On Wed, Apr 9, 2008 at 12:12 PM, Stuart Henderson  wrote:

quoted text
> On 2008-04-09, syl  wrote:

>  > I'm developping a firmware for an usb device for my enterprise

>  > and I try to be compatible with OpenBSD.

>

>  great (:

:)

quoted text
>

>

>  > The plug seems to works fine. The attach works and attach ucom0

>  > at umodem and umodem at uhub.

>  >

>  > But when I'm trying to do this :

>  > echo aaab > /dev/ttyU0

>

>  This is normal, try cuaU0. Also you might find "cu -l /dev/cuaU0"

>  is useful for testing..

>

>

I've just try cuaU0 and the trace is nice and more understandable.
But I think I've always got a problem with the setlinecoding
When I add trace into the kernel I saw that the function

umodem_set_line_coding failed and return the value 0x0f.

In the usbdi.h the 0x0f value equals to the USBD_TIMEOUT

define.
So I think I made a bad response to the SET_LINE_CODING.
What I am doing exactly on device side:
DEVICE RECV 8 BITS (ENDPOINT0) : 0x21 0x20 0x00 0x00 0x00 0x00 0x07 0x00

DEVICE RECV 7 BITS (ENDPOINT0) : 0x80 0x25 0x00 0x00 0x00 0x00 0x08
The first receive is do because I intercept an interrupt that there

are 8 bits available

on the endpoint0 after i read directly 7 bits for getting the end of

this request.

Does I forget something ? Does I send some ack to the host, or data on the

interrupt endpoint ?
thanks in advance.
--

Gallon sylvestre

Rathaxes Core Developper / LSE researcher

kernel developer for adeneo and OpenBSD fan

http://devsyl.blogspot.com/ | www.rathaxes.eu