Subject: Re: 230400 on com.c?
To: None <current-users@netbsd.org>
From: Herb Peyerl <hpeyerl@beer.org>
List: current-users
Date: 05/02/2002 08:08:34
Ben Harris <bjh21@netbsd.org>  wrote:
 > In article <200205021321.g42DLZj22729@grok.beer.org> you write:
 > >I have a piece of hardware that seems to be insisting I speak to it at
 > >230400.  I'm being annoyed by the vendor who claims that it just works
 > >under 'linux' if you do some stupid setserial thing.  It appears there's
 > >some hack under linux to set the 16550 at 230400.  In our com.c, the
 > >comspeed() function won't let it go to 230400 since the percentage of
 > >error is too high...
 > 
 > How about picking the nearest speed the 16550 can do to 230400 and asking
 > for that (i.e. whatever comspeed would have chosen but for the error)?

That would be 115200 but the device doesn't respond to that.

The more I think about it, the more confused this vendor is.  The card
is a wireless CDMA 1xRTT/IS95B modem. A pcmcia card with a 16550 on the
bus.  I'm almost positive the speed shouldn't matter because the other
end of the 550 (rxd/txd) probably doesn't even exist.  They've probably
only implemented the bus-half of the 16550 and the FIFO just drops 
data into and receives data from the CDMA controller.  I should be able to
talk to it at 300baud and still get full throughput.

I just need to find the right clue to speak to.

Thanks.