home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.atari.st.tech
- Path: sparky!uunet!usc!sdd.hp.com!ux1.cso.uiuc.edu!news.iastate.edu!vincent1.iastate.edu!sourada
- From: sourada@iastate.edu (Steven D Ourada)
- Subject: Interrupt interference
- Message-ID: <sourada.714027685@vincent1.iastate.edu>
- Sender: news@news.iastate.edu (USENET News System)
- Organization: Iowa State University, Ames IA
- Date: Mon, 17 Aug 1992 05:01:25 GMT
- Lines: 28
-
- I'm writing a routine to hang off the 50Hz interrupt. This routine often
- takes as much as 2 ms (long, yes, but unavoidable). The problem is, this
- seems to interfere pretty badly with timing-intensive things like serial
- and disk access.
-
- I've tried the few things I could think of to alleviate this problem:
- flock is checked at the entrance to the routine, if it indicates the disk
- is in use, the routine is exited immediately. If not, the IPL is lowered
- to 3 to allow other interrupts and work proceeds.
-
- This, apparently, is not enough. Most of the time, disk accesses don't
- cause problems, although this does happen sometimes. Serial port access
- seems to bomb fairly frequently, even at lowly 2400 bps.
-
- Finally, the obvious question: What can I do to ensure that my routine
- gets the time it needs without bringing down the whole system? Is there
- a uniform way to see if some time-critical operations are going on? Is
- there a way to allow these to go on without my routine interfering too
- badly?
-
- Later,
- Steven Ourada
-
- --
- -----------------
- Steven Ourada Member of the Students for Electronic Freedom
- Ask me how Iowa State University is censoring my Usenet access!
- sourada@iastate.edu "can't casts no shadow" -- cummings
-