On Mon, May 22, 2000 at 07:27:47PM -0600, Vern Hoxie wrote:
> >> On Fri, May 19, 2000 at 12:00:21PM -0500, Al Borchers wrote:
> >>>
> >>> Our driver has a problem with select--select reports our device is ready
> >>> for writing, but write sometimes returns -EAGAIN.
>
> My immediate reaction is that select() should not be used on serial
> ports. There are other options using the basic serial driver without
> the additional overhead inherant in the more complex select()
> function.
I agree that select should not be used, but since this is the driver, it
has to be able to handle the fact that some strange program might want
to call select.
The question is about the n_tty.c code calling chars_in_buffer in the
driver, and doing something different based on the fact that the driver
returns 256 or a smaller number. This happens when select() is called,
and the current usb-serial drivers handle this differently (visor,
whiteheat, omninet, and generic do something different than the others,
and it looks like those are wrong.)
Did that make sense?
thanks,
greg k-h
greg@(kroah|wirex).com
p.s. thanks for maintaining the serial_suite HOWTO. I've used it loads
of times and it's full of good information.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
