| VladimirShushkov | Hi, mates! Sorry for my english and maybe stupid queston. I dont use linux for a long time since the times of kernel 2.6.10 maybe. Now I have installed Ubuntu and I want to rebuild my kernel. I want to compile kernel 3.0. I tried to find ppa for this kernel. I just added ppa:kernel-ppa/ppa to my repository. But after apt-get update and trying to execute "apt-get install linux-headers" I didnt see linux-headers v.3.0, only 2.6.38. Maybe I | 07:22 |
|---|---|---|
| ohsix | why do you want to build your own kernel? | 07:24 |
| ohsix | you can install kernel packages from mainline that are built almost nightly, or you can install something like xorg-edgers which has a 3.0 kernel in it; to support the updated drivers in the same repo | 07:24 |
| jk- | VladimirShushkov: `apt-get install linux-headers-$version` ? | 07:26 |
| VladimirShushkov | ohsix: Why do you ask about reasons to build new kernel? Is it bad idea? In the past I used slackware linux and always built my own kernel after linux installation. I built some of kernel features right in kernel but some left in modules. Building of you own kernel give you some definition. Also in the past I found that some of kernel features (e.x. packet filtering features) were not included in kernel nor in modules. Maybe today is diff | 07:37 |
| VladimirShushkov | jk-: Sorry, I dont understand what do you mean. As I said after typing apt-get install linux-headers- and pressing tab I see all candidates for installation: root@WorkstationX:~# apt-get install linux-headers- | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-10 linux-headers-2.6.38-8-virtual | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-10-generic linux-headers-generic | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-10-server linux-headers-lbm-2.6.38-10-generic | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-10-virtual linux-headers-lbm-2.6.38-10-server | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-11 linux-headers-lbm-2.6.38-11-generic | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-11-generic linux-headers-lbm-2.6.38-11-server | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-11-server linux-headers-lbm-2.6.38-8-generic | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-11-virtual linux-headers-lbm-2.6.38-8-server | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-8 linux-headers-server | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-8-generic linux-headers-virtual | 07:41 |
| VladimirShushkov | linux-headers-2.6.38-8-server | 07:41 |
| smb | VladimirShushkov, I think it would be linux-headers-<something> while the something is the same as you see when you run uname -r | 07:47 |
| smb | (like generic or generic-pae or server) | 07:47 |
| smb | morning .+ | 07:50 |
| VladimirShushkov | ohsix: thanks, xorg-edgers is what i need. One stupid question. What do you mean saying mainline? Is it kernel.org? Because when I used slackware I just downloaded kernel source in tar.bz2 format and used it. | 07:51 |
| smb | VladimirShushkov, It is the same, but instead of taking certain tar version, the source is taken from git directly and then compiled with the same configuration options as the released kernel packages. Any ubuntu specific modules are missing, but those are less often needed. | 08:03 |
| Q-FUNK | howdy! any news on re-enabling vesafb on mainline kernels? | 08:05 |
| VladimirShushkov | smb: Thanks for your explanation. As I see mainline doesnt contain kernel 3.0 for natty, only for oneiric. If I will use oneiric kernel dont I break depencies? | 08:07 |
| smb | VladimirShushkov, There are no hard dependencies. The natty or oneiric postfix rather tells you which config options where used to build the kernel. There always may be problems but rather due to the fact that a newer kernel could potentially have changed interfaces to user-space. Though 3.0 in Natty is working | 08:09 |
| VladimirShushkov | smb: thank you! | 08:11 |
| smb | Q-FUNK, No (though that merely means "I have no clue whether there are news") | 08:11 |
| Q-FUNK | smb: ok :) | 08:48 |
| apw | ppisati, yo ... cve-tracker ? | 09:39 |
| ppisati | are you and smb mentally connected? :) | 09:42 |
| smb | ppisati, Didn't you know. ;-P | 09:42 |
| ppisati | good to know :) | 09:42 |
| ppisati | anyway | 09:42 |
| ppisati | https://code.launchpad.net/~ubuntu-security/ubuntu-cve-tracker/master? | 09:42 |
| ppisati | this one? | 09:42 |
| smb | Not if you want to change anything | 09:43 |
| smb | That is the other teams original copy | 09:43 |
| ppisati | https://code.launchpad.net/~canonical-kernel-team/ubuntu-cve-tracker/kernel-team | 09:43 |
| ppisati | ok | 09:43 |
| ppisati | this one | 09:43 |
| smb | Yup, looks better | 09:43 |
| smb | ppisati, Anything in special that you need to achieve? | 09:44 |
| ppisati | no no | 09:44 |
| ppisati | just flip a cve status | 09:45 |
| ppisati | seems lp and cve tracker are not connected yet | 09:45 |
| smb | ppisati, They should be. But its something that apw has to trigger (or any real person). So it probably just runs or ran | 09:46 |
| apw | ppisati, they are connected but there is security-team latency involved | 09:47 |
| apw | but there isn't any problem flipping the state if you want to | 09:48 |
| apw | though other than 'not-affected' there isn't many that make sense to apply manually | 09:48 |
| * apw watches his upgrade go at 160k/s ... sigh | 09:49 | |
| apw | i presume the mirrors are still being hammered | 09:52 |
| apw | ppisati, for -2479 do you know which commit added the the huge page support in error ? | 10:39 |
| ppisati | apw: huge page support was added in 71e3aac0724ffe8918992d76acfe3aad7d8724a5 | 10:41 |
| apw | ppisati, so we can represent when something is added too on the break-fix line and get the not-affected status for free ... /me tries it out on this cve | 10:42 |
| Kano | hi apw , why are daily builds older than rc9? | 11:17 |
| apw | older in date or older than commit list | 11:20 |
| Kano | they are not build | 11:20 |
| apw | that is not really an answer to the question | 11:21 |
| Kano | http://kernel.ubuntu.com/~kernel-ppa/mainline/daily/ | 11:21 |
| Kano | 10-04 is older than rc9 date | 11:21 |
| apw | but by the looks of it linus has not pushed his master bracnh correctly to kernel.org | 11:22 |
| Kano | but there are much newer commits there | 11:22 |
| apw | there == where ? | 11:22 |
| Kano | http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=summary | 11:22 |
| Kano | what should be wrong there? | 11:23 |
| apw | Date: Sun Feb 20 10:15:57 2011 -0800 | 11:24 |
| apw | interesting, the latest commit if you pull the branch from kernel.org is nothing like that in gitweb | 11:25 |
| Kano | really weird | 11:25 |
| Kano | well usually i parse web gits to dl snapshots | 11:27 |
| * apw switches mainline over to github again, we'll see if that changes anything | 11:28 | |
| Kano | any diff there? | 11:37 |
| apw | Kano, take a while before we know | 11:45 |
| apw | cking, yo ... i have just confiermed my 35s of black screen is a win7 thing, if i set acpi_osi="!Windows 2009" it no longer occurs :( | 12:38 |
| Kano | did anybody manage to suspend a system booted via uefi correctly? | 12:39 |
| apw | what sort of machine is it, ie is it real uefi | 12:40 |
| apw | and not broken mac near a bit efi | 12:40 |
| Kano | i7 with q67 | 12:41 |
| apw | you could easily be the first | 12:41 |
| apw | does the machine suspend if you enable bios emulation? | 12:41 |
| Kano | it does not even suspend correctly win7 64 bit sp1 via efi | 12:42 |
| Kano | it works fine booted via mbr | 12:42 |
| Kano | somehow suspend does not seem to be that easy with efi... | 12:43 |
| apw | nope, efi is a bit of a disaster at the moment | 12:43 |
| nusch | Is there avaiable or could someone build for me standard 3.0.0-12-generic with CONFIG_RCU_TRACE option enabled ? | 12:43 |
| apw | nusch, i doubt there is one available, i might be able to make you one, what you need it for | 12:44 |
| apw | (and which architectures) | 12:44 |
| Kano | bbl | 12:44 |
| nusch | I installed new RAM and got message about RCU detected CPU 0 stall | 12:44 |
| nusch | trying to examine where exactly it locks | 12:44 |
| apw | did it come out of the stall ? | 12:44 |
| nusch | yes | 12:45 |
| nusch | kernel runs about 30x slower than normally | 12:45 |
| nusch | looking at tty lines | 12:45 |
| apw | and you get constant stalls ? | 12:45 |
| apw | all from adding some new RAM ? | 12:45 |
| nusch | yes | 12:45 |
| nusch | 8GB instead of previous 4GB | 12:45 |
| nusch | this is x86_64 | 12:46 |
| apw | doesn't seem like a change we'd expect to affect the kernel | 12:46 |
| apw | none of those values are exactly remarkable | 12:46 |
| smb | If you assume the RAM has no flaw | 12:46 |
| apw | and you've run a memory check? | 12:46 |
| nusch | memtest looks ok | 12:47 |
| nusch | and it boots on another hardware without problems | 12:47 |
| nusch | some people sugested buggy BIOS, is it correctable by kernel? | 12:47 |
| apw | i wonder where it got put, in the memory map | 12:47 |
| apw | if it is a physical layout issue triggered by the bios, not so easy | 12:47 |
| apw | could you pastebin a dmesg from the kernel with 8G inserted | 12:48 |
| apw | i wonder if the mtrrs could be wrong | 12:48 |
| smb | Yeah either that or maybe the timing goes wrong. Some BIOS may decide something bad based on what is in SPD | 12:49 |
| nusch | I asked people at ##kernel their also suggested mtrr but said it's ok, here is my dmesg https://bugs.launchpad.net/ubuntu/+source/linux/+bug/875655 | 12:50 |
| ubot2 | Launchpad bug 875655 in linux "Kernel extremally slow after upgrade from 4 to 8GB RAM on x86_64 machine, "RCU detected CPU0 stall"" [Undecided,Confirmed] | 12:50 |
| nusch | it was run with kernel mem=3G, becuse otherwise unusable. Here is another http://pastebin.com/YtLNPiCc with 8GB inserted and no kernel mem= option | 12:51 |
| * smb wished that dmesg would not just print the hard to understand mask format... | 13:14 | |
| smb | Though it feels like only the first 3G and then 1G above 4G are cacheable... | 13:15 |
| nusch | smb, and what is the reason of such behaviour ? | 13:24 |
| smb | nusch, Don't know exact. Actually I start to wonder whether it matters that much in 64bit mode... My system here also only has the first 4G declared in mtrrs and does not care that much | 13:26 |
| nusch | smb, and what about PAT, if PAT is enabled doesn't it mean MTRR don't have to strore correct information ? | 13:27 |
| smb | Thinking along similar lines. But I am sufficiently in doubt about my knowledge here... | 13:28 |
| smb | nusch, Theoretically there could also be the problem that the module reports incorrect RAM timing (or the bios fails to make a sense of it). Unfortunately netbooks usually have nothing useful (in the sense of tunable) as a bios | 13:33 |
| nusch | I hope that's not striclty hardware issue, that's why I want check it with CONFIG_RCU_TRACE | 13:35 |
| smb | nusch, Maybe you could add the output of "sudo dmidecode" to the bug report. That may contain some more bits of info | 13:38 |
| nusch | smb, I have no access to this computer at the moment , can send evening, but if it helps I have DSDT dump there | 13:43 |
| smb | nusch, Don't think the DSDT helps us here. That would be rather static info, while the dmidecode could show some info about what the bios detected/has set up | 13:46 |
| nusch | ok, will try that this evening, BTW could I pass commands like dmidecode | tee to kernel init= parameter ? Because it was not possible to boot system to bash prompt anytime. I was only able to sysrq reset or poweroff after few seconds pressing keys | 13:49 |
| smb | nusch, It should be ok to run the command even with mem=3G. As long as the memory module is present the bios will configure it and dmidecode reports it. Just the kernel make no use of the additional memory | 13:52 |
| nusch | ok | 13:52 |
| * ogasawara back in 20 | 14:16 | |
| bjf | ogasawara: are you still handling the oneiric kernel or has the batton been passed to stable ? | 15:11 |
| ogasawara | bjf: should probably pass to stable now | 15:12 |
| === smb` is now known as smb | ||
| Q-FUNK | who updated the mainline kernel today? :) | 15:34 |
| evt | Can someone explain the "isra" suffix I see on some symbols via /proc/kallsyms? (ex: do_anonymous_page.isra.40()) | 15:56 |
| === firewave is now known as FireWaveJob | ||
| nusch | smb: here is my dmicodede you asked http://pastebin.com/bKeT9iuw | 16:46 |
| nusch | smb, sorry that one was without one module mounted, paste in a moment | 16:53 |
| pmatulis | jsalisbury: hello sir, do you still need any vm's on kvm-server-2? | 17:35 |
| DBO | hi, I am having issues with my realtek wireless. rtl8192ce is the wireless driver I am currently using. It seems to randomly stop talking to my access point at which point I have to reconnect to keep talking | 19:55 |
| DBO | it really makes pushing to bzr a frustrating experience :) | 19:56 |
| jjohansen | DBO: your lucky it doesn't oops your machine, what kernel version are you using | 19:59 |
| jjohansen | DBO: I have found the oneiric driver much better than (still not good), the natty one | 19:59 |
| DBO | oneiric latest | 20:00 |
| DBO | is there a better driver to use? | 20:00 |
| DBO | or something I can do to improve its stability and usefulness? | 20:00 |
| bjf | ogasawara: what's your timeframe on uploading the first precise kernel ? | 20:10 |
| ogasawara | bjf: am about to send out email for the proposed -generic and -server config flavors. Assuming I get Acks, I'd like to upload tomorrow. | 20:11 |
| bjf | ogasawara: cool | 20:11 |
| Q-FUNK | bjf: we have separate desktop and server flavors, now? are the desktop ones full preempt? | 20:12 |
| jjohansen | DBO: try black listing the rtl8192ce driver and see if another of the rtl drivers will handle it | 20:12 |
| bjf | Q-FUNK: that's a question for ogasawara | 20:12 |
| Q-FUNK | ΓΆΓΆ, yes | 20:13 |
| ogasawara | Q-FUNK: we have separate flavors at the moment, we're proposed merging them. the desktop uses the voluntary preemption model. | 20:13 |
| Q-FUNK | ogasawara: why merge them? | 20:14 |
| DBO | woops, xchat crashed | 20:14 |
| ogasawara | Q-FUNK: because aside form the preemption model and the scheduler, the configs are almost identical | 20:14 |
| bjf | Q-FUNK: this is being discussed on the ubuntu kernel team mailing list | 20:14 |
| Q-FUNK | ogasawara: the preemption model is already a significant factor. | 20:14 |
| ogasawara | Q-FUNK: that's why we're discussing on the mailing list | 20:15 |
| ogasawara | Q-FUNK: and I'd like to keep the discussion there rather than having separate discussion on IRC and email | 20:15 |
| Q-FUNK | ogasawara: ok. fair enough :) | 20:17 |
| Q-FUNK | ogasawara: have you and apw had the time to figure out why 'vesafb' is missing from default mainline builds, btw? | 20:18 |
| ogasawara | Q-FUNK: sorry I haven't looked as I thought you'd discussed with apw | 20:19 |
| ogasawara | Q-FUNK: as he is the primary care taker of the mainline builds | 20:19 |
| Q-FUNK | ogasawara: we did. sorry if the hierarchy of the kernel team seems unclear to me at times. | 20:19 |
| Q-FUNK | ogasawara: we however never came to a definite answer about it. I noticed a new mainline build toda, but vesafb was still missing. | 20:27 |
| DBO | jjohansen, okay I tried that, no joy, blacklisting rtlwifi and rtl8192ce results in no wifi :/ | 20:27 |
| jjohansen | DBO: :( | 20:28 |
| DBO | is it possible that the driver isn't dealing with network manager doing its "scans" | 20:29 |
| ogasawara | Q-FUNK: might be best to follow back up with him tomorrow? I think he reset the mainline builds to point to github earlier today. | 20:29 |
| Q-FUNK | ogasawara: that's possible. I'm not updated on the situation. I'm merely waiting for vesafb to return so that I can test mainline as jsalisbury requested on several bugs. | 20:34 |
| jjohansen | DBO: from what I have seen, no its the driver having problems | 20:38 |
| DBO | jjohansen, is there anything I can do? It kinda makes hacking on Unity hard :P | 20:39 |
| jjohansen | DBO: I found a usb wifi dongle fixes the proble :/ | 20:39 |
| DBO | jjohansen, are we dependent on a fix from realtek or can we FOSS our way to a win here? | 20:40 |
| jjohansen | DBO: well FOSS away, but as I hear it the driver is pretty poor and needs a lot of work. It certainly take someone who knows their way around the wireless driver stack | 20:41 |
| DBO | jjohansen, I wasn't intending to imply *I* would fix it, more looking for hope :) | 20:42 |
| DBO | I'll see if I cant order an intel wifi card and just be done with it | 20:42 |
| DBO | swapping out hardware is fun | 20:42 |
| jjohansen | DBO: hehe :) I can give you the hope that I have only seen it crash oneiric once, which is better than natty where it was taking out the machine every 30 min | 20:43 |
| DBO | jjohansen, I have to admit, I find it a bit discouraging that the driver is in such a poor state | 20:43 |
| jjohansen | yeah /me too | 20:44 |
| DBO | jjohansen, I've never done wifi driver work, but I am an experienced hacker, is this a whole other world of pain or could I potentially help out here? | 20:44 |
| jjohansen | possibly, its just a driver, nothing magical. You need to learn the wifi driver stack some and maybe have card specs | 20:45 |
| jjohansen | I've never had the time to chase it | 20:45 |
| DBO | last one and I promise to leave oyu be: http://www.newegg.com/Product/Product.aspx?Item=N82E16833106061 will that be trouble free? | 20:46 |
| === yofel_ is now known as yofel | ||
| === emma_ is now known as emma | ||
| === emma is now known as EMMA | ||
| === EMMA is now known as emma | ||
| === emma is now known as em | ||
| apw | Q-FUNK (and ogasawara), i know why it is missing at least, it can no longer be a module -- fixing it is on my list | 22:30 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!