=== setuid [i=japh@gnu-designs.com] has joined #ubuntu-kernel | ||
setuid | Well that was fun... seems the fglrx module doesn't built at all, and nothing I can find contains it, nor radeon or r300 modules for the kernel, so it all dies after I try to log in. | 12:05 |
---|---|---|
setuid | I get the gdm login/graphics/etc. but after I log in, it just recycles gdm, no X at all after login. | 12:06 |
setuid | Is there some magical mojo to get the fglrx kernel module built for a custom kernel? | 12:06 |
=== fabbione [i=fabbione@gordian.fabbione.net] has joined #ubuntu-kernel | ||
=== maks [n=max@baikonur.stro.at] has joined #ubuntu-kernel | ||
=== JanC [n=janc@lugwv/member/JanC] has joined #ubuntu-kernel | ||
=== mjg59_ [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel | ||
=== _Traxer|off [i=traxer@shell6.powershells.de] has joined #ubuntu-kernel | ||
=== lifeless [n=robertc@dsl-152.1.240.220.rns01-kent-syd.dsl.comindico.com.au] has joined #ubuntu-kernel | ||
=== chninkel [n=yann@alcyone.pleiades.fr.eu.org] has joined #ubuntu-kernel | ||
=== maks [n=max@213.239.196.228] has joined #ubuntu-kernel | ||
=== lamont [n=lamont@mib.fc.hp.com] has joined #ubuntu-kernel | ||
=== chninkel_ [n=yann@alcyone.pleiades.fr.eu.org] has joined #ubuntu-kernel | ||
=== lifeless [n=robertc@dsl-152.1.240.220.rns01-kent-syd.dsl.comindico.com.au] has joined #ubuntu-kernel | ||
=== jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel | ||
jbailey | benc, airlied: I noticed that the kernel from today has more r300 bits in it. | 12:21 |
jbailey | So I'm trying it again. =) | 12:21 |
BenC | jbailey: yeah, I'm hoping that will help you | 12:21 |
setuid | hrm, r300 you say? ;) | 12:34 |
=== ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | Kernl GIT tree info (updated): https://wiki.ubuntu.com/KernelGitGuide | 2.6.15-22.35 uploaded (May your computer boot happily) | ||
=== ..[topic/#ubuntu-kernel:BenC] : Ubuntu kernel development discussion ONLY | Kernel GIT tree info (updated): https://wiki.ubuntu.com/KernelGitGuide | 2.6.15-22.35 uploaded (May your computer boot happily) | ||
BenC | well, he goes a git-pull from latest kernel | 12:38 |
=== lamont [n=lamont@mib.fc.hp.com] has joined #ubuntu-kernel | ||
=== _mike_ [n=mike@220.157.65.127] has joined #ubuntu-kernel | ||
jbailey | BenC: *lol* | 12:43 |
jbailey | I'd say "sheesh, you barely stopped for a smoke", except I'm sure that's not true ;) | 12:43 |
BenC | this merge is getting ugly :) | 12:57 |
jbailey | Do you cherry pick changesets, or create new diffs when you do backports? | 01:01 |
=== fabbione [i=fabbione@195.22.207.162] has joined #ubuntu-kernel | ||
BenC | jbailey: a little of both | 01:09 |
jbailey | I'm really not looking forward to merging glibc with Debian again. | 01:10 |
BenC | some of the stuff can't be cherry picked, especially things that made the semphore -> mutex switch | 01:10 |
jbailey | Right. | 01:10 |
jbailey | Assuming you're targetting .17 for edgy? | 01:10 |
jbailey | I'm wondering if glibc should have support for the new syscalls. | 01:10 |
BenC | man, I know I've seen atleast 50 conflicts pass by so far | 01:10 |
=== maks [n=max@213.239.196.228] has joined #ubuntu-kernel | ||
=== fabbione [i=fabbione@gordian.fabbione.net] has joined #ubuntu-kernel | ||
=== JanC [n=janc@lugwv/member/JanC] has joined #ubuntu-kernel | ||
=== mjg59_ [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel | ||
BenC | yeah, maybe even .18 | 01:10 |
=== maks [n=max@baikonur.stro.at] has joined #ubuntu-kernel | ||
jbailey | 'k | 01:10 |
=== fabbione_ [i=fabbione@gordian.fabbione.net] has joined #ubuntu-kernel | ||
BenC | if .17 released within 4 weeks of edgy opening, then we'll target .18 | 01:11 |
jbailey | You're brave. | 01:11 |
jbailey | Aren't we theoretically still aiming for an October edgy release? | 01:11 |
BenC | Linus has been sticking with his 3 month release cycles pretty closely | 01:11 |
BenC | are we??! | 01:11 |
BenC | wow, that sounds like it will be a little insane | 01:12 |
jbailey | I haven't seen anything that says we're not. | 01:12 |
crimsun | yeah, we're shooting for a shortened one to get back on track | 01:12 |
crimsun | 6.06, 6.10, 7.04, 7.10, ... | 01:13 |
jbailey | Anyhow, off to the gym. | 01:14 |
BenC | shit, this merge is going to take a few days | 01:52 |
desrt | oh man | 01:55 |
desrt | edgy kernel already? | 01:55 |
=== mxpxpod_ [n=BryanFor@wuw-ojr3gmca.dybb.com] has joined #ubuntu-kernel | ||
BenC | 360 files conflicting | 01:57 |
=== mxpxpod__ [n=BryanFor@wuw-ojr3gmca.dybb.com] has joined #ubuntu-kernel | ||
cjb | This is for edgy? I think Fedora does things pretty well by rebasing to Linus every day. | 02:16 |
cjb | But I guess they stay closer to upstream than you guys. | 02:16 |
desrt | it takes a lot of guts to be as cool as benc :) | 02:17 |
cjb | he doesn't feel the peer pressure, huh :) | 02:17 |
cjb | Of course, Fedora also introduces weird bugs that break things that they then ignore the bug report on for a month, and then continue to ignore it after I find out what they've done wrong and give them a patch to fix it. *sigh* | 02:22 |
desrt | BenC; got a sec to explain something to me? | 02:23 |
airlied | cjb: kernel bugs? or other things? | 02:23 |
desrt | (or anyone who knows, really) | 02:24 |
cjb | airlied: Yeah, kernel, I was thinking of https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=190128 . | 02:24 |
desrt | when building a new kernel config, how do i decide if it needs a new ABI number and how do i affect the change? is this somehow automated? | 02:24 |
cjb | (Which is still broken in Rawhide.) | 02:24 |
airlied | I haven't seen davej online lately, he might be away.. :-) | 02:25 |
cjb | He's posting to lkml and pushing new kernels, so I think he is around, but I've heard that he's pretty swamped. | 02:25 |
airlied | they should have a hey I fixed this flag in bugzilla.. | 02:27 |
crimsun | desrt: yes, it's in debian/rules. | 02:27 |
desrt | crimsun; does kpkg automatically deal with it? | 02:28 |
crimsun | desrt: it's sed, grep, wc, and shell magic. | 02:28 |
cjb | airlied: That's a really good idea. I wouldn't have privileges to change state to it, though, unless they make it so that anyone can. Which would also be a good idea. | 02:30 |
=== chuck_ [n=chuck@CPE0006258ec6c1-CM000a73655d0e.cpe.net.cable.rogers.com] has joined #ubuntu-kernel | ||
jbailey | cjb: Every distro has their own approaches to kernel, toolchain, etc. | 02:36 |
jbailey | It's what makes us all interesting ;) | 02:36 |
desrt | jbailey; is there more of a formal office in montreal these days? | 02:37 |
=== mjg59 [n=mjg59@cavan.codon.org.uk] has joined #ubuntu-kernel | ||
cjb | Yeah. And then there's Xen, which rebases against -stable guaranteeing that *no* distro will be in a position to merge it, no matter how often they rebase to Linus. :) | 02:37 |
desrt | crimsun; i guess the thing that confuses me -- you have to know the kernel version number before you start building it | 02:39 |
desrt | crimsun; but it doesn't seem like you'd know the ABI version number until after building | 02:39 |
BenC | desrt: If you do the debian/rules build, it will tell you the ABI changed | 02:40 |
BenC | and if it does do: debian/rules bumpabi | 02:40 |
BenC | and be sure to edit the version in debian/changelog | 02:40 |
desrt | then i have to rebuild? | 02:40 |
BenC | debian/rules binary-debs | 02:40 |
BenC | is the target I use | 02:40 |
BenC | desrt: yeah | 02:40 |
desrt | crikey | 02:40 |
BenC | desrt: it will only build the first target though | 02:40 |
BenC | not the whole thing | 02:41 |
desrt | ok. if i'm fairly sure the ABI is gonna change then i can just bumpabi first? | 02:41 |
BenC | desrt: apt-get install ccache | 02:41 |
desrt | good call. | 02:41 |
BenC | yeah | 02:41 |
desrt | does that setup automagically? | 02:41 |
BenC | you have to add /usr/lib/ccache to the front of your PATH | 02:41 |
desrt | done & done | 02:41 |
desrt | so first, build | 02:42 |
BenC | maybe even set CCACHE_DIR to some place you want | 02:42 |
BenC | debian/rules binary-debs | 02:42 |
=== cjb symlinks gcc to 'ccacne gcc'. | ||
BenC | if the ABI changes, run debian/rules bumpabi | 02:42 |
desrt | erm.... | 02:42 |
desrt | there is no debian/ directory here? | 02:42 |
BenC | edit debian/changelog to reflect the ABI change | 02:42 |
BenC | then how can you have an existing ABI? :) | 02:42 |
desrt | where do i get debian/ from? | 02:43 |
jbailey | desrt: Yeah, we take possesion of it next week. | 02:43 |
BenC | apt-get source linux-source-2.6.15 | 02:43 |
desrt | oh. | 02:43 |
desrt | i thought install == source for linux-source | 02:43 |
BenC | well, it is, minus the debian directory that we use to build the packages | 02:43 |
desrt | ok. this actually makes quite a bit more sense now :) | 02:43 |
desrt | what of make-kpkg? | 02:44 |
desrt | jbailey; that's pretty sweet. where is it? | 02:44 |
chuck_ | bleah... | 02:44 |
desrt | BenC; so i take it the bump-abi rule just copies the new abi info into the debian/ directory? | 02:46 |
BenC | desrt: it just sed's the files needed to reflect the new version | 02:50 |
desrt | there's a stored copy of the 'old' abi information somewhere in debian/ for purposes of comparison, right? | 02:50 |
desrt | ok... so the -23 source package has a copy of the abi's for -22 but not for -23 | 02:54 |
jbailey | desrt: Near Lucien L'alier. | 03:04 |
jbailey | (mtro) | 03:04 |
=== _human_blip_ [n=mike@220.157.65.127] has joined #ubuntu-kernel | ||
desrt | jbailey; cool. | 04:21 |
=== jbailey [n=jbailey@modemcable139.249-203-24.mc.videotron.ca] has joined #ubuntu-kernel | ||
jbailey | BenC: X spun out and kill -9 didn't solve it. | 04:42 |
jbailey | So still some sort of problem Ah well. post-dapper problem. | 04:42 |
airlied | jbailey: still nothing causes the spin? | 04:44 |
airlied | just general usage? | 04:44 |
BenC | jbailey: that sucks | 04:44 |
jbailey | airlied: In this case it was in a screensaver. | 04:45 |
jbailey | I was afk, so I don't have more details. | 04:45 |
jbailey | In a screensaver, I can't really tell if it's thatit's changing, was about to DPMS, etc. | 04:45 |
airlied | hmm I'll probably have to run more screensavers.. you on 64-bit or SMP? | 04:45 |
jbailey | It's unfortunately a very poor testcase. | 04:45 |
jbailey | both. =) | 04:45 |
airlied | ah lovely... I've always wondered if we might have some SMP issues.. | 04:46 |
jbailey | airlied: If you have some sort of assault test that you want me to run, I can cheerfully do so. | 04:46 |
airlied | my only SMP machine has an X1300 in it, so I can't use it yet.. | 04:46 |
jbailey | Although perhaps not tonight. | 04:46 |
airlied | jbailey: I'd love to have one that's the problem we don't so finding these things is a real pita.. | 04:46 |
airlied | I might put a PCI radeon in my SMP box for a test... | 04:50 |
jbailey | YEah, I'm sure. | 04:53 |
jbailey | I don't know how to start troubleshooting this sort of thing. | 04:53 |
jbailey | If it's at all significant, kill -9 wouldn't take out Xorg | 04:54 |
airlied | yeah nothing will take it out once the GPU crashes.. | 04:54 |
airlied | it's also like playing whack a mole, you smack one bug, another bug the exact same pops up, it could be 5 bugs all different having the same symptoms.. | 04:55 |
jbailey | heh. | 04:58 |
jbailey | Are you part of upstream for this? | 04:58 |
airlied | jbailey: I'm the DRM maintainer for upstream.. | 05:01 |
jbailey | Ah, cool. | 05:02 |
jbailey | You seemed excessivly clueful for a random observer. =) | 05:02 |
airlied | I also try and fix r300 things because I hate ATI. :-P | 05:03 |
fabbione_ | morning guys | 05:08 |
jbailey | airlied: Eh, you express hatred in a way with which I am not familiar ;) | 05:08 |
jbailey | Heya Fabio! | 05:08 |
fabbione | hey Jeff | 05:08 |
=== jbailey goes and grabs some more food. | ||
fabbione | touch: cannot touch `/root/glibc-2.3.6/stamp-dir/check_sparc64': No space left on device | 05:09 |
fabbione | WTF | 05:09 |
fabbione | infinity: i need your binaries.. sort of now | 05:09 |
=== dilinger [n=dilinger@207.210.101.209] has joined #ubuntu-kernel | ||
=== doko [n=doko@201.160.17.131.cableonline.com.mx] has joined #ubuntu-kernel | ||
BenC | airlied: When I sync to 2.6.17-git, I can just revert and merge conflicts in drm to match Linus' tree, right? | 08:07 |
airlied | BenC: yes there should be nothing in your tree that isn't in Linus.. | 08:10 |
=== BenC [n=bcollins@debian/developer/bcollins] has joined #ubuntu-kernel | ||
infinity | mjg59: Too late to do anything about it now, but I think your "switch to 640x480 on some machines" hack could be circumvented with just tweaking the 640x400 timings instead. | 08:54 |
infinity | mjg59: If I'd known you were looking at doing that, I would have worked with you to test some stuff. I didn't have access to any hardware where the current (admittedly not-quite-right) timings didn't work, so I couldn't test better ones. :/ | 08:55 |
=== allee [n=ach@allee.exgal.mpe.mpg.de] has joined #ubuntu-kernel | ||
=== Keybuk [n=scott@quest.netsplit.com] has joined #ubuntu-kernel | ||
=== fabbione [i=fabbione@gordian.fabbione.net] has joined #ubuntu-kernel | ||
bluefoxicy | whew. | 09:31 |
bluefoxicy | that is ugly as shit. | 09:31 |
=== bluefoxicy is attempting to write a patch to control mmap() and stack randomization entropy levels from kernel parameters | ||
bluefoxicy | trivial but 1) i don't understand the code really so I may be off by 1; 2) I shoved two additions, a subtraction, a bit shift, and a modulus in one statement. | 09:33 |
bluefoxicy | at least the code appears robust enough to handle it even if I screw up. | 09:33 |
=== Keybuk [n=scott@quest.netsplit.com] has left #ubuntu-kernel [""] | ||
=== allee [n=ach@allee.exgal.mpe.mpg.de] has joined #ubuntu-kernel | ||
=== allee [n=ach@allee.exgal.mpe.mpg.de] has joined #ubuntu-kernel | ||
=== tuxmaniac [n=aanjhan@60.254.67.17] has joined #ubuntu-kernel | ||
=== Keybuk [n=scott@quest.netsplit.com] has joined #ubuntu-kernel | ||
infinity | BenC: Hrm, the CONFIG_RTC=y thing didn't go in, did it? | 02:38 |
Keybuk | oh, hurrah | 02:42 |
Keybuk | pleeeease put that in | 02:42 |
Keybuk | and GENRTC=y | 02:42 |
Keybuk | let more pointless modules bite the dust <g> | 02:42 |
infinity | It's less about pointless modules and more about SMP machines *requiring* the RTC driver to manipulate the hwclock in any sane fashion. | 02:44 |
BenC | infinity: damnit...that's two config options I forgot to add | 02:45 |
infinity | So hwclock and ntp don't blow up, among others. | 02:45 |
Keybuk | hwclock wouldn't blow up now anyway | 02:45 |
BenC | I forgot HUGEPAGES for PPC64 aswell | 02:45 |
=== Keybuk points at /etc/udev/rules.d/85-hwclock.rules | ||
BenC | I'll sneak it into the first update | 02:45 |
desrt | Keybuk; try to convince BenC to release a -24 for dapper :) | 02:46 |
BenC | no no | 02:46 |
=== desrt j/k | ||
BenC | -23.36 if anything :) | 02:46 |
desrt | fwiw i have a sneaky suspicion that hugepages changes the abi | 02:46 |
Keybuk | desrt: that wouldn't be an ABI change | 02:46 |
desrt | you certainly get new functions | 02:47 |
infinity | BenC: When you say "first update", you mean "the first update before release", not "after", right? :) | 02:47 |
Keybuk | yeah, let's have ANOTHER kernel ABI change before release | 02:47 |
Keybuk | they're SO MUCH FUN | 02:47 |
infinity | BenC: I suspect the RTC change especially won't fly after release. | 02:47 |
BenC | if HUGEPAGES changes the ABI, then powerpc will get an ignore file | 02:47 |
infinity | (Doubly-so, since we seem to build some udebs with rtc drivers in them?... Feh) | 02:47 |
BenC | I don't think it would affect modules anyway | 02:47 |
desrt | ignore file? | 02:48 |
BenC | desrt: echo "Yes" > debian/abi/powerpc.ignore | 02:48 |
BenC | so I don't have to bump the ABI | 02:48 |
desrt | ah. i see. cheater file :) | 02:48 |
infinity | Of course, we put those udebs in universe, so we know d-i isn't using them ANYWAY. | 02:48 |
BenC | if it only adds functions, it would ignore it anyway | 02:48 |
BenC | since new != important | 02:48 |
BenC | wait a second | 02:49 |
desrt | plus... the normal modules get upgraded along with -image and you don't really have a whole lot of linux-restricted-modules kicking around on ppc | 02:49 |
BenC | I'm not working...I don't have money in my budget for extraneous beer-to-super expenses | 02:49 |
Keybuk | BenC: fabbione promised to wipe clean the irc logs :) | 02:49 |
infinity | BenC: I'll cover your drink requirement. I'm sure I already owe several. | 02:49 |
BenC | lol | 02:49 |
infinity | Hell, I actually UPLOADED on my day off. | 02:50 |
infinity | So, I'm a lost cause. | 02:50 |
=== zyga [n=zyga@ubuntu/member/zyga] has joined #ubuntu-kernel | ||
Keybuk | infinity: and it wasn't exactly a subtle package either ;) | 02:50 |
=== desrt has been looking for someone willing to do an upload since last night :p | ||
BenC | I'm at 73 files remaining of 360 for the edgy 2.6.17 git merge | 02:50 |
zyga | hi | 02:50 |
zyga | I've got a kernel panic just a moment ago | 02:51 |
zyga | I've got pictures if anyone is interested | 02:51 |
desrt | BenC; btw.. is it normal that the kernel doesn't build-dep on kernel-wedge or m4? | 02:51 |
BenC | zyga: you're too late...all oopses are not considered features, which enable breaks to prevent hyper-tension and repetitive syndromes | 02:51 |
jbailey | s/not/now/ ? | 02:52 |
zyga | ? :D | 02:52 |
BenC | yes | 02:52 |
zyga | duh.. I need to run now | 02:52 |
zyga | I'll post the pictures and come back in 5 hours | 02:52 |
zyga | bbl | 02:52 |
BenC | desrt: the kernel, or the kernel source? | 02:52 |
desrt | BenC; source | 02:52 |
desrt | oh man i'm a dumbass. i did apt-get build-dep linux-source. disregard. | 02:53 |
BenC | apt-get install linux-kernel-devel | 02:53 |
BenC | that'll help a little too :) | 02:53 |
BenC | http://www.technologyevangelist.com/2006/05/ubuntu_linux_dapper.html | 02:54 |
desrt | so i'm still trying to figure out the abi thing | 02:54 |
=== BenC notes we are better than Vista | ||
BenC | and, our install is easier than MacOSX, which I think says a lot more | 02:54 |
desrt | the -23 source i downloaded contains the -22 abi files.... how does the build check if my build is the same as -23? | 02:54 |
BenC | desrt: it doesn't | 02:54 |
BenC | we have to add the -23 files in afterwards, they aren't important in that source file | 02:55 |
desrt | ah. so if you were building a kernel, you'd have them kicking around in there | 02:55 |
infinity | desrt: Grab the -23 ABI file from /boot/ :) | 02:55 |
desrt | ahah. truth revealed | 02:55 |
desrt | infinity; thx. i've been wondering :) | 02:56 |
BenC | desrt: debian/bin/getabis 2.6.15 23 | 02:56 |
BenC | that will get them _all_ | 02:56 |
infinity | BenC: Actually, that's a fair point. Maybe our shipping kernel source should contain the ABI files, for people trying to mangle derivative kernel builds. | 02:56 |
BenC | in the process, downloading a few hundred megs of .debs :) | 02:56 |
infinity | BenC: Which makes a good excuse for one last upload. :) | 02:56 |
desrt | FAILED | 02:56 |
desrt | :) | 02:56 |
BenC | desrt: debian/bin/getabis 2.6.15 23.35 | 02:57 |
infinity | BenC: The process is confusing enough for people to understand without also wondering where the files are. :) | 02:57 |
desrt | to hell with it | 02:57 |
BenC | infinity: Nah, ppl love a good file hunt :) | 02:57 |
=== desrt copies from /boot :) | ||
jbailey | BenC: I'm vaguely amused at comparisons between Linux distros and Vista. | 02:58 |
BenC | jbailey: I'm ammused that our TCP/IP stack needs optimizing | 02:58 |
jbailey | With a target release of midwinters, the likelyhood of xgl being stable is high. | 02:58 |
jbailey | Did davem just plan a redo of a large part of the IP stack? | 02:59 |
jbailey | I thought I saw something about doing channels and some other optimisation. | 03:00 |
infinity | Our IP stack has needed optimisation for eons, this is no surprise. | 03:00 |
jbailey | But just that all the flashing zing that Vista can pump out will probably be matched in every aspect that an end user will care about. | 03:00 |
infinity | The WinNT IP stack was lifted nearly verbatim from BSD, and has been very performant from day 1. | 03:00 |
infinity | But, yeah, most users don't give a hoot about a slightly better IP stack. | 03:01 |
infinity | And most server scenarios don't push data from RAM fast enough for those benchmarks to mean anything to them. | 03:01 |
infinity | (Some few here and there actually do, but most machines aren't bottlenecked elsewhere) | 03:01 |
infinity | s/aren't/are/ | 03:02 |
desrt | heh | 03:05 |
desrt | this dapper review has an entire paragraph devoted to "we should have shipped networkmanager by default" | 03:06 |
infinity | Yeah, WiFi is either not important at all, or THE MOST IMPORTANT THING EVER. It's a rather polarising thing. | 03:08 |
infinity | I'm still in the "not important at all" camp, despite having a small wireless network at home. | 03:08 |
desrt | it depends on if you have a laptop or not | 03:08 |
infinity | Mostly because when I want to get any real work done, I still switch to copper (and probably always will, since copper is always a few steps ahead) | 03:09 |
desrt | my laptop's inability to get onto my wireless network without me invoking a shellscript is sort of a sore spot | 03:11 |
desrt | it doesn't hurt me too much.... | 03:12 |
desrt | but it hurts my "look how awesome ubuntu is!" credibility :p | 03:12 |
infinity | Mine seems to do well anywhere with NM now. | 03:12 |
infinity | WEP, WPA, etc, etc. | 03:12 |
desrt | nm is terribly buggy on bigendian | 03:12 |
=== mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel | ||
infinity | So are many wireless drivers. :/ | 03:13 |
desrt | the opensource broadcom ones seem to work fairly well | 03:15 |
desrt | and the latest upgrade massively improved the range i get | 03:15 |
desrt | i used to barely be able to connect from the far corner of my office... now i get flawless signal from across the hall | 03:16 |
BenC | linux-source-2.6.15 (2.6.15-23.36) UNRELEASED; urgency=low | 03:22 |
BenC | Changes by Ben Collins | 03:22 |
BenC | * powerpc64: Enable HUGETLB | 03:22 |
BenC | * x86: Build-in rtc and genrtc on all but the 386 kernel. | 03:22 |
BenC | I'll try to get this in for dapper release, but no guarantees | 03:22 |
desrt | :) | 03:23 |
infinity | * amd64: Same as above | 03:23 |
infinity | ? | 03:23 |
BenC | good call | 03:23 |
infinity | I have no idea about RTC drivers on PPC, so I can't comment there, sadly. | 03:23 |
infinity | If you have an SMP G5, you could do some hwclock testing. | 03:23 |
desrt | what's the issue with having the driver built-in? | 03:23 |
desrt | kernel gets correct initial time without having to use hwclock in the initscript? | 03:24 |
infinity | But on amd64, we know we want it (heck, the installer does "echo rtc >> /etc/modules" on amd64, doesn't it?) | 03:24 |
infinity | desrt: No. | 03:24 |
infinity | desrt: hwclock and ntp (and other hwclock-using things) just plain don't work right on SMP systems without an rtc driver. | 03:24 |
infinity | desrt: So having it modular on our kernels that are now all SMP-enabled is just plain silly. | 03:25 |
desrt | btw: there is no 'rtc' module on ppc | 03:25 |
desrt | infinity; why not just load the module on boot? | 03:25 |
infinity | desrt: Because loading it on boot unconditionally it no different from having it built in? | 03:28 |
desrt | fair enough :) | 03:28 |
infinity | desrt: s/it no/is no/ | 03:28 |
infinity | Simplicity++ | 03:28 |
desrt | i thought the ubuntu kernel policy was "make everything a module that is humanly possible" | 03:28 |
infinity | I dislike adding more moving parts without a good reason. | 03:28 |
infinity | No, that's the Herbert Xu kernel policy. | 03:29 |
desrt | that's my kernel policy too :p | 03:29 |
infinity | Ours is more like "make things modular if there are conflicting ways to set up a system" | 03:29 |
desrt | or my pre-ubuntu-feeding-my-kernel policy | 03:29 |
infinity | You'll note we don't modularise the IP stack, for instance. | 03:29 |
infinity | Which CAN be done. | 03:29 |
infinity | But it's insane. | 03:30 |
desrt | wait.. the true test | 03:30 |
desrt | brb | 03:30 |
desrt | ok | 03:30 |
desrt | unix domain sockets -- not a module | 03:30 |
desrt | good job :) | 03:30 |
infinity | Yeah, I vaguely recall Xu building that one modular. | 03:31 |
desrt | who is Xu? i think i'd like this chap. | 03:31 |
infinity | Herbert Xu.. Was a Debian kernel maintainer for ages, worked for us on the kernel back in the warty and hoary days, and did some contract work for us in the breezy timeframe. | 03:32 |
infinity | Has nothing to do with us recently. | 03:33 |
fabbione | BenC: what about HUGEPAGE for sparc? | 03:56 |
fabbione | HUGETLB i mean | 03:56 |
BenC | I can do that too | 03:56 |
fabbione | is it worth? | 03:56 |
fabbione | or are we going to destabilize *? | 03:56 |
BenC | probably best to ask davem | 03:56 |
fabbione | i know he enables it in his kernels | 03:57 |
=== BenC is down 10 files to manually merge from the 360 he started with | ||
BenC | fabbione: might be a good idea then | 03:57 |
jbailey | BenC: 10 files?!? | 03:57 |
jbailey | slacker ;) | 03:57 |
=== jbailey hides | ||
BenC | it'd be a lot more if my perl skillz weren't soo good :P | 03:58 |
=== mgalvin [n=mgalvin@ubuntu/member/mgalvin] has joined #ubuntu-kernel | ||
=== BenC pushes a most likely unbuildable edgy 2.6.17-git kernel | ||
fabbione | BenC: cool :) | 04:40 |
=== ivoks [n=ivoks@ubuntu/member/ivoks] has joined #ubuntu-kernel | ||
BenC | yay, first 2.6.17-git build started! | 05:43 |
=== BenC waits for it to fail horribly | ||
BenC | if this works on the first try, I'll be worried | 05:43 |
crimsun | I expect it to break horribly in sound/ ;-) | 05:44 |
BenC | I reverted most of sound/ | 05:47 |
BenC | it's pretty much stock 2.7.17-git | 05:47 |
crimsun | oh excellent. | 05:47 |
BenC | except for my snd-power-mac forward port | 05:47 |
BenC | crimsun: do you go both ways...like if I send you patches can you submit them to alsa? :) | 05:48 |
crimsun | BenC: oh sure | 05:48 |
BenC | good, I need to get the snd-powermac changes merged | 05:48 |
BenC | benh has been bugging me to get it done for awhile | 05:48 |
infinity | BenC: Stop hitting on our volunteer contributors. It could be misconstrued as sexual harassment. :) | 06:39 |
BenC | hehe | 06:39 |
infinity | "Do you go both ways" indeed. | 06:39 |
jbailey | BenC: merged.. What did you take on doign sound? | 07:44 |
=== ivoks [n=ivoks@ubuntu/member/ivoks] has joined #ubuntu-kernel | ||
bluefoxicy | fuck that | 07:56 |
bluefoxicy | oops wrong window | 07:56 |
tuxmaniac | Guys isnt it linux-kernel-tree to download source from ubuntu breezy? | 08:06 |
crimsun | you probably mean linux-tree{-2.6.12} | 08:08 |
tuxmaniac | crimsun: yes.. linux-tree-2.6.12 will download the source I guess? I forgot and now in dapper its linux-source? | 08:13 |
=== lloydinho [n=andreas@130.225.236.251] has joined #ubuntu-kernel | ||
=== doko [n=doko@201.160.17.131.cableonline.com.mx] has joined #ubuntu-kernel | ||
=== doko_ [n=doko@201.160.19.41.cableonline.com.mx] has joined #ubuntu-kernel | ||
=== bluefoxicy [n=bluefox@c-68-33-112-13.hsd1.md.comcast.net] has joined #ubuntu-kernel | ||
bluefoxicy | woot! | 10:54 |
bluefoxicy | I wrote a patch to let me shift around entropy for mmap() and stack bases via kernel command line! ... now to build and test. | 10:55 |
=== frenkel [n=frank@j210092.upc-j.chello.nl] has joined #ubuntu-kernel | ||
=== frenkel [n=frank@j210092.upc-j.chello.nl] has left #ubuntu-kernel [] | ||
=== doko [n=doko@201.160.19.41.cableonline.com.mx] has joined #ubuntu-kernel | ||
=== frafu [n=frafu06@vodsl-8926.vo.lu] has joined #ubuntu-kernel | ||
zyga | back | 11:12 |
zyga | anyone awake? | 11:17 |
bluefoxicy | I'm awake | 11:24 |
bluefoxicy | And very bored. | 11:24 |
bluefoxicy | LD [M] drivers/media/video/zr36067.o | 11:24 |
bluefoxicy | CC [M] drivers/media/video/videocodec.o | 11:24 |
bluefoxicy | It seems I have much more boredom to come. | 11:24 |
=== bluefoxicy building a kernel package to test a patch | ||
zyga | I've got pictures of a kernel panic | 11:25 |
zyga | interested? :) | 11:25 |
bluefoxicy | ocrap, i have it set to build for like 59 different architectures | 11:25 |
bluefoxicy | abort, abort! | 11:25 |
zyga | (fspot is still working on that part though :) | 11:25 |
zyga | 59? :) | 11:25 |
bluefoxicy | didn't clean antyhing out | 11:26 |
bluefoxicy | http://rafb.net/paste/results/ImsI6i74.html <-- this is what I'm doing | 11:27 |
zyga | checking | 11:28 |
zyga | I'll post the pictures in a moment | 11:28 |
bluefoxicy | gotta love how I replace 2 lines of code with 23 lines ;) | 11:28 |
zyga | holy sh... :-) | 11:29 |
bluefoxicy | random_factor = get_random_int() % (4096 << (mmap_random_bits - 1)); <-- this is wrong :) | 11:29 |
zyga | I had just plugged an USB mic+headphones | 11:29 |
zyga | and a notification popup asked me to configure it... | 11:29 |
zyga | when did linux distros get that good! | 11:29 |
bluefoxicy | lol | 11:30 |
infinity | The more interesting thing isn't that it asked... It's whether or not it workes afterward. :) | 11:30 |
infinity | s/workes/worked/ | 11:30 |
=== bluefoxicy changes things directly inside the patch ;) | ||
zyga | good point | 11:31 |
zyga | well gstreamer seems to see it | 11:31 |
=== zyga is curious what kind of usb device it reports as | ||
zyga | hmm, bDeviceClass - 0 | 11:32 |
zyga | I don't know usb :P | 11:32 |
zyga | nyway | 11:32 |
bluefoxicy | lsusb | 11:32 |
zyga | pictures are ready | 11:32 |
zyga | bluefoxicy: I did use lsusb | 11:32 |
bluefoxicy | what'd you panic this time? | 11:32 |
=== bluefoxicy would imagine usb headphones == sound card | ||
zyga | bluefoxicy: probably right | 11:33 |
bluefoxicy | zyga: next april I should send a patch to lkml to fix panic() | 11:34 |
bluefoxicy | it occurs to me the kernel doens't actually panic | 11:34 |
zyga | #define panic() BUG() | 11:34 |
zyga | http://ubuntu.suxx.pl/2006--1/bugs/kernel-panic-0 | 11:34 |
zyga | (tla has left its marks on my soul) | 11:35 |
bluefoxicy | I'll fix this by having it mdelay() for a few seconds and then randomly printk() one of {"My mind is going","I'm scared, Dave; will I dream?","Daisy, daisy....","HELP ME!!!!!!} in a loop. | 11:35 |
zyga | hehe | 11:36 |
zyga | hrumpf... gnome-sound-properties doesn't remember the default sound device setting | 11:38 |
zyga | whoooa | 11:41 |
zyga | it works | 11:41 |
zyga | are my photos of any use? | 11:42 |
bluefoxicy | I see lots of numbers. | 11:43 |
bluefoxicy | so not to me :) | 11:43 |
zyga | there is a backtrace... | 11:43 |
zyga | http://ubuntu.suxx.pl/2006--1/bugs/kernel-panic-0/DSCN5461.JPG | 11:44 |
bluefoxicy | I know | 11:44 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!