Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!iuvax!purdue!decwrl!ucbvax!LIVERPOOL.AC.UK!USQB015
From: USQB015@LIVERPOOL.AC.UK (Mark Powell)
Newsgroups: comp.sys.atari.st
Subject: Re: A REAL ramdisk
Message-ID: <8904301910.AA07292@ucbvax.Berkeley.EDU>
Date: 30 Apr 89 12:47:15 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 24

I`m not into GEM enough to know whether a ramdisk that grows and shrinks
in "normal" TPA area is possible, but I know that to create such a ram
disk that survives reset is not possible. This is because reset survivable
ramdisks grab a chunk of ram at the top of memory on boot up. They then
alter a few OS variables, such as top of physical memory, so that the OS
thinks it has less memory than it really has and then re boot so that
the OS will set itself up for this reduced amount of memory. So, this
chunk of memory is left alone at the top of memory. The ramdisk then uses
this memory to store it's data. This method is reset survivable because
this OS variable that was altered by the ramdisk controller is not
affected by a warm start. Thus, every time the ST is reset the OS always
thinks it has this reduced amount of memory and doesn't touch this "protected"
portion of memory that the ram disk is using, using only yhe memory below
the ram disk. Thus, to resize the ram disk this OS variable must be altered
and the system rebooted so that the OS knows the new size of the memory.
This more or less prohibits a re-sizeable, reset survivable ramdisk. Unless,
of course you don't mind the computer resetting every time you add or delete
something from the ram disk.

                         Mark Powell

ARPANet : Hmm.
JANET   : usqb015@uk.ac.liv.ibm
USENET  : mcvax!ukc!ibm.liv.ac.uk!usqb015