Path: utzoo!attcan!uunet!mcsun!sunic!chalmers!tekno.chalmers.se!cth_co
From: cth_co@tekno.chalmers.se (CHRISTER OLSSON)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Request for info on Disk Caches
Message-ID: <7261@tekno.chalmers.se>
Date: 6 Feb 90 10:19:36 GMT
References: <1414.25C69DAC@busker.FIDONET.ORG>
Organization: Chalmers Univ. of Technology, Gothenburg, Sweden
Lines: 20

In article <1414.25C69DAC@busker.FIDONET.ORG>, Howard.Spindel@p8.f14.n105.z1.FIDONET.ORG (Howard Spindel) writes:
>> In an article of <27 Jan 90 00:25:32 GMT>, jmerrill@jarthur.Claremont.EDU (Jason Merrill) writes:>> 
>>I've found that disk caches intefere with communications; I have a 19200

-- stuff deleted --
 
> This problem may be somewhat alleviated by using 16550 UART chips as
> they have internal buffering.  The problem does not exist with 386
> machines.

Wrong! At least one 386 from IMC has the problem. (It's 25Mhz 80386).
Replacing the 16450 to 16550 solves the problem.
 
I think the problem is most common if you have an RLL or other hd-controller
with own bios. The interrupts to the AT-bios, the hd-bios and to the 
cache-driver takes too long time and you lose chars.
(You transfer some file, the transferprogram will save a block to disk.
Use the dos-interrupt for disk transfer, the cache-driver gets the block to
store it on the cache-memory. The cache-routine goes to the HD-bios and the
HD-bios works. If that takes more than 520 us, you lose chars.