home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!hellgate.utah.edu!cc.usu.edu!ivie
- From: ivie@cc.usu.edu (CP/M lives!)
- Newsgroups: comp.arch
- Subject: Re: PALcode instructions vs TRAP instructions
- Message-ID: <1992Aug16.195833.58188@cc.usu.edu>
- Date: 16 Aug 92 19:58:33 MDT
- References: <1992Aug10.211215.25582@nntpd.lkg.dec.com> <1992Aug12.094231.1767@uklirb.informatik.uni-kl.de> <1992Aug12.092211.58077@cc.usu.edu> <34303@cbmvax.commodore.com>
- Organization: Utah State University
- Lines: 21
-
- In article <34303@cbmvax.commodore.com>, jesup@cbmvax.commodore.com (Randell Jesup) writes:
- > ivie@cc.usu.edu (CP/M lives!) writes:
- >>Not really. The OS will believe the console code about how much memory is
- >>available. The console reports the memory size as smaller than physical by
- >>the amount of memory needed for the PALcode and the console. The OS will not
- >>try to use the memory thus reserved for PALcode.
- >
- > This would seem to be dangerous, because of wild pointers. Just
- > because the OS doesn't know about the memory doesn't mean that there won't
- > be a write to those locations. I do hope that PALcode fetches are checked
- > for illegal accesses.
-
- The problem with wild pointers should be solved as soon as the OS is up far
- enough to initialize the page tables. If the page tables don't include the
- console memory, the OS can't get at them.
-
- At least, that works on a VAX. I'm not sure whether the kernel runs mapped
- on an Alpha. Wouldn't work on a MIPS.
-
- Roger Ivie
- ivie@cc.usu.edu
-