Hi,

so, I made some new attempts, and here are my results with the Matrox
driver.

Indeed, the monitor error was simply a non-working mode (I used the SVGA
monitor). Doing:
 export GGI_DEFMODE="1024x768x16"
(for example, "800x600x16" works too) solved most problems with fb setup.

So, yes, I definitely run libggi graphical apps. "flying_ggis" are
currently popping on my (secondary head) G200-powered monitor while I'm
writing this on the (first head) Mystique-powered monitor in VGA console.
Note that it's necessary to swith off the light to actually see the
"./stars" at 1024x768 (what realism! :-)).

There are still some problems surrounding kii: I can start "monitest" or
"demo" (the first shows the "mire", the other only the start message), but
I cannot provide input to them. I get an indication on VT0 console while
switching back to the first head, saying that "event_device 79 unmapped".
 On the contrary I can actually 'q'uit e.g. "stars" with 'q', but in this
case, the kernel speaks of "event_device 15". In order to quit,
e.g. "stars", I must not play with ALT-Fx or ALTGR-Fx (once I lose the
input focus, I cannot recover it with ALT/ALTGR). I have to use CTRL-C.
 I don't yet understand why "monitest" or "demo" get mapped to such a high
event device number (apparently incorrect, at least I don't know how to go
to that graphical console). But maybe it is an error in these programs.
 I don't understand yet why I fail to switch back the focus to all
programs. Maybe I'm doing the bad ALT/ALTGR-Fx magic?
 I can CTRL-C "monitest" or "demo" from the original console that started
them. Then the G200 is usable for another program. However, if I kill
"monitest" or "rmmod" the driver, the head is left unusable (apparently,
the kernel thinks the [event?] device is still open, even another insmod
does not change that belief).

Starting "textdemo" locked that program (unable to stop it via CTRL-C), so
it locked my driver and I had to reboot. (And there were fsck errors at
that boot.)
 "showaccel2" had no input either and also caused a kernel oops (assert
failure). But it was the last in the serie of test so it may need further
investigation (later...)

So:
 - runnings programs: "flying_ggis", "stars", "slimy", "showaccel"
 - programs running but with no input: "monitest", "demo".
 - programs not running: "textdemo"
 - no input and a kernel oops: "showaccel2"

I'm going to commit my changes to the Matrox driver in kgi-wip today. I
think it's good enough for testing.

Rodolphe

PS: Note that "flying_ggis" on the G200 and simultaneously "stars" on the
Mystique via fbdev runs too.






Reply via email to