Xsss4hell | any way to install the latest stable kernel wihthout much hassle and hoff? I don't want to comile it, can't I just aptitude it? | 01:51 |
---|---|---|
=== asac_ is now known as asac | ||
=== mdomsch_ is now known as mdomsch | ||
kraut | moin | 07:19 |
=== \sh_away is now known as \sh | ||
=== \sh is now known as \sh_away | ||
=== doko_ is now known as doko | ||
Kano | btw. the amd64 kernel is suboptimal, it does not show the splash | 08:19 |
bullgard4 | Does the Intel M processor use 16- or 32-bit Littleendian? | 08:22 |
ln- | BenC: ping. can you tell me what "Declined for Dapper by Kees Cook" means in https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.15/+bug/65631 ? | 08:51 |
ubotu | Launchpad bug 65631 in linux-source-2.6.15 "skge driver broken: invalid call to spin_unlock causes system crash" [Medium,Fix committed] | 08:51 |
ln- | (someone told me to ping you) | 08:51 |
* laga summons rtg | 13:31 | |
zul | is there a community kernel team meeting today? | 13:32 |
laga | i don't knopw | 13:32 |
laga | know* | 13:32 |
=== \sh_away is now known as \sh | ||
laga | rtg: looked at the patches already? | 16:28 |
rtg | laga: aufs? not yet. | 16:33 |
laga | rtg: good. aufs upstream is willing to look at the ubuntu kernel and see what it takes to make aufs work with it | 16:34 |
rtg | laga: if the kernel patches simply expose some symbols, then it ought not be too difficult. | 16:35 |
laga | rtg: that's all they do i believe. i'm not sure what's needed to work around some API changes introduced by apparmor (IIRC). | 16:37 |
laga | only the aufs guy can tell, i suppose | 16:37 |
laga | http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=summary <- is that the kernel tree he should look at? | 16:40 |
rtg | laga: yep | 16:41 |
juliank | laga: the new aufs upload in Debian (today or tomorrow) will fix the issues, using the patch Kano proposed. | 16:49 |
laga | juliank: i just saw your email. | 16:50 |
laga | juliank: sorry if parts of my msg were redundant, i wrote it earlier today and didn't have the opportunity to send it till now | 16:51 |
juliank | laga: who plans to include aufs into l-u-m? | 16:51 |
laga | juliank: rtg said he'd be cool with doing that if i submitted a patch. granted, he has to review the kernel patches proposed by junjiro first | 16:52 |
laga | juliank: with the patches in your package, does the kernel need to be patched at all? (except for the lhash patch which I do need) | 16:53 |
juliank | laga: I'm running the normal ubuntu kernel and it works. But some patches may boost performance or improve security. | 16:54 |
laga | juliank: it depends on rtg then. | 16:55 |
* laga digs up lhash patch | 16:55 | |
rtg | when you guys are ready, send the patches (or a reference to them) on kernel-team@lists.ubuntu.com so I don't forget them. | 16:56 |
laga | juliank: which patches would you recommend? | 16:59 |
juliank | laga: I never used them, but I would say: splice, lhash [nfs] , put_filp [nfs] | 17:03 |
laga | sec_perm is not needed anymore with your patches. | 17:04 |
* laga adds question mark | 17:04 | |
juliank | laga: Right. It's patched like unionfs in l-u-m, but I don't know if this is the best way. | 17:04 |
laga | juliank: does the sec_perm patch actually apply to the ubuntu kernel? | 17:05 |
laga | juliank: have you verified that the ia_file patch doesn't break anything? just commenting out stuff looks scary to me :) | 17:05 |
juliank | laga: by Kano. upstream did the same, see his last e-mail. - The problem with upstream's patch (although they may be better) is the size. | 17:06 |
juliank | everyone: Is security_inode_permission enabled in the Ubuntu kernel? | 17:07 |
laga | juliank: the last email from upstream is from 6:00am, correct? | 17:08 |
juliank | laga: 17:34 (CET) | 17:09 |
laga | juliank: i must be blind. the only email i see from around that time is the one i wrote (17:37) | 17:13 |
juliank | laga: Maybe it's not on the list yet. | 17:14 |
juliank | the patch is at http://jak.kicks-ass.org/~jak/upstream.diff | 17:14 |
laga | juliank: or he only sent it to you. anyways, if junjiro did the same thing, why not | 17:14 |
juliank | The question is: Are the API changes caused by AppArmor only? See the simple patch from Kano at http://jak.kicks-ass.org/~jak/01_vserver_apparmor.dpatch | 17:15 |
laga | juliank: if we put it into l-u-m, does it really matter how big that patch is? | 17:17 |
laga | juliank: i believe the ia_file api change is caused by apparmor patches - at least i found an apparmor patch which removes that member. vanilla 2.6.24 still has it. | 17:17 |
juliank | laga: I know this. I'm talking about the other changes. | 17:18 |
laga | ah. | 17:18 |
laga | interesting. that ia_file patch was indeed there earlier | 17:18 |
juliank | laga: The aufs-source package will be kept and people who build aufs with it should use the same set of patches like l-u-m. | 17:22 |
laga | juliank: sounds good. but which patch are we going to use? :) | 17:23 |
juliank | laga: Kano's patch is much easier to maintain. it's 2-4 kb, whereas upstream's patch is about 57 kb. | 17:25 |
juliank | And Kanotix is a live disk, which means it should work | 17:25 |
juliank | After Alpha 4, we should start building disks using aufs instead of unionfs (but keep it as a fallback, like in debian-live) | 17:27 |
juliank | laga, rtg: I'll provide a patch to add fs/aufs to l-u-m, based on the code in the aufs-source package (new revision not uploaded yet) | 17:30 |
rtg | juliank: thanks, much appreciated. | 17:30 |
juliank | rtg: It would just be good to keep most of the configuration in http://jak.kicks-ass.org/~jak/conf.mk - should I simply add CONFIG_AUFS=m to config/[archs] and keep the config.mk file inside fs/aufs/ - it exports many variables and EXTRA_CFLAGS | 17:34 |
rtg | juliank: If you figure out how to build it under l-u-m, then I can go from there. | 17:37 |
laga | juliank: how does the lhash patch fit in? :) | 17:47 |
Dekans | hello guys | 18:23 |
Dekans | what are compilation options for a live CD kernel ? | 18:27 |
dade` | https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/151016 | 18:28 |
ubotu | Launchpad bug 151016 in linux-source-2.6.22 "New in 2.6.22-13: system takes a LONG time to resume from suspend" [Undecided,Confirmed] | 18:28 |
Mithrandir | Dekans: it's the stock -generic kernel | 18:28 |
dade` | is this fixed on any new hardy kernel ? | 18:28 |
Dekans | I would like to make a live CD with a vanilla patched kernel | 18:29 |
dade` | someone in the last comments says it's fixed in -14-2 but there is no confirmation | 18:30 |
Dekans | Mithrandir: Mithrandir is the /boot/config-`uname -r` good ? | 18:32 |
Mithrandir | Dekans: it should be, yes. | 18:32 |
Dekans | and I can easily the -generic kernel of a live CD ? | 18:33 |
Dekans | +replace | 18:33 |
dade` | nobody cares about me :( | 18:33 |
dade` | hmm, an easy way to install hardy's kernel on gutsy ? | 19:09 |
=== \sh is now known as \sh_away | ||
juliank | Any list moderator here? | 20:57 |
juliank | rtg: I finished all patches, but the post "[l-u-m] [PATCH] UBUNTU: Add aufs module version 0+20080129" needs to be approved by a moderator first | 20:59 |
rtg | juliank: just did | 20:59 |
laga | wow | 21:22 |
laga | juliank: thanks for the patches | 21:22 |
laga | rtg: thanks for comitting them | 21:22 |
alex_joni | does anyone happen to know when pci_get_device() was added into the kernel? | 22:04 |
alex_joni | (do newer 2.4 kernels have that? or only the pci_find_device()?) | 22:05 |
Kano | hi, it seems you forgot to add the aufs config to lum | 23:40 |
laga | heh | 23:40 |
laga | yeah | 23:40 |
Kano | just compiled and it was without | 23:40 |
laga | took me some time to fix that | 23:40 |
rtg | Kano: workin on it. | 23:40 |
laga | rtg: great. | 23:40 |
Kano | did you see my casper patch (similar to live-helper)? | 23:40 |
Kano | you should upload a fixed casper package too | 23:41 |
laga | Kano: julian might be interested in that. i think he's going to provide aufs patches for the live disks tomorrow | 23:41 |
rtg | laga: the patch didn't apply clean to debian/config/i386 or amd64 because of a prior patch from Fabio. Simple enough to fix. | 23:41 |
Kano | i did that yesterady already | 23:41 |
Kano | will just change the patch position | 23:42 |
Kano | why are my 2 other patches are not used? | 23:43 |
Kano | http://kanotix.com/files/fix/casper-aufs.patch | 23:45 |
Kano | thats for casper | 23:45 |
Kano | then you can use union=aufs | 23:45 |
laga | hum. NFS branch not supported with aufs. either i did something wrong or the autodetection is not working. | 23:45 |
rtg | lum pushed | 23:46 |
rtg | I'm outta here | 23:46 |
* laga makes a mental note to bug julian tomorrow and leaves | 23:46 | |
Kano | at least i have got now only 2 extra module packages besides lum... | 23:50 |
ln- | BenC: ping. can you tell me what "Declined for Dapper by Kees Cook" means in https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.15/+bug/65631 ? (someone told me to ping you) | 23:52 |
ubotu | Launchpad bug 65631 in linux-source-2.6.15 "skge driver broken: invalid call to spin_unlock causes system crash" [Medium,Fix committed] | 23:52 |
Kano | ln-: how about using sk98lin and blacklist skge? | 23:52 |
BenC | ln-: means it wont be fixed in dapper | 23:52 |
ln- | BenC: any idea why not? | 23:53 |
BenC | ln-: without reading it, I can only guess because it doesn't meet our Stable Release Updates policy | 23:53 |
ln- | BenC: it's a security vulnerability; https://bugs.launchpad.net/bugs/cve/2006-7229 | 23:54 |
BenC | ln-: looks like it is fixed in dapper already...probably why it was declined...already fixed | 23:55 |
ln- | BenC: but it isn't fixed! | 23:55 |
BenC | ln-: * [UBUNTU:drivers/net] drop invalid spin_unlock calls in skge (CVE-2006-7229) | 23:56 |
ubotu | The skge driver 1.5 in Linux kernel 2.6.15 on Ubuntu does not properly use the spin_lock and spin_unlock functions, which allows remote attackers to cause a denial of service (machine crash) via a flood of network traffic. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2006-7229) | 23:56 |
BenC | ln-: that is in the latest changelog for 2.6.15 | 23:56 |
BenC | ln-: if you can show that this is reproducible in 2.6.15-51.65 upload, then I'd say take it up with kees | 23:57 |
ln- | BenC: please take a look at the skge.log attachment at the end.. i.e. this: http://launchpadlibrarian.net/11562782/skge.log | 23:58 |
ln- | ok, i admit there is no spin_unlock visible in it, but still an excellent way to crash the system. | 23:59 |
BenC | ln-: I'm going to need to see the version of the linux-image-2.6.15-51-386 package for that log | 23:59 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!