jun 05 18:26:40 <jacques> gcc-cross-initial-3.4.4-r0/
Jun 05 18:27:18 <jbowler-away> Hum.  That's weird, I would expect the 3.4.4+cvs version from csl-arm (because it has a DEFAULT_PREFERENCE of 1 on armeb)
Jun 05 18:30:21 <jbowler-away> Here's the problem: ChangeSet@1.3190.1.323, 2005-06-05 21:11:14+01:00, pb@mill.internal.reciva.com
Jun 05 18:30:21 <jbowler-away>   shuffle some gcc DEFAULT_PREFERENCEs around
Jun 05 18:30:41 <jacques> argh
Jun 05 18:31:00 <jacques> why pb why?
Jun 05 18:31:48 <ka6sox> my thoughts exactly
Jun 05 18:32:41 <jbowler-away> So if it was ok last night before the shufffle and it's not now (assuming I didn't mess something up), then the shuffle is the clue ;-)
Jun 05 18:33:23 <jacques> yeah, if it's trying to use csl-arm (my build hasn't gotten there yet) that's gotta be the problem
Jun 05 18:33:43 <jacques> I don't trust csl-arm one iota for BE
Jun 05 18:34:29 <jbowler-away> Funny thing is I fixed it and tried again, then got the multiple symbol problem - I'll repeat with DEFAULT_PREFERENCE_nslu2 = -1
Jun 05 18:36:22 <jacques> yeah it looks like it's checking out gcc :-(
Jun 05 18:36:37 <jacques> that's gotta be it - I would bet $$$ that arm-csl has no fscking clue about BE
Jun 05 18:37:21 <jbowler-away> It will fail with an ld error something like 'don't know how to emulate linux-elf, I now about linux-elfb'
Jun 05 18:37:28 <jbowler-away> s/now/know/
Jun 05 18:39:12 <jacques> yeah, to be clear, when I say "no clue" I mean arm-csl is not tested with BE and does not have the necessary patches
Jun 05 18:39:20 <jacques> (I am speculating)
Jun 05 18:39:55 <jbowler-away> Adding DEFAULT_PREFERENCE_nslu2 = "-1" to the end of gcc_csl-arm-20050603.bb should stop it using the csl-arm version.
Jun 05 18:40:10 <jacques> OK I'll try that after my build fails
Jun 05 18:40:32 <jbowler-away> I'll tell you if mine fails first (I have that change) ;-)
Jun 05 18:41:24 <jacques> mine is still cvs co'ing gcc
Jun 05 18:43:36 <jacques> NOTE: package gcc-cross-initial-3.4.4+csl-arm-20050603-r0: task do_compile: started  :-\
Jun 05 18:43:50 <jacques> i wish pb was around
Jun 05 18:46:53 <jacques> NOTE: package glibc-2.3.2+cvs20040726-r17: task do_compile: started
Jun 05 18:47:06 * jacques gets ready for the "boom"
Jun 05 18:47:21 <jacques> NOTE: package glibc-2.3.2+cvs20040726-r17: task do_compile: failed
Jun 05 18:47:30 <jbowler-away> linux-elfb?
Jun 05 18:47:42 <jacques> /home/packages/oe/nslu2/bb_openslug_default/tmp/cross/lib/gcc/armeb-linux/3.4.4/../../../../armeb-linux/bin/ld: unrecognised emulation mode: armelf_linux
Jun 05 18:47:42 <jacques> Supported emulations: armelfb_linux armelfb
Jun 05 18:47:49 <jacques> it's missing a damn patch
Jun 05 18:48:00 <jacques> because nobody uses arm-csl for BE
Jun 05 18:55:04 <NAiL> ka6sox: ping?
Jun 05 18:56:30 <jacques> looks like arm-csl is missing all the BE patches 
Jun 05 18:57:01 <jacques> who knows what else - I see no softfloat patch either
Jun 05 19:10:15 <jbowler-away> jacques: my glibc-2.3.2+cvs20040726 build succeeded this time (with gcc-3.4.4), at least to do_package.
Jun 05 19:17:00 <jacques> jbowler-away, ok good
Jun 05 19:19:54 <jbowler-away> Do you know what the license on upslug is?
Jun 05 19:20:14 <jacques> no, I haven't looked at it in a really long time
Jun 05 19:20:55 <jbowler-away> It's not explicit in the source, but it's part of the sourceforge nslu2 project and that's "bsd, gpl" - I'm not sure which applies to upslug...
Jun 05 19:21:48 <jacques> :-\ it really should be spelled out - that's an omission
Jun 05 19:23:39 <jbowler-away> I guess the decision lies with Christopher_bbl: I'm trying to build it along with openslug, so I need to put something in the .bb file.
Jun 05 19:24:09 <jacques> this is vaguely familiar - I think the question has come up before
Jun 05 19:26:18 <jbowler-away> My guess is that 'bsd' is the preference but that 'gpl' is there because anything derived from gpl can't be un-gpl'ed.
Jun 05 19:29:57 <Christopher> back
Jun 05 19:30:03 <Christopher> for a bit
Jun 05 19:30:25 <jacques> Christopher, so what is the upslug license?
Jun 05 19:30:40 <Christopher> that's an interresting question, i had never thought about it
Jun 05 19:31:25 <Christopher> what is the other slug stuff like slugimage under?
Jun 05 19:32:12 <jacques> slugimage is bsd
Jun 05 19:34:31 <Christopher> hmmmm what's major difference between bsd and gpl?
Jun 05 19:34:39 <Christopher> or should I google it :-)
Jun 05 19:34:44 <jbowler-away> no copyleft (on bsd)
Jun 05 19:35:10 <Christopher> i c
Jun 05 19:37:25 <Christopher> then let's make upslug bsd
Jun 05 19:37:36 <jacques> BSD licensed code can be used in proprietary apps without the requirement of distributing the code - only attribution has to be maintained
Jun 05 19:37:43 <Christopher> do I need to do anything to make it that?
Jun 05 19:37:52 <jacques> no, just put it in the code header
Jun 05 19:37:58 <Christopher> oik
Jun 05 19:38:00 <jacques> probably copy from some other BSD code
Jun 05 19:38:08 <Christopher> sounds good to me
Jun 05 19:38:20 <Christopher> I'll try to get that done tomorrow evening
Jun 05 19:38:21 <Tiersten> You're okay with the binary only dist part?
Jun 05 19:38:51 <Christopher> i'll do a little more research and then decide tomorrow evening
Jun 05 19:39:04 <Tiersten> For example, sercomm are perfectly okay to do some changes and then just put up a binary on their website and not give out any code.  They just have to say that you wrote it originally
Jun 05 19:39:19 <Christopher> mmmm i c
Jun 05 19:40:16 <jbowler-away> www.opensource.org has both the relevant licenses.
Jun 05 19:40:20 <Christopher> I'll try to figure out the licencse then, now to  bed before my first day as a NASA intern
Jun 05 19:46:22 <Tiersten> good luck
Jun 05 19:47:29 <jbowler-away> jacques: openslug tries to build gcc-cross-4.0.0 too (and fails 'GMP wth MPFR support is required to build f95' in do_configure.)
Jun 05 19:48:57 <jbowler-away> I wonder how much F95 code we have in openslug...
Jun 05 20:00:54 <jbowler-away> The gmp build needs --enable-mpfr
Jun 05 20:09:33 <CIA-8> 03jbowler * 10upslug/ (Makefile upslug.c): 
Jun 05 20:09:33 <CIA-8> Correct -Wall -Os warnings in upslug.c
Jun 05 20:09:33 <CIA-8> Make the makefile work with bitbake (install target)
Jun 05 20:26:30 <jacques> jbowler-away, hmm, why's it trying to build gcc 4.0.0?  that will never work :-(
Jun 05 20:27:14 <jacques> I don't think oe should be defaulting to gcc4
Jun 05 20:27:19 <jacques> for any arch
Jun 05 20:28:13 <jbowler-away> It looks like pb has mpfr installed on his machine (so the build of the FORTRAN compiler works), the build itself isn't looking in staging at do_compile.
Jun 05 20:28:23 <jbowler-away> Still, it's a moot point if it won't work anyway.
Jun 05 20:28:35 <jbowler-away> I'll just put in a DEFAULT_PREFERENCE_nslu2 there too...
Jun 05 20:31:06 <jacques> looks like we have a lot to bring up with pb when he gets back
Jun 05 20:31:55 <jbowler-away> My build is back to 3.4.4 now
Jun 05 20:32:21 <jacques> good
Jun 05 20:34:31 <jacques> from this, http://gcc.gnu.org/ml/gcc/2005-06/msg00123.html it looks like gcc4 is not ready for production use
Jun 05 20:34:39 <jacques> (as we probably all guessed)
Jun 05 20:39:57 * jacques starts reading the thread and then stop as he realizes it's just another whiny flame war.
Jun 05 22:27:48 <ka6sox-away> NAiL, 
Jun 05 22:27:50 <ka6sox-away> pong
Jun 05 23:28:39 <mr_claus> morning
Jun 05 23:31:51 <rwhitby-web> jbowler-away: can you please push your PREFERENCE patches ?
Jun 05 23:31:55 <jbowler-away> bk changeset 1.3434 allows NSLU2 (openslug, unslung) builds to work again...
Jun 05 23:32:30 <jbowler-away> rwhitby-web - ;-) I'd pushed them about 10seconds before you spoke...
Jun 05 23:32:31 <rwhitby-web> cool - I didn't see any CIA so I wasn't sure if you'd pushed or not (can't pull from here)
Jun 05 23:32:53 <rwhitby-web> so we're back to 3.4.4 ?
Jun 05 23:33:16 <jbowler-away> Yes, 3.4.4 on cross-initial and cross (also the openslug-packages gcc builds will use that).
Jun 05 23:33:29 <rwhitby-web> thx
Jun 05 23:33:44 <rwhitby-web> BTW, what's your opinion on the maintenance mode vs upgrade mode question?
Jun 05 23:34:33 <jbowler-away> One thing about maintenance mode - if you use tmpfs rather than ext2/dd there's a lot more space available in the rootfs.
Jun 05 23:36:18 <jbowler-away> I'm not really familiar with the unslung situation.  On openslug it's possible to reflash from any non-jffs2 root (so long as there is sufficient space for the image.)
Jun 05 23:38:30 <jbowler-away> So far as I can see it's going to be really exceptional for an openslug user to run from jffs2, and such a user will probably want complete control of any upgrade process (i.e. having an automated 'upgrade' from jffs2 root is probably a waste of time.)
Jun 05 23:43:05 <rwhitby-web> the Unslung situation is that the Linksys web upgrade program (for which we do not have source code) goes and checks the free memory before it creates an 8MB upload temp ramdisk.
Jun 05 23:43:21 <mr_claus> the ldconfig package was checked in last week, i installed the package and now i get a segfault while logging in with ssh or telnet, the webui is not working too, perhaps a segfault too
Jun 05 23:43:24 <rwhitby-web> Unfortunately, it doesn't take into account memory that is temporarily in buffers or cache.
Jun 05 23:43:28 <mr_claus> could it be a problem with ldconfig?
Jun 05 23:43:33 <rwhitby-web> and so it refuses to upgrade
Jun 05 23:44:14 <jbowler-away> ;-)  Oops.
Jun 05 23:44:16 <rwhitby-web> jbowler-away: so we actually have plenty of space in RAM, but the linksys app refuses to use it
Jun 05 23:44:47 <rwhitby-web> and whether or not you get a successfull upgrade depends on how much RAM has been used during the boot for cache and buffers.
Jun 05 23:44:54 <jacques> ugh
Jun 05 23:45:12 <rwhitby-web> I personally have not been able to get a single maintenance mode upgrade to work once you have installed 4.20
Jun 05 23:45:40 <jbowler-away> You can probably free the memory by int main(void) { (void)malloc(8*1024*1024); exit(0); }
Jun 05 23:46:10 <jbowler-away> ... it might be necessary to touch the pages to be sure...
Jun 05 23:47:14 <rwhitby-web> jbowler-away: yeah, we used to do that, and I'm not sure if it's still in 4.20 or not.  That might be the problem.
Jun 05 23:47:39 <rwhitby-web> we had a upgrade-real.cgi 
Jun 05 23:48:48 <jbowler-away> WRT your original ML posting (which I have only just read - I'm a bit behind), I think it's fine so long as there is a reasonable way of making binary upslugs available.
Jun 05 23:49:29 <rwhitby-web> yeah, we would get people to create RPMs or .debs or .ebuilds or whatever for all the major distros
Jun 05 23:49:45 <jbowler-away> The RedBoot upgrade can be exited (safely) just by pulling the power plug, but obviously that won't help if mtdblock5 has been zeroed.
Jun 05 23:50:46 <jbowler-away> upslug in my experience used to be pretty flakey - with my changes I find it reliable, but I don't know how many people have tested the new version.
Jun 05 23:51:40 <rwhitby-web> I can't use upslug cause my linux development box is behind a VMWare NAT
Jun 05 23:51:51 <jacques> real men use tftp
Jun 05 23:51:58 <jbowler-away> Hum.  My openslug 3.4.4 build does not boot (it's died inside rcS.d).
Jun 05 23:52:06 <jacques> :-\
Jun 05 23:52:28 <rwhitby-web> jacques: yeah, if I have serial on the box I use tftp.  If I don't have serial, I zero mtdblock5 and use the SerComm tool
Jun 05 23:52:53 <rwhitby-web> I'm getting so used to the SerComm method that I'll probably start using that exclusively.
Jun 05 23:53:22 <jacques> jbowler-away, I hope the compiler isn't misbehaving - I noticed some strange differences in what appeared to be the only differing patch between 3.4.3 and 3.4.4
Jun 05 23:54:21 <jacques> I'm doing an openslug build right now, after pulling the latest gcc preference fixes
Jun 05 23:54:34 <jbowler-away> jacques: I had also made changes to the bootstrap because there was an order change in OE initscripts, but that shouldn't have caused this hang.
Jun 05 23:55:03 <jacques> jeez, you;d think they wouldn't change everything at once
Jun 05 23:55:53 <dyoung> rwhitby-web, someone already posted the Packages parser to the ML.
Jun 05 23:56:25 <dyoung> rwhitby-web, I also pretty much use the upslug method all the time for firmware changing now.
Jun 05 23:57:23 <jbowler-away> jacques: it was a very minor change, just moved hostname.sh to a lower level (S39 from S40).  I'm pretty sure it doesn't matter, my uclibc rootfs/etc and openslug rootfs/etc match.
Jun 05 23:57:51 <jbowler-away> jacques: do you have a serial console on your NSLU2?
Jun 05 23:58:21 <jacques> jbowler-away, I do indeed - on both
Jun 05 23:59:04 <rwhitby-web> dyoung: yeah, perlguru posted the perl script :-)
Jun 05 23:59:04 <jbowler-away> That's good.  I've just got a flashing LED ;-)  I'm going to add some codes on disk1/2 to see where it hangs but this is a slow process...
Jun 05 23:59:28 <dyoung> Oh, I was thinking that name was familiar....
Jun 05 23:59:29 <jacques> jbowler-away, well I did suggest morse... :-)
Jun 05 23:59:54 <rwhitby-web> dyoung: did you see all the r29 stuff there now
**** ENDING LOGGING AT Sun Jun 05 23:59:56 2005