Hi there,
I'm trying to get directfb running on my ubuntu 7.10 system on a T61
with a custom kernel 2.6.26-rc6 which contains support for the Intel
Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 0c)
in the intelfb module.
My kernel options at boot time are:
video=intelfb:accel=1,mtrr,vram=128,fixed,hwcursor=0 vga=0x318
I decided to start with 1024x768, though the T61 supports higher
resolutions of course.
fbset -i says:
mode "1024x768-60"
# D: 62.360 MHz, H: 47.530 kHz, V: 60.013 Hz
geometry 1024 768 1024 768 32
timings 16036 128 32 16 4 128 4
accel true
rgba 8/16,8/8,8/0,8/24
endmode
Frame buffer device information:
Name : Intel(R) 965GM
Address : 0xe0000000
Size : 8122368
Type : PACKED PIXELS
Visual : TRUECOLOR
XPanStep : 8
YPanStep : 1
YWrapStep : 0
LineLength : 4096
MMIO Address: 0xf8100000
MMIO Size : 524288
Accelerator : Intel 830M/845G/85x/865G
I had to remove intelfb from /etc/modprobe.d/blacklist-framebuffer and
to recreate the initramfs of the custom kernel after editing
/etc/initramfs-tools/modules as follows to get /dev/fb0 created at all
(though this is Ubuntu related of course):
fbcon
intel_agp
The fbcon is not a great experience though, the cursor isn't working
and in general this particular intelfb feels painfully slow.
Anyway, as I understand your getting started info in the wiki, a
working framebuffer should be a good starting point to get directfb
running. I fetched a copy from your git repository with the following
git tip:
commit eec017753d1e32ccf6c6f1ecebfdaeca9ee0b8cb
Author: Denis Oliver Kropp <[EMAIL PROTECTED](none)>
Date: Mon Jun 16 09:19:01 2008 +0200
[davinci] Update allocations during Lock().
I build this version with --enable-debug and installed it in
/usr/local as the default expects. I followed the directfbrc(5) man
page further and created the following /etc/directfbrc
system=fbdev
fbdev=/dev/fb0
mode=1024x768-60
depth=24
mouse-protocol=IMP/2
bg-color=FF00FF
no-vt
no-hardware
When I try to run dfbinfo however, I end up with a black screen --
note, if I have no X session running in a different vt I have to
reboot the box to see anything again on my screen (which I consider a
serious flaw, there should be some way to restore the console),
otherwise a vt switch to X and back provides me the following error
log of dfbinfo:
~~~~~~~~~~~~~~~~~~~~~~~~| DirectFB 1.2.0-rc1 |~~~~~~~~~~~~~~~~~~~~~~~~
(c) 2001-2008 The world wide DirectFB Open Source Community
(c) 2000-2004 Convergence (integrated media) GmbH
----------------------------------------------------------------
(*) DirectFB/Core: Single Application Core. (2008-06-18 09:57) [ DEBUG ]
(*) Direct/Thread: Started 'VT Switcher' (1514) [CRITICAL OTHER/OTHER
0/0] <8388608>...
(!) DirectFB/FBDev: Could not disable console acceleration!
--> Invalid argument
(!) DirectFB/Core: Could not initialize 'system_core' core!
--> Initialization error!
(!) [Main Thread 0.000] ( 1512) *** Assertion [(data) != NULL]
failed *** [input.c:535 in dfb_input_core_suspend()]
(!) [ 1512: 0.000] --> Caught signal 5 (unknown origin) <--
(!) [Main Thread 0.000] ( 1512) *** Assertion [(data) != NULL]
failed *** [input.c:535 in dfb_input_core_suspend()]
(!) [ 1512: 0.000] --> Caught signal 5 (unknown origin) <--
The backtrace of the coredump looks as follows:
#0 0xffffe430 in __kernel_vsyscall ()
#1 0xb7e87590 in raise () from /lib/tls/i686/cmov/libpthread.so.0
#2 0xb7ea17f1 in signal_handler (num=5, info=0xbfd9a45c,
foo=0xbfd9a4dc) at signals.c:411
#3 <signal handler called>
#4 0xffffe430 in __kernel_vsyscall ()
#5 0xb7e87590 in raise () from /lib/tls/i686/cmov/libpthread.so.0
#6 0xb7e9adaa in trap (domain=0xb7eab8f9 "Assertion") at debug.c:337
#7 0xb7f29700 in dfb_input_core_suspend (data=0x0) at input.c:535
#8 0xb7f1599a in dfb_core_shutdown (core=0x80530b8, emergency=<value
optimized out>) at core.c:914
#9 0xb7f15c5f in dfb_core_arena_shutdown (arena=0x0, ctx=0x80530b8,
emergency=true) at core.c:1094
#10 0xb7eb4a4d in fusion_arena_exit (arena=0x0, shutdown=0xb7f15b50
<dfb_core_arena_shutdown>, leave=0, ctx=0x80530b8, emergency=true,
ret_error=0x0) at arena.c:556
#11 0xb7f15401 in dfb_core_destroy (core=0x80530b8, emergency=true) at
core.c:433
#12 0xb7f1563c in dfb_core_signal_handler (num=5, addr=0x0,
ctx=0x80530b8) at core.c:895
#13 0xb7ea1660 in signal_handler (num=5, info=0xbfd9a9ac,
foo=0xbfd9aa2c) at signals.c:383
#14 <signal handler called>
#15 0xffffe430 in __kernel_vsyscall ()
#16 0xb7e87590 in raise () from /lib/tls/i686/cmov/libpthread.so.0
#17 0xb7e9adaa in trap (domain=0xb7eab8f9 "Assertion") at debug.c:337
#18 0xb7f29700 in dfb_input_core_suspend (data=0x0) at input.c:535
#19 0xb7f1599a in dfb_core_shutdown (core=0x80530b8, emergency=<value
optimized out>) at core.c:914
#20 0xb7f1674c in dfb_core_initialize (core=0x80530b8) at core.c:976
#21 0xb7f1694b in dfb_core_arena_initialize (arena=0x0, ctx=0x80530b8)
at core.c:1052
#22 0xb7eb4b81 in fusion_arena_enter (world=0x8052e28, name=0xb7f5d6d4
"DirectFB/Core", initialize=0xb7f167c0 <dfb_core_arena_initialize>,
join=0xb7f16460 <dfb_core_arena_join>, ctx=0x80530b8,
ret_arena=0x80530c8, ret_error=0xbfd9aeb8) at arena.c:509
#23 0xb7f16103 in dfb_core_create (ret_core=0xbfd9aef8) at core.c:337
#24 0xb7eccce9 in DirectFBCreate (interface=0x804b464) at directfb.c:193
#25 0x0804888f in main (argc=1, argv=0xbfd9afc4) at dfbinfo.c:92
I get similiar behavior if I'm using the directfb packages from
ubuntu, which are the following:
ii libcairo-directfb2 1.4.10-1ubuntu4.4
The Cairo 2D vector graphics library DirectF
ii libcairo-directfb2-dev 1.4.10-1ubuntu4.4
Development files for Cairo graphics library
ii libdirectfb-0.9-25 0.9.25.1-6ubuntu3
direct frame buffer graphics - shared librar
ii libdirectfb-bin 0.9.25.1-6ubuntu3
direct frame buffer graphics - binaries
ii libdirectfb-dev 0.9.25.1-6ubuntu3
direct frame buffer graphics library - devel
ii libdirectfb-extra 0.9.25.1-6ubuntu3
direct frame buffer graphics - extra provide
ii libgtk-directfb-2.0-0 2.12.0-1ubuntu3
The GTK+ graphical user interface library -
ii libgtk-directfb-2.0-dev 2.12.0-1ubuntu3
Development files for the GTK+ library - Dir
So my final question, how to proceed -- what's going wrong here?
Kind regards,
--Anselm
_______________________________________________
directfb-users mailing list
directfb-users@directfb.org
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
