[01:02] hello, I tried installing some video drivers for a second graphics card ihave and now all i have is a black screen upon bootup, does anyone know how to make linux go back to the default video drivers, as they were when i installed the system? === yofel_ is now known as yofel [01:04] Stericson: This depends on how you installed the new drivers. [01:04] I used ubuntu hardware gui [01:05] Then, from a terminal, ‘jockey-text’ will allow you to uninstall them in the same way. [01:06] I cannot even get to a terminal, when it boots up the screens go black. [01:06] right now I am in Windows, I can get to the linux file system [01:07] You *should* be able to boot into “recovery mode” and run ‘jockey-text’ from there. That'll be easiest. [01:08] I will try again however I did try to go into recovery mode and it seemed to do the same thing. [01:10] Hm. That shouldn't happen :( [01:11] well, I will tell you what I was doing, I have two graphic cards in my machine. One is a Nvidia, the other one is an ATI card [01:12] Oh, man. [01:12] so I tried to switch to the ATI card [01:12] installed the listed ATI drivers [01:12] So, at least one problem here is that the nvidia and ati drivers cannot coexist. [01:12] and then I get nothing after a reboot [01:12] not one Linux anyways [01:12] on* [01:12] lol [01:12] most of us tend to pick one card and leave the other out of the mix [01:13] otherwise it's like a traffic accident [01:13] (Well, actually, the *open source* nvidia and ati drivers coexist just fine; just not the binary drivers) [01:14] I have a four monitor setup right now on Windows, in the future I may get matching cards, but not at the moment. [01:14] I was hoping they would work as well as they do in Windows [01:14] That's something that's poorly supported right now, unfortunately. [01:14] which? [01:15] 4 monitors or two cards [01:15] Multi-card multi-monitor. [01:15] oh [01:15] both [01:15] lol [01:15] No; just the multi-card bit. [01:15] I see [01:15] X is happy to drive 6 displays from a single EyeInfinity card. Just not drive two cards at once (at least, not without a bunch of setup and bugs that don't normally get tested). [01:15] RAOF, if you had two cards in there at once, for instance two pci cards, which one would be picked as the primary display adapter? [01:15] well ideally i would like to have just the two on the ATI card but it is the secondary card is that a problem? [01:16] seemingly so because I get no display when I activated the ATI drivers [01:18] You should get *at least* a text mode boot when you select recovery mode. [01:18] But that will go through whatever card your BIOS thinks is primary, though. [01:19] yea I see the initial bootup, I even get to select recovery mode and watch it do some stuff, but then it goes awau [01:19] away* [01:23] So you don't get to a blue-background text menu? Oh, urgh. [01:23] * RAOF wonders what's happening there. [01:24] not sure, give me a few moments, I need to boot my laptop up which has Linux on it, then I can open xchat and get this computer back into linux to see what I can do. [01:25] ooh, new mesa bits in edgers [01:25] hmm, I have to restart X before I can test those, right? [01:25] yes [01:25] Thought so, even though libEGL dynamically loads them [01:26] Actually, no. [01:26] Direct rendered clients will pick that up immediately. [01:27] in which case no luck for me, I should probably file a bug about this :) [01:27] but now I have new bits I can point to when filing it [01:27] (Because the server doesn't actually do very much with GL rendering at all; just provides the DRI2 interface, which is stable) [01:27] Heh. [01:28] Your problem with getting GL and GLES to actually work? :) [01:28] Can someone with intel hardware (pre-sandybridge) see if any EGL/GLES apps actually work? [01:28] RAOF: For some odd reason even the demo that was working doesn't after a rebuild [01:29] You want that tested against edgers, or natty? [01:29] And now I seem to have to #define SUPPORT_X11 to get the right defines in eglplatform.h for things to build [01:29] RAOF: edgers [01:29] Bah. [01:29] RAOF: Well, see if it works in natty [01:29] It didn't with sandy bridge [01:30] I pulled mesa/demos from git and the EGL/GL demos work but the EGL/GLES and EGL/GLES2 demos all fail in pretty much the same way everything I've tried fails [01:30] Except for compiz which seems to fail in its own special way [01:30] lol [01:31] in natty it gives a symbol error, with edgers I seem to get a loaded but not loaded opengl plugin (which probably means I got the same error as the demos but failed to check for it) [01:35] Both gears_x11 and es2gears work for me. [01:36] so it's apparently either sandy bridge specific or lolapple [01:37] How do they fail in natty? [01:37] glcompbench does this: Error: eglGetConfigAttrib() failed with error: 12293 Error: eglCreateContext() failed with error: 12288 [01:37] the mesa demos all do: es2_info: es2_info.c:153: make_x_window: Assertion `num_configs > 0' failed. [01:38] Freakydeaky. [01:39] What's really weird is that means eglChooseConfig isn't failing [01:40] hmm yea so upon choosing recovery it does some scrolling of text and then it just dies, all of my displays go black [01:40] URGH. [01:41] That's really obnoxious; we shouldn't let that happen. [01:42] stericson__: What version of Ubuntu are you using? [01:42] wow I cannot even get into my bios, interesting [01:42] 10.04 [01:43] Can't get into your bios? Oooh. That might be caused by the same thing that stops recovery mode from working. [01:43] ah nvm [01:43] I got into it [01:43] weird :) [01:46] Could you edit the menu item for recovery mode (to do this press ‘e’) and add ‘vesafb.sucks=1’ to the end of the kernel line? It'll be the one with “root=/dev/sdaX” or simialr in it. [01:47] That should (hopefully!) stop the transition to the blank screen. [01:48] ok, one moment [01:49] so I have set root='(hd1, 7)' you want it to say set root='(hd1, 7)vesafb.sucks=1'? [01:50] RAOF, do you know if there is a python library wrapper thingee for editing kernel command lines in grub? [01:50] No; there should be a line a couple below that with “2.6.35-??-…” in it with root=/something or other. [01:51] like, in the apport hooks if i wanted to give the user a pushbutton to do something like disabling vesafb or whatnot [01:51] ah I see it [01:51] bryceh: Not that I know of, no. It'd be quite easy to do, though; you'd mess with /etc/default/grub instead. [01:52] hmmm that did not work, in fact, i saw less go across the screen this time. [01:53] :(\ [01:53] Do you have a monitor plugged in to each video card? [01:53] yea, two in each [01:53] maybe I should swap the slots the cards are in [01:53] *shrug* [01:56] hahaha [01:57] ? [01:57] I pulled my extra video card out [01:57] got the blue screen [01:57] which makes me think it was trying to output to the extra video card [01:57] that it strange [01:57] is* [01:58] *shrug* [01:58] That is strange. [01:58] yep indeed indeed [01:59] and I have graphics again [01:59] yea [02:00] so is there no way at all to get the two cards working together? [02:01] It should be possible; there's no GUI tool to do the configuration, though. [02:02] You'll have to remove the driver installed by the Hardware manager; that's the ATI proprietary driver which breaks your other driver ;) [02:03] Then faffing around with xorg.conf is what you'll need to do. [02:03] It'll look *something* like http://ubuntuforums.org/showthread.php?t=1489377 [02:04] You might be able to get away with just two “Device” sections; one with Driver “radeon”, one with Driver “nouveau”. [02:05] That example is for two ati cards, but (apart from the binary drivers breaking stuff) ati and nvidia should work similarly. [02:07] ok, cool, thanks for your help, I really appreciate it [02:09] also, one of my monitors is not being detected correctly, do you know how to manually add in the proper resolution for it? [02:09] Yup. [02:10] A combination of “cvt” and “xrandr” can do it at runtime, or… [02:10] wiki.debian.org/XStrikeForce/HowToRandR12 [02:10] thanks again :) [04:06] OK, so I ordered the sandybridge laptop today. It ought to arrive just about exactly too late to be any good for testing before Natty's released. === 50UAAGBPG is now known as Amaranth === lool- is now known as lool === lilstevie|lagche is now known as lilstevie [15:22] Amaranth: will see if I can figure out whats going on now, it's quite possible I've fudged up the gles packaging in edgers since things have changed so much [15:23] Sarvatt: it didn't work with stock natty either, except for some reason the application I compiled on a different machine some time back did still work [15:24] hrm [15:24] good point, I'm actually on stock natty atm on this machine, thought it was edgers [15:42] Amaranth: hrm well EGL_SOFTWARE=1 works if you just need a quick test [15:42] Sarvatt: True, but that's not useful for running compiz :) [15:42] 0.5fps is no fun when it's controlling your entire screen [15:48] the dri driver is only reporting opengl (0x8) here when I run it under EGL_LOG_LEVEL=verbose [15:50] Yeah, that's the root issue but it's rather weird how it fails, it seems perhaps this case was overlooked [15:51] None of the egl functions seem to ever return EGL_FALSE, they just don't do what they say they will so unless you do extra checks like the demos do you end up with fun crashes later [16:00] gles is only enabled for gallium drivers? [16:01] EGL_DRIVER=egl_gallium eglinfo is showing OpenGL_ES and OpenGL_ES2 client APIs, egl_dri only has OpenGL [16:02] yeah es2_info works under EGL_DRIVER=egl_gallium, hrm [16:02] might just need to configure mesa differently to have gles under egl_dri2 [16:03] checking on a non sandybridge intel.. [16:04] wonder why eglinfo isn't shipped in mesa-utils-extra [16:06] odd, 945 defaults to egl_gallium, sandybridge egl_dri2 [17:38] bryceh: YokoZar is updating ia32-libs now \o/ [17:40] Sarvatt: hello, do you mind if I pm you? :) [17:40] Heyo! whats up BlackZ? [17:40] sure thing [17:41] Sarvatt: thanks :) [17:43] Sarvatt, excellent [17:43] Sarvatt, I hear vorlon has been making scads of progress on multiarch as well [17:46] yeah I see tons of stuff for mesa in git, haven't had a chance to look at it yet though [18:20] Sarvatt, wouldn't it be quicker to build mesa in pbuilder to see if it will build before sending it into the ppa? [18:21] is that a joke? :P [18:21] no [18:22] it builds quicker locally than it does in the ppa system [18:34] i do the packaging on an atom server, the uploads are automated and the failed to build logs are useful for passing along upstream. if you use gmail you can just filter out messages containing "Archive: xorg-edgers PPA" and "State: Failed to build" if the mails bug you [18:35] mesa takes about 2 hours to build on that thing, i'm not adding a build step before uploading on it [18:35] oh, i wasn't aware of any of that [18:36] mesa takes 2 hours to build? [18:36] yup [18:36] on an atom, still about 40 minutes on this i7 laptop if i did move it over, all my sandybridge machines get reimaged too often to use one of those [18:37] 40 minutes on a system that fast is a lot [18:37] i wonder how much of that code is just old junk [19:30] hi. how can I force the use of the gallium driver instead of the classic mesa one? [20:10] bryceh, I've got a fix for bug 736523 [20:10] Launchpad bug 736523 in xserver-xorg-input-synaptics (Ubuntu) "cannot left click and drag with touchpad (regression) (affects: 4) (dups: 1) (heat: 24)" [High,In progress] https://launchpad.net/bugs/736523 [20:10] but it's a larger than normal change for this late in the cycle [20:10] I've put it in our ppa:utouch-team/unstable [20:11] but I'd like you (and/or others) to review it too before adding it to the packaging rep [20:11] repo [20:11] (when lp finishes making the debdiff I'll paste a link) [20:13] bryceh, https://launchpadlibrarian.net/67613705/xserver-xorg-input-synaptics_1.3.99%2Bgit20110116.0e27ce3a-0ubuntu8~utouch1_1.3.99%2Bgit20110116.0e27ce3a-0ubuntu9~utouch1.diff.gz [23:58] cnd, sorry just now getting around to looking at your patches [23:58] cnd, you might include a mention of the upstream status for them in the changelog entry