Bugzilla – Bug 1334
Compiz hangs system on i915, i965GM
Last modified: 2008-11-02 11:46:25 UTC
You need to log in before you can comment on or make changes to this bug.
Two users with i965GM and snv_85 and snv_86 report that the bundled Compiz consistently produces screen corruption and hangs the system when started. Another user has the same problem with snv_82 (non-bundled package) and an i915.
Looks like a effect of http://bugs.opensolaris.org/view_bug.do?bug_id=6673072
(In reply to comment #1) > Looks like a effect of http://bugs.opensolaris.org/view_bug.do?bug_id=6673072 Actually screen corruption is the same as bug 6673072. It is a tiling issue after upgrade to 2.2.1. It can be fixed by a mesa patch. But for the Compiz hang, it doesn't caused by the upgrading of driver. I also reproduced this issue on snv_81 at my GM965 today.
(In reply to comment #2) > (In reply to comment #1) > > Looks like a effect of http://bugs.opensolaris.org/view_bug.do?bug_id=6673072 > Actually screen corruption is the same as bug 6673072. It is a tiling issue > after upgrade to 2.2.1. It can be fixed by a mesa patch. > But for the Compiz hang, it doesn't caused by the upgrading of driver. I also > reproduced this issue on snv_81 at my GM965 today. Is the mesa "patch" needed for Intel driver 2.3.0 and Mesa 7.0.3 for the Intel GM965 chipsets and before?
(In reply to comment #3) > (In reply to comment #2) > > (In reply to comment #1) > > > Looks like a effect of http://bugs.opensolaris.org/view_bug.do?bug_id=6673072 > > Actually screen corruption is the same as bug 6673072. It is a tiling issue > > after upgrade to 2.2.1. It can be fixed by a mesa patch. > > But for the Compiz hang, it doesn't caused by the upgrading of driver. I also > > reproduced this issue on snv_81 at my GM965 today. > > Is the mesa "patch" needed for Intel driver 2.3.0 and Mesa 7.0.3 for the Intel > GM965 chipsets and before? This patch is needed for driver 2.2.1 and Mesa6.5.2.
*** Bug 1409 has been marked as a duplicate of this bug. ***
Just tried intel 2.3.0 driver. It still hangs the system.
(In reply to comment #6) > Just tried intel 2.3.0 driver. It still hangs the system. Erwann, You'd need the patch from Intel (Liang) for Mesa 6.5.2. ~ Ken
(In reply to comment #6) > Just tried intel 2.3.0 driver. It still hangs the system. http://dlc.sun.com/osol/x/downloads/mesa-7.0.2-Xorg-1.3/
(In reply to comment #8) > (In reply to comment #6) > > Just tried intel 2.3.0 driver. It still hangs the system. > > http://dlc.sun.com/osol/x/downloads/mesa-7.0.2-Xorg-1.3/ I once tried both binary and source of mesa-7.0.2-Xorg-1.3 on b85 + Toshiba M9, but failed. For the binary: 1. 64-bit failed to start X and hang after loading libdri.so, please check the Xorg.0.log.64. 32-bit is OK to start X. 2. libglx.so still failed to load. No DRI 3. Keyboard driver has bug. Wrong key map. For the source: X failed to start as referenced symbol not found. Have you already successfully run mesa7.0.2 + X1.3 on your machine?
(In reply to comment #9) > I once tried both binary and source of mesa-7.0.2-Xorg-1.3 on b85 + Toshiba M9, > but failed. > > For the binary: > 1. 64-bit failed to start X and hang after loading libdri.so, > please check the Xorg.0.log.64. 32-bit is OK to start X. > 2. libglx.so still failed to load. No DRI > 3. Keyboard driver has bug. Wrong key map. > For the source: > X failed to start as referenced symbol not found. > > Have you already successfully run mesa7.0.2 + X1.3 on your machine? You mentioned you had a patch that allows your snv_b85 Mesa 6.5.2/Intel 2.2.1 video driver configuration to work correctly (i.e. not hang system) on your Toshiba Tecra M9. Can you include that mesa patch on this bug report for Albert and Erwann to test?
(In reply to comment #10) > > You mentioned you had a patch that allows your snv_b85 Mesa 6.5.2/Intel 2.2.1 > video driver configuration to work correctly (i.e. not hang system) on your > Toshiba Tecra M9. Can you include that mesa patch on this bug report for Albert > and Erwann to test? The patch I mentioned is for bug 6673072, not for system hang. I think this bug (system hang) will be automatically fixed after upgrading to mesa-7.0.2+Xorg-1.3 OR X11R7.4. (because I once test compize on my selfbuild mesa7.0.2+X1.4) So I'm still waiting for the upgrade now.
somehow compiz is mostly working here with my i915, the only problem are the window shadows, I only see a white stripe instead of a shadow I am using my own Xorg.conf instead of the autodetected one mainly because another issue with the latam keyboard not being available any other way. I'm pasting the relevant configuration files, can anyone else confirm? nacho ------------------------------------------------------------------------------- nacho@saturn:/var/log$ cat /etc/release OpenSolaris 2008.11 snv_91 X86 Copyright 2008 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 12 June 2008 nacho@saturn:/var/log$ uname -a SunOS saturn 5.11 snv_91 i86pc i386 i86pc nacho@saturn:/var/log$ cat /etc/X11/xorg.conf Section "ServerLayout" Identifier "X.org Configured" Screen 0 "Screen0" 0 0 InputDevice "Mouse0" "CorePointer" InputDevice "Keyboard0" "CoreKeyboard" Option "AIGLX" "true" EndSection Section "Files" RgbPath "/usr/X11/lib/X11/rgb" ModulePath "/usr/X11/lib/modules/amd64" FontPath "/usr/X11/lib/X11/fonts/TrueType/" FontPath "/usr/X11/lib/X11/fonts/Type1/" FontPath "/usr/X11/lib/X11/fonts/Type1/sun/" FontPath "/usr/X11/lib/X11/fonts/F3bitmaps/" FontPath "/usr/X11/lib/X11/fonts/misc/" FontPath "/usr/X11/lib/X11/fonts/100dpi/" FontPath "/usr/X11/lib/X11/fonts/75dpi/" EndSection Section "Module" Load "dri" Load "dbe" Load "glx" Load "record" Load "extmod" Load "GLcore" Load "IA" Load "xtrap" Load "freetype" EndSection Section "InputDevice" Identifier "Keyboard0" Driver "kbd" Option "XkbRules" "xorg" Option "XkbModel" "pc105" Option "XkbLayout" "latam" EndSection Section "InputDevice" Identifier "Mouse0" Driver "mouse" Option "Protocol" "auto" Option "Device" "/dev/mouse" Option "ZAxisMapping" "4 5 6 7" EndSection Section "Monitor" #DisplaySize 300 190 # mm Identifier "Monitor0" VendorName "LPL" ModelName "0" EndSection Section "Device" ### Available Driver options are:- ### Values: <i>: integer, <f>: float, <bool>: "True"/"False", ### <string>: "String", <freq>: "<f> Hz/kHz/MHz" ### [arg]: arg optional #Option "NoAccel" # [<bool>] #Option "SWcursor" # [<bool>] #Option "ColorKey" # <i> #Option "CacheLines" # <i> #Option "Dac6Bit" # [<bool>] #Option "DRI" # [<bool>] #Option "NoDDC" # [<bool>] #Option "ShowCache" # [<bool>] #Option "XvMCSurfaces" # <i> #Option "PageFlip" # [<bool>] Identifier "Card0" Driver "intel" VendorName "Intel Corporation" BoardName "Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller" BusID "PCI:0:2:0" Option "XAANoOffscreenPixmaps" "true" Option "DRI" "true" EndSection Section "Screen" Identifier "Screen0" Device "Card0" Monitor "Monitor0" SubSection "Display" Viewport 0 0 Depth 1 EndSubSection SubSection "Display" Viewport 0 0 Depth 4 EndSubSection SubSection "Display" Viewport 0 0 Depth 8 EndSubSection SubSection "Display" Viewport 0 0 Depth 15 EndSubSection SubSection "Display" Viewport 0 0 Depth 16 EndSubSection SubSection "Display" Viewport 0 0 Depth 24 EndSubSection EndSection Section "DRI" Group 0 Mode 0666 EndSection Section "Extensions" Option "Composite" "Enable" EndSection --------------------------------------------------------------------------------
(In reply to comment #12) > somehow compiz is mostly working here with my i915, the only problem are the > window shadows, I only see a white stripe instead of a shadow > I am using my own Xorg.conf instead of the autodetected one mainly because > another issue with the latam keyboard not being available any other way. > I'm pasting the relevant configuration files, can anyone else confirm? > > nacho I try your Xorg.conf on my M9(GM965) with build90. The system still hang after using compize. Liang Kan
Compiz is still hanging on build 95. It seems to work for approximately 10 seconds and then freezes up completely necessitating a hard restart. I also see white bars instead of shadows. -Ryan
For the people seeing white borders around windows instead of drop shadows, this caused by a bug affecting several drivers including intel and radeon when the shadow radius is a power of two. The radius is 8 by default, IIRC. Changing the radius to anything that isn't an exact power of two (e.g. 8.5) makes shadows visible again. The setting can be found under Window Decorations in ccsm (CompizConfig Settings Manager). It might be worth changing the default shadow settings in the next release since the problem appears to be as yet unsolved upstream.
Hey everyone, There is now a solution to this bug (at least on my system, intel gma x3100). After posting my problem here I was contacted by Niveditha Rau who provided me with the binaries for mesa 7.0.3. After setting up the new version of mesa everything worked but was quite laggy. The problem was my xorg.conf was attemping to use exa instead of xaa for the Accel Method. The instructions for installing the mesa 7.0.3 binaries are 1. Download binaries from https://mywebspace.wisc.edu/xythoswfs/webui/_xy-21296258_1-t_eXtFolxS please note this is not a sun supported download site, Niveditha mentioned that he could post them on the opensolaris website but for the interim I will post them here as I know some people really want them. I will also try to attach them to this ticket 2. Run $ pfexec svcadm disable gdm This will place you into a non graphical environment so you can safely upgrade the mesa drivers 3. Unpack the tarballs into /tmp/ (this is where the script expects the files to be) 4. Run the install-bin.csh script 5. Update /etc/X11/xorg.conf to use XAA instead of EXA (what it runs by if you do not specifically tell it to use XAA) 6. Reboot computer or run $pfexec svcadm enable gdm to renter a graphical environment 7. Enjoy compiz! For the white shadows Niveditha told me (settings are in the window decoration plugin of compiz) For the borders, could you please try the following settings (this sounds like http://defect.opensolaris.org/bz/show_bug.cgi?id=1418) Shadow Radius: 18 Shadow Opacity: 0.8 Shadow Color: #00000 (Black) Shadow OffsetX: 16 Shadow OffsetY: 0 Which corrected the white shadows My Xorg.conf Section "ServerLayout" Identifier "X.org Configured" Screen 0 "Screen0" 0 0 InputDevice "Mouse0" "CorePointer" InputDevice "Keyboard0" "CoreKeyboard" Option "AIGLX" "true" EndSection Section "Files" RgbPath "/usr/X11/lib/X11/rgb" ModulePath "/usr/X11/lib/modules/amd64" FontPath "/usr/X11/lib/X11/fonts/TrueType/" FontPath "/usr/X11/lib/X11/fonts/Type1/" FontPath "/usr/X11/lib/X11/fonts/Type1/sun/" FontPath "/usr/X11/lib/X11/fonts/F3bitmaps/" FontPath "/usr/X11/lib/X11/fonts/misc/" FontPath "/usr/X11/lib/X11/fonts/100dpi/" FontPath "/usr/X11/lib/X11/fonts/75dpi/" EndSection Section "Module" Load "record" Load "xtrap" Load "dri" Load "IA" Load "GLcore" Load "dbe" Load "extmod" Load "freetype" Load "glx" EndSection Section "InputDevice" Identifier "Keyboard0" Driver "kbd" EndSection Section "InputDevice" Identifier "Mouse0" Driver "mouse" Option "Protocol" "auto" Option "Device" "/dev/mouse" Option "ZAxisMapping" "4 5 6 7" Option "EmulateWheel" "on" Option "EmulateWheelButton" "2" Option "EmulateWheelTimeout" "200" EndSection Section "Monitor" #DisplaySize 300 190 # mm Identifier "Monitor0" VendorName "AUO" ModelName "4444" EndSection Section "Device" ### Available Driver options are:- ### Values: <i>: integer, <f>: float, <bool>: "True"/"False", ### <string>: "String", <freq>: "<f> Hz/kHz/MHz" ### [arg]: arg optional #Option "NoAccel" # [<bool>] #Option "SWcursor" # [<bool>] #Option "ColorKey" # <i> #Option "CacheLines" # <i> #Option "Dac6Bit" # [<bool>] Option "DRI" "true" # [<bool>] #Option "NoDDC" # [<bool>] #Option "ShowCache" # [<bool>] #Option "XvMCSurfaces" # <i> Option "PageFlip" "true" # [<bool>] Option "TextureFromPizmap" "true" Option "XAANoOffscreenPixmaps" "true" Option "AccelMethod" "XAA" # Option "AccelMethod" "EXA" # Option "ExaNoComposite" "false" # Option "MigrationHeuristic" "greedy" Identifier "Card0" Driver "intel" VendorName "Intel Corporation" BoardName "Mobile GM965/GL960 Integrated Graphics Controller" BusID "PCI:0:2:0" EndSection Section "Screen" Option "RenderAccel" "True" Option "AllowGLXWithComposite" "True" Option "AddARGBGLXVisuals" "True" Identifier "Screen0" Device "Card0" Monitor "Monitor0" SubSection "Display" Viewport 0 0 Depth 1 EndSubSection SubSection "Display" Viewport 0 0 Depth 4 EndSubSection SubSection "Display" Viewport 0 0 Depth 8 EndSubSection SubSection "Display" Viewport 0 0 Depth 15 EndSubSection SubSection "Display" Viewport 0 0 Depth 16 EndSubSection SubSection "Display" Viewport 0 0 Depth 24 EndSubSection EndSection Section "DRI" Group 0 Mode 0666 EndSection Section "Extensions" Option "Composite" "Enable" EndSection
Created an attachment (id=436) [details] Install Script for the 7.0.3 mesa binaries
Created an attachment (id=438) [details] mesa 7.0.3 32 bit binaries
Created an attachment (id=439) [details] mesa 7.0.3 64 bit binaries
Should this be considered a blocker for the OpenSolaris 2008.11 release? Regards, Edward.
Those who have tested with the Mesa 7.0.3 binaries have had success running compiz. Our plan is to integrated Mesa 7.0.3 for the OpenSolaris 2008.11 which should resolve this issue.
Those who have tested with the test Mesa 7.0.3 binaries have had success running compiz. Our plan is to integrate Mesa 7.0.3 for the OpenSolaris 2008.11 which should resolve this issue.
Adding as a integration blocker then.
*** Bug 2983 has been marked as a duplicate of this bug. ***
FWIW, I'm running a system with an Intel i915GM, and am NOT experiencing this problem. A quick scanpci shows pciids 8086/{2590,2592} for the agptarget and i915 drivers, and apparently unbound 8086/2792. After a gnome-cleanup (for unrelated reasons), I used ccsm to change my window shadow to 8.5, because in the past I had experienced the white bar issue. I then used System->Preferences->Appearance to select "Extra" Visual Effects. After multiple logout/login cycles, I have not had problems. My system is running OpenSolaris, updated to build 98, with Mesa library version 6.5.2. My SUNWdrmr pkg reports what I would expect for build 98, and I don't know how to map that to an intel driver version.
Marking incomplete with a request for the Reporter to try build 99 when it comes out to verify that the Mesa upgrade addresses the issue.
(In reply to comment #26) > Marking incomplete with a request for the Reporter to try build 99 when it > comes out to verify that the Mesa upgrade addresses the issue. The problem is resolved with Compiz on snv_99 here with an i965GM, but had to use XAA (not the default Exa) for 2D accel to avoid 2D corruption. Someone with a i915 should verify this.
(In reply to comment #27) > (In reply to comment #26) > > Marking incomplete with a request for the Reporter to try build 99 when it > > comes out to verify that the Mesa upgrade addresses the issue. > > The problem is resolved with Compiz on snv_99 here with an i965GM, but had to > use XAA (not the default Exa) for 2D accel to avoid 2D corruption. Someone with > a i915 should verify this. I can validate that XAA is the way to go (until Xserver 1.5.1/Mesa 7.2/Intel 2.5) on the older Intel chips. Testing Compiz 0.7.8 with Intel 2.4.2/Mesa 7.0.4/Sun Xserver 1.3.0/XAA acceleration gave decent performance on Intel 845G.
Given that the upgrade of Mesa to 7.0.3 addressed this issue, I'm marking it closed TRACKEDINBUGSTER. For any new issues in this area, please open a new bug.