Re: [hatari-devel] Fixing some STE sound issue and doing 2.4.1 release

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

On 29.7.2022 17.33, Nicolas Pomarède wrote:
> I'm not familiar with Screen_BitplaneToChunky32(), can't tell if it's a 
> false positive due to the various casts used to convert screen.

It does not give error if VDI screen size is small enough.

And it does not happen with Videl code (used for Falcon & TT).

With "--vdi-planes 4", 640x200 is enough to trigger it, with 2 planes 
640x400, and with mode, 640x736.

It's strange though that if they are heap overwrites, why Valgrind does 
not report them.  And why even AddressSanitizer does not always report 
anything for those VDI sizes, but instead crashes at Hatari exit?

(My AddressSanitizer crashes at Hatari exit could be related to use of 
VDI mode.)

Maybe heap allocs go then to a different (non-neighbouring) addresses, 
add corrupt just internal malloc book-keeping?


=> If this is real heap corruption, this is something I think should be 
fixed before release, as it could explain all kinds of weird behavior 
[1], and VDI mode might not be only case where overwrites happen.


	- Eero

[1] MMU issue, those funny Atari-forum reports from users where setting 
Hatari options did not take effect...