Re: xemacs / libX11 / NVIDIA crashes on opteron under kernel 2.6.14-2

Lonni J Friedman writes: > I have a few questions Thanks for the quick response. > 0) Does this reproduce if you're not using RenderAccel? Yes. At least commenting out # Option "RenderAccel" "true" in the screen section of xorg.conf and restarting X doesn't fix the problem. Similar looking backtrace: Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 46912549568976 (LWP 1252)] 0x00002aaaabfb4ac0 in XCheckIfEvent () from /usr/X11R6/lib/libX11.so.6 (gdb) backtrace #0 0x00002aaaabfb4ac0 in XCheckIfEvent () from #/usr/X11R6/lib/libX11.so.6 #1 0x000000000048a0b4 in emacs_shell_event_handler () #2 0x000000000055c5ff in check_quit () #3 0x0000000000518265 in make_string_from_buffer () #4 0x00000000005196d5 in buffer_delete_range () #5 0x000000000046baa5 in Fdelete_region () #6 0x0000000000451333 in execute_rare_opcode () #7 0x00000000004521db in execute_rare_opcode () #8 0x0000000000453f70 in funcall_compiled_function () #9 0x000000000047be41 in Ffuncall () #10 0x0000000000451809 in execute_rare_opcode () #11 0x0000000000453f70 in funcall_compiled_function () #12 0x000000000047be41 in Ffuncall () #13 0x0000000000451809 in execute_rare_opcode () #14 0x0000000000453f70 in funcall_compiled_function () #15 0x000000000047be41 in Ffuncall () #16 0x0000000000451809 in execute_rare_opcode () #17 0x0000000000453bdb in Fbyte_code () #18 0x000000000047a3c1 in Feval () #19 0x00000000004770fe in condition_case_1 () #20 0x0000000000451444 in execute_rare_opcode () #21 0x00000000004521db in execute_rare_opcode () #22 0x0000000000453f70 in funcall_compiled_function () ---Type <return> to continue, or q <return> to quit--- #23 0x000000000047be41 in Ffuncall () #24 0x0000000000451809 in execute_rare_opcode () #25 0x0000000000453f70 in funcall_compiled_function () #26 0x000000000047be41 in Ffuncall () #27 0x000000000047c3cc in call1 () #28 0x00000000004c3fc1 in Fdispatch_event () #29 0x000000000045ac05 in Fcommand_loop_1 () #30 0x00000000004770fe in condition_case_1 () #31 0x000000000045a774 in Freally_early_error_handler () #32 0x00000000004745ce in internal_catch () #33 0x000000000045a8bc in initial_command_loop () #34 0x00000000004737ba in xemacs_21_4_18_x86_64_debian_linux () #35 0x0000000000474269 in main () (gdb) > 2) For what its worth, I'm able to run xemacs-21.4.18 with > nvidia-1.0-8174 in a 32bit environment without any problems. Is > there anything to suggest that this is a 64bit specific issue? Just that I am running a full 64bit environment. For what it is worth, I have to comment out the Load "glx" line in the Modules section in order to get X to start at all. > 3) Does the crash go away if you switch to using the 'nv' X driver? No, it doesn't. For some reason, the nv driver gives me very bad rendering at 1600x1200 (the native of my lcd screen), so I haven't been able to use it. But even giving that a try (at a lower resolution), I still get the crash. E.g. Program received signal SIGSEGV, Segmentation fault. [Switching to Thread 46912549568976 (LWP 1987)] 0x00002aaaabfb4ac0 in XCheckIfEvent () from /usr/X11R6/lib/libX11.so.6 (gdb) backtrace #0 0x00002aaaabfb4ac0 in XCheckIfEvent () from #/usr/X11R6/lib/libX11.so.6 #1 0x000000000048a0b4 in emacs_shell_event_handler () #2 0x000000000055cdc2 in check_what_happened () #3 0x000000000047bd95 in Ffuncall () #4 0x000000000047c3cc in call1 () #5 0x00000000004c3fc1 in Fdispatch_event () #6 0x000000000045ac05 in Fcommand_loop_1 () #7 0x00000000004770fe in condition_case_1 () #8 0x000000000045a774 in Freally_early_error_handler () #9 0x00000000004745ce in internal_catch () #10 0x000000000045a8bc in initial_command_loop () #11 0x00000000004737ba in xemacs_21_4_18_x86_64_debian_linux () #12 0x0000000000474269 in main () (gdb) I guess this suggests it's really an libx11 bug, not so much an NVIDIA problem. Larry