xf86-video-modeset regression with Xorg upgrade

Dominic Fandrey kamikaze at bsdforen.de
Tue Feb 21 13:05:32 UTC 2017


On Mon, 2017-02-13 at 12:57 +0100, Dominic Fandrey wrote:
> With Xorg as provided by the CFT it worked fine. This morning I updated
> to the version in the ports tree and programs using OpenGL segfault.
> 
> The intel driver still works fine.

I took the time to throw all versions but llvm39 off the system and
rebuild xorg-server, dri and libEGL.

So here is a bt of an affected program (yquake2):
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 805816000 (LWP 100679/quake2)]
0x0000000808647dcb in __driDriverGetExtensions_i915 () from /usr/local/lib/dri/i965_dri.so
Current language:  auto; currently minimal
(gdb) bt
#0  0x0000000808647dcb in __driDriverGetExtensions_i915 () from /usr/local/lib/dri/i965_dri.so
#1  0x00000008086336ad in __driDriverGetExtensions_i915 () from /usr/local/lib/dri/i965_dri.so
#2  0x000000000042ef7e in R_Clear () at src/client/refresh/r_main.c:872
#3  0x0000000000420121 in SCR_UpdateScreen () at src/client/cl_screen.c:1461
#4  0x000000000041a6d4 in CL_Frame (msec=<value optimized out>) at src/client/cl_main.c:899
#5  0x0000000000469cbe in Qcommon_Frame (msec=<value optimized out>) at src/common/misc.c:388
#6  0x000000000047b493 in main (argc=<value optimized out>, argv=<value optimized out>) at src/backends/unix/main.c:143
(gdb) frame 2
#2  0x000000000042ef7e in R_Clear () at src/client/refresh/r_main.c:872
872                             glClear(GL_DEPTH_BUFFER_BIT | stencilFlags);


This one is iouqake3-devel:
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 804416000 (LWP 100275/ioquake3.x86_64)]
0x0000000832a47dcb in __driDriverGetExtensions_i915 () from /usr/local/lib/dri/i965_dri.so
Current language:  auto; currently minimal
(gdb) bt
#0  0x0000000832a47dcb in __driDriverGetExtensions_i915 () from /usr/local/lib/dri/i965_dri.so
#1  0x0000000832a336ad in __driDriverGetExtensions_i915 () from /usr/local/lib/dri/i965_dri.so
#2  0x000000082de58822 in GLimp_SetMode (mode=-1, fullscreen=qfalse, noborder=qtrue) at code/sdl/sdl_glimp.c:487
#3  0x000000082de5745e in GLimp_StartDriverAndSetMode (mode=-1, fullscreen=qfalse, noborder=qtrue) at code/sdl/sdl_glimp.c:552
#4  0x000000082de570b0 in GLimp_Init () at code/sdl/sdl_glimp.c:776
#5  0x000000082de2ee72 in InitOpenGL () at code/renderergl1/tr_init.c:198
#6  0x000000082de2edc9 in R_Init () at code/renderergl1/tr_init.c:1231
#7  0x000000082de3a0c1 in RE_BeginRegistration (glconfigOut=0xe23cc0) at code/renderergl1/tr_model.c:878
#8  0x00000000004204ce in CL_InitRenderer () at code/client/cl_main.c:3118
#9  0x000000000041b2e4 in CL_StartHunkUsers (rendererOnly=qfalse) at code/client/cl_main.c:3147
#10 0x00000000004451c0 in Com_Init (commandLine=0x7fffffffe4d0 "") at code/qcommon/common.c:2832
#11 0x00000000004e92a2 in main (argc=1, argv=0x7fffffffe950) at code/sys/sys_main.c:746
(gdb) frame 2
#2  0x000000082de58822 in GLimp_SetMode (mode=-1, fullscreen=qfalse, noborder=qtrue) at code/sdl/sdl_glimp.c:487
487                     qglClear( GL_COLOR_BUFFER_BIT );


-- 
A: Because it fouls the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?




More information about the freebsd-x11 mailing list