wine users - Nov 2009 - Real Serial port vs USB Serial adaptor difference?

When printing some labels with the Seiko Smart Label Printer software I've
come across a difference when using real serial ports vs a usb to serial
adaptor.
Printing labels with not much content is fine. 
But when printing labels with more and longer lines of address via a real serial
port only about half the label prints.
There is no obvious errors encountered and from the Queue manager of the SLP
software it appears it thinks the label has been printed.
If I print the same label via a usb to serial adaptor it does print fine.
I have tried adjusting the speed and the various bit settings and handshaking in
both Linux and in the SLP software on the real serial port with no improvement.
The same real serial ports work using the Linux SLAP software (though it is
painfully slow to print).
And the same real serial ports can also be made to work with Hylafax without
issue.

k4king ok thinking wine talks to serial ports basically the same no matter if
they are usb or real serial ports.

Wonder if that painfully slow is the sign of problem.   This could all be the
pure voodoo of kernel internal locking k4king.  Ie doing X alone does not cause
problems but doing X + Y hello gremlins.

http://blog.fenrus.org/?p=5  Running this that is in 2.6.32 kernel to come out
would be useful to get some idea where the wall it running into is.  Also 2.6.32
removes file-system locking that blocks all devices using a particular lock
while file-system is being written to and those devices block file-system
ability to write.     USB serial don't have that lock.   Could be a
operational clash.

I know 2.6.32 is not a stable kernel yet.  Lack of timechart makes tracing down
this problems hard.

oiaohm wrote:
> Also 2.6.32 removes file-system locking that blocks all devices using a
particular lock while file-system is being written to and those devices block
file-system ability to write.
I wonder of that will finally make Linux usable when handling any large files
(see this bug for explanation: http://bugzilla.kernel.org/show_bug.cgi?id=12309
)?