From: Carl Fongheiser (cmf@ositos.UUCP)
Date: 04/25/93


From: cmf@ositos.UUCP (Carl Fongheiser)
Subject: Re: 16550 uarts and .99pl8
Date: Sun, 25 Apr 1993 17:11:18 GMT

In article <C5yv07.Eny@world.std.com> nigel%gamble@uunet.UU.NET writes:
>I don't have any problems with 0.99pl8 and my 16550A. But I recently
>read over in comp.sys.ibm.pc.hardware that the difference between a
>16550 and a 16550A is that the 16550 has a bug in the FIFO which can
>cause repeated characters. So the question is: has anyone who knows
>that they have a 16550A (and not a plain 16550) seen this repeated
>character problem?

The Linux serial driver doesn't turn on the FIFO for 16550's, only the
16550A's. National Semiconductor came up with an algorithm to reliably
distinguish between the two. I doubt many people have honest-to-goodness
16550's, anyway. They didn't last very long (due to the FIFO bug).
In any case, the Boca cards use a chip which is essentially 4 16550A's in
one chip. My Boca card was running with absolutely no problems under
0.99pl{4,5,6}; the problem only started cropping up with pl7.

Carl Fongheiser
ositos!cmf@vpnet.chi.il.us