Re: The behaviour of usbser.sys
- From: chris.aseltine@xxxxxxxxx
- Date: Mon, 02 Jul 2007 08:01:01 -0700
On Jul 1, 9:14 pm, Julian Zhang <daemon...@xxxxxxxxxxx> wrote:
chris.aselt...@xxxxxxxxx wrote:
You didn't post your exact SERIAL_STATE message, but I doubt it
matters.
Thank you for your answer.
My SERIAL_STATE is as follows:
bTxCarrier 1
bRxCarrier 1
(All other bits are set to zero.)
I suspect bRxCarrier (which is equivalent to DC signal of RS-232)
should not be 1.
That still isn't a SERIAL_STATE message. SERIAL_STATE is ten bytes in
length. Can you post all ten bytes?
If you are literally sending all zeroes except those two bits, that's
wrong.
If you can't open the port, I suspect you are not really handling the
CDC endpoint 0 zero requests correctly, as closing and opening the
port alone won't affect the bulk/interrupt behavior.
What's the function of zero-length requests? Where does related
specification located in the CDC spec. ?
Sorry, that's a typo. ("Endpoint 0 zero requests" -- should just have
said "endpoint zero requests" or "endpoint 0 requests".)
Meaning, SET_CONTROL_LINE_STATE, GET_LINE_CODING, SET_LINE_CODING, and
SET_BREAK (or whatever). I think you mentioned you already support
these.
I would take an analyzer trace to see what's going on with the bus
when you open the port. If you are opening the device in
Hyperterminal, you're going to see a steady stream of the above
requests as the application configures baud rate, etc.
If you are just STALLing these requests, or otherwise not handling
them correctly, you will see bad app behavior when you open the port.
.
- References:
- Re: The behaviour of usbser.sys
- From: Julian Zhang
- Re: The behaviour of usbser.sys
- Prev by Date: Re: Help! How are drivers for plug & play adapters removed?
- Next by Date: Re: Desktop Capture (Vista)
- Previous by thread: Re: The behaviour of usbser.sys
- Next by thread: Re: Desktop Capture (Vista)
- Index(es):
Relevant Pages
|
