On 09/03/2013 09:53, Helmut Karlowski wrote:
- scrp_write() calls Fsfirst() with trap #1As MiNT hooks into Trap 1 the MiNT-Fsfirst should be called by EmuTOS in theory. Or EmuTOS did not notice the redirection and things go wrong somehow?
Indeed, EmuTOS scrp_write() uses FreeMiNT's Fsfirst(). This brings me on a serious trail.Whenever the EmuTOS AES is entered with trap #2, it switches to a private process stack. Currently that stack is 1788 bytes. If I increase it to 2000 bytes, it does not crash anymore.
I suspect that FreeMiNT's Fsfirst() use much more stack than the EmuTOS one, so the private stack overflows.
I will add more traces to be sure of what really happens. -- Vincent Rivière