> Call me crazy, but just to be pedantic, run memtest86 on the machine. Then ...
Ran the full suite ... no problems like that. Big builds have never
caused trouble either, so I've not suspected the machine has any
particular problems that'd trigger this.
> At this point (in fact long before this) I would get out the o'scope
> and the PCI extender. Is the board locked up? The symptom here ...
I don't have those tools, unfortunately, or I would have done the same!
However, I can report some progress. I just reproduced a very
similar kernel hang using a slightly tweaked (add some delays)
usb-storage with "usb-ohci" and with "uhci". (I couldn't use the
"usb-uhci" driver -- the device wouldn't enumerate with it, kept
getting c/b timeouts on set_address.) Takes a little longer to show
up, which I'm assuming is related to the fact that the particular
I/O operation in question succeeded, rather than failing. (We're
in contact with the device manufacturer to find out why this is
failing at high speed, while succeeding at full speed.)
That suggests very strongly to me that the problem isn't my driver,
and it's time to use a more recent kernel, in the 2.4.5 series since
Matt mentioned that some usb-storage problems vanished in that
release. Here's crossing fingers that this is one of them!
- Dave
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
http://lists.sourceforge.net/lists/listinfo/linux-usb-devel
