Re: [coldsync-hackers] Re: Problem with ColdSync and Visor/USB

[Greg KH <greg@kroah.com>]

On Wed, Sep 20, 2000 at 11:13:07AM -0400, Andrew Arensburger wrote:
> On Wed, 20 Sep 2000, Nicolas Bouthors wrote:
> > Yet, there seems to be big problems. During the syncing, I have lots of
> > "Bad CRC" messages, and whatever I try, after a few seconds of transfer,
> > the Visor says "lost connection with the PC" (while coldsync doesn't
> > seem to notice anything). This makes it impossible to synchronise it
> > fully (although part of the data got through down and up from the Visor).
> > 
> > I tried both -S and -F options with no difference.
> > 
> > Is there a known bug relative to this behaviour ? Anything I could do ?

First off, if you are using a OHCI USB host controller, then all bets
are off.  Currently the Linux USB Visor driver doesn't work well with
OHCI at all, sorry.

If you are using UHCI, then read on...

> 	This is a known problem, but unfortunately, I don't know what the
> cause is. There are two things you can try:
> 	1) Rebuild with "-funsigned-char":
> 
> 	CFLAGS=-funsigned-char ./configure      (sh, bash)
> 	env CFLAGS=-funsigned-char ./configure  (csh, tcsh)
> 
> This seems to get rid of some compile-time warnings as to whether
> "char" is "unsigned char" or "signed char".
> 	2) Make sure the speed is correct. If you run
> 		coldsync -dsync:3
> when it establishes a connection, it will print something like the
> following:
> 
> 	===== Got a wakeup packet
> 	pconn->speed == 38400
> 	-> Setting speed to 38400 (38400)
> 
> Make sure that the two speeds match. The first line is what the Visor
> wants; the second line says what ColdSync is setting the speed to, as
> set by the .coldsyncrc .

The USB Visor driver doesn't care about the speed settings.  It just
merrily ignores them, as they really don't pertain to USB, so I don't
think that this is the problem.

> 	I think this is fairly consistent. That is, if a packet is cut off
> this way, then ColdSync won't acknowledge it, and the Visor will resend
> it. The second, third, etc. packets will also be truncated the same
> way. This leads me to believe that this is a software problem, and not
> line noise.

I've seen this.  It's as if coldsync is reading the same buffer again
and again from the tty driver.  It's strange, and I don't know what's up
with it.

> 	This could be a bug in ColdSync, but it could also be a bug in
> whatever it is that does the USB-to-serial translation (sorry, I can't
> seem to get to usbvisor.sourceforge.net to get the FAQ). Or it could be
> both: I've seen similar problems under Linux with a plain old serial Palm
> V.

Sorry about the web page being down, the maintainer of it has lost his
previous host, and is looking for a new one (I know...don't ask me, I
don't run it...)

> 	The best way of figuring this out would be to monitor what is
> coming in through the USB port and compare that to what is coming out of
> the pseudo-serial port, and also to what ColdSync reads (at the SLP level,
> in "libpconn/slp.c").
> 	If /dev/usb0 and /dev/usbSerial0 (or whatever) don't agree, then
> there's a problem with the USB-to-serial translation. If /dev/usbSerial0
> and ColdSync don't agree, then there's a problem in ColdSync.

(it's usually called /dev/ttyUSB1 but you can name it whatever you want
with mknod)

And if you see that there is a problem in the usb-to-serial translation,
_please_ let me know.

Turning on the debugging option in the usb-serial drivers is a good way
to see what it is receiving and sending out the USB port.

thanks,

greg k-h
(Linux USB Visor author...)

-- 
greg@(kroah|wirex).com

--
This message was sent through the coldsync-hackers mailing list.  To remove
yourself from this mailing list, send a message to majordomo@thedotin.net
with the words "unsubscribe coldsync-hackers" in the message body.  For more
information on Coldsync, send mail to coldsync-hackers-owner@thedotin.net.