Hi Everyone.
I started tweaking my 68K app with some ARM code and as a first example I
took BIG Pno test from CodeWarrior 9.3 update.
I see this example shows potential problem with Tungsten T/T3 devices and
there is a note that it solves them. However I can't find
in the code how was it accomplished.
The code worked fine on T3. Then I changed the code a bit to run not a small
inline routine but a huge Run() routine that has tons of
this inline funcs. The result was a fatal crash.
Can anyone please explain me that crash? I do not have it when I created the
same thing in 5.2 simulator.
This is a code from Big PNO test ARM part. Palm part is fine.
Also I think it will be super if anyone can explain me how does this sample
solve code caching flush problems? (With volatile??)
Best regards,
Platon
inline void RunCore()
{
// draw randomly-colored pixels on the screen
for (Coord y = 16; y < 160; ++y)
{
for (Coord x = 0; x < 160; ++x)
{
RGBColorType RGB, prevRGB;
RGB.index = 0;
RGB.r = (UInt8)SysRandom(0);
RGB.g = (UInt8)x;
RGB.b = (UInt8)y;
WinSetForeColorRGB(&RGB, &prevRGB);
WinDrawPixel(x, y);
}
}
}
static unsigned long Run()
{
RunCore(); RunCore(); RunCore(); RunCore(); RunCore();....
................. about 200 or 300 times
return 0;
}
// ------------------------
unsigned long PNO_Main(
const void *emulStateP,
void *userData68KP,
Call68KFuncType *call68KFuncP)
{
InitPACEInterface(emulStateP, call68KFuncP);
volatile int dontRun = 1; // I changed this to zero
^^^^^^^^^^^^^^^^^^
if (!dontRun)
{
return Run();
}
else
{
RunCore();
return errNone;
}
}
--
For information on using the Palm Developer Forums, or to unsubscribe, please
see http://www.palmos.com/dev/support/forums/
