[02:47] <Sarvatt> sheesh, device detection in drivers is such a mess
[03:01] <Sarvatt> apw: new nouveau kernel isnt working for me :(
[03:02] <Sarvatt> [   79.549480] nouveau 0000:05:00.0: display fifo hung :(
[03:02] <Sarvatt> [   79.565710] nouveau 0000:05:00.0: nouveau_fifo_free: freeing fifo 2
[03:02] <Sarvatt> [   82.565011] [TTM] GPU lockup dectected on engine 2 fence type 0x00000001
[03:02] <Sarvatt> [   82.566471] nouveau 0000:05:00.0: PFIFO_CACHE_ERROR - Ch 2/7 Mthd 0x1ffc Data 0xffffffff
[03:02] <Sarvatt> then a few thousand lines repeating :D
[03:09] <Sarvatt> oh i see, want calling nouveau.modeset=1 for that kernel. guess you cant use it any other what than in KMS with the kernel option :D
[03:24] <hyperair> hmm i wonder how well geforce4 mx support is in nouveau at the moment O_O
[03:58] <Sarvatt> looks like it supports every mx :D
[03:59] <Sarvatt> ugh, wasn't thinking, apw's kms kernel is versioned higher than 9.10 so alot of ati people on karmic probably just got a rude awakening to a non working setup since we dont have the drm for it
[04:41] <hyperair> Sarvatt: no 3d =(
[04:41] <Sarvatt> theres 3d!
[04:41] <hyperair> there is?
[04:41] <hyperair> but not for the lower models, right?
[04:41] <Sarvatt> as long as you dont mind glxinfo being the only thing you can run lol
[04:42] <hyperair> at least the nouveau status matrix didn't show any 3d support at all O_o
[04:42] <Sarvatt> ahh not sure
[04:42] <hyperair> bah!
[04:42] <hyperair> i'll be staying with my old outdated blob for the time being then
[04:42] <hyperair> no KMS for my desktop =(
[16:20] <crevette> superm1, hey, I seen bluez has udev manager now http://git.kernel.org/?p=bluetooth/bluez.git;a=commit;h=03971a2d726bc9a8630b2419508235a2920eb19f
[16:21] <crevette> thanks for pushing this uptream
[16:23] <superm1> crevette, yeah, as soon as the udev rules show up to match it, we should be in the clear
[16:24] <crevette> "in the clear"?
[16:35] <superm1> ready to disable the init script and switch to ondemand
[16:44] <crevette> superm1, I don't see any code that hook on udev, so I think the udev rules to start blutoothd is still necessary, but for stopping bluetoothd will stop by itw own
[16:46] <superm1> crevette, right i'm saying the init script would go away from rc.d
[16:48] <crevette> superm1, ah you would be extremist ? :)
[16:48] <crevette> you're true, karmic is for testing so let's test :)
[17:13] <bryce> morning
[17:25] <jbarnes> bryce: hi
[17:40] <Ng> bryce: when's the switch to kms on by default? :D
[17:41] <bryce> Ng: the kernel team is working on it; iirc they need to implement a whitelist/blacklist before switching it on.  apw may have more info
[17:42] <Ng> I say switch it on now and start collecting machines for the blacklist ;)
[17:42] <Ng> but I can say that because a) nobody will whine at me, b) it works for me ;)
[17:44] <Ng> my laptop has been up for a week now and apart from a couple of userspace things, I'm having a great time
[17:44] <Ng> at least two suspend/resume cycles a day, no X borkage or kernel borkage that I've noticed
[17:45]  * hyperair has black screens caused by gnome-screensaver
[17:45] <tjaalton> it was pain to use 2.6.28 on karmic, because 3g is broken on .30 :)
[17:45] <bryce> yeah it's quite solid on -intel for me as well.  we've had a few kms bugs in launchpad, but they're either solved now or nothing earth shattering
[17:45] <tjaalton> it doesn't work too well with the intel snapshot
[17:45] <bryce> but the blacklist/whitelist stuff is more for catching ! -intel cases
[17:46] <Ng> hyperair: where it suddenly switches the display off with dpms until you poke a key? I was wondering if that was g-p-m, but either way I'm seeing that
[17:46] <bryce> so we don't try to do kms on e.g. -nvidia, until it supports it
[17:46] <hyperair> Ng: that's another bug.
[17:46] <hyperair> Ng: the one i'm talking about is if close the lid, wait some time, then open it again, the screen won't come back on
[17:47] <Ng> hyperair: it is, but upstream think it's fixed in gpm 2.27.2
[17:47] <hyperair> no matter what you do
[17:47] <Ng> hyperair: ah
[17:47] <bryce> (xkeyboard-config 1.6 merge uploaded)
[17:47] <hyperair> +you
[17:47] <hyperair> Ng: it also happens if i close the lid faster than my notebook can suspend
[17:48] <apw> bryce, i wanted to check with you that my memory was correct, that the requied X bits to cope with and without KMS are enabled in karmic by default
[17:48] <Ng> hyperair: ouch. are you sure that's gnome-screensaver's fault? it shouldn't be doing screen on/off stuff, should it?
[17:48] <hyperair> Ng: i'm really really not sure, but someone on #intel-gfx said so
[17:48] <Ng> hyperair: I won't pretend to know more than them ;)
[17:49] <hyperair> Ng: he had slightly different symptoms.. i think just locking the screen for some time would trigger it
[17:49] <hyperair> Ng: and i'm not willing to try just yet =p
[17:49] <Ng> heh
[17:49] <hyperair> ssh works
[17:49] <hyperair> but you can't chvt or anything
[17:53] <james_w> <-- one more satisfied KMS user
[17:54] <bryce> apw: yes should be good to go.  Only -intel will do kms so far, so if you have a whitelist, that's the only one to be whitelisted.  
[17:55] <apw> no lists right now, just want to make sure that making the default kernel turn it on, and all my combo kernels the same, that userspace will work
[17:55] <apw> at least for intel
[17:56] <bryce> yes, there's no other changes on the X side we need for -intel for kms
[17:56] <bryce> erf, horrible grammar.  hope that made sense.
[17:56] <Ng> so it should be on for intel for alpha3?
[17:56] <bryce> Ng: yes
[17:56] <Ng> \o/
[17:57] <Ng> I'll test 855GM shortly after that is published :)
[18:19] <hyperair> bryce: speaking of KMS stuff, are you noticing KMS spitting flames with usplash?
[18:20] <hyperair> i find it especially fun to stare at a blank screen for 20 odd minutes while waiting for my fsck to complete
[18:20] <hyperair> starting from when i915 is loaded, ending when X is started
[18:34] <bryce> hyperair: no I hadn't noticed that
[18:35] <hyperair> hmm
[18:35] <hyperair> how strange =\
[20:23] <maco> hola. there was a wiki page for UXA testing where the graphics chips were laid out in a table and comments could be added. any intention of doing that again but for KMS?
[20:29] <bryce> maco -> https://wiki.ubuntu.com/X/KernelModeSetting
[20:30] <maco> bryce, ahead of me as usual....thanks
[20:45] <Sarvatt> hyperair: got a dmesg handy?
[20:45] <hyperair> Sarvatt: for the black screen thing?
[20:45] <Sarvatt> yep
[20:45] <hyperair> i don't remember seeing anything in dmesg
[20:45] <hyperair> lemme go poke my syslog
[20:46] <hyperair> hmm nothing
[20:46] <hyperair> lemme try reproduce the bug
[20:51] <Sarvatt> can I see it anyway even if theres no error? :D
[20:51] <Sarvatt> want to see the order of things and whats getting loaded in your case
[20:53] <hyperair> Sarvatt: strange, i can't reproduce it any more O_o
[20:53] <Sarvatt> i'll build you a new gnome-power-manager in a few minutes if you want, it wasnt updated for a long time there so ya got me to look and see it needed to be updated :) i dont see anything since karmic's that would fix your lid problem though
[20:53] <hyperair> Sarvatt: wait, are we talking about the gnome screensaver one or the usplash one
[20:53] <Sarvatt> usplash
[20:53] <hyperair> aah
[20:55] <Sarvatt> i've been having to build that myself since i need hal enabled in it for things to work right on all my machines
[20:55] <Sarvatt> (g-p-m)
[20:56] <hyperair> i see
[20:57] <hyperair> Sarvatt: anyway i'm not very concerned about the lid problem, since it's mostly intermittent and doesn't occur most of the time
[20:57] <hyperair> Sarvatt: the usplash bit.. i'm pasting by dmesg, but looks like pastebinit's taking some time O_o
[20:58] <hyperair> http://pastebin.com/f4639d2c5
[20:58] <hyperair> there we go
[20:59] <hyperair> don't mind the whole bunch of iptables stuff at the bottom. i was debugging my router's dhcp server
[20:59] <hyperair> and then i forgot to turn it off
[21:02] <hyperair> ....okay suddenly my screen blanked out on me for no apparent reason
[21:02] <hyperair> and it won't come back on
[21:02] <hyperair> wtf?
[21:04] <hyperair> well off i go to restart then
[21:04]  * hyperair sighs
[21:06] <Sarvatt> looks like it might be picking your TV to output to until X starts hyperair
[21:08] <hyperair> that's very strange
[21:08] <hyperair> oh and hell, i can't seem to get my comp to reboot now =.=
[21:08] <hyperair> reisub time i guess
[21:12] <Sarvatt> nevermind [   17.703535] allocated 1280x800 fb: 0x007df000, bo ffff88007d5ed0c0
[21:14] <Sarvatt> your drm isnt getting initialized until after i915, mine loads before and looks like this
[21:15] <Sarvatt> oh he isnt here, wish i could turn part messages on on a channel by channel basis :D
[21:16] <hyperair> i'm back!
[21:16] <Sarvatt> hyperair: edit your /etc/initramfs-tools/modules and add these on 3 seperate lines in this order
[21:16] <Sarvatt> intel_agp
[21:16] <Sarvatt> drm
[21:16] <Sarvatt> i915
[21:16] <hyperair> sudo vim /etc/initramfs-tools/modules
[21:17] <hyperair> oh shit
[21:17] <Sarvatt> could do i915 modeset=1 instead too
[21:17]  * hyperair facepalms
[21:17] <jbarnes> Sarvatt: has the xorg edgers kernel landed in the ppa yet?
[21:17] <Sarvatt> :D
[21:17] <hyperair> modeset=1 is in /etc/modprobe.d
[21:17] <Sarvatt> yeah it is in there, there is no metapackage for it yet though so its an opt-in thing to install it
[21:17] <jbarnes> Sarvatt: great
[21:18] <Sarvatt> i saw that bug report from the guy not having fbcon loaded, told him to install the kernel in edgers since it has KMS enabled in the kernel config so fbcon gets linked right
[21:20] <Sarvatt> hyperair: after you're done you need to update-initramfs too
[21:20] <hyperair> Sarvatt: yeah i'm running that
[21:21] <hyperair> Sarvatt: by the way, doesn't i915 automatically pull in drm and intel_agp when loaded?
[21:21] <cwillu> oh where, oh where did my working suspend go?
[21:27] <Sarvatt> its so much nicer having i915 built in because you get a high res KMS console like 2 seconds after the boot starts :D
[21:27] <Sarvatt> but usplash doesnt show the throbber/progress bar that way
[21:28] <superm1> is it still going to be possible to turn off modesetting if it turns out to be unstable on some of your hardware with i915.modeset=0 once it's fully built in?
[21:29] <bryce> another question is if we build-in i915, would we have to do that for nouveau and radeon as well?
[21:29] <tjaalton> I don't think keybuk wants those built in
[21:29] <bryce> and if so, then how do we handle if say -fglrx or -nvidia needs to unload those modules in order for them to do their thing
[21:30] <bryce> heya tjaalton
[21:30] <tjaalton> hi
[21:31] <bryce> can the module loading be made to occur sooner so we can get the benefit without having to build them in?
[21:31] <tjaalton> they are not going to be included in the initramfs, because it would slow down booting up
[21:32] <Sarvatt> superm1: about turning it off, you boot with "nomodeset"
[21:32] <superm1> Sarvatt, okay good
[21:33] <Sarvatt> and no we wouldnt have to do it for radeon, thats a seperate option in the staging area on the kernel config and only applies to those
[21:34] <Sarvatt> i dont think theres any chance in heck nouveau KMS is going to be in karmic, been testing it out the past few days and have it all working but its incompatable with radeon KMS right now and doesnt work at all with KMS disabled
[21:34] <tjaalton> what happens if you remove the module from the console? :)
[21:34] <tjaalton> while running kms
[21:39] <Sarvatt> Duke`: have you tried the updated xserver on xorg-edgers with no xorg.conf to see if it works right?
[23:08] <Duke`> Sarvatt: not yet, but I have to leave, sorry... I'll try tomorrow if I have time
[23:08] <Sarvatt> no worries, was just letting you know i updated it and it should be fixed now