home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.amiga.misc
- Path: watserv3.uwaterloo.ca!xenitec!focsys!wayne
- From: wayne@focus-systems.on.ca (Wayne Fisher)
- Subject: Re: OS features
- Message-ID: <DLnqBB.DuD@focus-systems.on.ca>
- Organization: Focus Automation Systems Inc.
- References: <4aj1tc$39r@candelo.dpie.gov.au> <13213430@sourcery.han.de> <wfblanDL60p0.D0y@netcom.com> <1058.6591T492T1743@cycor.ca>
- Date: Tue, 23 Jan 1996 23:23:35 GMT
-
- In article <1058.6591T492T1743@cycor.ca>, gcaine <gcaine@cycor.ca> wrote:
- >>I agree with Olaf here. The limitation that 'full' Memory Protection would
- >>bring and the performance hinderances are not what I want. Rather, a better
- >>way of singling out a bombed program and clearing it from memory without
- >>killing the whole system, and cleanly freeing up memory better than it does
- >>currently. I am quite happy with the OS as it stands, but don't take this
- >>to mean that I don't desire more. As I said, the basic sentiment of what
- >>Olaf said above is how I feel. Lets not worry about 'full' Memroy Protection
- >>at this point in the game... Maybe when we have a much faster CPU in the
- >>low-end machine (ie, PPC low-end), then any performance loss would be less
- >>important.
- >
- >If full memory protection is going to bring a performance hit, then I
- >don't agree that it should be added, even with the more powerful CPU's.
-
- I don't understan this line of thinking. With a MMU, you can add
- memory protection and virtual memory (not including paging to disk)
- with a minimal amount of overhead to the system. We're only talking
- single digit percentages here. And I hope you're not going to try and
- tell me that you can tell the difference between two machines where
- one is 2% faster than the other.
-
- The benifits of memory protection, (things like protecting one
- program's data from another and preventing writes to code, etc.) by
- far, out weigh the loss of performance. The stability of the machine
- will greatly improve. Not to mention the benifits to programmers when
- writing the code. At least IMHO.
-
- We wouldn't be able to use memory protection on older programs but all
- new ones should use it. Start with some protection and add more with
- each OS release. This would provide a set of steps to full protection.
- As the years go by, and as older programs get replaced with new ones,
- the percentage of programs that would break with protection,
- decreases.
-
- And while they're hooking in the MMU for memory protection, they can
- also add virtual memory. Note that I'm not including paging to disk,
- although that is a natural extension to it. By virtual memory, I mean
- giving all programs a contiguous address space independant of other
- programs. This allows the OS to map non-contiguous fragments of
- physical memory into a contiguous virtual block.
-
- Neither of these features affects the realtime nature of the machine.
-
- Now, we'd also require the ability for the machine to disable these
- features for machines without a MMU.
-
- Personally, I'd like to see AmigaOS 4.0 add the following:
-
- - protect code from writes.
- - protect unallocated memory from any accesses.
- - protect any ROMs from writes.
- - new programs' code and data would be protected from access by another
- process unless explicitly allowed by the program.
- - option to kill offending process or simply log the offence.
-
- >Keep the memory handling as it is, and then when there's faster CPU,
- >it will be able to do that much more. Why add something that's going
- >to change the whole concept of the Amiga?
-
- "much more"? I don't think a few percentages is "much more". And I
- don't see how it's going to change the whole concept of the Amiga. It
- just means that you can't pass pointers between processes and the
- machine becomes more stable.
-
- >Gary Caine Member: Team AMIGA
-
- Wayne
-
- --
- Wayne Fisher, Software Engineer | Focus Automation Systems, Inc.,
- wayne@focus-systems.on.ca | 3-554 Parkside Drive,
- #include <stdDisclaimer.h> | Waterloo, Ontario. N2L 5Z4
- We engineer innovative imaging solutions. | (519) 746-4918
-