[03:10] <zul> java is fun
[03:13] <crimsun> both fun and brain-stabbing heartache
[03:15] <zul> i have an interview tomorrow so i guess i should go over some stuff
[03:15] <crimsun> good luck!
[03:15] <zul> thanks
[03:15] <crimsun> technical interview, I presume?
[03:16] <zul> i suppose it is...junior/intermediate tech support for a mobile company
[03:16] <crimsun> oh, good
[03:17] <zul> better than not working :)
[03:17] <crimsun> indeed :)
[03:18] <zul> wha...transmeta going out of business?
[03:18] <crimsun> :/
[03:18] <zul> stupid slashdot..
[06:07] <fabbione> morning
[08:36] <fabbione> mjg59: <jgarzik> fabbione: that suspend patch isn't right for SCSI
[08:50] <fabbione> hey JaneW 
[09:17] <fabbione> hey JaneW (2nd attempt)
[09:22] <fabbione> infinity http://ftp.kernel.org/pub/linux/kernel/people/mbligh/abat/regression_matrix.html
[09:23] <JaneW> hi fabbione (2nd attempt)
[09:25] <fabbione> JaneW: we need a name today :)
[09:30] <JaneW> fabbione: Bashful Brazil
[09:30] <fabbione> JaneW: is that a nuts? :)
[09:33] <JaneW> Fabbione: Brazil nut yes it's a big one ;P
[09:35] <JaneW> http://www.rain-tree.com/brazilnu.htm
[09:47] <fabbione> great :)
[09:55] <mjg59> fabbione: Hmm. Bah.
[09:55] <mjg59> fabbione: To the extent that it breaks existing scsi, or that it just doesn't work on scsi?
[09:55] <fabbione> it breaks
[09:55] <mjg59> Ok
[09:55] <fabbione> and he doesn't have another patch yet
[09:56] <mjg59> Ok
[09:56] <mjg59> Well, we'll just have to hope he gets a move on :)
[09:56] <fabbione> well.. yeah
[11:12] <jbailey> fabbione: Is that in now?
[11:12] <fabbione> jbailey: just uploaded..
[11:12] <fabbione> but i have binaries handy
[11:12] <jbailey> I don't need to rush it yet.
[11:12] <fabbione> i can spare you to wait 10 hours to build :)
[11:12] <jbailey> Handy. =)  Please.
[11:12] <jbailey> ...
[11:12] <jbailey> 10 *hours*?
[11:13] <fabbione> people.u.c/~fabbione/
[11:13] <fabbione> jbailey: building 8 flavours take ages
[11:13] <fabbione> + we have build problems
[11:13] <fabbione> so it might take way more than 10 hours
[11:14] <fabbione> jbailey: note that i don't even know if that kernel will boot
[11:14] <fabbione> so keep the proper precautions
[11:14] <fabbione> s/keep/take
[11:16] <jbailey> Hey the damned thing unpacks at least, better than svenl's image did.
[11:16] <jbailey> (I still haven't hunted down why.  I get dpkg errors with his .deb)
[11:17] <fabbione> hahah ok :)
[11:18] <jbailey> Look like no joy so far.
[11:18] <chmj> morning 
[11:18] <jbailey> Lemme try it again without 'quiet'
[11:18] <jbailey> Heya Charles
[11:19] <fabbione> hey chmj 
[11:19] <fabbione> jbailey: is your machine lpar? (or something like that)
[11:19] <jbailey> dict lpar?
[11:19] <chmj> fabbione: I think someone has to specificly maintain external drivers 
[11:19] <fabbione> jbailey: dunno... i found a "typo fix" for ppc64 lpar boot
[11:19] <fabbione> jbailey: but i didn't include it
[11:20] <fabbione> chmj: inside or outside the kernel?
[11:20] <fabbione> jbailey: since it will be part of rc6
[11:20] <chmj> fabbione: I don't know about that 
[11:21] <jbailey> Oh?  Is he actually doing another rc release?
[11:21] <chmj> fabbione: some off these drivers are already patches from upstream 
[11:21] <fabbione> chmj: what external drivers you are talking about... inside our kernel or outside?
[11:21] <fabbione> jbailey: yes.. rc6 and release according to the schedule
[11:21] <jbailey> Looks like it means logical partitions
[11:21] <chmj> fabbione: the ones in external-modules
[11:21] <fabbione> jbailey: probably
[11:22] <fabbione> chmj: i am not sure understand 100% what you mean... why somebody should maintain them specifically?
[11:22] <chmj> fabbione: some off them come as patches 
[11:22] <fabbione> chmj: yes
[11:22] <fabbione> and?
[11:23] <chmj> patches have no version numbers 
[11:23] <fabbione> chmj: yes they do :)
[11:23] <fabbione> chmj: some of them in the code, others in the name
[11:23] <chmj> hmm 
[11:24] <chmj> fabbione: in the code, makes it hard for them to be automatically watched 
[11:24] <jbailey> Oh, I see.  http://www.technonics.com/ppc970.htm  - basically lpar is where you do vmwarish things in-chip.  I doubt this beast does that, it's a "consumer" g5 mac.
[11:24] <fabbione> chmj: i never said it was going to be an easy task
[11:26] <chmj> fabbione: wich is why I'm saying that someone will need to manually check that
[11:26] <chmj> s/wich/which/g
[11:26] <fabbione> chmj: well if it's only 2/3 drivers is ok
[11:26] <fabbione> just spits out a warning to check them
[11:26] <fabbione> so that people will remember
[11:26] <jbailey> Nope, still no joy - almost like the framebuffer isn't loaded.
[11:26] <chmj> ok 
[11:27] <fabbione> jbailey: FB doesn't build :/
[11:27] <jbailey> I get the Setup Done message, Starting Linux PPC64 2.6.12-1-powerpc64-smp blah blah down to smp_core99_probe smp_core99_kick_cpu smp_Core99_kick_cpu done  smp_core99_setup_cpu 0 done
[11:27] <jbailey> Right, lemme try ssh then. =)
[11:28] <fabbione> powerpc64-smp:CONFIG_FB=m
[11:28] <fabbione> no
[11:28] <fabbione> actually it is built
[11:28] <fabbione> ahhh i see
[11:28] <fabbione> it needs to be inline
[11:28] <fabbione> crap
[11:29] <jbailey> inline?
[11:29] <fabbione> CONFIG_FB=y
[11:29] <fabbione> instead of modular
[11:29] <jbailey> Right, but in the meantime I can modprobe it in the initrd
[11:29] <fabbione> yup
[11:29] <jbailey> It's not getting as far as getting ssh going, so there's some other issue, too.
[11:31] <fabbione> jbailey: probably there is more that needs modprobing
[11:31] <fabbione> if you can start adding fb stuff and see where it stops, that would be great
[11:31] <fabbione> i am pretty sure it's just question of config options
[11:32] <fabbione> but not having the frigging hw locally doesn't make it easier
[11:32] <jbailey> Iz no problem, boss.
[11:32] <fabbione> jbailey: thanks!
[11:34] <fabbione> jbailey: i suggest you to copy also the other FB modules
[11:35] <jbailey> Other than just radeonfb?
[11:36] <fabbione> there is a +CONFIG_FB_MACMODES=m
[11:36] <fabbione> i did it all modular, but problably it needs to go all inline
[11:41] <jbailey> Trying it with all those drivers in the initrd.
[11:42] <jbailey> No joy.
[11:43] <fabbione> crap
[11:44] <fabbione> are you actually modprobing them manually?
[11:44] <fabbione> or just hoping that they will get modprobed?
[11:44] <jbailey> No, putting them in the initrd which is set to modprobe them.
[11:44] <jbailey> I don't have enough connection to the machine at that point to modprobe anything.
[11:45] <jbailey> I don't think this box has a serial port.
[11:45] <fabbione> hmmmm
[11:45] <jbailey> On what speed box does it take 10 hours to build?
[11:46] <fabbione> jbailey: on davis
[11:46] <fabbione> but i can rebuild ppc64 only in not too long
[11:46] <fabbione> 10 hours is the whole package
[11:46] <fabbione> i am changing the config right now :)
[11:51] <fabbione> powerpc64-smp:CONFIG_FRAMEBUFFER_CONSOLE=m
[11:51] <fabbione> powerpc-smp:CONFIG_FRAMEBUFFER_CONSOLE=y
[11:51] <fabbione> yeah there is more :)
[12:00] <fabbione> ok it's building now
[12:01] <fabbione> it shouldn't take too long
[12:01] <jbailey> Cool, thanks.
[12:01] <fabbione> modulo gcc segfaults
[12:02] <fabbione> this will make the FB situtation similar to all the other ppc configs
[12:02] <fabbione> there might be more to change but at least you should be able to get the console
[12:03] <fabbione> i am going to get some lunch while it builds
[12:03] <fabbione> brb
[12:04] <fabbione> never mind.. it already failed on some drivers
[12:12] <jbailey> Cool - if I can see what's failing after this point, that would be lovely. ;)
[12:22] <fabbione> jbailey: i was thinking that perhaps mkinitrd will need ppc64 love too....
[12:22] <fabbione> what about glibc?=
[12:22] <fabbione> would a 64bit kernel search for 64 bit libc6?
[12:23] <jbailey> It shouldn't, no.
[12:24] <fabbione> jbailey: the image is up
[12:24] <jbailey> The first thing it does is fire up /linuxrc, which uses the shell script interpreter in the kernel to seek to /bin/sh, which is 32bit.
[12:24] <jbailey> Same filename as before?
[12:24] <fabbione> note that is the next version :)
[12:24] <fabbione> nope...
[12:24] <fabbione> 1.3
[12:24] <fabbione> instead of 1.2
[12:24] <jbailey> Thanks. =)
[12:25] <fabbione> no problem
[12:50] <fabbione> jbailey: new 1.3 image on people
[12:55] <fabbione> jbailey: tell me something while you are working on it.. your silence is killing me :)
[12:55] <jbailey> Lol, it's just installing.
[12:56] <jbailey> My connection to the data centre is crap.
[12:56] <fabbione> hehe ok
[12:56] <jbailey> booting.
[12:57] <jbailey> I still get a black screen
[12:57] <jbailey> Let's give it a sec and see if I can ssh in, though.
[12:57] <fabbione> crap
[12:57] <jbailey> hey
[12:58] <jbailey> X apparently works, even if the framebuffer doesn't. =)
[12:58] <jbailey> And I can ssh in now. =)
[12:58] <jbailey> I see that it's loading the radeonfb
[12:58] <fabbione> AHHHHHAHAHHAHAHAHA
[12:59] <jbailey> *lol*
[12:59] <fabbione> now...
[12:59] <fabbione> we need to understand why the FB sucks
[12:59] <fabbione> well...
[12:59] <fabbione> there is not much to understand.. it does :)
[01:00] <fabbione> jbailey: can stick dmesg output somewhere?
[01:00] <jbailey> $ file a.out
[01:00] <jbailey> a.out: ELF 64-bit MSB executable, cisco 7500, version 1 (SYSV), for GNU/Linux 2.6.0, dynamically linked (uses shared libs), not stripped
[01:00] <jbailey> jbailey@ppc64:/tmp$ ./a.out
[01:00] <jbailey> Hello, world!
[01:00] <fabbione> eheheh
[01:01] <jbailey> sure, need a couple minutes to wake Angie up.
[01:01] <fabbione> no problem
[01:01] <jbailey> And the fans appear to have gone nuts. =)
[01:02] <fabbione> i am not sure APM or PM is available yet
[01:02] <fabbione> i didn't have any option to enable it
[01:04] <jbailey> This thing displaces an amazing amount of air.
[01:04] <jbailey> Also sounds a bit like a jet engine. =)
[01:04] <fabbione> ehhe
[01:04] <jbailey> http://people.ubuntu.com/~jbailey/dmesg.txt
[01:05] <fabbione> [   32.536434]  Console: colour dummy device 80x25
[01:05] <jbailey> Want one from the working setup too?
[01:05] <fabbione> isn't this one from the working ppc64 kernel?
[01:06] <jbailey> I mean from the working ppc32
[01:06] <fabbione> [   32.614271]  Calibrating delay loop... 66.56 BogoMIPS (lpj=33280)
[01:06] <fabbione> interesting
[01:06] <fabbione> yes.. later is fine...
[01:06] <jbailey> I may as well fetch it now to make the 747 go away...
[01:06] <fabbione> how much RAM do you have in that system?????
[01:07] <jbailey> Umm, Looks like 4gb?
[01:07] <fabbione> [   32.763094]  PowerMac SMP probe found 2 cpus
[01:07] <fabbione> uhuuh
[01:07] <fabbione> it's even SMP :)
[01:07] <jbailey> Yes
[01:07] <fabbione> jbailey: [   32.614139]  Memory: 3999720k/6291456k
[01:07] <fabbione> it's more like 6GB
[01:07] <jbailey> $ cat /proc/meminfo
[01:07] <jbailey> MemTotal:      4007812 kB
[01:07] <fabbione> but our kernels don't have more than GB highmem enabled
[01:08] <jbailey> Ah.
[01:08] <fabbione> yeah what's in meminfo comes from what the kernel can actually use :)
[01:09] <fabbione> we need to talk with benh about the FB stuff
[01:09] <fabbione> he hangs on debian-kernel
[01:11] <fabbione> he is not around
[01:13] <jbailey> On ppc32: [   25.954678]  Console: colour dummy device 80x25
[01:13] <jbailey> [   26.045504]  Calibrating delay loop... 1662.97 BogoMIPS (lpj=831488)
[01:15] <jbailey> On the ppc32 there's an extra:
[01:15] <jbailey> [   27.434522]  Console: switching to colour frame buffer device 210x65
[01:15] <jbailey> in there.
[01:16] <fabbione> hmmm
[01:24] <zul> he3y
[01:25] <jbailey> zul: Know anything about framebuffers? =)
[01:25] <zul> no :)
[01:29] <jbailey> Hmm, there's an fbset package.
[01:29] <fabbione> jbailey: that's not what i am looking for
[01:29] <jbailey> But I guess the problem might be making the console point to it?
[01:29] <fabbione> i can't find that exact string
[01:29] <fabbione> so it's generated at runtime
[01:30] <fabbione> that means that we need to find what driver spits it
[01:30] <chmj> jbailey: issint fbset a fluxbox util
[01:30] <fabbione> and see why in 32 bit works and not in 64
[01:30] <jbailey> chmj: No idea. =)
[01:31] <chmj> jbailey: bah, framebuffer device maintanance program
[01:35] <fabbione> ah here is
[01:35] <jbailey> ben's on.
[01:39] <chmj> fabbione: can I give you the watch proggie ?
[01:39] <fabbione> chmj: i am waiting for it...
[01:39] <zul> damn ties
[01:41] <chmj> fabbione: email ?
[01:41] <fabbione> chmj: or commit it in baz?
[01:42] <chmj> ok 
[01:42] <fabbione> chmj: you have access to people, don't you?
[01:42] <chmj> yes 
[01:42] <fabbione> so just use sftp instead of http to checkout/commit
[01:42] <fabbione> put it inside the debian/ dir
[01:47] <chmj> must I do that from chainstrap ?
[01:47] <fabbione> nope
[01:47] <fabbione> ssh to people is allowed from outise via ssh proxy
[01:48] <chmj> ssh proxy ?
[01:48] <fabbione> chmj: check the wiki
[01:48] <fabbione> there is the doc on how to set that up
[01:59] <fabbione> bbl
[03:01] <chmj> there goes my first kerne-team commit :) 
[03:04] <fabbione> ./uscan-extmod --report .
[03:04] <fabbione> Processing ./kernel-extmod.kwatch ......
[03:04] <fabbione> uscan-extmod warning: could not open ./kernel-extmod.kwatch: No such file or directory
[03:05] <fabbione> A   passed.kwatch
[03:05] <fabbione> what's that?
[03:05] <chmj> usca-extmod --wathifile = passed.kwatch 
[03:05] <chmj> uscan-extmod --wathifile passed.kwatch 
[03:06] <chmj> the wathfile by default will be kernel-extmod.kwatch, one I got it looking pretty
[03:06] <chmj> s/one/once/
[03:06] <fabbione> ehhe nice :)
[03:07] <chmj> the script, or my typing ? 
[03:08] <fabbione> the script :)
[03:08] <chmj> :) 
[03:08] <fabbione> chmj: did already convert all the entries in external-drivers to the watch file?
[03:09] <fabbione> or only a few for sampling?
[03:09] <chmj> a lot 
[03:09] <fabbione> all of them or only a selection?
[03:09] <fabbione> before we spend time updating that file to death...
[03:09] <chmj> only a section, 
[03:09] <fabbione> ok
[03:09] <chmj> I'm trying to get the rest off the URLs 
[03:10] <fabbione> chmj: ok. you tell me when it's ready for "deplyoment", but it looks good already :)
[03:11] <chmj> did I mention I hate "You don't have permission to access <insert driver location folder here> on this server."
[03:11] <fabbione> ehehhe
[03:13] <jbailey> So who thinks that Apple using "intel chips" might mean itanium? =)
[03:13] <zul> ok wish me luck
[03:13] <jbailey> zul: Good luck, chuck!
[03:16] <fabbione> zul: come back as winner!
[03:16] <jbailey> fabbione: fbcon working right.
[03:16] <fabbione> jbailey: cool
[03:17] <jbailey> fans haven't gone nuts.
[03:17] <jbailey> x is up
[03:17] <fabbione> jbailey: and yes.. i am sure apple will go ia64 :)
[03:17] <chmj> zul: what u need luck for ? 
[03:17] <fabbione> goody
[03:17] <jbailey> ssh in.
[03:17] <jbailey> "...  and print!"
[03:23] <fabbione> chmj: you broke the baz archive!
[03:23] <fabbione> you need to fix your umask on people
[03:23] <fabbione> and change some permissions in the baz repo
[03:23] <fabbione> otherwise we are all doomed
[03:24] <fabbione> chmj: login to people and do what i say
[03:24] <fabbione> cd /home/lamont/public_html/Archives/kernel-team@ubuntu.com--2005/kernel-debian--pre1,3--2.6.11.93
[03:24] <fabbione> chmod 775 patch-4
[03:25] <fabbione> cd patch-4
[03:25] <fabbione> chmod 775 ++revision-lock
[03:25] <fabbione> cd ++revision-lock
[03:25] <fabbione> chmod 775 +contents
[03:26] <jbailey> Not find a chmod -R g+w 
[03:26] <jbailey> s/find/just/
[03:26] <fabbione> jbailey: it's the same....
[03:26] <jbailey> ?
[03:27] <fabbione> chmj: once you have done go back to your home and:
[03:28] <fabbione> edit .bash_profile
[03:28] <fabbione> change the umask to look like:
[03:28] <fabbione> umask 002
[03:28] <fabbione> then edit .bashrc and add the same umask line
[03:29] <fabbione> logout
[03:29] <fabbione> and login again
[03:29] <fabbione> that will fix
[03:30] <chmj> done 
[03:30] <fabbione> chmj: you need to change the permissions to all the dirs i told you
[03:30] <chmj> I did 
[03:30] <fabbione> nope
[03:31] <fabbione>  /home/lamont/public_html/Archives/kernel-team@ubuntu.com--2005/kernel-debian--pre1,3--2.6.11.93/
[03:31] <fabbione> 4 drwxr-xr-x   3 charles  warthogs 4096 Jun  6 13:57 patch-4
[03:31] <fabbione> that needs to be 775
[03:31] <fabbione> same for the directories inside
[03:32] <chmj> recursively ?
[03:33] <fabbione> chmj: you are missing the ++revision-lock directory inside patch-4/
[03:33] <fabbione> the others are or seem to be ok
[03:36] <chmj> ok, done 
[03:36] <fabbione> ok thanks
[03:37] <fabbione> yup
[03:37] <fabbione> did also change your netmask?
[03:37] <fabbione> anyway i need to go offline again
[03:37] <fabbione> bbl
[03:39] <jbailey> fabbione: s/netmask/umask/ ?
[03:40] <chmj> erm,bashrc and bash_profile, done 
[03:51] <fabbione> yeah that :)
[04:37] <jbailey> O
[04:37] <jbailey> I'm trying to imagine what the ugliest scenario the initramfs might have to cope with for booting, and whether I care.
[04:38] <jbailey> I'm adding the hooks in, but obviously somethings might care what order they're run in, like lvm before md, and cryptroot after both.
[04:49] <mjg59> Any idea when the -1.2 builds are likely to be done on x86?
[04:49] <jbailey> None.  And Fabio may have sent up -1.3 with the ppc64 changes from this morning.
[04:50] <mjg59> Ah
[04:50] <mjg59> Fair enough
[05:52] <fabbione> no i didn't push 1.3 to archive.
[05:52] <fabbione> 1.2 needs NEW love
[05:52] <fabbione> and i386 is FTBFS
[05:53] <mjg59> Ah, ok
[05:53] <fabbione> hell it did build here!
[05:53] <fabbione> Inconsistent kallsyms data
[05:53] <fabbione> Try setting CONFIG_KALLSYMS_EXTRA_PASS
[05:53] <fabbione> BAH
[05:53] <fabbione> make[2] : Leaving directory `/build/buildd/linux-source-2.6.12-2.6.11.93/debian/b
[05:53] <fabbione> uild/build-686-dbg'
[05:54] <fabbione> well this is tomorrow's stuff
[09:00] <lamont_r> fabbione: linux-source-2.6.12 won't build on hppa until expect-tcl8.3 arrives in the archive and gcc-3.4 is buildable
[09:03] <doko> lamont: expect-tcl8.3 was made only for you^H^H^Hhppa ;)
[09:04] <lamont_r> doko: yeah
[09:05] <lamont_r> but it's not in the ubuntu archive.  hence gcc-3.4 is pending...
[09:05] <doko> so build it, don't grumble ;-)
[09:05] <lamont_r> another option would be to disable the tests on hppa... :)
[09:05] <doko> it's not?
[09:05] <lamont_r> ah, is in universe
[09:05] <lamont_r> coolness
[10:02] <zul> well round one done
[10:11] <zul> miss a day...miss alot