/srv/irclogs.ubuntu.com/2005/04/16/#ubuntu-kernel.txt

jbaileyAm I reading the encrypted filesystems wiki page right that losetup will go away in favour of device-mapper?01:58
=== zul [~chuck@CPE0006258ec6c2-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel
zulevening02:02
dilingerjbailey: i'm told dmcrypt is quite easy to set up04:16
dilingeri should really set that up w/in the next week or so04:17
lamontjbailey: no one has told the maintainer of losetup about it going away... :-)04:25
lamontbut devicemapper is better04:25
dilingerjbailey: you still around?04:27
jbaileydilinger: I'm back.04:57
jbaileydilinger: The dmcrypt stuff does look easy enough to setup.04:57
jbaileyThe initrd-tools bits make some assumptions around devince-mapper bits that appear to be wrong.  It's mostly new stuff to me.04:58
jbaileylamont: Cool, thanks/.04:59
dilingerjbailey: yea, it gets triggered in cases when root is not dmcrypt, apparently05:10
dilingeralso, it has libdevmapper1.00 hardcoded in mkinitrd, which is just plain wrong05:11
dilingerfor breezy and sid, anyways (possibly sarge too, not sure).  warty still has libdevmapper1.00; other stuff has libdevmapper1.0105:11
jbaileyAh lovely.  That'll be useful knowledge for troubleshooting.05:16
jbaileyIt's one of the dusty corners of initrd that I haven't explored yet.05:16
=== dilinger nods
jbaileyWoohoo, just called the cops to file a noise complaint against the construction site across the street.05:21
jbailey23:21 is awful late for running a jack hammer.05:21
dilingerstick it to 'em05:24
fabbionemorning05:44
dilingerhiya05:45
lamontfabbione: you got the thing you were pinging me for?06:08
=== lamont begins an interesting practical excercise in the kevin bacon game
fabbionelamont: you mean the dir.old.gz?06:09
lamontyeah06:09
lamont!= bacon game06:09
fabbioneyeah i am checking it06:09
=== lamont lost contact info for someone who has gone to pain to become unreachable.
fabbionethe point is that for me is libmpfr-dev_2.1.0-2ubuntu1_sparc.deb06:10
fabbionei wonder why it did show up again.06:10
fabbionethis package has been built the 12th of Feb06:11
fabbioneand i did other gcc-4.0 built since than06:11
fabbionewith no problems06:11
fabbioneso i guess it was not a Build-Dep before?06:12
lamontdunno06:16
fabbioneneither do i06:17
fabbionei can't keep old logs for too long06:17
=== Lathiat [~lathiat@202.92.213.227] has joined #ubuntu-kernel
=== Mithrand1r [~tfheen@vawad.err.no] has joined #ubuntu-kernel
=== Mithrandir [~tfheen@vawad.err.no] has joined #ubuntu-kernel
=== zul [~chuck@198.62.158.205] has joined #ubuntu-kernel
zulmorning03:09
=== Byakhee [~Byakhee@84.119.73.118] has joined #ubuntu-kernel
zulhah...defeated palm yet again04:25
zulfabbione: you should take 2.6.10 patch it up to 2.6.11 and then all the release candidates04:59
lamontzul: evil05:40
lamonttake 2.6.12rc2 and bring forward the not-from-cvs patches05:41
lamonts/from-cvs/in-bk-head/05:41
zulit works05:41
zulbbl...lunch..05:41
lamontand is much easier05:41
fabbionere05:42
fabbionezul: no no ... clean orig and portforward the patches05:42
fabbioneplus we might have to check all of the non-stolen-from-head05:43
lamontand brings hppa's diff down to (according to that channel topic): 2.6.12-rc2-pa0 is <200k: 55k ad1889, 47k input, 26k compat, 50k split for Linus, 22k remaining05:43
fabbionesome of them have been merged upstream across the time05:43
fabbioneupdate all the drivers05:43
lamontfabbione: right - many of the non-"stolen-from-head" patches have since made it into head.05:43
fabbioneand so on...05:43
fabbionei was thinking to enfore a policy name05:43
fabbionestolen-from-head is there and ok05:43
fabbioneedriver-05:44
lamontah, right05:44
fabbionefor external drivers05:44
lamontoh, yes.  we have the patch-name policy documented on the wiki soon, yes?05:44
fabbionenot that i remember....05:44
=== lamont delegates the task to fabbione
fabbionebut also.. i want to get rid of dpatch05:44
fabbionethat's a must05:45
fabbionelamont: yeah that's fine for me...05:45
lamontreplace it with what?05:45
fabbionelamont: i was considering either dbs-ng 05:45
fabbioneor patch the code directly. BUT keeping track of all the applied patches in debian/patches05:46
lamontcdbs, maybe05:46
fabbionei don't think cdbs can help us at all05:46
lamontdbs gives me the 'run screaming from doogie code' stuff05:47
fabbionewe have a too customized build system05:47
lamontright05:47
fabbionelamont: no no.. not dbs05:47
fabbionedbs-ng...05:47
jbaileyfabbione: dilinger seemed to think that cdbs could do it.05:47
lamontcalling it dbs-ng was a really stupid thing to do then05:47
fabbionejbailey: i really don't think it can, but if you want to look at our rules file and get an idea05:48
jbaileyfabbione: 'kay.  Mind if I do that this weekend?05:48
fabbionelamont: i think xen 2.0.5 in experimental is packaged with dbs-ng05:48
jbaileyOr are you guys looking at it sooner?05:48
=== lamont is kinda busy with other stuff this week. :-)
fabbionejbailey: well.. asap is the better05:49
lamontfabbione: and he said that this weekend was asap for him... :-)05:49
fabbionelamont: yeah.. we will be busy releasing anyway05:49
zulfor the bk cutting edge tree my suggestion would be weekly snapshot06:16
zuland turn all debugging features on06:16
zulit could be too much uploading for one person if we did a daily snapshot06:16
fabbionenah.. we were talking about weekly06:17
fabbioneor every 2 weeks06:17
zulweekly would be better06:17
T-Bonelamont: we've gone over 90% up-to-date! :)06:56
zulholy crap gcc4 sucks06:56
fabbionezul: why?06:57
zulall of these warnings like every second line about pointer targets06:57
fabbioneT-Bone: but i can't see a single hppa.deb in the archive.. am i looking at the wrong place??? oh probably they are only on the .fr mirror :P06:57
T-Bonefabbione: we're not in the archive06:57
fabbioneT-Bone: i know..06:58
T-Bonefabbione: if you look at http://archive.slashdirt.org/ you'll find them06:58
zulheh...im french why do you think i have this outrageous access you silly king06:58
T-Bonefabbione: your sarcasm doesn't reach me :)06:58
=== fabbione sends a real bottle of wine to T-Bone
=== caldwell [~caldwell@cerberus.dtn.radian.biz] has joined #ubuntu-kernel
=== T-Bone teaches fabbione real rugby ;)
zulwell at least i fixed where acpi bombs using gcc406:59
fabbioneoh god.. today archive is really slow07:00
fabbionei only need to mirror and upgrade from warty to hoary07:00
fabbioneit's taking AGES07:00
T-Bonefabbione: tried the fr mirror?07:01
T-Bonefabbione: felt pretty fast last time i tested it07:01
fabbioneno.. it's lagged07:01
zulgah...07:01
fabbioneeverything != archive is lagged07:01
zuldrivers/acpi/processor_idle.c:57: error: static declaration of 'pm_idle_save' follows non-static declaration07:01
zulinclude/acpi/processor.h:181: error: previous declaration of 'pm_idle_save' was here07:01
zulmake[2] : *** [drivers/acpi/processor_idle.o]  Error 107:01
zulmake[1] : *** [drivers/acpi]  Error 207:01
zulmake: *** [drivers]  Error 207:01
fabbioneand for a buildd and my tests i need to be in sync as much as possible07:01
lamontzul: is error07:02
fabbionezul: what are you trying to build?07:02
T-Bonefabbione: i use archive.u.c for the buildd source, works just fine07:02
lamontmirror-missing| grep -v _ia64 | wc07:02
lamont     59      61    424907:02
lamontand then there's xorg and l-s-2.6.10 :-(07:02
T-Bonefabbione: i get around 20Mb/s out of it at min07:02
lamontthat's from the run I started yesterday07:02
fabbioneT-Bone: not from here07:03
T-Bonefabbione: get a real ISP :)07:03
fabbioneT-Bone: you get 20Mb from home?07:03
T-Bonelamont: i had to give-back l-s, segfaulted during the build of ppp modules07:03
T-Bonefabbione: from school07:03
fabbionetsk07:03
fabbionei am home with my standard 2Mb line07:03
fabbioneit still slow07:03
T-Bonefrom home i get 7-8Mb/s07:03
T-Bonefabbione: get a real ISP :)07:04
fabbioneT-Bone: i don't have fibers here yet...07:04
fabbionetoo far away from decent centrals07:04
fabbionebut we are voting next week to get them07:04
T-Bonefabbione: i have adsl at home, not fibers. EC-45 is good for school :)07:04
fabbioneinside the block/area counsil07:04
fabbioneEC-45 is peanuts....07:04
fabbioneT-Bone: www.seabone.net <- that's where i was working in italy07:05
T-Bonemy adsl is to be upgraded to 15Mb/s soon though07:05
T-Bonehuh, I-Tim07:05
fabbioneadsl can't go over 8Mb. it has to be some kind of other technology07:05
T-Boneadslv207:05
T-Boneyou're out of date dude07:05
fabbione"<T -Bone> my adsl"07:06
T-Boneyou see, we may not have the best wine, but we have the best ADSL provider07:06
fabbioneyou are imprecise07:06
T-Bonewww.free.fr07:06
T-Boneif you can read french07:06
T-Boneand i'm not imprecise07:06
T-Boneit is "ADSL v2", 20Mbps ATM throughput, 15Mbps IP07:06
fabbioneyeah right.. adsl != adslv207:06
T-Boneit can use longer range from the DSLAM as well, which is cool for remote people07:07
fabbioneif we will get the fiber.. sorry.. no EC45 or adslv2...07:07
fabbione1xFE07:07
T-Bonei think we are somewhat world leaders on the ADSL market. At least european leaders :)07:08
T-Bonefabbione: how much do you pay for that? :)07:08
fabbioneT-Bone: if we get fibers (again it depends from the vote on thursday) it will cost probably less than my actual adsl07:08
fabbioneISP here are very cheap07:08
T-Bonedoh07:08
T-Bonehow much do you pay for ADSL?07:09
fabbionehmmmm07:09
fabbionelet me check.. i can't remember :)07:09
T-Bonelol07:09
fabbionearound 78 euro/month07:09
T-BoneUGH!07:09
T-Bonehell that's expensive07:09
fabbionenot here07:09
T-Bonethat's about 3x what I pay07:09
fabbionelife in dk is more expensive that in 90% of EU countries07:10
T-Bonei pay 29 euro/month07:10
T-Bonefor 8Mbps07:10
=== lamont would pay 78 euro/month
fabbioneT-Bone: yes but you probably get 1/4 of the wage i get on average and so on07:10
T-Boneand will pay 29e/m for 15Mbps when upgraded07:10
fabbioneT-Bone: it is really impossible to compare prices07:10
fabbionetrust me07:10
fabbionethere are too many factors that influence the price of an ISP between 2 different countries07:11
lamontfabbione: that's because the french government owns everything in the country07:11
T-Bonefabbione: so what's the point of living in .dk if everything is 4x more expensive? Might as well live in London... ;)07:11
fabbioneT-Bone: she was blonde :)07:11
lamontgood reason07:11
fabbionelamont: eheh07:11
T-Bonelamont: actually the state provider sucks hell compared to the provider called "Free", which is world leader (.eu at least) on ADSL technology07:11
T-Bonefabbione: ROTFL07:11
lamontT-Bone: he's dead serious07:12
T-Bonelamont: I got that, yet i'm ROTFLMAO ;] 07:12
=== T-Bone fortunes btw
fabbionei don't joke when it goes to blondes 07:12
fabbionehahaha07:13
T-Bonec'mon ;] 07:13
T-Boneeverybody jokes about blondes. Even my blonde female friends do ;)07:13
zulfabbione: 2.6.10 with gcc407:13
fabbionei know.. that was the real joke T-bone07:14
fabbionezul: it's not worth spending time on it.. really07:14
zulheh...i was bored :)07:14
T-Bonefabbione: i was playing Dumber ;}07:14
zulT-Bone: that cant be too hard for you :)07:14
fabbioneT-Bone: ah sorry.. i couldn't make the difference from the other days07:14
=== fabbione runs away
T-Bonefabbione: that was easy heh? :}07:14
T-Bonefabbione: i really offered you that one =)07:15
fabbioneT-Bone: you just slamed it in front of me :)07:15
T-Bonefabbione: cause I knew you'd jump on it, damned Italian big mouth 8-))07:15
fabbioneT-Bone: ain't my fault if you want to play dumb :)07:16
fabbioneoh god guys.. i really love this channel07:16
T-Bonefabbione: you're such a good audience ;] 07:16
T-Bonefabbione: I think we all do. zul might tell, seems that with lamont you and I on the same chan, one must have fun ;)07:17
fabbioneehehhe07:17
zuli never have fun07:18
fabbionezul: that's because you are french :P07:18
T-Bonezul: poor little thing. Feel left alone? :}07:18
zulfabbione: shut your mouth!07:19
=== T-Bone ducks!
fabbioneahahahha07:19
=== T-Bone wonders what makes zul french :)
zulT-Bone: yeah your gentoo haters..07:19
T-Bonelol07:19
zulthere is nothing lower than being french...well maybe italians07:19
T-BoneROTFL07:19
T-Bonei guess that one is well deserved ;)07:20
=== T-Bone ^5s zul
zulheh...burn..:)07:20
fabbioneahahha07:20
=== T-Bone notices that all that is logged and will soon be parsed by Googlebots
fabbionelet them parse....07:21
fabbionewho cares..07:21
=== zul does his italian imitation...
T-Bonenow if someone has a little script that converts irc nicks to realname, and makes that available to said googlebots, well... ;}07:21
fabbionei am still in time to erase the logs07:21
zulthat is one spicy meatball07:21
fabbioneand let them hit a bunch of 404 :)07:22
T-Bonelol07:22
fabbionenow.. let me think..07:22
T-Bonefabbione: don't do that07:22
fabbioneshould i start adding random entries like:07:22
T-Bonefabbione: it's said to damage italian brains ;)07:22
fabbione<T-Bone> I SUCK..I REALLY DO :)07:22
T-BoneLOL07:22
zulhehe07:22
fabbionethat would kinda own you :P07:23
T-Bonefabbione: you don't need to add these, I usually add them myself ;)07:23
zulits called T-Bone...sublimial...sucks...messages07:23
fabbioneahahha07:24
T-Bonezul: you're not funny07:24
fabbionemy wife is asking me wth i am laughing so much...07:24
zulim all for equal opprtunity :)07:24
T-Bone8)07:24
T-Bonefabbione: better not tell :)07:24
fabbionei am not...07:24
T-Bonefabbione: tell her you're working hard! :)07:25
fabbionei am :)07:25
T-Bone8)07:25
fabbionelooking at the ubuntu mirror that is rsync07:25
fabbionethat's a hard job07:25
fabbionespecially when you need to wait for it to do the last test before closing the day07:26
fabbioneat least she is cooking dinner :)07:26
T-Bonelol07:26
T-Bonei hate you :)07:26
fabbionetime for food.. the mirror hasn't finished yet07:29
zulso umm...07:43
zullike e_whatevercrappydriver.dpatch?07:44
fabbioneedriver-crappydevice07:54
zulk07:56
zulare you going to do the baz dance for breezy...07:56
fabbioneyeah after hoary is released07:57
zulok..07:57
zuli could wait until then...maybe..07:57
fabbionebut we need to decide first how we want to proceed07:57
fabbionelamont: is it confirmed that breezy will be gcc-4.0?07:57
zulyeppers07:58
fabbionethan i would say that first we need to get an orig.tar.gz :)07:59
fabbionefrom 2.6.11pre2.6.12rc2 or whatever lamont suggested07:59
fabbioneand to test if everything is ok07:59
zul2.6.12-rc i can scrunge one up tonight07:59
zulketchup is good ;)07:59
fabbionezul: easy :)07:59
fabbionemy suggestion is:08:00
zuli know...i was volunteering :)08:00
fabbionea) unpack linux-source-2.6.10.orig.tar.gz08:00
fabbioneb) rename the dir to be 2.6.1208:00
fabbionec) repackit08:00
fabbioned) i will let you figure this one out :)08:01
fabbionee) mv our debian/ to the 2.6.1208:01
fabbionef) edit control.stub and changelog08:01
fabbioneand build it08:01
fabbionesee if all the packages get the right versions and so on08:01
fabbionefor one simple reason08:01
fabbionemake-kpkg doesn't like fancy names in the versions08:01
fabbioneand even if it builds, it will fail later to create the debs08:02
fabbioneand the udebs08:02
fabbioneonce you have this undercontrol08:02
fabbioneyou can use the .11 from vanilla08:02
zulthe patch file?08:02
fabbioneand add the patches to debian/patches to go up to 2.6.12rc2 via stolen-from-head that you really want as the first patch08:03
fabbionezul: yeah well go up to 2.6.12 rc in someways08:03
fabbioneonce you are there...08:03
fabbioneyou can have fun...08:03
zulor an ulcer08:03
fabbionebut just create the orig.tar.gz08:03
zulokie dokie08:04
fabbionewe will work in rediffing the patches in baz08:04
fabbionewe need to keep track of what we do and why08:04
zularchive is slooooow08:07
lamontfabbione: I'd be inclined to go with linux-source-2.6.12_2.6.11.90.orig.tar.gz08:19
fabbionelamont: that works for me08:19
lamontand use 2.6.11.90-1 as the head of changelog08:19
lamontleast amount of pain.08:19
fabbionelamont: it is ok, but we need to see how and if make-kpkg breaks08:19
lamontand if it _won't_ eat that for some reason, then well.... linux-source-2.6.12_2.6.11.orig.tar.gz has some flavor... :)08:20
fabbionei remember some horror from Mataro' when i did try to create packages for bk08:20
fabbioneehehhe08:20
zuluh that sucks...i seem to have used up 13 GB on my /home partition08:20
lamont /dev/hda8             96124904  83814404   7427548  92% /home08:20
lamontuh, yeah.  13GB.08:20
zulppphpth08:21
fabbioneAlloc PE / Size       137223 / 536.03 GB08:21
fabbioneFree  PE / Size       20144 / 78.69 GB08:21
fabbionehmm08:21
fabbionei am almost running out of space08:21
=== fabbione ducks
lamontfabbione: it doesn't matter how much space you start with, it fills up almost immediately.08:22
lamontand then we start pruning08:22
fabbionelamont: that's why i got 600 DVD for my weeding08:22
fabbionewedding even08:22
fabbioneto backup that mess :)08:22
fabbioneand reinstall08:22
lamontheh08:22
fabbionei have in mind more or less what i want08:23
T-Bonelamont: that only happens when you don't know how to sort data. My desk and main machine has 15GB total diskspace08:23
T-Boneand it's not full :)08:23
lamontT-Bone: no, that only happens when you add more disk everytime you run out of space.. :-)08:24
fabbioneehhehe08:24
T-Bonelamont: fair enough. Though i have about 780GB on my secondary machine, of which a good 40% is free :)08:25
T-Bonefabbione: you see, mine is bigger ;] 08:25
lamonthrm.. who was the openfirmware guru08:25
lamont?08:25
T-Bonelamont: svenl08:25
T-Bonelamont: i can help a little maybe08:25
fabbioneT-Bone: is that a raid5 ?08:25
T-Bonefabbione: no08:25
fabbioneor just plain JOD?08:25
lamontT-Bone: I just want to have it spew out the paths on the machine08:26
fabbioneeh mine is raid508:26
T-BoneJBOD08:26
fabbioneyeah JBOD08:26
T-Boneand even more evil08:26
fabbioneT-Bone: i am over that if i unroll the raids08:26
T-BoneIDE JBOD ;)08:26
fabbioneyeah mine is IDE too08:26
T-Bonelamont: depends on the of version08:26
T-Bonelamont: usually you'd try 'ls'08:26
lamont"no active package"08:26
T-Bonefabbione: lol, what's the point of RAID over IDE? :)08:26
T-Bonelamont: macintosh or pegasos?08:27
lamontT-Bone: OF 3.0.f2 built  4/23/9908:27
lamontMac G3408:27
lamontG308:27
T-Bonehold on08:27
T-Boneerg08:29
=== T-Bone is doing nasty things with OF
=== T-Bone curses yaboot stage 1 that interferes with his attempt to get OF to talk nicely to him :P
fabbionenight guys08:31
fabbionecya tomorrow morning08:31
fabbioneoh i forgot the topic08:31
=== ..[topic/#ubuntu-kernel:fabbione] : Ubuntu kernel development discussion | http://www.ubuntulinux.org/wiki/KernelTeam | http://people.u.c/~lamont/Archives/kernel-team@ubuntu.com--2005/ stable: kernel-debian--pre35--2.6.10 playground: kernel-debian--experimental--2.6.10 | There are no kernel bugs.. only broken hardware
fabbionethere08:32
zulc ya08:32
T-Bonelamont: hmm i'm confused, pegasos and macs OF are quite different. You want to ask svenl, definitely08:34
lamontok08:34
T-Bonelamont: fyi, -34 currently building, as well as xorg -1008:41
T-Boneand we're 91% up-to-date :)08:42
zulwtf?08:49
zuldh_clean: cannot read debian/control: No such file or directory08:50
zuland then my debian directory gets blown away...frigging 09:00
=== zul smacks T-Bone
T-Bonehey!09:00
T-Bonei did nothing :P09:00
zuli know but today you are my bitch :)09:00
T-Bonei'm a computer whore ;}09:12
T-Gonedinner time09:12
zulye...yes you are09:12
dilingerjbailey: cdbs2, maybe.  cdbs1 would be a mess11:35
dilingerjbailey: i use it for kernel modules, which is fairly simple, and it gets ugly11:36
jbaileydilinger: 'kay.  I've promised seb a prototype for udu.11:36
jbaileyLike self hosting and able to do nano.11:36
dilingerhttp://svn.debian.org/wsvn/kernel/trunk/kernel/source/kernel-source-nonfree-2.6.11-2.6.11/debian/rules?op=file&rev=0&sc=011:37
dilingerjbailey: neat11:37
jbaileydilinger: This isn't too bad.  A few of these things could get safely tossed into a kernel.mk file.11:37
dilingerjbailey: picture that multipled by 511:38
dilingeror something like that11:38
dilingeras it needs to build multiple images11:38
jbaileydilinger: Right, but is the result still better than it would be with other build helpers?11:38
jbaileyThis is only a page and easy to read.  With a helper and some clear commenting, it growing wouldn't necessarily be that bad.11:38
dilingerjbailey: well, what really sucks is that it uses kernel-package11:39
dilingerwhich doesn't really go well w/ cdbs11:39
dilingeryou end up having to do the debhelper stuff through hooks11:39
jbaileyOh ugh.11:39
dilingerso, debhelper.mk is next to useles11:39
dilingernot a cdbs problem, but a kernel-package.. um.. misfeature, i guess.11:40
jbailey'kay then.  I had been thinking in terms of the rules file actually doing the configure/make dep/ make/ make install passes itself.11:40
jbaileyIn which case a helper could do that, and then you're just missing a basic multipass.11:41
jbaileydilinger: In other news, I've discovered that cat'ing multiple .gz'ed initramfs' works, so now all we need is better bootloaders, and it's trivially to have the bootloaded assmebler the modules you want for loading. ;)11:42
jbaileyIn the meantime, just have to do that step by hand from the running system.11:43
jbaileyBut when/if bootloaders grow that ability it would be nice to just have those as nice add-ons and just need to generate a config CPIO file that gets added on to drop the configurations onto the initramfs.11:43
dilingernice11:44
dilingerjbailey: actually, it would be nice to replace kernel-package w/ a cdbs build11:44
jbaileydilinger: Does it get more complicated than I described above?  I haven't done a kernel compiler in.. erm.  a long time.11:44
jbaileyLike from scratch.11:45
jbaileyI've mostly just tweaked .config files and built new packages.11:45
dilingerit does, but all the complexity can be popped into headers11:45
dilingerand a lot of the complexity is backwards compat crap11:45
dilingerwhich we don't have to worry about for 2.6+11:45
jbaileyNice, that's handy.11:46
jbaileydilinger: How simple do you think it ought to wind up being in the end?  Some variables to refer to the config passes, and have debhelper pluck out all the bits for the various packages?11:54

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