Hi!
Please prevent me from getting crazy:
I have (pretty large) a rt-module that accesses some cpp-class members. The
class is "simple" in the sense that it uses no virtual members ansd so on,
just ordinary data-types and some simple instanciated templates that are
based on fixed length c-arrays.
What drives me crazy is that after some simple extensions the system doesn't
work any more! The whole machine just reboots. For example:
------- snip ----------
int Parse(unsigned int fifo)
{
int msg, nReturn = OK, laenge, error, nBewegung_ok, nZustand;
int nAchse, nAnzahl, nAusgang, nArt, nFall, nFunktion,
Vektor<int, 6> nvInkremente;
Vektor<double,3> Sollkraftrichtung;
// Without this array it works, otherwise it crashes! The array is not used
// elsewere"
int testarray[1000000];
if ((laenge = rtf_get(fifo, &msg, sizeof(msg))) == sizeof(msg))
{
error = FALSE;
nReturn = OK;
switch (msg)
(...)
---- snap -----
After three weeks I'm pretty shure that there is no null-pointer thing or
something that could create a "seg fault".
Sometimes (really spare), when I decrease the size of the array, the first
machine gives me a page fault in a kernel oops and hangs a few minutes later!
Without the array, everything works fine!!!
The interesting this is: A Celeron based machine is much less sensitive to
the crashes than a PIII! Sometimes, with the array shown above, a Celeron
runs for 5 minutes bevor crashing, while the PIII reboots even right away!
Is there anybody out there that has an idea about what could be the problem?
Is there something like a maximum size of a module? Do C++ class-related
problems occure above a certain module size? I'm helpless....
Thanks for every help
Maik
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
--
For more information on Real-Time Linux see:
http://www.rtlinux.org/
