/srv/irclogs.ubuntu.com/2007/12/11/#ubuntu-kernel.txt

Kanoapt-get source linux-image-$(uname -r)00:00
yuhongAnd yes, there are NX capable processors that does not support 64-bit.00:00
yuhongIntel Core for example.00:00
Kanothen look into debian/binary-custom.d/ 00:00
Kanoand try it00:00
yuhongI want the config to be the same as generic with PAE enabled.00:01
Kanothen use it00:01
yuhongWhich why I am not asking ubuntu-hardened.00:01
Kanoa bit tricky to modifify linux-ubuntu-modules, but not that hard too00:01
yuhongAnyway, what is the status of this.00:02
Kanoyuhong: instead of complaining compile your own kernel i would say ;)00:03
JanCyuhong: it won't happen for gutsy00:03
yuhongNo, too late anyway.00:03
JanCfor hardy, you would have to convince the developers with very good reasons00:04
yuhongI can, but NX capable processors are so common and XP SP2 enables PAE with a compatiblity hack that there should be a generic-pae kernel.00:04
Kanoyuhong: another way: compile debian bigmem kernel00:04
yuhongOn NX capable processors, that is.00:05
yuhongWhy doesn't ubuntu have that bigmem kernel.00:05
Kanoit is called different00:05
JanCyuhong: does such a kernel boot on all systems that the current kernel boots on?00:05
yuhongNo, that is why I am not removing the generic-nonpae kernel.00:06
Kanoyuhong: in theory i could even write a script to compile such a kernel ;)00:06
JanCI don't think there is space on the desktop install CD for multiple kernels...00:06
yuhongXP SP2 handle this by having the bootloader not use the PAE kernel if the processor is not PAE capable or PAE is not enabled.00:06
yuhongEither by NX or otherwise.00:07
Kanowell linux is not xp00:08
Kanoyour kernel has pae enabled or not,you have to select it00:08
Kanoalso when you would know better about xp then you know the installer write this boot.ini option00:09
yuhongEven in Windows, there are separate kernels for each combination of SMP, non SMP, PAE, and non-PAE.00:09
yuhongThe bootloader choose the kernel used.00:09
Kanothat means when you exchange the cpu from a cpu without nx to a cpu with nx you will NOT use nx00:09
Kanoit is not dynamic00:10
yuhongThis kind of switching used in NTLDR is probably not feasible for Linux.00:10
yuhongBut at least have a PAE kernel.00:10
Kanothere is one ;)00:11
yuhongYes, but it is for servers.00:11
Kanonobody helds you back to use it00:11
yuhongI am recommending a desktop PAE kernel because almost all current processors have NX00:12
Kanobest complile a yuhung kernel, makes you feel better :)00:12
yuhongAnd enabling NX increases security.00:13
Kanoi highly doubt that00:13
yuhongWindows enables PAE to take advantage of NX on current processors too.00:13
Kanoyuhong: is windows secure because of nx? thats a marketing gag00:14
yuhongThough with a compatiblity hack that limits physical address space to 4G.00:14
yuhongBecause of driver problems.00:14
yuhongI agree that Windows is secure not just because of NX.00:14
yuhongBut still, windows does it.00:14
Kanoyuhong: usually the apps are more insecure than the kernel, so just attack em00:15
yuhongCorrect, but NX can prevent app buffer overflows too.00:16
Kanoscripts that can access other files from a webserver or whatever00:16
Kanodo you know how many php mistakes are exploitable00:16
yuhongI know, but NX can prevent app buffer overflows too.00:17
Kanoi see really no reason to extra harden desktop systems00:17
yuhongBut as I said, Windows does it too.00:17
yuhongAnd as I said, NX can protect more than just the kernel.00:18
Kanoeven apparmor is useless against stupid users ;)00:18
yuhongMaybe, but Windows does it too.00:19
Kanoyuhong: every one with phyiscal access to a linux pc can install backdoors or change everything until you encrypt all00:19
yuhongCorrect.00:19
Kanosame applies to win00:19
yuhongcorrect.00:19
Kanoso what attacks are you fearing most?00:20
yuhongBut stilll, Windows does that switching to PAE to enable NX too.00:20
Kanoyou are just too funny ;)00:20
yuhongI am suggesting this not because of attacks.00:20
Kanobecause of win i know *g*00:20
yuhongBut because NX can enhance security, thus Windows does it too.00:21
yuhongI mean, Windows does that switching to PAE to enable NX too.00:21
Kanosure but it still did not get any more secure by this, good example *g*00:21
Kanoat least not xp00:22
yuhongWindows were forced to enable NX only for system components by default because of appcompat issues.00:23
Kanoyuhong: and you are using win now,right ;)00:23
yuhongThankfully Linux is open source, thus any issues with NX and PAE can be fixed.00:23
yuhongBut I am running Ubuntu in a VM.00:24
yuhongSo no need for the hack that limits PA like with Windows.00:24
Kanodo you know, that a kernel with slower timer works faster in a vm?00:26
yuhongAnd saves power on laptops as well.00:26
yuhongBut I am not talking about timers.00:26
Kanoin your case: just install the server kernel, desktop optimisation is not required at all00:27
yuhongBut what about the many people who have processors that have NX.00:27
Kanoyuhong: they dont care00:28
yuhongNo, but NX is enabled in Windows by default in these computers.00:28
Kanodid they notice this? no00:28
yuhongNo, but it does enhance security.00:29
Kanoyou must be a real hacker target *g*00:29
yuhongIn fact, AMD and Intel advertises NX by another name.00:29
yuhongAnd I don't think it is sensible to have NX not enabled in Ubuntu by default when NX is enabled in Windows by default.00:30
Kanodefault is impossibel for old cpus00:31
yuhongNo but at least provide a PAE kernel.00:31
zulwe do its the server kernel00:31
yuhongWindows deals with this by bootloader autodetection.00:32
yuhongBut it is impossible for Linux.00:32
Kanobest write a patch for grub ;)00:32
yuhongExcept maybe the Installer can automaticly install the PAE kernel on PAE machines and enable it by default on NX.00:32
Kanoyuhong: as you are in your vm: check: grep -o nx /proc/cpuinfo00:33
Kanodo you get that flag00:33
yuhongWait.00:34
Kanoif you get no output then your discussion is completely useless00:35
yuhongFor me that is.00:35
Kanoas you could not even boot that kernel 00:35
yuhongWhich is why I am not removing the nonpae kernel.00:35
Kanoso you want something just for others ;)00:38
yuhongAnd also my physical processor support PAE.00:38
Kanothen boot it phyiscal ;)00:39
yuhongBTW, Vista allow you to shrink a volume.00:39
Kanobut is is too stupid to install on a partition without active flag set which every win version could do before ;)00:40
yuhongBTW, I am downloading the alternate CD to see if that will work.00:44
Kanoyou can resize with gparted too00:45
yuhongBTW, GRUB 0.97 is older than GRUB 1.9500:47
Kanoyou can count00:48
Kanobut thats a completely differnet branch00:48
yuhongCorrect.00:48
Kanodo you need grub200:48
yuhongBut GRUB 2 is better, and support EFI booting, important for Intel Macs.00:49
Kanothats absolutely unimportant00:49
Kanoas intel macs can boot with normal grub as well00:49
yuhongAnd GRUB 1.95 is 1 year newer than GRUB 0.97.00:49
Kanoyou only need it for booting00:49
yuhongBut CSM booting on Intel Macs is a completely different can of worms.00:50
yuhongFor example, external HDs and netboot.00:50
Kanowhen you look at intel mac you can say that the integrated madwifi driver is not working with em but a newer svn snapshot would do. but grub is totally unimportant00:50
yuhongBTW, does the radeonhd driver work with native EFI booting.00:51
Kanoyou do not need to boot efi natively as the mac has bios emulation00:51
yuhongIf it doesn't it is not hard to modify it so it does.00:51
Kanoas this was needed for win00:51
Kanoso why play with efi if not needed?00:51
yuhongBut the BIOS emulation have it's own problems.00:52
Kanoit works00:52
yuhongFor example, external HDs and netboot.00:52
Kanodo you own a mac00:52
yuhongNo, but I was thinking about how limited BIOS emulations in intel macs are.00:52
yuhongI once wished I could own one.00:53
Kanoyou are thinking, but i know kanotix users with intel mac and nobody needed something with efi00:53
Kanoyou can even boot cd using grub00:54
yuhongSomeone mentioned booting DOS with Intel Mac's EFI emulation.00:54
Kanosure, should do as well00:54
Kanono need for grub200:54
yuhongBut that depends on how good BIOS emulation in Intel Macs are.00:54
yuhongI heard of keyboard problems for example.00:54
Kanowhen it can boot win it is good enough ;)00:54
yuhongFor Linux, probably yes, but for DOS, it is a different matter.00:55
yuhongBut what about external HDs and netboot?00:55
Kanoyuhong: i hope you know any dos command *g*00:55
yuhongI heard of problems getting Intel Macs' BIOS emulation to boot from anything other than an internal hard disk or CD.00:56
yuhongFor example, netboot and external HDs.00:56
Kanoand thats your problem, right?00:57
Kanoyou can not use nx, nor you own a mac, so whats your problem00:57
yuhongNo, but that is why EFI Linux booting should not be abandoned.00:57
yuhongI can use NX.00:58
Kanoyuhong: you boot itanium with efi00:58
yuhongIt is not my problem.00:58
Kanoia6400:58
yuhongCorrect.00:58
yuhongyou always boot itanium with EFI,00:59
yuhongI heard of problems getting Intel Macs' BIOS emulation to boot from external HDs or a network.00:59
Kanothats not your problem,dont care00:59
yuhongNot really my problem of course, and indeed my problem is with NX.01:00
Kanoit si not you can use a kernel with nx enabled, a 64 bit system or you can compile your own kenrel01:00
yuhongXserves does not have BIOS emulation.01:01
yuhongNo not for me.01:01
yuhongBut for the average users, it is.01:01
Kanoyou can be sure, that no average users needs that01:01
yuhongAnd Windows enables it seemlessly.01:02
yuhongExcept that Windows enables it automaticly.01:02
Kanoyuhong: how long do you use linux?01:02
Kano10 min?01:02
yuhongNot really, more than that.01:03
Kanoso how long01:03
yuhongBut I read about linux a lot.01:03
Kanohow long do you use it01:03
yuhongI ran a Linux VM on another computer for a while01:04
Kanonever native?01:04
yuhongNo, unfortunately, because of disk space.01:04
Kanothe desktop iso images are live cds, you do not neven need a hd01:05
yuhongIn any case, Windows enables NX automaticly on processors supporting it.01:06
Kanoin any case you are too stupid to boot a live cd ;)01:06
yuhongI am not.01:06
Kanothen why didnt you do it01:06
yuhongLinux can't, so a PAE kernel is the closest thing.01:06
yuhongI prefer to install in a VM.01:07
Kanoyou did not boot a live cd because you need pae?01:07
yuhongNo.01:07
Kanoso why not01:07
Kanohd space is not needed01:08
yuhongThat is why the non-PAE kernel must be the default on Live CDs.01:08
yuhongBut installing to a HD is a different matter.01:08
Kanowhy did you never boot it?01:08
yuhongJust easier to install in a VM.01:09
Kanousally a standard user wants speed01:09
Kanothen he wants to play games, 3d games can not be used in a vm01:09
yuhongI am not stupid, I am thinking of the average user and comparing it to Windows.01:09
yuhongNot always.01:09
Kanoan average user does not even know about pae but he does know that he wants to use 3d, for compiz, games or whatever01:10
yuhongCorrect.01:10
Kanonobody cares about pae01:10
yuhongAnd Windows enables NX automaticly.01:10
Kanobut no user cares about01:11
yuhongA PAE kernel is the closest thing  for linux.01:11
yuhongI agree for the average user, but keep in mind that both Intel and AMD advertised PAE once.01:11
yuhongI mean NX.01:12
Kanomarketing yes01:12
Kanoamd said it would stop virues, it didnt01:12
yuhongRequiring NX for PAE may be a good thing.01:12
Kanonothing changed for xp01:12
Kanoabsolutely nothing01:12
yuhongNo not completely, espricially because Windows XP does not enable NX by default for all programs.01:13
Kanohow many linux virues do you know01:13
yuhongBecause of appcompat.01:13
yuhongBut linux is open source.01:14
Kanoyou know nothing,but you knwo this best01:14
Kanosweet dreams01:14
yuhongNot true.01:14
yuhongRequiring NX for PAE may be a good thing.01:15
yuhongBecause it finally push transmeta and VIA to add PAE support.01:15
yuhongIn any case, should there be a PAE kernel in Linux because NX is enabled by default in Windows and NX requires PAE?01:26
yuhongI mean ubuntu.01:27
yuhongSo in the end, will a PAE desktop kernel be available in Ubuntu? Because PAE is required to enable NX.02:42
=== asac_ is now known as asac
=== imbrando1 is now known as imbrandon
=== \sh_away is now known as \sh
=== imbrando1 is now known as imbrandon
=== _fs is now known as fs
pkhI want to get an updated sierra driver included in the ubuntu kernel.  can someone point me in the right direction for the procedure of dealing with this?10:14
aboganipkh: This is a good to start: https://wiki.ubuntu.com/KernelTeam/KnowledgeBase10:17
aboganipkh: Particularly https://wiki.ubuntu.com/KernelGitGuide in your case.10:17
tjaaltonsweet, a new l-r-m for 2.6.24. it needs newer nvidia/fglrx though, AIUI10:18
pkhabogani: thanks.  will dive into it.10:20
pkhi'm not sure what to do.  I want to get an updated sierra driver into the kernel, but i'm not sure whether I should be working from the ubuntu end or the "linus' tree"?10:38
dade`where can i get ubuntu sources for kernel 2.6.22-12 ?11:26
amitkdade`: git clone git://kernel.ubuntu.com/ubuntu/ubuntu-gutsy.git OR apt-get install linux-source11:30
dade`no11:30
dade`gutsy uses -14 by default11:30
dade`i need -1211:31
dade`the -14 has bugs i can't afford11:31
amitkdade`: Once you get the git tree, you can roll back to whatever version you need using git commands11:32
dade`this is more interesting11:32
dade`I found this https://launchpad.net/ubuntu/+source/linux-source-2.6.22/2.6.22-12.3911:34
dade`to match the -12 image, do i need to apply the .diff.gz patches, or they're applied already ?11:35
dade`i guess I have to apply them.11:35
aboganidade`: In your shoes i prefer git-archive or git-bisect...11:36
dade`yes but i need to learn how to use them, and I just want to enable CONFIG_NO_HZ11:38
dade`is someone working on this https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/151016 ?11:39
ubotuLaunchpad 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] 11:39
zdzichu_I have similar bug, https://bugs.launchpad.net/ubuntu/+bug/13508311:43
ubotuLaunchpad bug 135083 in ubuntu "resume from STR is very slow on Thinkpad z61t" [Undecided,New] 11:43
zdzichu_with no comments11:43
=== \sh is now known as \sh_away
=== doko_ is now known as doko
lkolbeHi! I'm trying to build the ubuntu kernel from the git repository (git://kernel.ubuntu.com/ubuntu/ubuntu-gutsy.git). following the procedure in https://wiki.ubuntu.com/KernelMaintenance, fakeroot debian/rules binary-arch does work, but dpkg-buildpackage does not - it complains loudly about files in build/ and .git that it can't represent changes in there.14:10
lkolbee.g., i can't build a source-package directly out of the git-tree14:10
lkolbeI know I'm missing something obvious here, I just can't figure out what ...14:10
rtg_lkolbe: try 'debuild -b' to see if it complains about missing dependencies.14:17
lkolbehm, thanks. I did install missing dependencies and try to build again now. 14:23
* lkolbe thinks d'oh14:23
rtg_lkolbe: if it still complains then 'make mrproper;git checkout -f;debuild -b'14:24
lkolbeokay, thanks rtg_! I'll let you hear if it works out ...14:26
lkolbehm, okay, different problems - I try to use pbuilder for the kernel build, and it tries to make a source-package:14:36
lkolbe"dpkg-buildpackage: failure: dpkg-source -b ubuntu-gutsy gave error exit status 1", it compains about "newly created empty file"s and 'cannot  represent change to build/scripts/basic/fixdep' and all that14:37
rtg_lkolbe: what is your goal ? An installable binary deb or a source deb? "dpkg-buildpackage -S -sa -rfakeroot -I.git -I.gitignore -i'\.git.*'" to build a source package.14:38
lkolbeI'm trying to build kernel .debs that I can dupload to our local repository for distribution :)14:38
lamontlkolbe: debuild -i.git14:39
rtg_lkolbe: to make the build faster, restrict it to one flavour to begin with: 'fakeroot debian/rules binary-debs flavours=generic'14:39
rtg_lkolbe: you should also have ccache installed and setup.14:40
lamontor what rtg_ said first. :-)14:40
* lamont shuts up14:40
rtg_lamont: dude, I love to hear different ways of doing stuff :) Like rebase v.s. merge. I listened to Randal's talk. I've been having the same argument with BenC.14:41
lamont"if you publish your changes for consumption by other than upstream, rebasing is unkind."14:42
lkolbeThanks alot. I'm still waiting for dpkg-buildpackage -S .. to complete. If it works, I shoudl be able to use pbuilder :)14:42
lamontI think that summarizes randal's explaination nicely.14:42
* lamont considers anew actually running pbuilder at least once in his life.14:43
rtg_lamont: I happen to agree with him. I have yet to really grok the advantage of Ubuntu rebasing as opposed to merge. 14:43
rtg_lkolbe: you're on your own with pbuilder :)14:43
lamontrtg_: if it's your tree, then rebasing is love.  it's everyone based on your tree that hates life.14:43
lamontso, if you want to win the discussion with BenC, take over pushing ubuntu-hardy, and rebase it, once BenC is doing active development that isn't immediately pushed into the tree.14:44
lamontrebasing keeps the gitk mess prettier14:44
lamontgitk screws anyone downstream, since commits go *poof*14:44
rtg_lamont: all of the kernel devs push to the repo, but once in awhile (supposedly on a schedule now) we sync with Linus.14:45
lamontright.  and as long as BenC is pushing to the repo, and not doing devel in a repo based off of ubuntu's repo, he won't understand our pain14:45
BenCthe basic idea is that rebasing makes it easier for us to move from 2.6.22 to 2.6.2414:46
lamontrebasing to 2.6.24 I can certainly see.14:46
BenClamont: I don't manage the main git repo's anymore :P14:46
lamontregularly rebasing once we're there is just pain14:46
lamontBenC: heh. ok14:46
rtg_lamont: perhaps you should be basing your hppa devel off of Linus tree, then reconcile your patches with Ubuntu periodically. We ought to be pretty close most of the time.14:46
lamontrtg_: so much for our planning in secret, eh?14:46
* BenC rains on the parade14:47
lamontrtg_: actually, my ubuntu-hppa tree isn't that active at all.14:47
lamontit's more what I've run into elsewhere.14:47
rtg_lamont: hell, I'm trying to get into this open source jive. Most of my career has been developing in the dark.14:47
lamontand then there's the challenges when hppa has a needed fix that isn't in the main repo, and is having challenges getting in because of a checkpoint-freeze..  those can be painful, and yet I understand the reasons for the lamont-pain-causing reluctance.14:48
lamontit hurts me when I have to tell myself "no".14:48
rtg_lamont: so how many hppa users are there? As many as there are ia64 users ?14:48
lamontless.14:49
rtg_At last count I think there were 4 ia64 users.14:49
lamontthere exists a small possibility that we don't fit on one hand anymore. :0)14:49
lamontia64 desktop? sounds about right. ia64 server? more14:49
rtg_I think the xen community has a hard time, and there are a bunch more users there.14:49
lamonthppa?  lets see... me, jbailey, (did we get BenC hardware?), t-bone, and probably a couple more14:50
rtg_though it will get easier as more ofit is pulled into mainline kernel.14:50
lamontyeah.  and most of the hppa stuff is pushed into mainline quickly, so we really don't have much pain.14:50
lamontkyle for the win14:51
rtg_lamont: do you think jeff actually moved his to mt view?14:51
BenClamont: I haven't received anything lately...my ia64 is dead, and the a500 hasn't booted a recent kernel since edgy14:51
BenCgutsy might work, but I haven't tested it14:51
lamontrtg_: he got new hppa gear in mtview... ggg is there, and all is lover.14:51
lamontBenC: I need to poke our guy to get me the ia64 box so I can install it14:51
rtg_BenC: maybe we should slice the ia64 baggage out of the kernel if there is no serious maintenance effort.14:52
lamontBenC: would you like a J6K?  (same rack space as an A500, no MP interface, but it can run 32- or 64- bit kernels)14:52
lamontoh, and gutsy is a little unstable on SMP A500, hence the ugly 2.6.22 hppa64 config14:52
BenClamont: yeah, that would be nice14:52
BenCrtg_: only baggage there is some config files...not much overhead in that14:52
* lamont makes a note to pester the nice admin guy to schlep the box to his cube14:53
lamontrtg_: there is pretty serious ia64 maint effort. just not many users...14:53
lamontand BenC should have working ia64 gear before new years, modulo UPS love.14:53
lamont(the freight company, not the backup power)14:54
lamonthrm.. time to get ready and go to the office for a 9AM meeting14:54
rtg_lamont: they seem to have a hard time finding his house.14:54
lamontmebbe I should give them lat-longs?14:55
lamontanyway, gone.14:55
zulhey14:56
=== zdzichu_ is now known as zdzichuBG
lkolbeyippie - dpkg-buildpackage -S -sa -rfakeroot -I.git -I.gitignore -i'(\.git.*|build.*)' finally builds the source-package. Now on to pbuilder. ... :)14:58
rtg_zul: speaking of xen...14:58
zulrtg_14:58
zulwhere? 14:58
rtg_zul: it just came up in the discussion of rebase vs merge. 14:59
zulrtg_: ah yes i was reading the backscroll redhat is doing alot of it now for 2.6.24 so we can use dom0 with xen paravirt-ops15:00
rtg_zul: do you think you will still need a custom flavour for xen?15:01
zulyep15:01
zulrtg_: fyi http://fedoraproject.org/wiki/Features/XenPvops15:07
rtg_zul: huh. looks like a lot of work left.15:13
zulyep15:13
humboltoHow well is the the ubuntu kernel patched for xen? Would you recommend it for a production environment?15:22
zuldepends but you would be safer using the kernel from xensource15:24
humboltohave not seen any kernels there.15:34
FlashBikerhello everyone15:39
FlashBikeri have a problem installing ubuntu on my Acer Aspire 5520G15:40
FlashBikermy nvidia sata controller doesn`t seem to be recognised15:40
FlashBikeranyone alive in here?15:50
FlashBikeranyone alive in here?15:52
reynaldo_FlashBiker: what kernel version are you trying to boot with and what are the details of your specific sata controller?15:53
=== reynaldo_ is now known as reynaldo
FlashBiker2.6.2015:55
FlashBikerreynaldo: i`m using 2.6.20-15-generic15:56
reynaldoFlashBiker: do you think you can try to boot with gutsy (2.6.22(?)) and see if the problem is still present ?15:58
FlashBikeryes it still persists15:58
FlashBikerbut i was told that...15:58
FlashBikerin the modules.pcimap... there is no entry refering to my device15:59
FlashBikerthis is my device16:00
FlashBikerPCI bridge: sVIDIA Corporation Unknown device 0563 (rev a2)16:00
FlashBikeroir if i do lspci -n16:00
FlashBiker00.00.0 0500: 10de:0547 (rev a2)16:00
FlashBikerthis is the first entry in modules.pcimap16:01
zultry booting with all_generic_ide16:01
FlashBikernow?16:02
FlashBikerhow?16:02
zultry googling for all_generic_ide16:03
reynaldoFlashBiker: sorry I didn't managage to help you further, I was looking to see if there where reported bugs against your kernel version and the specific controller you have but all I found was a claimed to be fixed bug against a similard card: https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/14500516:07
ubotuLaunchpad bug 145005 in linux-source-2.6.22 "[gutsy] kernel 2.6.22-12 doesn't boot with nVidia MCP65 SATA Controller pci id [10de:045d]" [High,Fix released] 16:07
FlashBikerubotu: so there is no kernel that supports my hardware?16:09
=== \sh_away is now known as \sh
reynaldoFlashBiker: do what zul said, and yes, there seems to be no specific support for your card16:09
reynaldoalltough that doesnt mean you cant use it16:09
FlashBikeri`m trying it now16:10
cr3in /proc/modules, I noticed that some lines end with (P). I thought that each line in /proc/modules consisted of the columns: name, size, instances, dependencies, state and offset. so, what's that final (P) column?16:10
FlashBikerit seems to have done something...16:12
FlashBikerbut now my xservers is having some buggs16:12
* reynaldo just relized ubotu passed the turing test :-)16:13
reynaldorel/real16:18
=== \sh is now known as \sh_away
reynaldobe back in half an hour16:51
BenCamitk__, rtg_: ping17:11
amitk__pong17:11
rtg_BenC: yo17:11
BenCI'm late as usual, but this should be quick17:11
BenCcr3: ping17:11
cr3BenC: pong17:12
BenCzul, mjg59`: ping17:12
* BenC rounds up the gang17:12
zulBenC: yo17:12
BenCThis is just going to be a quick kernel meeting to supplement the lack of #ubuntu-meeting/fridge being scheduled like it needs to be17:12
BenCAgenda is simple: 1) Go over current state of hardy kernel. 2) cr3 to discuss cert process and kernel team, 3) Open floor for questions/comments17:14
rtg_BenC: 1) We are behind the eight ball. Linus has released -rc517:14
BenCSo the basic idea right now is that linux 2.6.24 and accompanying lum/lrm has been uploaded17:14
BenCrtg_: I'm more concerned with everything compiling than staying close in sync with Linus, ATM17:15
BenClots of things are in dep-wait right now, so I'm not sure what the real state is17:16
BenCThe objective is to get all this sorted out by the end of  the week, so we can have a 2.6.24 based Alpha 2 on the 20th17:16
BenCAny questions about this simple plan? :)17:17
zulnot here17:17
rtg_I guess the one core-dev we have on the kernel team is gonna be busy.17:18
zuler we have 3 core-devs17:18
BenConly one that will be doing linux-2.6.24 source uploads17:18
BenCOk17:20
BenCcr3: you have the floor17:20
cr3I would like to announce that the hwtest framework, which is used as the basis for hardware certification and server enablement, now supports automated tests.17:21
cr3I would also like to inform you guys that steve george has suggested that the support department should start solving cases by also creating a script to reproduce the problem encountered.17:21
cr3this script could then be added to hwtest to make sure the problem doesn't occur again17:22
BenCcr3: is there an API for these scripts, or are they just generic types...and are they run as root?17:22
cr3the purpose of sharing all this information is that the same concept could apply to the kernel team: if you encounter a problem which is easily scriptable in something runnable from the command-line, I could add it to hwtest very easily17:22
BenCcr3: and how do we get feedback from these scripts?17:23
cr3BenC: the simplest way is to just write a script callable from the command line which returns 0 on success and non-0 on failure and, yes, this is run as root.17:23
cr3BenC: currently, certification.canonical.com. the fact this is a private website is a concern shared by myself and other team members, such as liw who works on desktop testing.17:23
BenCcr3: it's a good start, thanks for the framework. We'll have to see if we can hash out how to make this useful for our engineering efforts17:24
cr3BenC: from lack of a more suitable alternative, we agreed to use this private website and improve it with the experience we will gain during hardy. this experience could then be reused by the launchpad team when developing their hardware database website or some other testing site if necessary.17:24
cr3BenC: do you have somekind of setup for running tests for hardware vendors?17:25
BenCcr3: we only do manual testing17:26
BenCboot, install, try out all the hardware17:26
rtg_cr3: some of this stuff is hard to automate, like audio.17:27
cr3rtg_: I also have manual tests for audio and video, which you could reuse.17:28
rtg_cr3: cool.17:28
cr3so, I mentionned how scripts written by the kernel team could be useful to me when performing my own test runs. could the suite be useful to the kernel team as well?17:29
BenCcr3: is this hardware testing, or basic OS testing?17:30
rtg_cr3: perhaps in cases where we are trying to reproduce a bug.17:30
cr3by the way, I don't want this script writing to incur more overhead than the gains you would have, so we could start by just emailing me scripts you might write on the corner of the table17:30
cr3BenC: hardware testing, but there will be some OS testing done by other teams17:31
cr3BenC: err, when you say "hardware testing", do you mean like checking that there are badblocks or checking compatibility of Ubuntu with the hardware?17:31
BenCcr3: we're only interesting in spending our time on hardware testing17:32
BenCcr3: compatibility, and functionality17:32
BenC*interested17:32
cr3right, that's what I'm concentrating on as well. I have recently added a feature which relates tests with specific devices. so, I should eventually be able to generate a view where you see all devices on a machine and the number of tests passed/failed for each device17:33
cr3the relations are expressed with simple boolean expressions: 'net.80203' in info.capabilities17:34
BenCexcellent17:34
cr3or: 'fpu' in cpuinfo.flags and cpuinfo.vendor_id == 'GenuineIntel'17:35
BenCcr3: Could you provide me an email with more detail on this framework?17:35
cr3BenC: sure, I have a bug to document the framework some more. mind if I fix this bug and then email you the corresponding documentation?17:35
BenCcr3: sounds good17:36
BenCI'll toss the info on kernel-team@ and see if we can get some discussion and ideas17:36
cr3maybe this darn project should have a website with a mailing list at some point17:36
BenCcr3: thanks for your time17:36
cr3BenC: thanks for the audience, we'll keep in touch17:37
BenCAny topics people want to bring up now?17:37
zulnope17:38
BenCOk, then this meeting is adjourned :)17:39
BenCthanks everyone17:40
lamontBenC: so, given 2ea 73 GB drives, you want raid1?17:42
lamontor two drives?17:42
BenClamont: raid117:42
lamontok.  /me finally found the machine that the nice man schlepped over several days ago17:42
sorenmjg59`: rtg_ and I are interested why you think https://bugs.edge.launchpad.net/ubuntu/+source/usplash/+bug/139453 is a kernel issue?17:43
ubotuLaunchpad bug 139453 in usplash "colors are left on the screen after blanking with usplash enabled" [High,Triaged] 17:43
lamontsoren: sounds, uh, colorful. :-)17:44
sorendoh..17:44
soren:)17:44
zulBenC: has anyone considered going through sourceforge and looking for drivers compared to feature requests on launchpad?17:49
BenCzul: no, are you volunteering? :)17:50
BenCstability is a major issue when taking in drivers like that17:50
zulsure i have time17:50
mjg59`soren: The kernel's responsible for blanking the screen on the console. If it's not doing so, that sounds like a kernel bug.17:54
sorenmjg59`: Um... Yes, I suppose it's as simple as that :) I just kind of hoped you'd have a bit more of an idea.17:55
mjg59`Ah, no. I never got beyond that :)17:55
sorenOk. :)17:57
=== rpereira_ is now known as rpereira
=== chuck_ is now known as zul
reynaldoBenC: you around ?19:10
BenCreynaldo: yeah20:33
reynaldoHi Ben, I'd like you to mentor me for the device manager work on launchpad20:35
reynaldobut I'm unsure what the procces to atain that entitles20:35
reynaldoi'm talking about https://wiki.ubuntu.com/DeviceManagerSpec20:36
reynaldoits the only 'mentoring available' item in the ubuntu-kernel group20:37
reynaldohope you can understand what I'm saying, I'm not a native speaker20:37
BenCreynaldo: basically, just get to work on it :)20:39
BenCreynaldo: if you have any questions, feel free to email me20:40
reynaldoYeah, Id like to know if I can trust the blueprint contents to be rigth20:41
reynaldothey suggest this to be moved, in the 'work with the gnome dudes intead' sence20:41
reynaldothe basic idea was to for hal-device-manager20:41
reynaldofork20:42
reynaldonow it seems is just a matter of helping gnome-device-manager's upstream development. that's what got me confused20:44
reynaldoI'm more of an 'invent all the wheels again, as long as all other wheels are square' kind of guy :)20:46
BenCreynaldo: actually, there's some discussion outside of the spec, with some other developers working to do the same thing20:55
BenCreynaldo: email me, and I'll get you in contact with them20:55
reynaldoBenC, done, thanks for your time. hope to be of some help :-)21:02
reynaldoback in an hour21:14

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