[ problems with  com.c driver on 486, fifoless UARTs, PPP] 

>Also, I have only noticed this problem since updating from 1.0A to 1.2
>and 1.2B.  The rate of input silo overflows has increased by a lot as
>a result of this upgrade, too.


This is more-or-less an i386 problem (note the Reply-To:), but:

If you're using 1.2B, what do the vmstat -i counters say?
If you're getting "held" interrupts, it's possible that you're
dropping characters.  Or, perhaps, getting gobs of stray  interrupts
from the other UART.  (This  happens rarely with UARTS, more so with
internal modems, I believe.)

>From you description, I can't tell if this happened with 1.2 or
not; you're certainly claiming it didn't happen with 1.2.

The other thing to try is to apply Onno van der Linden's patches
to the i386 serial driver, which  reportedly fix many problems...

Happy Xmas to those still hacking!