1 bit RGB😲
1/3 bit R, 1/3 bit G and 1/3 bit B, or 1/4 bit Alpha, 1/4 bit R, 1/4 bit G and 1/4 bit B? 🤣

I imagine you're running the example program on a 32-bit truecolor Workbench screen, which means your friend BitMap is also 32-bit truecolor. But the backBuffer BitMap you're allocating has a depth of only one bit. So the question is, what happens when you ask for a one-bit color-mapped BitMap, but pass a 32-bit truecolor friend BitMap?

Back in the OS 3.1 days even graphics cards typically supported planar displays when the depth was less than eight. So for compatibility with programs from that era, you may well be getting a one-plane planar BitMap from AllocBitMap(). Of course that's going to require planar to chunky conversion when blitted to a 32-bit BitMap, which is going to be slow.

The ring buffer is likely the first place where you'd hit problems with lack of cache coherency. However, if you're using an SI card or newer, then the problem could also be that their drivers can't handle big-endian.

IIRC, ACube did have a Linux driver that worked with GART enabled. I don't know for which cards, or how it worked.

I can think of one way to get GART working, and that's to mark all memory used for GART as non-cacheable (or disable the data cache entirely). The L2 cache might need to be disabled too. Needless to say, doing so would come with a serious performance penalty.

It's a Sam460 limitation. The Sam460 doesn't have cache coherency. I tried doing manual cache flushing everywhere where it's needed, but couldn't get it working. It would always lock up.

I tried use the reverse load and store instructions I was disappointed with results. If remember correct, I got as poor results as bit mask bits, then shift & Or result code. Yeh it might be I don’t have the instruction, or maybe some trap instruction interrupt bug, some forgot check what CPU has or has not the instructions, not sure how that works.

Using an endian-swapping version of GCC might help, like the one which was used for building native x86/x64 Amithlon software.

Hehe. I've seriously thought about designing my own GPU. I'd never be able to beat the big guys in raw performance, but could design something that's a lot easier to use (i.e., simple enough you could easily bit-bang it without a driver just like the old Amiga chipsets). However, I'd need a large enough niche market to make it financially viable.

I don't mind not seeing Uboot, there us not much to do after it's working. I actually would not mind to have an option not to see it at all, I guess Rx 550 should fix that 😁.

Looking through the adtools code I have seen it many times before and just never thought about it. In the code, the address of the Exec library is sought from address 0x4 - just like on classic machines.