[00:01] bryceh, there's no upstream status for them [00:01] because the support is built on top of the multitouch work [00:01] which doesn't exist in upstream synaptics yet :) [00:02] I can note this if you'd like [00:02] tbh, I'm not sure what to do with upstream x synaptics either, because they will need to resolve the issue without multitouch support [00:02] which will be rather difficult [00:04] bryceh, btw, in case you get any questions about input coordinate transformation (swap axes, rotate, etc.): https://wiki.ubuntu.com/X/InputCoordinateTransformation [00:05] ok, yeah just indicate they're dependent on ubuntu's MT framework or something, that should be enough [00:05] cnd, excellent thanks [00:05] cnd, I've read through the patches and roughly follow the changes, no comment on them since I'm not super familiar with this stuff but looks sane [00:07] bryceh, ok, cool [00:08] they're in our ppa so I hope they are getting some testing by a handful of people [00:08] cnd, and yeah it's too late for beta1 but go ahead and push into -synaptics git packaging tree [00:08] ok [00:08] in the meanwhile I'll load it on my laptops and netbook to dog food it a bit [00:08] I'd like to get them pushed as soon as beta 1 is released so we get as much exposure early on as possible :) [00:09] sounds good [00:09] it actually makes my trackpad soooo much better [00:09] that gave me a bit of a kick in my step today :) [00:09] :-) [00:09] bryceh: So I'm in Natty on may main laptop now and finding in much better for Intel 945 than Maverick. [00:12] ScottK, good to hear [00:12] ScottK, in terms of stability, performance, ...? [00:12] yes. [00:12] nice [00:12] Actually mostly performance and quality of the display. [00:13] I didn't have a lot of stability trouble since just after maverick release. [00:13] I just upgraded today, so it'll be a few days before I have a solid opinion on stability. [12:22] the spurious mouse button release/clicks turned out to be due to the microswitch. cleaned it up and can't reproduce it anymore :) [13:13] Is there a proprietary nvidia driver that supports Vanta? (1999) [13:16] I believe I'd need nvidia-71, but I can't find it in the archives. Should I try this? http://www.nvidia.com/object/linux_display_ia32_71.86.13.html [13:20] Or should I try the one from debian sid? http://packages.debian.org/sid/nvidia-glx-legacy-71xx === yofel_ is now known as yofel [13:21] you won't be able to use that with a halfway recent xserver. [13:21] there probably is no blob that supports both your card and current(ish) xserver [13:21] echo [13:22] alkisg: I think ~2.6.18 is the latest kernel that supports too [13:22] Ah :( That's on a chroot, to be served on netbooted clients (LTSP). I can build a hardy chroot, if needed. [13:22] I saw a -glx-71 package for hardy, would that do? [13:24] http://packages.ubuntu.com/hardy/nvidia-glx-legacy [13:26] the nvidia page shows that it should work, so.. maybe [13:27] Thank you guys, I'll give that a try [19:59] bryceh, didrocks said you might know somethign/be able to help about the new fglrx not working with unity from the daily build ppa [20:04] cnd: Sarvatt said it'd need a newer compiz? [20:04] that was tseliot [20:05] ah, sorry [20:05] http://fossplanet.com/f10/%5Bubuntu-x%5D-fglrx-support-natty-118210/ [20:06] tjaalton, yeah, I've got the new compiz [20:06] compiz seems to be running fine [20:06] it's the panel and the launcher that seem broken [20:10] + // Avoid detection of the compiz program by the fglrx driver. [20:10] + // Change program name to work around app detection in fglrx, [20:10] + // which is needed for older versions of compiz but breaks this [20:10] + // version of compiz. [20:10] sheesh [20:11] got my ati box doing another set of upgrades, lets see if I have the same problem.. [20:11] hahaha [20:11] Sarvatt, yeah [20:11] Sarvatt, I take it you weren't cc'd on that massive internal thread we had about that? [20:12] which one? was on one a few weeks ago when we set up the PPA [20:12] after that one X-) [20:12] woohoo, yeah I missed out :) [20:12] gotta love working on proprietary drivers :-/ [20:14] I'm going to get those fugly scrollbars forced on me if I use the unity PPA aren't I? :P [20:15] I think they're cute (but haven't tried using them yet) [20:19] man I'm getting sick of -intel gpu lockup bugs [20:20] want to just turn them off? they're firing when the system recovers fine now too [20:21] still? thought apw fixed that [20:21] vblank_mode=0 glxgears on a sandybridge system is a good example [20:21] no, we should leave them on, there's still some legit problems [20:21] dmesg flooded with hangcheck timer messages, crash report every 2-3 seconds in /var/crash/ [20:21] only visible problem is a stutter while it does the dump [20:22] any idea what causes it? [20:22] just inconsistent vblanking or something? [20:22] nope, apparently ickle can't reproduce it [20:22] compiz + sandybridge + vblank_mode=0 glxgears always triggers it here [20:23] well, you know we still gotta update the apport hook to stop using intel_gpu_dump [20:25] ok without adding the unity PPA choosing a Ubuntu session launches a classic session, adding the PPA now [20:30] err, classic desktop picking the ubuntu session with it too.. [20:34] unity works with fglrx if i start it manually, of course I can't get rid of gnome-panel thats over top of it [20:35] don't know whats up with that, unity was working fine without fglrx [20:36] of course i did the weekly 600mb upgrade before rebooting too [20:37] removed fglrx and unity session works again, very odd [20:38] lessee what .xsession-errors has to say, maybe it bails out of unity for some reason [20:40] out of all my systems, the one using r600g by far has the best experience in unity [20:41] thats strange, compiz doesn't even try to load with fglrx installed [20:42] http://paste.ubuntu.com/587508/ http://paste.ubuntu.com/587509/ [20:43] http://paste.ubuntu.com/587510/ [21:13] cnd: I don't know what bug you're actually having after discussing my fglrx/unity problem in #ubuntu-desktop.. [21:13] cnd: have you been dist-upgrading that machine? is it possible you dont have gnome-panel installed? [21:14] it tried to remove it a few times in the past month or two here [21:14] the unity session falls back to the classic gnome session because of a screwed up exit status after running /usr/lib/nux/unity_support_test -p [21:14] just with fglrx [21:15] Sarvatt, I still have the panel in unity 2d [21:15] cnd: but do you have the gnome-panel package installed? I dunno if unity-2d uses that [21:15] I'll check [21:15] unity_support_test -p says everything looks good [21:16] but unity_support_test alone returns 234 [21:16] instead of 0 [21:16] do a echo $? after running it [21:16] yeah thats what I get [21:16] I still have gnome-panel installed [21:17] so to test out unity, I have to hack up /usr/share/gnome-session/sessions/ubuntu.session [21:17] maybe if I install unity-2d the panels will be screwed too, lets see :) [21:17] and replace the IsRunnableHelper with /bin/true [21:18] didrocks said the unity_support_test will be fixed in tomorrows build [21:18] yeah [21:20] ah yea i see it tries to do a unity-2d session as the fallback first if its installed, thats probably where things are getting screwed up [21:20] oh, so if you uninstall unity-2d things work? [21:20] i just installed it to see if the panels are screwed [21:20] hasnt come up yet [21:21] the classic session fallback for sure works here [21:21] nope unity-2d works fine [21:21] yes, but does unity 3d still work after installing unity-2d? [21:21] unity 2d works fine here too [21:21] just not unity 3d [21:22] i mean choosing a ubuntu session which should start unity after unity-2d is installed to see if the fallback to unity-2d works does work so that theory is out the window [21:22] yeesh confusing myself [21:22] oh [21:23] was hoping the ubuntu session fallback to unity-2d might be where the panels were getting lost for you but it works here :) [21:24] yeah [22:04] Ok, now I think my computer is just screwing with me [22:04] Latest updates from edgers resulted in some king of GPU hang when starting compiz so I did a ppa-purge and now with stock natty stuff GLES is working again [22:05] yep latest update borked sandybridge here too [22:05] it was 5 days of updates in one though since libdrm was busted [22:06] Sarvatt: That part isn't surprising, the surprising part is GLES stuff randomly working again [22:08] Although I still can't run compiz due to "symbol lookup error: /usr/lib/egl/egl_dri2.so: undefined symbol: _glapi_get_proc_address" but oh well [22:09] Amaranth: check out https://launchpad.net/~sarvatt/+archive/vanhoof if you want mesa speeds not to suck [22:09] https://bugs.freedesktop.org/show_bug.cgi?id=32677 ? [22:09] Freedesktop bug 32677 in Other "egl_dri.so depends on _glapi_get_proc_address" [Normal,Resolved: fixed] [22:10] Sarvatt: https://bugs.freedesktop.org/show_bug.cgi?id=34476 [22:10] Freedesktop bug 34476 in Mesa core "/usr/lib/egl/egl_dri2.so needs libGL but not link to it" [Normal,New] [22:10] But they look like the same bug [22:10] Although I'm not sure why this only hits compiz [22:11] Oh, because core doens't link to libGL or libGLES, we dlopen our libopengl.so plugin [22:12] so you're hitting a bug where gles isn't working over egl_dri2 in 7.11, and missing that other bug because of libglapi, and hitting that other bug in 7.10.. ugh [22:14] libglapi-mesa was the fix for the _glapi_get_proc_address problem [22:16] Yeah, but that's only in 7.11 which breaks GLES, like you said :) [22:19] I suppose I can just patch compiz to work around this for now, I guess [22:21] i'm not seeing any obvious reason why gles is broken over dri, wonder if the shared dricore has anything to do with it [22:26] * Sarvatt squints at --enable-shared-glapi [22:31] Amaranth: is es2_info showing you softpipe results on 7.10? [22:31] it is here [22:31] Sarvatt: If I don't define EGL_DRIVER, yes [22:31] thats the same as EGL_DRIVER=egl_gallium es2_info on 7.11 [22:31] oh ok [22:32] yep egl_dri2 is actually working here too on stock natty [22:32] I've just stuck export EGL_DRIVER=egl_dri2 in my ~/.bashrc [22:32] Yeah, for me as well [22:33] i dont see -DHAVE_SHARED_GLAPI in the build and looking at the egl_dri2 source i'm thinking it needs it [22:34] going to try building edgers mesa with --enable-shared-glapi [22:34] Don't you already do that for edgers mesa? [22:34] RAOF pointed me to it because you build with that :) [22:35] I thought it was enabled by default, I'm not forcing it enabled or anything but it looks like egl_dri2 compiles differently if you do force it enabled.. [22:36] Um, yeah. Why aren't you building with --enable-shared-glapi? :) [22:37] it's building a shared glapi lib without it [22:38] The more you know™ [22:38] only ifeq ($(SHARED_GLAPI),1) isn't returning true so its not building egl_dri2 with it enabled? [22:38] * Sarvatt is confused [22:40] Ok, so let's try a build with --enable-shared-glapi :) [22:40] started off a build, lets see how it goes [22:40] Although since something else is killing SNB in general right now I'm not sure I'll be able to test it for you [22:41] ah hell yer right, i shoulda used the 0325 checkout [22:41] glad i still have the source handy [22:44] so much for taking off early because I had to be up at 7 for a call :P [22:45] eep [22:45] will know in about 20 minutes, this i7-2620M isnt very fast [22:46] Sarvatt: Is that by any chance a fruit product? :) [22:46] Also how can you say that's not fast? Compared to what? [22:46] got a i7-2820qm in another laptop thats tons faster, nope its not it's a dell :) [22:47] and 2500k and 2600k in the desktops blow them all away but i got discrete gpu's in those at the moment testing fglrx and nvidia.. [22:48] 2820M will only be faster for compiles [22:48] the 2620 is a dual core, mesa loves the quads [22:48] Which is what you're doing so yeah [22:48] Sarvatt, how fast is the sandybridge graphics in natty? [22:48] compared to previous intl stuff [22:49] phoronix had a benchmark, it's like 20% of the windows driver [22:49] pretty crappy until the new mesa gets uploaded, unity is slow as heck [22:49] I haven't noticed unity being slow [22:49] that sucks [22:49] Amaranth: do you not use the search to find apps? [22:50] typing in that is a lagfest [22:50] oh god [22:50] Amaranth: compare it to https://launchpad.net/~sarvatt/+archive/vanhoof, it'll blow your mind [22:50] Yeah, I made my dock 32px so I should shove 10 apps on there so I don't use the dash much [22:51] Yeah, 7.11 already has that patch so I've seen the difference [22:52] in windows its basically the same as a hd5470m discrete GPU, about the same as a 9600 on the nvidia side [22:53] It makes low end discrete GPUs a waste of money :) [22:54] funny story about that, an OEM was gonna ship a discrete model and had to change the discrete GPU at the last minute because the sandybridge side was faster :P [22:54] doesn't sound much different than a gma4500 [22:54] oh it blows that away, at least 10x faster [22:55] yay debs building now, didnt have to screw with .symbols [22:56] Ok so I changed compiz's call to dlopen from RTLD_LAZY to RTLD_LAZY|RTLD_GLOBAL as a temporary hack and now my desktop can run GLES builds of compiz too [22:57] hey do we still support hibernate? bug 745304 says he sees graphics corruption and wondering if I should wontfix it [22:57] Launchpad bug 745304 in xserver-xorg-video-intel (Ubuntu) (and 1 other project) "Graphics corruption after hibernate with Intel GMA 3150 chipset (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/745304 === cnd` is now known as cnd [22:59] yeah its still offered (unless you have an encrypted /home apparently) [22:59] ok [22:59] after a week of messing with different library versions and getting an ARM development board up and running so I can test this stuff I make a one-line change and it works just fine on my desktop [23:00] --enable-shared-glapi made no difference [23:01] Sarvatt: Does it not set the define or does the define not help? [23:01] yeah it sets the define right [23:01] will make that change in edgers regardless [23:06] Hi all. Should I expect hardware accelerated video for my crappy graphics card using the PPA? I have "Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)" It worked in 10.04, but after upgrading to 10.10 I've had no such luck. :( [23:08] jhammons: That should be well supported by basically every release of Ubuntu ever. [23:08] :( [23:08] That's what I thought. [23:08] Although if you've upgraded from 10.04 to 10.10 without disabling the xorg-edgers PPA (as it says to ☺) then it's quite possible you've broken stuff. [23:09] I wasn't using this PPA with 10.04 [23:12] Which PPA are you thinking of specifically, and why are you using it? [23:14] RAOF: I tried ubuntu-x-swat thinking newer xserver-xorg-video-intel would help. [23:15] RAOF: Have any troubleshooting advice? [23:15] Except this: kernel: [ 24.608957] [Firmware Bug]: Duplicate ACPI video bus devices for the same VGA controller, please try module parameter "video.allow_duplicates=1"if the current driver doesn't work. [23:15] Will try that. [23:16] jhammons: The contents of /var/log/Xorg.0.log in a pastebin is often useful. [23:16] RAOF: Okay. Thanks. [23:16] Also, dmesg. [23:17] the type of improvements you're probably after aren't to be had from xserver-xorg-video-intel [23:17] bryceh: Indeed. :) [23:17] that X driver is really pretty limited these days [23:17] I see.. [23:18] It can screw things up, but it's not really involved in much that you care about. [23:18] probably where you should be looking for bleeding edge intel stuff are the drm-intel-next kernel daily builds [23:18] http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/ [23:19] beware sometimes those have new bugs. Work backwards from current until you find a kernel you like [23:19] also, those are raw upstream kernels, without any of the normal ubuntu enhancements [23:21] I see. [23:21] at the moment there probably aren't huge differences between stock natty and upstream stuff in any case, but enough for it to be worth testing [23:22] I'm going to try the suggestion I found in dmesg output (see above). Here's my Xorg log. I'll be back in a minute. If it doesn't work, a quick gander at that would be appreciated: http://pastebin.com/UCwVaFQG [23:23] I think what it is referring to is my stupid dock for the laptop. [23:24] [ 29.354] (EE) Failed to initialize GLX extension (Compatible NVIDIA X driver not found) [23:24] Less binary nvidia driver, please :) [23:25] Hmm. [23:28] I'll be right back (hopefully). Thanks RAOF, bryceh. [23:31] Thanks for catching that RAOF! Works like a dream now. :) [23:31] If you kind folks ever have any issues with an Atlassian tool, hit me up: jhammons@atlassian.com -- thanks again!