A2000/030 Level 7 Interrupt

[DanScott]

We're trying to fix up our "The Fall" demo to work on all systems. This includes testing on as many real Amigas as we can.

Currently on an A2000/030 there's a Level 7 Interrupt being triggered that is being caught by our demo "kernel".

Does anyone know anything that could trigger a Level 7 on this hardware configuration?

[alpine9000]

Originally Posted by DanScott

We're trying to fix up our "The Fall" demo to work on all systems. This includes testing on as many real Amigas as we can.

Currently on an A2000/030 there's a Level 7 Interrupt being triggered that is being caught by our demo "kernel".

Does anyone know anything that could trigger a Level 7 on this hardware configuration?

Sorry I can't help you, I tried going down the path of handling all the various random interrupts that can be triggered by all kinds of things from network cards, expansion cards etc and in the end the only 100% reliable solution was to keep the OS running in the background and let the OS deal with it.

[Toni Wilen]

Unfortunately some hardware generate spurious interrupts. I remember seeing some setpatch/KS(?) update note that mentions NMI vector being changed to simple RTE due to these glitches. I have also seen it with ACA500+ and some 1200 accelerator combinations.

I don't think it is restricted to NMI but other spurious interrupts usually don't matter because most interrupt handlers check INTREQ which is always valid.

[DanScott]

Thanks Toni

Actually, we have had a spurious level 5 interrupt generated on another machine too.

We will patch all unwanted interrupt vectors to point to a simple RTE, this should be a workable solution.

Is $1f for Level 7 across all 680x0 processors, or is it specific to 68000 ? (that's a question from Leonard, who coded the demo system kernel)

[StingRay]

Originally Posted by DanScott

We will patch all unwanted interrupt vectors to point to a simple RTE, this should be a workable solution.

That should do the trick.

Originally Posted by DanScott

Is $1f for Level 7 across all 680x0 processors, or is it specific to 68000 ? (that's a question from Leonard, who coded the demo system kernel)

Level 7 vector is $7c.w which is valid for all 680x0 processors as long as the VBR is not moved.

[alpine9000]

So probably not an issue for a demo system, but something I ran into was for example I have a keyboard interrupt enabled and a network card is trying to interrupt on the same line, yet you can never acknowledge the network interrupt so it just bombards the system with interrupts and wastes CPU. Is there a non system friendly way of dealing with this or is just ignoring the interrupt the only way to deal with it?

[DanScott]

Originally Posted by StingRay

That should do the trick.



Level 7 vector is $7c.w which is valid for all 680x0 processors as long as the VBR is not moved.

Thanks, I think Leonard will do it this way (we had a little look at some other demos, to see what they were doing too)