[RFC] [PATCH] usbatm.[ch]: multiple changes

castet.matthieu at free.fr castet.matthieu at free.fr
Thu Apr 7 04:50:40 EDT 2005


Selon Roman Kagan <rkagan at mail.ru>:

> On Wed, Apr 06, 2005 at 11:16:08PM +0200, matthieu castet wrote:
> > Roman Kagan wrote:
> > >http://benoit.papillault.free.fr/usbsnoop/
> >
> > It show me that even if I select bulk mode it still use iso !!!
> Didn't expect the windows driver to give this level of control to the
> user...  Maybe it just didn't take immediate effect, and the modem has
> to be replugged for the driver to change the transfer mode?
The control is hiden in a debug windows (alt+d).
I don't know if it take imediate effect, but when I apply the change, the driver
reboot the modem.
Also I unplug the modem : I didn't find an easy way the stop the trace...
Note I test what they call iso mode and bulk mode. There are another mode call
iso/bulk, but it give me a blue screen ;)

> > For information windows driver use 16 urb of 265 with 12 Packets.
> Is the number of urbs published by the driver, or you figure it out from
> the usbsnoop.log?  I don't seem to be able to do it with mine, can you
> please tell me how?
It is from usbsnoop.log.
I count the number of iso urb sumited before receiving completion.

> > And the status of urb is c0000005...
> Ups...  AFAICT from
> http://cvs.sourceforge.net/viewcvs.py/mingw/w32api/include/ddk/usbdi.h
> that means USBD_HALTED(status) is true, and the status is equal to
> USBD_STATUS_DEV_NOT_RESPONDING.  BTW are you talking about the urb
> status or the iso packet status?  The sniffer version I've got doesn't
> seem to print the urb status (or at least I don't know where to look for
> it).
It is the iso packet status
I have also ErrorCount = 0000000c
But that's when at the beginning and the size is 0.

At the end of my log I have ErrorCount = 0 and IsoPacket.Status = ffffffff and
Maybe this erros just tell that the modem is not ready.
Also when looking more carefully about the IsoPacket.Length, I saw they change
the size from 265 to 795.
I will try to do a more complete log.


PS : when looking the sagem site, I saw that they advice to use bulck mode for
via chip :

More information about the Usbatm mailing list