NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/sys/amiga/programm/16191

Internet Message Format | 1992-11-21 | 1.6 KB

Path: sparky!uunet!spool.mu.edu!agate!darkstar.UCSC.EDU!cats.ucsc.edu!davids
From: [e-mail] (Dave Schreiber)
Newsgroups: comp.sys.amiga.programmer
Subject: Re: physical memory protection with MMU
Date: 22 Nov 1992 05:15:15 GMT
Organization: University of California; Santa Cruz
Lines: 25
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]>
NNTP-Posting-Host: si.ucsc.edu


In article <[e-mail]> [e-mail] (W L Anderson) writes:
>Since in a protected memory environment, you are going to have to
>keep  track of what memory is allocated by which process, and the
>problem seems to be with memory that is being  used  to  pass  to
>other tasks is not always allocated as MEMF_PUBLIC, why not put a
>hook into the EXEC message passing code so that when  a  task  is
>about  to  pass  a message, you check to see if it's MEMF_PUBLIC.
>If it isn't, make it MEMF_PUBLIC.

The drawback is that the message could contain a pointer to some other
chunk of memory that isn't declared MEMF_PUBLIC.  While Exec could make
the message's memory MEMF_PUBLIC if needed, since it doesn't enforce a
particular structure upon messages, it wouldn't know that the message
contained a pointer and thus couldn't change the type of the second
memory block.  The process that received the message would then crash
when it tried to access that memory.

>Tom Hayko
>[e-mail]



--
Dave Schreiber  "Look.  Don't touch."  [e-mail] (until 6/20/93)