ucom/uftdi dropping bytes, with debug logs

[Geoffrey Mainland]

M. Warner Losh wrote:
> In message: <e6vcnn$um$1 at sea.gmane.org>
>             Geoffrey Mainland <mainland at apeiron.net> writes:
> : I've been trying to debug a problem with a ftdi-based serial device
> : where bytes are being dropped. The device is a Moteiv tmote sky with a
> : 8U232BM chip. Reprogramming this device always fails due to dropped
> : bytes. I eventually set up a VMWare image (under Windows) to make
> : debugging a bit easier, as I was crashing my machine often, and lo and
> : behold the problem went away! It looks like VMWare's USB interface is
> : buffering the data, so the FreeBSD driver works just fine under VMWare!
> : 
> : I turned on debugging in the ucom and uftdi drivers. Logs of the
> : transfers under both VMWare and a native environment are posted at:
> : 
> : http://www.apeiron.net/~mainland/uftdi/vmware-log.txt.gz
> : 
> : and
> : 
> : http://www.apeiron.net/~mainland/uftdi/native-log.txt.gz
> : 
> : This is under a recent 6.1 stable (for both "machines").
> : 
> : As you can see, the VMWare transfer manages to read a lot more data...
> : 
> : How do I go about finding the real problem and fixing it? I know nothing
> : about the FreeBSD USB stack...
> 
> I have similar problems with FTDI, but every time that I think I've
> found one, I discover that I have a second 'tip' running that's eating
> some characters :-(  Are you sure that this isn't happening?  I know
> it is a "rookie" mistake, but I recently wasted a day on it, and I'm
> no rookie :-).
> 
> Anyway, I've also seen other weird behavior with my ftdi dongle, but
> I've never been able to reproduce it well enough to try to fix the
> problem...
> 
> Warner

Wow, thanks for the quick response!

Yes, I'm sure no second 'tip' is running. In any case, wouldn't the
reads from the second 'tip' still be logged? :)

The fact that this runs fine under VMWare makes me strongly suspect a
timing issue that is fixed by some sort of buffering at the VMWare
level. Where should I start to look to hack something in to test this?
The ucom driver seems to be setting up the read transfers that don't
complete on time.

Geoff