[02:42] <poolie> hi, is anyone around?
[02:43] <poolie> i'm really blocked from using one machine by bug 90689
[02:43] <poolie> would love some help
[02:43] <crimsun> bug 90689
[03:22] <Amaranth> ubugtu isn't here
[03:22] <Amaranth> why not?
[06:20] <poolie> fabbione: ok, interesting 
[06:20] <poolie> now i have
[06:20] <poolie> ata1: failed to recover some devices, retrying in 5 secs
[06:21] <poolie> and again, with 
[06:21] <poolie> qc timeout, failed to set xrefmode err mask 4 
[06:22] <fabbione> ok...
[06:22] <fabbione> was that with or without options?
[06:23] <fabbione> you could also try to add: pci=routeirq
[06:23] <poolie> that was 9, which is currently installed, with nosmp noreplacement
[06:24] <poolie> do you want me to try the dvd, or the installed image with those options?
[06:25] <fabbione> before you change kernel, please try to add pci=routeirq too
[06:25] <fabbione> then jump to 2.6.20-10 or -11
[06:25] <fabbione> (not sure -11 has been uploaded yet)
[06:25] <fabbione> start without options
[06:25] <fabbione> and add them one at a time to see which one changes behavioy
[06:26] <fabbione> behaviour
[06:26] <fabbione> i suggest you also take note of these tests and add the results to the bug
[06:26] <poolie> ok
[06:26] <poolie> good idea
[07:58] <dade`> the new kernel does not even boot on my macbook
[07:59] <dade`> gives an ata bad status when setting the xfer mode
[08:41] <\sh> guys, dapper, edgy and feisty herd5 have problems booting on a HP DL320s server
[08:42] <\sh> it always stops loading the kernel after recognizing the PS/2 Keyboard port...no oops nothing...how can I try to gather more debug information for a bug report? :)
[08:44] <\sh> funny thing, too, booting an iso via ilo2, doesn't bring this lock, but runs into an endless loop for udev finding many times the same usb hub for the virtual mouse/keyboard of ilo2...(dapper)
[08:55] <poolie> \sh: assume you've tried removing 'quiet' and adding 'verbose' to the boot parameters?
[08:58] <\sh> poolie: yepp
[08:59] <fabbione> BenC: latest git fails on sparc:
[08:59] <fabbione> ld: cannot open linker script file ubuntu/net/wireless/.tmp_wext-common.ver: No such file or directory
[08:59] <fabbione> make[5] : *** [ubuntu/net/wireless/wext-common.o]  Error 1
[08:59] <fabbione> make[4] : *** [ubuntu/net/wireless]  Error 2
[08:59] <fabbione> make[3] : *** [ubuntu/net]  Error 2
[08:59] <fabbione> make[2] : *** [ubuntu]  Error 2
[08:59] <fabbione> make[2] : *** Waiting for unfinished jobs....
[09:00] <poolie> fabbione: hi, still no luck with the current -10
[09:00] <fabbione> poolie: did you try all the options*?
[09:00] <poolie> fabbione: where do the device files come from in initramfs? i have no sda1
[09:01] <poolie> just going to try pci=routeirq
[09:01] <fabbione> poolie: they are generated by udev
[09:01] <fabbione> and there is a bug about it
[09:04] <poolie> ok, it also oopses with pci=noroute added
[09:05] <poolie> but i also get, earlier, mdadm: No devices listed in conf file were found. / failed to assemble all arrays 
[09:05] <poolie> which i guess is why it can't mount root, which is on raid
[09:07] <\sh> fabbione: do you know anyone from HP who is an ubuntu geek, best location uk or europe in general? :)
[09:07] <fabbione> poolie: cat /proc/partitons .. does it show the devices?
[09:08] <poolie> it's empty
[09:08] <fabbione> \sh: try one of our sysadmins...
[09:08] <fabbione> poolie: ok if that's empty there is no way to assemble a raid out of thin air... it means that disks have not been found at all
[09:08] <poolie> right
[09:08] <\sh> fabbione: elmo or znarl?
[09:08] <fabbione> and that's mostlikely related to the block subsystem OOPSing
[09:09] <fabbione> \sh: any of the sysadmins
[09:09] <poolie> that's what i thought
[09:09] <fabbione> poolie: we need to wait for ben/kyle
[09:09] <fabbione> i am not sure what else to try
[09:09] <fabbione> feh
 poolie: we need to wait for ben/kyle
[09:09] <fabbione> * poolie (n=mbp@ppp112-44.static.internode.on.net) has left #ubuntu-kernel ("parted")
 i am not sure what else to try
 feh
[09:09] <poolie> fabbione: ok, at least i have a laptop
[09:09] <poolie> thanks 
[09:09] <fabbione> no problem
[09:09] <poolie> i'm too used to ^w for delete-word :)
[09:10] <poolie> fabbione: strangely enough the current daily cd does boot just fine in recovery mode, brings up md and lvm automatically
[09:11] <fabbione> poolie: hmmm so you mean it boots but then install doesn't?
[09:12] <lifeless> well he has an install :)
[09:12] <lifeless> poolie: try chrooting into your system now
[09:14] <poolie> from the boot cd
[09:14] <poolie> ok, will take a while to come back up
[09:14] <lifeless> poolie: you could also try setting root= during the boot
[09:15] <poolie> to what?
[09:15] <poolie> oh i see
[09:15] <poolie> while booting from cd 
[09:15] <poolie> yep worth a shot
[09:22] <poolie> i wonder why the recovery mode needs you to set a hostname?
[09:39] <poolie> oh ffs
[09:43] <fabbione> BenC: that linker script error seems to show up only when forking the build with CONCURRENCY_LEVEL
[09:47] <poolie> lifeless: i'm chrooted into my root from the installer kernel, what did you want me to try?
[09:55] <lifeless> well you should be able to upgrade your kernel, udev, now
[09:56] <poolie> yeah, that's what i did
[09:56] <poolie> so the good news is i have a way to get updates on; but they don't fix it yet
[09:56] <lifeless> ok
[09:56] <lifeless> so the installer has the same kernel as you now, and udev. So it might be another package
[09:57] <poolie> well, the installer actually uses -9, which is quite a bit older the current one in the distro
[09:58] <poolie> or it may be starting things up in a different order from userspace that avoids this
[09:58] <poolie> but for now i will wait for guidance from benc i think
[09:59] <Mithrandir> poolie: try installing -9 from the repositories and see if that boots correctly?
[10:01] <poolie> i have, though maybe they updated -9?
[10:01] <poolie> do you know when it gets a new -9 number rather than a -9.1?
[10:03] <lifeless> yes
[10:03] <lifeless> the -X is the ubuntu serial number for all intents and purposes
[10:03] <Mithrandir> poolie: grab the correct set of binaries off https://beta.launchpad.net/ubuntu/+source/linux-source-2.6.20 ?
[10:04] <poolie> lifeless: but sometimes there are several uploads of -9?
[10:05] <poolie> Mithrandir: will that be different to what's in the archive?
[10:05] <Mithrandir> poolie: if you grab the one marked "removed", yes.
[10:06] <Mithrandir> poolie: -9 is the ABI version, it's only bumped when the upload changes the ABI.
[10:06] <Mithrandir> there was two uploads of -9
[10:06] <poolie> i see, thanks 
[10:07] <poolie> Mithrandir: and which do you think is correct?
[10:07] <poolie> it's broken in -7; it's broken in -10
[10:08] <Mithrandir> 2.6.20-9.16, probably.
[10:10] <poolie> ok, and you think that might work because it's the one used by the installer?
[10:10] <Mithrandir> yes
[10:10] <poolie> ok thanks
[12:25] <Nafallo> LP #92202 might be a beta target? :-)
[12:48] <Mithrandir> it should be easy enough to fix, at least.
[12:48] <Mithrandir> BenC: ^^ ; are you planning a kernel upload before beta freeze?
[02:23] <BenC> Mithrandir: planning one in about 1 hour
[02:23] <Mithrandir> ok
[02:23] <Nafallo> BenC: with qc-usb? :-)
[02:24] <BenC> Nafallo: Yeah, I got 92202 up and I'll get the source in so long as it builds :)
[02:24] <Nafallo> kewl! :-)
[02:33] <fabbione> BenC: do we have the atyfb fix for davem in the tree=?
[02:34] <BenC> fabbione: Yeah, I think I put it in after -10, so it will be in -11
[02:34] <fabbione> BenC: yeah it's there
[02:35] <fabbione> BenC: i am still a bit puzzled by that sparc FTBFS with CONCURRENCY_LEVEL
[02:35] <fabbione> it was reproducible only one
[02:35] <fabbione> once
[02:35] <BenC> fabbione: All of my sparc test builds are done with -j10
[02:36] <fabbione> hmm ok
[02:36] <fabbione> i did fire -j128
[02:36] <fabbione> but i am ok if it works for you
[02:36] <BenC> we'll blame that on you being plain evil to your boxes :)
[02:39] <fabbione> nah evil was -j4096 on davem's machine :)
 it works! it works at -j32!
 gimme a sec..
 sure
 OOPS.. it crashes at -j4096...
 screw you
[03:16] <BenC> fabbione: lol
[03:16] <BenC> fabbione: Tim and I did a plain -j on a 64Gig 2 x Quad Xeon box
[03:17] <BenC> some 10k processes
[03:17] <BenC> only used like 7gigs of mem
[03:17] <BenC> 4.5 minutes for bzImage AND ~1800 modules, MODPOST and all
[03:18] <BenC> doing -j12 actually finished 4 seconds faster
[03:18] <BenC> fabbione: let that be a lesson to your hw abusive ways :)
[03:31] <abogani> Mithrandir: May i disturb you?
[03:55] <Mithrandir> abogani: sure, what's up?
[03:56] <abogani> Mithrandir: About my spec (https://wiki.ubuntu.com/RealTime)....
[03:56] <abogani> I have created a linux-patch-realtime-preempt package. This package contain only the Ingo Molnar's patch (and not all linux kernel source) adapted to Ubuntu linux kernel (i use dh-kpathes method). Users can build kernel follow this http://kernel-handbook.alioth.debian.org/ch-common-tasks.html#s-common-getting     
[03:57] <abogani> It is a right way? Now exist a possibility that this package will be accepted in universe? What i should do for achieve this objective?
[04:04] <BenC> abogani: get someone to review it, and sponsor the upload, then buy Mith a few pints
[04:05] <BenC> abogani: That's something you'd want to check #ubuntu-motu for
[04:07] <Mithrandir> abogani: your approach seems sane, yes.
[04:07] <abogani> BenC, Mithrandir : Ok Thanks :-)
[04:09] <BenC> abogani: and if it helps, I'll give a nod for the package, considering you were very open to packaging suggestions and quick to change the entire build process to suit it
[04:10] <Amaranth> wouldn't it be better to do like Xen and have the package pull in the kernel source, apply the patch, and build a kernel for the binary package?
[04:10] <BenC> Amaranth: That's what he did
[04:10] <Amaranth> oh, i thought he was saying the end user would have to build the kernel
[04:10] <BenC> oh, maybe he did, I misread
[04:10] <BenC> I suggested using the xen method yesterday :)
[04:33] <abogani> BenC: I already seen Xen packaging approach.. but xen-source package contains a complete linux source copy! It is also _vanilla_ kernel! ...... Ok i'm confused. :-| .... What is exactly the xen package that i should use as example?
[04:34] <BenC> abogani: The 2.6.17 one in edgy
[04:34] <BenC> the one in feisty might be full source
[04:34] <abogani> BenC: Ahhhh
[04:35] <Mithrandir> uh, please don't do that.  It's deprecated and I will reject any new packages doing that.
[04:36] <Mithrandir> you're free to have a package that build-depends on the appropriate kernel source, includes a patch and builds linux-image-$whatever binaries, though
[04:36] <abogani> BenC: thank you for the patience :-)
[04:37] <BenC> Mithrandir: that's the approach he's working towards
[04:37] <Mithrandir> BenC: ok, good.
[05:28] <fabbione> BenC: the kernel doesn't fork more than 3500 or there about... there isn't enough source code :)
[05:37] <BenC> fabbione: We went from 120 procs to 10100 or so
[05:37] <BenC> top verified this number :)
[05:38] <BenC> fabbione: this was full ubuntu tree, so almost all modules turned on
[05:38] <BenC> $ find -name \*.c | wc -l
[05:38] <BenC> 9198
[05:39] <BenC> figure up gcc+cc1+as forking, plus some make procs
[05:39] <BenC> seems feasible
[05:48] <fabbione> hmmm oh right it was crap86
[05:48] <fabbione> we were on sparc
[06:22] <BenC> fabbione: I don't think sparc uses -pipe, so there's more forking on x86 builds
[09:15] <TheMuso> c
[09:15] <TheMuso> damnit