http://bugs.freedesktop.org/show_bug.cgi?id=3493
------- Comment #7 from [EMAIL PROTECTED] 2007-03-02 18:22 PST ------- This is extremely easy to reproduce with SPECviewperf 9.0.3. (I have it installed to test my i965 hardware in the same machine.) Although unlike last time, I didn't get any (EE) R128(0): R128CCEWaitForIdle: (DEBUG) CCE idle took i = 49 messages in my Xorg.0.log. X unfroze when I killed the viewperf process that caused the [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock message. Terminal output from a full SPECviewperf run: [EMAIL PROTECTED]:/usr/local/src/opengl/SPEC/SPECViewperf9.0$ ./Run_All.csh Running: 3dsmax-04.csh Terminated Running: catia.csh Writing PNG file '../results/catia/01.png'...done. Terminated Running: Ensight-03.sh Terminated Running: light-08.csh Terminated Running: maya-02.csh Writing PNG file '../results/maya/01.png'...done. Terminated Running: proe-04.csh Writing PNG file '../results/proe/proe01.png'...done. Terminated Running: suzuki-engine Writing PNG file '../results/sw/sw01.png'...done. Writing PNG file '../results/sw/sw01Full.png'...done. Writing PNG file '../results/sw/sw02.png'...done. Writing PNG file '../results/sw/sw02Full.png'...done. Illegal instruction Running: ugnx-01.csh Segmentation fault Segmentation fault Terminated Terminated Terminated Terminated Terminated Segmentation fault Terminated Terminated Terminated Running: tcvis-01.csh Terminated Run All Summary 3dsmax-04 Weighted Geometric Mean = 0.00000 catia-02 Weighted Geometric Mean = 0.00000 ensight-03 Weighted Geometric Mean = 0.00000 light-08 Weighted Geometric Mean = 0.00000 maya-02 Weighted Geometric Mean = 0.00000 proe-04 Weighted Geometric Mean = 0.00000 sw-01 Weighted Geometric Mean = 0.00000 ugnx-01 Weighted Geometric Mean = 0.00000 tcvis-01 Weighted Geometric Mean = 0.00000 kernel log: Mar 2 21:25:46 tesla kernel: [65315.110796] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:28:21 tesla kernel: [65470.003625] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:28:32 tesla kernel: [65481.086515] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:28:42 tesla kernel: [65491.348416] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:29:07 tesla kernel: [65515.740211] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:29:48 tesla kernel: [65557.313944] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:31:07 tesla kernel: [65636.105453] viewperf[16959] trap invalid opcode rip:2b2deb5fea1b rsp:7fffc0bc8ed8 error:0 Mar 2 21:31:40 tesla kernel: [65669.276782] viewperf[16970]: segfault at 0000000007048f70 rip 0000000007048f70 rsp 00007fffa4036dc8 error 15 Mar 2 21:32:42 tesla kernel: [65731.491660] viewperf[16972]: segfault at ffffffffcd242f70 rip ffffffffcd242f70 rsp 00007fffdde3fbe8 error 14 Mar 2 21:35:38 tesla kernel: [65907.385109] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:44:39 tesla kernel: [66448.322471] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:49:10 tesla kernel: [66719.224818] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:51:42 tesla kernel: [66870.685200] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 21:56:50 tesla kernel: [67178.608402] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 22:01:33 tesla kernel: [67461.398885] viewperf[17042]: segfault at ffffffffe9c1df70 rip ffffffffe9c1df70 rsp 00007fffc14641e8 error 14 Mar 2 22:01:42 tesla kernel: [67470.173738] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 22:07:58 tesla kernel: [67846.066669] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 22:09:18 tesla kernel: [67926.606019] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock Mar 2 22:09:40 tesla kernel: [67948.741423] [drm:drm_lock_take] *ERROR* 3 holds heavyweight lock ugnx locked up the most. Maybe it just has a test script that invokes viewperf many times, instead of the batch mode the others use. Every "Terminated" line in that output is from me doing killall viewperf after a new kernel log message, and lack of CPU load, indicates that it got stuck again. I have no idea why it would run an illegal instruction, or segfault. Maybe the SIGILL was just from a jump to an address in an executable memory region that wasn't meant to be amd64 instructions. BTW, I compiled SPECviewperf 9.0.3 from the source in the src directory. ./Configure, and choose Linux64, IIRC. Mesa, r128_dri, and libdrm are all built from git sources. -- Configure bugmail: http://bugs.freedesktop.org/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee. ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
