/srv/irclogs.ubuntu.com/2005/09/11/#ubuntu-kernel.txt

BenCdesrt: any idea how to fix it?12:09
desrtnope!12:09
desrti was hoping you would :)12:10
BenCok, I'll check into it then :)12:10
desrti don't know thing 1 about acpi inside the kernel12:10
BenCme either, but never too late to learn I guess12:10
desrtwell12:10
desrta week into october, for example, would be too late to learn :)12:10
jbaileyBenC: Cool, thanks.12:11
=== fs [i=fs@213.178.77.98] has joined #ubuntu-kernel
=== TheMuso [n=luke@dsl-202-173-132-131.nsw.westnet.com.au] has joined #ubuntu-kernel
=== chuck__ [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel
=== chuck__ is now known as zulx
=== zulx is now known as zul
zulevening03:50
=== jorgp2 [n=jorgp@bnet-dial-123.bartnet.net] has joined #ubuntu-kernel
=== chmj_ [n=chmj@wbs-146-147-243.telkomadsl.co.za] has joined #ubuntu-kernel
fabbionemorning06:17
jbaileyfabbione: Hey!06:19
jbaileyFeeling better?06:19
fabbioneyeah a little bit06:19
jbaileyCool.  Sucks to be sick.06:20
fabbionei need to do the Sucks Of the Code evaluation today06:20
jbailey*lol*06:20
fabbionethat's why i am back..06:20
fabbionebut i might disappear again06:20
fabbionei still feel like half shit06:20
jbaileyAnd I was hoping it was to see my smiling fave.06:20
jbaileyface.06:20
fabbioneahaha06:20
=== doko_ [n=doko@dsl-084-059-069-062.arcor-ip.net] has joined #ubuntu-kernel
=== JaneW [n=JaneW@wbs-146-144-158.telkomadsl.co.za] has joined #ubuntu-kernel
=== ubuntulog [n=warthylo@port49.ds1-van.adsl.cybercity.dk] has joined #ubuntu-kernel
=== Topic for #ubuntu-kernel: Ubuntu kernel development discussion ONLY | http://www.ubuntulinux.org/wiki/KernelTeam | There are no kernel bugs.. only broken hardware | http://people.u.c/~lamont/Archives/kernel-team@ubuntu.com--2005/ playground: kernel-debian--preX,13--2.6.12
=== Topic (#ubuntu-kernel): set by BenC at Wed Aug 31 23:13:54 2005
=== mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel
=== doko [n=doko@dsl-084-059-069-062.arcor-ip.net] has joined #ubuntu-kernel
=== mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel
=== spayne [n=spayne@i-195-137-120-148.freedom2surf.net] has joined #ubuntu-kernel
=== Seveas [n=seveas@ksl403-uva-132.wireless.uva.nl] has joined #ubuntu-kernel
=== spayne [n=spayne@i-195-137-120-148.freedom2surf.net] has joined #ubuntu-kernel
=== doko [n=doko@dsl-084-059-069-062.arcor-ip.net] has joined #ubuntu-kernel
=== JaneW [n=JaneW@wbs-146-144-158.telkomadsl.co.za] has joined #ubuntu-kernel
=== chmj [n=chmj@wbs-146-147-243.telkomadsl.co.za] has joined #ubuntu-kernel
=== jorgp2 [n=jorgp@bnet-dial-123.bartnet.net] has joined #ubuntu-kernel
=== BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-kernel
=== dilinger [i=dilinger@mouth.voxel.net] has joined #ubuntu-kernel
=== jbailey [n=jbailey@testhaus.cns.utoronto.ca] has joined #ubuntu-kernel
=== lamont [n=lamont@mix.mmjgroup.com] has joined #ubuntu-kernel
=== mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel
=== zul [n=chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel
zulheylo02:38
BenCyo03:04
jbaileyHey'all.03:05
jbaileyzul: Hey - You mentioned in a bug report that you had the bits to make VmWare and initramfs  happy together.  Do you know when those are going to land?03:05
zuljbailey: i got a bug in my code that i have work on tonight03:06
jbailey'k03:06
zulie: buslogic doesnt compile with my code change :)03:07
jbaileyThere are a surprising number of people who seem to run Breezy in vmware.  What I don't understand so far is how it worked before.03:07
jbaileyOr I wonder if they just looked at lspci directly or something.03:07
jbaileyI don't understand all of what generates the modalias files.  All I know is that they're lovely.03:07
zuljbailey: buslogic doesnt have hotplug so thats why its not detecting it i believe03:08
jbaileyHmm, but how did hotplug discover it before then?03:09
jbaileyI tried looking through hd-detect to see if there were crazy corner cases, but I didn't see any.03:09
zuli dont think it ever did because i seen some bug reports in hoary as well but fabbione closed the bugs saying that vmware isnt supported03:11
BenCjbailey: I installed breezy in vmware aswell03:14
jbaileyzul: Those seem to be installing vmware in breezy.03:15
BenCcould have been hoary, but I believe it was breezy03:15
BenCeither way, it worked at some point03:15
jbaileyWell, it ought to work, and hopefully with Chucks hack it will.03:15
BenCjbailey: I haven't had time to look at your script, but can you explain to me what it is supposed to do?03:16
BenCor what it is supposed to help do I guess03:17
jbaileyBenC: Sure, although it doesn't much matter - mdz deferred it to post-preview.03:17
zulBenC: with buslogic?03:17
jbaileyThe problelm I'm trying to solve is that when someone installs usplash, it should update the initrd.img file so that it goes in.03:17
jbaileyI've taken a generic approach to the problem, which is that anything that lives in the initrd.img file ought to have the ability to force it to be updated.03:18
jbaileySpecifically, that means usplash, lvm, evms, mdadm, or even initramfs itself.03:18
jbaileyMy issue is that I don't want to:03:19
jbailey1) Risk a non-bootable system.03:19
jbailey2) Overwrite a users change to their initrd.img file.03:19
jbaileyHowever, I have limited control over it - I didn't create the initrd.img-* files, I don't delete them, I merely provide the tool that others call.03:19
jbaileyThis script is basically a wrapper that allows me to manage it.03:20
jbaileySo  rather that calling mkinitramfs directly, the kernel should call update-initramfs -c.  Then I've created it.  I know what it's called, I have time to take a sha1sum.03:20
jbaileyWhen the kernel gets purged, it should call update-initramfs -d03:20
jbaileyAnything that wants it to be updated should call update-initramfs -u03:20
jbaileyYou need to generally specify a version number for create and delete.03:21
jbaileyYou *can* specify one for update, but otherwise it'll pick the default one (mdz wants the one pointed to by the initrd.img symlink03:21
jbaileyI then track things like, has it been changed?  Did this tool make it?03:21
BenCzul: the vmware install I use was loading some sort of fibre channel scsi controller for the disks03:22
zulhmmm...03:23
BenCjbailey: what if it's called several times during the course of a dist-upgrade, for example?03:24
zuli am pretty sure its buslogic though03:24
jbaileyBenC: Then in the current implementation, speed will suck, but not dramatically.03:25
jbaileyBenC: For breezy, I'd honestly only expect usplash to call it.  Some sort of delayed rebuild or something like that could always happen after.03:26
jbaileyBenC: Longer term, I'd like the kernel to create initramfs overlays that can just be linked in to help with the speed problem.03:27
BenCjbailey: yeah, I was thinking something that kept checking for the dpkg lock to clear, and hold a lock of it's own (to prevent multiple invocations)03:27
jbaileyUgh.  You're thinking parallel postinsts running? =(03:27
jbaileyOh.03:27
BenCno03:28
jbaileyOr if the user runs it by hand.03:28
jbaileySuck.03:28
BenClike, if 100 .deb's are installed and say 4 of them affect initrd, then the update should only occur once at the end03:28
jbaileyRight.03:29
jbaileyOr we can ask for a dpkg hack  for "do this once at the end"03:29
BenCplus, you don't want to be able to call the program manually during an upgrade03:29
jbaileyscrollkeeper need sthis too.03:29
BenCapt-get probably03:29
BenCpost-install check of some sort03:29
jbaileyIn fact, scrollkeeper sucks way worse for this.  At least with us it'll only get run 4 or 5 times at the *maximum*.  Gnome does it 30 or 40, IIRC.03:29
BenCprobably do something like if dpkg lock is held just have initramfs-update touch some file03:30
BenCadd initramfs-update (--finish?) to apt-get as a post-install script and let it check for the file03:30
BenCsimilar to how debconf stuff works pre-install with apt03:31
jbaileyI've just asked keybuk to come visit to tell us the One True Way. =)03:31
=== Keybuk [n=scott@syndicate.netsplit.com] has joined #ubuntu-kernel
BenCzul: Fusion MPT is what my vmware used for the disks03:32
Keybukhey03:32
BenCcdrom was via IDE03:32
jbaileyKeybuk: We're just discussing the problem that someone updating something multiple times in postinsts really sucks (like scrollkeeper, or initramfs-tools)03:32
jbaileySince initramfs-tools takes several seconds to run, it's annoying if it has to hapen several times in a dist-upgrade.03:32
BenCor even ldconfig for that matter03:32
jbaileyRight.03:32
Keybukindeed03:33
jbaileyIs there a Canonical way in dpkg / apt-get to say "Do this once, last"03:33
Keybuknope03:33
jbaileyAnd if not, could there be please? =)03:33
Keybukthere's two proposals for this03:33
BenCI was thinking of (ab)using apt-get's post/pre hooks03:33
Keybuk1) put "dpkg-trigger ldconfig" in postinst03:33
Keybuk  at the end of the run dpkg runs /etc/dpkg/triggers/ldconfig03:33
zulBenC: ah...well Fusion MPT has hotlplug though03:33
BenCzul: do different versions of vmware use different disk controllers?03:34
Keybuk2) libc6 declares that "if any package installs something into /usr/lib, run /var/lib/dpkg/info/libc6.trigger on completion"03:34
zulyeah i believe so i have vmware open right now gimme a sec03:34
jbaileyKeybuk: I like the trigger idea, since it leaves it in the control of the package, not as a side-effect.03:34
BenCKeybuk: I like the first idea better03:34
zulvmware 4 has buslogic and lsi logic for scsi adapter03:35
BenCzul: ah, I have v503:35
zul4 here..03:35
zuli think most people use 4 though 03:35
BenCKeybuk: would Debian be willing to accept either of those?03:36
BenCif not, it would make things seriously incompatible (especially if ldconfig uses it)03:36
KeybukI only ever talk about dpkg stuff in terms of Debian :)03:37
Keybukit's a spare-time thing03:37
BenCunless of course dpkg-trigger was self standing03:37
jbaileyKeybuk: It's important enough for Ubuntu.  Can we ask to get some time assigned to it?03:37
BenCbut then you lose the concept of "when dpkg quits, everything is done and the system can be rebooted"03:37
jbaileyWell, it could terminate by saying "release the triggers now", basically.03:38
Keybukwell, that's not true anyway03:38
Keybukapt has to run dpkg --configure -a  ;)03:38
BenCwell "when apt or dpkg quits" :)03:38
BenCpersonally I prefer using apt's hooks, but that can cause a lot of other problems if ppl don't use apt to install a package03:39
BenCzul: when you say that they don't support hotplug, how does that cause a problem? The vmware disks aren't hotplug, right?03:40
Keybukat the moment I don't really have a plan whether triggers would be 1.13 or 2.0 material03:40
jbaileyBenC: The issue I have with relying on apt-get's hooks is that I occasionally download a series of debs and dpkg -i them together. =)03:40
zulBenC: when i went through the buslogic code they didnt have the struct for the pci_id 03:41
zulmine problem is i just put it in the wrong place.03:42
BenCzul: ah, you mean for the host controller module itself to be loaded, ok03:42
Keybukmy most common pair of operations is dpkg --unpack and apt-get -f install03:42
zulBenC: exactly03:43
Keybukwhich would work with apt hooks, I guess03:43
BenCzul: shouldn't be hard to patch that up03:43
zulBenC: no it wasnt..03:43
zuljust have to make sure it compiles when i get home tonight03:44
BenCKeybuk: what's the plan on making sure that packages that would call dpkg-trigger are not installed by a dpkg that doesn't have that program?03:47
Keybuknone yet03:50
Keybukthat'd depend whether it was a 1.13 or 2.0 issue03:50
Keybuk1.13 would require a dpkg depends03:50
jbaileyProbabyl same as the epoch test?  If you use it, call a dpkg flag and if it sucks, then bail?03:50
jbaileyKeybuk: If we assume for the moment that none of us will be alive when 2.0 hits the streets....03:50
=== jbailey hides.
Keybuk<g>03:51
zulbah...i spent 3 hours on this yesterday on my day off and then they change the data on me03:53
zul...and on my birthday even03:53
BenChappy birthday03:58
zulthanks man04:00
zuli won 85 bucks at the casino last night04:01
BenCwhich casino?04:03
zulcasino de lac lemay04:03
BenCwhere do you live?04:03
zulottawa,ontario so it was just across the river04:04
=== BenC is going to Atlantic City this weekend
zul3 hours from jbailey 04:04
BenCBorgata is having the World Poker Tour event this month, and I'm playing in the $500 NL04:04
BenCzul: do they have poker at the casinos up there? I couldn't find any in montreal that had poker (slots and blackjack was all they showed)04:07
zulBenC: yeah the have grand prix poker and carribean, no texas holdem04:09
zuli wish they have texas holdem though :(04:09
BenCyou play any online poker?04:11
zulyep i am quite good at it as well 04:12
BenCwhich sites do you usually?04:12
zuljust pokerstars.net and sometimes yahoo04:13
BenCI play pokerstars, partypoker, doylesroom, but mostly I play on fulltiltpoker.com04:13
BenCwe'll have to do a headsup match one day :)04:14
zulwe should...but i would kick your butt ;)04:14
BenCheh, you've underestimated me :)04:14
BenCmy winnings for tournaments and cash games exceeds about $25k right now04:14
BenC(but that's counting the $12500 buy-in for the foxwoods wpt even that I won)04:15
zulok...maybe not :)04:15
BenCnah, I'm not that good, just play a lot of tournaments04:15
BenCI like tournies more than straight cash games04:16
zulyeah so do i 04:16
zulbut there isnt any real tournaments in ottawa 04:16
BenCbest I ever did at cash was the week I made $9k on doylesroom playing $20/$40 limit04:16
zulnot bad04:16
zuli have to go online to get my fix04:16
BenCme too04:16
BenCatlantic city is 6 hours away04:17
BenCwhat's your nick on pokerstars? I'll see if I can find you online sometime04:18
=== lamont__ [n=lamont@15.238.5.156] has joined #ubuntu-kernel
BenCyou should check out fulltilt sometime too, nice site04:18
zulBenC: chuck_sh4504:23
=== TheMuso [n=luke@dsl-202-173-132-131.nsw.westnet.com.au] has joined #ubuntu-kernel
zulhow original isnt it :)04:23
TheMusoI/c04:24
=== Seveas [n=seveas@seveas.demon.nl] has joined #ubuntu-kernel
=== TheMuso [n=luke@dsl-202-173-132-131.nsw.westnet.com.au] has joined #ubuntu-kernel
jbaileyzul: Go to the reservations up north. =)04:44
fabbionehey guys04:46
fabbioneBenC: you got mail.. that stuff is pretty urgent04:46
fabbionei think most of it is in .12.6 tho04:46
BenCwhen did you send it?04:47
BenCoh, that thing to Herbert and I?04:47
fabbionepitti did04:47
fabbioneyes04:47
BenCI couldn't decrypt it04:48
BenCused the wrong pgp key, I think04:48
fabbioneit's @debian.org04:48
BenCyeah, that key has been lost, I have a new one for ben.collins@ubuntu.com04:48
fabbioneBenC: ah..04:49
BenCI need to do my revocation04:49
fabbioneyou should tell to pitti04:49
BenCbut I need to find a floppy drive to do it04:49
fabbioneso he will resend04:49
BenCI did04:49
fabbioneperfect04:49
BenCbut haven't gotten anything back yet04:49
fabbioneBenC: do you still have any sparc hw around?04:55
fabbioneperhaps you can try/test the sparc port :)04:55
fabbionewe have some problems with initramfs/klibc04:55
fabbionebut if i test, there is no buildd going on04:55
fabbionehaving one machine only sucks04:56
BenCgot an e300005:07
BenCis there a cd I can download?05:07
fabbioneBenC: nope.. only net install atm05:08
BenCfabbione: do these things affect hoary?05:09
fabbioneBenC: mostlikely yes05:09
fabbioneHerbert will take care of warty05:09
fabbioneif you prefer i can do hoary05:10
fabbione(that's why i was in CC)05:10
fabbioneso you can keep going for breezy05:10
BenCthat'd probably be best, I'm not tooled to test or build hoary05:10
fabbioneok05:11
BenCthese fixes apply for 2.6.12?05:11
fabbioneyup they should05:12
fabbioneyou will need to check that yourself if pitti didn't already05:12
fabbionehmm no pitti didn't this time05:13
BenCI got the info now05:16
fabbioneBenC: ok. some of the stuff is till embargoed05:17
fabbioneso we are not allowed to publish the fixes anywhere05:18
BenCthe second one doesn't even have a fix yet05:18
BenCwell, it says external, so I've no idea where to get it05:18
fabbioneBenC: we will ask pitti.. 05:20
fabbioneBenC: for urgent i mean that we need to get it done asap, but not that we need to die for it ;)05:21
zulhmmm..05:21
fabbionesorry if i gave you the wrong idea about it05:22
fabbionehey zul05:22
zulhey fabbione05:27
=== JaneW [n=JaneW@wbs-146-144-158.telkomadsl.co.za] has joined #ubuntu-kernel
=== jorgp3 [n=jorgp@bnet-dial-229.bartnet.net] has joined #ubuntu-kernel
=== fred__ [n=fred@82-32-125-115.cable.ubr04.hawk.blueyonder.co.uk] has joined #ubuntu-kernel
sedakhello07:10
sedakis there anyone here ?07:11
=== Diziet [n=ian@xenophobe.extern.relativity.greenend.org.uk] has joined #ubuntu-kernel
sedaki made a package for a new driver07:11
desrtsedak; hm?07:12
sedaki'd like to know if someone can have a look at it to see if everything is allright07:12
sedaka package for module-assistant07:12
desrtsedak; you probably want to start out with #ubuntu-motu07:12
sedakfor rtl818007:12
sedakyes, i know07:12
sedakbut they don't do many package driver07:12
desrtah07:12
sedakso they can't really review it07:13
sedakdesrt, can you have a look at it ?07:13
desrtsedak; i'd know less than the universe folks :)07:13
desrtsedak; i mostly just hang around here to hassle fabbio and ben :)07:13
sedakthat a good thing to do :-)07:14
sedakwell, i need to see fabbione is around 07:14
sedakfabbione ?07:14
=== Diziet emails kernel-team@lists re the dpatch header for external-drivers-net_ppp-mppe-compression.dpatch.
fabbionesedak: ?07:19
sedakhello07:19
fabbioneDiziet: it's easier here on irc.. 07:20
fabbioneDiziet: what's wrong with that patch?07:20
sedakdo you have the time to see the package i made for the rtl8180 driver ?07:20
fabbionesedak: not now.. probably tomorrow. i am just coming back from a bad flu07:20
sedakok07:21
sedakhope you'll be fine again tomorrow :-)07:21
DizietThere wasn't any info in it about where it came from.07:22
DizietAlso, I only just found out that this channel existed ...07:22
fabbioneDiziet: see debian/external-modules07:22
fabbionethere is all you need there07:22
DizietJoy.  I just spent half the afternoon looking for that and it turns out to be right there, and you turn out to be right here to tell me.07:23
DizietIt would have been nice if someone on #ubuntu-devel could have maybe pointed me in the right direction.07:24
fabbioneDiziet: what's the reason of such search?07:24
Dizietbugzilla 1424407:24
DizietWhich I think turns out to be included in our 2.6.12 anyway.07:24
fabbioneNOTABUG07:25
fabbioneit is already inclueded07:25
DizietI had the feeling it was a wild goose chase.07:25
fabbioneif it doesn't work patches are welcome07:26
DizietI have no evidence that it doesn't work.07:26
lamont__Diziet: for cooked-and-on-the-table values of wild goose07:26
fabbioneeverything that starts as external_* is basically half supported..07:26
lamont__fabbione: we support the half that works???07:26
lamont__:-)07:26
DizietWell, it's cooked and on the table _now_ but I was chasing it across half the web.07:26
fabbionelamont__: clearly :)07:26
DizietI got badly distracted by the pptpclient sourceforge project who have some found-under-a-rock mppe patches.07:27
fabbioneunfortunatly we didn't manage to get linux-non-supported-modules in place07:27
fabbioneotherwise all that crap would land in that package but in main07:27
fabbioneto *cough*help*cough* our wonderful users07:27
=== fabbione goes to cook dinner...
zulfabbione: its in the debian/rules ;)07:28
fabbionezul: yeah.. in your branch that we never got to merge before FeatureFreeze07:32
fabbioneanyway. -> off07:32
zulbah07:34
DizietWell, thank you for your time.07:34
=== chmj [n=chmj@wbs-146-147-243.telkomadsl.co.za] has joined #ubuntu-kernel
jbaileyHey, have we killed devfs yet?08:01
jbaileyOr are we waiting until initramfs-tools supports evms?08:01
lamont__jbailey: I'd just like it to support my kernel.. :-)08:03
jbaileylamont__: Is initramfs still sucking for you?08:04
jbaileyLast I played in my parisc chroot, lvm seemed to be a larger problem. =)08:05
=== Diziet [n=ian@xenophobe.extern.relativity.greenend.org.uk] has left #ubuntu-kernel []
lamont__jbailey: haven't played with it recently on ia64, but it's still claiming that it's not gzipped data after it fetches the gzipped data from the disk...08:05
jbaileyThat totally doesn't make any sense to me. =(08:09
jbaileyIf the magic numbers match, it ought to have returned valid data.08:09
=== jbailey needs UML/Xen for ia64.
zulgar...stressed08:19
=== Keybuk [n=scott@syndicate.netsplit.com] has left #ubuntu-kernel ["Leaving"]
=== spayne [n=spayne@i-195-137-120-148.freedom2surf.net] has joined #ubuntu-kernel
=== ohphracku [n=zombie@cpe-66-87-4-181.ut.sprintbbd.net] has joined #ubuntu-kernel
=== jorgp2 [n=jorgp@bnet-dial-214.bartnet.net] has joined #ubuntu-kernel
=== ohphracku [n=zombie@cpe-66-87-4-181.ut.sprintbbd.net] has left #ubuntu-kernel ["Leaving"]
=== pornstar [i=void@82.53.127.210] has joined #ubuntu-kernel

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!