home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.lang.dylan
- Path: sparky!uunet!brunix!cs.brown.edu!mhw
- From: mhw@cs.brown.edu (Mark Weaver)
- Subject: Re: Is GC really even necessary??
- In-Reply-To: barmar@think.com's message of 20 Jan 1993 17:11:48 GMT
- Message-ID: <MHW.93Jan21005548@cslab0a.cs.brown.edu>
- Sender: news@cs.brown.edu
- Organization: Department of Computer Science, Brown University
- References: <9301192202.AA18001@gensym.com> <9301192325.AA02857@af9hp.us.oracle.com>
- <1jk14kINNeb@early-bird.think.com>
- Date: Thu, 21 Jan 1993 05:55:48 GMT
- Lines: 18
-
- In article <1jk14kINNeb@early-bird.think.com> barmar@think.com (Barry Margolin) writes:
- > In systems like Genera (the Symbolics Lisp Machine OS), these objects are
- > represented in OO fashion, with direct references to the data structures.
- > Therefore, the fact that they're collected in a linked list doesn't affect
- > the efficiency of accessing them. This type of solution, where pointers to
- > kernel objects are passed in from user code, only works well on systems
- > with no kernel protection (like Genera) or with capabilities; otherwise,
- > user code could synthesize pointers to kernel memory that it shouldn't have
- > access to.
-
- What is the problem with user code synthesizing pointers to kernel
- memory if there is hardware-level memory protection that prevents
- access to such memory?
- --
- --------------------------------------------------------------------
- Internet/CSnet: mhw@cs.brown.edu | Mark Weaver
- BITNET: mhw@browncs.bitnet | Box 2160, Brown University
- UUCP: uunet!brunix!mhw | Providence, RI 02912-2160
-