=== fmasi_afk is now known as fmasi [04:38] infinity: I'm traveling until Monday. Is the 3.5 SRU a big hurry? [04:43] xequence, you should be fine, the cycle is just starting we haven't even gotten all of *our* kernels in the pipe yet. [04:45] bjf: Thanks. Wasn't sure if it was the scheduled SRU. Next week then. === fmasi is now known as fmasi_afk [05:19] apw: smb`: morning! Do you guys have recommendations for an x86 machine brand (remote access of serial preferable) to use for kernel build and boot testing? [05:20] s/of/through === fmasi_afk is now known as fmasi [05:44] moin === fmasi is now known as fmasi_afk === fmasi_afk is now known as fmasi === gavin__ is now known as gavinguo === smb` is now known as smb [06:49] morning [06:53] amitk, Maybe not really machine brand. I'd probably look for a newer Xeon based board. You probably need to think a bit more storage controller as one older in one of our builders rather slows things down. Personally I like supermicro boards with IPMI controller for remote administration. === fmasi is now known as fmasi_afk === fmasi_afk is now known as fmasi === fmasi is now known as fmasi_afk === fmasi_afk is now known as fmasi [07:31] smb: thanks will look at something with a supermicro [07:32] * apw whines [07:35] smb: our main usecase is testing patches against x86 as we mangle various core subsystems [07:36] smb: so a server class machine isn't a requirements per-se. We're looking for wider coverage across x86 cpu generations [07:37] amitk, Yeah, I just said Xeon for compile time. If you are not need a lot and quick you probably can just go with some i5 or i7. [07:38] IPMI was what I found just too useful when working from somewhere else and having complete control over the box at home (reset, turn off). Intel has that AMT stuff but frankly it seems they don't wnat it to be used by "normal" people without Windows [07:39] apw, Whassup? Bad morning or just morning in general? [07:40] ugg morning, no [07:41] apw, Just go back to bed man. :) (in case you are not still) [08:17] * ppisati -> reboot === fmasi is now known as fmasi_afk [09:10] * apw has a long and intense chat with apt-cacher-ng about its performance [09:19] * apw gets his 12lb sledge out and wacks her a few times [09:21] apw, You need some lower expectations [09:22] smb, heh just discovered that nightly 'maintenance' has been failing for the last 4 months, so it has not been expiring volatile files [09:22] and after much f*king about with it, a manual zap of the cache was required to sort its head out [09:23] Heh, I could imagine I got the same here... Not really likely to find out one has too many files [09:26] no, except when they are volatile and don't get appropriatly timed out [09:26] so you get a release updater from april [09:28] apw, With your down, my pragmatic fix would have been rm -rf cache/* [09:29] yeah i did that, that was the sledge hammer [09:29] ah [09:54] and in fact its behaviour is random, sigh, a few refreshes and all is well, wtg [12:18] * henrix -> lunch === kentb-out is now known as kentb [14:03] rtg, yo ... goldfish, was there meta changes to go with your kernel changes, i don't see 'em in the repo [14:03] apw, not yet [14:04] rtg, do you have 'em done somewhere so i can base on top of it ? [14:04] even if they arn't final yet [14:05] apw, there is a goldfish branch in ubuntu-saucy-meta [14:05] rtg, yeah am based on that, but there is no i386 in there, if you anr't planning one that is ok i can ignore [14:05] apw, no, that I don't have yet [14:06] ok i'll ignore and we can cross that when we get there === kentb is now known as kentb-afk [14:31] * rtg relocates === kentb-afk is now known as kentb === fmasi_afk is now known as fmasi === fmasi is now known as fmasi_afk [15:11] apw, https://launchpad.net/~canonical-kernel-team/+archive/ppa/+build/4959398/+files/buildlog_ubuntu-saucy-i386.linux-goldfish_3.4.0-0.4%7Epre1_FAILEDTOBUILD.txt.gz [15:19] rtg yeah saw ... am fixing ... [15:30] ppisati, P and Q ti-omap kernels ? [15:31] apw, do you have Mir running on any radeon GPUs ? I tried it on my Lenovo last night. It fell over with a resounding thud. [15:32] bjf: i'm finishing one thing, than i'll do those [15:32] ppisati, cool, just a reminder [15:43] rtg, i don't think i have a radeon ... and i am unsure where i am running mir :) [15:43] how did you detect you were [15:45] apw, I set it up according to instructions on the wiki [15:54] rtg, oh ... i see :) got a pointer [16:00] rtg yes hang on [16:03] infinity, whats the story on reiserfs in debian land ? I can't seem to find a discussion about _why_ reisferfs has been removed. [16:03] rtg: Because jail? I have no idea. You'd probably want to talk to Ben about it. [16:04] infinity, just wondered if you any insights. was just reading the ubuntu-devel thread [16:04] rtg: It's not something I've been following because I have zero carefactor. [16:05] xnox is proposing to disable it in the kernel, but I'm not so sure. [16:05] -1 [16:05] xnox, yo ... what was the background to debian dropping reiser ... i can see that we should drop it in the installer ... but kernel side we carry a heap of things enabled cause they are there and nothing more [16:06] bjf: rtg: it's not supported in upstream kernel, hence the reasoning to remove it in the debian kernel config. Are you going to maintain/keep-building reiserfs in newer kernels and backport stacks? [16:06] apw: why should anyone use reiserfs today? [16:06] apw: or floppy disks? [16:07] xnox, how is it not supported ? There have been 3 commits in the 3.12 merge window [16:07] rtg: oh, wow =) [16:07] xnox, indeed but we have a floppy disk support just in case [16:07] apw: ok. one sec. [16:07] for that one special character somewhere in iceland who uses one [16:07] apw: who also someone has PAE capable kernel?! =) [16:08] s/someone/somehow/ [16:08] xnox, if it's in linus' tree there there is still "support". is it broken? is it trashing peoples data? [16:08] hey i don't have to agree with all the decisions we have to amke :) [16:10] hmm ikonia is not here, dangit [16:11] rtg: apw: bjf: so here is the original email from Ben Hutchings https://groups.google.com/forum/#!topic/linux.debian.maint.boot/QO187Szy2_o with the list of modules he is dropping in the debian kernel config. [16:12] rtg: apw: bjf: at the same time debian-installer dropped support for those as well (e.g. reiserfs, ufs [!kfreebsd-any] ) [16:12] xnox, that hardly seems to be sufficient justification. there are plenty of drivers that are 'barely maintained upstream' [16:13] rtg: ok. But is it ok for the installer to drop support for those filesystems in Ubuntu? [16:13] xnox, yes [16:13] and relevant udebs. [16:13] xnox, so that just says that they won't package reiser so that it could be used in the installer [16:13] whilst stock kernel keeps on building them. [16:13] xnox, it says nothing about whether they kernel is still using them [16:13] xnox, I'm fine with that as well since only the installer (really) uses udebs [16:13] producing them [16:14] xnox, perhaps we need a bug for this in which we can list the udebs you no longer need, and we can adjust as appropriate [16:15] apw: true. I didn't check if the kernel config was actually changed to drop those, or whether it was a pure udebs change. [16:15] xnox, the wording as i read it is talking only about udeb changes [16:15] xnox, its just a packaging change [16:16] and indeed perfectly sensible ones i suspect, who uses floppies for instalation media now [16:16] get your pile to 2000 floppies and start feeding [16:17] apw: first operating system I installed in my life was with a CD-ROM drive. [16:17] =/ [16:17] and my last three machine don't have CD-ROM drives & installed with usb-sticks. [16:20] rtg: bjf: apw: ok, sorry about the confusiong wrt. dropping installer support vs [ udebs as well [vs kernel config/modules as well]]. Please reply on the mailing-list that e.g. "kernel team will keep the config, but can drop udebs on request and doesn't mind not supporting those options in the installer" [16:20] or some such. [16:21] xnox, too late, I already did [16:21] reply on the list, I mean [16:21] rtg: thanks. I had some delay receiving it. So only received it after yours "too late, I already did" [16:22] =) thanks a lot. [16:22] no prblem, email isn't exactly real time [17:51] rtg, ok in the ckt PPA is a set of linux, linux-meta, linux-goldfish, and linux-meta-goldfish packages, with the proposed changes for linux-tools applied. they are all ~preN [17:52] apw, ack [17:52] rtg, i am basically happy the concept is there i think and if you have no objections would proposed to upload them (with basically no other changes) [17:52] so ... let me know :) [17:52] ok. gimme a bit [17:52] once they are in and good, i'll do the 'rest of the world' to match [17:52] rtg, no rush [17:53] rtg, if you are looking for 'why is it like this' type information the whole why is documented in the linked bug [17:53] apw, ack === sabayonuser is now known as tuxkalle [18:06] Hi, I am having an intermittent issue with a 3.10 kernel on 13.04, on two different machines, and even on arch too [18:07] It intermittenly just hangs on boot and does nothing [18:07] And these are different patchset 3.10 kernels too it happens on [18:07] Has anyone seen this on 3.10? How may I try and troubleshoot? [18:08] I know that yes, they are not official ubuntu kernels, but they are pacvkaged for ubuntu and lots of users use them. [18:08] And it also happened using a 3.10 mainline kernel from the ubuntu kernel ppa mirror [18:08] TsarObomba, i don't think a significant number of people ran with 3.10, though my feeling was 3.10 wasn't very good [18:08] i have had much less pain with 3.11 [18:10] good to hear [18:10] One of my favorite patchsets should have 3.11 soon [18:11] pf already does, but they dont package for debian or ubuntu, so i gotta wait for the guy who makes those packages to update his mirror [18:11] And liquorix (the guy from zen kernel, damentz, makes it) hasnt updated. they just did 3.10.3 last night [18:15] apw, what branches have you pushed with the tools packaging name changes ? [18:23] rtg, ok ubuntu-saucy master-next and goldfish, ubuntu-saucy-meta master and goldfish, all pushed [18:23] apw, cool, thanks [18:41] Does ubuntu have a ubuntupatched 3.10 kernel? [18:41] for 13.10 or something? [18:41] I know it has mainline, but id like to pull the .config from a ubuntu patched kernel [18:42] TsarObomba, the config in teh mainline kernel build was seeded from the ubuntu config at the time. there are also tags for all out kernels in the git repo including any 3.10 uploads we did, you can also get those same source trees from the archive [18:43] * rtg -> lunch [18:44] i found what i was lookiung for [18:45] im trying to recompile a 3.11 kernel [18:45] with a different patchset and make it for deb [19:04] apw: where could i maybe find the source for the deb of the mainline kernels? [19:04] like this: http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.10-saucy/ [19:04] I just want the .config from it [19:04] i guess i can open up the deb [19:06] TsarObomba, the config is applied as one of the patche sin the mainline kernel directory [19:06] eh [19:06] its ok [19:06] i just extracted the deb and pulled it [19:08] TsarObomba, we try and be very transparent to allow people to reproduce you [19:08] s/you/what we have done/ [19:09] yup [21:15] * rtg -> EOD [21:22] apw, ping === kentb is now known as kentb-out === lifeless_ is now known as lifeless [23:46] hi, i have a question which may no directly fit in here, but maybe someone can still answer it: I am currently working with mount namespaces and want to get a single directory to have write access while the parent directory should have read only access (The directories are on the same mount point). [23:46] is it possible to do a mount bind before unsharing the mount namespace so that i can remount the filesystem as readonly while the mount bind still keeps right access?