Niels Roest wrote:
> René Pavlík wrote:
>> Hi,
>>
>> I have also difficulties running multiple applications. Can you give
>> me a hint
>> what not to forget or point me to a working demo?
>>
>> I compiled DirectFB in Multi-process core and every time I wanted to
>> run the
>> second application (which actually run as single), it segfault'ed,
>> ended, or
>> anything else but not worked. Therefore I tried also the SaWMan and
>> again: I was
>> able to run the daemon (application manager) and then I wanted to
>> execute those
>> 2 given example slave apps and the first ran ok, but the second one ended
>> immediately.
>>
>> I'm using the SH7723 gfx driver.
>>
>>
> Well, a segfault can be caused by any number of reasons.
> It sounds like your 2 slave apps do not like each other.
I also received SIGBUS very often. What part of DirectFB can cause kernel to
invoke this signal? It must be an unaligned memory access. And I'm also warned
about it:
(!!!) *** WARNING [no alignment specified yet?] *** [gfxcard.c:2692 in
dfb_gfxcard_reserve_memory()]
(*) SH7722/LCD: Allocated 800x480 RGB16 Buffer (768000 bytes) at 0x0ff44800
(0x2a22e800)
(*) DirectFB/Graphics: Renesas SH7723 0.9 (Denis & Janine Kropp)
(*) DirectFB/Core/WM: Default 0.3 (directfb.org)
(!!!) *** WARNING [workaround] *** [surfacemanager.c:194 in
dfb_surfacemanager_allocate()]
> If you for instance try df_window, this will run twice (or any number of
> times) without a problem, does that work for you?
Finally I got it working. Now, when I run multiple df_window apps, I have one
cursor controlling all the windows from different processes.
df_andi caused segfault when about to connect to a master app which was using
the primary surface directly without any layer beneath it. Is the usage of
layers necessary for it to work? Or what is the key requirement?
Thanks a lot,
Rene Pavlik
_______________________________________________
directfb-users mailing list
directfb-users@directfb.org
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
