[Simh] More on Qbus VAX interrupts

[Bob Supnik]

The width of the vector read, and potential masking by the processor, is 
moot, because the Qbus spec defines vectors as 9b, with two low order 
0s. (The Unibus spec does as well.)

For hardware-generated vectors, the controllers provided a 6b or 7b 
jumper block (with two implicit low order 0s) to define the vector.

For dynamic vector devices, masking of the software-provided vector was 
up to the firmware. For example, the RQDX3 takes its vector 
right-justified (low order 0s shifted out), masks it to 7b, and then 
shifts it left by 2 to get a 9b vector with low order 0s.

The only exception I can find is the DELUA/DEQNA family, which has a 10b 
(<9:2>) vector field. PDP11 OS's typically started code at 1000(8), so a 
vector >= 1000 would cause real problems.

/Bob