Article View: comp.sys.ibm.pc.hardware.comm
Article #152Re: File transfer errors with new MB's
From: Michael O'Reilly
Date: Sun, 12 Mar 1995 06:58
Date: Sun, 12 Mar 1995 06:58
48 lines
3283 bytes
3283 bytes
Tod Masters (masters@ozemail.com.au) wrote:
> Justin Hawkins <jhawkins@tardis.apana.org.au> wrote:
> >I recently upgraded to a DX2/66 (from a 386SX-25). This was a
> >motherboard only upgrade, so I still have the (crappy) old IO card
> >installed (with 8250 UARTS). In DOS I find that now file transfers
> >at any speed have errors, whereas before, I didn't. I tracked it
> >down, it turns out that whenever disk IO occured, the file transfer
> >would get an error. In OS/2 this does not occur. Downloading to a
> >RAM disk under DOS allows error free transfers.
> The 8250 is a pretty simple UART and doesn't do any buffering in
> hardware - [ .. ] The 16450 and 16550
> both have hardware receive and transmit buffers (I think the 16450
> has 8 character buffers - don't know about the 16550). This gives
> the software more time to do other things between interrupts.
The 16540 does no buffering. The 16550A is the chip that does
buffering. It has 16 character receive and transmit buffers.
> I can't explain why it doesn't happen under OS/2 - I can't believe OS/2 has
> better interrupt latency than DOS!
DOS doesn't have an interrupt latency. It doesn't handle serial
interrupts in any way shape or form. The comms program is what handles
serial interrupts, and that varies wildly.
OS/2 is very likely to have significantly better interrupt latency
> >What causes this? I suspect it has something to do with the way DOS
> >does it's disk IO. Is there any way of curing this under DOS, as my
> >friend won't be able to upgrade his OS for a while yet.
It's fairly likely. Again, it's usually not DOS, but the BIOS that
takes care of actually talking to the disk, and it's wildly variable
again. The question is if the BIOS turns off interrupts while reading
data from the disk. (This is fairly likely). Likely OS/2 does the
normally right thing ,and doesn't disable interrupts ('normally right'
as there are some pretty broken IDE drives out there...)
Possible the BIOS on the new motherboard disables interrupts where the
old didn't?
Michael.
--
Michael O'Reilly @ iiNet Technologies, Internet Service providers.
Voice (09) 307 1183, Fax (09) 307 8414. Email michael@iinet.com.au
GCS d? au- a- v* c++ UL++++ L+++ E po--(+) b+++ D++ h* r++ u+
e+ m+ s+++/--- !n h-- f? g+ w t-- y+
Technical Details
Message-ID:
<3ju61r$9cm@classic.iinet.com.au>Path:
arc1-ca.pugleaf.net!bga.com!news.sprintlink.net!cs.utexas.edu!swrinde!ihnp4.ucsd.edu!munnari.oz.au!news.uwa.edu.au!yarrina.connect.com.au!classic.iinet.com.au!not-for-mailReferences:
<jhawkins00442950114215637@tardis.apana.org.au> <950312$014113$1157masters@ozemail.com.au>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: 105ed0,d76712452d396d8f X-Google-Attributes: gid105ed0,public X-Google-Thread: fc9bd,d76712452d396d8f X-Google-Attributes: gidfc9bd,public X-Google-ArrivalTime: 1995-03-11 23:53:58 PST Path: arc1-ca.pugleaf.net!bga.com!news.sprintlink.net!cs.utexas.edu!swrinde!ihnp4.ucsd.edu!munnari.oz.au!news.uwa.edu.au!yarrina.connect.com.au!classic.iinet.com.au!not-for-mail From: michael@iinet.com.au (Michael O'Reilly) Newsgroups: comp.sys.ibm.pc.hardware.comm,aus.comms Subject: Re: File transfer errors with new MB's Followup-To: comp.sys.ibm.pc.hardware.comm,aus.comms Date: 12 Mar 1995 14:58:03 +0800 Organization: iiNet Technologies Lines: 48 Message-ID: <3ju61r$9cm@classic.iinet.com.au> References: <jhawkins00442950114215637@tardis.apana.org.au> <950312$014113$1157masters@ozemail.com.au> NNTP-Posting-Host: classic.iinet.com.au X-Newsreader: TIN [UNIX 1.3 941213BETA PL0]