On Tue, 20 Jan 2026, Chad Jablonski wrote:
I plan to validate the above Radeon behavior in the future but I don't
have the best test environment set up for that card at the moment.
Zoltan if you've seen the dst_x behavior is required then we can modify
this but otherwise it felt safe to me to follow the docs for now.
Just an update on Radeon (R100) behavior. I've confirmed in my testing that it
also does _not_ update dst_x and dst_y after vram and host data sourced blits.
This contradicts the m6 register guide.
Interesting. I think I've added the update because I've found that some
guest needed that but I don't remember which. I mostly tested with MorphOS
and Linux. But it could be I've seen it in the docs and that's why I
implemented it, I don't remember. If it does not break MorphOS it could be
removed to match what hardware seem to do. I think most drivers will set
these on every blit anyway.
Regards,
BALATON Zoltan
