Greg:

On Mon, Apr 15, 2002 at 03:47:42PM -0700, Greg KH wrote:
> Um, you're missing a out endpoint, which is necessary for the generic
> usb-serial driver to write data to.  This is why writes will not work :)

Right.  D'oh!

Know what's wierd?  The previous cut-and-paste from usbview *did* have
an OUT endpoint.  After tinkering a bit, I've discovered that
sometimes it's there, sometimes it isn't.  Hmmm, that's more to look
into than I had a few minutes ago... thanks!  :^)

> Do you have a pointer to the arm-linux code on the device that
> implements the usb device side interface?

Not yet, I haven't written it.  The general idea is to continue using
usbserial, a'la the Visor driver.

I intend to use /dev/ttyUSB0 on the PC side, and mknod c 240 10
(SA1110 usb-char) on the arm-linux side.  At the application level, it
will (hopefully, critiques welcome) just look like a fast serial port,
which I *think* is how usbserial is intended to work.  That's how it
looks like it's working in visor.c, anyway.

Once I have that going, I'll figure out how to talk to the arm-linux
client from the Win32 world--- which I'm hoping won't require any more
work on the arm-linux side.  Not sure exactly how that's going to pan
out, I still have some research to do...

At any rate, if possible I'd like to test with minicom running at both
ends for now, until I'm sure that the low-level stuff is right.

Does this seem rational?



b.g.
-- 
Bill Gatliff
[EMAIL PROTECTED]

_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users

Reply via email to