On Fri, Jun 02, 2000 at 04:32:54PM +0000, Stefan Steuerwald wrote:
> > That you have encountered one of the ugliest errors on USB... The device
> is
> > simply not responding anymore (even no NAK). The "CRC/Timeout" should be
> > formatted as "crc/!!!TIMEOUT!!!".
>
> How do you conclude it is NOT a CRC error?
Hm, experience? All the crc/timeout messages I could inspect with the USB
analyzer were 'real' timeouts.
> I have a device exhibiting similar behaviour.
> Assuming it is a TIMEOUT: How long is the timeout? Does the USB spec define
> this value?
The value is defined somewhere in the spec (can't look it up in the moment)
and the HC itself checks this timeout. There's no way to adjust this.
> > I assume that the device's firmware gets so confused about the first bulk
> > message and the clear_halt, that it crashes. I know a similar behavior
> from
> > an Atmel hub, that crashes after a while when asked for string descriptors
> > it doesn't have. The CATC shows no unusual behavior from the HCD side.
>
> My device spits out numerous debugging messages on its slow RS232 interface.
> It does so during enumeration as well. This even prevented enumeration on
> a Windows 2000 host (UHCI), worked all right on Win 98 (UHCI).
> I will try to switch off all remaining debugging and see if the HCD
> "crc/!!!TIMEOUT!!!"
> goes away.
>
> Can the HCD distinguish between CRC and timeout errors?
> If so, we should introduce separate error codes.
No, the TD status flags have only one bit for both cases.
--
Georg Acher, [EMAIL PROTECTED]
http://www.in.tum.de/~acher/
"Oh no, not again !" The bowl of petunias
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
