home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!agate!stanford.edu!snorkelwacker.mit.edu!ai-lab!hal.gnu.ai.mit.edu!jimc
- From: jimc@hal.gnu.ai.mit.edu (James Carpenter)
- Newsgroups: comp.os.linux
- Subject: Re: Serial IRQ polling
- Date: 16 Dec 1992 03:26:54 GMT
- Organization: Massachusetts Institute of Technology
- Lines: 18
- Message-ID: <1gm7luINNcco@life.ai.mit.edu>
- NNTP-Posting-Host: hal.gnu.ai.mit.edu
-
- In article <ADAMS.92Dec16013812@PDV2.pdv2.fmr.maschinenbau.th-darmstadt.de> adams@pdv2.fmr.maschinenbau.th-darmstadt.de (Adams) writes:
- >There is no reasonable way to poll a device in a multi user or
- >realtime environment. Attaching to a timer
- >will lead to a desastrous performance loss, and hooking the
- >CPU will disable multitasking.
-
- It is interesting that Coherent could do this with no problem. There was a
- small decrease in speed but who cared... it ran unbelivably slow anyway.
-
- >If there would be enough interest, it would be neither too hard nor
- >too expensive to throw an expansion card together, handling 2-4 serial
- >ports without any intervention of CPU.....
-
- That sounds rather nice.
-
- -Jim Carpenter
- CARPENTER@Eisner.DECUS.Org
- jimc@gnu.ai.mit.edu
-