Linux-USB testing with multiple levels of hubs
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- virtual root hub
1.- ADS external self-powered 4-port hub
2. - Cherry kbd/bus-powered hub (4 available ports) [kbd works]
3. - Entrega hub4U self-powered 4-port hub
4. - Aspire keyboard/hub (bus-powered; 1 available port) [kbd works]
5. - Ark (FTDI) external self-powered 7-port hub
then tested one at a time:
+ HID keyboard: works
+ Kye/Genius mouse: works
+ Zoom CPiA camera: not found
(control messages had CRC errors) [usb-uhci]
+ Y-E DATA floppy drive: works
+ Zip-100 drive: works
+ Compaq modem: works
Summary: I didn't experience any single-device problems with the
virtual root hub + 5 levels of external hubs EXCEPT with the
Zoom CPiA camera.
For the cpia camera, I removed external hub #5 (Ark) and
connected the cpia camera to a kbd/hub (bus-powered).
This made the hub driver unhappy (lots of "port 2 over-current
change") messages, so I also removed self-powered hub #4
(Aspire kbd/hub). After this, I could connect the camera and
detect it without errors.
Then I ran the cpia_test program that Alan posted a few weeks
ago. With usb-uhci, I get "usb-uhci.c: interrupt, status 2,
frame# nnnn" about every 400 video (capture) frames.
usb-uhci elapsed time for 1000 video capture frames: 1 min:08 sec.
No problems today running camstream + cpia + usb-uhci thru 3
external hubs.
With uhci, I didn't get any kind of error messages printed.
uhci elapsed time for 1000 video capture frames: 1 min:07 sec.
I ran camstream + cpia + uhci for a few minutes and camstream
seemed to hang up. I can't Close (via window mgr) it or kill it
(via kill -9). Repeated one more time. I can't tell if this
is a uhci problem or a camstream problem. Johannes, if you
want me to look further at this, how would you suggest that
I do it? I do have a CATC $$$ analyzer.
As Johannes pointed out a few days ago, to really be USB compliant,
the USB software needs to detect and prevent device connections
that are this deep in the topology, but that's low priority IMO.
Anyway, this isn't exhaustive testing, it's a sample. I'm
pleased by the hub-depth handling and most devices working
at any depth but not surprised since there shouldn't be
any differences except for some small timing changes.
~Randy
___________________________________________________
|Randy Dunlap Intel Corp., DAL Sr. SW Engr.|
|randy.dunlap.at.intel.com 503-696-2055|
|NOTE: Any views presented here are mine alone |
|and may not represent the views of my employer. |
|_________________________________________________|
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
