[04:00] <mlankhorst> I may or may not have woken up from a dream about the lockdep self-test functions I was writing for the reservation code..
[04:16] <tjaalton> mlankhorst: how do we tell? :)
[04:16] <tjaalton> btw, nice proposal from aritger about a new opengl abi, and how it should fix the alternatives mess we have now
[04:27] <mlankhorst> tjaalton: well usually I wake up at a certain time, if I wake up earlier it may indicate something like that :-)
[04:47] <Sarvatt> mlankhorst: thats a crappy dream, a good one would have danvet looking at it and fixing it for you :P
[04:50] <Sarvatt> tjaalton: yeah that was a great read, will solve so many headaches
[04:51] <Sarvatt> also reinforced /usr/lib/libGL.so.1 being ABI we're not conforming to..
[04:56] <RAOF> Where's this discussion?
[04:56] <Sarvatt> mesa-dev
[04:56] <RAOF> (Also, yeah, we should really have libGL.so.1 available) ☺
[04:57] <Sarvatt> mid.gmane.org/20120912210932.GI3318@parker.nvidia.com
[05:01] <mlankhorst> So nvidia finally acknowledges drivers that aren't nvidia's own exists? :D
[05:04] <mlankhorst> (like i915, radeon)
[05:04] <RAOF> Looks very interesting.
[05:22] <tjaalton> oh right, we don't have .so.1 in any direct path..
[05:28] <mlankhorst> why do we use ld.so.conf hacks though instead of debian alternatives? is it because that one fails on multiarch?
[05:38] <tjaalton> we use alternatives, the ld.so.conf files
[05:43] <LLStarks> bug 1050202 looks so sexy.
[05:43] <ubottu> Launchpad bug 1050202 in x11-xserver-utils (Ubuntu) "Cherrypick xrandr list providers to complete DRI2 offloading graphics stack" [Undecided,New] https://launchpad.net/bugs/1050202
[05:50] <Sarvatt> mlankhorst: feel like doing an xorg/app/xrandr release? :P
[05:54] <LLStarks> Sarvatt, will the ddx's make the beta 2 freeze?
[05:54] <Sarvatt> already did
[05:54] <LLStarks> oh cool
[05:54] <Sarvatt> tjaalton uploaded them yesterday
[05:55] <JohnDoe_71Rus> Hi try build ./auto-xorg-git -p xserver-xorg-video-ati from http://bazaar.launchpad.net/~xorg-edgers/xorg-server/xorg-pkg-tools/view/head:/README.auto-xorg-git. get error http://pastebin.com/HFDFM13j What i missed ?
[05:55] <Sarvatt> JohnDoe_71Rus: sudo apt-get build-dep xserver-xorg-video-ati?
[05:58] <JohnDoe_71Rus> Sarvatt: Thx. But the same output. ubuntu server 10.04
[05:59] <Sarvatt> you need lucid-backports enabled and manually installing dh-autoreconf in that case
[05:59] <Sarvatt> its going to be a pain, not sure it even works with lucid's xserver, if it does you'll need libdrm updated for sure
[06:01] <ajmitch> would an nvidia driver prerelease like 304.48 be uploaded to quantal?
[06:07] <JohnDoe_71Rus> Sarvatt: Thx.  http://pastebin.com/UXWAGhK3 but enother err
[06:07] <JohnDoe_71Rus> Depending on what needed to build?
[06:10] <tjaalton> meh, is there a gcal to subscibe to in order to get the release schedule on my calendar?-)
[06:11] <tjaalton> oh, still a week to go until b2 freeze
[06:12] <LLStarks> http://fridge.ubuntu.com/calendars/ubuntu-release-calendar/
=== LLStarks is now known as erappleman
[06:13] <tjaalton> erappleman: cool, thanks :)
[06:20] <tjaalton> ok, testing the gpu-autobind patch..
[06:21] <Sarvatt> wish i got a t430s instead of an x1 carbon so i could mess with it too
[06:22] <tjaalton> haha
[06:22] <tjaalton> so you got it already?
[06:23] <Sarvatt> tomorrow, 1 month later, many delays
[06:23] <tjaalton> return it
[06:23] <erappleman> >Now at patch 602_autoconfig.diff
[06:23] <Sarvatt> i called to return it and they shipped mine while i was on the phone :P 15% restocking fee to return it
[06:24] <erappleman> let's see if this works with unity, xfce, shell, whatever
[06:25] <erappleman> building xorg server is a pita, i have to sideload ltmain.sh during autogen
[06:25] <erappleman> since it gets removed
[06:26] <tjaalton> i've had no problems building it
[06:28] <erappleman> i'm building from apt-source using dpkg-buildpackage -rfakeroot -b
[06:28] <tjaalton> build from the git branch
[06:29] <erappleman> ubuntu x git?
[06:29] <tjaalton> yes
[06:29] <tjaalton> well, debian pkg-xorg
[06:29] <tjaalton> ubuntu branch
[06:38] <Sarvatt> JohnDoe_71Rus: so you're getting stuck because of xsf dh helper not being available which was added after lucid, try taking the debian/ directory from x-x-v-ati in the lucid package, then dropping it in a newer checkout and disabling all the patches in debian/patches/series, you're still going to need a libdrm update that will break the package manager though without some accounting for the old libdrm-nouveau1 :(
[06:42] <RAOF> Hm. Should try that on the ati hybrid lappy
[06:43] <JohnDoe_71Rus> Sarvatt: so hard ... I want to build a new driver, because after update cairo-dock does not start in gl mode
[06:43] <tjaalton> RAOF: need a newer ddx from git. wonder when they'll do a new release..
[06:44] <RAOF> Heh
[06:44] <JohnDoe_71Rus> Radeon HD 46xx or HD47xx series I do not remember
[06:45] <tjaalton> damn quantal, mismatch of the local dns cache address in resolv.conf & dnsmasqd
[06:45] <Sarvatt> JohnDoe_71Rus: yeah I dont recommend it at all, if anything updating to 12.04 is really worth it
[06:45] <tjaalton> (127.0.0.1 vs. 127.0.1.1)
[06:46] <JohnDoe_71Rus> I tried to build from git. and run cairo with the builded files it worked. But I want to get a deb
[06:49] <JohnDoe_71Rus> yet have a problem with wifi AP at 12.04. can not update
[06:58] <erappleman> raof, really curious how deucher and airlie aren't releasing a new stable -ati. intel and nouveau beat em to the punch
[06:59] <RAOF> -ati goes ages without a release.
[06:59] <RAOF> So we end up pretty much ignoring their releases and just taking what looks good from git.
[07:00] <tjaalton> so it's probably time for that again :)
[07:00] <RAOF> Indeed :)
[07:00] <Sarvatt> erappleman: easy, ship git in fedora, http://pkgs.fedoraproject.org/cgit/xorg-x11-drv-ati.git/log/?h=f18
[07:00] <mlankhorst> oh modesetting 0.5.0 is released
[07:01] <tjaalton> orly? :)
[07:01] <erappleman> Sarvatt, fedora is upstream for -ati? no sid/experimental?
[07:04] <mlankhorst> yeah asked airlied yesterday if I could do a release for prime, so finally did it
[07:05] <Sarvatt> mailserver is lagging
[07:06] <Sarvatt> no tag on xorg-commit, no announce :P
[07:06] <mlankhorst> asdf
[07:07] <mlankhorst> stupid git-push, it pushed something now, not sure what
[07:07] <tjaalton> nothing happened when I plugged in the dp cable :(
[07:07] <tjaalton> where's my xrandr 1.4
[07:07] <Sarvatt> i saw your push but cant find and record of it, sure its just mail lagging
[07:08] <mlankhorst> woops, I accidentally pushed the debian tag to there somehow
[07:09] <tjaalton> the log also has "(EE) NOUVEAU(G0): Error creating GPU channel: -19"
[07:09] <mlankhorst> tjaalton: yeah gpu acceleration is disabled then
[07:09] <Sarvatt> git tag -d xserver-xorg-video-modesetting 0.3.0-1! :P
[07:10] <tjaalton> oh pushed a debian tag there :)
[07:11] <mlankhorst> Sarvatt: remotely?
[07:13] <Sarvatt> oops forgot a - there before 0.3.0-1, then push after :P
[07:13] <RAOF> tjaalton: What gpu are you on - is that the T420s with the nvd9 that doesn't quite work properly?
[07:13] <tjaalton> RAOF: yes
[07:13] <tjaalton> how is it not working?
[07:13] <tjaalton> I've not used it much
[07:13] <tjaalton> the nvidia one
[07:14] <Sarvatt> tjaalton: dont have discrete turned on in the bios maybe?
[07:14] <RAOF> Last time I tried - which was some time ago - nouveau didn't have the ctxprogs to do accel. And my understanding from lurking in #nouveau is that this is still the case. Plus some interesting extra bits.
[07:14] <tjaalton> Sarvatt: nouveau gets loaded, optimus turned on
[07:14] <mlankhorst> tjaalton: because on some cards it's buggy
[07:14] <tjaalton> ahh right
[07:14] <mlankhorst> s/some/most/
[07:14] <tjaalton> well I'd be happy to get something on the monitor :)
[07:15] <tjaalton> but don't have xrandr 1.4 to kick it manually
[07:15] <RAOF> tjaalton: ...also, before the laptop died, I found that nouveau didn't know how to drive the DP port on the nvd9  :)
[07:15] <tjaalton> so in my case the autobind patch isn't that useful?
[07:15] <tjaalton> RAOF: ha, ok that would explain things
[07:15] <erappleman> prime is still buggy. i have pretty much every new patch from dri-devel and xorg-devel, but everything still renders white if it renders at all.
[07:16] <RAOF> But I did submit an mmiotrace of nvidia setting up a DP connection, and darktama was working on it, so that might actually work now.
[07:16] <Sarvatt> too bad that laptop died :(
[07:16] <RAOF> Also, DP→{HDMI,DVI} worked; it was just native DP that didn't.
[07:16] <RAOF> Yeah, that laptop was the bees knees.
[07:17] <mlankhorst> RAOF: btw you can still work on radeon if you want to, that part is still not done :D
[07:17]  * RAOF may be a bit busy to hack on radeon :)
[07:17] <tjaalton> oh well, attaching the dvi with a dongle crashed the xserver
[07:17] <RAOF> mlankhorst: It's the cross-gpu synchronisation that's not done yet, right? You should be able to do DRI2 offload on radeon?
[07:18] <RAOF> tjaalton: Progress!
[07:18] <mlankhorst> RAOF: yeah, I was working on some lockdep stuff
[07:18] <tjaalton> :)
[07:18] <mlankhorst> lockdep should complain if you mess up your reservations
[07:19] <Sarvatt> if you have the patch mlankhorst dreamt of last night
[07:19] <mlankhorst> hey I had most of it working right before I went to bed
[07:20] <tjaalton> unplugging crashed it as well
[07:20] <RAOF> Both sides of hotpulg work!
[07:20] <mlankhorst> anyone feels like uploading modesetting 0.5.0?
[07:21] <tjaalton> sure
[07:21] <mlankhorst> that should just leave -ati only being prime incapable then :-)
=== JanC is now known as Guest57134
[07:27] <tjaalton> uploaded
[07:42] <tjaalton> get a nice bootsplash image on the display hooked to nouveau
[07:43] <erappleman> wish i could do that
[07:43] <erappleman> hdmi is connected to intel
[08:15] <mlankhorst> getting there, slowly
[08:16] <tjaalton> mlankhorst: wanna release xrandr 1.4?-)
[08:16] <RAOF> What's needed over an existing quantal install for playing around with prime? New -ati, new -intel, ??? ...?
[08:16] <tjaalton> btw, the dp-dvi dongle gets super hot..
[08:16] <tjaalton> just new -ati
[08:16] <RAOF> That's not an active dp-dvi, is it?
[08:17] <tjaalton> no
[08:17] <tjaalton> testing 3.6rc to see if dp works
[08:22] <mlankhorst> tjaalton: I think airlied should
[08:23] <tjaalton> ok
[08:41] <tjaalton> i'll build it from git to see how it works
[08:45] <tjaalton> we can also dump the current git to x11-xserver-utils/xrandr if we want
[08:52] <tjaalton> sigh, forgot again how to disable multiarch on quantal
[08:52] <tjaalton> removing /etc/dpkg/dpkg.conf.d/multiarch isn't enough
[08:53] <tjaalton> dunno how it keeps coming back on the chroot
[08:56] <tjaalton> powertop isn't too happy on quantal, calculates the power consumption wrong
[08:56] <tjaalton> i doubt it's taking 1,6W with nouveau loaded :)
[09:01] <tjaalton> hmm, might be decimal fail
[09:08] <mlankhorst> found a case lockdep doesn't handle, maybe I throw too much at it :-)
[09:16] <mlankhorst> so awesome though :D
[09:25] <tjaalton> hmm, semi-success with dp on 3.6rc
[09:26] <tjaalton> [   66.509680] [drm] nouveau 0000:01:00.0: DDC responded, but no EDID for DP-1
[10:49] <RAOF> Ah. hybrid radeon lappy doesn't prime, because radeon's not hooked up to any outputs so it gets unloaded.
[10:49] <RAOF> Oh, well.
[10:52] <tjaalton> ah.. 
[11:02] <mlankhorst> enough lockdep for now
=== broder_ is now known as broder
[12:07] <tjaalton> great, intel refuses to recognize my monitor now
[12:08] <mlankhorst> 'what monitor' ;)
[12:09] <tjaalton> indeed..
[12:11] <tjaalton> ha, monitor fail
[12:12] <tjaalton> nouveau got it so confused that I had to pull the power plug too
[12:13] <tjaalton> and compiz fail, every window was maximized when I got it back
[12:57] <tjaalton> mlankhorst: bugs? 
[13:06] <mlankhorst> I suppose so
[13:26] <tjaalton> bzr fun.. "different rich-root support"
[13:26] <tjaalton> so I won't push it then
[13:28] <mlankhorst> bzr kind of makes me want to track down that git client for it again
[13:33] <tjaalton> eh, another try and it succeeded
[13:46] <mlankhorst> tjaalton: does the reference patch for autobinding work?
[13:48] <tjaalton> mlankhorst: not sure, what should it do?-)
[13:48] <tjaalton> besides, no acceleration here so..
[13:48] <mlankhorst> oh thought you were testing with it
[13:49] <mlankhorst> just binds everything to the first display
[13:49] <tjaalton> I have it running
[13:49] <mlankhorst> without the need to do it manually
[13:53] <tjaalton> ok
[13:56] <tjaalton> trying to figure out how to driver the external screen
[13:56] <tjaalton> -r
=== wgrant_ is now known as wgrant
=== JanC_ is now known as JanC
[14:38] <tjaalton> anyone with an intel 915/945 willing to bisect mesa?
[14:58] <tjaalton> guess I just need to build seven sets of packages then :)
=== Azelphur_ is now known as Azelphur
[17:17] <mlankhorst> hehe :)
[17:18] <tjaalton> hum?
[17:18]  * mlankhorst imagines someone bisecting through a launchpad ppa
[17:18] <tjaalton> ah
[17:19] <tjaalton> how the heck can "c8a86f717f8e302 mesa: move IS_NEGATIVE() and DIFFERENT_SIGNS() to macros.h" break i915?
[17:19] <bryceh> mlankhorst, heh - https://wiki.ubuntu.com/X/Blueprints/FriendlyGitBisection
[17:20] <mlankhorst> tjaalton: USE_IEEE not being defined or something?
[17:20] <tjaalton> mlankhorst: maybe.. main/macros.h is included anyway
[17:20] <tjaalton> hmm
[17:21] <tjaalton> intel_mipmap_tree.c: In function 'intel_miptree_alloc_mcs':
[17:21] <tjaalton> intel_mipmap_tree.c:796:37: warning: 'format' may be used uninitialized in this function [-Wmaybe
[17:22] <tjaalton> -uninitialized]
[17:22] <tjaalton> nah
[17:22] <tjaalton> less guessing, more building
[17:22] <mlankhorst> hehe :)
[17:23]  * mlankhorst is thinking of using RCU to remove a set of locks
=== ajmitch_ is now known as ajmitch
[20:29] <tjaalton> nice, can update mesa again.. maybe not tomorrow though