[geeks] USB reference?

[Jonathan Patschke]

On Sun, 5 Apr 2026, Mouse via geeks wrote:

>> That said, USB, as specified, is a lovely I/O system.
>
> Well, in some respects.  As I understand it, for example, there is no
> way for a device to interrupt the host

That's correct and intentional.  It makes USB problematic for anything
which needs predictable latency.  It only works well for storage because
computers have plenty of CPU cycles to waste on polling.  Some host
controllers can be set up for programmed I/O, which makes things easier,
but that's more often going to be in an embedded context rather than
general-purpose computer.

> For that matter, the ... borderline fraud, I would call it in my less
> charitable moods, of naming it what they did when it's actually
> neither universal nor a bus.

A charitable take on it is that it is serial, and was intended to be a
replacement for a great many other things--many of which were actual
buses.  But, yes, it's a complex network and gets much more complex in
3.0 and _vastly_ more complex in 4.0.

> Thank you very much!

You're welcome.  I hope it helps.  In case you found yourself needing to
interface with almost any recent PC connectivity thing, they almost
always boil down to USB or PCI with strange firmware and a purpose-built
connector.  Knowing the basics of USB is foundational anymore.

-- 
Jonathan Patschke
Austin, TX
USA