[01:13] <RAOF> ScottK: I've probably asked you this before, but is there any way to stop kwin deciding that GL is dangerous when I'm deliberately crashing it?
[01:37] <RAOF> Oh, kwin, how I hate you.
[01:59] <RAOF> Grr, heisenbugs.
[03:54] <ScottK> RAOF: Edit ~/.kde/share/config/kwinrc and remove the line that says's it's crashed too many times.  You might pop onto #kubuntu-devel and ask mgraesslin.
[03:54] <RAOF> Yeah, that re-enables it, but doesn't stop it from setting that line again.
[05:20]  * ScottK wonders what happens if you chmod -w kwinrc?
[05:20] <ScottK> (not enough to try it though)
[05:36] <RAOF> Yeah, tried that :).  What happens is: kwin won't start until you've hit “ok” on the “I can't write to kwinrc!  Waaaaaa!” dialog box.
[05:36] <RAOF> Which could be worse, I guess :)
[14:36] <doctormo> I'm stuck using the nouvou driver since none of the proprietary nvidia drivers support the GeForce 8600M GT in maverick.
[14:36] <doctormo> The problem is that as well as loosing 3D, I'm stuck on 800x600 resoltuion.
[14:36] <doctormo> Any ideas on getting a better display configuration?
[14:44] <bjsnider> doctormo, there's a new nvidia driver release. i will be putting it in the x-updates ppa shortly. it may solve your problem
[14:44] <doctormo> bjsnider: Can you explain what went wrong?
[14:46] <bjsnider> well, the current 260 blob in maverick is a beta that didn't work with all hardware. beynd that, i don't work for nvidia
[14:47] <doctormo> Fair enough, so 260.19.12 should do the trick?
[14:47] <bjsnider> i don't know. i don't work for nvidia
[14:47] <bjsnider> it will just have to be tested
[14:49] <doctormo> bjsnider: It's a good thing you don't work for nvidia, their reputation with me is so low it's unplumbable.
[14:51] <ScottK> In comparison you have Intel that ships free drivers that in theory should work wonderfully.
[14:51] <ScottK> At least Intel keeps RAOF off the street and occupied.
[14:52] <bjsnider> are you saying there's something wrong with the way intel does things?
[14:52] <Sarvatt> doctormo: I'm going to guess you were hitting https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/656279 which made it seem like nvidia-current didn't work, did you see anything like that in your logs?
[14:52] <ubot4> Launchpad bug 656279 in nvidia-graphics-drivers (Ubuntu) "NVRM: os_raise_smp_barrier(), invalid context! (affects: 27) (heat: 144)" [Low,Confirmed]
[14:53] <Sarvatt> bjsnider: the complete lack of caring for stable driver releases is pretty bad to be honest
[14:53] <doctormo> Sarvatt: No os_raise_smp_barrier errors in my dmesg, just The NVIDIA GPU 0000:01:00.0 (PCI ID: 10de:0407) installed in this system is not supported by the 260.19.06 NVIDIA Linux graphics driver release.
[14:54] <bjsnider> Sarvatt, you mean from intel? or that i don't personally care about them?
[14:55] <bjsnider> i don't see anything in the changelog that says that bug is fixed with this blob
[14:56] <doctormo> bjsnider: Are you talking to me?
[14:56] <Sarvatt> intel, anything less than the latest git everything is unsupported pretty much, backporting fixes to the latest stable is a nightmare with all of the code refactoring
[14:57] <Sarvatt> doctormo: huh, thats odd, seems like its a really old problem too http://www.nvnews.net/vbulletin/showthread.php?t=127244
[14:57] <bjsnider> doctormo, well, it might not be fixed but nvidia doesn't always put all their changes in the changelogs
[14:58] <ScottK> bjsnider: No.  I'm saying I'm dissatisfied with the quality of the result.
[14:58] <tseliot> doctormo: that's not a real problem. I can reproduce it here and yet my system works without problems. Not that spamming dmesg is fine...
[14:59] <doctormo> tseliot: This machine goes into crisis and dumps into a terminal.
[14:59] <doctormo> At least this terminal is a visible terminal and not the weird flashing terminal when I first upgraded.
[15:00] <tseliot> doctormo: right. Do you have a dmesg and /var/log/Xorg.0.log I can have a look at?
[15:00] <bjsnider> Sarvatt, they fixed the customedid problem
[15:00] <tseliot> doctormo: after reproducing the problem, that is
[15:00] <doctormo> tseliot: I've reproduced it several times, reinstalled maverick 6 times, installed different drivers a few times. It's reproduced.
[15:02] <tseliot> doctormo: I don't doubt it. It's just that I need some data before I can discuss this with Nvidia. If you can get me those files and maybe file a bug report, I'll talk to them
[15:02] <bjsnider> doctormo, you reinstalled 6 times and the only real problem was the nvidia driver?
[15:02] <doctormo> bjsnider: So far, I haven't tried to suspend!
[15:07] <doctormo> tseliot: https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/661248 does that help?
[15:07] <ubot4> Launchpad bug 661248 in nvidia-graphics-drivers (Ubuntu) "All Nvidia drivers fail to load (affects: 1) (heat: 6)" [Undecided,New]
[15:08] <tseliot> doctormo: it says that your card is not supported by the driver: [   18.981092] NVRM: in this system is not supported by the 260.19.06 NVIDIA Linux
[15:08] <tseliot> which is weird
[15:09] <tseliot> doctormo: if you can also provide your /var/log/Xorg.0.log and your /var/log/Xorg.0.log.old the bug report would be complete
[15:09] <tseliot> but I guess the problem is clearer now
[15:11] <doctormo> tseliot: done
[15:12] <tseliot> doctormo: perfect, thanks. I'll subscribe them to the bug report and send them an email too
[15:12] <doctormo> thanks!
[15:15] <Sarvatt> it even says its supported - ftp://download.nvidia.com/XFree86/Linux-x86/260.19.12/README/supportedchips.html
[15:15] <Sarvatt> doctormo: has it ever worked with the blob?
[15:15] <doctormo> Sarvatt: It works in lucid, I have an external hdd with lucid and it works good.
[15:16] <bjsnider> i've heard this same thing a lot in the past week
[15:16] <tseliot> Sarvatt: this is why I want to contact them. It sounds like a regression to me
[15:16] <tseliot> and it's unlikely that they drop support for geforce 8xxx at this point
[15:18] <Sarvatt> I just saw that nvnews post with someone with the exact same device saying it didn't work every time in 180.xx some boots
[15:19] <Sarvatt> http://www.nvnews.net/vbulletin/showthread.php?t=127244
[15:21] <bjsnider> there have been a few people on irc with 8000 cards or chips that don't work
[15:24] <bjsnider> Sarvatt, you mean if someone finds a bug in the intel driver in lucid, and reports it to intel, they just say "go get the latest git snapshot of everything"?
[15:25] <Sarvatt> the fix it in a branch that will be the next unreleased kernel series
[15:26] <Sarvatt> eDP is the big problem right now and a lot of the fixes wont be in until 2.6.37
[15:29] <Sarvatt> workflow seems to be like,  get a bug report -> go through the code and see its a mess -> rewrite the subsystem -> fix the bug
[15:29] <Sarvatt> making backporting the fix to even the current rc's a huge deal
[15:31] <doctormo> Sarvatt: Shouldn't that be a fork? and is the new code better?
[15:32] <doctormo> branch*
[15:33] <bjsnider> they sound like overachieving perfectionists
[15:33] <ScottK> If that were true, you'd get to skip the "see it's a mess" stage.
[15:35] <bjsnider> but just because someone is a perfectionist doesn't mean they're writing perfect code
[15:36] <ScottK> True
[15:39] <doctormo> bjsnider: You mean to say that they want perfection, but don't actually know what that is.
[15:41] <Sarvatt> it's bad when I breathe a sigh of relief when an OEM uses ati or nvidia because the proprietary driver people actually care if things work on older releases, there is no way things like enabling arrandale support on jaunty would actually be considered for a SRU
[15:49] <doctormo> Sarvatt: Proprietary has it's own barrel of issues which I wouldn't wish on anyone. Can't we just have sensible development of all drivers in the open?
[15:53] <bjsnider> i think the backwards compatibility issue is one of the things luc verhaegen's been complaining about for awhile
[15:56] <libv> bjsnider: :)
[15:56] <bjsnider> doctormo, alright the new driver is built and published on i386, amd64 will be a few more minutes. it's in the x-updates ppa. it will be called nvidia-current by the system of course.
[15:56] <bjsnider> but i personaly think all regular users should be forced to use bleeding edge git snapshots and unreleased kernels
[15:59] <doctormo> bjsnider: Gun point or blackmail?
[16:07] <Sarvatt> doctormo: sorry, do you have a bug about the being stuck at 800x600 problem? also you can get 3D on nouveau by installing libgl1-mesa-dri-experimental
[16:07] <Sarvatt> doctormo: ah sorry I see the nouveau logs in your nvidia-graphics-drivers one, what the heck is going on there
[16:07] <Sarvatt> you're using vesa
[16:07] <Sarvatt> [   19.040082] [drm] nouveau 0000:01:00.0: Unsupported chipset 0xffffffff
[16:07] <Sarvatt> [   19.040216] nouveau 0000:01:00.0: PCI INT A disabled
[16:07] <Sarvatt> [   19.040223] nouveau: probe of 0000:01:00.0 failed with error -22
[16:08] <bjsnider> usupported chipset. that's basically what the blob said too
[16:08] <Sarvatt> yeah something really wacky going on with that machine
[16:12] <Sarvatt> lovely support site for it http://www.compal.com/asp/driver_dnd/e_download.asp?M_path=/Download/NB/JFL92_JFT02/BIOS
[16:32] <Sarvatt> [    20.164] (--) PCI:*(0:1:0:0) 10de:0407:0000:0000 rev 161, Mem @ 0xc6000000/16777216, 0xd0000000/268435456, 0xc4000000/33554432, I/O @ 0x00002000/128
[16:32] <Sarvatt> is it even valid to have no subsystem vendor/product id's?
[16:46] <doctormo> Sarvatt: Interesting catch with nouveau, so what do I have vesa?
[16:47] <Sarvatt> yeah thats why its only 800x600
[16:48] <tseliot> doctormo: they replied
[16:49] <tseliot> "That chipset number of 0xffffffff is a telltale sign that the GPU is not
[16:49] <tseliot> responding.  This is typically caused by kernel bugs, but could be
[16:49] <tseliot> caused by bad hardware, insufficient power supply, etc."
[16:49] <tseliot> that refers to this line from nouveau
[16:49] <tseliot> [ 19.040082] [drm] nouveau 0000:01:00.0: Unsupported chipset 0xffffffff
[16:51] <tseliot> doctormo: so my next question is, does the card currently work with nvidia driver 253 or with lucid's kernel?
[16:52] <doctormo> tseliot: It works perfectly with lucid and the nvidia current that came with that.
[16:53] <tseliot> doctormo: right. Does it work or did it use to work?
[16:53] <Sarvatt> can you attach a lspci -vvnn to the bug if you have it handy?
[16:53] <doctormo> It still works
[16:53] <tseliot> ok
[16:54] <tseliot> it might be a kernel issue
[16:54] <tseliot> rather than a hardware issue
[16:56] <doctormo> Sarvatt: Done
[16:57] <doctormo> tseliot: Interesting assessment, I wonder what would cause the error observed.
[16:57] <tseliot> doctormo: can you install lucid's kernel (image and headers) in maverick and see if it solves the problem, please?
[16:59] <doctormo> tseliot: package or not package?
[17:00] <tseliot> doctormo: you can use the packages from lucid if this is what you're asking
[17:00] <Sarvatt> 2.6.32-24 package + header packages + nvidia-current
[17:00] <Sarvatt> doctormo: did nouveau work on lucid for you?
[17:00] <doctormo> Sarvatt: Yes
[17:04] <doctormo> tseliot: The installation is impossible to complete, impossible dependency chain.
[17:04] <tseliot> doctormo: what's the exact error?
[17:04] <doctormo> Lots of linux-image needs linux-image
[17:04] <tseliot> I'm asking as I have a lucid kernel installed in maverick here
[17:06] <tseliot> doctormo: do not install the whole metapackage. Just install the image and headers, e.g. 
[17:06] <tseliot> linux-headers-2.6.32-25
[17:06] <tseliot> linux-headers-2.6.32-25-generic
[17:06] <doctormo> http://packages.ubuntu.com/lucid/linux-image
[17:06] <tseliot> linux-image-2.6.32-25-generic
[17:07] <tseliot> doctormo: you can get the packages that I've just recommended here: https://launchpad.net/ubuntu/+source/linux
[17:08] <tseliot> just 3 packages
[17:13] <doctormo> tseliot: I got lost fighting through a maze of links
[17:13] <tseliot> doctormo: I can give you the exact links if you tell me what architecture you're using. Is it 32bit or 64bit?
[17:13] <doctormo> 32bit
[17:19] <tseliot> doctormo: here you go:
[17:19] <tseliot> https://launchpad.net/ubuntu/+archive/primary/+files/linux-headers-2.6.32-25_2.6.32-25.44_all.deb
[17:19] <tseliot> https://launchpad.net/ubuntu/+archive/primary/+files/linux-headers-2.6.32-25-generic_2.6.32-25.44_i386.deb
[17:19] <tseliot> https://launchpad.net/ubuntu/+archive/primary/+files/linux-image-2.6.32-25-generic_2.6.32-25.44_i386.deb
[17:24] <doctormo> thanks!
[17:34] <doctormo> This is really weird trying to edit images in 800x600 on a widescreen. (doing this before rebooting)
[17:37] <tseliot> I bet it is ;)
[18:17] <bjsnider> he never mentioned if it solved the problem or not
[18:18] <bjsnider> doctormo, what happened?
[18:19] <doctormo> bjsnider: I got the bubbles redone, decided to replace most of them, but got fustrated with inkscape 0.48 removing features on me.
[18:25] <bjsnider> no, i was referring to booting from the old kernel
[18:36] <Sarvatt> doctormo: under lucid it was different
[18:36] <Sarvatt> 01:00.0 VGA compatible controller [0300]: nVidia Corporation G84 [GeForce 8600M GT] [10de:0407] (rev a1)
[18:36] <Sarvatt> 	Subsystem: COMPAL Electronics Inc Device [14c0:0025]
[18:36] <Sarvatt> just looked up one of your old bug reports
[18:39] <doctormo> Sarvatt: Interesting
[18:40] <bjsnider> Sarvatt, if that's the kernel's fault, why does it affect everybody? why just some systems?
[18:41] <bjsnider> i meant why doesn't it affect everybody
[18:41] <doctormo> Sarvatt: I'm trying to spot the difference, what is different?
[18:42] <Sarvatt> i still have no idea what the actual problem is, you just have no pci subsystem id's being reported anymore for it on 2.6.35
[18:43] <Sarvatt> its 10de:0407 0000:0000 now and was 10de:0407 14c0:0025 before
[18:43] <doctormo> Sarvatt: Perhaps this is a system76 issue then, that's what the computer is.
[18:44] <Sarvatt> it's actual a compal JFL92 that they put stickers on and resell :)
[18:46] <tseliot> doctormo: did you test the packages that I recommended?
[18:47]  * tseliot -> dinner
[18:49] <doctormo> About to reboot now.... Enjoy my editing while I do that http://divajutta.com/doctormo/ubunchu/ubunchu-07-ltr.pdf
[18:49] <bjsnider> Sarvatt, what log are you looking at for the pciid's?
[18:50] <Sarvatt> lspci from https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/578993 and the lspci/xorg.0.log's from https://bugs.edge.launchpad.net/ubuntu/+source/nvidia-graphics-drivers/+bug/661248
[18:50] <ubot4> Launchpad bug 578993 in linux (Ubuntu) "Doesn't suspend on System76 (affects: 1) (heat: 25)" [Undecided,Expired]
[19:05] <doctormo> bjsnider and tseliot: We're back, the compal stuff is back to normal and the nvidia driver is loaded correctly. I have full resolution.
[19:06] <doctormo> 3D confirmed working.
[19:08] <bjsnider> nothing else on the system changed except the kernel, so this must be a kernel issue
[19:12] <bjsnider> so i think a new bug should be created against the maverick kernel to alert the kernel team so they can investigate. you should also link the bug you're currently following to the new one
[19:17] <doctormo> bjsnider: Trying to report bug, but can't find the drop off point.
[19:17] <bjsnider> Sarvatt, where's the nvidia-graphcis-drivers patch you use for maverick?
[19:18] <bjsnider> doctormo, i think if it's the kernel you just call it "linux"
[19:25] <doctormo> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/661394
[19:25] <ubot4> Launchpad bug 661394 in linux (Ubuntu) "Compal/nvidia breakage in system76 laptops (affects: 1) (heat: 6)" [Undecided,New]
[20:50] <tseliot> doctormo: ah, so it's bug 661394, this is interesting
[20:50] <ubot4> Launchpad bug 661394 in linux (Ubuntu) "Compal/nvidia breakage in system76 laptops (affects: 2) (heat: 12)" [Medium,Triaged] https://launchpad.net/bugs/661394
[20:57] <bjsnider> i'm not convinced that this is really limited to system76 laptops
[21:08] <tseliot> definitely a kernel issue though
[21:09] <bjsnider> unquestionably but everybody w ho has the problem will blame the graphics drivers for it
[21:12] <tseliot> :/
[21:23] <Sarvatt> RAOF: regarding /usr/bin/X (FreeClientResources+0xed) [0x808f04d], http://cgit.freedesktop.org/xorg/xserver/commit/?h=server-1.9-branch&id=3b687814a5b999305042775a2bddc3c10706f6da ?
[21:25] <ScottK> Is that possibly related to 660152?
[21:25] <Sarvatt> yeah
[21:26] <ScottK> Cool.
[22:13] <Sarvatt> bjsnider: the acer aspire 5920G the other guy in the original bug has is compal with the same 8600M GT MXM2 card in it, its basically the same machine http://launchpadlibrarian.net/41765266/lspci-vnvn.log
[22:14] <bjsnider> yeah but that's just the last few hours
[22:15] <bjsnider> i'm talking about the people who've been in the maverick channel the past week or two and nvidia just won't work at all
[22:15] <bjsnider> i've got to put the lucid blob in the ppa, where's the patch for the build scripts?
[22:16] <Sarvatt> it might be mxm specific, there are quite a few laptops with mxm gpu's
[22:16] <Sarvatt> "We know Compal did some pretty funky things with the original 8600GT that came with this machine (all to save a few cents on a SPI flash) "
[22:16] <Sarvatt> thats great to hear
[22:17] <bjsnider> i guess that's an "edge case"?
[22:19] <bjsnider> why is system76 selling this junk
[22:30] <Sarvatt> http://www.mxm-upgrade.com/IFL90.html kind of explains it :)
[22:36] <Sarvatt> http://www.system76.com/product_info.php?cPath=28&products_id=99 = clevo W860CU, at least system76 doesn't charge *that* much more than buying the actual thing unlike most places