home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.programmer
- Path: sparky!uunet!think.com!sdd.hp.com!hp-col!fc.hp.com!koren
- From: koren@fc.hp.com (Steve Koren)
- Subject: Re: Low level context switch for tasks
- Sender: news@fc.hp.com (news daemon)
- Message-ID: <BxLz7I.CCI@fc.hp.com>
- Date: Thu, 12 Nov 1992 15:04:30 GMT
- References: <1992Nov12.120146.23099@rhrk.uni-kl.de>
- Organization: Hewlett-Packard Fort Collins Site
- X-Newsreader: Tin 1.1.3 PL5
- Lines: 20
-
- Martin Apel (SAGA) (apel@physik.uni-kl.de) wrote:
-
- > coprocessor context was saved for every task. Otherwise, if tasks
- > used inline fpu code, the registers wouldn't have been saved.
- > I think it's the same way in 2.0, as I'm using JM, which installs
- > such Switch/Launch vectors for cpu accounting.
-
- I'm trying to find another way to do this in JM other than taking over
- tc_Switch and tc_Launch. There looks to be some functions called
- "Switch()" and "Launch()" in exec. Rumor has it that you can
- SetFunction() these just like anything else, and hence do CPU accounting
- without having to take over entries in the task structure.
-
- Does anyone know anything about this, such as whether it works? I also
- can't find library offsets for Switch() and Launch() like I can for the
- other functions in exec.library. If I could setfunction these instead
- of what I'm doing now, I think I could make JM more inconspicuous to
- other programs.
-
- - steve
-