[03:54] <kayve> is it possible for me to use atomic_dec_and_test(atomic_t* v) in a user program?
[04:15] <kayve> is it possible for me to use atomic_dec_and_test(atomic_t* v) in a user program?
[07:12] <Keybuk> apw: any particular reason you've not uploaded a new linux-meta for lucid for the -8 ABI?
[10:25] <ghostcube> apw: startet again without installing anything before and with one day pausing pc
[10:25] <ghostcube> cam is dark again
[10:25] <ghostcube> :)
[10:25] <ghostcube> so all i can say its a bootup problem like you suggested from the beginning of this strange bug
[12:04] <indus> hi
[12:06] <indus> apw: hello 
[13:52] <tseliot> apw: it looks like there's no backlight interface for intel in /sys/class/backlight in karmic (with KMS). Do you know anything about this?
[13:53] <rtg> tseliot, apw is on vacation today
[13:53] <tseliot> rtg: ah, ok, thanks
[14:09] <Keybuk> rtg: why no -meta package update?
[14:09] <rtg> Keybuk, for the kernel? I suppose 'cause Andy hasn't done it yet.
[14:12] <Keybuk> I wondered if there was any reason for that
[14:15] <rtg> Keybuk, -meta is at 2.6.32.8.8, which references the latest kernel. Andy uploaded it about an hour ago (though he's supposed to be on holiday)
[14:18] <tseliot> smb: did my email convince you?
[14:20] <smb> tseliot, It sounded reasonable. I now only try to make up my mind whether the problem is grave enough to justify direct inclusion or I rather try to get it included in one of the next upstream stable releases and get it back from there.
[15:39] <rtg> smb, can you remind me what the ACPI knobs are for controlling C states?
[15:40] <rtg> I have a system that resets when it goes idle
[15:40] <smb> rtg, might be idle=poll,halt,... 
[15:41] <rtg> smb, isn't there a /proc/ or /sys knob?
[15:41] <smb> but maybe others... not sure about the later stuff I thought it was boot args
[15:42] <rtg> smb, looking in processor_idle.c. it appears to be a module param
[15:43] <smb> rtg processor.max_cstate=
[15:43] <rtg> smb, right
[15:44] <smb> rtg, At least does not show up in the parameters dir on my system...
[15:45] <rtg> smb, I'm reasonably convinced that this platform resets when it goes into deep C states, or is throttled to P0, both of which imply an ACPI/BIOS problem.
[15:46] <smb> rtg, Right, then I'd try either with the max_cstate=2 or idle=halt options on boot
[16:35] <casl> greetings. ubuntu 9.04 x86 32-bit -- after installing 2.6.28-17 last week, i am no longer able to VPN using cisco client (4.8.02); reboot to -16 kernel and i'm back in business. cisco_ipsec module loads OK, no errors, but connection is never negotiated. any suggestions on debugging this would be appreciated.
[16:47] <smb> casl, I am not sure I can help too much, but has the cisco module been recompiled for the new kernel?
[16:49] <casl> smb, thanks. yes. i re-run the cisco installer with each kernel update. has worked OK for last 6+ months. kernel module loads OK. no obvious errors reported. 
[16:53] <smb> casl, OK, cause the update had some connector callback changes that it needs to pick up. If the module has no debug option... maybe ethereal or was that wireshark nowadays...
[16:59] <casl> smb, thanks! i just realized that there are .o's in the installer directory that did not get rebuilt. going to reboot, make clean & see if that helps
[17:06] <casl> smb, THANK YOU!  "make clean" before the reinstall of the cisco vpn client seems to have picked up the needed changes!
[17:07] <tseliot> smb: opinions on my last emai?
[17:08] <smb> tseliot, maybe. err, if I would read it...
[17:08] <smb> tseliot, give me a few minutes
[17:08] <tseliot> smb: ok, thanks
[18:42] <bjaglin> hi, i am not sure if this is the right place, but i am having troubles building 2.6.32 under karmic. I am trying to build an unpatched kernel (to start with) from the ubuntu git repostiory, following https://help.ubuntu.com/community/Kernel/Compile. The compilation goes fine if I build latest ubuntu-karmic, but it just hangs always at the same place when trying to build ubuntu-lucid 
[18:42] <bjaglin> http://pastebin.com/f1530df67
[18:43] <bjaglin> is there some requirements on the compiler that Karmic does not fulfill for building this 2.6.32??
[18:43] <hyperair> i build my 2.6.32 kernel fine
[18:44] <hyperair> s/build/built
[18:45] <bjaglin> ok.. it's weird, because i don't get any error message, nor CPU usage once it hangs...  Always at: LD      ubuntu/dm-raid4-5/built-in.o - what is so special about it?
[18:59] <logari81> Hi, trying to compile the karmic-proposed kernel under pbuilder I get the following error
[18:59] <logari81> EE: Previous or current ABI file missing!
[19:00] <logari81> what am I missing?
[19:06] <rtg> bjaglin, use this page as reference: https://wiki.ubuntu.com/KernelTeam/KernelMaintenance
[19:13] <bjaglin> rtg: thanks, i already had a look at it actually, but it looks like I am facing a toolchain problem more than a config problem, since i managed to build 2.6.31 without any problem from the ubuntu git repo
[19:13] <rtg> bjaglin, I'm not aware of any toolchain isses wrt building lucid under karmic.
[19:20] <bjaglin> rtg: ok. Since it always hangs on the same file when linking, does it make sense to upgrade my binutils/gcc to the packages provided in lucid to exclude any problem there? the changelog for binutils between karmic and lucid does not look really relevant though...
[19:24] <rtg> bjaglin, I think that'll just make a mess. I think your best bet is to install a lucid chroot and build within that environment.
[19:37] <bjaglin> rtg: i'll try that, thanks
[20:45] <JFo> hi ogasawara_ I'm ready whenever you are.
[20:47] <ogasawara_> JFo: ok cool, I'll call ya in 2min
[20:47] <JFo> no problem
[21:27] <JFo> ogasawara_, think I lost you
[21:57] <bjaglin> rtg: found my problem, it had nothing to do with the toolchain... fsam7400 was recently moved from misc to its own directory in ubuntu/, and in the Makefile, there is a KERNELSRC=?/lib/modules/`uname -r`/build that was empty for me since I did not have the headers of my current kernel, and freezing the following awk call. isn't it a problem?
[22:01] <bjaglin> i am a pure newbie with kernel building, and please tell me if i am wrong but i find it strange that `uname -r` is used in a script, since the target kernel might be completely different than the kernel running, no?
[22:06] <bjaglin> i am referring to http://archives.free.net.ph/message/20091208.182445.ab0091af.en.html