/srv/irclogs.ubuntu.com/2008/06/19/#ubuntu-kernel.txt

baron1984I have a question about the kernel, if anyone can enlighten me for a moment00:56
baron19842.6.24-17 and -18 failed to resume from suspend entirely requiring me to cut power to the power supply itself before the computer would even boot again00:57
baron19842.6.24-19 resumes but makes a chirping noise from the speakers then says it failed to resume00:57
baron1984is there any harm in using suspend in light of this?00:58
baron1984I looked in the system log and all it says is: Jun 18 19:48:52 ryan-desktop gnome-power-manager: (ryan) suspend failed00:59
sisyphefrenchies here ?02:34
crimsun_BenC: the previous intrepid upload of alsa-driver (1.1ubuntu1) already contains maks_'s addition, which you noted per Kano in the changelog for 1.1ubuntu2.  What gives?04:31
=== pgraner__ is now known as pgraner
=== thegodfather is now known as fabbione
Kano32hi BenC , i have got a problem with the lrm package08:39
Kano32i raised the abi with08:39
Kano32DEBFULLNAME="Hotfix" DEBEMAIL="no@mail.ad" dch -v 2.6.26-2.0 Kernel ABI bump.08:40
Kano32then usually08:40
Kano32debian/rules clean08:40
Kano32should update the abi08:40
Kano32but it does fetch 1 not 2 as abi, it seems the lowest line is used not the top line of the changelog?08:40
Kano32sed -e 's/PKGVER/2.6.26/g' -e 's/ABINUM/1/g' >>08:42
Kano32this is exchanged08:42
Kano32but when i put the new abi in the last line of the changelog this is evalutated, something is really wrong there..08:42
Kano32the same worked with lum08:43
Kano32could somebody look at this please..08:54
Kano32stupid package08:57
Kano32hmm now i know it...08:58
Kano32you renamed the package from linux-restricted-modules to linux-restricted-modules-2.6.2608:58
Kano32then it can not work...08:59
=== CjNr18 is now known as CjNr19
saispohi all09:44
saispoit's normal in ubuntu-feisty branch there are no debian/abi/2.6.20.17-36/ ?09:45
xivulonat UDS I understood that that intrepid kernel will support suspend-to-disk also when swap is on file12:19
xivuloncan someone pls confirm that? 12:19
mjg59The kernel already supports suspending when swap is a file12:20
mjg59But there's no support for resuming12:20
xivulonmjg59: is that planned for intrepid?12:21
mjg59I'm afraid I don't know.12:26
cjwatsonBenC: bug 241295 is another udeb issue for you (or somebody)13:59
=== pgraner_ is now known as pgraner
xivuloncking, given your patch in ntfs-3g, do you think there is still any reason to use dm-loop (for intrepid)?14:30
xivulonwill that improve write performance?14:30
ckingxivulon: one I have a usable intrepid kernel on my 2 boxes here I can thoroughly benchmark it and compare it with the normal loopback14:31
ckingxivulon: dm-loop will give better performance if it has direct access to the block device - in the Wubi case it won't, so the difference is debatable14:32
xivulondoes it help with LVM(2) by any chance? that is in order to have resizable virtual disk14:33
ckingxivulon: I'm unsure of this.14:34
xivulonthx14:35
ckingxivulon: can you drop me an Email concerning the LVM(2) resizable virtual disk requirements so that I won't forget it and so I can mull over that question14:36
xivulonhttps://wiki.ubuntu.com/WubiIntrepid14:36
ckingexcellent!14:37
xivulonon that subject, do you think we can do anything on fsck for ntfs-3g?14:37
xivulonnot necessarily a full blown fsck, but at least to address most common (or easier) issues14:38
xivulonand turn off the dirty flag14:38
zulgah....I didnt notice the ubuntu directory in ubuntu-intrepidi git tree14:38
rtgzul: its the new LUM14:39
zulrtg: gotcha14:39
zuldrbd needs to be updated since I uploaded a new version today14:40
ckingxivulon: it makes me feel uncomfortable fsck'ing a ntfs volume. It's a userland tool that I have no personal experience of (yet! :-)14:40
xivuloncking, I know, but that is probably the biggest issue we have in wubi at the moment :(14:41
=== pgraner__ is now known as pgraner
ckingxivulon: I suppose resourcing the ntfs fsck fix is the question to ask.14:43
zulBenC: ping14:58
BenCzul: yo15:00
tseliotBenC, rtg: why is the -generic flavour of Intrepid's kernel Xen-enabled?15:00
zulBenC: for the drbd8 any reason why you chose the 8.2.6 version?15:01
BenCtseliot: so it can be used a guest in xen environments15:01
BenCzul: I chose the latest version I saw available15:01
zulBenC: ah because I just uploaded 8.0.12 version which was just merged from debian15:02
tseliotBenC: fglrx doesn't work with Xen kernels and I'm having problems with nvidia too15:02
BenCtseliot: tell fglrx to ignore that fact...I think if you do that, it works ok15:02
BenCtseliot: have you contacted nvidia to see if they can fix their end?15:03
zulI doubt they will :)15:03
tseliotBenC: I have contacted them because their new driver (there are 4 flavours now) doesn't build with realtime kernels but I haven't received any reply15:04
tseliotBenC: here's the error with the NVIDIA driver: http://pastebin.com/m3c096a1715:06
tseliotwith the -generic flavour15:06
emgenttseliot: heya15:08
tseliotBenC: I'm a bit concerned about this, since tjaalton and I should replace the lrm with the new packages in Intrepid15:08
tseliotemgent: hi15:08
tseliotBenC: I'm building with IGNORE_XEN_PRESENCE=1 IGNORE_CC_MISMATCH=115:17
BenCtseliot: yeah, new lrm for intrepid is uploaded, without nvidia and fglrx15:23
tseliotBenC: I'll try contacting NVidia on their forums and saying explictly that I'm the new maintainer. Let's see if they care to reply.15:24
BenCtseliot: that looks like something I can fix15:25
BenCtseliot: can you give me your wip packages?15:25
tseliotBenC: wip?15:25
tseliotah ok15:25
tseliotwork in progress15:25
tseliotyes, sure15:26
BenCcjwatson: does that af_packet bug imply that we should have another upload soon for alpha 1?15:26
tseliotBenC: I'll upload them right now to my webspace.15:27
BenCcjwatson: I can do one today15:27
=== lukehasnoname changed the topic of #ubuntu-kernel to: "Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | Latest kernel upload: 2.6.24-16.30 | Latest news: Intrepid plans open: https://wiki.ubuntu.com/KernelTeam/UDS/May2008 | Next meeting: April 22, 16:00 UTC | Kernel Team machine: http://kernel.ubuntu.com UPDATE THE TOPIC"
=== BenC changed the topic of #ubuntu-kernel to: Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | Latest kernel upload: 2.6.26-1.5 | Latest news: Intrepid plans open: https://wiki.ubuntu.com/KernelTeam/UDS/May2008 | Next meeting: June 24, 16:00 UTC | Kernel Team machine: http://kernel.ubuntu.com
zulhow do you just build the lum stuff in intrepid now?16:00
cjwatsonBenC: yes please - lack of DHCP is sort of fatal :)16:01
=== pgraner_ is now known as pgraner
BenCzul: there is no "lum stuff" in intrepid16:16
zulso I have to do the fakeroot debian/rules binary-generic ?16:16
BenCzul: there are ways of only building a portion of the kernel tree if that's what you are trying to do16:20
zulBenC: yes please :)16:20
BenCzul: mkdir ../build; cat debian/config/i386/config{,.generic} > ../build/.config; make O=`pwd`/../build prepare; make O=`pwd`/../build ubuntu/16:29
BenCthat will build just the stuff under ubuntu/16:29
BenCyou can even do things like ubuntu/aufs/16:30
zulBenC: thanks16:30
mst_Hey guys, I'm really stuck here: I cannot get Linux to see my IDE cdrom drive unless I boot with acpi=off -- I've tried just about every boot-time option with the stock Ubuntu kernel and just about every possible way of configuring things with a custom kernel to no avail... I need some help in figuring out what to do next to try and track down the problem 16:30
BenCmst_: the problem sounds like a BIOS issue16:31
BenCmst_: I suggest seeing if there's a BIOS update before doing further16:31
BenCmst_: the next step would be a bug report along with the usual info (dmesg with and without acpi=off, lspci -vvn)16:32
mst_BenC: I'm real frustrated because I can boot a LiveCD just fine -- I have the latest BIOS, its from Lenovo and it's real shitty... totally devoid of configuration options and info 16:32
BenCmst_: maybe just adding acpi=off /boot/grub/menu.lst (but I suspect without acpi you don't get suspend)16:32
BenCmst_: Oh, wait, you can boot the livecd fine (without acpi=off) but not your install?16:33
mst_BenC: yeah, acpi=off is not really a solution for a laptop system :-( 16:33
mst_BenC: yeah16:33
BenCmst_: which release is this?16:33
mst_BenC: 8.04 16:33
BenCis the installed kernel updated prior to having this problem? what if you boot the originally installed kernel (via grub menu)?16:34
mst_I started running 8.04 pre-release and this problem has persisted with every subsequent kernel update 16:35
mst_I suspect it has something to do with the fact that it's an IDE cdrom and a SATA hard disk16:35
BenCI can't see how a livecd boots fine, but the normal install doesn't work16:35
mst_the live cd boots, installs, and I can boot and everything works great, just no cdrom without "acpi=off" in the actual install 16:36
BenCmst_: does your BIOS have a way to set your hd controller to something like legacy/compatible/ahci?16:36
mst_and i've tried libata.noacpi=116:36
BenCmst_: but the livecd wouldn't be able to boot if it didn't find your cdrom :)16:36
mst_That's why I suspect that libata is taking over control of the IDE ports (verifed by cat /proc/ioports) but wont recognize my drive 16:37
BenClibata is supposed to do that16:37
mst_I tried compatable and achi mode, each with combined_mode=ide and combined_mode=libata 16:37
mst_Yeah, but even with libata.atapi_enabled=1 I get nada... and I've tried all of these options as parameters in /etc/initrd-tools/modules, and having made and used the new initrd 16:38
=== BenC changed the topic of #ubuntu-kernel to: Ubuntu kernel development discussion ONLY | Kernel Wiki: https://wiki.ubuntu.com/KernelTeam | Latest kernel upload: 2.6.26-2.6 | Latest news: Intrepid plans open: https://wiki.ubuntu.com/KernelTeam/UDS/May2008 | Next meeting: June 24, 16:00 UTC | Kernel Team machine: http://kernel.ubuntu.com
BenCatapi is enabled by default in libata16:39
BenCso you can likely stop chasing that route, since libata is going to be the way your drive is detected16:40
BenCmst_: I'd be interested in a dmesg from a livecd boot, and one from a normal install boot (without acpi=off) to compare16:40
mst_cool, so what I think I might need to do is tell the kenrel explicitly what ioport to use for the cdrom drive 16:40
BenCif it works in a livecd, it should work in the installed system16:41
mst_I'd be happy to colleced that data, should I go about doing that now? I can but them both up in my public http space on my schools server 16:41
BenCmst_: yes, please16:43
mst_Alright... I have some rebooting to :-), be back in a few minutes 16:43
lukehasnonameif you had it on a VM you wouldn't have to bring down your machine :)16:46
BenCalso probably couldn't reproduce the problem either :)16:47
mst_lukehasnoname, wouldn't the kernel just see virtual hardware? I have to be honest I don't know too mcuh about that 16:47
mst_alright rebooting...16:47
BenCcjwatson: new kernel+lrm uploaded, I'll get meta out when it's near finished (which should be quick)16:48
BenCcjwatson: it is an ABI bump16:48
cjwatsonok, thanks16:48
BenCcjwatson: builds for intrepid appear to be amd64=1:20hours and i386=2hours16:56
BenCjust FYI16:56
mst_Alright... 17:08
mst_BenC: you should be able to see 3 files at www.uvm.edu/~mtretin 17:08
mst_BenC: I did dmesg for the installed system in AHCI mode and in Combatible mode 17:08
mst_as well as AHCI mode for the liveCD 17:09
mst_BenC: are you able to access them? I don't really ever use the public_html directory... 17:19
BenCmst_: weird, it's just plain not even probing the device...I was expecting a failure msg of some kind17:22
BenCmst_: do you have 2.6.24-16-generic on the installed system? If so, can you give me ahci dmesg from that?17:22
BenCthat's what the livecd is booting, and I want to compare the same kernel on both sides17:22
mst_BenC: yep... just one reboot away 17:23
mst_brb17:23
bdmurrayrtg: I seem to remember talking about bug 213011 in Prague but don't recall what you said.17:23
BenChmm...ide-pci-generic.all-generic-ide17:23
BenCthat shouldn't be there either17:23
rtgbdmurray: looking...17:25
bdmurrayrtg: thanks17:26
rtgbdmurray: these guys need to be running the -19 kernel. there was a cosmetic fix that I think will help them out.17:27
mst_BenC: it should be there nw17:27
bdmurrayrtg: alright, I'll ping them17:27
rtgbdmurray: I'll add it to the report17:27
bdmurrayrtg: is that in -updates yet?17:27
rtgyep17:27
mst_BenC: thank so *very* much for all the help BTW17:28
bdmurrayI'm so confused running -proposed all the time17:28
rtgbdmurray: I hear you there.17:28
BenCmst_: np...I tried to catch you before you left out, but can you also remove "ide-pci-generic.all-generic-ide" from the kernel cmdline...I don't think it is an issue, but just to keep a consistent test, redo the last dmesg without that17:30
BenCmst_: and thank you for taking the time to track this down17:30
mst_BenC: no problem... just let me get vim up...17:31
mst_BenC: dmesg.2.6.16-generic.achi-mode.txt should now be w/o the bogus kernel parameter 17:37
mst_or argument... to be technical lol 17:37
tseliotBenC: check your mail or the kernel mailing list. I have sent you the links to the source of my packages17:38
mst_BenC: and as far as me taking the time to track this stuff down... I couldn't be happier that I'm finally talking to someone who isn't telling me to do "modprobe ide-cd" :-)17:39
=== mkrufky is now known as mkrufky-away
=== pgraner_ is now known as pgraner
BenCmst_: I'm at somewhat of a loss as to why this is happening18:07
BenCmst_: can you try "sudo modprobe sr_mod"?18:07
mst_BenC: still no /dev/scd0 18:08
mst_BenC: woudl a dmesg with the option acpi=off (when the cdrom appears as scd0) help? 18:09
mst_BenC: I would love to figure out exactly what the addresse for my cdrom is, then couldn't I try ide0=xxx at boot time? 18:10
BenCmst_: is there anything in dmesg when you modprobe sr_mod?18:13
BenCmst_: also, try "sudo modprobe -r ata_piix; sleep 1; sudo modprobe ata_piix"18:13
mst_BenC: modprobe sr_mode doesn't put anything interesting in dmesg, except possibly "Driver 'sr' needs updating - please use bus_type methods" 18:15
mst_BenC: removing and reinserting (using your command) give me some output, but I don't really know whether any of it is useful or not 18:16
mst_output in dmesg that is18:16
mst_no cdrom in either case though 18:18
mst_is there a way of forcing the kernel to probe more deeply, equiv. to hdX=cdrom? like scd0=cdrom? 18:19
BenCNot that I'm aware of18:32
BenCit's probing, just not seeing anything18:33
BenCWhat I'm concerned about is what's the difference between the livecd boot and the installed system...they should be identical18:33
BUGabundohi there18:34
BUGabundoI'm having trouble with linux-restricted-modules-2.6.24-19-generic18:35
BUGabundogetting this /var/lib/dpkg/info/linux-restricted-modules-2.6.24-19-generic.postinst: 7: lrm-manager: not found18:35
mst_BenC: So what's the next step? And, I know there is a boot parameter to tell the scsi system to probe more deeply for devices on a scsi channel... you think that might help? 18:39
mst_BenC: I'm going to try adding sr_mod to /etc/initramfs-tools/modules and hope that it does something different... I figure as long I i'm trying something I'm moving in some direction. Thanks for all the help though! 18:53
Kanoanybody interested in a live cd 2.6.26 with auto installer of nvidia 177.13 and fglrx 8-6? (ubuntu config -xen +highmem64g)19:11
=== mkrufky-away is now known as mkrufky
Kanoit seems you are unable to produce a intrepid live cd ;)19:21
infinitykees: Yes?19:58
keesinfinity: oh, meant the other one, but this works19:58
keesRejected:19:58
keesavm-fritz-firmware_2.6.20.17.29_amd64.deb control file lists section as main/restricted but changes file has main/misc.19:58
keesavm-fritz-firmware_2.6.20.17.29_i386.deb control file lists section as main/restricted but changes file has main/misc.19:58
keesthe "old" problem was that the Section was wrong in the debian/control file19:58
keesthat was fixed -- all releases now match:19:58
keesPackage: avm-fritz-firmware19:58
keesArchitecture: i386 amd6419:58
keesSection: restricted/misc19:58
infinitymain/restricted?  Ouch.19:59
infinityThat's busticated.19:59
keesI'm a bit at a loss for what's going on here19:59
infinityLooks like a broken debian/control to me....19:59
infinitykees: Where can I snag this source?20:00
keesinfinity: dak?20:01
infinitykees: Mmkay.20:01
rtgkees: will it still be there if its been rejected?20:01
* infinity goes to find it.20:01
keesrtg: it got rejected by soyuz, so it's still in dak20:01
infinityrtg: The source will be there regardless.20:01
infinityLa, la, la... Bandwidth, FTL.20:05
keesinfinity: isn't "restricted/misc" valid?20:05
infinitykees: Should be.20:05
cjwatsonmain/restricted sounds like somebody did "Section: restricted"20:05
keescjwatson: right, that's the old hiccup20:06
keesand the warning I saw was:20:06
kees2008-06-19 17:55:39 WARNING Unable to grok section 'restricted', overriding it  20:06
infinityNo, that's the current hiccup.20:06
keeswith misc                                                                       20:06
kees2008-06-19 17:55:39 WARNING Unable to grok section 'restricted', overriding it  20:06
keeswith misc                                                                       20:06
keeswhich matches the same problem.20:06
infinityadconrad@lucifer:~/linux-meta$ dpkg-deb -I avm-fritz-firmware_2.6.20.17.29_amd64.deb | grep "^ Section"20:06
infinity Section: restricted20:06
cjwatsondebian/control.in.autogenerated.noreally?20:07
infinityIt needs to be "Section: restricted/misc"20:07
keeslinux-meta-2.6.20.17.29$ grep fritz -A2 debian/control20:07
keesPackage: avm-fritz-firmware20:07
keesArchitecture: i386 amd6420:07
keesSection: restricted/misc20:07
infinitydebian/control is autogenerated.20:07
infinityIs it not?20:07
rtginfinity: I see it, in control.common20:07
keesgah, control.common20:07
rtgshould be 'Section: main/restricted' ?20:08
infinityNo!20:08
keesrestricted/misc20:08
infinityShould be "Section: restricted/misc"20:08
rtgok20:08
cjwatsonmain/restricted is Soyuz "helpfully" expanding what it thinks Section: restricted means for you20:08
cjwatsonthe main effect of this is to make it more obvious that it's a mistake, because it looks weird :)20:09
cjwatson(actually it's more likely just an internal normalised form rather than intentional helpfulness or otherwise)20:09
alex_joniwhat's Soyuz?20:09
BenCand evil robot that controls everything20:10
cjwatsonalex_joni: the archive management component of Launchpad20:10
alex_jonicjwatson: ah, ok.. thanks20:10
BenCwhat cjwatson said is more PC I guess20:10
alex_joniPC = polite .. ?20:10
alex_jonipolitically correct?20:11
infinityThat.20:11
keesalex_joni: "politically correct" which is the PC term for polite.  ;)20:11
BenCthat one20:11
alex_joniah right.. that one ;)20:11
infinitycjwatson: Anything without an explicit component is shoved into main, so "main/restricted" is certainly the "correct" way to interpret "Section: restricted"20:11
cjwatsonindeed20:12
infinity(This is true of Debian too, for that matter, Debian/dak just won't reject for such an error, since it completely ignores a package's control field anyway, once overrides are in place)20:13
infinitys/control field/Section control field/20:13
cjwatsonBenC: linux-crashdump-{generic,server} targeted at main or universe?20:14
keesrtg: are you re-uploading with the fix, or do you want me to poke it?20:16
rtgkees: just did it. linux-meta_2.6.20.17.3020:17
* kees hugs rtg 20:17
mst_BenC: Alright... so i've been trying more stuff, and according to the libata FAQ there are detection problems with intel IDE mode drives (what I've got i'm sure), they suggest to play with force_pcs in ata_piix but modifno ata_piix doesn't have it listed as a param? 20:20
mst_BenC: shouldn't that men i cant call it at boot time? 20:20
BenCmst_: you're looking at old documentation20:49
BenCmst_: there's something more basic going on here...if it works in livecd, it should work on normal system, and the only reason for it not to is an ordering or race problem20:51
=== evand_ is now known as evand
BenCbrb, I need to force panic my machine for fun and profit20:55
hwildewhen I run "top" what is meant by "buffers" in the memory display?21:39
hwildeMem: 507664k total, 475780k used, 31884k free, 92116k buffers 21:39

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