Interrupts now are not what they were used to be. Even Legacy interrupts are not a dedicated line, but a message that simulates it. So you can have as many as you want. I think all those numbers you see under ISA are placeholders, not actual things.

MSI-X devices can use up to 2048 interrupts.. crazy

https://en.wikipedia.org/wiki/Message_Signale … nterrupts#MSI-X

Of course I know. Interrupts don't do anything close to regular ISA interrupts, but it's still funny to see 511 of them.

I'm just glad that interrupts are no longer an issue since the introduction of PCI and newer buses.
The worst part was in that 'transitional' period, when you had a 486, and wanted a fast 32-bit HDD controller and multi-IO card, a few sound cards, network card, a scanning device and stuff like that, and you just had to find a way to cram everything into that small IO/IRQ/DMA/memory map space.

Scali wrote:

I'm just glad that interrupts are no longer an issue since the introduction of PCI and newer buses.
The worst part was in that 'transitional' period, when you had a 486, and wanted a fast 32-bit HDD controller and multi-IO card, a few sound cards, network card, a scanning device and stuff like that, and you just had to find a way to cram everything into that small IO/IRQ/DMA/memory map space.

Oh yes, remember it well, DMA channels were an even greater challenge.
I used to make a "cheat sheet" listing devices by the least choices of resources first, in order to resolve conflicts without eternal jumper fiddling.

Favourite "solve" resulted in:
DMA 0 - Sound
DMA 1 - Network card
DMA 2 - Floppy
DMA 3 - Scanner card
May have been slightly different, but the day was saved by the sound being able to use DMA 0 ... my second cheat sheet rule, if only one device CAN use resource x, then that device gets it, though that pretty much paralleled assigning the one with the most choices last.