What signals can cause a move instruction to stall indefinitely?

Initially I had only tried G0, but further experiments revealed it didn't matter if it was G0 or G1... though I'm not sure which I was using when I overrode the spindle-at-speed signal. I'd test it, but that's going to be a bit difficult for the time being as the spindle has stopped working completely.
The dead spindle is not a linuxcnc, it's hardware. the VFD shows the correct target frequency but has no measurable output. So that's another mystery to be solved, the only other clue is that a couple of times it started up and then died after a couple of seconds.

I wonder if this was the problem with the spindle-at-apeed signal too?

I don't think the two are connected. Previously while the spindle was working the spindle-at-speed pin worked correctly (assuming it is not supposed to be high when the spindle is off).
The way this has happened, and also the reports of issues with the previous vfd which we replaced (before my time working on the unit), make me think that it may be some sort of wiring issue.