"Dunlap, Randy" wrote:
>
> Hi,
>
> I was looking over the USB-IF (Implementers Forum)
> compliance testing spec yesterday and immediately
> had several questions.
> [http://www.usb.org/developers/complian.html
> and .pdf files from that page; and
> http://www.usb.org/developers/complian_testing.html,
> especially "Interoperability Guidelines" from
> this page]
I was actually thing about writing a Linux (or at least operating system
independent version) of this.
> 1. Has anyone tested bus-powered hubs instead of
> self-powered hubs?
I have, although the lack of decent power management in Linux USB is a
problem. Bus powered generally sucks except in the "built into another
device" case. Device recognition and performance varied according to how
much power I was drawing at turn-on - it worked much better if I plugged
in individual devices after the hub was plugged in. This may have been
been a function of the crappy TI chipset that all of my hubs use.
> 2. Has anyone tested 5 levels ("tiers") of hubs
> and devices past the last level?
Not enough hubs - I can do root hub, and four lower level hubs (of which
only two can be self powered - one is inside my keyboard, and the other
is limited by a shortage of plug-packs. I have no problems with a five
metre cable between two self powered hubs, which is my normal
configuration.
> 3. What's the most USB devices that anyone has
> tested (at least connected) simultaneously?
My demo at the Canberra Linux User Group meeting showed:
DSBR100
two keyboards
three mice
Zip250
two self powered hubs
audio
HID gamepad
cpia camera
cy3640 thermometer
I have a Zip100 and a KLSI ethernet device to test as well. I am going
to try and borrow another KLSI device as well. I will try and get
another dual mode hub.
> There are lots of combinations that we could test.
Any requests?
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
