=== Sarvatt_ is now known as Sarvatt [06:33] hi [06:35] My laptop on natty is making fun of me. The desktop freeze very too often, sometimes I can go to a tty and restart gdm, sometines I only can use the power button. Can you guys give me some advices or help me report or find help on this bug ? [06:43] YoBoY: install the kernel from -proposed [06:44] hum... normally I'm already on proposed, I check [06:44] should be abi version 9 [06:46] I'm on 2.6.38-9-generic [06:47] ok then [06:49] (i'm on the 64bit edition btw) [06:50] doesn't matter [06:54] tjaalton: You wanted to test llvm support on your r300-r500 chips? [07:16] RAOF: no, sis :P [07:17] i have some firegl card, but don't know which generation it is [07:19] tjaalton: Oooh, the ultimate in software-fallbacks! [07:20] Anyway, it's in git. Or, if you prefer, I'm sure I've got some prebuilt packages I could throw at you :) [07:20] YoBoY: looks like that kernel doesn't have the fix I was after.. [07:20] RAOF: yeah, 32bit ones would be nice [07:21] RAOF: you wrote the patch for i915 that fixes a sh*tload of these natty crashers? [07:21] tjaalton: By “crashes” you meen freezes? Yes, and I finally aligned all the knobs to get it into stable-queue. [07:22] RAOF: yeah that, cool [07:24] tjaalton: the fix will be available soon ? [07:24] YoBoY: well it's not even in -proposed yet [07:24] as it seems [07:25] It's in the 2.6.39 kernel, but not yet in any natty kernel. [07:25] right [07:35] there is a bug I can subscribe to follow the progress ? [07:36] yes [07:36] but I don't have it handy :P [07:37] there's also a ppa with a test kernel [07:37] bug #740126 [07:37] Launchpad bug 740126 in unity (Ubuntu Natty) (and 6 other projects) "Disabling an output can cause vblank events to be missed (affects: 76) (dups: 10) (heat: 380)" [High,Triaged] https://launchpad.net/bugs/740126 [07:46] I'm having trouble in natty where after coming out of a fullscreen window, compiz is completely b0rked. There are major issues flickering and it bounces between front and back buffers that are remnants of previous frames [07:47] on rv350 with xorg-edgers installed [07:48] restarting compiz or even X does not help, it's like the driver is stuck in b0rken mode [07:49] I'm thinking it's a kernel/radeon module bug or mesa but not sure which [07:49] soreau: It's not inconceivable that it's an X bug, either. [07:50] tjaalton: Hm. It seems I don't have an i386 build handy; I'll build one now. [07:50] RAOF: Right but how do I know where to file a report? [07:50] RAOF: got beefy hw for it? I have a dualcore here, so could build it myself too [07:51] tjaalton: No beefy hardware here; just a dualcore. [07:51] ok, so maybe I'll just pull from git and build it here [07:51] soreau: I'm *pretty [07:51] sure* it's reported on xorg-devel. [07:51] Also, I'm pretty ;) [07:52] Alrighty then :) [07:52] RAOF: You mean it's a known issue? [07:52] I saw mareko say something on dri-devel that might be the issue [07:52] I think it might be. [07:53] apparently it only affects r3xx and not r5xx [07:53] tjaalton: I'll leave you to build it then. [07:55] RAOF: Possibly OT: Do you know if ubuntu will run X in wayland or even package wayland in the next cycle (or two)? [07:55] it's packaged [07:56] Run X in wayland by default in the next two cycles? That's a big N O. [07:57] Wayland is, however, packaged, and should be runnable under X and on the console for O. We'll get a new snapshot that doesn't require cairo-gl, so nvidia users won't be shafted. [07:57] RAOF: I'm kinda thinking that wayland would make it's debut in ubuntu by just running X inside it but perhaps next year, there will be enough stable toolkits that it will be able to run native stuffs [07:57] but I'm merely speculating [07:57] Well, that's the plan. [07:58] However, 12.04 is an LTS; we're not going to throw Wayland under X in the LTS :) [07:58] ah ok [07:58] Also I'm curious what nvidia is planning to do WRT wayland. AFAIK, they have no plans to provide support for it in their driver [07:59] Will they ever cough up hw specs you think? [07:59] All this is discussed in one of the Wayland sessions at UDS; the notes are on the etherpad. [07:59] RAOF: Link me? [07:59] (Well, not the nvidia discussion; that's a whole different kettle of awkward) [08:00] I'm more curious to know than anything [08:00] * soreau doesn't even use nvidia [08:02] http://summit.ubuntu.com/uds-o/meeting/desktop-o-xorg-wayland-something-or-other/ [08:02] heh, wayland-something-or-other [08:02] RAOF: thanks [08:05] cool [08:06] I'm surprised how quickly wayland is taking off and gaining a large amount of support by everyone (except nvidia) [08:06] I wonder how compiz will fare ;) [08:06] Compiz is surprisingly X-agnostic. [08:07] Yea I know. And porting to gles is probably only a quarter of the work [08:07] Porting to EGL might require an entire rewrite [08:32] RAOF: hum, building this on natty is more work it seems, need to backport llvm and libdrm [08:33] tjaalton: Yeah, of course :). I thought you were oneiricing :) [08:33] not on the desktop, though I could build it on my laptop which is dualcore as well [08:33] booted it up [08:37] and I'll upgrade the other laptop too, no point in keeping it on natty [11:29] RAOF: ok, finally got it built, but I have a hook that tries to install them and that phase fails [11:40] RAOF: http://pastebin.com/7n1LDKez [11:40] probably nothing, just a symptom of how the hook installs stuff [11:40] it just runs 'dpkg -i $tmpdir/*.deb' [11:44] hmm right, since libgl1-mesa-swx11 conflicts with libgl1, which is provided by libgl1-mesa-glx [11:45] so nothing to worry about, I just need to disable the hook because the failure makes the pbuilder run fail, and so the debs are gone :) [11:54] tjaalton: Oh, ouwch! [12:14] so the thinkpad fan control doesn't work, and the kernel is reporting near-critical temperature.. [12:14] while building mesa [12:16] * RAOF fires off an i386 mesa build. [12:16] it should finish soon though :) [12:16] my build [12:16] If your laptop doesn't melt first! :) [12:17] llvm takes a surprising amount of time to build. [12:18] placed a frozen bottle of water under the machine :) [12:18] Or, rather, there's an annoying delta between build_time(on launchpad buildd) and build_time(on laptop) [12:19] For the longest time, stuff was quicker to build locally than on the buildd's. [12:22] ooh, dpkg-shlibdeps.. soon finished [12:22] CPU temp down to 81 [12:23] if i'm reading /proc/acpi/ibm/thermal right [14:23] ok, so how do I determine if llvmpipe is used or not? [14:23] swrast_dri.so is loaded [14:23] glxinfo | grep render [14:25] http://paste.ubuntu.com/614934 [14:26] you are not running llvmpie. [14:27] right [14:27] it should say "Gallium 0.4 on llvmpipe" or something like that [14:27] does it build a dri driver of its own? [14:28] maybe the packaging change wasn't complete [14:32] dunno, upstream it was/is called swrastg_dri.so [14:33] yeah, probably just needs to be included in the package then :) [17:45] is #ubuntu-x tracking glew? [17:45] is there any plan to update to 1.6? [17:46] does anybody know if bug #711401 got investigated after the version revert to figure if the issue was a nux or glew one? [17:46] Launchpad bug 711401 in unity (Ubuntu) (and 3 other projects) "update to glew 1.5.7 broke unity (dup-of: 711396)" [Undecided,Confirmed] https://launchpad.net/bugs/711401 [17:46] Launchpad bug 711396 in nux (Ubuntu Natty) (and 4 other projects) "segfault in nux::IOpenGLFrameBufferObject::Deactivate (affects: 42) (dups: 13) (heat: 178)" [Undecided,Fix released] https://launchpad.net/bugs/711396 === yofel_ is now known as yofel === soren_ is now known as soren === maxb_ is now known as maxb