IRQs and such (was Re: Serial interfaces)

On 7/27/10, Charlie Carothers <csquared3 at tx.rr.com> wrote: Can be done that way... In a sense, yes.... All the way back to 1970, DEC published lists in various system handbooks about their own breakdown of assigned CSRs (Control Status Register - typically the first I/O address that the card responds do, though there's no way to magically know how many addresses a card will answer at) an interrupt vectors for disk controllers, comms controllers, etc. For older devices, DEC would reserve up to four locations for a given interface, allowing for primary through quaternary cards to be installed on the same bus. Later, there was typically a primary address given, then a scheme to calculate a non-conflicting "floating" address, but by then folks were rarely installing four of anything except perhaps serial interfaces or MSCP disk interfaces. I can give an example of how this worked in practice by describing a 3rd-party device I used to make and maintain, the COMBOARD. It wasn't on DEC's list, so we picked an arbitrary CSR up in "unassigned" space. We gave our customers a list of eight preferred CSR addresses to use, though any address in the I/O page would work. The user set the CSR address on a 10-position DIP switch, then, once the machine was powered on and running, would load the driver and tell it, via setup files, what CSR the card(s) will respond at. Our product had a 16-bit double-one-way register we called "the window". If the PDP-11/VAX wrote a value, the 68000 on the card could read it, and if the 68000 wrote a value, the PDP-11/VAX could read it (a processor could never read back the value that it just wrote). One of the values passed during initialization was the vector to use (as stated in a config file that was written by the user, or by us as a factory default). The board would retain that value and drop it on the bus during an interrupt request. So for our product, the CSR was set by switches that were not meant to be changed, but the interrupt vector could be changed at every reboot. Real DEC devices did a number of different things, the older ones typically relying on jumpers or switches, the newer devices relying on some mechanism to assign or calculate the values dynamically. In practice, there weren't as many kinds of interfaces for the Unibus as there ended up being for the ISA bus, so it may sound like anarchy, but it wasn't. I think in 20-ish years of the hey-day of the Unibus, they only ever reassigned CSR addresses once or twice, but it was extremely unlikely that you'd have a particular obscure peripheral from 1970, literally, on the same bus as a stack of c. 1990 disk controllers and Ethernet interfaces. Electrically, they would work together, but logically, a user would have no need to use 20-year-old devices. The only "old" device we ever used on our VAX-11/750 was an LP11, an original line-printer interface. The rest of the cards in our machine were manufactured 10-15 years later. Some of our devices, then, had a DIP switch for CSR assignment, but not for vector assignment. Something older, like a PDP-11/34, was probably a different story. -ethan