| Re: [hatari-devel] Inconsistent behavior during Hatari startup and reset |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/hatari-devel Archives
]
Hi,
On 02/25/2018 09:45 PM, Uwe Seimet wrote:
Sorry, I forgot to mention that I mapped my GEMDOS drives to host paths,
i.e. the AUTO folder is on the host.
With the current sources I am observing some strange behavior.
1. When starting a program in the AUTO folder Hatari reports a
segmentation fault in SDLGui_CenterDig.
Why Hatari is presenting dialog at this point, is AUTO folder
program causing a double bus error (=emu halt + dialog)?
When I press F12 before the
segmentation fault occurs, and then just click on OK, as soon as the fault
occurs I am offered to enter the debugger. Looks as if it requires to
press F12 in order for SDL to be properly initialized.
SDL1 or SDL2?
2. When I add an ILLEGAL instruction to my code, which is executed in
the AUTO folder, Hatari reports on the console that there is an illegal
instruction, but simply continues with the next opcode.
If you expect to get to debugger, you need to tell Hatari that you
want debugger to be invoked at that point.
It's a bit annoying to do during TOS bootup, because TOS itself does
illegal exceptions during bootup, and Hatari's --debug-except autostart
option enables exception catching only after TOS has read INF file and
pexec()ed the autostart program (i.e. after AUTO programs are already
run).
Shouldn't there have been 4 bombs in this case?
TOS does some illegal opcodes itself during startup, to detect HW.
I guess TOS is ignoring all illegal instruction exceptions until it
has booted fully up?
- Eero