On Wed, Jan 24, 2007 at 06:14:14PM +0200, [EMAIL PROTECTED] wrote:
> Hi all,
>
> Relevant background
> --------------------------
>
> I want to use DFB on a MIPS simulator I am working on. I am using (actually a
> simulator of-, something akin to QEMU) an embedded MIPS24K processor and a
> 2.6.10 kernel with a working framebuffer device (that is, fbcon works for
> instance).
>
<snip>
>
> I was unable to make this work (telling the configure script to install to
> /mnt/usr/local makes the dynamic linking to be to /mnt/use/local/lib during
> run
> time, where it should link to /usr/local/lib and this is just one example of
> why installing to a non-defulat directory doesn't work). Should this be
> possible after all? / anyone managed to do this? please explain how.
'configure --prefix=/usr/local' + 'make DESTDIR=/mnt install' should do
the trick.
> The problem
> ---------------
>
> This seems to compile and install properly but when I try to run any DFB
> program
> (dfbinfo, df_dok, etc') I get the following:
>
> # dfbinfo
>
> =======================| DirectFB 1.0.0-rc2 |=======================
> (c) 2001-2006 United Cultures of Earth - go for outer space!
> (c) 2000-2004 Convergence (integrated media) GmbH
> ----------------------------------------------------------------
>
> (*) DirectFB/Core: Single Application Core. (2007-01-23 07:00)
> fb_ioctl - 1
> fb_ioctl - 2
> (!) [ 206: 0.000] --> Caught signal 1 (sent by pid 211, uid 0) <--
> Killed
>
> Using strace I saw that the problem is in VT code, specifically when doing a
> certain ioctl on tty0: (here the strace output)
> fb_ioctl - 1
> ioctl(3, 0x460f, 0x7fff7740) = 0
> fstat(3, {st_mode=S_IFCHR|0644, st_rdev=makedev(29, 0), ...}) = 0
> fb_ioctl - 2
> ioctl(3, 0x4610, 0x7fff77d8) = 0
> ioctl(4, 0x5606, 0x1) = 0
> ioctl(4, 0x5607, 0x1) = 0
> nanosleep({0, 40000000}, NULL) = 0
> open("/dev/tty1", O_RDWR) = 5
> write(5, "\33[?1;0;0c\0", 10) = 10
> ioctl(5, 0x4b3a, 0x1) = 0
> ioctl(4, 0x5471) = -1 ENOTTY (Inappropriate ioctl for
> device)
> ioctl(5, 0x5480) = 0
> rt_sigaction(SIGRT_9, {SIG_DFL}, {SIG_DFL}, 16) = 0
> rt_sigaction(SIGRT_10, {SIG_DFL}, {SIG_DFL}, 16) = 0
> ioctl(5, 0x5602, 0x7fff78a8) = 0
> old_mmap(NULL, 8388608, PROT_READ|PROT_WRITE|PROT_EXEC,
> MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2af47000
> mprotect(0x2af47000, 4096, PROT_NONE) = 0
> clone(child_stack=0x2b7466d0,
> flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|CLONE_SYSVSEM|CLONE_SETTLS|CLONE_PARENT_SETTID|CLONE_CHILD_CLEARTID|CLONE_DETACHED,
> parent_tidptr=0x2b746b78, tls=0x2b74df80, child_tidptr=0x2b746b78) = 223
> sched_yield(
.
> (!) [ 223: 0.000] --> Caught signal 1 (sent by pid 223, uid 0) <--
> ) = 0
> +++ killed by SIGKILL +++
>
> I forced the code not to initialize VT as if config->vt was "false", but I
> have
> no idea how this will affect DFB.
You could have simply used the no-vt directfbrc option :)
> Another problem I have is that afterwards, when initializing the input
> drivers,
> I get:
>
> # /usr/local/bin/dfbdump
<snip>
>
> (*) Direct/Thread: Running 'PS/2 Input' (INPUT, 272)...
> (*) DirectFB/Input: IMPS/2 Mouse 1.0 (directfb.org)
> (!) [ 269: 0.000] --> Caught signal 11 (at 0x4, invalid address) <--
>
<snip>
>
> (I have no joystick connected, nor do I have kernel support for one, nor do I
> want one. Id prefer if were possible to work without input at all at least for
> starters)
You can disable all input modules during configure or at run-time (see
disable-module in directfbrc(5)).
--
Ville Syrjälä
[EMAIL PROTECTED]
http://www.sci.fi/~syrjala/
_______________________________________________
directfb-users mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
