home *** CD-ROM | disk | FTP | other *** search
- Path: oreig.uji.es!ii201
- From: ii201@rossegat.uji.es (Villellas Guillen Oscar)
- Newsgroups: comp.sys.amiga.programmer
- Subject: Re: resourcetracking/Kill on A4000
- Date: 9 Feb 1996 12:58:43 GMT
- Organization: Universitat Jaume I. Castell≤ de la Plana. Spain
- Distribution: world
- Message-ID: <4ffge3$g7i@oreig.uji.es>
- References: <4e0554$hro@merkurius.lu.se> <4e2cqe$1i2j@rs18.hrz.th-darmstadt.de> <4eamsf$h5t@oak70.doc.ic.ac.uk>
- NNTP-Posting-Host: @rossegat.uji.es
- X-Newsreader: TIN [version 1.2 PL2]
-
- In article <4e2cqe$1i2j@rs18.hrz.th-darmstadt.de>, meixner@rbg.informatik.th-darmstadt.de (Matthias Meixner) writes:
- >Resource-tracking is not possible with the current version
- >of the OS. The problem is resource trading:
-
- >task "a" allocatest memory and sends it to task "b"
- >task "b" can now do with it what it wants.
- >If you had resourcetracking, the memory would be
- >freed if task "a" terminates and whereas task "b"
- >would still use it, which is a good way to bring a
- >system down.
-
- If the app. is rightly coded I think that its quite
- possible.
-
- The resource tracker, would have to look at the memory
- allocation flags, if memory is being allocated with a
- special flag it should be treated as shared memory and not
- enter in the memory deallocation when the program that
- allocates it dies. I think this and memory protection was
- the purpose of the MEMF_PUBLIC flag.(which I think was stated
- to be used when
- you allocated memory that other task had to access to).
-
-
- --
- //\\\
- /o \\
- +-----oooO--(_)--Oooo-----+ _ _
- | Oscar Villellas Guillen | ////
- +-------------------------+ _ _ ////
- | | | \\\X///
- ooO Ooo \XXX/ Always...
- ii201@rossegat.uji.es Intuition rules!
-