When starting Xgl, it segfaults. Bugs seems to be caused by evdev recognizing my keyboard and mouse not 100% correctly and assigning the multimedia-keys of my keyboard as mouse buttons, resulting in 99 buttons. MAX_BUTTONS is set to 64 in hw/xgl/glx/xglx.c , setting it to 99 fixes this bug. Stack trace: (gdb) start :1 Breakpoint 1 at 0x80a7ae0: file main.c, line 255. Starting program: /opt/Xgl/bin/Xgl :1 main (argc=2, argv=0xbfc7d444, envp=0xbfc7d450) at main.c:255 255 display = "0"; (gdb) continue Continuing. [Thread debugging using libthread_db enabled] [New Thread -1211271488 (LWP 32351)] Program received signal SIGSEGV, Segmentation fault. [Switching to Thread -1211271488 (LWP 32351)] 0x080863ab in InitButtonClassDeviceStruct (dev=0x44434241, numButtons=99, map=0xbfc7d2e7 "") at devices.c:590 590 dev->button = butc; (gdb) bt #0 0x080863ab in InitButtonClassDeviceStruct (dev=0x44434241, numButtons=99, map=0xbfc7d2e7 "") at devices.c:590 #1 0x08086b43 in InitPointerDeviceStruct (device=0x44434241, map=0xbfc7d2e7 "", numButtons=99, motionProc=0x81190fa <miPointerGetMotionEvents>, controlProc=0xb798f16e <xglxChangePointerControl>, numMotionEvents=256) at devices.c:813 #2 0xb798f282 in xglxPointerProc (pDevice=0x82f8bc8, onoff=0) at xglx.c:1325 #3 0x08085376 in InitAndStartDevices () at devices.c:177 #4 0x080a7fdf in main (argc=2, argv=0xbfc7d444, envp=0xbfc7d450) at main.c:400
Sorry about the phenomenal bug spam, guys. Adding xorg-team@ to the QA contact so bugs don't get lost in future.
I can partially confirm, and reproduce this bug. Not sure if it only involves XGL. Nor am I sure if it can be fixed as proposed elsewhere in the bug comments. When I perform "X" from the command line, the submitted strace occurs using the following configuration after which X segfaults (or so the strace says..). Originally I suspected it to be my own fault due to custom CFLAGS, but I can reproduce this on more than one linux distribution (gentoo and archlinux) which removes that possibility. Section "InputDevice" Identifier "TrackManOptical" Driver "evdev" Option "Protocol" "Auto" # cat /proc/bus/input/devices Option "Dev Name" "Logitech USB Receiver" Option "Dev Phys" "usb-0000:00:02.1-3/input0" # May use /dev/input/mice instead, but I have two USB mice. Option "Device" "/dev/input/event2" Option "Buttons" "10" Option "ZAxisMapping" "4 5 6 7 8 9 10" # This will make the button 8 the "lock" modifier. Option "DragLockButtons" "8" Option "AngleOffset" "-7" EndSection
Created attachment 10449 [details] strace output on archlinux compiled for x86
Created attachment 10450 [details] strace output on archlinux compiled for x86 (This time uploaded with proper headers)
The bug can _not_ be reproduced by: Section "InputDevice" Identifier "Keyboard1" Driver "evdev" Option "Protocol" "evdev" Option "Device" "/dev/input/event0" Option "XkbModel" "microsoftnek4k" Option "XkbLayout" "dvorak" Option "XkbOptions" "ctrl:swapcaps,altwin:super_win" #Option "XkbOptions" "ctrl:swapcaps" EndSection Because when commenting out the mouse section, but not this section, X does load.
Mass closure: XGL is gone, these are never getting fixed.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.