Le 15/09/2014 14:34, Bruce Dubbs a écrit :
Pierre Labastie wrote:
Le 15/09/2014 08:54, Bruce Dubbs a écrit :
Christopher Gregory wrote:
On Mon, September 15, 2014 2:32 pm, Bruce Dubbs wrote:
In a loosely related (or maybe unrelated) issue, I have a bit of a
problem testing with glxgears. I have an Intel i915 based video
card. When
I use the Xorg Intel Driver-2.99.916, the above window managers
(twm, openbox, and xfce) come up fine, but glxgears gives a garbled
screen.
Removing the configuration (and renaming the driver) for the Intel
driver brings up the screen using the fbdev driver and glxgears
displays
properly.
Does anyone know of anything specific needed when using the Intel
driver
to get glxgears to render properly?
I am running lxde right now, though it is on a systemd system.
After a
few minor glitches with the taskbar not appearing due to the package
upgrades upstream it is working perfectly.
As far as the intel driver is concearned, I too am using it on a i915
base
chipset and for me, glxgears works fine.
Are you using the Intel Driver-2.99.916? Glamor acceleration?
Experimenting a bit, I now have xorg.conf with:
Section "Module"
Load "dri2"
Load "glamoregl"
EndSection
Section "Device"
Identifier "intel"
Driver "intel"
# Option "AccelMethod" "glamor"
EndSection
With that, the Intel driver works and glxgears works. If I uncomment
the AccelMethod, I get a segfault when starting Xorg.
Of course we are now enabling glamor in the Xorg server, so I don't
know if the newest driver is consistent with that.
Xorg.0.log is giving me the following with "AccelMethod" commented out:
[ 34956.227] (II) LoadModule: "dri2"
[ 34956.228] (II) Module "dri2" already built-in
[ 34956.228] (II) LoadModule: "glamoregl"
[ 34956.228] (II) Loading /opt/xorg/lib/xorg/modules/libglamoregl.so
[ 34956.233] (II) Module glamoregl: vendor="X.Org Foundation"
[ 34956.233] compiled for 1.16.0, module version = 1.0.0
[ 34956.233] ABI class: X.Org ANSI C Emulation, version 0.4
What method does it fall back to when commenting out AccelMethod? It is
lower in the log (in the "AIGL" lines).
I have:
[ 34956.260] (II) AIGLX: enabled GLX_MESA_copy_sub_buffer
[ 34956.260] (II) AIGLX: enabled GLX_ARB_create_context
[ 34956.260] (II) AIGLX: enabled GLX_ARB_create_context_profile
[ 34956.260] (II) AIGLX: enabled GLX_EXT_create_context_es2_profile
[ 34956.260] (II) AIGLX: enabled GLX_INTEL_swap_event
[ 34956.260] (II) AIGLX: enabled GLX_SGI_swap_control and
GLX_MESA_swap_control
[ 34956.260] (II) AIGLX: enabled GLX_EXT_framebuffer_sRGB
[ 34956.260] (II) AIGLX: enabled GLX_ARB_fbconfig_float
[ 34956.260] (II) AIGLX: GLX_EXT_texture_from_pixmap backed by buffer
objects
[ 34956.260] (II) AIGLX: enabled GLX_ARB_create_context_robustness
[ 34956.260] (II) AIGLX: Loaded and initialized i965
The entire log is at http://hastebin.com/zumitijeve.coffee
OK, the AccelMethod is "sna" (as far as I understand from doc, there are
3 of them,
"uxa", "sna", "glamor"):
|[34956.240](II) intel(0): SNA initializedwith Haswell(gen7.5, gt2) backend
Do I understand correctly that this setting works, and that glxgears runs OK?
Well: dri3+sna
is not bad!
I do not have an intel GPU, so I may not be very helpful. I have just played a
lot with
llvm+mesa+xorg-server to test the nouveau driver, so here are a few (maybe
relevant) hints
below:
It may be a pain to test, but does it still segfault with glibc 2.19 (rebuild
llvm, mesa,
xorg-server and of course the driver at least)? There has been a bunch of
non standard headers included in glibc-2.20, which contradict some definitions
in
xorg-server.h (actually, there is a workaround on xorg-server page to include
xorg-server.h
first in glamor.h)
Also, it may be worthwhile to try to downgrade llvm and see whether you still
get the fault.
Regards
Pierre
|
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
