Raven. A homemade Atari-like computer

[dml]

DSP access on $ffb200 is uncached-imprecise which significantly improves cpu->dsp throughput when bursting some blocks of data (imprecise means it is allowed to use the cpu storebuffer)

Interesting - does that still work out to be safe given the port has 3 sequentially written port bytes? The FIFO transfer is triggered on the low byte so it is important that the other two bytes are written first (but in any order - so long as the low byte is last). If there is a single store buffer then it will always be safe. If it is multiword & unordered, it could transfer the wrong upper-bytes on each low-byte transfer. Worse - the transfer would appear to have worked because the correct number of transfers take place, just not with the correct data sometimes.

I'm probably borrowing trouble here and you figured this already - otherwise it looks very nice

[agranlund]

Interesting - does that still work out to be safe given the port has 3 sequentially written port bytes? The FIFO transfer is triggered on the low byte so it is important that the other two bytes are written first (but in any order - so long as the low byte is last). If there is a single store buffer then it will always be safe. If it is multiword & unordered, it could transfer the wrong upper-bytes on each low-byte transfer. Worse - the transfer would appear to have worked because the correct number of transfers take place, just not with the correct data sometimes.

I'm probably borrowing trouble here and you figured this already - otherwise it looks very nice

The storebuffer is, as far as I understand from the manual, an ordered FIFO of 4 entries, and although the actual writes happen asynchronously they are always in the order you wrote them from the cpu.
Any read or write to any uncached-precise page will flushes the storebuffer so if you need to read some status flag, or just flush and stall for the storebuffer to finish, you'll want to use the regular precise page.

Data validity has so far only been tested by using the method in dsp_ExecProg and seeing that the uploaded program works as expected, so more testing should happen for sure.


And now I'm wondering if I can take advantage of this type of thing for other stuff too..
Writes to gfxcard VRAM perhaps.

[dml]

The storebuffer is, as far as I understand from the manual, an ordered FIFO of 4 longwords, and although the actual writes happen asynchronously they are always in the order you wrote them from the cpu.

Ok my question is fully answered there an ordered fifo. Not a detail I remember, if I even did absorb it (I did not do much on 060 except fix some broken things TBH).

All looking very nice and pretty much absorbs any overhead from address stride on those 3 bytes.

[agranlund]

Pushed a new release package 20260902 (the previous latest is still there saved as version 20260406)
https://github.com/agranlund/raven/releases

To be fair, this is mostly under-the-hood type changes that I did quite a while ago plus the recently added support for revision A2.
So maybe not that much interesting from a user perspective. Most notably, from complexity / risk point of view:

* rvbios is now inside the rom rather than a prg in auto folder
this will have some other future benefits but the immediate one is reducing workload of detecting and/or supporting mismatched versions as rvbios and bootrom are often quite closely coupled.
if it encounters an existing rvbios.prg in the auto folder then that should automatically end up getting deleted
(there will eventually be a sensible way to put an override binary on disk too, which is useful for dev and testing rvbios separately without reflashing)

* some raven bios type settings are now stored in a sector of the flash rom rather than crammed into the few bytes of nvram that TOS is not using.
long term this will allow more flexibility on what kinds of persistent settings can be stored
ideally, you'll want to update the rom flasher firmware to be sure this entire area is kept if using it for updating.

* support for raven.A2
The rom should be able to auto detect board revision and work on both A1 and A2.
But let me know if there is trouble on A1 as I have not ran the very latest on mine.


The release notes:

260902:
rom:
- rvbios is now in rom.
(old c:\auto\rvbios.prg is automatically deleted)
- improved flash detection and init
- ability to flash ROM from bios setup screen
- ability to flash CKBD firmware from bios setup screen
- monitor can be launched from bios setup screen
- TOS terminal can be launched from bios setup screen
- wip pram support in flash
- bios confs now in pram instead of rtc-nvram
- fix emutos hang on boot if ikbd fifo contain stale data
- autodetect board revision A1 or A2
- (A2) YM volume control support
- (A2) Internal speaker support
- (A2) DSP support (wip)

firmware / hardware
- Rom programmer: support rom images with 16kb pram area (was 4kb)
- Added Rev.A2 firmwares
- Added Rev.A2 gerbers and bom

software:
- pgusinit: v3.9.0c for firmware v4.1.1 and above
- rvsnd: fixed midi xbios crash
- uiptool: fixed random ftp server bug
- (A2) dsp: compilers and some tools

[luciodra]

I had missed this update. There are a lot of interesting features.

[kodak80]

Just gonna put this here as it would be interesting to see if it works in the Raven...

Video, sound and networking a single ISA card.

https://pcmidi.eu/sonicvisionpro.html

SonicVision Pro is now available for pre-orders...

svpro-proto2.jpg (347.82 KiB) Viewed 97 times

[viking272]

A great idea, 370 euros though

[mikro]

I guess the graphics drivers would need to be written on Atari, too. Networking shouldn't be a problem and OPL3 perhaps neither (thanks to @nokturnal's library).

[agranlund]

I guess the graphics drivers would need to be written on Atari, too. Networking shouldn't be a problem and OPL3 perhaps neither (thanks to @nokturnal's library).

Cirrus GD5434 is fully supported with hw acceleration with the rvnova svga driver

[Atarian Computing]

A great idea, 370 euros though

That's what a GD5434 ISA card can fetch on eBay on its own. Here you get a new card, sound card + nic for the same price.

If the Raven had only a limited slots, this would be a no-brainer. If this had interleaved memory, I might get it down the road...