Loci - my Oric storage emulation project

[Sodiumlightbaby]

Bisect C mia_action PIO program timing revert.

[Dbug]

First boot gots all corrupted

Second boot displayed some 44444 with the usual fat font...

...and after rebooting I got the same one but with a thin font?

[Sodiumlightbaby]

Thanks. Just to be sure we're testing one thing at the time, could you revert the timing tests A and B? Looks like they are still present.

[Dbug]

Thanks. Just to be sure we're testing one thing at the time, could you revert the timing tests A and B? Looks like they are still present.

I've set the values back to what I had on 0.2.5, not sure if that's the default, is there a "reset to default" command?

[Sodiumlightbaby]

They work with 0.2.5 and that's the important part.

[Sodiumlightbaby]

Bisect D, reduced action loop (ACIA regs disabled).
Not confident this is enought to fix the issue, but I'm interested in if there is different behaviour compared to 0.3.1

[Dbug]

Can't say there's much difference.

When the Oric is cold, it's fully glitching:

When the machine is warm the boot menu works as expecting... for about 2 seconds before freezing

[Sodiumlightbaby]

This is very interesting. The last image you have actually shows the monitor system that was added in 0.3.1 kicking in, which requires a break to be sent to the uart, indicating perhaps a spurious writhe 0x03A1.

With the Bisect D firmware, could you try an long-press on the action button to activate the diag ROM? It should be fixed in that fw, and I'm curious if that is also unstable - in only uses the ROM emulation and none of the page 3 interface emulation.

[Dbug]

Long press seems to work nicely:

At this point if I short press the button, I see

Code: Select all

IRQ Trap on

]?Unable to open file (4)
Booting ROM from firmware

And that's it, the screen stays on the Oric Diag ROM screen.

After that I have to power off, power on, and after the short press I'm back to the boot screen that spins for a couple seconds and stops on

[Sodiumlightbaby]

That is very very helpful. So the issues are connected with IO page accesses creating problems for the ROM emulation. ROM emulation alone seems to work fine.

[iss]

@Dbug: What Oric are you using in LOCi tests (Oric-1, Atmos)? What's the ROM inside? How long is the ribbon cable (>15 cm)?

[Dbug]

Atmos, the cable is about 10cm, and it's an actual 1.1 ROM, not an eprom.

[iss]

Looks good to me and it should work!

[Sodiumlightbaby]

Bisect E, revert unguarding of read DMA trigger in the action loop. This bisect is based on the observations so far that the machine handles ROM emulation ok when no IO operations are happening.
If this has an effect, the loading and execution of the LOCI ROM should be much better but likely some hangs and issues will still trigger.

[Dbug]

On the very first boot, I got some 22222

On the second boot, it spinned a bit longer than before, but ended up frozen with a "a" on the left side of the screen.

Not sure if relevant, but during one of the cold boots the device was entirely black: No red LED, and the screen was completely blank, I had to power cycle the machine to get it back on.

Also, even when the Oric is frozen, doing a long press will properly show the "Booting diag ROM" message on the screen, except nothing happens on the Oric side.

EDIT: After letting the machine plugged for almost an hour more, the next time the spinner was spinning for almost 5 seconds before freezing.