[LAU] Analysing (debugging) USB port performance

[Jacek Konieczny]

On 23/09/2019 13.14, Lorenzo Sutton wrote:
> 2. Is there a way to change (e.g. improve the not-so-good USB port
> performance) OS/software wise, or is this usually hard-wired in notebooks?

There are ways to improve things a bit, but only so much can be done in
software, because part of that is, indeed, hard-wired in notebooks.

Such a computer may have more than one USB bus hardware, and the
external ports are only part of the things connected to those buses.
Webcam, touchpad, smartcard reader and other devices can be connected
internally via USB.  There may also be internal HUBs adding extra layer.

Some of the external USB ports share bus (and its IRQ) or internal port
with some of those devices and possibly other external ports.

For real-time audio the best choice would be a dedicated USB port on a
separate bus, not shared with anything else, but thing might be not
available.

And when a bus, or internal port is shared with a different device, the
kind of the device also matters. Some play nicer with audio devices,
some are worse.

I guess one can define all the rules how to choose the optimal port, but
simple trial and error is much more practical.

>   2a. Are IRQs relevant on laptops and if so can a whole USB port (or
> the device attached to it) be optimised from the OS?

irqbalance, rtirq, some kernel boot options and maybe other tools may be
used to optimize USB IRQ handling. This might be quite complex and still
will be limited by the hardware configuration.

Jacek