home *** CD-ROM | disk | FTP | other *** search
- Path: comma.rhein.de!serpens!not-for-mail
- From: mlelstv@serpens.rhein.de (Michael van Elst)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: AddIntServer + VERTB strangeness
- Date: 29 Mar 1996 19:29:50 +0100
- Organization: dis-
- Message-ID: <4jha6u$a7v@serpens.rhein.de>
- References: <199603281321.NAA02157@mailhost.unibol.com>
- NNTP-Posting-Host: serpens.rhein.de
-
- John Girvin <jgirvin@bfs.unibol.com> writes:
-
- >On 20 Mar 1996 19:27:12 +0100, mlelstv@serpens.rhein.de (Michael van Elst):
- >:>No, that's not all. You use the functions in a wrong way.
-
- >My code does not care how the interrupt mechanism calls the interrupt code,
- >only that my handler gets called at the correct time and that my handler is
- >the only handler that gets called. Both these features are (allegedly)
- >available as OS features, so I use the OS to get them. Tell me what I am
- >doing wrong and Ill change it!
-
- The vblank interrupt is a server interrupt. You are just allowed to
- add your own code to the chain (at any position but you should be aware
- of a graphics.library bug) and you must not abort the chain.
-
- >:>>Do you believe that everything must be either 100% or 0% OS with no middle
- >:>>ground? I dont!
- >:>But obviously it has to be so.
-
- >Why obviously? Whats wrong with using the OS to allocate HW resources you
- >need (exitting nicely if not available) then hitting that hardware directly?
-
- Maybe I was misunderstood. Allocating HW resources and then hitting the
- hardware is "using the OS".
-
- >If you stick to the published OS and HW interfaces (which I do) then surely
- >there is no problem provided these interfaces dont change?
-
- The only problem is that of portability. If you poke the hardware then
- people must use this hardware. If you use just OS functions then this
- adds an abstraction layer that keeps the software happy even when the
- hardware changes.
-
- >:>Something in between just causes problems because
- >:>you have to rely on the particular behaviour of the implementation or
- >:>even of a particular machine configuration.
-
- >What specifics of my kickstart/machine am I relying on by using the OS
- >interrupts in this way?
-
- Using the interrupts in this way is not allowed and it doesn't work
- anyway :)
-
- But if it would work you would rely on this instead of the specification
- which forbids it.
-
- >Im relying only on official interfaces and their
- >officially described and officially "guaranteed" official behaviour.
-
- Sorry, no :-/
-
- >:>That's why the only
- >:>acceptable solution is to obey to the rules set by the OS specification.
-
- >Im trying to.
-
- I see this. I'm not flaming you but you simply made a mistake.
-
- >The original question was about the OS specification saying
- >one thing ([re?]set Z after interrupt handler code to disable others in
- >chain) and then behaving in another way.
-
- The OS specification says that you must not abort the chain with your
- handler. This is special for VERTB, it is allowed for PORTS and EXTER.
-
- Regards,
- --
- Michael van Elst
-
- Internet: mlelstv@serpens.rhein.de
- "A potential Snark may lurk in every tree."
-