Re: [hatari-devel] Hatari segfault on cold boot with MMU + EmuTOS

[Nicolas Pomarède <npomarede@xxxxxxxxxxxx>]

Le 09/11/2018 à 23:45, Eero Tamminen a écrit :

> Great that you found it, memory corruptions are nasty to debug.

note that in that case I found that it was useful to use gdb's "step" 
feature combined with "display" to show the location or stack after each 
step. then add a "break" at the place that handles the quit key in 
Hatari, then run "step 1000" for example and reapeat until you get a 
crash. Luckily you should see better where Hatari crashed before the 
complete stack corruption.


> Why WinUAE CPU core uses variable with name "quit_program" for
> indicating emulated machine reset?

that's their choice :) This variable is really used to quit when set to 
UAE_QUIT, I think it evolved over time and also included RESET, which 
has a similar behaviour, ie exit the main cpu emulation loop and go back 
to the upper level that checks for cpu parameter changes and so on.

Nicolas