[08:29] <richardcavell> Hi all. I'm running Ubuntu 9.04 on a MacBook with Intel GMA950 chipset.  The drivers are pretty buggy.  I hope to fix them by upgrading from the stable release to something more edgy.  I'm thinking kernel 2.6.30 with KMS, and later builds of the intel linux drivers. Can anyone nominate an easy way to do this?
[08:30] <tjaalton> yeah, upgrade to karmic
[08:30] <tjaalton> or use the xorg-updates PPA
[08:31] <crevette> richardcavell, but karmic is still in development so it going to be unstable
[08:31] <richardcavell> crevette: well my present video system is unstable...
[08:32] <richardcavell> tjaalton: If I simply upgrade to Karmic will that give me kernel mode setting, and the latest intel drivers?  
[08:32] <tjaalton> yes
[08:32] <tjaalton> other things might get broken though, like crevette pointed out
[08:32] <richardcavell> The latest intel drivers (from intellinuxdrivers.org) are xf86-video-intel-2.8.0
[08:33] <richardcavell> I tried downloading the source and ./configure but it gives me an error and I don't know enough to fix it
[08:34] <richardcavell> checking for XORG... configure: error: Package requirements (xorg-server >= 1.6 xproto fontsproto ) were not met:
[08:34] <richardcavell> No package 'xorg-server' found
[08:34] <richardcavell> No package 'fontsproto' found
[08:34] <tjaalton> even if you got it working, it wouldn't be enough
[08:34] <tjaalton> since you need the new kernel
[08:35] <richardcavell> tjaalton: I can download binaries for the new kernel
[08:35] <tjaalton> so build the deb from karmic then?
[08:35] <richardcavell> tjaalton: but I can't find binaries for Ubuntu for the latest intel drivers.  2.6.3 is the latest stable version
[08:36] <richardcavell> looking at the changelogs for the intel drivers, it looks like heaps of bugs have been fixed between 2.6.3 and 2.8.0
[08:36] <tjaalton> yes, but in order to get 2.8.0 working you really need 2.6.31rc
[08:37] <tjaalton> .30 might work
[08:41] <richardcavell> I'll upgrade if it will fix my problems, but I'm not sure if it will
[08:41] <tjaalton> you could also try the xorg-edgers PPA and only pull -intel and it's deps
[08:41] <tjaalton> and the new kernel
[08:45] <richardcavell> tjaalton: that's what I'd prefer
[08:46] <tjaalton> and after installing what you need, disable it
[08:46] <tjaalton> so that the other crack won't be installed along with other updates
[08:47] <hyperair> has anyone gotten the rc4 kernel to not panic with intel + modesetting?
[08:47] <tjaalton> haven't tried it yet
[08:57] <richardcavell> Okay, I want to upgrade to a more recent set of drivers for my Intel GMA950 on Jaunty.  Here's what I'm planning to do: Install linux kernel 2.6.30-10 and everything in the xorg-edgers PPA.  Does this sound good?
[09:12] <tjaalton> don't install everything, just the driver and kernel
[09:12] <tjaalton> gone ->
[09:35] <RAOF> Look out ninjas.  Preliminary suspend/resume support for g80 (exactly) lands in xorg-edgers/nouveau.
[12:17] <Guest6103> Hi, guys. I have Jaunty running on a Macbook second-generation, which has an Intel GMA 950.  I was getting poor performance with the standard issue Jaunty drivers.  So I installed everything from the xorg-edgers PPA and now X won't start at all!  Any ideas?
[12:21] <Guest6103> Okay, can I ask this: I can't get X11 to work, but I can get to a root command line.  What's the metapackage that I need to force install via apt-get to give me the standard Jaunty stable xorg?
[14:33] <Sarvatt> well heck, I go out of my way to make sure edgers always works and am conservative about updates in it. if people see it that way i might as well just throw xorg 7.5 in there and really break things :)
[14:43] <Sarvatt> telling people to just install intel and its dependencies is kinda a bad idea, its not like the packaging scripts get all the deps right since its based on the older releases.. and back when xserver 1.6.1.902 was in there it wouldnt even work with the jaunty xserver because of how the dri2 abi was broken.. thats what x-updates is for
[14:47] <Sarvatt> guess i need to work on the hooks more, the xserver in there requires mesa 7.5 at least because it compiles in setTexBuffer2 support because its available at build time
[15:50] <Le-Chuck_ITA> Hi all, can someone take a look at bug #402982? Bryce do you have an intel card? Do you have the same bug?
[16:02] <Sarvatt> revert back to the 2.6.30-2 kernel
[16:02] <Sarvatt> err 2.6.31-2
[16:02] <Sarvatt> hopefully the fix makes it into 2.6.31-5, it didnt make it into 2.6.31-4
[16:05] <Sarvatt> lots of people are having that problem though
[16:06] <Sarvatt> started with 2.6.31-rc3 because of this http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=7662c8bd6545c12ac7b2b39e4554c3ba34789c50
[17:57] <jbarnes> tseliot: ping
[17:57] <tseliot> jbarnes: pong
[17:57] <jbarnes> tseliot: hi!
[17:57] <jbarnes> so
[17:57] <tseliot> hi :-)
[17:57] <jbarnes> I just upgraded my dell mini 12 to jaunty yesterday
[17:58] <jbarnes> and found (after some digging) the ubuntu mobile repo with the latest poulsbo drivers
[17:58] <jbarnes> they seem to work a bit, but they hang a lot and suspend/resume takes forever
[17:58] <jbarnes> are there any updates in the works or something I can subscribe to for status updates?
[17:59] <tseliot> yes, that's not what I worked on
[17:59] <tseliot> I don't think I can share the OEM repositories though
[17:59] <jbarnes> oh so the mobile repo has old stuff?
[17:59] <tseliot> I'm not sure, as I don't maintain that
[18:00] <tseliot> I can ask my boss to see if I can share the repo
[18:00] <jbarnes> thanks
[18:01] <jbarnes> or you could point me at someone else... I'd like to get this machine to be a little more stable
[18:01] <jbarnes> I can also ping some people internally
[18:02] <tseliot> jbarnes: do you know Michael Frey?
[18:02] <jbarnes> no I don't think so
[18:03] <tseliot> jbarnes: ok, let me ask around in our private oem chat room
[18:04] <jbarnes> thanks, I poked some UMG folks too
[18:04] <jbarnes> maybe I can get access to the bits we end up sending you guys :)
[18:06] <tseliot> jbarnes: I did some work so that it works with DKMS, add diversions, etc.
[18:06] <jbarnes> cool
[18:06] <jbarnes> so if you're ever able to include it in the jaunty/karmic repos it should be easy right? :)
[18:07] <tseliot> jbarnes: I don't think I can do that. Maybe I can upload them to a PPA. I'm discussing this with my team manager
[18:08] <jbarnes> oh right, I'm not sure how you handle the binary only stuff
[18:10] <tseliot> jbarnes: ok, I can set up a PPA
[18:11] <jbarnes> yay
[18:11]  * jbarnes hopes the bits are better than the ones he already has
[18:15] <tseliot> jbarnes: ok, I'll upload the packages here: https://launchpad.net/~albertomilone/+archive/poulsbo-graphics
[18:15] <jbarnes> cool
[18:15] <tseliot> it will take a while though
[18:15] <jbarnes> np I probably won't get a chance to update until tonight
[18:18] <bryce> well I've gotten stuck on my upload tool
[18:18] <jbarnes> tseliot: thanks a lot
[18:19] <bryce> launchpad doesn't give enough info about attachments unfortunately
[18:19] <jbarnes> all the poor suffering poulsbo users will be a bit happier now
[18:19] <tseliot> jbarnes: np ;)
[18:19] <jbarnes> tseliot: fwiw there are a few forum threads etc where you could announce the ppa
[18:19] <jbarnes> seems quite a few people are looking for it
[18:20] <tseliot> jbarnes: I don't want to make too much noise about it
[18:21] <jbarnes> yeah I suppose attracting attention to poulsbo is a bad thing generally :)
[18:21] <tseliot> hehe
[18:22] <ScottK> tseliot: Are you going to build poulsbo for Karmic?
[18:23] <tseliot> ScottK: I'm not even sure as to whether it builds with the kernels in Karmic
[18:23] <ScottK> OK.
[18:24] <tseliot> ScottK: BTW did my synaptics package work well for you?
[18:25] <ScottK> tseliot: I still didn't get my 10v back to normal after Alpha 3 testing, so I didn't try it yet.
[18:25] <ScottK> $WORK has been very busy this week.
[18:25] <tseliot> ok, np
[18:27] <ScottK> I saw you got some very good feedback on it though.
[18:28] <tseliot> right. Thanks apw :-)
[18:29] <apw> tseliot, heh :)  my 10v was a pig to use without that one package with it its not bad
[18:29] <tseliot> :-)
[18:29] <apw> i have large finger tips and i could do with the buttons making bigger
[18:29] <apw> or whatever its doing to stop it going  bonkers
[18:30] <apw> anyhow ... win over the previous option which was poking ones eyes out everytime you lifted the wrong finger first
[18:30] <tseliot> apw: have you tried the xinput property?
[18:30] <tseliot> "Synaptics Area"
[18:30]  * apw has only added the package
[18:30] <apw> there is more magic?
[18:31] <tseliot> apw: that will allow you to create an area outside of which movements, taps, etc. are ignored
[18:31] <apw> oh nice so i can deaden like the whole bottom or something
[18:31] <tseliot> apw: yes, exactly
[18:31] <apw> is that just in the driver, or something the pad does?
[18:32] <apw> as it would be nice to be able to say the bottom left triangle
[18:32] <tseliot> apw: I implemented that in the X driver (and it's in upstream now)
[18:33] <apw> so we could make it any shape in principle
[18:33] <tseliot> triangle?
[18:33] <tseliot> it works with rectangles right now
[18:33] <apw> the area lets me define a rectagle in the middle which is active basically yes?
[18:34] <apw> for the particular case of the 10v it would be nice to define two square which do not work out of the corners of the pad
[18:34] <tseliot> yes
[18:34] <apw> rather then lose all the bottom or all the sides or whatever
[18:34] <apw> anyhow thats all academic, it sounds like in principle we could do that if we wanted
[18:34] <apw> as its done in sw we have control over :)
[18:35] <tseliot> it's doable but it took me a while to get my patch accepted ;)
[18:35] <apw> heh yeah ...
[18:35] <tseliot> for now you can try with xinput set-int-prop $YOUR_DEVICE_ID "Synaptics Area" 32 0 0 0 4000
[18:35] <apw> the main bit i am benfiting from then is just the 'you couldn't get there that fast, i can't hear you' fix
[18:36] <apw> how does one find out ones DEVICE_ID
[18:36] <tseliot> note: 0 means that limit = physical edge
[18:36]  * apw strays out of his nice comfortable kernel cave
[18:36] <tseliot> xinput list
[18:36] <tseliot> and see where it says id
[18:37] <apw> what the heck is a macintosh button emulation when its at home
[18:37] <apw> tseliot, so why 5 numbers?
[18:37] <tseliot> 32 bit
[18:38] <apw> ahh
[18:38] <apw> stupid interfaces :)
[18:38] <tseliot> then you have left, right, top, bottom. It's all documented in the man page of synaptics
[18:38] <apw> so is that 4000, 4000 units from the bottom up or 4000 from the top down
[18:39] <tseliot> y is inverted, so top < bottom
[18:39] <apw> so to make the bottom band smaller i'd make it > 4000
[18:40] <tseliot> apw: right
[18:40] <tseliot> jbarnes: do things improve if you put this in a file in /etc/modprobe.d/ :
[18:40] <tseliot> options psb disable_vsync=1 no_fb=1
[18:41] <tseliot> that should help
[18:41] <jbarnes> ok cool I'll try that
[18:42] <tseliot> jbarnes: that should prevent it from hanging or make it hang less frequently at least
[19:54] <Sarvatt> might be a little easier to use like synclient AreaBottomEdge 4000 instead of figuring out the id
[21:46] <Sarvatt> oops, I meant synclient AreaButtomEdge=4000 earlier
[22:12]  * hyperair feels stupid
[22:12] <hyperair> turns out the panic was because i forgot to add an initrd
[22:25]  * hyperair watches ccache hits rise and kernel modules go whizzing by